diff options
author | Dave Airlie <airlied@redhat.com> | 2017-01-27 12:08:32 +1000 |
---|---|---|
committer | Dave Airlie <airlied@redhat.com> | 2017-01-27 12:08:32 +1000 |
commit | a7e2641aafe261bf70de01ff5fc68dea50468237 (patch) | |
tree | 1f4244a6a18574a5aa12bbc785bc285b0116a19e | |
parent | c4d79c22015564a9d922715434a0ec09d1deee30 (diff) | |
parent | add6329c728cdc705e076fc3d1663bc78bb50a66 (diff) | |
download | op-kernel-dev-a7e2641aafe261bf70de01ff5fc68dea50468237.zip op-kernel-dev-a7e2641aafe261bf70de01ff5fc68dea50468237.tar.gz |
Merge tag 'drm-intel-next-2017-01-23' of git://anongit.freedesktop.org/git/drm-intel into drm-next
Final block of feature work for 4.11:
- gen8 pd cleanup from Matthew Auld
- more cleanups for view/vma (Chris)
- dmc support on glk (Anusha Srivatsa)
- use core crc api (Tomue)
- track wedged requests using fence.error (Chris)
- lots of psr fixes (Nagaraju, Vathsala)
- dp mst support, acked for merging through drm-intel by Takashi
(Libin)
- huc loading support, including uapi for libva to use it (Anusha
Srivatsa)
* tag 'drm-intel-next-2017-01-23' of git://anongit.freedesktop.org/git/drm-intel: (111 commits)
drm/i915: Update DRIVER_DATE to 20170123
drm/i915: reinstate call to trace_i915_vma_bind
drm/i915: Assert that created vma has a whole number of pages
drm/i915: Assert the drm_mm_node is allocated when on the VM lists
drm/i915: Treat an error from i915_vma_instance() as unlikely
drm/i915: Reject vma creation larger than address space
drm/i915: Use common LRU inactive vma bumping for unpin_from_display
drm/i915: Do an unlocked wait before set-cache-level ioctl
drm/i915/huc: Assert that HuC vma is placed in GuC accessible range
drm/i915/huc: Avoid attempting to authenticate non-existent fw
drm/i915: Set adjustment to zero on Up/Down interrupts if freq is already max/min
drm/i915: Remove the double handling of 'flags from intel_mode_from_pipe_config()
drm/i915: Remove crtc->config usage from intel_modeset_readout_hw_state()
drm/i915: Release temporary load-detect state upon switching
drm/i915: Remove i915_gem_object_to_ggtt()
drm/i915: Remove i915_vma_create from VMA API
drm/i915: Add a check that the VMA instance we lookup matches the request
drm/i915: Rename some warts in the VMA API
drm/i915: Track pinned vma in intel_plane_state
drm/i915/get_params: Add HuC status to getparams
...
58 files changed, 3257 insertions, 1891 deletions
diff --git a/Documentation/sound/hd-audio/dp-mst.rst b/Documentation/sound/hd-audio/dp-mst.rst index 58b7243..1617459 100644 --- a/Documentation/sound/hd-audio/dp-mst.rst +++ b/Documentation/sound/hd-audio/dp-mst.rst @@ -19,6 +19,23 @@ PCM === To be added +Pin Initialization +================== +Each pin may have several device entries (virtual pins). On Intel platform, +the device entries number is dynamically changed. If DP MST hub is connected, +it is in DP MST mode, and the device entries number is 3. Otherwise, the +device entries number is 1. + +To simplify the implementation, all the device entries will be initialized +when bootup no matter whether it is in DP MST mode or not. + +Connection list +=============== +DP MST reuses connection list code. The code can be reused because +device entries on the same pin have the same connection list. + +This means DP MST gets the device entry connection list without the +device entry setting. Jack ==== diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile index 5196509..74ca2e8 100644 --- a/drivers/gpu/drm/i915/Makefile +++ b/drivers/gpu/drm/i915/Makefile @@ -56,7 +56,9 @@ i915-y += i915_cmd_parser.o \ # general-purpose microcontroller (GuC) support i915-y += intel_uc.o \ + intel_guc_log.o \ intel_guc_loader.o \ + intel_huc.o \ i915_guc_submission.o # autogenerated null render state diff --git a/drivers/gpu/drm/i915/gvt/aperture_gm.c b/drivers/gpu/drm/i915/gvt/aperture_gm.c index d7d828a..7311aea 100644 --- a/drivers/gpu/drm/i915/gvt/aperture_gm.c +++ b/drivers/gpu/drm/i915/gvt/aperture_gm.c @@ -41,47 +41,34 @@ static int alloc_gm(struct intel_vgpu *vgpu, bool high_gm) { struct intel_gvt *gvt = vgpu->gvt; struct drm_i915_private *dev_priv = gvt->dev_priv; - u32 alloc_flag, search_flag; + unsigned int flags; u64 start, end, size; struct drm_mm_node *node; - int retried = 0; int ret; if (high_gm) { - search_flag = DRM_MM_SEARCH_BELOW; - alloc_flag = DRM_MM_CREATE_TOP; node = &vgpu->gm.high_gm_node; size = vgpu_hidden_sz(vgpu); start = gvt_hidden_gmadr_base(gvt); end = gvt_hidden_gmadr_end(gvt); + flags = PIN_HIGH; } else { - search_flag = DRM_MM_SEARCH_DEFAULT; - alloc_flag = DRM_MM_CREATE_DEFAULT; node = &vgpu->gm.low_gm_node; size = vgpu_aperture_sz(vgpu); start = gvt_aperture_gmadr_base(gvt); end = gvt_aperture_gmadr_end(gvt); + flags = PIN_MAPPABLE; } mutex_lock(&dev_priv->drm.struct_mutex); -search_again: - ret = drm_mm_insert_node_in_range_generic(&dev_priv->ggtt.base.mm, - node, size, 4096, - I915_COLOR_UNEVICTABLE, - start, end, search_flag, - alloc_flag); - if (ret) { - ret = i915_gem_evict_something(&dev_priv->ggtt.base, - size, 4096, - I915_COLOR_UNEVICTABLE, - start, end, 0); - if (ret == 0 && ++retried < 3) - goto search_again; - - gvt_err("fail to alloc %s gm space from host, retried %d\n", - high_gm ? "high" : "low", retried); - } + ret = i915_gem_gtt_insert(&dev_priv->ggtt.base, node, + size, 4096, I915_COLOR_UNEVICTABLE, + start, end, flags); mutex_unlock(&dev_priv->drm.struct_mutex); + if (ret) + gvt_err("fail to alloc %s gm space from host\n", + high_gm ? "high" : "low"); + return ret; } diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 9d7b5a8..fa69d72 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -159,8 +159,35 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj) seq_printf(m, " (%sgtt offset: %08llx, size: %08llx", i915_vma_is_ggtt(vma) ? "g" : "pp", vma->node.start, vma->node.size); - if (i915_vma_is_ggtt(vma)) - seq_printf(m, ", type: %u", vma->ggtt_view.type); + if (i915_vma_is_ggtt(vma)) { + switch (vma->ggtt_view.type) { + case I915_GGTT_VIEW_NORMAL: + seq_puts(m, ", normal"); + break; + + case I915_GGTT_VIEW_PARTIAL: + seq_printf(m, ", partial [%08llx+%x]", + vma->ggtt_view.partial.offset << PAGE_SHIFT, + vma->ggtt_view.partial.size << PAGE_SHIFT); + break; + + case I915_GGTT_VIEW_ROTATED: + seq_printf(m, ", rotated [(%ux%u, stride=%u, offset=%u), (%ux%u, stride=%u, offset=%u)]", + vma->ggtt_view.rotated.plane[0].width, + vma->ggtt_view.rotated.plane[0].height, + vma->ggtt_view.rotated.plane[0].stride, + vma->ggtt_view.rotated.plane[0].offset, + vma->ggtt_view.rotated.plane[1].width, + vma->ggtt_view.rotated.plane[1].height, + vma->ggtt_view.rotated.plane[1].stride, + vma->ggtt_view.rotated.plane[1].offset); + break; + + default: + MISSING_CASE(vma->ggtt_view.type); + break; + } + } if (vma->fence) seq_printf(m, " , fence: %d%s", vma->fence->id, @@ -2325,10 +2352,40 @@ static int i915_llc(struct seq_file *m, void *data) return 0; } +static int i915_huc_load_status_info(struct seq_file *m, void *data) +{ + struct drm_i915_private *dev_priv = node_to_i915(m->private); + struct intel_uc_fw *huc_fw = &dev_priv->huc.fw; + + if (!HAS_HUC_UCODE(dev_priv)) + return 0; + + seq_puts(m, "HuC firmware status:\n"); + seq_printf(m, "\tpath: %s\n", huc_fw->path); + seq_printf(m, "\tfetch: %s\n", + intel_uc_fw_status_repr(huc_fw->fetch_status)); + seq_printf(m, "\tload: %s\n", + intel_uc_fw_status_repr(huc_fw->load_status)); + seq_printf(m, "\tversion wanted: %d.%d\n", + huc_fw->major_ver_wanted, huc_fw->minor_ver_wanted); + seq_printf(m, "\tversion found: %d.%d\n", + huc_fw->major_ver_found, huc_fw->minor_ver_found); + seq_printf(m, "\theader: offset is %d; size = %d\n", + huc_fw->header_offset, huc_fw->header_size); + seq_printf(m, "\tuCode: offset is %d; size = %d\n", + huc_fw->ucode_offset, huc_fw->ucode_size); + seq_printf(m, "\tRSA: offset is %d; size = %d\n", + huc_fw->rsa_offset, huc_fw->rsa_size); + + seq_printf(m, "\nHuC status 0x%08x:\n", I915_READ(HUC_STATUS2)); + + return 0; +} + static int i915_guc_load_status_info(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = node_to_i915(m->private); - struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; + struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; u32 tmp, i; if (!HAS_GUC_UCODE(dev_priv)) @@ -2336,15 +2393,15 @@ static int i915_guc_load_status_info(struct seq_file *m, void *data) seq_printf(m, "GuC firmware status:\n"); seq_printf(m, "\tpath: %s\n", - guc_fw->guc_fw_path); + guc_fw->path); seq_printf(m, "\tfetch: %s\n", - intel_guc_fw_status_repr(guc_fw->guc_fw_fetch_status)); + intel_uc_fw_status_repr(guc_fw->fetch_status)); seq_printf(m, "\tload: %s\n", - intel_guc_fw_status_repr(guc_fw->guc_fw_load_status)); + intel_uc_fw_status_repr(guc_fw->load_status)); seq_printf(m, "\tversion wanted: %d.%d\n", - guc_fw->guc_fw_major_wanted, guc_fw->guc_fw_minor_wanted); + guc_fw->major_ver_wanted, guc_fw->minor_ver_wanted); seq_printf(m, "\tversion found: %d.%d\n", - guc_fw->guc_fw_major_found, guc_fw->guc_fw_minor_found); + guc_fw->major_ver_found, guc_fw->minor_ver_found); seq_printf(m, "\theader: offset is %d; size = %d\n", guc_fw->header_offset, guc_fw->header_size); seq_printf(m, "\tuCode: offset is %d; size = %d\n", @@ -2532,6 +2589,29 @@ DEFINE_SIMPLE_ATTRIBUTE(i915_guc_log_control_fops, i915_guc_log_control_get, i915_guc_log_control_set, "%lld\n"); +static const char *psr2_live_status(u32 val) +{ + static const char * const live_status[] = { + "IDLE", + "CAPTURE", + "CAPTURE_FS", + "SLEEP", + "BUFON_FW", + "ML_UP", + "SU_STANDBY", + "FAST_SLEEP", + "DEEP_SLEEP", + "BUF_ON", + "TG_ON" + }; + + val = (val & EDP_PSR2_STATUS_STATE_MASK) >> EDP_PSR2_STATUS_STATE_SHIFT; + if (val < ARRAY_SIZE(live_status)) + return live_status[val]; + + return "unknown"; +} + static int i915_edp_psr_status(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = node_to_i915(m->private); @@ -2606,6 +2686,12 @@ 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) { + u32 psr2 = I915_READ(EDP_PSR2_STATUS_CTL); + + seq_printf(m, "EDP_PSR2_STATUS_CTL: %x [%s]\n", + psr2, psr2_live_status(psr2)); + } mutex_unlock(&dev_priv->psr.lock); intel_runtime_pm_put(dev_priv); @@ -4553,6 +4639,7 @@ static const struct drm_info_list i915_debugfs_list[] = { {"i915_guc_info", i915_guc_info, 0}, {"i915_guc_load_status", i915_guc_load_status_info, 0}, {"i915_guc_log_dump", i915_guc_log_dump, 0}, + {"i915_huc_load_status", i915_huc_load_status_info, 0}, {"i915_frequency_info", i915_frequency_info, 0}, {"i915_hangcheck_info", i915_hangcheck_info, 0}, {"i915_drpc_info", i915_drpc_info, 0}, diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 4d22b4b..4ae69eb 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -49,6 +49,7 @@ #include "i915_trace.h" #include "i915_vgpu.h" #include "intel_drv.h" +#include "intel_uc.h" static struct drm_driver driver; @@ -315,6 +316,12 @@ static int i915_getparam(struct drm_device *dev, void *data, case I915_PARAM_MIN_EU_IN_POOL: value = INTEL_INFO(dev_priv)->sseu.min_eu_in_pool; break; + case I915_PARAM_HUC_STATUS: + /* The register is already force-woken. We dont need + * any rpm here + */ + value = I915_READ(HUC_STATUS2) & HUC_FW_VERIFIED; + break; case I915_PARAM_MMAP_GTT_VERSION: /* Though we've started our numbering from 1, and so class all * earlier versions as 0, in effect their value is undefined as @@ -599,6 +606,7 @@ static int i915_load_modeset_init(struct drm_device *dev) if (ret) goto cleanup_irq; + intel_huc_init(dev_priv); intel_guc_init(dev_priv); ret = i915_gem_init(dev_priv); @@ -627,6 +635,7 @@ cleanup_gem: i915_gem_fini(dev_priv); cleanup_irq: intel_guc_fini(dev_priv); + intel_huc_fini(dev_priv); drm_irq_uninstall(dev); intel_teardown_gmbus(dev_priv); cleanup_csr: @@ -1114,7 +1123,7 @@ 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_register(dev_priv); + i915_guc_log_register(dev_priv); i915_setup_sysfs(dev_priv); /* Depends on sysfs having been initialized */ @@ -1158,7 +1167,7 @@ static void i915_driver_unregister(struct drm_i915_private *dev_priv) i915_perf_unregister(dev_priv); i915_teardown_sysfs(dev_priv); - i915_guc_unregister(dev_priv); + i915_guc_log_unregister(dev_priv); i915_debugfs_unregister(dev_priv); drm_dev_unregister(&dev_priv->drm); @@ -1314,6 +1323,7 @@ void i915_driver_unload(struct drm_device *dev) drain_workqueue(dev_priv->wq); intel_guc_fini(dev_priv); + intel_huc_fini(dev_priv); i915_gem_fini(dev_priv); intel_fbc_cleanup_cfb(dev_priv); @@ -1471,7 +1481,7 @@ static int i915_drm_suspend_late(struct drm_device *dev, bool hibernation) intel_display_set_init_power(dev_priv, false); - fw_csr = !IS_BROXTON(dev_priv) && + fw_csr = !IS_GEN9_LP(dev_priv) && suspend_to_idle(dev_priv) && dev_priv->csr.dmc_payload; /* * In case of firmware assisted context save/restore don't manually @@ -1484,7 +1494,7 @@ static int i915_drm_suspend_late(struct drm_device *dev, bool hibernation) intel_power_domains_suspend(dev_priv); ret = 0; - if (IS_BROXTON(dev_priv)) + if (IS_GEN9_LP(dev_priv)) bxt_enable_dc9(dev_priv); else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) hsw_enable_pc8(dev_priv); @@ -1692,7 +1702,7 @@ static int i915_drm_resume_early(struct drm_device *dev) intel_uncore_early_sanitize(dev_priv, true); - if (IS_BROXTON(dev_priv)) { + if (IS_GEN9_LP(dev_priv)) { if (!dev_priv->suspended_to_idle) gen9_sanitize_dc_state(dev_priv); bxt_disable_dc9(dev_priv); @@ -1702,7 +1712,7 @@ static int i915_drm_resume_early(struct drm_device *dev) intel_uncore_sanitize(dev_priv); - if (IS_BROXTON(dev_priv) || + if (IS_GEN9_LP(dev_priv) || !(dev_priv->suspended_to_idle && dev_priv->csr.dmc_payload)) intel_power_domains_init_hw(dev_priv, true); @@ -1728,25 +1738,9 @@ static int i915_resume_switcheroo(struct drm_device *dev) return i915_drm_resume(dev); } -static void disable_engines_irq(struct drm_i915_private *dev_priv) -{ - struct intel_engine_cs *engine; - enum intel_engine_id id; - - /* Ensure irq handler finishes, and not run again. */ - disable_irq(dev_priv->drm.irq); - for_each_engine(engine, dev_priv, id) - tasklet_kill(&engine->irq_tasklet); -} - -static void enable_engines_irq(struct drm_i915_private *dev_priv) -{ - enable_irq(dev_priv->drm.irq); -} - /** * i915_reset - reset chip after a hang - * @dev: drm device to reset + * @dev_priv: device private to reset * * Reset the chip. Useful if a hang is detected. Marks the device as wedged * on failure. @@ -1776,12 +1770,15 @@ void i915_reset(struct drm_i915_private *dev_priv) error->reset_count++; pr_notice("drm/i915: Resetting chip after gpu hang\n"); - i915_gem_reset_prepare(dev_priv); + disable_irq(dev_priv->drm.irq); + ret = i915_gem_reset_prepare(dev_priv); + if (ret) { + DRM_ERROR("GPU recovery failed\n"); + intel_gpu_reset(dev_priv, ALL_ENGINES); + goto error; + } - disable_engines_irq(dev_priv); ret = intel_gpu_reset(dev_priv, ALL_ENGINES); - enable_engines_irq(dev_priv); - if (ret) { if (ret != -ENODEV) DRM_ERROR("Failed to reset chip: %i\n", ret); @@ -1816,6 +1813,7 @@ void i915_reset(struct drm_i915_private *dev_priv) i915_queue_hangcheck(dev_priv); wakeup: + enable_irq(dev_priv->drm.irq); wake_up_bit(&error->flags, I915_RESET_IN_PROGRESS); return; @@ -2326,7 +2324,7 @@ static int intel_runtime_suspend(struct device *kdev) intel_runtime_pm_disable_interrupts(dev_priv); ret = 0; - if (IS_BROXTON(dev_priv)) { + if (IS_GEN9_LP(dev_priv)) { bxt_display_core_uninit(dev_priv); bxt_enable_dc9(dev_priv); } else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) { @@ -2411,7 +2409,7 @@ static int intel_runtime_resume(struct device *kdev) if (IS_GEN6(dev_priv)) intel_init_pch_refclk(dev_priv); - if (IS_BROXTON(dev_priv)) { + if (IS_GEN9_LP(dev_priv)) { bxt_disable_dc9(dev_priv); bxt_display_core_init(dev_priv, true); if (dev_priv->csr.dmc_payload && @@ -2549,8 +2547,8 @@ static const struct drm_ioctl_desc i915_ioctls[] = { DRM_IOCTL_DEF_DRV(I915_GEM_MMAP_GTT, i915_gem_mmap_gtt_ioctl, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(I915_GEM_SET_DOMAIN, i915_gem_set_domain_ioctl, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(I915_GEM_SW_FINISH, i915_gem_sw_finish_ioctl, DRM_RENDER_ALLOW), - DRM_IOCTL_DEF_DRV(I915_GEM_SET_TILING, i915_gem_set_tiling, DRM_RENDER_ALLOW), - DRM_IOCTL_DEF_DRV(I915_GEM_GET_TILING, i915_gem_get_tiling, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(I915_GEM_SET_TILING, i915_gem_set_tiling_ioctl, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(I915_GEM_GET_TILING, i915_gem_get_tiling_ioctl, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(I915_GEM_GET_APERTURE, i915_gem_get_aperture_ioctl, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(I915_GET_PIPE_FROM_CRTC_ID, intel_get_pipe_from_crtc_id, 0), DRM_IOCTL_DEF_DRV(I915_GEM_MADVISE, i915_gem_madvise_ioctl, DRM_RENDER_ALLOW), diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 52d01be..ead3b64 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -78,8 +78,8 @@ #define DRIVER_NAME "i915" #define DRIVER_DESC "Intel Graphics" -#define DRIVER_DATE "20170109" -#define DRIVER_TIMESTAMP 1483953121 +#define DRIVER_DATE "20170123" +#define DRIVER_TIMESTAMP 1485156432 #undef WARN_ON /* Many gcc seem to no see through this and fall over :( */ @@ -1069,6 +1069,8 @@ struct intel_fbc { struct work_struct underrun_work; struct intel_fbc_state_cache { + struct i915_vma *vma; + struct { unsigned int mode_flags; uint32_t hsw_bdw_pixel_rate; @@ -1082,15 +1084,14 @@ struct intel_fbc { } plane; struct { - u64 ilk_ggtt_offset; const struct drm_format_info *format; unsigned int stride; - int fence_reg; - unsigned int tiling_mode; } fb; } state_cache; struct intel_fbc_reg_params { + struct i915_vma *vma; + struct { enum pipe pipe; enum plane plane; @@ -1098,10 +1099,8 @@ struct intel_fbc { } crtc; struct { - u64 ggtt_offset; const struct drm_format_info *format; unsigned int stride; - int fence_reg; } fb; int cfb_size; @@ -1154,6 +1153,9 @@ struct i915_psr { bool psr2_support; bool aux_frame_sync; bool link_standby; + bool y_cord_support; + bool colorimetry_support; + bool alpm; }; enum intel_pch { @@ -1809,6 +1811,7 @@ struct intel_pipe_crc { enum intel_pipe_crc_source source; int head, tail; wait_queue_head_t wq; + int skipped; }; struct i915_frontbuffer_tracking { @@ -2069,6 +2072,7 @@ struct drm_i915_private { struct intel_gvt *gvt; + struct intel_huc huc; struct intel_guc guc; struct intel_csr csr; @@ -2843,6 +2847,7 @@ intel_info(const struct drm_i915_private *dev_priv) #define HAS_GUC(dev_priv) ((dev_priv)->info.has_guc) #define HAS_GUC_UCODE(dev_priv) (HAS_GUC(dev_priv)) #define HAS_GUC_SCHED(dev_priv) (HAS_GUC(dev_priv)) +#define HAS_HUC_UCODE(dev_priv) (HAS_GUC(dev_priv)) #define HAS_RESOURCE_STREAMER(dev_priv) ((dev_priv)->info.has_resource_streamer) @@ -3101,10 +3106,10 @@ int i915_gem_throttle_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); int i915_gem_madvise_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); -int i915_gem_set_tiling(struct drm_device *dev, void *data, - struct drm_file *file_priv); -int i915_gem_get_tiling(struct drm_device *dev, void *data, - struct drm_file *file_priv); +int i915_gem_set_tiling_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); +int i915_gem_get_tiling_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); void i915_gem_init_userptr(struct drm_i915_private *dev_priv); int i915_gem_userptr_ioctl(struct drm_device *dev, void *data, struct drm_file *file); @@ -3323,7 +3328,7 @@ static inline u32 i915_reset_count(struct i915_gpu_error *error) return READ_ONCE(error->reset_count); } -void i915_gem_reset_prepare(struct drm_i915_private *dev_priv); +int i915_gem_reset_prepare(struct drm_i915_private *dev_priv); void i915_gem_reset_finish(struct drm_i915_private *dev_priv); void i915_gem_set_wedged(struct drm_i915_private *dev_priv); void i915_gem_clflush_object(struct drm_i915_gem_object *obj, bool force); @@ -3360,11 +3365,6 @@ int i915_gem_object_attach_phys(struct drm_i915_gem_object *obj, int i915_gem_open(struct drm_device *dev, struct drm_file *file); void i915_gem_release(struct drm_device *dev, struct drm_file *file); -u64 i915_gem_get_ggtt_size(struct drm_i915_private *dev_priv, u64 size, - int tiling_mode); -u64 i915_gem_get_ggtt_alignment(struct drm_i915_private *dev_priv, u64 size, - int tiling_mode, bool fenced); - int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj, enum i915_cache_level cache_level); @@ -3374,36 +3374,12 @@ struct drm_gem_object *i915_gem_prime_import(struct drm_device *dev, struct dma_buf *i915_gem_prime_export(struct drm_device *dev, struct drm_gem_object *gem_obj, int flags); -struct i915_vma * -i915_gem_obj_to_vma(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view); - -struct i915_vma * -i915_gem_obj_lookup_or_create_vma(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view); - static inline struct i915_hw_ppgtt * i915_vm_to_ppgtt(struct i915_address_space *vm) { return container_of(vm, struct i915_hw_ppgtt, base); } -static inline struct i915_vma * -i915_gem_object_to_ggtt(struct drm_i915_gem_object *obj, - const struct i915_ggtt_view *view) -{ - return i915_gem_obj_to_vma(obj, &to_i915(obj->base.dev)->ggtt.base, view); -} - -static inline unsigned long -i915_gem_object_ggtt_offset(struct drm_i915_gem_object *o, - const struct i915_ggtt_view *view) -{ - return i915_ggtt_offset(i915_gem_object_to_ggtt(o, view)); -} - /* i915_gem_fence_reg.c */ int __must_check i915_vma_get_fence(struct i915_vma *vma); int __must_check i915_vma_put_fence(struct i915_vma *vma); @@ -3471,8 +3447,9 @@ int __must_check i915_gem_evict_something(struct i915_address_space *vm, unsigned cache_level, u64 start, u64 end, unsigned flags); -int __must_check i915_gem_evict_for_vma(struct i915_vma *vma, - unsigned int flags); +int __must_check i915_gem_evict_for_node(struct i915_address_space *vm, + struct drm_mm_node *node, + unsigned int flags); int i915_gem_evict_vm(struct i915_address_space *vm, bool do_idle); /* belongs in i915_gem_gtt.h */ @@ -3506,7 +3483,7 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv /* i915_gem_internal.c */ struct drm_i915_gem_object * i915_gem_object_create_internal(struct drm_i915_private *dev_priv, - unsigned int size); + phys_addr_t size); /* i915_gem_shrinker.c */ unsigned long i915_gem_shrink(struct drm_i915_private *dev_priv, @@ -3531,6 +3508,11 @@ static inline bool i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_objec i915_gem_object_is_tiled(obj); } +u32 i915_gem_fence_size(struct drm_i915_private *dev_priv, u32 size, + unsigned int tiling, unsigned int stride); +u32 i915_gem_fence_alignment(struct drm_i915_private *dev_priv, u32 size, + unsigned int tiling, unsigned int stride); + /* i915_debugfs.c */ #ifdef CONFIG_DEBUG_FS int i915_debugfs_register(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 dc00d9a..a07b627 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -1696,12 +1696,7 @@ i915_gem_mmap_ioctl(struct drm_device *dev, void *data, static unsigned int tile_row_pages(struct drm_i915_gem_object *obj) { - u64 size; - - size = i915_gem_object_get_stride(obj); - size *= i915_gem_object_get_tiling(obj) == I915_TILING_Y ? 32 : 8; - - return size >> PAGE_SHIFT; + return i915_gem_object_get_tile_row_size(obj) >> PAGE_SHIFT; } /** @@ -1754,6 +1749,29 @@ int i915_gem_mmap_gtt_version(void) return 1; } +static inline struct i915_ggtt_view +compute_partial_view(struct drm_i915_gem_object *obj, + pgoff_t page_offset, + unsigned int chunk) +{ + struct i915_ggtt_view view; + + if (i915_gem_object_is_tiled(obj)) + chunk = roundup(chunk, tile_row_pages(obj)); + + view.type = I915_GGTT_VIEW_PARTIAL; + view.partial.offset = rounddown(page_offset, chunk); + view.partial.size = + min_t(unsigned int, chunk, + (obj->base.size >> PAGE_SHIFT) - view.partial.offset); + + /* If the partial covers the entire object, just create a normal VMA. */ + if (chunk >= obj->base.size >> PAGE_SHIFT) + view.type = I915_GGTT_VIEW_NORMAL; + + return view; +} + /** * i915_gem_fault - fault a page into the GTT * @area: CPU VMA in question @@ -1830,26 +1848,9 @@ int i915_gem_fault(struct vm_area_struct *area, struct vm_fault *vmf) /* Now pin it into the GTT as needed */ vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, flags); if (IS_ERR(vma)) { - struct i915_ggtt_view view; - unsigned int chunk_size; - /* Use a partial view if it is bigger than available space */ - chunk_size = MIN_CHUNK_PAGES; - if (i915_gem_object_is_tiled(obj)) - chunk_size = roundup(chunk_size, tile_row_pages(obj)); - - memset(&view, 0, sizeof(view)); - view.type = I915_GGTT_VIEW_PARTIAL; - view.params.partial.offset = rounddown(page_offset, chunk_size); - view.params.partial.size = - min_t(unsigned int, chunk_size, - vma_pages(area) - view.params.partial.offset); - - /* If the partial covers the entire object, just create a - * normal VMA. - */ - if (chunk_size >= obj->base.size >> PAGE_SHIFT) - view.type = I915_GGTT_VIEW_NORMAL; + struct i915_ggtt_view view = + compute_partial_view(obj, page_offset, MIN_CHUNK_PAGES); /* Userspace is now writing through an untracked VMA, abandon * all hope that the hardware is able to track future writes. @@ -1878,7 +1879,7 @@ int i915_gem_fault(struct vm_area_struct *area, struct vm_fault *vmf) /* Finally, remap it using the new GTT offset */ ret = remap_io_mapping(area, - area->vm_start + (vma->ggtt_view.params.partial.offset << PAGE_SHIFT), + area->vm_start + (vma->ggtt_view.partial.offset << PAGE_SHIFT), (ggtt->mappable_base + vma->node.start) >> PAGE_SHIFT, min_t(u64, vma->size, area->vm_end - area->vm_start), &ggtt->mappable); @@ -2021,69 +2022,6 @@ void i915_gem_runtime_suspend(struct drm_i915_private *dev_priv) } } -/** - * i915_gem_get_ggtt_size - return required global GTT size for an object - * @dev_priv: i915 device - * @size: object size - * @tiling_mode: tiling mode - * - * Return the required global GTT size for an object, taking into account - * potential fence register mapping. - */ -u64 i915_gem_get_ggtt_size(struct drm_i915_private *dev_priv, - u64 size, int tiling_mode) -{ - u64 ggtt_size; - - GEM_BUG_ON(size == 0); - - if (INTEL_GEN(dev_priv) >= 4 || - tiling_mode == I915_TILING_NONE) - return size; - - /* Previous chips need a power-of-two fence region when tiling */ - if (IS_GEN3(dev_priv)) - ggtt_size = 1024*1024; - else - ggtt_size = 512*1024; - - while (ggtt_size < size) - ggtt_size <<= 1; - - return ggtt_size; -} - -/** - * i915_gem_get_ggtt_alignment - return required global GTT alignment - * @dev_priv: i915 device - * @size: object size - * @tiling_mode: tiling mode - * @fenced: is fenced alignment required or not - * - * Return the required global GTT alignment for an object, taking into account - * potential fence register mapping. - */ -u64 i915_gem_get_ggtt_alignment(struct drm_i915_private *dev_priv, u64 size, - int tiling_mode, bool fenced) -{ - GEM_BUG_ON(size == 0); - - /* - * Minimum alignment is 4k (GTT page size), but might be greater - * if a fence register is needed for the object. - */ - if (INTEL_GEN(dev_priv) >= 4 || - (!fenced && (IS_G33(dev_priv) || IS_PINEVIEW(dev_priv))) || - tiling_mode == I915_TILING_NONE) - return 4096; - - /* - * Previous chips need to be aligned to the size of the smallest - * fence register that can contain the object. - */ - return i915_gem_get_ggtt_size(dev_priv, size, tiling_mode); -} - static int i915_gem_object_create_mmap_offset(struct drm_i915_gem_object *obj) { struct drm_i915_private *dev_priv = to_i915(obj->base.dev); @@ -2666,13 +2604,52 @@ i915_gem_find_active_request(struct intel_engine_cs *engine) if (__i915_gem_request_completed(request)) continue; + GEM_BUG_ON(request->engine != engine); return request; } return NULL; } -static void reset_request(struct drm_i915_gem_request *request) +static bool engine_stalled(struct intel_engine_cs *engine) +{ + if (!engine->hangcheck.stalled) + return false; + + /* Check for possible seqno movement after hang declaration */ + if (engine->hangcheck.seqno != intel_engine_get_seqno(engine)) { + DRM_DEBUG_DRIVER("%s pardoned\n", engine->name); + return false; + } + + return true; +} + +int i915_gem_reset_prepare(struct drm_i915_private *dev_priv) +{ + struct intel_engine_cs *engine; + enum intel_engine_id id; + int err = 0; + + /* Ensure irq handler finishes, and not run again. */ + for_each_engine(engine, dev_priv, id) { + struct drm_i915_gem_request *request; + + tasklet_kill(&engine->irq_tasklet); + + if (engine_stalled(engine)) { + request = i915_gem_find_active_request(engine); + if (request && request->fence.error == -EIO) + err = -EIO; /* Previous reset failed! */ + } + } + + i915_gem_revoke_fences(dev_priv); + + return err; +} + +static void skip_request(struct drm_i915_gem_request *request) { void *vaddr = request->ring->vaddr; u32 head; @@ -2687,20 +2664,74 @@ static void reset_request(struct drm_i915_gem_request *request) head = 0; } memset(vaddr + head, 0, request->postfix - head); + + dma_fence_set_error(&request->fence, -EIO); } -void i915_gem_reset_prepare(struct drm_i915_private *dev_priv) +static void engine_skip_context(struct drm_i915_gem_request *request) { - i915_gem_revoke_fences(dev_priv); + struct intel_engine_cs *engine = request->engine; + struct i915_gem_context *hung_ctx = request->ctx; + struct intel_timeline *timeline; + unsigned long flags; + + timeline = i915_gem_context_lookup_timeline(hung_ctx, engine); + + spin_lock_irqsave(&engine->timeline->lock, flags); + spin_lock(&timeline->lock); + + list_for_each_entry_continue(request, &engine->timeline->requests, link) + if (request->ctx == hung_ctx) + skip_request(request); + + list_for_each_entry(request, &timeline->requests, link) + skip_request(request); + + spin_unlock(&timeline->lock); + spin_unlock_irqrestore(&engine->timeline->lock, flags); +} + +/* Returns true if the request was guilty of hang */ +static bool i915_gem_reset_request(struct drm_i915_gem_request *request) +{ + /* Read once and return the resolution */ + const bool guilty = engine_stalled(request->engine); + + /* The guilty request will get skipped on a hung engine. + * + * Users of client default contexts do not rely on logical + * state preserved between batches so it is safe to execute + * queued requests following the hang. Non default contexts + * rely on preserved state, so skipping a batch loses the + * evolution of the state and it needs to be considered corrupted. + * Executing more queued batches on top of corrupted state is + * risky. But we take the risk by trying to advance through + * the queued requests in order to make the client behaviour + * more predictable around resets, by not throwing away random + * amount of batches it has prepared for execution. Sophisticated + * clients can use gem_reset_stats_ioctl and dma fence status + * (exported via sync_file info ioctl on explicit fences) to observe + * when it loses the context state and should rebuild accordingly. + * + * The context ban, and ultimately the client ban, mechanism are safety + * valves if client submission ends up resulting in nothing more than + * subsequent hangs. + */ + + if (guilty) { + i915_gem_context_mark_guilty(request->ctx); + skip_request(request); + } else { + i915_gem_context_mark_innocent(request->ctx); + dma_fence_set_error(&request->fence, -EAGAIN); + } + + return guilty; } static void i915_gem_reset_engine(struct intel_engine_cs *engine) { struct drm_i915_gem_request *request; - struct i915_gem_context *hung_ctx; - struct intel_timeline *timeline; - unsigned long flags; - bool ring_hung; if (engine->irq_seqno_barrier) engine->irq_seqno_barrier(engine); @@ -2709,22 +2740,7 @@ static void i915_gem_reset_engine(struct intel_engine_cs *engine) if (!request) return; - hung_ctx = request->ctx; - - ring_hung = engine->hangcheck.stalled; - if (engine->hangcheck.seqno != intel_engine_get_seqno(engine)) { - DRM_DEBUG_DRIVER("%s pardoned, was guilty? %s\n", - engine->name, - yesno(ring_hung)); - ring_hung = false; - } - - if (ring_hung) - i915_gem_context_mark_guilty(hung_ctx); - else - i915_gem_context_mark_innocent(hung_ctx); - - if (!ring_hung) + if (!i915_gem_reset_request(request)) return; DRM_DEBUG_DRIVER("resetting %s to restart from tail of request 0x%x\n", @@ -2734,34 +2750,8 @@ static void i915_gem_reset_engine(struct intel_engine_cs *engine) engine->reset_hw(engine, request); /* If this context is now banned, skip all of its pending requests. */ - if (!i915_gem_context_is_banned(hung_ctx)) - return; - - /* Users of the default context do not rely on logical state - * preserved between batches. They have to emit full state on - * every batch and so it is safe to execute queued requests following - * the hang. - * - * Other contexts preserve state, now corrupt. We want to skip all - * queued requests that reference the corrupt context. - */ - if (i915_gem_context_is_default(hung_ctx)) - return; - - timeline = i915_gem_context_lookup_timeline(hung_ctx, engine); - - spin_lock_irqsave(&engine->timeline->lock, flags); - spin_lock(&timeline->lock); - - list_for_each_entry_continue(request, &engine->timeline->requests, link) - if (request->ctx == hung_ctx) - reset_request(request); - - list_for_each_entry(request, &timeline->requests, link) - reset_request(request); - - spin_unlock(&timeline->lock); - spin_unlock_irqrestore(&engine->timeline->lock, flags); + if (i915_gem_context_is_banned(request->ctx)) + engine_skip_context(request); } void i915_gem_reset_finish(struct drm_i915_private *dev_priv) @@ -2788,12 +2778,16 @@ void i915_gem_reset_finish(struct drm_i915_private *dev_priv) static void nop_submit_request(struct drm_i915_gem_request *request) { + dma_fence_set_error(&request->fence, -EIO); i915_gem_request_submit(request); intel_engine_init_global_seqno(request->engine, request->global_seqno); } -static void i915_gem_cleanup_engine(struct intel_engine_cs *engine) +static void engine_set_wedged(struct intel_engine_cs *engine) { + struct drm_i915_gem_request *request; + unsigned long flags; + /* We need to be sure that no thread is running the old callback as * we install the nop handler (otherwise we would submit a request * to hardware that will never complete). In order to prevent this @@ -2802,6 +2796,12 @@ static void i915_gem_cleanup_engine(struct intel_engine_cs *engine) */ engine->submit_request = nop_submit_request; + /* Mark all executing requests as skipped */ + spin_lock_irqsave(&engine->timeline->lock, flags); + list_for_each_entry(request, &engine->timeline->requests, link) + dma_fence_set_error(&request->fence, -EIO); + spin_unlock_irqrestore(&engine->timeline->lock, flags); + /* Mark all pending requests as complete so that any concurrent * (lockless) lookup doesn't try and wait upon the request as we * reset it. @@ -2837,7 +2837,7 @@ static int __i915_gem_set_wedged_BKL(void *data) enum intel_engine_id id; for_each_engine(engine, i915, id) - i915_gem_cleanup_engine(engine); + engine_set_wedged(engine); return 0; } @@ -3397,7 +3397,7 @@ int i915_gem_set_caching_ioctl(struct drm_device *dev, void *data, struct drm_i915_gem_caching *args = data; struct drm_i915_gem_object *obj; enum i915_cache_level level; - int ret; + int ret = 0; switch (args->caching) { case I915_CACHING_NONE: @@ -3422,20 +3422,29 @@ int i915_gem_set_caching_ioctl(struct drm_device *dev, void *data, return -EINVAL; } - ret = i915_mutex_lock_interruptible(dev); + obj = i915_gem_object_lookup(file, args->handle); + if (!obj) + return -ENOENT; + + if (obj->cache_level == level) + goto out; + + ret = i915_gem_object_wait(obj, + I915_WAIT_INTERRUPTIBLE, + MAX_SCHEDULE_TIMEOUT, + to_rps_client(file)); if (ret) - return ret; + goto out; - obj = i915_gem_object_lookup(file, args->handle); - if (!obj) { - ret = -ENOENT; - goto unlock; - } + ret = i915_mutex_lock_interruptible(dev); + if (ret) + goto out; ret = i915_gem_object_set_cache_level(obj, level); - i915_gem_object_put(obj); -unlock: mutex_unlock(&dev->struct_mutex); + +out: + i915_gem_object_put(obj); return ret; } @@ -3485,7 +3494,7 @@ i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj, * try to preserve the existing ABI). */ vma = ERR_PTR(-ENOSPC); - if (view->type == I915_GGTT_VIEW_NORMAL) + if (!view || view->type == I915_GGTT_VIEW_NORMAL) vma = i915_gem_object_ggtt_pin(obj, view, 0, alignment, PIN_MAPPABLE | PIN_NONBLOCK); if (IS_ERR(vma)) { @@ -3544,11 +3553,10 @@ i915_gem_object_unpin_from_display_plane(struct i915_vma *vma) return; if (--vma->obj->pin_display == 0) - vma->display_alignment = 0; + vma->display_alignment = I915_GTT_MIN_ALIGNMENT; /* Bump the LRU to try and avoid premature eviction whilst flipping */ - if (!i915_vma_is_active(vma)) - list_move_tail(&vma->vm_link, &vma->vm->inactive_list); + i915_gem_object_bump_inactive_ggtt(vma->obj); i915_vma_unpin(vma); } @@ -3679,8 +3687,8 @@ i915_gem_object_ggtt_pin(struct drm_i915_gem_object *obj, lockdep_assert_held(&obj->base.dev->struct_mutex); - vma = i915_gem_obj_lookup_or_create_vma(obj, vm, view); - if (IS_ERR(vma)) + vma = i915_vma_instance(obj, vm, view); + if (unlikely(IS_ERR(vma))) return vma; if (i915_vma_misplaced(vma, size, alignment, flags)) { @@ -3689,10 +3697,6 @@ i915_gem_object_ggtt_pin(struct drm_i915_gem_object *obj, return ERR_PTR(-ENOSPC); if (flags & PIN_MAPPABLE) { - u32 fence_size; - - fence_size = i915_gem_get_ggtt_size(dev_priv, vma->size, - i915_gem_object_get_tiling(obj)); /* If the required space is larger than the available * aperture, we will not able to find a slot for the * object and unbinding the object now will be in @@ -3700,7 +3704,7 @@ i915_gem_object_ggtt_pin(struct drm_i915_gem_object *obj, * the object in and out of the Global GTT and * waste a lot of cycles under the mutex. */ - if (fence_size > dev_priv->ggtt.mappable_end) + if (vma->fence_size > dev_priv->ggtt.mappable_end) return ERR_PTR(-E2BIG); /* If NONBLOCK is set the caller is optimistically @@ -3719,7 +3723,7 @@ i915_gem_object_ggtt_pin(struct drm_i915_gem_object *obj, * we could try to minimise harm to others. */ if (flags & PIN_NONBLOCK && - fence_size > dev_priv->ggtt.mappable_end / 2) + vma->fence_size > dev_priv->ggtt.mappable_end / 2) return ERR_PTR(-ENOSPC); } @@ -4193,7 +4197,8 @@ static void assert_kernel_context_is_current(struct drm_i915_private *dev_priv) enum intel_engine_id id; for_each_engine(engine, dev_priv, id) - GEM_BUG_ON(!i915_gem_context_is_kernel(engine->last_retired_context)); + GEM_BUG_ON(engine->last_retired_context && + !i915_gem_context_is_kernel(engine->last_retired_context)); } int i915_gem_suspend(struct drm_i915_private *dev_priv) diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index 40a6939..17f90c6 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -97,7 +97,7 @@ * part. It should be safe to decrease this, but it's more future proof as is. */ #define GEN6_CONTEXT_ALIGN (64<<10) -#define GEN7_CONTEXT_ALIGN 4096 +#define GEN7_CONTEXT_ALIGN I915_GTT_MIN_ALIGNMENT static size_t get_context_alignment(struct drm_i915_private *dev_priv) { @@ -205,27 +205,6 @@ alloc_context_obj(struct drm_i915_private *dev_priv, u64 size) return obj; } -static void i915_ppgtt_close(struct i915_address_space *vm) -{ - struct list_head *phases[] = { - &vm->active_list, - &vm->inactive_list, - &vm->unbound_list, - NULL, - }, **phase; - - GEM_BUG_ON(vm->closed); - vm->closed = true; - - for (phase = phases; *phase; phase++) { - struct i915_vma *vma, *vn; - - list_for_each_entry_safe(vma, vn, *phase, vm_link) - if (!i915_vma_is_closed(vma)) - i915_vma_close(vma); - } -} - static void context_close(struct i915_gem_context *ctx) { i915_gem_context_set_closed(ctx); @@ -290,7 +269,7 @@ __create_hw_context(struct drm_i915_private *dev_priv, goto err_out; } - vma = i915_vma_create(obj, &dev_priv->ggtt.base, NULL); + vma = i915_vma_instance(obj, &dev_priv->ggtt.base, NULL); if (IS_ERR(vma)) { i915_gem_object_put(obj); ret = PTR_ERR(vma); @@ -341,7 +320,7 @@ __create_hw_context(struct drm_i915_private *dev_priv, if (HAS_GUC(dev_priv) && i915.enable_guc_loading) ctx->ggtt_offset_bias = GUC_WOPCM_TOP; else - ctx->ggtt_offset_bias = 4096; + ctx->ggtt_offset_bias = I915_GTT_PAGE_SIZE; return ctx; @@ -456,7 +435,8 @@ int i915_gem_context_init(struct drm_i915_private *dev_priv) dev_priv->hw_context_size = 0; } else if (HAS_HW_CONTEXTS(dev_priv)) { dev_priv->hw_context_size = - round_up(get_context_size(dev_priv), 4096); + round_up(get_context_size(dev_priv), + I915_GTT_PAGE_SIZE); if (dev_priv->hw_context_size > (1<<20)) { DRM_DEBUG_DRIVER("Disabling HW Contexts; invalid size %d\n", dev_priv->hw_context_size); @@ -897,6 +877,26 @@ int i915_switch_context(struct drm_i915_gem_request *req) return do_rcs_switch(req); } +static bool engine_has_kernel_context(struct intel_engine_cs *engine) +{ + struct i915_gem_timeline *timeline; + + list_for_each_entry(timeline, &engine->i915->gt.timelines, link) { + struct intel_timeline *tl; + + if (timeline == &engine->i915->gt.global_timeline) + continue; + + tl = &timeline->engine[engine->id]; + if (i915_gem_active_peek(&tl->last_request, + &engine->i915->drm.struct_mutex)) + return false; + } + + return (!engine->last_retired_context || + i915_gem_context_is_kernel(engine->last_retired_context)); +} + int i915_gem_switch_to_kernel_context(struct drm_i915_private *dev_priv) { struct intel_engine_cs *engine; @@ -905,10 +905,15 @@ int i915_gem_switch_to_kernel_context(struct drm_i915_private *dev_priv) lockdep_assert_held(&dev_priv->drm.struct_mutex); + i915_gem_retire_requests(dev_priv); + for_each_engine(engine, dev_priv, id) { struct drm_i915_gem_request *req; int ret; + if (engine_has_kernel_context(engine)) + continue; + req = i915_gem_request_alloc(engine, dev_priv->kernel_context); if (IS_ERR(req)) return PTR_ERR(req); diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c index 026ebc5..a43e44e 100644 --- a/drivers/gpu/drm/i915/i915_gem_evict.c +++ b/drivers/gpu/drm/i915/i915_gem_evict.c @@ -231,7 +231,8 @@ found: /** * i915_gem_evict_for_vma - Evict vmas to make room for binding a new one - * @target: address space and range to evict for + * @vm: address space to evict from + * @target: range (and color) to evict for * @flags: additional flags to control the eviction algorithm * * This function will try to evict vmas that overlap the target node. @@ -239,18 +240,20 @@ found: * To clarify: This is for freeing up virtual address space, not for freeing * memory in e.g. the shrinker. */ -int i915_gem_evict_for_vma(struct i915_vma *target, unsigned int flags) +int i915_gem_evict_for_node(struct i915_address_space *vm, + struct drm_mm_node *target, + unsigned int flags) { LIST_HEAD(eviction_list); struct drm_mm_node *node; - u64 start = target->node.start; - u64 end = start + target->node.size; + u64 start = target->start; + u64 end = start + target->size; struct i915_vma *vma, *next; bool check_color; int ret = 0; - lockdep_assert_held(&target->vm->i915->drm.struct_mutex); - trace_i915_gem_evict_vma(target, flags); + lockdep_assert_held(&vm->i915->drm.struct_mutex); + trace_i915_gem_evict_node(vm, target, flags); /* Retire before we search the active list. Although we have * reasonable accuracy in our retirement lists, we may have @@ -258,18 +261,18 @@ int i915_gem_evict_for_vma(struct i915_vma *target, unsigned int flags) * retiring. */ if (!(flags & PIN_NONBLOCK)) - i915_gem_retire_requests(target->vm->i915); + i915_gem_retire_requests(vm->i915); - check_color = target->vm->mm.color_adjust; + check_color = vm->mm.color_adjust; if (check_color) { /* Expand search to cover neighbouring guard pages (or lack!) */ - if (start > target->vm->start) - start -= 4096; - if (end < target->vm->start + target->vm->total) - end += 4096; + if (start > vm->start) + start -= I915_GTT_PAGE_SIZE; + if (end < vm->start + vm->total) + end += I915_GTT_PAGE_SIZE; } - drm_mm_for_each_node_in_range(node, &target->vm->mm, start, end) { + drm_mm_for_each_node_in_range(node, &vm->mm, start, end) { /* If we find any non-objects (!vma), we cannot evict them */ if (node->color == I915_COLOR_UNEVICTABLE) { ret = -ENOSPC; @@ -285,12 +288,12 @@ int i915_gem_evict_for_vma(struct i915_vma *target, unsigned int flags) * those as well to make room for our guard pages. */ if (check_color) { - if (vma->node.start + vma->node.size == target->node.start) { - if (vma->node.color == target->node.color) + if (vma->node.start + vma->node.size == node->start) { + if (vma->node.color == node->color) continue; } - if (vma->node.start == target->node.start + target->node.size) { - if (vma->node.color == target->node.color) + if (vma->node.start == node->start + node->size) { + if (vma->node.color == node->color) continue; } } @@ -302,7 +305,7 @@ int i915_gem_evict_for_vma(struct i915_vma *target, unsigned int flags) } /* Overlap of objects in the same batch? */ - if (i915_vma_is_pinned(vma)) { + if (i915_vma_is_pinned(vma) || !list_empty(&vma->exec_list)) { ret = -ENOSPC; if (vma->exec_entry && vma->exec_entry->flags & EXEC_OBJECT_PINNED) diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index a5fe299..c66e905 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -184,7 +184,7 @@ eb_lookup_vmas(struct eb_vmas *eb, * from the (obj, vm) we don't run the risk of creating * duplicated vmas for the same vm. */ - vma = i915_gem_obj_lookup_or_create_vma(obj, vm, NULL); + vma = i915_vma_instance(obj, vm, NULL); if (unlikely(IS_ERR(vma))) { DRM_DEBUG("Failed to lookup VMA\n"); ret = PTR_ERR(vma); @@ -438,7 +438,7 @@ static void *reloc_iomap(struct drm_i915_gem_object *obj, memset(&cache->node, 0, sizeof(cache->node)); ret = drm_mm_insert_node_in_range_generic (&ggtt->base.mm, &cache->node, - 4096, 0, I915_COLOR_UNEVICTABLE, + PAGE_SIZE, 0, I915_COLOR_UNEVICTABLE, 0, ggtt->mappable_end, DRM_MM_SEARCH_DEFAULT, DRM_MM_CREATE_DEFAULT); @@ -851,8 +851,7 @@ eb_vma_misplaced(struct i915_vma *vma) WARN_ON(entry->flags & __EXEC_OBJECT_NEEDS_MAP && !i915_vma_is_ggtt(vma)); - if (entry->alignment && - vma->node.start & (entry->alignment - 1)) + if (entry->alignment && !IS_ALIGNED(vma->node.start, entry->alignment)) return true; if (vma->node.size < entry->pad_to_size) diff --git a/drivers/gpu/drm/i915/i915_gem_fence_reg.c b/drivers/gpu/drm/i915/i915_gem_fence_reg.c index 775059e..fadbe8f 100644 --- a/drivers/gpu/drm/i915/i915_gem_fence_reg.c +++ b/drivers/gpu/drm/i915/i915_gem_fence_reg.c @@ -77,16 +77,17 @@ static void i965_write_fence_reg(struct drm_i915_fence_reg *fence, val = 0; if (vma) { - unsigned int tiling = i915_gem_object_get_tiling(vma->obj); - bool is_y_tiled = tiling == I915_TILING_Y; unsigned int stride = i915_gem_object_get_stride(vma->obj); - u32 row_size = stride * (is_y_tiled ? 32 : 8); - u32 size = rounddown((u32)vma->node.size, row_size); - val = ((vma->node.start + size - 4096) & 0xfffff000) << 32; - val |= vma->node.start & 0xfffff000; + GEM_BUG_ON(!i915_vma_is_map_and_fenceable(vma)); + GEM_BUG_ON(!IS_ALIGNED(vma->node.start, I965_FENCE_PAGE)); + GEM_BUG_ON(!IS_ALIGNED(vma->fence_size, I965_FENCE_PAGE)); + GEM_BUG_ON(!IS_ALIGNED(stride, 128)); + + val = (vma->node.start + vma->fence_size - I965_FENCE_PAGE) << 32; + val |= vma->node.start; val |= (u64)((stride / 128) - 1) << fence_pitch_shift; - if (is_y_tiled) + if (i915_gem_object_get_tiling(vma->obj) == I915_TILING_Y) val |= BIT(I965_FENCE_TILING_Y_SHIFT); val |= I965_FENCE_REG_VALID; } @@ -122,31 +123,24 @@ static void i915_write_fence_reg(struct drm_i915_fence_reg *fence, unsigned int tiling = i915_gem_object_get_tiling(vma->obj); bool is_y_tiled = tiling == I915_TILING_Y; unsigned int stride = i915_gem_object_get_stride(vma->obj); - int pitch_val; - int tile_width; - WARN((vma->node.start & ~I915_FENCE_START_MASK) || - !is_power_of_2(vma->node.size) || - (vma->node.start & (vma->node.size - 1)), - "object 0x%08llx [fenceable? %d] not 1M or pot-size (0x%08llx) aligned\n", - vma->node.start, - i915_vma_is_map_and_fenceable(vma), - vma->node.size); + GEM_BUG_ON(!i915_vma_is_map_and_fenceable(vma)); + GEM_BUG_ON(vma->node.start & ~I915_FENCE_START_MASK); + GEM_BUG_ON(!is_power_of_2(vma->fence_size)); + GEM_BUG_ON(!IS_ALIGNED(vma->node.start, vma->fence_size)); if (is_y_tiled && HAS_128_BYTE_Y_TILING(fence->i915)) - tile_width = 128; + stride /= 128; else - tile_width = 512; - - /* Note: pitch better be a power of two tile widths */ - pitch_val = stride / tile_width; - pitch_val = ffs(pitch_val) - 1; + stride /= 512; + GEM_BUG_ON(!is_power_of_2(stride)); val = vma->node.start; if (is_y_tiled) val |= BIT(I830_FENCE_TILING_Y_SHIFT); - val |= I915_FENCE_SIZE_BITS(vma->node.size); - val |= pitch_val << I830_FENCE_PITCH_SHIFT; + val |= I915_FENCE_SIZE_BITS(vma->fence_size); + val |= ilog2(stride) << I830_FENCE_PITCH_SHIFT; + val |= I830_FENCE_REG_VALID; } @@ -166,25 +160,19 @@ static void i830_write_fence_reg(struct drm_i915_fence_reg *fence, val = 0; if (vma) { - unsigned int tiling = i915_gem_object_get_tiling(vma->obj); - bool is_y_tiled = tiling == I915_TILING_Y; unsigned int stride = i915_gem_object_get_stride(vma->obj); - u32 pitch_val; - WARN((vma->node.start & ~I830_FENCE_START_MASK) || - !is_power_of_2(vma->node.size) || - (vma->node.start & (vma->node.size - 1)), - "object 0x%08llx not 512K or pot-size 0x%08llx aligned\n", - vma->node.start, vma->node.size); - - pitch_val = stride / 128; - pitch_val = ffs(pitch_val) - 1; + GEM_BUG_ON(!i915_vma_is_map_and_fenceable(vma)); + GEM_BUG_ON(vma->node.start & ~I830_FENCE_START_MASK); + GEM_BUG_ON(!is_power_of_2(vma->fence_size)); + GEM_BUG_ON(!is_power_of_2(stride / 128)); + GEM_BUG_ON(!IS_ALIGNED(vma->node.start, vma->fence_size)); val = vma->node.start; - if (is_y_tiled) + if (i915_gem_object_get_tiling(vma->obj) == I915_TILING_Y) val |= BIT(I830_FENCE_TILING_Y_SHIFT); - val |= I830_FENCE_SIZE_BITS(vma->node.size); - val |= pitch_val << I830_FENCE_PITCH_SHIFT; + val |= I830_FENCE_SIZE_BITS(vma->fence_size); + val |= ilog2(stride / 128) << I830_FENCE_PITCH_SHIFT; val |= I830_FENCE_REG_VALID; } diff --git a/drivers/gpu/drm/i915/i915_gem_fence_reg.h b/drivers/gpu/drm/i915/i915_gem_fence_reg.h index 22c4a2d..99a31de 100644 --- a/drivers/gpu/drm/i915/i915_gem_fence_reg.h +++ b/drivers/gpu/drm/i915/i915_gem_fence_reg.h @@ -30,6 +30,8 @@ struct drm_i915_private; struct i915_vma; +#define I965_FENCE_PAGE 4096UL + struct drm_i915_fence_reg { struct list_head link; struct drm_i915_private *i915; diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index f698006..e808aad 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -23,10 +23,14 @@ * */ +#include <linux/log2.h> +#include <linux/random.h> #include <linux/seq_file.h> #include <linux/stop_machine.h> + #include <drm/drmP.h> #include <drm/i915_drm.h> + #include "i915_drv.h" #include "i915_vgpu.h" #include "i915_trace.h" @@ -99,12 +103,29 @@ static int i915_get_ggtt_vma_pages(struct i915_vma *vma); -const struct i915_ggtt_view i915_ggtt_view_normal = { - .type = I915_GGTT_VIEW_NORMAL, -}; -const struct i915_ggtt_view i915_ggtt_view_rotated = { - .type = I915_GGTT_VIEW_ROTATED, -}; +static void gen6_ggtt_invalidate(struct drm_i915_private *dev_priv) +{ + /* Note that as an uncached mmio write, this should flush the + * WCB of the writes into the GGTT before it triggers the invalidate. + */ + I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); +} + +static void guc_ggtt_invalidate(struct drm_i915_private *dev_priv) +{ + gen6_ggtt_invalidate(dev_priv); + I915_WRITE(GEN8_GTCR, GEN8_GTCR_INVALIDATE); +} + +static void gmch_ggtt_invalidate(struct drm_i915_private *dev_priv) +{ + intel_gtt_chipset_flush(); +} + +static inline void i915_ggtt_invalidate(struct drm_i915_private *i915) +{ + i915->ggtt.invalidate(i915); +} int intel_sanitize_enable_ppgtt(struct drm_i915_private *dev_priv, int enable_ppgtt) @@ -329,7 +350,7 @@ static int __setup_page_dma(struct drm_i915_private *dev_priv, return -ENOMEM; p->daddr = dma_map_page(kdev, - p->page, 0, 4096, PCI_DMA_BIDIRECTIONAL); + p->page, 0, PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); if (dma_mapping_error(kdev, p->daddr)) { __free_page(p->page); @@ -353,7 +374,7 @@ static void cleanup_page_dma(struct drm_i915_private *dev_priv, if (WARN_ON(!p->page)) return; - dma_unmap_page(&pdev->dev, p->daddr, 4096, PCI_DMA_BIDIRECTIONAL); + dma_unmap_page(&pdev->dev, p->daddr, PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); __free_page(p->page); memset(p, 0, sizeof(*p)); } @@ -626,10 +647,10 @@ static void gen8_initialize_pml4(struct i915_address_space *vm, } static void -gen8_setup_page_directory(struct i915_hw_ppgtt *ppgtt, - struct i915_page_directory_pointer *pdp, - struct i915_page_directory *pd, - int index) +gen8_setup_pdpe(struct i915_hw_ppgtt *ppgtt, + struct i915_page_directory_pointer *pdp, + struct i915_page_directory *pd, + int index) { gen8_ppgtt_pdpe_t *page_directorypo; @@ -642,10 +663,10 @@ gen8_setup_page_directory(struct i915_hw_ppgtt *ppgtt, } static void -gen8_setup_page_directory_pointer(struct i915_hw_ppgtt *ppgtt, - struct i915_pml4 *pml4, - struct i915_page_directory_pointer *pdp, - int index) +gen8_setup_pml4e(struct i915_hw_ppgtt *ppgtt, + struct i915_pml4 *pml4, + struct i915_page_directory_pointer *pdp, + int index) { gen8_ppgtt_pml4e_t *pagemap = kmap_px(pml4); @@ -793,9 +814,6 @@ static bool gen8_ppgtt_clear_pdp(struct i915_address_space *vm, struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm); struct i915_page_directory *pd; uint64_t pdpe; - gen8_ppgtt_pdpe_t *pdpe_vaddr; - gen8_ppgtt_pdpe_t scratch_pdpe = - gen8_pdpe_encode(px_dma(vm->scratch_pd), I915_CACHE_LLC); gen8_for_each_pdpe(pd, pdp, start, length, pdpe) { if (WARN_ON(!pdp->page_directory[pdpe])) @@ -803,11 +821,7 @@ static bool gen8_ppgtt_clear_pdp(struct i915_address_space *vm, if (gen8_ppgtt_clear_pd(vm, pd, start, length)) { __clear_bit(pdpe, pdp->used_pdpes); - if (USES_FULL_48BIT_PPGTT(dev_priv)) { - pdpe_vaddr = kmap_px(pdp); - pdpe_vaddr[pdpe] = scratch_pdpe; - kunmap_px(ppgtt, pdpe_vaddr); - } + gen8_setup_pdpe(ppgtt, pdp, vm->scratch_pd, pdpe); free_pd(vm->i915, pd); } } @@ -832,9 +846,6 @@ static void gen8_ppgtt_clear_pml4(struct i915_address_space *vm, struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm); struct i915_page_directory_pointer *pdp; uint64_t pml4e; - gen8_ppgtt_pml4e_t *pml4e_vaddr; - gen8_ppgtt_pml4e_t scratch_pml4e = - gen8_pml4e_encode(px_dma(vm->scratch_pdp), I915_CACHE_LLC); GEM_BUG_ON(!USES_FULL_48BIT_PPGTT(vm->i915)); @@ -844,9 +855,7 @@ static void gen8_ppgtt_clear_pml4(struct i915_address_space *vm, if (gen8_ppgtt_clear_pdp(vm, pdp, start, length)) { __clear_bit(pml4e, pml4->used_pml4es); - pml4e_vaddr = kmap_px(pml4); - pml4e_vaddr[pml4e] = scratch_pml4e; - kunmap_px(ppgtt, pml4e_vaddr); + gen8_setup_pml4e(ppgtt, pml4, vm->scratch_pdp, pml4e); free_pdp(vm->i915, pdp); } } @@ -1366,7 +1375,7 @@ static int gen8_alloc_va_range_3lvl(struct i915_address_space *vm, kunmap_px(ppgtt, page_directory); __set_bit(pdpe, pdp->used_pdpes); - gen8_setup_page_directory(ppgtt, pdp, pd, pdpe); + gen8_setup_pdpe(ppgtt, pdp, pd, pdpe); } free_gen8_temp_bitmaps(new_page_dirs, new_page_tables); @@ -1425,7 +1434,7 @@ static int gen8_alloc_va_range_4lvl(struct i915_address_space *vm, if (ret) goto err_out; - gen8_setup_page_directory_pointer(ppgtt, pml4, pdp, pml4e); + gen8_setup_pml4e(ppgtt, pml4, pdp, pml4e); } bitmap_or(pml4->used_pml4es, new_pdps, pml4->used_pml4es, @@ -2044,7 +2053,6 @@ static int gen6_ppgtt_allocate_page_directories(struct i915_hw_ppgtt *ppgtt) struct i915_address_space *vm = &ppgtt->base; struct drm_i915_private *dev_priv = ppgtt->base.i915; struct i915_ggtt *ggtt = &dev_priv->ggtt; - bool retried = false; int ret; /* PPGTT PDEs reside in the GGTT and consists of 512 entries. The @@ -2057,29 +2065,14 @@ static int gen6_ppgtt_allocate_page_directories(struct i915_hw_ppgtt *ppgtt) if (ret) return ret; -alloc: - ret = drm_mm_insert_node_in_range_generic(&ggtt->base.mm, &ppgtt->node, - GEN6_PD_SIZE, GEN6_PD_ALIGN, - I915_COLOR_UNEVICTABLE, - 0, ggtt->base.total, - DRM_MM_TOPDOWN); - if (ret == -ENOSPC && !retried) { - ret = i915_gem_evict_something(&ggtt->base, - GEN6_PD_SIZE, GEN6_PD_ALIGN, - I915_COLOR_UNEVICTABLE, - 0, ggtt->base.total, - 0); - if (ret) - goto err_out; - - retried = true; - goto alloc; - } - + ret = i915_gem_gtt_insert(&ggtt->base, &ppgtt->node, + GEN6_PD_SIZE, GEN6_PD_ALIGN, + I915_COLOR_UNEVICTABLE, + 0, ggtt->base.total, + PIN_HIGH); if (ret) goto err_out; - if (ppgtt->node.start < ggtt->mappable_end) DRM_DEBUG("Forced to use aperture for PDEs\n"); @@ -2267,6 +2260,27 @@ i915_ppgtt_create(struct drm_i915_private *dev_priv, return ppgtt; } +void i915_ppgtt_close(struct i915_address_space *vm) +{ + struct list_head *phases[] = { + &vm->active_list, + &vm->inactive_list, + &vm->unbound_list, + NULL, + }, **phase; + + GEM_BUG_ON(vm->closed); + vm->closed = true; + + for (phase = phases; *phase; phase++) { + struct i915_vma *vma, *vn; + + list_for_each_entry_safe(vma, vn, *phase, vm_link) + if (!i915_vma_is_closed(vma)) + i915_vma_close(vma); + } +} + void i915_ppgtt_release(struct kref *kref) { struct i915_hw_ppgtt *ppgtt = @@ -2331,16 +2345,6 @@ void i915_check_and_clear_faults(struct drm_i915_private *dev_priv) POSTING_READ(RING_FAULT_REG(dev_priv->engine[RCS])); } -static void i915_ggtt_flush(struct drm_i915_private *dev_priv) -{ - if (INTEL_INFO(dev_priv)->gen < 6) { - intel_gtt_chipset_flush(); - } else { - I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); - POSTING_READ(GFX_FLSH_CNTL_GEN6); - } -} - void i915_gem_suspend_gtt_mappings(struct drm_i915_private *dev_priv) { struct i915_ggtt *ggtt = &dev_priv->ggtt; @@ -2355,7 +2359,7 @@ void i915_gem_suspend_gtt_mappings(struct drm_i915_private *dev_priv) ggtt->base.clear_range(&ggtt->base, ggtt->base.start, ggtt->base.total); - i915_ggtt_flush(dev_priv); + i915_ggtt_invalidate(dev_priv); } int i915_gem_gtt_prepare_pages(struct drm_i915_gem_object *obj, @@ -2394,15 +2398,13 @@ static void gen8_ggtt_insert_page(struct i915_address_space *vm, enum i915_cache_level level, u32 unused) { - struct drm_i915_private *dev_priv = vm->i915; + struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); gen8_pte_t __iomem *pte = - (gen8_pte_t __iomem *)dev_priv->ggtt.gsm + - (offset >> PAGE_SHIFT); + (gen8_pte_t __iomem *)ggtt->gsm + (offset >> PAGE_SHIFT); gen8_set_pte(pte, gen8_pte_encode(addr, level)); - I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); - POSTING_READ(GFX_FLSH_CNTL_GEN6); + ggtt->invalidate(vm->i915); } static void gen8_ggtt_insert_entries(struct i915_address_space *vm, @@ -2410,7 +2412,6 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm, uint64_t start, enum i915_cache_level level, u32 unused) { - struct drm_i915_private *dev_priv = vm->i915; struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); struct sgt_iter sgt_iter; gen8_pte_t __iomem *gtt_entries; @@ -2439,8 +2440,7 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm, * want to flush the TLBs only after we're certain all the PTE updates * have finished. */ - I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); - POSTING_READ(GFX_FLSH_CNTL_GEN6); + ggtt->invalidate(vm->i915); } struct insert_entries { @@ -2475,15 +2475,13 @@ static void gen6_ggtt_insert_page(struct i915_address_space *vm, enum i915_cache_level level, u32 flags) { - struct drm_i915_private *dev_priv = vm->i915; + struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); gen6_pte_t __iomem *pte = - (gen6_pte_t __iomem *)dev_priv->ggtt.gsm + - (offset >> PAGE_SHIFT); + (gen6_pte_t __iomem *)ggtt->gsm + (offset >> PAGE_SHIFT); iowrite32(vm->pte_encode(addr, level, flags), pte); - I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); - POSTING_READ(GFX_FLSH_CNTL_GEN6); + ggtt->invalidate(vm->i915); } /* @@ -2497,7 +2495,6 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm, uint64_t start, enum i915_cache_level level, u32 flags) { - struct drm_i915_private *dev_priv = vm->i915; struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm); struct sgt_iter sgt_iter; gen6_pte_t __iomem *gtt_entries; @@ -2525,8 +2522,7 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm, * want to flush the TLBs only after we're certain all the PTE updates * have finished. */ - I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); - POSTING_READ(GFX_FLSH_CNTL_GEN6); + ggtt->invalidate(vm->i915); } static void nop_clear_range(struct i915_address_space *vm, @@ -2723,11 +2719,11 @@ static void i915_gtt_color_adjust(const struct drm_mm_node *node, u64 *end) { if (node->color != color) - *start += 4096; + *start += I915_GTT_PAGE_SIZE; node = list_next_entry(node, node_list); if (node->allocated && node->color != color) - *end -= 4096; + *end -= I915_GTT_PAGE_SIZE; } int i915_gem_init_ggtt(struct drm_i915_private *dev_priv) @@ -2754,7 +2750,7 @@ int i915_gem_init_ggtt(struct drm_i915_private *dev_priv) /* Reserve a mappable slot for our lockless error capture */ ret = drm_mm_insert_node_in_range_generic(&ggtt->base.mm, &ggtt->error_capture, - 4096, 0, + PAGE_SIZE, 0, I915_COLOR_UNEVICTABLE, 0, ggtt->mappable_end, 0, 0); @@ -3086,6 +3082,8 @@ static int gen8_gmch_probe(struct i915_ggtt *ggtt) if (IS_CHERRYVIEW(dev_priv)) ggtt->base.insert_entries = gen8_ggtt_insert_entries__BKL; + ggtt->invalidate = gen6_ggtt_invalidate; + return ggtt_probe_common(ggtt, size); } @@ -3123,6 +3121,8 @@ static int gen6_gmch_probe(struct i915_ggtt *ggtt) ggtt->base.unbind_vma = ggtt_unbind_vma; ggtt->base.cleanup = gen6_gmch_remove; + ggtt->invalidate = gen6_ggtt_invalidate; + if (HAS_EDRAM(dev_priv)) ggtt->base.pte_encode = iris_pte_encode; else if (IS_HASWELL(dev_priv)) @@ -3166,6 +3166,8 @@ static int i915_gmch_probe(struct i915_ggtt *ggtt) ggtt->base.unbind_vma = ggtt_unbind_vma; ggtt->base.cleanup = i915_gmch_remove; + ggtt->invalidate = gmch_ggtt_invalidate; + if (unlikely(ggtt->do_idle_maps)) DRM_INFO("applying Ironlake quirks for intel_iommu\n"); @@ -3284,6 +3286,16 @@ int i915_ggtt_enable_hw(struct drm_i915_private *dev_priv) return 0; } +void i915_ggtt_enable_guc(struct drm_i915_private *i915) +{ + i915->ggtt.invalidate = guc_ggtt_invalidate; +} + +void i915_ggtt_disable_guc(struct drm_i915_private *i915) +{ + i915->ggtt.invalidate = gen6_ggtt_invalidate; +} + void i915_gem_restore_gtt_mappings(struct drm_i915_private *dev_priv) { struct i915_ggtt *ggtt = &dev_priv->ggtt; @@ -3347,52 +3359,7 @@ void i915_gem_restore_gtt_mappings(struct drm_i915_private *dev_priv) } } - i915_ggtt_flush(dev_priv); -} - -struct i915_vma * -i915_gem_obj_to_vma(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view) -{ - struct rb_node *rb; - - rb = obj->vma_tree.rb_node; - while (rb) { - struct i915_vma *vma = rb_entry(rb, struct i915_vma, obj_node); - long cmp; - - cmp = i915_vma_compare(vma, vm, view); - if (cmp == 0) - return vma; - - if (cmp < 0) - rb = rb->rb_right; - else - rb = rb->rb_left; - } - - return NULL; -} - -struct i915_vma * -i915_gem_obj_lookup_or_create_vma(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view) -{ - struct i915_vma *vma; - - lockdep_assert_held(&obj->base.dev->struct_mutex); - GEM_BUG_ON(view && !i915_is_ggtt(vm)); - - vma = i915_gem_obj_to_vma(obj, vm, view); - if (!vma) { - vma = i915_vma_create(obj, vm, view); - GEM_BUG_ON(vma != i915_gem_obj_to_vma(obj, vm, view)); - } - - GEM_BUG_ON(i915_vma_is_closed(vma)); - return vma; + i915_ggtt_invalidate(dev_priv); } static struct scatterlist * @@ -3492,7 +3459,7 @@ intel_partial_pages(const struct i915_ggtt_view *view, { struct sg_table *st; struct scatterlist *sg, *iter; - unsigned int count = view->params.partial.size; + unsigned int count = view->partial.size; unsigned int offset; int ret = -ENOMEM; @@ -3504,9 +3471,7 @@ intel_partial_pages(const struct i915_ggtt_view *view, if (ret) goto err_sg_alloc; - iter = i915_gem_object_get_sg(obj, - view->params.partial.offset, - &offset); + iter = i915_gem_object_get_sg(obj, view->partial.offset, &offset); GEM_BUG_ON(!iter); sg = st->sgl; @@ -3558,7 +3523,8 @@ i915_get_ggtt_vma_pages(struct i915_vma *vma) vma->pages = vma->obj->mm.pages; else if (vma->ggtt_view.type == I915_GGTT_VIEW_ROTATED) vma->pages = - intel_rotate_fb_obj_pages(&vma->ggtt_view.params.rotated, vma->obj); + intel_rotate_fb_obj_pages(&vma->ggtt_view.rotated, + vma->obj); else if (vma->ggtt_view.type == I915_GGTT_VIEW_PARTIAL) vma->pages = intel_partial_pages(&vma->ggtt_view, vma->obj); else @@ -3579,3 +3545,207 @@ i915_get_ggtt_vma_pages(struct i915_vma *vma) return ret; } +/** + * i915_gem_gtt_reserve - reserve a node in an address_space (GTT) + * @vm: the &struct i915_address_space + * @node: the &struct drm_mm_node (typically i915_vma.mode) + * @size: how much space to allocate inside the GTT, + * must be #I915_GTT_PAGE_SIZE aligned + * @offset: where to insert inside the GTT, + * must be #I915_GTT_MIN_ALIGNMENT aligned, and the node + * (@offset + @size) must fit within the address space + * @color: color to apply to node, if this node is not from a VMA, + * color must be #I915_COLOR_UNEVICTABLE + * @flags: control search and eviction behaviour + * + * i915_gem_gtt_reserve() tries to insert the @node at the exact @offset inside + * the address space (using @size and @color). If the @node does not fit, it + * tries to evict any overlapping nodes from the GTT, including any + * neighbouring nodes if the colors do not match (to ensure guard pages between + * differing domains). See i915_gem_evict_for_node() for the gory details + * on the eviction algorithm. #PIN_NONBLOCK may used to prevent waiting on + * evicting active overlapping objects, and any overlapping node that is pinned + * or marked as unevictable will also result in failure. + * + * Returns: 0 on success, -ENOSPC if no suitable hole is found, -EINTR if + * asked to wait for eviction and interrupted. + */ +int i915_gem_gtt_reserve(struct i915_address_space *vm, + struct drm_mm_node *node, + u64 size, u64 offset, unsigned long color, + unsigned int flags) +{ + int err; + + GEM_BUG_ON(!size); + GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE)); + GEM_BUG_ON(!IS_ALIGNED(offset, I915_GTT_MIN_ALIGNMENT)); + GEM_BUG_ON(range_overflows(offset, size, vm->total)); + GEM_BUG_ON(vm == &vm->i915->mm.aliasing_ppgtt->base); + GEM_BUG_ON(drm_mm_node_allocated(node)); + + node->size = size; + node->start = offset; + node->color = color; + + err = drm_mm_reserve_node(&vm->mm, node); + if (err != -ENOSPC) + return err; + + err = i915_gem_evict_for_node(vm, node, flags); + if (err == 0) + err = drm_mm_reserve_node(&vm->mm, node); + + return err; +} + +static u64 random_offset(u64 start, u64 end, u64 len, u64 align) +{ + u64 range, addr; + + GEM_BUG_ON(range_overflows(start, len, end)); + GEM_BUG_ON(round_up(start, align) > round_down(end - len, align)); + + range = round_down(end - len, align) - round_up(start, align); + if (range) { + if (sizeof(unsigned long) == sizeof(u64)) { + addr = get_random_long(); + } else { + addr = get_random_int(); + if (range > U32_MAX) { + addr <<= 32; + addr |= get_random_int(); + } + } + div64_u64_rem(addr, range, &addr); + start += addr; + } + + return round_up(start, align); +} + +/** + * i915_gem_gtt_insert - insert a node into an address_space (GTT) + * @vm: the &struct i915_address_space + * @node: the &struct drm_mm_node (typically i915_vma.node) + * @size: how much space to allocate inside the GTT, + * must be #I915_GTT_PAGE_SIZE aligned + * @alignment: required alignment of starting offset, may be 0 but + * if specified, this must be a power-of-two and at least + * #I915_GTT_MIN_ALIGNMENT + * @color: color to apply to node + * @start: start of any range restriction inside GTT (0 for all), + * must be #I915_GTT_PAGE_SIZE aligned + * @end: end of any range restriction inside GTT (U64_MAX for all), + * must be #I915_GTT_PAGE_SIZE aligned if not U64_MAX + * @flags: control search and eviction behaviour + * + * i915_gem_gtt_insert() first searches for an available hole into which + * is can insert the node. The hole address is aligned to @alignment and + * its @size must then fit entirely within the [@start, @end] bounds. The + * nodes on either side of the hole must match @color, or else a guard page + * will be inserted between the two nodes (or the node evicted). If no + * suitable hole is found, first a victim is randomly selected and tested + * for eviction, otherwise then the LRU list of objects within the GTT + * is scanned to find the first set of replacement nodes to create the hole. + * Those old overlapping nodes are evicted from the GTT (and so must be + * rebound before any future use). Any node that is currently pinned cannot + * be evicted (see i915_vma_pin()). Similar if the node's VMA is currently + * active and #PIN_NONBLOCK is specified, that node is also skipped when + * searching for an eviction candidate. See i915_gem_evict_something() for + * the gory details on the eviction algorithm. + * + * Returns: 0 on success, -ENOSPC if no suitable hole is found, -EINTR if + * asked to wait for eviction and interrupted. + */ +int i915_gem_gtt_insert(struct i915_address_space *vm, + struct drm_mm_node *node, + u64 size, u64 alignment, unsigned long color, + u64 start, u64 end, unsigned int flags) +{ + u32 search_flag, alloc_flag; + u64 offset; + int err; + + lockdep_assert_held(&vm->i915->drm.struct_mutex); + GEM_BUG_ON(!size); + GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE)); + GEM_BUG_ON(alignment && !is_power_of_2(alignment)); + GEM_BUG_ON(alignment && !IS_ALIGNED(alignment, I915_GTT_MIN_ALIGNMENT)); + GEM_BUG_ON(start >= end); + GEM_BUG_ON(start > 0 && !IS_ALIGNED(start, I915_GTT_PAGE_SIZE)); + GEM_BUG_ON(end < U64_MAX && !IS_ALIGNED(end, I915_GTT_PAGE_SIZE)); + GEM_BUG_ON(vm == &vm->i915->mm.aliasing_ppgtt->base); + GEM_BUG_ON(drm_mm_node_allocated(node)); + + if (unlikely(range_overflows(start, size, end))) + return -ENOSPC; + + if (unlikely(round_up(start, alignment) > round_down(end - size, alignment))) + return -ENOSPC; + + if (flags & PIN_HIGH) { + search_flag = DRM_MM_SEARCH_BELOW; + alloc_flag = DRM_MM_CREATE_TOP; + } else { + search_flag = DRM_MM_SEARCH_DEFAULT; + alloc_flag = DRM_MM_CREATE_DEFAULT; + } + + /* We only allocate in PAGE_SIZE/GTT_PAGE_SIZE (4096) chunks, + * so we know that we always have a minimum alignment of 4096. + * The drm_mm range manager is optimised to return results + * with zero alignment, so where possible use the optimal + * path. + */ + BUILD_BUG_ON(I915_GTT_MIN_ALIGNMENT > I915_GTT_PAGE_SIZE); + if (alignment <= I915_GTT_MIN_ALIGNMENT) + alignment = 0; + + err = drm_mm_insert_node_in_range_generic(&vm->mm, node, + size, alignment, color, + start, end, + search_flag, alloc_flag); + if (err != -ENOSPC) + return err; + + /* No free space, pick a slot at random. + * + * There is a pathological case here using a GTT shared between + * mmap and GPU (i.e. ggtt/aliasing_ppgtt but not full-ppgtt): + * + * |<-- 256 MiB aperture -->||<-- 1792 MiB unmappable -->| + * (64k objects) (448k objects) + * + * Now imagine that the eviction LRU is ordered top-down (just because + * pathology meets real life), and that we need to evict an object to + * make room inside the aperture. The eviction scan then has to walk + * the 448k list before it finds one within range. And now imagine that + * it has to search for a new hole between every byte inside the memcpy, + * for several simultaneous clients. + * + * On a full-ppgtt system, if we have run out of available space, there + * will be lots and lots of objects in the eviction list! Again, + * searching that LRU list may be slow if we are also applying any + * range restrictions (e.g. restriction to low 4GiB) and so, for + * simplicity and similarilty between different GTT, try the single + * random replacement first. + */ + offset = random_offset(start, end, + size, alignment ?: I915_GTT_MIN_ALIGNMENT); + err = i915_gem_gtt_reserve(vm, node, size, offset, color, flags); + if (err != -ENOSPC) + return err; + + /* Randomly selected placement is pinned, do a search */ + err = i915_gem_evict_something(vm, size, alignment, color, + start, end, flags); + if (err) + return err; + + search_flag = DRM_MM_SEARCH_DEFAULT; + return drm_mm_insert_node_in_range_generic(&vm->mm, node, + size, alignment, color, + start, end, + search_flag, alloc_flag); +} diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h index 9e91d7e..3c5ef53 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.h +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h @@ -40,6 +40,9 @@ #include "i915_gem_timeline.h" #include "i915_gem_request.h" +#define I915_GTT_PAGE_SIZE 4096UL +#define I915_GTT_MIN_ALIGNMENT I915_GTT_PAGE_SIZE + #define I915_FENCE_REG_NONE -1 #define I915_MAX_NUM_FENCES 32 /* 32 fences + sign bit for FENCE_REG_NONE */ @@ -142,34 +145,57 @@ typedef uint64_t gen8_ppgtt_pml4e_t; struct sg_table; -enum i915_ggtt_view_type { - I915_GGTT_VIEW_NORMAL = 0, - I915_GGTT_VIEW_ROTATED, - I915_GGTT_VIEW_PARTIAL, -}; - struct intel_rotation_info { - struct { + struct intel_rotation_plane_info { /* tiles */ unsigned int width, height, stride, offset; } plane[2]; +} __packed; + +static inline void assert_intel_rotation_info_is_packed(void) +{ + BUILD_BUG_ON(sizeof(struct intel_rotation_info) != 8*sizeof(unsigned int)); +} + +struct intel_partial_info { + u64 offset; + unsigned int size; +} __packed; + +static inline void assert_intel_partial_info_is_packed(void) +{ + BUILD_BUG_ON(sizeof(struct intel_partial_info) != sizeof(u64) + sizeof(unsigned int)); +} + +enum i915_ggtt_view_type { + I915_GGTT_VIEW_NORMAL = 0, + I915_GGTT_VIEW_ROTATED = sizeof(struct intel_rotation_info), + I915_GGTT_VIEW_PARTIAL = sizeof(struct intel_partial_info), }; +static inline void assert_i915_ggtt_view_type_is_unique(void) +{ + /* As we encode the size of each branch inside the union into its type, + * we have to be careful that each branch has a unique size. + */ + switch ((enum i915_ggtt_view_type)0) { + case I915_GGTT_VIEW_NORMAL: + case I915_GGTT_VIEW_PARTIAL: + case I915_GGTT_VIEW_ROTATED: + /* gcc complains if these are identical cases */ + break; + } +} + struct i915_ggtt_view { enum i915_ggtt_view_type type; - union { - struct { - u64 offset; - unsigned int size; - } partial; + /* Members need to contain no holes/padding */ + struct intel_partial_info partial; struct intel_rotation_info rotated; - } params; + }; }; -extern const struct i915_ggtt_view i915_ggtt_view_normal; -extern const struct i915_ggtt_view i915_ggtt_view_rotated; - enum i915_cache_level; struct i915_vma; @@ -333,6 +359,7 @@ struct i915_ggtt { /** "Graphics Stolen Memory" holds the global PTEs */ void __iomem *gsm; + void (*invalidate)(struct drm_i915_private *dev_priv); bool do_idle_maps; @@ -501,6 +528,8 @@ i915_vm_to_ggtt(struct i915_address_space *vm) int i915_ggtt_probe_hw(struct drm_i915_private *dev_priv); int i915_ggtt_init_hw(struct drm_i915_private *dev_priv); int i915_ggtt_enable_hw(struct drm_i915_private *dev_priv); +void i915_ggtt_enable_guc(struct drm_i915_private *i915); +void i915_ggtt_disable_guc(struct drm_i915_private *i915); int i915_gem_init_ggtt(struct drm_i915_private *dev_priv); void i915_ggtt_cleanup_hw(struct drm_i915_private *dev_priv); @@ -509,6 +538,7 @@ void i915_ppgtt_release(struct kref *kref); struct i915_hw_ppgtt *i915_ppgtt_create(struct drm_i915_private *dev_priv, struct drm_i915_file_private *fpriv, const char *name); +void i915_ppgtt_close(struct i915_address_space *vm); static inline void i915_ppgtt_get(struct i915_hw_ppgtt *ppgtt) { if (ppgtt) @@ -529,6 +559,16 @@ int __must_check i915_gem_gtt_prepare_pages(struct drm_i915_gem_object *obj, void i915_gem_gtt_finish_pages(struct drm_i915_gem_object *obj, struct sg_table *pages); +int i915_gem_gtt_reserve(struct i915_address_space *vm, + struct drm_mm_node *node, + u64 size, u64 offset, unsigned long color, + unsigned int flags); + +int i915_gem_gtt_insert(struct i915_address_space *vm, + struct drm_mm_node *node, + u64 size, u64 alignment, unsigned long color, + u64 start, u64 end, unsigned int flags); + /* Flags used by pin/bind&friends. */ #define PIN_NONBLOCK BIT(0) #define PIN_MAPPABLE BIT(1) @@ -543,6 +583,6 @@ void i915_gem_gtt_finish_pages(struct drm_i915_gem_object *obj, #define PIN_HIGH BIT(9) #define PIN_OFFSET_BIAS BIT(10) #define PIN_OFFSET_FIXED BIT(11) -#define PIN_OFFSET_MASK (~4095) +#define PIN_OFFSET_MASK (-I915_GTT_PAGE_SIZE) #endif diff --git a/drivers/gpu/drm/i915/i915_gem_internal.c b/drivers/gpu/drm/i915/i915_gem_internal.c index 2222863..17ce53d 100644 --- a/drivers/gpu/drm/i915/i915_gem_internal.c +++ b/drivers/gpu/drm/i915/i915_gem_internal.c @@ -151,10 +151,16 @@ static const struct drm_i915_gem_object_ops i915_gem_object_internal_ops = { */ struct drm_i915_gem_object * i915_gem_object_create_internal(struct drm_i915_private *i915, - unsigned int size) + phys_addr_t size) { struct drm_i915_gem_object *obj; + GEM_BUG_ON(!size); + GEM_BUG_ON(!IS_ALIGNED(size, PAGE_SIZE)); + + if (overflows_type(size, obj->base.size)) + return ERR_PTR(-E2BIG); + obj = i915_gem_object_alloc(i915); if (!obj) return ERR_PTR(-ENOMEM); diff --git a/drivers/gpu/drm/i915/i915_gem_object.h b/drivers/gpu/drm/i915/i915_gem_object.h index 6a368de..290eaa7 100644 --- a/drivers/gpu/drm/i915/i915_gem_object.h +++ b/drivers/gpu/drm/i915/i915_gem_object.h @@ -317,6 +317,29 @@ i915_gem_object_get_stride(struct drm_i915_gem_object *obj) return obj->tiling_and_stride & STRIDE_MASK; } +static inline unsigned int +i915_gem_tile_height(unsigned int tiling) +{ + GEM_BUG_ON(!tiling); + return tiling == I915_TILING_Y ? 32 : 8; +} + +static inline unsigned int +i915_gem_object_get_tile_height(struct drm_i915_gem_object *obj) +{ + return i915_gem_tile_height(i915_gem_object_get_tiling(obj)); +} + +static inline unsigned int +i915_gem_object_get_tile_row_size(struct drm_i915_gem_object *obj) +{ + return (i915_gem_object_get_stride(obj) * + i915_gem_object_get_tile_height(obj)); +} + +int i915_gem_object_set_tiling(struct drm_i915_gem_object *obj, + unsigned int tiling, unsigned int stride); + static inline struct intel_engine_cs * i915_gem_object_last_write_engine(struct drm_i915_gem_object *obj) { diff --git a/drivers/gpu/drm/i915/i915_gem_render_state.c b/drivers/gpu/drm/i915/i915_gem_render_state.c index 5af19b0..b42c81b 100644 --- a/drivers/gpu/drm/i915/i915_gem_render_state.c +++ b/drivers/gpu/drm/i915/i915_gem_render_state.c @@ -187,20 +187,20 @@ int i915_gem_render_state_init(struct intel_engine_cs *engine) if (!rodata) return 0; - if (rodata->batch_items * 4 > 4096) + if (rodata->batch_items * 4 > PAGE_SIZE) return -EINVAL; so = kmalloc(sizeof(*so), GFP_KERNEL); if (!so) return -ENOMEM; - obj = i915_gem_object_create_internal(engine->i915, 4096); + obj = i915_gem_object_create_internal(engine->i915, PAGE_SIZE); if (IS_ERR(obj)) { ret = PTR_ERR(obj); goto err_free; } - so->vma = i915_vma_create(obj, &engine->i915->ggtt.base, NULL); + so->vma = i915_vma_instance(obj, &engine->i915->ggtt.base, NULL); if (IS_ERR(so->vma)) { ret = PTR_ERR(so->vma); goto err_obj; diff --git a/drivers/gpu/drm/i915/i915_gem_request.c b/drivers/gpu/drm/i915/i915_gem_request.c index 99056b9..72b7f7d 100644 --- a/drivers/gpu/drm/i915/i915_gem_request.c +++ b/drivers/gpu/drm/i915/i915_gem_request.c @@ -307,26 +307,6 @@ void i915_gem_request_retire_upto(struct drm_i915_gem_request *req) } while (tmp != req); } -static int i915_gem_check_wedge(struct drm_i915_private *dev_priv) -{ - struct i915_gpu_error *error = &dev_priv->gpu_error; - - if (i915_terminally_wedged(error)) - return -EIO; - - if (i915_reset_in_progress(error)) { - /* Non-interruptible callers can't handle -EAGAIN, hence return - * -EIO unconditionally for these. - */ - if (!dev_priv->mm.interruptible) - return -EIO; - - return -EAGAIN; - } - - return 0; -} - static int i915_gem_init_global_seqno(struct drm_i915_private *i915, u32 seqno) { struct i915_gem_timeline *timeline = &i915->gt.global_timeline; @@ -521,12 +501,10 @@ i915_gem_request_alloc(struct intel_engine_cs *engine, lockdep_assert_held(&dev_priv->drm.struct_mutex); /* ABI: Before userspace accesses the GPU (e.g. execbuffer), report - * EIO if the GPU is already wedged, or EAGAIN to drop the struct_mutex - * and restart. + * EIO if the GPU is already wedged. */ - ret = i915_gem_check_wedge(dev_priv); - if (ret) - return ERR_PTR(ret); + if (i915_terminally_wedged(&dev_priv->gpu_error)) + return ERR_PTR(-EIO); /* Pinning the contexts may generate requests in order to acquire * GGTT space, so do this first before we reserve a seqno for @@ -851,6 +829,13 @@ void __i915_add_request(struct drm_i915_gem_request *request, bool flush_caches) lockdep_assert_held(&request->i915->drm.struct_mutex); trace_i915_gem_request_add(request); + /* Make sure that no request gazumped us - if it was allocated after + * our i915_gem_request_alloc() and called __i915_add_request() before + * us, the timeline will hold its seqno which is later than ours. + */ + GEM_BUG_ON(i915_seqno_passed(timeline->last_submitted_seqno, + request->fence.seqno)); + /* * To ensure that this call will not fail, space for its emissions * should already have been reserved in the ring buffer. Let the ring diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c index f1a1d33..127d698 100644 --- a/drivers/gpu/drm/i915/i915_gem_stolen.c +++ b/drivers/gpu/drm/i915/i915_gem_stolen.c @@ -647,8 +647,9 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv stolen_offset, gtt_offset, size); /* KISS and expect everything to be page-aligned */ - if (WARN_ON(size == 0) || WARN_ON(size & 4095) || - WARN_ON(stolen_offset & 4095)) + if (WARN_ON(size == 0) || + WARN_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE)) || + WARN_ON(!IS_ALIGNED(stolen_offset, I915_GTT_MIN_ALIGNMENT))) return NULL; stolen = kzalloc(sizeof(*stolen), GFP_KERNEL); @@ -682,7 +683,7 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv if (ret) goto err; - vma = i915_gem_obj_lookup_or_create_vma(obj, &ggtt->base, NULL); + vma = i915_vma_instance(obj, &ggtt->base, NULL); if (IS_ERR(vma)) { ret = PTR_ERR(vma); goto err_pages; @@ -693,15 +694,16 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv * setting up the GTT space. The actual reservation will occur * later. */ - vma->node.start = gtt_offset; - vma->node.size = size; - - ret = drm_mm_reserve_node(&ggtt->base.mm, &vma->node); + ret = i915_gem_gtt_reserve(&ggtt->base, &vma->node, + size, gtt_offset, obj->cache_level, + 0); if (ret) { DRM_DEBUG_KMS("failed to allocate stolen GTT space\n"); goto err_pages; } + GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); + vma->pages = obj->mm.pages; vma->flags |= I915_VMA_GLOBAL_BIND; __i915_vma_set_map_and_fenceable(vma); diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c index 62ad375..b1361cf 100644 --- a/drivers/gpu/drm/i915/i915_gem_tiling.c +++ b/drivers/gpu/drm/i915/i915_gem_tiling.c @@ -34,8 +34,8 @@ /** * DOC: buffer object tiling * - * i915_gem_set_tiling() and i915_gem_get_tiling() is the userspace interface to - * declare fence register requirements. + * i915_gem_set_tiling_ioctl() and i915_gem_get_tiling_ioctl() is the userspace + * interface to declare fence register requirements. * * In principle GEM doesn't care at all about the internal data layout of an * object, and hence it also doesn't care about tiling or swizzling. There's two @@ -58,86 +58,147 @@ * invovlement. */ +/** + * i915_gem_fence_size - required global GTT size for a fence + * @i915: i915 device + * @size: object size + * @tiling: tiling mode + * @stride: tiling stride + * + * Return the required global GTT size for a fence (view of a tiled object), + * taking into account potential fence register mapping. + */ +u32 i915_gem_fence_size(struct drm_i915_private *i915, + u32 size, unsigned int tiling, unsigned int stride) +{ + u32 ggtt_size; + + GEM_BUG_ON(!size); + + if (tiling == I915_TILING_NONE) + return size; + + GEM_BUG_ON(!stride); + + if (INTEL_GEN(i915) >= 4) { + stride *= i915_gem_tile_height(tiling); + GEM_BUG_ON(!IS_ALIGNED(stride, I965_FENCE_PAGE)); + return roundup(size, stride); + } + + /* Previous chips need a power-of-two fence region when tiling */ + if (IS_GEN3(i915)) + ggtt_size = 1024*1024; + else + ggtt_size = 512*1024; + + while (ggtt_size < size) + ggtt_size <<= 1; + + return ggtt_size; +} + +/** + * i915_gem_fence_alignment - required global GTT alignment for a fence + * @i915: i915 device + * @size: object size + * @tiling: tiling mode + * @stride: tiling stride + * + * Return the required global GTT alignment for a fence (a view of a tiled + * object), taking into account potential fence register mapping. + */ +u32 i915_gem_fence_alignment(struct drm_i915_private *i915, u32 size, + unsigned int tiling, unsigned int stride) +{ + GEM_BUG_ON(!size); + + /* + * Minimum alignment is 4k (GTT page size), but might be greater + * if a fence register is needed for the object. + */ + if (tiling == I915_TILING_NONE) + return I915_GTT_MIN_ALIGNMENT; + + if (INTEL_GEN(i915) >= 4) + return I965_FENCE_PAGE; + + /* + * Previous chips need to be aligned to the size of the smallest + * fence register that can contain the object. + */ + return i915_gem_fence_size(i915, size, tiling, stride); +} + /* Check pitch constriants for all chips & tiling formats */ static bool -i915_tiling_ok(struct drm_i915_private *dev_priv, - int stride, int size, int tiling_mode) +i915_tiling_ok(struct drm_i915_gem_object *obj, + unsigned int tiling, unsigned int stride) { - int tile_width; + struct drm_i915_private *i915 = to_i915(obj->base.dev); + unsigned int tile_width; /* Linear is always fine */ - if (tiling_mode == I915_TILING_NONE) + if (tiling == I915_TILING_NONE) return true; - if (tiling_mode > I915_TILING_LAST) + if (tiling > I915_TILING_LAST) return false; - if (IS_GEN2(dev_priv) || - (tiling_mode == I915_TILING_Y && HAS_128_BYTE_Y_TILING(dev_priv))) - tile_width = 128; - else - tile_width = 512; - /* check maximum stride & object size */ /* i965+ stores the end address of the gtt mapping in the fence * reg, so dont bother to check the size */ - if (INTEL_GEN(dev_priv) >= 7) { + if (INTEL_GEN(i915) >= 7) { if (stride / 128 > GEN7_FENCE_MAX_PITCH_VAL) return false; - } else if (INTEL_GEN(dev_priv) >= 4) { + } else if (INTEL_GEN(i915) >= 4) { if (stride / 128 > I965_FENCE_MAX_PITCH_VAL) return false; } else { if (stride > 8192) return false; - if (IS_GEN3(dev_priv)) { - if (size > I830_FENCE_MAX_SIZE_VAL << 20) + if (IS_GEN3(i915)) { + if (obj->base.size > I830_FENCE_MAX_SIZE_VAL << 20) return false; } else { - if (size > I830_FENCE_MAX_SIZE_VAL << 19) + if (obj->base.size > I830_FENCE_MAX_SIZE_VAL << 19) return false; } } - if (stride < tile_width) + if (IS_GEN2(i915) || + (tiling == I915_TILING_Y && HAS_128_BYTE_Y_TILING(i915))) + tile_width = 128; + else + tile_width = 512; + + if (!IS_ALIGNED(stride, tile_width)) return false; /* 965+ just needs multiples of tile width */ - if (INTEL_GEN(dev_priv) >= 4) { - if (stride & (tile_width - 1)) - return false; + if (INTEL_GEN(i915) >= 4) return true; - } /* Pre-965 needs power of two tile widths */ - if (stride & (stride - 1)) - return false; - - return true; + return is_power_of_2(stride); } -static bool i915_vma_fence_prepare(struct i915_vma *vma, int tiling_mode) +static bool i915_vma_fence_prepare(struct i915_vma *vma, + int tiling_mode, unsigned int stride) { - struct drm_i915_private *dev_priv = vma->vm->i915; - u32 size; + struct drm_i915_private *i915 = vma->vm->i915; + u32 size, alignment; if (!i915_vma_is_map_and_fenceable(vma)) return true; - if (INTEL_GEN(dev_priv) == 3) { - if (vma->node.start & ~I915_FENCE_START_MASK) - return false; - } else { - if (vma->node.start & ~I830_FENCE_START_MASK) - return false; - } - - size = i915_gem_get_ggtt_size(dev_priv, vma->size, tiling_mode); + size = i915_gem_fence_size(i915, vma->size, tiling_mode, stride); if (vma->node.size < size) return false; - if (vma->node.start & (size - 1)) + alignment = i915_gem_fence_alignment(i915, vma->size, tiling_mode, stride); + if (!IS_ALIGNED(vma->node.start, alignment)) return false; return true; @@ -145,20 +206,20 @@ static bool i915_vma_fence_prepare(struct i915_vma *vma, int tiling_mode) /* Make the current GTT allocation valid for the change in tiling. */ static int -i915_gem_object_fence_prepare(struct drm_i915_gem_object *obj, int tiling_mode) +i915_gem_object_fence_prepare(struct drm_i915_gem_object *obj, + int tiling_mode, unsigned int stride) { - struct drm_i915_private *dev_priv = to_i915(obj->base.dev); struct i915_vma *vma; int ret; if (tiling_mode == I915_TILING_NONE) return 0; - if (INTEL_GEN(dev_priv) >= 4) - return 0; - list_for_each_entry(vma, &obj->vma_list, obj_link) { - if (i915_vma_fence_prepare(vma, tiling_mode)) + if (!i915_vma_is_ggtt(vma)) + break; + + if (i915_vma_fence_prepare(vma, tiling_mode, stride)) continue; ret = i915_vma_unbind(vma); @@ -169,8 +230,100 @@ i915_gem_object_fence_prepare(struct drm_i915_gem_object *obj, int tiling_mode) return 0; } +int +i915_gem_object_set_tiling(struct drm_i915_gem_object *obj, + unsigned int tiling, unsigned int stride) +{ + struct drm_i915_private *i915 = to_i915(obj->base.dev); + struct i915_vma *vma; + int err; + + /* Make sure we don't cross-contaminate obj->tiling_and_stride */ + BUILD_BUG_ON(I915_TILING_LAST & STRIDE_MASK); + + GEM_BUG_ON(!i915_tiling_ok(obj, tiling, stride)); + GEM_BUG_ON(!stride ^ (tiling == I915_TILING_NONE)); + lockdep_assert_held(&i915->drm.struct_mutex); + + if ((tiling | stride) == obj->tiling_and_stride) + return 0; + + if (obj->framebuffer_references) + return -EBUSY; + + /* We need to rebind the object if its current allocation + * no longer meets the alignment restrictions for its new + * tiling mode. Otherwise we can just leave it alone, but + * need to ensure that any fence register is updated before + * the next fenced (either through the GTT or by the BLT unit + * on older GPUs) access. + * + * After updating the tiling parameters, we then flag whether + * we need to update an associated fence register. Note this + * has to also include the unfenced register the GPU uses + * whilst executing a fenced command for an untiled object. + */ + + err = i915_gem_object_fence_prepare(obj, tiling, stride); + if (err) + return err; + + /* If the memory has unknown (i.e. varying) swizzling, we pin the + * pages to prevent them being swapped out and causing corruption + * due to the change in swizzling. + */ + mutex_lock(&obj->mm.lock); + if (obj->mm.pages && + obj->mm.madv == I915_MADV_WILLNEED && + i915->quirks & QUIRK_PIN_SWIZZLED_PAGES) { + if (tiling == I915_TILING_NONE) { + GEM_BUG_ON(!obj->mm.quirked); + __i915_gem_object_unpin_pages(obj); + obj->mm.quirked = false; + } + if (!i915_gem_object_is_tiled(obj)) { + GEM_BUG_ON(!obj->mm.quirked); + __i915_gem_object_pin_pages(obj); + obj->mm.quirked = true; + } + } + mutex_unlock(&obj->mm.lock); + + list_for_each_entry(vma, &obj->vma_list, obj_link) { + if (!i915_vma_is_ggtt(vma)) + break; + + vma->fence_size = + i915_gem_fence_size(i915, vma->size, tiling, stride); + vma->fence_alignment = + i915_gem_fence_alignment(i915, + vma->size, tiling, stride); + + if (vma->fence) + vma->fence->dirty = true; + } + + obj->tiling_and_stride = tiling | stride; + + /* Force the fence to be reacquired for GTT access */ + i915_gem_release_mmap(obj); + + /* Try to preallocate memory required to save swizzling on put-pages */ + if (i915_gem_object_needs_bit17_swizzle(obj)) { + if (!obj->bit_17) { + obj->bit_17 = kcalloc(BITS_TO_LONGS(obj->base.size >> PAGE_SHIFT), + sizeof(long), GFP_KERNEL); + } + } else { + kfree(obj->bit_17); + obj->bit_17 = NULL; + } + + return 0; +} + /** - * i915_gem_set_tiling - IOCTL handler to set tiling mode + * i915_gem_set_tiling_ioctl - IOCTL handler to set tiling mode * @dev: DRM device * @data: data pointer for the ioctl * @file: DRM file for the ioctl call @@ -184,30 +337,19 @@ i915_gem_object_fence_prepare(struct drm_i915_gem_object *obj, int tiling_mode) * Zero on success, negative errno on failure. */ int -i915_gem_set_tiling(struct drm_device *dev, void *data, - struct drm_file *file) +i915_gem_set_tiling_ioctl(struct drm_device *dev, void *data, + struct drm_file *file) { struct drm_i915_gem_set_tiling *args = data; - struct drm_i915_private *dev_priv = to_i915(dev); struct drm_i915_gem_object *obj; - int err = 0; - - /* Make sure we don't cross-contaminate obj->tiling_and_stride */ - BUILD_BUG_ON(I915_TILING_LAST & STRIDE_MASK); + int err; obj = i915_gem_object_lookup(file, args->handle); if (!obj) return -ENOENT; - if (!i915_tiling_ok(dev_priv, - args->stride, obj->base.size, args->tiling_mode)) { - i915_gem_object_put(obj); - return -EINVAL; - } - - mutex_lock(&dev->struct_mutex); - if (obj->pin_display || obj->framebuffer_references) { - err = -EBUSY; + if (!i915_tiling_ok(obj, args->tiling_mode, args->stride)) { + err = -EINVAL; goto err; } @@ -216,9 +358,9 @@ i915_gem_set_tiling(struct drm_device *dev, void *data, args->stride = 0; } else { if (args->tiling_mode == I915_TILING_X) - args->swizzle_mode = dev_priv->mm.bit_6_swizzle_x; + args->swizzle_mode = to_i915(dev)->mm.bit_6_swizzle_x; else - args->swizzle_mode = dev_priv->mm.bit_6_swizzle_y; + args->swizzle_mode = to_i915(dev)->mm.bit_6_swizzle_y; /* Hide bit 17 swizzling from the user. This prevents old Mesa * from aborting the application on sw fallbacks to bit 17, @@ -240,79 +382,24 @@ i915_gem_set_tiling(struct drm_device *dev, void *data, } } - if (args->tiling_mode != i915_gem_object_get_tiling(obj) || - args->stride != i915_gem_object_get_stride(obj)) { - /* We need to rebind the object if its current allocation - * no longer meets the alignment restrictions for its new - * tiling mode. Otherwise we can just leave it alone, but - * need to ensure that any fence register is updated before - * the next fenced (either through the GTT or by the BLT unit - * on older GPUs) access. - * - * After updating the tiling parameters, we then flag whether - * we need to update an associated fence register. Note this - * has to also include the unfenced register the GPU uses - * whilst executing a fenced command for an untiled object. - */ + err = mutex_lock_interruptible(&dev->struct_mutex); + if (err) + goto err; - err = i915_gem_object_fence_prepare(obj, args->tiling_mode); - if (!err) { - struct i915_vma *vma; - - mutex_lock(&obj->mm.lock); - if (obj->mm.pages && - obj->mm.madv == I915_MADV_WILLNEED && - dev_priv->quirks & QUIRK_PIN_SWIZZLED_PAGES) { - if (args->tiling_mode == I915_TILING_NONE) { - GEM_BUG_ON(!obj->mm.quirked); - __i915_gem_object_unpin_pages(obj); - obj->mm.quirked = false; - } - if (!i915_gem_object_is_tiled(obj)) { - GEM_BUG_ON(!obj->mm.quirked); - __i915_gem_object_pin_pages(obj); - obj->mm.quirked = true; - } - } - mutex_unlock(&obj->mm.lock); - - list_for_each_entry(vma, &obj->vma_list, obj_link) { - if (!vma->fence) - continue; - - vma->fence->dirty = true; - } - obj->tiling_and_stride = - args->stride | args->tiling_mode; - - /* Force the fence to be reacquired for GTT access */ - i915_gem_release_mmap(obj); - } - } - /* we have to maintain this existing ABI... */ + err = i915_gem_object_set_tiling(obj, args->tiling_mode, args->stride); + mutex_unlock(&dev->struct_mutex); + + /* We have to maintain this existing ABI... */ args->stride = i915_gem_object_get_stride(obj); args->tiling_mode = i915_gem_object_get_tiling(obj); - /* Try to preallocate memory required to save swizzling on put-pages */ - if (i915_gem_object_needs_bit17_swizzle(obj)) { - if (obj->bit_17 == NULL) { - obj->bit_17 = kcalloc(BITS_TO_LONGS(obj->base.size >> PAGE_SHIFT), - sizeof(long), GFP_KERNEL); - } - } else { - kfree(obj->bit_17); - obj->bit_17 = NULL; - } - err: i915_gem_object_put(obj); - mutex_unlock(&dev->struct_mutex); - return err; } /** - * i915_gem_get_tiling - IOCTL handler to get tiling mode + * i915_gem_get_tiling_ioctl - IOCTL handler to get tiling mode * @dev: DRM device * @data: data pointer for the ioctl * @file: DRM file for the ioctl call @@ -325,8 +412,8 @@ err: * Zero on success, negative errno on failure. */ int -i915_gem_get_tiling(struct drm_device *dev, void *data, - struct drm_file *file) +i915_gem_get_tiling_ioctl(struct drm_device *dev, void *data, + struct drm_file *file) { struct drm_i915_gem_get_tiling *args = data; struct drm_i915_private *dev_priv = to_i915(dev); diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index 396c6f0..9cd22cd 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -121,6 +121,7 @@ static void __i915_error_advance(struct drm_i915_error_state_buf *e, e->pos += len; } +__printf(2, 0) static void i915_error_vprintf(struct drm_i915_error_state_buf *e, const char *f, va_list args) { diff --git a/drivers/gpu/drm/i915/i915_guc_reg.h b/drivers/gpu/drm/i915/i915_guc_reg.h index 6a0adaf..35cf991 100644 --- a/drivers/gpu/drm/i915/i915_guc_reg.h +++ b/drivers/gpu/drm/i915/i915_guc_reg.h @@ -61,12 +61,18 @@ #define DMA_ADDRESS_SPACE_GTT (8 << 16) #define DMA_COPY_SIZE _MMIO(0xc310) #define DMA_CTRL _MMIO(0xc314) +#define HUC_UKERNEL (1<<9) #define UOS_MOVE (1<<4) #define START_DMA (1<<0) #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_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 */ diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c index 710fbb9..8ced9e2 100644 --- a/drivers/gpu/drm/i915/i915_guc_submission.c +++ b/drivers/gpu/drm/i915/i915_guc_submission.c @@ -22,8 +22,6 @@ * */ #include <linux/circ_buf.h> -#include <linux/debugfs.h> -#include <linux/relay.h> #include "i915_drv.h" #include "intel_uc.h" @@ -545,7 +543,7 @@ static void i915_guc_submit(struct drm_i915_gem_request *rq) */ /** - * guc_allocate_vma() - Allocate a GGTT VMA for GuC usage + * 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) * @@ -557,7 +555,7 @@ static void i915_guc_submit(struct drm_i915_gem_request *rq) * * Return: A i915_vma if successful, otherwise an ERR_PTR. */ -static struct i915_vma *guc_allocate_vma(struct intel_guc *guc, u32 size) +struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size) { struct drm_i915_private *dev_priv = guc_to_i915(guc); struct drm_i915_gem_object *obj; @@ -568,7 +566,7 @@ static struct i915_vma *guc_allocate_vma(struct intel_guc *guc, u32 size) if (IS_ERR(obj)) return ERR_CAST(obj); - vma = i915_vma_create(obj, &dev_priv->ggtt.base, NULL); + vma = i915_vma_instance(obj, &dev_priv->ggtt.base, NULL); if (IS_ERR(vma)) goto err; @@ -579,9 +577,6 @@ static struct i915_vma *guc_allocate_vma(struct intel_guc *guc, u32 size) goto err; } - /* Invalidate GuC TLB to let GuC take the latest updates to GTT. */ - I915_WRITE(GEN8_GTCR, GEN8_GTCR_INVALIDATE); - return vma; err: @@ -721,7 +716,7 @@ guc_client_alloc(struct drm_i915_private *dev_priv, } /* The first page is doorbell/proc_desc. Two followed pages are wq. */ - vma = guc_allocate_vma(guc, GUC_DB_SIZE + GUC_WQ_SIZE); + vma = intel_guc_allocate_vma(guc, GUC_DB_SIZE + GUC_WQ_SIZE); if (IS_ERR(vma)) goto err; @@ -777,488 +772,7 @@ err: return NULL; } -/* - * 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. - */ -static int subbuf_start_callback(struct rchan_buf *buf, - void *subbuf, - void *prev_subbuf, - size_t prev_padding) -{ - /* Use no-overwrite mode by default, where relay will stop accepting - * new data if there are no empty sub buffers left. - * There is no strict synchronization enforced by relay between Consumer - * and Producer. In overwrite mode, there is a possibility of getting - * inconsistent/garbled data, the producer could be writing on to the - * same sub buffer from which Consumer is reading. This can't be avoided - * unless Consumer is fast enough and can always run in tandem with - * Producer. - */ - if (relay_buf_full(buf)) - return 0; - - return 1; -} - -/* - * file_create() callback. Creates relay file in debugfs. - */ -static struct dentry *create_buf_file_callback(const char *filename, - struct dentry *parent, - umode_t mode, - struct rchan_buf *buf, - int *is_global) -{ - struct dentry *buf_file; - - /* This to enable the use of a single buffer for the relay channel and - * correspondingly have a single file exposed to User, through which - * it can collect the logs in order without any post-processing. - * Need to set 'is_global' even if parent is NULL for early logging. - */ - *is_global = 1; - - 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, - parent, buf, &relay_file_operations); - return buf_file; -} - -/* - * file_remove() default callback. Removes relay file in debugfs. - */ -static int remove_buf_file_callback(struct dentry *dentry) -{ - debugfs_remove(dentry); - return 0; -} - -/* relay channel callbacks */ -static struct rchan_callbacks relay_callbacks = { - .subbuf_start = subbuf_start_callback, - .create_buf_file = create_buf_file_callback, - .remove_buf_file = remove_buf_file_callback, -}; - -static void guc_log_remove_relay_file(struct intel_guc *guc) -{ - relay_close(guc->log.relay_chan); -} - -static int guc_log_create_relay_channel(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct rchan *guc_log_relay_chan; - size_t n_subbufs, subbuf_size; - - /* Keep the size of sub buffers same as shared log buffer */ - subbuf_size = guc->log.vma->obj->base.size; - - /* Store up to 8 snapshots, which is large enough to buffer sufficient - * boot time logs and provides enough leeway to User, in terms of - * latency, for consuming the logs from relay. Also doesn't take - * up too much memory. - */ - n_subbufs = 8; - - guc_log_relay_chan = relay_open(NULL, NULL, subbuf_size, - n_subbufs, &relay_callbacks, dev_priv); - if (!guc_log_relay_chan) { - DRM_ERROR("Couldn't create relay chan for GuC logging\n"); - return -ENOMEM; - } - - GEM_BUG_ON(guc_log_relay_chan->subbuf_size < subbuf_size); - guc->log.relay_chan = guc_log_relay_chan; - return 0; -} - -static int guc_log_create_relay_file(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct dentry *log_dir; - int ret; - - /* 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(guc->log.relay_chan, "guc_log", log_dir); - if (ret) { - 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 *guc) -{ - /* Make sure the updates made in the sub buffer are visible when - * Consumer sees the following update to offset inside the sub buffer. - */ - smp_wmb(); - - /* All data has been written, so now move the offset of sub buffer. */ - relay_reserve(guc->log.relay_chan, guc->log.vma->obj->base.size); - - /* Switch to the next sub buffer */ - relay_flush(guc->log.relay_chan); -} - -static void *guc_get_write_buffer(struct intel_guc *guc) -{ - if (!guc->log.relay_chan) - 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 - * buffers are full. Could have used the relay_write() to indirectly - * copy the data, but that would have been bit convoluted, as we need to - * write to only certain locations inside a sub buffer which cannot be - * done without using relay_reserve() along with relay_write(). So its - * better to use relay_reserve() alone. - */ - return relay_reserve(guc->log.relay_chan, 0); -} - -static bool -guc_check_log_buf_overflow(struct intel_guc *guc, - enum guc_log_buffer_type type, unsigned int full_cnt) -{ - unsigned int prev_full_cnt = guc->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; - - if (full_cnt < prev_full_cnt) { - /* buffer_full_cnt is a 4 bit counter */ - guc->log.total_overflow_count[type] += 16; - } - DRM_ERROR_RATELIMITED("GuC log buffer overflow\n"); - } - - return overflow; -} - -static unsigned int guc_get_log_buffer_size(enum guc_log_buffer_type type) -{ - switch (type) { - case GUC_ISR_LOG_BUFFER: - return (GUC_LOG_ISR_PAGES + 1) * PAGE_SIZE; - case GUC_DPC_LOG_BUFFER: - return (GUC_LOG_DPC_PAGES + 1) * PAGE_SIZE; - case GUC_CRASH_DUMP_LOG_BUFFER: - return (GUC_LOG_CRASH_PAGES + 1) * PAGE_SIZE; - default: - MISSING_CASE(type); - } - - return 0; -} - -static void guc_read_update_log_buffer(struct intel_guc *guc) -{ - 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; - struct guc_log_buffer_state log_buf_state_local; - enum guc_log_buffer_type type; - void *src_data, *dst_data; - bool new_overflow; - - if (WARN_ON(!guc->log.buf_addr)) - return; - - /* Get the pointer to shared GuC log buffer */ - log_buf_state = src_data = guc->log.buf_addr; - - /* Get the pointer to local buffer to store the logs */ - log_buf_snapshot_state = dst_data = guc_get_write_buffer(guc); - - /* Actual logs are present from the 2nd page */ - src_data += PAGE_SIZE; - dst_data += PAGE_SIZE; - - for (type = GUC_ISR_LOG_BUFFER; type < GUC_MAX_LOG_BUFFER; type++) { - /* Make a copy of the state structure, inside GuC log buffer - * (which is uncached mapped), on the stack to avoid reading - * from it multiple times. - */ - memcpy(&log_buf_state_local, log_buf_state, - sizeof(struct guc_log_buffer_state)); - buffer_size = guc_get_log_buffer_size(type); - read_offset = log_buf_state_local.read_ptr; - write_offset = log_buf_state_local.sampled_write_ptr; - 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); - - /* Update the state of shared log buffer */ - log_buf_state->read_ptr = write_offset; - log_buf_state->flush_to_file = 0; - log_buf_state++; - - if (unlikely(!log_buf_snapshot_state)) - continue; - - /* First copy the state structure in snapshot buffer */ - memcpy(log_buf_snapshot_state, &log_buf_state_local, - sizeof(struct guc_log_buffer_state)); - - /* The write pointer could have been updated by GuC firmware, - * after sending the flush interrupt to Host, for consistency - * set write pointer value to same value of sampled_write_ptr - * in the snapshot buffer. - */ - log_buf_snapshot_state->write_ptr = write_offset; - log_buf_snapshot_state++; - - /* Now copy the actual logs. */ - if (unlikely(new_overflow)) { - /* copy the whole buffer in case of overflow */ - read_offset = 0; - write_offset = buffer_size; - } else if (unlikely((read_offset > buffer_size) || - (write_offset > buffer_size))) { - DRM_ERROR("invalid log buffer state\n"); - /* copy whole buffer as offsets are unreliable */ - read_offset = 0; - write_offset = buffer_size; - } - - /* Just copy the newly written data */ - if (read_offset > write_offset) { - i915_memcpy_from_wc(dst_data, src_data, write_offset); - bytes_to_copy = buffer_size - read_offset; - } else { - bytes_to_copy = write_offset - read_offset; - } - i915_memcpy_from_wc(dst_data + read_offset, - src_data + read_offset, bytes_to_copy); - - src_data += buffer_size; - dst_data += buffer_size; - } - - if (log_buf_snapshot_state) - guc_move_to_next_buf(guc); - else { - /* Used rate limited to avoid deluge of messages, logs might be - * getting consumed by User at a slow rate. - */ - DRM_ERROR_RATELIMITED("no sub-buffer to capture logs\n"); - guc->log.capture_miss_count++; - } -} - -static void guc_capture_logs_work(struct work_struct *work) -{ - struct drm_i915_private *dev_priv = - container_of(work, struct drm_i915_private, guc.log.flush_work); - - i915_guc_capture_logs(dev_priv); -} - -static void guc_log_cleanup(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - - lockdep_assert_held(&dev_priv->drm.struct_mutex); - - /* First disable the flush interrupt */ - gen9_disable_guc_interrupts(dev_priv); - if (guc->log.flush_wq) - destroy_workqueue(guc->log.flush_wq); - - guc->log.flush_wq = NULL; - - if (guc->log.relay_chan) - guc_log_remove_relay_file(guc); - - guc->log.relay_chan = NULL; - - if (guc->log.buf_addr) - i915_gem_object_unpin_map(guc->log.vma->obj); - - guc->log.buf_addr = NULL; -} - -static int guc_log_create_extras(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - void *vaddr; - int ret; - - lockdep_assert_held(&dev_priv->drm.struct_mutex); - - /* Nothing to do */ - if (i915.guc_log_level < 0) - return 0; - - if (!guc->log.buf_addr) { - /* Create a WC (Uncached for read) vmalloc mapping of log - * 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); - if (IS_ERR(vaddr)) { - ret = PTR_ERR(vaddr); - DRM_ERROR("Couldn't map log buffer pages %d\n", ret); - return ret; - } - - guc->log.buf_addr = vaddr; - } - - if (!guc->log.relay_chan) { - /* 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. - */ - ret = guc_log_create_relay_channel(guc); - if (ret) - return ret; - } - - if (!guc->log.flush_wq) { - INIT_WORK(&guc->log.flush_work, guc_capture_logs_work); - - /* - * GuC log buffer flush work item has to do register access to - * send the ack to GuC and this work item, if not synced before - * suspend, can potentially get executed after the GFX device is - * suspended. - * By marking the WQ as freezable, we don't have to bother about - * flushing of this work item from the suspend hooks, the pending - * work item if any will be either executed before the suspend - * or scheduled later on resume. This way the handling of work - * item can be kept same between system suspend & rpm suspend. - */ - guc->log.flush_wq = alloc_ordered_workqueue("i915-guc_log", - WQ_HIGHPRI | WQ_FREEZABLE); - if (guc->log.flush_wq == NULL) { - DRM_ERROR("Couldn't allocate the wq for GuC logging\n"); - return -ENOMEM; - } - } - - return 0; -} - -static void guc_log_create(struct intel_guc *guc) -{ - struct i915_vma *vma; - unsigned long offset; - uint32_t size, flags; - - if (i915.guc_log_level > GUC_LOG_VERBOSITY_MAX) - i915.guc_log_level = GUC_LOG_VERBOSITY_MAX; - - /* The first page is to save log buffer state. Allocate one - * extra page for others in case for overlap */ - size = (1 + GUC_LOG_DPC_PAGES + 1 + - GUC_LOG_ISR_PAGES + 1 + - GUC_LOG_CRASH_PAGES + 1) << PAGE_SHIFT; - - vma = guc->log.vma; - if (!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())) { - /* logging will not be enabled */ - i915.guc_log_level = -1; - return; - } - - vma = guc_allocate_vma(guc, size); - if (IS_ERR(vma)) { - /* logging will be off */ - i915.guc_log_level = -1; - return; - } - - guc->log.vma = vma; - - if (guc_log_create_extras(guc)) { - guc_log_cleanup(guc); - i915_vma_unpin_and_release(&guc->log.vma); - i915.guc_log_level = -1; - return; - } - } - - /* each allocated unit is a page */ - flags = GUC_LOG_VALID | GUC_LOG_NOTIFY_ON_HALF_FULL | - (GUC_LOG_DPC_PAGES << GUC_LOG_DPC_SHIFT) | - (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 */ - guc->log.flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags; -} - -static int guc_log_late_setup(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - int ret; - - lockdep_assert_held(&dev_priv->drm.struct_mutex); - - if (i915.guc_log_level < 0) - return -EINVAL; - - /* If log_level was set as -1 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_create_extras(guc); - if (ret) - goto err; - - ret = guc_log_create_relay_file(guc); - if (ret) - goto err; - - return 0; -err: - guc_log_cleanup(guc); - /* logging will remain off */ - i915.guc_log_level = -1; - return ret; -} static void guc_policies_init(struct guc_policies *policies) { @@ -1301,7 +815,7 @@ static void guc_addon_create(struct intel_guc *guc) vma = guc->ads_vma; if (!vma) { - vma = guc_allocate_vma(guc, PAGE_ALIGN(size)); + vma = intel_guc_allocate_vma(guc, PAGE_ALIGN(size)); if (IS_ERR(vma)) return; @@ -1376,13 +890,13 @@ int i915_guc_submission_init(struct drm_i915_private *dev_priv) if (guc->ctx_pool_vma) return 0; /* already allocated */ - vma = guc_allocate_vma(guc, gemsize); + vma = intel_guc_allocate_vma(guc, gemsize); if (IS_ERR(vma)) return PTR_ERR(vma); guc->ctx_pool_vma = vma; ida_init(&guc->ctx_ids); - guc_log_create(guc); + intel_guc_log_create(guc); guc_addon_create(guc); guc->execbuf_client = guc_client_alloc(dev_priv, @@ -1484,7 +998,7 @@ int intel_guc_suspend(struct drm_i915_private *dev_priv) struct i915_gem_context *ctx; u32 data[3]; - if (guc->guc_fw.guc_fw_load_status != GUC_FIRMWARE_SUCCESS) + if (guc->fw.load_status != INTEL_UC_FIRMWARE_SUCCESS) return 0; gen9_disable_guc_interrupts(dev_priv); @@ -1511,7 +1025,7 @@ int intel_guc_resume(struct drm_i915_private *dev_priv) struct i915_gem_context *ctx; u32 data[3]; - if (guc->guc_fw.guc_fw_load_status != GUC_FIRMWARE_SUCCESS) + if (guc->fw.load_status != INTEL_UC_FIRMWARE_SUCCESS) return 0; if (i915.guc_log_level >= 0) @@ -1527,103 +1041,4 @@ int intel_guc_resume(struct drm_i915_private *dev_priv) return intel_guc_send(guc, data, ARRAY_SIZE(data)); } -void i915_guc_capture_logs(struct drm_i915_private *dev_priv) -{ - guc_read_update_log_buffer(&dev_priv->guc); - - /* Generally device is expected to be active only at this - * time, so get/put should be really quick. - */ - intel_runtime_pm_get(dev_priv); - intel_guc_log_flush_complete(&dev_priv->guc); - intel_runtime_pm_put(dev_priv); -} - -void i915_guc_flush_logs(struct drm_i915_private *dev_priv) -{ - if (!i915.enable_guc_submission || (i915.guc_log_level < 0)) - return; - - /* First disable the interrupts, will be renabled afterwards */ - gen9_disable_guc_interrupts(dev_priv); - - /* Before initiating the forceful flush, wait for any pending/ongoing - * flush to complete otherwise forceful flush may not actually happen. - */ - flush_work(&dev_priv->guc.log.flush_work); - - /* Ask GuC to update the log buffer state */ - intel_guc_log_flush(&dev_priv->guc); - - /* GuC would have updated log buffer by now, so capture it */ - i915_guc_capture_logs(dev_priv); -} - -void i915_guc_unregister(struct drm_i915_private *dev_priv) -{ - if (!i915.enable_guc_submission) - return; - mutex_lock(&dev_priv->drm.struct_mutex); - guc_log_cleanup(&dev_priv->guc); - mutex_unlock(&dev_priv->drm.struct_mutex); -} - -void i915_guc_register(struct drm_i915_private *dev_priv) -{ - if (!i915.enable_guc_submission) - return; - - mutex_lock(&dev_priv->drm.struct_mutex); - guc_log_late_setup(&dev_priv->guc); - mutex_unlock(&dev_priv->drm.struct_mutex); -} - -int i915_guc_log_control(struct drm_i915_private *dev_priv, u64 control_val) -{ - union guc_log_control log_param; - int ret; - - log_param.value = control_val; - - if (log_param.verbosity < GUC_LOG_VERBOSITY_MIN || - log_param.verbosity > GUC_LOG_VERBOSITY_MAX) - return -EINVAL; - - /* This combination doesn't make sense & won't have any effect */ - if (!log_param.logging_enabled && (i915.guc_log_level < 0)) - return 0; - - ret = intel_guc_log_control(&dev_priv->guc, log_param.value); - if (ret < 0) { - DRM_DEBUG_DRIVER("guc_logging_control action failed %d\n", ret); - return ret; - } - - i915.guc_log_level = log_param.verbosity; - - /* If log_level was set as -1 at boot time, then the relay channel file - * wouldn't have been created by now and interrupts also would not have - * been enabled. - */ - if (!dev_priv->guc.log.relay_chan) { - ret = guc_log_late_setup(&dev_priv->guc); - if (!ret) - gen9_enable_guc_interrupts(dev_priv); - } else if (!log_param.logging_enabled) { - /* 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. - */ - i915_guc_flush_logs(dev_priv); - - /* As logging is disabled, update log level to reflect that */ - i915.guc_log_level = -1; - } else { - /* In case interrupts were disabled, enable them now */ - gen9_enable_guc_interrupts(dev_priv); - } - - return ret; -} diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index a0e70f5..6fefc34 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -1170,6 +1170,9 @@ static void gen6_pm_rps_work(struct work_struct *work) adj *= 2; else /* CHV needs even encode values */ adj = IS_CHERRYVIEW(dev_priv) ? 2 : 1; + + if (new_delay >= dev_priv->rps.max_freq_softlimit) + adj = 0; /* * For better performance, jump directly * to RPe if we're below it. @@ -1191,6 +1194,9 @@ static void gen6_pm_rps_work(struct work_struct *work) adj *= 2; else /* CHV needs even encode values */ adj = IS_CHERRYVIEW(dev_priv) ? -2 : -1; + + if (new_delay <= dev_priv->rps.min_freq_softlimit) + adj = 0; } else { /* unknown event */ adj = 0; } @@ -1553,41 +1559,68 @@ static void display_pipe_crc_irq_handler(struct drm_i915_private *dev_priv, { struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[pipe]; struct intel_pipe_crc_entry *entry; + struct intel_crtc *crtc = intel_get_crtc_for_pipe(dev_priv, pipe); + struct drm_driver *driver = dev_priv->drm.driver; + uint32_t crcs[5]; int head, tail; spin_lock(&pipe_crc->lock); + if (pipe_crc->source) { + if (!pipe_crc->entries) { + spin_unlock(&pipe_crc->lock); + DRM_DEBUG_KMS("spurious interrupt\n"); + return; + } - if (!pipe_crc->entries) { - spin_unlock(&pipe_crc->lock); - DRM_DEBUG_KMS("spurious interrupt\n"); - return; - } - - head = pipe_crc->head; - tail = pipe_crc->tail; + head = pipe_crc->head; + tail = pipe_crc->tail; - if (CIRC_SPACE(head, tail, INTEL_PIPE_CRC_ENTRIES_NR) < 1) { - spin_unlock(&pipe_crc->lock); - DRM_ERROR("CRC buffer overflowing\n"); - return; - } + if (CIRC_SPACE(head, tail, INTEL_PIPE_CRC_ENTRIES_NR) < 1) { + spin_unlock(&pipe_crc->lock); + DRM_ERROR("CRC buffer overflowing\n"); + return; + } - entry = &pipe_crc->entries[head]; + entry = &pipe_crc->entries[head]; - entry->frame = dev_priv->drm.driver->get_vblank_counter(&dev_priv->drm, - pipe); - entry->crc[0] = crc0; - entry->crc[1] = crc1; - entry->crc[2] = crc2; - entry->crc[3] = crc3; - entry->crc[4] = crc4; + entry->frame = driver->get_vblank_counter(&dev_priv->drm, pipe); + entry->crc[0] = crc0; + entry->crc[1] = crc1; + entry->crc[2] = crc2; + entry->crc[3] = crc3; + entry->crc[4] = crc4; - head = (head + 1) & (INTEL_PIPE_CRC_ENTRIES_NR - 1); - pipe_crc->head = head; + head = (head + 1) & (INTEL_PIPE_CRC_ENTRIES_NR - 1); + pipe_crc->head = head; - spin_unlock(&pipe_crc->lock); + spin_unlock(&pipe_crc->lock); - wake_up_interruptible(&pipe_crc->wq); + wake_up_interruptible(&pipe_crc->wq); + } else { + /* + * For some not yet identified reason, the first CRC is + * bonkers. So let's just wait for the next vblank and read + * out the buggy result. + * + * On CHV sometimes the second CRC is bonkers as well, so + * don't trust that one either. + */ + if (pipe_crc->skipped == 0 || + (IS_CHERRYVIEW(dev_priv) && pipe_crc->skipped == 1)) { + pipe_crc->skipped++; + spin_unlock(&pipe_crc->lock); + return; + } + spin_unlock(&pipe_crc->lock); + crcs[0] = crc0; + crcs[1] = crc1; + crcs[2] = crc2; + crcs[3] = crc3; + crcs[4] = crc4; + drm_crtc_add_crc_entry(&crtc->base, true, + drm_accurate_vblank_count(&crtc->base), + crcs); + } } #else static inline void @@ -2703,12 +2736,13 @@ 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 + * @fmt: Error message format string + * * Do some basic checking of register state at error time and * dump it to the syslog. Also call i915_capture_error_state() to make * sure we get a record and make it available in debugfs. Fire a uevent * so userspace knows something bad happened (should trigger collection * of a ring dump etc.). - * @fmt: Error message format string */ void i915_handle_error(struct drm_i915_private *dev_priv, u32 engine_mask, diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 00970aa..72f9f36 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -3597,9 +3597,12 @@ enum { #define EDP_PSR_PERF_CNT_MASK 0xffffff #define EDP_PSR_DEBUG_CTL _MMIO(dev_priv->psr_mmio_base + 0x60) -#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_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_PSR2_CTL _MMIO(0x6f900) #define EDP_PSR2_ENABLE (1<<31) @@ -3614,6 +3617,11 @@ enum { #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_FRAMES_BEFORE_SU_ENTRY (1<<4) + +#define EDP_PSR2_STATUS_CTL _MMIO(0x6f940) +#define EDP_PSR2_STATUS_STATE_MASK (0xf<<28) +#define EDP_PSR2_STATUS_STATE_SHIFT 28 /* VGA port control */ #define ADPA _MMIO(0x61100) @@ -6446,6 +6454,12 @@ enum { #define BDW_DPRS_MASK_VBLANK_SRD (1 << 0) #define CHICKEN_PIPESL_1(pipe) _MMIO_PIPE(pipe, _CHICKEN_PIPESL_1_A, _CHICKEN_PIPESL_1_B) +#define CHICKEN_TRANS_A 0x420c0 +#define CHICKEN_TRANS_B 0x420c4 +#define CHICKEN_TRANS(trans) _MMIO_TRANS(trans, CHICKEN_TRANS_A, CHICKEN_TRANS_B) +#define PSR2_VSC_ENABLE_PROG_HEADER (1<<12) +#define PSR2_ADD_VERTICAL_LINE_COUNT (1<<15) + #define DISP_ARB_CTL _MMIO(0x45000) #define DISP_FBC_MEMORY_WAKE (1<<31) #define DISP_TILE_SURFACE_SWIZZLING (1<<13) diff --git a/drivers/gpu/drm/i915/i915_sw_fence.c b/drivers/gpu/drm/i915/i915_sw_fence.c index f5a8809..40f4e5e 100644 --- a/drivers/gpu/drm/i915/i915_sw_fence.c +++ b/drivers/gpu/drm/i915/i915_sw_fence.c @@ -63,6 +63,7 @@ static inline void debug_fence_destroy(struct i915_sw_fence *fence) static inline void debug_fence_free(struct i915_sw_fence *fence) { debug_object_free(fence, &i915_sw_fence_debug_descr); + smp_wmb(); /* flush the change in state before reallocation */ } static inline void debug_fence_assert(struct i915_sw_fence *fence) diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c index 40c0ac70..376ac95 100644 --- a/drivers/gpu/drm/i915/i915_sysfs.c +++ b/drivers/gpu/drm/i915/i915_sysfs.c @@ -58,7 +58,7 @@ static u32 calc_residency(struct drm_i915_private *dev_priv, if (I915_READ(VLV_COUNTER_CONTROL) & VLV_COUNT_RANGE_HIGH) units <<= 8; - } else if (IS_BROXTON(dev_priv)) { + } else if (IS_GEN9_LP(dev_priv)) { units = 1; div = 1200; /* 833.33ns */ } diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h index 18ae37c..4461df5 100644 --- a/drivers/gpu/drm/i915/i915_trace.h +++ b/drivers/gpu/drm/i915/i915_trace.h @@ -450,9 +450,9 @@ TRACE_EVENT(i915_gem_evict_vm, TP_printk("dev=%d, vm=%p", __entry->dev, __entry->vm) ); -TRACE_EVENT(i915_gem_evict_vma, - TP_PROTO(struct i915_vma *vma, unsigned int flags), - TP_ARGS(vma, flags), +TRACE_EVENT(i915_gem_evict_node, + TP_PROTO(struct i915_address_space *vm, struct drm_mm_node *node, unsigned int flags), + TP_ARGS(vm, node, flags), TP_STRUCT__entry( __field(u32, dev) @@ -464,11 +464,11 @@ TRACE_EVENT(i915_gem_evict_vma, ), TP_fast_assign( - __entry->dev = vma->vm->i915->drm.primary->index; - __entry->vm = vma->vm; - __entry->start = vma->node.start; - __entry->size = vma->node.size; - __entry->color = vma->node.color; + __entry->dev = vm->i915->drm.primary->index; + __entry->vm = vm; + __entry->start = node->start; + __entry->size = node->size; + __entry->color = node->color; __entry->flags = flags; ), diff --git a/drivers/gpu/drm/i915/i915_vgpu.c b/drivers/gpu/drm/i915/i915_vgpu.c index dae340c..d0abfd0 100644 --- a/drivers/gpu/drm/i915/i915_vgpu.c +++ b/drivers/gpu/drm/i915/i915_vgpu.c @@ -116,22 +116,20 @@ void intel_vgt_deballoon(struct drm_i915_private *dev_priv) memset(&bl_info, 0, sizeof(bl_info)); } -static int vgt_balloon_space(struct drm_mm *mm, +static int vgt_balloon_space(struct i915_ggtt *ggtt, struct drm_mm_node *node, unsigned long start, unsigned long end) { unsigned long size = end - start; - if (start == end) + if (start >= end) return -EINVAL; DRM_INFO("balloon space: range [ 0x%lx - 0x%lx ] %lu KiB.\n", start, end, size / 1024); - - node->start = start; - node->size = size; - - return drm_mm_reserve_node(mm, node); + return i915_gem_gtt_reserve(&ggtt->base, node, + size, start, I915_COLOR_UNEVICTABLE, + 0); } /** @@ -214,10 +212,8 @@ int intel_vgt_balloon(struct drm_i915_private *dev_priv) /* Unmappable graphic memory ballooning */ if (unmappable_base > ggtt->mappable_end) { - ret = vgt_balloon_space(&ggtt->base.mm, - &bl_info.space[2], - ggtt->mappable_end, - unmappable_base); + ret = vgt_balloon_space(ggtt, &bl_info.space[2], + ggtt->mappable_end, unmappable_base); if (ret) goto err; @@ -228,18 +224,15 @@ int intel_vgt_balloon(struct drm_i915_private *dev_priv) * because it is reserved to the guard page. */ if (unmappable_end < ggtt_end - PAGE_SIZE) { - ret = vgt_balloon_space(&ggtt->base.mm, - &bl_info.space[3], - unmappable_end, - ggtt_end - PAGE_SIZE); + ret = vgt_balloon_space(ggtt, &bl_info.space[3], + unmappable_end, ggtt_end - PAGE_SIZE); if (ret) goto err; } /* Mappable graphic memory ballooning */ if (mappable_base > ggtt->base.start) { - ret = vgt_balloon_space(&ggtt->base.mm, - &bl_info.space[0], + ret = vgt_balloon_space(ggtt, &bl_info.space[0], ggtt->base.start, mappable_base); if (ret) @@ -247,10 +240,8 @@ int intel_vgt_balloon(struct drm_i915_private *dev_priv) } if (mappable_end < ggtt->mappable_end) { - ret = vgt_balloon_space(&ggtt->base.mm, - &bl_info.space[1], - mappable_end, - ggtt->mappable_end); + ret = vgt_balloon_space(ggtt, &bl_info.space[1], + mappable_end, ggtt->mappable_end); if (ret) goto err; diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index 58f2483..307b22a 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -45,6 +45,7 @@ i915_vma_retire(struct i915_gem_active *active, if (i915_vma_is_active(vma)) return; + GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); list_move_tail(&vma->vm_link, &vma->vm->inactive_list); if (unlikely(i915_vma_is_closed(vma) && !i915_vma_is_pinned(vma))) WARN_ON(i915_vma_unbind(vma)); @@ -69,17 +70,15 @@ i915_vma_retire(struct i915_gem_active *active, } static struct i915_vma * -__i915_vma_create(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view) +vma_create(struct drm_i915_gem_object *obj, + struct i915_address_space *vm, + const struct i915_ggtt_view *view) { struct i915_vma *vma; struct rb_node *rb, **p; int i; - GEM_BUG_ON(vm->closed); - - vma = kmem_cache_zalloc(to_i915(obj->base.dev)->vmas, GFP_KERNEL); + vma = kmem_cache_zalloc(vm->i915->vmas, GFP_KERNEL); if (vma == NULL) return ERR_PTR(-ENOMEM); @@ -87,29 +86,50 @@ __i915_vma_create(struct drm_i915_gem_object *obj, for (i = 0; i < ARRAY_SIZE(vma->last_read); i++) init_request_active(&vma->last_read[i], i915_vma_retire); init_request_active(&vma->last_fence, NULL); - list_add(&vma->vm_link, &vm->unbound_list); vma->vm = vm; vma->obj = obj; vma->size = obj->base.size; + vma->display_alignment = I915_GTT_MIN_ALIGNMENT; if (view) { vma->ggtt_view = *view; if (view->type == I915_GGTT_VIEW_PARTIAL) { GEM_BUG_ON(range_overflows_t(u64, - view->params.partial.offset, - view->params.partial.size, + view->partial.offset, + view->partial.size, obj->base.size >> PAGE_SHIFT)); - vma->size = view->params.partial.size; + vma->size = view->partial.size; vma->size <<= PAGE_SHIFT; GEM_BUG_ON(vma->size >= obj->base.size); } else if (view->type == I915_GGTT_VIEW_ROTATED) { - vma->size = - intel_rotation_info_size(&view->params.rotated); + vma->size = intel_rotation_info_size(&view->rotated); vma->size <<= PAGE_SHIFT; } } + if (unlikely(vma->size > vm->total)) + goto err_vma; + + GEM_BUG_ON(!IS_ALIGNED(vma->size, I915_GTT_PAGE_SIZE)); + if (i915_is_ggtt(vm)) { + if (unlikely(overflows_type(vma->size, u32))) + goto err_vma; + + vma->fence_size = i915_gem_fence_size(vm->i915, vma->size, + i915_gem_object_get_tiling(obj), + i915_gem_object_get_stride(obj)); + if (unlikely(vma->fence_size < vma->size || /* overflow */ + vma->fence_size > vm->total)) + goto err_vma; + + GEM_BUG_ON(!IS_ALIGNED(vma->fence_size, I915_GTT_MIN_ALIGNMENT)); + + vma->fence_alignment = i915_gem_fence_alignment(vm->i915, vma->size, + i915_gem_object_get_tiling(obj), + i915_gem_object_get_stride(obj)); + GEM_BUG_ON(!is_power_of_2(vma->fence_alignment)); + vma->flags |= I915_VMA_GGTT; list_add(&vma->obj_link, &obj->vma_list); } else { @@ -131,20 +151,74 @@ __i915_vma_create(struct drm_i915_gem_object *obj, } rb_link_node(&vma->obj_node, rb, p); rb_insert_color(&vma->obj_node, &obj->vma_tree); + list_add(&vma->vm_link, &vm->unbound_list); return vma; + +err_vma: + kmem_cache_free(vm->i915->vmas, vma); + return ERR_PTR(-E2BIG); } +static struct i915_vma * +vma_lookup(struct drm_i915_gem_object *obj, + struct i915_address_space *vm, + const struct i915_ggtt_view *view) +{ + struct rb_node *rb; + + rb = obj->vma_tree.rb_node; + while (rb) { + struct i915_vma *vma = rb_entry(rb, struct i915_vma, obj_node); + long cmp; + + cmp = i915_vma_compare(vma, vm, view); + if (cmp == 0) + return vma; + + if (cmp < 0) + rb = rb->rb_right; + else + rb = rb->rb_left; + } + + return NULL; +} + +/** + * i915_vma_instance - return the singleton instance of the VMA + * @obj: parent &struct drm_i915_gem_object to be mapped + * @vm: address space in which the mapping is located + * @view: additional mapping requirements + * + * i915_vma_instance() looks up an existing VMA of the @obj in the @vm with + * the same @view characteristics. If a match is not found, one is created. + * Once created, the VMA is kept until either the object is freed, or the + * address space is closed. + * + * Must be called with struct_mutex held. + * + * Returns the vma, or an error pointer. + */ struct i915_vma * -i915_vma_create(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view) +i915_vma_instance(struct drm_i915_gem_object *obj, + struct i915_address_space *vm, + const struct i915_ggtt_view *view) { + struct i915_vma *vma; + lockdep_assert_held(&obj->base.dev->struct_mutex); GEM_BUG_ON(view && !i915_is_ggtt(vm)); - GEM_BUG_ON(i915_gem_obj_to_vma(obj, vm, view)); + GEM_BUG_ON(vm->closed); + + vma = vma_lookup(obj, vm, view); + if (!vma) + vma = vma_create(obj, vm, view); - return __i915_vma_create(obj, vm, view); + GEM_BUG_ON(!IS_ERR(vma) && i915_vma_is_closed(vma)); + GEM_BUG_ON(!IS_ERR(vma) && i915_vma_compare(vma, vm, view)); + GEM_BUG_ON(!IS_ERR(vma) && vma_lookup(obj, vm, view) != vma); + return vma; } /** @@ -195,6 +269,7 @@ int i915_vma_bind(struct i915_vma *vma, enum i915_cache_level cache_level, return ret; } + trace_i915_vma_bind(vma, bind_flags); ret = vma->vm->bind_vma(vma, cache_level, bind_flags); if (ret) return ret; @@ -258,7 +333,8 @@ i915_vma_misplaced(struct i915_vma *vma, u64 size, u64 alignment, u64 flags) if (vma->node.size < size) return true; - if (alignment && vma->node.start & (alignment - 1)) + GEM_BUG_ON(alignment && !is_power_of_2(alignment)); + if (alignment && !IS_ALIGNED(vma->node.start, alignment)) return true; if (flags & PIN_MAPPABLE && !i915_vma_is_map_and_fenceable(vma)) @@ -277,31 +353,24 @@ i915_vma_misplaced(struct i915_vma *vma, u64 size, u64 alignment, u64 flags) void __i915_vma_set_map_and_fenceable(struct i915_vma *vma) { - struct drm_i915_gem_object *obj = vma->obj; - struct drm_i915_private *dev_priv = to_i915(obj->base.dev); bool mappable, fenceable; - u32 fence_size, fence_alignment; - - fence_size = i915_gem_get_ggtt_size(dev_priv, - vma->size, - i915_gem_object_get_tiling(obj)); - fence_alignment = i915_gem_get_ggtt_alignment(dev_priv, - vma->size, - i915_gem_object_get_tiling(obj), - true); - - fenceable = (vma->node.size == fence_size && - (vma->node.start & (fence_alignment - 1)) == 0); - mappable = (vma->node.start + fence_size <= - dev_priv->ggtt.mappable_end); + GEM_BUG_ON(!i915_vma_is_ggtt(vma)); + GEM_BUG_ON(!vma->fence_size); /* * Explicitly disable for rotated VMA since the display does not * need the fence and the VMA is not accessible to other users. */ - if (mappable && fenceable && - vma->ggtt_view.type != I915_GGTT_VIEW_ROTATED) + if (vma->ggtt_view.type == I915_GGTT_VIEW_ROTATED) + return; + + fenceable = (vma->node.size >= vma->fence_size && + IS_ALIGNED(vma->node.start, vma->fence_alignment)); + + mappable = vma->node.start + vma->fence_size <= i915_vm_to_ggtt(vma->vm)->mappable_end; + + if (mappable && fenceable) vma->flags |= I915_VMA_CAN_FENCE; else vma->flags &= ~I915_VMA_CAN_FENCE; @@ -368,22 +437,26 @@ i915_vma_insert(struct i915_vma *vma, u64 size, u64 alignment, u64 flags) GEM_BUG_ON(drm_mm_node_allocated(&vma->node)); size = max(size, vma->size); - if (flags & PIN_MAPPABLE) - size = i915_gem_get_ggtt_size(dev_priv, size, - i915_gem_object_get_tiling(obj)); + alignment = max(alignment, vma->display_alignment); + if (flags & PIN_MAPPABLE) { + size = max_t(typeof(size), size, vma->fence_size); + alignment = max_t(typeof(alignment), + alignment, vma->fence_alignment); + } - alignment = max(max(alignment, vma->display_alignment), - i915_gem_get_ggtt_alignment(dev_priv, size, - i915_gem_object_get_tiling(obj), - flags & PIN_MAPPABLE)); + GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE)); + GEM_BUG_ON(!IS_ALIGNED(alignment, I915_GTT_MIN_ALIGNMENT)); + GEM_BUG_ON(!is_power_of_2(alignment)); start = flags & PIN_OFFSET_BIAS ? flags & PIN_OFFSET_MASK : 0; + GEM_BUG_ON(!IS_ALIGNED(start, I915_GTT_PAGE_SIZE)); end = vma->vm->total; if (flags & PIN_MAPPABLE) end = min_t(u64, end, dev_priv->ggtt.mappable_end); if (flags & PIN_ZONE_4G) - end = min_t(u64, end, (1ULL << 32) - PAGE_SIZE); + end = min_t(u64, end, (1ULL << 32) - I915_GTT_PAGE_SIZE); + GEM_BUG_ON(!IS_ALIGNED(end, I915_GTT_PAGE_SIZE)); /* If binding the object/GGTT view requires more space than the entire * aperture has, reject it early before evicting everything in a vain @@ -403,65 +476,28 @@ i915_vma_insert(struct i915_vma *vma, u64 size, u64 alignment, u64 flags) if (flags & PIN_OFFSET_FIXED) { u64 offset = flags & PIN_OFFSET_MASK; - if (offset & (alignment - 1) || + if (!IS_ALIGNED(offset, alignment) || range_overflows(offset, size, end)) { ret = -EINVAL; goto err_unpin; } - vma->node.start = offset; - vma->node.size = size; - vma->node.color = obj->cache_level; - ret = drm_mm_reserve_node(&vma->vm->mm, &vma->node); - if (ret) { - ret = i915_gem_evict_for_vma(vma, flags); - if (ret == 0) - ret = drm_mm_reserve_node(&vma->vm->mm, &vma->node); - if (ret) - goto err_unpin; - } + ret = i915_gem_gtt_reserve(vma->vm, &vma->node, + size, offset, obj->cache_level, + flags); + if (ret) + goto err_unpin; } else { - u32 search_flag, alloc_flag; - - if (flags & PIN_HIGH) { - search_flag = DRM_MM_SEARCH_BELOW; - alloc_flag = DRM_MM_CREATE_TOP; - } else { - search_flag = DRM_MM_SEARCH_DEFAULT; - alloc_flag = DRM_MM_CREATE_DEFAULT; - } - - /* We only allocate in PAGE_SIZE/GTT_PAGE_SIZE (4096) chunks, - * so we know that we always have a minimum alignment of 4096. - * The drm_mm range manager is optimised to return results - * with zero alignment, so where possible use the optimal - * path. - */ - if (alignment <= 4096) - alignment = 0; - -search_free: - ret = drm_mm_insert_node_in_range_generic(&vma->vm->mm, - &vma->node, - size, alignment, - obj->cache_level, - start, end, - search_flag, - alloc_flag); - if (ret) { - ret = i915_gem_evict_something(vma->vm, size, alignment, - obj->cache_level, - start, end, - flags); - if (ret == 0) - goto search_free; - + ret = i915_gem_gtt_insert(vma->vm, &vma->node, + size, alignment, obj->cache_level, + start, end, flags); + if (ret) goto err_unpin; - } GEM_BUG_ON(vma->node.start < start); GEM_BUG_ON(vma->node.start + vma->node.size > end); } + GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); GEM_BUG_ON(!i915_gem_valid_gtt_space(vma, obj->cache_level)); list_move_tail(&obj->global_link, &dev_priv->mm.bound_list); @@ -504,6 +540,7 @@ int __i915_vma_do_pin(struct i915_vma *vma, if ((bound ^ vma->flags) & I915_VMA_GLOBAL_BIND) __i915_vma_set_map_and_fenceable(vma); + GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); GEM_BUG_ON(i915_vma_misplaced(vma, size, alignment, flags)); return 0; diff --git a/drivers/gpu/drm/i915/i915_vma.h b/drivers/gpu/drm/i915/i915_vma.h index e3b2b3b..e39d922 100644 --- a/drivers/gpu/drm/i915/i915_vma.h +++ b/drivers/gpu/drm/i915/i915_vma.h @@ -55,6 +55,9 @@ struct i915_vma { u64 size; u64 display_alignment; + u32 fence_size; + u32 fence_alignment; + unsigned int flags; /** * How many users have pinned this object in GTT space. The following @@ -109,9 +112,9 @@ struct i915_vma { }; struct i915_vma * -i915_vma_create(struct drm_i915_gem_object *obj, - struct i915_address_space *vm, - const struct i915_ggtt_view *view); +i915_vma_instance(struct drm_i915_gem_object *obj, + struct i915_address_space *vm, + const struct i915_ggtt_view *view); void i915_vma_unpin_and_release(struct i915_vma **p_vma); @@ -196,15 +199,30 @@ i915_vma_compare(struct i915_vma *vma, if (cmp) return cmp; + BUILD_BUG_ON(I915_GGTT_VIEW_NORMAL != 0); + cmp = vma->ggtt_view.type; if (!view) - return vma->ggtt_view.type; + return cmp; - if (vma->ggtt_view.type != view->type) - return vma->ggtt_view.type - view->type; + cmp -= view->type; + if (cmp) + return cmp; - return memcmp(&vma->ggtt_view.params, - &view->params, - sizeof(view->params)); + /* ggtt_view.type also encodes its size so that we both distinguish + * different views using it as a "type" and also use a compact (no + * accessing of uninitialised padding bytes) memcmp without storing + * an extra parameter or adding more code. + * + * To ensure that the memcmp is valid for all branches of the union, + * even though the code looks like it is just comparing one branch, + * we assert above that all branches have the same address, and that + * each branch has a unique type/size. + */ + BUILD_BUG_ON(I915_GGTT_VIEW_NORMAL >= I915_GGTT_VIEW_PARTIAL); + BUILD_BUG_ON(I915_GGTT_VIEW_PARTIAL >= I915_GGTT_VIEW_ROTATED); + BUILD_BUG_ON(offsetof(typeof(*view), rotated) != + offsetof(typeof(*view), partial)); + return memcmp(&vma->ggtt_view.partial, &view->partial, view->type); } int i915_vma_bind(struct i915_vma *vma, enum i915_cache_level cache_level, @@ -229,8 +247,11 @@ i915_vma_pin(struct i915_vma *vma, u64 size, u64 alignment, u64 flags) /* Pin early to prevent the shrinker/eviction logic from destroying * our vma as we insert and bind. */ - if (likely(((++vma->flags ^ flags) & I915_VMA_BIND_MASK) == 0)) + if (likely(((++vma->flags ^ flags) & I915_VMA_BIND_MASK) == 0)) { + GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); + GEM_BUG_ON(i915_vma_misplaced(vma, size, alignment, flags)); return 0; + } return __i915_vma_do_pin(vma, size, alignment, flags); } diff --git a/drivers/gpu/drm/i915/intel_atomic_plane.c b/drivers/gpu/drm/i915/intel_atomic_plane.c index 4612ffd..41fd94e 100644 --- a/drivers/gpu/drm/i915/intel_atomic_plane.c +++ b/drivers/gpu/drm/i915/intel_atomic_plane.c @@ -85,6 +85,8 @@ intel_plane_duplicate_state(struct drm_plane *plane) __drm_atomic_helper_plane_duplicate_state(plane, state); + intel_state->vma = NULL; + return state; } @@ -100,6 +102,24 @@ void intel_plane_destroy_state(struct drm_plane *plane, struct drm_plane_state *state) { + struct i915_vma *vma; + + vma = fetch_and_zero(&to_intel_plane_state(state)->vma); + + /* + * FIXME: Normally intel_cleanup_plane_fb handles destruction of vma. + * We currently don't clear all planes during driver unload, so we have + * to be able to unpin vma here for now. + * + * Normally this can only happen during unload when kmscon is disabled + * and userspace doesn't attempt to set a framebuffer at all. + */ + if (vma) { + mutex_lock(&plane->dev->struct_mutex); + intel_unpin_fb_vma(vma); + mutex_unlock(&plane->dev->struct_mutex); + } + drm_atomic_helper_plane_destroy_state(plane, state); } diff --git a/drivers/gpu/drm/i915/intel_csr.c b/drivers/gpu/drm/i915/intel_csr.c index 9cbb8d8..0085bc7 100644 --- a/drivers/gpu/drm/i915/intel_csr.c +++ b/drivers/gpu/drm/i915/intel_csr.c @@ -34,6 +34,10 @@ * low-power state and comes back to normal. */ +#define I915_CSR_GLK "i915/glk_dmc_ver1_01.bin" +MODULE_FIRMWARE(I915_CSR_GLK); +#define GLK_CSR_VERSION_REQUIRED CSR_VERSION(1, 1) + #define I915_CSR_KBL "i915/kbl_dmc_ver1_01.bin" MODULE_FIRMWARE(I915_CSR_KBL); #define KBL_CSR_VERSION_REQUIRED CSR_VERSION(1, 1) @@ -286,7 +290,9 @@ static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv, csr->version = css_header->version; - if (IS_KABYLAKE(dev_priv)) { + if (IS_GEMINILAKE(dev_priv)) { + required_version = GLK_CSR_VERSION_REQUIRED; + } else if (IS_KABYLAKE(dev_priv)) { required_version = KBL_CSR_VERSION_REQUIRED; } else if (IS_SKYLAKE(dev_priv)) { required_version = SKL_CSR_VERSION_REQUIRED; @@ -435,7 +441,9 @@ void intel_csr_ucode_init(struct drm_i915_private *dev_priv) if (!HAS_CSR(dev_priv)) return; - if (IS_KABYLAKE(dev_priv)) + if (IS_GEMINILAKE(dev_priv)) + csr->fw_path = I915_CSR_GLK; + else if (IS_KABYLAKE(dev_priv)) csr->fw_path = I915_CSR_KBL; else if (IS_SKYLAKE(dev_priv)) csr->fw_path = I915_CSR_SKL; diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c index f642f6d..fcf8181 100644 --- a/drivers/gpu/drm/i915/intel_device_info.c +++ b/drivers/gpu/drm/i915/intel_device_info.c @@ -192,7 +192,7 @@ static void gen9_sseu_info_init(struct drm_i915_private *dev_priv) (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) && hweight8(sseu->slice_mask) > 1; sseu->has_subslice_pg = - IS_BROXTON(dev_priv) && sseu_subslice_total(sseu) > 1; + IS_GEN9_LP(dev_priv) && sseu_subslice_total(sseu) > 1; sseu->has_eu_pg = sseu->eu_per_subslice > 2; if (IS_BROXTON(dev_priv)) { diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index bc35f9b..0f4272f 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -2137,11 +2137,10 @@ intel_fill_fb_ggtt_view(struct i915_ggtt_view *view, const struct drm_framebuffer *fb, unsigned int rotation) { + view->type = I915_GGTT_VIEW_NORMAL; if (drm_rotation_90_or_270(rotation)) { - *view = i915_ggtt_view_rotated; - view->params.rotated = to_intel_framebuffer(fb)->rot_info; - } else { - *view = i915_ggtt_view_normal; + view->type = I915_GGTT_VIEW_ROTATED; + view->rotated = to_intel_framebuffer(fb)->rot_info; } } @@ -2235,24 +2234,19 @@ intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, unsigned int rotation) i915_vma_pin_fence(vma); } + i915_vma_get(vma); err: intel_runtime_pm_put(dev_priv); return vma; } -void intel_unpin_fb_obj(struct drm_framebuffer *fb, unsigned int rotation) +void intel_unpin_fb_vma(struct i915_vma *vma) { - struct drm_i915_gem_object *obj = intel_fb_obj(fb); - struct i915_ggtt_view view; - struct i915_vma *vma; - - WARN_ON(!mutex_is_locked(&obj->base.dev->struct_mutex)); - - intel_fill_fb_ggtt_view(&view, fb, rotation); - vma = i915_gem_object_to_ggtt(obj, &view); + lockdep_assert_held(&vma->vm->i915->drm.struct_mutex); i915_vma_unpin_fence(vma); i915_gem_object_unpin_from_display_plane(vma); + i915_vma_put(vma); } static int intel_fb_pitch(const struct drm_framebuffer *fb, int plane, @@ -2745,7 +2739,6 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, struct drm_device *dev = intel_crtc->base.dev; struct drm_i915_private *dev_priv = to_i915(dev); struct drm_crtc *c; - struct intel_crtc *i; struct drm_i915_gem_object *obj; struct drm_plane *primary = intel_crtc->base.primary; struct drm_plane_state *plane_state = primary->state; @@ -2770,20 +2763,20 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, * an fb with another CRTC instead */ for_each_crtc(dev, c) { - i = to_intel_crtc(c); + struct intel_plane_state *state; if (c == &intel_crtc->base) continue; - if (!i->active) + if (!to_intel_crtc(c)->active) continue; - fb = c->primary->fb; - if (!fb) + state = to_intel_plane_state(c->primary->state); + if (!state->vma) continue; - obj = intel_fb_obj(fb); - if (i915_gem_object_ggtt_offset(obj, NULL) == plane_config->base) { + if (intel_plane_ggtt_offset(state) == plane_config->base) { + fb = c->primary->fb; drm_framebuffer_reference(fb); goto valid_fb; } @@ -2796,7 +2789,7 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, * simplest solution is to just disable the primary plane now and * pretend the BIOS never had it enabled. */ - to_intel_plane_state(plane_state)->base.visible = false; + plane_state->visible = false; crtc_state->plane_mask &= ~(1 << drm_plane_index(primary)); intel_pre_disable_primary_noatomic(&intel_crtc->base); intel_plane->disable_plane(primary, &intel_crtc->base); @@ -2804,6 +2797,19 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, return; valid_fb: + mutex_lock(&dev->struct_mutex); + intel_state->vma = + intel_pin_and_fence_fb_obj(fb, primary->state->rotation); + mutex_unlock(&dev->struct_mutex); + if (IS_ERR(intel_state->vma)) { + DRM_ERROR("failed to pin boot fb on pipe %d: %li\n", + intel_crtc->pipe, PTR_ERR(intel_state->vma)); + + intel_state->vma = NULL; + drm_framebuffer_unreference(fb); + return; + } + plane_state->src_x = 0; plane_state->src_y = 0; plane_state->src_w = fb->width << 16; @@ -3099,13 +3105,13 @@ static void i9xx_update_primary_plane(struct drm_plane *primary, I915_WRITE(DSPSTRIDE(plane), fb->pitches[0]); if (INTEL_GEN(dev_priv) >= 4) { I915_WRITE(DSPSURF(plane), - intel_fb_gtt_offset(fb, rotation) + + intel_plane_ggtt_offset(plane_state) + intel_crtc->dspaddr_offset); I915_WRITE(DSPTILEOFF(plane), (y << 16) | x); I915_WRITE(DSPLINOFF(plane), linear_offset); } else { I915_WRITE(DSPADDR(plane), - intel_fb_gtt_offset(fb, rotation) + + intel_plane_ggtt_offset(plane_state) + intel_crtc->dspaddr_offset); } POSTING_READ(reg); @@ -3202,7 +3208,7 @@ static void ironlake_update_primary_plane(struct drm_plane *primary, I915_WRITE(DSPSTRIDE(plane), fb->pitches[0]); I915_WRITE(DSPSURF(plane), - intel_fb_gtt_offset(fb, rotation) + + intel_plane_ggtt_offset(plane_state) + intel_crtc->dspaddr_offset); if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) { I915_WRITE(DSPOFFSET(plane), (y << 16) | x); @@ -3225,23 +3231,6 @@ u32 intel_fb_stride_alignment(const struct drm_i915_private *dev_priv, } } -u32 intel_fb_gtt_offset(struct drm_framebuffer *fb, - unsigned int rotation) -{ - struct drm_i915_gem_object *obj = intel_fb_obj(fb); - struct i915_ggtt_view view; - struct i915_vma *vma; - - intel_fill_fb_ggtt_view(&view, fb, rotation); - - vma = i915_gem_object_to_ggtt(obj, &view); - if (WARN(!vma, "ggtt vma for display object not found! (view=%u)\n", - view.type)) - return -1; - - return i915_ggtt_offset(vma); -} - static void skl_detach_scaler(struct intel_crtc *intel_crtc, int id) { struct drm_device *dev = intel_crtc->base.dev; @@ -3437,7 +3426,7 @@ static void skylake_update_primary_plane(struct drm_plane *plane, } I915_WRITE(PLANE_SURF(pipe, plane_id), - intel_fb_gtt_offset(fb, rotation) + surf_addr); + intel_plane_ggtt_offset(plane_state) + surf_addr); POSTING_READ(PLANE_SURF(pipe, plane_id)); } @@ -3561,23 +3550,19 @@ void intel_prepare_reset(struct drm_i915_private *dev_priv) state = drm_atomic_helper_duplicate_state(dev, ctx); if (IS_ERR(state)) { ret = PTR_ERR(state); - state = NULL; DRM_ERROR("Duplicating state failed with %i\n", ret); - goto err; + return; } ret = drm_atomic_helper_disable_all(dev, ctx); if (ret) { DRM_ERROR("Suspending crtc's failed with %i\n", ret); - goto err; + drm_atomic_state_put(state); + return; } dev_priv->modeset_restore_state = state; state->acquire_ctx = ctx; - return; - -err: - drm_atomic_state_put(state); } void intel_finish_reset(struct drm_i915_private *dev_priv) @@ -6886,13 +6871,13 @@ static void intel_crtc_disable_noatomic(struct drm_crtc *crtc) if (!intel_crtc->active) return; - if (to_intel_plane_state(crtc->primary->state)->base.visible) { + if (crtc->primary->state->visible) { WARN_ON(intel_crtc->flip_work); intel_pre_disable_primary_noatomic(crtc); intel_crtc_disable_planes(crtc, 1 << drm_plane_index(crtc->primary)); - to_intel_plane_state(crtc->primary->state)->base.visible = false; + crtc->primary->state->visible = false; } state = drm_atomic_state_alloc(crtc->dev); @@ -8406,7 +8391,6 @@ void intel_mode_from_pipe_config(struct drm_display_mode *mode, mode->type = DRM_MODE_TYPE_DRIVER; mode->clock = pipe_config->base.adjusted_mode.crtc_clock; - mode->flags |= pipe_config->base.adjusted_mode.flags; mode->hsync = drm_mode_hsync(mode); mode->vrefresh = drm_mode_vrefresh(mode); @@ -11291,6 +11275,7 @@ found: } old->restore_state = restore_state; + drm_atomic_state_put(state); /* let the connector get through one full cycle before testing */ intel_wait_for_vblank(dev_priv, intel_crtc->pipe); @@ -11570,7 +11555,7 @@ static void intel_unpin_work_fn(struct work_struct *__work) flush_work(&work->mmio_work); mutex_lock(&dev->struct_mutex); - intel_unpin_fb_obj(work->old_fb, primary->state->rotation); + intel_unpin_fb_vma(work->old_vma); i915_gem_object_put(work->pending_flip_obj); mutex_unlock(&dev->struct_mutex); @@ -12280,8 +12265,10 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, goto cleanup_pending; } - work->gtt_offset = intel_fb_gtt_offset(fb, primary->state->rotation); - work->gtt_offset += intel_crtc->dspaddr_offset; + work->old_vma = to_intel_plane_state(primary->state)->vma; + to_intel_plane_state(primary->state)->vma = vma; + + work->gtt_offset = i915_ggtt_offset(vma) + intel_crtc->dspaddr_offset; work->rotation = crtc->primary->state->rotation; /* @@ -12336,7 +12323,8 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, cleanup_request: i915_add_request_no_flush(request); cleanup_unpin: - intel_unpin_fb_obj(fb, crtc->primary->state->rotation); + to_intel_plane_state(primary->state)->vma = work->old_vma; + intel_unpin_fb_vma(vma); cleanup_pending: atomic_dec(&intel_crtc->unpin_work_count); unlock: @@ -12466,7 +12454,7 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state, } was_visible = old_plane_state->base.visible; - visible = to_intel_plane_state(plane_state)->base.visible; + visible = plane_state->visible; if (!was_crtc_enabled && WARN_ON(was_visible)) was_visible = false; @@ -12482,7 +12470,7 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state, * only combine the results from all planes in the current place? */ if (!is_crtc_enabled) - to_intel_plane_state(plane_state)->base.visible = visible = false; + plane_state->visible = visible = false; if (!was_visible && !visible) return 0; @@ -14740,6 +14728,7 @@ static const struct drm_crtc_funcs intel_crtc_funcs = { .page_flip = intel_crtc_page_flip, .atomic_duplicate_state = intel_crtc_duplicate_state, .atomic_destroy_state = intel_crtc_destroy_state, + .set_crc_source = intel_crtc_set_crc_source, }; /** @@ -14836,6 +14825,8 @@ intel_prepare_plane_fb(struct drm_plane *plane, DRM_DEBUG_KMS("failed to pin object\n"); return PTR_ERR(vma); } + + to_intel_plane_state(new_state)->vma = vma; } return 0; @@ -14854,19 +14845,12 @@ void intel_cleanup_plane_fb(struct drm_plane *plane, struct drm_plane_state *old_state) { - struct drm_i915_private *dev_priv = to_i915(plane->dev); - struct intel_plane_state *old_intel_state; - struct drm_i915_gem_object *old_obj = intel_fb_obj(old_state->fb); - struct drm_i915_gem_object *obj = intel_fb_obj(plane->state->fb); - - old_intel_state = to_intel_plane_state(old_state); - - if (!obj && !old_obj) - return; + struct i915_vma *vma; - if (old_obj && (plane->type != DRM_PLANE_TYPE_CURSOR || - !INTEL_INFO(dev_priv)->cursor_needs_physical)) - intel_unpin_fb_obj(old_state->fb, old_state->rotation); + /* Should only be called after a successful intel_prepare_plane_fb()! */ + vma = fetch_and_zero(&to_intel_plane_state(old_state)->vma); + if (vma) + intel_unpin_fb_vma(vma); } int @@ -15018,6 +15002,7 @@ intel_legacy_cursor_update(struct drm_plane *plane, struct intel_plane *intel_plane = to_intel_plane(plane); struct drm_framebuffer *old_fb; struct drm_crtc_state *crtc_state = crtc->state; + struct i915_vma *old_vma; /* * When crtc is inactive or there is a modeset pending, @@ -15089,9 +15074,12 @@ intel_legacy_cursor_update(struct drm_plane *plane, ret = PTR_ERR(vma); goto out_unlock; } + + to_intel_plane_state(new_plane_state)->vma = vma; } old_fb = old_plane_state->fb; + old_vma = to_intel_plane_state(old_plane_state)->vma; i915_gem_track_fb(intel_fb_obj(old_fb), intel_fb_obj(fb), intel_plane->frontbuffer_bit); @@ -15101,6 +15089,7 @@ intel_legacy_cursor_update(struct drm_plane *plane, *to_intel_plane_state(old_plane_state) = *to_intel_plane_state(new_plane_state); new_plane_state->fence = NULL; new_plane_state->fb = old_fb; + to_intel_plane_state(new_plane_state)->vma = old_vma; intel_plane->update_plane(plane, to_intel_crtc_state(crtc->state), @@ -15339,7 +15328,7 @@ intel_update_cursor_plane(struct drm_plane *plane, if (!obj) addr = 0; else if (!INTEL_INFO(dev_priv)->cursor_needs_physical) - addr = i915_gem_object_ggtt_offset(obj, NULL); + addr = intel_plane_ggtt_offset(state); else addr = obj->phys_handle->busaddr; @@ -16843,7 +16832,7 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc) * Temporarily change the plane mapping and disable everything * ... */ plane = crtc->plane; - to_intel_plane_state(crtc->base.primary->state)->base.visible = true; + crtc->base.primary->state->visible = true; crtc->plane = !plane; intel_crtc_disable_noatomic(&crtc->base); crtc->plane = plane; @@ -16995,7 +16984,8 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) dev_priv->active_crtcs = 0; for_each_intel_crtc(dev, crtc) { - struct intel_crtc_state *crtc_state = crtc->config; + struct intel_crtc_state *crtc_state = + to_intel_crtc_state(crtc->base.state); __drm_atomic_helper_crtc_destroy_state(&crtc_state->base); memset(crtc_state, 0, sizeof(*crtc_state)); @@ -17014,7 +17004,7 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) DRM_DEBUG_KMS("[CRTC:%d:%s] hw state readout: %s\n", crtc->base.base.id, crtc->base.name, - enableddisabled(crtc->active)); + enableddisabled(crtc_state->base.active)); } for (i = 0; i < dev_priv->num_shared_dpll; i++) { @@ -17024,7 +17014,11 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) &pll->state.hw_state); pll->state.crtc_mask = 0; for_each_intel_crtc(dev, crtc) { - if (crtc->active && crtc->config->shared_dpll == pll) + struct intel_crtc_state *crtc_state = + to_intel_crtc_state(crtc->base.state); + + if (crtc_state->base.active && + crtc_state->shared_dpll == pll) pll->state.crtc_mask |= 1 << crtc->pipe; } pll->active_mask = pll->state.crtc_mask; @@ -17037,11 +17031,14 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) pipe = 0; if (encoder->get_hw_state(encoder, &pipe)) { + struct intel_crtc_state *crtc_state; + crtc = intel_get_crtc_for_pipe(dev_priv, pipe); + crtc_state = to_intel_crtc_state(crtc->base.state); encoder->base.crtc = &crtc->base; - crtc->config->output_types |= 1 << encoder->type; - encoder->get_config(encoder, crtc->config); + crtc_state->output_types |= 1 << encoder->type; + encoder->get_config(encoder, crtc_state); } else { encoder->base.crtc = NULL; } @@ -17082,14 +17079,16 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) } for_each_intel_crtc(dev, crtc) { + struct intel_crtc_state *crtc_state = + to_intel_crtc_state(crtc->base.state); int pixclk = 0; - crtc->base.hwmode = crtc->config->base.adjusted_mode; + crtc->base.hwmode = crtc_state->base.adjusted_mode; memset(&crtc->base.mode, 0, sizeof(crtc->base.mode)); - if (crtc->base.state->active) { - intel_mode_from_pipe_config(&crtc->base.mode, crtc->config); - intel_mode_from_pipe_config(&crtc->base.state->adjusted_mode, crtc->config); + if (crtc_state->base.active) { + intel_mode_from_pipe_config(&crtc->base.mode, crtc_state); + intel_mode_from_pipe_config(&crtc_state->base.adjusted_mode, crtc_state); WARN_ON(drm_atomic_set_mode_for_crtc(crtc->base.state, &crtc->base.mode)); /* @@ -17101,17 +17100,17 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) * set a flag to indicate that a full recalculation is * needed on the next commit. */ - crtc->base.state->mode.private_flags = I915_MODE_FLAG_INHERITED; + crtc_state->base.mode.private_flags = I915_MODE_FLAG_INHERITED; if (INTEL_GEN(dev_priv) >= 9 || IS_BROADWELL(dev_priv)) - pixclk = ilk_pipe_pixel_rate(crtc->config); + pixclk = ilk_pipe_pixel_rate(crtc_state); else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) - pixclk = crtc->config->base.adjusted_mode.crtc_clock; + pixclk = crtc_state->base.adjusted_mode.crtc_clock; else WARN_ON(dev_priv->display.modeset_calc_cdclk); /* pixel rate mustn't exceed 95% of cdclk with IPS on BDW */ - if (IS_BROADWELL(dev_priv) && crtc->config->ips_enabled) + if (IS_BROADWELL(dev_priv) && crtc_state->ips_enabled) pixclk = DIV_ROUND_UP(pixclk * 100, 95); drm_calc_timestamping_constants(&crtc->base, &crtc->base.hwmode); @@ -17120,7 +17119,7 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) dev_priv->min_pixclk[crtc->pipe] = pixclk; - intel_pipe_config_sanity_check(dev_priv, crtc->config); + intel_pipe_config_sanity_check(dev_priv, crtc_state); } } @@ -17221,47 +17220,19 @@ void intel_display_resume(struct drm_device *dev) if (ret) DRM_ERROR("Restoring old state failed with %i\n", ret); - drm_atomic_state_put(state); + if (state) + drm_atomic_state_put(state); } void intel_modeset_gem_init(struct drm_device *dev) { struct drm_i915_private *dev_priv = to_i915(dev); - struct drm_crtc *c; - struct drm_i915_gem_object *obj; intel_init_gt_powersave(dev_priv); intel_modeset_init_hw(dev); intel_setup_overlay(dev_priv); - - /* - * Make sure any fbs we allocated at startup are properly - * pinned & fenced. When we do the allocation it's too early - * for this. - */ - for_each_crtc(dev, c) { - struct i915_vma *vma; - - obj = intel_fb_obj(c->primary->fb); - if (obj == NULL) - continue; - - mutex_lock(&dev->struct_mutex); - vma = intel_pin_and_fence_fb_obj(c->primary->fb, - c->primary->state->rotation); - mutex_unlock(&dev->struct_mutex); - if (IS_ERR(vma)) { - DRM_ERROR("failed to pin boot fb on pipe %d\n", - to_intel_crtc(c)->pipe); - drm_framebuffer_unreference(c->primary->fb); - c->primary->fb = NULL; - c->primary->crtc = c->primary->state->crtc = NULL; - update_state_fb(c->primary); - c->state->plane_mask &= ~(1 << drm_plane_index(c->primary)); - } - } } int intel_connector_register(struct drm_connector *connector) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index fb12896..e80d620 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -3042,6 +3042,32 @@ intel_dp_get_link_status(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_ DP_LINK_STATUS_SIZE) == DP_LINK_STATUS_SIZE; } +static bool intel_dp_get_y_cord_status(struct intel_dp *intel_dp) +{ + uint8_t psr_caps = 0; + + drm_dp_dpcd_readb(&intel_dp->aux, DP_PSR_CAPS, &psr_caps); + return psr_caps & DP_PSR2_SU_Y_COORDINATE_REQUIRED; +} + +static bool intel_dp_get_colorimetry_status(struct intel_dp *intel_dp) +{ + uint8_t dprx = 0; + + drm_dp_dpcd_readb(&intel_dp->aux, + DP_DPRX_FEATURE_ENUMERATION_LIST, + &dprx); + return dprx & DP_VSC_SDP_EXT_FOR_COLORIMETRY_SUPPORTED; +} + +static bool intel_dp_get_alpm_status(struct intel_dp *intel_dp) +{ + uint8_t alpm_caps = 0; + + drm_dp_dpcd_readb(&intel_dp->aux, DP_RECEIVER_ALPM_CAP, &alpm_caps); + return alpm_caps & DP_ALPM_CAP; +} + /* These are source-specific values. */ uint8_t intel_dp_voltage_max(struct intel_dp *intel_dp) @@ -3412,7 +3438,7 @@ intel_dp_set_signal_levels(struct intel_dp *intel_dp) if (HAS_DDI(dev_priv)) { signal_levels = ddi_signal_levels(intel_dp); - if (IS_BROXTON(dev_priv)) + if (IS_GEN9_LP(dev_priv)) signal_levels = 0; else mask = DDI_BUF_EMP_MASK; @@ -3620,6 +3646,16 @@ intel_edp_init_dpcd(struct intel_dp *intel_dp) dev_priv->psr.psr2_support = dev_priv->psr.aux_frame_sync; 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 = + intel_dp_get_alpm_status(intel_dp); + } + } /* Read the eDP Display control capabilities registers */ diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 6b02dac..0cec001 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -376,6 +376,7 @@ struct intel_atomic_state { struct intel_plane_state { struct drm_plane_state base; struct drm_rect clip; + struct i915_vma *vma; struct { u32 offset; @@ -1067,6 +1068,7 @@ struct intel_flip_work { struct work_struct mmio_work; struct drm_crtc *crtc; + struct i915_vma *old_vma; struct drm_framebuffer *old_fb; struct drm_i915_gem_object *pending_flip_obj; struct drm_pending_vblank_event *event; @@ -1302,7 +1304,7 @@ void intel_release_load_detect_pipe(struct drm_connector *connector, struct drm_modeset_acquire_ctx *ctx); struct i915_vma * intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, unsigned int rotation); -void intel_unpin_fb_obj(struct drm_framebuffer *fb, unsigned int rotation); +void intel_unpin_fb_vma(struct i915_vma *vma); struct drm_framebuffer * __intel_framebuffer_create(struct drm_device *dev, struct drm_mode_fb_cmd2 *mode_cmd, @@ -1391,7 +1393,10 @@ void intel_mode_from_pipe_config(struct drm_display_mode *mode, int skl_update_scaler_crtc(struct intel_crtc_state *crtc_state); int skl_max_scale(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state); -u32 intel_fb_gtt_offset(struct drm_framebuffer *fb, unsigned int rotation); +static inline u32 intel_plane_ggtt_offset(const struct intel_plane_state *state) +{ + return i915_ggtt_offset(state->vma); +} u32 skl_plane_ctl_format(uint32_t pixel_format); u32 skl_plane_ctl_tiling(uint64_t fb_modifier); @@ -1880,5 +1885,11 @@ void lspcon_wait_pcon_mode(struct intel_lspcon *lspcon); /* intel_pipe_crc.c */ int intel_pipe_crc_create(struct drm_minor *minor); void intel_pipe_crc_cleanup(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); +#else +#define intel_crtc_set_crc_source NULL +#endif extern const struct file_operations i915_display_crc_ctl_fops; #endif /* __INTEL_DRV_H__ */ diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 97bbbc3..371acf1 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -264,7 +264,7 @@ int intel_engine_create_scratch(struct intel_engine_cs *engine, int size) return PTR_ERR(obj); } - vma = i915_vma_create(obj, &engine->i915->ggtt.base, NULL); + vma = i915_vma_instance(obj, &engine->i915->ggtt.base, NULL); if (IS_ERR(vma)) { ret = PTR_ERR(vma); goto err_unref; diff --git a/drivers/gpu/drm/i915/intel_fbc.c b/drivers/gpu/drm/i915/intel_fbc.c index 26a81a9..89fe5c8 100644 --- a/drivers/gpu/drm/i915/intel_fbc.c +++ b/drivers/gpu/drm/i915/intel_fbc.c @@ -173,7 +173,7 @@ static void i8xx_fbc_activate(struct drm_i915_private *dev_priv) if (IS_I945GM(dev_priv)) fbc_ctl |= FBC_CTL_C3_IDLE; /* 945 needs special SR handling */ fbc_ctl |= (cfb_pitch & 0xff) << FBC_CTL_STRIDE_SHIFT; - fbc_ctl |= params->fb.fence_reg; + fbc_ctl |= params->vma->fence->id; I915_WRITE(FBC_CONTROL, fbc_ctl); } @@ -193,8 +193,8 @@ static void g4x_fbc_activate(struct drm_i915_private *dev_priv) else dpfc_ctl |= DPFC_CTL_LIMIT_1X; - if (params->fb.fence_reg != I915_FENCE_REG_NONE) { - dpfc_ctl |= DPFC_CTL_FENCE_EN | params->fb.fence_reg; + if (params->vma->fence) { + dpfc_ctl |= DPFC_CTL_FENCE_EN | params->vma->fence->id; I915_WRITE(DPFC_FENCE_YOFF, params->crtc.fence_y_offset); } else { I915_WRITE(DPFC_FENCE_YOFF, 0); @@ -251,13 +251,14 @@ static void ilk_fbc_activate(struct drm_i915_private *dev_priv) break; } - if (params->fb.fence_reg != I915_FENCE_REG_NONE) { + if (params->vma->fence) { dpfc_ctl |= DPFC_CTL_FENCE_EN; if (IS_GEN5(dev_priv)) - dpfc_ctl |= params->fb.fence_reg; + dpfc_ctl |= params->vma->fence->id; if (IS_GEN6(dev_priv)) { I915_WRITE(SNB_DPFC_CTL_SA, - SNB_CPU_FENCE_ENABLE | params->fb.fence_reg); + SNB_CPU_FENCE_ENABLE | + params->vma->fence->id); I915_WRITE(DPFC_CPU_FENCE_OFFSET, params->crtc.fence_y_offset); } @@ -269,7 +270,8 @@ static void ilk_fbc_activate(struct drm_i915_private *dev_priv) } I915_WRITE(ILK_DPFC_FENCE_YOFF, params->crtc.fence_y_offset); - I915_WRITE(ILK_FBC_RT_BASE, params->fb.ggtt_offset | ILK_FBC_RT_VALID); + I915_WRITE(ILK_FBC_RT_BASE, + i915_ggtt_offset(params->vma) | ILK_FBC_RT_VALID); /* enable it... */ I915_WRITE(ILK_DPFC_CONTROL, dpfc_ctl | DPFC_CTL_EN); @@ -319,10 +321,11 @@ static void gen7_fbc_activate(struct drm_i915_private *dev_priv) break; } - if (params->fb.fence_reg != I915_FENCE_REG_NONE) { + if (params->vma->fence) { dpfc_ctl |= IVB_DPFC_CTL_FENCE_EN; I915_WRITE(SNB_DPFC_CTL_SA, - SNB_CPU_FENCE_ENABLE | params->fb.fence_reg); + SNB_CPU_FENCE_ENABLE | + params->vma->fence->id); I915_WRITE(DPFC_CPU_FENCE_OFFSET, params->crtc.fence_y_offset); } else { I915_WRITE(SNB_DPFC_CTL_SA,0); @@ -727,14 +730,6 @@ static bool intel_fbc_hw_tracking_covers_screen(struct intel_crtc *crtc) return effective_w <= max_w && effective_h <= max_h; } -/* XXX replace me when we have VMA tracking for intel_plane_state */ -static int get_fence_id(struct drm_framebuffer *fb) -{ - struct i915_vma *vma = i915_gem_object_to_ggtt(intel_fb_obj(fb), NULL); - - return vma && vma->fence ? vma->fence->id : I915_FENCE_REG_NONE; -} - static void intel_fbc_update_state_cache(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state, struct intel_plane_state *plane_state) @@ -743,7 +738,8 @@ static void intel_fbc_update_state_cache(struct intel_crtc *crtc, struct intel_fbc *fbc = &dev_priv->fbc; struct intel_fbc_state_cache *cache = &fbc->state_cache; struct drm_framebuffer *fb = plane_state->base.fb; - struct drm_i915_gem_object *obj; + + cache->vma = NULL; cache->crtc.mode_flags = crtc_state->base.adjusted_mode.flags; if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) @@ -758,16 +754,10 @@ static void intel_fbc_update_state_cache(struct intel_crtc *crtc, if (!cache->plane.visible) return; - obj = intel_fb_obj(fb); - - /* FIXME: We lack the proper locking here, so only run this on the - * platforms that need. */ - if (IS_GEN(dev_priv, 5, 6)) - cache->fb.ilk_ggtt_offset = i915_gem_object_ggtt_offset(obj, NULL); cache->fb.format = fb->format; cache->fb.stride = fb->pitches[0]; - cache->fb.fence_reg = get_fence_id(fb); - cache->fb.tiling_mode = i915_gem_object_get_tiling(obj); + + cache->vma = plane_state->vma; } static bool intel_fbc_can_activate(struct intel_crtc *crtc) @@ -784,7 +774,7 @@ static bool intel_fbc_can_activate(struct intel_crtc *crtc) return false; } - if (!cache->plane.visible) { + if (!cache->vma) { fbc->no_fbc_reason = "primary plane not visible"; return false; } @@ -807,8 +797,7 @@ static bool intel_fbc_can_activate(struct intel_crtc *crtc) * so have no fence associated with it) due to aperture constaints * at the time of pinning. */ - if (cache->fb.tiling_mode != I915_TILING_X || - cache->fb.fence_reg == I915_FENCE_REG_NONE) { + if (!cache->vma->fence) { fbc->no_fbc_reason = "framebuffer not tiled or fenced"; return false; } @@ -888,17 +877,16 @@ static void intel_fbc_get_reg_params(struct intel_crtc *crtc, * zero. */ memset(params, 0, sizeof(*params)); + params->vma = cache->vma; + params->crtc.pipe = crtc->pipe; params->crtc.plane = crtc->plane; params->crtc.fence_y_offset = get_crtc_fence_y_offset(crtc); params->fb.format = cache->fb.format; params->fb.stride = cache->fb.stride; - params->fb.fence_reg = cache->fb.fence_reg; params->cfb_size = intel_fbc_calculate_cfb_size(dev_priv, cache); - - params->fb.ggtt_offset = cache->fb.ilk_ggtt_offset; } static bool intel_fbc_reg_params_equal(struct intel_fbc_reg_params *params1, @@ -1296,7 +1284,7 @@ void intel_fbc_init_pipe_state(struct drm_i915_private *dev_priv) for_each_intel_crtc(&dev_priv->drm, crtc) if (intel_crtc_active(crtc) && - to_intel_plane_state(crtc->base.primary->state)->base.visible) + crtc->base.primary->state->visible) dev_priv->fbc.visible_pipes_mask |= (1 << crtc->pipe); } diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c index bdefa61..e0d9e72 100644 --- a/drivers/gpu/drm/i915/intel_fbdev.c +++ b/drivers/gpu/drm/i915/intel_fbdev.c @@ -284,7 +284,7 @@ static int intelfb_create(struct drm_fb_helper *helper, out_destroy_fbi: drm_fb_helper_release_fbi(helper); out_unpin: - intel_unpin_fb_obj(&ifbdev->fb->base, DRM_ROTATE_0); + intel_unpin_fb_vma(vma); out_unlock: mutex_unlock(&dev->struct_mutex); return ret; @@ -549,7 +549,7 @@ static void intel_fbdev_destroy(struct intel_fbdev *ifbdev) if (ifbdev->fb) { mutex_lock(&ifbdev->helper.dev->struct_mutex); - intel_unpin_fb_obj(&ifbdev->fb->base, DRM_ROTATE_0); + intel_unpin_fb_vma(ifbdev->vma); mutex_unlock(&ifbdev->helper.dev->struct_mutex); drm_framebuffer_remove(&ifbdev->fb->base); @@ -742,6 +742,9 @@ void intel_fbdev_initial_config_async(struct drm_device *dev) { struct intel_fbdev *ifbdev = to_i915(dev)->fbdev; + if (!ifbdev) + return; + ifbdev->cookie = async_schedule(intel_fbdev_initial_config, ifbdev); } diff --git a/drivers/gpu/drm/i915/intel_guc_fwif.h b/drivers/gpu/drm/i915/intel_guc_fwif.h index 3202b32..25691f0 100644 --- a/drivers/gpu/drm/i915/intel_guc_fwif.h +++ b/drivers/gpu/drm/i915/intel_guc_fwif.h @@ -145,7 +145,7 @@ * The GuC firmware layout looks like this: * * +-------------------------------+ - * | guc_css_header | + * | uc_css_header | * | | * | contains major/minor version | * +-------------------------------+ @@ -172,9 +172,16 @@ * 3. Length info of each component can be found in header, in dwords. * 4. Modulus and exponent key are not required by driver. They may not appear * in fw. So driver will load a truncated firmware in this case. + * + * HuC firmware layout is same as GuC firmware. + * + * HuC firmware css header is different. However, the only difference is where + * the version information is saved. The uc_css_header is unified to support + * both. Driver should get HuC version from uc_css_header.huc_sw_version, while + * uc_css_header.guc_sw_version for GuC. */ -struct guc_css_header { +struct uc_css_header { uint32_t module_type; /* header_size includes all non-uCode bits, including css_header, rsa * key, modulus key and exponent data. */ @@ -205,8 +212,16 @@ struct guc_css_header { char username[8]; char buildnumber[12]; - uint32_t device_id; - uint32_t guc_sw_version; + union { + struct { + uint32_t branch_client_version; + uint32_t sw_version; + } guc; + struct { + uint32_t sw_version; + uint32_t reserved; + } huc; + }; uint32_t prod_preprod_fw; uint32_t reserved[12]; uint32_t header_info; @@ -490,6 +505,7 @@ enum intel_guc_action { INTEL_GUC_ACTION_ENTER_S_STATE = 0x501, INTEL_GUC_ACTION_EXIT_S_STATE = 0x502, INTEL_GUC_ACTION_SLPC_REQUEST = 0x3003, + INTEL_GUC_ACTION_AUTHENTICATE_HUC = 0x4000, INTEL_GUC_ACTION_UK_LOG_ENABLE_LOGGING = 0x0E000, INTEL_GUC_ACTION_LIMIT }; diff --git a/drivers/gpu/drm/i915/intel_guc_loader.c b/drivers/gpu/drm/i915/intel_guc_loader.c index 35d5690..2f1cf9a 100644 --- a/drivers/gpu/drm/i915/intel_guc_loader.c +++ b/drivers/gpu/drm/i915/intel_guc_loader.c @@ -51,12 +51,6 @@ * 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. * - * Firmware log: - * Firmware log is enabled by setting i915.guc_log_level to non-negative level. - * Log data is printed out via reading debugfs i915_guc_log_dump. Reading from - * i915_guc_load_status will print out firmware loading status and scratch - * registers value. - * */ #define SKL_FW_MAJOR 6 @@ -81,16 +75,16 @@ MODULE_FIRMWARE(I915_BXT_GUC_UCODE); MODULE_FIRMWARE(I915_KBL_GUC_UCODE); /* User-friendly representation of an enum */ -const char *intel_guc_fw_status_repr(enum intel_guc_fw_status status) +const char *intel_uc_fw_status_repr(enum intel_uc_fw_status status) { switch (status) { - case GUC_FIRMWARE_FAIL: + case INTEL_UC_FIRMWARE_FAIL: return "FAIL"; - case GUC_FIRMWARE_NONE: + case INTEL_UC_FIRMWARE_NONE: return "NONE"; - case GUC_FIRMWARE_PENDING: + case INTEL_UC_FIRMWARE_PENDING: return "PENDING"; - case GUC_FIRMWARE_SUCCESS: + case INTEL_UC_FIRMWARE_SUCCESS: return "SUCCESS"; default: return "UNKNOWN!"; @@ -278,7 +272,7 @@ static inline bool guc_ucode_response(struct drm_i915_private *dev_priv, static int guc_ucode_xfer_dma(struct drm_i915_private *dev_priv, struct i915_vma *vma) { - struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; + struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; unsigned long offset; struct sg_table *sg = vma->pages; u32 status, rsa[UOS_RSA_SCRATCH_MAX_COUNT]; @@ -334,12 +328,12 @@ static int guc_ucode_xfer_dma(struct drm_i915_private *dev_priv, return ret; } -static u32 guc_wopcm_size(struct drm_i915_private *dev_priv) +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_BROXTON(dev_priv)) + if (IS_GEN9_LP(dev_priv)) wopcm_size -= BXT_GUC_WOPCM_RC6_RESERVED; return wopcm_size; @@ -350,29 +344,27 @@ static u32 guc_wopcm_size(struct drm_i915_private *dev_priv) */ static int guc_ucode_xfer(struct drm_i915_private *dev_priv) { - struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; + struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; struct i915_vma *vma; int ret; - ret = i915_gem_object_set_to_gtt_domain(guc_fw->guc_fw_obj, false); + ret = i915_gem_object_set_to_gtt_domain(guc_fw->obj, false); if (ret) { DRM_DEBUG_DRIVER("set-domain failed %d\n", ret); return ret; } - vma = i915_gem_object_ggtt_pin(guc_fw->guc_fw_obj, NULL, 0, 0, 0); + vma = i915_gem_object_ggtt_pin(guc_fw->obj, NULL, 0, 0, + PIN_OFFSET_BIAS | GUC_WOPCM_TOP); if (IS_ERR(vma)) { DRM_DEBUG_DRIVER("pin failed %d\n", (int)PTR_ERR(vma)); return PTR_ERR(vma); } - /* Invalidate GuC TLB to let GuC take the latest updates to GTT. */ - I915_WRITE(GEN8_GTCR, GEN8_GTCR_INVALIDATE); - intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); /* init WOPCM */ - I915_WRITE(GUC_WOPCM_SIZE, guc_wopcm_size(dev_priv)); + I915_WRITE(GUC_WOPCM_SIZE, intel_guc_wopcm_size(dev_priv)); I915_WRITE(DMA_GUC_WOPCM_OFFSET, GUC_WOPCM_OFFSET_VALUE); /* Enable MIA caching. GuC clock gating is disabled. */ @@ -388,7 +380,7 @@ static int guc_ucode_xfer(struct drm_i915_private *dev_priv) if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_B0)) I915_WRITE(GEN6_GFXPAUSE, 0x30FFF); - if (IS_BROXTON(dev_priv)) + if (IS_GEN9_LP(dev_priv)) I915_WRITE(GEN9LP_GT_PM_CONFIG, GT_DOORBELL_ENABLE); else I915_WRITE(GEN9_GT_PM_CONFIG, GT_DOORBELL_ENABLE); @@ -450,14 +442,14 @@ static int guc_hw_reset(struct drm_i915_private *dev_priv) */ int intel_guc_setup(struct drm_i915_private *dev_priv) { - struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; - const char *fw_path = guc_fw->guc_fw_path; + struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; + const char *fw_path = guc_fw->path; int retries, ret, err; DRM_DEBUG_DRIVER("GuC fw status: path %s, fetch %s, load %s\n", fw_path, - intel_guc_fw_status_repr(guc_fw->guc_fw_fetch_status), - intel_guc_fw_status_repr(guc_fw->guc_fw_load_status)); + intel_uc_fw_status_repr(guc_fw->fetch_status), + intel_uc_fw_status_repr(guc_fw->load_status)); /* Loading forbidden, or no firmware to load? */ if (!i915.enable_guc_loading) { @@ -475,10 +467,10 @@ int intel_guc_setup(struct drm_i915_private *dev_priv) } /* Fetch failed, or already fetched but failed to load? */ - if (guc_fw->guc_fw_fetch_status != GUC_FIRMWARE_SUCCESS) { + if (guc_fw->fetch_status != INTEL_UC_FIRMWARE_SUCCESS) { err = -EIO; goto fail; - } else if (guc_fw->guc_fw_load_status == GUC_FIRMWARE_FAIL) { + } else if (guc_fw->load_status == INTEL_UC_FIRMWARE_FAIL) { err = -ENOEXEC; goto fail; } @@ -486,11 +478,14 @@ int intel_guc_setup(struct drm_i915_private *dev_priv) guc_interrupts_release(dev_priv); gen9_reset_guc_interrupts(dev_priv); - guc_fw->guc_fw_load_status = GUC_FIRMWARE_PENDING; + /* We need to notify the guc whenever we change the GGTT */ + i915_ggtt_enable_guc(dev_priv); + + guc_fw->load_status = INTEL_UC_FIRMWARE_PENDING; DRM_DEBUG_DRIVER("GuC fw status: fetch %s, load %s\n", - intel_guc_fw_status_repr(guc_fw->guc_fw_fetch_status), - intel_guc_fw_status_repr(guc_fw->guc_fw_load_status)); + intel_uc_fw_status_repr(guc_fw->fetch_status), + intel_uc_fw_status_repr(guc_fw->load_status)); err = i915_guc_submission_init(dev_priv); if (err) @@ -511,6 +506,7 @@ int intel_guc_setup(struct drm_i915_private *dev_priv) if (err) goto fail; + intel_huc_load(dev_priv); err = guc_ucode_xfer(dev_priv); if (!err) break; @@ -522,11 +518,13 @@ int intel_guc_setup(struct drm_i915_private *dev_priv) "retry %d more time(s)\n", err, retries); } - guc_fw->guc_fw_load_status = GUC_FIRMWARE_SUCCESS; + guc_fw->load_status = INTEL_UC_FIRMWARE_SUCCESS; DRM_DEBUG_DRIVER("GuC fw status: fetch %s, load %s\n", - intel_guc_fw_status_repr(guc_fw->guc_fw_fetch_status), - intel_guc_fw_status_repr(guc_fw->guc_fw_load_status)); + intel_uc_fw_status_repr(guc_fw->fetch_status), + intel_uc_fw_status_repr(guc_fw->load_status)); + + intel_guc_auth_huc(dev_priv); if (i915.enable_guc_submission) { if (i915.guc_log_level >= 0) @@ -541,12 +539,13 @@ int intel_guc_setup(struct drm_i915_private *dev_priv) return 0; fail: - if (guc_fw->guc_fw_load_status == GUC_FIRMWARE_PENDING) - guc_fw->guc_fw_load_status = GUC_FIRMWARE_FAIL; + if (guc_fw->load_status == INTEL_UC_FIRMWARE_PENDING) + guc_fw->load_status = INTEL_UC_FIRMWARE_FAIL; guc_interrupts_release(dev_priv); i915_guc_submission_disable(dev_priv); i915_guc_submission_fini(dev_priv); + i915_ggtt_disable_guc(dev_priv); /* * We've failed to load the firmware :( @@ -587,93 +586,108 @@ fail: return ret; } -static void guc_fw_fetch(struct drm_i915_private *dev_priv, - struct intel_guc_fw *guc_fw) +void intel_uc_fw_fetch(struct drm_i915_private *dev_priv, + struct intel_uc_fw *uc_fw) { struct pci_dev *pdev = dev_priv->drm.pdev; struct drm_i915_gem_object *obj; const struct firmware *fw = NULL; - struct guc_css_header *css; + struct uc_css_header *css; size_t size; int err; - DRM_DEBUG_DRIVER("before requesting firmware: GuC fw fetch status %s\n", - intel_guc_fw_status_repr(guc_fw->guc_fw_fetch_status)); + DRM_DEBUG_DRIVER("before requesting firmware: uC fw fetch status %s\n", + intel_uc_fw_status_repr(uc_fw->fetch_status)); - err = request_firmware(&fw, guc_fw->guc_fw_path, &pdev->dev); + err = request_firmware(&fw, uc_fw->path, &pdev->dev); if (err) goto fail; if (!fw) goto fail; - DRM_DEBUG_DRIVER("fetch GuC fw from %s succeeded, fw %p\n", - guc_fw->guc_fw_path, fw); + DRM_DEBUG_DRIVER("fetch uC fw from %s succeeded, fw %p\n", + uc_fw->path, fw); /* Check the size of the blob before examining buffer contents */ - if (fw->size < sizeof(struct guc_css_header)) { + if (fw->size < sizeof(struct uc_css_header)) { DRM_NOTE("Firmware header is missing\n"); goto fail; } - css = (struct guc_css_header *)fw->data; + css = (struct uc_css_header *)fw->data; /* Firmware bits always start from header */ - guc_fw->header_offset = 0; - guc_fw->header_size = (css->header_size_dw - css->modulus_size_dw - + uc_fw->header_offset = 0; + uc_fw->header_size = (css->header_size_dw - css->modulus_size_dw - css->key_size_dw - css->exponent_size_dw) * sizeof(u32); - if (guc_fw->header_size != sizeof(struct guc_css_header)) { + if (uc_fw->header_size != sizeof(struct uc_css_header)) { DRM_NOTE("CSS header definition mismatch\n"); goto fail; } /* then, uCode */ - guc_fw->ucode_offset = guc_fw->header_offset + guc_fw->header_size; - guc_fw->ucode_size = (css->size_dw - css->header_size_dw) * sizeof(u32); + 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); /* now RSA */ if (css->key_size_dw != UOS_RSA_SCRATCH_MAX_COUNT) { DRM_NOTE("RSA key size is bad\n"); goto fail; } - guc_fw->rsa_offset = guc_fw->ucode_offset + guc_fw->ucode_size; - guc_fw->rsa_size = css->key_size_dw * sizeof(u32); + uc_fw->rsa_offset = uc_fw->ucode_offset + uc_fw->ucode_size; + uc_fw->rsa_size = css->key_size_dw * sizeof(u32); /* At least, it should have header, uCode and RSA. Size of all three. */ - size = guc_fw->header_size + guc_fw->ucode_size + guc_fw->rsa_size; + size = uc_fw->header_size + uc_fw->ucode_size + uc_fw->rsa_size; if (fw->size < size) { DRM_NOTE("Missing firmware components\n"); goto fail; } - /* Header and uCode will be loaded to WOPCM. Size of the two. */ - size = guc_fw->header_size + guc_fw->ucode_size; - if (size > guc_wopcm_size(dev_priv)) { - DRM_NOTE("Firmware is too large to fit in WOPCM\n"); - goto fail; - } - /* * The GuC firmware image has the version number embedded at a well-known * offset within the firmware blob; note that major / minor version are * TWO bytes each (i.e. u16), although all pointers and offsets are defined * in terms of bytes (u8). */ - guc_fw->guc_fw_major_found = css->guc_sw_version >> 16; - guc_fw->guc_fw_minor_found = css->guc_sw_version & 0xFFFF; - - if (guc_fw->guc_fw_major_found != guc_fw->guc_fw_major_wanted || - guc_fw->guc_fw_minor_found < guc_fw->guc_fw_minor_wanted) { - DRM_NOTE("GuC firmware version %d.%d, required %d.%d\n", - guc_fw->guc_fw_major_found, guc_fw->guc_fw_minor_found, - guc_fw->guc_fw_major_wanted, guc_fw->guc_fw_minor_wanted); + switch (uc_fw->fw) { + case INTEL_UC_FW_TYPE_GUC: + /* Header and uCode will be loaded to WOPCM. Size of the two. */ + size = uc_fw->header_size + uc_fw->ucode_size; + + /* Top 32k of WOPCM is reserved (8K stack + 24k RC6 context). */ + if (size > intel_guc_wopcm_size(dev_priv)) { + DRM_ERROR("Firmware is too large to fit in WOPCM\n"); + goto fail; + } + uc_fw->major_ver_found = css->guc.sw_version >> 16; + uc_fw->minor_ver_found = css->guc.sw_version & 0xFFFF; + break; + + case INTEL_UC_FW_TYPE_HUC: + uc_fw->major_ver_found = css->huc.sw_version >> 16; + uc_fw->minor_ver_found = css->huc.sw_version & 0xFFFF; + break; + + default: + DRM_ERROR("Unknown firmware type %d\n", uc_fw->fw); + err = -ENOEXEC; + goto fail; + } + + if (uc_fw->major_ver_found != uc_fw->major_ver_wanted || + uc_fw->minor_ver_found < uc_fw->minor_ver_wanted) { + DRM_NOTE("uC firmware version %d.%d, required %d.%d\n", + uc_fw->major_ver_found, uc_fw->minor_ver_found, + uc_fw->major_ver_wanted, uc_fw->minor_ver_wanted); err = -ENOEXEC; goto fail; } DRM_DEBUG_DRIVER("firmware version %d.%d OK (minimum %d.%d)\n", - guc_fw->guc_fw_major_found, guc_fw->guc_fw_minor_found, - guc_fw->guc_fw_major_wanted, guc_fw->guc_fw_minor_wanted); + uc_fw->major_ver_found, uc_fw->minor_ver_found, + uc_fw->major_ver_wanted, uc_fw->minor_ver_wanted); mutex_lock(&dev_priv->drm.struct_mutex); obj = i915_gem_object_create_from_data(dev_priv, fw->data, fw->size); @@ -683,31 +697,31 @@ static void guc_fw_fetch(struct drm_i915_private *dev_priv, goto fail; } - guc_fw->guc_fw_obj = obj; - guc_fw->guc_fw_size = fw->size; + uc_fw->obj = obj; + uc_fw->size = fw->size; - DRM_DEBUG_DRIVER("GuC fw fetch status SUCCESS, obj %p\n", - guc_fw->guc_fw_obj); + DRM_DEBUG_DRIVER("uC fw fetch status SUCCESS, obj %p\n", + uc_fw->obj); release_firmware(fw); - guc_fw->guc_fw_fetch_status = GUC_FIRMWARE_SUCCESS; + uc_fw->fetch_status = INTEL_UC_FIRMWARE_SUCCESS; return; fail: - DRM_WARN("Failed to fetch valid GuC firmware from %s (error %d)\n", - guc_fw->guc_fw_path, err); - DRM_DEBUG_DRIVER("GuC fw fetch status FAIL; err %d, fw %p, obj %p\n", - err, fw, guc_fw->guc_fw_obj); + DRM_WARN("Failed to fetch valid uC firmware from %s (error %d)\n", + uc_fw->path, err); + DRM_DEBUG_DRIVER("uC fw fetch status FAIL; err %d, fw %p, obj %p\n", + err, fw, uc_fw->obj); mutex_lock(&dev_priv->drm.struct_mutex); - obj = guc_fw->guc_fw_obj; + obj = uc_fw->obj; if (obj) i915_gem_object_put(obj); - guc_fw->guc_fw_obj = NULL; + uc_fw->obj = NULL; mutex_unlock(&dev_priv->drm.struct_mutex); release_firmware(fw); /* OK even if fw is NULL */ - guc_fw->guc_fw_fetch_status = GUC_FIRMWARE_FAIL; + uc_fw->fetch_status = INTEL_UC_FIRMWARE_FAIL; } /** @@ -721,7 +735,7 @@ fail: */ void intel_guc_init(struct drm_i915_private *dev_priv) { - struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; + struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; const char *fw_path; if (!HAS_GUC(dev_priv)) { @@ -739,23 +753,23 @@ void intel_guc_init(struct drm_i915_private *dev_priv) fw_path = NULL; } else if (IS_SKYLAKE(dev_priv)) { fw_path = I915_SKL_GUC_UCODE; - guc_fw->guc_fw_major_wanted = SKL_FW_MAJOR; - guc_fw->guc_fw_minor_wanted = SKL_FW_MINOR; + guc_fw->major_ver_wanted = SKL_FW_MAJOR; + guc_fw->minor_ver_wanted = SKL_FW_MINOR; } else if (IS_BROXTON(dev_priv)) { fw_path = I915_BXT_GUC_UCODE; - guc_fw->guc_fw_major_wanted = BXT_FW_MAJOR; - guc_fw->guc_fw_minor_wanted = BXT_FW_MINOR; + guc_fw->major_ver_wanted = BXT_FW_MAJOR; + guc_fw->minor_ver_wanted = BXT_FW_MINOR; } else if (IS_KABYLAKE(dev_priv)) { fw_path = I915_KBL_GUC_UCODE; - guc_fw->guc_fw_major_wanted = KBL_FW_MAJOR; - guc_fw->guc_fw_minor_wanted = KBL_FW_MINOR; + guc_fw->major_ver_wanted = KBL_FW_MAJOR; + guc_fw->minor_ver_wanted = KBL_FW_MINOR; } else { fw_path = ""; /* unknown device */ } - guc_fw->guc_fw_path = fw_path; - guc_fw->guc_fw_fetch_status = GUC_FIRMWARE_NONE; - guc_fw->guc_fw_load_status = GUC_FIRMWARE_NONE; + guc_fw->path = fw_path; + guc_fw->fetch_status = INTEL_UC_FIRMWARE_NONE; + guc_fw->load_status = INTEL_UC_FIRMWARE_NONE; /* Early (and silent) return if GuC loading is disabled */ if (!i915.enable_guc_loading) @@ -765,9 +779,9 @@ void intel_guc_init(struct drm_i915_private *dev_priv) if (*fw_path == '\0') return; - guc_fw->guc_fw_fetch_status = GUC_FIRMWARE_PENDING; + guc_fw->fetch_status = INTEL_UC_FIRMWARE_PENDING; DRM_DEBUG_DRIVER("GuC firmware pending, path %s\n", fw_path); - guc_fw_fetch(dev_priv, guc_fw); + intel_uc_fw_fetch(dev_priv, guc_fw); /* status must now be FAIL or SUCCESS */ } @@ -777,17 +791,17 @@ void intel_guc_init(struct drm_i915_private *dev_priv) */ void intel_guc_fini(struct drm_i915_private *dev_priv) { - struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; + struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; mutex_lock(&dev_priv->drm.struct_mutex); guc_interrupts_release(dev_priv); i915_guc_submission_disable(dev_priv); i915_guc_submission_fini(dev_priv); - if (guc_fw->guc_fw_obj) - i915_gem_object_put(guc_fw->guc_fw_obj); - guc_fw->guc_fw_obj = NULL; + if (guc_fw->obj) + i915_gem_object_put(guc_fw->obj); + guc_fw->obj = NULL; mutex_unlock(&dev_priv->drm.struct_mutex); - guc_fw->guc_fw_fetch_status = GUC_FIRMWARE_NONE; + guc_fw->fetch_status = INTEL_UC_FIRMWARE_NONE; } diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c new file mode 100644 index 0000000..5c0f9a4 --- /dev/null +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -0,0 +1,658 @@ +/* + * Copyright © 2014-2017 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. + * + */ +#include <linux/debugfs.h> +#include <linux/relay.h> +#include "i915_drv.h" + +static void guc_log_capture_logs(struct intel_guc *guc); + +/** + * DOC: GuC firmware log + * + * Firmware log is enabled by setting i915.guc_log_level to non-negative level. + * Log data is printed out via reading debugfs i915_guc_log_dump. Reading from + * i915_guc_load_status will print out firmware loading status and scratch + * registers value. + * + */ + +static int guc_log_flush_complete(struct intel_guc *guc) +{ + u32 action[] = { + INTEL_GUC_ACTION_LOG_BUFFER_FILE_FLUSH_COMPLETE + }; + + return intel_guc_send(guc, action, ARRAY_SIZE(action)); +} + +static int guc_log_flush(struct intel_guc *guc) +{ + u32 action[] = { + INTEL_GUC_ACTION_FORCE_LOG_BUFFER_FLUSH, + 0 + }; + + return intel_guc_send(guc, action, ARRAY_SIZE(action)); +} + +static int guc_log_control(struct intel_guc *guc, u32 control_val) +{ + u32 action[] = { + INTEL_GUC_ACTION_UK_LOG_ENABLE_LOGGING, + control_val + }; + + return intel_guc_send(guc, action, ARRAY_SIZE(action)); +} + + +/* + * 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. + */ +static int subbuf_start_callback(struct rchan_buf *buf, + void *subbuf, + void *prev_subbuf, + size_t prev_padding) +{ + /* Use no-overwrite mode by default, where relay will stop accepting + * new data if there are no empty sub buffers left. + * There is no strict synchronization enforced by relay between Consumer + * and Producer. In overwrite mode, there is a possibility of getting + * inconsistent/garbled data, the producer could be writing on to the + * same sub buffer from which Consumer is reading. This can't be avoided + * unless Consumer is fast enough and can always run in tandem with + * Producer. + */ + if (relay_buf_full(buf)) + return 0; + + return 1; +} + +/* + * file_create() callback. Creates relay file in debugfs. + */ +static struct dentry *create_buf_file_callback(const char *filename, + struct dentry *parent, + umode_t mode, + struct rchan_buf *buf, + int *is_global) +{ + struct dentry *buf_file; + + /* This to enable the use of a single buffer for the relay channel and + * correspondingly have a single file exposed to User, through which + * it can collect the logs in order without any post-processing. + * Need to set 'is_global' even if parent is NULL for early logging. + */ + *is_global = 1; + + 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, + parent, buf, &relay_file_operations); + return buf_file; +} + +/* + * file_remove() default callback. Removes relay file in debugfs. + */ +static int remove_buf_file_callback(struct dentry *dentry) +{ + debugfs_remove(dentry); + return 0; +} + +/* relay channel callbacks */ +static struct rchan_callbacks relay_callbacks = { + .subbuf_start = subbuf_start_callback, + .create_buf_file = create_buf_file_callback, + .remove_buf_file = remove_buf_file_callback, +}; + +static void guc_log_remove_relay_file(struct intel_guc *guc) +{ + relay_close(guc->log.relay_chan); +} + +static int guc_log_create_relay_channel(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + struct rchan *guc_log_relay_chan; + size_t n_subbufs, subbuf_size; + + /* Keep the size of sub buffers same as shared log buffer */ + subbuf_size = guc->log.vma->obj->base.size; + + /* Store up to 8 snapshots, which is large enough to buffer sufficient + * boot time logs and provides enough leeway to User, in terms of + * latency, for consuming the logs from relay. Also doesn't take + * up too much memory. + */ + n_subbufs = 8; + + guc_log_relay_chan = relay_open(NULL, NULL, subbuf_size, + n_subbufs, &relay_callbacks, dev_priv); + if (!guc_log_relay_chan) { + DRM_ERROR("Couldn't create relay chan for GuC logging\n"); + return -ENOMEM; + } + + GEM_BUG_ON(guc_log_relay_chan->subbuf_size < subbuf_size); + guc->log.relay_chan = guc_log_relay_chan; + return 0; +} + +static int guc_log_create_relay_file(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + struct dentry *log_dir; + int ret; + + /* 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(guc->log.relay_chan, "guc_log", log_dir); + if (ret) { + 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 *guc) +{ + /* Make sure the updates made in the sub buffer are visible when + * Consumer sees the following update to offset inside the sub buffer. + */ + smp_wmb(); + + /* All data has been written, so now move the offset of sub buffer. */ + relay_reserve(guc->log.relay_chan, guc->log.vma->obj->base.size); + + /* Switch to the next sub buffer */ + relay_flush(guc->log.relay_chan); +} + +static void *guc_get_write_buffer(struct intel_guc *guc) +{ + if (!guc->log.relay_chan) + 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 + * buffers are full. Could have used the relay_write() to indirectly + * copy the data, but that would have been bit convoluted, as we need to + * write to only certain locations inside a sub buffer which cannot be + * done without using relay_reserve() along with relay_write(). So its + * better to use relay_reserve() alone. + */ + return relay_reserve(guc->log.relay_chan, 0); +} + +static bool guc_check_log_buf_overflow(struct intel_guc *guc, + enum guc_log_buffer_type type, + unsigned int full_cnt) +{ + unsigned int prev_full_cnt = guc->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; + + if (full_cnt < prev_full_cnt) { + /* buffer_full_cnt is a 4 bit counter */ + guc->log.total_overflow_count[type] += 16; + } + DRM_ERROR_RATELIMITED("GuC log buffer overflow\n"); + } + + return overflow; +} + +static unsigned int guc_get_log_buffer_size(enum guc_log_buffer_type type) +{ + switch (type) { + case GUC_ISR_LOG_BUFFER: + return (GUC_LOG_ISR_PAGES + 1) * PAGE_SIZE; + case GUC_DPC_LOG_BUFFER: + return (GUC_LOG_DPC_PAGES + 1) * PAGE_SIZE; + case GUC_CRASH_DUMP_LOG_BUFFER: + return (GUC_LOG_CRASH_PAGES + 1) * PAGE_SIZE; + default: + MISSING_CASE(type); + } + + return 0; +} + +static void guc_read_update_log_buffer(struct intel_guc *guc) +{ + 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; + struct guc_log_buffer_state log_buf_state_local; + enum guc_log_buffer_type type; + void *src_data, *dst_data; + bool new_overflow; + + if (WARN_ON(!guc->log.buf_addr)) + return; + + /* Get the pointer to shared GuC log buffer */ + log_buf_state = src_data = guc->log.buf_addr; + + /* Get the pointer to local buffer to store the logs */ + log_buf_snapshot_state = dst_data = guc_get_write_buffer(guc); + + /* Actual logs are present from the 2nd page */ + src_data += PAGE_SIZE; + dst_data += PAGE_SIZE; + + for (type = GUC_ISR_LOG_BUFFER; type < GUC_MAX_LOG_BUFFER; type++) { + /* Make a copy of the state structure, inside GuC log buffer + * (which is uncached mapped), on the stack to avoid reading + * from it multiple times. + */ + memcpy(&log_buf_state_local, log_buf_state, + sizeof(struct guc_log_buffer_state)); + buffer_size = guc_get_log_buffer_size(type); + read_offset = log_buf_state_local.read_ptr; + write_offset = log_buf_state_local.sampled_write_ptr; + 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); + + /* Update the state of shared log buffer */ + log_buf_state->read_ptr = write_offset; + log_buf_state->flush_to_file = 0; + log_buf_state++; + + if (unlikely(!log_buf_snapshot_state)) + continue; + + /* First copy the state structure in snapshot buffer */ + memcpy(log_buf_snapshot_state, &log_buf_state_local, + sizeof(struct guc_log_buffer_state)); + + /* The write pointer could have been updated by GuC firmware, + * after sending the flush interrupt to Host, for consistency + * set write pointer value to same value of sampled_write_ptr + * in the snapshot buffer. + */ + log_buf_snapshot_state->write_ptr = write_offset; + log_buf_snapshot_state++; + + /* Now copy the actual logs. */ + if (unlikely(new_overflow)) { + /* copy the whole buffer in case of overflow */ + read_offset = 0; + write_offset = buffer_size; + } else if (unlikely((read_offset > buffer_size) || + (write_offset > buffer_size))) { + DRM_ERROR("invalid log buffer state\n"); + /* copy whole buffer as offsets are unreliable */ + read_offset = 0; + write_offset = buffer_size; + } + + /* Just copy the newly written data */ + if (read_offset > write_offset) { + i915_memcpy_from_wc(dst_data, src_data, write_offset); + bytes_to_copy = buffer_size - read_offset; + } else { + bytes_to_copy = write_offset - read_offset; + } + i915_memcpy_from_wc(dst_data + read_offset, + src_data + read_offset, bytes_to_copy); + + src_data += buffer_size; + dst_data += buffer_size; + } + + if (log_buf_snapshot_state) + guc_move_to_next_buf(guc); + else { + /* Used rate limited to avoid deluge of messages, logs might be + * getting consumed by User at a slow rate. + */ + DRM_ERROR_RATELIMITED("no sub-buffer to capture logs\n"); + guc->log.capture_miss_count++; + } +} + +static void guc_log_cleanup(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + + lockdep_assert_held(&dev_priv->drm.struct_mutex); + + /* First disable the flush interrupt */ + gen9_disable_guc_interrupts(dev_priv); + + if (guc->log.flush_wq) + destroy_workqueue(guc->log.flush_wq); + + guc->log.flush_wq = NULL; + + if (guc->log.relay_chan) + guc_log_remove_relay_file(guc); + + guc->log.relay_chan = NULL; + + if (guc->log.buf_addr) + i915_gem_object_unpin_map(guc->log.vma->obj); + + guc->log.buf_addr = NULL; +} + +static void capture_logs_work(struct work_struct *work) +{ + struct intel_guc *guc = + container_of(work, struct intel_guc, log.flush_work); + + guc_log_capture_logs(guc); +} + +static int guc_log_create_extras(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + void *vaddr; + int ret; + + lockdep_assert_held(&dev_priv->drm.struct_mutex); + + /* Nothing to do */ + if (i915.guc_log_level < 0) + return 0; + + if (!guc->log.buf_addr) { + /* Create a WC (Uncached for read) vmalloc mapping of log + * 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); + if (IS_ERR(vaddr)) { + ret = PTR_ERR(vaddr); + DRM_ERROR("Couldn't map log buffer pages %d\n", ret); + return ret; + } + + guc->log.buf_addr = vaddr; + } + + if (!guc->log.relay_chan) { + /* 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. + */ + ret = guc_log_create_relay_channel(guc); + if (ret) + return ret; + } + + if (!guc->log.flush_wq) { + INIT_WORK(&guc->log.flush_work, capture_logs_work); + + /* + * GuC log buffer flush work item has to do register access to + * send the ack to GuC and this work item, if not synced before + * suspend, can potentially get executed after the GFX device is + * suspended. + * By marking the WQ as freezable, we don't have to bother about + * flushing of this work item from the suspend hooks, the pending + * work item if any will be either executed before the suspend + * or scheduled later on resume. This way the handling of work + * item can be kept same between system suspend & rpm suspend. + */ + guc->log.flush_wq = alloc_ordered_workqueue("i915-guc_log", + WQ_HIGHPRI | WQ_FREEZABLE); + if (guc->log.flush_wq == NULL) { + DRM_ERROR("Couldn't allocate the wq for GuC logging\n"); + return -ENOMEM; + } + } + + return 0; +} + +void intel_guc_log_create(struct intel_guc *guc) +{ + struct i915_vma *vma; + unsigned long offset; + uint32_t size, flags; + + if (i915.guc_log_level > GUC_LOG_VERBOSITY_MAX) + i915.guc_log_level = GUC_LOG_VERBOSITY_MAX; + + /* The first page is to save log buffer state. Allocate one + * extra page for others in case for overlap */ + size = (1 + GUC_LOG_DPC_PAGES + 1 + + GUC_LOG_ISR_PAGES + 1 + + GUC_LOG_CRASH_PAGES + 1) << PAGE_SHIFT; + + vma = guc->log.vma; + if (!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())) { + /* logging will not be enabled */ + i915.guc_log_level = -1; + return; + } + + vma = intel_guc_allocate_vma(guc, size); + if (IS_ERR(vma)) { + /* logging will be off */ + i915.guc_log_level = -1; + return; + } + + guc->log.vma = vma; + + if (guc_log_create_extras(guc)) { + guc_log_cleanup(guc); + i915_vma_unpin_and_release(&guc->log.vma); + i915.guc_log_level = -1; + return; + } + } + + /* each allocated unit is a page */ + flags = GUC_LOG_VALID | GUC_LOG_NOTIFY_ON_HALF_FULL | + (GUC_LOG_DPC_PAGES << GUC_LOG_DPC_SHIFT) | + (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 */ + guc->log.flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags; +} + +static int guc_log_late_setup(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + int ret; + + lockdep_assert_held(&dev_priv->drm.struct_mutex); + + if (i915.guc_log_level < 0) + return -EINVAL; + + /* If log_level was set as -1 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_create_extras(guc); + if (ret) + goto err; + + ret = guc_log_create_relay_file(guc); + if (ret) + goto err; + + return 0; +err: + guc_log_cleanup(guc); + /* logging will remain off */ + i915.guc_log_level = -1; + return ret; +} + +static void guc_log_capture_logs(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + + guc_read_update_log_buffer(guc); + + /* Generally device is expected to be active only at this + * time, so get/put should be really quick. + */ + intel_runtime_pm_get(dev_priv); + guc_log_flush_complete(guc); + intel_runtime_pm_put(dev_priv); +} + +static void guc_flush_logs(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + + if (!i915.enable_guc_submission || (i915.guc_log_level < 0)) + return; + + /* First disable the interrupts, will be renabled afterwards */ + gen9_disable_guc_interrupts(dev_priv); + + /* Before initiating the forceful flush, wait for any pending/ongoing + * flush to complete otherwise forceful flush may not actually happen. + */ + flush_work(&guc->log.flush_work); + + /* Ask GuC to update the log buffer state */ + guc_log_flush(guc); + + /* GuC would have updated log buffer by now, so capture it */ + guc_log_capture_logs(guc); +} + +int i915_guc_log_control(struct drm_i915_private *dev_priv, u64 control_val) +{ + struct intel_guc *guc = &dev_priv->guc; + + union guc_log_control log_param; + int ret; + + log_param.value = control_val; + + if (log_param.verbosity < GUC_LOG_VERBOSITY_MIN || + log_param.verbosity > GUC_LOG_VERBOSITY_MAX) + return -EINVAL; + + /* This combination doesn't make sense & won't have any effect */ + if (!log_param.logging_enabled && (i915.guc_log_level < 0)) + return 0; + + ret = guc_log_control(guc, log_param.value); + if (ret < 0) { + DRM_DEBUG_DRIVER("guc_logging_control action failed %d\n", ret); + return ret; + } + + i915.guc_log_level = log_param.verbosity; + + /* If log_level was set as -1 at boot time, then the relay channel file + * wouldn't have been created by now and interrupts also would not have + * been enabled. + */ + if (!dev_priv->guc.log.relay_chan) { + ret = guc_log_late_setup(guc); + if (!ret) + gen9_enable_guc_interrupts(dev_priv); + } else if (!log_param.logging_enabled) { + /* 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); + + /* As logging is disabled, update log level to reflect that */ + i915.guc_log_level = -1; + } else { + /* In case interrupts were disabled, enable them now */ + gen9_enable_guc_interrupts(dev_priv); + } + + return ret; +} + +void i915_guc_log_register(struct drm_i915_private *dev_priv) +{ + if (!i915.enable_guc_submission) + return; + + mutex_lock(&dev_priv->drm.struct_mutex); + guc_log_late_setup(&dev_priv->guc); + mutex_unlock(&dev_priv->drm.struct_mutex); +} + +void i915_guc_log_unregister(struct drm_i915_private *dev_priv) +{ + if (!i915.enable_guc_submission) + return; + + mutex_lock(&dev_priv->drm.struct_mutex); + guc_log_cleanup(&dev_priv->guc); + mutex_unlock(&dev_priv->drm.struct_mutex); +} diff --git a/drivers/gpu/drm/i915/intel_huc.c b/drivers/gpu/drm/i915/intel_huc.c new file mode 100644 index 0000000..c144609 --- /dev/null +++ b/drivers/gpu/drm/i915/intel_huc.c @@ -0,0 +1,338 @@ +/* + * Copyright © 2016-2017 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. + * + */ +#include <linux/firmware.h> +#include "i915_drv.h" +#include "intel_uc.h" + +/** + * DOC: HuC Firmware + * + * Motivation: + * GEN9 introduces a new dedicated firmware for usage in media HEVC (High + * Efficiency Video Coding) operations. Userspace can use the firmware + * capabilities by adding HuC specific commands to batch buffers. + * + * Implementation: + * The same firmware loader is used as the GuC. However, the actual + * loading to HW is deferred until GEM initialization is done. + * + * Note that HuC firmware loading must be done before GuC loading. + */ + +#define BXT_HUC_FW_MAJOR 01 +#define BXT_HUC_FW_MINOR 07 +#define BXT_BLD_NUM 1398 + +#define SKL_HUC_FW_MAJOR 01 +#define SKL_HUC_FW_MINOR 07 +#define SKL_BLD_NUM 1398 + +#define KBL_HUC_FW_MAJOR 02 +#define KBL_HUC_FW_MINOR 00 +#define KBL_BLD_NUM 1810 + +#define HUC_FW_PATH(platform, major, minor, bld_num) \ + "i915/" __stringify(platform) "_huc_ver" __stringify(major) "_" \ + __stringify(minor) "_" __stringify(bld_num) ".bin" + +#define I915_SKL_HUC_UCODE HUC_FW_PATH(skl, SKL_HUC_FW_MAJOR, \ + SKL_HUC_FW_MINOR, SKL_BLD_NUM) +MODULE_FIRMWARE(I915_SKL_HUC_UCODE); + +#define I915_BXT_HUC_UCODE HUC_FW_PATH(bxt, BXT_HUC_FW_MAJOR, \ + BXT_HUC_FW_MINOR, BXT_BLD_NUM) +MODULE_FIRMWARE(I915_BXT_HUC_UCODE); + +#define I915_KBL_HUC_UCODE HUC_FW_PATH(kbl, KBL_HUC_FW_MAJOR, \ + KBL_HUC_FW_MINOR, KBL_BLD_NUM) +MODULE_FIRMWARE(I915_KBL_HUC_UCODE); + +/** + * huc_ucode_xfer() - DMA's the firmware + * @dev_priv: the drm_i915_private device + * + * Transfer the firmware image to RAM for execution by the microcontroller. + * + * Return: 0 on success, non-zero on failure + */ +static int huc_ucode_xfer(struct drm_i915_private *dev_priv) +{ + struct intel_uc_fw *huc_fw = &dev_priv->huc.fw; + struct i915_vma *vma; + unsigned long offset = 0; + u32 size; + int ret; + + ret = i915_gem_object_set_to_gtt_domain(huc_fw->obj, false); + if (ret) { + DRM_DEBUG_DRIVER("set-domain failed %d\n", ret); + return ret; + } + + vma = i915_gem_object_ggtt_pin(huc_fw->obj, NULL, 0, 0, + PIN_OFFSET_BIAS | GUC_WOPCM_TOP); + if (IS_ERR(vma)) { + DRM_DEBUG_DRIVER("pin failed %d\n", (int)PTR_ERR(vma)); + return PTR_ERR(vma); + } + + intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); + + /* init WOPCM */ + I915_WRITE(GUC_WOPCM_SIZE, intel_guc_wopcm_size(dev_priv)); + I915_WRITE(DMA_GUC_WOPCM_OFFSET, GUC_WOPCM_OFFSET_VALUE | + HUC_LOADING_AGENT_GUC); + + /* Set the source address for the uCode */ + offset = guc_ggtt_offset(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); + + /* Hardware doesn't look at destination address for HuC. Set it to 0, + * but still program the correct address space. + */ + I915_WRITE(DMA_ADDR_1_LOW, 0); + I915_WRITE(DMA_ADDR_1_HIGH, DMA_ADDRESS_SPACE_WOPCM); + + size = huc_fw->header_size + huc_fw->ucode_size; + I915_WRITE(DMA_COPY_SIZE, size); + + /* Start the DMA */ + I915_WRITE(DMA_CTRL, _MASKED_BIT_ENABLE(HUC_UKERNEL | START_DMA)); + + /* Wait for DMA to finish */ + ret = wait_for((I915_READ(DMA_CTRL) & START_DMA) == 0, 100); + + DRM_DEBUG_DRIVER("HuC DMA transfer wait over with ret %d\n", ret); + + /* Disable the bits once DMA is over */ + I915_WRITE(DMA_CTRL, _MASKED_BIT_DISABLE(HUC_UKERNEL)); + + intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); + + /* + * We keep the object pages for reuse during resume. But we can unpin it + * now that DMA has completed, so it doesn't continue to take up space. + */ + i915_vma_unpin(vma); + + return ret; +} + +/** + * intel_huc_init() - initiate HuC firmware loading request + * @dev_priv: the drm_i915_private device + * + * Called early during driver load, but after GEM is initialised. The loading + * will continue only when driver explicitly specify firmware name and version. + * All other cases are considered as INTEL_UC_FIRMWARE_NONE either because HW + * is not capable or driver yet support it. And there will be no error message + * for INTEL_UC_FIRMWARE_NONE cases. + * + * The DMA-copying to HW is done later when intel_huc_load() is called. + */ +void intel_huc_init(struct drm_i915_private *dev_priv) +{ + struct intel_huc *huc = &dev_priv->huc; + struct intel_uc_fw *huc_fw = &huc->fw; + const char *fw_path = NULL; + + huc_fw->path = NULL; + huc_fw->fetch_status = INTEL_UC_FIRMWARE_NONE; + huc_fw->load_status = INTEL_UC_FIRMWARE_NONE; + huc_fw->fw = INTEL_UC_FW_TYPE_HUC; + + if (!HAS_HUC_UCODE(dev_priv)) + return; + + if (IS_SKYLAKE(dev_priv)) { + fw_path = I915_SKL_HUC_UCODE; + huc_fw->major_ver_wanted = SKL_HUC_FW_MAJOR; + huc_fw->minor_ver_wanted = SKL_HUC_FW_MINOR; + } else if (IS_BROXTON(dev_priv)) { + fw_path = I915_BXT_HUC_UCODE; + huc_fw->major_ver_wanted = BXT_HUC_FW_MAJOR; + huc_fw->minor_ver_wanted = BXT_HUC_FW_MINOR; + } else if (IS_KABYLAKE(dev_priv)) { + fw_path = I915_KBL_HUC_UCODE; + huc_fw->major_ver_wanted = KBL_HUC_FW_MAJOR; + huc_fw->minor_ver_wanted = KBL_HUC_FW_MINOR; + } + + huc_fw->path = fw_path; + huc_fw->fetch_status = INTEL_UC_FIRMWARE_PENDING; + + DRM_DEBUG_DRIVER("HuC firmware pending, path %s\n", fw_path); + + WARN(huc_fw->path == NULL, "HuC present but no fw path\n"); + + intel_uc_fw_fetch(dev_priv, huc_fw); +} + +/** + * intel_huc_load() - load HuC uCode to device + * @dev_priv: the drm_i915_private device + * + * Called from guc_setup() during driver loading and also after a GPU reset. + * Be note that HuC loading must be done before GuC loading. + * + * The firmware image should have already been fetched into memory by the + * earlier call to intel_huc_init(), so here we need only check that + * is succeeded, and then transfer the image to the h/w. + * + * Return: non-zero code on error + */ +int intel_huc_load(struct drm_i915_private *dev_priv) +{ + struct intel_uc_fw *huc_fw = &dev_priv->huc.fw; + int err; + + if (huc_fw->fetch_status == INTEL_UC_FIRMWARE_NONE) + return 0; + + DRM_DEBUG_DRIVER("%s fw status: fetch %s, load %s\n", + huc_fw->path, + intel_uc_fw_status_repr(huc_fw->fetch_status), + intel_uc_fw_status_repr(huc_fw->load_status)); + + if (huc_fw->fetch_status == INTEL_UC_FIRMWARE_SUCCESS && + huc_fw->load_status == INTEL_UC_FIRMWARE_FAIL) + return -ENOEXEC; + + huc_fw->load_status = INTEL_UC_FIRMWARE_PENDING; + + switch (huc_fw->fetch_status) { + case INTEL_UC_FIRMWARE_FAIL: + /* something went wrong :( */ + err = -EIO; + goto fail; + + case INTEL_UC_FIRMWARE_NONE: + case INTEL_UC_FIRMWARE_PENDING: + default: + /* "can't happen" */ + WARN_ONCE(1, "HuC fw %s invalid fetch_status %s [%d]\n", + huc_fw->path, + intel_uc_fw_status_repr(huc_fw->fetch_status), + huc_fw->fetch_status); + err = -ENXIO; + goto fail; + + case INTEL_UC_FIRMWARE_SUCCESS: + break; + } + + err = huc_ucode_xfer(dev_priv); + if (err) + goto fail; + + huc_fw->load_status = INTEL_UC_FIRMWARE_SUCCESS; + + DRM_DEBUG_DRIVER("%s fw status: fetch %s, load %s\n", + huc_fw->path, + intel_uc_fw_status_repr(huc_fw->fetch_status), + intel_uc_fw_status_repr(huc_fw->load_status)); + + return 0; + +fail: + if (huc_fw->load_status == INTEL_UC_FIRMWARE_PENDING) + huc_fw->load_status = INTEL_UC_FIRMWARE_FAIL; + + DRM_ERROR("Failed to complete HuC uCode load with ret %d\n", err); + + return err; +} + +/** + * intel_huc_fini() - clean up resources allocated for HuC + * @dev_priv: the drm_i915_private device + * + * Cleans up by releasing the huc firmware GEM obj. + */ +void intel_huc_fini(struct drm_i915_private *dev_priv) +{ + struct intel_uc_fw *huc_fw = &dev_priv->huc.fw; + + mutex_lock(&dev_priv->drm.struct_mutex); + if (huc_fw->obj) + i915_gem_object_put(huc_fw->obj); + huc_fw->obj = NULL; + mutex_unlock(&dev_priv->drm.struct_mutex); + + huc_fw->fetch_status = INTEL_UC_FIRMWARE_NONE; +} + +/** + * intel_guc_auth_huc() - authenticate ucode + * @dev_priv: the drm_i915_device + * + * Triggers a HuC fw authentication request to the GuC via intel_guc_action_ + * authenticate_huc interface. + */ +void intel_guc_auth_huc(struct drm_i915_private *dev_priv) +{ + struct intel_guc *guc = &dev_priv->guc; + struct intel_huc *huc = &dev_priv->huc; + struct i915_vma *vma; + int ret; + u32 data[2]; + + if (huc->fw.load_status != INTEL_UC_FIRMWARE_SUCCESS) + return; + + vma = i915_gem_object_ggtt_pin(huc->fw.obj, NULL, 0, 0, + PIN_OFFSET_BIAS | GUC_WOPCM_TOP); + if (IS_ERR(vma)) { + DRM_ERROR("failed to pin huc fw object %d\n", + (int)PTR_ERR(vma)); + return; + } + + /* Specify auth action and where public signature is. */ + data[0] = INTEL_GUC_ACTION_AUTHENTICATE_HUC; + data[1] = guc_ggtt_offset(vma) + huc->fw.rsa_offset; + + ret = intel_guc_send(guc, data, ARRAY_SIZE(data)); + if (ret) { + DRM_ERROR("HuC: GuC did not ack Auth request %d\n", ret); + goto out; + } + + /* Check authentication status, it should be done by now */ + ret = intel_wait_for_register(dev_priv, + HUC_STATUS2, + HUC_FW_VERIFIED, + HUC_FW_VERIFIED, + 50); + + if (ret) { + DRM_ERROR("HuC: Authentication failed %d\n", ret); + goto out; + } + +out: + i915_vma_unpin(vma); +} + diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 656e0a3..432ee49 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -811,12 +811,6 @@ static int execlists_context_pin(struct intel_engine_cs *engine, ce->state->obj->mm.dirty = true; - /* Invalidate GuC TLB. */ - if (i915.enable_guc_submission) { - struct drm_i915_private *dev_priv = ctx->i915; - I915_WRITE(GEN8_GTCR, GEN8_GTCR_INVALIDATE); - } - i915_gem_context_get(ctx); return 0; @@ -1231,7 +1225,7 @@ static int lrc_setup_wa_ctx_obj(struct intel_engine_cs *engine, u32 size) if (IS_ERR(obj)) return PTR_ERR(obj); - vma = i915_vma_create(obj, &engine->i915->ggtt.base, NULL); + vma = i915_vma_instance(obj, &engine->i915->ggtt.base, NULL); if (IS_ERR(vma)) { err = PTR_ERR(vma); goto err; @@ -1917,7 +1911,7 @@ int logical_render_ring_init(struct intel_engine_cs *engine) engine->emit_breadcrumb = gen8_emit_breadcrumb_render; engine->emit_breadcrumb_sz = gen8_emit_breadcrumb_render_sz; - ret = intel_engine_create_scratch(engine, 4096); + ret = intel_engine_create_scratch(engine, PAGE_SIZE); if (ret) return ret; @@ -2093,19 +2087,12 @@ static void execlists_init_reg_state(u32 *reg_state, ASSIGN_CTX_REG(reg_state, CTX_PDP0_LDW, GEN8_RING_PDP_LDW(engine, 0), 0); - if (USES_FULL_48BIT_PPGTT(ppgtt->base.dev)) { + if (ppgtt && USES_FULL_48BIT_PPGTT(ppgtt->base.dev)) { /* 64b PPGTT (48bit canonical) * PDP0_DESCRIPTOR contains the base address to PML4 and * other PDP Descriptors are ignored. */ ASSIGN_CTX_PML4(ppgtt, reg_state); - } else { - /* 32b PPGTT - * PDP*_DESCRIPTOR contains the base address of space supported. - * With dynamic page allocation, PDPs may not be allocated at - * this point. Point the unallocated PDPs to the scratch page - */ - execlists_update_context_pdps(ppgtt, reg_state); } if (engine->id == RCS) { @@ -2199,7 +2186,8 @@ static int execlists_context_deferred_alloc(struct i915_gem_context *ctx, WARN_ON(ce->state); - context_size = round_up(intel_lr_context_size(engine), 4096); + context_size = round_up(intel_lr_context_size(engine), + I915_GTT_PAGE_SIZE); /* One extra page as the sharing data between driver and GuC */ context_size += PAGE_SIZE * LRC_PPHWSP_PN; @@ -2210,7 +2198,7 @@ static int execlists_context_deferred_alloc(struct i915_gem_context *ctx, return PTR_ERR(ctx_obj); } - vma = i915_vma_create(ctx_obj, &ctx->i915->ggtt.base, NULL); + vma = i915_vma_instance(ctx_obj, &ctx->i915->ggtt.base, NULL); if (IS_ERR(vma)) { ret = PTR_ERR(vma); goto error_deref_obj; diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h index 01ba36e..0c852c0 100644 --- a/drivers/gpu/drm/i915/intel_lrc.h +++ b/drivers/gpu/drm/i915/intel_lrc.h @@ -26,7 +26,7 @@ #include "intel_ringbuffer.h" -#define GEN8_LR_CONTEXT_ALIGN 4096 +#define GEN8_LR_CONTEXT_ALIGN I915_GTT_MIN_ALIGNMENT /* Execlists regs */ #define RING_ELSP(engine) _MMIO((engine)->mmio_base + 0x230) diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c index 4473a61..0608fad 100644 --- a/drivers/gpu/drm/i915/intel_overlay.c +++ b/drivers/gpu/drm/i915/intel_overlay.c @@ -811,8 +811,7 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay, if (ret != 0) return ret; - vma = i915_gem_object_pin_to_display_plane(new_bo, 0, - &i915_ggtt_view_normal); + vma = i915_gem_object_pin_to_display_plane(new_bo, 0, NULL); if (IS_ERR(vma)) return PTR_ERR(vma); diff --git a/drivers/gpu/drm/i915/intel_pipe_crc.c b/drivers/gpu/drm/i915/intel_pipe_crc.c index ef0c0e1..c0b1f99 100644 --- a/drivers/gpu/drm/i915/intel_pipe_crc.c +++ b/drivers/gpu/drm/i915/intel_pipe_crc.c @@ -560,14 +560,14 @@ static void hsw_trans_edp_pipe_A_crc_wa(struct drm_i915_private *dev_priv, state = drm_atomic_state_alloc(dev); if (!state) { ret = -ENOMEM; - goto out; + goto unlock; } state->acquire_ctx = drm_modeset_legacy_acquire_ctx(&crtc->base); pipe_config = intel_atomic_get_crtc_state(state, crtc); if (IS_ERR(pipe_config)) { ret = PTR_ERR(pipe_config); - goto out; + goto put_state; } pipe_config->pch_pfit.force_thru = enable; @@ -576,10 +576,12 @@ static void hsw_trans_edp_pipe_A_crc_wa(struct drm_i915_private *dev_priv, pipe_config->base.connectors_changed = true; ret = drm_atomic_commit(state); -out: + +put_state: + drm_atomic_state_put(state); +unlock: WARN(ret, "Toggling workaround to %i returns %i\n", enable, ret); drm_modeset_unlock_all(dev); - drm_atomic_state_put(state); } static int ivb_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, @@ -613,6 +615,22 @@ static int ivb_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, return 0; } +static int get_new_crc_ctl_reg(struct drm_i915_private *dev_priv, + enum pipe pipe, + enum intel_pipe_crc_source *source, u32 *val) +{ + if (IS_GEN2(dev_priv)) + return i8xx_pipe_crc_ctl_reg(source, val); + else if (INTEL_GEN(dev_priv) < 5) + return i9xx_pipe_crc_ctl_reg(dev_priv, pipe, source, val); + else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) + return vlv_pipe_crc_ctl_reg(dev_priv, pipe, source, val); + 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); +} + static int pipe_crc_set_source(struct drm_i915_private *dev_priv, enum pipe pipe, enum intel_pipe_crc_source source) @@ -636,17 +654,7 @@ static int pipe_crc_set_source(struct drm_i915_private *dev_priv, return -EIO; } - if (IS_GEN2(dev_priv)) - ret = i8xx_pipe_crc_ctl_reg(&source, &val); - else if (INTEL_GEN(dev_priv) < 5) - ret = i9xx_pipe_crc_ctl_reg(dev_priv, pipe, &source, &val); - else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) - ret = vlv_pipe_crc_ctl_reg(dev_priv, pipe, &source, &val); - else if (IS_GEN5(dev_priv) || IS_GEN6(dev_priv)) - ret = ilk_pipe_crc_ctl_reg(&source, &val); - else - ret = ivb_pipe_crc_ctl_reg(dev_priv, pipe, &source, &val); - + ret = get_new_crc_ctl_reg(dev_priv, pipe, &source, &val); if (ret != 0) goto out; @@ -687,7 +695,7 @@ static int pipe_crc_set_source(struct drm_i915_private *dev_priv, POSTING_READ(PIPE_CRC_CTL(pipe)); /* real source -> none transition */ - if (source == INTEL_PIPE_CRC_SOURCE_NONE) { + if (!source) { struct intel_pipe_crc_entry *entries; struct intel_crtc *crtc = intel_get_crtc_for_pipe(dev_priv, pipe); @@ -809,6 +817,11 @@ display_crc_ctl_parse_source(const char *buf, enum intel_pipe_crc_source *s) { int i; + if (!buf) { + *s = INTEL_PIPE_CRC_SOURCE_NONE; + return 0; + } + for (i = 0; i < ARRAY_SIZE(pipe_crc_sources); i++) if (!strcmp(buf, pipe_crc_sources[i])) { *s = i; @@ -937,3 +950,62 @@ void intel_pipe_crc_cleanup(struct drm_minor *minor) drm_debugfs_remove_files(info_list, 1, 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 intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[crtc->index]; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + enum intel_display_power_domain power_domain; + enum intel_pipe_crc_source source; + u32 val = 0; /* shut up gcc */ + int ret = 0; + + if (display_crc_ctl_parse_source(source_name, &source) < 0) { + DRM_DEBUG_DRIVER("unknown source %s\n", source_name); + return -EINVAL; + } + + power_domain = POWER_DOMAIN_PIPE(crtc->index); + if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) { + DRM_DEBUG_KMS("Trying to capture CRC while pipe is off\n"); + return -EIO; + } + + ret = get_new_crc_ctl_reg(dev_priv, crtc->index, &source, &val); + if (ret != 0) + goto out; + + if (source) { + /* + * When IPS gets enabled, the pipe CRC changes. Since IPS gets + * enabled and disabled dynamically based on package C states, + * user space can't make reliable use of the CRCs, so let's just + * completely disable it. + */ + hsw_disable_ips(intel_crtc); + } + + I915_WRITE(PIPE_CRC_CTL(crtc->index), val); + POSTING_READ(PIPE_CRC_CTL(crtc->index)); + + if (!source) { + if (IS_G4X(dev_priv)) + g4x_undo_pipe_scramble_reset(dev_priv, crtc->index); + else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) + vlv_undo_pipe_scramble_reset(dev_priv, crtc->index); + else if (IS_HASWELL(dev_priv) && crtc->index == PIPE_A) + hsw_trans_edp_pipe_A_crc_wa(dev_priv, false); + + hsw_enable_ips(intel_crtc); + } + + pipe_crc->skipped = 0; + *values_cnt = 5; + +out: + intel_display_power_put(dev_priv, power_domain); + + return ret; +} diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index 6aca8ff..c3780d0 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -122,13 +122,26 @@ static void vlv_psr_setup_vsc(struct intel_dp *intel_dp) static void skl_psr_setup_su_vsc(struct intel_dp *intel_dp) { struct edp_vsc_psr psr_vsc; + struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); + struct drm_device *dev = intel_dig_port->base.base.dev; + struct drm_i915_private *dev_priv = to_i915(dev); /* 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; psr_vsc.sdp_header.HB1 = 0x7; - psr_vsc.sdp_header.HB2 = 0x3; - psr_vsc.sdp_header.HB3 = 0xb; + if (dev_priv->psr.colorimetry_support && + dev_priv->psr.y_cord_support) { + psr_vsc.sdp_header.HB2 = 0x5; + psr_vsc.sdp_header.HB3 = 0x13; + } else if (dev_priv->psr.y_cord_support) { + 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; + } + intel_psr_write_vsc(intel_dp, &psr_vsc); } @@ -196,7 +209,11 @@ static void hsw_psr_enable_sink(struct intel_dp *intel_dp) 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, + DP_RECEIVER_ALPM_CONFIG, + DP_ALPM_ENABLE); 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); @@ -248,7 +265,7 @@ static void vlv_psr_activate(struct intel_dp *intel_dp) VLV_EDP_PSR_ACTIVE_ENTRY); } -static void hsw_psr_enable_source(struct intel_dp *intel_dp) +static void intel_enable_source_psr1(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; @@ -299,14 +316,31 @@ static void hsw_psr_enable_source(struct intel_dp *intel_dp) val |= EDP_PSR_TP1_TP2_SEL; I915_WRITE(EDP_PSR_CTL, val); +} - if (!dev_priv->psr.psr2_support) - return; +static void intel_enable_source_psr2(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); + /* + * Let's respect VBT in case VBT asks a higher idle_frame value. + * Let's use 6 as the minimum to cover all known cases including + * the off-by-one issue that HW has in some cases. Also there are + * cases where sink should be able to train + * with the 5 or 6 idle patterns. + */ + uint32_t idle_frames = max(6, dev_priv->vbt.psr.idle_frames); + uint32_t val; + + val = idle_frames << EDP_PSR_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 * good enough. */ - val = EDP_PSR2_ENABLE | EDP_SU_TRACK_ENABLE; + val |= EDP_PSR2_ENABLE | + EDP_SU_TRACK_ENABLE | + EDP_FRAMES_BEFORE_SU_ENTRY; if (dev_priv->vbt.psr.tp2_tp3_wakeup_time > 5) val |= EDP_PSR2_TP2_TIME_2500; @@ -320,6 +354,19 @@ static void hsw_psr_enable_source(struct intel_dp *intel_dp) I915_WRITE(EDP_PSR2_CTL, val); } +static void hsw_psr_enable_source(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); + + /* psr1 and psr2 are mutually exclusive.*/ + if (dev_priv->psr.psr2_support) + intel_enable_source_psr2(intel_dp); + else + intel_enable_source_psr1(intel_dp); +} + static bool intel_psr_match_conditions(struct intel_dp *intel_dp) { struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); @@ -387,6 +434,22 @@ static bool intel_psr_match_conditions(struct intel_dp *intel_dp) return false; } + /* PSR2 is restricted to work with panel resolutions upto 3200x2000 */ + if (intel_crtc->config->pipe_src_w > 3200 || + intel_crtc->config->pipe_src_h > 2000) { + dev_priv->psr.psr2_support = false; + return false; + } + + /* + * FIXME:enable psr2 only for y-cordinate psr2 panels + * After gtc implementation , remove this restriction. + */ + if (!dev_priv->psr.y_cord_support && dev_priv->psr.psr2_support) { + DRM_DEBUG_KMS("PSR2 disabled, panel does not support Y coordinate\n"); + return false; + } + dev_priv->psr.source_ok = true; return true; } @@ -397,7 +460,10 @@ 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); - WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE); + if (dev_priv->psr.psr2_support) + WARN_ON(I915_READ(EDP_PSR2_CTL) & EDP_PSR2_ENABLE); + else + WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE); WARN_ON(dev_priv->psr.active); lockdep_assert_held(&dev_priv->psr.lock); @@ -426,6 +492,8 @@ void intel_psr_enable(struct intel_dp *intel_dp) struct drm_device *dev = intel_dig_port->base.base.dev; struct drm_i915_private *dev_priv = to_i915(dev); struct intel_crtc *crtc = to_intel_crtc(intel_dig_port->base.base.crtc); + enum transcoder cpu_transcoder = crtc->config->cpu_transcoder; + u32 chicken; if (!HAS_PSR(dev_priv)) { DRM_DEBUG_KMS("PSR not supported on this platform\n"); @@ -449,26 +517,34 @@ void intel_psr_enable(struct intel_dp *intel_dp) dev_priv->psr.busy_frontbuffer_bits = 0; if (HAS_DDI(dev_priv)) { - hsw_psr_setup_vsc(intel_dp); - if (dev_priv->psr.psr2_support) { - /* PSR2 is restricted to work with panel resolutions upto 3200x2000 */ - if (crtc->config->pipe_src_w > 3200 || - crtc->config->pipe_src_h > 2000) - dev_priv->psr.psr2_support = false; - else - skl_psr_setup_su_vsc(intel_dp); + skl_psr_setup_su_vsc(intel_dp); + chicken = PSR2_VSC_ENABLE_PROG_HEADER; + if (dev_priv->psr.y_cord_support) + chicken |= PSR2_ADD_VERTICAL_LINE_COUNT; + I915_WRITE(CHICKEN_TRANS(cpu_transcoder), chicken); + I915_WRITE(EDP_PSR_DEBUG_CTL, + EDP_PSR_DEBUG_MASK_MEMUP | + EDP_PSR_DEBUG_MASK_HPD | + EDP_PSR_DEBUG_MASK_LPSP | + EDP_PSR_DEBUG_MASK_MAX_SLEEP | + EDP_PSR_DEBUG_MASK_DISP_REG_WRITE); + } else { + /* set up vsc header for psr1 */ + hsw_psr_setup_vsc(intel_dp); + /* + * Per Spec: Avoid continuous PSR exit by masking MEMUP + * and HPD. also mask LPSP to avoid dependency on other + * drivers that might block runtime_pm besides + * preventing other hw tracking issues now we can rely + * on frontbuffer tracking. + */ + I915_WRITE(EDP_PSR_DEBUG_CTL, + EDP_PSR_DEBUG_MASK_MEMUP | + EDP_PSR_DEBUG_MASK_HPD | + EDP_PSR_DEBUG_MASK_LPSP); } - /* - * Per Spec: Avoid continuous PSR exit by masking MEMUP and HPD. - * Also mask LPSP to avoid dependency on other drivers that - * might block runtime_pm besides preventing other hw tracking - * issues now we can rely on frontbuffer tracking. - */ - I915_WRITE(EDP_PSR_DEBUG_CTL, EDP_PSR_DEBUG_MASK_MEMUP | - EDP_PSR_DEBUG_MASK_HPD | EDP_PSR_DEBUG_MASK_LPSP); - /* Enable PSR on the panel */ hsw_psr_enable_sink(intel_dp); @@ -544,20 +620,42 @@ static void hsw_psr_disable(struct intel_dp *intel_dp) struct drm_i915_private *dev_priv = to_i915(dev); if (dev_priv->psr.active) { - I915_WRITE(EDP_PSR_CTL, - I915_READ(EDP_PSR_CTL) & ~EDP_PSR_ENABLE); + i915_reg_t psr_ctl; + 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_ctl = EDP_PSR2_CTL; + psr_status_mask = EDP_PSR2_STATUS_STATE_MASK; + + I915_WRITE(psr_ctl, + I915_READ(psr_ctl) & + ~(EDP_PSR2_ENABLE | EDP_SU_TRACK_ENABLE)); + + } else { + psr_ctl = EDP_PSR_STATUS_CTL; + psr_status_mask = EDP_PSR_STATUS_STATE_MASK; + + I915_WRITE(psr_ctl, + I915_READ(psr_ctl) & ~EDP_PSR_ENABLE); + } /* Wait till PSR is idle */ if (intel_wait_for_register(dev_priv, - EDP_PSR_STATUS_CTL, - EDP_PSR_STATUS_STATE_MASK, - 0, + psr_ctl, psr_status_mask, 0, 2000)) DRM_ERROR("Timed out waiting for PSR Idle State\n"); dev_priv->psr.active = false; } else { - WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE); + if (dev_priv->psr.psr2_support) + WARN_ON(I915_READ(EDP_PSR2_CTL) & EDP_PSR2_ENABLE); + else + WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE); } } @@ -608,13 +706,24 @@ static void intel_psr_work(struct work_struct *work) * and be ready for re-enable. */ if (HAS_DDI(dev_priv)) { - if (intel_wait_for_register(dev_priv, - EDP_PSR_STATUS_CTL, - EDP_PSR_STATUS_STATE_MASK, - 0, - 50)) { - DRM_ERROR("Timed out waiting for PSR Idle for re-enable\n"); - return; + if (dev_priv->psr.psr2_support) { + if (intel_wait_for_register(dev_priv, + EDP_PSR2_STATUS_CTL, + EDP_PSR2_STATUS_STATE_MASK, + 0, + 50)) { + DRM_ERROR("Timed out waiting for PSR2 Idle for re-enable\n"); + return; + } + } else { + if (intel_wait_for_register(dev_priv, + EDP_PSR_STATUS_CTL, + EDP_PSR_STATUS_STATE_MASK, + 0, + 50)) { + DRM_ERROR("Timed out waiting for PSR Idle for re-enable\n"); + return; + } } } else { if (intel_wait_for_register(dev_priv, @@ -656,11 +765,19 @@ static void intel_psr_exit(struct drm_i915_private *dev_priv) return; if (HAS_DDI(dev_priv)) { - val = I915_READ(EDP_PSR_CTL); - - WARN_ON(!(val & EDP_PSR_ENABLE)); - - I915_WRITE(EDP_PSR_CTL, val & ~EDP_PSR_ENABLE); + 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)); + I915_WRITE(EDP_PSR2_CTL, val & ~EDP_PSR2_ENABLE); + } else { + val = I915_READ(EDP_PSR_CTL); + WARN_ON(!(val & EDP_PSR_ENABLE)); + I915_WRITE(EDP_PSR_CTL, val & ~EDP_PSR_ENABLE); + } } else { val = I915_READ(VLV_PSRCTL(pipe)); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 7cb2ab4..69035e4 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1728,7 +1728,7 @@ static int init_status_page(struct intel_engine_cs *engine) void *vaddr; int ret; - obj = i915_gem_object_create_internal(engine->i915, 4096); + obj = i915_gem_object_create_internal(engine->i915, PAGE_SIZE); if (IS_ERR(obj)) { DRM_ERROR("Failed to allocate status page\n"); return PTR_ERR(obj); @@ -1738,7 +1738,7 @@ static int init_status_page(struct intel_engine_cs *engine) if (ret) goto err; - vma = i915_vma_create(obj, &engine->i915->ggtt.base, NULL); + vma = i915_vma_instance(obj, &engine->i915->ggtt.base, NULL); if (IS_ERR(vma)) { ret = PTR_ERR(vma); goto err; @@ -1769,7 +1769,7 @@ static int init_status_page(struct intel_engine_cs *engine) engine->status_page.vma = vma; engine->status_page.ggtt_offset = i915_ggtt_offset(vma); - engine->status_page.page_addr = memset(vaddr, 0, 4096); + engine->status_page.page_addr = memset(vaddr, 0, PAGE_SIZE); DRM_DEBUG_DRIVER("%s hws offset: 0x%08x\n", engine->name, i915_ggtt_offset(vma)); @@ -1872,7 +1872,7 @@ intel_ring_create_vma(struct drm_i915_private *dev_priv, int size) /* mark ring buffers as read-only from GPU side by default */ obj->gt_ro = 1; - vma = i915_vma_create(obj, &dev_priv->ggtt.base, NULL); + vma = i915_vma_instance(obj, &dev_priv->ggtt.base, NULL); if (IS_ERR(vma)) goto err; @@ -2041,7 +2041,7 @@ static int intel_init_ring_buffer(struct intel_engine_cs *engine) } /* Ring wraparound at offset 0 sometimes hangs. No idea why. */ - ret = intel_ring_pin(ring, 4096); + ret = intel_ring_pin(ring, I915_GTT_PAGE_SIZE); if (ret) { intel_ring_free(ring); goto error; @@ -2458,11 +2458,11 @@ static void intel_ring_init_semaphores(struct drm_i915_private *dev_priv, if (INTEL_GEN(dev_priv) >= 8 && !dev_priv->semaphore) { struct i915_vma *vma; - obj = i915_gem_object_create(dev_priv, 4096); + obj = i915_gem_object_create(dev_priv, PAGE_SIZE); if (IS_ERR(obj)) goto err; - vma = i915_vma_create(obj, &dev_priv->ggtt.base, NULL); + vma = i915_vma_instance(obj, &dev_priv->ggtt.base, NULL); if (IS_ERR(vma)) goto err_obj; @@ -2675,7 +2675,7 @@ int intel_init_render_ring_buffer(struct intel_engine_cs *engine) return ret; if (INTEL_GEN(dev_priv) >= 6) { - ret = intel_engine_create_scratch(engine, 4096); + ret = intel_engine_create_scratch(engine, PAGE_SIZE); if (ret) return ret; } else if (HAS_BROKEN_CS_TLB(dev_priv)) { diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c index 7031bc7..9ef5468 100644 --- a/drivers/gpu/drm/i915/intel_sprite.c +++ b/drivers/gpu/drm/i915/intel_sprite.c @@ -273,7 +273,7 @@ skl_update_plane(struct drm_plane *drm_plane, I915_WRITE(PLANE_CTL(pipe, plane_id), plane_ctl); I915_WRITE(PLANE_SURF(pipe, plane_id), - intel_fb_gtt_offset(fb, rotation) + surf_addr); + intel_plane_ggtt_offset(plane_state) + surf_addr); POSTING_READ(PLANE_SURF(pipe, plane_id)); } @@ -458,7 +458,7 @@ vlv_update_plane(struct drm_plane *dplane, I915_WRITE(SPSIZE(pipe, plane_id), (crtc_h << 16) | crtc_w); I915_WRITE(SPCNTR(pipe, plane_id), sprctl); I915_WRITE(SPSURF(pipe, plane_id), - intel_fb_gtt_offset(fb, rotation) + sprsurf_offset); + intel_plane_ggtt_offset(plane_state) + sprsurf_offset); POSTING_READ(SPSURF(pipe, plane_id)); } @@ -594,7 +594,7 @@ ivb_update_plane(struct drm_plane *plane, I915_WRITE(SPRSCALE(pipe), sprscale); I915_WRITE(SPRCTL(pipe), sprctl); I915_WRITE(SPRSURF(pipe), - intel_fb_gtt_offset(fb, rotation) + sprsurf_offset); + intel_plane_ggtt_offset(plane_state) + sprsurf_offset); POSTING_READ(SPRSURF(pipe)); } @@ -721,7 +721,7 @@ ilk_update_plane(struct drm_plane *plane, I915_WRITE(DVSSCALE(pipe), dvsscale); I915_WRITE(DVSCNTR(pipe), dvscntr); I915_WRITE(DVSSURF(pipe), - intel_fb_gtt_offset(fb, rotation) + dvssurf_offset); + intel_plane_ggtt_offset(plane_state) + dvssurf_offset); POSTING_READ(DVSSURF(pipe)); } diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index c6be352..c46bc85 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -114,29 +114,3 @@ int intel_guc_sample_forcewake(struct intel_guc *guc) return intel_guc_send(guc, action, ARRAY_SIZE(action)); } -int intel_guc_log_flush_complete(struct intel_guc *guc) -{ - u32 action[] = { INTEL_GUC_ACTION_LOG_BUFFER_FILE_FLUSH_COMPLETE }; - - return intel_guc_send(guc, action, ARRAY_SIZE(action)); -} - -int intel_guc_log_flush(struct intel_guc *guc) -{ - u32 action[] = { - INTEL_GUC_ACTION_FORCE_LOG_BUFFER_FLUSH, - 0 - }; - - return intel_guc_send(guc, action, ARRAY_SIZE(action)); -} - -int intel_guc_log_control(struct intel_guc *guc, u32 control_val) -{ - u32 action[] = { - INTEL_GUC_ACTION_UK_LOG_ENABLE_LOGGING, - control_val - }; - - return intel_guc_send(guc, action, ARRAY_SIZE(action)); -} diff --git a/drivers/gpu/drm/i915/intel_uc.h b/drivers/gpu/drm/i915/intel_uc.h index 9490a8e..d74f4d3 100644 --- a/drivers/gpu/drm/i915/intel_uc.h +++ b/drivers/gpu/drm/i915/intel_uc.h @@ -93,29 +93,35 @@ struct i915_guc_client { uint64_t submissions[I915_NUM_ENGINES]; }; -enum intel_guc_fw_status { - GUC_FIRMWARE_FAIL = -1, - GUC_FIRMWARE_NONE = 0, - GUC_FIRMWARE_PENDING, - GUC_FIRMWARE_SUCCESS +enum intel_uc_fw_status { + INTEL_UC_FIRMWARE_FAIL = -1, + INTEL_UC_FIRMWARE_NONE = 0, + INTEL_UC_FIRMWARE_PENDING, + INTEL_UC_FIRMWARE_SUCCESS +}; + +enum intel_uc_fw_type { + INTEL_UC_FW_TYPE_GUC, + INTEL_UC_FW_TYPE_HUC }; /* * This structure encapsulates all the data needed during the process * of fetching, caching, and loading the firmware image into the GuC. */ -struct intel_guc_fw { - const char * guc_fw_path; - size_t guc_fw_size; - struct drm_i915_gem_object * guc_fw_obj; - enum intel_guc_fw_status guc_fw_fetch_status; - enum intel_guc_fw_status guc_fw_load_status; - - uint16_t guc_fw_major_wanted; - uint16_t guc_fw_minor_wanted; - uint16_t guc_fw_major_found; - uint16_t guc_fw_minor_found; - +struct intel_uc_fw { + const char *path; + size_t size; + struct drm_i915_gem_object *obj; + enum intel_uc_fw_status fetch_status; + enum intel_uc_fw_status load_status; + + uint16_t major_ver_wanted; + uint16_t minor_ver_wanted; + uint16_t major_ver_found; + uint16_t minor_ver_found; + + enum intel_uc_fw_type fw; uint32_t header_size; uint32_t header_offset; uint32_t rsa_size; @@ -141,7 +147,7 @@ struct intel_guc_log { }; struct intel_guc { - struct intel_guc_fw guc_fw; + struct intel_uc_fw fw; struct intel_guc_log log; /* intel_guc_recv interrupt related state */ @@ -170,21 +176,28 @@ struct intel_guc { struct mutex send_mutex; }; +struct intel_huc { + /* Generic uC firmware management */ + struct intel_uc_fw fw; + + /* HuC-specific additions */ +}; + /* intel_uc.c */ void intel_uc_init_early(struct drm_i915_private *dev_priv); int intel_guc_send(struct intel_guc *guc, const u32 *action, u32 len); int intel_guc_sample_forcewake(struct intel_guc *guc); -int intel_guc_log_flush_complete(struct intel_guc *guc); -int intel_guc_log_flush(struct intel_guc *guc); -int intel_guc_log_control(struct intel_guc *guc, u32 control_val); /* intel_guc_loader.c */ extern void intel_guc_init(struct drm_i915_private *dev_priv); extern int intel_guc_setup(struct drm_i915_private *dev_priv); extern void intel_guc_fini(struct drm_i915_private *dev_priv); -extern const char *intel_guc_fw_status_repr(enum intel_guc_fw_status status); +extern const char *intel_uc_fw_status_repr(enum intel_uc_fw_status status); extern int intel_guc_suspend(struct drm_i915_private *dev_priv); extern int intel_guc_resume(struct drm_i915_private *dev_priv); +void intel_uc_fw_fetch(struct drm_i915_private *dev_priv, + struct intel_uc_fw *uc_fw); +u32 intel_guc_wopcm_size(struct drm_i915_private *dev_priv); /* i915_guc_submission.c */ int i915_guc_submission_init(struct drm_i915_private *dev_priv); @@ -193,10 +206,12 @@ int i915_guc_wq_reserve(struct drm_i915_gem_request *rq); void i915_guc_wq_unreserve(struct drm_i915_gem_request *request); void i915_guc_submission_disable(struct drm_i915_private *dev_priv); void i915_guc_submission_fini(struct drm_i915_private *dev_priv); -void i915_guc_capture_logs(struct drm_i915_private *dev_priv); -void i915_guc_flush_logs(struct drm_i915_private *dev_priv); -void i915_guc_register(struct drm_i915_private *dev_priv); -void i915_guc_unregister(struct drm_i915_private *dev_priv); +struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size); + +/* intel_guc_log.c */ +void intel_guc_log_create(struct intel_guc *guc); +void i915_guc_log_register(struct drm_i915_private *dev_priv); +void i915_guc_log_unregister(struct drm_i915_private *dev_priv); int i915_guc_log_control(struct drm_i915_private *dev_priv, u64 control_val); static inline u32 guc_ggtt_offset(struct i915_vma *vma) @@ -207,4 +222,10 @@ static inline u32 guc_ggtt_offset(struct i915_vma *vma) return offset; } +/* intel_huc.c */ +void intel_huc_init(struct drm_i915_private *dev_priv); +void intel_huc_fini(struct drm_i915_private *dev_priv); +int intel_huc_load(struct drm_i915_private *dev_priv); +void intel_guc_auth_huc(struct drm_i915_private *dev_priv); + #endif diff --git a/include/drm/drm_dp_helper.h b/include/drm/drm_dp_helper.h index 55bbeb0..0468135 100644 --- a/include/drm/drm_dp_helper.h +++ b/include/drm/drm_dp_helper.h @@ -194,7 +194,8 @@ # define DP_PSR_SETUP_TIME_0 (6 << 1) # define DP_PSR_SETUP_TIME_MASK (7 << 1) # define DP_PSR_SETUP_TIME_SHIFT 1 - +# define DP_PSR2_SU_Y_COORDINATE_REQUIRED (1 << 4) /* eDP 1.4a */ +# define DP_PSR2_SU_GRANULARITY_REQUIRED (1 << 5) /* eDP 1.4b */ /* * 0x80-0x8f describe downstream port capabilities, but there are two layouts * based on whether DP_DETAILED_CAP_INFO_AVAILABLE was set. If it was not, @@ -568,6 +569,16 @@ #define DP_RECEIVER_ALPM_STATUS 0x200b /* eDP 1.4 */ # define DP_ALPM_LOCK_TIMEOUT_ERROR (1 << 0) +#define DP_DPRX_FEATURE_ENUMERATION_LIST 0x2210 /* DP 1.3 */ +# define DP_GTC_CAP (1 << 0) /* DP 1.3 */ +# define DP_SST_SPLIT_SDP_CAP (1 << 1) /* DP 1.4 */ +# define DP_AV_SYNC_CAP (1 << 2) /* DP 1.3 */ +# define DP_VSC_SDP_EXT_FOR_COLORIMETRY_SUPPORTED (1 << 3) /* DP 1.3 */ +# define DP_VSC_EXT_VESA_SDP_SUPPORTED (1 << 4) /* DP 1.4 */ +# define DP_VSC_EXT_VESA_SDP_CHAINING_SUPPORTED (1 << 5) /* DP 1.4 */ +# define DP_VSC_EXT_CEA_SDP_SUPPORTED (1 << 6) /* DP 1.4 */ +# define DP_VSC_EXT_CEA_SDP_CHAINING_SUPPORTED (1 << 7) /* DP 1.4 */ + /* DP 1.2 Sideband message defines */ /* peer device type - DP 1.2a Table 2-92 */ #define DP_PEER_DEVICE_NONE 0x0 diff --git a/include/uapi/drm/i915_drm.h b/include/uapi/drm/i915_drm.h index da32c2f..57093b4 100644 --- a/include/uapi/drm/i915_drm.h +++ b/include/uapi/drm/i915_drm.h @@ -395,6 +395,7 @@ typedef struct drm_i915_irq_wait { * priorities and the driver will attempt to execute batches in priority order. */ #define I915_PARAM_HAS_SCHEDULER 41 +#define I915_PARAM_HUC_STATUS 42 typedef struct drm_i915_getparam { __s32 param; diff --git a/sound/pci/hda/hda_codec.c b/sound/pci/hda/hda_codec.c index 9913be8..8fd745c 100644 --- a/sound/pci/hda/hda_codec.c +++ b/sound/pci/hda/hda_codec.c @@ -311,9 +311,15 @@ int snd_hda_get_conn_index(struct hda_codec *codec, hda_nid_t mux, } EXPORT_SYMBOL_GPL(snd_hda_get_conn_index); - -/* return DEVLIST_LEN parameter of the given widget */ -static unsigned int get_num_devices(struct hda_codec *codec, hda_nid_t nid) +/** + * snd_hda_get_num_devices - get DEVLIST_LEN parameter of the given widget + * @codec: the HDA codec + * @nid: NID of the pin to parse + * + * Get the device entry number on the given widget. This is a feature of + * DP MST audio. Each pin can have several device entries in it. + */ +unsigned int snd_hda_get_num_devices(struct hda_codec *codec, hda_nid_t nid) { unsigned int wcaps = get_wcaps(codec, nid); unsigned int parm; @@ -327,6 +333,7 @@ static unsigned int get_num_devices(struct hda_codec *codec, hda_nid_t nid) parm = 0; return parm & AC_DEV_LIST_LEN_MASK; } +EXPORT_SYMBOL_GPL(snd_hda_get_num_devices); /** * snd_hda_get_devices - copy device list without cache @@ -344,7 +351,7 @@ int snd_hda_get_devices(struct hda_codec *codec, hda_nid_t nid, unsigned int parm; int i, dev_len, devices; - parm = get_num_devices(codec, nid); + parm = snd_hda_get_num_devices(codec, nid); if (!parm) /* not multi-stream capable */ return 0; @@ -368,6 +375,63 @@ int snd_hda_get_devices(struct hda_codec *codec, hda_nid_t nid, return devices; } +/** + * snd_hda_get_dev_select - get device entry select on the pin + * @codec: the HDA codec + * @nid: NID of the pin to get device entry select + * + * Get the devcie entry select on the pin. Return the device entry + * id selected on the pin. Return 0 means the first device entry + * is selected or MST is not supported. + */ +int snd_hda_get_dev_select(struct hda_codec *codec, hda_nid_t nid) +{ + /* not support dp_mst will always return 0, using first dev_entry */ + if (!codec->dp_mst) + return 0; + + return snd_hda_codec_read(codec, nid, 0, AC_VERB_GET_DEVICE_SEL, 0); +} +EXPORT_SYMBOL_GPL(snd_hda_get_dev_select); + +/** + * snd_hda_set_dev_select - set device entry select on the pin + * @codec: the HDA codec + * @nid: NID of the pin to set device entry select + * @dev_id: device entry id to be set + * + * Set the device entry select on the pin nid. + */ +int snd_hda_set_dev_select(struct hda_codec *codec, hda_nid_t nid, int dev_id) +{ + int ret, num_devices; + + /* not support dp_mst will always return 0, using first dev_entry */ + if (!codec->dp_mst) + return 0; + + /* AC_PAR_DEVLIST_LEN is 0 based. */ + num_devices = snd_hda_get_num_devices(codec, nid) + 1; + /* If Device List Length is 0 (num_device = 1), + * the pin is not multi stream capable. + * Do nothing in this case. + */ + if (num_devices == 1) + return 0; + + /* Behavior of setting index being equal to or greater than + * Device List Length is not predictable + */ + if (num_devices <= dev_id) + return -EINVAL; + + ret = snd_hda_codec_write(codec, nid, 0, + AC_VERB_SET_DEVICE_SEL, dev_id); + + return ret; +} +EXPORT_SYMBOL_GPL(snd_hda_set_dev_select); + /* * read widget caps for each widget and store in cache */ @@ -403,6 +467,10 @@ static int read_pin_defaults(struct hda_codec *codec) pin->nid = nid; pin->cfg = snd_hda_codec_read(codec, nid, 0, AC_VERB_GET_CONFIG_DEFAULT, 0); + /* + * all device entries are the same widget control so far + * fixme: if any codec is different, need fix here + */ pin->ctrl = snd_hda_codec_read(codec, nid, 0, AC_VERB_GET_PIN_WIDGET_CONTROL, 0); diff --git a/sound/pci/hda/hda_codec.h b/sound/pci/hda/hda_codec.h index 373fcad..f17f252 100644 --- a/sound/pci/hda/hda_codec.h +++ b/sound/pci/hda/hda_codec.h @@ -347,8 +347,11 @@ int snd_hda_override_conn_list(struct hda_codec *codec, hda_nid_t nid, int nums, const hda_nid_t *list); int snd_hda_get_conn_index(struct hda_codec *codec, hda_nid_t mux, hda_nid_t nid, int recursive); +unsigned int snd_hda_get_num_devices(struct hda_codec *codec, hda_nid_t nid); int snd_hda_get_devices(struct hda_codec *codec, hda_nid_t nid, u8 *dev_list, int max_devices); +int snd_hda_get_dev_select(struct hda_codec *codec, hda_nid_t nid); +int snd_hda_set_dev_select(struct hda_codec *codec, hda_nid_t nid, int dev_id); struct hda_verb { hda_nid_t nid; diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index cf9bc042..32105ce 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -76,6 +76,7 @@ struct hdmi_spec_per_cvt { struct hdmi_spec_per_pin { hda_nid_t pin_nid; + int dev_id; /* pin idx, different device entries on the same pin use the same idx */ int pin_nid_idx; int num_mux_nids; @@ -130,7 +131,23 @@ struct hdmi_spec { struct snd_array cvts; /* struct hdmi_spec_per_cvt */ hda_nid_t cvt_nids[4]; /* only for haswell fix */ + /* + * num_pins is the number of virtual pins + * for example, there are 3 pins, and each pin + * has 4 device entries, then the num_pins is 12 + */ int num_pins; + /* + * num_nids is the number of real pins + * In the above example, num_nids is 3 + */ + int num_nids; + /* + * dev_num is the number of device entries + * on each pin. + * In the above example, dev_num is 4 + */ + int dev_num; struct snd_array pins; /* struct hdmi_spec_per_pin */ struct hdmi_pcm pcm_rec[16]; struct mutex pcm_lock; @@ -217,14 +234,26 @@ union audio_infoframe { /* obtain hda_pcm object assigned to idx */ #define get_pcm_rec(spec, idx) (get_hdmi_pcm(spec, idx)->pcm) -static int pin_nid_to_pin_index(struct hda_codec *codec, hda_nid_t pin_nid) +static int pin_id_to_pin_index(struct hda_codec *codec, + hda_nid_t pin_nid, int dev_id) { struct hdmi_spec *spec = codec->spec; int pin_idx; + struct hdmi_spec_per_pin *per_pin; + + /* + * (dev_id == -1) means it is NON-MST pin + * return the first virtual pin on this port + */ + if (dev_id == -1) + dev_id = 0; - for (pin_idx = 0; pin_idx < spec->num_pins; pin_idx++) - if (get_pin(spec, pin_idx)->pin_nid == pin_nid) + for (pin_idx = 0; pin_idx < spec->num_pins; pin_idx++) { + per_pin = get_pin(spec, pin_idx); + if ((per_pin->pin_nid == pin_nid) && + (per_pin->dev_id == dev_id)) return pin_idx; + } codec_warn(codec, "HDMI: pin nid %d not registered\n", pin_nid); return -EINVAL; @@ -724,10 +753,11 @@ static void hdmi_setup_audio_infoframe(struct hda_codec *codec, static bool hdmi_present_sense(struct hdmi_spec_per_pin *per_pin, int repoll); -static void check_presence_and_report(struct hda_codec *codec, hda_nid_t nid) +static void check_presence_and_report(struct hda_codec *codec, hda_nid_t nid, + int dev_id) { struct hdmi_spec *spec = codec->spec; - int pin_idx = pin_nid_to_pin_index(codec, nid); + int pin_idx = pin_id_to_pin_index(codec, nid, dev_id); if (pin_idx < 0) return; @@ -738,7 +768,8 @@ static void check_presence_and_report(struct hda_codec *codec, hda_nid_t nid) static void jack_callback(struct hda_codec *codec, struct hda_jack_callback *jack) { - check_presence_and_report(codec, jack->nid); + /* hda_jack don't support DP MST */ + check_presence_and_report(codec, jack->nid, 0); } static void hdmi_intrinsic_event(struct hda_codec *codec, unsigned int res) @@ -747,6 +778,12 @@ static void hdmi_intrinsic_event(struct hda_codec *codec, unsigned int res) struct hda_jack_tbl *jack; int dev_entry = (res & AC_UNSOL_RES_DE) >> AC_UNSOL_RES_DE_SHIFT; + /* + * assume DP MST uses dyn_pcm_assign and acomp and + * never comes here + * if DP MST supports unsol event, below code need + * consider dev_entry + */ jack = snd_hda_jack_tbl_get_from_tag(codec, tag); if (!jack) return; @@ -757,7 +794,8 @@ static void hdmi_intrinsic_event(struct hda_codec *codec, unsigned int res) codec->addr, jack->nid, dev_entry, !!(res & AC_UNSOL_RES_IA), !!(res & AC_UNSOL_RES_PD), !!(res & AC_UNSOL_RES_ELDV)); - check_presence_and_report(codec, jack->nid); + /* hda_jack don't support DP MST */ + check_presence_and_report(codec, jack->nid, 0); } static void hdmi_non_intrinsic_event(struct hda_codec *codec, unsigned int res) @@ -970,28 +1008,60 @@ static int intel_cvt_id_to_mux_idx(struct hdmi_spec *spec, * by any other pins. */ static void intel_not_share_assigned_cvt(struct hda_codec *codec, - hda_nid_t pin_nid, int mux_idx) + hda_nid_t pin_nid, + int dev_id, int mux_idx) { struct hdmi_spec *spec = codec->spec; hda_nid_t nid; int cvt_idx, curr; struct hdmi_spec_per_cvt *per_cvt; + struct hdmi_spec_per_pin *per_pin; + int pin_idx; + + /* configure the pins connections */ + for (pin_idx = 0; pin_idx < spec->num_pins; pin_idx++) { + int dev_id_saved; + int dev_num; - /* configure all pins, including "no physical connection" ones */ - for_each_hda_codec_node(nid, codec) { - unsigned int wid_caps = get_wcaps(codec, nid); - unsigned int wid_type = get_wcaps_type(wid_caps); + per_pin = get_pin(spec, pin_idx); + /* + * pin not connected to monitor + * no need to operate on it + */ + if (!per_pin->pcm) + continue; - if (wid_type != AC_WID_PIN) + if ((per_pin->pin_nid == pin_nid) && + (per_pin->dev_id == dev_id)) continue; - if (nid == pin_nid) + /* + * if per_pin->dev_id >= dev_num, + * snd_hda_get_dev_select() will fail, + * and the following operation is unpredictable. + * So skip this situation. + */ + dev_num = snd_hda_get_num_devices(codec, per_pin->pin_nid) + 1; + if (per_pin->dev_id >= dev_num) continue; + nid = per_pin->pin_nid; + + /* + * Calling this function should not impact + * on the device entry selection + * So let's save the dev id for each pin, + * and restore it when return + */ + dev_id_saved = snd_hda_get_dev_select(codec, nid); + snd_hda_set_dev_select(codec, nid, per_pin->dev_id); curr = snd_hda_codec_read(codec, nid, 0, AC_VERB_GET_CONNECT_SEL, 0); - if (curr != mux_idx) + if (curr != mux_idx) { + snd_hda_set_dev_select(codec, nid, dev_id_saved); continue; + } + /* choose an unassigned converter. The conveters in the * connection list are in the same order as in the codec. @@ -1008,12 +1078,13 @@ static void intel_not_share_assigned_cvt(struct hda_codec *codec, break; } } + snd_hda_set_dev_select(codec, nid, dev_id_saved); } } /* A wrapper of intel_not_share_asigned_cvt() */ static void intel_not_share_assigned_cvt_nid(struct hda_codec *codec, - hda_nid_t pin_nid, hda_nid_t cvt_nid) + hda_nid_t pin_nid, int dev_id, hda_nid_t cvt_nid) { int mux_idx; struct hdmi_spec *spec = codec->spec; @@ -1025,7 +1096,7 @@ static void intel_not_share_assigned_cvt_nid(struct hda_codec *codec, */ mux_idx = intel_cvt_id_to_mux_idx(spec, cvt_nid); if (mux_idx >= 0) - intel_not_share_assigned_cvt(codec, pin_nid, mux_idx); + intel_not_share_assigned_cvt(codec, pin_nid, dev_id, mux_idx); } /* skeleton caller of pin_cvt_fixup ops */ @@ -1140,6 +1211,7 @@ static int hdmi_pcm_open(struct hda_pcm_stream *hinfo, per_pin->cvt_nid = per_cvt->cvt_nid; hinfo->nid = per_cvt->cvt_nid; + snd_hda_set_dev_select(codec, per_pin->pin_nid, per_pin->dev_id); snd_hda_codec_write_cache(codec, per_pin->pin_nid, 0, AC_VERB_SET_CONNECT_SEL, per_pin->mux_idx); @@ -1198,6 +1270,7 @@ static int hdmi_read_pin_conn(struct hda_codec *codec, int pin_idx) return -EINVAL; } + /* all the device entries on the same pin have the same conn list */ per_pin->num_mux_nids = snd_hda_get_connections(codec, pin_nid, per_pin->mux_nids, HDA_MAX_CONNECTIONS); @@ -1215,13 +1288,13 @@ static int hdmi_find_pcm_slot(struct hdmi_spec *spec, return per_pin->pin_nid_idx; /* have a second try; check the "reserved area" over num_pins */ - for (i = spec->num_pins; i < spec->pcm_used; i++) { + for (i = spec->num_nids; i < spec->pcm_used; i++) { if (!test_bit(i, &spec->pcm_bitmap)) return i; } /* the last try; check the empty slots in pins */ - for (i = 0; i < spec->num_pins; i++) { + for (i = 0; i < spec->num_nids; i++) { if (!test_bit(i, &spec->pcm_bitmap)) return i; } @@ -1296,10 +1369,13 @@ static void hdmi_pcm_setup_pin(struct hdmi_spec *spec, per_pin->cvt_nid = hinfo->nid; mux_idx = hdmi_get_pin_cvt_mux(spec, per_pin, hinfo->nid); - if (mux_idx < per_pin->num_mux_nids) + if (mux_idx < per_pin->num_mux_nids) { + snd_hda_set_dev_select(codec, per_pin->pin_nid, + per_pin->dev_id); snd_hda_codec_write_cache(codec, per_pin->pin_nid, 0, AC_VERB_SET_CONNECT_SEL, mux_idx); + } snd_hda_spdif_ctls_assign(codec, per_pin->pcm_idx, hinfo->nid); non_pcm = check_non_pcm_per_cvt(codec, hinfo->nid); @@ -1467,6 +1543,11 @@ static struct snd_jack *pin_idx_to_jack(struct hda_codec *codec, if (per_pin->pcm_idx >= 0 && spec->dyn_pcm_assign) jack = spec->pcm_rec[per_pin->pcm_idx].jack; else if (!spec->dyn_pcm_assign) { + /* + * jack tbl doesn't support DP MST + * DP MST will use dyn_pcm_assign, + * so DP MST will never come here + */ jack_tbl = snd_hda_jack_tbl_get(codec, per_pin->pin_nid); if (jack_tbl) jack = jack_tbl->jack; @@ -1485,9 +1566,9 @@ static void sync_eld_via_acomp(struct hda_codec *codec, mutex_lock(&per_pin->lock); eld->monitor_present = false; - size = snd_hdac_acomp_get_eld(&codec->core, per_pin->pin_nid, -1, - &eld->monitor_present, eld->eld_buffer, - ELD_MAX_SIZE); + size = snd_hdac_acomp_get_eld(&codec->core, per_pin->pin_nid, + per_pin->dev_id, &eld->monitor_present, + eld->eld_buffer, ELD_MAX_SIZE); if (size > 0) { size = min(size, ELD_MAX_SIZE); if (snd_hdmi_parse_eld(codec, &eld->info, @@ -1565,38 +1646,81 @@ static int hdmi_add_pin(struct hda_codec *codec, hda_nid_t pin_nid) int pin_idx; struct hdmi_spec_per_pin *per_pin; int err; + int dev_num, i; caps = snd_hda_query_pin_caps(codec, pin_nid); if (!(caps & (AC_PINCAP_HDMI | AC_PINCAP_DP))) return 0; + /* + * For DP MST audio, Configuration Default is the same for + * all device entries on the same pin + */ config = snd_hda_codec_get_pincfg(codec, pin_nid); if (get_defcfg_connect(config) == AC_JACK_PORT_NONE) return 0; - if (is_haswell_plus(codec)) - intel_haswell_fixup_connect_list(codec, pin_nid); - - pin_idx = spec->num_pins; - per_pin = snd_array_new(&spec->pins); - if (!per_pin) - return -ENOMEM; - - per_pin->pin_nid = pin_nid; - per_pin->non_pcm = false; - if (spec->dyn_pcm_assign) - per_pin->pcm_idx = -1; - else { - per_pin->pcm = get_hdmi_pcm(spec, pin_idx); - per_pin->pcm_idx = pin_idx; + /* + * To simplify the implementation, malloc all + * the virtual pins in the initialization statically + */ + if (is_haswell_plus(codec)) { + /* + * On Intel platforms, device entries number is + * changed dynamically. If there is a DP MST + * hub connected, the device entries number is 3. + * Otherwise, it is 1. + * Here we manually set dev_num to 3, so that + * we can initialize all the device entries when + * bootup statically. + */ + dev_num = 3; + spec->dev_num = 3; + } else if (spec->dyn_pcm_assign && codec->dp_mst) { + dev_num = snd_hda_get_num_devices(codec, pin_nid) + 1; + /* + * spec->dev_num is the maxinum number of device entries + * among all the pins + */ + spec->dev_num = (spec->dev_num > dev_num) ? + spec->dev_num : dev_num; + } else { + /* + * If the platform doesn't support DP MST, + * manually set dev_num to 1. This means + * the pin has only one device entry. + */ + dev_num = 1; + spec->dev_num = 1; } - per_pin->pin_nid_idx = pin_idx; - err = hdmi_read_pin_conn(codec, pin_idx); - if (err < 0) - return err; + for (i = 0; i < dev_num; i++) { + pin_idx = spec->num_pins; + per_pin = snd_array_new(&spec->pins); - spec->num_pins++; + if (!per_pin) + return -ENOMEM; + + if (spec->dyn_pcm_assign) { + per_pin->pcm = NULL; + per_pin->pcm_idx = -1; + } else { + per_pin->pcm = get_hdmi_pcm(spec, pin_idx); + per_pin->pcm_idx = pin_idx; + } + per_pin->pin_nid = pin_nid; + per_pin->pin_nid_idx = spec->num_nids; + per_pin->dev_id = i; + per_pin->non_pcm = false; + snd_hda_set_dev_select(codec, pin_nid, i); + if (is_haswell_plus(codec)) + intel_haswell_fixup_connect_list(codec, pin_nid); + err = hdmi_read_pin_conn(codec, pin_idx); + if (err < 0) + return err; + spec->num_pins++; + } + spec->num_nids++; return 0; } @@ -1744,7 +1868,7 @@ static int generic_hdmi_playback_pcm_prepare(struct hda_pcm_stream *hinfo, /* Call sync_audio_rate to set the N/CTS/M manually if necessary */ /* Todo: add DP1.2 MST audio support later */ if (codec_has_acomp(codec)) - snd_hdac_sync_audio_rate(&codec->core, pin_nid, -1, + snd_hdac_sync_audio_rate(&codec->core, pin_nid, per_pin->dev_id, runtime->rate); non_pcm = check_non_pcm_per_cvt(codec, cvt_nid); @@ -1762,6 +1886,7 @@ static int generic_hdmi_playback_pcm_prepare(struct hda_pcm_stream *hinfo, pinctl | PIN_OUT); } + /* snd_hda_set_dev_select() has been called before */ err = spec->ops.setup_stream(codec, cvt_nid, pin_nid, stream_tag, format); mutex_unlock(&spec->pcm_lock); @@ -1897,17 +2022,23 @@ static bool is_hdmi_pcm_attached(struct hdac_device *hdac, int pcm_idx) static int generic_hdmi_build_pcms(struct hda_codec *codec) { struct hdmi_spec *spec = codec->spec; - int pin_idx; + int idx; - for (pin_idx = 0; pin_idx < spec->num_pins; pin_idx++) { + /* + * for non-mst mode, pcm number is the same as before + * for DP MST mode, pcm number is (nid number + dev_num - 1) + * dev_num is the device entry number in a pin + * + */ + for (idx = 0; idx < spec->num_nids + spec->dev_num - 1; idx++) { struct hda_pcm *info; struct hda_pcm_stream *pstr; - info = snd_hda_codec_pcm_new(codec, "HDMI %d", pin_idx); + info = snd_hda_codec_pcm_new(codec, "HDMI %d", idx); if (!info) return -ENOMEM; - spec->pcm_rec[pin_idx].pcm = info; + spec->pcm_rec[idx].pcm = info; spec->pcm_used++; info->pcm_type = HDA_PCM_TYPE_HDMI; info->own_chmap = true; @@ -1915,6 +2046,9 @@ static int generic_hdmi_build_pcms(struct hda_codec *codec) pstr = &info->stream[SNDRV_PCM_STREAM_PLAYBACK]; pstr->substreams = 1; pstr->ops = generic_ops; + /* pcm number is less than 16 */ + if (spec->pcm_used >= 16) + break; /* other pstr fields are set in open */ } @@ -2070,7 +2204,9 @@ static int generic_hdmi_init(struct hda_codec *codec) for (pin_idx = 0; pin_idx < spec->num_pins; pin_idx++) { struct hdmi_spec_per_pin *per_pin = get_pin(spec, pin_idx); hda_nid_t pin_nid = per_pin->pin_nid; + int dev_id = per_pin->dev_id; + snd_hda_set_dev_select(codec, pin_nid, dev_id); hdmi_init_pin(codec, pin_nid); if (!codec_has_acomp(codec)) snd_hda_jack_detect_enable_callback(codec, pin_nid, @@ -2178,6 +2314,7 @@ static int alloc_generic_hdmi(struct hda_codec *codec) return -ENOMEM; spec->ops = generic_standard_hdmi_ops; + spec->dev_num = 1; /* initialize to 1 */ mutex_init(&spec->pcm_lock); snd_hdac_register_chmap_ops(&codec->core, &spec->chmap); @@ -2295,6 +2432,7 @@ static void intel_pin_eld_notify(void *audio_ptr, int port, int pipe) { struct hda_codec *codec = audio_ptr; int pin_nid; + int dev_id = pipe; /* we assume only from port-B to port-D */ if (port < 1 || port > 3) @@ -2321,7 +2459,7 @@ static void intel_pin_eld_notify(void *audio_ptr, int port, int pipe) return; snd_hdac_i915_set_bclk(&codec->bus->core); - check_presence_and_report(codec, pin_nid); + check_presence_and_report(codec, pin_nid, dev_id); } /* register i915 component pin_eld_notify callback */ @@ -2354,11 +2492,13 @@ static void i915_pin_cvt_fixup(struct hda_codec *codec, hda_nid_t cvt_nid) { if (per_pin) { + snd_hda_set_dev_select(codec, per_pin->pin_nid, + per_pin->dev_id); intel_verify_pin_cvt_connect(codec, per_pin); intel_not_share_assigned_cvt(codec, per_pin->pin_nid, - per_pin->mux_idx); + per_pin->dev_id, per_pin->mux_idx); } else { - intel_not_share_assigned_cvt_nid(codec, 0, cvt_nid); + intel_not_share_assigned_cvt_nid(codec, 0, 0, cvt_nid); } } @@ -2378,6 +2518,8 @@ static int patch_i915_hsw_hdmi(struct hda_codec *codec) if (err < 0) return err; spec = codec->spec; + codec->dp_mst = true; + spec->dyn_pcm_assign = true; intel_haswell_enable_all_pins(codec, true); intel_haswell_fixup_enable_dp12(codec); @@ -2389,7 +2531,6 @@ static int patch_i915_hsw_hdmi(struct hda_codec *codec) codec->core.link_power_control = 1; codec->patch_ops.set_power_state = haswell_set_power_state; - codec->dp_mst = true; codec->depop_delay = 0; codec->auto_runtime_pm = 1; |