From fa73055b8442c97b3ba7cd0aa57cd2ad32124201 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 7 Mar 2018 17:13:03 +0000 Subject: drm/i915: Only prune fences after wait-for-all Currently, we only allow ourselves to prune the fences so long as all the waits completed (i.e. all the fences we checked were signaled), and that the reservation snapshot did not change across the wait. However, if we only waited for a subset of the reservation object, i.e. just waiting for the last writer to complete as opposed to all readers as well, then we would erroneously conclude we could prune the fences as indeed although all of our waits were successful, they did not represent the totality of the reservation object. v2: We only need to check the shared fences due to construction (i.e. all of the shared fences will be later than the exclusive fence, if any). Fixes: e54ca9774777 ("drm/i915: Remove completed fences after a wait") Signed-off-by: Chris Wilson Cc: Joonas Lahtinen Cc: Matthew Auld Reviewed-by: Matthew Auld Link: https://patchwork.freedesktop.org/patch/msgid/20180307171303.29466-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_gem.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index a5bd073..ab88ca5 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -433,20 +433,28 @@ i915_gem_object_wait_reservation(struct reservation_object *resv, dma_fence_put(shared[i]); kfree(shared); + /* + * If both shared fences and an exclusive fence exist, + * then by construction the shared fences must be later + * than the exclusive fence. If we successfully wait for + * all the shared fences, we know that the exclusive fence + * must all be signaled. If all the shared fences are + * signaled, we can prune the array and recover the + * floating references on the fences/requests. + */ prune_fences = count && timeout >= 0; } else { excl = reservation_object_get_excl_rcu(resv); } - if (excl && timeout >= 0) { + if (excl && timeout >= 0) timeout = i915_gem_object_wait_fence(excl, flags, timeout, rps_client); - prune_fences = timeout >= 0; - } dma_fence_put(excl); - /* Oportunistically prune the fences iff we know they have *all* been + /* + * Opportunistically prune the fences iff we know they have *all* been * signaled and that the reservation object has not been changed (i.e. * no new fences have been added). */ -- cgit v1.1 From 033b7a230cfa759d1b582fa46bf2cd54db406cf3 Mon Sep 17 00:00:00 2001 From: Maarten Lankhorst Date: Thu, 8 Mar 2018 13:02:02 +0100 Subject: drm/i915: Handle pipe CRC around enabling/disabling pipe. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This will get rid of the following error: [ 74.730271] WARNING: CPU: 4 PID: 0 at drivers/gpu/drm/drm_vblank.c:614 drm_calc_vbltimestamp_from_scanoutpos+0x13e/0x2f0 [ 74.730311] Modules linked in: vgem snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic i915 x86_pkg_temp_thermal intel_powerclamp coretemp snd_hda_intel crct10dif_pclmul snd_hda_codec crc32_pclmul snd_hwdep broadcom ghash_clmulni_intel snd_hda_core bcm_phy_lib snd_pcm tg3 lpc_ich mei_me mei prime_numbers [ 74.730353] CPU: 4 PID: 0 Comm: swapper/4 Tainted: G U 4.16.0-rc2-CI-CI_DRM_3822+ #1 [ 74.730355] Hardware name: Dell Inc. XPS 8300 /0Y2MRG, BIOS A06 10/17/2011 [ 74.730359] RIP: 0010:drm_calc_vbltimestamp_from_scanoutpos+0x13e/0x2f0 [ 74.730361] RSP: 0018:ffff88022fb03d10 EFLAGS: 00010086 [ 74.730365] RAX: ffffffffa0291d20 RBX: ffff88021a180000 RCX: 0000000000000001 [ 74.730367] RDX: ffffffff820e7db8 RSI: 0000000000000001 RDI: ffffffff82068cea [ 74.730369] RBP: ffff88022fb03d70 R08: 0000000000000000 R09: ffffffff815d26d0 [ 74.730371] R10: 0000000000000000 R11: ffffffffa0161ca0 R12: 0000000000000001 [ 74.730373] R13: ffff880212448008 R14: ffff880212448330 R15: 0000000000000000 [ 74.730376] FS: 0000000000000000(0000) GS:ffff88022fb00000(0000) knlGS:0000000000000000 [ 74.730378] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 74.730380] CR2: 000055edcbec9000 CR3: 0000000002210001 CR4: 00000000000606e0 [ 74.730382] Call Trace: [ 74.730385] [ 74.730397] drm_get_last_vbltimestamp+0x36/0x50 [ 74.730401] drm_update_vblank_count+0x64/0x240 [ 74.730409] drm_crtc_accurate_vblank_count+0x41/0x90 [ 74.730453] display_pipe_crc_irq_handler+0x176/0x220 [i915] [ 74.730497] i9xx_pipe_crc_irq_handler+0xfe/0x150 [i915] [ 74.730537] ironlake_irq_handler+0x618/0xa30 [i915] [ 74.730548] __handle_irq_event_percpu+0x3c/0x340 [ 74.730556] handle_irq_event_percpu+0x1b/0x50 [ 74.730561] handle_irq_event+0x2f/0x50 [ 74.730566] handle_edge_irq+0xe4/0x1b0 [ 74.730572] handle_irq+0x11/0x20 [ 74.730576] do_IRQ+0x5e/0x120 [ 74.730584] common_interrupt+0x84/0x84 [ 74.730586] [ 74.730591] RIP: 0010:cpuidle_enter_state+0xaa/0x350 [ 74.730593] RSP: 0018:ffffc9000008beb8 EFLAGS: 00000212 ORIG_RAX: ffffffffffffffde [ 74.730597] RAX: ffff880226b80040 RBX: 000000000031fc3e RCX: 0000000000000001 [ 74.730599] RDX: 0000000000000000 RSI: ffffffff8210fb59 RDI: ffffffff820c02e7 [ 74.730601] RBP: 0000000000000004 R08: 00000000000040af R09: 0000000000000018 [ 74.730603] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000004 [ 74.730606] R13: ffffe8ffffd00430 R14: 0000001166120bf4 R15: ffffffff82294460 [ 74.730621] ? cpuidle_enter_state+0xa6/0x350 [ 74.730629] do_idle+0x188/0x1d0 [ 74.730636] cpu_startup_entry+0x14/0x20 [ 74.730641] start_secondary+0x129/0x160 [ 74.730646] secondary_startup_64+0xa5/0xb0 [ 74.730660] Code: e1 48 c7 c2 b8 7d 0e 82 be 01 00 00 00 48 c7 c7 ea 8c 06 82 e8 64 ec ff ff 48 8b 83 c8 07 00 00 48 83 78 28 00 0f 84 e2 fe ff ff <0f> 0b 45 31 ed e9 db fe ff ff 41 b8 d3 4d 62 10 89 c8 6a 03 41 [ 74.730754] ---[ end trace 14b1345705b68565 ]--- Changes since v1: - Don't try to apply CRC workaround when enabling pipe, it should already be enabled. Changes since v2: - Make crc functions for !DEBUGFS case inline. - Pass intel_crtc to crc functions. - Add comments to callsites. Changes since v3: - Cache selected source to pipe_crc->source. - Set pipe_crc->skipped to MIN_INT during disable to close a race condition. Changes since v4: - Handle fallout from setting pipe_crc->source in irq handler. Cc: Marta Löfstedt Reported-by: Marta Löfstedt Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105185 Signed-off-by: Maarten Lankhorst Link: https://patchwork.freedesktop.org/patch/msgid/20180308120202.52446-1-maarten.lankhorst@linux.intel.com Reviewed-by: Ville Syrjälä --- drivers/gpu/drm/i915/i915_irq.c | 4 +-- drivers/gpu/drm/i915/intel_display.c | 10 +++++++ drivers/gpu/drm/i915/intel_drv.h | 9 ++++++ drivers/gpu/drm/i915/intel_pipe_crc.c | 53 ++++++++++++++++++++++++++++++----- 4 files changed, 67 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index 633c187..babf81c 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -1627,7 +1627,7 @@ static void display_pipe_crc_irq_handler(struct drm_i915_private *dev_priv, int head, tail; spin_lock(&pipe_crc->lock); - if (pipe_crc->source) { + if (pipe_crc->source && !crtc->base.crc.opened) { if (!pipe_crc->entries) { spin_unlock(&pipe_crc->lock); DRM_DEBUG_KMS("spurious interrupt\n"); @@ -1667,7 +1667,7 @@ static void display_pipe_crc_irq_handler(struct drm_i915_private *dev_priv, * On GEN8+ sometimes the second CRC is bonkers as well, so * don't trust that one either. */ - if (pipe_crc->skipped == 0 || + if (pipe_crc->skipped <= 0 || (INTEL_GEN(dev_priv) >= 8 && pipe_crc->skipped == 1)) { pipe_crc->skipped++; spin_unlock(&pipe_crc->lock); diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index ceed082..f424fff 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -12145,6 +12145,9 @@ static void intel_update_crtc(struct drm_crtc *crtc, if (modeset) { update_scanline_offset(intel_crtc); dev_priv->display.crtc_enable(pipe_config, state); + + /* vblanks work again, re-enable pipe CRC. */ + intel_crtc_enable_pipe_crc(intel_crtc); } else { intel_pre_plane_update(to_intel_crtc_state(old_crtc_state), pipe_config); @@ -12325,6 +12328,13 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state) if (old_crtc_state->active) { intel_crtc_disable_planes(crtc, old_crtc_state->plane_mask); + + /* + * We need to disable pipe CRC before disabling the pipe, + * or we race against vblank off. + */ + intel_crtc_disable_pipe_crc(intel_crtc); + dev_priv->display.crtc_disable(to_intel_crtc_state(old_crtc_state), state); intel_crtc->active = false; intel_fbc_disable(intel_crtc); diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 37d5412..83e5ca8 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -2136,8 +2136,17 @@ int intel_pipe_crc_create(struct drm_minor *minor); #ifdef CONFIG_DEBUG_FS int intel_crtc_set_crc_source(struct drm_crtc *crtc, const char *source_name, size_t *values_cnt); +void intel_crtc_disable_pipe_crc(struct intel_crtc *crtc); +void intel_crtc_enable_pipe_crc(struct intel_crtc *crtc); #else #define intel_crtc_set_crc_source NULL +static inline void intel_crtc_disable_pipe_crc(struct intel_crtc *crtc) +{ +} + +static inline void intel_crtc_enable_pipe_crc(struct intel_crtc *crtc) +{ +} #endif extern const struct file_operations i915_display_crc_ctl_fops; #endif /* __INTEL_DRV_H__ */ diff --git a/drivers/gpu/drm/i915/intel_pipe_crc.c b/drivers/gpu/drm/i915/intel_pipe_crc.c index 1f5cd57..4f367c1 100644 --- a/drivers/gpu/drm/i915/intel_pipe_crc.c +++ b/drivers/gpu/drm/i915/intel_pipe_crc.c @@ -569,7 +569,8 @@ unlock: static int ivb_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, enum pipe pipe, enum intel_pipe_crc_source *source, - uint32_t *val) + uint32_t *val, + bool set_wa) { if (*source == INTEL_PIPE_CRC_SOURCE_AUTO) *source = INTEL_PIPE_CRC_SOURCE_PF; @@ -582,7 +583,7 @@ static int ivb_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, *val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_SPRITE_IVB; break; case INTEL_PIPE_CRC_SOURCE_PF: - if ((IS_HASWELL(dev_priv) || + if (set_wa && (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) && pipe == PIPE_A) hsw_pipe_A_crc_wa(dev_priv, true); @@ -600,7 +601,8 @@ static int ivb_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, static int get_new_crc_ctl_reg(struct drm_i915_private *dev_priv, enum pipe pipe, - enum intel_pipe_crc_source *source, u32 *val) + enum intel_pipe_crc_source *source, u32 *val, + bool set_wa) { if (IS_GEN2(dev_priv)) return i8xx_pipe_crc_ctl_reg(source, val); @@ -611,7 +613,7 @@ static int get_new_crc_ctl_reg(struct drm_i915_private *dev_priv, else if (IS_GEN5(dev_priv) || IS_GEN6(dev_priv)) return ilk_pipe_crc_ctl_reg(source, val); else - return ivb_pipe_crc_ctl_reg(dev_priv, pipe, source, val); + return ivb_pipe_crc_ctl_reg(dev_priv, pipe, source, val, set_wa); } static int pipe_crc_set_source(struct drm_i915_private *dev_priv, @@ -636,7 +638,7 @@ static int pipe_crc_set_source(struct drm_i915_private *dev_priv, return -EIO; } - ret = get_new_crc_ctl_reg(dev_priv, pipe, &source, &val); + ret = get_new_crc_ctl_reg(dev_priv, pipe, &source, &val, true); if (ret != 0) goto out; @@ -916,7 +918,7 @@ int intel_pipe_crc_create(struct drm_minor *minor) int intel_crtc_set_crc_source(struct drm_crtc *crtc, const char *source_name, size_t *values_cnt) { - struct drm_i915_private *dev_priv = crtc->dev->dev_private; + struct drm_i915_private *dev_priv = to_i915(crtc->dev); struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[crtc->index]; enum intel_display_power_domain power_domain; enum intel_pipe_crc_source source; @@ -934,10 +936,11 @@ int intel_crtc_set_crc_source(struct drm_crtc *crtc, const char *source_name, return -EIO; } - ret = get_new_crc_ctl_reg(dev_priv, crtc->index, &source, &val); + ret = get_new_crc_ctl_reg(dev_priv, crtc->index, &source, &val, true); if (ret != 0) goto out; + pipe_crc->source = source; I915_WRITE(PIPE_CRC_CTL(crtc->index), val); POSTING_READ(PIPE_CRC_CTL(crtc->index)); @@ -959,3 +962,39 @@ out: return ret; } + +void intel_crtc_enable_pipe_crc(struct intel_crtc *intel_crtc) +{ + struct drm_crtc *crtc = &intel_crtc->base; + struct drm_i915_private *dev_priv = to_i915(crtc->dev); + struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[crtc->index]; + u32 val = 0; + + if (!crtc->crc.opened) + return; + + if (get_new_crc_ctl_reg(dev_priv, crtc->index, &pipe_crc->source, &val, false) < 0) + return; + + /* Don't need pipe_crc->lock here, IRQs are not generated. */ + pipe_crc->skipped = 0; + + I915_WRITE(PIPE_CRC_CTL(crtc->index), val); + POSTING_READ(PIPE_CRC_CTL(crtc->index)); +} + +void intel_crtc_disable_pipe_crc(struct intel_crtc *intel_crtc) +{ + struct drm_crtc *crtc = &intel_crtc->base; + struct drm_i915_private *dev_priv = to_i915(crtc->dev); + struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[crtc->index]; + + /* Swallow crc's until we stop generating them. */ + spin_lock_irq(&pipe_crc->lock); + pipe_crc->skipped = INT_MIN; + spin_unlock_irq(&pipe_crc->lock); + + I915_WRITE(PIPE_CRC_CTL(crtc->index), 0); + POSTING_READ(PIPE_CRC_CTL(crtc->index)); + synchronize_irq(dev_priv->drm.irq); +} -- cgit v1.1 From 59cd31f177b34deb834a5c97478502741be1cf2e Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 8 Mar 2018 14:26:47 +0000 Subject: drm/i915: Kick the rps worker when changing the boost frequency The boost frequency is only applied from the RPS worker while someone is waiting on a request and requested a boost. As such, when the user wishes to change the frequency, we have to kick the worker in order to re-evaluate whether to apply the boost frequency. v2: Check num_waiters to decide if we should kick the worker to handle boosting. Fixes: 29ecd78d3b79 ("drm/i915: Define a separate variable and control for RPS waitboost frequency") Signed-off-by: Chris Wilson Cc: Mika Kuoppala Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180308142648.4016-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_sysfs.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c index b33d215..e5e6f6b 100644 --- a/drivers/gpu/drm/i915/i915_sysfs.c +++ b/drivers/gpu/drm/i915/i915_sysfs.c @@ -304,8 +304,9 @@ static ssize_t gt_boost_freq_mhz_store(struct device *kdev, { struct drm_i915_private *dev_priv = kdev_minor_to_i915(kdev); struct intel_rps *rps = &dev_priv->gt_pm.rps; - u32 val; + bool boost = false; ssize_t ret; + u32 val; ret = kstrtou32(buf, 0, &val); if (ret) @@ -317,8 +318,13 @@ static ssize_t gt_boost_freq_mhz_store(struct device *kdev, return -EINVAL; mutex_lock(&dev_priv->pcu_lock); - rps->boost_freq = val; + if (val != rps->boost_freq) { + rps->boost_freq = val; + boost = atomic_read(&rps->num_waiters); + } mutex_unlock(&dev_priv->pcu_lock); + if (boost) + schedule_work(&rps->work); return count; } -- cgit v1.1 From d586b5f4cf0792f69644d1aea171f82d029fb5ed Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 8 Mar 2018 14:26:48 +0000 Subject: drm/i915: Index the ring frequency table by HW frequency range When reporting the frequency table stored in the punit, report the full range and not just the user restricted frequency range. In the process keep the code to set the frequency table and read it the same. v3: As we haven't separated the sb_lock from the pcu_lock yet, there's a cycle between the pcu_lock and intel_runtime_pm_get. References: f936ec34dea8 ("drm/i915/skl: Updated the i915_ring_freq_table debugfs function") Signed-off-by: Chris Wilson Cc: Mika Kuoppala Reviewed-by: Mika Kuoppala #v1 Link: https://patchwork.freedesktop.org/patch/msgid/20180308142648.4016-2-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_debugfs.c | 13 ++++++------- drivers/gpu/drm/i915/intel_pm.c | 9 ++++----- 2 files changed, 10 insertions(+), 12 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 89f7ff2..d8bc1bb 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -1796,9 +1796,9 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused) { struct drm_i915_private *dev_priv = node_to_i915(m->private); struct intel_rps *rps = &dev_priv->gt_pm.rps; - int ret = 0; - int gpu_freq, ia_freq; unsigned int max_gpu_freq, min_gpu_freq; + int gpu_freq, ia_freq; + int ret; if (!HAS_LLC(dev_priv)) return -ENODEV; @@ -1809,13 +1809,12 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused) if (ret) goto out; + min_gpu_freq = rps->min_freq; + max_gpu_freq = rps->max_freq; if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { /* Convert GT frequency to 50 HZ units */ - min_gpu_freq = rps->min_freq_softlimit / GEN9_FREQ_SCALER; - max_gpu_freq = rps->max_freq_softlimit / GEN9_FREQ_SCALER; - } else { - min_gpu_freq = rps->min_freq_softlimit; - max_gpu_freq = rps->max_freq_softlimit; + min_gpu_freq /= GEN9_FREQ_SCALER; + max_gpu_freq /= GEN9_FREQ_SCALER; } seq_puts(m, "GPU freq (MHz)\tEffective CPU freq (MHz)\tEffective Ring freq (MHz)\n"); diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index b8da4dc..dd5ddb7 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -6918,13 +6918,12 @@ static void gen6_update_ring_freq(struct drm_i915_private *dev_priv) /* convert DDR frequency from units of 266.6MHz to bandwidth */ min_ring_freq = mult_frac(min_ring_freq, 8, 3); + min_gpu_freq = rps->min_freq; + max_gpu_freq = rps->max_freq; if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { /* Convert GT frequency to 50 HZ units */ - min_gpu_freq = rps->min_freq / GEN9_FREQ_SCALER; - max_gpu_freq = rps->max_freq / GEN9_FREQ_SCALER; - } else { - min_gpu_freq = rps->min_freq; - max_gpu_freq = rps->max_freq; + min_gpu_freq /= GEN9_FREQ_SCALER; + max_gpu_freq /= GEN9_FREQ_SCALER; } /* -- cgit v1.1 From 51f6b0f99cab765477a636443ce63295b76b9bb4 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 9 Mar 2018 01:08:08 +0000 Subject: drm/i915: Push irq_shift from gen8_cs_irq_handler() to caller Originally we were inlining gen8_cs_irq_handler() and so expected the compiler to constant-fold away the irq_shift (so we had hardcoded it as opposed to use engine->irq_shift). However, we dropped the inline given the proliferation of gen8_cs_irq_handler()s. If we pull the shifting of the iir into the caller, we can shrink the code still further: add/remove: 0/0 grow/shrink: 0/3 up/down: 0/-34 (-34) Function old new delta gen8_cs_irq_handler 123 118 -5 gen8_gt_irq_handler 261 248 -13 gen11_irq_handler 722 706 -16 v2: Drop gen11_cs_irq_handler now that it is a simple stub around gen8_cs_irq_handler (Daniele) References: 5d3d69d5c119 ("drm/i915: Stop inlining the execlists IRQ handler") Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Tvrtko Ursulin Cc: Daniele Ceraolo Spurio Reviewed-by: Daniele Ceraolo Spurio Link: https://patchwork.freedesktop.org/patch/msgid/20180309010808.11921-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_irq.c | 38 ++++++++++++++++---------------------- 1 file changed, 16 insertions(+), 22 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index babf81c..c8c29d8 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -1399,19 +1399,19 @@ static void snb_gt_irq_handler(struct drm_i915_private *dev_priv, } static void -gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir, int test_shift) +gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir) { struct intel_engine_execlists * const execlists = &engine->execlists; bool tasklet = false; - if (iir & (GT_CONTEXT_SWITCH_INTERRUPT << test_shift)) { + if (iir & GT_CONTEXT_SWITCH_INTERRUPT) { if (READ_ONCE(engine->execlists.active)) { __set_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); tasklet = true; } } - if (iir & (GT_RENDER_USER_INTERRUPT << test_shift)) { + if (iir & GT_RENDER_USER_INTERRUPT) { notify_ring(engine); tasklet |= USES_GUC_SUBMISSION(engine->i915); } @@ -1466,21 +1466,21 @@ static void gen8_gt_irq_handler(struct drm_i915_private *i915, { if (master_ctl & (GEN8_GT_RCS_IRQ | GEN8_GT_BCS_IRQ)) { gen8_cs_irq_handler(i915->engine[RCS], - gt_iir[0], GEN8_RCS_IRQ_SHIFT); + gt_iir[0] >> GEN8_RCS_IRQ_SHIFT); gen8_cs_irq_handler(i915->engine[BCS], - gt_iir[0], GEN8_BCS_IRQ_SHIFT); + gt_iir[0] >> GEN8_BCS_IRQ_SHIFT); } if (master_ctl & (GEN8_GT_VCS1_IRQ | GEN8_GT_VCS2_IRQ)) { gen8_cs_irq_handler(i915->engine[VCS], - gt_iir[1], GEN8_VCS1_IRQ_SHIFT); + gt_iir[1] >> GEN8_VCS1_IRQ_SHIFT); gen8_cs_irq_handler(i915->engine[VCS2], - gt_iir[1], GEN8_VCS2_IRQ_SHIFT); + gt_iir[1] >> GEN8_VCS2_IRQ_SHIFT); } if (master_ctl & GEN8_GT_VECS_IRQ) { gen8_cs_irq_handler(i915->engine[VECS], - gt_iir[3], GEN8_VECS_IRQ_SHIFT); + gt_iir[3] >> GEN8_VECS_IRQ_SHIFT); } if (master_ctl & (GEN8_GT_PM_IRQ | GEN8_GT_GUC_IRQ)) { @@ -2762,12 +2762,6 @@ static void __fini_wedge(struct wedge_me *w) (W)->i915; \ __fini_wedge((W))) -static __always_inline void -gen11_cs_irq_handler(struct intel_engine_cs * const engine, const u32 iir) -{ - gen8_cs_irq_handler(engine, iir, 0); -} - static void gen11_gt_engine_irq_handler(struct drm_i915_private * const i915, const unsigned int bank, @@ -2781,27 +2775,27 @@ gen11_gt_engine_irq_handler(struct drm_i915_private * const i915, switch (engine_n) { case GEN11_RCS0: - return gen11_cs_irq_handler(engine[RCS], iir); + return gen8_cs_irq_handler(engine[RCS], iir); case GEN11_BCS: - return gen11_cs_irq_handler(engine[BCS], iir); + return gen8_cs_irq_handler(engine[BCS], iir); } case 1: switch (engine_n) { case GEN11_VCS(0): - return gen11_cs_irq_handler(engine[_VCS(0)], iir); + return gen8_cs_irq_handler(engine[_VCS(0)], iir); case GEN11_VCS(1): - return gen11_cs_irq_handler(engine[_VCS(1)], iir); + return gen8_cs_irq_handler(engine[_VCS(1)], iir); case GEN11_VCS(2): - return gen11_cs_irq_handler(engine[_VCS(2)], iir); + return gen8_cs_irq_handler(engine[_VCS(2)], iir); case GEN11_VCS(3): - return gen11_cs_irq_handler(engine[_VCS(3)], iir); + return gen8_cs_irq_handler(engine[_VCS(3)], iir); case GEN11_VECS(0): - return gen11_cs_irq_handler(engine[_VECS(0)], iir); + return gen8_cs_irq_handler(engine[_VECS(0)], iir); case GEN11_VECS(1): - return gen11_cs_irq_handler(engine[_VECS(1)], iir); + return gen8_cs_irq_handler(engine[_VECS(1)], iir); } } } -- cgit v1.1 From 1e6aa7e55c28ecd842b8b4599e4273c2429ee061 Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Tue, 6 Mar 2018 12:41:55 +0200 Subject: drm/i915/icl: do not save DDI A/E sharing bit for ICL MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We don't want to preserve the DDI A 4 lane bit on ICL. Fixes: 3d2011cfa41f ("drm/i915/icl: remove port A/E lane sharing limitation.") Cc: Mahesh Kumar Cc: Paulo Zanoni Reviewed-by: Ville Syrjälä Signed-off-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20180306104155.3526-1-jani.nikula@intel.com --- drivers/gpu/drm/i915/intel_ddi.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index ac8fc2a..dbcf1a0 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -3080,9 +3080,12 @@ void intel_ddi_init(struct drm_i915_private *dev_priv, enum port port) intel_encoder->crtc_mask = (1 << 0) | (1 << 1) | (1 << 2); intel_encoder->cloneable = 0; - intel_dig_port->saved_port_bits = I915_READ(DDI_BUF_CTL(port)) & - (DDI_BUF_PORT_REVERSAL | - DDI_A_4_LANES); + if (INTEL_GEN(dev_priv) >= 11) + intel_dig_port->saved_port_bits = I915_READ(DDI_BUF_CTL(port)) & + DDI_BUF_PORT_REVERSAL; + else + intel_dig_port->saved_port_bits = I915_READ(DDI_BUF_CTL(port)) & + (DDI_BUF_PORT_REVERSAL | DDI_A_4_LANES); intel_dig_port->dp.output_reg = INVALID_MMIO_REG; intel_dig_port->max_lanes = intel_ddi_max_lanes(intel_dig_port); -- cgit v1.1 From 2d4ecace3a7861c6071235a6cc88067b8c3eec4a Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 7 Mar 2018 13:42:21 +0000 Subject: drm/i915: Finish the wait-for-wedge by retiring all the inflight requests Before we reset the GPU after marking the device as wedged, we wait for all the remaining requests to be completed (and marked as EIO). Afterwards, we should flush the request lists so the next batch start with the driver in an idle state. Signed-off-by: Chris Wilson Cc: Mika Kuoppala Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180307134226.25492-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_gem.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index ab88ca5..c3d6507 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -3281,7 +3281,8 @@ bool i915_gem_unset_wedged(struct drm_i915_private *i915) if (!test_bit(I915_WEDGED, &i915->gpu_error.flags)) return true; - /* Before unwedging, make sure that all pending operations + /* + * Before unwedging, make sure that all pending operations * are flushed and errored out - we may have requests waiting upon * third party fences. We marked all inflight requests as EIO, and * every execbuf since returned EIO, for consistency we want all @@ -3299,7 +3300,8 @@ bool i915_gem_unset_wedged(struct drm_i915_private *i915) if (!rq) continue; - /* We can't use our normal waiter as we want to + /* + * We can't use our normal waiter as we want to * avoid recursively trying to handle the current * reset. The basic dma_fence_default_wait() installs * a callback for dma_fence_signal(), which is @@ -3314,8 +3316,11 @@ bool i915_gem_unset_wedged(struct drm_i915_private *i915) return false; } } + i915_retire_requests(i915); + GEM_BUG_ON(i915->gt.active_requests); - /* Undo nop_submit_request. We prevent all new i915 requests from + /* + * Undo nop_submit_request. We prevent all new i915 requests from * being queued (by disallowing execbuf whilst wedged) so having * waited for all active requests above, we know the system is idle * and do not have to worry about a thread being inside -- cgit v1.1 From 36620032ceccb4bf07bbe780a3998e88a585ad69 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 7 Mar 2018 13:42:23 +0000 Subject: drm/i915: Update ring position from request on retiring When wedged, we do not update the ring->tail as we submit the requests causing us to leak the ring->space upon cleaning up the wedged driver. We can just use the value stored in rq->tail, and keep the submission backend details away from set-wedge. Signed-off-by: Chris Wilson Cc: Mika Kuoppala Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180307134226.25492-3-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_request.c | 2 +- drivers/gpu/drm/i915/intel_ringbuffer.c | 1 + 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index d437bea..75c8826 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -358,7 +358,7 @@ static void advance_ring(struct i915_request *request) * is just about to be. Either works, if we miss the last two * noops - they are safe to be replayed on a reset. */ - tail = READ_ONCE(request->ring->tail); + tail = READ_ONCE(request->tail); } else { tail = request->postfix; } diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 1d59952..88eeb64 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1593,6 +1593,7 @@ static noinline int wait_for_space(struct intel_ring *ring, unsigned int bytes) if (intel_ring_update_space(ring) >= bytes) return 0; + GEM_BUG_ON(list_empty(&ring->request_list)); list_for_each_entry(target, &ring->request_list, ring_link) { /* Would completion of this request free enough space? */ if (bytes <= __intel_ring_space(target->postfix, -- cgit v1.1 From ef5032a06a73ca5f40ce6975d956aa478536c411 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 7 Mar 2018 13:42:24 +0000 Subject: drm/i915: Include ring->emit in debugging Include ring->emit and ring->space alongside ring->(head,tail) when printing debug information. Signed-off-by: Chris Wilson Cc: Mika Kuoppala Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180307134226.25492-4-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_debugfs.c | 4 ++-- drivers/gpu/drm/i915/intel_engine_cs.c | 10 +++++++--- 2 files changed, 9 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index d8bc1bb..34d1252 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -1922,8 +1922,8 @@ static int i915_gem_framebuffer_info(struct seq_file *m, void *data) static void describe_ctx_ring(struct seq_file *m, struct intel_ring *ring) { - seq_printf(m, " (ringbuffer, space: %d, head: %u, tail: %u)", - ring->space, ring->head, ring->tail); + seq_printf(m, " (ringbuffer, space: %d, head: %u, tail: %u, emit: %u)", + ring->space, ring->head, ring->tail, ring->emit); } static int i915_context_status(struct seq_file *m, void *unused) diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 4ba139c..048cd01 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1929,12 +1929,16 @@ void intel_engine_dump(struct intel_engine_cs *engine, rq->head, rq->postfix, rq->tail, rq->batch ? upper_32_bits(rq->batch->node.start) : ~0u, rq->batch ? lower_32_bits(rq->batch->node.start) : ~0u); - drm_printf(m, "\t\tring->start: 0x%08x\n", + drm_printf(m, "\t\tring->start: 0x%08x\n", i915_ggtt_offset(rq->ring->vma)); - drm_printf(m, "\t\tring->head: 0x%08x\n", + drm_printf(m, "\t\tring->head: 0x%08x\n", rq->ring->head); - drm_printf(m, "\t\tring->tail: 0x%08x\n", + drm_printf(m, "\t\tring->tail: 0x%08x\n", rq->ring->tail); + drm_printf(m, "\t\tring->emit: 0x%08x\n", + rq->ring->emit); + drm_printf(m, "\t\tring->space: 0x%08x\n", + rq->ring->space); } rcu_read_unlock(); -- cgit v1.1 From 47650db02dd52267953df81438c93cf8a0eb0e5e Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 7 Mar 2018 13:42:25 +0000 Subject: drm/i915: Wrap engine->schedule in RCU locks for set-wedge protection Similar to the staging around handling of engine->submit_request, we need to stop adding to the execlists->queue prior to calling engine->cancel_requests. cancel_requests will move requests from the queue onto the timeline, so if we add a request onto the queue after that point, it will be lost. Fixes: af7a8ffad9c5 ("drm/i915: Use rcu instead of stop_machine in set_wedged") Signed-off-by: Chris Wilson Cc: Mika Kuoppala Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180307134226.25492-5-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_gem.c | 13 +++++++------ drivers/gpu/drm/i915/i915_request.c | 2 ++ 2 files changed, 9 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index c3d6507..50e165b 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -479,10 +479,11 @@ static void __fence_set_priority(struct dma_fence *fence, int prio) rq = to_request(fence); engine = rq->engine; - if (!engine->schedule) - return; - engine->schedule(rq, prio); + rcu_read_lock(); + if (engine->schedule) + engine->schedule(rq, prio); + rcu_read_unlock(); } static void fence_set_priority(struct dma_fence *fence, int prio) @@ -3222,8 +3223,11 @@ void i915_gem_set_wedged(struct drm_i915_private *i915) */ for_each_engine(engine, i915, id) { i915_gem_reset_prepare_engine(engine); + engine->submit_request = nop_submit_request; + engine->schedule = NULL; } + i915->caps.scheduler = 0; /* * Make sure no one is running the old callback before we proceed with @@ -3241,11 +3245,8 @@ void i915_gem_set_wedged(struct drm_i915_private *i915) * start to complete all requests. */ engine->submit_request = nop_complete_submit_request; - engine->schedule = NULL; } - i915->caps.scheduler = 0; - /* * Make sure no request can slip through without getting completed by * either this call here to intel_engine_init_global_seqno, or the one diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index 75c8826..2f62acd 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -1081,8 +1081,10 @@ void __i915_request_add(struct i915_request *request, bool flush_caches) * decide whether to preempt the entire chain so that it is ready to * run at the earliest possible convenience. */ + rcu_read_lock(); if (engine->schedule) engine->schedule(request, request->ctx->priority); + rcu_read_unlock(); local_bh_disable(); i915_sw_fence_commit(&request->submit); -- cgit v1.1 From 68ad361285a9cc73b259f59adbaafde196c15987 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 7 Mar 2018 13:42:26 +0000 Subject: drm/i915: Only call tasklet_kill() on the first prepare_reset tasklet_kill() will spin waiting for the current tasklet to be executed. However, if tasklet_disable() has been called, then the tasklet is never executed but permanently put back onto the runlist until tasklet_enable() is called. Ergo, we cannot use tasklet_kill() inside a disable/enable pair. This is the case when we call set-wedge from inside i915_reset(), and another request was submitted to us concurrent to the reset. Fixes: 963ddd63c314 ("drm/i915: Suspend submission tasklets around wedging") Signed-off-by: Chris Wilson Cc: Mika Kuoppala Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180307134226.25492-6-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_gem.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 50e165b..e58b741 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -2948,8 +2948,16 @@ i915_gem_reset_prepare_engine(struct intel_engine_cs *engine) * calling engine->init_hw() and also writing the ELSP. * Turning off the execlists->tasklet until the reset is over * prevents the race. + * + * Note that this needs to be a single atomic operation on the + * tasklet (flush existing tasks, prevent new tasks) to prevent + * a race between reset and set-wedged. It is not, so we do the best + * we can atm and make sure we don't lock the machine up in the more + * common case of recursively being called from set-wedged from inside + * i915_reset. */ - tasklet_kill(&engine->execlists.tasklet); + if (!atomic_read(&engine->execlists.tasklet.count)) + tasklet_kill(&engine->execlists.tasklet); tasklet_disable(&engine->execlists.tasklet); /* -- cgit v1.1 From df9471689685857c38d5e095652a1bc867ee11cf Mon Sep 17 00:00:00 2001 From: Lyude Paul Date: Thu, 8 Mar 2018 18:24:15 -0500 Subject: drm/i915: Remove unused DP_LINK_CHECK_TIMEOUT MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Lyude Paul Cc: Manasi Navare Cc: Ville Syrjälä Reviewed-by: Manasi Navare Link: https://patchwork.freedesktop.org/patch/msgid/20180308232421.14049-2-lyude@redhat.com --- drivers/gpu/drm/i915/intel_dp.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 9a4a51e..4dd1b22 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -43,7 +43,6 @@ #include #include "i915_drv.h" -#define DP_LINK_CHECK_TIMEOUT (10 * 1000) #define DP_DPRX_ESI_LEN 14 /* Compliance test status bits */ -- cgit v1.1 From 86aa82476cffdfa2cb85c2dc8b198e1675773982 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Thu, 8 Mar 2018 16:46:53 +0100 Subject: drm/i915/guc: Tidy guc_log_control MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We plan to decouple log runtime (mapping + relay) from verbosity control. Let's tidy the code now to reduce the churn in the following patches. v2: Tidy macros, keep debug messages, use helper var for enable, correct typo (Michał) Fix incorrect input validaction (Sagar) Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Michal Wajdeczko Cc: Sagar Arun Kamble Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180308154707.21716-1-michal.winiarski@intel.com --- drivers/gpu/drm/i915/i915_debugfs.c | 11 ++--- drivers/gpu/drm/i915/intel_guc_log.c | 80 +++++++++++++++++++++--------------- drivers/gpu/drm/i915/intel_guc_log.h | 3 +- 3 files changed, 53 insertions(+), 41 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 34d1252..c4cc8fe 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -2499,13 +2499,10 @@ static int i915_guc_log_control_get(void *data, u64 *val) { struct drm_i915_private *dev_priv = data; - if (!HAS_GUC(dev_priv)) + if (!USES_GUC(dev_priv)) return -ENODEV; - if (!dev_priv->guc.log.vma) - return -EINVAL; - - *val = i915_modparams.guc_log_level; + *val = intel_guc_log_control_get(&dev_priv->guc); return 0; } @@ -2514,10 +2511,10 @@ static int i915_guc_log_control_set(void *data, u64 val) { struct drm_i915_private *dev_priv = data; - if (!HAS_GUC(dev_priv)) + if (!USES_GUC(dev_priv)) return -ENODEV; - return intel_guc_log_control(&dev_priv->guc, val); + return intel_guc_log_control_set(&dev_priv->guc, val); } DEFINE_SIMPLE_ATTRIBUTE(i915_guc_log_control_fops, diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index c0c2e7d..7e59fb0 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -659,51 +659,63 @@ void intel_guc_log_destroy(struct intel_guc *guc) i915_vma_unpin_and_release(&guc->log.vma); } -int intel_guc_log_control(struct intel_guc *guc, u64 control_val) +int intel_guc_log_control_get(struct intel_guc *guc) +{ + GEM_BUG_ON(!guc->log.vma); + GEM_BUG_ON(i915_modparams.guc_log_level < 0); + + return i915_modparams.guc_log_level; +} + +#define GUC_LOG_LEVEL_DISABLED 0 +#define LOG_LEVEL_TO_ENABLED(x) ((x) > 0) +#define LOG_LEVEL_TO_VERBOSITY(x) ({ \ + typeof(x) _x = (x); \ + LOG_LEVEL_TO_ENABLED(_x) ? _x - 1 : 0; \ +}) +#define VERBOSITY_TO_LOG_LEVEL(x) ((x) + 1) +int intel_guc_log_control_set(struct intel_guc *guc, u64 val) { struct drm_i915_private *dev_priv = guc_to_i915(guc); - bool enable_logging = control_val > 0; - u32 verbosity; + bool enabled = LOG_LEVEL_TO_ENABLED(val); int ret; - if (!guc->log.vma) - return -ENODEV; + BUILD_BUG_ON(GUC_LOG_VERBOSITY_MIN != 0); + GEM_BUG_ON(!guc->log.vma); + GEM_BUG_ON(i915_modparams.guc_log_level < 0); - BUILD_BUG_ON(GUC_LOG_VERBOSITY_MIN); - if (control_val > 1 + GUC_LOG_VERBOSITY_MAX) + /* + * GuC is recognizing log levels starting from 0 to max, we're using 0 + * as indication that logging should be disabled. + */ + if (val < GUC_LOG_LEVEL_DISABLED || + val > VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX)) return -EINVAL; - /* This combination doesn't make sense & won't have any effect */ - if (!enable_logging && !i915_modparams.guc_log_level) - return 0; + mutex_lock(&dev_priv->drm.struct_mutex); - verbosity = enable_logging ? control_val - 1 : 0; + if (i915_modparams.guc_log_level == val) { + ret = 0; + goto out_unlock; + } - ret = mutex_lock_interruptible(&dev_priv->drm.struct_mutex); - if (ret) - return ret; intel_runtime_pm_get(dev_priv); - ret = guc_log_control(guc, enable_logging, verbosity); + ret = guc_log_control(guc, enabled, LOG_LEVEL_TO_VERBOSITY(val)); intel_runtime_pm_put(dev_priv); - mutex_unlock(&dev_priv->drm.struct_mutex); - - if (ret < 0) { - DRM_DEBUG_DRIVER("guc_logging_control action failed %d\n", ret); - return ret; + if (ret) { + DRM_DEBUG_DRIVER("guc_log_control action failed %d\n", ret); + goto out_unlock; } - if (enable_logging) { - i915_modparams.guc_log_level = 1 + verbosity; + i915_modparams.guc_log_level = val; - /* - * If log was disabled at boot time, then the relay channel file - * wouldn't have been created by now and interrupts also would - * not have been enabled. Try again now, just in case. - */ + mutex_unlock(&dev_priv->drm.struct_mutex); + + if (enabled && !guc_log_has_runtime(guc)) { ret = guc_log_late_setup(guc); - if (ret < 0) { + if (ret) { DRM_DEBUG_DRIVER("GuC log late setup failed %d\n", ret); - return ret; + goto out; } /* GuC logging is currently the only user of Guc2Host interrupts */ @@ -712,7 +724,7 @@ int intel_guc_log_control(struct intel_guc *guc, u64 control_val) gen9_enable_guc_interrupts(dev_priv); intel_runtime_pm_put(dev_priv); mutex_unlock(&dev_priv->drm.struct_mutex); - } else { + } else if (!enabled && guc_log_has_runtime(guc)) { /* * Once logging is disabled, GuC won't generate logs & send an * interrupt. But there could be some data in the log buffer @@ -720,11 +732,13 @@ int intel_guc_log_control(struct intel_guc *guc, u64 control_val) * buffer state and then collect the left over logs. */ guc_flush_logs(guc); - - /* As logging is disabled, update log level to reflect that */ - i915_modparams.guc_log_level = 0; } + return 0; + +out_unlock: + mutex_unlock(&dev_priv->drm.struct_mutex); +out: return ret; } diff --git a/drivers/gpu/drm/i915/intel_guc_log.h b/drivers/gpu/drm/i915/intel_guc_log.h index dab0e94..141ce9c 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.h +++ b/drivers/gpu/drm/i915/intel_guc_log.h @@ -64,7 +64,8 @@ void intel_guc_log_destroy(struct intel_guc *guc); void intel_guc_log_init_early(struct intel_guc *guc); int intel_guc_log_relay_create(struct intel_guc *guc); void intel_guc_log_relay_destroy(struct intel_guc *guc); -int intel_guc_log_control(struct intel_guc *guc, u64 control_val); +int intel_guc_log_control_get(struct intel_guc *guc); +int intel_guc_log_control_set(struct intel_guc *guc, u64 control_val); void i915_guc_log_register(struct drm_i915_private *dev_priv); void i915_guc_log_unregister(struct drm_i915_private *dev_priv); -- cgit v1.1 From 950724ba88521195b4aefd092c8a0337487be352 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Thu, 8 Mar 2018 16:46:54 +0100 Subject: drm/i915/guc: Create common entry points for log register/unregister MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We have many functions responsible for allocating different parts of GuC log runtime called from multiple places. Let's stick with keeping everything in guc_log_register instead. v2: Use more generic intel_uc_register name, keep using "misc" suffix (Michał) s/dev_priv/i915 (Sagar) Make guc_log_relay_* static (sparse) Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Michal Wajdeczko Cc: Sagar Arun Kamble Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180308154707.21716-2-michal.winiarski@intel.com --- drivers/gpu/drm/i915/i915_drv.c | 6 +- drivers/gpu/drm/i915/intel_guc_log.c | 156 ++++++++++++++--------------------- drivers/gpu/drm/i915/intel_guc_log.h | 6 +- drivers/gpu/drm/i915/intel_uc.c | 41 +++++---- drivers/gpu/drm/i915/intel_uc.h | 2 + 5 files changed, 95 insertions(+), 116 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index d7c4de4..987c677 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -1238,9 +1238,11 @@ static void i915_driver_register(struct drm_i915_private *dev_priv) /* Reveal our presence to userspace */ if (drm_dev_register(dev, 0) == 0) { i915_debugfs_register(dev_priv); - i915_guc_log_register(dev_priv); i915_setup_sysfs(dev_priv); + /* Depends on debugfs having been initialized */ + intel_uc_register(dev_priv); + /* Depends on sysfs having been initialized */ i915_perf_register(dev_priv); } else @@ -1298,7 +1300,7 @@ static void i915_driver_unregister(struct drm_i915_private *dev_priv) i915_pmu_unregister(dev_priv); i915_teardown_sysfs(dev_priv); - i915_guc_log_unregister(dev_priv); + intel_uc_unregister(dev_priv); drm_dev_unregister(&dev_priv->drm); i915_gem_shrinker_unregister(dev_priv); diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 7e59fb0..90b395f 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -443,7 +443,7 @@ void intel_guc_log_init_early(struct intel_guc *guc) INIT_WORK(&guc->log.runtime.flush_work, capture_logs_work); } -int intel_guc_log_relay_create(struct intel_guc *guc) +static int guc_log_relay_create(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); struct rchan *guc_log_relay_chan; @@ -496,7 +496,7 @@ err: return ret; } -void intel_guc_log_relay_destroy(struct intel_guc *guc) +static void guc_log_relay_destroy(struct intel_guc *guc) { mutex_lock(&guc->log.runtime.relay_lock); @@ -514,49 +514,6 @@ out_unlock: mutex_unlock(&guc->log.runtime.relay_lock); } -static int guc_log_late_setup(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - int ret; - - if (!guc_log_has_runtime(guc)) { - /* - * If log was disabled at boot time, then setup needed to handle - * log buffer flush interrupts would not have been done yet, so - * do that now. - */ - ret = intel_guc_log_relay_create(guc); - if (ret) - goto err; - - mutex_lock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_get(dev_priv); - ret = guc_log_runtime_create(guc); - intel_runtime_pm_put(dev_priv); - mutex_unlock(&dev_priv->drm.struct_mutex); - - if (ret) - goto err_relay; - } - - ret = guc_log_relay_file_create(guc); - if (ret) - goto err_runtime; - - return 0; - -err_runtime: - mutex_lock(&dev_priv->drm.struct_mutex); - guc_log_runtime_destroy(guc); - mutex_unlock(&dev_priv->drm.struct_mutex); -err_relay: - intel_guc_log_relay_destroy(guc); -err: - /* logging will remain off */ - i915_modparams.guc_log_level = 0; - return ret; -} - static void guc_log_capture_logs(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); @@ -576,16 +533,6 @@ static void guc_flush_logs(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); - if (!USES_GUC_SUBMISSION(dev_priv) || !i915_modparams.guc_log_level) - return; - - /* First disable the interrupts, will be renabled afterwards */ - mutex_lock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_get(dev_priv); - gen9_disable_guc_interrupts(dev_priv); - intel_runtime_pm_put(dev_priv); - mutex_unlock(&dev_priv->drm.struct_mutex); - /* * Before initiating the forceful flush, wait for any pending/ongoing * flush to complete otherwise forceful flush may not actually happen. @@ -628,12 +575,6 @@ int intel_guc_log_create(struct intel_guc *guc) guc->log.vma = vma; - if (i915_modparams.guc_log_level) { - ret = guc_log_runtime_create(guc); - if (ret < 0) - goto err_vma; - } - /* each allocated unit is a page */ flags = GUC_LOG_VALID | GUC_LOG_NOTIFY_ON_HALF_FULL | (GUC_LOG_DPC_PAGES << GUC_LOG_DPC_SHIFT) | @@ -645,8 +586,6 @@ int intel_guc_log_create(struct intel_guc *guc) return 0; -err_vma: - i915_vma_unpin_and_release(&guc->log.vma); err: /* logging will be off */ i915_modparams.guc_log_level = 0; @@ -712,26 +651,14 @@ int intel_guc_log_control_set(struct intel_guc *guc, u64 val) mutex_unlock(&dev_priv->drm.struct_mutex); if (enabled && !guc_log_has_runtime(guc)) { - ret = guc_log_late_setup(guc); + ret = intel_guc_log_register(guc); if (ret) { - DRM_DEBUG_DRIVER("GuC log late setup failed %d\n", ret); + /* logging will remain off */ + i915_modparams.guc_log_level = 0; goto out; } - - /* GuC logging is currently the only user of Guc2Host interrupts */ - mutex_lock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_get(dev_priv); - gen9_enable_guc_interrupts(dev_priv); - intel_runtime_pm_put(dev_priv); - mutex_unlock(&dev_priv->drm.struct_mutex); } else if (!enabled && guc_log_has_runtime(guc)) { - /* - * Once logging is disabled, GuC won't generate logs & send an - * interrupt. But there could be some data in the log buffer - * which is yet to be captured. So request GuC to update the log - * buffer state and then collect the left over logs. - */ - guc_flush_logs(guc); + intel_guc_log_unregister(guc); } return 0; @@ -742,29 +669,72 @@ out: return ret; } -void i915_guc_log_register(struct drm_i915_private *dev_priv) +int intel_guc_log_register(struct intel_guc *guc) { - if (!USES_GUC_SUBMISSION(dev_priv) || !i915_modparams.guc_log_level) - return; + struct drm_i915_private *i915 = guc_to_i915(guc); + int ret; + + GEM_BUG_ON(guc_log_has_runtime(guc)); + + /* + * If log was disabled at boot time, then setup needed to handle + * log buffer flush interrupts would not have been done yet, so + * do that now. + */ + ret = guc_log_relay_create(guc); + if (ret) + goto err; + + mutex_lock(&i915->drm.struct_mutex); + ret = guc_log_runtime_create(guc); + mutex_unlock(&i915->drm.struct_mutex); + + if (ret) + goto err_relay; + + ret = guc_log_relay_file_create(guc); + if (ret) + goto err_runtime; + + /* GuC logging is currently the only user of Guc2Host interrupts */ + mutex_lock(&i915->drm.struct_mutex); + intel_runtime_pm_get(i915); + gen9_enable_guc_interrupts(i915); + intel_runtime_pm_put(i915); + mutex_unlock(&i915->drm.struct_mutex); + + return 0; - guc_log_late_setup(&dev_priv->guc); +err_runtime: + mutex_lock(&i915->drm.struct_mutex); + guc_log_runtime_destroy(guc); + mutex_unlock(&i915->drm.struct_mutex); +err_relay: + guc_log_relay_destroy(guc); +err: + return ret; } -void i915_guc_log_unregister(struct drm_i915_private *dev_priv) +void intel_guc_log_unregister(struct intel_guc *guc) { - struct intel_guc *guc = &dev_priv->guc; + struct drm_i915_private *i915 = guc_to_i915(guc); - if (!USES_GUC_SUBMISSION(dev_priv)) - return; + /* + * Once logging is disabled, GuC won't generate logs & send an + * interrupt. But there could be some data in the log buffer + * which is yet to be captured. So request GuC to update the log + * buffer state and then collect the left over logs. + */ + guc_flush_logs(guc); - mutex_lock(&dev_priv->drm.struct_mutex); + mutex_lock(&i915->drm.struct_mutex); /* GuC logging is currently the only user of Guc2Host interrupts */ - intel_runtime_pm_get(dev_priv); - gen9_disable_guc_interrupts(dev_priv); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_get(i915); + gen9_disable_guc_interrupts(i915); + intel_runtime_pm_put(i915); guc_log_runtime_destroy(guc); - mutex_unlock(&dev_priv->drm.struct_mutex); + mutex_unlock(&i915->drm.struct_mutex); - intel_guc_log_relay_destroy(guc); + guc_log_relay_destroy(guc); } diff --git a/drivers/gpu/drm/i915/intel_guc_log.h b/drivers/gpu/drm/i915/intel_guc_log.h index 141ce9c..09dd2ef 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.h +++ b/drivers/gpu/drm/i915/intel_guc_log.h @@ -62,11 +62,9 @@ struct intel_guc_log { int intel_guc_log_create(struct intel_guc *guc); void intel_guc_log_destroy(struct intel_guc *guc); void intel_guc_log_init_early(struct intel_guc *guc); -int intel_guc_log_relay_create(struct intel_guc *guc); -void intel_guc_log_relay_destroy(struct intel_guc *guc); int intel_guc_log_control_get(struct intel_guc *guc); int intel_guc_log_control_set(struct intel_guc *guc, u64 control_val); -void i915_guc_log_register(struct drm_i915_private *dev_priv); -void i915_guc_log_unregister(struct drm_i915_private *dev_priv); +int intel_guc_log_register(struct intel_guc *guc); +void intel_guc_log_unregister(struct intel_guc *guc); #endif diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index e5bf0d3..1c1a00d 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -219,6 +219,28 @@ static void guc_free_load_err_log(struct intel_guc *guc) i915_gem_object_put(guc->load_err_log); } +int intel_uc_register(struct drm_i915_private *i915) +{ + int ret = 0; + + if (!USES_GUC(i915)) + return 0; + + if (i915_modparams.guc_log_level) + ret = intel_guc_log_register(&i915->guc); + + return ret; +} + +void intel_uc_unregister(struct drm_i915_private *i915) +{ + if (!USES_GUC(i915)) + return; + + if (i915_modparams.guc_log_level) + intel_guc_log_unregister(&i915->guc); +} + static int guc_enable_communication(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); @@ -249,23 +271,10 @@ int intel_uc_init_misc(struct drm_i915_private *dev_priv) return 0; ret = intel_guc_init_wq(guc); - if (ret) { - DRM_ERROR("Couldn't allocate workqueues for GuC\n"); - goto err; - } - - ret = intel_guc_log_relay_create(guc); - if (ret) { - DRM_ERROR("Couldn't allocate relay for GuC log\n"); - goto err_relay; - } + if (ret) + return ret; return 0; - -err_relay: - intel_guc_fini_wq(guc); -err: - return ret; } void intel_uc_fini_misc(struct drm_i915_private *dev_priv) @@ -276,8 +285,6 @@ void intel_uc_fini_misc(struct drm_i915_private *dev_priv) return; intel_guc_fini_wq(guc); - - intel_guc_log_relay_destroy(guc); } int intel_uc_init(struct drm_i915_private *dev_priv) diff --git a/drivers/gpu/drm/i915/intel_uc.h b/drivers/gpu/drm/i915/intel_uc.h index f76d51d..d6af984 100644 --- a/drivers/gpu/drm/i915/intel_uc.h +++ b/drivers/gpu/drm/i915/intel_uc.h @@ -31,6 +31,8 @@ void intel_uc_sanitize_options(struct drm_i915_private *dev_priv); void intel_uc_init_early(struct drm_i915_private *dev_priv); void intel_uc_init_mmio(struct drm_i915_private *dev_priv); +int intel_uc_register(struct drm_i915_private *dev_priv); +void intel_uc_unregister(struct drm_i915_private *dev_priv); void intel_uc_init_fw(struct drm_i915_private *dev_priv); void intel_uc_fini_fw(struct drm_i915_private *dev_priv); int intel_uc_init_misc(struct drm_i915_private *dev_priv); -- cgit v1.1 From 93bf8096c7daa78f0dabc811dad62fa98fe01742 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Thu, 8 Mar 2018 16:46:55 +0100 Subject: drm/i915/guc: Move GuC notification handling to separate function MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit To allow future code reuse. While here, fix comment style. v2: Notifications are a separate thing - rename the handler (Sagar) Suggested-by: Oscar Mateo Signed-off-by: Michal Wajdeczko Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Oscar Mateo Cc: Sagar Arun Kamble Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180308154707.21716-3-michal.winiarski@intel.com --- drivers/gpu/drm/i915/i915_irq.c | 33 ++------------------------------- drivers/gpu/drm/i915/intel_guc.c | 37 +++++++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_guc.h | 1 + 3 files changed, 40 insertions(+), 31 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index c8c29d8..828f310 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -1766,37 +1766,8 @@ static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir) static void gen9_guc_irq_handler(struct drm_i915_private *dev_priv, u32 gt_iir) { - if (gt_iir & GEN9_GUC_TO_HOST_INT_EVENT) { - /* Sample the log buffer flush related bits & clear them out now - * itself from the message identity register to minimize the - * probability of losing a flush interrupt, when there are back - * to back flush interrupts. - * There can be a new flush interrupt, for different log buffer - * type (like for ISR), whilst Host is handling one (for DPC). - * Since same bit is used in message register for ISR & DPC, it - * could happen that GuC sets the bit for 2nd interrupt but Host - * clears out the bit on handling the 1st interrupt. - */ - u32 msg, flush; - - msg = I915_READ(SOFT_SCRATCH(15)); - flush = msg & (INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED | - INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER); - if (flush) { - /* Clear the message bits that are handled */ - I915_WRITE(SOFT_SCRATCH(15), msg & ~flush); - - /* Handle flush interrupt in bottom half */ - queue_work(dev_priv->guc.log.runtime.flush_wq, - &dev_priv->guc.log.runtime.flush_work); - - dev_priv->guc.log.flush_interrupt_count++; - } else { - /* Not clearing of unhandled event bits won't result in - * re-triggering of the interrupt. - */ - } - } + if (gt_iir & GEN9_GUC_TO_HOST_INT_EVENT) + intel_guc_to_host_event_handler(&dev_priv->guc); } static void i9xx_pipestat_irq_reset(struct drm_i915_private *dev_priv) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index ff08ea0..25f9229 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -364,6 +364,43 @@ int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len) return ret; } +void intel_guc_to_host_event_handler(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + u32 msg, flush; + + /* + * Sample the log buffer flush related bits & clear them out now + * itself from the message identity register to minimize the + * probability of losing a flush interrupt, when there are back + * to back flush interrupts. + * There can be a new flush interrupt, for different log buffer + * type (like for ISR), whilst Host is handling one (for DPC). + * Since same bit is used in message register for ISR & DPC, it + * could happen that GuC sets the bit for 2nd interrupt but Host + * clears out the bit on handling the 1st interrupt. + */ + + msg = I915_READ(SOFT_SCRATCH(15)); + flush = msg & (INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED | + INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER); + if (flush) { + /* Clear the message bits that are handled */ + I915_WRITE(SOFT_SCRATCH(15), msg & ~flush); + + /* Handle flush interrupt in bottom half */ + queue_work(guc->log.runtime.flush_wq, + &guc->log.runtime.flush_work); + + guc->log.flush_interrupt_count++; + } else { + /* + * Not clearing of unhandled event bits won't result in + * re-triggering of the interrupt. + */ + } +} + int intel_guc_sample_forcewake(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h index b9424ac..6d5aebe 100644 --- a/drivers/gpu/drm/i915/intel_guc.h +++ b/drivers/gpu/drm/i915/intel_guc.h @@ -125,6 +125,7 @@ int intel_guc_init(struct intel_guc *guc); void intel_guc_fini(struct intel_guc *guc); int intel_guc_send_nop(struct intel_guc *guc, const u32 *action, u32 len); int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len); +void intel_guc_to_host_event_handler(struct intel_guc *guc); int intel_guc_sample_forcewake(struct intel_guc *guc); int intel_guc_auth_huc(struct intel_guc *guc, u32 rsa_offset); int intel_guc_suspend(struct intel_guc *guc); -- cgit v1.1 From ff491603ffec80d79b970d540f066535c8743796 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Thu, 8 Mar 2018 09:50:34 +0000 Subject: drm/i915: Include i915_reg.h in intel_ringbuffer.h Header intel_ringbuffer.h is using definitions from i915_reg.h but forget to include it. Remove this hidden dependency by explicitly include missing header. v2: add reminder (Chris) Signed-off-by: Michal Wajdeczko Cc: Chris Wilson Cc: Tvrtko Ursulin Reviewed-by: Tvrtko Ursulin Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180308095037.18264-2-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_ringbuffer.h | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 0320c2c..c31258d 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -7,6 +7,7 @@ #include "i915_gem_batch_pool.h" #include "i915_gem_timeline.h" +#include "i915_reg.h" /* FIXME split out i915_gpu_commands.h */ #include "i915_pmu.h" #include "i915_request.h" #include "i915_selftest.h" -- cgit v1.1 From c5781351450db8ff8374657a8c568772967f3795 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Thu, 8 Mar 2018 09:50:35 +0000 Subject: drm/i915: Change parameters order in i915_gem_batch_pool_init Function i915_gem_batch_pool_init() failed to follow obj-verb naming schema. Fix that by swapping function parameters. While here, change license text to SPDX format. v2: use intel_engine_init_batch_pool (Chris) as proxy (Michal) Signed-off-by: Michal Wajdeczko Cc: Chris Wilson Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180308095037.18264-3-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/i915_gem_batch_pool.c | 30 ++++++------------------------ drivers/gpu/drm/i915/i915_gem_batch_pool.h | 29 +++++------------------------ drivers/gpu/drm/i915/intel_engine_cs.c | 9 ++++++--- 3 files changed, 17 insertions(+), 51 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem_batch_pool.c b/drivers/gpu/drm/i915/i915_gem_batch_pool.c index d3cbe84..f3890b6 100644 --- a/drivers/gpu/drm/i915/i915_gem_batch_pool.c +++ b/drivers/gpu/drm/i915/i915_gem_batch_pool.c @@ -1,29 +1,11 @@ /* - * Copyright © 2014 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. + * SPDX-License-Identifier: MIT * + * Copyright © 2014-2018 Intel Corporation */ -#include "i915_drv.h" #include "i915_gem_batch_pool.h" +#include "i915_drv.h" /** * DOC: batch pool @@ -41,11 +23,11 @@ /** * i915_gem_batch_pool_init() - initialize a batch buffer pool - * @engine: the associated request submission engine * @pool: the batch buffer pool + * @engine: the associated request submission engine */ -void i915_gem_batch_pool_init(struct intel_engine_cs *engine, - struct i915_gem_batch_pool *pool) +void i915_gem_batch_pool_init(struct i915_gem_batch_pool *pool, + struct intel_engine_cs *engine) { int n; diff --git a/drivers/gpu/drm/i915/i915_gem_batch_pool.h b/drivers/gpu/drm/i915/i915_gem_batch_pool.h index 10d5ac4..56947da 100644 --- a/drivers/gpu/drm/i915/i915_gem_batch_pool.h +++ b/drivers/gpu/drm/i915/i915_gem_batch_pool.h @@ -1,31 +1,13 @@ /* - * Copyright © 2014 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. + * SPDX-License-Identifier: MIT * + * Copyright © 2014-2018 Intel Corporation */ #ifndef I915_GEM_BATCH_POOL_H #define I915_GEM_BATCH_POOL_H -#include "i915_drv.h" +#include struct intel_engine_cs; @@ -34,9 +16,8 @@ struct i915_gem_batch_pool { struct list_head cache_list[4]; }; -/* i915_gem_batch_pool.c */ -void i915_gem_batch_pool_init(struct intel_engine_cs *engine, - struct i915_gem_batch_pool *pool); +void i915_gem_batch_pool_init(struct i915_gem_batch_pool *pool, + struct intel_engine_cs *engine); void i915_gem_batch_pool_fini(struct i915_gem_batch_pool *pool); struct drm_i915_gem_object* i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool, size_t size); diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 048cd01..a2b1e9e 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -441,6 +441,11 @@ static void intel_engine_init_timeline(struct intel_engine_cs *engine) engine->timeline = &engine->i915->gt.global_timeline.engine[engine->id]; } +static void intel_engine_init_batch_pool(struct intel_engine_cs *engine) +{ + i915_gem_batch_pool_init(&engine->batch_pool, engine); +} + static bool csb_force_mmio(struct drm_i915_private *i915) { /* @@ -485,11 +490,9 @@ static void intel_engine_init_execlist(struct intel_engine_cs *engine) void intel_engine_setup_common(struct intel_engine_cs *engine) { intel_engine_init_execlist(engine); - intel_engine_init_timeline(engine); intel_engine_init_hangcheck(engine); - i915_gem_batch_pool_init(engine, &engine->batch_pool); - + intel_engine_init_batch_pool(engine); intel_engine_init_cmd_parser(engine); } -- cgit v1.1 From 058a9b43a37a2406a574752707c5346e7b6444f4 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Thu, 8 Mar 2018 09:50:36 +0000 Subject: drm/i915: Make header i915_pmu.h more robust Definitions in i915_pmu.h header depend on other types and declarations that were not explicitly included. Fix that by adding related headers and forward declarations. While here, change license text to SPDX format. v2: don't drop "intel_ringbuffer.h" (Tvrtko) Signed-off-by: Michal Wajdeczko Cc: Chris Wilson Cc: Tvrtko Ursulin Reviewed-by: Tvrtko Ursulin Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180308095037.18264-4-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/i915_pmu.c | 27 +++------------------------ drivers/gpu/drm/i915/i915_pmu.h | 30 ++++++++++-------------------- 2 files changed, 13 insertions(+), 44 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_pmu.c b/drivers/gpu/drm/i915/i915_pmu.c index 964467b..4bc7aef 100644 --- a/drivers/gpu/drm/i915/i915_pmu.c +++ b/drivers/gpu/drm/i915/i915_pmu.c @@ -1,33 +1,12 @@ /* - * Copyright © 2017 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. + * SPDX-License-Identifier: MIT * + * Copyright © 2017-2018 Intel Corporation */ -#include -#include - -#include "i915_drv.h" #include "i915_pmu.h" #include "intel_ringbuffer.h" +#include "i915_drv.h" /* Frequency for the sampling timer for events which need it. */ #define FREQUENCY 200 diff --git a/drivers/gpu/drm/i915/i915_pmu.h b/drivers/gpu/drm/i915/i915_pmu.h index aa1b1a9..2ba7352 100644 --- a/drivers/gpu/drm/i915/i915_pmu.h +++ b/drivers/gpu/drm/i915/i915_pmu.h @@ -1,29 +1,19 @@ /* - * Copyright © 2017 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. + * SPDX-License-Identifier: MIT * + * Copyright © 2017-2018 Intel Corporation */ + #ifndef __I915_PMU_H__ #define __I915_PMU_H__ +#include +#include +#include +#include + +struct drm_i915_private; + enum { __I915_SAMPLE_FREQ_ACT = 0, __I915_SAMPLE_FREQ_REQ, -- cgit v1.1 From d897a111940fe2d644172466914d7c97791bda05 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Thu, 8 Mar 2018 09:50:37 +0000 Subject: drm/i915: Move i915_gpu_error into its own header Error state management code was moved into separate .c unit but we didn't move related definitions into own header. v2: move also intel_display_error_state forward decl fix ("Prefer 'unsigned int' to bare use of 'unsigned'") warnings detected by checkpatch in moved code (Michal) Signed-off-by: Michal Wajdeczko Cc: Chris Wilson Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180308095037.18264-5-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/i915_drv.h | 332 +------------------------------ drivers/gpu/drm/i915/i915_gpu_error.c | 1 + drivers/gpu/drm/i915/i915_gpu_error.h | 356 ++++++++++++++++++++++++++++++++++ 3 files changed, 358 insertions(+), 331 deletions(-) create mode 100644 drivers/gpu/drm/i915/i915_gpu_error.h diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 6e740f6..d35f805 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -72,7 +72,7 @@ #include "i915_gem_object.h" #include "i915_gem_gtt.h" #include "i915_gem_timeline.h" - +#include "i915_gpu_error.h" #include "i915_request.h" #include "i915_vma.h" @@ -453,172 +453,6 @@ struct intel_csr { uint32_t allowed_dc_mask; }; -struct intel_display_error_state; - -struct i915_gpu_state { - struct kref ref; - ktime_t time; - ktime_t boottime; - ktime_t uptime; - - struct drm_i915_private *i915; - - char error_msg[128]; - bool simulated; - bool awake; - bool wakelock; - bool suspended; - int iommu; - u32 reset_count; - u32 suspend_count; - struct intel_device_info device_info; - struct intel_driver_caps driver_caps; - struct i915_params params; - - struct i915_error_uc { - struct intel_uc_fw guc_fw; - struct intel_uc_fw huc_fw; - struct drm_i915_error_object *guc_log; - } uc; - - /* Generic register state */ - u32 eir; - u32 pgtbl_er; - u32 ier; - u32 gtier[4], ngtier; - u32 ccid; - u32 derrmr; - u32 forcewake; - u32 error; /* gen6+ */ - u32 err_int; /* gen7 */ - u32 fault_data0; /* gen8, gen9 */ - u32 fault_data1; /* gen8, gen9 */ - u32 done_reg; - u32 gac_eco; - u32 gam_ecochk; - u32 gab_ctl; - u32 gfx_mode; - - u32 nfence; - u64 fence[I915_MAX_NUM_FENCES]; - struct intel_overlay_error_state *overlay; - struct intel_display_error_state *display; - - struct drm_i915_error_engine { - int engine_id; - /* Software tracked state */ - bool idle; - bool waiting; - int num_waiters; - unsigned long hangcheck_timestamp; - bool hangcheck_stalled; - enum intel_engine_hangcheck_action hangcheck_action; - struct i915_address_space *vm; - int num_requests; - u32 reset_count; - - /* position of active request inside the ring */ - u32 rq_head, rq_post, rq_tail; - - /* our own tracking of ring head and tail */ - u32 cpu_ring_head; - u32 cpu_ring_tail; - - u32 last_seqno; - - /* Register state */ - u32 start; - u32 tail; - u32 head; - u32 ctl; - u32 mode; - u32 hws; - u32 ipeir; - u32 ipehr; - u32 bbstate; - u32 instpm; - u32 instps; - u32 seqno; - u64 bbaddr; - u64 acthd; - u32 fault_reg; - u64 faddr; - u32 rc_psmi; /* sleep state */ - u32 semaphore_mboxes[I915_NUM_ENGINES - 1]; - struct intel_instdone instdone; - - struct drm_i915_error_context { - char comm[TASK_COMM_LEN]; - pid_t pid; - u32 handle; - u32 hw_id; - int priority; - int ban_score; - int active; - int guilty; - bool bannable; - } context; - - struct drm_i915_error_object { - u64 gtt_offset; - u64 gtt_size; - int page_count; - int unused; - u32 *pages[0]; - } *ringbuffer, *batchbuffer, *wa_batchbuffer, *ctx, *hws_page; - - struct drm_i915_error_object **user_bo; - long user_bo_count; - - struct drm_i915_error_object *wa_ctx; - struct drm_i915_error_object *default_state; - - struct drm_i915_error_request { - long jiffies; - pid_t pid; - u32 context; - int priority; - int ban_score; - u32 seqno; - u32 head; - u32 tail; - } *requests, execlist[EXECLIST_MAX_PORTS]; - unsigned int num_ports; - - struct drm_i915_error_waiter { - char comm[TASK_COMM_LEN]; - pid_t pid; - u32 seqno; - } *waiters; - - struct { - u32 gfx_mode; - union { - u64 pdp[4]; - u32 pp_dir_base; - }; - } vm_info; - } engine[I915_NUM_ENGINES]; - - struct drm_i915_error_buffer { - u32 size; - u32 name; - u32 rseqno[I915_NUM_ENGINES], wseqno; - u64 gtt_offset; - u32 read_domains; - u32 write_domain; - s32 fence_reg:I915_MAX_NUM_FENCE_BITS; - u32 tiling:2; - u32 dirty:1; - u32 purgeable:1; - u32 userptr:1; - s32 engine:4; - u32 cache_level:3; - } *active_bo[I915_NUM_ENGINES], *pinned_bo; - u32 active_bo_count[I915_NUM_ENGINES], pinned_bo_count; - struct i915_address_space *active_vm[I915_NUM_ENGINES]; -}; - enum i915_cache_level { I915_CACHE_NONE = 0, I915_CACHE_LLC, /* also used for snoopable memory on non-LLC */ @@ -1146,16 +980,6 @@ struct i915_gem_mm { u32 object_count; }; -struct drm_i915_error_state_buf { - struct drm_i915_private *i915; - unsigned bytes; - unsigned size; - int err; - u8 *buf; - loff_t start; - loff_t pos; -}; - #define I915_IDLE_ENGINES_TIMEOUT (200) /* in ms */ #define I915_RESET_TIMEOUT (10 * HZ) /* 10s */ @@ -1164,102 +988,6 @@ struct drm_i915_error_state_buf { #define I915_ENGINE_DEAD_TIMEOUT (4 * HZ) /* Seqno, head and subunits dead */ #define I915_SEQNO_DEAD_TIMEOUT (12 * HZ) /* Seqno dead with active head */ -struct i915_gpu_error { - /* For hangcheck timer */ -#define DRM_I915_HANGCHECK_PERIOD 1500 /* in ms */ -#define DRM_I915_HANGCHECK_JIFFIES msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD) - - struct delayed_work hangcheck_work; - - /* For reset and error_state handling. */ - spinlock_t lock; - /* Protected by the above dev->gpu_error.lock. */ - struct i915_gpu_state *first_error; - - atomic_t pending_fb_pin; - - unsigned long missed_irq_rings; - - /** - * State variable controlling the reset flow and count - * - * This is a counter which gets incremented when reset is triggered, - * - * Before the reset commences, the I915_RESET_BACKOFF bit is set - * meaning that any waiters holding onto the struct_mutex should - * relinquish the lock immediately in order for the reset to start. - * - * If reset is not completed succesfully, the I915_WEDGE bit is - * set meaning that hardware is terminally sour and there is no - * recovery. All waiters on the reset_queue will be woken when - * that happens. - * - * This counter is used by the wait_seqno code to notice that reset - * event happened and it needs to restart the entire ioctl (since most - * likely the seqno it waited for won't ever signal anytime soon). - * - * This is important for lock-free wait paths, where no contended lock - * naturally enforces the correct ordering between the bail-out of the - * waiter and the gpu reset work code. - */ - unsigned long reset_count; - - /** - * flags: Control various stages of the GPU reset - * - * #I915_RESET_BACKOFF - When we start a reset, we want to stop any - * other users acquiring the struct_mutex. To do this we set the - * #I915_RESET_BACKOFF bit in the error flags when we detect a reset - * and then check for that bit before acquiring the struct_mutex (in - * i915_mutex_lock_interruptible()?). I915_RESET_BACKOFF serves a - * secondary role in preventing two concurrent global reset attempts. - * - * #I915_RESET_HANDOFF - To perform the actual GPU reset, we need the - * struct_mutex. We try to acquire the struct_mutex in the reset worker, - * but it may be held by some long running waiter (that we cannot - * interrupt without causing trouble). Once we are ready to do the GPU - * reset, we set the I915_RESET_HANDOFF bit and wakeup any waiters. If - * they already hold the struct_mutex and want to participate they can - * inspect the bit and do the reset directly, otherwise the worker - * waits for the struct_mutex. - * - * #I915_RESET_ENGINE[num_engines] - Since the driver doesn't need to - * acquire the struct_mutex to reset an engine, we need an explicit - * flag to prevent two concurrent reset attempts in the same engine. - * As the number of engines continues to grow, allocate the flags from - * the most significant bits. - * - * #I915_WEDGED - If reset fails and we can no longer use the GPU, - * we set the #I915_WEDGED bit. Prior to command submission, e.g. - * i915_request_alloc(), this bit is checked and the sequence - * aborted (with -EIO reported to userspace) if set. - */ - unsigned long flags; -#define I915_RESET_BACKOFF 0 -#define I915_RESET_HANDOFF 1 -#define I915_RESET_MODESET 2 -#define I915_WEDGED (BITS_PER_LONG - 1) -#define I915_RESET_ENGINE (I915_WEDGED - I915_NUM_ENGINES) - - /** Number of times an engine has been reset */ - u32 reset_engine_count[I915_NUM_ENGINES]; - - /** - * Waitqueue to signal when a hang is detected. Used to for waiters - * to release the struct_mutex for the reset to procede. - */ - wait_queue_head_t wait_queue; - - /** - * Waitqueue to signal when the reset has completed. Used by clients - * that wait for dev_priv->mm.wedged to settle. - */ - wait_queue_head_t reset_queue; - - /* For missed irq/seqno simulation. */ - unsigned long test_irq_rings; -}; - enum modeset_restore { MODESET_ON_LID_OPEN, MODESET_DONE, @@ -3589,64 +3317,6 @@ static inline int i915_debugfs_connector_add(struct drm_connector *connector) static inline void intel_display_crc_init(struct drm_i915_private *dev_priv) {} #endif -/* i915_gpu_error.c */ -#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR) - -__printf(2, 3) -void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...); -int i915_error_state_to_str(struct drm_i915_error_state_buf *estr, - const struct i915_gpu_state *gpu); -int i915_error_state_buf_init(struct drm_i915_error_state_buf *eb, - struct drm_i915_private *i915, - size_t count, loff_t pos); -static inline void i915_error_state_buf_release( - struct drm_i915_error_state_buf *eb) -{ - kfree(eb->buf); -} - -struct i915_gpu_state *i915_capture_gpu_state(struct drm_i915_private *i915); -void i915_capture_error_state(struct drm_i915_private *dev_priv, - u32 engine_mask, - const char *error_msg); - -static inline struct i915_gpu_state * -i915_gpu_state_get(struct i915_gpu_state *gpu) -{ - kref_get(&gpu->ref); - return gpu; -} - -void __i915_gpu_state_free(struct kref *kref); -static inline void i915_gpu_state_put(struct i915_gpu_state *gpu) -{ - if (gpu) - kref_put(&gpu->ref, __i915_gpu_state_free); -} - -struct i915_gpu_state *i915_first_error_state(struct drm_i915_private *i915); -void i915_reset_error_state(struct drm_i915_private *i915); - -#else - -static inline void i915_capture_error_state(struct drm_i915_private *dev_priv, - u32 engine_mask, - const char *error_msg) -{ -} - -static inline struct i915_gpu_state * -i915_first_error_state(struct drm_i915_private *i915) -{ - return NULL; -} - -static inline void i915_reset_error_state(struct drm_i915_private *i915) -{ -} - -#endif - const char *i915_cache_level_str(struct drm_i915_private *i915, int type); /* i915_cmd_parser.c */ diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index f89ac7a8..effaf98 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -32,6 +32,7 @@ #include #include +#include "i915_gpu_error.h" #include "i915_drv.h" static inline const struct intel_engine_cs * diff --git a/drivers/gpu/drm/i915/i915_gpu_error.h b/drivers/gpu/drm/i915/i915_gpu_error.h new file mode 100644 index 0000000..ebbdf37 --- /dev/null +++ b/drivers/gpu/drm/i915/i915_gpu_error.h @@ -0,0 +1,356 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright � 2008-2018 Intel Corporation + */ + +#ifndef _I915_GPU_ERROR_H_ +#define _I915_GPU_ERROR_H_ + +#include +#include +#include + +#include + +#include "intel_device_info.h" +#include "intel_ringbuffer.h" +#include "intel_uc_fw.h" + +#include "i915_gem.h" +#include "i915_gem_gtt.h" +#include "i915_params.h" + +struct drm_i915_private; +struct intel_overlay_error_state; +struct intel_display_error_state; + +struct i915_gpu_state { + struct kref ref; + ktime_t time; + ktime_t boottime; + ktime_t uptime; + + struct drm_i915_private *i915; + + char error_msg[128]; + bool simulated; + bool awake; + bool wakelock; + bool suspended; + int iommu; + u32 reset_count; + u32 suspend_count; + struct intel_device_info device_info; + struct intel_driver_caps driver_caps; + struct i915_params params; + + struct i915_error_uc { + struct intel_uc_fw guc_fw; + struct intel_uc_fw huc_fw; + struct drm_i915_error_object *guc_log; + } uc; + + /* Generic register state */ + u32 eir; + u32 pgtbl_er; + u32 ier; + u32 gtier[4], ngtier; + u32 ccid; + u32 derrmr; + u32 forcewake; + u32 error; /* gen6+ */ + u32 err_int; /* gen7 */ + u32 fault_data0; /* gen8, gen9 */ + u32 fault_data1; /* gen8, gen9 */ + u32 done_reg; + u32 gac_eco; + u32 gam_ecochk; + u32 gab_ctl; + u32 gfx_mode; + + u32 nfence; + u64 fence[I915_MAX_NUM_FENCES]; + struct intel_overlay_error_state *overlay; + struct intel_display_error_state *display; + + struct drm_i915_error_engine { + int engine_id; + /* Software tracked state */ + bool idle; + bool waiting; + int num_waiters; + unsigned long hangcheck_timestamp; + bool hangcheck_stalled; + enum intel_engine_hangcheck_action hangcheck_action; + struct i915_address_space *vm; + int num_requests; + u32 reset_count; + + /* position of active request inside the ring */ + u32 rq_head, rq_post, rq_tail; + + /* our own tracking of ring head and tail */ + u32 cpu_ring_head; + u32 cpu_ring_tail; + + u32 last_seqno; + + /* Register state */ + u32 start; + u32 tail; + u32 head; + u32 ctl; + u32 mode; + u32 hws; + u32 ipeir; + u32 ipehr; + u32 bbstate; + u32 instpm; + u32 instps; + u32 seqno; + u64 bbaddr; + u64 acthd; + u32 fault_reg; + u64 faddr; + u32 rc_psmi; /* sleep state */ + u32 semaphore_mboxes[I915_NUM_ENGINES - 1]; + struct intel_instdone instdone; + + struct drm_i915_error_context { + char comm[TASK_COMM_LEN]; + pid_t pid; + u32 handle; + u32 hw_id; + int priority; + int ban_score; + int active; + int guilty; + bool bannable; + } context; + + struct drm_i915_error_object { + u64 gtt_offset; + u64 gtt_size; + int page_count; + int unused; + u32 *pages[0]; + } *ringbuffer, *batchbuffer, *wa_batchbuffer, *ctx, *hws_page; + + struct drm_i915_error_object **user_bo; + long user_bo_count; + + struct drm_i915_error_object *wa_ctx; + struct drm_i915_error_object *default_state; + + struct drm_i915_error_request { + long jiffies; + pid_t pid; + u32 context; + int priority; + int ban_score; + u32 seqno; + u32 head; + u32 tail; + } *requests, execlist[EXECLIST_MAX_PORTS]; + unsigned int num_ports; + + struct drm_i915_error_waiter { + char comm[TASK_COMM_LEN]; + pid_t pid; + u32 seqno; + } *waiters; + + struct { + u32 gfx_mode; + union { + u64 pdp[4]; + u32 pp_dir_base; + }; + } vm_info; + } engine[I915_NUM_ENGINES]; + + struct drm_i915_error_buffer { + u32 size; + u32 name; + u32 rseqno[I915_NUM_ENGINES], wseqno; + u64 gtt_offset; + u32 read_domains; + u32 write_domain; + s32 fence_reg:I915_MAX_NUM_FENCE_BITS; + u32 tiling:2; + u32 dirty:1; + u32 purgeable:1; + u32 userptr:1; + s32 engine:4; + u32 cache_level:3; + } *active_bo[I915_NUM_ENGINES], *pinned_bo; + u32 active_bo_count[I915_NUM_ENGINES], pinned_bo_count; + struct i915_address_space *active_vm[I915_NUM_ENGINES]; +}; + +struct i915_gpu_error { + /* For hangcheck timer */ +#define DRM_I915_HANGCHECK_PERIOD 1500 /* in ms */ +#define DRM_I915_HANGCHECK_JIFFIES msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD) + + struct delayed_work hangcheck_work; + + /* For reset and error_state handling. */ + spinlock_t lock; + /* Protected by the above dev->gpu_error.lock. */ + struct i915_gpu_state *first_error; + + atomic_t pending_fb_pin; + + unsigned long missed_irq_rings; + + /** + * State variable controlling the reset flow and count + * + * This is a counter which gets incremented when reset is triggered, + * + * Before the reset commences, the I915_RESET_BACKOFF bit is set + * meaning that any waiters holding onto the struct_mutex should + * relinquish the lock immediately in order for the reset to start. + * + * If reset is not completed successfully, the I915_WEDGE bit is + * set meaning that hardware is terminally sour and there is no + * recovery. All waiters on the reset_queue will be woken when + * that happens. + * + * This counter is used by the wait_seqno code to notice that reset + * event happened and it needs to restart the entire ioctl (since most + * likely the seqno it waited for won't ever signal anytime soon). + * + * This is important for lock-free wait paths, where no contended lock + * naturally enforces the correct ordering between the bail-out of the + * waiter and the gpu reset work code. + */ + unsigned long reset_count; + + /** + * flags: Control various stages of the GPU reset + * + * #I915_RESET_BACKOFF - When we start a reset, we want to stop any + * other users acquiring the struct_mutex. To do this we set the + * #I915_RESET_BACKOFF bit in the error flags when we detect a reset + * and then check for that bit before acquiring the struct_mutex (in + * i915_mutex_lock_interruptible()?). I915_RESET_BACKOFF serves a + * secondary role in preventing two concurrent global reset attempts. + * + * #I915_RESET_HANDOFF - To perform the actual GPU reset, we need the + * struct_mutex. We try to acquire the struct_mutex in the reset worker, + * but it may be held by some long running waiter (that we cannot + * interrupt without causing trouble). Once we are ready to do the GPU + * reset, we set the I915_RESET_HANDOFF bit and wakeup any waiters. If + * they already hold the struct_mutex and want to participate they can + * inspect the bit and do the reset directly, otherwise the worker + * waits for the struct_mutex. + * + * #I915_RESET_ENGINE[num_engines] - Since the driver doesn't need to + * acquire the struct_mutex to reset an engine, we need an explicit + * flag to prevent two concurrent reset attempts in the same engine. + * As the number of engines continues to grow, allocate the flags from + * the most significant bits. + * + * #I915_WEDGED - If reset fails and we can no longer use the GPU, + * we set the #I915_WEDGED bit. Prior to command submission, e.g. + * i915_request_alloc(), this bit is checked and the sequence + * aborted (with -EIO reported to userspace) if set. + */ + unsigned long flags; +#define I915_RESET_BACKOFF 0 +#define I915_RESET_HANDOFF 1 +#define I915_RESET_MODESET 2 +#define I915_WEDGED (BITS_PER_LONG - 1) +#define I915_RESET_ENGINE (I915_WEDGED - I915_NUM_ENGINES) + + /** Number of times an engine has been reset */ + u32 reset_engine_count[I915_NUM_ENGINES]; + + /** + * Waitqueue to signal when a hang is detected. Used to for waiters + * to release the struct_mutex for the reset to procede. + */ + wait_queue_head_t wait_queue; + + /** + * Waitqueue to signal when the reset has completed. Used by clients + * that wait for dev_priv->mm.wedged to settle. + */ + wait_queue_head_t reset_queue; + + /* For missed irq/seqno simulation. */ + unsigned long test_irq_rings; +}; + +struct drm_i915_error_state_buf { + struct drm_i915_private *i915; + unsigned int bytes; + unsigned int size; + int err; + u8 *buf; + loff_t start; + loff_t pos; +}; + +#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR) + +__printf(2, 3) +void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...); +int i915_error_state_to_str(struct drm_i915_error_state_buf *estr, + const struct i915_gpu_state *gpu); +int i915_error_state_buf_init(struct drm_i915_error_state_buf *eb, + struct drm_i915_private *i915, + size_t count, loff_t pos); + +static inline void +i915_error_state_buf_release(struct drm_i915_error_state_buf *eb) +{ + kfree(eb->buf); +} + +struct i915_gpu_state *i915_capture_gpu_state(struct drm_i915_private *i915); +void i915_capture_error_state(struct drm_i915_private *dev_priv, + u32 engine_mask, + const char *error_msg); + +static inline struct i915_gpu_state * +i915_gpu_state_get(struct i915_gpu_state *gpu) +{ + kref_get(&gpu->ref); + return gpu; +} + +void __i915_gpu_state_free(struct kref *kref); +static inline void i915_gpu_state_put(struct i915_gpu_state *gpu) +{ + if (gpu) + kref_put(&gpu->ref, __i915_gpu_state_free); +} + +struct i915_gpu_state *i915_first_error_state(struct drm_i915_private *i915); +void i915_reset_error_state(struct drm_i915_private *i915); + +#else + +static inline void i915_capture_error_state(struct drm_i915_private *dev_priv, + u32 engine_mask, + const char *error_msg) +{ +} + +static inline struct i915_gpu_state * +i915_first_error_state(struct drm_i915_private *i915) +{ + return NULL; +} + +static inline void i915_reset_error_state(struct drm_i915_private *i915) +{ +} + +#endif /* IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR) */ + +#endif /* _I915_GPU_ERROR_H_ */ -- cgit v1.1 From caa1fd660e6f6701255544a3586a498676527fa4 Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Thu, 8 Mar 2018 16:52:18 -0800 Subject: drm/i915/psr: Display WA 0884 applied broadly for more HW tracking. WA 0884:bxt:all,cnl:*:A - "When FBC is enabled with eDP PSR, the CPU host modify writes may not get updated on the Display as expected. WA: Write 0x00000000 to CUR_SURFLIVE_A with every CPU host modify write to trigger PSR exit." We can also find on spec other cases where they describe bogus writes to cursor registers to force PSR exit with HW tracking. And it was confirmed by HW engineers that this Wa can be safely applied for any frontbuffer activity. So let's use this more and more here instead of forcibly disable and re-enable PSR everytime that we have a simple reliable flush case. Other commits improve the fbcon/fbdev use a lot, but this approach is the only when where we can get a fully reliable console with no slowness or missed frames and PSR still enabled and active. v2: - Rebase on drm-tip - (DK) Add a comment to explain that WA tells about writing 0 to CUR_SURFLIVE_A but we write to CUR_SURFLIVE(pipe). v3: Wa doesn't work on PSR2. Cc: Dhinakaran Pandiyan Signed-off-by: Rodrigo Vivi Reviewed-by: Dhinakaran Pandiyan Link: https://patchwork.freedesktop.org/patch/msgid/20180309005218.26772-1-rodrigo.vivi@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 3 +++ drivers/gpu/drm/i915/intel_psr.c | 19 +++++++++++++++++-- 2 files changed, 20 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 9e76546..60febfb 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -6032,6 +6032,9 @@ enum { #define IVB_CURSOR_B_OFFSET 0x71080 #define IVB_CURSOR_C_OFFSET 0x72080 +#define _CUR_SURLIVE 0x700AC +#define CUR_SURLIVE(pipe) _CURSOR2(pipe, _CUR_SURLIVE) + /* Display A control */ #define _DSPACNTR 0x70180 #define DISPLAY_PLANE_ENABLE (1<<31) diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index 23175c5..975ebb5 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -1027,8 +1027,23 @@ void intel_psr_flush(struct drm_i915_private *dev_priv, dev_priv->psr.busy_frontbuffer_bits &= ~frontbuffer_bits; /* By definition flush = invalidate + flush */ - if (frontbuffer_bits) - intel_psr_exit(dev_priv); + if (frontbuffer_bits) { + if (dev_priv->psr.psr2_support || + IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { + intel_psr_exit(dev_priv); + } else { + /* + * Display WA #0884: all + * This documented WA for bxt can be safely applied + * broadly so we can force HW tracking to exit PSR + * instead of disabling and re-enabling. + * Workaround tells us to write 0 to CUR_SURLIVE_A, + * but it makes more sense write to the current active + * pipe. + */ + I915_WRITE(CUR_SURLIVE(pipe), 0); + } + } if (!dev_priv->psr.active && !dev_priv->psr.busy_frontbuffer_bits) if (!work_busy(&dev_priv->psr.work.work)) -- cgit v1.1 From 6f9ec414ec47eea3f3e2c5ad4c67b4265bbff2a3 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 8 Mar 2018 14:07:32 +0000 Subject: drm/i915: Remove the impedance mismatch around intel_engine_enable_signaling There is some redundancy between dma_fence->ops->enable_signaling (via i915_fence_enable_signaling) and our backend, intel_engine_enable_signaling() in that both levels recheck the fence status multiple times. If we convert intel_engine_enable_signaling() to return the information desired by dma_fence->ops->enable_signaling, we can reduce i915_fence_enable_signaling to a simple stub and avoid trying to reinterpret the same information. Signed-off-by: Chris Wilson Cc: Tvrtko Ursulin Cc: Mika Kuoppala Cc: Michal Winiarski Reviewed-by: Tvrtko Ursulin Link: https://patchwork.freedesktop.org/patch/msgid/20180308140732.25090-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_request.c | 6 +----- drivers/gpu/drm/i915/intel_breadcrumbs.c | 21 +++++++++++++-------- drivers/gpu/drm/i915/intel_ringbuffer.h | 2 +- 3 files changed, 15 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index 2f62acd..1810fa1 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -59,11 +59,7 @@ static bool i915_fence_signaled(struct dma_fence *fence) static bool i915_fence_enable_signaling(struct dma_fence *fence) { - if (i915_fence_signaled(fence)) - return false; - - intel_engine_enable_signaling(to_request(fence), true); - return !i915_fence_signaled(fence); + return intel_engine_enable_signaling(to_request(fence), true); } static signed long i915_fence_wait(struct dma_fence *fence, diff --git a/drivers/gpu/drm/i915/intel_breadcrumbs.c b/drivers/gpu/drm/i915/intel_breadcrumbs.c index 1f79e7a..671a6d6 100644 --- a/drivers/gpu/drm/i915/intel_breadcrumbs.c +++ b/drivers/gpu/drm/i915/intel_breadcrumbs.c @@ -730,10 +730,11 @@ static void insert_signal(struct intel_breadcrumbs *b, list_add(&request->signaling.link, &iter->signaling.link); } -void intel_engine_enable_signaling(struct i915_request *request, bool wakeup) +bool intel_engine_enable_signaling(struct i915_request *request, bool wakeup) { struct intel_engine_cs *engine = request->engine; struct intel_breadcrumbs *b = &engine->breadcrumbs; + struct intel_wait *wait = &request->signaling.wait; u32 seqno; /* @@ -750,12 +751,12 @@ void intel_engine_enable_signaling(struct i915_request *request, bool wakeup) seqno = i915_request_global_seqno(request); if (!seqno) /* will be enabled later upon execution */ - return; + return true; - GEM_BUG_ON(request->signaling.wait.seqno); - request->signaling.wait.tsk = b->signaler; - request->signaling.wait.request = request; - request->signaling.wait.seqno = seqno; + GEM_BUG_ON(wait->seqno); + wait->tsk = b->signaler; + wait->request = request; + wait->seqno = seqno; /* * Add ourselves into the list of waiters, but registering our @@ -768,11 +769,15 @@ void intel_engine_enable_signaling(struct i915_request *request, bool wakeup) */ spin_lock(&b->rb_lock); insert_signal(b, request, seqno); - wakeup &= __intel_engine_add_wait(engine, &request->signaling.wait); + wakeup &= __intel_engine_add_wait(engine, wait); spin_unlock(&b->rb_lock); - if (wakeup) + if (wakeup) { wake_up_process(b->signaler); + return !intel_wait_complete(wait); + } + + return true; } void intel_engine_cancel_signaling(struct i915_request *request) diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index c31258d..81cdbbf 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -940,7 +940,7 @@ bool intel_engine_add_wait(struct intel_engine_cs *engine, struct intel_wait *wait); void intel_engine_remove_wait(struct intel_engine_cs *engine, struct intel_wait *wait); -void intel_engine_enable_signaling(struct i915_request *request, bool wakeup); +bool intel_engine_enable_signaling(struct i915_request *request, bool wakeup); void intel_engine_cancel_signaling(struct i915_request *request); static inline bool intel_engine_has_waiter(const struct intel_engine_cs *engine) -- cgit v1.1 From 3c33fc7c1af9a3426eff9015e5bab08a21a5fa9d Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 12 Mar 2018 13:03:06 +0000 Subject: drm/i915/uc: Sanitize uC options early We are sanitizing uC related modparams together with other driver modparams in intel_sanitize_options called from i915_driver_init_hw, but this is too late for us as we will want to use USES_GUC/USES_HUC macros at earlier stage. Since our sanitizing does not require any MMIO access, we can do it in intel_uc_init_early right after we resolve firmware names. Signed-off-by: Michal Wajdeczko Cc: Sagar Arun Kamble Cc: Chris Wilson Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180312130308.22952-2-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/i915_drv.c | 2 -- drivers/gpu/drm/i915/intel_uc.c | 6 ++++-- drivers/gpu/drm/i915/intel_uc.h | 1 - 3 files changed, 4 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 987c677..0126b22 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -1074,8 +1074,6 @@ static void intel_sanitize_options(struct drm_i915_private *dev_priv) i915_modparams.enable_ppgtt); DRM_DEBUG_DRIVER("ppgtt mode: %i\n", i915_modparams.enable_ppgtt); - intel_uc_sanitize_options(dev_priv); - intel_gvt_sanitize_options(dev_priv); } diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 1c1a00d..6dec6d6 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -83,7 +83,7 @@ static int __get_default_guc_log_level(struct drm_i915_private *dev_priv) } /** - * intel_uc_sanitize_options - sanitize uC related modparam options + * sanitize_options_early - sanitize uC related modparam options * @dev_priv: device private * * In case of "enable_guc" option this function will attempt to modify @@ -99,7 +99,7 @@ static int __get_default_guc_log_level(struct drm_i915_private *dev_priv) * unless GuC is enabled on given platform and the driver is compiled with * debug config when this modparam will default to "enable(1..4)". */ -void intel_uc_sanitize_options(struct drm_i915_private *dev_priv) +static void sanitize_options_early(struct drm_i915_private *dev_priv) { struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; struct intel_uc_fw *huc_fw = &dev_priv->huc.fw; @@ -163,6 +163,8 @@ void intel_uc_init_early(struct drm_i915_private *dev_priv) { intel_guc_init_early(&dev_priv->guc); intel_huc_init_early(&dev_priv->huc); + + sanitize_options_early(dev_priv); } void intel_uc_init_fw(struct drm_i915_private *dev_priv) diff --git a/drivers/gpu/drm/i915/intel_uc.h b/drivers/gpu/drm/i915/intel_uc.h index d6af984..49b5b2f 100644 --- a/drivers/gpu/drm/i915/intel_uc.h +++ b/drivers/gpu/drm/i915/intel_uc.h @@ -28,7 +28,6 @@ #include "intel_huc.h" #include "i915_params.h" -void intel_uc_sanitize_options(struct drm_i915_private *dev_priv); void intel_uc_init_early(struct drm_i915_private *dev_priv); void intel_uc_init_mmio(struct drm_i915_private *dev_priv); int intel_uc_register(struct drm_i915_private *dev_priv); -- cgit v1.1 From c37d57282033067edf60e044229a4b4f367cc81b Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 12 Mar 2018 13:03:07 +0000 Subject: drm/i915/uc: Sanitize uC together with GEM Instead of dancing around uC on reset/suspend/resume scenarios, explicitly sanitize uC when we sanitize GEM to force uC reload and start from known beginning. v2: don't forget about reset path (Daniele) sanitize uc before gem initiated full reset (Daniele) v3: drop redundant disable_communication in init_hw (Daniele) Signed-off-by: Michal Wajdeczko Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Chris Wilson Cc: Michel Thierry Reviewed-by: Daniele Ceraolo Spurio Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180312130308.22952-3-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/i915_gem.c | 2 ++ drivers/gpu/drm/i915/intel_guc.h | 6 ++++++ drivers/gpu/drm/i915/intel_huc.h | 6 ++++++ drivers/gpu/drm/i915/intel_uc.c | 19 ++++++++++++++++++- drivers/gpu/drm/i915/intel_uc.h | 1 + drivers/gpu/drm/i915/intel_uc_fw.h | 6 ++++++ 6 files changed, 39 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index e58b741..05b0724 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -2998,6 +2998,7 @@ int i915_gem_reset_prepare(struct drm_i915_private *dev_priv) } i915_gem_revoke_fences(dev_priv); + intel_uc_sanitize(dev_priv); return err; } @@ -4978,6 +4979,7 @@ int i915_gem_suspend(struct drm_i915_private *dev_priv) * machines is a good idea, we don't - just in case it leaves the * machine in an unusable condition. */ + intel_uc_sanitize(dev_priv); i915_gem_sanitize(dev_priv); intel_runtime_pm_put(dev_priv); diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h index 6d5aebe..d878160 100644 --- a/drivers/gpu/drm/i915/intel_guc.h +++ b/drivers/gpu/drm/i915/intel_guc.h @@ -133,4 +133,10 @@ int intel_guc_resume(struct intel_guc *guc); struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size); u32 intel_guc_wopcm_size(struct drm_i915_private *dev_priv); +static inline int intel_guc_sanitize(struct intel_guc *guc) +{ + intel_uc_fw_sanitize(&guc->fw); + return 0; +} + #endif diff --git a/drivers/gpu/drm/i915/intel_huc.h b/drivers/gpu/drm/i915/intel_huc.h index 5d6e804..b185850 100644 --- a/drivers/gpu/drm/i915/intel_huc.h +++ b/drivers/gpu/drm/i915/intel_huc.h @@ -38,4 +38,10 @@ struct intel_huc { void intel_huc_init_early(struct intel_huc *huc); int intel_huc_auth(struct intel_huc *huc); +static inline int intel_huc_sanitize(struct intel_huc *huc) +{ + intel_uc_fw_sanitize(&huc->fw); + return 0; +} + #endif diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 6dec6d6..9d5ffd7 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -334,6 +334,24 @@ void intel_uc_fini(struct drm_i915_private *dev_priv) intel_guc_fini(guc); } +void intel_uc_sanitize(struct drm_i915_private *i915) +{ + struct intel_guc *guc = &i915->guc; + struct intel_huc *huc = &i915->huc; + + if (!USES_GUC(i915)) + return; + + GEM_BUG_ON(!HAS_GUC(i915)); + + guc_disable_communication(guc); + + intel_huc_sanitize(huc); + intel_guc_sanitize(guc); + + __intel_uc_reset_hw(i915); +} + int intel_uc_init_hw(struct drm_i915_private *dev_priv) { struct intel_guc *guc = &dev_priv->guc; @@ -345,7 +363,6 @@ int intel_uc_init_hw(struct drm_i915_private *dev_priv) GEM_BUG_ON(!HAS_GUC(dev_priv)); - guc_disable_communication(guc); gen9_reset_guc_interrupts(dev_priv); /* init WOPCM */ diff --git a/drivers/gpu/drm/i915/intel_uc.h b/drivers/gpu/drm/i915/intel_uc.h index 49b5b2f..0a2b413 100644 --- a/drivers/gpu/drm/i915/intel_uc.h +++ b/drivers/gpu/drm/i915/intel_uc.h @@ -36,6 +36,7 @@ void intel_uc_init_fw(struct drm_i915_private *dev_priv); void intel_uc_fini_fw(struct drm_i915_private *dev_priv); int intel_uc_init_misc(struct drm_i915_private *dev_priv); void intel_uc_fini_misc(struct drm_i915_private *dev_priv); +void intel_uc_sanitize(struct drm_i915_private *dev_priv); int intel_uc_init_hw(struct drm_i915_private *dev_priv); void intel_uc_fini_hw(struct drm_i915_private *dev_priv); int intel_uc_init(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/intel_uc_fw.h b/drivers/gpu/drm/i915/intel_uc_fw.h index d5fd460..2601521 100644 --- a/drivers/gpu/drm/i915/intel_uc_fw.h +++ b/drivers/gpu/drm/i915/intel_uc_fw.h @@ -115,6 +115,12 @@ static inline bool intel_uc_fw_is_selected(struct intel_uc_fw *uc_fw) return uc_fw->path != NULL; } +static inline void intel_uc_fw_sanitize(struct intel_uc_fw *uc_fw) +{ + if (uc_fw->load_status == INTEL_UC_FIRMWARE_SUCCESS) + uc_fw->load_status = INTEL_UC_FIRMWARE_PENDING; +} + void intel_uc_fw_fetch(struct drm_i915_private *dev_priv, struct intel_uc_fw *uc_fw); int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, -- cgit v1.1 From 7aa0b14ede643fb7c33aaa8e0041de04a0d6f278 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 13 Mar 2018 00:40:54 +0000 Subject: drm/i915: Remove variable length arrays from sseu debugfs printers In order to enable -Wvla to prevent new variable length arrays being used in i915.ko, we first must remove the existing VLA. Inside i915_print_sseu_info(), VLA are used as the actual size of the sseu depends on platform. Replace the VLA with the maximum required. Signed-off-by: Chris Wilson Cc: Lionel Landwerlin Cc: Tvrtko Ursulin Cc: Matthew Auld Reviewed-by: Tvrtko Ursulin Link: https://patchwork.freedesktop.org/patch/msgid/20180313004055.25411-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_debugfs.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index c4cc8fe..0eac7dc 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -4312,9 +4312,10 @@ DEFINE_SIMPLE_ATTRIBUTE(i915_cache_sharing_fops, static void cherryview_sseu_device_status(struct drm_i915_private *dev_priv, struct sseu_dev_info *sseu) { - int ss_max = 2; +#define SS_MAX 2 + const int ss_max = SS_MAX; + u32 sig1[SS_MAX], sig2[SS_MAX]; int ss; - u32 sig1[ss_max], sig2[ss_max]; sig1[0] = I915_READ(CHV_POWER_SS0_SIG1); sig1[1] = I915_READ(CHV_POWER_SS1_SIG1); @@ -4338,15 +4339,15 @@ static void cherryview_sseu_device_status(struct drm_i915_private *dev_priv, sseu->eu_per_subslice = max_t(unsigned int, sseu->eu_per_subslice, eu_cnt); } +#undef SS_MAX } static void gen10_sseu_device_status(struct drm_i915_private *dev_priv, struct sseu_dev_info *sseu) { const struct intel_device_info *info = INTEL_INFO(dev_priv); + u32 s_reg[6], eu_reg[2 * 4], eu_mask[2]; int s, ss; - u32 s_reg[info->sseu.max_slices]; - u32 eu_reg[2 * info->sseu.max_subslices], eu_mask[2]; for (s = 0; s < info->sseu.max_slices; s++) { /* @@ -4399,9 +4400,8 @@ static void gen9_sseu_device_status(struct drm_i915_private *dev_priv, struct sseu_dev_info *sseu) { const struct intel_device_info *info = INTEL_INFO(dev_priv); + u32 s_reg[3], eu_reg[2 * 4], eu_mask[2]; int s, ss; - u32 s_reg[info->sseu.max_slices]; - u32 eu_reg[2 * info->sseu.max_subslices], eu_mask[2]; for (s = 0; s < info->sseu.max_slices; s++) { s_reg[s] = I915_READ(GEN9_SLICE_PGCTL_ACK(s)); -- cgit v1.1 From c5c2b11894f4f862cf243b955ac59bb1a5fe61b9 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 13 Mar 2018 00:40:55 +0000 Subject: drm/i915: Warn against variable length arrays VLA are strongly discouraged in the kernel due to ambiguity they impose on the limited stack space and security concerns over manipulating the stack frame. Add -Wvla to our compiler flags so that CI rejects them. Signed-off-by: Chris Wilson Cc: Jani Nikula Cc: Joonas Lahtinen Acked-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20180313004055.25411-2-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile index 4eee91a..fcb8a7b 100644 --- a/drivers/gpu/drm/i915/Makefile +++ b/drivers/gpu/drm/i915/Makefile @@ -12,7 +12,7 @@ # Note the danger in using -Wall -Wextra is that when CI updates gcc we # will most likely get a sudden build breakage... Hopefully we will fix # new warnings before CI updates! -subdir-ccflags-y := -Wall -Wextra +subdir-ccflags-y := -Wall -Wextra -Wvla subdir-ccflags-y += $(call cc-disable-warning, unused-parameter) subdir-ccflags-y += $(call cc-disable-warning, type-limits) subdir-ccflags-y += $(call cc-disable-warning, missing-field-initializers) -- cgit v1.1 From c7fb3c6c1893fddbbd39e13066489050c29397c1 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 13 Mar 2018 11:31:49 +0000 Subject: drm/i915: Use sseu size for determining eu_regs[] eu_regs[] is written 2*max_slices times (like s_reg[]) but oddly read 2*max_slices + max_subslices/2 times. Allocate the array large enough for the writes to avoid overwriting our stack and worry about the logic later. Fixes: 7aa0b14ede64 ("drm/i915: Remove variable length arrays from sseu debugfs printers") Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105479 Signed-off-by: Chris Wilson Cc: Lionel Landwerlin Cc: Tvrtko Ursulin Cc: Matthew Auld Reviewed-by: Lionel Landwerlin Link: https://patchwork.freedesktop.org/patch/msgid/20180313113149.1094-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_debugfs.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 0eac7dc..bc3f7d5 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -4345,8 +4345,9 @@ static void cherryview_sseu_device_status(struct drm_i915_private *dev_priv, static void gen10_sseu_device_status(struct drm_i915_private *dev_priv, struct sseu_dev_info *sseu) { +#define SS_MAX 6 const struct intel_device_info *info = INTEL_INFO(dev_priv); - u32 s_reg[6], eu_reg[2 * 4], eu_mask[2]; + u32 s_reg[SS_MAX], eu_reg[2 * SS_MAX], eu_mask[2]; int s, ss; for (s = 0; s < info->sseu.max_slices; s++) { @@ -4394,13 +4395,15 @@ static void gen10_sseu_device_status(struct drm_i915_private *dev_priv, eu_cnt); } } +#undef SS_MAX } static void gen9_sseu_device_status(struct drm_i915_private *dev_priv, struct sseu_dev_info *sseu) { +#define SS_MAX 3 const struct intel_device_info *info = INTEL_INFO(dev_priv); - u32 s_reg[3], eu_reg[2 * 4], eu_mask[2]; + u32 s_reg[SS_MAX], eu_reg[2 * SS_MAX], eu_mask[2]; int s, ss; for (s = 0; s < info->sseu.max_slices; s++) { @@ -4448,6 +4451,7 @@ static void gen9_sseu_device_status(struct drm_i915_private *dev_priv, eu_cnt); } } +#undef SS_MAX } static void broadwell_sseu_device_status(struct drm_i915_private *dev_priv, -- cgit v1.1 From 07bcd99b80477cc4f1b878afb3dec26877fa0ed0 Mon Sep 17 00:00:00 2001 From: Dhinakaran Pandiyan Date: Tue, 6 Mar 2018 19:34:18 -0800 Subject: drm/i915/frontbuffer: Pull frontbuffer_flush out of gem_obj_pin_to_display MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit i915_gem_obj_pin_to_display() calls frontbuffer_flush with origin set to DIRTYFB. The callers however are at a vantage point to decide if hardware frontbuffer tracking can do the flush for us. For example, legacy cursor updates, like flips, write to MMIO registers, which then triggers PSR flush by the hardware. Moving frontbuffer_flush out will enable us to skip a software initiated flush by setting origin to FLIP. Thanks to Chris for the idea. v2: Rebased due to Ville adding intel_plane_pin_fb(). Minor code reordering as fb_obj_flush doesn't need struct_mutex (Chris) Cc: Chris Wilson Cc: Ville Syrjälä Cc: Paulo Zanoni Signed-off-by: Dhinakaran Pandiyan Reviewed-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180307033420.3086-1-dhinakaran.pandiyan@intel.com --- drivers/gpu/drm/i915/i915_gem.c | 9 ++++----- drivers/gpu/drm/i915/intel_display.c | 9 +++++++-- drivers/gpu/drm/i915/intel_fbdev.c | 5 +++-- drivers/gpu/drm/i915/intel_overlay.c | 1 + 4 files changed, 15 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 05b0724..58f8cf7 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -4094,9 +4094,10 @@ out: } /* - * Prepare buffer for display plane (scanout, cursors, etc). - * Can be called from an uninterruptible phase (modesetting) and allows - * any flushes to be pipelined (for pageflips). + * Prepare buffer for display plane (scanout, cursors, etc). Can be called from + * an uninterruptible phase (modesetting) and allows any flushes to be pipelined + * (for pageflips). We only flush the caches while preparing the buffer for + * display, the callers are responsible for frontbuffer flush. */ struct i915_vma * i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj, @@ -4152,9 +4153,7 @@ i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj, vma->display_alignment = max_t(u64, vma->display_alignment, alignment); - /* Treat this as an end-of-frame, like intel_user_framebuffer_dirty() */ __i915_gem_object_flush_for_display(obj); - intel_fb_obj_flush(obj, ORIGIN_DIRTYFB); /* It should now be out of any other write domains, and we can update * the domain values for our changes. diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index f424fff..1b2a402 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -2858,6 +2858,9 @@ valid_fb: return; } + obj = intel_fb_obj(fb); + intel_fb_obj_flush(obj, ORIGIN_DIRTYFB); + plane_state->src_x = 0; plane_state->src_y = 0; plane_state->src_w = fb->width << 16; @@ -2871,7 +2874,6 @@ valid_fb: intel_state->base.src = drm_plane_state_src(plane_state); intel_state->base.dst = drm_plane_state_dest(plane_state); - obj = intel_fb_obj(fb); if (i915_gem_object_is_tiled(obj)) dev_priv->preserve_bios_swizzle = true; @@ -12793,6 +12795,8 @@ intel_prepare_plane_fb(struct drm_plane *plane, if (ret) return ret; + intel_fb_obj_flush(obj, ORIGIN_DIRTYFB); + if (!new_state->fence) { /* implicit fencing */ struct dma_fence *fence; @@ -13186,8 +13190,9 @@ intel_legacy_cursor_update(struct drm_plane *plane, if (ret) goto out_unlock; - old_fb = old_plane_state->fb; + intel_fb_obj_flush(intel_fb_obj(fb), ORIGIN_DIRTYFB); + old_fb = old_plane_state->fb; i915_gem_track_fb(intel_fb_obj(old_fb), intel_fb_obj(fb), intel_plane->frontbuffer_bit); diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c index 6f12adc..65a3313 100644 --- a/drivers/gpu/drm/i915/intel_fbdev.c +++ b/drivers/gpu/drm/i915/intel_fbdev.c @@ -221,6 +221,9 @@ static int intelfb_create(struct drm_fb_helper *helper, goto out_unlock; } + fb = &ifbdev->fb->base; + intel_fb_obj_flush(intel_fb_obj(fb), ORIGIN_DIRTYFB); + info = drm_fb_helper_alloc_fbi(helper); if (IS_ERR(info)) { DRM_ERROR("Failed to allocate fb_info\n"); @@ -230,8 +233,6 @@ static int intelfb_create(struct drm_fb_helper *helper, info->par = helper; - fb = &ifbdev->fb->base; - ifbdev->helper.fb = fb; strcpy(info->fix.id, "inteldrmfb"); diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c index 36671a9..c2f10d8 100644 --- a/drivers/gpu/drm/i915/intel_overlay.c +++ b/drivers/gpu/drm/i915/intel_overlay.c @@ -807,6 +807,7 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay, ret = PTR_ERR(vma); goto out_pin_section; } + intel_fb_obj_flush(new_bo, ORIGIN_DIRTYFB); ret = i915_vma_put_fence(vma); if (ret) -- cgit v1.1 From a694e226fbaefbf3101982e54ca2f014292c540f Mon Sep 17 00:00:00 2001 From: Dhinakaran Pandiyan Date: Tue, 6 Mar 2018 19:34:19 -0800 Subject: drm/i915/frontbuffer: HW tracking for cursor moves to fix PSR lags. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit DRM_IOCTL_MODE_CURSOR results in frontbuffer flush before the cursor plane MMIOs are written to. But this flush should not be necessary for PSR as hardware tracking triggers PSR exit when MMIOs are written. As for FBC, the spec says "Flips or changes to plane size and panning" cause FBC to be nuked. Use origin == ORIGIN_FLIP so that features can ignore cursor updates in their frontbuffer_flush implementations. /sys/kernel/debug/dri/0/i915_fbc_status shows "Compressing: yes" when I move the cursor around. v3: Use ORIGIN_FLIP now that pin_to_display does not flush frontbuffer. v2: Update comment in i915_gem_object_pin_to_display_plane. (Chris) Cc: Paulo Zanoni Cc: Ville Syrjälä Cc: Chris Wilson Cc: Rodrigo Vivi Signed-off-by: Dhinakaran Pandiyan Reviewed-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180307033420.3086-2-dhinakaran.pandiyan@intel.com --- drivers/gpu/drm/i915/intel_display.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 1b2a402..a7bfa23 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -13190,7 +13190,7 @@ intel_legacy_cursor_update(struct drm_plane *plane, if (ret) goto out_unlock; - intel_fb_obj_flush(intel_fb_obj(fb), ORIGIN_DIRTYFB); + intel_fb_obj_flush(intel_fb_obj(fb), ORIGIN_FLIP); old_fb = old_plane_state->fb; i915_gem_track_fb(intel_fb_obj(old_fb), intel_fb_obj(fb), -- cgit v1.1 From 5baf63cc4d7879474221c5a32e4c2adc7ed33add Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Tue, 6 Mar 2018 19:34:20 -0800 Subject: drm/i915/psr: Use more PSR HW tracking. So far we are using frontbuffer tracking for everything and ignoring that PSR has a HW capable HW tracking for many modern usages of GPU on Core platforms and newer Atom ones. One reason for that is that we were trying to keep same infrastructure in place for VLV/CHV than the rest of platforms. But also because when this infrastructure was created the front-buffer-tracking origin wasn't that good and stable how it is today after Paulo reworked it to attend FBC cases. However this PSR implementation without HW tracking died on gen8LP. And newer platforms are starting to demand more HW tracking specially with PSR2 cases in mind. By disabling and re-enabling PSR totally every time we believe someone is going to change the front buffer content we don't allow PSR HW tracking to do this job and specially compromising the whole idea of PSR2 case where the HW tracking detect only the damaged area and do a partial screen update. So, from now on, on the platforms that has hw_tracking let's rely more on HW tracking. This also is the case in used by other drivers and more validated by SV teams. So I hope that this will lead us to less misterious bugs. v2: Only do this for platform that actually has hw tracking. v3 from DK Do this only for flips, small gradual changes are better. Cc: Dhinakaran Pandiyan Cc: Jim Bride Cc: Vathsala Nagaraju Signed-off-by: Rodrigo Vivi Signed-off-by: Dhinakaran Pandiyan Reviewed-by: Jose Roberto de Souza Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180307033420.3086-3-dhinakaran.pandiyan@intel.com --- drivers/gpu/drm/i915/i915_drv.h | 1 + drivers/gpu/drm/i915/intel_drv.h | 3 ++- drivers/gpu/drm/i915/intel_frontbuffer.c | 2 +- drivers/gpu/drm/i915/intel_psr.c | 10 +++++++++- 4 files changed, 13 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index d35f805..74b0e9d 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -606,6 +606,7 @@ struct i915_psr { bool y_cord_support; bool colorimetry_support; bool alpm; + bool has_hw_tracking; void (*enable_source)(struct intel_dp *, const struct intel_crtc_state *); diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 83e5ca8..de6db91 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -1876,7 +1876,8 @@ void intel_psr_enable(struct intel_dp *intel_dp, void intel_psr_disable(struct intel_dp *intel_dp, const struct intel_crtc_state *old_crtc_state); void intel_psr_invalidate(struct drm_i915_private *dev_priv, - unsigned frontbuffer_bits); + unsigned frontbuffer_bits, + enum fb_op_origin origin); void intel_psr_flush(struct drm_i915_private *dev_priv, unsigned frontbuffer_bits, enum fb_op_origin origin); diff --git a/drivers/gpu/drm/i915/intel_frontbuffer.c b/drivers/gpu/drm/i915/intel_frontbuffer.c index 3a8d3d0..7fff0a0 100644 --- a/drivers/gpu/drm/i915/intel_frontbuffer.c +++ b/drivers/gpu/drm/i915/intel_frontbuffer.c @@ -80,7 +80,7 @@ void __intel_fb_obj_invalidate(struct drm_i915_gem_object *obj, } might_sleep(); - intel_psr_invalidate(dev_priv, frontbuffer_bits); + intel_psr_invalidate(dev_priv, frontbuffer_bits, origin); intel_edp_drrs_invalidate(dev_priv, frontbuffer_bits); intel_fbc_invalidate(dev_priv, frontbuffer_bits, origin); } diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index 975ebb5..a079b62 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -957,6 +957,7 @@ void intel_psr_single_frame_update(struct drm_i915_private *dev_priv, * intel_psr_invalidate - Invalidade PSR * @dev_priv: i915 device * @frontbuffer_bits: frontbuffer plane tracking bits + * @origin: which operation caused the invalidate * * Since the hardware frontbuffer tracking has gaps we need to integrate * with the software frontbuffer tracking. This function gets called every @@ -966,7 +967,7 @@ void intel_psr_single_frame_update(struct drm_i915_private *dev_priv, * Dirty frontbuffers relevant to PSR are tracked in busy_frontbuffer_bits." */ void intel_psr_invalidate(struct drm_i915_private *dev_priv, - unsigned frontbuffer_bits) + unsigned frontbuffer_bits, enum fb_op_origin origin) { struct drm_crtc *crtc; enum pipe pipe; @@ -974,6 +975,9 @@ void intel_psr_invalidate(struct drm_i915_private *dev_priv, if (!CAN_PSR(dev_priv)) return; + if (dev_priv->psr.has_hw_tracking && origin == ORIGIN_FLIP) + return; + mutex_lock(&dev_priv->psr.lock); if (!dev_priv->psr.enabled) { mutex_unlock(&dev_priv->psr.lock); @@ -1014,6 +1018,9 @@ void intel_psr_flush(struct drm_i915_private *dev_priv, if (!CAN_PSR(dev_priv)) return; + if (dev_priv->psr.has_hw_tracking && origin == ORIGIN_FLIP) + return; + mutex_lock(&dev_priv->psr.lock); if (!dev_priv->psr.enabled) { mutex_unlock(&dev_priv->psr.lock); @@ -1105,6 +1112,7 @@ void intel_psr_init(struct drm_i915_private *dev_priv) dev_priv->psr.activate = vlv_psr_activate; dev_priv->psr.setup_vsc = vlv_psr_setup_vsc; } else { + dev_priv->psr.has_hw_tracking = true; dev_priv->psr.enable_source = hsw_psr_enable_source; dev_priv->psr.disable_source = hsw_psr_disable; dev_priv->psr.enable_sink = hsw_psr_enable_sink; -- cgit v1.1 From be74229bd5456729a9e81dabc8aac2fb58a69492 Mon Sep 17 00:00:00 2001 From: Dhinakaran Pandiyan Date: Mon, 12 Mar 2018 21:42:11 -0700 Subject: drm/i915/psr: Remove PSR active flag from debugfs The flag becomes misleading with flips and cursor moves not modifying it's state as HW takes care of exiting PSR (when HW tracking is enabled) Cc: Rodrigo Vivi Signed-off-by: Dhinakaran Pandiyan Acked-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180313044211.27105-1-dhinakaran.pandiyan@intel.com --- drivers/gpu/drm/i915/i915_debugfs.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index bc3f7d5..972014b 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -2565,7 +2565,6 @@ static int i915_edp_psr_status(struct seq_file *m, void *data) mutex_lock(&dev_priv->psr.lock); seq_printf(m, "Enabled: %s\n", yesno((bool)dev_priv->psr.enabled)); - seq_printf(m, "Active: %s\n", yesno(dev_priv->psr.active)); seq_printf(m, "Busy frontbuffer bits: 0x%03x\n", dev_priv->psr.busy_frontbuffer_bits); seq_printf(m, "Re-enable work scheduled: %s\n", -- cgit v1.1 From a8ada068a5025d738c870851d023b80cf6be0c95 Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Mon, 12 Mar 2018 14:05:28 -0700 Subject: drm/i915: Move CUR SURFLIVE definition to a better place. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit No functional change. But let's keep definitions clean and cursor related register definitions together. v2: Fix caps x no caps on same reg. Change name to match original reg name. (by Ville). Also fix name on code s/surlive/surflive and on subject s/cur_surlife/cur surflive/. Suggested-by: Ville Syrjälä Cc: Ville Syrjälä Signed-off-by: Rodrigo Vivi Reviewed-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180312210528.7905-1-rodrigo.vivi@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 5 ++--- drivers/gpu/drm/i915/intel_psr.c | 4 ++-- 2 files changed, 4 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 60febfb..38d4be4 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -6009,6 +6009,7 @@ enum { #define CURSIZE _MMIO(0x700a0) /* 845/865 */ #define _CUR_FBC_CTL_A 0x700a0 /* ivb+ */ #define CUR_FBC_CTL_EN (1 << 31) +#define _CURASURFLIVE 0x700ac /* g4x+ */ #define _CURBCNTR 0x700c0 #define _CURBBASE 0x700c4 #define _CURBPOS 0x700c8 @@ -6025,6 +6026,7 @@ enum { #define CURBASE(pipe) _CURSOR2(pipe, _CURABASE) #define CURPOS(pipe) _CURSOR2(pipe, _CURAPOS) #define CUR_FBC_CTL(pipe) _CURSOR2(pipe, _CUR_FBC_CTL_A) +#define CURSURFLIVE(pipe) _CURSOR2(pipe, _CURASURFLIVE) #define CURSOR_A_OFFSET 0x70080 #define CURSOR_B_OFFSET 0x700c0 @@ -6032,9 +6034,6 @@ enum { #define IVB_CURSOR_B_OFFSET 0x71080 #define IVB_CURSOR_C_OFFSET 0x72080 -#define _CUR_SURLIVE 0x700AC -#define CUR_SURLIVE(pipe) _CURSOR2(pipe, _CUR_SURLIVE) - /* Display A control */ #define _DSPACNTR 0x70180 #define DISPLAY_PLANE_ENABLE (1<<31) diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index a079b62..317cb4a 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -1044,11 +1044,11 @@ void intel_psr_flush(struct drm_i915_private *dev_priv, * This documented WA for bxt can be safely applied * broadly so we can force HW tracking to exit PSR * instead of disabling and re-enabling. - * Workaround tells us to write 0 to CUR_SURLIVE_A, + * Workaround tells us to write 0 to CUR_SURFLIVE_A, * but it makes more sense write to the current active * pipe. */ - I915_WRITE(CUR_SURLIVE(pipe), 0); + I915_WRITE(CURSURFLIVE(pipe), 0); } } -- cgit v1.1 From 629820fcd0ddbb7955a37c075e82756da69ea908 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 9 Mar 2018 10:11:14 +0000 Subject: drm/i915: Show GEM_TRACE when detecting a failed GPU idle If we timeout waiting for the GPU to idle, something went seriously wrong. We currently dump the engine state, but we can also dump the ftrace buffer showing our last operations (when available). In passing, note that since commit 559e040f1f08 ("drm/i915: Show the GPU state when declaring wedged", we now show the engine state twice, once in detecting the failed idle and then again on declaring wedged. v2: ftrace_dump() takes a parameter specifying whether to dump all cpu buffers or the local cpu's. Signed-off-by: Chris Wilson Cc: Joonas Lahtinen Cc: Mika Kuoppala Reviewed-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/20180309101114.1138-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_gem.c | 11 +---------- drivers/gpu/drm/i915/i915_gem.h | 2 ++ 2 files changed, 3 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 58f8cf7..d0624c5 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -3672,16 +3672,7 @@ static int wait_for_engines(struct drm_i915_private *i915) if (wait_for(intel_engines_are_idle(i915), I915_IDLE_ENGINES_TIMEOUT)) { dev_err(i915->drm.dev, "Failed to idle engines, declaring wedged!\n"); - if (drm_debug & DRM_UT_DRIVER) { - struct drm_printer p = drm_debug_printer(__func__); - struct intel_engine_cs *engine; - enum intel_engine_id id; - - for_each_engine(engine, i915, id) - intel_engine_dump(engine, &p, - "%s\n", engine->name); - } - + GEM_TRACE_DUMP(); i915_gem_set_wedged(i915); return -EIO; } diff --git a/drivers/gpu/drm/i915/i915_gem.h b/drivers/gpu/drm/i915/i915_gem.h index f54c4ff..8922344 100644 --- a/drivers/gpu/drm/i915/i915_gem.h +++ b/drivers/gpu/drm/i915/i915_gem.h @@ -53,8 +53,10 @@ #if IS_ENABLED(CONFIG_DRM_I915_TRACE_GEM) #define GEM_TRACE(...) trace_printk(__VA_ARGS__) +#define GEM_TRACE_DUMP() ftrace_dump(DUMP_ALL) #else #define GEM_TRACE(...) do { } while (0) +#define GEM_TRACE_DUMP() do { } while (0) #endif #define I915_NUM_ENGINES 8 -- cgit v1.1 From 62801bf615679293957ecdf37cc093a18158a201 Mon Sep 17 00:00:00 2001 From: Dhinakaran Pandiyan Date: Mon, 12 Mar 2018 21:09:54 -0700 Subject: drm/i915/psr: Comment to clarify SRD_DEBUG is called PSR_MASK SKL+ What was called SRD_DEBUG(0x6F860) on HSW and BDW was renamed to PSR_MASK SKL onwards, add a note next to the macro definition. There is also a different PSR_DEBUG on SKL+ to add to the confusion. Signed-off-by: Dhinakaran Pandiyan Acked-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180313040954.6289-1-dhinakaran.pandiyan@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 38d4be4..761bd3a 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -4180,13 +4180,13 @@ enum { #define EDP_PSR_PERF_CNT _MMIO(dev_priv->psr_mmio_base + 0x44) #define EDP_PSR_PERF_CNT_MASK 0xffffff -#define EDP_PSR_DEBUG _MMIO(dev_priv->psr_mmio_base + 0x60) +#define EDP_PSR_DEBUG _MMIO(dev_priv->psr_mmio_base + 0x60) /* PSR_MASK on SKL+ */ #define EDP_PSR_DEBUG_MASK_MAX_SLEEP (1<<28) #define EDP_PSR_DEBUG_MASK_LPSP (1<<27) #define EDP_PSR_DEBUG_MASK_MEMUP (1<<26) #define EDP_PSR_DEBUG_MASK_HPD (1<<25) #define EDP_PSR_DEBUG_MASK_DISP_REG_WRITE (1<<16) -#define EDP_PSR_DEBUG_EXIT_ON_PIXEL_UNDERRUN (1<<15) +#define EDP_PSR_DEBUG_EXIT_ON_PIXEL_UNDERRUN (1<<15) /* SKL+ */ #define EDP_PSR2_CTL _MMIO(0x6f900) #define EDP_PSR2_ENABLE (1<<31) -- cgit v1.1 From 3c009e3c468a9c095343febca7de6397c0e78b7d Mon Sep 17 00:00:00 2001 From: Jackie Li Date: Tue, 13 Mar 2018 17:32:49 -0700 Subject: drm/i915/guc: Rename guc_ggtt_offset to intel_guc_ggtt_offset GuC related exported functions should start with "intel_guc_" prefix and pass intel_guc as the first parameter since its GuC related. Current guc_ggtt_offset() failed to follow this code convention and this is a problem for future patches that needs to access intel_guc data to verify the GGTT offset against the GuC WOPCM top. This patch renames the guc_ggtt_offset to intel_guc_ggtt_offset and updates the related code to pass intel_guc pointer to this function call, so that we can have a unified coding style for GuC code and also enable the future patches to get GuC related data from intel_guc to do the offset verification. Meanwhile, this patch also moves the GUC_GGTT_TOP from intel_guc_regs.h to intel_guc.h since it is not GuC register related definition. v8: - Fixed coding style issues and moved GUC_GGTT_TOP to intel_guc.h (Sagar) - Updated commit message to explain to reason and motivation to add intel_guc as the first parameter of intel_guc_ggtt_offset (Chris) v9: - Fixed code alignment issue due to line break (Chris) v10: - Removed unnecessary comments, redundant code and avoided reuse variable to avoid potential issues (Joonas) v13: - Updated the ordering of s-o-b/cc/r-b tags (Sagar) Signed-off-by: Jackie Li Cc: Michal Wajdeczko Cc: Sagar Arun Kamble Cc: Chris Wilson Cc: Joonas Lahtinen Reviewed-by: Sagar Arun Kamble (v8) Reviewed-by: Joonas Lahtinen (v9) Reviewed-by: Michal Wajdeczko (v11) Reviewed-by: Joonas Lahtinen (v12) Reviewed-by: Joonas Lahtinen Signed-off-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/1520987574-19351-1-git-send-email-yaodong.li@intel.com --- drivers/gpu/drm/i915/intel_guc.c | 11 ++++++----- drivers/gpu/drm/i915/intel_guc.h | 14 ++++++++++++-- drivers/gpu/drm/i915/intel_guc_ads.c | 8 ++++---- drivers/gpu/drm/i915/intel_guc_ct.c | 5 +++-- drivers/gpu/drm/i915/intel_guc_fw.c | 2 +- drivers/gpu/drm/i915/intel_guc_log.c | 2 +- drivers/gpu/drm/i915/intel_guc_reg.h | 3 --- drivers/gpu/drm/i915/intel_guc_submission.c | 10 +++++----- drivers/gpu/drm/i915/intel_huc.c | 3 ++- drivers/gpu/drm/i915/intel_huc_fw.c | 3 ++- 10 files changed, 36 insertions(+), 25 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index 25f9229..7846384 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -269,8 +269,9 @@ void intel_guc_init_params(struct intel_guc *guc) /* If GuC submission is enabled, set up additional parameters here */ if (USES_GUC_SUBMISSION(dev_priv)) { - u32 ads = guc_ggtt_offset(guc->ads_vma) >> PAGE_SHIFT; - u32 pgs = guc_ggtt_offset(dev_priv->guc.stage_desc_pool); + u32 ads = intel_guc_ggtt_offset(guc, + guc->ads_vma) >> PAGE_SHIFT; + u32 pgs = intel_guc_ggtt_offset(guc, guc->stage_desc_pool); u32 ctx_in_16 = GUC_MAX_STAGE_DESCRIPTORS / 16; params[GUC_CTL_DEBUG] |= ads << GUC_ADS_ADDR_SHIFT; @@ -447,7 +448,7 @@ int intel_guc_suspend(struct intel_guc *guc) u32 data[] = { INTEL_GUC_ACTION_ENTER_S_STATE, GUC_POWER_D1, /* any value greater than GUC_POWER_D0 */ - guc_ggtt_offset(guc->shared_data) + intel_guc_ggtt_offset(guc, guc->shared_data) }; return intel_guc_send(guc, data, ARRAY_SIZE(data)); @@ -471,7 +472,7 @@ int intel_guc_reset_engine(struct intel_guc *guc, data[3] = 0; data[4] = 0; data[5] = guc->execbuf_client->stage_id; - data[6] = guc_ggtt_offset(guc->shared_data); + data[6] = intel_guc_ggtt_offset(guc, guc->shared_data); return intel_guc_send(guc, data, ARRAY_SIZE(data)); } @@ -485,7 +486,7 @@ int intel_guc_resume(struct intel_guc *guc) u32 data[] = { INTEL_GUC_ACTION_EXIT_S_STATE, GUC_POWER_D0, - guc_ggtt_offset(guc->shared_data) + intel_guc_ggtt_offset(guc, guc->shared_data) }; return intel_guc_send(guc, data, ARRAY_SIZE(data)); diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h index d878160..a1be04e 100644 --- a/drivers/gpu/drm/i915/intel_guc.h +++ b/drivers/gpu/drm/i915/intel_guc.h @@ -100,13 +100,23 @@ static inline void intel_guc_notify(struct intel_guc *guc) guc->notify(guc); } -/* +/* GuC addresses above GUC_GGTT_TOP also don't map through the GTT */ +#define GUC_GGTT_TOP 0xFEE00000 + +/** + * intel_guc_ggtt_offset() - Get and validate the GGTT offset of @vma + * @guc: intel_guc structure. + * @vma: i915 graphics virtual memory area. + * * GuC does not allow any gfx GGTT address that falls into range [0, WOPCM_TOP), * which is reserved for Boot ROM, SRAM and WOPCM. Currently this top address is * 512K. In order to exclude 0-512K address space from GGTT, all gfx objects * used by GuC is pinned with PIN_OFFSET_BIAS along with size of WOPCM. + * + * Return: GGTT offset that meets the GuC gfx address requirement. */ -static inline u32 guc_ggtt_offset(struct i915_vma *vma) +static inline u32 intel_guc_ggtt_offset(struct intel_guc *guc, + struct i915_vma *vma) { u32 offset = i915_ggtt_offset(vma); diff --git a/drivers/gpu/drm/i915/intel_guc_ads.c b/drivers/gpu/drm/i915/intel_guc_ads.c index ac62753..334cb52 100644 --- a/drivers/gpu/drm/i915/intel_guc_ads.c +++ b/drivers/gpu/drm/i915/intel_guc_ads.c @@ -75,7 +75,7 @@ static void guc_policies_init(struct guc_policies *policies) int intel_guc_ads_create(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct i915_vma *vma; + struct i915_vma *vma, *kernel_ctx_vma; struct page *page; /* The ads obj includes the struct itself and buffers passed to GuC */ struct { @@ -121,9 +121,9 @@ int intel_guc_ads_create(struct intel_guc *guc) * to find it. Note that we have to skip our header (1 page), * because our GuC shared data is there. */ + kernel_ctx_vma = dev_priv->kernel_context->engine[RCS].state; blob->ads.golden_context_lrca = - guc_ggtt_offset(dev_priv->kernel_context->engine[RCS].state) + - skipped_offset; + intel_guc_ggtt_offset(guc, kernel_ctx_vma) + skipped_offset; /* * The GuC expects us to exclude the portion of the context image that @@ -135,7 +135,7 @@ int intel_guc_ads_create(struct intel_guc *guc) blob->ads.eng_state_size[engine->guc_id] = engine->context_size - skipped_size; - base = guc_ggtt_offset(vma); + base = intel_guc_ggtt_offset(guc, vma); blob->ads.scheduler_policies = base + ptr_offset(blob, policies); blob->ads.reg_state_buffer = base + ptr_offset(blob, reg_state_buffer); blob->ads.reg_state_addr = base + ptr_offset(blob, reg_state); diff --git a/drivers/gpu/drm/i915/intel_guc_ct.c b/drivers/gpu/drm/i915/intel_guc_ct.c index 24ad557..0a0d3d5 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/intel_guc_ct.c @@ -156,7 +156,8 @@ static int ctch_init(struct intel_guc *guc, err = PTR_ERR(blob); goto err_vma; } - DRM_DEBUG_DRIVER("CT: vma base=%#x\n", guc_ggtt_offset(ctch->vma)); + DRM_DEBUG_DRIVER("CT: vma base=%#x\n", + intel_guc_ggtt_offset(guc, ctch->vma)); /* store pointers to desc and cmds */ for (i = 0; i < ARRAY_SIZE(ctch->ctbs); i++) { @@ -202,7 +203,7 @@ static int ctch_open(struct intel_guc *guc, } /* vma should be already allocated and map'ed */ - base = guc_ggtt_offset(ctch->vma); + base = intel_guc_ggtt_offset(guc, ctch->vma); /* (re)initialize descriptors * cmds buffers are in the second half of the blob page diff --git a/drivers/gpu/drm/i915/intel_guc_fw.c b/drivers/gpu/drm/i915/intel_guc_fw.c index d07f2b9..978668c 100644 --- a/drivers/gpu/drm/i915/intel_guc_fw.c +++ b/drivers/gpu/drm/i915/intel_guc_fw.c @@ -165,7 +165,7 @@ static int guc_xfer_ucode(struct intel_guc *guc, struct i915_vma *vma) I915_WRITE(DMA_COPY_SIZE, guc_fw->header_size + guc_fw->ucode_size); /* Set the source address for the new blob */ - offset = guc_ggtt_offset(vma) + guc_fw->header_offset; + offset = intel_guc_ggtt_offset(guc, vma) + guc_fw->header_offset; I915_WRITE(DMA_ADDR_0_LOW, lower_32_bits(offset)); I915_WRITE(DMA_ADDR_0_HIGH, upper_32_bits(offset) & 0xFFFF); diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 90b395f..b9c7bd7 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -581,7 +581,7 @@ int intel_guc_log_create(struct intel_guc *guc) (GUC_LOG_ISR_PAGES << GUC_LOG_ISR_SHIFT) | (GUC_LOG_CRASH_PAGES << GUC_LOG_CRASH_SHIFT); - offset = guc_ggtt_offset(vma) >> PAGE_SHIFT; /* in pages */ + offset = intel_guc_ggtt_offset(guc, vma) >> PAGE_SHIFT; guc->log.flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags; return 0; diff --git a/drivers/gpu/drm/i915/intel_guc_reg.h b/drivers/gpu/drm/i915/intel_guc_reg.h index 19a9247..711e9e9 100644 --- a/drivers/gpu/drm/i915/intel_guc_reg.h +++ b/drivers/gpu/drm/i915/intel_guc_reg.h @@ -80,9 +80,6 @@ #define GUC_WOPCM_TOP (0x80 << 12) /* 512KB */ #define BXT_GUC_WOPCM_RC6_RESERVED (0x10 << 12) /* 64KB */ -/* GuC addresses above GUC_GGTT_TOP also don't map through the GTT */ -#define GUC_GGTT_TOP 0xFEE00000 - #define GEN8_GT_PM_CONFIG _MMIO(0x138140) #define GEN9LP_GT_PM_CONFIG _MMIO(0x138140) #define GEN9_GT_PM_CONFIG _MMIO(0x13816c) diff --git a/drivers/gpu/drm/i915/intel_guc_submission.c b/drivers/gpu/drm/i915/intel_guc_submission.c index 8a8ad2f..33af293 100644 --- a/drivers/gpu/drm/i915/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/intel_guc_submission.c @@ -386,8 +386,8 @@ static void guc_stage_desc_init(struct intel_guc *guc, lrc->context_desc = lower_32_bits(ce->lrc_desc); /* The state page is after PPHWSP */ - lrc->ring_lrca = - guc_ggtt_offset(ce->state) + LRC_STATE_PN * PAGE_SIZE; + lrc->ring_lrca = intel_guc_ggtt_offset(guc, ce->state) + + LRC_STATE_PN * PAGE_SIZE; /* XXX: In direct submission, the GuC wants the HW context id * here. In proxy submission, it wants the stage id @@ -395,7 +395,7 @@ static void guc_stage_desc_init(struct intel_guc *guc, lrc->context_id = (client->stage_id << GUC_ELC_CTXID_OFFSET) | (guc_engine_id << GUC_ELC_ENGINE_OFFSET); - lrc->ring_begin = guc_ggtt_offset(ce->ring->vma); + lrc->ring_begin = intel_guc_ggtt_offset(guc, ce->ring->vma); lrc->ring_end = lrc->ring_begin + ce->ring->size - 1; lrc->ring_next_free_location = lrc->ring_begin; lrc->ring_current_tail_pointer_value = 0; @@ -411,7 +411,7 @@ static void guc_stage_desc_init(struct intel_guc *guc, * The doorbell, process descriptor, and workqueue are all parts * of the client object, which the GuC will reference via the GGTT */ - gfx_addr = guc_ggtt_offset(client->vma); + gfx_addr = intel_guc_ggtt_offset(guc, client->vma); desc->db_trigger_phy = sg_dma_address(client->vma->pages->sgl) + client->doorbell_offset; desc->db_trigger_cpu = ptr_to_u64(__get_doorbell(client)); @@ -584,7 +584,7 @@ static void inject_preempt_context(struct work_struct *work) data[3] = engine->guc_id; data[4] = guc->execbuf_client->priority; data[5] = guc->execbuf_client->stage_id; - data[6] = guc_ggtt_offset(guc->shared_data); + data[6] = intel_guc_ggtt_offset(guc, guc->shared_data); if (WARN_ON(intel_guc_send(guc, data, ARRAY_SIZE(data)))) { execlists_clear_active(&engine->execlists, diff --git a/drivers/gpu/drm/i915/intel_huc.c b/drivers/gpu/drm/i915/intel_huc.c index 65e2afb..858c954 100644 --- a/drivers/gpu/drm/i915/intel_huc.c +++ b/drivers/gpu/drm/i915/intel_huc.c @@ -63,7 +63,8 @@ int intel_huc_auth(struct intel_huc *huc) } ret = intel_guc_auth_huc(guc, - guc_ggtt_offset(vma) + huc->fw.rsa_offset); + intel_guc_ggtt_offset(guc, vma) + + huc->fw.rsa_offset); if (ret) { DRM_ERROR("HuC: GuC did not ack Auth request %d\n", ret); goto fail_unpin; diff --git a/drivers/gpu/drm/i915/intel_huc_fw.c b/drivers/gpu/drm/i915/intel_huc_fw.c index c66afa9..bb0f8b7 100644 --- a/drivers/gpu/drm/i915/intel_huc_fw.c +++ b/drivers/gpu/drm/i915/intel_huc_fw.c @@ -118,7 +118,8 @@ static int huc_fw_xfer(struct intel_uc_fw *huc_fw, struct i915_vma *vma) intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); /* Set the source address for the uCode */ - offset = guc_ggtt_offset(vma) + huc_fw->header_offset; + offset = intel_guc_ggtt_offset(&dev_priv->guc, vma) + + huc_fw->header_offset; I915_WRITE(DMA_ADDR_0_LOW, lower_32_bits(offset)); I915_WRITE(DMA_ADDR_0_HIGH, upper_32_bits(offset) & 0xFFFF); -- cgit v1.1 From 6b0478fb722ae638ad747e17251e90bbf1b7969b Mon Sep 17 00:00:00 2001 From: Jackie Li Date: Tue, 13 Mar 2018 17:32:50 -0700 Subject: drm/i915: Implement dynamic GuC WOPCM offset and size calculation Hardware may have specific restrictions on GuC WOPCM offset and size. On Gen9, the value of the GuC WOPCM size register needs to be larger than the value of GuC WOPCM offset register + a Gen9 specific offset (144KB) for reserved GuC WOPCM. Fail to enforce such a restriction on GuC WOPCM size will lead to GuC firmware execution failures. On the other hand, with current static GuC WOPCM offset and size values (512KB for both offset and size), the GuC WOPCM size verification will fail on Gen9 even if it can be fixed by lowering the GuC WOPCM offset by calculating its value based on HuC firmware size (which is likely less than 200KB on Gen9), so that we can have a GuC WOPCM size value which is large enough to pass the GuC WOPCM size check. This patch updates the reserved GuC WOPCM size for RC6 context on Gen9 to 24KB to strictly align with the Gen9 GuC WOPCM layout. It also adds support to verify the GuC WOPCM size aganist the Gen9 hardware restrictions. To meet all above requirements, let's provide dynamic partitioning of the WOPCM that will be based on platform specific HuC/GuC firmware sizes. v2: - Removed intel_wopcm_init (Ville/Sagar/Joonas) - Renamed and Moved the intel_wopcm_partition into intel_guc (Sagar) - Removed unnecessary function calls (Joonas) - Init GuC WOPCM partition as soon as firmware fetching is completed v3: - Fixed indentation issues (Chris) - Removed layering violation code (Chris/Michal) - Created separat files for GuC wopcm code (Michal) - Used inline function to avoid code duplication (Michal) v4: - Preset the GuC WOPCM top during early GuC init (Chris) - Fail intel_uc_init_hw() as soon as GuC WOPCM partitioning failed v5: - Moved GuC DMA WOPCM register updating code into intel_wopcm.c - Took care of the locking status before writing to GuC DMA Write-Once registers. (Joonas) v6: - Made sure the GuC WOPCM size to be multiple of 4K (4K aligned) v8: - Updated comments and fixed naming issues (Sagar/Joonas) - Updated commit message to include more description about the hardware restriction on GuC WOPCM size (Sagar) v9: - Minor changes variable names and code comments (Sagar) - Added detailed GuC WOPCM layout drawing (Sagar/Michal) - Refined macro definitions to be reader friendly (Michal) - Removed redundent check to valid flag (Michal) - Unified first parameter for exported GuC WOPCM functions (Michal) - Refined the name and parameter list of hardware restriction checking functions (Michal) v10: - Used shorter function name for internal functions (Joonas) - Moved init-ealry function into c file (Joonas) - Consolidated and removed redundant size checks (Joonas/Michal) - Removed unnecessary unlikely() from code which is only called once during boot (Joonas) - More fixes to kernel-doc format and content (Michal) - Avoided the use of PAGE_MASK for 4K pages (Michal) - Added error log messages to error paths (Michal) v11: - Replaced intel_guc_wopcm with more generic intel_wopcm and attached intel_wopcm to drm_i915_private instead intel_guc (Michal) - dynamic calculation of GuC non-wopcm memory start (a.k.a WOPCM Top offset from GuC WOPCM base) (Michal) - Moved WOPCM marco definitions into .c source file (Michal) - Exported WOPCM layout diagram as kernel-doc (Michal) v12: - Updated naming, function kernel-doc to align with new changes (Michal) v13: - Updated the ordering of s-o-b/cc/r-b tags (Sagar) - Corrected one tense error in comment (Sagar) - Corrected typos and removed spurious comments (Joonas) Bspec: 12690 Signed-off-by: Jackie Li Cc: Michal Wajdeczko Cc: Sagar Arun Kamble Cc: Sujaritha Sundaresan Cc: Daniele Ceraolo Spurio Cc: John Spotswood Cc: Oscar Mateo Cc: Chris Wilson Cc: Joonas Lahtinen Reviewed-by: Sagar Arun Kamble (v8) Reviewed-by: Joonas Lahtinen (v9) Reviewed-by: Michal Wajdeczko (v11) Reviewed-by: Joonas Lahtinen (v12) Reviewed-by: Joonas Lahtinen Signed-off-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/1520987574-19351-2-git-send-email-yaodong.li@intel.com --- drivers/gpu/drm/i915/Makefile | 3 +- drivers/gpu/drm/i915/i915_drv.c | 1 + drivers/gpu/drm/i915/i915_drv.h | 8 ++ drivers/gpu/drm/i915/i915_gem.c | 4 + drivers/gpu/drm/i915/i915_gem_context.c | 5 +- drivers/gpu/drm/i915/intel_guc.c | 66 +++++++++--- drivers/gpu/drm/i915/intel_guc.h | 18 ++-- drivers/gpu/drm/i915/intel_guc_reg.h | 8 +- drivers/gpu/drm/i915/intel_huc.c | 2 +- drivers/gpu/drm/i915/intel_uc.c | 6 +- drivers/gpu/drm/i915/intel_uc_fw.c | 13 +-- drivers/gpu/drm/i915/intel_uc_fw.h | 16 +++ drivers/gpu/drm/i915/intel_wopcm.c | 182 ++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_wopcm.h | 30 ++++++ 14 files changed, 321 insertions(+), 41 deletions(-) create mode 100644 drivers/gpu/drm/i915/intel_wopcm.c create mode 100644 drivers/gpu/drm/i915/intel_wopcm.h diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile index fcb8a7b..552e43e 100644 --- a/drivers/gpu/drm/i915/Makefile +++ b/drivers/gpu/drm/i915/Makefile @@ -79,7 +79,8 @@ i915-y += i915_cmd_parser.o \ intel_lrc.o \ intel_mocs.o \ intel_ringbuffer.o \ - intel_uncore.o + intel_uncore.o \ + intel_wopcm.o # general-purpose microcontroller (GuC) support i915-y += intel_uc.o \ diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 0126b22..f03555e 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -919,6 +919,7 @@ static int i915_driver_init_early(struct drm_i915_private *dev_priv, mutex_init(&dev_priv->wm.wm_mutex); mutex_init(&dev_priv->pps_mutex); + intel_wopcm_init_early(&dev_priv->wopcm); intel_uc_init_early(dev_priv); i915_memcpy_init_early(dev_priv); diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 74b0e9d..e27ba8f 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -64,6 +64,7 @@ #include "intel_opregion.h" #include "intel_ringbuffer.h" #include "intel_uncore.h" +#include "intel_wopcm.h" #include "intel_uc.h" #include "i915_gem.h" @@ -1589,6 +1590,8 @@ struct drm_i915_private { struct intel_gvt *gvt; + struct intel_wopcm wopcm; + struct intel_huc huc; struct intel_guc guc; @@ -2121,6 +2124,11 @@ static inline struct drm_i915_private *kdev_to_i915(struct device *kdev) return to_i915(dev_get_drvdata(kdev)); } +static inline struct drm_i915_private *wopcm_to_i915(struct intel_wopcm *wopcm) +{ + return container_of(wopcm, struct drm_i915_private, wopcm); +} + static inline struct drm_i915_private *guc_to_i915(struct intel_guc *guc) { return container_of(guc, struct drm_i915_private, guc); diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index d0624c5..51faa65 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -5294,6 +5294,10 @@ int i915_gem_init(struct drm_i915_private *dev_priv) if (ret) return ret; + ret = intel_wopcm_init(&dev_priv->wopcm); + if (ret) + return ret; + ret = intel_uc_init_misc(dev_priv); if (ret) return ret; diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index f2cbea7..5cfac02 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -318,12 +318,13 @@ __create_hw_context(struct drm_i915_private *dev_priv, ctx->desc_template = default_desc_template(dev_priv, dev_priv->mm.aliasing_ppgtt); - /* GuC requires the ring to be placed above GUC_WOPCM_TOP. If GuC is not + /* + * GuC requires the ring to be placed in Non-WOPCM memory. If GuC is not * present or not in use we still need a small bias as ring wraparound * at offset 0 sometimes hangs. No idea why. */ if (USES_GUC(dev_priv)) - ctx->ggtt_offset_bias = GUC_WOPCM_TOP; + ctx->ggtt_offset_bias = dev_priv->guc.ggtt_pin_bias; else ctx->ggtt_offset_bias = I915_GTT_PAGE_SIZE; diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index 7846384..3eb516e 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -493,6 +493,57 @@ int intel_guc_resume(struct intel_guc *guc) } /** + * DOC: GuC Address Space + * + * The layout of GuC address space is shown as below: + * + * +==============> +====================+ <== GUC_GGTT_TOP + * ^ | | + * | | | + * | | DRAM | + * | | Memory | + * | | | + * GuC | | + * Address +========> +====================+ <== WOPCM Top + * Space ^ | HW contexts RSVD | + * | | | WOPCM | + * | | +==> +--------------------+ <== GuC WOPCM Top + * | GuC ^ | | + * | GGTT | | | + * | Pin GuC | GuC | + * | Bias WOPCM | WOPCM | + * | | Size | | + * | | | | | + * v v v | | + * +=====+=====+==> +====================+ <== GuC WOPCM Base + * | Non-GuC WOPCM | + * | (HuC/Reserved) | + * +====================+ <== WOPCM Base + * + * The lower part [0, GuC ggtt_pin_bias) is mapped to WOPCM which consists of + * GuC WOPCM and WOPCM reserved for other usage (e.g.RC6 context). The value of + * the GuC ggtt_pin_bias is determined by the actually GuC WOPCM size which is + * set in GUC_WOPCM_SIZE register. + */ + +/** + * intel_guc_init_ggtt_pin_bias() - Initialize the GuC ggtt_pin_bias value. + * @guc: intel_guc structure. + * + * This function will calculate and initialize the ggtt_pin_bias value based on + * overall WOPCM size and GuC WOPCM size. + */ +void intel_guc_init_ggtt_pin_bias(struct intel_guc *guc) +{ + struct drm_i915_private *i915 = guc_to_i915(guc); + + GEM_BUG_ON(!i915->wopcm.size); + GEM_BUG_ON(i915->wopcm.size < i915->wopcm.guc.base); + + guc->ggtt_pin_bias = i915->wopcm.size - i915->wopcm.guc.base; +} + +/** * intel_guc_allocate_vma() - Allocate a GGTT VMA for GuC usage * @guc: the guc * @size: size of area to allocate (both virtual space and memory) @@ -500,7 +551,7 @@ int intel_guc_resume(struct intel_guc *guc) * This is a wrapper to create an object for use with the GuC. In order to * use it inside the GuC, an object needs to be pinned lifetime, so we allocate * both some backing storage and a range inside the Global GTT. We must pin - * it in the GGTT somewhere other than than [0, GUC_WOPCM_TOP) because that + * it in the GGTT somewhere other than than [0, GUC ggtt_pin_bias) because that * range is reserved inside GuC. * * Return: A i915_vma if successful, otherwise an ERR_PTR. @@ -521,7 +572,7 @@ struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size) goto err; ret = i915_vma_pin(vma, 0, PAGE_SIZE, - PIN_GLOBAL | PIN_OFFSET_BIAS | GUC_WOPCM_TOP); + PIN_GLOBAL | PIN_OFFSET_BIAS | guc->ggtt_pin_bias); if (ret) { vma = ERR_PTR(ret); goto err; @@ -533,14 +584,3 @@ err: i915_gem_object_put(obj); return vma; } - -u32 intel_guc_wopcm_size(struct drm_i915_private *dev_priv) -{ - u32 wopcm_size = GUC_WOPCM_TOP; - - /* On BXT, the top of WOPCM is reserved for RC6 context */ - if (IS_GEN9_LP(dev_priv)) - wopcm_size -= BXT_GUC_WOPCM_RC6_RESERVED; - - return wopcm_size; -} diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h index a1be04e..cdb649a9 100644 --- a/drivers/gpu/drm/i915/intel_guc.h +++ b/drivers/gpu/drm/i915/intel_guc.h @@ -49,6 +49,9 @@ struct intel_guc { struct intel_guc_log log; struct intel_guc_ct ct; + /* Offset where Non-WOPCM memory starts. */ + u32 ggtt_pin_bias; + /* Log snapshot if GuC errors during load */ struct drm_i915_gem_object *load_err_log; @@ -108,19 +111,20 @@ static inline void intel_guc_notify(struct intel_guc *guc) * @guc: intel_guc structure. * @vma: i915 graphics virtual memory area. * - * GuC does not allow any gfx GGTT address that falls into range [0, WOPCM_TOP), - * which is reserved for Boot ROM, SRAM and WOPCM. Currently this top address is - * 512K. In order to exclude 0-512K address space from GGTT, all gfx objects - * used by GuC is pinned with PIN_OFFSET_BIAS along with size of WOPCM. + * GuC does not allow any gfx GGTT address that falls into range + * [0, GuC ggtt_pin_bias), which is reserved for Boot ROM, SRAM and WOPCM. + * Currently, in order to exclude [0, GuC ggtt_pin_bias) address space from + * GGTT, all gfx objects used by GuC are allocated with intel_guc_allocate_vma() + * and pinned with PIN_OFFSET_BIAS along with the value of GuC ggtt_pin_bias. * - * Return: GGTT offset that meets the GuC gfx address requirement. + * Return: GGTT offset of the @vma. */ static inline u32 intel_guc_ggtt_offset(struct intel_guc *guc, struct i915_vma *vma) { u32 offset = i915_ggtt_offset(vma); - GEM_BUG_ON(offset < GUC_WOPCM_TOP); + GEM_BUG_ON(offset < guc->ggtt_pin_bias); GEM_BUG_ON(range_overflows_t(u64, offset, vma->size, GUC_GGTT_TOP)); return offset; @@ -129,6 +133,7 @@ static inline u32 intel_guc_ggtt_offset(struct intel_guc *guc, void intel_guc_init_early(struct intel_guc *guc); void intel_guc_init_send_regs(struct intel_guc *guc); void intel_guc_init_params(struct intel_guc *guc); +void intel_guc_init_ggtt_pin_bias(struct intel_guc *guc); int intel_guc_init_wq(struct intel_guc *guc); void intel_guc_fini_wq(struct intel_guc *guc); int intel_guc_init(struct intel_guc *guc); @@ -141,7 +146,6 @@ int intel_guc_auth_huc(struct intel_guc *guc, u32 rsa_offset); int intel_guc_suspend(struct intel_guc *guc); int intel_guc_resume(struct intel_guc *guc); struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size); -u32 intel_guc_wopcm_size(struct drm_i915_private *dev_priv); static inline int intel_guc_sanitize(struct intel_guc *guc) { diff --git a/drivers/gpu/drm/i915/intel_guc_reg.h b/drivers/gpu/drm/i915/intel_guc_reg.h index 711e9e9..01963d0 100644 --- a/drivers/gpu/drm/i915/intel_guc_reg.h +++ b/drivers/gpu/drm/i915/intel_guc_reg.h @@ -68,17 +68,15 @@ #define DMA_GUC_WOPCM_OFFSET _MMIO(0xc340) #define HUC_LOADING_AGENT_VCR (0<<1) #define HUC_LOADING_AGENT_GUC (1<<1) -#define GUC_WOPCM_OFFSET_VALUE 0x80000 /* 512KB */ +#define GUC_WOPCM_OFFSET_SHIFT 14 #define GUC_MAX_IDLE_COUNT _MMIO(0xC3E4) #define HUC_STATUS2 _MMIO(0xD3B0) #define HUC_FW_VERIFIED (1<<7) -/* Defines WOPCM space available to GuC firmware */ #define GUC_WOPCM_SIZE _MMIO(0xc050) -/* GuC addresses below GUC_WOPCM_TOP don't map through the GTT */ -#define GUC_WOPCM_TOP (0x80 << 12) /* 512KB */ -#define BXT_GUC_WOPCM_RC6_RESERVED (0x10 << 12) /* 64KB */ +#define GUC_WOPCM_SIZE_SHIFT 12 +#define GUC_WOPCM_SIZE_MASK (0xfffff << GUC_WOPCM_SIZE_SHIFT) #define GEN8_GT_PM_CONFIG _MMIO(0x138140) #define GEN9LP_GT_PM_CONFIG _MMIO(0x138140) diff --git a/drivers/gpu/drm/i915/intel_huc.c b/drivers/gpu/drm/i915/intel_huc.c index 858c954..1d6c47b 100644 --- a/drivers/gpu/drm/i915/intel_huc.c +++ b/drivers/gpu/drm/i915/intel_huc.c @@ -55,7 +55,7 @@ int intel_huc_auth(struct intel_huc *huc) return -ENOEXEC; vma = i915_gem_object_ggtt_pin(huc->fw.obj, NULL, 0, 0, - PIN_OFFSET_BIAS | GUC_WOPCM_TOP); + PIN_OFFSET_BIAS | guc->ggtt_pin_bias); if (IS_ERR(vma)) { ret = PTR_ERR(vma); DRM_ERROR("HuC: Failed to pin huc fw object %d\n", ret); diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 9d5ffd7..ed5a6fc 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -272,6 +272,8 @@ int intel_uc_init_misc(struct drm_i915_private *dev_priv) if (!USES_GUC(dev_priv)) return 0; + intel_guc_init_ggtt_pin_bias(guc); + ret = intel_guc_init_wq(guc); if (ret) return ret; @@ -366,9 +368,9 @@ int intel_uc_init_hw(struct drm_i915_private *dev_priv) gen9_reset_guc_interrupts(dev_priv); /* init WOPCM */ - I915_WRITE(GUC_WOPCM_SIZE, intel_guc_wopcm_size(dev_priv)); + I915_WRITE(GUC_WOPCM_SIZE, dev_priv->wopcm.guc.size); I915_WRITE(DMA_GUC_WOPCM_OFFSET, - GUC_WOPCM_OFFSET_VALUE | HUC_LOADING_AGENT_GUC); + dev_priv->wopcm.guc.base | HUC_LOADING_AGENT_GUC); /* WaEnableuKernelHeaderValidFix:skl */ /* WaEnableGuCBootHashCheckNotSet:skl,bxt,kbl */ diff --git a/drivers/gpu/drm/i915/intel_uc_fw.c b/drivers/gpu/drm/i915/intel_uc_fw.c index 3ec0ce5..30c7324 100644 --- a/drivers/gpu/drm/i915/intel_uc_fw.c +++ b/drivers/gpu/drm/i915/intel_uc_fw.c @@ -95,15 +95,6 @@ void intel_uc_fw_fetch(struct drm_i915_private *dev_priv, uc_fw->ucode_offset = uc_fw->header_offset + uc_fw->header_size; uc_fw->ucode_size = (css->size_dw - css->header_size_dw) * sizeof(u32); - /* Header and uCode will be loaded to WOPCM */ - size = uc_fw->header_size + uc_fw->ucode_size; - if (size > intel_guc_wopcm_size(dev_priv)) { - DRM_WARN("%s: Firmware is too large to fit in WOPCM\n", - intel_uc_fw_type_repr(uc_fw->type)); - err = -E2BIG; - goto fail; - } - /* now RSA */ if (css->key_size_dw != UOS_RSA_SCRATCH_COUNT) { DRM_WARN("%s: Mismatched firmware RSA key size (%u)\n", @@ -208,6 +199,7 @@ int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, int (*xfer)(struct intel_uc_fw *uc_fw, struct i915_vma *vma)) { + struct drm_i915_private *i915 = to_i915(uc_fw->obj->base.dev); struct i915_vma *vma; int err; @@ -231,7 +223,8 @@ int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, } vma = i915_gem_object_ggtt_pin(uc_fw->obj, NULL, 0, 0, - PIN_OFFSET_BIAS | GUC_WOPCM_TOP); + PIN_OFFSET_BIAS | + i915->guc.ggtt_pin_bias); if (IS_ERR(vma)) { err = PTR_ERR(vma); DRM_DEBUG_DRIVER("%s fw ggtt-pin err=%d\n", diff --git a/drivers/gpu/drm/i915/intel_uc_fw.h b/drivers/gpu/drm/i915/intel_uc_fw.h index 2601521..dc33b12 100644 --- a/drivers/gpu/drm/i915/intel_uc_fw.h +++ b/drivers/gpu/drm/i915/intel_uc_fw.h @@ -121,6 +121,22 @@ static inline void intel_uc_fw_sanitize(struct intel_uc_fw *uc_fw) uc_fw->load_status = INTEL_UC_FIRMWARE_PENDING; } +/** + * intel_uc_fw_get_upload_size() - Get size of firmware needed to be uploaded. + * @uc_fw: uC firmware. + * + * Get the size of the firmware and header that will be uploaded to WOPCM. + * + * Return: Upload firmware size, or zero on firmware fetch failure. + */ +static inline u32 intel_uc_fw_get_upload_size(struct intel_uc_fw *uc_fw) +{ + if (uc_fw->fetch_status != INTEL_UC_FIRMWARE_SUCCESS) + return 0; + + return uc_fw->header_size + uc_fw->ucode_size; +} + void intel_uc_fw_fetch(struct drm_i915_private *dev_priv, struct intel_uc_fw *uc_fw); int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, diff --git a/drivers/gpu/drm/i915/intel_wopcm.c b/drivers/gpu/drm/i915/intel_wopcm.c new file mode 100644 index 0000000..7b150d5 --- /dev/null +++ b/drivers/gpu/drm/i915/intel_wopcm.c @@ -0,0 +1,182 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2017-2018 Intel Corporation + */ + +#include "intel_wopcm.h" +#include "i915_drv.h" + +/** + * DOC: WOPCM Layout + * + * The layout of the WOPCM will be fixed after writing to GuC WOPCM size and + * offset registers whose are calculated are determined by size of HuC/GuC + * firmware size and set of hw requirements/restrictions as shown below: + * + * +=========> +====================+ <== WOPCM Top + * ^ | HW contexts RSVD | + * | +===> +====================+ <== GuC WOPCM Top + * | ^ | | + * | | | | + * | | | | + * | GuC | | + * | WOPCM | | + * | Size +--------------------+ + * WOPCM | | GuC FW RSVD | + * | | +--------------------+ + * | | | GuC Stack RSVD | + * | | +------------------- + + * | v | GuC WOPCM RSVD | + * | +===> +====================+ <== GuC WOPCM base + * | | WOPCM RSVD | + * | +------------------- + <== HuC Firmware Top + * v | HuC FW | + * +=========> +====================+ <== WOPCM Base + * + * GuC accessible WOPCM starts at GuC WOPCM base and ends at GuC WOPCM top. + * The top part of the WOPCM is reserved for hardware contexts (e.g. RC6 + * context). + */ + +/* Default WOPCM size 1MB. */ +#define GEN9_WOPCM_SIZE (1024 * 1024) +/* 16KB WOPCM (RSVD WOPCM) is reserved from HuC firmware top. */ +#define WOPCM_RESERVED_SIZE (16 * 1024) + +/* 16KB reserved at the beginning of GuC WOPCM. */ +#define GUC_WOPCM_RESERVED (16 * 1024) +/* 8KB from GUC_WOPCM_RESERVED is reserved for GuC stack. */ +#define GUC_WOPCM_STACK_RESERVED (8 * 1024) + +/* GuC WOPCM Offset value needs to be aligned to 16KB. */ +#define GUC_WOPCM_OFFSET_ALIGNMENT (1UL << GUC_WOPCM_OFFSET_SHIFT) + +/* 24KB at the end of WOPCM is reserved for RC6 CTX on BXT. */ +#define BXT_WOPCM_RC6_CTX_RESERVED (24 * 1024) + +/* 128KB from GUC_WOPCM_RESERVED is reserved for FW on Gen9. */ +#define GEN9_GUC_FW_RESERVED (128 * 1024) +#define GEN9_GUC_WOPCM_OFFSET (GUC_WOPCM_RESERVED + GEN9_GUC_FW_RESERVED) + +/** + * intel_wopcm_init_early() - Early initialization of the WOPCM. + * @wopcm: pointer to intel_wopcm. + * + * Setup the size of WOPCM which will be used by later on WOPCM partitioning. + */ +void intel_wopcm_init_early(struct intel_wopcm *wopcm) +{ + wopcm->size = GEN9_WOPCM_SIZE; + + DRM_DEBUG_DRIVER("WOPCM size: %uKiB\n", wopcm->size / 1024); +} + +static inline u32 context_reserved_size(struct drm_i915_private *i915) +{ + if (IS_GEN9_LP(i915)) + return BXT_WOPCM_RC6_CTX_RESERVED; + else + return 0; +} + +static inline int gen9_check_dword_gap(u32 guc_wopcm_base, u32 guc_wopcm_size) +{ + u32 offset; + + /* + * GuC WOPCM size shall be at least a dword larger than the offset from + * WOPCM base (GuC WOPCM offset from WOPCM base + GEN9_GUC_WOPCM_OFFSET) + * due to hardware limitation on Gen9. + */ + offset = guc_wopcm_base + GEN9_GUC_WOPCM_OFFSET; + if (offset > guc_wopcm_size || + (guc_wopcm_size - offset) < sizeof(u32)) { + DRM_ERROR("GuC WOPCM size %uKiB is too small. %uKiB needed.\n", + guc_wopcm_size / 1024, + (u32)(offset + sizeof(u32)) / 1024); + return -E2BIG; + } + + return 0; +} + +static inline int check_hw_restriction(struct drm_i915_private *i915, + u32 guc_wopcm_base, u32 guc_wopcm_size) +{ + int err = 0; + + if (IS_GEN9(i915)) + err = gen9_check_dword_gap(guc_wopcm_base, guc_wopcm_size); + + return err; +} + +/** + * intel_wopcm_init() - Initialize the WOPCM structure. + * @wopcm: pointer to intel_wopcm. + * + * This function will partition WOPCM space based on GuC and HuC firmware sizes + * and will allocate max remaining for use by GuC. This function will also + * enforce platform dependent hardware restrictions on GuC WOPCM offset and + * size. It will fail the WOPCM init if any of these checks were failed, so that + * the following GuC firmware uploading would be aborted. + * + * Return: 0 on success, non-zero error code on failure. + */ +int intel_wopcm_init(struct intel_wopcm *wopcm) +{ + struct drm_i915_private *i915 = wopcm_to_i915(wopcm); + u32 guc_fw_size = intel_uc_fw_get_upload_size(&i915->guc.fw); + u32 huc_fw_size = intel_uc_fw_get_upload_size(&i915->huc.fw); + u32 ctx_rsvd = context_reserved_size(i915); + u32 guc_wopcm_base; + u32 guc_wopcm_size; + u32 guc_wopcm_rsvd; + int err; + + GEM_BUG_ON(!wopcm->size); + + if (guc_fw_size >= wopcm->size) { + DRM_ERROR("GuC FW (%uKiB) is too big to fit in WOPCM.", + guc_fw_size / 1024); + return -E2BIG; + } + + if (huc_fw_size >= wopcm->size) { + DRM_ERROR("HuC FW (%uKiB) is too big to fit in WOPCM.", + huc_fw_size / 1024); + return -E2BIG; + } + + guc_wopcm_base = ALIGN(huc_fw_size + WOPCM_RESERVED_SIZE, + GUC_WOPCM_OFFSET_ALIGNMENT); + if ((guc_wopcm_base + ctx_rsvd) >= wopcm->size) { + DRM_ERROR("GuC WOPCM base (%uKiB) is too big.\n", + guc_wopcm_base / 1024); + return -E2BIG; + } + + guc_wopcm_size = wopcm->size - guc_wopcm_base - ctx_rsvd; + guc_wopcm_size &= GUC_WOPCM_SIZE_MASK; + + DRM_DEBUG_DRIVER("Calculated GuC WOPCM Region: [%uKiB, %uKiB)\n", + guc_wopcm_base / 1024, guc_wopcm_size / 1024); + + guc_wopcm_rsvd = GUC_WOPCM_RESERVED + GUC_WOPCM_STACK_RESERVED; + if ((guc_fw_size + guc_wopcm_rsvd) > guc_wopcm_size) { + DRM_ERROR("Need %uKiB WOPCM for GuC, %uKiB available.\n", + (guc_fw_size + guc_wopcm_rsvd) / 1024, + guc_wopcm_size / 1024); + return -E2BIG; + } + + err = check_hw_restriction(i915, guc_wopcm_base, guc_wopcm_size); + if (err) + return err; + + wopcm->guc.base = guc_wopcm_base; + wopcm->guc.size = guc_wopcm_size; + + return 0; +} diff --git a/drivers/gpu/drm/i915/intel_wopcm.h b/drivers/gpu/drm/i915/intel_wopcm.h new file mode 100644 index 0000000..93c402c --- /dev/null +++ b/drivers/gpu/drm/i915/intel_wopcm.h @@ -0,0 +1,30 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2017-2018 Intel Corporation + */ + +#ifndef _INTEL_WOPCM_H_ +#define _INTEL_WOPCM_H_ + +#include + +/** + * struct intel_wopcm - Overall WOPCM info and WOPCM regions. + * @size: Size of overall WOPCM. + * @guc: GuC WOPCM Region info. + * @guc.base: GuC WOPCM base which is offset from WOPCM base. + * @guc.size: Size of the GuC WOPCM region. + */ +struct intel_wopcm { + u32 size; + struct { + u32 base; + u32 size; + } guc; +}; + +void intel_wopcm_init_early(struct intel_wopcm *wopcm); +int intel_wopcm_init(struct intel_wopcm *wopcm); + +#endif -- cgit v1.1 From 5cbc1e2f48086ea08a37b3b75aa481ae9842af39 Mon Sep 17 00:00:00 2001 From: Jackie Li Date: Tue, 13 Mar 2018 17:32:51 -0700 Subject: drm/i915: Add support to return CNL specific reserved WOPCM size CNL has its specific reserved GuC WOPCM size for RC6 and other hardware contexts. This patch updates the code to return CNL specific reserved GuC WOPCM size for RC6 and other hardware contexts so that the GuC WOPCM size can be calculated correctly for CNL. v9: - Created a new patch for these changes originally made in v8 4/6 patch of this series (Sagar/Michal) v10: - Used if-else ladder to the returning of context sizes (Joonas) v11: - Removed GUC_ prefix from context size macro (Michal) v13: - Updated the ordering of s-o-b/cc/r-b tags (Sagar) Bspec: 12690 Signed-off-by: Jackie Li Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Cc: Chris Wilson Cc: Joonas Lahtinen Reviewed-by: Joonas Lahtinen (v9) Reviewed-by: Michal Wajdeczko (v11) Reviewed-by: Joonas Lahtinen (v12) Reviewed-by: Joonas Lahtinen Signed-off-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/1520987574-19351-3-git-send-email-yaodong.li@intel.com --- drivers/gpu/drm/i915/intel_wopcm.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_wopcm.c b/drivers/gpu/drm/i915/intel_wopcm.c index 7b150d5..a29e0c9 100644 --- a/drivers/gpu/drm/i915/intel_wopcm.c +++ b/drivers/gpu/drm/i915/intel_wopcm.c @@ -54,6 +54,8 @@ /* 24KB at the end of WOPCM is reserved for RC6 CTX on BXT. */ #define BXT_WOPCM_RC6_CTX_RESERVED (24 * 1024) +/* 36KB WOPCM reserved at the end of WOPCM on CNL. */ +#define CNL_WOPCM_HW_CTX_RESERVED (36 * 1024) /* 128KB from GUC_WOPCM_RESERVED is reserved for FW on Gen9. */ #define GEN9_GUC_FW_RESERVED (128 * 1024) @@ -76,6 +78,8 @@ static inline u32 context_reserved_size(struct drm_i915_private *i915) { if (IS_GEN9_LP(i915)) return BXT_WOPCM_RC6_CTX_RESERVED; + else if (INTEL_GEN(i915) >= 10) + return CNL_WOPCM_HW_CTX_RESERVED; else return 0; } -- cgit v1.1 From 96c83d35a26f7ba06a1623adaf872b56d54fe093 Mon Sep 17 00:00:00 2001 From: Jackie Li Date: Tue, 13 Mar 2018 17:32:52 -0700 Subject: drm/i915: Add HuC firmware size related restriction for Gen9 and CNL A0 On CNL A0 and Gen9, there's a hardware restriction that requires the available GuC WOPCM size to be larger than or equal to HuC firmware size. This patch adds new verification code to ensure the available GuC WOPCM size to be larger than or equal to HuC firmware size on both Gen9 and CNL A0. v6: - Extended HuC FW size check against GuC WOPCM size to all Gen9 and CNL A0 platforms v7: - Fixed patch format issues v8: - Renamed variables and functions to avoid ambiguity (Joonas) - Updated commit message and comments to be more comprehensive (Sagar) v9: - Moved code that is not related to restriction check into a separate patch and updated the commit message accordingly (Sagar/Michal) - Avoided to call uc_get_fw_size for better layer isolation (Michal) v10: - Shorten function names and reorganized size_check code to have clear isolation (Joonas) - Removed unnecessary comments (Joonas) v11: - Fixed logic error in size check (Michal) v12: - Add space between "HuC FW" and "(%uKiB)" in error message (Michal) v13: - Updated the ordering of s-o-b/cc/r-b tags (Sagar) BSpec: 10875 Signed-off-by: Jackie Li Cc: Michal Wajdeczko Cc: Sagar Arun Kamble Cc: John Spotswood Cc: Jeff McGee Cc: Chris Wilson Cc: Joonas Lahtinen Reviewed-by: Sagar Arun Kamble (v8) Reviewed-by: Michal Wajdeczko (v11) Reviewed-by: Joonas Lahtinen (v12) Reviewed-by: Joonas Lahtinen Signed-off-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/1520987574-19351-4-git-send-email-yaodong.li@intel.com --- drivers/gpu/drm/i915/intel_wopcm.c | 27 +++++++++++++++++++++++++-- 1 file changed, 25 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_wopcm.c b/drivers/gpu/drm/i915/intel_wopcm.c index a29e0c9..1fd1125 100644 --- a/drivers/gpu/drm/i915/intel_wopcm.c +++ b/drivers/gpu/drm/i915/intel_wopcm.c @@ -105,14 +105,36 @@ static inline int gen9_check_dword_gap(u32 guc_wopcm_base, u32 guc_wopcm_size) return 0; } +static inline int gen9_check_huc_fw_fits(u32 guc_wopcm_size, u32 huc_fw_size) +{ + /* + * On Gen9 & CNL A0, hardware requires the total available GuC WOPCM + * size to be larger than or equal to HuC firmware size. Otherwise, + * firmware uploading would fail. + */ + if (huc_fw_size > guc_wopcm_size - GUC_WOPCM_RESERVED) { + DRM_ERROR("HuC FW (%uKiB) won't fit in GuC WOPCM (%uKiB).\n", + huc_fw_size / 1024, + (guc_wopcm_size - GUC_WOPCM_RESERVED) / 1024); + return -E2BIG; + } + + return 0; +} + static inline int check_hw_restriction(struct drm_i915_private *i915, - u32 guc_wopcm_base, u32 guc_wopcm_size) + u32 guc_wopcm_base, u32 guc_wopcm_size, + u32 huc_fw_size) { int err = 0; if (IS_GEN9(i915)) err = gen9_check_dword_gap(guc_wopcm_base, guc_wopcm_size); + if (!err && + (IS_GEN9(i915) || IS_CNL_REVID(i915, CNL_REVID_A0, CNL_REVID_A0))) + err = gen9_check_huc_fw_fits(guc_wopcm_size, huc_fw_size); + return err; } @@ -175,7 +197,8 @@ int intel_wopcm_init(struct intel_wopcm *wopcm) return -E2BIG; } - err = check_hw_restriction(i915, guc_wopcm_base, guc_wopcm_size); + err = check_hw_restriction(i915, guc_wopcm_base, guc_wopcm_size, + huc_fw_size); if (err) return err; -- cgit v1.1 From f08e2035cc089caf52ce57e855d96ba6ba90c71a Mon Sep 17 00:00:00 2001 From: Jackie Li Date: Tue, 13 Mar 2018 17:32:53 -0700 Subject: drm/i915/guc: Check the locking status of GuC WOPCM registers GuC WOPCM registers are write-once registers. Current driver code accesses these registers without checking the accessibility to these registers which will lead to unpredictable driver behaviors if these registers were touch by other components (such as faulty BIOS code). This patch moves the GuC WOPCM registers updating code into intel_wopcm.c and adds check before and after the update to GuC WOPCM registers so that we can make sure the driver is in a known state after writing to these write-once registers. v6: - Made sure module reloading won't bug the kernel while doing locking status checking v7: - Fixed patch format issues v8: - Fixed coding style issue on register lock bit macro definition (Sagar) v9: - Avoided to use redundant !! to cast uint to bool (Chris) - Return error code instead of GEM_BUG_ON for locked with invalid register values case (Sagar) - Updated guc_wopcm_hw_init to use guc_wopcm as first parameter (Michal) - Added code to set and validate the HuC_LOADING_AGENT_GUC bit in GuC WOPCM offset register based on the presence of HuC firmware (Michal) - Use bit fields instead of macros for GuC WOPCM flags (Michal) v10: - Refined variable names, removed redundant comments (Joonas) - Introduced lockable_reg to handle the write once register write and propagate the write error to caller (Joonas) - Used lockable_reg abstraction to avoid locking bit check on generic i915_reg_t (Michal) - Added log message for error paths (Michal) - Removed hw_updated flag and only relies on real hardware status v11: - Replaced lockable_reg with simplified function (Michal) - Used new macros for locking bits of WOPCM size/offset registers instead of using BIT(0) directly (Michal) - use intel_wopcm_init_hw() called from intel_gem_init_hw() to do GuC WOPCM register setup instead of calling from intel_uc_init_hw() (Michal) v12: - Updated function kernel-doc to align with code changes (Michal) - Updated code to use wopcm pointer directly (Michal) v13: - Updated the ordering of s-o-b/cc/r-b tags (Sagar) BSpec: 10875, 10833 Signed-off-by: Jackie Li Cc: Michal Wajdeczko Cc: Chris Wilson Cc: Joonas Lahtinen Reviewed-by: Michal Wajdeczko (v11) Reviewed-by: Joonas Lahtinen (v12) Reviewed-by: Joonas Lahtinen Signed-off-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/1520987574-19351-5-git-send-email-yaodong.li@intel.com --- drivers/gpu/drm/i915/i915_gem.c | 6 ++++ drivers/gpu/drm/i915/intel_guc_reg.h | 3 ++ drivers/gpu/drm/i915/intel_uc.c | 5 --- drivers/gpu/drm/i915/intel_wopcm.c | 64 ++++++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_wopcm.h | 1 + 5 files changed, 74 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 51faa65..13d4b0e 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -5137,6 +5137,12 @@ int i915_gem_init_hw(struct drm_i915_private *dev_priv) goto out; } + ret = intel_wopcm_init_hw(&dev_priv->wopcm); + if (ret) { + DRM_ERROR("Enabling WOPCM failed (%d)\n", ret); + goto out; + } + /* We can't enable contexts until all firmware is loaded */ ret = intel_uc_init_hw(dev_priv); if (ret) { diff --git a/drivers/gpu/drm/i915/intel_guc_reg.h b/drivers/gpu/drm/i915/intel_guc_reg.h index 01963d0..d860847 100644 --- a/drivers/gpu/drm/i915/intel_guc_reg.h +++ b/drivers/gpu/drm/i915/intel_guc_reg.h @@ -66,15 +66,18 @@ #define UOS_MOVE (1<<4) #define START_DMA (1<<0) #define DMA_GUC_WOPCM_OFFSET _MMIO(0xc340) +#define GUC_WOPCM_OFFSET_VALID (1<<0) #define HUC_LOADING_AGENT_VCR (0<<1) #define HUC_LOADING_AGENT_GUC (1<<1) #define GUC_WOPCM_OFFSET_SHIFT 14 +#define GUC_WOPCM_OFFSET_MASK (0x3ffff << GUC_WOPCM_OFFSET_SHIFT) #define GUC_MAX_IDLE_COUNT _MMIO(0xC3E4) #define HUC_STATUS2 _MMIO(0xD3B0) #define HUC_FW_VERIFIED (1<<7) #define GUC_WOPCM_SIZE _MMIO(0xc050) +#define GUC_WOPCM_SIZE_LOCKED (1<<0) #define GUC_WOPCM_SIZE_SHIFT 12 #define GUC_WOPCM_SIZE_MASK (0xfffff << GUC_WOPCM_SIZE_SHIFT) diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index ed5a6fc..6316548 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -367,11 +367,6 @@ int intel_uc_init_hw(struct drm_i915_private *dev_priv) gen9_reset_guc_interrupts(dev_priv); - /* init WOPCM */ - I915_WRITE(GUC_WOPCM_SIZE, dev_priv->wopcm.guc.size); - I915_WRITE(DMA_GUC_WOPCM_OFFSET, - dev_priv->wopcm.guc.base | HUC_LOADING_AGENT_GUC); - /* WaEnableuKernelHeaderValidFix:skl */ /* WaEnableGuCBootHashCheckNotSet:skl,bxt,kbl */ if (IS_GEN9(dev_priv)) diff --git a/drivers/gpu/drm/i915/intel_wopcm.c b/drivers/gpu/drm/i915/intel_wopcm.c index 1fd1125..4117886 100644 --- a/drivers/gpu/drm/i915/intel_wopcm.c +++ b/drivers/gpu/drm/i915/intel_wopcm.c @@ -207,3 +207,67 @@ int intel_wopcm_init(struct intel_wopcm *wopcm) return 0; } + +static inline int write_and_verify(struct drm_i915_private *dev_priv, + i915_reg_t reg, u32 val, u32 mask, + u32 locked_bit) +{ + u32 reg_val; + + GEM_BUG_ON(val & ~mask); + + I915_WRITE(reg, val); + + reg_val = I915_READ(reg); + + return (reg_val & mask) != (val | locked_bit) ? -EIO : 0; +} + +/** + * intel_wopcm_init_hw() - Setup GuC WOPCM registers. + * @wopcm: pointer to intel_wopcm. + * + * Setup the GuC WOPCM size and offset registers with the calculated values. It + * will verify the register values to make sure the registers are locked with + * correct values. + * + * Return: 0 on success. -EIO if registers were locked with incorrect values. + */ +int intel_wopcm_init_hw(struct intel_wopcm *wopcm) +{ + struct drm_i915_private *dev_priv = wopcm_to_i915(wopcm); + u32 huc_agent; + u32 mask; + int err; + + if (!USES_GUC(dev_priv)) + return 0; + + GEM_BUG_ON(!HAS_GUC(dev_priv)); + GEM_BUG_ON(!wopcm->guc.size); + GEM_BUG_ON(!wopcm->guc.base); + + err = write_and_verify(dev_priv, GUC_WOPCM_SIZE, wopcm->guc.size, + GUC_WOPCM_SIZE_MASK | GUC_WOPCM_SIZE_LOCKED, + GUC_WOPCM_SIZE_LOCKED); + if (err) + goto err_out; + + huc_agent = USES_HUC(dev_priv) ? HUC_LOADING_AGENT_GUC : 0; + mask = GUC_WOPCM_OFFSET_MASK | GUC_WOPCM_OFFSET_VALID | huc_agent; + err = write_and_verify(dev_priv, DMA_GUC_WOPCM_OFFSET, + wopcm->guc.base | huc_agent, mask, + GUC_WOPCM_OFFSET_VALID); + if (err) + goto err_out; + + return 0; + +err_out: + DRM_ERROR("Failed to init WOPCM registers:\n"); + DRM_ERROR("DMA_GUC_WOPCM_OFFSET=%#x\n", + I915_READ(DMA_GUC_WOPCM_OFFSET)); + DRM_ERROR("GUC_WOPCM_SIZE=%#x\n", I915_READ(GUC_WOPCM_SIZE)); + + return err; +} diff --git a/drivers/gpu/drm/i915/intel_wopcm.h b/drivers/gpu/drm/i915/intel_wopcm.h index 93c402c..6298910 100644 --- a/drivers/gpu/drm/i915/intel_wopcm.h +++ b/drivers/gpu/drm/i915/intel_wopcm.h @@ -26,5 +26,6 @@ struct intel_wopcm { void intel_wopcm_init_early(struct intel_wopcm *wopcm); int intel_wopcm_init(struct intel_wopcm *wopcm); +int intel_wopcm_init_hw(struct intel_wopcm *wopcm); #endif -- cgit v1.1 From ab2681512b4c10d65a0fc412de0ce09bc4166edd Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 14 Mar 2018 10:16:30 +0000 Subject: drm/i915: Check rq->timeline before deference Not only is the context suspect to disappearing, but so is it's timeline. Under a lockless inspection of the requests for debugging from intel_engine_dump(), the context may already have been freed and we have to check before chasing the dangling pointer. [28033.681755] Modules linked in: vgem snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic x86_pkg_temp_thermal intel_powerclamp coretemp snd_hda_intel crct10dif_pclmul crc32_pclmul snd_hda_codec snd_hwdep snd_hda_core ghash_clmulni_intel snd_pcm mei_me mei i915 r8169 mii prime_numbers i2c_hid [28033.681796] CPU: 3 PID: 3058 Comm: gem_exec_schedu Tainted: G U 4.16.0-rc5+ #9 [28033.681804] Hardware name: Acer Aspire E5-575G/Ironman_SK , BIOS V1.12 08/02/2016 [28033.681834] RIP: 0010:print_request+0x2b/0xb0 [i915] [28033.681840] RSP: 0018:ffffc90004afbc18 EFLAGS: 00010202 [28033.681847] RAX: 6b6b6b6b6b6b6b6b RBX: ffff8801921b5a40 RCX: 0000000000000006 [28033.681854] RDX: ffffc90004afbc60 RSI: ffff8801921b5a40 RDI: 0000000000000004 [28033.681861] RBP: ffffc90004afbd80 R08: 0000000000000000 R09: 0000000000000001 [28033.681868] R10: ffffc90004afbbd0 R11: ffffc90004afbc73 R12: ffffc90004afbc60 [28033.681875] R13: ffffc90004afbd80 R14: ffff8801d40ec670 R15: ffff8801921b5a40 [28033.681883] FS: 00007fbba5f6c8c0(0000) GS:ffff8801e8400000(0000) knlGS:0000000000000000 [28033.681891] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [28033.681897] CR2: 00007fbba5f8f000 CR3: 00000001b2efa002 CR4: 00000000003606e0 [28033.681904] Call Trace: [28033.681932] intel_engine_print_registers+0x6a7/0x930 [i915] [28033.681962] intel_engine_dump+0x30d/0x740 [i915] [28033.681971] ? seq_printf+0x3a/0x50 [28033.681995] i915_engine_info+0xb8/0xe0 [i915] [28033.682003] ? drm_get_color_range_name+0x20/0x20 [28033.682010] seq_read+0xe1/0x440 [28033.682018] full_proxy_read+0x51/0x80 [28033.682025] __vfs_read+0x21/0x130 [28033.682031] ? do_sys_open+0x134/0x220 [28033.682037] ? kmem_cache_free+0x177/0x2b0 [28033.682043] vfs_read+0xa1/0x150 [28033.682049] SyS_read+0x40/0xa0 [28033.682055] do_syscall_64+0x6b/0x1b0 [28033.682063] entry_SYSCALL_64_after_hwframe+0x42/0xb7 [28033.682069] RIP: 0033:0x7fbba4655d11 [28033.682074] RSP: 002b:00007ffd8c49da58 EFLAGS: 00000246 ORIG_RAX: 0000000000000000 [28033.682082] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fbba4655d11 [28033.682089] RDX: 000000000000003f RSI: 00005647bfbfc260 RDI: 0000000000000006 [28033.682096] RBP: 000000000000003f R08: 00000000ffffffff R09: 0000000000000000 [28033.682104] R10: 0000000000000000 R11: 0000000000000246 R12: 00005647bfbfc260 [28033.682111] R13: 0000000000000006 R14: 0000000000000000 R15: 00005647bfbfc260 [28033.682119] Code: 41 55 41 54 49 89 d4 55 53 48 89 fd 48 8b 86 c8 00 00 00 48 8b 3d d6 1e 14 e2 48 89 f3 48 2b be a8 02 00 00 48 8b 80 b0 00 00 00 <4c> 8b 68 18 e8 bc 80 02 e1 8b 8b 70 02 00 00 8b b3 28 02 00 00 [28033.682206] RIP: print_request+0x2b/0xb0 [i915] RSP: ffffc90004afbc18 Reported-by: Mika Kuoppala Signed-off-by: Chris Wilson Cc: Mika Kuoppala Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180314101630.8933-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/intel_engine_cs.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index a2b1e9e..f22c5f7 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1716,13 +1716,15 @@ static void print_request(struct drm_printer *m, struct i915_request *rq, const char *prefix) { + const char *name = rq->fence.ops->get_timeline_name(&rq->fence); + drm_printf(m, "%s%x%s [%llx:%x] prio=%d @ %dms: %s\n", prefix, rq->global_seqno, i915_request_completed(rq) ? "!" : "", rq->fence.context, rq->fence.seqno, rq->priotree.priority, jiffies_to_msecs(jiffies - rq->emitted_jiffies), - rq->timeline->common->name); + name); } static void hexdump(struct drm_printer *m, const void *buf, size_t len) -- cgit v1.1 From ad055fb8e010e4ff37f66aeed1d380329bddce67 Mon Sep 17 00:00:00 2001 From: Tvrtko Ursulin Date: Wed, 14 Mar 2018 08:05:35 +0000 Subject: drm/i915/pmu: Work around compiler warnings on some kernel configs Arnd Bergman reports: """ The conditional spinlock confuses gcc into thinking the 'flags' value might contain uninitialized data: drivers/gpu/drm/i915/i915_pmu.c: In function '__i915_pmu_event_read': arch/x86/include/asm/paravirt_types.h:573:3: error: 'flags' may be used uninitialized in this function [-Werror=maybe-uninitialized] The code is correct, but it's easy to see how the compiler gets confused here. This avoids the problem by pulling the lock outside of the function into its only caller. """ On deeper look it seems this is caused by paravirt spinlocks implementation when CONFIG_PARAVIRT_DEBUG is set, which by being complicated, manages to convince gcc locked parameter can be changed externally (impossible). Work around it by removing the conditional locking parameters altogether. (It was never the most elegant code anyway.) Slight penalty we now pay is an additional irqsave spin lock/unlock cycle on the event enable path. But since enable is not a fast path, that is preferrable to the alternative solution which was doing MMIO under irqsave spinlock. Signed-off-by: Tvrtko Ursulin Reported-by: Arnd Bergmann Fixes: 1fe699e30113 ("drm/i915/pmu: Fix sleep under atomic in RC6 readout") Cc: Tvrtko Ursulin Cc: Chris Wilson Cc: Imre Deak Cc: Jani Nikula Cc: Joonas Lahtinen Cc: Rodrigo Vivi Cc: David Airlie Cc: intel-gfx@lists.freedesktop.org Cc: dri-devel@lists.freedesktop.org Reviewed-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180314080535.17490-1-tvrtko.ursulin@linux.intel.com --- drivers/gpu/drm/i915/i915_pmu.c | 32 +++++++++++++------------------- 1 file changed, 13 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_pmu.c b/drivers/gpu/drm/i915/i915_pmu.c index 4bc7aef..11fb76b 100644 --- a/drivers/gpu/drm/i915/i915_pmu.c +++ b/drivers/gpu/drm/i915/i915_pmu.c @@ -412,7 +412,7 @@ static u64 __get_rc6(struct drm_i915_private *i915) return val; } -static u64 get_rc6(struct drm_i915_private *i915, bool locked) +static u64 get_rc6(struct drm_i915_private *i915) { #if IS_ENABLED(CONFIG_PM) unsigned long flags; @@ -428,8 +428,7 @@ static u64 get_rc6(struct drm_i915_private *i915, bool locked) * previously. */ - if (!locked) - spin_lock_irqsave(&i915->pmu.lock, flags); + spin_lock_irqsave(&i915->pmu.lock, flags); if (val >= i915->pmu.sample[__I915_SAMPLE_RC6_ESTIMATED].cur) { i915->pmu.sample[__I915_SAMPLE_RC6_ESTIMATED].cur = 0; @@ -438,12 +437,10 @@ static u64 get_rc6(struct drm_i915_private *i915, bool locked) val = i915->pmu.sample[__I915_SAMPLE_RC6_ESTIMATED].cur; } - if (!locked) - spin_unlock_irqrestore(&i915->pmu.lock, flags); + spin_unlock_irqrestore(&i915->pmu.lock, flags); } else { struct pci_dev *pdev = i915->drm.pdev; struct device *kdev = &pdev->dev; - unsigned long flags2; /* * We are runtime suspended. @@ -452,10 +449,8 @@ static u64 get_rc6(struct drm_i915_private *i915, bool locked) * on top of the last known real value, as the approximated RC6 * counter value. */ - if (!locked) - spin_lock_irqsave(&i915->pmu.lock, flags); - - spin_lock_irqsave(&kdev->power.lock, flags2); + spin_lock_irqsave(&i915->pmu.lock, flags); + spin_lock(&kdev->power.lock); if (!i915->pmu.sample[__I915_SAMPLE_RC6_ESTIMATED].cur) i915->pmu.suspended_jiffies_last = @@ -465,14 +460,13 @@ static u64 get_rc6(struct drm_i915_private *i915, bool locked) i915->pmu.suspended_jiffies_last; val += jiffies - kdev->power.accounting_timestamp; - spin_unlock_irqrestore(&kdev->power.lock, flags2); + spin_unlock(&kdev->power.lock); val = jiffies_to_nsecs(val); val += i915->pmu.sample[__I915_SAMPLE_RC6].cur; i915->pmu.sample[__I915_SAMPLE_RC6_ESTIMATED].cur = val; - if (!locked) - spin_unlock_irqrestore(&i915->pmu.lock, flags); + spin_unlock_irqrestore(&i915->pmu.lock, flags); } return val; @@ -481,7 +475,7 @@ static u64 get_rc6(struct drm_i915_private *i915, bool locked) #endif } -static u64 __i915_pmu_event_read(struct perf_event *event, bool locked) +static u64 __i915_pmu_event_read(struct perf_event *event) { struct drm_i915_private *i915 = container_of(event->pmu, typeof(*i915), pmu.base); @@ -519,7 +513,7 @@ static u64 __i915_pmu_event_read(struct perf_event *event, bool locked) val = count_interrupts(i915); break; case I915_PMU_RC6_RESIDENCY: - val = get_rc6(i915, locked); + val = get_rc6(i915); break; } } @@ -534,7 +528,7 @@ static void i915_pmu_event_read(struct perf_event *event) again: prev = local64_read(&hwc->prev_count); - new = __i915_pmu_event_read(event, false); + new = __i915_pmu_event_read(event); if (local64_cmpxchg(&hwc->prev_count, prev, new) != prev) goto again; @@ -584,14 +578,14 @@ static void i915_pmu_enable(struct perf_event *event) engine->pmu.enable_count[sample]++; } + spin_unlock_irqrestore(&i915->pmu.lock, flags); + /* * Store the current counter value so we can report the correct delta * for all listeners. Even when the event was already enabled and has * an existing non-zero value. */ - local64_set(&event->hw.prev_count, __i915_pmu_event_read(event, true)); - - spin_unlock_irqrestore(&i915->pmu.lock, flags); + local64_set(&event->hw.prev_count, __i915_pmu_event_read(event)); } static void i915_pmu_disable(struct perf_event *event) -- cgit v1.1 From 4635b573634c7028043244dbc1141ef57341deb2 Mon Sep 17 00:00:00 2001 From: Mahesh Kumar Date: Wed, 14 Mar 2018 13:36:52 +0530 Subject: drm/i915/cnl; Add macro to get PORT_TX register This patch creates a new macro to get PORT_TX register for any given DW. This removes the need of defining register address for each port & DW. Changes since V1: - Use underscope prefix, as macro isn't returning an mmio reg(Lucas) - Merge patch 1 & 2 of the series Changes since V2: - remove _MMIO_PORT6_LN macro (Rodrigo) Signed-off-by: Mahesh Kumar Reviewed-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180314080653.9444-2-mahesh1.kumar@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 137 ++++++++++++---------------------------- 1 file changed, 39 insertions(+), 98 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 761bd3a..5a2a3d6 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -154,8 +154,6 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define _PLL(pll, a, b) ((a) + (pll)*((b)-(a))) #define _MMIO_PLL(pll, a, b) _MMIO(_PLL(pll, a, b)) #define _MMIO_PORT6(port, a, b, c, d, e, f) _MMIO(_PICK(port, a, b, c, d, e, f)) -#define _MMIO_PORT6_LN(port, ln, a0, a1, b, c, d, e, f) \ - _MMIO(_PICK(port, a0, b, c, d, e, f) + (ln * (a1 - a0))) #define _PHY3(phy, ...) _PICK(phy, __VA_ARGS__) #define _MMIO_PHY3(phy, a, b, c) _MMIO(_PHY3(phy, a, b, c)) @@ -1964,30 +1962,36 @@ enum i915_power_well_id { _CNL_PORT_PCS_DW1_LN0_F) #define COMMON_KEEPER_EN (1 << 26) -#define _CNL_PORT_TX_DW2_GRP_AE 0x162348 -#define _CNL_PORT_TX_DW2_GRP_B 0x1623C8 -#define _CNL_PORT_TX_DW2_GRP_C 0x162B48 -#define _CNL_PORT_TX_DW2_GRP_D 0x162BC8 -#define _CNL_PORT_TX_DW2_GRP_F 0x162A48 -#define _CNL_PORT_TX_DW2_LN0_AE 0x162448 -#define _CNL_PORT_TX_DW2_LN0_B 0x162648 -#define _CNL_PORT_TX_DW2_LN0_C 0x162C48 -#define _CNL_PORT_TX_DW2_LN0_D 0x162E48 -#define _CNL_PORT_TX_DW2_LN0_F 0x162848 -#define CNL_PORT_TX_DW2_GRP(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW2_GRP_AE, \ - _CNL_PORT_TX_DW2_GRP_B, \ - _CNL_PORT_TX_DW2_GRP_C, \ - _CNL_PORT_TX_DW2_GRP_D, \ - _CNL_PORT_TX_DW2_GRP_AE, \ - _CNL_PORT_TX_DW2_GRP_F) -#define CNL_PORT_TX_DW2_LN0(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW2_LN0_AE, \ - _CNL_PORT_TX_DW2_LN0_B, \ - _CNL_PORT_TX_DW2_LN0_C, \ - _CNL_PORT_TX_DW2_LN0_D, \ - _CNL_PORT_TX_DW2_LN0_AE, \ - _CNL_PORT_TX_DW2_LN0_F) +/* CNL Port TX registers */ +#define _CNL_PORT_TX_AE_GRP_OFFSET 0x162340 +#define _CNL_PORT_TX_B_GRP_OFFSET 0x1623C0 +#define _CNL_PORT_TX_C_GRP_OFFSET 0x162B40 +#define _CNL_PORT_TX_D_GRP_OFFSET 0x162BC0 +#define _CNL_PORT_TX_F_GRP_OFFSET 0x162A40 +#define _CNL_PORT_TX_AE_LN0_OFFSET 0x162440 +#define _CNL_PORT_TX_B_LN0_OFFSET 0x162640 +#define _CNL_PORT_TX_C_LN0_OFFSET 0x162C40 +#define _CNL_PORT_TX_D_LN0_OFFSET 0x162E40 +#define _CNL_PORT_TX_F_LN0_OFFSET 0x162840 +#define _CNL_PORT_TX_DW_GRP(port, dw) (_PICK((port), \ + _CNL_PORT_TX_AE_GRP_OFFSET, \ + _CNL_PORT_TX_B_GRP_OFFSET, \ + _CNL_PORT_TX_B_GRP_OFFSET, \ + _CNL_PORT_TX_D_GRP_OFFSET, \ + _CNL_PORT_TX_AE_GRP_OFFSET, \ + _CNL_PORT_TX_F_GRP_OFFSET) + \ + 4*(dw)) +#define _CNL_PORT_TX_DW_LN0(port, dw) (_PICK((port), \ + _CNL_PORT_TX_AE_LN0_OFFSET, \ + _CNL_PORT_TX_B_LN0_OFFSET, \ + _CNL_PORT_TX_B_LN0_OFFSET, \ + _CNL_PORT_TX_D_LN0_OFFSET, \ + _CNL_PORT_TX_AE_LN0_OFFSET, \ + _CNL_PORT_TX_F_LN0_OFFSET) + \ + 4*(dw)) + +#define CNL_PORT_TX_DW2_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP((port), 2)) +#define CNL_PORT_TX_DW2_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0((port), 2)) #define SWING_SEL_UPPER(x) ((x >> 3) << 15) #define SWING_SEL_UPPER_MASK (1 << 15) #define SWING_SEL_LOWER(x) ((x & 0x7) << 11) @@ -1995,32 +1999,13 @@ enum i915_power_well_id { #define RCOMP_SCALAR(x) ((x) << 0) #define RCOMP_SCALAR_MASK (0xFF << 0) -#define _CNL_PORT_TX_DW4_GRP_AE 0x162350 -#define _CNL_PORT_TX_DW4_GRP_B 0x1623D0 -#define _CNL_PORT_TX_DW4_GRP_C 0x162B50 -#define _CNL_PORT_TX_DW4_GRP_D 0x162BD0 -#define _CNL_PORT_TX_DW4_GRP_F 0x162A50 #define _CNL_PORT_TX_DW4_LN0_AE 0x162450 #define _CNL_PORT_TX_DW4_LN1_AE 0x1624D0 -#define _CNL_PORT_TX_DW4_LN0_B 0x162650 -#define _CNL_PORT_TX_DW4_LN0_C 0x162C50 -#define _CNL_PORT_TX_DW4_LN0_D 0x162E50 -#define _CNL_PORT_TX_DW4_LN0_F 0x162850 -#define CNL_PORT_TX_DW4_GRP(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW4_GRP_AE, \ - _CNL_PORT_TX_DW4_GRP_B, \ - _CNL_PORT_TX_DW4_GRP_C, \ - _CNL_PORT_TX_DW4_GRP_D, \ - _CNL_PORT_TX_DW4_GRP_AE, \ - _CNL_PORT_TX_DW4_GRP_F) -#define CNL_PORT_TX_DW4_LN(port, ln) _MMIO_PORT6_LN(port, ln, \ - _CNL_PORT_TX_DW4_LN0_AE, \ - _CNL_PORT_TX_DW4_LN1_AE, \ - _CNL_PORT_TX_DW4_LN0_B, \ - _CNL_PORT_TX_DW4_LN0_C, \ - _CNL_PORT_TX_DW4_LN0_D, \ - _CNL_PORT_TX_DW4_LN0_AE, \ - _CNL_PORT_TX_DW4_LN0_F) +#define CNL_PORT_TX_DW4_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP((port), 4)) +#define CNL_PORT_TX_DW4_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0((port), 4)) +#define CNL_PORT_TX_DW4_LN(port, ln) _MMIO(_CNL_PORT_TX_DW_LN0((port), 4) + \ + (ln * (_CNL_PORT_TX_DW4_LN1_AE - \ + _CNL_PORT_TX_DW4_LN0_AE))) #define LOADGEN_SELECT (1 << 31) #define POST_CURSOR_1(x) ((x) << 12) #define POST_CURSOR_1_MASK (0x3F << 12) @@ -2029,30 +2014,8 @@ enum i915_power_well_id { #define CURSOR_COEFF(x) ((x) << 0) #define CURSOR_COEFF_MASK (0x3F << 0) -#define _CNL_PORT_TX_DW5_GRP_AE 0x162354 -#define _CNL_PORT_TX_DW5_GRP_B 0x1623D4 -#define _CNL_PORT_TX_DW5_GRP_C 0x162B54 -#define _CNL_PORT_TX_DW5_GRP_D 0x162BD4 -#define _CNL_PORT_TX_DW5_GRP_F 0x162A54 -#define _CNL_PORT_TX_DW5_LN0_AE 0x162454 -#define _CNL_PORT_TX_DW5_LN0_B 0x162654 -#define _CNL_PORT_TX_DW5_LN0_C 0x162C54 -#define _CNL_PORT_TX_DW5_LN0_D 0x162E54 -#define _CNL_PORT_TX_DW5_LN0_F 0x162854 -#define CNL_PORT_TX_DW5_GRP(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW5_GRP_AE, \ - _CNL_PORT_TX_DW5_GRP_B, \ - _CNL_PORT_TX_DW5_GRP_C, \ - _CNL_PORT_TX_DW5_GRP_D, \ - _CNL_PORT_TX_DW5_GRP_AE, \ - _CNL_PORT_TX_DW5_GRP_F) -#define CNL_PORT_TX_DW5_LN0(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW5_LN0_AE, \ - _CNL_PORT_TX_DW5_LN0_B, \ - _CNL_PORT_TX_DW5_LN0_C, \ - _CNL_PORT_TX_DW5_LN0_D, \ - _CNL_PORT_TX_DW5_LN0_AE, \ - _CNL_PORT_TX_DW5_LN0_F) +#define CNL_PORT_TX_DW5_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP((port), 5)) +#define CNL_PORT_TX_DW5_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0((port), 5)) #define TX_TRAINING_EN (1 << 31) #define TAP3_DISABLE (1 << 29) #define SCALING_MODE_SEL(x) ((x) << 18) @@ -2060,30 +2023,8 @@ enum i915_power_well_id { #define RTERM_SELECT(x) ((x) << 3) #define RTERM_SELECT_MASK (0x7 << 3) -#define _CNL_PORT_TX_DW7_GRP_AE 0x16235C -#define _CNL_PORT_TX_DW7_GRP_B 0x1623DC -#define _CNL_PORT_TX_DW7_GRP_C 0x162B5C -#define _CNL_PORT_TX_DW7_GRP_D 0x162BDC -#define _CNL_PORT_TX_DW7_GRP_F 0x162A5C -#define _CNL_PORT_TX_DW7_LN0_AE 0x16245C -#define _CNL_PORT_TX_DW7_LN0_B 0x16265C -#define _CNL_PORT_TX_DW7_LN0_C 0x162C5C -#define _CNL_PORT_TX_DW7_LN0_D 0x162E5C -#define _CNL_PORT_TX_DW7_LN0_F 0x16285C -#define CNL_PORT_TX_DW7_GRP(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW7_GRP_AE, \ - _CNL_PORT_TX_DW7_GRP_B, \ - _CNL_PORT_TX_DW7_GRP_C, \ - _CNL_PORT_TX_DW7_GRP_D, \ - _CNL_PORT_TX_DW7_GRP_AE, \ - _CNL_PORT_TX_DW7_GRP_F) -#define CNL_PORT_TX_DW7_LN0(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW7_LN0_AE, \ - _CNL_PORT_TX_DW7_LN0_B, \ - _CNL_PORT_TX_DW7_LN0_C, \ - _CNL_PORT_TX_DW7_LN0_D, \ - _CNL_PORT_TX_DW7_LN0_AE, \ - _CNL_PORT_TX_DW7_LN0_F) +#define CNL_PORT_TX_DW7_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP((port), 7)) +#define CNL_PORT_TX_DW7_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0((port), 7)) #define N_SCALAR(x) ((x) << 24) #define N_SCALAR_MASK (0x7F << 24) -- cgit v1.1 From da9cb11f76623b99f2d5e6aa68f43d6ef714a7de Mon Sep 17 00:00:00 2001 From: Mahesh Kumar Date: Wed, 14 Mar 2018 13:36:53 +0530 Subject: drm/i915/cnl: Kill _MMIO_PORT6 macro This patch replaces use of remaining _MMIO_PORT6 macro and removes the macro. Changes Since V1: - Rebase Signed-off-by: Mahesh Kumar Reviewed-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180314080653.9444-3-mahesh1.kumar@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 5a2a3d6..d965b4a 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -153,7 +153,6 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define _MMIO_PORT3(pipe, a, b, c) _MMIO(_PICK(pipe, a, b, c)) #define _PLL(pll, a, b) ((a) + (pll)*((b)-(a))) #define _MMIO_PLL(pll, a, b) _MMIO(_PLL(pll, a, b)) -#define _MMIO_PORT6(port, a, b, c, d, e, f) _MMIO(_PICK(port, a, b, c, d, e, f)) #define _PHY3(phy, ...) _PICK(phy, __VA_ARGS__) #define _MMIO_PHY3(phy, a, b, c) _MMIO(_PHY3(phy, a, b, c)) @@ -1946,20 +1945,21 @@ enum i915_power_well_id { #define _CNL_PORT_PCS_DW1_LN0_C 0x162C04 #define _CNL_PORT_PCS_DW1_LN0_D 0x162E04 #define _CNL_PORT_PCS_DW1_LN0_F 0x162804 -#define CNL_PORT_PCS_DW1_GRP(port) _MMIO_PORT6(port, \ +#define CNL_PORT_PCS_DW1_GRP(port) _MMIO(_PICK(port, \ _CNL_PORT_PCS_DW1_GRP_AE, \ _CNL_PORT_PCS_DW1_GRP_B, \ _CNL_PORT_PCS_DW1_GRP_C, \ _CNL_PORT_PCS_DW1_GRP_D, \ _CNL_PORT_PCS_DW1_GRP_AE, \ - _CNL_PORT_PCS_DW1_GRP_F) -#define CNL_PORT_PCS_DW1_LN0(port) _MMIO_PORT6(port, \ + _CNL_PORT_PCS_DW1_GRP_F)) + +#define CNL_PORT_PCS_DW1_LN0(port) _MMIO(_PICK(port, \ _CNL_PORT_PCS_DW1_LN0_AE, \ _CNL_PORT_PCS_DW1_LN0_B, \ _CNL_PORT_PCS_DW1_LN0_C, \ _CNL_PORT_PCS_DW1_LN0_D, \ _CNL_PORT_PCS_DW1_LN0_AE, \ - _CNL_PORT_PCS_DW1_LN0_F) + _CNL_PORT_PCS_DW1_LN0_F)) #define COMMON_KEEPER_EN (1 << 26) /* CNL Port TX registers */ -- cgit v1.1 From 80b216b98b0cd4f10303863c062b7ab7d117ada7 Mon Sep 17 00:00:00 2001 From: Daniele Ceraolo Spurio Date: Wed, 14 Mar 2018 11:26:50 -0700 Subject: drm/i915: store all mmio bases in intel_engines The mmio bases we're currently storing in the intel_engines array are only valid for a subset of gens, so we need to ignore them and use different values in some cases. Instead of doing that, we can have a table of [starting gen, mmio base] pairs for each engine in intel_engines and select the correct one based on the gen we're running on in a consistent way. v2: document that the list goes in reverse order, update starting gen for render (Chris) v3: starting gen for render back to 1 to make our life easier with selftests (Chris) Cc: Chris Wilson Cc: Mika Kuoppala Cc: Tvrtko Ursulin Signed-off-by: Daniele Ceraolo Spurio Reviewed-by: Chris Wilson #v2 Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180314182653.26981-1-daniele.ceraolospurio@intel.com --- drivers/gpu/drm/i915/intel_engine_cs.c | 78 +++++++++++++++++++++------------ drivers/gpu/drm/i915/intel_ringbuffer.c | 1 - 2 files changed, 50 insertions(+), 29 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index f22c5f7..71eac57 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -81,12 +81,17 @@ static const struct engine_class_info intel_engine_classes[] = { }, }; +#define MAX_MMIO_BASES 3 struct engine_info { unsigned int hw_id; unsigned int uabi_id; u8 class; u8 instance; - u32 mmio_base; + /* mmio bases table *must* be sorted in reverse gen order */ + struct engine_mmio_base { + u32 gen : 8; + u32 base : 24; + } mmio_bases[MAX_MMIO_BASES]; unsigned irq_shift; }; @@ -96,7 +101,9 @@ static const struct engine_info intel_engines[] = { .uabi_id = I915_EXEC_RENDER, .class = RENDER_CLASS, .instance = 0, - .mmio_base = RENDER_RING_BASE, + .mmio_bases = { + { .gen = 1, .base = RENDER_RING_BASE } + }, .irq_shift = GEN8_RCS_IRQ_SHIFT, }, [BCS] = { @@ -104,7 +111,9 @@ static const struct engine_info intel_engines[] = { .uabi_id = I915_EXEC_BLT, .class = COPY_ENGINE_CLASS, .instance = 0, - .mmio_base = BLT_RING_BASE, + .mmio_bases = { + { .gen = 6, .base = BLT_RING_BASE } + }, .irq_shift = GEN8_BCS_IRQ_SHIFT, }, [VCS] = { @@ -112,7 +121,11 @@ static const struct engine_info intel_engines[] = { .uabi_id = I915_EXEC_BSD, .class = VIDEO_DECODE_CLASS, .instance = 0, - .mmio_base = GEN6_BSD_RING_BASE, + .mmio_bases = { + { .gen = 11, .base = GEN11_BSD_RING_BASE }, + { .gen = 6, .base = GEN6_BSD_RING_BASE }, + { .gen = 4, .base = BSD_RING_BASE } + }, .irq_shift = GEN8_VCS1_IRQ_SHIFT, }, [VCS2] = { @@ -120,7 +133,10 @@ static const struct engine_info intel_engines[] = { .uabi_id = I915_EXEC_BSD, .class = VIDEO_DECODE_CLASS, .instance = 1, - .mmio_base = GEN8_BSD2_RING_BASE, + .mmio_bases = { + { .gen = 11, .base = GEN11_BSD2_RING_BASE }, + { .gen = 8, .base = GEN8_BSD2_RING_BASE } + }, .irq_shift = GEN8_VCS2_IRQ_SHIFT, }, [VCS3] = { @@ -128,7 +144,9 @@ static const struct engine_info intel_engines[] = { .uabi_id = I915_EXEC_BSD, .class = VIDEO_DECODE_CLASS, .instance = 2, - .mmio_base = GEN11_BSD3_RING_BASE, + .mmio_bases = { + { .gen = 11, .base = GEN11_BSD3_RING_BASE } + }, .irq_shift = 0, /* not used */ }, [VCS4] = { @@ -136,7 +154,9 @@ static const struct engine_info intel_engines[] = { .uabi_id = I915_EXEC_BSD, .class = VIDEO_DECODE_CLASS, .instance = 3, - .mmio_base = GEN11_BSD4_RING_BASE, + .mmio_bases = { + { .gen = 11, .base = GEN11_BSD4_RING_BASE } + }, .irq_shift = 0, /* not used */ }, [VECS] = { @@ -144,7 +164,10 @@ static const struct engine_info intel_engines[] = { .uabi_id = I915_EXEC_VEBOX, .class = VIDEO_ENHANCEMENT_CLASS, .instance = 0, - .mmio_base = VEBOX_RING_BASE, + .mmio_bases = { + { .gen = 11, .base = GEN11_VEBOX_RING_BASE }, + { .gen = 7, .base = VEBOX_RING_BASE } + }, .irq_shift = GEN8_VECS_IRQ_SHIFT, }, [VECS2] = { @@ -152,7 +175,9 @@ static const struct engine_info intel_engines[] = { .uabi_id = I915_EXEC_VEBOX, .class = VIDEO_ENHANCEMENT_CLASS, .instance = 1, - .mmio_base = GEN11_VEBOX2_RING_BASE, + .mmio_bases = { + { .gen = 11, .base = GEN11_VEBOX2_RING_BASE } + }, .irq_shift = 0, /* not used */ }, }; @@ -223,6 +248,21 @@ __intel_engine_context_size(struct drm_i915_private *dev_priv, u8 class) } } +static u32 __engine_mmio_base(struct drm_i915_private *i915, + const struct engine_mmio_base *bases) +{ + int i; + + for (i = 0; i < MAX_MMIO_BASES; i++) + if (INTEL_GEN(i915) >= bases[i].gen) + break; + + GEM_BUG_ON(i == MAX_MMIO_BASES); + GEM_BUG_ON(!bases[i].base); + + return bases[i].base; +} + static int intel_engine_setup(struct drm_i915_private *dev_priv, enum intel_engine_id id) @@ -257,25 +297,7 @@ intel_engine_setup(struct drm_i915_private *dev_priv, class_info->name, info->instance) >= sizeof(engine->name)); engine->hw_id = engine->guc_id = info->hw_id; - if (INTEL_GEN(dev_priv) >= 11) { - switch (engine->id) { - case VCS: - engine->mmio_base = GEN11_BSD_RING_BASE; - break; - case VCS2: - engine->mmio_base = GEN11_BSD2_RING_BASE; - break; - case VECS: - engine->mmio_base = GEN11_VEBOX_RING_BASE; - break; - default: - /* take the original value for all other engines */ - engine->mmio_base = info->mmio_base; - break; - } - } else { - engine->mmio_base = info->mmio_base; - } + engine->mmio_base = __engine_mmio_base(dev_priv, info->mmio_bases); engine->irq_shift = info->irq_shift; engine->class = info->class; engine->instance = info->instance; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 88eeb64..3b47876 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -2080,7 +2080,6 @@ int intel_init_bsd_ring_buffer(struct intel_engine_cs *engine) engine->emit_flush = gen6_bsd_ring_flush; engine->irq_enable_mask = GT_BSD_USER_INTERRUPT; } else { - engine->mmio_base = BSD_RING_BASE; engine->emit_flush = bsd_ring_flush; if (IS_GEN5(dev_priv)) engine->irq_enable_mask = ILK_BSD_USER_INTERRUPT; -- cgit v1.1 From 74419daaae6c1dafe9cc5d4d0c92c17982f4eebd Mon Sep 17 00:00:00 2001 From: Daniele Ceraolo Spurio Date: Wed, 14 Mar 2018 11:26:51 -0700 Subject: drm/i915: add a selftest for the mmio_bases table Check that the entries are in reverse gen order and that all entries with gen > 0 have an mmio base set. v2: loop forward, simplify logic, use i915_subtests (Chris) Suggested-by: Chris Wilson Cc: Chris Wilson Signed-off-by: Daniele Ceraolo Spurio Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180314182653.26981-2-daniele.ceraolospurio@intel.com --- drivers/gpu/drm/i915/intel_engine_cs.c | 16 +++--- .../gpu/drm/i915/selftests/i915_mock_selftests.h | 1 + drivers/gpu/drm/i915/selftests/intel_engine_cs.c | 58 ++++++++++++++++++++++ 3 files changed, 69 insertions(+), 6 deletions(-) create mode 100644 drivers/gpu/drm/i915/selftests/intel_engine_cs.c diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 71eac57..8fda811 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -263,16 +263,21 @@ static u32 __engine_mmio_base(struct drm_i915_private *i915, return bases[i].base; } +static void __sprint_engine_name(char *name, const struct engine_info *info) +{ + WARN_ON(snprintf(name, INTEL_ENGINE_CS_MAX_NAME, "%s%u", + intel_engine_classes[info->class].name, + info->instance) >= INTEL_ENGINE_CS_MAX_NAME); +} + static int intel_engine_setup(struct drm_i915_private *dev_priv, enum intel_engine_id id) { const struct engine_info *info = &intel_engines[id]; - const struct engine_class_info *class_info; struct intel_engine_cs *engine; GEM_BUG_ON(info->class >= ARRAY_SIZE(intel_engine_classes)); - class_info = &intel_engine_classes[info->class]; BUILD_BUG_ON(MAX_ENGINE_CLASS >= BIT(GEN11_ENGINE_CLASS_WIDTH)); BUILD_BUG_ON(MAX_ENGINE_INSTANCE >= BIT(GEN11_ENGINE_INSTANCE_WIDTH)); @@ -293,9 +298,7 @@ intel_engine_setup(struct drm_i915_private *dev_priv, engine->id = id; engine->i915 = dev_priv; - WARN_ON(snprintf(engine->name, sizeof(engine->name), "%s%u", - class_info->name, info->instance) >= - sizeof(engine->name)); + __sprint_engine_name(engine->name, info); engine->hw_id = engine->guc_id = info->hw_id; engine->mmio_base = __engine_mmio_base(dev_priv, info->mmio_bases); engine->irq_shift = info->irq_shift; @@ -303,7 +306,7 @@ intel_engine_setup(struct drm_i915_private *dev_priv, engine->instance = info->instance; engine->uabi_id = info->uabi_id; - engine->uabi_class = class_info->uabi_class; + engine->uabi_class = intel_engine_classes[info->class].uabi_class; engine->context_size = __intel_engine_context_size(dev_priv, engine->class); @@ -2140,4 +2143,5 @@ void intel_disable_engine_stats(struct intel_engine_cs *engine) #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) #include "selftests/mock_engine.c" +#include "selftests/intel_engine_cs.c" #endif diff --git a/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h b/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h index 9a48aa44..d16d741 100644 --- a/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h +++ b/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h @@ -14,6 +14,7 @@ selftest(fence, i915_sw_fence_mock_selftests) selftest(scatterlist, scatterlist_mock_selftests) selftest(syncmap, i915_syncmap_mock_selftests) selftest(uncore, intel_uncore_mock_selftests) +selftest(engine, intel_engine_cs_mock_selftests) selftest(breadcrumbs, intel_breadcrumbs_mock_selftests) selftest(timelines, i915_gem_timeline_mock_selftests) selftest(requests, i915_request_mock_selftests) diff --git a/drivers/gpu/drm/i915/selftests/intel_engine_cs.c b/drivers/gpu/drm/i915/selftests/intel_engine_cs.c new file mode 100644 index 0000000..cfaa6b29 --- /dev/null +++ b/drivers/gpu/drm/i915/selftests/intel_engine_cs.c @@ -0,0 +1,58 @@ +/* + * SPDX-License-Identifier: GPL-2.0 + * + * Copyright © 2018 Intel Corporation + */ + +#include "../i915_selftest.h" + +static int intel_mmio_bases_check(void *arg) +{ + int i, j; + + for (i = 0; i < ARRAY_SIZE(intel_engines); i++) { + const struct engine_info *info = &intel_engines[i]; + char name[INTEL_ENGINE_CS_MAX_NAME]; + u8 prev = U8_MAX; + + __sprint_engine_name(name, info); + + for (j = 0; j < MAX_MMIO_BASES; j++) { + u8 gen = info->mmio_bases[j].gen; + u32 base = info->mmio_bases[j].base; + + if (gen >= prev) { + pr_err("%s: %s: mmio base for gen %x " + "is before the one for gen %x\n", + __func__, name, prev, gen); + return -EINVAL; + } + + if (gen == 0) + break; + + if (!base) { + pr_err("%s: %s: invalid mmio base (%x) " + "for gen %x at entry %u\n", + __func__, name, base, gen, j); + return -EINVAL; + } + + prev = gen; + } + + pr_info("%s: min gen supported for %s = %d\n", + __func__, name, prev); + } + + return 0; +} + +int intel_engine_cs_mock_selftests(void) +{ + static const struct i915_subtest tests[] = { + SUBTEST(intel_mmio_bases_check), + }; + + return i915_subtests(tests, NULL); +} -- cgit v1.1 From 210060edc216ebd6330ee4fded5a01547d938642 Mon Sep 17 00:00:00 2001 From: Daniele Ceraolo Spurio Date: Wed, 14 Mar 2018 11:26:52 -0700 Subject: drm/i915: use engine->irq_keep_mask when resetting irqs The "reset" value and the "keep" value are the same. While we are here, add a TODO for gen11 interrupt reset Suggested-by: Chris Wilson Cc: Chris Wilson Signed-off-by: Daniele Ceraolo Spurio Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180314182653.26981-3-daniele.ceraolospurio@intel.com --- drivers/gpu/drm/i915/intel_lrc.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 3a69b36..5e8f689 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -1666,6 +1666,10 @@ static void reset_irq(struct intel_engine_cs *engine) struct drm_i915_private *dev_priv = engine->i915; int i; + /* TODO: correctly reset irqs for gen11 */ + if (WARN_ON_ONCE(INTEL_GEN(engine->i915) >= 11)) + return; + GEM_BUG_ON(engine->id >= ARRAY_SIZE(gtiir)); /* @@ -1677,11 +1681,11 @@ static void reset_irq(struct intel_engine_cs *engine) */ for (i = 0; i < 2; i++) { I915_WRITE(GEN8_GT_IIR(gtiir[engine->id]), - GT_CONTEXT_SWITCH_INTERRUPT << engine->irq_shift); + engine->irq_keep_mask); POSTING_READ(GEN8_GT_IIR(gtiir[engine->id])); } GEM_BUG_ON(I915_READ(GEN8_GT_IIR(gtiir[engine->id])) & - (GT_CONTEXT_SWITCH_INTERRUPT << engine->irq_shift)); + engine->irq_keep_mask); clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); } -- cgit v1.1 From fa6f071d54fb3658c7012634b8e4035c8d3a25bc Mon Sep 17 00:00:00 2001 From: Daniele Ceraolo Spurio Date: Wed, 14 Mar 2018 11:26:53 -0700 Subject: drm/i915: move gen8 irq shifts to intel_lrc.c The only usage outside the intel_lrc.c file is in the ringbuffer init, but the irq mask calculated there is then overwritten for all engines that have a non-zero shift, so we can drop it. This change is not aimed at code saving but at removing from intel_engines information that does not apply to all gens that have the engine. When checking without the temporary WARN_ON, code size is basically unchanged. v2: make the irq_shifts array static const v3: rebase, move irq_shifts array to logical_ring_default_irqs v4: move array inside the if and use u8 for it (Chris) Suggested-by: Michel Thierry Signed-off-by: Daniele Ceraolo Spurio Cc: Chris Wilson Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180314182653.26981-4-daniele.ceraolospurio@intel.com --- drivers/gpu/drm/i915/intel_engine_cs.c | 10 ---------- drivers/gpu/drm/i915/intel_lrc.c | 15 ++++++++++++++- drivers/gpu/drm/i915/intel_ringbuffer.c | 4 ++-- drivers/gpu/drm/i915/intel_ringbuffer.h | 1 - 4 files changed, 16 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 8fda811..337dfa5 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -92,7 +92,6 @@ struct engine_info { u32 gen : 8; u32 base : 24; } mmio_bases[MAX_MMIO_BASES]; - unsigned irq_shift; }; static const struct engine_info intel_engines[] = { @@ -104,7 +103,6 @@ static const struct engine_info intel_engines[] = { .mmio_bases = { { .gen = 1, .base = RENDER_RING_BASE } }, - .irq_shift = GEN8_RCS_IRQ_SHIFT, }, [BCS] = { .hw_id = BCS_HW, @@ -114,7 +112,6 @@ static const struct engine_info intel_engines[] = { .mmio_bases = { { .gen = 6, .base = BLT_RING_BASE } }, - .irq_shift = GEN8_BCS_IRQ_SHIFT, }, [VCS] = { .hw_id = VCS_HW, @@ -126,7 +123,6 @@ static const struct engine_info intel_engines[] = { { .gen = 6, .base = GEN6_BSD_RING_BASE }, { .gen = 4, .base = BSD_RING_BASE } }, - .irq_shift = GEN8_VCS1_IRQ_SHIFT, }, [VCS2] = { .hw_id = VCS2_HW, @@ -137,7 +133,6 @@ static const struct engine_info intel_engines[] = { { .gen = 11, .base = GEN11_BSD2_RING_BASE }, { .gen = 8, .base = GEN8_BSD2_RING_BASE } }, - .irq_shift = GEN8_VCS2_IRQ_SHIFT, }, [VCS3] = { .hw_id = VCS3_HW, @@ -147,7 +142,6 @@ static const struct engine_info intel_engines[] = { .mmio_bases = { { .gen = 11, .base = GEN11_BSD3_RING_BASE } }, - .irq_shift = 0, /* not used */ }, [VCS4] = { .hw_id = VCS4_HW, @@ -157,7 +151,6 @@ static const struct engine_info intel_engines[] = { .mmio_bases = { { .gen = 11, .base = GEN11_BSD4_RING_BASE } }, - .irq_shift = 0, /* not used */ }, [VECS] = { .hw_id = VECS_HW, @@ -168,7 +161,6 @@ static const struct engine_info intel_engines[] = { { .gen = 11, .base = GEN11_VEBOX_RING_BASE }, { .gen = 7, .base = VEBOX_RING_BASE } }, - .irq_shift = GEN8_VECS_IRQ_SHIFT, }, [VECS2] = { .hw_id = VECS2_HW, @@ -178,7 +170,6 @@ static const struct engine_info intel_engines[] = { .mmio_bases = { { .gen = 11, .base = GEN11_VEBOX2_RING_BASE } }, - .irq_shift = 0, /* not used */ }, }; @@ -301,7 +292,6 @@ intel_engine_setup(struct drm_i915_private *dev_priv, __sprint_engine_name(engine->name, info); engine->hw_id = engine->guc_id = info->hw_id; engine->mmio_base = __engine_mmio_base(dev_priv, info->mmio_bases); - engine->irq_shift = info->irq_shift; engine->class = info->class; engine->instance = info->instance; diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 5e8f689..53f1c00 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -2118,7 +2118,20 @@ logical_ring_default_vfuncs(struct intel_engine_cs *engine) static inline void logical_ring_default_irqs(struct intel_engine_cs *engine) { - unsigned shift = engine->irq_shift; + unsigned int shift = 0; + + if (INTEL_GEN(engine->i915) < 11) { + const u8 irq_shifts[] = { + [RCS] = GEN8_RCS_IRQ_SHIFT, + [BCS] = GEN8_BCS_IRQ_SHIFT, + [VCS] = GEN8_VCS1_IRQ_SHIFT, + [VCS2] = GEN8_VCS2_IRQ_SHIFT, + [VECS] = GEN8_VECS_IRQ_SHIFT, + }; + + shift = irq_shifts[engine->id]; + } + engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT << shift; engine->irq_keep_mask = GT_CONTEXT_SWITCH_INTERRUPT << shift; } diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 3b47876..72d6167 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1944,8 +1944,6 @@ static void intel_ring_init_semaphores(struct drm_i915_private *dev_priv, static void intel_ring_init_irq(struct drm_i915_private *dev_priv, struct intel_engine_cs *engine) { - engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT << engine->irq_shift; - if (INTEL_GEN(dev_priv) >= 6) { engine->irq_enable = gen6_irq_enable; engine->irq_disable = gen6_irq_disable; @@ -2030,6 +2028,8 @@ int intel_init_render_ring_buffer(struct intel_engine_cs *engine) if (HAS_L3_DPF(dev_priv)) engine->irq_keep_mask = GT_RENDER_L3_PARITY_ERROR_INTERRUPT; + engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT; + if (INTEL_GEN(dev_priv) >= 6) { engine->init_context = intel_rcs_ctx_init; engine->emit_flush = gen7_render_ring_flush; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 81cdbbf..80fae80 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -331,7 +331,6 @@ struct intel_engine_cs { u8 instance; u32 context_size; u32 mmio_base; - unsigned int irq_shift; struct intel_ring *buffer; struct intel_timeline *timeline; -- cgit v1.1 From c080363fcdc3015e4cb9b5582afe2cd3aa890630 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Tue, 13 Mar 2018 23:19:20 +0000 Subject: drm/i915: Split GPU commands definitions into separate header We should not mix MMIO with MI_INSTR definitions. v2: sanitize comment, change include order (Chris) Suggested-by: Chris Wilson Signed-off-by: Michal Wajdeczko Cc: Chris Wilson Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180313124109.39216-1-michal.wajdeczko@intel.com Link: https://patchwork.freedesktop.org/patch/msgid/20180313231920.6932-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_reg.h | 263 ---------------------------- drivers/gpu/drm/i915/intel_gpu_commands.h | 274 ++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_ringbuffer.h | 3 +- 3 files changed, 276 insertions(+), 264 deletions(-) create mode 100644 drivers/gpu/drm/i915/intel_gpu_commands.h diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index d965b4a..dbcb882 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -427,145 +427,6 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define VGA_CR_INDEX_CGA 0x3d4 #define VGA_CR_DATA_CGA 0x3d5 -/* - * Instruction field definitions used by the command parser - */ -#define INSTR_CLIENT_SHIFT 29 -#define INSTR_MI_CLIENT 0x0 -#define INSTR_BC_CLIENT 0x2 -#define INSTR_RC_CLIENT 0x3 -#define INSTR_SUBCLIENT_SHIFT 27 -#define INSTR_SUBCLIENT_MASK 0x18000000 -#define INSTR_MEDIA_SUBCLIENT 0x2 -#define INSTR_26_TO_24_MASK 0x7000000 -#define INSTR_26_TO_24_SHIFT 24 - -/* - * Memory interface instructions used by the kernel - */ -#define MI_INSTR(opcode, flags) (((opcode) << 23) | (flags)) -/* Many MI commands use bit 22 of the header dword for GGTT vs PPGTT */ -#define MI_GLOBAL_GTT (1<<22) - -#define MI_NOOP MI_INSTR(0, 0) -#define MI_USER_INTERRUPT MI_INSTR(0x02, 0) -#define MI_WAIT_FOR_EVENT MI_INSTR(0x03, 0) -#define MI_WAIT_FOR_OVERLAY_FLIP (1<<16) -#define MI_WAIT_FOR_PLANE_B_FLIP (1<<6) -#define MI_WAIT_FOR_PLANE_A_FLIP (1<<2) -#define MI_WAIT_FOR_PLANE_A_SCANLINES (1<<1) -#define MI_FLUSH MI_INSTR(0x04, 0) -#define MI_READ_FLUSH (1 << 0) -#define MI_EXE_FLUSH (1 << 1) -#define MI_NO_WRITE_FLUSH (1 << 2) -#define MI_SCENE_COUNT (1 << 3) /* just increment scene count */ -#define MI_END_SCENE (1 << 4) /* flush binner and incr scene count */ -#define MI_INVALIDATE_ISP (1 << 5) /* invalidate indirect state pointers */ -#define MI_REPORT_HEAD MI_INSTR(0x07, 0) -#define MI_ARB_ON_OFF MI_INSTR(0x08, 0) -#define MI_ARB_ENABLE (1<<0) -#define MI_ARB_DISABLE (0<<0) -#define MI_BATCH_BUFFER_END MI_INSTR(0x0a, 0) -#define MI_SUSPEND_FLUSH MI_INSTR(0x0b, 0) -#define MI_SUSPEND_FLUSH_EN (1<<0) -#define MI_SET_APPID MI_INSTR(0x0e, 0) -#define MI_OVERLAY_FLIP MI_INSTR(0x11, 0) -#define MI_OVERLAY_CONTINUE (0x0<<21) -#define MI_OVERLAY_ON (0x1<<21) -#define MI_OVERLAY_OFF (0x2<<21) -#define MI_LOAD_SCAN_LINES_INCL MI_INSTR(0x12, 0) -#define MI_DISPLAY_FLIP MI_INSTR(0x14, 2) -#define MI_DISPLAY_FLIP_I915 MI_INSTR(0x14, 1) -#define MI_DISPLAY_FLIP_PLANE(n) ((n) << 20) -/* IVB has funny definitions for which plane to flip. */ -#define MI_DISPLAY_FLIP_IVB_PLANE_A (0 << 19) -#define MI_DISPLAY_FLIP_IVB_PLANE_B (1 << 19) -#define MI_DISPLAY_FLIP_IVB_SPRITE_A (2 << 19) -#define MI_DISPLAY_FLIP_IVB_SPRITE_B (3 << 19) -#define MI_DISPLAY_FLIP_IVB_PLANE_C (4 << 19) -#define MI_DISPLAY_FLIP_IVB_SPRITE_C (5 << 19) -/* SKL ones */ -#define MI_DISPLAY_FLIP_SKL_PLANE_1_A (0 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_1_B (1 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_1_C (2 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_2_A (4 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_2_B (5 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_2_C (6 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_3_A (7 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_3_B (8 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_3_C (9 << 8) -#define MI_SEMAPHORE_MBOX MI_INSTR(0x16, 1) /* gen6, gen7 */ -#define MI_SEMAPHORE_GLOBAL_GTT (1<<22) -#define MI_SEMAPHORE_UPDATE (1<<21) -#define MI_SEMAPHORE_COMPARE (1<<20) -#define MI_SEMAPHORE_REGISTER (1<<18) -#define MI_SEMAPHORE_SYNC_VR (0<<16) /* RCS wait for VCS (RVSYNC) */ -#define MI_SEMAPHORE_SYNC_VER (1<<16) /* RCS wait for VECS (RVESYNC) */ -#define MI_SEMAPHORE_SYNC_BR (2<<16) /* RCS wait for BCS (RBSYNC) */ -#define MI_SEMAPHORE_SYNC_BV (0<<16) /* VCS wait for BCS (VBSYNC) */ -#define MI_SEMAPHORE_SYNC_VEV (1<<16) /* VCS wait for VECS (VVESYNC) */ -#define MI_SEMAPHORE_SYNC_RV (2<<16) /* VCS wait for RCS (VRSYNC) */ -#define MI_SEMAPHORE_SYNC_RB (0<<16) /* BCS wait for RCS (BRSYNC) */ -#define MI_SEMAPHORE_SYNC_VEB (1<<16) /* BCS wait for VECS (BVESYNC) */ -#define MI_SEMAPHORE_SYNC_VB (2<<16) /* BCS wait for VCS (BVSYNC) */ -#define MI_SEMAPHORE_SYNC_BVE (0<<16) /* VECS wait for BCS (VEBSYNC) */ -#define MI_SEMAPHORE_SYNC_VVE (1<<16) /* VECS wait for VCS (VEVSYNC) */ -#define MI_SEMAPHORE_SYNC_RVE (2<<16) /* VECS wait for RCS (VERSYNC) */ -#define MI_SEMAPHORE_SYNC_INVALID (3<<16) -#define MI_SEMAPHORE_SYNC_MASK (3<<16) -#define MI_SET_CONTEXT MI_INSTR(0x18, 0) -#define MI_MM_SPACE_GTT (1<<8) -#define MI_MM_SPACE_PHYSICAL (0<<8) -#define MI_SAVE_EXT_STATE_EN (1<<3) -#define MI_RESTORE_EXT_STATE_EN (1<<2) -#define MI_FORCE_RESTORE (1<<1) -#define MI_RESTORE_INHIBIT (1<<0) -#define HSW_MI_RS_SAVE_STATE_EN (1<<3) -#define HSW_MI_RS_RESTORE_STATE_EN (1<<2) -#define MI_SEMAPHORE_SIGNAL MI_INSTR(0x1b, 0) /* GEN8+ */ -#define MI_SEMAPHORE_TARGET(engine) ((engine)<<15) -#define MI_SEMAPHORE_WAIT MI_INSTR(0x1c, 2) /* GEN8+ */ -#define MI_SEMAPHORE_POLL (1<<15) -#define MI_SEMAPHORE_SAD_GTE_SDD (1<<12) -#define MI_STORE_DWORD_IMM MI_INSTR(0x20, 1) -#define MI_STORE_DWORD_IMM_GEN4 MI_INSTR(0x20, 2) -#define MI_MEM_VIRTUAL (1 << 22) /* 945,g33,965 */ -#define MI_USE_GGTT (1 << 22) /* g4x+ */ -#define MI_STORE_DWORD_INDEX MI_INSTR(0x21, 1) -#define MI_STORE_DWORD_INDEX_SHIFT 2 -/* Official intel docs are somewhat sloppy concerning MI_LOAD_REGISTER_IMM: - * - Always issue a MI_NOOP _before_ the MI_LOAD_REGISTER_IMM - otherwise hw - * simply ignores the register load under certain conditions. - * - One can actually load arbitrary many arbitrary registers: Simply issue x - * address/value pairs. Don't overdue it, though, x <= 2^4 must hold! - */ -#define MI_LOAD_REGISTER_IMM(x) MI_INSTR(0x22, 2*(x)-1) -#define MI_LRI_FORCE_POSTED (1<<12) -#define MI_STORE_REGISTER_MEM MI_INSTR(0x24, 1) -#define MI_STORE_REGISTER_MEM_GEN8 MI_INSTR(0x24, 2) -#define MI_SRM_LRM_GLOBAL_GTT (1<<22) -#define MI_FLUSH_DW MI_INSTR(0x26, 1) /* for GEN6 */ -#define MI_FLUSH_DW_STORE_INDEX (1<<21) -#define MI_INVALIDATE_TLB (1<<18) -#define MI_FLUSH_DW_OP_STOREDW (1<<14) -#define MI_FLUSH_DW_OP_MASK (3<<14) -#define MI_FLUSH_DW_NOTIFY (1<<8) -#define MI_INVALIDATE_BSD (1<<7) -#define MI_FLUSH_DW_USE_GTT (1<<2) -#define MI_FLUSH_DW_USE_PPGTT (0<<2) -#define MI_LOAD_REGISTER_MEM MI_INSTR(0x29, 1) -#define MI_LOAD_REGISTER_MEM_GEN8 MI_INSTR(0x29, 2) -#define MI_BATCH_BUFFER MI_INSTR(0x30, 1) -#define MI_BATCH_NON_SECURE (1) -/* for snb/ivb/vlv this also means "batch in ppgtt" when ppgtt is enabled. */ -#define MI_BATCH_NON_SECURE_I965 (1<<8) -#define MI_BATCH_PPGTT_HSW (1<<8) -#define MI_BATCH_NON_SECURE_HSW (1<<13) -#define MI_BATCH_BUFFER_START MI_INSTR(0x31, 0) -#define MI_BATCH_GTT (2<<6) /* aliased with (1<<7) on gen4 */ -#define MI_BATCH_BUFFER_START_GEN8 MI_INSTR(0x31, 1) -#define MI_BATCH_RESOURCE_STREAMER (1<<10) - #define MI_PREDICATE_SRC0 _MMIO(0x2400) #define MI_PREDICATE_SRC0_UDW _MMIO(0x2400 + 4) #define MI_PREDICATE_SRC1 _MMIO(0x2408) @@ -576,130 +437,6 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define LOWER_SLICE_DISABLED (0<<0) /* - * 3D instructions used by the kernel - */ -#define GFX_INSTR(opcode, flags) ((0x3 << 29) | ((opcode) << 24) | (flags)) - -#define GEN9_MEDIA_POOL_STATE ((0x3 << 29) | (0x2 << 27) | (0x5 << 16) | 4) -#define GEN9_MEDIA_POOL_ENABLE (1 << 31) -#define GFX_OP_RASTER_RULES ((0x3<<29)|(0x7<<24)) -#define GFX_OP_SCISSOR ((0x3<<29)|(0x1c<<24)|(0x10<<19)) -#define SC_UPDATE_SCISSOR (0x1<<1) -#define SC_ENABLE_MASK (0x1<<0) -#define SC_ENABLE (0x1<<0) -#define GFX_OP_LOAD_INDIRECT ((0x3<<29)|(0x1d<<24)|(0x7<<16)) -#define GFX_OP_SCISSOR_INFO ((0x3<<29)|(0x1d<<24)|(0x81<<16)|(0x1)) -#define SCI_YMIN_MASK (0xffff<<16) -#define SCI_XMIN_MASK (0xffff<<0) -#define SCI_YMAX_MASK (0xffff<<16) -#define SCI_XMAX_MASK (0xffff<<0) -#define GFX_OP_SCISSOR_ENABLE ((0x3<<29)|(0x1c<<24)|(0x10<<19)) -#define GFX_OP_SCISSOR_RECT ((0x3<<29)|(0x1d<<24)|(0x81<<16)|1) -#define GFX_OP_COLOR_FACTOR ((0x3<<29)|(0x1d<<24)|(0x1<<16)|0x0) -#define GFX_OP_STIPPLE ((0x3<<29)|(0x1d<<24)|(0x83<<16)) -#define GFX_OP_MAP_INFO ((0x3<<29)|(0x1d<<24)|0x4) -#define GFX_OP_DESTBUFFER_VARS ((0x3<<29)|(0x1d<<24)|(0x85<<16)|0x0) -#define GFX_OP_DESTBUFFER_INFO ((0x3<<29)|(0x1d<<24)|(0x8e<<16)|1) -#define GFX_OP_DRAWRECT_INFO ((0x3<<29)|(0x1d<<24)|(0x80<<16)|(0x3)) -#define GFX_OP_DRAWRECT_INFO_I965 ((0x7900<<16)|0x2) - -#define COLOR_BLT_CMD (2<<29 | 0x40<<22 | (5-2)) -#define SRC_COPY_BLT_CMD ((2<<29)|(0x43<<22)|4) -#define XY_SRC_COPY_BLT_CMD ((2<<29)|(0x53<<22)|6) -#define XY_MONO_SRC_COPY_IMM_BLT ((2<<29)|(0x71<<22)|5) -#define BLT_WRITE_A (2<<20) -#define BLT_WRITE_RGB (1<<20) -#define BLT_WRITE_RGBA (BLT_WRITE_RGB | BLT_WRITE_A) -#define BLT_DEPTH_8 (0<<24) -#define BLT_DEPTH_16_565 (1<<24) -#define BLT_DEPTH_16_1555 (2<<24) -#define BLT_DEPTH_32 (3<<24) -#define BLT_ROP_SRC_COPY (0xcc<<16) -#define BLT_ROP_COLOR_COPY (0xf0<<16) -#define XY_SRC_COPY_BLT_SRC_TILED (1<<15) /* 965+ only */ -#define XY_SRC_COPY_BLT_DST_TILED (1<<11) /* 965+ only */ -#define CMD_OP_DISPLAYBUFFER_INFO ((0x0<<29)|(0x14<<23)|2) -#define ASYNC_FLIP (1<<22) -#define DISPLAY_PLANE_A (0<<20) -#define DISPLAY_PLANE_B (1<<20) -#define GFX_OP_PIPE_CONTROL(len) ((0x3<<29)|(0x3<<27)|(0x2<<24)|((len)-2)) -#define PIPE_CONTROL_FLUSH_L3 (1<<27) -#define PIPE_CONTROL_GLOBAL_GTT_IVB (1<<24) /* gen7+ */ -#define PIPE_CONTROL_MMIO_WRITE (1<<23) -#define PIPE_CONTROL_STORE_DATA_INDEX (1<<21) -#define PIPE_CONTROL_CS_STALL (1<<20) -#define PIPE_CONTROL_TLB_INVALIDATE (1<<18) -#define PIPE_CONTROL_MEDIA_STATE_CLEAR (1<<16) -#define PIPE_CONTROL_QW_WRITE (1<<14) -#define PIPE_CONTROL_POST_SYNC_OP_MASK (3<<14) -#define PIPE_CONTROL_DEPTH_STALL (1<<13) -#define PIPE_CONTROL_WRITE_FLUSH (1<<12) -#define PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH (1<<12) /* gen6+ */ -#define PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE (1<<11) /* MBZ on Ironlake */ -#define PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE (1<<10) /* GM45+ only */ -#define PIPE_CONTROL_INDIRECT_STATE_DISABLE (1<<9) -#define PIPE_CONTROL_NOTIFY (1<<8) -#define PIPE_CONTROL_FLUSH_ENABLE (1<<7) /* gen7+ */ -#define PIPE_CONTROL_DC_FLUSH_ENABLE (1<<5) -#define PIPE_CONTROL_VF_CACHE_INVALIDATE (1<<4) -#define PIPE_CONTROL_CONST_CACHE_INVALIDATE (1<<3) -#define PIPE_CONTROL_STATE_CACHE_INVALIDATE (1<<2) -#define PIPE_CONTROL_STALL_AT_SCOREBOARD (1<<1) -#define PIPE_CONTROL_DEPTH_CACHE_FLUSH (1<<0) -#define PIPE_CONTROL_GLOBAL_GTT (1<<2) /* in addr dword */ - -/* - * Commands used only by the command parser - */ -#define MI_SET_PREDICATE MI_INSTR(0x01, 0) -#define MI_ARB_CHECK MI_INSTR(0x05, 0) -#define MI_RS_CONTROL MI_INSTR(0x06, 0) -#define MI_URB_ATOMIC_ALLOC MI_INSTR(0x09, 0) -#define MI_PREDICATE MI_INSTR(0x0C, 0) -#define MI_RS_CONTEXT MI_INSTR(0x0F, 0) -#define MI_TOPOLOGY_FILTER MI_INSTR(0x0D, 0) -#define MI_LOAD_SCAN_LINES_EXCL MI_INSTR(0x13, 0) -#define MI_URB_CLEAR MI_INSTR(0x19, 0) -#define MI_UPDATE_GTT MI_INSTR(0x23, 0) -#define MI_CLFLUSH MI_INSTR(0x27, 0) -#define MI_REPORT_PERF_COUNT MI_INSTR(0x28, 0) -#define MI_REPORT_PERF_COUNT_GGTT (1<<0) -#define MI_LOAD_REGISTER_REG MI_INSTR(0x2A, 0) -#define MI_RS_STORE_DATA_IMM MI_INSTR(0x2B, 0) -#define MI_LOAD_URB_MEM MI_INSTR(0x2C, 0) -#define MI_STORE_URB_MEM MI_INSTR(0x2D, 0) -#define MI_CONDITIONAL_BATCH_BUFFER_END MI_INSTR(0x36, 0) - -#define PIPELINE_SELECT ((0x3<<29)|(0x1<<27)|(0x1<<24)|(0x4<<16)) -#define GFX_OP_3DSTATE_VF_STATISTICS ((0x3<<29)|(0x1<<27)|(0x0<<24)|(0xB<<16)) -#define MEDIA_VFE_STATE ((0x3<<29)|(0x2<<27)|(0x0<<24)|(0x0<<16)) -#define MEDIA_VFE_STATE_MMIO_ACCESS_MASK (0x18) -#define GPGPU_OBJECT ((0x3<<29)|(0x2<<27)|(0x1<<24)|(0x4<<16)) -#define GPGPU_WALKER ((0x3<<29)|(0x2<<27)|(0x1<<24)|(0x5<<16)) -#define GFX_OP_3DSTATE_DX9_CONSTANTF_VS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x39<<16)) -#define GFX_OP_3DSTATE_DX9_CONSTANTF_PS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x3A<<16)) -#define GFX_OP_3DSTATE_SO_DECL_LIST \ - ((0x3<<29)|(0x3<<27)|(0x1<<24)|(0x17<<16)) - -#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_VS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x43<<16)) -#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_GS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x44<<16)) -#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_HS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x45<<16)) -#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_DS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x46<<16)) -#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_PS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x47<<16)) - -#define MFX_WAIT ((0x3<<29)|(0x1<<27)|(0x0<<16)) - -#define COLOR_BLT ((0x2<<29)|(0x40<<22)) -#define SRC_COPY_BLT ((0x2<<29)|(0x43<<22)) - -/* * Registers used only by the command parser */ #define BCS_SWCTRL _MMIO(0x22200) diff --git a/drivers/gpu/drm/i915/intel_gpu_commands.h b/drivers/gpu/drm/i915/intel_gpu_commands.h new file mode 100644 index 0000000..105e2a9 --- /dev/null +++ b/drivers/gpu/drm/i915/intel_gpu_commands.h @@ -0,0 +1,274 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright � 2003-2018 Intel Corporation + */ + +#ifndef _INTEL_GPU_COMMANDS_H_ +#define _INTEL_GPU_COMMANDS_H_ + +/* + * Instruction field definitions used by the command parser + */ +#define INSTR_CLIENT_SHIFT 29 +#define INSTR_MI_CLIENT 0x0 +#define INSTR_BC_CLIENT 0x2 +#define INSTR_RC_CLIENT 0x3 +#define INSTR_SUBCLIENT_SHIFT 27 +#define INSTR_SUBCLIENT_MASK 0x18000000 +#define INSTR_MEDIA_SUBCLIENT 0x2 +#define INSTR_26_TO_24_MASK 0x7000000 +#define INSTR_26_TO_24_SHIFT 24 + +/* + * Memory interface instructions used by the kernel + */ +#define MI_INSTR(opcode, flags) (((opcode) << 23) | (flags)) +/* Many MI commands use bit 22 of the header dword for GGTT vs PPGTT */ +#define MI_GLOBAL_GTT (1<<22) + +#define MI_NOOP MI_INSTR(0, 0) +#define MI_USER_INTERRUPT MI_INSTR(0x02, 0) +#define MI_WAIT_FOR_EVENT MI_INSTR(0x03, 0) +#define MI_WAIT_FOR_OVERLAY_FLIP (1<<16) +#define MI_WAIT_FOR_PLANE_B_FLIP (1<<6) +#define MI_WAIT_FOR_PLANE_A_FLIP (1<<2) +#define MI_WAIT_FOR_PLANE_A_SCANLINES (1<<1) +#define MI_FLUSH MI_INSTR(0x04, 0) +#define MI_READ_FLUSH (1 << 0) +#define MI_EXE_FLUSH (1 << 1) +#define MI_NO_WRITE_FLUSH (1 << 2) +#define MI_SCENE_COUNT (1 << 3) /* just increment scene count */ +#define MI_END_SCENE (1 << 4) /* flush binner and incr scene count */ +#define MI_INVALIDATE_ISP (1 << 5) /* invalidate indirect state pointers */ +#define MI_REPORT_HEAD MI_INSTR(0x07, 0) +#define MI_ARB_ON_OFF MI_INSTR(0x08, 0) +#define MI_ARB_ENABLE (1<<0) +#define MI_ARB_DISABLE (0<<0) +#define MI_BATCH_BUFFER_END MI_INSTR(0x0a, 0) +#define MI_SUSPEND_FLUSH MI_INSTR(0x0b, 0) +#define MI_SUSPEND_FLUSH_EN (1<<0) +#define MI_SET_APPID MI_INSTR(0x0e, 0) +#define MI_OVERLAY_FLIP MI_INSTR(0x11, 0) +#define MI_OVERLAY_CONTINUE (0x0<<21) +#define MI_OVERLAY_ON (0x1<<21) +#define MI_OVERLAY_OFF (0x2<<21) +#define MI_LOAD_SCAN_LINES_INCL MI_INSTR(0x12, 0) +#define MI_DISPLAY_FLIP MI_INSTR(0x14, 2) +#define MI_DISPLAY_FLIP_I915 MI_INSTR(0x14, 1) +#define MI_DISPLAY_FLIP_PLANE(n) ((n) << 20) +/* IVB has funny definitions for which plane to flip. */ +#define MI_DISPLAY_FLIP_IVB_PLANE_A (0 << 19) +#define MI_DISPLAY_FLIP_IVB_PLANE_B (1 << 19) +#define MI_DISPLAY_FLIP_IVB_SPRITE_A (2 << 19) +#define MI_DISPLAY_FLIP_IVB_SPRITE_B (3 << 19) +#define MI_DISPLAY_FLIP_IVB_PLANE_C (4 << 19) +#define MI_DISPLAY_FLIP_IVB_SPRITE_C (5 << 19) +/* SKL ones */ +#define MI_DISPLAY_FLIP_SKL_PLANE_1_A (0 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_1_B (1 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_1_C (2 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_2_A (4 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_2_B (5 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_2_C (6 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_3_A (7 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_3_B (8 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_3_C (9 << 8) +#define MI_SEMAPHORE_MBOX MI_INSTR(0x16, 1) /* gen6, gen7 */ +#define MI_SEMAPHORE_GLOBAL_GTT (1<<22) +#define MI_SEMAPHORE_UPDATE (1<<21) +#define MI_SEMAPHORE_COMPARE (1<<20) +#define MI_SEMAPHORE_REGISTER (1<<18) +#define MI_SEMAPHORE_SYNC_VR (0<<16) /* RCS wait for VCS (RVSYNC) */ +#define MI_SEMAPHORE_SYNC_VER (1<<16) /* RCS wait for VECS (RVESYNC) */ +#define MI_SEMAPHORE_SYNC_BR (2<<16) /* RCS wait for BCS (RBSYNC) */ +#define MI_SEMAPHORE_SYNC_BV (0<<16) /* VCS wait for BCS (VBSYNC) */ +#define MI_SEMAPHORE_SYNC_VEV (1<<16) /* VCS wait for VECS (VVESYNC) */ +#define MI_SEMAPHORE_SYNC_RV (2<<16) /* VCS wait for RCS (VRSYNC) */ +#define MI_SEMAPHORE_SYNC_RB (0<<16) /* BCS wait for RCS (BRSYNC) */ +#define MI_SEMAPHORE_SYNC_VEB (1<<16) /* BCS wait for VECS (BVESYNC) */ +#define MI_SEMAPHORE_SYNC_VB (2<<16) /* BCS wait for VCS (BVSYNC) */ +#define MI_SEMAPHORE_SYNC_BVE (0<<16) /* VECS wait for BCS (VEBSYNC) */ +#define MI_SEMAPHORE_SYNC_VVE (1<<16) /* VECS wait for VCS (VEVSYNC) */ +#define MI_SEMAPHORE_SYNC_RVE (2<<16) /* VECS wait for RCS (VERSYNC) */ +#define MI_SEMAPHORE_SYNC_INVALID (3<<16) +#define MI_SEMAPHORE_SYNC_MASK (3<<16) +#define MI_SET_CONTEXT MI_INSTR(0x18, 0) +#define MI_MM_SPACE_GTT (1<<8) +#define MI_MM_SPACE_PHYSICAL (0<<8) +#define MI_SAVE_EXT_STATE_EN (1<<3) +#define MI_RESTORE_EXT_STATE_EN (1<<2) +#define MI_FORCE_RESTORE (1<<1) +#define MI_RESTORE_INHIBIT (1<<0) +#define HSW_MI_RS_SAVE_STATE_EN (1<<3) +#define HSW_MI_RS_RESTORE_STATE_EN (1<<2) +#define MI_SEMAPHORE_SIGNAL MI_INSTR(0x1b, 0) /* GEN8+ */ +#define MI_SEMAPHORE_TARGET(engine) ((engine)<<15) +#define MI_SEMAPHORE_WAIT MI_INSTR(0x1c, 2) /* GEN8+ */ +#define MI_SEMAPHORE_POLL (1<<15) +#define MI_SEMAPHORE_SAD_GTE_SDD (1<<12) +#define MI_STORE_DWORD_IMM MI_INSTR(0x20, 1) +#define MI_STORE_DWORD_IMM_GEN4 MI_INSTR(0x20, 2) +#define MI_MEM_VIRTUAL (1 << 22) /* 945,g33,965 */ +#define MI_USE_GGTT (1 << 22) /* g4x+ */ +#define MI_STORE_DWORD_INDEX MI_INSTR(0x21, 1) +#define MI_STORE_DWORD_INDEX_SHIFT 2 +/* + * Official intel docs are somewhat sloppy concerning MI_LOAD_REGISTER_IMM: + * - Always issue a MI_NOOP _before_ the MI_LOAD_REGISTER_IMM - otherwise hw + * simply ignores the register load under certain conditions. + * - One can actually load arbitrary many arbitrary registers: Simply issue x + * address/value pairs. Don't overdue it, though, x <= 2^4 must hold! + */ +#define MI_LOAD_REGISTER_IMM(x) MI_INSTR(0x22, 2*(x)-1) +#define MI_LRI_FORCE_POSTED (1<<12) +#define MI_STORE_REGISTER_MEM MI_INSTR(0x24, 1) +#define MI_STORE_REGISTER_MEM_GEN8 MI_INSTR(0x24, 2) +#define MI_SRM_LRM_GLOBAL_GTT (1<<22) +#define MI_FLUSH_DW MI_INSTR(0x26, 1) /* for GEN6 */ +#define MI_FLUSH_DW_STORE_INDEX (1<<21) +#define MI_INVALIDATE_TLB (1<<18) +#define MI_FLUSH_DW_OP_STOREDW (1<<14) +#define MI_FLUSH_DW_OP_MASK (3<<14) +#define MI_FLUSH_DW_NOTIFY (1<<8) +#define MI_INVALIDATE_BSD (1<<7) +#define MI_FLUSH_DW_USE_GTT (1<<2) +#define MI_FLUSH_DW_USE_PPGTT (0<<2) +#define MI_LOAD_REGISTER_MEM MI_INSTR(0x29, 1) +#define MI_LOAD_REGISTER_MEM_GEN8 MI_INSTR(0x29, 2) +#define MI_BATCH_BUFFER MI_INSTR(0x30, 1) +#define MI_BATCH_NON_SECURE (1) +/* for snb/ivb/vlv this also means "batch in ppgtt" when ppgtt is enabled. */ +#define MI_BATCH_NON_SECURE_I965 (1<<8) +#define MI_BATCH_PPGTT_HSW (1<<8) +#define MI_BATCH_NON_SECURE_HSW (1<<13) +#define MI_BATCH_BUFFER_START MI_INSTR(0x31, 0) +#define MI_BATCH_GTT (2<<6) /* aliased with (1<<7) on gen4 */ +#define MI_BATCH_BUFFER_START_GEN8 MI_INSTR(0x31, 1) +#define MI_BATCH_RESOURCE_STREAMER (1<<10) + +/* + * 3D instructions used by the kernel + */ +#define GFX_INSTR(opcode, flags) ((0x3 << 29) | ((opcode) << 24) | (flags)) + +#define GEN9_MEDIA_POOL_STATE ((0x3 << 29) | (0x2 << 27) | (0x5 << 16) | 4) +#define GEN9_MEDIA_POOL_ENABLE (1 << 31) +#define GFX_OP_RASTER_RULES ((0x3<<29)|(0x7<<24)) +#define GFX_OP_SCISSOR ((0x3<<29)|(0x1c<<24)|(0x10<<19)) +#define SC_UPDATE_SCISSOR (0x1<<1) +#define SC_ENABLE_MASK (0x1<<0) +#define SC_ENABLE (0x1<<0) +#define GFX_OP_LOAD_INDIRECT ((0x3<<29)|(0x1d<<24)|(0x7<<16)) +#define GFX_OP_SCISSOR_INFO ((0x3<<29)|(0x1d<<24)|(0x81<<16)|(0x1)) +#define SCI_YMIN_MASK (0xffff<<16) +#define SCI_XMIN_MASK (0xffff<<0) +#define SCI_YMAX_MASK (0xffff<<16) +#define SCI_XMAX_MASK (0xffff<<0) +#define GFX_OP_SCISSOR_ENABLE ((0x3<<29)|(0x1c<<24)|(0x10<<19)) +#define GFX_OP_SCISSOR_RECT ((0x3<<29)|(0x1d<<24)|(0x81<<16)|1) +#define GFX_OP_COLOR_FACTOR ((0x3<<29)|(0x1d<<24)|(0x1<<16)|0x0) +#define GFX_OP_STIPPLE ((0x3<<29)|(0x1d<<24)|(0x83<<16)) +#define GFX_OP_MAP_INFO ((0x3<<29)|(0x1d<<24)|0x4) +#define GFX_OP_DESTBUFFER_VARS ((0x3<<29)|(0x1d<<24)|(0x85<<16)|0x0) +#define GFX_OP_DESTBUFFER_INFO ((0x3<<29)|(0x1d<<24)|(0x8e<<16)|1) +#define GFX_OP_DRAWRECT_INFO ((0x3<<29)|(0x1d<<24)|(0x80<<16)|(0x3)) +#define GFX_OP_DRAWRECT_INFO_I965 ((0x7900<<16)|0x2) + +#define COLOR_BLT_CMD (2<<29 | 0x40<<22 | (5-2)) +#define SRC_COPY_BLT_CMD ((2<<29)|(0x43<<22)|4) +#define XY_SRC_COPY_BLT_CMD ((2<<29)|(0x53<<22)|6) +#define XY_MONO_SRC_COPY_IMM_BLT ((2<<29)|(0x71<<22)|5) +#define BLT_WRITE_A (2<<20) +#define BLT_WRITE_RGB (1<<20) +#define BLT_WRITE_RGBA (BLT_WRITE_RGB | BLT_WRITE_A) +#define BLT_DEPTH_8 (0<<24) +#define BLT_DEPTH_16_565 (1<<24) +#define BLT_DEPTH_16_1555 (2<<24) +#define BLT_DEPTH_32 (3<<24) +#define BLT_ROP_SRC_COPY (0xcc<<16) +#define BLT_ROP_COLOR_COPY (0xf0<<16) +#define XY_SRC_COPY_BLT_SRC_TILED (1<<15) /* 965+ only */ +#define XY_SRC_COPY_BLT_DST_TILED (1<<11) /* 965+ only */ +#define CMD_OP_DISPLAYBUFFER_INFO ((0x0<<29)|(0x14<<23)|2) +#define ASYNC_FLIP (1<<22) +#define DISPLAY_PLANE_A (0<<20) +#define DISPLAY_PLANE_B (1<<20) +#define GFX_OP_PIPE_CONTROL(len) ((0x3<<29)|(0x3<<27)|(0x2<<24)|((len)-2)) +#define PIPE_CONTROL_FLUSH_L3 (1<<27) +#define PIPE_CONTROL_GLOBAL_GTT_IVB (1<<24) /* gen7+ */ +#define PIPE_CONTROL_MMIO_WRITE (1<<23) +#define PIPE_CONTROL_STORE_DATA_INDEX (1<<21) +#define PIPE_CONTROL_CS_STALL (1<<20) +#define PIPE_CONTROL_TLB_INVALIDATE (1<<18) +#define PIPE_CONTROL_MEDIA_STATE_CLEAR (1<<16) +#define PIPE_CONTROL_QW_WRITE (1<<14) +#define PIPE_CONTROL_POST_SYNC_OP_MASK (3<<14) +#define PIPE_CONTROL_DEPTH_STALL (1<<13) +#define PIPE_CONTROL_WRITE_FLUSH (1<<12) +#define PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH (1<<12) /* gen6+ */ +#define PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE (1<<11) /* MBZ on ILK */ +#define PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE (1<<10) /* GM45+ only */ +#define PIPE_CONTROL_INDIRECT_STATE_DISABLE (1<<9) +#define PIPE_CONTROL_NOTIFY (1<<8) +#define PIPE_CONTROL_FLUSH_ENABLE (1<<7) /* gen7+ */ +#define PIPE_CONTROL_DC_FLUSH_ENABLE (1<<5) +#define PIPE_CONTROL_VF_CACHE_INVALIDATE (1<<4) +#define PIPE_CONTROL_CONST_CACHE_INVALIDATE (1<<3) +#define PIPE_CONTROL_STATE_CACHE_INVALIDATE (1<<2) +#define PIPE_CONTROL_STALL_AT_SCOREBOARD (1<<1) +#define PIPE_CONTROL_DEPTH_CACHE_FLUSH (1<<0) +#define PIPE_CONTROL_GLOBAL_GTT (1<<2) /* in addr dword */ + +/* + * Commands used only by the command parser + */ +#define MI_SET_PREDICATE MI_INSTR(0x01, 0) +#define MI_ARB_CHECK MI_INSTR(0x05, 0) +#define MI_RS_CONTROL MI_INSTR(0x06, 0) +#define MI_URB_ATOMIC_ALLOC MI_INSTR(0x09, 0) +#define MI_PREDICATE MI_INSTR(0x0C, 0) +#define MI_RS_CONTEXT MI_INSTR(0x0F, 0) +#define MI_TOPOLOGY_FILTER MI_INSTR(0x0D, 0) +#define MI_LOAD_SCAN_LINES_EXCL MI_INSTR(0x13, 0) +#define MI_URB_CLEAR MI_INSTR(0x19, 0) +#define MI_UPDATE_GTT MI_INSTR(0x23, 0) +#define MI_CLFLUSH MI_INSTR(0x27, 0) +#define MI_REPORT_PERF_COUNT MI_INSTR(0x28, 0) +#define MI_REPORT_PERF_COUNT_GGTT (1<<0) +#define MI_LOAD_REGISTER_REG MI_INSTR(0x2A, 0) +#define MI_RS_STORE_DATA_IMM MI_INSTR(0x2B, 0) +#define MI_LOAD_URB_MEM MI_INSTR(0x2C, 0) +#define MI_STORE_URB_MEM MI_INSTR(0x2D, 0) +#define MI_CONDITIONAL_BATCH_BUFFER_END MI_INSTR(0x36, 0) + +#define PIPELINE_SELECT ((0x3<<29)|(0x1<<27)|(0x1<<24)|(0x4<<16)) +#define GFX_OP_3DSTATE_VF_STATISTICS ((0x3<<29)|(0x1<<27)|(0x0<<24)|(0xB<<16)) +#define MEDIA_VFE_STATE ((0x3<<29)|(0x2<<27)|(0x0<<24)|(0x0<<16)) +#define MEDIA_VFE_STATE_MMIO_ACCESS_MASK (0x18) +#define GPGPU_OBJECT ((0x3<<29)|(0x2<<27)|(0x1<<24)|(0x4<<16)) +#define GPGPU_WALKER ((0x3<<29)|(0x2<<27)|(0x1<<24)|(0x5<<16)) +#define GFX_OP_3DSTATE_DX9_CONSTANTF_VS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x39<<16)) +#define GFX_OP_3DSTATE_DX9_CONSTANTF_PS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x3A<<16)) +#define GFX_OP_3DSTATE_SO_DECL_LIST \ + ((0x3<<29)|(0x3<<27)|(0x1<<24)|(0x17<<16)) + +#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_VS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x43<<16)) +#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_GS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x44<<16)) +#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_HS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x45<<16)) +#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_DS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x46<<16)) +#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_PS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x47<<16)) + +#define MFX_WAIT ((0x3<<29)|(0x1<<27)|(0x0<<16)) + +#define COLOR_BLT ((0x2<<29)|(0x40<<22)) +#define SRC_COPY_BLT ((0x2<<29)|(0x43<<22)) + +#endif /* _INTEL_GPU_COMMANDS_H_ */ diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 80fae80..1f50727 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -7,10 +7,11 @@ #include "i915_gem_batch_pool.h" #include "i915_gem_timeline.h" -#include "i915_reg.h" /* FIXME split out i915_gpu_commands.h */ +#include "i915_reg.h" #include "i915_pmu.h" #include "i915_request.h" #include "i915_selftest.h" +#include "intel_gpu_commands.h" struct drm_printer; -- cgit v1.1 From 56b9a8b083870162310fe37d4b1b5597eb983bae Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Wed, 14 Mar 2018 14:45:39 +0000 Subject: drm/i915/guc: Update syntax of GuC log functions MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We moved GuC log related data and code to separate files and definition but we didn't change functions syntax to follow object-verb pattern. Let's fix that before we continue with next round of code refactoring. v2: rebased Signed-off-by: Michal Wajdeczko Cc: Michal Winiarski Cc: Chris Wilson Reviewed-by: Michał Winiarski Link: https://patchwork.freedesktop.org/patch/msgid/20180314144539.11152-1-michal.wajdeczko@intel.com [ickle: checkpatch booleans] Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/i915_debugfs.c | 4 +- drivers/gpu/drm/i915/intel_guc.c | 8 +- drivers/gpu/drm/i915/intel_guc_log.c | 203 +++++++++++++++++++---------------- drivers/gpu/drm/i915/intel_guc_log.h | 18 ++-- drivers/gpu/drm/i915/intel_uc.c | 4 +- 5 files changed, 126 insertions(+), 111 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 972014b..298a3aa 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -2502,7 +2502,7 @@ static int i915_guc_log_control_get(void *data, u64 *val) if (!USES_GUC(dev_priv)) return -ENODEV; - *val = intel_guc_log_control_get(&dev_priv->guc); + *val = intel_guc_log_control_get(&dev_priv->guc.log); return 0; } @@ -2514,7 +2514,7 @@ static int i915_guc_log_control_set(void *data, u64 val) if (!USES_GUC(dev_priv)) return -ENODEV; - return intel_guc_log_control_set(&dev_priv->guc, val); + return intel_guc_log_control_set(&dev_priv->guc.log, val); } DEFINE_SIMPLE_ATTRIBUTE(i915_guc_log_control_fops, diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index 3eb516e..e70bf65 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -64,7 +64,7 @@ void intel_guc_init_early(struct intel_guc *guc) { intel_guc_fw_init_early(guc); intel_guc_ct_init_early(&guc->ct); - intel_guc_log_init_early(guc); + intel_guc_log_init_early(&guc->log); mutex_init(&guc->send_mutex); guc->send = intel_guc_send_nop; @@ -169,7 +169,7 @@ int intel_guc_init(struct intel_guc *guc) return ret; GEM_BUG_ON(!guc->shared_data); - ret = intel_guc_log_create(guc); + ret = intel_guc_log_create(&guc->log); if (ret) goto err_shared; @@ -184,7 +184,7 @@ int intel_guc_init(struct intel_guc *guc) return 0; err_log: - intel_guc_log_destroy(guc); + intel_guc_log_destroy(&guc->log); err_shared: guc_shared_data_destroy(guc); return ret; @@ -196,7 +196,7 @@ void intel_guc_fini(struct intel_guc *guc) i915_ggtt_disable_guc(dev_priv); intel_guc_ads_destroy(guc); - intel_guc_log_destroy(guc); + intel_guc_log_destroy(&guc->log); guc_shared_data_destroy(guc); } diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index b9c7bd7..bfb9a68 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -22,13 +22,10 @@ * */ -#include -#include - #include "intel_guc_log.h" #include "i915_drv.h" -static void guc_log_capture_logs(struct intel_guc *guc); +static void guc_log_capture_logs(struct intel_guc_log *log); /** * DOC: GuC firmware log @@ -74,6 +71,11 @@ static int guc_log_control(struct intel_guc *guc, bool enable, u32 verbosity) return intel_guc_send(guc, action, ARRAY_SIZE(action)); } +static inline struct intel_guc *log_to_guc(struct intel_guc_log *log) +{ + return container_of(log, struct intel_guc, log); +} + /* * Sub buffer switch callback. Called whenever relay has to switch to a new * sub buffer, relay stays on the same sub buffer if 0 is returned. @@ -149,8 +151,9 @@ static struct rchan_callbacks relay_callbacks = { .remove_buf_file = remove_buf_file_callback, }; -static int guc_log_relay_file_create(struct intel_guc *guc) +static int guc_log_relay_file_create(struct intel_guc_log *log) { + struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); struct dentry *log_dir; int ret; @@ -158,7 +161,7 @@ static int guc_log_relay_file_create(struct intel_guc *guc) if (!i915_modparams.guc_log_level) return 0; - mutex_lock(&guc->log.runtime.relay_lock); + mutex_lock(&log->runtime.relay_lock); /* For now create the log file in /sys/kernel/debug/dri/0 dir */ log_dir = dev_priv->drm.primary->debugfs_root; @@ -181,7 +184,8 @@ static int guc_log_relay_file_create(struct intel_guc *guc) goto out_unlock; } - ret = relay_late_setup_files(guc->log.runtime.relay_chan, "guc_log", log_dir); + ret = relay_late_setup_files(log->runtime.relay_chan, "guc_log", + log_dir); if (ret < 0 && ret != -EEXIST) { DRM_ERROR("Couldn't associate relay chan with file %d\n", ret); goto out_unlock; @@ -190,18 +194,18 @@ static int guc_log_relay_file_create(struct intel_guc *guc) ret = 0; out_unlock: - mutex_unlock(&guc->log.runtime.relay_lock); + mutex_unlock(&log->runtime.relay_lock); return ret; } -static bool guc_log_has_relay(struct intel_guc *guc) +static bool guc_log_has_relay(struct intel_guc_log *log) { - lockdep_assert_held(&guc->log.runtime.relay_lock); + lockdep_assert_held(&log->runtime.relay_lock); - return guc->log.runtime.relay_chan != NULL; + return log->runtime.relay_chan; } -static void guc_move_to_next_buf(struct intel_guc *guc) +static void guc_move_to_next_buf(struct intel_guc_log *log) { /* * Make sure the updates made in the sub buffer are visible when @@ -209,19 +213,19 @@ static void guc_move_to_next_buf(struct intel_guc *guc) */ smp_wmb(); - if (!guc_log_has_relay(guc)) + if (!guc_log_has_relay(log)) return; /* All data has been written, so now move the offset of sub buffer. */ - relay_reserve(guc->log.runtime.relay_chan, guc->log.vma->obj->base.size); + relay_reserve(log->runtime.relay_chan, log->vma->obj->base.size); /* Switch to the next sub buffer */ - relay_flush(guc->log.runtime.relay_chan); + relay_flush(log->runtime.relay_chan); } -static void *guc_get_write_buffer(struct intel_guc *guc) +static void *guc_get_write_buffer(struct intel_guc_log *log) { - if (!guc_log_has_relay(guc)) + if (!guc_log_has_relay(log)) return NULL; /* @@ -233,25 +237,25 @@ static void *guc_get_write_buffer(struct intel_guc *guc) * done without using relay_reserve() along with relay_write(). So its * better to use relay_reserve() alone. */ - return relay_reserve(guc->log.runtime.relay_chan, 0); + return relay_reserve(log->runtime.relay_chan, 0); } -static bool guc_check_log_buf_overflow(struct intel_guc *guc, +static bool guc_check_log_buf_overflow(struct intel_guc_log *log, enum guc_log_buffer_type type, unsigned int full_cnt) { - unsigned int prev_full_cnt = guc->log.prev_overflow_count[type]; + unsigned int prev_full_cnt = log->prev_overflow_count[type]; bool overflow = false; if (full_cnt != prev_full_cnt) { overflow = true; - guc->log.prev_overflow_count[type] = full_cnt; - guc->log.total_overflow_count[type] += full_cnt - prev_full_cnt; + log->prev_overflow_count[type] = full_cnt; + log->total_overflow_count[type] += full_cnt - prev_full_cnt; if (full_cnt < prev_full_cnt) { /* buffer_full_cnt is a 4 bit counter */ - guc->log.total_overflow_count[type] += 16; + log->total_overflow_count[type] += 16; } DRM_ERROR_RATELIMITED("GuC log buffer overflow\n"); } @@ -275,7 +279,7 @@ static unsigned int guc_get_log_buffer_size(enum guc_log_buffer_type type) return 0; } -static void guc_read_update_log_buffer(struct intel_guc *guc) +static void guc_read_update_log_buffer(struct intel_guc_log *log) { unsigned int buffer_size, read_offset, write_offset, bytes_to_copy, full_cnt; struct guc_log_buffer_state *log_buf_state, *log_buf_snapshot_state; @@ -284,16 +288,16 @@ static void guc_read_update_log_buffer(struct intel_guc *guc) void *src_data, *dst_data; bool new_overflow; - if (WARN_ON(!guc->log.runtime.buf_addr)) + if (WARN_ON(!log->runtime.buf_addr)) return; /* Get the pointer to shared GuC log buffer */ - log_buf_state = src_data = guc->log.runtime.buf_addr; + log_buf_state = src_data = log->runtime.buf_addr; - mutex_lock(&guc->log.runtime.relay_lock); + mutex_lock(&log->runtime.relay_lock); /* Get the pointer to local buffer to store the logs */ - log_buf_snapshot_state = dst_data = guc_get_write_buffer(guc); + log_buf_snapshot_state = dst_data = guc_get_write_buffer(log); if (unlikely(!log_buf_snapshot_state)) { /* @@ -301,8 +305,8 @@ static void guc_read_update_log_buffer(struct intel_guc *guc) * getting consumed by User at a slow rate. */ DRM_ERROR_RATELIMITED("no sub-buffer to capture logs\n"); - guc->log.capture_miss_count++; - mutex_unlock(&guc->log.runtime.relay_lock); + log->capture_miss_count++; + mutex_unlock(&log->runtime.relay_lock); return; } @@ -325,8 +329,8 @@ static void guc_read_update_log_buffer(struct intel_guc *guc) full_cnt = log_buf_state_local.buffer_full_cnt; /* Bookkeeping stuff */ - guc->log.flush_count[type] += log_buf_state_local.flush_to_file; - new_overflow = guc_check_log_buf_overflow(guc, type, full_cnt); + log->flush_count[type] += log_buf_state_local.flush_to_file; + new_overflow = guc_check_log_buf_overflow(log, type, full_cnt); /* Update the state of shared log buffer */ log_buf_state->read_ptr = write_offset; @@ -373,38 +377,39 @@ static void guc_read_update_log_buffer(struct intel_guc *guc) dst_data += buffer_size; } - guc_move_to_next_buf(guc); + guc_move_to_next_buf(log); - mutex_unlock(&guc->log.runtime.relay_lock); + mutex_unlock(&log->runtime.relay_lock); } static void capture_logs_work(struct work_struct *work) { - struct intel_guc *guc = - container_of(work, struct intel_guc, log.runtime.flush_work); + struct intel_guc_log *log = + container_of(work, struct intel_guc_log, runtime.flush_work); - guc_log_capture_logs(guc); + guc_log_capture_logs(log); } -static bool guc_log_has_runtime(struct intel_guc *guc) +static bool guc_log_has_runtime(struct intel_guc_log *log) { - return guc->log.runtime.buf_addr != NULL; + return log->runtime.buf_addr; } -static int guc_log_runtime_create(struct intel_guc *guc) +static int guc_log_runtime_create(struct intel_guc_log *log) { + struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); void *vaddr; int ret; lockdep_assert_held(&dev_priv->drm.struct_mutex); - if (!guc->log.vma) + if (!log->vma) return -ENODEV; - GEM_BUG_ON(guc_log_has_runtime(guc)); + GEM_BUG_ON(guc_log_has_runtime(log)); - ret = i915_gem_object_set_to_wc_domain(guc->log.vma->obj, true); + ret = i915_gem_object_set_to_wc_domain(log->vma->obj, true); if (ret) return ret; @@ -413,38 +418,39 @@ static int guc_log_runtime_create(struct intel_guc *guc) * buffer pages, so that we can directly get the data * (up-to-date) from memory. */ - vaddr = i915_gem_object_pin_map(guc->log.vma->obj, I915_MAP_WC); + vaddr = i915_gem_object_pin_map(log->vma->obj, I915_MAP_WC); if (IS_ERR(vaddr)) { DRM_ERROR("Couldn't map log buffer pages %d\n", ret); return PTR_ERR(vaddr); } - guc->log.runtime.buf_addr = vaddr; + log->runtime.buf_addr = vaddr; return 0; } -static void guc_log_runtime_destroy(struct intel_guc *guc) +static void guc_log_runtime_destroy(struct intel_guc_log *log) { /* * It's possible that the runtime stuff was never allocated because * GuC log was disabled at the boot time. */ - if (!guc_log_has_runtime(guc)) + if (!guc_log_has_runtime(log)) return; - i915_gem_object_unpin_map(guc->log.vma->obj); - guc->log.runtime.buf_addr = NULL; + i915_gem_object_unpin_map(log->vma->obj); + log->runtime.buf_addr = NULL; } -void intel_guc_log_init_early(struct intel_guc *guc) +void intel_guc_log_init_early(struct intel_guc_log *log) { - mutex_init(&guc->log.runtime.relay_lock); - INIT_WORK(&guc->log.runtime.flush_work, capture_logs_work); + mutex_init(&log->runtime.relay_lock); + INIT_WORK(&log->runtime.flush_work, capture_logs_work); } -static int guc_log_relay_create(struct intel_guc *guc) +static int guc_log_relay_create(struct intel_guc_log *log) { + struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); struct rchan *guc_log_relay_chan; size_t n_subbufs, subbuf_size; @@ -453,9 +459,9 @@ static int guc_log_relay_create(struct intel_guc *guc) if (!i915_modparams.guc_log_level) return 0; - mutex_lock(&guc->log.runtime.relay_lock); + mutex_lock(&log->runtime.relay_lock); - GEM_BUG_ON(guc_log_has_relay(guc)); + GEM_BUG_ON(guc_log_has_relay(log)); /* Keep the size of sub buffers same as shared log buffer */ subbuf_size = GUC_LOG_SIZE; @@ -483,42 +489,43 @@ static int guc_log_relay_create(struct intel_guc *guc) } GEM_BUG_ON(guc_log_relay_chan->subbuf_size < subbuf_size); - guc->log.runtime.relay_chan = guc_log_relay_chan; + log->runtime.relay_chan = guc_log_relay_chan; - mutex_unlock(&guc->log.runtime.relay_lock); + mutex_unlock(&log->runtime.relay_lock); return 0; err: - mutex_unlock(&guc->log.runtime.relay_lock); + mutex_unlock(&log->runtime.relay_lock); /* logging will be off */ i915_modparams.guc_log_level = 0; return ret; } -static void guc_log_relay_destroy(struct intel_guc *guc) +static void guc_log_relay_destroy(struct intel_guc_log *log) { - mutex_lock(&guc->log.runtime.relay_lock); + mutex_lock(&log->runtime.relay_lock); /* * It's possible that the relay was never allocated because * GuC log was disabled at the boot time. */ - if (!guc_log_has_relay(guc)) + if (!guc_log_has_relay(log)) goto out_unlock; - relay_close(guc->log.runtime.relay_chan); - guc->log.runtime.relay_chan = NULL; + relay_close(log->runtime.relay_chan); + log->runtime.relay_chan = NULL; out_unlock: - mutex_unlock(&guc->log.runtime.relay_lock); + mutex_unlock(&log->runtime.relay_lock); } -static void guc_log_capture_logs(struct intel_guc *guc) +static void guc_log_capture_logs(struct intel_guc_log *log) { + struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); - guc_read_update_log_buffer(guc); + guc_read_update_log_buffer(log); /* * Generally device is expected to be active only at this @@ -529,15 +536,16 @@ static void guc_log_capture_logs(struct intel_guc *guc) intel_runtime_pm_put(dev_priv); } -static void guc_flush_logs(struct intel_guc *guc) +static void guc_flush_logs(struct intel_guc_log *log) { + struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); /* * Before initiating the forceful flush, wait for any pending/ongoing * flush to complete otherwise forceful flush may not actually happen. */ - flush_work(&guc->log.runtime.flush_work); + flush_work(&log->runtime.flush_work); /* Ask GuC to update the log buffer state */ intel_runtime_pm_get(dev_priv); @@ -545,17 +553,18 @@ static void guc_flush_logs(struct intel_guc *guc) intel_runtime_pm_put(dev_priv); /* GuC would have updated log buffer by now, so capture it */ - guc_log_capture_logs(guc); + guc_log_capture_logs(log); } -int intel_guc_log_create(struct intel_guc *guc) +int intel_guc_log_create(struct intel_guc_log *log) { + struct intel_guc *guc = log_to_guc(log); struct i915_vma *vma; unsigned long offset; u32 flags; int ret; - GEM_BUG_ON(guc->log.vma); + GEM_BUG_ON(log->vma); /* * We require SSE 4.1 for fast reads from the GuC log buffer and @@ -573,7 +582,7 @@ int intel_guc_log_create(struct intel_guc *guc) goto err; } - guc->log.vma = vma; + log->vma = vma; /* each allocated unit is a page */ flags = GUC_LOG_VALID | GUC_LOG_NOTIFY_ON_HALF_FULL | @@ -582,7 +591,7 @@ int intel_guc_log_create(struct intel_guc *guc) (GUC_LOG_CRASH_PAGES << GUC_LOG_CRASH_SHIFT); offset = intel_guc_ggtt_offset(guc, vma) >> PAGE_SHIFT; - guc->log.flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags; + log->flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags; return 0; @@ -592,15 +601,15 @@ err: return ret; } -void intel_guc_log_destroy(struct intel_guc *guc) +void intel_guc_log_destroy(struct intel_guc_log *log) { - guc_log_runtime_destroy(guc); - i915_vma_unpin_and_release(&guc->log.vma); + guc_log_runtime_destroy(log); + i915_vma_unpin_and_release(&log->vma); } -int intel_guc_log_control_get(struct intel_guc *guc) +int intel_guc_log_control_get(struct intel_guc_log *log) { - GEM_BUG_ON(!guc->log.vma); + GEM_BUG_ON(!log->vma); GEM_BUG_ON(i915_modparams.guc_log_level < 0); return i915_modparams.guc_log_level; @@ -613,14 +622,15 @@ int intel_guc_log_control_get(struct intel_guc *guc) LOG_LEVEL_TO_ENABLED(_x) ? _x - 1 : 0; \ }) #define VERBOSITY_TO_LOG_LEVEL(x) ((x) + 1) -int intel_guc_log_control_set(struct intel_guc *guc, u64 val) +int intel_guc_log_control_set(struct intel_guc_log *log, u64 val) { + struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); bool enabled = LOG_LEVEL_TO_ENABLED(val); int ret; BUILD_BUG_ON(GUC_LOG_VERBOSITY_MIN != 0); - GEM_BUG_ON(!guc->log.vma); + GEM_BUG_ON(!log->vma); GEM_BUG_ON(i915_modparams.guc_log_level < 0); /* @@ -650,15 +660,15 @@ int intel_guc_log_control_set(struct intel_guc *guc, u64 val) mutex_unlock(&dev_priv->drm.struct_mutex); - if (enabled && !guc_log_has_runtime(guc)) { - ret = intel_guc_log_register(guc); + if (enabled && !guc_log_has_runtime(log)) { + ret = intel_guc_log_register(log); if (ret) { /* logging will remain off */ i915_modparams.guc_log_level = 0; goto out; } - } else if (!enabled && guc_log_has_runtime(guc)) { - intel_guc_log_unregister(guc); + } else if (!enabled && guc_log_has_runtime(log)) { + intel_guc_log_unregister(log); } return 0; @@ -669,30 +679,31 @@ out: return ret; } -int intel_guc_log_register(struct intel_guc *guc) +int intel_guc_log_register(struct intel_guc_log *log) { + struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *i915 = guc_to_i915(guc); int ret; - GEM_BUG_ON(guc_log_has_runtime(guc)); + GEM_BUG_ON(guc_log_has_runtime(log)); /* * If log was disabled at boot time, then setup needed to handle * log buffer flush interrupts would not have been done yet, so * do that now. */ - ret = guc_log_relay_create(guc); + ret = guc_log_relay_create(log); if (ret) goto err; mutex_lock(&i915->drm.struct_mutex); - ret = guc_log_runtime_create(guc); + ret = guc_log_runtime_create(log); mutex_unlock(&i915->drm.struct_mutex); if (ret) goto err_relay; - ret = guc_log_relay_file_create(guc); + ret = guc_log_relay_file_create(log); if (ret) goto err_runtime; @@ -707,16 +718,17 @@ int intel_guc_log_register(struct intel_guc *guc) err_runtime: mutex_lock(&i915->drm.struct_mutex); - guc_log_runtime_destroy(guc); + guc_log_runtime_destroy(log); mutex_unlock(&i915->drm.struct_mutex); err_relay: - guc_log_relay_destroy(guc); + guc_log_relay_destroy(log); err: return ret; } -void intel_guc_log_unregister(struct intel_guc *guc) +void intel_guc_log_unregister(struct intel_guc_log *log) { + struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *i915 = guc_to_i915(guc); /* @@ -725,16 +737,17 @@ void intel_guc_log_unregister(struct intel_guc *guc) * which is yet to be captured. So request GuC to update the log * buffer state and then collect the left over logs. */ - guc_flush_logs(guc); + guc_flush_logs(log); mutex_lock(&i915->drm.struct_mutex); + /* GuC logging is currently the only user of Guc2Host interrupts */ intel_runtime_pm_get(i915); gen9_disable_guc_interrupts(i915); intel_runtime_pm_put(i915); - guc_log_runtime_destroy(guc); + guc_log_runtime_destroy(log); mutex_unlock(&i915->drm.struct_mutex); - guc_log_relay_destroy(guc); + guc_log_relay_destroy(log); } diff --git a/drivers/gpu/drm/i915/intel_guc_log.h b/drivers/gpu/drm/i915/intel_guc_log.h index 09dd2ef..6264bd5 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.h +++ b/drivers/gpu/drm/i915/intel_guc_log.h @@ -25,11 +25,12 @@ #ifndef _INTEL_GUC_LOG_H_ #define _INTEL_GUC_LOG_H_ +#include +#include #include #include "intel_guc_fwif.h" -struct drm_i915_private; struct intel_guc; /* @@ -59,12 +60,13 @@ struct intel_guc_log { u32 flush_count[GUC_MAX_LOG_BUFFER]; }; -int intel_guc_log_create(struct intel_guc *guc); -void intel_guc_log_destroy(struct intel_guc *guc); -void intel_guc_log_init_early(struct intel_guc *guc); -int intel_guc_log_control_get(struct intel_guc *guc); -int intel_guc_log_control_set(struct intel_guc *guc, u64 control_val); -int intel_guc_log_register(struct intel_guc *guc); -void intel_guc_log_unregister(struct intel_guc *guc); +void intel_guc_log_init_early(struct intel_guc_log *log); +int intel_guc_log_create(struct intel_guc_log *log); +int intel_guc_log_register(struct intel_guc_log *log); +void intel_guc_log_unregister(struct intel_guc_log *log); +void intel_guc_log_destroy(struct intel_guc_log *log); + +int intel_guc_log_control_get(struct intel_guc_log *log); +int intel_guc_log_control_set(struct intel_guc_log *log, u64 control); #endif diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 6316548..104c03a 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -229,7 +229,7 @@ int intel_uc_register(struct drm_i915_private *i915) return 0; if (i915_modparams.guc_log_level) - ret = intel_guc_log_register(&i915->guc); + ret = intel_guc_log_register(&i915->guc.log); return ret; } @@ -240,7 +240,7 @@ void intel_uc_unregister(struct drm_i915_private *i915) return; if (i915_modparams.guc_log_level) - intel_guc_log_unregister(&i915->guc); + intel_guc_log_unregister(&i915->guc.log); } static int guc_enable_communication(struct intel_guc *guc) -- cgit v1.1 From 7fb96dac6755d053d3a540aff55fe0064001bdf6 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Thu, 15 Mar 2018 15:28:47 +0000 Subject: drm/i915/guc: Fix build break on config without DEBUG_FS In commit 56b9a8b08387 ("drm/i915/guc: Update syntax of GuC log functions") we accidentally removed debugfs.h header where needed stub functions were defined. Reported-by: Mike Lothian Signed-off-by: Michal Wajdeczko Cc: Mike Lothian Cc: Chris Wilson Acked-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180315152848.40476-1-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc_log.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index bfb9a68..1c2127b 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -22,6 +22,8 @@ * */ +#include + #include "intel_guc_log.h" #include "i915_drv.h" -- cgit v1.1 From d9b13c4dde6cacd8f2c4385cd6d293b0ac622e0b Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 15 Mar 2018 13:14:50 +0000 Subject: drm/i915: Trace GEM steps between submit and wedging We still have an odd race with wedging/unwedging as shown by igt/gem_eio that defies expectations. Add some more trace_printks to try and visualize the flow over the precipice. Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Tvrtko Ursulin Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180315131451.4060-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_gem.c | 14 ++++++++++++++ drivers/gpu/drm/i915/i915_request.c | 23 +++++++++++++++++++++++ 2 files changed, 37 insertions(+) diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 13d4b0e..2fbd622 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -3193,6 +3193,9 @@ void i915_gem_reset_finish(struct drm_i915_private *dev_priv) static void nop_submit_request(struct i915_request *request) { + GEM_TRACE("%s fence %llx:%d -> -EIO\n", + request->engine->name, + request->fence.context, request->fence.seqno); dma_fence_set_error(&request->fence, -EIO); i915_request_submit(request); @@ -3202,6 +3205,9 @@ static void nop_complete_submit_request(struct i915_request *request) { unsigned long flags; + GEM_TRACE("%s fence %llx:%d -> -EIO\n", + request->engine->name, + request->fence.context, request->fence.seqno); dma_fence_set_error(&request->fence, -EIO); spin_lock_irqsave(&request->engine->timeline->lock, flags); @@ -3215,6 +3221,8 @@ void i915_gem_set_wedged(struct drm_i915_private *i915) struct intel_engine_cs *engine; enum intel_engine_id id; + GEM_TRACE("start\n"); + if (drm_debug & DRM_UT_DRIVER) { struct drm_printer p = drm_debug_printer(__func__); @@ -3279,6 +3287,8 @@ void i915_gem_set_wedged(struct drm_i915_private *i915) i915_gem_reset_finish_engine(engine); } + GEM_TRACE("end\n"); + wake_up_all(&i915->gpu_error.reset_queue); } @@ -3291,6 +3301,8 @@ bool i915_gem_unset_wedged(struct drm_i915_private *i915) if (!test_bit(I915_WEDGED, &i915->gpu_error.flags)) return true; + GEM_TRACE("start\n"); + /* * Before unwedging, make sure that all pending operations * are flushed and errored out - we may have requests waiting upon @@ -3341,6 +3353,8 @@ bool i915_gem_unset_wedged(struct drm_i915_private *i915) intel_engines_reset_default_submission(i915); i915_gem_contexts_lost(i915); + GEM_TRACE("end\n"); + smp_mb__before_atomic(); /* complete takeover before enabling execbuf */ clear_bit(I915_WEDGED, &i915->gpu_error.flags); diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index 1810fa1..43c7134 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -207,11 +207,16 @@ static int reset_all_global_seqno(struct drm_i915_private *i915, u32 seqno) if (ret) return ret; + GEM_BUG_ON(i915->gt.active_requests); + /* If the seqno wraps around, we need to clear the breadcrumb rbtree */ for_each_engine(engine, i915, id) { struct i915_gem_timeline *timeline; struct intel_timeline *tl = engine->timeline; + GEM_TRACE("%s seqno %d -> %d\n", + engine->name, tl->seqno, seqno); + if (!i915_seqno_passed(seqno, tl->seqno)) { /* Flush any waiters before we reuse the seqno */ intel_engine_disarm_breadcrumbs(engine); @@ -381,6 +386,11 @@ static void i915_request_retire(struct i915_request *request) struct intel_engine_cs *engine = request->engine; struct i915_gem_active *active, *next; + GEM_TRACE("%s(%d) fence %llx:%d, global_seqno %d\n", + engine->name, intel_engine_get_seqno(engine), + request->fence.context, request->fence.seqno, + request->global_seqno); + lockdep_assert_held(&request->i915->drm.struct_mutex); GEM_BUG_ON(!i915_sw_fence_signaled(&request->submit)); GEM_BUG_ON(!i915_request_completed(request)); @@ -488,6 +498,11 @@ void __i915_request_submit(struct i915_request *request) struct intel_timeline *timeline; u32 seqno; + GEM_TRACE("%s fence %llx:%d -> global_seqno %d\n", + request->engine->name, + request->fence.context, request->fence.seqno, + engine->timeline->seqno); + GEM_BUG_ON(!irqs_disabled()); lockdep_assert_held(&engine->timeline->lock); @@ -537,6 +552,11 @@ void __i915_request_unsubmit(struct i915_request *request) struct intel_engine_cs *engine = request->engine; struct intel_timeline *timeline; + GEM_TRACE("%s fence %llx:%d <- global_seqno %d\n", + request->engine->name, + request->fence.context, request->fence.seqno, + request->global_seqno); + GEM_BUG_ON(!irqs_disabled()); lockdep_assert_held(&engine->timeline->lock); @@ -996,6 +1016,9 @@ void __i915_request_add(struct i915_request *request, bool flush_caches) u32 *cs; int err; + GEM_TRACE("%s fence %llx:%d\n", + engine->name, request->fence.context, request->fence.seqno); + lockdep_assert_held(&request->i915->drm.struct_mutex); trace_i915_request_add(request); -- cgit v1.1 From ac697ae8013a7c7301174c9c3b02a92fe418b7ea Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 15 Mar 2018 15:10:15 +0000 Subject: drm/i915: Stop engines when declaring the machine wedged MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If we fail to reset the GPU, we declare the machine wedged. However, the GPU may well still be running in the background with an in-flight request. So despite our efforts in cleaning up the request queue and faking the breadcrumb in the HWSP, the GPU may eventually write the in-flght seqno there breaking all of our assumptions and throwing the driver into a deep turmoil, wedging beyond wedged. To avoid this we ideally want to reset the GPU. Since that has already failed, make sure the rings have the stop bit set instead. This is part of the normal GPU reset sequence, but that is actually disabled by igt/gem_eio to force the wedged state. If we assume the worst, we must poke at the bit again before we give up. v2: Move the intel_gpu_reset() from set-wedged in the reset error path into i915_gem_set_wedged() itself. Even if the reset fails (e.g. if it is disabled by gem_eio), it still tries to make sure the engines are stopped. For i915_gem_set_wedged() callers from outside of i915_reset(), this should make sure the GPU is disabled while the driver is marked as being wedged. Testcase: igt/gem_eio Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Michał Winiarski Cc: Michal Wajdeczko Cc: Michel Thierry Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180315151015.22741-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_drv.c | 1 - drivers/gpu/drm/i915/i915_gem.c | 3 +++ 2 files changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index f03555e..3df5193 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -1995,7 +1995,6 @@ taint: error: i915_gem_set_wedged(i915); i915_retire_requests(i915); - intel_gpu_reset(i915, ALL_ENGINES); goto finish; } diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 2fbd622..802df8e 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -3246,6 +3246,9 @@ void i915_gem_set_wedged(struct drm_i915_private *i915) } i915->caps.scheduler = 0; + /* Even if the GPU reset fails, it should still stop the engines */ + intel_gpu_reset(i915, ALL_ENGINES); + /* * Make sure no one is running the old callback before we proceed with * cancelling requests and resetting the completion tracking. Otherwise -- cgit v1.1 From 0c65dfd1a84142887c810fc11573e2edb8df87f6 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 12 Mar 2018 16:52:04 +0000 Subject: drm/i915/stolen: Switch from DEBUG_KMS to DEBUG_DRIVER MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit i915_gem_stolen is an allocator for the reserved portion of memory ("stolen" from the system by the BIOS). It is not tied to KMS but central to the driver, so prefer DRM_DEBUG_DRIVER. Signed-off-by: Chris Wilson Reviewed-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180312165206.31772-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_gem_stolen.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c index 62aa679..f11a4b9 100644 --- a/drivers/gpu/drm/i915/i915_gem_stolen.c +++ b/drivers/gpu/drm/i915/i915_gem_stolen.c @@ -121,8 +121,8 @@ static int i915_adjust_stolen(struct drm_i915_private *dev_priv, if (stolen[0].start != stolen[1].start || stolen[0].end != stolen[1].end) { - DRM_DEBUG_KMS("GTT within stolen memory at %pR\n", &ggtt_res); - DRM_DEBUG_KMS("Stolen memory adjusted to %pR\n", dsm); + DRM_DEBUG_DRIVER("GTT within stolen memory at %pR\n", &ggtt_res); + DRM_DEBUG_DRIVER("Stolen memory adjusted to %pR\n", dsm); } } @@ -406,9 +406,9 @@ int i915_gem_init_stolen(struct drm_i915_private *dev_priv) * memory, so just consider the start. */ reserved_total = stolen_top - reserved_base; - DRM_DEBUG_KMS("Memory reserved for graphics device: %lluK, usable: %lluK\n", - (u64)resource_size(&dev_priv->dsm) >> 10, - ((u64)resource_size(&dev_priv->dsm) - reserved_total) >> 10); + DRM_DEBUG_DRIVER("Memory reserved for graphics device: %lluK, usable: %lluK\n", + (u64)resource_size(&dev_priv->dsm) >> 10, + ((u64)resource_size(&dev_priv->dsm) - reserved_total) >> 10); stolen_usable_start = 0; /* WaSkipStolenMemoryFirstPage:bdw+ */ @@ -580,8 +580,8 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv lockdep_assert_held(&dev_priv->drm.struct_mutex); - DRM_DEBUG_KMS("creating preallocated stolen object: stolen_offset=%pa, gtt_offset=%pa, size=%pa\n", - &stolen_offset, >t_offset, &size); + DRM_DEBUG_DRIVER("creating preallocated stolen object: stolen_offset=%pa, gtt_offset=%pa, size=%pa\n", + &stolen_offset, >t_offset, &size); /* KISS and expect everything to be page-aligned */ if (WARN_ON(size == 0) || @@ -599,14 +599,14 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv ret = drm_mm_reserve_node(&dev_priv->mm.stolen, stolen); mutex_unlock(&dev_priv->mm.stolen_lock); if (ret) { - DRM_DEBUG_KMS("failed to allocate stolen space\n"); + DRM_DEBUG_DRIVER("failed to allocate stolen space\n"); kfree(stolen); return NULL; } obj = _i915_gem_object_create_stolen(dev_priv, stolen); if (obj == NULL) { - DRM_DEBUG_KMS("failed to allocate stolen object\n"); + DRM_DEBUG_DRIVER("failed to allocate stolen object\n"); i915_gem_stolen_remove_node(dev_priv, stolen); kfree(stolen); return NULL; @@ -635,7 +635,7 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv size, gtt_offset, obj->cache_level, 0); if (ret) { - DRM_DEBUG_KMS("failed to allocate stolen GTT space\n"); + DRM_DEBUG_DRIVER("failed to allocate stolen GTT space\n"); goto err_pages; } -- cgit v1.1 From 0efb656147e04f26433de5a399d1b03bf00e4ed6 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 12 Mar 2018 16:52:05 +0000 Subject: drm/i915/stolen: Checkpatch cleansing MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In the next patch, we will introduce a new vlv_get_stolen_reserved, so before we do, make sure checkpatch is happy with the surrounding code. Sneak in some debug output while we are here. Signed-off-by: Chris Wilson Cc: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180312165206.31772-2-chris@chris-wilson.co.uk Reviewed-by: Ville Syrjälä --- drivers/gpu/drm/i915/i915_gem_stolen.c | 40 ++++++++++++++++++++++++---------- 1 file changed, 28 insertions(+), 12 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c index f11a4b9..5a57a17 100644 --- a/drivers/gpu/drm/i915/i915_gem_stolen.c +++ b/drivers/gpu/drm/i915/i915_gem_stolen.c @@ -174,13 +174,17 @@ void i915_gem_cleanup_stolen(struct drm_device *dev) } static void g4x_get_stolen_reserved(struct drm_i915_private *dev_priv, - resource_size_t *base, resource_size_t *size) + resource_size_t *base, + resource_size_t *size) { - uint32_t reg_val = I915_READ(IS_GM45(dev_priv) ? - CTG_STOLEN_RESERVED : - ELK_STOLEN_RESERVED); + u32 reg_val = I915_READ(IS_GM45(dev_priv) ? + CTG_STOLEN_RESERVED : + ELK_STOLEN_RESERVED); resource_size_t stolen_top = dev_priv->dsm.end + 1; + DRM_DEBUG_DRIVER("%s_STOLEN_RESERVED = %08x\n", + IS_GM45(dev_priv) ? "CTG" : "ELK", reg_val); + if ((reg_val & G4X_STOLEN_RESERVED_ENABLE) == 0) { *base = 0; *size = 0; @@ -208,9 +212,12 @@ static void g4x_get_stolen_reserved(struct drm_i915_private *dev_priv, } static void gen6_get_stolen_reserved(struct drm_i915_private *dev_priv, - resource_size_t *base, resource_size_t *size) + resource_size_t *base, + resource_size_t *size) { - uint32_t reg_val = I915_READ(GEN6_STOLEN_RESERVED); + u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED); + + DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); if ((reg_val & GEN6_STOLEN_RESERVED_ENABLE) == 0) { *base = 0; @@ -240,9 +247,12 @@ static void gen6_get_stolen_reserved(struct drm_i915_private *dev_priv, } static void gen7_get_stolen_reserved(struct drm_i915_private *dev_priv, - resource_size_t *base, resource_size_t *size) + resource_size_t *base, + resource_size_t *size) { - uint32_t reg_val = I915_READ(GEN6_STOLEN_RESERVED); + u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED); + + DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); if ((reg_val & GEN6_STOLEN_RESERVED_ENABLE) == 0) { *base = 0; @@ -266,9 +276,12 @@ static void gen7_get_stolen_reserved(struct drm_i915_private *dev_priv, } static void chv_get_stolen_reserved(struct drm_i915_private *dev_priv, - resource_size_t *base, resource_size_t *size) + resource_size_t *base, + resource_size_t *size) { - uint32_t reg_val = I915_READ(GEN6_STOLEN_RESERVED); + u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED); + + DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); if ((reg_val & GEN6_STOLEN_RESERVED_ENABLE) == 0) { *base = 0; @@ -298,11 +311,14 @@ static void chv_get_stolen_reserved(struct drm_i915_private *dev_priv, } static void bdw_get_stolen_reserved(struct drm_i915_private *dev_priv, - resource_size_t *base, resource_size_t *size) + resource_size_t *base, + resource_size_t *size) { - uint32_t reg_val = I915_READ(GEN6_STOLEN_RESERVED); + u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED); resource_size_t stolen_top; + DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); + if ((reg_val & GEN6_STOLEN_RESERVED_ENABLE) == 0) { *base = 0; *size = 0; -- cgit v1.1 From 957d32feaf04d2a67fd506743e5789359480d574 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 12 Mar 2018 16:52:06 +0000 Subject: drm/i915/stolen: Deduce base of reserved portion as top-size on vlv MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit On Valleyview, the HW deduces the base of the reserved portion of stolen memory as being (top - size) and the address field within GEN6_STOLEN_RESERVED is set to 0. Add yet another GEN6_STOLEN_RESERVED reader to cope with the subtly different path required for vlv. v2: Avoid using reserved_base = reserved_size = 0 as the invalid condition as that typically falls outside of the stolen region, provoking a consistency error. Signed-off-by: Chris Wilson Cc: Ville Syrjälä Cc: Imre Deak Reviewed-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180312165206.31772-3-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_gem_stolen.c | 103 ++++++++++++++++++--------------- 1 file changed, 56 insertions(+), 47 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c index 5a57a17..af915d0 100644 --- a/drivers/gpu/drm/i915/i915_gem_stolen.c +++ b/drivers/gpu/drm/i915/i915_gem_stolen.c @@ -185,11 +185,8 @@ static void g4x_get_stolen_reserved(struct drm_i915_private *dev_priv, DRM_DEBUG_DRIVER("%s_STOLEN_RESERVED = %08x\n", IS_GM45(dev_priv) ? "CTG" : "ELK", reg_val); - if ((reg_val & G4X_STOLEN_RESERVED_ENABLE) == 0) { - *base = 0; - *size = 0; + if ((reg_val & G4X_STOLEN_RESERVED_ENABLE) == 0) return; - } /* * Whether ILK really reuses the ELK register for this is unclear. @@ -197,18 +194,13 @@ static void g4x_get_stolen_reserved(struct drm_i915_private *dev_priv, */ WARN(IS_GEN5(dev_priv), "ILK stolen reserved found? 0x%08x\n", reg_val); - *base = (reg_val & G4X_STOLEN_RESERVED_ADDR2_MASK) << 16; + if (!(reg_val & G4X_STOLEN_RESERVED_ADDR2_MASK)) + return; + *base = (reg_val & G4X_STOLEN_RESERVED_ADDR2_MASK) << 16; WARN_ON((reg_val & G4X_STOLEN_RESERVED_ADDR1_MASK) < *base); - /* On these platforms, the register doesn't have a size field, so the - * size is the distance between the base and the top of the stolen - * memory. We also have the genuine case where base is zero and there's - * nothing reserved. */ - if (*base == 0) - *size = 0; - else - *size = stolen_top - *base; + *size = stolen_top - *base; } static void gen6_get_stolen_reserved(struct drm_i915_private *dev_priv, @@ -219,11 +211,8 @@ static void gen6_get_stolen_reserved(struct drm_i915_private *dev_priv, DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); - if ((reg_val & GEN6_STOLEN_RESERVED_ENABLE) == 0) { - *base = 0; - *size = 0; + if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE)) return; - } *base = reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK; @@ -246,6 +235,33 @@ static void gen6_get_stolen_reserved(struct drm_i915_private *dev_priv, } } +static void vlv_get_stolen_reserved(struct drm_i915_private *dev_priv, + resource_size_t *base, + resource_size_t *size) +{ + u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED); + resource_size_t stolen_top = dev_priv->dsm.end + 1; + + DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); + + if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE)) + return; + + switch (reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK) { + default: + MISSING_CASE(reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK); + case GEN7_STOLEN_RESERVED_1M: + *size = 1024 * 1024; + break; + } + + /* + * On vlv, the ADDR_MASK portion is left as 0 and HW deduces the + * reserved location as (top - size). + */ + *base = stolen_top - *size; +} + static void gen7_get_stolen_reserved(struct drm_i915_private *dev_priv, resource_size_t *base, resource_size_t *size) @@ -254,11 +270,8 @@ static void gen7_get_stolen_reserved(struct drm_i915_private *dev_priv, DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); - if ((reg_val & GEN6_STOLEN_RESERVED_ENABLE) == 0) { - *base = 0; - *size = 0; + if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE)) return; - } *base = reg_val & GEN7_STOLEN_RESERVED_ADDR_MASK; @@ -283,11 +296,8 @@ static void chv_get_stolen_reserved(struct drm_i915_private *dev_priv, DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); - if ((reg_val & GEN6_STOLEN_RESERVED_ENABLE) == 0) { - *base = 0; - *size = 0; + if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE)) return; - } *base = reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK; @@ -315,28 +325,18 @@ static void bdw_get_stolen_reserved(struct drm_i915_private *dev_priv, resource_size_t *size) { u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED); - resource_size_t stolen_top; + resource_size_t stolen_top = dev_priv->dsm.end + 1; DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); - if ((reg_val & GEN6_STOLEN_RESERVED_ENABLE) == 0) { - *base = 0; - *size = 0; + if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE)) return; - } - stolen_top = dev_priv->dsm.end + 1; + if (!(reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK)) + return; *base = reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK; - - /* On these platforms, the register doesn't have a size field, so the - * size is the distance between the base and the top of the stolen - * memory. We also have the genuine case where base is zero and there's - * nothing reserved. */ - if (*base == 0) - *size = 0; - else - *size = stolen_top - *base; + *size = stolen_top - *base; } int i915_gem_init_stolen(struct drm_i915_private *dev_priv) @@ -369,7 +369,7 @@ int i915_gem_init_stolen(struct drm_i915_private *dev_priv) GEM_BUG_ON(dev_priv->dsm.end <= dev_priv->dsm.start); stolen_top = dev_priv->dsm.end + 1; - reserved_base = 0; + reserved_base = stolen_top; reserved_size = 0; switch (INTEL_GEN(dev_priv)) { @@ -389,8 +389,12 @@ int i915_gem_init_stolen(struct drm_i915_private *dev_priv) &reserved_base, &reserved_size); break; case 7: - gen7_get_stolen_reserved(dev_priv, - &reserved_base, &reserved_size); + if (IS_VALLEYVIEW(dev_priv)) + vlv_get_stolen_reserved(dev_priv, + &reserved_base, &reserved_size); + else + gen7_get_stolen_reserved(dev_priv, + &reserved_base, &reserved_size); break; default: if (IS_LP(dev_priv)) @@ -402,11 +406,16 @@ int i915_gem_init_stolen(struct drm_i915_private *dev_priv) break; } - /* It is possible for the reserved base to be zero, but the register - * field for size doesn't have a zero option. */ - if (reserved_base == 0) { - reserved_size = 0; + /* + * Our expectation is that the reserved space is at the top of the + * stolen region and *never* at the bottom. If we see !reserved_base, + * it likely means we failed to read the registers correctly. + */ + if (!reserved_base) { + DRM_ERROR("inconsistent reservation %pa + %pa; ignoring\n", + &reserved_base, &reserved_size); reserved_base = stolen_top; + reserved_size = 0; } dev_priv->dsm_reserved = -- cgit v1.1 From 1947fd133cf0f58b171adc8565685c1a06de07b7 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Mon, 5 Mar 2018 19:41:22 +0200 Subject: drm/i915: Don't initialize plane_to_crtc_mapping[] on SKL+ MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We don't use the enum i9xx_plane_id namespace on SKL+ anymore, so do not initialize the related plane_to_crtc_mapping[] table either. Actually the only remaining user of that table is the pre-g4x watermark code, but no harm in initializing the table on all pre-SKL platforms. Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180305174122.17273-1-ville.syrjala@linux.intel.com Reviewed-by: Dhinakaran Pandiyan --- drivers/gpu/drm/i915/intel_display.c | 15 +++++++++++---- 1 file changed, 11 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index a7bfa23..b31b806 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -13570,10 +13570,17 @@ static int intel_crtc_init(struct drm_i915_private *dev_priv, enum pipe pipe) /* initialize shared scalers */ intel_crtc_init_scalers(intel_crtc, crtc_state); - BUG_ON(pipe >= ARRAY_SIZE(dev_priv->plane_to_crtc_mapping) || - dev_priv->plane_to_crtc_mapping[primary->i9xx_plane] != NULL); - dev_priv->plane_to_crtc_mapping[primary->i9xx_plane] = intel_crtc; - dev_priv->pipe_to_crtc_mapping[intel_crtc->pipe] = intel_crtc; + BUG_ON(pipe >= ARRAY_SIZE(dev_priv->pipe_to_crtc_mapping) || + dev_priv->pipe_to_crtc_mapping[pipe] != NULL); + dev_priv->pipe_to_crtc_mapping[pipe] = intel_crtc; + + if (INTEL_GEN(dev_priv) < 9) { + enum i9xx_plane_id i9xx_plane = primary->i9xx_plane; + + BUG_ON(i9xx_plane >= ARRAY_SIZE(dev_priv->plane_to_crtc_mapping) || + dev_priv->plane_to_crtc_mapping[i9xx_plane] != NULL); + dev_priv->plane_to_crtc_mapping[i9xx_plane] = intel_crtc; + } drm_crtc_helper_add(&intel_crtc->base, &intel_helper_funcs); -- cgit v1.1 From e4006713d16567c203ba710f6a2b709ed6107db5 Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Fri, 16 Mar 2018 16:12:13 +0200 Subject: i915: Re-use DEFINE_SHOW_ATTRIBUTE() macro ...instead of open coding file operations followed by custom ->open() callbacks per each attribute. Reviewed-by: Chris Wilson Signed-off-by: Andy Shevchenko Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180316141213.38774-1-andriy.shevchenko@linux.intel.com --- drivers/gpu/drm/i915/gvt/debugfs.c | 13 +------ drivers/gpu/drm/i915/i915_debugfs.c | 76 ++++++------------------------------- 2 files changed, 12 insertions(+), 77 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/debugfs.c b/drivers/gpu/drm/i915/gvt/debugfs.c index 32a66df..f7d0078 100644 --- a/drivers/gpu/drm/i915/gvt/debugfs.c +++ b/drivers/gpu/drm/i915/gvt/debugfs.c @@ -122,18 +122,7 @@ static int vgpu_mmio_diff_show(struct seq_file *s, void *unused) seq_printf(s, "Total: %d, Diff: %d\n", param.total, param.diff); return 0; } - -static int vgpu_mmio_diff_open(struct inode *inode, struct file *file) -{ - return single_open(file, vgpu_mmio_diff_show, inode->i_private); -} - -static const struct file_operations vgpu_mmio_diff_fops = { - .open = vgpu_mmio_diff_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release, -}; +DEFINE_SHOW_ATTRIBUTE(vgpu_mmio_diff); /** * intel_gvt_debugfs_add_vgpu - register debugfs entries for a vGPU diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 298a3aa..5378863 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -3562,7 +3562,8 @@ static ssize_t i915_displayport_test_active_write(struct file *file, static int i915_displayport_test_active_show(struct seq_file *m, void *data) { - struct drm_device *dev = m->private; + struct drm_i915_private *dev_priv = m->private; + struct drm_device *dev = &dev_priv->drm; struct drm_connector *connector; struct drm_connector_list_iter conn_iter; struct intel_dp *intel_dp; @@ -3596,10 +3597,8 @@ static int i915_displayport_test_active_show(struct seq_file *m, void *data) static int i915_displayport_test_active_open(struct inode *inode, struct file *file) { - struct drm_i915_private *dev_priv = inode->i_private; - return single_open(file, i915_displayport_test_active_show, - &dev_priv->drm); + inode->i_private); } static const struct file_operations i915_displayport_test_active_fops = { @@ -3613,7 +3612,8 @@ static const struct file_operations i915_displayport_test_active_fops = { static int i915_displayport_test_data_show(struct seq_file *m, void *data) { - struct drm_device *dev = m->private; + struct drm_i915_private *dev_priv = m->private; + struct drm_device *dev = &dev_priv->drm; struct drm_connector *connector; struct drm_connector_list_iter conn_iter; struct intel_dp *intel_dp; @@ -3652,26 +3652,12 @@ static int i915_displayport_test_data_show(struct seq_file *m, void *data) return 0; } -static int i915_displayport_test_data_open(struct inode *inode, - struct file *file) -{ - struct drm_i915_private *dev_priv = inode->i_private; - - return single_open(file, i915_displayport_test_data_show, - &dev_priv->drm); -} - -static const struct file_operations i915_displayport_test_data_fops = { - .owner = THIS_MODULE, - .open = i915_displayport_test_data_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release -}; +DEFINE_SHOW_ATTRIBUTE(i915_displayport_test_data); static int i915_displayport_test_type_show(struct seq_file *m, void *data) { - struct drm_device *dev = m->private; + struct drm_i915_private *dev_priv = m->private; + struct drm_device *dev = &dev_priv->drm; struct drm_connector *connector; struct drm_connector_list_iter conn_iter; struct intel_dp *intel_dp; @@ -3698,23 +3684,7 @@ static int i915_displayport_test_type_show(struct seq_file *m, void *data) return 0; } - -static int i915_displayport_test_type_open(struct inode *inode, - struct file *file) -{ - struct drm_i915_private *dev_priv = inode->i_private; - - return single_open(file, i915_displayport_test_type_show, - &dev_priv->drm); -} - -static const struct file_operations i915_displayport_test_type_fops = { - .owner = THIS_MODULE, - .open = i915_displayport_test_type_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release -}; +DEFINE_SHOW_ATTRIBUTE(i915_displayport_test_type); static void wm_latency_show(struct seq_file *m, const uint16_t wm[8]) { @@ -4875,19 +4845,7 @@ static int i915_dpcd_show(struct seq_file *m, void *data) return 0; } - -static int i915_dpcd_open(struct inode *inode, struct file *file) -{ - return single_open(file, i915_dpcd_show, inode->i_private); -} - -static const struct file_operations i915_dpcd_fops = { - .owner = THIS_MODULE, - .open = i915_dpcd_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release, -}; +DEFINE_SHOW_ATTRIBUTE(i915_dpcd); static int i915_panel_show(struct seq_file *m, void *data) { @@ -4909,19 +4867,7 @@ static int i915_panel_show(struct seq_file *m, void *data) return 0; } - -static int i915_panel_open(struct inode *inode, struct file *file) -{ - return single_open(file, i915_panel_show, inode->i_private); -} - -static const struct file_operations i915_panel_fops = { - .owner = THIS_MODULE, - .open = i915_panel_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release, -}; +DEFINE_SHOW_ATTRIBUTE(i915_panel); /** * i915_debugfs_connector_add - add i915 specific connector debugfs files -- cgit v1.1 From 3b358cdaf3319521efdf19ff07918bcc4d57013e Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Fri, 2 Mar 2018 11:56:56 +0200 Subject: drm/i915: Kill the remaining CHV HBR2 leftovers MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit AFAIK CHV was supposed to have HBR2 originally, but in the end the feature was dropped. We still have some code leftovers from those early days. Eliminate them. The extra bit for the training pattern seems to be dead in the hardware. I can set it (in fact I can set almost any reserved bit in the registers) but it doesn't seem to interfere with the operation of the hardware. Either that or I'm very lucky that my displays complete link training with the incorrect pattern being sent out. Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180302095656.19662-1-ville.syrjala@linux.intel.com Reviewed-by: Dhinakaran Pandiyan --- drivers/gpu/drm/i915/i915_reg.h | 2 -- drivers/gpu/drm/i915/intel_dp.c | 20 ++++---------------- 2 files changed, 4 insertions(+), 18 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index dbcb882..1b48d50 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -4943,8 +4943,6 @@ enum { #define DP_LINK_TRAIN_OFF (3 << 28) #define DP_LINK_TRAIN_MASK (3 << 28) #define DP_LINK_TRAIN_SHIFT 28 -#define DP_LINK_TRAIN_PAT_3_CHV (1 << 14) -#define DP_LINK_TRAIN_MASK_CHV ((3 << 28)|(1<<14)) /* CPT Link training mode */ #define DP_LINK_TRAIN_PAT_1_CPT (0 << 8) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 4dd1b22..62f82c4 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -91,8 +91,6 @@ static const struct dp_link_dpll chv_dpll[] = { { .p1 = 4, .p2 = 2, .n = 1, .m1 = 2, .m2 = 0x819999a } }, { 270000, /* m2_int = 27, m2_fraction = 0 */ { .p1 = 4, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x6c00000 } }, - { 540000, /* m2_int = 27, m2_fraction = 0 */ - { .p1 = 2, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x6c00000 } } }; /** @@ -2900,10 +2898,7 @@ _intel_dp_set_link_train(struct intel_dp *intel_dp, } } else { - if (IS_CHERRYVIEW(dev_priv)) - *DP &= ~DP_LINK_TRAIN_MASK_CHV; - else - *DP &= ~DP_LINK_TRAIN_MASK; + *DP &= ~DP_LINK_TRAIN_MASK; switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) { case DP_TRAINING_PATTERN_DISABLE: @@ -2916,12 +2911,8 @@ _intel_dp_set_link_train(struct intel_dp *intel_dp, *DP |= DP_LINK_TRAIN_PAT_2; break; case DP_TRAINING_PATTERN_3: - if (IS_CHERRYVIEW(dev_priv)) { - *DP |= DP_LINK_TRAIN_PAT_3_CHV; - } else { - DRM_DEBUG_KMS("TPS3 not supported, using TPS2 instead\n"); - *DP |= DP_LINK_TRAIN_PAT_2; - } + DRM_DEBUG_KMS("TPS3 not supported, using TPS2 instead\n"); + *DP |= DP_LINK_TRAIN_PAT_2; break; } } @@ -3660,10 +3651,7 @@ intel_dp_link_down(struct intel_encoder *encoder, DP &= ~DP_LINK_TRAIN_MASK_CPT; DP |= DP_LINK_TRAIN_PAT_IDLE_CPT; } else { - if (IS_CHERRYVIEW(dev_priv)) - DP &= ~DP_LINK_TRAIN_MASK_CHV; - else - DP &= ~DP_LINK_TRAIN_MASK; + DP &= ~DP_LINK_TRAIN_MASK; DP |= DP_LINK_TRAIN_PAT_IDLE; } I915_WRITE(intel_dp->output_reg, DP); -- cgit v1.1 From ad260ab32a4d94fa974f58262f8000472d34fd5b Mon Sep 17 00:00:00 2001 From: Dhinakaran Pandiyan Date: Tue, 13 Mar 2018 22:48:25 -0700 Subject: drm/i915/dp: Write to SET_POWER dpcd to enable MST hub. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If bios sets up an MST output and hardware state readout code sees this is an SST configuration, when disabling the encoder we end up calling ->post_disable_dp() hook instead of the MST version. Consequently, we write to the DP_SET_POWER dpcd to set it D3 state. Further along when we try enable the encoder in MST mode, POWER_UP_PHY transaction fails to power up the MST hub. This results in continuous link training failures which keep the system busy delaying boot. We could identify bios MST boot discrepancy and handle it accordingly but a simple way to solve this is to write to the DP_SET_POWER dpcd for MST too. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105470 Cc: Ville Syrjälä Cc: Jani Nikula Reviewed-by: Ville Syrjälä Reported-by: Laura Abbott Cc: stable@vger.kernel.org Fixes: 5ea2355a100a ("drm/i915/mst: Use MST sideband message transactions for dpms control") Tested-by: Laura Abbott Signed-off-by: Dhinakaran Pandiyan Signed-off-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20180314054825.1718-1-dhinakaran.pandiyan@intel.com --- drivers/gpu/drm/i915/intel_ddi.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index dbcf1a0..8c2d778 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -2205,8 +2205,7 @@ static void intel_ddi_pre_enable_dp(struct intel_encoder *encoder, intel_prepare_dp_ddi_buffers(encoder, crtc_state); intel_ddi_init_dp_buf_reg(encoder); - if (!is_mst) - intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON); + intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON); intel_dp_start_link_train(intel_dp); if (port != PORT_A || INTEL_GEN(dev_priv) >= 9) intel_dp_stop_link_train(intel_dp); @@ -2304,14 +2303,12 @@ static void intel_ddi_post_disable_dp(struct intel_encoder *encoder, struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_digital_port *dig_port = enc_to_dig_port(&encoder->base); struct intel_dp *intel_dp = &dig_port->dp; - bool is_mst = intel_crtc_has_type(old_crtc_state, INTEL_OUTPUT_DP_MST); /* * Power down sink before disabling the port, otherwise we end * up getting interrupts from the sink on detecting link loss. */ - if (!is_mst) - intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_OFF); + intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_OFF); intel_disable_ddi_buf(encoder); -- cgit v1.1 From eacd8391f977d3800cc41a026f9f81fce210a78c Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Mon, 19 Mar 2018 10:53:36 +0100 Subject: drm/i915/guc: Keep GuC interrupts enabled when using GuC MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The GuC log contains a separate space used for crash dump. We even get a separate notification for it. While we're not handling crash differently yet, it makes sense to decouple the two right now to simplify the following patches. v2: Move guc_log_flush_irq_disable up to avoid movement in following patches (Sagar). v3: s/guc_log_flush_irq_*/guc_flush_log_msg_*, rebase after mass rename Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Reviewed-by: Sagar Arun Kamble (v2) Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319095348.9716-1-michal.winiarski@intel.com --- drivers/gpu/drm/i915/intel_guc.c | 25 ++++++++++--------------- drivers/gpu/drm/i915/intel_guc.h | 2 ++ drivers/gpu/drm/i915/intel_guc_log.c | 31 +++++++++++++++++++------------ drivers/gpu/drm/i915/intel_uc.c | 14 +++++--------- 4 files changed, 36 insertions(+), 36 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index e70bf65..3af6035 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -67,6 +67,7 @@ void intel_guc_init_early(struct intel_guc *guc) intel_guc_log_init_early(&guc->log); mutex_init(&guc->send_mutex); + spin_lock_init(&guc->irq_lock); guc->send = intel_guc_send_nop; guc->notify = gen8_guc_raise_irq; } @@ -368,7 +369,7 @@ int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len) void intel_guc_to_host_event_handler(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); - u32 msg, flush; + u32 msg, val; /* * Sample the log buffer flush related bits & clear them out now @@ -381,24 +382,18 @@ void intel_guc_to_host_event_handler(struct intel_guc *guc) * could happen that GuC sets the bit for 2nd interrupt but Host * clears out the bit on handling the 1st interrupt. */ - - msg = I915_READ(SOFT_SCRATCH(15)); - flush = msg & (INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED | - INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER); - if (flush) { - /* Clear the message bits that are handled */ - I915_WRITE(SOFT_SCRATCH(15), msg & ~flush); - - /* Handle flush interrupt in bottom half */ + spin_lock(&guc->irq_lock); + val = I915_READ(SOFT_SCRATCH(15)); + msg = val & guc->msg_enabled_mask; + I915_WRITE(SOFT_SCRATCH(15), val & ~msg); + spin_unlock(&guc->irq_lock); + + if (msg & (INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | + INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED)) { queue_work(guc->log.runtime.flush_wq, &guc->log.runtime.flush_work); guc->log.flush_interrupt_count++; - } else { - /* - * Not clearing of unhandled event bits won't result in - * re-triggering of the interrupt. - */ } } diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h index cdb649a9..9a95d15 100644 --- a/drivers/gpu/drm/i915/intel_guc.h +++ b/drivers/gpu/drm/i915/intel_guc.h @@ -56,7 +56,9 @@ struct intel_guc { struct drm_i915_gem_object *load_err_log; /* intel_guc_recv interrupt related state */ + spinlock_t irq_lock; bool interrupts_enabled; + unsigned int msg_enabled_mask; struct i915_vma *ads_vma; struct i915_vma *stage_desc_pool; diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 1c2127b..1e209fc 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -73,6 +73,22 @@ static int guc_log_control(struct intel_guc *guc, bool enable, u32 verbosity) return intel_guc_send(guc, action, ARRAY_SIZE(action)); } +static void guc_flush_log_msg_enable(struct intel_guc *guc) +{ + spin_lock_irq(&guc->irq_lock); + guc->msg_enabled_mask |= INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | + INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED; + spin_unlock_irq(&guc->irq_lock); +} + +static void guc_flush_log_msg_disable(struct intel_guc *guc) +{ + spin_lock_irq(&guc->irq_lock); + guc->msg_enabled_mask &= ~(INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | + INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED); + spin_unlock_irq(&guc->irq_lock); +} + static inline struct intel_guc *log_to_guc(struct intel_guc_log *log) { return container_of(log, struct intel_guc, log); @@ -709,12 +725,7 @@ int intel_guc_log_register(struct intel_guc_log *log) if (ret) goto err_runtime; - /* GuC logging is currently the only user of Guc2Host interrupts */ - mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); - gen9_enable_guc_interrupts(i915); - intel_runtime_pm_put(i915); - mutex_unlock(&i915->drm.struct_mutex); + guc_flush_log_msg_enable(guc); return 0; @@ -733,6 +744,8 @@ void intel_guc_log_unregister(struct intel_guc_log *log) struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *i915 = guc_to_i915(guc); + guc_flush_log_msg_disable(guc); + /* * Once logging is disabled, GuC won't generate logs & send an * interrupt. But there could be some data in the log buffer @@ -742,12 +755,6 @@ void intel_guc_log_unregister(struct intel_guc_log *log) guc_flush_logs(log); mutex_lock(&i915->drm.struct_mutex); - - /* GuC logging is currently the only user of Guc2Host interrupts */ - intel_runtime_pm_get(i915); - gen9_disable_guc_interrupts(i915); - intel_runtime_pm_put(i915); - guc_log_runtime_destroy(log); mutex_unlock(&i915->drm.struct_mutex); diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 104c03a..765b86a 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -247,6 +247,8 @@ static int guc_enable_communication(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); + gen9_enable_guc_interrupts(dev_priv); + if (HAS_GUC_CT(dev_priv)) return intel_guc_enable_ct(guc); @@ -261,6 +263,8 @@ static void guc_disable_communication(struct intel_guc *guc) if (HAS_GUC_CT(dev_priv)) intel_guc_disable_ct(guc); + gen9_disable_guc_interrupts(dev_priv); + guc->send = intel_guc_send_nop; } @@ -413,12 +417,9 @@ int intel_uc_init_hw(struct drm_i915_private *dev_priv) } if (USES_GUC_SUBMISSION(dev_priv)) { - if (i915_modparams.guc_log_level) - gen9_enable_guc_interrupts(dev_priv); - ret = intel_guc_submission_enable(guc); if (ret) - goto err_interrupts; + goto err_communication; } dev_info(dev_priv->drm.dev, "GuC firmware version %u.%u\n", @@ -433,8 +434,6 @@ int intel_uc_init_hw(struct drm_i915_private *dev_priv) /* * We've failed to load the firmware :( */ -err_interrupts: - gen9_disable_guc_interrupts(dev_priv); err_communication: guc_disable_communication(guc); err_log_capture: @@ -464,9 +463,6 @@ void intel_uc_fini_hw(struct drm_i915_private *dev_priv) intel_guc_submission_disable(guc); guc_disable_communication(guc); - - if (USES_GUC_SUBMISSION(dev_priv)) - gen9_disable_guc_interrupts(dev_priv); } int intel_uc_suspend(struct drm_i915_private *i915) -- cgit v1.1 From b813d50e869aeda09ccf22e8d869e61b63389e4d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Mon, 19 Mar 2018 10:53:37 +0100 Subject: drm/i915/guc: Log runtime should consist of both mapping and relay MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Currently, we're treating relay and mapping of GuC log as a separate concepts. We're also using inconsistent locking, sometimes using relay_lock, sometimes using struct mutex. Let's correct that. Anything touching the runtime is now serialized using runtime.lock, while we're still using struct mutex as inner lock for mapping. We're still racy in setting the log level - but we'll take care of that in the following patches. v2: Tidy locking (Sagar) v3: Remove obsoleted comment (Sagar) Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319095348.9716-2-michal.winiarski@intel.com --- drivers/gpu/drm/i915/intel_guc_log.c | 125 +++++++++++------------------------ drivers/gpu/drm/i915/intel_guc_log.h | 3 +- 2 files changed, 38 insertions(+), 90 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 1e209fc..b82866b 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -176,10 +176,7 @@ static int guc_log_relay_file_create(struct intel_guc_log *log) struct dentry *log_dir; int ret; - if (!i915_modparams.guc_log_level) - return 0; - - mutex_lock(&log->runtime.relay_lock); + lockdep_assert_held(&log->runtime.lock); /* For now create the log file in /sys/kernel/debug/dri/0 dir */ log_dir = dev_priv->drm.primary->debugfs_root; @@ -198,29 +195,17 @@ static int guc_log_relay_file_create(struct intel_guc_log *log) */ if (!log_dir) { DRM_ERROR("Debugfs dir not available yet for GuC log file\n"); - ret = -ENODEV; - goto out_unlock; + return -ENODEV; } ret = relay_late_setup_files(log->runtime.relay_chan, "guc_log", log_dir); if (ret < 0 && ret != -EEXIST) { DRM_ERROR("Couldn't associate relay chan with file %d\n", ret); - goto out_unlock; + return ret; } - ret = 0; - -out_unlock: - mutex_unlock(&log->runtime.relay_lock); - return ret; -} - -static bool guc_log_has_relay(struct intel_guc_log *log) -{ - lockdep_assert_held(&log->runtime.relay_lock); - - return log->runtime.relay_chan; + return 0; } static void guc_move_to_next_buf(struct intel_guc_log *log) @@ -231,9 +216,6 @@ static void guc_move_to_next_buf(struct intel_guc_log *log) */ smp_wmb(); - if (!guc_log_has_relay(log)) - return; - /* All data has been written, so now move the offset of sub buffer. */ relay_reserve(log->runtime.relay_chan, log->vma->obj->base.size); @@ -243,9 +225,6 @@ static void guc_move_to_next_buf(struct intel_guc_log *log) static void *guc_get_write_buffer(struct intel_guc_log *log) { - if (!guc_log_has_relay(log)) - return NULL; - /* * Just get the base address of a new sub buffer and copy data into it * ourselves. NULL will be returned in no-overwrite mode, if all sub @@ -306,14 +285,14 @@ static void guc_read_update_log_buffer(struct intel_guc_log *log) void *src_data, *dst_data; bool new_overflow; + mutex_lock(&log->runtime.lock); + if (WARN_ON(!log->runtime.buf_addr)) - return; + goto out_unlock; /* Get the pointer to shared GuC log buffer */ log_buf_state = src_data = log->runtime.buf_addr; - mutex_lock(&log->runtime.relay_lock); - /* Get the pointer to local buffer to store the logs */ log_buf_snapshot_state = dst_data = guc_get_write_buffer(log); @@ -324,9 +303,8 @@ static void guc_read_update_log_buffer(struct intel_guc_log *log) */ DRM_ERROR_RATELIMITED("no sub-buffer to capture logs\n"); log->capture_miss_count++; - mutex_unlock(&log->runtime.relay_lock); - return; + goto out_unlock; } /* Actual logs are present from the 2nd page */ @@ -397,7 +375,8 @@ static void guc_read_update_log_buffer(struct intel_guc_log *log) guc_move_to_next_buf(log); - mutex_unlock(&log->runtime.relay_lock); +out_unlock: + mutex_unlock(&log->runtime.lock); } static void capture_logs_work(struct work_struct *work) @@ -413,21 +392,21 @@ static bool guc_log_has_runtime(struct intel_guc_log *log) return log->runtime.buf_addr; } -static int guc_log_runtime_create(struct intel_guc_log *log) +static int guc_log_map(struct intel_guc_log *log) { struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); void *vaddr; int ret; - lockdep_assert_held(&dev_priv->drm.struct_mutex); + lockdep_assert_held(&log->runtime.lock); if (!log->vma) return -ENODEV; - GEM_BUG_ON(guc_log_has_runtime(log)); - + mutex_lock(&dev_priv->drm.struct_mutex); ret = i915_gem_object_set_to_wc_domain(log->vma->obj, true); + mutex_unlock(&dev_priv->drm.struct_mutex); if (ret) return ret; @@ -447,14 +426,9 @@ static int guc_log_runtime_create(struct intel_guc_log *log) return 0; } -static void guc_log_runtime_destroy(struct intel_guc_log *log) +static void guc_log_unmap(struct intel_guc_log *log) { - /* - * It's possible that the runtime stuff was never allocated because - * GuC log was disabled at the boot time. - */ - if (!guc_log_has_runtime(log)) - return; + lockdep_assert_held(&log->runtime.lock); i915_gem_object_unpin_map(log->vma->obj); log->runtime.buf_addr = NULL; @@ -462,7 +436,7 @@ static void guc_log_runtime_destroy(struct intel_guc_log *log) void intel_guc_log_init_early(struct intel_guc_log *log) { - mutex_init(&log->runtime.relay_lock); + mutex_init(&log->runtime.lock); INIT_WORK(&log->runtime.flush_work, capture_logs_work); } @@ -474,12 +448,7 @@ static int guc_log_relay_create(struct intel_guc_log *log) size_t n_subbufs, subbuf_size; int ret; - if (!i915_modparams.guc_log_level) - return 0; - - mutex_lock(&log->runtime.relay_lock); - - GEM_BUG_ON(guc_log_has_relay(log)); + lockdep_assert_held(&log->runtime.lock); /* Keep the size of sub buffers same as shared log buffer */ subbuf_size = GUC_LOG_SIZE; @@ -509,12 +478,9 @@ static int guc_log_relay_create(struct intel_guc_log *log) GEM_BUG_ON(guc_log_relay_chan->subbuf_size < subbuf_size); log->runtime.relay_chan = guc_log_relay_chan; - mutex_unlock(&log->runtime.relay_lock); - return 0; err: - mutex_unlock(&log->runtime.relay_lock); /* logging will be off */ i915_modparams.guc_log_level = 0; return ret; @@ -522,20 +488,10 @@ err: static void guc_log_relay_destroy(struct intel_guc_log *log) { - mutex_lock(&log->runtime.relay_lock); - - /* - * It's possible that the relay was never allocated because - * GuC log was disabled at the boot time. - */ - if (!guc_log_has_relay(log)) - goto out_unlock; + lockdep_assert_held(&log->runtime.lock); relay_close(log->runtime.relay_chan); log->runtime.relay_chan = NULL; - -out_unlock: - mutex_unlock(&log->runtime.relay_lock); } static void guc_log_capture_logs(struct intel_guc_log *log) @@ -621,7 +577,6 @@ err: void intel_guc_log_destroy(struct intel_guc_log *log) { - guc_log_runtime_destroy(log); i915_vma_unpin_and_release(&log->vma); } @@ -699,52 +654,43 @@ out: int intel_guc_log_register(struct intel_guc_log *log) { - struct intel_guc *guc = log_to_guc(log); - struct drm_i915_private *i915 = guc_to_i915(guc); int ret; + mutex_lock(&log->runtime.lock); + GEM_BUG_ON(guc_log_has_runtime(log)); - /* - * If log was disabled at boot time, then setup needed to handle - * log buffer flush interrupts would not have been done yet, so - * do that now. - */ ret = guc_log_relay_create(log); if (ret) goto err; - mutex_lock(&i915->drm.struct_mutex); - ret = guc_log_runtime_create(log); - mutex_unlock(&i915->drm.struct_mutex); - + ret = guc_log_map(log); if (ret) goto err_relay; ret = guc_log_relay_file_create(log); if (ret) - goto err_runtime; + goto err_unmap; - guc_flush_log_msg_enable(guc); + guc_flush_log_msg_enable(log_to_guc(log)); + + mutex_unlock(&log->runtime.lock); return 0; -err_runtime: - mutex_lock(&i915->drm.struct_mutex); - guc_log_runtime_destroy(log); - mutex_unlock(&i915->drm.struct_mutex); +err_unmap: + guc_log_unmap(log); err_relay: guc_log_relay_destroy(log); err: + mutex_unlock(&log->runtime.lock); + return ret; } void intel_guc_log_unregister(struct intel_guc_log *log) { - struct intel_guc *guc = log_to_guc(log); - struct drm_i915_private *i915 = guc_to_i915(guc); - - guc_flush_log_msg_disable(guc); + guc_flush_log_msg_disable(log_to_guc(log)); /* * Once logging is disabled, GuC won't generate logs & send an @@ -754,9 +700,12 @@ void intel_guc_log_unregister(struct intel_guc_log *log) */ guc_flush_logs(log); - mutex_lock(&i915->drm.struct_mutex); - guc_log_runtime_destroy(log); - mutex_unlock(&i915->drm.struct_mutex); + mutex_lock(&log->runtime.lock); + + GEM_BUG_ON(!guc_log_has_runtime(log)); + guc_log_unmap(log); guc_log_relay_destroy(log); + + mutex_unlock(&log->runtime.lock); } diff --git a/drivers/gpu/drm/i915/intel_guc_log.h b/drivers/gpu/drm/i915/intel_guc_log.h index 6264bd5..e0ea6250 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.h +++ b/drivers/gpu/drm/i915/intel_guc_log.h @@ -49,8 +49,7 @@ struct intel_guc_log { struct workqueue_struct *flush_wq; struct work_struct flush_work; struct rchan *relay_chan; - /* To serialize the access to relay_chan */ - struct mutex relay_lock; + struct mutex lock; } runtime; /* logging related stats */ u32 capture_miss_count; -- cgit v1.1 From 2b47733045aaf883c275c3bdbe3b503137144f6e Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Mon, 19 Mar 2018 10:53:38 +0100 Subject: drm/i915/guc: Merge log relay file and channel creation MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We have all the information we need at relay_open call time. Since there's no reason to split the process into relay_open and relay_late_setup_files, let's remove the extra code. v2: Remove obsoleted comments (Sagar) v3: There was one obsolete comment left (Sagar) Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319095348.9716-3-michal.winiarski@intel.com --- drivers/gpu/drm/i915/intel_guc_log.c | 65 +++--------------------------------- 1 file changed, 5 insertions(+), 60 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index b82866b..767c0d0 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -141,14 +141,7 @@ static struct dentry *create_buf_file_callback(const char *filename, if (!parent) return NULL; - /* - * Not using the channel filename passed as an argument, since for each - * channel relay appends the corresponding CPU number to the filename - * passed in relay_open(). This should be fine as relay just needs a - * dentry of the file associated with the channel buffer and that file's - * name need not be same as the filename passed as an argument. - */ - buf_file = debugfs_create_file("guc_log", mode, + buf_file = debugfs_create_file(filename, mode, parent, buf, &relay_file_operations); return buf_file; } @@ -169,45 +162,6 @@ static struct rchan_callbacks relay_callbacks = { .remove_buf_file = remove_buf_file_callback, }; -static int guc_log_relay_file_create(struct intel_guc_log *log) -{ - struct intel_guc *guc = log_to_guc(log); - struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct dentry *log_dir; - int ret; - - lockdep_assert_held(&log->runtime.lock); - - /* For now create the log file in /sys/kernel/debug/dri/0 dir */ - log_dir = dev_priv->drm.primary->debugfs_root; - - /* - * If /sys/kernel/debug/dri/0 location do not exist, then debugfs is - * not mounted and so can't create the relay file. - * The relay API seems to fit well with debugfs only, for availing relay - * there are 3 requirements which can be met for debugfs file only in a - * straightforward/clean manner :- - * i) Need the associated dentry pointer of the file, while opening the - * relay channel. - * ii) Should be able to use 'relay_file_operations' fops for the file. - * iii) Set the 'i_private' field of file's inode to the pointer of - * relay channel buffer. - */ - if (!log_dir) { - DRM_ERROR("Debugfs dir not available yet for GuC log file\n"); - return -ENODEV; - } - - ret = relay_late_setup_files(log->runtime.relay_chan, "guc_log", - log_dir); - if (ret < 0 && ret != -EEXIST) { - DRM_ERROR("Couldn't associate relay chan with file %d\n", ret); - return ret; - } - - return 0; -} - static void guc_move_to_next_buf(struct intel_guc_log *log) { /* @@ -461,13 +415,10 @@ static int guc_log_relay_create(struct intel_guc_log *log) */ n_subbufs = 8; - /* - * Create a relay channel, so that we have buffers for storing - * the GuC firmware logs, the channel will be linked with a file - * later on when debugfs is registered. - */ - guc_log_relay_chan = relay_open(NULL, NULL, subbuf_size, - n_subbufs, &relay_callbacks, dev_priv); + guc_log_relay_chan = relay_open("guc_log", + dev_priv->drm.primary->debugfs_root, + subbuf_size, n_subbufs, + &relay_callbacks, dev_priv); if (!guc_log_relay_chan) { DRM_ERROR("Couldn't create relay chan for GuC logging\n"); @@ -668,18 +619,12 @@ int intel_guc_log_register(struct intel_guc_log *log) if (ret) goto err_relay; - ret = guc_log_relay_file_create(log); - if (ret) - goto err_unmap; - guc_flush_log_msg_enable(log_to_guc(log)); mutex_unlock(&log->runtime.lock); return 0; -err_unmap: - guc_log_unmap(log); err_relay: guc_log_relay_destroy(log); err: -- cgit v1.1 From d3fbf9437b22bd663e292d5d5e9f8e37c8eed208 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Mon, 19 Mar 2018 10:53:39 +0100 Subject: drm/i915/guc: Flush directly in log unregister MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Having both guc_flush_logs and guc_log_flush functions is confusing. While we could just rename things, guc_flush_logs implementation is quite simple. Let's get rid of it and move its content to unregister. v2: s/dev_priv/i915 (Sagar) Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319095348.9716-4-michal.winiarski@intel.com --- drivers/gpu/drm/i915/intel_guc_log.c | 38 +++++++++++++++--------------------- 1 file changed, 16 insertions(+), 22 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 767c0d0..72a71bc 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -461,26 +461,6 @@ static void guc_log_capture_logs(struct intel_guc_log *log) intel_runtime_pm_put(dev_priv); } -static void guc_flush_logs(struct intel_guc_log *log) -{ - struct intel_guc *guc = log_to_guc(log); - struct drm_i915_private *dev_priv = guc_to_i915(guc); - - /* - * Before initiating the forceful flush, wait for any pending/ongoing - * flush to complete otherwise forceful flush may not actually happen. - */ - flush_work(&log->runtime.flush_work); - - /* Ask GuC to update the log buffer state */ - intel_runtime_pm_get(dev_priv); - guc_log_flush(guc); - intel_runtime_pm_put(dev_priv); - - /* GuC would have updated log buffer by now, so capture it */ - guc_log_capture_logs(log); -} - int intel_guc_log_create(struct intel_guc_log *log) { struct intel_guc *guc = log_to_guc(log); @@ -635,7 +615,16 @@ err: void intel_guc_log_unregister(struct intel_guc_log *log) { - guc_flush_log_msg_disable(log_to_guc(log)); + struct intel_guc *guc = log_to_guc(log); + struct drm_i915_private *i915 = guc_to_i915(guc); + + guc_flush_log_msg_disable(guc); + + /* + * Before initiating the forceful flush, wait for any pending/ongoing + * flush to complete otherwise forceful flush may not actually happen. + */ + flush_work(&log->runtime.flush_work); /* * Once logging is disabled, GuC won't generate logs & send an @@ -643,7 +632,12 @@ void intel_guc_log_unregister(struct intel_guc_log *log) * which is yet to be captured. So request GuC to update the log * buffer state and then collect the left over logs. */ - guc_flush_logs(log); + intel_runtime_pm_get(i915); + guc_log_flush(guc); + intel_runtime_pm_put(i915); + + /* GuC would have updated log buffer by now, so capture it */ + guc_log_capture_logs(log); mutex_lock(&log->runtime.lock); -- cgit v1.1 From 4977a287b9e7c3dbe156bf28f8771b758060ee3e Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Mon, 19 Mar 2018 10:53:40 +0100 Subject: drm/i915/guc: Split relay control and GuC log level MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Those two concepts are really separate. Since GuC is writing data into its own buffer and we even provide a way for userspace to read directly from it using i915_guc_log_dump debugfs, there's no real reason to tie log level with relay creation. Let's create a separate debugfs, giving userspace a way to create a relay on demand, when it wants to read a continuous log rather than a snapshot. v2: Don't touch guc_log_level on relay creation error, adjust locking after rebase, s/dev_priv/i915, pass guc to file->private_data (Sagar) Use struct_mutex rather than runtime.lock for set_log_level v3: Tidy ordering of definitions (Sagar) Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319095348.9716-5-michal.winiarski@intel.com --- drivers/gpu/drm/i915/i915_debugfs.c | 56 +++++++++++++++++++++++---- drivers/gpu/drm/i915/i915_drv.c | 4 -- drivers/gpu/drm/i915/intel_guc_log.c | 75 +++++++++++++++--------------------- drivers/gpu/drm/i915/intel_guc_log.h | 9 +++-- drivers/gpu/drm/i915/intel_uc.c | 22 ----------- drivers/gpu/drm/i915/intel_uc.h | 2 - 6 files changed, 84 insertions(+), 84 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 5378863..e857a94 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -2495,32 +2495,73 @@ static int i915_guc_log_dump(struct seq_file *m, void *data) return 0; } -static int i915_guc_log_control_get(void *data, u64 *val) +static int i915_guc_log_level_get(void *data, u64 *val) { struct drm_i915_private *dev_priv = data; if (!USES_GUC(dev_priv)) return -ENODEV; - *val = intel_guc_log_control_get(&dev_priv->guc.log); + *val = intel_guc_log_level_get(&dev_priv->guc.log); return 0; } -static int i915_guc_log_control_set(void *data, u64 val) +static int i915_guc_log_level_set(void *data, u64 val) { struct drm_i915_private *dev_priv = data; if (!USES_GUC(dev_priv)) return -ENODEV; - return intel_guc_log_control_set(&dev_priv->guc.log, val); + return intel_guc_log_level_set(&dev_priv->guc.log, val); } -DEFINE_SIMPLE_ATTRIBUTE(i915_guc_log_control_fops, - i915_guc_log_control_get, i915_guc_log_control_set, +DEFINE_SIMPLE_ATTRIBUTE(i915_guc_log_level_fops, + i915_guc_log_level_get, i915_guc_log_level_set, "%lld\n"); +static int i915_guc_log_relay_open(struct inode *inode, struct file *file) +{ + struct drm_i915_private *dev_priv = inode->i_private; + + if (!USES_GUC(dev_priv)) + return -ENODEV; + + file->private_data = &dev_priv->guc.log; + + return intel_guc_log_relay_open(&dev_priv->guc.log); +} + +static ssize_t +i915_guc_log_relay_write(struct file *filp, + const char __user *ubuf, + size_t cnt, + loff_t *ppos) +{ + struct intel_guc_log *log = filp->private_data; + + intel_guc_log_relay_flush(log); + + return cnt; +} + +static int i915_guc_log_relay_release(struct inode *inode, struct file *file) +{ + struct drm_i915_private *dev_priv = inode->i_private; + + intel_guc_log_relay_close(&dev_priv->guc.log); + + return 0; +} + +static const struct file_operations i915_guc_log_relay_fops = { + .owner = THIS_MODULE, + .open = i915_guc_log_relay_open, + .write = i915_guc_log_relay_write, + .release = i915_guc_log_relay_release, +}; + static const char *psr2_live_status(u32 val) { static const char * const live_status[] = { @@ -4748,7 +4789,8 @@ static const struct i915_debugfs_files { {"i915_dp_test_data", &i915_displayport_test_data_fops}, {"i915_dp_test_type", &i915_displayport_test_type_fops}, {"i915_dp_test_active", &i915_displayport_test_active_fops}, - {"i915_guc_log_control", &i915_guc_log_control_fops}, + {"i915_guc_log_level", &i915_guc_log_level_fops}, + {"i915_guc_log_relay", &i915_guc_log_relay_fops}, {"i915_hpd_storm_ctl", &i915_hpd_storm_ctl_fops}, {"i915_ipc_status", &i915_ipc_status_fops}, {"i915_drrs_ctl", &i915_drrs_ctl_fops} diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 3df5193..1021bf4 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -1239,9 +1239,6 @@ static void i915_driver_register(struct drm_i915_private *dev_priv) i915_debugfs_register(dev_priv); i915_setup_sysfs(dev_priv); - /* Depends on debugfs having been initialized */ - intel_uc_register(dev_priv); - /* Depends on sysfs having been initialized */ i915_perf_register(dev_priv); } else @@ -1299,7 +1296,6 @@ static void i915_driver_unregister(struct drm_i915_private *dev_priv) i915_pmu_unregister(dev_priv); i915_teardown_sysfs(dev_priv); - intel_uc_unregister(dev_priv); drm_dev_unregister(&dev_priv->drm); i915_gem_shrinker_unregister(dev_priv); diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 72a71bc..20254dd 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -423,18 +423,13 @@ static int guc_log_relay_create(struct intel_guc_log *log) DRM_ERROR("Couldn't create relay chan for GuC logging\n"); ret = -ENOMEM; - goto err; + return ret; } GEM_BUG_ON(guc_log_relay_chan->subbuf_size < subbuf_size); log->runtime.relay_chan = guc_log_relay_chan; return 0; - -err: - /* logging will be off */ - i915_modparams.guc_log_level = 0; - return ret; } static void guc_log_relay_destroy(struct intel_guc_log *log) @@ -511,7 +506,7 @@ void intel_guc_log_destroy(struct intel_guc_log *log) i915_vma_unpin_and_release(&log->vma); } -int intel_guc_log_control_get(struct intel_guc_log *log) +int intel_guc_log_level_get(struct intel_guc_log *log) { GEM_BUG_ON(!log->vma); GEM_BUG_ON(i915_modparams.guc_log_level < 0); @@ -526,11 +521,10 @@ int intel_guc_log_control_get(struct intel_guc_log *log) LOG_LEVEL_TO_ENABLED(_x) ? _x - 1 : 0; \ }) #define VERBOSITY_TO_LOG_LEVEL(x) ((x) + 1) -int intel_guc_log_control_set(struct intel_guc_log *log, u64 val) +int intel_guc_log_level_set(struct intel_guc_log *log, u64 val) { struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); - bool enabled = LOG_LEVEL_TO_ENABLED(val); int ret; BUILD_BUG_ON(GUC_LOG_VERBOSITY_MIN != 0); @@ -553,7 +547,8 @@ int intel_guc_log_control_set(struct intel_guc_log *log, u64 val) } intel_runtime_pm_get(dev_priv); - ret = guc_log_control(guc, enabled, LOG_LEVEL_TO_VERBOSITY(val)); + ret = guc_log_control(guc, LOG_LEVEL_TO_ENABLED(val), + LOG_LEVEL_TO_VERBOSITY(val)); intel_runtime_pm_put(dev_priv); if (ret) { DRM_DEBUG_DRIVER("guc_log_control action failed %d\n", ret); @@ -562,89 +557,79 @@ int intel_guc_log_control_set(struct intel_guc_log *log, u64 val) i915_modparams.guc_log_level = val; - mutex_unlock(&dev_priv->drm.struct_mutex); - - if (enabled && !guc_log_has_runtime(log)) { - ret = intel_guc_log_register(log); - if (ret) { - /* logging will remain off */ - i915_modparams.guc_log_level = 0; - goto out; - } - } else if (!enabled && guc_log_has_runtime(log)) { - intel_guc_log_unregister(log); - } - - return 0; - out_unlock: mutex_unlock(&dev_priv->drm.struct_mutex); -out: + return ret; } -int intel_guc_log_register(struct intel_guc_log *log) +int intel_guc_log_relay_open(struct intel_guc_log *log) { int ret; mutex_lock(&log->runtime.lock); - GEM_BUG_ON(guc_log_has_runtime(log)); + if (guc_log_has_runtime(log)) { + ret = -EEXIST; + goto out_unlock; + } ret = guc_log_relay_create(log); if (ret) - goto err; + goto out_unlock; ret = guc_log_map(log); if (ret) - goto err_relay; + goto out_relay; + + mutex_unlock(&log->runtime.lock); guc_flush_log_msg_enable(log_to_guc(log)); - mutex_unlock(&log->runtime.lock); + /* + * When GuC is logging without us relaying to userspace, we're ignoring + * the flush notification. This means that we need to unconditionally + * flush on relay enabling, since GuC only notifies us once. + */ + queue_work(log->runtime.flush_wq, &log->runtime.flush_work); return 0; -err_relay: +out_relay: guc_log_relay_destroy(log); -err: +out_unlock: mutex_unlock(&log->runtime.lock); return ret; } -void intel_guc_log_unregister(struct intel_guc_log *log) +void intel_guc_log_relay_flush(struct intel_guc_log *log) { struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *i915 = guc_to_i915(guc); - guc_flush_log_msg_disable(guc); - /* * Before initiating the forceful flush, wait for any pending/ongoing * flush to complete otherwise forceful flush may not actually happen. */ flush_work(&log->runtime.flush_work); - /* - * Once logging is disabled, GuC won't generate logs & send an - * interrupt. But there could be some data in the log buffer - * which is yet to be captured. So request GuC to update the log - * buffer state and then collect the left over logs. - */ intel_runtime_pm_get(i915); guc_log_flush(guc); intel_runtime_pm_put(i915); /* GuC would have updated log buffer by now, so capture it */ guc_log_capture_logs(log); +} - mutex_lock(&log->runtime.lock); +void intel_guc_log_relay_close(struct intel_guc_log *log) +{ + guc_flush_log_msg_disable(log_to_guc(log)); + flush_work(&log->runtime.flush_work); + mutex_lock(&log->runtime.lock); GEM_BUG_ON(!guc_log_has_runtime(log)); - guc_log_unmap(log); guc_log_relay_destroy(log); - mutex_unlock(&log->runtime.lock); } diff --git a/drivers/gpu/drm/i915/intel_guc_log.h b/drivers/gpu/drm/i915/intel_guc_log.h index e0ea6250..3cf911e 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.h +++ b/drivers/gpu/drm/i915/intel_guc_log.h @@ -61,11 +61,12 @@ struct intel_guc_log { void intel_guc_log_init_early(struct intel_guc_log *log); int intel_guc_log_create(struct intel_guc_log *log); -int intel_guc_log_register(struct intel_guc_log *log); -void intel_guc_log_unregister(struct intel_guc_log *log); void intel_guc_log_destroy(struct intel_guc_log *log); -int intel_guc_log_control_get(struct intel_guc_log *log); -int intel_guc_log_control_set(struct intel_guc_log *log, u64 control); +int intel_guc_log_level_get(struct intel_guc_log *log); +int intel_guc_log_level_set(struct intel_guc_log *log, u64 control_val); +int intel_guc_log_relay_open(struct intel_guc_log *log); +void intel_guc_log_relay_flush(struct intel_guc_log *log); +void intel_guc_log_relay_close(struct intel_guc_log *log); #endif diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 765b86a..9bb40cd 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -221,28 +221,6 @@ static void guc_free_load_err_log(struct intel_guc *guc) i915_gem_object_put(guc->load_err_log); } -int intel_uc_register(struct drm_i915_private *i915) -{ - int ret = 0; - - if (!USES_GUC(i915)) - return 0; - - if (i915_modparams.guc_log_level) - ret = intel_guc_log_register(&i915->guc.log); - - return ret; -} - -void intel_uc_unregister(struct drm_i915_private *i915) -{ - if (!USES_GUC(i915)) - return; - - if (i915_modparams.guc_log_level) - intel_guc_log_unregister(&i915->guc.log); -} - static int guc_enable_communication(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); diff --git a/drivers/gpu/drm/i915/intel_uc.h b/drivers/gpu/drm/i915/intel_uc.h index 0a2b413..937e611 100644 --- a/drivers/gpu/drm/i915/intel_uc.h +++ b/drivers/gpu/drm/i915/intel_uc.h @@ -30,8 +30,6 @@ void intel_uc_init_early(struct drm_i915_private *dev_priv); void intel_uc_init_mmio(struct drm_i915_private *dev_priv); -int intel_uc_register(struct drm_i915_private *dev_priv); -void intel_uc_unregister(struct drm_i915_private *dev_priv); void intel_uc_init_fw(struct drm_i915_private *dev_priv); void intel_uc_fini_fw(struct drm_i915_private *dev_priv); int intel_uc_init_misc(struct drm_i915_private *dev_priv); -- cgit v1.1 From b8299c71d4e15972ba507a9b8cdba5653cd247d5 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Mon, 19 Mar 2018 10:53:41 +0100 Subject: drm/i915/guc: Move check for fast memcpy_wc to relay creation MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We only need those fast memcpy_wc when we're using relay to read continuous GuC log. Let's prevent the user from creating a relay if we know we won't be able to keep up with GuC. v2: Adjust the return value (Michał) Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319095348.9716-6-michal.winiarski@intel.com --- drivers/gpu/drm/i915/intel_guc_log.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 20254dd..db89999 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -466,16 +466,6 @@ int intel_guc_log_create(struct intel_guc_log *log) GEM_BUG_ON(log->vma); - /* - * We require SSE 4.1 for fast reads from the GuC log buffer and - * it should be present on the chipsets supporting GuC based - * submisssions. - */ - if (WARN_ON(!i915_has_memcpy_from_wc())) { - ret = -EINVAL; - goto err; - } - vma = intel_guc_allocate_vma(guc, GUC_LOG_SIZE); if (IS_ERR(vma)) { ret = PTR_ERR(vma); @@ -574,6 +564,16 @@ int intel_guc_log_relay_open(struct intel_guc_log *log) goto out_unlock; } + /* + * We require SSE 4.1 for fast reads from the GuC log buffer and + * it should be present on the chipsets supporting GuC based + * submisssions. + */ + if (!i915_has_memcpy_from_wc()) { + ret = -ENXIO; + goto out_unlock; + } + ret = guc_log_relay_create(log); if (ret) goto out_unlock; -- cgit v1.1 From 6a96be2448a446efb1ac67974535fd4b33df3d48 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Mon, 19 Mar 2018 10:53:42 +0100 Subject: drm/i915/guc: Get rid of GuC log runtime MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Runtime is not a very good name. Let's also move counting relay overflows inside relay struct. v2: Rename things rather than remove the struct (Chris) Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319095348.9716-7-michal.winiarski@intel.com --- drivers/gpu/drm/i915/i915_debugfs.c | 4 +-- drivers/gpu/drm/i915/intel_guc.c | 15 +++++---- drivers/gpu/drm/i915/intel_guc_log.c | 64 ++++++++++++++++++------------------ drivers/gpu/drm/i915/intel_guc_log.h | 7 ++-- 4 files changed, 45 insertions(+), 45 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index e857a94..d3d4d1b 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -2347,8 +2347,8 @@ static void i915_guc_log_info(struct seq_file *m, seq_printf(m, "\tTotal flush interrupt count: %u\n", guc->log.flush_interrupt_count); - seq_printf(m, "\tCapture miss count: %u\n", - guc->log.capture_miss_count); + seq_printf(m, "\tRelay full count: %u\n", + guc->log.relay.full_count); } static void i915_guc_client_info(struct seq_file *m, diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index 3af6035..70d118b 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -87,9 +87,10 @@ int intel_guc_init_wq(struct intel_guc *guc) * or scheduled later on resume. This way the handling of work * item can be kept same between system suspend & rpm suspend. */ - guc->log.runtime.flush_wq = alloc_ordered_workqueue("i915-guc_log", - WQ_HIGHPRI | WQ_FREEZABLE); - if (!guc->log.runtime.flush_wq) { + guc->log.relay.flush_wq = + alloc_ordered_workqueue("i915-guc_log", + WQ_HIGHPRI | WQ_FREEZABLE); + if (!guc->log.relay.flush_wq) { DRM_ERROR("Couldn't allocate workqueue for GuC log\n"); return -ENOMEM; } @@ -112,7 +113,7 @@ int intel_guc_init_wq(struct intel_guc *guc) guc->preempt_wq = alloc_ordered_workqueue("i915-guc_preempt", WQ_HIGHPRI); if (!guc->preempt_wq) { - destroy_workqueue(guc->log.runtime.flush_wq); + destroy_workqueue(guc->log.relay.flush_wq); DRM_ERROR("Couldn't allocate workqueue for GuC " "preemption\n"); return -ENOMEM; @@ -130,7 +131,7 @@ void intel_guc_fini_wq(struct intel_guc *guc) USES_GUC_SUBMISSION(dev_priv)) destroy_workqueue(guc->preempt_wq); - destroy_workqueue(guc->log.runtime.flush_wq); + destroy_workqueue(guc->log.relay.flush_wq); } static int guc_shared_data_create(struct intel_guc *guc) @@ -390,8 +391,8 @@ void intel_guc_to_host_event_handler(struct intel_guc *guc) if (msg & (INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED)) { - queue_work(guc->log.runtime.flush_wq, - &guc->log.runtime.flush_work); + queue_work(guc->log.relay.flush_wq, + &guc->log.relay.flush_work); guc->log.flush_interrupt_count++; } diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index db89999..c220c28 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -171,10 +171,10 @@ static void guc_move_to_next_buf(struct intel_guc_log *log) smp_wmb(); /* All data has been written, so now move the offset of sub buffer. */ - relay_reserve(log->runtime.relay_chan, log->vma->obj->base.size); + relay_reserve(log->relay.channel, log->vma->obj->base.size); /* Switch to the next sub buffer */ - relay_flush(log->runtime.relay_chan); + relay_flush(log->relay.channel); } static void *guc_get_write_buffer(struct intel_guc_log *log) @@ -188,7 +188,7 @@ static void *guc_get_write_buffer(struct intel_guc_log *log) * done without using relay_reserve() along with relay_write(). So its * better to use relay_reserve() alone. */ - return relay_reserve(log->runtime.relay_chan, 0); + return relay_reserve(log->relay.channel, 0); } static bool guc_check_log_buf_overflow(struct intel_guc_log *log, @@ -239,13 +239,13 @@ static void guc_read_update_log_buffer(struct intel_guc_log *log) void *src_data, *dst_data; bool new_overflow; - mutex_lock(&log->runtime.lock); + mutex_lock(&log->relay.lock); - if (WARN_ON(!log->runtime.buf_addr)) + if (WARN_ON(!log->relay.buf_addr)) goto out_unlock; /* Get the pointer to shared GuC log buffer */ - log_buf_state = src_data = log->runtime.buf_addr; + log_buf_state = src_data = log->relay.buf_addr; /* Get the pointer to local buffer to store the logs */ log_buf_snapshot_state = dst_data = guc_get_write_buffer(log); @@ -256,7 +256,7 @@ static void guc_read_update_log_buffer(struct intel_guc_log *log) * getting consumed by User at a slow rate. */ DRM_ERROR_RATELIMITED("no sub-buffer to capture logs\n"); - log->capture_miss_count++; + log->relay.full_count++; goto out_unlock; } @@ -330,20 +330,20 @@ static void guc_read_update_log_buffer(struct intel_guc_log *log) guc_move_to_next_buf(log); out_unlock: - mutex_unlock(&log->runtime.lock); + mutex_unlock(&log->relay.lock); } static void capture_logs_work(struct work_struct *work) { struct intel_guc_log *log = - container_of(work, struct intel_guc_log, runtime.flush_work); + container_of(work, struct intel_guc_log, relay.flush_work); guc_log_capture_logs(log); } -static bool guc_log_has_runtime(struct intel_guc_log *log) +static bool guc_log_relay_enabled(struct intel_guc_log *log) { - return log->runtime.buf_addr; + return log->relay.buf_addr; } static int guc_log_map(struct intel_guc_log *log) @@ -353,7 +353,7 @@ static int guc_log_map(struct intel_guc_log *log) void *vaddr; int ret; - lockdep_assert_held(&log->runtime.lock); + lockdep_assert_held(&log->relay.lock); if (!log->vma) return -ENODEV; @@ -375,23 +375,23 @@ static int guc_log_map(struct intel_guc_log *log) return PTR_ERR(vaddr); } - log->runtime.buf_addr = vaddr; + log->relay.buf_addr = vaddr; return 0; } static void guc_log_unmap(struct intel_guc_log *log) { - lockdep_assert_held(&log->runtime.lock); + lockdep_assert_held(&log->relay.lock); i915_gem_object_unpin_map(log->vma->obj); - log->runtime.buf_addr = NULL; + log->relay.buf_addr = NULL; } void intel_guc_log_init_early(struct intel_guc_log *log) { - mutex_init(&log->runtime.lock); - INIT_WORK(&log->runtime.flush_work, capture_logs_work); + mutex_init(&log->relay.lock); + INIT_WORK(&log->relay.flush_work, capture_logs_work); } static int guc_log_relay_create(struct intel_guc_log *log) @@ -402,7 +402,7 @@ static int guc_log_relay_create(struct intel_guc_log *log) size_t n_subbufs, subbuf_size; int ret; - lockdep_assert_held(&log->runtime.lock); + lockdep_assert_held(&log->relay.lock); /* Keep the size of sub buffers same as shared log buffer */ subbuf_size = GUC_LOG_SIZE; @@ -427,17 +427,17 @@ static int guc_log_relay_create(struct intel_guc_log *log) } GEM_BUG_ON(guc_log_relay_chan->subbuf_size < subbuf_size); - log->runtime.relay_chan = guc_log_relay_chan; + log->relay.channel = guc_log_relay_chan; return 0; } static void guc_log_relay_destroy(struct intel_guc_log *log) { - lockdep_assert_held(&log->runtime.lock); + lockdep_assert_held(&log->relay.lock); - relay_close(log->runtime.relay_chan); - log->runtime.relay_chan = NULL; + relay_close(log->relay.channel); + log->relay.channel = NULL; } static void guc_log_capture_logs(struct intel_guc_log *log) @@ -557,9 +557,9 @@ int intel_guc_log_relay_open(struct intel_guc_log *log) { int ret; - mutex_lock(&log->runtime.lock); + mutex_lock(&log->relay.lock); - if (guc_log_has_runtime(log)) { + if (guc_log_relay_enabled(log)) { ret = -EEXIST; goto out_unlock; } @@ -582,7 +582,7 @@ int intel_guc_log_relay_open(struct intel_guc_log *log) if (ret) goto out_relay; - mutex_unlock(&log->runtime.lock); + mutex_unlock(&log->relay.lock); guc_flush_log_msg_enable(log_to_guc(log)); @@ -591,14 +591,14 @@ int intel_guc_log_relay_open(struct intel_guc_log *log) * the flush notification. This means that we need to unconditionally * flush on relay enabling, since GuC only notifies us once. */ - queue_work(log->runtime.flush_wq, &log->runtime.flush_work); + queue_work(log->relay.flush_wq, &log->relay.flush_work); return 0; out_relay: guc_log_relay_destroy(log); out_unlock: - mutex_unlock(&log->runtime.lock); + mutex_unlock(&log->relay.lock); return ret; } @@ -612,7 +612,7 @@ void intel_guc_log_relay_flush(struct intel_guc_log *log) * Before initiating the forceful flush, wait for any pending/ongoing * flush to complete otherwise forceful flush may not actually happen. */ - flush_work(&log->runtime.flush_work); + flush_work(&log->relay.flush_work); intel_runtime_pm_get(i915); guc_log_flush(guc); @@ -625,11 +625,11 @@ void intel_guc_log_relay_flush(struct intel_guc_log *log) void intel_guc_log_relay_close(struct intel_guc_log *log) { guc_flush_log_msg_disable(log_to_guc(log)); - flush_work(&log->runtime.flush_work); + flush_work(&log->relay.flush_work); - mutex_lock(&log->runtime.lock); - GEM_BUG_ON(!guc_log_has_runtime(log)); + mutex_lock(&log->relay.lock); + GEM_BUG_ON(!guc_log_relay_enabled(log)); guc_log_unmap(log); guc_log_relay_destroy(log); - mutex_unlock(&log->runtime.lock); + mutex_unlock(&log->relay.lock); } diff --git a/drivers/gpu/drm/i915/intel_guc_log.h b/drivers/gpu/drm/i915/intel_guc_log.h index 3cf911e..db35e54 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.h +++ b/drivers/gpu/drm/i915/intel_guc_log.h @@ -43,16 +43,15 @@ struct intel_guc; struct intel_guc_log { u32 flags; struct i915_vma *vma; - /* The runtime stuff gets created only when GuC logging gets enabled */ struct { void *buf_addr; struct workqueue_struct *flush_wq; struct work_struct flush_work; - struct rchan *relay_chan; + struct rchan *channel; struct mutex lock; - } runtime; + u32 full_count; + } relay; /* logging related stats */ - u32 capture_miss_count; u32 flush_interrupt_count; u32 prev_overflow_count[GUC_MAX_LOG_BUFFER]; u32 total_overflow_count[GUC_MAX_LOG_BUFFER]; -- cgit v1.1 From db5579934f2fc8d916fe29355ac0c716acf1d921 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Mon, 19 Mar 2018 10:53:43 +0100 Subject: drm/i915/guc: Always print log stats in i915_guc_info when using GuC MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit While some of the content in this file is related to GuC submission only, that's not the case with log related statistics. Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319095348.9716-8-michal.winiarski@intel.com --- drivers/gpu/drm/i915/i915_debugfs.c | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index d3d4d1b..5584736 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -2330,7 +2330,7 @@ static void i915_guc_log_info(struct seq_file *m, { struct intel_guc *guc = &dev_priv->guc; - seq_puts(m, "\nGuC logging stats:\n"); + seq_puts(m, "GuC logging stats:\n"); seq_printf(m, "\tISR: flush count %10u, overflow count %10u\n", guc->log.flush_count[GUC_ISR_LOG_BUFFER], @@ -2378,14 +2378,19 @@ static int i915_guc_info(struct seq_file *m, void *data) struct drm_i915_private *dev_priv = node_to_i915(m->private); const struct intel_guc *guc = &dev_priv->guc; - if (!USES_GUC_SUBMISSION(dev_priv)) + if (!USES_GUC(dev_priv)) return -ENODEV; + i915_guc_log_info(m, dev_priv); + + if (!USES_GUC_SUBMISSION(dev_priv)) + return 0; + GEM_BUG_ON(!guc->execbuf_client); - seq_printf(m, "Doorbell map:\n"); + seq_printf(m, "\nDoorbell map:\n"); seq_printf(m, "\t%*pb\n", GUC_NUM_DOORBELLS, guc->doorbell_bitmap); - seq_printf(m, "Doorbell next cacheline: 0x%x\n\n", guc->db_cacheline); + seq_printf(m, "Doorbell next cacheline: 0x%x\n", guc->db_cacheline); seq_printf(m, "\nGuC execbuf client @ %p:\n", guc->execbuf_client); i915_guc_client_info(m, dev_priv, guc->execbuf_client); @@ -2395,8 +2400,6 @@ static int i915_guc_info(struct seq_file *m, void *data) i915_guc_client_info(m, dev_priv, guc->preempt_client); } - i915_guc_log_info(m, dev_priv); - /* Add more as required ... */ return 0; -- cgit v1.1 From 5e24e4a240770008ed46d90d6571ec27b5e2bd5b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Mon, 19 Mar 2018 10:53:44 +0100 Subject: drm/i915/guc: Don't print out relay statistics when relay is disabled MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If nobody has enabled the relay, we're not comunicating with GuC, which means that the stats don't have any meaning. Let's also remove interrupt counter and tidy the debugfs formatting. v2: Correct stats accounting (Sagar) v3: Corrected one more error in stats accounting, move relay_enabled (Sagar) Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319095348.9716-9-michal.winiarski@intel.com --- drivers/gpu/drm/i915/i915_debugfs.c | 49 +++++++++++++++++++++++------------- drivers/gpu/drm/i915/intel_guc.c | 5 +--- drivers/gpu/drm/i915/intel_guc_log.c | 26 +++++++++---------- drivers/gpu/drm/i915/intel_guc_log.h | 10 +++++--- 4 files changed, 52 insertions(+), 38 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 5584736..964ea1a 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -2325,30 +2325,45 @@ static int i915_guc_load_status_info(struct seq_file *m, void *data) return 0; } -static void i915_guc_log_info(struct seq_file *m, - struct drm_i915_private *dev_priv) +static const char * +stringify_guc_log_type(enum guc_log_buffer_type type) { - struct intel_guc *guc = &dev_priv->guc; - - seq_puts(m, "GuC logging stats:\n"); + switch (type) { + case GUC_ISR_LOG_BUFFER: + return "ISR"; + case GUC_DPC_LOG_BUFFER: + return "DPC"; + case GUC_CRASH_DUMP_LOG_BUFFER: + return "CRASH"; + default: + MISSING_CASE(type); + } - seq_printf(m, "\tISR: flush count %10u, overflow count %10u\n", - guc->log.flush_count[GUC_ISR_LOG_BUFFER], - guc->log.total_overflow_count[GUC_ISR_LOG_BUFFER]); + return ""; +} - seq_printf(m, "\tDPC: flush count %10u, overflow count %10u\n", - guc->log.flush_count[GUC_DPC_LOG_BUFFER], - guc->log.total_overflow_count[GUC_DPC_LOG_BUFFER]); +static void i915_guc_log_info(struct seq_file *m, + struct drm_i915_private *dev_priv) +{ + struct intel_guc_log *log = &dev_priv->guc.log; + enum guc_log_buffer_type type; - seq_printf(m, "\tCRASH: flush count %10u, overflow count %10u\n", - guc->log.flush_count[GUC_CRASH_DUMP_LOG_BUFFER], - guc->log.total_overflow_count[GUC_CRASH_DUMP_LOG_BUFFER]); + if (!intel_guc_log_relay_enabled(log)) { + seq_puts(m, "GuC log relay disabled\n"); + return; + } - seq_printf(m, "\tTotal flush interrupt count: %u\n", - guc->log.flush_interrupt_count); + seq_puts(m, "GuC logging stats:\n"); seq_printf(m, "\tRelay full count: %u\n", - guc->log.relay.full_count); + log->relay.full_count); + + for (type = GUC_ISR_LOG_BUFFER; type < GUC_MAX_LOG_BUFFER; type++) { + seq_printf(m, "\t%s:\tflush count %10u, overflow count %10u\n", + stringify_guc_log_type(type), + log->stats[type].flush, + log->stats[type].sampled_overflow); + } } static void i915_guc_client_info(struct seq_file *m, diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index 70d118b..eeda1aa 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -390,12 +390,9 @@ void intel_guc_to_host_event_handler(struct intel_guc *guc) spin_unlock(&guc->irq_lock); if (msg & (INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | - INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED)) { + INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED)) queue_work(guc->log.relay.flush_wq, &guc->log.relay.flush_work); - - guc->log.flush_interrupt_count++; - } } int intel_guc_sample_forcewake(struct intel_guc *guc) diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index c220c28..3180645 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -195,18 +195,18 @@ static bool guc_check_log_buf_overflow(struct intel_guc_log *log, enum guc_log_buffer_type type, unsigned int full_cnt) { - unsigned int prev_full_cnt = log->prev_overflow_count[type]; + unsigned int prev_full_cnt = log->stats[type].sampled_overflow; bool overflow = false; if (full_cnt != prev_full_cnt) { overflow = true; - log->prev_overflow_count[type] = full_cnt; - log->total_overflow_count[type] += full_cnt - prev_full_cnt; + log->stats[type].overflow = full_cnt; + log->stats[type].sampled_overflow += full_cnt - prev_full_cnt; if (full_cnt < prev_full_cnt) { /* buffer_full_cnt is a 4 bit counter */ - log->total_overflow_count[type] += 16; + log->stats[type].sampled_overflow += 16; } DRM_ERROR_RATELIMITED("GuC log buffer overflow\n"); } @@ -241,7 +241,7 @@ static void guc_read_update_log_buffer(struct intel_guc_log *log) mutex_lock(&log->relay.lock); - if (WARN_ON(!log->relay.buf_addr)) + if (WARN_ON(!intel_guc_log_relay_enabled(log))) goto out_unlock; /* Get the pointer to shared GuC log buffer */ @@ -279,7 +279,7 @@ static void guc_read_update_log_buffer(struct intel_guc_log *log) full_cnt = log_buf_state_local.buffer_full_cnt; /* Bookkeeping stuff */ - log->flush_count[type] += log_buf_state_local.flush_to_file; + log->stats[type].flush += log_buf_state_local.flush_to_file; new_overflow = guc_check_log_buf_overflow(log, type, full_cnt); /* Update the state of shared log buffer */ @@ -341,11 +341,6 @@ static void capture_logs_work(struct work_struct *work) guc_log_capture_logs(log); } -static bool guc_log_relay_enabled(struct intel_guc_log *log) -{ - return log->relay.buf_addr; -} - static int guc_log_map(struct intel_guc_log *log) { struct intel_guc *guc = log_to_guc(log); @@ -553,13 +548,18 @@ out_unlock: return ret; } +bool intel_guc_log_relay_enabled(const struct intel_guc_log *log) +{ + return log->relay.buf_addr; +} + int intel_guc_log_relay_open(struct intel_guc_log *log) { int ret; mutex_lock(&log->relay.lock); - if (guc_log_relay_enabled(log)) { + if (intel_guc_log_relay_enabled(log)) { ret = -EEXIST; goto out_unlock; } @@ -628,7 +628,7 @@ void intel_guc_log_relay_close(struct intel_guc_log *log) flush_work(&log->relay.flush_work); mutex_lock(&log->relay.lock); - GEM_BUG_ON(!guc_log_relay_enabled(log)); + GEM_BUG_ON(!intel_guc_log_relay_enabled(log)); guc_log_unmap(log); guc_log_relay_destroy(log); mutex_unlock(&log->relay.lock); diff --git a/drivers/gpu/drm/i915/intel_guc_log.h b/drivers/gpu/drm/i915/intel_guc_log.h index db35e54..9ec5703 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.h +++ b/drivers/gpu/drm/i915/intel_guc_log.h @@ -52,10 +52,11 @@ struct intel_guc_log { u32 full_count; } relay; /* logging related stats */ - u32 flush_interrupt_count; - u32 prev_overflow_count[GUC_MAX_LOG_BUFFER]; - u32 total_overflow_count[GUC_MAX_LOG_BUFFER]; - u32 flush_count[GUC_MAX_LOG_BUFFER]; + struct { + u32 sampled_overflow; + u32 overflow; + u32 flush; + } stats[GUC_MAX_LOG_BUFFER]; }; void intel_guc_log_init_early(struct intel_guc_log *log); @@ -64,6 +65,7 @@ void intel_guc_log_destroy(struct intel_guc_log *log); int intel_guc_log_level_get(struct intel_guc_log *log); int intel_guc_log_level_set(struct intel_guc_log *log, u64 control_val); +bool intel_guc_log_relay_enabled(const struct intel_guc_log *log); int intel_guc_log_relay_open(struct intel_guc_log *log); void intel_guc_log_relay_flush(struct intel_guc_log *log); void intel_guc_log_relay_close(struct intel_guc_log *log); -- cgit v1.1 From cb5d64e9f13e0dd817c3ae2dbe73c3b8c6c13f95 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Mon, 19 Mar 2018 10:53:45 +0100 Subject: drm/i915/guc: Allow user to control default GuC logging MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit While both naming and actual log enable logic in GuC interface are confusing, we can simply expose the default log as yet another log level. GuC logic aside, from i915 point of view we now have the following GuC log levels: 0 Log disabled 1 Non-verbose log 2-5 Verbose log v2: Adjust naming after rebase. v3: Fixed the log_level logic error introduced on rebase. Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319095348.9716-10-michal.winiarski@intel.com --- drivers/gpu/drm/i915/intel_guc.c | 24 +++++++++++++++--------- drivers/gpu/drm/i915/intel_guc_fwif.h | 5 +++-- drivers/gpu/drm/i915/intel_guc_log.c | 18 +++++++----------- drivers/gpu/drm/i915/intel_guc_log.h | 15 +++++++++++++++ drivers/gpu/drm/i915/intel_uc.c | 14 +++++++++----- 5 files changed, 49 insertions(+), 27 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index eeda1aa..dc16392 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -222,17 +222,23 @@ static u32 get_core_family(struct drm_i915_private *dev_priv) } } -static u32 get_log_verbosity_flags(void) +static u32 get_log_control_flags(void) { - if (i915_modparams.guc_log_level > 0) { - u32 verbosity = i915_modparams.guc_log_level - 1; + u32 level = i915_modparams.guc_log_level; + u32 flags = 0; - GEM_BUG_ON(verbosity > GUC_LOG_VERBOSITY_MAX); - return verbosity << GUC_LOG_VERBOSITY_SHIFT; - } + GEM_BUG_ON(level < 0); + + if (!GUC_LOG_LEVEL_TO_ENABLED(level)) + flags |= GUC_LOG_DEFAULT_DISABLED; + + if (!GUC_LOG_LEVEL_TO_VERBOSE(level)) + flags |= GUC_LOG_DISABLED; + else + flags |= GUC_LOG_LEVEL_TO_VERBOSITY(level) << + GUC_LOG_VERBOSITY_SHIFT; - GEM_BUG_ON(i915_modparams.enable_guc < 0); - return GUC_LOG_DISABLED; + return flags; } /* @@ -267,7 +273,7 @@ void intel_guc_init_params(struct intel_guc *guc) params[GUC_CTL_LOG_PARAMS] = guc->log.flags; - params[GUC_CTL_DEBUG] = get_log_verbosity_flags(); + params[GUC_CTL_DEBUG] = get_log_control_flags(); /* If GuC submission is enabled, set up additional parameters here */ if (USES_GUC_SUBMISSION(dev_priv)) { diff --git a/drivers/gpu/drm/i915/intel_guc_fwif.h b/drivers/gpu/drm/i915/intel_guc_fwif.h index 6a10aa6..4971685 100644 --- a/drivers/gpu/drm/i915/intel_guc_fwif.h +++ b/drivers/gpu/drm/i915/intel_guc_fwif.h @@ -127,7 +127,7 @@ #define GUC_PROFILE_ENABLED (1 << 7) #define GUC_WQ_TRACK_ENABLED (1 << 8) #define GUC_ADS_ENABLED (1 << 9) -#define GUC_DEBUG_RESERVED (1 << 10) +#define GUC_LOG_DEFAULT_DISABLED (1 << 10) #define GUC_ADS_ADDR_SHIFT 11 #define GUC_ADS_ADDR_MASK 0xfffff800 @@ -539,7 +539,8 @@ union guc_log_control { u32 logging_enabled:1; u32 reserved1:3; u32 verbosity:4; - u32 reserved2:24; + u32 default_logging:1; + u32 reserved2:23; }; u32 value; } __packed; diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 3180645..4cb422c 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -57,12 +57,14 @@ static int guc_log_flush(struct intel_guc *guc) return intel_guc_send(guc, action, ARRAY_SIZE(action)); } -static int guc_log_control(struct intel_guc *guc, bool enable, u32 verbosity) +static int guc_log_control(struct intel_guc *guc, bool enable, + bool default_logging, u32 verbosity) { union guc_log_control control_val = { { .logging_enabled = enable, .verbosity = verbosity, + .default_logging = default_logging, }, }; u32 action[] = { @@ -499,13 +501,6 @@ int intel_guc_log_level_get(struct intel_guc_log *log) return i915_modparams.guc_log_level; } -#define GUC_LOG_LEVEL_DISABLED 0 -#define LOG_LEVEL_TO_ENABLED(x) ((x) > 0) -#define LOG_LEVEL_TO_VERBOSITY(x) ({ \ - typeof(x) _x = (x); \ - LOG_LEVEL_TO_ENABLED(_x) ? _x - 1 : 0; \ -}) -#define VERBOSITY_TO_LOG_LEVEL(x) ((x) + 1) int intel_guc_log_level_set(struct intel_guc_log *log, u64 val) { struct intel_guc *guc = log_to_guc(log); @@ -521,7 +516,7 @@ int intel_guc_log_level_set(struct intel_guc_log *log, u64 val) * as indication that logging should be disabled. */ if (val < GUC_LOG_LEVEL_DISABLED || - val > VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX)) + val > GUC_VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX)) return -EINVAL; mutex_lock(&dev_priv->drm.struct_mutex); @@ -532,8 +527,9 @@ int intel_guc_log_level_set(struct intel_guc_log *log, u64 val) } intel_runtime_pm_get(dev_priv); - ret = guc_log_control(guc, LOG_LEVEL_TO_ENABLED(val), - LOG_LEVEL_TO_VERBOSITY(val)); + ret = guc_log_control(guc, GUC_LOG_LEVEL_TO_VERBOSE(val), + GUC_LOG_LEVEL_TO_ENABLED(val), + GUC_LOG_LEVEL_TO_VERBOSITY(val)); intel_runtime_pm_put(dev_priv); if (ret) { DRM_DEBUG_DRIVER("guc_log_control action failed %d\n", ret); diff --git a/drivers/gpu/drm/i915/intel_guc_log.h b/drivers/gpu/drm/i915/intel_guc_log.h index 9ec5703..af1532c 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.h +++ b/drivers/gpu/drm/i915/intel_guc_log.h @@ -40,6 +40,21 @@ struct intel_guc; #define GUC_LOG_SIZE ((1 + GUC_LOG_DPC_PAGES + 1 + GUC_LOG_ISR_PAGES + \ 1 + GUC_LOG_CRASH_PAGES + 1) << PAGE_SHIFT) +/* + * While we're using plain log level in i915, GuC controls are much more... + * "elaborate"? We have a couple of bits for verbosity, separate bit for actual + * log enabling, and separate bit for default logging - which "conveniently" + * ignores the enable bit. + */ +#define GUC_LOG_LEVEL_DISABLED 0 +#define GUC_LOG_LEVEL_TO_ENABLED(x) ((x) > 0) +#define GUC_LOG_LEVEL_TO_VERBOSE(x) ((x) > 1) +#define GUC_LOG_LEVEL_TO_VERBOSITY(x) ({ \ + typeof(x) _x = (x); \ + GUC_LOG_LEVEL_TO_VERBOSE(_x) ? _x - 2 : 0; \ +}) +#define GUC_VERBOSITY_TO_LOG_LEVEL(x) ((x) + 2) + struct intel_guc_log { u32 flags; struct i915_vma *vma; diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 9bb40cd..ad178552 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -75,7 +75,8 @@ static int __get_default_guc_log_level(struct drm_i915_private *dev_priv) if (HAS_GUC(dev_priv) && intel_uc_is_using_guc() && (IS_ENABLED(CONFIG_DRM_I915_DEBUG) || IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))) - guc_log_level = 1 + GUC_LOG_VERBOSITY_MAX; + guc_log_level = + GUC_VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX); /* Any platform specific fine-tuning can be done here */ @@ -142,17 +143,20 @@ static void sanitize_options_early(struct drm_i915_private *dev_priv) i915_modparams.guc_log_level = 0; } - if (i915_modparams.guc_log_level > 1 + GUC_LOG_VERBOSITY_MAX) { + if (i915_modparams.guc_log_level > + GUC_VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX)) { DRM_WARN("Incompatible option detected: %s=%d, %s!\n", "guc_log_level", i915_modparams.guc_log_level, "verbosity too high"); - i915_modparams.guc_log_level = 1 + GUC_LOG_VERBOSITY_MAX; + i915_modparams.guc_log_level = + GUC_VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX); } - DRM_DEBUG_DRIVER("guc_log_level=%d (enabled:%s verbosity:%d)\n", + DRM_DEBUG_DRIVER("guc_log_level=%d (enabled:%s, verbose:%s, verbosity:%d)\n", i915_modparams.guc_log_level, yesno(i915_modparams.guc_log_level), - i915_modparams.guc_log_level - 1); + yesno(GUC_LOG_LEVEL_TO_VERBOSE(i915_modparams.guc_log_level)), + GUC_LOG_LEVEL_TO_VERBOSITY(i915_modparams.guc_log_level)); /* Make sure that sanitization was done */ GEM_BUG_ON(i915_modparams.enable_guc < 0); -- cgit v1.1 From 9605d1ce7c6bcb673b6893ac12b565f1bde8f0bc Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Mon, 19 Mar 2018 10:53:46 +0100 Subject: drm/i915/guc: Default to non-verbose GuC logging MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Now that we've decoupled logging from relay, GuC log level is only controlling the GuC behavior - there shouldn't be any impact on i915 behaviour. We're only going to see a single extra interrupt when log will get half full. That, and the fact that we're seeing igt/gem_exec_nop/basic-series failing with non-verbose logging being disabled. v2: Bring back the "auto" guc_log_level, now that we fixed the log Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319095348.9716-11-michal.winiarski@intel.com --- drivers/gpu/drm/i915/i915_params.h | 2 +- drivers/gpu/drm/i915/intel_uc.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_params.h b/drivers/gpu/drm/i915/i915_params.h index 430f5f9..c963603 100644 --- a/drivers/gpu/drm/i915/i915_params.h +++ b/drivers/gpu/drm/i915/i915_params.h @@ -48,7 +48,7 @@ struct drm_printer; param(int, enable_ips, 1) \ param(int, invert_brightness, 0) \ param(int, enable_guc, 0) \ - param(int, guc_log_level, 0) \ + param(int, guc_log_level, -1) \ param(char *, guc_firmware_path, NULL) \ param(char *, huc_firmware_path, NULL) \ param(int, mmio_debug, 0) \ diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index ad178552..34e847d 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -69,7 +69,7 @@ static int __get_platform_enable_guc(struct drm_i915_private *dev_priv) static int __get_default_guc_log_level(struct drm_i915_private *dev_priv) { - int guc_log_level = 0; /* disabled */ + int guc_log_level = 1; /* non-verbose */ /* Enable if we're running on platform with GuC and debug config */ if (HAS_GUC(dev_priv) && intel_uc_is_using_guc() && -- cgit v1.1 From feb06c151fade9ecaa3dd410d792cce26e8b10de Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Mon, 19 Mar 2018 10:53:47 +0100 Subject: drm/i915/guc: Demote GuC error messages MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We're using those functions in selftests, and the callers are expected to do the error handling anyways. Let's demote all GuC actions and doorbell creation to DEBUG_DRIVER. Signed-off-by: Michał Winiarski Cc: Michal Wajdeczko Cc: Michel Thierry Cc: Chris Wilson Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319095348.9716-12-michal.winiarski@intel.com --- drivers/gpu/drm/i915/intel_guc.c | 7 ++++--- drivers/gpu/drm/i915/intel_guc_submission.c | 4 ++-- 2 files changed, 6 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index dc16392..ee5230c 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -362,9 +362,10 @@ int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len) if (ret != -ETIMEDOUT) ret = -EIO; - DRM_WARN("INTEL_GUC_SEND: Action 0x%X failed;" - " ret=%d status=0x%08X response=0x%08X\n", - action[0], ret, status, I915_READ(SOFT_SCRATCH(15))); + DRM_DEBUG_DRIVER("INTEL_GUC_SEND: Action 0x%X failed;" + " ret=%d status=0x%08X response=0x%08X\n", + action[0], ret, status, + I915_READ(SOFT_SCRATCH(15))); } intel_uncore_forcewake_put(dev_priv, guc->send_regs.fw_domains); diff --git a/drivers/gpu/drm/i915/intel_guc_submission.c b/drivers/gpu/drm/i915/intel_guc_submission.c index 33af293..207cda0 100644 --- a/drivers/gpu/drm/i915/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/intel_guc_submission.c @@ -231,8 +231,8 @@ static int create_doorbell(struct intel_guc_client *client) if (ret) { __destroy_doorbell(client); __update_doorbell_desc(client, GUC_DOORBELL_INVALID); - DRM_ERROR("Couldn't create client %u doorbell: %d\n", - client->stage_id, ret); + DRM_DEBUG_DRIVER("Couldn't create client %u doorbell: %d\n", + client->stage_id, ret); return ret; } -- cgit v1.1 From 46b863325c2f58b564463c4d6e66ee0d4f2f3244 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 19 Mar 2018 12:35:28 +0000 Subject: drm/i915: Prefer memset64() when filling the iomap As the ringbuffer may exist inside stolen memory, our access to it may be via the GTT iomap. This implies we may only have WC access for which the conventional memset() substitution of rep stos performs very badly, so switch to the rep mov[dq] variants when available. Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Matthew Auld Reviewed-by: Matthew Auld Link: https://patchwork.freedesktop.org/patch/msgid/20180319123528.28249-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/intel_ringbuffer.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 72d6167..04d9d9a 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1693,17 +1693,18 @@ u32 *intel_ring_begin(struct i915_request *rq, unsigned int num_dwords) need_wrap &= ~1; GEM_BUG_ON(need_wrap > ring->space); GEM_BUG_ON(ring->emit + need_wrap > ring->size); + GEM_BUG_ON(!IS_ALIGNED(need_wrap, sizeof(u64))); /* Fill the tail with MI_NOOP */ - memset(ring->vaddr + ring->emit, 0, need_wrap); - ring->emit = 0; + memset64(ring->vaddr + ring->emit, 0, need_wrap / sizeof(u64)); ring->space -= need_wrap; + ring->emit = 0; } GEM_BUG_ON(ring->emit > ring->size - bytes); GEM_BUG_ON(ring->space < bytes); cs = ring->vaddr + ring->emit; - GEM_DEBUG_EXEC(memset(cs, POISON_INUSE, bytes)); + GEM_DEBUG_EXEC(memset32(cs, POISON_INUSE, bytes / sizeof(*cs))); ring->emit += bytes; ring->space -= bytes; -- cgit v1.1 From 873d66fb9b1d4f4cd441f84068abb5457c60f127 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 16 Mar 2018 21:49:59 +0000 Subject: drm/i915: Trim error mask to known engines MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit For the convenience of userspace passing in an arbitrary reset mask, remove unknown engines from the set of engines that are to be reset. This means that we always follow a per-engine reset with a full-device reset when userspace writes -1 into debugfs/i915_wedged. Reported-by: Michał Winiarski Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Michał Winiarski Link: https://patchwork.freedesktop.org/patch/msgid/20180316215001.12391-1-chris@chris-wilson.co.uk Reviewed-by: Michel Thierry --- drivers/gpu/drm/i915/i915_irq.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index 828f310..44eef35 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -2985,6 +2985,7 @@ void i915_handle_error(struct drm_i915_private *dev_priv, */ intel_runtime_pm_get(dev_priv); + engine_mask &= INTEL_INFO(dev_priv)->ring_mask; i915_capture_error_state(dev_priv, engine_mask, error_msg); i915_clear_error_registers(dev_priv); -- cgit v1.1 From 91b00dff56856ea4f87aa9fb3f6c90dbb5adfc55 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Tue, 20 Mar 2018 13:50:09 +0100 Subject: drm/i915: Select STACKDEPOT for DRM_I915_DEBUG select in Kconfig isn't recursive, we need to select the stuff our selects select, too. Fix that. Signed-off-by: Daniel Vetter Reviewed-by: Chris Wilson Signed-off-by: Daniel Vetter Link: https://patchwork.freedesktop.org/patch/msgid/20180320125009.2305-1-daniel.vetter@ffwll.ch --- drivers/gpu/drm/i915/Kconfig.debug | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/i915/Kconfig.debug b/drivers/gpu/drm/i915/Kconfig.debug index 108d21f..dd5bf63 100644 --- a/drivers/gpu/drm/i915/Kconfig.debug +++ b/drivers/gpu/drm/i915/Kconfig.debug @@ -25,6 +25,7 @@ config DRM_I915_DEBUG select X86_MSR # used by igt/pm_rpm select DRM_VGEM # used by igt/prime_vgem (dmabuf interop checks) select DRM_DEBUG_MM if DRM=y + select STACKDEPOT if DRM=y # for DRM_DEBUG_MM select DRM_DEBUG_MM_SELFTEST select SW_SYNC # signaling validation framework (igt/syncobj*) select DRM_I915_SW_FENCE_DEBUG_OBJECTS -- cgit v1.1 From 26376a7e74d2deff445a72a2cfbfef084c28e4bc Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Fri, 16 Mar 2018 14:14:49 +0200 Subject: drm/i915/icl: Check for fused-off VDBOX and VEBOX instances In Gen11, the Video Decode engines (aka VDBOX, aka VCS, aka BSD) and the Video Enhancement engines (aka VEBOX, aka VECS) could be fused off. Also, each VDBOX and VEBOX has its own power well, which only exist if the related engine exists in the HW. Unfortunately, we have a Catch-22 situation going on: we need the blitter forcewake to read the register with the fuse info, but we cannot initialize the forcewake domains without knowin about the engines present in the HW. We workaround this problem by allowing the initialization of all forcewake domains and then pruning the fused off ones, as per the fuse information. Bspec: 20680 v2: We were shifting incorrectly for vebox disable (Vinay) v3: Assert mmio is ready and warn if we have attempted to initialize forcewake for fused-off engines (Paulo) v4: - Use INTEL_GEN in new code (Tvrtko) - Shorter local variable (Tvrtko, Michal) - Keep "if (!...) continue" style (Tvrtko) - No unnecessary BUG_ON (Tvrtko) - WARN_ON and cleanup if wrong mask (Tvrtko, Michal) - Use I915_READ_FW (Michal) - Use I915_MAX_VCS/VECS macros (Michal) v5: Rebased by Rodrigo fixing conflicts on top of: "drm/i915: Simplify intel_engines_init" v6: Fix v5. Remove info->num_rings. (by Oscar) v7: Rebase (Rodrigo). v8: - s/intel_device_info_fused_off_engines/ intel_device_info_init_mmio (Chris) - Make vdbox_disable & vebox_disable local variables (Chris) v9: - Move function declaration to intel_device_info.h (Michal) - Missing indent in bit fields definitions (Michal) - When RC6 is enabled by BIOS, the fuse register cannot be read until the blitter powerwell is awake. Shuffle where the fuse is read, prune the forcewake domains after the fact and change the commit message accordingly (Vinay, Sagar, Chris). v10: - Improved commit message (Sagar) - New line in header file (Sagar) - Specify the message in fw_domain_reset applies to ICL+ (Sagar) Cc: Paulo Zanoni Cc: Vinay Belgaumkar Cc: Tvrtko Ursulin Cc: Michal Wajdeczko Cc: Chris Wilson Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Signed-off-by: Rodrigo Vivi Signed-off-by: Oscar Mateo Reviewed-by: Sagar Arun Kamble Link: https://patchwork.freedesktop.org/patch/msgid/20180316121456.11577-1-mika.kuoppala@linux.intel.com [Mika: soothe checkpatch on commit msg] Signed-off-by: Mika Kuoppala --- drivers/gpu/drm/i915/i915_drv.c | 4 +++ drivers/gpu/drm/i915/i915_reg.h | 5 +++ drivers/gpu/drm/i915/intel_device_info.c | 47 +++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_device_info.h | 2 ++ drivers/gpu/drm/i915/intel_uncore.c | 56 ++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_uncore.h | 1 + 6 files changed, 115 insertions(+) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 1021bf4..ba5f150 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -1033,6 +1033,10 @@ static int i915_driver_init_mmio(struct drm_i915_private *dev_priv) intel_uncore_init(dev_priv); + intel_device_info_init_mmio(dev_priv); + + intel_uncore_prune(dev_priv); + intel_uc_init_mmio(dev_priv); ret = intel_engines_init_mmio(dev_priv); diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 1b48d50..429de0a 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -2545,6 +2545,11 @@ enum i915_power_well_id { #define GEN10_EU_DISABLE3 _MMIO(0x9140) #define GEN10_EU_DIS_SS_MASK 0xff +#define GEN11_GT_VEBOX_VDBOX_DISABLE _MMIO(0x9140) +#define GEN11_GT_VDBOX_DISABLE_MASK 0xff +#define GEN11_GT_VEBOX_DISABLE_SHIFT 16 +#define GEN11_GT_VEBOX_DISABLE_MASK (0xff << GEN11_GT_VEBOX_DISABLE_SHIFT) + #define GEN6_BSD_SLEEP_PSMI_CONTROL _MMIO(0x12050) #define GEN6_BSD_SLEEP_MSG_DISABLE (1 << 0) #define GEN6_BSD_SLEEP_FLUSH_DISABLE (1 << 2) diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c index 3dd350f..4babfc6 100644 --- a/drivers/gpu/drm/i915/intel_device_info.c +++ b/drivers/gpu/drm/i915/intel_device_info.c @@ -780,3 +780,50 @@ void intel_driver_caps_print(const struct intel_driver_caps *caps, { drm_printf(p, "scheduler: %x\n", caps->scheduler); } + +/* + * Determine which engines are fused off in our particular hardware. Since the + * fuse register is in the blitter powerwell, we need forcewake to be ready at + * this point (but later we need to prune the forcewake domains for engines that + * are indeed fused off). + */ +void intel_device_info_init_mmio(struct drm_i915_private *dev_priv) +{ + struct intel_device_info *info = mkwrite_device_info(dev_priv); + u8 vdbox_disable, vebox_disable; + u32 media_fuse; + int i; + + if (INTEL_GEN(dev_priv) < 11) + return; + + media_fuse = I915_READ(GEN11_GT_VEBOX_VDBOX_DISABLE); + + vdbox_disable = media_fuse & GEN11_GT_VDBOX_DISABLE_MASK; + vebox_disable = (media_fuse & GEN11_GT_VEBOX_DISABLE_MASK) >> + GEN11_GT_VEBOX_DISABLE_SHIFT; + + DRM_DEBUG_DRIVER("vdbox disable: %04x\n", vdbox_disable); + for (i = 0; i < I915_MAX_VCS; i++) { + if (!HAS_ENGINE(dev_priv, _VCS(i))) + continue; + + if (!(BIT(i) & vdbox_disable)) + continue; + + info->ring_mask &= ~ENGINE_MASK(_VCS(i)); + DRM_DEBUG_DRIVER("vcs%u fused off\n", i); + } + + DRM_DEBUG_DRIVER("vebox disable: %04x\n", vebox_disable); + for (i = 0; i < I915_MAX_VECS; i++) { + if (!HAS_ENGINE(dev_priv, _VECS(i))) + continue; + + if (!(BIT(i) & vebox_disable)) + continue; + + info->ring_mask &= ~ENGINE_MASK(_VECS(i)); + DRM_DEBUG_DRIVER("vecs%u fused off\n", i); + } +} diff --git a/drivers/gpu/drm/i915/intel_device_info.h b/drivers/gpu/drm/i915/intel_device_info.h index 0835752..0cbb922 100644 --- a/drivers/gpu/drm/i915/intel_device_info.h +++ b/drivers/gpu/drm/i915/intel_device_info.h @@ -247,6 +247,8 @@ void intel_device_info_dump_runtime(const struct intel_device_info *info, void intel_device_info_dump_topology(const struct sseu_dev_info *sseu, struct drm_printer *p); +void intel_device_info_init_mmio(struct drm_i915_private *dev_priv); + void intel_driver_caps_print(const struct intel_driver_caps *caps, struct drm_printer *p); diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c index 4df7c2ef..4c616d0 100644 --- a/drivers/gpu/drm/i915/intel_uncore.c +++ b/drivers/gpu/drm/i915/intel_uncore.c @@ -62,6 +62,11 @@ static inline void fw_domain_reset(struct drm_i915_private *i915, const struct intel_uncore_forcewake_domain *d) { + /* + * We don't really know if the powerwell for the forcewake domain we are + * trying to reset here does exist at this point (engines could be fused + * off in ICL+), so no waiting for acks + */ __raw_i915_write32(i915, d->reg_set, i915->uncore.fw_reset); } @@ -1353,6 +1358,23 @@ static void fw_domain_init(struct drm_i915_private *dev_priv, fw_domain_reset(dev_priv, d); } +static void fw_domain_fini(struct drm_i915_private *dev_priv, + enum forcewake_domain_id domain_id) +{ + struct intel_uncore_forcewake_domain *d; + + if (WARN_ON(domain_id >= FW_DOMAIN_ID_COUNT)) + return; + + d = &dev_priv->uncore.fw_domain[domain_id]; + + WARN_ON(d->wake_count); + WARN_ON(hrtimer_cancel(&d->timer)); + memset(d, 0, sizeof(*d)); + + dev_priv->uncore.fw_domains &= ~BIT(domain_id); +} + static void intel_uncore_fw_domains_init(struct drm_i915_private *dev_priv) { if (INTEL_GEN(dev_priv) <= 5 || intel_vgpu_active(dev_priv)) @@ -1565,6 +1587,40 @@ void intel_uncore_init(struct drm_i915_private *dev_priv) &dev_priv->uncore.pmic_bus_access_nb); } +/* + * We might have detected that some engines are fused off after we initialized + * the forcewake domains. Prune them, to make sure they only reference existing + * engines. + */ +void intel_uncore_prune(struct drm_i915_private *dev_priv) +{ + if (INTEL_GEN(dev_priv) >= 11) { + enum forcewake_domains fw_domains = dev_priv->uncore.fw_domains; + enum forcewake_domain_id domain_id; + int i; + + for (i = 0; i < I915_MAX_VCS; i++) { + domain_id = FW_DOMAIN_ID_MEDIA_VDBOX0 + i; + + if (HAS_ENGINE(dev_priv, _VCS(i))) + continue; + + if (fw_domains & BIT(domain_id)) + fw_domain_fini(dev_priv, domain_id); + } + + for (i = 0; i < I915_MAX_VECS; i++) { + domain_id = FW_DOMAIN_ID_MEDIA_VEBOX0 + i; + + if (HAS_ENGINE(dev_priv, _VECS(i))) + continue; + + if (fw_domains & BIT(domain_id)) + fw_domain_fini(dev_priv, domain_id); + } + } +} + void intel_uncore_fini(struct drm_i915_private *dev_priv) { /* Paranoia: make sure we have disabled everything before we exit. */ diff --git a/drivers/gpu/drm/i915/intel_uncore.h b/drivers/gpu/drm/i915/intel_uncore.h index dfdf444..47478d6 100644 --- a/drivers/gpu/drm/i915/intel_uncore.h +++ b/drivers/gpu/drm/i915/intel_uncore.h @@ -140,6 +140,7 @@ struct intel_uncore { void intel_uncore_sanitize(struct drm_i915_private *dev_priv); void intel_uncore_init(struct drm_i915_private *dev_priv); +void intel_uncore_prune(struct drm_i915_private *dev_priv); bool intel_uncore_unclaimed_mmio(struct drm_i915_private *dev_priv); bool intel_uncore_arm_unclaimed_mmio_detection(struct drm_i915_private *dev_priv); void intel_uncore_fini(struct drm_i915_private *dev_priv); -- cgit v1.1 From d53d5ffb9b937ae08402d5aec5c44fb9be409afb Mon Sep 17 00:00:00 2001 From: Oscar Mateo Date: Fri, 16 Mar 2018 14:14:50 +0200 Subject: drm/i915/icl: Enable the extra video decode and enhancement boxes for Icelake 11 Icelake 11 has one vebox and two vdboxes (0 and 2). Bspec: 21140 v2: Split out in two (Daniele) Cc: Daniele Ceraolo Spurio Signed-off-by: Oscar Mateo Reviewed-by: Michel Thierry Link: https://patchwork.freedesktop.org/patch/msgid/20180316121456.11577-2-mika.kuoppala@linux.intel.com Signed-off-by: Mika Kuoppala --- drivers/gpu/drm/i915/i915_pci.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/i915/i915_pci.c b/drivers/gpu/drm/i915/i915_pci.c index 062e91b..4364922 100644 --- a/drivers/gpu/drm/i915/i915_pci.c +++ b/drivers/gpu/drm/i915/i915_pci.c @@ -602,6 +602,7 @@ static const struct intel_device_info intel_icelake_11_info = { PLATFORM(INTEL_ICELAKE), .is_alpha_support = 1, .has_resource_streamer = 0, + .ring_mask = RENDER_RING | BLT_RING | VEBOX_RING | BSD_RING | BSD3_RING, }; #undef GEN -- cgit v1.1 From d3d57927995f872e5786ff6ae517a6c3e7a94d75 Mon Sep 17 00:00:00 2001 From: Kelvin Gardiner Date: Fri, 16 Mar 2018 14:14:51 +0200 Subject: drm/i915/icl: Update subslice define for ICL 11 ICL 11 has a greater number of maximum subslices. This patch reflects this. v2: GEN11 updates to MCR_SELECTOR (Oscar) v3: Copypasta error in the new defines (Lionel) Bspec: 21139 BSpec: 21108 Signed-off-by: Kelvin Gardiner Reviewed-by: Oscar Mateo (v1) Reviewed-by: Daniele Ceraolo Spurio (v1) Signed-off-by: Oscar Mateo Reviewed-by: Lionel Landwerlin Link: https://patchwork.freedesktop.org/patch/msgid/20180316121456.11577-3-mika.kuoppala@linux.intel.com Signed-off-by: Mika Kuoppala --- drivers/gpu/drm/i915/i915_reg.h | 4 ++++ drivers/gpu/drm/i915/intel_engine_cs.c | 22 ++++++++++++++++++---- drivers/gpu/drm/i915/intel_ringbuffer.h | 2 +- 3 files changed, 23 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 429de0a..699292e 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -2151,6 +2151,10 @@ enum i915_power_well_id { #define GEN8_MCR_SLICE_MASK GEN8_MCR_SLICE(3) #define GEN8_MCR_SUBSLICE(subslice) (((subslice) & 3) << 24) #define GEN8_MCR_SUBSLICE_MASK GEN8_MCR_SUBSLICE(3) +#define GEN11_MCR_SLICE(slice) (((slice) & 0xf) << 27) +#define GEN11_MCR_SLICE_MASK GEN11_MCR_SLICE(0xf) +#define GEN11_MCR_SUBSLICE(subslice) (((subslice) & 0x7) << 24) +#define GEN11_MCR_SUBSLICE_MASK GEN11_MCR_SUBSLICE(0x7) #define RING_IPEIR(base) _MMIO((base)+0x64) #define RING_IPEHR(base) _MMIO((base)+0x68) /* diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 337dfa5..de09fa4 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -800,10 +800,24 @@ static inline uint32_t read_subslice_reg(struct drm_i915_private *dev_priv, int slice, int subslice, i915_reg_t reg) { + uint32_t mcr_slice_subslice_mask; + uint32_t mcr_slice_subslice_select; uint32_t mcr; uint32_t ret; enum forcewake_domains fw_domains; + if (INTEL_GEN(dev_priv) >= 11) { + mcr_slice_subslice_mask = GEN11_MCR_SLICE_MASK | + GEN11_MCR_SUBSLICE_MASK; + mcr_slice_subslice_select = GEN11_MCR_SLICE(slice) | + GEN11_MCR_SUBSLICE(subslice); + } else { + mcr_slice_subslice_mask = GEN8_MCR_SLICE_MASK | + GEN8_MCR_SUBSLICE_MASK; + mcr_slice_subslice_select = GEN8_MCR_SLICE(slice) | + GEN8_MCR_SUBSLICE(subslice); + } + fw_domains = intel_uncore_forcewake_for_reg(dev_priv, reg, FW_REG_READ); fw_domains |= intel_uncore_forcewake_for_reg(dev_priv, @@ -818,14 +832,14 @@ read_subslice_reg(struct drm_i915_private *dev_priv, int slice, * The HW expects the slice and sublice selectors to be reset to 0 * after reading out the registers. */ - WARN_ON_ONCE(mcr & (GEN8_MCR_SLICE_MASK | GEN8_MCR_SUBSLICE_MASK)); - mcr &= ~(GEN8_MCR_SLICE_MASK | GEN8_MCR_SUBSLICE_MASK); - mcr |= GEN8_MCR_SLICE(slice) | GEN8_MCR_SUBSLICE(subslice); + WARN_ON_ONCE(mcr & mcr_slice_subslice_mask); + mcr &= ~mcr_slice_subslice_mask; + mcr |= mcr_slice_subslice_select; I915_WRITE_FW(GEN8_MCR_SELECTOR, mcr); ret = I915_READ_FW(reg); - mcr &= ~(GEN8_MCR_SLICE_MASK | GEN8_MCR_SUBSLICE_MASK); + mcr &= ~mcr_slice_subslice_mask; I915_WRITE_FW(GEN8_MCR_SELECTOR, mcr); intel_uncore_forcewake_put__locked(dev_priv, fw_domains); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 1f50727..a02c7b3 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -86,7 +86,7 @@ hangcheck_action_to_str(const enum intel_engine_hangcheck_action a) } #define I915_MAX_SLICES 3 -#define I915_MAX_SUBSLICES 3 +#define I915_MAX_SUBSLICES 8 #define instdone_slice_mask(dev_priv__) \ (INTEL_GEN(dev_priv__) == 7 ? \ -- cgit v1.1 From 03380d173a697475c747e4cd6ea2be739005dedc Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Micha=C5=82=20Winiarski?= Date: Tue, 20 Mar 2018 12:55:17 +0100 Subject: drm/i915/guc: Don't try to enable GuC logging when we're not using GuC MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When changing the default values for guc_log_level, we accidentally left the log enabled on non-guc platforms. Let's fix that. v2: Define the levels used and remove (now obsolete) comments (Chris) v3: Use "IS" rather than "TO" for booleans (Chris) Fixes: 9605d1ce7c6b ("drm/i915/guc: Default to non-verbose GuC logging") Reported-by: Chris Wilson Signed-off-by: Michał Winiarski Cc: Chris Wilson Cc: Sagar Arun Kamble Cc: Michal Wajdeczko Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180320115517.20423-1-michal.winiarski@intel.com --- drivers/gpu/drm/i915/intel_guc.c | 4 ++-- drivers/gpu/drm/i915/intel_guc_log.c | 7 +++---- drivers/gpu/drm/i915/intel_guc_log.h | 12 +++++++----- drivers/gpu/drm/i915/intel_uc.c | 23 +++++++++++------------ 4 files changed, 23 insertions(+), 23 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index ee5230c..4b7c9c6 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -229,10 +229,10 @@ static u32 get_log_control_flags(void) GEM_BUG_ON(level < 0); - if (!GUC_LOG_LEVEL_TO_ENABLED(level)) + if (!GUC_LOG_LEVEL_IS_ENABLED(level)) flags |= GUC_LOG_DEFAULT_DISABLED; - if (!GUC_LOG_LEVEL_TO_VERBOSE(level)) + if (!GUC_LOG_LEVEL_IS_VERBOSE(level)) flags |= GUC_LOG_DISABLED; else flags |= GUC_LOG_LEVEL_TO_VERBOSITY(level) << diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 4cb422c..ae9b256 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -515,8 +515,7 @@ int intel_guc_log_level_set(struct intel_guc_log *log, u64 val) * GuC is recognizing log levels starting from 0 to max, we're using 0 * as indication that logging should be disabled. */ - if (val < GUC_LOG_LEVEL_DISABLED || - val > GUC_VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX)) + if (val < GUC_LOG_LEVEL_DISABLED || val > GUC_LOG_LEVEL_MAX) return -EINVAL; mutex_lock(&dev_priv->drm.struct_mutex); @@ -527,8 +526,8 @@ int intel_guc_log_level_set(struct intel_guc_log *log, u64 val) } intel_runtime_pm_get(dev_priv); - ret = guc_log_control(guc, GUC_LOG_LEVEL_TO_VERBOSE(val), - GUC_LOG_LEVEL_TO_ENABLED(val), + ret = guc_log_control(guc, GUC_LOG_LEVEL_IS_VERBOSE(val), + GUC_LOG_LEVEL_IS_ENABLED(val), GUC_LOG_LEVEL_TO_VERBOSITY(val)); intel_runtime_pm_put(dev_priv); if (ret) { diff --git a/drivers/gpu/drm/i915/intel_guc_log.h b/drivers/gpu/drm/i915/intel_guc_log.h index af1532c..1b0d2fa 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.h +++ b/drivers/gpu/drm/i915/intel_guc_log.h @@ -46,14 +46,16 @@ struct intel_guc; * log enabling, and separate bit for default logging - which "conveniently" * ignores the enable bit. */ -#define GUC_LOG_LEVEL_DISABLED 0 -#define GUC_LOG_LEVEL_TO_ENABLED(x) ((x) > 0) -#define GUC_LOG_LEVEL_TO_VERBOSE(x) ((x) > 1) +#define GUC_LOG_LEVEL_DISABLED 0 +#define GUC_LOG_LEVEL_NON_VERBOSE 1 +#define GUC_LOG_LEVEL_IS_ENABLED(x) ((x) > GUC_LOG_LEVEL_DISABLED) +#define GUC_LOG_LEVEL_IS_VERBOSE(x) ((x) > GUC_LOG_LEVEL_NON_VERBOSE) #define GUC_LOG_LEVEL_TO_VERBOSITY(x) ({ \ typeof(x) _x = (x); \ - GUC_LOG_LEVEL_TO_VERBOSE(_x) ? _x - 2 : 0; \ + GUC_LOG_LEVEL_IS_VERBOSE(_x) ? _x - 2 : 0; \ }) -#define GUC_VERBOSITY_TO_LOG_LEVEL(x) ((x) + 2) +#define GUC_VERBOSITY_TO_LOG_LEVEL(x) ((x) + 2) +#define GUC_LOG_LEVEL_MAX GUC_VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX) struct intel_guc_log { u32 flags; diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 34e847d..2befcaf 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -69,14 +69,15 @@ static int __get_platform_enable_guc(struct drm_i915_private *dev_priv) static int __get_default_guc_log_level(struct drm_i915_private *dev_priv) { - int guc_log_level = 1; /* non-verbose */ + int guc_log_level; - /* Enable if we're running on platform with GuC and debug config */ - if (HAS_GUC(dev_priv) && intel_uc_is_using_guc() && - (IS_ENABLED(CONFIG_DRM_I915_DEBUG) || - IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))) - guc_log_level = - GUC_VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX); + if (!HAS_GUC(dev_priv) || !intel_uc_is_using_guc()) + guc_log_level = GUC_LOG_LEVEL_DISABLED; + else if (IS_ENABLED(CONFIG_DRM_I915_DEBUG) || + IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)) + guc_log_level = GUC_LOG_LEVEL_MAX; + else + guc_log_level = GUC_LOG_LEVEL_NON_VERBOSE; /* Any platform specific fine-tuning can be done here */ @@ -143,19 +144,17 @@ static void sanitize_options_early(struct drm_i915_private *dev_priv) i915_modparams.guc_log_level = 0; } - if (i915_modparams.guc_log_level > - GUC_VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX)) { + if (i915_modparams.guc_log_level > GUC_LOG_LEVEL_MAX) { DRM_WARN("Incompatible option detected: %s=%d, %s!\n", "guc_log_level", i915_modparams.guc_log_level, "verbosity too high"); - i915_modparams.guc_log_level = - GUC_VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX); + i915_modparams.guc_log_level = GUC_LOG_LEVEL_MAX; } DRM_DEBUG_DRIVER("guc_log_level=%d (enabled:%s, verbose:%s, verbosity:%d)\n", i915_modparams.guc_log_level, yesno(i915_modparams.guc_log_level), - yesno(GUC_LOG_LEVEL_TO_VERBOSE(i915_modparams.guc_log_level)), + yesno(GUC_LOG_LEVEL_IS_VERBOSE(i915_modparams.guc_log_level)), GUC_LOG_LEVEL_TO_VERBOSITY(i915_modparams.guc_log_level)); /* Make sure that sanitization was done */ -- cgit v1.1 From ca98317b89428e6ac17be0938b467ed78654dd56 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 20 Mar 2018 10:04:48 +0000 Subject: drm/i915: Specify which engines to reset following semaphore/event lockups If the GPU is stuck waiting for an event or for a semaphore, we need to reset the GPU in order to recover. We have to tell the reset routine which engines we want reset, but we were still using the old interface and declaring it as "not-fatal". Fixes: 14b730fcb8d9 ("drm/i915/tdr: Prepare error handler to accept mask of hung engines") Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Michel Thierry Reviewed-by: Michel Thierry Link: https://patchwork.freedesktop.org/patch/msgid/20180320100449.1360-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/intel_hangcheck.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_hangcheck.c b/drivers/gpu/drm/i915/intel_hangcheck.c index 42e45ae..c8ea510 100644 --- a/drivers/gpu/drm/i915/intel_hangcheck.c +++ b/drivers/gpu/drm/i915/intel_hangcheck.c @@ -246,7 +246,7 @@ engine_stuck(struct intel_engine_cs *engine, u64 acthd) */ tmp = I915_READ_CTL(engine); if (tmp & RING_WAIT) { - i915_handle_error(dev_priv, 0, + i915_handle_error(dev_priv, BIT(engine->id), "Kicking stuck wait on %s", engine->name); I915_WRITE_CTL(engine, tmp); @@ -258,7 +258,7 @@ engine_stuck(struct intel_engine_cs *engine, u64 acthd) default: return ENGINE_DEAD; case 1: - i915_handle_error(dev_priv, 0, + i915_handle_error(dev_priv, ALL_ENGINES, "Kicking stuck semaphore on %s", engine->name); I915_WRITE_CTL(engine, tmp); -- cgit v1.1 From ce80075470f6328e487389262c95af092d421ffc Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 20 Mar 2018 10:04:49 +0000 Subject: drm/i915: Add control flags to i915_handle_error() Not all callers want the GPU error to handled in the same way, so expose a control parameter. In the first instance, some callers do not want the heavyweight error capture so add a bit to request the state to be captured and saved. v2: Pass msg down to i915_reset/i915_reset_engine so that we include the reason for the reset in the dev_notice(), superseding the earlier option to not print that notice. v3: Stash the reason inside the i915->gpu_error to handover to the direct reset from the blocking waiter. Signed-off-by: Chris Wilson Cc: Jeff McGee Cc: Mika Kuoppala Cc: Michel Thierry Reviewed-by: Michel Thierry Link: https://patchwork.freedesktop.org/patch/msgid/20180320100449.1360-2-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_debugfs.c | 4 +- drivers/gpu/drm/i915/i915_drv.c | 17 ++++---- drivers/gpu/drm/i915/i915_drv.h | 10 ++--- drivers/gpu/drm/i915/i915_gpu_error.h | 3 ++ drivers/gpu/drm/i915/i915_irq.c | 55 ++++++++++++++---------- drivers/gpu/drm/i915/i915_request.c | 2 +- drivers/gpu/drm/i915/intel_hangcheck.c | 13 +++--- drivers/gpu/drm/i915/selftests/intel_hangcheck.c | 13 +++--- 8 files changed, 62 insertions(+), 55 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 964ea1a..7816cd5 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -4011,8 +4011,8 @@ i915_wedged_set(void *data, u64 val) engine->hangcheck.stalled = true; } - i915_handle_error(i915, val, "Manually set wedged engine mask = %llx", - val); + i915_handle_error(i915, val, I915_ERROR_CAPTURE, + "Manually set wedged engine mask = %llx", val); wait_on_bit(&i915->gpu_error.flags, I915_RESET_HANDOFF, diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index ba5f150..3f637ab 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -1873,7 +1873,6 @@ static int i915_resume_switcheroo(struct drm_device *dev) /** * i915_reset - reset chip after a hang * @i915: #drm_i915_private to reset - * @flags: Instructions * * Reset the chip. Useful if a hang is detected. Marks the device as wedged * on failure. @@ -1888,7 +1887,7 @@ static int i915_resume_switcheroo(struct drm_device *dev) * - re-init interrupt state * - re-init display */ -void i915_reset(struct drm_i915_private *i915, unsigned int flags) +void i915_reset(struct drm_i915_private *i915) { struct i915_gpu_error *error = &i915->gpu_error; int ret; @@ -1905,8 +1904,9 @@ void i915_reset(struct drm_i915_private *i915, unsigned int flags) if (!i915_gem_unset_wedged(i915)) goto wakeup; - if (!(flags & I915_RESET_QUIET)) - dev_notice(i915->drm.dev, "Resetting chip after gpu hang\n"); + if (error->reason) + dev_notice(i915->drm.dev, + "Resetting chip for %s\n", error->reason); error->reset_count++; disable_irq(i915->drm.irq); @@ -2007,7 +2007,7 @@ static inline int intel_gt_reset_engine(struct drm_i915_private *dev_priv, /** * i915_reset_engine - reset GPU engine to recover from a hang * @engine: engine to reset - * @flags: options + * @msg: reason for GPU reset; or NULL for no dev_notice() * * Reset a specific GPU engine. Useful if a hang is detected. * Returns zero on successful reset or otherwise an error code. @@ -2017,7 +2017,7 @@ static inline int intel_gt_reset_engine(struct drm_i915_private *dev_priv, * - reset engine (which will force the engine to idle) * - re-init/configure engine */ -int i915_reset_engine(struct intel_engine_cs *engine, unsigned int flags) +int i915_reset_engine(struct intel_engine_cs *engine, const char *msg) { struct i915_gpu_error *error = &engine->i915->gpu_error; struct i915_request *active_request; @@ -2032,10 +2032,9 @@ int i915_reset_engine(struct intel_engine_cs *engine, unsigned int flags) goto out; } - if (!(flags & I915_RESET_QUIET)) { + if (msg) dev_notice(engine->i915->drm.dev, - "Resetting %s after gpu hang\n", engine->name); - } + "Resetting %s for %s\n", engine->name, msg); error->reset_engine_count[engine->id]++; if (!engine->i915->guc.execbuf_client) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index e27ba8f..c9c3b2b 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -2700,10 +2700,8 @@ extern void i915_driver_unload(struct drm_device *dev); extern int intel_gpu_reset(struct drm_i915_private *dev_priv, u32 engine_mask); extern bool intel_has_gpu_reset(struct drm_i915_private *dev_priv); -#define I915_RESET_QUIET BIT(0) -extern void i915_reset(struct drm_i915_private *i915, unsigned int flags); -extern int i915_reset_engine(struct intel_engine_cs *engine, - unsigned int flags); +extern void i915_reset(struct drm_i915_private *i915); +extern int i915_reset_engine(struct intel_engine_cs *engine, const char *msg); extern bool intel_has_reset_engine(struct drm_i915_private *dev_priv); extern int intel_reset_guc(struct drm_i915_private *dev_priv); @@ -2751,10 +2749,12 @@ static inline void i915_queue_hangcheck(struct drm_i915_private *dev_priv) &dev_priv->gpu_error.hangcheck_work, delay); } -__printf(3, 4) +__printf(4, 5) void i915_handle_error(struct drm_i915_private *dev_priv, u32 engine_mask, + unsigned long flags, const char *fmt, ...); +#define I915_ERROR_CAPTURE BIT(0) extern void intel_irq_init(struct drm_i915_private *dev_priv); extern void intel_irq_fini(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/i915_gpu_error.h b/drivers/gpu/drm/i915/i915_gpu_error.h index ebbdf37..ac57606 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.h +++ b/drivers/gpu/drm/i915/i915_gpu_error.h @@ -269,6 +269,9 @@ struct i915_gpu_error { /** Number of times an engine has been reset */ u32 reset_engine_count[I915_NUM_ENGINES]; + /** Reason for the current *global* reset */ + const char *reason; + /** * Waitqueue to signal when a hang is detected. Used to for waiters * to release the struct_mutex for the reset to procede. diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index 44eef35..fa73107 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -2877,15 +2877,10 @@ static irqreturn_t gen11_irq_handler(int irq, void *arg) return IRQ_HANDLED; } -/** - * i915_reset_device - do process context error handling work - * @dev_priv: i915 device private - * - * Fire an error uevent so userspace can see that a hang or error - * was detected. - */ -static void i915_reset_device(struct drm_i915_private *dev_priv) +static void i915_reset_device(struct drm_i915_private *dev_priv, + const char *msg) { + struct i915_gpu_error *error = &dev_priv->gpu_error; struct kobject *kobj = &dev_priv->drm.primary->kdev->kobj; char *error_event[] = { I915_ERROR_UEVENT "=1", NULL }; char *reset_event[] = { I915_RESET_UEVENT "=1", NULL }; @@ -2901,29 +2896,32 @@ static void i915_reset_device(struct drm_i915_private *dev_priv) i915_wedge_on_timeout(&w, dev_priv, 5*HZ) { intel_prepare_reset(dev_priv); + error->reason = msg; + /* Signal that locked waiters should reset the GPU */ - set_bit(I915_RESET_HANDOFF, &dev_priv->gpu_error.flags); - wake_up_all(&dev_priv->gpu_error.wait_queue); + set_bit(I915_RESET_HANDOFF, &error->flags); + wake_up_all(&error->wait_queue); /* Wait for anyone holding the lock to wakeup, without * blocking indefinitely on struct_mutex. */ do { if (mutex_trylock(&dev_priv->drm.struct_mutex)) { - i915_reset(dev_priv, 0); + i915_reset(dev_priv); mutex_unlock(&dev_priv->drm.struct_mutex); } - } while (wait_on_bit_timeout(&dev_priv->gpu_error.flags, + } while (wait_on_bit_timeout(&error->flags, I915_RESET_HANDOFF, TASK_UNINTERRUPTIBLE, 1)); + error->reason = NULL; + intel_finish_reset(dev_priv); } - if (!test_bit(I915_WEDGED, &dev_priv->gpu_error.flags)) - kobject_uevent_env(kobj, - KOBJ_CHANGE, reset_done_event); + if (!test_bit(I915_WEDGED, &error->flags)) + kobject_uevent_env(kobj, KOBJ_CHANGE, reset_done_event); } static void i915_clear_error_registers(struct drm_i915_private *dev_priv) @@ -2955,6 +2953,7 @@ static void i915_clear_error_registers(struct drm_i915_private *dev_priv) * i915_handle_error - handle a gpu error * @dev_priv: i915 device private * @engine_mask: mask representing engines that are hung + * @flags: control flags * @fmt: Error message format string * * Do some basic checking of register state at error time and @@ -2965,16 +2964,23 @@ static void i915_clear_error_registers(struct drm_i915_private *dev_priv) */ void i915_handle_error(struct drm_i915_private *dev_priv, u32 engine_mask, + unsigned long flags, const char *fmt, ...) { struct intel_engine_cs *engine; unsigned int tmp; - va_list args; char error_msg[80]; + char *msg = NULL; - va_start(args, fmt); - vscnprintf(error_msg, sizeof(error_msg), fmt, args); - va_end(args); + if (fmt) { + va_list args; + + va_start(args, fmt); + vscnprintf(error_msg, sizeof(error_msg), fmt, args); + va_end(args); + + msg = error_msg; + } /* * In most cases it's guaranteed that we get here with an RPM @@ -2986,8 +2992,11 @@ void i915_handle_error(struct drm_i915_private *dev_priv, intel_runtime_pm_get(dev_priv); engine_mask &= INTEL_INFO(dev_priv)->ring_mask; - i915_capture_error_state(dev_priv, engine_mask, error_msg); - i915_clear_error_registers(dev_priv); + + if (flags & I915_ERROR_CAPTURE) { + i915_capture_error_state(dev_priv, engine_mask, msg); + i915_clear_error_registers(dev_priv); + } /* * Try engine reset when available. We fall back to full reset if @@ -3000,7 +3009,7 @@ void i915_handle_error(struct drm_i915_private *dev_priv, &dev_priv->gpu_error.flags)) continue; - if (i915_reset_engine(engine, 0) == 0) + if (i915_reset_engine(engine, msg) == 0) engine_mask &= ~intel_engine_flag(engine); clear_bit(I915_RESET_ENGINE + engine->id, @@ -3030,7 +3039,7 @@ void i915_handle_error(struct drm_i915_private *dev_priv, TASK_UNINTERRUPTIBLE); } - i915_reset_device(dev_priv); + i915_reset_device(dev_priv, msg); for_each_engine(engine, dev_priv, tmp) { clear_bit(I915_RESET_ENGINE + engine->id, diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index 43c7134..2325886 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -1229,7 +1229,7 @@ static bool __i915_wait_request_check_and_reset(struct i915_request *request) return false; __set_current_state(TASK_RUNNING); - i915_reset(request->i915, 0); + i915_reset(request->i915); return true; } diff --git a/drivers/gpu/drm/i915/intel_hangcheck.c b/drivers/gpu/drm/i915/intel_hangcheck.c index c8ea510..fd0ffb8 100644 --- a/drivers/gpu/drm/i915/intel_hangcheck.c +++ b/drivers/gpu/drm/i915/intel_hangcheck.c @@ -246,9 +246,8 @@ engine_stuck(struct intel_engine_cs *engine, u64 acthd) */ tmp = I915_READ_CTL(engine); if (tmp & RING_WAIT) { - i915_handle_error(dev_priv, BIT(engine->id), - "Kicking stuck wait on %s", - engine->name); + i915_handle_error(dev_priv, BIT(engine->id), 0, + "stuck wait on %s", engine->name); I915_WRITE_CTL(engine, tmp); return ENGINE_WAIT_KICK; } @@ -258,8 +257,8 @@ engine_stuck(struct intel_engine_cs *engine, u64 acthd) default: return ENGINE_DEAD; case 1: - i915_handle_error(dev_priv, ALL_ENGINES, - "Kicking stuck semaphore on %s", + i915_handle_error(dev_priv, ALL_ENGINES, 0, + "stuck semaphore on %s", engine->name); I915_WRITE_CTL(engine, tmp); return ENGINE_WAIT_KICK; @@ -386,13 +385,13 @@ static void hangcheck_declare_hang(struct drm_i915_private *i915, if (stuck != hung) hung &= ~stuck; len = scnprintf(msg, sizeof(msg), - "%s on ", stuck == hung ? "No progress" : "Hang"); + "%s on ", stuck == hung ? "no progress" : "hang"); for_each_engine_masked(engine, i915, hung, tmp) len += scnprintf(msg + len, sizeof(msg) - len, "%s, ", engine->name); msg[len-2] = '\0'; - return i915_handle_error(i915, hung, "%s", msg); + return i915_handle_error(i915, hung, I915_ERROR_CAPTURE, "%s", msg); } /* diff --git a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c index df7898c..4372826 100644 --- a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c +++ b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c @@ -433,7 +433,7 @@ static int igt_global_reset(void *arg) mutex_lock(&i915->drm.struct_mutex); reset_count = i915_reset_count(&i915->gpu_error); - i915_reset(i915, I915_RESET_QUIET); + i915_reset(i915); if (i915_reset_count(&i915->gpu_error) == reset_count) { pr_err("No GPU reset recorded!\n"); @@ -518,7 +518,7 @@ static int __igt_reset_engine(struct drm_i915_private *i915, bool active) engine->hangcheck.seqno = intel_engine_get_seqno(engine); - err = i915_reset_engine(engine, I915_RESET_QUIET); + err = i915_reset_engine(engine, NULL); if (err) { pr_err("i915_reset_engine failed\n"); break; @@ -725,7 +725,7 @@ static int __igt_reset_engine_others(struct drm_i915_private *i915, engine->hangcheck.seqno = intel_engine_get_seqno(engine); - err = i915_reset_engine(engine, I915_RESET_QUIET); + err = i915_reset_engine(engine, NULL); if (err) { pr_err("i915_reset_engine(%s:%s) failed, err=%d\n", engine->name, active ? "active" : "idle", err); @@ -865,7 +865,6 @@ static int igt_wait_reset(void *arg) __func__, rq->fence.seqno, hws_seqno(&h, rq)); intel_engine_dump(rq->engine, &p, "%s\n", rq->engine->name); - i915_reset(i915, 0); i915_gem_set_wedged(i915); err = -EIO; @@ -962,7 +961,6 @@ static int igt_reset_queue(void *arg) i915_request_put(rq); i915_request_put(prev); - i915_reset(i915, 0); i915_gem_set_wedged(i915); err = -EIO; @@ -971,7 +969,7 @@ static int igt_reset_queue(void *arg) reset_count = fake_hangcheck(prev); - i915_reset(i915, I915_RESET_QUIET); + i915_reset(i915); GEM_BUG_ON(test_bit(I915_RESET_HANDOFF, &i915->gpu_error.flags)); @@ -1069,7 +1067,6 @@ static int igt_handle_error(void *arg) __func__, rq->fence.seqno, hws_seqno(&h, rq)); intel_engine_dump(rq->engine, &p, "%s\n", rq->engine->name); - i915_reset(i915, 0); i915_gem_set_wedged(i915); err = -EIO; @@ -1084,7 +1081,7 @@ static int igt_handle_error(void *arg) engine->hangcheck.stalled = true; engine->hangcheck.seqno = intel_engine_get_seqno(engine); - i915_handle_error(i915, intel_engine_flag(engine), "%s", __func__); + i915_handle_error(i915, intel_engine_flag(engine), 0, NULL); xchg(&i915->gpu_error.first_error, error); -- cgit v1.1 From 8b5eb5e2b5d2ddf9185e55669f22ea87d28f4e90 Mon Sep 17 00:00:00 2001 From: Kelvin Gardiner Date: Tue, 20 Mar 2018 12:45:21 -0700 Subject: drm/i915/icl: Added ICL 11 slice, subslice and EU fuse detection This patch adds support to detect ICL, slice, subslice and EU fuse settings. Add addresses for ICL 11 slice, subslice and EU fuses registers. These register addresses are the same as previous platforms but the format and / or the meaning of the information is different. Therefore Gen11 defines for these registers are added. Bspec: 9731 Bspec: 20643 Bspec: 20673 v2: Update fusing information storage after introducing the new query uAPI (Lionel) v3 (Oscar): - The maximum number of slices in ICL 11 is 1 - The subslice disable fuse can potentially store information in all bits - GEN_MAX_SUBSLICES has to be increased to 8 - Don't trust the slice enabled fuse outside the max number of expected slices - Indentation fix and some reordering and renaming of local variables v4: Use single space after Cc tag Cc: Mika Kuoppala Signed-off-by: Kelvin Gardiner Signed-off-by: Lionel Landwerlin Signed-off-by: Oscar Mateo Reviewed-by: Lionel Landwerlin Signed-off-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/1521575121-9577-1-git-send-email-oscar.mateo@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 8 ++++++ drivers/gpu/drm/i915/intel_device_info.c | 43 +++++++++++++++++++++++++++++++- drivers/gpu/drm/i915/intel_device_info.h | 2 +- 3 files changed, 51 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 699292e..bac3e92 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -2554,6 +2554,14 @@ enum i915_power_well_id { #define GEN11_GT_VEBOX_DISABLE_SHIFT 16 #define GEN11_GT_VEBOX_DISABLE_MASK (0xff << GEN11_GT_VEBOX_DISABLE_SHIFT) +#define GEN11_EU_DISABLE _MMIO(0x9134) +#define GEN11_EU_DIS_MASK 0xFF + +#define GEN11_GT_SLICE_ENABLE _MMIO(0x9138) +#define GEN11_GT_S_ENA_MASK 0xFF + +#define GEN11_GT_SUBSLICE_DISABLE _MMIO(0x913C) + #define GEN6_BSD_SLEEP_PSMI_CONTROL _MMIO(0x12050) #define GEN6_BSD_SLEEP_MSG_DISABLE (1 << 0) #define GEN6_BSD_SLEEP_FLUSH_DISABLE (1 << 2) diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c index 4babfc6..a504281 100644 --- a/drivers/gpu/drm/i915/intel_device_info.c +++ b/drivers/gpu/drm/i915/intel_device_info.c @@ -158,6 +158,45 @@ static u16 compute_eu_total(const struct sseu_dev_info *sseu) return total; } +static void gen11_sseu_info_init(struct drm_i915_private *dev_priv) +{ + struct sseu_dev_info *sseu = &mkwrite_device_info(dev_priv)->sseu; + u8 s_en; + u32 ss_en, ss_en_mask; + u8 eu_en; + int s; + + sseu->max_slices = 1; + sseu->max_subslices = 8; + sseu->max_eus_per_subslice = 8; + + s_en = I915_READ(GEN11_GT_SLICE_ENABLE) & GEN11_GT_S_ENA_MASK; + ss_en = ~I915_READ(GEN11_GT_SUBSLICE_DISABLE); + ss_en_mask = BIT(sseu->max_subslices) - 1; + eu_en = ~(I915_READ(GEN11_EU_DISABLE) & GEN11_EU_DIS_MASK); + + for (s = 0; s < sseu->max_slices; s++) { + if (s_en & BIT(s)) { + int ss_idx = sseu->max_subslices * s; + int ss; + + sseu->slice_mask |= BIT(s); + sseu->subslice_mask[s] = (ss_en >> ss_idx) & ss_en_mask; + for (ss = 0; ss < sseu->max_subslices; ss++) { + if (sseu->subslice_mask[s] & BIT(ss)) + sseu_set_eus(sseu, s, ss, eu_en); + } + } + } + sseu->eu_per_subslice = hweight8(eu_en); + sseu->eu_total = compute_eu_total(sseu); + + /* ICL has no power gating restrictions. */ + sseu->has_slice_pg = 1; + sseu->has_subslice_pg = 1; + sseu->has_eu_pg = 1; +} + static void gen10_sseu_info_init(struct drm_i915_private *dev_priv) { struct sseu_dev_info *sseu = &mkwrite_device_info(dev_priv)->sseu; @@ -768,8 +807,10 @@ void intel_device_info_runtime_init(struct intel_device_info *info) broadwell_sseu_info_init(dev_priv); else if (INTEL_GEN(dev_priv) == 9) gen9_sseu_info_init(dev_priv); - else if (INTEL_GEN(dev_priv) >= 10) + else if (INTEL_GEN(dev_priv) == 10) gen10_sseu_info_init(dev_priv); + else if (INTEL_INFO(dev_priv)->gen >= 11) + gen11_sseu_info_init(dev_priv); /* Initialize command stream timestamp frequency */ info->cs_timestamp_frequency_khz = read_timestamp_frequency(dev_priv); diff --git a/drivers/gpu/drm/i915/intel_device_info.h b/drivers/gpu/drm/i915/intel_device_info.h index 0cbb922..933e316 100644 --- a/drivers/gpu/drm/i915/intel_device_info.h +++ b/drivers/gpu/drm/i915/intel_device_info.h @@ -114,7 +114,7 @@ enum intel_platform { func(has_ipc); #define GEN_MAX_SLICES (6) /* CNL upper bound */ -#define GEN_MAX_SUBSLICES (7) +#define GEN_MAX_SUBSLICES (8) /* ICL upper bound */ struct sseu_dev_info { u8 slice_mask; -- cgit v1.1 From fa265275910f9d2396f8656317196c830878bd40 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Wed, 14 Mar 2018 20:04:29 +0000 Subject: drm/i915/huc: Check HuC status in dedicated function We try to keep all HuC related code in dedicated file. There is no need to peek HuC register directly during handling getparam ioctl. Signed-off-by: Michal Wajdeczko Cc: Michel Thierry Cc: Rodrigo Vivi Cc: Anusha Srivatsa Reviewed-by: Michel Thierry Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180314200429.40132-1-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/i915_drv.c | 6 +++--- drivers/gpu/drm/i915/intel_huc.c | 25 +++++++++++++++++++++++++ drivers/gpu/drm/i915/intel_huc.h | 1 + 3 files changed, 29 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 3f637ab..a7d3275 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -377,9 +377,9 @@ static int i915_getparam_ioctl(struct drm_device *dev, void *data, value = INTEL_INFO(dev_priv)->sseu.min_eu_in_pool; break; case I915_PARAM_HUC_STATUS: - intel_runtime_pm_get(dev_priv); - value = I915_READ(HUC_STATUS2) & HUC_FW_VERIFIED; - intel_runtime_pm_put(dev_priv); + value = intel_huc_check_status(&dev_priv->huc); + if (value < 0) + return value; break; case I915_PARAM_MMAP_GTT_VERSION: /* Though we've started our numbering from 1, and so class all diff --git a/drivers/gpu/drm/i915/intel_huc.c b/drivers/gpu/drm/i915/intel_huc.c index 1d6c47b..2912852 100644 --- a/drivers/gpu/drm/i915/intel_huc.c +++ b/drivers/gpu/drm/i915/intel_huc.c @@ -92,3 +92,28 @@ fail: DRM_ERROR("HuC: Authentication failed %d\n", ret); return ret; } + +/** + * intel_huc_check_status() - check HuC status + * @huc: intel_huc structure + * + * This function reads status register to verify if HuC + * firmware was successfully loaded. + * + * Returns positive value if HuC firmware is loaded and verified + * and -ENODEV if HuC is not present. + */ +int intel_huc_check_status(struct intel_huc *huc) +{ + struct drm_i915_private *dev_priv = huc_to_i915(huc); + u32 status; + + if (!HAS_HUC(dev_priv)) + return -ENODEV; + + intel_runtime_pm_get(dev_priv); + status = I915_READ(HUC_STATUS2) & HUC_FW_VERIFIED; + intel_runtime_pm_put(dev_priv); + + return status; +} diff --git a/drivers/gpu/drm/i915/intel_huc.h b/drivers/gpu/drm/i915/intel_huc.h index b185850..aa85490 100644 --- a/drivers/gpu/drm/i915/intel_huc.h +++ b/drivers/gpu/drm/i915/intel_huc.h @@ -37,6 +37,7 @@ struct intel_huc { void intel_huc_init_early(struct intel_huc *huc); int intel_huc_auth(struct intel_huc *huc); +int intel_huc_check_status(struct intel_huc *huc); static inline int intel_huc_sanitize(struct intel_huc *huc) { -- cgit v1.1 From 7beae44d7b295323d8416526fd612ee166851baf Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Tue, 20 Mar 2018 18:14:17 +0000 Subject: drm/i915/guc: Unify naming of private GuC action functions MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We should avoid using guc_log prefix for functions that don't operate on GuC log, but rather request action from the GuC. Better to use guc_action prefix. v2: rebase + naming compromise v3: rebase Signed-off-by: Michal Wajdeczko Cc: Michal Winiarski Cc: Sagar Arun Kamble Reviewed-by: Michał Winiarski Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180320181419.35576-1-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc_log.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index ae9b256..957f7ed 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -38,7 +38,7 @@ static void guc_log_capture_logs(struct intel_guc_log *log); * registers value. */ -static int guc_log_flush_complete(struct intel_guc *guc) +static int guc_action_flush_log_complete(struct intel_guc *guc) { u32 action[] = { INTEL_GUC_ACTION_LOG_BUFFER_FILE_FLUSH_COMPLETE @@ -47,7 +47,7 @@ static int guc_log_flush_complete(struct intel_guc *guc) return intel_guc_send(guc, action, ARRAY_SIZE(action)); } -static int guc_log_flush(struct intel_guc *guc) +static int guc_action_flush_log(struct intel_guc *guc) { u32 action[] = { INTEL_GUC_ACTION_FORCE_LOG_BUFFER_FLUSH, @@ -57,8 +57,8 @@ static int guc_log_flush(struct intel_guc *guc) return intel_guc_send(guc, action, ARRAY_SIZE(action)); } -static int guc_log_control(struct intel_guc *guc, bool enable, - bool default_logging, u32 verbosity) +static int guc_action_control_log(struct intel_guc *guc, bool enable, + bool default_logging, u32 verbosity) { union guc_log_control control_val = { { @@ -449,7 +449,7 @@ static void guc_log_capture_logs(struct intel_guc_log *log) * time, so get/put should be really quick. */ intel_runtime_pm_get(dev_priv); - guc_log_flush_complete(guc); + guc_action_flush_log_complete(guc); intel_runtime_pm_put(dev_priv); } @@ -526,9 +526,9 @@ int intel_guc_log_level_set(struct intel_guc_log *log, u64 val) } intel_runtime_pm_get(dev_priv); - ret = guc_log_control(guc, GUC_LOG_LEVEL_IS_VERBOSE(val), - GUC_LOG_LEVEL_IS_ENABLED(val), - GUC_LOG_LEVEL_TO_VERBOSITY(val)); + ret = guc_action_control_log(guc, GUC_LOG_LEVEL_IS_VERBOSE(val), + GUC_LOG_LEVEL_IS_ENABLED(val), + GUC_LOG_LEVEL_TO_VERBOSITY(val)); intel_runtime_pm_put(dev_priv); if (ret) { DRM_DEBUG_DRIVER("guc_log_control action failed %d\n", ret); @@ -610,7 +610,7 @@ void intel_guc_log_relay_flush(struct intel_guc_log *log) flush_work(&log->relay.flush_work); intel_runtime_pm_get(i915); - guc_log_flush(guc); + guc_action_flush_log(guc); intel_runtime_pm_put(i915); /* GuC would have updated log buffer by now, so capture it */ -- cgit v1.1 From 154374c331b01acbaf6a6957a7f9e65192f6a459 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Tue, 20 Mar 2018 18:14:18 +0000 Subject: drm/i915/guc: Drop union guc_log_control MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Usually we use shift/mask macros for bit field definitions. Union guc_log_control was not following that pattern. Additional bonus: add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-25 (-25) Function old new delta intel_guc_log_level_set 388 363 -25 v2: prevent out-of-range verbosity (MichalWi) Signed-off-by: Michal Wajdeczko Cc: Michal Winiarski Cc: Sagar Arun Kamble Reviewed-by: MichaĹ Winiarski Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180320181419.35576-2-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc_fwif.h | 16 +++++----------- drivers/gpu/drm/i915/intel_guc_log.c | 13 +++++-------- 2 files changed, 10 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc_fwif.h b/drivers/gpu/drm/i915/intel_guc_fwif.h index 4971685..72941bd 100644 --- a/drivers/gpu/drm/i915/intel_guc_fwif.h +++ b/drivers/gpu/drm/i915/intel_guc_fwif.h @@ -534,17 +534,6 @@ struct guc_log_buffer_state { u32 version; } __packed; -union guc_log_control { - struct { - u32 logging_enabled:1; - u32 reserved1:3; - u32 verbosity:4; - u32 default_logging:1; - u32 reserved2:23; - }; - u32 value; -} __packed; - struct guc_ctx_report { u32 report_return_status; u32 reserved1[64]; @@ -603,6 +592,11 @@ enum intel_guc_report_status { INTEL_GUC_REPORT_STATUS_COMPLETE = 0x4, }; +#define GUC_LOG_CONTROL_LOGGING_ENABLED (1 << 0) +#define GUC_LOG_CONTROL_VERBOSITY_SHIFT 4 +#define GUC_LOG_CONTROL_VERBOSITY_MASK (0xF << GUC_LOG_CONTROL_VERBOSITY_SHIFT) +#define GUC_LOG_CONTROL_DEFAULT_LOGGING (1 << 8) + /* * The GuC sends its response to a command by overwriting the * command in SS0. The response is distinguishable from a command diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 957f7ed..188d390 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -60,18 +60,15 @@ static int guc_action_flush_log(struct intel_guc *guc) static int guc_action_control_log(struct intel_guc *guc, bool enable, bool default_logging, u32 verbosity) { - union guc_log_control control_val = { - { - .logging_enabled = enable, - .verbosity = verbosity, - .default_logging = default_logging, - }, - }; u32 action[] = { INTEL_GUC_ACTION_UK_LOG_ENABLE_LOGGING, - control_val.value + (enable ? GUC_LOG_CONTROL_LOGGING_ENABLED : 0) | + (verbosity << GUC_LOG_CONTROL_VERBOSITY_SHIFT) | + (default_logging ? GUC_LOG_CONTROL_DEFAULT_LOGGING : 0) }; + GEM_BUG_ON(verbosity > GUC_LOG_VERBOSITY_MAX); + return intel_guc_send(guc, action, ARRAY_SIZE(action)); } -- cgit v1.1 From bc598425eb18d40a18f18d67f7c460189c43a3af Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Tue, 20 Mar 2018 18:14:19 +0000 Subject: drm/i915/guc: Move enable/disable msg functions to GuC header MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit While today we are modifying GuC enabled msg mask only in GuC log, this code should be defined as generic GuC to allow future code reuse. Signed-off-by: Michal Wajdeczko Cc: Michal Winiarski Cc: Sagar Arun Kamble Cc: Chris Wilson Reviewed-by: Michał Winiarski Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180320181419.35576-3-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc.h | 14 ++++++++++++++ drivers/gpu/drm/i915/intel_guc_log.c | 26 ++++++++++++-------------- 2 files changed, 26 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h index 9a95d15..13f3d1d 100644 --- a/drivers/gpu/drm/i915/intel_guc.h +++ b/drivers/gpu/drm/i915/intel_guc.h @@ -155,4 +155,18 @@ static inline int intel_guc_sanitize(struct intel_guc *guc) return 0; } +static inline void intel_guc_enable_msg(struct intel_guc *guc, u32 mask) +{ + spin_lock_irq(&guc->irq_lock); + guc->msg_enabled_mask |= mask; + spin_unlock_irq(&guc->irq_lock); +} + +static inline void intel_guc_disable_msg(struct intel_guc *guc, u32 mask) +{ + spin_lock_irq(&guc->irq_lock); + guc->msg_enabled_mask &= ~mask; + spin_unlock_irq(&guc->irq_lock); +} + #endif diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 188d390..a401f7e 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -72,25 +72,23 @@ static int guc_action_control_log(struct intel_guc *guc, bool enable, return intel_guc_send(guc, action, ARRAY_SIZE(action)); } -static void guc_flush_log_msg_enable(struct intel_guc *guc) +static inline struct intel_guc *log_to_guc(struct intel_guc_log *log) { - spin_lock_irq(&guc->irq_lock); - guc->msg_enabled_mask |= INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | - INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED; - spin_unlock_irq(&guc->irq_lock); + return container_of(log, struct intel_guc, log); } -static void guc_flush_log_msg_disable(struct intel_guc *guc) +static void guc_log_enable_flush_events(struct intel_guc_log *log) { - spin_lock_irq(&guc->irq_lock); - guc->msg_enabled_mask &= ~(INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | - INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED); - spin_unlock_irq(&guc->irq_lock); + intel_guc_enable_msg(log_to_guc(log), + INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | + INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED); } -static inline struct intel_guc *log_to_guc(struct intel_guc_log *log) +static void guc_log_disable_flush_events(struct intel_guc_log *log) { - return container_of(log, struct intel_guc, log); + intel_guc_disable_msg(log_to_guc(log), + INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | + INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED); } /* @@ -576,7 +574,7 @@ int intel_guc_log_relay_open(struct intel_guc_log *log) mutex_unlock(&log->relay.lock); - guc_flush_log_msg_enable(log_to_guc(log)); + guc_log_enable_flush_events(log); /* * When GuC is logging without us relaying to userspace, we're ignoring @@ -616,7 +614,7 @@ void intel_guc_log_relay_flush(struct intel_guc_log *log) void intel_guc_log_relay_close(struct intel_guc_log *log) { - guc_flush_log_msg_disable(log_to_guc(log)); + guc_log_disable_flush_events(log); flush_work(&log->relay.flush_work); mutex_lock(&log->relay.lock); -- cgit v1.1 From e9c7e651798b340a175e18fc70ba41c7008d0760 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 19 Mar 2018 12:50:49 +0000 Subject: drm/i915/guc: Handle GuC log flush event in dedicated function MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We already try to keep all GuC log related code in separate file, handling flush event should be placed there too. This will also allow future code reuse. v2: rebased Signed-off-by: Michal Wajdeczko Cc: Michal Winiarski Cc: Sagar Arun Kamble Cc: Chris Wilson Cc: Oscar Mateo Reviewed-by: Michał Winiarski Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180319125049.48932-1-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc.c | 3 +-- drivers/gpu/drm/i915/intel_guc_log.c | 5 +++++ drivers/gpu/drm/i915/intel_guc_log.h | 2 ++ 3 files changed, 8 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index 4b7c9c6..8f93f5b 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -398,8 +398,7 @@ void intel_guc_to_host_event_handler(struct intel_guc *guc) if (msg & (INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED)) - queue_work(guc->log.relay.flush_wq, - &guc->log.relay.flush_work); + intel_guc_log_handle_flush_event(&guc->log); } int intel_guc_sample_forcewake(struct intel_guc *guc) diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index a401f7e..401e170 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -623,3 +623,8 @@ void intel_guc_log_relay_close(struct intel_guc_log *log) guc_log_relay_destroy(log); mutex_unlock(&log->relay.lock); } + +void intel_guc_log_handle_flush_event(struct intel_guc_log *log) +{ + queue_work(log->relay.flush_wq, &log->relay.flush_work); +} diff --git a/drivers/gpu/drm/i915/intel_guc_log.h b/drivers/gpu/drm/i915/intel_guc_log.h index 1b0d2fa..fa80535 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.h +++ b/drivers/gpu/drm/i915/intel_guc_log.h @@ -87,4 +87,6 @@ int intel_guc_log_relay_open(struct intel_guc_log *log); void intel_guc_log_relay_flush(struct intel_guc_log *log); void intel_guc_log_relay_close(struct intel_guc_log *log); +void intel_guc_log_handle_flush_event(struct intel_guc_log *log); + #endif -- cgit v1.1 From d871bfd0089f50e3010f361e804d290abe67119c Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Tue, 20 Mar 2018 16:20:20 +0000 Subject: drm/i915/guc: Unify parameters of public CT functions There is no need to mix parameter types in public CT functions as we can always accept intel_guc_ct. v2: fix 'Return' doc, s/dev_priv/i915 (Sagar) Signed-off-by: Michal Wajdeczko Cc: Sagar Arun Kamble Cc: Chris Wilson Reviewed-by: Sagar Arun Kamble Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180320162020.38672-1-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc_ct.c | 41 ++++++++++++++++++++++++------------- drivers/gpu/drm/i915/intel_guc_ct.h | 6 ++---- drivers/gpu/drm/i915/intel_uc.c | 4 ++-- 3 files changed, 31 insertions(+), 20 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc_ct.c b/drivers/gpu/drm/i915/intel_guc_ct.c index 0a0d3d5..a726283 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/intel_guc_ct.c @@ -28,12 +28,21 @@ enum { CTB_SEND = 0, CTB_RECV = 1 }; enum { CTB_OWNER_HOST = 0 }; +/** + * intel_guc_ct_init_early - Initialize CT state without requiring device access + * @ct: pointer to CT struct + */ void intel_guc_ct_init_early(struct intel_guc_ct *ct) { /* we're using static channel owners */ ct->host_channel.owner = CTB_OWNER_HOST; } +static inline struct intel_guc *ct_to_guc(struct intel_guc_ct *ct) +{ + return container_of(ct, struct intel_guc, ct); +} + static inline const char *guc_ct_buffer_type_to_str(u32 type) { switch (type) { @@ -416,19 +425,21 @@ static int intel_guc_send_ct(struct intel_guc *guc, const u32 *action, u32 len) } /** - * Enable buffer based command transport + * intel_guc_ct_enable - Enable buffer based command transport. + * @ct: pointer to CT struct + * * Shall only be called for platforms with HAS_GUC_CT. - * @guc: the guc - * return: 0 on success - * non-zero on failure + * + * Return: 0 on success, a negative errno code on failure. */ -int intel_guc_enable_ct(struct intel_guc *guc) +int intel_guc_ct_enable(struct intel_guc_ct *ct) { - struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct intel_guc_ct_channel *ctch = &guc->ct.host_channel; + struct intel_guc *guc = ct_to_guc(ct); + struct drm_i915_private *i915 = guc_to_i915(guc); + struct intel_guc_ct_channel *ctch = &ct->host_channel; int err; - GEM_BUG_ON(!HAS_GUC_CT(dev_priv)); + GEM_BUG_ON(!HAS_GUC_CT(i915)); err = ctch_open(guc, ctch); if (unlikely(err)) @@ -441,16 +452,18 @@ int intel_guc_enable_ct(struct intel_guc *guc) } /** - * Disable buffer based command transport. + * intel_guc_ct_disable - Disable buffer based command transport. + * @ct: pointer to CT struct + * * Shall only be called for platforms with HAS_GUC_CT. - * @guc: the guc */ -void intel_guc_disable_ct(struct intel_guc *guc) +void intel_guc_ct_disable(struct intel_guc_ct *ct) { - struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct intel_guc_ct_channel *ctch = &guc->ct.host_channel; + struct intel_guc *guc = ct_to_guc(ct); + struct drm_i915_private *i915 = guc_to_i915(guc); + struct intel_guc_ct_channel *ctch = &ct->host_channel; - GEM_BUG_ON(!HAS_GUC_CT(dev_priv)); + GEM_BUG_ON(!HAS_GUC_CT(i915)); if (!ctch_is_open(ctch)) return; diff --git a/drivers/gpu/drm/i915/intel_guc_ct.h b/drivers/gpu/drm/i915/intel_guc_ct.h index 6d97f36..595c8ad 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.h +++ b/drivers/gpu/drm/i915/intel_guc_ct.h @@ -78,9 +78,7 @@ struct intel_guc_ct { }; void intel_guc_ct_init_early(struct intel_guc_ct *ct); - -/* XXX: move to intel_uc.h ? don't fit there either */ -int intel_guc_enable_ct(struct intel_guc *guc); -void intel_guc_disable_ct(struct intel_guc *guc); +int intel_guc_ct_enable(struct intel_guc_ct *ct); +void intel_guc_ct_disable(struct intel_guc_ct *ct); #endif /* _INTEL_GUC_CT_H_ */ diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 2befcaf..34f8a2c 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -231,7 +231,7 @@ static int guc_enable_communication(struct intel_guc *guc) gen9_enable_guc_interrupts(dev_priv); if (HAS_GUC_CT(dev_priv)) - return intel_guc_enable_ct(guc); + return intel_guc_ct_enable(&guc->ct); guc->send = intel_guc_send_mmio; return 0; @@ -242,7 +242,7 @@ static void guc_disable_communication(struct intel_guc *guc) struct drm_i915_private *dev_priv = guc_to_i915(guc); if (HAS_GUC_CT(dev_priv)) - intel_guc_disable_ct(guc); + intel_guc_ct_disable(&guc->ct); gen9_disable_guc_interrupts(dev_priv); -- cgit v1.1 From 9153e6b7c85edbc89e874e5c83f86217c53dcfaf Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 21 Mar 2018 09:10:27 +0000 Subject: drm/i915/execlists: Use a locked clear_bit() for synchronisation with interrupt We were relying on the uncached reads when processing the CSB to provide ourselves with the serialisation with the interrupt handler (so we could detect new interrupts in the middle of processing the old one). However, in commit 767a983ab255 ("drm/i915/execlists: Read the context-status HEAD from the HWSP") those uncached reads were eliminated (on one path at least) and along with them our serialisation. The result is that we would very rarely miss notification of a new interrupt and leave a context-switch unprocessed, hanging the GPU. Fixes: 767a983ab255 ("drm/i915/execlists: Read the context-status HEAD from the HWSP") Signed-off-by: Chris Wilson Cc: Michel Thierry Cc: Tvrtko Ursulin Cc: Mika Kuoppala Reviewed-by: Tvrtko Ursulin Reviewed-by: Mika Kuoppala Reviewed-by: Michel Thierry Link: https://patchwork.freedesktop.org/patch/msgid/20180321091027.21034-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/intel_lrc.c | 21 ++++++++------------- 1 file changed, 8 insertions(+), 13 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 53f1c00..67b6a0f 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -831,7 +831,8 @@ static void execlists_submission_tasklet(unsigned long data) struct drm_i915_private *dev_priv = engine->i915; bool fw = false; - /* We can skip acquiring intel_runtime_pm_get() here as it was taken + /* + * We can skip acquiring intel_runtime_pm_get() here as it was taken * on our behalf by the request (see i915_gem_mark_busy()) and it will * not be relinquished until the device is idle (see * i915_gem_idle_work_handler()). As a precaution, we make sure @@ -840,7 +841,8 @@ static void execlists_submission_tasklet(unsigned long data) */ GEM_BUG_ON(!dev_priv->gt.awake); - /* Prefer doing test_and_clear_bit() as a two stage operation to avoid + /* + * Prefer doing test_and_clear_bit() as a two stage operation to avoid * imposing the cost of a locked atomic transaction when submitting a * new request (outside of the context-switch interrupt). */ @@ -856,17 +858,10 @@ static void execlists_submission_tasklet(unsigned long data) execlists->csb_head = -1; /* force mmio read of CSB ptrs */ } - /* The write will be ordered by the uncached read (itself - * a memory barrier), so we do not need another in the form - * of a locked instruction. The race between the interrupt - * handler and the split test/clear is harmless as we order - * our clear before the CSB read. If the interrupt arrived - * first between the test and the clear, we read the updated - * CSB and clear the bit. If the interrupt arrives as we read - * the CSB or later (i.e. after we had cleared the bit) the bit - * is set and we do a new loop. - */ - __clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); + /* Clear before reading to catch new interrupts */ + clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); + smp_mb__after_atomic(); + if (unlikely(execlists->csb_head == -1)) { /* following a reset */ if (!fw) { intel_uncore_forcewake_get(dev_priv, -- cgit v1.1 From b90eed08d8d0f07f9f08074645d4470e121ff6f5 Mon Sep 17 00:00:00 2001 From: Dhinakaran Pandiyan Date: Mon, 12 Mar 2018 20:46:45 -0700 Subject: drm/i915/psr: Move PSR aux setup to it's own function. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Non-functional change useful for the following patch. Cc: Rodrigo Vivi Cc: José Roberto de Souza Signed-off-by: Dhinakaran Pandiyan Reviewed-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180313034646.3721-1-dhinakaran.pandiyan@intel.com --- drivers/gpu/drm/i915/intel_psr.c | 31 ++++++++++++++++++++----------- 1 file changed, 20 insertions(+), 11 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index 317cb4a..2c001f4 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -246,7 +246,7 @@ static i915_reg_t psr_aux_data_reg(struct drm_i915_private *dev_priv, return EDP_PSR_AUX_DATA(index); } -static void hsw_psr_enable_sink(struct intel_dp *intel_dp) +static void hsw_psr_setup_aux(struct intel_dp *intel_dp) { struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); struct drm_device *dev = dig_port->base.base.dev; @@ -267,6 +267,24 @@ static void hsw_psr_enable_sink(struct intel_dp *intel_dp) BUILD_BUG_ON(sizeof(aux_msg) > 20); aux_clock_divider = intel_dp->get_aux_clock_divider(intel_dp, 0); + aux_ctl_reg = psr_aux_ctl_reg(dev_priv, port); + + /* Setup AUX registers */ + for (i = 0; i < sizeof(aux_msg); i += 4) + I915_WRITE(psr_aux_data_reg(dev_priv, port, i >> 2), + intel_dp_pack_aux(&aux_msg[i], sizeof(aux_msg) - i)); + + aux_ctl = intel_dp->get_aux_send_ctl(intel_dp, 0, sizeof(aux_msg), + aux_clock_divider); + I915_WRITE(aux_ctl_reg, aux_ctl); +} + +static void hsw_psr_enable_sink(struct intel_dp *intel_dp) +{ + struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); + struct drm_device *dev = dig_port->base.base.dev; + struct drm_i915_private *dev_priv = to_i915(dev); + /* Enable AUX frame sync at sink */ if (dev_priv->psr.aux_frame_sync) @@ -285,16 +303,7 @@ static void hsw_psr_enable_sink(struct intel_dp *intel_dp) drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, DP_PSR_ENABLE); - aux_ctl_reg = psr_aux_ctl_reg(dev_priv, port); - - /* Setup AUX registers */ - for (i = 0; i < sizeof(aux_msg); i += 4) - I915_WRITE(psr_aux_data_reg(dev_priv, port, i >> 2), - intel_dp_pack_aux(&aux_msg[i], sizeof(aux_msg) - i)); - - aux_ctl = intel_dp->get_aux_send_ctl(intel_dp, 0, sizeof(aux_msg), - aux_clock_divider); - I915_WRITE(aux_ctl_reg, aux_ctl); + hsw_psr_setup_aux(intel_dp); } static void vlv_psr_enable_source(struct intel_dp *intel_dp, -- cgit v1.1 From d544e918ff132488770ab2cb6b03e2af69497d1c Mon Sep 17 00:00:00 2001 From: Dhinakaran Pandiyan Date: Mon, 12 Mar 2018 20:46:46 -0700 Subject: drm/i915/psr: Remove open-coded PSR AUX transactions for SKL+ MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit HSW and BDW have SRD_AUX_{CTL, STATUS} registers that the driver needs to setup for the HW to use whenever exiting PSR. SKL+ hardware use hardcoded values for the same and do not need any registers to be setup. So, use drm_dp_dpcd_writeb() for a one-time write during PSR enable and setup the PSR aux registers on HSW and BDW for later use by HW. We also end up writing to reserved bits in SRD_AUX_CTL by reusing intel_dp->get_aux_send_ctl() for HSW and BDW, fix this. Since the AUX register setup is source side programming, move the call to enable_source() from enable_sink(). Cc: José Roberto de Souza Cc: Rodrigo Vivi Signed-off-by: Dhinakaran Pandiyan Reviewed-by: Jose Roberto de Souza Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180313034646.3721-2-dhinakaran.pandiyan@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 6 +++++ drivers/gpu/drm/i915/intel_psr.c | 55 ++++++++++++++++------------------------ 2 files changed, 28 insertions(+), 33 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index bac3e92..4e31dff 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -3846,6 +3846,12 @@ enum { #define EDP_PSR_IDLE_FRAME_SHIFT 0 #define EDP_PSR_AUX_CTL _MMIO(dev_priv->psr_mmio_base + 0x10) +#define EDP_PSR_AUX_CTL_TIME_OUT_MASK (3 << 26) +#define EDP_PSR_AUX_CTL_MESSAGE_SIZE_MASK (0x1f << 20) +#define EDP_PSR_AUX_CTL_PRECHARGE_2US_MASK (0xf << 16) +#define EDP_PSR_AUX_CTL_ERROR_INTERRUPT (1 << 11) +#define EDP_PSR_AUX_CTL_BIT_CLOCK_2X_MASK (0x7ff) + #define EDP_PSR_AUX_DATA(i) _MMIO(dev_priv->psr_mmio_base + 0x14 + (i) * 4) /* 5 registers */ #define EDP_PSR_STATUS _MMIO(dev_priv->psr_mmio_base + 0x40) diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index 2c001f4..b8e083e 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -228,31 +228,12 @@ static void vlv_psr_enable_sink(struct intel_dp *intel_dp) DP_PSR_ENABLE | DP_PSR_MAIN_LINK_ACTIVE); } -static i915_reg_t psr_aux_ctl_reg(struct drm_i915_private *dev_priv, - enum port port) -{ - if (INTEL_GEN(dev_priv) >= 9) - return DP_AUX_CH_CTL(port); - else - return EDP_PSR_AUX_CTL; -} - -static i915_reg_t psr_aux_data_reg(struct drm_i915_private *dev_priv, - enum port port, int index) -{ - if (INTEL_GEN(dev_priv) >= 9) - return DP_AUX_CH_DATA(port, index); - else - return EDP_PSR_AUX_DATA(index); -} - static void hsw_psr_setup_aux(struct intel_dp *intel_dp) { struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - struct drm_device *dev = dig_port->base.base.dev; - struct drm_i915_private *dev_priv = to_i915(dev); - uint32_t aux_clock_divider; - i915_reg_t aux_ctl_reg; + struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev); + u32 aux_clock_divider, aux_ctl; + int i; static const uint8_t aux_msg[] = { [0] = DP_AUX_NATIVE_WRITE << 4, [1] = DP_SET_POWER >> 8, @@ -260,23 +241,25 @@ static void hsw_psr_setup_aux(struct intel_dp *intel_dp) [3] = 1 - 1, [4] = DP_SET_POWER_D0, }; - enum port port = dig_port->base.port; - u32 aux_ctl; - int i; + u32 psr_aux_mask = EDP_PSR_AUX_CTL_TIME_OUT_MASK | + EDP_PSR_AUX_CTL_MESSAGE_SIZE_MASK | + EDP_PSR_AUX_CTL_PRECHARGE_2US_MASK | + EDP_PSR_AUX_CTL_BIT_CLOCK_2X_MASK; BUILD_BUG_ON(sizeof(aux_msg) > 20); - - aux_clock_divider = intel_dp->get_aux_clock_divider(intel_dp, 0); - aux_ctl_reg = psr_aux_ctl_reg(dev_priv, port); - - /* Setup AUX registers */ for (i = 0; i < sizeof(aux_msg); i += 4) - I915_WRITE(psr_aux_data_reg(dev_priv, port, i >> 2), + I915_WRITE(EDP_PSR_AUX_DATA(i >> 2), intel_dp_pack_aux(&aux_msg[i], sizeof(aux_msg) - i)); + aux_clock_divider = intel_dp->get_aux_clock_divider(intel_dp, 0); + + /* Start with bits set for DDI_AUX_CTL register */ aux_ctl = intel_dp->get_aux_send_ctl(intel_dp, 0, sizeof(aux_msg), aux_clock_divider); - I915_WRITE(aux_ctl_reg, aux_ctl); + + /* Select only valid bits for SRD_AUX_CTL */ + aux_ctl &= psr_aux_mask; + I915_WRITE(EDP_PSR_AUX_CTL, aux_ctl); } static void hsw_psr_enable_sink(struct intel_dp *intel_dp) @@ -303,7 +286,7 @@ static void hsw_psr_enable_sink(struct intel_dp *intel_dp) drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, DP_PSR_ENABLE); - hsw_psr_setup_aux(intel_dp); + drm_dp_dpcd_writeb(&intel_dp->aux, DP_SET_POWER, DP_SET_POWER_D0); } static void vlv_psr_enable_source(struct intel_dp *intel_dp, @@ -599,6 +582,12 @@ static void hsw_psr_enable_source(struct intel_dp *intel_dp, psr_aux_io_power_get(intel_dp); + /* Only HSW and BDW have PSR AUX registers that need to be setup. SKL+ + * use hardcoded values PSR AUX transactions + */ + if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) + hsw_psr_setup_aux(intel_dp); + if (dev_priv->psr.psr2_support) { chicken = PSR2_VSC_ENABLE_PROG_HEADER; if (dev_priv->psr.y_cord_support) -- cgit v1.1 From 3ae7fb202d86b7847f237daa474f3946bdc3b0c6 Mon Sep 17 00:00:00 2001 From: Haneen Mohammed Date: Tue, 20 Mar 2018 09:37:49 -0400 Subject: drm: Remove drm_property_{un/reference}_blob aliases This patch remove the compatibility aliases drm_property_{reference/unreference}_blob of drm_property_blob_{get/put} since all callers have been converted to the prefered _{get/put}. Remove the helpers from the semantic patch drm-get-put-cocci. Signed-off-by: Haneen Mohammed Signed-off-by: Sean Paul Link: https://patchwork.freedesktop.org/patch/msgid/20180320133749.GA11695@haneen-VirtualBox --- include/drm/drm_property.h | 26 -------------------------- scripts/coccinelle/api/drm-get-put.cocci | 10 ---------- 2 files changed, 36 deletions(-) diff --git a/include/drm/drm_property.h b/include/drm/drm_property.h index d1423c7f..ab8167b 100644 --- a/include/drm/drm_property.h +++ b/include/drm/drm_property.h @@ -281,32 +281,6 @@ struct drm_property_blob *drm_property_blob_get(struct drm_property_blob *blob); void drm_property_blob_put(struct drm_property_blob *blob); /** - * drm_property_reference_blob - acquire a blob property reference - * @blob: DRM blob property - * - * This is a compatibility alias for drm_property_blob_get() and should not be - * used by new code. - */ -static inline struct drm_property_blob * -drm_property_reference_blob(struct drm_property_blob *blob) -{ - return drm_property_blob_get(blob); -} - -/** - * drm_property_unreference_blob - release a blob property reference - * @blob: DRM blob property - * - * This is a compatibility alias for drm_property_blob_put() and should not be - * used by new code. - */ -static inline void -drm_property_unreference_blob(struct drm_property_blob *blob) -{ - drm_property_blob_put(blob); -} - -/** * drm_property_find - find property object * @dev: DRM device * @file_priv: drm file to check for lease against. diff --git a/scripts/coccinelle/api/drm-get-put.cocci b/scripts/coccinelle/api/drm-get-put.cocci index ceb71ea..3a09c97 100644 --- a/scripts/coccinelle/api/drm-get-put.cocci +++ b/scripts/coccinelle/api/drm-get-put.cocci @@ -40,12 +40,6 @@ expression object; - drm_gem_object_unreference_unlocked(object) + drm_gem_object_put_unlocked(object) | -- drm_property_reference_blob(object) -+ drm_property_blob_get(object) -| -- drm_property_unreference_blob(object) -+ drm_property_blob_put(object) -| - drm_dev_unref(object) + drm_dev_put(object) ) @@ -72,10 +66,6 @@ __drm_gem_object_unreference(object) | drm_gem_object_unreference_unlocked(object) | -drm_property_unreference_blob@p(object) -| -drm_property_reference_blob@p(object) -| drm_dev_unref@p(object) ) -- cgit v1.1 From 0e59c209f4ccf9f9d505babdb04731294e18c4ed Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 22 Mar 2018 11:00:59 +0000 Subject: drm/i915: Fix tracing of submit seqno We pre-increment the timeline->seqno when handing it to the request, make sure the GEM_TRACE takes this into account. Otherwise, it appears that we go backwards over a preemption point: 1d..1 157681077us : __i915_request_unsubmit: vcs0 fence 75e:3 <- global_seqno 17 0d.s1 157681113us : __i915_request_submit: vcs0 fence 75e:3 -> global_seqno 16 Fixes: d9b13c4dde6c ("drm/i915: Trace GEM steps between submit and wedging") Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Tvrtko Ursulin Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180322110059.4467-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_request.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index 2325886..f1b81fe 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -501,7 +501,7 @@ void __i915_request_submit(struct i915_request *request) GEM_TRACE("%s fence %llx:%d -> global_seqno %d\n", request->engine->name, request->fence.context, request->fence.seqno, - engine->timeline->seqno); + engine->timeline->seqno + 1); GEM_BUG_ON(!irqs_disabled()); lockdep_assert_held(&engine->timeline->lock); -- cgit v1.1 From 4ccfee92f4b6fbbedee1eb68f110a66f03edf7c6 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 22 Mar 2018 13:10:34 +0000 Subject: drm/i915: Remove local timeline var from submit/unsubmit Both request_submit and request_unsubmit deal with transferring the request from the client's timeline onto the execution timeline and back again. As both functions deal with a pair of timeline's, using a shorthand for just one of them is slightly confusing, especially as the different functions use the shorthand for the alternate timeline. Instead, use the full version of each timeline so it should be easier to keep track of the transfer between the request/client and the engine. v2: Refactor the common lock+list_move v3: Be clear we require the other timeline list to be locked as well. Signed-off-by: Chris Wilson Cc: Mika Kuoppala Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180322131034.6036-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_request.c | 30 +++++++++++++++--------------- 1 file changed, 15 insertions(+), 15 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index f1b81fe..2314a26 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -492,10 +492,20 @@ static u32 timeline_get_seqno(struct intel_timeline *tl) return ++tl->seqno; } +static void move_to_timeline(struct i915_request *request, + struct intel_timeline *timeline) +{ + GEM_BUG_ON(request->timeline == request->engine->timeline); + lockdep_assert_held(&request->engine->timeline->lock); + + spin_lock(&request->timeline->lock); + list_move_tail(&request->link, &timeline->requests); + spin_unlock(&request->timeline->lock); +} + void __i915_request_submit(struct i915_request *request) { struct intel_engine_cs *engine = request->engine; - struct intel_timeline *timeline; u32 seqno; GEM_TRACE("%s fence %llx:%d -> global_seqno %d\n", @@ -506,12 +516,9 @@ void __i915_request_submit(struct i915_request *request) GEM_BUG_ON(!irqs_disabled()); lockdep_assert_held(&engine->timeline->lock); - /* Transfer from per-context onto the global per-engine timeline */ - timeline = engine->timeline; - GEM_BUG_ON(timeline == request->timeline); GEM_BUG_ON(request->global_seqno); - seqno = timeline_get_seqno(timeline); + seqno = timeline_get_seqno(engine->timeline); GEM_BUG_ON(!seqno); GEM_BUG_ON(i915_seqno_passed(intel_engine_get_seqno(engine), seqno)); @@ -525,9 +532,8 @@ void __i915_request_submit(struct i915_request *request) engine->emit_breadcrumb(request, request->ring->vaddr + request->postfix); - spin_lock(&request->timeline->lock); - list_move_tail(&request->link, &timeline->requests); - spin_unlock(&request->timeline->lock); + /* Transfer from per-context onto the global per-engine timeline */ + move_to_timeline(request, engine->timeline); trace_i915_request_execute(request); @@ -550,7 +556,6 @@ void i915_request_submit(struct i915_request *request) void __i915_request_unsubmit(struct i915_request *request) { struct intel_engine_cs *engine = request->engine; - struct intel_timeline *timeline; GEM_TRACE("%s fence %llx:%d <- global_seqno %d\n", request->engine->name, @@ -578,12 +583,7 @@ void __i915_request_unsubmit(struct i915_request *request) spin_unlock(&request->lock); /* Transfer back from the global per-engine timeline to per-context */ - timeline = request->timeline; - GEM_BUG_ON(timeline == engine->timeline); - - spin_lock(&timeline->lock); - list_move(&request->link, &timeline->requests); - spin_unlock(&timeline->lock); + move_to_timeline(request, request->timeline); /* * We don't need to wake_up any waiters on request->execute, they -- cgit v1.1 From 0ade43909d599bdde0d60e0c79e6d73479d65ffa Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 22 Mar 2018 07:49:08 +0000 Subject: drm/i915/selftests: Include the trace as a debug aide If we fail to reset the GPU in a timely fashion, dump the GEM trace so that we can see what operations were in flight when the GPU got stuck. v2: There's more than one timeout that deserves tracing! v3: Silence checkpatch by not even using a product at all! Signed-off-by: Chris Wilson Reviewed-by: Mika Kuoppala Reviewed-by: Jeff McGee Link: https://patchwork.freedesktop.org/patch/msgid/20180322074908.10838-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/selftests/intel_hangcheck.c | 23 ++++++++++++++++++++--- 1 file changed, 20 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c index 4372826..9b235da 100644 --- a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c +++ b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c @@ -260,8 +260,11 @@ static void wedge_me(struct work_struct *work) { struct wedge_me *w = container_of(work, typeof(*w), work.work); - pr_err("%pS timed out, cancelling all further testing.\n", - w->symbol); + pr_err("%pS timed out, cancelling all further testing.\n", w->symbol); + + GEM_TRACE("%pS timed out.\n", w->symbol); + GEM_TRACE_DUMP(); + i915_gem_set_wedged(w->i915); } @@ -621,9 +624,19 @@ static int active_engine(void *data) mutex_unlock(&engine->i915->drm.struct_mutex); if (old) { - i915_request_wait(old, 0, MAX_SCHEDULE_TIMEOUT); + if (i915_request_wait(old, 0, HZ) < 0) { + GEM_TRACE("%s timed out.\n", engine->name); + GEM_TRACE_DUMP(); + + i915_gem_set_wedged(engine->i915); + i915_request_put(old); + err = -EIO; + break; + } i915_request_put(old); } + + cond_resched(); } for (count = 0; count < ARRAY_SIZE(rq); count++) @@ -1126,6 +1139,10 @@ int intel_hangcheck_live_selftests(struct drm_i915_private *i915) err = i915_subtests(tests, i915); + mutex_lock(&i915->drm.struct_mutex); + flush_test(i915, I915_WAIT_LOCKED); + mutex_unlock(&i915->drm.struct_mutex); + i915_modparams.enable_hangcheck = saved_hangcheck; intel_runtime_pm_put(i915); -- cgit v1.1 From a90507d60763ee1067cf217614a8bb2ab43aca1a Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 22 Mar 2018 07:35:31 +0000 Subject: drm/i915/selftests: Stress resets-vs-request-priority Watch what happens if we try to reset with a queue of requests with varying priorities -- that may need reordering or preemption across the reset. v2: Tweak priorities to avoid starving the hanging thread. Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180322073533.5313-2-chris@chris-wilson.co.uk Reviewed-by: Jeff McGee --- drivers/gpu/drm/i915/selftests/intel_hangcheck.c | 189 +++++++++++++++-------- 1 file changed, 126 insertions(+), 63 deletions(-) diff --git a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c index 9b235da..9e4e0ad 100644 --- a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c +++ b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c @@ -25,6 +25,7 @@ #include #include "../i915_selftest.h" +#include "i915_random.h" #include "mock_context.h" #include "mock_drm.h" @@ -486,6 +487,8 @@ static int __igt_reset_engine(struct drm_i915_private *i915, bool active) set_bit(I915_RESET_ENGINE + id, &i915->gpu_error.flags); do { + u32 seqno = intel_engine_get_seqno(engine); + if (active) { struct i915_request *rq; @@ -514,12 +517,13 @@ static int __igt_reset_engine(struct drm_i915_private *i915, bool active) break; } + GEM_BUG_ON(!rq->global_seqno); + seqno = rq->global_seqno - 1; i915_request_put(rq); } engine->hangcheck.stalled = true; - engine->hangcheck.seqno = - intel_engine_get_seqno(engine); + engine->hangcheck.seqno = seqno; err = i915_reset_engine(engine, NULL); if (err) { @@ -576,11 +580,25 @@ static int igt_reset_active_engine(void *arg) return __igt_reset_engine(arg, true); } +struct active_engine { + struct task_struct *task; + struct intel_engine_cs *engine; + unsigned long resets; + unsigned int flags; +}; + +#define TEST_ACTIVE BIT(0) +#define TEST_OTHERS BIT(1) +#define TEST_SELF BIT(2) +#define TEST_PRIORITY BIT(3) + static int active_engine(void *data) { - struct intel_engine_cs *engine = data; - struct i915_request *rq[2] = {}; - struct i915_gem_context *ctx[2]; + I915_RND_STATE(prng); + struct active_engine *arg = data; + struct intel_engine_cs *engine = arg->engine; + struct i915_request *rq[8] = {}; + struct i915_gem_context *ctx[ARRAY_SIZE(rq)]; struct drm_file *file; unsigned long count = 0; int err = 0; @@ -589,25 +607,20 @@ static int active_engine(void *data) if (IS_ERR(file)) return PTR_ERR(file); - mutex_lock(&engine->i915->drm.struct_mutex); - ctx[0] = live_context(engine->i915, file); - mutex_unlock(&engine->i915->drm.struct_mutex); - if (IS_ERR(ctx[0])) { - err = PTR_ERR(ctx[0]); - goto err_file; - } - - mutex_lock(&engine->i915->drm.struct_mutex); - ctx[1] = live_context(engine->i915, file); - mutex_unlock(&engine->i915->drm.struct_mutex); - if (IS_ERR(ctx[1])) { - err = PTR_ERR(ctx[1]); - i915_gem_context_put(ctx[0]); - goto err_file; + for (count = 0; count < ARRAY_SIZE(ctx); count++) { + mutex_lock(&engine->i915->drm.struct_mutex); + ctx[count] = live_context(engine->i915, file); + mutex_unlock(&engine->i915->drm.struct_mutex); + if (IS_ERR(ctx[count])) { + err = PTR_ERR(ctx[count]); + while (--count) + i915_gem_context_put(ctx[count]); + goto err_file; + } } while (!kthread_should_stop()) { - unsigned int idx = count++ & 1; + unsigned int idx = count++ & (ARRAY_SIZE(rq) - 1); struct i915_request *old = rq[idx]; struct i915_request *new; @@ -619,6 +632,10 @@ static int active_engine(void *data) break; } + if (arg->flags & TEST_PRIORITY) + ctx[idx]->priority = + i915_prandom_u32_max_state(512, &prng); + rq[idx] = i915_request_get(new); i915_request_add(new); mutex_unlock(&engine->i915->drm.struct_mutex); @@ -647,8 +664,9 @@ err_file: return err; } -static int __igt_reset_engine_others(struct drm_i915_private *i915, - bool active) +static int __igt_reset_engines(struct drm_i915_private *i915, + const char *test_name, + unsigned int flags) { struct intel_engine_cs *engine, *other; enum intel_engine_id id, tmp; @@ -662,50 +680,61 @@ static int __igt_reset_engine_others(struct drm_i915_private *i915, if (!intel_has_reset_engine(i915)) return 0; - if (active) { + if (flags & TEST_ACTIVE) { mutex_lock(&i915->drm.struct_mutex); err = hang_init(&h, i915); mutex_unlock(&i915->drm.struct_mutex); if (err) return err; + + if (flags & TEST_PRIORITY) + h.ctx->priority = 1024; } for_each_engine(engine, i915, id) { - struct task_struct *threads[I915_NUM_ENGINES] = {}; - unsigned long resets[I915_NUM_ENGINES]; + struct active_engine threads[I915_NUM_ENGINES] = {}; unsigned long global = i915_reset_count(&i915->gpu_error); - unsigned long count = 0; + unsigned long count = 0, reported; IGT_TIMEOUT(end_time); - if (active && !intel_engine_can_store_dword(engine)) + if (flags & TEST_ACTIVE && + !intel_engine_can_store_dword(engine)) continue; memset(threads, 0, sizeof(threads)); for_each_engine(other, i915, tmp) { struct task_struct *tsk; - resets[tmp] = i915_reset_engine_count(&i915->gpu_error, - other); + threads[tmp].resets = + i915_reset_engine_count(&i915->gpu_error, + other); - if (other == engine) + if (!(flags & TEST_OTHERS)) continue; - tsk = kthread_run(active_engine, other, + if (other == engine && !(flags & TEST_SELF)) + continue; + + threads[tmp].engine = other; + threads[tmp].flags = flags; + + tsk = kthread_run(active_engine, &threads[tmp], "igt/%s", other->name); if (IS_ERR(tsk)) { err = PTR_ERR(tsk); goto unwind; } - threads[tmp] = tsk; + threads[tmp].task = tsk; get_task_struct(tsk); } set_bit(I915_RESET_ENGINE + id, &i915->gpu_error.flags); do { - if (active) { - struct i915_request *rq; + u32 seqno = intel_engine_get_seqno(engine); + struct i915_request *rq = NULL; + if (flags & TEST_ACTIVE) { mutex_lock(&i915->drm.struct_mutex); rq = hang_create_request(&h, engine); if (IS_ERR(rq)) { @@ -731,33 +760,38 @@ static int __igt_reset_engine_others(struct drm_i915_private *i915, break; } - i915_request_put(rq); + GEM_BUG_ON(!rq->global_seqno); + seqno = rq->global_seqno - 1; } engine->hangcheck.stalled = true; - engine->hangcheck.seqno = - intel_engine_get_seqno(engine); + engine->hangcheck.seqno = seqno; err = i915_reset_engine(engine, NULL); if (err) { - pr_err("i915_reset_engine(%s:%s) failed, err=%d\n", - engine->name, active ? "active" : "idle", err); + pr_err("i915_reset_engine(%s:%s): failed, err=%d\n", + engine->name, test_name, err); break; } engine->hangcheck.stalled = false; count++; + + if (rq) { + i915_request_wait(rq, 0, MAX_SCHEDULE_TIMEOUT); + i915_request_put(rq); + } } while (time_before(jiffies, end_time)); clear_bit(I915_RESET_ENGINE + id, &i915->gpu_error.flags); pr_info("i915_reset_engine(%s:%s): %lu resets\n", - engine->name, active ? "active" : "idle", count); - - if (i915_reset_engine_count(&i915->gpu_error, engine) - - resets[engine->id] != (active ? count : 0)) { - pr_err("i915_reset_engine(%s:%s): reset %lu times, but reported %lu\n", - engine->name, active ? "active" : "idle", count, - i915_reset_engine_count(&i915->gpu_error, - engine) - resets[engine->id]); + engine->name, test_name, count); + + reported = i915_reset_engine_count(&i915->gpu_error, engine); + reported -= threads[engine->id].resets; + if (reported != (flags & TEST_ACTIVE ? count : 0)) { + pr_err("i915_reset_engine(%s:%s): reset %lu times, but reported %lu, expected %lu reported\n", + engine->name, test_name, count, reported, + (flags & TEST_ACTIVE ? count : 0)); if (!err) err = -EINVAL; } @@ -766,24 +800,26 @@ unwind: for_each_engine(other, i915, tmp) { int ret; - if (!threads[tmp]) + if (!threads[tmp].task) continue; - ret = kthread_stop(threads[tmp]); + ret = kthread_stop(threads[tmp].task); if (ret) { pr_err("kthread for other engine %s failed, err=%d\n", other->name, ret); if (!err) err = ret; } - put_task_struct(threads[tmp]); + put_task_struct(threads[tmp].task); - if (resets[tmp] != i915_reset_engine_count(&i915->gpu_error, - other)) { + if (other != engine && + threads[tmp].resets != + i915_reset_engine_count(&i915->gpu_error, other)) { pr_err("Innocent engine %s was reset (count=%ld)\n", other->name, i915_reset_engine_count(&i915->gpu_error, - other) - resets[tmp]); + other) - + threads[tmp].resets); if (!err) err = -EINVAL; } @@ -807,7 +843,7 @@ unwind: if (i915_terminally_wedged(&i915->gpu_error)) err = -EIO; - if (active) { + if (flags & TEST_ACTIVE) { mutex_lock(&i915->drm.struct_mutex); hang_fini(&h); mutex_unlock(&i915->drm.struct_mutex); @@ -816,14 +852,42 @@ unwind: return err; } -static int igt_reset_idle_engine_others(void *arg) +static int igt_reset_engines(void *arg) { - return __igt_reset_engine_others(arg, false); -} + static const struct { + const char *name; + unsigned int flags; + } phases[] = { + { "idle", 0 }, + { "active", TEST_ACTIVE }, + { "others-idle", TEST_OTHERS }, + { "others-active", TEST_OTHERS | TEST_ACTIVE }, + { + "others-priority", + TEST_OTHERS | TEST_ACTIVE | TEST_PRIORITY + }, + { + "self-priority", + TEST_OTHERS | TEST_ACTIVE | TEST_PRIORITY | TEST_SELF, + }, + { } + }; + struct drm_i915_private *i915 = arg; + typeof(*phases) *p; + int err; -static int igt_reset_active_engine_others(void *arg) -{ - return __igt_reset_engine_others(arg, true); + for (p = phases; p->name; p++) { + if (p->flags & TEST_PRIORITY) { + if (!(i915->caps.scheduler & I915_SCHEDULER_CAP_PRIORITY)) + continue; + } + + err = __igt_reset_engines(arg, p->name, p->flags); + if (err) + return err; + } + + return 0; } static u32 fake_hangcheck(struct i915_request *rq) @@ -1122,8 +1186,7 @@ int intel_hangcheck_live_selftests(struct drm_i915_private *i915) SUBTEST(igt_hang_sanitycheck), SUBTEST(igt_reset_idle_engine), SUBTEST(igt_reset_active_engine), - SUBTEST(igt_reset_idle_engine_others), - SUBTEST(igt_reset_active_engine_others), + SUBTEST(igt_reset_engines), SUBTEST(igt_wait_reset), SUBTEST(igt_reset_queue), SUBTEST(igt_handle_error), -- cgit v1.1 From 1c645bf4378f1539df57c6228f5c4957c130324a Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 22 Mar 2018 07:35:32 +0000 Subject: drm/i915: Use full serialisation around engine->irq_posted MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Using engine->irq_posted for execlists, we are not always serialised by the tasklet as we supposed. On the reset paths, the tasklet is disabled and ignored. Instead, we manipulate the engine->irq_posted directly to account for the reset, but if an interrupt fired before the reset and so wrote to engine->irq_posted, that write may not be flushed from the local CPU's cacheline until much later as the tasklet is already active and so does not generate a mb(). To correctly serialise the interrupt with reset, we need serialisation on the set_bit() itself. And at last Mika can be happy. Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Michał Winiarski CC: Michel Thierry Cc: Jeff McGee Cc: Tvrtko Ursulin Reviewed-by: Mika Kuoppala Reviewed-by: Jeff McGee Link: https://patchwork.freedesktop.org/patch/msgid/20180322073533.5313-3-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/i915_irq.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index fa73107..27aee25 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -1405,10 +1405,9 @@ gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir) bool tasklet = false; if (iir & GT_CONTEXT_SWITCH_INTERRUPT) { - if (READ_ONCE(engine->execlists.active)) { - __set_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); - tasklet = true; - } + if (READ_ONCE(engine->execlists.active)) + tasklet = !test_and_set_bit(ENGINE_IRQ_EXECLIST, + &engine->irq_posted); } if (iir & GT_RENDER_USER_INTERRUPT) { -- cgit v1.1 From 0f36a85c3bd5e0dfcbb49af203a96a933dae86cf Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 22 Mar 2018 07:35:33 +0000 Subject: drm/i915: Flush pending interrupt following a GPU reset MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit After resetting the GPU (or subset of engines), call synchronize_irq() to flush any pending irq before proceeding with the cleanup. For a device level reset, we disable the interupts around the reset, but when resetting just one engine, we have to avoid such global disabling. This leaves us open to an interrupt arriving for the engine as we try to reset it. We already do try to flush the IIR following the reset, but we have to ensure that the in-flight interrupt does not land after we start cleaning up after the reset; enter synchronize_irq(). As it current stands, we very rarely, but fatally, see sequences such as: 2.... 57964564us : execlists_reset_prepare: rcs0 2.... 57964613us : execlists_reset: rcs0 seqno=424 0d.h1 57964615us : gen8_cs_irq_handler: rcs0 CS active=1 2d..1 57964617us : __i915_request_unsubmit: rcs0 fence 29:1056 <- global_seqno 1060 2.... 57964703us : execlists_reset_finish: rcs0 0..s. 57964705us : execlists_submission_tasklet: rcs0 awake?=1, active=0, irq-posted?=1 v2: Move the sync into the execlists reset handler so that we coordinate the flush with disabling the interrupt handling and canceling the pending interrupt. v3: Just use synchronize_hardirq() to avoid the might_sleep(), we do not yet have threaded-irq to worry about. Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Michel Thierry Cc: Michał Winiarski Cc: Jeff McGee Link: https://patchwork.freedesktop.org/patch/msgid/20180322073533.5313-4-chris@chris-wilson.co.uk Reviewed-by: Jeff McGee --- drivers/gpu/drm/i915/intel_lrc.c | 7 ++++--- drivers/gpu/drm/i915/intel_uncore.c | 4 +++- 2 files changed, 7 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 67b6a0f..ce09c5a 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -805,6 +805,10 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) spin_unlock(&engine->timeline->lock); + /* Mark all CS interrupts as complete */ + smp_store_mb(execlists->active, 0); + synchronize_hardirq(engine->i915->drm.irq); + /* * The port is checked prior to scheduling a tasklet, but * just in case we have suspended the tasklet to do the @@ -813,9 +817,6 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) */ clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); - /* Mark all CS interrupts as complete */ - execlists->active = 0; - local_irq_restore(flags); } diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c index 4c616d0..f37ecfc 100644 --- a/drivers/gpu/drm/i915/intel_uncore.c +++ b/drivers/gpu/drm/i915/intel_uncore.c @@ -2116,8 +2116,10 @@ int intel_gpu_reset(struct drm_i915_private *dev_priv, unsigned engine_mask) i915_stop_engines(dev_priv, engine_mask); ret = -ENODEV; - if (reset) + if (reset) { + GEM_TRACE("engine_mask=%x\n", engine_mask); ret = reset(dev_priv, engine_mask); + } if (ret != -ETIMEDOUT) break; -- cgit v1.1 From 66c1f77ae2b773f349c1ea1312d69a6ab775cc26 Mon Sep 17 00:00:00 2001 From: Mika Kuoppala Date: Tue, 20 Mar 2018 17:17:33 +0200 Subject: drm/i915: Avoid setting ring freq on invalid rps freqs Looping through rps frequencies when both min and max are zero ends up into an endless loop. This can happen during hardware enablement. Bail out early if rps frequencies are not correctly set yet. Cc: Chris Wilson Signed-off-by: Mika Kuoppala Reviewed-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180320151734.11761-1-mika.kuoppala@linux.intel.com --- drivers/gpu/drm/i915/intel_pm.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index dd5ddb7..19e82aa 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -6890,15 +6890,18 @@ static void gen6_enable_rps(struct drm_i915_private *dev_priv) static void gen6_update_ring_freq(struct drm_i915_private *dev_priv) { struct intel_rps *rps = &dev_priv->gt_pm.rps; - int min_freq = 15; + const int min_freq = 15; + const int scaling_factor = 180; unsigned int gpu_freq; unsigned int max_ia_freq, min_ring_freq; unsigned int max_gpu_freq, min_gpu_freq; - int scaling_factor = 180; struct cpufreq_policy *policy; WARN_ON(!mutex_is_locked(&dev_priv->pcu_lock)); + if (rps->max_freq <= rps->min_freq) + return; + policy = cpufreq_cpu_get(0); if (policy) { max_ia_freq = policy->cpuinfo.max_freq; @@ -6932,7 +6935,7 @@ static void gen6_update_ring_freq(struct drm_i915_private *dev_priv) * the PCU should use as a reference to determine the ring frequency. */ for (gpu_freq = max_gpu_freq; gpu_freq >= min_gpu_freq; gpu_freq--) { - int diff = max_gpu_freq - gpu_freq; + const int diff = max_gpu_freq - gpu_freq; unsigned int ia_freq = 0, ring_freq = 0; if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { -- cgit v1.1 From 0ef904bb3a6d9dc2d81301e84116eaff880412f2 Mon Sep 17 00:00:00 2001 From: Tvrtko Ursulin Date: Wed, 21 Mar 2018 10:32:28 +0000 Subject: drm/i915: Skip logging impossible slices Log up to sseu->max_slices instead basing on ARRAY_SIZE since to avoid printing impossible and empty slices for a platform. Also compact slice total and slice mask into one log line. Signed-off-by: Tvrtko Ursulin Cc: Lionel Landwerlin Reviewed-by: Lionel Landwerlin Link: https://patchwork.freedesktop.org/patch/msgid/20180321103228.32205-1-tvrtko.ursulin@linux.intel.com --- drivers/gpu/drm/i915/intel_device_info.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c index a504281..0d1509e 100644 --- a/drivers/gpu/drm/i915/intel_device_info.c +++ b/drivers/gpu/drm/i915/intel_device_info.c @@ -83,11 +83,11 @@ static void sseu_dump(const struct sseu_dev_info *sseu, struct drm_printer *p) { int s; - drm_printf(p, "slice mask: %04x\n", sseu->slice_mask); - drm_printf(p, "slice total: %u\n", hweight8(sseu->slice_mask)); + drm_printf(p, "slice total: %u, mask=%04x\n", + hweight8(sseu->slice_mask), sseu->slice_mask); drm_printf(p, "subslice total: %u\n", sseu_subslice_total(sseu)); - for (s = 0; s < ARRAY_SIZE(sseu->subslice_mask); s++) { - drm_printf(p, "slice%d %u subslices mask=%04x\n", + for (s = 0; s < sseu->max_slices; s++) { + drm_printf(p, "slice%d: %u subslices, mask=%04x\n", s, hweight8(sseu->subslice_mask[s]), sseu->subslice_mask[s]); } -- cgit v1.1 From 277ab5abc68df2f6f8fac7a46e50105b6648f432 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 17:47:07 +0200 Subject: drm/i915: Don't spew errors when resetting HDMI scrambling/bit clock ratio fails MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When we're disabling the HDMI link we try to reset the scrambling and TMDS bit clock ratio back to the default values. This will fail if the sink has already been disconnected. Thus we should not print an error message when resetting the scrambling/TMDS bit clock ratio fail during disable. During enable we do want the error, and during disable we may still want to know what happended for debug purposes so let's use DRM_DEBUG_KMS() there. v2: Remember them consts v3: Go back to just one function and print the errors/debugs from callers (Shashank) Cc: Shashank Sharma Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105644 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105655 Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322154707.22103-1-ville.syrjala@linux.intel.com Reviewed-by: Shashank Sharma --- drivers/gpu/drm/i915/intel_ddi.c | 19 ++++++++++++------- drivers/gpu/drm/i915/intel_drv.h | 2 +- drivers/gpu/drm/i915/intel_hdmi.c | 40 ++++++++++++++++----------------------- 3 files changed, 29 insertions(+), 32 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index 8c2d778..c449619 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -2424,12 +2424,14 @@ static void intel_enable_ddi_hdmi(struct intel_encoder *encoder, { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_digital_port *dig_port = enc_to_dig_port(&encoder->base); + struct drm_connector *connector = conn_state->connector; enum port port = encoder->port; - intel_hdmi_handle_sink_scrambling(encoder, - conn_state->connector, - crtc_state->hdmi_high_tmds_clock_ratio, - crtc_state->hdmi_scrambling); + if (!intel_hdmi_handle_sink_scrambling(encoder, connector, + crtc_state->hdmi_high_tmds_clock_ratio, + crtc_state->hdmi_scrambling)) + DRM_ERROR("[CONNECTOR:%d:%s] Failed to configure sink scrambling/TMDS bit clock ratio\n", + connector->base.id, connector->name); /* Display WA #1143: skl,kbl,cfl */ if (IS_GEN9_BC(dev_priv)) { @@ -2520,13 +2522,16 @@ static void intel_disable_ddi_hdmi(struct intel_encoder *encoder, const struct intel_crtc_state *old_crtc_state, const struct drm_connector_state *old_conn_state) { + struct drm_connector *connector = old_conn_state->connector; + if (old_crtc_state->has_audio) intel_audio_codec_disable(encoder, old_crtc_state, old_conn_state); - intel_hdmi_handle_sink_scrambling(encoder, - old_conn_state->connector, - false, false); + if (!intel_hdmi_handle_sink_scrambling(encoder, connector, + false, false)) + DRM_DEBUG_KMS("[CONNECTOR:%d:%s] Failed to reset sink scrambling/TMDS bit clock ratio\n", + connector->base.id, connector->name); } static void intel_disable_ddi(struct intel_encoder *encoder, diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index de6db91..b79a01b 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -1782,7 +1782,7 @@ struct intel_hdmi *enc_to_intel_hdmi(struct drm_encoder *encoder); bool intel_hdmi_compute_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config, struct drm_connector_state *conn_state); -void intel_hdmi_handle_sink_scrambling(struct intel_encoder *intel_encoder, +bool intel_hdmi_handle_sink_scrambling(struct intel_encoder *encoder, struct drm_connector *connector, bool high_tmds_clock_ratio, bool scrambling); diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c index 1baef4a..ee929f3 100644 --- a/drivers/gpu/drm/i915/intel_hdmi.c +++ b/drivers/gpu/drm/i915/intel_hdmi.c @@ -2082,41 +2082,33 @@ intel_hdmi_add_properties(struct intel_hdmi *intel_hdmi, struct drm_connector *c * it enables scrambling. This should be called before enabling the HDMI * 2.0 port, as the sink can choose to disable the scrambling if it doesn't * detect a scrambled clock within 100 ms. + * + * Returns: + * True on success, false on failure. */ -void intel_hdmi_handle_sink_scrambling(struct intel_encoder *encoder, +bool intel_hdmi_handle_sink_scrambling(struct intel_encoder *encoder, struct drm_connector *connector, bool high_tmds_clock_ratio, bool scrambling) { + struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base); - struct drm_i915_private *dev_priv = connector->dev->dev_private; struct drm_scrambling *sink_scrambling = - &connector->display_info.hdmi.scdc.scrambling; - struct i2c_adapter *adptr = intel_gmbus_get_adapter(dev_priv, - intel_hdmi->ddc_bus); - bool ret; + &connector->display_info.hdmi.scdc.scrambling; + struct i2c_adapter *adapter = + intel_gmbus_get_adapter(dev_priv, intel_hdmi->ddc_bus); if (!sink_scrambling->supported) - return; - - DRM_DEBUG_KMS("Setting sink scrambling for enc:%s connector:%s\n", - encoder->base.name, connector->name); + return true; - /* Set TMDS bit clock ratio to 1/40 or 1/10 */ - ret = drm_scdc_set_high_tmds_clock_ratio(adptr, high_tmds_clock_ratio); - if (!ret) { - DRM_ERROR("Set TMDS ratio failed\n"); - return; - } - - /* Enable/disable sink scrambling */ - ret = drm_scdc_set_scrambling(adptr, scrambling); - if (!ret) { - DRM_ERROR("Set sink scrambling failed\n"); - return; - } + DRM_DEBUG_KMS("[CONNECTOR:%d:%s] scrambling=%s, TMDS bit clock ratio=1/%d\n", + connector->base.id, connector->name, + yesno(scrambling), high_tmds_clock_ratio ? 40 : 10); - DRM_DEBUG_KMS("sink scrambling handled\n"); + /* Set TMDS bit clock ratio to 1/40 or 1/10, and enable/disable scrambling */ + return drm_scdc_set_high_tmds_clock_ratio(adapter, + high_tmds_clock_ratio) && + drm_scdc_set_scrambling(adapter, scrambling); } static u8 chv_port_to_ddc_pin(struct drm_i915_private *dev_priv, enum port port) -- cgit v1.1 From 28e0e8ac27409e1f2f2abd226548f0ebeef19ad8 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Piotr=20Pi=C3=B3rkowski?= Date: Fri, 23 Mar 2018 12:23:18 +0100 Subject: drm/i915/guc: Fix null pointer dereference when GuC FW is not available MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If GuC firmware is not available on the system and we load i915 with enable GuC, then we hit this null pointer dereference issue: [ 71.098873] BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 [ 71.098938] IP: intel_uc_fw_upload+0x1f/0x360 [i915] [ 71.098947] PGD 0 P4D 0 [ 71.098956] Oops: 0000 [#1] PREEMPT SMP PTI [ 71.098965] Modules linked in: i915(O+) netconsole x86_pkg_temp_thermal intel_powerclamp coretemp crct10dif_pclmul crc32_pclmul ghash_clmulni_intel mei_me i2c_i801 prime_numbers mei [last unloaded: i915] [ 71.099005] CPU: 2 PID: 1167 Comm: insmod Tainted: G U W O 4.16.0-rc1+ #337 [ 71.099018] Hardware name: /NUC6i5SYB, BIOS SYSKLi35.86A.0065.2018.0103.1000 01/03/2018 [ 71.099077] RIP: 0010:intel_uc_fw_upload+0x1f/0x360 [i915] [ 71.099087] RSP: 0018:ffffc90000417aa0 EFLAGS: 00010282 [ 71.099097] RAX: 0000000000000000 RBX: ffff88084cad12f8 RCX: ffffffffa03e9357 [ 71.099108] RDX: 0000000000000002 RSI: ffffffffa034dba0 RDI: ffff88084cad12f8 [ 71.099118] RBP: 0000000000000002 R08: ffff88085344ca90 R09: 0000000000000001 [ 71.099128] R10: 0000000000000000 R11: 0000000000000000 R12: ffff88084cad0000 [ 71.099139] R13: ffffffffa034dba0 R14: 00000000fffffff5 R15: ffff88084cad12b0 [ 71.099151] FS: 00007f7f24ae2740(0000) GS:ffff88085e200000(0000) knlGS:0000000000000000 [ 71.099162] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 71.099171] CR2: 0000000000000008 CR3: 0000000855f48001 CR4: 00000000003606e0 [ 71.099182] Call Trace: [ 71.099246] intel_uc_init_hw+0xc8/0x520 [i915] [ 71.099303] i915_gem_init_hw+0x11f/0x2d0 [i915] [ 71.099364] i915_gem_init+0x2b9/0x640 [i915] [ 71.099413] i915_driver_load+0xb74/0x1110 [i915] [ 71.099462] i915_pci_probe+0x2e/0x90 [i915] [ 71.099476] pci_device_probe+0xa1/0x130 [ 71.099488] driver_probe_device+0x302/0x470 [ 71.099502] __driver_attach+0xb9/0xe0 [ 71.099513] ? driver_probe_device+0x470/0x470 [ 71.099525] ? driver_probe_device+0x470/0x470 [ 71.099538] bus_for_each_dev+0x64/0x90 [ 71.099550] bus_add_driver+0x164/0x260 [ 71.099561] ? 0xffffffffa04d6000 [ 71.099572] driver_register+0x57/0xc0 [ 71.099582] ? 0xffffffffa04d6000 [ 71.099593] do_one_initcall+0x3b/0x160 [ 71.099606] ? kmem_cache_alloc_trace+0x1c3/0x2a0 [ 71.099621] do_init_module+0x5b/0x1f9 [ 71.099635] load_module+0x2467/0x2a70 [ 71.099654] ? SyS_finit_module+0xbd/0xe0 [ 71.099668] SyS_finit_module+0xbd/0xe0 [ 71.099682] do_syscall_64+0x73/0x1c0 [ 71.099694] entry_SYSCALL_64_after_hwframe+0x26/0x9b [ 71.099706] RIP: 0033:0x7f7f23fb40d9 [ 71.099717] RSP: 002b:00007ffda7d67ed8 EFLAGS: 00000246 ORIG_RAX: 0000000000000139 [ 71.099734] RAX: ffffffffffffffda RBX: 000055f96e2a8870 RCX: 00007f7f23fb40d9 [ 71.099748] RDX: 0000000000000000 RSI: 000055f96e2a8260 RDI: 0000000000000003 [ 71.099763] RBP: 000055f96e2a8260 R08: 0000000000000000 R09: 00007ffda7d68088 [ 71.099777] R10: 0000000000000003 R11: 0000000000000246 R12: 0000000000000000 [ 71.099791] R13: 000055f96e2a8830 R14: 0000000000000000 R15: 000055f96e2a8260 [ 71.099810] Code: 00 00 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 41 55 41 54 49 89 f5 55 53 48 c7 c1 57 93 3e a0 48 8b 47 10 48 89 fb 4c 8b 07 <48> 8b 68 08 8b 47 28 85 c0 74 15 83 f8 01 48 c7 c1 5b 93 3e a0 [ 71.100004] RIP: intel_uc_fw_upload+0x1f/0x360 [i915] RSP: ffffc90000417aa0 [ 71.100020] CR2: 0000000000000008 [ 71.100031] ---[ end trace d8ac93c30ceff5b2 ]-- Fixes: 6b0478fb722a ("drm/i915: Implement dynamic GuC WOPCM offset and size calculation") v2: don't assume it is always GuC FW (Michal) v3: added a new variable to avoid exceeding the number of characters in the line (Michal) Signed-off-by: Piotr Piórkowski Reported-by: Radoslaw Szwichtenberg Cc: Michał Winiarski Cc: Michal Wajdeczko Cc: Sagar Arun Kamble Cc: Joonas Lahtinen Cc: Jackie Li Cc: Radoslaw Szwichtenberg Reviewed-by: Michal Wajdeczko Reviewed-by: Jackie Li Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180323112319.16293-1-piotr.piorkowski@intel.com --- drivers/gpu/drm/i915/intel_uc_fw.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_uc_fw.c b/drivers/gpu/drm/i915/intel_uc_fw.c index 30c7324..6e8e0b5 100644 --- a/drivers/gpu/drm/i915/intel_uc_fw.c +++ b/drivers/gpu/drm/i915/intel_uc_fw.c @@ -199,8 +199,8 @@ int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, int (*xfer)(struct intel_uc_fw *uc_fw, struct i915_vma *vma)) { - struct drm_i915_private *i915 = to_i915(uc_fw->obj->base.dev); struct i915_vma *vma; + u32 ggtt_pin_bias; int err; DRM_DEBUG_DRIVER("%s fw load %s\n", @@ -222,9 +222,9 @@ int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, goto fail; } + ggtt_pin_bias = to_i915(uc_fw->obj->base.dev)->guc.ggtt_pin_bias; vma = i915_gem_object_ggtt_pin(uc_fw->obj, NULL, 0, 0, - PIN_OFFSET_BIAS | - i915->guc.ggtt_pin_bias); + PIN_OFFSET_BIAS | ggtt_pin_bias); if (IS_ERR(vma)) { err = PTR_ERR(vma); DRM_DEBUG_DRIVER("%s fw ggtt-pin err=%d\n", -- cgit v1.1 From a0de908d44fb67500b7c45bd8325f316496227db Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Fri, 23 Mar 2018 12:34:49 +0000 Subject: drm/i915: Reorder early initialization In upcoming patch, we want to perform more actions in early initialization of the uC. This reordering will help resolve new dependencies that will be introduced by future patch. v2: s/i915_gem_load_init/i915_gem_init_early (Chris) v3: s/i915_gem_load_cleanup/i915_gem_cleanup_early (Michal) Signed-off-by: Michal Wajdeczko Cc: Chris Wilson Cc: Tvrtko Ursulin Cc: Joonas Lahtinen Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180323123451.59244-1-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/i915_drv.c | 17 ++++++++--------- drivers/gpu/drm/i915/i915_drv.h | 4 ++-- drivers/gpu/drm/i915/i915_gem.c | 5 ++--- 3 files changed, 12 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index a7d3275..2561974 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -919,17 +919,21 @@ static int i915_driver_init_early(struct drm_i915_private *dev_priv, mutex_init(&dev_priv->wm.wm_mutex); mutex_init(&dev_priv->pps_mutex); - intel_wopcm_init_early(&dev_priv->wopcm); - intel_uc_init_early(dev_priv); i915_memcpy_init_early(dev_priv); ret = i915_workqueues_init(dev_priv); if (ret < 0) goto err_engines; + ret = i915_gem_init_early(dev_priv); + if (ret < 0) + goto err_workqueues; + /* This must be called before any calls to HAS_PCH_* */ intel_detect_pch(dev_priv); + intel_wopcm_init_early(&dev_priv->wopcm); + intel_uc_init_early(dev_priv); intel_pm_setup(dev_priv); intel_init_dpio(dev_priv); intel_power_domains_init(dev_priv); @@ -938,18 +942,13 @@ static int i915_driver_init_early(struct drm_i915_private *dev_priv, intel_init_display_hooks(dev_priv); intel_init_clock_gating_hooks(dev_priv); intel_init_audio_hooks(dev_priv); - ret = i915_gem_load_init(dev_priv); - if (ret < 0) - goto err_irq; - intel_display_crc_init(dev_priv); intel_detect_preproduction_hw(dev_priv); return 0; -err_irq: - intel_irq_fini(dev_priv); +err_workqueues: i915_workqueues_cleanup(dev_priv); err_engines: i915_engines_cleanup(dev_priv); @@ -962,8 +961,8 @@ err_engines: */ static void i915_driver_cleanup_early(struct drm_i915_private *dev_priv) { - i915_gem_load_cleanup(dev_priv); intel_irq_fini(dev_priv); + i915_gem_cleanup_early(dev_priv); i915_workqueues_cleanup(dev_priv); i915_engines_cleanup(dev_priv); } diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index c9c3b2b..28ab918 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -2869,8 +2869,8 @@ int i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data, int i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); void i915_gem_sanitize(struct drm_i915_private *i915); -int i915_gem_load_init(struct drm_i915_private *dev_priv); -void i915_gem_load_cleanup(struct drm_i915_private *dev_priv); +int i915_gem_init_early(struct drm_i915_private *dev_priv); +void i915_gem_cleanup_early(struct drm_i915_private *dev_priv); void i915_gem_load_init_fences(struct drm_i915_private *dev_priv); int i915_gem_freeze(struct drm_i915_private *dev_priv); int i915_gem_freeze_late(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 802df8e..9650a7b 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -5502,8 +5502,7 @@ static void i915_gem_init__mm(struct drm_i915_private *i915) INIT_WORK(&i915->mm.free_work, __i915_gem_free_work); } -int -i915_gem_load_init(struct drm_i915_private *dev_priv) +int i915_gem_init_early(struct drm_i915_private *dev_priv) { int err = -ENOMEM; @@ -5578,7 +5577,7 @@ err_out: return err; } -void i915_gem_load_cleanup(struct drm_i915_private *dev_priv) +void i915_gem_cleanup_early(struct drm_i915_private *dev_priv) { i915_gem_drain_freed_objects(dev_priv); GEM_BUG_ON(!llist_empty(&dev_priv->mm.free_list)); -- cgit v1.1 From 8c650aefb82d559aa0e1b7c0c36346b906481106 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Fri, 23 Mar 2018 12:34:50 +0000 Subject: drm/i915/uc: Fetch uC firmware in init_early We were fetching uC firmwares in separate uc_init_fw step, while there is no reason why we can't fetch them during init_early. This will also simplify upcoming patches, as size of the firmware may be used for register initialization. Signed-off-by: Michal Wajdeczko Cc: Michal Winiarski Cc: Sagar Arun Kamble Cc: Chris Wilson Cc: Joonas Lahtinen Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180323123451.59244-2-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/i915_drv.c | 8 ++------ drivers/gpu/drm/i915/intel_guc_fw.c | 5 ++--- drivers/gpu/drm/i915/intel_huc_fw.c | 5 ++--- drivers/gpu/drm/i915/intel_uc.c | 37 ++++++++++++++++++------------------- drivers/gpu/drm/i915/intel_uc.h | 3 +-- 5 files changed, 25 insertions(+), 33 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 2561974..db22337 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -692,11 +692,9 @@ static int i915_load_modeset_init(struct drm_device *dev) if (ret) goto cleanup_irq; - intel_uc_init_fw(dev_priv); - ret = i915_gem_init(dev_priv); if (ret) - goto cleanup_uc; + goto cleanup_irq; intel_setup_overlay(dev_priv); @@ -716,8 +714,6 @@ cleanup_gem: if (i915_gem_suspend(dev_priv)) DRM_ERROR("failed to idle hardware; continuing to unload!\n"); i915_gem_fini(dev_priv); -cleanup_uc: - intel_uc_fini_fw(dev_priv); cleanup_irq: drm_irq_uninstall(dev); intel_teardown_gmbus(dev_priv); @@ -962,6 +958,7 @@ err_engines: static void i915_driver_cleanup_early(struct drm_i915_private *dev_priv) { intel_irq_fini(dev_priv); + intel_uc_cleanup_early(dev_priv); i915_gem_cleanup_early(dev_priv); i915_workqueues_cleanup(dev_priv); i915_engines_cleanup(dev_priv); @@ -1457,7 +1454,6 @@ void i915_driver_unload(struct drm_device *dev) i915_reset_error_state(dev_priv); i915_gem_fini(dev_priv); - intel_uc_fini_fw(dev_priv); intel_fbc_cleanup_cfb(dev_priv); intel_power_domains_fini(dev_priv); diff --git a/drivers/gpu/drm/i915/intel_guc_fw.c b/drivers/gpu/drm/i915/intel_guc_fw.c index 978668c..a9e6fcc 100644 --- a/drivers/gpu/drm/i915/intel_guc_fw.c +++ b/drivers/gpu/drm/i915/intel_guc_fw.c @@ -275,9 +275,8 @@ static int guc_fw_xfer(struct intel_uc_fw *guc_fw, struct i915_vma *vma) * Called from intel_uc_init_hw() during driver load, resume from sleep and * after a GPU reset. * - * The firmware image should have already been fetched into memory by the - * earlier call to intel_uc_init_fw(), so here we need to only check that - * fetch succeeded, and then transfer the image to the h/w. + * The firmware image should have already been fetched into memory, so only + * check that fetch succeeded, and then transfer the image to the h/w. * * Return: non-zero code on error */ diff --git a/drivers/gpu/drm/i915/intel_huc_fw.c b/drivers/gpu/drm/i915/intel_huc_fw.c index bb0f8b7..f93d238 100644 --- a/drivers/gpu/drm/i915/intel_huc_fw.c +++ b/drivers/gpu/drm/i915/intel_huc_fw.c @@ -155,9 +155,8 @@ static int huc_fw_xfer(struct intel_uc_fw *huc_fw, struct i915_vma *vma) * Called from intel_uc_init_hw() during driver load, resume from sleep and * after a GPU reset. Note that HuC must be loaded before GuC. * - * The firmware image should have already been fetched into memory by the - * earlier call to intel_uc_init_fw(), so here we need to only check that - * fetch succeeded, and then transfer the image to the h/w. + * The firmware image should have already been fetched into memory, so only + * check that fetch succeeded, and then transfer the image to the h/w. * * Return: non-zero code on error */ diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 34f8a2c..4aad844 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -162,36 +162,35 @@ static void sanitize_options_early(struct drm_i915_private *dev_priv) GEM_BUG_ON(i915_modparams.guc_log_level < 0); } -void intel_uc_init_early(struct drm_i915_private *dev_priv) +void intel_uc_init_early(struct drm_i915_private *i915) { - intel_guc_init_early(&dev_priv->guc); - intel_huc_init_early(&dev_priv->huc); + struct intel_guc *guc = &i915->guc; + struct intel_huc *huc = &i915->huc; - sanitize_options_early(dev_priv); -} + intel_guc_init_early(guc); + intel_huc_init_early(huc); -void intel_uc_init_fw(struct drm_i915_private *dev_priv) -{ - if (!USES_GUC(dev_priv)) - return; + sanitize_options_early(i915); - if (USES_HUC(dev_priv)) - intel_uc_fw_fetch(dev_priv, &dev_priv->huc.fw); + if (USES_GUC(i915)) + intel_uc_fw_fetch(i915, &guc->fw); - intel_uc_fw_fetch(dev_priv, &dev_priv->guc.fw); + if (USES_HUC(i915)) + intel_uc_fw_fetch(i915, &huc->fw); } -void intel_uc_fini_fw(struct drm_i915_private *dev_priv) +void intel_uc_cleanup_early(struct drm_i915_private *i915) { - if (!USES_GUC(dev_priv)) - return; + struct intel_guc *guc = &i915->guc; + struct intel_huc *huc = &i915->huc; - intel_uc_fw_fini(&dev_priv->guc.fw); + if (USES_HUC(i915)) + intel_uc_fw_fini(&huc->fw); - if (USES_HUC(dev_priv)) - intel_uc_fw_fini(&dev_priv->huc.fw); + if (USES_GUC(i915)) + intel_uc_fw_fini(&guc->fw); - guc_free_load_err_log(&dev_priv->guc); + guc_free_load_err_log(guc); } /** diff --git a/drivers/gpu/drm/i915/intel_uc.h b/drivers/gpu/drm/i915/intel_uc.h index 937e611..25d73ad 100644 --- a/drivers/gpu/drm/i915/intel_uc.h +++ b/drivers/gpu/drm/i915/intel_uc.h @@ -29,9 +29,8 @@ #include "i915_params.h" void intel_uc_init_early(struct drm_i915_private *dev_priv); +void intel_uc_cleanup_early(struct drm_i915_private *dev_priv); void intel_uc_init_mmio(struct drm_i915_private *dev_priv); -void intel_uc_init_fw(struct drm_i915_private *dev_priv); -void intel_uc_fini_fw(struct drm_i915_private *dev_priv); int intel_uc_init_misc(struct drm_i915_private *dev_priv); void intel_uc_fini_misc(struct drm_i915_private *dev_priv); void intel_uc_sanitize(struct drm_i915_private *dev_priv); -- cgit v1.1 From 46b3617dfec875c1414c6ccbfcab371c97735562 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 23 Mar 2018 10:18:24 +0000 Subject: drm/i915: Actually flush interrupts on reset not just wedging MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Commit 0f36a85c3bd5 ("drm/i915: Flush pending interrupt following a GPU reset") got confused and only applied the flush to the set-wedge path (which itself is proving troublesome), but we also need the serialisation on the regular reset path. Oops. Move the interrupt into reset_irq() and make it common to the reset and final set-wedge. v2: reset_irq() after port cancellation, as we assert that execlists->active is sane for cancellation (and is being reset by reset_irq). References: 0f36a85c3bd5 ("drm/i915: Flush pending interrupt following a GPU reset") Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Michel Thierry Cc: Michał Winiarski Cc: Jeff McGee Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180323101824.14645-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/intel_lrc.c | 107 +++++++++++++++++++-------------------- 1 file changed, 53 insertions(+), 54 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index ce09c5a..b4ab06b 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -740,6 +740,57 @@ execlists_cancel_port_requests(struct intel_engine_execlists * const execlists) } } +static void clear_gtiir(struct intel_engine_cs *engine) +{ + static const u8 gtiir[] = { + [RCS] = 0, + [BCS] = 0, + [VCS] = 1, + [VCS2] = 1, + [VECS] = 3, + }; + struct drm_i915_private *dev_priv = engine->i915; + int i; + + /* TODO: correctly reset irqs for gen11 */ + if (WARN_ON_ONCE(INTEL_GEN(engine->i915) >= 11)) + return; + + GEM_BUG_ON(engine->id >= ARRAY_SIZE(gtiir)); + + /* + * Clear any pending interrupt state. + * + * We do it twice out of paranoia that some of the IIR are + * double buffered, and so if we only reset it once there may + * still be an interrupt pending. + */ + for (i = 0; i < 2; i++) { + I915_WRITE(GEN8_GT_IIR(gtiir[engine->id]), + engine->irq_keep_mask); + POSTING_READ(GEN8_GT_IIR(gtiir[engine->id])); + } + GEM_BUG_ON(I915_READ(GEN8_GT_IIR(gtiir[engine->id])) & + engine->irq_keep_mask); +} + +static void reset_irq(struct intel_engine_cs *engine) +{ + /* Mark all CS interrupts as complete */ + smp_store_mb(engine->execlists.active, 0); + synchronize_hardirq(engine->i915->drm.irq); + + clear_gtiir(engine); + + /* + * The port is checked prior to scheduling a tasklet, but + * just in case we have suspended the tasklet to do the + * wedging make sure that when it wakes, it decides there + * is no work to do by clearing the irq_posted bit. + */ + clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); +} + static void execlists_cancel_requests(struct intel_engine_cs *engine) { struct intel_engine_execlists * const execlists = &engine->execlists; @@ -767,6 +818,7 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) /* Cancel the requests on the HW and clear the ELSP tracker. */ execlists_cancel_port_requests(execlists); + reset_irq(engine); spin_lock(&engine->timeline->lock); @@ -805,18 +857,6 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) spin_unlock(&engine->timeline->lock); - /* Mark all CS interrupts as complete */ - smp_store_mb(execlists->active, 0); - synchronize_hardirq(engine->i915->drm.irq); - - /* - * The port is checked prior to scheduling a tasklet, but - * just in case we have suspended the tasklet to do the - * wedging make sure that when it wakes, it decides there - * is no work to do by clearing the irq_posted bit. - */ - clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); - local_irq_restore(flags); } @@ -1566,14 +1606,6 @@ static int intel_init_workaround_bb(struct intel_engine_cs *engine) return ret; } -static u8 gtiir[] = { - [RCS] = 0, - [BCS] = 0, - [VCS] = 1, - [VCS2] = 1, - [VECS] = 3, -}; - static void enable_execlists(struct intel_engine_cs *engine) { struct drm_i915_private *dev_priv = engine->i915; @@ -1657,35 +1689,6 @@ static int gen9_init_render_ring(struct intel_engine_cs *engine) return init_workarounds_ring(engine); } -static void reset_irq(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int i; - - /* TODO: correctly reset irqs for gen11 */ - if (WARN_ON_ONCE(INTEL_GEN(engine->i915) >= 11)) - return; - - GEM_BUG_ON(engine->id >= ARRAY_SIZE(gtiir)); - - /* - * Clear any pending interrupt state. - * - * We do it twice out of paranoia that some of the IIR are double - * buffered, and if we only reset it once there may still be - * an interrupt pending. - */ - for (i = 0; i < 2; i++) { - I915_WRITE(GEN8_GT_IIR(gtiir[engine->id]), - engine->irq_keep_mask); - POSTING_READ(GEN8_GT_IIR(gtiir[engine->id])); - } - GEM_BUG_ON(I915_READ(GEN8_GT_IIR(gtiir[engine->id])) & - engine->irq_keep_mask); - - clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); -} - static void reset_common_ring(struct intel_engine_cs *engine, struct i915_request *request) { @@ -1699,8 +1702,6 @@ static void reset_common_ring(struct intel_engine_cs *engine, /* See execlists_cancel_requests() for the irq/spinlock split. */ local_irq_save(flags); - reset_irq(engine); - /* * Catch up with any missed context-switch interrupts. * @@ -1711,15 +1712,13 @@ static void reset_common_ring(struct intel_engine_cs *engine, * requests were completed. */ execlists_cancel_port_requests(execlists); + reset_irq(engine); /* Push back any incomplete requests for replay after the reset. */ spin_lock(&engine->timeline->lock); __unwind_incomplete_requests(engine); spin_unlock(&engine->timeline->lock); - /* Mark all CS interrupts as complete */ - execlists->active = 0; - local_irq_restore(flags); /* -- cgit v1.1 From 0f90603c33bdf6575cfdc81edd53f3f13ba166fb Mon Sep 17 00:00:00 2001 From: Imre Deak Date: Thu, 22 Mar 2018 16:36:42 +0200 Subject: drm/i915: Fix hibernation with ACPI S0 target state MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit After commit dd9f31c7a3887950cbd0d49eb9d43f7a1518a356 Author: Imre Deak Date: Wed Aug 16 17:46:07 2017 +0300 drm/i915/gen9+: Set same power state before hibernation image save/restore during hibernation/suspend the power domain functionality got disabled, after which resume could leave it incorrectly disabled if the ACPI target state was S0 during suspend and i915 was not loaded by the loader kernel. This was caused by not considering if we resumed from hibernation as the condition for power domains reiniting. Fix this by simply tracking if we suspended power domains during system suspend and reinit power domains accordingly during resume. This will result in reiniting power domains always when resuming from hibernation, regardless of the platform and whether or not i915 is loaded by the loader kernel. The reason we didn't catch this earlier is that the enabled/disabled state of power domains during PMSG_FREEZE/PMSG_QUIESCE is platform and kernel config dependent: on my SKL the target state is S4 during PMSG_FREEZE and (with the driver loaded in the loader kernel) S0 during PMSG_QUIESCE. On the reporter's machine it's S0 during PMSG_FREEZE but (contrary to this) power domains are not initialized during PMSG_QUIESCE since i915 is not loaded in the loader kernel, or it's loaded but without the DMC firmware being available. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105196 Reported-and-tested-by: amn-bas@hotmail.com Fixes: dd9f31c7a388 ("drm/i915/gen9+: Set same power state before hibernation image save/restore") Cc: amn-bas@hotmail.com Cc: Ville Syrjälä Cc: Signed-off-by: Imre Deak Reviewed-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322143642.26883-1-imre.deak@intel.com --- drivers/gpu/drm/i915/i915_drv.c | 22 ++++++++++------------ drivers/gpu/drm/i915/i915_drv.h | 2 +- 2 files changed, 11 insertions(+), 13 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index db22337..d354627 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -1607,15 +1607,12 @@ static int i915_drm_suspend_late(struct drm_device *dev, bool hibernation) { struct drm_i915_private *dev_priv = to_i915(dev); struct pci_dev *pdev = dev_priv->drm.pdev; - bool fw_csr; int ret; disable_rpm_wakeref_asserts(dev_priv); intel_display_set_init_power(dev_priv, false); - fw_csr = !IS_GEN9_LP(dev_priv) && !hibernation && - suspend_to_idle(dev_priv) && dev_priv->csr.dmc_payload; /* * In case of firmware assisted context save/restore don't manually * deinit the power domains. This also means the CSR/DMC firmware will @@ -1623,8 +1620,11 @@ static int i915_drm_suspend_late(struct drm_device *dev, bool hibernation) * also enable deeper system power states that would be blocked if the * firmware was inactive. */ - if (!fw_csr) + if (IS_GEN9_LP(dev_priv) || hibernation || !suspend_to_idle(dev_priv) || + dev_priv->csr.dmc_payload == NULL) { intel_power_domains_suspend(dev_priv); + dev_priv->power_domains_suspended = true; + } ret = 0; if (IS_GEN9_LP(dev_priv)) @@ -1636,8 +1636,10 @@ static int i915_drm_suspend_late(struct drm_device *dev, bool hibernation) if (ret) { DRM_ERROR("Suspend complete failed: %d\n", ret); - if (!fw_csr) + if (dev_priv->power_domains_suspended) { intel_power_domains_init_hw(dev_priv, true); + dev_priv->power_domains_suspended = false; + } goto out; } @@ -1658,8 +1660,6 @@ static int i915_drm_suspend_late(struct drm_device *dev, bool hibernation) if (!(hibernation && INTEL_GEN(dev_priv) < 6)) pci_set_power_state(pdev, PCI_D3hot); - dev_priv->suspended_to_idle = suspend_to_idle(dev_priv); - out: enable_rpm_wakeref_asserts(dev_priv); @@ -1826,8 +1826,7 @@ static int i915_drm_resume_early(struct drm_device *dev) intel_uncore_resume_early(dev_priv); if (IS_GEN9_LP(dev_priv)) { - if (!dev_priv->suspended_to_idle) - gen9_sanitize_dc_state(dev_priv); + gen9_sanitize_dc_state(dev_priv); bxt_disable_dc9(dev_priv); } else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) { hsw_disable_pc8(dev_priv); @@ -1835,8 +1834,7 @@ static int i915_drm_resume_early(struct drm_device *dev) intel_uncore_sanitize(dev_priv); - if (IS_GEN9_LP(dev_priv) || - !(dev_priv->suspended_to_idle && dev_priv->csr.dmc_payload)) + if (dev_priv->power_domains_suspended) intel_power_domains_init_hw(dev_priv, true); else intel_display_set_init_power(dev_priv, true); @@ -1846,7 +1844,7 @@ static int i915_drm_resume_early(struct drm_device *dev) enable_rpm_wakeref_asserts(dev_priv); out: - dev_priv->suspended_to_idle = false; + dev_priv->power_domains_suspended = false; return ret; } diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 28ab918..299b240 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -1851,7 +1851,7 @@ struct drm_i915_private { u32 bxt_phy_grc; u32 suspend_count; - bool suspended_to_idle; + bool power_domains_suspended; struct i915_suspend_saved_registers regfile; struct vlv_s0ix_state vlv_s0ix_state; -- cgit v1.1 From e52482dec8366a98ac380b3bdc1a4abb8a390914 Mon Sep 17 00:00:00 2001 From: Felix Kuehling Date: Fri, 23 Mar 2018 15:32:28 -0400 Subject: drm/amdgpu: Add MMU notifier type for KFD userptr This commit adds the notion of MMU notifier types GFX and HSA. GFX continues to work like MMU notifiers did before. HSA adds support for KFD userptr BOs. The implementation of KFD userptr eviction is a stub for now. Signed-off-by: Felix Kuehling Reviewed-by: Oded Gabbay Signed-off-by: Oded Gabbay --- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h | 1 + drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 7 ++ drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c | 94 ++++++++++++++++++++---- drivers/gpu/drm/amd/amdgpu/amdgpu_mn.h | 11 ++- 5 files changed, 97 insertions(+), 18 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h index c2c2bea..83e0c5c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h @@ -104,6 +104,7 @@ void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev); void amdgpu_amdkfd_device_init(struct amdgpu_device *adev); void amdgpu_amdkfd_device_fini(struct amdgpu_device *adev); +int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm); int amdgpu_amdkfd_submit_ib(struct kgd_dev *kgd, enum kgd_engine_type engine, uint32_t vmid, uint64_t gpu_addr, uint32_t *ib_cmd, uint32_t ib_len); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index 1d6e147..2463ff6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -1418,6 +1418,13 @@ bo_reserve_failed: return ret; } +int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, + struct mm_struct *mm) +{ + /* TODO */ + return 0; +} + /** amdgpu_amdkfd_gpuvm_restore_process_bos - Restore all BOs for the given * KFD process identified by process_info * diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index dc34b50..8e66f37 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -536,7 +536,7 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, if (p->bo_list) { amdgpu_bo_list_get_list(p->bo_list, &p->validated); if (p->bo_list->first_userptr != p->bo_list->num_entries) - p->mn = amdgpu_mn_get(p->adev); + p->mn = amdgpu_mn_get(p->adev, AMDGPU_MN_TYPE_GFX); } INIT_LIST_HEAD(&duplicates); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c index bd67f4c..f2ed18e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c @@ -36,12 +36,14 @@ #include #include "amdgpu.h" +#include "amdgpu_amdkfd.h" struct amdgpu_mn { /* constant after initialisation */ struct amdgpu_device *adev; struct mm_struct *mm; struct mmu_notifier mn; + enum amdgpu_mn_type type; /* only used on destruction */ struct work_struct work; @@ -185,7 +187,7 @@ static void amdgpu_mn_invalidate_node(struct amdgpu_mn_node *node, } /** - * amdgpu_mn_invalidate_range_start - callback to notify about mm change + * amdgpu_mn_invalidate_range_start_gfx - callback to notify about mm change * * @mn: our notifier * @mn: the mm this callback is about @@ -195,10 +197,10 @@ static void amdgpu_mn_invalidate_node(struct amdgpu_mn_node *node, * We block for all BOs between start and end to be idle and * unmap them by move them into system domain again. */ -static void amdgpu_mn_invalidate_range_start(struct mmu_notifier *mn, - struct mm_struct *mm, - unsigned long start, - unsigned long end) +static void amdgpu_mn_invalidate_range_start_gfx(struct mmu_notifier *mn, + struct mm_struct *mm, + unsigned long start, + unsigned long end) { struct amdgpu_mn *rmn = container_of(mn, struct amdgpu_mn, mn); struct interval_tree_node *it; @@ -220,6 +222,49 @@ static void amdgpu_mn_invalidate_range_start(struct mmu_notifier *mn, } /** + * amdgpu_mn_invalidate_range_start_hsa - callback to notify about mm change + * + * @mn: our notifier + * @mn: the mm this callback is about + * @start: start of updated range + * @end: end of updated range + * + * We temporarily evict all BOs between start and end. This + * necessitates evicting all user-mode queues of the process. The BOs + * are restorted in amdgpu_mn_invalidate_range_end_hsa. + */ +static void amdgpu_mn_invalidate_range_start_hsa(struct mmu_notifier *mn, + struct mm_struct *mm, + unsigned long start, + unsigned long end) +{ + struct amdgpu_mn *rmn = container_of(mn, struct amdgpu_mn, mn); + struct interval_tree_node *it; + + /* notification is exclusive, but interval is inclusive */ + end -= 1; + + amdgpu_mn_read_lock(rmn); + + it = interval_tree_iter_first(&rmn->objects, start, end); + while (it) { + struct amdgpu_mn_node *node; + struct amdgpu_bo *bo; + + node = container_of(it, struct amdgpu_mn_node, it); + it = interval_tree_iter_next(it, start, end); + + list_for_each_entry(bo, &node->bos, mn_list) { + struct kgd_mem *mem = bo->kfd_bo; + + if (amdgpu_ttm_tt_affect_userptr(bo->tbo.ttm, + start, end)) + amdgpu_amdkfd_evict_userptr(mem, mm); + } + } +} + +/** * amdgpu_mn_invalidate_range_end - callback to notify about mm change * * @mn: our notifier @@ -239,23 +284,39 @@ static void amdgpu_mn_invalidate_range_end(struct mmu_notifier *mn, amdgpu_mn_read_unlock(rmn); } -static const struct mmu_notifier_ops amdgpu_mn_ops = { - .release = amdgpu_mn_release, - .invalidate_range_start = amdgpu_mn_invalidate_range_start, - .invalidate_range_end = amdgpu_mn_invalidate_range_end, +static const struct mmu_notifier_ops amdgpu_mn_ops[] = { + [AMDGPU_MN_TYPE_GFX] = { + .release = amdgpu_mn_release, + .invalidate_range_start = amdgpu_mn_invalidate_range_start_gfx, + .invalidate_range_end = amdgpu_mn_invalidate_range_end, + }, + [AMDGPU_MN_TYPE_HSA] = { + .release = amdgpu_mn_release, + .invalidate_range_start = amdgpu_mn_invalidate_range_start_hsa, + .invalidate_range_end = amdgpu_mn_invalidate_range_end, + }, }; +/* Low bits of any reasonable mm pointer will be unused due to struct + * alignment. Use these bits to make a unique key from the mm pointer + * and notifier type. + */ +#define AMDGPU_MN_KEY(mm, type) ((unsigned long)(mm) + (type)) + /** * amdgpu_mn_get - create notifier context * * @adev: amdgpu device pointer + * @type: type of MMU notifier context * * Creates a notifier context for current->mm. */ -struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev) +struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev, + enum amdgpu_mn_type type) { struct mm_struct *mm = current->mm; struct amdgpu_mn *rmn; + unsigned long key = AMDGPU_MN_KEY(mm, type); int r; mutex_lock(&adev->mn_lock); @@ -264,8 +325,8 @@ struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev) return ERR_PTR(-EINTR); } - hash_for_each_possible(adev->mn_hash, rmn, node, (unsigned long)mm) - if (rmn->mm == mm) + hash_for_each_possible(adev->mn_hash, rmn, node, key) + if (AMDGPU_MN_KEY(rmn->mm, rmn->type) == key) goto release_locks; rmn = kzalloc(sizeof(*rmn), GFP_KERNEL); @@ -276,8 +337,9 @@ struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev) rmn->adev = adev; rmn->mm = mm; - rmn->mn.ops = &amdgpu_mn_ops; init_rwsem(&rmn->lock); + rmn->type = type; + rmn->mn.ops = &amdgpu_mn_ops[type]; rmn->objects = RB_ROOT_CACHED; mutex_init(&rmn->read_lock); atomic_set(&rmn->recursion, 0); @@ -286,7 +348,7 @@ struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev) if (r) goto free_rmn; - hash_add(adev->mn_hash, &rmn->node, (unsigned long)mm); + hash_add(adev->mn_hash, &rmn->node, AMDGPU_MN_KEY(mm, type)); release_locks: up_write(&mm->mmap_sem); @@ -315,12 +377,14 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr) { unsigned long end = addr + amdgpu_bo_size(bo) - 1; struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); + enum amdgpu_mn_type type = + bo->kfd_bo ? AMDGPU_MN_TYPE_HSA : AMDGPU_MN_TYPE_GFX; struct amdgpu_mn *rmn; struct amdgpu_mn_node *node = NULL; struct list_head bos; struct interval_tree_node *it; - rmn = amdgpu_mn_get(adev); + rmn = amdgpu_mn_get(adev, type); if (IS_ERR(rmn)) return PTR_ERR(rmn); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.h index d0095a3..eb0f432 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.h @@ -29,16 +29,23 @@ */ struct amdgpu_mn; +enum amdgpu_mn_type { + AMDGPU_MN_TYPE_GFX, + AMDGPU_MN_TYPE_HSA, +}; + #if defined(CONFIG_MMU_NOTIFIER) void amdgpu_mn_lock(struct amdgpu_mn *mn); void amdgpu_mn_unlock(struct amdgpu_mn *mn); -struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev); +struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev, + enum amdgpu_mn_type type); int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr); void amdgpu_mn_unregister(struct amdgpu_bo *bo); #else static inline void amdgpu_mn_lock(struct amdgpu_mn *mn) {} static inline void amdgpu_mn_unlock(struct amdgpu_mn *mn) {} -static inline struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev) +static inline struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev, + enum amdgpu_mn_type type) { return NULL; } -- cgit v1.1 From 0919195f2b0d7437cb0de49b8975fdd7b5575490 Mon Sep 17 00:00:00 2001 From: Felix Kuehling Date: Fri, 23 Mar 2018 15:32:29 -0400 Subject: drm/amdgpu: Enable amdgpu_ttm_tt_get_user_pages in worker threads This commit allows amdgpu_ttm_tt_get_user_pages to work in a worker thread rather than regular process context. This will be used when KFD userptr BOs are restored after an MMU-notifier eviction. v2: Manage task reference with get_task_struct/put_task_struct Signed-off-by: Felix Kuehling Acked-by: Oded Gabbay Acked-by: Alex Deucher Signed-off-by: Oded Gabbay --- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 38 +++++++++++++++++++++++++-------- 1 file changed, 29 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index 205da3f..c713d30 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -695,7 +695,7 @@ struct amdgpu_ttm_tt { struct ttm_dma_tt ttm; u64 offset; uint64_t userptr; - struct mm_struct *usermm; + struct task_struct *usertask; uint32_t userflags; spinlock_t guptasklock; struct list_head guptasks; @@ -706,14 +706,18 @@ struct amdgpu_ttm_tt { int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages) { struct amdgpu_ttm_tt *gtt = (void *)ttm; + struct mm_struct *mm = gtt->usertask->mm; unsigned int flags = 0; unsigned pinned = 0; int r; + if (!mm) /* Happens during process shutdown */ + return -ESRCH; + if (!(gtt->userflags & AMDGPU_GEM_USERPTR_READONLY)) flags |= FOLL_WRITE; - down_read(¤t->mm->mmap_sem); + down_read(&mm->mmap_sem); if (gtt->userflags & AMDGPU_GEM_USERPTR_ANONONLY) { /* check that we only use anonymous memory @@ -721,9 +725,9 @@ int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages) unsigned long end = gtt->userptr + ttm->num_pages * PAGE_SIZE; struct vm_area_struct *vma; - vma = find_vma(gtt->usermm, gtt->userptr); + vma = find_vma(mm, gtt->userptr); if (!vma || vma->vm_file || vma->vm_end < end) { - up_read(¤t->mm->mmap_sem); + up_read(&mm->mmap_sem); return -EPERM; } } @@ -739,7 +743,12 @@ int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages) list_add(&guptask.list, >t->guptasks); spin_unlock(>t->guptasklock); - r = get_user_pages(userptr, num_pages, flags, p, NULL); + if (mm == current->mm) + r = get_user_pages(userptr, num_pages, flags, p, NULL); + else + r = get_user_pages_remote(gtt->usertask, + mm, userptr, num_pages, + flags, p, NULL, NULL); spin_lock(>t->guptasklock); list_del(&guptask.list); @@ -752,12 +761,12 @@ int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages) } while (pinned < ttm->num_pages); - up_read(¤t->mm->mmap_sem); + up_read(&mm->mmap_sem); return 0; release_pages: release_pages(pages, pinned); - up_read(¤t->mm->mmap_sem); + up_read(&mm->mmap_sem); return r; } @@ -978,6 +987,9 @@ static void amdgpu_ttm_backend_destroy(struct ttm_tt *ttm) { struct amdgpu_ttm_tt *gtt = (void *)ttm; + if (gtt->usertask) + put_task_struct(gtt->usertask); + ttm_dma_tt_fini(>t->ttm); kfree(gtt); } @@ -1079,8 +1091,13 @@ int amdgpu_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr, return -EINVAL; gtt->userptr = addr; - gtt->usermm = current->mm; gtt->userflags = flags; + + if (gtt->usertask) + put_task_struct(gtt->usertask); + gtt->usertask = current->group_leader; + get_task_struct(gtt->usertask); + spin_lock_init(>t->guptasklock); INIT_LIST_HEAD(>t->guptasks); atomic_set(>t->mmu_invalidations, 0); @@ -1096,7 +1113,10 @@ struct mm_struct *amdgpu_ttm_tt_get_usermm(struct ttm_tt *ttm) if (gtt == NULL) return NULL; - return gtt->usermm; + if (gtt->usertask == NULL) + return NULL; + + return gtt->usertask->mm; } bool amdgpu_ttm_tt_affect_userptr(struct ttm_tt *ttm, unsigned long start, -- cgit v1.1 From 6e08e0995b8f339fd2a7ee4fa11f17396405ef60 Mon Sep 17 00:00:00 2001 From: Felix Kuehling Date: Fri, 23 Mar 2018 15:32:30 -0400 Subject: drm/amdgpu: Avoid reclaim while holding locks taken in MMU notifier MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When an MMU notifier runs in memory reclaim context, it can deadlock trying to take locks that are already held in the thread causing the memory reclaim. The solution is to avoid memory reclaim while holding locks that are taken in MMU notifiers. This commit fixes kmalloc while holding rmn->lock by moving the call outside the lock. The GFX MMU notifier also locks reservation objects. I have no good solution for avoiding reclaim while holding reservation objects. The HSA MMU notifier will not lock any reservation objects. v2: Moved allocation outside lock instead of using GFP_NOIO Signed-off-by: Felix Kuehling Acked-by: Oded Gabbay Reviewed-by: Christian König Signed-off-by: Oded Gabbay --- drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c index f2ed18e..83e344f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c @@ -380,7 +380,7 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr) enum amdgpu_mn_type type = bo->kfd_bo ? AMDGPU_MN_TYPE_HSA : AMDGPU_MN_TYPE_GFX; struct amdgpu_mn *rmn; - struct amdgpu_mn_node *node = NULL; + struct amdgpu_mn_node *node = NULL, *new_node; struct list_head bos; struct interval_tree_node *it; @@ -388,6 +388,10 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr) if (IS_ERR(rmn)) return PTR_ERR(rmn); + new_node = kmalloc(sizeof(*new_node), GFP_KERNEL); + if (!new_node) + return -ENOMEM; + INIT_LIST_HEAD(&bos); down_write(&rmn->lock); @@ -401,13 +405,10 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr) list_splice(&node->bos, &bos); } - if (!node) { - node = kmalloc(sizeof(struct amdgpu_mn_node), GFP_KERNEL); - if (!node) { - up_write(&rmn->lock); - return -ENOMEM; - } - } + if (!node) + node = new_node; + else + kfree(new_node); bo->mn = rmn; -- cgit v1.1 From d1853f42b63da94fa0147091d22bf5675b0ff89b Mon Sep 17 00:00:00 2001 From: Felix Kuehling Date: Fri, 23 Mar 2018 15:32:31 -0400 Subject: drm/amdkfd: GFP_NOIO while holding locks taken in MMU notifier When an MMU notifier runs in memory reclaim context, it can deadlock trying to take locks that are already held in the thread causing the memory reclaim. The solution is to avoid memory reclaim while holding locks that are taken in MMU notifiers by using GFP_NOIO. This commit fixes memory allocations done while holding the dqm->lock which is needed in the MMU notifier (dqm->ops.evict_process_queues). Signed-off-by: Felix Kuehling Reviewed-by: Oded Gabbay Signed-off-by: Oded Gabbay --- drivers/gpu/drm/amd/amdkfd/kfd_device.c | 2 +- drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c | 2 +- drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c index 334669996..0434f65 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c @@ -652,7 +652,7 @@ int kfd_gtt_sa_allocate(struct kfd_dev *kfd, unsigned int size, if (size > kfd->gtt_sa_num_of_chunks * kfd->gtt_sa_chunk_size) return -ENOMEM; - *mem_obj = kmalloc(sizeof(struct kfd_mem_obj), GFP_KERNEL); + *mem_obj = kmalloc(sizeof(struct kfd_mem_obj), GFP_NOIO); if ((*mem_obj) == NULL) return -ENOMEM; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c index c00c325..2bc49c6 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_cik.c @@ -412,7 +412,7 @@ struct mqd_manager *mqd_manager_init_cik(enum KFD_MQD_TYPE type, if (WARN_ON(type >= KFD_MQD_TYPE_MAX)) return NULL; - mqd = kzalloc(sizeof(*mqd), GFP_KERNEL); + mqd = kzalloc(sizeof(*mqd), GFP_NOIO); if (!mqd) return NULL; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c index 89e4242..481307b 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_vi.c @@ -394,7 +394,7 @@ struct mqd_manager *mqd_manager_init_vi(enum KFD_MQD_TYPE type, if (WARN_ON(type >= KFD_MQD_TYPE_MAX)) return NULL; - mqd = kzalloc(sizeof(*mqd), GFP_KERNEL); + mqd = kzalloc(sizeof(*mqd), GFP_NOIO); if (!mqd) return NULL; -- cgit v1.1 From 6b95e7973a136181e37446bd29b0b2e2f0d2d653 Mon Sep 17 00:00:00 2001 From: Felix Kuehling Date: Fri, 23 Mar 2018 15:32:32 -0400 Subject: drm/amdkfd: Add quiesce_mm and resume_mm to kgd2kfd_calls These interfaces allow KGD to stop and resume all GPU user mode queue access to a process address space. This is needed for handling MMU notifiers of userptrs mapped for GPU access in KFD VMs. Signed-off-by: Felix Kuehling Reviewed-by: Oded Gabbay Signed-off-by: Oded Gabbay --- drivers/gpu/drm/amd/amdkfd/kfd_device.c | 38 +++++++++++++++++++++++++ drivers/gpu/drm/amd/amdkfd/kfd_module.c | 2 ++ drivers/gpu/drm/amd/amdkfd/kfd_priv.h | 4 +++ drivers/gpu/drm/amd/amdkfd/kfd_process.c | 10 +++---- drivers/gpu/drm/amd/include/kgd_kfd_interface.h | 6 ++++ 5 files changed, 55 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c index 0434f65..7b57995 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c @@ -541,6 +541,44 @@ void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry) spin_unlock(&kfd->interrupt_lock); } +int kgd2kfd_quiesce_mm(struct mm_struct *mm) +{ + struct kfd_process *p; + int r; + + /* Because we are called from arbitrary context (workqueue) as opposed + * to process context, kfd_process could attempt to exit while we are + * running so the lookup function increments the process ref count. + */ + p = kfd_lookup_process_by_mm(mm); + if (!p) + return -ESRCH; + + r = kfd_process_evict_queues(p); + + kfd_unref_process(p); + return r; +} + +int kgd2kfd_resume_mm(struct mm_struct *mm) +{ + struct kfd_process *p; + int r; + + /* Because we are called from arbitrary context (workqueue) as opposed + * to process context, kfd_process could attempt to exit while we are + * running so the lookup function increments the process ref count. + */ + p = kfd_lookup_process_by_mm(mm); + if (!p) + return -ESRCH; + + r = kfd_process_restore_queues(p); + + kfd_unref_process(p); + return r; +} + /** kgd2kfd_schedule_evict_and_restore_process - Schedules work queue that will * prepare for safe eviction of KFD BOs that belong to the specified * process. diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_module.c b/drivers/gpu/drm/amd/amdkfd/kfd_module.c index e0c07d2..45bc458 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_module.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_module.c @@ -43,6 +43,8 @@ static const struct kgd2kfd_calls kgd2kfd = { .interrupt = kgd2kfd_interrupt, .suspend = kgd2kfd_suspend, .resume = kgd2kfd_resume, + .quiesce_mm = kgd2kfd_quiesce_mm, + .resume_mm = kgd2kfd_resume_mm, .schedule_evict_and_restore_process = kgd2kfd_schedule_evict_and_restore_process, }; diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h index 96a9cc0..4d5c49e 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h @@ -512,6 +512,8 @@ struct qcm_process_device { /* Approx. time before evicting the process again */ #define PROCESS_ACTIVE_TIME_MS 10 +int kgd2kfd_quiesce_mm(struct mm_struct *mm); +int kgd2kfd_resume_mm(struct mm_struct *mm); int kgd2kfd_schedule_evict_and_restore_process(struct mm_struct *mm, struct dma_fence *fence); @@ -681,6 +683,8 @@ struct kfd_process *kfd_get_process(const struct task_struct *); struct kfd_process *kfd_lookup_process_by_pasid(unsigned int pasid); struct kfd_process *kfd_lookup_process_by_mm(const struct mm_struct *mm); void kfd_unref_process(struct kfd_process *p); +int kfd_process_evict_queues(struct kfd_process *p); +int kfd_process_restore_queues(struct kfd_process *p); void kfd_suspend_all_processes(void); int kfd_resume_all_processes(void); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c index 1711ad0..2791e72 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c @@ -808,7 +808,7 @@ struct kfd_process *kfd_lookup_process_by_mm(const struct mm_struct *mm) * Eviction is reference-counted per process-device. This means multiple * evictions from different sources can be nested safely. */ -static int process_evict_queues(struct kfd_process *p) +int kfd_process_evict_queues(struct kfd_process *p) { struct kfd_process_device *pdd; int r = 0; @@ -844,7 +844,7 @@ fail: } /* process_restore_queues - Restore all user queues of a process */ -static int process_restore_queues(struct kfd_process *p) +int kfd_process_restore_queues(struct kfd_process *p) { struct kfd_process_device *pdd; int r, ret = 0; @@ -886,7 +886,7 @@ static void evict_process_worker(struct work_struct *work) flush_delayed_work(&p->restore_work); pr_debug("Started evicting pasid %d\n", p->pasid); - ret = process_evict_queues(p); + ret = kfd_process_evict_queues(p); if (!ret) { dma_fence_signal(p->ef); dma_fence_put(p->ef); @@ -946,7 +946,7 @@ static void restore_process_worker(struct work_struct *work) return; } - ret = process_restore_queues(p); + ret = kfd_process_restore_queues(p); if (!ret) pr_debug("Finished restoring pasid %d\n", p->pasid); else @@ -963,7 +963,7 @@ void kfd_suspend_all_processes(void) cancel_delayed_work_sync(&p->eviction_work); cancel_delayed_work_sync(&p->restore_work); - if (process_evict_queues(p)) + if (kfd_process_evict_queues(p)) pr_err("Failed to suspend process %d\n", p->pasid); dma_fence_signal(p->ef); dma_fence_put(p->ef); diff --git a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h index 237289a..286cfe7 100644 --- a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h +++ b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h @@ -382,6 +382,10 @@ struct kfd2kgd_calls { * * @resume: Notifies amdkfd about a resume action done to a kgd device * + * @quiesce_mm: Quiesce all user queue access to specified MM address space + * + * @resume_mm: Resume user queue access to specified MM address space + * * @schedule_evict_and_restore_process: Schedules work queue that will prepare * for safe eviction of KFD BOs that belong to the specified process. * @@ -399,6 +403,8 @@ struct kgd2kfd_calls { void (*interrupt)(struct kfd_dev *kfd, const void *ih_ring_entry); void (*suspend)(struct kfd_dev *kfd); int (*resume)(struct kfd_dev *kfd); + int (*quiesce_mm)(struct mm_struct *mm); + int (*resume_mm)(struct mm_struct *mm); int (*schedule_evict_and_restore_process)(struct mm_struct *mm, struct dma_fence *fence); }; -- cgit v1.1 From 5ae0283e831a94c714fce61063e4724baf364ef3 Mon Sep 17 00:00:00 2001 From: Felix Kuehling Date: Fri, 23 Mar 2018 15:32:33 -0400 Subject: drm/amdgpu: Add userptr support for KFD This adds support for allocating, mapping, unmapping and freeing userptr BOs, and for handling MMU notifiers. v2: updated a comment Signed-off-by: Felix Kuehling Reviewed-by: Oded Gabbay Signed-off-by: Oded Gabbay --- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h | 11 + drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 567 ++++++++++++++++++++++- 2 files changed, 554 insertions(+), 24 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h index 83e0c5c..c3024b1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h @@ -28,6 +28,7 @@ #include #include #include +#include #include #include #include "amdgpu_sync.h" @@ -59,7 +60,9 @@ struct kgd_mem { uint32_t mapping_flags; + atomic_t invalid; struct amdkfd_process_info *process_info; + struct page **user_pages; struct amdgpu_sync sync; @@ -84,6 +87,9 @@ struct amdkfd_process_info { struct list_head vm_list_head; /* List head for all KFD BOs that belong to a KFD process. */ struct list_head kfd_bo_list; + /* List of userptr BOs that are valid or invalid */ + struct list_head userptr_valid_list; + struct list_head userptr_inval_list; /* Lock to protect kfd_bo_list */ struct mutex lock; @@ -91,6 +97,11 @@ struct amdkfd_process_info { unsigned int n_vms; /* Eviction Fence */ struct amdgpu_amdkfd_fence *eviction_fence; + + /* MMU-notifier related fields */ + atomic_t evicted_bos; + struct delayed_work restore_userptr_work; + struct pid *pid; }; int amdgpu_amdkfd_init(void); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index 2463ff6..5296e24 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -23,6 +23,7 @@ #define pr_fmt(fmt) "kfd2kgd: " fmt #include +#include #include #include "amdgpu_object.h" #include "amdgpu_vm.h" @@ -33,10 +34,20 @@ */ #define VI_BO_SIZE_ALIGN (0x8000) +/* BO flag to indicate a KFD userptr BO */ +#define AMDGPU_AMDKFD_USERPTR_BO (1ULL << 63) + +/* Userptr restore delay, just long enough to allow consecutive VM + * changes to accumulate + */ +#define AMDGPU_USERPTR_RESTORE_DELAY_MS 1 + /* Impose limit on how much memory KFD can use */ static struct { uint64_t max_system_mem_limit; + uint64_t max_userptr_mem_limit; int64_t system_mem_used; + int64_t userptr_mem_used; spinlock_t mem_limit_lock; } kfd_mem_limit; @@ -57,6 +68,7 @@ static const char * const domain_bit_to_string[] = { #define domain_string(domain) domain_bit_to_string[ffs(domain)-1] +static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work); static inline struct amdgpu_device *get_amdgpu_device(struct kgd_dev *kgd) @@ -78,6 +90,7 @@ static bool check_if_add_bo_to_vm(struct amdgpu_vm *avm, /* Set memory usage limits. Current, limits are * System (kernel) memory - 3/8th System RAM + * Userptr memory - 3/4th System RAM */ void amdgpu_amdkfd_gpuvm_init_mem_limits(void) { @@ -90,8 +103,10 @@ void amdgpu_amdkfd_gpuvm_init_mem_limits(void) spin_lock_init(&kfd_mem_limit.mem_limit_lock); kfd_mem_limit.max_system_mem_limit = (mem >> 1) - (mem >> 3); - pr_debug("Kernel memory limit %lluM\n", - (kfd_mem_limit.max_system_mem_limit >> 20)); + kfd_mem_limit.max_userptr_mem_limit = mem - (mem >> 2); + pr_debug("Kernel memory limit %lluM, userptr limit %lluM\n", + (kfd_mem_limit.max_system_mem_limit >> 20), + (kfd_mem_limit.max_userptr_mem_limit >> 20)); } static int amdgpu_amdkfd_reserve_system_mem_limit(struct amdgpu_device *adev, @@ -111,6 +126,16 @@ static int amdgpu_amdkfd_reserve_system_mem_limit(struct amdgpu_device *adev, goto err_no_mem; } kfd_mem_limit.system_mem_used += (acc_size + size); + } else if (domain == AMDGPU_GEM_DOMAIN_CPU) { + if ((kfd_mem_limit.system_mem_used + acc_size > + kfd_mem_limit.max_system_mem_limit) || + (kfd_mem_limit.userptr_mem_used + (size + acc_size) > + kfd_mem_limit.max_userptr_mem_limit)) { + ret = -ENOMEM; + goto err_no_mem; + } + kfd_mem_limit.system_mem_used += acc_size; + kfd_mem_limit.userptr_mem_used += size; } err_no_mem: spin_unlock(&kfd_mem_limit.mem_limit_lock); @@ -126,10 +151,16 @@ static void unreserve_system_mem_limit(struct amdgpu_device *adev, sizeof(struct amdgpu_bo)); spin_lock(&kfd_mem_limit.mem_limit_lock); - if (domain == AMDGPU_GEM_DOMAIN_GTT) + if (domain == AMDGPU_GEM_DOMAIN_GTT) { kfd_mem_limit.system_mem_used -= (acc_size + size); + } else if (domain == AMDGPU_GEM_DOMAIN_CPU) { + kfd_mem_limit.system_mem_used -= acc_size; + kfd_mem_limit.userptr_mem_used -= size; + } WARN_ONCE(kfd_mem_limit.system_mem_used < 0, "kfd system memory accounting unbalanced"); + WARN_ONCE(kfd_mem_limit.userptr_mem_used < 0, + "kfd userptr memory accounting unbalanced"); spin_unlock(&kfd_mem_limit.mem_limit_lock); } @@ -138,12 +169,17 @@ void amdgpu_amdkfd_unreserve_system_memory_limit(struct amdgpu_bo *bo) { spin_lock(&kfd_mem_limit.mem_limit_lock); - if (bo->preferred_domains == AMDGPU_GEM_DOMAIN_GTT) { + if (bo->flags & AMDGPU_AMDKFD_USERPTR_BO) { + kfd_mem_limit.system_mem_used -= bo->tbo.acc_size; + kfd_mem_limit.userptr_mem_used -= amdgpu_bo_size(bo); + } else if (bo->preferred_domains == AMDGPU_GEM_DOMAIN_GTT) { kfd_mem_limit.system_mem_used -= (bo->tbo.acc_size + amdgpu_bo_size(bo)); } WARN_ONCE(kfd_mem_limit.system_mem_used < 0, "kfd system memory accounting unbalanced"); + WARN_ONCE(kfd_mem_limit.userptr_mem_used < 0, + "kfd userptr memory accounting unbalanced"); spin_unlock(&kfd_mem_limit.mem_limit_lock); } @@ -506,7 +542,8 @@ static void remove_bo_from_vm(struct amdgpu_device *adev, } static void add_kgd_mem_to_kfd_bo_list(struct kgd_mem *mem, - struct amdkfd_process_info *process_info) + struct amdkfd_process_info *process_info, + bool userptr) { struct ttm_validate_buffer *entry = &mem->validate_list; struct amdgpu_bo *bo = mem->bo; @@ -515,8 +552,93 @@ static void add_kgd_mem_to_kfd_bo_list(struct kgd_mem *mem, entry->shared = true; entry->bo = &bo->tbo; mutex_lock(&process_info->lock); - list_add_tail(&entry->head, &process_info->kfd_bo_list); + if (userptr) + list_add_tail(&entry->head, &process_info->userptr_valid_list); + else + list_add_tail(&entry->head, &process_info->kfd_bo_list); + mutex_unlock(&process_info->lock); +} + +/* Initializes user pages. It registers the MMU notifier and validates + * the userptr BO in the GTT domain. + * + * The BO must already be on the userptr_valid_list. Otherwise an + * eviction and restore may happen that leaves the new BO unmapped + * with the user mode queues running. + * + * Takes the process_info->lock to protect against concurrent restore + * workers. + * + * Returns 0 for success, negative errno for errors. + */ +static int init_user_pages(struct kgd_mem *mem, struct mm_struct *mm, + uint64_t user_addr) +{ + struct amdkfd_process_info *process_info = mem->process_info; + struct amdgpu_bo *bo = mem->bo; + struct ttm_operation_ctx ctx = { true, false }; + int ret = 0; + + mutex_lock(&process_info->lock); + + ret = amdgpu_ttm_tt_set_userptr(bo->tbo.ttm, user_addr, 0); + if (ret) { + pr_err("%s: Failed to set userptr: %d\n", __func__, ret); + goto out; + } + + ret = amdgpu_mn_register(bo, user_addr); + if (ret) { + pr_err("%s: Failed to register MMU notifier: %d\n", + __func__, ret); + goto out; + } + + /* If no restore worker is running concurrently, user_pages + * should not be allocated + */ + WARN(mem->user_pages, "Leaking user_pages array"); + + mem->user_pages = kvmalloc_array(bo->tbo.ttm->num_pages, + sizeof(struct page *), + GFP_KERNEL | __GFP_ZERO); + if (!mem->user_pages) { + pr_err("%s: Failed to allocate pages array\n", __func__); + ret = -ENOMEM; + goto unregister_out; + } + + ret = amdgpu_ttm_tt_get_user_pages(bo->tbo.ttm, mem->user_pages); + if (ret) { + pr_err("%s: Failed to get user pages: %d\n", __func__, ret); + goto free_out; + } + + amdgpu_ttm_tt_set_user_pages(bo->tbo.ttm, mem->user_pages); + + ret = amdgpu_bo_reserve(bo, true); + if (ret) { + pr_err("%s: Failed to reserve BO\n", __func__); + goto release_out; + } + amdgpu_ttm_placement_from_domain(bo, mem->domain); + ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); + if (ret) + pr_err("%s: failed to validate BO\n", __func__); + amdgpu_bo_unreserve(bo); + +release_out: + if (ret) + release_pages(mem->user_pages, bo->tbo.ttm->num_pages); +free_out: + kvfree(mem->user_pages); + mem->user_pages = NULL; +unregister_out: + if (ret) + amdgpu_mn_unregister(bo); +out: mutex_unlock(&process_info->lock); + return ret; } /* Reserving a BO and its page table BOs must happen atomically to @@ -748,7 +870,8 @@ static int update_gpuvm_pte(struct amdgpu_device *adev, } static int map_bo_to_gpuvm(struct amdgpu_device *adev, - struct kfd_bo_va_list *entry, struct amdgpu_sync *sync) + struct kfd_bo_va_list *entry, struct amdgpu_sync *sync, + bool no_update_pte) { int ret; @@ -762,6 +885,9 @@ static int map_bo_to_gpuvm(struct amdgpu_device *adev, return ret; } + if (no_update_pte) + return 0; + ret = update_gpuvm_pte(adev, entry, sync); if (ret) { pr_err("update_gpuvm_pte() failed\n"); @@ -820,6 +946,8 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info, mutex_init(&info->lock); INIT_LIST_HEAD(&info->vm_list_head); INIT_LIST_HEAD(&info->kfd_bo_list); + INIT_LIST_HEAD(&info->userptr_valid_list); + INIT_LIST_HEAD(&info->userptr_inval_list); info->eviction_fence = amdgpu_amdkfd_fence_create(dma_fence_context_alloc(1), @@ -830,6 +958,11 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info, goto create_evict_fence_fail; } + info->pid = get_task_pid(current->group_leader, PIDTYPE_PID); + atomic_set(&info->evicted_bos, 0); + INIT_DELAYED_WORK(&info->restore_userptr_work, + amdgpu_amdkfd_restore_userptr_worker); + *process_info = info; *ef = dma_fence_get(&info->eviction_fence->base); } @@ -872,6 +1005,7 @@ reserve_pd_fail: dma_fence_put(*ef); *ef = NULL; *process_info = NULL; + put_pid(info->pid); create_evict_fence_fail: mutex_destroy(&info->lock); kfree(info); @@ -967,8 +1101,12 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev, /* Release per-process resources when last compute VM is destroyed */ if (!process_info->n_vms) { WARN_ON(!list_empty(&process_info->kfd_bo_list)); + WARN_ON(!list_empty(&process_info->userptr_valid_list)); + WARN_ON(!list_empty(&process_info->userptr_inval_list)); dma_fence_put(&process_info->eviction_fence->base); + cancel_delayed_work_sync(&process_info->restore_userptr_work); + put_pid(process_info->pid); mutex_destroy(&process_info->lock); kfree(process_info); } @@ -1003,9 +1141,10 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( { struct amdgpu_device *adev = get_amdgpu_device(kgd); struct amdgpu_vm *avm = (struct amdgpu_vm *)vm; + uint64_t user_addr = 0; struct amdgpu_bo *bo; int byte_align; - u32 alloc_domain; + u32 domain, alloc_domain; u64 alloc_flags; uint32_t mapping_flags; int ret; @@ -1014,14 +1153,21 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( * Check on which domain to allocate BO */ if (flags & ALLOC_MEM_FLAGS_VRAM) { - alloc_domain = AMDGPU_GEM_DOMAIN_VRAM; + domain = alloc_domain = AMDGPU_GEM_DOMAIN_VRAM; alloc_flags = AMDGPU_GEM_CREATE_VRAM_CLEARED; alloc_flags |= (flags & ALLOC_MEM_FLAGS_PUBLIC) ? AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED : AMDGPU_GEM_CREATE_NO_CPU_ACCESS; } else if (flags & ALLOC_MEM_FLAGS_GTT) { - alloc_domain = AMDGPU_GEM_DOMAIN_GTT; + domain = alloc_domain = AMDGPU_GEM_DOMAIN_GTT; alloc_flags = 0; + } else if (flags & ALLOC_MEM_FLAGS_USERPTR) { + domain = AMDGPU_GEM_DOMAIN_GTT; + alloc_domain = AMDGPU_GEM_DOMAIN_CPU; + alloc_flags = 0; + if (!offset || !*offset) + return -EINVAL; + user_addr = *offset; } else { return -EINVAL; } @@ -1078,18 +1224,34 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu( } bo->kfd_bo = *mem; (*mem)->bo = bo; + if (user_addr) + bo->flags |= AMDGPU_AMDKFD_USERPTR_BO; (*mem)->va = va; - (*mem)->domain = alloc_domain; + (*mem)->domain = domain; (*mem)->mapped_to_gpu_memory = 0; (*mem)->process_info = avm->process_info; - add_kgd_mem_to_kfd_bo_list(*mem, avm->process_info); + add_kgd_mem_to_kfd_bo_list(*mem, avm->process_info, user_addr); + + if (user_addr) { + ret = init_user_pages(*mem, current->mm, user_addr); + if (ret) { + mutex_lock(&avm->process_info->lock); + list_del(&(*mem)->validate_list.head); + mutex_unlock(&avm->process_info->lock); + goto allocate_init_user_pages_failed; + } + } if (offset) *offset = amdgpu_bo_mmap_offset(bo); return 0; +allocate_init_user_pages_failed: + amdgpu_bo_unref(&bo); + /* Don't unreserve system mem limit twice */ + goto err_reserve_system_mem; err_bo_create: unreserve_system_mem_limit(adev, size, alloc_domain); err_reserve_system_mem: @@ -1122,12 +1284,24 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu( * be freed anyway */ + /* No more MMU notifiers */ + amdgpu_mn_unregister(mem->bo); + /* Make sure restore workers don't access the BO any more */ bo_list_entry = &mem->validate_list; mutex_lock(&process_info->lock); list_del(&bo_list_entry->head); mutex_unlock(&process_info->lock); + /* Free user pages if necessary */ + if (mem->user_pages) { + pr_debug("%s: Freeing user_pages array\n", __func__); + if (mem->user_pages[0]) + release_pages(mem->user_pages, + mem->bo->tbo.ttm->num_pages); + kvfree(mem->user_pages); + } + ret = reserve_bo_and_cond_vms(mem, NULL, BO_VM_ALL, &ctx); if (unlikely(ret)) return ret; @@ -1173,21 +1347,32 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu( struct kfd_bo_va_list *bo_va_entry = NULL; struct kfd_bo_va_list *bo_va_entry_aql = NULL; unsigned long bo_size; - - /* Make sure restore is not running concurrently. - */ - mutex_lock(&mem->process_info->lock); - - mutex_lock(&mem->lock); + bool is_invalid_userptr = false; bo = mem->bo; - if (!bo) { pr_err("Invalid BO when mapping memory to GPU\n"); - ret = -EINVAL; - goto out; + return -EINVAL; + } + + /* Make sure restore is not running concurrently. Since we + * don't map invalid userptr BOs, we rely on the next restore + * worker to do the mapping + */ + mutex_lock(&mem->process_info->lock); + + /* Lock mmap-sem. If we find an invalid userptr BO, we can be + * sure that the MMU notifier is no longer running + * concurrently and the queues are actually stopped + */ + if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) { + down_write(¤t->mm->mmap_sem); + is_invalid_userptr = atomic_read(&mem->invalid); + up_write(¤t->mm->mmap_sem); } + mutex_lock(&mem->lock); + domain = mem->domain; bo_size = bo->tbo.mem.size; @@ -1200,6 +1385,14 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu( if (unlikely(ret)) goto out; + /* Userptr can be marked as "not invalid", but not actually be + * validated yet (still in the system domain). In that case + * the queues are still stopped and we can leave mapping for + * the next restore worker + */ + if (bo->tbo.mem.mem_type == TTM_PL_SYSTEM) + is_invalid_userptr = true; + if (check_if_add_bo_to_vm(avm, mem)) { ret = add_bo_to_vm(adev, mem, avm, false, &bo_va_entry); @@ -1217,7 +1410,8 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu( goto add_bo_to_vm_failed; } - if (mem->mapped_to_gpu_memory == 0) { + if (mem->mapped_to_gpu_memory == 0 && + !amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) { /* Validate BO only once. The eviction fence gets added to BO * the first time it is mapped. Validate will wait for all * background evictions to complete. @@ -1235,7 +1429,8 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu( entry->va, entry->va + bo_size, entry); - ret = map_bo_to_gpuvm(adev, entry, ctx.sync); + ret = map_bo_to_gpuvm(adev, entry, ctx.sync, + is_invalid_userptr); if (ret) { pr_err("Failed to map radeon bo to gpuvm\n"); goto map_bo_to_gpuvm_failed; @@ -1418,13 +1613,337 @@ bo_reserve_failed: return ret; } +/* Evict a userptr BO by stopping the queues if necessary + * + * Runs in MMU notifier, may be in RECLAIM_FS context. This means it + * cannot do any memory allocations, and cannot take any locks that + * are held elsewhere while allocating memory. Therefore this is as + * simple as possible, using atomic counters. + * + * It doesn't do anything to the BO itself. The real work happens in + * restore, where we get updated page addresses. This function only + * ensures that GPU access to the BO is stopped. + */ int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm) { - /* TODO */ + struct amdkfd_process_info *process_info = mem->process_info; + int invalid, evicted_bos; + int r = 0; + + invalid = atomic_inc_return(&mem->invalid); + evicted_bos = atomic_inc_return(&process_info->evicted_bos); + if (evicted_bos == 1) { + /* First eviction, stop the queues */ + r = kgd2kfd->quiesce_mm(mm); + if (r) + pr_err("Failed to quiesce KFD\n"); + schedule_delayed_work(&process_info->restore_userptr_work, + msecs_to_jiffies(AMDGPU_USERPTR_RESTORE_DELAY_MS)); + } + + return r; +} + +/* Update invalid userptr BOs + * + * Moves invalidated (evicted) userptr BOs from userptr_valid_list to + * userptr_inval_list and updates user pages for all BOs that have + * been invalidated since their last update. + */ +static int update_invalid_user_pages(struct amdkfd_process_info *process_info, + struct mm_struct *mm) +{ + struct kgd_mem *mem, *tmp_mem; + struct amdgpu_bo *bo; + struct ttm_operation_ctx ctx = { false, false }; + int invalid, ret; + + /* Move all invalidated BOs to the userptr_inval_list and + * release their user pages by migration to the CPU domain + */ + list_for_each_entry_safe(mem, tmp_mem, + &process_info->userptr_valid_list, + validate_list.head) { + if (!atomic_read(&mem->invalid)) + continue; /* BO is still valid */ + + bo = mem->bo; + + if (amdgpu_bo_reserve(bo, true)) + return -EAGAIN; + amdgpu_ttm_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU); + ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); + amdgpu_bo_unreserve(bo); + if (ret) { + pr_err("%s: Failed to invalidate userptr BO\n", + __func__); + return -EAGAIN; + } + + list_move_tail(&mem->validate_list.head, + &process_info->userptr_inval_list); + } + + if (list_empty(&process_info->userptr_inval_list)) + return 0; /* All evicted userptr BOs were freed */ + + /* Go through userptr_inval_list and update any invalid user_pages */ + list_for_each_entry(mem, &process_info->userptr_inval_list, + validate_list.head) { + invalid = atomic_read(&mem->invalid); + if (!invalid) + /* BO hasn't been invalidated since the last + * revalidation attempt. Keep its BO list. + */ + continue; + + bo = mem->bo; + + if (!mem->user_pages) { + mem->user_pages = + kvmalloc_array(bo->tbo.ttm->num_pages, + sizeof(struct page *), + GFP_KERNEL | __GFP_ZERO); + if (!mem->user_pages) { + pr_err("%s: Failed to allocate pages array\n", + __func__); + return -ENOMEM; + } + } else if (mem->user_pages[0]) { + release_pages(mem->user_pages, bo->tbo.ttm->num_pages); + } + + /* Get updated user pages */ + ret = amdgpu_ttm_tt_get_user_pages(bo->tbo.ttm, + mem->user_pages); + if (ret) { + mem->user_pages[0] = NULL; + pr_info("%s: Failed to get user pages: %d\n", + __func__, ret); + /* Pretend it succeeded. It will fail later + * with a VM fault if the GPU tries to access + * it. Better than hanging indefinitely with + * stalled user mode queues. + */ + } + + /* Mark the BO as valid unless it was invalidated + * again concurrently + */ + if (atomic_cmpxchg(&mem->invalid, invalid, 0) != invalid) + return -EAGAIN; + } + return 0; } +/* Validate invalid userptr BOs + * + * Validates BOs on the userptr_inval_list, and moves them back to the + * userptr_valid_list. Also updates GPUVM page tables with new page + * addresses and waits for the page table updates to complete. + */ +static int validate_invalid_user_pages(struct amdkfd_process_info *process_info) +{ + struct amdgpu_bo_list_entry *pd_bo_list_entries; + struct list_head resv_list, duplicates; + struct ww_acquire_ctx ticket; + struct amdgpu_sync sync; + + struct amdgpu_vm *peer_vm; + struct kgd_mem *mem, *tmp_mem; + struct amdgpu_bo *bo; + struct ttm_operation_ctx ctx = { false, false }; + int i, ret; + + pd_bo_list_entries = kcalloc(process_info->n_vms, + sizeof(struct amdgpu_bo_list_entry), + GFP_KERNEL); + if (!pd_bo_list_entries) { + pr_err("%s: Failed to allocate PD BO list entries\n", __func__); + return -ENOMEM; + } + + INIT_LIST_HEAD(&resv_list); + INIT_LIST_HEAD(&duplicates); + + /* Get all the page directory BOs that need to be reserved */ + i = 0; + list_for_each_entry(peer_vm, &process_info->vm_list_head, + vm_list_node) + amdgpu_vm_get_pd_bo(peer_vm, &resv_list, + &pd_bo_list_entries[i++]); + /* Add the userptr_inval_list entries to resv_list */ + list_for_each_entry(mem, &process_info->userptr_inval_list, + validate_list.head) { + list_add_tail(&mem->resv_list.head, &resv_list); + mem->resv_list.bo = mem->validate_list.bo; + mem->resv_list.shared = mem->validate_list.shared; + } + + /* Reserve all BOs and page tables for validation */ + ret = ttm_eu_reserve_buffers(&ticket, &resv_list, false, &duplicates); + WARN(!list_empty(&duplicates), "Duplicates should be empty"); + if (ret) + goto out; + + amdgpu_sync_create(&sync); + + /* Avoid triggering eviction fences when unmapping invalid + * userptr BOs (waits for all fences, doesn't use + * FENCE_OWNER_VM) + */ + list_for_each_entry(peer_vm, &process_info->vm_list_head, + vm_list_node) + amdgpu_amdkfd_remove_eviction_fence(peer_vm->root.base.bo, + process_info->eviction_fence, + NULL, NULL); + + ret = process_validate_vms(process_info); + if (ret) + goto unreserve_out; + + /* Validate BOs and update GPUVM page tables */ + list_for_each_entry_safe(mem, tmp_mem, + &process_info->userptr_inval_list, + validate_list.head) { + struct kfd_bo_va_list *bo_va_entry; + + bo = mem->bo; + + /* Copy pages array and validate the BO if we got user pages */ + if (mem->user_pages[0]) { + amdgpu_ttm_tt_set_user_pages(bo->tbo.ttm, + mem->user_pages); + amdgpu_ttm_placement_from_domain(bo, mem->domain); + ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); + if (ret) { + pr_err("%s: failed to validate BO\n", __func__); + goto unreserve_out; + } + } + + /* Validate succeeded, now the BO owns the pages, free + * our copy of the pointer array. Put this BO back on + * the userptr_valid_list. If we need to revalidate + * it, we need to start from scratch. + */ + kvfree(mem->user_pages); + mem->user_pages = NULL; + list_move_tail(&mem->validate_list.head, + &process_info->userptr_valid_list); + + /* Update mapping. If the BO was not validated + * (because we couldn't get user pages), this will + * clear the page table entries, which will result in + * VM faults if the GPU tries to access the invalid + * memory. + */ + list_for_each_entry(bo_va_entry, &mem->bo_va_list, bo_list) { + if (!bo_va_entry->is_mapped) + continue; + + ret = update_gpuvm_pte((struct amdgpu_device *) + bo_va_entry->kgd_dev, + bo_va_entry, &sync); + if (ret) { + pr_err("%s: update PTE failed\n", __func__); + /* make sure this gets validated again */ + atomic_inc(&mem->invalid); + goto unreserve_out; + } + } + } + + /* Update page directories */ + ret = process_update_pds(process_info, &sync); + +unreserve_out: + list_for_each_entry(peer_vm, &process_info->vm_list_head, + vm_list_node) + amdgpu_bo_fence(peer_vm->root.base.bo, + &process_info->eviction_fence->base, true); + ttm_eu_backoff_reservation(&ticket, &resv_list); + amdgpu_sync_wait(&sync, false); + amdgpu_sync_free(&sync); +out: + kfree(pd_bo_list_entries); + + return ret; +} + +/* Worker callback to restore evicted userptr BOs + * + * Tries to update and validate all userptr BOs. If successful and no + * concurrent evictions happened, the queues are restarted. Otherwise, + * reschedule for another attempt later. + */ +static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work) +{ + struct delayed_work *dwork = to_delayed_work(work); + struct amdkfd_process_info *process_info = + container_of(dwork, struct amdkfd_process_info, + restore_userptr_work); + struct task_struct *usertask; + struct mm_struct *mm; + int evicted_bos; + + evicted_bos = atomic_read(&process_info->evicted_bos); + if (!evicted_bos) + return; + + /* Reference task and mm in case of concurrent process termination */ + usertask = get_pid_task(process_info->pid, PIDTYPE_PID); + if (!usertask) + return; + mm = get_task_mm(usertask); + if (!mm) { + put_task_struct(usertask); + return; + } + + mutex_lock(&process_info->lock); + + if (update_invalid_user_pages(process_info, mm)) + goto unlock_out; + /* userptr_inval_list can be empty if all evicted userptr BOs + * have been freed. In that case there is nothing to validate + * and we can just restart the queues. + */ + if (!list_empty(&process_info->userptr_inval_list)) { + if (atomic_read(&process_info->evicted_bos) != evicted_bos) + goto unlock_out; /* Concurrent eviction, try again */ + + if (validate_invalid_user_pages(process_info)) + goto unlock_out; + } + /* Final check for concurrent evicton and atomic update. If + * another eviction happens after successful update, it will + * be a first eviction that calls quiesce_mm. The eviction + * reference counting inside KFD will handle this case. + */ + if (atomic_cmpxchg(&process_info->evicted_bos, evicted_bos, 0) != + evicted_bos) + goto unlock_out; + evicted_bos = 0; + if (kgd2kfd->resume_mm(mm)) { + pr_err("%s: Failed to resume KFD\n", __func__); + /* No recovery from this failure. Probably the CP is + * hanging. No point trying again. + */ + } +unlock_out: + mutex_unlock(&process_info->lock); + mmput(mm); + put_task_struct(usertask); + + /* If validation failed, reschedule another attempt */ + if (evicted_bos) + schedule_delayed_work(&process_info->restore_userptr_work, + msecs_to_jiffies(AMDGPU_USERPTR_RESTORE_DELAY_MS)); +} + /** amdgpu_amdkfd_gpuvm_restore_process_bos - Restore all BOs for the given * KFD process identified by process_info * -- cgit v1.1 From 5bb975de3f279c6577fb54334cdd7e55c47a362c Mon Sep 17 00:00:00 2001 From: Manasi Navare Date: Fri, 23 Mar 2018 10:24:13 -0700 Subject: drm/i915/icl: Add register definitions for Combo PHY vswing sequences. This patch defines register definitions required for ICL voltage vswing programming for Combo PHY DDI Ports. It uses the same bit definitions and macros as the CNL voltage swing sequences. v8 (from Paulo): * Rebase. v7: * Kill _MMIIO_PORT2_LN (Paulo) v6: * Replace some spaces with TAB (Paulo) v5: * Use _PORT instead of _PICK (Paulo) * Remove DW7 defs for ICL, not used (Paulo) v4: * Rebase after _PICK was used instead of _PORT3 * Use _PICK for _MMIO_PORT2 since address of B is less than address of A so cant use the math (Paulo) v3: * Make changes to the existing macro in a diff patch (Paulo) v2: * Add new defs fro ICL regs (Paulo) Cc: Jani Nikula Cc: Rodrigo Vivi Reviewed-by: Paulo Zanoni Signed-off-by: Manasi Navare Signed-off-by: Paulo Zanoni Link: https://patchwork.freedesktop.org/patch/msgid/20180323172419.24911-2-paulo.r.zanoni@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 44 +++++++++++++++++++++++++++++++++++++++++ 1 file changed, 44 insertions(+) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 4e31dff..407ee5c 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -1697,6 +1697,16 @@ enum i915_power_well_id { _CNL_PORT_PCS_DW1_LN0_D, \ _CNL_PORT_PCS_DW1_LN0_AE, \ _CNL_PORT_PCS_DW1_LN0_F)) +#define _ICL_PORT_PCS_DW1_GRP_A 0x162604 +#define _ICL_PORT_PCS_DW1_GRP_B 0x6C604 +#define _ICL_PORT_PCS_DW1_LN0_A 0x162804 +#define _ICL_PORT_PCS_DW1_LN0_B 0x6C804 +#define ICL_PORT_PCS_DW1_GRP(port) _MMIO_PORT(port,\ + _ICL_PORT_PCS_DW1_GRP_A, \ + _ICL_PORT_PCS_DW1_GRP_B) +#define ICL_PORT_PCS_DW1_LN0(port) _MMIO_PORT(port, \ + _ICL_PORT_PCS_DW1_LN0_A, \ + _ICL_PORT_PCS_DW1_LN0_B) #define COMMON_KEEPER_EN (1 << 26) /* CNL Port TX registers */ @@ -1729,6 +1739,16 @@ enum i915_power_well_id { #define CNL_PORT_TX_DW2_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP((port), 2)) #define CNL_PORT_TX_DW2_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0((port), 2)) +#define _ICL_PORT_TX_DW2_GRP_A 0x162688 +#define _ICL_PORT_TX_DW2_GRP_B 0x6C688 +#define _ICL_PORT_TX_DW2_LN0_A 0x162888 +#define _ICL_PORT_TX_DW2_LN0_B 0x6C888 +#define ICL_PORT_TX_DW2_GRP(port) _MMIO_PORT(port, \ + _ICL_PORT_TX_DW2_GRP_A, \ + _ICL_PORT_TX_DW2_GRP_B) +#define ICL_PORT_TX_DW2_LN0(port) _MMIO_PORT(port, \ + _ICL_PORT_TX_DW2_LN0_A, \ + _ICL_PORT_TX_DW2_LN0_B) #define SWING_SEL_UPPER(x) ((x >> 3) << 15) #define SWING_SEL_UPPER_MASK (1 << 15) #define SWING_SEL_LOWER(x) ((x & 0x7) << 11) @@ -1743,6 +1763,19 @@ enum i915_power_well_id { #define CNL_PORT_TX_DW4_LN(port, ln) _MMIO(_CNL_PORT_TX_DW_LN0((port), 4) + \ (ln * (_CNL_PORT_TX_DW4_LN1_AE - \ _CNL_PORT_TX_DW4_LN0_AE))) +#define _ICL_PORT_TX_DW4_GRP_A 0x162690 +#define _ICL_PORT_TX_DW4_GRP_B 0x6C690 +#define _ICL_PORT_TX_DW4_LN0_A 0x162890 +#define _ICL_PORT_TX_DW4_LN1_A 0x162990 +#define _ICL_PORT_TX_DW4_LN0_B 0x6C890 +#define ICL_PORT_TX_DW4_GRP(port) _MMIO_PORT(port, \ + _ICL_PORT_TX_DW4_GRP_A, \ + _ICL_PORT_TX_DW4_GRP_B) +#define ICL_PORT_TX_DW4_LN(port, ln) _MMIO(_PORT(port, \ + _ICL_PORT_TX_DW4_LN0_A, \ + _ICL_PORT_TX_DW4_LN0_B) + \ + (ln * (_ICL_PORT_TX_DW4_LN1_A - \ + _ICL_PORT_TX_DW4_LN0_A))) #define LOADGEN_SELECT (1 << 31) #define POST_CURSOR_1(x) ((x) << 12) #define POST_CURSOR_1_MASK (0x3F << 12) @@ -1753,7 +1786,18 @@ enum i915_power_well_id { #define CNL_PORT_TX_DW5_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP((port), 5)) #define CNL_PORT_TX_DW5_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0((port), 5)) +#define _ICL_PORT_TX_DW5_GRP_A 0x162694 +#define _ICL_PORT_TX_DW5_GRP_B 0x6C694 +#define _ICL_PORT_TX_DW5_LN0_A 0x162894 +#define _ICL_PORT_TX_DW5_LN0_B 0x6C894 +#define ICL_PORT_TX_DW5_GRP(port) _MMIO_PORT(port, \ + _ICL_PORT_TX_DW5_GRP_A, \ + _ICL_PORT_TX_DW5_GRP_B) +#define ICL_PORT_TX_DW5_LN0(port) _MMIO_PORT(port, \ + _ICL_PORT_TX_DW5_LN0_A, \ + _ICL_PORT_TX_DW5_LN0_B) #define TX_TRAINING_EN (1 << 31) +#define TAP2_DISABLE (1 << 30) #define TAP3_DISABLE (1 << 29) #define SCALING_MODE_SEL(x) ((x) << 18) #define SCALING_MODE_SEL_MASK (0x7 << 18) -- cgit v1.1 From 19b904f8df5c6c1418769de35bf238ef72e49814 Mon Sep 17 00:00:00 2001 From: Manasi Navare Date: Fri, 23 Mar 2018 10:24:14 -0700 Subject: drm/i915/icl: Add Combo PHY DDI Buffer translation tables for Icelake. These tables are used on voltage vswing sequence initialization on Icelake. The swing_sel on the spec's table is defined in a 4 bits binary like 1010. However the register bits are split in upper 1 bit swing_sel and lower 3 bits swing sel. In this table here we store this value as a single value in hex like it is mentioned in the Bspec and split it to the upper and lower bit values only while programming the registers. For instance: b1010 is written as 0xA and then while writing to the register, the upper 1 bit is obtained by (0xA & 0x8) and shifting by appropriate bits while lower 3 bits are obtained by (0xA & 0x7) and shifting by appropriate bits. Some of the columns need to be updated after the spec is updated. v5 (from Paulo): * Checkpatch fixes. v4 (from Paulo): * Fix minor typo * Coding style conformance v3: * Get rid of HDMI/DVI tables, same as DP (Paulo) * Use combo_phy in ddi buf trans table defs (Paulo) v2: * Added DW4_scaling_hex column to the translation tables (Rodrigo) Cc: Jani Nikula Cc: Rodrigo Vivi Reviewed-by: Paulo Zanoni Signed-off-by: Manasi Navare Signed-off-by: Paulo Zanoni Link: https://patchwork.freedesktop.org/patch/msgid/20180323172419.24911-3-paulo.r.zanoni@intel.com --- drivers/gpu/drm/i915/intel_ddi.c | 99 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 99 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index c449619..229b9d5 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -493,6 +493,105 @@ static const struct cnl_ddi_buf_trans cnl_ddi_translations_edp_1_05V[] = { { 0x2, 0x7F, 0x3F, 0x00, 0x00 }, /* 400 400 0.0 */ }; +struct icl_combo_phy_ddi_buf_trans { + u32 dw2_swing_select; + u32 dw2_swing_scalar; + u32 dw4_scaling; +}; + +/* Voltage Swing Programming for VccIO 0.85V for DP */ +static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_dp_hdmi_0_85V[] = { + /* Voltage mV db */ + { 0x2, 0x98, 0x0018 }, /* 400 0.0 */ + { 0x2, 0x98, 0x3015 }, /* 400 3.5 */ + { 0x2, 0x98, 0x6012 }, /* 400 6.0 */ + { 0x2, 0x98, 0x900F }, /* 400 9.5 */ + { 0xB, 0x70, 0x0018 }, /* 600 0.0 */ + { 0xB, 0x70, 0x3015 }, /* 600 3.5 */ + { 0xB, 0x70, 0x6012 }, /* 600 6.0 */ + { 0x5, 0x00, 0x0018 }, /* 800 0.0 */ + { 0x5, 0x00, 0x3015 }, /* 800 3.5 */ + { 0x6, 0x98, 0x0018 }, /* 1200 0.0 */ +}; + +/* FIXME - After table is updated in Bspec */ +/* Voltage Swing Programming for VccIO 0.85V for eDP */ +static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_edp_0_85V[] = { + /* Voltage mV db */ + { 0x0, 0x00, 0x00 }, /* 200 0.0 */ + { 0x0, 0x00, 0x00 }, /* 200 1.5 */ + { 0x0, 0x00, 0x00 }, /* 200 4.0 */ + { 0x0, 0x00, 0x00 }, /* 200 6.0 */ + { 0x0, 0x00, 0x00 }, /* 250 0.0 */ + { 0x0, 0x00, 0x00 }, /* 250 1.5 */ + { 0x0, 0x00, 0x00 }, /* 250 4.0 */ + { 0x0, 0x00, 0x00 }, /* 300 0.0 */ + { 0x0, 0x00, 0x00 }, /* 300 1.5 */ + { 0x0, 0x00, 0x00 }, /* 350 0.0 */ +}; + +/* Voltage Swing Programming for VccIO 0.95V for DP */ +static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_dp_hdmi_0_95V[] = { + /* Voltage mV db */ + { 0x2, 0x98, 0x0018 }, /* 400 0.0 */ + { 0x2, 0x98, 0x3015 }, /* 400 3.5 */ + { 0x2, 0x98, 0x6012 }, /* 400 6.0 */ + { 0x2, 0x98, 0x900F }, /* 400 9.5 */ + { 0x4, 0x98, 0x0018 }, /* 600 0.0 */ + { 0x4, 0x98, 0x3015 }, /* 600 3.5 */ + { 0x4, 0x98, 0x6012 }, /* 600 6.0 */ + { 0x5, 0x76, 0x0018 }, /* 800 0.0 */ + { 0x5, 0x76, 0x3015 }, /* 800 3.5 */ + { 0x6, 0x98, 0x0018 }, /* 1200 0.0 */ +}; + +/* FIXME - After table is updated in Bspec */ +/* Voltage Swing Programming for VccIO 0.95V for eDP */ +static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_edp_0_95V[] = { + /* Voltage mV db */ + { 0x0, 0x00, 0x00 }, /* 200 0.0 */ + { 0x0, 0x00, 0x00 }, /* 200 1.5 */ + { 0x0, 0x00, 0x00 }, /* 200 4.0 */ + { 0x0, 0x00, 0x00 }, /* 200 6.0 */ + { 0x0, 0x00, 0x00 }, /* 250 0.0 */ + { 0x0, 0x00, 0x00 }, /* 250 1.5 */ + { 0x0, 0x00, 0x00 }, /* 250 4.0 */ + { 0x0, 0x00, 0x00 }, /* 300 0.0 */ + { 0x0, 0x00, 0x00 }, /* 300 1.5 */ + { 0x0, 0x00, 0x00 }, /* 350 0.0 */ +}; + +/* Voltage Swing Programming for VccIO 1.05V for DP */ +static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_dp_hdmi_1_05V[] = { + /* Voltage mV db */ + { 0x2, 0x98, 0x0018 }, /* 400 0.0 */ + { 0x2, 0x98, 0x3015 }, /* 400 3.5 */ + { 0x2, 0x98, 0x6012 }, /* 400 6.0 */ + { 0x2, 0x98, 0x900F }, /* 400 9.5 */ + { 0x4, 0x98, 0x0018 }, /* 600 0.0 */ + { 0x4, 0x98, 0x3015 }, /* 600 3.5 */ + { 0x4, 0x98, 0x6012 }, /* 600 6.0 */ + { 0x5, 0x71, 0x0018 }, /* 800 0.0 */ + { 0x5, 0x71, 0x3015 }, /* 800 3.5 */ + { 0x6, 0x98, 0x0018 }, /* 1200 0.0 */ +}; + +/* FIXME - After table is updated in Bspec */ +/* Voltage Swing Programming for VccIO 1.05V for eDP */ +static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_edp_1_05V[] = { + /* Voltage mV db */ + { 0x0, 0x00, 0x00 }, /* 200 0.0 */ + { 0x0, 0x00, 0x00 }, /* 200 1.5 */ + { 0x0, 0x00, 0x00 }, /* 200 4.0 */ + { 0x0, 0x00, 0x00 }, /* 200 6.0 */ + { 0x0, 0x00, 0x00 }, /* 250 0.0 */ + { 0x0, 0x00, 0x00 }, /* 250 1.5 */ + { 0x0, 0x00, 0x00 }, /* 250 4.0 */ + { 0x0, 0x00, 0x00 }, /* 300 0.0 */ + { 0x0, 0x00, 0x00 }, /* 300 1.5 */ + { 0x0, 0x00, 0x00 }, /* 350 0.0 */ +}; + static const struct ddi_buf_trans * bdw_get_buf_trans_edp(struct drm_i915_private *dev_priv, int *n_entries) { -- cgit v1.1 From c92f47b5ec977a31c72a3c3514ae460b3dd725ff Mon Sep 17 00:00:00 2001 From: Manasi Navare Date: Fri, 23 Mar 2018 10:24:15 -0700 Subject: drm/i915/icl: Add register defs for voltage swing sequences for MG PHY DDI On Icelake platform, MG PHY is used when operating in DP alternate mode or the legacy HDMI or DP modes. DDI Ports C, D, E, F are MG PHY DDI ports on ICL. This patch adds the necessary voltage swing programming related register definitions and macros for MG PHY DDI ports. v4 (from Paulo): * Use _PORT instead of _PICK * Change some mask names to our current coding standards * Stay under 80 columns v3: * Rebase on new revision of patches v2: * Remove whitespaces in the #defines (Paulo) Cc: Rodrigo Vivi Cc: Jani Nikula Reviewed-by: Paulo Zanoni Signed-off-by: Manasi Navare Signed-off-by: Paulo Zanoni Link: https://patchwork.freedesktop.org/patch/msgid/20180323172419.24911-4-paulo.r.zanoni@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 116 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 116 insertions(+) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 407ee5c..aa001dd 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -1809,6 +1809,122 @@ enum i915_power_well_id { #define N_SCALAR(x) ((x) << 24) #define N_SCALAR_MASK (0x7F << 24) +#define _ICL_MG_PHY_PORT_LN(port, ln, ln0p1, ln0p2, ln1p1) \ + _MMIO(_PORT((port) - PORT_C, ln0p1, ln0p2) + (ln) * ((ln1p1) - (ln0p1))) + +#define _ICL_MG_TX_LINK_PARAMS_TX1LN0_PORT1 0x16812C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN1_PORT1 0x16852C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN0_PORT2 0x16912C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN1_PORT2 0x16952C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN0_PORT3 0x16A12C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN1_PORT3 0x16A52C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN0_PORT4 0x16B12C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN1_PORT4 0x16B52C +#define ICL_PORT_MG_TX1_LINK_PARAMS(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_LINK_PARAMS_TX1LN0_PORT1, \ + _ICL_MG_TX_LINK_PARAMS_TX1LN0_PORT2, \ + _ICL_MG_TX_LINK_PARAMS_TX1LN1_PORT1) + +#define _ICL_MG_TX_LINK_PARAMS_TX2LN0_PORT1 0x1680AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN1_PORT1 0x1684AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN0_PORT2 0x1690AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN1_PORT2 0x1694AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN0_PORT3 0x16A0AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN1_PORT3 0x16A4AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN0_PORT4 0x16B0AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN1_PORT4 0x16B4AC +#define ICL_PORT_MG_TX2_LINK_PARAMS(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_LINK_PARAMS_TX2LN0_PORT1, \ + _ICL_MG_TX_LINK_PARAMS_TX2LN0_PORT2, \ + _ICL_MG_TX_LINK_PARAMS_TX2LN1_PORT1) +#define CRI_USE_FS32 (1 << 5) + +#define _ICL_MG_TX_PISO_READLOAD_TX1LN0_PORT1 0x16814C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN1_PORT1 0x16854C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN0_PORT2 0x16914C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN1_PORT2 0x16954C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN0_PORT3 0x16A14C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN1_PORT3 0x16A54C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN0_PORT4 0x16B14C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN1_PORT4 0x16B54C +#define ICL_PORT_MG_TX1_PISO_READLOAD(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_PISO_READLOAD_TX1LN0_PORT1, \ + _ICL_MG_TX_PISO_READLOAD_TX1LN0_PORT2, \ + _ICL_MG_TX_PISO_READLOAD_TX1LN1_PORT1) + +#define _ICL_MG_TX_PISO_READLOAD_TX2LN0_PORT1 0x1680CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN1_PORT1 0x1684CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN0_PORT2 0x1690CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN1_PORT2 0x1694CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN0_PORT3 0x16A0CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN1_PORT3 0x16A4CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN0_PORT4 0x16B0CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN1_PORT4 0x16B4CC +#define ICL_PORT_MG_TX2_PISO_READLOAD(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_PISO_READLOAD_TX2LN0_PORT1, \ + _ICL_MG_TX_PISO_READLOAD_TX2LN0_PORT2, \ + _ICL_MG_TX_PISO_READLOAD_TX2LN1_PORT1) +#define CRI_CALCINIT (1 << 1) + +#define _ICL_MG_TX_SWINGCTRL_TX1LN0_PORT1 0x168148 +#define _ICL_MG_TX_SWINGCTRL_TX1LN1_PORT1 0x168548 +#define _ICL_MG_TX_SWINGCTRL_TX1LN0_PORT2 0x169148 +#define _ICL_MG_TX_SWINGCTRL_TX1LN1_PORT2 0x169548 +#define _ICL_MG_TX_SWINGCTRL_TX1LN0_PORT3 0x16A148 +#define _ICL_MG_TX_SWINGCTRL_TX1LN1_PORT3 0x16A548 +#define _ICL_MG_TX_SWINGCTRL_TX1LN0_PORT4 0x16B148 +#define _ICL_MG_TX_SWINGCTRL_TX1LN1_PORT4 0x16B548 +#define ICL_PORT_MG_TX1_SWINGCTRL(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_SWINGCTRL_TX1LN0_PORT1, \ + _ICL_MG_TX_SWINGCTRL_TX1LN0_PORT2, \ + _ICL_MG_TX_SWINGCTRL_TX1LN1_PORT1) + +#define _ICL_MG_TX_SWINGCTRL_TX2LN0_PORT1 0x1680C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN1_PORT1 0x1684C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN0_PORT2 0x1690C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN1_PORT2 0x1694C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN0_PORT3 0x16A0C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN1_PORT3 0x16A4C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN0_PORT4 0x16B0C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN1_PORT4 0x16B4C8 +#define ICL_PORT_MG_TX2_SWINGCTRL(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_SWINGCTRL_TX2LN0_PORT1, \ + _ICL_MG_TX_SWINGCTRL_TX2LN0_PORT2, \ + _ICL_MG_TX_SWINGCTRL_TX2LN1_PORT1) +#define CRI_TXDEEMPH_OVERRIDE_17_12(x) ((x) << 0) +#define CRI_TXDEEMPH_OVERRIDE_17_12_MASK (0x3F << 0) + +#define _ICL_MG_TX_DRVCTRL_TX1LN0_PORT1 0x168144 +#define _ICL_MG_TX_DRVCTRL_TX1LN1_PORT1 0x168544 +#define _ICL_MG_TX_DRVCTRL_TX1LN0_PORT2 0x169144 +#define _ICL_MG_TX_DRVCTRL_TX1LN1_PORT2 0x169544 +#define _ICL_MG_TX_DRVCTRL_TX1LN0_PORT3 0x16A144 +#define _ICL_MG_TX_DRVCTRL_TX1LN1_PORT3 0x16A544 +#define _ICL_MG_TX_DRVCTRL_TX1LN0_PORT4 0x16B144 +#define _ICL_MG_TX_DRVCTRL_TX1LN1_PORT4 0x16B544 +#define ICL_PORT_MG_TX1_DRVCTRL(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_DRVCTRL_TX1LN0_PORT1, \ + _ICL_MG_TX_DRVCTRL_TX1LN0_PORT2, \ + _ICL_MG_TX_DRVCTRL_TX1LN1_PORT1) + +#define _ICL_MG_TX_DRVCTRL_TX2LN0_PORT1 0x1680C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN1_PORT1 0x1684C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN0_PORT2 0x1690C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN1_PORT2 0x1694C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN0_PORT3 0x16A0C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN1_PORT3 0x16A4C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN0_PORT4 0x16B0C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN1_PORT4 0x16B4C4 +#define ICL_PORT_MG_TX2_DRVCTRL(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_DRVCTRL_TX2LN0_PORT1, \ + _ICL_MG_TX_DRVCTRL_TX2LN0_PORT2, \ + _ICL_MG_TX_DRVCTRL_TX2LN1_PORT1) +#define CRI_TXDEEMPH_OVERRIDE_11_6(x) ((x) << 24) +#define CRI_TXDEEMPH_OVERRIDE_11_6_MASK (0x3F << 24) +#define CRI_TXDEEMPH_OVERRIDE_EN (1 << 22) +#define CRI_TXDEEMPH_OVERRIDE_5_0(x) ((x) << 16) +#define CRI_TXDEEMPH_OVERRIDE_5_0_MASK (0x3F << 16) + /* The spec defines this only for BXT PHY0, but lets assume that this * would exist for PHY1 too if it had a second channel. */ -- cgit v1.1 From cd96bea7ba90c45c8d1d315433c78021e56ec8c7 Mon Sep 17 00:00:00 2001 From: Manasi Navare Date: Fri, 23 Mar 2018 10:24:16 -0700 Subject: drm/i915/icl: Add Voltage swing table for MG PHY DDI Buffer This table is used for voltage swing programming sequence during DDI Buffer initialization for MG PHY DDI Buffers on Icelake. v2 (from Paulo): * Fix white space issues. Cc: Rodrigo Vivi Cc: Jani Nikula Reviewed-by: Paulo Zanoni Signed-off-by: Manasi Navare Signed-off-by: Paulo Zanoni Link: https://patchwork.freedesktop.org/patch/msgid/20180323172419.24911-5-paulo.r.zanoni@intel.com --- drivers/gpu/drm/i915/intel_ddi.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index 229b9d5..359acbf 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -592,6 +592,26 @@ static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_e { 0x0, 0x00, 0x00 }, /* 350 0.0 */ }; +struct icl_mg_phy_ddi_buf_trans { + u32 cri_txdeemph_override_5_0; + u32 cri_txdeemph_override_11_6; + u32 cri_txdeemph_override_17_12; +}; + +static const struct icl_mg_phy_ddi_buf_trans icl_mg_phy_ddi_translations[] = { + /* Voltage swing pre-emphasis */ + { 0x0, 0x1B, 0x00 }, /* 0 0 */ + { 0x0, 0x23, 0x08 }, /* 0 1 */ + { 0x0, 0x2D, 0x12 }, /* 0 2 */ + { 0x0, 0x00, 0x00 }, /* 0 3 */ + { 0x0, 0x23, 0x00 }, /* 1 0 */ + { 0x0, 0x2B, 0x09 }, /* 1 1 */ + { 0x0, 0x2E, 0x11 }, /* 1 2 */ + { 0x0, 0x2F, 0x00 }, /* 2 0 */ + { 0x0, 0x33, 0x0C }, /* 2 1 */ + { 0x0, 0x00, 0x00 }, /* 3 0 */ +}; + static const struct ddi_buf_trans * bdw_get_buf_trans_edp(struct drm_i915_private *dev_priv, int *n_entries) { -- cgit v1.1 From 96ae48311ebc23cfe4f929754dc1e1cfb0d031b0 Mon Sep 17 00:00:00 2001 From: Dhinakaran Pandiyan Date: Fri, 23 Mar 2018 10:24:17 -0700 Subject: drm/i915/icl: HPD pin for port F Extend enum hpd_pin to port F so that we can start using this for ICL. v2: Rebase. Cc: Rodrigo Vivi Cc: Paulo Zanoni Reviewed-by: Rodrigo Vivi Signed-off-by: Dhinakaran Pandiyan Signed-off-by: Paulo Zanoni Link: https://patchwork.freedesktop.org/patch/msgid/20180323172419.24911-6-paulo.r.zanoni@intel.com --- drivers/gpu/drm/i915/i915_drv.h | 1 + drivers/gpu/drm/i915/intel_hotplug.c | 3 +++ 2 files changed, 4 insertions(+) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 299b240..800230b 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -262,6 +262,7 @@ enum hpd_pin { HPD_PORT_C, HPD_PORT_D, HPD_PORT_E, + HPD_PORT_F, HPD_NUM_PINS }; diff --git a/drivers/gpu/drm/i915/intel_hotplug.c b/drivers/gpu/drm/i915/intel_hotplug.c index 0e3d3e8..43aa92b 100644 --- a/drivers/gpu/drm/i915/intel_hotplug.c +++ b/drivers/gpu/drm/i915/intel_hotplug.c @@ -100,6 +100,8 @@ enum port intel_hpd_pin_to_port(struct drm_i915_private *dev_priv, if (IS_CNL_WITH_PORT_F(dev_priv)) return PORT_F; return PORT_E; + case HPD_PORT_F: + return PORT_F; default: return PORT_NONE; /* no port for this pin */ } @@ -132,6 +134,7 @@ enum hpd_pin intel_hpd_pin_default(struct drm_i915_private *dev_priv, case PORT_F: if (IS_CNL_WITH_PORT_F(dev_priv)) return HPD_PORT_E; + return HPD_PORT_F; default: MISSING_CASE(port); return HPD_NONE; -- cgit v1.1 From 323301af974cdd4b797e5b54f5c418554f39d1fa Mon Sep 17 00:00:00 2001 From: Nabendu Maiti Date: Fri, 23 Mar 2018 10:24:18 -0700 Subject: drm/i915/icl: Added 5k source scaling support for Gen11 platform Gen11 supports upto 5k source scaling v2: Re-factoring of code as per review v3: Corrected max Vertical size and indentation v4: Added max Vertical dst size in same patch Reviewed-by: Paulo Zanoni Signed-off-by: Nabendu Maiti Signed-off-by: Paulo Zanoni Link: https://patchwork.freedesktop.org/patch/msgid/20180323172419.24911-7-paulo.r.zanoni@intel.com --- drivers/gpu/drm/i915/intel_display.c | 11 +++++++---- drivers/gpu/drm/i915/intel_drv.h | 4 ++++ 2 files changed, 11 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index b31b806..d5b3c7e 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -4756,10 +4756,13 @@ skl_update_scaler(struct intel_crtc_state *crtc_state, bool force_detach, /* range checks */ if (src_w < SKL_MIN_SRC_W || src_h < SKL_MIN_SRC_H || - dst_w < SKL_MIN_DST_W || dst_h < SKL_MIN_DST_H || - - src_w > SKL_MAX_SRC_W || src_h > SKL_MAX_SRC_H || - dst_w > SKL_MAX_DST_W || dst_h > SKL_MAX_DST_H) { + dst_w < SKL_MIN_DST_W || dst_h < SKL_MIN_DST_H || + (IS_GEN11(dev_priv) && + (src_w > ICL_MAX_SRC_W || src_h > ICL_MAX_SRC_H || + dst_w > ICL_MAX_DST_W || dst_h > ICL_MAX_DST_H)) || + (!IS_GEN11(dev_priv) && + (src_w > SKL_MAX_SRC_W || src_h > SKL_MAX_SRC_H || + dst_w > SKL_MAX_DST_W || dst_h > SKL_MAX_DST_H))) { DRM_DEBUG_KMS("scaler_user index %u.%u: src %ux%u dst %ux%u " "size is out of scaler range\n", intel_crtc->pipe, scaler_user, src_w, src_h, dst_w, dst_h); diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index b79a01b..d2935ac 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -548,6 +548,10 @@ struct intel_initial_plane_config { #define SKL_MAX_DST_W 4096 #define SKL_MIN_DST_H 8 #define SKL_MAX_DST_H 4096 +#define ICL_MAX_SRC_W 5120 +#define ICL_MAX_SRC_H 4096 +#define ICL_MAX_DST_W 5120 +#define ICL_MAX_DST_H 4096 struct intel_scaler { int in_use; -- cgit v1.1 From 7487508eff1fe787573aa6e0f3daaa6b12bd4520 Mon Sep 17 00:00:00 2001 From: Paulo Zanoni Date: Fri, 23 Mar 2018 12:58:53 -0700 Subject: drm/i915: protect macro parameters in SWING_SEL_{UPP,LO}WER Protect the macro parameters with parens in order to avoid priority issues on macro evaluation when the macro argument is not a single operand. This is not a problem today, but it could be in the future. I found this while reviewing a patch that introduces new callers for the macros. v2: Rebase. Reference: commit 04416108ccea ("drm/i915/cnl: Add registers related to voltage swing sequences.") Cc: Rodrigo Vivi Reviewed-by: Rodrigo Vivi Signed-off-by: Paulo Zanoni Link: https://patchwork.freedesktop.org/patch/msgid/20180323195853.4599-1-paulo.r.zanoni@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index aa001dd..b0c55f9 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -1749,9 +1749,9 @@ enum i915_power_well_id { #define ICL_PORT_TX_DW2_LN0(port) _MMIO_PORT(port, \ _ICL_PORT_TX_DW2_LN0_A, \ _ICL_PORT_TX_DW2_LN0_B) -#define SWING_SEL_UPPER(x) ((x >> 3) << 15) +#define SWING_SEL_UPPER(x) (((x) >> 3) << 15) #define SWING_SEL_UPPER_MASK (1 << 15) -#define SWING_SEL_LOWER(x) ((x & 0x7) << 11) +#define SWING_SEL_LOWER(x) (((x) & 0x7) << 11) #define SWING_SEL_LOWER_MASK (0x7 << 11) #define RCOMP_SCALAR(x) ((x) << 0) #define RCOMP_SCALAR_MASK (0xFF << 0) -- cgit v1.1 From aec06c7606be1635d82c16dda66282c2010bcd39 Mon Sep 17 00:00:00 2001 From: Arushi Singhal Date: Sun, 25 Mar 2018 23:31:32 +0530 Subject: gpu: drm/lease:: Use list_{next/prev}_entry instead of list_entry It's better to use list_entry instead of list_{next/prev}_entry as it makes the code more clear to read. This patch replace list_entry with list_{next/prev}_entry. Signed-off-by: Arushi Singhal Signed-off-by: Daniel Vetter Link: https://patchwork.freedesktop.org/patch/msgid/1522000893-5331-2-git-send-email-arushisinghal19971997@gmail.com --- drivers/gpu/drm/drm_lease.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_lease.c b/drivers/gpu/drm/drm_lease.c index d345563..50c73c0 100644 --- a/drivers/gpu/drm/drm_lease.c +++ b/drivers/gpu/drm/drm_lease.c @@ -340,7 +340,7 @@ static void _drm_lease_revoke(struct drm_master *top) break; /* Over */ - master = list_entry(master->lessee_list.next, struct drm_master, lessee_list); + master = list_next_entry(master, lessee_list); } } } -- cgit v1.1 From 3f07f28b9712605d1adb589344ced72e8397dc8a Mon Sep 17 00:00:00 2001 From: Arushi Singhal Date: Sun, 25 Mar 2018 23:31:33 +0530 Subject: gpu: drm: nouveau: Use list_{next/prev}_entry instead of list_entry It's better to use list_entry instead of list_{next/prev}_entry as it makes the code more clear to read. This patch replace list_entry with list_{next/prev}_entry. Signed-off-by: Arushi Singhal Acked-by: Ben Skeggs Signed-off-by: Daniel Vetter Link: https://patchwork.freedesktop.org/patch/msgid/1522000893-5331-3-git-send-email-arushisinghal19971997@gmail.com --- drivers/gpu/drm/nouveau/nvkm/subdev/clk/base.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/base.c index e4c8d31..81c3567 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/base.c @@ -134,7 +134,7 @@ nvkm_cstate_find_best(struct nvkm_clk *clk, struct nvkm_pstate *pstate, nvkm_volt_map(volt, volt->max2_id, clk->temp)); for (cstate = start; &cstate->head != &pstate->list; - cstate = list_entry(cstate->head.prev, typeof(*cstate), head)) { + cstate = list_prev_entry(cstate, head)) { if (nvkm_cstate_valid(clk, cstate, max_volt, clk->temp)) break; } -- cgit v1.1 From 3903117609f40a0bc21b8c48533ac07a574c4cdc Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 22 Mar 2018 09:02:33 +0100 Subject: drm/gem: Document that handle_create must be the last step It published the gem object to userspace, by that point other threads can guess the id and start using it. And gem IDs are _very_ easy to guess (it's just an idr). Since gem objects is the only thing we allow drivers to create themselves (all the kms/prime/syncobj stuff is handled by the core) no other functions seem to be in need of this clarification. Motivated by reviewing the xen-front kms driver. Cc: Oleksandr Andrushchenko Reviewed-by: Oleksandr Andrushchenko Signed-off-by: Daniel Vetter Link: https://patchwork.freedesktop.org/patch/msgid/20180322080233.17266-1-daniel.vetter@ffwll.ch --- drivers/gpu/drm/drm_gem.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c index 4975ba9..4a16d7b 100644 --- a/drivers/gpu/drm/drm_gem.c +++ b/drivers/gpu/drm/drm_gem.c @@ -436,9 +436,12 @@ err_unref: * @obj: object to register * @handlep: pionter to return the created handle to the caller * - * Create a handle for this object. This adds a handle reference - * to the object, which includes a regular reference count. Callers - * will likely want to dereference the object afterwards. + * Create a handle for this object. This adds a handle reference to the object, + * which includes a regular reference count. Callers will likely want to + * dereference the object afterwards. + * + * Since this publishes @obj to userspace it must be fully set up by this point, + * drivers must call this last in their buffer object creation callbacks. */ int drm_gem_handle_create(struct drm_file *file_priv, struct drm_gem_object *obj, -- cgit v1.1 From 8017e422af3f5ab90210bfbd17f9a8d8d56db289 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Fri, 23 Mar 2018 20:25:37 +0200 Subject: drm/scdc-helper: Convert errors into debug messages MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Since we may attempt to reconfigure SCDC when the sink has already been disconnected we probably shouldn't scare the user with errors in dmesg that are 100% expected in that case. Just leave it up to the caller whether to print an error message or not, and just output debug messages from the helper itself. Cc: Shashank Sharma Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180323182537.30784-1-ville.syrjala@linux.intel.com Reviewed-by: Shashank Sharma --- drivers/gpu/drm/drm_scdc_helper.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/drm_scdc_helper.c b/drivers/gpu/drm/drm_scdc_helper.c index 657ea5a..870e25f 100644 --- a/drivers/gpu/drm/drm_scdc_helper.c +++ b/drivers/gpu/drm/drm_scdc_helper.c @@ -141,7 +141,7 @@ bool drm_scdc_get_scrambling_status(struct i2c_adapter *adapter) ret = drm_scdc_readb(adapter, SCDC_SCRAMBLER_STATUS, &status); if (ret < 0) { - DRM_ERROR("Failed to read scrambling status: %d\n", ret); + DRM_DEBUG_KMS("Failed to read scrambling status: %d\n", ret); return false; } @@ -168,7 +168,7 @@ bool drm_scdc_set_scrambling(struct i2c_adapter *adapter, bool enable) ret = drm_scdc_readb(adapter, SCDC_TMDS_CONFIG, &config); if (ret < 0) { - DRM_ERROR("Failed to read TMDS config: %d\n", ret); + DRM_DEBUG_KMS("Failed to read TMDS config: %d\n", ret); return false; } @@ -179,7 +179,7 @@ bool drm_scdc_set_scrambling(struct i2c_adapter *adapter, bool enable) ret = drm_scdc_writeb(adapter, SCDC_TMDS_CONFIG, config); if (ret < 0) { - DRM_ERROR("Failed to enable scrambling: %d\n", ret); + DRM_DEBUG_KMS("Failed to enable scrambling: %d\n", ret); return false; } @@ -223,7 +223,7 @@ bool drm_scdc_set_high_tmds_clock_ratio(struct i2c_adapter *adapter, bool set) ret = drm_scdc_readb(adapter, SCDC_TMDS_CONFIG, &config); if (ret < 0) { - DRM_ERROR("Failed to read TMDS config: %d\n", ret); + DRM_DEBUG_KMS("Failed to read TMDS config: %d\n", ret); return false; } @@ -234,7 +234,7 @@ bool drm_scdc_set_high_tmds_clock_ratio(struct i2c_adapter *adapter, bool set) ret = drm_scdc_writeb(adapter, SCDC_TMDS_CONFIG, config); if (ret < 0) { - DRM_ERROR("Failed to set TMDS clock ratio: %d\n", ret); + DRM_DEBUG_KMS("Failed to set TMDS clock ratio: %d\n", ret); return false; } -- cgit v1.1 From a7d2a87e99deb5481b5dd723408c42f460de25a3 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 22 Mar 2018 11:51:28 +0100 Subject: drm/tinydrm: Use gem_free_object_unlocked MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit tinydrm doesn't use dev->struct_mutex and therefore has no need to use gem_free_object. Signed-off-by: Daniel Vetter Cc: "Noralf Trønnes" Acked-by: Noralf Trønnes Link: https://patchwork.freedesktop.org/patch/msgid/20180322105133.11211-2-daniel.vetter@ffwll.ch --- drivers/gpu/drm/tinydrm/core/tinydrm-core.c | 2 +- include/drm/tinydrm/tinydrm.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/tinydrm/core/tinydrm-core.c b/drivers/gpu/drm/tinydrm/core/tinydrm-core.c index 4c661627..24a33bf 100644 --- a/drivers/gpu/drm/tinydrm/core/tinydrm-core.c +++ b/drivers/gpu/drm/tinydrm/core/tinydrm-core.c @@ -91,7 +91,7 @@ EXPORT_SYMBOL(tinydrm_gem_cma_prime_import_sg_table); * GEM object state and frees the memory used to store the object itself using * drm_gem_cma_free_object(). It also handles PRIME buffers which has the kernel * virtual address set by tinydrm_gem_cma_prime_import_sg_table(). Drivers - * can use this as their &drm_driver->gem_free_object callback. + * can use this as their &drm_driver->gem_free_object_unlocked callback. */ void tinydrm_gem_cma_free_object(struct drm_gem_object *gem_obj) { diff --git a/include/drm/tinydrm/tinydrm.h b/include/drm/tinydrm/tinydrm.h index 07a9a11..77a93ec 100644 --- a/include/drm/tinydrm/tinydrm.h +++ b/include/drm/tinydrm/tinydrm.h @@ -41,7 +41,7 @@ pipe_to_tinydrm(struct drm_simple_display_pipe *pipe) * the &drm_driver structure. */ #define TINYDRM_GEM_DRIVER_OPS \ - .gem_free_object = tinydrm_gem_cma_free_object, \ + .gem_free_object_unlocked = tinydrm_gem_cma_free_object, \ .gem_print_info = drm_gem_cma_print_info, \ .gem_vm_ops = &drm_gem_cma_vm_ops, \ .prime_handle_to_fd = drm_gem_prime_handle_to_fd, \ -- cgit v1.1 From eed7ec52f214bac2f25395ccaad610fbeb842a6e Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Sat, 24 Mar 2018 12:58:29 +0000 Subject: drm/i915/execlists: Clear user-active flag on preemption completion MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When cancelling the requests and clearing out the ports following a successful preemption completion, also clear the active flag. I had assumed that all preemptions would be followed by an immediate dequeue (preserving the active user flag), but under rare circumstances we may be triggering a preemption for the second port only for it to have completed before the preemotion kicks in; leaving execlists->active set even though the system is now idle. We can clear the flag inside the common execlists_cancel_port_requests() as the other users also expect the semantics of active being cleared. Fixes: f6322eddaff7 ("drm/i915/preemption: Allow preemption between submission ports") Signed-off-by: Chris Wilson Cc: Michał Winiarski Cc: Michel Thierry Cc: Mika Kuoppala Cc: Tvrtko Ursulin Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180324125829.27026-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/intel_lrc.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index b4ab06b..9c84af5 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -577,6 +577,8 @@ static void execlists_dequeue(struct intel_engine_cs *engine) * know the next preemption status we see corresponds * to this ELSP update. */ + GEM_BUG_ON(!execlists_is_active(execlists, + EXECLISTS_ACTIVE_USER)); GEM_BUG_ON(!port_count(&port[0])); if (port_count(&port[0]) > 1) goto unlock; @@ -738,6 +740,8 @@ execlists_cancel_port_requests(struct intel_engine_execlists * const execlists) memset(port, 0, sizeof(*port)); port++; } + + execlists_clear_active(execlists, EXECLISTS_ACTIVE_USER); } static void clear_gtiir(struct intel_engine_cs *engine) @@ -1042,6 +1046,11 @@ static void execlists_submission_tasklet(unsigned long data) if (fw) intel_uncore_forcewake_put(dev_priv, execlists->fw_domains); + + /* If the engine is now idle, so should be the flag; and vice versa. */ + GEM_BUG_ON(execlists_is_active(&engine->execlists, + EXECLISTS_ACTIVE_USER) == + !port_isset(engine->execlists.port)); } static void queue_request(struct intel_engine_cs *engine, -- cgit v1.1 From 9040871336db2c2e379402d7ba9c275d6495b9e0 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 26 Mar 2018 12:50:36 +0100 Subject: drm/i915: Include submission tasklet state in engine dump For the off-chance we have an interrupt posted and haven't processed the CSB. v2: Include tasklet enable/disable state for good measure. Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180326115044.2505-4-chris@chris-wilson.co.uk Reviewed-by: Mika Kuoppala --- drivers/gpu/drm/i915/intel_engine_cs.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index de09fa4..12486d8 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -1859,12 +1859,15 @@ static void intel_engine_print_registers(const struct intel_engine_cs *engine, ptr = I915_READ(RING_CONTEXT_STATUS_PTR(engine)); read = GEN8_CSB_READ_PTR(ptr); write = GEN8_CSB_WRITE_PTR(ptr); - drm_printf(m, "\tExeclist CSB read %d [%d cached], write %d [%d from hws], interrupt posted? %s\n", + drm_printf(m, "\tExeclist CSB read %d [%d cached], write %d [%d from hws], interrupt posted? %s, tasklet queued? %s (%s)\n", read, execlists->csb_head, write, intel_read_status_page(engine, intel_hws_csb_write_index(engine->i915)), yesno(test_bit(ENGINE_IRQ_EXECLIST, - &engine->irq_posted))); + &engine->irq_posted)), + yesno(test_bit(TASKLET_STATE_SCHED, + &engine->execlists.tasklet.state)), + enableddisabled(!atomic_read(&engine->execlists.tasklet.count))); if (read >= GEN8_CSB_ENTRIES) read = 0; if (write >= GEN8_CSB_ENTRIES) -- cgit v1.1 From ae2f5c009335f4d273d2431539de91c8f28d4736 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 26 Mar 2018 12:50:34 +0100 Subject: drm/i915/execlists: Avoid kicking the submission too early for rescheduling MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If the request is still waiting on external fences, it has not yet been submitted to the HW queue and so we can forgo kicking the submission tasklet when re-evaluating its priority. This should have no impact other than reducing the number of tasklet wakeups under signal heavy workloads (e.g. switching between engines). v2: Use prebaked container_of() References: f6322eddaff7 ("drm/i915/preemption: Allow preemption between submission ports") Signed-off-by: Chris Wilson Cc: Michał Winiarski Cc: Michel Thierry Cc: Mika Kuoppala Cc: Tvrtko Ursulin Link: https://patchwork.freedesktop.org/patch/msgid/20180326115044.2505-2-chris@chris-wilson.co.uk Reviewed-by: Mika Kuoppala --- drivers/gpu/drm/i915/intel_lrc.c | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 9c84af5..ba7f783 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -1060,12 +1060,16 @@ static void queue_request(struct intel_engine_cs *engine, list_add_tail(&pt->link, &lookup_priolist(engine, pt, prio)->requests); } +static void __submit_queue(struct intel_engine_cs *engine, int prio) +{ + engine->execlists.queue_priority = prio; + tasklet_hi_schedule(&engine->execlists.tasklet); +} + static void submit_queue(struct intel_engine_cs *engine, int prio) { - if (prio > engine->execlists.queue_priority) { - engine->execlists.queue_priority = prio; - tasklet_hi_schedule(&engine->execlists.tasklet); - } + if (prio > engine->execlists.queue_priority) + __submit_queue(engine, prio); } static void execlists_submit_request(struct i915_request *request) @@ -1198,7 +1202,10 @@ static void execlists_schedule(struct i915_request *request, int prio) __list_del_entry(&pt->link); queue_request(engine, pt, prio); } - submit_queue(engine, prio); + + if (prio > engine->execlists.queue_priority && + i915_sw_fence_done(&pt_to_request(pt)->submit)) + __submit_queue(engine, prio); } spin_unlock_irq(&engine->timeline->lock); -- cgit v1.1 From 57bdff48a0a553468f16a65149d51213b5f25fee Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Mon, 19 Mar 2018 10:37:20 -0700 Subject: drm/i915: Reword warning for missing cases MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In some places we end up converting switch statements to a series of if/else, particularly when introducing helper functions to handle a group of cases. It's tempting to either leave a wrong warning (since now we don't have a switch case anymore) or to convert to WARN(1, ...), but we can just provide a better message and avoid the doubt when such conversions arrise. Introducing a warning inside i915_driver_load() just for tests we get: [ 4535.233717] Missing case (ret == 0) [ 4535.233868] WARNING: CPU: 1 PID: 795 at drivers/gpu/drm/i915/i915_drv.c:1341 i915_driver_load+0x42/0x10e0 [i915] which is clear enough. v2: remove __func__ since this is already on the warning. Cc: Chris Wilson Cc: Ville Syrjälä Cc: Daniel Vetter Signed-off-by: Lucas De Marchi Reviewed-by: Chris Wilson Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180319173720.6974-1-lucas.demarchi@intel.com --- drivers/gpu/drm/i915/i915_utils.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_utils.h b/drivers/gpu/drm/i915/i915_utils.h index 51dbfe5..0695717 100644 --- a/drivers/gpu/drm/i915/i915_utils.h +++ b/drivers/gpu/drm/i915/i915_utils.h @@ -40,8 +40,8 @@ #undef WARN_ON_ONCE #define WARN_ON_ONCE(x) WARN_ONCE((x), "%s", "WARN_ON_ONCE(" __stringify(x) ")") -#define MISSING_CASE(x) WARN(1, "Missing switch case (%lu) in %s\n", \ - (long)(x), __func__) +#define MISSING_CASE(x) WARN(1, "Missing case (%s == %ld)\n", \ + __stringify(x), (long)(x)) #if GCC_VERSION >= 70000 #define add_overflows(A, B) \ -- cgit v1.1 From 47aa1e73e72e5e0f5a07e30b84478461195845f5 Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Tue, 20 Mar 2018 15:06:31 -0700 Subject: drm/i915: move dpll_info to header MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This will allow the struct to be embedded in intel_shared_dpll. Signed-off-by: Lucas De Marchi Reviewed-by: Ville Syrjälä Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180320220637.21480-2-lucas.demarchi@intel.com --- drivers/gpu/drm/i915/intel_dpll_mgr.c | 7 ------- drivers/gpu/drm/i915/intel_dpll_mgr.h | 10 ++++++++++ 2 files changed, 10 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.c b/drivers/gpu/drm/i915/intel_dpll_mgr.c index 51c5ae4..52d6e73 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.c +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.c @@ -1877,13 +1877,6 @@ static void intel_ddi_pll_init(struct drm_device *dev) } } -struct dpll_info { - const char *name; - const int id; - const struct intel_shared_dpll_funcs *funcs; - uint32_t flags; -}; - struct intel_dpll_mgr { const struct dpll_info *dpll_info; diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.h b/drivers/gpu/drm/i915/intel_dpll_mgr.h index f24ccf4..e99d638 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.h +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.h @@ -206,6 +206,16 @@ struct intel_shared_dpll_funcs { }; /** + * struct dpll_info - display PLL platform specific info + */ +struct dpll_info { + const char *name; + const int id; + const struct intel_shared_dpll_funcs *funcs; + uint32_t flags; +}; + +/** * struct intel_shared_dpll - display PLL with tracked state and users */ struct intel_shared_dpll { -- cgit v1.1 From e30379637fc712991ec8fff1f79344f384ace0aa Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Tue, 20 Mar 2018 15:06:32 -0700 Subject: drm/i915: add dpll_info inside intel_shared_dpll MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This way we can stop copying fields from dpll_info to intel_shared_dpll one by one. The migration of each field will come on separate patches. Signed-off-by: Lucas De Marchi Reviewed-by: Ville Syrjälä Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180320220637.21480-3-lucas.demarchi@intel.com --- drivers/gpu/drm/i915/intel_dpll_mgr.c | 1 + drivers/gpu/drm/i915/intel_dpll_mgr.h | 5 +++++ 2 files changed, 6 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.c b/drivers/gpu/drm/i915/intel_dpll_mgr.c index 52d6e73..30a9ac5 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.c +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.c @@ -2410,6 +2410,7 @@ void intel_shared_dpll_init(struct drm_device *dev) for (i = 0; dpll_info[i].id >= 0; i++) { WARN_ON(i != dpll_info[i].id); + dev_priv->shared_dplls[i].info = &dpll_info[i]; dev_priv->shared_dplls[i].id = dpll_info[i].id; dev_priv->shared_dplls[i].name = dpll_info[i].name; diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.h b/drivers/gpu/drm/i915/intel_dpll_mgr.h index e99d638..bd2d365 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.h +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.h @@ -253,6 +253,11 @@ struct intel_shared_dpll { */ struct intel_shared_dpll_funcs funcs; + /** + * @info: platform specific info + */ + const struct dpll_info *info; + #define INTEL_DPLL_ALWAYS_ON (1 << 0) /** * @flags: -- cgit v1.1 From ee1398ba01b308015e568d46fd62a38ceebd7abe Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Tue, 20 Mar 2018 15:06:33 -0700 Subject: drm/i915: use funcs from intel_shared_dpll.info MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Replace all users of pll->funcs.* to use pll->info->funcs->*. The extra indirection here is not on any critical path and we can leave all const data together. Signed-off-by: Lucas De Marchi Reviewed-by: Ville Syrjälä Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180320220637.21480-4-lucas.demarchi@intel.com --- drivers/gpu/drm/i915/intel_display.c | 16 ++++++++-------- drivers/gpu/drm/i915/intel_dpll_mgr.c | 9 ++++----- drivers/gpu/drm/i915/intel_dpll_mgr.h | 8 +++----- 3 files changed, 15 insertions(+), 18 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index d5b3c7e..cee9ea3 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -8763,8 +8763,8 @@ static bool ironlake_get_pipe_config(struct intel_crtc *crtc, intel_get_shared_dpll_by_id(dev_priv, pll_id); pll = pipe_config->shared_dpll; - WARN_ON(!pll->funcs.get_hw_state(dev_priv, pll, - &pipe_config->dpll_hw_state)); + WARN_ON(!pll->info->funcs->get_hw_state(dev_priv, pll, + &pipe_config->dpll_hw_state)); tmp = pipe_config->dpll_hw_state.dpll; pipe_config->pixel_multiplier = @@ -9240,8 +9240,8 @@ static void haswell_get_ddi_port_state(struct intel_crtc *crtc, pll = pipe_config->shared_dpll; if (pll) { - WARN_ON(!pll->funcs.get_hw_state(dev_priv, pll, - &pipe_config->dpll_hw_state)); + WARN_ON(!pll->info->funcs->get_hw_state(dev_priv, pll, + &pipe_config->dpll_hw_state)); } /* @@ -11655,7 +11655,7 @@ verify_single_dpll_state(struct drm_i915_private *dev_priv, DRM_DEBUG_KMS("%s\n", pll->name); - active = pll->funcs.get_hw_state(dev_priv, pll, &dpll_hw_state); + active = pll->info->funcs->get_hw_state(dev_priv, pll, &dpll_hw_state); if (!(pll->flags & INTEL_DPLL_ALWAYS_ON)) { I915_STATE_WARN(!pll->on && pll->active_mask, @@ -15128,8 +15128,8 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) for (i = 0; i < dev_priv->num_shared_dpll; i++) { struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i]; - pll->on = pll->funcs.get_hw_state(dev_priv, pll, - &pll->state.hw_state); + pll->on = pll->info->funcs->get_hw_state(dev_priv, pll, + &pll->state.hw_state); pll->state.crtc_mask = 0; for_each_intel_crtc(dev, crtc) { struct intel_crtc_state *crtc_state = @@ -15318,7 +15318,7 @@ intel_modeset_setup_hw_state(struct drm_device *dev, DRM_DEBUG_KMS("%s enabled but not in use, disabling\n", pll->name); - pll->funcs.disable(dev_priv, pll); + pll->info->funcs->disable(dev_priv, pll); pll->on = false; } diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.c b/drivers/gpu/drm/i915/intel_dpll_mgr.c index 30a9ac5..24d9aa1 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.c +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.c @@ -118,7 +118,7 @@ void assert_shared_dpll(struct drm_i915_private *dev_priv, if (WARN(!pll, "asserting DPLL %s with no DPLL\n", onoff(state))) return; - cur_state = pll->funcs.get_hw_state(dev_priv, pll, &hw_state); + cur_state = pll->info->funcs->get_hw_state(dev_priv, pll, &hw_state); I915_STATE_WARN(cur_state != state, "%s assertion failure (expected %s, current %s)\n", pll->name, onoff(state), onoff(cur_state)); @@ -147,7 +147,7 @@ void intel_prepare_shared_dpll(struct intel_crtc *crtc) WARN_ON(pll->on); assert_shared_dpll_disabled(dev_priv, pll); - pll->funcs.prepare(dev_priv, pll); + pll->info->funcs->prepare(dev_priv, pll); } mutex_unlock(&dev_priv->dpll_lock); } @@ -190,7 +190,7 @@ void intel_enable_shared_dpll(struct intel_crtc *crtc) WARN_ON(pll->on); DRM_DEBUG_KMS("enabling %s\n", pll->name); - pll->funcs.enable(dev_priv, pll); + pll->info->funcs->enable(dev_priv, pll); pll->on = true; out: @@ -232,7 +232,7 @@ void intel_disable_shared_dpll(struct intel_crtc *crtc) goto out; DRM_DEBUG_KMS("disabling %s\n", pll->name); - pll->funcs.disable(dev_priv, pll); + pll->info->funcs->disable(dev_priv, pll); pll->on = false; out: @@ -2414,7 +2414,6 @@ void intel_shared_dpll_init(struct drm_device *dev) dev_priv->shared_dplls[i].id = dpll_info[i].id; dev_priv->shared_dplls[i].name = dpll_info[i].name; - dev_priv->shared_dplls[i].funcs = *dpll_info[i].funcs; dev_priv->shared_dplls[i].flags = dpll_info[i].flags; } diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.h b/drivers/gpu/drm/i915/intel_dpll_mgr.h index bd2d365..f493822 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.h +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.h @@ -211,6 +211,9 @@ struct intel_shared_dpll_funcs { struct dpll_info { const char *name; const int id; + /** + * @funcs: platform specific hooks + */ const struct intel_shared_dpll_funcs *funcs; uint32_t flags; }; @@ -249,11 +252,6 @@ struct intel_shared_dpll { enum intel_dpll_id id; /** - * @funcs: platform specific hooks - */ - struct intel_shared_dpll_funcs funcs; - - /** * @info: platform specific info */ const struct dpll_info *info; -- cgit v1.1 From 72f775fa284886893bec4a189ed38ac30e2535aa Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Tue, 20 Mar 2018 15:06:34 -0700 Subject: drm/i915: use name from intel_shared_dpll.info MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Replace all users of pll->name to use pll->info->name. Signed-off-by: Lucas De Marchi Reviewed-by: Ville Syrjälä Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180320220637.21480-5-lucas.demarchi@intel.com --- drivers/gpu/drm/i915/i915_debugfs.c | 3 ++- drivers/gpu/drm/i915/intel_display.c | 7 ++++--- drivers/gpu/drm/i915/intel_dpll_mgr.c | 26 ++++++++++++++------------ drivers/gpu/drm/i915/intel_dpll_mgr.h | 8 +++----- 4 files changed, 23 insertions(+), 21 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 7816cd5..057fe12 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -3285,7 +3285,8 @@ static int i915_shared_dplls_info(struct seq_file *m, void *unused) for (i = 0; i < dev_priv->num_shared_dpll; i++) { struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i]; - seq_printf(m, "DPLL%i: %s, id: %i\n", i, pll->name, pll->id); + seq_printf(m, "DPLL%i: %s, id: %i\n", i, pll->info->name, + pll->id); seq_printf(m, " crtc_mask: 0x%08x, active: 0x%x, on: %s\n", pll->state.crtc_mask, pll->active_mask, yesno(pll->on)); seq_printf(m, " tracked hardware state:\n"); diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index cee9ea3..5f79444 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -11653,7 +11653,7 @@ verify_single_dpll_state(struct drm_i915_private *dev_priv, memset(&dpll_hw_state, 0, sizeof(dpll_hw_state)); - DRM_DEBUG_KMS("%s\n", pll->name); + DRM_DEBUG_KMS("%s\n", pll->info->name); active = pll->info->funcs->get_hw_state(dev_priv, pll, &dpll_hw_state); @@ -15142,7 +15142,7 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) pll->active_mask = pll->state.crtc_mask; DRM_DEBUG_KMS("%s hw state readout: crtc_mask 0x%08x, on %i\n", - pll->name, pll->state.crtc_mask, pll->on); + pll->info->name, pll->state.crtc_mask, pll->on); } for_each_intel_encoder(dev, encoder) { @@ -15316,7 +15316,8 @@ intel_modeset_setup_hw_state(struct drm_device *dev, if (!pll->on || pll->active_mask) continue; - DRM_DEBUG_KMS("%s enabled but not in use, disabling\n", pll->name); + DRM_DEBUG_KMS("%s enabled but not in use, disabling\n", + pll->info->name); pll->info->funcs->disable(dev_priv, pll); pll->on = false; diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.c b/drivers/gpu/drm/i915/intel_dpll_mgr.c index 24d9aa1..ed46ade 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.c +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.c @@ -121,7 +121,7 @@ void assert_shared_dpll(struct drm_i915_private *dev_priv, cur_state = pll->info->funcs->get_hw_state(dev_priv, pll, &hw_state); I915_STATE_WARN(cur_state != state, "%s assertion failure (expected %s, current %s)\n", - pll->name, onoff(state), onoff(cur_state)); + pll->info->name, onoff(state), onoff(cur_state)); } /** @@ -143,7 +143,7 @@ void intel_prepare_shared_dpll(struct intel_crtc *crtc) mutex_lock(&dev_priv->dpll_lock); WARN_ON(!pll->state.crtc_mask); if (!pll->active_mask) { - DRM_DEBUG_DRIVER("setting up %s\n", pll->name); + DRM_DEBUG_DRIVER("setting up %s\n", pll->info->name); WARN_ON(pll->on); assert_shared_dpll_disabled(dev_priv, pll); @@ -179,7 +179,7 @@ void intel_enable_shared_dpll(struct intel_crtc *crtc) pll->active_mask |= crtc_mask; DRM_DEBUG_KMS("enable %s (active %x, on? %d) for crtc %d\n", - pll->name, pll->active_mask, pll->on, + pll->info->name, pll->active_mask, pll->on, crtc->base.base.id); if (old_mask) { @@ -189,7 +189,7 @@ void intel_enable_shared_dpll(struct intel_crtc *crtc) } WARN_ON(pll->on); - DRM_DEBUG_KMS("enabling %s\n", pll->name); + DRM_DEBUG_KMS("enabling %s\n", pll->info->name); pll->info->funcs->enable(dev_priv, pll); pll->on = true; @@ -221,7 +221,7 @@ void intel_disable_shared_dpll(struct intel_crtc *crtc) goto out; DRM_DEBUG_KMS("disable %s (active %x, on? %d) for crtc %d\n", - pll->name, pll->active_mask, pll->on, + pll->info->name, pll->active_mask, pll->on, crtc->base.base.id); assert_shared_dpll_enabled(dev_priv, pll); @@ -231,7 +231,7 @@ void intel_disable_shared_dpll(struct intel_crtc *crtc) if (pll->active_mask) goto out; - DRM_DEBUG_KMS("disabling %s\n", pll->name); + DRM_DEBUG_KMS("disabling %s\n", pll->info->name); pll->info->funcs->disable(dev_priv, pll); pll->on = false; @@ -263,7 +263,8 @@ intel_find_shared_dpll(struct intel_crtc *crtc, &shared_dpll[i].hw_state, sizeof(crtc_state->dpll_hw_state)) == 0) { DRM_DEBUG_KMS("[CRTC:%d:%s] sharing existing %s (crtc mask 0x%08x, active %x)\n", - crtc->base.base.id, crtc->base.name, pll->name, + crtc->base.base.id, crtc->base.name, + pll->info->name, shared_dpll[i].crtc_mask, pll->active_mask); return pll; @@ -275,7 +276,8 @@ intel_find_shared_dpll(struct intel_crtc *crtc, pll = &dev_priv->shared_dplls[i]; if (shared_dpll[i].crtc_mask == 0) { DRM_DEBUG_KMS("[CRTC:%d:%s] allocated %s\n", - crtc->base.base.id, crtc->base.name, pll->name); + crtc->base.base.id, crtc->base.name, + pll->info->name); return pll; } } @@ -298,7 +300,7 @@ intel_reference_shared_dpll(struct intel_shared_dpll *pll, crtc_state->dpll_hw_state; crtc_state->shared_dpll = pll; - DRM_DEBUG_DRIVER("using %s for pipe %c\n", pll->name, + DRM_DEBUG_DRIVER("using %s for pipe %c\n", pll->info->name, pipe_name(crtc->pipe)); shared_dpll[pll->id].crtc_mask |= 1 << crtc->pipe; @@ -429,7 +431,8 @@ ibx_get_dpll(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state, pll = &dev_priv->shared_dplls[i]; DRM_DEBUG_KMS("[CRTC:%d:%s] using pre-allocated %s\n", - crtc->base.base.id, crtc->base.name, pll->name); + crtc->base.base.id, crtc->base.name, + pll->info->name); } else { pll = intel_find_shared_dpll(crtc, crtc_state, DPLL_ID_PCH_PLL_A, @@ -1824,7 +1827,7 @@ bxt_get_dpll(struct intel_crtc *crtc, pll = intel_get_shared_dpll_by_id(dev_priv, i); DRM_DEBUG_KMS("[CRTC:%d:%s] using pre-allocated %s\n", - crtc->base.base.id, crtc->base.name, pll->name); + crtc->base.base.id, crtc->base.name, pll->info->name); intel_reference_shared_dpll(pll, crtc_state); @@ -2413,7 +2416,6 @@ void intel_shared_dpll_init(struct drm_device *dev) dev_priv->shared_dplls[i].info = &dpll_info[i]; dev_priv->shared_dplls[i].id = dpll_info[i].id; - dev_priv->shared_dplls[i].name = dpll_info[i].name; dev_priv->shared_dplls[i].flags = dpll_info[i].flags; } diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.h b/drivers/gpu/drm/i915/intel_dpll_mgr.h index f493822..e5ed3e0 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.h +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.h @@ -209,6 +209,9 @@ struct intel_shared_dpll_funcs { * struct dpll_info - display PLL platform specific info */ struct dpll_info { + /** + * @name: DPLL name; used for logging + */ const char *name; const int id; /** @@ -241,11 +244,6 @@ struct intel_shared_dpll { bool on; /** - * @name: DPLL name; used for logging - */ - const char *name; - - /** * @id: unique indentifier for this DPLL; should match the index in the * dev_priv->shared_dplls array */ -- cgit v1.1 From 0823eb9c52f9e47564277e5bdd6b78cd95cf9f9c Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Tue, 20 Mar 2018 15:06:35 -0700 Subject: drm/i915: use id from intel_shared_dpll.info MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Replace all users of pll->id to use pll->info->id. In functions using this more than once it was preferred to add an id variable to make the code easier to read. Signed-off-by: Lucas De Marchi Reviewed-by: Ville Syrjälä Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180320220637.21480-6-lucas.demarchi@intel.com --- drivers/gpu/drm/i915/i915_debugfs.c | 2 +- drivers/gpu/drm/i915/intel_ddi.c | 8 +- drivers/gpu/drm/i915/intel_dpll_mgr.c | 160 +++++++++++++++++++--------------- drivers/gpu/drm/i915/intel_dpll_mgr.h | 10 +-- 4 files changed, 98 insertions(+), 82 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 057fe12..ff90577 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -3286,7 +3286,7 @@ static int i915_shared_dplls_info(struct seq_file *m, void *unused) struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i]; seq_printf(m, "DPLL%i: %s, id: %i\n", i, pll->info->name, - pll->id); + pll->info->id); seq_printf(m, " crtc_mask: 0x%08x, active: 0x%x, on: %s\n", pll->state.crtc_mask, pll->active_mask, yesno(pll->on)); seq_printf(m, " tracked hardware state:\n"); diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index 359acbf..a6672a9 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -994,7 +994,7 @@ static void intel_wait_ddi_buf_idle(struct drm_i915_private *dev_priv, static uint32_t hsw_pll_to_ddi_pll_sel(const struct intel_shared_dpll *pll) { - switch (pll->id) { + switch (pll->info->id) { case DPLL_ID_WRPLL1: return PORT_CLK_SEL_WRPLL1; case DPLL_ID_WRPLL2: @@ -1008,7 +1008,7 @@ static uint32_t hsw_pll_to_ddi_pll_sel(const struct intel_shared_dpll *pll) case DPLL_ID_LCPLL_2700: return PORT_CLK_SEL_LCPLL_2700; default: - MISSING_CASE(pll->id); + MISSING_CASE(pll->info->id); return PORT_CLK_SEL_NONE; } } @@ -2250,7 +2250,7 @@ static void intel_ddi_clk_select(struct intel_encoder *encoder, /* Configure DPCLKA_CFGCR0 to map the DPLL to the DDI. */ val = I915_READ(DPCLKA_CFGCR0); val &= ~DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(port); - val |= DPCLKA_CFGCR0_DDI_CLK_SEL(pll->id, port); + val |= DPCLKA_CFGCR0_DDI_CLK_SEL(pll->info->id, port); I915_WRITE(DPCLKA_CFGCR0, val); /* @@ -2267,7 +2267,7 @@ static void intel_ddi_clk_select(struct intel_encoder *encoder, val &= ~(DPLL_CTRL2_DDI_CLK_OFF(port) | DPLL_CTRL2_DDI_CLK_SEL_MASK(port)); - val |= (DPLL_CTRL2_DDI_CLK_SEL(pll->id, port) | + val |= (DPLL_CTRL2_DDI_CLK_SEL(pll->info->id, port) | DPLL_CTRL2_DDI_SEL_OVERRIDE(port)); I915_WRITE(DPLL_CTRL2, val); diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.c b/drivers/gpu/drm/i915/intel_dpll_mgr.c index ed46ade..48466b1 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.c +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.c @@ -291,19 +291,19 @@ intel_reference_shared_dpll(struct intel_shared_dpll *pll, { struct intel_shared_dpll_state *shared_dpll; struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); - enum intel_dpll_id i = pll->id; + const enum intel_dpll_id id = pll->info->id; shared_dpll = intel_atomic_get_shared_dpll_state(crtc_state->base.state); - if (shared_dpll[i].crtc_mask == 0) - shared_dpll[i].hw_state = + if (shared_dpll[id].crtc_mask == 0) + shared_dpll[id].hw_state = crtc_state->dpll_hw_state; crtc_state->shared_dpll = pll; DRM_DEBUG_DRIVER("using %s for pipe %c\n", pll->info->name, pipe_name(crtc->pipe)); - shared_dpll[pll->id].crtc_mask |= 1 << crtc->pipe; + shared_dpll[id].crtc_mask |= 1 << crtc->pipe; } /** @@ -343,15 +343,16 @@ static bool ibx_pch_dpll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll, struct intel_dpll_hw_state *hw_state) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) return false; - val = I915_READ(PCH_DPLL(pll->id)); + val = I915_READ(PCH_DPLL(id)); hw_state->dpll = val; - hw_state->fp0 = I915_READ(PCH_FP0(pll->id)); - hw_state->fp1 = I915_READ(PCH_FP1(pll->id)); + hw_state->fp0 = I915_READ(PCH_FP0(id)); + hw_state->fp1 = I915_READ(PCH_FP1(id)); intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); @@ -361,8 +362,10 @@ static bool ibx_pch_dpll_get_hw_state(struct drm_i915_private *dev_priv, static void ibx_pch_dpll_prepare(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { - I915_WRITE(PCH_FP0(pll->id), pll->state.hw_state.fp0); - I915_WRITE(PCH_FP1(pll->id), pll->state.hw_state.fp1); + const enum intel_dpll_id id = pll->info->id; + + I915_WRITE(PCH_FP0(id), pll->state.hw_state.fp0); + I915_WRITE(PCH_FP1(id), pll->state.hw_state.fp1); } static void ibx_assert_pch_refclk_enabled(struct drm_i915_private *dev_priv) @@ -381,13 +384,15 @@ static void ibx_assert_pch_refclk_enabled(struct drm_i915_private *dev_priv) static void ibx_pch_dpll_enable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { + const enum intel_dpll_id id = pll->info->id; + /* PCH refclock must be enabled first */ ibx_assert_pch_refclk_enabled(dev_priv); - I915_WRITE(PCH_DPLL(pll->id), pll->state.hw_state.dpll); + I915_WRITE(PCH_DPLL(id), pll->state.hw_state.dpll); /* Wait for the clocks to stabilize. */ - POSTING_READ(PCH_DPLL(pll->id)); + POSTING_READ(PCH_DPLL(id)); udelay(150); /* The pixel multiplier can only be updated once the @@ -395,14 +400,15 @@ static void ibx_pch_dpll_enable(struct drm_i915_private *dev_priv, * * So write it again. */ - I915_WRITE(PCH_DPLL(pll->id), pll->state.hw_state.dpll); - POSTING_READ(PCH_DPLL(pll->id)); + I915_WRITE(PCH_DPLL(id), pll->state.hw_state.dpll); + POSTING_READ(PCH_DPLL(id)); udelay(200); } static void ibx_pch_dpll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { + const enum intel_dpll_id id = pll->info->id; struct drm_device *dev = &dev_priv->drm; struct intel_crtc *crtc; @@ -412,8 +418,8 @@ static void ibx_pch_dpll_disable(struct drm_i915_private *dev_priv, assert_pch_transcoder_disabled(dev_priv, crtc->pipe); } - I915_WRITE(PCH_DPLL(pll->id), 0); - POSTING_READ(PCH_DPLL(pll->id)); + I915_WRITE(PCH_DPLL(id), 0); + POSTING_READ(PCH_DPLL(id)); udelay(200); } @@ -469,8 +475,10 @@ static const struct intel_shared_dpll_funcs ibx_pch_dpll_funcs = { static void hsw_ddi_wrpll_enable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { - I915_WRITE(WRPLL_CTL(pll->id), pll->state.hw_state.wrpll); - POSTING_READ(WRPLL_CTL(pll->id)); + const enum intel_dpll_id id = pll->info->id; + + I915_WRITE(WRPLL_CTL(id), pll->state.hw_state.wrpll); + POSTING_READ(WRPLL_CTL(id)); udelay(20); } @@ -485,11 +493,12 @@ static void hsw_ddi_spll_enable(struct drm_i915_private *dev_priv, static void hsw_ddi_wrpll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; - val = I915_READ(WRPLL_CTL(pll->id)); - I915_WRITE(WRPLL_CTL(pll->id), val & ~WRPLL_PLL_ENABLE); - POSTING_READ(WRPLL_CTL(pll->id)); + val = I915_READ(WRPLL_CTL(id)); + I915_WRITE(WRPLL_CTL(id), val & ~WRPLL_PLL_ENABLE); + POSTING_READ(WRPLL_CTL(id)); } static void hsw_ddi_spll_disable(struct drm_i915_private *dev_priv, @@ -506,12 +515,13 @@ static bool hsw_ddi_wrpll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll, struct intel_dpll_hw_state *hw_state) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) return false; - val = I915_READ(WRPLL_CTL(pll->id)); + val = I915_READ(WRPLL_CTL(id)); hw_state->wrpll = val; intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); @@ -917,13 +927,15 @@ static const struct skl_dpll_regs skl_dpll_regs[4] = { static void skl_ddi_pll_write_ctrl1(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; val = I915_READ(DPLL_CTRL1); - val &= ~(DPLL_CTRL1_HDMI_MODE(pll->id) | DPLL_CTRL1_SSC(pll->id) | - DPLL_CTRL1_LINK_RATE_MASK(pll->id)); - val |= pll->state.hw_state.ctrl1 << (pll->id * 6); + val &= ~(DPLL_CTRL1_HDMI_MODE(id) | + DPLL_CTRL1_SSC(id) | + DPLL_CTRL1_LINK_RATE_MASK(id)); + val |= pll->state.hw_state.ctrl1 << (id * 6); I915_WRITE(DPLL_CTRL1, val); POSTING_READ(DPLL_CTRL1); @@ -933,24 +945,25 @@ static void skl_ddi_pll_enable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { const struct skl_dpll_regs *regs = skl_dpll_regs; + const enum intel_dpll_id id = pll->info->id; skl_ddi_pll_write_ctrl1(dev_priv, pll); - I915_WRITE(regs[pll->id].cfgcr1, pll->state.hw_state.cfgcr1); - I915_WRITE(regs[pll->id].cfgcr2, pll->state.hw_state.cfgcr2); - POSTING_READ(regs[pll->id].cfgcr1); - POSTING_READ(regs[pll->id].cfgcr2); + I915_WRITE(regs[id].cfgcr1, pll->state.hw_state.cfgcr1); + I915_WRITE(regs[id].cfgcr2, pll->state.hw_state.cfgcr2); + POSTING_READ(regs[id].cfgcr1); + POSTING_READ(regs[id].cfgcr2); /* the enable bit is always bit 31 */ - I915_WRITE(regs[pll->id].ctl, - I915_READ(regs[pll->id].ctl) | LCPLL_PLL_ENABLE); + I915_WRITE(regs[id].ctl, + I915_READ(regs[id].ctl) | LCPLL_PLL_ENABLE); if (intel_wait_for_register(dev_priv, DPLL_STATUS, - DPLL_LOCK(pll->id), - DPLL_LOCK(pll->id), + DPLL_LOCK(id), + DPLL_LOCK(id), 5)) - DRM_ERROR("DPLL %d not locked\n", pll->id); + DRM_ERROR("DPLL %d not locked\n", id); } static void skl_ddi_dpll0_enable(struct drm_i915_private *dev_priv, @@ -963,11 +976,12 @@ static void skl_ddi_pll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { const struct skl_dpll_regs *regs = skl_dpll_regs; + const enum intel_dpll_id id = pll->info->id; /* the enable bit is always bit 31 */ - I915_WRITE(regs[pll->id].ctl, - I915_READ(regs[pll->id].ctl) & ~LCPLL_PLL_ENABLE); - POSTING_READ(regs[pll->id].ctl); + I915_WRITE(regs[id].ctl, + I915_READ(regs[id].ctl) & ~LCPLL_PLL_ENABLE); + POSTING_READ(regs[id].ctl); } static void skl_ddi_dpll0_disable(struct drm_i915_private *dev_priv, @@ -981,6 +995,7 @@ static bool skl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, { uint32_t val; const struct skl_dpll_regs *regs = skl_dpll_regs; + const enum intel_dpll_id id = pll->info->id; bool ret; if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) @@ -988,17 +1003,17 @@ static bool skl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, ret = false; - val = I915_READ(regs[pll->id].ctl); + val = I915_READ(regs[id].ctl); if (!(val & LCPLL_PLL_ENABLE)) goto out; val = I915_READ(DPLL_CTRL1); - hw_state->ctrl1 = (val >> (pll->id * 6)) & 0x3f; + hw_state->ctrl1 = (val >> (id * 6)) & 0x3f; /* avoid reading back stale values if HDMI mode is not enabled */ - if (val & DPLL_CTRL1_HDMI_MODE(pll->id)) { - hw_state->cfgcr1 = I915_READ(regs[pll->id].cfgcr1); - hw_state->cfgcr2 = I915_READ(regs[pll->id].cfgcr2); + if (val & DPLL_CTRL1_HDMI_MODE(id)) { + hw_state->cfgcr1 = I915_READ(regs[id].cfgcr1); + hw_state->cfgcr2 = I915_READ(regs[id].cfgcr2); } ret = true; @@ -1014,6 +1029,7 @@ static bool skl_ddi_dpll0_get_hw_state(struct drm_i915_private *dev_priv, { uint32_t val; const struct skl_dpll_regs *regs = skl_dpll_regs; + const enum intel_dpll_id id = pll->info->id; bool ret; if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) @@ -1022,12 +1038,12 @@ static bool skl_ddi_dpll0_get_hw_state(struct drm_i915_private *dev_priv, ret = false; /* DPLL0 is always enabled since it drives CDCLK */ - val = I915_READ(regs[pll->id].ctl); + val = I915_READ(regs[id].ctl); if (WARN_ON(!(val & LCPLL_PLL_ENABLE))) goto out; val = I915_READ(DPLL_CTRL1); - hw_state->ctrl1 = (val >> (pll->id * 6)) & 0x3f; + hw_state->ctrl1 = (val >> (id * 6)) & 0x3f; ret = true; @@ -1427,7 +1443,7 @@ static void bxt_ddi_pll_enable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { uint32_t temp; - enum port port = (enum port)pll->id; /* 1:1 port->PLL mapping */ + enum port port = (enum port)pll->info->id; /* 1:1 port->PLL mapping */ enum dpio_phy phy; enum dpio_channel ch; @@ -1546,7 +1562,7 @@ static void bxt_ddi_pll_enable(struct drm_i915_private *dev_priv, static void bxt_ddi_pll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { - enum port port = (enum port)pll->id; /* 1:1 port->PLL mapping */ + enum port port = (enum port)pll->info->id; /* 1:1 port->PLL mapping */ uint32_t temp; temp = I915_READ(BXT_PORT_PLL_ENABLE(port)); @@ -1569,7 +1585,7 @@ static bool bxt_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll, struct intel_dpll_hw_state *hw_state) { - enum port port = (enum port)pll->id; /* 1:1 port->PLL mapping */ + enum port port = (enum port)pll->info->id; /* 1:1 port->PLL mapping */ uint32_t val; bool ret; enum dpio_phy phy; @@ -1949,38 +1965,39 @@ static const struct intel_dpll_mgr bxt_pll_mgr = { static void cnl_ddi_pll_enable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; /* 1. Enable DPLL power in DPLL_ENABLE. */ - val = I915_READ(CNL_DPLL_ENABLE(pll->id)); + val = I915_READ(CNL_DPLL_ENABLE(id)); val |= PLL_POWER_ENABLE; - I915_WRITE(CNL_DPLL_ENABLE(pll->id), val); + I915_WRITE(CNL_DPLL_ENABLE(id), val); /* 2. Wait for DPLL power state enabled in DPLL_ENABLE. */ if (intel_wait_for_register(dev_priv, - CNL_DPLL_ENABLE(pll->id), + CNL_DPLL_ENABLE(id), PLL_POWER_STATE, PLL_POWER_STATE, 5)) - DRM_ERROR("PLL %d Power not enabled\n", pll->id); + DRM_ERROR("PLL %d Power not enabled\n", id); /* * 3. Configure DPLL_CFGCR0 to set SSC enable/disable, * select DP mode, and set DP link rate. */ val = pll->state.hw_state.cfgcr0; - I915_WRITE(CNL_DPLL_CFGCR0(pll->id), val); + I915_WRITE(CNL_DPLL_CFGCR0(id), val); /* 4. Reab back to ensure writes completed */ - POSTING_READ(CNL_DPLL_CFGCR0(pll->id)); + POSTING_READ(CNL_DPLL_CFGCR0(id)); /* 3. Configure DPLL_CFGCR0 */ /* Avoid touch CFGCR1 if HDMI mode is not enabled */ if (pll->state.hw_state.cfgcr0 & DPLL_CFGCR0_HDMI_MODE) { val = pll->state.hw_state.cfgcr1; - I915_WRITE(CNL_DPLL_CFGCR1(pll->id), val); + I915_WRITE(CNL_DPLL_CFGCR1(id), val); /* 4. Reab back to ensure writes completed */ - POSTING_READ(CNL_DPLL_CFGCR1(pll->id)); + POSTING_READ(CNL_DPLL_CFGCR1(id)); } /* @@ -1993,17 +2010,17 @@ static void cnl_ddi_pll_enable(struct drm_i915_private *dev_priv, */ /* 6. Enable DPLL in DPLL_ENABLE. */ - val = I915_READ(CNL_DPLL_ENABLE(pll->id)); + val = I915_READ(CNL_DPLL_ENABLE(id)); val |= PLL_ENABLE; - I915_WRITE(CNL_DPLL_ENABLE(pll->id), val); + I915_WRITE(CNL_DPLL_ENABLE(id), val); /* 7. Wait for PLL lock status in DPLL_ENABLE. */ if (intel_wait_for_register(dev_priv, - CNL_DPLL_ENABLE(pll->id), + CNL_DPLL_ENABLE(id), PLL_LOCK, PLL_LOCK, 5)) - DRM_ERROR("PLL %d not locked\n", pll->id); + DRM_ERROR("PLL %d not locked\n", id); /* * 8. If the frequency will result in a change to the voltage @@ -2023,6 +2040,7 @@ static void cnl_ddi_pll_enable(struct drm_i915_private *dev_priv, static void cnl_ddi_pll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; /* @@ -2040,17 +2058,17 @@ static void cnl_ddi_pll_disable(struct drm_i915_private *dev_priv, */ /* 3. Disable DPLL through DPLL_ENABLE. */ - val = I915_READ(CNL_DPLL_ENABLE(pll->id)); + val = I915_READ(CNL_DPLL_ENABLE(id)); val &= ~PLL_ENABLE; - I915_WRITE(CNL_DPLL_ENABLE(pll->id), val); + I915_WRITE(CNL_DPLL_ENABLE(id), val); /* 4. Wait for PLL not locked status in DPLL_ENABLE. */ if (intel_wait_for_register(dev_priv, - CNL_DPLL_ENABLE(pll->id), + CNL_DPLL_ENABLE(id), PLL_LOCK, 0, 5)) - DRM_ERROR("PLL %d locked\n", pll->id); + DRM_ERROR("PLL %d locked\n", id); /* * 5. If the frequency will result in a change to the voltage @@ -2062,23 +2080,24 @@ static void cnl_ddi_pll_disable(struct drm_i915_private *dev_priv, */ /* 6. Disable DPLL power in DPLL_ENABLE. */ - val = I915_READ(CNL_DPLL_ENABLE(pll->id)); + val = I915_READ(CNL_DPLL_ENABLE(id)); val &= ~PLL_POWER_ENABLE; - I915_WRITE(CNL_DPLL_ENABLE(pll->id), val); + I915_WRITE(CNL_DPLL_ENABLE(id), val); /* 7. Wait for DPLL power state disabled in DPLL_ENABLE. */ if (intel_wait_for_register(dev_priv, - CNL_DPLL_ENABLE(pll->id), + CNL_DPLL_ENABLE(id), PLL_POWER_STATE, 0, 5)) - DRM_ERROR("PLL %d Power not disabled\n", pll->id); + DRM_ERROR("PLL %d Power not disabled\n", id); } static bool cnl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll, struct intel_dpll_hw_state *hw_state) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; bool ret; @@ -2087,16 +2106,16 @@ static bool cnl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, ret = false; - val = I915_READ(CNL_DPLL_ENABLE(pll->id)); + val = I915_READ(CNL_DPLL_ENABLE(id)); if (!(val & PLL_ENABLE)) goto out; - val = I915_READ(CNL_DPLL_CFGCR0(pll->id)); + val = I915_READ(CNL_DPLL_CFGCR0(id)); hw_state->cfgcr0 = val; /* avoid reading back stale values if HDMI mode is not enabled */ if (val & DPLL_CFGCR0_HDMI_MODE) { - hw_state->cfgcr1 = I915_READ(CNL_DPLL_CFGCR1(pll->id)); + hw_state->cfgcr1 = I915_READ(CNL_DPLL_CFGCR1(id)); } ret = true; @@ -2415,7 +2434,6 @@ void intel_shared_dpll_init(struct drm_device *dev) WARN_ON(i != dpll_info[i].id); dev_priv->shared_dplls[i].info = &dpll_info[i]; - dev_priv->shared_dplls[i].id = dpll_info[i].id; dev_priv->shared_dplls[i].flags = dpll_info[i].flags; } @@ -2476,7 +2494,7 @@ void intel_release_shared_dpll(struct intel_shared_dpll *dpll, struct intel_shared_dpll_state *shared_dpll_state; shared_dpll_state = intel_atomic_get_shared_dpll_state(state); - shared_dpll_state[dpll->id].crtc_mask &= ~(1 << crtc->pipe); + shared_dpll_state[dpll->info->id].crtc_mask &= ~(1 << crtc->pipe); } /** diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.h b/drivers/gpu/drm/i915/intel_dpll_mgr.h index e5ed3e0..7c95ecc 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.h +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.h @@ -213,6 +213,10 @@ struct dpll_info { * @name: DPLL name; used for logging */ const char *name; + /** + * @id: unique indentifier for this DPLL; should match the index in the + * dev_priv->shared_dplls array + */ const int id; /** * @funcs: platform specific hooks @@ -244,12 +248,6 @@ struct intel_shared_dpll { bool on; /** - * @id: unique indentifier for this DPLL; should match the index in the - * dev_priv->shared_dplls array - */ - enum intel_dpll_id id; - - /** * @info: platform specific info */ const struct dpll_info *info; -- cgit v1.1 From 5cd281f679f3ccb517f87589fe4070a3f8014d08 Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Tue, 20 Mar 2018 15:06:36 -0700 Subject: drm/i915: use flags from dpll_info embedded in intel_shared_dpll MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Replace all users of pll->flags to use pll->info.flags. Signed-off-by: Lucas De Marchi Reviewed-by: Ville Syrjälä Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180320220637.21480-7-lucas.demarchi@intel.com --- drivers/gpu/drm/i915/intel_display.c | 2 +- drivers/gpu/drm/i915/intel_dpll_mgr.c | 2 -- drivers/gpu/drm/i915/intel_dpll_mgr.h | 18 ++++++++---------- 3 files changed, 9 insertions(+), 13 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 5f79444..64dd88e 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -11657,7 +11657,7 @@ verify_single_dpll_state(struct drm_i915_private *dev_priv, active = pll->info->funcs->get_hw_state(dev_priv, pll, &dpll_hw_state); - if (!(pll->flags & INTEL_DPLL_ALWAYS_ON)) { + if (!(pll->info->flags & INTEL_DPLL_ALWAYS_ON)) { I915_STATE_WARN(!pll->on && pll->active_mask, "pll in active use but not on in sw tracking\n"); I915_STATE_WARN(pll->on && !pll->active_mask, diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.c b/drivers/gpu/drm/i915/intel_dpll_mgr.c index 48466b1..bda69e1 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.c +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.c @@ -2433,8 +2433,6 @@ void intel_shared_dpll_init(struct drm_device *dev) for (i = 0; dpll_info[i].id >= 0; i++) { WARN_ON(i != dpll_info[i].id); dev_priv->shared_dplls[i].info = &dpll_info[i]; - - dev_priv->shared_dplls[i].flags = dpll_info[i].flags; } dev_priv->dpll_mgr = dpll_mgr; diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.h b/drivers/gpu/drm/i915/intel_dpll_mgr.h index 7c95ecc..e4c01e4 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.h +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.h @@ -222,6 +222,14 @@ struct dpll_info { * @funcs: platform specific hooks */ const struct intel_shared_dpll_funcs *funcs; +#define INTEL_DPLL_ALWAYS_ON (1 << 0) + /** + * @flags: + * + * INTEL_DPLL_ALWAYS_ON + * Inform the state checker that the DPLL is kept enabled even if + * not in use by any CRTC. + */ uint32_t flags; }; @@ -251,16 +259,6 @@ struct intel_shared_dpll { * @info: platform specific info */ const struct dpll_info *info; - -#define INTEL_DPLL_ALWAYS_ON (1 << 0) - /** - * @flags: - * - * INTEL_DPLL_ALWAYS_ON - * Inform the state checker that the DPLL is kept enabled even if - * not in use by any CRTC. - */ - uint32_t flags; }; #define SKL_DPLL0 0 -- cgit v1.1 From 7fd9e829931349ad417579a718213c2b8369bff4 Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Tue, 20 Mar 2018 15:06:37 -0700 Subject: drm/i915: reorder dpll_info members MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Remove 4-bytes hole in this struct an reorder tables accordingly. This also changes the last element of the tables to be more future-proof. Signed-off-by: Lucas De Marchi Reviewed-by: Ville Syrjälä Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180320220637.21480-8-lucas.demarchi@intel.com --- drivers/gpu/drm/i915/intel_dpll_mgr.c | 48 +++++++++++++++++------------------ drivers/gpu/drm/i915/intel_dpll_mgr.h | 13 ++++++---- 2 files changed, 32 insertions(+), 29 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.c b/drivers/gpu/drm/i915/intel_dpll_mgr.c index bda69e1..d5e114e 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.c +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.c @@ -1908,9 +1908,9 @@ struct intel_dpll_mgr { }; static const struct dpll_info pch_plls[] = { - { "PCH DPLL A", DPLL_ID_PCH_PLL_A, &ibx_pch_dpll_funcs, 0 }, - { "PCH DPLL B", DPLL_ID_PCH_PLL_B, &ibx_pch_dpll_funcs, 0 }, - { NULL, -1, NULL, 0 }, + { "PCH DPLL A", &ibx_pch_dpll_funcs, DPLL_ID_PCH_PLL_A, 0 }, + { "PCH DPLL B", &ibx_pch_dpll_funcs, DPLL_ID_PCH_PLL_B, 0 }, + { }, }; static const struct intel_dpll_mgr pch_pll_mgr = { @@ -1920,13 +1920,13 @@ static const struct intel_dpll_mgr pch_pll_mgr = { }; static const struct dpll_info hsw_plls[] = { - { "WRPLL 1", DPLL_ID_WRPLL1, &hsw_ddi_wrpll_funcs, 0 }, - { "WRPLL 2", DPLL_ID_WRPLL2, &hsw_ddi_wrpll_funcs, 0 }, - { "SPLL", DPLL_ID_SPLL, &hsw_ddi_spll_funcs, 0 }, - { "LCPLL 810", DPLL_ID_LCPLL_810, &hsw_ddi_lcpll_funcs, INTEL_DPLL_ALWAYS_ON }, - { "LCPLL 1350", DPLL_ID_LCPLL_1350, &hsw_ddi_lcpll_funcs, INTEL_DPLL_ALWAYS_ON }, - { "LCPLL 2700", DPLL_ID_LCPLL_2700, &hsw_ddi_lcpll_funcs, INTEL_DPLL_ALWAYS_ON }, - { NULL, -1, NULL, }, + { "WRPLL 1", &hsw_ddi_wrpll_funcs, DPLL_ID_WRPLL1, 0 }, + { "WRPLL 2", &hsw_ddi_wrpll_funcs, DPLL_ID_WRPLL2, 0 }, + { "SPLL", &hsw_ddi_spll_funcs, DPLL_ID_SPLL, 0 }, + { "LCPLL 810", &hsw_ddi_lcpll_funcs, DPLL_ID_LCPLL_810, INTEL_DPLL_ALWAYS_ON }, + { "LCPLL 1350", &hsw_ddi_lcpll_funcs, DPLL_ID_LCPLL_1350, INTEL_DPLL_ALWAYS_ON }, + { "LCPLL 2700", &hsw_ddi_lcpll_funcs, DPLL_ID_LCPLL_2700, INTEL_DPLL_ALWAYS_ON }, + { }, }; static const struct intel_dpll_mgr hsw_pll_mgr = { @@ -1936,11 +1936,11 @@ static const struct intel_dpll_mgr hsw_pll_mgr = { }; static const struct dpll_info skl_plls[] = { - { "DPLL 0", DPLL_ID_SKL_DPLL0, &skl_ddi_dpll0_funcs, INTEL_DPLL_ALWAYS_ON }, - { "DPLL 1", DPLL_ID_SKL_DPLL1, &skl_ddi_pll_funcs, 0 }, - { "DPLL 2", DPLL_ID_SKL_DPLL2, &skl_ddi_pll_funcs, 0 }, - { "DPLL 3", DPLL_ID_SKL_DPLL3, &skl_ddi_pll_funcs, 0 }, - { NULL, -1, NULL, }, + { "DPLL 0", &skl_ddi_dpll0_funcs, DPLL_ID_SKL_DPLL0, INTEL_DPLL_ALWAYS_ON }, + { "DPLL 1", &skl_ddi_pll_funcs, DPLL_ID_SKL_DPLL1, 0 }, + { "DPLL 2", &skl_ddi_pll_funcs, DPLL_ID_SKL_DPLL2, 0 }, + { "DPLL 3", &skl_ddi_pll_funcs, DPLL_ID_SKL_DPLL3, 0 }, + { }, }; static const struct intel_dpll_mgr skl_pll_mgr = { @@ -1950,10 +1950,10 @@ static const struct intel_dpll_mgr skl_pll_mgr = { }; static const struct dpll_info bxt_plls[] = { - { "PORT PLL A", DPLL_ID_SKL_DPLL0, &bxt_ddi_pll_funcs, 0 }, - { "PORT PLL B", DPLL_ID_SKL_DPLL1, &bxt_ddi_pll_funcs, 0 }, - { "PORT PLL C", DPLL_ID_SKL_DPLL2, &bxt_ddi_pll_funcs, 0 }, - { NULL, -1, NULL, }, + { "PORT PLL A", &bxt_ddi_pll_funcs, DPLL_ID_SKL_DPLL0, 0 }, + { "PORT PLL B", &bxt_ddi_pll_funcs, DPLL_ID_SKL_DPLL1, 0 }, + { "PORT PLL C", &bxt_ddi_pll_funcs, DPLL_ID_SKL_DPLL2, 0 }, + { }, }; static const struct intel_dpll_mgr bxt_pll_mgr = { @@ -2387,10 +2387,10 @@ static const struct intel_shared_dpll_funcs cnl_ddi_pll_funcs = { }; static const struct dpll_info cnl_plls[] = { - { "DPLL 0", DPLL_ID_SKL_DPLL0, &cnl_ddi_pll_funcs, 0 }, - { "DPLL 1", DPLL_ID_SKL_DPLL1, &cnl_ddi_pll_funcs, 0 }, - { "DPLL 2", DPLL_ID_SKL_DPLL2, &cnl_ddi_pll_funcs, 0 }, - { NULL, -1, NULL, }, + { "DPLL 0", &cnl_ddi_pll_funcs, DPLL_ID_SKL_DPLL0, 0 }, + { "DPLL 1", &cnl_ddi_pll_funcs, DPLL_ID_SKL_DPLL1, 0 }, + { "DPLL 2", &cnl_ddi_pll_funcs, DPLL_ID_SKL_DPLL2, 0 }, + { }, }; static const struct intel_dpll_mgr cnl_pll_mgr = { @@ -2430,7 +2430,7 @@ void intel_shared_dpll_init(struct drm_device *dev) dpll_info = dpll_mgr->dpll_info; - for (i = 0; dpll_info[i].id >= 0; i++) { + for (i = 0; dpll_info[i].name; i++) { WARN_ON(i != dpll_info[i].id); dev_priv->shared_dplls[i].info = &dpll_info[i]; } diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.h b/drivers/gpu/drm/i915/intel_dpll_mgr.h index e4c01e4..4febfaa 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.h +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.h @@ -213,15 +213,18 @@ struct dpll_info { * @name: DPLL name; used for logging */ const char *name; - /** - * @id: unique indentifier for this DPLL; should match the index in the - * dev_priv->shared_dplls array - */ - const int id; + /** * @funcs: platform specific hooks */ const struct intel_shared_dpll_funcs *funcs; + + /** + * @id: unique indentifier for this DPLL; should match the index in the + * dev_priv->shared_dplls array + */ + enum intel_dpll_id id; + #define INTEL_DPLL_ALWAYS_ON (1 << 0) /** * @flags: -- cgit v1.1 From 7056a2bccc3b5afc51f9b35b30a46f0d9219968d Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Mon, 19 Mar 2018 16:19:32 +0200 Subject: drm/dp/mst: Fix off-by-one typo when dump payload table It seems there is a classical off-by-one typo from the beginning when commit ad7f8a1f9ced ("drm/helper: add Displayport multi-stream helper (v0.6)") introduced a new helper. Fix a typo by introducing a macro constant. Cc: Dave Airlie Signed-off-by: Andy Shevchenko Signed-off-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20180319141932.37290-1-andriy.shevchenko@linux.intel.com --- drivers/gpu/drm/drm_dp_mst_topology.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/drm_dp_mst_topology.c b/drivers/gpu/drm/drm_dp_mst_topology.c index 6fac412..6588306 100644 --- a/drivers/gpu/drm/drm_dp_mst_topology.c +++ b/drivers/gpu/drm/drm_dp_mst_topology.c @@ -2941,12 +2941,14 @@ static void drm_dp_mst_dump_mstb(struct seq_file *m, } } +#define DP_PAYLOAD_TABLE_SIZE 64 + static bool dump_dp_payload_table(struct drm_dp_mst_topology_mgr *mgr, char *buf) { int i; - for (i = 0; i < 64; i += 16) { + for (i = 0; i < DP_PAYLOAD_TABLE_SIZE; i += 16) { if (drm_dp_dpcd_read(mgr->aux, DP_PAYLOAD_TABLE_UPDATE_STATUS + i, &buf[i], 16) != 16) @@ -3015,7 +3017,7 @@ void drm_dp_mst_dump_topology(struct seq_file *m, mutex_lock(&mgr->lock); if (mgr->mst_primary) { - u8 buf[64]; + u8 buf[DP_PAYLOAD_TABLE_SIZE]; int ret; ret = drm_dp_dpcd_read(mgr->aux, DP_DPCD_REV, buf, DP_RECEIVER_CAP_SIZE); @@ -3033,8 +3035,7 @@ void drm_dp_mst_dump_topology(struct seq_file *m, seq_printf(m, " revision: hw: %x.%x sw: %x.%x\n", buf[0x9] >> 4, buf[0x9] & 0xf, buf[0xa], buf[0xb]); if (dump_dp_payload_table(mgr, buf)) - seq_printf(m, "payload table: %*ph\n", 63, buf); - + seq_printf(m, "payload table: %*ph\n", DP_PAYLOAD_TABLE_SIZE, buf); } mutex_unlock(&mgr->lock); -- cgit v1.1 From d1a9d710d12485710d424daaa2430fe93bc767f8 Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Tue, 27 Mar 2018 23:47:20 +0300 Subject: drm: prefer inline over __inline__ Remove last users of __inline__. Reviewed-by: Chris Wilson Signed-off-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20180327204722.31246-1-jani.nikula@intel.com --- include/drm/drmP.h | 5 ++--- include/drm/drm_legacy.h | 4 ++-- 2 files changed, 4 insertions(+), 5 deletions(-) diff --git a/include/drm/drmP.h b/include/drm/drmP.h index c6666cd..4bbef06 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -123,8 +123,7 @@ static inline bool drm_drv_uses_atomic_modeset(struct drm_device *dev) #define DRM_SWITCH_POWER_CHANGING 2 #define DRM_SWITCH_POWER_DYNAMIC_OFF 3 -static __inline__ int drm_core_check_feature(struct drm_device *dev, - int feature) +static inline int drm_core_check_feature(struct drm_device *dev, int feature) { return ((dev->driver->driver_features & feature) ? 1 : 0); } @@ -143,7 +142,7 @@ static __inline__ int drm_core_check_feature(struct drm_device *dev, /*@}*/ /* returns true if currently okay to sleep */ -static __inline__ bool drm_can_sleep(void) +static inline bool drm_can_sleep(void) { if (in_atomic() || in_dbg_master() || irqs_disabled()) return false; diff --git a/include/drm/drm_legacy.h b/include/drm/drm_legacy.h index cf0e7d8..8fad66f 100644 --- a/include/drm/drm_legacy.h +++ b/include/drm/drm_legacy.h @@ -194,8 +194,8 @@ void drm_legacy_ioremap(struct drm_local_map *map, struct drm_device *dev); void drm_legacy_ioremap_wc(struct drm_local_map *map, struct drm_device *dev); void drm_legacy_ioremapfree(struct drm_local_map *map, struct drm_device *dev); -static __inline__ struct drm_local_map *drm_legacy_findmap(struct drm_device *dev, - unsigned int token) +static inline struct drm_local_map *drm_legacy_findmap(struct drm_device *dev, + unsigned int token) { struct drm_map_list *_entry; list_for_each_entry(_entry, &dev->maplist, head) -- cgit v1.1 From 885a31cb6c752d5403adc6389894c27560fc6e6c Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Tue, 27 Mar 2018 23:47:21 +0300 Subject: drm: remove old documentation comment cruft from drmP.h Throw out the leftovers. Reviewed-by: Chris Wilson Signed-off-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20180327204722.31246-2-jani.nikula@intel.com --- include/drm/drmP.h | 21 --------------------- 1 file changed, 21 deletions(-) diff --git a/include/drm/drmP.h b/include/drm/drmP.h index 4bbef06..b5d52a3 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -95,14 +95,6 @@ struct dma_buf_attachment; struct pci_dev; struct pci_controller; -/***********************************************************************/ -/** \name DRM template customization defaults */ -/*@{*/ - -/***********************************************************************/ -/** \name Internal types and structures */ -/*@{*/ - #define DRM_IF_VERSION(maj, min) (maj << 16 | min) /** @@ -128,19 +120,6 @@ static inline int drm_core_check_feature(struct drm_device *dev, int feature) return ((dev->driver->driver_features & feature) ? 1 : 0); } -/******************************************************************/ -/** \name Internal function definitions */ -/*@{*/ - - /* Driver support (drm_drv.h) */ - -/* - * These are exported to drivers so that they can implement fencing using - * DMA quiscent + idle. DMA quiescent usually requires the hardware lock. - */ - -/*@}*/ - /* returns true if currently okay to sleep */ static inline bool drm_can_sleep(void) { -- cgit v1.1 From f4392860b4fe55d7d7cadaa64743c9b2466e4fd8 Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Tue, 27 Mar 2018 23:47:22 +0300 Subject: drm: make drm_core_check_feature() bool that it is Bool is the more appropriate return type here, use it. Reviewed-by: Chris Wilson Signed-off-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20180327204722.31246-3-jani.nikula@intel.com --- include/drm/drmP.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/include/drm/drmP.h b/include/drm/drmP.h index b5d52a3..f5099c1 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -115,9 +115,9 @@ static inline bool drm_drv_uses_atomic_modeset(struct drm_device *dev) #define DRM_SWITCH_POWER_CHANGING 2 #define DRM_SWITCH_POWER_DYNAMIC_OFF 3 -static inline int drm_core_check_feature(struct drm_device *dev, int feature) +static inline bool drm_core_check_feature(struct drm_device *dev, int feature) { - return ((dev->driver->driver_features & feature) ? 1 : 0); + return dev->driver->driver_features & feature; } /* returns true if currently okay to sleep */ -- cgit v1.1 From 49efffc7fbd48d5ea3d0dd60c218c7502d4a179d Mon Sep 17 00:00:00 2001 From: Peter Ujfalusi Date: Wed, 21 Mar 2018 12:20:24 +0200 Subject: drm: Add drm_mode_config->normalize_zpos boolean Instead of drivers duplicating the drm_atomic_helper_check() code to be able to normalize the zpos they can use the normalize_zpos flag to let the drm core to do it. Signed-off-by: Peter Ujfalusi Signed-off-by: Tomi Valkeinen Link: https://patchwork.freedesktop.org/patch/msgid/20180321102029.15248-2-peter.ujfalusi@ti.com --- drivers/gpu/drm/drm_atomic_helper.c | 11 +++++++++++ include/drm/drm_mode_config.h | 8 ++++++++ include/drm/drm_plane.h | 4 ++-- 3 files changed, 21 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index c356545..d63c806 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -875,6 +875,11 @@ EXPORT_SYMBOL(drm_atomic_helper_check_planes); * functions depend upon an updated adjusted_mode.clock to e.g. properly compute * watermarks. * + * Note that zpos normalization will add all enable planes to the state which + * might not desired for some drivers. + * For example enable/disable of a cursor plane which have fixed zpos value + * would trigger all other enabled planes to be forced to the state change. + * * RETURNS: * Zero for success or -errno */ @@ -887,6 +892,12 @@ int drm_atomic_helper_check(struct drm_device *dev, if (ret) return ret; + if (dev->mode_config.normalize_zpos) { + ret = drm_atomic_normalize_zpos(dev, state); + if (ret) + return ret; + } + ret = drm_atomic_helper_check_planes(dev, state); if (ret) return ret; diff --git a/include/drm/drm_mode_config.h b/include/drm/drm_mode_config.h index 7569f22..33b3a96 100644 --- a/include/drm/drm_mode_config.h +++ b/include/drm/drm_mode_config.h @@ -796,6 +796,14 @@ struct drm_mode_config { bool allow_fb_modifiers; /** + * @normalize_zpos: + * + * If true the drm core will call drm_atomic_normalize_zpos() as part of + * atomic mode checking from drm_atomic_helper_check() + */ + bool normalize_zpos; + + /** * @modifiers_property: Plane property to list support modifier/format * combination. */ diff --git a/include/drm/drm_plane.h b/include/drm/drm_plane.h index f7bf4a4..d6da26d 100644 --- a/include/drm/drm_plane.h +++ b/include/drm/drm_plane.h @@ -51,8 +51,8 @@ struct drm_modeset_acquire_ctx; * plane with a lower ID. * @normalized_zpos: normalized value of zpos: unique, range from 0 to N-1 * where N is the number of active planes for given crtc. Note that - * the driver must call drm_atomic_normalize_zpos() to update this before - * it can be trusted. + * the driver must set drm_mode_config.normalize_zpos or call + * drm_atomic_normalize_zpos() to update this before it can be trusted. * @src: clipped source coordinates of the plane (in 16.16) * @dst: clipped destination coordinates of the plane * @state: backpointer to global drm_atomic_state -- cgit v1.1 From a7da5cfe0cd6d36af6dc05ee4aa3e506c88e8f0e Mon Sep 17 00:00:00 2001 From: Peter Ujfalusi Date: Wed, 21 Mar 2018 12:20:25 +0200 Subject: drm/exynos: Let core take care of normalizing the zpos Instead of re-implementing the drm_atomic_helper_check() locally with just adding drm_atomic_normalize_zpos() into it, set the drm_mode_config->normalize_zpos. Signed-off-by: Peter Ujfalusi CC: Inki Dae CC: Joonyoung Shim CC: Seung-Woo Kim CC: Kyungmin Park Acked-by: Daniel Vetter Acked-by: Inki Dae Signed-off-by: Tomi Valkeinen Link: https://patchwork.freedesktop.org/patch/msgid/20180321102029.15248-3-peter.ujfalusi@ti.com --- drivers/gpu/drm/exynos/exynos_drm_drv.c | 20 -------------------- drivers/gpu/drm/exynos/exynos_drm_drv.h | 1 - drivers/gpu/drm/exynos/exynos_drm_fb.c | 4 +++- 3 files changed, 3 insertions(+), 22 deletions(-) diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c index a518e9c..39284bb 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_drv.c +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c @@ -37,26 +37,6 @@ #define DRIVER_MAJOR 1 #define DRIVER_MINOR 0 -int exynos_atomic_check(struct drm_device *dev, - struct drm_atomic_state *state) -{ - int ret; - - ret = drm_atomic_helper_check_modeset(dev, state); - if (ret) - return ret; - - ret = drm_atomic_normalize_zpos(dev, state); - if (ret) - return ret; - - ret = drm_atomic_helper_check_planes(dev, state); - if (ret) - return ret; - - return ret; -} - static int exynos_drm_open(struct drm_device *dev, struct drm_file *file) { struct drm_exynos_file_private *file_priv; diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h index df2262f..075957c 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_drv.h +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h @@ -275,7 +275,6 @@ static inline int exynos_dpi_bind(struct drm_device *dev, int exynos_atomic_commit(struct drm_device *dev, struct drm_atomic_state *state, bool nonblock); -int exynos_atomic_check(struct drm_device *dev, struct drm_atomic_state *state); extern struct platform_driver fimd_driver; diff --git a/drivers/gpu/drm/exynos/exynos_drm_fb.c b/drivers/gpu/drm/exynos/exynos_drm_fb.c index 0faaf82..2379d73 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fb.c +++ b/drivers/gpu/drm/exynos/exynos_drm_fb.c @@ -206,7 +206,7 @@ static struct drm_mode_config_helper_funcs exynos_drm_mode_config_helpers = { static const struct drm_mode_config_funcs exynos_drm_mode_config_funcs = { .fb_create = exynos_user_fb_create, .output_poll_changed = drm_fb_helper_output_poll_changed, - .atomic_check = exynos_atomic_check, + .atomic_check = drm_atomic_helper_check, .atomic_commit = drm_atomic_helper_commit, }; @@ -227,4 +227,6 @@ void exynos_drm_mode_config_init(struct drm_device *dev) dev->mode_config.helper_private = &exynos_drm_mode_config_helpers; dev->mode_config.allow_fb_modifiers = true; + + dev->mode_config.normalize_zpos = true; } -- cgit v1.1 From a18301b9f556101b4b87cd83b050c553652e91e4 Mon Sep 17 00:00:00 2001 From: Peter Ujfalusi Date: Wed, 21 Mar 2018 12:20:26 +0200 Subject: drm/tegra: Let core take care of normalizing the zpos Set the drm_mode_config->normalize_zpos and call the generic drm_atomic_helper_check() instead of duplicating it within tegra_atomic_check(). Call tegra_display_hub_atomic_check() after the drm_atomic_helpre_check() returned without error. Signed-off-by: Peter Ujfalusi CC: Thierry Reding Signed-off-by: Tomi Valkeinen Link: https://patchwork.freedesktop.org/patch/msgid/20180321102029.15248-4-peter.ujfalusi@ti.com --- drivers/gpu/drm/tegra/drm.c | 21 ++++----------------- 1 file changed, 4 insertions(+), 17 deletions(-) diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index e20e013..ac11211 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -38,26 +38,11 @@ static int tegra_atomic_check(struct drm_device *drm, { int err; - err = drm_atomic_helper_check_modeset(drm, state); + err = drm_atomic_helper_check(drm, state); if (err < 0) return err; - err = tegra_display_hub_atomic_check(drm, state); - if (err < 0) - return err; - - err = drm_atomic_normalize_zpos(drm, state); - if (err < 0) - return err; - - err = drm_atomic_helper_check_planes(drm, state); - if (err < 0) - return err; - - if (state->legacy_cursor_update) - state->async_update = !drm_atomic_helper_async_check(drm, state); - - return 0; + return tegra_display_hub_atomic_check(drm, state); } static const struct drm_mode_config_funcs tegra_drm_mode_config_funcs = { @@ -151,6 +136,8 @@ static int tegra_drm_load(struct drm_device *drm, unsigned long flags) drm->mode_config.allow_fb_modifiers = true; + drm->mode_config.normalize_zpos = true; + drm->mode_config.funcs = &tegra_drm_mode_config_funcs; drm->mode_config.helper_private = &tegra_drm_mode_config_helpers; -- cgit v1.1 From 352f9a8419df87b925ccc7fc56f1a75aa2290e93 Mon Sep 17 00:00:00 2001 From: Peter Ujfalusi Date: Wed, 21 Mar 2018 12:20:27 +0200 Subject: drm/sti: Let core take care of normalizing the zpos Instead of re-implementing the drm_atomic_helper_check() locally with just adding drm_atomic_normalize_zpos() into it, set the drm_mode_config->normalize_zpos. Note: the drm_atomic_helper_check() now includes if (state->legacy_cursor_update) state->async_update = !drm_atomic_helper_async_check(drm, state); which was added after the driver moved away from using it (38d868e41c4b9250d5a115c049dc2d48f4909581 drm: Don't force all planes to be added to the state due to zpos) Signed-off-by: Peter Ujfalusi CC: Benjamin Gaignard CC: Vincent Abriou Acked-by: Daniel Vetter Acked-by: Benjamin Gaignard Signed-off-by: Tomi Valkeinen Link: https://patchwork.freedesktop.org/patch/msgid/20180321102029.15248-5-peter.ujfalusi@ti.com --- drivers/gpu/drm/sti/sti_drv.c | 24 +++--------------------- 1 file changed, 3 insertions(+), 21 deletions(-) diff --git a/drivers/gpu/drm/sti/sti_drv.c b/drivers/gpu/drm/sti/sti_drv.c index 55b6967..90c46b4 100644 --- a/drivers/gpu/drm/sti/sti_drv.c +++ b/drivers/gpu/drm/sti/sti_drv.c @@ -119,30 +119,10 @@ err: return ret; } -static int sti_atomic_check(struct drm_device *dev, - struct drm_atomic_state *state) -{ - int ret; - - ret = drm_atomic_helper_check_modeset(dev, state); - if (ret) - return ret; - - ret = drm_atomic_normalize_zpos(dev, state); - if (ret) - return ret; - - ret = drm_atomic_helper_check_planes(dev, state); - if (ret) - return ret; - - return ret; -} - static const struct drm_mode_config_funcs sti_mode_config_funcs = { .fb_create = drm_gem_fb_create, .output_poll_changed = drm_fb_helper_output_poll_changed, - .atomic_check = sti_atomic_check, + .atomic_check = drm_atomic_helper_check, .atomic_commit = drm_atomic_helper_commit, }; @@ -160,6 +140,8 @@ static void sti_mode_config_init(struct drm_device *dev) dev->mode_config.max_height = STI_MAX_FB_HEIGHT; dev->mode_config.funcs = &sti_mode_config_funcs; + + dev->mode_config.normalize_zpos = true; } DEFINE_DRM_GEM_CMA_FOPS(sti_driver_fops); -- cgit v1.1 From 75def7785f4901b65a89dc99ea9506b1395242fa Mon Sep 17 00:00:00 2001 From: Peter Ujfalusi Date: Wed, 21 Mar 2018 12:20:28 +0200 Subject: drm: rcar-du: Let core take care of normalizing the zpos Set the drm_mode_config->normalize_zpos and call drm_atomic_helper_check() from rcar_du_atomic_check() instead of re implementing the function locally. Signed-off-by: Peter Ujfalusi CC: Laurent Pinchart Acked-by: Daniel Vetter Signed-off-by: Tomi Valkeinen Link: https://patchwork.freedesktop.org/patch/msgid/20180321102029.15248-6-peter.ujfalusi@ti.com --- drivers/gpu/drm/rcar-du/rcar_du_kms.c | 11 ++--------- 1 file changed, 2 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c index 0329b35..ab59d20 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c @@ -233,15 +233,7 @@ static int rcar_du_atomic_check(struct drm_device *dev, struct rcar_du_device *rcdu = dev->dev_private; int ret; - ret = drm_atomic_helper_check_modeset(dev, state); - if (ret) - return ret; - - ret = drm_atomic_normalize_zpos(dev, state); - if (ret) - return ret; - - ret = drm_atomic_helper_check_planes(dev, state); + ret = drm_atomic_helper_check(dev, state); if (ret) return ret; @@ -529,6 +521,7 @@ int rcar_du_modeset_init(struct rcar_du_device *rcdu) dev->mode_config.min_height = 0; dev->mode_config.max_width = 4095; dev->mode_config.max_height = 2047; + dev->mode_config.normalize_zpos = true; dev->mode_config.funcs = &rcar_du_mode_config_funcs; dev->mode_config.helper_private = &rcar_du_mode_config_helper; -- cgit v1.1 From 23936ba940fbccf08f9f61d7c6d39ba0feb383bc Mon Sep 17 00:00:00 2001 From: Peter Ujfalusi Date: Wed, 21 Mar 2018 12:20:29 +0200 Subject: drm/omap: Use normalized zpos for plane placement Planes with identical zpos value will result undefined behavior: disappearing planes, screen flickering and it is not supported by the hardware. Use normalized zpos to make sure that we don't encounter invalid configuration. Signed-off-by: Peter Ujfalusi CC: Tomi Valkeinen Acked-by: Daniel Vetter Reviewed-by: Tomi Valkeinen Signed-off-by: Tomi Valkeinen Link: https://patchwork.freedesktop.org/patch/msgid/20180321102029.15248-7-peter.ujfalusi@ti.com --- drivers/gpu/drm/omapdrm/omap_drv.c | 3 +++ drivers/gpu/drm/omapdrm/omap_plane.c | 2 +- 2 files changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/omapdrm/omap_drv.c b/drivers/gpu/drm/omapdrm/omap_drv.c index 3632854..ef3b0e3 100644 --- a/drivers/gpu/drm/omapdrm/omap_drv.c +++ b/drivers/gpu/drm/omapdrm/omap_drv.c @@ -319,6 +319,9 @@ static int omap_modeset_init(struct drm_device *dev) dev->mode_config.max_width = 8192; dev->mode_config.max_height = 8192; + /* We want the zpos to be normalized */ + dev->mode_config.normalize_zpos = true; + dev->mode_config.funcs = &omap_mode_config_funcs; dev->mode_config.helper_private = &omap_mode_config_helper_funcs; diff --git a/drivers/gpu/drm/omapdrm/omap_plane.c b/drivers/gpu/drm/omapdrm/omap_plane.c index 2899435..161233c 100644 --- a/drivers/gpu/drm/omapdrm/omap_plane.c +++ b/drivers/gpu/drm/omapdrm/omap_plane.c @@ -65,7 +65,7 @@ static void omap_plane_atomic_update(struct drm_plane *plane, info.rotation_type = OMAP_DSS_ROT_NONE; info.rotation = DRM_MODE_ROTATE_0; info.global_alpha = 0xff; - info.zorder = state->zpos; + info.zorder = state->normalized_zpos; /* update scanout: */ omap_framebuffer_update_scanout(state->fb, state, &info); -- cgit v1.1 From fbe6f8f2a648584b97beeaaaeff75b795fb3c6cb Mon Sep 17 00:00:00 2001 From: Yaodong Li Date: Thu, 22 Mar 2018 16:59:22 -0700 Subject: drm/i915: Use correct reST syntax for WOPCM and GuC kernel-doc diagrams GuC Address Space and WOPCM Layout diagrams won't be generated correctly by sphinx build if not using proper reST syntax. This patch uses reST literal blocks to make sure GuC Address Space and WOPCM Layout diagrams to be generated correctly, and it also corrects some errors in the diagram description. v2: - Fixed errors in diagram description v3: - Updated GuC Address Space kernel-doc based on Michal's suggestion v4: - Added WOPCM layout and GuC address space docs into i915.rst (Joonas) Signed-off-by: Jackie Li Cc: Michal Wajdeczko Cc: Sagar Arun Kamble Cc: Joonas Lahtinen Reviewed-by: Joonas Lahtinen Signed-off-by: Joonas Lahtinen Link: https://patchwork.freedesktop.org/patch/msgid/1521763162-11424-1-git-send-email-yaodong.li@intel.com --- Documentation/gpu/i915.rst | 15 ++++++++++ drivers/gpu/drm/i915/intel_guc.c | 56 ++++++++++++++++++++------------------ drivers/gpu/drm/i915/intel_wopcm.c | 44 ++++++++++++++++-------------- 3 files changed, 67 insertions(+), 48 deletions(-) diff --git a/Documentation/gpu/i915.rst b/Documentation/gpu/i915.rst index 41dc881..7ecad71 100644 --- a/Documentation/gpu/i915.rst +++ b/Documentation/gpu/i915.rst @@ -335,6 +335,15 @@ objects, which has the goal to make space in gpu virtual address spaces. .. kernel-doc:: drivers/gpu/drm/i915/i915_gem_shrinker.c :internal: +WOPCM +===== + +WOPCM Layout +------------ + +.. kernel-doc:: drivers/gpu/drm/i915/intel_wopcm.c + :doc: WOPCM Layout + GuC === @@ -359,6 +368,12 @@ GuC Firmware Layout .. kernel-doc:: drivers/gpu/drm/i915/intel_guc_fwif.h :doc: GuC Firmware Layout +GuC Address Space +----------------- + +.. kernel-doc:: drivers/gpu/drm/i915/intel_guc.c + :doc: GuC Address Space + Tracing ======= diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index 8f93f5b..c5f64c7 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -494,35 +494,37 @@ int intel_guc_resume(struct intel_guc *guc) /** * DOC: GuC Address Space * - * The layout of GuC address space is shown as below: + * The layout of GuC address space is shown below: * - * +==============> +====================+ <== GUC_GGTT_TOP - * ^ | | - * | | | - * | | DRAM | - * | | Memory | - * | | | - * GuC | | - * Address +========> +====================+ <== WOPCM Top - * Space ^ | HW contexts RSVD | - * | | | WOPCM | - * | | +==> +--------------------+ <== GuC WOPCM Top - * | GuC ^ | | - * | GGTT | | | - * | Pin GuC | GuC | - * | Bias WOPCM | WOPCM | - * | | Size | | - * | | | | | - * v v v | | - * +=====+=====+==> +====================+ <== GuC WOPCM Base - * | Non-GuC WOPCM | - * | (HuC/Reserved) | - * +====================+ <== WOPCM Base + * :: * - * The lower part [0, GuC ggtt_pin_bias) is mapped to WOPCM which consists of - * GuC WOPCM and WOPCM reserved for other usage (e.g.RC6 context). The value of - * the GuC ggtt_pin_bias is determined by the actually GuC WOPCM size which is - * set in GUC_WOPCM_SIZE register. + * +==============> +====================+ <== GUC_GGTT_TOP + * ^ | | + * | | | + * | | DRAM | + * | | Memory | + * | | | + * GuC | | + * Address +========> +====================+ <== WOPCM Top + * Space ^ | HW contexts RSVD | + * | | | WOPCM | + * | | +==> +--------------------+ <== GuC WOPCM Top + * | GuC ^ | | + * | GGTT | | | + * | Pin GuC | GuC | + * | Bias WOPCM | WOPCM | + * | | Size | | + * | | | | | + * v v v | | + * +=====+=====+==> +====================+ <== GuC WOPCM Base + * | Non-GuC WOPCM | + * | (HuC/Reserved) | + * +====================+ <== WOPCM Base + * + * The lower part of GuC Address Space [0, ggtt_pin_bias) is mapped to WOPCM + * while upper part of GuC Address Space [ggtt_pin_bias, GUC_GGTT_TOP) is mapped + * to DRAM. The value of the GuC ggtt_pin_bias is determined by WOPCM size and + * actual GuC WOPCM size. */ /** diff --git a/drivers/gpu/drm/i915/intel_wopcm.c b/drivers/gpu/drm/i915/intel_wopcm.c index 4117886..74bf76f 100644 --- a/drivers/gpu/drm/i915/intel_wopcm.c +++ b/drivers/gpu/drm/i915/intel_wopcm.c @@ -11,28 +11,30 @@ * DOC: WOPCM Layout * * The layout of the WOPCM will be fixed after writing to GuC WOPCM size and - * offset registers whose are calculated are determined by size of HuC/GuC - * firmware size and set of hw requirements/restrictions as shown below: + * offset registers whose values are calculated and determined by HuC/GuC + * firmware size and set of hardware requirements/restrictions as shown below: * - * +=========> +====================+ <== WOPCM Top - * ^ | HW contexts RSVD | - * | +===> +====================+ <== GuC WOPCM Top - * | ^ | | - * | | | | - * | | | | - * | GuC | | - * | WOPCM | | - * | Size +--------------------+ - * WOPCM | | GuC FW RSVD | - * | | +--------------------+ - * | | | GuC Stack RSVD | - * | | +------------------- + - * | v | GuC WOPCM RSVD | - * | +===> +====================+ <== GuC WOPCM base - * | | WOPCM RSVD | - * | +------------------- + <== HuC Firmware Top - * v | HuC FW | - * +=========> +====================+ <== WOPCM Base + * :: + * + * +=========> +====================+ <== WOPCM Top + * ^ | HW contexts RSVD | + * | +===> +====================+ <== GuC WOPCM Top + * | ^ | | + * | | | | + * | | | | + * | GuC | | + * | WOPCM | | + * | Size +--------------------+ + * WOPCM | | GuC FW RSVD | + * | | +--------------------+ + * | | | GuC Stack RSVD | + * | | +------------------- + + * | v | GuC WOPCM RSVD | + * | +===> +====================+ <== GuC WOPCM base + * | | WOPCM RSVD | + * | +------------------- + <== HuC Firmware Top + * v | HuC FW | + * +=========> +====================+ <== WOPCM Base * * GuC accessible WOPCM starts at GuC WOPCM base and ends at GuC WOPCM top. * The top part of the WOPCM is reserved for hardware contexts (e.g. RC6 -- cgit v1.1 From 606f1fc5cf2c27e86f2ea03d293e77fac789fae5 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Tue, 27 Mar 2018 10:23:52 +0200 Subject: staging/vboxvideo: Use gem_free_object_unlocked vboxvideo doesn't use dev->struct_mutex and therefore has no need to use gem_free_object. Signed-off-by: Daniel Vetter Cc: Greg Kroah-Hartman Cc: Hans de Goede Cc: Michael Thayer Cc: Colin Ian King Cc: Daniel Vetter Cc: Stephen Rothwell Reviewed-by: Greg Kroah-Hartman Link: https://patchwork.freedesktop.org/patch/msgid/20180327082356.24516-1-daniel.vetter@ffwll.ch --- drivers/staging/vboxvideo/vbox_drv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/staging/vboxvideo/vbox_drv.c b/drivers/staging/vboxvideo/vbox_drv.c index e18642e..f6d26be 100644 --- a/drivers/staging/vboxvideo/vbox_drv.c +++ b/drivers/staging/vboxvideo/vbox_drv.c @@ -242,7 +242,7 @@ static struct drm_driver driver = { .minor = DRIVER_MINOR, .patchlevel = DRIVER_PATCHLEVEL, - .gem_free_object = vbox_gem_free_object, + .gem_free_object_unlocked = vbox_gem_free_object, .dumb_create = vbox_dumb_create, .dumb_map_offset = vbox_dumb_mmap_offset, .dumb_destroy = drm_gem_dumb_destroy, -- cgit v1.1 From fcb1e57f79c0ecf4c85fecd2294a469367cbddd0 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Tue, 27 Mar 2018 10:23:53 +0200 Subject: drm/rockchip: fixup comment for gem_free_object_unlocked MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Just want to clean out all grep hits. gem_free_object is deprecated. Signed-off-by: Daniel Vetter Cc: Sandy Huang Cc: "Heiko Stübner" Cc: linux-arm-kernel@lists.infradead.org Cc: linux-rockchip@lists.infradead.org Reviewed-by: Sean Paul Link: https://patchwork.freedesktop.org/patch/msgid/20180327082356.24516-2-daniel.vetter@ffwll.ch --- drivers/gpu/drm/rockchip/rockchip_drm_gem.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_gem.c b/drivers/gpu/drm/rockchip/rockchip_drm_gem.c index 074db7a..a8db758 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_gem.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_gem.c @@ -357,8 +357,8 @@ err_free_rk_obj: } /* - * rockchip_gem_free_object - (struct drm_driver)->gem_free_object callback - * function + * rockchip_gem_free_object - (struct drm_driver)->gem_free_object_unlocked + * callback function */ void rockchip_gem_free_object(struct drm_gem_object *obj) { -- cgit v1.1 From ae358dacd217370cc362f1674712c4e9246ace8d Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Tue, 27 Mar 2018 10:23:54 +0200 Subject: drm/udl: Get rid of dev->struct_mutex usage It's only used to protect our page list, and only when we know we have a full reference. This means none of these code paths can ever race with the final unref, and hence we do not need dev->struct_mutex serialization and can simply switch to our own locking. For more context the only magic the locked gem_free_object provides is that it prevents concurrent final unref (and destruction) of gem objects while anyone is holding dev->struct_mutex. This was used by i915 (and other drivers) to implement eviction handling with less headaches. Signed-off-by: Daniel Vetter Cc: Dave Airlie Reviewed-by: Sean Paul Link: https://patchwork.freedesktop.org/patch/msgid/20180327082356.24516-3-daniel.vetter@ffwll.ch --- drivers/gpu/drm/udl/udl_dmabuf.c | 5 +++-- drivers/gpu/drm/udl/udl_drv.c | 2 +- drivers/gpu/drm/udl/udl_drv.h | 2 ++ drivers/gpu/drm/udl/udl_gem.c | 5 +++-- drivers/gpu/drm/udl/udl_main.c | 2 ++ 5 files changed, 11 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/udl/udl_dmabuf.c b/drivers/gpu/drm/udl/udl_dmabuf.c index 2867ed1..0a20695 100644 --- a/drivers/gpu/drm/udl/udl_dmabuf.c +++ b/drivers/gpu/drm/udl/udl_dmabuf.c @@ -76,6 +76,7 @@ static struct sg_table *udl_map_dma_buf(struct dma_buf_attachment *attach, struct udl_drm_dmabuf_attachment *udl_attach = attach->priv; struct udl_gem_object *obj = to_udl_bo(attach->dmabuf->priv); struct drm_device *dev = obj->base.dev; + struct udl_device *udl = dev->dev_private; struct scatterlist *rd, *wr; struct sg_table *sgt = NULL; unsigned int i; @@ -112,7 +113,7 @@ static struct sg_table *udl_map_dma_buf(struct dma_buf_attachment *attach, return ERR_PTR(-ENOMEM); } - mutex_lock(&dev->struct_mutex); + mutex_lock(&udl->gem_lock); rd = obj->sg->sgl; wr = sgt->sgl; @@ -137,7 +138,7 @@ static struct sg_table *udl_map_dma_buf(struct dma_buf_attachment *attach, attach->priv = udl_attach; err_unlock: - mutex_unlock(&dev->struct_mutex); + mutex_unlock(&udl->gem_lock); return sgt; } diff --git a/drivers/gpu/drm/udl/udl_drv.c b/drivers/gpu/drm/udl/udl_drv.c index 3c45a30..9ef515d 100644 --- a/drivers/gpu/drm/udl/udl_drv.c +++ b/drivers/gpu/drm/udl/udl_drv.c @@ -53,7 +53,7 @@ static struct drm_driver driver = { .unload = udl_driver_unload, /* gem hooks */ - .gem_free_object = udl_gem_free_object, + .gem_free_object_unlocked = udl_gem_free_object, .gem_vm_ops = &udl_gem_vm_ops, .dumb_create = udl_dumb_create, diff --git a/drivers/gpu/drm/udl/udl_drv.h b/drivers/gpu/drm/udl/udl_drv.h index 2a75ab8..55c0cc3 100644 --- a/drivers/gpu/drm/udl/udl_drv.h +++ b/drivers/gpu/drm/udl/udl_drv.h @@ -54,6 +54,8 @@ struct udl_device { struct usb_device *udev; struct drm_crtc *crtc; + struct mutex gem_lock; + int sku_pixel_limit; struct urb_list urbs; diff --git a/drivers/gpu/drm/udl/udl_gem.c b/drivers/gpu/drm/udl/udl_gem.c index dee6bd9..9a15cce 100644 --- a/drivers/gpu/drm/udl/udl_gem.c +++ b/drivers/gpu/drm/udl/udl_gem.c @@ -214,9 +214,10 @@ int udl_gem_mmap(struct drm_file *file, struct drm_device *dev, { struct udl_gem_object *gobj; struct drm_gem_object *obj; + struct udl_device *udl = dev->dev_private; int ret = 0; - mutex_lock(&dev->struct_mutex); + mutex_lock(&udl->gem_lock); obj = drm_gem_object_lookup(file, handle); if (obj == NULL) { ret = -ENOENT; @@ -236,6 +237,6 @@ int udl_gem_mmap(struct drm_file *file, struct drm_device *dev, out: drm_gem_object_put(&gobj->base); unlock: - mutex_unlock(&dev->struct_mutex); + mutex_unlock(&udl->gem_lock); return ret; } diff --git a/drivers/gpu/drm/udl/udl_main.c b/drivers/gpu/drm/udl/udl_main.c index f1ec452..d518de8 100644 --- a/drivers/gpu/drm/udl/udl_main.c +++ b/drivers/gpu/drm/udl/udl_main.c @@ -324,6 +324,8 @@ int udl_driver_load(struct drm_device *dev, unsigned long flags) udl->ddev = dev; dev->dev_private = udl; + mutex_init(&udl->gem_lock); + if (!udl_parse_vendor_descriptor(dev, udl->udev)) { ret = -ENODEV; DRM_ERROR("firmware not recognized. Assume incompatible device\n"); -- cgit v1.1 From 0c9c7fd00e17907efb35697ecb9f2df39a0b536c Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 22:27:37 +0200 Subject: drm/simple-kms-helper: Plumb plane state to the enable hook MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit tinydrm enable hook wants to play around with the new fb in .atomic_enable(), thus we'll need access to the plane state. Performed with coccinelle: @r1@ identifier F =~ ".*enable$"; identifier P, CS; @@ F( struct drm_simple_display_pipe *P ,struct drm_crtc_state *CS + ,struct drm_plane_state *plane_state ) { ... } @@ struct drm_simple_display_pipe *P; expression E; @@ { + struct drm_plane *plane; ... + plane = &P->plane; P->funcs->enable(P ,E + ,plane->state ); ... } @@ identifier P, CS; @@ struct drm_simple_display_pipe_funcs { ... void (*enable)(struct drm_simple_display_pipe *P ,struct drm_crtc_state *CS + ,struct drm_plane_state *plane_state ); ... }; v2: Pimp the commit message (David) Cc: Marek Vasut Cc: Eric Anholt Cc: David Lechner Cc: "Noralf Trønnes" Cc: Linus Walleij Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322202738.25817-1-ville.syrjala@linux.intel.com Reviewed-by: Noralf Trønnes --- drivers/gpu/drm/drm_simple_kms_helper.c | 4 +++- drivers/gpu/drm/mxsfb/mxsfb_drv.c | 3 ++- drivers/gpu/drm/pl111/pl111_display.c | 3 ++- drivers/gpu/drm/tinydrm/ili9225.c | 3 ++- drivers/gpu/drm/tinydrm/mi0283qt.c | 3 ++- drivers/gpu/drm/tinydrm/repaper.c | 3 ++- drivers/gpu/drm/tinydrm/st7586.c | 3 ++- drivers/gpu/drm/tinydrm/st7735r.c | 3 ++- drivers/gpu/drm/tve200/tve200_display.c | 3 ++- include/drm/drm_simple_kms_helper.h | 3 ++- 10 files changed, 21 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/drm_simple_kms_helper.c b/drivers/gpu/drm/drm_simple_kms_helper.c index 987a353..7a00455 100644 --- a/drivers/gpu/drm/drm_simple_kms_helper.c +++ b/drivers/gpu/drm/drm_simple_kms_helper.c @@ -64,13 +64,15 @@ static int drm_simple_kms_crtc_check(struct drm_crtc *crtc, static void drm_simple_kms_crtc_enable(struct drm_crtc *crtc, struct drm_crtc_state *old_state) { + struct drm_plane *plane; struct drm_simple_display_pipe *pipe; pipe = container_of(crtc, struct drm_simple_display_pipe, crtc); if (!pipe->funcs || !pipe->funcs->enable) return; - pipe->funcs->enable(pipe, crtc->state); + plane = &pipe->plane; + pipe->funcs->enable(pipe, crtc->state, plane->state); } static void drm_simple_kms_crtc_disable(struct drm_crtc *crtc, diff --git a/drivers/gpu/drm/mxsfb/mxsfb_drv.c b/drivers/gpu/drm/mxsfb/mxsfb_drv.c index 5cae8db..b9c7507 100644 --- a/drivers/gpu/drm/mxsfb/mxsfb_drv.c +++ b/drivers/gpu/drm/mxsfb/mxsfb_drv.c @@ -99,7 +99,8 @@ static const struct drm_mode_config_funcs mxsfb_mode_config_funcs = { }; static void mxsfb_pipe_enable(struct drm_simple_display_pipe *pipe, - struct drm_crtc_state *crtc_state) + struct drm_crtc_state *crtc_state, + struct drm_plane_state *plane_state) { struct mxsfb_drm_private *mxsfb = drm_pipe_to_mxsfb_drm_private(pipe); diff --git a/drivers/gpu/drm/pl111/pl111_display.c b/drivers/gpu/drm/pl111/pl111_display.c index 3106464..1fee578 100644 --- a/drivers/gpu/drm/pl111/pl111_display.c +++ b/drivers/gpu/drm/pl111/pl111_display.c @@ -120,7 +120,8 @@ static int pl111_display_check(struct drm_simple_display_pipe *pipe, } static void pl111_display_enable(struct drm_simple_display_pipe *pipe, - struct drm_crtc_state *cstate) + struct drm_crtc_state *cstate, + struct drm_plane_state *plane_state) { struct drm_crtc *crtc = &pipe->crtc; struct drm_plane *plane = &pipe->plane; diff --git a/drivers/gpu/drm/tinydrm/ili9225.c b/drivers/gpu/drm/tinydrm/ili9225.c index a075950..089d227 100644 --- a/drivers/gpu/drm/tinydrm/ili9225.c +++ b/drivers/gpu/drm/tinydrm/ili9225.c @@ -176,7 +176,8 @@ static const struct drm_framebuffer_funcs ili9225_fb_funcs = { }; static void ili9225_pipe_enable(struct drm_simple_display_pipe *pipe, - struct drm_crtc_state *crtc_state) + struct drm_crtc_state *crtc_state, + struct drm_plane_state *plane_state) { struct tinydrm_device *tdev = pipe_to_tinydrm(pipe); struct mipi_dbi *mipi = mipi_dbi_from_tinydrm(tdev); diff --git a/drivers/gpu/drm/tinydrm/mi0283qt.c b/drivers/gpu/drm/tinydrm/mi0283qt.c index d8ed6e6..82ad9b6 100644 --- a/drivers/gpu/drm/tinydrm/mi0283qt.c +++ b/drivers/gpu/drm/tinydrm/mi0283qt.c @@ -49,7 +49,8 @@ #define ILI9341_MADCTL_MY BIT(7) static void mi0283qt_enable(struct drm_simple_display_pipe *pipe, - struct drm_crtc_state *crtc_state) + struct drm_crtc_state *crtc_state, + struct drm_plane_state *plane_state) { struct tinydrm_device *tdev = pipe_to_tinydrm(pipe); struct mipi_dbi *mipi = mipi_dbi_from_tinydrm(tdev); diff --git a/drivers/gpu/drm/tinydrm/repaper.c b/drivers/gpu/drm/tinydrm/repaper.c index 7574063..33b4a71 100644 --- a/drivers/gpu/drm/tinydrm/repaper.c +++ b/drivers/gpu/drm/tinydrm/repaper.c @@ -659,7 +659,8 @@ static void power_off(struct repaper_epd *epd) } static void repaper_pipe_enable(struct drm_simple_display_pipe *pipe, - struct drm_crtc_state *crtc_state) + struct drm_crtc_state *crtc_state, + struct drm_plane_state *plane_state) { struct tinydrm_device *tdev = pipe_to_tinydrm(pipe); struct repaper_epd *epd = epd_from_tinydrm(tdev); diff --git a/drivers/gpu/drm/tinydrm/st7586.c b/drivers/gpu/drm/tinydrm/st7586.c index a6396ef..bb08b29 100644 --- a/drivers/gpu/drm/tinydrm/st7586.c +++ b/drivers/gpu/drm/tinydrm/st7586.c @@ -175,7 +175,8 @@ static const struct drm_framebuffer_funcs st7586_fb_funcs = { }; static void st7586_pipe_enable(struct drm_simple_display_pipe *pipe, - struct drm_crtc_state *crtc_state) + struct drm_crtc_state *crtc_state, + struct drm_plane_state *plane_state) { struct tinydrm_device *tdev = pipe_to_tinydrm(pipe); struct mipi_dbi *mipi = mipi_dbi_from_tinydrm(tdev); diff --git a/drivers/gpu/drm/tinydrm/st7735r.c b/drivers/gpu/drm/tinydrm/st7735r.c index 67d197e..19b28f8 100644 --- a/drivers/gpu/drm/tinydrm/st7735r.c +++ b/drivers/gpu/drm/tinydrm/st7735r.c @@ -37,7 +37,8 @@ #define ST7735R_MV BIT(5) static void jd_t18003_t01_pipe_enable(struct drm_simple_display_pipe *pipe, - struct drm_crtc_state *crtc_state) + struct drm_crtc_state *crtc_state, + struct drm_plane_state *plane_state) { struct tinydrm_device *tdev = pipe_to_tinydrm(pipe); struct mipi_dbi *mipi = mipi_dbi_from_tinydrm(tdev); diff --git a/drivers/gpu/drm/tve200/tve200_display.c b/drivers/gpu/drm/tve200/tve200_display.c index db397fc..108f3b2 100644 --- a/drivers/gpu/drm/tve200/tve200_display.c +++ b/drivers/gpu/drm/tve200/tve200_display.c @@ -120,7 +120,8 @@ static int tve200_display_check(struct drm_simple_display_pipe *pipe, } static void tve200_display_enable(struct drm_simple_display_pipe *pipe, - struct drm_crtc_state *cstate) + struct drm_crtc_state *cstate, + struct drm_plane_state *plane_state) { struct drm_crtc *crtc = &pipe->crtc; struct drm_plane *plane = &pipe->plane; diff --git a/include/drm/drm_simple_kms_helper.h b/include/drm/drm_simple_kms_helper.h index 1b4e352..b027937 100644 --- a/include/drm/drm_simple_kms_helper.h +++ b/include/drm/drm_simple_kms_helper.h @@ -64,7 +64,8 @@ struct drm_simple_display_pipe_funcs { * This hook is optional. */ void (*enable)(struct drm_simple_display_pipe *pipe, - struct drm_crtc_state *crtc_state); + struct drm_crtc_state *crtc_state, + struct drm_plane_state *plane_state); /** * @disable: * -- cgit v1.1 From e85d30060eddccfcfbf7fdbd61a23cfbda05cc59 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Fri, 23 Mar 2018 17:35:09 +0200 Subject: drm/tinydrm: Make fb_dirty into a lower level hook MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit mipi_dbi_enable_flush() wants to call the fb->dirty() hook from the bowels of the .atomic_enable() hook. That prevents us from taking the plane mutex in fb->dirty() unless we also plumb down the acquire context. Instead it seems simpler to split the fb->dirty() into a tinydrm specific lower level hook that can be called from mipi_dbi_enable_flush() and from a generic higher level tinydrm_fb_dirty() helper. As we don't have a tinydrm specific vfuncs table we'll just stick it into tinydrm_device directly for now. v2: Deal with the fb->dirty() in tinydrm_display_pipe_update() as well (Noralf) Cc: "Noralf Trønnes" Cc: David Lechner Signed-off-by: Ville Syrjälä Reviewed-by: Noralf Trønnes Link: https://patchwork.freedesktop.org/patch/msgid/20180323153509.15287-1-ville.syrjala@linux.intel.com Reviewed-by: Noralf Trønnes Tested-by: Noralf Trønnes --- drivers/gpu/drm/tinydrm/core/tinydrm-helpers.c | 30 ++++++++++++++++++++++++++ drivers/gpu/drm/tinydrm/core/tinydrm-pipe.c | 5 ++--- drivers/gpu/drm/tinydrm/ili9225.c | 23 ++++++-------------- drivers/gpu/drm/tinydrm/mi0283qt.c | 2 +- drivers/gpu/drm/tinydrm/mipi-dbi.c | 30 ++++++++++---------------- drivers/gpu/drm/tinydrm/repaper.c | 28 ++++++++---------------- drivers/gpu/drm/tinydrm/st7586.c | 23 ++++++-------------- drivers/gpu/drm/tinydrm/st7735r.c | 2 +- include/drm/tinydrm/mipi-dbi.h | 4 +++- include/drm/tinydrm/tinydrm-helpers.h | 5 +++++ include/drm/tinydrm/tinydrm.h | 4 ++++ 11 files changed, 78 insertions(+), 78 deletions(-) diff --git a/drivers/gpu/drm/tinydrm/core/tinydrm-helpers.c b/drivers/gpu/drm/tinydrm/core/tinydrm-helpers.c index d1c3ce9..dcd3901 100644 --- a/drivers/gpu/drm/tinydrm/core/tinydrm-helpers.c +++ b/drivers/gpu/drm/tinydrm/core/tinydrm-helpers.c @@ -78,6 +78,36 @@ bool tinydrm_merge_clips(struct drm_clip_rect *dst, } EXPORT_SYMBOL(tinydrm_merge_clips); +int tinydrm_fb_dirty(struct drm_framebuffer *fb, + struct drm_file *file_priv, + unsigned int flags, unsigned int color, + struct drm_clip_rect *clips, + unsigned int num_clips) +{ + struct tinydrm_device *tdev = fb->dev->dev_private; + struct drm_plane *plane = &tdev->pipe.plane; + int ret = 0; + + drm_modeset_lock(&plane->mutex, NULL); + + /* fbdev can flush even when we're not interested */ + if (plane->state->fb == fb) { + mutex_lock(&tdev->dirty_lock); + ret = tdev->fb_dirty(fb, file_priv, flags, + color, clips, num_clips); + mutex_unlock(&tdev->dirty_lock); + } + + drm_modeset_unlock(&plane->mutex); + + if (ret) + dev_err_once(fb->dev->dev, + "Failed to update display %d\n", ret); + + return ret; +} +EXPORT_SYMBOL(tinydrm_fb_dirty); + /** * tinydrm_memcpy - Copy clip buffer * @dst: Destination buffer diff --git a/drivers/gpu/drm/tinydrm/core/tinydrm-pipe.c b/drivers/gpu/drm/tinydrm/core/tinydrm-pipe.c index 11ae950..e68b528 100644 --- a/drivers/gpu/drm/tinydrm/core/tinydrm-pipe.c +++ b/drivers/gpu/drm/tinydrm/core/tinydrm-pipe.c @@ -125,9 +125,8 @@ void tinydrm_display_pipe_update(struct drm_simple_display_pipe *pipe, struct drm_crtc *crtc = &tdev->pipe.crtc; if (fb && (fb != old_state->fb)) { - pipe->plane.fb = fb; - if (fb->funcs->dirty) - fb->funcs->dirty(fb, NULL, 0, 0, NULL, 0); + if (tdev->fb_dirty) + tdev->fb_dirty(fb, NULL, 0, 0, NULL, 0); } if (crtc->state->event) { diff --git a/drivers/gpu/drm/tinydrm/ili9225.c b/drivers/gpu/drm/tinydrm/ili9225.c index 089d227..0874e87 100644 --- a/drivers/gpu/drm/tinydrm/ili9225.c +++ b/drivers/gpu/drm/tinydrm/ili9225.c @@ -88,14 +88,8 @@ static int ili9225_fb_dirty(struct drm_framebuffer *fb, bool full; void *tr; - mutex_lock(&tdev->dirty_lock); - if (!mipi->enabled) - goto out_unlock; - - /* fbdev can flush even when we're not interested */ - if (tdev->pipe.plane.fb != fb) - goto out_unlock; + return 0; full = tinydrm_merge_clips(&clip, clips, num_clips, flags, fb->width, fb->height); @@ -108,7 +102,7 @@ static int ili9225_fb_dirty(struct drm_framebuffer *fb, tr = mipi->tx_buf; ret = mipi_dbi_buf_copy(mipi->tx_buf, fb, &clip, swap); if (ret) - goto out_unlock; + return ret; } else { tr = cma_obj->vaddr; } @@ -159,20 +153,13 @@ static int ili9225_fb_dirty(struct drm_framebuffer *fb, ret = mipi_dbi_command_buf(mipi, ILI9225_WRITE_DATA_TO_GRAM, tr, (clip.x2 - clip.x1) * (clip.y2 - clip.y1) * 2); -out_unlock: - mutex_unlock(&tdev->dirty_lock); - - if (ret) - dev_err_once(fb->dev->dev, "Failed to update display %d\n", - ret); - return ret; } static const struct drm_framebuffer_funcs ili9225_fb_funcs = { .destroy = drm_gem_fb_destroy, .create_handle = drm_gem_fb_create_handle, - .dirty = ili9225_fb_dirty, + .dirty = tinydrm_fb_dirty, }; static void ili9225_pipe_enable(struct drm_simple_display_pipe *pipe, @@ -269,7 +256,7 @@ static void ili9225_pipe_enable(struct drm_simple_display_pipe *pipe, ili9225_command(mipi, ILI9225_DISPLAY_CONTROL_1, 0x1017); - mipi_dbi_enable_flush(mipi); + mipi_dbi_enable_flush(mipi, crtc_state, plane_state); } static void ili9225_pipe_disable(struct drm_simple_display_pipe *pipe) @@ -342,6 +329,8 @@ static int ili9225_init(struct device *dev, struct mipi_dbi *mipi, if (ret) return ret; + tdev->fb_dirty = ili9225_fb_dirty; + ret = tinydrm_display_pipe_init(tdev, pipe_funcs, DRM_MODE_CONNECTOR_VIRTUAL, ili9225_formats, diff --git a/drivers/gpu/drm/tinydrm/mi0283qt.c b/drivers/gpu/drm/tinydrm/mi0283qt.c index 82ad9b6..4e6d2ee 100644 --- a/drivers/gpu/drm/tinydrm/mi0283qt.c +++ b/drivers/gpu/drm/tinydrm/mi0283qt.c @@ -127,7 +127,7 @@ static void mi0283qt_enable(struct drm_simple_display_pipe *pipe, msleep(100); out_enable: - mipi_dbi_enable_flush(mipi); + mipi_dbi_enable_flush(mipi, crtc_state, plane_state); } static const struct drm_simple_display_pipe_funcs mi0283qt_pipe_funcs = { diff --git a/drivers/gpu/drm/tinydrm/mipi-dbi.c b/drivers/gpu/drm/tinydrm/mipi-dbi.c index 9e90381..4d1fb31 100644 --- a/drivers/gpu/drm/tinydrm/mipi-dbi.c +++ b/drivers/gpu/drm/tinydrm/mipi-dbi.c @@ -219,14 +219,8 @@ static int mipi_dbi_fb_dirty(struct drm_framebuffer *fb, bool full; void *tr; - mutex_lock(&tdev->dirty_lock); - if (!mipi->enabled) - goto out_unlock; - - /* fbdev can flush even when we're not interested */ - if (tdev->pipe.plane.fb != fb) - goto out_unlock; + return 0; full = tinydrm_merge_clips(&clip, clips, num_clips, flags, fb->width, fb->height); @@ -239,7 +233,7 @@ static int mipi_dbi_fb_dirty(struct drm_framebuffer *fb, tr = mipi->tx_buf; ret = mipi_dbi_buf_copy(mipi->tx_buf, fb, &clip, swap); if (ret) - goto out_unlock; + return ret; } else { tr = cma_obj->vaddr; } @@ -254,20 +248,13 @@ static int mipi_dbi_fb_dirty(struct drm_framebuffer *fb, ret = mipi_dbi_command_buf(mipi, MIPI_DCS_WRITE_MEMORY_START, tr, (clip.x2 - clip.x1) * (clip.y2 - clip.y1) * 2); -out_unlock: - mutex_unlock(&tdev->dirty_lock); - - if (ret) - dev_err_once(fb->dev->dev, "Failed to update display %d\n", - ret); - return ret; } static const struct drm_framebuffer_funcs mipi_dbi_fb_funcs = { .destroy = drm_gem_fb_destroy, .create_handle = drm_gem_fb_create_handle, - .dirty = mipi_dbi_fb_dirty, + .dirty = tinydrm_fb_dirty, }; /** @@ -278,13 +265,16 @@ static const struct drm_framebuffer_funcs mipi_dbi_fb_funcs = { * enables the backlight. Drivers can use this in their * &drm_simple_display_pipe_funcs->enable callback. */ -void mipi_dbi_enable_flush(struct mipi_dbi *mipi) +void mipi_dbi_enable_flush(struct mipi_dbi *mipi, + struct drm_crtc_state *crtc_state, + struct drm_plane_state *plane_state) { - struct drm_framebuffer *fb = mipi->tinydrm.pipe.plane.fb; + struct tinydrm_device *tdev = &mipi->tinydrm; + struct drm_framebuffer *fb = plane_state->fb; mipi->enabled = true; if (fb) - fb->funcs->dirty(fb, NULL, 0, 0, NULL, 0); + tdev->fb_dirty(fb, NULL, 0, 0, NULL, 0); backlight_enable(mipi->backlight); } @@ -381,6 +371,8 @@ int mipi_dbi_init(struct device *dev, struct mipi_dbi *mipi, if (ret) return ret; + tdev->fb_dirty = mipi_dbi_fb_dirty; + /* TODO: Maybe add DRM_MODE_CONNECTOR_SPI */ ret = tinydrm_display_pipe_init(tdev, pipe_funcs, DRM_MODE_CONNECTOR_VIRTUAL, diff --git a/drivers/gpu/drm/tinydrm/repaper.c b/drivers/gpu/drm/tinydrm/repaper.c index 33b4a71..bb6f80a 100644 --- a/drivers/gpu/drm/tinydrm/repaper.c +++ b/drivers/gpu/drm/tinydrm/repaper.c @@ -540,14 +540,8 @@ static int repaper_fb_dirty(struct drm_framebuffer *fb, clip.y1 = 0; clip.y2 = fb->height; - mutex_lock(&tdev->dirty_lock); - if (!epd->enabled) - goto out_unlock; - - /* fbdev can flush even when we're not interested */ - if (tdev->pipe.plane.fb != fb) - goto out_unlock; + return 0; repaper_get_temperature(epd); @@ -555,16 +549,14 @@ static int repaper_fb_dirty(struct drm_framebuffer *fb, epd->factored_stage_time); buf = kmalloc(fb->width * fb->height, GFP_KERNEL); - if (!buf) { - ret = -ENOMEM; - goto out_unlock; - } + if (!buf) + return -ENOMEM; if (import_attach) { ret = dma_buf_begin_cpu_access(import_attach->dmabuf, DMA_FROM_DEVICE); if (ret) - goto out_unlock; + goto out_free; } tinydrm_xrgb8888_to_gray8(buf, cma_obj->vaddr, fb, &clip); @@ -573,7 +565,7 @@ static int repaper_fb_dirty(struct drm_framebuffer *fb, ret = dma_buf_end_cpu_access(import_attach->dmabuf, DMA_FROM_DEVICE); if (ret) - goto out_unlock; + goto out_free; } repaper_gray8_to_mono_reversed(buf, fb->width, fb->height); @@ -625,11 +617,7 @@ static int repaper_fb_dirty(struct drm_framebuffer *fb, } } -out_unlock: - mutex_unlock(&tdev->dirty_lock); - - if (ret) - DRM_DEV_ERROR(fb->dev->dev, "Failed to update display (%d)\n", ret); +out_free: kfree(buf); return ret; @@ -638,7 +626,7 @@ out_unlock: static const struct drm_framebuffer_funcs repaper_fb_funcs = { .destroy = drm_gem_fb_destroy, .create_handle = drm_gem_fb_create_handle, - .dirty = repaper_fb_dirty, + .dirty = tinydrm_fb_dirty, }; static void power_off(struct repaper_epd *epd) @@ -1070,6 +1058,8 @@ static int repaper_probe(struct spi_device *spi) if (ret) return ret; + tdev->fb_dirty = repaper_fb_dirty; + ret = tinydrm_display_pipe_init(tdev, &repaper_pipe_funcs, DRM_MODE_CONNECTOR_VIRTUAL, repaper_formats, diff --git a/drivers/gpu/drm/tinydrm/st7586.c b/drivers/gpu/drm/tinydrm/st7586.c index bb08b29..22644b8 100644 --- a/drivers/gpu/drm/tinydrm/st7586.c +++ b/drivers/gpu/drm/tinydrm/st7586.c @@ -120,14 +120,8 @@ static int st7586_fb_dirty(struct drm_framebuffer *fb, int start, end; int ret = 0; - mutex_lock(&tdev->dirty_lock); - if (!mipi->enabled) - goto out_unlock; - - /* fbdev can flush even when we're not interested */ - if (tdev->pipe.plane.fb != fb) - goto out_unlock; + return 0; tinydrm_merge_clips(&clip, clips, num_clips, flags, fb->width, fb->height); @@ -141,7 +135,7 @@ static int st7586_fb_dirty(struct drm_framebuffer *fb, ret = st7586_buf_copy(mipi->tx_buf, fb, &clip); if (ret) - goto out_unlock; + return ret; /* Pixels are packed 3 per byte */ start = clip.x1 / 3; @@ -158,20 +152,13 @@ static int st7586_fb_dirty(struct drm_framebuffer *fb, (u8 *)mipi->tx_buf, (end - start) * (clip.y2 - clip.y1)); -out_unlock: - mutex_unlock(&tdev->dirty_lock); - - if (ret) - dev_err_once(fb->dev->dev, "Failed to update display %d\n", - ret); - return ret; } static const struct drm_framebuffer_funcs st7586_fb_funcs = { .destroy = drm_gem_fb_destroy, .create_handle = drm_gem_fb_create_handle, - .dirty = st7586_fb_dirty, + .dirty = tinydrm_fb_dirty, }; static void st7586_pipe_enable(struct drm_simple_display_pipe *pipe, @@ -238,7 +225,7 @@ static void st7586_pipe_enable(struct drm_simple_display_pipe *pipe, mipi_dbi_command(mipi, MIPI_DCS_SET_DISPLAY_ON); - mipi_dbi_enable_flush(mipi); + mipi_dbi_enable_flush(mipi, crtc_state, plane_state); } static void st7586_pipe_disable(struct drm_simple_display_pipe *pipe) @@ -278,6 +265,8 @@ static int st7586_init(struct device *dev, struct mipi_dbi *mipi, if (ret) return ret; + tdev->fb_dirty = st7586_fb_dirty; + ret = tinydrm_display_pipe_init(tdev, pipe_funcs, DRM_MODE_CONNECTOR_VIRTUAL, st7586_formats, diff --git a/drivers/gpu/drm/tinydrm/st7735r.c b/drivers/gpu/drm/tinydrm/st7735r.c index 19b28f8..189a078 100644 --- a/drivers/gpu/drm/tinydrm/st7735r.c +++ b/drivers/gpu/drm/tinydrm/st7735r.c @@ -99,7 +99,7 @@ static void jd_t18003_t01_pipe_enable(struct drm_simple_display_pipe *pipe, msleep(20); - mipi_dbi_enable_flush(mipi); + mipi_dbi_enable_flush(mipi, crtc_state, plane_state); } static const struct drm_simple_display_pipe_funcs jd_t18003_t01_pipe_funcs = { diff --git a/include/drm/tinydrm/mipi-dbi.h b/include/drm/tinydrm/mipi-dbi.h index 44e824a..b8ba588 100644 --- a/include/drm/tinydrm/mipi-dbi.h +++ b/include/drm/tinydrm/mipi-dbi.h @@ -67,7 +67,9 @@ int mipi_dbi_init(struct device *dev, struct mipi_dbi *mipi, const struct drm_simple_display_pipe_funcs *pipe_funcs, struct drm_driver *driver, const struct drm_display_mode *mode, unsigned int rotation); -void mipi_dbi_enable_flush(struct mipi_dbi *mipi); +void mipi_dbi_enable_flush(struct mipi_dbi *mipi, + struct drm_crtc_state *crtc_state, + struct drm_plane_state *plan_state); void mipi_dbi_pipe_disable(struct drm_simple_display_pipe *pipe); void mipi_dbi_hw_reset(struct mipi_dbi *mipi); bool mipi_dbi_display_is_on(struct mipi_dbi *mipi); diff --git a/include/drm/tinydrm/tinydrm-helpers.h b/include/drm/tinydrm/tinydrm-helpers.h index 0a4ddbc..5b96f0b 100644 --- a/include/drm/tinydrm/tinydrm-helpers.h +++ b/include/drm/tinydrm/tinydrm-helpers.h @@ -36,6 +36,11 @@ static inline bool tinydrm_machine_little_endian(void) bool tinydrm_merge_clips(struct drm_clip_rect *dst, struct drm_clip_rect *src, unsigned int num_clips, unsigned int flags, u32 max_width, u32 max_height); +int tinydrm_fb_dirty(struct drm_framebuffer *fb, + struct drm_file *file_priv, + unsigned int flags, unsigned int color, + struct drm_clip_rect *clips, + unsigned int num_clips); void tinydrm_memcpy(void *dst, void *vaddr, struct drm_framebuffer *fb, struct drm_clip_rect *clip); void tinydrm_swab16(u16 *dst, void *vaddr, struct drm_framebuffer *fb, diff --git a/include/drm/tinydrm/tinydrm.h b/include/drm/tinydrm/tinydrm.h index 77a93ec..6e2b960 100644 --- a/include/drm/tinydrm/tinydrm.h +++ b/include/drm/tinydrm/tinydrm.h @@ -26,6 +26,10 @@ struct tinydrm_device { struct drm_simple_display_pipe pipe; struct mutex dirty_lock; const struct drm_framebuffer_funcs *fb_funcs; + int (*fb_dirty)(struct drm_framebuffer *framebuffer, + struct drm_file *file_priv, unsigned flags, + unsigned color, struct drm_clip_rect *clips, + unsigned num_clips); }; static inline struct tinydrm_device * -- cgit v1.1 From d775a7b1840ddc96e7f25af20989ff43f2809436 Mon Sep 17 00:00:00 2001 From: Paulo Zanoni Date: Tue, 9 Jan 2018 21:28:35 -0200 Subject: drm/i915/gen11: add support for reading the timestamp frequency The only thing that differs here is that the crystal clock freq now has four possible values. This patch gets rid of the "Unknown gen, unable to compute..." message at boot for gen11. Reviewed-by: Lionel Landwerlin Signed-off-by: Paulo Zanoni Link: https://patchwork.freedesktop.org/patch/msgid/20180109232835.11478-18-paulo.r.zanoni@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 6 +++ drivers/gpu/drm/i915/intel_device_info.c | 71 +++++++++++++++++++++++++------- 2 files changed, 61 insertions(+), 16 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index b0c55f9..5a53d0e 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -861,6 +861,12 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK (1 << GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT) #define GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ 0 #define GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ 1 +#define GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT 3 +#define GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK (0x7 << GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT) +#define GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ 0 +#define GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ 1 +#define GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_38_4_MHZ 2 +#define GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_25_MHZ 3 #define GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_SHIFT 1 #define GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_MASK (0x3 << GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_SHIFT) diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c index 0d1509e..a32ba72 100644 --- a/drivers/gpu/drm/i915/intel_device_info.c +++ b/drivers/gpu/drm/i915/intel_device_info.c @@ -596,6 +596,52 @@ static u32 read_reference_ts_freq(struct drm_i915_private *dev_priv) return base_freq + frac_freq; } +static u32 gen10_get_crystal_clock_freq(struct drm_i915_private *dev_priv, + u32 rpm_config_reg) +{ + u32 f19_2_mhz = 19200; + u32 f24_mhz = 24000; + u32 crystal_clock = (rpm_config_reg & + GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK) >> + GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT; + + switch (crystal_clock) { + case GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ: + return f19_2_mhz; + case GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ: + return f24_mhz; + default: + MISSING_CASE(crystal_clock); + return 0; + } +} + +static u32 gen11_get_crystal_clock_freq(struct drm_i915_private *dev_priv, + u32 rpm_config_reg) +{ + u32 f19_2_mhz = 19200; + u32 f24_mhz = 24000; + u32 f25_mhz = 25000; + u32 f38_4_mhz = 38400; + u32 crystal_clock = (rpm_config_reg & + GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK) >> + GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT; + + switch (crystal_clock) { + case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ: + return f24_mhz; + case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ: + return f19_2_mhz; + case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_38_4_MHZ: + return f38_4_mhz; + case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_25_MHZ: + return f25_mhz; + default: + MISSING_CASE(crystal_clock); + return 0; + } +} + static u32 read_timestamp_frequency(struct drm_i915_private *dev_priv) { u32 f12_5_mhz = 12500; @@ -636,10 +682,9 @@ static u32 read_timestamp_frequency(struct drm_i915_private *dev_priv) } return freq; - } else if (INTEL_GEN(dev_priv) <= 10) { + } else if (INTEL_GEN(dev_priv) <= 11) { u32 ctc_reg = I915_READ(CTC_MODE); u32 freq = 0; - u32 rpm_config_reg = 0; /* First figure out the reference frequency. There are 2 ways * we can compute the frequency, either through the @@ -649,20 +694,14 @@ static u32 read_timestamp_frequency(struct drm_i915_private *dev_priv) if ((ctc_reg & CTC_SOURCE_PARAMETER_MASK) == CTC_SOURCE_DIVIDE_LOGIC) { freq = read_reference_ts_freq(dev_priv); } else { - u32 crystal_clock; - - rpm_config_reg = I915_READ(RPM_CONFIG0); - crystal_clock = (rpm_config_reg & - GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK) >> - GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT; - switch (crystal_clock) { - case GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ: - freq = f19_2_mhz; - break; - case GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ: - freq = f24_mhz; - break; - } + u32 rpm_config_reg = I915_READ(RPM_CONFIG0); + + if (INTEL_GEN(dev_priv) <= 10) + freq = gen10_get_crystal_clock_freq(dev_priv, + rpm_config_reg); + else + freq = gen11_get_crystal_clock_freq(dev_priv, + rpm_config_reg); /* Now figure out how the command stream's timestamp * register increments from this frequency (it might -- cgit v1.1 From c216e90686105e5b9fdbb22f6cfcc38334e432cc Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 27 Mar 2018 22:01:36 +0100 Subject: drm/i915/execlists: Reset ring registers on rebinding contexts Tvrtko uncovered a fun issue with recovering from a wedge device. In his tests, he wedged the driver by injecting an unrecoverable hang whilst a batch was spinning. As we reset the gpu in the middle of the spinner, when resumed it would continue on from the next instruction in the ring and write it's breadcrumb. However, on wedging we updated our bookkeeping to indicate that the GPU had completed executing and would restart from after the breadcrumb; so the emission of the stale breadcrumb from before the reset came as a bit of a surprise. A simple fix is to when rebinding the context into the GPU, we update the ring register state in the context image to match our bookkeeping. We already have to update the RING_START and RING_TAIL, so updating RING_HEAD as well is trivial. This works because whenever we unbind the context, we keep the bookkeeping in check; and on wedging we unbind all contexts. Testcase: igt/gem_eio Signed-off-by: Chris Wilson Cc: Tvrtko Ursulin Cc: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180327210136.16750-1-chris@chris-wilson.co.uk Tested-by: Tvrtko Ursulin Reviewed-by: Mika Kuoppala Reviewed-by: Tvrtko Ursulin --- drivers/gpu/drm/i915/intel_lrc.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index ba7f783..6546342 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -1272,6 +1272,7 @@ execlists_context_pin(struct intel_engine_cs *engine, ce->lrc_reg_state = vaddr + LRC_STATE_PN * PAGE_SIZE; ce->lrc_reg_state[CTX_RING_BUFFER_START+1] = i915_ggtt_offset(ce->ring->vma); + ce->lrc_reg_state[CTX_RING_HEAD+1] = ce->ring->head; ce->state->obj->pin_global++; i915_gem_context_get(ctx); -- cgit v1.1 From 4d82a174847b1b2d5e541ca7040398ac56efed7b Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 26 Mar 2018 19:48:18 +0000 Subject: drm/i915/guc: Add documentation for MMIO based communication As we are going to extend our use of MMIO based communication, try to explain its mechanics and update corresponding definitions. v2: fix checkpatch MACRO_ARG_REUSE Signed-off-by: Michal Wajdeczko Cc: Daniele Ceraolo Spurio Cc: Sagar Arun Kamble Cc: Kelvin Gardiner Reviewed-by: Michel Thierry #1 Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180326194829.58836-2-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc.c | 20 ++++---- drivers/gpu/drm/i915/intel_guc_ct.c | 2 +- drivers/gpu/drm/i915/intel_guc_fwif.h | 88 ++++++++++++++++++++++++++++------- 3 files changed, 82 insertions(+), 28 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index c5f64c7..78e68b1 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -329,6 +329,9 @@ int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len) GEM_BUG_ON(!len); GEM_BUG_ON(len > guc->send_regs.count); + /* We expect only action code */ + GEM_BUG_ON(*action & ~INTEL_GUC_MSG_CODE_MASK); + /* If CT is available, we expect to use MMIO only during init/fini */ GEM_BUG_ON(HAS_GUC_CT(dev_priv) && *action != INTEL_GUC_ACTION_REGISTER_COMMAND_TRANSPORT_BUFFER && @@ -350,18 +353,15 @@ int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len) */ ret = __intel_wait_for_register_fw(dev_priv, guc_send_reg(guc, 0), - INTEL_GUC_RECV_MASK, - INTEL_GUC_RECV_MASK, + INTEL_GUC_MSG_TYPE_MASK, + INTEL_GUC_MSG_TYPE_RESPONSE << + INTEL_GUC_MSG_TYPE_SHIFT, 10, 10, &status); - if (status != INTEL_GUC_STATUS_SUCCESS) { - /* - * Either the GuC explicitly returned an error (which - * we convert to -EIO here) or no response at all was - * received within the timeout limit (-ETIMEDOUT) - */ - if (ret != -ETIMEDOUT) - ret = -EIO; + /* If GuC explicitly returned an error, convert it to -EIO */ + if (!ret && !INTEL_GUC_MSG_IS_RESPONSE_SUCCESS(status)) + ret = -EIO; + if (ret) { DRM_DEBUG_DRIVER("INTEL_GUC_SEND: Action 0x%X failed;" " ret=%d status=0x%08X response=0x%08X\n", action[0], ret, status, diff --git a/drivers/gpu/drm/i915/intel_guc_ct.c b/drivers/gpu/drm/i915/intel_guc_ct.c index a726283..1dafa7a 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/intel_guc_ct.c @@ -398,7 +398,7 @@ static int ctch_send(struct intel_guc *guc, err = wait_for_response(desc, fence, status); if (unlikely(err)) return err; - if (*status != INTEL_GUC_STATUS_SUCCESS) + if (!INTEL_GUC_MSG_IS_RESPONSE_SUCCESS(*status)) return -EIO; return 0; } diff --git a/drivers/gpu/drm/i915/intel_guc_fwif.h b/drivers/gpu/drm/i915/intel_guc_fwif.h index 72941bd..83143e8 100644 --- a/drivers/gpu/drm/i915/intel_guc_fwif.h +++ b/drivers/gpu/drm/i915/intel_guc_fwif.h @@ -560,7 +560,68 @@ struct guc_shared_ctx_data { struct guc_ctx_report preempt_ctx_report[GUC_MAX_ENGINES_NUM]; } __packed; -/* This Action will be programmed in C180 - SOFT_SCRATCH_O_REG */ +/** + * DOC: MMIO based communication + * + * The MMIO based communication between Host and GuC uses software scratch + * registers, where first register holds data treated as message header, + * and other registers are used to hold message payload. + * + * For Gen9+, GuC uses software scratch registers 0xC180-0xC1B8 + * + * +-----------+---------+---------+---------+ + * | MMIO[0] | MMIO[1] | ... | MMIO[n] | + * +-----------+---------+---------+---------+ + * | header | optional payload | + * +======+====+=========+=========+=========+ + * | 31:28|type| | | | + * +------+----+ | | | + * | 27:16|data| | | | + * +------+----+ | | | + * | 15:0|code| | | | + * +------+----+---------+---------+---------+ + * + * The message header consists of: + * + * - **type**, indicates message type + * - **code**, indicates message code, is specific for **type** + * - **data**, indicates message data, optional, depends on **code** + * + * The following message **types** are supported: + * + * - **REQUEST**, indicates Host-to-GuC request, requested GuC action code + * must be priovided in **code** field. Optional action specific parameters + * can be provided in remaining payload registers or **data** field. + * + * - **RESPONSE**, indicates GuC-to-Host response from earlier GuC request, + * action response status will be provided in **code** field. Optional + * response data can be returned in remaining payload registers or **data** + * field. + */ + +#define INTEL_GUC_MSG_TYPE_SHIFT 28 +#define INTEL_GUC_MSG_TYPE_MASK (0xF << INTEL_GUC_MSG_TYPE_SHIFT) +#define INTEL_GUC_MSG_DATA_SHIFT 16 +#define INTEL_GUC_MSG_DATA_MASK (0xFFF << INTEL_GUC_MSG_DATA_SHIFT) +#define INTEL_GUC_MSG_CODE_SHIFT 0 +#define INTEL_GUC_MSG_CODE_MASK (0xFFFF << INTEL_GUC_MSG_CODE_SHIFT) + +#define __INTEL_GUC_MSG_GET(T, m) \ + (((m) & INTEL_GUC_MSG_ ## T ## _MASK) >> INTEL_GUC_MSG_ ## T ## _SHIFT) +#define INTEL_GUC_MSG_TO_TYPE(m) __INTEL_GUC_MSG_GET(TYPE, m) +#define INTEL_GUC_MSG_TO_DATA(m) __INTEL_GUC_MSG_GET(DATA, m) +#define INTEL_GUC_MSG_TO_CODE(m) __INTEL_GUC_MSG_GET(CODE, m) + +enum intel_guc_msg_type { + INTEL_GUC_MSG_TYPE_REQUEST = 0x0, + INTEL_GUC_MSG_TYPE_RESPONSE = 0xF, +}; + +#define __INTEL_GUC_MSG_TYPE_IS(T, m) \ + (INTEL_GUC_MSG_TO_TYPE(m) == INTEL_GUC_MSG_TYPE_ ## T) +#define INTEL_GUC_MSG_IS_REQUEST(m) __INTEL_GUC_MSG_TYPE_IS(REQUEST, m) +#define INTEL_GUC_MSG_IS_RESPONSE(m) __INTEL_GUC_MSG_TYPE_IS(RESPONSE, m) + enum intel_guc_action { INTEL_GUC_ACTION_DEFAULT = 0x0, INTEL_GUC_ACTION_REQUEST_PREEMPTION = 0x2, @@ -597,24 +658,17 @@ enum intel_guc_report_status { #define GUC_LOG_CONTROL_VERBOSITY_MASK (0xF << GUC_LOG_CONTROL_VERBOSITY_SHIFT) #define GUC_LOG_CONTROL_DEFAULT_LOGGING (1 << 8) -/* - * The GuC sends its response to a command by overwriting the - * command in SS0. The response is distinguishable from a command - * by the fact that all the MASK bits are set. The remaining bits - * give more detail. - */ -#define INTEL_GUC_RECV_MASK ((u32)0xF0000000) -#define INTEL_GUC_RECV_IS_RESPONSE(x) ((u32)(x) >= INTEL_GUC_RECV_MASK) -#define INTEL_GUC_RECV_STATUS(x) (INTEL_GUC_RECV_MASK | (x)) - -/* GUC will return status back to SOFT_SCRATCH_O_REG */ -enum intel_guc_status { - INTEL_GUC_STATUS_SUCCESS = INTEL_GUC_RECV_STATUS(0x0), - INTEL_GUC_STATUS_ALLOCATE_DOORBELL_FAIL = INTEL_GUC_RECV_STATUS(0x10), - INTEL_GUC_STATUS_DEALLOCATE_DOORBELL_FAIL = INTEL_GUC_RECV_STATUS(0x20), - INTEL_GUC_STATUS_GENERIC_FAIL = INTEL_GUC_RECV_STATUS(0x0000F000) +enum intel_guc_response_status { + INTEL_GUC_RESPONSE_STATUS_SUCCESS = 0x0, + INTEL_GUC_RESPONSE_STATUS_GENERIC_FAIL = 0xF000, }; +#define INTEL_GUC_MSG_IS_RESPONSE_SUCCESS(m) \ + (typecheck(u32, (m)) && \ + ((m) & (INTEL_GUC_MSG_TYPE_MASK | INTEL_GUC_MSG_CODE_MASK)) == \ + ((INTEL_GUC_MSG_TYPE_RESPONSE << INTEL_GUC_MSG_TYPE_SHIFT) | \ + (INTEL_GUC_RESPONSE_STATUS_SUCCESS << INTEL_GUC_MSG_CODE_SHIFT))) + /* This action will be programmed in C1BC - SOFT_SCRATCH_15_REG */ enum intel_guc_recv_message { INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED = BIT(1), -- cgit v1.1 From b839a869dfc9f01aab72c5dd26cb7a7f2e264201 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 26 Mar 2018 19:48:19 +0000 Subject: drm/i915/guc: Add support for data reporting in GuC responses GuC may return additional data in the response message. Format and meaning of this data is action specific. We will use this non-negative data as a new success return value. Currently used actions don't return data that way yet. v2: fix prohibited space after '~' (Michel) update commit message (Daniele) v3: rebase Signed-off-by: Michal Wajdeczko Cc: Oscar Mateo Cc: Michel Thierry Cc: Daniele Ceraolo Spurio Reviewed-by: Michel Thierry Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180326194829.58836-3-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc.c | 3 +++ drivers/gpu/drm/i915/intel_guc_ct.c | 14 ++++++++------ 2 files changed, 11 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index 78e68b1..1af32a0 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -366,6 +366,9 @@ int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len) " ret=%d status=0x%08X response=0x%08X\n", action[0], ret, status, I915_READ(SOFT_SCRATCH(15))); + } else { + /* Use data from the GuC response as our return value */ + ret = INTEL_GUC_MSG_TO_DATA(status); } intel_uncore_forcewake_put(dev_priv, guc->send_regs.fw_domains); diff --git a/drivers/gpu/drm/i915/intel_guc_ct.c b/drivers/gpu/drm/i915/intel_guc_ct.c index 1dafa7a..fa52259 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/intel_guc_ct.c @@ -400,7 +400,9 @@ static int ctch_send(struct intel_guc *guc, return err; if (!INTEL_GUC_MSG_IS_RESPONSE_SUCCESS(*status)) return -EIO; - return 0; + + /* Use data from the GuC status as our return value */ + return INTEL_GUC_MSG_TO_DATA(*status); } /* @@ -410,18 +412,18 @@ static int intel_guc_send_ct(struct intel_guc *guc, const u32 *action, u32 len) { struct intel_guc_ct_channel *ctch = &guc->ct.host_channel; u32 status = ~0; /* undefined */ - int err; + int ret; mutex_lock(&guc->send_mutex); - err = ctch_send(guc, ctch, action, len, &status); - if (unlikely(err)) { + ret = ctch_send(guc, ctch, action, len, &status); + if (unlikely(ret < 0)) { DRM_ERROR("CT: send action %#X failed; err=%d status=%#X\n", - action[0], err, status); + action[0], ret, status); } mutex_unlock(&guc->send_mutex); - return err; + return ret; } /** -- cgit v1.1 From e09af3a6a62dfe0a9c4bc931c81447a3969177fb Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 26 Mar 2018 19:48:20 +0000 Subject: drm/i915/guc: Prepare send() function to accept bigger response This is a preparation step for the upcoming patches. We already can return some small data decoded from the command status, but we will need more in the future. v2: add explicit response buf size v3: squash with helper patch Signed-off-by: Michal Wajdeczko Cc: Oscar Mateo Cc: Michel Thierry Cc: Daniele Ceraolo Spurio Reviewed-by: Michel Thierry Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180326194829.58836-4-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc.c | 6 ++++-- drivers/gpu/drm/i915/intel_guc.h | 18 ++++++++++++++---- drivers/gpu/drm/i915/intel_guc_ct.c | 7 ++++--- 3 files changed, 22 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index 1af32a0..ba5a962 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -310,7 +310,8 @@ void intel_guc_init_params(struct intel_guc *guc) intel_uncore_forcewake_put(dev_priv, FORCEWAKE_BLITTER); } -int intel_guc_send_nop(struct intel_guc *guc, const u32 *action, u32 len) +int intel_guc_send_nop(struct intel_guc *guc, const u32 *action, u32 len, + u32 *response_buf, u32 response_buf_size) { WARN(1, "Unexpected send: action=%#x\n", *action); return -ENODEV; @@ -319,7 +320,8 @@ int intel_guc_send_nop(struct intel_guc *guc, const u32 *action, u32 len) /* * This function implements the MMIO based host to GuC interface. */ -int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len) +int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len, + u32 *response_buf, u32 response_buf_size) { struct drm_i915_private *dev_priv = guc_to_i915(guc); u32 status; diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h index 13f3d1d..7ee0732 100644 --- a/drivers/gpu/drm/i915/intel_guc.h +++ b/drivers/gpu/drm/i915/intel_guc.h @@ -88,7 +88,8 @@ struct intel_guc { struct mutex send_mutex; /* GuC's FW specific send function */ - int (*send)(struct intel_guc *guc, const u32 *data, u32 len); + int (*send)(struct intel_guc *guc, const u32 *data, u32 len, + u32 *response_buf, u32 response_buf_size); /* GuC's FW specific notify function */ void (*notify)(struct intel_guc *guc); @@ -97,7 +98,14 @@ struct intel_guc { static inline int intel_guc_send(struct intel_guc *guc, const u32 *action, u32 len) { - return guc->send(guc, action, len); + return guc->send(guc, action, len, NULL, 0); +} + +static inline int +intel_guc_send_and_receive(struct intel_guc *guc, const u32 *action, u32 len, + u32 *response_buf, u32 response_buf_size) +{ + return guc->send(guc, action, len, response_buf, response_buf_size); } static inline void intel_guc_notify(struct intel_guc *guc) @@ -140,8 +148,10 @@ int intel_guc_init_wq(struct intel_guc *guc); void intel_guc_fini_wq(struct intel_guc *guc); int intel_guc_init(struct intel_guc *guc); void intel_guc_fini(struct intel_guc *guc); -int intel_guc_send_nop(struct intel_guc *guc, const u32 *action, u32 len); -int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len); +int intel_guc_send_nop(struct intel_guc *guc, const u32 *action, u32 len, + u32 *response_buf, u32 response_buf_size); +int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len, + u32 *response_buf, u32 response_buf_size); void intel_guc_to_host_event_handler(struct intel_guc *guc); int intel_guc_sample_forcewake(struct intel_guc *guc); int intel_guc_auth_huc(struct intel_guc *guc, u32 rsa_offset); diff --git a/drivers/gpu/drm/i915/intel_guc_ct.c b/drivers/gpu/drm/i915/intel_guc_ct.c index fa52259..a54bf58 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/intel_guc_ct.c @@ -88,7 +88,7 @@ static int guc_action_register_ct_buffer(struct intel_guc *guc, int err; /* Can't use generic send(), CT registration must go over MMIO */ - err = intel_guc_send_mmio(guc, action, ARRAY_SIZE(action)); + err = intel_guc_send_mmio(guc, action, ARRAY_SIZE(action), NULL, 0); if (err) DRM_ERROR("CT: register %s buffer failed; err=%d\n", guc_ct_buffer_type_to_str(type), err); @@ -107,7 +107,7 @@ static int guc_action_deregister_ct_buffer(struct intel_guc *guc, int err; /* Can't use generic send(), CT deregistration must go over MMIO */ - err = intel_guc_send_mmio(guc, action, ARRAY_SIZE(action)); + err = intel_guc_send_mmio(guc, action, ARRAY_SIZE(action), NULL, 0); if (err) DRM_ERROR("CT: deregister %s buffer failed; owner=%d err=%d\n", guc_ct_buffer_type_to_str(type), owner, err); @@ -408,7 +408,8 @@ static int ctch_send(struct intel_guc *guc, /* * Command Transport (CT) buffer based GuC send function. */ -static int intel_guc_send_ct(struct intel_guc *guc, const u32 *action, u32 len) +static int intel_guc_send_ct(struct intel_guc *guc, const u32 *action, u32 len, + u32 *response_buf, u32 response_buf_size) { struct intel_guc_ct_channel *ctch = &guc->ct.host_channel; u32 status = ~0; /* undefined */ -- cgit v1.1 From f6a70b59705109edb935674d36a5065fe2a5b4ad Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 26 Mar 2018 19:48:21 +0000 Subject: drm/i915/guc: Implement response handling in send_mmio() We're using data encoded in the status MMIO as return value from send function, but GuC may also write more data in remaining MMIO regs. Let's copy content of these registers to the buffer provided by caller. v2: new line (Michel) v3: updated commit message Signed-off-by: Michal Wajdeczko Cc: Daniele Ceraolo Spurio Cc: Oscar Mateo Reviewed-by: Michel Thierry Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180326194829.58836-5-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index ba5a962..b83a5ad 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -368,11 +368,20 @@ int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len, " ret=%d status=0x%08X response=0x%08X\n", action[0], ret, status, I915_READ(SOFT_SCRATCH(15))); - } else { - /* Use data from the GuC response as our return value */ - ret = INTEL_GUC_MSG_TO_DATA(status); + goto out; } + if (response_buf) { + int count = min(response_buf_size, guc->send_regs.count - 1); + + for (i = 0; i < count; i++) + response_buf[i] = I915_READ(guc_send_reg(guc, i + 1)); + } + + /* Use data from the GuC response as our return value */ + ret = INTEL_GUC_MSG_TO_DATA(status); + +out: intel_uncore_forcewake_put(dev_priv, guc->send_regs.fw_domains); mutex_unlock(&guc->send_mutex); -- cgit v1.1 From 769bfbf943235c8c54e189a0d12bf8d3195af0c5 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 26 Mar 2018 19:48:22 +0000 Subject: drm/i915/guc: Make event handler a virtual function On platforms with CTB based GuC communications, we will handle GuC events in a different way. Let's make event handler a virtual function to allow easy switch between those variants. Credits-to: Oscar Mateo Signed-off-by: Michal Wajdeczko Cc: Daniele Ceraolo Spurio Cc: Oscar Mateo Reviewed-by: Michel Thierry Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180326194829.58836-6-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc.c | 8 +++++++- drivers/gpu/drm/i915/intel_guc.h | 10 ++++++++++ drivers/gpu/drm/i915/intel_uc.c | 2 ++ 3 files changed, 19 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index b83a5ad..411c8e9 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -69,6 +69,7 @@ void intel_guc_init_early(struct intel_guc *guc) mutex_init(&guc->send_mutex); spin_lock_init(&guc->irq_lock); guc->send = intel_guc_send_nop; + guc->handler = intel_guc_to_host_event_handler_nop; guc->notify = gen8_guc_raise_irq; } @@ -317,6 +318,11 @@ int intel_guc_send_nop(struct intel_guc *guc, const u32 *action, u32 len, return -ENODEV; } +void intel_guc_to_host_event_handler_nop(struct intel_guc *guc) +{ + WARN(1, "Unexpected event: no suitable handler\n"); +} + /* * This function implements the MMIO based host to GuC interface. */ @@ -388,7 +394,7 @@ out: return ret; } -void intel_guc_to_host_event_handler(struct intel_guc *guc) +void intel_guc_to_host_event_handler_mmio(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); u32 msg, val; diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h index 7ee0732..6dc109a 100644 --- a/drivers/gpu/drm/i915/intel_guc.h +++ b/drivers/gpu/drm/i915/intel_guc.h @@ -91,6 +91,9 @@ struct intel_guc { int (*send)(struct intel_guc *guc, const u32 *data, u32 len, u32 *response_buf, u32 response_buf_size); + /* GuC's FW specific event handler function */ + void (*handler)(struct intel_guc *guc); + /* GuC's FW specific notify function */ void (*notify)(struct intel_guc *guc); }; @@ -113,6 +116,11 @@ static inline void intel_guc_notify(struct intel_guc *guc) guc->notify(guc); } +static inline void intel_guc_to_host_event_handler(struct intel_guc *guc) +{ + guc->handler(guc); +} + /* GuC addresses above GUC_GGTT_TOP also don't map through the GTT */ #define GUC_GGTT_TOP 0xFEE00000 @@ -153,6 +161,8 @@ int intel_guc_send_nop(struct intel_guc *guc, const u32 *action, u32 len, int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len, u32 *response_buf, u32 response_buf_size); void intel_guc_to_host_event_handler(struct intel_guc *guc); +void intel_guc_to_host_event_handler_nop(struct intel_guc *guc); +void intel_guc_to_host_event_handler_mmio(struct intel_guc *guc); int intel_guc_sample_forcewake(struct intel_guc *guc); int intel_guc_auth_huc(struct intel_guc *guc, u32 rsa_offset); int intel_guc_suspend(struct intel_guc *guc); diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 4aad844..081e424 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -233,6 +233,7 @@ static int guc_enable_communication(struct intel_guc *guc) return intel_guc_ct_enable(&guc->ct); guc->send = intel_guc_send_mmio; + guc->handler = intel_guc_to_host_event_handler_mmio; return 0; } @@ -246,6 +247,7 @@ static void guc_disable_communication(struct intel_guc *guc) gen9_disable_guc_interrupts(dev_priv); guc->send = intel_guc_send_nop; + guc->handler = intel_guc_to_host_event_handler_nop; } int intel_uc_init_misc(struct drm_i915_private *dev_priv) -- cgit v1.1 From 1d407096002becab2fd5b19253cee0de65aab668 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 26 Mar 2018 19:48:23 +0000 Subject: drm/i915/guc: Prepare to handle messages from CT RECV buffer GuC can respond to our commands not only by updating SEND buffer descriptor, but can also send a response message over RECV buffer. Guc can also send unsolicited request messages over RECV buffer. Let's start reading those messages and make placeholders for actual response/request handlers. v2: misc improvements (Michal) v3: change response detection (Michal) invalid status is protocol error (Michal) v4: rebase v5: fix checkpatch (Michel) don't use fields before check (Jani) add some documentation (Michal) Signed-off-by: Michal Wajdeczko Cc: Oscar Mateo Cc: Michel Thierry Cc: Daniele Ceraolo Spurio Reviewed-by: Michel Thierry # 4.5 Cc: Jani Nikula Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180326194829.58836-7-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc_ct.c | 184 +++++++++++++++++++++++++++++++++++- 1 file changed, 183 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/intel_guc_ct.c b/drivers/gpu/drm/i915/intel_guc_ct.c index a54bf58..14f55de 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/intel_guc_ct.c @@ -273,6 +273,24 @@ static u32 ctch_get_next_fence(struct intel_guc_ct_channel *ctch) return ++ctch->next_fence; } +/** + * DOC: CTB Host to GuC request + * + * Format of the CTB Host to GuC request message is as follows:: + * + * +------------+---------+---------+---------+---------+ + * | msg[0] | [1] | [2] | ... | [n-1] | + * +------------+---------+---------+---------+---------+ + * | MESSAGE | MESSAGE PAYLOAD | + * + HEADER +---------+---------+---------+---------+ + * | | 0 | 1 | ... | n | + * +============+=========+=========+=========+=========+ + * | len >= 1 | FENCE | request specific data | + * +------+-----+---------+---------+---------+---------+ + * + * ^-----------------len-------------------^ + */ + static int ctb_write(struct intel_guc_ct_buffer *ctb, const u32 *action, u32 len /* in dwords */, @@ -305,7 +323,8 @@ static int ctb_write(struct intel_guc_ct_buffer *ctb, if (unlikely(used + len + 1 >= size)) return -ENOSPC; - /* Write the message. The format is the following: + /* + * Write the message. The format is the following: * DW0: header (including action code) * DW1: fence * DW2+: action data @@ -427,6 +446,167 @@ static int intel_guc_send_ct(struct intel_guc *guc, const u32 *action, u32 len, return ret; } +static inline unsigned int ct_header_get_len(u32 header) +{ + return (header >> GUC_CT_MSG_LEN_SHIFT) & GUC_CT_MSG_LEN_MASK; +} + +static inline unsigned int ct_header_get_action(u32 header) +{ + return (header >> GUC_CT_MSG_ACTION_SHIFT) & GUC_CT_MSG_ACTION_MASK; +} + +static inline bool ct_header_is_response(u32 header) +{ + return ct_header_get_action(header) == INTEL_GUC_ACTION_DEFAULT; +} + +static int ctb_read(struct intel_guc_ct_buffer *ctb, u32 *data) +{ + struct guc_ct_buffer_desc *desc = ctb->desc; + u32 head = desc->head / 4; /* in dwords */ + u32 tail = desc->tail / 4; /* in dwords */ + u32 size = desc->size / 4; /* in dwords */ + u32 *cmds = ctb->cmds; + s32 available; /* in dwords */ + unsigned int len; + unsigned int i; + + GEM_BUG_ON(desc->size % 4); + GEM_BUG_ON(desc->head % 4); + GEM_BUG_ON(desc->tail % 4); + GEM_BUG_ON(tail >= size); + GEM_BUG_ON(head >= size); + + /* tail == head condition indicates empty */ + available = tail - head; + if (unlikely(available == 0)) + return -ENODATA; + + /* beware of buffer wrap case */ + if (unlikely(available < 0)) + available += size; + GEM_BUG_ON(available < 0); + + data[0] = cmds[head]; + head = (head + 1) % size; + + /* message len with header */ + len = ct_header_get_len(data[0]) + 1; + if (unlikely(len > (u32)available)) { + DRM_ERROR("CT: incomplete message %*phn %*phn %*phn\n", + 4, data, + 4 * (head + available - 1 > size ? + size - head : available - 1), &cmds[head], + 4 * (head + available - 1 > size ? + available - 1 - size + head : 0), &cmds[0]); + return -EPROTO; + } + + for (i = 1; i < len; i++) { + data[i] = cmds[head]; + head = (head + 1) % size; + } + + desc->head = head * 4; + return 0; +} + +/** + * DOC: CTB GuC to Host response + * + * Format of the CTB GuC to Host response message is as follows:: + * + * +------------+---------+---------+---------+---------+---------+ + * | msg[0] | [1] | [2] | [3] | ... | [n-1] | + * +------------+---------+---------+---------+---------+---------+ + * | MESSAGE | MESSAGE PAYLOAD | + * + HEADER +---------+---------+---------+---------+---------+ + * | | 0 | 1 | 2 | ... | n | + * +============+=========+=========+=========+=========+=========+ + * | len >= 2 | FENCE | STATUS | response specific data | + * +------+-----+---------+---------+---------+---------+---------+ + * + * ^-----------------------len-----------------------^ + */ + +static int ct_handle_response(struct intel_guc_ct *ct, const u32 *msg) +{ + u32 header = msg[0]; + u32 len = ct_header_get_len(header); + u32 msglen = len + 1; /* total message length including header */ + u32 fence; + u32 status; + + GEM_BUG_ON(!ct_header_is_response(header)); + + /* Response payload shall at least include fence and status */ + if (unlikely(len < 2)) { + DRM_ERROR("CT: corrupted response %*phn\n", 4 * msglen, msg); + return -EPROTO; + } + + fence = msg[1]; + status = msg[2]; + + /* Format of the status follows RESPONSE message */ + if (unlikely(!INTEL_GUC_MSG_IS_RESPONSE(status))) { + DRM_ERROR("CT: corrupted response %*phn\n", 4 * msglen, msg); + return -EPROTO; + } + + /* XXX */ + return 0; +} + +static int ct_handle_request(struct intel_guc_ct *ct, const u32 *msg) +{ + u32 header = msg[0]; + + GEM_BUG_ON(ct_header_is_response(header)); + + /* XXX */ + return 0; +} + +static void ct_process_host_channel(struct intel_guc_ct *ct) +{ + struct intel_guc_ct_channel *ctch = &ct->host_channel; + struct intel_guc_ct_buffer *ctb = &ctch->ctbs[CTB_RECV]; + u32 msg[GUC_CT_MSG_LEN_MASK + 1]; /* one extra dw for the header */ + int err = 0; + + if (!ctch_is_open(ctch)) + return; + + do { + err = ctb_read(ctb, msg); + if (err) + break; + + if (ct_header_is_response(msg[0])) + err = ct_handle_response(ct, msg); + else + err = ct_handle_request(ct, msg); + } while (!err); + + if (GEM_WARN_ON(err == -EPROTO)) { + DRM_ERROR("CT: corrupted message detected!\n"); + ctb->desc->is_in_error = 1; + } +} + +/* + * When we're communicating with the GuC over CT, GuC uses events + * to notify us about new messages being posted on the RECV buffer. + */ +static void intel_guc_to_host_event_handler_ct(struct intel_guc *guc) +{ + struct intel_guc_ct *ct = &guc->ct; + + ct_process_host_channel(ct); +} + /** * intel_guc_ct_enable - Enable buffer based command transport. * @ct: pointer to CT struct @@ -450,6 +630,7 @@ int intel_guc_ct_enable(struct intel_guc_ct *ct) /* Switch into cmd transport buffer based send() */ guc->send = intel_guc_send_ct; + guc->handler = intel_guc_to_host_event_handler_ct; DRM_INFO("CT: %s\n", enableddisabled(true)); return 0; } @@ -475,5 +656,6 @@ void intel_guc_ct_disable(struct intel_guc_ct *ct) /* Disable send */ guc->send = intel_guc_send_nop; + guc->handler = intel_guc_to_host_event_handler_nop; DRM_INFO("CT: %s\n", enableddisabled(false)); } -- cgit v1.1 From 24827cd0dd86a3158abc90ee64dcaf18aa843970 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 26 Mar 2018 19:48:24 +0000 Subject: drm/i915/guc: Use better name for helper wait function In next patch we will introduce another way of waiting for the response that will use RECV buffer. To avoid misleading names, rename old wait function to reflect the fact that it is based on descriptor update. v2: fix comment style (Michal) v3: use more specific name (Michel) Signed-off-by: Michal Wajdeczko Reviewed-by: Michel Thierry Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180326194829.58836-8-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc_ct.c | 25 +++++++++++++++++-------- 1 file changed, 17 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc_ct.c b/drivers/gpu/drm/i915/intel_guc_ct.c index 14f55de..2d805a2 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/intel_guc_ct.c @@ -351,16 +351,25 @@ static int ctb_write(struct intel_guc_ct_buffer *ctb, return 0; } -/* Wait for the response from the GuC. +/** + * wait_for_ctb_desc_update - Wait for the CT buffer descriptor update. + * @desc: buffer descriptor * @fence: response fence * @status: placeholder for status - * return: 0 response received (status is valid) - * -ETIMEDOUT no response within hardcoded timeout - * -EPROTO no response, ct buffer was in error + * + * Guc will update CT buffer descriptor with new fence and status + * after processing the command identified by the fence. Wait for + * specified fence and then read from the descriptor status of the + * command. + * + * Return: + * * 0 response received (status is valid) + * * -ETIMEDOUT no response within hardcoded timeout + * * -EPROTO no response, CT buffer is in error */ -static int wait_for_response(struct guc_ct_buffer_desc *desc, - u32 fence, - u32 *status) +static int wait_for_ctb_desc_update(struct guc_ct_buffer_desc *desc, + u32 fence, + u32 *status) { int err; @@ -414,7 +423,7 @@ static int ctch_send(struct intel_guc *guc, intel_guc_notify(guc); - err = wait_for_response(desc, fence, status); + err = wait_for_ctb_desc_update(desc, fence, status); if (unlikely(err)) return err; if (!INTEL_GUC_MSG_IS_RESPONSE_SUCCESS(*status)) -- cgit v1.1 From 9ef4c75e06006eb041c0bcc578548e3cdb8f9e03 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Tue, 27 Mar 2018 12:14:39 +0000 Subject: drm/i915/guc: Implement response handling in send_ct() Instead of returning small data in response status dword, GuC may append longer data as response message payload. If caller provides response buffer, we will copy received data and use number of received data dwords as new success return value. We will WARN if response from GuC does not match caller expectation. v2: fix timeout and checkpatch warnings (Michal) v3: fix checkpatch again (Michel) update wait function name (Michal) no need for spinlock_irqsave (MichalWi) no magic numbers (MichalWi) must check before use (Jani) add some more documentation (Michal) v4: update documentation (Michal) Signed-off-by: Michal Wajdeczko Cc: Oscar Mateo Cc: Michel Thierry Cc: Daniele Ceraolo Spurio Reviewed-by: Michel Thierry #2.5 Cc: Michal Winiarski Cc: Jani Nikula Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180327121439.70096-1-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc_ct.c | 142 ++++++++++++++++++++++++++++++---- drivers/gpu/drm/i915/intel_guc_ct.h | 6 ++ drivers/gpu/drm/i915/intel_guc_fwif.h | 52 +++++++++++++ 3 files changed, 186 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc_ct.c b/drivers/gpu/drm/i915/intel_guc_ct.c index 2d805a2..41b071c 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/intel_guc_ct.c @@ -24,6 +24,14 @@ #include "i915_drv.h" #include "intel_guc_ct.h" +struct ct_request { + struct list_head link; + u32 fence; + u32 status; + u32 response_len; + u32 *response_buf; +}; + enum { CTB_SEND = 0, CTB_RECV = 1 }; enum { CTB_OWNER_HOST = 0 }; @@ -36,6 +44,9 @@ void intel_guc_ct_init_early(struct intel_guc_ct *ct) { /* we're using static channel owners */ ct->host_channel.owner = CTB_OWNER_HOST; + + spin_lock_init(&ct->lock); + INIT_LIST_HEAD(&ct->pending_requests); } static inline struct intel_guc *ct_to_guc(struct intel_guc_ct *ct) @@ -294,7 +305,8 @@ static u32 ctch_get_next_fence(struct intel_guc_ct_channel *ctch) static int ctb_write(struct intel_guc_ct_buffer *ctb, const u32 *action, u32 len /* in dwords */, - u32 fence) + u32 fence, + bool want_response) { struct guc_ct_buffer_desc *desc = ctb->desc; u32 head = desc->head / 4; /* in dwords */ @@ -331,6 +343,7 @@ static int ctb_write(struct intel_guc_ct_buffer *ctb, */ header = (len << GUC_CT_MSG_LEN_SHIFT) | (GUC_CT_MSG_WRITE_FENCE_TO_DESC) | + (want_response ? GUC_CT_MSG_SEND_STATUS : 0) | (action[0] << GUC_CT_MSG_ACTION_SHIFT); cmds[tail] = header; @@ -401,36 +414,108 @@ static int wait_for_ctb_desc_update(struct guc_ct_buffer_desc *desc, return err; } -static int ctch_send(struct intel_guc *guc, +/** + * wait_for_ct_request_update - Wait for CT request state update. + * @req: pointer to pending request + * @status: placeholder for status + * + * For each sent request, Guc shall send bac CT response message. + * Our message handler will update status of tracked request once + * response message with given fence is received. Wait here and + * check for valid response status value. + * + * Return: + * * 0 response received (status is valid) + * * -ETIMEDOUT no response within hardcoded timeout + */ +static int wait_for_ct_request_update(struct ct_request *req, u32 *status) +{ + int err; + + /* + * Fast commands should complete in less than 10us, so sample quickly + * up to that length of time, then switch to a slower sleep-wait loop. + * No GuC command should ever take longer than 10ms. + */ +#define done INTEL_GUC_MSG_IS_RESPONSE(READ_ONCE(req->status)) + err = wait_for_us(done, 10); + if (err) + err = wait_for(done, 10); +#undef done + + if (unlikely(err)) + DRM_ERROR("CT: fence %u err %d\n", req->fence, err); + + *status = req->status; + return err; +} + +static int ctch_send(struct intel_guc_ct *ct, struct intel_guc_ct_channel *ctch, const u32 *action, u32 len, + u32 *response_buf, + u32 response_buf_size, u32 *status) { struct intel_guc_ct_buffer *ctb = &ctch->ctbs[CTB_SEND]; struct guc_ct_buffer_desc *desc = ctb->desc; + struct ct_request request; + unsigned long flags; u32 fence; int err; GEM_BUG_ON(!ctch_is_open(ctch)); GEM_BUG_ON(!len); GEM_BUG_ON(len & ~GUC_CT_MSG_LEN_MASK); + GEM_BUG_ON(!response_buf && response_buf_size); fence = ctch_get_next_fence(ctch); - err = ctb_write(ctb, action, len, fence); + request.fence = fence; + request.status = 0; + request.response_len = response_buf_size; + request.response_buf = response_buf; + + spin_lock_irqsave(&ct->lock, flags); + list_add_tail(&request.link, &ct->pending_requests); + spin_unlock_irqrestore(&ct->lock, flags); + + err = ctb_write(ctb, action, len, fence, !!response_buf); if (unlikely(err)) - return err; + goto unlink; - intel_guc_notify(guc); + intel_guc_notify(ct_to_guc(ct)); - err = wait_for_ctb_desc_update(desc, fence, status); + if (response_buf) + err = wait_for_ct_request_update(&request, status); + else + err = wait_for_ctb_desc_update(desc, fence, status); if (unlikely(err)) - return err; - if (!INTEL_GUC_MSG_IS_RESPONSE_SUCCESS(*status)) - return -EIO; + goto unlink; + + if (!INTEL_GUC_MSG_IS_RESPONSE_SUCCESS(*status)) { + err = -EIO; + goto unlink; + } + + if (response_buf) { + /* There shall be no data in the status */ + WARN_ON(INTEL_GUC_MSG_TO_DATA(request.status)); + /* Return actual response len */ + err = request.response_len; + } else { + /* There shall be no response payload */ + WARN_ON(request.response_len); + /* Return data decoded from the status dword */ + err = INTEL_GUC_MSG_TO_DATA(*status); + } - /* Use data from the GuC status as our return value */ - return INTEL_GUC_MSG_TO_DATA(*status); +unlink: + spin_lock_irqsave(&ct->lock, flags); + list_del(&request.link); + spin_unlock_irqrestore(&ct->lock, flags); + + return err; } /* @@ -439,13 +524,15 @@ static int ctch_send(struct intel_guc *guc, static int intel_guc_send_ct(struct intel_guc *guc, const u32 *action, u32 len, u32 *response_buf, u32 response_buf_size) { - struct intel_guc_ct_channel *ctch = &guc->ct.host_channel; + struct intel_guc_ct *ct = &guc->ct; + struct intel_guc_ct_channel *ctch = &ct->host_channel; u32 status = ~0; /* undefined */ int ret; mutex_lock(&guc->send_mutex); - ret = ctch_send(guc, ctch, action, len, &status); + ret = ctch_send(ct, ctch, action, len, response_buf, response_buf_size, + &status); if (unlikely(ret < 0)) { DRM_ERROR("CT: send action %#X failed; err=%d status=%#X\n", action[0], ret, status); @@ -546,8 +633,12 @@ static int ct_handle_response(struct intel_guc_ct *ct, const u32 *msg) u32 msglen = len + 1; /* total message length including header */ u32 fence; u32 status; + u32 datalen; + struct ct_request *req; + bool found = false; GEM_BUG_ON(!ct_header_is_response(header)); + GEM_BUG_ON(!in_irq()); /* Response payload shall at least include fence and status */ if (unlikely(len < 2)) { @@ -557,6 +648,7 @@ static int ct_handle_response(struct intel_guc_ct *ct, const u32 *msg) fence = msg[1]; status = msg[2]; + datalen = len - 2; /* Format of the status follows RESPONSE message */ if (unlikely(!INTEL_GUC_MSG_IS_RESPONSE(status))) { @@ -564,7 +656,29 @@ static int ct_handle_response(struct intel_guc_ct *ct, const u32 *msg) return -EPROTO; } - /* XXX */ + spin_lock(&ct->lock); + list_for_each_entry(req, &ct->pending_requests, link) { + if (unlikely(fence != req->fence)) { + DRM_DEBUG_DRIVER("CT: request %u awaits response\n", + req->fence); + continue; + } + if (unlikely(datalen > req->response_len)) { + DRM_ERROR("CT: response %u too long %*phn\n", + req->fence, 4 * msglen, msg); + datalen = 0; + } + if (datalen) + memcpy(req->response_buf, msg + 3, 4 * datalen); + req->response_len = datalen; + WRITE_ONCE(req->status, status); + found = true; + break; + } + spin_unlock(&ct->lock); + + if (!found) + DRM_ERROR("CT: unsolicited response %*phn\n", 4 * msglen, msg); return 0; } diff --git a/drivers/gpu/drm/i915/intel_guc_ct.h b/drivers/gpu/drm/i915/intel_guc_ct.h index 595c8ad..fac6e53 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.h +++ b/drivers/gpu/drm/i915/intel_guc_ct.h @@ -75,6 +75,12 @@ struct intel_guc_ct_channel { struct intel_guc_ct { struct intel_guc_ct_channel host_channel; /* other channels are tbd */ + + /** @lock: protects pending requests list */ + spinlock_t lock; + + /** @pending_requests: list of requests waiting for response */ + struct list_head pending_requests; }; void intel_guc_ct_init_early(struct intel_guc_ct *ct); diff --git a/drivers/gpu/drm/i915/intel_guc_fwif.h b/drivers/gpu/drm/i915/intel_guc_fwif.h index 83143e8..d73673f 100644 --- a/drivers/gpu/drm/i915/intel_guc_fwif.h +++ b/drivers/gpu/drm/i915/intel_guc_fwif.h @@ -327,6 +327,58 @@ struct guc_stage_desc { u64 desc_private; } __packed; +/** + * DOC: CTB based communication + * + * The CTB (command transport buffer) communication between Host and GuC + * is based on u32 data stream written to the shared buffer. One buffer can + * be used to transmit data only in one direction (one-directional channel). + * + * Current status of the each buffer is stored in the buffer descriptor. + * Buffer descriptor holds tail and head fields that represents active data + * stream. The tail field is updated by the data producer (sender), and head + * field is updated by the data consumer (receiver):: + * + * +------------+ + * | DESCRIPTOR | +=================+============+========+ + * +============+ | | MESSAGE(s) | | + * | address |--------->+=================+============+========+ + * +------------+ + * | head | ^-----head--------^ + * +------------+ + * | tail | ^---------tail-----------------^ + * +------------+ + * | size | ^---------------size--------------------^ + * +------------+ + * + * Each message in data stream starts with the single u32 treated as a header, + * followed by optional set of u32 data that makes message specific payload:: + * + * +------------+---------+---------+---------+ + * | MESSAGE | + * +------------+---------+---------+---------+ + * | msg[0] | [1] | ... | [n-1] | + * +------------+---------+---------+---------+ + * | MESSAGE | MESSAGE PAYLOAD | + * + HEADER +---------+---------+---------+ + * | | 0 | ... | n | + * +======+=====+=========+=========+=========+ + * | 31:16| code| | | | + * +------+-----+ | | | + * | 15:5|flags| | | | + * +------+-----+ | | | + * | 4:0| len| | | | + * +------+-----+---------+---------+---------+ + * + * ^-------------len-------------^ + * + * The message header consists of: + * + * - **len**, indicates length of the message payload (in u32) + * - **code**, indicates message code + * - **flags**, holds various bits to control message handling + */ + /* * Describes single command transport buffer. * Used by both guc-master and clients. -- cgit v1.1 From 6c77a2b05841a4082f746bf58345209df57175c9 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 26 Mar 2018 19:48:26 +0000 Subject: drm/i915/guc: Prepare to process incoming requests from CT Requests are read from CT in the irq handler, but actual processing will be done in the work thread. Processing of specific actions will be added in the upcoming patches. v2: don't use GEM_BUG_ON (Chris) don't kmalloc too large buffer (Michal) v3: rebased v4: don't name it 'dispatch' (Michel) and fix checkpatch add some documentation (Michal) Signed-off-by: Michal Wajdeczko Cc: Oscar Mateo Cc: Michel Thierry Cc: Daniele Ceraolo Spurio Reviewed-by: Michel Thierry Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180326194829.58836-10-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc_ct.c | 95 ++++++++++++++++++++++++++++++++++++- drivers/gpu/drm/i915/intel_guc_ct.h | 6 +++ 2 files changed, 100 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/intel_guc_ct.c b/drivers/gpu/drm/i915/intel_guc_ct.c index 41b071c..aa810ad 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/intel_guc_ct.c @@ -32,10 +32,17 @@ struct ct_request { u32 *response_buf; }; +struct ct_incoming_request { + struct list_head link; + u32 msg[]; +}; + enum { CTB_SEND = 0, CTB_RECV = 1 }; enum { CTB_OWNER_HOST = 0 }; +static void ct_incoming_request_worker_func(struct work_struct *w); + /** * intel_guc_ct_init_early - Initialize CT state without requiring device access * @ct: pointer to CT struct @@ -47,6 +54,8 @@ void intel_guc_ct_init_early(struct intel_guc_ct *ct) spin_lock_init(&ct->lock); INIT_LIST_HEAD(&ct->pending_requests); + INIT_LIST_HEAD(&ct->incoming_requests); + INIT_WORK(&ct->worker, ct_incoming_request_worker_func); } static inline struct intel_guc *ct_to_guc(struct intel_guc_ct *ct) @@ -682,13 +691,97 @@ static int ct_handle_response(struct intel_guc_ct *ct, const u32 *msg) return 0; } +static void ct_process_request(struct intel_guc_ct *ct, + u32 action, u32 len, const u32 *payload) +{ + switch (action) { + default: + DRM_ERROR("CT: unexpected request %x %*phn\n", + action, 4 * len, payload); + break; + } +} + +static bool ct_process_incoming_requests(struct intel_guc_ct *ct) +{ + unsigned long flags; + struct ct_incoming_request *request; + u32 header; + u32 *payload; + bool done; + + spin_lock_irqsave(&ct->lock, flags); + request = list_first_entry_or_null(&ct->incoming_requests, + struct ct_incoming_request, link); + if (request) + list_del(&request->link); + done = !!list_empty(&ct->incoming_requests); + spin_unlock_irqrestore(&ct->lock, flags); + + if (!request) + return true; + + header = request->msg[0]; + payload = &request->msg[1]; + ct_process_request(ct, + ct_header_get_action(header), + ct_header_get_len(header), + payload); + + kfree(request); + return done; +} + +static void ct_incoming_request_worker_func(struct work_struct *w) +{ + struct intel_guc_ct *ct = container_of(w, struct intel_guc_ct, worker); + bool done; + + done = ct_process_incoming_requests(ct); + if (!done) + queue_work(system_unbound_wq, &ct->worker); +} + +/** + * DOC: CTB GuC to Host request + * + * Format of the CTB GuC to Host request message is as follows:: + * + * +------------+---------+---------+---------+---------+---------+ + * | msg[0] | [1] | [2] | [3] | ... | [n-1] | + * +------------+---------+---------+---------+---------+---------+ + * | MESSAGE | MESSAGE PAYLOAD | + * + HEADER +---------+---------+---------+---------+---------+ + * | | 0 | 1 | 2 | ... | n | + * +============+=========+=========+=========+=========+=========+ + * | len | request specific data | + * +------+-----+---------+---------+---------+---------+---------+ + * + * ^-----------------------len-----------------------^ + */ + static int ct_handle_request(struct intel_guc_ct *ct, const u32 *msg) { u32 header = msg[0]; + u32 len = ct_header_get_len(header); + u32 msglen = len + 1; /* total message length including header */ + struct ct_incoming_request *request; + unsigned long flags; GEM_BUG_ON(ct_header_is_response(header)); - /* XXX */ + request = kmalloc(sizeof(*request) + 4 * msglen, GFP_ATOMIC); + if (unlikely(!request)) { + DRM_ERROR("CT: dropping request %*phn\n", 4 * msglen, msg); + return 0; /* XXX: -ENOMEM ? */ + } + memcpy(request->msg, msg, 4 * msglen); + + spin_lock_irqsave(&ct->lock, flags); + list_add_tail(&request->link, &ct->incoming_requests); + spin_unlock_irqrestore(&ct->lock, flags); + + queue_work(system_unbound_wq, &ct->worker); return 0; } diff --git a/drivers/gpu/drm/i915/intel_guc_ct.h b/drivers/gpu/drm/i915/intel_guc_ct.h index fac6e53..d774895a 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.h +++ b/drivers/gpu/drm/i915/intel_guc_ct.h @@ -81,6 +81,12 @@ struct intel_guc_ct { /** @pending_requests: list of requests waiting for response */ struct list_head pending_requests; + + /** @incoming_requests: list of incoming requests */ + struct list_head incoming_requests; + + /** @worker: worker for handling incoming requests */ + struct work_struct worker; }; void intel_guc_ct_init_early(struct intel_guc_ct *ct); -- cgit v1.1 From b6b0166d49f04ace13885a7ef42cb4a63ecf1f02 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Tue, 27 Mar 2018 21:41:24 +0000 Subject: drm/i915/guc: Handle default action received over CT When running on platform with CTB based GuC communication enabled, GuC to Host event data will be delivered as CT request message. However, content of the data[1] of this CT message follows format of the scratch register used in MMIO based communication, so some code reuse is still possible. v2: filter disabled messages (Daniele) Signed-off-by: Michal Wajdeczko Cc: Oscar Mateo Reviewed-by: Michel Thierry #1 Cc: Daniele Ceraolo Spurio Acked-by: Daniele Ceraolo Spurio Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180327214124.70680-1-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/intel_guc.c | 8 ++++++++ drivers/gpu/drm/i915/intel_guc.h | 1 + drivers/gpu/drm/i915/intel_guc_ct.c | 9 +++++++++ 3 files changed, 18 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index 411c8e9..a00a59a 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -416,6 +416,14 @@ void intel_guc_to_host_event_handler_mmio(struct intel_guc *guc) I915_WRITE(SOFT_SCRATCH(15), val & ~msg); spin_unlock(&guc->irq_lock); + intel_guc_to_host_process_recv_msg(guc, msg); +} + +void intel_guc_to_host_process_recv_msg(struct intel_guc *guc, u32 msg) +{ + /* Make sure to handle only enabled messages */ + msg &= guc->msg_enabled_mask; + if (msg & (INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED)) intel_guc_log_handle_flush_event(&guc->log); diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h index 6dc109a..f1265e1 100644 --- a/drivers/gpu/drm/i915/intel_guc.h +++ b/drivers/gpu/drm/i915/intel_guc.h @@ -163,6 +163,7 @@ int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len, void intel_guc_to_host_event_handler(struct intel_guc *guc); void intel_guc_to_host_event_handler_nop(struct intel_guc *guc); void intel_guc_to_host_event_handler_mmio(struct intel_guc *guc); +void intel_guc_to_host_process_recv_msg(struct intel_guc *guc, u32 msg); int intel_guc_sample_forcewake(struct intel_guc *guc); int intel_guc_auth_huc(struct intel_guc *guc, u32 rsa_offset); int intel_guc_suspend(struct intel_guc *guc); diff --git a/drivers/gpu/drm/i915/intel_guc_ct.c b/drivers/gpu/drm/i915/intel_guc_ct.c index aa810ad..e837084 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/intel_guc_ct.c @@ -694,8 +694,17 @@ static int ct_handle_response(struct intel_guc_ct *ct, const u32 *msg) static void ct_process_request(struct intel_guc_ct *ct, u32 action, u32 len, const u32 *payload) { + struct intel_guc *guc = ct_to_guc(ct); + switch (action) { + case INTEL_GUC_ACTION_DEFAULT: + if (unlikely(len < 1)) + goto fail_unexpected; + intel_guc_to_host_process_recv_msg(guc, *payload); + break; + default: +fail_unexpected: DRM_ERROR("CT: unexpected request %x %*phn\n", action, 4 * len, payload); break; -- cgit v1.1 From 0a015ff9730c169aa0d581e3f7727752ba3ff5b3 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 26 Mar 2018 19:48:28 +0000 Subject: drm/i915/guc: Trace messages from CT while in debug During debug we may want to investigate all communication from the Guc. Add proper tracing macros in debug config. v2: convert remaining DRM_DEBUG into new CT_DEBUG (Michal) v3: use dedicated Kconfig (Daniele) v4: checkpatch Signed-off-by: Michal Wajdeczko Cc: Daniele Ceraolo Spurio Cc: Chris Wilson Acked-by: Daniele Ceraolo Spurio Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180326194829.58836-12-michal.wajdeczko@intel.com --- drivers/gpu/drm/i915/Kconfig.debug | 12 +++++++++++ drivers/gpu/drm/i915/intel_guc_ct.c | 43 ++++++++++++++++++++++++++----------- 2 files changed, 43 insertions(+), 12 deletions(-) diff --git a/drivers/gpu/drm/i915/Kconfig.debug b/drivers/gpu/drm/i915/Kconfig.debug index dd5bf63..80efee1 100644 --- a/drivers/gpu/drm/i915/Kconfig.debug +++ b/drivers/gpu/drm/i915/Kconfig.debug @@ -90,6 +90,18 @@ config DRM_I915_SW_FENCE_CHECK_DAG If in doubt, say "N". +config DRM_I915_DEBUG_GUC + bool "Enable additional driver debugging for GuC" + depends on DRM_I915 + default n + help + Choose this option to turn on extra driver debugging that may affect + performance but will help resolve GuC related issues. + + Recommended for driver developers only. + + If in doubt, say "N". + config DRM_I915_SELFTEST bool "Enable selftests upon driver load" depends on DRM_I915 diff --git a/drivers/gpu/drm/i915/intel_guc_ct.c b/drivers/gpu/drm/i915/intel_guc_ct.c index e837084..990141d 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/intel_guc_ct.c @@ -24,6 +24,12 @@ #include "i915_drv.h" #include "intel_guc_ct.h" +#ifdef CONFIG_DRM_I915_DEBUG_GUC +#define CT_DEBUG_DRIVER(...) DRM_DEBUG_DRIVER(__VA_ARGS__) +#else +#define CT_DEBUG_DRIVER(...) do { } while (0) +#endif + struct ct_request { struct list_head link; u32 fence; @@ -78,8 +84,8 @@ static inline const char *guc_ct_buffer_type_to_str(u32 type) static void guc_ct_buffer_desc_init(struct guc_ct_buffer_desc *desc, u32 cmds_addr, u32 size, u32 owner) { - DRM_DEBUG_DRIVER("CT: desc %p init addr=%#x size=%u owner=%u\n", - desc, cmds_addr, size, owner); + CT_DEBUG_DRIVER("CT: desc %p init addr=%#x size=%u owner=%u\n", + desc, cmds_addr, size, owner); memset(desc, 0, sizeof(*desc)); desc->addr = cmds_addr; desc->size = size; @@ -88,8 +94,8 @@ static void guc_ct_buffer_desc_init(struct guc_ct_buffer_desc *desc, static void guc_ct_buffer_desc_reset(struct guc_ct_buffer_desc *desc) { - DRM_DEBUG_DRIVER("CT: desc %p reset head=%u tail=%u\n", - desc, desc->head, desc->tail); + CT_DEBUG_DRIVER("CT: desc %p reset head=%u tail=%u\n", + desc, desc->head, desc->tail); desc->head = 0; desc->tail = 0; desc->is_in_error = 0; @@ -185,8 +191,8 @@ static int ctch_init(struct intel_guc *guc, err = PTR_ERR(blob); goto err_vma; } - DRM_DEBUG_DRIVER("CT: vma base=%#x\n", - intel_guc_ggtt_offset(guc, ctch->vma)); + CT_DEBUG_DRIVER("CT: vma base=%#x\n", + intel_guc_ggtt_offset(guc, ctch->vma)); /* store pointers to desc and cmds */ for (i = 0; i < ARRAY_SIZE(ctch->ctbs); i++) { @@ -200,8 +206,8 @@ static int ctch_init(struct intel_guc *guc, err_vma: i915_vma_unpin_and_release(&ctch->vma); err_out: - DRM_DEBUG_DRIVER("CT: channel %d initialization failed; err=%d\n", - ctch->owner, err); + CT_DEBUG_DRIVER("CT: channel %d initialization failed; err=%d\n", + ctch->owner, err); return err; } @@ -221,8 +227,8 @@ static int ctch_open(struct intel_guc *guc, int err; int i; - DRM_DEBUG_DRIVER("CT: channel %d reopen=%s\n", - ctch->owner, yesno(ctch_is_open(ctch))); + CT_DEBUG_DRIVER("CT: channel %d reopen=%s\n", + ctch->owner, yesno(ctch_is_open(ctch))); if (!ctch->vma) { err = ctch_init(guc, ctch); @@ -355,6 +361,10 @@ static int ctb_write(struct intel_guc_ct_buffer *ctb, (want_response ? GUC_CT_MSG_SEND_STATUS : 0) | (action[0] << GUC_CT_MSG_ACTION_SHIFT); + CT_DEBUG_DRIVER("CT: writing %*phn %*phn %*phn\n", + 4, &header, 4, &fence, + 4 * (len - 1), &action[1]); + cmds[tail] = header; tail = (tail + 1) % size; @@ -545,6 +555,9 @@ static int intel_guc_send_ct(struct intel_guc *guc, const u32 *action, u32 len, if (unlikely(ret < 0)) { DRM_ERROR("CT: send action %#X failed; err=%d status=%#X\n", action[0], ret, status); + } else if (unlikely(ret)) { + CT_DEBUG_DRIVER("CT: send action %#x returned %d (%#x)\n", + action[0], ret, ret); } mutex_unlock(&guc->send_mutex); @@ -591,6 +604,7 @@ static int ctb_read(struct intel_guc_ct_buffer *ctb, u32 *data) /* beware of buffer wrap case */ if (unlikely(available < 0)) available += size; + CT_DEBUG_DRIVER("CT: available %d (%u:%u)\n", available, head, tail); GEM_BUG_ON(available < 0); data[0] = cmds[head]; @@ -612,6 +626,7 @@ static int ctb_read(struct intel_guc_ct_buffer *ctb, u32 *data) data[i] = cmds[head]; head = (head + 1) % size; } + CT_DEBUG_DRIVER("CT: received %*phn\n", 4 * len, data); desc->head = head * 4; return 0; @@ -665,11 +680,13 @@ static int ct_handle_response(struct intel_guc_ct *ct, const u32 *msg) return -EPROTO; } + CT_DEBUG_DRIVER("CT: response fence %u status %#x\n", fence, status); + spin_lock(&ct->lock); list_for_each_entry(req, &ct->pending_requests, link) { if (unlikely(fence != req->fence)) { - DRM_DEBUG_DRIVER("CT: request %u awaits response\n", - req->fence); + CT_DEBUG_DRIVER("CT: request %u awaits response\n", + req->fence); continue; } if (unlikely(datalen > req->response_len)) { @@ -696,6 +713,8 @@ static void ct_process_request(struct intel_guc_ct *ct, { struct intel_guc *guc = ct_to_guc(ct); + CT_DEBUG_DRIVER("CT: request %x %*phn\n", action, 4 * len, payload); + switch (action) { case INTEL_GUC_ACTION_DEFAULT: if (unlikely(len < 1)) -- cgit v1.1 From 790861cc34f872015806cef311d5c64cc3167a0d Mon Sep 17 00:00:00 2001 From: Samuel Li Date: Mon, 26 Mar 2018 15:22:25 -0400 Subject: drm: add parameter explanation for some gem dmabuf_ops To reduce some warnings. Reviewed-by: Alex Deucher Signed-off-by: Samuel Li Signed-off-by: Alex Deucher Link: https://patchwork.freedesktop.org/patch/msgid/1522092145-12645-1-git-send-email-Samuel.Li@amd.com --- drivers/gpu/drm/drm_prime.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c index 7856a9b..caf675e 100644 --- a/drivers/gpu/drm/drm_prime.c +++ b/drivers/gpu/drm/drm_prime.c @@ -331,6 +331,9 @@ EXPORT_SYMBOL(drm_gem_map_dma_buf); /** * drm_gem_unmap_dma_buf - unmap_dma_buf implementation for GEM + * @attach: attachment to unmap buffer from + * @sgt: scatterlist info of the buffer to unmap + * @dir: direction of DMA transfer * * Not implemented. The unmap is done at drm_gem_map_detach(). This can be * used as the &dma_buf_ops.unmap_dma_buf callback. @@ -429,6 +432,8 @@ EXPORT_SYMBOL(drm_gem_dmabuf_vunmap); /** * drm_gem_dmabuf_kmap_atomic - map_atomic implementation for GEM + * @dma_buf: buffer to be mapped + * @page_num: page number within the buffer * * Not implemented. This can be used as the &dma_buf_ops.map_atomic callback. */ @@ -441,6 +446,9 @@ EXPORT_SYMBOL(drm_gem_dmabuf_kmap_atomic); /** * drm_gem_dmabuf_kunmap_atomic - unmap_atomic implementation for GEM + * @dma_buf: buffer to be unmapped + * @page_num: page number within the buffer + * @addr: virtual address of the buffer * * Not implemented. This can be used as the &dma_buf_ops.unmap_atomic callback. */ @@ -453,6 +461,8 @@ EXPORT_SYMBOL(drm_gem_dmabuf_kunmap_atomic); /** * drm_gem_dmabuf_kmap - map implementation for GEM + * @dma_buf: buffer to be mapped + * @page_num: page number within the buffer * * Not implemented. This can be used as the &dma_buf_ops.map callback. */ @@ -464,6 +474,9 @@ EXPORT_SYMBOL(drm_gem_dmabuf_kmap); /** * drm_gem_dmabuf_kunmap - unmap implementation for GEM + * @dma_buf: buffer to be unmapped + * @page_num: page number within the buffer + * @addr: virtual address of the buffer * * Not implemented. This can be used as the &dma_buf_ops.unmap callback. */ -- cgit v1.1 From bee330f3d67273a68dcb99f59480d59553c008b2 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Noralf=20Tr=C3=B8nnes?= Date: Wed, 28 Mar 2018 10:38:35 +0300 Subject: drm: Use srcu to protect drm_device.unplugged MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Use srcu to protect drm_device.unplugged in a race free manner. Drivers can use drm_dev_enter()/drm_dev_exit() to protect and mark sections preventing access to device resources that are not available after the device is gone. Suggested-by: Daniel Vetter Signed-off-by: Noralf Trønnes Signed-off-by: Oleksandr Andrushchenko Reviewed-by: Oleksandr Andrushchenko Tested-by: Oleksandr Andrushchenko Cc: intel-gfx@lists.freedesktop.org Reviewed-by: Daniel Vetter Link: https://patchwork.freedesktop.org/patch/msgid/1522222715-11814-1-git-send-email-andr2000@gmail.com --- drivers/gpu/drm/drm_drv.c | 54 ++++++++++++++++++++++++++++++++++++++++++----- include/drm/drm_device.h | 9 +++++++- include/drm/drm_drv.h | 15 +++++++++---- 3 files changed, 68 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c index a1b9338..32a83b4 100644 --- a/drivers/gpu/drm/drm_drv.c +++ b/drivers/gpu/drm/drm_drv.c @@ -32,6 +32,7 @@ #include #include #include +#include #include #include @@ -75,6 +76,8 @@ static bool drm_core_init_complete = false; static struct dentry *drm_debugfs_root; +DEFINE_STATIC_SRCU(drm_unplug_srcu); + /* * DRM Minors * A DRM device can provide several char-dev interfaces on the DRM-Major. Each @@ -318,18 +321,51 @@ void drm_put_dev(struct drm_device *dev) } EXPORT_SYMBOL(drm_put_dev); -static void drm_device_set_unplugged(struct drm_device *dev) +/** + * drm_dev_enter - Enter device critical section + * @dev: DRM device + * @idx: Pointer to index that will be passed to the matching drm_dev_exit() + * + * This function marks and protects the beginning of a section that should not + * be entered after the device has been unplugged. The section end is marked + * with drm_dev_exit(). Calls to this function can be nested. + * + * Returns: + * True if it is OK to enter the section, false otherwise. + */ +bool drm_dev_enter(struct drm_device *dev, int *idx) +{ + *idx = srcu_read_lock(&drm_unplug_srcu); + + if (dev->unplugged) { + srcu_read_unlock(&drm_unplug_srcu, *idx); + return false; + } + + return true; +} +EXPORT_SYMBOL(drm_dev_enter); + +/** + * drm_dev_exit - Exit device critical section + * @idx: index returned from drm_dev_enter() + * + * This function marks the end of a section that should not be entered after + * the device has been unplugged. + */ +void drm_dev_exit(int idx) { - smp_wmb(); - atomic_set(&dev->unplugged, 1); + srcu_read_unlock(&drm_unplug_srcu, idx); } +EXPORT_SYMBOL(drm_dev_exit); /** * drm_dev_unplug - unplug a DRM device * @dev: DRM device * * This unplugs a hotpluggable DRM device, which makes it inaccessible to - * userspace operations. Entry-points can use drm_dev_is_unplugged(). This + * userspace operations. Entry-points can use drm_dev_enter() and + * drm_dev_exit() to protect device resources in a race free manner. This * essentially unregisters the device like drm_dev_unregister(), but can be * called while there are still open users of @dev. */ @@ -338,10 +374,18 @@ void drm_dev_unplug(struct drm_device *dev) drm_dev_unregister(dev); mutex_lock(&drm_global_mutex); - drm_device_set_unplugged(dev); if (dev->open_count == 0) drm_dev_put(dev); mutex_unlock(&drm_global_mutex); + + /* + * After synchronizing any critical read section is guaranteed to see + * the new value of ->unplugged, and any critical section which might + * still have seen the old value of ->unplugged is guaranteed to have + * finished. + */ + dev->unplugged = true; + synchronize_srcu(&drm_unplug_srcu); } EXPORT_SYMBOL(drm_dev_unplug); diff --git a/include/drm/drm_device.h b/include/drm/drm_device.h index 7c4fa32..3a0eac2 100644 --- a/include/drm/drm_device.h +++ b/include/drm/drm_device.h @@ -46,7 +46,14 @@ struct drm_device { /* currently active master for this device. Protected by master_mutex */ struct drm_master *master; - atomic_t unplugged; /**< Flag whether dev is dead */ + /** + * @unplugged: + * + * Flag to tell if the device has been unplugged. + * See drm_dev_enter() and drm_dev_is_unplugged(). + */ + bool unplugged; + struct inode *anon_inode; /**< inode for private address-space */ char *unique; /**< unique name of the device */ /*@} */ diff --git a/include/drm/drm_drv.h b/include/drm/drm_drv.h index d32b688..ff7312c 100644 --- a/include/drm/drm_drv.h +++ b/include/drm/drm_drv.h @@ -623,6 +623,8 @@ void drm_dev_get(struct drm_device *dev); void drm_dev_put(struct drm_device *dev); void drm_dev_unref(struct drm_device *dev); void drm_put_dev(struct drm_device *dev); +bool drm_dev_enter(struct drm_device *dev, int *idx); +void drm_dev_exit(int idx); void drm_dev_unplug(struct drm_device *dev); /** @@ -634,11 +636,16 @@ void drm_dev_unplug(struct drm_device *dev); * unplugged, these two functions guarantee that any store before calling * drm_dev_unplug() is visible to callers of this function after it completes */ -static inline int drm_dev_is_unplugged(struct drm_device *dev) +static inline bool drm_dev_is_unplugged(struct drm_device *dev) { - int ret = atomic_read(&dev->unplugged); - smp_rmb(); - return ret; + int idx; + + if (drm_dev_enter(dev, &idx)) { + drm_dev_exit(idx); + return false; + } + + return true; } -- cgit v1.1 From 4d07f6c40a764748f3c30c65d9f8e75a36fd1ad9 Mon Sep 17 00:00:00 2001 From: Michel Thierry Date: Wed, 28 Mar 2018 13:58:50 -0700 Subject: drm/i915/guc: enable guc interrupts unconditionally in uc_resume MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Probably lost while rebasing commit eacd8391f977 ("drm/i915/guc: Keep GuC interrupts enabled when using GuC"). Not really needed since i915_gem_init_hw is called before uc_resume, but it brings symmetry to uc_suspend. Signed-off-by: Michel Thierry Cc: Michał Winiarski Reviewed-by: Michał Winiarski Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180328205851.16188-1-michel.thierry@intel.com --- drivers/gpu/drm/i915/intel_uc.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 081e424..1cffaf7 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -480,8 +480,7 @@ int intel_uc_resume(struct drm_i915_private *i915) if (guc->fw.load_status != INTEL_UC_FIRMWARE_SUCCESS) return 0; - if (i915_modparams.guc_log_level) - gen9_enable_guc_interrupts(i915); + gen9_enable_guc_interrupts(i915); err = intel_guc_resume(guc); if (err) { -- cgit v1.1 From e770276079fd6e1088a255dee182a3c09a2d7aa9 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 27 Mar 2018 22:01:57 +0100 Subject: drm/i915: Include the HW breadcrumb whenever we trace the global_seqno When we include a request's global_seqno in a GEM_TRACE it often helps to know how that relates to the current breadcrumb as seen by the hardware. Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20180327210157.16896-3-chris@chris-wilson.co.uk Reviewed-by: Tvrtko Ursulin --- drivers/gpu/drm/i915/i915_request.c | 28 +++++++++++++++++----------- drivers/gpu/drm/i915/intel_lrc.c | 6 ++++-- 2 files changed, 21 insertions(+), 13 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index 2314a26..5852428 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -214,8 +214,11 @@ static int reset_all_global_seqno(struct drm_i915_private *i915, u32 seqno) struct i915_gem_timeline *timeline; struct intel_timeline *tl = engine->timeline; - GEM_TRACE("%s seqno %d -> %d\n", - engine->name, tl->seqno, seqno); + GEM_TRACE("%s seqno %d (current %d) -> %d\n", + engine->name, + tl->seqno, + intel_engine_get_seqno(engine), + seqno); if (!i915_seqno_passed(seqno, tl->seqno)) { /* Flush any waiters before we reuse the seqno */ @@ -386,10 +389,11 @@ static void i915_request_retire(struct i915_request *request) struct intel_engine_cs *engine = request->engine; struct i915_gem_active *active, *next; - GEM_TRACE("%s(%d) fence %llx:%d, global_seqno %d\n", - engine->name, intel_engine_get_seqno(engine), + GEM_TRACE("%s fence %llx:%d, global_seqno %d, current %d\n", + engine->name, request->fence.context, request->fence.seqno, - request->global_seqno); + request->global_seqno, + intel_engine_get_seqno(engine)); lockdep_assert_held(&request->i915->drm.struct_mutex); GEM_BUG_ON(!i915_sw_fence_signaled(&request->submit)); @@ -508,10 +512,11 @@ void __i915_request_submit(struct i915_request *request) struct intel_engine_cs *engine = request->engine; u32 seqno; - GEM_TRACE("%s fence %llx:%d -> global_seqno %d\n", - request->engine->name, + GEM_TRACE("%s fence %llx:%d -> global_seqno %d, current %d\n", + engine->name, request->fence.context, request->fence.seqno, - engine->timeline->seqno + 1); + engine->timeline->seqno + 1, + intel_engine_get_seqno(engine)); GEM_BUG_ON(!irqs_disabled()); lockdep_assert_held(&engine->timeline->lock); @@ -557,10 +562,11 @@ void __i915_request_unsubmit(struct i915_request *request) { struct intel_engine_cs *engine = request->engine; - GEM_TRACE("%s fence %llx:%d <- global_seqno %d\n", - request->engine->name, + GEM_TRACE("%s fence %llx:%d <- global_seqno %d, current %d\n", + engine->name, request->fence.context, request->fence.seqno, - request->global_seqno); + request->global_seqno, + intel_engine_get_seqno(engine)); GEM_BUG_ON(!irqs_disabled()); lockdep_assert_held(&engine->timeline->lock); diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 6546342..f60b61b 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -454,10 +454,11 @@ static void execlists_submit_ports(struct intel_engine_cs *engine) desc = execlists_update_context(rq); GEM_DEBUG_EXEC(port[n].context_id = upper_32_bits(desc)); - GEM_TRACE("%s in[%d]: ctx=%d.%d, seqno=%x, prio=%d\n", + GEM_TRACE("%s in[%d]: ctx=%d.%d, seqno=%d (current %d), prio=%d\n", engine->name, n, port[n].context_id, count, rq->global_seqno, + intel_engine_get_seqno(engine), rq_prio(rq)); } else { GEM_BUG_ON(!n); @@ -999,10 +1000,11 @@ static void execlists_submission_tasklet(unsigned long data) EXECLISTS_ACTIVE_USER)); rq = port_unpack(port, &count); - GEM_TRACE("%s out[0]: ctx=%d.%d, seqno=%x, prio=%d\n", + GEM_TRACE("%s out[0]: ctx=%d.%d, seqno=%d (current %d), prio=%d\n", engine->name, port->context_id, count, rq ? rq->global_seqno : 0, + intel_engine_get_seqno(engine), rq ? rq_prio(rq) : 0); /* Check the context/desc id for this event matches */ -- cgit v1.1 From 1de401c08fa805f3ac34604af1d43f48aeb17eb4 Mon Sep 17 00:00:00 2001 From: Lionel Landwerlin Date: Mon, 26 Mar 2018 14:39:48 +0100 Subject: drm/i915/perf: enable perf support on ICL No significant changes from either context offsets, nor report formats, nor register whitelist. v2: Also drop slice/unslice clock ratio changes (Matt) Signed-off-by: Lionel Landwerlin Reviewed-by: Matthew Auld Link: https://patchwork.freedesktop.org/patch/msgid/20180326133949.12469-3-lionel.g.landwerlin@intel.com --- drivers/gpu/drm/i915/Makefile | 3 +- drivers/gpu/drm/i915/i915_oa_icl.c | 118 +++++++++++++++++++++++++++++++++++++ drivers/gpu/drm/i915/i915_oa_icl.h | 34 +++++++++++ drivers/gpu/drm/i915/i915_perf.c | 7 ++- 4 files changed, 159 insertions(+), 3 deletions(-) create mode 100644 drivers/gpu/drm/i915/i915_oa_icl.c create mode 100644 drivers/gpu/drm/i915/i915_oa_icl.h diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile index 552e43e..0c79c19 100644 --- a/drivers/gpu/drm/i915/Makefile +++ b/drivers/gpu/drm/i915/Makefile @@ -172,7 +172,8 @@ i915-y += i915_perf.o \ i915_oa_glk.o \ i915_oa_cflgt2.o \ i915_oa_cflgt3.o \ - i915_oa_cnl.o + i915_oa_cnl.o \ + i915_oa_icl.o ifeq ($(CONFIG_DRM_I915_GVT),y) i915-y += intel_gvt.o diff --git a/drivers/gpu/drm/i915/i915_oa_icl.c b/drivers/gpu/drm/i915/i915_oa_icl.c new file mode 100644 index 0000000..a566792 --- /dev/null +++ b/drivers/gpu/drm/i915/i915_oa_icl.c @@ -0,0 +1,118 @@ +/* + * Autogenerated file by GPU Top : https://github.com/rib/gputop + * DO NOT EDIT manually! + * + * + * Copyright (c) 2015 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + */ + +#include + +#include "i915_drv.h" +#include "i915_oa_icl.h" + +static const struct i915_oa_reg b_counter_config_test_oa[] = { + { _MMIO(0x2740), 0x00000000 }, + { _MMIO(0x2710), 0x00000000 }, + { _MMIO(0x2714), 0xf0800000 }, + { _MMIO(0x2720), 0x00000000 }, + { _MMIO(0x2724), 0xf0800000 }, + { _MMIO(0x2770), 0x00000004 }, + { _MMIO(0x2774), 0x0000ffff }, + { _MMIO(0x2778), 0x00000003 }, + { _MMIO(0x277c), 0x0000ffff }, + { _MMIO(0x2780), 0x00000007 }, + { _MMIO(0x2784), 0x0000ffff }, + { _MMIO(0x2788), 0x00100002 }, + { _MMIO(0x278c), 0x0000fff7 }, + { _MMIO(0x2790), 0x00100002 }, + { _MMIO(0x2794), 0x0000ffcf }, + { _MMIO(0x2798), 0x00100082 }, + { _MMIO(0x279c), 0x0000ffef }, + { _MMIO(0x27a0), 0x001000c2 }, + { _MMIO(0x27a4), 0x0000ffe7 }, + { _MMIO(0x27a8), 0x00100001 }, + { _MMIO(0x27ac), 0x0000ffe7 }, +}; + +static const struct i915_oa_reg flex_eu_config_test_oa[] = { +}; + +static const struct i915_oa_reg mux_config_test_oa[] = { + { _MMIO(0xd04), 0x00000200 }, + { _MMIO(0x9840), 0x00000000 }, + { _MMIO(0x9884), 0x00000000 }, + { _MMIO(0x9888), 0x10060000 }, + { _MMIO(0x9888), 0x22060000 }, + { _MMIO(0x9888), 0x16060000 }, + { _MMIO(0x9888), 0x24060000 }, + { _MMIO(0x9888), 0x18060000 }, + { _MMIO(0x9888), 0x1a060000 }, + { _MMIO(0x9888), 0x12060000 }, + { _MMIO(0x9888), 0x14060000 }, + { _MMIO(0x9888), 0x10060000 }, + { _MMIO(0x9888), 0x22060000 }, + { _MMIO(0x9884), 0x00000003 }, + { _MMIO(0x9888), 0x16130000 }, + { _MMIO(0x9888), 0x24000001 }, + { _MMIO(0x9888), 0x0e130056 }, + { _MMIO(0x9888), 0x10130000 }, + { _MMIO(0x9888), 0x1a130000 }, + { _MMIO(0x9888), 0x541f0001 }, + { _MMIO(0x9888), 0x181f0000 }, + { _MMIO(0x9888), 0x4c1f0000 }, + { _MMIO(0x9888), 0x301f0000 }, +}; + +static ssize_t +show_test_oa_id(struct device *kdev, struct device_attribute *attr, char *buf) +{ + return sprintf(buf, "1\n"); +} + +void +i915_perf_load_test_config_icl(struct drm_i915_private *dev_priv) +{ + strlcpy(dev_priv->perf.oa.test_config.uuid, + "a291665e-244b-4b76-9b9a-01de9d3c8068", + sizeof(dev_priv->perf.oa.test_config.uuid)); + dev_priv->perf.oa.test_config.id = 1; + + dev_priv->perf.oa.test_config.mux_regs = mux_config_test_oa; + dev_priv->perf.oa.test_config.mux_regs_len = ARRAY_SIZE(mux_config_test_oa); + + dev_priv->perf.oa.test_config.b_counter_regs = b_counter_config_test_oa; + dev_priv->perf.oa.test_config.b_counter_regs_len = ARRAY_SIZE(b_counter_config_test_oa); + + dev_priv->perf.oa.test_config.flex_regs = flex_eu_config_test_oa; + dev_priv->perf.oa.test_config.flex_regs_len = ARRAY_SIZE(flex_eu_config_test_oa); + + dev_priv->perf.oa.test_config.sysfs_metric.name = "a291665e-244b-4b76-9b9a-01de9d3c8068"; + dev_priv->perf.oa.test_config.sysfs_metric.attrs = dev_priv->perf.oa.test_config.attrs; + + dev_priv->perf.oa.test_config.attrs[0] = &dev_priv->perf.oa.test_config.sysfs_metric_id.attr; + + dev_priv->perf.oa.test_config.sysfs_metric_id.attr.name = "id"; + dev_priv->perf.oa.test_config.sysfs_metric_id.attr.mode = 0444; + dev_priv->perf.oa.test_config.sysfs_metric_id.show = show_test_oa_id; +} diff --git a/drivers/gpu/drm/i915/i915_oa_icl.h b/drivers/gpu/drm/i915/i915_oa_icl.h new file mode 100644 index 0000000..ae1c24a --- /dev/null +++ b/drivers/gpu/drm/i915/i915_oa_icl.h @@ -0,0 +1,34 @@ +/* + * Autogenerated file by GPU Top : https://github.com/rib/gputop + * DO NOT EDIT manually! + * + * + * Copyright (c) 2015 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + */ + +#ifndef __I915_OA_ICL_H__ +#define __I915_OA_ICL_H__ + +extern void i915_perf_load_test_config_icl(struct drm_i915_private *dev_priv); + +#endif diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index abaca6e..30444bb 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -209,6 +209,7 @@ #include "i915_oa_cflgt2.h" #include "i915_oa_cflgt3.h" #include "i915_oa_cnl.h" +#include "i915_oa_icl.h" /* HW requires this to be a power of two, between 128k and 16M, though driver * is currently generally designed assuming the largest 16M size is used such @@ -1840,7 +1841,7 @@ static int gen8_enable_metric_set(struct drm_i915_private *dev_priv, * be read back from automatically triggered reports, as part of the * RPT_ID field. */ - if (IS_GEN9(dev_priv) || IS_GEN10(dev_priv)) { + if (IS_GEN(dev_priv, 9, 11)) { I915_WRITE(GEN8_OA_DEBUG, _MASKED_BIT_ENABLE(GEN9_OA_DEBUG_DISABLE_CLK_RATIO_REPORTS | GEN9_OA_DEBUG_INCLUDE_CLK_RATIO)); @@ -2935,6 +2936,8 @@ void i915_perf_register(struct drm_i915_private *dev_priv) i915_perf_load_test_config_cflgt3(dev_priv); } else if (IS_CANNONLAKE(dev_priv)) { i915_perf_load_test_config_cnl(dev_priv); + } else if (IS_ICELAKE(dev_priv)) { + i915_perf_load_test_config_icl(dev_priv); } if (dev_priv->perf.oa.test_config.id == 0) @@ -3467,7 +3470,7 @@ void i915_perf_init(struct drm_i915_private *dev_priv) dev_priv->perf.oa.gen8_valid_ctx_bit = (1<<16); } - } else if (IS_GEN10(dev_priv)) { + } else if (IS_GEN(dev_priv, 10, 11)) { dev_priv->perf.oa.ops.is_valid_b_counter_reg = gen7_is_valid_b_counter_addr; dev_priv->perf.oa.ops.is_valid_mux_reg = -- cgit v1.1 From b6dd47b9c82d619195370c38b7579fe18a8f6055 Mon Sep 17 00:00:00 2001 From: Lionel Landwerlin Date: Mon, 26 Mar 2018 10:08:22 +0100 Subject: drm/i915/perf: check the value of PROP_SAMPLE_OA uapi parameter We've been a bit loose about this opening parameter. We should only add the flag for writing OA reports when the value of this parameter is != 0. Signed-off-by: Lionel Landwerlin Reviewed-by: Matthew Auld Link: https://patchwork.freedesktop.org/patch/msgid/20180326090831.22686-3-lionel.g.landwerlin@intel.com --- drivers/gpu/drm/i915/i915_perf.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 30444bb..21a985b 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -2746,7 +2746,8 @@ static int read_properties_unlocked(struct drm_i915_private *dev_priv, props->ctx_handle = value; break; case DRM_I915_PERF_PROP_SAMPLE_OA: - props->sample_flags |= SAMPLE_OA_REPORT; + if (value) + props->sample_flags |= SAMPLE_OA_REPORT; break; case DRM_I915_PERF_PROP_OA_METRICS_SET: if (value == 0) { -- cgit v1.1 From 11051303344bd4b3334a7b456f6152f4d6032c61 Mon Sep 17 00:00:00 2001 From: Lionel Landwerlin Date: Mon, 26 Mar 2018 10:08:23 +0100 Subject: drm/i915/perf: simplify OA unit enabling on gen7 In commit d79651522e89c ("drm/i915: Enable i915 perf stream for Haswell OA unit") the enable/disable vfunc hadn't appear yet and the same function would deal with enabling/disabling the OA unit. This was split later on for gen8 but the gen7 retained some code that isn't actually useful anymore. Signed-off-by: Lionel Landwerlin Reviewed-by: Matthew Auld Link: https://patchwork.freedesktop.org/patch/msgid/20180326090831.22686-4-lionel.g.landwerlin@intel.com --- drivers/gpu/drm/i915/i915_perf.c | 34 +++++++++++++++------------------- 1 file changed, 15 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 21a985b..f1af58a 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -1886,6 +1886,13 @@ static void gen10_disable_metric_set(struct drm_i915_private *dev_priv) static void gen7_oa_enable(struct drm_i915_private *dev_priv) { + struct i915_gem_context *ctx = + dev_priv->perf.oa.exclusive_stream->ctx; + u32 ctx_id = dev_priv->perf.oa.specific_ctx_id; + bool periodic = dev_priv->perf.oa.periodic; + u32 period_exponent = dev_priv->perf.oa.period_exponent; + u32 report_format = dev_priv->perf.oa.oa_buffer.format; + /* * Reset buf pointers so we don't forward reports from before now. * @@ -1897,25 +1904,14 @@ static void gen7_oa_enable(struct drm_i915_private *dev_priv) */ gen7_init_oa_buffer(dev_priv); - if (dev_priv->perf.oa.exclusive_stream->enabled) { - struct i915_gem_context *ctx = - dev_priv->perf.oa.exclusive_stream->ctx; - u32 ctx_id = dev_priv->perf.oa.specific_ctx_id; - - bool periodic = dev_priv->perf.oa.periodic; - u32 period_exponent = dev_priv->perf.oa.period_exponent; - u32 report_format = dev_priv->perf.oa.oa_buffer.format; - - I915_WRITE(GEN7_OACONTROL, - (ctx_id & GEN7_OACONTROL_CTX_MASK) | - (period_exponent << - GEN7_OACONTROL_TIMER_PERIOD_SHIFT) | - (periodic ? GEN7_OACONTROL_TIMER_ENABLE : 0) | - (report_format << GEN7_OACONTROL_FORMAT_SHIFT) | - (ctx ? GEN7_OACONTROL_PER_CTX_ENABLE : 0) | - GEN7_OACONTROL_ENABLE); - } else - I915_WRITE(GEN7_OACONTROL, 0); + I915_WRITE(GEN7_OACONTROL, + (ctx_id & GEN7_OACONTROL_CTX_MASK) | + (period_exponent << + GEN7_OACONTROL_TIMER_PERIOD_SHIFT) | + (periodic ? GEN7_OACONTROL_TIMER_ENABLE : 0) | + (report_format << GEN7_OACONTROL_FORMAT_SHIFT) | + (ctx ? GEN7_OACONTROL_PER_CTX_ENABLE : 0) | + GEN7_OACONTROL_ENABLE); } static void gen8_oa_enable(struct drm_i915_private *dev_priv) -- cgit v1.1 From 53744104bec47a3357622f9d6815fda1c0d0f842 Mon Sep 17 00:00:00 2001 From: Lionel Landwerlin Date: Mon, 26 Mar 2018 10:08:25 +0100 Subject: drm/i915/perf: remove empty line This was added by mistake in commit 28964cf25ee67 ("drm/i915/perf: disable NOA logic when not used"). Signed-off-by: Lionel Landwerlin Reviewed-by: Matthew Auld Link: https://patchwork.freedesktop.org/patch/msgid/20180326090831.22686-6-lionel.g.landwerlin@intel.com --- drivers/gpu/drm/i915/i915_perf.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index f1af58a..b5c65ab 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -1871,7 +1871,6 @@ static void gen8_disable_metric_set(struct drm_i915_private *dev_priv) I915_WRITE(GDT_CHICKEN_BITS, (I915_READ(GDT_CHICKEN_BITS) & ~GT_NOA_ENABLE)); - } static void gen10_disable_metric_set(struct drm_i915_private *dev_priv) -- cgit v1.1 From b82ed43de5c01e41a7e5d756da3287fc66e7600b Mon Sep 17 00:00:00 2001 From: Lionel Landwerlin Date: Mon, 26 Mar 2018 10:08:26 +0100 Subject: drm/i915: rename PPGTT/GGTT fields OA registers We had a generic field name used across 2 registers but it feels like it's clearer we make it obvious what register this field belongs to. Signed-off-by: Lionel Landwerlin Reviewed-by: Matthew Auld Link: https://patchwork.freedesktop.org/patch/msgid/20180326090831.22686-7-lionel.g.landwerlin@intel.com --- drivers/gpu/drm/i915/i915_perf.c | 7 ++++--- drivers/gpu/drm/i915/i915_reg.h | 6 +++--- 2 files changed, 7 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index b5c65ab..d41a252 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -1043,7 +1043,7 @@ static int gen7_append_oa_reports(struct i915_perf_stream *stream, I915_WRITE(GEN7_OASTATUS2, ((head & GEN7_OASTATUS2_HEAD_MASK) | - OA_MEM_SELECT_GGTT)); + GEN7_OASTATUS2_MEM_SELECT_GGTT)); dev_priv->perf.oa.oa_buffer.head = head; spin_unlock_irqrestore(&dev_priv->perf.oa.oa_buffer.ptr_lock, flags); @@ -1333,7 +1333,8 @@ static void gen7_init_oa_buffer(struct drm_i915_private *dev_priv) /* Pre-DevBDW: OABUFFER must be set with counters off, * before OASTATUS1, but after OASTATUS2 */ - I915_WRITE(GEN7_OASTATUS2, gtt_offset | OA_MEM_SELECT_GGTT); /* head */ + I915_WRITE(GEN7_OASTATUS2, + gtt_offset | GEN7_OASTATUS2_MEM_SELECT_GGTT); /* head */ dev_priv->perf.oa.oa_buffer.head = gtt_offset; I915_WRITE(GEN7_OABUFFER, gtt_offset); @@ -1393,7 +1394,7 @@ static void gen8_init_oa_buffer(struct drm_i915_private *dev_priv) * bit." */ I915_WRITE(GEN8_OABUFFER, gtt_offset | - OABUFFER_SIZE_16M | OA_MEM_SELECT_GGTT); + OABUFFER_SIZE_16M | GEN8_OABUFFER_MEM_SELECT_GGTT); I915_WRITE(GEN8_OATAILPTR, gtt_offset & GEN8_OATAILPTR_MASK); /* Mark that we need updated tail pointers to read from... */ diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 5a53d0e..b926520 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -536,6 +536,7 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define GEN8_OABUFFER_UDW _MMIO(0x23b4) #define GEN8_OABUFFER _MMIO(0x2b14) +#define GEN8_OABUFFER_MEM_SELECT_GGTT (1 << 0) /* 0: PPGTT, 1: GGTT */ #define GEN7_OASTATUS1 _MMIO(0x2364) #define GEN7_OASTATUS1_TAIL_MASK 0xffffffc0 @@ -544,7 +545,8 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define GEN7_OASTATUS1_REPORT_LOST (1<<0) #define GEN7_OASTATUS2 _MMIO(0x2368) -#define GEN7_OASTATUS2_HEAD_MASK 0xffffffc0 +#define GEN7_OASTATUS2_HEAD_MASK 0xffffffc0 +#define GEN7_OASTATUS2_MEM_SELECT_GGTT (1 << 0) /* 0: PPGTT, 1: GGTT */ #define GEN8_OASTATUS _MMIO(0x2b08) #define GEN8_OASTATUS_OVERRUN_STATUS (1<<3) @@ -566,8 +568,6 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define OABUFFER_SIZE_8M (6<<3) #define OABUFFER_SIZE_16M (7<<3) -#define OA_MEM_SELECT_GGTT (1<<0) - /* * Flexible, Aggregate EU Counter Registers. * Note: these aren't contiguous -- cgit v1.1 From 9bd9be666008499bb9071e9e5472ded24e522f0b Mon Sep 17 00:00:00 2001 From: Lionel Landwerlin Date: Mon, 26 Mar 2018 10:08:28 +0100 Subject: drm/i915/perf: add more debug message on perf open & configs This will make it easier to spot issues related to config creation/usage. Signed-off-by: Lionel Landwerlin Reviewed-by: Matthew Auld Link: https://patchwork.freedesktop.org/patch/msgid/20180326090831.22686-9-lionel.g.landwerlin@intel.com --- drivers/gpu/drm/i915/i915_perf.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index d41a252..bfc906c 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -2096,13 +2096,17 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, if (stream->ctx) { ret = oa_get_render_ctx_id(stream); - if (ret) + if (ret) { + DRM_DEBUG("Invalid context id to filter with\n"); return ret; + } } ret = get_oa_config(dev_priv, props->metrics_set, &stream->oa_config); - if (ret) + if (ret) { + DRM_DEBUG("Invalid OA config id=%i\n", props->metrics_set); goto err_config; + } /* PRM - observability performance counters: * @@ -2129,8 +2133,10 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, ret = dev_priv->perf.oa.ops.enable_metric_set(dev_priv, stream->oa_config); - if (ret) + if (ret) { + DRM_DEBUG("Unable to enable metric set\n"); goto err_enable; + } stream->ops = &i915_oa_stream_ops; @@ -3292,6 +3298,8 @@ int i915_perf_add_config_ioctl(struct drm_device *dev, void *data, mutex_unlock(&dev_priv->perf.metrics_lock); + DRM_DEBUG("Added config %s id=%i\n", oa_config->uuid, oa_config->id); + return oa_config->id; sysfs_err: @@ -3348,6 +3356,9 @@ int i915_perf_remove_config_ioctl(struct drm_device *dev, void *data, &oa_config->sysfs_metric); idr_remove(&dev_priv->perf.metrics_idr, *arg); + + DRM_DEBUG("Removed config %s id=%i\n", oa_config->uuid, oa_config->id); + put_oa_config(dev_priv, oa_config); config_err: -- cgit v1.1 From 5e9cfeba6abb7e1a3f240bd24eb29178f0b83716 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 17:22:51 +0200 Subject: drm/atomic-helper: Drop plane->fb references only for drm_atomic_helper_shutdown() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit drm_atomic_helper_shutdown() needs to release the reference held by plane->fb. Since commit 49d70aeaeca8 ("drm/atomic-helper: Fix leak in disable_all") we're doing that by calling drm_atomic_clean_old_fb() in drm_atomic_helper_disable_all(). This also leaves plane->fb == NULL afterwards. However, since drm_atomic_helper_disable_all() is also used by the i915 gpu reset code drm_atomic_helper_commit_duplicated_state() then has to undo the damage and put the correct plane->fb pointers back in (and also adjust the ref counts to match again as well). That approach doesn't work so well for load detection as nothing sets up the plane->old_fb pointers for us. This causes us to leak an extra reference for each plane->fb when drm_atomic_helper_commit_duplicated_state() calls drm_atomic_clean_old_fb() after load detection. To fix this let's call drm_atomic_clean_old_fb() only for drm_atomic_helper_shutdown() as that's the only time we need to actually drop the plane->fb references. In all the other cases (load detection, gpu reset) we want to leave plane->fb alone. v2: Don't inflict the clean_old_fbs bool to drivers (Daniel) v3: Squash in the revert and rewrite the commit msg (Daniel) Cc: martin.peres@free.fr Cc: chris@chris-wilson.co.uk Cc: Dave Airlie Cc: Maarten Lankhorst Cc: Daniel Vetter Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322152313.6561-3-ville.syrjala@linux.intel.com Reviewed-by: Maarten Lankhorst #pre-squash Reviewed-by: Daniel Vetter --- drivers/gpu/drm/drm_atomic_helper.c | 78 ++++++++++++++++++------------------- 1 file changed, 39 insertions(+), 39 deletions(-) diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index d63c806..ef4ddfe 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -2892,31 +2892,9 @@ commit: return 0; } -/** - * drm_atomic_helper_disable_all - disable all currently active outputs - * @dev: DRM device - * @ctx: lock acquisition context - * - * Loops through all connectors, finding those that aren't turned off and then - * turns them off by setting their DPMS mode to OFF and deactivating the CRTC - * that they are connected to. - * - * This is used for example in suspend/resume to disable all currently active - * functions when suspending. If you just want to shut down everything at e.g. - * driver unload, look at drm_atomic_helper_shutdown(). - * - * Note that if callers haven't already acquired all modeset locks this might - * return -EDEADLK, which must be handled by calling drm_modeset_backoff(). - * - * Returns: - * 0 on success or a negative error code on failure. - * - * See also: - * drm_atomic_helper_suspend(), drm_atomic_helper_resume() and - * drm_atomic_helper_shutdown(). - */ -int drm_atomic_helper_disable_all(struct drm_device *dev, - struct drm_modeset_acquire_ctx *ctx) +static int __drm_atomic_helper_disable_all(struct drm_device *dev, + struct drm_modeset_acquire_ctx *ctx, + bool clean_old_fbs) { struct drm_atomic_state *state; struct drm_connector_state *conn_state; @@ -2968,8 +2946,11 @@ int drm_atomic_helper_disable_all(struct drm_device *dev, goto free; drm_atomic_set_fb_for_plane(plane_state, NULL); - plane_mask |= BIT(drm_plane_index(plane)); - plane->old_fb = plane->fb; + + if (clean_old_fbs) { + plane->old_fb = plane->fb; + plane_mask |= BIT(drm_plane_index(plane)); + } } ret = drm_atomic_commit(state); @@ -2980,6 +2961,34 @@ free: return ret; } +/** + * drm_atomic_helper_disable_all - disable all currently active outputs + * @dev: DRM device + * @ctx: lock acquisition context + * + * Loops through all connectors, finding those that aren't turned off and then + * turns them off by setting their DPMS mode to OFF and deactivating the CRTC + * that they are connected to. + * + * This is used for example in suspend/resume to disable all currently active + * functions when suspending. If you just want to shut down everything at e.g. + * driver unload, look at drm_atomic_helper_shutdown(). + * + * Note that if callers haven't already acquired all modeset locks this might + * return -EDEADLK, which must be handled by calling drm_modeset_backoff(). + * + * Returns: + * 0 on success or a negative error code on failure. + * + * See also: + * drm_atomic_helper_suspend(), drm_atomic_helper_resume() and + * drm_atomic_helper_shutdown(). + */ +int drm_atomic_helper_disable_all(struct drm_device *dev, + struct drm_modeset_acquire_ctx *ctx) +{ + return __drm_atomic_helper_disable_all(dev, ctx, false); +} EXPORT_SYMBOL(drm_atomic_helper_disable_all); /** @@ -3002,7 +3011,7 @@ void drm_atomic_helper_shutdown(struct drm_device *dev) while (1) { ret = drm_modeset_lock_all_ctx(dev, &ctx); if (!ret) - ret = drm_atomic_helper_disable_all(dev, &ctx); + ret = __drm_atomic_helper_disable_all(dev, &ctx, true); if (ret != -EDEADLK) break; @@ -3106,16 +3115,11 @@ int drm_atomic_helper_commit_duplicated_state(struct drm_atomic_state *state, struct drm_connector_state *new_conn_state; struct drm_crtc *crtc; struct drm_crtc_state *new_crtc_state; - unsigned plane_mask = 0; - struct drm_device *dev = state->dev; - int ret; state->acquire_ctx = ctx; - for_each_new_plane_in_state(state, plane, new_plane_state, i) { - plane_mask |= BIT(drm_plane_index(plane)); + for_each_new_plane_in_state(state, plane, new_plane_state, i) state->planes[i].old_state = plane->state; - } for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) state->crtcs[i].old_state = crtc->state; @@ -3123,11 +3127,7 @@ int drm_atomic_helper_commit_duplicated_state(struct drm_atomic_state *state, for_each_new_connector_in_state(state, connector, new_conn_state, i) state->connectors[i].old_state = connector->state; - ret = drm_atomic_commit(state); - if (plane_mask) - drm_atomic_clean_old_fb(dev, plane_mask, ret); - - return ret; + return drm_atomic_commit(state); } EXPORT_SYMBOL(drm_atomic_helper_commit_duplicated_state); -- cgit v1.1 From 7e7de761af2e4760cc5ad2968c37c2814317ede6 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 17:22:53 +0200 Subject: drm: Clear crtc->primary->crtc when disabling the crtc via setcrtc() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Keep the primary->crtc in sync with the state->crtc (also with primary->fb and state->fb) when disabling the crtc (and thus also the primary) via setcrtc(). Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322152313.6561-4-ville.syrjala@linux.intel.com Reviewed-by: Daniel Vetter Reviewed-by: Maarten Lankhorst --- drivers/gpu/drm/drm_crtc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index 0358388..7a973ada 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -471,7 +471,7 @@ static int __drm_mode_set_config_internal(struct drm_mode_set *set, ret = crtc->funcs->set_config(set, ctx); if (ret == 0) { - crtc->primary->crtc = crtc; + crtc->primary->crtc = fb ? crtc : NULL; crtc->primary->fb = fb; } -- cgit v1.1 From 5e78d01fa1a7f1794dfc44a6eb77bfec7a8d590d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 17:22:54 +0200 Subject: drm/atomic-helper: WARN if legacy plane fb pointers are bogus when committing duplicated state MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit drm_atomic_helper_commit_duplicated_state() should only be called resume/reset/load_detect paths where plane->old_fb should always be NULL and plane->fb should be equal to the new_plane_state->fb. Assert that is indeed the case. Cc: martin.peres@free.fr Cc: chris@chris-wilson.co.uk Cc: Dave Airlie (v1) Cc: Maarten Lankhorst Cc: Daniel Vetter Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322152313.6561-5-ville.syrjala@linux.intel.com Reviewed-by: Daniel Vetter Reviewed-by: Maarten Lankhorst --- drivers/gpu/drm/drm_atomic_helper.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index ef4ddfe..fe09d62 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -3118,8 +3118,13 @@ int drm_atomic_helper_commit_duplicated_state(struct drm_atomic_state *state, state->acquire_ctx = ctx; - for_each_new_plane_in_state(state, plane, new_plane_state, i) + for_each_new_plane_in_state(state, plane, new_plane_state, i) { + WARN_ON(plane->crtc != new_plane_state->crtc); + WARN_ON(plane->fb != new_plane_state->fb); + WARN_ON(plane->old_fb); + state->planes[i].old_state = plane->state; + } for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) state->crtcs[i].old_state = crtc->state; -- cgit v1.1 From 64c32b490333c9ccb05b172997c4f2f940c5d4d1 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 17:22:55 +0200 Subject: drm: Add local 'plane' variable for primary/cursor planes MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Make the code a bit more readable by storing the plane pointer in a local variable rather than having to do crtc->{primary,cursor} all the time. Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322152313.6561-6-ville.syrjala@linux.intel.com Reviewed-by: Daniel Vetter Reviewed-by: Maarten Lankhorst --- drivers/gpu/drm/drm_crtc.c | 32 +++++++++++++++++++------------- drivers/gpu/drm/drm_plane.c | 32 ++++++++++++++++++-------------- 2 files changed, 37 insertions(+), 27 deletions(-) diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index 7a973ada..8552ed4 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -402,6 +402,7 @@ int drm_mode_getcrtc(struct drm_device *dev, { struct drm_mode_crtc *crtc_resp = data; struct drm_crtc *crtc; + struct drm_plane *plane; if (!drm_core_check_feature(dev, DRIVER_MODESET)) return -EINVAL; @@ -410,21 +411,23 @@ int drm_mode_getcrtc(struct drm_device *dev, if (!crtc) return -ENOENT; + plane = crtc->primary; + crtc_resp->gamma_size = crtc->gamma_size; - drm_modeset_lock(&crtc->primary->mutex, NULL); - if (crtc->primary->state && crtc->primary->state->fb) - crtc_resp->fb_id = crtc->primary->state->fb->base.id; - else if (!crtc->primary->state && crtc->primary->fb) - crtc_resp->fb_id = crtc->primary->fb->base.id; + drm_modeset_lock(&plane->mutex, NULL); + if (plane->state && plane->state->fb) + crtc_resp->fb_id = plane->state->fb->base.id; + else if (!plane->state && plane->fb) + crtc_resp->fb_id = plane->fb->base.id; else crtc_resp->fb_id = 0; - if (crtc->primary->state) { - crtc_resp->x = crtc->primary->state->src_x >> 16; - crtc_resp->y = crtc->primary->state->src_y >> 16; + if (plane->state) { + crtc_resp->x = plane->state->src_x >> 16; + crtc_resp->y = plane->state->src_y >> 16; } - drm_modeset_unlock(&crtc->primary->mutex); + drm_modeset_unlock(&plane->mutex); drm_modeset_lock(&crtc->mutex, NULL); if (crtc->state) { @@ -554,6 +557,7 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data, struct drm_mode_config *config = &dev->mode_config; struct drm_mode_crtc *crtc_req = data; struct drm_crtc *crtc; + struct drm_plane *plane; struct drm_connector **connector_set = NULL, *connector; struct drm_framebuffer *fb = NULL; struct drm_display_mode *mode = NULL; @@ -580,6 +584,8 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data, } DRM_DEBUG_KMS("[CRTC:%d:%s]\n", crtc->base.id, crtc->name); + plane = crtc->primary; + mutex_lock(&crtc->dev->mode_config.mutex); drm_modeset_acquire_init(&ctx, DRM_MODESET_ACQUIRE_INTERRUPTIBLE); retry: @@ -590,12 +596,12 @@ retry: /* If we have a mode we need a framebuffer. */ /* If we pass -1, set the mode with the currently bound fb */ if (crtc_req->fb_id == -1) { - if (!crtc->primary->fb) { + if (!plane->fb) { DRM_DEBUG_KMS("CRTC doesn't have current FB\n"); ret = -EINVAL; goto out; } - fb = crtc->primary->fb; + fb = plane->fb; /* Make refcounting symmetric with the lookup path. */ drm_framebuffer_get(fb); } else { @@ -627,8 +633,8 @@ retry: * match real hardware capabilities. Skip the check in that * case. */ - if (!crtc->primary->format_default) { - ret = drm_plane_check_pixel_format(crtc->primary, + if (!plane->format_default) { + ret = drm_plane_check_pixel_format(plane, fb->format->format, fb->modifier); if (ret) { diff --git a/drivers/gpu/drm/drm_plane.c b/drivers/gpu/drm/drm_plane.c index 6d2a6e4..38e2a62 100644 --- a/drivers/gpu/drm/drm_plane.c +++ b/drivers/gpu/drm/drm_plane.c @@ -756,6 +756,7 @@ static int drm_mode_cursor_universal(struct drm_crtc *crtc, struct drm_modeset_acquire_ctx *ctx) { struct drm_device *dev = crtc->dev; + struct drm_plane *plane = crtc->cursor; struct drm_framebuffer *fb = NULL; struct drm_mode_fb_cmd2 fbreq = { .width = req->width, @@ -769,8 +770,8 @@ static int drm_mode_cursor_universal(struct drm_crtc *crtc, uint32_t src_w = 0, src_h = 0; int ret = 0; - BUG_ON(!crtc->cursor); - WARN_ON(crtc->cursor->crtc != crtc && crtc->cursor->crtc != NULL); + BUG_ON(!plane); + WARN_ON(plane->crtc != crtc && plane->crtc != NULL); /* * Obtain fb we'll be using (either new or existing) and take an extra @@ -790,7 +791,7 @@ static int drm_mode_cursor_universal(struct drm_crtc *crtc, fb = NULL; } } else { - fb = crtc->cursor->fb; + fb = plane->fb; if (fb) drm_framebuffer_get(fb); } @@ -810,7 +811,7 @@ static int drm_mode_cursor_universal(struct drm_crtc *crtc, src_h = fb->height << 16; } - ret = __setplane_internal(crtc->cursor, crtc, fb, + ret = __setplane_internal(plane, crtc, fb, crtc_x, crtc_y, crtc_w, crtc_h, 0, 0, src_w, src_h, ctx); @@ -931,6 +932,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev, { struct drm_mode_crtc_page_flip_target *page_flip = data; struct drm_crtc *crtc; + struct drm_plane *plane; struct drm_framebuffer *fb = NULL; struct drm_pending_vblank_event *e = NULL; u32 target_vblank = page_flip->sequence; @@ -959,6 +961,8 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev, if (!crtc) return -ENOENT; + plane = crtc->primary; + if (crtc->funcs->page_flip_target) { u32 current_vblank; int r; @@ -1003,11 +1007,11 @@ retry: ret = drm_modeset_lock(&crtc->mutex, &ctx); if (ret) goto out; - ret = drm_modeset_lock(&crtc->primary->mutex, &ctx); + ret = drm_modeset_lock(&plane->mutex, &ctx); if (ret) goto out; - if (crtc->primary->fb == NULL) { + if (plane->fb == NULL) { /* The framebuffer is currently unbound, presumably * due to a hotplug event, that userspace has not * yet discovered. @@ -1023,7 +1027,7 @@ retry: } if (crtc->state) { - const struct drm_plane_state *state = crtc->primary->state; + const struct drm_plane_state *state = plane->state; ret = drm_framebuffer_check_src_coords(state->src_x, state->src_y, @@ -1036,7 +1040,7 @@ retry: if (ret) goto out; - if (crtc->primary->fb->format != fb->format) { + if (plane->fb->format != fb->format) { DRM_DEBUG_KMS("Page flip is not allowed to change frame buffer format.\n"); ret = -EINVAL; goto out; @@ -1060,7 +1064,7 @@ retry: } } - crtc->primary->old_fb = crtc->primary->fb; + plane->old_fb = plane->fb; if (crtc->funcs->page_flip_target) ret = crtc->funcs->page_flip_target(crtc, fb, e, page_flip->flags, @@ -1073,9 +1077,9 @@ retry: if (page_flip->flags & DRM_MODE_PAGE_FLIP_EVENT) drm_event_cancel_free(dev, &e->base); /* Keep the old fb, don't unref it. */ - crtc->primary->old_fb = NULL; + plane->old_fb = NULL; } else { - crtc->primary->fb = fb; + plane->fb = fb; /* Unref only the old framebuffer. */ fb = NULL; } @@ -1083,9 +1087,9 @@ retry: out: if (fb) drm_framebuffer_put(fb); - if (crtc->primary->old_fb) - drm_framebuffer_put(crtc->primary->old_fb); - crtc->primary->old_fb = NULL; + if (plane->old_fb) + drm_framebuffer_put(plane->old_fb); + plane->old_fb = NULL; if (ret == -EDEADLK) { ret = drm_modeset_backoff(&ctx); -- cgit v1.1 From bf2d5eb902e29d6e7c25540b85242744591b703e Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 17:22:56 +0200 Subject: drm: Adjust whitespace for legibility MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Add a bit of whitespace here and there to make the code look a bit more structured. Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322152313.6561-7-ville.syrjala@linux.intel.com Acked-by: Daniel Vetter Reviewed-by: Maarten Lankhorst --- drivers/gpu/drm/drm_crtc.c | 4 +++- drivers/gpu/drm/drm_plane.c | 6 +++++- 2 files changed, 8 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index 8552ed4..537ffaa 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -434,13 +434,13 @@ int drm_mode_getcrtc(struct drm_device *dev, if (crtc->state->enable) { drm_mode_convert_to_umode(&crtc_resp->mode, &crtc->state->mode); crtc_resp->mode_valid = 1; - } else { crtc_resp->mode_valid = 0; } } else { crtc_resp->x = crtc->x; crtc_resp->y = crtc->y; + if (crtc->enabled) { drm_mode_convert_to_umode(&crtc_resp->mode, &crtc->mode); crtc_resp->mode_valid = 1; @@ -592,6 +592,7 @@ retry: ret = drm_modeset_lock_all_ctx(crtc->dev, &ctx); if (ret) goto out; + if (crtc_req->mode_valid) { /* If we have a mode we need a framebuffer. */ /* If we pass -1, set the mode with the currently bound fb */ @@ -601,6 +602,7 @@ retry: ret = -EINVAL; goto out; } + fb = plane->fb; /* Make refcounting symmetric with the lookup path. */ drm_framebuffer_get(fb); diff --git a/drivers/gpu/drm/drm_plane.c b/drivers/gpu/drm/drm_plane.c index 38e2a62..bedceca 100644 --- a/drivers/gpu/drm/drm_plane.c +++ b/drivers/gpu/drm/drm_plane.c @@ -785,6 +785,7 @@ static int drm_mode_cursor_universal(struct drm_crtc *crtc, DRM_DEBUG_KMS("failed to wrap cursor buffer in drm framebuffer\n"); return PTR_ERR(fb); } + fb->hot_x = req->hot_x; fb->hot_y = req->hot_y; } else { @@ -1035,7 +1036,8 @@ retry: state->src_h, fb); } else { - ret = drm_crtc_check_viewport(crtc, crtc->x, crtc->y, &crtc->mode, fb); + ret = drm_crtc_check_viewport(crtc, crtc->x, crtc->y, + &crtc->mode, fb); } if (ret) goto out; @@ -1052,10 +1054,12 @@ retry: ret = -ENOMEM; goto out; } + e->event.base.type = DRM_EVENT_FLIP_COMPLETE; e->event.base.length = sizeof(e->event); e->event.vbl.user_data = page_flip->user_data; e->event.vbl.crtc_id = crtc->base.id; + ret = drm_event_reserve_init(dev, file_priv, &e->base, &e->event.base); if (ret) { kfree(e); -- cgit v1.1 From 23a5e1fb349d80d8c0eb97644dea16d2d7cfac26 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 17:22:57 +0200 Subject: drm: Make the fb refcount handover less magic MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Instead of assigning the plane->fb pointer and clearing the fb pointer to hand over the reference, let's just do it by grabbing another referece for plane->fb and let fb keep its original one. Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322152313.6561-8-ville.syrjala@linux.intel.com Reviewed-by: Daniel Vetter Reviewed-by: Maarten Lankhorst --- drivers/gpu/drm/drm_plane.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/gpu/drm/drm_plane.c b/drivers/gpu/drm/drm_plane.c index bedceca..008f945 100644 --- a/drivers/gpu/drm/drm_plane.c +++ b/drivers/gpu/drm/drm_plane.c @@ -1084,8 +1084,7 @@ retry: plane->old_fb = NULL; } else { plane->fb = fb; - /* Unref only the old framebuffer. */ - fb = NULL; + drm_framebuffer_get(fb); } out: -- cgit v1.1 From a36c027db57b6a33970c5c830a0d143f0e98c248 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 17:22:58 +0200 Subject: drm: Use plane->state->fb over plane->fb MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Stop looking at plane->fb on atomic drivers. Use plane->state->fb instead. Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322152313.6561-9-ville.syrjala@linux.intel.com Reviewed-by: Daniel Vetter Reviewed-by: Maarten Lankhorst --- drivers/gpu/drm/drm_atomic_helper.c | 2 +- drivers/gpu/drm/drm_crtc.c | 11 +++++++++-- drivers/gpu/drm/drm_plane.c | 19 ++++++++++++++----- 3 files changed, 24 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index fe09d62..ee03c1e 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -2670,7 +2670,7 @@ int drm_atomic_helper_disable_plane(struct drm_plane *plane, goto fail; } - if (plane_state->crtc && (plane == plane->crtc->cursor)) + if (plane_state->crtc && plane_state->crtc->cursor == plane) plane_state->state->legacy_cursor_update = true; ret = __drm_atomic_helper_disable_plane(plane, plane_state); diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index 537ffaa..a231dd5 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -597,13 +597,20 @@ retry: /* If we have a mode we need a framebuffer. */ /* If we pass -1, set the mode with the currently bound fb */ if (crtc_req->fb_id == -1) { - if (!plane->fb) { + struct drm_framebuffer *old_fb; + + if (plane->state) + old_fb = plane->state->fb; + else + old_fb = plane->fb; + + if (!old_fb) { DRM_DEBUG_KMS("CRTC doesn't have current FB\n"); ret = -EINVAL; goto out; } - fb = plane->fb; + fb = old_fb; /* Make refcounting symmetric with the lookup path. */ drm_framebuffer_get(fb); } else { diff --git a/drivers/gpu/drm/drm_plane.c b/drivers/gpu/drm/drm_plane.c index 008f945..0350544 100644 --- a/drivers/gpu/drm/drm_plane.c +++ b/drivers/gpu/drm/drm_plane.c @@ -792,7 +792,11 @@ static int drm_mode_cursor_universal(struct drm_crtc *crtc, fb = NULL; } } else { - fb = plane->fb; + if (plane->state) + fb = plane->state->fb; + else + fb = plane->fb; + if (fb) drm_framebuffer_get(fb); } @@ -934,7 +938,7 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev, struct drm_mode_crtc_page_flip_target *page_flip = data; struct drm_crtc *crtc; struct drm_plane *plane; - struct drm_framebuffer *fb = NULL; + struct drm_framebuffer *fb = NULL, *old_fb; struct drm_pending_vblank_event *e = NULL; u32 target_vblank = page_flip->sequence; struct drm_modeset_acquire_ctx ctx; @@ -1012,7 +1016,12 @@ retry: if (ret) goto out; - if (plane->fb == NULL) { + if (plane->state) + old_fb = plane->state->fb; + else + old_fb = plane->fb; + + if (old_fb == NULL) { /* The framebuffer is currently unbound, presumably * due to a hotplug event, that userspace has not * yet discovered. @@ -1027,7 +1036,7 @@ retry: goto out; } - if (crtc->state) { + if (plane->state) { const struct drm_plane_state *state = plane->state; ret = drm_framebuffer_check_src_coords(state->src_x, @@ -1042,7 +1051,7 @@ retry: if (ret) goto out; - if (plane->fb->format != fb->format) { + if (old_fb->format != fb->format) { DRM_DEBUG_KMS("Page flip is not allowed to change frame buffer format.\n"); ret = -EINVAL; goto out; -- cgit v1.1 From 8bc20f6594a10d3e69b37fc223e43da3804673ef Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 17:22:59 +0200 Subject: drm/i915: Stop consulting plane->fb MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We want to get rid of plane->fb on atomic drivers. Stop looking at it. Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322152313.6561-10-ville.syrjala@linux.intel.com Reviewed-by: Maarten Lankhorst --- drivers/gpu/drm/i915/intel_display.c | 2 +- drivers/gpu/drm/i915/intel_fbdev.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 3b48fd2..54d4c36 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -2824,7 +2824,7 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, continue; if (intel_plane_ggtt_offset(state) == plane_config->base) { - fb = c->primary->fb; + fb = state->base.fb; drm_framebuffer_get(fb); goto valid_fb; } diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c index 6f12adc..89592ec 100644 --- a/drivers/gpu/drm/i915/intel_fbdev.c +++ b/drivers/gpu/drm/i915/intel_fbdev.c @@ -640,7 +640,7 @@ static bool intel_fbdev_init_bios(struct drm_device *dev, if (!crtc->state->active) continue; - WARN(!crtc->primary->fb, + WARN(!crtc->primary->state->fb, "re-used BIOS config but lost an fb on crtc %d\n", crtc->base.id); } -- cgit v1.1 From bf4a7a227511410cd6d6ca623e9ed17e5fc5f30b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 17:23:01 +0200 Subject: drm/sti: Stop consulting plane->fb MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We want to get rid of plane->fb on atomic drivers. Stop looking at it. Cc: Benjamin Gaignard Cc: Vincent Abriou Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322152313.6561-12-ville.syrjala@linux.intel.com Acked-by: Benjamin Gaignard Reviewed-by: Maarten Lankhorst --- drivers/gpu/drm/sti/sti_plane.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/sti/sti_plane.c b/drivers/gpu/drm/sti/sti_plane.c index b074609..b48cd86 100644 --- a/drivers/gpu/drm/sti/sti_plane.c +++ b/drivers/gpu/drm/sti/sti_plane.c @@ -40,6 +40,7 @@ void sti_plane_update_fps(struct sti_plane *plane, bool new_frame, bool new_field) { + struct drm_plane_state *state = plane->drm_plane.state; ktime_t now; struct sti_fps_info *fps; int fpks, fipks, ms_since_last, num_frames, num_fields; @@ -66,14 +67,14 @@ void sti_plane_update_fps(struct sti_plane *plane, fps->last_timestamp = now; fps->last_frame_counter = fps->curr_frame_counter; - if (plane->drm_plane.fb) { + if (state->fb) { fpks = (num_frames * 1000000) / ms_since_last; snprintf(plane->fps_info.fps_str, FPS_LENGTH, "%-8s %4dx%-4d %.4s @ %3d.%-3.3d fps (%s)", plane->drm_plane.name, - plane->drm_plane.fb->width, - plane->drm_plane.fb->height, - (char *)&plane->drm_plane.fb->format->format, + state->fb->width, + state->fb->height, + (char *)&state->fb->format->format, fpks / 1000, fpks % 1000, sti_plane_to_str(plane)); } -- cgit v1.1 From 2bb01c4cc0592ff69323077c093a67f45155d72f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 17:23:02 +0200 Subject: drm/vmwgfx: Stop consulting plane->fb MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We want to get rid of plane->fb on atomic drivers. Stop looking at it. Cc: VMware Graphics Cc: Sinclair Yeh Cc: Thomas Hellstrom Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322152313.6561-13-ville.syrjala@linux.intel.com Reviewed-by: Thomas Hellstrom Reviewed-by: Maarten Lankhorst --- drivers/gpu/drm/vmwgfx/vmwgfx_kms.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c index 34ecc27..9fdb3ec 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c @@ -385,9 +385,9 @@ vmw_du_cursor_plane_atomic_update(struct drm_plane *plane, hotspot_x = du->hotspot_x; hotspot_y = du->hotspot_y; - if (plane->fb) { - hotspot_x += plane->fb->hot_x; - hotspot_y += plane->fb->hot_y; + if (plane->state->fb) { + hotspot_x += plane->state->fb->hot_x; + hotspot_y += plane->state->fb->hot_y; } du->cursor_surface = vps->surf; -- cgit v1.1 From 6f6887dad5405bc87fc26eddca1fad64d04360fc Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Mon, 26 Mar 2018 15:14:42 +0300 Subject: drm/zte: Stop consulting plane->crtc MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We want to get rid of plane->crtc on atomic drivers. Stop looking at it. v2: Use old_state->crtc (Maarten) v3: s/fb/crtc/ in commit message to actually match the patch (Shawn) Cc: Shawn Guo Cc: Maarten Lankhorst Signed-off-by: Ville Syrjälä Acked-by: Shawn Guo Link: https://patchwork.freedesktop.org/patch/msgid/20180326121442.32009-1-ville.syrjala@linux.intel.com Reviewed-by: Maarten Lankhorst --- drivers/gpu/drm/zte/zx_plane.c | 2 +- drivers/gpu/drm/zte/zx_vou.c | 5 +++-- drivers/gpu/drm/zte/zx_vou.h | 3 ++- 3 files changed, 6 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/zte/zx_plane.c b/drivers/gpu/drm/zte/zx_plane.c index 94545ad..d1931f5 100644 --- a/drivers/gpu/drm/zte/zx_plane.c +++ b/drivers/gpu/drm/zte/zx_plane.c @@ -268,7 +268,7 @@ static void zx_plane_atomic_disable(struct drm_plane *plane, struct zx_plane *zplane = to_zx_plane(plane); void __iomem *hbsc = zplane->hbsc; - zx_vou_layer_disable(plane); + zx_vou_layer_disable(plane, old_state); /* Disable HBSC block */ zx_writel_mask(hbsc + HBSC_CTRL0, HBSC_CTRL_EN, 0); diff --git a/drivers/gpu/drm/zte/zx_vou.c b/drivers/gpu/drm/zte/zx_vou.c index 7491813..442311d 100644 --- a/drivers/gpu/drm/zte/zx_vou.c +++ b/drivers/gpu/drm/zte/zx_vou.c @@ -627,9 +627,10 @@ void zx_vou_layer_enable(struct drm_plane *plane) zx_writel_mask(vou->osd + OSD_CTRL0, bits->enable, bits->enable); } -void zx_vou_layer_disable(struct drm_plane *plane) +void zx_vou_layer_disable(struct drm_plane *plane, + struct drm_plane_state *old_state) { - struct zx_crtc *zcrtc = to_zx_crtc(plane->crtc); + struct zx_crtc *zcrtc = to_zx_crtc(old_state->crtc); struct zx_vou_hw *vou = zcrtc->vou; struct zx_plane *zplane = to_zx_plane(plane); const struct vou_layer_bits *bits = zplane->bits; diff --git a/drivers/gpu/drm/zte/zx_vou.h b/drivers/gpu/drm/zte/zx_vou.h index 97d72bf..5b7f84f 100644 --- a/drivers/gpu/drm/zte/zx_vou.h +++ b/drivers/gpu/drm/zte/zx_vou.h @@ -62,6 +62,7 @@ void zx_vou_config_dividers(struct drm_crtc *crtc, struct vou_div_config *configs, int num); void zx_vou_layer_enable(struct drm_plane *plane); -void zx_vou_layer_disable(struct drm_plane *plane); +void zx_vou_layer_disable(struct drm_plane *plane, + struct drm_plane_state *old_state); #endif /* __ZX_VOU_H__ */ -- cgit v1.1 From be90cc318b4c8d5c3999eb25783d8055e6a67855 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 17:23:12 +0200 Subject: drm/i915: Restore planes after load detection MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Actually turn the planes back on after were done with the load detection. Fixes: 20bdc112bbe4 ("drm/i915: Disable all planes for load detection, v2.") Cc: Maarten Lankhorst Cc: Daniel Vetter Reviewed-by: Maarten Lankhorst Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322152313.6561-23-ville.syrjala@linux.intel.com --- drivers/gpu/drm/i915/intel_display.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 54d4c36..182f9bf 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -9974,6 +9974,8 @@ found: ret = PTR_ERR_OR_ZERO(drm_atomic_get_connector_state(restore_state, connector)); if (!ret) ret = PTR_ERR_OR_ZERO(drm_atomic_get_crtc_state(restore_state, crtc)); + if (!ret) + ret = drm_atomic_add_affected_planes(restore_state, crtc); if (ret) { DRM_DEBUG_KMS("Failed to create a copy of old state to restore: %i\n", ret); goto fail; -- cgit v1.1 From 4165791d29f64e01860a064f3c649447dbac41c3 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 22 Mar 2018 19:41:35 +0200 Subject: drm/i915: Make force_load_detect effective even w/ DMI quirks/hotplug MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When doing forced load detection testing we should totally ignore any hotplug status for the connector. This is mostly relevant for machines where we already ignore the hotplug status based on the DMI quirks. On other machines we would currently skip the force load detection tests on account of the connector already being connected. v2: Drop the other force_load_detect check since it's useless now (Maarten) Cc: Maarten Lankhorst Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180322174135.5982-1-ville.syrjala@linux.intel.com Reviewed-by: Maarten Lankhorst --- drivers/gpu/drm/i915/intel_crt.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c index c0a8805..de0e223 100644 --- a/drivers/gpu/drm/i915/intel_crt.c +++ b/drivers/gpu/drm/i915/intel_crt.c @@ -748,6 +748,11 @@ intel_crt_detect(struct drm_connector *connector, connector->base.id, connector->name, force); + if (i915_modparams.load_detect_test) { + intel_display_power_get(dev_priv, intel_encoder->power_domain); + goto load_detect; + } + /* Skip machines without VGA that falsely report hotplug events */ if (dmi_check_system(intel_spurious_crt_detect)) return connector_status_disconnected; @@ -776,11 +781,12 @@ intel_crt_detect(struct drm_connector *connector, * broken monitor (without edid) to work behind a broken kvm (that fails * to have the right resistors for HP detection) needs to fix this up. * For now just bail out. */ - if (I915_HAS_HOTPLUG(dev_priv) && !i915_modparams.load_detect_test) { + if (I915_HAS_HOTPLUG(dev_priv)) { status = connector_status_disconnected; goto out; } +load_detect: if (!force) { status = connector->status; goto out; -- cgit v1.1 From 5807e1c21dbd56c87f0e86ae6fe49ec745660c0d Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 29 Mar 2018 23:45:18 +0100 Subject: drm/i915: Avoid sleeping inside per-engine reset MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Only sleep and repeat when asked for a full device reset (ALL_ENGINES) and avoid using sleeping waits when asked for a per-engine reset. The goal is to be able to use a per-engine reset from hardirq/softirq/timer context. A consequence is that our individual wait timeouts are a thousand times shorter, on the order of a hundred microseconds rather than hundreds of millisecond. This may make hitting the timeouts more common, but hopefully the fallover to the full-device reset will be sufficient to pick up the pieces. Note, that the sleeps inside older gen (pre-gen8) have been left as they are only used in full device reset mode. Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Michał Winiarski CC: Michel Thierry Cc: Jeff McGee Cc: Tvrtko Ursulin Reviewed-by: Michel Thierry Link: https://patchwork.freedesktop.org/patch/msgid/20180329224519.13598-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/intel_uncore.c | 51 ++++++++++++++++++++++++------------- 1 file changed, 34 insertions(+), 17 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c index f37ecfc..a0d7e0c 100644 --- a/drivers/gpu/drm/i915/intel_uncore.c +++ b/drivers/gpu/drm/i915/intel_uncore.c @@ -1702,11 +1702,10 @@ static void gen3_stop_engine(struct intel_engine_cs *engine) const i915_reg_t mode = RING_MI_MODE(base); I915_WRITE_FW(mode, _MASKED_BIT_ENABLE(STOP_RING)); - if (intel_wait_for_register_fw(dev_priv, - mode, - MODE_IDLE, - MODE_IDLE, - 500)) + if (__intel_wait_for_register_fw(dev_priv, + mode, MODE_IDLE, MODE_IDLE, + 500, 0, + NULL)) DRM_DEBUG_DRIVER("%s: timed out on STOP_RING\n", engine->name); @@ -1860,9 +1859,10 @@ static int gen6_hw_domain_reset(struct drm_i915_private *dev_priv, __raw_i915_write32(dev_priv, GEN6_GDRST, hw_domain_mask); /* Wait for the device to ack the reset requests */ - err = intel_wait_for_register_fw(dev_priv, - GEN6_GDRST, hw_domain_mask, 0, - 500); + err = __intel_wait_for_register_fw(dev_priv, + GEN6_GDRST, hw_domain_mask, 0, + 500, 0, + NULL); if (err) DRM_DEBUG_DRIVER("Wait for 0x%08x engines reset failed\n", hw_domain_mask); @@ -2027,11 +2027,12 @@ static int gen8_reset_engine_start(struct intel_engine_cs *engine) I915_WRITE_FW(RING_RESET_CTL(engine->mmio_base), _MASKED_BIT_ENABLE(RESET_CTL_REQUEST_RESET)); - ret = intel_wait_for_register_fw(dev_priv, - RING_RESET_CTL(engine->mmio_base), - RESET_CTL_READY_TO_RESET, - RESET_CTL_READY_TO_RESET, - 700); + ret = __intel_wait_for_register_fw(dev_priv, + RING_RESET_CTL(engine->mmio_base), + RESET_CTL_READY_TO_RESET, + RESET_CTL_READY_TO_RESET, + 700, 0, + NULL); if (ret) DRM_ERROR("%s: reset request timeout\n", engine->name); @@ -2094,15 +2095,31 @@ int intel_gpu_reset(struct drm_i915_private *dev_priv, unsigned engine_mask) int retry; int ret; - might_sleep(); + /* + * We want to perform per-engine reset from atomic context (e.g. + * softirq), which imposes the constraint that we cannot sleep. + * However, experience suggests that spending a bit of time waiting + * for a reset helps in various cases, so for a full-device reset + * we apply the opposite rule and wait if we want to. As we should + * always follow up a failed per-engine reset with a full device reset, + * being a little faster, stricter and more error prone for the + * atomic case seems an acceptable compromise. + * + * Unfortunately this leads to a bimodal routine, when the goal was + * to have a single reset function that worked for resetting any + * number of engines simultaneously. + */ + might_sleep_if(engine_mask == ALL_ENGINES); - /* If the power well sleeps during the reset, the reset + /* + * If the power well sleeps during the reset, the reset * request may be dropped and never completes (causing -EIO). */ intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); for (retry = 0; retry < 3; retry++) { - /* We stop engines, otherwise we might get failed reset and a + /* + * We stop engines, otherwise we might get failed reset and a * dead gpu (on elk). Also as modern gpu as kbl can suffer * from system hang if batchbuffer is progressing when * the reset is issued, regardless of READY_TO_RESET ack. @@ -2120,7 +2137,7 @@ int intel_gpu_reset(struct drm_i915_private *dev_priv, unsigned engine_mask) GEM_TRACE("engine_mask=%x\n", engine_mask); ret = reset(dev_priv, engine_mask); } - if (ret != -ETIMEDOUT) + if (ret != -ETIMEDOUT || engine_mask != ALL_ENGINES) break; cond_resched(); -- cgit v1.1 From 3df82dd43be4b6efde20f819d5829c8ed5e95476 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 29 Mar 2018 23:45:19 +0100 Subject: drm/i915: Only warn for might_sleep() before a slow wait_for_register As intel_wait_for_register_fw() may use, and if successful only use, a busy-wait loop, the might_sleep() warning is a little over-zealous. Restrict it to a might_sleep_if() a slow timeout is specified (and so the caller authorises use of a usleep). Signed-off-by: Chris Wilson Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180329224519.13598-2-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/intel_uncore.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c index a0d7e0c..e7540bb 100644 --- a/drivers/gpu/drm/i915/intel_uncore.c +++ b/drivers/gpu/drm/i915/intel_uncore.c @@ -1996,7 +1996,7 @@ int __intel_wait_for_register(struct drm_i915_private *dev_priv, u32 reg_value; int ret; - might_sleep(); + might_sleep_if(slow_timeout_ms); spin_lock_irq(&dev_priv->uncore.lock); intel_uncore_forcewake_get__locked(dev_priv, fw); @@ -2008,7 +2008,7 @@ int __intel_wait_for_register(struct drm_i915_private *dev_priv, intel_uncore_forcewake_put__locked(dev_priv, fw); spin_unlock_irq(&dev_priv->uncore.lock); - if (ret) + if (ret && slow_timeout_ms) ret = __wait_for(reg_value = I915_READ_NOTRACE(reg), (reg_value & mask) == value, slow_timeout_ms * 1000, 10, 1000); -- cgit v1.1 From 487da6172f5678496699bec685797dc816f6a131 Mon Sep 17 00:00:00 2001 From: Daniel Stone Date: Fri, 23 Mar 2018 13:45:51 +0000 Subject: drm: Reshuffle getfb error returns MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Make it a little more clear what's going on inside of getfb, and also make it easier to add alternate paths to get a handle in future. Signed-off-by: Daniel Stone Reviewed-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20180323134553.15993-3-daniels@collabora.com --- drivers/gpu/drm/drm_framebuffer.c | 34 ++++++++++++++++++---------------- 1 file changed, 18 insertions(+), 16 deletions(-) diff --git a/drivers/gpu/drm/drm_framebuffer.c b/drivers/gpu/drm/drm_framebuffer.c index ad67203..8c4d32a 100644 --- a/drivers/gpu/drm/drm_framebuffer.c +++ b/drivers/gpu/drm/drm_framebuffer.c @@ -468,29 +468,31 @@ int drm_mode_getfb(struct drm_device *dev, goto out; } + if (!fb->funcs->create_handle) { + ret = -ENODEV; + goto out; + } + r->height = fb->height; r->width = fb->width; r->depth = fb->format->depth; r->bpp = fb->format->cpp[0] * 8; r->pitch = fb->pitches[0]; - if (fb->funcs->create_handle) { - if (drm_is_current_master(file_priv) || capable(CAP_SYS_ADMIN) || - drm_is_control_client(file_priv)) { - ret = fb->funcs->create_handle(fb, file_priv, - &r->handle); - } else { - /* GET_FB() is an unprivileged ioctl so we must not - * return a buffer-handle to non-master processes! For - * backwards-compatibility reasons, we cannot make - * GET_FB() privileged, so just return an invalid handle - * for non-masters. */ - r->handle = 0; - ret = 0; - } - } else { - ret = -ENODEV; + + /* GET_FB() is an unprivileged ioctl so we must not return a + * buffer-handle to non-master processes! For + * backwards-compatibility reasons, we cannot make GET_FB() privileged, + * so just return an invalid handle for non-masters. + */ + if (!drm_is_current_master(file_priv) && !capable(CAP_SYS_ADMIN) && + !drm_is_control_client(file_priv)) { + r->handle = 0; + ret = 0; + goto out; } + ret = fb->funcs->create_handle(fb, file_priv, &r->handle); + out: drm_framebuffer_put(fb); -- cgit v1.1 From 4f212e40468650e220c1770876c7f25b8e0c1ff5 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jos=C3=A9=20Roberto=20de=20Souza?= Date: Wed, 28 Mar 2018 15:30:37 -0700 Subject: drm: Add DP PSR2 sink enable bit MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit To comply with eDP1.4a this bit should be set when enabling PSR2. Signed-off-by: José Roberto de Souza Reviewed-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180328223046.16125-1-jose.souza@intel.com --- include/drm/drm_dp_helper.h | 1 + 1 file changed, 1 insertion(+) diff --git a/include/drm/drm_dp_helper.h b/include/drm/drm_dp_helper.h index 4de97e9..a627145 100644 --- a/include/drm/drm_dp_helper.h +++ b/include/drm/drm_dp_helper.h @@ -477,6 +477,7 @@ # define DP_PSR_FRAME_CAPTURE (1 << 3) # define DP_PSR_SELECTIVE_UPDATE (1 << 4) # define DP_PSR_IRQ_HPD_WITH_CRC_ERRORS (1 << 5) +# define DP_PSR_ENABLE_PSR2 (1 << 6) /* eDP 1.4a */ #define DP_ADAPTER_CTRL 0x1a0 # define DP_ADAPTER_CTRL_FORCE_LOAD_SENSE (1 << 0) -- cgit v1.1 From fe36948afb08e361d12b2878348778aeaba74134 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jos=C3=A9=20Roberto=20de=20Souza?= Date: Wed, 28 Mar 2018 15:30:38 -0700 Subject: drm: Add DP last received PSR SDP VSC register and bits MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This is a register to help debug what is in the last SDP VSC packet revived by sink. Signed-off-by: José Roberto de Souza Reviewed-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180328223046.16125-2-jose.souza@intel.com --- include/drm/drm_dp_helper.h | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/include/drm/drm_dp_helper.h b/include/drm/drm_dp_helper.h index a627145..c6853f0 100644 --- a/include/drm/drm_dp_helper.h +++ b/include/drm/drm_dp_helper.h @@ -794,6 +794,15 @@ # define DP_LAST_ACTUAL_SYNCHRONIZATION_LATENCY_MASK (0xf << 4) # define DP_LAST_ACTUAL_SYNCHRONIZATION_LATENCY_SHIFT 4 +#define DP_LAST_RECEIVED_PSR_SDP 0x200a /* eDP 1.2 */ +# define DP_PSR_STATE_BIT (1 << 0) /* eDP 1.2 */ +# define DP_UPDATE_RFB_BIT (1 << 1) /* eDP 1.2 */ +# define DP_CRC_VALID_BIT (1 << 2) /* eDP 1.2 */ +# define DP_SU_VALID (1 << 3) /* eDP 1.4 */ +# define DP_FIRST_SCAN_LINE_SU_REGION (1 << 4) /* eDP 1.4 */ +# define DP_LAST_SCAN_LINE_SU_REGION (1 << 5) /* eDP 1.4 */ +# define DP_Y_COORDINATE_VALID (1 << 6) /* eDP 1.4a */ + #define DP_RECEIVER_ALPM_STATUS 0x200b /* eDP 1.4 */ # define DP_ALPM_LOCK_TIMEOUT_ERROR (1 << 0) -- cgit v1.1 From 6ce9b78a7388c37b33293666aa6fc61c177046e6 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jos=C3=A9=20Roberto=20de=20Souza?= Date: Wed, 28 Mar 2018 15:30:39 -0700 Subject: drm/i915/psr: Nuke aux frame sync MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit eDP spec states that aux frame is required to do PSR2 selective update but i915 don't fully implement it. It sends the aux frame sync messages but the value is always zero as the GTC is not enabled in driver. Through tests was findout that pannels can do selective update when the y-coordinate is also included in SDP, that is why it is required to run PSR2 in i915. A dummy value is not useful at all to sink, so removing everything related to aux frame sync, if GTC is enabled we can bring this back. Cc: Vathsala Nagaraju Acked-by: Rodrigo Vivi Reviewed-by: Dhinakaran Pandiyan Signed-off-by: José Roberto de Souza Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180328223046.16125-3-jose.souza@intel.com --- drivers/gpu/drm/i915/i915_drv.h | 1 - drivers/gpu/drm/i915/intel_psr.c | 24 +----------------------- 2 files changed, 1 insertion(+), 24 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 800230b..fade902 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -603,7 +603,6 @@ struct i915_psr { struct delayed_work work; unsigned busy_frontbuffer_bits; bool psr2_support; - bool aux_frame_sync; bool link_standby; bool y_cord_support; bool colorimetry_support; diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index b8e083e..c0a6f63 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -137,16 +137,9 @@ void intel_psr_init_dpcd(struct intel_dp *intel_dp) if (INTEL_GEN(dev_priv) >= 9 && (intel_dp->psr_dpcd[0] & DP_PSR2_IS_SUPPORTED)) { - uint8_t frame_sync_cap; dev_priv->psr.sink_support = true; - if (drm_dp_dpcd_readb(&intel_dp->aux, - DP_SINK_DEVICE_AUX_FRAME_SYNC_CAP, - &frame_sync_cap) != 1) - frame_sync_cap = 0; - dev_priv->psr.aux_frame_sync = frame_sync_cap & DP_AUX_FRAME_SYNC_CAP; - /* PSR2 needs frame sync as well */ - dev_priv->psr.psr2_support = dev_priv->psr.aux_frame_sync; + dev_priv->psr.psr2_support = true; DRM_DEBUG_KMS("PSR2 %s on sink", dev_priv->psr.psr2_support ? "supported" : "not supported"); @@ -268,12 +261,6 @@ static void hsw_psr_enable_sink(struct intel_dp *intel_dp) struct drm_device *dev = dig_port->base.base.dev; struct drm_i915_private *dev_priv = to_i915(dev); - - /* Enable AUX frame sync at sink */ - if (dev_priv->psr.aux_frame_sync) - drm_dp_dpcd_writeb(&intel_dp->aux, - DP_SINK_DEVICE_AUX_FRAME_SYNC_CONF, - DP_AUX_FRAME_SYNC_ENABLE); /* Enable ALPM at sink for psr2 */ if (dev_priv->psr.psr2_support && dev_priv->psr.alpm) drm_dp_dpcd_writeb(&intel_dp->aux, @@ -712,11 +699,6 @@ static void hsw_psr_disable(struct intel_dp *intel_dp, i915_reg_t psr_status; u32 psr_status_mask; - if (dev_priv->psr.aux_frame_sync) - drm_dp_dpcd_writeb(&intel_dp->aux, - DP_SINK_DEVICE_AUX_FRAME_SYNC_CONF, - 0); - if (dev_priv->psr.psr2_support) { psr_status = EDP_PSR2_STATUS; psr_status_mask = EDP_PSR2_STATUS_STATE_MASK; @@ -860,10 +842,6 @@ static void intel_psr_exit(struct drm_i915_private *dev_priv) return; if (HAS_DDI(dev_priv)) { - if (dev_priv->psr.aux_frame_sync) - drm_dp_dpcd_writeb(&intel_dp->aux, - DP_SINK_DEVICE_AUX_FRAME_SYNC_CONF, - 0); if (dev_priv->psr.psr2_support) { val = I915_READ(EDP_PSR2_CTL); WARN_ON(!(val & EDP_PSR2_ENABLE)); -- cgit v1.1 From aee3bac0a3a89ea12644533142ba69eebb602e4c Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jos=C3=A9=20Roberto=20de=20Souza?= Date: Wed, 28 Mar 2018 15:30:40 -0700 Subject: drm/i915/psr: Tie PSR2 support to Y coordinate requirement MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Although i915 don't implement aux sync frame through tests was findout that pannels can do selective update when the y-coordinate is also included in SDP, that is why it is required to run PSR2 in i915. So moving to only one place the sink requirements that the actual driver needs to enable PSR2. Also intel_psr2_config_valid() is called every time the crtc config is computed, wasting some time every time it was checking for Y coordinate requirement. This allow us to nuke y_cord_support and some of VSC setup code that was handling a scenario that would never happen(PSR2 without Y coordinate). Also here renaming intel_dp_get_y_cord_status() to intel_dp_get_y_coord_required() as it more accurate to the name and function of bit according to eDP spec. Cc: Rodrigo Vivi Reviewed-by: Dhinakaran Pandiyan Signed-off-by: José Roberto de Souza Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180328223046.16125-4-jose.souza@intel.com --- drivers/gpu/drm/i915/i915_drv.h | 1 - drivers/gpu/drm/i915/intel_psr.c | 46 +++++++++++++++++----------------------- 2 files changed, 19 insertions(+), 28 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index fade902..92cf6f4e 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -604,7 +604,6 @@ struct i915_psr { unsigned busy_frontbuffer_bits; bool psr2_support; bool link_standby; - bool y_cord_support; bool colorimetry_support; bool alpm; bool has_hw_tracking; diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index c0a6f63..fb2d0fe 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -93,7 +93,7 @@ static void psr_aux_io_power_put(struct intel_dp *intel_dp) intel_display_power_put(dev_priv, psr_aux_domain(intel_dp)); } -static bool intel_dp_get_y_cord_status(struct intel_dp *intel_dp) +static bool intel_dp_get_y_coord_required(struct intel_dp *intel_dp) { uint8_t psr_caps = 0; @@ -130,22 +130,29 @@ void intel_psr_init_dpcd(struct intel_dp *intel_dp) drm_dp_dpcd_read(&intel_dp->aux, DP_PSR_SUPPORT, intel_dp->psr_dpcd, sizeof(intel_dp->psr_dpcd)); - if (intel_dp->psr_dpcd[0] & DP_PSR_IS_SUPPORTED) { + if (intel_dp->psr_dpcd[0]) { dev_priv->psr.sink_support = true; DRM_DEBUG_KMS("Detected EDP PSR Panel.\n"); } if (INTEL_GEN(dev_priv) >= 9 && - (intel_dp->psr_dpcd[0] & DP_PSR2_IS_SUPPORTED)) { - - dev_priv->psr.sink_support = true; - dev_priv->psr.psr2_support = true; + (intel_dp->psr_dpcd[0] == DP_PSR2_WITH_Y_COORD_IS_SUPPORTED)) { + /* + * All panels that supports PSR version 03h (PSR2 + + * Y-coordinate) can handle Y-coordinates in VSC but we are + * only sure that it is going to be used when required by the + * panel. This way panel is capable to do selective update + * without a aux frame sync. + * + * To support PSR version 02h and PSR version 03h without + * Y-coordinate requirement panels we would need to enable + * GTC first. + */ + dev_priv->psr.psr2_support = intel_dp_get_y_coord_required(intel_dp); DRM_DEBUG_KMS("PSR2 %s on sink", dev_priv->psr.psr2_support ? "supported" : "not supported"); if (dev_priv->psr.psr2_support) { - dev_priv->psr.y_cord_support = - intel_dp_get_y_cord_status(intel_dp); dev_priv->psr.colorimetry_support = intel_dp_get_colorimetry_status(intel_dp); dev_priv->psr.alpm = @@ -191,16 +198,12 @@ static void hsw_psr_setup_vsc(struct intel_dp *intel_dp, memset(&psr_vsc, 0, sizeof(psr_vsc)); psr_vsc.sdp_header.HB0 = 0; psr_vsc.sdp_header.HB1 = 0x7; - if (dev_priv->psr.colorimetry_support && - dev_priv->psr.y_cord_support) { + if (dev_priv->psr.colorimetry_support) { psr_vsc.sdp_header.HB2 = 0x5; psr_vsc.sdp_header.HB3 = 0x13; - } else if (dev_priv->psr.y_cord_support) { + } else { psr_vsc.sdp_header.HB2 = 0x4; psr_vsc.sdp_header.HB3 = 0xe; - } else { - psr_vsc.sdp_header.HB2 = 0x3; - psr_vsc.sdp_header.HB3 = 0xc; } } else { /* Prepare VSC packet as per EDP 1.3 spec, Table 3.10 */ @@ -457,15 +460,6 @@ static bool intel_psr2_config_valid(struct intel_dp *intel_dp, return false; } - /* - * FIXME:enable psr2 only for y-cordinate psr2 panels - * After gtc implementation , remove this restriction. - */ - if (!dev_priv->psr.y_cord_support) { - DRM_DEBUG_KMS("PSR2 not enabled, panel does not support Y coordinate\n"); - return false; - } - return true; } @@ -565,7 +559,6 @@ static void hsw_psr_enable_source(struct intel_dp *intel_dp, struct drm_device *dev = dig_port->base.base.dev; struct drm_i915_private *dev_priv = to_i915(dev); enum transcoder cpu_transcoder = crtc_state->cpu_transcoder; - u32 chicken; psr_aux_io_power_get(intel_dp); @@ -576,9 +569,8 @@ static void hsw_psr_enable_source(struct intel_dp *intel_dp, hsw_psr_setup_aux(intel_dp); if (dev_priv->psr.psr2_support) { - chicken = PSR2_VSC_ENABLE_PROG_HEADER; - if (dev_priv->psr.y_cord_support) - chicken |= PSR2_ADD_VERTICAL_LINE_COUNT; + u32 chicken = PSR2_VSC_ENABLE_PROG_HEADER + | PSR2_ADD_VERTICAL_LINE_COUNT; I915_WRITE(CHICKEN_TRANS(cpu_transcoder), chicken); I915_WRITE(EDP_PSR_DEBUG, -- cgit v1.1 From 5e87325f5c57ba59cc6908bf38efd40146d7ad9c Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jos=C3=A9=20Roberto=20de=20Souza?= Date: Wed, 28 Mar 2018 15:30:41 -0700 Subject: drm/i915/psr/cnl: Enable Y-coordinate support in source MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit For Geminilake and Cannonlake+ the Y-coordinate support must be enabled in PSR2_CTL too. Spec: 7713 and 7720 Cc: Dhinakaran Pandiyan Reviewed-by: Rodrigo Vivi Signed-off-by: José Roberto de Souza Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180328223046.16125-5-jose.souza@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 3 +++ drivers/gpu/drm/i915/intel_psr.c | 16 ++++++++++++---- 2 files changed, 15 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index b926520..6566f6b 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -4058,6 +4058,8 @@ enum { #define EDP_PSR2_CTL _MMIO(0x6f900) #define EDP_PSR2_ENABLE (1<<31) #define EDP_SU_TRACK_ENABLE (1<<30) +#define EDP_Y_COORDINATE_VALID (1<<26) /* GLK and CNL+ */ +#define EDP_Y_COORDINATE_ENABLE (1<<25) /* GLK and CNL+ */ #define EDP_MAX_SU_DISABLE_TIME(t) ((t)<<20) #define EDP_MAX_SU_DISABLE_TIME_MASK (0x1f<<20) #define EDP_PSR2_TP2_TIME_500 (0<<8) @@ -7042,6 +7044,7 @@ enum { #define CHICKEN_TRANS_A 0x420c0 #define CHICKEN_TRANS_B 0x420c4 #define CHICKEN_TRANS(trans) _MMIO_TRANS(trans, CHICKEN_TRANS_A, CHICKEN_TRANS_B) +#define VSC_DATA_SEL_SOFTWARE_CONTROL (1<<25) /* GLK and CNL+ */ #define DDI_TRAINING_OVERRIDE_ENABLE (1<<19) #define DDI_TRAINING_OVERRIDE_VALUE (1<<18) #define DDIE_TRAINING_OVERRIDE_ENABLE (1<<17) /* CHICKEN_TRANS_A only */ diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index fb2d0fe..84e1f8b 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -386,8 +386,10 @@ static void hsw_activate_psr2(struct intel_dp *intel_dp) /* FIXME: selective update is probably totally broken because it doesn't * mesh at all with our frontbuffer tracking. And the hw alone isn't * good enough. */ - val |= EDP_PSR2_ENABLE | - EDP_SU_TRACK_ENABLE; + val |= EDP_PSR2_ENABLE | EDP_SU_TRACK_ENABLE; + if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv)) { + val |= EDP_Y_COORDINATE_VALID | EDP_Y_COORDINATE_ENABLE; + } if (drm_dp_dpcd_readb(&intel_dp->aux, DP_SYNCHRONIZATION_LATENCY_IN_SINK, @@ -569,8 +571,14 @@ static void hsw_psr_enable_source(struct intel_dp *intel_dp, hsw_psr_setup_aux(intel_dp); if (dev_priv->psr.psr2_support) { - u32 chicken = PSR2_VSC_ENABLE_PROG_HEADER - | PSR2_ADD_VERTICAL_LINE_COUNT; + u32 chicken = I915_READ(CHICKEN_TRANS(cpu_transcoder)); + + if (INTEL_GEN(dev_priv) == 9 && !IS_GEMINILAKE(dev_priv)) + chicken |= (PSR2_VSC_ENABLE_PROG_HEADER + | PSR2_ADD_VERTICAL_LINE_COUNT); + + else + chicken &= ~VSC_DATA_SEL_SOFTWARE_CONTROL; I915_WRITE(CHICKEN_TRANS(cpu_transcoder), chicken); I915_WRITE(EDP_PSR_DEBUG, -- cgit v1.1 From 95f28d2ec75ac388a8cc988e2e5496ce4adef4e4 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jos=C3=A9=20Roberto=20de=20Souza?= Date: Wed, 28 Mar 2018 15:30:42 -0700 Subject: drm/i915/psr: Do not override PSR2 sink support MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Sink can support our PSR2 requirements but userspace can request a resolution that PSR2 hardware do not support, in this case it was overwritten the PSR2 sink support. Adding another flag here, this way if requested resolution changed to a value that PSR2 hardware can handle, PSR2 can be enabled. Cc: Dhinakaran Pandiyan Reviewed-by: Rodrigo Vivi Signed-off-by: José Roberto de Souza Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180328223046.16125-6-jose.souza@intel.com --- drivers/gpu/drm/i915/i915_debugfs.c | 4 ++-- drivers/gpu/drm/i915/i915_drv.h | 3 ++- drivers/gpu/drm/i915/intel_psr.c | 33 +++++++++++++++++---------------- 3 files changed, 21 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index ff90577..1dba2c4 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -2630,7 +2630,7 @@ static int i915_edp_psr_status(struct seq_file *m, void *data) yesno(work_busy(&dev_priv->psr.work.work))); if (HAS_DDI(dev_priv)) { - if (dev_priv->psr.psr2_support) + if (dev_priv->psr.psr2_enabled) enabled = I915_READ(EDP_PSR2_CTL) & EDP_PSR2_ENABLE; else enabled = I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE; @@ -2678,7 +2678,7 @@ static int i915_edp_psr_status(struct seq_file *m, void *data) seq_printf(m, "Performance_Counter: %u\n", psrperf); } - if (dev_priv->psr.psr2_support) { + if (dev_priv->psr.psr2_enabled) { u32 psr2 = I915_READ(EDP_PSR2_STATUS); seq_printf(m, "EDP_PSR2_STATUS: %x [%s]\n", diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 92cf6f4e..46cae09 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -602,11 +602,12 @@ struct i915_psr { bool active; struct delayed_work work; unsigned busy_frontbuffer_bits; - bool psr2_support; + bool sink_psr2_support; bool link_standby; bool colorimetry_support; bool alpm; bool has_hw_tracking; + bool psr2_enabled; void (*enable_source)(struct intel_dp *, const struct intel_crtc_state *); diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index 84e1f8b..5efddd9 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -148,11 +148,12 @@ void intel_psr_init_dpcd(struct intel_dp *intel_dp) * Y-coordinate requirement panels we would need to enable * GTC first. */ - dev_priv->psr.psr2_support = intel_dp_get_y_coord_required(intel_dp); - DRM_DEBUG_KMS("PSR2 %s on sink", - dev_priv->psr.psr2_support ? "supported" : "not supported"); + dev_priv->psr.sink_psr2_support = + intel_dp_get_y_coord_required(intel_dp); + DRM_DEBUG_KMS("PSR2 %s on sink", dev_priv->psr.sink_psr2_support + ? "supported" : "not supported"); - if (dev_priv->psr.psr2_support) { + if (dev_priv->psr.sink_psr2_support) { dev_priv->psr.colorimetry_support = intel_dp_get_colorimetry_status(intel_dp); dev_priv->psr.alpm = @@ -193,7 +194,7 @@ static void hsw_psr_setup_vsc(struct intel_dp *intel_dp, struct drm_i915_private *dev_priv = to_i915(intel_dig_port->base.base.dev); struct edp_vsc_psr psr_vsc; - if (dev_priv->psr.psr2_support) { + if (dev_priv->psr.psr2_enabled) { /* Prepare VSC Header for SU as per EDP 1.4 spec, Table 6.11 */ memset(&psr_vsc, 0, sizeof(psr_vsc)); psr_vsc.sdp_header.HB0 = 0; @@ -265,7 +266,7 @@ static void hsw_psr_enable_sink(struct intel_dp *intel_dp) struct drm_i915_private *dev_priv = to_i915(dev); /* Enable ALPM at sink for psr2 */ - if (dev_priv->psr.psr2_support && dev_priv->psr.alpm) + if (dev_priv->psr.psr2_enabled && dev_priv->psr.alpm) drm_dp_dpcd_writeb(&intel_dp->aux, DP_RECEIVER_ALPM_CONFIG, DP_ALPM_ENABLE); @@ -424,7 +425,7 @@ static void hsw_psr_activate(struct intel_dp *intel_dp) */ /* psr1 and psr2 are mutually exclusive.*/ - if (dev_priv->psr.psr2_support) + if (dev_priv->psr.psr2_enabled) hsw_activate_psr2(intel_dp); else hsw_activate_psr1(intel_dp); @@ -444,7 +445,7 @@ static bool intel_psr2_config_valid(struct intel_dp *intel_dp, * dynamically during PSR enable, and extracted from sink * caps during eDP detection. */ - if (!dev_priv->psr.psr2_support) + if (!dev_priv->psr.sink_psr2_support) return false; if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv)) { @@ -543,7 +544,7 @@ static void intel_psr_activate(struct intel_dp *intel_dp) struct drm_device *dev = intel_dig_port->base.base.dev; struct drm_i915_private *dev_priv = to_i915(dev); - if (dev_priv->psr.psr2_support) + if (dev_priv->psr.psr2_enabled) WARN_ON(I915_READ(EDP_PSR2_CTL) & EDP_PSR2_ENABLE); else WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE); @@ -570,7 +571,7 @@ static void hsw_psr_enable_source(struct intel_dp *intel_dp, if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) hsw_psr_setup_aux(intel_dp); - if (dev_priv->psr.psr2_support) { + if (dev_priv->psr.psr2_enabled) { u32 chicken = I915_READ(CHICKEN_TRANS(cpu_transcoder)); if (INTEL_GEN(dev_priv) == 9 && !IS_GEMINILAKE(dev_priv)) @@ -629,7 +630,7 @@ void intel_psr_enable(struct intel_dp *intel_dp, goto unlock; } - dev_priv->psr.psr2_support = crtc_state->has_psr2; + dev_priv->psr.psr2_enabled = crtc_state->has_psr2; dev_priv->psr.busy_frontbuffer_bits = 0; dev_priv->psr.setup_vsc(intel_dp, crtc_state); @@ -699,7 +700,7 @@ static void hsw_psr_disable(struct intel_dp *intel_dp, i915_reg_t psr_status; u32 psr_status_mask; - if (dev_priv->psr.psr2_support) { + if (dev_priv->psr.psr2_enabled) { psr_status = EDP_PSR2_STATUS; psr_status_mask = EDP_PSR2_STATUS_STATE_MASK; @@ -723,7 +724,7 @@ static void hsw_psr_disable(struct intel_dp *intel_dp, dev_priv->psr.active = false; } else { - if (dev_priv->psr.psr2_support) + if (dev_priv->psr.psr2_enabled) WARN_ON(I915_READ(EDP_PSR2_CTL) & EDP_PSR2_ENABLE); else WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE); @@ -783,7 +784,7 @@ static void intel_psr_work(struct work_struct *work) * and be ready for re-enable. */ if (HAS_DDI(dev_priv)) { - if (dev_priv->psr.psr2_support) { + if (dev_priv->psr.psr2_enabled) { if (intel_wait_for_register(dev_priv, EDP_PSR2_STATUS, EDP_PSR2_STATUS_STATE_MASK, @@ -842,7 +843,7 @@ static void intel_psr_exit(struct drm_i915_private *dev_priv) return; if (HAS_DDI(dev_priv)) { - if (dev_priv->psr.psr2_support) { + if (dev_priv->psr.psr2_enabled) { val = I915_READ(EDP_PSR2_CTL); WARN_ON(!(val & EDP_PSR2_ENABLE)); I915_WRITE(EDP_PSR2_CTL, val & ~EDP_PSR2_ENABLE); @@ -1011,7 +1012,7 @@ void intel_psr_flush(struct drm_i915_private *dev_priv, /* By definition flush = invalidate + flush */ if (frontbuffer_bits) { - if (dev_priv->psr.psr2_support || + if (dev_priv->psr.psr2_enabled || IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { intel_psr_exit(dev_priv); } else { -- cgit v1.1 From fe36181be371f3d98441cc23ccbfa89783fa55b1 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jos=C3=A9=20Roberto=20de=20Souza?= Date: Wed, 28 Mar 2018 15:30:43 -0700 Subject: drm/i915/psr: Use PSR2 macro for PSR2 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Cosmetic change. Reviewed-by: Rodrigo Vivi Signed-off-by: José Roberto de Souza Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180328223046.16125-7-jose.souza@intel.com --- drivers/gpu/drm/i915/i915_reg.h | 3 ++- drivers/gpu/drm/i915/intel_psr.c | 2 +- 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 6566f6b..17b8691 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -4069,8 +4069,9 @@ enum { #define EDP_PSR2_TP2_TIME_MASK (3<<8) #define EDP_PSR2_FRAME_BEFORE_SU_SHIFT 4 #define EDP_PSR2_FRAME_BEFORE_SU_MASK (0xf<<4) -#define EDP_PSR2_IDLE_MASK 0xf #define EDP_PSR2_FRAME_BEFORE_SU(a) ((a)<<4) +#define EDP_PSR2_IDLE_FRAME_MASK 0xf +#define EDP_PSR2_IDLE_FRAME_SHIFT 0 #define EDP_PSR2_STATUS _MMIO(0x6f940) #define EDP_PSR2_STATUS_STATE_MASK (0xf<<28) diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index 5efddd9..bec455e 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -382,7 +382,7 @@ static void hsw_activate_psr2(struct intel_dp *intel_dp) uint32_t val; uint8_t sink_latency; - val = idle_frames << EDP_PSR_IDLE_FRAME_SHIFT; + val = idle_frames << EDP_PSR2_IDLE_FRAME_SHIFT; /* FIXME: selective update is probably totally broken because it doesn't * mesh at all with our frontbuffer tracking. And the hw alone isn't -- cgit v1.1 From 26e5378d115501a7cab25fdfc6ab10ccb5e4106f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jos=C3=A9=20Roberto=20de=20Souza?= Date: Wed, 28 Mar 2018 15:30:44 -0700 Subject: drm/i915/psr: Cache sink synchronization latency MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This value do not change overtime so better cache it than fetch it every PSR enable. Cc: Dhinakaran Pandiyan Reviewed-by: Rodrigo Vivi Signed-off-by: José Roberto de Souza Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180328223046.16125-8-jose.souza@intel.com --- drivers/gpu/drm/i915/i915_drv.h | 1 + drivers/gpu/drm/i915/intel_psr.c | 28 ++++++++++++++++------------ 2 files changed, 17 insertions(+), 12 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 46cae09..5373b17 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -608,6 +608,7 @@ struct i915_psr { bool alpm; bool has_hw_tracking; bool psr2_enabled; + u8 sink_sync_latency; void (*enable_source)(struct intel_dp *, const struct intel_crtc_state *); diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index bec455e..d079cf0 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -122,6 +122,18 @@ static bool intel_dp_get_alpm_status(struct intel_dp *intel_dp) return alpm_caps & DP_ALPM_CAP; } +static u8 intel_dp_get_sink_sync_latency(struct intel_dp *intel_dp) +{ + u8 val = 0; + + if (drm_dp_dpcd_readb(&intel_dp->aux, + DP_SYNCHRONIZATION_LATENCY_IN_SINK, &val) == 1) + val &= DP_MAX_RESYNC_FRAME_COUNT_MASK; + else + DRM_ERROR("Unable to get sink synchronization latency\n"); + return val; +} + void intel_psr_init_dpcd(struct intel_dp *intel_dp) { struct drm_i915_private *dev_priv = @@ -158,6 +170,8 @@ void intel_psr_init_dpcd(struct intel_dp *intel_dp) intel_dp_get_colorimetry_status(intel_dp); dev_priv->psr.alpm = intel_dp_get_alpm_status(intel_dp); + dev_priv->psr.sink_sync_latency = + intel_dp_get_sink_sync_latency(intel_dp); } } } @@ -379,10 +393,7 @@ static void hsw_activate_psr2(struct intel_dp *intel_dp) * with the 5 or 6 idle patterns. */ uint32_t idle_frames = max(6, dev_priv->vbt.psr.idle_frames); - uint32_t val; - uint8_t sink_latency; - - val = idle_frames << EDP_PSR2_IDLE_FRAME_SHIFT; + u32 val = idle_frames << EDP_PSR2_IDLE_FRAME_SHIFT; /* FIXME: selective update is probably totally broken because it doesn't * mesh at all with our frontbuffer tracking. And the hw alone isn't @@ -392,14 +403,7 @@ static void hsw_activate_psr2(struct intel_dp *intel_dp) val |= EDP_Y_COORDINATE_VALID | EDP_Y_COORDINATE_ENABLE; } - if (drm_dp_dpcd_readb(&intel_dp->aux, - DP_SYNCHRONIZATION_LATENCY_IN_SINK, - &sink_latency) == 1) { - sink_latency &= DP_MAX_RESYNC_FRAME_COUNT_MASK; - } else { - sink_latency = 0; - } - val |= EDP_PSR2_FRAME_BEFORE_SU(sink_latency + 1); + val |= EDP_PSR2_FRAME_BEFORE_SU(dev_priv->psr.sink_sync_latency + 1); if (dev_priv->vbt.psr.tp2_tp3_wakeup_time > 5) val |= EDP_PSR2_TP2_TIME_2500; -- cgit v1.1 From 4df4925b1b26f285aa76f89d95db3388a2d55281 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jos=C3=A9=20Roberto=20de=20Souza?= Date: Wed, 28 Mar 2018 15:30:45 -0700 Subject: drm/i915/psr: Set DPCD PSR2 enable bit when needed MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In the 2 eDP1.4a pannels tested set or not set bit have no effect but is better set it and comply with specification. Signed-off-by: José Roberto de Souza Cc: Dhinakaran Pandiyan Reviewed-by: Rodrigo Vivi Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20180328223046.16125-9-jose.souza@intel.com --- drivers/gpu/drm/i915/intel_psr.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index d079cf0..2d53f73 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -278,18 +278,19 @@ static void hsw_psr_enable_sink(struct intel_dp *intel_dp) struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); struct drm_device *dev = dig_port->base.base.dev; struct drm_i915_private *dev_priv = to_i915(dev); + u8 dpcd_val = DP_PSR_ENABLE; /* Enable ALPM at sink for psr2 */ if (dev_priv->psr.psr2_enabled && dev_priv->psr.alpm) drm_dp_dpcd_writeb(&intel_dp->aux, DP_RECEIVER_ALPM_CONFIG, DP_ALPM_ENABLE); + + if (dev_priv->psr.psr2_enabled) + dpcd_val |= DP_PSR_ENABLE_PSR2; if (dev_priv->psr.link_standby) - drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, - DP_PSR_ENABLE | DP_PSR_MAIN_LINK_ACTIVE); - else - drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, - DP_PSR_ENABLE); + dpcd_val |= DP_PSR_MAIN_LINK_ACTIVE; + drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, dpcd_val); drm_dp_dpcd_writeb(&intel_dp->aux, DP_SET_POWER, DP_SET_POWER_D0); } -- cgit v1.1 From d93ae190e2c95276caceb3642e6d541d93bba705 Mon Sep 17 00:00:00 2001 From: Russell King Date: Thu, 17 Nov 2016 23:38:29 +0000 Subject: drm/i2c: tda998x: move mutex/waitqueue/timer/work init early Move the mutex, waitqueue, timer and detect work initialisation early in the driver's initialisation, rather than being after we've registered the CEC device. Acked-by: Hans Verkuil Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index cd3f087..8340715 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -1475,7 +1475,11 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv) u32 video; int rev_lo, rev_hi, ret; - mutex_init(&priv->audio_mutex); /* Protect access from audio thread */ + mutex_init(&priv->mutex); /* protect the page access */ + mutex_init(&priv->audio_mutex); /* protect access from audio thread */ + init_waitqueue_head(&priv->edid_delay_waitq); + timer_setup(&priv->edid_delay_timer, tda998x_edid_delay_done, 0); + INIT_WORK(&priv->detect_work, tda998x_detect_work); priv->vip_cntrl_0 = VIP_CNTRL_0_SWAP_A(2) | VIP_CNTRL_0_SWAP_B(3); priv->vip_cntrl_1 = VIP_CNTRL_1_SWAP_C(0) | VIP_CNTRL_1_SWAP_D(1); @@ -1489,11 +1493,6 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv) if (!priv->cec) return -ENODEV; - mutex_init(&priv->mutex); /* protect the page access */ - init_waitqueue_head(&priv->edid_delay_waitq); - timer_setup(&priv->edid_delay_timer, tda998x_edid_delay_done, 0); - INIT_WORK(&priv->detect_work, tda998x_detect_work); - /* wake up the device: */ cec_write(priv, REG_CEC_ENAMODS, CEC_ENAMODS_EN_RXSENS | CEC_ENAMODS_EN_HDMI); -- cgit v1.1 From 6a765c3fe5497359c11536dfbdcf7526ccb2a33f Mon Sep 17 00:00:00 2001 From: Russell King Date: Thu, 17 Nov 2016 23:49:43 +0000 Subject: drm/i2c: tda998x: fix error cleanup paths If tda998x_get_audio_ports() fails, and we requested the interrupt, we fail to free the interrupt before returning failure. Rework the failure cleanup code and exit paths so that we always clean up properly after an error, and always propagate the error code. Acked-by: Hans Verkuil Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 31 ++++++++++++++++++------------- 1 file changed, 18 insertions(+), 13 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 8340715..2a99930 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -1501,10 +1501,15 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv) /* read version: */ rev_lo = reg_read(priv, REG_VERSION_LSB); + if (rev_lo < 0) { + dev_err(&client->dev, "failed to read version: %d\n", rev_lo); + return rev_lo; + } + rev_hi = reg_read(priv, REG_VERSION_MSB); - if (rev_lo < 0 || rev_hi < 0) { - ret = rev_lo < 0 ? rev_lo : rev_hi; - goto fail; + if (rev_hi < 0) { + dev_err(&client->dev, "failed to read version: %d\n", rev_hi); + return rev_hi; } priv->rev = rev_lo | rev_hi << 8; @@ -1528,7 +1533,7 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv) default: dev_err(&client->dev, "found unsupported device: %04x\n", priv->rev); - goto fail; + return -ENXIO; } /* after reset, enable DDC: */ @@ -1566,7 +1571,7 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv) dev_err(&client->dev, "failed to request IRQ#%u: %d\n", client->irq, ret); - goto fail; + goto err_irq; } /* enable HPD irq */ @@ -1589,19 +1594,19 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv) ret = tda998x_get_audio_ports(priv, np); if (ret) - goto fail; + goto err_audio; if (priv->audio_port[0].format != AFMT_UNUSED) tda998x_audio_codec_init(priv, &client->dev); return 0; -fail: - /* if encoder_init fails, the encoder slave is never registered, - * so cleanup here: - */ - if (priv->cec) - i2c_unregister_device(priv->cec); - return -ENXIO; + +err_audio: + if (client->irq) + free_irq(client->irq, priv); +err_irq: + i2c_unregister_device(priv->cec); + return ret; } static void tda998x_encoder_prepare(struct drm_encoder *encoder) -- cgit v1.1 From 101e996b8d321514c45136bef4dd0b1d3c577036 Mon Sep 17 00:00:00 2001 From: Russell King Date: Thu, 17 Nov 2016 23:40:26 +0000 Subject: drm/i2c: tda998x: move CEC device initialisation later We no longer use the CEC client to access the CEC part itself, so we can move this later in the initialisation sequence. Acked-by: Hans Verkuil Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 2a99930..7f2762f 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -1489,9 +1489,6 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv) priv->cec_addr = 0x34 + (client->addr & 0x03); priv->current_page = 0xff; priv->hdmi = client; - priv->cec = i2c_new_dummy(client->adapter, priv->cec_addr); - if (!priv->cec) - return -ENODEV; /* wake up the device: */ cec_write(priv, REG_CEC_ENAMODS, @@ -1578,6 +1575,12 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv) cec_write(priv, REG_CEC_RXSHPDINTENA, CEC_RXSHPDLEV_HPD); } + priv->cec = i2c_new_dummy(client->adapter, priv->cec_addr); + if (!priv->cec) { + ret = -ENODEV; + goto fail; + } + /* enable EDID read irq: */ reg_set(priv, REG_INT_FLAGS_2, INT_FLAGS_2_EDID_BLK_RD); @@ -1594,14 +1597,14 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv) ret = tda998x_get_audio_ports(priv, np); if (ret) - goto err_audio; + goto fail; if (priv->audio_port[0].format != AFMT_UNUSED) tda998x_audio_codec_init(priv, &client->dev); return 0; -err_audio: +fail: if (client->irq) free_irq(client->irq, priv); err_irq: -- cgit v1.1 From ba8975f15bb93d7f3ebd995a41c2e4b6945bad78 Mon Sep 17 00:00:00 2001 From: Russell King Date: Sat, 11 Mar 2017 11:12:22 +0000 Subject: drm/i2c: tda998x: always disable and clear interrupts at probe Always disable and clear interrupts at probe time to ensure that the TDA998x is in a sane state. This ensures that the interrupt line, which is also the CEC clock calibration signal, is always deasserted. Acked-by: Hans Verkuil Signed-off-by: Russell King --- drivers/gpu/drm/i2c/tda998x_drv.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 7f2762f..16e0439 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -1546,6 +1546,15 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv) cec_write(priv, REG_CEC_FRO_IM_CLK_CTRL, CEC_FRO_IM_CLK_CTRL_GHOST_DIS | CEC_FRO_IM_CLK_CTRL_IMCLK_SEL); + /* ensure interrupts are disabled */ + cec_write(priv, REG_CEC_RXSHPDINTENA, 0); + + /* clear pending interrupts */ + cec_read(priv, REG_CEC_RXSHPDINT); + reg_read(priv, REG_INT_FLAGS_0); + reg_read(priv, REG_INT_FLAGS_1); + reg_read(priv, REG_INT_FLAGS_2); + /* initialize the optional IRQ */ if (client->irq) { unsigned long irq_flags; @@ -1553,11 +1562,6 @@ static int tda998x_create(struct i2c_client *client, struct tda998x_priv *priv) /* init read EDID waitqueue and HDP work */ init_waitqueue_head(&priv->wq_edid); - /* clear pending interrupts */ - reg_read(priv, REG_INT_FLAGS_0); - reg_read(priv, REG_INT_FLAGS_1); - reg_read(priv, REG_INT_FLAGS_2); - irq_flags = irqd_get_trigger_type(irq_get_irq_data(client->irq)); irq_flags |= IRQF_SHARED | IRQF_ONESHOT; -- cgit v1.1 From f26052079070cd0e8940717a7a7cb7cdd512ac05 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Sat, 31 Mar 2018 14:06:26 +0100 Subject: drm/i915/execlists: Track begin/end of execlists submission sequences We would like to start doing some bookkeeping at the beginning, between contexts and at the end of execlists submission. We already mark the beginning and end using EXECLISTS_ACTIVE_USER, to provide an indication when the HW is idle. This give us a pair of sequence points we can then expand on for further bookkeeping. v2: Refactor guc submission to share the same begin/end. Signed-off-by: Chris Wilson Cc: Mika Kuoppala Cc: Francisco Jerez Reviewed-by: Francisco Jerez Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20180331130626.10712-1-chris@chris-wilson.co.uk --- drivers/gpu/drm/i915/intel_guc_submission.c | 17 ++++++---- drivers/gpu/drm/i915/intel_lrc.c | 50 ++++++++++++++++++++++------- drivers/gpu/drm/i915/intel_ringbuffer.h | 15 ++++++++- 3 files changed, 63 insertions(+), 19 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_guc_submission.c b/drivers/gpu/drm/i915/intel_guc_submission.c index 207cda0..749f279 100644 --- a/drivers/gpu/drm/i915/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/intel_guc_submission.c @@ -728,7 +728,7 @@ done: execlists->first = rb; if (submit) { port_assign(port, last); - execlists_set_active(execlists, EXECLISTS_ACTIVE_USER); + execlists_user_begin(execlists, execlists->port); guc_submit(engine); } @@ -748,17 +748,20 @@ static void guc_submission_tasklet(unsigned long data) struct execlist_port *port = execlists->port; struct i915_request *rq; - rq = port_request(&port[0]); + rq = port_request(port); while (rq && i915_request_completed(rq)) { trace_i915_request_out(rq); i915_request_put(rq); - execlists_port_complete(execlists, port); - - rq = port_request(&port[0]); + port = execlists_port_complete(execlists, port); + if (port_isset(port)) { + execlists_user_begin(execlists, port); + rq = port_request(port); + } else { + execlists_user_end(execlists); + rq = NULL; + } } - if (!rq) - execlists_clear_active(execlists, EXECLISTS_ACTIVE_USER); if (execlists_is_active(execlists, EXECLISTS_ACTIVE_PREEMPT) && intel_read_status_page(engine, I915_GEM_HWS_PREEMPT_INDEX) == diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index f60b61b..4d08875 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -374,6 +374,19 @@ execlists_context_status_change(struct i915_request *rq, unsigned long status) status, rq); } +inline void +execlists_user_begin(struct intel_engine_execlists *execlists, + const struct execlist_port *port) +{ + execlists_set_active_once(execlists, EXECLISTS_ACTIVE_USER); +} + +inline void +execlists_user_end(struct intel_engine_execlists *execlists) +{ + execlists_clear_active(execlists, EXECLISTS_ACTIVE_USER); +} + static inline void execlists_context_schedule_in(struct i915_request *rq) { @@ -711,7 +724,7 @@ unlock: spin_unlock_irq(&engine->timeline->lock); if (submit) { - execlists_set_active(execlists, EXECLISTS_ACTIVE_USER); + execlists_user_begin(execlists, execlists->port); execlists_submit_ports(engine); } @@ -742,7 +755,7 @@ execlists_cancel_port_requests(struct intel_engine_execlists * const execlists) port++; } - execlists_clear_active(execlists, EXECLISTS_ACTIVE_USER); + execlists_user_end(execlists); } static void clear_gtiir(struct intel_engine_cs *engine) @@ -873,7 +886,7 @@ static void execlists_submission_tasklet(unsigned long data) { struct intel_engine_cs * const engine = (struct intel_engine_cs *)data; struct intel_engine_execlists * const execlists = &engine->execlists; - struct execlist_port * const port = execlists->port; + struct execlist_port *port = execlists->port; struct drm_i915_private *dev_priv = engine->i915; bool fw = false; @@ -1012,10 +1025,28 @@ static void execlists_submission_tasklet(unsigned long data) GEM_BUG_ON(count == 0); if (--count == 0) { + /* + * On the final event corresponding to the + * submission of this context, we expect either + * an element-switch event or a completion + * event (and on completion, the active-idle + * marker). No more preemptions, lite-restore + * or otherwise. + */ GEM_BUG_ON(status & GEN8_CTX_STATUS_PREEMPTED); GEM_BUG_ON(port_isset(&port[1]) && !(status & GEN8_CTX_STATUS_ELEMENT_SWITCH)); + GEM_BUG_ON(!port_isset(&port[1]) && + !(status & GEN8_CTX_STATUS_ACTIVE_IDLE)); + + /* + * We rely on the hardware being strongly + * ordered, that the breadcrumb write is + * coherent (visible from the CPU) before the + * user interrupt and CSB is processed. + */ GEM_BUG_ON(!i915_request_completed(rq)); + execlists_context_schedule_out(rq); trace_i915_request_out(rq); i915_request_put(rq); @@ -1023,17 +1054,14 @@ static void execlists_submission_tasklet(unsigned long data) GEM_TRACE("%s completed ctx=%d\n", engine->name, port->context_id); - execlists_port_complete(execlists, port); + port = execlists_port_complete(execlists, port); + if (port_isset(port)) + execlists_user_begin(execlists, port); + else + execlists_user_end(execlists); } else { port_set(port, port_pack(rq, count)); } - - /* After the final element, the hw should be idle */ - GEM_BUG_ON(port_count(port) == 0 && - !(status & GEN8_CTX_STATUS_ACTIVE_IDLE)); - if (port_count(port) == 0) - execlists_clear_active(execlists, - EXECLISTS_ACTIVE_USER); } if (head != execlists->csb_head) { diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index a02c7b3..40461e2 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -638,6 +638,13 @@ execlists_set_active(struct intel_engine_execlists *execlists, __set_bit(bit, (unsigned long *)&execlists->active); } +static inline bool +execlists_set_active_once(struct intel_engine_execlists *execlists, + unsigned int bit) +{ + return !__test_and_set_bit(bit, (unsigned long *)&execlists->active); +} + static inline void execlists_clear_active(struct intel_engine_execlists *execlists, unsigned int bit) @@ -652,6 +659,10 @@ execlists_is_active(const struct intel_engine_execlists *execlists, return test_bit(bit, (unsigned long *)&execlists->active); } +void execlists_user_begin(struct intel_engine_execlists *execlists, + const struct execlist_port *port); +void execlists_user_end(struct intel_engine_execlists *execlists); + void execlists_cancel_port_requests(struct intel_engine_execlists * const execlists); @@ -664,7 +675,7 @@ execlists_num_ports(const struct intel_engine_execlists * const execlists) return execlists->port_mask + 1; } -static inline void +static inline struct execlist_port * execlists_port_complete(struct intel_engine_execlists * const execlists, struct execlist_port * const port) { @@ -675,6 +686,8 @@ execlists_port_complete(struct intel_engine_execlists * const execlists, memmove(port, port + 1, m * sizeof(struct execlist_port)); memset(port + m, 0, sizeof(struct execlist_port)); + + return port; } static inline unsigned int -- cgit v1.1 From c575b7eeb89f94356997abd62d6d5a0590e259b7 Mon Sep 17 00:00:00 2001 From: Oleksandr Andrushchenko Date: Tue, 3 Apr 2018 14:23:17 +0300 Subject: drm/xen-front: Add support for Xen PV display frontend Add support for Xen para-virtualized frontend display driver. Accompanying backend [1] is implemented as a user-space application and its helper library [2], capable of running as a Weston client or DRM master. Configuration of both backend and frontend is done via Xen guest domain configuration options [3]. Driver limitations: 1. Only primary plane without additional properties is supported. 2. Only one video mode supported which resolution is configured via XenStore. 3. All CRTCs operate at fixed frequency of 60Hz. 1. Implement Xen bus state machine for the frontend driver according to the state diagram and recovery flow from display para-virtualized protocol: xen/interface/io/displif.h. 2. Read configuration values from Xen store according to xen/interface/io/displif.h protocol: - read connector(s) configuration - read buffer allocation mode (backend/frontend) 3. Handle Xen event channels: - create for all configured connectors and publish corresponding ring references and event channels in Xen store, so backend can connect - implement event channels interrupt handlers - create and destroy event channels with respect to Xen bus state 4. Implement shared buffer handling according to the para-virtualized display device protocol at xen/interface/io/displif.h: - handle page directories according to displif protocol: - allocate and share page directories - grant references to the required set of pages for the page directory - allocate xen balllooned pages via Xen balloon driver with alloc_xenballooned_pages/free_xenballooned_pages - grant references to the required set of pages for the shared buffer itself - implement pages map/unmap for the buffers allocated by the backend (gnttab_map_refs/gnttab_unmap_refs) 5. Implement kernel modesetiing/connector handling using DRM simple KMS helper pipeline: - implement KMS part of the driver with the help of DRM simple pipepline helper which is possible due to the fact that the para-virtualized driver only supports a single (primary) plane: - initialize connectors according to XenStore configuration - handle frame done events from the backend - create and destroy frame buffers and propagate those to the backend - propagate set/reset mode configuration to the backend on display enable/disable callbacks - send page flip request to the backend and implement logic for reporting backend IO errors on prepare fb callback - implement virtual connector handling: - support only pixel formats suitable for single plane modes - make sure the connector is always connected - support a single video mode as per para-virtualized driver configuration 6. Implement GEM handling depending on driver mode of operation: depending on the requirements for the para-virtualized environment, namely requirements dictated by the accompanying DRM/(v)GPU drivers running in both host and guest environments, number of operating modes of para-virtualized display driver are supported: - display buffers can be allocated by either frontend driver or backend - display buffers can be allocated to be contiguous in memory or not Note! Frontend driver itself has no dependency on contiguous memory for its operation. 6.1. Buffers allocated by the frontend driver. The below modes of operation are configured at compile-time via frontend driver's kernel configuration. 6.1.1. Front driver configured to use GEM CMA helpers This use-case is useful when used with accompanying DRM/vGPU driver in guest domain which was designed to only work with contiguous buffers, e.g. DRM driver based on GEM CMA helpers: such drivers can only import contiguous PRIME buffers, thus requiring frontend driver to provide such. In order to implement this mode of operation para-virtualized frontend driver can be configured to use GEM CMA helpers. 6.1.2. Front driver doesn't use GEM CMA If accompanying drivers can cope with non-contiguous memory then, to lower pressure on CMA subsystem of the kernel, driver can allocate buffers from system memory. Note! If used with accompanying DRM/(v)GPU drivers this mode of operation may require IOMMU support on the platform, so accompanying DRM/vGPU hardware can still reach display buffer memory while importing PRIME buffers from the frontend driver. 6.2. Buffers allocated by the backend This mode of operation is run-time configured via guest domain configuration through XenStore entries. For systems which do not provide IOMMU support, but having specific requirements for display buffers it is possible to allocate such buffers at backend side and share those with the frontend. For example, if host domain is 1:1 mapped and has DRM/GPU hardware expecting physically contiguous memory, this allows implementing zero-copying use-cases. Note, while using this scenario the following should be considered: a) If guest domain dies then pages/grants received from the backend cannot be claimed back b) Misbehaving guest may send too many requests to the backend exhausting its grant references and memory (consider this from security POV). Note! Configuration options 1.1 (contiguous display buffers) and 2 (backend allocated buffers) are not supported at the same time. 7. Handle communication with the backend: - send requests and wait for the responses according to the displif protocol - serialize access to the communication channel - time-out used for backend communication is set to 3000 ms - manage display buffers shared with the backend [1] https://github.com/xen-troops/displ_be [2] https://github.com/xen-troops/libxenbe [3] https://xenbits.xen.org/gitweb/?p=xen.git;a=blob;f=docs/man/xl.cfg.pod.5.in;h=a699367779e2ae1212ff8f638eff0206ec1a1cc9;hb=refs/heads/master#l1257 Signed-off-by: Oleksandr Andrushchenko Reviewed-by: Boris Ostrovsky Reviewed-by: Daniel Vetter Link: https://patchwork.freedesktop.org/patch/msgid/20180403112317.28751-2-andr2000@gmail.com --- Documentation/gpu/drivers.rst | 1 + Documentation/gpu/xen-front.rst | 43 ++ drivers/gpu/drm/Kconfig | 2 + drivers/gpu/drm/Makefile | 1 + drivers/gpu/drm/xen/Kconfig | 30 + drivers/gpu/drm/xen/Makefile | 16 + drivers/gpu/drm/xen/xen_drm_front.c | 882 ++++++++++++++++++++++++++++ drivers/gpu/drm/xen/xen_drm_front.h | 188 ++++++ drivers/gpu/drm/xen/xen_drm_front_cfg.c | 77 +++ drivers/gpu/drm/xen/xen_drm_front_cfg.h | 37 ++ drivers/gpu/drm/xen/xen_drm_front_conn.c | 115 ++++ drivers/gpu/drm/xen/xen_drm_front_conn.h | 27 + drivers/gpu/drm/xen/xen_drm_front_evtchnl.c | 387 ++++++++++++ drivers/gpu/drm/xen/xen_drm_front_evtchnl.h | 81 +++ drivers/gpu/drm/xen/xen_drm_front_gem.c | 314 ++++++++++ drivers/gpu/drm/xen/xen_drm_front_gem.h | 43 ++ drivers/gpu/drm/xen/xen_drm_front_gem_cma.c | 79 +++ drivers/gpu/drm/xen/xen_drm_front_kms.c | 372 ++++++++++++ drivers/gpu/drm/xen/xen_drm_front_kms.h | 26 + drivers/gpu/drm/xen/xen_drm_front_shbuf.c | 436 ++++++++++++++ drivers/gpu/drm/xen/xen_drm_front_shbuf.h | 72 +++ 21 files changed, 3229 insertions(+) create mode 100644 Documentation/gpu/xen-front.rst create mode 100644 drivers/gpu/drm/xen/Kconfig create mode 100644 drivers/gpu/drm/xen/Makefile create mode 100644 drivers/gpu/drm/xen/xen_drm_front.c create mode 100644 drivers/gpu/drm/xen/xen_drm_front.h create mode 100644 drivers/gpu/drm/xen/xen_drm_front_cfg.c create mode 100644 drivers/gpu/drm/xen/xen_drm_front_cfg.h create mode 100644 drivers/gpu/drm/xen/xen_drm_front_conn.c create mode 100644 drivers/gpu/drm/xen/xen_drm_front_conn.h create mode 100644 drivers/gpu/drm/xen/xen_drm_front_evtchnl.c create mode 100644 drivers/gpu/drm/xen/xen_drm_front_evtchnl.h create mode 100644 drivers/gpu/drm/xen/xen_drm_front_gem.c create mode 100644 drivers/gpu/drm/xen/xen_drm_front_gem.h create mode 100644 drivers/gpu/drm/xen/xen_drm_front_gem_cma.c create mode 100644 drivers/gpu/drm/xen/xen_drm_front_kms.c create mode 100644 drivers/gpu/drm/xen/xen_drm_front_kms.h create mode 100644 drivers/gpu/drm/xen/xen_drm_front_shbuf.c create mode 100644 drivers/gpu/drm/xen/xen_drm_front_shbuf.h diff --git a/Documentation/gpu/drivers.rst b/Documentation/gpu/drivers.rst index e8c8441..d3ab6ab 100644 --- a/Documentation/gpu/drivers.rst +++ b/Documentation/gpu/drivers.rst @@ -12,6 +12,7 @@ GPU Driver Documentation tve200 vc4 bridge/dw-hdmi + xen-front .. only:: subproject and html diff --git a/Documentation/gpu/xen-front.rst b/Documentation/gpu/xen-front.rst new file mode 100644 index 0000000..009d942 --- /dev/null +++ b/Documentation/gpu/xen-front.rst @@ -0,0 +1,43 @@ +==================================================== + drm/xen-front Xen para-virtualized frontend driver +==================================================== + +This frontend driver implements Xen para-virtualized display +according to the display protocol described at +include/xen/interface/io/displif.h + +Driver modes of operation in terms of display buffers used +========================================================== + +.. kernel-doc:: drivers/gpu/drm/xen/xen_drm_front.h + :doc: Driver modes of operation in terms of display buffers used + +Buffers allocated by the frontend driver +---------------------------------------- + +.. kernel-doc:: drivers/gpu/drm/xen/xen_drm_front.h + :doc: Buffers allocated by the frontend driver + +With GEM CMA helpers +~~~~~~~~~~~~~~~~~~~~ + +.. kernel-doc:: drivers/gpu/drm/xen/xen_drm_front.h + :doc: With GEM CMA helpers + +Without GEM CMA helpers +~~~~~~~~~~~~~~~~~~~~~~~ + +.. kernel-doc:: drivers/gpu/drm/xen/xen_drm_front.h + :doc: Without GEM CMA helpers + +Buffers allocated by the backend +-------------------------------- + +.. kernel-doc:: drivers/gpu/drm/xen/xen_drm_front.h + :doc: Buffers allocated by the backend + +Driver limitations +================== + +.. kernel-doc:: drivers/gpu/drm/xen/xen_drm_front.h + :doc: Driver limitations diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig index deeefa7..757825a 100644 --- a/drivers/gpu/drm/Kconfig +++ b/drivers/gpu/drm/Kconfig @@ -289,6 +289,8 @@ source "drivers/gpu/drm/pl111/Kconfig" source "drivers/gpu/drm/tve200/Kconfig" +source "drivers/gpu/drm/xen/Kconfig" + # Keep legacy drivers last menuconfig DRM_LEGACY diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile index 50093ff..9d66657 100644 --- a/drivers/gpu/drm/Makefile +++ b/drivers/gpu/drm/Makefile @@ -103,3 +103,4 @@ obj-$(CONFIG_DRM_MXSFB) += mxsfb/ obj-$(CONFIG_DRM_TINYDRM) += tinydrm/ obj-$(CONFIG_DRM_PL111) += pl111/ obj-$(CONFIG_DRM_TVE200) += tve200/ +obj-$(CONFIG_DRM_XEN) += xen/ diff --git a/drivers/gpu/drm/xen/Kconfig b/drivers/gpu/drm/xen/Kconfig new file mode 100644 index 0000000..4f4abc9 --- /dev/null +++ b/drivers/gpu/drm/xen/Kconfig @@ -0,0 +1,30 @@ +config DRM_XEN + bool "DRM Support for Xen guest OS" + depends on XEN + help + Choose this option if you want to enable DRM support + for Xen. + +config DRM_XEN_FRONTEND + tristate "Para-virtualized frontend driver for Xen guest OS" + depends on DRM_XEN + depends on DRM + select DRM_KMS_HELPER + select VIDEOMODE_HELPERS + select XEN_XENBUS_FRONTEND + help + Choose this option if you want to enable a para-virtualized + frontend DRM/KMS driver for Xen guest OSes. + +config DRM_XEN_FRONTEND_CMA + bool "Use DRM CMA to allocate dumb buffers" + depends on DRM_XEN_FRONTEND + select DRM_KMS_CMA_HELPER + select DRM_GEM_CMA_HELPER + help + Use DRM CMA helpers to allocate display buffers. + This is useful for the use-cases when guest driver needs to + share or export buffers to other drivers which only expect + contiguous buffers. + Note: in this mode driver cannot use buffers allocated + by the backend. diff --git a/drivers/gpu/drm/xen/Makefile b/drivers/gpu/drm/xen/Makefile new file mode 100644 index 0000000..352730d --- /dev/null +++ b/drivers/gpu/drm/xen/Makefile @@ -0,0 +1,16 @@ +# SPDX-License-Identifier: GPL-2.0 OR MIT + +drm_xen_front-objs := xen_drm_front.o \ + xen_drm_front_kms.o \ + xen_drm_front_conn.o \ + xen_drm_front_evtchnl.o \ + xen_drm_front_shbuf.o \ + xen_drm_front_cfg.o + +ifeq ($(CONFIG_DRM_XEN_FRONTEND_CMA),y) + drm_xen_front-objs += xen_drm_front_gem_cma.o +else + drm_xen_front-objs += xen_drm_front_gem.o +endif + +obj-$(CONFIG_DRM_XEN_FRONTEND) += drm_xen_front.o diff --git a/drivers/gpu/drm/xen/xen_drm_front.c b/drivers/gpu/drm/xen/xen_drm_front.c new file mode 100644 index 0000000..4a08b77 --- /dev/null +++ b/drivers/gpu/drm/xen/xen_drm_front.c @@ -0,0 +1,882 @@ +// SPDX-License-Identifier: GPL-2.0 OR MIT + +/* + * Xen para-virtual DRM device + * + * Copyright (C) 2016-2018 EPAM Systems Inc. + * + * Author: Oleksandr Andrushchenko + */ + +#include +#include +#include +#include +#include + +#include + +#include +#include +#include + +#include + +#include "xen_drm_front.h" +#include "xen_drm_front_cfg.h" +#include "xen_drm_front_evtchnl.h" +#include "xen_drm_front_gem.h" +#include "xen_drm_front_kms.h" +#include "xen_drm_front_shbuf.h" + +struct xen_drm_front_dbuf { + struct list_head list; + u64 dbuf_cookie; + u64 fb_cookie; + struct xen_drm_front_shbuf *shbuf; +}; + +static int dbuf_add_to_list(struct xen_drm_front_info *front_info, + struct xen_drm_front_shbuf *shbuf, u64 dbuf_cookie) +{ + struct xen_drm_front_dbuf *dbuf; + + dbuf = kzalloc(sizeof(*dbuf), GFP_KERNEL); + if (!dbuf) + return -ENOMEM; + + dbuf->dbuf_cookie = dbuf_cookie; + dbuf->shbuf = shbuf; + list_add(&dbuf->list, &front_info->dbuf_list); + return 0; +} + +static struct xen_drm_front_dbuf *dbuf_get(struct list_head *dbuf_list, + u64 dbuf_cookie) +{ + struct xen_drm_front_dbuf *buf, *q; + + list_for_each_entry_safe(buf, q, dbuf_list, list) + if (buf->dbuf_cookie == dbuf_cookie) + return buf; + + return NULL; +} + +static void dbuf_flush_fb(struct list_head *dbuf_list, u64 fb_cookie) +{ + struct xen_drm_front_dbuf *buf, *q; + + list_for_each_entry_safe(buf, q, dbuf_list, list) + if (buf->fb_cookie == fb_cookie) + xen_drm_front_shbuf_flush(buf->shbuf); +} + +static void dbuf_free(struct list_head *dbuf_list, u64 dbuf_cookie) +{ + struct xen_drm_front_dbuf *buf, *q; + + list_for_each_entry_safe(buf, q, dbuf_list, list) + if (buf->dbuf_cookie == dbuf_cookie) { + list_del(&buf->list); + xen_drm_front_shbuf_unmap(buf->shbuf); + xen_drm_front_shbuf_free(buf->shbuf); + kfree(buf); + break; + } +} + +static void dbuf_free_all(struct list_head *dbuf_list) +{ + struct xen_drm_front_dbuf *buf, *q; + + list_for_each_entry_safe(buf, q, dbuf_list, list) { + list_del(&buf->list); + xen_drm_front_shbuf_unmap(buf->shbuf); + xen_drm_front_shbuf_free(buf->shbuf); + kfree(buf); + } +} + +static struct xendispl_req * +be_prepare_req(struct xen_drm_front_evtchnl *evtchnl, u8 operation) +{ + struct xendispl_req *req; + + req = RING_GET_REQUEST(&evtchnl->u.req.ring, + evtchnl->u.req.ring.req_prod_pvt); + req->operation = operation; + req->id = evtchnl->evt_next_id++; + evtchnl->evt_id = req->id; + return req; +} + +static int be_stream_do_io(struct xen_drm_front_evtchnl *evtchnl, + struct xendispl_req *req) +{ + reinit_completion(&evtchnl->u.req.completion); + if (unlikely(evtchnl->state != EVTCHNL_STATE_CONNECTED)) + return -EIO; + + xen_drm_front_evtchnl_flush(evtchnl); + return 0; +} + +static int be_stream_wait_io(struct xen_drm_front_evtchnl *evtchnl) +{ + if (wait_for_completion_timeout(&evtchnl->u.req.completion, + msecs_to_jiffies(XEN_DRM_FRONT_WAIT_BACK_MS)) <= 0) + return -ETIMEDOUT; + + return evtchnl->u.req.resp_status; +} + +int xen_drm_front_mode_set(struct xen_drm_front_drm_pipeline *pipeline, + u32 x, u32 y, u32 width, u32 height, + u32 bpp, u64 fb_cookie) +{ + struct xen_drm_front_evtchnl *evtchnl; + struct xen_drm_front_info *front_info; + struct xendispl_req *req; + unsigned long flags; + int ret; + + front_info = pipeline->drm_info->front_info; + evtchnl = &front_info->evt_pairs[pipeline->index].req; + if (unlikely(!evtchnl)) + return -EIO; + + mutex_lock(&evtchnl->u.req.req_io_lock); + + spin_lock_irqsave(&front_info->io_lock, flags); + req = be_prepare_req(evtchnl, XENDISPL_OP_SET_CONFIG); + req->op.set_config.x = x; + req->op.set_config.y = y; + req->op.set_config.width = width; + req->op.set_config.height = height; + req->op.set_config.bpp = bpp; + req->op.set_config.fb_cookie = fb_cookie; + + ret = be_stream_do_io(evtchnl, req); + spin_unlock_irqrestore(&front_info->io_lock, flags); + + if (ret == 0) + ret = be_stream_wait_io(evtchnl); + + mutex_unlock(&evtchnl->u.req.req_io_lock); + return ret; +} + +static int be_dbuf_create_int(struct xen_drm_front_info *front_info, + u64 dbuf_cookie, u32 width, u32 height, + u32 bpp, u64 size, struct page **pages, + struct sg_table *sgt) +{ + struct xen_drm_front_evtchnl *evtchnl; + struct xen_drm_front_shbuf *shbuf; + struct xendispl_req *req; + struct xen_drm_front_shbuf_cfg buf_cfg; + unsigned long flags; + int ret; + + evtchnl = &front_info->evt_pairs[GENERIC_OP_EVT_CHNL].req; + if (unlikely(!evtchnl)) + return -EIO; + + memset(&buf_cfg, 0, sizeof(buf_cfg)); + buf_cfg.xb_dev = front_info->xb_dev; + buf_cfg.pages = pages; + buf_cfg.size = size; + buf_cfg.sgt = sgt; + buf_cfg.be_alloc = front_info->cfg.be_alloc; + + shbuf = xen_drm_front_shbuf_alloc(&buf_cfg); + if (!shbuf) + return -ENOMEM; + + ret = dbuf_add_to_list(front_info, shbuf, dbuf_cookie); + if (ret < 0) { + xen_drm_front_shbuf_free(shbuf); + return ret; + } + + mutex_lock(&evtchnl->u.req.req_io_lock); + + spin_lock_irqsave(&front_info->io_lock, flags); + req = be_prepare_req(evtchnl, XENDISPL_OP_DBUF_CREATE); + req->op.dbuf_create.gref_directory = + xen_drm_front_shbuf_get_dir_start(shbuf); + req->op.dbuf_create.buffer_sz = size; + req->op.dbuf_create.dbuf_cookie = dbuf_cookie; + req->op.dbuf_create.width = width; + req->op.dbuf_create.height = height; + req->op.dbuf_create.bpp = bpp; + if (buf_cfg.be_alloc) + req->op.dbuf_create.flags |= XENDISPL_DBUF_FLG_REQ_ALLOC; + + ret = be_stream_do_io(evtchnl, req); + spin_unlock_irqrestore(&front_info->io_lock, flags); + + if (ret < 0) + goto fail; + + ret = be_stream_wait_io(evtchnl); + if (ret < 0) + goto fail; + + ret = xen_drm_front_shbuf_map(shbuf); + if (ret < 0) + goto fail; + + mutex_unlock(&evtchnl->u.req.req_io_lock); + return 0; + +fail: + mutex_unlock(&evtchnl->u.req.req_io_lock); + dbuf_free(&front_info->dbuf_list, dbuf_cookie); + return ret; +} + +int xen_drm_front_dbuf_create_from_sgt(struct xen_drm_front_info *front_info, + u64 dbuf_cookie, u32 width, u32 height, + u32 bpp, u64 size, struct sg_table *sgt) +{ + return be_dbuf_create_int(front_info, dbuf_cookie, width, height, + bpp, size, NULL, sgt); +} + +int xen_drm_front_dbuf_create_from_pages(struct xen_drm_front_info *front_info, + u64 dbuf_cookie, u32 width, u32 height, + u32 bpp, u64 size, struct page **pages) +{ + return be_dbuf_create_int(front_info, dbuf_cookie, width, height, + bpp, size, pages, NULL); +} + +static int xen_drm_front_dbuf_destroy(struct xen_drm_front_info *front_info, + u64 dbuf_cookie) +{ + struct xen_drm_front_evtchnl *evtchnl; + struct xendispl_req *req; + unsigned long flags; + bool be_alloc; + int ret; + + evtchnl = &front_info->evt_pairs[GENERIC_OP_EVT_CHNL].req; + if (unlikely(!evtchnl)) + return -EIO; + + be_alloc = front_info->cfg.be_alloc; + + /* + * For the backend allocated buffer release references now, so backend + * can free the buffer. + */ + if (be_alloc) + dbuf_free(&front_info->dbuf_list, dbuf_cookie); + + mutex_lock(&evtchnl->u.req.req_io_lock); + + spin_lock_irqsave(&front_info->io_lock, flags); + req = be_prepare_req(evtchnl, XENDISPL_OP_DBUF_DESTROY); + req->op.dbuf_destroy.dbuf_cookie = dbuf_cookie; + + ret = be_stream_do_io(evtchnl, req); + spin_unlock_irqrestore(&front_info->io_lock, flags); + + if (ret == 0) + ret = be_stream_wait_io(evtchnl); + + /* + * Do this regardless of communication status with the backend: + * if we cannot remove remote resources remove what we can locally. + */ + if (!be_alloc) + dbuf_free(&front_info->dbuf_list, dbuf_cookie); + + mutex_unlock(&evtchnl->u.req.req_io_lock); + return ret; +} + +int xen_drm_front_fb_attach(struct xen_drm_front_info *front_info, + u64 dbuf_cookie, u64 fb_cookie, u32 width, + u32 height, u32 pixel_format) +{ + struct xen_drm_front_evtchnl *evtchnl; + struct xen_drm_front_dbuf *buf; + struct xendispl_req *req; + unsigned long flags; + int ret; + + evtchnl = &front_info->evt_pairs[GENERIC_OP_EVT_CHNL].req; + if (unlikely(!evtchnl)) + return -EIO; + + buf = dbuf_get(&front_info->dbuf_list, dbuf_cookie); + if (!buf) + return -EINVAL; + + buf->fb_cookie = fb_cookie; + + mutex_lock(&evtchnl->u.req.req_io_lock); + + spin_lock_irqsave(&front_info->io_lock, flags); + req = be_prepare_req(evtchnl, XENDISPL_OP_FB_ATTACH); + req->op.fb_attach.dbuf_cookie = dbuf_cookie; + req->op.fb_attach.fb_cookie = fb_cookie; + req->op.fb_attach.width = width; + req->op.fb_attach.height = height; + req->op.fb_attach.pixel_format = pixel_format; + + ret = be_stream_do_io(evtchnl, req); + spin_unlock_irqrestore(&front_info->io_lock, flags); + + if (ret == 0) + ret = be_stream_wait_io(evtchnl); + + mutex_unlock(&evtchnl->u.req.req_io_lock); + return ret; +} + +int xen_drm_front_fb_detach(struct xen_drm_front_info *front_info, + u64 fb_cookie) +{ + struct xen_drm_front_evtchnl *evtchnl; + struct xendispl_req *req; + unsigned long flags; + int ret; + + evtchnl = &front_info->evt_pairs[GENERIC_OP_EVT_CHNL].req; + if (unlikely(!evtchnl)) + return -EIO; + + mutex_lock(&evtchnl->u.req.req_io_lock); + + spin_lock_irqsave(&front_info->io_lock, flags); + req = be_prepare_req(evtchnl, XENDISPL_OP_FB_DETACH); + req->op.fb_detach.fb_cookie = fb_cookie; + + ret = be_stream_do_io(evtchnl, req); + spin_unlock_irqrestore(&front_info->io_lock, flags); + + if (ret == 0) + ret = be_stream_wait_io(evtchnl); + + mutex_unlock(&evtchnl->u.req.req_io_lock); + return ret; +} + +int xen_drm_front_page_flip(struct xen_drm_front_info *front_info, + int conn_idx, u64 fb_cookie) +{ + struct xen_drm_front_evtchnl *evtchnl; + struct xendispl_req *req; + unsigned long flags; + int ret; + + if (unlikely(conn_idx >= front_info->num_evt_pairs)) + return -EINVAL; + + dbuf_flush_fb(&front_info->dbuf_list, fb_cookie); + evtchnl = &front_info->evt_pairs[conn_idx].req; + + mutex_lock(&evtchnl->u.req.req_io_lock); + + spin_lock_irqsave(&front_info->io_lock, flags); + req = be_prepare_req(evtchnl, XENDISPL_OP_PG_FLIP); + req->op.pg_flip.fb_cookie = fb_cookie; + + ret = be_stream_do_io(evtchnl, req); + spin_unlock_irqrestore(&front_info->io_lock, flags); + + if (ret == 0) + ret = be_stream_wait_io(evtchnl); + + mutex_unlock(&evtchnl->u.req.req_io_lock); + return ret; +} + +void xen_drm_front_on_frame_done(struct xen_drm_front_info *front_info, + int conn_idx, u64 fb_cookie) +{ + struct xen_drm_front_drm_info *drm_info = front_info->drm_info; + + if (unlikely(conn_idx >= front_info->cfg.num_connectors)) + return; + + xen_drm_front_kms_on_frame_done(&drm_info->pipeline[conn_idx], + fb_cookie); +} + +static int xen_drm_drv_dumb_create(struct drm_file *filp, + struct drm_device *dev, + struct drm_mode_create_dumb *args) +{ + struct xen_drm_front_drm_info *drm_info = dev->dev_private; + struct drm_gem_object *obj; + int ret; + + /* + * Dumb creation is a two stage process: first we create a fully + * constructed GEM object which is communicated to the backend, and + * only after that we can create GEM's handle. This is done so, + * because of the possible races: once you create a handle it becomes + * immediately visible to user-space, so the latter can try accessing + * object without pages etc. + * For details also see drm_gem_handle_create + */ + args->pitch = DIV_ROUND_UP(args->width * args->bpp, 8); + args->size = args->pitch * args->height; + + obj = xen_drm_front_gem_create(dev, args->size); + if (IS_ERR_OR_NULL(obj)) { + ret = PTR_ERR(obj); + goto fail; + } + + /* + * In case of CONFIG_DRM_XEN_FRONTEND_CMA gem_obj is constructed + * via DRM CMA helpers and doesn't have ->pages allocated + * (xendrm_gem_get_pages will return NULL), but instead can provide + * sg table + */ + if (xen_drm_front_gem_get_pages(obj)) + ret = xen_drm_front_dbuf_create_from_pages(drm_info->front_info, + xen_drm_front_dbuf_to_cookie(obj), + args->width, args->height, args->bpp, + args->size, + xen_drm_front_gem_get_pages(obj)); + else + ret = xen_drm_front_dbuf_create_from_sgt(drm_info->front_info, + xen_drm_front_dbuf_to_cookie(obj), + args->width, args->height, args->bpp, + args->size, + xen_drm_front_gem_get_sg_table(obj)); + if (ret) + goto fail_backend; + + /* This is the tail of GEM object creation */ + ret = drm_gem_handle_create(filp, obj, &args->handle); + if (ret) + goto fail_handle; + + /* Drop reference from allocate - handle holds it now */ + drm_gem_object_put_unlocked(obj); + return 0; + +fail_handle: + xen_drm_front_dbuf_destroy(drm_info->front_info, + xen_drm_front_dbuf_to_cookie(obj)); +fail_backend: + /* drop reference from allocate */ + drm_gem_object_put_unlocked(obj); +fail: + DRM_ERROR("Failed to create dumb buffer: %d\n", ret); + return ret; +} + +static void xen_drm_drv_free_object_unlocked(struct drm_gem_object *obj) +{ + struct xen_drm_front_drm_info *drm_info = obj->dev->dev_private; + int idx; + + if (drm_dev_enter(obj->dev, &idx)) { + xen_drm_front_dbuf_destroy(drm_info->front_info, + xen_drm_front_dbuf_to_cookie(obj)); + drm_dev_exit(idx); + } else { + dbuf_free(&drm_info->front_info->dbuf_list, + xen_drm_front_dbuf_to_cookie(obj)); + } + + xen_drm_front_gem_free_object_unlocked(obj); +} + +static void xen_drm_drv_release(struct drm_device *dev) +{ + struct xen_drm_front_drm_info *drm_info = dev->dev_private; + struct xen_drm_front_info *front_info = drm_info->front_info; + + xen_drm_front_kms_fini(drm_info); + + drm_atomic_helper_shutdown(dev); + drm_mode_config_cleanup(dev); + + drm_dev_fini(dev); + kfree(dev); + + if (front_info->cfg.be_alloc) + xenbus_switch_state(front_info->xb_dev, + XenbusStateInitialising); + + kfree(drm_info); +} + +static const struct file_operations xen_drm_dev_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, +#ifdef CONFIG_COMPAT + .compat_ioctl = drm_compat_ioctl, +#endif + .poll = drm_poll, + .read = drm_read, + .llseek = no_llseek, +#ifdef CONFIG_DRM_XEN_FRONTEND_CMA + .mmap = drm_gem_cma_mmap, +#else + .mmap = xen_drm_front_gem_mmap, +#endif +}; + +static const struct vm_operations_struct xen_drm_drv_vm_ops = { + .open = drm_gem_vm_open, + .close = drm_gem_vm_close, +}; + +static struct drm_driver xen_drm_driver = { + .driver_features = DRIVER_GEM | DRIVER_MODESET | + DRIVER_PRIME | DRIVER_ATOMIC, + .release = xen_drm_drv_release, + .gem_vm_ops = &xen_drm_drv_vm_ops, + .gem_free_object_unlocked = xen_drm_drv_free_object_unlocked, + .prime_handle_to_fd = drm_gem_prime_handle_to_fd, + .prime_fd_to_handle = drm_gem_prime_fd_to_handle, + .gem_prime_import = drm_gem_prime_import, + .gem_prime_export = drm_gem_prime_export, + .gem_prime_import_sg_table = xen_drm_front_gem_import_sg_table, + .gem_prime_get_sg_table = xen_drm_front_gem_get_sg_table, + .dumb_create = xen_drm_drv_dumb_create, + .fops = &xen_drm_dev_fops, + .name = "xendrm-du", + .desc = "Xen PV DRM Display Unit", + .date = "20180221", + .major = 1, + .minor = 0, + +#ifdef CONFIG_DRM_XEN_FRONTEND_CMA + .gem_prime_vmap = drm_gem_cma_prime_vmap, + .gem_prime_vunmap = drm_gem_cma_prime_vunmap, + .gem_prime_mmap = drm_gem_cma_prime_mmap, +#else + .gem_prime_vmap = xen_drm_front_gem_prime_vmap, + .gem_prime_vunmap = xen_drm_front_gem_prime_vunmap, + .gem_prime_mmap = xen_drm_front_gem_prime_mmap, +#endif +}; + +static int xen_drm_drv_init(struct xen_drm_front_info *front_info) +{ + struct device *dev = &front_info->xb_dev->dev; + struct xen_drm_front_drm_info *drm_info; + struct drm_device *drm_dev; + int ret; + + DRM_INFO("Creating %s\n", xen_drm_driver.desc); + + drm_info = kzalloc(sizeof(*drm_info), GFP_KERNEL); + if (!drm_info) { + ret = -ENOMEM; + goto fail; + } + + drm_info->front_info = front_info; + front_info->drm_info = drm_info; + + drm_dev = drm_dev_alloc(&xen_drm_driver, dev); + if (!drm_dev) { + ret = -ENOMEM; + goto fail; + } + + drm_info->drm_dev = drm_dev; + + drm_dev->dev_private = drm_info; + + ret = xen_drm_front_kms_init(drm_info); + if (ret) { + DRM_ERROR("Failed to initialize DRM/KMS, ret %d\n", ret); + goto fail_modeset; + } + + ret = drm_dev_register(drm_dev, 0); + if (ret) + goto fail_register; + + DRM_INFO("Initialized %s %d.%d.%d %s on minor %d\n", + xen_drm_driver.name, xen_drm_driver.major, + xen_drm_driver.minor, xen_drm_driver.patchlevel, + xen_drm_driver.date, drm_dev->primary->index); + + return 0; + +fail_register: + drm_dev_unregister(drm_dev); +fail_modeset: + drm_kms_helper_poll_fini(drm_dev); + drm_mode_config_cleanup(drm_dev); +fail: + kfree(drm_info); + return ret; +} + +static void xen_drm_drv_fini(struct xen_drm_front_info *front_info) +{ + struct xen_drm_front_drm_info *drm_info = front_info->drm_info; + struct drm_device *dev; + + if (!drm_info) + return; + + dev = drm_info->drm_dev; + if (!dev) + return; + + /* Nothing to do if device is already unplugged */ + if (drm_dev_is_unplugged(dev)) + return; + + drm_kms_helper_poll_fini(dev); + drm_dev_unplug(dev); + + front_info->drm_info = NULL; + + xen_drm_front_evtchnl_free_all(front_info); + dbuf_free_all(&front_info->dbuf_list); + + /* + * If we are not using backend allocated buffers, then tell the + * backend we are ready to (re)initialize. Otherwise, wait for + * drm_driver.release. + */ + if (!front_info->cfg.be_alloc) + xenbus_switch_state(front_info->xb_dev, + XenbusStateInitialising); +} + +static int displback_initwait(struct xen_drm_front_info *front_info) +{ + struct xen_drm_front_cfg *cfg = &front_info->cfg; + int ret; + + cfg->front_info = front_info; + ret = xen_drm_front_cfg_card(front_info, cfg); + if (ret < 0) + return ret; + + DRM_INFO("Have %d conector(s)\n", cfg->num_connectors); + /* Create event channels for all connectors and publish */ + ret = xen_drm_front_evtchnl_create_all(front_info); + if (ret < 0) + return ret; + + return xen_drm_front_evtchnl_publish_all(front_info); +} + +static int displback_connect(struct xen_drm_front_info *front_info) +{ + xen_drm_front_evtchnl_set_state(front_info, EVTCHNL_STATE_CONNECTED); + return xen_drm_drv_init(front_info); +} + +static void displback_disconnect(struct xen_drm_front_info *front_info) +{ + if (!front_info->drm_info) + return; + + /* Tell the backend to wait until we release the DRM driver. */ + xenbus_switch_state(front_info->xb_dev, XenbusStateReconfiguring); + + xen_drm_drv_fini(front_info); +} + +static void displback_changed(struct xenbus_device *xb_dev, + enum xenbus_state backend_state) +{ + struct xen_drm_front_info *front_info = dev_get_drvdata(&xb_dev->dev); + int ret; + + DRM_DEBUG("Backend state is %s, front is %s\n", + xenbus_strstate(backend_state), + xenbus_strstate(xb_dev->state)); + + switch (backend_state) { + case XenbusStateReconfiguring: + /* fall through */ + case XenbusStateReconfigured: + /* fall through */ + case XenbusStateInitialised: + break; + + case XenbusStateInitialising: + if (xb_dev->state == XenbusStateReconfiguring) + break; + + /* recovering after backend unexpected closure */ + displback_disconnect(front_info); + break; + + case XenbusStateInitWait: + if (xb_dev->state == XenbusStateReconfiguring) + break; + + /* recovering after backend unexpected closure */ + displback_disconnect(front_info); + if (xb_dev->state != XenbusStateInitialising) + break; + + ret = displback_initwait(front_info); + if (ret < 0) + xenbus_dev_fatal(xb_dev, ret, "initializing frontend"); + else + xenbus_switch_state(xb_dev, XenbusStateInitialised); + break; + + case XenbusStateConnected: + if (xb_dev->state != XenbusStateInitialised) + break; + + ret = displback_connect(front_info); + if (ret < 0) { + displback_disconnect(front_info); + xenbus_dev_fatal(xb_dev, ret, "connecting backend"); + } else { + xenbus_switch_state(xb_dev, XenbusStateConnected); + } + break; + + case XenbusStateClosing: + /* + * in this state backend starts freeing resources, + * so let it go into closed state, so we can also + * remove ours + */ + break; + + case XenbusStateUnknown: + /* fall through */ + case XenbusStateClosed: + if (xb_dev->state == XenbusStateClosed) + break; + + displback_disconnect(front_info); + break; + } +} + +static int xen_drv_probe(struct xenbus_device *xb_dev, + const struct xenbus_device_id *id) +{ + struct xen_drm_front_info *front_info; + struct device *dev = &xb_dev->dev; + int ret; + + /* + * The device is not spawn from a device tree, so arch_setup_dma_ops + * is not called, thus leaving the device with dummy DMA ops. + * This makes the device return error on PRIME buffer import, which + * is not correct: to fix this call of_dma_configure() with a NULL + * node to set default DMA ops. + */ + dev->bus->force_dma = true; + dev->coherent_dma_mask = DMA_BIT_MASK(32); + ret = of_dma_configure(dev, NULL); + if (ret < 0) { + DRM_ERROR("Cannot setup DMA ops, ret %d", ret); + return ret; + } + + front_info = devm_kzalloc(&xb_dev->dev, + sizeof(*front_info), GFP_KERNEL); + if (!front_info) + return -ENOMEM; + + front_info->xb_dev = xb_dev; + spin_lock_init(&front_info->io_lock); + INIT_LIST_HEAD(&front_info->dbuf_list); + dev_set_drvdata(&xb_dev->dev, front_info); + + return xenbus_switch_state(xb_dev, XenbusStateInitialising); +} + +static int xen_drv_remove(struct xenbus_device *dev) +{ + struct xen_drm_front_info *front_info = dev_get_drvdata(&dev->dev); + int to = 100; + + xenbus_switch_state(dev, XenbusStateClosing); + + /* + * On driver removal it is disconnected from XenBus, + * so no backend state change events come via .otherend_changed + * callback. This prevents us from exiting gracefully, e.g. + * signaling the backend to free event channels, waiting for its + * state to change to XenbusStateClosed and cleaning at our end. + * Normally when front driver removed backend will finally go into + * XenbusStateInitWait state. + * + * Workaround: read backend's state manually and wait with time-out. + */ + while ((xenbus_read_unsigned(front_info->xb_dev->otherend, "state", + XenbusStateUnknown) != XenbusStateInitWait) && + to--) + msleep(10); + + if (!to) { + unsigned int state; + + state = xenbus_read_unsigned(front_info->xb_dev->otherend, + "state", XenbusStateUnknown); + DRM_ERROR("Backend state is %s while removing driver\n", + xenbus_strstate(state)); + } + + xen_drm_drv_fini(front_info); + xenbus_frontend_closed(dev); + return 0; +} + +static const struct xenbus_device_id xen_driver_ids[] = { + { XENDISPL_DRIVER_NAME }, + { "" } +}; + +static struct xenbus_driver xen_driver = { + .ids = xen_driver_ids, + .probe = xen_drv_probe, + .remove = xen_drv_remove, + .otherend_changed = displback_changed, +}; + +static int __init xen_drv_init(void) +{ + /* At the moment we only support case with XEN_PAGE_SIZE == PAGE_SIZE */ + if (XEN_PAGE_SIZE != PAGE_SIZE) { + DRM_ERROR(XENDISPL_DRIVER_NAME ": different kernel and Xen page sizes are not supported: XEN_PAGE_SIZE (%lu) != PAGE_SIZE (%lu)\n", + XEN_PAGE_SIZE, PAGE_SIZE); + return -ENODEV; + } + + if (!xen_domain()) + return -ENODEV; + + if (!xen_has_pv_devices()) + return -ENODEV; + + DRM_INFO("Registering XEN PV " XENDISPL_DRIVER_NAME "\n"); + return xenbus_register_frontend(&xen_driver); +} + +static void __exit xen_drv_fini(void) +{ + DRM_INFO("Unregistering XEN PV " XENDISPL_DRIVER_NAME "\n"); + xenbus_unregister_driver(&xen_driver); +} + +module_init(xen_drv_init); +module_exit(xen_drv_fini); + +MODULE_DESCRIPTION("Xen para-virtualized display device frontend"); +MODULE_LICENSE("GPL"); +MODULE_ALIAS("xen:" XENDISPL_DRIVER_NAME); diff --git a/drivers/gpu/drm/xen/xen_drm_front.h b/drivers/gpu/drm/xen/xen_drm_front.h new file mode 100644 index 0000000..16554b2 --- /dev/null +++ b/drivers/gpu/drm/xen/xen_drm_front.h @@ -0,0 +1,188 @@ +/* SPDX-License-Identifier: GPL-2.0 OR MIT */ + +/* + * Xen para-virtual DRM device + * + * Copyright (C) 2016-2018 EPAM Systems Inc. + * + * Author: Oleksandr Andrushchenko + */ + +#ifndef __XEN_DRM_FRONT_H_ +#define __XEN_DRM_FRONT_H_ + +#include +#include + +#include + +#include "xen_drm_front_cfg.h" + +/** + * DOC: Driver modes of operation in terms of display buffers used + * + * Depending on the requirements for the para-virtualized environment, namely + * requirements dictated by the accompanying DRM/(v)GPU drivers running in both + * host and guest environments, number of operating modes of para-virtualized + * display driver are supported: + * + * - display buffers can be allocated by either frontend driver or backend + * - display buffers can be allocated to be contiguous in memory or not + * + * Note! Frontend driver itself has no dependency on contiguous memory for + * its operation. + */ + +/** + * DOC: Buffers allocated by the frontend driver + * + * The below modes of operation are configured at compile-time via + * frontend driver's kernel configuration: + */ + +/** + * DOC: With GEM CMA helpers + * + * This use-case is useful when used with accompanying DRM/vGPU driver in + * guest domain which was designed to only work with contiguous buffers, + * e.g. DRM driver based on GEM CMA helpers: such drivers can only import + * contiguous PRIME buffers, thus requiring frontend driver to provide + * such. In order to implement this mode of operation para-virtualized + * frontend driver can be configured to use GEM CMA helpers. + */ + +/** + * DOC: Without GEM CMA helpers + * + * If accompanying drivers can cope with non-contiguous memory then, to + * lower pressure on CMA subsystem of the kernel, driver can allocate + * buffers from system memory. + * + * Note! If used with accompanying DRM/(v)GPU drivers this mode of operation + * may require IOMMU support on the platform, so accompanying DRM/vGPU + * hardware can still reach display buffer memory while importing PRIME + * buffers from the frontend driver. + */ + +/** + * DOC: Buffers allocated by the backend + * + * This mode of operation is run-time configured via guest domain configuration + * through XenStore entries. + * + * For systems which do not provide IOMMU support, but having specific + * requirements for display buffers it is possible to allocate such buffers + * at backend side and share those with the frontend. + * For example, if host domain is 1:1 mapped and has DRM/GPU hardware expecting + * physically contiguous memory, this allows implementing zero-copying + * use-cases. + * + * Note, while using this scenario the following should be considered: + * + * #. If guest domain dies then pages/grants received from the backend + * cannot be claimed back + * + * #. Misbehaving guest may send too many requests to the + * backend exhausting its grant references and memory + * (consider this from security POV) + */ + +/** + * DOC: Driver limitations + * + * #. Only primary plane without additional properties is supported. + * + * #. Only one video mode per connector supported which is configured + * via XenStore. + * + * #. All CRTCs operate at fixed frequency of 60Hz. + */ + +/* timeout in ms to wait for backend to respond */ +#define XEN_DRM_FRONT_WAIT_BACK_MS 3000 + +#ifndef GRANT_INVALID_REF +/* + * Note on usage of grant reference 0 as invalid grant reference: + * grant reference 0 is valid, but never exposed to a PV driver, + * because of the fact it is already in use/reserved by the PV console. + */ +#define GRANT_INVALID_REF 0 +#endif + +struct xen_drm_front_info { + struct xenbus_device *xb_dev; + struct xen_drm_front_drm_info *drm_info; + + /* to protect data between backend IO code and interrupt handler */ + spinlock_t io_lock; + + int num_evt_pairs; + struct xen_drm_front_evtchnl_pair *evt_pairs; + struct xen_drm_front_cfg cfg; + + /* display buffers */ + struct list_head dbuf_list; +}; + +struct xen_drm_front_drm_pipeline { + struct xen_drm_front_drm_info *drm_info; + + int index; + + struct drm_simple_display_pipe pipe; + + struct drm_connector conn; + /* These are only for connector mode checking */ + int width, height; + + struct drm_pending_vblank_event *pending_event; + + struct delayed_work pflip_to_worker; + + bool conn_connected; +}; + +struct xen_drm_front_drm_info { + struct xen_drm_front_info *front_info; + struct drm_device *drm_dev; + + struct xen_drm_front_drm_pipeline pipeline[XEN_DRM_FRONT_MAX_CRTCS]; +}; + +static inline u64 xen_drm_front_fb_to_cookie(struct drm_framebuffer *fb) +{ + return (u64)fb; +} + +static inline u64 xen_drm_front_dbuf_to_cookie(struct drm_gem_object *gem_obj) +{ + return (u64)gem_obj; +} + +int xen_drm_front_mode_set(struct xen_drm_front_drm_pipeline *pipeline, + u32 x, u32 y, u32 width, u32 height, + u32 bpp, u64 fb_cookie); + +int xen_drm_front_dbuf_create_from_sgt(struct xen_drm_front_info *front_info, + u64 dbuf_cookie, u32 width, u32 height, + u32 bpp, u64 size, struct sg_table *sgt); + +int xen_drm_front_dbuf_create_from_pages(struct xen_drm_front_info *front_info, + u64 dbuf_cookie, u32 width, u32 height, + u32 bpp, u64 size, struct page **pages); + +int xen_drm_front_fb_attach(struct xen_drm_front_info *front_info, + u64 dbuf_cookie, u64 fb_cookie, u32 width, + u32 height, u32 pixel_format); + +int xen_drm_front_fb_detach(struct xen_drm_front_info *front_info, + u64 fb_cookie); + +int xen_drm_front_page_flip(struct xen_drm_front_info *front_info, + int conn_idx, u64 fb_cookie); + +void xen_drm_front_on_frame_done(struct xen_drm_front_info *front_info, + int conn_idx, u64 fb_cookie); + +#endif /* __XEN_DRM_FRONT_H_ */ diff --git a/drivers/gpu/drm/xen/xen_drm_front_cfg.c b/drivers/gpu/drm/xen/xen_drm_front_cfg.c new file mode 100644 index 0000000..5baf2b9 --- /dev/null +++ b/drivers/gpu/drm/xen/xen_drm_front_cfg.c @@ -0,0 +1,77 @@ +// SPDX-License-Identifier: GPL-2.0 OR MIT + +/* + * Xen para-virtual DRM device + * + * Copyright (C) 2016-2018 EPAM Systems Inc. + * + * Author: Oleksandr Andrushchenko + */ + +#include + +#include + +#include +#include + +#include "xen_drm_front.h" +#include "xen_drm_front_cfg.h" + +static int cfg_connector(struct xen_drm_front_info *front_info, + struct xen_drm_front_cfg_connector *connector, + const char *path, int index) +{ + char *connector_path; + + connector_path = devm_kasprintf(&front_info->xb_dev->dev, + GFP_KERNEL, "%s/%d", path, index); + if (!connector_path) + return -ENOMEM; + + if (xenbus_scanf(XBT_NIL, connector_path, XENDISPL_FIELD_RESOLUTION, + "%d" XENDISPL_RESOLUTION_SEPARATOR "%d", + &connector->width, &connector->height) < 0) { + /* either no entry configured or wrong resolution set */ + connector->width = 0; + connector->height = 0; + return -EINVAL; + } + + connector->xenstore_path = connector_path; + + DRM_INFO("Connector %s: resolution %dx%d\n", + connector_path, connector->width, connector->height); + return 0; +} + +int xen_drm_front_cfg_card(struct xen_drm_front_info *front_info, + struct xen_drm_front_cfg *cfg) +{ + struct xenbus_device *xb_dev = front_info->xb_dev; + int ret, i; + + if (xenbus_read_unsigned(front_info->xb_dev->nodename, + XENDISPL_FIELD_BE_ALLOC, 0)) { + DRM_INFO("Backend can provide display buffers\n"); + cfg->be_alloc = true; + } + + cfg->num_connectors = 0; + for (i = 0; i < ARRAY_SIZE(cfg->connectors); i++) { + ret = cfg_connector(front_info, &cfg->connectors[i], + xb_dev->nodename, i); + if (ret < 0) + break; + cfg->num_connectors++; + } + + if (!cfg->num_connectors) { + DRM_ERROR("No connector(s) configured at %s\n", + xb_dev->nodename); + return -ENODEV; + } + + return 0; +} + diff --git a/drivers/gpu/drm/xen/xen_drm_front_cfg.h b/drivers/gpu/drm/xen/xen_drm_front_cfg.h new file mode 100644 index 0000000..aa8490b --- /dev/null +++ b/drivers/gpu/drm/xen/xen_drm_front_cfg.h @@ -0,0 +1,37 @@ +/* SPDX-License-Identifier: GPL-2.0 OR MIT */ + +/* + * Xen para-virtual DRM device + * + * Copyright (C) 2016-2018 EPAM Systems Inc. + * + * Author: Oleksandr Andrushchenko + */ + +#ifndef __XEN_DRM_FRONT_CFG_H_ +#define __XEN_DRM_FRONT_CFG_H_ + +#include + +#define XEN_DRM_FRONT_MAX_CRTCS 4 + +struct xen_drm_front_cfg_connector { + int width; + int height; + char *xenstore_path; +}; + +struct xen_drm_front_cfg { + struct xen_drm_front_info *front_info; + /* number of connectors in this configuration */ + int num_connectors; + /* connector configurations */ + struct xen_drm_front_cfg_connector connectors[XEN_DRM_FRONT_MAX_CRTCS]; + /* set if dumb buffers are allocated externally on backend side */ + bool be_alloc; +}; + +int xen_drm_front_cfg_card(struct xen_drm_front_info *front_info, + struct xen_drm_front_cfg *cfg); + +#endif /* __XEN_DRM_FRONT_CFG_H_ */ diff --git a/drivers/gpu/drm/xen/xen_drm_front_conn.c b/drivers/gpu/drm/xen/xen_drm_front_conn.c new file mode 100644 index 0000000..c91ae53 --- /dev/null +++ b/drivers/gpu/drm/xen/xen_drm_front_conn.c @@ -0,0 +1,115 @@ +// SPDX-License-Identifier: GPL-2.0 OR MIT + +/* + * Xen para-virtual DRM device + * + * Copyright (C) 2016-2018 EPAM Systems Inc. + * + * Author: Oleksandr Andrushchenko + */ + +#include +#include + +#include