diff options
Diffstat (limited to 'drivers/gpu/drm/i915')
| -rw-r--r-- | drivers/gpu/drm/i915/display/intel_display.c | 11 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/display/intel_display.h | 1 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/display/intel_fbc.c | 65 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/display/intel_hdmi.c | 10 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/gt/intel_context.c | 12 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/gt/intel_lrc.c | 19 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/gt/selftest_rps.c | 8 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/gt/shaders/README | 46 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm | 119 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm | 117 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/gvt/debugfs.c | 2 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/gvt/handlers.c | 24 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/gvt/mmio_context.h | 6 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/gvt/reg.h | 5 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/i915_debugfs.c | 2 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/i915_drv.h | 7 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/i915_perf.c | 1 | ||||
| -rw-r--r-- | drivers/gpu/drm/i915/i915_vma.c | 15 |
18 files changed, 381 insertions, 89 deletions
diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c index 9ea1a397d1b5..26996e1839e2 100644 --- a/drivers/gpu/drm/i915/display/intel_display.c +++ b/drivers/gpu/drm/i915/display/intel_display.c @@ -3822,6 +3822,17 @@ skl_check_main_ccs_coordinates(struct intel_plane_state *plane_state, return true; } +unsigned int +intel_plane_fence_y_offset(const struct intel_plane_state *plane_state) +{ + int x = 0, y = 0; + + intel_plane_adjust_aligned_offset(&x, &y, plane_state, 0, + plane_state->color_plane[0].offset, 0); + + return y; +} + static int skl_check_main_surface(struct intel_plane_state *plane_state) { struct drm_i915_private *dev_priv = to_i915(plane_state->uapi.plane->dev); diff --git a/drivers/gpu/drm/i915/display/intel_display.h b/drivers/gpu/drm/i915/display/intel_display.h index efb4da205ea2..3a06f72c9859 100644 --- a/drivers/gpu/drm/i915/display/intel_display.h +++ b/drivers/gpu/drm/i915/display/intel_display.h @@ -608,6 +608,7 @@ unsigned int i9xx_plane_max_stride(struct intel_plane *plane, u32 pixel_format, u64 modifier, unsigned int rotation); int bdw_get_pipemisc_bpp(struct intel_crtc *crtc); +unsigned int intel_plane_fence_y_offset(const struct intel_plane_state *plane_state); struct intel_display_error_state * intel_display_capture_error_state(struct drm_i915_private *dev_priv); diff --git a/drivers/gpu/drm/i915/display/intel_fbc.c b/drivers/gpu/drm/i915/display/intel_fbc.c index 1c26673acb2d..412572f88b67 100644 --- a/drivers/gpu/drm/i915/display/intel_fbc.c +++ b/drivers/gpu/drm/i915/display/intel_fbc.c @@ -48,19 +48,6 @@ #include "intel_frontbuffer.h" /* - * In some platforms where the CRTC's x:0/y:0 coordinates doesn't match the - * frontbuffer's x:0/y:0 coordinates we lie to the hardware about the plane's - * origin so the x and y offsets can actually fit the registers. As a - * consequence, the fence doesn't really start exactly at the display plane - * address we program because it starts at the real start of the buffer, so we - * have to take this into consideration here. - */ -static unsigned int get_crtc_fence_y_offset(struct intel_fbc *fbc) -{ - return fbc->state_cache.plane.y - fbc->state_cache.plane.adjusted_y; -} - -/* * For SKL+, the plane source size used by the hardware is based on the value we * write to the PLANE_SIZE register. For BDW-, the hardware looks at the value * we wrote to PIPESRC. @@ -141,7 +128,7 @@ static void i8xx_fbc_activate(struct drm_i915_private *dev_priv) fbc_ctl2 |= FBC_CTL_CPU_FENCE; intel_de_write(dev_priv, FBC_CONTROL2, fbc_ctl2); intel_de_write(dev_priv, FBC_FENCE_OFF, - params->crtc.fence_y_offset); + params->fence_y_offset); } /* enable it... */ @@ -175,7 +162,7 @@ static void g4x_fbc_activate(struct drm_i915_private *dev_priv) if (params->fence_id >= 0) { dpfc_ctl |= DPFC_CTL_FENCE_EN | params->fence_id; intel_de_write(dev_priv, DPFC_FENCE_YOFF, - params->crtc.fence_y_offset); + params->fence_y_offset); } else { intel_de_write(dev_priv, DPFC_FENCE_YOFF, 0); } @@ -243,7 +230,7 @@ static void ilk_fbc_activate(struct drm_i915_private *dev_priv) intel_de_write(dev_priv, SNB_DPFC_CTL_SA, SNB_CPU_FENCE_ENABLE | params->fence_id); intel_de_write(dev_priv, DPFC_CPU_FENCE_OFFSET, - params->crtc.fence_y_offset); + params->fence_y_offset); } } else { if (IS_GEN(dev_priv, 6)) { @@ -253,7 +240,7 @@ static void ilk_fbc_activate(struct drm_i915_private *dev_priv) } intel_de_write(dev_priv, ILK_DPFC_FENCE_YOFF, - params->crtc.fence_y_offset); + params->fence_y_offset); /* enable it... */ intel_de_write(dev_priv, ILK_DPFC_CONTROL, dpfc_ctl | DPFC_CTL_EN); @@ -320,7 +307,7 @@ static void gen7_fbc_activate(struct drm_i915_private *dev_priv) intel_de_write(dev_priv, SNB_DPFC_CTL_SA, SNB_CPU_FENCE_ENABLE | params->fence_id); intel_de_write(dev_priv, DPFC_CPU_FENCE_OFFSET, - params->crtc.fence_y_offset); + params->fence_y_offset); } else if (dev_priv->ggtt.num_fences) { intel_de_write(dev_priv, SNB_DPFC_CTL_SA, 0); intel_de_write(dev_priv, DPFC_CPU_FENCE_OFFSET, 0); @@ -631,8 +618,8 @@ static bool rotation_is_valid(struct drm_i915_private *dev_priv, /* * For some reason, the hardware tracking starts looking at whatever we * programmed as the display plane base address register. It does not look at - * the X and Y offset registers. That's why we look at the crtc->adjusted{x,y} - * variables instead of just looking at the pipe/plane size. + * the X and Y offset registers. That's why we include the src x/y offsets + * instead of just looking at the plane size. */ static bool intel_fbc_hw_tracking_covers_screen(struct intel_crtc *crtc) { @@ -705,7 +692,6 @@ static void intel_fbc_update_state_cache(struct intel_crtc *crtc, cache->plane.src_h = drm_rect_height(&plane_state->uapi.src) >> 16; cache->plane.adjusted_x = plane_state->color_plane[0].x; cache->plane.adjusted_y = plane_state->color_plane[0].y; - cache->plane.y = plane_state->uapi.src.y1 >> 16; cache->plane.pixel_blend_mode = plane_state->hw.pixel_blend_mode; @@ -713,6 +699,8 @@ static void intel_fbc_update_state_cache(struct intel_crtc *crtc, cache->fb.stride = fb->pitches[0]; cache->fb.modifier = fb->modifier; + cache->fence_y_offset = intel_plane_fence_y_offset(plane_state); + drm_WARN_ON(&dev_priv->drm, plane_state->flags & PLANE_HAS_FENCE && !plane_state->vma->fence); @@ -731,6 +719,25 @@ static bool intel_fbc_cfb_size_changed(struct drm_i915_private *dev_priv) fbc->compressed_fb.size * fbc->threshold; } +static u16 intel_fbc_gen9_wa_cfb_stride(struct drm_i915_private *dev_priv) +{ + struct intel_fbc *fbc = &dev_priv->fbc; + struct intel_fbc_state_cache *cache = &fbc->state_cache; + + if ((IS_GEN9_BC(dev_priv) || IS_BROXTON(dev_priv)) && + cache->fb.modifier != I915_FORMAT_MOD_X_TILED) + return DIV_ROUND_UP(cache->plane.src_w, 32 * fbc->threshold) * 8; + else + return 0; +} + +static bool intel_fbc_gen9_wa_cfb_stride_changed(struct drm_i915_private *dev_priv) +{ + struct intel_fbc *fbc = &dev_priv->fbc; + + return fbc->params.gen9_wa_cfb_stride != intel_fbc_gen9_wa_cfb_stride(dev_priv); +} + static bool intel_fbc_can_enable(struct drm_i915_private *dev_priv) { struct intel_fbc *fbc = &dev_priv->fbc; @@ -883,12 +890,13 @@ static void intel_fbc_get_reg_params(struct intel_crtc *crtc, memset(params, 0, sizeof(*params)); params->fence_id = cache->fence_id; + params->fence_y_offset = cache->fence_y_offset; params->crtc.pipe = crtc->pipe; params->crtc.i9xx_plane = to_intel_plane(crtc->base.primary)->i9xx_plane; - params->crtc.fence_y_offset = get_crtc_fence_y_offset(fbc); params->fb.format = cache->fb.format; + params->fb.modifier = cache->fb.modifier; params->fb.stride = cache->fb.stride; params->cfb_size = intel_fbc_calculate_cfb_size(dev_priv, cache); @@ -918,6 +926,9 @@ static bool intel_fbc_can_flip_nuke(const struct intel_crtc_state *crtc_state) if (params->fb.format != cache->fb.format) return false; + if (params->fb.modifier != cache->fb.modifier) + return false; + if (params->fb.stride != cache->fb.stride) return false; @@ -1197,7 +1208,8 @@ void intel_fbc_enable(struct intel_atomic_state *state, if (fbc->crtc) { if (fbc->crtc != crtc || - !intel_fbc_cfb_size_changed(dev_priv)) + (!intel_fbc_cfb_size_changed(dev_priv) && + !intel_fbc_gen9_wa_cfb_stride_changed(dev_priv))) goto out; __intel_fbc_disable(dev_priv); @@ -1219,12 +1231,7 @@ void intel_fbc_enable(struct intel_atomic_state *state, goto out; } - if ((IS_GEN9_BC(dev_priv) || IS_BROXTON(dev_priv)) && - plane_state->hw.fb->modifier != I915_FORMAT_MOD_X_TILED) - cache->gen9_wa_cfb_stride = - DIV_ROUND_UP(cache->plane.src_w, 32 * fbc->threshold) * 8; - else - cache->gen9_wa_cfb_stride = 0; + cache->gen9_wa_cfb_stride = intel_fbc_gen9_wa_cfb_stride(dev_priv); drm_dbg_kms(&dev_priv->drm, "Enabling FBC on pipe %c\n", pipe_name(crtc->pipe)); diff --git a/drivers/gpu/drm/i915/display/intel_hdmi.c b/drivers/gpu/drm/i915/display/intel_hdmi.c index 010f37240710..95b6d9457910 100644 --- a/drivers/gpu/drm/i915/display/intel_hdmi.c +++ b/drivers/gpu/drm/i915/display/intel_hdmi.c @@ -2867,19 +2867,13 @@ intel_hdmi_connector_register(struct drm_connector *connector) return ret; } -static void intel_hdmi_destroy(struct drm_connector *connector) +static void intel_hdmi_connector_unregister(struct drm_connector *connector) { struct cec_notifier *n = intel_attached_hdmi(to_intel_connector(connector))->cec_notifier; cec_notifier_conn_unregister(n); - intel_connector_destroy(connector); -} - -static void intel_hdmi_connector_unregister(struct drm_connector *connector) -{ intel_hdmi_remove_i2c_symlink(connector); - intel_connector_unregister(connector); } @@ -2891,7 +2885,7 @@ static const struct drm_connector_funcs intel_hdmi_connector_funcs = { .atomic_set_property = intel_digital_connector_atomic_set_property, .late_register = intel_hdmi_connector_register, .early_unregister = intel_hdmi_connector_unregister, - .destroy = intel_hdmi_destroy, + .destroy = intel_connector_destroy, .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, .atomic_duplicate_state = intel_digital_connector_duplicate_state, }; diff --git a/drivers/gpu/drm/i915/gt/intel_context.c b/drivers/gpu/drm/i915/gt/intel_context.c index e4aece20bc80..52db2bde44a3 100644 --- a/drivers/gpu/drm/i915/gt/intel_context.c +++ b/drivers/gpu/drm/i915/gt/intel_context.c @@ -204,25 +204,25 @@ static int __ring_active(struct intel_ring *ring) { int err; - err = i915_active_acquire(&ring->vma->active); + err = intel_ring_pin(ring); if (err) return err; - err = intel_ring_pin(ring); + err = i915_active_acquire(&ring->vma->active); if (err) - goto err_active; + goto err_pin; return 0; -err_active: - i915_active_release(&ring->vma->active); +err_pin: + intel_ring_unpin(ring); return err; } static void __ring_retire(struct intel_ring *ring) { - intel_ring_unpin(ring); i915_active_release(&ring->vma->active); + intel_ring_unpin(ring); } __i915_active_call diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c index 7c3d8ef4a47c..cb07e1d2a353 100644 --- a/drivers/gpu/drm/i915/gt/intel_lrc.c +++ b/drivers/gpu/drm/i915/gt/intel_lrc.c @@ -5396,13 +5396,8 @@ static void virtual_engine_initial_hint(struct virtual_engine *ve) * typically be the first we inspect for submission. */ swp = prandom_u32_max(ve->num_siblings); - if (!swp) - return; - - swap(ve->siblings[swp], ve->siblings[0]); - if (!intel_engine_has_relative_mmio(ve->siblings[0])) - virtual_update_register_offsets(ve->context.lrc_reg_state, - ve->siblings[0]); + if (swp) + swap(ve->siblings[swp], ve->siblings[0]); } static int virtual_context_alloc(struct intel_context *ce) @@ -5415,15 +5410,9 @@ static int virtual_context_alloc(struct intel_context *ce) static int virtual_context_pin(struct intel_context *ce) { struct virtual_engine *ve = container_of(ce, typeof(*ve), context); - int err; /* Note: we must use a real engine class for setting up reg state */ - err = __execlists_context_pin(ce, ve->siblings[0]); - if (err) - return err; - - virtual_engine_initial_hint(ve); - return 0; + return __execlists_context_pin(ce, ve->siblings[0]); } static void virtual_context_enter(struct intel_context *ce) @@ -5688,6 +5677,7 @@ intel_execlists_create_virtual(struct intel_engine_cs **siblings, intel_engine_init_active(&ve->base, ENGINE_VIRTUAL); intel_engine_init_breadcrumbs(&ve->base); intel_engine_init_execlists(&ve->base); + ve->base.breadcrumbs.irq_armed = true; /* fake HW, used for irq_work */ ve->base.cops = &virtual_context_ops; ve->base.request_alloc = execlists_request_alloc; @@ -5769,6 +5759,7 @@ intel_execlists_create_virtual(struct intel_engine_cs **siblings, ve->base.flags |= I915_ENGINE_IS_VIRTUAL; + virtual_engine_initial_hint(ve); return &ve->context; err_put: diff --git a/drivers/gpu/drm/i915/gt/selftest_rps.c b/drivers/gpu/drm/i915/gt/selftest_rps.c index 5049c3dd08a6..c91981e75ebf 100644 --- a/drivers/gpu/drm/i915/gt/selftest_rps.c +++ b/drivers/gpu/drm/i915/gt/selftest_rps.c @@ -44,9 +44,9 @@ static int cmp_u64(const void *A, const void *B) { const u64 *a = A, *b = B; - if (a < b) + if (*a < *b) return -1; - else if (a > b) + else if (*a > *b) return 1; else return 0; @@ -56,9 +56,9 @@ static int cmp_u32(const void *A, const void *B) { const u32 *a = A, *b = B; - if (a < b) + if (*a < *b) return -1; - else if (a > b) + else if (*a > *b) return 1; else return 0; diff --git a/drivers/gpu/drm/i915/gt/shaders/README b/drivers/gpu/drm/i915/gt/shaders/README new file mode 100644 index 000000000000..e7e96d7073c7 --- /dev/null +++ b/drivers/gpu/drm/i915/gt/shaders/README @@ -0,0 +1,46 @@ +ASM sources for auto generated shaders +====================================== + +The i915/gt/hsw_clear_kernel.c and i915/gt/ivb_clear_kernel.c files contain +pre-compiled batch chunks that will clear any residual render cache during +context switch. + +They are generated from their respective platform ASM files present on +i915/gt/shaders/clear_kernel directory. + +The generated .c files should never be modified directly. Instead, any modification +needs to be done on the on their respective ASM files and build instructions below +needes to be followed. + +Building +======== + +Environment +----------- + +IGT GPU tool scripts and the Mesa's i965 instruction assembler tool are used +on building. + +Please make sure your Mesa tool is compiled with "-Dtools=intel" and +"-Ddri-drivers=i965", and run this script from IGT source root directory" + +The instructions bellow assume: + * IGT gpu tools source code is located on your home directory (~) as ~/igt + * Mesa source code is located on your home directory (~) as ~/mesa + and built under the ~/mesa/build directory + * Linux kernel source code is under your home directory (~) as ~/linux + +Instructions +------------ + +~ $ cp ~/linux/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm \ + ~/igt/lib/i915/shaders/clear_kernel/ivb.asm +~ $ cd ~/igt +igt $ ./scripts/generate_clear_kernel.sh -g ivb \ + -m ~/mesa/build/src/intel/tools/i965_asm + +~ $ cp ~/linux/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm \ + ~/igt/lib/i915/shaders/clear_kernel/hsw.asm +~ $ cd ~/igt +igt $ ./scripts/generate_clear_kernel.sh -g hsw \ + -m ~/mesa/build/src/intel/tools/i965_asm
\ No newline at end of file diff --git a/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm b/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm new file mode 100644 index 000000000000..5fdf384bb621 --- /dev/null +++ b/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm @@ -0,0 +1,119 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright © 2020 Intel Corporation + */ + +/* + * Kernel for PAVP buffer clear. + * + * 1. Clear all 64 GRF registers assigned to the kernel with designated value; + * 2. Write 32x16 block of all "0" to render target buffer which indirectly clears + * 512 bytes of Render Cache. + */ + +/* Store designated "clear GRF" value */ +mov(1) f0.1<1>UW g1.2<0,1,0>UW { align1 1N }; + +/** + * Curbe Format + * + * DW 1.0 - Block Offset to write Render Cache + * DW 1.1 [15:0] - Clear Word + * DW 1.2 - Delay iterations + * DW 1.3 - Enable Instrumentation (only for debug) + * DW 1.4 - Rsvd (intended for context ID) + * DW 1.5 - [31:16]:SliceCount, [15:0]:SubSlicePerSliceCount + * DW 1.6 - Rsvd MBZ (intended for Enable Wait on Total Thread Count) + * DW 1.7 - Rsvd MBZ (inteded for Total Thread Count) + * + * Binding Table + * + * BTI 0: 2D Surface to help clear L3 (Render/Data Cache) + * BTI 1: Wait/Instrumentation Buffer + * Size : (SliceCount * SubSliceCount * 16 EUs/SubSlice) rows * (16 threads/EU) cols (Format R32_UINT) + * Expected to be initialized to 0 by driver/another kernel + * Layout: + * RowN: Histogram for EU-N: (SliceID*SubSlicePerSliceCount + SSID)*16 + EUID [assume max 16 EUs / SS] + * Col-k[DW-k]: Threads Executed on ThreadID-k for EU-N + */ +add(1) g1.2<1>UD g1.2<0,1,0>UD 0x00000001UD { align1 1N }; /* Loop count to delay kernel: Init to (g1.2 + 1) */ +cmp.z.f0.0(1) null<1>UD g1.3<0,1,0>UD 0x00000000UD { align1 1N }; +(+f0.0) jmpi(1) 352D { align1 WE_all 1N }; + +/** + * State Register has info on where this thread is running + * IVB: sr0.0 :: [15:13]: MBZ, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID + * HSW: sr0.0 :: 15: MBZ, [14:13]: SliceID, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID + */ +mov(8) g3<1>UD 0x00000000UD { align1 1Q }; +shr(1) g3<1>D sr0<0,1,0>D 12D { align1 1N }; +and(1) g3<1>D g3<0,1,0>D 1D { align1 1N }; /* g3 has HSID */ +shr(1) g3.1<1>D sr0<0,1,0>D 13D { align1 1N }; +and(1) g3.1<1>D g3.1<0,1,0>D 3D { align1 1N }; /* g3.1 has sliceID */ +mul(1) g3.5<1>D g3.1<0,1,0>D g1.10<0,1,0>UW { align1 1N }; +add(1) g3<1>D g3<0,1,0>D g3.5<0,1,0>D { align1 1N }; /* g3 = sliceID * SubSlicePerSliceCount + HSID */ +shr(1) g3.2<1>D sr0<0,1,0>D 8D { align1 1N }; +and(1) g3.2<1>D g3.2<0,1,0>D 15D { align1 1N }; /* g3.2 = EUID */ +mul(1) g3.4<1>D g3<0,1,0>D 16D { align1 1N }; +add(1) g3.2<1>D g3.2<0,1,0>D g3.4<0,1,0>D { align1 1N }; /* g3.2 now points to EU row number (Y-pixel = V address ) in instrumentation surf */ + +mov(8) g5<1>UD 0x00000000UD { align1 1Q }; +and(1) g3.3<1>D sr0<0,1,0>D 7D { align1 1N }; +mul(1) g3.3<1>D g3.3<0,1,0>D 4D { align1 1N }; + +mov(8) g4<1>UD g0<8,8,1>UD { align1 1Q }; /* Initialize message header with g0 */ +mov(1) g4<1>UD g3.3<0,1,0>UD { align1 1N }; /* Block offset */ +mov(1) g4.1<1>UD g3.2<0,1,0>UD { align1 1N }; /* Block offset */ +mov(1) g4.2<1>UD 0x00000003UD { align1 1N }; /* Block size (1 row x 4 bytes) */ +and(1) g4.3<1>UD g4.3<0,1,0>UW 0xffffffffUD { align1 1N }; + +/* Media block read to fetch current value at specified location in instrumentation buffer */ +sendc(8) g5<1>UD g4<8,8,1>F 0x02190001 + + render MsgDesc: media block read MsgCtrl = 0x0 Surface = 1 mlen 1 rlen 1 { align1 1Q }; +add(1) g5<1>D g5<0,1,0>D 1D { align1 1N }; + +/* Media block write for updated value at specified location in instrumentation buffer */ +sendc(8) g5<1>UD g4<8,8,1>F 0x040a8001 + render MsgDesc: media block write MsgCtrl = 0x0 Surface = 1 mlen 2 rlen 0 { align1 1Q }; + +/* Delay thread for specified parameter */ +add.nz.f0.0(1) g1.2<1>UD g1.2<0,1,0>UD -1D { align1 1N }; +(+f0.0) jmpi(1) -32D { align1 WE_all 1N }; + +/* Store designated "clear GRF" value */ +mov(1) f0.1<1>UW g1.2<0,1,0>UW { align1 1N }; + +/* Initialize looping parameters */ +mov(1) a0<1>D 0D { align1 1N }; /* Initialize a0.0:w=0 */ +mov(1) a0.4<1>W 127W { align1 1N }; /* Loop count. Each loop contains 16 GRF's */ + +/* Write 32x16 all "0" block */ +mov(8) g2<1>UD g0<8,8,1>UD { align1 1Q }; +mov(8) g127<1>UD g0<8,8,1>UD { align1 1Q }; +mov(2) g2<1>UD g1<2,2,1>UW { align1 1N }; +mov(1) g2.2<1>UD 0x000f000fUD { align1 1N }; /* Block size (16x16) */ +and(1) g2.3<1>UD g2.3<0,1,0>UW 0xffffffefUD { align1 1N }; +mov(16) g3<1>UD 0x00000000UD { align1 1H }; +mov(16) g4<1>UD 0x00000000UD { align1 1H }; +mov(16) g5<1>UD 0x00000000UD { align1 1H }; +mov(16) g6<1>UD 0x00000000UD { align1 1H }; +mov(16) g7<1>UD 0x00000000UD { align1 1H }; +mov(16) g8<1>UD 0x00000000UD { align1 1H }; +mov(16) g9<1>UD 0x00000000UD { align1 1H }; +mov(16) g10<1>UD 0x00000000UD { align1 1H }; +sendc(8) null<1>UD g2<8,8,1>F 0x120a8000 + render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q }; +add(1) g2<1>UD g1<0,1,0>UW 0x0010UW { align1 1N }; +sendc(8) null<1>UD g2<8,8,1>F 0x120a8000 + render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q }; + +/* Now, clear all GRF registers */ +add.nz.f0.0(1) a0.4<1>W a0.4<0,1,0>W -1W { align1 1N }; +mov(16) g[a0]<1>UW f0.1<0,1,0>UW { align1 1H }; +add(1) a0<1>D a0<0,1,0>D 32D { align1 1N }; +(+f0.0) jmpi(1) -64D { align1 WE_all 1N }; + +/* Terminante the thread */ +sendc(8) null<1>UD g127<8,8,1>F 0x82000010 + thread_spawner MsgDesc: mlen 1 rlen 0 { align1 1Q EOT }; diff --git a/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm b/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm new file mode 100644 index 000000000000..97c7ac9e3854 --- /dev/null +++ b/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm @@ -0,0 +1,117 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright © 2020 Intel Corporation + */ + +/* + * Kernel for PAVP buffer clear. + * + * 1. Clear all 64 GRF registers assigned to the kernel with designated value; + * 2. Write 32x16 block of all "0" to render target buffer which indirectly clears + * 512 bytes of Render Cache. + */ + +/* Store designated "clear GRF" value */ +mov(1) f0.1<1>UW g1.2<0,1,0>UW { align1 1N }; + +/** + * Curbe Format + * + * DW 1.0 - Block Offset to write Render Cache + * DW 1.1 [15:0] - Clear Word + * DW 1.2 - Delay iterations + * DW 1.3 - Enable Instrumentation (only for debug) + * DW 1.4 - Rsvd (intended for context ID) + * DW 1.5 - [31:16]:SliceCount, [15:0]:SubSlicePerSliceCount + * DW 1.6 - Rsvd MBZ (intended for Enable Wait on Total Thread Count) + * DW 1.7 - Rsvd MBZ (inteded for Total Thread Count) + * + * Binding Table + * + * BTI 0: 2D Surface to help clear L3 (Render/Data Cache) + * BTI 1: Wait/Instrumentation Buffer + * Size : (SliceCount * SubSliceCount * 16 EUs/SubSlice) rows * (16 threads/EU) cols (Format R32_UINT) + * Expected to be initialized to 0 by driver/another kernel + * Layout : + * RowN: Histogram for EU-N: (SliceID*SubSlicePerSliceCount + SSID)*16 + EUID [assume max 16 EUs / SS] + * Col-k[DW-k]: Threads Executed on ThreadID-k for EU-N + */ +add(1) g1.2<1>UD g1.2<0,1,0>UD 0x00000001UD { align1 1N }; /* Loop count to delay kernel: Init to (g1.2 + 1) */ +cmp.z.f0.0(1) null<1>UD g1.3<0,1,0>UD 0x00000000UD { align1 1N }; +(+f0.0) jmpi(1) 44D { align1 WE_all 1N }; + +/** + * State Register has info on where this thread is running + * IVB: sr0.0 :: [15:13]: MBZ, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID + * HSW: sr0.0 :: 15: MBZ, [14:13]: SliceID, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID + */ +mov(8) g3<1>UD 0x00000000UD { align1 1Q }; +shr(1) g3<1>D sr0<0,1,0>D 12D { align1 1N }; +and(1) g3<1>D g3<0,1,0>D 1D { align1 1N }; /* g3 has HSID */ +shr(1) g3.1<1>D sr0<0,1,0>D 13D { align1 1N }; +and(1) g3.1<1>D g3.1<0,1,0>D 3D { align1 1N }; /* g3.1 has sliceID */ +mul(1) g3.5<1>D g3.1<0,1,0>D g1.10<0,1,0>UW { align1 1N }; +add(1) g3<1>D g3<0,1,0>D g3.5<0,1,0>D { align1 1N }; /* g3 = sliceID * SubSlicePerSliceCount + HSID */ +shr(1) g3.2<1>D sr0<0,1,0>D 8D { align1 1N }; +and(1) g3.2<1>D g3.2<0,1,0>D 15D { align1 1N }; /* g3.2 = EUID */ +mul(1) g3.4<1>D g3<0,1,0>D 16D { align1 1N }; +add(1) g3.2<1>D g3.2<0,1,0>D g3.4<0,1,0>D { align1 1N }; /* g3.2 now points to EU row number (Y-pixel = V address ) in instrumentation surf */ + +mov(8) g5<1>UD 0x00000000UD { align1 1Q }; +and(1) g3.3<1>D sr0<0,1,0>D 7D { align1 1N }; +mul(1) g3.3<1>D g3.3<0,1,0>D 4D { align1 1N }; + +mov(8) g4<1>UD g0<8,8,1>UD { align1 1Q }; /* Initialize message header with g0 */ +mov(1) g4<1>UD g3.3<0,1,0>UD { align1 1N }; /* Block offset */ +mov(1) g4.1<1>UD g3.2<0,1,0>UD { align1 1N }; /* Block offset */ +mov(1) g4.2<1>UD 0x00000003UD { align1 1N }; /* Block size (1 row x 4 bytes) */ +and(1) g4.3<1>UD g4.3<0,1,0>UW 0xffffffffUD { align1 1N }; + +/* Media block read to fetch current value at specified location in instrumentation buffer */ +sendc(8) g5<1>UD g4<8,8,1>F 0x02190001 + render MsgDesc: media block read MsgCtrl = 0x0 Surface = 1 mlen 1 rlen 1 { align1 1Q }; +add(1) g5<1>D g5<0,1,0>D 1D { align1 1N }; + +/* Media block write for updated value at specified location in instrumentation buffer */ +sendc(8) g5<1>UD g4<8,8,1>F 0x040a8001 + render MsgDesc: media block write MsgCtrl = 0x0 Surface = 1 mlen 2 rlen 0 { align1 1Q }; +/* Delay thread for specified parameter */ +add.nz.f0.0(1) g1.2<1>UD g1.2<0,1,0>UD -1D { align1 1N }; +(+f0.0) jmpi(1) -4D { align1 WE_all 1N }; + +/* Store designated "clear GRF" value */ +mov(1) f0.1<1>UW g1.2<0,1,0>UW { align1 1N }; + +/* Initialize looping parameters */ +mov(1) a0<1>D 0D { align1 1N }; /* Initialize a0.0:w=0 */ +mov(1) a0.4<1>W 127W { align1 1N }; /* Loop count. Each loop contains 16 GRF's */ + +/* Write 32x16 all "0" block */ +mov(8) g2<1>UD g0<8,8,1>UD { align1 1Q }; +mov(8) g127<1>UD g0<8,8,1>UD { align1 1Q }; +mov(2) g2<1>UD g1<2,2,1>UW { align1 1N }; +mov(1) g2.2<1>UD 0x000f000fUD { align1 1N }; /* Block size (16x16) */ +and(1) g2.3<1>UD g2.3<0,1,0>UW 0xffffffefUD { align1 1N }; +mov(16) g3<1>UD 0x00000000UD { align1 1H }; +mov(16) g4<1>UD 0x00000000UD { align1 1H }; +mov(16) g5<1>UD 0x00000000UD { align1 1H }; +mov(16) g6<1>UD 0x00000000UD { align1 1H }; +mov(16) g7<1>UD 0x00000000UD { align1 1H }; +mov(16) g8<1>UD 0x00000000UD { align1 1H }; +mov(16) g9<1>UD 0x00000000UD { align1 1H }; +mov(16) g10<1>UD 0x00000000UD { align1 1H }; +sendc(8) null<1>UD g2<8,8,1>F 0x120a8000 + render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q }; +add(1) g2<1>UD g1<0,1,0>UW 0x0010UW { align1 1N }; +sendc(8) null<1>UD g2<8,8,1>F 0x120a8000 + render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q }; + +/* Now, clear all GRF registers */ +add.nz.f0.0(1) a0.4<1>W a0.4<0,1,0>W -1W { align1 1N }; +mov(16) g[a0]<1>UW f0.1<0,1,0>UW { align1 1H }; +add(1) a0<1>D a0<0,1,0>D 32D { align1 1N }; +(+f0.0) jmpi(1) -8D { align1 WE_all 1N }; + +/* Terminante the thread */ +sendc(8) null<1>UD g127<8,8,1>F 0x82000010 + thread_spawner MsgDesc: mlen 1 rlen 0 { align1 1Q EOT }; diff --git a/drivers/gpu/drm/i915/gvt/debugfs.c b/drivers/gpu/drm/i915/gvt/debugfs.c index ec47d4114554..62e6a14ad58e 100644 --- a/drivers/gpu/drm/i915/gvt/debugfs.c +++ b/drivers/gpu/drm/i915/gvt/debugfs.c @@ -66,7 +66,7 @@ static inline int mmio_diff_handler(struct intel_gvt *gvt, vreg = vgpu_vreg(param->vgpu, offset); if (preg != vreg) { - node = kmalloc(sizeof(*node), GFP_KERNEL); + node = kmalloc(sizeof(*node), GFP_ATOMIC); if (!node) return -ENOMEM; diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index 3e88e3b5c43a..fadd2adb8030 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -1726,13 +1726,13 @@ static int ring_mode_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, (*(u32 *)p_data) &= ~_MASKED_BIT_ENABLE(2); write_vreg(vgpu, offset, p_data, bytes); - if (data & _MASKED_BIT_ENABLE(1)) { + if (IS_MASKED_BITS_ENABLED(data, 1)) { enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST); return 0; } if (IS_COFFEELAKE(vgpu->gvt->gt->i915) && - data & _MASKED_BIT_ENABLE(2)) { + IS_MASKED_BITS_ENABLED(data, 2)) { enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST); return 0; } @@ -1741,14 +1741,14 @@ static int ring_mode_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, * pvinfo, if not, we will treat this guest as non-gvtg-aware * guest, and stop emulating its cfg space, mmio, gtt, etc. */ - if (((data & _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE)) || - (data & _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE))) - && !vgpu->pv_notified) { + if ((IS_MASKED_BITS_ENABLED(data, GFX_PPGTT_ENABLE) || + IS_MASKED_BITS_ENABLED(data, GFX_RUN_LIST_ENABLE)) && + !vgpu->pv_notified) { enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST); return 0; } - if ((data & _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE)) - || (data & _MASKED_BIT_DISABLE(GFX_RUN_LIST_ENABLE))) { + if (IS_MASKED_BITS_ENABLED(data, GFX_RUN_LIST_ENABLE) || + IS_MASKED_BITS_DISABLED(data, GFX_RUN_LIST_ENABLE)) { enable_execlist = !!(data & GFX_RUN_LIST_ENABLE); gvt_dbg_core("EXECLIST %s on ring %s\n", @@ -1809,7 +1809,7 @@ static int ring_reset_ctl_write(struct intel_vgpu *vgpu, write_vreg(vgpu, offset, p_data, bytes); data = vgpu_vreg(vgpu, offset); - if (data & _MASKED_BIT_ENABLE(RESET_CTL_REQUEST_RESET)) + if (IS_MASKED_BITS_ENABLED(data, RESET_CTL_REQUEST_RESET)) data |= RESET_CTL_READY_TO_RESET; else if (data & _MASKED_BIT_DISABLE(RESET_CTL_REQUEST_RESET)) data &= ~RESET_CTL_READY_TO_RESET; @@ -1827,7 +1827,8 @@ static int csfe_chicken1_mmio_write(struct intel_vgpu *vgpu, (*(u32 *)p_data) &= ~_MASKED_BIT_ENABLE(0x18); write_vreg(vgpu, offset, p_data, bytes); - if (data & _MASKED_BIT_ENABLE(0x10) || data & _MASKED_BIT_ENABLE(0x8)) + if (IS_MASKED_BITS_ENABLED(data, 0x10) || + IS_MASKED_BITS_ENABLED(data, 0x8)) enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST); return 0; @@ -3055,6 +3056,7 @@ static int init_skl_mmio_info(struct intel_gvt *gvt) MMIO_D(_MMIO(0x72380), D_SKL_PLUS); MMIO_D(_MMIO(0x7239c), D_SKL_PLUS); MMIO_D(_MMIO(_PLANE_SURF_3_A), D_SKL_PLUS); + MMIO_D(_MMIO(_PLANE_SURF_3_B), D_SKL_PLUS); MMIO_D(CSR_SSP_BASE, D_SKL_PLUS); MMIO_D(CSR_HTP_SKL, D_SKL_PLUS); @@ -3131,8 +3133,8 @@ static int init_skl_mmio_info(struct intel_gvt *gvt) MMIO_DFH(GEN9_WM_CHICKEN3, D_SKL_PLUS, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL); - MMIO_D(GAMT_CHKN_BIT_REG, D_KBL); - MMIO_D(GEN9_CTX_PREEMPT_REG, D_KBL | D_SKL); + MMIO_D(GAMT_CHKN_BIT_REG, D_KBL | D_CFL); + MMIO_D(GEN9_CTX_PREEMPT_REG, D_SKL_PLUS); return 0; } diff --git a/drivers/gpu/drm/i915/gvt/mmio_context.h b/drivers/gpu/drm/i915/gvt/mmio_context.h index 970704b18f23..3b25e7fe32f6 100644 --- a/drivers/gpu/drm/i915/gvt/mmio_context.h +++ b/drivers/gpu/drm/i915/gvt/mmio_context.h @@ -54,8 +54,8 @@ bool is_inhibit_context(struct intel_context *ce); int intel_vgpu_restore_inhibit_context(struct intel_vgpu *vgpu, struct i915_request *req); -#define IS_RESTORE_INHIBIT(a) \ - (_MASKED_BIT_ENABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT) == \ - ((a) & _MASKED_BIT_ENABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT))) + +#define IS_RESTORE_INHIBIT(a) \ + IS_MASKED_BITS_ENABLED(a, CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT) #endif diff --git a/drivers/gpu/drm/i915/gvt/reg.h b/drivers/gpu/drm/i915/gvt/reg.h index 5b66e14c5b7b..b88e033cbed4 100644 --- a/drivers/gpu/drm/i915/gvt/reg.h +++ b/drivers/gpu/drm/i915/gvt/reg.h @@ -94,6 +94,11 @@ #define GFX_MODE_BIT_SET_IN_MASK(val, bit) \ ((((bit) & 0xffff0000) == 0) && !!((val) & (((bit) << 16)))) +#define IS_MASKED_BITS_ENABLED(_val, _b) \ + (((_val) & _MASKED_BIT_ENABLE(_b)) == _MASKED_BIT_ENABLE(_b)) +#define IS_MASKED_BITS_DISABLED(_val, _b) \ + ((_val) & _MASKED_BIT_DISABLE(_b)) + #define FORCEWAKE_RENDER_GEN9_REG 0xa278 #define FORCEWAKE_ACK_RENDER_GEN9_REG 0x0D84 #define FORCEWAKE_BLITTER_GEN9_REG 0xa188 diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index bca036ac6621..e7532e7d74e9 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -230,7 +230,7 @@ static int per_file_stats(int id, void *ptr, void *data) struct file_stats *stats = data; struct i915_vma *vma; - if (!kref_get_unless_zero(&obj->base.refcount)) + if (IS_ERR_OR_NULL(obj) || !kref_get_unless_zero(&obj->base.refcount)) return 0; stats->count++; diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index adb9bf34cf97..ae99a9190200 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -410,8 +410,6 @@ struct intel_fbc { int adjusted_x; int adjusted_y; - int y; - u16 pixel_blend_mode; } plane; @@ -420,6 +418,8 @@ struct intel_fbc { unsigned int stride; u64 modifier; } fb; + + unsigned int fence_y_offset; u16 gen9_wa_cfb_stride; s8 fence_id; } state_cache; @@ -435,15 +435,16 @@ struct intel_fbc { struct { enum pipe pipe; enum i9xx_plane_id i9xx_plane; - unsigned int fence_y_offset; } crtc; struct { const struct drm_format_info *format; unsigned int stride; + u64 modifier; } fb; int cfb_size; + unsigned int fence_y_offset; u16 gen9_wa_cfb_stride; s8 fence_id; bool plane_visible; diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 25329b7600c9..014f34c047d5 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -1592,6 +1592,7 @@ static u32 *save_restore_register(struct i915_perf_stream *stream, u32 *cs, u32 d; cmd = save ? MI_STORE_REGISTER_MEM : MI_LOAD_REGISTER_MEM; + cmd |= MI_SRM_LRM_GLOBAL_GTT; if (INTEL_GEN(stream->perf->i915) >= 8) cmd++; diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index fc14ebf9a0b7..1f9cd33b35cb 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -104,6 +104,7 @@ vma_create(struct drm_i915_gem_object *obj, struct i915_address_space *vm, const struct i915_ggtt_view *view) { + struct i915_vma *pos = ERR_PTR(-E2BIG); struct i915_vma *vma; struct rb_node *rb, **p; @@ -184,7 +185,6 @@ vma_create(struct drm_i915_gem_object *obj, rb = NULL; p = &obj->vma.tree.rb_node; while (*p) { - struct i915_vma *pos; long cmp; rb = *p; @@ -196,16 +196,12 @@ vma_create(struct drm_i915_gem_object *obj, * and dispose of ours. */ cmp = i915_vma_compare(pos, vm, view); - if (cmp == 0) { - spin_unlock(&obj->vma.lock); - i915_vma_free(vma); - return pos; - } - if (cmp < 0) p = &rb->rb_right; - else + else if (cmp > 0) p = &rb->rb_left; + else + goto err_unlock; } rb_link_node(&vma->obj_node, rb, p); rb_insert_color(&vma->obj_node, &obj->vma.tree); @@ -228,8 +224,9 @@ vma_create(struct drm_i915_gem_object *obj, err_unlock: spin_unlock(&obj->vma.lock); err_vma: + i915_vm_put(vm); i915_vma_free(vma); - return ERR_PTR(-E2BIG); + return pos; } static struct i915_vma * |