diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2018-06-06 08:16:33 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2018-06-06 08:16:33 -0700 |
commit | 135c5504a600ff9b06e321694fbcac78a9530cd4 (patch) | |
tree | 8d22ed739b0e85954010a964a9aeadf3c692c977 /drivers/gpu/drm/i915 | |
parent | af6c5d5e01ad9f2c9ca38cccaae6b5d67ddd241f (diff) | |
parent | 568cf2e6aa0c762f14d2d0d481a006f93c63ab7a (diff) | |
download | op-kernel-dev-135c5504a600ff9b06e321694fbcac78a9530cd4.zip op-kernel-dev-135c5504a600ff9b06e321694fbcac78a9530cd4.tar.gz |
Merge tag 'drm-next-2018-06-06-1' of git://anongit.freedesktop.org/drm/drm
Pull drm updates from Dave Airlie:
"This starts to support NVIDIA volta hardware with nouveau, and adds
amdgpu support for the GPU in the Kabylake-G (the intel + radeon
single package chip), along with some initial Intel icelake enabling.
Summary:
New Drivers:
- v3d - driver for broadcom V3D V3.x+ hardware
- xen-front - XEN PV display frontend
core:
- handle zpos normalization in the core
- stop looking at legacy pointers in atomic paths
- improved scheduler documentation
- improved aspect ratio validation
- aspect ratio support for 64:27 and 256:135
- drop unused control node code.
i915:
- Icelake (ICL) enabling
- GuC/HuC refactoring
- PSR/PSR2 enabling and fixes
- DPLL management refactoring
- DP MST fixes
- NV12 enabling
- HDCP improvements
- GEM/Execlist/reset improvements
- GVT improvements
- stolen memory first 4k fix
amdgpu:
- Vega 20 support
- VEGAM support (Kabylake-G)
- preOS scanout buffer reservation
- power management gfxoff support for raven
- SR-IOV fixes
- Vega10 power profiles and clock voltage control
- scatter/gather display support on CZ/ST
amdkfd:
- GFX9 dGPU support
- userptr memory mapping
nouveau:
- major refactoring for Volta GV100 support
tda998x:
- HDMI i2c CEC support
etnaviv:
- removed unused logging code
- license text cleanups
- MMU handling improvements
- timeout fence fix for 50 days uptime
tegra:
- IOMMU support in gr2d/gr3d drivers
- zpos support
vc4:
- syncobj support
- CTM, plane alpha and async cursor support
analogix_dp:
- HPD and aux chan fixes
sun4i:
- MIPI DSI support
tilcdc:
- clock divider fixes for OMAP-l138 LCDK board
rcar-du:
- R8A77965 support
- dma-buf fences fixes
- hardware indexed crtc/du group handling
- generic zplane property support
atmel-hclcdc:
- generic zplane property support
mediatek:
- use generic video mode function
exynos:
- S5PV210 FIMD variant support
- IPP v2 framework
- more HW overlays support"
* tag 'drm-next-2018-06-06-1' of git://anongit.freedesktop.org/drm/drm: (1286 commits)
drm/amdgpu: fix 32-bit build warning
drm/exynos: fimc: signedness bug in fimc_setup_clocks()
drm/exynos: scaler: fix static checker warning
drm/amdgpu: Use dev_info() to report amdkfd is not supported for this ASIC
drm/amd/display: Remove use of division operator for long longs
drm/amdgpu: Update GFX info structure to match what vega20 used
drm/amdgpu/pp: remove duplicate assignment
drm/sched: add rcu_barrier after entity fini
drm/amdgpu: move VM BOs on LRU again
drm/amdgpu: consistenly use VM moved flag
drm/amdgpu: kmap PDs/PTs in amdgpu_vm_update_directories
drm/amdgpu: further optimize amdgpu_vm_handle_moved
drm/amdgpu: cleanup amdgpu_vm_validate_pt_bos v2
drm/amdgpu: rework VM state machine lock handling v2
drm/amdgpu: Add runtime VCN PG support
drm/amdgpu: Enable VCN static PG by default on RV
drm/amdgpu: Add VCN static PG support on RV
drm/amdgpu: Enable VCN CG by default on RV
drm/amdgpu: Add static CG control for VCN on RV
drm/exynos: Fix default value for zpos plane property
...
Diffstat (limited to 'drivers/gpu/drm/i915')
129 files changed, 11316 insertions, 5392 deletions
diff --git a/drivers/gpu/drm/i915/Kconfig.debug b/drivers/gpu/drm/i915/Kconfig.debug index 108d21f..9de8b1c 100644 --- a/drivers/gpu/drm/i915/Kconfig.debug +++ b/drivers/gpu/drm/i915/Kconfig.debug @@ -25,7 +25,8 @@ config DRM_I915_DEBUG select X86_MSR # used by igt/pm_rpm select DRM_VGEM # used by igt/prime_vgem (dmabuf interop checks) select DRM_DEBUG_MM if DRM=y - select DRM_DEBUG_MM_SELFTEST + select STACKDEPOT if DRM=y # for DRM_DEBUG_MM + select DRM_DEBUG_SELFTEST select SW_SYNC # signaling validation framework (igt/syncobj*) select DRM_I915_SW_FENCE_DEBUG_OBJECTS select DRM_I915_SELFTEST @@ -89,6 +90,18 @@ config DRM_I915_SW_FENCE_CHECK_DAG If in doubt, say "N". +config DRM_I915_DEBUG_GUC + bool "Enable additional driver debugging for GuC" + depends on DRM_I915 + default n + help + Choose this option to turn on extra driver debugging that may affect + performance but will help resolve GuC related issues. + + Recommended for driver developers only. + + If in doubt, say "N". + config DRM_I915_SELFTEST bool "Enable selftests upon driver load" depends on DRM_I915 diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile index 4eee91a..4c6adae 100644 --- a/drivers/gpu/drm/i915/Makefile +++ b/drivers/gpu/drm/i915/Makefile @@ -12,12 +12,16 @@ # Note the danger in using -Wall -Wextra is that when CI updates gcc we # will most likely get a sudden build breakage... Hopefully we will fix # new warnings before CI updates! -subdir-ccflags-y := -Wall -Wextra +subdir-ccflags-y := -Wall -Wextra -Wvla subdir-ccflags-y += $(call cc-disable-warning, unused-parameter) subdir-ccflags-y += $(call cc-disable-warning, type-limits) subdir-ccflags-y += $(call cc-disable-warning, missing-field-initializers) subdir-ccflags-y += $(call cc-disable-warning, implicit-fallthrough) subdir-ccflags-y += $(call cc-disable-warning, unused-but-set-variable) +# clang warnings +subdir-ccflags-y += $(call cc-disable-warning, sign-compare) +subdir-ccflags-y += $(call cc-disable-warning, sometimes-uninitialized) +subdir-ccflags-y += $(call cc-disable-warning, initializer-overrides) subdir-ccflags-$(CONFIG_DRM_I915_WERROR) += -Werror # Fine grained warnings disable @@ -43,7 +47,8 @@ i915-y := i915_drv.o \ intel_csr.o \ intel_device_info.o \ intel_pm.o \ - intel_runtime_pm.o + intel_runtime_pm.o \ + intel_workarounds.o i915-$(CONFIG_COMPAT) += i915_ioc32.o i915-$(CONFIG_DEBUG_FS) += i915_debugfs.o intel_pipe_crc.o @@ -66,11 +71,11 @@ i915-y += i915_cmd_parser.o \ i915_gem_shrinker.o \ i915_gem_stolen.o \ i915_gem_tiling.o \ - i915_gem_timeline.o \ i915_gem_userptr.o \ i915_gemfs.o \ i915_query.o \ i915_request.o \ + i915_timeline.o \ i915_trace_points.o \ i915_vma.o \ intel_breadcrumbs.o \ @@ -79,7 +84,8 @@ i915-y += i915_cmd_parser.o \ intel_lrc.o \ intel_mocs.o \ intel_ringbuffer.o \ - intel_uncore.o + intel_uncore.o \ + intel_wopcm.o # general-purpose microcontroller (GuC) support i915-y += intel_uc.o \ @@ -152,7 +158,8 @@ i915-y += dvo_ch7017.o \ i915-$(CONFIG_DRM_I915_CAPTURE_ERROR) += i915_gpu_error.o i915-$(CONFIG_DRM_I915_SELFTEST) += \ selftests/i915_random.o \ - selftests/i915_selftest.o + selftests/i915_selftest.o \ + selftests/igt_flush_test.o # virtual gpu code i915-y += i915_vgpu.o @@ -171,7 +178,8 @@ i915-y += i915_perf.o \ i915_oa_glk.o \ i915_oa_cflgt2.o \ i915_oa_cflgt3.o \ - i915_oa_cnl.o + i915_oa_cnl.o \ + i915_oa_icl.o ifeq ($(CONFIG_DRM_I915_GVT),y) i915-y += intel_gvt.o diff --git a/drivers/gpu/drm/i915/gvt/cmd_parser.c b/drivers/gpu/drm/i915/gvt/cmd_parser.c index d85939b..718ca08 100644 --- a/drivers/gpu/drm/i915/gvt/cmd_parser.c +++ b/drivers/gpu/drm/i915/gvt/cmd_parser.c @@ -813,15 +813,31 @@ static inline bool is_force_nonpriv_mmio(unsigned int offset) } static int force_nonpriv_reg_handler(struct parser_exec_state *s, - unsigned int offset, unsigned int index) + unsigned int offset, unsigned int index, char *cmd) { struct intel_gvt *gvt = s->vgpu->gvt; - unsigned int data = cmd_val(s, index + 1); + unsigned int data; + u32 ring_base; + u32 nopid; + struct drm_i915_private *dev_priv = s->vgpu->gvt->dev_priv; + + if (!strcmp(cmd, "lri")) + data = cmd_val(s, index + 1); + else { + gvt_err("Unexpected forcenonpriv 0x%x write from cmd %s\n", + offset, cmd); + return -EINVAL; + } + + ring_base = dev_priv->engine[s->ring_id]->mmio_base; + nopid = i915_mmio_reg_offset(RING_NOPID(ring_base)); - if (!intel_gvt_in_force_nonpriv_whitelist(gvt, data)) { + if (!intel_gvt_in_force_nonpriv_whitelist(gvt, data) && + data != nopid) { gvt_err("Unexpected forcenonpriv 0x%x LRI write, value=0x%x\n", offset, data); - return -EPERM; + patch_value(s, cmd_ptr(s, index), nopid); + return 0; } return 0; } @@ -869,7 +885,7 @@ static int cmd_reg_handler(struct parser_exec_state *s, return -EINVAL; if (is_force_nonpriv_mmio(offset) && - force_nonpriv_reg_handler(s, offset, index)) + force_nonpriv_reg_handler(s, offset, index, cmd)) return -EPERM; if (offset == i915_mmio_reg_offset(DERRMR) || @@ -1604,7 +1620,8 @@ static int batch_buffer_needs_scan(struct parser_exec_state *s) if (IS_BROADWELL(gvt->dev_priv) || IS_SKYLAKE(gvt->dev_priv) || IS_KABYLAKE(gvt->dev_priv)) { /* BDW decides privilege based on address space */ - if (cmd_val(s, 0) & (1 << 8)) + if (cmd_val(s, 0) & (1 << 8) && + !(s->vgpu->scan_nonprivbb & (1 << s->ring_id))) return 0; } return 1; @@ -1618,6 +1635,8 @@ static int find_bb_size(struct parser_exec_state *s, unsigned long *bb_size) bool bb_end = false; struct intel_vgpu *vgpu = s->vgpu; u32 cmd; + struct intel_vgpu_mm *mm = (s->buf_addr_type == GTT_BUFFER) ? + s->vgpu->gtt.ggtt_mm : s->workload->shadow_mm; *bb_size = 0; @@ -1629,18 +1648,22 @@ static int find_bb_size(struct parser_exec_state *s, unsigned long *bb_size) cmd = cmd_val(s, 0); info = get_cmd_info(s->vgpu->gvt, cmd, s->ring_id); if (info == NULL) { - gvt_vgpu_err("unknown cmd 0x%x, opcode=0x%x\n", - cmd, get_opcode(cmd, s->ring_id)); + gvt_vgpu_err("unknown cmd 0x%x, opcode=0x%x, addr_type=%s, ring %d, workload=%p\n", + cmd, get_opcode(cmd, s->ring_id), + (s->buf_addr_type == PPGTT_BUFFER) ? + "ppgtt" : "ggtt", s->ring_id, s->workload); return -EBADRQC; } do { - if (copy_gma_to_hva(s->vgpu, s->vgpu->gtt.ggtt_mm, + if (copy_gma_to_hva(s->vgpu, mm, gma, gma + 4, &cmd) < 0) return -EFAULT; info = get_cmd_info(s->vgpu->gvt, cmd, s->ring_id); if (info == NULL) { - gvt_vgpu_err("unknown cmd 0x%x, opcode=0x%x\n", - cmd, get_opcode(cmd, s->ring_id)); + gvt_vgpu_err("unknown cmd 0x%x, opcode=0x%x, addr_type=%s, ring %d, workload=%p\n", + cmd, get_opcode(cmd, s->ring_id), + (s->buf_addr_type == PPGTT_BUFFER) ? + "ppgtt" : "ggtt", s->ring_id, s->workload); return -EBADRQC; } @@ -1666,6 +1689,9 @@ static int perform_bb_shadow(struct parser_exec_state *s) unsigned long gma = 0; unsigned long bb_size; int ret = 0; + struct intel_vgpu_mm *mm = (s->buf_addr_type == GTT_BUFFER) ? + s->vgpu->gtt.ggtt_mm : s->workload->shadow_mm; + unsigned long gma_start_offset = 0; /* get the start gm address of the batch buffer */ gma = get_gma_bb_from_cmd(s, 1); @@ -1680,8 +1706,24 @@ static int perform_bb_shadow(struct parser_exec_state *s) if (!bb) return -ENOMEM; + bb->ppgtt = (s->buf_addr_type == GTT_BUFFER) ? false : true; + + /* the gma_start_offset stores the batch buffer's start gma's + * offset relative to page boundary. so for non-privileged batch + * buffer, the shadowed gem object holds exactly the same page + * layout as original gem object. This is for the convience of + * replacing the whole non-privilged batch buffer page to this + * shadowed one in PPGTT at the same gma address. (this replacing + * action is not implemented yet now, but may be necessary in + * future). + * for prileged batch buffer, we just change start gma address to + * that of shadowed page. + */ + if (bb->ppgtt) + gma_start_offset = gma & ~I915_GTT_PAGE_MASK; + bb->obj = i915_gem_object_create(s->vgpu->gvt->dev_priv, - roundup(bb_size, PAGE_SIZE)); + roundup(bb_size + gma_start_offset, PAGE_SIZE)); if (IS_ERR(bb->obj)) { ret = PTR_ERR(bb->obj); goto err_free_bb; @@ -1702,9 +1744,9 @@ static int perform_bb_shadow(struct parser_exec_state *s) bb->clflush &= ~CLFLUSH_BEFORE; } - ret = copy_gma_to_hva(s->vgpu, s->vgpu->gtt.ggtt_mm, + ret = copy_gma_to_hva(s->vgpu, mm, gma, gma + bb_size, - bb->va); + bb->va + gma_start_offset); if (ret < 0) { gvt_vgpu_err("fail to copy guest ring buffer\n"); ret = -EFAULT; @@ -1730,7 +1772,7 @@ static int perform_bb_shadow(struct parser_exec_state *s) * buffer's gma in pair. After all, we don't want to pin the shadow * buffer here (too early). */ - s->ip_va = bb->va; + s->ip_va = bb->va + gma_start_offset; s->ip_gma = gma; return 0; err_unmap: @@ -2469,15 +2511,18 @@ static int cmd_parser_exec(struct parser_exec_state *s) info = get_cmd_info(s->vgpu->gvt, cmd, s->ring_id); if (info == NULL) { - gvt_vgpu_err("unknown cmd 0x%x, opcode=0x%x\n", - cmd, get_opcode(cmd, s->ring_id)); + gvt_vgpu_err("unknown cmd 0x%x, opcode=0x%x, addr_type=%s, ring %d, workload=%p\n", + cmd, get_opcode(cmd, s->ring_id), + (s->buf_addr_type == PPGTT_BUFFER) ? + "ppgtt" : "ggtt", s->ring_id, s->workload); return -EBADRQC; } s->info = info; trace_gvt_command(vgpu->id, s->ring_id, s->ip_gma, s->ip_va, - cmd_length(s), s->buf_type); + cmd_length(s), s->buf_type, s->buf_addr_type, + s->workload, info->name); if (info->handler) { ret = info->handler(s); diff --git a/drivers/gpu/drm/i915/gvt/debugfs.c b/drivers/gpu/drm/i915/gvt/debugfs.c index 32a66df..2ec89bc 100644 --- a/drivers/gpu/drm/i915/gvt/debugfs.c +++ b/drivers/gpu/drm/i915/gvt/debugfs.c @@ -122,18 +122,69 @@ static int vgpu_mmio_diff_show(struct seq_file *s, void *unused) seq_printf(s, "Total: %d, Diff: %d\n", param.total, param.diff); return 0; } +DEFINE_SHOW_ATTRIBUTE(vgpu_mmio_diff); -static int vgpu_mmio_diff_open(struct inode *inode, struct file *file) +static int +vgpu_scan_nonprivbb_get(void *data, u64 *val) { - return single_open(file, vgpu_mmio_diff_show, inode->i_private); + struct intel_vgpu *vgpu = (struct intel_vgpu *)data; + *val = vgpu->scan_nonprivbb; + return 0; } -static const struct file_operations vgpu_mmio_diff_fops = { - .open = vgpu_mmio_diff_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release, -}; +/* + * set/unset bit engine_id of vgpu->scan_nonprivbb to turn on/off scanning + * of non-privileged batch buffer. e.g. + * if vgpu->scan_nonprivbb=3, then it will scan non-privileged batch buffer + * on engine 0 and 1. + */ +static int +vgpu_scan_nonprivbb_set(void *data, u64 val) +{ + struct intel_vgpu *vgpu = (struct intel_vgpu *)data; + struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv; + enum intel_engine_id id; + char buf[128], *s; + int len; + + val &= (1 << I915_NUM_ENGINES) - 1; + + if (vgpu->scan_nonprivbb == val) + return 0; + + if (!val) + goto done; + + len = sprintf(buf, + "gvt: vgpu %d turns on non-privileged batch buffers scanning on Engines:", + vgpu->id); + + s = buf + len; + + for (id = 0; id < I915_NUM_ENGINES; id++) { + struct intel_engine_cs *engine; + + engine = dev_priv->engine[id]; + if (engine && (val & (1 << id))) { + len = snprintf(s, 4, "%d, ", engine->id); + s += len; + } else + val &= ~(1 << id); + } + + if (val) + sprintf(s, "low performance expected."); + + pr_warn("%s\n", buf); + +done: + vgpu->scan_nonprivbb = val; + return 0; +} + +DEFINE_SIMPLE_ATTRIBUTE(vgpu_scan_nonprivbb_fops, + vgpu_scan_nonprivbb_get, vgpu_scan_nonprivbb_set, + "0x%llx\n"); /** * intel_gvt_debugfs_add_vgpu - register debugfs entries for a vGPU @@ -162,6 +213,11 @@ int intel_gvt_debugfs_add_vgpu(struct intel_vgpu *vgpu) if (!ent) return -ENOMEM; + ent = debugfs_create_file("scan_nonprivbb", 0644, vgpu->debugfs, + vgpu, &vgpu_scan_nonprivbb_fops); + if (!ent) + return -ENOMEM; + return 0; } diff --git a/drivers/gpu/drm/i915/gvt/gvt.h b/drivers/gpu/drm/i915/gvt/gvt.h index efacd8a..05d15a0 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.h +++ b/drivers/gpu/drm/i915/gvt/gvt.h @@ -99,7 +99,6 @@ struct intel_vgpu_fence { struct intel_vgpu_mmio { void *vreg; void *sreg; - bool disable_warn_untrack; }; #define INTEL_GVT_MAX_BAR_NUM 4 @@ -226,6 +225,7 @@ struct intel_vgpu { struct completion vblank_done; + u32 scan_nonprivbb; }; /* validating GM healthy status*/ diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index a33c1c3e..4b6532f 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -191,6 +191,8 @@ static int sanitize_fence_mmio_access(struct intel_vgpu *vgpu, unsigned int max_fence = vgpu_fence_sz(vgpu); if (fence_num >= max_fence) { + gvt_vgpu_err("access oob fence reg %d/%d\n", + fence_num, max_fence); /* When guest access oob fence regs without access * pv_info first, we treat guest not supporting GVT, @@ -200,11 +202,6 @@ static int sanitize_fence_mmio_access(struct intel_vgpu *vgpu, enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST); - if (!vgpu->mmio.disable_warn_untrack) { - gvt_vgpu_err("found oob fence register access\n"); - gvt_vgpu_err("total fence %d, access fence %d\n", - max_fence, fence_num); - } memset(p_data, 0, bytes); return -EINVAL; } @@ -477,22 +474,28 @@ static int force_nonpriv_write(struct intel_vgpu *vgpu, unsigned int offset, void *p_data, unsigned int bytes) { u32 reg_nonpriv = *(u32 *)p_data; + int ring_id = intel_gvt_render_mmio_to_ring_id(vgpu->gvt, offset); + u32 ring_base; + struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv; int ret = -EINVAL; - if ((bytes != 4) || ((offset & (bytes - 1)) != 0)) { - gvt_err("vgpu(%d) Invalid FORCE_NONPRIV offset %x(%dB)\n", - vgpu->id, offset, bytes); + if ((bytes != 4) || ((offset & (bytes - 1)) != 0) || ring_id < 0) { + gvt_err("vgpu(%d) ring %d Invalid FORCE_NONPRIV offset %x(%dB)\n", + vgpu->id, ring_id, offset, bytes); return ret; } - if (in_whitelist(reg_nonpriv)) { + ring_base = dev_priv->engine[ring_id]->mmio_base; + + if (in_whitelist(reg_nonpriv) || + reg_nonpriv == i915_mmio_reg_offset(RING_NOPID(ring_base))) { ret = intel_vgpu_default_mmio_write(vgpu, offset, p_data, bytes); - } else { - gvt_err("vgpu(%d) Invalid FORCE_NONPRIV write %x\n", - vgpu->id, reg_nonpriv); - } - return ret; + } else + gvt_err("vgpu(%d) Invalid FORCE_NONPRIV write %x at offset %x\n", + vgpu->id, reg_nonpriv, offset); + + return 0; } static int ddi_buf_ctl_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, @@ -3092,9 +3095,7 @@ int intel_vgpu_mmio_reg_rw(struct intel_vgpu *vgpu, unsigned int offset, */ mmio_info = find_mmio_info(gvt, offset); if (!mmio_info) { - if (!vgpu->mmio.disable_warn_untrack) - gvt_vgpu_err("untracked MMIO %08x len %d\n", - offset, bytes); + gvt_dbg_mmio("untracked MMIO %08x len %d\n", offset, bytes); goto default_rw; } diff --git a/drivers/gpu/drm/i915/gvt/mmio.c b/drivers/gpu/drm/i915/gvt/mmio.c index 11b71b3..e4960af 100644 --- a/drivers/gpu/drm/i915/gvt/mmio.c +++ b/drivers/gpu/drm/i915/gvt/mmio.c @@ -244,8 +244,6 @@ void intel_vgpu_reset_mmio(struct intel_vgpu *vgpu, bool dmlr) /* set the bit 0:2(Core C-State ) to C0 */ vgpu_vreg_t(vgpu, GEN6_GT_CORE_STATUS) = 0; - - vgpu->mmio.disable_warn_untrack = false; } else { #define GVT_GEN8_MMIO_RESET_OFFSET (0x44200) /* only reset the engine related, so starting with 0x44200 diff --git a/drivers/gpu/drm/i915/gvt/mmio_context.c b/drivers/gpu/drm/i915/gvt/mmio_context.c index a5bac83..0f94955 100644 --- a/drivers/gpu/drm/i915/gvt/mmio_context.c +++ b/drivers/gpu/drm/i915/gvt/mmio_context.c @@ -448,7 +448,7 @@ static void switch_mocs(struct intel_vgpu *pre, struct intel_vgpu *next, bool is_inhibit_context(struct i915_gem_context *ctx, int ring_id) { - u32 *reg_state = ctx->engine[ring_id].lrc_reg_state; + u32 *reg_state = ctx->__engine[ring_id].lrc_reg_state; u32 inhibit_mask = _MASKED_BIT_ENABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT); diff --git a/drivers/gpu/drm/i915/gvt/sched_policy.c b/drivers/gpu/drm/i915/gvt/sched_policy.c index 75b7bc7..d053cbe 100644 --- a/drivers/gpu/drm/i915/gvt/sched_policy.c +++ b/drivers/gpu/drm/i915/gvt/sched_policy.c @@ -53,7 +53,6 @@ struct vgpu_sched_data { bool active; ktime_t sched_in_time; - ktime_t sched_out_time; ktime_t sched_time; ktime_t left_ts; ktime_t allocated_ts; @@ -66,17 +65,22 @@ struct gvt_sched_data { struct hrtimer timer; unsigned long period; struct list_head lru_runq_head; + ktime_t expire_time; }; -static void vgpu_update_timeslice(struct intel_vgpu *pre_vgpu) +static void vgpu_update_timeslice(struct intel_vgpu *vgpu, ktime_t cur_time) { ktime_t delta_ts; - struct vgpu_sched_data *vgpu_data = pre_vgpu->sched_data; + struct vgpu_sched_data *vgpu_data; - delta_ts = vgpu_data->sched_out_time - vgpu_data->sched_in_time; + if (!vgpu || vgpu == vgpu->gvt->idle_vgpu) + return; - vgpu_data->sched_time += delta_ts; - vgpu_data->left_ts -= delta_ts; + vgpu_data = vgpu->sched_data; + delta_ts = ktime_sub(cur_time, vgpu_data->sched_in_time); + vgpu_data->sched_time = ktime_add(vgpu_data->sched_time, delta_ts); + vgpu_data->left_ts = ktime_sub(vgpu_data->left_ts, delta_ts); + vgpu_data->sched_in_time = cur_time; } #define GVT_TS_BALANCE_PERIOD_MS 100 @@ -150,11 +154,7 @@ static void try_to_schedule_next_vgpu(struct intel_gvt *gvt) } cur_time = ktime_get(); - if (scheduler->current_vgpu) { - vgpu_data = scheduler->current_vgpu->sched_data; - vgpu_data->sched_out_time = cur_time; - vgpu_update_timeslice(scheduler->current_vgpu); - } + vgpu_update_timeslice(scheduler->current_vgpu, cur_time); vgpu_data = scheduler->next_vgpu->sched_data; vgpu_data->sched_in_time = cur_time; @@ -226,17 +226,22 @@ out: void intel_gvt_schedule(struct intel_gvt *gvt) { struct gvt_sched_data *sched_data = gvt->scheduler.sched_data; - static uint64_t timer_check; + ktime_t cur_time; mutex_lock(&gvt->lock); + cur_time = ktime_get(); if (test_and_clear_bit(INTEL_GVT_REQUEST_SCHED, (void *)&gvt->service_request)) { - if (!(timer_check++ % GVT_TS_BALANCE_PERIOD_MS)) + if (cur_time >= sched_data->expire_time) { gvt_balance_timeslice(sched_data); + sched_data->expire_time = ktime_add_ms( + cur_time, GVT_TS_BALANCE_PERIOD_MS); + } } clear_bit(INTEL_GVT_REQUEST_EVENT_SCHED, (void *)&gvt->service_request); + vgpu_update_timeslice(gvt->scheduler.current_vgpu, cur_time); tbs_sched_func(sched_data); mutex_unlock(&gvt->lock); diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c index 638abe8..c2d183b 100644 --- a/drivers/gpu/drm/i915/gvt/scheduler.c +++ b/drivers/gpu/drm/i915/gvt/scheduler.c @@ -58,7 +58,7 @@ static void update_shadow_pdps(struct intel_vgpu_workload *workload) int ring_id = workload->ring_id; struct i915_gem_context *shadow_ctx = vgpu->submission.shadow_ctx; struct drm_i915_gem_object *ctx_obj = - shadow_ctx->engine[ring_id].state->obj; + shadow_ctx->__engine[ring_id].state->obj; struct execlist_ring_context *shadow_ring_context; struct page *page; @@ -97,7 +97,7 @@ static void sr_oa_regs(struct intel_vgpu_workload *workload, i915_mmio_reg_offset(EU_PERF_CNTL6), }; - if (!workload || !reg_state || workload->ring_id != RCS) + if (workload->ring_id != RCS) return; if (save) { @@ -130,7 +130,7 @@ static int populate_shadow_context(struct intel_vgpu_workload *workload) int ring_id = workload->ring_id; struct i915_gem_context *shadow_ctx = vgpu->submission.shadow_ctx; struct drm_i915_gem_object *ctx_obj = - shadow_ctx->engine[ring_id].state->obj; + shadow_ctx->__engine[ring_id].state->obj; struct execlist_ring_context *shadow_ring_context; struct page *page; void *dst; @@ -283,7 +283,7 @@ static int shadow_context_status_change(struct notifier_block *nb, static void shadow_context_descriptor_update(struct i915_gem_context *ctx, struct intel_engine_cs *engine) { - struct intel_context *ce = &ctx->engine[engine->id]; + struct intel_context *ce = to_intel_context(ctx, engine); u64 desc = 0; desc = ce->lrc_desc; @@ -389,7 +389,7 @@ int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload) * shadow_ctx pages invalid. So gvt need to pin itself. After update * the guest context, gvt can unpin the shadow_ctx safely. */ - ring = engine->context_pin(engine, shadow_ctx); + ring = intel_context_pin(shadow_ctx, engine); if (IS_ERR(ring)) { ret = PTR_ERR(ring); gvt_vgpu_err("fail to pin shadow context\n"); @@ -403,7 +403,7 @@ int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload) return 0; err_unpin: - engine->context_unpin(engine, shadow_ctx); + intel_context_unpin(shadow_ctx, engine); err_shadow: release_shadow_wa_ctx(&workload->wa_ctx); err_scan: @@ -437,7 +437,7 @@ static int intel_gvt_generate_request(struct intel_vgpu_workload *workload) return 0; err_unpin: - engine->context_unpin(engine, shadow_ctx); + intel_context_unpin(shadow_ctx, engine); release_shadow_wa_ctx(&workload->wa_ctx); return ret; } @@ -452,12 +452,6 @@ static int prepare_shadow_batch_buffer(struct intel_vgpu_workload *workload) int ret; list_for_each_entry(bb, &workload->shadow_bb, list) { - bb->vma = i915_gem_object_ggtt_pin(bb->obj, NULL, 0, 0, 0); - if (IS_ERR(bb->vma)) { - ret = PTR_ERR(bb->vma); - goto err; - } - /* For privilge batch buffer and not wa_ctx, the bb_start_cmd_va * is only updated into ring_scan_buffer, not real ring address * allocated in later copy_workload_to_ring_buffer. pls be noted @@ -469,25 +463,53 @@ static int prepare_shadow_batch_buffer(struct intel_vgpu_workload *workload) bb->bb_start_cmd_va = workload->shadow_ring_buffer_va + bb->bb_offset; - /* relocate shadow batch buffer */ - bb->bb_start_cmd_va[1] = i915_ggtt_offset(bb->vma); - if (gmadr_bytes == 8) - bb->bb_start_cmd_va[2] = 0; + if (bb->ppgtt) { + /* for non-priv bb, scan&shadow is only for + * debugging purpose, so the content of shadow bb + * is the same as original bb. Therefore, + * here, rather than switch to shadow bb's gma + * address, we directly use original batch buffer's + * gma address, and send original bb to hardware + * directly + */ + if (bb->clflush & CLFLUSH_AFTER) { + drm_clflush_virt_range(bb->va, + bb->obj->base.size); + bb->clflush &= ~CLFLUSH_AFTER; + } + i915_gem_obj_finish_shmem_access(bb->obj); + bb->accessing = false; + + } else { + bb->vma = i915_gem_object_ggtt_pin(bb->obj, + NULL, 0, 0, 0); + if (IS_ERR(bb->vma)) { + ret = PTR_ERR(bb->vma); + goto err; + } - /* No one is going to touch shadow bb from now on. */ - if (bb->clflush & CLFLUSH_AFTER) { - drm_clflush_virt_range(bb->va, bb->obj->base.size); - bb->clflush &= ~CLFLUSH_AFTER; - } + /* relocate shadow batch buffer */ + bb->bb_start_cmd_va[1] = i915_ggtt_offset(bb->vma); + if (gmadr_bytes == 8) + bb->bb_start_cmd_va[2] = 0; - ret = i915_gem_object_set_to_gtt_domain(bb->obj, false); - if (ret) - goto err; + /* No one is going to touch shadow bb from now on. */ + if (bb->clflush & CLFLUSH_AFTER) { + drm_clflush_virt_range(bb->va, + bb->obj->base.size); + bb->clflush &= ~CLFLUSH_AFTER; + } - i915_gem_obj_finish_shmem_access(bb->obj); - bb->accessing = false; + ret = i915_gem_object_set_to_gtt_domain(bb->obj, + false); + if (ret) + goto err; - i915_vma_move_to_active(bb->vma, workload->req, 0); + i915_gem_obj_finish_shmem_access(bb->obj); + bb->accessing = false; + + i915_vma_move_to_active(bb->vma, workload->req, 0); + } } return 0; err: @@ -504,7 +526,7 @@ static int update_wa_ctx_2_shadow_ctx(struct intel_shadow_wa_ctx *wa_ctx) struct intel_vgpu_submission *s = &workload->vgpu->submission; struct i915_gem_context *shadow_ctx = s->shadow_ctx; struct drm_i915_gem_object *ctx_obj = - shadow_ctx->engine[ring_id].state->obj; + shadow_ctx->__engine[ring_id].state->obj; struct execlist_ring_context *shadow_ring_context; struct page *page; @@ -666,7 +688,7 @@ static int dispatch_workload(struct intel_vgpu_workload *workload) ret = prepare_workload(workload); if (ret) { - engine->context_unpin(engine, shadow_ctx); + intel_context_unpin(shadow_ctx, engine); goto out; } @@ -749,7 +771,7 @@ static void update_guest_context(struct intel_vgpu_workload *workload) struct i915_gem_context *shadow_ctx = s->shadow_ctx; int ring_id = workload->ring_id; struct drm_i915_gem_object *ctx_obj = - shadow_ctx->engine[ring_id].state->obj; + shadow_ctx->__engine[ring_id].state->obj; struct execlist_ring_context *shadow_ring_context; struct page *page; void *src; @@ -876,7 +898,7 @@ static void complete_current_workload(struct intel_gvt *gvt, int ring_id) } mutex_lock(&dev_priv->drm.struct_mutex); /* unpin shadow ctx as the shadow_ctx update is done */ - engine->context_unpin(engine, s->shadow_ctx); + intel_context_unpin(s->shadow_ctx, engine); mutex_unlock(&dev_priv->drm.struct_mutex); } @@ -1134,9 +1156,6 @@ int intel_vgpu_setup_submission(struct intel_vgpu *vgpu) if (IS_ERR(s->shadow_ctx)) return PTR_ERR(s->shadow_ctx); - if (HAS_LOGICAL_RING_PREEMPTION(vgpu->gvt->dev_priv)) - s->shadow_ctx->priority = INT_MAX; - bitmap_zero(s->shadow_ctx_desc_updated, I915_NUM_ENGINES); s->workloads = kmem_cache_create_usercopy("gvt-g_vgpu_workload", diff --git a/drivers/gpu/drm/i915/gvt/scheduler.h b/drivers/gpu/drm/i915/gvt/scheduler.h index 486ed57..6c64478 100644 --- a/drivers/gpu/drm/i915/gvt/scheduler.h +++ b/drivers/gpu/drm/i915/gvt/scheduler.h @@ -125,6 +125,7 @@ struct intel_vgpu_shadow_bb { unsigned int clflush; bool accessing; unsigned long bb_offset; + bool ppgtt; }; #define workload_q_head(vgpu, ring_id) \ diff --git a/drivers/gpu/drm/i915/gvt/trace.h b/drivers/gpu/drm/i915/gvt/trace.h index 82093f1..1fd6420 100644 --- a/drivers/gpu/drm/i915/gvt/trace.h +++ b/drivers/gpu/drm/i915/gvt/trace.h @@ -224,19 +224,25 @@ TRACE_EVENT(oos_sync, TP_printk("%s", __entry->buf) ); +#define GVT_CMD_STR_LEN 40 TRACE_EVENT(gvt_command, - TP_PROTO(u8 vgpu_id, u8 ring_id, u32 ip_gma, u32 *cmd_va, u32 cmd_len, - u32 buf_type), + TP_PROTO(u8 vgpu_id, u8 ring_id, u32 ip_gma, u32 *cmd_va, + u32 cmd_len, u32 buf_type, u32 buf_addr_type, + void *workload, char *cmd_name), - TP_ARGS(vgpu_id, ring_id, ip_gma, cmd_va, cmd_len, buf_type), + TP_ARGS(vgpu_id, ring_id, ip_gma, cmd_va, cmd_len, buf_type, + buf_addr_type, workload, cmd_name), TP_STRUCT__entry( __field(u8, vgpu_id) __field(u8, ring_id) __field(u32, ip_gma) __field(u32, buf_type) + __field(u32, buf_addr_type) __field(u32, cmd_len) + __field(void*, workload) __dynamic_array(u32, raw_cmd, cmd_len) + __array(char, cmd_name, GVT_CMD_STR_LEN) ), TP_fast_assign( @@ -244,17 +250,25 @@ TRACE_EVENT(gvt_command, __entry->ring_id = ring_id; __entry->ip_gma = ip_gma; __entry->buf_type = buf_type; + __entry->buf_addr_type = buf_addr_type; __entry->cmd_len = cmd_len; + __entry->workload = workload; + snprintf(__entry->cmd_name, GVT_CMD_STR_LEN, "%s", cmd_name); memcpy(__get_dynamic_array(raw_cmd), cmd_va, cmd_len * sizeof(*cmd_va)); ), - TP_printk("vgpu%d ring %d: buf_type %u, ip_gma %08x, raw cmd %s", + TP_printk("vgpu%d ring %d: address_type %u, buf_type %u, ip_gma %08x,cmd (name=%s,len=%u,raw cmd=%s), workload=%p\n", __entry->vgpu_id, __entry->ring_id, + __entry->buf_addr_type, __entry->buf_type, __entry->ip_gma, - __print_array(__get_dynamic_array(raw_cmd), __entry->cmd_len, 4)) + __entry->cmd_name, + __entry->cmd_len, + __print_array(__get_dynamic_array(raw_cmd), + __entry->cmd_len, 4), + __entry->workload) ); #define GVT_TEMP_STR_LEN 10 diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 89f7ff2..13e7b9e 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -377,16 +377,19 @@ static void print_batch_pool_stats(struct seq_file *m, print_file_stats(m, "[k]batch pool", stats); } -static int per_file_ctx_stats(int id, void *ptr, void *data) +static int per_file_ctx_stats(int idx, void *ptr, void *data) { struct i915_gem_context *ctx = ptr; - int n; + struct intel_engine_cs *engine; + enum intel_engine_id id; + + for_each_engine(engine, ctx->i915, id) { + struct intel_context *ce = to_intel_context(ctx, engine); - for (n = 0; n < ARRAY_SIZE(ctx->engine); n++) { - if (ctx->engine[n].state) - per_file_stats(0, ctx->engine[n].state->obj, data); - if (ctx->engine[n].ring) - per_file_stats(0, ctx->engine[n].ring->vma->obj, data); + if (ce->state) + per_file_stats(0, ce->state->obj, data); + if (ce->ring) + per_file_stats(0, ce->ring->vma->obj, data); } return 0; @@ -1215,20 +1218,20 @@ static int i915_frequency_info(struct seq_file *m, void *unused) max_freq = (IS_GEN9_LP(dev_priv) ? rp_state_cap >> 0 : rp_state_cap >> 16) & 0xff; max_freq *= (IS_GEN9_BC(dev_priv) || - IS_CANNONLAKE(dev_priv) ? GEN9_FREQ_SCALER : 1); + INTEL_GEN(dev_priv) >= 10 ? GEN9_FREQ_SCALER : 1); seq_printf(m, "Lowest (RPN) frequency: %dMHz\n", intel_gpu_freq(dev_priv, max_freq)); max_freq = (rp_state_cap & 0xff00) >> 8; max_freq *= (IS_GEN9_BC(dev_priv) || - IS_CANNONLAKE(dev_priv) ? GEN9_FREQ_SCALER : 1); + INTEL_GEN(dev_priv) >= 10 ? GEN9_FREQ_SCALER : 1); seq_printf(m, "Nominal (RP1) frequency: %dMHz\n", intel_gpu_freq(dev_priv, max_freq)); max_freq = (IS_GEN9_LP(dev_priv) ? rp_state_cap >> 16 : rp_state_cap >> 0) & 0xff; max_freq *= (IS_GEN9_BC(dev_priv) || - IS_CANNONLAKE(dev_priv) ? GEN9_FREQ_SCALER : 1); + INTEL_GEN(dev_priv) >= 10 ? GEN9_FREQ_SCALER : 1); seq_printf(m, "Max non-overclocked (RP0) frequency: %dMHz\n", intel_gpu_freq(dev_priv, max_freq)); seq_printf(m, "Max overclocked frequency: %dMHz\n", @@ -1340,10 +1343,9 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused) struct rb_node *rb; seq_printf(m, "%s:\n", engine->name); - seq_printf(m, "\tseqno = %x [current %x, last %x], inflight %d\n", + seq_printf(m, "\tseqno = %x [current %x, last %x]\n", engine->hangcheck.seqno, seqno[id], - intel_engine_last_submit(engine), - engine->timeline->inflight_seqnos); + intel_engine_last_submit(engine)); seq_printf(m, "\twaiters? %s, fake irq active? %s, stalled? %s\n", yesno(intel_engine_has_waiter(engine)), yesno(test_bit(engine->id, @@ -1796,9 +1798,9 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused) { struct drm_i915_private *dev_priv = node_to_i915(m->private); struct intel_rps *rps = &dev_priv->gt_pm.rps; - int ret = 0; - int gpu_freq, ia_freq; unsigned int max_gpu_freq, min_gpu_freq; + int gpu_freq, ia_freq; + int ret; if (!HAS_LLC(dev_priv)) return -ENODEV; @@ -1809,13 +1811,12 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused) if (ret) goto out; - if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { + min_gpu_freq = rps->min_freq; + max_gpu_freq = rps->max_freq; + if (IS_GEN9_BC(dev_priv) || INTEL_GEN(dev_priv) >= 10) { /* Convert GT frequency to 50 HZ units */ - min_gpu_freq = rps->min_freq_softlimit / GEN9_FREQ_SCALER; - max_gpu_freq = rps->max_freq_softlimit / GEN9_FREQ_SCALER; - } else { - min_gpu_freq = rps->min_freq_softlimit; - max_gpu_freq = rps->max_freq_softlimit; + min_gpu_freq /= GEN9_FREQ_SCALER; + max_gpu_freq /= GEN9_FREQ_SCALER; } seq_puts(m, "GPU freq (MHz)\tEffective CPU freq (MHz)\tEffective Ring freq (MHz)\n"); @@ -1828,7 +1829,7 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused) seq_printf(m, "%d\t\t%d\t\t\t\t%d\n", intel_gpu_freq(dev_priv, (gpu_freq * (IS_GEN9_BC(dev_priv) || - IS_CANNONLAKE(dev_priv) ? + INTEL_GEN(dev_priv) >= 10 ? GEN9_FREQ_SCALER : 1))), ((ia_freq >> 0) & 0xff) * 100, ((ia_freq >> 8) & 0xff) * 100); @@ -1923,8 +1924,8 @@ static int i915_gem_framebuffer_info(struct seq_file *m, void *data) static void describe_ctx_ring(struct seq_file *m, struct intel_ring *ring) { - seq_printf(m, " (ringbuffer, space: %d, head: %u, tail: %u)", - ring->space, ring->head, ring->tail); + seq_printf(m, " (ringbuffer, space: %d, head: %u, tail: %u, emit: %u)", + ring->space, ring->head, ring->tail, ring->emit); } static int i915_context_status(struct seq_file *m, void *unused) @@ -1961,7 +1962,8 @@ static int i915_context_status(struct seq_file *m, void *unused) seq_putc(m, '\n'); for_each_engine(engine, dev_priv, id) { - struct intel_context *ce = &ctx->engine[engine->id]; + struct intel_context *ce = + to_intel_context(ctx, engine); seq_printf(m, "%s: ", engine->name); if (ce->state) @@ -2326,30 +2328,45 @@ static int i915_guc_load_status_info(struct seq_file *m, void *data) return 0; } -static void i915_guc_log_info(struct seq_file *m, - struct drm_i915_private *dev_priv) +static const char * +stringify_guc_log_type(enum guc_log_buffer_type type) { - struct intel_guc *guc = &dev_priv->guc; + switch (type) { + case GUC_ISR_LOG_BUFFER: + return "ISR"; + case GUC_DPC_LOG_BUFFER: + return "DPC"; + case GUC_CRASH_DUMP_LOG_BUFFER: + return "CRASH"; + default: + MISSING_CASE(type); + } - seq_puts(m, "\nGuC logging stats:\n"); + return ""; +} - seq_printf(m, "\tISR: flush count %10u, overflow count %10u\n", - guc->log.flush_count[GUC_ISR_LOG_BUFFER], - guc->log.total_overflow_count[GUC_ISR_LOG_BUFFER]); +static void i915_guc_log_info(struct seq_file *m, + struct drm_i915_private *dev_priv) +{ + struct intel_guc_log *log = &dev_priv->guc.log; + enum guc_log_buffer_type type; - seq_printf(m, "\tDPC: flush count %10u, overflow count %10u\n", - guc->log.flush_count[GUC_DPC_LOG_BUFFER], - guc->log.total_overflow_count[GUC_DPC_LOG_BUFFER]); + if (!intel_guc_log_relay_enabled(log)) { + seq_puts(m, "GuC log relay disabled\n"); + return; + } - seq_printf(m, "\tCRASH: flush count %10u, overflow count %10u\n", - guc->log.flush_count[GUC_CRASH_DUMP_LOG_BUFFER], - guc->log.total_overflow_count[GUC_CRASH_DUMP_LOG_BUFFER]); + seq_puts(m, "GuC logging stats:\n"); - seq_printf(m, "\tTotal flush interrupt count: %u\n", - guc->log.flush_interrupt_count); + seq_printf(m, "\tRelay full count: %u\n", + log->relay.full_count); - seq_printf(m, "\tCapture miss count: %u\n", - guc->log.capture_miss_count); + for (type = GUC_ISR_LOG_BUFFER; type < GUC_MAX_LOG_BUFFER; type++) { + seq_printf(m, "\t%s:\tflush count %10u, overflow count %10u\n", + stringify_guc_log_type(type), + log->stats[type].flush, + log->stats[type].sampled_overflow); + } } static void i915_guc_client_info(struct seq_file *m, @@ -2379,14 +2396,19 @@ static int i915_guc_info(struct seq_file *m, void *data) struct drm_i915_private *dev_priv = node_to_i915(m->private); const struct intel_guc *guc = &dev_priv->guc; - if (!USES_GUC_SUBMISSION(dev_priv)) + if (!USES_GUC(dev_priv)) return -ENODEV; + i915_guc_log_info(m, dev_priv); + + if (!USES_GUC_SUBMISSION(dev_priv)) + return 0; + GEM_BUG_ON(!guc->execbuf_client); - seq_printf(m, "Doorbell map:\n"); + seq_printf(m, "\nDoorbell map:\n"); seq_printf(m, "\t%*pb\n", GUC_NUM_DOORBELLS, guc->doorbell_bitmap); - seq_printf(m, "Doorbell next cacheline: 0x%x\n\n", guc->db_cacheline); + seq_printf(m, "Doorbell next cacheline: 0x%x\n", guc->db_cacheline); seq_printf(m, "\nGuC execbuf client @ %p:\n", guc->execbuf_client); i915_guc_client_info(m, dev_priv, guc->execbuf_client); @@ -2396,8 +2418,6 @@ static int i915_guc_info(struct seq_file *m, void *data) i915_guc_client_info(m, dev_priv, guc->preempt_client); } - i915_guc_log_info(m, dev_priv); - /* Add more as required ... */ return 0; @@ -2496,35 +2516,73 @@ static int i915_guc_log_dump(struct seq_file *m, void *data) return 0; } -static int i915_guc_log_control_get(void *data, u64 *val) +static int i915_guc_log_level_get(void *data, u64 *val) { struct drm_i915_private *dev_priv = data; - if (!HAS_GUC(dev_priv)) + if (!USES_GUC(dev_priv)) return -ENODEV; - if (!dev_priv->guc.log.vma) - return -EINVAL; - - *val = i915_modparams.guc_log_level; + *val = intel_guc_log_level_get(&dev_priv->guc.log); return 0; } -static int i915_guc_log_control_set(void *data, u64 val) +static int i915_guc_log_level_set(void *data, u64 val) { struct drm_i915_private *dev_priv = data; - if (!HAS_GUC(dev_priv)) + if (!USES_GUC(dev_priv)) return -ENODEV; - return intel_guc_log_control(&dev_priv->guc, val); + return intel_guc_log_level_set(&dev_priv->guc.log, val); } -DEFINE_SIMPLE_ATTRIBUTE(i915_guc_log_control_fops, - i915_guc_log_control_get, i915_guc_log_control_set, +DEFINE_SIMPLE_ATTRIBUTE(i915_guc_log_level_fops, + i915_guc_log_level_get, i915_guc_log_level_set, "%lld\n"); +static int i915_guc_log_relay_open(struct inode *inode, struct file *file) +{ + struct drm_i915_private *dev_priv = inode->i_private; + + if (!USES_GUC(dev_priv)) + return -ENODEV; + + file->private_data = &dev_priv->guc.log; + + return intel_guc_log_relay_open(&dev_priv->guc.log); +} + +static ssize_t +i915_guc_log_relay_write(struct file *filp, + const char __user *ubuf, + size_t cnt, + loff_t *ppos) +{ + struct intel_guc_log *log = filp->private_data; + + intel_guc_log_relay_flush(log); + + return cnt; +} + +static int i915_guc_log_relay_release(struct inode *inode, struct file *file) +{ + struct drm_i915_private *dev_priv = inode->i_private; + + intel_guc_log_relay_close(&dev_priv->guc.log); + + return 0; +} + +static const struct file_operations i915_guc_log_relay_fops = { + .owner = THIS_MODULE, + .open = i915_guc_log_relay_open, + .write = i915_guc_log_relay_write, + .release = i915_guc_log_relay_release, +}; + static const char *psr2_live_status(u32 val) { static const char * const live_status[] = { @@ -2548,6 +2606,26 @@ static const char *psr2_live_status(u32 val) return "unknown"; } +static const char *psr_sink_status(u8 val) +{ + static const char * const sink_status[] = { + "inactive", + "transition to active, capture and display", + "active, display from RFB", + "active, capture and display on sink device timings", + "transition to inactive, capture and display, timing re-sync", + "reserved", + "reserved", + "sink internal error" + }; + + val &= DP_PSR_SINK_STATE_MASK; + if (val < ARRAY_SIZE(sink_status)) + return sink_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); @@ -2569,14 +2647,13 @@ static int i915_edp_psr_status(struct seq_file *m, void *data) mutex_lock(&dev_priv->psr.lock); seq_printf(m, "Enabled: %s\n", yesno((bool)dev_priv->psr.enabled)); - seq_printf(m, "Active: %s\n", yesno(dev_priv->psr.active)); seq_printf(m, "Busy frontbuffer bits: 0x%03x\n", dev_priv->psr.busy_frontbuffer_bits); seq_printf(m, "Re-enable work scheduled: %s\n", yesno(work_busy(&dev_priv->psr.work.work))); if (HAS_DDI(dev_priv)) { - if (dev_priv->psr.psr2_support) + if (dev_priv->psr.psr2_enabled) enabled = I915_READ(EDP_PSR2_CTL) & EDP_PSR2_ENABLE; else enabled = I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE; @@ -2624,18 +2701,67 @@ static int i915_edp_psr_status(struct seq_file *m, void *data) seq_printf(m, "Performance_Counter: %u\n", psrperf); } - if (dev_priv->psr.psr2_support) { + if (dev_priv->psr.psr2_enabled) { u32 psr2 = I915_READ(EDP_PSR2_STATUS); seq_printf(m, "EDP_PSR2_STATUS: %x [%s]\n", psr2, psr2_live_status(psr2)); } + + if (dev_priv->psr.enabled) { + struct drm_dp_aux *aux = &dev_priv->psr.enabled->aux; + u8 val; + + if (drm_dp_dpcd_readb(aux, DP_PSR_STATUS, &val) == 1) + seq_printf(m, "Sink PSR status: 0x%x [%s]\n", val, + psr_sink_status(val)); + } mutex_unlock(&dev_priv->psr.lock); + if (READ_ONCE(dev_priv->psr.debug)) { + seq_printf(m, "Last attempted entry at: %lld\n", + dev_priv->psr.last_entry_attempt); + seq_printf(m, "Last exit at: %lld\n", + dev_priv->psr.last_exit); + } + intel_runtime_pm_put(dev_priv); return 0; } +static int +i915_edp_psr_debug_set(void *data, u64 val) +{ + struct drm_i915_private *dev_priv = data; + + if (!CAN_PSR(dev_priv)) + return -ENODEV; + + DRM_DEBUG_KMS("PSR debug %s\n", enableddisabled(val)); + + intel_runtime_pm_get(dev_priv); + intel_psr_irq_control(dev_priv, !!val); + intel_runtime_pm_put(dev_priv); + + return 0; +} + +static int +i915_edp_psr_debug_get(void *data, u64 *val) +{ + struct drm_i915_private *dev_priv = data; + + if (!CAN_PSR(dev_priv)) + return -ENODEV; + + *val = READ_ONCE(dev_priv->psr.debug); + return 0; +} + +DEFINE_SIMPLE_ATTRIBUTE(i915_edp_psr_debug_fops, + i915_edp_psr_debug_get, i915_edp_psr_debug_set, + "%llu\n"); + static int i915_sink_crc(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = node_to_i915(m->private); @@ -3231,7 +3357,8 @@ static int i915_shared_dplls_info(struct seq_file *m, void *unused) for (i = 0; i < dev_priv->num_shared_dpll; i++) { struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i]; - seq_printf(m, "DPLL%i: %s, id: %i\n", i, pll->name, pll->id); + seq_printf(m, "DPLL%i: %s, id: %i\n", i, pll->info->name, + pll->info->id); seq_printf(m, " crtc_mask: 0x%08x, active: 0x%x, on: %s\n", pll->state.crtc_mask, pll->active_mask, yesno(pll->on)); seq_printf(m, " tracked hardware state:\n"); @@ -3241,6 +3368,28 @@ static int i915_shared_dplls_info(struct seq_file *m, void *unused) seq_printf(m, " fp0: 0x%08x\n", pll->state.hw_state.fp0); seq_printf(m, " fp1: 0x%08x\n", pll->state.hw_state.fp1); seq_printf(m, " wrpll: 0x%08x\n", pll->state.hw_state.wrpll); + seq_printf(m, " cfgcr0: 0x%08x\n", pll->state.hw_state.cfgcr0); + seq_printf(m, " cfgcr1: 0x%08x\n", pll->state.hw_state.cfgcr1); + seq_printf(m, " mg_refclkin_ctl: 0x%08x\n", + pll->state.hw_state.mg_refclkin_ctl); + seq_printf(m, " mg_clktop2_coreclkctl1: 0x%08x\n", + pll->state.hw_state.mg_clktop2_coreclkctl1); + seq_printf(m, " mg_clktop2_hsclkctl: 0x%08x\n", + pll->state.hw_state.mg_clktop2_hsclkctl); + seq_printf(m, " mg_pll_div0: 0x%08x\n", + pll->state.hw_state.mg_pll_div0); + seq_printf(m, " mg_pll_div1: 0x%08x\n", + pll->state.hw_state.mg_pll_div1); + seq_printf(m, " mg_pll_lf: 0x%08x\n", + pll->state.hw_state.mg_pll_lf); + seq_printf(m, " mg_pll_frac_lock: 0x%08x\n", + pll->state.hw_state.mg_pll_frac_lock); + seq_printf(m, " mg_pll_ssc: 0x%08x\n", + pll->state.hw_state.mg_pll_ssc); + seq_printf(m, " mg_pll_bias: 0x%08x\n", + pll->state.hw_state.mg_pll_bias); + seq_printf(m, " mg_pll_tdc_coldst_bias: 0x%08x\n", + pll->state.hw_state.mg_pll_tdc_coldst_bias); } drm_modeset_unlock_all(dev); @@ -3249,24 +3398,13 @@ static int i915_shared_dplls_info(struct seq_file *m, void *unused) static int i915_wa_registers(struct seq_file *m, void *unused) { - int i; - int ret; - struct intel_engine_cs *engine; struct drm_i915_private *dev_priv = node_to_i915(m->private); - struct drm_device *dev = &dev_priv->drm; struct i915_workarounds *workarounds = &dev_priv->workarounds; - enum intel_engine_id id; - - ret = mutex_lock_interruptible(&dev->struct_mutex); - if (ret) - return ret; + int i; intel_runtime_pm_get(dev_priv); seq_printf(m, "Workarounds applied: %d\n", workarounds->count); - for_each_engine(engine, dev_priv, id) - seq_printf(m, "HW whitelist count for %s: %d\n", - engine->name, workarounds->hw_whitelist_count[id]); for (i = 0; i < workarounds->count; ++i) { i915_reg_t addr; u32 mask, value, read; @@ -3282,7 +3420,6 @@ static int i915_wa_registers(struct seq_file *m, void *unused) } intel_runtime_pm_put(dev_priv); - mutex_unlock(&dev->struct_mutex); return 0; } @@ -3567,7 +3704,8 @@ static ssize_t i915_displayport_test_active_write(struct file *file, static int i915_displayport_test_active_show(struct seq_file *m, void *data) { - struct drm_device *dev = m->private; + struct drm_i915_private *dev_priv = m->private; + struct drm_device *dev = &dev_priv->drm; struct drm_connector *connector; struct drm_connector_list_iter conn_iter; struct intel_dp *intel_dp; @@ -3601,10 +3739,8 @@ static int i915_displayport_test_active_show(struct seq_file *m, void *data) static int i915_displayport_test_active_open(struct inode *inode, struct file *file) { - struct drm_i915_private *dev_priv = inode->i_private; - return single_open(file, i915_displayport_test_active_show, - &dev_priv->drm); + inode->i_private); } static const struct file_operations i915_displayport_test_active_fops = { @@ -3618,7 +3754,8 @@ static const struct file_operations i915_displayport_test_active_fops = { static int i915_displayport_test_data_show(struct seq_file *m, void *data) { - struct drm_device *dev = m->private; + struct drm_i915_private *dev_priv = m->private; + struct drm_device *dev = &dev_priv->drm; struct drm_connector *connector; struct drm_connector_list_iter conn_iter; struct intel_dp *intel_dp; @@ -3657,26 +3794,12 @@ static int i915_displayport_test_data_show(struct seq_file *m, void *data) return 0; } -static int i915_displayport_test_data_open(struct inode *inode, - struct file *file) -{ - struct drm_i915_private *dev_priv = inode->i_private; - - return single_open(file, i915_displayport_test_data_show, - &dev_priv->drm); -} - -static const struct file_operations i915_displayport_test_data_fops = { - .owner = THIS_MODULE, - .open = i915_displayport_test_data_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release -}; +DEFINE_SHOW_ATTRIBUTE(i915_displayport_test_data); static int i915_displayport_test_type_show(struct seq_file *m, void *data) { - struct drm_device *dev = m->private; + struct drm_i915_private *dev_priv = m->private; + struct drm_device *dev = &dev_priv->drm; struct drm_connector *connector; struct drm_connector_list_iter conn_iter; struct intel_dp *intel_dp; @@ -3703,23 +3826,7 @@ static int i915_displayport_test_type_show(struct seq_file *m, void *data) return 0; } - -static int i915_displayport_test_type_open(struct inode *inode, - struct file *file) -{ - struct drm_i915_private *dev_priv = inode->i_private; - - return single_open(file, i915_displayport_test_type_show, - &dev_priv->drm); -} - -static const struct file_operations i915_displayport_test_type_fops = { - .owner = THIS_MODULE, - .open = i915_displayport_test_type_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release -}; +DEFINE_SHOW_ATTRIBUTE(i915_displayport_test_type); static void wm_latency_show(struct seq_file *m, const uint16_t wm[8]) { @@ -3987,8 +4094,8 @@ i915_wedged_set(void *data, u64 val) engine->hangcheck.stalled = true; } - i915_handle_error(i915, val, "Manually set wedged engine mask = %llx", - val); + i915_handle_error(i915, val, I915_ERROR_CAPTURE, + "Manually set wedged engine mask = %llx", val); wait_on_bit(&i915->gpu_error.flags, I915_RESET_HANDOFF, @@ -4152,119 +4259,6 @@ DEFINE_SIMPLE_ATTRIBUTE(i915_drop_caches_fops, "0x%08llx\n"); static int -i915_max_freq_get(void *data, u64 *val) -{ - struct drm_i915_private *dev_priv = data; - - if (INTEL_GEN(dev_priv) < 6) - return -ENODEV; - - *val = intel_gpu_freq(dev_priv, dev_priv->gt_pm.rps.max_freq_softlimit); - return 0; -} - -static int -i915_max_freq_set(void *data, u64 val) -{ - struct drm_i915_private *dev_priv = data; - struct intel_rps *rps = &dev_priv->gt_pm.rps; - u32 hw_max, hw_min; - int ret; - - if (INTEL_GEN(dev_priv) < 6) - return -ENODEV; - - DRM_DEBUG_DRIVER("Manually setting max freq to %llu\n", val); - - ret = mutex_lock_interruptible(&dev_priv->pcu_lock); - if (ret) - return ret; - - /* - * Turbo will still be enabled, but won't go above the set value. - */ - val = intel_freq_opcode(dev_priv, val); - - hw_max = rps->max_freq; - hw_min = rps->min_freq; - - if (val < hw_min || val > hw_max || val < rps->min_freq_softlimit) { - mutex_unlock(&dev_priv->pcu_lock); - return -EINVAL; - } - - rps->max_freq_softlimit = val; - - if (intel_set_rps(dev_priv, val)) - DRM_DEBUG_DRIVER("failed to update RPS to new softlimit\n"); - - mutex_unlock(&dev_priv->pcu_lock); - - return 0; -} - -DEFINE_SIMPLE_ATTRIBUTE(i915_max_freq_fops, - i915_max_freq_get, i915_max_freq_set, - "%llu\n"); - -static int -i915_min_freq_get(void *data, u64 *val) -{ - struct drm_i915_private *dev_priv = data; - - if (INTEL_GEN(dev_priv) < 6) - return -ENODEV; - - *val = intel_gpu_freq(dev_priv, dev_priv->gt_pm.rps.min_freq_softlimit); - return 0; -} - -static int -i915_min_freq_set(void *data, u64 val) -{ - struct drm_i915_private *dev_priv = data; - struct intel_rps *rps = &dev_priv->gt_pm.rps; - u32 hw_max, hw_min; - int ret; - - if (INTEL_GEN(dev_priv) < 6) - return -ENODEV; - - DRM_DEBUG_DRIVER("Manually setting min freq to %llu\n", val); - - ret = mutex_lock_interruptible(&dev_priv->pcu_lock); - if (ret) - return ret; - - /* - * Turbo will still be enabled, but won't go below the set value. - */ - val = intel_freq_opcode(dev_priv, val); - - hw_max = rps->max_freq; - hw_min = rps->min_freq; - - if (val < hw_min || - val > hw_max || val > rps->max_freq_softlimit) { - mutex_unlock(&dev_priv->pcu_lock); - return -EINVAL; - } - - rps->min_freq_softlimit = val; - - if (intel_set_rps(dev_priv, val)) - DRM_DEBUG_DRIVER("failed to update RPS to new softlimit\n"); - - mutex_unlock(&dev_priv->pcu_lock); - - return 0; -} - -DEFINE_SIMPLE_ATTRIBUTE(i915_min_freq_fops, - i915_min_freq_get, i915_min_freq_set, - "%llu\n"); - -static int i915_cache_sharing_get(void *data, u64 *val) { struct drm_i915_private *dev_priv = data; @@ -4316,9 +4310,10 @@ DEFINE_SIMPLE_ATTRIBUTE(i915_cache_sharing_fops, static void cherryview_sseu_device_status(struct drm_i915_private *dev_priv, struct sseu_dev_info *sseu) { - int ss_max = 2; +#define SS_MAX 2 + const int ss_max = SS_MAX; + u32 sig1[SS_MAX], sig2[SS_MAX]; int ss; - u32 sig1[ss_max], sig2[ss_max]; sig1[0] = I915_READ(CHV_POWER_SS0_SIG1); sig1[1] = I915_READ(CHV_POWER_SS1_SIG1); @@ -4342,15 +4337,16 @@ static void cherryview_sseu_device_status(struct drm_i915_private *dev_priv, sseu->eu_per_subslice = max_t(unsigned int, sseu->eu_per_subslice, eu_cnt); } +#undef SS_MAX } static void gen10_sseu_device_status(struct drm_i915_private *dev_priv, struct sseu_dev_info *sseu) { +#define SS_MAX 6 const struct intel_device_info *info = INTEL_INFO(dev_priv); + u32 s_reg[SS_MAX], eu_reg[2 * SS_MAX], eu_mask[2]; int s, ss; - u32 s_reg[info->sseu.max_slices]; - u32 eu_reg[2 * info->sseu.max_subslices], eu_mask[2]; for (s = 0; s < info->sseu.max_slices; s++) { /* @@ -4397,15 +4393,16 @@ static void gen10_sseu_device_status(struct drm_i915_private *dev_priv, eu_cnt); } } +#undef SS_MAX } static void gen9_sseu_device_status(struct drm_i915_private *dev_priv, struct sseu_dev_info *sseu) { +#define SS_MAX 3 const struct intel_device_info *info = INTEL_INFO(dev_priv); + u32 s_reg[SS_MAX], eu_reg[2 * SS_MAX], eu_mask[2]; int s, ss; - u32 s_reg[info->sseu.max_slices]; - u32 eu_reg[2 * info->sseu.max_subslices], eu_mask[2]; for (s = 0; s < info->sseu.max_slices; s++) { s_reg[s] = I915_READ(GEN9_SLICE_PGCTL_ACK(s)); @@ -4452,6 +4449,7 @@ static void gen9_sseu_device_status(struct drm_i915_private *dev_priv, eu_cnt); } } +#undef SS_MAX } static void broadwell_sseu_device_status(struct drm_i915_private *dev_priv, @@ -4703,6 +4701,67 @@ static int i915_drrs_ctl_set(void *data, u64 val) DEFINE_SIMPLE_ATTRIBUTE(i915_drrs_ctl_fops, NULL, i915_drrs_ctl_set, "%llu\n"); +static ssize_t +i915_fifo_underrun_reset_write(struct file *filp, + const char __user *ubuf, + size_t cnt, loff_t *ppos) +{ + struct drm_i915_private *dev_priv = filp->private_data; + struct intel_crtc *intel_crtc; + struct drm_device *dev = &dev_priv->drm; + int ret; + bool reset; + + ret = kstrtobool_from_user(ubuf, cnt, &reset); + if (ret) + return ret; + + if (!reset) + return cnt; + + for_each_intel_crtc(dev, intel_crtc) { + struct drm_crtc_commit *commit; + struct intel_crtc_state *crtc_state; + + ret = drm_modeset_lock_single_interruptible(&intel_crtc->base.mutex); + if (ret) + return ret; + + crtc_state = to_intel_crtc_state(intel_crtc->base.state); + commit = crtc_state->base.commit; + if (commit) { + ret = wait_for_completion_interruptible(&commit->hw_done); + if (!ret) + ret = wait_for_completion_interruptible(&commit->flip_done); + } + + if (!ret && crtc_state->base.active) { + DRM_DEBUG_KMS("Re-arming FIFO underruns on pipe %c\n", + pipe_name(intel_crtc->pipe)); + + intel_crtc_arm_fifo_underrun(intel_crtc, crtc_state); + } + + drm_modeset_unlock(&intel_crtc->base.mutex); + + if (ret) + return ret; + } + + ret = intel_fbc_reset_underrun(dev_priv); + if (ret) + return ret; + + return cnt; +} + +static const struct file_operations i915_fifo_underrun_reset_ops = { + .owner = THIS_MODULE, + .open = simple_open, + .write = i915_fifo_underrun_reset_write, + .llseek = default_llseek, +}; + static const struct drm_info_list i915_debugfs_list[] = { {"i915_capabilities", i915_capabilities, 0}, {"i915_gem_objects", i915_gem_object_info, 0}, @@ -4760,8 +4819,6 @@ static const struct i915_debugfs_files { const struct file_operations *fops; } i915_debugfs_files[] = { {"i915_wedged", &i915_wedged_fops}, - {"i915_max_freq", &i915_max_freq_fops}, - {"i915_min_freq", &i915_min_freq_fops}, {"i915_cache_sharing", &i915_cache_sharing_fops}, {"i915_ring_missed_irq", &i915_ring_missed_irq_fops}, {"i915_ring_test_irq", &i915_ring_test_irq_fops}, @@ -4770,6 +4827,7 @@ static const struct i915_debugfs_files { {"i915_error_state", &i915_error_state_fops}, {"i915_gpu_info", &i915_gpu_info_fops}, #endif + {"i915_fifo_underrun_reset", &i915_fifo_underrun_reset_ops}, {"i915_next_seqno", &i915_next_seqno_fops}, {"i915_display_crc_ctl", &i915_display_crc_ctl_fops}, {"i915_pri_wm_latency", &i915_pri_wm_latency_fops}, @@ -4779,10 +4837,12 @@ static const struct i915_debugfs_files { {"i915_dp_test_data", &i915_displayport_test_data_fops}, {"i915_dp_test_type", &i915_displayport_test_type_fops}, {"i915_dp_test_active", &i915_displayport_test_active_fops}, - {"i915_guc_log_control", &i915_guc_log_control_fops}, + {"i915_guc_log_level", &i915_guc_log_level_fops}, + {"i915_guc_log_relay", &i915_guc_log_relay_fops}, {"i915_hpd_storm_ctl", &i915_hpd_storm_ctl_fops}, {"i915_ipc_status", &i915_ipc_status_fops}, - {"i915_drrs_ctl", &i915_drrs_ctl_fops} + {"i915_drrs_ctl", &i915_drrs_ctl_fops}, + {"i915_edp_psr_debug", &i915_edp_psr_debug_fops} }; int i915_debugfs_register(struct drm_i915_private *dev_priv) @@ -4876,19 +4936,7 @@ static int i915_dpcd_show(struct seq_file *m, void *data) return 0; } - -static int i915_dpcd_open(struct inode *inode, struct file *file) -{ - return single_open(file, i915_dpcd_show, inode->i_private); -} - -static const struct file_operations i915_dpcd_fops = { - .owner = THIS_MODULE, - .open = i915_dpcd_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release, -}; +DEFINE_SHOW_ATTRIBUTE(i915_dpcd); static int i915_panel_show(struct seq_file *m, void *data) { @@ -4910,19 +4958,7 @@ static int i915_panel_show(struct seq_file *m, void *data) return 0; } - -static int i915_panel_open(struct inode *inode, struct file *file) -{ - return single_open(file, i915_panel_show, inode->i_private); -} - -static const struct file_operations i915_panel_fops = { - .owner = THIS_MODULE, - .open = i915_panel_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release, -}; +DEFINE_SHOW_ATTRIBUTE(i915_panel); /** * i915_debugfs_connector_add - add i915 specific connector debugfs files diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 3b4daaf..9c449b8 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -101,7 +101,13 @@ __i915_printk(struct drm_i915_private *dev_priv, const char *level, __builtin_return_address(0), &vaf); if (is_error && !shown_bug_once) { - dev_notice(kdev, "%s", FDO_BUG_MSG); + /* + * Ask the user to file a bug report for the error, except + * if they may have caused the bug by fiddling with unsafe + * module parameters. + */ + if (!test_taint(TAINT_USER)) + dev_notice(kdev, "%s", FDO_BUG_MSG); shown_bug_once = true; } @@ -377,9 +383,9 @@ static int i915_getparam_ioctl(struct drm_device *dev, void *data, value = INTEL_INFO(dev_priv)->sseu.min_eu_in_pool; break; case I915_PARAM_HUC_STATUS: - intel_runtime_pm_get(dev_priv); - value = I915_READ(HUC_STATUS2) & HUC_FW_VERIFIED; - intel_runtime_pm_put(dev_priv); + value = intel_huc_check_status(&dev_priv->huc); + if (value < 0) + return value; break; case I915_PARAM_MMAP_GTT_VERSION: /* Though we've started our numbering from 1, and so class all @@ -695,11 +701,9 @@ static int i915_load_modeset_init(struct drm_device *dev) if (ret) goto cleanup_irq; - intel_uc_init_fw(dev_priv); - ret = i915_gem_init(dev_priv); if (ret) - goto cleanup_uc; + goto cleanup_irq; intel_setup_overlay(dev_priv); @@ -719,8 +723,6 @@ cleanup_gem: if (i915_gem_suspend(dev_priv)) DRM_ERROR("failed to idle hardware; continuing to unload!\n"); i915_gem_fini(dev_priv); -cleanup_uc: - intel_uc_fini_fw(dev_priv); cleanup_irq: drm_irq_uninstall(dev); intel_teardown_gmbus(dev_priv); @@ -922,16 +924,21 @@ static int i915_driver_init_early(struct drm_i915_private *dev_priv, mutex_init(&dev_priv->wm.wm_mutex); mutex_init(&dev_priv->pps_mutex); - intel_uc_init_early(dev_priv); i915_memcpy_init_early(dev_priv); ret = i915_workqueues_init(dev_priv); if (ret < 0) goto err_engines; + ret = i915_gem_init_early(dev_priv); + if (ret < 0) + goto err_workqueues; + /* This must be called before any calls to HAS_PCH_* */ intel_detect_pch(dev_priv); + intel_wopcm_init_early(&dev_priv->wopcm); + intel_uc_init_early(dev_priv); intel_pm_setup(dev_priv); intel_init_dpio(dev_priv); intel_power_domains_init(dev_priv); @@ -940,18 +947,13 @@ static int i915_driver_init_early(struct drm_i915_private *dev_priv, intel_init_display_hooks(dev_priv); intel_init_clock_gating_hooks(dev_priv); intel_init_audio_hooks(dev_priv); - ret = i915_gem_load_init(dev_priv); - if (ret < 0) - goto err_irq; - intel_display_crc_init(dev_priv); intel_detect_preproduction_hw(dev_priv); return 0; -err_irq: - intel_irq_fini(dev_priv); +err_workqueues: i915_workqueues_cleanup(dev_priv); err_engines: i915_engines_cleanup(dev_priv); @@ -964,8 +966,9 @@ err_engines: */ static void i915_driver_cleanup_early(struct drm_i915_private *dev_priv) { - i915_gem_load_cleanup(dev_priv); intel_irq_fini(dev_priv); + intel_uc_cleanup_early(dev_priv); + i915_gem_cleanup_early(dev_priv); i915_workqueues_cleanup(dev_priv); i915_engines_cleanup(dev_priv); } @@ -1035,6 +1038,10 @@ static int i915_driver_init_mmio(struct drm_i915_private *dev_priv) intel_uncore_init(dev_priv); + intel_device_info_init_mmio(dev_priv); + + intel_uncore_prune(dev_priv); + intel_uc_init_mmio(dev_priv); ret = intel_engines_init_mmio(dev_priv); @@ -1077,8 +1084,6 @@ static void intel_sanitize_options(struct drm_i915_private *dev_priv) i915_modparams.enable_ppgtt); DRM_DEBUG_DRIVER("ppgtt mode: %i\n", i915_modparams.enable_ppgtt); - intel_uc_sanitize_options(dev_priv); - intel_gvt_sanitize_options(dev_priv); } @@ -1244,7 +1249,6 @@ static void i915_driver_register(struct drm_i915_private *dev_priv) /* Reveal our presence to userspace */ if (drm_dev_register(dev, 0) == 0) { i915_debugfs_register(dev_priv); - i915_guc_log_register(dev_priv); i915_setup_sysfs(dev_priv); /* Depends on sysfs having been initialized */ @@ -1304,7 +1308,6 @@ static void i915_driver_unregister(struct drm_i915_private *dev_priv) i915_pmu_unregister(dev_priv); i915_teardown_sysfs(dev_priv); - i915_guc_log_unregister(dev_priv); drm_dev_unregister(&dev_priv->drm); i915_gem_shrinker_unregister(dev_priv); @@ -1463,7 +1466,6 @@ void i915_driver_unload(struct drm_device *dev) i915_reset_error_state(dev_priv); i915_gem_fini(dev_priv); - intel_uc_fini_fw(dev_priv); intel_fbc_cleanup_cfb(dev_priv); intel_power_domains_fini(dev_priv); @@ -1876,7 +1878,8 @@ static int i915_resume_switcheroo(struct drm_device *dev) /** * i915_reset - reset chip after a hang * @i915: #drm_i915_private to reset - * @flags: Instructions + * @stalled_mask: mask of the stalled engines with the guilty requests + * @reason: user error message for why we are resetting * * Reset the chip. Useful if a hang is detected. Marks the device as wedged * on failure. @@ -1891,12 +1894,16 @@ static int i915_resume_switcheroo(struct drm_device *dev) * - re-init interrupt state * - re-init display */ -void i915_reset(struct drm_i915_private *i915, unsigned int flags) +void i915_reset(struct drm_i915_private *i915, + unsigned int stalled_mask, + const char *reason) { struct i915_gpu_error *error = &i915->gpu_error; int ret; int i; + GEM_TRACE("flags=%lx\n", error->flags); + might_sleep(); lockdep_assert_held(&i915->drm.struct_mutex); GEM_BUG_ON(!test_bit(I915_RESET_BACKOFF, &error->flags)); @@ -1908,8 +1915,8 @@ void i915_reset(struct drm_i915_private *i915, unsigned int flags) if (!i915_gem_unset_wedged(i915)) goto wakeup; - if (!(flags & I915_RESET_QUIET)) - dev_notice(i915->drm.dev, "Resetting chip after gpu hang\n"); + if (reason) + dev_notice(i915->drm.dev, "Resetting chip for %s\n", reason); error->reset_count++; disable_irq(i915->drm.irq); @@ -1952,7 +1959,7 @@ void i915_reset(struct drm_i915_private *i915, unsigned int flags) goto error; } - i915_gem_reset(i915); + i915_gem_reset(i915, stalled_mask); intel_overlay_reset(i915); /* @@ -1998,7 +2005,6 @@ taint: error: i915_gem_set_wedged(i915); i915_retire_requests(i915); - intel_gpu_reset(i915, ALL_ENGINES); goto finish; } @@ -2011,7 +2017,7 @@ static inline int intel_gt_reset_engine(struct drm_i915_private *dev_priv, /** * i915_reset_engine - reset GPU engine to recover from a hang * @engine: engine to reset - * @flags: options + * @msg: reason for GPU reset; or NULL for no dev_notice() * * Reset a specific GPU engine. Useful if a hang is detected. * Returns zero on successful reset or otherwise an error code. @@ -2021,12 +2027,13 @@ static inline int intel_gt_reset_engine(struct drm_i915_private *dev_priv, * - reset engine (which will force the engine to idle) * - re-init/configure engine */ -int i915_reset_engine(struct intel_engine_cs *engine, unsigned int flags) +int i915_reset_engine(struct intel_engine_cs *engine, const char *msg) { struct i915_gpu_error *error = &engine->i915->gpu_error; struct i915_request *active_request; int ret; + GEM_TRACE("%s flags=%lx\n", engine->name, error->flags); GEM_BUG_ON(!test_bit(I915_RESET_ENGINE + engine->id, &error->flags)); active_request = i915_gem_reset_prepare_engine(engine); @@ -2036,10 +2043,9 @@ int i915_reset_engine(struct intel_engine_cs *engine, unsigned int flags) goto out; } - if (!(flags & I915_RESET_QUIET)) { + if (msg) dev_notice(engine->i915->drm.dev, - "Resetting %s after gpu hang\n", engine->name); - } + "Resetting %s for %s\n", engine->name, msg); error->reset_engine_count[engine->id]++; if (!engine->i915->guc.execbuf_client) @@ -2059,7 +2065,7 @@ int i915_reset_engine(struct intel_engine_cs *engine, unsigned int flags) * active request and can drop it, adjust head to skip the offending * request to resume executing remaining requests in the queue. */ - i915_gem_reset_engine(engine, active_request); + i915_gem_reset_engine(engine, active_request, true); /* * The engine and its registers (and workarounds in case of render) @@ -2468,10 +2474,13 @@ static void vlv_wait_for_gt_wells(struct drm_i915_private *dev_priv, /* * RC6 transitioning can be delayed up to 2 msec (see * valleyview_enable_rps), use 3 msec for safety. + * + * This can fail to turn off the rc6 if the GPU is stuck after a failed + * reset and we are trying to force the machine to sleep. */ if (vlv_wait_for_pw_status(dev_priv, mask, val)) - DRM_ERROR("timeout waiting for GT wells to go %s\n", - onoff(wait_for_on)); + DRM_DEBUG_DRIVER("timeout waiting for GT wells to go %s\n", + onoff(wait_for_on)); } static void vlv_check_no_gt_access(struct drm_i915_private *dev_priv) @@ -2822,10 +2831,10 @@ static const struct drm_ioctl_desc i915_ioctls[] = { 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_ioctl, 0), DRM_IOCTL_DEF_DRV(I915_GEM_MADVISE, i915_gem_madvise_ioctl, DRM_RENDER_ALLOW), - DRM_IOCTL_DEF_DRV(I915_OVERLAY_PUT_IMAGE, intel_overlay_put_image_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF_DRV(I915_OVERLAY_ATTRS, intel_overlay_attrs_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF_DRV(I915_SET_SPRITE_COLORKEY, intel_sprite_set_colorkey_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW), - DRM_IOCTL_DEF_DRV(I915_GET_SPRITE_COLORKEY, drm_noop, DRM_MASTER|DRM_CONTROL_ALLOW), + DRM_IOCTL_DEF_DRV(I915_OVERLAY_PUT_IMAGE, intel_overlay_put_image_ioctl, DRM_MASTER), + DRM_IOCTL_DEF_DRV(I915_OVERLAY_ATTRS, intel_overlay_attrs_ioctl, DRM_MASTER), + DRM_IOCTL_DEF_DRV(I915_SET_SPRITE_COLORKEY, intel_sprite_set_colorkey_ioctl, DRM_MASTER), + DRM_IOCTL_DEF_DRV(I915_GET_SPRITE_COLORKEY, drm_noop, DRM_MASTER), DRM_IOCTL_DEF_DRV(I915_GEM_WAIT, i915_gem_wait_ioctl, DRM_AUTH|DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_CREATE, i915_gem_context_create_ioctl, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(I915_GEM_CONTEXT_DESTROY, i915_gem_context_destroy_ioctl, DRM_RENDER_ALLOW), diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index ce18b6c..34c125e 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -64,6 +64,7 @@ #include "intel_opregion.h" #include "intel_ringbuffer.h" #include "intel_uncore.h" +#include "intel_wopcm.h" #include "intel_uc.h" #include "i915_gem.h" @@ -71,9 +72,10 @@ #include "i915_gem_fence_reg.h" #include "i915_gem_object.h" #include "i915_gem_gtt.h" -#include "i915_gem_timeline.h" - +#include "i915_gpu_error.h" #include "i915_request.h" +#include "i915_scheduler.h" +#include "i915_timeline.h" #include "i915_vma.h" #include "intel_gvt.h" @@ -83,8 +85,8 @@ #define DRIVER_NAME "i915" #define DRIVER_DESC "Intel Graphics" -#define DRIVER_DATE "20180308" -#define DRIVER_TIMESTAMP 1520513379 +#define DRIVER_DATE "20180514" +#define DRIVER_TIMESTAMP 1526300884 /* Use I915_STATE_WARN(x) and I915_STATE_WARN_ON() (rather than WARN() and * WARN_ON()) for hw state sanity checks to check for unexpected conditions @@ -261,6 +263,7 @@ enum hpd_pin { HPD_PORT_C, HPD_PORT_D, HPD_PORT_E, + HPD_PORT_F, HPD_NUM_PINS }; @@ -453,172 +456,6 @@ struct intel_csr { uint32_t allowed_dc_mask; }; -struct intel_display_error_state; - -struct i915_gpu_state { - struct kref ref; - ktime_t time; - ktime_t boottime; - ktime_t uptime; - - struct drm_i915_private *i915; - - char error_msg[128]; - bool simulated; - bool awake; - bool wakelock; - bool suspended; - int iommu; - u32 reset_count; - u32 suspend_count; - struct intel_device_info device_info; - struct intel_driver_caps driver_caps; - struct i915_params params; - - struct i915_error_uc { - struct intel_uc_fw guc_fw; - struct intel_uc_fw huc_fw; - struct drm_i915_error_object *guc_log; - } uc; - - /* Generic register state */ - u32 eir; - u32 pgtbl_er; - u32 ier; - u32 gtier[4], ngtier; - u32 ccid; - u32 derrmr; - u32 forcewake; - u32 error; /* gen6+ */ - u32 err_int; /* gen7 */ - u32 fault_data0; /* gen8, gen9 */ - u32 fault_data1; /* gen8, gen9 */ - u32 done_reg; - u32 gac_eco; - u32 gam_ecochk; - u32 gab_ctl; - u32 gfx_mode; - - u32 nfence; - u64 fence[I915_MAX_NUM_FENCES]; - struct intel_overlay_error_state *overlay; - struct intel_display_error_state *display; - - struct drm_i915_error_engine { - int engine_id; - /* Software tracked state */ - bool idle; - bool waiting; - int num_waiters; - unsigned long hangcheck_timestamp; - bool hangcheck_stalled; - enum intel_engine_hangcheck_action hangcheck_action; - struct i915_address_space *vm; - int num_requests; - u32 reset_count; - - /* position of active request inside the ring */ - u32 rq_head, rq_post, rq_tail; - - /* our own tracking of ring head and tail */ - u32 cpu_ring_head; - u32 cpu_ring_tail; - - u32 last_seqno; - - /* Register state */ - u32 start; - u32 tail; - u32 head; - u32 ctl; - u32 mode; - u32 hws; - u32 ipeir; - u32 ipehr; - u32 bbstate; - u32 instpm; - u32 instps; - u32 seqno; - u64 bbaddr; - u64 acthd; - u32 fault_reg; - u64 faddr; - u32 rc_psmi; /* sleep state */ - u32 semaphore_mboxes[I915_NUM_ENGINES - 1]; - struct intel_instdone instdone; - - struct drm_i915_error_context { - char comm[TASK_COMM_LEN]; - pid_t pid; - u32 handle; - u32 hw_id; - int priority; - int ban_score; - int active; - int guilty; - bool bannable; - } context; - - struct drm_i915_error_object { - u64 gtt_offset; - u64 gtt_size; - int page_count; - int unused; - u32 *pages[0]; - } *ringbuffer, *batchbuffer, *wa_batchbuffer, *ctx, *hws_page; - - struct drm_i915_error_object **user_bo; - long user_bo_count; - - struct drm_i915_error_object *wa_ctx; - struct drm_i915_error_object *default_state; - - struct drm_i915_error_request { - long jiffies; - pid_t pid; - u32 context; - int priority; - int ban_score; - u32 seqno; - u32 head; - u32 tail; - } *requests, execlist[EXECLIST_MAX_PORTS]; - unsigned int num_ports; - - struct drm_i915_error_waiter { - char comm[TASK_COMM_LEN]; - pid_t pid; - u32 seqno; - } *waiters; - - struct { - u32 gfx_mode; - union { - u64 pdp[4]; - u32 pp_dir_base; - }; - } vm_info; - } engine[I915_NUM_ENGINES]; - - struct drm_i915_error_buffer { - u32 size; - u32 name; - u32 rseqno[I915_NUM_ENGINES], wseqno; - u64 gtt_offset; - u32 read_domains; - u32 write_domain; - s32 fence_reg:I915_MAX_NUM_FENCE_BITS; - u32 tiling:2; - u32 dirty:1; - u32 purgeable:1; - u32 userptr:1; - s32 engine:4; - u32 cache_level:3; - } *active_bo[I915_NUM_ENGINES], *pinned_bo; - u32 active_bo_count[I915_NUM_ENGINES], pinned_bo_count; - struct i915_address_space *active_vm[I915_NUM_ENGINES]; -}; - enum i915_cache_level { I915_CACHE_NONE = 0, I915_CACHE_LLC, /* also used for snoopable memory on non-LLC */ @@ -766,12 +603,16 @@ struct i915_psr { bool active; struct delayed_work work; unsigned busy_frontbuffer_bits; - bool psr2_support; - bool aux_frame_sync; + bool sink_psr2_support; bool link_standby; - bool y_cord_support; bool colorimetry_support; bool alpm; + bool has_hw_tracking; + bool psr2_enabled; + u8 sink_sync_latency; + bool debug; + ktime_t last_entry_attempt; + ktime_t last_exit; void (*enable_source)(struct intel_dp *, const struct intel_crtc_state *); @@ -1146,16 +987,6 @@ struct i915_gem_mm { u32 object_count; }; -struct drm_i915_error_state_buf { - struct drm_i915_private *i915; - unsigned bytes; - unsigned size; - int err; - u8 *buf; - loff_t start; - loff_t pos; -}; - #define I915_IDLE_ENGINES_TIMEOUT (200) /* in ms */ #define I915_RESET_TIMEOUT (10 * HZ) /* 10s */ @@ -1164,102 +995,6 @@ struct drm_i915_error_state_buf { #define I915_ENGINE_DEAD_TIMEOUT (4 * HZ) /* Seqno, head and subunits dead */ #define I915_SEQNO_DEAD_TIMEOUT (12 * HZ) /* Seqno dead with active head */ -struct i915_gpu_error { - /* For hangcheck timer */ -#define DRM_I915_HANGCHECK_PERIOD 1500 /* in ms */ -#define DRM_I915_HANGCHECK_JIFFIES msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD) - - struct delayed_work hangcheck_work; - - /* For reset and error_state handling. */ - spinlock_t lock; - /* Protected by the above dev->gpu_error.lock. */ - struct i915_gpu_state *first_error; - - atomic_t pending_fb_pin; - - unsigned long missed_irq_rings; - - /** - * State variable controlling the reset flow and count - * - * This is a counter which gets incremented when reset is triggered, - * - * Before the reset commences, the I915_RESET_BACKOFF bit is set - * meaning that any waiters holding onto the struct_mutex should - * relinquish the lock immediately in order for the reset to start. - * - * If reset is not completed succesfully, the I915_WEDGE bit is - * set meaning that hardware is terminally sour and there is no - * recovery. All waiters on the reset_queue will be woken when - * that happens. - * - * This counter is used by the wait_seqno code to notice that reset - * event happened and it needs to restart the entire ioctl (since most - * likely the seqno it waited for won't ever signal anytime soon). - * - * This is important for lock-free wait paths, where no contended lock - * naturally enforces the correct ordering between the bail-out of the - * waiter and the gpu reset work code. - */ - unsigned long reset_count; - - /** - * flags: Control various stages of the GPU reset - * - * #I915_RESET_BACKOFF - When we start a reset, we want to stop any - * other users acquiring the struct_mutex. To do this we set the - * #I915_RESET_BACKOFF bit in the error flags when we detect a reset - * and then check for that bit before acquiring the struct_mutex (in - * i915_mutex_lock_interruptible()?). I915_RESET_BACKOFF serves a - * secondary role in preventing two concurrent global reset attempts. - * - * #I915_RESET_HANDOFF - To perform the actual GPU reset, we need the - * struct_mutex. We try to acquire the struct_mutex in the reset worker, - * but it may be held by some long running waiter (that we cannot - * interrupt without causing trouble). Once we are ready to do the GPU - * reset, we set the I915_RESET_HANDOFF bit and wakeup any waiters. If - * they already hold the struct_mutex and want to participate they can - * inspect the bit and do the reset directly, otherwise the worker - * waits for the struct_mutex. - * - * #I915_RESET_ENGINE[num_engines] - Since the driver doesn't need to - * acquire the struct_mutex to reset an engine, we need an explicit - * flag to prevent two concurrent reset attempts in the same engine. - * As the number of engines continues to grow, allocate the flags from - * the most significant bits. - * - * #I915_WEDGED - If reset fails and we can no longer use the GPU, - * we set the #I915_WEDGED bit. Prior to command submission, e.g. - * i915_request_alloc(), this bit is checked and the sequence - * aborted (with -EIO reported to userspace) if set. - */ - unsigned long flags; -#define I915_RESET_BACKOFF 0 -#define I915_RESET_HANDOFF 1 -#define I915_RESET_MODESET 2 -#define I915_WEDGED (BITS_PER_LONG - 1) -#define I915_RESET_ENGINE (I915_WEDGED - I915_NUM_ENGINES) - - /** Number of times an engine has been reset */ - u32 reset_engine_count[I915_NUM_ENGINES]; - - /** - * Waitqueue to signal when a hang is detected. Used to for waiters - * to release the struct_mutex for the reset to procede. - */ - wait_queue_head_t wait_queue; - - /** - * Waitqueue to signal when the reset has completed. Used by clients - * that wait for dev_priv->mm.wedged to settle. - */ - wait_queue_head_t reset_queue; - - /* For missed irq/seqno simulation. */ - unsigned long test_irq_rings; -}; - enum modeset_restore { MODESET_ON_LID_OPEN, MODESET_DONE, @@ -1338,6 +1073,7 @@ struct intel_vbt_data { } edp; struct { + bool enable; bool full_link; bool require_aux_wakeup; int idle_frames; @@ -1451,11 +1187,13 @@ static inline bool skl_ddb_entry_equal(const struct skl_ddb_entry *e1, } struct skl_ddb_allocation { - struct skl_ddb_entry plane[I915_MAX_PIPES][I915_MAX_PLANES]; /* packed/uv */ - struct skl_ddb_entry y_plane[I915_MAX_PIPES][I915_MAX_PLANES]; + /* packed/y */ + struct skl_ddb_entry plane[I915_MAX_PIPES][I915_MAX_PLANES]; + struct skl_ddb_entry uv_plane[I915_MAX_PIPES][I915_MAX_PLANES]; + u8 enabled_slices; /* GEN11 has configurable 2 slices */ }; -struct skl_wm_values { +struct skl_ddb_values { unsigned dirty_pipes; struct skl_ddb_allocation ddb; }; @@ -1470,6 +1208,7 @@ struct skl_wm_level { struct skl_wm_params { bool x_tiled, y_tiled; bool rc_surface; + bool is_planar; uint32_t width; uint8_t cpp; uint32_t plane_pixel_rate; @@ -1564,7 +1303,6 @@ struct i915_wa_reg { struct i915_workarounds { struct i915_wa_reg reg[I915_MAX_WA_REGS]; u32 count; - u32 hw_whitelist_count[I915_NUM_ENGINES]; }; struct i915_virtual_gpu { @@ -1860,6 +1598,8 @@ struct drm_i915_private { struct intel_gvt *gvt; + struct intel_wopcm wopcm; + struct intel_huc huc; struct intel_guc guc; @@ -2152,7 +1892,7 @@ struct drm_i915_private { /* current hardware state */ union { struct ilk_wm_values hw; - struct skl_wm_values skl_hw; + struct skl_ddb_values skl_hw; struct vlv_wm_values vlv; struct g4x_wm_values g4x; }; @@ -2321,8 +2061,11 @@ struct drm_i915_private { void (*cleanup_engine)(struct intel_engine_cs *engine); struct list_head timelines; - struct i915_gem_timeline global_timeline; + + struct list_head active_rings; + struct list_head closed_vma; u32 active_requests; + u32 request_serial; /** * Is the GPU currently considered idle, or busy executing @@ -2392,6 +2135,11 @@ static inline struct drm_i915_private *kdev_to_i915(struct device *kdev) return to_i915(dev_get_drvdata(kdev)); } +static inline struct drm_i915_private *wopcm_to_i915(struct intel_wopcm *wopcm) +{ + return container_of(wopcm, struct drm_i915_private, wopcm); +} + static inline struct drm_i915_private *guc_to_i915(struct intel_guc *guc) { return container_of(guc, struct drm_i915_private, guc); @@ -2411,8 +2159,10 @@ static inline struct drm_i915_private *huc_to_i915(struct intel_huc *huc) /* Iterator over subset of engines selected by mask */ #define for_each_engine_masked(engine__, dev_priv__, mask__, tmp__) \ - for (tmp__ = mask__ & INTEL_INFO(dev_priv__)->ring_mask; \ - tmp__ ? (engine__ = (dev_priv__)->engine[__mask_next_bit(tmp__)]), 1 : 0; ) + for ((tmp__) = (mask__) & INTEL_INFO(dev_priv__)->ring_mask; \ + (tmp__) ? \ + ((engine__) = (dev_priv__)->engine[__mask_next_bit(tmp__)]), 1 : \ + 0;) enum hdmi_force_audio { HDMI_AUDIO_OFF_DVI = -2, /* no aux data for HDMI-DVI converter */ @@ -2720,6 +2470,15 @@ intel_info(const struct drm_i915_private *dev_priv) #define IS_CNL_REVID(p, since, until) \ (IS_CANNONLAKE(p) && IS_REVID(p, since, until)) +#define ICL_REVID_A0 0x0 +#define ICL_REVID_A2 0x1 +#define ICL_REVID_B0 0x3 +#define ICL_REVID_B2 0x4 +#define ICL_REVID_C0 0x5 + +#define IS_ICL_REVID(p, since, until) \ + (IS_ICELAKE(p) && IS_REVID(p, since, until)) + /* * The genX designation typically refers to the render engine, so render * capability related checks should use IS_GEN, while display and other checks @@ -2963,10 +2722,11 @@ extern void i915_driver_unload(struct drm_device *dev); extern int intel_gpu_reset(struct drm_i915_private *dev_priv, u32 engine_mask); extern bool intel_has_gpu_reset(struct drm_i915_private *dev_priv); -#define I915_RESET_QUIET BIT(0) -extern void i915_reset(struct drm_i915_private *i915, unsigned int flags); +extern void i915_reset(struct drm_i915_private *i915, + unsigned int stalled_mask, + const char *reason); extern int i915_reset_engine(struct intel_engine_cs *engine, - unsigned int flags); + const char *reason); extern bool intel_has_reset_engine(struct drm_i915_private *dev_priv); extern int intel_reset_guc(struct drm_i915_private *dev_priv); @@ -3014,10 +2774,12 @@ static inline void i915_queue_hangcheck(struct drm_i915_private *dev_priv) &dev_priv->gpu_error.hangcheck_work, delay); } -__printf(3, 4) +__printf(4, 5) void i915_handle_error(struct drm_i915_private *dev_priv, u32 engine_mask, + unsigned long flags, const char *fmt, ...); +#define I915_ERROR_CAPTURE BIT(0) extern void intel_irq_init(struct drm_i915_private *dev_priv); extern void intel_irq_fini(struct drm_i915_private *dev_priv); @@ -3132,8 +2894,8 @@ int i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data, int i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); void i915_gem_sanitize(struct drm_i915_private *i915); -int i915_gem_load_init(struct drm_i915_private *dev_priv); -void i915_gem_load_cleanup(struct drm_i915_private *dev_priv); +int i915_gem_init_early(struct drm_i915_private *dev_priv); +void i915_gem_cleanup_early(struct drm_i915_private *dev_priv); void i915_gem_load_init_fences(struct drm_i915_private *dev_priv); int i915_gem_freeze(struct drm_i915_private *dev_priv); int i915_gem_freeze_late(struct drm_i915_private *dev_priv); @@ -3388,13 +3150,15 @@ static inline u32 i915_reset_engine_count(struct i915_gpu_error *error, struct i915_request * i915_gem_reset_prepare_engine(struct intel_engine_cs *engine); int i915_gem_reset_prepare(struct drm_i915_private *dev_priv); -void i915_gem_reset(struct drm_i915_private *dev_priv); +void i915_gem_reset(struct drm_i915_private *dev_priv, + unsigned int stalled_mask); void i915_gem_reset_finish_engine(struct intel_engine_cs *engine); void i915_gem_reset_finish(struct drm_i915_private *dev_priv); void i915_gem_set_wedged(struct drm_i915_private *dev_priv); bool i915_gem_unset_wedged(struct drm_i915_private *dev_priv); void i915_gem_reset_engine(struct intel_engine_cs *engine, - struct i915_request *request); + struct i915_request *request, + bool stalled); void i915_gem_init_mmio(struct drm_i915_private *i915); int __must_check i915_gem_init(struct drm_i915_private *dev_priv); @@ -3412,7 +3176,7 @@ int i915_gem_object_wait(struct drm_i915_gem_object *obj, struct intel_rps_client *rps); int i915_gem_object_wait_priority(struct drm_i915_gem_object *obj, unsigned int flags, - int priority); + const struct i915_sched_attr *attr); #define I915_PRIORITY_DISPLAY I915_PRIORITY_MAX int __must_check @@ -3481,16 +3245,6 @@ i915_gem_context_lookup(struct drm_i915_file_private *file_priv, u32 id) return ctx; } -static inline struct intel_timeline * -i915_gem_context_lookup_timeline(struct i915_gem_context *ctx, - struct intel_engine_cs *engine) -{ - struct i915_address_space *vm; - - vm = ctx->ppgtt ? &ctx->ppgtt->base : &ctx->i915->ggtt.base; - return &vm->timeline.engine[engine->id]; -} - int i915_perf_open_ioctl(struct drm_device *dev, void *data, struct drm_file *file); int i915_perf_add_config_ioctl(struct drm_device *dev, void *data, @@ -3589,64 +3343,6 @@ static inline int i915_debugfs_connector_add(struct drm_connector *connector) static inline void intel_display_crc_init(struct drm_i915_private *dev_priv) {} #endif -/* i915_gpu_error.c */ -#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR) - -__printf(2, 3) -void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...); -int i915_error_state_to_str(struct drm_i915_error_state_buf *estr, - const struct i915_gpu_state *gpu); -int i915_error_state_buf_init(struct drm_i915_error_state_buf *eb, - struct drm_i915_private *i915, - size_t count, loff_t pos); -static inline void i915_error_state_buf_release( - struct drm_i915_error_state_buf *eb) -{ - kfree(eb->buf); -} - -struct i915_gpu_state *i915_capture_gpu_state(struct drm_i915_private *i915); -void i915_capture_error_state(struct drm_i915_private *dev_priv, - u32 engine_mask, - const char *error_msg); - -static inline struct i915_gpu_state * -i915_gpu_state_get(struct i915_gpu_state *gpu) -{ - kref_get(&gpu->ref); - return gpu; -} - -void __i915_gpu_state_free(struct kref *kref); -static inline void i915_gpu_state_put(struct i915_gpu_state *gpu) -{ - if (gpu) - kref_put(&gpu->ref, __i915_gpu_state_free); -} - -struct i915_gpu_state *i915_first_error_state(struct drm_i915_private *i915); -void i915_reset_error_state(struct drm_i915_private *i915); - -#else - -static inline void i915_capture_error_state(struct drm_i915_private *dev_priv, - u32 engine_mask, - const char *error_msg) -{ -} - -static inline struct i915_gpu_state * -i915_first_error_state(struct drm_i915_private *i915) -{ - return NULL; -} - -static inline void i915_reset_error_state(struct drm_i915_private *i915) -{ -} - -#endif - const char *i915_cache_level_str(struct drm_i915_private *i915, int type); /* i915_cmd_parser.c */ diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 7b5a9d7..0a20701 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -35,6 +35,7 @@ #include "intel_drv.h" #include "intel_frontbuffer.h" #include "intel_mocs.h" +#include "intel_workarounds.h" #include "i915_gemfs.h" #include <linux/dma-fence-array.h> #include <linux/kthread.h> @@ -136,6 +137,102 @@ int i915_mutex_lock_interruptible(struct drm_device *dev) return 0; } +static u32 __i915_gem_park(struct drm_i915_private *i915) +{ + lockdep_assert_held(&i915->drm.struct_mutex); + GEM_BUG_ON(i915->gt.active_requests); + GEM_BUG_ON(!list_empty(&i915->gt.active_rings)); + + if (!i915->gt.awake) + return I915_EPOCH_INVALID; + + GEM_BUG_ON(i915->gt.epoch == I915_EPOCH_INVALID); + + /* + * Be paranoid and flush a concurrent interrupt to make sure + * we don't reactivate any irq tasklets after parking. + * + * FIXME: Note that even though we have waited for execlists to be idle, + * there may still be an in-flight interrupt even though the CSB + * is now empty. synchronize_irq() makes sure that a residual interrupt + * is completed before we continue, but it doesn't prevent the HW from + * raising a spurious interrupt later. To complete the shield we should + * coordinate disabling the CS irq with flushing the interrupts. + */ + synchronize_irq(i915->drm.irq); + + intel_engines_park(i915); + i915_timelines_park(i915); + + i915_pmu_gt_parked(i915); + i915_vma_parked(i915); + + i915->gt.awake = false; + + if (INTEL_GEN(i915) >= 6) + gen6_rps_idle(i915); + + intel_display_power_put(i915, POWER_DOMAIN_GT_IRQ); + + intel_runtime_pm_put(i915); + + return i915->gt.epoch; +} + +void i915_gem_park(struct drm_i915_private *i915) +{ + lockdep_assert_held(&i915->drm.struct_mutex); + GEM_BUG_ON(i915->gt.active_requests); + + if (!i915->gt.awake) + return; + + /* Defer the actual call to __i915_gem_park() to prevent ping-pongs */ + mod_delayed_work(i915->wq, &i915->gt.idle_work, msecs_to_jiffies(100)); +} + +void i915_gem_unpark(struct drm_i915_private *i915) +{ + lockdep_assert_held(&i915->drm.struct_mutex); + GEM_BUG_ON(!i915->gt.active_requests); + + if (i915->gt.awake) + return; + + intel_runtime_pm_get_noresume(i915); + + /* + * It seems that the DMC likes to transition between the DC states a lot + * when there are no connected displays (no active power domains) during + * command submission. + * + * This activity has negative impact on the performance of the chip with + * huge latencies observed in the interrupt handler and elsewhere. + * + * Work around it by grabbing a GT IRQ power domain whilst there is any + * GT activity, preventing any DC state transitions. + */ + intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ); + + i915->gt.awake = true; + if (unlikely(++i915->gt.epoch == 0)) /* keep 0 as invalid */ + i915->gt.epoch = 1; + + intel_enable_gt_powersave(i915); + i915_update_gfx_val(i915); + if (INTEL_GEN(i915) >= 6) + gen6_rps_busy(i915); + i915_pmu_gt_unparked(i915); + + intel_engines_unpark(i915); + + i915_queue_hangcheck(i915); + + queue_delayed_work(i915->wq, + &i915->gt.retire_work, + round_jiffies_up_relative(HZ)); +} + int i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data, struct drm_file *file) @@ -469,7 +566,8 @@ i915_gem_object_wait_reservation(struct reservation_object *resv, return timeout; } -static void __fence_set_priority(struct dma_fence *fence, int prio) +static void __fence_set_priority(struct dma_fence *fence, + const struct i915_sched_attr *attr) { struct i915_request *rq; struct intel_engine_cs *engine; @@ -480,13 +578,16 @@ static void __fence_set_priority(struct dma_fence *fence, int prio) rq = to_request(fence); engine = rq->engine; - rcu_read_lock(); + local_bh_disable(); + rcu_read_lock(); /* RCU serialisation for set-wedged protection */ if (engine->schedule) - engine->schedule(rq, prio); + engine->schedule(rq, attr); rcu_read_unlock(); + local_bh_enable(); /* kick the tasklets if queues were reprioritised */ } -static void fence_set_priority(struct dma_fence *fence, int prio) +static void fence_set_priority(struct dma_fence *fence, + const struct i915_sched_attr *attr) { /* Recurse once into a fence-array */ if (dma_fence_is_array(fence)) { @@ -494,16 +595,16 @@ static void fence_set_priority(struct dma_fence *fence, int prio) int i; for (i = 0; i < array->num_fences; i++) - __fence_set_priority(array->fences[i], prio); + __fence_set_priority(array->fences[i], attr); } else { - __fence_set_priority(fence, prio); + __fence_set_priority(fence, attr); } } int i915_gem_object_wait_priority(struct drm_i915_gem_object *obj, unsigned int flags, - int prio) + const struct i915_sched_attr *attr) { struct dma_fence *excl; @@ -518,7 +619,7 @@ i915_gem_object_wait_priority(struct drm_i915_gem_object *obj, return ret; for (i = 0; i < count; i++) { - fence_set_priority(shared[i], prio); + fence_set_priority(shared[i], attr); dma_fence_put(shared[i]); } @@ -528,7 +629,7 @@ i915_gem_object_wait_priority(struct drm_i915_gem_object *obj, } if (excl) { - fence_set_priority(excl, prio); + fence_set_priority(excl, attr); dma_fence_put(excl); } return 0; @@ -2879,8 +2980,8 @@ i915_gem_find_active_request(struct intel_engine_cs *engine) * extra delay for a recent interrupt is pointless. Hence, we do * not need an engine->irq_seqno_barrier() before the seqno reads. */ - spin_lock_irqsave(&engine->timeline->lock, flags); - list_for_each_entry(request, &engine->timeline->requests, link) { + spin_lock_irqsave(&engine->timeline.lock, flags); + list_for_each_entry(request, &engine->timeline.requests, link) { if (__i915_request_completed(request, request->global_seqno)) continue; @@ -2891,25 +2992,11 @@ i915_gem_find_active_request(struct intel_engine_cs *engine) active = request; break; } - spin_unlock_irqrestore(&engine->timeline->lock, flags); + spin_unlock_irqrestore(&engine->timeline.lock, flags); return active; } -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; -} - /* * Ensure irq handler finishes, and not run again. * Also return the active request so that we only search for it once. @@ -2998,6 +3085,7 @@ int i915_gem_reset_prepare(struct drm_i915_private *dev_priv) } i915_gem_revoke_fences(dev_priv); + intel_uc_sanitize(dev_priv); return err; } @@ -3025,15 +3113,15 @@ static void engine_skip_context(struct i915_request *request) { struct intel_engine_cs *engine = request->engine; struct i915_gem_context *hung_ctx = request->ctx; - struct intel_timeline *timeline; + struct i915_timeline *timeline = request->timeline; unsigned long flags; - timeline = i915_gem_context_lookup_timeline(hung_ctx, engine); + GEM_BUG_ON(timeline == &engine->timeline); - spin_lock_irqsave(&engine->timeline->lock, flags); - spin_lock(&timeline->lock); + spin_lock_irqsave(&engine->timeline.lock, flags); + spin_lock_nested(&timeline->lock, SINGLE_DEPTH_NESTING); - list_for_each_entry_continue(request, &engine->timeline->requests, link) + list_for_each_entry_continue(request, &engine->timeline.requests, link) if (request->ctx == hung_ctx) skip_request(request); @@ -3041,13 +3129,14 @@ static void engine_skip_context(struct i915_request *request) skip_request(request); spin_unlock(&timeline->lock); - spin_unlock_irqrestore(&engine->timeline->lock, flags); + spin_unlock_irqrestore(&engine->timeline.lock, flags); } /* Returns the request if it was guilty of the hang */ static struct i915_request * i915_gem_reset_request(struct intel_engine_cs *engine, - struct i915_request *request) + struct i915_request *request, + bool stalled) { /* The guilty request will get skipped on a hung engine. * @@ -3070,7 +3159,15 @@ i915_gem_reset_request(struct intel_engine_cs *engine, * subsequent hangs. */ - if (engine_stalled(engine)) { + if (i915_request_completed(request)) { + GEM_TRACE("%s pardoned global=%d (fence %llx:%d), current %d\n", + engine->name, request->global_seqno, + request->fence.context, request->fence.seqno, + intel_engine_get_seqno(engine)); + stalled = false; + } + + if (stalled) { i915_gem_context_mark_guilty(request->ctx); skip_request(request); @@ -3089,11 +3186,11 @@ i915_gem_reset_request(struct intel_engine_cs *engine, dma_fence_set_error(&request->fence, -EAGAIN); /* Rewind the engine to replay the incomplete rq */ - spin_lock_irq(&engine->timeline->lock); + spin_lock_irq(&engine->timeline.lock); request = list_prev_entry(request, link); - if (&request->link == &engine->timeline->requests) + if (&request->link == &engine->timeline.requests) request = NULL; - spin_unlock_irq(&engine->timeline->lock); + spin_unlock_irq(&engine->timeline.lock); } } @@ -3101,7 +3198,8 @@ i915_gem_reset_request(struct intel_engine_cs *engine, } void i915_gem_reset_engine(struct intel_engine_cs *engine, - struct i915_request *request) + struct i915_request *request, + bool stalled) { /* * Make sure this write is visible before we re-enable the interrupt @@ -3111,7 +3209,7 @@ void i915_gem_reset_engine(struct intel_engine_cs *engine, smp_store_mb(engine->irq_posted, 0); if (request) - request = i915_gem_reset_request(engine, request); + request = i915_gem_reset_request(engine, request, stalled); if (request) { DRM_DEBUG_DRIVER("resetting %s to restart from tail of request 0x%x\n", @@ -3122,7 +3220,8 @@ void i915_gem_reset_engine(struct intel_engine_cs *engine, engine->reset_hw(engine, request); } -void i915_gem_reset(struct drm_i915_private *dev_priv) +void i915_gem_reset(struct drm_i915_private *dev_priv, + unsigned int stalled_mask) { struct intel_engine_cs *engine; enum intel_engine_id id; @@ -3134,10 +3233,12 @@ void i915_gem_reset(struct drm_i915_private *dev_priv) for_each_engine(engine, dev_priv, id) { struct i915_gem_context *ctx; - i915_gem_reset_engine(engine, engine->hangcheck.active_request); + i915_gem_reset_engine(engine, + engine->hangcheck.active_request, + stalled_mask & ENGINE_MASK(id)); ctx = fetch_and_zero(&engine->last_retired_context); if (ctx) - engine->context_unpin(engine, ctx); + intel_context_unpin(ctx, engine); /* * Ostensibily, we always want a context loaded for powersaving, @@ -3160,13 +3261,6 @@ void i915_gem_reset(struct drm_i915_private *dev_priv) } i915_gem_restore_fences(dev_priv); - - if (dev_priv->gt.awake) { - intel_sanitize_gt_powersave(dev_priv); - intel_enable_gt_powersave(dev_priv); - if (INTEL_GEN(dev_priv) >= 6) - gen6_rps_busy(dev_priv); - } } void i915_gem_reset_finish_engine(struct intel_engine_cs *engine) @@ -3192,6 +3286,9 @@ void i915_gem_reset_finish(struct drm_i915_private *dev_priv) static void nop_submit_request(struct i915_request *request) { + GEM_TRACE("%s fence %llx:%d -> -EIO\n", + request->engine->name, + request->fence.context, request->fence.seqno); dma_fence_set_error(&request->fence, -EIO); i915_request_submit(request); @@ -3201,12 +3298,15 @@ static void nop_complete_submit_request(struct i915_request *request) { unsigned long flags; + GEM_TRACE("%s fence %llx:%d -> -EIO\n", + request->engine->name, + request->fence.context, request->fence.seqno); dma_fence_set_error(&request->fence, -EIO); - spin_lock_irqsave(&request->engine->timeline->lock, flags); + spin_lock_irqsave(&request->engine->timeline.lock, flags); __i915_request_submit(request); intel_engine_init_global_seqno(request->engine, request->global_seqno); - spin_unlock_irqrestore(&request->engine->timeline->lock, flags); + spin_unlock_irqrestore(&request->engine->timeline.lock, flags); } void i915_gem_set_wedged(struct drm_i915_private *i915) @@ -3214,7 +3314,9 @@ void i915_gem_set_wedged(struct drm_i915_private *i915) struct intel_engine_cs *engine; enum intel_engine_id id; - if (drm_debug & DRM_UT_DRIVER) { + GEM_TRACE("start\n"); + + if (GEM_SHOW_DEBUG()) { struct drm_printer p = drm_debug_printer(__func__); for_each_engine(engine, i915, id) @@ -3237,6 +3339,9 @@ void i915_gem_set_wedged(struct drm_i915_private *i915) } i915->caps.scheduler = 0; + /* Even if the GPU reset fails, it should still stop the engines */ + intel_gpu_reset(i915, ALL_ENGINES); + /* * Make sure no one is running the old callback before we proceed with * cancelling requests and resetting the completion tracking. Otherwise @@ -3270,27 +3375,31 @@ void i915_gem_set_wedged(struct drm_i915_private *i915) * (lockless) lookup doesn't try and wait upon the request as we * reset it. */ - spin_lock_irqsave(&engine->timeline->lock, flags); + spin_lock_irqsave(&engine->timeline.lock, flags); intel_engine_init_global_seqno(engine, intel_engine_last_submit(engine)); - spin_unlock_irqrestore(&engine->timeline->lock, flags); + spin_unlock_irqrestore(&engine->timeline.lock, flags); i915_gem_reset_finish_engine(engine); } + GEM_TRACE("end\n"); + wake_up_all(&i915->gpu_error.reset_queue); } bool i915_gem_unset_wedged(struct drm_i915_private *i915) { - struct i915_gem_timeline *tl; - int i; + struct i915_timeline *tl; lockdep_assert_held(&i915->drm.struct_mutex); if (!test_bit(I915_WEDGED, &i915->gpu_error.flags)) return true; - /* Before unwedging, make sure that all pending operations + GEM_TRACE("start\n"); + + /* + * Before unwedging, make sure that all pending operations * are flushed and errored out - we may have requests waiting upon * third party fences. We marked all inflight requests as EIO, and * every execbuf since returned EIO, for consistency we want all @@ -3300,31 +3409,33 @@ bool i915_gem_unset_wedged(struct drm_i915_private *i915) * No more can be submitted until we reset the wedged bit. */ list_for_each_entry(tl, &i915->gt.timelines, link) { - for (i = 0; i < ARRAY_SIZE(tl->engine); i++) { - struct i915_request *rq; + struct i915_request *rq; - rq = i915_gem_active_peek(&tl->engine[i].last_request, - &i915->drm.struct_mutex); - if (!rq) - continue; + rq = i915_gem_active_peek(&tl->last_request, + &i915->drm.struct_mutex); + if (!rq) + continue; - /* We can't use our normal waiter as we want to - * avoid recursively trying to handle the current - * reset. The basic dma_fence_default_wait() installs - * a callback for dma_fence_signal(), which is - * triggered by our nop handler (indirectly, the - * callback enables the signaler thread which is - * woken by the nop_submit_request() advancing the seqno - * and when the seqno passes the fence, the signaler - * then signals the fence waking us up). - */ - if (dma_fence_default_wait(&rq->fence, true, - MAX_SCHEDULE_TIMEOUT) < 0) - return false; - } + /* + * We can't use our normal waiter as we want to + * avoid recursively trying to handle the current + * reset. The basic dma_fence_default_wait() installs + * a callback for dma_fence_signal(), which is + * triggered by our nop handler (indirectly, the + * callback enables the signaler thread which is + * woken by the nop_submit_request() advancing the seqno + * and when the seqno passes the fence, the signaler + * then signals the fence waking us up). + */ + if (dma_fence_default_wait(&rq->fence, true, + MAX_SCHEDULE_TIMEOUT) < 0) + return false; } + i915_retire_requests(i915); + GEM_BUG_ON(i915->gt.active_requests); - /* Undo nop_submit_request. We prevent all new i915 requests from + /* + * Undo nop_submit_request. We prevent all new i915 requests from * being queued (by disallowing execbuf whilst wedged) so having * waited for all active requests above, we know the system is idle * and do not have to worry about a thread being inside @@ -3335,6 +3446,8 @@ bool i915_gem_unset_wedged(struct drm_i915_private *i915) intel_engines_reset_default_submission(i915); i915_gem_contexts_lost(i915); + GEM_TRACE("end\n"); + smp_mb__before_atomic(); /* complete takeover before enabling execbuf */ clear_bit(I915_WEDGED, &i915->gpu_error.flags); @@ -3473,36 +3586,9 @@ i915_gem_idle_work_handler(struct work_struct *work) if (new_requests_since_last_retire(dev_priv)) goto out_unlock; - /* - * Be paranoid and flush a concurrent interrupt to make sure - * we don't reactivate any irq tasklets after parking. - * - * FIXME: Note that even though we have waited for execlists to be idle, - * there may still be an in-flight interrupt even though the CSB - * is now empty. synchronize_irq() makes sure that a residual interrupt - * is completed before we continue, but it doesn't prevent the HW from - * raising a spurious interrupt later. To complete the shield we should - * coordinate disabling the CS irq with flushing the interrupts. - */ - synchronize_irq(dev_priv->drm.irq); - - intel_engines_park(dev_priv); - i915_gem_timelines_park(dev_priv); + epoch = __i915_gem_park(dev_priv); - i915_pmu_gt_parked(dev_priv); - - GEM_BUG_ON(!dev_priv->gt.awake); - dev_priv->gt.awake = false; - epoch = dev_priv->gt.epoch; - GEM_BUG_ON(epoch == I915_EPOCH_INVALID); rearm_hangcheck = false; - - if (INTEL_GEN(dev_priv) >= 6) - gen6_rps_idle(dev_priv); - - intel_display_power_put(dev_priv, POWER_DOMAIN_GT_IRQ); - - intel_runtime_pm_put(dev_priv); out_unlock: mutex_unlock(&dev_priv->drm.struct_mutex); @@ -3648,17 +3734,9 @@ i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file) return ret; } -static int wait_for_timeline(struct i915_gem_timeline *tl, unsigned int flags) +static int wait_for_timeline(struct i915_timeline *tl, unsigned int flags) { - int ret, i; - - for (i = 0; i < ARRAY_SIZE(tl->engine); i++) { - ret = i915_gem_active_wait(&tl->engine[i].last_request, flags); - if (ret) - return ret; - } - - return 0; + return i915_gem_active_wait(&tl->last_request, flags); } static int wait_for_engines(struct drm_i915_private *i915) @@ -3666,16 +3744,7 @@ static int wait_for_engines(struct drm_i915_private *i915) if (wait_for(intel_engines_are_idle(i915), I915_IDLE_ENGINES_TIMEOUT)) { dev_err(i915->drm.dev, "Failed to idle engines, declaring wedged!\n"); - if (drm_debug & DRM_UT_DRIVER) { - struct drm_printer p = drm_debug_printer(__func__); - struct intel_engine_cs *engine; - enum intel_engine_id id; - - for_each_engine(engine, i915, id) - intel_engine_dump(engine, &p, - "%s\n", engine->name); - } - + GEM_TRACE_DUMP(); i915_gem_set_wedged(i915); return -EIO; } @@ -3685,30 +3754,37 @@ static int wait_for_engines(struct drm_i915_private *i915) int i915_gem_wait_for_idle(struct drm_i915_private *i915, unsigned int flags) { - int ret; - /* If the device is asleep, we have no requests outstanding */ if (!READ_ONCE(i915->gt.awake)) return 0; if (flags & I915_WAIT_LOCKED) { - struct i915_gem_timeline *tl; + struct i915_timeline *tl; + int err; lockdep_assert_held(&i915->drm.struct_mutex); list_for_each_entry(tl, &i915->gt.timelines, link) { - ret = wait_for_timeline(tl, flags); - if (ret) - return ret; + err = wait_for_timeline(tl, flags); + if (err) + return err; } i915_retire_requests(i915); - ret = wait_for_engines(i915); + return wait_for_engines(i915); } else { - ret = wait_for_timeline(&i915->gt.global_timeline, flags); - } + struct intel_engine_cs *engine; + enum intel_engine_id id; + int err; - return ret; + for_each_engine(engine, i915, id) { + err = wait_for_timeline(&engine->timeline, flags); + if (err) + return err; + } + + return 0; + } } static void __i915_gem_object_flush_for_display(struct drm_i915_gem_object *obj) @@ -4088,9 +4164,10 @@ out: } /* - * Prepare buffer for display plane (scanout, cursors, etc). - * Can be called from an uninterruptible phase (modesetting) and allows - * any flushes to be pipelined (for pageflips). + * Prepare buffer for display plane (scanout, cursors, etc). Can be called from + * an uninterruptible phase (modesetting) and allows any flushes to be pipelined + * (for pageflips). We only flush the caches while preparing the buffer for + * display, the callers are responsible for frontbuffer flush. */ struct i915_vma * i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj, @@ -4146,9 +4223,7 @@ i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj, vma->display_alignment = max_t(u64, vma->display_alignment, alignment); - /* Treat this as an end-of-frame, like intel_user_framebuffer_dirty() */ __i915_gem_object_flush_for_display(obj); - intel_fb_obj_flush(obj, ORIGIN_DIRTYFB); /* It should now be out of any other write domains, and we can update * the domain values for our changes. @@ -4723,7 +4798,7 @@ static void __i915_gem_free_objects(struct drm_i915_private *i915, &obj->vma_list, obj_link) { GEM_BUG_ON(i915_vma_is_active(vma)); vma->flags &= ~I915_VMA_PIN_MASK; - i915_vma_close(vma); + i915_vma_destroy(vma); } GEM_BUG_ON(!list_empty(&obj->vma_list)); GEM_BUG_ON(!RB_EMPTY_ROOT(&obj->vma_tree)); @@ -4878,7 +4953,7 @@ static void assert_kernel_context_is_current(struct drm_i915_private *i915) enum intel_engine_id id; for_each_engine(engine, i915, id) { - GEM_BUG_ON(__i915_gem_active_peek(&engine->timeline->last_request)); + GEM_BUG_ON(__i915_gem_active_peek(&engine->timeline.last_request)); GEM_BUG_ON(engine->last_retired_context != kernel_context); } } @@ -4973,6 +5048,7 @@ int i915_gem_suspend(struct drm_i915_private *dev_priv) * machines is a good idea, we don't - just in case it leaves the * machine in an unusable condition. */ + intel_uc_sanitize(dev_priv); i915_gem_sanitize(dev_priv); intel_runtime_pm_put(dev_priv); @@ -5118,6 +5194,8 @@ int i915_gem_init_hw(struct drm_i915_private *dev_priv) } } + intel_gt_workarounds_apply(dev_priv); + i915_gem_init_swizzling(dev_priv); /* @@ -5140,6 +5218,12 @@ int i915_gem_init_hw(struct drm_i915_private *dev_priv) goto out; } + ret = intel_wopcm_init_hw(&dev_priv->wopcm); + if (ret) { + DRM_ERROR("Enabling WOPCM failed (%d)\n", ret); + goto out; + } + /* We can't enable contexts until all firmware is loaded */ ret = intel_uc_init_hw(dev_priv); if (ret) { @@ -5207,7 +5291,7 @@ static int __intel_engines_record_defaults(struct drm_i915_private *i915) for_each_engine(engine, i915, id) { struct i915_vma *state; - state = ctx->engine[id].state; + state = to_intel_context(ctx, engine)->state; if (!state) continue; @@ -5297,6 +5381,10 @@ int i915_gem_init(struct drm_i915_private *dev_priv) if (ret) return ret; + ret = intel_wopcm_init(&dev_priv->wopcm); + if (ret) + return ret; + ret = intel_uc_init_misc(dev_priv); if (ret) return ret; @@ -5478,8 +5566,7 @@ static void i915_gem_init__mm(struct drm_i915_private *i915) INIT_WORK(&i915->mm.free_work, __i915_gem_free_work); } -int -i915_gem_load_init(struct drm_i915_private *dev_priv) +int i915_gem_init_early(struct drm_i915_private *dev_priv) { int err = -ENOMEM; @@ -5512,12 +5599,9 @@ i915_gem_load_init(struct drm_i915_private *dev_priv) if (!dev_priv->priorities) goto err_dependencies; - mutex_lock(&dev_priv->drm.struct_mutex); INIT_LIST_HEAD(&dev_priv->gt.timelines); - err = i915_gem_timeline_init__global(dev_priv); - mutex_unlock(&dev_priv->drm.struct_mutex); - if (err) - goto err_priorities; + INIT_LIST_HEAD(&dev_priv->gt.active_rings); + INIT_LIST_HEAD(&dev_priv->gt.closed_vma); i915_gem_init__mm(dev_priv); @@ -5538,8 +5622,6 @@ i915_gem_load_init(struct drm_i915_private *dev_priv) return 0; -err_priorities: - kmem_cache_destroy(dev_priv->priorities); err_dependencies: kmem_cache_destroy(dev_priv->dependencies); err_requests: @@ -5554,17 +5636,13 @@ err_out: return err; } -void i915_gem_load_cleanup(struct drm_i915_private *dev_priv) +void i915_gem_cleanup_early(struct drm_i915_private *dev_priv) { i915_gem_drain_freed_objects(dev_priv); GEM_BUG_ON(!llist_empty(&dev_priv->mm.free_list)); GEM_BUG_ON(atomic_read(&dev_priv->mm.free_count)); WARN_ON(dev_priv->mm.object_count); - - mutex_lock(&dev_priv->drm.struct_mutex); - i915_gem_timeline_fini(&dev_priv->gt.global_timeline); WARN_ON(!list_empty(&dev_priv->gt.timelines)); - mutex_unlock(&dev_priv->drm.struct_mutex); kmem_cache_destroy(dev_priv->priorities); kmem_cache_destroy(dev_priv->dependencies); diff --git a/drivers/gpu/drm/i915/i915_gem.h b/drivers/gpu/drm/i915/i915_gem.h index f54c4ff..5259204 100644 --- a/drivers/gpu/drm/i915/i915_gem.h +++ b/drivers/gpu/drm/i915/i915_gem.h @@ -27,7 +27,12 @@ #include <linux/bug.h> +struct drm_i915_private; + #ifdef CONFIG_DRM_I915_DEBUG_GEM + +#define GEM_SHOW_DEBUG() (drm_debug & DRM_UT_DRIVER) + #define GEM_BUG_ON(condition) do { if (unlikely((condition))) { \ pr_err("%s:%d GEM_BUG_ON(%s)\n", \ __func__, __LINE__, __stringify(condition)); \ @@ -43,6 +48,9 @@ #define GEM_DEBUG_BUG_ON(expr) GEM_BUG_ON(expr) #else + +#define GEM_SHOW_DEBUG() (0) + #define GEM_BUG_ON(expr) BUILD_BUG_ON_INVALID(expr) #define GEM_WARN_ON(expr) (BUILD_BUG_ON_INVALID(expr), 0) @@ -53,10 +61,15 @@ #if IS_ENABLED(CONFIG_DRM_I915_TRACE_GEM) #define GEM_TRACE(...) trace_printk(__VA_ARGS__) +#define GEM_TRACE_DUMP() ftrace_dump(DUMP_ALL) #else #define GEM_TRACE(...) do { } while (0) +#define GEM_TRACE_DUMP() do { } while (0) #endif #define I915_NUM_ENGINES 8 +void i915_gem_park(struct drm_i915_private *i915); +void i915_gem_unpark(struct drm_i915_private *i915); + #endif /* __I915_GEM_H__ */ diff --git a/drivers/gpu/drm/i915/i915_gem_batch_pool.c b/drivers/gpu/drm/i915/i915_gem_batch_pool.c index d3cbe84..f3890b6 100644 --- a/drivers/gpu/drm/i915/i915_gem_batch_pool.c +++ b/drivers/gpu/drm/i915/i915_gem_batch_pool.c @@ -1,29 +1,11 @@ /* - * Copyright © 2014 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. + * SPDX-License-Identifier: MIT * + * Copyright © 2014-2018 Intel Corporation */ -#include "i915_drv.h" #include "i915_gem_batch_pool.h" +#include "i915_drv.h" /** * DOC: batch pool @@ -41,11 +23,11 @@ /** * i915_gem_batch_pool_init() - initialize a batch buffer pool - * @engine: the associated request submission engine * @pool: the batch buffer pool + * @engine: the associated request submission engine */ -void i915_gem_batch_pool_init(struct intel_engine_cs *engine, - struct i915_gem_batch_pool *pool) +void i915_gem_batch_pool_init(struct i915_gem_batch_pool *pool, + struct intel_engine_cs *engine) { int n; diff --git a/drivers/gpu/drm/i915/i915_gem_batch_pool.h b/drivers/gpu/drm/i915/i915_gem_batch_pool.h index 10d5ac4..56947da 100644 --- a/drivers/gpu/drm/i915/i915_gem_batch_pool.h +++ b/drivers/gpu/drm/i915/i915_gem_batch_pool.h @@ -1,31 +1,13 @@ /* - * Copyright © 2014 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. + * SPDX-License-Identifier: MIT * + * Copyright © 2014-2018 Intel Corporation */ #ifndef I915_GEM_BATCH_POOL_H #define I915_GEM_BATCH_POOL_H -#include "i915_drv.h" +#include <linux/types.h> struct intel_engine_cs; @@ -34,9 +16,8 @@ struct i915_gem_batch_pool { struct list_head cache_list[4]; }; -/* i915_gem_batch_pool.c */ -void i915_gem_batch_pool_init(struct intel_engine_cs *engine, - struct i915_gem_batch_pool *pool); +void i915_gem_batch_pool_init(struct i915_gem_batch_pool *pool, + struct intel_engine_cs *engine); void i915_gem_batch_pool_fini(struct i915_gem_batch_pool *pool); struct drm_i915_gem_object* i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool, size_t size); diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index f2cbea7..33f8a4b 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -90,6 +90,7 @@ #include <drm/i915_drm.h> #include "i915_drv.h" #include "i915_trace.h" +#include "intel_workarounds.h" #define ALL_L3_SLICES(dev) (1 << NUM_L3_SLICES(dev)) - 1 @@ -116,15 +117,15 @@ static void lut_close(struct i915_gem_context *ctx) static void i915_gem_context_free(struct i915_gem_context *ctx) { - int i; + unsigned int n; lockdep_assert_held(&ctx->i915->drm.struct_mutex); GEM_BUG_ON(!i915_gem_context_is_closed(ctx)); i915_ppgtt_put(ctx->ppgtt); - for (i = 0; i < I915_NUM_ENGINES; i++) { - struct intel_context *ce = &ctx->engine[i]; + for (n = 0; n < ARRAY_SIZE(ctx->__engine); n++) { + struct intel_context *ce = &ctx->__engine[n]; if (!ce->state) continue; @@ -280,7 +281,7 @@ __create_hw_context(struct drm_i915_private *dev_priv, kref_init(&ctx->ref); list_add_tail(&ctx->link, &dev_priv->contexts.list); ctx->i915 = dev_priv; - ctx->priority = I915_PRIORITY_NORMAL; + ctx->sched.priority = I915_PRIORITY_NORMAL; INIT_RADIX_TREE(&ctx->handles_vma, GFP_KERNEL); INIT_LIST_HEAD(&ctx->handles_list); @@ -318,12 +319,13 @@ __create_hw_context(struct drm_i915_private *dev_priv, ctx->desc_template = default_desc_template(dev_priv, dev_priv->mm.aliasing_ppgtt); - /* GuC requires the ring to be placed above GUC_WOPCM_TOP. If GuC is not + /* + * GuC requires the ring to be placed in Non-WOPCM memory. If GuC is not * present or not in use we still need a small bias as ring wraparound * at offset 0 sometimes hangs. No idea why. */ if (USES_GUC(dev_priv)) - ctx->ggtt_offset_bias = GUC_WOPCM_TOP; + ctx->ggtt_offset_bias = dev_priv->guc.ggtt_pin_bias; else ctx->ggtt_offset_bias = I915_GTT_PAGE_SIZE; @@ -429,7 +431,7 @@ i915_gem_context_create_kernel(struct drm_i915_private *i915, int prio) return ctx; i915_gem_context_clear_bannable(ctx); - ctx->priority = prio; + ctx->sched.priority = prio; ctx->ring_size = PAGE_SIZE; GEM_BUG_ON(!i915_gem_context_is_kernel(ctx)); @@ -458,11 +460,16 @@ static bool needs_preempt_context(struct drm_i915_private *i915) int i915_gem_contexts_init(struct drm_i915_private *dev_priv) { struct i915_gem_context *ctx; + int ret; /* Reassure ourselves we are only called once */ GEM_BUG_ON(dev_priv->kernel_context); GEM_BUG_ON(dev_priv->preempt_context); + ret = intel_ctx_workarounds_init(dev_priv); + if (ret) + return ret; + INIT_LIST_HEAD(&dev_priv->contexts.list); INIT_WORK(&dev_priv->contexts.free_work, contexts_free_worker); init_llist_head(&dev_priv->contexts.free_list); @@ -514,7 +521,7 @@ void i915_gem_contexts_lost(struct drm_i915_private *dev_priv) if (!engine->last_retired_context) continue; - engine->context_unpin(engine, engine->last_retired_context); + intel_context_unpin(engine->last_retired_context, engine); engine->last_retired_context = NULL; } } @@ -570,19 +577,29 @@ void i915_gem_context_close(struct drm_file *file) idr_destroy(&file_priv->context_idr); } -static bool engine_has_idle_kernel_context(struct intel_engine_cs *engine) +static struct i915_request * +last_request_on_engine(struct i915_timeline *timeline, + struct intel_engine_cs *engine) { - struct i915_gem_timeline *timeline; + struct i915_request *rq; - list_for_each_entry(timeline, &engine->i915->gt.timelines, link) { - struct intel_timeline *tl; + if (timeline == &engine->timeline) + return NULL; - if (timeline == &engine->i915->gt.global_timeline) - continue; + rq = i915_gem_active_raw(&timeline->last_request, + &engine->i915->drm.struct_mutex); + if (rq && rq->engine == engine) + return rq; + + return NULL; +} - tl = &timeline->engine[engine->id]; - if (i915_gem_active_peek(&tl->last_request, - &engine->i915->drm.struct_mutex)) +static bool engine_has_idle_kernel_context(struct intel_engine_cs *engine) +{ + struct i915_timeline *timeline; + + list_for_each_entry(timeline, &engine->i915->gt.timelines, link) { + if (last_request_on_engine(timeline, engine)) return false; } @@ -592,7 +609,7 @@ static bool engine_has_idle_kernel_context(struct intel_engine_cs *engine) int i915_gem_switch_to_kernel_context(struct drm_i915_private *dev_priv) { struct intel_engine_cs *engine; - struct i915_gem_timeline *timeline; + struct i915_timeline *timeline; enum intel_engine_id id; lockdep_assert_held(&dev_priv->drm.struct_mutex); @@ -612,11 +629,8 @@ int i915_gem_switch_to_kernel_context(struct drm_i915_private *dev_priv) /* Queue this switch after all other activity */ list_for_each_entry(timeline, &dev_priv->gt.timelines, link) { struct i915_request *prev; - struct intel_timeline *tl; - tl = &timeline->engine[engine->id]; - prev = i915_gem_active_raw(&tl->last_request, - &dev_priv->drm.struct_mutex); + prev = last_request_on_engine(timeline, engine); if (prev) i915_sw_fence_await_sw_fence_gfp(&rq->submit, &prev->submit, @@ -746,7 +760,7 @@ int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data, args->value = i915_gem_context_is_bannable(ctx); break; case I915_CONTEXT_PARAM_PRIORITY: - args->value = ctx->priority; + args->value = ctx->sched.priority; break; default: ret = -EINVAL; @@ -819,7 +833,7 @@ int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data, !capable(CAP_SYS_NICE)) ret = -EPERM; else - ctx->priority = priority; + ctx->sched.priority = priority; } break; diff --git a/drivers/gpu/drm/i915/i915_gem_context.h b/drivers/gpu/drm/i915/i915_gem_context.h index 7854262..ace3b12 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.h +++ b/drivers/gpu/drm/i915/i915_gem_context.h @@ -137,18 +137,7 @@ struct i915_gem_context { */ u32 user_handle; - /** - * @priority: execution and service priority - * - * All clients are equal, but some are more equal than others! - * - * Requests from a context with a greater (more positive) value of - * @priority will be executed before those with a lower @priority - * value, forming a simple QoS. - * - * The &drm_i915_private.kernel_context is assigned the lowest priority. - */ - int priority; + struct i915_sched_attr sched; /** ggtt_offset_bias: placement restriction for context objects */ u32 ggtt_offset_bias; @@ -160,7 +149,7 @@ struct i915_gem_context { u32 *lrc_reg_state; u64 lrc_desc; int pin_count; - } engine[I915_NUM_ENGINES]; + } __engine[I915_NUM_ENGINES]; /** ring_size: size for allocating the per-engine ring buffer */ u32 ring_size; @@ -267,6 +256,34 @@ static inline bool i915_gem_context_is_kernel(struct i915_gem_context *ctx) return !ctx->file_priv; } +static inline struct intel_context * +to_intel_context(struct i915_gem_context *ctx, + const struct intel_engine_cs *engine) +{ + return &ctx->__engine[engine->id]; +} + +static inline struct intel_ring * +intel_context_pin(struct i915_gem_context *ctx, struct intel_engine_cs *engine) +{ + return engine->context_pin(engine, ctx); +} + +static inline void __intel_context_pin(struct i915_gem_context *ctx, + const struct intel_engine_cs *engine) +{ + struct intel_context *ce = to_intel_context(ctx, engine); + + GEM_BUG_ON(!ce->pin_count); + ce->pin_count++; +} + +static inline void intel_context_unpin(struct i915_gem_context *ctx, + struct intel_engine_cs *engine) +{ + engine->context_unpin(engine, ctx); +} + /* i915_gem_context.c */ int __must_check i915_gem_contexts_init(struct drm_i915_private *dev_priv); void i915_gem_contexts_lost(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 0414228..f627a8c 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -81,6 +81,35 @@ enum { * but this remains just a hint as the kernel may choose a new location for * any object in the future. * + * At the level of talking to the hardware, submitting a batchbuffer for the + * GPU to execute is to add content to a buffer from which the HW + * command streamer is reading. + * + * 1. Add a command to load the HW context. For Logical Ring Contexts, i.e. + * Execlists, this command is not placed on the same buffer as the + * remaining items. + * + * 2. Add a command to invalidate caches to the buffer. + * + * 3. Add a batchbuffer start command to the buffer; the start command is + * essentially a token together with the GPU address of the batchbuffer + * to be executed. + * + * 4. Add a pipeline flush to the buffer. + * + * 5. Add a memory write command to the buffer to record when the GPU + * is done executing the batchbuffer. The memory write writes the + * global sequence number of the request, ``i915_request::global_seqno``; + * the i915 driver uses the current value in the register to determine + * if the GPU has completed the batchbuffer. + * + * 6. Add a user interrupt command to the buffer. This command instructs + * the GPU to issue an interrupt when the command, pipeline flush and + * memory write are completed. + * + * 7. Inform the hardware of the additional commands added to the buffer + * (by updating the tail pointer). + * * Processing an execbuf ioctl is conceptually split up into a few phases. * * 1. Validation - Ensure all the pointers, handles and flags are valid. @@ -733,7 +762,8 @@ static int eb_lookup_vmas(struct i915_execbuffer *eb) } /* transfer ref to ctx */ - vma->open_count++; + if (!vma->open_count++) + i915_vma_reopen(vma); list_add(&lut->obj_link, &obj->lut_list); list_add(&lut->ctx_link, &eb->ctx->handles_list); lut->ctx = eb->ctx; diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index 21d72f6..996ab2a 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -110,7 +110,8 @@ i915_get_ggtt_vma_pages(struct i915_vma *vma); static void gen6_ggtt_invalidate(struct drm_i915_private *dev_priv) { - /* Note that as an uncached mmio write, this should flush the + /* + * Note that as an uncached mmio write, this will flush the * WCB of the writes into the GGTT before it triggers the invalidate. */ I915_WRITE(GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN); @@ -1161,6 +1162,27 @@ static void gen8_ppgtt_insert_huge_entries(struct i915_vma *vma, vaddr[idx.pde] |= GEN8_PDE_IPS_64K; kunmap_atomic(vaddr); page_size = I915_GTT_PAGE_SIZE_64K; + + /* + * We write all 4K page entries, even when using 64K + * pages. In order to verify that the HW isn't cheating + * by using the 4K PTE instead of the 64K PTE, we want + * to remove all the surplus entries. If the HW skipped + * the 64K PTE, it will read/write into the scratch page + * instead - which we detect as missing results during + * selftests. + */ + if (I915_SELFTEST_ONLY(vma->vm->scrub_64K)) { + u16 i; + + encode = pte_encode | vma->vm->scratch_page.daddr; + vaddr = kmap_atomic_px(pd->page_table[idx.pde]); + + for (i = 1; i < index; i += 16) + memset64(vaddr + i, encode, 15); + + kunmap_atomic(vaddr); + } } vma->page_sizes.gtt |= page_size; @@ -2111,8 +2133,6 @@ static void i915_address_space_init(struct i915_address_space *vm, struct drm_i915_private *dev_priv, const char *name) { - i915_gem_timeline_init(dev_priv, &vm->timeline, name); - drm_mm_init(&vm->mm, 0, vm->total); vm->mm.head_node.color = I915_COLOR_UNEVICTABLE; @@ -2129,7 +2149,6 @@ static void i915_address_space_fini(struct i915_address_space *vm) if (pagevec_count(&vm->free_pages)) vm_free_pages_release(vm, true); - i915_gem_timeline_fini(&vm->timeline); drm_mm_takedown(&vm->mm); list_del(&vm->global_link); } @@ -2140,15 +2159,15 @@ static void gtt_write_workarounds(struct drm_i915_private *dev_priv) * called on driver load and after a GPU reset, so you can place * workarounds here even if they get overwritten by GPU reset. */ - /* WaIncreaseDefaultTLBEntries:chv,bdw,skl,bxt,kbl,glk,cfl,cnl */ + /* WaIncreaseDefaultTLBEntries:chv,bdw,skl,bxt,kbl,glk,cfl,cnl,icl */ if (IS_BROADWELL(dev_priv)) I915_WRITE(GEN8_L3_LRA_1_GPGPU, GEN8_L3_LRA_1_GPGPU_DEFAULT_VALUE_BDW); else if (IS_CHERRYVIEW(dev_priv)) I915_WRITE(GEN8_L3_LRA_1_GPGPU, GEN8_L3_LRA_1_GPGPU_DEFAULT_VALUE_CHV); - else if (IS_GEN9_BC(dev_priv) || IS_GEN10(dev_priv)) - I915_WRITE(GEN8_L3_LRA_1_GPGPU, GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_SKL); else if (IS_GEN9_LP(dev_priv)) I915_WRITE(GEN8_L3_LRA_1_GPGPU, GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_BXT); + else if (INTEL_GEN(dev_priv) >= 9) + I915_WRITE(GEN8_L3_LRA_1_GPGPU, GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_SKL); /* * To support 64K PTEs we need to first enable the use of the @@ -2222,6 +2241,12 @@ i915_ppgtt_create(struct drm_i915_private *dev_priv, void i915_ppgtt_close(struct i915_address_space *vm) { + GEM_BUG_ON(vm->closed); + vm->closed = true; +} + +static void ppgtt_destroy_vma(struct i915_address_space *vm) +{ struct list_head *phases[] = { &vm->active_list, &vm->inactive_list, @@ -2229,15 +2254,12 @@ void i915_ppgtt_close(struct i915_address_space *vm) 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); + i915_vma_destroy(vma); } } @@ -2248,7 +2270,8 @@ void i915_ppgtt_release(struct kref *kref) trace_i915_ppgtt_release(&ppgtt->base); - /* vmas should already be unbound and destroyed */ + ppgtt_destroy_vma(&ppgtt->base); + GEM_BUG_ON(!list_empty(&ppgtt->base.active_list)); GEM_BUG_ON(!list_empty(&ppgtt->base.inactive_list)); GEM_BUG_ON(!list_empty(&ppgtt->base.unbound_list)); @@ -2417,11 +2440,9 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm, for_each_sgt_dma(addr, sgt_iter, vma->pages) gen8_set_pte(gtt_entries++, pte_encode | addr); - wmb(); - - /* This next bit makes the above posting read even more important. We - * want to flush the TLBs only after we're certain all the PTE updates - * have finished. + /* + * We want to flush the TLBs only after we're certain all the PTE + * updates have finished. */ ggtt->invalidate(vm->i915); } @@ -2459,11 +2480,10 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm, dma_addr_t addr; for_each_sgt_dma(addr, iter, vma->pages) iowrite32(vm->pte_encode(addr, level, flags), &entries[i++]); - wmb(); - /* This next bit makes the above posting read even more important. We - * want to flush the TLBs only after we're certain all the PTE updates - * have finished. + /* + * We want to flush the TLBs only after we're certain all the PTE + * updates have finished. */ ggtt->invalidate(vm->i915); } @@ -3325,14 +3345,10 @@ static int gen8_gmch_probe(struct i915_ggtt *ggtt) DRM_ERROR("Can't set DMA mask/consistent mask (%d)\n", err); pci_read_config_word(pdev, SNB_GMCH_CTRL, &snb_gmch_ctl); - - if (INTEL_GEN(dev_priv) >= 9) { - size = gen8_get_total_gtt_size(snb_gmch_ctl); - } else if (IS_CHERRYVIEW(dev_priv)) { + if (IS_CHERRYVIEW(dev_priv)) size = chv_get_total_gtt_size(snb_gmch_ctl); - } else { + else size = gen8_get_total_gtt_size(snb_gmch_ctl); - } ggtt->base.total = (size / sizeof(gen8_pte_t)) << PAGE_SHIFT; ggtt->base.cleanup = gen6_gmch_remove; diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h index 6efc017..aec4f73 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.h +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h @@ -38,10 +38,9 @@ #include <linux/mm.h> #include <linux/pagevec.h> -#include "i915_gem_timeline.h" - #include "i915_request.h" #include "i915_selftest.h" +#include "i915_timeline.h" #define I915_GTT_PAGE_SIZE_4K BIT(12) #define I915_GTT_PAGE_SIZE_64K BIT(16) @@ -257,7 +256,6 @@ struct i915_pml4 { struct i915_address_space { struct drm_mm mm; - struct i915_gem_timeline timeline; struct drm_i915_private *i915; struct device *dma; /* Every address space belongs to a struct file - except for the global @@ -344,6 +342,7 @@ struct i915_address_space { void (*clear_pages)(struct i915_vma *vma); I915_SELFTEST_DECLARE(struct fault_attr fault_attr); + I915_SELFTEST_DECLARE(bool scrub_64K); }; #define i915_is_ggtt(V) (!(V)->file) diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c index 62aa679..ad949cc 100644 --- a/drivers/gpu/drm/i915/i915_gem_stolen.c +++ b/drivers/gpu/drm/i915/i915_gem_stolen.c @@ -51,6 +51,10 @@ int i915_gem_stolen_insert_node_in_range(struct drm_i915_private *dev_priv, if (!drm_mm_initialized(&dev_priv->mm.stolen)) return -ENODEV; + /* WaSkipStolenMemoryFirstPage:bdw+ */ + if (INTEL_GEN(dev_priv) >= 8 && start < 4096) + start = 4096; + mutex_lock(&dev_priv->mm.stolen_lock); ret = drm_mm_insert_node_in_range(&dev_priv->mm.stolen, node, size, alignment, 0, @@ -121,8 +125,8 @@ static int i915_adjust_stolen(struct drm_i915_private *dev_priv, if (stolen[0].start != stolen[1].start || stolen[0].end != stolen[1].end) { - DRM_DEBUG_KMS("GTT within stolen memory at %pR\n", &ggtt_res); - DRM_DEBUG_KMS("Stolen memory adjusted to %pR\n", dsm); + DRM_DEBUG_DRIVER("GTT within stolen memory at %pR\n", &ggtt_res); + DRM_DEBUG_DRIVER("Stolen memory adjusted to %pR\n", dsm); } } @@ -174,18 +178,19 @@ void i915_gem_cleanup_stolen(struct drm_device *dev) } static void g4x_get_stolen_reserved(struct drm_i915_private *dev_priv, - resource_size_t *base, resource_size_t *size) + resource_size_t *base, + resource_size_t *size) { - uint32_t reg_val = I915_READ(IS_GM45(dev_priv) ? - CTG_STOLEN_RESERVED : - ELK_STOLEN_RESERVED); + u32 reg_val = I915_READ(IS_GM45(dev_priv) ? + CTG_STOLEN_RESERVED : + ELK_STOLEN_RESERVED); resource_size_t stolen_top = dev_priv->dsm.end + 1; - if ((reg_val & G4X_STOLEN_RESERVED_ENABLE) == 0) { - *base = 0; - *size = 0; + DRM_DEBUG_DRIVER("%s_STOLEN_RESERVED = %08x\n", + IS_GM45(dev_priv) ? "CTG" : "ELK", reg_val); + + if ((reg_val & G4X_STOLEN_RESERVED_ENABLE) == 0) return; - } /* * Whether ILK really reuses the ELK register for this is unclear. @@ -193,30 +198,25 @@ static void g4x_get_stolen_reserved(struct drm_i915_private *dev_priv, */ WARN(IS_GEN5(dev_priv), "ILK stolen reserved found? 0x%08x\n", reg_val); - *base = (reg_val & G4X_STOLEN_RESERVED_ADDR2_MASK) << 16; + if (!(reg_val & G4X_STOLEN_RESERVED_ADDR2_MASK)) + return; + *base = (reg_val & G4X_STOLEN_RESERVED_ADDR2_MASK) << 16; WARN_ON((reg_val & G4X_STOLEN_RESERVED_ADDR1_MASK) < *base); - /* On these platforms, the register doesn't have a size field, so the - * size is the distance between the base and the top of the stolen - * memory. We also have the genuine case where base is zero and there's - * nothing reserved. */ - if (*base == 0) - *size = 0; - else - *size = stolen_top - *base; + *size = stolen_top - *base; } static void gen6_get_stolen_reserved(struct drm_i915_private *dev_priv, - resource_size_t *base, resource_size_t *size) + resource_size_t *base, + resource_size_t *size) { - uint32_t reg_val = I915_READ(GEN6_STOLEN_RESERVED); + u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED); + + DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); - if ((reg_val & GEN6_STOLEN_RESERVED_ENABLE) == 0) { - *base = 0; - *size = 0; + if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE)) return; - } *base = reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK; @@ -239,17 +239,44 @@ static void gen6_get_stolen_reserved(struct drm_i915_private *dev_priv, } } -static void gen7_get_stolen_reserved(struct drm_i915_private *dev_priv, - resource_size_t *base, resource_size_t *size) +static void vlv_get_stolen_reserved(struct drm_i915_private *dev_priv, + resource_size_t *base, + resource_size_t *size) { - uint32_t reg_val = I915_READ(GEN6_STOLEN_RESERVED); + u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED); + resource_size_t stolen_top = dev_priv->dsm.end + 1; + + DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); - if ((reg_val & GEN6_STOLEN_RESERVED_ENABLE) == 0) { - *base = 0; - *size = 0; + if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE)) return; + + switch (reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK) { + default: + MISSING_CASE(reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK); + case GEN7_STOLEN_RESERVED_1M: + *size = 1024 * 1024; + break; } + /* + * On vlv, the ADDR_MASK portion is left as 0 and HW deduces the + * reserved location as (top - size). + */ + *base = stolen_top - *size; +} + +static void gen7_get_stolen_reserved(struct drm_i915_private *dev_priv, + resource_size_t *base, + resource_size_t *size) +{ + u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED); + + DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); + + if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE)) + return; + *base = reg_val & GEN7_STOLEN_RESERVED_ADDR_MASK; switch (reg_val & GEN7_STOLEN_RESERVED_SIZE_MASK) { @@ -266,15 +293,15 @@ static void gen7_get_stolen_reserved(struct drm_i915_private *dev_priv, } static void chv_get_stolen_reserved(struct drm_i915_private *dev_priv, - resource_size_t *base, resource_size_t *size) + resource_size_t *base, + resource_size_t *size) { - uint32_t reg_val = I915_READ(GEN6_STOLEN_RESERVED); + u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED); - if ((reg_val & GEN6_STOLEN_RESERVED_ENABLE) == 0) { - *base = 0; - *size = 0; + DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); + + if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE)) return; - } *base = reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK; @@ -298,36 +325,28 @@ static void chv_get_stolen_reserved(struct drm_i915_private *dev_priv, } static void bdw_get_stolen_reserved(struct drm_i915_private *dev_priv, - resource_size_t *base, resource_size_t *size) + resource_size_t *base, + resource_size_t *size) { - uint32_t reg_val = I915_READ(GEN6_STOLEN_RESERVED); - resource_size_t stolen_top; + u32 reg_val = I915_READ(GEN6_STOLEN_RESERVED); + resource_size_t stolen_top = dev_priv->dsm.end + 1; - if ((reg_val & GEN6_STOLEN_RESERVED_ENABLE) == 0) { - *base = 0; - *size = 0; + DRM_DEBUG_DRIVER("GEN6_STOLEN_RESERVED = %08x\n", reg_val); + + if (!(reg_val & GEN6_STOLEN_RESERVED_ENABLE)) return; - } - stolen_top = dev_priv->dsm.end + 1; + if (!(reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK)) + return; *base = reg_val & GEN6_STOLEN_RESERVED_ADDR_MASK; - - /* On these platforms, the register doesn't have a size field, so the - * size is the distance between the base and the top of the stolen - * memory. We also have the genuine case where base is zero and there's - * nothing reserved. */ - if (*base == 0) - *size = 0; - else - *size = stolen_top - *base; + *size = stolen_top - *base; } int i915_gem_init_stolen(struct drm_i915_private *dev_priv) { resource_size_t reserved_base, stolen_top; resource_size_t reserved_total, reserved_size; - resource_size_t stolen_usable_start; mutex_init(&dev_priv->mm.stolen_lock); @@ -353,7 +372,7 @@ int i915_gem_init_stolen(struct drm_i915_private *dev_priv) GEM_BUG_ON(dev_priv->dsm.end <= dev_priv->dsm.start); stolen_top = dev_priv->dsm.end + 1; - reserved_base = 0; + reserved_base = stolen_top; reserved_size = 0; switch (INTEL_GEN(dev_priv)) { @@ -373,8 +392,12 @@ int i915_gem_init_stolen(struct drm_i915_private *dev_priv) &reserved_base, &reserved_size); break; case 7: - gen7_get_stolen_reserved(dev_priv, - &reserved_base, &reserved_size); + if (IS_VALLEYVIEW(dev_priv)) + vlv_get_stolen_reserved(dev_priv, + &reserved_base, &reserved_size); + else + gen7_get_stolen_reserved(dev_priv, + &reserved_base, &reserved_size); break; default: if (IS_LP(dev_priv)) @@ -386,11 +409,16 @@ int i915_gem_init_stolen(struct drm_i915_private *dev_priv) break; } - /* It is possible for the reserved base to be zero, but the register - * field for size doesn't have a zero option. */ - if (reserved_base == 0) { - reserved_size = 0; + /* + * Our expectation is that the reserved space is at the top of the + * stolen region and *never* at the bottom. If we see !reserved_base, + * it likely means we failed to read the registers correctly. + */ + if (!reserved_base) { + DRM_ERROR("inconsistent reservation %pa + %pa; ignoring\n", + &reserved_base, &reserved_size); reserved_base = stolen_top; + reserved_size = 0; } dev_priv->dsm_reserved = @@ -406,21 +434,15 @@ int i915_gem_init_stolen(struct drm_i915_private *dev_priv) * memory, so just consider the start. */ reserved_total = stolen_top - reserved_base; - DRM_DEBUG_KMS("Memory reserved for graphics device: %lluK, usable: %lluK\n", - (u64)resource_size(&dev_priv->dsm) >> 10, - ((u64)resource_size(&dev_priv->dsm) - reserved_total) >> 10); - - stolen_usable_start = 0; - /* WaSkipStolenMemoryFirstPage:bdw+ */ - if (INTEL_GEN(dev_priv) >= 8) - stolen_usable_start = 4096; + DRM_DEBUG_DRIVER("Memory reserved for graphics device: %lluK, usable: %lluK\n", + (u64)resource_size(&dev_priv->dsm) >> 10, + ((u64)resource_size(&dev_priv->dsm) - reserved_total) >> 10); dev_priv->stolen_usable_size = - resource_size(&dev_priv->dsm) - reserved_total - stolen_usable_start; + resource_size(&dev_priv->dsm) - reserved_total; /* Basic memrange allocator for stolen space. */ - drm_mm_init(&dev_priv->mm.stolen, stolen_usable_start, - dev_priv->stolen_usable_size); + drm_mm_init(&dev_priv->mm.stolen, 0, dev_priv->stolen_usable_size); return 0; } @@ -580,8 +602,8 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv lockdep_assert_held(&dev_priv->drm.struct_mutex); - DRM_DEBUG_KMS("creating preallocated stolen object: stolen_offset=%pa, gtt_offset=%pa, size=%pa\n", - &stolen_offset, >t_offset, &size); + DRM_DEBUG_DRIVER("creating preallocated stolen object: stolen_offset=%pa, gtt_offset=%pa, size=%pa\n", + &stolen_offset, >t_offset, &size); /* KISS and expect everything to be page-aligned */ if (WARN_ON(size == 0) || @@ -599,14 +621,14 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv ret = drm_mm_reserve_node(&dev_priv->mm.stolen, stolen); mutex_unlock(&dev_priv->mm.stolen_lock); if (ret) { - DRM_DEBUG_KMS("failed to allocate stolen space\n"); + DRM_DEBUG_DRIVER("failed to allocate stolen space\n"); kfree(stolen); return NULL; } obj = _i915_gem_object_create_stolen(dev_priv, stolen); if (obj == NULL) { - DRM_DEBUG_KMS("failed to allocate stolen object\n"); + DRM_DEBUG_DRIVER("failed to allocate stolen object\n"); i915_gem_stolen_remove_node(dev_priv, stolen); kfree(stolen); return NULL; @@ -635,7 +657,7 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv size, gtt_offset, obj->cache_level, 0); if (ret) { - DRM_DEBUG_KMS("failed to allocate stolen GTT space\n"); + DRM_DEBUG_DRIVER("failed to allocate stolen GTT space\n"); goto err_pages; } diff --git a/drivers/gpu/drm/i915/i915_gem_timeline.c b/drivers/gpu/drm/i915/i915_gem_timeline.c deleted file mode 100644 index e9fd876..0000000 --- a/drivers/gpu/drm/i915/i915_gem_timeline.c +++ /dev/null @@ -1,154 +0,0 @@ -/* - * Copyright © 2016 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 "i915_drv.h" -#include "i915_syncmap.h" - -static void __intel_timeline_init(struct intel_timeline *tl, - struct i915_gem_timeline *parent, - u64 context, - struct lock_class_key *lockclass, - const char *lockname) -{ - tl->fence_context = context; - tl->common = parent; - spin_lock_init(&tl->lock); - lockdep_set_class_and_name(&tl->lock, lockclass, lockname); - init_request_active(&tl->last_request, NULL); - INIT_LIST_HEAD(&tl->requests); - i915_syncmap_init(&tl->sync); -} - -static void __intel_timeline_fini(struct intel_timeline *tl) -{ - GEM_BUG_ON(!list_empty(&tl->requests)); - - i915_syncmap_free(&tl->sync); -} - -static int __i915_gem_timeline_init(struct drm_i915_private *i915, - struct i915_gem_timeline *timeline, - const char *name, - struct lock_class_key *lockclass, - const char *lockname) -{ - unsigned int i; - u64 fences; - - lockdep_assert_held(&i915->drm.struct_mutex); - - /* - * Ideally we want a set of engines on a single leaf as we expect - * to mostly be tracking synchronisation between engines. It is not - * a huge issue if this is not the case, but we may want to mitigate - * any page crossing penalties if they become an issue. - */ - BUILD_BUG_ON(KSYNCMAP < I915_NUM_ENGINES); - - timeline->i915 = i915; - timeline->name = kstrdup(name ?: "[kernel]", GFP_KERNEL); - if (!timeline->name) - return -ENOMEM; - - list_add(&timeline->link, &i915->gt.timelines); - - /* Called during early_init before we know how many engines there are */ - fences = dma_fence_context_alloc(ARRAY_SIZE(timeline->engine)); - for (i = 0; i < ARRAY_SIZE(timeline->engine); i++) - __intel_timeline_init(&timeline->engine[i], - timeline, fences++, - lockclass, lockname); - - return 0; -} - -int i915_gem_timeline_init(struct drm_i915_private *i915, - struct i915_gem_timeline *timeline, - const char *name) -{ - static struct lock_class_key class; - - return __i915_gem_timeline_init(i915, timeline, name, - &class, "&timeline->lock"); -} - -int i915_gem_timeline_init__global(struct drm_i915_private *i915) -{ - static struct lock_class_key class; - - return __i915_gem_timeline_init(i915, - &i915->gt.global_timeline, - "[execution]", - &class, "&global_timeline->lock"); -} - -/** - * i915_gem_timelines_park - called when the driver idles - * @i915: the drm_i915_private device - * - * When the driver is completely idle, we know that all of our sync points - * have been signaled and our tracking is then entirely redundant. Any request - * to wait upon an older sync point will be completed instantly as we know - * the fence is signaled and therefore we will not even look them up in the - * sync point map. - */ -void i915_gem_timelines_park(struct drm_i915_private *i915) -{ - struct i915_gem_timeline *timeline; - int i; - - lockdep_assert_held(&i915->drm.struct_mutex); - - list_for_each_entry(timeline, &i915->gt.timelines, link) { - for (i = 0; i < ARRAY_SIZE(timeline->engine); i++) { - struct intel_timeline *tl = &timeline->engine[i]; - - /* - * All known fences are completed so we can scrap - * the current sync point tracking and start afresh, - * any attempt to wait upon a previous sync point - * will be skipped as the fence was signaled. - */ - i915_syncmap_free(&tl->sync); - } - } -} - -void i915_gem_timeline_fini(struct i915_gem_timeline *timeline) -{ - int i; - - lockdep_assert_held(&timeline->i915->drm.struct_mutex); - - for (i = 0; i < ARRAY_SIZE(timeline->engine); i++) - __intel_timeline_fini(&timeline->engine[i]); - - list_del(&timeline->link); - kfree(timeline->name); -} - -#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) -#include "selftests/mock_timeline.c" -#include "selftests/i915_gem_timeline.c" -#endif diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index f89ac7a8..df234dc 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -32,6 +32,7 @@ #include <linux/zlib.h> #include <drm/drm_print.h> +#include "i915_gpu_error.h" #include "i915_drv.h" static inline const struct intel_engine_cs * @@ -403,16 +404,17 @@ static const char *bannable(const struct drm_i915_error_context *ctx) static void error_print_request(struct drm_i915_error_state_buf *m, const char *prefix, - const struct drm_i915_error_request *erq) + const struct drm_i915_error_request *erq, + const unsigned long epoch) { if (!erq->seqno) return; - err_printf(m, "%s pid %d, ban score %d, seqno %8x:%08x, prio %d, emitted %dms ago, head %08x, tail %08x\n", + err_printf(m, "%s pid %d, ban score %d, seqno %8x:%08x, prio %d, emitted %dms, start %08x, head %08x, tail %08x\n", prefix, erq->pid, erq->ban_score, - erq->context, erq->seqno, erq->priority, - jiffies_to_msecs(jiffies - erq->jiffies), - erq->head, erq->tail); + erq->context, erq->seqno, erq->sched_attr.priority, + jiffies_to_msecs(erq->jiffies - epoch), + erq->start, erq->head, erq->tail); } static void error_print_context(struct drm_i915_error_state_buf *m, @@ -421,12 +423,13 @@ static void error_print_context(struct drm_i915_error_state_buf *m, { err_printf(m, "%s%s[%d] user_handle %d hw_id %d, prio %d, ban score %d%s guilty %d active %d\n", header, ctx->comm, ctx->pid, ctx->handle, ctx->hw_id, - ctx->priority, ctx->ban_score, bannable(ctx), + ctx->sched_attr.priority, ctx->ban_score, bannable(ctx), ctx->guilty, ctx->active); } static void error_print_engine(struct drm_i915_error_state_buf *m, - const struct drm_i915_error_engine *ee) + const struct drm_i915_error_engine *ee, + const unsigned long epoch) { int n; @@ -496,14 +499,15 @@ static void error_print_engine(struct drm_i915_error_state_buf *m, err_printf(m, " hangcheck stall: %s\n", yesno(ee->hangcheck_stalled)); err_printf(m, " hangcheck action: %s\n", hangcheck_action_to_str(ee->hangcheck_action)); - err_printf(m, " hangcheck action timestamp: %lu, %u ms ago\n", + err_printf(m, " hangcheck action timestamp: %dms (%lu%s)\n", + jiffies_to_msecs(ee->hangcheck_timestamp - epoch), ee->hangcheck_timestamp, - jiffies_to_msecs(jiffies - ee->hangcheck_timestamp)); + ee->hangcheck_timestamp == epoch ? "; epoch" : ""); err_printf(m, " engine reset count: %u\n", ee->reset_count); for (n = 0; n < ee->num_ports; n++) { err_printf(m, " ELSP[%d]:", n); - error_print_request(m, " ", &ee->execlist[n]); + error_print_request(m, " ", &ee->execlist[n], epoch); } error_print_context(m, " Active context: ", &ee->context); @@ -649,6 +653,11 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m, ts = ktime_to_timespec64(error->uptime); err_printf(m, "Uptime: %lld s %ld us\n", (s64)ts.tv_sec, ts.tv_nsec / NSEC_PER_USEC); + err_printf(m, "Epoch: %lu jiffies (%u HZ)\n", error->epoch, HZ); + err_printf(m, "Capture: %lu jiffies; %d ms ago, %d ms after epoch\n", + error->capture, + jiffies_to_msecs(jiffies - error->capture), + jiffies_to_msecs(error->capture - error->epoch)); for (i = 0; i < ARRAY_SIZE(error->engine); i++) { if (error->engine[i].hangcheck_stalled && @@ -709,7 +718,7 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m, for (i = 0; i < ARRAY_SIZE(error->engine); i++) { if (error->engine[i].engine_id != -1) - error_print_engine(m, &error->engine[i]); + error_print_engine(m, &error->engine[i], error->epoch); } for (i = 0; i < ARRAY_SIZE(error->active_vm); i++) { @@ -768,7 +777,9 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m, dev_priv->engine[i]->name, ee->num_requests); for (j = 0; j < ee->num_requests; j++) - error_print_request(m, " ", &ee->requests[j]); + error_print_request(m, " ", + &ee->requests[j], + error->epoch); } if (IS_ERR(ee->waiters)) { @@ -1277,10 +1288,11 @@ static void record_request(struct i915_request *request, struct drm_i915_error_request *erq) { erq->context = request->ctx->hw_id; - erq->priority = request->priotree.priority; + erq->sched_attr = request->sched.attr; erq->ban_score = atomic_read(&request->ctx->ban_score); erq->seqno = request->global_seqno; erq->jiffies = request->emitted_jiffies; + erq->start = i915_ggtt_offset(request->ring->vma); erq->head = request->head; erq->tail = request->tail; @@ -1298,7 +1310,7 @@ static void engine_record_requests(struct intel_engine_cs *engine, count = 0; request = first; - list_for_each_entry_from(request, &engine->timeline->requests, link) + list_for_each_entry_from(request, &engine->timeline.requests, link) count++; if (!count) return; @@ -1311,7 +1323,7 @@ static void engine_record_requests(struct intel_engine_cs *engine, count = 0; request = first; - list_for_each_entry_from(request, &engine->timeline->requests, link) { + list_for_each_entry_from(request, &engine->timeline.requests, link) { if (count >= ee->num_requests) { /* * If the ring request list was changed in @@ -1371,7 +1383,7 @@ static void record_context(struct drm_i915_error_context *e, e->handle = ctx->user_handle; e->hw_id = ctx->hw_id; - e->priority = ctx->priority; + e->sched_attr = ctx->sched; e->ban_score = atomic_read(&ctx->ban_score); e->bannable = i915_gem_context_is_bannable(ctx); e->guilty = atomic_read(&ctx->guilty_count); @@ -1471,7 +1483,8 @@ static void gem_record_rings(struct i915_gpu_state *error) ee->ctx = i915_error_object_create(i915, - request->ctx->engine[i].state); + to_intel_context(request->ctx, + engine)->state); error->simulated |= i915_gem_context_no_error_capture(request->ctx); @@ -1734,6 +1747,22 @@ static void capture_params(struct i915_gpu_state *error) #undef DUP } +static unsigned long capture_find_epoch(const struct i915_gpu_state *error) +{ + unsigned long epoch = error->capture; + int i; + + for (i = 0; i < ARRAY_SIZE(error->engine); i++) { + const struct drm_i915_error_engine *ee = &error->engine[i]; + + if (ee->hangcheck_stalled && + time_before(ee->hangcheck_timestamp, epoch)) + epoch = ee->hangcheck_timestamp; + } + + return epoch; +} + static int capture(void *data) { struct i915_gpu_state *error = data; @@ -1742,6 +1771,7 @@ static int capture(void *data) error->boottime = ktime_get_boottime(); error->uptime = ktime_sub(ktime_get(), error->i915->gt.last_init_time); + error->capture = jiffies; capture_params(error); capture_gen_state(error); @@ -1755,6 +1785,8 @@ static int capture(void *data) error->overlay = intel_overlay_capture_error_state(error->i915); error->display = intel_display_capture_error_state(error->i915); + error->epoch = capture_find_epoch(error); + return 0; } diff --git a/drivers/gpu/drm/i915/i915_gpu_error.h b/drivers/gpu/drm/i915/i915_gpu_error.h new file mode 100644 index 0000000..dac0f8c --- /dev/null +++ b/drivers/gpu/drm/i915/i915_gpu_error.h @@ -0,0 +1,366 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright � 2008-2018 Intel Corporation + */ + +#ifndef _I915_GPU_ERROR_H_ +#define _I915_GPU_ERROR_H_ + +#include <linux/kref.h> +#include <linux/ktime.h> +#include <linux/sched.h> + +#include <drm/drm_mm.h> + +#include "intel_device_info.h" +#include "intel_ringbuffer.h" +#include "intel_uc_fw.h" + +#include "i915_gem.h" +#include "i915_gem_gtt.h" +#include "i915_params.h" +#include "i915_scheduler.h" + +struct drm_i915_private; +struct intel_overlay_error_state; +struct intel_display_error_state; + +struct i915_gpu_state { + struct kref ref; + ktime_t time; + ktime_t boottime; + ktime_t uptime; + unsigned long capture; + unsigned long epoch; + + struct drm_i915_private *i915; + + char error_msg[128]; + bool simulated; + bool awake; + bool wakelock; + bool suspended; + int iommu; + u32 reset_count; + u32 suspend_count; + struct intel_device_info device_info; + struct intel_driver_caps driver_caps; + struct i915_params params; + + struct i915_error_uc { + struct intel_uc_fw guc_fw; + struct intel_uc_fw huc_fw; + struct drm_i915_error_object *guc_log; + } uc; + + /* Generic register state */ + u32 eir; + u32 pgtbl_er; + u32 ier; + u32 gtier[4], ngtier; + u32 ccid; + u32 derrmr; + u32 forcewake; + u32 error; /* gen6+ */ + u32 err_int; /* gen7 */ + u32 fault_data0; /* gen8, gen9 */ + u32 fault_data1; /* gen8, gen9 */ + u32 done_reg; + u32 gac_eco; + u32 gam_ecochk; + u32 gab_ctl; + u32 gfx_mode; + + u32 nfence; + u64 fence[I915_MAX_NUM_FENCES]; + struct intel_overlay_error_state *overlay; + struct intel_display_error_state *display; + + struct drm_i915_error_engine { + int engine_id; + /* Software tracked state */ + bool idle; + bool waiting; + int num_waiters; + unsigned long hangcheck_timestamp; + bool hangcheck_stalled; + enum intel_engine_hangcheck_action hangcheck_action; + struct i915_address_space *vm; + int num_requests; + u32 reset_count; + + /* position of active request inside the ring */ + u32 rq_head, rq_post, rq_tail; + + /* our own tracking of ring head and tail */ + u32 cpu_ring_head; + u32 cpu_ring_tail; + + u32 last_seqno; + + /* Register state */ + u32 start; + u32 tail; + u32 head; + u32 ctl; + u32 mode; + u32 hws; + u32 ipeir; + u32 ipehr; + u32 bbstate; + u32 instpm; + u32 instps; + u32 seqno; + u64 bbaddr; + u64 acthd; + u32 fault_reg; + u64 faddr; + u32 rc_psmi; /* sleep state */ + u32 semaphore_mboxes[I915_NUM_ENGINES - 1]; + struct intel_instdone instdone; + + struct drm_i915_error_context { + char comm[TASK_COMM_LEN]; + pid_t pid; + u32 handle; + u32 hw_id; + int ban_score; + int active; + int guilty; + bool bannable; + struct i915_sched_attr sched_attr; + } context; + + struct drm_i915_error_object { + u64 gtt_offset; + u64 gtt_size; + int page_count; + int unused; + u32 *pages[0]; + } *ringbuffer, *batchbuffer, *wa_batchbuffer, *ctx, *hws_page; + + struct drm_i915_error_object **user_bo; + long user_bo_count; + + struct drm_i915_error_object *wa_ctx; + struct drm_i915_error_object *default_state; + + struct drm_i915_error_request { + long jiffies; + pid_t pid; + u32 context; + int ban_score; + u32 seqno; + u32 start; + u32 head; + u32 tail; + struct i915_sched_attr sched_attr; + } *requests, execlist[EXECLIST_MAX_PORTS]; + unsigned int num_ports; + + struct drm_i915_error_waiter { + char comm[TASK_COMM_LEN]; + pid_t pid; + u32 seqno; + } *waiters; + + struct { + u32 gfx_mode; + union { + u64 pdp[4]; + u32 pp_dir_base; + }; + } vm_info; + } engine[I915_NUM_ENGINES]; + + struct drm_i915_error_buffer { + u32 size; + u32 name; + u32 rseqno[I915_NUM_ENGINES], wseqno; + u64 gtt_offset; + u32 read_domains; + u32 write_domain; + s32 fence_reg:I915_MAX_NUM_FENCE_BITS; + u32 tiling:2; + u32 dirty:1; + u32 purgeable:1; + u32 userptr:1; + s32 engine:4; + u32 cache_level:3; + } *active_bo[I915_NUM_ENGINES], *pinned_bo; + u32 active_bo_count[I915_NUM_ENGINES], pinned_bo_count; + struct i915_address_space *active_vm[I915_NUM_ENGINES]; +}; + +struct i915_gpu_error { + /* For hangcheck timer */ +#define DRM_I915_HANGCHECK_PERIOD 1500 /* in ms */ +#define DRM_I915_HANGCHECK_JIFFIES msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD) + + struct delayed_work hangcheck_work; + + /* For reset and error_state handling. */ + spinlock_t lock; + /* Protected by the above dev->gpu_error.lock. */ + struct i915_gpu_state *first_error; + + atomic_t pending_fb_pin; + + unsigned long missed_irq_rings; + + /** + * State variable controlling the reset flow and count + * + * This is a counter which gets incremented when reset is triggered, + * + * Before the reset commences, the I915_RESET_BACKOFF bit is set + * meaning that any waiters holding onto the struct_mutex should + * relinquish the lock immediately in order for the reset to start. + * + * If reset is not completed successfully, the I915_WEDGE bit is + * set meaning that hardware is terminally sour and there is no + * recovery. All waiters on the reset_queue will be woken when + * that happens. + * + * This counter is used by the wait_seqno code to notice that reset + * event happened and it needs to restart the entire ioctl (since most + * likely the seqno it waited for won't ever signal anytime soon). + * + * This is important for lock-free wait paths, where no contended lock + * naturally enforces the correct ordering between the bail-out of the + * waiter and the gpu reset work code. + */ + unsigned long reset_count; + + /** + * flags: Control various stages of the GPU reset + * + * #I915_RESET_BACKOFF - When we start a reset, we want to stop any + * other users acquiring the struct_mutex. To do this we set the + * #I915_RESET_BACKOFF bit in the error flags when we detect a reset + * and then check for that bit before acquiring the struct_mutex (in + * i915_mutex_lock_interruptible()?). I915_RESET_BACKOFF serves a + * secondary role in preventing two concurrent global reset attempts. + * + * #I915_RESET_HANDOFF - To perform the actual GPU reset, we need the + * struct_mutex. We try to acquire the struct_mutex in the reset worker, + * but it may be held by some long running waiter (that we cannot + * interrupt without causing trouble). Once we are ready to do the GPU + * reset, we set the I915_RESET_HANDOFF bit and wakeup any waiters. If + * they already hold the struct_mutex and want to participate they can + * inspect the bit and do the reset directly, otherwise the worker + * waits for the struct_mutex. + * + * #I915_RESET_ENGINE[num_engines] - Since the driver doesn't need to + * acquire the struct_mutex to reset an engine, we need an explicit + * flag to prevent two concurrent reset attempts in the same engine. + * As the number of engines continues to grow, allocate the flags from + * the most significant bits. + * + * #I915_WEDGED - If reset fails and we can no longer use the GPU, + * we set the #I915_WEDGED bit. Prior to command submission, e.g. + * i915_request_alloc(), this bit is checked and the sequence + * aborted (with -EIO reported to userspace) if set. + */ + unsigned long flags; +#define I915_RESET_BACKOFF 0 +#define I915_RESET_HANDOFF 1 +#define I915_RESET_MODESET 2 +#define I915_WEDGED (BITS_PER_LONG - 1) +#define I915_RESET_ENGINE (I915_WEDGED - I915_NUM_ENGINES) + + /** Number of times an engine has been reset */ + u32 reset_engine_count[I915_NUM_ENGINES]; + + /** Set of stalled engines with guilty requests, in the current reset */ + u32 stalled_mask; + + /** Reason for the current *global* reset */ + const char *reason; + + /** + * Waitqueue to signal when a hang is detected. Used to for waiters + * to release the struct_mutex for the reset to procede. + */ + wait_queue_head_t wait_queue; + + /** + * Waitqueue to signal when the reset has completed. Used by clients + * that wait for dev_priv->mm.wedged to settle. + */ + wait_queue_head_t reset_queue; + + /* For missed irq/seqno simulation. */ + unsigned long test_irq_rings; +}; + +struct drm_i915_error_state_buf { + struct drm_i915_private *i915; + unsigned int bytes; + unsigned int size; + int err; + u8 *buf; + loff_t start; + loff_t pos; +}; + +#if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR) + +__printf(2, 3) +void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...); +int i915_error_state_to_str(struct drm_i915_error_state_buf *estr, + const struct i915_gpu_state *gpu); +int i915_error_state_buf_init(struct drm_i915_error_state_buf *eb, + struct drm_i915_private *i915, + size_t count, loff_t pos); + +static inline void +i915_error_state_buf_release(struct drm_i915_error_state_buf *eb) +{ + kfree(eb->buf); +} + +struct i915_gpu_state *i915_capture_gpu_state(struct drm_i915_private *i915); +void i915_capture_error_state(struct drm_i915_private *dev_priv, + u32 engine_mask, + const char *error_msg); + +static inline struct i915_gpu_state * +i915_gpu_state_get(struct i915_gpu_state *gpu) +{ + kref_get(&gpu->ref); + return gpu; +} + +void __i915_gpu_state_free(struct kref *kref); +static inline void i915_gpu_state_put(struct i915_gpu_state *gpu) +{ + if (gpu) + kref_put(&gpu->ref, __i915_gpu_state_free); +} + +struct i915_gpu_state *i915_first_error_state(struct drm_i915_private *i915); +void i915_reset_error_state(struct drm_i915_private *i915); + +#else + +static inline void i915_capture_error_state(struct drm_i915_private *dev_priv, + u32 engine_mask, + const char *error_msg) +{ +} + +static inline struct i915_gpu_state * +i915_first_error_state(struct drm_i915_private *i915) +{ + return NULL; +} + +static inline void i915_reset_error_state(struct drm_i915_private *i915) +{ +} + +#endif /* IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR) */ + +#endif /* _I915_GPU_ERROR_H_ */ diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index 633c187..f9bc3aa 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -243,6 +243,41 @@ void i915_hotplug_interrupt_update(struct drm_i915_private *dev_priv, spin_unlock_irq(&dev_priv->irq_lock); } +static u32 +gen11_gt_engine_identity(struct drm_i915_private * const i915, + const unsigned int bank, const unsigned int bit); + +bool gen11_reset_one_iir(struct drm_i915_private * const i915, + const unsigned int bank, + const unsigned int bit) +{ + void __iomem * const regs = i915->regs; + u32 dw; + + lockdep_assert_held(&i915->irq_lock); + + dw = raw_reg_read(regs, GEN11_GT_INTR_DW(bank)); + if (dw & BIT(bit)) { + /* + * According to the BSpec, DW_IIR bits cannot be cleared without + * first servicing the Selector & Shared IIR registers. + */ + gen11_gt_engine_identity(i915, bank, bit); + + /* + * We locked GT INT DW by reading it. If we want to (try + * to) recover from this succesfully, we need to clear + * our bit, otherwise we are locking the register for + * everybody. + */ + raw_reg_write(regs, GEN11_GT_INTR_DW(bank), BIT(bit)); + + return true; + } + + return false; +} + /** * ilk_update_display_irq - update DEIMR * @dev_priv: driver private @@ -308,17 +343,29 @@ void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask) static i915_reg_t gen6_pm_iir(struct drm_i915_private *dev_priv) { + WARN_ON_ONCE(INTEL_GEN(dev_priv) >= 11); + return INTEL_GEN(dev_priv) >= 8 ? GEN8_GT_IIR(2) : GEN6_PMIIR; } static i915_reg_t gen6_pm_imr(struct drm_i915_private *dev_priv) { - return INTEL_GEN(dev_priv) >= 8 ? GEN8_GT_IMR(2) : GEN6_PMIMR; + if (INTEL_GEN(dev_priv) >= 11) + return GEN11_GPM_WGBOXPERF_INTR_MASK; + else if (INTEL_GEN(dev_priv) >= 8) + return GEN8_GT_IMR(2); + else + return GEN6_PMIMR; } static i915_reg_t gen6_pm_ier(struct drm_i915_private *dev_priv) { - return INTEL_GEN(dev_priv) >= 8 ? GEN8_GT_IER(2) : GEN6_PMIER; + if (INTEL_GEN(dev_priv) >= 11) + return GEN11_GPM_WGBOXPERF_INTR_ENABLE; + else if (INTEL_GEN(dev_priv) >= 8) + return GEN8_GT_IER(2); + else + return GEN6_PMIER; } /** @@ -400,6 +447,18 @@ static void gen6_disable_pm_irq(struct drm_i915_private *dev_priv, u32 disable_m /* though a barrier is missing here, but don't really need a one */ } +void gen11_reset_rps_interrupts(struct drm_i915_private *dev_priv) +{ + spin_lock_irq(&dev_priv->irq_lock); + + while (gen11_reset_one_iir(dev_priv, 0, GEN11_GTPM)) + ; + + dev_priv->gt_pm.rps.pm_iir = 0; + + spin_unlock_irq(&dev_priv->irq_lock); +} + void gen6_reset_rps_interrupts(struct drm_i915_private *dev_priv) { spin_lock_irq(&dev_priv->irq_lock); @@ -415,12 +474,14 @@ void gen6_enable_rps_interrupts(struct drm_i915_private *dev_priv) if (READ_ONCE(rps->interrupts_enabled)) return; - if (WARN_ON_ONCE(IS_GEN11(dev_priv))) - return; - spin_lock_irq(&dev_priv->irq_lock); WARN_ON_ONCE(rps->pm_iir); - WARN_ON_ONCE(I915_READ(gen6_pm_iir(dev_priv)) & dev_priv->pm_rps_events); + + if (INTEL_GEN(dev_priv) >= 11) + WARN_ON_ONCE(gen11_reset_one_iir(dev_priv, 0, GEN11_GTPM)); + else + WARN_ON_ONCE(I915_READ(gen6_pm_iir(dev_priv)) & dev_priv->pm_rps_events); + rps->interrupts_enabled = true; gen6_enable_pm_irq(dev_priv, dev_priv->pm_rps_events); @@ -434,9 +495,6 @@ void gen6_disable_rps_interrupts(struct drm_i915_private *dev_priv) if (!READ_ONCE(rps->interrupts_enabled)) return; - if (WARN_ON_ONCE(IS_GEN11(dev_priv))) - return; - spin_lock_irq(&dev_priv->irq_lock); rps->interrupts_enabled = false; @@ -453,7 +511,10 @@ void gen6_disable_rps_interrupts(struct drm_i915_private *dev_priv) * state of the worker can be discarded. */ cancel_work_sync(&rps->work); - gen6_reset_rps_interrupts(dev_priv); + if (INTEL_GEN(dev_priv) >= 11) + gen11_reset_rps_interrupts(dev_priv); + else + gen6_reset_rps_interrupts(dev_priv); } void gen9_reset_guc_interrupts(struct drm_i915_private *dev_priv) @@ -1399,19 +1460,18 @@ static void snb_gt_irq_handler(struct drm_i915_private *dev_priv, } static void -gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir, int test_shift) +gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir) { struct intel_engine_execlists * const execlists = &engine->execlists; bool tasklet = false; - if (iir & (GT_CONTEXT_SWITCH_INTERRUPT << test_shift)) { - if (READ_ONCE(engine->execlists.active)) { - __set_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); - tasklet = true; - } + if (iir & GT_CONTEXT_SWITCH_INTERRUPT) { + if (READ_ONCE(engine->execlists.active)) + tasklet = !test_and_set_bit(ENGINE_IRQ_EXECLIST, + &engine->irq_posted); } - if (iir & (GT_RENDER_USER_INTERRUPT << test_shift)) { + if (iir & GT_RENDER_USER_INTERRUPT) { notify_ring(engine); tasklet |= USES_GUC_SUBMISSION(engine->i915); } @@ -1466,21 +1526,21 @@ static void gen8_gt_irq_handler(struct drm_i915_private *i915, { if (master_ctl & (GEN8_GT_RCS_IRQ | GEN8_GT_BCS_IRQ)) { gen8_cs_irq_handler(i915->engine[RCS], - gt_iir[0], GEN8_RCS_IRQ_SHIFT); + gt_iir[0] >> GEN8_RCS_IRQ_SHIFT); gen8_cs_irq_handler(i915->engine[BCS], - gt_iir[0], GEN8_BCS_IRQ_SHIFT); + gt_iir[0] >> GEN8_BCS_IRQ_SHIFT); } if (master_ctl & (GEN8_GT_VCS1_IRQ | GEN8_GT_VCS2_IRQ)) { gen8_cs_irq_handler(i915->engine[VCS], - gt_iir[1], GEN8_VCS1_IRQ_SHIFT); + gt_iir[1] >> GEN8_VCS1_IRQ_SHIFT); gen8_cs_irq_handler(i915->engine[VCS2], - gt_iir[1], GEN8_VCS2_IRQ_SHIFT); + gt_iir[1] >> GEN8_VCS2_IRQ_SHIFT); } if (master_ctl & GEN8_GT_VECS_IRQ) { gen8_cs_irq_handler(i915->engine[VECS], - gt_iir[3], GEN8_VECS_IRQ_SHIFT); + gt_iir[3] >> GEN8_VECS_IRQ_SHIFT); } if (master_ctl & (GEN8_GT_PM_IRQ | GEN8_GT_GUC_IRQ)) { @@ -1627,7 +1687,7 @@ static void display_pipe_crc_irq_handler(struct drm_i915_private *dev_priv, int head, tail; spin_lock(&pipe_crc->lock); - if (pipe_crc->source) { + if (pipe_crc->source && !crtc->base.crc.opened) { if (!pipe_crc->entries) { spin_unlock(&pipe_crc->lock); DRM_DEBUG_KMS("spurious interrupt\n"); @@ -1667,7 +1727,7 @@ static void display_pipe_crc_irq_handler(struct drm_i915_private *dev_priv, * On GEN8+ sometimes the second CRC is bonkers as well, so * don't trust that one either. */ - if (pipe_crc->skipped == 0 || + if (pipe_crc->skipped <= 0 || (INTEL_GEN(dev_priv) >= 8 && pipe_crc->skipped == 1)) { pipe_crc->skipped++; spin_unlock(&pipe_crc->lock); @@ -1766,37 +1826,8 @@ static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir) static void gen9_guc_irq_handler(struct drm_i915_private *dev_priv, u32 gt_iir) { - if (gt_iir & GEN9_GUC_TO_HOST_INT_EVENT) { - /* Sample the log buffer flush related bits & clear them out now - * itself from the message identity register to minimize the - * probability of losing a flush interrupt, when there are back - * to back flush interrupts. - * There can be a new flush interrupt, for different log buffer - * type (like for ISR), whilst Host is handling one (for DPC). - * Since same bit is used in message register for ISR & DPC, it - * could happen that GuC sets the bit for 2nd interrupt but Host - * clears out the bit on handling the 1st interrupt. - */ - u32 msg, flush; - - msg = I915_READ(SOFT_SCRATCH(15)); - flush = msg & (INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED | - INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER); - if (flush) { - /* Clear the message bits that are handled */ - I915_WRITE(SOFT_SCRATCH(15), msg & ~flush); - - /* Handle flush interrupt in bottom half */ - queue_work(dev_priv->guc.log.runtime.flush_wq, - &dev_priv->guc.log.runtime.flush_work); - - dev_priv->guc.log.flush_interrupt_count++; - } else { - /* Not clearing of unhandled event bits won't result in - * re-triggering of the interrupt. - */ - } - } + if (gt_iir & GEN9_GUC_TO_HOST_INT_EVENT) + intel_guc_to_host_event_handler(&dev_priv->guc); } static void i9xx_pipestat_irq_reset(struct drm_i915_private *dev_priv) @@ -2433,6 +2464,13 @@ static void ivb_display_irq_handler(struct drm_i915_private *dev_priv, if (de_iir & DE_ERR_INT_IVB) ivb_err_int_handler(dev_priv); + if (de_iir & DE_EDP_PSR_INT_HSW) { + u32 psr_iir = I915_READ(EDP_PSR_IIR); + + intel_psr_irq_handler(dev_priv, psr_iir); + I915_WRITE(EDP_PSR_IIR, psr_iir); + } + if (de_iir & DE_AUX_CHANNEL_A_IVB) dp_aux_irq_handler(dev_priv); @@ -2562,11 +2600,25 @@ gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl) if (master_ctl & GEN8_DE_MISC_IRQ) { iir = I915_READ(GEN8_DE_MISC_IIR); if (iir) { + bool found = false; + I915_WRITE(GEN8_DE_MISC_IIR, iir); ret = IRQ_HANDLED; - if (iir & GEN8_DE_MISC_GSE) + + if (iir & GEN8_DE_MISC_GSE) { intel_opregion_asle_intr(dev_priv); - else + found = true; + } + + if (iir & GEN8_DE_EDP_PSR) { + u32 psr_iir = I915_READ(EDP_PSR_IIR); + + intel_psr_irq_handler(dev_priv, psr_iir); + I915_WRITE(EDP_PSR_IIR, psr_iir); + found = true; + } + + if (!found) DRM_ERROR("Unexpected DE Misc interrupt\n"); } else @@ -2762,58 +2814,16 @@ static void __fini_wedge(struct wedge_me *w) (W)->i915; \ __fini_wedge((W))) -static __always_inline void -gen11_cs_irq_handler(struct intel_engine_cs * const engine, const u32 iir) -{ - gen8_cs_irq_handler(engine, iir, 0); -} - -static void -gen11_gt_engine_irq_handler(struct drm_i915_private * const i915, - const unsigned int bank, - const unsigned int engine_n, - const u16 iir) -{ - struct intel_engine_cs ** const engine = i915->engine; - - switch (bank) { - case 0: - switch (engine_n) { - - case GEN11_RCS0: - return gen11_cs_irq_handler(engine[RCS], iir); - - case GEN11_BCS: - return gen11_cs_irq_handler(engine[BCS], iir); - } - case 1: - switch (engine_n) { - - case GEN11_VCS(0): - return gen11_cs_irq_handler(engine[_VCS(0)], iir); - case GEN11_VCS(1): - return gen11_cs_irq_handler(engine[_VCS(1)], iir); - case GEN11_VCS(2): - return gen11_cs_irq_handler(engine[_VCS(2)], iir); - case GEN11_VCS(3): - return gen11_cs_irq_handler(engine[_VCS(3)], iir); - - case GEN11_VECS(0): - return gen11_cs_irq_handler(engine[_VECS(0)], iir); - case GEN11_VECS(1): - return gen11_cs_irq_handler(engine[_VECS(1)], iir); - } - } -} - static u32 -gen11_gt_engine_intr(struct drm_i915_private * const i915, - const unsigned int bank, const unsigned int bit) +gen11_gt_engine_identity(struct drm_i915_private * const i915, + const unsigned int bank, const unsigned int bit) { void __iomem * const regs = i915->regs; u32 timeout_ts; u32 ident; + lockdep_assert_held(&i915->irq_lock); + raw_reg_write(regs, GEN11_IIR_REG_SELECTOR(bank), BIT(bit)); /* @@ -2835,42 +2845,101 @@ gen11_gt_engine_intr(struct drm_i915_private * const i915, raw_reg_write(regs, GEN11_INTR_IDENTITY_REG(bank), GEN11_INTR_DATA_VALID); - return ident & GEN11_INTR_ENGINE_MASK; + return ident; } static void -gen11_gt_irq_handler(struct drm_i915_private * const i915, - const u32 master_ctl) +gen11_other_irq_handler(struct drm_i915_private * const i915, + const u8 instance, const u16 iir) +{ + if (instance == OTHER_GTPM_INSTANCE) + return gen6_rps_irq_handler(i915, iir); + + WARN_ONCE(1, "unhandled other interrupt instance=0x%x, iir=0x%x\n", + instance, iir); +} + +static void +gen11_engine_irq_handler(struct drm_i915_private * const i915, + const u8 class, const u8 instance, const u16 iir) +{ + struct intel_engine_cs *engine; + + if (instance <= MAX_ENGINE_INSTANCE) + engine = i915->engine_class[class][instance]; + else + engine = NULL; + + if (likely(engine)) + return gen8_cs_irq_handler(engine, iir); + + WARN_ONCE(1, "unhandled engine interrupt class=0x%x, instance=0x%x\n", + class, instance); +} + +static void +gen11_gt_identity_handler(struct drm_i915_private * const i915, + const u32 identity) +{ + const u8 class = GEN11_INTR_ENGINE_CLASS(identity); + const u8 instance = GEN11_INTR_ENGINE_INSTANCE(identity); + const u16 intr = GEN11_INTR_ENGINE_INTR(identity); + + if (unlikely(!intr)) + return; + + if (class <= COPY_ENGINE_CLASS) + return gen11_engine_irq_handler(i915, class, instance, intr); + + if (class == OTHER_CLASS) + return gen11_other_irq_handler(i915, instance, intr); + + WARN_ONCE(1, "unknown interrupt class=0x%x, instance=0x%x, intr=0x%x\n", + class, instance, intr); +} + +static void +gen11_gt_bank_handler(struct drm_i915_private * const i915, + const unsigned int bank) { void __iomem * const regs = i915->regs; - unsigned int bank; + unsigned long intr_dw; + unsigned int bit; - for (bank = 0; bank < 2; bank++) { - unsigned long intr_dw; - unsigned int bit; + lockdep_assert_held(&i915->irq_lock); - if (!(master_ctl & GEN11_GT_DW_IRQ(bank))) - continue; + intr_dw = raw_reg_read(regs, GEN11_GT_INTR_DW(bank)); - intr_dw = raw_reg_read(regs, GEN11_GT_INTR_DW(bank)); + if (unlikely(!intr_dw)) { + DRM_ERROR("GT_INTR_DW%u blank!\n", bank); + return; + } - if (unlikely(!intr_dw)) { - DRM_ERROR("GT_INTR_DW%u blank!\n", bank); - continue; - } + for_each_set_bit(bit, &intr_dw, 32) { + const u32 ident = gen11_gt_engine_identity(i915, + bank, bit); - for_each_set_bit(bit, &intr_dw, 32) { - const u16 iir = gen11_gt_engine_intr(i915, bank, bit); + gen11_gt_identity_handler(i915, ident); + } - if (unlikely(!iir)) - continue; + /* Clear must be after shared has been served for engine */ + raw_reg_write(regs, GEN11_GT_INTR_DW(bank), intr_dw); +} - gen11_gt_engine_irq_handler(i915, bank, bit, iir); - } +static void +gen11_gt_irq_handler(struct drm_i915_private * const i915, + const u32 master_ctl) +{ + unsigned int bank; - /* Clear must be after shared has been served for engine */ - raw_reg_write(regs, GEN11_GT_INTR_DW(bank), intr_dw); + spin_lock(&i915->irq_lock); + + for (bank = 0; bank < 2; bank++) { + if (master_ctl & GEN11_GT_DW_IRQ(bank)) + gen11_gt_bank_handler(i915, bank); } + + spin_unlock(&i915->irq_lock); } static irqreturn_t gen11_irq_handler(int irq, void *arg) @@ -2912,15 +2981,11 @@ static irqreturn_t gen11_irq_handler(int irq, void *arg) return IRQ_HANDLED; } -/** - * i915_reset_device - do process context error handling work - * @dev_priv: i915 device private - * - * Fire an error uevent so userspace can see that a hang or error - * was detected. - */ -static void i915_reset_device(struct drm_i915_private *dev_priv) +static void i915_reset_device(struct drm_i915_private *dev_priv, + u32 engine_mask, + const char *reason) { + struct i915_gpu_error *error = &dev_priv->gpu_error; struct kobject *kobj = &dev_priv->drm.primary->kdev->kobj; char *error_event[] = { I915_ERROR_UEVENT "=1", NULL }; char *reset_event[] = { I915_RESET_UEVENT "=1", NULL }; @@ -2936,29 +3001,35 @@ static void i915_reset_device(struct drm_i915_private *dev_priv) i915_wedge_on_timeout(&w, dev_priv, 5*HZ) { intel_prepare_reset(dev_priv); + error->reason = reason; + error->stalled_mask = engine_mask; + /* Signal that locked waiters should reset the GPU */ - set_bit(I915_RESET_HANDOFF, &dev_priv->gpu_error.flags); - wake_up_all(&dev_priv->gpu_error.wait_queue); + smp_mb__before_atomic(); + set_bit(I915_RESET_HANDOFF, &error->flags); + wake_up_all(&error->wait_queue); /* Wait for anyone holding the lock to wakeup, without * blocking indefinitely on struct_mutex. */ do { if (mutex_trylock(&dev_priv->drm.struct_mutex)) { - i915_reset(dev_priv, 0); + i915_reset(dev_priv, engine_mask, reason); mutex_unlock(&dev_priv->drm.struct_mutex); } - } while (wait_on_bit_timeout(&dev_priv->gpu_error.flags, + } while (wait_on_bit_timeout(&error->flags, I915_RESET_HANDOFF, TASK_UNINTERRUPTIBLE, 1)); + error->stalled_mask = 0; + error->reason = NULL; + intel_finish_reset(dev_priv); } - if (!test_bit(I915_WEDGED, &dev_priv->gpu_error.flags)) - kobject_uevent_env(kobj, - KOBJ_CHANGE, reset_done_event); + if (!test_bit(I915_WEDGED, &error->flags)) + kobject_uevent_env(kobj, KOBJ_CHANGE, reset_done_event); } static void i915_clear_error_registers(struct drm_i915_private *dev_priv) @@ -2990,6 +3061,7 @@ static void i915_clear_error_registers(struct drm_i915_private *dev_priv) * i915_handle_error - handle a gpu error * @dev_priv: i915 device private * @engine_mask: mask representing engines that are hung + * @flags: control flags * @fmt: Error message format string * * Do some basic checking of register state at error time and @@ -3000,16 +3072,23 @@ static void i915_clear_error_registers(struct drm_i915_private *dev_priv) */ void i915_handle_error(struct drm_i915_private *dev_priv, u32 engine_mask, + unsigned long flags, const char *fmt, ...) { struct intel_engine_cs *engine; unsigned int tmp; - va_list args; char error_msg[80]; + char *msg = NULL; - va_start(args, fmt); - vscnprintf(error_msg, sizeof(error_msg), fmt, args); - va_end(args); + if (fmt) { + va_list args; + + va_start(args, fmt); + vscnprintf(error_msg, sizeof(error_msg), fmt, args); + va_end(args); + + msg = error_msg; + } /* * In most cases it's guaranteed that we get here with an RPM @@ -3020,8 +3099,12 @@ void i915_handle_error(struct drm_i915_private *dev_priv, */ intel_runtime_pm_get(dev_priv); - i915_capture_error_state(dev_priv, engine_mask, error_msg); - i915_clear_error_registers(dev_priv); + engine_mask &= INTEL_INFO(dev_priv)->ring_mask; + + if (flags & I915_ERROR_CAPTURE) { + i915_capture_error_state(dev_priv, engine_mask, msg); + i915_clear_error_registers(dev_priv); + } /* * Try engine reset when available. We fall back to full reset if @@ -3034,7 +3117,7 @@ void i915_handle_error(struct drm_i915_private *dev_priv, &dev_priv->gpu_error.flags)) continue; - if (i915_reset_engine(engine, 0) == 0) + if (i915_reset_engine(engine, msg) == 0) engine_mask &= ~intel_engine_flag(engine); clear_bit(I915_RESET_ENGINE + engine->id, @@ -3064,7 +3147,7 @@ void i915_handle_error(struct drm_i915_private *dev_priv, TASK_UNINTERRUPTIBLE); } - i915_reset_device(dev_priv); + i915_reset_device(dev_priv, engine_mask, msg); for_each_engine(engine, dev_priv, tmp) { clear_bit(I915_RESET_ENGINE + engine->id, @@ -3286,6 +3369,11 @@ static void ironlake_irq_reset(struct drm_device *dev) if (IS_GEN7(dev_priv)) I915_WRITE(GEN7_ERR_INT, 0xffffffff); + if (IS_HASWELL(dev_priv)) { + I915_WRITE(EDP_PSR_IMR, 0xffffffff); + I915_WRITE(EDP_PSR_IIR, 0xffffffff); + } + gen5_gt_irq_reset(dev_priv); ibx_irq_reset(dev_priv); @@ -3324,6 +3412,9 @@ static void gen8_irq_reset(struct drm_device *dev) gen8_gt_irq_reset(dev_priv); + I915_WRITE(EDP_PSR_IMR, 0xffffffff); + I915_WRITE(EDP_PSR_IIR, 0xffffffff); + for_each_pipe(dev_priv, pipe) if (intel_display_power_is_enabled(dev_priv, POWER_DOMAIN_PIPE(pipe))) @@ -3349,6 +3440,9 @@ static void gen11_gt_irq_reset(struct drm_i915_private *dev_priv) I915_WRITE(GEN11_VCS0_VCS1_INTR_MASK, ~0); I915_WRITE(GEN11_VCS2_VCS3_INTR_MASK, ~0); I915_WRITE(GEN11_VECS0_VECS1_INTR_MASK, ~0); + + I915_WRITE(GEN11_GPM_WGBOXPERF_INTR_ENABLE, 0); + I915_WRITE(GEN11_GPM_WGBOXPERF_INTR_MASK, ~0); } static void gen11_irq_reset(struct drm_device *dev) @@ -3697,6 +3791,12 @@ static int ironlake_irq_postinstall(struct drm_device *dev) DE_DP_A_HOTPLUG); } + if (IS_HASWELL(dev_priv)) { + gen3_assert_iir_is_zero(dev_priv, EDP_PSR_IIR); + intel_psr_irq_control(dev_priv, dev_priv->psr.debug); + display_mask |= DE_EDP_PSR_INT_HSW; + } + dev_priv->irq_mask = ~display_mask; ibx_irq_pre_postinstall(dev); @@ -3807,7 +3907,7 @@ static void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv) uint32_t de_pipe_enables; u32 de_port_masked = GEN8_AUX_CHANNEL_A; u32 de_port_enables; - u32 de_misc_masked = GEN8_DE_MISC_GSE; + u32 de_misc_masked = GEN8_DE_MISC_GSE | GEN8_DE_EDP_PSR; enum pipe pipe; if (INTEL_GEN(dev_priv) >= 9) { @@ -3832,6 +3932,9 @@ static void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv) else if (IS_BROADWELL(dev_priv)) de_port_enables |= GEN8_PORT_DP_A_HOTPLUG; + gen3_assert_iir_is_zero(dev_priv, EDP_PSR_IIR); + intel_psr_irq_control(dev_priv, dev_priv->psr.debug); + for_each_pipe(dev_priv, pipe) { dev_priv->de_irq_mask[pipe] = ~de_pipe_masked; @@ -3887,7 +3990,14 @@ static void gen11_gt_irq_postinstall(struct drm_i915_private *dev_priv) I915_WRITE(GEN11_VCS2_VCS3_INTR_MASK, ~(irqs | irqs << 16)); I915_WRITE(GEN11_VECS0_VECS1_INTR_MASK, ~(irqs | irqs << 16)); - dev_priv->pm_imr = 0xffffffff; /* TODO */ + /* + * RPS interrupts will get enabled/disabled on demand when RPS itself + * is enabled/disabled. + */ + dev_priv->pm_ier = 0x0; + dev_priv->pm_imr = ~dev_priv->pm_ier; + I915_WRITE(GEN11_GPM_WGBOXPERF_INTR_ENABLE, 0); + I915_WRITE(GEN11_GPM_WGBOXPERF_INTR_MASK, ~0); } static int gen11_irq_postinstall(struct drm_device *dev) diff --git a/drivers/gpu/drm/i915/i915_oa_icl.c b/drivers/gpu/drm/i915/i915_oa_icl.c new file mode 100644 index 0000000..a566792 --- /dev/null +++ b/drivers/gpu/drm/i915/i915_oa_icl.c @@ -0,0 +1,118 @@ +/* + * Autogenerated file by GPU Top : https://github.com/rib/gputop + * DO NOT EDIT manually! + * + * + * Copyright (c) 2015 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + */ + +#include <linux/sysfs.h> + +#include "i915_drv.h" +#include "i915_oa_icl.h" + +static const struct i915_oa_reg b_counter_config_test_oa[] = { + { _MMIO(0x2740), 0x00000000 }, + { _MMIO(0x2710), 0x00000000 }, + { _MMIO(0x2714), 0xf0800000 }, + { _MMIO(0x2720), 0x00000000 }, + { _MMIO(0x2724), 0xf0800000 }, + { _MMIO(0x2770), 0x00000004 }, + { _MMIO(0x2774), 0x0000ffff }, + { _MMIO(0x2778), 0x00000003 }, + { _MMIO(0x277c), 0x0000ffff }, + { _MMIO(0x2780), 0x00000007 }, + { _MMIO(0x2784), 0x0000ffff }, + { _MMIO(0x2788), 0x00100002 }, + { _MMIO(0x278c), 0x0000fff7 }, + { _MMIO(0x2790), 0x00100002 }, + { _MMIO(0x2794), 0x0000ffcf }, + { _MMIO(0x2798), 0x00100082 }, + { _MMIO(0x279c), 0x0000ffef }, + { _MMIO(0x27a0), 0x001000c2 }, + { _MMIO(0x27a4), 0x0000ffe7 }, + { _MMIO(0x27a8), 0x00100001 }, + { _MMIO(0x27ac), 0x0000ffe7 }, +}; + +static const struct i915_oa_reg flex_eu_config_test_oa[] = { +}; + +static const struct i915_oa_reg mux_config_test_oa[] = { + { _MMIO(0xd04), 0x00000200 }, + { _MMIO(0x9840), 0x00000000 }, + { _MMIO(0x9884), 0x00000000 }, + { _MMIO(0x9888), 0x10060000 }, + { _MMIO(0x9888), 0x22060000 }, + { _MMIO(0x9888), 0x16060000 }, + { _MMIO(0x9888), 0x24060000 }, + { _MMIO(0x9888), 0x18060000 }, + { _MMIO(0x9888), 0x1a060000 }, + { _MMIO(0x9888), 0x12060000 }, + { _MMIO(0x9888), 0x14060000 }, + { _MMIO(0x9888), 0x10060000 }, + { _MMIO(0x9888), 0x22060000 }, + { _MMIO(0x9884), 0x00000003 }, + { _MMIO(0x9888), 0x16130000 }, + { _MMIO(0x9888), 0x24000001 }, + { _MMIO(0x9888), 0x0e130056 }, + { _MMIO(0x9888), 0x10130000 }, + { _MMIO(0x9888), 0x1a130000 }, + { _MMIO(0x9888), 0x541f0001 }, + { _MMIO(0x9888), 0x181f0000 }, + { _MMIO(0x9888), 0x4c1f0000 }, + { _MMIO(0x9888), 0x301f0000 }, +}; + +static ssize_t +show_test_oa_id(struct device *kdev, struct device_attribute *attr, char *buf) +{ + return sprintf(buf, "1\n"); +} + +void +i915_perf_load_test_config_icl(struct drm_i915_private *dev_priv) +{ + strlcpy(dev_priv->perf.oa.test_config.uuid, + "a291665e-244b-4b76-9b9a-01de9d3c8068", + sizeof(dev_priv->perf.oa.test_config.uuid)); + dev_priv->perf.oa.test_config.id = 1; + + dev_priv->perf.oa.test_config.mux_regs = mux_config_test_oa; + dev_priv->perf.oa.test_config.mux_regs_len = ARRAY_SIZE(mux_config_test_oa); + + dev_priv->perf.oa.test_config.b_counter_regs = b_counter_config_test_oa; + dev_priv->perf.oa.test_config.b_counter_regs_len = ARRAY_SIZE(b_counter_config_test_oa); + + dev_priv->perf.oa.test_config.flex_regs = flex_eu_config_test_oa; + dev_priv->perf.oa.test_config.flex_regs_len = ARRAY_SIZE(flex_eu_config_test_oa); + + dev_priv->perf.oa.test_config.sysfs_metric.name = "a291665e-244b-4b76-9b9a-01de9d3c8068"; + dev_priv->perf.oa.test_config.sysfs_metric.attrs = dev_priv->perf.oa.test_config.attrs; + + dev_priv->perf.oa.test_config.attrs[0] = &dev_priv->perf.oa.test_config.sysfs_metric_id.attr; + + dev_priv->perf.oa.test_config.sysfs_metric_id.attr.name = "id"; + dev_priv->perf.oa.test_config.sysfs_metric_id.attr.mode = 0444; + dev_priv->perf.oa.test_config.sysfs_metric_id.show = show_test_oa_id; +} diff --git a/drivers/gpu/drm/i915/i915_oa_icl.h b/drivers/gpu/drm/i915/i915_oa_icl.h new file mode 100644 index 0000000..ae1c24a --- /dev/null +++ b/drivers/gpu/drm/i915/i915_oa_icl.h @@ -0,0 +1,34 @@ +/* + * Autogenerated file by GPU Top : https://github.com/rib/gputop + * DO NOT EDIT manually! + * + * + * Copyright (c) 2015 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + */ + +#ifndef __I915_OA_ICL_H__ +#define __I915_OA_ICL_H__ + +extern void i915_perf_load_test_config_icl(struct drm_i915_private *dev_priv); + +#endif diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c index 08108ce..66ea355 100644 --- a/drivers/gpu/drm/i915/i915_params.c +++ b/drivers/gpu/drm/i915/i915_params.c @@ -164,6 +164,9 @@ i915_param_named_unsafe(guc_firmware_path, charp, 0400, i915_param_named_unsafe(huc_firmware_path, charp, 0400, "HuC firmware path to use instead of the default one"); +i915_param_named_unsafe(dmc_firmware_path, charp, 0400, + "DMC firmware path to use instead of the default one"); + i915_param_named_unsafe(enable_dp_mst, bool, 0600, "Enable multi-stream transport (MST) for new DisplayPort sinks. (default: true)"); diff --git a/drivers/gpu/drm/i915/i915_params.h b/drivers/gpu/drm/i915/i915_params.h index 430f5f9..6684025 100644 --- a/drivers/gpu/drm/i915/i915_params.h +++ b/drivers/gpu/drm/i915/i915_params.h @@ -48,9 +48,10 @@ struct drm_printer; param(int, enable_ips, 1) \ param(int, invert_brightness, 0) \ param(int, enable_guc, 0) \ - param(int, guc_log_level, 0) \ + param(int, guc_log_level, -1) \ param(char *, guc_firmware_path, NULL) \ param(char *, huc_firmware_path, NULL) \ + param(char *, dmc_firmware_path, NULL) \ param(int, mmio_debug, 0) \ param(int, edp_vswing, 0) \ param(int, reset, 2) \ diff --git a/drivers/gpu/drm/i915/i915_pci.c b/drivers/gpu/drm/i915/i915_pci.c index 062e91b..4364922 100644 --- a/drivers/gpu/drm/i915/i915_pci.c +++ b/drivers/gpu/drm/i915/i915_pci.c @@ -602,6 +602,7 @@ static const struct intel_device_info intel_icelake_11_info = { PLATFORM(INTEL_ICELAKE), .is_alpha_support = 1, .has_resource_streamer = 0, + .ring_mask = RENDER_RING | BLT_RING | VEBOX_RING | BSD_RING | BSD3_RING, }; #undef GEN diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index abaca6e..019bd2d 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -209,6 +209,7 @@ #include "i915_oa_cflgt2.h" #include "i915_oa_cflgt3.h" #include "i915_oa_cnl.h" +#include "i915_oa_icl.h" /* HW requires this to be a power of two, between 128k and 16M, though driver * is currently generally designed assuming the largest 16M size is used such @@ -1042,7 +1043,7 @@ static int gen7_append_oa_reports(struct i915_perf_stream *stream, I915_WRITE(GEN7_OASTATUS2, ((head & GEN7_OASTATUS2_HEAD_MASK) | - OA_MEM_SELECT_GGTT)); + GEN7_OASTATUS2_MEM_SELECT_GGTT)); dev_priv->perf.oa.oa_buffer.head = head; spin_unlock_irqrestore(&dev_priv->perf.oa.oa_buffer.ptr_lock, flags); @@ -1233,7 +1234,7 @@ static int oa_get_render_ctx_id(struct i915_perf_stream *stream) * * NB: implied RCS engine... */ - ring = engine->context_pin(engine, stream->ctx); + ring = intel_context_pin(stream->ctx, engine); mutex_unlock(&dev_priv->drm.struct_mutex); if (IS_ERR(ring)) return PTR_ERR(ring); @@ -1245,7 +1246,7 @@ static int oa_get_render_ctx_id(struct i915_perf_stream *stream) * with gen8+ and execlists */ dev_priv->perf.oa.specific_ctx_id = - i915_ggtt_offset(stream->ctx->engine[engine->id].state); + i915_ggtt_offset(to_intel_context(stream->ctx, engine)->state); } return 0; @@ -1270,7 +1271,7 @@ static void oa_put_render_ctx_id(struct i915_perf_stream *stream) mutex_lock(&dev_priv->drm.struct_mutex); dev_priv->perf.oa.specific_ctx_id = INVALID_CTX_ID; - engine->context_unpin(engine, stream->ctx); + intel_context_unpin(stream->ctx, engine); mutex_unlock(&dev_priv->drm.struct_mutex); } @@ -1332,7 +1333,8 @@ static void gen7_init_oa_buffer(struct drm_i915_private *dev_priv) /* Pre-DevBDW: OABUFFER must be set with counters off, * before OASTATUS1, but after OASTATUS2 */ - I915_WRITE(GEN7_OASTATUS2, gtt_offset | OA_MEM_SELECT_GGTT); /* head */ + I915_WRITE(GEN7_OASTATUS2, + gtt_offset | GEN7_OASTATUS2_MEM_SELECT_GGTT); /* head */ dev_priv->perf.oa.oa_buffer.head = gtt_offset; I915_WRITE(GEN7_OABUFFER, gtt_offset); @@ -1392,7 +1394,7 @@ static void gen8_init_oa_buffer(struct drm_i915_private *dev_priv) * bit." */ I915_WRITE(GEN8_OABUFFER, gtt_offset | - OABUFFER_SIZE_16M | OA_MEM_SELECT_GGTT); + OABUFFER_SIZE_16M | GEN8_OABUFFER_MEM_SELECT_GGTT); I915_WRITE(GEN8_OATAILPTR, gtt_offset & GEN8_OATAILPTR_MASK); /* Mark that we need updated tail pointers to read from... */ @@ -1693,7 +1695,7 @@ static int gen8_switch_to_updated_kernel_context(struct drm_i915_private *dev_pr const struct i915_oa_config *oa_config) { struct intel_engine_cs *engine = dev_priv->engine[RCS]; - struct i915_gem_timeline *timeline; + struct i915_timeline *timeline; struct i915_request *rq; int ret; @@ -1714,15 +1716,11 @@ static int gen8_switch_to_updated_kernel_context(struct drm_i915_private *dev_pr /* Queue this switch after all other activity */ list_for_each_entry(timeline, &dev_priv->gt.timelines, link) { struct i915_request *prev; - struct intel_timeline *tl; - tl = &timeline->engine[engine->id]; - prev = i915_gem_active_raw(&tl->last_request, + prev = i915_gem_active_raw(&timeline->last_request, &dev_priv->drm.struct_mutex); if (prev) - i915_sw_fence_await_sw_fence_gfp(&rq->submit, - &prev->submit, - GFP_KERNEL); + i915_request_await_dma_fence(rq, &prev->fence); } i915_request_add(rq); @@ -1757,6 +1755,7 @@ static int gen8_switch_to_updated_kernel_context(struct drm_i915_private *dev_pr static int gen8_configure_all_contexts(struct drm_i915_private *dev_priv, const struct i915_oa_config *oa_config) { + struct intel_engine_cs *engine = dev_priv->engine[RCS]; struct i915_gem_context *ctx; int ret; unsigned int wait_flags = I915_WAIT_LOCKED; @@ -1787,7 +1786,7 @@ static int gen8_configure_all_contexts(struct drm_i915_private *dev_priv, /* Update all contexts now that we've stalled the submission. */ list_for_each_entry(ctx, &dev_priv->contexts.list, link) { - struct intel_context *ce = &ctx->engine[RCS]; + struct intel_context *ce = to_intel_context(ctx, engine); u32 *regs; /* OA settings will be set upon first use */ @@ -1840,7 +1839,7 @@ static int gen8_enable_metric_set(struct drm_i915_private *dev_priv, * be read back from automatically triggered reports, as part of the * RPT_ID field. */ - if (IS_GEN9(dev_priv) || IS_GEN10(dev_priv)) { + if (IS_GEN(dev_priv, 9, 11)) { I915_WRITE(GEN8_OA_DEBUG, _MASKED_BIT_ENABLE(GEN9_OA_DEBUG_DISABLE_CLK_RATIO_REPORTS | GEN9_OA_DEBUG_INCLUDE_CLK_RATIO)); @@ -1870,7 +1869,6 @@ static void gen8_disable_metric_set(struct drm_i915_private *dev_priv) I915_WRITE(GDT_CHICKEN_BITS, (I915_READ(GDT_CHICKEN_BITS) & ~GT_NOA_ENABLE)); - } static void gen10_disable_metric_set(struct drm_i915_private *dev_priv) @@ -1885,6 +1883,13 @@ static void gen10_disable_metric_set(struct drm_i915_private *dev_priv) static void gen7_oa_enable(struct drm_i915_private *dev_priv) { + struct i915_gem_context *ctx = + dev_priv->perf.oa.exclusive_stream->ctx; + u32 ctx_id = dev_priv->perf.oa.specific_ctx_id; + bool periodic = dev_priv->perf.oa.periodic; + u32 period_exponent = dev_priv->perf.oa.period_exponent; + u32 report_format = dev_priv->perf.oa.oa_buffer.format; + /* * Reset buf pointers so we don't forward reports from before now. * @@ -1896,25 +1901,14 @@ static void gen7_oa_enable(struct drm_i915_private *dev_priv) */ gen7_init_oa_buffer(dev_priv); - if (dev_priv->perf.oa.exclusive_stream->enabled) { - struct i915_gem_context *ctx = - dev_priv->perf.oa.exclusive_stream->ctx; - u32 ctx_id = dev_priv->perf.oa.specific_ctx_id; - - bool periodic = dev_priv->perf.oa.periodic; - u32 period_exponent = dev_priv->perf.oa.period_exponent; - u32 report_format = dev_priv->perf.oa.oa_buffer.format; - - I915_WRITE(GEN7_OACONTROL, - (ctx_id & GEN7_OACONTROL_CTX_MASK) | - (period_exponent << - GEN7_OACONTROL_TIMER_PERIOD_SHIFT) | - (periodic ? GEN7_OACONTROL_TIMER_ENABLE : 0) | - (report_format << GEN7_OACONTROL_FORMAT_SHIFT) | - (ctx ? GEN7_OACONTROL_PER_CTX_ENABLE : 0) | - GEN7_OACONTROL_ENABLE); - } else - I915_WRITE(GEN7_OACONTROL, 0); + I915_WRITE(GEN7_OACONTROL, + (ctx_id & GEN7_OACONTROL_CTX_MASK) | + (period_exponent << + GEN7_OACONTROL_TIMER_PERIOD_SHIFT) | + (periodic ? GEN7_OACONTROL_TIMER_ENABLE : 0) | + (report_format << GEN7_OACONTROL_FORMAT_SHIFT) | + (ctx ? GEN7_OACONTROL_PER_CTX_ENABLE : 0) | + GEN7_OACONTROL_ENABLE); } static void gen8_oa_enable(struct drm_i915_private *dev_priv) @@ -1966,11 +1960,19 @@ static void i915_oa_stream_enable(struct i915_perf_stream *stream) static void gen7_oa_disable(struct drm_i915_private *dev_priv) { I915_WRITE(GEN7_OACONTROL, 0); + if (intel_wait_for_register(dev_priv, + GEN7_OACONTROL, GEN7_OACONTROL_ENABLE, 0, + 50)) + DRM_ERROR("wait for OA to be disabled timed out\n"); } static void gen8_oa_disable(struct drm_i915_private *dev_priv) { I915_WRITE(GEN8_OACONTROL, 0); + if (intel_wait_for_register(dev_priv, + GEN8_OACONTROL, GEN8_OA_COUNTER_ENABLE, 0, + 50)) + DRM_ERROR("wait for OA to be disabled timed out\n"); } /** @@ -2099,13 +2101,17 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, if (stream->ctx) { ret = oa_get_render_ctx_id(stream); - if (ret) + if (ret) { + DRM_DEBUG("Invalid context id to filter with\n"); return ret; + } } ret = get_oa_config(dev_priv, props->metrics_set, &stream->oa_config); - if (ret) + if (ret) { + DRM_DEBUG("Invalid OA config id=%i\n", props->metrics_set); goto err_config; + } /* PRM - observability performance counters: * @@ -2132,8 +2138,10 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, ret = dev_priv->perf.oa.ops.enable_metric_set(dev_priv, stream->oa_config); - if (ret) + if (ret) { + DRM_DEBUG("Unable to enable metric set\n"); goto err_enable; + } stream->ops = &i915_oa_stream_ops; @@ -2745,7 +2753,8 @@ static int read_properties_unlocked(struct drm_i915_private *dev_priv, props->ctx_handle = value; break; case DRM_I915_PERF_PROP_SAMPLE_OA: - props->sample_flags |= SAMPLE_OA_REPORT; + if (value) + props->sample_flags |= SAMPLE_OA_REPORT; break; case DRM_I915_PERF_PROP_OA_METRICS_SET: if (value == 0) { @@ -2935,6 +2944,8 @@ void i915_perf_register(struct drm_i915_private *dev_priv) i915_perf_load_test_config_cflgt3(dev_priv); } else if (IS_CANNONLAKE(dev_priv)) { i915_perf_load_test_config_cnl(dev_priv); + } else if (IS_ICELAKE(dev_priv)) { + i915_perf_load_test_config_icl(dev_priv); } if (dev_priv->perf.oa.test_config.id == 0) @@ -3292,6 +3303,8 @@ int i915_perf_add_config_ioctl(struct drm_device *dev, void *data, mutex_unlock(&dev_priv->perf.metrics_lock); + DRM_DEBUG("Added config %s id=%i\n", oa_config->uuid, oa_config->id); + return oa_config->id; sysfs_err: @@ -3348,6 +3361,9 @@ int i915_perf_remove_config_ioctl(struct drm_device *dev, void *data, &oa_config->sysfs_metric); idr_remove(&dev_priv->perf.metrics_idr, *arg); + + DRM_DEBUG("Removed config %s id=%i\n", oa_config->uuid, oa_config->id); + put_oa_config(dev_priv, oa_config); config_err: @@ -3467,7 +3483,7 @@ void i915_perf_init(struct drm_i915_private *dev_priv) dev_priv->perf.oa.gen8_valid_ctx_bit = (1<<16); } - } else if (IS_GEN10(dev_priv)) { + } else if (IS_GEN(dev_priv, 10, 11)) { dev_priv->perf.oa.ops.is_valid_b_counter_reg = gen7_is_valid_b_counter_addr; dev_priv->perf.oa.ops.is_valid_mux_reg = diff --git a/drivers/gpu/drm/i915/i915_pmu.c b/drivers/gpu/drm/i915/i915_pmu.c index f0519e3..dc87797 100644 --- a/drivers/gpu/drm/i915/i915_pmu.c +++ b/drivers/gpu/drm/i915/i915_pmu.c @@ -1,33 +1,12 @@ /* - * Copyright © 2017 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. + * SPDX-License-Identifier: MIT * + * Copyright © 2017-2018 Intel Corporation */ -#include <linux/perf_event.h> -#include <linux/pm_runtime.h> - -#include "i915_drv.h" #include "i915_pmu.h" #include "intel_ringbuffer.h" +#include "i915_drv.h" /* Frequency for the sampling timer for events which need it. */ #define FREQUENCY 200 diff --git a/drivers/gpu/drm/i915/i915_pmu.h b/drivers/gpu/drm/i915/i915_pmu.h index aa1b1a9..2ba7352 100644 --- a/drivers/gpu/drm/i915/i915_pmu.h +++ b/drivers/gpu/drm/i915/i915_pmu.h @@ -1,29 +1,19 @@ /* - * Copyright © 2017 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. + * SPDX-License-Identifier: MIT * + * Copyright © 2017-2018 Intel Corporation */ + #ifndef __I915_PMU_H__ #define __I915_PMU_H__ +#include <linux/hrtimer.h> +#include <linux/perf_event.h> +#include <linux/spinlock_types.h> +#include <drm/i915_drm.h> + +struct drm_i915_private; + enum { __I915_SAMPLE_FREQ_ACT = 0, __I915_SAMPLE_FREQ_REQ, diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 8a69a92..f11bb21 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -153,9 +153,6 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define _MMIO_PORT3(pipe, a, b, c) _MMIO(_PICK(pipe, a, b, c)) #define _PLL(pll, a, b) ((a) + (pll)*((b)-(a))) #define _MMIO_PLL(pll, a, b) _MMIO(_PLL(pll, a, b)) -#define _MMIO_PORT6(port, a, b, c, d, e, f) _MMIO(_PICK(port, a, b, c, d, e, f)) -#define _MMIO_PORT6_LN(port, ln, a0, a1, b, c, d, e, f) \ - _MMIO(_PICK(port, a0, b, c, d, e, f) + (ln * (a1 - a0))) #define _PHY3(phy, ...) _PICK(phy, __VA_ARGS__) #define _MMIO_PHY3(phy, a, b, c) _MMIO(_PHY3(phy, a, b, c)) @@ -191,6 +188,7 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define OTHER_CLASS 4 #define MAX_ENGINE_CLASS 4 +#define OTHER_GTPM_INSTANCE 1 #define MAX_ENGINE_INSTANCE 3 /* PCI config space */ @@ -304,6 +302,17 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define GEN6_GRDOM_VECS (1 << 4) #define GEN9_GRDOM_GUC (1 << 5) #define GEN8_GRDOM_MEDIA2 (1 << 7) +/* GEN11 changed all bit defs except for FULL & RENDER */ +#define GEN11_GRDOM_FULL GEN6_GRDOM_FULL +#define GEN11_GRDOM_RENDER GEN6_GRDOM_RENDER +#define GEN11_GRDOM_BLT (1 << 2) +#define GEN11_GRDOM_GUC (1 << 3) +#define GEN11_GRDOM_MEDIA (1 << 5) +#define GEN11_GRDOM_MEDIA2 (1 << 6) +#define GEN11_GRDOM_MEDIA3 (1 << 7) +#define GEN11_GRDOM_MEDIA4 (1 << 8) +#define GEN11_GRDOM_VECS (1 << 13) +#define GEN11_GRDOM_VECS2 (1 << 14) #define RING_PP_DIR_BASE(engine) _MMIO((engine)->mmio_base+0x228) #define RING_PP_DIR_BASE_READ(engine) _MMIO((engine)->mmio_base+0x518) @@ -430,145 +439,6 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define VGA_CR_INDEX_CGA 0x3d4 #define VGA_CR_DATA_CGA 0x3d5 -/* - * Instruction field definitions used by the command parser - */ -#define INSTR_CLIENT_SHIFT 29 -#define INSTR_MI_CLIENT 0x0 -#define INSTR_BC_CLIENT 0x2 -#define INSTR_RC_CLIENT 0x3 -#define INSTR_SUBCLIENT_SHIFT 27 -#define INSTR_SUBCLIENT_MASK 0x18000000 -#define INSTR_MEDIA_SUBCLIENT 0x2 -#define INSTR_26_TO_24_MASK 0x7000000 -#define INSTR_26_TO_24_SHIFT 24 - -/* - * Memory interface instructions used by the kernel - */ -#define MI_INSTR(opcode, flags) (((opcode) << 23) | (flags)) -/* Many MI commands use bit 22 of the header dword for GGTT vs PPGTT */ -#define MI_GLOBAL_GTT (1<<22) - -#define MI_NOOP MI_INSTR(0, 0) -#define MI_USER_INTERRUPT MI_INSTR(0x02, 0) -#define MI_WAIT_FOR_EVENT MI_INSTR(0x03, 0) -#define MI_WAIT_FOR_OVERLAY_FLIP (1<<16) -#define MI_WAIT_FOR_PLANE_B_FLIP (1<<6) -#define MI_WAIT_FOR_PLANE_A_FLIP (1<<2) -#define MI_WAIT_FOR_PLANE_A_SCANLINES (1<<1) -#define MI_FLUSH MI_INSTR(0x04, 0) -#define MI_READ_FLUSH (1 << 0) -#define MI_EXE_FLUSH (1 << 1) -#define MI_NO_WRITE_FLUSH (1 << 2) -#define MI_SCENE_COUNT (1 << 3) /* just increment scene count */ -#define MI_END_SCENE (1 << 4) /* flush binner and incr scene count */ -#define MI_INVALIDATE_ISP (1 << 5) /* invalidate indirect state pointers */ -#define MI_REPORT_HEAD MI_INSTR(0x07, 0) -#define MI_ARB_ON_OFF MI_INSTR(0x08, 0) -#define MI_ARB_ENABLE (1<<0) -#define MI_ARB_DISABLE (0<<0) -#define MI_BATCH_BUFFER_END MI_INSTR(0x0a, 0) -#define MI_SUSPEND_FLUSH MI_INSTR(0x0b, 0) -#define MI_SUSPEND_FLUSH_EN (1<<0) -#define MI_SET_APPID MI_INSTR(0x0e, 0) -#define MI_OVERLAY_FLIP MI_INSTR(0x11, 0) -#define MI_OVERLAY_CONTINUE (0x0<<21) -#define MI_OVERLAY_ON (0x1<<21) -#define MI_OVERLAY_OFF (0x2<<21) -#define MI_LOAD_SCAN_LINES_INCL MI_INSTR(0x12, 0) -#define MI_DISPLAY_FLIP MI_INSTR(0x14, 2) -#define MI_DISPLAY_FLIP_I915 MI_INSTR(0x14, 1) -#define MI_DISPLAY_FLIP_PLANE(n) ((n) << 20) -/* IVB has funny definitions for which plane to flip. */ -#define MI_DISPLAY_FLIP_IVB_PLANE_A (0 << 19) -#define MI_DISPLAY_FLIP_IVB_PLANE_B (1 << 19) -#define MI_DISPLAY_FLIP_IVB_SPRITE_A (2 << 19) -#define MI_DISPLAY_FLIP_IVB_SPRITE_B (3 << 19) -#define MI_DISPLAY_FLIP_IVB_PLANE_C (4 << 19) -#define MI_DISPLAY_FLIP_IVB_SPRITE_C (5 << 19) -/* SKL ones */ -#define MI_DISPLAY_FLIP_SKL_PLANE_1_A (0 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_1_B (1 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_1_C (2 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_2_A (4 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_2_B (5 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_2_C (6 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_3_A (7 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_3_B (8 << 8) -#define MI_DISPLAY_FLIP_SKL_PLANE_3_C (9 << 8) -#define MI_SEMAPHORE_MBOX MI_INSTR(0x16, 1) /* gen6, gen7 */ -#define MI_SEMAPHORE_GLOBAL_GTT (1<<22) -#define MI_SEMAPHORE_UPDATE (1<<21) -#define MI_SEMAPHORE_COMPARE (1<<20) -#define MI_SEMAPHORE_REGISTER (1<<18) -#define MI_SEMAPHORE_SYNC_VR (0<<16) /* RCS wait for VCS (RVSYNC) */ -#define MI_SEMAPHORE_SYNC_VER (1<<16) /* RCS wait for VECS (RVESYNC) */ -#define MI_SEMAPHORE_SYNC_BR (2<<16) /* RCS wait for BCS (RBSYNC) */ -#define MI_SEMAPHORE_SYNC_BV (0<<16) /* VCS wait for BCS (VBSYNC) */ -#define MI_SEMAPHORE_SYNC_VEV (1<<16) /* VCS wait for VECS (VVESYNC) */ -#define MI_SEMAPHORE_SYNC_RV (2<<16) /* VCS wait for RCS (VRSYNC) */ -#define MI_SEMAPHORE_SYNC_RB (0<<16) /* BCS wait for RCS (BRSYNC) */ -#define MI_SEMAPHORE_SYNC_VEB (1<<16) /* BCS wait for VECS (BVESYNC) */ -#define MI_SEMAPHORE_SYNC_VB (2<<16) /* BCS wait for VCS (BVSYNC) */ -#define MI_SEMAPHORE_SYNC_BVE (0<<16) /* VECS wait for BCS (VEBSYNC) */ -#define MI_SEMAPHORE_SYNC_VVE (1<<16) /* VECS wait for VCS (VEVSYNC) */ -#define MI_SEMAPHORE_SYNC_RVE (2<<16) /* VECS wait for RCS (VERSYNC) */ -#define MI_SEMAPHORE_SYNC_INVALID (3<<16) -#define MI_SEMAPHORE_SYNC_MASK (3<<16) -#define MI_SET_CONTEXT MI_INSTR(0x18, 0) -#define MI_MM_SPACE_GTT (1<<8) -#define MI_MM_SPACE_PHYSICAL (0<<8) -#define MI_SAVE_EXT_STATE_EN (1<<3) -#define MI_RESTORE_EXT_STATE_EN (1<<2) -#define MI_FORCE_RESTORE (1<<1) -#define MI_RESTORE_INHIBIT (1<<0) -#define HSW_MI_RS_SAVE_STATE_EN (1<<3) -#define HSW_MI_RS_RESTORE_STATE_EN (1<<2) -#define MI_SEMAPHORE_SIGNAL MI_INSTR(0x1b, 0) /* GEN8+ */ -#define MI_SEMAPHORE_TARGET(engine) ((engine)<<15) -#define MI_SEMAPHORE_WAIT MI_INSTR(0x1c, 2) /* GEN8+ */ -#define MI_SEMAPHORE_POLL (1<<15) -#define MI_SEMAPHORE_SAD_GTE_SDD (1<<12) -#define MI_STORE_DWORD_IMM MI_INSTR(0x20, 1) -#define MI_STORE_DWORD_IMM_GEN4 MI_INSTR(0x20, 2) -#define MI_MEM_VIRTUAL (1 << 22) /* 945,g33,965 */ -#define MI_USE_GGTT (1 << 22) /* g4x+ */ -#define MI_STORE_DWORD_INDEX MI_INSTR(0x21, 1) -#define MI_STORE_DWORD_INDEX_SHIFT 2 -/* Official intel docs are somewhat sloppy concerning MI_LOAD_REGISTER_IMM: - * - Always issue a MI_NOOP _before_ the MI_LOAD_REGISTER_IMM - otherwise hw - * simply ignores the register load under certain conditions. - * - One can actually load arbitrary many arbitrary registers: Simply issue x - * address/value pairs. Don't overdue it, though, x <= 2^4 must hold! - */ -#define MI_LOAD_REGISTER_IMM(x) MI_INSTR(0x22, 2*(x)-1) -#define MI_LRI_FORCE_POSTED (1<<12) -#define MI_STORE_REGISTER_MEM MI_INSTR(0x24, 1) -#define MI_STORE_REGISTER_MEM_GEN8 MI_INSTR(0x24, 2) -#define MI_SRM_LRM_GLOBAL_GTT (1<<22) -#define MI_FLUSH_DW MI_INSTR(0x26, 1) /* for GEN6 */ -#define MI_FLUSH_DW_STORE_INDEX (1<<21) -#define MI_INVALIDATE_TLB (1<<18) -#define MI_FLUSH_DW_OP_STOREDW (1<<14) -#define MI_FLUSH_DW_OP_MASK (3<<14) -#define MI_FLUSH_DW_NOTIFY (1<<8) -#define MI_INVALIDATE_BSD (1<<7) -#define MI_FLUSH_DW_USE_GTT (1<<2) -#define MI_FLUSH_DW_USE_PPGTT (0<<2) -#define MI_LOAD_REGISTER_MEM MI_INSTR(0x29, 1) -#define MI_LOAD_REGISTER_MEM_GEN8 MI_INSTR(0x29, 2) -#define MI_BATCH_BUFFER MI_INSTR(0x30, 1) -#define MI_BATCH_NON_SECURE (1) -/* for snb/ivb/vlv this also means "batch in ppgtt" when ppgtt is enabled. */ -#define MI_BATCH_NON_SECURE_I965 (1<<8) -#define MI_BATCH_PPGTT_HSW (1<<8) -#define MI_BATCH_NON_SECURE_HSW (1<<13) -#define MI_BATCH_BUFFER_START MI_INSTR(0x31, 0) -#define MI_BATCH_GTT (2<<6) /* aliased with (1<<7) on gen4 */ -#define MI_BATCH_BUFFER_START_GEN8 MI_INSTR(0x31, 1) -#define MI_BATCH_RESOURCE_STREAMER (1<<10) - #define MI_PREDICATE_SRC0 _MMIO(0x2400) #define MI_PREDICATE_SRC0_UDW _MMIO(0x2400 + 4) #define MI_PREDICATE_SRC1 _MMIO(0x2408) @@ -579,130 +449,6 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define LOWER_SLICE_DISABLED (0<<0) /* - * 3D instructions used by the kernel - */ -#define GFX_INSTR(opcode, flags) ((0x3 << 29) | ((opcode) << 24) | (flags)) - -#define GEN9_MEDIA_POOL_STATE ((0x3 << 29) | (0x2 << 27) | (0x5 << 16) | 4) -#define GEN9_MEDIA_POOL_ENABLE (1 << 31) -#define GFX_OP_RASTER_RULES ((0x3<<29)|(0x7<<24)) -#define GFX_OP_SCISSOR ((0x3<<29)|(0x1c<<24)|(0x10<<19)) -#define SC_UPDATE_SCISSOR (0x1<<1) -#define SC_ENABLE_MASK (0x1<<0) -#define SC_ENABLE (0x1<<0) -#define GFX_OP_LOAD_INDIRECT ((0x3<<29)|(0x1d<<24)|(0x7<<16)) -#define GFX_OP_SCISSOR_INFO ((0x3<<29)|(0x1d<<24)|(0x81<<16)|(0x1)) -#define SCI_YMIN_MASK (0xffff<<16) -#define SCI_XMIN_MASK (0xffff<<0) -#define SCI_YMAX_MASK (0xffff<<16) -#define SCI_XMAX_MASK (0xffff<<0) -#define GFX_OP_SCISSOR_ENABLE ((0x3<<29)|(0x1c<<24)|(0x10<<19)) -#define GFX_OP_SCISSOR_RECT ((0x3<<29)|(0x1d<<24)|(0x81<<16)|1) -#define GFX_OP_COLOR_FACTOR ((0x3<<29)|(0x1d<<24)|(0x1<<16)|0x0) -#define GFX_OP_STIPPLE ((0x3<<29)|(0x1d<<24)|(0x83<<16)) -#define GFX_OP_MAP_INFO ((0x3<<29)|(0x1d<<24)|0x4) -#define GFX_OP_DESTBUFFER_VARS ((0x3<<29)|(0x1d<<24)|(0x85<<16)|0x0) -#define GFX_OP_DESTBUFFER_INFO ((0x3<<29)|(0x1d<<24)|(0x8e<<16)|1) -#define GFX_OP_DRAWRECT_INFO ((0x3<<29)|(0x1d<<24)|(0x80<<16)|(0x3)) -#define GFX_OP_DRAWRECT_INFO_I965 ((0x7900<<16)|0x2) - -#define COLOR_BLT_CMD (2<<29 | 0x40<<22 | (5-2)) -#define SRC_COPY_BLT_CMD ((2<<29)|(0x43<<22)|4) -#define XY_SRC_COPY_BLT_CMD ((2<<29)|(0x53<<22)|6) -#define XY_MONO_SRC_COPY_IMM_BLT ((2<<29)|(0x71<<22)|5) -#define BLT_WRITE_A (2<<20) -#define BLT_WRITE_RGB (1<<20) -#define BLT_WRITE_RGBA (BLT_WRITE_RGB | BLT_WRITE_A) -#define BLT_DEPTH_8 (0<<24) -#define BLT_DEPTH_16_565 (1<<24) -#define BLT_DEPTH_16_1555 (2<<24) -#define BLT_DEPTH_32 (3<<24) -#define BLT_ROP_SRC_COPY (0xcc<<16) -#define BLT_ROP_COLOR_COPY (0xf0<<16) -#define XY_SRC_COPY_BLT_SRC_TILED (1<<15) /* 965+ only */ -#define XY_SRC_COPY_BLT_DST_TILED (1<<11) /* 965+ only */ -#define CMD_OP_DISPLAYBUFFER_INFO ((0x0<<29)|(0x14<<23)|2) -#define ASYNC_FLIP (1<<22) -#define DISPLAY_PLANE_A (0<<20) -#define DISPLAY_PLANE_B (1<<20) -#define GFX_OP_PIPE_CONTROL(len) ((0x3<<29)|(0x3<<27)|(0x2<<24)|((len)-2)) -#define PIPE_CONTROL_FLUSH_L3 (1<<27) -#define PIPE_CONTROL_GLOBAL_GTT_IVB (1<<24) /* gen7+ */ -#define PIPE_CONTROL_MMIO_WRITE (1<<23) -#define PIPE_CONTROL_STORE_DATA_INDEX (1<<21) -#define PIPE_CONTROL_CS_STALL (1<<20) -#define PIPE_CONTROL_TLB_INVALIDATE (1<<18) -#define PIPE_CONTROL_MEDIA_STATE_CLEAR (1<<16) -#define PIPE_CONTROL_QW_WRITE (1<<14) -#define PIPE_CONTROL_POST_SYNC_OP_MASK (3<<14) -#define PIPE_CONTROL_DEPTH_STALL (1<<13) -#define PIPE_CONTROL_WRITE_FLUSH (1<<12) -#define PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH (1<<12) /* gen6+ */ -#define PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE (1<<11) /* MBZ on Ironlake */ -#define PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE (1<<10) /* GM45+ only */ -#define PIPE_CONTROL_INDIRECT_STATE_DISABLE (1<<9) -#define PIPE_CONTROL_NOTIFY (1<<8) -#define PIPE_CONTROL_FLUSH_ENABLE (1<<7) /* gen7+ */ -#define PIPE_CONTROL_DC_FLUSH_ENABLE (1<<5) -#define PIPE_CONTROL_VF_CACHE_INVALIDATE (1<<4) -#define PIPE_CONTROL_CONST_CACHE_INVALIDATE (1<<3) -#define PIPE_CONTROL_STATE_CACHE_INVALIDATE (1<<2) -#define PIPE_CONTROL_STALL_AT_SCOREBOARD (1<<1) -#define PIPE_CONTROL_DEPTH_CACHE_FLUSH (1<<0) -#define PIPE_CONTROL_GLOBAL_GTT (1<<2) /* in addr dword */ - -/* - * Commands used only by the command parser - */ -#define MI_SET_PREDICATE MI_INSTR(0x01, 0) -#define MI_ARB_CHECK MI_INSTR(0x05, 0) -#define MI_RS_CONTROL MI_INSTR(0x06, 0) -#define MI_URB_ATOMIC_ALLOC MI_INSTR(0x09, 0) -#define MI_PREDICATE MI_INSTR(0x0C, 0) -#define MI_RS_CONTEXT MI_INSTR(0x0F, 0) -#define MI_TOPOLOGY_FILTER MI_INSTR(0x0D, 0) -#define MI_LOAD_SCAN_LINES_EXCL MI_INSTR(0x13, 0) -#define MI_URB_CLEAR MI_INSTR(0x19, 0) -#define MI_UPDATE_GTT MI_INSTR(0x23, 0) -#define MI_CLFLUSH MI_INSTR(0x27, 0) -#define MI_REPORT_PERF_COUNT MI_INSTR(0x28, 0) -#define MI_REPORT_PERF_COUNT_GGTT (1<<0) -#define MI_LOAD_REGISTER_REG MI_INSTR(0x2A, 0) -#define MI_RS_STORE_DATA_IMM MI_INSTR(0x2B, 0) -#define MI_LOAD_URB_MEM MI_INSTR(0x2C, 0) -#define MI_STORE_URB_MEM MI_INSTR(0x2D, 0) -#define MI_CONDITIONAL_BATCH_BUFFER_END MI_INSTR(0x36, 0) - -#define PIPELINE_SELECT ((0x3<<29)|(0x1<<27)|(0x1<<24)|(0x4<<16)) -#define GFX_OP_3DSTATE_VF_STATISTICS ((0x3<<29)|(0x1<<27)|(0x0<<24)|(0xB<<16)) -#define MEDIA_VFE_STATE ((0x3<<29)|(0x2<<27)|(0x0<<24)|(0x0<<16)) -#define MEDIA_VFE_STATE_MMIO_ACCESS_MASK (0x18) -#define GPGPU_OBJECT ((0x3<<29)|(0x2<<27)|(0x1<<24)|(0x4<<16)) -#define GPGPU_WALKER ((0x3<<29)|(0x2<<27)|(0x1<<24)|(0x5<<16)) -#define GFX_OP_3DSTATE_DX9_CONSTANTF_VS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x39<<16)) -#define GFX_OP_3DSTATE_DX9_CONSTANTF_PS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x3A<<16)) -#define GFX_OP_3DSTATE_SO_DECL_LIST \ - ((0x3<<29)|(0x3<<27)|(0x1<<24)|(0x17<<16)) - -#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_VS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x43<<16)) -#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_GS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x44<<16)) -#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_HS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x45<<16)) -#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_DS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x46<<16)) -#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_PS \ - ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x47<<16)) - -#define MFX_WAIT ((0x3<<29)|(0x1<<27)|(0x0<<16)) - -#define COLOR_BLT ((0x2<<29)|(0x40<<22)) -#define SRC_COPY_BLT ((0x2<<29)|(0x43<<22)) - -/* * Registers used only by the command parser */ #define BCS_SWCTRL _MMIO(0x22200) @@ -802,6 +548,7 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define GEN8_OABUFFER_UDW _MMIO(0x23b4) #define GEN8_OABUFFER _MMIO(0x2b14) +#define GEN8_OABUFFER_MEM_SELECT_GGTT (1 << 0) /* 0: PPGTT, 1: GGTT */ #define GEN7_OASTATUS1 _MMIO(0x2364) #define GEN7_OASTATUS1_TAIL_MASK 0xffffffc0 @@ -810,7 +557,8 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define GEN7_OASTATUS1_REPORT_LOST (1<<0) #define GEN7_OASTATUS2 _MMIO(0x2368) -#define GEN7_OASTATUS2_HEAD_MASK 0xffffffc0 +#define GEN7_OASTATUS2_HEAD_MASK 0xffffffc0 +#define GEN7_OASTATUS2_MEM_SELECT_GGTT (1 << 0) /* 0: PPGTT, 1: GGTT */ #define GEN8_OASTATUS _MMIO(0x2b08) #define GEN8_OASTATUS_OVERRUN_STATUS (1<<3) @@ -832,8 +580,6 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define OABUFFER_SIZE_8M (6<<3) #define OABUFFER_SIZE_16M (7<<3) -#define OA_MEM_SELECT_GGTT (1<<0) - /* * Flexible, Aggregate EU Counter Registers. * Note: these aren't contiguous @@ -1127,6 +873,12 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg) #define GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK (1 << GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT) #define GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ 0 #define GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ 1 +#define GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT 3 +#define GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK (0x7 << GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT) +#define GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ 0 +#define GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ 1 +#define GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_38_4_MHZ 2 +#define GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_25_MHZ 3 #define GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_SHIFT 1 #define GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_MASK (0x3 << GEN10_RPM_CONFIG0_CTC_SHIFT_PARAMETER_SHIFT) @@ -1948,79 +1700,100 @@ enum i915_power_well_id { #define _CNL_PORT_PCS_DW1_LN0_C 0x162C04 #define _CNL_PORT_PCS_DW1_LN0_D 0x162E04 #define _CNL_PORT_PCS_DW1_LN0_F 0x162804 -#define CNL_PORT_PCS_DW1_GRP(port) _MMIO_PORT6(port, \ +#define CNL_PORT_PCS_DW1_GRP(port) _MMIO(_PICK(port, \ _CNL_PORT_PCS_DW1_GRP_AE, \ _CNL_PORT_PCS_DW1_GRP_B, \ _CNL_PORT_PCS_DW1_GRP_C, \ _CNL_PORT_PCS_DW1_GRP_D, \ _CNL_PORT_PCS_DW1_GRP_AE, \ - _CNL_PORT_PCS_DW1_GRP_F) -#define CNL_PORT_PCS_DW1_LN0(port) _MMIO_PORT6(port, \ + _CNL_PORT_PCS_DW1_GRP_F)) + +#define CNL_PORT_PCS_DW1_LN0(port) _MMIO(_PICK(port, \ _CNL_PORT_PCS_DW1_LN0_AE, \ _CNL_PORT_PCS_DW1_LN0_B, \ _CNL_PORT_PCS_DW1_LN0_C, \ _CNL_PORT_PCS_DW1_LN0_D, \ _CNL_PORT_PCS_DW1_LN0_AE, \ - _CNL_PORT_PCS_DW1_LN0_F) + _CNL_PORT_PCS_DW1_LN0_F)) +#define _ICL_PORT_PCS_DW1_GRP_A 0x162604 +#define _ICL_PORT_PCS_DW1_GRP_B 0x6C604 +#define _ICL_PORT_PCS_DW1_LN0_A 0x162804 +#define _ICL_PORT_PCS_DW1_LN0_B 0x6C804 +#define ICL_PORT_PCS_DW1_GRP(port) _MMIO_PORT(port,\ + _ICL_PORT_PCS_DW1_GRP_A, \ + _ICL_PORT_PCS_DW1_GRP_B) +#define ICL_PORT_PCS_DW1_LN0(port) _MMIO_PORT(port, \ + _ICL_PORT_PCS_DW1_LN0_A, \ + _ICL_PORT_PCS_DW1_LN0_B) #define COMMON_KEEPER_EN (1 << 26) -#define _CNL_PORT_TX_DW2_GRP_AE 0x162348 -#define _CNL_PORT_TX_DW2_GRP_B 0x1623C8 -#define _CNL_PORT_TX_DW2_GRP_C 0x162B48 -#define _CNL_PORT_TX_DW2_GRP_D 0x162BC8 -#define _CNL_PORT_TX_DW2_GRP_F 0x162A48 -#define _CNL_PORT_TX_DW2_LN0_AE 0x162448 -#define _CNL_PORT_TX_DW2_LN0_B 0x162648 -#define _CNL_PORT_TX_DW2_LN0_C 0x162C48 -#define _CNL_PORT_TX_DW2_LN0_D 0x162E48 -#define _CNL_PORT_TX_DW2_LN0_F 0x162848 -#define CNL_PORT_TX_DW2_GRP(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW2_GRP_AE, \ - _CNL_PORT_TX_DW2_GRP_B, \ - _CNL_PORT_TX_DW2_GRP_C, \ - _CNL_PORT_TX_DW2_GRP_D, \ - _CNL_PORT_TX_DW2_GRP_AE, \ - _CNL_PORT_TX_DW2_GRP_F) -#define CNL_PORT_TX_DW2_LN0(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW2_LN0_AE, \ - _CNL_PORT_TX_DW2_LN0_B, \ - _CNL_PORT_TX_DW2_LN0_C, \ - _CNL_PORT_TX_DW2_LN0_D, \ - _CNL_PORT_TX_DW2_LN0_AE, \ - _CNL_PORT_TX_DW2_LN0_F) -#define SWING_SEL_UPPER(x) ((x >> 3) << 15) +/* CNL Port TX registers */ +#define _CNL_PORT_TX_AE_GRP_OFFSET 0x162340 +#define _CNL_PORT_TX_B_GRP_OFFSET 0x1623C0 +#define _CNL_PORT_TX_C_GRP_OFFSET 0x162B40 +#define _CNL_PORT_TX_D_GRP_OFFSET 0x162BC0 +#define _CNL_PORT_TX_F_GRP_OFFSET 0x162A40 +#define _CNL_PORT_TX_AE_LN0_OFFSET 0x162440 +#define _CNL_PORT_TX_B_LN0_OFFSET 0x162640 +#define _CNL_PORT_TX_C_LN0_OFFSET 0x162C40 +#define _CNL_PORT_TX_D_LN0_OFFSET 0x162E40 +#define _CNL_PORT_TX_F_LN0_OFFSET 0x162840 +#define _CNL_PORT_TX_DW_GRP(port, dw) (_PICK((port), \ + _CNL_PORT_TX_AE_GRP_OFFSET, \ + _CNL_PORT_TX_B_GRP_OFFSET, \ + _CNL_PORT_TX_B_GRP_OFFSET, \ + _CNL_PORT_TX_D_GRP_OFFSET, \ + _CNL_PORT_TX_AE_GRP_OFFSET, \ + _CNL_PORT_TX_F_GRP_OFFSET) + \ + 4*(dw)) +#define _CNL_PORT_TX_DW_LN0(port, dw) (_PICK((port), \ + _CNL_PORT_TX_AE_LN0_OFFSET, \ + _CNL_PORT_TX_B_LN0_OFFSET, \ + _CNL_PORT_TX_B_LN0_OFFSET, \ + _CNL_PORT_TX_D_LN0_OFFSET, \ + _CNL_PORT_TX_AE_LN0_OFFSET, \ + _CNL_PORT_TX_F_LN0_OFFSET) + \ + 4*(dw)) + +#define CNL_PORT_TX_DW2_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP((port), 2)) +#define CNL_PORT_TX_DW2_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0((port), 2)) +#define _ICL_PORT_TX_DW2_GRP_A 0x162688 +#define _ICL_PORT_TX_DW2_GRP_B 0x6C688 +#define _ICL_PORT_TX_DW2_LN0_A 0x162888 +#define _ICL_PORT_TX_DW2_LN0_B 0x6C888 +#define ICL_PORT_TX_DW2_GRP(port) _MMIO_PORT(port, \ + _ICL_PORT_TX_DW2_GRP_A, \ + _ICL_PORT_TX_DW2_GRP_B) +#define ICL_PORT_TX_DW2_LN0(port) _MMIO_PORT(port, \ + _ICL_PORT_TX_DW2_LN0_A, \ + _ICL_PORT_TX_DW2_LN0_B) +#define SWING_SEL_UPPER(x) (((x) >> 3) << 15) #define SWING_SEL_UPPER_MASK (1 << 15) -#define SWING_SEL_LOWER(x) ((x & 0x7) << 11) +#define SWING_SEL_LOWER(x) (((x) & 0x7) << 11) #define SWING_SEL_LOWER_MASK (0x7 << 11) #define RCOMP_SCALAR(x) ((x) << 0) #define RCOMP_SCALAR_MASK (0xFF << 0) -#define _CNL_PORT_TX_DW4_GRP_AE 0x162350 -#define _CNL_PORT_TX_DW4_GRP_B 0x1623D0 -#define _CNL_PORT_TX_DW4_GRP_C 0x162B50 -#define _CNL_PORT_TX_DW4_GRP_D 0x162BD0 -#define _CNL_PORT_TX_DW4_GRP_F 0x162A50 #define _CNL_PORT_TX_DW4_LN0_AE 0x162450 #define _CNL_PORT_TX_DW4_LN1_AE 0x1624D0 -#define _CNL_PORT_TX_DW4_LN0_B 0x162650 -#define _CNL_PORT_TX_DW4_LN0_C 0x162C50 -#define _CNL_PORT_TX_DW4_LN0_D 0x162E50 -#define _CNL_PORT_TX_DW4_LN0_F 0x162850 -#define CNL_PORT_TX_DW4_GRP(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW4_GRP_AE, \ - _CNL_PORT_TX_DW4_GRP_B, \ - _CNL_PORT_TX_DW4_GRP_C, \ - _CNL_PORT_TX_DW4_GRP_D, \ - _CNL_PORT_TX_DW4_GRP_AE, \ - _CNL_PORT_TX_DW4_GRP_F) -#define CNL_PORT_TX_DW4_LN(port, ln) _MMIO_PORT6_LN(port, ln, \ - _CNL_PORT_TX_DW4_LN0_AE, \ - _CNL_PORT_TX_DW4_LN1_AE, \ - _CNL_PORT_TX_DW4_LN0_B, \ - _CNL_PORT_TX_DW4_LN0_C, \ - _CNL_PORT_TX_DW4_LN0_D, \ - _CNL_PORT_TX_DW4_LN0_AE, \ - _CNL_PORT_TX_DW4_LN0_F) +#define CNL_PORT_TX_DW4_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP((port), 4)) +#define CNL_PORT_TX_DW4_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0((port), 4)) +#define CNL_PORT_TX_DW4_LN(port, ln) _MMIO(_CNL_PORT_TX_DW_LN0((port), 4) + \ + (ln * (_CNL_PORT_TX_DW4_LN1_AE - \ + _CNL_PORT_TX_DW4_LN0_AE))) +#define _ICL_PORT_TX_DW4_GRP_A 0x162690 +#define _ICL_PORT_TX_DW4_GRP_B 0x6C690 +#define _ICL_PORT_TX_DW4_LN0_A 0x162890 +#define _ICL_PORT_TX_DW4_LN1_A 0x162990 +#define _ICL_PORT_TX_DW4_LN0_B 0x6C890 +#define ICL_PORT_TX_DW4_GRP(port) _MMIO_PORT(port, \ + _ICL_PORT_TX_DW4_GRP_A, \ + _ICL_PORT_TX_DW4_GRP_B) +#define ICL_PORT_TX_DW4_LN(port, ln) _MMIO(_PORT(port, \ + _ICL_PORT_TX_DW4_LN0_A, \ + _ICL_PORT_TX_DW4_LN0_B) + \ + (ln * (_ICL_PORT_TX_DW4_LN1_A - \ + _ICL_PORT_TX_DW4_LN0_A))) #define LOADGEN_SELECT (1 << 31) #define POST_CURSOR_1(x) ((x) << 12) #define POST_CURSOR_1_MASK (0x3F << 12) @@ -2029,64 +1802,147 @@ enum i915_power_well_id { #define CURSOR_COEFF(x) ((x) << 0) #define CURSOR_COEFF_MASK (0x3F << 0) -#define _CNL_PORT_TX_DW5_GRP_AE 0x162354 -#define _CNL_PORT_TX_DW5_GRP_B 0x1623D4 -#define _CNL_PORT_TX_DW5_GRP_C 0x162B54 -#define _CNL_PORT_TX_DW5_GRP_D 0x162BD4 -#define _CNL_PORT_TX_DW5_GRP_F 0x162A54 -#define _CNL_PORT_TX_DW5_LN0_AE 0x162454 -#define _CNL_PORT_TX_DW5_LN0_B 0x162654 -#define _CNL_PORT_TX_DW5_LN0_C 0x162C54 -#define _CNL_PORT_TX_DW5_LN0_D 0x162E54 -#define _CNL_PORT_TX_DW5_LN0_F 0x162854 -#define CNL_PORT_TX_DW5_GRP(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW5_GRP_AE, \ - _CNL_PORT_TX_DW5_GRP_B, \ - _CNL_PORT_TX_DW5_GRP_C, \ - _CNL_PORT_TX_DW5_GRP_D, \ - _CNL_PORT_TX_DW5_GRP_AE, \ - _CNL_PORT_TX_DW5_GRP_F) -#define CNL_PORT_TX_DW5_LN0(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW5_LN0_AE, \ - _CNL_PORT_TX_DW5_LN0_B, \ - _CNL_PORT_TX_DW5_LN0_C, \ - _CNL_PORT_TX_DW5_LN0_D, \ - _CNL_PORT_TX_DW5_LN0_AE, \ - _CNL_PORT_TX_DW5_LN0_F) +#define CNL_PORT_TX_DW5_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP((port), 5)) +#define CNL_PORT_TX_DW5_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0((port), 5)) +#define _ICL_PORT_TX_DW5_GRP_A 0x162694 +#define _ICL_PORT_TX_DW5_GRP_B 0x6C694 +#define _ICL_PORT_TX_DW5_LN0_A 0x162894 +#define _ICL_PORT_TX_DW5_LN0_B 0x6C894 +#define ICL_PORT_TX_DW5_GRP(port) _MMIO_PORT(port, \ + _ICL_PORT_TX_DW5_GRP_A, \ + _ICL_PORT_TX_DW5_GRP_B) +#define ICL_PORT_TX_DW5_LN0(port) _MMIO_PORT(port, \ + _ICL_PORT_TX_DW5_LN0_A, \ + _ICL_PORT_TX_DW5_LN0_B) #define TX_TRAINING_EN (1 << 31) +#define TAP2_DISABLE (1 << 30) #define TAP3_DISABLE (1 << 29) #define SCALING_MODE_SEL(x) ((x) << 18) #define SCALING_MODE_SEL_MASK (0x7 << 18) #define RTERM_SELECT(x) ((x) << 3) #define RTERM_SELECT_MASK (0x7 << 3) -#define _CNL_PORT_TX_DW7_GRP_AE 0x16235C -#define _CNL_PORT_TX_DW7_GRP_B 0x1623DC -#define _CNL_PORT_TX_DW7_GRP_C 0x162B5C -#define _CNL_PORT_TX_DW7_GRP_D 0x162BDC -#define _CNL_PORT_TX_DW7_GRP_F 0x162A5C -#define _CNL_PORT_TX_DW7_LN0_AE 0x16245C -#define _CNL_PORT_TX_DW7_LN0_B 0x16265C -#define _CNL_PORT_TX_DW7_LN0_C 0x162C5C -#define _CNL_PORT_TX_DW7_LN0_D 0x162E5C -#define _CNL_PORT_TX_DW7_LN0_F 0x16285C -#define CNL_PORT_TX_DW7_GRP(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW7_GRP_AE, \ - _CNL_PORT_TX_DW7_GRP_B, \ - _CNL_PORT_TX_DW7_GRP_C, \ - _CNL_PORT_TX_DW7_GRP_D, \ - _CNL_PORT_TX_DW7_GRP_AE, \ - _CNL_PORT_TX_DW7_GRP_F) -#define CNL_PORT_TX_DW7_LN0(port) _MMIO_PORT6(port, \ - _CNL_PORT_TX_DW7_LN0_AE, \ - _CNL_PORT_TX_DW7_LN0_B, \ - _CNL_PORT_TX_DW7_LN0_C, \ - _CNL_PORT_TX_DW7_LN0_D, \ - _CNL_PORT_TX_DW7_LN0_AE, \ - _CNL_PORT_TX_DW7_LN0_F) +#define CNL_PORT_TX_DW7_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP((port), 7)) +#define CNL_PORT_TX_DW7_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0((port), 7)) #define N_SCALAR(x) ((x) << 24) #define N_SCALAR_MASK (0x7F << 24) +#define _ICL_MG_PHY_PORT_LN(port, ln, ln0p1, ln0p2, ln1p1) \ + _MMIO(_PORT((port) - PORT_C, ln0p1, ln0p2) + (ln) * ((ln1p1) - (ln0p1))) + +#define _ICL_MG_TX_LINK_PARAMS_TX1LN0_PORT1 0x16812C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN1_PORT1 0x16852C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN0_PORT2 0x16912C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN1_PORT2 0x16952C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN0_PORT3 0x16A12C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN1_PORT3 0x16A52C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN0_PORT4 0x16B12C +#define _ICL_MG_TX_LINK_PARAMS_TX1LN1_PORT4 0x16B52C +#define ICL_PORT_MG_TX1_LINK_PARAMS(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_LINK_PARAMS_TX1LN0_PORT1, \ + _ICL_MG_TX_LINK_PARAMS_TX1LN0_PORT2, \ + _ICL_MG_TX_LINK_PARAMS_TX1LN1_PORT1) + +#define _ICL_MG_TX_LINK_PARAMS_TX2LN0_PORT1 0x1680AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN1_PORT1 0x1684AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN0_PORT2 0x1690AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN1_PORT2 0x1694AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN0_PORT3 0x16A0AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN1_PORT3 0x16A4AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN0_PORT4 0x16B0AC +#define _ICL_MG_TX_LINK_PARAMS_TX2LN1_PORT4 0x16B4AC +#define ICL_PORT_MG_TX2_LINK_PARAMS(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_LINK_PARAMS_TX2LN0_PORT1, \ + _ICL_MG_TX_LINK_PARAMS_TX2LN0_PORT2, \ + _ICL_MG_TX_LINK_PARAMS_TX2LN1_PORT1) +#define CRI_USE_FS32 (1 << 5) + +#define _ICL_MG_TX_PISO_READLOAD_TX1LN0_PORT1 0x16814C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN1_PORT1 0x16854C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN0_PORT2 0x16914C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN1_PORT2 0x16954C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN0_PORT3 0x16A14C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN1_PORT3 0x16A54C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN0_PORT4 0x16B14C +#define _ICL_MG_TX_PISO_READLOAD_TX1LN1_PORT4 0x16B54C +#define ICL_PORT_MG_TX1_PISO_READLOAD(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_PISO_READLOAD_TX1LN0_PORT1, \ + _ICL_MG_TX_PISO_READLOAD_TX1LN0_PORT2, \ + _ICL_MG_TX_PISO_READLOAD_TX1LN1_PORT1) + +#define _ICL_MG_TX_PISO_READLOAD_TX2LN0_PORT1 0x1680CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN1_PORT1 0x1684CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN0_PORT2 0x1690CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN1_PORT2 0x1694CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN0_PORT3 0x16A0CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN1_PORT3 0x16A4CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN0_PORT4 0x16B0CC +#define _ICL_MG_TX_PISO_READLOAD_TX2LN1_PORT4 0x16B4CC +#define ICL_PORT_MG_TX2_PISO_READLOAD(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_PISO_READLOAD_TX2LN0_PORT1, \ + _ICL_MG_TX_PISO_READLOAD_TX2LN0_PORT2, \ + _ICL_MG_TX_PISO_READLOAD_TX2LN1_PORT1) +#define CRI_CALCINIT (1 << 1) + +#define _ICL_MG_TX_SWINGCTRL_TX1LN0_PORT1 0x168148 +#define _ICL_MG_TX_SWINGCTRL_TX1LN1_PORT1 0x168548 +#define _ICL_MG_TX_SWINGCTRL_TX1LN0_PORT2 0x169148 +#define _ICL_MG_TX_SWINGCTRL_TX1LN1_PORT2 0x169548 +#define _ICL_MG_TX_SWINGCTRL_TX1LN0_PORT3 0x16A148 +#define _ICL_MG_TX_SWINGCTRL_TX1LN1_PORT3 0x16A548 +#define _ICL_MG_TX_SWINGCTRL_TX1LN0_PORT4 0x16B148 +#define _ICL_MG_TX_SWINGCTRL_TX1LN1_PORT4 0x16B548 +#define ICL_PORT_MG_TX1_SWINGCTRL(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_SWINGCTRL_TX1LN0_PORT1, \ + _ICL_MG_TX_SWINGCTRL_TX1LN0_PORT2, \ + _ICL_MG_TX_SWINGCTRL_TX1LN1_PORT1) + +#define _ICL_MG_TX_SWINGCTRL_TX2LN0_PORT1 0x1680C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN1_PORT1 0x1684C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN0_PORT2 0x1690C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN1_PORT2 0x1694C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN0_PORT3 0x16A0C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN1_PORT3 0x16A4C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN0_PORT4 0x16B0C8 +#define _ICL_MG_TX_SWINGCTRL_TX2LN1_PORT4 0x16B4C8 +#define ICL_PORT_MG_TX2_SWINGCTRL(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_SWINGCTRL_TX2LN0_PORT1, \ + _ICL_MG_TX_SWINGCTRL_TX2LN0_PORT2, \ + _ICL_MG_TX_SWINGCTRL_TX2LN1_PORT1) +#define CRI_TXDEEMPH_OVERRIDE_17_12(x) ((x) << 0) +#define CRI_TXDEEMPH_OVERRIDE_17_12_MASK (0x3F << 0) + +#define _ICL_MG_TX_DRVCTRL_TX1LN0_PORT1 0x168144 +#define _ICL_MG_TX_DRVCTRL_TX1LN1_PORT1 0x168544 +#define _ICL_MG_TX_DRVCTRL_TX1LN0_PORT2 0x169144 +#define _ICL_MG_TX_DRVCTRL_TX1LN1_PORT2 0x169544 +#define _ICL_MG_TX_DRVCTRL_TX1LN0_PORT3 0x16A144 +#define _ICL_MG_TX_DRVCTRL_TX1LN1_PORT3 0x16A544 +#define _ICL_MG_TX_DRVCTRL_TX1LN0_PORT4 0x16B144 +#define _ICL_MG_TX_DRVCTRL_TX1LN1_PORT4 0x16B544 +#define ICL_PORT_MG_TX1_DRVCTRL(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_DRVCTRL_TX1LN0_PORT1, \ + _ICL_MG_TX_DRVCTRL_TX1LN0_PORT2, \ + _ICL_MG_TX_DRVCTRL_TX1LN1_PORT1) + +#define _ICL_MG_TX_DRVCTRL_TX2LN0_PORT1 0x1680C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN1_PORT1 0x1684C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN0_PORT2 0x1690C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN1_PORT2 0x1694C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN0_PORT3 0x16A0C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN1_PORT3 0x16A4C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN0_PORT4 0x16B0C4 +#define _ICL_MG_TX_DRVCTRL_TX2LN1_PORT4 0x16B4C4 +#define ICL_PORT_MG_TX2_DRVCTRL(port, ln) \ + _ICL_MG_PHY_PORT_LN(port, ln, _ICL_MG_TX_DRVCTRL_TX2LN0_PORT1, \ + _ICL_MG_TX_DRVCTRL_TX2LN0_PORT2, \ + _ICL_MG_TX_DRVCTRL_TX2LN1_PORT1) +#define CRI_TXDEEMPH_OVERRIDE_11_6(x) ((x) << 24) +#define CRI_TXDEEMPH_OVERRIDE_11_6_MASK (0x3F << 24) +#define CRI_TXDEEMPH_OVERRIDE_EN (1 << 22) +#define CRI_TXDEEMPH_OVERRIDE_5_0(x) ((x) << 16) +#define CRI_TXDEEMPH_OVERRIDE_5_0_MASK (0x3F << 16) + /* The spec defines this only for BXT PHY0, but lets assume that this * would exist for PHY1 too if it had a second channel. */ @@ -2473,6 +2329,10 @@ enum i915_power_well_id { #define GEN8_MCR_SLICE_MASK GEN8_MCR_SLICE(3) #define GEN8_MCR_SUBSLICE(subslice) (((subslice) & 3) << 24) #define GEN8_MCR_SUBSLICE_MASK GEN8_MCR_SUBSLICE(3) +#define GEN11_MCR_SLICE(slice) (((slice) & 0xf) << 27) +#define GEN11_MCR_SLICE_MASK GEN11_MCR_SLICE(0xf) +#define GEN11_MCR_SUBSLICE(subslice) (((subslice) & 0x7) << 24) +#define GEN11_MCR_SUBSLICE_MASK GEN11_MCR_SUBSLICE(0x7) #define RING_IPEIR(base) _MMIO((base)+0x64) #define RING_IPEHR(base) _MMIO((base)+0x68) /* @@ -2867,6 +2727,19 @@ enum i915_power_well_id { #define GEN10_EU_DISABLE3 _MMIO(0x9140) #define GEN10_EU_DIS_SS_MASK 0xff +#define GEN11_GT_VEBOX_VDBOX_DISABLE _MMIO(0x9140) +#define GEN11_GT_VDBOX_DISABLE_MASK 0xff +#define GEN11_GT_VEBOX_DISABLE_SHIFT 16 +#define GEN11_GT_VEBOX_DISABLE_MASK (0xff << GEN11_GT_VEBOX_DISABLE_SHIFT) + +#define GEN11_EU_DISABLE _MMIO(0x9134) +#define GEN11_EU_DIS_MASK 0xFF + +#define GEN11_GT_SLICE_ENABLE _MMIO(0x9138) +#define GEN11_GT_S_ENA_MASK 0xFF + +#define GEN11_GT_SUBSLICE_DISABLE _MMIO(0x913C) + #define GEN6_BSD_SLEEP_PSMI_CONTROL _MMIO(0x12050) #define GEN6_BSD_SLEEP_MSG_DISABLE (1 << 0) #define GEN6_BSD_SLEEP_FLUSH_DISABLE (1 << 2) @@ -3951,6 +3824,9 @@ enum { #define _CLKGATE_DIS_PSL_A 0x46520 #define _CLKGATE_DIS_PSL_B 0x46524 #define _CLKGATE_DIS_PSL_C 0x46528 +#define DUPS1_GATING_DIS (1 << 15) +#define DUPS2_GATING_DIS (1 << 19) +#define DUPS3_GATING_DIS (1 << 23) #define DPF_GATING_DIS (1 << 10) #define DPF_RAM_GATING_DIS (1 << 9) #define DPFR_GATING_DIS (1 << 8) @@ -3964,6 +3840,7 @@ enum { #define SLICE_UNIT_LEVEL_CLKGATE _MMIO(0x94d4) #define SARBUNIT_CLKGATE_DIS (1 << 5) #define RCCUNIT_CLKGATE_DIS (1 << 7) +#define MSCUNIT_CLKGATE_DIS (1 << 10) #define SUBSLICE_UNIT_LEVEL_CLKGATE _MMIO(0x9524) #define GWUNIT_CLKGATE_DIS (1 << 16) @@ -3971,6 +3848,9 @@ enum { #define UNSLICE_UNIT_LEVEL_CLKGATE _MMIO(0x9434) #define VFUNIT_CLKGATE_DIS (1 << 20) +#define INF_UNIT_LEVEL_CLKGATE _MMIO(0x9560) +#define CGPSF_CLKGATE_DIS (1 << 3) + /* * Display engine regs */ @@ -4150,7 +4030,20 @@ enum { #define EDP_PSR_TP1_TIME_0us (3<<4) #define EDP_PSR_IDLE_FRAME_SHIFT 0 +/* Bspec claims those aren't shifted but stay at 0x64800 */ +#define EDP_PSR_IMR _MMIO(0x64834) +#define EDP_PSR_IIR _MMIO(0x64838) +#define EDP_PSR_ERROR(trans) (1 << (((trans) * 8 + 10) & 31)) +#define EDP_PSR_POST_EXIT(trans) (1 << (((trans) * 8 + 9) & 31)) +#define EDP_PSR_PRE_ENTRY(trans) (1 << (((trans) * 8 + 8) & 31)) + #define EDP_PSR_AUX_CTL _MMIO(dev_priv->psr_mmio_base + 0x10) +#define EDP_PSR_AUX_CTL_TIME_OUT_MASK (3 << 26) +#define EDP_PSR_AUX_CTL_MESSAGE_SIZE_MASK (0x1f << 20) +#define EDP_PSR_AUX_CTL_PRECHARGE_2US_MASK (0xf << 16) +#define EDP_PSR_AUX_CTL_ERROR_INTERRUPT (1 << 11) +#define EDP_PSR_AUX_CTL_BIT_CLOCK_2X_MASK (0x7ff) + #define EDP_PSR_AUX_DATA(i) _MMIO(dev_priv->psr_mmio_base + 0x14 + (i) * 4) /* 5 registers */ #define EDP_PSR_STATUS _MMIO(dev_priv->psr_mmio_base + 0x40) @@ -4180,17 +4073,19 @@ enum { #define EDP_PSR_PERF_CNT _MMIO(dev_priv->psr_mmio_base + 0x44) #define EDP_PSR_PERF_CNT_MASK 0xffffff -#define EDP_PSR_DEBUG _MMIO(dev_priv->psr_mmio_base + 0x60) +#define EDP_PSR_DEBUG _MMIO(dev_priv->psr_mmio_base + 0x60) /* PSR_MASK on SKL+ */ #define EDP_PSR_DEBUG_MASK_MAX_SLEEP (1<<28) #define EDP_PSR_DEBUG_MASK_LPSP (1<<27) #define EDP_PSR_DEBUG_MASK_MEMUP (1<<26) #define EDP_PSR_DEBUG_MASK_HPD (1<<25) #define EDP_PSR_DEBUG_MASK_DISP_REG_WRITE (1<<16) -#define EDP_PSR_DEBUG_EXIT_ON_PIXEL_UNDERRUN (1<<15) +#define EDP_PSR_DEBUG_EXIT_ON_PIXEL_UNDERRUN (1<<15) /* SKL+ */ #define EDP_PSR2_CTL _MMIO(0x6f900) #define EDP_PSR2_ENABLE (1<<31) #define EDP_SU_TRACK_ENABLE (1<<30) +#define EDP_Y_COORDINATE_VALID (1<<26) /* GLK and CNL+ */ +#define EDP_Y_COORDINATE_ENABLE (1<<25) /* GLK and CNL+ */ #define EDP_MAX_SU_DISABLE_TIME(t) ((t)<<20) #define EDP_MAX_SU_DISABLE_TIME_MASK (0x1f<<20) #define EDP_PSR2_TP2_TIME_500 (0<<8) @@ -4200,8 +4095,32 @@ enum { #define EDP_PSR2_TP2_TIME_MASK (3<<8) #define EDP_PSR2_FRAME_BEFORE_SU_SHIFT 4 #define EDP_PSR2_FRAME_BEFORE_SU_MASK (0xf<<4) -#define EDP_PSR2_IDLE_MASK 0xf #define EDP_PSR2_FRAME_BEFORE_SU(a) ((a)<<4) +#define EDP_PSR2_IDLE_FRAME_MASK 0xf +#define EDP_PSR2_IDLE_FRAME_SHIFT 0 + +#define _PSR_EVENT_TRANS_A 0x60848 +#define _PSR_EVENT_TRANS_B 0x61848 +#define _PSR_EVENT_TRANS_C 0x62848 +#define _PSR_EVENT_TRANS_D 0x63848 +#define _PSR_EVENT_TRANS_EDP 0x6F848 +#define PSR_EVENT(trans) _MMIO_TRANS2(trans, _PSR_EVENT_TRANS_A) +#define PSR_EVENT_PSR2_WD_TIMER_EXPIRE (1 << 17) +#define PSR_EVENT_PSR2_DISABLED (1 << 16) +#define PSR_EVENT_SU_DIRTY_FIFO_UNDERRUN (1 << 15) +#define PSR_EVENT_SU_CRC_FIFO_UNDERRUN (1 << 14) +#define PSR_EVENT_GRAPHICS_RESET (1 << 12) +#define PSR_EVENT_PCH_INTERRUPT (1 << 11) +#define PSR_EVENT_MEMORY_UP (1 << 10) +#define PSR_EVENT_FRONT_BUFFER_MODIFY (1 << 9) +#define PSR_EVENT_WD_TIMER_EXPIRE (1 << 8) +#define PSR_EVENT_PIPE_REGISTERS_UPDATE (1 << 6) +#define PSR_EVENT_REGISTER_UPDATE (1 << 5) +#define PSR_EVENT_HDCP_ENABLE (1 << 4) +#define PSR_EVENT_KVMR_SESSION_ENABLE (1 << 3) +#define PSR_EVENT_VBI_ENABLE (1 << 2) +#define PSR_EVENT_LPSP_MODE_EXIT (1 << 1) +#define PSR_EVENT_PSR_DISABLE (1 << 0) #define EDP_PSR2_STATUS _MMIO(0x6f940) #define EDP_PSR2_STATUS_STATE_MASK (0xf<<28) @@ -5265,8 +5184,6 @@ enum { #define DP_LINK_TRAIN_OFF (3 << 28) #define DP_LINK_TRAIN_MASK (3 << 28) #define DP_LINK_TRAIN_SHIFT 28 -#define DP_LINK_TRAIN_PAT_3_CHV (1 << 14) -#define DP_LINK_TRAIN_MASK_CHV ((3 << 28)|(1<<14)) /* CPT Link training mode */ #define DP_LINK_TRAIN_PAT_1_CPT (0 << 8) @@ -6009,6 +5926,7 @@ enum { #define CURSIZE _MMIO(0x700a0) /* 845/865 */ #define _CUR_FBC_CTL_A 0x700a0 /* ivb+ */ #define CUR_FBC_CTL_EN (1 << 31) +#define _CURASURFLIVE 0x700ac /* g4x+ */ #define _CURBCNTR 0x700c0 #define _CURBBASE 0x700c4 #define _CURBPOS 0x700c8 @@ -6025,6 +5943,7 @@ enum { #define CURBASE(pipe) _CURSOR2(pipe, _CURABASE) #define CURPOS(pipe) _CURSOR2(pipe, _CURAPOS) #define CUR_FBC_CTL(pipe) _CURSOR2(pipe, _CUR_FBC_CTL_A) +#define CURSURFLIVE(pipe) _CURSOR2(pipe, _CURASURFLIVE) #define CURSOR_A_OFFSET 0x70080 #define CURSOR_B_OFFSET 0x700c0 @@ -6492,9 +6411,9 @@ enum { #define _PLANE_COLOR_CTL_1_A 0x701CC /* GLK+ */ #define _PLANE_COLOR_CTL_2_A 0x702CC /* GLK+ */ #define _PLANE_COLOR_CTL_3_A 0x703CC /* GLK+ */ -#define PLANE_COLOR_PIPE_GAMMA_ENABLE (1 << 30) +#define PLANE_COLOR_PIPE_GAMMA_ENABLE (1 << 30) /* Pre-ICL */ #define PLANE_COLOR_YUV_RANGE_CORRECTION_DISABLE (1 << 28) -#define PLANE_COLOR_PIPE_CSC_ENABLE (1 << 23) +#define PLANE_COLOR_PIPE_CSC_ENABLE (1 << 23) /* Pre-ICL */ #define PLANE_COLOR_CSC_MODE_BYPASS (0 << 17) #define PLANE_COLOR_CSC_MODE_YUV601_TO_RGB709 (1 << 17) #define PLANE_COLOR_CSC_MODE_YUV709_TO_RGB709 (2 << 17) @@ -6589,6 +6508,9 @@ enum { #define _PLANE_BUF_CFG_1_B 0x7127c #define _PLANE_BUF_CFG_2_B 0x7137c +#define SKL_DDB_ENTRY_MASK 0x3FF +#define ICL_DDB_ENTRY_MASK 0x7FF +#define DDB_ENTRY_END_SHIFT 16 #define _PLANE_BUF_CFG_1(pipe) \ _PIPE(pipe, _PLANE_BUF_CFG_1_A, _PLANE_BUF_CFG_1_B) #define _PLANE_BUF_CFG_2(pipe) \ @@ -6779,6 +6701,8 @@ enum { #define PS_SCALER_MODE_MASK (3 << 28) #define PS_SCALER_MODE_DYN (0 << 28) #define PS_SCALER_MODE_HQ (1 << 28) +#define SKL_PS_SCALER_MODE_NV12 (2 << 28) +#define PS_SCALER_MODE_PLANAR (1 << 29) #define PS_PLANE_SEL_MASK (7 << 25) #define PS_PLANE_SEL(plane) (((plane) + 1) << 25) #define PS_FILTER_MASK (3 << 23) @@ -6950,6 +6874,7 @@ enum { #define DE_PCH_EVENT_IVB (1<<28) #define DE_DP_A_HOTPLUG_IVB (1<<27) #define DE_AUX_CHANNEL_A_IVB (1<<26) +#define DE_EDP_PSR_INT_HSW (1<<19) #define DE_SPRITEC_FLIP_DONE_IVB (1<<14) #define DE_PLANEC_FLIP_DONE_IVB (1<<13) #define DE_PIPEC_VBLANK_IVB (1<<10) @@ -7074,6 +6999,7 @@ enum { #define GEN8_DE_MISC_IIR _MMIO(0x44468) #define GEN8_DE_MISC_IER _MMIO(0x4446c) #define GEN8_DE_MISC_GSE (1 << 27) +#define GEN8_DE_EDP_PSR (1 << 19) #define GEN8_PCU_ISR _MMIO(0x444e0) #define GEN8_PCU_IMR _MMIO(0x444e4) @@ -7117,7 +7043,9 @@ enum { #define GEN11_INTR_IDENTITY_REG0 _MMIO(0x190060) #define GEN11_INTR_IDENTITY_REG1 _MMIO(0x190064) #define GEN11_INTR_DATA_VALID (1 << 31) -#define GEN11_INTR_ENGINE_MASK (0xffff) +#define GEN11_INTR_ENGINE_CLASS(x) (((x) & GENMASK(18, 16)) >> 16) +#define GEN11_INTR_ENGINE_INSTANCE(x) (((x) & GENMASK(25, 20)) >> 20) +#define GEN11_INTR_ENGINE_INTR(x) ((x) & 0xffff) #define GEN11_INTR_IDENTITY_REG(x) _MMIO(0x190060 + (x * 4)) @@ -7197,6 +7125,7 @@ enum { #define CHICKEN_TRANS_A 0x420c0 #define CHICKEN_TRANS_B 0x420c4 #define CHICKEN_TRANS(trans) _MMIO_TRANS(trans, CHICKEN_TRANS_A, CHICKEN_TRANS_B) +#define VSC_DATA_SEL_SOFTWARE_CONTROL (1<<25) /* GLK and CNL+ */ #define DDI_TRAINING_OVERRIDE_ENABLE (1<<19) #define DDI_TRAINING_OVERRIDE_VALUE (1<<18) #define DDIE_TRAINING_OVERRIDE_ENABLE (1<<17) /* CHICKEN_TRANS_A only */ @@ -7301,18 +7230,22 @@ enum { #define GEN7_L3CNTLREG3 _MMIO(0xB024) #define GEN7_L3_CHICKEN_MODE_REGISTER _MMIO(0xB030) -#define GEN7_WA_L3_CHICKEN_MODE 0x20000000 +#define GEN7_WA_L3_CHICKEN_MODE 0x20000000 +#define GEN10_L3_CHICKEN_MODE_REGISTER _MMIO(0xB114) +#define GEN11_I2M_WRITE_DISABLE (1 << 28) #define GEN7_L3SQCREG4 _MMIO(0xb034) #define L3SQ_URB_READ_CAM_MATCH_DISABLE (1<<27) #define GEN8_L3SQCREG4 _MMIO(0xb118) -#define GEN8_LQSC_RO_PERF_DIS (1<<27) -#define GEN8_LQSC_FLUSH_COHERENT_LINES (1<<21) +#define GEN11_LQSC_CLEAN_EVICT_DISABLE (1 << 6) +#define GEN8_LQSC_RO_PERF_DIS (1 << 27) +#define GEN8_LQSC_FLUSH_COHERENT_LINES (1 << 21) /* GEN8 chicken */ #define HDC_CHICKEN0 _MMIO(0x7300) #define CNL_HDC_CHICKEN0 _MMIO(0xE5F0) +#define ICL_HDC_MODE _MMIO(0xE5F4) #define HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE (1<<15) #define HDC_FENCE_DEST_SLM_DISABLE (1<<14) #define HDC_DONOT_FETCH_MEM_WHEN_MASKED (1<<11) @@ -8327,8 +8260,30 @@ enum { #define GEN8_DOP_CLOCK_GATE_GUC_ENABLE (1<<4) #define GEN8_DOP_CLOCK_GATE_MEDIA_ENABLE (1<<6) -#define GEN8_GARBCNTL _MMIO(0xB004) -#define GEN9_GAPS_TSV_CREDIT_DISABLE (1<<7) +#define GEN8_GARBCNTL _MMIO(0xB004) +#define GEN9_GAPS_TSV_CREDIT_DISABLE (1 << 7) +#define GEN11_ARBITRATION_PRIO_ORDER_MASK (0x3f << 22) +#define GEN11_HASH_CTRL_EXCL_MASK (0x7f << 0) +#define GEN11_HASH_CTRL_EXCL_BIT0 (1 << 0) + +#define GEN11_GLBLINVL _MMIO(0xB404) +#define GEN11_BANK_HASH_ADDR_EXCL_MASK (0x7f << 5) +#define GEN11_BANK_HASH_ADDR_EXCL_BIT0 (1 << 5) + +#define GEN10_DFR_RATIO_EN_AND_CHICKEN _MMIO(0x9550) +#define DFR_DISABLE (1 << 9) + +#define GEN11_GACB_PERF_CTRL _MMIO(0x4B80) +#define GEN11_HASH_CTRL_MASK (0x3 << 12 | 0xf << 0) +#define GEN11_HASH_CTRL_BIT0 (1 << 0) +#define GEN11_HASH_CTRL_BIT4 (1 << 12) + +#define GEN11_LSN_UNSLCVC _MMIO(0xB43C) +#define GEN11_LSN_UNSLCVC_GAFS_HALF_CL2_MAXALLOC (1 << 9) +#define GEN11_LSN_UNSLCVC_GAFS_HALF_SF_MAXALLOC (1 << 7) + +#define GAMW_ECO_DEV_RW_IA_REG _MMIO(0x4080) +#define GAMW_ECO_DEV_CTX_RELOAD_DISABLE (1 << 7) /* IVYBRIDGE DPF */ #define GEN7_L3CDERRST1(slice) _MMIO(0xB008 + (slice) * 0x200) /* L3CD Error Status 1 */ @@ -8837,6 +8792,12 @@ enum skl_power_gate { #define PORT_CLK_SEL_NONE (7<<29) #define PORT_CLK_SEL_MASK (7<<29) +/* On ICL+ this is the same as PORT_CLK_SEL, but all bits change. */ +#define DDI_CLK_SEL(port) PORT_CLK_SEL(port) +#define DDI_CLK_SEL_NONE (0x0 << 28) +#define DDI_CLK_SEL_MG (0x8 << 28) +#define DDI_CLK_SEL_MASK (0xF << 28) + /* Transcoder clock selection */ #define _TRANS_CLK_SEL_A 0x46140 #define _TRANS_CLK_SEL_B 0x46144 @@ -8967,6 +8928,7 @@ enum skl_power_gate { * CNL Clocks */ #define DPCLKA_CFGCR0 _MMIO(0x6C200) +#define DPCLKA_CFGCR0_ICL _MMIO(0x164280) #define DPCLKA_CFGCR0_DDI_CLK_OFF(port) (1 << ((port) == PORT_F ? 23 : \ (port)+10)) #define DPCLKA_CFGCR0_DDI_CLK_SEL_SHIFT(port) ((port) == PORT_F ? 21 : \ @@ -8983,10 +8945,141 @@ enum skl_power_gate { #define PLL_POWER_STATE (1 << 26) #define CNL_DPLL_ENABLE(pll) _MMIO_PLL(pll, DPLL0_ENABLE, DPLL1_ENABLE) +#define _MG_PLL1_ENABLE 0x46030 +#define _MG_PLL2_ENABLE 0x46034 +#define _MG_PLL3_ENABLE 0x46038 +#define _MG_PLL4_ENABLE 0x4603C +/* Bits are the same as DPLL0_ENABLE */ +#define MG_PLL_ENABLE(port) _MMIO_PORT((port) - PORT_C, _MG_PLL1_ENABLE, \ + _MG_PLL2_ENABLE) + +#define _MG_REFCLKIN_CTL_PORT1 0x16892C +#define _MG_REFCLKIN_CTL_PORT2 0x16992C +#define _MG_REFCLKIN_CTL_PORT3 0x16A92C +#define _MG_REFCLKIN_CTL_PORT4 0x16B92C +#define MG_REFCLKIN_CTL_OD_2_MUX(x) ((x) << 8) +#define MG_REFCLKIN_CTL(port) _MMIO_PORT((port) - PORT_C, \ + _MG_REFCLKIN_CTL_PORT1, \ + _MG_REFCLKIN_CTL_PORT2) + +#define _MG_CLKTOP2_CORECLKCTL1_PORT1 0x1688D8 +#define _MG_CLKTOP2_CORECLKCTL1_PORT2 0x1698D8 +#define _MG_CLKTOP2_CORECLKCTL1_PORT3 0x16A8D8 +#define _MG_CLKTOP2_CORECLKCTL1_PORT4 0x16B8D8 +#define MG_CLKTOP2_CORECLKCTL1_B_DIVRATIO(x) ((x) << 16) +#define MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO(x) ((x) << 8) +#define MG_CLKTOP2_CORECLKCTL1(port) _MMIO_PORT((port) - PORT_C, \ + _MG_CLKTOP2_CORECLKCTL1_PORT1, \ + _MG_CLKTOP2_CORECLKCTL1_PORT2) + +#define _MG_CLKTOP2_HSCLKCTL_PORT1 0x1688D4 +#define _MG_CLKTOP2_HSCLKCTL_PORT2 0x1698D4 +#define _MG_CLKTOP2_HSCLKCTL_PORT3 0x16A8D4 +#define _MG_CLKTOP2_HSCLKCTL_PORT4 0x16B8D4 +#define MG_CLKTOP2_HSCLKCTL_CORE_INPUTSEL(x) ((x) << 16) +#define MG_CLKTOP2_HSCLKCTL_TLINEDRV_CLKSEL(x) ((x) << 14) +#define MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO(x) ((x) << 12) +#define MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO(x) ((x) << 8) +#define MG_CLKTOP2_HSCLKCTL(port) _MMIO_PORT((port) - PORT_C, \ + _MG_CLKTOP2_HSCLKCTL_PORT1, \ + _MG_CLKTOP2_HSCLKCTL_PORT2) + +#define _MG_PLL_DIV0_PORT1 0x168A00 +#define _MG_PLL_DIV0_PORT2 0x169A00 +#define _MG_PLL_DIV0_PORT3 0x16AA00 +#define _MG_PLL_DIV0_PORT4 0x16BA00 +#define MG_PLL_DIV0_FRACNEN_H (1 << 30) +#define MG_PLL_DIV0_FBDIV_FRAC(x) ((x) << 8) +#define MG_PLL_DIV0_FBDIV_INT(x) ((x) << 0) +#define MG_PLL_DIV0(port) _MMIO_PORT((port) - PORT_C, _MG_PLL_DIV0_PORT1, \ + _MG_PLL_DIV0_PORT2) + +#define _MG_PLL_DIV1_PORT1 0x168A04 +#define _MG_PLL_DIV1_PORT2 0x169A04 +#define _MG_PLL_DIV1_PORT3 0x16AA04 +#define _MG_PLL_DIV1_PORT4 0x16BA04 +#define MG_PLL_DIV1_IREF_NDIVRATIO(x) ((x) << 16) +#define MG_PLL_DIV1_DITHER_DIV_1 (0 << 12) +#define MG_PLL_DIV1_DITHER_DIV_2 (1 << 12) +#define MG_PLL_DIV1_DITHER_DIV_4 (2 << 12) +#define MG_PLL_DIV1_DITHER_DIV_8 (3 << 12) +#define MG_PLL_DIV1_NDIVRATIO(x) ((x) << 4) +#define MG_PLL_DIV1_FBPREDIV(x) ((x) << 0) +#define MG_PLL_DIV1(port) _MMIO_PORT((port) - PORT_C, _MG_PLL_DIV1_PORT1, \ + _MG_PLL_DIV1_PORT2) + +#define _MG_PLL_LF_PORT1 0x168A08 +#define _MG_PLL_LF_PORT2 0x169A08 +#define _MG_PLL_LF_PORT3 0x16AA08 +#define _MG_PLL_LF_PORT4 0x16BA08 +#define MG_PLL_LF_TDCTARGETCNT(x) ((x) << 24) +#define MG_PLL_LF_AFCCNTSEL_256 (0 << 20) +#define MG_PLL_LF_AFCCNTSEL_512 (1 << 20) +#define MG_PLL_LF_GAINCTRL(x) ((x) << 16) +#define MG_PLL_LF_INT_COEFF(x) ((x) << 8) +#define MG_PLL_LF_PROP_COEFF(x) ((x) << 0) +#define MG_PLL_LF(port) _MMIO_PORT((port) - PORT_C, _MG_PLL_LF_PORT1, \ + _MG_PLL_LF_PORT2) + +#define _MG_PLL_FRAC_LOCK_PORT1 0x168A0C +#define _MG_PLL_FRAC_LOCK_PORT2 0x169A0C +#define _MG_PLL_FRAC_LOCK_PORT3 0x16AA0C +#define _MG_PLL_FRAC_LOCK_PORT4 0x16BA0C +#define MG_PLL_FRAC_LOCK_TRUELOCK_CRIT_32 (1 << 18) +#define MG_PLL_FRAC_LOCK_EARLYLOCK_CRIT_32 (1 << 16) +#define MG_PLL_FRAC_LOCK_LOCKTHRESH(x) ((x) << 11) +#define MG_PLL_FRAC_LOCK_DCODITHEREN (1 << 10) +#define MG_PLL_FRAC_LOCK_FEEDFWRDCAL_EN (1 << 8) +#define MG_PLL_FRAC_LOCK_FEEDFWRDGAIN(x) ((x) << 0) +#define MG_PLL_FRAC_LOCK(port) _MMIO_PORT((port) - PORT_C, \ + _MG_PLL_FRAC_LOCK_PORT1, \ + _MG_PLL_FRAC_LOCK_PORT2) + +#define _MG_PLL_SSC_PORT1 0x168A10 +#define _MG_PLL_SSC_PORT2 0x169A10 +#define _MG_PLL_SSC_PORT3 0x16AA10 +#define _MG_PLL_SSC_PORT4 0x16BA10 +#define MG_PLL_SSC_EN (1 << 28) +#define MG_PLL_SSC_TYPE(x) ((x) << 26) +#define MG_PLL_SSC_STEPLENGTH(x) ((x) << 16) +#define MG_PLL_SSC_STEPNUM(x) ((x) << 10) +#define MG_PLL_SSC_FLLEN (1 << 9) +#define MG_PLL_SSC_STEPSIZE(x) ((x) << 0) +#define MG_PLL_SSC(port) _MMIO_PORT((port) - PORT_C, _MG_PLL_SSC_PORT1, \ + _MG_PLL_SSC_PORT2) + +#define _MG_PLL_BIAS_PORT1 0x168A14 +#define _MG_PLL_BIAS_PORT2 0x169A14 +#define _MG_PLL_BIAS_PORT3 0x16AA14 +#define _MG_PLL_BIAS_PORT4 0x16BA14 +#define MG_PLL_BIAS_BIAS_GB_SEL(x) ((x) << 30) +#define MG_PLL_BIAS_INIT_DCOAMP(x) ((x) << 24) +#define MG_PLL_BIAS_BIAS_BONUS(x) ((x) << 16) +#define MG_PLL_BIAS_BIASCAL_EN (1 << 15) +#define MG_PLL_BIAS_CTRIM(x) ((x) << 8) +#define MG_PLL_BIAS_VREF_RDAC(x) ((x) << 5) +#define MG_PLL_BIAS_IREFTRIM(x) ((x) << 0) +#define MG_PLL_BIAS(port) _MMIO_PORT((port) - PORT_C, _MG_PLL_BIAS_PORT1, \ + _MG_PLL_BIAS_PORT2) + +#define _MG_PLL_TDC_COLDST_BIAS_PORT1 0x168A18 +#define _MG_PLL_TDC_COLDST_BIAS_PORT2 0x169A18 +#define _MG_PLL_TDC_COLDST_BIAS_PORT3 0x16AA18 +#define _MG_PLL_TDC_COLDST_BIAS_PORT4 0x16BA18 +#define MG_PLL_TDC_COLDST_IREFINT_EN (1 << 27) +#define MG_PLL_TDC_COLDST_REFBIAS_START_PULSE_W(x) ((x) << 17) +#define MG_PLL_TDC_COLDST_COLDSTART (1 << 16) +#define MG_PLL_TDC_TDCOVCCORR_EN (1 << 2) +#define MG_PLL_TDC_TDCSEL(x) ((x) << 0) +#define MG_PLL_TDC_COLDST_BIAS(port) _MMIO_PORT((port) - PORT_C, \ + _MG_PLL_TDC_COLDST_BIAS_PORT1, \ + _MG_PLL_TDC_COLDST_BIAS_PORT2) + #define _CNL_DPLL0_CFGCR0 0x6C000 #define _CNL_DPLL1_CFGCR0 0x6C080 #define DPLL_CFGCR0_HDMI_MODE (1 << 30) #define DPLL_CFGCR0_SSC_ENABLE (1 << 29) +#define DPLL_CFGCR0_SSC_ENABLE_ICL (1 << 25) #define DPLL_CFGCR0_LINK_RATE_MASK (0xf << 25) #define DPLL_CFGCR0_LINK_RATE_2700 (0 << 25) #define DPLL_CFGCR0_LINK_RATE_1350 (1 << 25) @@ -9020,8 +9113,19 @@ enum skl_power_gate { #define DPLL_CFGCR1_PDIV_5 (4 << 2) #define DPLL_CFGCR1_PDIV_7 (8 << 2) #define DPLL_CFGCR1_CENTRAL_FREQ (3 << 0) +#define DPLL_CFGCR1_CENTRAL_FREQ_8400 (3 << 0) #define CNL_DPLL_CFGCR1(pll) _MMIO_PLL(pll, _CNL_DPLL0_CFGCR1, _CNL_DPLL1_CFGCR1) +#define _ICL_DPLL0_CFGCR0 0x164000 +#define _ICL_DPLL1_CFGCR0 0x164080 +#define ICL_DPLL_CFGCR0(pll) _MMIO_PLL(pll, _ICL_DPLL0_CFGCR0, \ + _ICL_DPLL1_CFGCR0) + +#define _ICL_DPLL0_CFGCR1 0x164004 +#define _ICL_DPLL1_CFGCR1 0x164084 +#define ICL_DPLL_CFGCR1(pll) _MMIO_PLL(pll, _ICL_DPLL0_CFGCR1, \ + _ICL_DPLL1_CFGCR1) + /* BXT display engine PLL */ #define BXT_DE_PLL_CTL _MMIO(0x6d000) #define BXT_DE_PLL_RATIO(x) (x) /* {60,65,100} * 19.2MHz */ @@ -9793,6 +9897,13 @@ enum skl_power_gate { #define GEN9_MFX1_MOCS(i) _MMIO(0xca00 + (i) * 4) /* Media 1 MOCS registers */ #define GEN9_VEBOX_MOCS(i) _MMIO(0xcb00 + (i) * 4) /* Video MOCS registers */ #define GEN9_BLT_MOCS(i) _MMIO(0xcc00 + (i) * 4) /* Blitter MOCS registers */ +/* Media decoder 2 MOCS registers */ +#define GEN11_MFX2_MOCS(i) _MMIO(0x10000 + (i) * 4) + +#define GEN10_SCRATCH_LNCF2 _MMIO(0xb0a0) +#define PMFLUSHDONE_LNICRSDROP (1 << 20) +#define PMFLUSH_GAPL3UNBLOCK (1 << 21) +#define PMFLUSHDONE_LNEBLK (1 << 22) /* gamt regs */ #define GEN8_L3_LRA_1_GPGPU _MMIO(0x4dd4) diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index 282f576..8928894 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -49,7 +49,7 @@ static const char *i915_fence_get_timeline_name(struct dma_fence *fence) if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags)) return "signaled"; - return to_request(fence)->timeline->common->name; + return to_request(fence)->timeline->name; } static bool i915_fence_signaled(struct dma_fence *fence) @@ -59,11 +59,7 @@ static bool i915_fence_signaled(struct dma_fence *fence) static bool i915_fence_enable_signaling(struct dma_fence *fence) { - if (i915_fence_signaled(fence)) - return false; - - intel_engine_enable_signaling(to_request(fence), true); - return !i915_fence_signaled(fence); + return intel_engine_enable_signaling(to_request(fence), true); } static signed long i915_fence_wait(struct dma_fence *fence, @@ -129,22 +125,22 @@ i915_dependency_free(struct drm_i915_private *i915, } static void -__i915_priotree_add_dependency(struct i915_priotree *pt, - struct i915_priotree *signal, - struct i915_dependency *dep, - unsigned long flags) +__i915_sched_node_add_dependency(struct i915_sched_node *node, + struct i915_sched_node *signal, + struct i915_dependency *dep, + unsigned long flags) { INIT_LIST_HEAD(&dep->dfs_link); list_add(&dep->wait_link, &signal->waiters_list); - list_add(&dep->signal_link, &pt->signalers_list); + list_add(&dep->signal_link, &node->signalers_list); dep->signaler = signal; dep->flags = flags; } static int -i915_priotree_add_dependency(struct drm_i915_private *i915, - struct i915_priotree *pt, - struct i915_priotree *signal) +i915_sched_node_add_dependency(struct drm_i915_private *i915, + struct i915_sched_node *node, + struct i915_sched_node *signal) { struct i915_dependency *dep; @@ -152,16 +148,18 @@ i915_priotree_add_dependency(struct drm_i915_private *i915, if (!dep) return -ENOMEM; - __i915_priotree_add_dependency(pt, signal, dep, I915_DEPENDENCY_ALLOC); + __i915_sched_node_add_dependency(node, signal, dep, + I915_DEPENDENCY_ALLOC); return 0; } static void -i915_priotree_fini(struct drm_i915_private *i915, struct i915_priotree *pt) +i915_sched_node_fini(struct drm_i915_private *i915, + struct i915_sched_node *node) { - struct i915_dependency *dep, *next; + struct i915_dependency *dep, *tmp; - GEM_BUG_ON(!list_empty(&pt->link)); + GEM_BUG_ON(!list_empty(&node->link)); /* * Everyone we depended upon (the fences we wait to be signaled) @@ -169,8 +167,8 @@ i915_priotree_fini(struct drm_i915_private *i915, struct i915_priotree *pt) * However, retirement is run independently on each timeline and * so we may be called out-of-order. */ - list_for_each_entry_safe(dep, next, &pt->signalers_list, signal_link) { - GEM_BUG_ON(!i915_priotree_signaled(dep->signaler)); + list_for_each_entry_safe(dep, tmp, &node->signalers_list, signal_link) { + GEM_BUG_ON(!i915_sched_node_signaled(dep->signaler)); GEM_BUG_ON(!list_empty(&dep->dfs_link)); list_del(&dep->wait_link); @@ -179,8 +177,8 @@ i915_priotree_fini(struct drm_i915_private *i915, struct i915_priotree *pt) } /* Remove ourselves from everyone who depends upon us */ - list_for_each_entry_safe(dep, next, &pt->waiters_list, wait_link) { - GEM_BUG_ON(dep->signaler != pt); + list_for_each_entry_safe(dep, tmp, &node->waiters_list, wait_link) { + GEM_BUG_ON(dep->signaler != node); GEM_BUG_ON(!list_empty(&dep->dfs_link)); list_del(&dep->signal_link); @@ -190,17 +188,18 @@ i915_priotree_fini(struct drm_i915_private *i915, struct i915_priotree *pt) } static void -i915_priotree_init(struct i915_priotree *pt) +i915_sched_node_init(struct i915_sched_node *node) { - INIT_LIST_HEAD(&pt->signalers_list); - INIT_LIST_HEAD(&pt->waiters_list); - INIT_LIST_HEAD(&pt->link); - pt->priority = I915_PRIORITY_INVALID; + INIT_LIST_HEAD(&node->signalers_list); + INIT_LIST_HEAD(&node->waiters_list); + INIT_LIST_HEAD(&node->link); + node->attr.priority = I915_PRIORITY_INVALID; } static int reset_all_global_seqno(struct drm_i915_private *i915, u32 seqno) { struct intel_engine_cs *engine; + struct i915_timeline *timeline; enum intel_engine_id id; int ret; @@ -211,30 +210,37 @@ static int reset_all_global_seqno(struct drm_i915_private *i915, u32 seqno) if (ret) return ret; + GEM_BUG_ON(i915->gt.active_requests); + /* If the seqno wraps around, we need to clear the breadcrumb rbtree */ for_each_engine(engine, i915, id) { - struct i915_gem_timeline *timeline; - struct intel_timeline *tl = engine->timeline; + GEM_TRACE("%s seqno %d (current %d) -> %d\n", + engine->name, + engine->timeline.seqno, + intel_engine_get_seqno(engine), + seqno); - if (!i915_seqno_passed(seqno, tl->seqno)) { + if (!i915_seqno_passed(seqno, engine->timeline.seqno)) { /* Flush any waiters before we reuse the seqno */ intel_engine_disarm_breadcrumbs(engine); + intel_engine_init_hangcheck(engine); GEM_BUG_ON(!list_empty(&engine->breadcrumbs.signals)); } /* Check we are idle before we fiddle with hw state! */ GEM_BUG_ON(!intel_engine_is_idle(engine)); - GEM_BUG_ON(i915_gem_active_isset(&engine->timeline->last_request)); + GEM_BUG_ON(i915_gem_active_isset(&engine->timeline.last_request)); /* Finally reset hw state */ intel_engine_init_global_seqno(engine, seqno); - tl->seqno = seqno; - - list_for_each_entry(timeline, &i915->gt.timelines, link) - memset(timeline->engine[id].global_sync, 0, - sizeof(timeline->engine[id].global_sync)); + engine->timeline.seqno = seqno; } + list_for_each_entry(timeline, &i915->gt.timelines, link) + memset(timeline->global_sync, 0, sizeof(timeline->global_sync)); + + i915->gt.request_serial = seqno; + return 0; } @@ -251,83 +257,37 @@ int i915_gem_set_global_seqno(struct drm_device *dev, u32 seqno) return reset_all_global_seqno(i915, seqno - 1); } -static void mark_busy(struct drm_i915_private *i915) +static int reserve_gt(struct drm_i915_private *i915) { - if (i915->gt.awake) - return; - - GEM_BUG_ON(!i915->gt.active_requests); - - intel_runtime_pm_get_noresume(i915); + int ret; /* - * It seems that the DMC likes to transition between the DC states a lot - * when there are no connected displays (no active power domains) during - * command submission. - * - * This activity has negative impact on the performance of the chip with - * huge latencies observed in the interrupt handler and elsewhere. + * Reservation is fine until we may need to wrap around * - * Work around it by grabbing a GT IRQ power domain whilst there is any - * GT activity, preventing any DC state transitions. + * By incrementing the serial for every request, we know that no + * individual engine may exceed that serial (as each is reset to 0 + * on any wrap). This protects even the most pessimistic of migrations + * of every request from all engines onto just one. */ - intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ); - - i915->gt.awake = true; - if (unlikely(++i915->gt.epoch == 0)) /* keep 0 as invalid */ - i915->gt.epoch = 1; - - intel_enable_gt_powersave(i915); - i915_update_gfx_val(i915); - if (INTEL_GEN(i915) >= 6) - gen6_rps_busy(i915); - i915_pmu_gt_unparked(i915); - - intel_engines_unpark(i915); - - i915_queue_hangcheck(i915); - - queue_delayed_work(i915->wq, - &i915->gt.retire_work, - round_jiffies_up_relative(HZ)); -} - -static int reserve_engine(struct intel_engine_cs *engine) -{ - struct drm_i915_private *i915 = engine->i915; - u32 active = ++engine->timeline->inflight_seqnos; - u32 seqno = engine->timeline->seqno; - int ret; - - /* Reservation is fine until we need to wrap around */ - if (unlikely(add_overflows(seqno, active))) { + while (unlikely(++i915->gt.request_serial == 0)) { ret = reset_all_global_seqno(i915, 0); if (ret) { - engine->timeline->inflight_seqnos--; + i915->gt.request_serial--; return ret; } } if (!i915->gt.active_requests++) - mark_busy(i915); + i915_gem_unpark(i915); return 0; } -static void unreserve_engine(struct intel_engine_cs *engine) +static void unreserve_gt(struct drm_i915_private *i915) { - struct drm_i915_private *i915 = engine->i915; - - if (!--i915->gt.active_requests) { - /* Cancel the mark_busy() from our reserve_engine() */ - GEM_BUG_ON(!i915->gt.awake); - mod_delayed_work(i915->wq, - &i915->gt.idle_work, - msecs_to_jiffies(100)); - } - - GEM_BUG_ON(!engine->timeline->inflight_seqnos); - engine->timeline->inflight_seqnos--; + GEM_BUG_ON(!i915->gt.active_requests); + if (!--i915->gt.active_requests) + i915_gem_park(i915); } void i915_gem_retire_noop(struct i915_gem_active *active, @@ -338,6 +298,7 @@ void i915_gem_retire_noop(struct i915_gem_active *active, static void advance_ring(struct i915_request *request) { + struct intel_ring *ring = request->ring; unsigned int tail; /* @@ -349,7 +310,8 @@ static void advance_ring(struct i915_request *request) * Note this requires that we are always called in request * completion order. */ - if (list_is_last(&request->ring_link, &request->ring->request_list)) { + GEM_BUG_ON(!list_is_first(&request->ring_link, &ring->request_list)); + if (list_is_last(&request->ring_link, &ring->request_list)) { /* * We may race here with execlists resubmitting this request * as we retire it. The resubmission will move the ring->tail @@ -358,13 +320,14 @@ static void advance_ring(struct i915_request *request) * is just about to be. Either works, if we miss the last two * noops - they are safe to be replayed on a reset. */ - tail = READ_ONCE(request->ring->tail); + tail = READ_ONCE(request->tail); + list_del(&ring->active_link); } else { tail = request->postfix; } - list_del(&request->ring_link); + list_del_init(&request->ring_link); - request->ring->head = tail; + ring->head = tail; } static void free_capture_list(struct i915_request *request) @@ -380,25 +343,84 @@ static void free_capture_list(struct i915_request *request) } } +static void __retire_engine_request(struct intel_engine_cs *engine, + struct i915_request *rq) +{ + GEM_TRACE("%s(%s) fence %llx:%d, global=%d, current %d\n", + __func__, engine->name, + rq->fence.context, rq->fence.seqno, + rq->global_seqno, + intel_engine_get_seqno(engine)); + + GEM_BUG_ON(!i915_request_completed(rq)); + + local_irq_disable(); + + spin_lock(&engine->timeline.lock); + GEM_BUG_ON(!list_is_first(&rq->link, &engine->timeline.requests)); + list_del_init(&rq->link); + spin_unlock(&engine->timeline.lock); + + spin_lock(&rq->lock); + if (!test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags)) + dma_fence_signal_locked(&rq->fence); + if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &rq->fence.flags)) + intel_engine_cancel_signaling(rq); + if (rq->waitboost) { + GEM_BUG_ON(!atomic_read(&rq->i915->gt_pm.rps.num_waiters)); + atomic_dec(&rq->i915->gt_pm.rps.num_waiters); + } + spin_unlock(&rq->lock); + + local_irq_enable(); + + /* + * The backing object for the context is done after switching to the + * *next* context. Therefore we cannot retire the previous context until + * the next context has already started running. However, since we + * cannot take the required locks at i915_request_submit() we + * defer the unpinning of the active context to now, retirement of + * the subsequent request. + */ + if (engine->last_retired_context) + intel_context_unpin(engine->last_retired_context, engine); + engine->last_retired_context = rq->ctx; +} + +static void __retire_engine_upto(struct intel_engine_cs *engine, + struct i915_request *rq) +{ + struct i915_request *tmp; + + if (list_empty(&rq->link)) + return; + + do { + tmp = list_first_entry(&engine->timeline.requests, + typeof(*tmp), link); + + GEM_BUG_ON(tmp->engine != engine); + __retire_engine_request(engine, tmp); + } while (tmp != rq); +} + static void i915_request_retire(struct i915_request *request) { - struct intel_engine_cs *engine = request->engine; struct i915_gem_active *active, *next; + GEM_TRACE("%s fence %llx:%d, global=%d, current %d\n", + request->engine->name, + request->fence.context, request->fence.seqno, + request->global_seqno, + intel_engine_get_seqno(request->engine)); + lockdep_assert_held(&request->i915->drm.struct_mutex); GEM_BUG_ON(!i915_sw_fence_signaled(&request->submit)); GEM_BUG_ON(!i915_request_completed(request)); - GEM_BUG_ON(!request->i915->gt.active_requests); trace_i915_request_retire(request); - spin_lock_irq(&engine->timeline->lock); - list_del_init(&request->link); - spin_unlock_irq(&engine->timeline->lock); - - unreserve_engine(request->engine); advance_ring(request); - free_capture_list(request); /* @@ -434,73 +456,74 @@ static void i915_request_retire(struct i915_request *request) /* Retirement decays the ban score as it is a sign of ctx progress */ atomic_dec_if_positive(&request->ctx->ban_score); + intel_context_unpin(request->ctx, request->engine); - /* - * The backing object for the context is done after switching to the - * *next* context. Therefore we cannot retire the previous context until - * the next context has already started running. However, since we - * cannot take the required locks at i915_request_submit() we - * defer the unpinning of the active context to now, retirement of - * the subsequent request. - */ - if (engine->last_retired_context) - engine->context_unpin(engine, engine->last_retired_context); - engine->last_retired_context = request->ctx; + __retire_engine_upto(request->engine, request); - spin_lock_irq(&request->lock); - if (!test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &request->fence.flags)) - dma_fence_signal_locked(&request->fence); - if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &request->fence.flags)) - intel_engine_cancel_signaling(request); - if (request->waitboost) { - GEM_BUG_ON(!atomic_read(&request->i915->gt_pm.rps.num_waiters)); - atomic_dec(&request->i915->gt_pm.rps.num_waiters); - } - spin_unlock_irq(&request->lock); + unreserve_gt(request->i915); - i915_priotree_fini(request->i915, &request->priotree); + i915_sched_node_fini(request->i915, &request->sched); i915_request_put(request); } void i915_request_retire_upto(struct i915_request *rq) { - struct intel_engine_cs *engine = rq->engine; + struct intel_ring *ring = rq->ring; struct i915_request *tmp; + GEM_TRACE("%s fence %llx:%d, global=%d, current %d\n", + rq->engine->name, + rq->fence.context, rq->fence.seqno, + rq->global_seqno, + intel_engine_get_seqno(rq->engine)); + lockdep_assert_held(&rq->i915->drm.struct_mutex); GEM_BUG_ON(!i915_request_completed(rq)); - if (list_empty(&rq->link)) + if (list_empty(&rq->ring_link)) return; do { - tmp = list_first_entry(&engine->timeline->requests, - typeof(*tmp), link); + tmp = list_first_entry(&ring->request_list, + typeof(*tmp), ring_link); i915_request_retire(tmp); } while (tmp != rq); } -static u32 timeline_get_seqno(struct intel_timeline *tl) +static u32 timeline_get_seqno(struct i915_timeline *tl) { return ++tl->seqno; } +static void move_to_timeline(struct i915_request *request, + struct i915_timeline *timeline) +{ + GEM_BUG_ON(request->timeline == &request->engine->timeline); + lockdep_assert_held(&request->engine->timeline.lock); + + spin_lock_nested(&request->timeline->lock, SINGLE_DEPTH_NESTING); + list_move_tail(&request->link, &timeline->requests); + spin_unlock(&request->timeline->lock); +} + void __i915_request_submit(struct i915_request *request) { struct intel_engine_cs *engine = request->engine; - struct intel_timeline *timeline; u32 seqno; + GEM_TRACE("%s fence %llx:%d -> global=%d, current %d\n", + engine->name, + request->fence.context, request->fence.seqno, + engine->timeline.seqno + 1, + intel_engine_get_seqno(engine)); + GEM_BUG_ON(!irqs_disabled()); - lockdep_assert_held(&engine->timeline->lock); + lockdep_assert_held(&engine->timeline.lock); - /* Transfer from per-context onto the global per-engine timeline */ - timeline = engine->timeline; - GEM_BUG_ON(timeline == request->timeline); GEM_BUG_ON(request->global_seqno); - seqno = timeline_get_seqno(timeline); + seqno = timeline_get_seqno(&engine->timeline); GEM_BUG_ON(!seqno); GEM_BUG_ON(i915_seqno_passed(intel_engine_get_seqno(engine), seqno)); @@ -514,9 +537,8 @@ void __i915_request_submit(struct i915_request *request) engine->emit_breadcrumb(request, request->ring->vaddr + request->postfix); - spin_lock(&request->timeline->lock); - list_move_tail(&request->link, &timeline->requests); - spin_unlock(&request->timeline->lock); + /* Transfer from per-context onto the global per-engine timeline */ + move_to_timeline(request, &engine->timeline); trace_i915_request_execute(request); @@ -529,30 +551,35 @@ void i915_request_submit(struct i915_request *request) unsigned long flags; /* Will be called from irq-context when using foreign fences. */ - spin_lock_irqsave(&engine->timeline->lock, flags); + spin_lock_irqsave(&engine->timeline.lock, flags); __i915_request_submit(request); - spin_unlock_irqrestore(&engine->timeline->lock, flags); + spin_unlock_irqrestore(&engine->timeline.lock, flags); } void __i915_request_unsubmit(struct i915_request *request) { struct intel_engine_cs *engine = request->engine; - struct intel_timeline *timeline; + + GEM_TRACE("%s fence %llx:%d <- global=%d, current %d\n", + engine->name, + request->fence.context, request->fence.seqno, + request->global_seqno, + intel_engine_get_seqno(engine)); GEM_BUG_ON(!irqs_disabled()); - lockdep_assert_held(&engine->timeline->lock); + lockdep_assert_held(&engine->timeline.lock); /* * Only unwind in reverse order, required so that the per-context list * is kept in seqno/ring order. */ GEM_BUG_ON(!request->global_seqno); - GEM_BUG_ON(request->global_seqno != engine->timeline->seqno); + GEM_BUG_ON(request->global_seqno != engine->timeline.seqno); GEM_BUG_ON(i915_seqno_passed(intel_engine_get_seqno(engine), request->global_seqno)); - engine->timeline->seqno--; + engine->timeline.seqno--; /* We may be recursing from the signal callback of another i915 fence */ spin_lock_nested(&request->lock, SINGLE_DEPTH_NESTING); @@ -562,12 +589,7 @@ void __i915_request_unsubmit(struct i915_request *request) spin_unlock(&request->lock); /* Transfer back from the global per-engine timeline to per-context */ - timeline = request->timeline; - GEM_BUG_ON(timeline == engine->timeline); - - spin_lock(&timeline->lock); - list_move(&request->link, &timeline->requests); - spin_unlock(&timeline->lock); + move_to_timeline(request, request->timeline); /* * We don't need to wake_up any waiters on request->execute, they @@ -584,11 +606,11 @@ void i915_request_unsubmit(struct i915_request *request) unsigned long flags; /* Will be called from irq-context when using foreign fences. */ - spin_lock_irqsave(&engine->timeline->lock, flags); + spin_lock_irqsave(&engine->timeline.lock, flags); __i915_request_unsubmit(request); - spin_unlock_irqrestore(&engine->timeline->lock, flags); + spin_unlock_irqrestore(&engine->timeline.lock, flags); } static int __i915_sw_fence_call @@ -659,12 +681,12 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) * GGTT space, so do this first before we reserve a seqno for * ourselves. */ - ring = engine->context_pin(engine, ctx); + ring = intel_context_pin(ctx, engine); if (IS_ERR(ring)) return ERR_CAST(ring); GEM_BUG_ON(!ring); - ret = reserve_engine(engine); + ret = reserve_gt(i915); if (ret) goto err_unpin; @@ -672,10 +694,10 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) if (ret) goto err_unreserve; - /* Move the oldest request to the slab-cache (if not in use!) */ - rq = list_first_entry_or_null(&engine->timeline->requests, - typeof(*rq), link); - if (rq && i915_request_completed(rq)) + /* Move our oldest request to the slab-cache (if not in use!) */ + rq = list_first_entry(&ring->request_list, typeof(*rq), ring_link); + if (!list_is_last(&rq->ring_link, &ring->request_list) && + i915_request_completed(rq)) i915_request_retire(rq); /* @@ -735,8 +757,13 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) } } - rq->timeline = i915_gem_context_lookup_timeline(ctx, engine); - GEM_BUG_ON(rq->timeline == engine->timeline); + INIT_LIST_HEAD(&rq->active_list); + rq->i915 = i915; + rq->engine = engine; + rq->ctx = ctx; + rq->ring = ring; + rq->timeline = ring->timeline; + GEM_BUG_ON(rq->timeline == &engine->timeline); spin_lock_init(&rq->lock); dma_fence_init(&rq->fence, @@ -749,13 +776,7 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) i915_sw_fence_init(&i915_request_get(rq)->submit, submit_notify); init_waitqueue_head(&rq->execute); - i915_priotree_init(&rq->priotree); - - INIT_LIST_HEAD(&rq->active_list); - rq->i915 = i915; - rq->engine = engine; - rq->ctx = ctx; - rq->ring = ring; + i915_sched_node_init(&rq->sched); /* No zalloc, must clear what we need by hand */ rq->global_seqno = 0; @@ -792,6 +813,9 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) if (ret) goto err_unwind; + /* Keep a second pin for the dual retirement along engine and ring */ + __intel_context_pin(rq->ctx, engine); + /* Check that we didn't interrupt ourselves with a new request */ GEM_BUG_ON(rq->timeline->seqno != rq->fence.seqno); return rq; @@ -801,14 +825,14 @@ err_unwind: /* Make sure we didn't add ourselves to external state before freeing */ GEM_BUG_ON(!list_empty(&rq->active_list)); - GEM_BUG_ON(!list_empty(&rq->priotree.signalers_list)); - GEM_BUG_ON(!list_empty(&rq->priotree.waiters_list)); + GEM_BUG_ON(!list_empty(&rq->sched.signalers_list)); + GEM_BUG_ON(!list_empty(&rq->sched.waiters_list)); kmem_cache_free(i915->requests, rq); err_unreserve: - unreserve_engine(engine); + unreserve_gt(i915); err_unpin: - engine->context_unpin(engine, ctx); + intel_context_unpin(ctx, engine); return ERR_PTR(ret); } @@ -824,9 +848,9 @@ i915_request_await_request(struct i915_request *to, struct i915_request *from) return 0; if (to->engine->schedule) { - ret = i915_priotree_add_dependency(to->i915, - &to->priotree, - &from->priotree); + ret = i915_sched_node_add_dependency(to->i915, + &to->sched, + &from->sched); if (ret < 0) return ret; } @@ -904,7 +928,7 @@ i915_request_await_dma_fence(struct i915_request *rq, struct dma_fence *fence) /* Squash repeated waits to the same timelines */ if (fence->context != rq->i915->mm.unordered_timeline && - intel_timeline_sync_is_later(rq->timeline, fence)) + i915_timeline_sync_is_later(rq->timeline, fence)) continue; if (dma_fence_is_i915(fence)) @@ -918,7 +942,7 @@ i915_request_await_dma_fence(struct i915_request *rq, struct dma_fence *fence) /* Record the latest fence used against each timeline */ if (fence->context != rq->i915->mm.unordered_timeline) - intel_timeline_sync_set(rq->timeline, fence); + i915_timeline_sync_set(rq->timeline, fence); } while (--nchild); return 0; @@ -995,11 +1019,14 @@ void __i915_request_add(struct i915_request *request, bool flush_caches) { struct intel_engine_cs *engine = request->engine; struct intel_ring *ring = request->ring; - struct intel_timeline *timeline = request->timeline; + struct i915_timeline *timeline = request->timeline; struct i915_request *prev; u32 *cs; int err; + GEM_TRACE("%s fence %llx:%d\n", + engine->name, request->fence.context, request->fence.seqno); + lockdep_assert_held(&request->i915->drm.struct_mutex); trace_i915_request_add(request); @@ -1054,10 +1081,10 @@ void __i915_request_add(struct i915_request *request, bool flush_caches) i915_sw_fence_await_sw_fence(&request->submit, &prev->submit, &request->submitq); if (engine->schedule) - __i915_priotree_add_dependency(&request->priotree, - &prev->priotree, - &request->dep, - 0); + __i915_sched_node_add_dependency(&request->sched, + &prev->sched, + &request->dep, + 0); } spin_lock_irq(&timeline->lock); @@ -1068,6 +1095,8 @@ void __i915_request_add(struct i915_request *request, bool flush_caches) i915_gem_active_set(&timeline->last_request, request); list_add_tail(&request->ring_link, &ring->request_list); + if (list_is_first(&request->ring_link, &ring->request_list)) + list_add(&ring->active_link, &request->i915->gt.active_rings); request->emitted_jiffies = jiffies; /* @@ -1081,12 +1110,11 @@ void __i915_request_add(struct i915_request *request, bool flush_caches) * decide whether to preempt the entire chain so that it is ready to * run at the earliest possible convenience. */ - rcu_read_lock(); + local_bh_disable(); + rcu_read_lock(); /* RCU serialisation for set-wedged protection */ if (engine->schedule) - engine->schedule(request, request->ctx->priority); + engine->schedule(request, &request->ctx->sched); rcu_read_unlock(); - - local_bh_disable(); i915_sw_fence_commit(&request->submit); local_bh_enable(); /* Kick the execlists tasklet if just scheduled */ @@ -1206,11 +1234,13 @@ static bool __i915_spin_request(const struct i915_request *rq, static bool __i915_wait_request_check_and_reset(struct i915_request *request) { - if (likely(!i915_reset_handoff(&request->i915->gpu_error))) + struct i915_gpu_error *error = &request->i915->gpu_error; + + if (likely(!i915_reset_handoff(error))) return false; __set_current_state(TASK_RUNNING); - i915_reset(request->i915, 0); + i915_reset(request->i915, error->stalled_mask, error->reason); return true; } @@ -1373,38 +1403,30 @@ complete: return timeout; } -static void engine_retire_requests(struct intel_engine_cs *engine) +static void ring_retire_requests(struct intel_ring *ring) { struct i915_request *request, *next; - u32 seqno = intel_engine_get_seqno(engine); - LIST_HEAD(retire); - spin_lock_irq(&engine->timeline->lock); list_for_each_entry_safe(request, next, - &engine->timeline->requests, link) { - if (!i915_seqno_passed(seqno, request->global_seqno)) + &ring->request_list, ring_link) { + if (!i915_request_completed(request)) break; - list_move_tail(&request->link, &retire); - } - spin_unlock_irq(&engine->timeline->lock); - - list_for_each_entry_safe(request, next, &retire, link) i915_request_retire(request); + } } void i915_retire_requests(struct drm_i915_private *i915) { - struct intel_engine_cs *engine; - enum intel_engine_id id; + struct intel_ring *ring, *tmp; lockdep_assert_held(&i915->drm.struct_mutex); if (!i915->gt.active_requests) return; - for_each_engine(engine, i915, id) - engine_retire_requests(engine); + list_for_each_entry_safe(ring, tmp, &i915->gt.active_rings, active_link) + ring_retire_requests(ring); } #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) diff --git a/drivers/gpu/drm/i915/i915_request.h b/drivers/gpu/drm/i915/i915_request.h index 7d6eb82..eddbd424 100644 --- a/drivers/gpu/drm/i915/i915_request.h +++ b/drivers/gpu/drm/i915/i915_request.h @@ -28,13 +28,16 @@ #include <linux/dma-fence.h> #include "i915_gem.h" +#include "i915_scheduler.h" #include "i915_sw_fence.h" +#include "i915_scheduler.h" #include <uapi/drm/i915_drm.h> struct drm_file; struct drm_i915_gem_object; struct i915_request; +struct i915_timeline; struct intel_wait { struct rb_node node; @@ -48,44 +51,6 @@ struct intel_signal_node { struct list_head link; }; -struct i915_dependency { - struct i915_priotree *signaler; - struct list_head signal_link; - struct list_head wait_link; - struct list_head dfs_link; - unsigned long flags; -#define I915_DEPENDENCY_ALLOC BIT(0) -}; - -/* - * "People assume that time is a strict progression of cause to effect, but - * actually, from a nonlinear, non-subjective viewpoint, it's more like a big - * ball of wibbly-wobbly, timey-wimey ... stuff." -The Doctor, 2015 - * - * Requests exist in a complex web of interdependencies. Each request - * has to wait for some other request to complete before it is ready to be run - * (e.g. we have to wait until the pixels have been rendering into a texture - * before we can copy from it). We track the readiness of a request in terms - * of fences, but we also need to keep the dependency tree for the lifetime - * of the request (beyond the life of an individual fence). We use the tree - * at various points to reorder the requests whilst keeping the requests - * in order with respect to their various dependencies. - */ -struct i915_priotree { - struct list_head signalers_list; /* those before us, we depend upon */ - struct list_head waiters_list; /* those after us, they depend upon us */ - struct list_head link; - int priority; -}; - -enum { - I915_PRIORITY_MIN = I915_CONTEXT_MIN_USER_PRIORITY - 1, - I915_PRIORITY_NORMAL = I915_CONTEXT_DEFAULT_PRIORITY, - I915_PRIORITY_MAX = I915_CONTEXT_MAX_USER_PRIORITY + 1, - - I915_PRIORITY_INVALID = INT_MIN -}; - struct i915_capture_list { struct i915_capture_list *next; struct i915_vma *vma; @@ -131,7 +96,7 @@ struct i915_request { struct i915_gem_context *ctx; struct intel_engine_cs *engine; struct intel_ring *ring; - struct intel_timeline *timeline; + struct i915_timeline *timeline; struct intel_signal_node signaling; /* @@ -154,7 +119,7 @@ struct i915_request { * to retirement), i.e. bidirectional dependency information for the * request not tied to individual fences. */ - struct i915_priotree priotree; + struct i915_sched_node sched; struct i915_dependency dep; /** @@ -343,10 +308,10 @@ static inline bool i915_request_started(const struct i915_request *rq) seqno - 1); } -static inline bool i915_priotree_signaled(const struct i915_priotree *pt) +static inline bool i915_sched_node_signaled(const struct i915_sched_node *node) { const struct i915_request *rq = - container_of(pt, const struct i915_request, priotree); + container_of(node, const struct i915_request, sched); return i915_request_completed(rq); } diff --git a/drivers/gpu/drm/i915/i915_scheduler.h b/drivers/gpu/drm/i915/i915_scheduler.h new file mode 100644 index 0000000..70a4222 --- /dev/null +++ b/drivers/gpu/drm/i915/i915_scheduler.h @@ -0,0 +1,72 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2018 Intel Corporation + */ + +#ifndef _I915_SCHEDULER_H_ +#define _I915_SCHEDULER_H_ + +#include <linux/bitops.h> + +#include <uapi/drm/i915_drm.h> + +enum { + I915_PRIORITY_MIN = I915_CONTEXT_MIN_USER_PRIORITY - 1, + I915_PRIORITY_NORMAL = I915_CONTEXT_DEFAULT_PRIORITY, + I915_PRIORITY_MAX = I915_CONTEXT_MAX_USER_PRIORITY + 1, + + I915_PRIORITY_INVALID = INT_MIN +}; + +struct i915_sched_attr { + /** + * @priority: execution and service priority + * + * All clients are equal, but some are more equal than others! + * + * Requests from a context with a greater (more positive) value of + * @priority will be executed before those with a lower @priority + * value, forming a simple QoS. + * + * The &drm_i915_private.kernel_context is assigned the lowest priority. + */ + int priority; +}; + +/* + * "People assume that time is a strict progression of cause to effect, but + * actually, from a nonlinear, non-subjective viewpoint, it's more like a big + * ball of wibbly-wobbly, timey-wimey ... stuff." -The Doctor, 2015 + * + * Requests exist in a complex web of interdependencies. Each request + * has to wait for some other request to complete before it is ready to be run + * (e.g. we have to wait until the pixels have been rendering into a texture + * before we can copy from it). We track the readiness of a request in terms + * of fences, but we also need to keep the dependency tree for the lifetime + * of the request (beyond the life of an individual fence). We use the tree + * at various points to reorder the requests whilst keeping the requests + * in order with respect to their various dependencies. + * + * There is no active component to the "scheduler". As we know the dependency + * DAG of each request, we are able to insert it into a sorted queue when it + * is ready, and are able to reorder its portion of the graph to accommodate + * dynamic priority changes. + */ +struct i915_sched_node { + struct list_head signalers_list; /* those before us, we depend upon */ + struct list_head waiters_list; /* those after us, they depend upon us */ + struct list_head link; + struct i915_sched_attr attr; +}; + +struct i915_dependency { + struct i915_sched_node *signaler; + struct list_head signal_link; + struct list_head wait_link; + struct list_head dfs_link; + unsigned long flags; +#define I915_DEPENDENCY_ALLOC BIT(0) +}; + +#endif /* _I915_SCHEDULER_H_ */ diff --git a/drivers/gpu/drm/i915/i915_timeline.c b/drivers/gpu/drm/i915/i915_timeline.c new file mode 100644 index 0000000..4667cc0 --- /dev/null +++ b/drivers/gpu/drm/i915/i915_timeline.c @@ -0,0 +1,105 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2016-2018 Intel Corporation + */ + +#include "i915_drv.h" + +#include "i915_timeline.h" +#include "i915_syncmap.h" + +void i915_timeline_init(struct drm_i915_private *i915, + struct i915_timeline *timeline, + const char *name) +{ + lockdep_assert_held(&i915->drm.struct_mutex); + + /* + * Ideally we want a set of engines on a single leaf as we expect + * to mostly be tracking synchronisation between engines. It is not + * a huge issue if this is not the case, but we may want to mitigate + * any page crossing penalties if they become an issue. + */ + BUILD_BUG_ON(KSYNCMAP < I915_NUM_ENGINES); + + timeline->name = name; + + list_add(&timeline->link, &i915->gt.timelines); + + /* Called during early_init before we know how many engines there are */ + + timeline->fence_context = dma_fence_context_alloc(1); + + spin_lock_init(&timeline->lock); + + init_request_active(&timeline->last_request, NULL); + INIT_LIST_HEAD(&timeline->requests); + + i915_syncmap_init(&timeline->sync); +} + +/** + * i915_timelines_park - called when the driver idles + * @i915: the drm_i915_private device + * + * When the driver is completely idle, we know that all of our sync points + * have been signaled and our tracking is then entirely redundant. Any request + * to wait upon an older sync point will be completed instantly as we know + * the fence is signaled and therefore we will not even look them up in the + * sync point map. + */ +void i915_timelines_park(struct drm_i915_private *i915) +{ + struct i915_timeline *timeline; + + lockdep_assert_held(&i915->drm.struct_mutex); + + list_for_each_entry(timeline, &i915->gt.timelines, link) { + /* + * All known fences are completed so we can scrap + * the current sync point tracking and start afresh, + * any attempt to wait upon a previous sync point + * will be skipped as the fence was signaled. + */ + i915_syncmap_free(&timeline->sync); + } +} + +void i915_timeline_fini(struct i915_timeline *timeline) +{ + GEM_BUG_ON(!list_empty(&timeline->requests)); + + i915_syncmap_free(&timeline->sync); + + list_del(&timeline->link); +} + +struct i915_timeline * +i915_timeline_create(struct drm_i915_private *i915, const char *name) +{ + struct i915_timeline *timeline; + + timeline = kzalloc(sizeof(*timeline), GFP_KERNEL); + if (!timeline) + return ERR_PTR(-ENOMEM); + + i915_timeline_init(i915, timeline, name); + kref_init(&timeline->kref); + + return timeline; +} + +void __i915_timeline_free(struct kref *kref) +{ + struct i915_timeline *timeline = + container_of(kref, typeof(*timeline), kref); + + i915_timeline_fini(timeline); + kfree(timeline); +} + +#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) +#include "selftests/mock_timeline.c" +#include "selftests/i915_timeline.c" +#endif diff --git a/drivers/gpu/drm/i915/i915_gem_timeline.h b/drivers/gpu/drm/i915/i915_timeline.h index 33e01bf..dc2a463 100644 --- a/drivers/gpu/drm/i915/i915_gem_timeline.h +++ b/drivers/gpu/drm/i915/i915_timeline.h @@ -22,27 +22,20 @@ * */ -#ifndef I915_GEM_TIMELINE_H -#define I915_GEM_TIMELINE_H +#ifndef I915_TIMELINE_H +#define I915_TIMELINE_H #include <linux/list.h> +#include <linux/kref.h> #include "i915_request.h" #include "i915_syncmap.h" #include "i915_utils.h" -struct i915_gem_timeline; - -struct intel_timeline { +struct i915_timeline { u64 fence_context; u32 seqno; - /** - * Count of outstanding requests, from the time they are constructed - * to the moment they are retired. Loosely coupled to hardware. - */ - u32 inflight_seqnos; - spinlock_t lock; /** @@ -77,47 +70,57 @@ struct intel_timeline { */ u32 global_sync[I915_NUM_ENGINES]; - struct i915_gem_timeline *common; -}; - -struct i915_gem_timeline { struct list_head link; - - struct drm_i915_private *i915; const char *name; - struct intel_timeline engine[I915_NUM_ENGINES]; + struct kref kref; }; -int i915_gem_timeline_init(struct drm_i915_private *i915, - struct i915_gem_timeline *tl, - const char *name); -int i915_gem_timeline_init__global(struct drm_i915_private *i915); -void i915_gem_timelines_park(struct drm_i915_private *i915); -void i915_gem_timeline_fini(struct i915_gem_timeline *tl); +void i915_timeline_init(struct drm_i915_private *i915, + struct i915_timeline *tl, + const char *name); +void i915_timeline_fini(struct i915_timeline *tl); + +struct i915_timeline * +i915_timeline_create(struct drm_i915_private *i915, const char *name); -static inline int __intel_timeline_sync_set(struct intel_timeline *tl, - u64 context, u32 seqno) +static inline struct i915_timeline * +i915_timeline_get(struct i915_timeline *timeline) +{ + kref_get(&timeline->kref); + return timeline; +} + +void __i915_timeline_free(struct kref *kref); +static inline void i915_timeline_put(struct i915_timeline *timeline) +{ + kref_put(&timeline->kref, __i915_timeline_free); +} + +static inline int __i915_timeline_sync_set(struct i915_timeline *tl, + u64 context, u32 seqno) { return i915_syncmap_set(&tl->sync, context, seqno); } -static inline int intel_timeline_sync_set(struct intel_timeline *tl, - const struct dma_fence *fence) +static inline int i915_timeline_sync_set(struct i915_timeline *tl, + const struct dma_fence *fence) { - return __intel_timeline_sync_set(tl, fence->context, fence->seqno); + return __i915_timeline_sync_set(tl, fence->context, fence->seqno); } -static inline bool __intel_timeline_sync_is_later(struct intel_timeline *tl, - u64 context, u32 seqno) +static inline bool __i915_timeline_sync_is_later(struct i915_timeline *tl, + u64 context, u32 seqno) { return i915_syncmap_is_later(&tl->sync, context, seqno); } -static inline bool intel_timeline_sync_is_later(struct intel_timeline *tl, - const struct dma_fence *fence) +static inline bool i915_timeline_sync_is_later(struct i915_timeline *tl, + const struct dma_fence *fence) { - return __intel_timeline_sync_is_later(tl, fence->context, fence->seqno); + return __i915_timeline_sync_is_later(tl, fence->context, fence->seqno); } +void i915_timelines_park(struct drm_i915_private *i915); + #endif diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h index 408827b..8cc3a25 100644 --- a/drivers/gpu/drm/i915/i915_trace.h +++ b/drivers/gpu/drm/i915/i915_trace.h @@ -679,45 +679,68 @@ DEFINE_EVENT(i915_request, i915_request_execute, TP_ARGS(rq) ); -DECLARE_EVENT_CLASS(i915_request_hw, - TP_PROTO(struct i915_request *rq, unsigned int port), - TP_ARGS(rq, port), - - TP_STRUCT__entry( - __field(u32, dev) - __field(u32, hw_id) - __field(u32, ring) - __field(u32, ctx) - __field(u32, seqno) - __field(u32, global_seqno) - __field(u32, port) - ), - - TP_fast_assign( - __entry->dev = rq->i915->drm.primary->index; - __entry->hw_id = rq->ctx->hw_id; - __entry->ring = rq->engine->id; - __entry->ctx = rq->fence.context; - __entry->seqno = rq->fence.seqno; - __entry->global_seqno = rq->global_seqno; - __entry->port = port; - ), - - TP_printk("dev=%u, hw_id=%u, ring=%u, ctx=%u, seqno=%u, global=%u, port=%u", - __entry->dev, __entry->hw_id, __entry->ring, - __entry->ctx, __entry->seqno, - __entry->global_seqno, __entry->port) -); +TRACE_EVENT(i915_request_in, + TP_PROTO(struct i915_request *rq, unsigned int port), + TP_ARGS(rq, port), + + TP_STRUCT__entry( + __field(u32, dev) + __field(u32, hw_id) + __field(u32, ring) + __field(u32, ctx) + __field(u32, seqno) + __field(u32, global_seqno) + __field(u32, port) + __field(u32, prio) + ), + + TP_fast_assign( + __entry->dev = rq->i915->drm.primary->index; + __entry->hw_id = rq->ctx->hw_id; + __entry->ring = rq->engine->id; + __entry->ctx = rq->fence.context; + __entry->seqno = rq->fence.seqno; + __entry->global_seqno = rq->global_seqno; + __entry->prio = rq->sched.attr.priority; + __entry->port = port; + ), -DEFINE_EVENT(i915_request_hw, i915_request_in, - TP_PROTO(struct i915_request *rq, unsigned int port), - TP_ARGS(rq, port) + TP_printk("dev=%u, hw_id=%u, ring=%u, ctx=%u, seqno=%u, prio=%u, global=%u, port=%u", + __entry->dev, __entry->hw_id, __entry->ring, __entry->ctx, + __entry->seqno, __entry->prio, __entry->global_seqno, + __entry->port) ); -DEFINE_EVENT(i915_request, i915_request_out, - TP_PROTO(struct i915_request *rq), - TP_ARGS(rq) +TRACE_EVENT(i915_request_out, + TP_PROTO(struct i915_request *rq), + TP_ARGS(rq), + + TP_STRUCT__entry( + __field(u32, dev) + __field(u32, hw_id) + __field(u32, ring) + __field(u32, ctx) + __field(u32, seqno) + __field(u32, global_seqno) + __field(u32, completed) + ), + + TP_fast_assign( + __entry->dev = rq->i915->drm.primary->index; + __entry->hw_id = rq->ctx->hw_id; + __entry->ring = rq->engine->id; + __entry->ctx = rq->fence.context; + __entry->seqno = rq->fence.seqno; + __entry->global_seqno = rq->global_seqno; + __entry->completed = i915_request_completed(rq); + ), + + TP_printk("dev=%u, hw_id=%u, ring=%u, ctx=%u, seqno=%u, global=%u, completed?=%u", + __entry->dev, __entry->hw_id, __entry->ring, + __entry->ctx, __entry->seqno, + __entry->global_seqno, __entry->completed) ); + #else #if !defined(TRACE_HEADER_MULTI_READ) static inline void @@ -811,42 +834,6 @@ DEFINE_EVENT(i915_request, i915_request_wait_end, TP_ARGS(rq) ); -TRACE_EVENT(i915_flip_request, - TP_PROTO(int plane, struct drm_i915_gem_object *obj), - - TP_ARGS(plane, obj), - - TP_STRUCT__entry( - __field(int, plane) - __field(struct drm_i915_gem_object *, obj) - ), - - TP_fast_assign( - __entry->plane = plane; - __entry->obj = obj; - ), - - TP_printk("plane=%d, obj=%p", __entry->plane, __entry->obj) -); - -TRACE_EVENT(i915_flip_complete, - TP_PROTO(int plane, struct drm_i915_gem_object *obj), - - TP_ARGS(plane, obj), - - TP_STRUCT__entry( - __field(int, plane) - __field(struct drm_i915_gem_object *, obj) - ), - - TP_fast_assign( - __entry->plane = plane; - __entry->obj = obj; - ), - - TP_printk("plane=%d, obj=%p", __entry->plane, __entry->obj) -); - TRACE_EVENT_CONDITION(i915_reg_rw, TP_PROTO(bool write, i915_reg_t reg, u64 val, int len, bool trace), diff --git a/drivers/gpu/drm/i915/i915_utils.h b/drivers/gpu/drm/i915/i915_utils.h index 51dbfe5..00165ad 100644 --- a/drivers/gpu/drm/i915/i915_utils.h +++ b/drivers/gpu/drm/i915/i915_utils.h @@ -40,8 +40,8 @@ #undef WARN_ON_ONCE #define WARN_ON_ONCE(x) WARN_ONCE((x), "%s", "WARN_ON_ONCE(" __stringify(x) ")") -#define MISSING_CASE(x) WARN(1, "Missing switch case (%lu) in %s\n", \ - (long)(x), __func__) +#define MISSING_CASE(x) WARN(1, "Missing case (%s == %ld)\n", \ + __stringify(x), (long)(x)) #if GCC_VERSION >= 70000 #define add_overflows(A, B) \ @@ -120,6 +120,12 @@ static inline u64 ptr_to_u64(const void *ptr) #include <linux/list.h> +static inline int list_is_first(const struct list_head *list, + const struct list_head *head) +{ + return head->next == list; +} + static inline void __list_del_many(struct list_head *head, struct list_head *first) { diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index 4bda3bd..9324d47 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -46,8 +46,6 @@ i915_vma_retire(struct i915_gem_active *active, struct i915_request *rq) 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)); GEM_BUG_ON(!i915_gem_object_is_active(obj)); if (--obj->active_count) @@ -232,7 +230,6 @@ i915_vma_instance(struct drm_i915_gem_object *obj, if (!vma) vma = 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; @@ -684,13 +681,43 @@ err_unpin: return ret; } -static void i915_vma_destroy(struct i915_vma *vma) +void i915_vma_close(struct i915_vma *vma) +{ + lockdep_assert_held(&vma->vm->i915->drm.struct_mutex); + + GEM_BUG_ON(i915_vma_is_closed(vma)); + vma->flags |= I915_VMA_CLOSED; + + /* + * We defer actually closing, unbinding and destroying the VMA until + * the next idle point, or if the object is freed in the meantime. By + * postponing the unbind, we allow for it to be resurrected by the + * client, avoiding the work required to rebind the VMA. This is + * advantageous for DRI, where the client/server pass objects + * between themselves, temporarily opening a local VMA to the + * object, and then closing it again. The same object is then reused + * on the next frame (or two, depending on the depth of the swap queue) + * causing us to rebind the VMA once more. This ends up being a lot + * of wasted work for the steady state. + */ + list_add_tail(&vma->closed_link, &vma->vm->i915->gt.closed_vma); +} + +void i915_vma_reopen(struct i915_vma *vma) +{ + lockdep_assert_held(&vma->vm->i915->drm.struct_mutex); + + if (vma->flags & I915_VMA_CLOSED) { + vma->flags &= ~I915_VMA_CLOSED; + list_del(&vma->closed_link); + } +} + +static void __i915_vma_destroy(struct i915_vma *vma) { int i; GEM_BUG_ON(vma->node.allocated); - GEM_BUG_ON(i915_vma_is_active(vma)); - GEM_BUG_ON(!i915_vma_is_closed(vma)); GEM_BUG_ON(vma->fence); for (i = 0; i < ARRAY_SIZE(vma->last_read); i++) @@ -699,6 +726,7 @@ static void i915_vma_destroy(struct i915_vma *vma) list_del(&vma->obj_link); list_del(&vma->vm_link); + rb_erase(&vma->obj_node, &vma->obj->vma_tree); if (!i915_vma_is_ggtt(vma)) i915_ppgtt_put(i915_vm_to_ppgtt(vma->vm)); @@ -706,15 +734,30 @@ static void i915_vma_destroy(struct i915_vma *vma) kmem_cache_free(to_i915(vma->obj->base.dev)->vmas, vma); } -void i915_vma_close(struct i915_vma *vma) +void i915_vma_destroy(struct i915_vma *vma) { - GEM_BUG_ON(i915_vma_is_closed(vma)); - vma->flags |= I915_VMA_CLOSED; + lockdep_assert_held(&vma->vm->i915->drm.struct_mutex); - rb_erase(&vma->obj_node, &vma->obj->vma_tree); + GEM_BUG_ON(i915_vma_is_active(vma)); + GEM_BUG_ON(i915_vma_is_pinned(vma)); + + if (i915_vma_is_closed(vma)) + list_del(&vma->closed_link); + + WARN_ON(i915_vma_unbind(vma)); + __i915_vma_destroy(vma); +} + +void i915_vma_parked(struct drm_i915_private *i915) +{ + struct i915_vma *vma, *next; - if (!i915_vma_is_active(vma) && !i915_vma_is_pinned(vma)) - WARN_ON(i915_vma_unbind(vma)); + list_for_each_entry_safe(vma, next, &i915->gt.closed_vma, closed_link) { + GEM_BUG_ON(!i915_vma_is_closed(vma)); + i915_vma_destroy(vma); + } + + GEM_BUG_ON(!list_empty(&i915->gt.closed_vma)); } static void __i915_vma_iounmap(struct i915_vma *vma) @@ -804,7 +847,7 @@ int i915_vma_unbind(struct i915_vma *vma) return -EBUSY; if (!drm_mm_node_allocated(&vma->node)) - goto destroy; + return 0; GEM_BUG_ON(obj->bind_count == 0); GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj)); @@ -841,10 +884,6 @@ int i915_vma_unbind(struct i915_vma *vma) i915_vma_remove(vma); -destroy: - if (unlikely(i915_vma_is_closed(vma))) - i915_vma_destroy(vma); - return 0; } diff --git a/drivers/gpu/drm/i915/i915_vma.h b/drivers/gpu/drm/i915/i915_vma.h index 8c50220..fc4294c 100644 --- a/drivers/gpu/drm/i915/i915_vma.h +++ b/drivers/gpu/drm/i915/i915_vma.h @@ -119,6 +119,8 @@ struct i915_vma { /** This vma's place in the eviction list */ struct list_head evict_link; + struct list_head closed_link; + /** * Used for performing relocations during execbuffer insertion. */ @@ -285,6 +287,8 @@ void i915_vma_revoke_mmap(struct i915_vma *vma); int __must_check i915_vma_unbind(struct i915_vma *vma); void i915_vma_unlink_ctx(struct i915_vma *vma); void i915_vma_close(struct i915_vma *vma); +void i915_vma_reopen(struct i915_vma *vma); +void i915_vma_destroy(struct i915_vma *vma); int __i915_vma_do_pin(struct i915_vma *vma, u64 size, u64 alignment, u64 flags); @@ -408,6 +412,8 @@ i915_vma_unpin_fence(struct i915_vma *vma) __i915_vma_unpin_fence(vma); } +void i915_vma_parked(struct drm_i915_private *i915); + #define for_each_until(cond) if (cond) break; else /** diff --git a/drivers/gpu/drm/i915/intel_atomic.c b/drivers/gpu/drm/i915/intel_atomic.c index e9fb6920..40285d1 100644 --- a/drivers/gpu/drm/i915/intel_atomic.c +++ b/drivers/gpu/drm/i915/intel_atomic.c @@ -227,6 +227,7 @@ int intel_atomic_setup_scalers(struct drm_i915_private *dev_priv, struct intel_crtc_scaler_state *scaler_state = &crtc_state->scaler_state; struct drm_atomic_state *drm_state = crtc_state->base.state; + struct intel_atomic_state *intel_state = to_intel_atomic_state(drm_state); int num_scalers_need; int i, j; @@ -304,8 +305,8 @@ int intel_atomic_setup_scalers(struct drm_i915_private *dev_priv, continue; } - plane_state = intel_atomic_get_existing_plane_state(drm_state, - intel_plane); + plane_state = intel_atomic_get_new_plane_state(intel_state, + intel_plane); scaler_id = &plane_state->scaler_id; } @@ -328,8 +329,18 @@ int intel_atomic_setup_scalers(struct drm_i915_private *dev_priv, } /* set scaler mode */ - if (IS_GEMINILAKE(dev_priv) || IS_CANNONLAKE(dev_priv)) { - scaler_state->scalers[*scaler_id].mode = 0; + if ((INTEL_GEN(dev_priv) >= 9) && + plane_state && plane_state->base.fb && + plane_state->base.fb->format->format == + DRM_FORMAT_NV12) { + if (INTEL_GEN(dev_priv) == 9 && + !IS_GEMINILAKE(dev_priv) && + !IS_SKYLAKE(dev_priv)) + scaler_state->scalers[*scaler_id].mode = + SKL_PS_SCALER_MODE_NV12; + else + scaler_state->scalers[*scaler_id].mode = + PS_SCALER_MODE_PLANAR; } else if (num_scalers_need == 1 && intel_crtc->pipe != PIPE_C) { /* * when only 1 scaler is in use on either pipe A or B, diff --git a/drivers/gpu/drm/i915/intel_atomic_plane.c b/drivers/gpu/drm/i915/intel_atomic_plane.c index 7481ce8..6d06878 100644 --- a/drivers/gpu/drm/i915/intel_atomic_plane.c +++ b/drivers/gpu/drm/i915/intel_atomic_plane.c @@ -183,11 +183,16 @@ int intel_plane_atomic_check_with_state(const struct intel_crtc_state *old_crtc_ } /* FIXME pre-g4x don't work like this */ - if (intel_state->base.visible) + if (state->visible) crtc_state->active_planes |= BIT(intel_plane->id); else crtc_state->active_planes &= ~BIT(intel_plane->id); + if (state->visible && state->fb->format->format == DRM_FORMAT_NV12) + crtc_state->nv12_planes |= BIT(intel_plane->id); + else + crtc_state->nv12_planes &= ~BIT(intel_plane->id); + return intel_plane_atomic_calc_changes(old_crtc_state, &crtc_state->base, old_plane_state, diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c index 447b721..54270bd 100644 --- a/drivers/gpu/drm/i915/intel_bios.c +++ b/drivers/gpu/drm/i915/intel_bios.c @@ -530,6 +530,7 @@ parse_driver_features(struct drm_i915_private *dev_priv, */ if (!driver->drrs_enabled) dev_priv->vbt.drrs_type = DRRS_NOT_SUPPORTED; + dev_priv->vbt.psr.enable = driver->psr_enabled; } static void @@ -1215,10 +1216,8 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port, { struct child_device_config *it, *child = NULL; struct ddi_vbt_port_info *info = &dev_priv->vbt.ddi_port_info[port]; - uint8_t hdmi_level_shift; int i, j; bool is_dvi, is_hdmi, is_dp, is_edp, is_crt; - uint8_t aux_channel, ddc_pin; /* Each DDI port can have more than one value on the "DVO Port" field, * so look for all the possible values for each port. */ @@ -1255,8 +1254,6 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port, if (!child) return; - aux_channel = child->aux_channel; - is_dvi = child->device_type & DEVICE_TYPE_TMDS_DVI_SIGNALING; is_dp = child->device_type & DEVICE_TYPE_DISPLAYPORT_OUTPUT; is_crt = child->device_type & DEVICE_TYPE_ANALOG_OUTPUT; @@ -1270,13 +1267,6 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port, is_hdmi = false; } - if (port == PORT_A && is_dvi) { - DRM_DEBUG_KMS("VBT claims port A supports DVI%s, ignoring\n", - is_hdmi ? "/HDMI" : ""); - is_dvi = false; - is_hdmi = false; - } - info->supports_dvi = is_dvi; info->supports_hdmi = is_hdmi; info->supports_dp = is_dp; @@ -1302,6 +1292,8 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port, DRM_DEBUG_KMS("Port %c is internal DP\n", port_name(port)); if (is_dvi) { + u8 ddc_pin; + ddc_pin = map_ddc_pin(dev_priv, child->ddc_pin); if (intel_gmbus_is_valid_pin(dev_priv, ddc_pin)) { info->alternate_ddc_pin = ddc_pin; @@ -1314,14 +1306,14 @@ static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port, } if (is_dp) { - info->alternate_aux_channel = aux_channel; + info->alternate_aux_channel = child->aux_channel; sanitize_aux_ch(dev_priv, port); } if (bdb_version >= 158) { /* The VBT HDMI level shift values match the table we have. */ - hdmi_level_shift = child->hdmi_level_shifter_value; + u8 hdmi_level_shift = child->hdmi_level_shifter_value; DRM_DEBUG_KMS("VBT HDMI level shift for port %c: %d\n", port_name(port), hdmi_level_shift); diff --git a/drivers/gpu/drm/i915/intel_breadcrumbs.c b/drivers/gpu/drm/i915/intel_breadcrumbs.c index 1f79e7a..18e643d 100644 --- a/drivers/gpu/drm/i915/intel_breadcrumbs.c +++ b/drivers/gpu/drm/i915/intel_breadcrumbs.c @@ -82,7 +82,7 @@ static unsigned long wait_timeout(void) static noinline void missed_breadcrumb(struct intel_engine_cs *engine) { - if (drm_debug & DRM_UT_DRIVER) { + if (GEM_SHOW_DEBUG()) { struct drm_printer p = drm_debug_printer(__func__); intel_engine_dump(engine, &p, @@ -130,11 +130,12 @@ static void intel_breadcrumbs_hangcheck(struct timer_list *t) static void intel_breadcrumbs_fake_irq(struct timer_list *t) { - struct intel_engine_cs *engine = from_timer(engine, t, - breadcrumbs.fake_irq); + struct intel_engine_cs *engine = + from_timer(engine, t, breadcrumbs.fake_irq); struct intel_breadcrumbs *b = &engine->breadcrumbs; - /* The timer persists in case we cannot enable interrupts, + /* + * The timer persists in case we cannot enable interrupts, * or if we have previously seen seqno/interrupt incoherency * ("missed interrupt" syndrome, better known as a "missed breadcrumb"). * Here the worker will wake up every jiffie in order to kick the @@ -148,6 +149,12 @@ static void intel_breadcrumbs_fake_irq(struct timer_list *t) if (!b->irq_armed) return; + /* If the user has disabled the fake-irq, restore the hangchecking */ + if (!test_bit(engine->id, &engine->i915->gpu_error.missed_irq_rings)) { + mod_timer(&b->hangcheck, wait_timeout()); + return; + } + mod_timer(&b->fake_irq, jiffies + 1); } @@ -730,10 +737,11 @@ static void insert_signal(struct intel_breadcrumbs *b, list_add(&request->signaling.link, &iter->signaling.link); } -void intel_engine_enable_signaling(struct i915_request *request, bool wakeup) +bool intel_engine_enable_signaling(struct i915_request *request, bool wakeup) { struct intel_engine_cs *engine = request->engine; struct intel_breadcrumbs *b = &engine->breadcrumbs; + struct intel_wait *wait = &request->signaling.wait; u32 seqno; /* @@ -750,12 +758,12 @@ void intel_engine_enable_signaling(struct i915_request *request, bool wakeup) seqno = i915_request_global_seqno(request); if (!seqno) /* will be enabled later upon execution */ - return; + return true; - GEM_BUG_ON(request->signaling.wait.seqno); - request->signaling.wait.tsk = b->signaler; - request->signaling.wait.request = request; - request->signaling.wait.seqno = seqno; + GEM_BUG_ON(wait->seqno); + wait->tsk = b->signaler; + wait->request = request; + wait->seqno = seqno; /* * Add ourselves into the list of waiters, but registering our @@ -768,11 +776,15 @@ void intel_engine_enable_signaling(struct i915_request *request, bool wakeup) */ spin_lock(&b->rb_lock); insert_signal(b, request, seqno); - wakeup &= __intel_engine_add_wait(engine, &request->signaling.wait); + wakeup &= __intel_engine_add_wait(engine, wait); spin_unlock(&b->rb_lock); - if (wakeup) + if (wakeup) { wake_up_process(b->signaler); + return !intel_wait_complete(wait); + } + + return true; } void intel_engine_cancel_signaling(struct i915_request *request) @@ -826,8 +838,8 @@ static void cancel_fake_irq(struct intel_engine_cs *engine) { struct intel_breadcrumbs *b = &engine->breadcrumbs; + del_timer_sync(&b->fake_irq); /* may queue b->hangcheck */ del_timer_sync(&b->hangcheck); - del_timer_sync(&b->fake_irq); clear_bit(engine->id, &engine->i915->gpu_error.missed_irq_rings); } @@ -835,15 +847,22 @@ void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine) { struct intel_breadcrumbs *b = &engine->breadcrumbs; - cancel_fake_irq(engine); spin_lock_irq(&b->irq_lock); + /* + * Leave the fake_irq timer enabled (if it is running), but clear the + * bit so that it turns itself off on its next wake up and goes back + * to the long hangcheck interval if still required. + */ + clear_bit(engine->id, &engine->i915->gpu_error.missed_irq_rings); + if (b->irq_enabled) irq_enable(engine); else irq_disable(engine); - /* We set the IRQ_BREADCRUMB bit when we enable the irq presuming the + /* + * We set the IRQ_BREADCRUMB bit when we enable the irq presuming the * GPU is active and may have already executed the MI_USER_INTERRUPT * before the CPU is ready to receive. However, the engine is currently * idle (we haven't started it yet), there is no possibility for a @@ -852,9 +871,6 @@ void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine) */ clear_bit(ENGINE_IRQ_BREADCRUMB, &engine->irq_posted); - if (b->irq_armed) - enable_fake_irq(b); - spin_unlock_irq(&b->irq_lock); } diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c index c0a8805..de0e223 100644 --- a/drivers/gpu/drm/i915/intel_crt.c +++ b/drivers/gpu/drm/i915/intel_crt.c @@ -748,6 +748,11 @@ intel_crt_detect(struct drm_connector *connector, connector->base.id, connector->name, force); + if (i915_modparams.load_detect_test) { + intel_display_power_get(dev_priv, intel_encoder->power_domain); + goto load_detect; + } + /* Skip machines without VGA that falsely report hotplug events */ if (dmi_check_system(intel_spurious_crt_detect)) return connector_status_disconnected; @@ -776,11 +781,12 @@ intel_crt_detect(struct drm_connector *connector, * broken monitor (without edid) to work behind a broken kvm (that fails * to have the right resistors for HP detection) needs to fix this up. * For now just bail out. */ - if (I915_HAS_HOTPLUG(dev_priv) && !i915_modparams.load_detect_test) { + if (I915_HAS_HOTPLUG(dev_priv)) { status = connector_status_disconnected; goto out; } +load_detect: if (!force) { status = connector->status; goto out; diff --git a/drivers/gpu/drm/i915/intel_csr.c b/drivers/gpu/drm/i915/intel_csr.c index f9550ea..cf9b600 100644 --- a/drivers/gpu/drm/i915/intel_csr.c +++ b/drivers/gpu/drm/i915/intel_csr.c @@ -298,7 +298,10 @@ static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv, csr->version = css_header->version; - if (IS_CANNONLAKE(dev_priv)) { + if (csr->fw_path == i915_modparams.dmc_firmware_path) { + /* Bypass version check for firmware override. */ + required_version = csr->version; + } else if (IS_CANNONLAKE(dev_priv)) { required_version = CNL_CSR_VERSION_REQUIRED; } else if (IS_GEMINILAKE(dev_priv)) { required_version = GLK_CSR_VERSION_REQUIRED; @@ -453,7 +456,9 @@ void intel_csr_ucode_init(struct drm_i915_private *dev_priv) if (!HAS_CSR(dev_priv)) return; - if (IS_CANNONLAKE(dev_priv)) + if (i915_modparams.dmc_firmware_path) + csr->fw_path = i915_modparams.dmc_firmware_path; + else if (IS_CANNONLAKE(dev_priv)) csr->fw_path = I915_CSR_CNL; else if (IS_GEMINILAKE(dev_priv)) csr->fw_path = I915_CSR_GLK; diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index 8c2d778..b98ac054 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -493,6 +493,125 @@ static const struct cnl_ddi_buf_trans cnl_ddi_translations_edp_1_05V[] = { { 0x2, 0x7F, 0x3F, 0x00, 0x00 }, /* 400 400 0.0 */ }; +struct icl_combo_phy_ddi_buf_trans { + u32 dw2_swing_select; + u32 dw2_swing_scalar; + u32 dw4_scaling; +}; + +/* Voltage Swing Programming for VccIO 0.85V for DP */ +static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_dp_hdmi_0_85V[] = { + /* Voltage mV db */ + { 0x2, 0x98, 0x0018 }, /* 400 0.0 */ + { 0x2, 0x98, 0x3015 }, /* 400 3.5 */ + { 0x2, 0x98, 0x6012 }, /* 400 6.0 */ + { 0x2, 0x98, 0x900F }, /* 400 9.5 */ + { 0xB, 0x70, 0x0018 }, /* 600 0.0 */ + { 0xB, 0x70, 0x3015 }, /* 600 3.5 */ + { 0xB, 0x70, 0x6012 }, /* 600 6.0 */ + { 0x5, 0x00, 0x0018 }, /* 800 0.0 */ + { 0x5, 0x00, 0x3015 }, /* 800 3.5 */ + { 0x6, 0x98, 0x0018 }, /* 1200 0.0 */ +}; + +/* FIXME - After table is updated in Bspec */ +/* Voltage Swing Programming for VccIO 0.85V for eDP */ +static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_edp_0_85V[] = { + /* Voltage mV db */ + { 0x0, 0x00, 0x00 }, /* 200 0.0 */ + { 0x0, 0x00, 0x00 }, /* 200 1.5 */ + { 0x0, 0x00, 0x00 }, /* 200 4.0 */ + { 0x0, 0x00, 0x00 }, /* 200 6.0 */ + { 0x0, 0x00, 0x00 }, /* 250 0.0 */ + { 0x0, 0x00, 0x00 }, /* 250 1.5 */ + { 0x0, 0x00, 0x00 }, /* 250 4.0 */ + { 0x0, 0x00, 0x00 }, /* 300 0.0 */ + { 0x0, 0x00, 0x00 }, /* 300 1.5 */ + { 0x0, 0x00, 0x00 }, /* 350 0.0 */ +}; + +/* Voltage Swing Programming for VccIO 0.95V for DP */ +static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_dp_hdmi_0_95V[] = { + /* Voltage mV db */ + { 0x2, 0x98, 0x0018 }, /* 400 0.0 */ + { 0x2, 0x98, 0x3015 }, /* 400 3.5 */ + { 0x2, 0x98, 0x6012 }, /* 400 6.0 */ + { 0x2, 0x98, 0x900F }, /* 400 9.5 */ + { 0x4, 0x98, 0x0018 }, /* 600 0.0 */ + { 0x4, 0x98, 0x3015 }, /* 600 3.5 */ + { 0x4, 0x98, 0x6012 }, /* 600 6.0 */ + { 0x5, 0x76, 0x0018 }, /* 800 0.0 */ + { 0x5, 0x76, 0x3015 }, /* 800 3.5 */ + { 0x6, 0x98, 0x0018 }, /* 1200 0.0 */ +}; + +/* FIXME - After table is updated in Bspec */ +/* Voltage Swing Programming for VccIO 0.95V for eDP */ +static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_edp_0_95V[] = { + /* Voltage mV db */ + { 0x0, 0x00, 0x00 }, /* 200 0.0 */ + { 0x0, 0x00, 0x00 }, /* 200 1.5 */ + { 0x0, 0x00, 0x00 }, /* 200 4.0 */ + { 0x0, 0x00, 0x00 }, /* 200 6.0 */ + { 0x0, 0x00, 0x00 }, /* 250 0.0 */ + { 0x0, 0x00, 0x00 }, /* 250 1.5 */ + { 0x0, 0x00, 0x00 }, /* 250 4.0 */ + { 0x0, 0x00, 0x00 }, /* 300 0.0 */ + { 0x0, 0x00, 0x00 }, /* 300 1.5 */ + { 0x0, 0x00, 0x00 }, /* 350 0.0 */ +}; + +/* Voltage Swing Programming for VccIO 1.05V for DP */ +static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_dp_hdmi_1_05V[] = { + /* Voltage mV db */ + { 0x2, 0x98, 0x0018 }, /* 400 0.0 */ + { 0x2, 0x98, 0x3015 }, /* 400 3.5 */ + { 0x2, 0x98, 0x6012 }, /* 400 6.0 */ + { 0x2, 0x98, 0x900F }, /* 400 9.5 */ + { 0x4, 0x98, 0x0018 }, /* 600 0.0 */ + { 0x4, 0x98, 0x3015 }, /* 600 3.5 */ + { 0x4, 0x98, 0x6012 }, /* 600 6.0 */ + { 0x5, 0x71, 0x0018 }, /* 800 0.0 */ + { 0x5, 0x71, 0x3015 }, /* 800 3.5 */ + { 0x6, 0x98, 0x0018 }, /* 1200 0.0 */ +}; + +/* FIXME - After table is updated in Bspec */ +/* Voltage Swing Programming for VccIO 1.05V for eDP */ +static const struct icl_combo_phy_ddi_buf_trans icl_combo_phy_ddi_translations_edp_1_05V[] = { + /* Voltage mV db */ + { 0x0, 0x00, 0x00 }, /* 200 0.0 */ + { 0x0, 0x00, 0x00 }, /* 200 1.5 */ + { 0x0, 0x00, 0x00 }, /* 200 4.0 */ + { 0x0, 0x00, 0x00 }, /* 200 6.0 */ + { 0x0, 0x00, 0x00 }, /* 250 0.0 */ + { 0x0, 0x00, 0x00 }, /* 250 1.5 */ + { 0x0, 0x00, 0x00 }, /* 250 4.0 */ + { 0x0, 0x00, 0x00 }, /* 300 0.0 */ + { 0x0, 0x00, 0x00 }, /* 300 1.5 */ + { 0x0, 0x00, 0x00 }, /* 350 0.0 */ +}; + +struct icl_mg_phy_ddi_buf_trans { + u32 cri_txdeemph_override_5_0; + u32 cri_txdeemph_override_11_6; + u32 cri_txdeemph_override_17_12; +}; + +static const struct icl_mg_phy_ddi_buf_trans icl_mg_phy_ddi_translations[] = { + /* Voltage swing pre-emphasis */ + { 0x0, 0x1B, 0x00 }, /* 0 0 */ + { 0x0, 0x23, 0x08 }, /* 0 1 */ + { 0x0, 0x2D, 0x12 }, /* 0 2 */ + { 0x0, 0x00, 0x00 }, /* 0 3 */ + { 0x0, 0x23, 0x00 }, /* 1 0 */ + { 0x0, 0x2B, 0x09 }, /* 1 1 */ + { 0x0, 0x2E, 0x11 }, /* 1 2 */ + { 0x0, 0x2F, 0x00 }, /* 2 0 */ + { 0x0, 0x33, 0x0C }, /* 2 1 */ + { 0x0, 0x00, 0x00 }, /* 3 0 */ +}; + static const struct ddi_buf_trans * bdw_get_buf_trans_edp(struct drm_i915_private *dev_priv, int *n_entries) { @@ -751,6 +870,45 @@ cnl_get_buf_trans_edp(struct drm_i915_private *dev_priv, int *n_entries) } } +static const struct icl_combo_phy_ddi_buf_trans * +icl_get_combo_buf_trans(struct drm_i915_private *dev_priv, enum port port, + int type, int *n_entries) +{ + u32 voltage = I915_READ(ICL_PORT_COMP_DW3(port)) & VOLTAGE_INFO_MASK; + + if (type == INTEL_OUTPUT_EDP && dev_priv->vbt.edp.low_vswing) { + switch (voltage) { + case VOLTAGE_INFO_0_85V: + *n_entries = ARRAY_SIZE(icl_combo_phy_ddi_translations_edp_0_85V); + return icl_combo_phy_ddi_translations_edp_0_85V; + case VOLTAGE_INFO_0_95V: + *n_entries = ARRAY_SIZE(icl_combo_phy_ddi_translations_edp_0_95V); + return icl_combo_phy_ddi_translations_edp_0_95V; + case VOLTAGE_INFO_1_05V: + *n_entries = ARRAY_SIZE(icl_combo_phy_ddi_translations_edp_1_05V); + return icl_combo_phy_ddi_translations_edp_1_05V; + default: + MISSING_CASE(voltage); + return NULL; + } + } else { + switch (voltage) { + case VOLTAGE_INFO_0_85V: + *n_entries = ARRAY_SIZE(icl_combo_phy_ddi_translations_dp_hdmi_0_85V); + return icl_combo_phy_ddi_translations_dp_hdmi_0_85V; + case VOLTAGE_INFO_0_95V: + *n_entries = ARRAY_SIZE(icl_combo_phy_ddi_translations_dp_hdmi_0_95V); + return icl_combo_phy_ddi_translations_dp_hdmi_0_95V; + case VOLTAGE_INFO_1_05V: + *n_entries = ARRAY_SIZE(icl_combo_phy_ddi_translations_dp_hdmi_1_05V); + return icl_combo_phy_ddi_translations_dp_hdmi_1_05V; + default: + MISSING_CASE(voltage); + return NULL; + } + } +} + static int intel_ddi_hdmi_level(struct drm_i915_private *dev_priv, enum port port) { int n_entries, level, default_entry; @@ -875,7 +1033,7 @@ static void intel_wait_ddi_buf_idle(struct drm_i915_private *dev_priv, static uint32_t hsw_pll_to_ddi_pll_sel(const struct intel_shared_dpll *pll) { - switch (pll->id) { + switch (pll->info->id) { case DPLL_ID_WRPLL1: return PORT_CLK_SEL_WRPLL1; case DPLL_ID_WRPLL2: @@ -889,11 +1047,30 @@ static uint32_t hsw_pll_to_ddi_pll_sel(const struct intel_shared_dpll *pll) case DPLL_ID_LCPLL_2700: return PORT_CLK_SEL_LCPLL_2700; default: - MISSING_CASE(pll->id); + MISSING_CASE(pll->info->id); return PORT_CLK_SEL_NONE; } } +static uint32_t icl_pll_to_ddi_pll_sel(struct intel_encoder *encoder, + const struct intel_shared_dpll *pll) +{ + const enum intel_dpll_id id = pll->info->id; + + switch (id) { + default: + MISSING_CASE(id); + case DPLL_ID_ICL_DPLL0: + case DPLL_ID_ICL_DPLL1: + return DDI_CLK_SEL_NONE; + case DPLL_ID_ICL_MGPLL1: + case DPLL_ID_ICL_MGPLL2: + case DPLL_ID_ICL_MGPLL3: + case DPLL_ID_ICL_MGPLL4: + return DDI_CLK_SEL_MG; + } +} + /* Starting with Haswell, different DDI ports can work in FDI mode for * connection to the PCH-located connectors. For this, it is necessary to train * both the DDI port and PCH receiver for the desired DDI buffer settings. @@ -1906,7 +2083,13 @@ u8 intel_ddi_dp_voltage_max(struct intel_encoder *encoder) enum port port = encoder->port; int n_entries; - if (IS_CANNONLAKE(dev_priv)) { + if (IS_ICELAKE(dev_priv)) { + if (port == PORT_A || port == PORT_B) + icl_get_combo_buf_trans(dev_priv, port, encoder->type, + &n_entries); + else + n_entries = ARRAY_SIZE(icl_mg_phy_ddi_translations); + } else if (IS_CANNONLAKE(dev_priv)) { if (encoder->type == INTEL_OUTPUT_EDP) cnl_get_buf_trans_edp(dev_priv, &n_entries); else @@ -2063,6 +2246,146 @@ static void cnl_ddi_vswing_sequence(struct intel_encoder *encoder, I915_WRITE(CNL_PORT_TX_DW5_GRP(port), val); } +static void icl_ddi_combo_vswing_program(struct drm_i915_private *dev_priv, + u32 level, enum port port, int type) +{ + const struct icl_combo_phy_ddi_buf_trans *ddi_translations = NULL; + u32 n_entries, val; + int ln; + + ddi_translations = icl_get_combo_buf_trans(dev_priv, port, type, + &n_entries); + if (!ddi_translations) + return; + + if (level >= n_entries) { + DRM_DEBUG_KMS("DDI translation not found for level %d. Using %d instead.", level, n_entries - 1); + level = n_entries - 1; + } + + /* Set PORT_TX_DW5 Rterm Sel to 110b. */ + val = I915_READ(ICL_PORT_TX_DW5_LN0(port)); + val &= ~RTERM_SELECT_MASK; + val |= RTERM_SELECT(0x6); + I915_WRITE(ICL_PORT_TX_DW5_GRP(port), val); + + /* Program PORT_TX_DW5 */ + val = I915_READ(ICL_PORT_TX_DW5_LN0(port)); + /* Set DisableTap2 and DisableTap3 if MIPI DSI + * Clear DisableTap2 and DisableTap3 for all other Ports + */ + if (type == INTEL_OUTPUT_DSI) { + val |= TAP2_DISABLE; + val |= TAP3_DISABLE; + } else { + val &= ~TAP2_DISABLE; + val &= ~TAP3_DISABLE; + } + I915_WRITE(ICL_PORT_TX_DW5_GRP(port), val); + + /* Program PORT_TX_DW2 */ + val = I915_READ(ICL_PORT_TX_DW2_LN0(port)); + val &= ~(SWING_SEL_LOWER_MASK | SWING_SEL_UPPER_MASK | + RCOMP_SCALAR_MASK); + val |= SWING_SEL_UPPER(ddi_translations[level].dw2_swing_select); + val |= SWING_SEL_LOWER(ddi_translations[level].dw2_swing_select); + /* Program Rcomp scalar for every table entry */ + val |= RCOMP_SCALAR(ddi_translations[level].dw2_swing_scalar); + I915_WRITE(ICL_PORT_TX_DW2_GRP(port), val); + + /* Program PORT_TX_DW4 */ + /* We cannot write to GRP. It would overwrite individual loadgen. */ + for (ln = 0; ln <= 3; ln++) { + val = I915_READ(ICL_PORT_TX_DW4_LN(port, ln)); + val &= ~(POST_CURSOR_1_MASK | POST_CURSOR_2_MASK | + CURSOR_COEFF_MASK); + val |= ddi_translations[level].dw4_scaling; + I915_WRITE(ICL_PORT_TX_DW4_LN(port, ln), val); + } +} + +static void icl_combo_phy_ddi_vswing_sequence(struct intel_encoder *encoder, + u32 level, + enum intel_output_type type) +{ + struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + enum port port = encoder->port; + int width = 0; + int rate = 0; + u32 val; + int ln = 0; + + if (type == INTEL_OUTPUT_HDMI) { + width = 4; + /* Rate is always < than 6GHz for HDMI */ + } else { + struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base); + + width = intel_dp->lane_count; + rate = intel_dp->link_rate; + } + + /* + * 1. If port type is eDP or DP, + * set PORT_PCS_DW1 cmnkeeper_enable to 1b, + * else clear to 0b. + */ + val = I915_READ(ICL_PORT_PCS_DW1_LN0(port)); + if (type == INTEL_OUTPUT_HDMI) + val &= ~COMMON_KEEPER_EN; + else + val |= COMMON_KEEPER_EN; + I915_WRITE(ICL_PORT_PCS_DW1_GRP(port), val); + + /* 2. Program loadgen select */ + /* + * Program PORT_TX_DW4_LN depending on Bit rate and used lanes + * <= 6 GHz and 4 lanes (LN0=0, LN1=1, LN2=1, LN3=1) + * <= 6 GHz and 1,2 lanes (LN0=0, LN1=1, LN2=1, LN3=0) + * > 6 GHz (LN0=0, LN1=0, LN2=0, LN3=0) + */ + for (ln = 0; ln <= 3; ln++) { + val = I915_READ(ICL_PORT_TX_DW4_LN(port, ln)); + val &= ~LOADGEN_SELECT; + + if ((rate <= 600000 && width == 4 && ln >= 1) || + (rate <= 600000 && width < 4 && (ln == 1 || ln == 2))) { + val |= LOADGEN_SELECT; + } + I915_WRITE(ICL_PORT_TX_DW4_LN(port, ln), val); + } + + /* 3. Set PORT_CL_DW5 SUS Clock Config to 11b */ + val = I915_READ(ICL_PORT_CL_DW5(port)); + val |= SUS_CLOCK_CONFIG; + I915_WRITE(ICL_PORT_CL_DW5(port), val); + + /* 4. Clear training enable to change swing values */ + val = I915_READ(ICL_PORT_TX_DW5_LN0(port)); + val &= ~TX_TRAINING_EN; + I915_WRITE(ICL_PORT_TX_DW5_GRP(port), val); + + /* 5. Program swing and de-emphasis */ + icl_ddi_combo_vswing_program(dev_priv, level, port, type); + + /* 6. Set training enable to trigger update */ + val = I915_READ(ICL_PORT_TX_DW5_LN0(port)); + val |= TX_TRAINING_EN; + I915_WRITE(ICL_PORT_TX_DW5_GRP(port), val); +} + +static void icl_ddi_vswing_sequence(struct intel_encoder *encoder, u32 level, + enum intel_output_type type) +{ + enum port port = encoder->port; + + if (port == PORT_A || port == PORT_B) + icl_combo_phy_ddi_vswing_sequence(encoder, level, type); + else + /* Not Implemented Yet */ + WARN_ON(1); +} + static uint32_t translate_signal_level(int signal_levels) { int i; @@ -2094,7 +2417,9 @@ u32 bxt_signal_levels(struct intel_dp *intel_dp) struct intel_encoder *encoder = &dport->base; int level = intel_ddi_dp_level(intel_dp); - if (IS_CANNONLAKE(dev_priv)) + if (IS_ICELAKE(dev_priv)) + icl_ddi_vswing_sequence(encoder, level, encoder->type); + else if (IS_CANNONLAKE(dev_priv)) cnl_ddi_vswing_sequence(encoder, level, encoder->type); else bxt_ddi_vswing_sequence(encoder, level, encoder->type); @@ -2115,6 +2440,69 @@ uint32_t ddi_signal_levels(struct intel_dp *intel_dp) return DDI_BUF_TRANS_SELECT(level); } +void icl_map_plls_to_ports(struct drm_crtc *crtc, + struct intel_crtc_state *crtc_state, + struct drm_atomic_state *old_state) +{ + struct intel_shared_dpll *pll = crtc_state->shared_dpll; + struct drm_i915_private *dev_priv = to_i915(crtc->dev); + struct drm_connector_state *conn_state; + struct drm_connector *conn; + int i; + + for_each_new_connector_in_state(old_state, conn, conn_state, i) { + struct intel_encoder *encoder = + to_intel_encoder(conn_state->best_encoder); + enum port port = encoder->port; + uint32_t val; + + if (conn_state->crtc != crtc) + continue; + + mutex_lock(&dev_priv->dpll_lock); + + val = I915_READ(DPCLKA_CFGCR0_ICL); + WARN_ON((val & DPCLKA_CFGCR0_DDI_CLK_OFF(port)) == 0); + + if (port == PORT_A || port == PORT_B) { + val &= ~DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(port); + val |= DPCLKA_CFGCR0_DDI_CLK_SEL(pll->info->id, port); + I915_WRITE(DPCLKA_CFGCR0_ICL, val); + POSTING_READ(DPCLKA_CFGCR0_ICL); + } + + val &= ~DPCLKA_CFGCR0_DDI_CLK_OFF(port); + I915_WRITE(DPCLKA_CFGCR0_ICL, val); + + mutex_unlock(&dev_priv->dpll_lock); + } +} + +void icl_unmap_plls_to_ports(struct drm_crtc *crtc, + struct intel_crtc_state *crtc_state, + struct drm_atomic_state *old_state) +{ + struct drm_i915_private *dev_priv = to_i915(crtc->dev); + struct drm_connector_state *old_conn_state; + struct drm_connector *conn; + int i; + + for_each_old_connector_in_state(old_state, conn, old_conn_state, i) { + struct intel_encoder *encoder = + to_intel_encoder(old_conn_state->best_encoder); + enum port port = encoder->port; + + if (old_conn_state->crtc != crtc) + continue; + + mutex_lock(&dev_priv->dpll_lock); + I915_WRITE(DPCLKA_CFGCR0_ICL, + I915_READ(DPCLKA_CFGCR0_ICL) | + DPCLKA_CFGCR0_DDI_CLK_OFF(port)); + mutex_unlock(&dev_priv->dpll_lock); + } +} + static void intel_ddi_clk_select(struct intel_encoder *encoder, const struct intel_shared_dpll *pll) { @@ -2127,11 +2515,15 @@ static void intel_ddi_clk_select(struct intel_encoder *encoder, mutex_lock(&dev_priv->dpll_lock); - if (IS_CANNONLAKE(dev_priv)) { + if (IS_ICELAKE(dev_priv)) { + if (port >= PORT_C) + I915_WRITE(DDI_CLK_SEL(port), + icl_pll_to_ddi_pll_sel(encoder, pll)); + } else if (IS_CANNONLAKE(dev_priv)) { /* Configure DPCLKA_CFGCR0 to map the DPLL to the DDI. */ val = I915_READ(DPCLKA_CFGCR0); val &= ~DPCLKA_CFGCR0_DDI_CLK_SEL_MASK(port); - val |= DPCLKA_CFGCR0_DDI_CLK_SEL(pll->id, port); + val |= DPCLKA_CFGCR0_DDI_CLK_SEL(pll->info->id, port); I915_WRITE(DPCLKA_CFGCR0, val); /* @@ -2148,7 +2540,7 @@ static void intel_ddi_clk_select(struct intel_encoder *encoder, val &= ~(DPLL_CTRL2_DDI_CLK_OFF(port) | DPLL_CTRL2_DDI_CLK_SEL_MASK(port)); - val |= (DPLL_CTRL2_DDI_CLK_SEL(pll->id, port) | + val |= (DPLL_CTRL2_DDI_CLK_SEL(pll->info->id, port) | DPLL_CTRL2_DDI_SEL_OVERRIDE(port)); I915_WRITE(DPLL_CTRL2, val); @@ -2165,14 +2557,18 @@ static void intel_ddi_clk_disable(struct intel_encoder *encoder) struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); enum port port = encoder->port; - if (IS_CANNONLAKE(dev_priv)) + if (IS_ICELAKE(dev_priv)) { + if (port >= PORT_C) + I915_WRITE(DDI_CLK_SEL(port), DDI_CLK_SEL_NONE); + } else if (IS_CANNONLAKE(dev_priv)) { I915_WRITE(DPCLKA_CFGCR0, I915_READ(DPCLKA_CFGCR0) | DPCLKA_CFGCR0_DDI_CLK_OFF(port)); - else if (IS_GEN9_BC(dev_priv)) + } else if (IS_GEN9_BC(dev_priv)) { I915_WRITE(DPLL_CTRL2, I915_READ(DPLL_CTRL2) | DPLL_CTRL2_DDI_CLK_OFF(port)); - else if (INTEL_GEN(dev_priv) < 9) + } else if (INTEL_GEN(dev_priv) < 9) { I915_WRITE(PORT_CLK_SEL(port), PORT_CLK_SEL_NONE); + } } static void intel_ddi_pre_enable_dp(struct intel_encoder *encoder, @@ -2197,7 +2593,9 @@ static void intel_ddi_pre_enable_dp(struct intel_encoder *encoder, intel_display_power_get(dev_priv, dig_port->ddi_io_power_domain); - if (IS_CANNONLAKE(dev_priv)) + if (IS_ICELAKE(dev_priv)) + icl_ddi_vswing_sequence(encoder, level, encoder->type); + else if (IS_CANNONLAKE(dev_priv)) cnl_ddi_vswing_sequence(encoder, level, encoder->type); else if (IS_GEN9_LP(dev_priv)) bxt_ddi_vswing_sequence(encoder, level, encoder->type); @@ -2205,7 +2603,8 @@ static void intel_ddi_pre_enable_dp(struct intel_encoder *encoder, intel_prepare_dp_ddi_buffers(encoder, crtc_state); intel_ddi_init_dp_buf_reg(encoder); - intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON); + if (!is_mst) + intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON); intel_dp_start_link_train(intel_dp); if (port != PORT_A || INTEL_GEN(dev_priv) >= 9) intel_dp_stop_link_train(intel_dp); @@ -2227,7 +2626,9 @@ static void intel_ddi_pre_enable_hdmi(struct intel_encoder *encoder, intel_display_power_get(dev_priv, dig_port->ddi_io_power_domain); - if (IS_CANNONLAKE(dev_priv)) + if (IS_ICELAKE(dev_priv)) + icl_ddi_vswing_sequence(encoder, level, INTEL_OUTPUT_HDMI); + else if (IS_CANNONLAKE(dev_priv)) cnl_ddi_vswing_sequence(encoder, level, INTEL_OUTPUT_HDMI); else if (IS_GEN9_LP(dev_priv)) bxt_ddi_vswing_sequence(encoder, level, INTEL_OUTPUT_HDMI); @@ -2303,12 +2704,15 @@ static void intel_ddi_post_disable_dp(struct intel_encoder *encoder, struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_digital_port *dig_port = enc_to_dig_port(&encoder->base); struct intel_dp *intel_dp = &dig_port->dp; + bool is_mst = intel_crtc_has_type(old_crtc_state, + INTEL_OUTPUT_DP_MST); /* * Power down sink before disabling the port, otherwise we end * up getting interrupts from the sink on detecting link loss. */ - intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_OFF); + if (!is_mst) + intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_OFF); intel_disable_ddi_buf(encoder); @@ -2424,12 +2828,14 @@ static void intel_enable_ddi_hdmi(struct intel_encoder *encoder, { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_digital_port *dig_port = enc_to_dig_port(&encoder->base); + struct drm_connector *connector = conn_state->connector; enum port port = encoder->port; - intel_hdmi_handle_sink_scrambling(encoder, - conn_state->connector, - crtc_state->hdmi_high_tmds_clock_ratio, - crtc_state->hdmi_scrambling); + if (!intel_hdmi_handle_sink_scrambling(encoder, connector, + crtc_state->hdmi_high_tmds_clock_ratio, + crtc_state->hdmi_scrambling)) + DRM_ERROR("[CONNECTOR:%d:%s] Failed to configure sink scrambling/TMDS bit clock ratio\n", + connector->base.id, connector->name); /* Display WA #1143: skl,kbl,cfl */ if (IS_GEN9_BC(dev_priv)) { @@ -2520,13 +2926,16 @@ static void intel_disable_ddi_hdmi(struct intel_encoder *encoder, const struct intel_crtc_state *old_crtc_state, const struct drm_connector_state *old_conn_state) { + struct drm_connector *connector = old_conn_state->connector; + if (old_crtc_state->has_audio) intel_audio_codec_disable(encoder, old_crtc_state, old_conn_state); - intel_hdmi_handle_sink_scrambling(encoder, - old_conn_state->connector, - false, false); + if (!intel_hdmi_handle_sink_scrambling(encoder, connector, + false, false)) + DRM_DEBUG_KMS("[CONNECTOR:%d:%s] Failed to reset sink scrambling/TMDS bit clock ratio\n", + connector->base.id, connector->name); } static void intel_disable_ddi(struct intel_encoder *encoder, diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c index 3dd350f..0fd13df 100644 --- a/drivers/gpu/drm/i915/intel_device_info.c +++ b/drivers/gpu/drm/i915/intel_device_info.c @@ -83,11 +83,11 @@ static void sseu_dump(const struct sseu_dev_info *sseu, struct drm_printer *p) { int s; - drm_printf(p, "slice mask: %04x\n", sseu->slice_mask); - drm_printf(p, "slice total: %u\n", hweight8(sseu->slice_mask)); + drm_printf(p, "slice total: %u, mask=%04x\n", + hweight8(sseu->slice_mask), sseu->slice_mask); drm_printf(p, "subslice total: %u\n", sseu_subslice_total(sseu)); - for (s = 0; s < ARRAY_SIZE(sseu->subslice_mask); s++) { - drm_printf(p, "slice%d %u subslices mask=%04x\n", + for (s = 0; s < sseu->max_slices; s++) { + drm_printf(p, "slice%d: %u subslices, mask=%04x\n", s, hweight8(sseu->subslice_mask[s]), sseu->subslice_mask[s]); } @@ -158,6 +158,45 @@ static u16 compute_eu_total(const struct sseu_dev_info *sseu) return total; } +static void gen11_sseu_info_init(struct drm_i915_private *dev_priv) +{ + struct sseu_dev_info *sseu = &mkwrite_device_info(dev_priv)->sseu; + u8 s_en; + u32 ss_en, ss_en_mask; + u8 eu_en; + int s; + + sseu->max_slices = 1; + sseu->max_subslices = 8; + sseu->max_eus_per_subslice = 8; + + s_en = I915_READ(GEN11_GT_SLICE_ENABLE) & GEN11_GT_S_ENA_MASK; + ss_en = ~I915_READ(GEN11_GT_SUBSLICE_DISABLE); + ss_en_mask = BIT(sseu->max_subslices) - 1; + eu_en = ~(I915_READ(GEN11_EU_DISABLE) & GEN11_EU_DIS_MASK); + + for (s = 0; s < sseu->max_slices; s++) { + if (s_en & BIT(s)) { + int ss_idx = sseu->max_subslices * s; + int ss; + + sseu->slice_mask |= BIT(s); + sseu->subslice_mask[s] = (ss_en >> ss_idx) & ss_en_mask; + for (ss = 0; ss < sseu->max_subslices; ss++) { + if (sseu->subslice_mask[s] & BIT(ss)) + sseu_set_eus(sseu, s, ss, eu_en); + } + } + } + sseu->eu_per_subslice = hweight8(eu_en); + sseu->eu_total = compute_eu_total(sseu); + + /* ICL has no power gating restrictions. */ + sseu->has_slice_pg = 1; + sseu->has_subslice_pg = 1; + sseu->has_eu_pg = 1; +} + static void gen10_sseu_info_init(struct drm_i915_private *dev_priv) { struct sseu_dev_info *sseu = &mkwrite_device_info(dev_priv)->sseu; @@ -557,6 +596,52 @@ static u32 read_reference_ts_freq(struct drm_i915_private *dev_priv) return base_freq + frac_freq; } +static u32 gen10_get_crystal_clock_freq(struct drm_i915_private *dev_priv, + u32 rpm_config_reg) +{ + u32 f19_2_mhz = 19200; + u32 f24_mhz = 24000; + u32 crystal_clock = (rpm_config_reg & + GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK) >> + GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT; + + switch (crystal_clock) { + case GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ: + return f19_2_mhz; + case GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ: + return f24_mhz; + default: + MISSING_CASE(crystal_clock); + return 0; + } +} + +static u32 gen11_get_crystal_clock_freq(struct drm_i915_private *dev_priv, + u32 rpm_config_reg) +{ + u32 f19_2_mhz = 19200; + u32 f24_mhz = 24000; + u32 f25_mhz = 25000; + u32 f38_4_mhz = 38400; + u32 crystal_clock = (rpm_config_reg & + GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK) >> + GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT; + + switch (crystal_clock) { + case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ: + return f24_mhz; + case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ: + return f19_2_mhz; + case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_38_4_MHZ: + return f38_4_mhz; + case GEN11_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_25_MHZ: + return f25_mhz; + default: + MISSING_CASE(crystal_clock); + return 0; + } +} + static u32 read_timestamp_frequency(struct drm_i915_private *dev_priv) { u32 f12_5_mhz = 12500; @@ -597,10 +682,9 @@ static u32 read_timestamp_frequency(struct drm_i915_private *dev_priv) } return freq; - } else if (INTEL_GEN(dev_priv) <= 10) { + } else if (INTEL_GEN(dev_priv) <= 11) { u32 ctc_reg = I915_READ(CTC_MODE); u32 freq = 0; - u32 rpm_config_reg = 0; /* First figure out the reference frequency. There are 2 ways * we can compute the frequency, either through the @@ -610,20 +694,14 @@ static u32 read_timestamp_frequency(struct drm_i915_private *dev_priv) if ((ctc_reg & CTC_SOURCE_PARAMETER_MASK) == CTC_SOURCE_DIVIDE_LOGIC) { freq = read_reference_ts_freq(dev_priv); } else { - u32 crystal_clock; - - rpm_config_reg = I915_READ(RPM_CONFIG0); - crystal_clock = (rpm_config_reg & - GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_MASK) >> - GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_SHIFT; - switch (crystal_clock) { - case GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_19_2_MHZ: - freq = f19_2_mhz; - break; - case GEN9_RPM_CONFIG0_CRYSTAL_CLOCK_FREQ_24_MHZ: - freq = f24_mhz; - break; - } + u32 rpm_config_reg = I915_READ(RPM_CONFIG0); + + if (INTEL_GEN(dev_priv) <= 10) + freq = gen10_get_crystal_clock_freq(dev_priv, + rpm_config_reg); + else + freq = gen11_get_crystal_clock_freq(dev_priv, + rpm_config_reg); /* Now figure out how the command stream's timestamp * register increments from this frequency (it might @@ -768,8 +846,10 @@ void intel_device_info_runtime_init(struct intel_device_info *info) broadwell_sseu_info_init(dev_priv); else if (INTEL_GEN(dev_priv) == 9) gen9_sseu_info_init(dev_priv); - else if (INTEL_GEN(dev_priv) >= 10) + else if (INTEL_GEN(dev_priv) == 10) gen10_sseu_info_init(dev_priv); + else if (INTEL_GEN(dev_priv) >= 11) + gen11_sseu_info_init(dev_priv); /* Initialize command stream timestamp frequency */ info->cs_timestamp_frequency_khz = read_timestamp_frequency(dev_priv); @@ -780,3 +860,50 @@ void intel_driver_caps_print(const struct intel_driver_caps *caps, { drm_printf(p, "scheduler: %x\n", caps->scheduler); } + +/* + * Determine which engines are fused off in our particular hardware. Since the + * fuse register is in the blitter powerwell, we need forcewake to be ready at + * this point (but later we need to prune the forcewake domains for engines that + * are indeed fused off). + */ +void intel_device_info_init_mmio(struct drm_i915_private *dev_priv) +{ + struct intel_device_info *info = mkwrite_device_info(dev_priv); + u8 vdbox_disable, vebox_disable; + u32 media_fuse; + int i; + + if (INTEL_GEN(dev_priv) < 11) + return; + + media_fuse = I915_READ(GEN11_GT_VEBOX_VDBOX_DISABLE); + + vdbox_disable = media_fuse & GEN11_GT_VDBOX_DISABLE_MASK; + vebox_disable = (media_fuse & GEN11_GT_VEBOX_DISABLE_MASK) >> + GEN11_GT_VEBOX_DISABLE_SHIFT; + + DRM_DEBUG_DRIVER("vdbox disable: %04x\n", vdbox_disable); + for (i = 0; i < I915_MAX_VCS; i++) { + if (!HAS_ENGINE(dev_priv, _VCS(i))) + continue; + + if (!(BIT(i) & vdbox_disable)) + continue; + + info->ring_mask &= ~ENGINE_MASK(_VCS(i)); + DRM_DEBUG_DRIVER("vcs%u fused off\n", i); + } + + DRM_DEBUG_DRIVER("vebox disable: %04x\n", vebox_disable); + for (i = 0; i < I915_MAX_VECS; i++) { + if (!HAS_ENGINE(dev_priv, _VECS(i))) + continue; + + if (!(BIT(i) & vebox_disable)) + continue; + + info->ring_mask &= ~ENGINE_MASK(_VECS(i)); + DRM_DEBUG_DRIVER("vecs%u fused off\n", i); + } +} diff --git a/drivers/gpu/drm/i915/intel_device_info.h b/drivers/gpu/drm/i915/intel_device_info.h index 0835752..933e316 100644 --- a/drivers/gpu/drm/i915/intel_device_info.h +++ b/drivers/gpu/drm/i915/intel_device_info.h @@ -114,7 +114,7 @@ enum intel_platform { func(has_ipc); #define GEN_MAX_SLICES (6) /* CNL upper bound */ -#define GEN_MAX_SUBSLICES (7) +#define GEN_MAX_SUBSLICES (8) /* ICL upper bound */ struct sseu_dev_info { u8 slice_mask; @@ -247,6 +247,8 @@ void intel_device_info_dump_runtime(const struct intel_device_info *info, void intel_device_info_dump_topology(const struct sseu_dev_info *sseu, struct drm_printer *p); +void intel_device_info_init_mmio(struct drm_i915_private *dev_priv); + void intel_driver_caps_print(const struct intel_driver_caps *caps, struct drm_printer *p); diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 56004ff..ad588d5 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -88,6 +88,22 @@ static const uint32_t skl_primary_formats[] = { DRM_FORMAT_VYUY, }; +static const uint32_t skl_pri_planar_formats[] = { + DRM_FORMAT_C8, + DRM_FORMAT_RGB565, + DRM_FORMAT_XRGB8888, + DRM_FORMAT_XBGR8888, + DRM_FORMAT_ARGB8888, + DRM_FORMAT_ABGR8888, + DRM_FORMAT_XRGB2101010, + DRM_FORMAT_XBGR2101010, + DRM_FORMAT_YUYV, + DRM_FORMAT_YVYU, + DRM_FORMAT_UYVY, + DRM_FORMAT_VYUY, + DRM_FORMAT_NV12, +}; + static const uint64_t skl_format_modifiers_noccs[] = { I915_FORMAT_MOD_Yf_TILED, I915_FORMAT_MOD_Y_TILED, @@ -488,6 +504,33 @@ static const struct intel_limit intel_limits_bxt = { .p2 = { .p2_slow = 1, .p2_fast = 20 }, }; +static void +skl_wa_528(struct drm_i915_private *dev_priv, int pipe, bool enable) +{ + if (IS_SKYLAKE(dev_priv) || IS_BROXTON(dev_priv)) + return; + + if (enable) + I915_WRITE(CHICKEN_PIPESL_1(pipe), HSW_FBCQ_DIS); + else + I915_WRITE(CHICKEN_PIPESL_1(pipe), 0); +} + +static void +skl_wa_clkgate(struct drm_i915_private *dev_priv, int pipe, bool enable) +{ + if (IS_SKYLAKE(dev_priv) || IS_BROXTON(dev_priv)) + return; + + if (enable) + I915_WRITE(CLKGATE_DIS_PSL(pipe), + DUPS1_GATING_DIS | DUPS2_GATING_DIS); + else + I915_WRITE(CLKGATE_DIS_PSL(pipe), + I915_READ(CLKGATE_DIS_PSL(pipe)) & + ~(DUPS1_GATING_DIS | DUPS2_GATING_DIS)); +} + static bool needs_modeset(const struct drm_crtc_state *state) { @@ -2657,11 +2700,13 @@ static int i9xx_format_to_fourcc(int format) } } -static int skl_format_to_fourcc(int format, bool rgb_order, bool alpha) +int skl_format_to_fourcc(int format, bool rgb_order, bool alpha) { switch (format) { case PLANE_CTL_FORMAT_RGB_565: return DRM_FORMAT_RGB565; + case PLANE_CTL_FORMAT_NV12: + return DRM_FORMAT_NV12; default: case PLANE_CTL_FORMAT_XRGB_8888: if (rgb_order) { @@ -2824,7 +2869,7 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, continue; if (intel_plane_ggtt_offset(state) == plane_config->base) { - fb = c->primary->fb; + fb = state->base.fb; drm_framebuffer_get(fb); goto valid_fb; } @@ -2858,6 +2903,9 @@ valid_fb: return; } + obj = intel_fb_obj(fb); + intel_fb_obj_flush(obj, ORIGIN_DIRTYFB); + plane_state->src_x = 0; plane_state->src_y = 0; plane_state->src_w = fb->width << 16; @@ -2871,7 +2919,6 @@ valid_fb: intel_state->base.src = drm_plane_state_src(plane_state); intel_state->base.dst = drm_plane_state_dest(plane_state); - obj = intel_fb_obj(fb); if (i915_gem_object_is_tiled(obj)) dev_priv->preserve_bios_swizzle = true; @@ -3071,6 +3118,29 @@ static int skl_check_main_surface(const struct intel_crtc_state *crtc_state, return 0; } +static int +skl_check_nv12_surface(const struct intel_crtc_state *crtc_state, + struct intel_plane_state *plane_state) +{ + /* Display WA #1106 */ + if (plane_state->base.rotation != + (DRM_MODE_REFLECT_X | DRM_MODE_ROTATE_90) && + plane_state->base.rotation != DRM_MODE_ROTATE_270) + return 0; + + /* + * src coordinates are rotated here. + * We check height but report it as width + */ + if (((drm_rect_height(&plane_state->base.src) >> 16) % 4) != 0) { + DRM_DEBUG_KMS("src width must be multiple " + "of 4 for rotated NV12\n"); + return -EINVAL; + } + + return 0; +} + static int skl_check_nv12_aux_surface(struct intel_plane_state *plane_state) { const struct drm_framebuffer *fb = plane_state->base.fb; @@ -3154,6 +3224,9 @@ int skl_check_plane_surface(const struct intel_crtc_state *crtc_state, * the main surface setup depends on it. */ if (fb->format->format == DRM_FORMAT_NV12) { + ret = skl_check_nv12_surface(crtc_state, plane_state); + if (ret) + return ret; ret = skl_check_nv12_aux_surface(plane_state); if (ret) return ret; @@ -3464,6 +3537,8 @@ static u32 skl_plane_ctl_format(uint32_t pixel_format) return PLANE_CTL_FORMAT_YUV422 | PLANE_CTL_YUV422_UYVY; case DRM_FORMAT_VYUY: return PLANE_CTL_FORMAT_YUV422 | PLANE_CTL_YUV422_VYUY; + case DRM_FORMAT_NV12: + return PLANE_CTL_FORMAT_NV12; default: MISSING_CASE(pixel_format); } @@ -3602,15 +3677,24 @@ u32 skl_plane_ctl(const struct intel_crtc_state *crtc_state, u32 glk_plane_color_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { + struct drm_i915_private *dev_priv = + to_i915(plane_state->base.plane->dev); const struct drm_framebuffer *fb = plane_state->base.fb; u32 plane_color_ctl = 0; - plane_color_ctl |= PLANE_COLOR_PIPE_GAMMA_ENABLE; - plane_color_ctl |= PLANE_COLOR_PIPE_CSC_ENABLE; + if (INTEL_GEN(dev_priv) < 11) { + plane_color_ctl |= PLANE_COLOR_PIPE_GAMMA_ENABLE; + plane_color_ctl |= PLANE_COLOR_PIPE_CSC_ENABLE; + } plane_color_ctl |= PLANE_COLOR_PLANE_GAMMA_DISABLE; plane_color_ctl |= glk_plane_color_ctl_alpha(fb->format->format); if (intel_format_is_yuv(fb->format->format)) { + if (fb->format->format == DRM_FORMAT_NV12) { + plane_color_ctl |= + PLANE_COLOR_CSC_MODE_YUV709_TO_RGB709; + goto out; + } if (plane_state->base.color_encoding == DRM_COLOR_YCBCR_BT709) plane_color_ctl |= PLANE_COLOR_CSC_MODE_YUV709_TO_RGB709; else @@ -3619,7 +3703,7 @@ u32 glk_plane_color_ctl(const struct intel_crtc_state *crtc_state, if (plane_state->base.color_range == DRM_COLOR_YCBCR_FULL_RANGE) plane_color_ctl |= PLANE_COLOR_YUV_RANGE_CORRECTION_DISABLE; } - +out: return plane_color_ctl; } @@ -3675,7 +3759,6 @@ void intel_prepare_reset(struct drm_i915_private *dev_priv) struct drm_atomic_state *state; int ret; - /* reset doesn't touch the display */ if (!i915_modparams.force_reset_modeset_test && !gpu_reset_clobbers_display(dev_priv)) @@ -3729,19 +3812,17 @@ void intel_finish_reset(struct drm_i915_private *dev_priv) { struct drm_device *dev = &dev_priv->drm; struct drm_modeset_acquire_ctx *ctx = &dev_priv->reset_ctx; - struct drm_atomic_state *state = dev_priv->modeset_restore_state; + struct drm_atomic_state *state; int ret; /* reset doesn't touch the display */ - if (!i915_modparams.force_reset_modeset_test && - !gpu_reset_clobbers_display(dev_priv)) + if (!test_bit(I915_RESET_MODESET, &dev_priv->gpu_error.flags)) return; + state = fetch_and_zero(&dev_priv->modeset_restore_state); if (!state) goto unlock; - dev_priv->modeset_restore_state = NULL; - /* reset doesn't touch the display */ if (!gpu_reset_clobbers_display(dev_priv)) { /* for testing only restore the display */ @@ -4703,7 +4784,9 @@ static void cpt_verify_modeset(struct drm_device *dev, int pipe) static int skl_update_scaler(struct intel_crtc_state *crtc_state, bool force_detach, unsigned int scaler_user, int *scaler_id, - int src_w, int src_h, int dst_w, int dst_h) + int src_w, int src_h, int dst_w, int dst_h, + bool plane_scaler_check, + uint32_t pixel_format) { struct intel_crtc_scaler_state *scaler_state = &crtc_state->scaler_state; @@ -4721,6 +4804,10 @@ skl_update_scaler(struct intel_crtc_state *crtc_state, bool force_detach, */ need_scaling = src_w != dst_w || src_h != dst_h; + if (plane_scaler_check) + if (pixel_format == DRM_FORMAT_NV12) + need_scaling = true; + if (crtc_state->ycbcr420 && scaler_user == SKL_CRTC_INDEX) need_scaling = true; @@ -4760,12 +4847,21 @@ skl_update_scaler(struct intel_crtc_state *crtc_state, bool force_detach, return 0; } + if (plane_scaler_check && pixel_format == DRM_FORMAT_NV12 && + (src_h < SKL_MIN_YUV_420_SRC_H || src_w < SKL_MIN_YUV_420_SRC_W)) { + DRM_DEBUG_KMS("NV12: src dimensions not met\n"); + return -EINVAL; + } + /* range checks */ if (src_w < SKL_MIN_SRC_W || src_h < SKL_MIN_SRC_H || - dst_w < SKL_MIN_DST_W || dst_h < SKL_MIN_DST_H || - - src_w > SKL_MAX_SRC_W || src_h > SKL_MAX_SRC_H || - dst_w > SKL_MAX_DST_W || dst_h > SKL_MAX_DST_H) { + dst_w < SKL_MIN_DST_W || dst_h < SKL_MIN_DST_H || + (IS_GEN11(dev_priv) && + (src_w > ICL_MAX_SRC_W || src_h > ICL_MAX_SRC_H || + dst_w > ICL_MAX_DST_W || dst_h > ICL_MAX_DST_H)) || + (!IS_GEN11(dev_priv) && + (src_w > SKL_MAX_SRC_W || src_h > SKL_MAX_SRC_H || + dst_w > SKL_MAX_DST_W || dst_h > SKL_MAX_DST_H))) { DRM_DEBUG_KMS("scaler_user index %u.%u: src %ux%u dst %ux%u " "size is out of scaler range\n", intel_crtc->pipe, scaler_user, src_w, src_h, dst_w, dst_h); @@ -4796,9 +4892,10 @@ int skl_update_scaler_crtc(struct intel_crtc_state *state) const struct drm_display_mode *adjusted_mode = &state->base.adjusted_mode; return skl_update_scaler(state, !state->base.active, SKL_CRTC_INDEX, - &state->scaler_state.scaler_id, - state->pipe_src_w, state->pipe_src_h, - adjusted_mode->crtc_hdisplay, adjusted_mode->crtc_vdisplay); + &state->scaler_state.scaler_id, + state->pipe_src_w, state->pipe_src_h, + adjusted_mode->crtc_hdisplay, + adjusted_mode->crtc_vdisplay, false, 0); } /** @@ -4827,7 +4924,8 @@ static int skl_update_scaler_plane(struct intel_crtc_state *crtc_state, drm_rect_width(&plane_state->base.src) >> 16, drm_rect_height(&plane_state->base.src) >> 16, drm_rect_width(&plane_state->base.dst), - drm_rect_height(&plane_state->base.dst)); + drm_rect_height(&plane_state->base.dst), + fb ? true : false, fb ? fb->format->format : 0); if (ret || plane_state->scaler_id < 0) return ret; @@ -4853,6 +4951,7 @@ static int skl_update_scaler_plane(struct intel_crtc_state *crtc_state, case DRM_FORMAT_YVYU: case DRM_FORMAT_UYVY: case DRM_FORMAT_VYUY: + case DRM_FORMAT_NV12: break; default: DRM_DEBUG_KMS("[PLANE:%d:%s] FB:%d unsupported scaling format 0x%x\n", @@ -5096,16 +5195,34 @@ static bool hsw_post_update_enable_ips(const struct intel_crtc_state *old_crtc_s return !old_crtc_state->ips_enabled; } +static bool needs_nv12_wa(struct drm_i915_private *dev_priv, + const struct intel_crtc_state *crtc_state) +{ + if (!crtc_state->nv12_planes) + return false; + + if (IS_SKYLAKE(dev_priv) || IS_BROXTON(dev_priv)) + return false; + + if ((INTEL_GEN(dev_priv) == 9 && !IS_GEMINILAKE(dev_priv)) || + IS_CANNONLAKE(dev_priv)) + return true; + + return false; +} + static void intel_post_plane_update(struct intel_crtc_state *old_crtc_state) { struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->base.crtc); + struct drm_device *dev = crtc->base.dev; + struct drm_i915_private *dev_priv = to_i915(dev); struct drm_atomic_state *old_state = old_crtc_state->base.state; struct intel_crtc_state *pipe_config = intel_atomic_get_new_crtc_state(to_intel_atomic_state(old_state), crtc); struct drm_plane *primary = crtc->base.primary; - struct drm_plane_state *old_pri_state = - drm_atomic_get_existing_plane_state(old_state, primary); + struct drm_plane_state *old_primary_state = + drm_atomic_get_old_plane_state(old_state, primary); intel_frontbuffer_flip(to_i915(crtc->base.dev), pipe_config->fb_bits); @@ -5115,20 +5232,24 @@ static void intel_post_plane_update(struct intel_crtc_state *old_crtc_state) if (hsw_post_update_enable_ips(old_crtc_state, pipe_config)) hsw_enable_ips(pipe_config); - if (old_pri_state) { - struct intel_plane_state *primary_state = - intel_atomic_get_new_plane_state(to_intel_atomic_state(old_state), - to_intel_plane(primary)); - struct intel_plane_state *old_primary_state = - to_intel_plane_state(old_pri_state); + if (old_primary_state) { + struct drm_plane_state *new_primary_state = + drm_atomic_get_new_plane_state(old_state, primary); intel_fbc_post_update(crtc); - if (primary_state->base.visible && + if (new_primary_state->visible && (needs_modeset(&pipe_config->base) || - !old_primary_state->base.visible)) + !old_primary_state->visible)) intel_post_enable_primary(&crtc->base, pipe_config); } + + /* Display WA 827 */ + if (needs_nv12_wa(dev_priv, old_crtc_state) && + !needs_nv12_wa(dev_priv, pipe_config)) { + skl_wa_clkgate(dev_priv, crtc->pipe, false); + skl_wa_528(dev_priv, crtc->pipe, false); + } } static void intel_pre_plane_update(struct intel_crtc_state *old_crtc_state, @@ -5139,8 +5260,8 @@ static void intel_pre_plane_update(struct intel_crtc_state *old_crtc_state, struct drm_i915_private *dev_priv = to_i915(dev); struct drm_atomic_state *old_state = old_crtc_state->base.state; struct drm_plane *primary = crtc->base.primary; - struct drm_plane_state *old_pri_state = - drm_atomic_get_existing_plane_state(old_state, primary); + struct drm_plane_state *old_primary_state = + drm_atomic_get_old_plane_state(old_state, primary); bool modeset = needs_modeset(&pipe_config->base); struct intel_atomic_state *old_intel_state = to_intel_atomic_state(old_state); @@ -5148,23 +5269,28 @@ static void intel_pre_plane_update(struct intel_crtc_state *old_crtc_state, if (hsw_pre_update_disable_ips(old_crtc_state, pipe_config)) hsw_disable_ips(old_crtc_state); - if (old_pri_state) { - struct intel_plane_state *primary_state = + if (old_primary_state) { + struct intel_plane_state *new_primary_state = intel_atomic_get_new_plane_state(old_intel_state, to_intel_plane(primary)); - struct intel_plane_state *old_primary_state = - to_intel_plane_state(old_pri_state); - intel_fbc_pre_update(crtc, pipe_config, primary_state); + intel_fbc_pre_update(crtc, pipe_config, new_primary_state); /* * Gen2 reports pipe underruns whenever all planes are disabled. * So disable underrun reporting before all the planes get disabled. */ - if (IS_GEN2(dev_priv) && old_primary_state->base.visible && - (modeset || !primary_state->base.visible)) + if (IS_GEN2(dev_priv) && old_primary_state->visible && + (modeset || !new_primary_state->base.visible)) intel_set_cpu_fifo_underrun_reporting(dev_priv, crtc->pipe, false); } + /* Display WA 827 */ + if (!needs_nv12_wa(dev_priv, old_crtc_state) && + needs_nv12_wa(dev_priv, pipe_config)) { + skl_wa_clkgate(dev_priv, crtc->pipe, true); + skl_wa_528(dev_priv, crtc->pipe, true); + } + /* * Vblank time updates from the shadow to live plane control register * are blocked if the memory self-refresh mode is active at that @@ -5499,6 +5625,9 @@ static void haswell_crtc_enable(struct intel_crtc_state *pipe_config, if (intel_crtc->config->shared_dpll) intel_enable_shared_dpll(intel_crtc); + if (INTEL_GEN(dev_priv) >= 11) + icl_map_plls_to_ports(crtc, pipe_config, old_state); + if (intel_crtc_has_dp_encoder(intel_crtc->config)) intel_dp_set_m_n(intel_crtc, M1_N1); @@ -5696,6 +5825,9 @@ static void haswell_crtc_disable(struct intel_crtc_state *old_crtc_state, intel_ddi_disable_pipe_clock(intel_crtc->config); intel_encoders_post_disable(crtc, old_crtc_state, old_state); + + if (INTEL_GEN(dev_priv) >= 11) + icl_unmap_plls_to_ports(crtc, old_crtc_state, old_state); } static void i9xx_pfit_enable(struct intel_crtc *crtc) @@ -8766,8 +8898,8 @@ static bool ironlake_get_pipe_config(struct intel_crtc *crtc, intel_get_shared_dpll_by_id(dev_priv, pll_id); pll = pipe_config->shared_dpll; - WARN_ON(!pll->funcs.get_hw_state(dev_priv, pll, - &pipe_config->dpll_hw_state)); + WARN_ON(!pll->info->funcs->get_hw_state(dev_priv, pll, + &pipe_config->dpll_hw_state)); tmp = pipe_config->dpll_hw_state.dpll; pipe_config->pixel_multiplier = @@ -9243,8 +9375,8 @@ static void haswell_get_ddi_port_state(struct intel_crtc *crtc, pll = pipe_config->shared_dpll; if (pll) { - WARN_ON(!pll->funcs.get_hw_state(dev_priv, pll, - &pipe_config->dpll_hw_state)); + WARN_ON(!pll->info->funcs->get_hw_state(dev_priv, pll, + &pipe_config->dpll_hw_state)); } /* @@ -9974,6 +10106,8 @@ found: ret = PTR_ERR_OR_ZERO(drm_atomic_get_connector_state(restore_state, connector)); if (!ret) ret = PTR_ERR_OR_ZERO(drm_atomic_get_crtc_state(restore_state, crtc)); + if (!ret) + ret = drm_atomic_add_affected_planes(restore_state, crtc); if (ret) { DRM_DEBUG_KMS("Failed to create a copy of old state to restore: %i\n", ret); goto fail; @@ -10773,7 +10907,7 @@ static bool check_digital_port_conflicts(struct drm_atomic_state *state) struct drm_connector_state *connector_state; struct intel_encoder *encoder; - connector_state = drm_atomic_get_existing_connector_state(state, connector); + connector_state = drm_atomic_get_new_connector_state(state, connector); if (!connector_state) connector_state = connector->state; @@ -11085,39 +11219,42 @@ intel_pipe_config_compare(struct drm_i915_private *dev_priv, (current_config->base.mode.private_flags & I915_MODE_FLAG_INHERITED) && !(pipe_config->base.mode.private_flags & I915_MODE_FLAG_INHERITED); -#define PIPE_CONF_CHECK_X(name) \ +#define PIPE_CONF_CHECK_X(name) do { \ if (current_config->name != pipe_config->name) { \ pipe_config_err(adjust, __stringify(name), \ "(expected 0x%08x, found 0x%08x)\n", \ current_config->name, \ pipe_config->name); \ ret = false; \ - } + } \ +} while (0) -#define PIPE_CONF_CHECK_I(name) \ +#define PIPE_CONF_CHECK_I(name) do { \ if (current_config->name != pipe_config->name) { \ pipe_config_err(adjust, __stringify(name), \ "(expected %i, found %i)\n", \ current_config->name, \ pipe_config->name); \ ret = false; \ - } + } \ +} while (0) -#define PIPE_CONF_CHECK_BOOL(name) \ +#define PIPE_CONF_CHECK_BOOL(name) do { \ if (current_config->name != pipe_config->name) { \ pipe_config_err(adjust, __stringify(name), \ "(expected %s, found %s)\n", \ yesno(current_config->name), \ yesno(pipe_config->name)); \ ret = false; \ - } + } \ +} while (0) /* * Checks state where we only read out the enabling, but not the entire * state itself (like full infoframes or ELD for audio). These states * require a full modeset on bootup to fix up. */ -#define PIPE_CONF_CHECK_BOOL_INCOMPLETE(name) \ +#define PIPE_CONF_CHECK_BOOL_INCOMPLETE(name) do { \ if (!fixup_inherited || (!current_config->name && !pipe_config->name)) { \ PIPE_CONF_CHECK_BOOL(name); \ } else { \ @@ -11126,18 +11263,20 @@ intel_pipe_config_compare(struct drm_i915_private *dev_priv, yesno(current_config->name), \ yesno(pipe_config->name)); \ ret = false; \ - } + } \ +} while (0) -#define PIPE_CONF_CHECK_P(name) \ +#define PIPE_CONF_CHECK_P(name) do { \ if (current_config->name != pipe_config->name) { \ pipe_config_err(adjust, __stringify(name), \ "(expected %p, found %p)\n", \ current_config->name, \ pipe_config->name); \ ret = false; \ - } + } \ +} while (0) -#define PIPE_CONF_CHECK_M_N(name) \ +#define PIPE_CONF_CHECK_M_N(name) do { \ if (!intel_compare_link_m_n(¤t_config->name, \ &pipe_config->name,\ adjust)) { \ @@ -11155,14 +11294,15 @@ intel_pipe_config_compare(struct drm_i915_private *dev_priv, pipe_config->name.link_m, \ pipe_config->name.link_n); \ ret = false; \ - } + } \ +} while (0) /* This is required for BDW+ where there is only one set of registers for * switching between high and low RR. * This macro can be used whenever a comparison has to be made between one * hw state and multiple sw state variables. */ -#define PIPE_CONF_CHECK_M_N_ALT(name, alt_name) \ +#define PIPE_CONF_CHECK_M_N_ALT(name, alt_name) do { \ if (!intel_compare_link_m_n(¤t_config->name, \ &pipe_config->name, adjust) && \ !intel_compare_link_m_n(¤t_config->alt_name, \ @@ -11187,9 +11327,10 @@ intel_pipe_config_compare(struct drm_i915_private *dev_priv, pipe_config->name.link_m, \ pipe_config->name.link_n); \ ret = false; \ - } + } \ +} while (0) -#define PIPE_CONF_CHECK_FLAGS(name, mask) \ +#define PIPE_CONF_CHECK_FLAGS(name, mask) do { \ if ((current_config->name ^ pipe_config->name) & (mask)) { \ pipe_config_err(adjust, __stringify(name), \ "(%x) (expected %i, found %i)\n", \ @@ -11197,16 +11338,18 @@ intel_pipe_config_compare(struct drm_i915_private *dev_priv, current_config->name & (mask), \ pipe_config->name & (mask)); \ ret = false; \ - } + } \ +} while (0) -#define PIPE_CONF_CHECK_CLOCK_FUZZY(name) \ +#define PIPE_CONF_CHECK_CLOCK_FUZZY(name) do { \ if (!intel_fuzzy_clock_check(current_config->name, pipe_config->name)) { \ pipe_config_err(adjust, __stringify(name), \ "(expected %i, found %i)\n", \ current_config->name, \ pipe_config->name); \ ret = false; \ - } + } \ +} while (0) #define PIPE_CONF_QUIRK(quirk) \ ((current_config->quirks | pipe_config->quirks) & (quirk)) @@ -11315,6 +11458,16 @@ intel_pipe_config_compare(struct drm_i915_private *dev_priv, PIPE_CONF_CHECK_X(dpll_hw_state.pll9); PIPE_CONF_CHECK_X(dpll_hw_state.pll10); PIPE_CONF_CHECK_X(dpll_hw_state.pcsdw12); + PIPE_CONF_CHECK_X(dpll_hw_state.mg_refclkin_ctl); + PIPE_CONF_CHECK_X(dpll_hw_state.mg_clktop2_coreclkctl1); + PIPE_CONF_CHECK_X(dpll_hw_state.mg_clktop2_hsclkctl); + PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_div0); + PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_div1); + PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_lf); + PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_frac_lock); + PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_ssc); + PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_bias); + PIPE_CONF_CHECK_X(dpll_hw_state.mg_pll_tdc_coldst_bias); PIPE_CONF_CHECK_X(dsi_pll.ctrl); PIPE_CONF_CHECK_X(dsi_pll.div); @@ -11378,6 +11531,11 @@ static void verify_wm_state(struct drm_crtc *crtc, skl_ddb_get_hw_state(dev_priv, &hw_ddb); sw_ddb = &dev_priv->wm.skl_hw.ddb; + if (INTEL_GEN(dev_priv) >= 11) + if (hw_ddb.enabled_slices != sw_ddb->enabled_slices) + DRM_ERROR("mismatch in DBUF Slices (expected %u, got %u)\n", + sw_ddb->enabled_slices, + hw_ddb.enabled_slices); /* planes */ for_each_universal_plane(dev_priv, pipe, plane) { hw_plane_wm = &hw_wm.planes[plane]; @@ -11643,11 +11801,11 @@ verify_single_dpll_state(struct drm_i915_private *dev_priv, memset(&dpll_hw_state, 0, sizeof(dpll_hw_state)); - DRM_DEBUG_KMS("%s\n", pll->name); + DRM_DEBUG_KMS("%s\n", pll->info->name); - active = pll->funcs.get_hw_state(dev_priv, pll, &dpll_hw_state); + active = pll->info->funcs->get_hw_state(dev_priv, pll, &dpll_hw_state); - if (!(pll->flags & INTEL_DPLL_ALWAYS_ON)) { + if (!(pll->info->flags & INTEL_DPLL_ALWAYS_ON)) { I915_STATE_WARN(!pll->on && pll->active_mask, "pll in active use but not on in sw tracking\n"); I915_STATE_WARN(pll->on && !pll->active_mask, @@ -12136,20 +12294,23 @@ static void intel_update_crtc(struct drm_crtc *crtc, struct intel_crtc *intel_crtc = to_intel_crtc(crtc); struct intel_crtc_state *pipe_config = to_intel_crtc_state(new_crtc_state); bool modeset = needs_modeset(new_crtc_state); + struct intel_plane_state *new_plane_state = + intel_atomic_get_new_plane_state(to_intel_atomic_state(state), + to_intel_plane(crtc->primary)); if (modeset) { update_scanline_offset(intel_crtc); dev_priv->display.crtc_enable(pipe_config, state); + + /* vblanks work again, re-enable pipe CRC. */ + intel_crtc_enable_pipe_crc(intel_crtc); } else { intel_pre_plane_update(to_intel_crtc_state(old_crtc_state), pipe_config); } - if (drm_atomic_get_existing_plane_state(state, crtc->primary)) { - intel_fbc_enable( - intel_crtc, pipe_config, - to_intel_plane_state(crtc->primary->state)); - } + if (new_plane_state) + intel_fbc_enable(intel_crtc, pipe_config, new_plane_state); drm_atomic_helper_commit_planes_on_crtc(old_crtc_state); } @@ -12181,6 +12342,8 @@ static void skl_update_crtcs(struct drm_atomic_state *state) bool progress; enum pipe pipe; int i; + u8 hw_enabled_slices = dev_priv->wm.skl_hw.ddb.enabled_slices; + u8 required_slices = intel_state->wm_results.ddb.enabled_slices; const struct skl_ddb_entry *entries[I915_MAX_PIPES] = {}; @@ -12189,6 +12352,10 @@ static void skl_update_crtcs(struct drm_atomic_state *state) if (new_crtc_state->active) entries[i] = &to_intel_crtc_state(old_crtc_state)->wm.skl.ddb; + /* If 2nd DBuf slice required, enable it here */ + if (INTEL_GEN(dev_priv) >= 11 && required_slices > hw_enabled_slices) + icl_dbuf_slices_update(dev_priv, required_slices); + /* * Whenever the number of active pipes changes, we need to make sure we * update the pipes in the right order so that their ddb allocations @@ -12239,6 +12406,10 @@ static void skl_update_crtcs(struct drm_atomic_state *state) progress = true; } } while (progress); + + /* If 2nd DBuf slice is no more required disable it */ + if (INTEL_GEN(dev_priv) >= 11 && required_slices < hw_enabled_slices) + icl_dbuf_slices_update(dev_priv, required_slices); } static void intel_atomic_helper_free_state(struct drm_i915_private *dev_priv) @@ -12320,6 +12491,13 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state) if (old_crtc_state->active) { intel_crtc_disable_planes(crtc, old_crtc_state->plane_mask); + + /* + * We need to disable pipe CRC before disabling the pipe, + * or we race against vblank off. + */ + intel_crtc_disable_pipe_crc(intel_crtc); + dev_priv->display.crtc_disable(to_intel_crtc_state(old_crtc_state), state); intel_crtc->active = false; intel_fbc_disable(intel_crtc); @@ -12695,6 +12873,15 @@ static void intel_plane_unpin_fb(struct intel_plane_state *old_plane_state) intel_unpin_fb_vma(vma, old_plane_state->flags); } +static void fb_obj_bump_render_priority(struct drm_i915_gem_object *obj) +{ + struct i915_sched_attr attr = { + .priority = I915_PRIORITY_DISPLAY, + }; + + i915_gem_object_wait_priority(obj, 0, &attr); +} + /** * intel_prepare_plane_fb - Prepare fb for usage on plane * @plane: drm plane to prepare for @@ -12723,8 +12910,8 @@ intel_prepare_plane_fb(struct drm_plane *plane, if (old_obj) { struct drm_crtc_state *crtc_state = - drm_atomic_get_existing_crtc_state(new_state->state, - plane->state->crtc); + drm_atomic_get_new_crtc_state(new_state->state, + plane->state->crtc); /* Big Hammer, we also need to ensure that any pending * MI_WAIT_FOR_EVENT inside a user batch buffer on the @@ -12771,13 +12958,15 @@ intel_prepare_plane_fb(struct drm_plane *plane, ret = intel_plane_pin_fb(to_intel_plane_state(new_state)); - i915_gem_object_wait_priority(obj, 0, I915_PRIORITY_DISPLAY); + fb_obj_bump_render_priority(obj); mutex_unlock(&dev_priv->drm.struct_mutex); i915_gem_object_unpin_pages(obj); if (ret) return ret; + intel_fb_obj_flush(obj, ORIGIN_DIRTYFB); + if (!new_state->fence) { /* implicit fencing */ struct dma_fence *fence; @@ -12822,11 +13011,13 @@ intel_cleanup_plane_fb(struct drm_plane *plane, } int -skl_max_scale(struct intel_crtc *intel_crtc, struct intel_crtc_state *crtc_state) +skl_max_scale(struct intel_crtc *intel_crtc, + struct intel_crtc_state *crtc_state, + uint32_t pixel_format) { struct drm_i915_private *dev_priv; - int max_scale; - int crtc_clock, max_dotclk; + int max_scale, mult; + int crtc_clock, max_dotclk, tmpclk1, tmpclk2; if (!intel_crtc || !crtc_state->base.enable) return DRM_PLANE_HELPER_NO_SCALING; @@ -12848,8 +13039,10 @@ skl_max_scale(struct intel_crtc *intel_crtc, struct intel_crtc_state *crtc_state * or * cdclk/crtc_clock */ - max_scale = min((1 << 16) * 3 - 1, - (1 << 8) * ((max_dotclk << 8) / crtc_clock)); + mult = pixel_format == DRM_FORMAT_NV12 ? 2 : 3; + tmpclk1 = (1 << 16) * mult - 1; + tmpclk2 = (1 << 8) * ((max_dotclk << 8) / crtc_clock); + max_scale = min(tmpclk1, tmpclk2); return max_scale; } @@ -12865,12 +13058,16 @@ intel_check_primary_plane(struct intel_plane *plane, int max_scale = DRM_PLANE_HELPER_NO_SCALING; bool can_position = false; int ret; + uint32_t pixel_format = 0; if (INTEL_GEN(dev_priv) >= 9) { /* use scaler when colorkey is not required */ if (!state->ckey.flags) { min_scale = 1; - max_scale = skl_max_scale(to_intel_crtc(crtc), crtc_state); + if (state->base.fb) + pixel_format = state->base.fb->format->format; + max_scale = skl_max_scale(to_intel_crtc(crtc), + crtc_state, pixel_format); } can_position = true; } @@ -12943,10 +13140,25 @@ out: intel_cstate); } +void intel_crtc_arm_fifo_underrun(struct intel_crtc *crtc, + struct intel_crtc_state *crtc_state) +{ + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + + if (!IS_GEN2(dev_priv)) + intel_set_cpu_fifo_underrun_reporting(dev_priv, crtc->pipe, true); + + if (crtc_state->has_pch_encoder) { + enum pipe pch_transcoder = + intel_crtc_pch_transcoder(crtc); + + intel_set_pch_fifo_underrun_reporting(dev_priv, pch_transcoder, true); + } +} + static void intel_finish_crtc_commit(struct drm_crtc *crtc, struct drm_crtc_state *old_crtc_state) { - struct drm_i915_private *dev_priv = to_i915(crtc->dev); struct intel_crtc *intel_crtc = to_intel_crtc(crtc); struct intel_atomic_state *old_intel_state = to_intel_atomic_state(old_crtc_state->state); @@ -12957,17 +13169,8 @@ static void intel_finish_crtc_commit(struct drm_crtc *crtc, if (new_crtc_state->update_pipe && !needs_modeset(&new_crtc_state->base) && - old_crtc_state->mode.private_flags & I915_MODE_FLAG_INHERITED) { - if (!IS_GEN2(dev_priv)) - intel_set_cpu_fifo_underrun_reporting(dev_priv, intel_crtc->pipe, true); - - if (new_crtc_state->has_pch_encoder) { - enum pipe pch_transcoder = - intel_crtc_pch_transcoder(intel_crtc); - - intel_set_pch_fifo_underrun_reporting(dev_priv, pch_transcoder, true); - } - } + old_crtc_state->mode.private_flags & I915_MODE_FLAG_INHERITED) + intel_crtc_arm_fifo_underrun(intel_crtc, new_crtc_state); } /** @@ -13031,6 +13234,7 @@ static bool skl_mod_supported(uint32_t format, uint64_t modifier) case DRM_FORMAT_YVYU: case DRM_FORMAT_UYVY: case DRM_FORMAT_VYUY: + case DRM_FORMAT_NV12: if (modifier == I915_FORMAT_MOD_Yf_TILED) return true; /* fall through */ @@ -13165,8 +13369,9 @@ intel_legacy_cursor_update(struct drm_plane *plane, if (ret) goto out_unlock; - old_fb = old_plane_state->fb; + intel_fb_obj_flush(intel_fb_obj(fb), ORIGIN_FLIP); + old_fb = old_plane_state->fb; i915_gem_track_fb(intel_fb_obj(old_fb), intel_fb_obj(fb), intel_plane->frontbuffer_bit); @@ -13237,6 +13442,30 @@ static bool skl_plane_has_fbc(struct drm_i915_private *dev_priv, return pipe == PIPE_A && plane_id == PLANE_PRIMARY; } +bool skl_plane_has_planar(struct drm_i915_private *dev_priv, + enum pipe pipe, enum plane_id plane_id) +{ + if (plane_id == PLANE_PRIMARY) { + if (IS_SKYLAKE(dev_priv) || IS_BROXTON(dev_priv)) + return false; + else if ((INTEL_GEN(dev_priv) == 9 && pipe == PIPE_C) && + !IS_GEMINILAKE(dev_priv)) + return false; + } else if (plane_id >= PLANE_SPRITE0) { + if (plane_id == PLANE_CURSOR) + return false; + if (IS_GEMINILAKE(dev_priv) || INTEL_GEN(dev_priv) == 10) { + if (plane_id != PLANE_SPRITE0) + return false; + } else { + if (plane_id != PLANE_SPRITE0 || pipe == PIPE_C || + IS_SKYLAKE(dev_priv) || IS_BROXTON(dev_priv)) + return false; + } + } + return true; +} + static struct intel_plane * intel_primary_plane_create(struct drm_i915_private *dev_priv, enum pipe pipe) { @@ -13297,8 +13526,13 @@ intel_primary_plane_create(struct drm_i915_private *dev_priv, enum pipe pipe) primary->check_plane = intel_check_primary_plane; if (INTEL_GEN(dev_priv) >= 9) { - intel_primary_formats = skl_primary_formats; - num_formats = ARRAY_SIZE(skl_primary_formats); + if (skl_plane_has_planar(dev_priv, pipe, PLANE_PRIMARY)) { + intel_primary_formats = skl_pri_planar_formats; + num_formats = ARRAY_SIZE(skl_pri_planar_formats); + } else { + intel_primary_formats = skl_primary_formats; + num_formats = ARRAY_SIZE(skl_primary_formats); + } if (skl_plane_has_ccs(dev_priv, pipe, PLANE_PRIMARY)) modifiers = skl_format_modifiers_ccs; @@ -13553,10 +13787,17 @@ static int intel_crtc_init(struct drm_i915_private *dev_priv, enum pipe pipe) /* initialize shared scalers */ intel_crtc_init_scalers(intel_crtc, crtc_state); - BUG_ON(pipe >= ARRAY_SIZE(dev_priv->plane_to_crtc_mapping) || - dev_priv->plane_to_crtc_mapping[primary->i9xx_plane] != NULL); - dev_priv->plane_to_crtc_mapping[primary->i9xx_plane] = intel_crtc; - dev_priv->pipe_to_crtc_mapping[intel_crtc->pipe] = intel_crtc; + BUG_ON(pipe >= ARRAY_SIZE(dev_priv->pipe_to_crtc_mapping) || + dev_priv->pipe_to_crtc_mapping[pipe] != NULL); + dev_priv->pipe_to_crtc_mapping[pipe] = intel_crtc; + + if (INTEL_GEN(dev_priv) < 9) { + enum i9xx_plane_id i9xx_plane = primary->i9xx_plane; + + BUG_ON(i9xx_plane >= ARRAY_SIZE(dev_priv->plane_to_crtc_mapping) || + dev_priv->plane_to_crtc_mapping[i9xx_plane] != NULL); + dev_priv->plane_to_crtc_mapping[i9xx_plane] = intel_crtc; + } drm_crtc_helper_add(&intel_crtc->base, &intel_helper_funcs); @@ -14112,6 +14353,20 @@ static int intel_framebuffer_init(struct intel_framebuffer *intel_fb, goto err; } break; + case DRM_FORMAT_NV12: + if (mode_cmd->modifier[0] == I915_FORMAT_MOD_Y_TILED_CCS || + mode_cmd->modifier[0] == I915_FORMAT_MOD_Yf_TILED_CCS) { + DRM_DEBUG_KMS("RC not to be enabled with NV12\n"); + goto err; + } + if (INTEL_GEN(dev_priv) < 9 || IS_SKYLAKE(dev_priv) || + IS_BROXTON(dev_priv)) { + DRM_DEBUG_KMS("unsupported pixel format: %s\n", + drm_get_format_name(mode_cmd->pixel_format, + &format_name)); + goto err; + } + break; default: DRM_DEBUG_KMS("unsupported pixel format: %s\n", drm_get_format_name(mode_cmd->pixel_format, &format_name)); @@ -14124,6 +14379,14 @@ static int intel_framebuffer_init(struct intel_framebuffer *intel_fb, drm_helper_mode_fill_fb_struct(&dev_priv->drm, fb, mode_cmd); + if (fb->format->format == DRM_FORMAT_NV12 && + (fb->width < SKL_MIN_YUV_420_SRC_W || + fb->height < SKL_MIN_YUV_420_SRC_H || + (fb->width % 4) != 0 || (fb->height % 4) != 0)) { + DRM_DEBUG_KMS("src dimensions not correct for NV12\n"); + return -EINVAL; + } + for (i = 0; i < fb->format->num_planes; i++) { u32 stride_alignment; @@ -15101,8 +15364,8 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) for (i = 0; i < dev_priv->num_shared_dpll; i++) { struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i]; - pll->on = pll->funcs.get_hw_state(dev_priv, pll, - &pll->state.hw_state); + pll->on = pll->info->funcs->get_hw_state(dev_priv, pll, + &pll->state.hw_state); pll->state.crtc_mask = 0; for_each_intel_crtc(dev, crtc) { struct intel_crtc_state *crtc_state = @@ -15115,7 +15378,7 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) pll->active_mask = pll->state.crtc_mask; DRM_DEBUG_KMS("%s hw state readout: crtc_mask 0x%08x, on %i\n", - pll->name, pll->state.crtc_mask, pll->on); + pll->info->name, pll->state.crtc_mask, pll->on); } for_each_intel_encoder(dev, encoder) { @@ -15291,9 +15554,10 @@ intel_modeset_setup_hw_state(struct drm_device *dev, if (!pll->on || pll->active_mask) continue; - DRM_DEBUG_KMS("%s enabled but not in use, disabling\n", pll->name); + DRM_DEBUG_KMS("%s enabled but not in use, disabling\n", + pll->info->name); - pll->funcs.disable(dev_priv, pll); + pll->info->funcs->disable(dev_priv, pll); pll->on = false; } diff --git a/drivers/gpu/drm/i915/intel_display.h b/drivers/gpu/drm/i915/intel_display.h index 4e7418b..2ef3161 100644 --- a/drivers/gpu/drm/i915/intel_display.h +++ b/drivers/gpu/drm/i915/intel_display.h @@ -218,6 +218,10 @@ struct intel_link_m_n { for ((__p) = 0; (__p) < INTEL_INFO(__dev_priv)->num_pipes; (__p)++) \ for_each_if((__mask) & BIT(__p)) +#define for_each_cpu_transcoder_masked(__dev_priv, __t, __mask) \ + for ((__t) = 0; (__t) < I915_MAX_TRANSCODERS; (__t)++) \ + for_each_if ((__mask) & (1 << (__t))) + #define for_each_universal_plane(__dev_priv, __pipe, __p) \ for ((__p) = 0; \ (__p) < INTEL_INFO(__dev_priv)->num_sprites[(__pipe)] + 1; \ diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index b7b4cfd..dde92e4 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -43,7 +43,6 @@ #include <drm/i915_drm.h> #include "i915_drv.h" -#define DP_LINK_CHECK_TIMEOUT (10 * 1000) #define DP_DPRX_ESI_LEN 14 /* Compliance test status bits */ @@ -92,8 +91,6 @@ static const struct dp_link_dpll chv_dpll[] = { { .p1 = 4, .p2 = 2, .n = 1, .m1 = 2, .m2 = 0x819999a } }, { 270000, /* m2_int = 27, m2_fraction = 0 */ { .p1 = 4, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x6c00000 } }, - { 540000, /* m2_int = 27, m2_fraction = 0 */ - { .p1 = 2, .p2 = 1, .n = 1, .m1 = 2, .m2 = 0x6c00000 } } }; /** @@ -1650,9 +1647,17 @@ void intel_dp_compute_rate(struct intel_dp *intel_dp, int port_clock, } } +struct link_config_limits { + int min_clock, max_clock; + int min_lane_count, max_lane_count; + int min_bpp, max_bpp; +}; + static int intel_dp_compute_bpp(struct intel_dp *intel_dp, struct intel_crtc_state *pipe_config) { + struct drm_i915_private *dev_priv = to_i915(intel_dp_to_dev(intel_dp)); + struct intel_connector *intel_connector = intel_dp->attached_connector; int bpp, bpc; bpp = pipe_config->pipe_bpp; @@ -1661,13 +1666,16 @@ static int intel_dp_compute_bpp(struct intel_dp *intel_dp, if (bpc > 0) bpp = min(bpp, 3*bpc); - /* For DP Compliance we override the computed bpp for the pipe */ - if (intel_dp->compliance.test_data.bpc != 0) { - pipe_config->pipe_bpp = 3*intel_dp->compliance.test_data.bpc; - pipe_config->dither_force_disable = pipe_config->pipe_bpp == 6*3; - DRM_DEBUG_KMS("Setting pipe_bpp to %d\n", - pipe_config->pipe_bpp); + if (intel_dp_is_edp(intel_dp)) { + /* Get bpp from vbt only for panels that dont have bpp in edid */ + if (intel_connector->base.display_info.bpc == 0 && + dev_priv->vbt.edp.bpp && dev_priv->vbt.edp.bpp < bpp) { + DRM_DEBUG_KMS("clamping bpp for eDP panel to BIOS-provided %i\n", + dev_priv->vbt.edp.bpp); + bpp = dev_priv->vbt.edp.bpp; + } } + return bpp; } @@ -1688,6 +1696,142 @@ static bool intel_edp_compare_alt_mode(struct drm_display_mode *m1, return bres; } +/* Adjust link config limits based on compliance test requests. */ +static void +intel_dp_adjust_compliance_config(struct intel_dp *intel_dp, + struct intel_crtc_state *pipe_config, + struct link_config_limits *limits) +{ + /* For DP Compliance we override the computed bpp for the pipe */ + if (intel_dp->compliance.test_data.bpc != 0) { + int bpp = 3 * intel_dp->compliance.test_data.bpc; + + limits->min_bpp = limits->max_bpp = bpp; + pipe_config->dither_force_disable = bpp == 6 * 3; + + DRM_DEBUG_KMS("Setting pipe_bpp to %d\n", bpp); + } + + /* Use values requested by Compliance Test Request */ + if (intel_dp->compliance.test_type == DP_TEST_LINK_TRAINING) { + int index; + + /* Validate the compliance test data since max values + * might have changed due to link train fallback. + */ + if (intel_dp_link_params_valid(intel_dp, intel_dp->compliance.test_link_rate, + intel_dp->compliance.test_lane_count)) { + index = intel_dp_rate_index(intel_dp->common_rates, + intel_dp->num_common_rates, + intel_dp->compliance.test_link_rate); + if (index >= 0) + limits->min_clock = limits->max_clock = index; + limits->min_lane_count = limits->max_lane_count = + intel_dp->compliance.test_lane_count; + } + } +} + +/* Optimize link config in order: max bpp, min clock, min lanes */ +static bool +intel_dp_compute_link_config_wide(struct intel_dp *intel_dp, + struct intel_crtc_state *pipe_config, + const struct link_config_limits *limits) +{ + struct drm_display_mode *adjusted_mode = &pipe_config->base.adjusted_mode; + int bpp, clock, lane_count; + int mode_rate, link_clock, link_avail; + + for (bpp = limits->max_bpp; bpp >= limits->min_bpp; bpp -= 2 * 3) { + mode_rate = intel_dp_link_required(adjusted_mode->crtc_clock, + bpp); + + for (clock = limits->min_clock; clock <= limits->max_clock; clock++) { + for (lane_count = limits->min_lane_count; + lane_count <= limits->max_lane_count; + lane_count <<= 1) { + link_clock = intel_dp->common_rates[clock]; + link_avail = intel_dp_max_data_rate(link_clock, + lane_count); + + if (mode_rate <= link_avail) { + pipe_config->lane_count = lane_count; + pipe_config->pipe_bpp = bpp; + pipe_config->port_clock = link_clock; + + return true; + } + } + } + } + + return false; +} + +static bool +intel_dp_compute_link_config(struct intel_encoder *encoder, + struct intel_crtc_state *pipe_config) +{ + struct drm_display_mode *adjusted_mode = &pipe_config->base.adjusted_mode; + struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base); + struct link_config_limits limits; + int common_len; + + common_len = intel_dp_common_len_rate_limit(intel_dp, + intel_dp->max_link_rate); + + /* No common link rates between source and sink */ + WARN_ON(common_len <= 0); + + limits.min_clock = 0; + limits.max_clock = common_len - 1; + + limits.min_lane_count = 1; + limits.max_lane_count = intel_dp_max_lane_count(intel_dp); + + limits.min_bpp = 6 * 3; + limits.max_bpp = intel_dp_compute_bpp(intel_dp, pipe_config); + + if (intel_dp_is_edp(intel_dp)) { + /* + * Use the maximum clock and number of lanes the eDP panel + * advertizes being capable of. The panels are generally + * designed to support only a single clock and lane + * configuration, and typically these values correspond to the + * native resolution of the panel. + */ + limits.min_lane_count = limits.max_lane_count; + limits.min_clock = limits.max_clock; + } + + intel_dp_adjust_compliance_config(intel_dp, pipe_config, &limits); + + DRM_DEBUG_KMS("DP link computation with max lane count %i " + "max rate %d max bpp %d pixel clock %iKHz\n", + limits.max_lane_count, + intel_dp->common_rates[limits.max_clock], + limits.max_bpp, adjusted_mode->crtc_clock); + + /* + * Optimize for slow and wide. This is the place to add alternative + * optimization policy. + */ + if (!intel_dp_compute_link_config_wide(intel_dp, pipe_config, &limits)) + return false; + + DRM_DEBUG_KMS("DP lane count %d clock %d bpp %d\n", + pipe_config->lane_count, pipe_config->port_clock, + pipe_config->pipe_bpp); + + DRM_DEBUG_KMS("DP link rate required %i available %i\n", + intel_dp_link_required(adjusted_mode->crtc_clock, + pipe_config->pipe_bpp), + intel_dp_max_data_rate(pipe_config->port_clock, + pipe_config->lane_count)); + + return true; +} + bool intel_dp_compute_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config, @@ -1701,27 +1845,9 @@ intel_dp_compute_config(struct intel_encoder *encoder, struct intel_connector *intel_connector = intel_dp->attached_connector; struct intel_digital_connector_state *intel_conn_state = to_intel_digital_connector_state(conn_state); - int lane_count, clock; - int min_lane_count = 1; - int max_lane_count = intel_dp_max_lane_count(intel_dp); - /* Conveniently, the link BW constants become indices with a shift...*/ - int min_clock = 0; - int max_clock; - int bpp, mode_rate; - int link_avail, link_clock; - int common_len; - uint8_t link_bw, rate_select; bool reduce_m_n = drm_dp_has_quirk(&intel_dp->desc, DP_DPCD_QUIRK_LIMITED_M_N); - common_len = intel_dp_common_len_rate_limit(intel_dp, - intel_dp->max_link_rate); - - /* No common link rates between source and sink */ - WARN_ON(common_len <= 0); - - max_clock = common_len - 1; - if (HAS_PCH_SPLIT(dev_priv) && !HAS_DDI(dev_priv) && port != PORT_A) pipe_config->has_pch_encoder = true; @@ -1747,6 +1873,7 @@ intel_dp_compute_config(struct intel_encoder *encoder, if (INTEL_GEN(dev_priv) >= 9) { int ret; + ret = skl_update_scaler_crtc(pipe_config); if (ret) return ret; @@ -1767,75 +1894,9 @@ intel_dp_compute_config(struct intel_encoder *encoder, if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK) return false; - /* Use values requested by Compliance Test Request */ - if (intel_dp->compliance.test_type == DP_TEST_LINK_TRAINING) { - int index; - - /* Validate the compliance test data since max values - * might have changed due to link train fallback. - */ - if (intel_dp_link_params_valid(intel_dp, intel_dp->compliance.test_link_rate, - intel_dp->compliance.test_lane_count)) { - index = intel_dp_rate_index(intel_dp->common_rates, - intel_dp->num_common_rates, - intel_dp->compliance.test_link_rate); - if (index >= 0) - min_clock = max_clock = index; - min_lane_count = max_lane_count = intel_dp->compliance.test_lane_count; - } - } - DRM_DEBUG_KMS("DP link computation with max lane count %i " - "max bw %d pixel clock %iKHz\n", - max_lane_count, intel_dp->common_rates[max_clock], - adjusted_mode->crtc_clock); - - /* Walk through all bpp values. Luckily they're all nicely spaced with 2 - * bpc in between. */ - bpp = intel_dp_compute_bpp(intel_dp, pipe_config); - if (intel_dp_is_edp(intel_dp)) { - - /* Get bpp from vbt only for panels that dont have bpp in edid */ - if (intel_connector->base.display_info.bpc == 0 && - (dev_priv->vbt.edp.bpp && dev_priv->vbt.edp.bpp < bpp)) { - DRM_DEBUG_KMS("clamping bpp for eDP panel to BIOS-provided %i\n", - dev_priv->vbt.edp.bpp); - bpp = dev_priv->vbt.edp.bpp; - } - - /* - * Use the maximum clock and number of lanes the eDP panel - * advertizes being capable of. The panels are generally - * designed to support only a single clock and lane - * configuration, and typically these values correspond to the - * native resolution of the panel. - */ - min_lane_count = max_lane_count; - min_clock = max_clock; - } - - for (; bpp >= 6*3; bpp -= 2*3) { - mode_rate = intel_dp_link_required(adjusted_mode->crtc_clock, - bpp); - - for (clock = min_clock; clock <= max_clock; clock++) { - for (lane_count = min_lane_count; - lane_count <= max_lane_count; - lane_count <<= 1) { - - link_clock = intel_dp->common_rates[clock]; - link_avail = intel_dp_max_data_rate(link_clock, - lane_count); - - if (mode_rate <= link_avail) { - goto found; - } - } - } - } - - return false; + if (!intel_dp_compute_link_config(encoder, pipe_config)) + return false; -found: if (intel_conn_state->broadcast_rgb == INTEL_BROADCAST_RGB_AUTO) { /* * See: @@ -1843,7 +1904,7 @@ found: * VESA DisplayPort Ver.1.2a - 5.1.1.1 Video Colorimetry */ pipe_config->limited_color_range = - bpp != 18 && + pipe_config->pipe_bpp != 18 && drm_default_rgb_quant_range(adjusted_mode) == HDMI_QUANTIZATION_RANGE_LIMITED; } else { @@ -1851,21 +1912,7 @@ found: intel_conn_state->broadcast_rgb == INTEL_BROADCAST_RGB_LIMITED; } - pipe_config->lane_count = lane_count; - - pipe_config->pipe_bpp = bpp; - pipe_config->port_clock = intel_dp->common_rates[clock]; - - intel_dp_compute_rate(intel_dp, pipe_config->port_clock, - &link_bw, &rate_select); - - DRM_DEBUG_KMS("DP link bw %02x rate select %02x lane count %d clock %d bpp %d\n", - link_bw, rate_select, pipe_config->lane_count, - pipe_config->port_clock, bpp); - DRM_DEBUG_KMS("DP link bw required %i available %i\n", - mode_rate, link_avail); - - intel_link_compute_m_n(bpp, lane_count, + intel_link_compute_m_n(pipe_config->pipe_bpp, pipe_config->lane_count, adjusted_mode->crtc_clock, pipe_config->port_clock, &pipe_config->dp_m_n, @@ -1874,11 +1921,12 @@ found: if (intel_connector->panel.downclock_mode != NULL && dev_priv->drrs.type == SEAMLESS_DRRS_SUPPORT) { pipe_config->has_drrs = true; - intel_link_compute_m_n(bpp, lane_count, - intel_connector->panel.downclock_mode->clock, - pipe_config->port_clock, - &pipe_config->dp_m2_n2, - reduce_m_n); + intel_link_compute_m_n(pipe_config->pipe_bpp, + pipe_config->lane_count, + intel_connector->panel.downclock_mode->clock, + pipe_config->port_clock, + &pipe_config->dp_m2_n2, + reduce_m_n); } if (!HAS_DDI(dev_priv)) @@ -2881,10 +2929,7 @@ _intel_dp_set_link_train(struct intel_dp *intel_dp, } } else { - if (IS_CHERRYVIEW(dev_priv)) - *DP &= ~DP_LINK_TRAIN_MASK_CHV; - else - *DP &= ~DP_LINK_TRAIN_MASK; + *DP &= ~DP_LINK_TRAIN_MASK; switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) { case DP_TRAINING_PATTERN_DISABLE: @@ -2897,12 +2942,8 @@ _intel_dp_set_link_train(struct intel_dp *intel_dp, *DP |= DP_LINK_TRAIN_PAT_2; break; case DP_TRAINING_PATTERN_3: - if (IS_CHERRYVIEW(dev_priv)) { - *DP |= DP_LINK_TRAIN_PAT_3_CHV; - } else { - DRM_DEBUG_KMS("TPS3 not supported, using TPS2 instead\n"); - *DP |= DP_LINK_TRAIN_PAT_2; - } + DRM_DEBUG_KMS("TPS3 not supported, using TPS2 instead\n"); + *DP |= DP_LINK_TRAIN_PAT_2; break; } } @@ -3641,10 +3682,7 @@ intel_dp_link_down(struct intel_encoder *encoder, DP &= ~DP_LINK_TRAIN_MASK_CPT; DP |= DP_LINK_TRAIN_PAT_IDLE_CPT; } else { - if (IS_CHERRYVIEW(dev_priv)) - DP &= ~DP_LINK_TRAIN_MASK_CHV; - else - DP &= ~DP_LINK_TRAIN_MASK; + DP &= ~DP_LINK_TRAIN_MASK; DP |= DP_LINK_TRAIN_PAT_IDLE; } I915_WRITE(intel_dp->output_reg, DP); diff --git a/drivers/gpu/drm/i915/intel_dp_link_training.c b/drivers/gpu/drm/i915/intel_dp_link_training.c index f59b59b..3fcaa98 100644 --- a/drivers/gpu/drm/i915/intel_dp_link_training.c +++ b/drivers/gpu/drm/i915/intel_dp_link_training.c @@ -139,6 +139,11 @@ intel_dp_link_training_clock_recovery(struct intel_dp *intel_dp) intel_dp_compute_rate(intel_dp, intel_dp->link_rate, &link_bw, &rate_select); + if (link_bw) + DRM_DEBUG_KMS("Using LINK_BW_SET value %02x\n", link_bw); + else + DRM_DEBUG_KMS("Using LINK_RATE_SET value %02x\n", rate_select); + /* Write the link configuration data */ link_config[0] = link_bw; link_config[1] = intel_dp->lane_count; diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c index c3de091..9e6956c 100644 --- a/drivers/gpu/drm/i915/intel_dp_mst.c +++ b/drivers/gpu/drm/i915/intel_dp_mst.c @@ -180,9 +180,11 @@ static void intel_mst_post_disable_dp(struct intel_encoder *encoder, intel_dp->active_mst_links--; intel_mst->connector = NULL; - if (intel_dp->active_mst_links == 0) + if (intel_dp->active_mst_links == 0) { + intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_OFF); intel_dig_port->base.post_disable(&intel_dig_port->base, old_crtc_state, NULL); + } DRM_DEBUG_KMS("active links %d\n", intel_dp->active_mst_links); } @@ -223,7 +225,11 @@ static void intel_mst_pre_enable_dp(struct intel_encoder *encoder, DRM_DEBUG_KMS("active links %d\n", intel_dp->active_mst_links); + if (intel_dp->active_mst_links == 0) + intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON); + drm_dp_send_power_updown_phy(&intel_dp->mst_mgr, connector->port, true); + if (intel_dp->active_mst_links == 0) intel_dig_port->base.pre_enable(&intel_dig_port->base, pipe_config, NULL); diff --git a/drivers/gpu/drm/i915/intel_dpio_phy.c b/drivers/gpu/drm/i915/intel_dpio_phy.c index c8e9e44..00b3ab6 100644 --- a/drivers/gpu/drm/i915/intel_dpio_phy.c +++ b/drivers/gpu/drm/i915/intel_dpio_phy.c @@ -380,13 +380,14 @@ static void _bxt_ddi_phy_init(struct drm_i915_private *dev_priv, * all 1s. Eventually they become accessible as they power up, then * the reserved bit will give the default 0. Poll on the reserved bit * becoming 0 to find when the PHY is accessible. - * HW team confirmed that the time to reach phypowergood status is - * anywhere between 50 us and 100us. + * The flag should get set in 100us according to the HW team, but + * use 1ms due to occasional timeouts observed with that. */ - if (wait_for_us(((I915_READ(BXT_PORT_CL1CM_DW0(phy)) & - (PHY_RESERVED | PHY_POWER_GOOD)) == PHY_POWER_GOOD), 100)) { + if (intel_wait_for_register_fw(dev_priv, BXT_PORT_CL1CM_DW0(phy), + PHY_RESERVED | PHY_POWER_GOOD, + PHY_POWER_GOOD, + 1)) DRM_ERROR("timeout during PHY%d power on\n", phy); - } /* Program PLL Rcomp code offset */ val = I915_READ(BXT_PORT_CL1CM_DW9(phy)); diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.c b/drivers/gpu/drm/i915/intel_dpll_mgr.c index 51c5ae4..383fbc1 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.c +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.c @@ -118,10 +118,10 @@ void assert_shared_dpll(struct drm_i915_private *dev_priv, if (WARN(!pll, "asserting DPLL %s with no DPLL\n", onoff(state))) return; - cur_state = pll->funcs.get_hw_state(dev_priv, pll, &hw_state); + cur_state = pll->info->funcs->get_hw_state(dev_priv, pll, &hw_state); I915_STATE_WARN(cur_state != state, "%s assertion failure (expected %s, current %s)\n", - pll->name, onoff(state), onoff(cur_state)); + pll->info->name, onoff(state), onoff(cur_state)); } /** @@ -143,11 +143,11 @@ void intel_prepare_shared_dpll(struct intel_crtc *crtc) mutex_lock(&dev_priv->dpll_lock); WARN_ON(!pll->state.crtc_mask); if (!pll->active_mask) { - DRM_DEBUG_DRIVER("setting up %s\n", pll->name); + DRM_DEBUG_DRIVER("setting up %s\n", pll->info->name); WARN_ON(pll->on); assert_shared_dpll_disabled(dev_priv, pll); - pll->funcs.prepare(dev_priv, pll); + pll->info->funcs->prepare(dev_priv, pll); } mutex_unlock(&dev_priv->dpll_lock); } @@ -179,7 +179,7 @@ void intel_enable_shared_dpll(struct intel_crtc *crtc) pll->active_mask |= crtc_mask; DRM_DEBUG_KMS("enable %s (active %x, on? %d) for crtc %d\n", - pll->name, pll->active_mask, pll->on, + pll->info->name, pll->active_mask, pll->on, crtc->base.base.id); if (old_mask) { @@ -189,8 +189,8 @@ void intel_enable_shared_dpll(struct intel_crtc *crtc) } WARN_ON(pll->on); - DRM_DEBUG_KMS("enabling %s\n", pll->name); - pll->funcs.enable(dev_priv, pll); + DRM_DEBUG_KMS("enabling %s\n", pll->info->name); + pll->info->funcs->enable(dev_priv, pll); pll->on = true; out: @@ -221,7 +221,7 @@ void intel_disable_shared_dpll(struct intel_crtc *crtc) goto out; DRM_DEBUG_KMS("disable %s (active %x, on? %d) for crtc %d\n", - pll->name, pll->active_mask, pll->on, + pll->info->name, pll->active_mask, pll->on, crtc->base.base.id); assert_shared_dpll_enabled(dev_priv, pll); @@ -231,8 +231,8 @@ void intel_disable_shared_dpll(struct intel_crtc *crtc) if (pll->active_mask) goto out; - DRM_DEBUG_KMS("disabling %s\n", pll->name); - pll->funcs.disable(dev_priv, pll); + DRM_DEBUG_KMS("disabling %s\n", pll->info->name); + pll->info->funcs->disable(dev_priv, pll); pll->on = false; out: @@ -263,7 +263,8 @@ intel_find_shared_dpll(struct intel_crtc *crtc, &shared_dpll[i].hw_state, sizeof(crtc_state->dpll_hw_state)) == 0) { DRM_DEBUG_KMS("[CRTC:%d:%s] sharing existing %s (crtc mask 0x%08x, active %x)\n", - crtc->base.base.id, crtc->base.name, pll->name, + crtc->base.base.id, crtc->base.name, + pll->info->name, shared_dpll[i].crtc_mask, pll->active_mask); return pll; @@ -275,7 +276,8 @@ intel_find_shared_dpll(struct intel_crtc *crtc, pll = &dev_priv->shared_dplls[i]; if (shared_dpll[i].crtc_mask == 0) { DRM_DEBUG_KMS("[CRTC:%d:%s] allocated %s\n", - crtc->base.base.id, crtc->base.name, pll->name); + crtc->base.base.id, crtc->base.name, + pll->info->name); return pll; } } @@ -289,19 +291,19 @@ intel_reference_shared_dpll(struct intel_shared_dpll *pll, { struct intel_shared_dpll_state *shared_dpll; struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); - enum intel_dpll_id i = pll->id; + const enum intel_dpll_id id = pll->info->id; shared_dpll = intel_atomic_get_shared_dpll_state(crtc_state->base.state); - if (shared_dpll[i].crtc_mask == 0) - shared_dpll[i].hw_state = + if (shared_dpll[id].crtc_mask == 0) + shared_dpll[id].hw_state = crtc_state->dpll_hw_state; crtc_state->shared_dpll = pll; - DRM_DEBUG_DRIVER("using %s for pipe %c\n", pll->name, + DRM_DEBUG_DRIVER("using %s for pipe %c\n", pll->info->name, pipe_name(crtc->pipe)); - shared_dpll[pll->id].crtc_mask |= 1 << crtc->pipe; + shared_dpll[id].crtc_mask |= 1 << crtc->pipe; } /** @@ -341,15 +343,16 @@ static bool ibx_pch_dpll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll, struct intel_dpll_hw_state *hw_state) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) return false; - val = I915_READ(PCH_DPLL(pll->id)); + val = I915_READ(PCH_DPLL(id)); hw_state->dpll = val; - hw_state->fp0 = I915_READ(PCH_FP0(pll->id)); - hw_state->fp1 = I915_READ(PCH_FP1(pll->id)); + hw_state->fp0 = I915_READ(PCH_FP0(id)); + hw_state->fp1 = I915_READ(PCH_FP1(id)); intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); @@ -359,8 +362,10 @@ static bool ibx_pch_dpll_get_hw_state(struct drm_i915_private *dev_priv, static void ibx_pch_dpll_prepare(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { - I915_WRITE(PCH_FP0(pll->id), pll->state.hw_state.fp0); - I915_WRITE(PCH_FP1(pll->id), pll->state.hw_state.fp1); + const enum intel_dpll_id id = pll->info->id; + + I915_WRITE(PCH_FP0(id), pll->state.hw_state.fp0); + I915_WRITE(PCH_FP1(id), pll->state.hw_state.fp1); } static void ibx_assert_pch_refclk_enabled(struct drm_i915_private *dev_priv) @@ -379,13 +384,15 @@ static void ibx_assert_pch_refclk_enabled(struct drm_i915_private *dev_priv) static void ibx_pch_dpll_enable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { + const enum intel_dpll_id id = pll->info->id; + /* PCH refclock must be enabled first */ ibx_assert_pch_refclk_enabled(dev_priv); - I915_WRITE(PCH_DPLL(pll->id), pll->state.hw_state.dpll); + I915_WRITE(PCH_DPLL(id), pll->state.hw_state.dpll); /* Wait for the clocks to stabilize. */ - POSTING_READ(PCH_DPLL(pll->id)); + POSTING_READ(PCH_DPLL(id)); udelay(150); /* The pixel multiplier can only be updated once the @@ -393,14 +400,15 @@ static void ibx_pch_dpll_enable(struct drm_i915_private *dev_priv, * * So write it again. */ - I915_WRITE(PCH_DPLL(pll->id), pll->state.hw_state.dpll); - POSTING_READ(PCH_DPLL(pll->id)); + I915_WRITE(PCH_DPLL(id), pll->state.hw_state.dpll); + POSTING_READ(PCH_DPLL(id)); udelay(200); } static void ibx_pch_dpll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { + const enum intel_dpll_id id = pll->info->id; struct drm_device *dev = &dev_priv->drm; struct intel_crtc *crtc; @@ -410,8 +418,8 @@ static void ibx_pch_dpll_disable(struct drm_i915_private *dev_priv, assert_pch_transcoder_disabled(dev_priv, crtc->pipe); } - I915_WRITE(PCH_DPLL(pll->id), 0); - POSTING_READ(PCH_DPLL(pll->id)); + I915_WRITE(PCH_DPLL(id), 0); + POSTING_READ(PCH_DPLL(id)); udelay(200); } @@ -429,7 +437,8 @@ ibx_get_dpll(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state, pll = &dev_priv->shared_dplls[i]; DRM_DEBUG_KMS("[CRTC:%d:%s] using pre-allocated %s\n", - crtc->base.base.id, crtc->base.name, pll->name); + crtc->base.base.id, crtc->base.name, + pll->info->name); } else { pll = intel_find_shared_dpll(crtc, crtc_state, DPLL_ID_PCH_PLL_A, @@ -466,8 +475,10 @@ static const struct intel_shared_dpll_funcs ibx_pch_dpll_funcs = { static void hsw_ddi_wrpll_enable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { - I915_WRITE(WRPLL_CTL(pll->id), pll->state.hw_state.wrpll); - POSTING_READ(WRPLL_CTL(pll->id)); + const enum intel_dpll_id id = pll->info->id; + + I915_WRITE(WRPLL_CTL(id), pll->state.hw_state.wrpll); + POSTING_READ(WRPLL_CTL(id)); udelay(20); } @@ -482,11 +493,12 @@ static void hsw_ddi_spll_enable(struct drm_i915_private *dev_priv, static void hsw_ddi_wrpll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; - val = I915_READ(WRPLL_CTL(pll->id)); - I915_WRITE(WRPLL_CTL(pll->id), val & ~WRPLL_PLL_ENABLE); - POSTING_READ(WRPLL_CTL(pll->id)); + val = I915_READ(WRPLL_CTL(id)); + I915_WRITE(WRPLL_CTL(id), val & ~WRPLL_PLL_ENABLE); + POSTING_READ(WRPLL_CTL(id)); } static void hsw_ddi_spll_disable(struct drm_i915_private *dev_priv, @@ -503,12 +515,13 @@ static bool hsw_ddi_wrpll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll, struct intel_dpll_hw_state *hw_state) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) return false; - val = I915_READ(WRPLL_CTL(pll->id)); + val = I915_READ(WRPLL_CTL(id)); hw_state->wrpll = val; intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); @@ -914,13 +927,15 @@ static const struct skl_dpll_regs skl_dpll_regs[4] = { static void skl_ddi_pll_write_ctrl1(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; val = I915_READ(DPLL_CTRL1); - val &= ~(DPLL_CTRL1_HDMI_MODE(pll->id) | DPLL_CTRL1_SSC(pll->id) | - DPLL_CTRL1_LINK_RATE_MASK(pll->id)); - val |= pll->state.hw_state.ctrl1 << (pll->id * 6); + val &= ~(DPLL_CTRL1_HDMI_MODE(id) | + DPLL_CTRL1_SSC(id) | + DPLL_CTRL1_LINK_RATE_MASK(id)); + val |= pll->state.hw_state.ctrl1 << (id * 6); I915_WRITE(DPLL_CTRL1, val); POSTING_READ(DPLL_CTRL1); @@ -930,24 +945,25 @@ static void skl_ddi_pll_enable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { const struct skl_dpll_regs *regs = skl_dpll_regs; + const enum intel_dpll_id id = pll->info->id; skl_ddi_pll_write_ctrl1(dev_priv, pll); - I915_WRITE(regs[pll->id].cfgcr1, pll->state.hw_state.cfgcr1); - I915_WRITE(regs[pll->id].cfgcr2, pll->state.hw_state.cfgcr2); - POSTING_READ(regs[pll->id].cfgcr1); - POSTING_READ(regs[pll->id].cfgcr2); + I915_WRITE(regs[id].cfgcr1, pll->state.hw_state.cfgcr1); + I915_WRITE(regs[id].cfgcr2, pll->state.hw_state.cfgcr2); + POSTING_READ(regs[id].cfgcr1); + POSTING_READ(regs[id].cfgcr2); /* the enable bit is always bit 31 */ - I915_WRITE(regs[pll->id].ctl, - I915_READ(regs[pll->id].ctl) | LCPLL_PLL_ENABLE); + I915_WRITE(regs[id].ctl, + I915_READ(regs[id].ctl) | LCPLL_PLL_ENABLE); if (intel_wait_for_register(dev_priv, DPLL_STATUS, - DPLL_LOCK(pll->id), - DPLL_LOCK(pll->id), + DPLL_LOCK(id), + DPLL_LOCK(id), 5)) - DRM_ERROR("DPLL %d not locked\n", pll->id); + DRM_ERROR("DPLL %d not locked\n", id); } static void skl_ddi_dpll0_enable(struct drm_i915_private *dev_priv, @@ -960,11 +976,12 @@ static void skl_ddi_pll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { const struct skl_dpll_regs *regs = skl_dpll_regs; + const enum intel_dpll_id id = pll->info->id; /* the enable bit is always bit 31 */ - I915_WRITE(regs[pll->id].ctl, - I915_READ(regs[pll->id].ctl) & ~LCPLL_PLL_ENABLE); - POSTING_READ(regs[pll->id].ctl); + I915_WRITE(regs[id].ctl, + I915_READ(regs[id].ctl) & ~LCPLL_PLL_ENABLE); + POSTING_READ(regs[id].ctl); } static void skl_ddi_dpll0_disable(struct drm_i915_private *dev_priv, @@ -978,6 +995,7 @@ static bool skl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, { uint32_t val; const struct skl_dpll_regs *regs = skl_dpll_regs; + const enum intel_dpll_id id = pll->info->id; bool ret; if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) @@ -985,17 +1003,17 @@ static bool skl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, ret = false; - val = I915_READ(regs[pll->id].ctl); + val = I915_READ(regs[id].ctl); if (!(val & LCPLL_PLL_ENABLE)) goto out; val = I915_READ(DPLL_CTRL1); - hw_state->ctrl1 = (val >> (pll->id * 6)) & 0x3f; + hw_state->ctrl1 = (val >> (id * 6)) & 0x3f; /* avoid reading back stale values if HDMI mode is not enabled */ - if (val & DPLL_CTRL1_HDMI_MODE(pll->id)) { - hw_state->cfgcr1 = I915_READ(regs[pll->id].cfgcr1); - hw_state->cfgcr2 = I915_READ(regs[pll->id].cfgcr2); + if (val & DPLL_CTRL1_HDMI_MODE(id)) { + hw_state->cfgcr1 = I915_READ(regs[id].cfgcr1); + hw_state->cfgcr2 = I915_READ(regs[id].cfgcr2); } ret = true; @@ -1011,6 +1029,7 @@ static bool skl_ddi_dpll0_get_hw_state(struct drm_i915_private *dev_priv, { uint32_t val; const struct skl_dpll_regs *regs = skl_dpll_regs; + const enum intel_dpll_id id = pll->info->id; bool ret; if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) @@ -1019,12 +1038,12 @@ static bool skl_ddi_dpll0_get_hw_state(struct drm_i915_private *dev_priv, ret = false; /* DPLL0 is always enabled since it drives CDCLK */ - val = I915_READ(regs[pll->id].ctl); + val = I915_READ(regs[id].ctl); if (WARN_ON(!(val & LCPLL_PLL_ENABLE))) goto out; val = I915_READ(DPLL_CTRL1); - hw_state->ctrl1 = (val >> (pll->id * 6)) & 0x3f; + hw_state->ctrl1 = (val >> (id * 6)) & 0x3f; ret = true; @@ -1424,7 +1443,7 @@ static void bxt_ddi_pll_enable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { uint32_t temp; - enum port port = (enum port)pll->id; /* 1:1 port->PLL mapping */ + enum port port = (enum port)pll->info->id; /* 1:1 port->PLL mapping */ enum dpio_phy phy; enum dpio_channel ch; @@ -1543,7 +1562,7 @@ static void bxt_ddi_pll_enable(struct drm_i915_private *dev_priv, static void bxt_ddi_pll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { - enum port port = (enum port)pll->id; /* 1:1 port->PLL mapping */ + enum port port = (enum port)pll->info->id; /* 1:1 port->PLL mapping */ uint32_t temp; temp = I915_READ(BXT_PORT_PLL_ENABLE(port)); @@ -1566,7 +1585,7 @@ static bool bxt_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll, struct intel_dpll_hw_state *hw_state) { - enum port port = (enum port)pll->id; /* 1:1 port->PLL mapping */ + enum port port = (enum port)pll->info->id; /* 1:1 port->PLL mapping */ uint32_t val; bool ret; enum dpio_phy phy; @@ -1824,7 +1843,7 @@ bxt_get_dpll(struct intel_crtc *crtc, pll = intel_get_shared_dpll_by_id(dev_priv, i); DRM_DEBUG_KMS("[CRTC:%d:%s] using pre-allocated %s\n", - crtc->base.base.id, crtc->base.name, pll->name); + crtc->base.base.id, crtc->base.name, pll->info->name); intel_reference_shared_dpll(pll, crtc_state); @@ -1877,13 +1896,6 @@ static void intel_ddi_pll_init(struct drm_device *dev) } } -struct dpll_info { - const char *name; - const int id; - const struct intel_shared_dpll_funcs *funcs; - uint32_t flags; -}; - struct intel_dpll_mgr { const struct dpll_info *dpll_info; @@ -1896,9 +1908,9 @@ struct intel_dpll_mgr { }; static const struct dpll_info pch_plls[] = { - { "PCH DPLL A", DPLL_ID_PCH_PLL_A, &ibx_pch_dpll_funcs, 0 }, - { "PCH DPLL B", DPLL_ID_PCH_PLL_B, &ibx_pch_dpll_funcs, 0 }, - { NULL, -1, NULL, 0 }, + { "PCH DPLL A", &ibx_pch_dpll_funcs, DPLL_ID_PCH_PLL_A, 0 }, + { "PCH DPLL B", &ibx_pch_dpll_funcs, DPLL_ID_PCH_PLL_B, 0 }, + { }, }; static const struct intel_dpll_mgr pch_pll_mgr = { @@ -1908,13 +1920,13 @@ static const struct intel_dpll_mgr pch_pll_mgr = { }; static const struct dpll_info hsw_plls[] = { - { "WRPLL 1", DPLL_ID_WRPLL1, &hsw_ddi_wrpll_funcs, 0 }, - { "WRPLL 2", DPLL_ID_WRPLL2, &hsw_ddi_wrpll_funcs, 0 }, - { "SPLL", DPLL_ID_SPLL, &hsw_ddi_spll_funcs, 0 }, - { "LCPLL 810", DPLL_ID_LCPLL_810, &hsw_ddi_lcpll_funcs, INTEL_DPLL_ALWAYS_ON }, - { "LCPLL 1350", DPLL_ID_LCPLL_1350, &hsw_ddi_lcpll_funcs, INTEL_DPLL_ALWAYS_ON }, - { "LCPLL 2700", DPLL_ID_LCPLL_2700, &hsw_ddi_lcpll_funcs, INTEL_DPLL_ALWAYS_ON }, - { NULL, -1, NULL, }, + { "WRPLL 1", &hsw_ddi_wrpll_funcs, DPLL_ID_WRPLL1, 0 }, + { "WRPLL 2", &hsw_ddi_wrpll_funcs, DPLL_ID_WRPLL2, 0 }, + { "SPLL", &hsw_ddi_spll_funcs, DPLL_ID_SPLL, 0 }, + { "LCPLL 810", &hsw_ddi_lcpll_funcs, DPLL_ID_LCPLL_810, INTEL_DPLL_ALWAYS_ON }, + { "LCPLL 1350", &hsw_ddi_lcpll_funcs, DPLL_ID_LCPLL_1350, INTEL_DPLL_ALWAYS_ON }, + { "LCPLL 2700", &hsw_ddi_lcpll_funcs, DPLL_ID_LCPLL_2700, INTEL_DPLL_ALWAYS_ON }, + { }, }; static const struct intel_dpll_mgr hsw_pll_mgr = { @@ -1924,11 +1936,11 @@ static const struct intel_dpll_mgr hsw_pll_mgr = { }; static const struct dpll_info skl_plls[] = { - { "DPLL 0", DPLL_ID_SKL_DPLL0, &skl_ddi_dpll0_funcs, INTEL_DPLL_ALWAYS_ON }, - { "DPLL 1", DPLL_ID_SKL_DPLL1, &skl_ddi_pll_funcs, 0 }, - { "DPLL 2", DPLL_ID_SKL_DPLL2, &skl_ddi_pll_funcs, 0 }, - { "DPLL 3", DPLL_ID_SKL_DPLL3, &skl_ddi_pll_funcs, 0 }, - { NULL, -1, NULL, }, + { "DPLL 0", &skl_ddi_dpll0_funcs, DPLL_ID_SKL_DPLL0, INTEL_DPLL_ALWAYS_ON }, + { "DPLL 1", &skl_ddi_pll_funcs, DPLL_ID_SKL_DPLL1, 0 }, + { "DPLL 2", &skl_ddi_pll_funcs, DPLL_ID_SKL_DPLL2, 0 }, + { "DPLL 3", &skl_ddi_pll_funcs, DPLL_ID_SKL_DPLL3, 0 }, + { }, }; static const struct intel_dpll_mgr skl_pll_mgr = { @@ -1938,10 +1950,10 @@ static const struct intel_dpll_mgr skl_pll_mgr = { }; static const struct dpll_info bxt_plls[] = { - { "PORT PLL A", DPLL_ID_SKL_DPLL0, &bxt_ddi_pll_funcs, 0 }, - { "PORT PLL B", DPLL_ID_SKL_DPLL1, &bxt_ddi_pll_funcs, 0 }, - { "PORT PLL C", DPLL_ID_SKL_DPLL2, &bxt_ddi_pll_funcs, 0 }, - { NULL, -1, NULL, }, + { "PORT PLL A", &bxt_ddi_pll_funcs, DPLL_ID_SKL_DPLL0, 0 }, + { "PORT PLL B", &bxt_ddi_pll_funcs, DPLL_ID_SKL_DPLL1, 0 }, + { "PORT PLL C", &bxt_ddi_pll_funcs, DPLL_ID_SKL_DPLL2, 0 }, + { }, }; static const struct intel_dpll_mgr bxt_pll_mgr = { @@ -1953,38 +1965,39 @@ static const struct intel_dpll_mgr bxt_pll_mgr = { static void cnl_ddi_pll_enable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; /* 1. Enable DPLL power in DPLL_ENABLE. */ - val = I915_READ(CNL_DPLL_ENABLE(pll->id)); + val = I915_READ(CNL_DPLL_ENABLE(id)); val |= PLL_POWER_ENABLE; - I915_WRITE(CNL_DPLL_ENABLE(pll->id), val); + I915_WRITE(CNL_DPLL_ENABLE(id), val); /* 2. Wait for DPLL power state enabled in DPLL_ENABLE. */ if (intel_wait_for_register(dev_priv, - CNL_DPLL_ENABLE(pll->id), + CNL_DPLL_ENABLE(id), PLL_POWER_STATE, PLL_POWER_STATE, 5)) - DRM_ERROR("PLL %d Power not enabled\n", pll->id); + DRM_ERROR("PLL %d Power not enabled\n", id); /* * 3. Configure DPLL_CFGCR0 to set SSC enable/disable, * select DP mode, and set DP link rate. */ val = pll->state.hw_state.cfgcr0; - I915_WRITE(CNL_DPLL_CFGCR0(pll->id), val); + I915_WRITE(CNL_DPLL_CFGCR0(id), val); /* 4. Reab back to ensure writes completed */ - POSTING_READ(CNL_DPLL_CFGCR0(pll->id)); + POSTING_READ(CNL_DPLL_CFGCR0(id)); /* 3. Configure DPLL_CFGCR0 */ /* Avoid touch CFGCR1 if HDMI mode is not enabled */ if (pll->state.hw_state.cfgcr0 & DPLL_CFGCR0_HDMI_MODE) { val = pll->state.hw_state.cfgcr1; - I915_WRITE(CNL_DPLL_CFGCR1(pll->id), val); + I915_WRITE(CNL_DPLL_CFGCR1(id), val); /* 4. Reab back to ensure writes completed */ - POSTING_READ(CNL_DPLL_CFGCR1(pll->id)); + POSTING_READ(CNL_DPLL_CFGCR1(id)); } /* @@ -1997,17 +2010,17 @@ static void cnl_ddi_pll_enable(struct drm_i915_private *dev_priv, */ /* 6. Enable DPLL in DPLL_ENABLE. */ - val = I915_READ(CNL_DPLL_ENABLE(pll->id)); + val = I915_READ(CNL_DPLL_ENABLE(id)); val |= PLL_ENABLE; - I915_WRITE(CNL_DPLL_ENABLE(pll->id), val); + I915_WRITE(CNL_DPLL_ENABLE(id), val); /* 7. Wait for PLL lock status in DPLL_ENABLE. */ if (intel_wait_for_register(dev_priv, - CNL_DPLL_ENABLE(pll->id), + CNL_DPLL_ENABLE(id), PLL_LOCK, PLL_LOCK, 5)) - DRM_ERROR("PLL %d not locked\n", pll->id); + DRM_ERROR("PLL %d not locked\n", id); /* * 8. If the frequency will result in a change to the voltage @@ -2027,6 +2040,7 @@ static void cnl_ddi_pll_enable(struct drm_i915_private *dev_priv, static void cnl_ddi_pll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; /* @@ -2044,17 +2058,17 @@ static void cnl_ddi_pll_disable(struct drm_i915_private *dev_priv, */ /* 3. Disable DPLL through DPLL_ENABLE. */ - val = I915_READ(CNL_DPLL_ENABLE(pll->id)); + val = I915_READ(CNL_DPLL_ENABLE(id)); val &= ~PLL_ENABLE; - I915_WRITE(CNL_DPLL_ENABLE(pll->id), val); + I915_WRITE(CNL_DPLL_ENABLE(id), val); /* 4. Wait for PLL not locked status in DPLL_ENABLE. */ if (intel_wait_for_register(dev_priv, - CNL_DPLL_ENABLE(pll->id), + CNL_DPLL_ENABLE(id), PLL_LOCK, 0, 5)) - DRM_ERROR("PLL %d locked\n", pll->id); + DRM_ERROR("PLL %d locked\n", id); /* * 5. If the frequency will result in a change to the voltage @@ -2066,23 +2080,24 @@ static void cnl_ddi_pll_disable(struct drm_i915_private *dev_priv, */ /* 6. Disable DPLL power in DPLL_ENABLE. */ - val = I915_READ(CNL_DPLL_ENABLE(pll->id)); + val = I915_READ(CNL_DPLL_ENABLE(id)); val &= ~PLL_POWER_ENABLE; - I915_WRITE(CNL_DPLL_ENABLE(pll->id), val); + I915_WRITE(CNL_DPLL_ENABLE(id), val); /* 7. Wait for DPLL power state disabled in DPLL_ENABLE. */ if (intel_wait_for_register(dev_priv, - CNL_DPLL_ENABLE(pll->id), + CNL_DPLL_ENABLE(id), PLL_POWER_STATE, 0, 5)) - DRM_ERROR("PLL %d Power not disabled\n", pll->id); + DRM_ERROR("PLL %d Power not disabled\n", id); } static bool cnl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll, struct intel_dpll_hw_state *hw_state) { + const enum intel_dpll_id id = pll->info->id; uint32_t val; bool ret; @@ -2091,16 +2106,16 @@ static bool cnl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, ret = false; - val = I915_READ(CNL_DPLL_ENABLE(pll->id)); + val = I915_READ(CNL_DPLL_ENABLE(id)); if (!(val & PLL_ENABLE)) goto out; - val = I915_READ(CNL_DPLL_CFGCR0(pll->id)); + val = I915_READ(CNL_DPLL_CFGCR0(id)); hw_state->cfgcr0 = val; /* avoid reading back stale values if HDMI mode is not enabled */ if (val & DPLL_CFGCR0_HDMI_MODE) { - hw_state->cfgcr1 = I915_READ(CNL_DPLL_CFGCR1(pll->id)); + hw_state->cfgcr1 = I915_READ(CNL_DPLL_CFGCR1(id)); } ret = true; @@ -2203,6 +2218,7 @@ cnl_ddi_calculate_wrpll(int clock, struct skl_wrpll_params *wrpll_params) { u32 afe_clock = clock * 5; + uint32_t ref_clock; u32 dco_min = 7998000; u32 dco_max = 10000000; u32 dco_mid = (dco_min + dco_max) / 2; @@ -2235,8 +2251,17 @@ cnl_ddi_calculate_wrpll(int clock, cnl_wrpll_get_multipliers(best_div, &pdiv, &qdiv, &kdiv); - cnl_wrpll_params_populate(wrpll_params, best_dco, - dev_priv->cdclk.hw.ref, pdiv, qdiv, kdiv); + ref_clock = dev_priv->cdclk.hw.ref; + + /* + * For ICL, the spec states: if reference frequency is 38.4, use 19.2 + * because the DPLL automatically divides that by 2. + */ + if (IS_ICELAKE(dev_priv) && ref_clock == 38400) + ref_clock = 19200; + + cnl_wrpll_params_populate(wrpll_params, best_dco, ref_clock, pdiv, qdiv, + kdiv); return true; } @@ -2372,10 +2397,10 @@ static const struct intel_shared_dpll_funcs cnl_ddi_pll_funcs = { }; static const struct dpll_info cnl_plls[] = { - { "DPLL 0", DPLL_ID_SKL_DPLL0, &cnl_ddi_pll_funcs, 0 }, - { "DPLL 1", DPLL_ID_SKL_DPLL1, &cnl_ddi_pll_funcs, 0 }, - { "DPLL 2", DPLL_ID_SKL_DPLL2, &cnl_ddi_pll_funcs, 0 }, - { NULL, -1, NULL, }, + { "DPLL 0", &cnl_ddi_pll_funcs, DPLL_ID_SKL_DPLL0, 0 }, + { "DPLL 1", &cnl_ddi_pll_funcs, DPLL_ID_SKL_DPLL1, 0 }, + { "DPLL 2", &cnl_ddi_pll_funcs, DPLL_ID_SKL_DPLL2, 0 }, + { }, }; static const struct intel_dpll_mgr cnl_pll_mgr = { @@ -2384,6 +2409,644 @@ static const struct intel_dpll_mgr cnl_pll_mgr = { .dump_hw_state = cnl_dump_hw_state, }; +/* + * These values alrea already adjusted: they're the bits we write to the + * registers, not the logical values. + */ +static const struct skl_wrpll_params icl_dp_combo_pll_24MHz_values[] = { + { .dco_integer = 0x151, .dco_fraction = 0x4000, /* [0]: 5.4 */ + .pdiv = 0x2 /* 3 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0}, + { .dco_integer = 0x151, .dco_fraction = 0x4000, /* [1]: 2.7 */ + .pdiv = 0x2 /* 3 */, .kdiv = 2, .qdiv_mode = 0, .qdiv_ratio = 0}, + { .dco_integer = 0x151, .dco_fraction = 0x4000, /* [2]: 1.62 */ + .pdiv = 0x4 /* 5 */, .kdiv = 2, .qdiv_mode = 0, .qdiv_ratio = 0}, + { .dco_integer = 0x151, .dco_fraction = 0x4000, /* [3]: 3.24 */ + .pdiv = 0x4 /* 5 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0}, + { .dco_integer = 0x168, .dco_fraction = 0x0000, /* [4]: 2.16 */ + .pdiv = 0x1 /* 2 */, .kdiv = 2, .qdiv_mode = 1, .qdiv_ratio = 2}, + { .dco_integer = 0x168, .dco_fraction = 0x0000, /* [5]: 4.32 */ + .pdiv = 0x1 /* 2 */, .kdiv = 2, .qdiv_mode = 0, .qdiv_ratio = 0}, + { .dco_integer = 0x195, .dco_fraction = 0x0000, /* [6]: 6.48 */ + .pdiv = 0x2 /* 3 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0}, + { .dco_integer = 0x151, .dco_fraction = 0x4000, /* [7]: 8.1 */ + .pdiv = 0x1 /* 2 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0}, +}; + +/* Also used for 38.4 MHz values. */ +static const struct skl_wrpll_params icl_dp_combo_pll_19_2MHz_values[] = { + { .dco_integer = 0x1A5, .dco_fraction = 0x7000, /* [0]: 5.4 */ + .pdiv = 0x2 /* 3 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0}, + { .dco_integer = 0x1A5, .dco_fraction = 0x7000, /* [1]: 2.7 */ + .pdiv = 0x2 /* 3 */, .kdiv = 2, .qdiv_mode = 0, .qdiv_ratio = 0}, + { .dco_integer = 0x1A5, .dco_fraction = 0x7000, /* [2]: 1.62 */ + .pdiv = 0x4 /* 5 */, .kdiv = 2, .qdiv_mode = 0, .qdiv_ratio = 0}, + { .dco_integer = 0x1A5, .dco_fraction = 0x7000, /* [3]: 3.24 */ + .pdiv = 0x4 /* 5 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0}, + { .dco_integer = 0x1C2, .dco_fraction = 0x0000, /* [4]: 2.16 */ + .pdiv = 0x1 /* 2 */, .kdiv = 2, .qdiv_mode = 1, .qdiv_ratio = 2}, + { .dco_integer = 0x1C2, .dco_fraction = 0x0000, /* [5]: 4.32 */ + .pdiv = 0x1 /* 2 */, .kdiv = 2, .qdiv_mode = 0, .qdiv_ratio = 0}, + { .dco_integer = 0x1FA, .dco_fraction = 0x2000, /* [6]: 6.48 */ + .pdiv = 0x2 /* 3 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0}, + { .dco_integer = 0x1A5, .dco_fraction = 0x7000, /* [7]: 8.1 */ + .pdiv = 0x1 /* 2 */, .kdiv = 1, .qdiv_mode = 0, .qdiv_ratio = 0}, +}; + +static bool icl_calc_dp_combo_pll(struct drm_i915_private *dev_priv, int clock, + struct skl_wrpll_params *pll_params) +{ + const struct skl_wrpll_params *params; + + params = dev_priv->cdclk.hw.ref == 24000 ? + icl_dp_combo_pll_24MHz_values : + icl_dp_combo_pll_19_2MHz_values; + + switch (clock) { + case 540000: + *pll_params = params[0]; + break; + case 270000: + *pll_params = params[1]; + break; + case 162000: + *pll_params = params[2]; + break; + case 324000: + *pll_params = params[3]; + break; + case 216000: + *pll_params = params[4]; + break; + case 432000: + *pll_params = params[5]; + break; + case 648000: + *pll_params = params[6]; + break; + case 810000: + *pll_params = params[7]; + break; + default: + MISSING_CASE(clock); + return false; + } + + return true; +} + +static bool icl_calc_dpll_state(struct intel_crtc_state *crtc_state, + struct intel_encoder *encoder, int clock, + struct intel_dpll_hw_state *pll_state) +{ + struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + uint32_t cfgcr0, cfgcr1; + struct skl_wrpll_params pll_params = { 0 }; + bool ret; + + if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI)) + ret = cnl_ddi_calculate_wrpll(clock, dev_priv, &pll_params); + else + ret = icl_calc_dp_combo_pll(dev_priv, clock, &pll_params); + + if (!ret) + return false; + + cfgcr0 = DPLL_CFGCR0_DCO_FRACTION(pll_params.dco_fraction) | + pll_params.dco_integer; + + cfgcr1 = DPLL_CFGCR1_QDIV_RATIO(pll_params.qdiv_ratio) | + DPLL_CFGCR1_QDIV_MODE(pll_params.qdiv_mode) | + DPLL_CFGCR1_KDIV(pll_params.kdiv) | + DPLL_CFGCR1_PDIV(pll_params.pdiv) | + DPLL_CFGCR1_CENTRAL_FREQ_8400; + + pll_state->cfgcr0 = cfgcr0; + pll_state->cfgcr1 = cfgcr1; + return true; +} + +static enum port icl_mg_pll_id_to_port(enum intel_dpll_id id) +{ + return id - DPLL_ID_ICL_MGPLL1 + PORT_C; +} + +static enum intel_dpll_id icl_port_to_mg_pll_id(enum port port) +{ + return port - PORT_C + DPLL_ID_ICL_MGPLL1; +} + +static bool icl_mg_pll_find_divisors(int clock_khz, bool is_dp, bool use_ssc, + uint32_t *target_dco_khz, + struct intel_dpll_hw_state *state) +{ + uint32_t dco_min_freq, dco_max_freq; + int div1_vals[] = {7, 5, 3, 2}; + unsigned int i; + int div2; + + dco_min_freq = is_dp ? 8100000 : use_ssc ? 8000000 : 7992000; + dco_max_freq = is_dp ? 8100000 : 10000000; + + for (i = 0; i < ARRAY_SIZE(div1_vals); i++) { + int div1 = div1_vals[i]; + + for (div2 = 10; div2 > 0; div2--) { + int dco = div1 * div2 * clock_khz * 5; + int a_divratio, tlinedrv, inputsel, hsdiv; + + if (dco < dco_min_freq || dco > dco_max_freq) + continue; + + if (div2 >= 2) { + a_divratio = is_dp ? 10 : 5; + tlinedrv = 2; + } else { + a_divratio = 5; + tlinedrv = 0; + } + inputsel = is_dp ? 0 : 1; + + switch (div1) { + default: + MISSING_CASE(div1); + case 2: + hsdiv = 0; + break; + case 3: + hsdiv = 1; + break; + case 5: + hsdiv = 2; + break; + case 7: + hsdiv = 3; + break; + } + + *target_dco_khz = dco; + + state->mg_refclkin_ctl = MG_REFCLKIN_CTL_OD_2_MUX(1); + + state->mg_clktop2_coreclkctl1 = + MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO(a_divratio); + + state->mg_clktop2_hsclkctl = + MG_CLKTOP2_HSCLKCTL_TLINEDRV_CLKSEL(tlinedrv) | + MG_CLKTOP2_HSCLKCTL_CORE_INPUTSEL(inputsel) | + MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO(hsdiv) | + MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO(div2); + + return true; + } + } + + return false; +} + +/* + * The specification for this function uses real numbers, so the math had to be + * adapted to integer-only calculation, that's why it looks so different. + */ +static bool icl_calc_mg_pll_state(struct intel_crtc_state *crtc_state, + struct intel_encoder *encoder, int clock, + struct intel_dpll_hw_state *pll_state) +{ + struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + int refclk_khz = dev_priv->cdclk.hw.ref; + uint32_t dco_khz, m1div, m2div_int, m2div_rem, m2div_frac; + uint32_t iref_ndiv, iref_trim, iref_pulse_w; + uint32_t prop_coeff, int_coeff; + uint32_t tdc_targetcnt, feedfwgain; + uint64_t ssc_stepsize, ssc_steplen, ssc_steplog; + uint64_t tmp; + bool use_ssc = false; + bool is_dp = !intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI); + + if (!icl_mg_pll_find_divisors(clock, is_dp, use_ssc, &dco_khz, + pll_state)) { + DRM_DEBUG_KMS("Failed to find divisors for clock %d\n", clock); + return false; + } + + m1div = 2; + m2div_int = dco_khz / (refclk_khz * m1div); + if (m2div_int > 255) { + m1div = 4; + m2div_int = dco_khz / (refclk_khz * m1div); + if (m2div_int > 255) { + DRM_DEBUG_KMS("Failed to find mdiv for clock %d\n", + clock); + return false; + } + } + m2div_rem = dco_khz % (refclk_khz * m1div); + + tmp = (uint64_t)m2div_rem * (1 << 22); + do_div(tmp, refclk_khz * m1div); + m2div_frac = tmp; + + switch (refclk_khz) { + case 19200: + iref_ndiv = 1; + iref_trim = 28; + iref_pulse_w = 1; + break; + case 24000: + iref_ndiv = 1; + iref_trim = 25; + iref_pulse_w = 2; + break; + case 38400: + iref_ndiv = 2; + iref_trim = 28; + iref_pulse_w = 1; + break; + default: + MISSING_CASE(refclk_khz); + return false; + } + + /* + * tdc_res = 0.000003 + * tdc_targetcnt = int(2 / (tdc_res * 8 * 50 * 1.1) / refclk_mhz + 0.5) + * + * The multiplication by 1000 is due to refclk MHz to KHz conversion. It + * was supposed to be a division, but we rearranged the operations of + * the formula to avoid early divisions so we don't multiply the + * rounding errors. + * + * 0.000003 * 8 * 50 * 1.1 = 0.00132, also known as 132 / 100000, which + * we also rearrange to work with integers. + * + * The 0.5 transformed to 5 results in a multiplication by 10 and the + * last division by 10. + */ + tdc_targetcnt = (2 * 1000 * 100000 * 10 / (132 * refclk_khz) + 5) / 10; + + /* + * Here we divide dco_khz by 10 in order to allow the dividend to fit in + * 32 bits. That's not a problem since we round the division down + * anyway. + */ + feedfwgain = (use_ssc || m2div_rem > 0) ? + m1div * 1000000 * 100 / (dco_khz * 3 / 10) : 0; + + if (dco_khz >= 9000000) { + prop_coeff = 5; + int_coeff = 10; + } else { + prop_coeff = 4; + int_coeff = 8; + } + + if (use_ssc) { + tmp = (uint64_t)dco_khz * 47 * 32; + do_div(tmp, refclk_khz * m1div * 10000); + ssc_stepsize = tmp; + + tmp = (uint64_t)dco_khz * 1000; + ssc_steplen = DIV_ROUND_UP_ULL(tmp, 32 * 2 * 32); + } else { + ssc_stepsize = 0; + ssc_steplen = 0; + } + ssc_steplog = 4; + + pll_state->mg_pll_div0 = (m2div_rem > 0 ? MG_PLL_DIV0_FRACNEN_H : 0) | + MG_PLL_DIV0_FBDIV_FRAC(m2div_frac) | + MG_PLL_DIV0_FBDIV_INT(m2div_int); + + pll_state->mg_pll_div1 = MG_PLL_DIV1_IREF_NDIVRATIO(iref_ndiv) | + MG_PLL_DIV1_DITHER_DIV_2 | + MG_PLL_DIV1_NDIVRATIO(1) | + MG_PLL_DIV1_FBPREDIV(m1div); + + pll_state->mg_pll_lf = MG_PLL_LF_TDCTARGETCNT(tdc_targetcnt) | + MG_PLL_LF_AFCCNTSEL_512 | + MG_PLL_LF_GAINCTRL(1) | + MG_PLL_LF_INT_COEFF(int_coeff) | + MG_PLL_LF_PROP_COEFF(prop_coeff); + + pll_state->mg_pll_frac_lock = MG_PLL_FRAC_LOCK_TRUELOCK_CRIT_32 | + MG_PLL_FRAC_LOCK_EARLYLOCK_CRIT_32 | + MG_PLL_FRAC_LOCK_LOCKTHRESH(10) | + MG_PLL_FRAC_LOCK_DCODITHEREN | + MG_PLL_FRAC_LOCK_FEEDFWRDGAIN(feedfwgain); + if (use_ssc || m2div_rem > 0) + pll_state->mg_pll_frac_lock |= MG_PLL_FRAC_LOCK_FEEDFWRDCAL_EN; + + pll_state->mg_pll_ssc = (use_ssc ? MG_PLL_SSC_EN : 0) | + MG_PLL_SSC_TYPE(2) | + MG_PLL_SSC_STEPLENGTH(ssc_steplen) | + MG_PLL_SSC_STEPNUM(ssc_steplog) | + MG_PLL_SSC_FLLEN | + MG_PLL_SSC_STEPSIZE(ssc_stepsize); + + pll_state->mg_pll_tdc_coldst_bias = MG_PLL_TDC_COLDST_COLDSTART; + + if (refclk_khz != 38400) { + pll_state->mg_pll_tdc_coldst_bias |= + MG_PLL_TDC_COLDST_IREFINT_EN | + MG_PLL_TDC_COLDST_REFBIAS_START_PULSE_W(iref_pulse_w) | + MG_PLL_TDC_COLDST_COLDSTART | + MG_PLL_TDC_TDCOVCCORR_EN | + MG_PLL_TDC_TDCSEL(3); + + pll_state->mg_pll_bias = MG_PLL_BIAS_BIAS_GB_SEL(3) | + MG_PLL_BIAS_INIT_DCOAMP(0x3F) | + MG_PLL_BIAS_BIAS_BONUS(10) | + MG_PLL_BIAS_BIASCAL_EN | + MG_PLL_BIAS_CTRIM(12) | + MG_PLL_BIAS_VREF_RDAC(4) | + MG_PLL_BIAS_IREFTRIM(iref_trim); + } + + return true; +} + +static struct intel_shared_dpll * +icl_get_dpll(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state, + struct intel_encoder *encoder) +{ + struct intel_shared_dpll *pll; + struct intel_dpll_hw_state pll_state = {}; + enum port port = encoder->port; + enum intel_dpll_id min, max; + int clock = crtc_state->port_clock; + bool ret; + + switch (port) { + case PORT_A: + case PORT_B: + min = DPLL_ID_ICL_DPLL0; + max = DPLL_ID_ICL_DPLL1; + ret = icl_calc_dpll_state(crtc_state, encoder, clock, + &pll_state); + break; + case PORT_C: + case PORT_D: + case PORT_E: + case PORT_F: + min = icl_port_to_mg_pll_id(port); + max = min; + ret = icl_calc_mg_pll_state(crtc_state, encoder, clock, + &pll_state); + break; + default: + MISSING_CASE(port); + return NULL; + } + + if (!ret) { + DRM_DEBUG_KMS("Could not calculate PLL state.\n"); + return NULL; + } + + crtc_state->dpll_hw_state = pll_state; + + pll = intel_find_shared_dpll(crtc, crtc_state, min, max); + if (!pll) { + DRM_DEBUG_KMS("No PLL selected\n"); + return NULL; + } + + intel_reference_shared_dpll(pll, crtc_state); + + return pll; +} + +static i915_reg_t icl_pll_id_to_enable_reg(enum intel_dpll_id id) +{ + switch (id) { + default: + MISSING_CASE(id); + case DPLL_ID_ICL_DPLL0: + case DPLL_ID_ICL_DPLL1: + return CNL_DPLL_ENABLE(id); + case DPLL_ID_ICL_MGPLL1: + case DPLL_ID_ICL_MGPLL2: + case DPLL_ID_ICL_MGPLL3: + case DPLL_ID_ICL_MGPLL4: + return MG_PLL_ENABLE(icl_mg_pll_id_to_port(id)); + } +} + +static bool icl_pll_get_hw_state(struct drm_i915_private *dev_priv, + struct intel_shared_dpll *pll, + struct intel_dpll_hw_state *hw_state) +{ + const enum intel_dpll_id id = pll->info->id; + uint32_t val; + enum port port; + bool ret = false; + + if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) + return false; + + val = I915_READ(icl_pll_id_to_enable_reg(id)); + if (!(val & PLL_ENABLE)) + goto out; + + switch (id) { + case DPLL_ID_ICL_DPLL0: + case DPLL_ID_ICL_DPLL1: + hw_state->cfgcr0 = I915_READ(ICL_DPLL_CFGCR0(id)); + hw_state->cfgcr1 = I915_READ(ICL_DPLL_CFGCR1(id)); + break; + case DPLL_ID_ICL_MGPLL1: + case DPLL_ID_ICL_MGPLL2: + case DPLL_ID_ICL_MGPLL3: + case DPLL_ID_ICL_MGPLL4: + port = icl_mg_pll_id_to_port(id); + hw_state->mg_refclkin_ctl = I915_READ(MG_REFCLKIN_CTL(port)); + hw_state->mg_clktop2_coreclkctl1 = + I915_READ(MG_CLKTOP2_CORECLKCTL1(port)); + hw_state->mg_clktop2_hsclkctl = + I915_READ(MG_CLKTOP2_HSCLKCTL(port)); + hw_state->mg_pll_div0 = I915_READ(MG_PLL_DIV0(port)); + hw_state->mg_pll_div1 = I915_READ(MG_PLL_DIV1(port)); + hw_state->mg_pll_lf = I915_READ(MG_PLL_LF(port)); + hw_state->mg_pll_frac_lock = I915_READ(MG_PLL_FRAC_LOCK(port)); + hw_state->mg_pll_ssc = I915_READ(MG_PLL_SSC(port)); + hw_state->mg_pll_bias = I915_READ(MG_PLL_BIAS(port)); + hw_state->mg_pll_tdc_coldst_bias = + I915_READ(MG_PLL_TDC_COLDST_BIAS(port)); + break; + default: + MISSING_CASE(id); + } + + ret = true; +out: + intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); + return ret; +} + +static void icl_dpll_write(struct drm_i915_private *dev_priv, + struct intel_shared_dpll *pll) +{ + struct intel_dpll_hw_state *hw_state = &pll->state.hw_state; + const enum intel_dpll_id id = pll->info->id; + + I915_WRITE(ICL_DPLL_CFGCR0(id), hw_state->cfgcr0); + I915_WRITE(ICL_DPLL_CFGCR1(id), hw_state->cfgcr1); + POSTING_READ(ICL_DPLL_CFGCR1(id)); +} + +static void icl_mg_pll_write(struct drm_i915_private *dev_priv, + struct intel_shared_dpll *pll) +{ + struct intel_dpll_hw_state *hw_state = &pll->state.hw_state; + enum port port = icl_mg_pll_id_to_port(pll->info->id); + + I915_WRITE(MG_REFCLKIN_CTL(port), hw_state->mg_refclkin_ctl); + I915_WRITE(MG_CLKTOP2_CORECLKCTL1(port), + hw_state->mg_clktop2_coreclkctl1); + I915_WRITE(MG_CLKTOP2_HSCLKCTL(port), hw_state->mg_clktop2_hsclkctl); + I915_WRITE(MG_PLL_DIV0(port), hw_state->mg_pll_div0); + I915_WRITE(MG_PLL_DIV1(port), hw_state->mg_pll_div1); + I915_WRITE(MG_PLL_LF(port), hw_state->mg_pll_lf); + I915_WRITE(MG_PLL_FRAC_LOCK(port), hw_state->mg_pll_frac_lock); + I915_WRITE(MG_PLL_SSC(port), hw_state->mg_pll_ssc); + I915_WRITE(MG_PLL_BIAS(port), hw_state->mg_pll_bias); + I915_WRITE(MG_PLL_TDC_COLDST_BIAS(port), + hw_state->mg_pll_tdc_coldst_bias); + POSTING_READ(MG_PLL_TDC_COLDST_BIAS(port)); +} + +static void icl_pll_enable(struct drm_i915_private *dev_priv, + struct intel_shared_dpll *pll) +{ + const enum intel_dpll_id id = pll->info->id; + i915_reg_t enable_reg = icl_pll_id_to_enable_reg(id); + uint32_t val; + + val = I915_READ(enable_reg); + val |= PLL_POWER_ENABLE; + I915_WRITE(enable_reg, val); + + /* + * The spec says we need to "wait" but it also says it should be + * immediate. + */ + if (intel_wait_for_register(dev_priv, enable_reg, PLL_POWER_STATE, + PLL_POWER_STATE, 1)) + DRM_ERROR("PLL %d Power not enabled\n", id); + + switch (id) { + case DPLL_ID_ICL_DPLL0: + case DPLL_ID_ICL_DPLL1: + icl_dpll_write(dev_priv, pll); + break; + case DPLL_ID_ICL_MGPLL1: + case DPLL_ID_ICL_MGPLL2: + case DPLL_ID_ICL_MGPLL3: + case DPLL_ID_ICL_MGPLL4: + icl_mg_pll_write(dev_priv, pll); + break; + default: + MISSING_CASE(id); + } + + /* + * DVFS pre sequence would be here, but in our driver the cdclk code + * paths should already be setting the appropriate voltage, hence we do + * nothign here. + */ + + val = I915_READ(enable_reg); + val |= PLL_ENABLE; + I915_WRITE(enable_reg, val); + + if (intel_wait_for_register(dev_priv, enable_reg, PLL_LOCK, PLL_LOCK, + 1)) /* 600us actually. */ + DRM_ERROR("PLL %d not locked\n", id); + + /* DVFS post sequence would be here. See the comment above. */ +} + +static void icl_pll_disable(struct drm_i915_private *dev_priv, + struct intel_shared_dpll *pll) +{ + const enum intel_dpll_id id = pll->info->id; + i915_reg_t enable_reg = icl_pll_id_to_enable_reg(id); + uint32_t val; + + /* The first steps are done by intel_ddi_post_disable(). */ + + /* + * DVFS pre sequence would be here, but in our driver the cdclk code + * paths should already be setting the appropriate voltage, hence we do + * nothign here. + */ + + val = I915_READ(enable_reg); + val &= ~PLL_ENABLE; + I915_WRITE(enable_reg, val); + + /* Timeout is actually 1us. */ + if (intel_wait_for_register(dev_priv, enable_reg, PLL_LOCK, 0, 1)) + DRM_ERROR("PLL %d locked\n", id); + + /* DVFS post sequence would be here. See the comment above. */ + + val = I915_READ(enable_reg); + val &= ~PLL_POWER_ENABLE; + I915_WRITE(enable_reg, val); + + /* + * The spec says we need to "wait" but it also says it should be + * immediate. + */ + if (intel_wait_for_register(dev_priv, enable_reg, PLL_POWER_STATE, 0, + 1)) + DRM_ERROR("PLL %d Power not disabled\n", id); +} + +static void icl_dump_hw_state(struct drm_i915_private *dev_priv, + struct intel_dpll_hw_state *hw_state) +{ + DRM_DEBUG_KMS("dpll_hw_state: cfgcr0: 0x%x, cfgcr1: 0x%x, " + "mg_refclkin_ctl: 0x%x, hg_clktop2_coreclkctl1: 0x%x, " + "mg_clktop2_hsclkctl: 0x%x, mg_pll_div0: 0x%x, " + "mg_pll_div2: 0x%x, mg_pll_lf: 0x%x, " + "mg_pll_frac_lock: 0x%x, mg_pll_ssc: 0x%x, " + "mg_pll_bias: 0x%x, mg_pll_tdc_coldst_bias: 0x%x\n", + hw_state->cfgcr0, hw_state->cfgcr1, + hw_state->mg_refclkin_ctl, + hw_state->mg_clktop2_coreclkctl1, + hw_state->mg_clktop2_hsclkctl, + hw_state->mg_pll_div0, + hw_state->mg_pll_div1, + hw_state->mg_pll_lf, + hw_state->mg_pll_frac_lock, + hw_state->mg_pll_ssc, + hw_state->mg_pll_bias, + hw_state->mg_pll_tdc_coldst_bias); +} + +static const struct intel_shared_dpll_funcs icl_pll_funcs = { + .enable = icl_pll_enable, + .disable = icl_pll_disable, + .get_hw_state = icl_pll_get_hw_state, +}; + +static const struct dpll_info icl_plls[] = { + { "DPLL 0", &icl_pll_funcs, DPLL_ID_ICL_DPLL0, 0 }, + { "DPLL 1", &icl_pll_funcs, DPLL_ID_ICL_DPLL1, 0 }, + { "MG PLL 1", &icl_pll_funcs, DPLL_ID_ICL_MGPLL1, 0 }, + { "MG PLL 2", &icl_pll_funcs, DPLL_ID_ICL_MGPLL2, 0 }, + { "MG PLL 3", &icl_pll_funcs, DPLL_ID_ICL_MGPLL3, 0 }, + { "MG PLL 4", &icl_pll_funcs, DPLL_ID_ICL_MGPLL4, 0 }, + { }, +}; + +static const struct intel_dpll_mgr icl_pll_mgr = { + .dpll_info = icl_plls, + .get_dpll = icl_get_dpll, + .dump_hw_state = icl_dump_hw_state, +}; + /** * intel_shared_dpll_init - Initialize shared DPLLs * @dev: drm device @@ -2397,7 +3060,9 @@ void intel_shared_dpll_init(struct drm_device *dev) const struct dpll_info *dpll_info; int i; - if (IS_CANNONLAKE(dev_priv)) + if (IS_ICELAKE(dev_priv)) + dpll_mgr = &icl_pll_mgr; + else if (IS_CANNONLAKE(dev_priv)) dpll_mgr = &cnl_pll_mgr; else if (IS_GEN9_BC(dev_priv)) dpll_mgr = &skl_pll_mgr; @@ -2415,13 +3080,9 @@ void intel_shared_dpll_init(struct drm_device *dev) dpll_info = dpll_mgr->dpll_info; - for (i = 0; dpll_info[i].id >= 0; i++) { + for (i = 0; dpll_info[i].name; i++) { WARN_ON(i != dpll_info[i].id); - - dev_priv->shared_dplls[i].id = dpll_info[i].id; - dev_priv->shared_dplls[i].name = dpll_info[i].name; - dev_priv->shared_dplls[i].funcs = *dpll_info[i].funcs; - dev_priv->shared_dplls[i].flags = dpll_info[i].flags; + dev_priv->shared_dplls[i].info = &dpll_info[i]; } dev_priv->dpll_mgr = dpll_mgr; @@ -2481,7 +3142,7 @@ void intel_release_shared_dpll(struct intel_shared_dpll *dpll, struct intel_shared_dpll_state *shared_dpll_state; shared_dpll_state = intel_atomic_get_shared_dpll_state(state); - shared_dpll_state[dpll->id].crtc_mask &= ~(1 << crtc->pipe); + shared_dpll_state[dpll->info->id].crtc_mask &= ~(1 << crtc->pipe); } /** diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.h b/drivers/gpu/drm/i915/intel_dpll_mgr.h index f24ccf4..7a0cd56 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.h +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.h @@ -103,6 +103,32 @@ enum intel_dpll_id { * @DPLL_ID_SKL_DPLL3: SKL and later DPLL3 */ DPLL_ID_SKL_DPLL3 = 3, + + + /** + * @DPLL_ID_ICL_DPLL0: ICL combo PHY DPLL0 + */ + DPLL_ID_ICL_DPLL0 = 0, + /** + * @DPLL_ID_ICL_DPLL1: ICL combo PHY DPLL1 + */ + DPLL_ID_ICL_DPLL1 = 1, + /** + * @DPLL_ID_ICL_MGPLL1: ICL MG PLL 1 port 1 (C) + */ + DPLL_ID_ICL_MGPLL1 = 2, + /** + * @DPLL_ID_ICL_MGPLL2: ICL MG PLL 1 port 2 (D) + */ + DPLL_ID_ICL_MGPLL2 = 3, + /** + * @DPLL_ID_ICL_MGPLL3: ICL MG PLL 1 port 3 (E) + */ + DPLL_ID_ICL_MGPLL3 = 4, + /** + * @DPLL_ID_ICL_MGPLL4: ICL MG PLL 1 port 4 (F) + */ + DPLL_ID_ICL_MGPLL4 = 5, }; #define I915_NUM_PLLS 6 @@ -135,6 +161,21 @@ struct intel_dpll_hw_state { /* bxt */ uint32_t ebb0, ebb4, pll0, pll1, pll2, pll3, pll6, pll8, pll9, pll10, pcsdw12; + + /* + * ICL uses the following, already defined: + * uint32_t cfgcr0, cfgcr1; + */ + uint32_t mg_refclkin_ctl; + uint32_t mg_clktop2_coreclkctl1; + uint32_t mg_clktop2_hsclkctl; + uint32_t mg_pll_div0; + uint32_t mg_pll_div1; + uint32_t mg_pll_lf; + uint32_t mg_pll_frac_lock; + uint32_t mg_pll_ssc; + uint32_t mg_pll_bias; + uint32_t mg_pll_tdc_coldst_bias; }; /** @@ -206,6 +247,37 @@ struct intel_shared_dpll_funcs { }; /** + * struct dpll_info - display PLL platform specific info + */ +struct dpll_info { + /** + * @name: DPLL name; used for logging + */ + const char *name; + + /** + * @funcs: platform specific hooks + */ + const struct intel_shared_dpll_funcs *funcs; + + /** + * @id: unique indentifier for this DPLL; should match the index in the + * dev_priv->shared_dplls array + */ + enum intel_dpll_id id; + +#define INTEL_DPLL_ALWAYS_ON (1 << 0) + /** + * @flags: + * + * INTEL_DPLL_ALWAYS_ON + * Inform the state checker that the DPLL is kept enabled even if + * not in use by any CRTC. + */ + uint32_t flags; +}; + +/** * struct intel_shared_dpll - display PLL with tracked state and users */ struct intel_shared_dpll { @@ -228,30 +300,9 @@ struct intel_shared_dpll { bool on; /** - * @name: DPLL name; used for logging + * @info: platform specific info */ - const char *name; - - /** - * @id: unique indentifier for this DPLL; should match the index in the - * dev_priv->shared_dplls array - */ - enum intel_dpll_id id; - - /** - * @funcs: platform specific hooks - */ - struct intel_shared_dpll_funcs funcs; - -#define INTEL_DPLL_ALWAYS_ON (1 << 0) - /** - * @flags: - * - * INTEL_DPLL_ALWAYS_ON - * Inform the state checker that the DPLL is kept enabled even if - * not in use by any CRTC. - */ - uint32_t flags; + const struct dpll_info *info; }; #define SKL_DPLL0 0 diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index a80fbad..d7dbca1 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -56,6 +56,8 @@ for (;;) { \ const bool expired__ = ktime_after(ktime_get_raw(), end__); \ OP; \ + /* Guarantee COND check prior to timeout */ \ + barrier(); \ if (COND) { \ ret__ = 0; \ break; \ @@ -96,6 +98,8 @@ u64 now = local_clock(); \ if (!(ATOMIC)) \ preempt_enable(); \ + /* Guarantee COND check prior to timeout */ \ + barrier(); \ if (COND) { \ ret = 0; \ break; \ @@ -140,6 +144,10 @@ #define KHz(x) (1000 * (x)) #define MHz(x) KHz(1000 * (x)) +#define KBps(x) (1000 * (x)) +#define MBps(x) KBps(1000 * (x)) +#define GBps(x) ((u64)1000 * MBps((x))) + /* * Display related stuff */ @@ -482,7 +490,7 @@ struct intel_atomic_state { bool skip_intermediate_wm; /* Gen9+ only */ - struct skl_wm_values wm_results; + struct skl_ddb_values wm_results; struct i915_sw_fence commit_ready; @@ -548,6 +556,12 @@ struct intel_initial_plane_config { #define SKL_MAX_DST_W 4096 #define SKL_MIN_DST_H 8 #define SKL_MAX_DST_H 4096 +#define ICL_MAX_SRC_W 5120 +#define ICL_MAX_SRC_H 4096 +#define ICL_MAX_DST_W 5120 +#define ICL_MAX_DST_H 4096 +#define SKL_MIN_YUV_420_SRC_W 16 +#define SKL_MIN_YUV_420_SRC_H 16 struct intel_scaler { int in_use; @@ -598,7 +612,9 @@ struct intel_pipe_wm { struct skl_plane_wm { struct skl_wm_level wm[8]; + struct skl_wm_level uv_wm[8]; struct skl_wm_level trans_wm; + bool is_planar; }; struct skl_pipe_wm { @@ -874,6 +890,7 @@ struct intel_crtc_state { /* bitmask of visible planes (enum plane_id) */ u8 active_planes; + u8 nv12_planes; /* HDMI scrambling status */ bool hdmi_scrambling; @@ -1321,10 +1338,14 @@ void intel_check_cpu_fifo_underruns(struct drm_i915_private *dev_priv); void intel_check_pch_fifo_underruns(struct drm_i915_private *dev_priv); /* i915_irq.c */ +bool gen11_reset_one_iir(struct drm_i915_private * const i915, + const unsigned int bank, + const unsigned int bit); void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask); void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask); void gen6_mask_pm_irq(struct drm_i915_private *dev_priv, u32 mask); void gen6_unmask_pm_irq(struct drm_i915_private *dev_priv, u32 mask); +void gen11_reset_rps_interrupts(struct drm_i915_private *dev_priv); void gen6_reset_rps_interrupts(struct drm_i915_private *dev_priv); void gen6_enable_rps_interrupts(struct drm_i915_private *dev_priv); void gen6_disable_rps_interrupts(struct drm_i915_private *dev_priv); @@ -1389,6 +1410,12 @@ uint32_t ddi_signal_levels(struct intel_dp *intel_dp); u8 intel_ddi_dp_voltage_max(struct intel_encoder *encoder); int intel_ddi_toggle_hdcp_signalling(struct intel_encoder *intel_encoder, bool enable); +void icl_map_plls_to_ports(struct drm_crtc *crtc, + struct intel_crtc_state *crtc_state, + struct drm_atomic_state *old_state); +void icl_unmap_plls_to_ports(struct drm_crtc *crtc, + struct intel_crtc_state *crtc_state, + struct drm_atomic_state *old_state); unsigned int intel_fb_align_height(const struct drm_framebuffer *fb, int plane, unsigned int height); @@ -1571,8 +1598,6 @@ void bxt_enable_dc9(struct drm_i915_private *dev_priv); void bxt_disable_dc9(struct drm_i915_private *dev_priv); void gen9_enable_dc5(struct drm_i915_private *dev_priv); unsigned int skl_cdclk_get_vco(unsigned int freq); -void skl_enable_dc6(struct drm_i915_private *dev_priv); -void skl_disable_dc6(struct drm_i915_private *dev_priv); void intel_dp_get_m_n(struct intel_crtc *crtc, struct intel_crtc_state *pipe_config); void intel_dp_set_m_n(struct intel_crtc *crtc, enum link_m_n_set m_n); @@ -1588,9 +1613,12 @@ void hsw_disable_ips(const struct intel_crtc_state *crtc_state); enum intel_display_power_domain intel_port_to_power_domain(enum port port); void intel_mode_from_pipe_config(struct drm_display_mode *mode, struct intel_crtc_state *pipe_config); +void intel_crtc_arm_fifo_underrun(struct intel_crtc *crtc, + struct intel_crtc_state *crtc_state); 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); +int skl_max_scale(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state, + uint32_t pixel_format); static inline u32 intel_plane_ggtt_offset(const struct intel_plane_state *state) { @@ -1607,6 +1635,7 @@ u32 skl_plane_stride(const struct drm_framebuffer *fb, int plane, int skl_check_plane_surface(const struct intel_crtc_state *crtc_state, struct intel_plane_state *plane_state); int i9xx_check_plane_surface(struct intel_plane_state *plane_state); +int skl_format_to_fourcc(int format, bool rgb_order, bool alpha); /* intel_csr.c */ void intel_csr_ucode_init(struct drm_i915_private *); @@ -1773,6 +1802,7 @@ void intel_fbc_flush(struct drm_i915_private *dev_priv, unsigned int frontbuffer_bits, enum fb_op_origin origin); void intel_fbc_cleanup_cfb(struct drm_i915_private *dev_priv); void intel_fbc_handle_fifo_underrun_irq(struct drm_i915_private *dev_priv); +int intel_fbc_reset_underrun(struct drm_i915_private *dev_priv); /* intel_hdmi.c */ void intel_hdmi_init(struct drm_i915_private *dev_priv, i915_reg_t hdmi_reg, @@ -1783,7 +1813,7 @@ struct intel_hdmi *enc_to_intel_hdmi(struct drm_encoder *encoder); bool intel_hdmi_compute_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config, struct drm_connector_state *conn_state); -void intel_hdmi_handle_sink_scrambling(struct intel_encoder *intel_encoder, +bool intel_hdmi_handle_sink_scrambling(struct intel_encoder *encoder, struct drm_connector *connector, bool high_tmds_clock_ratio, bool scrambling); @@ -1877,7 +1907,8 @@ void intel_psr_enable(struct intel_dp *intel_dp, void intel_psr_disable(struct intel_dp *intel_dp, const struct intel_crtc_state *old_crtc_state); void intel_psr_invalidate(struct drm_i915_private *dev_priv, - unsigned frontbuffer_bits); + unsigned frontbuffer_bits, + enum fb_op_origin origin); void intel_psr_flush(struct drm_i915_private *dev_priv, unsigned frontbuffer_bits, enum fb_op_origin origin); @@ -1886,6 +1917,8 @@ void intel_psr_single_frame_update(struct drm_i915_private *dev_priv, unsigned frontbuffer_bits); void intel_psr_compute_config(struct intel_dp *intel_dp, struct intel_crtc_state *crtc_state); +void intel_psr_irq_control(struct drm_i915_private *dev_priv, bool debug); +void intel_psr_irq_handler(struct drm_i915_private *dev_priv, u32 psr_iir); /* intel_runtime_pm.c */ int intel_power_domains_init(struct drm_i915_private *); @@ -1909,6 +1942,8 @@ bool intel_display_power_get_if_enabled(struct drm_i915_private *dev_priv, enum intel_display_power_domain domain); void intel_display_power_put(struct drm_i915_private *dev_priv, enum intel_display_power_domain domain); +void icl_dbuf_slices_update(struct drm_i915_private *dev_priv, + u8 req_slices); static inline void assert_rpm_device_not_suspended(struct drm_i915_private *dev_priv) @@ -2046,6 +2081,9 @@ void skl_disable_plane(struct intel_plane *plane, struct intel_crtc *crtc); bool skl_plane_get_hw_state(struct intel_plane *plane); bool skl_plane_has_ccs(struct drm_i915_private *dev_priv, enum pipe pipe, enum plane_id plane_id); +bool intel_format_is_yuv(uint32_t format); +bool skl_plane_has_planar(struct drm_i915_private *dev_priv, + enum pipe pipe, enum plane_id plane_id); /* intel_tv.c */ void intel_tv_init(struct drm_i915_private *dev_priv); @@ -2082,31 +2120,6 @@ intel_atomic_get_crtc_state(struct drm_atomic_state *state, return to_intel_crtc_state(crtc_state); } -static inline struct intel_crtc_state * -intel_atomic_get_existing_crtc_state(struct drm_atomic_state *state, - struct intel_crtc *crtc) -{ - struct drm_crtc_state *crtc_state; - - crtc_state = drm_atomic_get_existing_crtc_state(state, &crtc->base); - - if (crtc_state) - return to_intel_crtc_state(crtc_state); - else - return NULL; -} - -static inline struct intel_plane_state * -intel_atomic_get_existing_plane_state(struct drm_atomic_state *state, - struct intel_plane *plane) -{ - struct drm_plane_state *plane_state; - - plane_state = drm_atomic_get_existing_plane_state(state, &plane->base); - - return to_intel_plane_state(plane_state); -} - int intel_atomic_setup_scalers(struct drm_i915_private *dev_priv, struct intel_crtc *intel_crtc, struct intel_crtc_state *crtc_state); @@ -2138,8 +2151,17 @@ int intel_pipe_crc_create(struct drm_minor *minor); #ifdef CONFIG_DEBUG_FS int intel_crtc_set_crc_source(struct drm_crtc *crtc, const char *source_name, size_t *values_cnt); +void intel_crtc_disable_pipe_crc(struct intel_crtc *crtc); +void intel_crtc_enable_pipe_crc(struct intel_crtc *crtc); #else #define intel_crtc_set_crc_source NULL +static inline void intel_crtc_disable_pipe_crc(struct intel_crtc *crtc) +{ +} + +static inline void intel_crtc_enable_pipe_crc(struct intel_crtc *crtc) +{ +} #endif extern const struct file_operations i915_display_crc_ctl_fops; #endif /* __INTEL_DRV_H__ */ diff --git a/drivers/gpu/drm/i915/intel_dsi_vbt.c b/drivers/gpu/drm/i915/intel_dsi_vbt.c index 91c07b0..4d6ffa7 100644 --- a/drivers/gpu/drm/i915/intel_dsi_vbt.c +++ b/drivers/gpu/drm/i915/intel_dsi_vbt.c @@ -647,6 +647,11 @@ bool intel_dsi_vbt_init(struct intel_dsi *intel_dsi, u16 panel_id) /* prepare count */ prepare_cnt = DIV_ROUND_UP(ths_prepare_ns * ui_den, ui_num * mul); + if (prepare_cnt > PREPARE_CNT_MAX) { + DRM_DEBUG_KMS("prepare count too high %u\n", prepare_cnt); + prepare_cnt = PREPARE_CNT_MAX; + } + /* exit zero count */ exit_zero_cnt = DIV_ROUND_UP( (ths_prepare_hszero - ths_prepare_ns) * ui_den, @@ -662,32 +667,29 @@ bool intel_dsi_vbt_init(struct intel_dsi *intel_dsi, u16 panel_id) if (exit_zero_cnt < (55 * ui_den / ui_num) && (55 * ui_den) % ui_num) exit_zero_cnt += 1; + if (exit_zero_cnt > EXIT_ZERO_CNT_MAX) { + DRM_DEBUG_KMS("exit zero count too high %u\n", exit_zero_cnt); + exit_zero_cnt = EXIT_ZERO_CNT_MAX; + } + /* clk zero count */ clk_zero_cnt = DIV_ROUND_UP( (tclk_prepare_clkzero - ths_prepare_ns) * ui_den, ui_num * mul); + if (clk_zero_cnt > CLK_ZERO_CNT_MAX) { + DRM_DEBUG_KMS("clock zero count too high %u\n", clk_zero_cnt); + clk_zero_cnt = CLK_ZERO_CNT_MAX; + } + /* trail count */ tclk_trail_ns = max(mipi_config->tclk_trail, mipi_config->ths_trail); trail_cnt = DIV_ROUND_UP(tclk_trail_ns * ui_den, ui_num * mul); - if (prepare_cnt > PREPARE_CNT_MAX || - exit_zero_cnt > EXIT_ZERO_CNT_MAX || - clk_zero_cnt > CLK_ZERO_CNT_MAX || - trail_cnt > TRAIL_CNT_MAX) - DRM_DEBUG_DRIVER("Values crossing maximum limits, restricting to max values\n"); - - if (prepare_cnt > PREPARE_CNT_MAX) - prepare_cnt = PREPARE_CNT_MAX; - - if (exit_zero_cnt > EXIT_ZERO_CNT_MAX) - exit_zero_cnt = EXIT_ZERO_CNT_MAX; - - if (clk_zero_cnt > CLK_ZERO_CNT_MAX) - clk_zero_cnt = CLK_ZERO_CNT_MAX; - - if (trail_cnt > TRAIL_CNT_MAX) + if (trail_cnt > TRAIL_CNT_MAX) { + DRM_DEBUG_KMS("trail count too high %u\n", trail_cnt); trail_cnt = TRAIL_CNT_MAX; + } /* B080 */ intel_dsi->dphy_reg = exit_zero_cnt << 24 | trail_cnt << 16 | diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index f7c2582..6bfd7e3 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -81,13 +81,17 @@ static const struct engine_class_info intel_engine_classes[] = { }, }; +#define MAX_MMIO_BASES 3 struct engine_info { unsigned int hw_id; unsigned int uabi_id; u8 class; u8 instance; - u32 mmio_base; - unsigned irq_shift; + /* mmio bases table *must* be sorted in reverse gen order */ + struct engine_mmio_base { + u32 gen : 8; + u32 base : 24; + } mmio_bases[MAX_MMIO_BASES]; }; static const struct engine_info intel_engines[] = { @@ -96,64 +100,76 @@ static const struct engine_info intel_engines[] = { .uabi_id = I915_EXEC_RENDER, .class = RENDER_CLASS, .instance = 0, - .mmio_base = RENDER_RING_BASE, - .irq_shift = GEN8_RCS_IRQ_SHIFT, + .mmio_bases = { + { .gen = 1, .base = RENDER_RING_BASE } + }, }, [BCS] = { .hw_id = BCS_HW, .uabi_id = I915_EXEC_BLT, .class = COPY_ENGINE_CLASS, .instance = 0, - .mmio_base = BLT_RING_BASE, - .irq_shift = GEN8_BCS_IRQ_SHIFT, + .mmio_bases = { + { .gen = 6, .base = BLT_RING_BASE } + }, }, [VCS] = { .hw_id = VCS_HW, .uabi_id = I915_EXEC_BSD, .class = VIDEO_DECODE_CLASS, .instance = 0, - .mmio_base = GEN6_BSD_RING_BASE, - .irq_shift = GEN8_VCS1_IRQ_SHIFT, + .mmio_bases = { + { .gen = 11, .base = GEN11_BSD_RING_BASE }, + { .gen = 6, .base = GEN6_BSD_RING_BASE }, + { .gen = 4, .base = BSD_RING_BASE } + }, }, [VCS2] = { .hw_id = VCS2_HW, .uabi_id = I915_EXEC_BSD, .class = VIDEO_DECODE_CLASS, .instance = 1, - .mmio_base = GEN8_BSD2_RING_BASE, - .irq_shift = GEN8_VCS2_IRQ_SHIFT, + .mmio_bases = { + { .gen = 11, .base = GEN11_BSD2_RING_BASE }, + { .gen = 8, .base = GEN8_BSD2_RING_BASE } + }, }, [VCS3] = { .hw_id = VCS3_HW, .uabi_id = I915_EXEC_BSD, .class = VIDEO_DECODE_CLASS, .instance = 2, - .mmio_base = GEN11_BSD3_RING_BASE, - .irq_shift = 0, /* not used */ + .mmio_bases = { + { .gen = 11, .base = GEN11_BSD3_RING_BASE } + }, }, [VCS4] = { .hw_id = VCS4_HW, .uabi_id = I915_EXEC_BSD, .class = VIDEO_DECODE_CLASS, .instance = 3, - .mmio_base = GEN11_BSD4_RING_BASE, - .irq_shift = 0, /* not used */ + .mmio_bases = { + { .gen = 11, .base = GEN11_BSD4_RING_BASE } + }, }, [VECS] = { .hw_id = VECS_HW, .uabi_id = I915_EXEC_VEBOX, .class = VIDEO_ENHANCEMENT_CLASS, .instance = 0, - .mmio_base = VEBOX_RING_BASE, - .irq_shift = GEN8_VECS_IRQ_SHIFT, + .mmio_bases = { + { .gen = 11, .base = GEN11_VEBOX_RING_BASE }, + { .gen = 7, .base = VEBOX_RING_BASE } + }, }, [VECS2] = { .hw_id = VECS2_HW, .uabi_id = I915_EXEC_VEBOX, .class = VIDEO_ENHANCEMENT_CLASS, .instance = 1, - .mmio_base = GEN11_VEBOX2_RING_BASE, - .irq_shift = 0, /* not used */ + .mmio_bases = { + { .gen = 11, .base = GEN11_VEBOX2_RING_BASE } + }, }, }; @@ -223,16 +239,36 @@ __intel_engine_context_size(struct drm_i915_private *dev_priv, u8 class) } } +static u32 __engine_mmio_base(struct drm_i915_private *i915, + const struct engine_mmio_base *bases) +{ + int i; + + for (i = 0; i < MAX_MMIO_BASES; i++) + if (INTEL_GEN(i915) >= bases[i].gen) + break; + + GEM_BUG_ON(i == MAX_MMIO_BASES); + GEM_BUG_ON(!bases[i].base); + + return bases[i].base; +} + +static void __sprint_engine_name(char *name, const struct engine_info *info) +{ + WARN_ON(snprintf(name, INTEL_ENGINE_CS_MAX_NAME, "%s%u", + intel_engine_classes[info->class].name, + info->instance) >= INTEL_ENGINE_CS_MAX_NAME); +} + static int intel_engine_setup(struct drm_i915_private *dev_priv, enum intel_engine_id id) { const struct engine_info *info = &intel_engines[id]; - const struct engine_class_info *class_info; struct intel_engine_cs *engine; GEM_BUG_ON(info->class >= ARRAY_SIZE(intel_engine_classes)); - class_info = &intel_engine_classes[info->class]; BUILD_BUG_ON(MAX_ENGINE_CLASS >= BIT(GEN11_ENGINE_CLASS_WIDTH)); BUILD_BUG_ON(MAX_ENGINE_INSTANCE >= BIT(GEN11_ENGINE_INSTANCE_WIDTH)); @@ -253,35 +289,14 @@ intel_engine_setup(struct drm_i915_private *dev_priv, engine->id = id; engine->i915 = dev_priv; - WARN_ON(snprintf(engine->name, sizeof(engine->name), "%s%u", - class_info->name, info->instance) >= - sizeof(engine->name)); + __sprint_engine_name(engine->name, info); engine->hw_id = engine->guc_id = info->hw_id; - if (INTEL_GEN(dev_priv) >= 11) { - switch (engine->id) { - case VCS: - engine->mmio_base = GEN11_BSD_RING_BASE; - break; - case VCS2: - engine->mmio_base = GEN11_BSD2_RING_BASE; - break; - case VECS: - engine->mmio_base = GEN11_VEBOX_RING_BASE; - break; - default: - /* take the original value for all other engines */ - engine->mmio_base = info->mmio_base; - break; - } - } else { - engine->mmio_base = info->mmio_base; - } - engine->irq_shift = info->irq_shift; + engine->mmio_base = __engine_mmio_base(dev_priv, info->mmio_bases); engine->class = info->class; engine->instance = info->instance; engine->uabi_id = info->uabi_id; - engine->uabi_class = class_info->uabi_class; + engine->uabi_class = intel_engine_classes[info->class].uabi_class; engine->context_size = __intel_engine_context_size(dev_priv, engine->class); @@ -291,7 +306,7 @@ intel_engine_setup(struct drm_i915_private *dev_priv, /* Nothing to do here, execute in order of dependencies */ engine->schedule = NULL; - spin_lock_init(&engine->stats.lock); + seqlock_init(&engine->stats.lock); ATOMIC_INIT_NOTIFIER_HEAD(&engine->context_status_notifier); @@ -436,21 +451,13 @@ void intel_engine_init_global_seqno(struct intel_engine_cs *engine, u32 seqno) GEM_BUG_ON(intel_engine_get_seqno(engine) != seqno); } -static void intel_engine_init_timeline(struct intel_engine_cs *engine) +static void intel_engine_init_batch_pool(struct intel_engine_cs *engine) { - engine->timeline = &engine->i915->gt.global_timeline.engine[engine->id]; + i915_gem_batch_pool_init(&engine->batch_pool, engine); } static bool csb_force_mmio(struct drm_i915_private *i915) { - /* - * IOMMU adds unpredictable latency causing the CSB write (from the - * GPU into the HWSP) to only be visible some time after the interrupt - * (missed breadcrumb syndrome). - */ - if (intel_vtd_active()) - return true; - /* Older GVT emulation depends upon intercepting CSB mmio */ if (intel_vgpu_active(i915) && !intel_vgpu_has_hwsp_emulation(i915)) return true; @@ -484,12 +491,11 @@ static void intel_engine_init_execlist(struct intel_engine_cs *engine) */ void intel_engine_setup_common(struct intel_engine_cs *engine) { - intel_engine_init_execlist(engine); + i915_timeline_init(engine->i915, &engine->timeline, engine->name); - intel_engine_init_timeline(engine); + intel_engine_init_execlist(engine); intel_engine_init_hangcheck(engine); - i915_gem_batch_pool_init(engine, &engine->batch_pool); - + intel_engine_init_batch_pool(engine); intel_engine_init_cmd_parser(engine); } @@ -520,8 +526,6 @@ int intel_engine_create_scratch(struct intel_engine_cs *engine, int size) goto err_unref; engine->scratch = vma; - DRM_DEBUG_DRIVER("%s pipe control offset: 0x%08x\n", - engine->name, i915_ggtt_offset(vma)); return 0; err_unref: @@ -615,9 +619,6 @@ 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, PAGE_SIZE); - - DRM_DEBUG_DRIVER("%s hws offset: 0x%08x\n", - engine->name, i915_ggtt_offset(vma)); return 0; err_unpin: @@ -669,7 +670,7 @@ int intel_engine_init_common(struct intel_engine_cs *engine) * be available. To avoid this we always pin the default * context. */ - ring = engine->context_pin(engine, engine->i915->kernel_context); + ring = intel_context_pin(engine->i915->kernel_context, engine); if (IS_ERR(ring)) return PTR_ERR(ring); @@ -678,8 +679,7 @@ int intel_engine_init_common(struct intel_engine_cs *engine) * we can interrupt the engine at any time. */ if (engine->i915->preempt_context) { - ring = engine->context_pin(engine, - engine->i915->preempt_context); + ring = intel_context_pin(engine->i915->preempt_context, engine); if (IS_ERR(ring)) { ret = PTR_ERR(ring); goto err_unpin_kernel; @@ -703,9 +703,9 @@ err_breadcrumbs: intel_engine_fini_breadcrumbs(engine); err_unpin_preempt: if (engine->i915->preempt_context) - engine->context_unpin(engine, engine->i915->preempt_context); + intel_context_unpin(engine->i915->preempt_context, engine); err_unpin_kernel: - engine->context_unpin(engine, engine->i915->kernel_context); + intel_context_unpin(engine->i915->kernel_context, engine); return ret; } @@ -733,8 +733,10 @@ void intel_engine_cleanup_common(struct intel_engine_cs *engine) i915_gem_object_put(engine->default_state); if (engine->i915->preempt_context) - engine->context_unpin(engine, engine->i915->preempt_context); - engine->context_unpin(engine, engine->i915->kernel_context); + intel_context_unpin(engine->i915->preempt_context, engine); + intel_context_unpin(engine->i915->kernel_context, engine); + + i915_timeline_fini(&engine->timeline); } u64 intel_engine_get_active_head(const struct intel_engine_cs *engine) @@ -782,10 +784,24 @@ static inline uint32_t read_subslice_reg(struct drm_i915_private *dev_priv, int slice, int subslice, i915_reg_t reg) { + uint32_t mcr_slice_subslice_mask; + uint32_t mcr_slice_subslice_select; uint32_t mcr; uint32_t ret; enum forcewake_domains fw_domains; + if (INTEL_GEN(dev_priv) >= 11) { + mcr_slice_subslice_mask = GEN11_MCR_SLICE_MASK | + GEN11_MCR_SUBSLICE_MASK; + mcr_slice_subslice_select = GEN11_MCR_SLICE(slice) | + GEN11_MCR_SUBSLICE(subslice); + } else { + mcr_slice_subslice_mask = GEN8_MCR_SLICE_MASK | + GEN8_MCR_SUBSLICE_MASK; + mcr_slice_subslice_select = GEN8_MCR_SLICE(slice) | + GEN8_MCR_SUBSLICE(subslice); + } + fw_domains = intel_uncore_forcewake_for_reg(dev_priv, reg, FW_REG_READ); fw_domains |= intel_uncore_forcewake_for_reg(dev_priv, @@ -800,14 +816,14 @@ read_subslice_reg(struct drm_i915_private *dev_priv, int slice, * The HW expects the slice and sublice selectors to be reset to 0 * after reading out the registers. */ - WARN_ON_ONCE(mcr & (GEN8_MCR_SLICE_MASK | GEN8_MCR_SUBSLICE_MASK)); - mcr &= ~(GEN8_MCR_SLICE_MASK | GEN8_MCR_SUBSLICE_MASK); - mcr |= GEN8_MCR_SLICE(slice) | GEN8_MCR_SUBSLICE(subslice); + WARN_ON_ONCE(mcr & mcr_slice_subslice_mask); + mcr &= ~mcr_slice_subslice_mask; + mcr |= mcr_slice_subslice_select; I915_WRITE_FW(GEN8_MCR_SELECTOR, mcr); ret = I915_READ_FW(reg); - mcr &= ~(GEN8_MCR_SLICE_MASK | GEN8_MCR_SUBSLICE_MASK); + mcr &= ~mcr_slice_subslice_mask; I915_WRITE_FW(GEN8_MCR_SELECTOR, mcr); intel_uncore_forcewake_put__locked(dev_priv, fw_domains); @@ -871,644 +887,6 @@ void intel_engine_get_instdone(struct intel_engine_cs *engine, } } -static int wa_add(struct drm_i915_private *dev_priv, - i915_reg_t addr, - const u32 mask, const u32 val) -{ - const u32 idx = dev_priv->workarounds.count; - - if (WARN_ON(idx >= I915_MAX_WA_REGS)) - return -ENOSPC; - - dev_priv->workarounds.reg[idx].addr = addr; - dev_priv->workarounds.reg[idx].value = val; - dev_priv->workarounds.reg[idx].mask = mask; - - dev_priv->workarounds.count++; - - return 0; -} - -#define WA_REG(addr, mask, val) do { \ - const int r = wa_add(dev_priv, (addr), (mask), (val)); \ - if (r) \ - return r; \ - } while (0) - -#define WA_SET_BIT_MASKED(addr, mask) \ - WA_REG(addr, (mask), _MASKED_BIT_ENABLE(mask)) - -#define WA_CLR_BIT_MASKED(addr, mask) \ - WA_REG(addr, (mask), _MASKED_BIT_DISABLE(mask)) - -#define WA_SET_FIELD_MASKED(addr, mask, value) \ - WA_REG(addr, mask, _MASKED_FIELD(mask, value)) - -static int wa_ring_whitelist_reg(struct intel_engine_cs *engine, - i915_reg_t reg) -{ - struct drm_i915_private *dev_priv = engine->i915; - struct i915_workarounds *wa = &dev_priv->workarounds; - const uint32_t index = wa->hw_whitelist_count[engine->id]; - - if (WARN_ON(index >= RING_MAX_NONPRIV_SLOTS)) - return -EINVAL; - - I915_WRITE(RING_FORCE_TO_NONPRIV(engine->mmio_base, index), - i915_mmio_reg_offset(reg)); - wa->hw_whitelist_count[engine->id]++; - - return 0; -} - -static int gen8_init_workarounds(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - - WA_SET_BIT_MASKED(INSTPM, INSTPM_FORCE_ORDERING); - - /* WaDisableAsyncFlipPerfMode:bdw,chv */ - WA_SET_BIT_MASKED(MI_MODE, ASYNC_FLIP_PERF_DISABLE); - - /* WaDisablePartialInstShootdown:bdw,chv */ - WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, - PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE); - - /* Use Force Non-Coherent whenever executing a 3D context. This is a - * workaround for for a possible hang in the unlikely event a TLB - * invalidation occurs during a PSD flush. - */ - /* WaForceEnableNonCoherent:bdw,chv */ - /* WaHdcDisableFetchWhenMasked:bdw,chv */ - WA_SET_BIT_MASKED(HDC_CHICKEN0, - HDC_DONOT_FETCH_MEM_WHEN_MASKED | - HDC_FORCE_NON_COHERENT); - - /* From the Haswell PRM, Command Reference: Registers, CACHE_MODE_0: - * "The Hierarchical Z RAW Stall Optimization allows non-overlapping - * polygons in the same 8x4 pixel/sample area to be processed without - * stalling waiting for the earlier ones to write to Hierarchical Z - * buffer." - * - * This optimization is off by default for BDW and CHV; turn it on. - */ - WA_CLR_BIT_MASKED(CACHE_MODE_0_GEN7, HIZ_RAW_STALL_OPT_DISABLE); - - /* Wa4x4STCOptimizationDisable:bdw,chv */ - WA_SET_BIT_MASKED(CACHE_MODE_1, GEN8_4x4_STC_OPTIMIZATION_DISABLE); - - /* - * BSpec recommends 8x4 when MSAA is used, - * however in practice 16x4 seems fastest. - * - * Note that PS/WM thread counts depend on the WIZ hashing - * disable bit, which we don't touch here, but it's good - * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM). - */ - WA_SET_FIELD_MASKED(GEN7_GT_MODE, - GEN6_WIZ_HASHING_MASK, - GEN6_WIZ_HASHING_16x4); - - return 0; -} - -static int bdw_init_workarounds(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int ret; - - ret = gen8_init_workarounds(engine); - if (ret) - return ret; - - /* WaDisableThreadStallDopClockGating:bdw (pre-production) */ - WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE); - - /* WaDisableDopClockGating:bdw - * - * Also see the related UCGTCL1 write in broadwell_init_clock_gating() - * to disable EUTC clock gating. - */ - WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2, - DOP_CLOCK_GATING_DISABLE); - - WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, - GEN8_SAMPLER_POWER_BYPASS_DIS); - - WA_SET_BIT_MASKED(HDC_CHICKEN0, - /* WaForceContextSaveRestoreNonCoherent:bdw */ - HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT | - /* WaDisableFenceDestinationToSLM:bdw (pre-prod) */ - (IS_BDW_GT3(dev_priv) ? HDC_FENCE_DEST_SLM_DISABLE : 0)); - - return 0; -} - -static int chv_init_workarounds(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int ret; - - ret = gen8_init_workarounds(engine); - if (ret) - return ret; - - /* WaDisableThreadStallDopClockGating:chv */ - WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE); - - /* Improve HiZ throughput on CHV. */ - WA_SET_BIT_MASKED(HIZ_CHICKEN, CHV_HZ_8X8_MODE_IN_1X); - - return 0; -} - -static int gen9_init_workarounds(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int ret; - - /* WaConextSwitchWithConcurrentTLBInvalidate:skl,bxt,kbl,glk,cfl */ - I915_WRITE(GEN9_CSFE_CHICKEN1_RCS, _MASKED_BIT_ENABLE(GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE)); - - /* WaEnableLbsSlaRetryTimerDecrement:skl,bxt,kbl,glk,cfl */ - I915_WRITE(BDW_SCRATCH1, I915_READ(BDW_SCRATCH1) | - GEN9_LBS_SLA_RETRY_TIMER_DECREMENT_ENABLE); - - /* WaDisableKillLogic:bxt,skl,kbl */ - if (!IS_COFFEELAKE(dev_priv)) - I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | - ECOCHK_DIS_TLB); - - if (HAS_LLC(dev_priv)) { - /* WaCompressedResourceSamplerPbeMediaNewHashMode:skl,kbl - * - * Must match Display Engine. See - * WaCompressedResourceDisplayNewHashMode. - */ - WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, - GEN9_PBE_COMPRESSED_HASH_SELECTION); - WA_SET_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN7, - GEN9_SAMPLER_HASH_COMPRESSED_READ_ADDR); - - I915_WRITE(MMCD_MISC_CTRL, - I915_READ(MMCD_MISC_CTRL) | - MMCD_PCLA | - MMCD_HOTSPOT_EN); - } - - /* WaClearFlowControlGpgpuContextSave:skl,bxt,kbl,glk,cfl */ - /* WaDisablePartialInstShootdown:skl,bxt,kbl,glk,cfl */ - WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, - FLOW_CONTROL_ENABLE | - PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE); - - /* Syncing dependencies between camera and graphics:skl,bxt,kbl */ - if (!IS_COFFEELAKE(dev_priv)) - WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, - GEN9_DISABLE_OCL_OOB_SUPPRESS_LOGIC); - - /* WaEnableYV12BugFixInHalfSliceChicken7:skl,bxt,kbl,glk,cfl */ - /* WaEnableSamplerGPGPUPreemptionSupport:skl,bxt,kbl,cfl */ - WA_SET_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN7, - GEN9_ENABLE_YV12_BUGFIX | - GEN9_ENABLE_GPGPU_PREEMPTION); - - /* Wa4x4STCOptimizationDisable:skl,bxt,kbl,glk,cfl */ - /* WaDisablePartialResolveInVc:skl,bxt,kbl,cfl */ - WA_SET_BIT_MASKED(CACHE_MODE_1, (GEN8_4x4_STC_OPTIMIZATION_DISABLE | - GEN9_PARTIAL_RESOLVE_IN_VC_DISABLE)); - - /* WaCcsTlbPrefetchDisable:skl,bxt,kbl,glk,cfl */ - WA_CLR_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN5, - GEN9_CCS_TLB_PREFETCH_ENABLE); - - /* WaForceContextSaveRestoreNonCoherent:skl,bxt,kbl,cfl */ - WA_SET_BIT_MASKED(HDC_CHICKEN0, - HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT | - HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE); - - /* WaForceEnableNonCoherent and WaDisableHDCInvalidation are - * both tied to WaForceContextSaveRestoreNonCoherent - * in some hsds for skl. We keep the tie for all gen9. The - * documentation is a bit hazy and so we want to get common behaviour, - * even though there is no clear evidence we would need both on kbl/bxt. - * This area has been source of system hangs so we play it safe - * and mimic the skl regardless of what bspec says. - * - * Use Force Non-Coherent whenever executing a 3D context. This - * is a workaround for a possible hang in the unlikely event - * a TLB invalidation occurs during a PSD flush. - */ - - /* WaForceEnableNonCoherent:skl,bxt,kbl,cfl */ - WA_SET_BIT_MASKED(HDC_CHICKEN0, - HDC_FORCE_NON_COHERENT); - - /* WaDisableHDCInvalidation:skl,bxt,kbl,cfl */ - I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | - BDW_DISABLE_HDC_INVALIDATION); - - /* WaDisableSamplerPowerBypassForSOPingPong:skl,bxt,kbl,cfl */ - if (IS_SKYLAKE(dev_priv) || - IS_KABYLAKE(dev_priv) || - IS_COFFEELAKE(dev_priv)) - WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, - GEN8_SAMPLER_POWER_BYPASS_DIS); - - /* WaDisableSTUnitPowerOptimization:skl,bxt,kbl,glk,cfl */ - WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN2, GEN8_ST_PO_DISABLE); - - /* WaProgramL3SqcReg1DefaultForPerf:bxt,glk */ - if (IS_GEN9_LP(dev_priv)) { - u32 val = I915_READ(GEN8_L3SQCREG1); - - val &= ~L3_PRIO_CREDITS_MASK; - val |= L3_GENERAL_PRIO_CREDITS(62) | L3_HIGH_PRIO_CREDITS(2); - I915_WRITE(GEN8_L3SQCREG1, val); - } - - /* WaOCLCoherentLineFlush:skl,bxt,kbl,cfl */ - I915_WRITE(GEN8_L3SQCREG4, (I915_READ(GEN8_L3SQCREG4) | - GEN8_LQSC_FLUSH_COHERENT_LINES)); - - /* - * Supporting preemption with fine-granularity requires changes in the - * batch buffer programming. Since we can't break old userspace, we - * need to set our default preemption level to safe value. Userspace is - * still able to use more fine-grained preemption levels, since in - * WaEnablePreemptionGranularityControlByUMD we're whitelisting the - * per-ctx register. As such, WaDisable{3D,GPGPU}MidCmdPreemption are - * not real HW workarounds, but merely a way to start using preemption - * while maintaining old contract with userspace. - */ - - /* WaDisable3DMidCmdPreemption:skl,bxt,glk,cfl,[cnl] */ - WA_CLR_BIT_MASKED(GEN8_CS_CHICKEN1, GEN9_PREEMPT_3D_OBJECT_LEVEL); - - /* WaDisableGPGPUMidCmdPreemption:skl,bxt,blk,cfl,[cnl] */ - WA_SET_FIELD_MASKED(GEN8_CS_CHICKEN1, GEN9_PREEMPT_GPGPU_LEVEL_MASK, - GEN9_PREEMPT_GPGPU_COMMAND_LEVEL); - - /* WaClearHIZ_WM_CHICKEN3:bxt,glk */ - if (IS_GEN9_LP(dev_priv)) - WA_SET_BIT_MASKED(GEN9_WM_CHICKEN3, GEN9_FACTOR_IN_CLR_VAL_HIZ); - - /* WaVFEStateAfterPipeControlwithMediaStateClear:skl,bxt,glk,cfl */ - ret = wa_ring_whitelist_reg(engine, GEN9_CTX_PREEMPT_REG); - if (ret) - return ret; - - /* WaEnablePreemptionGranularityControlByUMD:skl,bxt,kbl,cfl,[cnl] */ - I915_WRITE(GEN7_FF_SLICE_CS_CHICKEN1, - _MASKED_BIT_ENABLE(GEN9_FFSC_PERCTX_PREEMPT_CTRL)); - ret = wa_ring_whitelist_reg(engine, GEN8_CS_CHICKEN1); - if (ret) - return ret; - - /* WaAllowUMDToModifyHDCChicken1:skl,bxt,kbl,glk,cfl */ - ret = wa_ring_whitelist_reg(engine, GEN8_HDC_CHICKEN1); - if (ret) - return ret; - - return 0; -} - -static int skl_tune_iz_hashing(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - u8 vals[3] = { 0, 0, 0 }; - unsigned int i; - - for (i = 0; i < 3; i++) { - u8 ss; - - /* - * Only consider slices where one, and only one, subslice has 7 - * EUs - */ - if (!is_power_of_2(INTEL_INFO(dev_priv)->sseu.subslice_7eu[i])) - continue; - - /* - * subslice_7eu[i] != 0 (because of the check above) and - * ss_max == 4 (maximum number of subslices possible per slice) - * - * -> 0 <= ss <= 3; - */ - ss = ffs(INTEL_INFO(dev_priv)->sseu.subslice_7eu[i]) - 1; - vals[i] = 3 - ss; - } - - if (vals[0] == 0 && vals[1] == 0 && vals[2] == 0) - return 0; - - /* Tune IZ hashing. See intel_device_info_runtime_init() */ - WA_SET_FIELD_MASKED(GEN7_GT_MODE, - GEN9_IZ_HASHING_MASK(2) | - GEN9_IZ_HASHING_MASK(1) | - GEN9_IZ_HASHING_MASK(0), - GEN9_IZ_HASHING(2, vals[2]) | - GEN9_IZ_HASHING(1, vals[1]) | - GEN9_IZ_HASHING(0, vals[0])); - - return 0; -} - -static int skl_init_workarounds(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int ret; - - ret = gen9_init_workarounds(engine); - if (ret) - return ret; - - /* WaEnableGapsTsvCreditFix:skl */ - I915_WRITE(GEN8_GARBCNTL, (I915_READ(GEN8_GARBCNTL) | - GEN9_GAPS_TSV_CREDIT_DISABLE)); - - /* WaDisableGafsUnitClkGating:skl */ - I915_WRITE(GEN7_UCGCTL4, (I915_READ(GEN7_UCGCTL4) | - GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE)); - - /* WaInPlaceDecompressionHang:skl */ - if (IS_SKL_REVID(dev_priv, SKL_REVID_H0, REVID_FOREVER)) - I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, - (I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | - GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS)); - - /* WaDisableLSQCROPERFforOCL:skl */ - ret = wa_ring_whitelist_reg(engine, GEN8_L3SQCREG4); - if (ret) - return ret; - - return skl_tune_iz_hashing(engine); -} - -static int bxt_init_workarounds(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int ret; - - ret = gen9_init_workarounds(engine); - if (ret) - return ret; - - /* WaDisableThreadStallDopClockGating:bxt */ - WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, - STALL_DOP_GATING_DISABLE); - - /* WaDisablePooledEuLoadBalancingFix:bxt */ - I915_WRITE(FF_SLICE_CS_CHICKEN2, - _MASKED_BIT_ENABLE(GEN9_POOLED_EU_LOAD_BALANCING_FIX_DISABLE)); - - /* WaToEnableHwFixForPushConstHWBug:bxt */ - WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, - GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); - - /* WaInPlaceDecompressionHang:bxt */ - I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, - (I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | - GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS)); - - return 0; -} - -static int cnl_init_workarounds(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int ret; - - /* WaDisableI2mCycleOnWRPort:cnl (pre-prod) */ - if (IS_CNL_REVID(dev_priv, CNL_REVID_B0, CNL_REVID_B0)) - I915_WRITE(GAMT_CHKN_BIT_REG, - (I915_READ(GAMT_CHKN_BIT_REG) | - GAMT_CHKN_DISABLE_I2M_CYCLE_ON_WR_PORT)); - - /* WaForceContextSaveRestoreNonCoherent:cnl */ - WA_SET_BIT_MASKED(CNL_HDC_CHICKEN0, - HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT); - - /* WaThrottleEUPerfToAvoidTDBackPressure:cnl(pre-prod) */ - if (IS_CNL_REVID(dev_priv, CNL_REVID_B0, CNL_REVID_B0)) - WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, THROTTLE_12_5); - - /* WaDisableReplayBufferBankArbitrationOptimization:cnl */ - WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, - GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); - - /* WaDisableEnhancedSBEVertexCaching:cnl (pre-prod) */ - if (IS_CNL_REVID(dev_priv, 0, CNL_REVID_B0)) - WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, - GEN8_CSC2_SBE_VUE_CACHE_CONSERVATIVE); - - /* WaInPlaceDecompressionHang:cnl */ - I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, - (I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | - GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS)); - - /* WaPushConstantDereferenceHoldDisable:cnl */ - WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2, PUSH_CONSTANT_DEREF_DISABLE); - - /* FtrEnableFastAnisoL1BankingFix: cnl */ - WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, CNL_FAST_ANISO_L1_BANKING_FIX); - - /* WaDisable3DMidCmdPreemption:cnl */ - WA_CLR_BIT_MASKED(GEN8_CS_CHICKEN1, GEN9_PREEMPT_3D_OBJECT_LEVEL); - - /* WaDisableGPGPUMidCmdPreemption:cnl */ - WA_SET_FIELD_MASKED(GEN8_CS_CHICKEN1, GEN9_PREEMPT_GPGPU_LEVEL_MASK, - GEN9_PREEMPT_GPGPU_COMMAND_LEVEL); - - /* WaEnablePreemptionGranularityControlByUMD:cnl */ - I915_WRITE(GEN7_FF_SLICE_CS_CHICKEN1, - _MASKED_BIT_ENABLE(GEN9_FFSC_PERCTX_PREEMPT_CTRL)); - ret= wa_ring_whitelist_reg(engine, GEN8_CS_CHICKEN1); - if (ret) - return ret; - - /* WaDisableEarlyEOT:cnl */ - WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, DISABLE_EARLY_EOT); - - return 0; -} - -static int kbl_init_workarounds(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int ret; - - ret = gen9_init_workarounds(engine); - if (ret) - return ret; - - /* WaEnableGapsTsvCreditFix:kbl */ - I915_WRITE(GEN8_GARBCNTL, (I915_READ(GEN8_GARBCNTL) | - GEN9_GAPS_TSV_CREDIT_DISABLE)); - - /* WaDisableDynamicCreditSharing:kbl */ - if (IS_KBL_REVID(dev_priv, 0, KBL_REVID_B0)) - I915_WRITE(GAMT_CHKN_BIT_REG, - (I915_READ(GAMT_CHKN_BIT_REG) | - GAMT_CHKN_DISABLE_DYNAMIC_CREDIT_SHARING)); - - /* WaDisableFenceDestinationToSLM:kbl (pre-prod) */ - if (IS_KBL_REVID(dev_priv, KBL_REVID_A0, KBL_REVID_A0)) - WA_SET_BIT_MASKED(HDC_CHICKEN0, - HDC_FENCE_DEST_SLM_DISABLE); - - /* WaToEnableHwFixForPushConstHWBug:kbl */ - if (IS_KBL_REVID(dev_priv, KBL_REVID_C0, REVID_FOREVER)) - WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, - GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); - - /* WaDisableGafsUnitClkGating:kbl */ - I915_WRITE(GEN7_UCGCTL4, (I915_READ(GEN7_UCGCTL4) | - GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE)); - - /* WaDisableSbeCacheDispatchPortSharing:kbl */ - WA_SET_BIT_MASKED( - GEN7_HALF_SLICE_CHICKEN1, - GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE); - - /* WaInPlaceDecompressionHang:kbl */ - I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, - (I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | - GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS)); - - /* WaDisableLSQCROPERFforOCL:kbl */ - ret = wa_ring_whitelist_reg(engine, GEN8_L3SQCREG4); - if (ret) - return ret; - - return 0; -} - -static int glk_init_workarounds(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int ret; - - ret = gen9_init_workarounds(engine); - if (ret) - return ret; - - /* WA #0862: Userspace has to set "Barrier Mode" to avoid hangs. */ - ret = wa_ring_whitelist_reg(engine, GEN9_SLICE_COMMON_ECO_CHICKEN1); - if (ret) - return ret; - - /* WaToEnableHwFixForPushConstHWBug:glk */ - WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, - GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); - - return 0; -} - -static int cfl_init_workarounds(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int ret; - - ret = gen9_init_workarounds(engine); - if (ret) - return ret; - - /* WaEnableGapsTsvCreditFix:cfl */ - I915_WRITE(GEN8_GARBCNTL, (I915_READ(GEN8_GARBCNTL) | - GEN9_GAPS_TSV_CREDIT_DISABLE)); - - /* WaToEnableHwFixForPushConstHWBug:cfl */ - WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, - GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); - - /* WaDisableGafsUnitClkGating:cfl */ - I915_WRITE(GEN7_UCGCTL4, (I915_READ(GEN7_UCGCTL4) | - GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE)); - - /* WaDisableSbeCacheDispatchPortSharing:cfl */ - WA_SET_BIT_MASKED( - GEN7_HALF_SLICE_CHICKEN1, - GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE); - - /* WaInPlaceDecompressionHang:cfl */ - I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, - (I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | - GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS)); - - return 0; -} - -int init_workarounds_ring(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int err; - - if (GEM_WARN_ON(engine->id != RCS)) - return -EINVAL; - - dev_priv->workarounds.count = 0; - dev_priv->workarounds.hw_whitelist_count[engine->id] = 0; - - if (IS_BROADWELL(dev_priv)) - err = bdw_init_workarounds(engine); - else if (IS_CHERRYVIEW(dev_priv)) - err = chv_init_workarounds(engine); - else if (IS_SKYLAKE(dev_priv)) - err = skl_init_workarounds(engine); - else if (IS_BROXTON(dev_priv)) - err = bxt_init_workarounds(engine); - else if (IS_KABYLAKE(dev_priv)) - err = kbl_init_workarounds(engine); - else if (IS_GEMINILAKE(dev_priv)) - err = glk_init_workarounds(engine); - else if (IS_COFFEELAKE(dev_priv)) - err = cfl_init_workarounds(engine); - else if (IS_CANNONLAKE(dev_priv)) - err = cnl_init_workarounds(engine); - else - err = 0; - if (err) - return err; - - DRM_DEBUG_DRIVER("%s: Number of context specific w/a: %d\n", - engine->name, dev_priv->workarounds.count); - return 0; -} - -int intel_ring_workarounds_emit(struct i915_request *rq) -{ - struct i915_workarounds *w = &rq->i915->workarounds; - u32 *cs; - int ret, i; - - if (w->count == 0) - return 0; - - ret = rq->engine->emit_flush(rq, EMIT_BARRIER); - if (ret) - return ret; - - cs = intel_ring_begin(rq, w->count * 2 + 2); - if (IS_ERR(cs)) - return PTR_ERR(cs); - - *cs++ = MI_LOAD_REGISTER_IMM(w->count); - for (i = 0; i < w->count; i++) { - *cs++ = i915_mmio_reg_offset(w->reg[i].addr); - *cs++ = w->reg[i].value; - } - *cs++ = MI_NOOP; - - intel_ring_advance(rq, cs); - - ret = rq->engine->emit_flush(rq, EMIT_BARRIER); - if (ret) - return ret; - - return 0; -} - static bool ring_is_idle(struct intel_engine_cs *engine) { struct drm_i915_private *dev_priv = engine->i915; @@ -1611,7 +989,7 @@ bool intel_engine_has_kernel_context(const struct intel_engine_cs *engine) * the last request that remains in the timeline. When idle, it is * the last executed context as tracked by retirement. */ - rq = __i915_gem_active_peek(&engine->timeline->last_request); + rq = __i915_gem_active_peek(&engine->timeline.last_request); if (rq) return rq->ctx == kernel_context; else @@ -1659,6 +1037,9 @@ void intel_engines_park(struct drm_i915_private *i915) intel_engine_dump(engine, &p, NULL); } + /* Must be reset upon idling, or we may miss the busy wakeup. */ + GEM_BUG_ON(engine->execlists.queue_priority != INT_MIN); + if (engine->park) engine->park(engine); @@ -1681,6 +1062,8 @@ void intel_engines_unpark(struct drm_i915_private *i915) for_each_engine(engine, i915, id) { if (engine->unpark) engine->unpark(engine); + + intel_engine_init_hangcheck(engine); } } @@ -1713,17 +1096,37 @@ unsigned int intel_engines_has_context_isolation(struct drm_i915_private *i915) return which; } +static int print_sched_attr(struct drm_i915_private *i915, + const struct i915_sched_attr *attr, + char *buf, int x, int len) +{ + if (attr->priority == I915_PRIORITY_INVALID) + return x; + + x += snprintf(buf + x, len - x, + " prio=%d", attr->priority); + + return x; +} + static void print_request(struct drm_printer *m, struct i915_request *rq, const char *prefix) { - drm_printf(m, "%s%x%s [%llx:%x] prio=%d @ %dms: %s\n", prefix, + const char *name = rq->fence.ops->get_timeline_name(&rq->fence); + char buf[80]; + int x = 0; + + x = print_sched_attr(rq->i915, &rq->sched.attr, buf, x, sizeof(buf)); + + drm_printf(m, "%s%x%s [%llx:%x]%s @ %dms: %s\n", + prefix, rq->global_seqno, i915_request_completed(rq) ? "!" : "", rq->fence.context, rq->fence.seqno, - rq->priotree.priority, + buf, jiffies_to_msecs(jiffies - rq->emitted_jiffies), - rq->timeline->common->name); + name); } static void hexdump(struct drm_printer *m, const void *buf, size_t len) @@ -1829,12 +1232,15 @@ static void intel_engine_print_registers(const struct intel_engine_cs *engine, ptr = I915_READ(RING_CONTEXT_STATUS_PTR(engine)); read = GEN8_CSB_READ_PTR(ptr); write = GEN8_CSB_WRITE_PTR(ptr); - drm_printf(m, "\tExeclist CSB read %d [%d cached], write %d [%d from hws], interrupt posted? %s\n", + drm_printf(m, "\tExeclist CSB read %d [%d cached], write %d [%d from hws], interrupt posted? %s, tasklet queued? %s (%s)\n", read, execlists->csb_head, write, intel_read_status_page(engine, intel_hws_csb_write_index(engine->i915)), yesno(test_bit(ENGINE_IRQ_EXECLIST, - &engine->irq_posted))); + &engine->irq_posted)), + yesno(test_bit(TASKLET_STATE_SCHED, + &engine->execlists.tasklet.state)), + enableddisabled(!atomic_read(&engine->execlists.tasklet.count))); if (read >= GEN8_CSB_ENTRIES) read = 0; if (write >= GEN8_CSB_ENTRIES) @@ -1861,8 +1267,9 @@ static void intel_engine_print_registers(const struct intel_engine_cs *engine, char hdr[80]; snprintf(hdr, sizeof(hdr), - "\t\tELSP[%d] count=%d, rq: ", - idx, count); + "\t\tELSP[%d] count=%d, ring->start=%08x, rq: ", + idx, count, + i915_ggtt_offset(rq->ring->vma)); print_request(m, rq, hdr); } else { drm_printf(m, "\t\tELSP[%d] idle\n", idx); @@ -1884,11 +1291,13 @@ void intel_engine_dump(struct intel_engine_cs *engine, struct drm_printer *m, const char *header, ...) { + const int MAX_REQUESTS_TO_SHOW = 8; struct intel_breadcrumbs * const b = &engine->breadcrumbs; const struct intel_engine_execlists * const execlists = &engine->execlists; struct i915_gpu_error * const error = &engine->i915->gpu_error; - struct i915_request *rq; + struct i915_request *rq, *last; struct rb_node *rb; + int count; if (header) { va_list ap; @@ -1901,12 +1310,11 @@ void intel_engine_dump(struct intel_engine_cs *engine, if (i915_terminally_wedged(&engine->i915->gpu_error)) drm_printf(m, "*** WEDGED ***\n"); - drm_printf(m, "\tcurrent seqno %x, last %x, hangcheck %x [%d ms], inflight %d\n", + drm_printf(m, "\tcurrent seqno %x, last %x, hangcheck %x [%d ms]\n", intel_engine_get_seqno(engine), intel_engine_last_submit(engine), engine->hangcheck.seqno, - jiffies_to_msecs(jiffies - engine->hangcheck.action_timestamp), - engine->timeline->inflight_seqnos); + jiffies_to_msecs(jiffies - engine->hangcheck.action_timestamp)); drm_printf(m, "\tReset count: %d (global %d)\n", i915_reset_engine_count(error, engine), i915_reset_count(error)); @@ -1915,14 +1323,14 @@ void intel_engine_dump(struct intel_engine_cs *engine, drm_printf(m, "\tRequests:\n"); - rq = list_first_entry(&engine->timeline->requests, + rq = list_first_entry(&engine->timeline.requests, struct i915_request, link); - if (&rq->link != &engine->timeline->requests) + if (&rq->link != &engine->timeline.requests) print_request(m, rq, "\t\tfirst "); - rq = list_last_entry(&engine->timeline->requests, + rq = list_last_entry(&engine->timeline.requests, struct i915_request, link); - if (&rq->link != &engine->timeline->requests) + if (&rq->link != &engine->timeline.requests) print_request(m, rq, "\t\tlast "); rq = i915_gem_find_active_request(engine); @@ -1933,12 +1341,16 @@ void intel_engine_dump(struct intel_engine_cs *engine, rq->head, rq->postfix, rq->tail, rq->batch ? upper_32_bits(rq->batch->node.start) : ~0u, rq->batch ? lower_32_bits(rq->batch->node.start) : ~0u); - drm_printf(m, "\t\tring->start: 0x%08x\n", + drm_printf(m, "\t\tring->start: 0x%08x\n", i915_ggtt_offset(rq->ring->vma)); - drm_printf(m, "\t\tring->head: 0x%08x\n", + drm_printf(m, "\t\tring->head: 0x%08x\n", rq->ring->head); - drm_printf(m, "\t\tring->tail: 0x%08x\n", + drm_printf(m, "\t\tring->tail: 0x%08x\n", rq->ring->tail); + drm_printf(m, "\t\tring->emit: 0x%08x\n", + rq->ring->emit); + drm_printf(m, "\t\tring->space: 0x%08x\n", + rq->ring->space); } rcu_read_unlock(); @@ -1950,18 +1362,49 @@ void intel_engine_dump(struct intel_engine_cs *engine, drm_printf(m, "\tDevice is asleep; skipping register dump\n"); } - spin_lock_irq(&engine->timeline->lock); - list_for_each_entry(rq, &engine->timeline->requests, link) - print_request(m, rq, "\t\tE "); + spin_lock_irq(&engine->timeline.lock); + + last = NULL; + count = 0; + list_for_each_entry(rq, &engine->timeline.requests, link) { + if (count++ < MAX_REQUESTS_TO_SHOW - 1) + print_request(m, rq, "\t\tE "); + else + last = rq; + } + if (last) { + if (count > MAX_REQUESTS_TO_SHOW) { + drm_printf(m, + "\t\t...skipping %d executing requests...\n", + count - MAX_REQUESTS_TO_SHOW); + } + print_request(m, last, "\t\tE "); + } + + last = NULL; + count = 0; drm_printf(m, "\t\tQueue priority: %d\n", execlists->queue_priority); for (rb = execlists->first; rb; rb = rb_next(rb)) { struct i915_priolist *p = rb_entry(rb, typeof(*p), node); - list_for_each_entry(rq, &p->requests, priotree.link) - print_request(m, rq, "\t\tQ "); + list_for_each_entry(rq, &p->requests, sched.link) { + if (count++ < MAX_REQUESTS_TO_SHOW - 1) + print_request(m, rq, "\t\tQ "); + else + last = rq; + } + } + if (last) { + if (count > MAX_REQUESTS_TO_SHOW) { + drm_printf(m, + "\t\t...skipping %d queued requests...\n", + count - MAX_REQUESTS_TO_SHOW); + } + print_request(m, last, "\t\tQ "); } - spin_unlock_irq(&engine->timeline->lock); + + spin_unlock_irq(&engine->timeline.lock); spin_lock_irq(&b->rb_lock); for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) { @@ -2026,7 +1469,7 @@ int intel_enable_engine_stats(struct intel_engine_cs *engine) return -ENODEV; tasklet_disable(&execlists->tasklet); - spin_lock_irqsave(&engine->stats.lock, flags); + write_seqlock_irqsave(&engine->stats.lock, flags); if (unlikely(engine->stats.enabled == ~0)) { err = -EBUSY; @@ -2050,7 +1493,7 @@ int intel_enable_engine_stats(struct intel_engine_cs *engine) } unlock: - spin_unlock_irqrestore(&engine->stats.lock, flags); + write_sequnlock_irqrestore(&engine->stats.lock, flags); tasklet_enable(&execlists->tasklet); return err; @@ -2079,12 +1522,13 @@ static ktime_t __intel_engine_get_busy_time(struct intel_engine_cs *engine) */ ktime_t intel_engine_get_busy_time(struct intel_engine_cs *engine) { + unsigned int seq; ktime_t total; - unsigned long flags; - spin_lock_irqsave(&engine->stats.lock, flags); - total = __intel_engine_get_busy_time(engine); - spin_unlock_irqrestore(&engine->stats.lock, flags); + do { + seq = read_seqbegin(&engine->stats.lock); + total = __intel_engine_get_busy_time(engine); + } while (read_seqretry(&engine->stats.lock, seq)); return total; } @@ -2102,15 +1546,16 @@ void intel_disable_engine_stats(struct intel_engine_cs *engine) if (!intel_engine_supports_stats(engine)) return; - spin_lock_irqsave(&engine->stats.lock, flags); + write_seqlock_irqsave(&engine->stats.lock, flags); WARN_ON_ONCE(engine->stats.enabled == 0); if (--engine->stats.enabled == 0) { engine->stats.total = __intel_engine_get_busy_time(engine); engine->stats.active = 0; } - spin_unlock_irqrestore(&engine->stats.lock, flags); + write_sequnlock_irqrestore(&engine->stats.lock, flags); } #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) #include "selftests/mock_engine.c" +#include "selftests/intel_engine_cs.c" #endif diff --git a/drivers/gpu/drm/i915/intel_fbc.c b/drivers/gpu/drm/i915/intel_fbc.c index 707d49c..b431b67 100644 --- a/drivers/gpu/drm/i915/intel_fbc.c +++ b/drivers/gpu/drm/i915/intel_fbc.c @@ -1272,6 +1272,34 @@ out: mutex_unlock(&fbc->lock); } +/* + * intel_fbc_reset_underrun - reset FBC fifo underrun status. + * @dev_priv: i915 device instance + * + * See intel_fbc_handle_fifo_underrun_irq(). For automated testing we + * want to re-enable FBC after an underrun to increase test coverage. + */ +int intel_fbc_reset_underrun(struct drm_i915_private *dev_priv) +{ + int ret; + + cancel_work_sync(&dev_priv->fbc.underrun_work); + + ret = mutex_lock_interruptible(&dev_priv->fbc.lock); + if (ret) + return ret; + + if (dev_priv->fbc.underrun_detected) { + DRM_DEBUG_KMS("Re-allowing FBC after fifo underrun\n"); + dev_priv->fbc.no_fbc_reason = "FIFO underrun cleared"; + } + + dev_priv->fbc.underrun_detected = false; + mutex_unlock(&dev_priv->fbc.lock); + + return 0; +} + /** * intel_fbc_handle_fifo_underrun_irq - disable FBC when we get a FIFO underrun * @dev_priv: i915 device instance diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c index 6467a5c..e9e02b5 100644 --- a/drivers/gpu/drm/i915/intel_fbdev.c +++ b/drivers/gpu/drm/i915/intel_fbdev.c @@ -221,6 +221,9 @@ static int intelfb_create(struct drm_fb_helper *helper, goto out_unlock; } + fb = &ifbdev->fb->base; + intel_fb_obj_flush(intel_fb_obj(fb), ORIGIN_DIRTYFB); + info = drm_fb_helper_alloc_fbi(helper); if (IS_ERR(info)) { DRM_ERROR("Failed to allocate fb_info\n"); @@ -230,8 +233,6 @@ static int intelfb_create(struct drm_fb_helper *helper, info->par = helper; - fb = &ifbdev->fb->base; - ifbdev->helper.fb = fb; strcpy(info->fix.id, "inteldrmfb"); @@ -640,7 +641,7 @@ static bool intel_fbdev_init_bios(struct drm_device *dev, if (!crtc->state->active) continue; - WARN(!crtc->primary->fb, + WARN(!crtc->primary->state->fb, "re-used BIOS config but lost an fb on crtc %d\n", crtc->base.id); } diff --git a/drivers/gpu/drm/i915/intel_frontbuffer.c b/drivers/gpu/drm/i915/intel_frontbuffer.c index 3a8d3d0..7fff0a0 100644 --- a/drivers/gpu/drm/i915/intel_frontbuffer.c +++ b/drivers/gpu/drm/i915/intel_frontbuffer.c @@ -80,7 +80,7 @@ void __intel_fb_obj_invalidate(struct drm_i915_gem_object *obj, } might_sleep(); - intel_psr_invalidate(dev_priv, frontbuffer_bits); + intel_psr_invalidate(dev_priv, frontbuffer_bits, origin); intel_edp_drrs_invalidate(dev_priv, frontbuffer_bits); intel_fbc_invalidate(dev_priv, frontbuffer_bits, origin); } diff --git a/drivers/gpu/drm/i915/intel_gpu_commands.h b/drivers/gpu/drm/i915/intel_gpu_commands.h new file mode 100644 index 0000000..105e2a9 --- /dev/null +++ b/drivers/gpu/drm/i915/intel_gpu_commands.h @@ -0,0 +1,274 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright � 2003-2018 Intel Corporation + */ + +#ifndef _INTEL_GPU_COMMANDS_H_ +#define _INTEL_GPU_COMMANDS_H_ + +/* + * Instruction field definitions used by the command parser + */ +#define INSTR_CLIENT_SHIFT 29 +#define INSTR_MI_CLIENT 0x0 +#define INSTR_BC_CLIENT 0x2 +#define INSTR_RC_CLIENT 0x3 +#define INSTR_SUBCLIENT_SHIFT 27 +#define INSTR_SUBCLIENT_MASK 0x18000000 +#define INSTR_MEDIA_SUBCLIENT 0x2 +#define INSTR_26_TO_24_MASK 0x7000000 +#define INSTR_26_TO_24_SHIFT 24 + +/* + * Memory interface instructions used by the kernel + */ +#define MI_INSTR(opcode, flags) (((opcode) << 23) | (flags)) +/* Many MI commands use bit 22 of the header dword for GGTT vs PPGTT */ +#define MI_GLOBAL_GTT (1<<22) + +#define MI_NOOP MI_INSTR(0, 0) +#define MI_USER_INTERRUPT MI_INSTR(0x02, 0) +#define MI_WAIT_FOR_EVENT MI_INSTR(0x03, 0) +#define MI_WAIT_FOR_OVERLAY_FLIP (1<<16) +#define MI_WAIT_FOR_PLANE_B_FLIP (1<<6) +#define MI_WAIT_FOR_PLANE_A_FLIP (1<<2) +#define MI_WAIT_FOR_PLANE_A_SCANLINES (1<<1) +#define MI_FLUSH MI_INSTR(0x04, 0) +#define MI_READ_FLUSH (1 << 0) +#define MI_EXE_FLUSH (1 << 1) +#define MI_NO_WRITE_FLUSH (1 << 2) +#define MI_SCENE_COUNT (1 << 3) /* just increment scene count */ +#define MI_END_SCENE (1 << 4) /* flush binner and incr scene count */ +#define MI_INVALIDATE_ISP (1 << 5) /* invalidate indirect state pointers */ +#define MI_REPORT_HEAD MI_INSTR(0x07, 0) +#define MI_ARB_ON_OFF MI_INSTR(0x08, 0) +#define MI_ARB_ENABLE (1<<0) +#define MI_ARB_DISABLE (0<<0) +#define MI_BATCH_BUFFER_END MI_INSTR(0x0a, 0) +#define MI_SUSPEND_FLUSH MI_INSTR(0x0b, 0) +#define MI_SUSPEND_FLUSH_EN (1<<0) +#define MI_SET_APPID MI_INSTR(0x0e, 0) +#define MI_OVERLAY_FLIP MI_INSTR(0x11, 0) +#define MI_OVERLAY_CONTINUE (0x0<<21) +#define MI_OVERLAY_ON (0x1<<21) +#define MI_OVERLAY_OFF (0x2<<21) +#define MI_LOAD_SCAN_LINES_INCL MI_INSTR(0x12, 0) +#define MI_DISPLAY_FLIP MI_INSTR(0x14, 2) +#define MI_DISPLAY_FLIP_I915 MI_INSTR(0x14, 1) +#define MI_DISPLAY_FLIP_PLANE(n) ((n) << 20) +/* IVB has funny definitions for which plane to flip. */ +#define MI_DISPLAY_FLIP_IVB_PLANE_A (0 << 19) +#define MI_DISPLAY_FLIP_IVB_PLANE_B (1 << 19) +#define MI_DISPLAY_FLIP_IVB_SPRITE_A (2 << 19) +#define MI_DISPLAY_FLIP_IVB_SPRITE_B (3 << 19) +#define MI_DISPLAY_FLIP_IVB_PLANE_C (4 << 19) +#define MI_DISPLAY_FLIP_IVB_SPRITE_C (5 << 19) +/* SKL ones */ +#define MI_DISPLAY_FLIP_SKL_PLANE_1_A (0 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_1_B (1 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_1_C (2 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_2_A (4 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_2_B (5 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_2_C (6 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_3_A (7 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_3_B (8 << 8) +#define MI_DISPLAY_FLIP_SKL_PLANE_3_C (9 << 8) +#define MI_SEMAPHORE_MBOX MI_INSTR(0x16, 1) /* gen6, gen7 */ +#define MI_SEMAPHORE_GLOBAL_GTT (1<<22) +#define MI_SEMAPHORE_UPDATE (1<<21) +#define MI_SEMAPHORE_COMPARE (1<<20) +#define MI_SEMAPHORE_REGISTER (1<<18) +#define MI_SEMAPHORE_SYNC_VR (0<<16) /* RCS wait for VCS (RVSYNC) */ +#define MI_SEMAPHORE_SYNC_VER (1<<16) /* RCS wait for VECS (RVESYNC) */ +#define MI_SEMAPHORE_SYNC_BR (2<<16) /* RCS wait for BCS (RBSYNC) */ +#define MI_SEMAPHORE_SYNC_BV (0<<16) /* VCS wait for BCS (VBSYNC) */ +#define MI_SEMAPHORE_SYNC_VEV (1<<16) /* VCS wait for VECS (VVESYNC) */ +#define MI_SEMAPHORE_SYNC_RV (2<<16) /* VCS wait for RCS (VRSYNC) */ +#define MI_SEMAPHORE_SYNC_RB (0<<16) /* BCS wait for RCS (BRSYNC) */ +#define MI_SEMAPHORE_SYNC_VEB (1<<16) /* BCS wait for VECS (BVESYNC) */ +#define MI_SEMAPHORE_SYNC_VB (2<<16) /* BCS wait for VCS (BVSYNC) */ +#define MI_SEMAPHORE_SYNC_BVE (0<<16) /* VECS wait for BCS (VEBSYNC) */ +#define MI_SEMAPHORE_SYNC_VVE (1<<16) /* VECS wait for VCS (VEVSYNC) */ +#define MI_SEMAPHORE_SYNC_RVE (2<<16) /* VECS wait for RCS (VERSYNC) */ +#define MI_SEMAPHORE_SYNC_INVALID (3<<16) +#define MI_SEMAPHORE_SYNC_MASK (3<<16) +#define MI_SET_CONTEXT MI_INSTR(0x18, 0) +#define MI_MM_SPACE_GTT (1<<8) +#define MI_MM_SPACE_PHYSICAL (0<<8) +#define MI_SAVE_EXT_STATE_EN (1<<3) +#define MI_RESTORE_EXT_STATE_EN (1<<2) +#define MI_FORCE_RESTORE (1<<1) +#define MI_RESTORE_INHIBIT (1<<0) +#define HSW_MI_RS_SAVE_STATE_EN (1<<3) +#define HSW_MI_RS_RESTORE_STATE_EN (1<<2) +#define MI_SEMAPHORE_SIGNAL MI_INSTR(0x1b, 0) /* GEN8+ */ +#define MI_SEMAPHORE_TARGET(engine) ((engine)<<15) +#define MI_SEMAPHORE_WAIT MI_INSTR(0x1c, 2) /* GEN8+ */ +#define MI_SEMAPHORE_POLL (1<<15) +#define MI_SEMAPHORE_SAD_GTE_SDD (1<<12) +#define MI_STORE_DWORD_IMM MI_INSTR(0x20, 1) +#define MI_STORE_DWORD_IMM_GEN4 MI_INSTR(0x20, 2) +#define MI_MEM_VIRTUAL (1 << 22) /* 945,g33,965 */ +#define MI_USE_GGTT (1 << 22) /* g4x+ */ +#define MI_STORE_DWORD_INDEX MI_INSTR(0x21, 1) +#define MI_STORE_DWORD_INDEX_SHIFT 2 +/* + * Official intel docs are somewhat sloppy concerning MI_LOAD_REGISTER_IMM: + * - Always issue a MI_NOOP _before_ the MI_LOAD_REGISTER_IMM - otherwise hw + * simply ignores the register load under certain conditions. + * - One can actually load arbitrary many arbitrary registers: Simply issue x + * address/value pairs. Don't overdue it, though, x <= 2^4 must hold! + */ +#define MI_LOAD_REGISTER_IMM(x) MI_INSTR(0x22, 2*(x)-1) +#define MI_LRI_FORCE_POSTED (1<<12) +#define MI_STORE_REGISTER_MEM MI_INSTR(0x24, 1) +#define MI_STORE_REGISTER_MEM_GEN8 MI_INSTR(0x24, 2) +#define MI_SRM_LRM_GLOBAL_GTT (1<<22) +#define MI_FLUSH_DW MI_INSTR(0x26, 1) /* for GEN6 */ +#define MI_FLUSH_DW_STORE_INDEX (1<<21) +#define MI_INVALIDATE_TLB (1<<18) +#define MI_FLUSH_DW_OP_STOREDW (1<<14) +#define MI_FLUSH_DW_OP_MASK (3<<14) +#define MI_FLUSH_DW_NOTIFY (1<<8) +#define MI_INVALIDATE_BSD (1<<7) +#define MI_FLUSH_DW_USE_GTT (1<<2) +#define MI_FLUSH_DW_USE_PPGTT (0<<2) +#define MI_LOAD_REGISTER_MEM MI_INSTR(0x29, 1) +#define MI_LOAD_REGISTER_MEM_GEN8 MI_INSTR(0x29, 2) +#define MI_BATCH_BUFFER MI_INSTR(0x30, 1) +#define MI_BATCH_NON_SECURE (1) +/* for snb/ivb/vlv this also means "batch in ppgtt" when ppgtt is enabled. */ +#define MI_BATCH_NON_SECURE_I965 (1<<8) +#define MI_BATCH_PPGTT_HSW (1<<8) +#define MI_BATCH_NON_SECURE_HSW (1<<13) +#define MI_BATCH_BUFFER_START MI_INSTR(0x31, 0) +#define MI_BATCH_GTT (2<<6) /* aliased with (1<<7) on gen4 */ +#define MI_BATCH_BUFFER_START_GEN8 MI_INSTR(0x31, 1) +#define MI_BATCH_RESOURCE_STREAMER (1<<10) + +/* + * 3D instructions used by the kernel + */ +#define GFX_INSTR(opcode, flags) ((0x3 << 29) | ((opcode) << 24) | (flags)) + +#define GEN9_MEDIA_POOL_STATE ((0x3 << 29) | (0x2 << 27) | (0x5 << 16) | 4) +#define GEN9_MEDIA_POOL_ENABLE (1 << 31) +#define GFX_OP_RASTER_RULES ((0x3<<29)|(0x7<<24)) +#define GFX_OP_SCISSOR ((0x3<<29)|(0x1c<<24)|(0x10<<19)) +#define SC_UPDATE_SCISSOR (0x1<<1) +#define SC_ENABLE_MASK (0x1<<0) +#define SC_ENABLE (0x1<<0) +#define GFX_OP_LOAD_INDIRECT ((0x3<<29)|(0x1d<<24)|(0x7<<16)) +#define GFX_OP_SCISSOR_INFO ((0x3<<29)|(0x1d<<24)|(0x81<<16)|(0x1)) +#define SCI_YMIN_MASK (0xffff<<16) +#define SCI_XMIN_MASK (0xffff<<0) +#define SCI_YMAX_MASK (0xffff<<16) +#define SCI_XMAX_MASK (0xffff<<0) +#define GFX_OP_SCISSOR_ENABLE ((0x3<<29)|(0x1c<<24)|(0x10<<19)) +#define GFX_OP_SCISSOR_RECT ((0x3<<29)|(0x1d<<24)|(0x81<<16)|1) +#define GFX_OP_COLOR_FACTOR ((0x3<<29)|(0x1d<<24)|(0x1<<16)|0x0) +#define GFX_OP_STIPPLE ((0x3<<29)|(0x1d<<24)|(0x83<<16)) +#define GFX_OP_MAP_INFO ((0x3<<29)|(0x1d<<24)|0x4) +#define GFX_OP_DESTBUFFER_VARS ((0x3<<29)|(0x1d<<24)|(0x85<<16)|0x0) +#define GFX_OP_DESTBUFFER_INFO ((0x3<<29)|(0x1d<<24)|(0x8e<<16)|1) +#define GFX_OP_DRAWRECT_INFO ((0x3<<29)|(0x1d<<24)|(0x80<<16)|(0x3)) +#define GFX_OP_DRAWRECT_INFO_I965 ((0x7900<<16)|0x2) + +#define COLOR_BLT_CMD (2<<29 | 0x40<<22 | (5-2)) +#define SRC_COPY_BLT_CMD ((2<<29)|(0x43<<22)|4) +#define XY_SRC_COPY_BLT_CMD ((2<<29)|(0x53<<22)|6) +#define XY_MONO_SRC_COPY_IMM_BLT ((2<<29)|(0x71<<22)|5) +#define BLT_WRITE_A (2<<20) +#define BLT_WRITE_RGB (1<<20) +#define BLT_WRITE_RGBA (BLT_WRITE_RGB | BLT_WRITE_A) +#define BLT_DEPTH_8 (0<<24) +#define BLT_DEPTH_16_565 (1<<24) +#define BLT_DEPTH_16_1555 (2<<24) +#define BLT_DEPTH_32 (3<<24) +#define BLT_ROP_SRC_COPY (0xcc<<16) +#define BLT_ROP_COLOR_COPY (0xf0<<16) +#define XY_SRC_COPY_BLT_SRC_TILED (1<<15) /* 965+ only */ +#define XY_SRC_COPY_BLT_DST_TILED (1<<11) /* 965+ only */ +#define CMD_OP_DISPLAYBUFFER_INFO ((0x0<<29)|(0x14<<23)|2) +#define ASYNC_FLIP (1<<22) +#define DISPLAY_PLANE_A (0<<20) +#define DISPLAY_PLANE_B (1<<20) +#define GFX_OP_PIPE_CONTROL(len) ((0x3<<29)|(0x3<<27)|(0x2<<24)|((len)-2)) +#define PIPE_CONTROL_FLUSH_L3 (1<<27) +#define PIPE_CONTROL_GLOBAL_GTT_IVB (1<<24) /* gen7+ */ +#define PIPE_CONTROL_MMIO_WRITE (1<<23) +#define PIPE_CONTROL_STORE_DATA_INDEX (1<<21) +#define PIPE_CONTROL_CS_STALL (1<<20) +#define PIPE_CONTROL_TLB_INVALIDATE (1<<18) +#define PIPE_CONTROL_MEDIA_STATE_CLEAR (1<<16) +#define PIPE_CONTROL_QW_WRITE (1<<14) +#define PIPE_CONTROL_POST_SYNC_OP_MASK (3<<14) +#define PIPE_CONTROL_DEPTH_STALL (1<<13) +#define PIPE_CONTROL_WRITE_FLUSH (1<<12) +#define PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH (1<<12) /* gen6+ */ +#define PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE (1<<11) /* MBZ on ILK */ +#define PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE (1<<10) /* GM45+ only */ +#define PIPE_CONTROL_INDIRECT_STATE_DISABLE (1<<9) +#define PIPE_CONTROL_NOTIFY (1<<8) +#define PIPE_CONTROL_FLUSH_ENABLE (1<<7) /* gen7+ */ +#define PIPE_CONTROL_DC_FLUSH_ENABLE (1<<5) +#define PIPE_CONTROL_VF_CACHE_INVALIDATE (1<<4) +#define PIPE_CONTROL_CONST_CACHE_INVALIDATE (1<<3) +#define PIPE_CONTROL_STATE_CACHE_INVALIDATE (1<<2) +#define PIPE_CONTROL_STALL_AT_SCOREBOARD (1<<1) +#define PIPE_CONTROL_DEPTH_CACHE_FLUSH (1<<0) +#define PIPE_CONTROL_GLOBAL_GTT (1<<2) /* in addr dword */ + +/* + * Commands used only by the command parser + */ +#define MI_SET_PREDICATE MI_INSTR(0x01, 0) +#define MI_ARB_CHECK MI_INSTR(0x05, 0) +#define MI_RS_CONTROL MI_INSTR(0x06, 0) +#define MI_URB_ATOMIC_ALLOC MI_INSTR(0x09, 0) +#define MI_PREDICATE MI_INSTR(0x0C, 0) +#define MI_RS_CONTEXT MI_INSTR(0x0F, 0) +#define MI_TOPOLOGY_FILTER MI_INSTR(0x0D, 0) +#define MI_LOAD_SCAN_LINES_EXCL MI_INSTR(0x13, 0) +#define MI_URB_CLEAR MI_INSTR(0x19, 0) +#define MI_UPDATE_GTT MI_INSTR(0x23, 0) +#define MI_CLFLUSH MI_INSTR(0x27, 0) +#define MI_REPORT_PERF_COUNT MI_INSTR(0x28, 0) +#define MI_REPORT_PERF_COUNT_GGTT (1<<0) +#define MI_LOAD_REGISTER_REG MI_INSTR(0x2A, 0) +#define MI_RS_STORE_DATA_IMM MI_INSTR(0x2B, 0) +#define MI_LOAD_URB_MEM MI_INSTR(0x2C, 0) +#define MI_STORE_URB_MEM MI_INSTR(0x2D, 0) +#define MI_CONDITIONAL_BATCH_BUFFER_END MI_INSTR(0x36, 0) + +#define PIPELINE_SELECT ((0x3<<29)|(0x1<<27)|(0x1<<24)|(0x4<<16)) +#define GFX_OP_3DSTATE_VF_STATISTICS ((0x3<<29)|(0x1<<27)|(0x0<<24)|(0xB<<16)) +#define MEDIA_VFE_STATE ((0x3<<29)|(0x2<<27)|(0x0<<24)|(0x0<<16)) +#define MEDIA_VFE_STATE_MMIO_ACCESS_MASK (0x18) +#define GPGPU_OBJECT ((0x3<<29)|(0x2<<27)|(0x1<<24)|(0x4<<16)) +#define GPGPU_WALKER ((0x3<<29)|(0x2<<27)|(0x1<<24)|(0x5<<16)) +#define GFX_OP_3DSTATE_DX9_CONSTANTF_VS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x39<<16)) +#define GFX_OP_3DSTATE_DX9_CONSTANTF_PS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x3A<<16)) +#define GFX_OP_3DSTATE_SO_DECL_LIST \ + ((0x3<<29)|(0x3<<27)|(0x1<<24)|(0x17<<16)) + +#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_VS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x43<<16)) +#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_GS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x44<<16)) +#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_HS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x45<<16)) +#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_DS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x46<<16)) +#define GFX_OP_3DSTATE_BINDING_TABLE_EDIT_PS \ + ((0x3<<29)|(0x3<<27)|(0x0<<24)|(0x47<<16)) + +#define MFX_WAIT ((0x3<<29)|(0x1<<27)|(0x0<<16)) + +#define COLOR_BLT ((0x2<<29)|(0x40<<22)) +#define SRC_COPY_BLT ((0x2<<29)|(0x43<<22)) + +#endif /* _INTEL_GPU_COMMANDS_H_ */ diff --git a/drivers/gpu/drm/i915/intel_guc.c b/drivers/gpu/drm/i915/intel_guc.c index ff08ea0..116f4cc 100644 --- a/drivers/gpu/drm/i915/intel_guc.c +++ b/drivers/gpu/drm/i915/intel_guc.c @@ -64,10 +64,12 @@ void intel_guc_init_early(struct intel_guc *guc) { intel_guc_fw_init_early(guc); intel_guc_ct_init_early(&guc->ct); - intel_guc_log_init_early(guc); + intel_guc_log_init_early(&guc->log); mutex_init(&guc->send_mutex); + spin_lock_init(&guc->irq_lock); guc->send = intel_guc_send_nop; + guc->handler = intel_guc_to_host_event_handler_nop; guc->notify = gen8_guc_raise_irq; } @@ -86,9 +88,10 @@ int intel_guc_init_wq(struct intel_guc *guc) * or scheduled later on resume. This way the handling of work * item can be kept same between system suspend & rpm suspend. */ - guc->log.runtime.flush_wq = alloc_ordered_workqueue("i915-guc_log", - WQ_HIGHPRI | WQ_FREEZABLE); - if (!guc->log.runtime.flush_wq) { + guc->log.relay.flush_wq = + alloc_ordered_workqueue("i915-guc_log", + WQ_HIGHPRI | WQ_FREEZABLE); + if (!guc->log.relay.flush_wq) { DRM_ERROR("Couldn't allocate workqueue for GuC log\n"); return -ENOMEM; } @@ -111,7 +114,7 @@ int intel_guc_init_wq(struct intel_guc *guc) guc->preempt_wq = alloc_ordered_workqueue("i915-guc_preempt", WQ_HIGHPRI); if (!guc->preempt_wq) { - destroy_workqueue(guc->log.runtime.flush_wq); + destroy_workqueue(guc->log.relay.flush_wq); DRM_ERROR("Couldn't allocate workqueue for GuC " "preemption\n"); return -ENOMEM; @@ -129,7 +132,7 @@ void intel_guc_fini_wq(struct intel_guc *guc) USES_GUC_SUBMISSION(dev_priv)) destroy_workqueue(guc->preempt_wq); - destroy_workqueue(guc->log.runtime.flush_wq); + destroy_workqueue(guc->log.relay.flush_wq); } static int guc_shared_data_create(struct intel_guc *guc) @@ -169,7 +172,7 @@ int intel_guc_init(struct intel_guc *guc) return ret; GEM_BUG_ON(!guc->shared_data); - ret = intel_guc_log_create(guc); + ret = intel_guc_log_create(&guc->log); if (ret) goto err_shared; @@ -184,7 +187,7 @@ int intel_guc_init(struct intel_guc *guc) return 0; err_log: - intel_guc_log_destroy(guc); + intel_guc_log_destroy(&guc->log); err_shared: guc_shared_data_destroy(guc); return ret; @@ -196,41 +199,27 @@ void intel_guc_fini(struct intel_guc *guc) i915_ggtt_disable_guc(dev_priv); intel_guc_ads_destroy(guc); - intel_guc_log_destroy(guc); + intel_guc_log_destroy(&guc->log); guc_shared_data_destroy(guc); } -static u32 get_gt_type(struct drm_i915_private *dev_priv) +static u32 get_log_control_flags(void) { - /* XXX: GT type based on PCI device ID? field seems unused by fw */ - return 0; -} - -static u32 get_core_family(struct drm_i915_private *dev_priv) -{ - u32 gen = INTEL_GEN(dev_priv); + u32 level = i915_modparams.guc_log_level; + u32 flags = 0; - switch (gen) { - case 9: - return GUC_CORE_FAMILY_GEN9; - - default: - MISSING_CASE(gen); - return GUC_CORE_FAMILY_UNKNOWN; - } -} + GEM_BUG_ON(level < 0); -static u32 get_log_verbosity_flags(void) -{ - if (i915_modparams.guc_log_level > 0) { - u32 verbosity = i915_modparams.guc_log_level - 1; + if (!GUC_LOG_LEVEL_IS_ENABLED(level)) + flags |= GUC_LOG_DEFAULT_DISABLED; - GEM_BUG_ON(verbosity > GUC_LOG_VERBOSITY_MAX); - return verbosity << GUC_LOG_VERBOSITY_SHIFT; - } + if (!GUC_LOG_LEVEL_IS_VERBOSE(level)) + flags |= GUC_LOG_DISABLED; + else + flags |= GUC_LOG_LEVEL_TO_VERBOSITY(level) << + GUC_LOG_VERBOSITY_SHIFT; - GEM_BUG_ON(i915_modparams.enable_guc < 0); - return GUC_LOG_DISABLED; + return flags; } /* @@ -246,10 +235,6 @@ void intel_guc_init_params(struct intel_guc *guc) memset(params, 0, sizeof(params)); - params[GUC_CTL_DEVICE_INFO] |= - (get_gt_type(dev_priv) << GUC_CTL_GT_TYPE_SHIFT) | - (get_core_family(dev_priv) << GUC_CTL_CORE_FAMILY_SHIFT); - /* * GuC ARAT increment is 10 ns. GuC default scheduler quantum is one * second. This ARAR is calculated by: @@ -265,12 +250,13 @@ void intel_guc_init_params(struct intel_guc *guc) params[GUC_CTL_LOG_PARAMS] = guc->log.flags; - params[GUC_CTL_DEBUG] = get_log_verbosity_flags(); + params[GUC_CTL_DEBUG] = get_log_control_flags(); /* If GuC submission is enabled, set up additional parameters here */ if (USES_GUC_SUBMISSION(dev_priv)) { - u32 ads = guc_ggtt_offset(guc->ads_vma) >> PAGE_SHIFT; - u32 pgs = guc_ggtt_offset(dev_priv->guc.stage_desc_pool); + u32 ads = intel_guc_ggtt_offset(guc, + guc->ads_vma) >> PAGE_SHIFT; + u32 pgs = intel_guc_ggtt_offset(guc, guc->stage_desc_pool); u32 ctx_in_16 = GUC_MAX_STAGE_DESCRIPTORS / 16; params[GUC_CTL_DEBUG] |= ads << GUC_ADS_ADDR_SHIFT; @@ -301,16 +287,23 @@ void intel_guc_init_params(struct intel_guc *guc) intel_uncore_forcewake_put(dev_priv, FORCEWAKE_BLITTER); } -int intel_guc_send_nop(struct intel_guc *guc, const u32 *action, u32 len) +int intel_guc_send_nop(struct intel_guc *guc, const u32 *action, u32 len, + u32 *response_buf, u32 response_buf_size) { WARN(1, "Unexpected send: action=%#x\n", *action); return -ENODEV; } +void intel_guc_to_host_event_handler_nop(struct intel_guc *guc) +{ + WARN(1, "Unexpected event: no suitable handler\n"); +} + /* * This function implements the MMIO based host to GuC interface. */ -int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len) +int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len, + u32 *response_buf, u32 response_buf_size) { struct drm_i915_private *dev_priv = guc_to_i915(guc); u32 status; @@ -320,6 +313,9 @@ int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len) GEM_BUG_ON(!len); GEM_BUG_ON(len > guc->send_regs.count); + /* We expect only action code */ + GEM_BUG_ON(*action & ~INTEL_GUC_MSG_CODE_MASK); + /* If CT is available, we expect to use MMIO only during init/fini */ GEM_BUG_ON(HAS_GUC_CT(dev_priv) && *action != INTEL_GUC_ACTION_REGISTER_COMMAND_TRANSPORT_BUFFER && @@ -341,29 +337,74 @@ int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len) */ ret = __intel_wait_for_register_fw(dev_priv, guc_send_reg(guc, 0), - INTEL_GUC_RECV_MASK, - INTEL_GUC_RECV_MASK, + INTEL_GUC_MSG_TYPE_MASK, + INTEL_GUC_MSG_TYPE_RESPONSE << + INTEL_GUC_MSG_TYPE_SHIFT, 10, 10, &status); - if (status != INTEL_GUC_STATUS_SUCCESS) { - /* - * Either the GuC explicitly returned an error (which - * we convert to -EIO here) or no response at all was - * received within the timeout limit (-ETIMEDOUT) - */ - if (ret != -ETIMEDOUT) - ret = -EIO; - - DRM_WARN("INTEL_GUC_SEND: Action 0x%X failed;" - " ret=%d status=0x%08X response=0x%08X\n", - action[0], ret, status, I915_READ(SOFT_SCRATCH(15))); + /* If GuC explicitly returned an error, convert it to -EIO */ + if (!ret && !INTEL_GUC_MSG_IS_RESPONSE_SUCCESS(status)) + ret = -EIO; + + if (ret) { + DRM_DEBUG_DRIVER("INTEL_GUC_SEND: Action 0x%X failed;" + " ret=%d status=0x%08X response=0x%08X\n", + action[0], ret, status, + I915_READ(SOFT_SCRATCH(15))); + goto out; } + if (response_buf) { + int count = min(response_buf_size, guc->send_regs.count - 1); + + for (i = 0; i < count; i++) + response_buf[i] = I915_READ(guc_send_reg(guc, i + 1)); + } + + /* Use data from the GuC response as our return value */ + ret = INTEL_GUC_MSG_TO_DATA(status); + +out: intel_uncore_forcewake_put(dev_priv, guc->send_regs.fw_domains); mutex_unlock(&guc->send_mutex); return ret; } +void intel_guc_to_host_event_handler_mmio(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + u32 msg, val; + + /* + * Sample the log buffer flush related bits & clear them out now + * itself from the message identity register to minimize the + * probability of losing a flush interrupt, when there are back + * to back flush interrupts. + * There can be a new flush interrupt, for different log buffer + * type (like for ISR), whilst Host is handling one (for DPC). + * Since same bit is used in message register for ISR & DPC, it + * could happen that GuC sets the bit for 2nd interrupt but Host + * clears out the bit on handling the 1st interrupt. + */ + spin_lock(&guc->irq_lock); + val = I915_READ(SOFT_SCRATCH(15)); + msg = val & guc->msg_enabled_mask; + I915_WRITE(SOFT_SCRATCH(15), val & ~msg); + spin_unlock(&guc->irq_lock); + + intel_guc_to_host_process_recv_msg(guc, msg); +} + +void intel_guc_to_host_process_recv_msg(struct intel_guc *guc, u32 msg) +{ + /* Make sure to handle only enabled messages */ + msg &= guc->msg_enabled_mask; + + if (msg & (INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | + INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED)) + intel_guc_log_handle_flush_event(&guc->log); +} + int intel_guc_sample_forcewake(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); @@ -410,7 +451,7 @@ int intel_guc_suspend(struct intel_guc *guc) u32 data[] = { INTEL_GUC_ACTION_ENTER_S_STATE, GUC_POWER_D1, /* any value greater than GUC_POWER_D0 */ - guc_ggtt_offset(guc->shared_data) + intel_guc_ggtt_offset(guc, guc->shared_data) }; return intel_guc_send(guc, data, ARRAY_SIZE(data)); @@ -434,7 +475,7 @@ int intel_guc_reset_engine(struct intel_guc *guc, data[3] = 0; data[4] = 0; data[5] = guc->execbuf_client->stage_id; - data[6] = guc_ggtt_offset(guc->shared_data); + data[6] = intel_guc_ggtt_offset(guc, guc->shared_data); return intel_guc_send(guc, data, ARRAY_SIZE(data)); } @@ -448,13 +489,66 @@ int intel_guc_resume(struct intel_guc *guc) u32 data[] = { INTEL_GUC_ACTION_EXIT_S_STATE, GUC_POWER_D0, - guc_ggtt_offset(guc->shared_data) + intel_guc_ggtt_offset(guc, guc->shared_data) }; return intel_guc_send(guc, data, ARRAY_SIZE(data)); } /** + * DOC: GuC Address Space + * + * The layout of GuC address space is shown below: + * + * :: + * + * +==============> +====================+ <== GUC_GGTT_TOP + * ^ | | + * | | | + * | | DRAM | + * | | Memory | + * | | | + * GuC | | + * Address +========> +====================+ <== WOPCM Top + * Space ^ | HW contexts RSVD | + * | | | WOPCM | + * | | +==> +--------------------+ <== GuC WOPCM Top + * | GuC ^ | | + * | GGTT | | | + * | Pin GuC | GuC | + * | Bias WOPCM | WOPCM | + * | | Size | | + * | | | | | + * v v v | | + * +=====+=====+==> +====================+ <== GuC WOPCM Base + * | Non-GuC WOPCM | + * | (HuC/Reserved) | + * +====================+ <== WOPCM Base + * + * The lower part of GuC Address Space [0, ggtt_pin_bias) is mapped to WOPCM + * while upper part of GuC Address Space [ggtt_pin_bias, GUC_GGTT_TOP) is mapped + * to DRAM. The value of the GuC ggtt_pin_bias is determined by WOPCM size and + * actual GuC WOPCM size. + */ + +/** + * intel_guc_init_ggtt_pin_bias() - Initialize the GuC ggtt_pin_bias value. + * @guc: intel_guc structure. + * + * This function will calculate and initialize the ggtt_pin_bias value based on + * overall WOPCM size and GuC WOPCM size. + */ +void intel_guc_init_ggtt_pin_bias(struct intel_guc *guc) +{ + struct drm_i915_private *i915 = guc_to_i915(guc); + + GEM_BUG_ON(!i915->wopcm.size); + GEM_BUG_ON(i915->wopcm.size < i915->wopcm.guc.base); + + guc->ggtt_pin_bias = i915->wopcm.size - i915->wopcm.guc.base; +} + +/** * intel_guc_allocate_vma() - Allocate a GGTT VMA for GuC usage * @guc: the guc * @size: size of area to allocate (both virtual space and memory) @@ -462,7 +556,7 @@ int intel_guc_resume(struct intel_guc *guc) * This is a wrapper to create an object for use with the GuC. In order to * use it inside the GuC, an object needs to be pinned lifetime, so we allocate * both some backing storage and a range inside the Global GTT. We must pin - * it in the GGTT somewhere other than than [0, GUC_WOPCM_TOP) because that + * it in the GGTT somewhere other than than [0, GUC ggtt_pin_bias) because that * range is reserved inside GuC. * * Return: A i915_vma if successful, otherwise an ERR_PTR. @@ -483,7 +577,7 @@ struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size) goto err; ret = i915_vma_pin(vma, 0, PAGE_SIZE, - PIN_GLOBAL | PIN_OFFSET_BIAS | GUC_WOPCM_TOP); + PIN_GLOBAL | PIN_OFFSET_BIAS | guc->ggtt_pin_bias); if (ret) { vma = ERR_PTR(ret); goto err; @@ -495,14 +589,3 @@ err: i915_gem_object_put(obj); return vma; } - -u32 intel_guc_wopcm_size(struct drm_i915_private *dev_priv) -{ - u32 wopcm_size = GUC_WOPCM_TOP; - - /* On BXT, the top of WOPCM is reserved for RC6 context */ - if (IS_GEN9_LP(dev_priv)) - wopcm_size -= BXT_GUC_WOPCM_RC6_RESERVED; - - return wopcm_size; -} diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h index b9424ac..f1265e1 100644 --- a/drivers/gpu/drm/i915/intel_guc.h +++ b/drivers/gpu/drm/i915/intel_guc.h @@ -49,11 +49,16 @@ struct intel_guc { struct intel_guc_log log; struct intel_guc_ct ct; + /* Offset where Non-WOPCM memory starts. */ + u32 ggtt_pin_bias; + /* Log snapshot if GuC errors during load */ struct drm_i915_gem_object *load_err_log; /* intel_guc_recv interrupt related state */ + spinlock_t irq_lock; bool interrupts_enabled; + unsigned int msg_enabled_mask; struct i915_vma *ads_vma; struct i915_vma *stage_desc_pool; @@ -83,7 +88,11 @@ struct intel_guc { struct mutex send_mutex; /* GuC's FW specific send function */ - int (*send)(struct intel_guc *guc, const u32 *data, u32 len); + int (*send)(struct intel_guc *guc, const u32 *data, u32 len, + u32 *response_buf, u32 response_buf_size); + + /* GuC's FW specific event handler function */ + void (*handler)(struct intel_guc *guc); /* GuC's FW specific notify function */ void (*notify)(struct intel_guc *guc); @@ -92,7 +101,14 @@ struct intel_guc { static inline int intel_guc_send(struct intel_guc *guc, const u32 *action, u32 len) { - return guc->send(guc, action, len); + return guc->send(guc, action, len, NULL, 0); +} + +static inline int +intel_guc_send_and_receive(struct intel_guc *guc, const u32 *action, u32 len, + u32 *response_buf, u32 response_buf_size) +{ + return guc->send(guc, action, len, response_buf, response_buf_size); } static inline void intel_guc_notify(struct intel_guc *guc) @@ -100,17 +116,33 @@ static inline void intel_guc_notify(struct intel_guc *guc) guc->notify(guc); } -/* - * GuC does not allow any gfx GGTT address that falls into range [0, WOPCM_TOP), - * which is reserved for Boot ROM, SRAM and WOPCM. Currently this top address is - * 512K. In order to exclude 0-512K address space from GGTT, all gfx objects - * used by GuC is pinned with PIN_OFFSET_BIAS along with size of WOPCM. +static inline void intel_guc_to_host_event_handler(struct intel_guc *guc) +{ + guc->handler(guc); +} + +/* GuC addresses above GUC_GGTT_TOP also don't map through the GTT */ +#define GUC_GGTT_TOP 0xFEE00000 + +/** + * intel_guc_ggtt_offset() - Get and validate the GGTT offset of @vma + * @guc: intel_guc structure. + * @vma: i915 graphics virtual memory area. + * + * GuC does not allow any gfx GGTT address that falls into range + * [0, GuC ggtt_pin_bias), which is reserved for Boot ROM, SRAM and WOPCM. + * Currently, in order to exclude [0, GuC ggtt_pin_bias) address space from + * GGTT, all gfx objects used by GuC are allocated with intel_guc_allocate_vma() + * and pinned with PIN_OFFSET_BIAS along with the value of GuC ggtt_pin_bias. + * + * Return: GGTT offset of the @vma. */ -static inline u32 guc_ggtt_offset(struct i915_vma *vma) +static inline u32 intel_guc_ggtt_offset(struct intel_guc *guc, + struct i915_vma *vma) { u32 offset = i915_ggtt_offset(vma); - GEM_BUG_ON(offset < GUC_WOPCM_TOP); + GEM_BUG_ON(offset < guc->ggtt_pin_bias); GEM_BUG_ON(range_overflows_t(u64, offset, vma->size, GUC_GGTT_TOP)); return offset; @@ -119,17 +151,43 @@ static inline u32 guc_ggtt_offset(struct i915_vma *vma) void intel_guc_init_early(struct intel_guc *guc); void intel_guc_init_send_regs(struct intel_guc *guc); void intel_guc_init_params(struct intel_guc *guc); +void intel_guc_init_ggtt_pin_bias(struct intel_guc *guc); int intel_guc_init_wq(struct intel_guc *guc); void intel_guc_fini_wq(struct intel_guc *guc); int intel_guc_init(struct intel_guc *guc); void intel_guc_fini(struct intel_guc *guc); -int intel_guc_send_nop(struct intel_guc *guc, const u32 *action, u32 len); -int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len); +int intel_guc_send_nop(struct intel_guc *guc, const u32 *action, u32 len, + u32 *response_buf, u32 response_buf_size); +int intel_guc_send_mmio(struct intel_guc *guc, const u32 *action, u32 len, + u32 *response_buf, u32 response_buf_size); +void intel_guc_to_host_event_handler(struct intel_guc *guc); +void intel_guc_to_host_event_handler_nop(struct intel_guc *guc); +void intel_guc_to_host_event_handler_mmio(struct intel_guc *guc); +void intel_guc_to_host_process_recv_msg(struct intel_guc *guc, u32 msg); int intel_guc_sample_forcewake(struct intel_guc *guc); int intel_guc_auth_huc(struct intel_guc *guc, u32 rsa_offset); int intel_guc_suspend(struct intel_guc *guc); int intel_guc_resume(struct intel_guc *guc); struct i915_vma *intel_guc_allocate_vma(struct intel_guc *guc, u32 size); -u32 intel_guc_wopcm_size(struct drm_i915_private *dev_priv); + +static inline int intel_guc_sanitize(struct intel_guc *guc) +{ + intel_uc_fw_sanitize(&guc->fw); + return 0; +} + +static inline void intel_guc_enable_msg(struct intel_guc *guc, u32 mask) +{ + spin_lock_irq(&guc->irq_lock); + guc->msg_enabled_mask |= mask; + spin_unlock_irq(&guc->irq_lock); +} + +static inline void intel_guc_disable_msg(struct intel_guc *guc, u32 mask) +{ + spin_lock_irq(&guc->irq_lock); + guc->msg_enabled_mask &= ~mask; + spin_unlock_irq(&guc->irq_lock); +} #endif diff --git a/drivers/gpu/drm/i915/intel_guc_ads.c b/drivers/gpu/drm/i915/intel_guc_ads.c index ac62753..dcaa3fb7 100644 --- a/drivers/gpu/drm/i915/intel_guc_ads.c +++ b/drivers/gpu/drm/i915/intel_guc_ads.c @@ -75,7 +75,7 @@ static void guc_policies_init(struct guc_policies *policies) int intel_guc_ads_create(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct i915_vma *vma; + struct i915_vma *vma, *kernel_ctx_vma; struct page *page; /* The ads obj includes the struct itself and buffers passed to GuC */ struct { @@ -121,9 +121,10 @@ int intel_guc_ads_create(struct intel_guc *guc) * to find it. Note that we have to skip our header (1 page), * because our GuC shared data is there. */ + kernel_ctx_vma = to_intel_context(dev_priv->kernel_context, + dev_priv->engine[RCS])->state; blob->ads.golden_context_lrca = - guc_ggtt_offset(dev_priv->kernel_context->engine[RCS].state) + - skipped_offset; + intel_guc_ggtt_offset(guc, kernel_ctx_vma) + skipped_offset; /* * The GuC expects us to exclude the portion of the context image that @@ -135,7 +136,7 @@ int intel_guc_ads_create(struct intel_guc *guc) blob->ads.eng_state_size[engine->guc_id] = engine->context_size - skipped_size; - base = guc_ggtt_offset(vma); + base = intel_guc_ggtt_offset(guc, vma); blob->ads.scheduler_policies = base + ptr_offset(blob, policies); blob->ads.reg_state_buffer = base + ptr_offset(blob, reg_state_buffer); blob->ads.reg_state_addr = base + ptr_offset(blob, reg_state); diff --git a/drivers/gpu/drm/i915/intel_guc_ct.c b/drivers/gpu/drm/i915/intel_guc_ct.c index 24ad557..371b600 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/intel_guc_ct.c @@ -24,14 +24,49 @@ #include "i915_drv.h" #include "intel_guc_ct.h" +#ifdef CONFIG_DRM_I915_DEBUG_GUC +#define CT_DEBUG_DRIVER(...) DRM_DEBUG_DRIVER(__VA_ARGS__) +#else +#define CT_DEBUG_DRIVER(...) do { } while (0) +#endif + +struct ct_request { + struct list_head link; + u32 fence; + u32 status; + u32 response_len; + u32 *response_buf; +}; + +struct ct_incoming_request { + struct list_head link; + u32 msg[]; +}; + enum { CTB_SEND = 0, CTB_RECV = 1 }; enum { CTB_OWNER_HOST = 0 }; +static void ct_incoming_request_worker_func(struct work_struct *w); + +/** + * intel_guc_ct_init_early - Initialize CT state without requiring device access + * @ct: pointer to CT struct + */ void intel_guc_ct_init_early(struct intel_guc_ct *ct) { /* we're using static channel owners */ ct->host_channel.owner = CTB_OWNER_HOST; + + spin_lock_init(&ct->lock); + INIT_LIST_HEAD(&ct->pending_requests); + INIT_LIST_HEAD(&ct->incoming_requests); + INIT_WORK(&ct->worker, ct_incoming_request_worker_func); +} + +static inline struct intel_guc *ct_to_guc(struct intel_guc_ct *ct) +{ + return container_of(ct, struct intel_guc, ct); } static inline const char *guc_ct_buffer_type_to_str(u32 type) @@ -49,8 +84,8 @@ static inline const char *guc_ct_buffer_type_to_str(u32 type) static void guc_ct_buffer_desc_init(struct guc_ct_buffer_desc *desc, u32 cmds_addr, u32 size, u32 owner) { - DRM_DEBUG_DRIVER("CT: desc %p init addr=%#x size=%u owner=%u\n", - desc, cmds_addr, size, owner); + CT_DEBUG_DRIVER("CT: desc %p init addr=%#x size=%u owner=%u\n", + desc, cmds_addr, size, owner); memset(desc, 0, sizeof(*desc)); desc->addr = cmds_addr; desc->size = size; @@ -59,8 +94,8 @@ static void guc_ct_buffer_desc_init(struct guc_ct_buffer_desc *desc, static void guc_ct_buffer_desc_reset(struct guc_ct_buffer_desc *desc) { - DRM_DEBUG_DRIVER("CT: desc %p reset head=%u tail=%u\n", - desc, desc->head, desc->tail); + CT_DEBUG_DRIVER("CT: desc %p reset head=%u tail=%u\n", + desc, desc->head, desc->tail); desc->head = 0; desc->tail = 0; desc->is_in_error = 0; @@ -79,7 +114,7 @@ static int guc_action_register_ct_buffer(struct intel_guc *guc, int err; /* Can't use generic send(), CT registration must go over MMIO */ - err = intel_guc_send_mmio(guc, action, ARRAY_SIZE(action)); + err = intel_guc_send_mmio(guc, action, ARRAY_SIZE(action), NULL, 0); if (err) DRM_ERROR("CT: register %s buffer failed; err=%d\n", guc_ct_buffer_type_to_str(type), err); @@ -98,7 +133,7 @@ static int guc_action_deregister_ct_buffer(struct intel_guc *guc, int err; /* Can't use generic send(), CT deregistration must go over MMIO */ - err = intel_guc_send_mmio(guc, action, ARRAY_SIZE(action)); + err = intel_guc_send_mmio(guc, action, ARRAY_SIZE(action), NULL, 0); if (err) DRM_ERROR("CT: deregister %s buffer failed; owner=%d err=%d\n", guc_ct_buffer_type_to_str(type), owner, err); @@ -156,7 +191,8 @@ static int ctch_init(struct intel_guc *guc, err = PTR_ERR(blob); goto err_vma; } - DRM_DEBUG_DRIVER("CT: vma base=%#x\n", guc_ggtt_offset(ctch->vma)); + CT_DEBUG_DRIVER("CT: vma base=%#x\n", + intel_guc_ggtt_offset(guc, ctch->vma)); /* store pointers to desc and cmds */ for (i = 0; i < ARRAY_SIZE(ctch->ctbs); i++) { @@ -170,8 +206,8 @@ static int ctch_init(struct intel_guc *guc, err_vma: i915_vma_unpin_and_release(&ctch->vma); err_out: - DRM_DEBUG_DRIVER("CT: channel %d initialization failed; err=%d\n", - ctch->owner, err); + CT_DEBUG_DRIVER("CT: channel %d initialization failed; err=%d\n", + ctch->owner, err); return err; } @@ -191,8 +227,8 @@ static int ctch_open(struct intel_guc *guc, int err; int i; - DRM_DEBUG_DRIVER("CT: channel %d reopen=%s\n", - ctch->owner, yesno(ctch_is_open(ctch))); + CT_DEBUG_DRIVER("CT: channel %d reopen=%s\n", + ctch->owner, yesno(ctch_is_open(ctch))); if (!ctch->vma) { err = ctch_init(guc, ctch); @@ -202,7 +238,7 @@ static int ctch_open(struct intel_guc *guc, } /* vma should be already allocated and map'ed */ - base = guc_ggtt_offset(ctch->vma); + base = intel_guc_ggtt_offset(guc, ctch->vma); /* (re)initialize descriptors * cmds buffers are in the second half of the blob page @@ -263,10 +299,29 @@ static u32 ctch_get_next_fence(struct intel_guc_ct_channel *ctch) return ++ctch->next_fence; } +/** + * DOC: CTB Host to GuC request + * + * Format of the CTB Host to GuC request message is as follows:: + * + * +------------+---------+---------+---------+---------+ + * | msg[0] | [1] | [2] | ... | [n-1] | + * +------------+---------+---------+---------+---------+ + * | MESSAGE | MESSAGE PAYLOAD | + * + HEADER +---------+---------+---------+---------+ + * | | 0 | 1 | ... | n | + * +============+=========+=========+=========+=========+ + * | len >= 1 | FENCE | request specific data | + * +------+-----+---------+---------+---------+---------+ + * + * ^-----------------len-------------------^ + */ + static int ctb_write(struct intel_guc_ct_buffer *ctb, const u32 *action, u32 len /* in dwords */, - u32 fence) + u32 fence, + bool want_response) { struct guc_ct_buffer_desc *desc = ctb->desc; u32 head = desc->head / 4; /* in dwords */ @@ -295,15 +350,21 @@ static int ctb_write(struct intel_guc_ct_buffer *ctb, if (unlikely(used + len + 1 >= size)) return -ENOSPC; - /* Write the message. The format is the following: + /* + * Write the message. The format is the following: * DW0: header (including action code) * DW1: fence * DW2+: action data */ header = (len << GUC_CT_MSG_LEN_SHIFT) | (GUC_CT_MSG_WRITE_FENCE_TO_DESC) | + (want_response ? GUC_CT_MSG_SEND_STATUS : 0) | (action[0] << GUC_CT_MSG_ACTION_SHIFT); + CT_DEBUG_DRIVER("CT: writing %*ph %*ph %*ph\n", + 4, &header, 4, &fence, + 4 * (len - 1), &action[1]); + cmds[tail] = header; tail = (tail + 1) % size; @@ -322,16 +383,25 @@ static int ctb_write(struct intel_guc_ct_buffer *ctb, return 0; } -/* Wait for the response from the GuC. +/** + * wait_for_ctb_desc_update - Wait for the CT buffer descriptor update. + * @desc: buffer descriptor * @fence: response fence * @status: placeholder for status - * return: 0 response received (status is valid) - * -ETIMEDOUT no response within hardcoded timeout - * -EPROTO no response, ct buffer was in error + * + * Guc will update CT buffer descriptor with new fence and status + * after processing the command identified by the fence. Wait for + * specified fence and then read from the descriptor status of the + * command. + * + * Return: + * * 0 response received (status is valid) + * * -ETIMEDOUT no response within hardcoded timeout + * * -EPROTO no response, CT buffer is in error */ -static int wait_for_response(struct guc_ct_buffer_desc *desc, - u32 fence, - u32 *status) +static int wait_for_ctb_desc_update(struct guc_ct_buffer_desc *desc, + u32 fence, + u32 *status) { int err; @@ -363,71 +433,440 @@ static int wait_for_response(struct guc_ct_buffer_desc *desc, return err; } -static int ctch_send(struct intel_guc *guc, +/** + * wait_for_ct_request_update - Wait for CT request state update. + * @req: pointer to pending request + * @status: placeholder for status + * + * For each sent request, Guc shall send bac CT response message. + * Our message handler will update status of tracked request once + * response message with given fence is received. Wait here and + * check for valid response status value. + * + * Return: + * * 0 response received (status is valid) + * * -ETIMEDOUT no response within hardcoded timeout + */ +static int wait_for_ct_request_update(struct ct_request *req, u32 *status) +{ + int err; + + /* + * Fast commands should complete in less than 10us, so sample quickly + * up to that length of time, then switch to a slower sleep-wait loop. + * No GuC command should ever take longer than 10ms. + */ +#define done INTEL_GUC_MSG_IS_RESPONSE(READ_ONCE(req->status)) + err = wait_for_us(done, 10); + if (err) + err = wait_for(done, 10); +#undef done + + if (unlikely(err)) + DRM_ERROR("CT: fence %u err %d\n", req->fence, err); + + *status = req->status; + return err; +} + +static int ctch_send(struct intel_guc_ct *ct, struct intel_guc_ct_channel *ctch, const u32 *action, u32 len, + u32 *response_buf, + u32 response_buf_size, u32 *status) { struct intel_guc_ct_buffer *ctb = &ctch->ctbs[CTB_SEND]; struct guc_ct_buffer_desc *desc = ctb->desc; + struct ct_request request; + unsigned long flags; u32 fence; int err; GEM_BUG_ON(!ctch_is_open(ctch)); GEM_BUG_ON(!len); GEM_BUG_ON(len & ~GUC_CT_MSG_LEN_MASK); + GEM_BUG_ON(!response_buf && response_buf_size); fence = ctch_get_next_fence(ctch); - err = ctb_write(ctb, action, len, fence); + request.fence = fence; + request.status = 0; + request.response_len = response_buf_size; + request.response_buf = response_buf; + + spin_lock_irqsave(&ct->lock, flags); + list_add_tail(&request.link, &ct->pending_requests); + spin_unlock_irqrestore(&ct->lock, flags); + + err = ctb_write(ctb, action, len, fence, !!response_buf); if (unlikely(err)) - return err; + goto unlink; - intel_guc_notify(guc); + intel_guc_notify(ct_to_guc(ct)); - err = wait_for_response(desc, fence, status); + if (response_buf) + err = wait_for_ct_request_update(&request, status); + else + err = wait_for_ctb_desc_update(desc, fence, status); if (unlikely(err)) - return err; - if (*status != INTEL_GUC_STATUS_SUCCESS) - return -EIO; - return 0; + goto unlink; + + if (!INTEL_GUC_MSG_IS_RESPONSE_SUCCESS(*status)) { + err = -EIO; + goto unlink; + } + + if (response_buf) { + /* There shall be no data in the status */ + WARN_ON(INTEL_GUC_MSG_TO_DATA(request.status)); + /* Return actual response len */ + err = request.response_len; + } else { + /* There shall be no response payload */ + WARN_ON(request.response_len); + /* Return data decoded from the status dword */ + err = INTEL_GUC_MSG_TO_DATA(*status); + } + +unlink: + spin_lock_irqsave(&ct->lock, flags); + list_del(&request.link); + spin_unlock_irqrestore(&ct->lock, flags); + + return err; } /* * Command Transport (CT) buffer based GuC send function. */ -static int intel_guc_send_ct(struct intel_guc *guc, const u32 *action, u32 len) +static int intel_guc_send_ct(struct intel_guc *guc, const u32 *action, u32 len, + u32 *response_buf, u32 response_buf_size) { - struct intel_guc_ct_channel *ctch = &guc->ct.host_channel; + struct intel_guc_ct *ct = &guc->ct; + struct intel_guc_ct_channel *ctch = &ct->host_channel; u32 status = ~0; /* undefined */ - int err; + int ret; mutex_lock(&guc->send_mutex); - err = ctch_send(guc, ctch, action, len, &status); - if (unlikely(err)) { + ret = ctch_send(ct, ctch, action, len, response_buf, response_buf_size, + &status); + if (unlikely(ret < 0)) { DRM_ERROR("CT: send action %#X failed; err=%d status=%#X\n", - action[0], err, status); + action[0], ret, status); + } else if (unlikely(ret)) { + CT_DEBUG_DRIVER("CT: send action %#x returned %d (%#x)\n", + action[0], ret, ret); } mutex_unlock(&guc->send_mutex); - return err; + return ret; +} + +static inline unsigned int ct_header_get_len(u32 header) +{ + return (header >> GUC_CT_MSG_LEN_SHIFT) & GUC_CT_MSG_LEN_MASK; +} + +static inline unsigned int ct_header_get_action(u32 header) +{ + return (header >> GUC_CT_MSG_ACTION_SHIFT) & GUC_CT_MSG_ACTION_MASK; +} + +static inline bool ct_header_is_response(u32 header) +{ + return ct_header_get_action(header) == INTEL_GUC_ACTION_DEFAULT; +} + +static int ctb_read(struct intel_guc_ct_buffer *ctb, u32 *data) +{ + struct guc_ct_buffer_desc *desc = ctb->desc; + u32 head = desc->head / 4; /* in dwords */ + u32 tail = desc->tail / 4; /* in dwords */ + u32 size = desc->size / 4; /* in dwords */ + u32 *cmds = ctb->cmds; + s32 available; /* in dwords */ + unsigned int len; + unsigned int i; + + GEM_BUG_ON(desc->size % 4); + GEM_BUG_ON(desc->head % 4); + GEM_BUG_ON(desc->tail % 4); + GEM_BUG_ON(tail >= size); + GEM_BUG_ON(head >= size); + + /* tail == head condition indicates empty */ + available = tail - head; + if (unlikely(available == 0)) + return -ENODATA; + + /* beware of buffer wrap case */ + if (unlikely(available < 0)) + available += size; + CT_DEBUG_DRIVER("CT: available %d (%u:%u)\n", available, head, tail); + GEM_BUG_ON(available < 0); + + data[0] = cmds[head]; + head = (head + 1) % size; + + /* message len with header */ + len = ct_header_get_len(data[0]) + 1; + if (unlikely(len > (u32)available)) { + DRM_ERROR("CT: incomplete message %*ph %*ph %*ph\n", + 4, data, + 4 * (head + available - 1 > size ? + size - head : available - 1), &cmds[head], + 4 * (head + available - 1 > size ? + available - 1 - size + head : 0), &cmds[0]); + return -EPROTO; + } + + for (i = 1; i < len; i++) { + data[i] = cmds[head]; + head = (head + 1) % size; + } + CT_DEBUG_DRIVER("CT: received %*ph\n", 4 * len, data); + + desc->head = head * 4; + return 0; } /** - * Enable buffer based command transport + * DOC: CTB GuC to Host response + * + * Format of the CTB GuC to Host response message is as follows:: + * + * +------------+---------+---------+---------+---------+---------+ + * | msg[0] | [1] | [2] | [3] | ... | [n-1] | + * +------------+---------+---------+---------+---------+---------+ + * | MESSAGE | MESSAGE PAYLOAD | + * + HEADER +---------+---------+---------+---------+---------+ + * | | 0 | 1 | 2 | ... | n | + * +============+=========+=========+=========+=========+=========+ + * | len >= 2 | FENCE | STATUS | response specific data | + * +------+-----+---------+---------+---------+---------+---------+ + * + * ^-----------------------len-----------------------^ + */ + +static int ct_handle_response(struct intel_guc_ct *ct, const u32 *msg) +{ + u32 header = msg[0]; + u32 len = ct_header_get_len(header); + u32 msglen = len + 1; /* total message length including header */ + u32 fence; + u32 status; + u32 datalen; + struct ct_request *req; + bool found = false; + + GEM_BUG_ON(!ct_header_is_response(header)); + GEM_BUG_ON(!in_irq()); + + /* Response payload shall at least include fence and status */ + if (unlikely(len < 2)) { + DRM_ERROR("CT: corrupted response %*ph\n", 4 * msglen, msg); + return -EPROTO; + } + + fence = msg[1]; + status = msg[2]; + datalen = len - 2; + + /* Format of the status follows RESPONSE message */ + if (unlikely(!INTEL_GUC_MSG_IS_RESPONSE(status))) { + DRM_ERROR("CT: corrupted response %*ph\n", 4 * msglen, msg); + return -EPROTO; + } + + CT_DEBUG_DRIVER("CT: response fence %u status %#x\n", fence, status); + + spin_lock(&ct->lock); + list_for_each_entry(req, &ct->pending_requests, link) { + if (unlikely(fence != req->fence)) { + CT_DEBUG_DRIVER("CT: request %u awaits response\n", + req->fence); + continue; + } + if (unlikely(datalen > req->response_len)) { + DRM_ERROR("CT: response %u too long %*ph\n", + req->fence, 4 * msglen, msg); + datalen = 0; + } + if (datalen) + memcpy(req->response_buf, msg + 3, 4 * datalen); + req->response_len = datalen; + WRITE_ONCE(req->status, status); + found = true; + break; + } + spin_unlock(&ct->lock); + + if (!found) + DRM_ERROR("CT: unsolicited response %*ph\n", 4 * msglen, msg); + return 0; +} + +static void ct_process_request(struct intel_guc_ct *ct, + u32 action, u32 len, const u32 *payload) +{ + struct intel_guc *guc = ct_to_guc(ct); + + CT_DEBUG_DRIVER("CT: request %x %*ph\n", action, 4 * len, payload); + + switch (action) { + case INTEL_GUC_ACTION_DEFAULT: + if (unlikely(len < 1)) + goto fail_unexpected; + intel_guc_to_host_process_recv_msg(guc, *payload); + break; + + default: +fail_unexpected: + DRM_ERROR("CT: unexpected request %x %*ph\n", + action, 4 * len, payload); + break; + } +} + +static bool ct_process_incoming_requests(struct intel_guc_ct *ct) +{ + unsigned long flags; + struct ct_incoming_request *request; + u32 header; + u32 *payload; + bool done; + + spin_lock_irqsave(&ct->lock, flags); + request = list_first_entry_or_null(&ct->incoming_requests, + struct ct_incoming_request, link); + if (request) + list_del(&request->link); + done = !!list_empty(&ct->incoming_requests); + spin_unlock_irqrestore(&ct->lock, flags); + + if (!request) + return true; + + header = request->msg[0]; + payload = &request->msg[1]; + ct_process_request(ct, + ct_header_get_action(header), + ct_header_get_len(header), + payload); + + kfree(request); + return done; +} + +static void ct_incoming_request_worker_func(struct work_struct *w) +{ + struct intel_guc_ct *ct = container_of(w, struct intel_guc_ct, worker); + bool done; + + done = ct_process_incoming_requests(ct); + if (!done) + queue_work(system_unbound_wq, &ct->worker); +} + +/** + * DOC: CTB GuC to Host request + * + * Format of the CTB GuC to Host request message is as follows:: + * + * +------------+---------+---------+---------+---------+---------+ + * | msg[0] | [1] | [2] | [3] | ... | [n-1] | + * +------------+---------+---------+---------+---------+---------+ + * | MESSAGE | MESSAGE PAYLOAD | + * + HEADER +---------+---------+---------+---------+---------+ + * | | 0 | 1 | 2 | ... | n | + * +============+=========+=========+=========+=========+=========+ + * | len | request specific data | + * +------+-----+---------+---------+---------+---------+---------+ + * + * ^-----------------------len-----------------------^ + */ + +static int ct_handle_request(struct intel_guc_ct *ct, const u32 *msg) +{ + u32 header = msg[0]; + u32 len = ct_header_get_len(header); + u32 msglen = len + 1; /* total message length including header */ + struct ct_incoming_request *request; + unsigned long flags; + + GEM_BUG_ON(ct_header_is_response(header)); + + request = kmalloc(sizeof(*request) + 4 * msglen, GFP_ATOMIC); + if (unlikely(!request)) { + DRM_ERROR("CT: dropping request %*ph\n", 4 * msglen, msg); + return 0; /* XXX: -ENOMEM ? */ + } + memcpy(request->msg, msg, 4 * msglen); + + spin_lock_irqsave(&ct->lock, flags); + list_add_tail(&request->link, &ct->incoming_requests); + spin_unlock_irqrestore(&ct->lock, flags); + + queue_work(system_unbound_wq, &ct->worker); + return 0; +} + +static void ct_process_host_channel(struct intel_guc_ct *ct) +{ + struct intel_guc_ct_channel *ctch = &ct->host_channel; + struct intel_guc_ct_buffer *ctb = &ctch->ctbs[CTB_RECV]; + u32 msg[GUC_CT_MSG_LEN_MASK + 1]; /* one extra dw for the header */ + int err = 0; + + if (!ctch_is_open(ctch)) + return; + + do { + err = ctb_read(ctb, msg); + if (err) + break; + + if (ct_header_is_response(msg[0])) + err = ct_handle_response(ct, msg); + else + err = ct_handle_request(ct, msg); + } while (!err); + + if (GEM_WARN_ON(err == -EPROTO)) { + DRM_ERROR("CT: corrupted message detected!\n"); + ctb->desc->is_in_error = 1; + } +} + +/* + * When we're communicating with the GuC over CT, GuC uses events + * to notify us about new messages being posted on the RECV buffer. + */ +static void intel_guc_to_host_event_handler_ct(struct intel_guc *guc) +{ + struct intel_guc_ct *ct = &guc->ct; + + ct_process_host_channel(ct); +} + +/** + * intel_guc_ct_enable - Enable buffer based command transport. + * @ct: pointer to CT struct + * * Shall only be called for platforms with HAS_GUC_CT. - * @guc: the guc - * return: 0 on success - * non-zero on failure + * + * Return: 0 on success, a negative errno code on failure. */ -int intel_guc_enable_ct(struct intel_guc *guc) +int intel_guc_ct_enable(struct intel_guc_ct *ct) { - struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct intel_guc_ct_channel *ctch = &guc->ct.host_channel; + struct intel_guc *guc = ct_to_guc(ct); + struct drm_i915_private *i915 = guc_to_i915(guc); + struct intel_guc_ct_channel *ctch = &ct->host_channel; int err; - GEM_BUG_ON(!HAS_GUC_CT(dev_priv)); + GEM_BUG_ON(!HAS_GUC_CT(i915)); err = ctch_open(guc, ctch); if (unlikely(err)) @@ -435,21 +874,24 @@ int intel_guc_enable_ct(struct intel_guc *guc) /* Switch into cmd transport buffer based send() */ guc->send = intel_guc_send_ct; + guc->handler = intel_guc_to_host_event_handler_ct; DRM_INFO("CT: %s\n", enableddisabled(true)); return 0; } /** - * Disable buffer based command transport. + * intel_guc_ct_disable - Disable buffer based command transport. + * @ct: pointer to CT struct + * * Shall only be called for platforms with HAS_GUC_CT. - * @guc: the guc */ -void intel_guc_disable_ct(struct intel_guc *guc) +void intel_guc_ct_disable(struct intel_guc_ct *ct) { - struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct intel_guc_ct_channel *ctch = &guc->ct.host_channel; + struct intel_guc *guc = ct_to_guc(ct); + struct drm_i915_private *i915 = guc_to_i915(guc); + struct intel_guc_ct_channel *ctch = &ct->host_channel; - GEM_BUG_ON(!HAS_GUC_CT(dev_priv)); + GEM_BUG_ON(!HAS_GUC_CT(i915)); if (!ctch_is_open(ctch)) return; @@ -458,5 +900,6 @@ void intel_guc_disable_ct(struct intel_guc *guc) /* Disable send */ guc->send = intel_guc_send_nop; + guc->handler = intel_guc_to_host_event_handler_nop; DRM_INFO("CT: %s\n", enableddisabled(false)); } diff --git a/drivers/gpu/drm/i915/intel_guc_ct.h b/drivers/gpu/drm/i915/intel_guc_ct.h index 6d97f36..d774895a 100644 --- a/drivers/gpu/drm/i915/intel_guc_ct.h +++ b/drivers/gpu/drm/i915/intel_guc_ct.h @@ -75,12 +75,22 @@ struct intel_guc_ct_channel { struct intel_guc_ct { struct intel_guc_ct_channel host_channel; /* other channels are tbd */ + + /** @lock: protects pending requests list */ + spinlock_t lock; + + /** @pending_requests: list of requests waiting for response */ + struct list_head pending_requests; + + /** @incoming_requests: list of incoming requests */ + struct list_head incoming_requests; + + /** @worker: worker for handling incoming requests */ + struct work_struct worker; }; void intel_guc_ct_init_early(struct intel_guc_ct *ct); - -/* XXX: move to intel_uc.h ? don't fit there either */ -int intel_guc_enable_ct(struct intel_guc *guc); -void intel_guc_disable_ct(struct intel_guc *guc); +int intel_guc_ct_enable(struct intel_guc_ct *ct); +void intel_guc_ct_disable(struct intel_guc_ct *ct); #endif /* _INTEL_GUC_CT_H_ */ diff --git a/drivers/gpu/drm/i915/intel_guc_fw.c b/drivers/gpu/drm/i915/intel_guc_fw.c index d07f2b9..a9e6fcc 100644 --- a/drivers/gpu/drm/i915/intel_guc_fw.c +++ b/drivers/gpu/drm/i915/intel_guc_fw.c @@ -165,7 +165,7 @@ static int guc_xfer_ucode(struct intel_guc *guc, struct i915_vma *vma) I915_WRITE(DMA_COPY_SIZE, guc_fw->header_size + guc_fw->ucode_size); /* Set the source address for the new blob */ - offset = guc_ggtt_offset(vma) + guc_fw->header_offset; + offset = intel_guc_ggtt_offset(guc, vma) + guc_fw->header_offset; I915_WRITE(DMA_ADDR_0_LOW, lower_32_bits(offset)); I915_WRITE(DMA_ADDR_0_HIGH, upper_32_bits(offset) & 0xFFFF); @@ -275,9 +275,8 @@ static int guc_fw_xfer(struct intel_uc_fw *guc_fw, struct i915_vma *vma) * Called from intel_uc_init_hw() during driver load, resume from sleep and * after a GPU reset. * - * The firmware image should have already been fetched into memory by the - * earlier call to intel_uc_init_fw(), so here we need to only check that - * fetch succeeded, and then transfer the image to the h/w. + * The firmware image should have already been fetched into memory, so only + * check that fetch succeeded, and then transfer the image to the h/w. * * Return: non-zero code on error */ diff --git a/drivers/gpu/drm/i915/intel_guc_fwif.h b/drivers/gpu/drm/i915/intel_guc_fwif.h index 6a10aa6..0867ba7 100644 --- a/drivers/gpu/drm/i915/intel_guc_fwif.h +++ b/drivers/gpu/drm/i915/intel_guc_fwif.h @@ -23,9 +23,6 @@ #ifndef _INTEL_GUC_FWIF_H #define _INTEL_GUC_FWIF_H -#define GUC_CORE_FAMILY_GEN9 12 -#define GUC_CORE_FAMILY_UNKNOWN 0x7fffffff - #define GUC_CLIENT_PRIORITY_KMD_HIGH 0 #define GUC_CLIENT_PRIORITY_HIGH 1 #define GUC_CLIENT_PRIORITY_KMD_NORMAL 2 @@ -82,8 +79,6 @@ #define GUC_CTL_ARAT_LOW 2 #define GUC_CTL_DEVICE_INFO 3 -#define GUC_CTL_GT_TYPE_SHIFT 0 -#define GUC_CTL_CORE_FAMILY_SHIFT 7 #define GUC_CTL_LOG_PARAMS 4 #define GUC_LOG_VALID (1 << 0) @@ -127,7 +122,7 @@ #define GUC_PROFILE_ENABLED (1 << 7) #define GUC_WQ_TRACK_ENABLED (1 << 8) #define GUC_ADS_ENABLED (1 << 9) -#define GUC_DEBUG_RESERVED (1 << 10) +#define GUC_LOG_DEFAULT_DISABLED (1 << 10) #define GUC_ADS_ADDR_SHIFT 11 #define GUC_ADS_ADDR_MASK 0xfffff800 @@ -327,6 +322,58 @@ struct guc_stage_desc { u64 desc_private; } __packed; +/** + * DOC: CTB based communication + * + * The CTB (command transport buffer) communication between Host and GuC + * is based on u32 data stream written to the shared buffer. One buffer can + * be used to transmit data only in one direction (one-directional channel). + * + * Current status of the each buffer is stored in the buffer descriptor. + * Buffer descriptor holds tail and head fields that represents active data + * stream. The tail field is updated by the data producer (sender), and head + * field is updated by the data consumer (receiver):: + * + * +------------+ + * | DESCRIPTOR | +=================+============+========+ + * +============+ | | MESSAGE(s) | | + * | address |--------->+=================+============+========+ + * +------------+ + * | head | ^-----head--------^ + * +------------+ + * | tail | ^---------tail-----------------^ + * +------------+ + * | size | ^---------------size--------------------^ + * +------------+ + * + * Each message in data stream starts with the single u32 treated as a header, + * followed by optional set of u32 data that makes message specific payload:: + * + * +------------+---------+---------+---------+ + * | MESSAGE | + * +------------+---------+---------+---------+ + * | msg[0] | [1] | ... | [n-1] | + * +------------+---------+---------+---------+ + * | MESSAGE | MESSAGE PAYLOAD | + * + HEADER +---------+---------+---------+ + * | | 0 | ... | n | + * +======+=====+=========+=========+=========+ + * | 31:16| code| | | | + * +------+-----+ | | | + * | 15:5|flags| | | | + * +------+-----+ | | | + * | 4:0| len| | | | + * +------+-----+---------+---------+---------+ + * + * ^-------------len-------------^ + * + * The message header consists of: + * + * - **len**, indicates length of the message payload (in u32) + * - **code**, indicates message code + * - **flags**, holds various bits to control message handling + */ + /* * Describes single command transport buffer. * Used by both guc-master and clients. @@ -534,16 +581,6 @@ struct guc_log_buffer_state { u32 version; } __packed; -union guc_log_control { - struct { - u32 logging_enabled:1; - u32 reserved1:3; - u32 verbosity:4; - u32 reserved2:24; - }; - u32 value; -} __packed; - struct guc_ctx_report { u32 report_return_status; u32 reserved1[64]; @@ -570,7 +607,68 @@ struct guc_shared_ctx_data { struct guc_ctx_report preempt_ctx_report[GUC_MAX_ENGINES_NUM]; } __packed; -/* This Action will be programmed in C180 - SOFT_SCRATCH_O_REG */ +/** + * DOC: MMIO based communication + * + * The MMIO based communication between Host and GuC uses software scratch + * registers, where first register holds data treated as message header, + * and other registers are used to hold message payload. + * + * For Gen9+, GuC uses software scratch registers 0xC180-0xC1B8 + * + * +-----------+---------+---------+---------+ + * | MMIO[0] | MMIO[1] | ... | MMIO[n] | + * +-----------+---------+---------+---------+ + * | header | optional payload | + * +======+====+=========+=========+=========+ + * | 31:28|type| | | | + * +------+----+ | | | + * | 27:16|data| | | | + * +------+----+ | | | + * | 15:0|code| | | | + * +------+----+---------+---------+---------+ + * + * The message header consists of: + * + * - **type**, indicates message type + * - **code**, indicates message code, is specific for **type** + * - **data**, indicates message data, optional, depends on **code** + * + * The following message **types** are supported: + * + * - **REQUEST**, indicates Host-to-GuC request, requested GuC action code + * must be priovided in **code** field. Optional action specific parameters + * can be provided in remaining payload registers or **data** field. + * + * - **RESPONSE**, indicates GuC-to-Host response from earlier GuC request, + * action response status will be provided in **code** field. Optional + * response data can be returned in remaining payload registers or **data** + * field. + */ + +#define INTEL_GUC_MSG_TYPE_SHIFT 28 +#define INTEL_GUC_MSG_TYPE_MASK (0xF << INTEL_GUC_MSG_TYPE_SHIFT) +#define INTEL_GUC_MSG_DATA_SHIFT 16 +#define INTEL_GUC_MSG_DATA_MASK (0xFFF << INTEL_GUC_MSG_DATA_SHIFT) +#define INTEL_GUC_MSG_CODE_SHIFT 0 +#define INTEL_GUC_MSG_CODE_MASK (0xFFFF << INTEL_GUC_MSG_CODE_SHIFT) + +#define __INTEL_GUC_MSG_GET(T, m) \ + (((m) & INTEL_GUC_MSG_ ## T ## _MASK) >> INTEL_GUC_MSG_ ## T ## _SHIFT) +#define INTEL_GUC_MSG_TO_TYPE(m) __INTEL_GUC_MSG_GET(TYPE, m) +#define INTEL_GUC_MSG_TO_DATA(m) __INTEL_GUC_MSG_GET(DATA, m) +#define INTEL_GUC_MSG_TO_CODE(m) __INTEL_GUC_MSG_GET(CODE, m) + +enum intel_guc_msg_type { + INTEL_GUC_MSG_TYPE_REQUEST = 0x0, + INTEL_GUC_MSG_TYPE_RESPONSE = 0xF, +}; + +#define __INTEL_GUC_MSG_TYPE_IS(T, m) \ + (INTEL_GUC_MSG_TO_TYPE(m) == INTEL_GUC_MSG_TYPE_ ## T) +#define INTEL_GUC_MSG_IS_REQUEST(m) __INTEL_GUC_MSG_TYPE_IS(REQUEST, m) +#define INTEL_GUC_MSG_IS_RESPONSE(m) __INTEL_GUC_MSG_TYPE_IS(RESPONSE, m) + enum intel_guc_action { INTEL_GUC_ACTION_DEFAULT = 0x0, INTEL_GUC_ACTION_REQUEST_PREEMPTION = 0x2, @@ -602,24 +700,22 @@ enum intel_guc_report_status { INTEL_GUC_REPORT_STATUS_COMPLETE = 0x4, }; -/* - * The GuC sends its response to a command by overwriting the - * command in SS0. The response is distinguishable from a command - * by the fact that all the MASK bits are set. The remaining bits - * give more detail. - */ -#define INTEL_GUC_RECV_MASK ((u32)0xF0000000) -#define INTEL_GUC_RECV_IS_RESPONSE(x) ((u32)(x) >= INTEL_GUC_RECV_MASK) -#define INTEL_GUC_RECV_STATUS(x) (INTEL_GUC_RECV_MASK | (x)) - -/* GUC will return status back to SOFT_SCRATCH_O_REG */ -enum intel_guc_status { - INTEL_GUC_STATUS_SUCCESS = INTEL_GUC_RECV_STATUS(0x0), - INTEL_GUC_STATUS_ALLOCATE_DOORBELL_FAIL = INTEL_GUC_RECV_STATUS(0x10), - INTEL_GUC_STATUS_DEALLOCATE_DOORBELL_FAIL = INTEL_GUC_RECV_STATUS(0x20), - INTEL_GUC_STATUS_GENERIC_FAIL = INTEL_GUC_RECV_STATUS(0x0000F000) +#define GUC_LOG_CONTROL_LOGGING_ENABLED (1 << 0) +#define GUC_LOG_CONTROL_VERBOSITY_SHIFT 4 +#define GUC_LOG_CONTROL_VERBOSITY_MASK (0xF << GUC_LOG_CONTROL_VERBOSITY_SHIFT) +#define GUC_LOG_CONTROL_DEFAULT_LOGGING (1 << 8) + +enum intel_guc_response_status { + INTEL_GUC_RESPONSE_STATUS_SUCCESS = 0x0, + INTEL_GUC_RESPONSE_STATUS_GENERIC_FAIL = 0xF000, }; +#define INTEL_GUC_MSG_IS_RESPONSE_SUCCESS(m) \ + (typecheck(u32, (m)) && \ + ((m) & (INTEL_GUC_MSG_TYPE_MASK | INTEL_GUC_MSG_CODE_MASK)) == \ + ((INTEL_GUC_MSG_TYPE_RESPONSE << INTEL_GUC_MSG_TYPE_SHIFT) | \ + (INTEL_GUC_RESPONSE_STATUS_SUCCESS << INTEL_GUC_MSG_CODE_SHIFT))) + /* This action will be programmed in C1BC - SOFT_SCRATCH_15_REG */ enum intel_guc_recv_message { INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED = BIT(1), diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index c0c2e7d..401e170 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -23,12 +23,11 @@ */ #include <linux/debugfs.h> -#include <linux/relay.h> #include "intel_guc_log.h" #include "i915_drv.h" -static void guc_log_capture_logs(struct intel_guc *guc); +static void guc_log_capture_logs(struct intel_guc_log *log); /** * DOC: GuC firmware log @@ -39,7 +38,7 @@ static void guc_log_capture_logs(struct intel_guc *guc); * registers value. */ -static int guc_log_flush_complete(struct intel_guc *guc) +static int guc_action_flush_log_complete(struct intel_guc *guc) { u32 action[] = { INTEL_GUC_ACTION_LOG_BUFFER_FILE_FLUSH_COMPLETE @@ -48,7 +47,7 @@ static int guc_log_flush_complete(struct intel_guc *guc) return intel_guc_send(guc, action, ARRAY_SIZE(action)); } -static int guc_log_flush(struct intel_guc *guc) +static int guc_action_flush_log(struct intel_guc *guc) { u32 action[] = { INTEL_GUC_ACTION_FORCE_LOG_BUFFER_FLUSH, @@ -58,22 +57,40 @@ static int guc_log_flush(struct intel_guc *guc) return intel_guc_send(guc, action, ARRAY_SIZE(action)); } -static int guc_log_control(struct intel_guc *guc, bool enable, u32 verbosity) +static int guc_action_control_log(struct intel_guc *guc, bool enable, + bool default_logging, u32 verbosity) { - union guc_log_control control_val = { - { - .logging_enabled = enable, - .verbosity = verbosity, - }, - }; u32 action[] = { INTEL_GUC_ACTION_UK_LOG_ENABLE_LOGGING, - control_val.value + (enable ? GUC_LOG_CONTROL_LOGGING_ENABLED : 0) | + (verbosity << GUC_LOG_CONTROL_VERBOSITY_SHIFT) | + (default_logging ? GUC_LOG_CONTROL_DEFAULT_LOGGING : 0) }; + GEM_BUG_ON(verbosity > GUC_LOG_VERBOSITY_MAX); + return intel_guc_send(guc, action, ARRAY_SIZE(action)); } +static inline struct intel_guc *log_to_guc(struct intel_guc_log *log) +{ + return container_of(log, struct intel_guc, log); +} + +static void guc_log_enable_flush_events(struct intel_guc_log *log) +{ + intel_guc_enable_msg(log_to_guc(log), + INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | + INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED); +} + +static void guc_log_disable_flush_events(struct intel_guc_log *log) +{ + intel_guc_disable_msg(log_to_guc(log), + INTEL_GUC_RECV_MSG_FLUSH_LOG_BUFFER | + INTEL_GUC_RECV_MSG_CRASH_DUMP_POSTED); +} + /* * 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. @@ -121,14 +138,7 @@ static struct dentry *create_buf_file_callback(const char *filename, if (!parent) return NULL; - /* - * Not using the channel filename passed as an argument, since for each - * channel relay appends the corresponding CPU number to the filename - * passed in relay_open(). This should be fine as relay just needs a - * dentry of the file associated with the channel buffer and that file's - * name need not be same as the filename passed as an argument. - */ - buf_file = debugfs_create_file("guc_log", mode, + buf_file = debugfs_create_file(filename, mode, parent, buf, &relay_file_operations); return buf_file; } @@ -149,59 +159,7 @@ static struct rchan_callbacks relay_callbacks = { .remove_buf_file = remove_buf_file_callback, }; -static int guc_log_relay_file_create(struct intel_guc *guc) -{ - struct drm_i915_private *dev_priv = guc_to_i915(guc); - struct dentry *log_dir; - int ret; - - if (!i915_modparams.guc_log_level) - return 0; - - mutex_lock(&guc->log.runtime.relay_lock); - - /* For now create the log file in /sys/kernel/debug/dri/0 dir */ - log_dir = dev_priv->drm.primary->debugfs_root; - - /* - * 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"); - ret = -ENODEV; - goto out_unlock; - } - - ret = relay_late_setup_files(guc->log.runtime.relay_chan, "guc_log", log_dir); - if (ret < 0 && ret != -EEXIST) { - DRM_ERROR("Couldn't associate relay chan with file %d\n", ret); - goto out_unlock; - } - - ret = 0; - -out_unlock: - mutex_unlock(&guc->log.runtime.relay_lock); - return ret; -} - -static bool guc_log_has_relay(struct intel_guc *guc) -{ - lockdep_assert_held(&guc->log.runtime.relay_lock); - - return guc->log.runtime.relay_chan != NULL; -} - -static void guc_move_to_next_buf(struct intel_guc *guc) +static void guc_move_to_next_buf(struct intel_guc_log *log) { /* * Make sure the updates made in the sub buffer are visible when @@ -209,21 +167,15 @@ static void guc_move_to_next_buf(struct intel_guc *guc) */ smp_wmb(); - if (!guc_log_has_relay(guc)) - return; - /* All data has been written, so now move the offset of sub buffer. */ - relay_reserve(guc->log.runtime.relay_chan, guc->log.vma->obj->base.size); + relay_reserve(log->relay.channel, log->vma->obj->base.size); /* Switch to the next sub buffer */ - relay_flush(guc->log.runtime.relay_chan); + relay_flush(log->relay.channel); } -static void *guc_get_write_buffer(struct intel_guc *guc) +static void *guc_get_write_buffer(struct intel_guc_log *log) { - if (!guc_log_has_relay(guc)) - 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 @@ -233,25 +185,25 @@ static void *guc_get_write_buffer(struct intel_guc *guc) * done without using relay_reserve() along with relay_write(). So its * better to use relay_reserve() alone. */ - return relay_reserve(guc->log.runtime.relay_chan, 0); + return relay_reserve(log->relay.channel, 0); } -static bool guc_check_log_buf_overflow(struct intel_guc *guc, +static bool guc_check_log_buf_overflow(struct intel_guc_log *log, enum guc_log_buffer_type type, unsigned int full_cnt) { - unsigned int prev_full_cnt = guc->log.prev_overflow_count[type]; + unsigned int prev_full_cnt = log->stats[type].sampled_overflow; bool overflow = false; if (full_cnt != prev_full_cnt) { overflow = true; - guc->log.prev_overflow_count[type] = full_cnt; - guc->log.total_overflow_count[type] += full_cnt - prev_full_cnt; + log->stats[type].overflow = full_cnt; + log->stats[type].sampled_overflow += full_cnt - prev_full_cnt; if (full_cnt < prev_full_cnt) { /* buffer_full_cnt is a 4 bit counter */ - guc->log.total_overflow_count[type] += 16; + log->stats[type].sampled_overflow += 16; } DRM_ERROR_RATELIMITED("GuC log buffer overflow\n"); } @@ -275,7 +227,7 @@ static unsigned int guc_get_log_buffer_size(enum guc_log_buffer_type type) return 0; } -static void guc_read_update_log_buffer(struct intel_guc *guc) +static void guc_read_update_log_buffer(struct intel_guc_log *log) { unsigned int buffer_size, read_offset, write_offset, bytes_to_copy, full_cnt; struct guc_log_buffer_state *log_buf_state, *log_buf_snapshot_state; @@ -284,16 +236,16 @@ static void guc_read_update_log_buffer(struct intel_guc *guc) void *src_data, *dst_data; bool new_overflow; - if (WARN_ON(!guc->log.runtime.buf_addr)) - return; + mutex_lock(&log->relay.lock); - /* Get the pointer to shared GuC log buffer */ - log_buf_state = src_data = guc->log.runtime.buf_addr; + if (WARN_ON(!intel_guc_log_relay_enabled(log))) + goto out_unlock; - mutex_lock(&guc->log.runtime.relay_lock); + /* Get the pointer to shared GuC log buffer */ + log_buf_state = src_data = log->relay.buf_addr; /* Get the pointer to local buffer to store the logs */ - log_buf_snapshot_state = dst_data = guc_get_write_buffer(guc); + log_buf_snapshot_state = dst_data = guc_get_write_buffer(log); if (unlikely(!log_buf_snapshot_state)) { /* @@ -301,10 +253,9 @@ static void guc_read_update_log_buffer(struct intel_guc *guc) * getting consumed by User at a slow rate. */ DRM_ERROR_RATELIMITED("no sub-buffer to capture logs\n"); - guc->log.capture_miss_count++; - mutex_unlock(&guc->log.runtime.relay_lock); + log->relay.full_count++; - return; + goto out_unlock; } /* Actual logs are present from the 2nd page */ @@ -325,8 +276,8 @@ static void guc_read_update_log_buffer(struct intel_guc *guc) full_cnt = log_buf_state_local.buffer_full_cnt; /* Bookkeeping stuff */ - guc->log.flush_count[type] += log_buf_state_local.flush_to_file; - new_overflow = guc_check_log_buf_overflow(guc, type, full_cnt); + log->stats[type].flush += log_buf_state_local.flush_to_file; + new_overflow = guc_check_log_buf_overflow(log, type, full_cnt); /* Update the state of shared log buffer */ log_buf_state->read_ptr = write_offset; @@ -373,38 +324,35 @@ static void guc_read_update_log_buffer(struct intel_guc *guc) dst_data += buffer_size; } - guc_move_to_next_buf(guc); + guc_move_to_next_buf(log); - mutex_unlock(&guc->log.runtime.relay_lock); +out_unlock: + mutex_unlock(&log->relay.lock); } static void capture_logs_work(struct work_struct *work) { - struct intel_guc *guc = - container_of(work, struct intel_guc, log.runtime.flush_work); - - guc_log_capture_logs(guc); -} + struct intel_guc_log *log = + container_of(work, struct intel_guc_log, relay.flush_work); -static bool guc_log_has_runtime(struct intel_guc *guc) -{ - return guc->log.runtime.buf_addr != NULL; + guc_log_capture_logs(log); } -static int guc_log_runtime_create(struct intel_guc *guc) +static int guc_log_map(struct intel_guc_log *log) { + struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); void *vaddr; int ret; - lockdep_assert_held(&dev_priv->drm.struct_mutex); + lockdep_assert_held(&log->relay.lock); - if (!guc->log.vma) + if (!log->vma) return -ENODEV; - GEM_BUG_ON(guc_log_has_runtime(guc)); - - ret = i915_gem_object_set_to_wc_domain(guc->log.vma->obj, true); + mutex_lock(&dev_priv->drm.struct_mutex); + ret = i915_gem_object_set_to_wc_domain(log->vma->obj, true); + mutex_unlock(&dev_priv->drm.struct_mutex); if (ret) return ret; @@ -413,49 +361,40 @@ static int guc_log_runtime_create(struct intel_guc *guc) * buffer pages, so that we can directly get the data * (up-to-date) from memory. */ - vaddr = i915_gem_object_pin_map(guc->log.vma->obj, I915_MAP_WC); + vaddr = i915_gem_object_pin_map(log->vma->obj, I915_MAP_WC); if (IS_ERR(vaddr)) { DRM_ERROR("Couldn't map log buffer pages %d\n", ret); return PTR_ERR(vaddr); } - guc->log.runtime.buf_addr = vaddr; + log->relay.buf_addr = vaddr; return 0; } -static void guc_log_runtime_destroy(struct intel_guc *guc) +static void guc_log_unmap(struct intel_guc_log *log) { - /* - * It's possible that the runtime stuff was never allocated because - * GuC log was disabled at the boot time. - */ - if (!guc_log_has_runtime(guc)) - return; + lockdep_assert_held(&log->relay.lock); - i915_gem_object_unpin_map(guc->log.vma->obj); - guc->log.runtime.buf_addr = NULL; + i915_gem_object_unpin_map(log->vma->obj); + log->relay.buf_addr = NULL; } -void intel_guc_log_init_early(struct intel_guc *guc) +void intel_guc_log_init_early(struct intel_guc_log *log) { - mutex_init(&guc->log.runtime.relay_lock); - INIT_WORK(&guc->log.runtime.flush_work, capture_logs_work); + mutex_init(&log->relay.lock); + INIT_WORK(&log->relay.flush_work, capture_logs_work); } -int intel_guc_log_relay_create(struct intel_guc *guc) +static int guc_log_relay_create(struct intel_guc_log *log) { + struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); struct rchan *guc_log_relay_chan; size_t n_subbufs, subbuf_size; int ret; - if (!i915_modparams.guc_log_level) - return 0; - - mutex_lock(&guc->log.runtime.relay_lock); - - GEM_BUG_ON(guc_log_has_relay(guc)); + lockdep_assert_held(&log->relay.lock); /* Keep the size of sub buffers same as shared log buffer */ subbuf_size = GUC_LOG_SIZE; @@ -468,157 +407,56 @@ int intel_guc_log_relay_create(struct intel_guc *guc) */ n_subbufs = 8; - /* - * Create a relay channel, so that we have buffers for storing - * the GuC firmware logs, the channel will be linked with a file - * later on when debugfs is registered. - */ - guc_log_relay_chan = relay_open(NULL, NULL, subbuf_size, - n_subbufs, &relay_callbacks, dev_priv); + guc_log_relay_chan = relay_open("guc_log", + dev_priv->drm.primary->debugfs_root, + subbuf_size, n_subbufs, + &relay_callbacks, dev_priv); if (!guc_log_relay_chan) { DRM_ERROR("Couldn't create relay chan for GuC logging\n"); ret = -ENOMEM; - goto err; + return ret; } GEM_BUG_ON(guc_log_relay_chan->subbuf_size < subbuf_size); - guc->log.runtime.relay_chan = guc_log_relay_chan; - - mutex_unlock(&guc->log.runtime.relay_lock); + log->relay.channel = guc_log_relay_chan; return 0; - -err: - mutex_unlock(&guc->log.runtime.relay_lock); - /* logging will be off */ - i915_modparams.guc_log_level = 0; - return ret; -} - -void intel_guc_log_relay_destroy(struct intel_guc *guc) -{ - mutex_lock(&guc->log.runtime.relay_lock); - - /* - * It's possible that the relay was never allocated because - * GuC log was disabled at the boot time. - */ - if (!guc_log_has_relay(guc)) - goto out_unlock; - - relay_close(guc->log.runtime.relay_chan); - guc->log.runtime.relay_chan = NULL; - -out_unlock: - mutex_unlock(&guc->log.runtime.relay_lock); } -static int guc_log_late_setup(struct intel_guc *guc) +static void guc_log_relay_destroy(struct intel_guc_log *log) { - struct drm_i915_private *dev_priv = guc_to_i915(guc); - int ret; - - if (!guc_log_has_runtime(guc)) { - /* - * If log was disabled at boot time, then setup needed to handle - * log buffer flush interrupts would not have been done yet, so - * do that now. - */ - ret = intel_guc_log_relay_create(guc); - if (ret) - goto err; - - mutex_lock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_get(dev_priv); - ret = guc_log_runtime_create(guc); - intel_runtime_pm_put(dev_priv); - mutex_unlock(&dev_priv->drm.struct_mutex); - - if (ret) - goto err_relay; - } - - ret = guc_log_relay_file_create(guc); - if (ret) - goto err_runtime; - - return 0; + lockdep_assert_held(&log->relay.lock); -err_runtime: - mutex_lock(&dev_priv->drm.struct_mutex); - guc_log_runtime_destroy(guc); - mutex_unlock(&dev_priv->drm.struct_mutex); -err_relay: - intel_guc_log_relay_destroy(guc); -err: - /* logging will remain off */ - i915_modparams.guc_log_level = 0; - return ret; + relay_close(log->relay.channel); + log->relay.channel = NULL; } -static void guc_log_capture_logs(struct intel_guc *guc) +static void guc_log_capture_logs(struct intel_guc_log *log) { + struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); - guc_read_update_log_buffer(guc); + guc_read_update_log_buffer(log); /* * Generally device is expected to be active only at this * 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 (!USES_GUC_SUBMISSION(dev_priv) || !i915_modparams.guc_log_level) - return; - - /* First disable the interrupts, will be renabled afterwards */ - mutex_lock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_get(dev_priv); - gen9_disable_guc_interrupts(dev_priv); - intel_runtime_pm_put(dev_priv); - mutex_unlock(&dev_priv->drm.struct_mutex); - - /* - * Before initiating the forceful flush, wait for any pending/ongoing - * flush to complete otherwise forceful flush may not actually happen. - */ - flush_work(&guc->log.runtime.flush_work); - - /* Ask GuC to update the log buffer state */ - intel_runtime_pm_get(dev_priv); - guc_log_flush(guc); + guc_action_flush_log_complete(guc); intel_runtime_pm_put(dev_priv); - - /* GuC would have updated log buffer by now, so capture it */ - guc_log_capture_logs(guc); } -int intel_guc_log_create(struct intel_guc *guc) +int intel_guc_log_create(struct intel_guc_log *log) { + struct intel_guc *guc = log_to_guc(log); struct i915_vma *vma; unsigned long offset; u32 flags; int ret; - GEM_BUG_ON(guc->log.vma); - - /* - * We require SSE 4.1 for fast reads from the GuC log buffer and - * it should be present on the chipsets supporting GuC based - * submisssions. - */ - if (WARN_ON(!i915_has_memcpy_from_wc())) { - ret = -EINVAL; - goto err; - } + GEM_BUG_ON(log->vma); vma = intel_guc_allocate_vma(guc, GUC_LOG_SIZE); if (IS_ERR(vma)) { @@ -626,13 +464,7 @@ int intel_guc_log_create(struct intel_guc *guc) goto err; } - guc->log.vma = vma; - - if (i915_modparams.guc_log_level) { - ret = guc_log_runtime_create(guc); - if (ret < 0) - goto err_vma; - } + log->vma = vma; /* each allocated unit is a page */ flags = GUC_LOG_VALID | GUC_LOG_NOTIFY_ON_HALF_FULL | @@ -640,117 +472,159 @@ int intel_guc_log_create(struct intel_guc *guc) (GUC_LOG_ISR_PAGES << GUC_LOG_ISR_SHIFT) | (GUC_LOG_CRASH_PAGES << GUC_LOG_CRASH_SHIFT); - offset = guc_ggtt_offset(vma) >> PAGE_SHIFT; /* in pages */ - guc->log.flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags; + offset = intel_guc_ggtt_offset(guc, vma) >> PAGE_SHIFT; + log->flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags; return 0; -err_vma: - i915_vma_unpin_and_release(&guc->log.vma); err: /* logging will be off */ i915_modparams.guc_log_level = 0; return ret; } -void intel_guc_log_destroy(struct intel_guc *guc) +void intel_guc_log_destroy(struct intel_guc_log *log) +{ + i915_vma_unpin_and_release(&log->vma); +} + +int intel_guc_log_level_get(struct intel_guc_log *log) { - guc_log_runtime_destroy(guc); - i915_vma_unpin_and_release(&guc->log.vma); + GEM_BUG_ON(!log->vma); + GEM_BUG_ON(i915_modparams.guc_log_level < 0); + + return i915_modparams.guc_log_level; } -int intel_guc_log_control(struct intel_guc *guc, u64 control_val) +int intel_guc_log_level_set(struct intel_guc_log *log, u64 val) { + struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); - bool enable_logging = control_val > 0; - u32 verbosity; int ret; - if (!guc->log.vma) - return -ENODEV; + BUILD_BUG_ON(GUC_LOG_VERBOSITY_MIN != 0); + GEM_BUG_ON(!log->vma); + GEM_BUG_ON(i915_modparams.guc_log_level < 0); - BUILD_BUG_ON(GUC_LOG_VERBOSITY_MIN); - if (control_val > 1 + GUC_LOG_VERBOSITY_MAX) + /* + * GuC is recognizing log levels starting from 0 to max, we're using 0 + * as indication that logging should be disabled. + */ + if (val < GUC_LOG_LEVEL_DISABLED || val > GUC_LOG_LEVEL_MAX) return -EINVAL; - /* This combination doesn't make sense & won't have any effect */ - if (!enable_logging && !i915_modparams.guc_log_level) - return 0; + mutex_lock(&dev_priv->drm.struct_mutex); - verbosity = enable_logging ? control_val - 1 : 0; + if (i915_modparams.guc_log_level == val) { + ret = 0; + goto out_unlock; + } - ret = mutex_lock_interruptible(&dev_priv->drm.struct_mutex); - if (ret) - return ret; intel_runtime_pm_get(dev_priv); - ret = guc_log_control(guc, enable_logging, verbosity); + ret = guc_action_control_log(guc, GUC_LOG_LEVEL_IS_VERBOSE(val), + GUC_LOG_LEVEL_IS_ENABLED(val), + GUC_LOG_LEVEL_TO_VERBOSITY(val)); intel_runtime_pm_put(dev_priv); + if (ret) { + DRM_DEBUG_DRIVER("guc_log_control action failed %d\n", ret); + goto out_unlock; + } + + i915_modparams.guc_log_level = val; + +out_unlock: mutex_unlock(&dev_priv->drm.struct_mutex); - if (ret < 0) { - DRM_DEBUG_DRIVER("guc_logging_control action failed %d\n", ret); - return ret; - } + return ret; +} - if (enable_logging) { - i915_modparams.guc_log_level = 1 + verbosity; +bool intel_guc_log_relay_enabled(const struct intel_guc_log *log) +{ + return log->relay.buf_addr; +} - /* - * If log was disabled at boot time, then the relay channel file - * wouldn't have been created by now and interrupts also would - * not have been enabled. Try again now, just in case. - */ - ret = guc_log_late_setup(guc); - if (ret < 0) { - DRM_DEBUG_DRIVER("GuC log late setup failed %d\n", ret); - return ret; - } +int intel_guc_log_relay_open(struct intel_guc_log *log) +{ + int ret; - /* GuC logging is currently the only user of Guc2Host interrupts */ - mutex_lock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_get(dev_priv); - gen9_enable_guc_interrupts(dev_priv); - intel_runtime_pm_put(dev_priv); - mutex_unlock(&dev_priv->drm.struct_mutex); - } else { - /* - * Once logging is disabled, GuC won't generate logs & send an - * interrupt. But there could be some data in the log buffer - * which is yet to be captured. So request GuC to update the log - * buffer state and then collect the left over logs. - */ - guc_flush_logs(guc); + mutex_lock(&log->relay.lock); - /* As logging is disabled, update log level to reflect that */ - i915_modparams.guc_log_level = 0; + if (intel_guc_log_relay_enabled(log)) { + ret = -EEXIST; + goto out_unlock; } - return ret; -} + /* + * We require SSE 4.1 for fast reads from the GuC log buffer and + * it should be present on the chipsets supporting GuC based + * submisssions. + */ + if (!i915_has_memcpy_from_wc()) { + ret = -ENXIO; + goto out_unlock; + } -void i915_guc_log_register(struct drm_i915_private *dev_priv) -{ - if (!USES_GUC_SUBMISSION(dev_priv) || !i915_modparams.guc_log_level) - return; + ret = guc_log_relay_create(log); + if (ret) + goto out_unlock; + + ret = guc_log_map(log); + if (ret) + goto out_relay; - guc_log_late_setup(&dev_priv->guc); + mutex_unlock(&log->relay.lock); + + guc_log_enable_flush_events(log); + + /* + * When GuC is logging without us relaying to userspace, we're ignoring + * the flush notification. This means that we need to unconditionally + * flush on relay enabling, since GuC only notifies us once. + */ + queue_work(log->relay.flush_wq, &log->relay.flush_work); + + return 0; + +out_relay: + guc_log_relay_destroy(log); +out_unlock: + mutex_unlock(&log->relay.lock); + + return ret; } -void i915_guc_log_unregister(struct drm_i915_private *dev_priv) +void intel_guc_log_relay_flush(struct intel_guc_log *log) { - struct intel_guc *guc = &dev_priv->guc; + struct intel_guc *guc = log_to_guc(log); + struct drm_i915_private *i915 = guc_to_i915(guc); + + /* + * Before initiating the forceful flush, wait for any pending/ongoing + * flush to complete otherwise forceful flush may not actually happen. + */ + flush_work(&log->relay.flush_work); - if (!USES_GUC_SUBMISSION(dev_priv)) - return; + intel_runtime_pm_get(i915); + guc_action_flush_log(guc); + intel_runtime_pm_put(i915); - mutex_lock(&dev_priv->drm.struct_mutex); - /* GuC logging is currently the only user of Guc2Host interrupts */ - intel_runtime_pm_get(dev_priv); - gen9_disable_guc_interrupts(dev_priv); - intel_runtime_pm_put(dev_priv); + /* GuC would have updated log buffer by now, so capture it */ + guc_log_capture_logs(log); +} - guc_log_runtime_destroy(guc); - mutex_unlock(&dev_priv->drm.struct_mutex); +void intel_guc_log_relay_close(struct intel_guc_log *log) +{ + guc_log_disable_flush_events(log); + flush_work(&log->relay.flush_work); + + mutex_lock(&log->relay.lock); + GEM_BUG_ON(!intel_guc_log_relay_enabled(log)); + guc_log_unmap(log); + guc_log_relay_destroy(log); + mutex_unlock(&log->relay.lock); +} - intel_guc_log_relay_destroy(guc); +void intel_guc_log_handle_flush_event(struct intel_guc_log *log) +{ + queue_work(log->relay.flush_wq, &log->relay.flush_work); } diff --git a/drivers/gpu/drm/i915/intel_guc_log.h b/drivers/gpu/drm/i915/intel_guc_log.h index dab0e94..fa80535 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.h +++ b/drivers/gpu/drm/i915/intel_guc_log.h @@ -25,11 +25,12 @@ #ifndef _INTEL_GUC_LOG_H_ #define _INTEL_GUC_LOG_H_ +#include <linux/mutex.h> +#include <linux/relay.h> #include <linux/workqueue.h> #include "intel_guc_fwif.h" -struct drm_i915_private; struct intel_guc; /* @@ -39,33 +40,53 @@ struct intel_guc; #define GUC_LOG_SIZE ((1 + GUC_LOG_DPC_PAGES + 1 + GUC_LOG_ISR_PAGES + \ 1 + GUC_LOG_CRASH_PAGES + 1) << PAGE_SHIFT) +/* + * While we're using plain log level in i915, GuC controls are much more... + * "elaborate"? We have a couple of bits for verbosity, separate bit for actual + * log enabling, and separate bit for default logging - which "conveniently" + * ignores the enable bit. + */ +#define GUC_LOG_LEVEL_DISABLED 0 +#define GUC_LOG_LEVEL_NON_VERBOSE 1 +#define GUC_LOG_LEVEL_IS_ENABLED(x) ((x) > GUC_LOG_LEVEL_DISABLED) +#define GUC_LOG_LEVEL_IS_VERBOSE(x) ((x) > GUC_LOG_LEVEL_NON_VERBOSE) +#define GUC_LOG_LEVEL_TO_VERBOSITY(x) ({ \ + typeof(x) _x = (x); \ + GUC_LOG_LEVEL_IS_VERBOSE(_x) ? _x - 2 : 0; \ +}) +#define GUC_VERBOSITY_TO_LOG_LEVEL(x) ((x) + 2) +#define GUC_LOG_LEVEL_MAX GUC_VERBOSITY_TO_LOG_LEVEL(GUC_LOG_VERBOSITY_MAX) + struct intel_guc_log { u32 flags; struct i915_vma *vma; - /* The runtime stuff gets created only when GuC logging gets enabled */ struct { void *buf_addr; struct workqueue_struct *flush_wq; struct work_struct flush_work; - struct rchan *relay_chan; - /* To serialize the access to relay_chan */ - struct mutex relay_lock; - } runtime; + struct rchan *channel; + struct mutex lock; + u32 full_count; + } relay; /* logging related stats */ - u32 capture_miss_count; - u32 flush_interrupt_count; - u32 prev_overflow_count[GUC_MAX_LOG_BUFFER]; - u32 total_overflow_count[GUC_MAX_LOG_BUFFER]; - u32 flush_count[GUC_MAX_LOG_BUFFER]; + struct { + u32 sampled_overflow; + u32 overflow; + u32 flush; + } stats[GUC_MAX_LOG_BUFFER]; }; -int intel_guc_log_create(struct intel_guc *guc); -void intel_guc_log_destroy(struct intel_guc *guc); -void intel_guc_log_init_early(struct intel_guc *guc); -int intel_guc_log_relay_create(struct intel_guc *guc); -void intel_guc_log_relay_destroy(struct intel_guc *guc); -int intel_guc_log_control(struct intel_guc *guc, u64 control_val); -void i915_guc_log_register(struct drm_i915_private *dev_priv); -void i915_guc_log_unregister(struct drm_i915_private *dev_priv); +void intel_guc_log_init_early(struct intel_guc_log *log); +int intel_guc_log_create(struct intel_guc_log *log); +void intel_guc_log_destroy(struct intel_guc_log *log); + +int intel_guc_log_level_get(struct intel_guc_log *log); +int intel_guc_log_level_set(struct intel_guc_log *log, u64 control_val); +bool intel_guc_log_relay_enabled(const struct intel_guc_log *log); +int intel_guc_log_relay_open(struct intel_guc_log *log); +void intel_guc_log_relay_flush(struct intel_guc_log *log); +void intel_guc_log_relay_close(struct intel_guc_log *log); + +void intel_guc_log_handle_flush_event(struct intel_guc_log *log); #endif diff --git a/drivers/gpu/drm/i915/intel_guc_reg.h b/drivers/gpu/drm/i915/intel_guc_reg.h index 19a9247..d860847 100644 --- a/drivers/gpu/drm/i915/intel_guc_reg.h +++ b/drivers/gpu/drm/i915/intel_guc_reg.h @@ -66,22 +66,20 @@ #define UOS_MOVE (1<<4) #define START_DMA (1<<0) #define DMA_GUC_WOPCM_OFFSET _MMIO(0xc340) +#define GUC_WOPCM_OFFSET_VALID (1<<0) #define HUC_LOADING_AGENT_VCR (0<<1) #define HUC_LOADING_AGENT_GUC (1<<1) -#define GUC_WOPCM_OFFSET_VALUE 0x80000 /* 512KB */ +#define GUC_WOPCM_OFFSET_SHIFT 14 +#define GUC_WOPCM_OFFSET_MASK (0x3ffff << GUC_WOPCM_OFFSET_SHIFT) #define GUC_MAX_IDLE_COUNT _MMIO(0xC3E4) #define HUC_STATUS2 _MMIO(0xD3B0) #define HUC_FW_VERIFIED (1<<7) -/* Defines WOPCM space available to GuC firmware */ #define GUC_WOPCM_SIZE _MMIO(0xc050) -/* GuC addresses below GUC_WOPCM_TOP don't map through the GTT */ -#define GUC_WOPCM_TOP (0x80 << 12) /* 512KB */ -#define BXT_GUC_WOPCM_RC6_RESERVED (0x10 << 12) /* 64KB */ - -/* GuC addresses above GUC_GGTT_TOP also don't map through the GTT */ -#define GUC_GGTT_TOP 0xFEE00000 +#define GUC_WOPCM_SIZE_LOCKED (1<<0) +#define GUC_WOPCM_SIZE_SHIFT 12 +#define GUC_WOPCM_SIZE_MASK (0xfffff << GUC_WOPCM_SIZE_SHIFT) #define GEN8_GT_PM_CONFIG _MMIO(0x138140) #define GEN9LP_GT_PM_CONFIG _MMIO(0x138140) diff --git a/drivers/gpu/drm/i915/intel_guc_submission.c b/drivers/gpu/drm/i915/intel_guc_submission.c index 8a8ad2f..2feb650 100644 --- a/drivers/gpu/drm/i915/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/intel_guc_submission.c @@ -124,9 +124,17 @@ static int reserve_doorbell(struct intel_guc_client *client) return 0; } +static bool has_doorbell(struct intel_guc_client *client) +{ + if (client->doorbell_id == GUC_DOORBELL_INVALID) + return false; + + return test_bit(client->doorbell_id, client->guc->doorbell_bitmap); +} + static void unreserve_doorbell(struct intel_guc_client *client) { - GEM_BUG_ON(client->doorbell_id == GUC_DOORBELL_INVALID); + GEM_BUG_ON(!has_doorbell(client)); __clear_bit(client->doorbell_id, client->guc->doorbell_bitmap); client->doorbell_id = GUC_DOORBELL_INVALID; @@ -184,14 +192,6 @@ static struct guc_doorbell_info *__get_doorbell(struct intel_guc_client *client) return client->vaddr + client->doorbell_offset; } -static bool has_doorbell(struct intel_guc_client *client) -{ - if (client->doorbell_id == GUC_DOORBELL_INVALID) - return false; - - return test_bit(client->doorbell_id, client->guc->doorbell_bitmap); -} - static void __create_doorbell(struct intel_guc_client *client) { struct guc_doorbell_info *doorbell; @@ -207,7 +207,6 @@ static void __destroy_doorbell(struct intel_guc_client *client) struct guc_doorbell_info *doorbell; u16 db_id = client->doorbell_id; - doorbell = __get_doorbell(client); doorbell->db_status = GUC_DOORBELL_DISABLED; doorbell->cookie = 0; @@ -224,6 +223,9 @@ static int create_doorbell(struct intel_guc_client *client) { int ret; + if (WARN_ON(!has_doorbell(client))) + return -ENODEV; /* internal setup error, should never happen */ + __update_doorbell_desc(client, client->doorbell_id); __create_doorbell(client); @@ -231,8 +233,8 @@ static int create_doorbell(struct intel_guc_client *client) if (ret) { __destroy_doorbell(client); __update_doorbell_desc(client, GUC_DOORBELL_INVALID); - DRM_ERROR("Couldn't create client %u doorbell: %d\n", - client->stage_id, ret); + DRM_DEBUG_DRIVER("Couldn't create client %u doorbell: %d\n", + client->stage_id, ret); return ret; } @@ -362,7 +364,7 @@ static void guc_stage_desc_init(struct intel_guc *guc, desc->db_id = client->doorbell_id; for_each_engine_masked(engine, dev_priv, client->engines, tmp) { - struct intel_context *ce = &ctx->engine[engine->id]; + struct intel_context *ce = to_intel_context(ctx, engine); u32 guc_engine_id = engine->guc_id; struct guc_execlist_context *lrc = &desc->lrc[guc_engine_id]; @@ -386,8 +388,8 @@ static void guc_stage_desc_init(struct intel_guc *guc, lrc->context_desc = lower_32_bits(ce->lrc_desc); /* The state page is after PPHWSP */ - lrc->ring_lrca = - guc_ggtt_offset(ce->state) + LRC_STATE_PN * PAGE_SIZE; + lrc->ring_lrca = intel_guc_ggtt_offset(guc, ce->state) + + LRC_STATE_PN * PAGE_SIZE; /* XXX: In direct submission, the GuC wants the HW context id * here. In proxy submission, it wants the stage id @@ -395,7 +397,7 @@ static void guc_stage_desc_init(struct intel_guc *guc, lrc->context_id = (client->stage_id << GUC_ELC_CTXID_OFFSET) | (guc_engine_id << GUC_ELC_ENGINE_OFFSET); - lrc->ring_begin = guc_ggtt_offset(ce->ring->vma); + lrc->ring_begin = intel_guc_ggtt_offset(guc, ce->ring->vma); lrc->ring_end = lrc->ring_begin + ce->ring->size - 1; lrc->ring_next_free_location = lrc->ring_begin; lrc->ring_current_tail_pointer_value = 0; @@ -411,7 +413,7 @@ static void guc_stage_desc_init(struct intel_guc *guc, * The doorbell, process descriptor, and workqueue are all parts * of the client object, which the GuC will reference via the GGTT */ - gfx_addr = guc_ggtt_offset(client->vma); + gfx_addr = intel_guc_ggtt_offset(guc, client->vma); desc->db_trigger_phy = sg_dma_address(client->vma->pages->sgl) + client->doorbell_offset; desc->db_trigger_cpu = ptr_to_u64(__get_doorbell(client)); @@ -584,7 +586,7 @@ static void inject_preempt_context(struct work_struct *work) data[3] = engine->guc_id; data[4] = guc->execbuf_client->priority; data[5] = guc->execbuf_client->stage_id; - data[6] = guc_ggtt_offset(guc->shared_data); + data[6] = intel_guc_ggtt_offset(guc, guc->shared_data); if (WARN_ON(intel_guc_send(guc, data, ARRAY_SIZE(data)))) { execlists_clear_active(&engine->execlists, @@ -657,7 +659,17 @@ static void port_assign(struct execlist_port *port, struct i915_request *rq) port_set(port, i915_request_get(rq)); } -static void guc_dequeue(struct intel_engine_cs *engine) +static inline int rq_prio(const struct i915_request *rq) +{ + return rq->sched.attr.priority; +} + +static inline int port_prio(const struct execlist_port *port) +{ + return rq_prio(port_request(port)); +} + +static bool __guc_dequeue(struct intel_engine_cs *engine) { struct intel_engine_execlists * const execlists = &engine->execlists; struct execlist_port *port = execlists->port; @@ -667,28 +679,29 @@ static void guc_dequeue(struct intel_engine_cs *engine) bool submit = false; struct rb_node *rb; - spin_lock_irq(&engine->timeline->lock); + lockdep_assert_held(&engine->timeline.lock); + rb = execlists->first; GEM_BUG_ON(rb_first(&execlists->queue) != rb); if (port_isset(port)) { - if (engine->i915->preempt_context) { + if (intel_engine_has_preemption(engine)) { struct guc_preempt_work *preempt_work = &engine->i915->guc.preempt_work[engine->id]; + int prio = execlists->queue_priority; - if (execlists->queue_priority > - max(port_request(port)->priotree.priority, 0)) { + if (__execlists_need_preempt(prio, port_prio(port))) { execlists_set_active(execlists, EXECLISTS_ACTIVE_PREEMPT); queue_work(engine->i915->guc.preempt_wq, &preempt_work->work); - goto unlock; + return false; } } port++; if (port_isset(port)) - goto unlock; + return false; } GEM_BUG_ON(port_isset(port)); @@ -696,11 +709,11 @@ static void guc_dequeue(struct intel_engine_cs *engine) struct i915_priolist *p = to_priolist(rb); struct i915_request *rq, *rn; - list_for_each_entry_safe(rq, rn, &p->requests, priotree.link) { + list_for_each_entry_safe(rq, rn, &p->requests, sched.link) { if (last && rq->ctx != last->ctx) { if (port == last_port) { __list_del_many(&p->requests, - &rq->priotree.link); + &rq->sched.link); goto done; } @@ -709,7 +722,7 @@ static void guc_dequeue(struct intel_engine_cs *engine) port++; } - INIT_LIST_HEAD(&rq->priotree.link); + INIT_LIST_HEAD(&rq->sched.link); __i915_request_submit(rq); trace_i915_request_in(rq, port_index(port, execlists)); @@ -726,19 +739,34 @@ static void guc_dequeue(struct intel_engine_cs *engine) done: execlists->queue_priority = rb ? to_priolist(rb)->priority : INT_MIN; execlists->first = rb; - if (submit) { + if (submit) port_assign(port, last); - execlists_set_active(execlists, EXECLISTS_ACTIVE_USER); - guc_submit(engine); - } + if (last) + execlists_user_begin(execlists, execlists->port); /* We must always keep the beast fed if we have work piled up */ GEM_BUG_ON(port_isset(execlists->port) && !execlists_is_active(execlists, EXECLISTS_ACTIVE_USER)); GEM_BUG_ON(execlists->first && !port_isset(execlists->port)); -unlock: - spin_unlock_irq(&engine->timeline->lock); + return submit; +} + +static void guc_dequeue(struct intel_engine_cs *engine) +{ + unsigned long flags; + bool submit; + + local_irq_save(flags); + + spin_lock(&engine->timeline.lock); + submit = __guc_dequeue(engine); + spin_unlock(&engine->timeline.lock); + + if (submit) + guc_submit(engine); + + local_irq_restore(flags); } static void guc_submission_tasklet(unsigned long data) @@ -748,17 +776,20 @@ static void guc_submission_tasklet(unsigned long data) struct execlist_port *port = execlists->port; struct i915_request *rq; - rq = port_request(&port[0]); + rq = port_request(port); while (rq && i915_request_completed(rq)) { trace_i915_request_out(rq); i915_request_put(rq); - execlists_port_complete(execlists, port); - - rq = port_request(&port[0]); + port = execlists_port_complete(execlists, port); + if (port_isset(port)) { + execlists_user_begin(execlists, port); + rq = port_request(port); + } else { + execlists_user_end(execlists); + rq = NULL; + } } - if (!rq) - execlists_clear_active(execlists, EXECLISTS_ACTIVE_USER); if (execlists_is_active(execlists, EXECLISTS_ACTIVE_PREEMPT) && intel_read_status_page(engine, I915_GEM_HWS_PREEMPT_INDEX) == @@ -977,7 +1008,8 @@ static void guc_fill_preempt_context(struct intel_guc *guc) enum intel_engine_id id; for_each_engine(engine, dev_priv, id) { - struct intel_context *ce = &client->owner->engine[id]; + struct intel_context *ce = + to_intel_context(client->owner, engine); u32 addr = intel_hws_preempt_done_address(engine); u32 *cs; diff --git a/drivers/gpu/drm/i915/intel_hangcheck.c b/drivers/gpu/drm/i915/intel_hangcheck.c index c8ea510..d47e346 100644 --- a/drivers/gpu/drm/i915/intel_hangcheck.c +++ b/drivers/gpu/drm/i915/intel_hangcheck.c @@ -246,9 +246,8 @@ engine_stuck(struct intel_engine_cs *engine, u64 acthd) */ tmp = I915_READ_CTL(engine); if (tmp & RING_WAIT) { - i915_handle_error(dev_priv, BIT(engine->id), - "Kicking stuck wait on %s", - engine->name); + i915_handle_error(dev_priv, BIT(engine->id), 0, + "stuck wait on %s", engine->name); I915_WRITE_CTL(engine, tmp); return ENGINE_WAIT_KICK; } @@ -258,8 +257,8 @@ engine_stuck(struct intel_engine_cs *engine, u64 acthd) default: return ENGINE_DEAD; case 1: - i915_handle_error(dev_priv, ALL_ENGINES, - "Kicking stuck semaphore on %s", + i915_handle_error(dev_priv, ALL_ENGINES, 0, + "stuck semaphore on %s", engine->name); I915_WRITE_CTL(engine, tmp); return ENGINE_WAIT_KICK; @@ -357,7 +356,7 @@ static void hangcheck_accumulate_sample(struct intel_engine_cs *engine, break; case ENGINE_DEAD: - if (drm_debug & DRM_UT_DRIVER) { + if (GEM_SHOW_DEBUG()) { struct drm_printer p = drm_debug_printer("hangcheck"); intel_engine_dump(engine, &p, "%s\n", engine->name); } @@ -386,13 +385,13 @@ static void hangcheck_declare_hang(struct drm_i915_private *i915, if (stuck != hung) hung &= ~stuck; len = scnprintf(msg, sizeof(msg), - "%s on ", stuck == hung ? "No progress" : "Hang"); + "%s on ", stuck == hung ? "no progress" : "hang"); for_each_engine_masked(engine, i915, hung, tmp) len += scnprintf(msg + len, sizeof(msg) - len, "%s, ", engine->name); msg[len-2] = '\0'; - return i915_handle_error(i915, hung, "%s", msg); + return i915_handle_error(i915, hung, I915_ERROR_CAPTURE, "%s", msg); } /* @@ -453,6 +452,7 @@ static void i915_hangcheck_elapsed(struct work_struct *work) void intel_engine_init_hangcheck(struct intel_engine_cs *engine) { memset(&engine->hangcheck, 0, sizeof(engine->hangcheck)); + engine->hangcheck.action_timestamp = jiffies; } void intel_hangcheck_init(struct drm_i915_private *i915) diff --git a/drivers/gpu/drm/i915/intel_hdcp.c b/drivers/gpu/drm/i915/intel_hdcp.c index 14ca5d3..2db5da5 100644 --- a/drivers/gpu/drm/i915/intel_hdcp.c +++ b/drivers/gpu/drm/i915/intel_hdcp.c @@ -37,6 +37,43 @@ static int intel_hdcp_poll_ksv_fifo(struct intel_digital_port *intel_dig_port, return 0; } +static bool hdcp_key_loadable(struct drm_i915_private *dev_priv) +{ + struct i915_power_domains *power_domains = &dev_priv->power_domains; + struct i915_power_well *power_well; + enum i915_power_well_id id; + bool enabled = false; + + /* + * On HSW and BDW, Display HW loads the Key as soon as Display resumes. + * On all BXT+, SW can load the keys only when the PW#1 is turned on. + */ + if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) + id = HSW_DISP_PW_GLOBAL; + else + id = SKL_DISP_PW_1; + + mutex_lock(&power_domains->lock); + + /* PG1 (power well #1) needs to be enabled */ + for_each_power_well(dev_priv, power_well) { + if (power_well->id == id) { + enabled = power_well->ops->is_enabled(dev_priv, + power_well); + break; + } + } + mutex_unlock(&power_domains->lock); + + /* + * Another req for hdcp key loadability is enabled state of pll for + * cdclk. Without active crtc we wont land here. So we are assuming that + * cdclk is already on. + */ + + return enabled; +} + static void intel_hdcp_clear_keys(struct drm_i915_private *dev_priv) { I915_WRITE(HDCP_KEY_CONF, HDCP_CLEAR_KEYS_TRIGGER); @@ -142,53 +179,17 @@ bool intel_hdcp_is_ksv_valid(u8 *ksv) return true; } -/* Implements Part 2 of the HDCP authorization procedure */ static -int intel_hdcp_auth_downstream(struct intel_digital_port *intel_dig_port, - const struct intel_hdcp_shim *shim) +int intel_hdcp_validate_v_prime(struct intel_digital_port *intel_dig_port, + const struct intel_hdcp_shim *shim, + u8 *ksv_fifo, u8 num_downstream, u8 *bstatus) { struct drm_i915_private *dev_priv; u32 vprime, sha_text, sha_leftovers, rep_ctl; - u8 bstatus[2], num_downstream, *ksv_fifo; int ret, i, j, sha_idx; dev_priv = intel_dig_port->base.base.dev->dev_private; - ret = intel_hdcp_poll_ksv_fifo(intel_dig_port, shim); - if (ret) { - DRM_ERROR("KSV list failed to become ready (%d)\n", ret); - return ret; - } - - ret = shim->read_bstatus(intel_dig_port, bstatus); - if (ret) - return ret; - - if (DRM_HDCP_MAX_DEVICE_EXCEEDED(bstatus[0]) || - DRM_HDCP_MAX_CASCADE_EXCEEDED(bstatus[1])) { - DRM_ERROR("Max Topology Limit Exceeded\n"); - return -EPERM; - } - - /* - * When repeater reports 0 device count, HDCP1.4 spec allows disabling - * the HDCP encryption. That implies that repeater can't have its own - * display. As there is no consumption of encrypted content in the - * repeater with 0 downstream devices, we are failing the - * authentication. - */ - num_downstream = DRM_HDCP_NUM_DOWNSTREAM(bstatus[0]); - if (num_downstream == 0) - return -EINVAL; - - ksv_fifo = kzalloc(num_downstream * DRM_HDCP_KSV_LEN, GFP_KERNEL); - if (!ksv_fifo) - return -ENOMEM; - - ret = shim->read_ksv_fifo(intel_dig_port, num_downstream, ksv_fifo); - if (ret) - return ret; - /* Process V' values from the receiver */ for (i = 0; i < DRM_HDCP_V_PRIME_NUM_PARTS; i++) { ret = shim->read_v_prime_part(intel_dig_port, i, &vprime); @@ -353,7 +354,8 @@ int intel_hdcp_auth_downstream(struct intel_digital_port *intel_dig_port, return ret; sha_idx += sizeof(sha_text); } else { - DRM_ERROR("Invalid number of leftovers %d\n", sha_leftovers); + DRM_DEBUG_KMS("Invalid number of leftovers %d\n", + sha_leftovers); return -EINVAL; } @@ -381,17 +383,83 @@ int intel_hdcp_auth_downstream(struct intel_digital_port *intel_dig_port, if (intel_wait_for_register(dev_priv, HDCP_REP_CTL, HDCP_SHA1_COMPLETE, HDCP_SHA1_COMPLETE, 1)) { - DRM_ERROR("Timed out waiting for SHA1 complete\n"); + DRM_DEBUG_KMS("Timed out waiting for SHA1 complete\n"); return -ETIMEDOUT; } if (!(I915_READ(HDCP_REP_CTL) & HDCP_SHA1_V_MATCH)) { - DRM_ERROR("SHA-1 mismatch, HDCP failed\n"); + DRM_DEBUG_KMS("SHA-1 mismatch, HDCP failed\n"); return -ENXIO; } + return 0; +} + +/* Implements Part 2 of the HDCP authorization procedure */ +static +int intel_hdcp_auth_downstream(struct intel_digital_port *intel_dig_port, + const struct intel_hdcp_shim *shim) +{ + u8 bstatus[2], num_downstream, *ksv_fifo; + int ret, i, tries = 3; + + ret = intel_hdcp_poll_ksv_fifo(intel_dig_port, shim); + if (ret) { + DRM_ERROR("KSV list failed to become ready (%d)\n", ret); + return ret; + } + + ret = shim->read_bstatus(intel_dig_port, bstatus); + if (ret) + return ret; + + if (DRM_HDCP_MAX_DEVICE_EXCEEDED(bstatus[0]) || + DRM_HDCP_MAX_CASCADE_EXCEEDED(bstatus[1])) { + DRM_ERROR("Max Topology Limit Exceeded\n"); + return -EPERM; + } + + /* + * When repeater reports 0 device count, HDCP1.4 spec allows disabling + * the HDCP encryption. That implies that repeater can't have its own + * display. As there is no consumption of encrypted content in the + * repeater with 0 downstream devices, we are failing the + * authentication. + */ + num_downstream = DRM_HDCP_NUM_DOWNSTREAM(bstatus[0]); + if (num_downstream == 0) + return -EINVAL; + + ksv_fifo = kzalloc(num_downstream * DRM_HDCP_KSV_LEN, GFP_KERNEL); + if (!ksv_fifo) + return -ENOMEM; + + ret = shim->read_ksv_fifo(intel_dig_port, num_downstream, ksv_fifo); + if (ret) + goto err; + + /* + * When V prime mismatches, DP Spec mandates re-read of + * V prime atleast twice. + */ + for (i = 0; i < tries; i++) { + ret = intel_hdcp_validate_v_prime(intel_dig_port, shim, + ksv_fifo, num_downstream, + bstatus); + if (!ret) + break; + } + + if (i == tries) { + DRM_ERROR("V Prime validation failed.(%d)\n", ret); + goto err; + } + DRM_DEBUG_KMS("HDCP is enabled (%d downstream devices)\n", num_downstream); - return 0; + ret = 0; +err: + kfree(ksv_fifo); + return ret; } /* Implements Part 1 of the HDCP authorization procedure */ @@ -506,15 +574,26 @@ static int intel_hdcp_auth(struct intel_digital_port *intel_dig_port, */ wait_remaining_ms_from_jiffies(r0_prime_gen_start, 300); - ri.reg = 0; - ret = shim->read_ri_prime(intel_dig_port, ri.shim); - if (ret) - return ret; - I915_WRITE(PORT_HDCP_RPRIME(port), ri.reg); + tries = 3; - /* Wait for Ri prime match */ - if (wait_for(I915_READ(PORT_HDCP_STATUS(port)) & - (HDCP_STATUS_RI_MATCH | HDCP_STATUS_ENC), 1)) { + /* + * DP HDCP Spec mandates the two more reattempt to read R0, incase + * of R0 mismatch. + */ + for (i = 0; i < tries; i++) { + ri.reg = 0; + ret = shim->read_ri_prime(intel_dig_port, ri.shim); + if (ret) + return ret; + I915_WRITE(PORT_HDCP_RPRIME(port), ri.reg); + + /* Wait for Ri prime match */ + if (!wait_for(I915_READ(PORT_HDCP_STATUS(port)) & + (HDCP_STATUS_RI_MATCH | HDCP_STATUS_ENC), 1)) + break; + } + + if (i == tries) { DRM_ERROR("Timed out waiting for Ri prime match (%x)\n", I915_READ(PORT_HDCP_STATUS(port))); return -ETIMEDOUT; @@ -580,8 +659,8 @@ static int _intel_hdcp_enable(struct intel_connector *connector) DRM_DEBUG_KMS("[%s:%d] HDCP is being enabled...\n", connector->base.name, connector->base.base.id); - if (!(I915_READ(SKL_FUSE_STATUS) & SKL_FUSE_PG_DIST_STATUS(1))) { - DRM_ERROR("PG1 is disabled, cannot load keys\n"); + if (!hdcp_key_loadable(dev_priv)) { + DRM_ERROR("HDCP key Load is not possible\n"); return -ENXIO; } diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c index 1baef4a..ee929f3 100644 --- a/drivers/gpu/drm/i915/intel_hdmi.c +++ b/drivers/gpu/drm/i915/intel_hdmi.c @@ -2082,41 +2082,33 @@ intel_hdmi_add_properties(struct intel_hdmi *intel_hdmi, struct drm_connector *c * it enables scrambling. This should be called before enabling the HDMI * 2.0 port, as the sink can choose to disable the scrambling if it doesn't * detect a scrambled clock within 100 ms. + * + * Returns: + * True on success, false on failure. */ -void intel_hdmi_handle_sink_scrambling(struct intel_encoder *encoder, +bool intel_hdmi_handle_sink_scrambling(struct intel_encoder *encoder, struct drm_connector *connector, bool high_tmds_clock_ratio, bool scrambling) { + struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base); - struct drm_i915_private *dev_priv = connector->dev->dev_private; struct drm_scrambling *sink_scrambling = - &connector->display_info.hdmi.scdc.scrambling; - struct i2c_adapter *adptr = intel_gmbus_get_adapter(dev_priv, - intel_hdmi->ddc_bus); - bool ret; + &connector->display_info.hdmi.scdc.scrambling; + struct i2c_adapter *adapter = + intel_gmbus_get_adapter(dev_priv, intel_hdmi->ddc_bus); if (!sink_scrambling->supported) - return; - - DRM_DEBUG_KMS("Setting sink scrambling for enc:%s connector:%s\n", - encoder->base.name, connector->name); + return true; - /* Set TMDS bit clock ratio to 1/40 or 1/10 */ - ret = drm_scdc_set_high_tmds_clock_ratio(adptr, high_tmds_clock_ratio); - if (!ret) { - DRM_ERROR("Set TMDS ratio failed\n"); - return; - } - - /* Enable/disable sink scrambling */ - ret = drm_scdc_set_scrambling(adptr, scrambling); - if (!ret) { - DRM_ERROR("Set sink scrambling failed\n"); - return; - } + DRM_DEBUG_KMS("[CONNECTOR:%d:%s] scrambling=%s, TMDS bit clock ratio=1/%d\n", + connector->base.id, connector->name, + yesno(scrambling), high_tmds_clock_ratio ? 40 : 10); - DRM_DEBUG_KMS("sink scrambling handled\n"); + /* Set TMDS bit clock ratio to 1/40 or 1/10, and enable/disable scrambling */ + return drm_scdc_set_high_tmds_clock_ratio(adapter, + high_tmds_clock_ratio) && + drm_scdc_set_scrambling(adapter, scrambling); } static u8 chv_port_to_ddc_pin(struct drm_i915_private *dev_priv, enum port port) diff --git a/drivers/gpu/drm/i915/intel_hotplug.c b/drivers/gpu/drm/i915/intel_hotplug.c index 0e3d3e8..43aa92b 100644 --- a/drivers/gpu/drm/i915/intel_hotplug.c +++ b/drivers/gpu/drm/i915/intel_hotplug.c @@ -100,6 +100,8 @@ enum port intel_hpd_pin_to_port(struct drm_i915_private *dev_priv, if (IS_CNL_WITH_PORT_F(dev_priv)) return PORT_F; return PORT_E; + case HPD_PORT_F: + return PORT_F; default: return PORT_NONE; /* no port for this pin */ } @@ -132,6 +134,7 @@ enum hpd_pin intel_hpd_pin_default(struct drm_i915_private *dev_priv, case PORT_F: if (IS_CNL_WITH_PORT_F(dev_priv)) return HPD_PORT_E; + return HPD_PORT_F; default: MISSING_CASE(port); return HPD_NONE; diff --git a/drivers/gpu/drm/i915/intel_huc.c b/drivers/gpu/drm/i915/intel_huc.c index 65e2afb..2912852 100644 --- a/drivers/gpu/drm/i915/intel_huc.c +++ b/drivers/gpu/drm/i915/intel_huc.c @@ -55,7 +55,7 @@ int intel_huc_auth(struct intel_huc *huc) return -ENOEXEC; vma = i915_gem_object_ggtt_pin(huc->fw.obj, NULL, 0, 0, - PIN_OFFSET_BIAS | GUC_WOPCM_TOP); + PIN_OFFSET_BIAS | guc->ggtt_pin_bias); if (IS_ERR(vma)) { ret = PTR_ERR(vma); DRM_ERROR("HuC: Failed to pin huc fw object %d\n", ret); @@ -63,7 +63,8 @@ int intel_huc_auth(struct intel_huc *huc) } ret = intel_guc_auth_huc(guc, - guc_ggtt_offset(vma) + huc->fw.rsa_offset); + intel_guc_ggtt_offset(guc, vma) + + huc->fw.rsa_offset); if (ret) { DRM_ERROR("HuC: GuC did not ack Auth request %d\n", ret); goto fail_unpin; @@ -91,3 +92,28 @@ fail: DRM_ERROR("HuC: Authentication failed %d\n", ret); return ret; } + +/** + * intel_huc_check_status() - check HuC status + * @huc: intel_huc structure + * + * This function reads status register to verify if HuC + * firmware was successfully loaded. + * + * Returns positive value if HuC firmware is loaded and verified + * and -ENODEV if HuC is not present. + */ +int intel_huc_check_status(struct intel_huc *huc) +{ + struct drm_i915_private *dev_priv = huc_to_i915(huc); + u32 status; + + if (!HAS_HUC(dev_priv)) + return -ENODEV; + + intel_runtime_pm_get(dev_priv); + status = I915_READ(HUC_STATUS2) & HUC_FW_VERIFIED; + intel_runtime_pm_put(dev_priv); + + return status; +} diff --git a/drivers/gpu/drm/i915/intel_huc.h b/drivers/gpu/drm/i915/intel_huc.h index 5d6e804..aa85490 100644 --- a/drivers/gpu/drm/i915/intel_huc.h +++ b/drivers/gpu/drm/i915/intel_huc.h @@ -37,5 +37,12 @@ struct intel_huc { void intel_huc_init_early(struct intel_huc *huc); int intel_huc_auth(struct intel_huc *huc); +int intel_huc_check_status(struct intel_huc *huc); + +static inline int intel_huc_sanitize(struct intel_huc *huc) +{ + intel_uc_fw_sanitize(&huc->fw); + return 0; +} #endif diff --git a/drivers/gpu/drm/i915/intel_huc_fw.c b/drivers/gpu/drm/i915/intel_huc_fw.c index c66afa9..f93d238 100644 --- a/drivers/gpu/drm/i915/intel_huc_fw.c +++ b/drivers/gpu/drm/i915/intel_huc_fw.c @@ -118,7 +118,8 @@ static int huc_fw_xfer(struct intel_uc_fw *huc_fw, struct i915_vma *vma) intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); /* Set the source address for the uCode */ - offset = guc_ggtt_offset(vma) + huc_fw->header_offset; + offset = intel_guc_ggtt_offset(&dev_priv->guc, vma) + + huc_fw->header_offset; I915_WRITE(DMA_ADDR_0_LOW, lower_32_bits(offset)); I915_WRITE(DMA_ADDR_0_HIGH, upper_32_bits(offset) & 0xFFFF); @@ -154,9 +155,8 @@ static int huc_fw_xfer(struct intel_uc_fw *huc_fw, struct i915_vma *vma) * Called from intel_uc_init_hw() during driver load, resume from sleep and * after a GPU reset. Note that HuC must be loaded before GuC. * - * The firmware image should have already been fetched into memory by the - * earlier call to intel_uc_init_fw(), so here we need to only check that - * fetch succeeded, and then transfer the image to the h/w. + * The firmware image should have already been fetched into memory, so only + * check that fetch succeeded, and then transfer the image to the h/w. * * Return: non-zero code on error */ diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 8704f7f..15434ca 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -139,6 +139,7 @@ #include "i915_gem_render_state.h" #include "intel_lrc_reg.h" #include "intel_mocs.h" +#include "intel_workarounds.h" #define RING_EXECLIST_QFULL (1 << 0x2) #define RING_EXECLIST1_VALID (1 << 0x3) @@ -176,14 +177,16 @@ static inline struct i915_priolist *to_priolist(struct rb_node *rb) static inline int rq_prio(const struct i915_request *rq) { - return rq->priotree.priority; + return rq->sched.attr.priority; } static inline bool need_preempt(const struct intel_engine_cs *engine, const struct i915_request *last, int prio) { - return engine->i915->preempt_context && prio > max(rq_prio(last), 0); + return (intel_engine_has_preemption(engine) && + __execlists_need_preempt(prio, rq_prio(last)) && + !i915_request_completed(last)); } /** @@ -221,7 +224,7 @@ static void intel_lr_context_descriptor_update(struct i915_gem_context *ctx, struct intel_engine_cs *engine) { - struct intel_context *ce = &ctx->engine[engine->id]; + struct intel_context *ce = to_intel_context(ctx, engine); u64 desc; BUILD_BUG_ON(MAX_CONTEXT_HW_ID > (BIT(GEN8_CTX_ID_WIDTH))); @@ -255,9 +258,7 @@ intel_lr_context_descriptor_update(struct i915_gem_context *ctx, } static struct i915_priolist * -lookup_priolist(struct intel_engine_cs *engine, - struct i915_priotree *pt, - int prio) +lookup_priolist(struct intel_engine_cs *engine, int prio) { struct intel_engine_execlists * const execlists = &engine->execlists; struct i915_priolist *p; @@ -328,10 +329,10 @@ static void __unwind_incomplete_requests(struct intel_engine_cs *engine) struct i915_priolist *uninitialized_var(p); int last_prio = I915_PRIORITY_INVALID; - lockdep_assert_held(&engine->timeline->lock); + lockdep_assert_held(&engine->timeline.lock); list_for_each_entry_safe_reverse(rq, rn, - &engine->timeline->requests, + &engine->timeline.requests, link) { if (i915_request_completed(rq)) return; @@ -342,10 +343,11 @@ static void __unwind_incomplete_requests(struct intel_engine_cs *engine) GEM_BUG_ON(rq_prio(rq) == I915_PRIORITY_INVALID); if (rq_prio(rq) != last_prio) { last_prio = rq_prio(rq); - p = lookup_priolist(engine, &rq->priotree, last_prio); + p = lookup_priolist(engine, last_prio); } - list_add(&rq->priotree.link, &p->requests); + GEM_BUG_ON(p->priority != rq_prio(rq)); + list_add(&rq->sched.link, &p->requests); } } @@ -354,10 +356,13 @@ execlists_unwind_incomplete_requests(struct intel_engine_execlists *execlists) { struct intel_engine_cs *engine = container_of(execlists, typeof(*engine), execlists); + unsigned long flags; + + spin_lock_irqsave(&engine->timeline.lock, flags); - spin_lock_irq(&engine->timeline->lock); __unwind_incomplete_requests(engine); - spin_unlock_irq(&engine->timeline->lock); + + spin_unlock_irqrestore(&engine->timeline.lock, flags); } static inline void @@ -374,6 +379,19 @@ execlists_context_status_change(struct i915_request *rq, unsigned long status) status, rq); } +inline void +execlists_user_begin(struct intel_engine_execlists *execlists, + const struct execlist_port *port) +{ + execlists_set_active_once(execlists, EXECLISTS_ACTIVE_USER); +} + +inline void +execlists_user_end(struct intel_engine_execlists *execlists) +{ + execlists_clear_active(execlists, EXECLISTS_ACTIVE_USER); +} + static inline void execlists_context_schedule_in(struct i915_request *rq) { @@ -382,10 +400,11 @@ execlists_context_schedule_in(struct i915_request *rq) } static inline void -execlists_context_schedule_out(struct i915_request *rq) +execlists_context_schedule_out(struct i915_request *rq, unsigned long status) { intel_engine_context_out(rq->engine); - execlists_context_status_change(rq, INTEL_CONTEXT_SCHEDULE_OUT); + execlists_context_status_change(rq, status); + trace_i915_request_out(rq); } static void @@ -399,7 +418,7 @@ execlists_update_context_pdps(struct i915_hw_ppgtt *ppgtt, u32 *reg_state) static u64 execlists_update_context(struct i915_request *rq) { - struct intel_context *ce = &rq->ctx->engine[rq->engine->id]; + struct intel_context *ce = to_intel_context(rq->ctx, rq->engine); struct i915_hw_ppgtt *ppgtt = rq->ctx->ppgtt ?: rq->i915->mm.aliasing_ppgtt; u32 *reg_state = ce->lrc_reg_state; @@ -454,10 +473,12 @@ static void execlists_submit_ports(struct intel_engine_cs *engine) desc = execlists_update_context(rq); GEM_DEBUG_EXEC(port[n].context_id = upper_32_bits(desc)); - GEM_TRACE("%s in[%d]: ctx=%d.%d, seqno=%x, prio=%d\n", + GEM_TRACE("%s in[%d]: ctx=%d.%d, global=%d (fence %llx:%d) (current %d), prio=%d\n", engine->name, n, port[n].context_id, count, rq->global_seqno, + rq->fence.context, rq->fence.seqno, + intel_engine_get_seqno(engine), rq_prio(rq)); } else { GEM_BUG_ON(!n); @@ -506,7 +527,7 @@ static void inject_preempt_context(struct intel_engine_cs *engine) { struct intel_engine_execlists *execlists = &engine->execlists; struct intel_context *ce = - &engine->i915->preempt_context->engine[engine->id]; + to_intel_context(engine->i915->preempt_context, engine); unsigned int n; GEM_BUG_ON(execlists->preempt_complete_status != @@ -535,7 +556,7 @@ static void inject_preempt_context(struct intel_engine_cs *engine) execlists_set_active(&engine->execlists, EXECLISTS_ACTIVE_PREEMPT); } -static void execlists_dequeue(struct intel_engine_cs *engine) +static bool __execlists_dequeue(struct intel_engine_cs *engine) { struct intel_engine_execlists * const execlists = &engine->execlists; struct execlist_port *port = execlists->port; @@ -545,6 +566,8 @@ static void execlists_dequeue(struct intel_engine_cs *engine) struct rb_node *rb; bool submit = false; + lockdep_assert_held(&engine->timeline.lock); + /* Hardware submission is through 2 ports. Conceptually each port * has a (RING_START, RING_HEAD, RING_TAIL) tuple. RING_START is * static for a context, and unique to each, so we only execute @@ -566,7 +589,6 @@ static void execlists_dequeue(struct intel_engine_cs *engine) * and context switches) submission. */ - spin_lock_irq(&engine->timeline->lock); rb = execlists->first; GEM_BUG_ON(rb_first(&execlists->queue) != rb); @@ -581,7 +603,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine) EXECLISTS_ACTIVE_USER)); GEM_BUG_ON(!port_count(&port[0])); if (port_count(&port[0]) > 1) - goto unlock; + return false; /* * If we write to ELSP a second time before the HW has had @@ -591,11 +613,11 @@ static void execlists_dequeue(struct intel_engine_cs *engine) * the HW to indicate that it has had a chance to respond. */ if (!execlists_is_active(execlists, EXECLISTS_ACTIVE_HWACK)) - goto unlock; + return false; if (need_preempt(engine, last, execlists->queue_priority)) { inject_preempt_context(engine); - goto unlock; + return false; } /* @@ -620,7 +642,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine) * priorities of the ports haven't been switch. */ if (port_count(&port[1])) - goto unlock; + return false; /* * WaIdleLiteRestore:bdw,skl @@ -637,7 +659,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine) struct i915_priolist *p = to_priolist(rb); struct i915_request *rq, *rn; - list_for_each_entry_safe(rq, rn, &p->requests, priotree.link) { + list_for_each_entry_safe(rq, rn, &p->requests, sched.link) { /* * Can we combine this request with the current port? * It has to be the same context/ringbuffer and not @@ -657,7 +679,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine) */ if (port == last_port) { __list_del_many(&p->requests, - &rq->priotree.link); + &rq->sched.link); goto done; } @@ -671,7 +693,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine) if (ctx_single_port_submission(last->ctx) || ctx_single_port_submission(rq->ctx)) { __list_del_many(&p->requests, - &rq->priotree.link); + &rq->sched.link); goto done; } @@ -684,7 +706,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine) GEM_BUG_ON(port_isset(port)); } - INIT_LIST_HEAD(&rq->priotree.link); + INIT_LIST_HEAD(&rq->sched.link); __i915_request_submit(rq); trace_i915_request_in(rq, port_index(port, execlists)); last = rq; @@ -697,8 +719,27 @@ static void execlists_dequeue(struct intel_engine_cs *engine) if (p->priority != I915_PRIORITY_NORMAL) kmem_cache_free(engine->i915->priorities, p); } + done: - execlists->queue_priority = rb ? to_priolist(rb)->priority : INT_MIN; + /* + * Here be a bit of magic! Or sleight-of-hand, whichever you prefer. + * + * We choose queue_priority such that if we add a request of greater + * priority than this, we kick the submission tasklet to decide on + * the right order of submitting the requests to hardware. We must + * also be prepared to reorder requests as they are in-flight on the + * HW. We derive the queue_priority then as the first "hole" in + * the HW submission ports and if there are no available slots, + * the priority of the lowest executing request, i.e. last. + * + * When we do receive a higher priority request ready to run from the + * user, see queue_request(), the queue_priority is bumped to that + * request triggering preemption on the next dequeue (or subsequent + * interrupt for secondary ports). + */ + execlists->queue_priority = + port != execlists->port ? rq_prio(last) : INT_MIN; + execlists->first = rb; if (submit) port_assign(port, last); @@ -706,13 +747,25 @@ done: /* We must always keep the beast fed if we have work piled up */ GEM_BUG_ON(execlists->first && !port_isset(execlists->port)); -unlock: - spin_unlock_irq(&engine->timeline->lock); + /* Re-evaluate the executing context setup after each preemptive kick */ + if (last) + execlists_user_begin(execlists, execlists->port); - if (submit) { - execlists_set_active(execlists, EXECLISTS_ACTIVE_USER); + return submit; +} + +static void execlists_dequeue(struct intel_engine_cs *engine) +{ + struct intel_engine_execlists * const execlists = &engine->execlists; + unsigned long flags; + bool submit; + + spin_lock_irqsave(&engine->timeline.lock, flags); + submit = __execlists_dequeue(engine); + spin_unlock_irqrestore(&engine->timeline.lock, flags); + + if (submit) execlists_submit_ports(engine); - } GEM_BUG_ON(port_isset(execlists->port) && !execlists_is_active(execlists, EXECLISTS_ACTIVE_USER)); @@ -727,13 +780,18 @@ execlists_cancel_port_requests(struct intel_engine_execlists * const execlists) while (num_ports-- && port_isset(port)) { struct i915_request *rq = port_request(port); - GEM_BUG_ON(!execlists->active); - intel_engine_context_out(rq->engine); + GEM_TRACE("%s:port%u global=%d (fence %llx:%d), (current %d)\n", + rq->engine->name, + (unsigned int)(port - execlists->port), + rq->global_seqno, + rq->fence.context, rq->fence.seqno, + intel_engine_get_seqno(rq->engine)); - execlists_context_status_change(rq, - i915_request_completed(rq) ? - INTEL_CONTEXT_SCHEDULE_OUT : - INTEL_CONTEXT_SCHEDULE_PREEMPTED); + GEM_BUG_ON(!execlists->active); + execlists_context_schedule_out(rq, + i915_request_completed(rq) ? + INTEL_CONTEXT_SCHEDULE_OUT : + INTEL_CONTEXT_SCHEDULE_PREEMPTED); i915_request_put(rq); @@ -741,7 +799,82 @@ execlists_cancel_port_requests(struct intel_engine_execlists * const execlists) port++; } - execlists_clear_active(execlists, EXECLISTS_ACTIVE_USER); + execlists_user_end(execlists); +} + +static void clear_gtiir(struct intel_engine_cs *engine) +{ + struct drm_i915_private *dev_priv = engine->i915; + int i; + + /* + * Clear any pending interrupt state. + * + * We do it twice out of paranoia that some of the IIR are + * double buffered, and so if we only reset it once there may + * still be an interrupt pending. + */ + if (INTEL_GEN(dev_priv) >= 11) { + static const struct { + u8 bank; + u8 bit; + } gen11_gtiir[] = { + [RCS] = {0, GEN11_RCS0}, + [BCS] = {0, GEN11_BCS}, + [_VCS(0)] = {1, GEN11_VCS(0)}, + [_VCS(1)] = {1, GEN11_VCS(1)}, + [_VCS(2)] = {1, GEN11_VCS(2)}, + [_VCS(3)] = {1, GEN11_VCS(3)}, + [_VECS(0)] = {1, GEN11_VECS(0)}, + [_VECS(1)] = {1, GEN11_VECS(1)}, + }; + unsigned long irqflags; + + GEM_BUG_ON(engine->id >= ARRAY_SIZE(gen11_gtiir)); + + spin_lock_irqsave(&dev_priv->irq_lock, irqflags); + for (i = 0; i < 2; i++) { + gen11_reset_one_iir(dev_priv, + gen11_gtiir[engine->id].bank, + gen11_gtiir[engine->id].bit); + } + spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + } else { + static const u8 gtiir[] = { + [RCS] = 0, + [BCS] = 0, + [VCS] = 1, + [VCS2] = 1, + [VECS] = 3, + }; + + GEM_BUG_ON(engine->id >= ARRAY_SIZE(gtiir)); + + for (i = 0; i < 2; i++) { + I915_WRITE(GEN8_GT_IIR(gtiir[engine->id]), + engine->irq_keep_mask); + POSTING_READ(GEN8_GT_IIR(gtiir[engine->id])); + } + GEM_BUG_ON(I915_READ(GEN8_GT_IIR(gtiir[engine->id])) & + engine->irq_keep_mask); + } +} + +static void reset_irq(struct intel_engine_cs *engine) +{ + /* Mark all CS interrupts as complete */ + smp_store_mb(engine->execlists.active, 0); + synchronize_hardirq(engine->i915->drm.irq); + + clear_gtiir(engine); + + /* + * The port is checked prior to scheduling a tasklet, but + * just in case we have suspended the tasklet to do the + * wedging make sure that when it wakes, it decides there + * is no work to do by clearing the irq_posted bit. + */ + clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); } static void execlists_cancel_requests(struct intel_engine_cs *engine) @@ -751,7 +884,8 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) struct rb_node *rb; unsigned long flags; - GEM_TRACE("%s\n", engine->name); + GEM_TRACE("%s current %d\n", + engine->name, intel_engine_get_seqno(engine)); /* * Before we call engine->cancel_requests(), we should have exclusive @@ -771,11 +905,12 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) /* Cancel the requests on the HW and clear the ELSP tracker. */ execlists_cancel_port_requests(execlists); + reset_irq(engine); - spin_lock(&engine->timeline->lock); + spin_lock(&engine->timeline.lock); /* Mark all executing requests as skipped. */ - list_for_each_entry(rq, &engine->timeline->requests, link) { + list_for_each_entry(rq, &engine->timeline.requests, link) { GEM_BUG_ON(!rq->global_seqno); if (!i915_request_completed(rq)) dma_fence_set_error(&rq->fence, -EIO); @@ -786,8 +921,8 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) while (rb) { struct i915_priolist *p = to_priolist(rb); - list_for_each_entry_safe(rq, rn, &p->requests, priotree.link) { - INIT_LIST_HEAD(&rq->priotree.link); + list_for_each_entry_safe(rq, rn, &p->requests, sched.link) { + INIT_LIST_HEAD(&rq->sched.link); dma_fence_set_error(&rq->fence, -EIO); __i915_request_submit(rq); @@ -807,18 +942,7 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) execlists->first = NULL; GEM_BUG_ON(port_isset(execlists->port)); - spin_unlock(&engine->timeline->lock); - - /* - * The port is checked prior to scheduling a tasklet, but - * just in case we have suspended the tasklet to do the - * wedging make sure that when it wakes, it decides there - * is no work to do by clearing the irq_posted bit. - */ - clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); - - /* Mark all CS interrupts as complete */ - execlists->active = 0; + spin_unlock(&engine->timeline.lock); local_irq_restore(flags); } @@ -831,7 +955,7 @@ static void execlists_submission_tasklet(unsigned long data) { struct intel_engine_cs * const engine = (struct intel_engine_cs *)data; struct intel_engine_execlists * const execlists = &engine->execlists; - struct execlist_port * const port = execlists->port; + struct execlist_port *port = execlists->port; struct drm_i915_private *dev_priv = engine->i915; bool fw = false; @@ -959,10 +1083,13 @@ static void execlists_submission_tasklet(unsigned long data) EXECLISTS_ACTIVE_USER)); rq = port_unpack(port, &count); - GEM_TRACE("%s out[0]: ctx=%d.%d, seqno=%x, prio=%d\n", + GEM_TRACE("%s out[0]: ctx=%d.%d, global=%d (fence %llx:%d) (current %d), prio=%d\n", engine->name, port->context_id, count, rq ? rq->global_seqno : 0, + rq ? rq->fence.context : 0, + rq ? rq->fence.seqno : 0, + intel_engine_get_seqno(engine), rq ? rq_prio(rq) : 0); /* Check the context/desc id for this event matches */ @@ -970,28 +1097,43 @@ static void execlists_submission_tasklet(unsigned long data) GEM_BUG_ON(count == 0); if (--count == 0) { + /* + * On the final event corresponding to the + * submission of this context, we expect either + * an element-switch event or a completion + * event (and on completion, the active-idle + * marker). No more preemptions, lite-restore + * or otherwise. + */ GEM_BUG_ON(status & GEN8_CTX_STATUS_PREEMPTED); GEM_BUG_ON(port_isset(&port[1]) && !(status & GEN8_CTX_STATUS_ELEMENT_SWITCH)); + GEM_BUG_ON(!port_isset(&port[1]) && + !(status & GEN8_CTX_STATUS_ACTIVE_IDLE)); + + /* + * We rely on the hardware being strongly + * ordered, that the breadcrumb write is + * coherent (visible from the CPU) before the + * user interrupt and CSB is processed. + */ GEM_BUG_ON(!i915_request_completed(rq)); - execlists_context_schedule_out(rq); - trace_i915_request_out(rq); + + execlists_context_schedule_out(rq, + INTEL_CONTEXT_SCHEDULE_OUT); i915_request_put(rq); GEM_TRACE("%s completed ctx=%d\n", engine->name, port->context_id); - execlists_port_complete(execlists, port); + port = execlists_port_complete(execlists, port); + if (port_isset(port)) + execlists_user_begin(execlists, port); + else + execlists_user_end(execlists); } else { port_set(port, port_pack(rq, count)); } - - /* After the final element, the hw should be idle */ - GEM_BUG_ON(port_count(port) == 0 && - !(status & GEN8_CTX_STATUS_ACTIVE_IDLE)); - if (port_count(port) == 0) - execlists_clear_active(execlists, - EXECLISTS_ACTIVE_USER); } if (head != execlists->csb_head) { @@ -1014,18 +1156,23 @@ static void execlists_submission_tasklet(unsigned long data) } static void queue_request(struct intel_engine_cs *engine, - struct i915_priotree *pt, + struct i915_sched_node *node, int prio) { - list_add_tail(&pt->link, &lookup_priolist(engine, pt, prio)->requests); + list_add_tail(&node->link, + &lookup_priolist(engine, prio)->requests); +} + +static void __submit_queue(struct intel_engine_cs *engine, int prio) +{ + engine->execlists.queue_priority = prio; + tasklet_hi_schedule(&engine->execlists.tasklet); } static void submit_queue(struct intel_engine_cs *engine, int prio) { - if (prio > engine->execlists.queue_priority) { - engine->execlists.queue_priority = prio; - tasklet_hi_schedule(&engine->execlists.tasklet); - } + if (prio > engine->execlists.queue_priority) + __submit_queue(engine, prio); } static void execlists_submit_request(struct i915_request *request) @@ -1034,42 +1181,45 @@ static void execlists_submit_request(struct i915_request *request) unsigned long flags; /* Will be called from irq-context when using foreign fences. */ - spin_lock_irqsave(&engine->timeline->lock, flags); + spin_lock_irqsave(&engine->timeline.lock, flags); - queue_request(engine, &request->priotree, rq_prio(request)); + queue_request(engine, &request->sched, rq_prio(request)); submit_queue(engine, rq_prio(request)); GEM_BUG_ON(!engine->execlists.first); - GEM_BUG_ON(list_empty(&request->priotree.link)); + GEM_BUG_ON(list_empty(&request->sched.link)); - spin_unlock_irqrestore(&engine->timeline->lock, flags); + spin_unlock_irqrestore(&engine->timeline.lock, flags); } -static struct i915_request *pt_to_request(struct i915_priotree *pt) +static struct i915_request *sched_to_request(struct i915_sched_node *node) { - return container_of(pt, struct i915_request, priotree); + return container_of(node, struct i915_request, sched); } static struct intel_engine_cs * -pt_lock_engine(struct i915_priotree *pt, struct intel_engine_cs *locked) +sched_lock_engine(struct i915_sched_node *node, struct intel_engine_cs *locked) { - struct intel_engine_cs *engine = pt_to_request(pt)->engine; + struct intel_engine_cs *engine = sched_to_request(node)->engine; GEM_BUG_ON(!locked); if (engine != locked) { - spin_unlock(&locked->timeline->lock); - spin_lock(&engine->timeline->lock); + spin_unlock(&locked->timeline.lock); + spin_lock(&engine->timeline.lock); } return engine; } -static void execlists_schedule(struct i915_request *request, int prio) +static void execlists_schedule(struct i915_request *request, + const struct i915_sched_attr *attr) { - struct intel_engine_cs *engine; + struct i915_priolist *uninitialized_var(pl); + struct intel_engine_cs *engine, *last; struct i915_dependency *dep, *p; struct i915_dependency stack; + const int prio = attr->priority; LIST_HEAD(dfs); GEM_BUG_ON(prio == I915_PRIORITY_INVALID); @@ -1077,23 +1227,23 @@ static void execlists_schedule(struct i915_request *request, int prio) if (i915_request_completed(request)) return; - if (prio <= READ_ONCE(request->priotree.priority)) + if (prio <= READ_ONCE(request->sched.attr.priority)) return; /* Need BKL in order to use the temporary link inside i915_dependency */ lockdep_assert_held(&request->i915->drm.struct_mutex); - stack.signaler = &request->priotree; + stack.signaler = &request->sched; list_add(&stack.dfs_link, &dfs); /* * Recursively bump all dependent priorities to match the new request. * * A naive approach would be to use recursion: - * static void update_priorities(struct i915_priotree *pt, prio) { - * list_for_each_entry(dep, &pt->signalers_list, signal_link) + * static void update_priorities(struct i915_sched_node *node, prio) { + * list_for_each_entry(dep, &node->signalers_list, signal_link) * update_priorities(dep->signal, prio) - * queue_request(pt); + * queue_request(node); * } * but that may have unlimited recursion depth and so runs a very * real risk of overunning the kernel stack. Instead, we build @@ -1105,7 +1255,7 @@ static void execlists_schedule(struct i915_request *request, int prio) * last element in the list is the request we must execute first. */ list_for_each_entry(dep, &dfs, dfs_link) { - struct i915_priotree *pt = dep->signaler; + struct i915_sched_node *node = dep->signaler; /* * Within an engine, there can be no cycle, but we may @@ -1113,14 +1263,14 @@ static void execlists_schedule(struct i915_request *request, int prio) * (redundant dependencies are not eliminated) and across * engines. */ - list_for_each_entry(p, &pt->signalers_list, signal_link) { + list_for_each_entry(p, &node->signalers_list, signal_link) { GEM_BUG_ON(p == dep); /* no cycles! */ - if (i915_priotree_signaled(p->signaler)) + if (i915_sched_node_signaled(p->signaler)) continue; - GEM_BUG_ON(p->signaler->priority < pt->priority); - if (prio > READ_ONCE(p->signaler->priority)) + GEM_BUG_ON(p->signaler->attr.priority < node->attr.priority); + if (prio > READ_ONCE(p->signaler->attr.priority)) list_move_tail(&p->dfs_link, &dfs); } } @@ -1131,37 +1281,45 @@ static void execlists_schedule(struct i915_request *request, int prio) * execlists_submit_request()), we can set our own priority and skip * acquiring the engine locks. */ - if (request->priotree.priority == I915_PRIORITY_INVALID) { - GEM_BUG_ON(!list_empty(&request->priotree.link)); - request->priotree.priority = prio; + if (request->sched.attr.priority == I915_PRIORITY_INVALID) { + GEM_BUG_ON(!list_empty(&request->sched.link)); + request->sched.attr = *attr; if (stack.dfs_link.next == stack.dfs_link.prev) return; __list_del_entry(&stack.dfs_link); } + last = NULL; engine = request->engine; - spin_lock_irq(&engine->timeline->lock); + spin_lock_irq(&engine->timeline.lock); /* Fifo and depth-first replacement ensure our deps execute before us */ list_for_each_entry_safe_reverse(dep, p, &dfs, dfs_link) { - struct i915_priotree *pt = dep->signaler; + struct i915_sched_node *node = dep->signaler; INIT_LIST_HEAD(&dep->dfs_link); - engine = pt_lock_engine(pt, engine); + engine = sched_lock_engine(node, engine); - if (prio <= pt->priority) + if (prio <= node->attr.priority) continue; - pt->priority = prio; - if (!list_empty(&pt->link)) { - __list_del_entry(&pt->link); - queue_request(engine, pt, prio); + node->attr.priority = prio; + if (!list_empty(&node->link)) { + if (last != engine) { + pl = lookup_priolist(engine, prio); + last = engine; + } + GEM_BUG_ON(pl->priority != prio); + list_move_tail(&node->link, &pl->requests); } - submit_queue(engine, prio); + + if (prio > engine->execlists.queue_priority && + i915_sw_fence_done(&sched_to_request(node)->submit)) + __submit_queue(engine, prio); } - spin_unlock_irq(&engine->timeline->lock); + spin_unlock_irq(&engine->timeline.lock); } static int __context_pin(struct i915_gem_context *ctx, struct i915_vma *vma) @@ -1191,7 +1349,7 @@ static struct intel_ring * execlists_context_pin(struct intel_engine_cs *engine, struct i915_gem_context *ctx) { - struct intel_context *ce = &ctx->engine[engine->id]; + struct intel_context *ce = to_intel_context(ctx, engine); void *vaddr; int ret; @@ -1225,6 +1383,7 @@ execlists_context_pin(struct intel_engine_cs *engine, ce->lrc_reg_state = vaddr + LRC_STATE_PN * PAGE_SIZE; ce->lrc_reg_state[CTX_RING_BUFFER_START+1] = i915_ggtt_offset(ce->ring->vma); + ce->lrc_reg_state[CTX_RING_HEAD+1] = ce->ring->head; ce->state->obj->pin_global++; i915_gem_context_get(ctx); @@ -1243,7 +1402,7 @@ err: static void execlists_context_unpin(struct intel_engine_cs *engine, struct i915_gem_context *ctx) { - struct intel_context *ce = &ctx->engine[engine->id]; + struct intel_context *ce = to_intel_context(ctx, engine); lockdep_assert_held(&ctx->i915->drm.struct_mutex); GEM_BUG_ON(ce->pin_count == 0); @@ -1262,8 +1421,8 @@ static void execlists_context_unpin(struct intel_engine_cs *engine, static int execlists_request_alloc(struct i915_request *request) { - struct intel_engine_cs *engine = request->engine; - struct intel_context *ce = &request->ctx->engine[engine->id]; + struct intel_context *ce = + to_intel_context(request->ctx, request->engine); int ret; GEM_BUG_ON(!ce->pin_count); @@ -1523,6 +1682,8 @@ static int intel_init_workaround_bb(struct intel_engine_cs *engine) return -EINVAL; switch (INTEL_GEN(engine->i915)) { + case 11: + return 0; case 10: wa_bb_fn[0] = gen10_init_indirectctx_bb; wa_bb_fn[1] = NULL; @@ -1575,14 +1736,6 @@ static int intel_init_workaround_bb(struct intel_engine_cs *engine) return ret; } -static u8 gtiir[] = { - [RCS] = 0, - [BCS] = 0, - [VCS] = 1, - [VCS2] = 1, - [VECS] = 3, -}; - static void enable_execlists(struct intel_engine_cs *engine) { struct drm_i915_private *dev_priv = engine->i915; @@ -1642,6 +1795,8 @@ static int gen8_init_render_ring(struct intel_engine_cs *engine) if (ret) return ret; + intel_whitelist_workarounds_apply(engine); + /* We need to disable the AsyncFlip performance optimisations in order * to use MI_WAIT_FOR_EVENT within the CS. It should already be * programmed to '1' on all products. @@ -1652,7 +1807,7 @@ static int gen8_init_render_ring(struct intel_engine_cs *engine) I915_WRITE(INSTPM, _MASKED_BIT_ENABLE(INSTPM_FORCE_ORDERING)); - return init_workarounds_ring(engine); + return 0; } static int gen9_init_render_ring(struct intel_engine_cs *engine) @@ -1663,49 +1818,25 @@ static int gen9_init_render_ring(struct intel_engine_cs *engine) if (ret) return ret; - return init_workarounds_ring(engine); -} - -static void reset_irq(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int i; - - GEM_BUG_ON(engine->id >= ARRAY_SIZE(gtiir)); - - /* - * Clear any pending interrupt state. - * - * We do it twice out of paranoia that some of the IIR are double - * buffered, and if we only reset it once there may still be - * an interrupt pending. - */ - for (i = 0; i < 2; i++) { - I915_WRITE(GEN8_GT_IIR(gtiir[engine->id]), - GT_CONTEXT_SWITCH_INTERRUPT << engine->irq_shift); - POSTING_READ(GEN8_GT_IIR(gtiir[engine->id])); - } - GEM_BUG_ON(I915_READ(GEN8_GT_IIR(gtiir[engine->id])) & - (GT_CONTEXT_SWITCH_INTERRUPT << engine->irq_shift)); + intel_whitelist_workarounds_apply(engine); - clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); + return 0; } static void reset_common_ring(struct intel_engine_cs *engine, struct i915_request *request) { struct intel_engine_execlists * const execlists = &engine->execlists; - struct intel_context *ce; unsigned long flags; + u32 *regs; - GEM_TRACE("%s seqno=%x\n", - engine->name, request ? request->global_seqno : 0); + GEM_TRACE("%s request global=%x, current=%d\n", + engine->name, request ? request->global_seqno : 0, + intel_engine_get_seqno(engine)); /* See execlists_cancel_requests() for the irq/spinlock split. */ local_irq_save(flags); - reset_irq(engine); - /* * Catch up with any missed context-switch interrupts. * @@ -1716,14 +1847,12 @@ static void reset_common_ring(struct intel_engine_cs *engine, * requests were completed. */ execlists_cancel_port_requests(execlists); + reset_irq(engine); /* Push back any incomplete requests for replay after the reset. */ - spin_lock(&engine->timeline->lock); + spin_lock(&engine->timeline.lock); __unwind_incomplete_requests(engine); - spin_unlock(&engine->timeline->lock); - - /* Mark all CS interrupts as complete */ - execlists->active = 0; + spin_unlock(&engine->timeline.lock); local_irq_restore(flags); @@ -1749,14 +1878,24 @@ static void reset_common_ring(struct intel_engine_cs *engine, * future request will be after userspace has had the opportunity * to recreate its own state. */ - ce = &request->ctx->engine[engine->id]; - execlists_init_reg_state(ce->lrc_reg_state, - request->ctx, engine, ce->ring); + regs = to_intel_context(request->ctx, engine)->lrc_reg_state; + if (engine->default_state) { + void *defaults; + + defaults = i915_gem_object_pin_map(engine->default_state, + I915_MAP_WB); + if (!IS_ERR(defaults)) { + memcpy(regs, /* skip restoring the vanilla PPHWSP */ + defaults + LRC_STATE_PN * PAGE_SIZE, + engine->context_size - PAGE_SIZE); + i915_gem_object_unpin_map(engine->default_state); + } + } + execlists_init_reg_state(regs, request->ctx, engine, request->ring); /* Move the RING_HEAD onto the breadcrumb, past the hanging batch */ - ce->lrc_reg_state[CTX_RING_BUFFER_START+1] = - i915_ggtt_offset(ce->ring->vma); - ce->lrc_reg_state[CTX_RING_HEAD+1] = request->postfix; + regs[CTX_RING_BUFFER_START + 1] = i915_ggtt_offset(request->ring->vma); + regs[CTX_RING_HEAD + 1] = request->postfix; request->ring->head = request->postfix; intel_ring_update_space(request->ring); @@ -1817,7 +1956,7 @@ static int gen8_emit_bb_start(struct i915_request *rq, rq->ctx->ppgtt->pd_dirty_rings &= ~intel_engine_flag(rq->engine); } - cs = intel_ring_begin(rq, 4); + cs = intel_ring_begin(rq, 6); if (IS_ERR(cs)) return PTR_ERR(cs); @@ -1846,6 +1985,9 @@ static int gen8_emit_bb_start(struct i915_request *rq, (flags & I915_DISPATCH_RS ? MI_BATCH_RESOURCE_STREAMER : 0); *cs++ = lower_32_bits(offset); *cs++ = upper_32_bits(offset); + + *cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE; + *cs++ = MI_NOOP; intel_ring_advance(rq, cs); return 0; @@ -1988,7 +2130,7 @@ static void gen8_emit_breadcrumb(struct i915_request *request, u32 *cs) cs = gen8_emit_ggtt_write(cs, request->global_seqno, intel_hws_seqno_address(request->engine)); *cs++ = MI_USER_INTERRUPT; - *cs++ = MI_NOOP; + *cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE; request->tail = intel_ring_offset(request, cs); assert_ring_tail_valid(request->ring, request->tail); @@ -2004,7 +2146,7 @@ static void gen8_emit_breadcrumb_rcs(struct i915_request *request, u32 *cs) cs = gen8_emit_ggtt_write_rcs(cs, request->global_seqno, intel_hws_seqno_address(request->engine)); *cs++ = MI_USER_INTERRUPT; - *cs++ = MI_NOOP; + *cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE; request->tail = intel_ring_offset(request, cs); assert_ring_tail_valid(request->ring, request->tail); @@ -2016,7 +2158,7 @@ static int gen8_init_rcs_context(struct i915_request *rq) { int ret; - ret = intel_ring_workarounds_emit(rq); + ret = intel_ctx_workarounds_emit(rq); if (ret) return ret; @@ -2076,11 +2218,13 @@ static void execlists_set_default_submission(struct intel_engine_cs *engine) engine->unpark = NULL; engine->flags |= I915_ENGINE_SUPPORTS_STATS; + if (engine->i915->preempt_context) + engine->flags |= I915_ENGINE_HAS_PREEMPTION; engine->i915->caps.scheduler = I915_SCHEDULER_CAP_ENABLED | I915_SCHEDULER_CAP_PRIORITY; - if (engine->i915->preempt_context) + if (intel_engine_has_preemption(engine)) engine->i915->caps.scheduler |= I915_SCHEDULER_CAP_PREEMPTION; } @@ -2119,7 +2263,20 @@ logical_ring_default_vfuncs(struct intel_engine_cs *engine) static inline void logical_ring_default_irqs(struct intel_engine_cs *engine) { - unsigned shift = engine->irq_shift; + unsigned int shift = 0; + + if (INTEL_GEN(engine->i915) < 11) { + const u8 irq_shifts[] = { + [RCS] = GEN8_RCS_IRQ_SHIFT, + [BCS] = GEN8_BCS_IRQ_SHIFT, + [VCS] = GEN8_VCS1_IRQ_SHIFT, + [VCS2] = GEN8_VCS2_IRQ_SHIFT, + [VECS] = GEN8_VECS_IRQ_SHIFT, + }; + + shift = irq_shifts[engine->id]; + } + engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT << shift; engine->irq_keep_mask = GT_CONTEXT_SWITCH_INTERRUPT << shift; } @@ -2175,9 +2332,13 @@ static int logical_ring_init(struct intel_engine_cs *engine) } engine->execlists.preempt_complete_status = ~0u; - if (engine->i915->preempt_context) + if (engine->i915->preempt_context) { + struct intel_context *ce = + to_intel_context(engine->i915->preempt_context, engine); + engine->execlists.preempt_complete_status = - upper_32_bits(engine->i915->preempt_context->engine[engine->id].lrc_desc); + upper_32_bits(ce->lrc_desc); + } return 0; @@ -2431,8 +2592,10 @@ populate_lr_context(struct i915_gem_context *ctx, defaults = i915_gem_object_pin_map(engine->default_state, I915_MAP_WB); - if (IS_ERR(defaults)) - return PTR_ERR(defaults); + if (IS_ERR(defaults)) { + ret = PTR_ERR(defaults); + goto err_unpin_ctx; + } memcpy(vaddr + start, defaults + start, engine->context_size); i915_gem_object_unpin_map(engine->default_state); @@ -2450,19 +2613,20 @@ populate_lr_context(struct i915_gem_context *ctx, _MASKED_BIT_ENABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT | CTX_CTRL_ENGINE_CTX_SAVE_INHIBIT); +err_unpin_ctx: i915_gem_object_unpin_map(ctx_obj); - - return 0; + return ret; } static int execlists_context_deferred_alloc(struct i915_gem_context *ctx, struct intel_engine_cs *engine) { struct drm_i915_gem_object *ctx_obj; - struct intel_context *ce = &ctx->engine[engine->id]; + struct intel_context *ce = to_intel_context(ctx, engine); struct i915_vma *vma; uint32_t context_size; struct intel_ring *ring; + struct i915_timeline *timeline; int ret; if (ce->state) @@ -2478,8 +2642,8 @@ static int execlists_context_deferred_alloc(struct i915_gem_context *ctx, ctx_obj = i915_gem_object_create(ctx->i915, context_size); if (IS_ERR(ctx_obj)) { - DRM_DEBUG_DRIVER("Alloc LRC backing obj failed.\n"); - return PTR_ERR(ctx_obj); + ret = PTR_ERR(ctx_obj); + goto error_deref_obj; } vma = i915_vma_instance(ctx_obj, &ctx->i915->ggtt.base, NULL); @@ -2488,7 +2652,14 @@ static int execlists_context_deferred_alloc(struct i915_gem_context *ctx, goto error_deref_obj; } - ring = intel_engine_create_ring(engine, ctx->ring_size); + timeline = i915_timeline_create(ctx->i915, ctx->name); + if (IS_ERR(timeline)) { + ret = PTR_ERR(timeline); + goto error_deref_obj; + } + + ring = intel_engine_create_ring(engine, timeline, ctx->ring_size); + i915_timeline_put(timeline); if (IS_ERR(ring)) { ret = PTR_ERR(ring); goto error_deref_obj; @@ -2530,7 +2701,8 @@ void intel_lr_context_resume(struct drm_i915_private *dev_priv) */ list_for_each_entry(ctx, &dev_priv->contexts.list, link) { for_each_engine(engine, dev_priv, id) { - struct intel_context *ce = &ctx->engine[engine->id]; + struct intel_context *ce = + to_intel_context(ctx, engine); u32 *reg; if (!ce->state) @@ -2552,3 +2724,7 @@ void intel_lr_context_resume(struct drm_i915_private *dev_priv) } } } + +#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) +#include "selftests/intel_lrc.c" +#endif diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h index 59d7b86..4ec7d8d 100644 --- a/drivers/gpu/drm/i915/intel_lrc.h +++ b/drivers/gpu/drm/i915/intel_lrc.h @@ -108,7 +108,7 @@ static inline uint64_t intel_lr_context_descriptor(struct i915_gem_context *ctx, struct intel_engine_cs *engine) { - return ctx->engine[engine->id].lrc_desc; + return to_intel_context(ctx, engine)->lrc_desc; } #endif /* _INTEL_LRC_H_ */ diff --git a/drivers/gpu/drm/i915/intel_mocs.c b/drivers/gpu/drm/i915/intel_mocs.c index c0b34b7..9f0bd6a4 100644 --- a/drivers/gpu/drm/i915/intel_mocs.c +++ b/drivers/gpu/drm/i915/intel_mocs.c @@ -178,7 +178,8 @@ static bool get_mocs_settings(struct drm_i915_private *dev_priv, { bool result = false; - if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { + if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv) || + IS_ICELAKE(dev_priv)) { table->size = ARRAY_SIZE(skylake_mocs_table); table->table = skylake_mocs_table; result = true; @@ -217,6 +218,8 @@ static i915_reg_t mocs_register(enum intel_engine_id engine_id, int index) return GEN9_VEBOX_MOCS(index); case VCS2: return GEN9_MFX1_MOCS(index); + case VCS3: + return GEN11_MFX2_MOCS(index); default: MISSING_CASE(engine_id); return INVALID_MMIO_REG; diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c index 36671a9..c2f10d8 100644 --- a/drivers/gpu/drm/i915/intel_overlay.c +++ b/drivers/gpu/drm/i915/intel_overlay.c @@ -807,6 +807,7 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay, ret = PTR_ERR(vma); goto out_pin_section; } + intel_fb_obj_flush(new_bo, ORIGIN_DIRTYFB); ret = i915_vma_put_fence(vma); if (ret) diff --git a/drivers/gpu/drm/i915/intel_pipe_crc.c b/drivers/gpu/drm/i915/intel_pipe_crc.c index 1f5cd57..39a4e4e 100644 --- a/drivers/gpu/drm/i915/intel_pipe_crc.c +++ b/drivers/gpu/drm/i915/intel_pipe_crc.c @@ -569,7 +569,8 @@ unlock: static int ivb_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, enum pipe pipe, enum intel_pipe_crc_source *source, - uint32_t *val) + uint32_t *val, + bool set_wa) { if (*source == INTEL_PIPE_CRC_SOURCE_AUTO) *source = INTEL_PIPE_CRC_SOURCE_PF; @@ -582,7 +583,7 @@ static int ivb_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, *val = PIPE_CRC_ENABLE | PIPE_CRC_SOURCE_SPRITE_IVB; break; case INTEL_PIPE_CRC_SOURCE_PF: - if ((IS_HASWELL(dev_priv) || + if (set_wa && (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) && pipe == PIPE_A) hsw_pipe_A_crc_wa(dev_priv, true); @@ -600,7 +601,8 @@ static int ivb_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, static int get_new_crc_ctl_reg(struct drm_i915_private *dev_priv, enum pipe pipe, - enum intel_pipe_crc_source *source, u32 *val) + enum intel_pipe_crc_source *source, u32 *val, + bool set_wa) { if (IS_GEN2(dev_priv)) return i8xx_pipe_crc_ctl_reg(source, val); @@ -611,7 +613,7 @@ static int get_new_crc_ctl_reg(struct drm_i915_private *dev_priv, else if (IS_GEN5(dev_priv) || IS_GEN6(dev_priv)) return ilk_pipe_crc_ctl_reg(source, val); else - return ivb_pipe_crc_ctl_reg(dev_priv, pipe, source, val); + return ivb_pipe_crc_ctl_reg(dev_priv, pipe, source, val, set_wa); } static int pipe_crc_set_source(struct drm_i915_private *dev_priv, @@ -636,7 +638,7 @@ static int pipe_crc_set_source(struct drm_i915_private *dev_priv, return -EIO; } - ret = get_new_crc_ctl_reg(dev_priv, pipe, &source, &val); + ret = get_new_crc_ctl_reg(dev_priv, pipe, &source, &val, true); if (ret != 0) goto out; @@ -764,13 +766,12 @@ display_crc_ctl_parse_object(const char *buf, enum intel_pipe_crc_object *o) { int i; - for (i = 0; i < ARRAY_SIZE(pipe_crc_objects); i++) - if (!strcmp(buf, pipe_crc_objects[i])) { - *o = i; - return 0; - } + i = match_string(pipe_crc_objects, ARRAY_SIZE(pipe_crc_objects), buf); + if (i < 0) + return i; - return -EINVAL; + *o = i; + return 0; } static int display_crc_ctl_parse_pipe(struct drm_i915_private *dev_priv, @@ -796,13 +797,12 @@ display_crc_ctl_parse_source(const char *buf, enum intel_pipe_crc_source *s) return 0; } - for (i = 0; i < ARRAY_SIZE(pipe_crc_sources); i++) - if (!strcmp(buf, pipe_crc_sources[i])) { - *s = i; - return 0; - } + i = match_string(pipe_crc_sources, ARRAY_SIZE(pipe_crc_sources), buf); + if (i < 0) + return i; - return -EINVAL; + *s = i; + return 0; } static int display_crc_ctl_parse(struct drm_i915_private *dev_priv, @@ -916,7 +916,7 @@ int intel_pipe_crc_create(struct drm_minor *minor) int intel_crtc_set_crc_source(struct drm_crtc *crtc, const char *source_name, size_t *values_cnt) { - struct drm_i915_private *dev_priv = crtc->dev->dev_private; + struct drm_i915_private *dev_priv = to_i915(crtc->dev); struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[crtc->index]; enum intel_display_power_domain power_domain; enum intel_pipe_crc_source source; @@ -934,10 +934,11 @@ int intel_crtc_set_crc_source(struct drm_crtc *crtc, const char *source_name, return -EIO; } - ret = get_new_crc_ctl_reg(dev_priv, crtc->index, &source, &val); + ret = get_new_crc_ctl_reg(dev_priv, crtc->index, &source, &val, true); if (ret != 0) goto out; + pipe_crc->source = source; I915_WRITE(PIPE_CRC_CTL(crtc->index), val); POSTING_READ(PIPE_CRC_CTL(crtc->index)); @@ -959,3 +960,39 @@ out: return ret; } + +void intel_crtc_enable_pipe_crc(struct intel_crtc *intel_crtc) +{ + struct drm_crtc *crtc = &intel_crtc->base; + struct drm_i915_private *dev_priv = to_i915(crtc->dev); + struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[crtc->index]; + u32 val = 0; + + if (!crtc->crc.opened) + return; + + if (get_new_crc_ctl_reg(dev_priv, crtc->index, &pipe_crc->source, &val, false) < 0) + return; + + /* Don't need pipe_crc->lock here, IRQs are not generated. */ + pipe_crc->skipped = 0; + + I915_WRITE(PIPE_CRC_CTL(crtc->index), val); + POSTING_READ(PIPE_CRC_CTL(crtc->index)); +} + +void intel_crtc_disable_pipe_crc(struct intel_crtc *intel_crtc) +{ + struct drm_crtc *crtc = &intel_crtc->base; + struct drm_i915_private *dev_priv = to_i915(crtc->dev); + struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[crtc->index]; + + /* Swallow crc's until we stop generating them. */ + spin_lock_irq(&pipe_crc->lock); + pipe_crc->skipped = INT_MIN; + spin_unlock_irq(&pipe_crc->lock); + + I915_WRITE(PIPE_CRC_CTL(crtc->index), 0); + POSTING_READ(PIPE_CRC_CTL(crtc->index)); + synchronize_irq(dev_priv->drm.irq); +} diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index b8da4dc..b85229e 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -3567,6 +3567,23 @@ bool ilk_disable_lp_wm(struct drm_device *dev) return _ilk_disable_lp_wm(dev_priv, WM_DIRTY_LP_ALL); } +static u8 intel_enabled_dbuf_slices_num(struct drm_i915_private *dev_priv) +{ + u8 enabled_slices; + + /* Slice 1 will always be enabled */ + enabled_slices = 1; + + /* Gen prior to GEN11 have only one DBuf slice */ + if (INTEL_GEN(dev_priv) < 11) + return enabled_slices; + + if (I915_READ(DBUF_CTL_S2) & DBUF_POWER_STATE) + enabled_slices++; + + return enabled_slices; +} + /* * FIXME: We still don't have the proper code detect if we need to apply the WA, * so assume we'll always need it in order to avoid underruns. @@ -3754,9 +3771,42 @@ bool intel_can_enable_sagv(struct drm_atomic_state *state) return true; } +static unsigned int intel_get_ddb_size(struct drm_i915_private *dev_priv, + const struct intel_crtc_state *cstate, + const unsigned int total_data_rate, + const int num_active, + struct skl_ddb_allocation *ddb) +{ + const struct drm_display_mode *adjusted_mode; + u64 total_data_bw; + u16 ddb_size = INTEL_INFO(dev_priv)->ddb_size; + + WARN_ON(ddb_size == 0); + + if (INTEL_GEN(dev_priv) < 11) + return ddb_size - 4; /* 4 blocks for bypass path allocation */ + + adjusted_mode = &cstate->base.adjusted_mode; + total_data_bw = (u64)total_data_rate * drm_mode_vrefresh(adjusted_mode); + + /* + * 12GB/s is maximum BW supported by single DBuf slice. + */ + if (total_data_bw >= GBps(12) || num_active > 1) { + ddb->enabled_slices = 2; + } else { + ddb->enabled_slices = 1; + ddb_size /= 2; + } + + return ddb_size; +} + static void skl_ddb_get_pipe_allocation_limits(struct drm_device *dev, const struct intel_crtc_state *cstate, + const unsigned int total_data_rate, + struct skl_ddb_allocation *ddb, struct skl_ddb_entry *alloc, /* out */ int *num_active /* out */) { @@ -3779,11 +3829,8 @@ skl_ddb_get_pipe_allocation_limits(struct drm_device *dev, else *num_active = hweight32(dev_priv->active_crtcs); - ddb_size = INTEL_INFO(dev_priv)->ddb_size; - WARN_ON(ddb_size == 0); - - if (INTEL_GEN(dev_priv) < 11) - ddb_size -= 4; /* 4 blocks for bypass path allocation */ + ddb_size = intel_get_ddb_size(dev_priv, cstate, total_data_rate, + *num_active, ddb); /* * If the state doesn't change the active CRTC's, then there's @@ -3817,14 +3864,64 @@ static unsigned int skl_cursor_allocation(int num_active) return 8; } -static void skl_ddb_entry_init_from_hw(struct skl_ddb_entry *entry, u32 reg) +static void skl_ddb_entry_init_from_hw(struct drm_i915_private *dev_priv, + struct skl_ddb_entry *entry, u32 reg) { - entry->start = reg & 0x3ff; - entry->end = (reg >> 16) & 0x3ff; + u16 mask; + + if (INTEL_GEN(dev_priv) >= 11) + mask = ICL_DDB_ENTRY_MASK; + else + mask = SKL_DDB_ENTRY_MASK; + entry->start = reg & mask; + entry->end = (reg >> DDB_ENTRY_END_SHIFT) & mask; + if (entry->end) entry->end += 1; } +static void +skl_ddb_get_hw_plane_state(struct drm_i915_private *dev_priv, + const enum pipe pipe, + const enum plane_id plane_id, + struct skl_ddb_allocation *ddb /* out */) +{ + u32 val, val2 = 0; + int fourcc, pixel_format; + + /* Cursor doesn't support NV12/planar, so no extra calculation needed */ + if (plane_id == PLANE_CURSOR) { + val = I915_READ(CUR_BUF_CFG(pipe)); + skl_ddb_entry_init_from_hw(dev_priv, + &ddb->plane[pipe][plane_id], val); + return; + } + + val = I915_READ(PLANE_CTL(pipe, plane_id)); + + /* No DDB allocated for disabled planes */ + if (!(val & PLANE_CTL_ENABLE)) + return; + + pixel_format = val & PLANE_CTL_FORMAT_MASK; + fourcc = skl_format_to_fourcc(pixel_format, + val & PLANE_CTL_ORDER_RGBX, + val & PLANE_CTL_ALPHA_MASK); + + val = I915_READ(PLANE_BUF_CFG(pipe, plane_id)); + val2 = I915_READ(PLANE_NV12_BUF_CFG(pipe, plane_id)); + + if (fourcc == DRM_FORMAT_NV12) { + skl_ddb_entry_init_from_hw(dev_priv, + &ddb->plane[pipe][plane_id], val2); + skl_ddb_entry_init_from_hw(dev_priv, + &ddb->uv_plane[pipe][plane_id], val); + } else { + skl_ddb_entry_init_from_hw(dev_priv, + &ddb->plane[pipe][plane_id], val); + } +} + void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv, struct skl_ddb_allocation *ddb /* out */) { @@ -3832,6 +3929,8 @@ void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv, memset(ddb, 0, sizeof(*ddb)); + ddb->enabled_slices = intel_enabled_dbuf_slices_num(dev_priv); + for_each_intel_crtc(&dev_priv->drm, crtc) { enum intel_display_power_domain power_domain; enum plane_id plane_id; @@ -3841,16 +3940,9 @@ void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv, if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) continue; - for_each_plane_id_on_crtc(crtc, plane_id) { - u32 val; - - if (plane_id != PLANE_CURSOR) - val = I915_READ(PLANE_BUF_CFG(pipe, plane_id)); - else - val = I915_READ(CUR_BUF_CFG(pipe)); - - skl_ddb_entry_init_from_hw(&ddb->plane[pipe][plane_id], val); - } + for_each_plane_id_on_crtc(crtc, plane_id) + skl_ddb_get_hw_plane_state(dev_priv, pipe, + plane_id, ddb); intel_display_power_put(dev_priv, power_domain); } @@ -4009,9 +4101,9 @@ int skl_check_pipe_max_pixel_rate(struct intel_crtc *intel_crtc, static unsigned int skl_plane_relative_data_rate(const struct intel_crtc_state *cstate, const struct drm_plane_state *pstate, - int y) + const int plane) { - struct intel_plane *plane = to_intel_plane(pstate->plane); + struct intel_plane *intel_plane = to_intel_plane(pstate->plane); struct intel_plane_state *intel_pstate = to_intel_plane_state(pstate); uint32_t data_rate; uint32_t width = 0, height = 0; @@ -4025,9 +4117,9 @@ skl_plane_relative_data_rate(const struct intel_crtc_state *cstate, fb = pstate->fb; format = fb->format->format; - if (plane->id == PLANE_CURSOR) + if (intel_plane->id == PLANE_CURSOR) return 0; - if (y && format != DRM_FORMAT_NV12) + if (plane == 1 && format != DRM_FORMAT_NV12) return 0; /* @@ -4038,19 +4130,14 @@ skl_plane_relative_data_rate(const struct intel_crtc_state *cstate, width = drm_rect_width(&intel_pstate->base.src) >> 16; height = drm_rect_height(&intel_pstate->base.src) >> 16; - /* for planar format */ - if (format == DRM_FORMAT_NV12) { - if (y) /* y-plane data rate */ - data_rate = width * height * - fb->format->cpp[0]; - else /* uv-plane data rate */ - data_rate = (width / 2) * (height / 2) * - fb->format->cpp[1]; - } else { - /* for packed formats */ - data_rate = width * height * fb->format->cpp[0]; + /* UV plane does 1/2 pixel sub-sampling */ + if (plane == 1 && format == DRM_FORMAT_NV12) { + width /= 2; + height /= 2; } + data_rate = width * height * fb->format->cpp[plane]; + down_scale_amount = skl_plane_downscale_amount(cstate, intel_pstate); return mul_round_up_u32_fixed16(data_rate, down_scale_amount); @@ -4063,8 +4150,8 @@ skl_plane_relative_data_rate(const struct intel_crtc_state *cstate, */ static unsigned int skl_get_total_relative_data_rate(struct intel_crtc_state *intel_cstate, - unsigned *plane_data_rate, - unsigned *plane_y_data_rate) + unsigned int *plane_data_rate, + unsigned int *uv_plane_data_rate) { struct drm_crtc_state *cstate = &intel_cstate->base; struct drm_atomic_state *state = cstate->state; @@ -4080,17 +4167,17 @@ skl_get_total_relative_data_rate(struct intel_crtc_state *intel_cstate, enum plane_id plane_id = to_intel_plane(plane)->id; unsigned int rate; - /* packed/uv */ + /* packed/y */ rate = skl_plane_relative_data_rate(intel_cstate, pstate, 0); plane_data_rate[plane_id] = rate; total_data_rate += rate; - /* y-plane */ + /* uv-plane */ rate = skl_plane_relative_data_rate(intel_cstate, pstate, 1); - plane_y_data_rate[plane_id] = rate; + uv_plane_data_rate[plane_id] = rate; total_data_rate += rate; } @@ -4099,8 +4186,7 @@ skl_get_total_relative_data_rate(struct intel_crtc_state *intel_cstate, } static uint16_t -skl_ddb_min_alloc(const struct drm_plane_state *pstate, - const int y) +skl_ddb_min_alloc(const struct drm_plane_state *pstate, const int plane) { struct drm_framebuffer *fb = pstate->fb; struct intel_plane_state *intel_pstate = to_intel_plane_state(pstate); @@ -4111,8 +4197,8 @@ skl_ddb_min_alloc(const struct drm_plane_state *pstate, if (WARN_ON(!fb)) return 0; - /* For packed formats, no y-plane, return 0 */ - if (y && fb->format->format != DRM_FORMAT_NV12) + /* For packed formats, and uv-plane, return 0 */ + if (plane == 1 && fb->format->format != DRM_FORMAT_NV12) return 0; /* For Non Y-tile return 8-blocks */ @@ -4131,15 +4217,12 @@ skl_ddb_min_alloc(const struct drm_plane_state *pstate, src_h = drm_rect_height(&intel_pstate->base.src) >> 16; /* Halve UV plane width and height for NV12 */ - if (fb->format->format == DRM_FORMAT_NV12 && !y) { + if (plane == 1) { src_w /= 2; src_h /= 2; } - if (fb->format->format == DRM_FORMAT_NV12 && !y) - plane_bpp = fb->format->cpp[1]; - else - plane_bpp = fb->format->cpp[0]; + plane_bpp = fb->format->cpp[plane]; if (drm_rotation_90_or_270(pstate->rotation)) { switch (plane_bpp) { @@ -4167,7 +4250,7 @@ skl_ddb_min_alloc(const struct drm_plane_state *pstate, static void skl_ddb_calc_min(const struct intel_crtc_state *cstate, int num_active, - uint16_t *minimum, uint16_t *y_minimum) + uint16_t *minimum, uint16_t *uv_minimum) { const struct drm_plane_state *pstate; struct drm_plane *plane; @@ -4182,7 +4265,7 @@ skl_ddb_calc_min(const struct intel_crtc_state *cstate, int num_active, continue; minimum[plane_id] = skl_ddb_min_alloc(pstate, 0); - y_minimum[plane_id] = skl_ddb_min_alloc(pstate, 1); + uv_minimum[plane_id] = skl_ddb_min_alloc(pstate, 1); } minimum[PLANE_CURSOR] = skl_cursor_allocation(num_active); @@ -4200,17 +4283,17 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, struct skl_ddb_entry *alloc = &cstate->wm.skl.ddb; uint16_t alloc_size, start; uint16_t minimum[I915_MAX_PLANES] = {}; - uint16_t y_minimum[I915_MAX_PLANES] = {}; + uint16_t uv_minimum[I915_MAX_PLANES] = {}; unsigned int total_data_rate; enum plane_id plane_id; int num_active; - unsigned plane_data_rate[I915_MAX_PLANES] = {}; - unsigned plane_y_data_rate[I915_MAX_PLANES] = {}; + unsigned int plane_data_rate[I915_MAX_PLANES] = {}; + unsigned int uv_plane_data_rate[I915_MAX_PLANES] = {}; uint16_t total_min_blocks = 0; /* Clear the partitioning for disabled planes. */ memset(ddb->plane[pipe], 0, sizeof(ddb->plane[pipe])); - memset(ddb->y_plane[pipe], 0, sizeof(ddb->y_plane[pipe])); + memset(ddb->uv_plane[pipe], 0, sizeof(ddb->uv_plane[pipe])); if (WARN_ON(!state)) return 0; @@ -4220,12 +4303,16 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, return 0; } - skl_ddb_get_pipe_allocation_limits(dev, cstate, alloc, &num_active); + total_data_rate = skl_get_total_relative_data_rate(cstate, + plane_data_rate, + uv_plane_data_rate); + skl_ddb_get_pipe_allocation_limits(dev, cstate, total_data_rate, ddb, + alloc, &num_active); alloc_size = skl_ddb_entry_size(alloc); if (alloc_size == 0) return 0; - skl_ddb_calc_min(cstate, num_active, minimum, y_minimum); + skl_ddb_calc_min(cstate, num_active, minimum, uv_minimum); /* * 1. Allocate the mininum required blocks for each active plane @@ -4235,7 +4322,7 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, for_each_plane_id_on_crtc(intel_crtc, plane_id) { total_min_blocks += minimum[plane_id]; - total_min_blocks += y_minimum[plane_id]; + total_min_blocks += uv_minimum[plane_id]; } if (total_min_blocks > alloc_size) { @@ -4255,16 +4342,13 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, * * FIXME: we may not allocate every single block here. */ - total_data_rate = skl_get_total_relative_data_rate(cstate, - plane_data_rate, - plane_y_data_rate); if (total_data_rate == 0) return 0; start = alloc->start; for_each_plane_id_on_crtc(intel_crtc, plane_id) { - unsigned int data_rate, y_data_rate; - uint16_t plane_blocks, y_plane_blocks = 0; + unsigned int data_rate, uv_data_rate; + uint16_t plane_blocks, uv_plane_blocks; if (plane_id == PLANE_CURSOR) continue; @@ -4288,21 +4372,20 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, start += plane_blocks; - /* - * allocation for y_plane part of planar format: - */ - y_data_rate = plane_y_data_rate[plane_id]; + /* Allocate DDB for UV plane for planar format/NV12 */ + uv_data_rate = uv_plane_data_rate[plane_id]; - y_plane_blocks = y_minimum[plane_id]; - y_plane_blocks += div_u64((uint64_t)alloc_size * y_data_rate, - total_data_rate); + uv_plane_blocks = uv_minimum[plane_id]; + uv_plane_blocks += div_u64((uint64_t)alloc_size * uv_data_rate, + total_data_rate); - if (y_data_rate) { - ddb->y_plane[pipe][plane_id].start = start; - ddb->y_plane[pipe][plane_id].end = start + y_plane_blocks; + if (uv_data_rate) { + ddb->uv_plane[pipe][plane_id].start = start; + ddb->uv_plane[pipe][plane_id].end = + start + uv_plane_blocks; } - start += y_plane_blocks; + start += uv_plane_blocks; } return 0; @@ -4398,7 +4481,7 @@ static int skl_compute_plane_wm_params(const struct drm_i915_private *dev_priv, struct intel_crtc_state *cstate, const struct intel_plane_state *intel_pstate, - struct skl_wm_params *wp) + struct skl_wm_params *wp, int plane_id) { struct intel_plane *plane = to_intel_plane(intel_pstate->base.plane); const struct drm_plane_state *pstate = &intel_pstate->base; @@ -4411,6 +4494,12 @@ skl_compute_plane_wm_params(const struct drm_i915_private *dev_priv, if (!intel_wm_plane_visible(cstate, intel_pstate)) return 0; + /* only NV12 format has two planes */ + if (plane_id == 1 && fb->format->format != DRM_FORMAT_NV12) { + DRM_DEBUG_KMS("Non NV12 format have single plane\n"); + return -EINVAL; + } + wp->y_tiled = fb->modifier == I915_FORMAT_MOD_Y_TILED || fb->modifier == I915_FORMAT_MOD_Yf_TILED || fb->modifier == I915_FORMAT_MOD_Y_TILED_CCS || @@ -4418,6 +4507,7 @@ skl_compute_plane_wm_params(const struct drm_i915_private *dev_priv, wp->x_tiled = fb->modifier == I915_FORMAT_MOD_X_TILED; wp->rc_surface = fb->modifier == I915_FORMAT_MOD_Y_TILED_CCS || fb->modifier == I915_FORMAT_MOD_Yf_TILED_CCS; + wp->is_planar = fb->format->format == DRM_FORMAT_NV12; if (plane->id == PLANE_CURSOR) { wp->width = intel_pstate->base.crtc_w; @@ -4430,8 +4520,10 @@ skl_compute_plane_wm_params(const struct drm_i915_private *dev_priv, wp->width = drm_rect_width(&intel_pstate->base.src) >> 16; } - wp->cpp = (fb->format->format == DRM_FORMAT_NV12) ? fb->format->cpp[1] : - fb->format->cpp[0]; + if (plane_id == 1 && wp->is_planar) + wp->width /= 2; + + wp->cpp = fb->format->cpp[plane_id]; wp->plane_pixel_rate = skl_adjusted_plane_pixel_rate(cstate, intel_pstate); @@ -4499,9 +4591,8 @@ static int skl_compute_plane_wm(const struct drm_i915_private *dev_priv, uint16_t ddb_allocation, int level, const struct skl_wm_params *wp, - uint16_t *out_blocks, /* out */ - uint8_t *out_lines, /* out */ - bool *enabled /* out */) + const struct skl_wm_level *result_prev, + struct skl_wm_level *result /* out */) { const struct drm_plane_state *pstate = &intel_pstate->base; uint32_t latency = dev_priv->wm.skl_latency[level]; @@ -4515,7 +4606,7 @@ static int skl_compute_plane_wm(const struct drm_i915_private *dev_priv, if (latency == 0 || !intel_wm_plane_visible(cstate, intel_pstate)) { - *enabled = false; + result->plane_en = false; return 0; } @@ -4568,6 +4659,15 @@ static int skl_compute_plane_wm(const struct drm_i915_private *dev_priv, } else { res_blocks++; } + + /* + * Make sure result blocks for higher latency levels are atleast + * as high as level below the current level. + * Assumption in DDB algorithm optimization for special cases. + * Also covers Display WA #1125 for RC. + */ + if (result_prev->plane_res_b > res_blocks) + res_blocks = result_prev->plane_res_b; } if (INTEL_GEN(dev_priv) >= 11) { @@ -4596,7 +4696,7 @@ static int skl_compute_plane_wm(const struct drm_i915_private *dev_priv, if ((level > 0 && res_lines > 31) || res_blocks >= ddb_allocation || min_disp_buf_needed >= ddb_allocation) { - *enabled = false; + result->plane_en = false; /* * If there are no valid level 0 watermarks, then we can't @@ -4615,10 +4715,21 @@ static int skl_compute_plane_wm(const struct drm_i915_private *dev_priv, } } + /* + * Display WA #826 (SKL:ALL, BXT:ALL) & #1059 (CNL:A) + * disable wm level 1-7 on NV12 planes + */ + if (wp->is_planar && level >= 1 && + (IS_SKYLAKE(dev_priv) || IS_BROXTON(dev_priv) || + IS_CNL_REVID(dev_priv, CNL_REVID_A0, CNL_REVID_A0))) { + result->plane_en = false; + return 0; + } + /* The number of lines are ignored for the level 0 watermark. */ - *out_lines = level ? res_lines : 0; - *out_blocks = res_blocks; - *enabled = true; + result->plane_res_b = res_blocks; + result->plane_res_l = res_lines; + result->plane_en = true; return 0; } @@ -4629,7 +4740,8 @@ skl_compute_wm_levels(const struct drm_i915_private *dev_priv, struct intel_crtc_state *cstate, const struct intel_plane_state *intel_pstate, const struct skl_wm_params *wm_params, - struct skl_plane_wm *wm) + struct skl_plane_wm *wm, + int plane_id) { struct intel_crtc *intel_crtc = to_intel_crtc(cstate->base.crtc); struct drm_plane *plane = intel_pstate->base.plane; @@ -4637,15 +4749,26 @@ skl_compute_wm_levels(const struct drm_i915_private *dev_priv, uint16_t ddb_blocks; enum pipe pipe = intel_crtc->pipe; int level, max_level = ilk_wm_max_level(dev_priv); + enum plane_id intel_plane_id = intel_plane->id; int ret; if (WARN_ON(!intel_pstate->base.fb)) return -EINVAL; - ddb_blocks = skl_ddb_entry_size(&ddb->plane[pipe][intel_plane->id]); + ddb_blocks = plane_id ? + skl_ddb_entry_size(&ddb->uv_plane[pipe][intel_plane_id]) : + skl_ddb_entry_size(&ddb->plane[pipe][intel_plane_id]); for (level = 0; level <= max_level; level++) { - struct skl_wm_level *result = &wm->wm[level]; + struct skl_wm_level *result = plane_id ? &wm->uv_wm[level] : + &wm->wm[level]; + struct skl_wm_level *result_prev; + + if (level) + result_prev = plane_id ? &wm->uv_wm[level - 1] : + &wm->wm[level - 1]; + else + result_prev = plane_id ? &wm->uv_wm[0] : &wm->wm[0]; ret = skl_compute_plane_wm(dev_priv, cstate, @@ -4653,13 +4776,15 @@ skl_compute_wm_levels(const struct drm_i915_private *dev_priv, ddb_blocks, level, wm_params, - &result->plane_res_b, - &result->plane_res_l, - &result->plane_en); + result_prev, + result); if (ret) return ret; } + if (intel_pstate->base.fb->format->format == DRM_FORMAT_NV12) + wm->is_planar = true; + return 0; } @@ -4769,20 +4894,39 @@ static int skl_build_pipe_wm(struct intel_crtc_state *cstate, wm = &pipe_wm->planes[plane_id]; ddb_blocks = skl_ddb_entry_size(&ddb->plane[pipe][plane_id]); - memset(&wm_params, 0, sizeof(struct skl_wm_params)); ret = skl_compute_plane_wm_params(dev_priv, cstate, - intel_pstate, &wm_params); + intel_pstate, &wm_params, 0); if (ret) return ret; ret = skl_compute_wm_levels(dev_priv, ddb, cstate, - intel_pstate, &wm_params, wm); + intel_pstate, &wm_params, wm, 0); if (ret) return ret; + skl_compute_transition_wm(cstate, &wm_params, &wm->wm[0], ddb_blocks, &wm->trans_wm); + + /* uv plane watermarks must also be validated for NV12/Planar */ + if (wm_params.is_planar) { + memset(&wm_params, 0, sizeof(struct skl_wm_params)); + wm->is_planar = true; + + ret = skl_compute_plane_wm_params(dev_priv, cstate, + intel_pstate, + &wm_params, 1); + if (ret) + return ret; + + ret = skl_compute_wm_levels(dev_priv, ddb, cstate, + intel_pstate, &wm_params, + wm, 1); + if (ret) + return ret; + } } + pipe_wm->linetime = skl_compute_linetime_wm(cstate); return 0; @@ -4833,10 +4977,21 @@ static void skl_write_plane_wm(struct intel_crtc *intel_crtc, skl_ddb_entry_write(dev_priv, PLANE_BUF_CFG(pipe, plane_id), &ddb->plane[pipe][plane_id]); - if (INTEL_GEN(dev_priv) < 11) + if (INTEL_GEN(dev_priv) >= 11) + return skl_ddb_entry_write(dev_priv, + PLANE_BUF_CFG(pipe, plane_id), + &ddb->plane[pipe][plane_id]); + if (wm->is_planar) { + skl_ddb_entry_write(dev_priv, PLANE_BUF_CFG(pipe, plane_id), + &ddb->uv_plane[pipe][plane_id]); skl_ddb_entry_write(dev_priv, PLANE_NV12_BUF_CFG(pipe, plane_id), - &ddb->y_plane[pipe][plane_id]); + &ddb->plane[pipe][plane_id]); + } else { + skl_ddb_entry_write(dev_priv, PLANE_BUF_CFG(pipe, plane_id), + &ddb->plane[pipe][plane_id]); + I915_WRITE(PLANE_NV12_BUF_CFG(pipe, plane_id), 0x0); + } } static void skl_write_cursor_wm(struct intel_crtc *intel_crtc, @@ -4944,15 +5099,13 @@ skl_ddb_add_affected_planes(struct intel_crtc_state *cstate) struct drm_plane *plane; enum pipe pipe = intel_crtc->pipe; - WARN_ON(!drm_atomic_get_existing_crtc_state(state, crtc)); - drm_for_each_plane_mask(plane, dev, cstate->base.plane_mask) { enum plane_id plane_id = to_intel_plane(plane)->id; if (skl_ddb_entry_equal(&cur_ddb->plane[pipe][plane_id], &new_ddb->plane[pipe][plane_id]) && - skl_ddb_entry_equal(&cur_ddb->y_plane[pipe][plane_id], - &new_ddb->y_plane[pipe][plane_id])) + skl_ddb_entry_equal(&cur_ddb->uv_plane[pipe][plane_id], + &new_ddb->uv_plane[pipe][plane_id])) continue; plane_state = drm_atomic_get_plane_state(state, plane); @@ -4966,69 +5119,16 @@ skl_ddb_add_affected_planes(struct intel_crtc_state *cstate) static int skl_compute_ddb(struct drm_atomic_state *state) { - struct drm_device *dev = state->dev; - struct drm_i915_private *dev_priv = to_i915(dev); + const struct drm_i915_private *dev_priv = to_i915(state->dev); struct intel_atomic_state *intel_state = to_intel_atomic_state(state); - struct intel_crtc *intel_crtc; struct skl_ddb_allocation *ddb = &intel_state->wm_results.ddb; - uint32_t realloc_pipes = pipes_modified(state); - int ret; - - /* - * If this is our first atomic update following hardware readout, - * we can't trust the DDB that the BIOS programmed for us. Let's - * pretend that all pipes switched active status so that we'll - * ensure a full DDB recompute. - */ - if (dev_priv->wm.distrust_bios_wm) { - ret = drm_modeset_lock(&dev->mode_config.connection_mutex, - state->acquire_ctx); - if (ret) - return ret; - - intel_state->active_pipe_changes = ~0; - - /* - * We usually only initialize intel_state->active_crtcs if we - * we're doing a modeset; make sure this field is always - * initialized during the sanitization process that happens - * on the first commit too. - */ - if (!intel_state->modeset) - intel_state->active_crtcs = dev_priv->active_crtcs; - } - - /* - * If the modeset changes which CRTC's are active, we need to - * recompute the DDB allocation for *all* active pipes, even - * those that weren't otherwise being modified in any way by this - * atomic commit. Due to the shrinking of the per-pipe allocations - * when new active CRTC's are added, it's possible for a pipe that - * we were already using and aren't changing at all here to suddenly - * become invalid if its DDB needs exceeds its new allocation. - * - * Note that if we wind up doing a full DDB recompute, we can't let - * any other display updates race with this transaction, so we need - * to grab the lock on *all* CRTC's. - */ - if (intel_state->active_pipe_changes) { - realloc_pipes = ~0; - intel_state->wm_results.dirty_pipes = ~0; - } + struct intel_crtc *crtc; + struct intel_crtc_state *cstate; + int ret, i; - /* - * We're not recomputing for the pipes not included in the commit, so - * make sure we start with the current state. - */ memcpy(ddb, &dev_priv->wm.skl_hw.ddb, sizeof(*ddb)); - for_each_intel_crtc_mask(dev, intel_crtc, realloc_pipes) { - struct intel_crtc_state *cstate; - - cstate = intel_atomic_get_crtc_state(state, intel_crtc); - if (IS_ERR(cstate)) - return PTR_ERR(cstate); - + for_each_new_intel_crtc_in_state(intel_state, crtc, cstate, i) { ret = skl_allocate_pipe_ddb(cstate, ddb); if (ret) return ret; @@ -5042,14 +5142,15 @@ skl_compute_ddb(struct drm_atomic_state *state) } static void -skl_copy_wm_for_pipe(struct skl_wm_values *dst, - struct skl_wm_values *src, - enum pipe pipe) +skl_copy_ddb_for_pipe(struct skl_ddb_values *dst, + struct skl_ddb_values *src, + enum pipe pipe) { - memcpy(dst->ddb.y_plane[pipe], src->ddb.y_plane[pipe], - sizeof(dst->ddb.y_plane[pipe])); + memcpy(dst->ddb.uv_plane[pipe], src->ddb.uv_plane[pipe], + sizeof(dst->ddb.uv_plane[pipe])); memcpy(dst->ddb.plane[pipe], src->ddb.plane[pipe], sizeof(dst->ddb.plane[pipe])); + dst->ddb.enabled_slices = src->ddb.enabled_slices; } static void @@ -5090,23 +5191,23 @@ skl_print_wm_changes(const struct drm_atomic_state *state) } static int -skl_compute_wm(struct drm_atomic_state *state) +skl_ddb_add_affected_pipes(struct drm_atomic_state *state, bool *changed) { - struct drm_crtc *crtc; - struct drm_crtc_state *cstate; - struct intel_atomic_state *intel_state = to_intel_atomic_state(state); - struct skl_wm_values *results = &intel_state->wm_results; struct drm_device *dev = state->dev; - struct skl_pipe_wm *pipe_wm; - bool changed = false; + const struct drm_i915_private *dev_priv = to_i915(dev); + const struct drm_crtc *crtc; + const struct drm_crtc_state *cstate; + struct intel_crtc *intel_crtc; + struct intel_atomic_state *intel_state = to_intel_atomic_state(state); + uint32_t realloc_pipes = pipes_modified(state); int ret, i; /* * When we distrust bios wm we always need to recompute to set the * expected DDB allocations for each CRTC. */ - if (to_i915(dev)->wm.distrust_bios_wm) - changed = true; + if (dev_priv->wm.distrust_bios_wm) + (*changed) = true; /* * If this transaction isn't actually touching any CRTC's, don't @@ -5117,14 +5218,86 @@ skl_compute_wm(struct drm_atomic_state *state) * hold _all_ CRTC state mutexes. */ for_each_new_crtc_in_state(state, crtc, cstate, i) - changed = true; + (*changed) = true; - if (!changed) + if (!*changed) return 0; + /* + * If this is our first atomic update following hardware readout, + * we can't trust the DDB that the BIOS programmed for us. Let's + * pretend that all pipes switched active status so that we'll + * ensure a full DDB recompute. + */ + if (dev_priv->wm.distrust_bios_wm) { + ret = drm_modeset_lock(&dev->mode_config.connection_mutex, + state->acquire_ctx); + if (ret) + return ret; + + intel_state->active_pipe_changes = ~0; + + /* + * We usually only initialize intel_state->active_crtcs if we + * we're doing a modeset; make sure this field is always + * initialized during the sanitization process that happens + * on the first commit too. + */ + if (!intel_state->modeset) + intel_state->active_crtcs = dev_priv->active_crtcs; + } + + /* + * If the modeset changes which CRTC's are active, we need to + * recompute the DDB allocation for *all* active pipes, even + * those that weren't otherwise being modified in any way by this + * atomic commit. Due to the shrinking of the per-pipe allocations + * when new active CRTC's are added, it's possible for a pipe that + * we were already using and aren't changing at all here to suddenly + * become invalid if its DDB needs exceeds its new allocation. + * + * Note that if we wind up doing a full DDB recompute, we can't let + * any other display updates race with this transaction, so we need + * to grab the lock on *all* CRTC's. + */ + if (intel_state->active_pipe_changes) { + realloc_pipes = ~0; + intel_state->wm_results.dirty_pipes = ~0; + } + + /* + * We're not recomputing for the pipes not included in the commit, so + * make sure we start with the current state. + */ + for_each_intel_crtc_mask(dev, intel_crtc, realloc_pipes) { + struct intel_crtc_state *cstate; + + cstate = intel_atomic_get_crtc_state(state, intel_crtc); + if (IS_ERR(cstate)) + return PTR_ERR(cstate); + } + + return 0; +} + +static int +skl_compute_wm(struct drm_atomic_state *state) +{ + struct drm_crtc *crtc; + struct drm_crtc_state *cstate; + struct intel_atomic_state *intel_state = to_intel_atomic_state(state); + struct skl_ddb_values *results = &intel_state->wm_results; + struct skl_pipe_wm *pipe_wm; + bool changed = false; + int ret, i; + /* Clear all dirty flags */ results->dirty_pipes = 0; + ret = skl_ddb_add_affected_pipes(state, &changed); + if (ret || !changed) + return ret; + ret = skl_compute_ddb(state); if (ret) return ret; @@ -5197,8 +5370,8 @@ static void skl_initial_wm(struct intel_atomic_state *state, struct intel_crtc *intel_crtc = to_intel_crtc(cstate->base.crtc); struct drm_device *dev = intel_crtc->base.dev; struct drm_i915_private *dev_priv = to_i915(dev); - struct skl_wm_values *results = &state->wm_results; - struct skl_wm_values *hw_vals = &dev_priv->wm.skl_hw; + struct skl_ddb_values *results = &state->wm_results; + struct skl_ddb_values *hw_vals = &dev_priv->wm.skl_hw; enum pipe pipe = intel_crtc->pipe; if ((results->dirty_pipes & drm_crtc_mask(&intel_crtc->base)) == 0) @@ -5209,7 +5382,7 @@ static void skl_initial_wm(struct intel_atomic_state *state, if (cstate->base.active_changed) skl_atomic_update_crtc_wm(state, cstate); - skl_copy_wm_for_pipe(hw_vals, results, pipe); + skl_copy_ddb_for_pipe(hw_vals, results, pipe); mutex_unlock(&dev_priv->wm.wm_mutex); } @@ -5341,7 +5514,7 @@ void skl_pipe_wm_get_hw_state(struct drm_crtc *crtc, void skl_wm_get_hw_state(struct drm_device *dev) { struct drm_i915_private *dev_priv = to_i915(dev); - struct skl_wm_values *hw = &dev_priv->wm.skl_hw; + struct skl_ddb_values *hw = &dev_priv->wm.skl_hw; struct skl_ddb_allocation *ddb = &dev_priv->wm.skl_hw.ddb; struct drm_crtc *crtc; struct intel_crtc *intel_crtc; @@ -5362,8 +5535,12 @@ void skl_wm_get_hw_state(struct drm_device *dev) /* Fully recompute DDB on first atomic commit */ dev_priv->wm.distrust_bios_wm = true; } else { - /* Easy/common case; just sanitize DDB now if everything off */ - memset(ddb, 0, sizeof(*ddb)); + /* + * Easy/common case; just sanitize DDB now if everything off + * Keep dbuf slice info intact + */ + memset(ddb->plane, 0, sizeof(ddb->plane)); + memset(ddb->uv_plane, 0, sizeof(ddb->uv_plane)); } } @@ -6572,7 +6749,7 @@ static void gen6_init_rps_frequencies(struct drm_i915_private *dev_priv) rps->efficient_freq = rps->rp1_freq; if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv) || - IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { + IS_GEN9_BC(dev_priv) || INTEL_GEN(dev_priv) >= 10) { u32 ddcc_status = 0; if (sandybridge_pcode_read(dev_priv, @@ -6585,7 +6762,7 @@ static void gen6_init_rps_frequencies(struct drm_i915_private *dev_priv) rps->max_freq); } - if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { + if (IS_GEN9_BC(dev_priv) || INTEL_GEN(dev_priv) >= 10) { /* Store the frequency values in 16.66 MHZ units, which is * the natural hardware unit for SKL */ @@ -6890,15 +7067,18 @@ static void gen6_enable_rps(struct drm_i915_private *dev_priv) static void gen6_update_ring_freq(struct drm_i915_private *dev_priv) { struct intel_rps *rps = &dev_priv->gt_pm.rps; - int min_freq = 15; + const int min_freq = 15; + const int scaling_factor = 180; unsigned int gpu_freq; unsigned int max_ia_freq, min_ring_freq; unsigned int max_gpu_freq, min_gpu_freq; - int scaling_factor = 180; struct cpufreq_policy *policy; WARN_ON(!mutex_is_locked(&dev_priv->pcu_lock)); + if (rps->max_freq <= rps->min_freq) + return; + policy = cpufreq_cpu_get(0); if (policy) { max_ia_freq = policy->cpuinfo.max_freq; @@ -6918,13 +7098,12 @@ static void gen6_update_ring_freq(struct drm_i915_private *dev_priv) /* convert DDR frequency from units of 266.6MHz to bandwidth */ min_ring_freq = mult_frac(min_ring_freq, 8, 3); - if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { + min_gpu_freq = rps->min_freq; + max_gpu_freq = rps->max_freq; + if (IS_GEN9_BC(dev_priv) || INTEL_GEN(dev_priv) >= 10) { /* Convert GT frequency to 50 HZ units */ - min_gpu_freq = rps->min_freq / GEN9_FREQ_SCALER; - max_gpu_freq = rps->max_freq / GEN9_FREQ_SCALER; - } else { - min_gpu_freq = rps->min_freq; - max_gpu_freq = rps->max_freq; + min_gpu_freq /= GEN9_FREQ_SCALER; + max_gpu_freq /= GEN9_FREQ_SCALER; } /* @@ -6933,10 +7112,10 @@ static void gen6_update_ring_freq(struct drm_i915_private *dev_priv) * the PCU should use as a reference to determine the ring frequency. */ for (gpu_freq = max_gpu_freq; gpu_freq >= min_gpu_freq; gpu_freq--) { - int diff = max_gpu_freq - gpu_freq; + const int diff = max_gpu_freq - gpu_freq; unsigned int ia_freq = 0, ring_freq = 0; - if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { + if (IS_GEN9_BC(dev_priv) || INTEL_GEN(dev_priv) >= 10) { /* * ring_freq = 2 * GT. ring_freq is in 100MHz units * No floor required for ring frequency on SKL. @@ -8026,10 +8205,10 @@ void intel_sanitize_gt_powersave(struct drm_i915_private *dev_priv) dev_priv->gt_pm.rc6.enabled = true; /* force RC6 disabling */ intel_disable_gt_powersave(dev_priv); - if (INTEL_GEN(dev_priv) < 11) - gen6_reset_rps_interrupts(dev_priv); + if (INTEL_GEN(dev_priv) >= 11) + gen11_reset_rps_interrupts(dev_priv); else - WARN_ON_ONCE(1); + gen6_reset_rps_interrupts(dev_priv); } static inline void intel_disable_llc_pstate(struct drm_i915_private *i915) @@ -8142,8 +8321,6 @@ static void intel_enable_rps(struct drm_i915_private *dev_priv) cherryview_enable_rps(dev_priv); } else if (IS_VALLEYVIEW(dev_priv)) { valleyview_enable_rps(dev_priv); - } else if (WARN_ON_ONCE(INTEL_GEN(dev_priv) >= 11)) { - /* TODO */ } else if (INTEL_GEN(dev_priv) >= 9) { gen9_enable_rps(dev_priv); } else if (IS_BROADWELL(dev_priv)) { @@ -8487,6 +8664,13 @@ static void gen8_set_l3sqc_credits(struct drm_i915_private *dev_priv, I915_WRITE(GEN7_MISCCPCTL, misccpctl); } +static void icl_init_clock_gating(struct drm_i915_private *dev_priv) +{ + /* This is not an Wa. Enable to reduce Sampler power */ + I915_WRITE(GEN10_DFR_RATIO_EN_AND_CHICKEN, + I915_READ(GEN10_DFR_RATIO_EN_AND_CHICKEN) & ~DFR_DISABLE); +} + static void cnp_init_clock_gating(struct drm_i915_private *dev_priv) { if (!HAS_PCH_CNP(dev_priv)) @@ -9013,7 +9197,9 @@ static void nop_init_clock_gating(struct drm_i915_private *dev_priv) */ void intel_init_clock_gating_hooks(struct drm_i915_private *dev_priv) { - if (IS_CANNONLAKE(dev_priv)) + if (IS_ICELAKE(dev_priv)) + dev_priv->display.init_clock_gating = icl_init_clock_gating; + else if (IS_CANNONLAKE(dev_priv)) dev_priv->display.init_clock_gating = cnl_init_clock_gating; else if (IS_COFFEELAKE(dev_priv)) dev_priv->display.init_clock_gating = cfl_init_clock_gating; diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index 23175c5..db27f2f 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -93,7 +93,115 @@ static void psr_aux_io_power_put(struct intel_dp *intel_dp) intel_display_power_put(dev_priv, psr_aux_domain(intel_dp)); } -static bool intel_dp_get_y_cord_status(struct intel_dp *intel_dp) +void intel_psr_irq_control(struct drm_i915_private *dev_priv, bool debug) +{ + u32 debug_mask, mask; + + /* No PSR interrupts on VLV/CHV */ + if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) + return; + + mask = EDP_PSR_ERROR(TRANSCODER_EDP); + debug_mask = EDP_PSR_POST_EXIT(TRANSCODER_EDP) | + EDP_PSR_PRE_ENTRY(TRANSCODER_EDP); + + if (INTEL_GEN(dev_priv) >= 8) { + mask |= EDP_PSR_ERROR(TRANSCODER_A) | + EDP_PSR_ERROR(TRANSCODER_B) | + EDP_PSR_ERROR(TRANSCODER_C); + + debug_mask |= EDP_PSR_POST_EXIT(TRANSCODER_A) | + EDP_PSR_PRE_ENTRY(TRANSCODER_A) | + EDP_PSR_POST_EXIT(TRANSCODER_B) | + EDP_PSR_PRE_ENTRY(TRANSCODER_B) | + EDP_PSR_POST_EXIT(TRANSCODER_C) | + EDP_PSR_PRE_ENTRY(TRANSCODER_C); + } + + if (debug) + mask |= debug_mask; + + WRITE_ONCE(dev_priv->psr.debug, debug); + I915_WRITE(EDP_PSR_IMR, ~mask); +} + +static void psr_event_print(u32 val, bool psr2_enabled) +{ + DRM_DEBUG_KMS("PSR exit events: 0x%x\n", val); + if (val & PSR_EVENT_PSR2_WD_TIMER_EXPIRE) + DRM_DEBUG_KMS("\tPSR2 watchdog timer expired\n"); + if ((val & PSR_EVENT_PSR2_DISABLED) && psr2_enabled) + DRM_DEBUG_KMS("\tPSR2 disabled\n"); + if (val & PSR_EVENT_SU_DIRTY_FIFO_UNDERRUN) + DRM_DEBUG_KMS("\tSU dirty FIFO underrun\n"); + if (val & PSR_EVENT_SU_CRC_FIFO_UNDERRUN) + DRM_DEBUG_KMS("\tSU CRC FIFO underrun\n"); + if (val & PSR_EVENT_GRAPHICS_RESET) + DRM_DEBUG_KMS("\tGraphics reset\n"); + if (val & PSR_EVENT_PCH_INTERRUPT) + DRM_DEBUG_KMS("\tPCH interrupt\n"); + if (val & PSR_EVENT_MEMORY_UP) + DRM_DEBUG_KMS("\tMemory up\n"); + if (val & PSR_EVENT_FRONT_BUFFER_MODIFY) + DRM_DEBUG_KMS("\tFront buffer modification\n"); + if (val & PSR_EVENT_WD_TIMER_EXPIRE) + DRM_DEBUG_KMS("\tPSR watchdog timer expired\n"); + if (val & PSR_EVENT_PIPE_REGISTERS_UPDATE) + DRM_DEBUG_KMS("\tPIPE registers updated\n"); + if (val & PSR_EVENT_REGISTER_UPDATE) + DRM_DEBUG_KMS("\tRegister updated\n"); + if (val & PSR_EVENT_HDCP_ENABLE) + DRM_DEBUG_KMS("\tHDCP enabled\n"); + if (val & PSR_EVENT_KVMR_SESSION_ENABLE) + DRM_DEBUG_KMS("\tKVMR session enabled\n"); + if (val & PSR_EVENT_VBI_ENABLE) + DRM_DEBUG_KMS("\tVBI enabled\n"); + if (val & PSR_EVENT_LPSP_MODE_EXIT) + DRM_DEBUG_KMS("\tLPSP mode exited\n"); + if ((val & PSR_EVENT_PSR_DISABLE) && !psr2_enabled) + DRM_DEBUG_KMS("\tPSR disabled\n"); +} + +void intel_psr_irq_handler(struct drm_i915_private *dev_priv, u32 psr_iir) +{ + u32 transcoders = BIT(TRANSCODER_EDP); + enum transcoder cpu_transcoder; + ktime_t time_ns = ktime_get(); + + if (INTEL_GEN(dev_priv) >= 8) + transcoders |= BIT(TRANSCODER_A) | + BIT(TRANSCODER_B) | + BIT(TRANSCODER_C); + + for_each_cpu_transcoder_masked(dev_priv, cpu_transcoder, transcoders) { + /* FIXME: Exit PSR and link train manually when this happens. */ + if (psr_iir & EDP_PSR_ERROR(cpu_transcoder)) + DRM_DEBUG_KMS("[transcoder %s] PSR aux error\n", + transcoder_name(cpu_transcoder)); + + if (psr_iir & EDP_PSR_PRE_ENTRY(cpu_transcoder)) { + dev_priv->psr.last_entry_attempt = time_ns; + DRM_DEBUG_KMS("[transcoder %s] PSR entry attempt in 2 vblanks\n", + transcoder_name(cpu_transcoder)); + } + + if (psr_iir & EDP_PSR_POST_EXIT(cpu_transcoder)) { + dev_priv->psr.last_exit = time_ns; + DRM_DEBUG_KMS("[transcoder %s] PSR exit completed\n", + transcoder_name(cpu_transcoder)); + + if (INTEL_GEN(dev_priv) >= 9) { + u32 val = I915_READ(PSR_EVENT(cpu_transcoder)); + bool psr2_enabled = dev_priv->psr.psr2_enabled; + + I915_WRITE(PSR_EVENT(cpu_transcoder), val); + psr_event_print(val, psr2_enabled); + } + } + } +} + +static bool intel_dp_get_y_coord_required(struct intel_dp *intel_dp) { uint8_t psr_caps = 0; @@ -122,6 +230,18 @@ static bool intel_dp_get_alpm_status(struct intel_dp *intel_dp) return alpm_caps & DP_ALPM_CAP; } +static u8 intel_dp_get_sink_sync_latency(struct intel_dp *intel_dp) +{ + u8 val = 0; + + if (drm_dp_dpcd_readb(&intel_dp->aux, + DP_SYNCHRONIZATION_LATENCY_IN_SINK, &val) == 1) + val &= DP_MAX_RESYNC_FRAME_COUNT_MASK; + else + DRM_ERROR("Unable to get sink synchronization latency\n"); + return val; +} + void intel_psr_init_dpcd(struct intel_dp *intel_dp) { struct drm_i915_private *dev_priv = @@ -130,33 +250,36 @@ void intel_psr_init_dpcd(struct intel_dp *intel_dp) drm_dp_dpcd_read(&intel_dp->aux, DP_PSR_SUPPORT, intel_dp->psr_dpcd, sizeof(intel_dp->psr_dpcd)); - if (intel_dp->psr_dpcd[0] & DP_PSR_IS_SUPPORTED) { + if (intel_dp->psr_dpcd[0]) { dev_priv->psr.sink_support = true; DRM_DEBUG_KMS("Detected EDP PSR Panel.\n"); } if (INTEL_GEN(dev_priv) >= 9 && - (intel_dp->psr_dpcd[0] & DP_PSR2_IS_SUPPORTED)) { - uint8_t frame_sync_cap; + (intel_dp->psr_dpcd[0] == DP_PSR2_WITH_Y_COORD_IS_SUPPORTED)) { + /* + * All panels that supports PSR version 03h (PSR2 + + * Y-coordinate) can handle Y-coordinates in VSC but we are + * only sure that it is going to be used when required by the + * panel. This way panel is capable to do selective update + * without a aux frame sync. + * + * To support PSR version 02h and PSR version 03h without + * Y-coordinate requirement panels we would need to enable + * GTC first. + */ + dev_priv->psr.sink_psr2_support = + intel_dp_get_y_coord_required(intel_dp); + DRM_DEBUG_KMS("PSR2 %s on sink", dev_priv->psr.sink_psr2_support + ? "supported" : "not supported"); - dev_priv->psr.sink_support = true; - if (drm_dp_dpcd_readb(&intel_dp->aux, - DP_SINK_DEVICE_AUX_FRAME_SYNC_CAP, - &frame_sync_cap) != 1) - frame_sync_cap = 0; - dev_priv->psr.aux_frame_sync = frame_sync_cap & DP_AUX_FRAME_SYNC_CAP; - /* PSR2 needs frame sync as well */ - dev_priv->psr.psr2_support = dev_priv->psr.aux_frame_sync; - 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); + if (dev_priv->psr.sink_psr2_support) { dev_priv->psr.colorimetry_support = intel_dp_get_colorimetry_status(intel_dp); dev_priv->psr.alpm = intel_dp_get_alpm_status(intel_dp); + dev_priv->psr.sink_sync_latency = + intel_dp_get_sink_sync_latency(intel_dp); } } } @@ -193,21 +316,17 @@ static void hsw_psr_setup_vsc(struct intel_dp *intel_dp, struct drm_i915_private *dev_priv = to_i915(intel_dig_port->base.base.dev); struct edp_vsc_psr psr_vsc; - if (dev_priv->psr.psr2_support) { + if (dev_priv->psr.psr2_enabled) { /* Prepare VSC Header for SU as per EDP 1.4 spec, Table 6.11 */ memset(&psr_vsc, 0, sizeof(psr_vsc)); psr_vsc.sdp_header.HB0 = 0; psr_vsc.sdp_header.HB1 = 0x7; - if (dev_priv->psr.colorimetry_support && - dev_priv->psr.y_cord_support) { + if (dev_priv->psr.colorimetry_support) { psr_vsc.sdp_header.HB2 = 0x5; psr_vsc.sdp_header.HB3 = 0x13; - } else if (dev_priv->psr.y_cord_support) { + } else { psr_vsc.sdp_header.HB2 = 0x4; psr_vsc.sdp_header.HB3 = 0xe; - } else { - psr_vsc.sdp_header.HB2 = 0x3; - psr_vsc.sdp_header.HB3 = 0xc; } } else { /* Prepare VSC packet as per EDP 1.3 spec, Table 3.10 */ @@ -228,31 +347,12 @@ static void vlv_psr_enable_sink(struct intel_dp *intel_dp) DP_PSR_ENABLE | DP_PSR_MAIN_LINK_ACTIVE); } -static i915_reg_t psr_aux_ctl_reg(struct drm_i915_private *dev_priv, - enum port port) -{ - if (INTEL_GEN(dev_priv) >= 9) - return DP_AUX_CH_CTL(port); - else - return EDP_PSR_AUX_CTL; -} - -static i915_reg_t psr_aux_data_reg(struct drm_i915_private *dev_priv, - enum port port, int index) -{ - if (INTEL_GEN(dev_priv) >= 9) - return DP_AUX_CH_DATA(port, index); - else - return EDP_PSR_AUX_DATA(index); -} - -static void hsw_psr_enable_sink(struct intel_dp *intel_dp) +static void hsw_psr_setup_aux(struct intel_dp *intel_dp) { struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - struct drm_device *dev = dig_port->base.base.dev; - struct drm_i915_private *dev_priv = to_i915(dev); - uint32_t aux_clock_divider; - i915_reg_t aux_ctl_reg; + struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev); + u32 aux_clock_divider, aux_ctl; + int i; static const uint8_t aux_msg[] = { [0] = DP_AUX_NATIVE_WRITE << 4, [1] = DP_SET_POWER >> 8, @@ -260,41 +360,47 @@ static void hsw_psr_enable_sink(struct intel_dp *intel_dp) [3] = 1 - 1, [4] = DP_SET_POWER_D0, }; - enum port port = dig_port->base.port; - u32 aux_ctl; - int i; + u32 psr_aux_mask = EDP_PSR_AUX_CTL_TIME_OUT_MASK | + EDP_PSR_AUX_CTL_MESSAGE_SIZE_MASK | + EDP_PSR_AUX_CTL_PRECHARGE_2US_MASK | + EDP_PSR_AUX_CTL_BIT_CLOCK_2X_MASK; BUILD_BUG_ON(sizeof(aux_msg) > 20); + for (i = 0; i < sizeof(aux_msg); i += 4) + I915_WRITE(EDP_PSR_AUX_DATA(i >> 2), + intel_dp_pack_aux(&aux_msg[i], sizeof(aux_msg) - i)); aux_clock_divider = intel_dp->get_aux_clock_divider(intel_dp, 0); - /* Enable AUX frame sync at sink */ - if (dev_priv->psr.aux_frame_sync) - drm_dp_dpcd_writeb(&intel_dp->aux, - DP_SINK_DEVICE_AUX_FRAME_SYNC_CONF, - DP_AUX_FRAME_SYNC_ENABLE); + /* Start with bits set for DDI_AUX_CTL register */ + aux_ctl = intel_dp->get_aux_send_ctl(intel_dp, 0, sizeof(aux_msg), + aux_clock_divider); + + /* Select only valid bits for SRD_AUX_CTL */ + aux_ctl &= psr_aux_mask; + I915_WRITE(EDP_PSR_AUX_CTL, aux_ctl); +} + +static void hsw_psr_enable_sink(struct intel_dp *intel_dp) +{ + struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); + struct drm_device *dev = dig_port->base.base.dev; + struct drm_i915_private *dev_priv = to_i915(dev); + u8 dpcd_val = DP_PSR_ENABLE; + /* Enable ALPM at sink for psr2 */ - if (dev_priv->psr.psr2_support && dev_priv->psr.alpm) + if (dev_priv->psr.psr2_enabled && dev_priv->psr.alpm) drm_dp_dpcd_writeb(&intel_dp->aux, DP_RECEIVER_ALPM_CONFIG, DP_ALPM_ENABLE); - if (dev_priv->psr.link_standby) - drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, - DP_PSR_ENABLE | DP_PSR_MAIN_LINK_ACTIVE); - else - drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, - DP_PSR_ENABLE); - - aux_ctl_reg = psr_aux_ctl_reg(dev_priv, port); - /* Setup AUX registers */ - for (i = 0; i < sizeof(aux_msg); i += 4) - I915_WRITE(psr_aux_data_reg(dev_priv, port, i >> 2), - intel_dp_pack_aux(&aux_msg[i], sizeof(aux_msg) - i)); + if (dev_priv->psr.psr2_enabled) + dpcd_val |= DP_PSR_ENABLE_PSR2; + if (dev_priv->psr.link_standby) + dpcd_val |= DP_PSR_MAIN_LINK_ACTIVE; + drm_dp_dpcd_writeb(&intel_dp->aux, DP_PSR_EN_CFG, dpcd_val); - aux_ctl = intel_dp->get_aux_send_ctl(intel_dp, 0, sizeof(aux_msg), - aux_clock_divider); - I915_WRITE(aux_ctl_reg, aux_ctl); + drm_dp_dpcd_writeb(&intel_dp->aux, DP_SET_POWER, DP_SET_POWER_D0); } static void vlv_psr_enable_source(struct intel_dp *intel_dp, @@ -396,25 +502,16 @@ static void hsw_activate_psr2(struct intel_dp *intel_dp) * with the 5 or 6 idle patterns. */ uint32_t idle_frames = max(6, dev_priv->vbt.psr.idle_frames); - uint32_t val; - uint8_t sink_latency; - - val = idle_frames << EDP_PSR_IDLE_FRAME_SHIFT; + u32 val = idle_frames << EDP_PSR2_IDLE_FRAME_SHIFT; /* FIXME: selective update is probably totally broken because it doesn't * mesh at all with our frontbuffer tracking. And the hw alone isn't * good enough. */ - val |= EDP_PSR2_ENABLE | - EDP_SU_TRACK_ENABLE; + val |= EDP_PSR2_ENABLE | EDP_SU_TRACK_ENABLE; + if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv)) + val |= EDP_Y_COORDINATE_ENABLE; - if (drm_dp_dpcd_readb(&intel_dp->aux, - DP_SYNCHRONIZATION_LATENCY_IN_SINK, - &sink_latency) == 1) { - sink_latency &= DP_MAX_RESYNC_FRAME_COUNT_MASK; - } else { - sink_latency = 0; - } - val |= EDP_PSR2_FRAME_BEFORE_SU(sink_latency + 1); + val |= EDP_PSR2_FRAME_BEFORE_SU(dev_priv->psr.sink_sync_latency + 1); if (dev_priv->vbt.psr.tp2_tp3_wakeup_time > 5) val |= EDP_PSR2_TP2_TIME_2500; @@ -440,7 +537,7 @@ static void hsw_psr_activate(struct intel_dp *intel_dp) */ /* psr1 and psr2 are mutually exclusive.*/ - if (dev_priv->psr.psr2_support) + if (dev_priv->psr.psr2_enabled) hsw_activate_psr2(intel_dp); else hsw_activate_psr1(intel_dp); @@ -460,7 +557,7 @@ static bool intel_psr2_config_valid(struct intel_dp *intel_dp, * dynamically during PSR enable, and extracted from sink * caps during eDP detection. */ - if (!dev_priv->psr.psr2_support) + if (!dev_priv->psr.sink_psr2_support) return false; if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv)) { @@ -478,15 +575,6 @@ static bool intel_psr2_config_valid(struct intel_dp *intel_dp, return false; } - /* - * FIXME:enable psr2 only for y-cordinate psr2 panels - * After gtc implementation , remove this restriction. - */ - if (!dev_priv->psr.y_cord_support) { - DRM_DEBUG_KMS("PSR2 not enabled, panel does not support Y coordinate\n"); - return false; - } - return true; } @@ -568,7 +656,7 @@ static void intel_psr_activate(struct intel_dp *intel_dp) struct drm_device *dev = intel_dig_port->base.base.dev; struct drm_i915_private *dev_priv = to_i915(dev); - if (dev_priv->psr.psr2_support) + if (dev_priv->psr.psr2_enabled) WARN_ON(I915_READ(EDP_PSR2_CTL) & EDP_PSR2_ENABLE); else WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE); @@ -586,14 +674,24 @@ static void hsw_psr_enable_source(struct intel_dp *intel_dp, struct drm_device *dev = dig_port->base.base.dev; struct drm_i915_private *dev_priv = to_i915(dev); enum transcoder cpu_transcoder = crtc_state->cpu_transcoder; - u32 chicken; psr_aux_io_power_get(intel_dp); - if (dev_priv->psr.psr2_support) { - chicken = PSR2_VSC_ENABLE_PROG_HEADER; - if (dev_priv->psr.y_cord_support) - chicken |= PSR2_ADD_VERTICAL_LINE_COUNT; + /* Only HSW and BDW have PSR AUX registers that need to be setup. SKL+ + * use hardcoded values PSR AUX transactions + */ + if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) + hsw_psr_setup_aux(intel_dp); + + if (dev_priv->psr.psr2_enabled) { + u32 chicken = I915_READ(CHICKEN_TRANS(cpu_transcoder)); + + if (INTEL_GEN(dev_priv) == 9 && !IS_GEMINILAKE(dev_priv)) + chicken |= (PSR2_VSC_ENABLE_PROG_HEADER + | PSR2_ADD_VERTICAL_LINE_COUNT); + + else + chicken &= ~VSC_DATA_SEL_SOFTWARE_CONTROL; I915_WRITE(CHICKEN_TRANS(cpu_transcoder), chicken); I915_WRITE(EDP_PSR_DEBUG, @@ -613,7 +711,8 @@ static void hsw_psr_enable_source(struct intel_dp *intel_dp, I915_WRITE(EDP_PSR_DEBUG, EDP_PSR_DEBUG_MASK_MEMUP | EDP_PSR_DEBUG_MASK_HPD | - EDP_PSR_DEBUG_MASK_LPSP); + EDP_PSR_DEBUG_MASK_LPSP | + EDP_PSR_DEBUG_MASK_DISP_REG_WRITE); } } @@ -644,7 +743,7 @@ void intel_psr_enable(struct intel_dp *intel_dp, goto unlock; } - dev_priv->psr.psr2_support = crtc_state->has_psr2; + dev_priv->psr.psr2_enabled = crtc_state->has_psr2; dev_priv->psr.busy_frontbuffer_bits = 0; dev_priv->psr.setup_vsc(intel_dp, crtc_state); @@ -714,12 +813,7 @@ static void hsw_psr_disable(struct intel_dp *intel_dp, i915_reg_t psr_status; u32 psr_status_mask; - if (dev_priv->psr.aux_frame_sync) - drm_dp_dpcd_writeb(&intel_dp->aux, - DP_SINK_DEVICE_AUX_FRAME_SYNC_CONF, - 0); - - if (dev_priv->psr.psr2_support) { + if (dev_priv->psr.psr2_enabled) { psr_status = EDP_PSR2_STATUS; psr_status_mask = EDP_PSR2_STATUS_STATE_MASK; @@ -743,7 +837,7 @@ static void hsw_psr_disable(struct intel_dp *intel_dp, dev_priv->psr.active = false; } else { - if (dev_priv->psr.psr2_support) + if (dev_priv->psr.psr2_enabled) WARN_ON(I915_READ(EDP_PSR2_CTL) & EDP_PSR2_ENABLE); else WARN_ON(I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE); @@ -789,53 +883,59 @@ void intel_psr_disable(struct intel_dp *intel_dp, cancel_delayed_work_sync(&dev_priv->psr.work); } -static void intel_psr_work(struct work_struct *work) +static bool psr_wait_for_idle(struct drm_i915_private *dev_priv) { - struct drm_i915_private *dev_priv = - container_of(work, typeof(*dev_priv), psr.work.work); - struct intel_dp *intel_dp = dev_priv->psr.enabled; - struct drm_crtc *crtc = dp_to_dig_port(intel_dp)->base.base.crtc; - enum pipe pipe = to_intel_crtc(crtc)->pipe; + struct intel_dp *intel_dp; + i915_reg_t reg; + u32 mask; + int err; + + intel_dp = dev_priv->psr.enabled; + if (!intel_dp) + return false; - /* We have to make sure PSR is ready for re-enable - * otherwise it keeps disabled until next full enable/disable cycle. - * PSR might take some time to get fully disabled - * and be ready for re-enable. - */ if (HAS_DDI(dev_priv)) { - if (dev_priv->psr.psr2_support) { - if (intel_wait_for_register(dev_priv, - EDP_PSR2_STATUS, - EDP_PSR2_STATUS_STATE_MASK, - 0, - 50)) { - DRM_ERROR("Timed out waiting for PSR2 Idle for re-enable\n"); - return; - } + if (dev_priv->psr.psr2_enabled) { + reg = EDP_PSR2_STATUS; + mask = EDP_PSR2_STATUS_STATE_MASK; } else { - if (intel_wait_for_register(dev_priv, - EDP_PSR_STATUS, - EDP_PSR_STATUS_STATE_MASK, - 0, - 50)) { - DRM_ERROR("Timed out waiting for PSR Idle for re-enable\n"); - return; - } + reg = EDP_PSR_STATUS; + mask = EDP_PSR_STATUS_STATE_MASK; } } else { - if (intel_wait_for_register(dev_priv, - VLV_PSRSTAT(pipe), - VLV_EDP_PSR_IN_TRANS, - 0, - 1)) { - DRM_ERROR("Timed out waiting for PSR Idle for re-enable\n"); - return; - } + struct drm_crtc *crtc = + dp_to_dig_port(intel_dp)->base.base.crtc; + enum pipe pipe = to_intel_crtc(crtc)->pipe; + + reg = VLV_PSRSTAT(pipe); + mask = VLV_EDP_PSR_IN_TRANS; } + + mutex_unlock(&dev_priv->psr.lock); + + err = intel_wait_for_register(dev_priv, reg, mask, 0, 50); + if (err) + DRM_ERROR("Timed out waiting for PSR Idle for re-enable\n"); + + /* After the unlocked wait, verify that PSR is still wanted! */ mutex_lock(&dev_priv->psr.lock); - intel_dp = dev_priv->psr.enabled; + return err == 0 && dev_priv->psr.enabled; +} - if (!intel_dp) +static void intel_psr_work(struct work_struct *work) +{ + struct drm_i915_private *dev_priv = + container_of(work, typeof(*dev_priv), psr.work.work); + + mutex_lock(&dev_priv->psr.lock); + + /* + * We have to make sure PSR is ready for re-enable + * otherwise it keeps disabled until next full enable/disable cycle. + * PSR might take some time to get fully disabled + * and be ready for re-enable. + */ + if (!psr_wait_for_idle(dev_priv)) goto unlock; /* @@ -846,7 +946,7 @@ static void intel_psr_work(struct work_struct *work) if (dev_priv->psr.busy_frontbuffer_bits) goto unlock; - intel_psr_activate(intel_dp); + intel_psr_activate(dev_priv->psr.enabled); unlock: mutex_unlock(&dev_priv->psr.lock); } @@ -862,11 +962,7 @@ static void intel_psr_exit(struct drm_i915_private *dev_priv) return; if (HAS_DDI(dev_priv)) { - if (dev_priv->psr.aux_frame_sync) - drm_dp_dpcd_writeb(&intel_dp->aux, - DP_SINK_DEVICE_AUX_FRAME_SYNC_CONF, - 0); - if (dev_priv->psr.psr2_support) { + if (dev_priv->psr.psr2_enabled) { val = I915_READ(EDP_PSR2_CTL); WARN_ON(!(val & EDP_PSR2_ENABLE)); I915_WRITE(EDP_PSR2_CTL, val & ~EDP_PSR2_ENABLE); @@ -957,6 +1053,7 @@ void intel_psr_single_frame_update(struct drm_i915_private *dev_priv, * intel_psr_invalidate - Invalidade PSR * @dev_priv: i915 device * @frontbuffer_bits: frontbuffer plane tracking bits + * @origin: which operation caused the invalidate * * Since the hardware frontbuffer tracking has gaps we need to integrate * with the software frontbuffer tracking. This function gets called every @@ -966,7 +1063,7 @@ void intel_psr_single_frame_update(struct drm_i915_private *dev_priv, * Dirty frontbuffers relevant to PSR are tracked in busy_frontbuffer_bits." */ void intel_psr_invalidate(struct drm_i915_private *dev_priv, - unsigned frontbuffer_bits) + unsigned frontbuffer_bits, enum fb_op_origin origin) { struct drm_crtc *crtc; enum pipe pipe; @@ -974,6 +1071,9 @@ void intel_psr_invalidate(struct drm_i915_private *dev_priv, if (!CAN_PSR(dev_priv)) return; + if (dev_priv->psr.has_hw_tracking && origin == ORIGIN_FLIP) + return; + mutex_lock(&dev_priv->psr.lock); if (!dev_priv->psr.enabled) { mutex_unlock(&dev_priv->psr.lock); @@ -1014,6 +1114,9 @@ void intel_psr_flush(struct drm_i915_private *dev_priv, if (!CAN_PSR(dev_priv)) return; + if (dev_priv->psr.has_hw_tracking && origin == ORIGIN_FLIP) + return; + mutex_lock(&dev_priv->psr.lock); if (!dev_priv->psr.enabled) { mutex_unlock(&dev_priv->psr.lock); @@ -1027,8 +1130,23 @@ void intel_psr_flush(struct drm_i915_private *dev_priv, dev_priv->psr.busy_frontbuffer_bits &= ~frontbuffer_bits; /* By definition flush = invalidate + flush */ - if (frontbuffer_bits) - intel_psr_exit(dev_priv); + if (frontbuffer_bits) { + if (dev_priv->psr.psr2_enabled || + IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { + intel_psr_exit(dev_priv); + } else { + /* + * Display WA #0884: all + * This documented WA for bxt can be safely applied + * broadly so we can force HW tracking to exit PSR + * instead of disabling and re-enabling. + * Workaround tells us to write 0 to CUR_SURFLIVE_A, + * but it makes more sense write to the current active + * pipe. + */ + I915_WRITE(CURSURFLIVE(pipe), 0); + } + } if (!dev_priv->psr.active && !dev_priv->psr.busy_frontbuffer_bits) if (!work_busy(&dev_priv->psr.work.work)) @@ -1055,9 +1173,12 @@ void intel_psr_init(struct drm_i915_private *dev_priv) if (!dev_priv->psr.sink_support) return; - /* Per platform default: all disabled. */ - if (i915_modparams.enable_psr == -1) + if (i915_modparams.enable_psr == -1) { + i915_modparams.enable_psr = dev_priv->vbt.psr.enable; + + /* Per platform default: all disabled. */ i915_modparams.enable_psr = 0; + } /* Set link_standby x link_off defaults */ if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) @@ -1090,6 +1211,7 @@ void intel_psr_init(struct drm_i915_private *dev_priv) dev_priv->psr.activate = vlv_psr_activate; dev_priv->psr.setup_vsc = vlv_psr_setup_vsc; } else { + dev_priv->psr.has_hw_tracking = true; dev_priv->psr.enable_source = hsw_psr_enable_source; dev_priv->psr.disable_source = hsw_psr_disable; dev_priv->psr.enable_sink = hsw_psr_enable_sink; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 1d59952..8f19349 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -36,6 +36,7 @@ #include "i915_gem_render_state.h" #include "i915_trace.h" #include "intel_drv.h" +#include "intel_workarounds.h" /* Rough estimate of the typical request size, performing a flush, * set-context and then emitting the batch. @@ -557,7 +558,8 @@ static void reset_ring_common(struct intel_engine_cs *engine, */ if (request) { struct drm_i915_private *dev_priv = request->i915; - struct intel_context *ce = &request->ctx->engine[engine->id]; + struct intel_context *ce = to_intel_context(request->ctx, + engine); struct i915_hw_ppgtt *ppgtt; if (ce->state) { @@ -599,7 +601,7 @@ static int intel_rcs_ctx_init(struct i915_request *rq) { int ret; - ret = intel_ring_workarounds_emit(rq); + ret = intel_ctx_workarounds_emit(rq); if (ret != 0) return ret; @@ -617,6 +619,8 @@ static int init_render_ring(struct intel_engine_cs *engine) if (ret) return ret; + intel_whitelist_workarounds_apply(engine); + /* WaTimedSingleVertexDispatch:cl,bw,ctg,elk,ilk,snb */ if (IS_GEN(dev_priv, 4, 6)) I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(VS_TIMER_DISPATCH)); @@ -658,7 +662,7 @@ static int init_render_ring(struct intel_engine_cs *engine) if (INTEL_GEN(dev_priv) >= 6) I915_WRITE_IMR(engine, ~engine->irq_keep_mask); - return init_workarounds_ring(engine); + return 0; } static u32 *gen6_signal(struct i915_request *rq, u32 *cs) @@ -693,17 +697,17 @@ static void cancel_requests(struct intel_engine_cs *engine) struct i915_request *request; unsigned long flags; - spin_lock_irqsave(&engine->timeline->lock, flags); + spin_lock_irqsave(&engine->timeline.lock, flags); /* Mark all submitted requests as skipped. */ - list_for_each_entry(request, &engine->timeline->requests, link) { + list_for_each_entry(request, &engine->timeline.requests, link) { GEM_BUG_ON(!request->global_seqno); if (!i915_request_completed(request)) dma_fence_set_error(&request->fence, -EIO); } /* Remaining _unready_ requests will be nop'ed when submitted */ - spin_unlock_irqrestore(&engine->timeline->lock, flags); + spin_unlock_irqrestore(&engine->timeline.lock, flags); } static void i9xx_submit_request(struct i915_request *request) @@ -1062,7 +1066,6 @@ err: void intel_ring_reset(struct intel_ring *ring, u32 tail) { - GEM_BUG_ON(!list_empty(&ring->request_list)); ring->tail = tail; ring->head = tail; ring->emit = tail; @@ -1114,19 +1117,24 @@ err: } struct intel_ring * -intel_engine_create_ring(struct intel_engine_cs *engine, int size) +intel_engine_create_ring(struct intel_engine_cs *engine, + struct i915_timeline *timeline, + int size) { struct intel_ring *ring; struct i915_vma *vma; GEM_BUG_ON(!is_power_of_2(size)); GEM_BUG_ON(RING_CTL_SIZE(size) & ~RING_NR_PAGES); + GEM_BUG_ON(timeline == &engine->timeline); + lockdep_assert_held(&engine->i915->drm.struct_mutex); ring = kzalloc(sizeof(*ring), GFP_KERNEL); if (!ring) return ERR_PTR(-ENOMEM); INIT_LIST_HEAD(&ring->request_list); + ring->timeline = i915_timeline_get(timeline); ring->size = size; /* Workaround an erratum on the i830 which causes a hang if @@ -1157,12 +1165,13 @@ intel_ring_free(struct intel_ring *ring) i915_vma_close(ring->vma); __i915_gem_object_release_unless_active(obj); + i915_timeline_put(ring->timeline); kfree(ring); } -static int context_pin(struct i915_gem_context *ctx) +static int context_pin(struct intel_context *ce) { - struct i915_vma *vma = ctx->engine[RCS].state; + struct i915_vma *vma = ce->state; int ret; /* @@ -1253,7 +1262,7 @@ static struct intel_ring * intel_ring_context_pin(struct intel_engine_cs *engine, struct i915_gem_context *ctx) { - struct intel_context *ce = &ctx->engine[engine->id]; + struct intel_context *ce = to_intel_context(ctx, engine); int ret; lockdep_assert_held(&ctx->i915->drm.struct_mutex); @@ -1275,7 +1284,7 @@ intel_ring_context_pin(struct intel_engine_cs *engine, } if (ce->state) { - ret = context_pin(ctx); + ret = context_pin(ce); if (ret) goto err; @@ -1296,7 +1305,7 @@ err: static void intel_ring_context_unpin(struct intel_engine_cs *engine, struct i915_gem_context *ctx) { - struct intel_context *ce = &ctx->engine[engine->id]; + struct intel_context *ce = to_intel_context(ctx, engine); lockdep_assert_held(&ctx->i915->drm.struct_mutex); GEM_BUG_ON(ce->pin_count == 0); @@ -1315,6 +1324,7 @@ static void intel_ring_context_unpin(struct intel_engine_cs *engine, static int intel_init_ring_buffer(struct intel_engine_cs *engine) { struct intel_ring *ring; + struct i915_timeline *timeline; int err; intel_engine_setup_common(engine); @@ -1323,7 +1333,14 @@ static int intel_init_ring_buffer(struct intel_engine_cs *engine) if (err) goto err; - ring = intel_engine_create_ring(engine, 32 * PAGE_SIZE); + timeline = i915_timeline_create(engine->i915, engine->name); + if (IS_ERR(timeline)) { + err = PTR_ERR(timeline); + goto err; + } + + ring = intel_engine_create_ring(engine, timeline, 32 * PAGE_SIZE); + i915_timeline_put(timeline); if (IS_ERR(ring)) { err = PTR_ERR(ring); goto err; @@ -1424,7 +1441,7 @@ static inline int mi_set_context(struct i915_request *rq, u32 flags) *cs++ = MI_NOOP; *cs++ = MI_SET_CONTEXT; - *cs++ = i915_ggtt_offset(rq->ctx->engine[RCS].state) | flags; + *cs++ = i915_ggtt_offset(to_intel_context(rq->ctx, engine)->state) | flags; /* * w/a: MI_SET_CONTEXT must always be followed by MI_NOOP * WaMiSetContext_Hang:snb,ivb,vlv @@ -1515,7 +1532,7 @@ static int switch_context(struct i915_request *rq) hw_flags = MI_FORCE_RESTORE; } - if (to_ctx->engine[engine->id].state && + if (to_intel_context(to_ctx, engine)->state && (to_ctx != from_ctx || hw_flags & MI_FORCE_RESTORE)) { GEM_BUG_ON(engine->id != RCS); @@ -1563,7 +1580,7 @@ static int ring_request_alloc(struct i915_request *request) { int ret; - GEM_BUG_ON(!request->ctx->engine[request->engine->id].pin_count); + GEM_BUG_ON(!to_intel_context(request->ctx, request->engine)->pin_count); /* Flush enough space to reduce the likelihood of waiting after * we start building the request - in which case we will just @@ -1593,6 +1610,7 @@ static noinline int wait_for_space(struct intel_ring *ring, unsigned int bytes) if (intel_ring_update_space(ring) >= bytes) return 0; + GEM_BUG_ON(list_empty(&ring->request_list)); list_for_each_entry(target, &ring->request_list, ring_link) { /* Would completion of this request free enough space? */ if (bytes <= __intel_ring_space(target->postfix, @@ -1692,17 +1710,18 @@ u32 *intel_ring_begin(struct i915_request *rq, unsigned int num_dwords) need_wrap &= ~1; GEM_BUG_ON(need_wrap > ring->space); GEM_BUG_ON(ring->emit + need_wrap > ring->size); + GEM_BUG_ON(!IS_ALIGNED(need_wrap, sizeof(u64))); /* Fill the tail with MI_NOOP */ - memset(ring->vaddr + ring->emit, 0, need_wrap); - ring->emit = 0; + memset64(ring->vaddr + ring->emit, 0, need_wrap / sizeof(u64)); ring->space -= need_wrap; + ring->emit = 0; } GEM_BUG_ON(ring->emit > ring->size - bytes); GEM_BUG_ON(ring->space < bytes); cs = ring->vaddr + ring->emit; - GEM_DEBUG_EXEC(memset(cs, POISON_INUSE, bytes)); + GEM_DEBUG_EXEC(memset32(cs, POISON_INUSE, bytes / sizeof(*cs))); ring->emit += bytes; ring->space -= bytes; @@ -1712,22 +1731,24 @@ u32 *intel_ring_begin(struct i915_request *rq, unsigned int num_dwords) /* Align the ring tail to a cacheline boundary */ int intel_ring_cacheline_align(struct i915_request *rq) { - int num_dwords = (rq->ring->emit & (CACHELINE_BYTES - 1)) / sizeof(u32); - u32 *cs; + int num_dwords; + void *cs; + num_dwords = (rq->ring->emit & (CACHELINE_BYTES - 1)) / sizeof(u32); if (num_dwords == 0) return 0; - num_dwords = CACHELINE_BYTES / sizeof(u32) - num_dwords; + num_dwords = CACHELINE_DWORDS - num_dwords; + GEM_BUG_ON(num_dwords & 1); + cs = intel_ring_begin(rq, num_dwords); if (IS_ERR(cs)) return PTR_ERR(cs); - while (num_dwords--) - *cs++ = MI_NOOP; - + memset64(cs, (u64)MI_NOOP << 32 | MI_NOOP, num_dwords / 2); intel_ring_advance(rq, cs); + GEM_BUG_ON(rq->ring->emit & (CACHELINE_BYTES - 1)); return 0; } @@ -1943,8 +1964,6 @@ static void intel_ring_init_semaphores(struct drm_i915_private *dev_priv, static void intel_ring_init_irq(struct drm_i915_private *dev_priv, struct intel_engine_cs *engine) { - engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT << engine->irq_shift; - if (INTEL_GEN(dev_priv) >= 6) { engine->irq_enable = gen6_irq_enable; engine->irq_disable = gen6_irq_disable; @@ -2029,6 +2048,8 @@ int intel_init_render_ring_buffer(struct intel_engine_cs *engine) if (HAS_L3_DPF(dev_priv)) engine->irq_keep_mask = GT_RENDER_L3_PARITY_ERROR_INTERRUPT; + engine->irq_enable_mask = GT_RENDER_USER_INTERRUPT; + if (INTEL_GEN(dev_priv) >= 6) { engine->init_context = intel_rcs_ctx_init; engine->emit_flush = gen7_render_ring_flush; @@ -2079,7 +2100,6 @@ int intel_init_bsd_ring_buffer(struct intel_engine_cs *engine) engine->emit_flush = gen6_bsd_ring_flush; engine->irq_enable_mask = GT_BSD_USER_INTERRUPT; } else { - engine->mmio_base = BSD_RING_BASE; engine->emit_flush = bsd_ring_flush; if (IS_GEN5(dev_priv)) engine->irq_enable_mask = ILK_BSD_USER_INTERRUPT; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 0320c2c..010750e 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -3,15 +3,19 @@ #define _INTEL_RINGBUFFER_H_ #include <linux/hashtable.h> +#include <linux/seqlock.h> #include "i915_gem_batch_pool.h" -#include "i915_gem_timeline.h" +#include "i915_reg.h" #include "i915_pmu.h" #include "i915_request.h" #include "i915_selftest.h" +#include "i915_timeline.h" +#include "intel_gpu_commands.h" struct drm_printer; +struct i915_sched_attr; #define I915_CMD_HASH_ORDER 9 @@ -84,7 +88,7 @@ hangcheck_action_to_str(const enum intel_engine_hangcheck_action a) } #define I915_MAX_SLICES 3 -#define I915_MAX_SUBSLICES 3 +#define I915_MAX_SUBSLICES 8 #define instdone_slice_mask(dev_priv__) \ (INTEL_GEN(dev_priv__) == 7 ? \ @@ -125,7 +129,9 @@ struct intel_ring { struct i915_vma *vma; void *vaddr; + struct i915_timeline *timeline; struct list_head request_list; + struct list_head active_link; u32 head; u32 tail; @@ -330,10 +336,10 @@ struct intel_engine_cs { u8 instance; u32 context_size; u32 mmio_base; - unsigned int irq_shift; struct intel_ring *buffer; - struct intel_timeline *timeline; + + struct i915_timeline timeline; struct drm_i915_gem_object *default_state; @@ -459,7 +465,8 @@ struct intel_engine_cs { * * Called under the struct_mutex. */ - void (*schedule)(struct i915_request *request, int priority); + void (*schedule)(struct i915_request *request, + const struct i915_sched_attr *attr); /* * Cancel all requests on the hardware, or queued for execution. @@ -561,6 +568,7 @@ struct intel_engine_cs { #define I915_ENGINE_NEEDS_CMD_PARSER BIT(0) #define I915_ENGINE_SUPPORTS_STATS BIT(1) +#define I915_ENGINE_HAS_PREEMPTION BIT(2) unsigned int flags; /* @@ -591,7 +599,7 @@ struct intel_engine_cs { /** * @lock: Lock protecting the below fields. */ - spinlock_t lock; + seqlock_t lock; /** * @enabled: Reference count indicating number of listeners. */ @@ -620,16 +628,29 @@ struct intel_engine_cs { } stats; }; -static inline bool intel_engine_needs_cmd_parser(struct intel_engine_cs *engine) +static inline bool +intel_engine_needs_cmd_parser(const struct intel_engine_cs *engine) { return engine->flags & I915_ENGINE_NEEDS_CMD_PARSER; } -static inline bool intel_engine_supports_stats(struct intel_engine_cs *engine) +static inline bool +intel_engine_supports_stats(const struct intel_engine_cs *engine) { return engine->flags & I915_ENGINE_SUPPORTS_STATS; } +static inline bool +intel_engine_has_preemption(const struct intel_engine_cs *engine) +{ + return engine->flags & I915_ENGINE_HAS_PREEMPTION; +} + +static inline bool __execlists_need_preempt(int prio, int last) +{ + return prio > max(0, last); +} + static inline void execlists_set_active(struct intel_engine_execlists *execlists, unsigned int bit) @@ -637,6 +658,13 @@ execlists_set_active(struct intel_engine_execlists *execlists, __set_bit(bit, (unsigned long *)&execlists->active); } +static inline bool +execlists_set_active_once(struct intel_engine_execlists *execlists, + unsigned int bit) +{ + return !__test_and_set_bit(bit, (unsigned long *)&execlists->active); +} + static inline void execlists_clear_active(struct intel_engine_execlists *execlists, unsigned int bit) @@ -651,6 +679,10 @@ execlists_is_active(const struct intel_engine_execlists *execlists, return test_bit(bit, (unsigned long *)&execlists->active); } +void execlists_user_begin(struct intel_engine_execlists *execlists, + const struct execlist_port *port); +void execlists_user_end(struct intel_engine_execlists *execlists); + void execlists_cancel_port_requests(struct intel_engine_execlists * const execlists); @@ -663,7 +695,7 @@ execlists_num_ports(const struct intel_engine_execlists * const execlists) return execlists->port_mask + 1; } -static inline void +static inline struct execlist_port * execlists_port_complete(struct intel_engine_execlists * const execlists, struct execlist_port * const port) { @@ -674,6 +706,8 @@ execlists_port_complete(struct intel_engine_execlists * const execlists, memmove(port, port + 1, m * sizeof(struct execlist_port)); memset(port + m, 0, sizeof(struct execlist_port)); + + return port; } static inline unsigned int @@ -736,7 +770,9 @@ intel_write_status_page(struct intel_engine_cs *engine, int reg, u32 value) #define CNL_HWS_CSB_WRITE_INDEX 0x2f struct intel_ring * -intel_engine_create_ring(struct intel_engine_cs *engine, int size); +intel_engine_create_ring(struct intel_engine_cs *engine, + struct i915_timeline *timeline, + int size); int intel_ring_pin(struct intel_ring *ring, struct drm_i915_private *i915, unsigned int offset_bias); @@ -854,12 +890,9 @@ static inline u32 intel_engine_last_submit(struct intel_engine_cs *engine) * wtih serialising this hint with anything, so document it as * a hint and nothing more. */ - return READ_ONCE(engine->timeline->seqno); + return READ_ONCE(engine->timeline.seqno); } -int init_workarounds_ring(struct intel_engine_cs *engine); -int intel_ring_workarounds_emit(struct i915_request *rq); - void intel_engine_get_instdone(struct intel_engine_cs *engine, struct intel_instdone *instdone); @@ -939,7 +972,7 @@ bool intel_engine_add_wait(struct intel_engine_cs *engine, struct intel_wait *wait); void intel_engine_remove_wait(struct intel_engine_cs *engine, struct intel_wait *wait); -void intel_engine_enable_signaling(struct i915_request *request, bool wakeup); +bool intel_engine_enable_signaling(struct i915_request *request, bool wakeup); void intel_engine_cancel_signaling(struct i915_request *request); static inline bool intel_engine_has_waiter(const struct intel_engine_cs *engine) @@ -1037,7 +1070,7 @@ static inline void intel_engine_context_in(struct intel_engine_cs *engine) if (READ_ONCE(engine->stats.enabled) == 0) return; - spin_lock_irqsave(&engine->stats.lock, flags); + write_seqlock_irqsave(&engine->stats.lock, flags); if (engine->stats.enabled > 0) { if (engine->stats.active++ == 0) @@ -1045,7 +1078,7 @@ static inline void intel_engine_context_in(struct intel_engine_cs *engine) GEM_BUG_ON(engine->stats.active == 0); } - spin_unlock_irqrestore(&engine->stats.lock, flags); + write_sequnlock_irqrestore(&engine->stats.lock, flags); } static inline void intel_engine_context_out(struct intel_engine_cs *engine) @@ -1055,7 +1088,7 @@ static inline void intel_engine_context_out(struct intel_engine_cs *engine) if (READ_ONCE(engine->stats.enabled) == 0) return; - spin_lock_irqsave(&engine->stats.lock, flags); + write_seqlock_irqsave(&engine->stats.lock, flags); if (engine->stats.enabled > 0) { ktime_t last; @@ -1082,7 +1115,7 @@ static inline void intel_engine_context_out(struct intel_engine_cs *engine) } } - spin_unlock_irqrestore(&engine->stats.lock, flags); + write_sequnlock_irqrestore(&engine->stats.lock, flags); } int intel_enable_engine_stats(struct intel_engine_cs *engine); diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c index 66de4b2..53a6eaa 100644 --- a/drivers/gpu/drm/i915/intel_runtime_pm.c +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c @@ -542,6 +542,29 @@ void gen9_sanitize_dc_state(struct drm_i915_private *dev_priv) dev_priv->csr.dc_state = val; } +/** + * gen9_set_dc_state - set target display C power state + * @dev_priv: i915 device instance + * @state: target DC power state + * - DC_STATE_DISABLE + * - DC_STATE_EN_UPTO_DC5 + * - DC_STATE_EN_UPTO_DC6 + * - DC_STATE_EN_DC9 + * + * Signal to DMC firmware/HW the target DC power state passed in @state. + * DMC/HW can turn off individual display clocks and power rails when entering + * a deeper DC power state (higher in number) and turns these back when exiting + * that state to a shallower power state (lower in number). The HW will decide + * when to actually enter a given state on an on-demand basis, for instance + * depending on the active state of display pipes. The state of display + * registers backed by affected power rails are saved/restored as needed. + * + * Based on the above enabling a deeper DC power state is asynchronous wrt. + * enabling it. Disabling a deeper power state is synchronous: for instance + * setting %DC_STATE_DISABLE won't complete until all HW resources are turned + * back on and register state is restored. This is guaranteed by the MMIO write + * to DC_STATE_EN blocking until the state is restored. + */ static void gen9_set_dc_state(struct drm_i915_private *dev_priv, uint32_t state) { uint32_t val; @@ -635,7 +658,7 @@ static void assert_can_enable_dc6(struct drm_i915_private *dev_priv) assert_csr_loaded(dev_priv); } -void skl_enable_dc6(struct drm_i915_private *dev_priv) +static void skl_enable_dc6(struct drm_i915_private *dev_priv) { assert_can_enable_dc6(dev_priv); @@ -649,13 +672,6 @@ void skl_enable_dc6(struct drm_i915_private *dev_priv) gen9_set_dc_state(dev_priv, DC_STATE_EN_UPTO_DC6); } -void skl_disable_dc6(struct drm_i915_private *dev_priv) -{ - DRM_DEBUG_KMS("Disabling DC6\n"); - - gen9_set_dc_state(dev_priv, DC_STATE_DISABLE); -} - static void hsw_power_well_sync_hw(struct drm_i915_private *dev_priv, struct i915_power_well *power_well) { @@ -2626,32 +2642,69 @@ static void intel_power_domains_sync_hw(struct drm_i915_private *dev_priv) mutex_unlock(&power_domains->lock); } -static void gen9_dbuf_enable(struct drm_i915_private *dev_priv) +static inline +bool intel_dbuf_slice_set(struct drm_i915_private *dev_priv, + i915_reg_t reg, bool enable) { - I915_WRITE(DBUF_CTL, I915_READ(DBUF_CTL) | DBUF_POWER_REQUEST); - POSTING_READ(DBUF_CTL); + u32 val, status; + val = I915_READ(reg); + val = enable ? (val | DBUF_POWER_REQUEST) : (val & ~DBUF_POWER_REQUEST); + I915_WRITE(reg, val); + POSTING_READ(reg); udelay(10); - if (!(I915_READ(DBUF_CTL) & DBUF_POWER_STATE)) - DRM_ERROR("DBuf power enable timeout\n"); + status = I915_READ(reg) & DBUF_POWER_STATE; + if ((enable && !status) || (!enable && status)) { + DRM_ERROR("DBus power %s timeout!\n", + enable ? "enable" : "disable"); + return false; + } + return true; +} + +static void gen9_dbuf_enable(struct drm_i915_private *dev_priv) +{ + intel_dbuf_slice_set(dev_priv, DBUF_CTL, true); } static void gen9_dbuf_disable(struct drm_i915_private *dev_priv) { - I915_WRITE(DBUF_CTL, I915_READ(DBUF_CTL) & ~DBUF_POWER_REQUEST); - POSTING_READ(DBUF_CTL); + intel_dbuf_slice_set(dev_priv, DBUF_CTL, false); +} - udelay(10); +static u8 intel_dbuf_max_slices(struct drm_i915_private *dev_priv) +{ + if (INTEL_GEN(dev_priv) < 11) + return 1; + return 2; +} - if (I915_READ(DBUF_CTL) & DBUF_POWER_STATE) - DRM_ERROR("DBuf power disable timeout!\n"); +void icl_dbuf_slices_update(struct drm_i915_private *dev_priv, + u8 req_slices) +{ + u8 hw_enabled_slices = dev_priv->wm.skl_hw.ddb.enabled_slices; + u32 val; + bool ret; + + if (req_slices > intel_dbuf_max_slices(dev_priv)) { + DRM_ERROR("Invalid number of dbuf slices requested\n"); + return; + } + + if (req_slices == hw_enabled_slices || req_slices == 0) + return; + + val = I915_READ(DBUF_CTL_S2); + if (req_slices > hw_enabled_slices) + ret = intel_dbuf_slice_set(dev_priv, DBUF_CTL_S2, true); + else + ret = intel_dbuf_slice_set(dev_priv, DBUF_CTL_S2, false); + + if (ret) + dev_priv->wm.skl_hw.ddb.enabled_slices = req_slices; } -/* - * TODO: we shouldn't always enable DBUF_CTL_S2, we should only enable it when - * needed and keep it disabled as much as possible. - */ static void icl_dbuf_enable(struct drm_i915_private *dev_priv) { I915_WRITE(DBUF_CTL_S1, I915_READ(DBUF_CTL_S1) | DBUF_POWER_REQUEST); @@ -2663,6 +2716,8 @@ static void icl_dbuf_enable(struct drm_i915_private *dev_priv) if (!(I915_READ(DBUF_CTL_S1) & DBUF_POWER_STATE) || !(I915_READ(DBUF_CTL_S2) & DBUF_POWER_STATE)) DRM_ERROR("DBuf power enable timeout\n"); + else + dev_priv->wm.skl_hw.ddb.enabled_slices = 2; } static void icl_dbuf_disable(struct drm_i915_private *dev_priv) @@ -2676,6 +2731,8 @@ static void icl_dbuf_disable(struct drm_i915_private *dev_priv) if ((I915_READ(DBUF_CTL_S1) & DBUF_POWER_STATE) || (I915_READ(DBUF_CTL_S2) & DBUF_POWER_STATE)) DRM_ERROR("DBuf power disable timeout!\n"); + else + dev_priv->wm.skl_hw.ddb.enabled_slices = 0; } static void icl_mbus_init(struct drm_i915_private *dev_priv) diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c index 96e213e..2500502 100644 --- a/drivers/gpu/drm/i915/intel_sdvo.c +++ b/drivers/gpu/drm/i915/intel_sdvo.c @@ -2779,9 +2779,8 @@ static bool intel_sdvo_tv_create_property(struct intel_sdvo *intel_sdvo, return false; for (i = 0; i < intel_sdvo_connector->format_supported_num; i++) - drm_property_add_enum( - intel_sdvo_connector->tv_format, i, - i, tv_format_names[intel_sdvo_connector->tv_format_supported[i]]); + drm_property_add_enum(intel_sdvo_connector->tv_format, i, + tv_format_names[intel_sdvo_connector->tv_format_supported[i]]); intel_sdvo_connector->base.base.state->tv.mode = intel_sdvo_connector->tv_format_supported[0]; drm_object_attach_property(&intel_sdvo_connector->base.base.base, diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c index dbdcf85..ee23613f 100644 --- a/drivers/gpu/drm/i915/intel_sprite.c +++ b/drivers/gpu/drm/i915/intel_sprite.c @@ -48,6 +48,7 @@ bool intel_format_is_yuv(u32 format) case DRM_FORMAT_UYVY: case DRM_FORMAT_VYUY: case DRM_FORMAT_YVYU: + case DRM_FORMAT_NV12: return true; default: return false; @@ -130,7 +131,7 @@ void intel_pipe_update_start(const struct intel_crtc_state *new_crtc_state) if (scanline < min || scanline > max) break; - if (timeout <= 0) { + if (!timeout) { DRM_ERROR("Potential atomic update failure on pipe %c\n", pipe_name(crtc->pipe)); break; @@ -935,20 +936,11 @@ intel_check_sprite_plane(struct intel_plane *plane, struct drm_i915_private *dev_priv = to_i915(plane->base.dev); struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); struct drm_framebuffer *fb = state->base.fb; - int crtc_x, crtc_y; - unsigned int crtc_w, crtc_h; - uint32_t src_x, src_y, src_w, src_h; - struct drm_rect *src = &state->base.src; - struct drm_rect *dst = &state->base.dst; - struct drm_rect clip = {}; int max_stride = INTEL_GEN(dev_priv) >= 9 ? 32768 : 16384; - int hscale, vscale; int max_scale, min_scale; bool can_scale; int ret; - - *src = drm_plane_state_src(&state->base); - *dst = drm_plane_state_dest(&state->base); + uint32_t pixel_format = 0; if (!fb) { state->base.visible = false; @@ -969,11 +961,14 @@ intel_check_sprite_plane(struct intel_plane *plane, /* setup can_scale, min_scale, max_scale */ if (INTEL_GEN(dev_priv) >= 9) { + if (state->base.fb) + pixel_format = state->base.fb->format->format; /* use scaler when colorkey is not required */ if (!state->ckey.flags) { can_scale = 1; min_scale = 1; - max_scale = skl_max_scale(crtc, crtc_state); + max_scale = + skl_max_scale(crtc, crtc_state, pixel_format); } else { can_scale = 0; min_scale = DRM_PLANE_HELPER_NO_SCALING; @@ -985,64 +980,19 @@ intel_check_sprite_plane(struct intel_plane *plane, min_scale = plane->can_scale ? 1 : (1 << 16); } - /* - * FIXME the following code does a bunch of fuzzy adjustments to the - * coordinates and sizes. We probably need some way to decide whether - * more strict checking should be done instead. - */ - drm_rect_rotate(src, fb->width << 16, fb->height << 16, - state->base.rotation); - - hscale = drm_rect_calc_hscale_relaxed(src, dst, min_scale, max_scale); - BUG_ON(hscale < 0); - - vscale = drm_rect_calc_vscale_relaxed(src, dst, min_scale, max_scale); - BUG_ON(vscale < 0); - - if (crtc_state->base.enable) - drm_mode_get_hv_timing(&crtc_state->base.mode, - &clip.x2, &clip.y2); - - state->base.visible = drm_rect_clip_scaled(src, dst, &clip, hscale, vscale); - - crtc_x = dst->x1; - crtc_y = dst->y1; - crtc_w = drm_rect_width(dst); - crtc_h = drm_rect_height(dst); + ret = drm_atomic_helper_check_plane_state(&state->base, + &crtc_state->base, + min_scale, max_scale, + true, true); + if (ret) + return ret; if (state->base.visible) { - /* check again in case clipping clamped the results */ - hscale = drm_rect_calc_hscale(src, dst, min_scale, max_scale); - if (hscale < 0) { - DRM_DEBUG_KMS("Horizontal scaling factor out of limits\n"); - drm_rect_debug_print("src: ", src, true); - drm_rect_debug_print("dst: ", dst, false); - - return hscale; - } - - vscale = drm_rect_calc_vscale(src, dst, min_scale, max_scale); - if (vscale < 0) { - DRM_DEBUG_KMS("Vertical scaling factor out of limits\n"); - drm_rect_debug_print("src: ", src, true); - drm_rect_debug_print("dst: ", dst, false); - - return vscale; - } - - /* Make the source viewport size an exact multiple of the scaling factors. */ - drm_rect_adjust_size(src, - drm_rect_width(dst) * hscale - drm_rect_width(src), - drm_rect_height(dst) * vscale - drm_rect_height(src)); - - drm_rect_rotate_inv(src, fb->width << 16, fb->height << 16, - state->base.rotation); - - /* sanity check to make sure the src viewport wasn't enlarged */ - WARN_ON(src->x1 < (int) state->base.src_x || - src->y1 < (int) state->base.src_y || - src->x2 > (int) state->base.src_x + state->base.src_w || - src->y2 > (int) state->base.src_y + state->base.src_h); + struct drm_rect *src = &state->base.src; + struct drm_rect *dst = &state->base.dst; + unsigned int crtc_w = drm_rect_width(dst); + unsigned int crtc_h = drm_rect_height(dst); + uint32_t src_x, src_y, src_w, src_h; /* * Hardware doesn't handle subpixel coordinates. @@ -1055,58 +1005,40 @@ intel_check_sprite_plane(struct intel_plane *plane, src_y = src->y1 >> 16; src_h = drm_rect_height(src) >> 16; - if (intel_format_is_yuv(fb->format->format)) { - src_x &= ~1; - src_w &= ~1; - - /* - * Must keep src and dst the - * same if we can't scale. - */ - if (!can_scale) - crtc_w &= ~1; + src->x1 = src_x << 16; + src->x2 = (src_x + src_w) << 16; + src->y1 = src_y << 16; + src->y2 = (src_y + src_h) << 16; - if (crtc_w == 0) - state->base.visible = false; + if (intel_format_is_yuv(fb->format->format) && + fb->format->format != DRM_FORMAT_NV12 && + (src_x % 2 || src_w % 2)) { + DRM_DEBUG_KMS("src x/w (%u, %u) must be a multiple of 2 for YUV planes\n", + src_x, src_w); + return -EINVAL; } - } - - /* Check size restrictions when scaling */ - if (state->base.visible && (src_w != crtc_w || src_h != crtc_h)) { - unsigned int width_bytes; - int cpp = fb->format->cpp[0]; - WARN_ON(!can_scale); + /* Check size restrictions when scaling */ + if (src_w != crtc_w || src_h != crtc_h) { + unsigned int width_bytes; + int cpp = fb->format->cpp[0]; - /* FIXME interlacing min height is 6 */ + WARN_ON(!can_scale); - if (crtc_w < 3 || crtc_h < 3) - state->base.visible = false; + width_bytes = ((src_x * cpp) & 63) + src_w * cpp; - if (src_w < 3 || src_h < 3) - state->base.visible = false; - - width_bytes = ((src_x * cpp) & 63) + src_w * cpp; - - if (INTEL_GEN(dev_priv) < 9 && (src_w > 2048 || src_h > 2048 || - width_bytes > 4096 || fb->pitches[0] > 4096)) { - DRM_DEBUG_KMS("Source dimensions exceed hardware limits\n"); - return -EINVAL; + /* FIXME interlacing min height is 6 */ + if (INTEL_GEN(dev_priv) < 9 && ( + src_w < 3 || src_h < 3 || + src_w > 2048 || src_h > 2048 || + crtc_w < 3 || crtc_h < 3 || + width_bytes > 4096 || fb->pitches[0] > 4096)) { + DRM_DEBUG_KMS("Source dimensions exceed hardware limits\n"); + return -EINVAL; + } } } - if (state->base.visible) { - src->x1 = src_x << 16; - src->x2 = (src_x + src_w) << 16; - src->y1 = src_y << 16; - src->y2 = (src_y + src_h) << 16; - } - - dst->x1 = crtc_x; - dst->x2 = crtc_x + crtc_w; - dst->y1 = crtc_y; - dst->y2 = crtc_y + crtc_h; - if (INTEL_GEN(dev_priv) >= 9) { ret = skl_check_plane_surface(crtc_state, state); if (ret) @@ -1248,6 +1180,19 @@ static uint32_t skl_plane_formats[] = { DRM_FORMAT_VYUY, }; +static uint32_t skl_planar_formats[] = { + DRM_FORMAT_RGB565, + DRM_FORMAT_ABGR8888, + DRM_FORMAT_ARGB8888, + DRM_FORMAT_XBGR8888, + DRM_FORMAT_XRGB8888, + DRM_FORMAT_YUYV, + DRM_FORMAT_YVYU, + DRM_FORMAT_UYVY, + DRM_FORMAT_VYUY, + DRM_FORMAT_NV12, +}; + static const uint64_t skl_plane_format_modifiers_noccs[] = { I915_FORMAT_MOD_Yf_TILED, I915_FORMAT_MOD_Y_TILED, @@ -1342,6 +1287,7 @@ static bool skl_mod_supported(uint32_t format, uint64_t modifier) case DRM_FORMAT_YVYU: case DRM_FORMAT_UYVY: case DRM_FORMAT_VYUY: + case DRM_FORMAT_NV12: if (modifier == I915_FORMAT_MOD_Yf_TILED) return true; /* fall through */ @@ -1441,8 +1387,14 @@ intel_sprite_plane_create(struct drm_i915_private *dev_priv, intel_plane->disable_plane = skl_disable_plane; intel_plane->get_hw_state = skl_plane_get_hw_state; - plane_formats = skl_plane_formats; - num_plane_formats = ARRAY_SIZE(skl_plane_formats); + if (skl_plane_has_planar(dev_priv, pipe, + PLANE_SPRITE0 + plane)) { + plane_formats = skl_planar_formats; + num_plane_formats = ARRAY_SIZE(skl_planar_formats); + } else { + plane_formats = skl_plane_formats; + num_plane_formats = ARRAY_SIZE(skl_plane_formats); + } if (skl_plane_has_ccs(dev_priv, pipe, PLANE_SPRITE0 + plane)) modifiers = skl_plane_format_modifiers_ccs; diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index e5bf0d3..1cffaf7 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -69,13 +69,15 @@ static int __get_platform_enable_guc(struct drm_i915_private *dev_priv) static int __get_default_guc_log_level(struct drm_i915_private *dev_priv) { - int guc_log_level = 0; /* disabled */ + int guc_log_level; - /* Enable if we're running on platform with GuC and debug config */ - if (HAS_GUC(dev_priv) && intel_uc_is_using_guc() && - (IS_ENABLED(CONFIG_DRM_I915_DEBUG) || - IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))) - guc_log_level = 1 + GUC_LOG_VERBOSITY_MAX; + if (!HAS_GUC(dev_priv) || !intel_uc_is_using_guc()) + guc_log_level = GUC_LOG_LEVEL_DISABLED; + else if (IS_ENABLED(CONFIG_DRM_I915_DEBUG) || + IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)) + guc_log_level = GUC_LOG_LEVEL_MAX; + else + guc_log_level = GUC_LOG_LEVEL_NON_VERBOSE; /* Any platform specific fine-tuning can be done here */ @@ -83,7 +85,7 @@ static int __get_default_guc_log_level(struct drm_i915_private *dev_priv) } /** - * intel_uc_sanitize_options - sanitize uC related modparam options + * sanitize_options_early - sanitize uC related modparam options * @dev_priv: device private * * In case of "enable_guc" option this function will attempt to modify @@ -99,7 +101,7 @@ static int __get_default_guc_log_level(struct drm_i915_private *dev_priv) * unless GuC is enabled on given platform and the driver is compiled with * debug config when this modparam will default to "enable(1..4)". */ -void intel_uc_sanitize_options(struct drm_i915_private *dev_priv) +static void sanitize_options_early(struct drm_i915_private *dev_priv) { struct intel_uc_fw *guc_fw = &dev_priv->guc.fw; struct intel_uc_fw *huc_fw = &dev_priv->huc.fw; @@ -142,51 +144,53 @@ void intel_uc_sanitize_options(struct drm_i915_private *dev_priv) i915_modparams.guc_log_level = 0; } - if (i915_modparams.guc_log_level > 1 + GUC_LOG_VERBOSITY_MAX) { + if (i915_modparams.guc_log_level > GUC_LOG_LEVEL_MAX) { DRM_WARN("Incompatible option detected: %s=%d, %s!\n", "guc_log_level", i915_modparams.guc_log_level, "verbosity too high"); - i915_modparams.guc_log_level = 1 + GUC_LOG_VERBOSITY_MAX; + i915_modparams.guc_log_level = GUC_LOG_LEVEL_MAX; } - DRM_DEBUG_DRIVER("guc_log_level=%d (enabled:%s verbosity:%d)\n", + DRM_DEBUG_DRIVER("guc_log_level=%d (enabled:%s, verbose:%s, verbosity:%d)\n", i915_modparams.guc_log_level, yesno(i915_modparams.guc_log_level), - i915_modparams.guc_log_level - 1); + yesno(GUC_LOG_LEVEL_IS_VERBOSE(i915_modparams.guc_log_level)), + GUC_LOG_LEVEL_TO_VERBOSITY(i915_modparams.guc_log_level)); /* Make sure that sanitization was done */ GEM_BUG_ON(i915_modparams.enable_guc < 0); GEM_BUG_ON(i915_modparams.guc_log_level < 0); } -void intel_uc_init_early(struct drm_i915_private *dev_priv) +void intel_uc_init_early(struct drm_i915_private *i915) { - intel_guc_init_early(&dev_priv->guc); - intel_huc_init_early(&dev_priv->huc); -} + struct intel_guc *guc = &i915->guc; + struct intel_huc *huc = &i915->huc; -void intel_uc_init_fw(struct drm_i915_private *dev_priv) -{ - if (!USES_GUC(dev_priv)) - return; + intel_guc_init_early(guc); + intel_huc_init_early(huc); - if (USES_HUC(dev_priv)) - intel_uc_fw_fetch(dev_priv, &dev_priv->huc.fw); + sanitize_options_early(i915); - intel_uc_fw_fetch(dev_priv, &dev_priv->guc.fw); + if (USES_GUC(i915)) + intel_uc_fw_fetch(i915, &guc->fw); + + if (USES_HUC(i915)) + intel_uc_fw_fetch(i915, &huc->fw); } -void intel_uc_fini_fw(struct drm_i915_private *dev_priv) +void intel_uc_cleanup_early(struct drm_i915_private *i915) { - if (!USES_GUC(dev_priv)) - return; + struct intel_guc *guc = &i915->guc; + struct intel_huc *huc = &i915->huc; - intel_uc_fw_fini(&dev_priv->guc.fw); + if (USES_HUC(i915)) + intel_uc_fw_fini(&huc->fw); - if (USES_HUC(dev_priv)) - intel_uc_fw_fini(&dev_priv->huc.fw); + if (USES_GUC(i915)) + intel_uc_fw_fini(&guc->fw); - guc_free_load_err_log(&dev_priv->guc); + guc_free_load_err_log(guc); } /** @@ -223,10 +227,13 @@ static int guc_enable_communication(struct intel_guc *guc) { struct drm_i915_private *dev_priv = guc_to_i915(guc); + gen9_enable_guc_interrupts(dev_priv); + if (HAS_GUC_CT(dev_priv)) - return intel_guc_enable_ct(guc); + return intel_guc_ct_enable(&guc->ct); guc->send = intel_guc_send_mmio; + guc->handler = intel_guc_to_host_event_handler_mmio; return 0; } @@ -235,9 +242,12 @@ static void guc_disable_communication(struct intel_guc *guc) struct drm_i915_private *dev_priv = guc_to_i915(guc); if (HAS_GUC_CT(dev_priv)) - intel_guc_disable_ct(guc); + intel_guc_ct_disable(&guc->ct); + + gen9_disable_guc_interrupts(dev_priv); guc->send = intel_guc_send_nop; + guc->handler = intel_guc_to_host_event_handler_nop; } int intel_uc_init_misc(struct drm_i915_private *dev_priv) @@ -248,24 +258,13 @@ int intel_uc_init_misc(struct drm_i915_private *dev_priv) if (!USES_GUC(dev_priv)) return 0; - ret = intel_guc_init_wq(guc); - if (ret) { - DRM_ERROR("Couldn't allocate workqueues for GuC\n"); - goto err; - } + intel_guc_init_ggtt_pin_bias(guc); - ret = intel_guc_log_relay_create(guc); - if (ret) { - DRM_ERROR("Couldn't allocate relay for GuC log\n"); - goto err_relay; - } + ret = intel_guc_init_wq(guc); + if (ret) + return ret; return 0; - -err_relay: - intel_guc_fini_wq(guc); -err: - return ret; } void intel_uc_fini_misc(struct drm_i915_private *dev_priv) @@ -276,8 +275,6 @@ void intel_uc_fini_misc(struct drm_i915_private *dev_priv) return; intel_guc_fini_wq(guc); - - intel_guc_log_relay_destroy(guc); } int intel_uc_init(struct drm_i915_private *dev_priv) @@ -325,6 +322,24 @@ void intel_uc_fini(struct drm_i915_private *dev_priv) intel_guc_fini(guc); } +void intel_uc_sanitize(struct drm_i915_private *i915) +{ + struct intel_guc *guc = &i915->guc; + struct intel_huc *huc = &i915->huc; + + if (!USES_GUC(i915)) + return; + + GEM_BUG_ON(!HAS_GUC(i915)); + + guc_disable_communication(guc); + + intel_huc_sanitize(huc); + intel_guc_sanitize(guc); + + __intel_uc_reset_hw(i915); +} + int intel_uc_init_hw(struct drm_i915_private *dev_priv) { struct intel_guc *guc = &dev_priv->guc; @@ -336,14 +351,8 @@ int intel_uc_init_hw(struct drm_i915_private *dev_priv) GEM_BUG_ON(!HAS_GUC(dev_priv)); - guc_disable_communication(guc); gen9_reset_guc_interrupts(dev_priv); - /* init WOPCM */ - 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); - /* WaEnableuKernelHeaderValidFix:skl */ /* WaEnableGuCBootHashCheckNotSet:skl,bxt,kbl */ if (IS_GEN9(dev_priv)) @@ -390,12 +399,9 @@ int intel_uc_init_hw(struct drm_i915_private *dev_priv) } if (USES_GUC_SUBMISSION(dev_priv)) { - if (i915_modparams.guc_log_level) - gen9_enable_guc_interrupts(dev_priv); - ret = intel_guc_submission_enable(guc); if (ret) - goto err_interrupts; + goto err_communication; } dev_info(dev_priv->drm.dev, "GuC firmware version %u.%u\n", @@ -410,8 +416,6 @@ int intel_uc_init_hw(struct drm_i915_private *dev_priv) /* * We've failed to load the firmware :( */ -err_interrupts: - gen9_disable_guc_interrupts(dev_priv); err_communication: guc_disable_communication(guc); err_log_capture: @@ -441,9 +445,6 @@ void intel_uc_fini_hw(struct drm_i915_private *dev_priv) intel_guc_submission_disable(guc); guc_disable_communication(guc); - - if (USES_GUC_SUBMISSION(dev_priv)) - gen9_disable_guc_interrupts(dev_priv); } int intel_uc_suspend(struct drm_i915_private *i915) @@ -479,8 +480,7 @@ int intel_uc_resume(struct drm_i915_private *i915) if (guc->fw.load_status != INTEL_UC_FIRMWARE_SUCCESS) return 0; - if (i915_modparams.guc_log_level) - gen9_enable_guc_interrupts(i915); + gen9_enable_guc_interrupts(i915); err = intel_guc_resume(guc); if (err) { diff --git a/drivers/gpu/drm/i915/intel_uc.h b/drivers/gpu/drm/i915/intel_uc.h index f76d51d..25d73ad 100644 --- a/drivers/gpu/drm/i915/intel_uc.h +++ b/drivers/gpu/drm/i915/intel_uc.h @@ -28,13 +28,12 @@ #include "intel_huc.h" #include "i915_params.h" -void intel_uc_sanitize_options(struct drm_i915_private *dev_priv); void intel_uc_init_early(struct drm_i915_private *dev_priv); +void intel_uc_cleanup_early(struct drm_i915_private *dev_priv); void intel_uc_init_mmio(struct drm_i915_private *dev_priv); -void intel_uc_init_fw(struct drm_i915_private *dev_priv); -void intel_uc_fini_fw(struct drm_i915_private *dev_priv); int intel_uc_init_misc(struct drm_i915_private *dev_priv); void intel_uc_fini_misc(struct drm_i915_private *dev_priv); +void intel_uc_sanitize(struct drm_i915_private *dev_priv); int intel_uc_init_hw(struct drm_i915_private *dev_priv); void intel_uc_fini_hw(struct drm_i915_private *dev_priv); int intel_uc_init(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/intel_uc_fw.c b/drivers/gpu/drm/i915/intel_uc_fw.c index 3ec0ce5..6e8e0b5 100644 --- a/drivers/gpu/drm/i915/intel_uc_fw.c +++ b/drivers/gpu/drm/i915/intel_uc_fw.c @@ -95,15 +95,6 @@ void intel_uc_fw_fetch(struct drm_i915_private *dev_priv, uc_fw->ucode_offset = uc_fw->header_offset + uc_fw->header_size; uc_fw->ucode_size = (css->size_dw - css->header_size_dw) * sizeof(u32); - /* Header and uCode will be loaded to WOPCM */ - size = uc_fw->header_size + uc_fw->ucode_size; - if (size > intel_guc_wopcm_size(dev_priv)) { - DRM_WARN("%s: Firmware is too large to fit in WOPCM\n", - intel_uc_fw_type_repr(uc_fw->type)); - err = -E2BIG; - goto fail; - } - /* now RSA */ if (css->key_size_dw != UOS_RSA_SCRATCH_COUNT) { DRM_WARN("%s: Mismatched firmware RSA key size (%u)\n", @@ -209,6 +200,7 @@ int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, struct i915_vma *vma)) { struct i915_vma *vma; + u32 ggtt_pin_bias; int err; DRM_DEBUG_DRIVER("%s fw load %s\n", @@ -230,8 +222,9 @@ int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, goto fail; } + ggtt_pin_bias = to_i915(uc_fw->obj->base.dev)->guc.ggtt_pin_bias; vma = i915_gem_object_ggtt_pin(uc_fw->obj, NULL, 0, 0, - PIN_OFFSET_BIAS | GUC_WOPCM_TOP); + PIN_OFFSET_BIAS | ggtt_pin_bias); if (IS_ERR(vma)) { err = PTR_ERR(vma); DRM_DEBUG_DRIVER("%s fw ggtt-pin err=%d\n", diff --git a/drivers/gpu/drm/i915/intel_uc_fw.h b/drivers/gpu/drm/i915/intel_uc_fw.h index d5fd460..87910aa 100644 --- a/drivers/gpu/drm/i915/intel_uc_fw.h +++ b/drivers/gpu/drm/i915/intel_uc_fw.h @@ -30,7 +30,7 @@ struct drm_i915_private; struct i915_vma; /* Home of GuC, HuC and DMC firmwares */ -#define INTEL_UC_FIRMWARE_URL "https://01.org/linuxgraphics/downloads/firmware" +#define INTEL_UC_FIRMWARE_URL "https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/i915" enum intel_uc_fw_status { INTEL_UC_FIRMWARE_FAIL = -1, @@ -115,6 +115,28 @@ static inline bool intel_uc_fw_is_selected(struct intel_uc_fw *uc_fw) return uc_fw->path != NULL; } +static inline void intel_uc_fw_sanitize(struct intel_uc_fw *uc_fw) +{ + if (uc_fw->load_status == INTEL_UC_FIRMWARE_SUCCESS) + uc_fw->load_status = INTEL_UC_FIRMWARE_PENDING; +} + +/** + * intel_uc_fw_get_upload_size() - Get size of firmware needed to be uploaded. + * @uc_fw: uC firmware. + * + * Get the size of the firmware and header that will be uploaded to WOPCM. + * + * Return: Upload firmware size, or zero on firmware fetch failure. + */ +static inline u32 intel_uc_fw_get_upload_size(struct intel_uc_fw *uc_fw) +{ + if (uc_fw->fetch_status != INTEL_UC_FIRMWARE_SUCCESS) + return 0; + + return uc_fw->header_size + uc_fw->ucode_size; +} + void intel_uc_fw_fetch(struct drm_i915_private *dev_priv, struct intel_uc_fw *uc_fw); int intel_uc_fw_upload(struct intel_uc_fw *uc_fw, diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c index 4df7c2ef..448293e 100644 --- a/drivers/gpu/drm/i915/intel_uncore.c +++ b/drivers/gpu/drm/i915/intel_uncore.c @@ -62,6 +62,11 @@ static inline void fw_domain_reset(struct drm_i915_private *i915, const struct intel_uncore_forcewake_domain *d) { + /* + * We don't really know if the powerwell for the forcewake domain we are + * trying to reset here does exist at this point (engines could be fused + * off in ICL+), so no waiting for acks + */ __raw_i915_write32(i915, d->reg_set, i915->uncore.fw_reset); } @@ -134,7 +139,9 @@ fw_domain_wait_ack_with_fallback(const struct drm_i915_private *i915, * in the hope that the original ack will be delivered along with * the fallback ack. * - * This workaround is described in HSDES #1604254524 + * This workaround is described in HSDES #1604254524 and it's known as: + * WaRsForcewakeAddDelayForAck:skl,bxt,kbl,glk,cfl,cnl,icl + * although the name is a bit misleading. */ pass = 1; @@ -1353,6 +1360,23 @@ static void fw_domain_init(struct drm_i915_private *dev_priv, fw_domain_reset(dev_priv, d); } +static void fw_domain_fini(struct drm_i915_private *dev_priv, + enum forcewake_domain_id domain_id) +{ + struct intel_uncore_forcewake_domain *d; + + if (WARN_ON(domain_id >= FW_DOMAIN_ID_COUNT)) + return; + + d = &dev_priv->uncore.fw_domain[domain_id]; + + WARN_ON(d->wake_count); + WARN_ON(hrtimer_cancel(&d->timer)); + memset(d, 0, sizeof(*d)); + + dev_priv->uncore.fw_domains &= ~BIT(domain_id); +} + static void intel_uncore_fw_domains_init(struct drm_i915_private *dev_priv) { if (INTEL_GEN(dev_priv) <= 5 || intel_vgpu_active(dev_priv)) @@ -1372,7 +1396,8 @@ static void intel_uncore_fw_domains_init(struct drm_i915_private *dev_priv) if (INTEL_GEN(dev_priv) >= 11) { int i; - dev_priv->uncore.funcs.force_wake_get = fw_domains_get; + dev_priv->uncore.funcs.force_wake_get = + fw_domains_get_with_fallback; dev_priv->uncore.funcs.force_wake_put = fw_domains_put; fw_domain_init(dev_priv, FW_DOMAIN_ID_RENDER, FORCEWAKE_RENDER_GEN9, @@ -1565,6 +1590,40 @@ void intel_uncore_init(struct drm_i915_private *dev_priv) &dev_priv->uncore.pmic_bus_access_nb); } +/* + * We might have detected that some engines are fused off after we initialized + * the forcewake domains. Prune them, to make sure they only reference existing + * engines. + */ +void intel_uncore_prune(struct drm_i915_private *dev_priv) +{ + if (INTEL_GEN(dev_priv) >= 11) { + enum forcewake_domains fw_domains = dev_priv->uncore.fw_domains; + enum forcewake_domain_id domain_id; + int i; + + for (i = 0; i < I915_MAX_VCS; i++) { + domain_id = FW_DOMAIN_ID_MEDIA_VDBOX0 + i; + + if (HAS_ENGINE(dev_priv, _VCS(i))) + continue; + + if (fw_domains & BIT(domain_id)) + fw_domain_fini(dev_priv, domain_id); + } + + for (i = 0; i < I915_MAX_VECS; i++) { + domain_id = FW_DOMAIN_ID_MEDIA_VEBOX0 + i; + + if (HAS_ENGINE(dev_priv, _VECS(i))) + continue; + + if (fw_domains & BIT(domain_id)) + fw_domain_fini(dev_priv, domain_id); + } + } +} + void intel_uncore_fini(struct drm_i915_private *dev_priv) { /* Paranoia: make sure we have disabled everything before we exit. */ @@ -1646,11 +1705,10 @@ static void gen3_stop_engine(struct intel_engine_cs *engine) const i915_reg_t mode = RING_MI_MODE(base); I915_WRITE_FW(mode, _MASKED_BIT_ENABLE(STOP_RING)); - if (intel_wait_for_register_fw(dev_priv, - mode, - MODE_IDLE, - MODE_IDLE, - 500)) + if (__intel_wait_for_register_fw(dev_priv, + mode, MODE_IDLE, MODE_IDLE, + 500, 0, + NULL)) DRM_DEBUG_DRIVER("%s: timed out on STOP_RING\n", engine->name); @@ -1804,9 +1862,10 @@ static int gen6_hw_domain_reset(struct drm_i915_private *dev_priv, __raw_i915_write32(dev_priv, GEN6_GDRST, hw_domain_mask); /* Wait for the device to ack the reset requests */ - err = intel_wait_for_register_fw(dev_priv, - GEN6_GDRST, hw_domain_mask, 0, - 500); + err = __intel_wait_for_register_fw(dev_priv, + GEN6_GDRST, hw_domain_mask, 0, + 500, 0, + NULL); if (err) DRM_DEBUG_DRIVER("Wait for 0x%08x engines reset failed\n", hw_domain_mask); @@ -1854,6 +1913,50 @@ static int gen6_reset_engines(struct drm_i915_private *dev_priv, } /** + * gen11_reset_engines - reset individual engines + * @dev_priv: i915 device + * @engine_mask: mask of intel_ring_flag() engines or ALL_ENGINES for full reset + * + * This function will reset the individual engines that are set in engine_mask. + * If you provide ALL_ENGINES as mask, full global domain reset will be issued. + * + * Note: It is responsibility of the caller to handle the difference between + * asking full domain reset versus reset for all available individual engines. + * + * Returns 0 on success, nonzero on error. + */ +static int gen11_reset_engines(struct drm_i915_private *dev_priv, + unsigned engine_mask) +{ + struct intel_engine_cs *engine; + const u32 hw_engine_mask[I915_NUM_ENGINES] = { + [RCS] = GEN11_GRDOM_RENDER, + [BCS] = GEN11_GRDOM_BLT, + [VCS] = GEN11_GRDOM_MEDIA, + [VCS2] = GEN11_GRDOM_MEDIA2, + [VCS3] = GEN11_GRDOM_MEDIA3, + [VCS4] = GEN11_GRDOM_MEDIA4, + [VECS] = GEN11_GRDOM_VECS, + [VECS2] = GEN11_GRDOM_VECS2, + }; + u32 hw_mask; + + BUILD_BUG_ON(VECS2 + 1 != I915_NUM_ENGINES); + + if (engine_mask == ALL_ENGINES) { + hw_mask = GEN11_GRDOM_FULL; + } else { + unsigned int tmp; + + hw_mask = 0; + for_each_engine_masked(engine, dev_priv, engine_mask, tmp) + hw_mask |= hw_engine_mask[engine->id]; + } + + return gen6_hw_domain_reset(dev_priv, hw_mask); +} + +/** * __intel_wait_for_register_fw - wait until register matches expected state * @dev_priv: the i915 device * @reg: the register to read @@ -1940,7 +2043,7 @@ int __intel_wait_for_register(struct drm_i915_private *dev_priv, u32 reg_value; int ret; - might_sleep(); + might_sleep_if(slow_timeout_ms); spin_lock_irq(&dev_priv->uncore.lock); intel_uncore_forcewake_get__locked(dev_priv, fw); @@ -1952,7 +2055,7 @@ int __intel_wait_for_register(struct drm_i915_private *dev_priv, intel_uncore_forcewake_put__locked(dev_priv, fw); spin_unlock_irq(&dev_priv->uncore.lock); - if (ret) + if (ret && slow_timeout_ms) ret = __wait_for(reg_value = I915_READ_NOTRACE(reg), (reg_value & mask) == value, slow_timeout_ms * 1000, 10, 1000); @@ -1971,11 +2074,12 @@ static int gen8_reset_engine_start(struct intel_engine_cs *engine) I915_WRITE_FW(RING_RESET_CTL(engine->mmio_base), _MASKED_BIT_ENABLE(RESET_CTL_REQUEST_RESET)); - ret = intel_wait_for_register_fw(dev_priv, - RING_RESET_CTL(engine->mmio_base), - RESET_CTL_READY_TO_RESET, - RESET_CTL_READY_TO_RESET, - 700); + ret = __intel_wait_for_register_fw(dev_priv, + RING_RESET_CTL(engine->mmio_base), + RESET_CTL_READY_TO_RESET, + RESET_CTL_READY_TO_RESET, + 700, 0, + NULL); if (ret) DRM_ERROR("%s: reset request timeout\n", engine->name); @@ -2000,7 +2104,10 @@ static int gen8_reset_engines(struct drm_i915_private *dev_priv, if (gen8_reset_engine_start(engine)) goto not_ready; - return gen6_reset_engines(dev_priv, engine_mask); + if (INTEL_GEN(dev_priv) >= 11) + return gen11_reset_engines(dev_priv, engine_mask); + else + return gen6_reset_engines(dev_priv, engine_mask); not_ready: for_each_engine_masked(engine, dev_priv, engine_mask, tmp) @@ -2038,15 +2145,31 @@ int intel_gpu_reset(struct drm_i915_private *dev_priv, unsigned engine_mask) int retry; int ret; - might_sleep(); + /* + * We want to perform per-engine reset from atomic context (e.g. + * softirq), which imposes the constraint that we cannot sleep. + * However, experience suggests that spending a bit of time waiting + * for a reset helps in various cases, so for a full-device reset + * we apply the opposite rule and wait if we want to. As we should + * always follow up a failed per-engine reset with a full device reset, + * being a little faster, stricter and more error prone for the + * atomic case seems an acceptable compromise. + * + * Unfortunately this leads to a bimodal routine, when the goal was + * to have a single reset function that worked for resetting any + * number of engines simultaneously. + */ + might_sleep_if(engine_mask == ALL_ENGINES); - /* If the power well sleeps during the reset, the reset + /* + * If the power well sleeps during the reset, the reset * request may be dropped and never completes (causing -EIO). */ intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); for (retry = 0; retry < 3; retry++) { - /* We stop engines, otherwise we might get failed reset and a + /* + * We stop engines, otherwise we might get failed reset and a * dead gpu (on elk). Also as modern gpu as kbl can suffer * from system hang if batchbuffer is progressing when * the reset is issued, regardless of READY_TO_RESET ack. @@ -2060,9 +2183,11 @@ int intel_gpu_reset(struct drm_i915_private *dev_priv, unsigned engine_mask) i915_stop_engines(dev_priv, engine_mask); ret = -ENODEV; - if (reset) + if (reset) { + GEM_TRACE("engine_mask=%x\n", engine_mask); ret = reset(dev_priv, engine_mask); - if (ret != -ETIMEDOUT) + } + if (ret != -ETIMEDOUT || engine_mask != ALL_ENGINES) break; cond_resched(); @@ -2085,12 +2210,14 @@ bool intel_has_reset_engine(struct drm_i915_private *dev_priv) int intel_reset_guc(struct drm_i915_private *dev_priv) { + u32 guc_domain = INTEL_GEN(dev_priv) >= 11 ? GEN11_GRDOM_GUC : + GEN9_GRDOM_GUC; int ret; GEM_BUG_ON(!HAS_GUC(dev_priv)); intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); - ret = gen6_hw_domain_reset(dev_priv, GEN9_GRDOM_GUC); + ret = gen6_hw_domain_reset(dev_priv, guc_domain); intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); return ret; diff --git a/drivers/gpu/drm/i915/intel_uncore.h b/drivers/gpu/drm/i915/intel_uncore.h index dfdf444..47478d6 100644 --- a/drivers/gpu/drm/i915/intel_uncore.h +++ b/drivers/gpu/drm/i915/intel_uncore.h @@ -140,6 +140,7 @@ struct intel_uncore { void intel_uncore_sanitize(struct drm_i915_private *dev_priv); void intel_uncore_init(struct drm_i915_private *dev_priv); +void intel_uncore_prune(struct drm_i915_private *dev_priv); bool intel_uncore_unclaimed_mmio(struct drm_i915_private *dev_priv); bool intel_uncore_arm_unclaimed_mmio_detection(struct drm_i915_private *dev_priv); void intel_uncore_fini(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/intel_wopcm.c b/drivers/gpu/drm/i915/intel_wopcm.c new file mode 100644 index 0000000..74bf76f --- /dev/null +++ b/drivers/gpu/drm/i915/intel_wopcm.c @@ -0,0 +1,275 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2017-2018 Intel Corporation + */ + +#include "intel_wopcm.h" +#include "i915_drv.h" + +/** + * DOC: WOPCM Layout + * + * The layout of the WOPCM will be fixed after writing to GuC WOPCM size and + * offset registers whose values are calculated and determined by HuC/GuC + * firmware size and set of hardware requirements/restrictions as shown below: + * + * :: + * + * +=========> +====================+ <== WOPCM Top + * ^ | HW contexts RSVD | + * | +===> +====================+ <== GuC WOPCM Top + * | ^ | | + * | | | | + * | | | | + * | GuC | | + * | WOPCM | | + * | Size +--------------------+ + * WOPCM | | GuC FW RSVD | + * | | +--------------------+ + * | | | GuC Stack RSVD | + * | | +------------------- + + * | v | GuC WOPCM RSVD | + * | +===> +====================+ <== GuC WOPCM base + * | | WOPCM RSVD | + * | +------------------- + <== HuC Firmware Top + * v | HuC FW | + * +=========> +====================+ <== WOPCM Base + * + * GuC accessible WOPCM starts at GuC WOPCM base and ends at GuC WOPCM top. + * The top part of the WOPCM is reserved for hardware contexts (e.g. RC6 + * context). + */ + +/* Default WOPCM size 1MB. */ +#define GEN9_WOPCM_SIZE (1024 * 1024) +/* 16KB WOPCM (RSVD WOPCM) is reserved from HuC firmware top. */ +#define WOPCM_RESERVED_SIZE (16 * 1024) + +/* 16KB reserved at the beginning of GuC WOPCM. */ +#define GUC_WOPCM_RESERVED (16 * 1024) +/* 8KB from GUC_WOPCM_RESERVED is reserved for GuC stack. */ +#define GUC_WOPCM_STACK_RESERVED (8 * 1024) + +/* GuC WOPCM Offset value needs to be aligned to 16KB. */ +#define GUC_WOPCM_OFFSET_ALIGNMENT (1UL << GUC_WOPCM_OFFSET_SHIFT) + +/* 24KB at the end of WOPCM is reserved for RC6 CTX on BXT. */ +#define BXT_WOPCM_RC6_CTX_RESERVED (24 * 1024) +/* 36KB WOPCM reserved at the end of WOPCM on CNL. */ +#define CNL_WOPCM_HW_CTX_RESERVED (36 * 1024) + +/* 128KB from GUC_WOPCM_RESERVED is reserved for FW on Gen9. */ +#define GEN9_GUC_FW_RESERVED (128 * 1024) +#define GEN9_GUC_WOPCM_OFFSET (GUC_WOPCM_RESERVED + GEN9_GUC_FW_RESERVED) + +/** + * intel_wopcm_init_early() - Early initialization of the WOPCM. + * @wopcm: pointer to intel_wopcm. + * + * Setup the size of WOPCM which will be used by later on WOPCM partitioning. + */ +void intel_wopcm_init_early(struct intel_wopcm *wopcm) +{ + wopcm->size = GEN9_WOPCM_SIZE; + + DRM_DEBUG_DRIVER("WOPCM size: %uKiB\n", wopcm->size / 1024); +} + +static inline u32 context_reserved_size(struct drm_i915_private *i915) +{ + if (IS_GEN9_LP(i915)) + return BXT_WOPCM_RC6_CTX_RESERVED; + else if (INTEL_GEN(i915) >= 10) + return CNL_WOPCM_HW_CTX_RESERVED; + else + return 0; +} + +static inline int gen9_check_dword_gap(u32 guc_wopcm_base, u32 guc_wopcm_size) +{ + u32 offset; + + /* + * GuC WOPCM size shall be at least a dword larger than the offset from + * WOPCM base (GuC WOPCM offset from WOPCM base + GEN9_GUC_WOPCM_OFFSET) + * due to hardware limitation on Gen9. + */ + offset = guc_wopcm_base + GEN9_GUC_WOPCM_OFFSET; + if (offset > guc_wopcm_size || + (guc_wopcm_size - offset) < sizeof(u32)) { + DRM_ERROR("GuC WOPCM size %uKiB is too small. %uKiB needed.\n", + guc_wopcm_size / 1024, + (u32)(offset + sizeof(u32)) / 1024); + return -E2BIG; + } + + return 0; +} + +static inline int gen9_check_huc_fw_fits(u32 guc_wopcm_size, u32 huc_fw_size) +{ + /* + * On Gen9 & CNL A0, hardware requires the total available GuC WOPCM + * size to be larger than or equal to HuC firmware size. Otherwise, + * firmware uploading would fail. + */ + if (huc_fw_size > guc_wopcm_size - GUC_WOPCM_RESERVED) { + DRM_ERROR("HuC FW (%uKiB) won't fit in GuC WOPCM (%uKiB).\n", + huc_fw_size / 1024, + (guc_wopcm_size - GUC_WOPCM_RESERVED) / 1024); + return -E2BIG; + } + + return 0; +} + +static inline int check_hw_restriction(struct drm_i915_private *i915, + u32 guc_wopcm_base, u32 guc_wopcm_size, + u32 huc_fw_size) +{ + int err = 0; + + if (IS_GEN9(i915)) + err = gen9_check_dword_gap(guc_wopcm_base, guc_wopcm_size); + + if (!err && + (IS_GEN9(i915) || IS_CNL_REVID(i915, CNL_REVID_A0, CNL_REVID_A0))) + err = gen9_check_huc_fw_fits(guc_wopcm_size, huc_fw_size); + + return err; +} + +/** + * intel_wopcm_init() - Initialize the WOPCM structure. + * @wopcm: pointer to intel_wopcm. + * + * This function will partition WOPCM space based on GuC and HuC firmware sizes + * and will allocate max remaining for use by GuC. This function will also + * enforce platform dependent hardware restrictions on GuC WOPCM offset and + * size. It will fail the WOPCM init if any of these checks were failed, so that + * the following GuC firmware uploading would be aborted. + * + * Return: 0 on success, non-zero error code on failure. + */ +int intel_wopcm_init(struct intel_wopcm *wopcm) +{ + struct drm_i915_private *i915 = wopcm_to_i915(wopcm); + u32 guc_fw_size = intel_uc_fw_get_upload_size(&i915->guc.fw); + u32 huc_fw_size = intel_uc_fw_get_upload_size(&i915->huc.fw); + u32 ctx_rsvd = context_reserved_size(i915); + u32 guc_wopcm_base; + u32 guc_wopcm_size; + u32 guc_wopcm_rsvd; + int err; + + GEM_BUG_ON(!wopcm->size); + + if (guc_fw_size >= wopcm->size) { + DRM_ERROR("GuC FW (%uKiB) is too big to fit in WOPCM.", + guc_fw_size / 1024); + return -E2BIG; + } + + if (huc_fw_size >= wopcm->size) { + DRM_ERROR("HuC FW (%uKiB) is too big to fit in WOPCM.", + huc_fw_size / 1024); + return -E2BIG; + } + + guc_wopcm_base = ALIGN(huc_fw_size + WOPCM_RESERVED_SIZE, + GUC_WOPCM_OFFSET_ALIGNMENT); + if ((guc_wopcm_base + ctx_rsvd) >= wopcm->size) { + DRM_ERROR("GuC WOPCM base (%uKiB) is too big.\n", + guc_wopcm_base / 1024); + return -E2BIG; + } + + guc_wopcm_size = wopcm->size - guc_wopcm_base - ctx_rsvd; + guc_wopcm_size &= GUC_WOPCM_SIZE_MASK; + + DRM_DEBUG_DRIVER("Calculated GuC WOPCM Region: [%uKiB, %uKiB)\n", + guc_wopcm_base / 1024, guc_wopcm_size / 1024); + + guc_wopcm_rsvd = GUC_WOPCM_RESERVED + GUC_WOPCM_STACK_RESERVED; + if ((guc_fw_size + guc_wopcm_rsvd) > guc_wopcm_size) { + DRM_ERROR("Need %uKiB WOPCM for GuC, %uKiB available.\n", + (guc_fw_size + guc_wopcm_rsvd) / 1024, + guc_wopcm_size / 1024); + return -E2BIG; + } + + err = check_hw_restriction(i915, guc_wopcm_base, guc_wopcm_size, + huc_fw_size); + if (err) + return err; + + wopcm->guc.base = guc_wopcm_base; + wopcm->guc.size = guc_wopcm_size; + + return 0; +} + +static inline int write_and_verify(struct drm_i915_private *dev_priv, + i915_reg_t reg, u32 val, u32 mask, + u32 locked_bit) +{ + u32 reg_val; + + GEM_BUG_ON(val & ~mask); + + I915_WRITE(reg, val); + + reg_val = I915_READ(reg); + + return (reg_val & mask) != (val | locked_bit) ? -EIO : 0; +} + +/** + * intel_wopcm_init_hw() - Setup GuC WOPCM registers. + * @wopcm: pointer to intel_wopcm. + * + * Setup the GuC WOPCM size and offset registers with the calculated values. It + * will verify the register values to make sure the registers are locked with + * correct values. + * + * Return: 0 on success. -EIO if registers were locked with incorrect values. + */ +int intel_wopcm_init_hw(struct intel_wopcm *wopcm) +{ + struct drm_i915_private *dev_priv = wopcm_to_i915(wopcm); + u32 huc_agent; + u32 mask; + int err; + + if (!USES_GUC(dev_priv)) + return 0; + + GEM_BUG_ON(!HAS_GUC(dev_priv)); + GEM_BUG_ON(!wopcm->guc.size); + GEM_BUG_ON(!wopcm->guc.base); + + err = write_and_verify(dev_priv, GUC_WOPCM_SIZE, wopcm->guc.size, + GUC_WOPCM_SIZE_MASK | GUC_WOPCM_SIZE_LOCKED, + GUC_WOPCM_SIZE_LOCKED); + if (err) + goto err_out; + + huc_agent = USES_HUC(dev_priv) ? HUC_LOADING_AGENT_GUC : 0; + mask = GUC_WOPCM_OFFSET_MASK | GUC_WOPCM_OFFSET_VALID | huc_agent; + err = write_and_verify(dev_priv, DMA_GUC_WOPCM_OFFSET, + wopcm->guc.base | huc_agent, mask, + GUC_WOPCM_OFFSET_VALID); + if (err) + goto err_out; + + return 0; + +err_out: + DRM_ERROR("Failed to init WOPCM registers:\n"); + DRM_ERROR("DMA_GUC_WOPCM_OFFSET=%#x\n", + I915_READ(DMA_GUC_WOPCM_OFFSET)); + DRM_ERROR("GUC_WOPCM_SIZE=%#x\n", I915_READ(GUC_WOPCM_SIZE)); + + return err; +} diff --git a/drivers/gpu/drm/i915/intel_wopcm.h b/drivers/gpu/drm/i915/intel_wopcm.h new file mode 100644 index 0000000..6298910 --- /dev/null +++ b/drivers/gpu/drm/i915/intel_wopcm.h @@ -0,0 +1,31 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2017-2018 Intel Corporation + */ + +#ifndef _INTEL_WOPCM_H_ +#define _INTEL_WOPCM_H_ + +#include <linux/types.h> + +/** + * struct intel_wopcm - Overall WOPCM info and WOPCM regions. + * @size: Size of overall WOPCM. + * @guc: GuC WOPCM Region info. + * @guc.base: GuC WOPCM base which is offset from WOPCM base. + * @guc.size: Size of the GuC WOPCM region. + */ +struct intel_wopcm { + u32 size; + struct { + u32 base; + u32 size; + } guc; +}; + +void intel_wopcm_init_early(struct intel_wopcm *wopcm); +int intel_wopcm_init(struct intel_wopcm *wopcm); +int intel_wopcm_init_hw(struct intel_wopcm *wopcm); + +#endif diff --git a/drivers/gpu/drm/i915/intel_workarounds.c b/drivers/gpu/drm/i915/intel_workarounds.c new file mode 100644 index 0000000..2df3538 --- /dev/null +++ b/drivers/gpu/drm/i915/intel_workarounds.c @@ -0,0 +1,949 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2014-2018 Intel Corporation + */ + +#include "i915_drv.h" +#include "intel_workarounds.h" + +/** + * DOC: Hardware workarounds + * + * This file is intended as a central place to implement most [1]_ of the + * required workarounds for hardware to work as originally intended. They fall + * in five basic categories depending on how/when they are applied: + * + * - Workarounds that touch registers that are saved/restored to/from the HW + * context image. The list is emitted (via Load Register Immediate commands) + * everytime a new context is created. + * - GT workarounds. The list of these WAs is applied whenever these registers + * revert to default values (on GPU reset, suspend/resume [2]_, etc..). + * - Display workarounds. The list is applied during display clock-gating + * initialization. + * - Workarounds that whitelist a privileged register, so that UMDs can manage + * them directly. This is just a special case of a MMMIO workaround (as we + * write the list of these to/be-whitelisted registers to some special HW + * registers). + * - Workaround batchbuffers, that get executed automatically by the hardware + * on every HW context restore. + * + * .. [1] Please notice that there are other WAs that, due to their nature, + * cannot be applied from a central place. Those are peppered around the rest + * of the code, as needed. + * + * .. [2] Technically, some registers are powercontext saved & restored, so they + * survive a suspend/resume. In practice, writing them again is not too + * costly and simplifies things. We can revisit this in the future. + * + * Layout + * '''''' + * + * Keep things in this file ordered by WA type, as per the above (context, GT, + * display, register whitelist, batchbuffer). Then, inside each type, keep the + * following order: + * + * - Infrastructure functions and macros + * - WAs per platform in standard gen/chrono order + * - Public functions to init or apply the given workaround type. + */ + +static int wa_add(struct drm_i915_private *dev_priv, + i915_reg_t addr, + const u32 mask, const u32 val) +{ + const unsigned int idx = dev_priv->workarounds.count; + + if (WARN_ON(idx >= I915_MAX_WA_REGS)) + return -ENOSPC; + + dev_priv->workarounds.reg[idx].addr = addr; + dev_priv->workarounds.reg[idx].value = val; + dev_priv->workarounds.reg[idx].mask = mask; + + dev_priv->workarounds.count++; + + return 0; +} + +#define WA_REG(addr, mask, val) do { \ + const int r = wa_add(dev_priv, (addr), (mask), (val)); \ + if (r) \ + return r; \ + } while (0) + +#define WA_SET_BIT_MASKED(addr, mask) \ + WA_REG(addr, (mask), _MASKED_BIT_ENABLE(mask)) + +#define WA_CLR_BIT_MASKED(addr, mask) \ + WA_REG(addr, (mask), _MASKED_BIT_DISABLE(mask)) + +#define WA_SET_FIELD_MASKED(addr, mask, value) \ + WA_REG(addr, (mask), _MASKED_FIELD(mask, value)) + +static int gen8_ctx_workarounds_init(struct drm_i915_private *dev_priv) +{ + WA_SET_BIT_MASKED(INSTPM, INSTPM_FORCE_ORDERING); + + /* WaDisableAsyncFlipPerfMode:bdw,chv */ + WA_SET_BIT_MASKED(MI_MODE, ASYNC_FLIP_PERF_DISABLE); + + /* WaDisablePartialInstShootdown:bdw,chv */ + WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, + PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE); + + /* Use Force Non-Coherent whenever executing a 3D context. This is a + * workaround for for a possible hang in the unlikely event a TLB + * invalidation occurs during a PSD flush. + */ + /* WaForceEnableNonCoherent:bdw,chv */ + /* WaHdcDisableFetchWhenMasked:bdw,chv */ + WA_SET_BIT_MASKED(HDC_CHICKEN0, + HDC_DONOT_FETCH_MEM_WHEN_MASKED | + HDC_FORCE_NON_COHERENT); + + /* From the Haswell PRM, Command Reference: Registers, CACHE_MODE_0: + * "The Hierarchical Z RAW Stall Optimization allows non-overlapping + * polygons in the same 8x4 pixel/sample area to be processed without + * stalling waiting for the earlier ones to write to Hierarchical Z + * buffer." + * + * This optimization is off by default for BDW and CHV; turn it on. + */ + WA_CLR_BIT_MASKED(CACHE_MODE_0_GEN7, HIZ_RAW_STALL_OPT_DISABLE); + + /* Wa4x4STCOptimizationDisable:bdw,chv */ + WA_SET_BIT_MASKED(CACHE_MODE_1, GEN8_4x4_STC_OPTIMIZATION_DISABLE); + + /* + * BSpec recommends 8x4 when MSAA is used, + * however in practice 16x4 seems fastest. + * + * Note that PS/WM thread counts depend on the WIZ hashing + * disable bit, which we don't touch here, but it's good + * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM). + */ + WA_SET_FIELD_MASKED(GEN7_GT_MODE, + GEN6_WIZ_HASHING_MASK, + GEN6_WIZ_HASHING_16x4); + + return 0; +} + +static int bdw_ctx_workarounds_init(struct drm_i915_private *dev_priv) +{ + int ret; + + ret = gen8_ctx_workarounds_init(dev_priv); + if (ret) + return ret; + + /* WaDisableThreadStallDopClockGating:bdw (pre-production) */ + WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE); + + /* WaDisableDopClockGating:bdw + * + * Also see the related UCGTCL1 write in broadwell_init_clock_gating() + * to disable EUTC clock gating. + */ + WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2, + DOP_CLOCK_GATING_DISABLE); + + WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, + GEN8_SAMPLER_POWER_BYPASS_DIS); + + WA_SET_BIT_MASKED(HDC_CHICKEN0, + /* WaForceContextSaveRestoreNonCoherent:bdw */ + HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT | + /* WaDisableFenceDestinationToSLM:bdw (pre-prod) */ + (IS_BDW_GT3(dev_priv) ? HDC_FENCE_DEST_SLM_DISABLE : 0)); + + return 0; +} + +static int chv_ctx_workarounds_init(struct drm_i915_private *dev_priv) +{ + int ret; + + ret = gen8_ctx_workarounds_init(dev_priv); + if (ret) + return ret; + + /* WaDisableThreadStallDopClockGating:chv */ + WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE); + + /* Improve HiZ throughput on CHV. */ + WA_SET_BIT_MASKED(HIZ_CHICKEN, CHV_HZ_8X8_MODE_IN_1X); + + return 0; +} + +static int gen9_ctx_workarounds_init(struct drm_i915_private *dev_priv) +{ + if (HAS_LLC(dev_priv)) { + /* WaCompressedResourceSamplerPbeMediaNewHashMode:skl,kbl + * + * Must match Display Engine. See + * WaCompressedResourceDisplayNewHashMode. + */ + WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, + GEN9_PBE_COMPRESSED_HASH_SELECTION); + WA_SET_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN7, + GEN9_SAMPLER_HASH_COMPRESSED_READ_ADDR); + } + + /* WaClearFlowControlGpgpuContextSave:skl,bxt,kbl,glk,cfl */ + /* WaDisablePartialInstShootdown:skl,bxt,kbl,glk,cfl */ + WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, + FLOW_CONTROL_ENABLE | + PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE); + + /* Syncing dependencies between camera and graphics:skl,bxt,kbl */ + if (!IS_COFFEELAKE(dev_priv)) + WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, + GEN9_DISABLE_OCL_OOB_SUPPRESS_LOGIC); + + /* WaEnableYV12BugFixInHalfSliceChicken7:skl,bxt,kbl,glk,cfl */ + /* WaEnableSamplerGPGPUPreemptionSupport:skl,bxt,kbl,cfl */ + WA_SET_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN7, + GEN9_ENABLE_YV12_BUGFIX | + GEN9_ENABLE_GPGPU_PREEMPTION); + + /* Wa4x4STCOptimizationDisable:skl,bxt,kbl,glk,cfl */ + /* WaDisablePartialResolveInVc:skl,bxt,kbl,cfl */ + WA_SET_BIT_MASKED(CACHE_MODE_1, + GEN8_4x4_STC_OPTIMIZATION_DISABLE | + GEN9_PARTIAL_RESOLVE_IN_VC_DISABLE); + + /* WaCcsTlbPrefetchDisable:skl,bxt,kbl,glk,cfl */ + WA_CLR_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN5, + GEN9_CCS_TLB_PREFETCH_ENABLE); + + /* WaForceContextSaveRestoreNonCoherent:skl,bxt,kbl,cfl */ + WA_SET_BIT_MASKED(HDC_CHICKEN0, + HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT | + HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE); + + /* WaForceEnableNonCoherent and WaDisableHDCInvalidation are + * both tied to WaForceContextSaveRestoreNonCoherent + * in some hsds for skl. We keep the tie for all gen9. The + * documentation is a bit hazy and so we want to get common behaviour, + * even though there is no clear evidence we would need both on kbl/bxt. + * This area has been source of system hangs so we play it safe + * and mimic the skl regardless of what bspec says. + * + * Use Force Non-Coherent whenever executing a 3D context. This + * is a workaround for a possible hang in the unlikely event + * a TLB invalidation occurs during a PSD flush. + */ + + /* WaForceEnableNonCoherent:skl,bxt,kbl,cfl */ + WA_SET_BIT_MASKED(HDC_CHICKEN0, + HDC_FORCE_NON_COHERENT); + + /* WaDisableSamplerPowerBypassForSOPingPong:skl,bxt,kbl,cfl */ + if (IS_SKYLAKE(dev_priv) || + IS_KABYLAKE(dev_priv) || + IS_COFFEELAKE(dev_priv)) + WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, + GEN8_SAMPLER_POWER_BYPASS_DIS); + + /* WaDisableSTUnitPowerOptimization:skl,bxt,kbl,glk,cfl */ + WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN2, GEN8_ST_PO_DISABLE); + + /* + * Supporting preemption with fine-granularity requires changes in the + * batch buffer programming. Since we can't break old userspace, we + * need to set our default preemption level to safe value. Userspace is + * still able to use more fine-grained preemption levels, since in + * WaEnablePreemptionGranularityControlByUMD we're whitelisting the + * per-ctx register. As such, WaDisable{3D,GPGPU}MidCmdPreemption are + * not real HW workarounds, but merely a way to start using preemption + * while maintaining old contract with userspace. + */ + + /* WaDisable3DMidCmdPreemption:skl,bxt,glk,cfl,[cnl] */ + WA_CLR_BIT_MASKED(GEN8_CS_CHICKEN1, GEN9_PREEMPT_3D_OBJECT_LEVEL); + + /* WaDisableGPGPUMidCmdPreemption:skl,bxt,blk,cfl,[cnl] */ + WA_SET_FIELD_MASKED(GEN8_CS_CHICKEN1, + GEN9_PREEMPT_GPGPU_LEVEL_MASK, + GEN9_PREEMPT_GPGPU_COMMAND_LEVEL); + + /* WaClearHIZ_WM_CHICKEN3:bxt,glk */ + if (IS_GEN9_LP(dev_priv)) + WA_SET_BIT_MASKED(GEN9_WM_CHICKEN3, GEN9_FACTOR_IN_CLR_VAL_HIZ); + + return 0; +} + +static int skl_tune_iz_hashing(struct drm_i915_private *dev_priv) +{ + u8 vals[3] = { 0, 0, 0 }; + unsigned int i; + + for (i = 0; i < 3; i++) { + u8 ss; + + /* + * Only consider slices where one, and only one, subslice has 7 + * EUs + */ + if (!is_power_of_2(INTEL_INFO(dev_priv)->sseu.subslice_7eu[i])) + continue; + + /* + * subslice_7eu[i] != 0 (because of the check above) and + * ss_max == 4 (maximum number of subslices possible per slice) + * + * -> 0 <= ss <= 3; + */ + ss = ffs(INTEL_INFO(dev_priv)->sseu.subslice_7eu[i]) - 1; + vals[i] = 3 - ss; + } + + if (vals[0] == 0 && vals[1] == 0 && vals[2] == 0) + return 0; + + /* Tune IZ hashing. See intel_device_info_runtime_init() */ + WA_SET_FIELD_MASKED(GEN7_GT_MODE, + GEN9_IZ_HASHING_MASK(2) | + GEN9_IZ_HASHING_MASK(1) | + GEN9_IZ_HASHING_MASK(0), + GEN9_IZ_HASHING(2, vals[2]) | + GEN9_IZ_HASHING(1, vals[1]) | + GEN9_IZ_HASHING(0, vals[0])); + + return 0; +} + +static int skl_ctx_workarounds_init(struct drm_i915_private *dev_priv) +{ + int ret; + + ret = gen9_ctx_workarounds_init(dev_priv); + if (ret) + return ret; + + return skl_tune_iz_hashing(dev_priv); +} + +static int bxt_ctx_workarounds_init(struct drm_i915_private *dev_priv) +{ + int ret; + + ret = gen9_ctx_workarounds_init(dev_priv); + if (ret) + return ret; + + /* WaDisableThreadStallDopClockGating:bxt */ + WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, + STALL_DOP_GATING_DISABLE); + + /* WaToEnableHwFixForPushConstHWBug:bxt */ + WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, + GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); + + return 0; +} + +static int kbl_ctx_workarounds_init(struct drm_i915_private *dev_priv) +{ + int ret; + + ret = gen9_ctx_workarounds_init(dev_priv); + if (ret) + return ret; + + /* WaDisableFenceDestinationToSLM:kbl (pre-prod) */ + if (IS_KBL_REVID(dev_priv, KBL_REVID_A0, KBL_REVID_A0)) + WA_SET_BIT_MASKED(HDC_CHICKEN0, + HDC_FENCE_DEST_SLM_DISABLE); + + /* WaToEnableHwFixForPushConstHWBug:kbl */ + if (IS_KBL_REVID(dev_priv, KBL_REVID_C0, REVID_FOREVER)) + WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, + GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); + + /* WaDisableSbeCacheDispatchPortSharing:kbl */ + WA_SET_BIT_MASKED(GEN7_HALF_SLICE_CHICKEN1, + GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE); + + return 0; +} + +static int glk_ctx_workarounds_init(struct drm_i915_private *dev_priv) +{ + int ret; + + ret = gen9_ctx_workarounds_init(dev_priv); + if (ret) + return ret; + + /* WaToEnableHwFixForPushConstHWBug:glk */ + WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, + GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); + + return 0; +} + +static int cfl_ctx_workarounds_init(struct drm_i915_private *dev_priv) +{ + int ret; + + ret = gen9_ctx_workarounds_init(dev_priv); + if (ret) + return ret; + + /* WaToEnableHwFixForPushConstHWBug:cfl */ + WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, + GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); + + /* WaDisableSbeCacheDispatchPortSharing:cfl */ + WA_SET_BIT_MASKED(GEN7_HALF_SLICE_CHICKEN1, + GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE); + + return 0; +} + +static int cnl_ctx_workarounds_init(struct drm_i915_private *dev_priv) +{ + /* WaForceContextSaveRestoreNonCoherent:cnl */ + WA_SET_BIT_MASKED(CNL_HDC_CHICKEN0, + HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT); + + /* WaThrottleEUPerfToAvoidTDBackPressure:cnl(pre-prod) */ + if (IS_CNL_REVID(dev_priv, CNL_REVID_B0, CNL_REVID_B0)) + WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, THROTTLE_12_5); + + /* WaDisableReplayBufferBankArbitrationOptimization:cnl */ + WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, + GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION); + + /* WaDisableEnhancedSBEVertexCaching:cnl (pre-prod) */ + if (IS_CNL_REVID(dev_priv, 0, CNL_REVID_B0)) + WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2, + GEN8_CSC2_SBE_VUE_CACHE_CONSERVATIVE); + + /* WaPushConstantDereferenceHoldDisable:cnl */ + WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2, PUSH_CONSTANT_DEREF_DISABLE); + + /* FtrEnableFastAnisoL1BankingFix:cnl */ + WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3, CNL_FAST_ANISO_L1_BANKING_FIX); + + /* WaDisable3DMidCmdPreemption:cnl */ + WA_CLR_BIT_MASKED(GEN8_CS_CHICKEN1, GEN9_PREEMPT_3D_OBJECT_LEVEL); + + /* WaDisableGPGPUMidCmdPreemption:cnl */ + WA_SET_FIELD_MASKED(GEN8_CS_CHICKEN1, + GEN9_PREEMPT_GPGPU_LEVEL_MASK, + GEN9_PREEMPT_GPGPU_COMMAND_LEVEL); + + /* WaDisableEarlyEOT:cnl */ + WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, DISABLE_EARLY_EOT); + + return 0; +} + +static int icl_ctx_workarounds_init(struct drm_i915_private *dev_priv) +{ + /* Wa_1604370585:icl (pre-prod) + * Formerly known as WaPushConstantDereferenceHoldDisable + */ + if (IS_ICL_REVID(dev_priv, ICL_REVID_A0, ICL_REVID_B0)) + WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2, + PUSH_CONSTANT_DEREF_DISABLE); + + /* WaForceEnableNonCoherent:icl + * This is not the same workaround as in early Gen9 platforms, where + * lacking this could cause system hangs, but coherency performance + * overhead is high and only a few compute workloads really need it + * (the register is whitelisted in hardware now, so UMDs can opt in + * for coherency if they have a good reason). + */ + WA_SET_BIT_MASKED(ICL_HDC_MODE, HDC_FORCE_NON_COHERENT); + + return 0; +} + +int intel_ctx_workarounds_init(struct drm_i915_private *dev_priv) +{ + int err = 0; + + dev_priv->workarounds.count = 0; + + if (INTEL_GEN(dev_priv) < 8) + err = 0; + else if (IS_BROADWELL(dev_priv)) + err = bdw_ctx_workarounds_init(dev_priv); + else if (IS_CHERRYVIEW(dev_priv)) + err = chv_ctx_workarounds_init(dev_priv); + else if (IS_SKYLAKE(dev_priv)) + err = skl_ctx_workarounds_init(dev_priv); + else if (IS_BROXTON(dev_priv)) + err = bxt_ctx_workarounds_init(dev_priv); + else if (IS_KABYLAKE(dev_priv)) + err = kbl_ctx_workarounds_init(dev_priv); + else if (IS_GEMINILAKE(dev_priv)) + err = glk_ctx_workarounds_init(dev_priv); + else if (IS_COFFEELAKE(dev_priv)) + err = cfl_ctx_workarounds_init(dev_priv); + else if (IS_CANNONLAKE(dev_priv)) + err = cnl_ctx_workarounds_init(dev_priv); + else if (IS_ICELAKE(dev_priv)) + err = icl_ctx_workarounds_init(dev_priv); + else + MISSING_CASE(INTEL_GEN(dev_priv)); + if (err) + return err; + + DRM_DEBUG_DRIVER("Number of context specific w/a: %d\n", + dev_priv->workarounds.count); + return 0; +} + +int intel_ctx_workarounds_emit(struct i915_request *rq) +{ + struct i915_workarounds *w = &rq->i915->workarounds; + u32 *cs; + int ret, i; + + if (w->count == 0) + return 0; + + ret = rq->engine->emit_flush(rq, EMIT_BARRIER); + if (ret) + return ret; + + cs = intel_ring_begin(rq, (w->count * 2 + 2)); + if (IS_ERR(cs)) + return PTR_ERR(cs); + + *cs++ = MI_LOAD_REGISTER_IMM(w->count); + for (i = 0; i < w->count; i++) { + *cs++ = i915_mmio_reg_offset(w->reg[i].addr); + *cs++ = w->reg[i].value; + } + *cs++ = MI_NOOP; + + intel_ring_advance(rq, cs); + + ret = rq->engine->emit_flush(rq, EMIT_BARRIER); + if (ret) + return ret; + + return 0; +} + +static void bdw_gt_workarounds_apply(struct drm_i915_private *dev_priv) +{ +} + +static void chv_gt_workarounds_apply(struct drm_i915_private *dev_priv) +{ +} + +static void gen9_gt_workarounds_apply(struct drm_i915_private *dev_priv) +{ + /* WaContextSwitchWithConcurrentTLBInvalidate:skl,bxt,kbl,glk,cfl */ + I915_WRITE(GEN9_CSFE_CHICKEN1_RCS, + _MASKED_BIT_ENABLE(GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE)); + + /* WaEnableLbsSlaRetryTimerDecrement:skl,bxt,kbl,glk,cfl */ + I915_WRITE(BDW_SCRATCH1, I915_READ(BDW_SCRATCH1) | + GEN9_LBS_SLA_RETRY_TIMER_DECREMENT_ENABLE); + + /* WaDisableKillLogic:bxt,skl,kbl */ + if (!IS_COFFEELAKE(dev_priv)) + I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | + ECOCHK_DIS_TLB); + + if (HAS_LLC(dev_priv)) { + /* WaCompressedResourceSamplerPbeMediaNewHashMode:skl,kbl + * + * Must match Display Engine. See + * WaCompressedResourceDisplayNewHashMode. + */ + I915_WRITE(MMCD_MISC_CTRL, + I915_READ(MMCD_MISC_CTRL) | + MMCD_PCLA | + MMCD_HOTSPOT_EN); + } + + /* WaDisableHDCInvalidation:skl,bxt,kbl,cfl */ + I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | + BDW_DISABLE_HDC_INVALIDATION); + + /* WaProgramL3SqcReg1DefaultForPerf:bxt,glk */ + if (IS_GEN9_LP(dev_priv)) { + u32 val = I915_READ(GEN8_L3SQCREG1); + + val &= ~L3_PRIO_CREDITS_MASK; + val |= L3_GENERAL_PRIO_CREDITS(62) | L3_HIGH_PRIO_CREDITS(2); + I915_WRITE(GEN8_L3SQCREG1, val); + } + + /* WaOCLCoherentLineFlush:skl,bxt,kbl,cfl */ + I915_WRITE(GEN8_L3SQCREG4, + I915_READ(GEN8_L3SQCREG4) | GEN8_LQSC_FLUSH_COHERENT_LINES); + + /* WaEnablePreemptionGranularityControlByUMD:skl,bxt,kbl,cfl,[cnl] */ + I915_WRITE(GEN7_FF_SLICE_CS_CHICKEN1, + _MASKED_BIT_ENABLE(GEN9_FFSC_PERCTX_PREEMPT_CTRL)); +} + +static void skl_gt_workarounds_apply(struct drm_i915_private *dev_priv) +{ + gen9_gt_workarounds_apply(dev_priv); + + /* WaEnableGapsTsvCreditFix:skl */ + I915_WRITE(GEN8_GARBCNTL, + I915_READ(GEN8_GARBCNTL) | GEN9_GAPS_TSV_CREDIT_DISABLE); + + /* WaDisableGafsUnitClkGating:skl */ + I915_WRITE(GEN7_UCGCTL4, + I915_READ(GEN7_UCGCTL4) | GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE); + + /* WaInPlaceDecompressionHang:skl */ + if (IS_SKL_REVID(dev_priv, SKL_REVID_H0, REVID_FOREVER)) + I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, + I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); +} + +static void bxt_gt_workarounds_apply(struct drm_i915_private *dev_priv) +{ + gen9_gt_workarounds_apply(dev_priv); + + /* WaDisablePooledEuLoadBalancingFix:bxt */ + I915_WRITE(FF_SLICE_CS_CHICKEN2, + _MASKED_BIT_ENABLE(GEN9_POOLED_EU_LOAD_BALANCING_FIX_DISABLE)); + + /* WaInPlaceDecompressionHang:bxt */ + I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, + I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); +} + +static void kbl_gt_workarounds_apply(struct drm_i915_private *dev_priv) +{ + gen9_gt_workarounds_apply(dev_priv); + + /* WaEnableGapsTsvCreditFix:kbl */ + I915_WRITE(GEN8_GARBCNTL, + I915_READ(GEN8_GARBCNTL) | GEN9_GAPS_TSV_CREDIT_DISABLE); + + /* WaDisableDynamicCreditSharing:kbl */ + if (IS_KBL_REVID(dev_priv, 0, KBL_REVID_B0)) + I915_WRITE(GAMT_CHKN_BIT_REG, + I915_READ(GAMT_CHKN_BIT_REG) | + GAMT_CHKN_DISABLE_DYNAMIC_CREDIT_SHARING); + + /* WaDisableGafsUnitClkGating:kbl */ + I915_WRITE(GEN7_UCGCTL4, + I915_READ(GEN7_UCGCTL4) | GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE); + + /* WaInPlaceDecompressionHang:kbl */ + I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, + I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); +} + +static void glk_gt_workarounds_apply(struct drm_i915_private *dev_priv) +{ + gen9_gt_workarounds_apply(dev_priv); +} + +static void cfl_gt_workarounds_apply(struct drm_i915_private *dev_priv) +{ + gen9_gt_workarounds_apply(dev_priv); + + /* WaEnableGapsTsvCreditFix:cfl */ + I915_WRITE(GEN8_GARBCNTL, + I915_READ(GEN8_GARBCNTL) | GEN9_GAPS_TSV_CREDIT_DISABLE); + + /* WaDisableGafsUnitClkGating:cfl */ + I915_WRITE(GEN7_UCGCTL4, + I915_READ(GEN7_UCGCTL4) | GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE); + + /* WaInPlaceDecompressionHang:cfl */ + I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, + I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); +} + +static void cnl_gt_workarounds_apply(struct drm_i915_private *dev_priv) +{ + /* WaDisableI2mCycleOnWRPort:cnl (pre-prod) */ + if (IS_CNL_REVID(dev_priv, CNL_REVID_B0, CNL_REVID_B0)) + I915_WRITE(GAMT_CHKN_BIT_REG, + I915_READ(GAMT_CHKN_BIT_REG) | + GAMT_CHKN_DISABLE_I2M_CYCLE_ON_WR_PORT); + + /* WaInPlaceDecompressionHang:cnl */ + I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, + I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); + + /* WaEnablePreemptionGranularityControlByUMD:cnl */ + I915_WRITE(GEN7_FF_SLICE_CS_CHICKEN1, + _MASKED_BIT_ENABLE(GEN9_FFSC_PERCTX_PREEMPT_CTRL)); +} + +static void icl_gt_workarounds_apply(struct drm_i915_private *dev_priv) +{ + /* This is not an Wa. Enable for better image quality */ + I915_WRITE(_3D_CHICKEN3, + _MASKED_BIT_ENABLE(_3D_CHICKEN3_AA_LINE_QUALITY_FIX_ENABLE)); + + /* WaInPlaceDecompressionHang:icl */ + I915_WRITE(GEN9_GAMT_ECO_REG_RW_IA, I915_READ(GEN9_GAMT_ECO_REG_RW_IA) | + GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS); + + /* WaPipelineFlushCoherentLines:icl */ + I915_WRITE(GEN8_L3SQCREG4, I915_READ(GEN8_L3SQCREG4) | + GEN8_LQSC_FLUSH_COHERENT_LINES); + + /* Wa_1405543622:icl + * Formerly known as WaGAPZPriorityScheme + */ + I915_WRITE(GEN8_GARBCNTL, I915_READ(GEN8_GARBCNTL) | + GEN11_ARBITRATION_PRIO_ORDER_MASK); + + /* Wa_1604223664:icl + * Formerly known as WaL3BankAddressHashing + */ + I915_WRITE(GEN8_GARBCNTL, + (I915_READ(GEN8_GARBCNTL) & ~GEN11_HASH_CTRL_EXCL_MASK) | + GEN11_HASH_CTRL_EXCL_BIT0); + I915_WRITE(GEN11_GLBLINVL, + (I915_READ(GEN11_GLBLINVL) & ~GEN11_BANK_HASH_ADDR_EXCL_MASK) | + GEN11_BANK_HASH_ADDR_EXCL_BIT0); + + /* WaModifyGamTlbPartitioning:icl */ + I915_WRITE(GEN11_GACB_PERF_CTRL, + (I915_READ(GEN11_GACB_PERF_CTRL) & ~GEN11_HASH_CTRL_MASK) | + GEN11_HASH_CTRL_BIT0 | GEN11_HASH_CTRL_BIT4); + + /* Wa_1405733216:icl + * Formerly known as WaDisableCleanEvicts + */ + I915_WRITE(GEN8_L3SQCREG4, I915_READ(GEN8_L3SQCREG4) | + GEN11_LQSC_CLEAN_EVICT_DISABLE); + + /* Wa_1405766107:icl + * Formerly known as WaCL2SFHalfMaxAlloc + */ + I915_WRITE(GEN11_LSN_UNSLCVC, I915_READ(GEN11_LSN_UNSLCVC) | + GEN11_LSN_UNSLCVC_GAFS_HALF_SF_MAXALLOC | + GEN11_LSN_UNSLCVC_GAFS_HALF_CL2_MAXALLOC); + + /* Wa_220166154:icl + * Formerly known as WaDisCtxReload + */ + I915_WRITE(GAMW_ECO_DEV_RW_IA_REG, I915_READ(GAMW_ECO_DEV_RW_IA_REG) | + GAMW_ECO_DEV_CTX_RELOAD_DISABLE); + + /* Wa_1405779004:icl (pre-prod) */ + if (IS_ICL_REVID(dev_priv, ICL_REVID_A0, ICL_REVID_A0)) + I915_WRITE(SLICE_UNIT_LEVEL_CLKGATE, + I915_READ(SLICE_UNIT_LEVEL_CLKGATE) | + MSCUNIT_CLKGATE_DIS); + + /* Wa_1406680159:icl */ + I915_WRITE(SUBSLICE_UNIT_LEVEL_CLKGATE, + I915_READ(SUBSLICE_UNIT_LEVEL_CLKGATE) | + GWUNIT_CLKGATE_DIS); + + /* Wa_1604302699:icl */ + I915_WRITE(GEN10_L3_CHICKEN_MODE_REGISTER, + I915_READ(GEN10_L3_CHICKEN_MODE_REGISTER) | + GEN11_I2M_WRITE_DISABLE); + + /* Wa_1406838659:icl (pre-prod) */ + if (IS_ICL_REVID(dev_priv, ICL_REVID_A0, ICL_REVID_B0)) + I915_WRITE(INF_UNIT_LEVEL_CLKGATE, + I915_READ(INF_UNIT_LEVEL_CLKGATE) | + CGPSF_CLKGATE_DIS); + + /* WaForwardProgressSoftReset:icl */ + I915_WRITE(GEN10_SCRATCH_LNCF2, + I915_READ(GEN10_SCRATCH_LNCF2) | + PMFLUSHDONE_LNICRSDROP | + PMFLUSH_GAPL3UNBLOCK | + PMFLUSHDONE_LNEBLK); +} + +void intel_gt_workarounds_apply(struct drm_i915_private *dev_priv) +{ + if (INTEL_GEN(dev_priv) < 8) + return; + else if (IS_BROADWELL(dev_priv)) + bdw_gt_workarounds_apply(dev_priv); + else if (IS_CHERRYVIEW(dev_priv)) + chv_gt_workarounds_apply(dev_priv); + else if (IS_SKYLAKE(dev_priv)) + skl_gt_workarounds_apply(dev_priv); + else if (IS_BROXTON(dev_priv)) + bxt_gt_workarounds_apply(dev_priv); + else if (IS_KABYLAKE(dev_priv)) + kbl_gt_workarounds_apply(dev_priv); + else if (IS_GEMINILAKE(dev_priv)) + glk_gt_workarounds_apply(dev_priv); + else if (IS_COFFEELAKE(dev_priv)) + cfl_gt_workarounds_apply(dev_priv); + else if (IS_CANNONLAKE(dev_priv)) + cnl_gt_workarounds_apply(dev_priv); + else if (IS_ICELAKE(dev_priv)) + icl_gt_workarounds_apply(dev_priv); + else + MISSING_CASE(INTEL_GEN(dev_priv)); +} + +struct whitelist { + i915_reg_t reg[RING_MAX_NONPRIV_SLOTS]; + unsigned int count; + u32 nopid; +}; + +static void whitelist_reg(struct whitelist *w, i915_reg_t reg) +{ + if (GEM_WARN_ON(w->count >= RING_MAX_NONPRIV_SLOTS)) + return; + + w->reg[w->count++] = reg; +} + +static void bdw_whitelist_build(struct whitelist *w) +{ +} + +static void chv_whitelist_build(struct whitelist *w) +{ +} + +static void gen9_whitelist_build(struct whitelist *w) +{ + /* WaVFEStateAfterPipeControlwithMediaStateClear:skl,bxt,glk,cfl */ + whitelist_reg(w, GEN9_CTX_PREEMPT_REG); + + /* WaEnablePreemptionGranularityControlByUMD:skl,bxt,kbl,cfl,[cnl] */ + whitelist_reg(w, GEN8_CS_CHICKEN1); + + /* WaAllowUMDToModifyHDCChicken1:skl,bxt,kbl,glk,cfl */ + whitelist_reg(w, GEN8_HDC_CHICKEN1); +} + +static void skl_whitelist_build(struct whitelist *w) +{ + gen9_whitelist_build(w); + + /* WaDisableLSQCROPERFforOCL:skl */ + whitelist_reg(w, GEN8_L3SQCREG4); +} + +static void bxt_whitelist_build(struct whitelist *w) +{ + gen9_whitelist_build(w); +} + +static void kbl_whitelist_build(struct whitelist *w) +{ + gen9_whitelist_build(w); + + /* WaDisableLSQCROPERFforOCL:kbl */ + whitelist_reg(w, GEN8_L3SQCREG4); +} + +static void glk_whitelist_build(struct whitelist *w) +{ + gen9_whitelist_build(w); + + /* WA #0862: Userspace has to set "Barrier Mode" to avoid hangs. */ + whitelist_reg(w, GEN9_SLICE_COMMON_ECO_CHICKEN1); +} + +static void cfl_whitelist_build(struct whitelist *w) +{ + gen9_whitelist_build(w); +} + +static void cnl_whitelist_build(struct whitelist *w) +{ + /* WaEnablePreemptionGranularityControlByUMD:cnl */ + whitelist_reg(w, GEN8_CS_CHICKEN1); +} + +static void icl_whitelist_build(struct whitelist *w) +{ +} + +static struct whitelist *whitelist_build(struct intel_engine_cs *engine, + struct whitelist *w) +{ + struct drm_i915_private *i915 = engine->i915; + + GEM_BUG_ON(engine->id != RCS); + + w->count = 0; + w->nopid = i915_mmio_reg_offset(RING_NOPID(engine->mmio_base)); + + if (INTEL_GEN(i915) < 8) + return NULL; + else if (IS_BROADWELL(i915)) + bdw_whitelist_build(w); + else if (IS_CHERRYVIEW(i915)) + chv_whitelist_build(w); + else if (IS_SKYLAKE(i915)) + skl_whitelist_build(w); + else if (IS_BROXTON(i915)) + bxt_whitelist_build(w); + else if (IS_KABYLAKE(i915)) + kbl_whitelist_build(w); + else if (IS_GEMINILAKE(i915)) + glk_whitelist_build(w); + else if (IS_COFFEELAKE(i915)) + cfl_whitelist_build(w); + else if (IS_CANNONLAKE(i915)) + cnl_whitelist_build(w); + else if (IS_ICELAKE(i915)) + icl_whitelist_build(w); + else + MISSING_CASE(INTEL_GEN(i915)); + + return w; +} + +static void whitelist_apply(struct intel_engine_cs *engine, + const struct whitelist *w) +{ + struct drm_i915_private *dev_priv = engine->i915; + const u32 base = engine->mmio_base; + unsigned int i; + + if (!w) + return; + + intel_uncore_forcewake_get(engine->i915, FORCEWAKE_ALL); + + for (i = 0; i < w->count; i++) + I915_WRITE_FW(RING_FORCE_TO_NONPRIV(base, i), + i915_mmio_reg_offset(w->reg[i])); + + /* And clear the rest just in case of garbage */ + for (; i < RING_MAX_NONPRIV_SLOTS; i++) + I915_WRITE_FW(RING_FORCE_TO_NONPRIV(base, i), w->nopid); + + intel_uncore_forcewake_put(engine->i915, FORCEWAKE_ALL); +} + +void intel_whitelist_workarounds_apply(struct intel_engine_cs *engine) +{ + struct whitelist w; + + whitelist_apply(engine, whitelist_build(engine, &w)); +} + +#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) +#include "selftests/intel_workarounds.c" +#endif diff --git a/drivers/gpu/drm/i915/intel_workarounds.h b/drivers/gpu/drm/i915/intel_workarounds.h new file mode 100644 index 0000000..b11d0623 --- /dev/null +++ b/drivers/gpu/drm/i915/intel_workarounds.h @@ -0,0 +1,17 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2014-2018 Intel Corporation + */ + +#ifndef _I915_WORKAROUNDS_H_ +#define _I915_WORKAROUNDS_H_ + +int intel_ctx_workarounds_init(struct drm_i915_private *dev_priv); +int intel_ctx_workarounds_emit(struct i915_request *rq); + +void intel_gt_workarounds_apply(struct drm_i915_private *dev_priv); + +void intel_whitelist_workarounds_apply(struct intel_engine_cs *engine); + +#endif diff --git a/drivers/gpu/drm/i915/selftests/huge_pages.c b/drivers/gpu/drm/i915/selftests/huge_pages.c index 05bbef3..91c7291 100644 --- a/drivers/gpu/drm/i915/selftests/huge_pages.c +++ b/drivers/gpu/drm/i915/selftests/huge_pages.c @@ -1091,7 +1091,7 @@ static int __igt_write_huge(struct i915_gem_context *ctx, out_vma_unpin: i915_vma_unpin(vma); out_vma_close: - i915_vma_close(vma); + i915_vma_destroy(vma); return err; } @@ -1757,6 +1757,9 @@ int i915_gem_huge_page_live_selftests(struct drm_i915_private *dev_priv) goto out_unlock; } + if (ctx->ppgtt) + ctx->ppgtt->base.scrub_64K = true; + err = i915_subtests(tests, ctx); out_unlock: diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_context.c b/drivers/gpu/drm/i915/selftests/i915_gem_context.c index 7ecaed5..ddb03f0 100644 --- a/drivers/gpu/drm/i915/selftests/i915_gem_context.c +++ b/drivers/gpu/drm/i915/selftests/i915_gem_context.c @@ -23,6 +23,7 @@ */ #include "../i915_selftest.h" +#include "igt_flush_test.h" #include "mock_drm.h" #include "huge_gem_object.h" @@ -411,6 +412,8 @@ static int igt_ctx_exec(void *arg) } out_unlock: + if (igt_flush_test(i915, I915_WAIT_LOCKED)) + err = -EIO; mutex_unlock(&i915->drm.struct_mutex); mock_file_free(i915, file); diff --git a/drivers/gpu/drm/i915/selftests/i915_live_selftests.h b/drivers/gpu/drm/i915/selftests/i915_live_selftests.h index 9c76f03..a00e2bd 100644 --- a/drivers/gpu/drm/i915/selftests/i915_live_selftests.h +++ b/drivers/gpu/drm/i915/selftests/i915_live_selftests.h @@ -11,6 +11,7 @@ */ selftest(sanitycheck, i915_live_sanitycheck) /* keep first (igt selfcheck) */ selftest(uncore, intel_uncore_live_selftests) +selftest(workarounds, intel_workarounds_live_selftests) selftest(requests, i915_request_live_selftests) selftest(objects, i915_gem_object_live_selftests) selftest(dmabuf, i915_gem_dmabuf_live_selftests) @@ -20,4 +21,5 @@ selftest(evict, i915_gem_evict_live_selftests) selftest(hugepages, i915_gem_huge_page_live_selftests) selftest(contexts, i915_gem_context_live_selftests) selftest(hangcheck, intel_hangcheck_live_selftests) +selftest(execlists, intel_execlists_live_selftests) selftest(guc, intel_guc_live_selftest) diff --git a/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h b/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h index 9a48aa44..d16d741 100644 --- a/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h +++ b/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h @@ -14,6 +14,7 @@ selftest(fence, i915_sw_fence_mock_selftests) selftest(scatterlist, scatterlist_mock_selftests) selftest(syncmap, i915_syncmap_mock_selftests) selftest(uncore, intel_uncore_mock_selftests) +selftest(engine, intel_engine_cs_mock_selftests) selftest(breadcrumbs, intel_breadcrumbs_mock_selftests) selftest(timelines, i915_gem_timeline_mock_selftests) selftest(requests, i915_request_mock_selftests) diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_timeline.c b/drivers/gpu/drm/i915/selftests/i915_timeline.c index 3000e6a..19f1c6a 100644 --- a/drivers/gpu/drm/i915/selftests/i915_gem_timeline.c +++ b/drivers/gpu/drm/i915/selftests/i915_timeline.c @@ -1,25 +1,7 @@ /* - * Copyright © 2017 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. + * SPDX-License-Identifier: MIT * + * Copyright © 2017-2018 Intel Corporation */ #include "../i915_selftest.h" @@ -35,21 +17,21 @@ struct __igt_sync { bool set; }; -static int __igt_sync(struct intel_timeline *tl, +static int __igt_sync(struct i915_timeline *tl, u64 ctx, const struct __igt_sync *p, const char *name) { int ret; - if (__intel_timeline_sync_is_later(tl, ctx, p->seqno) != p->expected) { + if (__i915_timeline_sync_is_later(tl, ctx, p->seqno) != p->expected) { pr_err("%s: %s(ctx=%llu, seqno=%u) expected passed %s but failed\n", name, p->name, ctx, p->seqno, yesno(p->expected)); return -EINVAL; } if (p->set) { - ret = __intel_timeline_sync_set(tl, ctx, p->seqno); + ret = __i915_timeline_sync_set(tl, ctx, p->seqno); if (ret) return ret; } @@ -77,37 +59,31 @@ static int igt_sync(void *arg) { "unwrap", UINT_MAX, true, false }, {}, }, *p; - struct intel_timeline *tl; + struct i915_timeline tl; int order, offset; int ret = -ENODEV; - tl = mock_timeline(0); - if (!tl) - return -ENOMEM; - + mock_timeline_init(&tl, 0); for (p = pass; p->name; p++) { for (order = 1; order < 64; order++) { for (offset = -1; offset <= (order > 1); offset++) { u64 ctx = BIT_ULL(order) + offset; - ret = __igt_sync(tl, ctx, p, "1"); + ret = __igt_sync(&tl, ctx, p, "1"); if (ret) goto out; } } } - mock_timeline_destroy(tl); - - tl = mock_timeline(0); - if (!tl) - return -ENOMEM; + mock_timeline_fini(&tl); + mock_timeline_init(&tl, 0); for (order = 1; order < 64; order++) { for (offset = -1; offset <= (order > 1); offset++) { u64 ctx = BIT_ULL(order) + offset; for (p = pass; p->name; p++) { - ret = __igt_sync(tl, ctx, p, "2"); + ret = __igt_sync(&tl, ctx, p, "2"); if (ret) goto out; } @@ -115,7 +91,7 @@ static int igt_sync(void *arg) } out: - mock_timeline_destroy(tl); + mock_timeline_fini(&tl); return ret; } @@ -127,15 +103,13 @@ static unsigned int random_engine(struct rnd_state *rnd) static int bench_sync(void *arg) { struct rnd_state prng; - struct intel_timeline *tl; + struct i915_timeline tl; unsigned long end_time, count; u64 prng32_1M; ktime_t kt; int order, last_order; - tl = mock_timeline(0); - if (!tl) - return -ENOMEM; + mock_timeline_init(&tl, 0); /* Lookups from cache are very fast and so the random number generation * and the loop itself becomes a significant factor in the per-iteration @@ -167,7 +141,7 @@ static int bench_sync(void *arg) do { u64 id = i915_prandom_u64_state(&prng); - __intel_timeline_sync_set(tl, id, 0); + __i915_timeline_sync_set(&tl, id, 0); count++; } while (!time_after(jiffies, end_time)); kt = ktime_sub(ktime_get(), kt); @@ -182,8 +156,8 @@ static int bench_sync(void *arg) while (end_time--) { u64 id = i915_prandom_u64_state(&prng); - if (!__intel_timeline_sync_is_later(tl, id, 0)) { - mock_timeline_destroy(tl); + if (!__i915_timeline_sync_is_later(&tl, id, 0)) { + mock_timeline_fini(&tl); pr_err("Lookup of %llu failed\n", id); return -EINVAL; } @@ -193,19 +167,17 @@ static int bench_sync(void *arg) pr_info("%s: %lu random lookups, %lluns/lookup\n", __func__, count, (long long)div64_ul(ktime_to_ns(kt), count)); - mock_timeline_destroy(tl); + mock_timeline_fini(&tl); cond_resched(); - tl = mock_timeline(0); - if (!tl) - return -ENOMEM; + mock_timeline_init(&tl, 0); /* Benchmark setting the first N (in order) contexts */ count = 0; kt = ktime_get(); end_time = jiffies + HZ/10; do { - __intel_timeline_sync_set(tl, count++, 0); + __i915_timeline_sync_set(&tl, count++, 0); } while (!time_after(jiffies, end_time)); kt = ktime_sub(ktime_get(), kt); pr_info("%s: %lu in-order insertions, %lluns/insert\n", @@ -215,9 +187,9 @@ static int bench_sync(void *arg) end_time = count; kt = ktime_get(); while (end_time--) { - if (!__intel_timeline_sync_is_later(tl, end_time, 0)) { + if (!__i915_timeline_sync_is_later(&tl, end_time, 0)) { pr_err("Lookup of %lu failed\n", end_time); - mock_timeline_destroy(tl); + mock_timeline_fini(&tl); return -EINVAL; } } @@ -225,12 +197,10 @@ static int bench_sync(void *arg) pr_info("%s: %lu in-order lookups, %lluns/lookup\n", __func__, count, (long long)div64_ul(ktime_to_ns(kt), count)); - mock_timeline_destroy(tl); + mock_timeline_fini(&tl); cond_resched(); - tl = mock_timeline(0); - if (!tl) - return -ENOMEM; + mock_timeline_init(&tl, 0); /* Benchmark searching for a random context id and maybe changing it */ prandom_seed_state(&prng, i915_selftest.random_seed); @@ -241,8 +211,8 @@ static int bench_sync(void *arg) u32 id = random_engine(&prng); u32 seqno = prandom_u32_state(&prng); - if (!__intel_timeline_sync_is_later(tl, id, seqno)) - __intel_timeline_sync_set(tl, id, seqno); + if (!__i915_timeline_sync_is_later(&tl, id, seqno)) + __i915_timeline_sync_set(&tl, id, seqno); count++; } while (!time_after(jiffies, end_time)); @@ -250,7 +220,7 @@ static int bench_sync(void *arg) kt = ktime_sub_ns(kt, (count * prng32_1M * 2) >> 20); pr_info("%s: %lu repeated insert/lookups, %lluns/op\n", __func__, count, (long long)div64_ul(ktime_to_ns(kt), count)); - mock_timeline_destroy(tl); + mock_timeline_fini(&tl); cond_resched(); /* Benchmark searching for a known context id and changing the seqno */ @@ -258,9 +228,7 @@ static int bench_sync(void *arg) ({ int tmp = last_order; last_order = order; order += tmp; })) { unsigned int mask = BIT(order) - 1; - tl = mock_timeline(0); - if (!tl) - return -ENOMEM; + mock_timeline_init(&tl, 0); count = 0; kt = ktime_get(); @@ -272,8 +240,8 @@ static int bench_sync(void *arg) */ u64 id = (u64)(count & mask) << order; - __intel_timeline_sync_is_later(tl, id, 0); - __intel_timeline_sync_set(tl, id, 0); + __i915_timeline_sync_is_later(&tl, id, 0); + __i915_timeline_sync_set(&tl, id, 0); count++; } while (!time_after(jiffies, end_time)); @@ -281,7 +249,7 @@ static int bench_sync(void *arg) pr_info("%s: %lu cyclic/%d insert/lookups, %lluns/op\n", __func__, count, order, (long long)div64_ul(ktime_to_ns(kt), count)); - mock_timeline_destroy(tl); + mock_timeline_fini(&tl); cond_resched(); } diff --git a/drivers/gpu/drm/i915/selftests/i915_vma.c b/drivers/gpu/drm/i915/selftests/i915_vma.c index eb89e30..e90f972 100644 --- a/drivers/gpu/drm/i915/selftests/i915_vma.c +++ b/drivers/gpu/drm/i915/selftests/i915_vma.c @@ -81,7 +81,7 @@ checked_vma_instance(struct drm_i915_gem_object *obj, } if (i915_vma_compare(vma, vm, view)) { - pr_err("i915_vma_compare failed with create parmaters!\n"); + pr_err("i915_vma_compare failed with create parameters!\n"); return ERR_PTR(-EINVAL); } diff --git a/drivers/gpu/drm/i915/selftests/igt_flush_test.c b/drivers/gpu/drm/i915/selftests/igt_flush_test.c new file mode 100644 index 0000000..0d06f55 --- /dev/null +++ b/drivers/gpu/drm/i915/selftests/igt_flush_test.c @@ -0,0 +1,70 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2018 Intel Corporation + */ + +#include "../i915_drv.h" + +#include "../i915_selftest.h" +#include "igt_flush_test.h" + +struct wedge_me { + struct delayed_work work; + struct drm_i915_private *i915; + const void *symbol; +}; + +static void wedge_me(struct work_struct *work) +{ + struct wedge_me *w = container_of(work, typeof(*w), work.work); + + pr_err("%pS timed out, cancelling all further testing.\n", w->symbol); + + GEM_TRACE("%pS timed out.\n", w->symbol); + GEM_TRACE_DUMP(); + + i915_gem_set_wedged(w->i915); +} + +static void __init_wedge(struct wedge_me *w, + struct drm_i915_private *i915, + long timeout, + const void *symbol) +{ + w->i915 = i915; + w->symbol = symbol; + + INIT_DELAYED_WORK_ONSTACK(&w->work, wedge_me); + schedule_delayed_work(&w->work, timeout); +} + +static void __fini_wedge(struct wedge_me *w) +{ + cancel_delayed_work_sync(&w->work); + destroy_delayed_work_on_stack(&w->work); + w->i915 = NULL; +} + +#define wedge_on_timeout(W, DEV, TIMEOUT) \ + for (__init_wedge((W), (DEV), (TIMEOUT), __builtin_return_address(0)); \ + (W)->i915; \ + __fini_wedge((W))) + +int igt_flush_test(struct drm_i915_private *i915, unsigned int flags) +{ + struct wedge_me w; + + cond_resched(); + + if (flags & I915_WAIT_LOCKED && + i915_gem_switch_to_kernel_context(i915)) { + pr_err("Failed to switch back to kernel context; declaring wedged\n"); + i915_gem_set_wedged(i915); + } + + wedge_on_timeout(&w, i915, HZ) + i915_gem_wait_for_idle(i915, flags); + + return i915_terminally_wedged(&i915->gpu_error) ? -EIO : 0; +} diff --git a/drivers/gpu/drm/i915/selftests/igt_flush_test.h b/drivers/gpu/drm/i915/selftests/igt_flush_test.h new file mode 100644 index 0000000..63e0099 --- /dev/null +++ b/drivers/gpu/drm/i915/selftests/igt_flush_test.h @@ -0,0 +1,14 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2018 Intel Corporation + */ + +#ifndef IGT_FLUSH_TEST_H +#define IGT_FLUSH_TEST_H + +struct drm_i915_private; + +int igt_flush_test(struct drm_i915_private *i915, unsigned int flags); + +#endif /* IGT_FLUSH_TEST_H */ diff --git a/drivers/gpu/drm/i915/selftests/intel_breadcrumbs.c b/drivers/gpu/drm/i915/selftests/intel_breadcrumbs.c index 4658002..d6926e7 100644 --- a/drivers/gpu/drm/i915/selftests/intel_breadcrumbs.c +++ b/drivers/gpu/drm/i915/selftests/intel_breadcrumbs.c @@ -412,10 +412,11 @@ static int igt_wakeup(void *arg) * that they are ready for the next test. We wait until all * threads are complete and waiting for us (i.e. not a seqno). */ - err = wait_var_event_timeout(&done, !atomic_read(&done), 10 * HZ); - if (err) { + if (!wait_var_event_timeout(&done, + !atomic_read(&done), 10 * HZ)) { pr_err("Timed out waiting for %d remaining waiters\n", atomic_read(&done)); + err = -ETIMEDOUT; break; } diff --git a/drivers/gpu/drm/i915/selftests/intel_engine_cs.c b/drivers/gpu/drm/i915/selftests/intel_engine_cs.c new file mode 100644 index 0000000..cfaa6b29 --- /dev/null +++ b/drivers/gpu/drm/i915/selftests/intel_engine_cs.c @@ -0,0 +1,58 @@ +/* + * SPDX-License-Identifier: GPL-2.0 + * + * Copyright © 2018 Intel Corporation + */ + +#include "../i915_selftest.h" + +static int intel_mmio_bases_check(void *arg) +{ + int i, j; + + for (i = 0; i < ARRAY_SIZE(intel_engines); i++) { + const struct engine_info *info = &intel_engines[i]; + char name[INTEL_ENGINE_CS_MAX_NAME]; + u8 prev = U8_MAX; + + __sprint_engine_name(name, info); + + for (j = 0; j < MAX_MMIO_BASES; j++) { + u8 gen = info->mmio_bases[j].gen; + u32 base = info->mmio_bases[j].base; + + if (gen >= prev) { + pr_err("%s: %s: mmio base for gen %x " + "is before the one for gen %x\n", + __func__, name, prev, gen); + return -EINVAL; + } + + if (gen == 0) + break; + + if (!base) { + pr_err("%s: %s: invalid mmio base (%x) " + "for gen %x at entry %u\n", + __func__, name, base, gen, j); + return -EINVAL; + } + + prev = gen; + } + + pr_info("%s: min gen supported for %s = %d\n", + __func__, name, prev); + } + + return 0; +} + +int intel_engine_cs_mock_selftests(void) +{ + static const struct i915_subtest tests[] = { + SUBTEST(intel_mmio_bases_check), + }; + + return i915_subtests(tests, NULL); +} diff --git a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c index df7898c..438e0b0 100644 --- a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c +++ b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c @@ -25,10 +25,14 @@ #include <linux/kthread.h> #include "../i915_selftest.h" +#include "i915_random.h" +#include "igt_flush_test.h" #include "mock_context.h" #include "mock_drm.h" +#define IGT_IDLE_TIMEOUT 50 /* ms; time to wait after flushing between tests */ + struct hang { struct drm_i915_private *i915; struct drm_i915_gem_object *hws; @@ -250,58 +254,6 @@ static u32 hws_seqno(const struct hang *h, const struct i915_request *rq) return READ_ONCE(h->seqno[rq->fence.context % (PAGE_SIZE/sizeof(u32))]); } -struct wedge_me { - struct delayed_work work; - struct drm_i915_private *i915; - const void *symbol; -}; - -static void wedge_me(struct work_struct *work) -{ - struct wedge_me *w = container_of(work, typeof(*w), work.work); - - pr_err("%pS timed out, cancelling all further testing.\n", - w->symbol); - i915_gem_set_wedged(w->i915); -} - -static void __init_wedge(struct wedge_me *w, - struct drm_i915_private *i915, - long timeout, - const void *symbol) -{ - w->i915 = i915; - w->symbol = symbol; - - INIT_DELAYED_WORK_ONSTACK(&w->work, wedge_me); - schedule_delayed_work(&w->work, timeout); -} - -static void __fini_wedge(struct wedge_me *w) -{ - cancel_delayed_work_sync(&w->work); - destroy_delayed_work_on_stack(&w->work); - w->i915 = NULL; -} - -#define wedge_on_timeout(W, DEV, TIMEOUT) \ - for (__init_wedge((W), (DEV), (TIMEOUT), __builtin_return_address(0)); \ - (W)->i915; \ - __fini_wedge((W))) - -static noinline int -flush_test(struct drm_i915_private *i915, unsigned int flags) -{ - struct wedge_me w; - - cond_resched(); - - wedge_on_timeout(&w, i915, HZ) - i915_gem_wait_for_idle(i915, flags); - - return i915_terminally_wedged(&i915->gpu_error) ? -EIO : 0; -} - static void hang_fini(struct hang *h) { *h->batch = MI_BATCH_BUFFER_END; @@ -315,10 +267,10 @@ static void hang_fini(struct hang *h) kernel_context_close(h->ctx); - flush_test(h->i915, I915_WAIT_LOCKED); + igt_flush_test(h->i915, I915_WAIT_LOCKED); } -static bool wait_for_hang(struct hang *h, struct i915_request *rq) +static bool wait_until_running(struct hang *h, struct i915_request *rq) { return !(wait_for_us(i915_seqno_passed(hws_seqno(h, rq), rq->fence.seqno), @@ -433,7 +385,7 @@ static int igt_global_reset(void *arg) mutex_lock(&i915->drm.struct_mutex); reset_count = i915_reset_count(&i915->gpu_error); - i915_reset(i915, I915_RESET_QUIET); + i915_reset(i915, ALL_ENGINES, NULL); if (i915_reset_count(&i915->gpu_error) == reset_count) { pr_err("No GPU reset recorded!\n"); @@ -450,6 +402,11 @@ static int igt_global_reset(void *arg) return err; } +static bool wait_for_idle(struct intel_engine_cs *engine) +{ + return wait_for(intel_engine_is_idle(engine), IGT_IDLE_TIMEOUT) == 0; +} + static int __igt_reset_engine(struct drm_i915_private *i915, bool active) { struct intel_engine_cs *engine; @@ -477,12 +434,21 @@ static int __igt_reset_engine(struct drm_i915_private *i915, bool active) if (active && !intel_engine_can_store_dword(engine)) continue; + if (!wait_for_idle(engine)) { + pr_err("%s failed to idle before reset\n", + engine->name); + err = -EIO; + break; + } + reset_count = i915_reset_count(&i915->gpu_error); reset_engine_count = i915_reset_engine_count(&i915->gpu_error, engine); set_bit(I915_RESET_ENGINE + id, &i915->gpu_error.flags); do { + u32 seqno = intel_engine_get_seqno(engine); + if (active) { struct i915_request *rq; @@ -498,7 +464,7 @@ static int __igt_reset_engine(struct drm_i915_private *i915, bool active) __i915_request_add(rq, true); mutex_unlock(&i915->drm.struct_mutex); - if (!wait_for_hang(&h, rq)) { + if (!wait_until_running(&h, rq)) { struct drm_printer p = drm_info_printer(i915->drm.dev); pr_err("%s: Failed to start request %x, at %x\n", @@ -511,14 +477,12 @@ static int __igt_reset_engine(struct drm_i915_private *i915, bool active) break; } + GEM_BUG_ON(!rq->global_seqno); + seqno = rq->global_seqno - 1; i915_request_put(rq); } - engine->hangcheck.stalled = true; - engine->hangcheck.seqno = - intel_engine_get_seqno(engine); - - err = i915_reset_engine(engine, I915_RESET_QUIET); + err = i915_reset_engine(engine, NULL); if (err) { pr_err("i915_reset_engine failed\n"); break; @@ -539,14 +503,25 @@ static int __igt_reset_engine(struct drm_i915_private *i915, bool active) break; } - engine->hangcheck.stalled = false; + if (!wait_for_idle(engine)) { + struct drm_printer p = + drm_info_printer(i915->drm.dev); + + pr_err("%s failed to idle after reset\n", + engine->name); + intel_engine_dump(engine, &p, + "%s\n", engine->name); + + err = -EIO; + break; + } } while (time_before(jiffies, end_time)); clear_bit(I915_RESET_ENGINE + id, &i915->gpu_error.flags); if (err) break; - err = flush_test(i915, 0); + err = igt_flush_test(i915, 0); if (err) break; } @@ -573,11 +548,25 @@ static int igt_reset_active_engine(void *arg) return __igt_reset_engine(arg, true); } +struct active_engine { + struct task_struct *task; + struct intel_engine_cs *engine; + unsigned long resets; + unsigned int flags; +}; + +#define TEST_ACTIVE BIT(0) +#define TEST_OTHERS BIT(1) +#define TEST_SELF BIT(2) +#define TEST_PRIORITY BIT(3) + static int active_engine(void *data) { - struct intel_engine_cs *engine = data; - struct i915_request *rq[2] = {}; - struct i915_gem_context *ctx[2]; + I915_RND_STATE(prng); + struct active_engine *arg = data; + struct intel_engine_cs *engine = arg->engine; + struct i915_request *rq[8] = {}; + struct i915_gem_context *ctx[ARRAY_SIZE(rq)]; struct drm_file *file; unsigned long count = 0; int err = 0; @@ -586,25 +575,20 @@ static int active_engine(void *data) if (IS_ERR(file)) return PTR_ERR(file); - mutex_lock(&engine->i915->drm.struct_mutex); - ctx[0] = live_context(engine->i915, file); - mutex_unlock(&engine->i915->drm.struct_mutex); - if (IS_ERR(ctx[0])) { - err = PTR_ERR(ctx[0]); - goto err_file; - } - - mutex_lock(&engine->i915->drm.struct_mutex); - ctx[1] = live_context(engine->i915, file); - mutex_unlock(&engine->i915->drm.struct_mutex); - if (IS_ERR(ctx[1])) { - err = PTR_ERR(ctx[1]); - i915_gem_context_put(ctx[0]); - goto err_file; + for (count = 0; count < ARRAY_SIZE(ctx); count++) { + mutex_lock(&engine->i915->drm.struct_mutex); + ctx[count] = live_context(engine->i915, file); + mutex_unlock(&engine->i915->drm.struct_mutex); + if (IS_ERR(ctx[count])) { + err = PTR_ERR(ctx[count]); + while (--count) + i915_gem_context_put(ctx[count]); + goto err_file; + } } while (!kthread_should_stop()) { - unsigned int idx = count++ & 1; + unsigned int idx = count++ & (ARRAY_SIZE(rq) - 1); struct i915_request *old = rq[idx]; struct i915_request *new; @@ -616,14 +600,28 @@ static int active_engine(void *data) break; } + if (arg->flags & TEST_PRIORITY) + ctx[idx]->sched.priority = + i915_prandom_u32_max_state(512, &prng); + rq[idx] = i915_request_get(new); i915_request_add(new); mutex_unlock(&engine->i915->drm.struct_mutex); if (old) { - i915_request_wait(old, 0, MAX_SCHEDULE_TIMEOUT); + if (i915_request_wait(old, 0, HZ) < 0) { + GEM_TRACE("%s timed out.\n", engine->name); + GEM_TRACE_DUMP(); + + i915_gem_set_wedged(engine->i915); + i915_request_put(old); + err = -EIO; + break; + } i915_request_put(old); } + + cond_resched(); } for (count = 0; count < ARRAY_SIZE(rq); count++) @@ -634,8 +632,9 @@ err_file: return err; } -static int __igt_reset_engine_others(struct drm_i915_private *i915, - bool active) +static int __igt_reset_engines(struct drm_i915_private *i915, + const char *test_name, + unsigned int flags) { struct intel_engine_cs *engine, *other; enum intel_engine_id id, tmp; @@ -649,50 +648,68 @@ static int __igt_reset_engine_others(struct drm_i915_private *i915, if (!intel_has_reset_engine(i915)) return 0; - if (active) { + if (flags & TEST_ACTIVE) { mutex_lock(&i915->drm.struct_mutex); err = hang_init(&h, i915); mutex_unlock(&i915->drm.struct_mutex); if (err) return err; + + if (flags & TEST_PRIORITY) + h.ctx->sched.priority = 1024; } for_each_engine(engine, i915, id) { - struct task_struct *threads[I915_NUM_ENGINES] = {}; - unsigned long resets[I915_NUM_ENGINES]; + struct active_engine threads[I915_NUM_ENGINES] = {}; unsigned long global = i915_reset_count(&i915->gpu_error); - unsigned long count = 0; + unsigned long count = 0, reported; IGT_TIMEOUT(end_time); - if (active && !intel_engine_can_store_dword(engine)) + if (flags & TEST_ACTIVE && + !intel_engine_can_store_dword(engine)) continue; + if (!wait_for_idle(engine)) { + pr_err("i915_reset_engine(%s:%s): failed to idle before reset\n", + engine->name, test_name); + err = -EIO; + break; + } + memset(threads, 0, sizeof(threads)); for_each_engine(other, i915, tmp) { struct task_struct *tsk; - resets[tmp] = i915_reset_engine_count(&i915->gpu_error, - other); + threads[tmp].resets = + i915_reset_engine_count(&i915->gpu_error, + other); + + if (!(flags & TEST_OTHERS)) + continue; - if (other == engine) + if (other == engine && !(flags & TEST_SELF)) continue; - tsk = kthread_run(active_engine, other, + threads[tmp].engine = other; + threads[tmp].flags = flags; + + tsk = kthread_run(active_engine, &threads[tmp], "igt/%s", other->name); if (IS_ERR(tsk)) { err = PTR_ERR(tsk); goto unwind; } - threads[tmp] = tsk; + threads[tmp].task = tsk; get_task_struct(tsk); } set_bit(I915_RESET_ENGINE + id, &i915->gpu_error.flags); do { - if (active) { - struct i915_request *rq; + u32 seqno = intel_engine_get_seqno(engine); + struct i915_request *rq = NULL; + if (flags & TEST_ACTIVE) { mutex_lock(&i915->drm.struct_mutex); rq = hang_create_request(&h, engine); if (IS_ERR(rq)) { @@ -705,7 +722,7 @@ static int __igt_reset_engine_others(struct drm_i915_private *i915, __i915_request_add(rq, true); mutex_unlock(&i915->drm.struct_mutex); - if (!wait_for_hang(&h, rq)) { + if (!wait_until_running(&h, rq)) { struct drm_printer p = drm_info_printer(i915->drm.dev); pr_err("%s: Failed to start request %x, at %x\n", @@ -718,33 +735,48 @@ static int __igt_reset_engine_others(struct drm_i915_private *i915, break; } - i915_request_put(rq); + GEM_BUG_ON(!rq->global_seqno); + seqno = rq->global_seqno - 1; } - engine->hangcheck.stalled = true; - engine->hangcheck.seqno = - intel_engine_get_seqno(engine); - - err = i915_reset_engine(engine, I915_RESET_QUIET); + err = i915_reset_engine(engine, NULL); if (err) { - pr_err("i915_reset_engine(%s:%s) failed, err=%d\n", - engine->name, active ? "active" : "idle", err); + pr_err("i915_reset_engine(%s:%s): failed, err=%d\n", + engine->name, test_name, err); break; } - engine->hangcheck.stalled = false; count++; + + if (rq) { + i915_request_wait(rq, 0, MAX_SCHEDULE_TIMEOUT); + i915_request_put(rq); + } + + if (!(flags & TEST_SELF) && !wait_for_idle(engine)) { + struct drm_printer p = + drm_info_printer(i915->drm.dev); + + pr_err("i915_reset_engine(%s:%s):" + " failed to idle after reset\n", + engine->name, test_name); + intel_engine_dump(engine, &p, + "%s\n", engine->name); + + err = -EIO; + break; + } } while (time_before(jiffies, end_time)); clear_bit(I915_RESET_ENGINE + id, &i915->gpu_error.flags); pr_info("i915_reset_engine(%s:%s): %lu resets\n", - engine->name, active ? "active" : "idle", count); - - if (i915_reset_engine_count(&i915->gpu_error, engine) - - resets[engine->id] != (active ? count : 0)) { - pr_err("i915_reset_engine(%s:%s): reset %lu times, but reported %lu\n", - engine->name, active ? "active" : "idle", count, - i915_reset_engine_count(&i915->gpu_error, - engine) - resets[engine->id]); + engine->name, test_name, count); + + reported = i915_reset_engine_count(&i915->gpu_error, engine); + reported -= threads[engine->id].resets; + if (reported != (flags & TEST_ACTIVE ? count : 0)) { + pr_err("i915_reset_engine(%s:%s): reset %lu times, but reported %lu, expected %lu reported\n", + engine->name, test_name, count, reported, + (flags & TEST_ACTIVE ? count : 0)); if (!err) err = -EINVAL; } @@ -753,24 +785,26 @@ unwind: for_each_engine(other, i915, tmp) { int ret; - if (!threads[tmp]) + if (!threads[tmp].task) continue; - ret = kthread_stop(threads[tmp]); + ret = kthread_stop(threads[tmp].task); if (ret) { pr_err("kthread for other engine %s failed, err=%d\n", other->name, ret); if (!err) err = ret; } - put_task_struct(threads[tmp]); + put_task_struct(threads[tmp].task); - if (resets[tmp] != i915_reset_engine_count(&i915->gpu_error, - other)) { + if (other != engine && + threads[tmp].resets != + i915_reset_engine_count(&i915->gpu_error, other)) { pr_err("Innocent engine %s was reset (count=%ld)\n", other->name, i915_reset_engine_count(&i915->gpu_error, - other) - resets[tmp]); + other) - + threads[tmp].resets); if (!err) err = -EINVAL; } @@ -786,7 +820,7 @@ unwind: if (err) break; - err = flush_test(i915, 0); + err = igt_flush_test(i915, 0); if (err) break; } @@ -794,7 +828,7 @@ unwind: if (i915_terminally_wedged(&i915->gpu_error)) err = -EIO; - if (active) { + if (flags & TEST_ACTIVE) { mutex_lock(&i915->drm.struct_mutex); hang_fini(&h); mutex_unlock(&i915->drm.struct_mutex); @@ -803,27 +837,56 @@ unwind: return err; } -static int igt_reset_idle_engine_others(void *arg) +static int igt_reset_engines(void *arg) { - return __igt_reset_engine_others(arg, false); -} + static const struct { + const char *name; + unsigned int flags; + } phases[] = { + { "idle", 0 }, + { "active", TEST_ACTIVE }, + { "others-idle", TEST_OTHERS }, + { "others-active", TEST_OTHERS | TEST_ACTIVE }, + { + "others-priority", + TEST_OTHERS | TEST_ACTIVE | TEST_PRIORITY + }, + { + "self-priority", + TEST_OTHERS | TEST_ACTIVE | TEST_PRIORITY | TEST_SELF, + }, + { } + }; + struct drm_i915_private *i915 = arg; + typeof(*phases) *p; + int err; -static int igt_reset_active_engine_others(void *arg) -{ - return __igt_reset_engine_others(arg, true); + for (p = phases; p->name; p++) { + if (p->flags & TEST_PRIORITY) { + if (!(i915->caps.scheduler & I915_SCHEDULER_CAP_PRIORITY)) + continue; + } + + err = __igt_reset_engines(arg, p->name, p->flags); + if (err) + return err; + } + + return 0; } -static u32 fake_hangcheck(struct i915_request *rq) +static u32 fake_hangcheck(struct i915_request *rq, u32 mask) { - u32 reset_count; + struct i915_gpu_error *error = &rq->i915->gpu_error; + u32 reset_count = i915_reset_count(error); - rq->engine->hangcheck.stalled = true; - rq->engine->hangcheck.seqno = intel_engine_get_seqno(rq->engine); + error->stalled_mask = mask; - reset_count = i915_reset_count(&rq->i915->gpu_error); + /* set_bit() must be after we have setup the backchannel (mask) */ + smp_mb__before_atomic(); + set_bit(I915_RESET_HANDOFF, &error->flags); - set_bit(I915_RESET_HANDOFF, &rq->i915->gpu_error.flags); - wake_up_all(&rq->i915->gpu_error.wait_queue); + wake_up_all(&error->wait_queue); return reset_count; } @@ -858,21 +921,20 @@ static int igt_wait_reset(void *arg) i915_request_get(rq); __i915_request_add(rq, true); - if (!wait_for_hang(&h, rq)) { + if (!wait_until_running(&h, rq)) { struct drm_printer p = drm_info_printer(i915->drm.dev); pr_err("%s: Failed to start request %x, at %x\n", __func__, rq->fence.seqno, hws_seqno(&h, rq)); intel_engine_dump(rq->engine, &p, "%s\n", rq->engine->name); - i915_reset(i915, 0); i915_gem_set_wedged(i915); err = -EIO; goto out_rq; } - reset_count = fake_hangcheck(rq); + reset_count = fake_hangcheck(rq, ALL_ENGINES); timeout = i915_request_wait(rq, I915_WAIT_LOCKED, 10); if (timeout < 0) { @@ -903,6 +965,23 @@ unlock: return err; } +static int wait_for_others(struct drm_i915_private *i915, + struct intel_engine_cs *exclude) +{ + struct intel_engine_cs *engine; + enum intel_engine_id id; + + for_each_engine(engine, i915, id) { + if (engine == exclude) + continue; + + if (!wait_for_idle(engine)) + return -EIO; + } + + return 0; +} + static int igt_reset_queue(void *arg) { struct drm_i915_private *i915 = arg; @@ -951,27 +1030,49 @@ static int igt_reset_queue(void *arg) i915_request_get(rq); __i915_request_add(rq, true); - if (!wait_for_hang(&h, prev)) { + /* + * XXX We don't handle resetting the kernel context + * very well. If we trigger a device reset twice in + * quick succession while the kernel context is + * executing, we may end up skipping the breadcrumb. + * This is really only a problem for the selftest as + * normally there is a large interlude between resets + * (hangcheck), or we focus on resetting just one + * engine and so avoid repeatedly resetting innocents. + */ + err = wait_for_others(i915, engine); + if (err) { + pr_err("%s(%s): Failed to idle other inactive engines after device reset\n", + __func__, engine->name); + i915_request_put(rq); + i915_request_put(prev); + + GEM_TRACE_DUMP(); + i915_gem_set_wedged(i915); + goto fini; + } + + if (!wait_until_running(&h, prev)) { struct drm_printer p = drm_info_printer(i915->drm.dev); - pr_err("%s: Failed to start request %x, at %x\n", - __func__, prev->fence.seqno, hws_seqno(&h, prev)); - intel_engine_dump(prev->engine, &p, - "%s\n", prev->engine->name); + pr_err("%s(%s): Failed to start request %x, at %x\n", + __func__, engine->name, + prev->fence.seqno, hws_seqno(&h, prev)); + intel_engine_dump(engine, &p, + "%s\n", engine->name); i915_request_put(rq); i915_request_put(prev); - i915_reset(i915, 0); i915_gem_set_wedged(i915); err = -EIO; goto fini; } - reset_count = fake_hangcheck(prev); + reset_count = fake_hangcheck(prev, ENGINE_MASK(id)); - i915_reset(i915, I915_RESET_QUIET); + i915_reset(i915, ENGINE_MASK(id), NULL); GEM_BUG_ON(test_bit(I915_RESET_HANDOFF, &i915->gpu_error.flags)); @@ -1013,7 +1114,7 @@ static int igt_reset_queue(void *arg) i915_request_put(prev); - err = flush_test(i915, I915_WAIT_LOCKED); + err = igt_flush_test(i915, I915_WAIT_LOCKED); if (err) break; } @@ -1044,7 +1145,7 @@ static int igt_handle_error(void *arg) if (!intel_has_reset_engine(i915)) return 0; - if (!intel_engine_can_store_dword(i915->engine[RCS])) + if (!engine || !intel_engine_can_store_dword(engine)) return 0; mutex_lock(&i915->drm.struct_mutex); @@ -1062,14 +1163,13 @@ static int igt_handle_error(void *arg) i915_request_get(rq); __i915_request_add(rq, true); - if (!wait_for_hang(&h, rq)) { + if (!wait_until_running(&h, rq)) { struct drm_printer p = drm_info_printer(i915->drm.dev); pr_err("%s: Failed to start request %x, at %x\n", __func__, rq->fence.seqno, hws_seqno(&h, rq)); intel_engine_dump(rq->engine, &p, "%s\n", rq->engine->name); - i915_reset(i915, 0); i915_gem_set_wedged(i915); err = -EIO; @@ -1081,10 +1181,7 @@ static int igt_handle_error(void *arg) /* Temporarily disable error capture */ error = xchg(&i915->gpu_error.first_error, (void *)-1); - engine->hangcheck.stalled = true; - engine->hangcheck.seqno = intel_engine_get_seqno(engine); - - i915_handle_error(i915, intel_engine_flag(engine), "%s", __func__); + i915_handle_error(i915, ENGINE_MASK(engine->id), 0, NULL); xchg(&i915->gpu_error.first_error, error); @@ -1112,8 +1209,7 @@ int intel_hangcheck_live_selftests(struct drm_i915_private *i915) SUBTEST(igt_hang_sanitycheck), SUBTEST(igt_reset_idle_engine), SUBTEST(igt_reset_active_engine), - SUBTEST(igt_reset_idle_engine_others), - SUBTEST(igt_reset_active_engine_others), + SUBTEST(igt_reset_engines), SUBTEST(igt_wait_reset), SUBTEST(igt_reset_queue), SUBTEST(igt_handle_error), @@ -1129,6 +1225,10 @@ int intel_hangcheck_live_selftests(struct drm_i915_private *i915) err = i915_subtests(tests, i915); + mutex_lock(&i915->drm.struct_mutex); + igt_flush_test(i915, I915_WAIT_LOCKED); + mutex_unlock(&i915->drm.struct_mutex); + i915_modparams.enable_hangcheck = saved_hangcheck; intel_runtime_pm_put(i915); diff --git a/drivers/gpu/drm/i915/selftests/intel_lrc.c b/drivers/gpu/drm/i915/selftests/intel_lrc.c new file mode 100644 index 0000000..1b8a071 --- /dev/null +++ b/drivers/gpu/drm/i915/selftests/intel_lrc.c @@ -0,0 +1,459 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2018 Intel Corporation + */ + +#include "../i915_selftest.h" +#include "igt_flush_test.h" + +#include "mock_context.h" + +struct spinner { + struct drm_i915_private *i915; + struct drm_i915_gem_object *hws; + struct drm_i915_gem_object *obj; + u32 *batch; + void *seqno; +}; + +static int spinner_init(struct spinner *spin, struct drm_i915_private *i915) +{ + unsigned int mode; + void *vaddr; + int err; + + GEM_BUG_ON(INTEL_GEN(i915) < 8); + + memset(spin, 0, sizeof(*spin)); + spin->i915 = i915; + + spin->hws = i915_gem_object_create_internal(i915, PAGE_SIZE); + if (IS_ERR(spin->hws)) { + err = PTR_ERR(spin->hws); + goto err; + } + + spin->obj = i915_gem_object_create_internal(i915, PAGE_SIZE); + if (IS_ERR(spin->obj)) { + err = PTR_ERR(spin->obj); + goto err_hws; + } + + i915_gem_object_set_cache_level(spin->hws, I915_CACHE_LLC); + vaddr = i915_gem_object_pin_map(spin->hws, I915_MAP_WB); + if (IS_ERR(vaddr)) { + err = PTR_ERR(vaddr); + goto err_obj; + } + spin->seqno = memset(vaddr, 0xff, PAGE_SIZE); + + mode = HAS_LLC(i915) ? I915_MAP_WB : I915_MAP_WC; + vaddr = i915_gem_object_pin_map(spin->obj, mode); + if (IS_ERR(vaddr)) { + err = PTR_ERR(vaddr); + goto err_unpin_hws; + } + spin->batch = vaddr; + + return 0; + +err_unpin_hws: + i915_gem_object_unpin_map(spin->hws); +err_obj: + i915_gem_object_put(spin->obj); +err_hws: + i915_gem_object_put(spin->hws); +err: + return err; +} + +static unsigned int seqno_offset(u64 fence) +{ + return offset_in_page(sizeof(u32) * fence); +} + +static u64 hws_address(const struct i915_vma *hws, + const struct i915_request *rq) +{ + return hws->node.start + seqno_offset(rq->fence.context); +} + +static int emit_recurse_batch(struct spinner *spin, + struct i915_request *rq, + u32 arbitration_command) +{ + struct i915_address_space *vm = &rq->ctx->ppgtt->base; + struct i915_vma *hws, *vma; + u32 *batch; + int err; + + vma = i915_vma_instance(spin->obj, vm, NULL); + if (IS_ERR(vma)) + return PTR_ERR(vma); + + hws = i915_vma_instance(spin->hws, vm, NULL); + if (IS_ERR(hws)) + return PTR_ERR(hws); + + err = i915_vma_pin(vma, 0, 0, PIN_USER); + if (err) + return err; + + err = i915_vma_pin(hws, 0, 0, PIN_USER); + if (err) + goto unpin_vma; + + i915_vma_move_to_active(vma, rq, 0); + if (!i915_gem_object_has_active_reference(vma->obj)) { + i915_gem_object_get(vma->obj); + i915_gem_object_set_active_reference(vma->obj); + } + + i915_vma_move_to_active(hws, rq, 0); + if (!i915_gem_object_has_active_reference(hws->obj)) { + i915_gem_object_get(hws->obj); + i915_gem_object_set_active_reference(hws->obj); + } + + batch = spin->batch; + + *batch++ = MI_STORE_DWORD_IMM_GEN4; + *batch++ = lower_32_bits(hws_address(hws, rq)); + *batch++ = upper_32_bits(hws_address(hws, rq)); + *batch++ = rq->fence.seqno; + + *batch++ = arbitration_command; + + *batch++ = MI_BATCH_BUFFER_START | 1 << 8 | 1; + *batch++ = lower_32_bits(vma->node.start); + *batch++ = upper_32_bits(vma->node.start); + *batch++ = MI_BATCH_BUFFER_END; /* not reached */ + + i915_gem_chipset_flush(spin->i915); + + err = rq->engine->emit_bb_start(rq, vma->node.start, PAGE_SIZE, 0); + + i915_vma_unpin(hws); +unpin_vma: + i915_vma_unpin(vma); + return err; +} + +static struct i915_request * +spinner_create_request(struct spinner *spin, + struct i915_gem_context *ctx, + struct intel_engine_cs *engine, + u32 arbitration_command) +{ + struct i915_request *rq; + int err; + + rq = i915_request_alloc(engine, ctx); + if (IS_ERR(rq)) + return rq; + + err = emit_recurse_batch(spin, rq, arbitration_command); + if (err) { + __i915_request_add(rq, false); + return ERR_PTR(err); + } + + return rq; +} + +static u32 hws_seqno(const struct spinner *spin, const struct i915_request *rq) +{ + u32 *seqno = spin->seqno + seqno_offset(rq->fence.context); + + return READ_ONCE(*seqno); +} + +static void spinner_end(struct spinner *spin) +{ + *spin->batch = MI_BATCH_BUFFER_END; + i915_gem_chipset_flush(spin->i915); +} + +static void spinner_fini(struct spinner *spin) +{ + spinner_end(spin); + + i915_gem_object_unpin_map(spin->obj); + i915_gem_object_put(spin->obj); + + i915_gem_object_unpin_map(spin->hws); + i915_gem_object_put(spin->hws); +} + +static bool wait_for_spinner(struct spinner *spin, struct i915_request *rq) +{ + if (!wait_event_timeout(rq->execute, + READ_ONCE(rq->global_seqno), + msecs_to_jiffies(10))) + return false; + + return !(wait_for_us(i915_seqno_passed(hws_seqno(spin, rq), + rq->fence.seqno), + 10) && + wait_for(i915_seqno_passed(hws_seqno(spin, rq), + rq->fence.seqno), + 1000)); +} + +static int live_sanitycheck(void *arg) +{ + struct drm_i915_private *i915 = arg; + struct intel_engine_cs *engine; + struct i915_gem_context *ctx; + enum intel_engine_id id; + struct spinner spin; + int err = -ENOMEM; + + if (!HAS_LOGICAL_RING_CONTEXTS(i915)) + return 0; + + mutex_lock(&i915->drm.struct_mutex); + + if (spinner_init(&spin, i915)) + goto err_unlock; + + ctx = kernel_context(i915); + if (!ctx) + goto err_spin; + + for_each_engine(engine, i915, id) { + struct i915_request *rq; + + rq = spinner_create_request(&spin, ctx, engine, MI_NOOP); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + goto err_ctx; + } + + i915_request_add(rq); + if (!wait_for_spinner(&spin, rq)) { + GEM_TRACE("spinner failed to start\n"); + GEM_TRACE_DUMP(); + i915_gem_set_wedged(i915); + err = -EIO; + goto err_ctx; + } + + spinner_end(&spin); + if (igt_flush_test(i915, I915_WAIT_LOCKED)) { + err = -EIO; + goto err_ctx; + } + } + + err = 0; +err_ctx: + kernel_context_close(ctx); +err_spin: + spinner_fini(&spin); +err_unlock: + igt_flush_test(i915, I915_WAIT_LOCKED); + mutex_unlock(&i915->drm.struct_mutex); + return err; +} + +static int live_preempt(void *arg) +{ + struct drm_i915_private *i915 = arg; + struct i915_gem_context *ctx_hi, *ctx_lo; + struct spinner spin_hi, spin_lo; + struct intel_engine_cs *engine; + enum intel_engine_id id; + int err = -ENOMEM; + + if (!HAS_LOGICAL_RING_PREEMPTION(i915)) + return 0; + + mutex_lock(&i915->drm.struct_mutex); + + if (spinner_init(&spin_hi, i915)) + goto err_unlock; + + if (spinner_init(&spin_lo, i915)) + goto err_spin_hi; + + ctx_hi = kernel_context(i915); + if (!ctx_hi) + goto err_spin_lo; + ctx_hi->sched.priority = I915_CONTEXT_MAX_USER_PRIORITY; + + ctx_lo = kernel_context(i915); + if (!ctx_lo) + goto err_ctx_hi; + ctx_lo->sched.priority = I915_CONTEXT_MIN_USER_PRIORITY; + + for_each_engine(engine, i915, id) { + struct i915_request *rq; + + rq = spinner_create_request(&spin_lo, ctx_lo, engine, + MI_ARB_CHECK); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + goto err_ctx_lo; + } + + i915_request_add(rq); + if (!wait_for_spinner(&spin_lo, rq)) { + GEM_TRACE("lo spinner failed to start\n"); + GEM_TRACE_DUMP(); + i915_gem_set_wedged(i915); + err = -EIO; + goto err_ctx_lo; + } + + rq = spinner_create_request(&spin_hi, ctx_hi, engine, + MI_ARB_CHECK); + if (IS_ERR(rq)) { + spinner_end(&spin_lo); + err = PTR_ERR(rq); + goto err_ctx_lo; + } + + i915_request_add(rq); + if (!wait_for_spinner(&spin_hi, rq)) { + GEM_TRACE("hi spinner failed to start\n"); + GEM_TRACE_DUMP(); + i915_gem_set_wedged(i915); + err = -EIO; + goto err_ctx_lo; + } + + spinner_end(&spin_hi); + spinner_end(&spin_lo); + if (igt_flush_test(i915, I915_WAIT_LOCKED)) { + err = -EIO; + goto err_ctx_lo; + } + } + + err = 0; +err_ctx_lo: + kernel_context_close(ctx_lo); +err_ctx_hi: + kernel_context_close(ctx_hi); +err_spin_lo: + spinner_fini(&spin_lo); +err_spin_hi: + spinner_fini(&spin_hi); +err_unlock: + igt_flush_test(i915, I915_WAIT_LOCKED); + mutex_unlock(&i915->drm.struct_mutex); + return err; +} + +static int live_late_preempt(void *arg) +{ + struct drm_i915_private *i915 = arg; + struct i915_gem_context *ctx_hi, *ctx_lo; + struct spinner spin_hi, spin_lo; + struct intel_engine_cs *engine; + struct i915_sched_attr attr = {}; + enum intel_engine_id id; + int err = -ENOMEM; + + if (!HAS_LOGICAL_RING_PREEMPTION(i915)) + return 0; + + mutex_lock(&i915->drm.struct_mutex); + + if (spinner_init(&spin_hi, i915)) + goto err_unlock; + + if (spinner_init(&spin_lo, i915)) + goto err_spin_hi; + + ctx_hi = kernel_context(i915); + if (!ctx_hi) + goto err_spin_lo; + + ctx_lo = kernel_context(i915); + if (!ctx_lo) + goto err_ctx_hi; + + for_each_engine(engine, i915, id) { + struct i915_request *rq; + + rq = spinner_create_request(&spin_lo, ctx_lo, engine, + MI_ARB_CHECK); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + goto err_ctx_lo; + } + + i915_request_add(rq); + if (!wait_for_spinner(&spin_lo, rq)) { + pr_err("First context failed to start\n"); + goto err_wedged; + } + + rq = spinner_create_request(&spin_hi, ctx_hi, engine, MI_NOOP); + if (IS_ERR(rq)) { + spinner_end(&spin_lo); + err = PTR_ERR(rq); + goto err_ctx_lo; + } + + i915_request_add(rq); + if (wait_for_spinner(&spin_hi, rq)) { + pr_err("Second context overtook first?\n"); + goto err_wedged; + } + + attr.priority = I915_PRIORITY_MAX; + engine->schedule(rq, &attr); + + if (!wait_for_spinner(&spin_hi, rq)) { + pr_err("High priority context failed to preempt the low priority context\n"); + GEM_TRACE_DUMP(); + goto err_wedged; + } + + spinner_end(&spin_hi); + spinner_end(&spin_lo); + if (igt_flush_test(i915, I915_WAIT_LOCKED)) { + err = -EIO; + goto err_ctx_lo; + } + } + + err = 0; +err_ctx_lo: + kernel_context_close(ctx_lo); +err_ctx_hi: + kernel_context_close(ctx_hi); +err_spin_lo: + spinner_fini(&spin_lo); +err_spin_hi: + spinner_fini(&spin_hi); +err_unlock: + igt_flush_test(i915, I915_WAIT_LOCKED); + mutex_unlock(&i915->drm.struct_mutex); + return err; + +err_wedged: + spinner_end(&spin_hi); + spinner_end(&spin_lo); + i915_gem_set_wedged(i915); + err = -EIO; + goto err_ctx_lo; +} + +int intel_execlists_live_selftests(struct drm_i915_private *i915) +{ + static const struct i915_subtest tests[] = { + SUBTEST(live_sanitycheck), + SUBTEST(live_preempt), + SUBTEST(live_late_preempt), + }; + + if (!HAS_EXECLISTS(i915)) + return 0; + + return i915_subtests(tests, i915); +} diff --git a/drivers/gpu/drm/i915/selftests/intel_workarounds.c b/drivers/gpu/drm/i915/selftests/intel_workarounds.c new file mode 100644 index 0000000..17444a3 --- /dev/null +++ b/drivers/gpu/drm/i915/selftests/intel_workarounds.c @@ -0,0 +1,291 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2018 Intel Corporation + */ + +#include "../i915_selftest.h" + +#include "mock_context.h" + +static struct drm_i915_gem_object * +read_nonprivs(struct i915_gem_context *ctx, struct intel_engine_cs *engine) +{ + struct drm_i915_gem_object *result; + struct i915_request *rq; + struct i915_vma *vma; + const u32 base = engine->mmio_base; + u32 srm, *cs; + int err; + int i; + + result = i915_gem_object_create_internal(engine->i915, PAGE_SIZE); + if (IS_ERR(result)) + return result; + + i915_gem_object_set_cache_level(result, I915_CACHE_LLC); + + cs = i915_gem_object_pin_map(result, I915_MAP_WB); + if (IS_ERR(cs)) { + err = PTR_ERR(cs); + goto err_obj; + } + memset(cs, 0xc5, PAGE_SIZE); + i915_gem_object_unpin_map(result); + + vma = i915_vma_instance(result, &engine->i915->ggtt.base, NULL); + if (IS_ERR(vma)) { + err = PTR_ERR(vma); + goto err_obj; + } + + err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL); + if (err) + goto err_obj; + + rq = i915_request_alloc(engine, ctx); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + goto err_pin; + } + + srm = MI_STORE_REGISTER_MEM | MI_SRM_LRM_GLOBAL_GTT; + if (INTEL_GEN(ctx->i915) >= 8) + srm++; + + cs = intel_ring_begin(rq, 4 * RING_MAX_NONPRIV_SLOTS); + if (IS_ERR(cs)) { + err = PTR_ERR(cs); + goto err_req; + } + + for (i = 0; i < RING_MAX_NONPRIV_SLOTS; i++) { + *cs++ = srm; + *cs++ = i915_mmio_reg_offset(RING_FORCE_TO_NONPRIV(base, i)); + *cs++ = i915_ggtt_offset(vma) + sizeof(u32) * i; + *cs++ = 0; + } + intel_ring_advance(rq, cs); + + i915_vma_move_to_active(vma, rq, EXEC_OBJECT_WRITE); + reservation_object_lock(vma->resv, NULL); + reservation_object_add_excl_fence(vma->resv, &rq->fence); + reservation_object_unlock(vma->resv); + + i915_gem_object_get(result); + i915_gem_object_set_active_reference(result); + + __i915_request_add(rq, true); + i915_vma_unpin(vma); + + return result; + +err_req: + i915_request_add(rq); +err_pin: + i915_vma_unpin(vma); +err_obj: + i915_gem_object_put(result); + return ERR_PTR(err); +} + +static u32 get_whitelist_reg(const struct whitelist *w, unsigned int i) +{ + return i < w->count ? i915_mmio_reg_offset(w->reg[i]) : w->nopid; +} + +static void print_results(const struct whitelist *w, const u32 *results) +{ + unsigned int i; + + for (i = 0; i < RING_MAX_NONPRIV_SLOTS; i++) { + u32 expected = get_whitelist_reg(w, i); + u32 actual = results[i]; + + pr_info("RING_NONPRIV[%d]: expected 0x%08x, found 0x%08x\n", + i, expected, actual); + } +} + +static int check_whitelist(const struct whitelist *w, + struct i915_gem_context *ctx, + struct intel_engine_cs *engine) +{ + struct drm_i915_gem_object *results; + u32 *vaddr; + int err; + int i; + + results = read_nonprivs(ctx, engine); + if (IS_ERR(results)) + return PTR_ERR(results); + + err = i915_gem_object_set_to_cpu_domain(results, false); + if (err) + goto out_put; + + vaddr = i915_gem_object_pin_map(results, I915_MAP_WB); + if (IS_ERR(vaddr)) { + err = PTR_ERR(vaddr); + goto out_put; + } + + for (i = 0; i < RING_MAX_NONPRIV_SLOTS; i++) { + u32 expected = get_whitelist_reg(w, i); + u32 actual = vaddr[i]; + + if (expected != actual) { + print_results(w, vaddr); + pr_err("Invalid RING_NONPRIV[%d], expected 0x%08x, found 0x%08x\n", + i, expected, actual); + + err = -EINVAL; + break; + } + } + + i915_gem_object_unpin_map(results); +out_put: + i915_gem_object_put(results); + return err; +} + +static int do_device_reset(struct intel_engine_cs *engine) +{ + i915_reset(engine->i915, ENGINE_MASK(engine->id), NULL); + return 0; +} + +static int do_engine_reset(struct intel_engine_cs *engine) +{ + return i915_reset_engine(engine, NULL); +} + +static int switch_to_scratch_context(struct intel_engine_cs *engine) +{ + struct i915_gem_context *ctx; + struct i915_request *rq; + + ctx = kernel_context(engine->i915); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + rq = i915_request_alloc(engine, ctx); + kernel_context_close(ctx); + if (IS_ERR(rq)) + return PTR_ERR(rq); + + i915_request_add(rq); + + return 0; +} + +static int check_whitelist_across_reset(struct intel_engine_cs *engine, + int (*reset)(struct intel_engine_cs *), + const struct whitelist *w, + const char *name) +{ + struct i915_gem_context *ctx; + int err; + + ctx = kernel_context(engine->i915); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + err = check_whitelist(w, ctx, engine); + if (err) { + pr_err("Invalid whitelist *before* %s reset!\n", name); + goto out; + } + + err = switch_to_scratch_context(engine); + if (err) + goto out; + + err = reset(engine); + if (err) { + pr_err("%s reset failed\n", name); + goto out; + } + + err = check_whitelist(w, ctx, engine); + if (err) { + pr_err("Whitelist not preserved in context across %s reset!\n", + name); + goto out; + } + + kernel_context_close(ctx); + + ctx = kernel_context(engine->i915); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + err = check_whitelist(w, ctx, engine); + if (err) { + pr_err("Invalid whitelist *after* %s reset in fresh context!\n", + name); + goto out; + } + +out: + kernel_context_close(ctx); + return err; +} + +static int live_reset_whitelist(void *arg) +{ + struct drm_i915_private *i915 = arg; + struct intel_engine_cs *engine = i915->engine[RCS]; + struct i915_gpu_error *error = &i915->gpu_error; + struct whitelist w; + int err = 0; + + /* If we reset the gpu, we should not lose the RING_NONPRIV */ + + if (!engine) + return 0; + + if (!whitelist_build(engine, &w)) + return 0; + + pr_info("Checking %d whitelisted registers (RING_NONPRIV)\n", w.count); + + set_bit(I915_RESET_BACKOFF, &error->flags); + set_bit(I915_RESET_ENGINE + engine->id, &error->flags); + + if (intel_has_reset_engine(i915)) { + err = check_whitelist_across_reset(engine, + do_engine_reset, &w, + "engine"); + if (err) + goto out; + } + + if (intel_has_gpu_reset(i915)) { + err = check_whitelist_across_reset(engine, + do_device_reset, &w, + "device"); + if (err) + goto out; + } + +out: + clear_bit(I915_RESET_ENGINE + engine->id, &error->flags); + clear_bit(I915_RESET_BACKOFF, &error->flags); + return err; +} + +int intel_workarounds_live_selftests(struct drm_i915_private *i915) +{ + static const struct i915_subtest tests[] = { + SUBTEST(live_reset_whitelist), + }; + int err; + + mutex_lock(&i915->drm.struct_mutex); + err = i915_subtests(tests, i915); + mutex_unlock(&i915->drm.struct_mutex); + + return err; +} diff --git a/drivers/gpu/drm/i915/selftests/mock_engine.c b/drivers/gpu/drm/i915/selftests/mock_engine.c index 78a89ef..26bf29d 100644 --- a/drivers/gpu/drm/i915/selftests/mock_engine.c +++ b/drivers/gpu/drm/i915/selftests/mock_engine.c @@ -25,6 +25,11 @@ #include "mock_engine.h" #include "mock_request.h" +struct mock_ring { + struct intel_ring base; + struct i915_timeline timeline; +}; + static struct mock_request *first_request(struct mock_engine *engine) { return list_first_entry_or_null(&engine->hw_queue, @@ -71,14 +76,21 @@ static struct intel_ring * mock_context_pin(struct intel_engine_cs *engine, struct i915_gem_context *ctx) { - i915_gem_context_get(ctx); + struct intel_context *ce = to_intel_context(ctx, engine); + + if (!ce->pin_count++) + i915_gem_context_get(ctx); + return engine->buffer; } static void mock_context_unpin(struct intel_engine_cs *engine, struct i915_gem_context *ctx) { - i915_gem_context_put(ctx); + struct intel_context *ce = to_intel_context(ctx, engine); + + if (!--ce->pin_count) + i915_gem_context_put(ctx); } static int mock_request_alloc(struct i915_request *request) @@ -125,7 +137,7 @@ static void mock_submit_request(struct i915_request *request) static struct intel_ring *mock_ring(struct intel_engine_cs *engine) { const unsigned long sz = PAGE_SIZE / 2; - struct intel_ring *ring; + struct mock_ring *ring; BUILD_BUG_ON(MIN_SPACE_FOR_ADD_REQUEST > sz); @@ -133,14 +145,25 @@ static struct intel_ring *mock_ring(struct intel_engine_cs *engine) if (!ring) return NULL; - ring->size = sz; - ring->effective_size = sz; - ring->vaddr = (void *)(ring + 1); + i915_timeline_init(engine->i915, &ring->timeline, engine->name); + + ring->base.size = sz; + ring->base.effective_size = sz; + ring->base.vaddr = (void *)(ring + 1); + ring->base.timeline = &ring->timeline; - INIT_LIST_HEAD(&ring->request_list); - intel_ring_update_space(ring); + INIT_LIST_HEAD(&ring->base.request_list); + intel_ring_update_space(&ring->base); - return ring; + return &ring->base; +} + +static void mock_ring_free(struct intel_ring *base) +{ + struct mock_ring *ring = container_of(base, typeof(*ring), base); + + i915_timeline_fini(&ring->timeline); + kfree(ring); } struct intel_engine_cs *mock_engine(struct drm_i915_private *i915, @@ -155,12 +178,6 @@ struct intel_engine_cs *mock_engine(struct drm_i915_private *i915, if (!engine) return NULL; - engine->base.buffer = mock_ring(&engine->base); - if (!engine->base.buffer) { - kfree(engine); - return NULL; - } - /* minimal engine setup for requests */ engine->base.i915 = i915; snprintf(engine->base.name, sizeof(engine->base.name), "%s", name); @@ -174,9 +191,7 @@ struct intel_engine_cs *mock_engine(struct drm_i915_private *i915, engine->base.emit_breadcrumb = mock_emit_breadcrumb; engine->base.submit_request = mock_submit_request; - engine->base.timeline = - &i915->gt.global_timeline.engine[engine->base.id]; - + i915_timeline_init(i915, &engine->base.timeline, engine->base.name); intel_engine_init_breadcrumbs(&engine->base); engine->base.breadcrumbs.mock = true; /* prevent touching HW for irqs */ @@ -185,7 +200,17 @@ struct intel_engine_cs *mock_engine(struct drm_i915_private *i915, timer_setup(&engine->hw_delay, hw_delay_complete, 0); INIT_LIST_HEAD(&engine->hw_queue); + engine->base.buffer = mock_ring(&engine->base); + if (!engine->base.buffer) + goto err_breadcrumbs; + return &engine->base; + +err_breadcrumbs: + intel_engine_fini_breadcrumbs(&engine->base); + i915_timeline_fini(&engine->base.timeline); + kfree(engine); + return NULL; } void mock_engine_flush(struct intel_engine_cs *engine) @@ -217,10 +242,12 @@ void mock_engine_free(struct intel_engine_cs *engine) GEM_BUG_ON(timer_pending(&mock->hw_delay)); if (engine->last_retired_context) - engine->context_unpin(engine, engine->last_retired_context); + intel_context_unpin(engine->last_retired_context, engine); + + mock_ring_free(engine->buffer); intel_engine_fini_breadcrumbs(engine); + i915_timeline_fini(&engine->timeline); - kfree(engine->buffer); kfree(engine); } diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_device.c b/drivers/gpu/drm/i915/selftests/mock_gem_device.c index e6d4b88..94baedf 100644 --- a/drivers/gpu/drm/i915/selftests/mock_gem_device.c +++ b/drivers/gpu/drm/i915/selftests/mock_gem_device.c @@ -44,6 +44,7 @@ void mock_device_flush(struct drm_i915_private *i915) mock_engine_flush(engine); i915_retire_requests(i915); + GEM_BUG_ON(i915->gt.active_requests); } static void mock_device_release(struct drm_device *dev) @@ -72,8 +73,8 @@ static void mock_device_release(struct drm_device *dev) mutex_lock(&i915->drm.struct_mutex); mock_fini_ggtt(i915); - i915_gem_timeline_fini(&i915->gt.global_timeline); mutex_unlock(&i915->drm.struct_mutex); + WARN_ON(!list_empty(&i915->gt.timelines)); destroy_workqueue(i915->wq); @@ -223,26 +224,25 @@ struct drm_i915_private *mock_gem_device(void) if (!i915->priorities) goto err_dependencies; - mutex_lock(&i915->drm.struct_mutex); INIT_LIST_HEAD(&i915->gt.timelines); - err = i915_gem_timeline_init__global(i915); - if (err) { - mutex_unlock(&i915->drm.struct_mutex); - goto err_priorities; - } + INIT_LIST_HEAD(&i915->gt.active_rings); + INIT_LIST_HEAD(&i915->gt.closed_vma); + + mutex_lock(&i915->drm.struct_mutex); mock_init_ggtt(i915); - mutex_unlock(&i915->drm.struct_mutex); mkwrite_device_info(i915)->ring_mask = BIT(0); i915->engine[RCS] = mock_engine(i915, "mock", RCS); if (!i915->engine[RCS]) - goto err_priorities; + goto err_unlock; i915->kernel_context = mock_context(i915, NULL); if (!i915->kernel_context) goto err_engine; + mutex_unlock(&i915->drm.struct_mutex); + WARN_ON(i915_gemfs_init(i915)); return i915; @@ -250,7 +250,8 @@ struct drm_i915_private *mock_gem_device(void) err_engine: for_each_engine(engine, i915, id) mock_engine_free(engine); -err_priorities: +err_unlock: + mutex_unlock(&i915->drm.struct_mutex); kmem_cache_destroy(i915->priorities); err_dependencies: kmem_cache_destroy(i915->dependencies); diff --git a/drivers/gpu/drm/i915/selftests/mock_gtt.c b/drivers/gpu/drm/i915/selftests/mock_gtt.c index e96873f..36c1120 100644 --- a/drivers/gpu/drm/i915/selftests/mock_gtt.c +++ b/drivers/gpu/drm/i915/selftests/mock_gtt.c @@ -76,7 +76,6 @@ mock_ppgtt(struct drm_i915_private *i915, INIT_LIST_HEAD(&ppgtt->base.global_link); drm_mm_init(&ppgtt->base.mm, 0, ppgtt->base.total); - i915_gem_timeline_init(i915, &ppgtt->base.timeline, name); ppgtt->base.clear_range = nop_clear_range; ppgtt->base.insert_page = mock_insert_page; diff --git a/drivers/gpu/drm/i915/selftests/mock_timeline.c b/drivers/gpu/drm/i915/selftests/mock_timeline.c index 47b1f47..dcf3b16 100644 --- a/drivers/gpu/drm/i915/selftests/mock_timeline.c +++ b/drivers/gpu/drm/i915/selftests/mock_timeline.c @@ -1,45 +1,28 @@ /* - * Copyright © 2017 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. + * SPDX-License-Identifier: MIT * + * Copyright © 2017-2018 Intel Corporation */ +#include "../i915_timeline.h" + #include "mock_timeline.h" -struct intel_timeline *mock_timeline(u64 context) +void mock_timeline_init(struct i915_timeline *timeline, u64 context) { - static struct lock_class_key class; - struct intel_timeline *tl; + timeline->fence_context = context; + + spin_lock_init(&timeline->lock); - tl = kzalloc(sizeof(*tl), GFP_KERNEL); - if (!tl) - return NULL; + init_request_active(&timeline->last_request, NULL); + INIT_LIST_HEAD(&timeline->requests); - __intel_timeline_init(tl, NULL, context, &class, "mock"); + i915_syncmap_init(&timeline->sync); - return tl; + INIT_LIST_HEAD(&timeline->link); } -void mock_timeline_destroy(struct intel_timeline *tl) +void mock_timeline_fini(struct i915_timeline *timeline) { - __intel_timeline_fini(tl); - kfree(tl); + i915_timeline_fini(timeline); } diff --git a/drivers/gpu/drm/i915/selftests/mock_timeline.h b/drivers/gpu/drm/i915/selftests/mock_timeline.h index c27ff46..b6deaa6 100644 --- a/drivers/gpu/drm/i915/selftests/mock_timeline.h +++ b/drivers/gpu/drm/i915/selftests/mock_timeline.h @@ -1,33 +1,15 @@ /* - * Copyright © 2017 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. + * SPDX-License-Identifier: MIT * + * Copyright © 2017-2018 Intel Corporation */ #ifndef __MOCK_TIMELINE__ #define __MOCK_TIMELINE__ -#include "../i915_gem_timeline.h" +struct i915_timeline; -struct intel_timeline *mock_timeline(u64 context); -void mock_timeline_destroy(struct intel_timeline *tl); +void mock_timeline_init(struct i915_timeline *timeline, u64 context); +void mock_timeline_fini(struct i915_timeline *timeline); #endif /* !__MOCK_TIMELINE__ */ |