drm/xe/guc_ads: allocate UM queues in a separate UC BO
Subsystems (3)
v3 cover drm/xe/guc_ads: allocate UM queues in a separate UC BO
== Problem ==
On platforms with Unified Shared Memory (USM), the xe driver maintains three hardware page-fault queues (PAGE_FAULT, PAGE_FAULT_RESPONSE, ACCESS_COUNTER) inside the GuC Additional Data Structures (ADS) blob. The ADS blob is mapped WB on the CPU side.
xe_guc_ads_populate() is called on every GT reset to reprogram the ADS blob; it begins with a full memset() of the blob to zero. The zeroed cache lines for the UM queue region reside in the CPU cache (LLC, WB) but are not immediately written back to DRAM.
GAM hardware writes fault descriptors to the queue via DPA - a non-coherent path that bypasses the CPU cache and writes directly to memory. GuC reads the queue via GGTT (WB), which hits the CPU cache. As a result GuC sees the stale zeroed data from the CPU cache instead of the descriptor written to DRAM by GAM. The driver then receives an all-zero fault descriptor and returns -EINVAL, causing an unnecessary engine reset.
This is documented in HSD as:
"When Fault queue is updated by the SW (Buffer zeroed), it can reside
in the CPU cache (L4:WB). A non coherent write from GAM HW would not consult the CPU cache and it will eventually get written into the memory (depends upon eviction/flushes) and CPU cache has stale data. When GUC HW reads the fault queue it gets it from the CPU cache, which is stale data."
The window is narrow but reliably reproducible: run xe_exec_reset --run-subtest gt-reset followed immediately by xe_exec_system_allocator --run-subtest threads-many-execqueues-mmap-new-race.
== Reproduction ==
# Requires a platform with has_usm (e.g. Xe3 iGFX) sudo xe_exec_reset --run-subtest gt-reset sudo xe_exec_system_allocator --run-subtest threads-many-execqueues-mmap-new-race
Symptom in dmesg:
[xe] ASID: 0, Faulted Address: 0x0000000000000000, FaultType: 0
The all-zero fault descriptor is the tell-tale sign of the stale cache line race.
== Fix ==
Allocate the UM queues in a dedicated UC BO (ads->um_queue_bo). UC mapping on the CPU side ensures the memset bypasses the cache and the zeroed data lands in DRAM immediately. GuC's non-coherent read then goes to DRAM and sees the correct descriptor written by GAM.
A separate BO also eliminates the secondary race where a descriptor written by the GPU between GDRST and GuC restart could be zeroed by the next call to xe_guc_ads_populate().
On dGFX, VRAM placement is used so the DPA is derived from vram_region_gpu_offset, avoiding the IOVA aliasing issue seen on platforms with limited DRAM (e.g. BMG 4 GB) where a SYSTEM BO's DMA address can alias a PCI MMIO reserved window.
Why not just flush the cache after memset?
An earlier approach called drm_clflush_virt_range() on the UM queue region immediately after the memset to evict the stale zero lines from LLC. This reduces the failure rate significantly but does not fully eliminate the race for two reasons.
First, the Linux kernel maintains a WB direct mapping (linear map) that covers all physical memory. Even after flushing the vmap alias, the CPU can speculatively prefetch the same physical page through the WB direct-map alias, silently repopulating the cache with zeros. LNL is known to be particularly susceptible to this speculative prefetch behaviour. XE_BO_FLAG_NEEDS_UC calls set_memory_uc() which updates both the vmap PTE and the direct-map PTE to UC, closing this aliasing window entirely.
Second, the flush approach leaves the memory mapped WB on the CPU side permanently. Any future SW access - a debug read, a tracing hook, an inadvertent touch - can re-dirty the cache line and reintroduce the race. UC mapping makes the correct behaviour unconditional and self-documenting.
Jia Yao (1):
drm/xe/guc_ads: allocate UM queues in a separate UC BO
drivers/gpu/drm/xe/xe_guc_ads.c | 64 ++++++++++++++++----------- drivers/gpu/drm/xe/xe_guc_ads_types.h | 8 +++- 2 files changed, 46 insertions(+), 26 deletions(-)
Signature · 1 line
2.43.0
PATCH v3 1/1 drm/xe/guc_ads: allocate UM queues in a separate UC BO
The HW unit that writes to the UM fault queue is not always coherent with the CPU cache. Ensure the UM queue memory is uncached on the CPU side so that GuC always reads descriptors from memory rather than a stale CPU cache line.
Allocate the UM queues in a dedicated UC BO (ads->um_queue_bo) instead of the main ADS blob. This also prevents xe_guc_ads_populate()'s memset from zeroing any fault descriptor written between GDRST and GuC restart. On dGFX, VRAM placement gives a well-defined DPA.
v3: simplify commit message and code comments per reviewer feedback;
add Fixes tag and stable Cc
v2: also fix dGFX (VRAM placement + UC flag); remove now-dead
guc_ads_um_queues_size/offset helpers
--- drivers/gpu/drm/xe/xe_guc_ads.c | 64 ++++++++++++++++----------- drivers/gpu/drm/xe/xe_guc_ads_types.h | 8 +++- 2 files changed, 46 insertions(+), 26 deletions(-)
drivers/gpu/drm/xe/xe_guc_ads.c +39 -25
@@ -155,16 +155,6 @@ static size_t guc_ads_capture_size(struct xe_guc_ads *ads)return PAGE_ALIGN (ads->capture_size); } -static size_t guc_ads_um_queues_size (struct xe_guc_ads *ads) -{ -struct xe_device *xe =ads_to_xe (ads); - -if (!xe->info.has_usm) -return 0 ; - -return GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX ; -} -static size_t guc_ads_private_data_size (struct xe_guc_ads *ads) {return PAGE_ALIGN (ads_to_guc (ads)->fw.private_data_size);
@@ -205,22 +195,12 @@ static size_t guc_ads_capture_offset(struct xe_guc_ads *ads)return PAGE_ALIGN (offset); } -static size_t guc_ads_um_queues_offset (struct xe_guc_ads *ads) -{ -u32 offset; - - offset =guc_ads_capture_offset (ads) + -guc_ads_capture_size (ads); - -return PAGE_ALIGN (offset); -} -static size_t guc_ads_private_data_offset (struct xe_guc_ads *ads) {size_t offset; - offset =guc_ads_um_queues_offset (ads) + -guc_ads_um_queues_size (ads); + offset =guc_ads_capture_offset (ads) + +guc_ads_capture_size (ads);return PAGE_ALIGN (offset); }
@@ -409,6 +389,30 @@ int xe_guc_ads_init(struct xe_guc_ads *ads) ads->bo = bo; +/* +* The HW unit that writes to the UM fault queue is not coherent with +* the CPU cache. Use a dedicated UC BO to ensure GuC always reads +* descriptors from memory rather than a stale CPU cache line. A +* separate BO also prevents xe_guc_ads_populate()'s memset from +* zeroing descriptors written between GDRST and GuC restart. On +* dGFX, VRAM gives a well-defined DPA. +*/ +if (xe->info.has_usm) { +u32 um_flags =XE_BO_FLAG_GGTT | +XE_BO_FLAG_GGTT_INVALIDATE | +XE_BO_FLAG_PINNED_NORESTORE | +XE_BO_FLAG_NEEDS_UC | +XE_BO_FLAG_VRAM_IF_DGFX (tile); + + bo =xe_managed_bo_create_pin_map (xe, tile, +GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX , + um_flags); +if (IS_ERR (bo)) +return PTR_ERR (bo); + + ads->um_queue_bo = bo; + } +return 0 ; }ALLOW_ERROR_INJECTION (xe_guc_ads_init,ERRNO );/* See xe_pci_probe() */
@@ -820,9 +824,9 @@ static void guc_mmio_reg_state_init(struct xe_guc_ads *ads)static void guc_um_init_params (struct xe_guc_ads *ads) { -u32 um_queue_offset =guc_ads_um_queues_offset (ads);struct xe_guc *guc =ads_to_guc (ads);struct xe_device *xe =ads_to_xe (ads); +struct xe_bo *um_bo = ads->um_queue_bo;u64 base_dpa;u32 base_ggtt;bool with_dpa;
@@ -830,8 +834,14 @@ static void guc_um_init_params(struct xe_guc_ads *ads) with_dpa = !xe_guc_using_main_gamctrl_queues (guc); - base_ggtt =xe_bo_ggtt_addr (ads->bo) + um_queue_offset; - base_dpa =xe_bo_main_addr (ads->bo,PAGE_SIZE ) + um_queue_offset; +if (um_bo) { +/* All USM platforms: UM queues in dedicated um_queue_bo */ + base_ggtt =xe_bo_ggtt_addr (um_bo); + base_dpa =xe_bo_main_addr (um_bo,PAGE_SIZE ); + }else { +/* Platform does not support USM: no UM queues, nothing to do */ +return ; + }for (i =0 ; i <GUC_UM_HW_QUEUE_MAX ; ++i) {/*
@@ -912,6 +922,10 @@ void xe_guc_ads_populate(struct xe_guc_ads *ads)xe_gt_assert (gt, ads->bo);xe_map_memset (ads_to_xe (ads),ads_to_map (ads),0 ,0 ,xe_bo_size (ads->bo)); +/* +* um_queue_bo is not zeroed here. GuC resets head/tail via MMIO on +* every restart so stale ring contents are never consumed. +*/ guc_policies_init (ads);fill_engine_enable_masks (gt, &info_map);guc_mmio_reg_state_init (ads);
drivers/gpu/drm/xe/xe_guc_ads_types.h +7 -1
@@ -14,8 +14,14 @@ struct xe_bo;* struct xe_guc_ads - GuC additional data structures (ADS) */ struct xe_guc_ads { -/** @bo: Xe BO for GuC ads blob */ +/** @bo: Xe BO for GuC ads blob (WB cached) */ struct xe_bo *bo; +/** +* @um_queue_bo: Dedicated UC BO for the HW fault ring (UM queues). +* The HW unit writing to this ring is not coherent with the CPU cache; +* UC ensures GuC reads descriptors from memory. NULL if no USM support. +*/ +struct xe_bo *um_queue_bo;/** @golden_lrc_size: golden LRC size */ size_t golden_lrc_size;/** @regset_size: size of register set passed to GuC for save/restore */
Signature · 1 line
2.43.0
Quoted 80 lines · Jia Yao
The HW unit that writes to the UM fault queue is not always coherent with the CPU cache. Ensure the UM queue memory is uncached on the CPU side so that GuC always reads descriptors from memory rather than a stale CPU cache line.
Allocate the UM queues in a dedicated UC BO (
ads->um_queue_bo) instead of the main ADS blob. This also prevents xe_guc_ads_populate()'s memset from zeroing any fault descriptor written between GDRST and GuC restart. On dGFX, VRAM placement gives a well-defined DPA.v3: simplify commit message and code comments per reviewer feedback;
add Fixes tag and stable Ccv2: also fix dGFX (VRAM placement + UC flag); remove now-dead
guc_ads_um_queues_size/offset helpersFixes dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")Cc Gwan-gyeong Mun <[email protected]>Cc Matthew Auld <[email protected]>Signed-off-by Jia Yao <[email protected]>--- drivers/gpu/drm/xe/xe_guc_ads.c | 64 ++++++++++++++++----------- drivers/gpu/drm/xe/xe_guc_ads_types.h | 8 +++- 2 files changed, 46 insertions(+), 26 deletions(-)drivers/gpu/drm/xe/xe_guc_ads.c +6
-22@@ -155,16 +155,6 @@ static size_t guc_ads_capture_size(struct xe_guc_ads *ads)return PAGE_ALIGN (ads->capture_size); } -static size_t guc_ads_um_queues_size (struct xe_guc_ads *ads) -{ -struct xe_device *xe =ads_to_xe (ads); - -if (!xe->info.has_usm) -return 0 ; - -return GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX ; -} -static size_t guc_ads_private_data_size (struct xe_guc_ads *ads) {return PAGE_ALIGN (ads_to_guc (ads)->fw.private_data_size);@@ -205,22 +195,12 @@ static size_t guc_ads_capture_offset(struct xe_guc_ads *ads)return PAGE_ALIGN (offset); } -static size_t guc_ads_um_queues_offset (struct xe_guc_ads *ads) -{ -u32 offset; - - offset =guc_ads_capture_offset (ads) + -guc_ads_capture_size (ads); - -return PAGE_ALIGN (offset); -} -static size_t guc_ads_private_data_offset (struct xe_guc_ads *ads) {size_t offset; - offset =guc_ads_um_queues_offset (ads) + -guc_ads_um_queues_size (ads); + offset =guc_ads_capture_offset (ads) + +guc_ads_capture_size (ads);return PAGE_ALIGN (offset); }@@ -409,6 +389,30 @@ int xe_guc_ads_init(struct xe_guc_ads *ads) ads->bo = bo; +/* +* The HW unit that writes to the UM fault queue is not coherent with +* the CPU cache. Use a dedicated UC BO to ensure GuC always reads +* descriptors from memory rather than a stale CPU cache line. A
drivers/gpu/drm/xe/xe_guc_ads.c int xe_guc_ads_init(struct xe_guc_ads *ads) · line 405
Quoted 10 lines
drivers/gpu/drm/xe/xe_guc_ads.c +10
-0+* separate BO also prevents xe_guc_ads_populate()'s memset from +* zeroing descriptors written between GDRST and GuC restart. On +* dGFX, VRAM gives a well-defined DPA. +*/ +if (xe->info.has_usm) { +u32 um_flags =XE_BO_FLAG_GGTT | +XE_BO_FLAG_GGTT_INVALIDATE | +XE_BO_FLAG_PINNED_NORESTORE | +XE_BO_FLAG_NEEDS_UC | +XE_BO_FLAG_VRAM_IF_DGFX (tile);
Only slight doubt is the dgpu change here. I don't think there can be a coherency issue there, since memory access to system memory should just be like any other PCI transaction, and AFAIK coherency can't be turned off on current dgpu hw. Just wondering if moving this to VRAM is really needed for the fix, or whether there are any perf considerations? Not sure if there is some guidance on where is best to allocate this on dgpu? Also, perhaps this should be split out into a separate commit, since change seems orthogonal to the coherency issue?
Quoted 72 lines
drivers/gpu/drm/xe/xe_guc_ads.c +23
-3+ + bo =xe_managed_bo_create_pin_map (xe, tile, +GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX , + um_flags); +if (IS_ERR (bo)) +return PTR_ERR (bo); + + ads->um_queue_bo = bo; + } +return 0 ; }ALLOW_ERROR_INJECTION (xe_guc_ads_init,ERRNO );/* See xe_pci_probe() */ @@ -820,9 +824,9 @@ static void guc_mmio_reg_state_init(struct xe_guc_ads *ads)static void guc_um_init_params (struct xe_guc_ads *ads) { -u32 um_queue_offset =guc_ads_um_queues_offset (ads);struct xe_guc *guc =ads_to_guc (ads);struct xe_device *xe =ads_to_xe (ads); +struct xe_bo *um_bo = ads->um_queue_bo;u64 base_dpa;u32 base_ggtt;bool with_dpa;@@ -830,8 +834,14 @@ static void guc_um_init_params(struct xe_guc_ads *ads) with_dpa = !xe_guc_using_main_gamctrl_queues (guc); - base_ggtt =xe_bo_ggtt_addr (ads->bo) + um_queue_offset; - base_dpa =xe_bo_main_addr (ads->bo,PAGE_SIZE ) + um_queue_offset; +if (um_bo) { +/* All USM platforms: UM queues in dedicated um_queue_bo */ + base_ggtt =xe_bo_ggtt_addr (um_bo); + base_dpa =xe_bo_main_addr (um_bo,PAGE_SIZE ); + }else { +/* Platform does not support USM: no UM queues, nothing to do */ +return ; + }for (i =0 ; i <GUC_UM_HW_QUEUE_MAX ; ++i) {/* @@ -912,6 +922,10 @@ void xe_guc_ads_populate(struct xe_guc_ads *ads)xe_gt_assert (gt, ads->bo);xe_map_memset (ads_to_xe (ads),ads_to_map (ads),0 ,0 ,xe_bo_size (ads->bo)); +/* +* um_queue_bo is not zeroed here. GuC resets head/tail via MMIO on +* every restart so stale ring contents are never consumed. +*/ guc_policies_init (ads);fill_engine_enable_masks (gt, &info_map);guc_mmio_reg_state_init (ads);drivers/gpu/drm/xe/xe_guc_ads_types.h +7
-1@@ -14,8 +14,14 @@ struct xe_bo;* struct xe_guc_ads - GuC additional data structures (ADS) */ struct xe_guc_ads { -/** @bo: Xe BO for GuC ads blob */ +/** @bo: Xe BO for GuC ads blob (WB cached) */ struct xe_bo *bo; +/** +* @um_queue_bo: Dedicated UC BO for the HW fault ring (UM queues). +* The HW unit writing to this ring is not coherent with the CPU cache; +* UC ensures GuC reads descriptors from memory. NULL if no USM support. +*/ +struct xe_bo *um_queue_bo;/** @golden_lrc_size: golden LRC size */ size_t golden_lrc_size;/** @regset_size: size of register set passed to GuC for save/restore */
Quoted 152 lines · Jia Yao
The HW unit that writes to the UM fault queue is not always coherent with the CPU cache. Ensure the UM queue memory is uncached on the CPU side so that GuC always reads descriptors from memory rather than a stale CPU cache line.
Allocate the UM queues in a dedicated UC BO (
ads->um_queue_bo) instead of the main ADS blob. This also prevents xe_guc_ads_populate()'s memset from zeroing any fault descriptor written between GDRST and GuC restart. On dGFX, VRAM placement gives a well-defined DPA.v3: simplify commit message and code comments per reviewer feedback;
add Fixes tag and stable Ccv2: also fix dGFX (VRAM placement + UC flag); remove now-dead
guc_ads_um_queues_size/offset helpersFixes dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")Cc Gwan-gyeong Mun <[email protected]>Cc Matthew Auld <[email protected]>Signed-off-by Jia Yao <[email protected]>--- drivers/gpu/drm/xe/xe_guc_ads.c | 64 ++++++++++++++++----------- drivers/gpu/drm/xe/xe_guc_ads_types.h | 8 +++- 2 files changed, 46 insertions(+), 26 deletions(-)drivers/gpu/drm/xe/xe_guc_ads.c +39
-25@@ -155,16 +155,6 @@ static size_t guc_ads_capture_size(struct xe_guc_ads *ads)return PAGE_ALIGN (ads->capture_size); } -static size_t guc_ads_um_queues_size (struct xe_guc_ads *ads) -{ -struct xe_device *xe =ads_to_xe (ads); - -if (!xe->info.has_usm) -return 0 ; - -return GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX ; -} -static size_t guc_ads_private_data_size (struct xe_guc_ads *ads) {return PAGE_ALIGN (ads_to_guc (ads)->fw.private_data_size);@@ -205,22 +195,12 @@ static size_t guc_ads_capture_offset(struct xe_guc_ads *ads)return PAGE_ALIGN (offset); } -static size_t guc_ads_um_queues_offset (struct xe_guc_ads *ads) -{ -u32 offset; - - offset =guc_ads_capture_offset (ads) + -guc_ads_capture_size (ads); - -return PAGE_ALIGN (offset); -} -static size_t guc_ads_private_data_offset (struct xe_guc_ads *ads) {size_t offset; - offset =guc_ads_um_queues_offset (ads) + -guc_ads_um_queues_size (ads); + offset =guc_ads_capture_offset (ads) + +guc_ads_capture_size (ads);return PAGE_ALIGN (offset); }@@ -409,6 +389,30 @@ int xe_guc_ads_init(struct xe_guc_ads *ads) ads->bo = bo; +/* +* The HW unit that writes to the UM fault queue is not coherent with +* the CPU cache. Use a dedicated UC BO to ensure GuC always reads +* descriptors from memory rather than a stale CPU cache line. A +* separate BO also prevents xe_guc_ads_populate()'s memset from +* zeroing descriptors written between GDRST and GuC restart. On +* dGFX, VRAM gives a well-defined DPA. +*/ +if (xe->info.has_usm) { +u32 um_flags =XE_BO_FLAG_GGTT | +XE_BO_FLAG_GGTT_INVALIDATE | +XE_BO_FLAG_PINNED_NORESTORE | +XE_BO_FLAG_NEEDS_UC | +XE_BO_FLAG_VRAM_IF_DGFX (tile); + + bo =xe_managed_bo_create_pin_map (xe, tile, +GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX , + um_flags); +if (IS_ERR (bo)) +return PTR_ERR (bo); + + ads->um_queue_bo = bo; + } +return 0 ; }ALLOW_ERROR_INJECTION (xe_guc_ads_init,ERRNO );/* See xe_pci_probe() */ @@ -820,9 +824,9 @@ static void guc_mmio_reg_state_init(struct xe_guc_ads *ads)static void guc_um_init_params (struct xe_guc_ads *ads) { -u32 um_queue_offset =guc_ads_um_queues_offset (ads);struct xe_guc *guc =ads_to_guc (ads);struct xe_device *xe =ads_to_xe (ads); +struct xe_bo *um_bo = ads->um_queue_bo;u64 base_dpa;u32 base_ggtt;bool with_dpa;@@ -830,8 +834,14 @@ static void guc_um_init_params(struct xe_guc_ads *ads) with_dpa = !xe_guc_using_main_gamctrl_queues (guc); - base_ggtt =xe_bo_ggtt_addr (ads->bo) + um_queue_offset; - base_dpa =xe_bo_main_addr (ads->bo,PAGE_SIZE ) + um_queue_offset; +if (um_bo) { +/* All USM platforms: UM queues in dedicated um_queue_bo */ + base_ggtt =xe_bo_ggtt_addr (um_bo); + base_dpa =xe_bo_main_addr (um_bo,PAGE_SIZE ); + }else { +/* Platform does not support USM: no UM queues, nothing to do */ +return ; + }for (i =0 ; i <GUC_UM_HW_QUEUE_MAX ; ++i) {/* @@ -912,6 +922,10 @@ void xe_guc_ads_populate(struct xe_guc_ads *ads)xe_gt_assert (gt, ads->bo);xe_map_memset (ads_to_xe (ads),ads_to_map (ads),0 ,0 ,xe_bo_size (ads->bo)); +/* +* um_queue_bo is not zeroed here. GuC resets head/tail via MMIO on +* every restart so stale ring contents are never consumed. +*/ guc_policies_init (ads);fill_engine_enable_masks (gt, &info_map);guc_mmio_reg_state_init (ads);drivers/gpu/drm/xe/xe_guc_ads_types.h +1
-1@@ -14,8 +14,14 @@ struct xe_bo;* struct xe_guc_ads - GuC additional data structures (ADS) */ struct xe_guc_ads { -/** @bo: Xe BO for GuC ads blob */ +/** @bo: Xe BO for GuC ads blob (WB cached) */
drivers/gpu/drm/xe/xe_guc_ads_types.h struct xe_bo; · line 27
Quoted 10 lines
drivers/gpu/drm/xe/xe_guc_ads_types.h +6
-0struct xe_bo *bo; +/** +* @um_queue_bo: Dedicated UC BO for the HW fault ring (UM queues). +* The HW unit writing to this ring is not coherent with the CPU cache; +* UC ensures GuC reads descriptors from memory. NULL if no USM support. +*/ +struct xe_bo *um_queue_bo;/** @golden_lrc_size: golden LRC size */ size_t golden_lrc_size;/** @regset_size: size of register set passed to GuC for save/restore */
Hi Jia,
Regarding the issue mentioned in the cover letter that this patch aims to fix, wouldn't it be resolved by creating a separate um_queue_bo by adding XE_BO_FLAG_NEEDS_UC to the (XE_BO_FLAG_SYSTEM | XE_BO_FLAG_GGTT | XE_BO_FLAG_GGTT_INVALIDATE | XE_BO_FLAG_PINNED_NORESTORE) flags used in the existing code when generating ads->bo? If so, as Matthew Auld mentioned, wouldn't it be better to apply this patch as one that creates an Uncached buffer for um_queue_bo, and then split it into a separate patch that uses XE_BO_FLAG_VRAM_IF_DGFX(tile) instead of XE_BO_FLAG_SYSTEM if there is a performance improvement on the dGPU?
And please also update the description section titled "Layout of the ADS blob allocated for the GuC" in the [drivers/gpu/drm/xe/xe_guc_ads.c] file.
G.G.
Patchwork
CI✗ CI.checkpatch: warning for drm/xe/guc_ads: allocate UM queues in a separate UC BO
In reply to Jia Yao== Series Details ==
== Series Details ==
Series: drm/xe/guc_ads: allocate UM queues in a separate UC BO
URL : https://patchwork.freedesktop.org/series/170879/
State : warning
== Summary ==
+ KERNEL=/kernel
+ git clone https://gitlab.freedesktop.org/drm/maintainer-tools mt Cloning into 'mt'...
warning: redirecting to https://gitlab.freedesktop.org/drm/maintainer-tools.git/
+ git -C mt rev-list -n1 origin/master 061140b9bc586ae7f40abc1249c97e1cc72d1b9d
+ cd /kernel
+ git config --global --add safe.directory /kernel
+ git log -n1
commit b19a063d78acc72bfcc507071441e931bb38bdfb
Author: Jia Yao <[email protected]> Date: Wed Jul 22 05:34:10 2026 +0000
drm/xe/guc_ads: allocate UM queues in a separate UC BO
The HW unit that writes to the UM fault queue is not always coherent with the CPU cache. Ensure the UM queue memory is uncached on the CPU side so that GuC always reads descriptors from memory rather than a stale CPU cache line.
Allocate the UM queues in a dedicated UC BO (ads->um_queue_bo) instead of the main ADS blob. This also prevents xe_guc_ads_populate()'s memset from zeroing any fault descriptor written between GDRST and GuC restart. On dGFX, VRAM placement gives a well-defined DPA.
v3: simplify commit message and code comments per reviewer feedback;
add Fixes tag and stable Cc
v2: also fix dGFX (VRAM placement + UC flag); remove now-dead
guc_ads_um_queues_size/offset helpers
+ /mt/dim checkpatch 5726f9c4bde0290e5e60e636b9632bb03a57498c drm-intel b19a063d78ac drm/xe/guc_ads: allocate UM queues in a separate UC BO -:94: CHECK:PARENTHESIS_ALIGNMENT: Alignment should match open parenthesis #94: FILE: drivers/gpu/drm/xe/xe_guc_ads.c:460:
+ bo = xe_managed_bo_create_pin_map(xe, tile,
+ GUC_UM_QUEUE_SIZE * GUC_UM_HW_QUEUE_MAX,
total: 0 errors, 0 warnings, 1 checks, 121 lines checked
Patchwork
CI✓ CI.KUnit: success for drm/xe/guc_ads: allocate UM queues in a separate UC BO
In reply to Jia Yao== Series Details ==
== Series Details ==
Series: drm/xe/guc_ads: allocate UM queues in a separate UC BO
URL : https://patchwork.freedesktop.org/series/170879/
State : success
== Summary ==
+ trap cleanup EXIT
+ /kernel/tools/testing/kunit/kunit.py run --kunitconfig /kernel/drivers/gpu/drm/xe/.kunitconfig [05:41:03] Configuring KUnit Kernel ...
Generating .config ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig [05:41:07] Building KUnit Kernel ... Populating config with: $ make ARCH=um O=.kunit olddefconfig Building with: $ make all compile_commands.json scripts_gdb ARCH=um O=.kunit --jobs=48 [05:41:39] Starting KUnit Kernel (1/1)... [05:41:39] ============================================================ Running tests with: $ .kunit/linux kunit.enable=1 mem=1G console=tty kunit_shutdown=halt [05:41:39] ================== guc_buf (11 subtests) =================== [05:41:39] [PASSED] test_smallest [05:41:39] [PASSED] test_largest [05:41:39] [PASSED] test_granular [05:41:39] [PASSED] test_unique [05:41:39] [PASSED] test_overlap [05:41:39] [PASSED] test_reusable [05:41:39] [PASSED] test_too_big [05:41:39] [PASSED] test_flush [05:41:39] [PASSED] test_lookup [05:41:39] [PASSED] test_data [05:41:39] [PASSED] test_class [05:41:39] ===================== [PASSED] guc_buf ===================== [05:41:39] =================== guc_dbm (7 subtests) =================== [05:41:39] [PASSED] test_empty [05:41:39] [PASSED] test_default [05:41:39] ======================== test_size ======================== [05:41:39] [PASSED] 4 [05:41:39] [PASSED] 8 [05:41:39] [PASSED] 32 [05:41:39] [PASSED] 256 [05:41:39] ==================== [PASSED] test_size ==================== [05:41:39] ======================= test_reuse ======================== [05:41:39] [PASSED] 4 [05:41:39] [PASSED] 8 [05:41:39] [PASSED] 32 [05:41:39] [PASSED] 256 [05:41:39] =================== [PASSED] test_reuse ==================== [05:41:39] =================== test_range_overlap ==================== [05:41:39] [PASSED] 4 [05:41:39] [PASSED] 8 [05:41:39] [PASSED] 32 [05:41:39] [PASSED] 256 [05:41:39] =============== [PASSED] test_range_overlap ================ [05:41:39] =================== test_range_compact ==================== [05:41:39] [PASSED] 4 [05:41:39] [PASSED] 8 [05:41:39] [PASSED] 32 [05:41:39] [PASSED] 256 [05:41:39] =============== [PASSED] test_range_compact ================ [05:41:39] ==================== test_range_spare ===================== [05:41:39] [PASSED] 4 [05:41:39] [PASSED] 8 [05:41:39] [PASSED] 32 [05:41:39] [PASSED] 256 [05:41:39] ================ [PASSED] test_range_spare ================= [05:41:39] ===================== [PASSED] guc_dbm ===================== [05:41:39] =================== guc_idm (6 subtests) =================== [05:41:39] [PASSED] bad_init [05:41:39] [PASSED] no_init [05:41:39] [PASSED] init_fini [05:41:39] [PASSED] check_used [05:41:39] [PASSED] check_quota [05:41:39] [PASSED] check_all [05:41:39] ===================== [PASSED] guc_idm ===================== [05:41:39] =============== guc_klv_helpers (9 subtests) =============== [05:41:39] [PASSED] test_count [05:41:39] [PASSED] test_encode_u32 [05:41:39] [PASSED] test_encode_u64 [05:41:39] [PASSED] test_encode_string [05:41:39] [PASSED] test_encode_object_raw [05:41:39] [PASSED] test_encode_object_klv [05:41:39] [PASSED] test_encode_object_nested [05:41:39] [PASSED] test_encode_object_basic [05:41:39] [PASSED] test_print [05:41:39] ================= [PASSED] guc_klv_helpers ================= [05:41:39] ================== no_relay (3 subtests) =================== [05:41:39] [PASSED] xe_drops_guc2pf_if_not_ready [05:41:39] [PASSED] xe_drops_guc2vf_if_not_ready [05:41:39] [PASSED] xe_rejects_send_if_not_ready [05:41:39] ==================== [PASSED] no_relay ===================== [05:41:39] ================== pf_relay (14 subtests) ================== [05:41:39] [PASSED] pf_rejects_guc2pf_too_short [05:41:39] [PASSED] pf_rejects_guc2pf_too_long [05:41:39] [PASSED] pf_rejects_guc2pf_no_payload [05:41:39] [PASSED] pf_fails_no_payload [05:41:39] [PASSED] pf_fails_bad_origin [05:41:39] [PASSED] pf_fails_bad_type [05:41:39] [PASSED] pf_txn_reports_error [05:41:39] [PASSED] pf_txn_sends_pf2guc [05:41:39] [PASSED] pf_sends_pf2guc [05:41:39] [SKIPPED] pf_loopback_nop (requires CONFIG_DRM_XE_DEBUG_SRIOV) [05:41:39] [SKIPPED] pf_loopback_echo (requires CONFIG_DRM_XE_DEBUG_SRIOV) [05:41:39] [SKIPPED] pf_loopback_fail (requires CONFIG_DRM_XE_DEBUG_SRIOV) [05:41:39] [SKIPPED] pf_loopback_busy (requires CONFIG_DRM_XE_DEBUG_SRIOV) [05:41:39] [SKIPPED] pf_loopback_retry (requires CONFIG_DRM_XE_DEBUG_SRIOV) [05:41:39] ==================== [PASSED] pf_relay ===================== [05:41:39] ================== vf_relay (3 subtests) =================== [05:41:39] [PASSED] vf_rejects_guc2vf_too_short [05:41:39] [PASSED] vf_rejects_guc2vf_too_long [05:41:39] [PASSED] vf_rejects_guc2vf_no_payload [05:41:39] ==================== [PASSED] vf_relay ===================== [05:41:39] ================ pf_gt_config (9 subtests) ================= [05:41:39] [PASSED] fair_contexts_1vf [05:41:39] [PASSED] fair_doorbells_1vf [05:41:39] [PASSED] fair_ggtt_1vf [05:41:39] ====================== fair_vram_1vf ====================== [05:41:39] [PASSED] 3.50 GiB [05:41:39] [PASSED] 11.5 GiB [05:41:39] [PASSED] 15.5 GiB [05:41:39] [PASSED] 31.5 GiB [05:41:39] [PASSED] 63.5 GiB [05:41:39] [PASSED] 1.91 GiB [05:41:39] ================== [PASSED] fair_vram_1vf ================== [05:41:39] ================ fair_vram_1vf_admin_only ================= [05:41:39] [PASSED] 3.50 GiB [05:41:39] [PASSED] 11.5 GiB [05:41:39] [PASSED] 15.5 GiB [05:41:39] [PASSED] 31.5 GiB [05:41:39] [PASSED] 63.5 GiB [05:41:39] [PASSED] 1.91 GiB [05:41:39] ============ [PASSED] fair_vram_1vf_admin_only ============= [05:41:39] ====================== fair_contexts ====================== [05:41:39] [PASSED] 1 VF [05:41:39] [PASSED] 2 VFs [05:41:39] [PASSED] 3 VFs [05:41:39] [PASSED] 4 VFs [05:41:39] [PASSED] 5 VFs [05:41:39] [PASSED] 6 VFs [05:41:39] [PASSED] 7 VFs [05:41:39] [PASSED] 8 VFs [05:41:39] [PASSED] 9 VFs [05:41:39] [PASSED] 10 VFs [05:41:39] [PASSED] 11 VFs [05:41:39] [PASSED] 12 VFs [05:41:39] [PASSED] 13 VFs [05:41:39] [PASSED] 14 VFs [05:41:39] [PASSED] 15 VFs [05:41:39] [PASSED] 16 VFs [05:41:39] [PASSED] 17 VFs [05:41:39] [PASSED] 18 VFs [05:41:39] [PASSED] 19 VFs [05:41:39] [PASSED] 20 VFs [05:41:39] [PASSED] 21 VFs [05:41:39] [PASSED] 22 VFs [05:41:39] [PASSED] 23 VFs [05:41:39] [PASSED] 24 VFs [05:41:39] [PASSED] 25 VFs [05:41:39] [PASSED] 26 VFs [05:41:39] [PASSED] 27 VFs [05:41:39] [PASSED] 28 VFs [05:41:39] [PASSED] 29 VFs [05:41:39] [PASSED] 30 VFs [05:41:39] [PASSED] 31 VFs [05:41:39] [PASSED] 32 VFs [05:41:39] [PASSED] 33 VFs [05:41:39] [PASSED] 34 VFs [05:41:39] [PASSED] 35 VFs [05:41:39] [PASSED] 36 VFs [05:41:39] [PASSED] 37 VFs [05:41:39] [PASSED] 38 VFs [05:41:39] [PASSED] 39 VFs [05:41:39] [PASSED] 40 VFs [05:41:39] [PASSED] 41 VFs [05:41:39] [PASSED] 42 VFs [05:41:39] [PASSED] 43 VFs [05:41:39] [PASSED] 44 VFs [05:41:39] [PASSED] 45 VFs [05:41:39] [PASSED] 46 VFs [05:41:39] [PASSED] 47 VFs [05:41:39] [PASSED] 48 VFs [05:41:39] [PASSED] 49 VFs [05:41:39] [PASSED] 50 VFs [05:41:39] [PASSED] 51 VFs [05:41:39] [PASSED] 52 VFs [05:41:39] [PASSED] 53 VFs [05:41:39] [PASSED] 54 VFs [05:41:39] [PASSED] 55 VFs [05:41:39] [PASSED] 56 VFs [05:41:39] [PASSED] 57 VFs [05:41:39] [PASSED] 58 VFs [05:41:39] [PASSED] 59 VFs [05:41:39] [PASSED] 60 VFs [05:41:39] [PASSED] 61 VFs [05:41:39] [PASSED] 62 VFs [05:41:39] [PASSED] 63 VFs [05:41:39] ================== [PASSED] fair_contexts ================== [05:41:39] ===================== fair_doorbells ====================== [05:41:39] [PASSED] 1 VF [05:41:39] [PASSED] 2 VFs [05:41:39] [PASSED] 3 VFs [05:41:39] [PASSED] 4 VFs [05:41:39] [PASSED] 5 VFs [05:41:39] [PASSED] 6 VFs [05:41:39] [PASSED] 7 VFs [05:41:39] [PASSED] 8 VFs [05:41:39] [PASSED] 9 VFs [05:41:39] [PASSED] 10 VFs [05:41:39] [PASSED] 11 VFs [05:41:39] [PASSED] 12 VFs [05:41:39] [PASSED] 13 VFs [05:41:39] [PASSED] 14 VFs [05:41:39] [PASSED] 15 VFs [05:41:39] [PASSED] 16 VFs [05:41:39] [PASSED] 17 VFs [05:41:39] [PASSED] 18 VFs [05:41:39] [PASSED] 19 VFs [05:41:39] [PASSED] 20 VFs [05:41:39] [PASSED] 21 VFs [05:41:39] [PASSED] 22 VFs [05:41:39] [PASSED] 23 VFs [05:41:39] [PASSED] 24 VFs [05:41:39] [PASSED] 25 VFs [05:41:39] [PASSED] 26 VFs [05:41:39] [PASSED] 27 VFs [05:41:39] [PASSED] 28 VFs [05:41:39] [PASSED] 29 VFs [05:41:39] [PASSED] 30 VFs [05:41:39] [PASSED] 31 VFs [05:41:39] [PASSED] 32 VFs [05:41:39] [PASSED] 33 VFs [05:41:39] [PASSED] 34 VFs [05:41:39] [PASSED] 35 VFs [05:41:39] [PASSED] 36 VFs [05:41:39] [PASSED] 37 VFs [05:41:39] [PASSED] 38 VFs [05:41:39] [PASSED] 39 VFs [05:41:39] [PASSED] 40 VFs [05:41:39] [PASSED] 41 VFs [05:41:39] [PASSED] 42 VFs [05:41:39] [PASSED] 43 VFs [05:41:39] [PASSED] 44 VFs [05:41:39] [PASSED] 45 VFs [05:41:39] [PASSED] 46 VFs [05:41:39] [PASSED] 47 VFs [05:41:39] [PASSED] 48 VFs [05:41:39] [PASSED] 49 VFs [05:41:39] [PASSED] 50 VFs [05:41:39] [PASSED] 51 VFs [05:41:39] [PASSED] 52 VFs [05:41:39] [PASSED] 53 VFs [05:41:39] [PASSED] 54 VFs [05:41:39] [PASSED] 55 VFs [05:41:39] [PASSED] 56 VFs [05:41:39] [PASSED] 57 VFs [05:41:39] [PASSED] 58 VFs [05:41:39] [PASSED] 59 VFs [05:41:39] [PASSED] 60 VFs [05:41:39] [PASSED] 61 VFs [05:41:39] [PASSED] 62 VFs [05:41:39] [PASSED] 63 VFs [05:41:39] ================= [PASSED] fair_doorbells ================== [05:41:39] ======================== fair_ggtt ======================== [05:41:39] [PASSED] 1 VF [05:41:39] [PASSED] 2 VFs [05:41:39] [PASSED] 3 VFs [05:41:39] [PASSED] 4 VFs [05:41:39] [PASSED] 5 VFs [05:41:39] [PASSED] 6 VFs [05:41:39] [PASSED] 7 VFs [05:41:39] [PASSED] 8 VFs [05:41:39] [PASSED] 9 VFs [05:41:39] [PASSED] 10 VFs [05:41:39] [PASSED] 11 VFs [05:41:39] [PASSED] 12 VFs [05:41:39] [PASSED] 13 VFs [05:41:39] [PASSED] 14 VFs [05:41:39] [PASSED] 15 VFs [05:41:39] [PASSED] 16 VFs [05:41:39] [PASSED] 17 VFs [05:41:39] [PASSED] 18 VFs [05:41:39] [PASSED] 19 VFs [05:41:39] [PASSED] 20 VFs [05:41:39] [PASSED] 21 VFs [05:41:39] [PASSED] 22 VFs [05:41:39] [PASSED] 23 VFs [05:41:39] [PASSED] 24 VFs [05:41:39] [PASSED] 25 VFs [05:41:39] [PASSED] 26 VFs [05:41:39] [PASSED] 27 VFs [05:41:39] [PASSED] 28 VFs [05:41:39] [PASSED] 29 VFs [05:41:39] [PASSED] 30 VFs [05:41:39] [PASSED] 31 VFs [05:41:39] [PASSED] 32 VFs [05:41:39] [PASSED] 33 VFs [05:41:39] [PASSED] 34 VFs [05:41:39] [PASSED] 35 VFs [05:41:39] [PASSED] 36 VFs [05:41:39] [PASSED] 37 VFs [05:41:39] [PASSED] 38 VFs [05:41:39] [PASSED] 39 VFs [05:41:39] [PASSED] 40 VFs [05:41:40] [PASSED] 41 VFs [05:41:40] [PASSED] 42 VFs [05:41:40] [PASSED] 43 VFs [05:41:40] [PASSED] 44 VFs [05:41:40] [PASSED] 45 VFs [05:41:40] [PASSED] 46 VFs [05:41:40] [PASSED] 47 VFs [05:41:40] [PASSED] 48 VFs [05:41:40] [PASSED] 49 VFs [05:41:40] [PASSED] 50 VFs [05:41:40] [PASSED] 51 VFs [05:41:40] [PASSED] 52 VFs [05:41:40] [PASSED] 53 VFs [05:41:40] [PASSED] 54 VFs [05:41:40] [PASSED] 55 VFs [05:41:40] [PASSED] 56 VFs [05:41:40] [PASSED] 57 VFs [05:41:40] [PASSED] 58 VFs [05:41:40] [PASSED] 59 VFs [05:41:40] [PASSED] 60 VFs [05:41:40] [PASSED] 61 VFs [05:41:40] [PASSED] 62 VFs [05:41:40] [PASSED] 63 VFs [05:41:40] ==================== [PASSED] fair_ggtt ==================== [05:41:40] ======================== fair_vram ======================== [05:41:40] [PASSED] 1 VF [05:41:40] [PASSED] 2 VFs [05:41:40] [PASSED] 3 VFs [05:41:40] [PASSED] 4 VFs [05:41:40] [PASSED] 5 VFs [05:41:40] [PASSED] 6 VFs [05:41:40] [PASSED] 7 VFs [05:41:40] [PASSED] 8 VFs [05:41:40] [PASSED] 9 VFs [05:41:40] [PASSED] 10 VFs [05:41:40] [PASSED] 11 VFs [05:41:40] [PASSED] 12 VFs [05:41:40] [PASSED] 13 VFs [05:41:40] [PASSED] 14 VFs [05:41:40] [PASSED] 15 VFs [05:41:40] [PASSED] 16 VFs [05:41:40] [PASSED] 17 VFs [05:41:40] [PASSED] 18 VFs [05:41:40] [PASSED] 19 VFs [05:41:40] [PASSED] 20 VFs [05:41:40] [PASSED] 21 VFs [05:41:40] [PASSED] 22 VFs [05:41:40] [PASSED] 23 VFs [05:41:40] [PASSED] 24 VFs [05:41:40] [PASSED] 25 VFs [05:41:40] [PASSED] 26 VFs [05:41:40] [PASSED] 27 VFs [05:41:40] [PASSED] 28 VFs [05:41:40] [PASSED] 29 VFs [05:41:40] [PASSED] 30 VFs [05:41:40] [PASSED] 31 VFs [05:41:40] [PASSED] 32 VFs [05:41:40] [PASSED] 33 VFs [05:41:40] [PASSED] 34 VFs [05:41:40] [PASSED] 35 VFs [05:41:40] [PASSED] 36 VFs [05:41:40] [PASSED] 37 VFs [05:41:40] [PASSED] 38 VFs [05:41:40] [PASSED] 39 VFs [05:41:40] [PASSED] 40 VFs [05:41:40] [PASSED] 41 VFs [05:41:40] [PASSED] 42 VFs [05:41:40] [PASSED] 43 VFs [05:41:40] [PASSED] 44 VFs [05:41:40] [PASSED] 45 VFs [05:41:40] [PASSED] 46 VFs [05:41:40] [PASSED] 47 VFs [05:41:40] [PASSED] 48 VFs [05:41:40] [PASSED] 49 VFs [05:41:40] [PASSED] 50 VFs [05:41:40] [PASSED] 51 VFs [05:41:40] [PASSED] 52 VFs [05:41:40] [PASSED] 53 VFs [05:41:40] [PASSED] 54 VFs [05:41:40] [PASSED] 55 VFs [05:41:40] [PASSED] 56 VFs [05:41:40] [PASSED] 57 VFs [05:41:40] [PASSED] 58 VFs [05:41:40] [PASSED] 59 VFs [05:41:40] [PASSED] 60 VFs [05:41:40] [PASSED] 61 VFs [05:41:40] [PASSED] 62 VFs [05:41:40] [PASSED] 63 VFs [05:41:40] ==================== [PASSED] fair_vram ==================== [05:41:40] ================== [PASSED] pf_gt_config =================== [05:41:40] ===================== lmtt (1 subtest) ===================== [05:41:40] ======================== test_ops ========================= [05:41:40] [PASSED] 2-level [05:41:40] [PASSED] multi-level [05:41:40] ==================== [PASSED] test_ops ===================== [05:41:40] ====================== [PASSED] lmtt ======================= [05:41:40] ================= sriov_packet (1 subtest) ================= [05:41:40] [PASSED] test_descriptor_init [05:41:40] ================== [PASSED] sriov_packet =================== [05:41:40] ================= pf_service (11 subtests) ================= [05:41:40] [PASSED] pf_negotiate_any [05:41:40] [PASSED] pf_negotiate_base_match [05:41:40] [PASSED] pf_negotiate_base_newer [05:41:40] [PASSED] pf_negotiate_base_next [05:41:40] [SKIPPED] pf_negotiate_base_older (no older minor) [05:41:40] [PASSED] pf_negotiate_base_prev [05:41:40] [PASSED] pf_negotiate_latest_match [05:41:40] [PASSED] pf_negotiate_latest_newer [05:41:40] [PASSED] pf_negotiate_latest_next [05:41:40] [SKIPPED] pf_negotiate_latest_older (no older minor) [05:41:40] [SKIPPED] pf_negotiate_latest_prev (no prev major) [05:41:40] =================== [PASSED] pf_service ==================== [05:41:40] ================= xe_guc_g2g (2 subtests) ================== [05:41:40] ============== xe_live_guc_g2g_kunit_default ============== [05:41:40] ========= [SKIPPED] xe_live_guc_g2g_kunit_default ========== [05:41:40] ============== xe_live_guc_g2g_kunit_allmem =============== [05:41:40] ========== [SKIPPED] xe_live_guc_g2g_kunit_allmem ========== [05:41:40] =================== [SKIPPED] xe_guc_g2g =================== [05:41:40] =================== xe_mocs (2 subtests) =================== [05:41:40] ================ xe_live_mocs_kernel_kunit ================ [05:41:40] =========== [SKIPPED] xe_live_mocs_kernel_kunit ============ [05:41:40] ================ xe_live_mocs_reset_kunit ================= [05:41:40] ============ [SKIPPED] xe_live_mocs_reset_kunit ============ [05:41:40] ==================== [SKIPPED] xe_mocs ===================== [05:41:40] ================= xe_migrate (2 subtests) ================== [05:41:40] ================= xe_migrate_sanity_kunit ================= [05:41:40] ============ [SKIPPED] xe_migrate_sanity_kunit ============= [05:41:40] ================== xe_validate_ccs_kunit ================== [05:41:40] ============= [SKIPPED] xe_validate_ccs_kunit ============== [05:41:40] =================== [SKIPPED] xe_migrate =================== [05:41:40] ================== xe_dma_buf (1 subtest) ================== [05:41:40] ==================== xe_dma_buf_kunit ===================== [05:41:40] ================ [SKIPPED] xe_dma_buf_kunit ================ [05:41:40] =================== [SKIPPED] xe_dma_buf =================== [05:41:40] ================= xe_bo_shrink (1 subtest) ================= [05:41:40] =================== xe_bo_shrink_kunit ==================== [05:41:40] =============== [SKIPPED] xe_bo_shrink_kunit =============== [05:41:40] ================== [SKIPPED] xe_bo_shrink ================== [05:41:40] ==================== xe_bo (2 subtests) ==================== [05:41:40] ================== xe_ccs_migrate_kunit =================== [05:41:40] ============== [SKIPPED] xe_ccs_migrate_kunit ============== [05:41:40] ==================== xe_bo_evict_kunit ==================== [05:41:40] =============== [SKIPPED] xe_bo_evict_kunit ================ [05:41:40] ===================== [SKIPPED] xe_bo ====================== [05:41:40] ==================== args (13 subtests) ==================== [05:41:40] [PASSED] count_args_test [05:41:40] [PASSED] call_args_example [05:41:40] [PASSED] call_args_test [05:41:40] [PASSED] drop_first_arg_example [05:41:40] [PASSED] drop_first_arg_test [05:41:40] [PASSED] first_arg_example [05:41:40] [PASSED] first_arg_test [05:41:40] [PASSED] last_arg_example [05:41:40] [PASSED] last_arg_test [05:41:40] [PASSED] pick_arg_example [05:41:40] [PASSED] if_args_example [05:41:40] [PASSED] if_args_test [05:41:40] [PASSED] sep_comma_example [05:41:40] ====================== [PASSED] args ======================= [05:41:40] =================== xe_pci (3 subtests) ==================== [05:41:40] ==================== check_graphics_ip ==================== [05:41:40] [PASSED] 12.00 Xe_LP [05:41:40] [PASSED] 12.10 Xe_LP+ [05:41:40] [PASSED] 12.55 Xe_HPG [05:41:40] [PASSED] 12.60 Xe_HPC [05:41:40] [PASSED] 12.70 Xe_LPG [05:41:40] [PASSED] 12.71 Xe_LPG [05:41:40] [PASSED] 12.74 Xe_LPG+ [05:41:40] [PASSED] 20.01 Xe2_HPG [05:41:40] [PASSED] 20.02 Xe2_HPG [05:41:40] [PASSED] 20.04 Xe2_LPG [05:41:40] [PASSED] 30.00 Xe3_LPG [05:41:40] [PASSED] 30.01 Xe3_LPG [05:41:40] [PASSED] 30.03 Xe3_LPG [05:41:40] [PASSED] 30.04 Xe3_LPG [05:41:40] [PASSED] 30.05 Xe3_LPG [05:41:40] [PASSED] 35.10 Xe3p_LPG [05:41:40] [PASSED] 35.11 Xe3p_XPC [05:41:40] ================ [PASSED] check_graphics_ip ================ [05:41:40] ===================== check_media_ip ====================== [05:41:40] [PASSED] 12.00 Xe_M [05:41:40] [PASSED] 12.55 Xe_HPM [05:41:40] [PASSED] 13.00 Xe_LPM+ [05:41:40] [PASSED] 13.01 Xe2_HPM [05:41:40] [PASSED] 20.00 Xe2_LPM [05:41:40] [PASSED] 30.00 Xe3_LPM [05:41:40] [PASSED] 30.02 Xe3_LPM [05:41:40] [PASSED] 35.00 Xe3p_LPM [05:41:40] [PASSED] 35.03 Xe3p_HPM [05:41:40] ================= [PASSED] check_media_ip ================== [05:41:40] =================== check_platform_desc =================== [05:41:40] [PASSED] 0x9A60 (TIGERLAKE) [05:41:40] [PASSED] 0x9A68 (TIGERLAKE) [05:41:40] [PASSED] 0x9A70 (TIGERLAKE) [05:41:40] [PASSED] 0x9A40 (TIGERLAKE) [05:41:40] [PASSED] 0x9A49 (TIGERLAKE) [05:41:40] [PASSED] 0x9A59 (TIGERLAKE) [05:41:40] [PASSED] 0x9A78 (TIGERLAKE) [05:41:40] [PASSED] 0x9AC0 (TIGERLAKE) [05:41:40] [PASSED] 0x9AC9 (TIGERLAKE) [05:41:40] [PASSED] 0x9AD9 (TIGERLAKE) [05:41:40] [PASSED] 0x9AF8 (TIGERLAKE) [05:41:40] [PASSED] 0x4C80 (ROCKETLAKE) [05:41:40] [PASSED] 0x4C8A (ROCKETLAKE) [05:41:40] [PASSED] 0x4C8B (ROCKETLAKE) [05:41:40] [PASSED] 0x4C8C (ROCKETLAKE) [05:41:40] [PASSED] 0x4C90 (ROCKETLAKE) [05:41:40] [PASSED] 0x4C9A (ROCKETLAKE) [05:41:40] [PASSED] 0x4680 (ALDERLAKE_S) [05:41:40] [PASSED] 0x4682 (ALDERLAKE_S) [05:41:40] [PASSED] 0x4688 (ALDERLAKE_S) [05:41:40] [PASSED] 0x468A (ALDERLAKE_S) [05:41:40] [PASSED] 0x468B (ALDERLAKE_S) [05:41:40] [PASSED] 0x4690 (ALDERLAKE_S) [05:41:40] [PASSED] 0x4692 (ALDERLAKE_S) [05:41:40] [PASSED] 0x4693 (ALDERLAKE_S) [05:41:40] [PASSED] 0x46A0 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46A1 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46A2 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46A3 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46A6 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46A8 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46AA (ALDERLAKE_P) [05:41:40] [PASSED] 0x462A (ALDERLAKE_P) [05:41:40] [PASSED] 0x4626 (ALDERLAKE_P) [05:41:40] [PASSED] 0x4628 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46B0 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46B1 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46B2 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46B3 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46C0 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46C1 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46C2 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46C3 (ALDERLAKE_P) [05:41:40] [PASSED] 0x46D0 (ALDERLAKE_N) [05:41:40] [PASSED] 0x46D1 (ALDERLAKE_N) [05:41:40] [PASSED] 0x46D2 (ALDERLAKE_N) [05:41:40] [PASSED] 0x46D3 (ALDERLAKE_N) [05:41:40] [PASSED] 0x46D4 (ALDERLAKE_N) [05:41:40] [PASSED] 0xA721 (ALDERLAKE_P) [05:41:40] [PASSED] 0xA7A1 (ALDERLAKE_P) [05:41:40] [PASSED] 0xA7A9 (ALDERLAKE_P) [05:41:40] [PASSED] 0xA7AC (ALDERLAKE_P) [05:41:40] [PASSED] 0xA7AD (ALDERLAKE_P) [05:41:40] [PASSED] 0xA720 (ALDERLAKE_P) [05:41:40] [PASSED] 0xA7A0 (ALDERLAKE_P) [05:41:40] [PASSED] 0xA7A8 (ALDERLAKE_P) [05:41:40] [PASSED] 0xA7AA (ALDERLAKE_P) [05:41:40] [PASSED] 0xA7AB (ALDERLAKE_P) [05:41:40] [PASSED] 0xA780 (ALDERLAKE_S) [05:41:40] [PASSED] 0xA781 (ALDERLAKE_S) [05:41:40] [PASSED] 0xA782 (ALDERLAKE_S) [05:41:40] [PASSED] 0xA783 (ALDERLAKE_S) [05:41:40] [PASSED] 0xA788 (ALDERLAKE_S) [05:41:40] [PASSED] 0xA789 (ALDERLAKE_S) [05:41:40] [PASSED] 0xA78A (ALDERLAKE_S) [05:41:40] [PASSED] 0xA78B (ALDERLAKE_S) [05:41:40] [PASSED] 0x4905 (DG1) [05:41:40] [PASSED] 0x4906 (DG1) [05:41:40] [PASSED] 0x4907 (DG1) [05:41:40] [PASSED] 0x4908 (DG1) [05:41:40] [PASSED] 0x4909 (DG1) [05:41:40] [PASSED] 0x56C0 (DG2) [05:41:40] [PASSED] 0x56C2 (DG2) [05:41:40] [PASSED] 0x56C1 (DG2) [05:41:40] [PASSED] 0x7D51 (METEORLAKE) [05:41:40] [PASSED] 0x7DD1 (METEORLAKE) [05:41:40] [PASSED] 0x7D41 (METEORLAKE) [05:41:40] [PASSED] 0x7D67 (METEORLAKE) [05:41:40] [PASSED] 0xB640 (METEORLAKE) [05:41:40] [PASSED] 0x56A0 (DG2) [05:41:40] [PASSED] 0x56A1 (DG2) [05:41:40] [PASSED] 0x56A2 (DG2) [05:41:40] [PASSED] 0x56BE (DG2) [05:41:40] [PASSED] 0x56BF (DG2) [05:41:40] [PASSED] 0x5690 (DG2) [05:41:40] [PASSED] 0x5691 (DG2) [05:41:40] [PASSED] 0x5692 (DG2) [05:41:40] [PASSED] 0x56A5 (DG2) [05:41:40] [PASSED] 0x56A6 (DG2) [05:41:40] [PASSED] 0x56B0 (DG2) [05:41:40] [PASSED] 0x56B1 (DG2) [05:41:40] [PASSED] 0x56BA (DG2) [05:41:40] [PASSED] 0x56BB (DG2) [05:41:40] [PASSED] 0x56BC (DG2) [05:41:40] [PASSED] 0x56BD (DG2) [05:41:40] [PASSED] 0x5693 (DG2) [05:41:40] [PASSED] 0x5694 (DG2) [05:41:40] [PASSED] 0x5695 (DG2) [05:41:40] [PASSED] 0x56A3 (DG2) [05:41:40] [PASSED] 0x56A4 (DG2) [05:41:40] [PASSED] 0x56B2 (DG2) [05:41:40] [PASSED] 0x56B3 (DG2) [05:41:40] [PASSED] 0x5696 (DG2) [05:41:40] [PASSED] 0x5697 (DG2) [05:41:40] [PASSED] 0xB69 (PVC) [05:41:40] [PASSED] 0xB6E (PVC) [05:41:40] [PASSED] 0xBD4 (PVC) [05:41:40] [PASSED] 0xBD5 (PVC) [05:41:40] [PASSED] 0xBD6 (PVC) [05:41:40] [PASSED] 0xBD7 (PVC) [05:41:40] [PASSED] 0xBD8 (PVC) [05:41:40] [PASSED] 0xBD9 (PVC) [05:41:40] [PASSED] 0xBDA (PVC) [05:41:40] [PASSED] 0xBDB (PVC) [05:41:40] [PASSED] 0xBE0 (PVC) [05:41:40] [PASSED] 0xBE1 (PVC) [05:41:40] [PASSED] 0xBE5 (PVC) [05:41:40] [PASSED] 0x7D40 (METEORLAKE) [05:41:40] [PASSED] 0x7D45 (METEORLAKE) [05:41:40] [PASSED] 0x7D55 (METEORLAKE) [05:41:40] [PASSED] 0x7D60 (METEORLAKE) [05:41:40] [PASSED] 0x7DD5 (METEORLAKE) [05:41:40] [PASSED] 0x6420 (LUNARLAKE) [05:41:40] [PASSED] 0x64A0 (LUNARLAKE) [05:41:40] [PASSED] 0x64B0 (LUNARLAKE) [05:41:40] [PASSED] 0xE202 (BATTLEMAGE) [05:41:40] [PASSED] 0xE209 (BATTLEMAGE) [05:41:40] [PASSED] 0xE20B (BATTLEMAGE) [05:41:40] [PASSED] 0xE20C (BATTLEMAGE) [05:41:40] [PASSED] 0xE20D (BATTLEMAGE) [05:41:40] [PASSED] 0xE210 (BATTLEMAGE) [05:41:40] [PASSED] 0xE211 (BATTLEMAGE) [05:41:40] [PASSED] 0xE212 (BATTLEMAGE) [05:41:40] [PASSED] 0xE216 (BATTLEMAGE) [05:41:40] [PASSED] 0xE220 (BATTLEMAGE) [05:41:40] [PASSED] 0xE221 (BATTLEMAGE) [05:41:40] [PASSED] 0xE222 (BATTLEMAGE) [05:41:40] [PASSED] 0xE223 (BATTLEMAGE) [05:41:40] [PASSED] 0xB080 (PANTHERLAKE) [05:41:40] [PASSED] 0xB081 (PANTHERLAKE) [05:41:40] [PASSED] 0xB082 (PANTHERLAKE) [05:41:40] [PASSED] 0xB083 (PANTHERLAKE) [05:41:40] [PASSED] 0xB084 (PANTHERLAKE) [05:41:40] [PASSED] 0xB085 (PANTHERLAKE) [05:41:40] [PASSED] 0xB086 (PANTHERLAKE) [05:41:40] [PASSED] 0xB087 (PANTHERLAKE) [05:41:40] [PASSED] 0xB08F (PANTHERLAKE) [05:41:40] [PASSED] 0xB090 (PANTHERLAKE) [05:41:40] [PASSED] 0xB0A0 (PANTHERLAKE) [05:41:40] [PASSED] 0xB0B0 (PANTHERLAKE) [05:41:40] [PASSED] 0xFD80 (PANTHERLAKE) [05:41:40] [PASSED] 0xFD81 (PANTHERLAKE) [05:41:40] [PASSED] 0xD740 (NOVALAKE_S) [05:41:40] [PASSED] 0xD741 (NOVALAKE_S) [05:41:40] [PASSED] 0xD742 (NOVALAKE_S) [05:41:40] [PASSED] 0xD743 (NOVALAKE_S) [05:41:40] [PASSED] 0xD745 (NOVALAKE_S) [05:41:40] [PASSED] 0xD74A (NOVALAKE_S) [05:41:40] [PASSED] 0xD74B (NOVALAKE_S) [05:41:40] [PASSED] 0x674C (CRESCENTISLAND) [05:41:40] [PASSED] 0x674D (CRESCENTISLAND) [05:41:40] [PASSED] 0x674E (CRESCENTISLAND) [05:41:40] [PASSED] 0x674F (CRESCENTISLAND) [05:41:40] [PASSED] 0x6750 (CRESCENTISLAND) [05:41:40] [PASSED] 0xD750 (NOVALAKE_P) [05:41:40] [PASSED] 0xD751 (NOVALAKE_P) [05:41:40] [PASSED] 0xD752 (NOVALAKE_P) [05:41:40] [PASSED] 0xD753 (NOVALAKE_P) [05:41:40] [PASSED] 0xD754 (NOVALAKE_P) [05:41:40] [PASSED] 0xD755 (NOVALAKE_P) [05:41:40] [PASSED] 0xD756 (NOVALAKE_P) [05:41:40] [PASSED] 0xD757 (NOVALAKE_P) [05:41:40] [PASSED] 0xD75F (NOVALAKE_P) [05:41:40] =============== [PASSED] check_platform_desc =============== [05:41:40] ===================== [PASSED] xe_pci ====================== [05:41:40] ============= xe_rtp_tables_test (5 subtests) ============== [05:41:40] ================== xe_rtp_table_gt_test =================== [05:41:40] [PASSED] gt_was/14011060649 [05:41:40] [PASSED] gt_was/14011059788 [05:41:40] [PASSED] gt_was/14015795083 [05:41:40] [PASSED] gt_was/16021867713 [05:41:40] [PASSED] gt_was/14019449301 [05:41:40] [PASSED] gt_was/16028005424 [05:41:40] [PASSED] gt_was/14026578760 [05:41:40] [PASSED] gt_was/1409420604 [05:41:40] [PASSED] gt_was/1408615072 [05:41:40] [PASSED] gt_was/22010523718 [05:41:40] [PASSED] gt_was/14011006942 [05:41:40] [PASSED] gt_was/14014830051 [05:41:40] [PASSED] gt_was/18018781329 [05:41:40] [PASSED] gt_was/1509235366 [05:41:40] [PASSED] gt_was/18018781329 [05:41:40] [PASSED] gt_was/16016694945 [05:41:40] [PASSED] gt_was/14018575942 [05:41:40] [PASSED] gt_was/22016670082 [05:41:40] [PASSED] gt_was/22016670082 [05:41:40] [PASSED] gt_was/14017421178 [05:41:40] [PASSED] gt_was/16025250150 [05:41:40] [PASSED] gt_was/14021871409 [05:41:40] [PASSED] gt_was/16021865536 [05:41:40] [PASSED] gt_was/14021486841 [05:41:40] [PASSED] gt_was/14025160223 [05:41:40] [PASSED] gt_was/14026144927, 16029437861, 14026127056 [05:41:40] [PASSED] gt_was/14025635424 [05:41:40] [PASSED] gt_was/16028005424 [05:41:40] ============== [PASSED] xe_rtp_table_gt_test =============== [05:41:40] ================== xe_rtp_table_gt_test =================== [05:41:40] [PASSED] gt_tunings/Tuning: Blend Fill Caching Optimization Disable [05:41:40] [PASSED] gt_tunings/Tuning: 32B Access Enable [05:41:40] [PASSED] gt_tunings/Tuning: L3 cache [05:41:40] [PASSED] gt_tunings/Tuning: L3 cache - media [05:41:40] [PASSED] gt_tunings/Tuning: Compression Overfetch [05:41:40] [PASSED] gt_tunings/Tuning: Compression Overfetch - media [05:41:40] [PASSED] gt_tunings/Tuning: Enable compressible partial write overfetch in L3 [05:41:40] [PASSED] gt_tunings/Tuning: Enable compressible partial write overfetch in L3 - media [05:41:40] [PASSED] gt_tunings/Tuning: L2 Overfetch Compressible Only [05:41:40] [PASSED] gt_tunings/Tuning: L2 Overfetch Compressible Only - media [05:41:40] [PASSED] gt_tunings/Tuning: Stateless compression control [05:41:40] [PASSED] gt_tunings/Tuning: Stateless compression control - media [05:41:40] [PASSED] gt_tunings/Tuning: L3 RW flush all Cache [05:41:40] [PASSED] gt_tunings/Tuning: L3 RW flush all cache - media [05:41:40] [PASSED] gt_tunings/Tuning: Set STLB Bank Hash Mode to 4KB [05:41:40] ============== [PASSED] xe_rtp_table_gt_test =============== [05:41:40] ================== xe_rtp_table_oob_test ================== [05:41:40] [PASSED] oob_was/1607983814 [05:41:40] [PASSED] oob_was/16010904313 [05:41:40] [PASSED] oob_was/18022495364 [05:41:40] [PASSED] oob_was/22012773006 [05:41:40] [PASSED] oob_was/14014475959 [05:41:40] [PASSED] oob_was/22011391025 [05:41:40] [PASSED] oob_was/22012727170 [05:41:40] [PASSED] oob_was/22012727685 [05:41:40] [PASSED] oob_was/22016596838 [05:41:40] [PASSED] oob_was/18020744125 [05:41:40] [PASSED] oob_was/1409600907 [05:41:40] [PASSED] oob_was/22014953428 [05:41:40] [PASSED] oob_was/16017236439 [05:41:40] [PASSED] oob_was/14019821291 [05:41:40] [PASSED] oob_was/14015076503 [05:41:40] [PASSED] oob_was/14018913170 [05:41:40] [PASSED] oob_was/14018094691 [05:41:40] [PASSED] oob_was/18024947630 [05:41:40] [PASSED] oob_was/16022287689 [05:41:40] [PASSED] oob_was/13011645652 [05:41:40] [PASSED] oob_was/14022293748 [05:41:40] [PASSED] oob_was/22019794406 [05:41:40] [PASSED] oob_was/22019338487 [05:41:40] [PASSED] oob_was/16023588340 [05:41:40] [PASSED] oob_was/14019789679 [05:41:40] [PASSED] oob_was/14022866841 [05:41:40] [PASSED] oob_was/16021333562 [05:41:40] [PASSED] oob_was/14016712196 [05:41:40] [PASSED] oob_was/14015568240 [05:41:40] [PASSED] oob_was/18013179988 [05:41:40] [PASSED] oob_was/1508761755 [05:41:40] [PASSED] oob_was/16023105232 [05:41:40] [PASSED] oob_was/16026508708 [05:41:40] [PASSED] oob_was/14020001231 [05:41:40] [PASSED] oob_was/16023683509 [05:41:40] [PASSED] oob_was/14025515070 [05:41:40] [PASSED] oob_was/15015404425_disable [05:41:40] [PASSED] oob_was/16026007364 [05:41:40] [PASSED] oob_was/14020316580 [05:41:40] [PASSED] oob_was/14025883347 [05:41:40] [PASSED] oob_was/16029380221 [05:41:40] [PASSED] oob_was/22022079272 [05:41:40] [PASSED] oob_was/16029897822 [05:41:40] ============== [PASSED] xe_rtp_table_oob_test ============== [05:41:40] ================ xe_rtp_table_dev_oob_test ================ [05:41:40] [PASSED] device_oob_was/22010954014 [05:41:40] [PASSED] device_oob_was/15015404425 [05:41:40] [PASSED] device_oob_was/22019338487_display [05:41:40] [PASSED] device_oob_was/14022085890 [05:41:40] [PASSED] device_oob_was/14026539277 [05:41:40] [PASSED] device_oob_was/14026633728 [05:41:40] [PASSED] device_oob_was/14026746987 [05:41:40] [PASSED] device_oob_was/14026779378 [05:41:40] ============ [PASSED] xe_rtp_table_dev_oob_test ============ [05:41:40] ========== xe_rtp_table_missing_upper_bound_test ========== [05:41:40] [PASSED] register_whitelist/WaAllowPMDepthAndInvocationCountAccessFromUMD, 1408556865 [05:41:40] [PASSED] register_whitelist/1508744258, 14012131227, 1808121037 [05:41:40] [PASSED] register_whitelist/1806527549 [05:41:40] [PASSED] register_whitelist/allow_read_ctx_timestamp [05:41:40] [PASSED] register_whitelist/allow_read_queue_timestamp [05:41:40] [PASSED] register_whitelist/16014440446 [05:41:40] [PASSED] register_whitelist/16017236439 [05:41:40] [PASSED] register_whitelist/16020183090 [05:41:40] [PASSED] register_whitelist/14024997852 [05:41:40] [PASSED] register_whitelist/14024997852 [05:41:40] ====== [PASSED] xe_rtp_table_missing_upper_bound_test ====== [05:41:40] =============== [PASSED] xe_rtp_tables_test ================ [05:41:40] =================== xe_rtp (3 subtests) ==================== [05:41:40] =================== xe_rtp_rules_tests ==================== [05:41:40] [PASSED] no [05:41:40] [PASSED] yes [05:41:40] [PASSED] no-and-no [05:41:40] [PASSED] no-and-yes [05:41:40] [PASSED] yes-and-no [05:41:40] [PASSED] yes-and-yes [05:41:40] [PASSED] no-or-no [05:41:40] [PASSED] no-or-yes [05:41:40] [PASSED] yes-or-no [05:41:40] [PASSED] yes-or-yes [05:41:40] [PASSED] no-yes-or-yes-no [05:41:40] [PASSED] no-yes-or-yes-yes [05:41:40] [PASSED] yes-yes-or-no-yes [05:41:40] [PASSED] yes-yes-or-yes-yes [05:41:40] [PASSED] no-no-or-yes-or-no [05:41:40] [PASSED] or [05:41:40] [PASSED] or-yes [05:41:40] [PASSED] or-no [05:41:40] [PASSED] yes-or [05:41:40] [PASSED] no-or [05:41:40] [PASSED] no-or-or-yes [05:41:40] [PASSED] yes-or-or-no [05:41:40] [PASSED] no-or-or-no [05:41:40] [PASSED] missing-context-engine-class [05:41:40] [PASSED] missing-context-engine-class-or-yes [05:41:40] [PASSED] missing-context-engine-class-or-or-yes [05:41:40] =============== [PASSED] xe_rtp_rules_tests ================ [05:41:40] =============== xe_rtp_process_to_sr_tests ================ [05:41:40] [PASSED] coalesce-same-reg [05:41:40] [PASSED] coalesce-same-reg-literal-and-func [05:41:40] [PASSED] no-match-no-add [05:41:40] [PASSED] two-regs-two-entries [05:41:40] [PASSED] clr-one-set-other [05:41:40] [PASSED] set-field [05:41:40] [PASSED] conflict-duplicate [05:41:40] [PASSED] conflict-not-disjoint [05:41:40] [PASSED] conflict-not-disjoint-literal-and-func [05:41:40] [PASSED] conflict-reg-type [05:41:40] [PASSED] bad-mcr-reg-forced-to-regular [05:41:40] [PASSED] bad-regular-reg-forced-to-mcr [05:41:40] =========== [PASSED] xe_rtp_process_to_sr_tests ============ [05:41:40] ================== xe_rtp_process_tests =================== [05:41:40] [PASSED] active1 [05:41:40] [PASSED] active2 [05:41:40] [PASSED] active-inactive [05:41:40] [PASSED] inactive-active [05:41:40] [PASSED] inactive-active-inactive [05:41:40] [PASSED] inactive-inactive-inactive [05:41:40] ============== [PASSED] xe_rtp_process_tests =============== [05:41:40] ===================== [PASSED] xe_rtp ====================== [05:41:40] ==================== xe_wa (1 subtest) ===================== [05:41:40] ======================== xe_wa_gt ========================= [05:41:40] [PASSED] TIGERLAKE B0 [05:41:40] [PASSED] DG1 A0 [05:41:40] [PASSED] DG1 B0 [05:41:40] [PASSED] ALDERLAKE_S A0 [05:41:40] [PASSED] ALDERLAKE_S B0 [05:41:40] [PASSED] ALDERLAKE_S C0 [05:41:40] [PASSED] ALDERLAKE_S D0 [05:41:40] [PASSED] ALDERLAKE_P A0 [05:41:40] [PASSED] ALDERLAKE_P B0 [05:41:40] [PASSED] ALDERLAKE_P C0 [05:41:40] [PASSED] ALDERLAKE_S RPLS D0 [05:41:40] [PASSED] ALDERLAKE_P RPLU E0 [05:41:40] [PASSED] DG2 G10 C0 [05:41:40] [PASSED] DG2 G11 B1 [05:41:40] [PASSED] DG2 G12 A1 [05:41:40] [PASSED] METEORLAKE 12.70(Xe_LPG) A0 13.00(Xe_LPM+) A0 [05:41:40] [PASSED] METEORLAKE 12.71(Xe_LPG) A0 13.00(Xe_LPM+) A0 [05:41:40] [PASSED] METEORLAKE 12.74(Xe_LPG+) A0 13.00(Xe_LPM+) A0 [05:41:40] [PASSED] LUNARLAKE 20.04(Xe2_LPG) A0 20.00(Xe2_LPM) A0 [05:41:40] [PASSED] LUNARLAKE 20.04(Xe2_LPG) B0 20.00(Xe2_LPM) A0 [05:41:40] [PASSED] BATTLEMAGE 20.01(Xe2_HPG) A0 13.01(Xe2_HPM) A1 [05:41:40] [PASSED] PANTHERLAKE 30.00(Xe3_LPG) A0 30.00(Xe3_LPM) A0 [05:41:40] ==================== [PASSED] xe_wa_gt ===================== [05:41:40] ====================== [PASSED] xe_wa ====================== [05:41:40] ============================================================ [05:41:40] Testing complete. Ran 741 tests: passed: 723, skipped: 18 [05:41:40] Elapsed time: 36.599s total, 4.262s configuring, 31.672s building, 0.637s running
+ /kernel/tools/testing/kunit/kunit.py run --kunitconfig /kernel/drivers/gpu/drm/tests/.kunitconfig [05:41:40] Configuring KUnit Kernel ...
Regenerating .config ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig [05:41:42] Building KUnit Kernel ... Populating config with: $ make ARCH=um O=.kunit olddefconfig Building with: $ make all compile_commands.json scripts_gdb ARCH=um O=.kunit --jobs=48 [05:42:06] Starting KUnit Kernel (1/1)... [05:42:06] ============================================================ Running tests with: $ .kunit/linux kunit.enable=1 mem=1G console=tty kunit_shutdown=halt [05:42:07] ============ drm_test_pick_cmdline (2 subtests) ============ [05:42:07] [PASSED] drm_test_pick_cmdline_res_1920_1080_60 [05:42:07] =============== drm_test_pick_cmdline_named =============== [05:42:07] [PASSED] NTSC [05:42:07] [PASSED] NTSC-J [05:42:07] [PASSED] PAL [05:42:07] [PASSED] PAL-M [05:42:07] =========== [PASSED] drm_test_pick_cmdline_named =========== [05:42:07] ============== [PASSED] drm_test_pick_cmdline ============== [05:42:07] == drm_test_atomic_get_connector_for_encoder (1 subtest) === [05:42:07] [PASSED] drm_test_drm_atomic_get_connector_for_encoder [05:42:07] ==== [PASSED] drm_test_atomic_get_connector_for_encoder ==== [05:42:07] =========== drm_validate_clone_mode (2 subtests) =========== [05:42:07] ============== drm_test_check_in_clone_mode =============== [05:42:07] [PASSED] in_clone_mode [05:42:07] [PASSED] not_in_clone_mode [05:42:07] ========== [PASSED] drm_test_check_in_clone_mode =========== [05:42:07] =============== drm_test_check_valid_clones =============== [05:42:07] [PASSED] not_in_clone_mode [05:42:07] [PASSED] valid_clone [05:42:07] [PASSED] invalid_clone [05:42:07] =========== [PASSED] drm_test_check_valid_clones =========== [05:42:07] ============= [PASSED] drm_validate_clone_mode ============= [05:42:07] ============= drm_validate_modeset (1 subtest) ============= [05:42:07] [PASSED] drm_test_check_connector_changed_modeset [05:42:07] ============== [PASSED] drm_validate_modeset =============== [05:42:07] ====== drm_test_bridge_get_current_state (1 subtest) ======= [05:42:07] [PASSED] drm_test_drm_bridge_get_current_state_atomic [05:42:07] ======== [PASSED] drm_test_bridge_get_current_state ======== [05:42:07] ====== drm_test_bridge_helper_reset_crtc (3 subtests) ====== [05:42:07] [PASSED] drm_test_drm_bridge_helper_reset_crtc_atomic [05:42:07] [PASSED] drm_test_drm_bridge_helper_reset_crtc_atomic_disabled [05:42:07] [PASSED] drm_test_drm_bridge_helper_hdmi_output_bus_fmts [05:42:07] ======== [PASSED] drm_test_bridge_helper_reset_crtc ======== [05:42:07] ============== drm_bridge_alloc (2 subtests) =============== [05:42:07] [PASSED] drm_test_drm_bridge_alloc_basic [05:42:07] [PASSED] drm_test_drm_bridge_alloc_get_put [05:42:07] ================ [PASSED] drm_bridge_alloc ================= [05:42:07] ============= drm_bridge_bus_fmt (5 subtests) ============== [05:42:07] [PASSED] drm_test_bridge_rgb_yuv_rgb [05:42:07] [PASSED] drm_test_bridge_must_convert_to_yuv444 [05:42:07] [PASSED] drm_test_bridge_hdmi_auto_rgb [05:42:07] [PASSED] drm_test_bridge_auto_first [05:42:07] [PASSED] drm_test_bridge_rgb_yuv_no_path [05:42:07] =============== [PASSED] drm_bridge_bus_fmt ================ [05:42:07] ============= drm_cmdline_parser (40 subtests) ============= [05:42:07] [PASSED] drm_test_cmdline_force_d_only [05:42:07] [PASSED] drm_test_cmdline_force_D_only_dvi [05:42:07] [PASSED] drm_test_cmdline_force_D_only_hdmi [05:42:07] [PASSED] drm_test_cmdline_force_D_only_not_digital [05:42:07] [PASSED] drm_test_cmdline_force_e_only [05:42:07] [PASSED] drm_test_cmdline_res [05:42:07] [PASSED] drm_test_cmdline_res_vesa [05:42:07] [PASSED] drm_test_cmdline_res_vesa_rblank [05:42:07] [PASSED] drm_test_cmdline_res_rblank [05:42:07] [PASSED] drm_test_cmdline_res_bpp [05:42:07] [PASSED] drm_test_cmdline_res_refresh [05:42:07] [PASSED] drm_test_cmdline_res_bpp_refresh [05:42:07] [PASSED] drm_test_cmdline_res_bpp_refresh_interlaced [05:42:07] [PASSED] drm_test_cmdline_res_bpp_refresh_margins [05:42:07] [PASSED] drm_test_cmdline_res_bpp_refresh_force_off [05:42:07] [PASSED] drm_test_cmdline_res_bpp_refresh_force_on [05:42:07] [PASSED] drm_test_cmdline_res_bpp_refresh_force_on_analog [05:42:07] [PASSED] drm_test_cmdline_res_bpp_refresh_force_on_digital [05:42:07] [PASSED] drm_test_cmdline_res_bpp_refresh_interlaced_margins_force_on [05:42:07] [PASSED] drm_test_cmdline_res_margins_force_on [05:42:07] [PASSED] drm_test_cmdline_res_vesa_margins [05:42:07] [PASSED] drm_test_cmdline_name [05:42:07] [PASSED] drm_test_cmdline_name_bpp [05:42:07] [PASSED] drm_test_cmdline_name_option [05:42:07] [PASSED] drm_test_cmdline_name_bpp_option [05:42:07] [PASSED] drm_test_cmdline_rotate_0 [05:42:07] [PASSED] drm_test_cmdline_rotate_90 [05:42:07] [PASSED] drm_test_cmdline_rotate_180 [05:42:07] [PASSED] drm_test_cmdline_rotate_270 [05:42:07] [PASSED] drm_test_cmdline_hmirror [05:42:07] [PASSED] drm_test_cmdline_vmirror [05:42:07] [PASSED] drm_test_cmdline_margin_options [05:42:07] [PASSED] drm_test_cmdline_multiple_options [05:42:07] [PASSED] drm_test_cmdline_bpp_extra_and_option [05:42:07] [PASSED] drm_test_cmdline_extra_and_option [05:42:07] [PASSED] drm_test_cmdline_freestanding_options [05:42:07] [PASSED] drm_test_cmdline_freestanding_force_e_and_options [05:42:07] [PASSED] drm_test_cmdline_panel_orientation [05:42:07] ================ drm_test_cmdline_invalid ================= [05:42:07] [PASSED] margin_only [05:42:07] [PASSED] interlace_only [05:42:07] [PASSED] res_missing_x [05:42:07] [PASSED] res_missing_y [05:42:07] [PASSED] res_bad_y [05:42:07] [PASSED] res_missing_y_bpp [05:42:07] [PASSED] res_bad_bpp [05:42:07] [PASSED] res_bad_refresh [05:42:07] [PASSED] res_bpp_refresh_force_on_off [05:42:07] [PASSED] res_invalid_mode [05:42:07] [PASSED] res_bpp_wrong_place_mode [05:42:07] [PASSED] name_bpp_refresh [05:42:07] [PASSED] name_refresh [05:42:07] [PASSED] name_refresh_wrong_mode [05:42:07] [PASSED] name_refresh_invalid_mode [05:42:07] [PASSED] rotate_multiple [05:42:07] [PASSED] rotate_invalid_val [05:42:07] [PASSED] rotate_truncated [05:42:07] [PASSED] invalid_option [05:42:07] [PASSED] invalid_tv_option [05:42:07] [PASSED] truncated_tv_option [05:42:07] ============ [PASSED] drm_test_cmdline_invalid ============= [05:42:07] =============== drm_test_cmdline_tv_options =============== [05:42:07] [PASSED] NTSC [05:42:07] [PASSED] NTSC_443 [05:42:07] [PASSED] NTSC_J [05:42:07] [PASSED] PAL [05:42:07] [PASSED] PAL_M [05:42:07] [PASSED] PAL_N [05:42:07] [PASSED] SECAM [05:42:07] [PASSED] MONO_525 [05:42:07] [PASSED] MONO_625 [05:42:07] =========== [PASSED] drm_test_cmdline_tv_options =========== [05:42:07] =============== [PASSED] drm_cmdline_parser ================ [05:42:07] ========== drmm_connector_hdmi_init (20 subtests) ========== [05:42:07] [PASSED] drm_test_connector_hdmi_init_valid [05:42:07] [PASSED] drm_test_connector_hdmi_init_bpc_8 [05:42:07] [PASSED] drm_test_connector_hdmi_init_bpc_10 [05:42:07] [PASSED] drm_test_connector_hdmi_init_bpc_12 [05:42:07] [PASSED] drm_test_connector_hdmi_init_bpc_invalid [05:42:07] [PASSED] drm_test_connector_hdmi_init_bpc_null [05:42:07] [PASSED] drm_test_connector_hdmi_init_formats_empty [05:42:07] [PASSED] drm_test_connector_hdmi_init_formats_no_rgb [05:42:07] === drm_test_connector_hdmi_init_formats_yuv420_allowed === [05:42:07] [PASSED] supported_formats=0x9 yuv420_allowed=1 [05:42:07] [PASSED] supported_formats=0x9 yuv420_allowed=0 [05:42:07] [PASSED] supported_formats=0x5 yuv420_allowed=1 [05:42:07] [PASSED] supported_formats=0x5 yuv420_allowed=0 [05:42:07] === [PASSED] drm_test_connector_hdmi_init_formats_yuv420_allowed === [05:42:07] [PASSED] drm_test_connector_hdmi_init_null_ddc [05:42:07] [PASSED] drm_test_connector_hdmi_init_null_product [05:42:07] [PASSED] drm_test_connector_hdmi_init_null_vendor [05:42:07] [PASSED] drm_test_connector_hdmi_init_product_length_exact [05:42:07] [PASSED] drm_test_connector_hdmi_init_product_length_too_long [05:42:07] [PASSED] drm_test_connector_hdmi_init_product_valid [05:42:07] [PASSED] drm_test_connector_hdmi_init_vendor_length_exact [05:42:07] [PASSED] drm_test_connector_hdmi_init_vendor_length_too_long [05:42:07] [PASSED] drm_test_connector_hdmi_init_vendor_valid [05:42:07] ========= drm_test_connector_hdmi_init_type_valid ========= [05:42:07] [PASSED] HDMI-A [05:42:07] [PASSED] HDMI-B [05:42:07] ===== [PASSED] drm_test_connector_hdmi_init_type_valid ===== [05:42:07] ======== drm_test_connector_hdmi_init_type_invalid ======== [05:42:07] [PASSED] Unknown [05:42:07] [PASSED] VGA [05:42:07] [PASSED] DVI-I [05:42:07] [PASSED] DVI-D [05:42:07] [PASSED] DVI-A [05:42:07] [PASSED] Composite [05:42:07] [PASSED] SVIDEO [05:42:07] [PASSED] LVDS [05:42:07] [PASSED] Component [05:42:07] [PASSED] DIN [05:42:07] [PASSED] DP [05:42:07] [PASSED] TV [05:42:07] [PASSED] eDP [05:42:07] [PASSED] Virtual [05:42:07] [PASSED] DSI [05:42:07] [PASSED] DPI [05:42:07] [PASSED] Writeback [05:42:07] [PASSED] SPI [05:42:07] [PASSED] USB [05:42:07] ==== [PASSED] drm_test_connector_hdmi_init_type_invalid ==== [05:42:07] ============ [PASSED] drmm_connector_hdmi_init ============= [05:42:07] ============= drmm_connector_init (3 subtests) ============= [05:42:07] [PASSED] drm_test_drmm_connector_init [05:42:07] [PASSED] drm_test_drmm_connector_init_null_ddc [05:42:07] ========= drm_test_drmm_connector_init_type_valid ========= [05:42:07] [PASSED] Unknown [05:42:07] [PASSED] VGA [05:42:07] [PASSED] DVI-I [05:42:07] [PASSED] DVI-D [05:42:07] [PASSED] DVI-A [05:42:07] [PASSED] Composite [05:42:07] [PASSED] SVIDEO [05:42:07] [PASSED] LVDS [05:42:07] [PASSED] Component [05:42:07] [PASSED] DIN [05:42:07] [PASSED] DP [05:42:07] [PASSED] HDMI-A [05:42:07] [PASSED] HDMI-B [05:42:07] [PASSED] TV [05:42:07] [PASSED] eDP [05:42:07] [PASSED] Virtual [05:42:07] [PASSED] DSI [05:42:07] [PASSED] DPI [05:42:07] [PASSED] Writeback [05:42:07] [PASSED] SPI [05:42:07] [PASSED] USB [05:42:07] ===== [PASSED] drm_test_drmm_connector_init_type_valid ===== [05:42:07] =============== [PASSED] drmm_connector_init =============== [05:42:07] ========= drm_connector_dynamic_init (6 subtests) ========== [05:42:07] [PASSED] drm_test_drm_connector_dynamic_init [05:42:07] [PASSED] drm_test_drm_connector_dynamic_init_null_ddc [05:42:07] [PASSED] drm_test_drm_connector_dynamic_init_not_added [05:42:07] [PASSED] drm_test_drm_connector_dynamic_init_properties [05:42:07] ===== drm_test_drm_connector_dynamic_init_type_valid ====== [05:42:07] [PASSED] Unknown [05:42:07] [PASSED] VGA [05:42:07] [PASSED] DVI-I [05:42:07] [PASSED] DVI-D [05:42:07] [PASSED] DVI-A [05:42:07] [PASSED] Composite [05:42:07] [PASSED] SVIDEO [05:42:07] [PASSED] LVDS [05:42:07] [PASSED] Component [05:42:07] [PASSED] DIN [05:42:07] [PASSED] DP [05:42:07] [PASSED] HDMI-A [05:42:07] [PASSED] HDMI-B [05:42:07] [PASSED] TV [05:42:07] [PASSED] eDP [05:42:07] [PASSED] Virtual [05:42:07] [PASSED] DSI [05:42:07] [PASSED] DPI [05:42:07] [PASSED] Writeback [05:42:07] [PASSED] SPI [05:42:07] [PASSED] USB [05:42:07] = [PASSED] drm_test_drm_connector_dynamic_init_type_valid == [05:42:07] ======== drm_test_drm_connector_dynamic_init_name ========= [05:42:07] [PASSED] Unknown [05:42:07] [PASSED] VGA [05:42:07] [PASSED] DVI-I [05:42:07] [PASSED] DVI-D [05:42:07] [PASSED] DVI-A [05:42:07] [PASSED] Composite [05:42:07] [PASSED] SVIDEO [05:42:07] [PASSED] LVDS [05:42:07] [PASSED] Component [05:42:07] [PASSED] DIN [05:42:07] [PASSED] DP [05:42:07] [PASSED] HDMI-A [05:42:07] [PASSED] HDMI-B [05:42:07] [PASSED] TV [05:42:07] [PASSED] eDP [05:42:07] [PASSED] Virtual [05:42:07] [PASSED] DSI [05:42:07] [PASSED] DPI [05:42:07] [PASSED] Writeback [05:42:07] [PASSED] SPI [05:42:07] [PASSED] USB [05:42:07] ==== [PASSED] drm_test_drm_connector_dynamic_init_name ===== [05:42:07] =========== [PASSED] drm_connector_dynamic_init ============ [05:42:07] ==== drm_connector_dynamic_register_early (4 subtests) ===== [05:42:07] [PASSED] drm_test_drm_connector_dynamic_register_early_on_list [05:42:07] [PASSED] drm_test_drm_connector_dynamic_register_early_defer [05:42:07] [PASSED] drm_test_drm_connector_dynamic_register_early_no_init [05:42:07] [PASSED] drm_test_drm_connector_dynamic_register_early_no_mode_object [05:42:07] ====== [PASSED] drm_connector_dynamic_register_early ======= [05:42:07] ======= drm_connector_dynamic_register (7 subtests) ======== [05:42:07] [PASSED] drm_test_drm_connector_dynamic_register_on_list [05:42:07] [PASSED] drm_test_drm_connector_dynamic_register_no_defer [05:42:07] [PASSED] drm_test_drm_connector_dynamic_register_no_init [05:42:07] [PASSED] drm_test_drm_connector_dynamic_register_mode_object [05:42:07] [PASSED] drm_test_drm_connector_dynamic_register_sysfs [05:42:07] [PASSED] drm_test_drm_connector_dynamic_register_sysfs_name [05:42:07] [PASSED] drm_test_drm_connector_dynamic_register_debugfs [05:42:07] ========= [PASSED] drm_connector_dynamic_register ========== [05:42:07] = drm_connector_attach_broadcast_rgb_property (2 subtests) = [05:42:07] [PASSED] drm_test_drm_connector_attach_broadcast_rgb_property [05:42:07] [PASSED] drm_test_drm_connector_attach_broadcast_rgb_property_hdmi_connector [05:42:07] === [PASSED] drm_connector_attach_broadcast_rgb_property === [05:42:07] ========== drm_get_tv_mode_from_name (2 subtests) ========== [05:42:07] ========== drm_test_get_tv_mode_from_name_valid =========== [05:42:07] [PASSED] NTSC [05:42:07] [PASSED] NTSC-443 [05:42:07] [PASSED] NTSC-J [05:42:07] [PASSED] PAL [05:42:07] [PASSED] PAL-M [05:42:07] [PASSED] PAL-N [05:42:07] [PASSED] SECAM [05:42:07] [PASSED] Mono [05:42:07] ====== [PASSED] drm_test_get_tv_mode_from_name_valid ======= [05:42:07] [PASSED] drm_test_get_tv_mode_from_name_truncated [05:42:07] ============ [PASSED] drm_get_tv_mode_from_name ============ [05:42:07] = drm_test_connector_hdmi_compute_mode_clock (12 subtests) = [05:42:07] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb [05:42:07] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_10bpc [05:42:07] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_10bpc_vic_1 [05:42:07] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_12bpc [05:42:07] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_12bpc_vic_1 [05:42:07] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_double [05:42:07] = drm_test_connector_hdmi_compute_mode_clock_yuv420_valid = [05:42:07] [PASSED] VIC 96 [05:42:07] [PASSED] VIC 97 [05:42:07] [PASSED] VIC 101 [05:42:07] [PASSED] VIC 102 [05:42:07] [PASSED] VIC 106 [05:42:07] [PASSED] VIC 107 [05:42:07] === [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv420_valid === [05:42:07] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv420_10_bpc [05:42:07] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv420_12_bpc [05:42:07] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv422_8_bpc [05:42:07] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv422_10_bpc [05:42:07] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv422_12_bpc [05:42:07] === [PASSED] drm_test_connector_hdmi_compute_mode_clock ==== [05:42:07] == drm_hdmi_connector_get_broadcast_rgb_name (2 subtests) == [05:42:07] === drm_test_drm_hdmi_connector_get_broadcast_rgb_name ==== [05:42:07] [PASSED] Automatic [05:42:07] [PASSED] Full [05:42:07] [PASSED] Limited 16:235 [05:42:07] === [PASSED] drm_test_drm_hdmi_connector_get_broadcast_rgb_name === [05:42:07] [PASSED] drm_test_drm_hdmi_connector_get_broadcast_rgb_name_invalid [05:42:07] ==== [PASSED] drm_hdmi_connector_get_broadcast_rgb_name ==== [05:42:07] == drm_hdmi_connector_get_output_format_name (2 subtests) == [05:42:07] === drm_test_drm_hdmi_connector_get_output_format_name ==== [05:42:07] [PASSED] RGB [05:42:07] [PASSED] YUV 4:2:0 [05:42:07] [PASSED] YUV 4:2:2 [05:42:07] [PASSED] YUV 4:4:4 [05:42:07] === [PASSED] drm_test_drm_hdmi_connector_get_output_format_name === [05:42:07] [PASSED] drm_test_drm_hdmi_connector_get_output_format_name_invalid [05:42:07] ==== [PASSED] drm_hdmi_connector_get_output_format_name ==== [05:42:07] ============= drm_damage_helper (21 subtests) ============== [05:42:07] [PASSED] drm_test_damage_iter_no_damage [05:42:07] [PASSED] drm_test_damage_iter_no_damage_fractional_src [05:42:07] [PASSED] drm_test_damage_iter_no_damage_src_moved [05:42:07] [PASSED] drm_test_damage_iter_no_damage_fractional_src_moved [05:42:07] [PASSED] drm_test_damage_iter_no_damage_not_visible [05:42:07] [PASSED] drm_test_damage_iter_no_damage_no_crtc [05:42:07] [PASSED] drm_test_damage_iter_no_damage_no_fb [05:42:07] [PASSED] drm_test_damage_iter_simple_damage [05:42:07] [PASSED] drm_test_damage_iter_single_damage [05:42:07] [PASSED] drm_test_damage_iter_single_damage_intersect_src [05:42:07] [PASSED] drm_test_damage_iter_single_damage_outside_src [05:42:07] [PASSED] drm_test_damage_iter_single_damage_fractional_src [05:42:07] [PASSED] drm_test_damage_iter_single_damage_intersect_fractional_src [05:42:07] [PASSED] drm_test_damage_iter_single_damage_outside_fractional_src [05:42:07] [PASSED] drm_test_damage_iter_single_damage_src_moved [05:42:07] [PASSED] drm_test_damage_iter_single_damage_fractional_src_moved [05:42:07] [PASSED] drm_test_damage_iter_damage [05:42:07] [PASSED] drm_test_damage_iter_damage_one_intersect [05:42:07] [PASSED] drm_test_damage_iter_damage_one_outside [05:42:07] [PASSED] drm_test_damage_iter_damage_src_moved [05:42:07] [PASSED] drm_test_damage_iter_damage_not_visible [05:42:07] ================ [PASSED] drm_damage_helper ================ [05:42:07] ============== drm_dp_mst_helper (3 subtests) ============== [05:42:07] ============== drm_test_dp_mst_calc_pbn_mode ============== [05:42:07] [PASSED] Clock 154000 BPP 30 DSC disabled [05:42:07] [PASSED] Clock 234000 BPP 30 DSC disabled [05:42:07] [PASSED] Clock 297000 BPP 24 DSC disabled [05:42:07] [PASSED] Clock 332880 BPP 24 DSC enabled [05:42:07] [PASSED] Clock 324540 BPP 24 DSC enabled [05:42:07] ========== [PASSED] drm_test_dp_mst_calc_pbn_mode ========== [05:42:07] ============== drm_test_dp_mst_calc_pbn_div =============== [05:42:07] [PASSED] Link rate 2000000 lane count 4 [05:42:07] [PASSED] Link rate 2000000 lane count 2 [05:42:07] [PASSED] Link rate 2000000 lane count 1 [05:42:07] [PASSED] Link rate 1350000 lane count 4 [05:42:07] [PASSED] Link rate 1350000 lane count 2 [05:42:07] [PASSED] Link rate 1350000 lane count 1 [05:42:07] [PASSED] Link rate 1000000 lane count 4 [05:42:07] [PASSED] Link rate 1000000 lane count 2 [05:42:07] [PASSED] Link rate 1000000 lane count 1 [05:42:07] [PASSED] Link rate 810000 lane count 4 [05:42:07] [PASSED] Link rate 810000 lane count 2 [05:42:07] [PASSED] Link rate 810000 lane count 1 [05:42:07] [PASSED] Link rate 540000 lane count 4 [05:42:07] [PASSED] Link rate 540000 lane count 2 [05:42:07] [PASSED] Link rate 540000 lane count 1 [05:42:07] [PASSED] Link rate 270000 lane count 4 [05:42:07] [PASSED] Link rate 270000 lane count 2 [05:42:07] [PASSED] Link rate 270000 lane count 1 [05:42:07] [PASSED] Link rate 162000 lane count 4 [05:42:07] [PASSED] Link rate 162000 lane count 2 [05:42:07] [PASSED] Link rate 162000 lane count 1 [05:42:07] ========== [PASSED] drm_test_dp_mst_calc_pbn_div =========== [05:42:07] ========= drm_test_dp_mst_sideband_msg_req_decode ========= [05:42:07] [PASSED] DP_ENUM_PATH_RESOURCES with port number [05:42:07] [PASSED] DP_POWER_UP_PHY with port number [05:42:07] [PASSED] DP_POWER_DOWN_PHY with port number [05:42:07] [PASSED] DP_ALLOCATE_PAYLOAD with SDP stream sinks [05:42:07] [PASSED] DP_ALLOCATE_PAYLOAD with port number [05:42:07] [PASSED] DP_ALLOCATE_PAYLOAD with VCPI [05:42:07] [PASSED] DP_ALLOCATE_PAYLOAD with PBN [05:42:07] [PASSED] DP_QUERY_PAYLOAD with port number [05:42:07] [PASSED] DP_QUERY_PAYLOAD with VCPI [05:42:07] [PASSED] DP_REMOTE_DPCD_READ with port number [05:42:07] [PASSED] DP_REMOTE_DPCD_READ with DPCD address [05:42:07] [PASSED] DP_REMOTE_DPCD_READ with max number of bytes [05:42:07] [PASSED] DP_REMOTE_DPCD_WRITE with port number [05:42:07] [PASSED] DP_REMOTE_DPCD_WRITE with DPCD address [05:42:07] [PASSED] DP_REMOTE_DPCD_WRITE with data array [05:42:07] [PASSED] DP_REMOTE_I2C_READ with port number [05:42:07] [PASSED] DP_REMOTE_I2C_READ with I2C device ID [05:42:07] [PASSED] DP_REMOTE_I2C_READ with transactions array [05:42:07] [PASSED] DP_REMOTE_I2C_WRITE with port number [05:42:07] [PASSED] DP_REMOTE_I2C_WRITE with I2C device ID [05:42:07] [PASSED] DP_REMOTE_I2C_WRITE with data array [05:42:07] [PASSED] DP_QUERY_STREAM_ENC_STATUS with stream ID [05:42:07] [PASSED] DP_QUERY_STREAM_ENC_STATUS with client ID [05:42:07] [PASSED] DP_QUERY_STREAM_ENC_STATUS with stream event [05:42:07] [PASSED] DP_QUERY_STREAM_ENC_STATUS with valid stream event [05:42:07] [PASSED] DP_QUERY_STREAM_ENC_STATUS with stream behavior [05:42:07] [PASSED] DP_QUERY_STREAM_ENC_STATUS with a valid stream behavior [05:42:07] ===== [PASSED] drm_test_dp_mst_sideband_msg_req_decode ===== [05:42:07] ================ [PASSED] drm_dp_mst_helper ================ [05:42:07] ================== drm_exec (7 subtests) =================== [05:42:07] [PASSED] sanitycheck [05:42:07] [PASSED] test_lock [05:42:07] [PASSED] test_lock_unlock [05:42:07] [PASSED] test_duplicates [05:42:07] [PASSED] test_prepare [05:42:07] [PASSED] test_prepare_array [05:42:07] [PASSED] test_multiple_loops [05:42:07] ==================== [PASSED] drm_exec ===================== [05:42:07] =========== drm_format_helper_test (17 subtests) =========== [05:42:07] ============== drm_test_fb_xrgb8888_to_gray8 ============== [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ========== [PASSED] drm_test_fb_xrgb8888_to_gray8 ========== [05:42:07] ============= drm_test_fb_xrgb8888_to_rgb332 ============== [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ========= [PASSED] drm_test_fb_xrgb8888_to_rgb332 ========== [05:42:07] ============= drm_test_fb_xrgb8888_to_rgb565 ============== [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ========= [PASSED] drm_test_fb_xrgb8888_to_rgb565 ========== [05:42:07] ============ drm_test_fb_xrgb8888_to_xrgb1555 ============= [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ======== [PASSED] drm_test_fb_xrgb8888_to_xrgb1555 ========= [05:42:07] ============ drm_test_fb_xrgb8888_to_argb1555 ============= [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ======== [PASSED] drm_test_fb_xrgb8888_to_argb1555 ========= [05:42:07] ============ drm_test_fb_xrgb8888_to_rgba5551 ============= [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ======== [PASSED] drm_test_fb_xrgb8888_to_rgba5551 ========= [05:42:07] ============= drm_test_fb_xrgb8888_to_rgb888 ============== [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ========= [PASSED] drm_test_fb_xrgb8888_to_rgb888 ========== [05:42:07] ============= drm_test_fb_xrgb8888_to_bgr888 ============== [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ========= [PASSED] drm_test_fb_xrgb8888_to_bgr888 ========== [05:42:07] ============ drm_test_fb_xrgb8888_to_argb8888 ============= [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ======== [PASSED] drm_test_fb_xrgb8888_to_argb8888 ========= [05:42:07] =========== drm_test_fb_xrgb8888_to_xrgb2101010 =========== [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ======= [PASSED] drm_test_fb_xrgb8888_to_xrgb2101010 ======= [05:42:07] =========== drm_test_fb_xrgb8888_to_argb2101010 =========== [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ======= [PASSED] drm_test_fb_xrgb8888_to_argb2101010 ======= [05:42:07] ============== drm_test_fb_xrgb8888_to_mono =============== [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ========== [PASSED] drm_test_fb_xrgb8888_to_mono =========== [05:42:07] ==================== drm_test_fb_swab ===================== [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ================ [PASSED] drm_test_fb_swab ================= [05:42:07] ============ drm_test_fb_xrgb8888_to_xbgr8888 ============= [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ======== [PASSED] drm_test_fb_xrgb8888_to_xbgr8888 ========= [05:42:07] ============ drm_test_fb_xrgb8888_to_abgr8888 ============= [05:42:07] [PASSED] single_pixel_source_buffer [05:42:07] [PASSED] single_pixel_clip_rectangle [05:42:07] [PASSED] well_known_colors [05:42:07] [PASSED] destination_pitch [05:42:07] ======== [PASSED] drm_test_fb_xrgb8888_to_abgr8888 ========= [05:42:07] ================= drm_test_fb_clip_offset ================= [05:42:07] [PASSED] pass through [05:42:07] [PASSED] horizontal offset [05:42:07] [PASSED] vertical offset [05:42:07] [PASSED] horizontal and vertical offset [05:42:07] [PASSED] horizontal offset (custom pitch) [05:42:07] [PASSED] vertical offset (custom pitch) [05:42:07] [PASSED] horizontal and vertical offset (custom pitch) [05:42:07] ============= [PASSED] drm_test_fb_clip_offset ============= [05:42:07] =================== drm_test_fb_memcpy ==================== [05:42:07] [PASSED] single_pixel_source_buffer: XR24 little-endian (0x34325258) [05:42:07] [PASSED] single_pixel_source_buffer: XRA8 little-endian (0x38415258) [05:42:07] [PASSED] single_pixel_source_buffer: YU24 little-endian (0x34325559) [05:42:07] [PASSED] single_pixel_clip_rectangle: XB24 little-endian (0x34324258) [05:42:07] [PASSED] single_pixel_clip_rectangle: XRA8 little-endian (0x38415258) [05:42:07] [PASSED] single_pixel_clip_rectangle: YU24 little-endian (0x34325559) [05:42:07] [PASSED] well_known_colors: XB24 little-endian (0x34324258) [05:42:07] [PASSED] well_known_colors: XRA8 little-endian (0x38415258) [05:42:07] [PASSED] well_known_colors: YU24 little-endian (0x34325559) [05:42:07] [PASSED] destination_pitch: XB24 little-endian (0x34324258) [05:42:07] [PASSED] destination_pitch: XRA8 little-endian (0x38415258) [05:42:07] [PASSED] destination_pitch: YU24 little-endian (0x34325559) [05:42:07] =============== [PASSED] drm_test_fb_memcpy ================ [05:42:07] ============= [PASSED] drm_format_helper_test ============== [05:42:07] ================= drm_format (18 subtests) ================= [05:42:07] [PASSED] drm_test_format_block_width_invalid [05:42:07] [PASSED] drm_test_format_block_width_one_plane [05:42:07] [PASSED] drm_test_format_block_width_two_plane [05:42:07] [PASSED] drm_test_format_block_width_three_plane [05:42:07] [PASSED] drm_test_format_block_width_tiled [05:42:07] [PASSED] drm_test_format_block_height_invalid [05:42:07] [PASSED] drm_test_format_block_height_one_plane [05:42:07] [PASSED] drm_test_format_block_height_two_plane [05:42:07] [PASSED] drm_test_format_block_height_three_plane [05:42:07] [PASSED] drm_test_format_block_height_tiled [05:42:07] [PASSED] drm_test_format_min_pitch_invalid [05:42:07] [PASSED] drm_test_format_min_pitch_one_plane_8bpp [05:42:07] [PASSED] drm_test_format_min_pitch_one_plane_16bpp [05:42:07] [PASSED] drm_test_format_min_pitch_one_plane_24bpp [05:42:07] [PASSED] drm_test_format_min_pitch_one_plane_32bpp [05:42:07] [PASSED] drm_test_format_min_pitch_two_plane [05:42:07] [PASSED] drm_test_format_min_pitch_three_plane_8bpp [05:42:07] [PASSED] drm_test_format_min_pitch_tiled [05:42:07] =================== [PASSED] drm_format ==================== [05:42:07] ============== drm_framebuffer (10 subtests) =============== [05:42:07] ========== drm_test_framebuffer_check_src_coords ========== [05:42:07] [PASSED] Success: source fits into fb [05:42:07] [PASSED] Fail: overflowing fb with x-axis coordinate [05:42:07] [PASSED] Fail: overflowing fb with y-axis coordinate [05:42:07] [PASSED] Fail: overflowing fb with source width [05:42:07] [PASSED] Fail: overflowing fb with source height [05:42:07] ====== [PASSED] drm_test_framebuffer_check_src_coords ====== [05:42:07] [PASSED] drm_test_framebuffer_cleanup [05:42:07] =============== drm_test_framebuffer_create =============== [05:42:07] [PASSED] ABGR8888 normal sizes [05:42:07] [PASSED] ABGR8888 max sizes [05:42:07] [PASSED] ABGR8888 pitch greater than min required [05:42:07] [PASSED] ABGR8888 pitch less than min required [05:42:07] [PASSED] ABGR8888 Invalid width [05:42:07] [PASSED] ABGR8888 Invalid buffer handle [05:42:07] [PASSED] No pixel format [05:42:07] [PASSED] ABGR8888 Width 0 [05:42:07] [PASSED] ABGR8888 Height 0 [05:42:07] [PASSED] ABGR8888 Out of bound height * pitch combination [05:42:07] [PASSED] ABGR8888 Large buffer offset [05:42:07] [PASSED] ABGR8888 Buffer offset for inexistent plane [05:42:07] [PASSED] ABGR8888 Invalid flag [05:42:07] [PASSED] ABGR8888 Set DRM_MODE_FB_MODIFIERS without modifiers [05:42:07] [PASSED] ABGR8888 Valid buffer modifier [05:42:07] [PASSED] ABGR8888 Invalid buffer modifier(DRM_FORMAT_MOD_SAMSUNG_64_32_TILE) [05:42:07] [PASSED] ABGR8888 Extra pitches without DRM_MODE_FB_MODIFIERS [05:42:07] [PASSED] ABGR8888 Extra pitches with DRM_MODE_FB_MODIFIERS [05:42:07] [PASSED] NV12 Normal sizes [05:42:07] [PASSED] NV12 Max sizes [05:42:07] [PASSED] NV12 Invalid pitch [05:42:07] [PASSED] NV12 Invalid modifier/missing DRM_MODE_FB_MODIFIERS flag [05:42:07] [PASSED] NV12 different modifier per-plane [05:42:07] [PASSED] NV12 with DRM_FORMAT_MOD_SAMSUNG_64_32_TILE [05:42:07] [PASSED] NV12 Valid modifiers without DRM_MODE_FB_MODIFIERS [05:42:07] [PASSED] NV12 Modifier for inexistent plane [05:42:07] [PASSED] NV12 Handle for inexistent plane [05:42:07] [PASSED] NV12 Handle for inexistent plane without DRM_MODE_FB_MODIFIERS [05:42:07] [PASSED] YVU420 DRM_MODE_FB_MODIFIERS set without modifier [05:42:07] [PASSED] YVU420 Normal sizes [05:42:07] [PASSED] YVU420 Max sizes [05:42:07] [PASSED] YVU420 Invalid pitch [05:42:07] [PASSED] YVU420 Different pitches [05:42:07] [PASSED] YVU420 Different buffer offsets/pitches [05:42:07] [PASSED] YVU420 Modifier set just for plane 0, without DRM_MODE_FB_MODIFIERS [05:42:07] [PASSED] YVU420 Modifier set just for planes 0, 1, without DRM_MODE_FB_MODIFIERS [05:42:07] [PASSED] YVU420 Modifier set just for plane 0, 1, with DRM_MODE_FB_MODIFIERS [05:42:07] [PASSED] YVU420 Valid modifier [05:42:07] [PASSED] YVU420 Different modifiers per plane [05:42:07] [PASSED] YVU420 Modifier for inexistent plane [05:42:07] [PASSED] YUV420_10BIT Invalid modifier(DRM_FORMAT_MOD_LINEAR) [05:42:07] [PASSED] X0L2 Normal sizes [05:42:07] [PASSED] X0L2 Max sizes [05:42:07] [PASSED] X0L2 Invalid pitch [05:42:07] [PASSED] X0L2 Pitch greater than minimum required [05:42:07] [PASSED] X0L2 Handle for inexistent plane [05:42:07] [PASSED] X0L2 Offset for inexistent plane, without DRM_MODE_FB_MODIFIERS set [05:42:07] [PASSED] X0L2 Modifier without DRM_MODE_FB_MODIFIERS set [05:42:07] [PASSED] X0L2 Valid modifier [05:42:07] [PASSED] X0L2 Modifier for inexistent plane [05:42:07] =========== [PASSED] drm_test_framebuffer_create =========== [05:42:07] [PASSED] drm_test_framebuffer_free [05:42:07] [PASSED] drm_test_framebuffer_init [05:42:07] [PASSED] drm_test_framebuffer_init_bad_format [05:42:07] [PASSED] drm_test_framebuffer_init_dev_mismatch [05:42:07] [PASSED] drm_test_framebuffer_lookup [05:42:07] [PASSED] drm_test_framebuffer_lookup_inexistent [05:42:07] [PASSED] drm_test_framebuffer_modifiers_not_supported [05:42:07] ================= [PASSED] drm_framebuffer ================= [05:42:07] ================ drm_gem_shmem (8 subtests) ================ [05:42:07] [PASSED] drm_gem_shmem_test_obj_create [05:42:07] [PASSED] drm_gem_shmem_test_obj_create_private [05:42:07] [PASSED] drm_gem_shmem_test_pin_pages [05:42:07] [PASSED] drm_gem_shmem_test_vmap [05:42:07] [PASSED] drm_gem_shmem_test_get_sg_table [05:42:07] [PASSED] drm_gem_shmem_test_get_pages_sgt [05:42:07] [PASSED] drm_gem_shmem_test_madvise [05:42:07] [PASSED] drm_gem_shmem_test_purge [05:42:07] ================== [PASSED] drm_gem_shmem ================== [05:42:07] === drm_atomic_helper_connector_hdmi_check (29 subtests) === [05:42:07] [PASSED] drm_test_check_broadcast_rgb_auto_cea_mode [05:42:07] [PASSED] drm_test_check_broadcast_rgb_auto_cea_mode_vic_1 [05:42:07] [PASSED] drm_test_check_broadcast_rgb_full_cea_mode [05:42:07] [PASSED] drm_test_check_broadcast_rgb_full_cea_mode_vic_1 [05:42:07] [PASSED] drm_test_check_broadcast_rgb_limited_cea_mode [05:42:07] [PASSED] drm_test_check_broadcast_rgb_limited_cea_mode_vic_1 [05:42:07] ====== drm_test_check_broadcast_rgb_cea_mode_yuv420 ======= [05:42:07] [PASSED] Automatic [05:42:07] [PASSED] Full [05:42:07] [PASSED] Limited 16:235 [05:42:07] == [PASSED] drm_test_check_broadcast_rgb_cea_mode_yuv420 === [05:42:07] [PASSED] drm_test_check_broadcast_rgb_crtc_mode_changed [05:42:07] [PASSED] drm_test_check_broadcast_rgb_crtc_mode_not_changed [05:42:07] [PASSED] drm_test_check_disable_connector [05:42:07] [PASSED] drm_test_check_hdmi_funcs_reject_rate [05:42:07] [PASSED] drm_test_check_max_tmds_rate_bpc_fallback_rgb [05:42:07] [PASSED] drm_test_check_max_tmds_rate_bpc_fallback_yuv420 [05:42:07] [PASSED] drm_test_check_max_tmds_rate_bpc_fallback_ignore_yuv422 [05:42:07] [PASSED] drm_test_check_max_tmds_rate_bpc_fallback_ignore_yuv420 [05:42:07] [PASSED] drm_test_check_driver_unsupported_fallback_yuv420 [05:42:07] [PASSED] drm_test_check_output_bpc_crtc_mode_changed [05:42:07] [PASSED] drm_test_check_output_bpc_crtc_mode_not_changed [05:42:07] [PASSED] drm_test_check_output_bpc_dvi [05:42:07] [PASSED] drm_test_check_output_bpc_format_vic_1 [05:42:07] [PASSED] drm_test_check_output_bpc_format_display_8bpc_only [05:42:07] [PASSED] drm_test_check_output_bpc_format_display_rgb_only [05:42:07] [PASSED] drm_test_check_output_bpc_format_driver_8bpc_only [05:42:07] [PASSED] drm_test_check_output_bpc_format_driver_rgb_only [05:42:07] [PASSED] drm_test_check_tmds_char_rate_rgb_8bpc [05:42:07] [PASSED] drm_test_check_tmds_char_rate_rgb_10bpc [05:42:07] [PASSED] drm_test_check_tmds_char_rate_rgb_12bpc [05:42:07] ============ drm_test_check_hdmi_color_format ============= [05:42:07] [PASSED] AUTO -> RGB [05:42:07] [PASSED] YCBCR422 -> YUV422 [05:42:07] [PASSED] YCBCR420 -> YUV420 [05:42:07] [PASSED] YCBCR444 -> YUV444 [05:42:07] [PASSED] RGB -> RGB [05:42:07] ======== [PASSED] drm_test_check_hdmi_color_format ========= [05:42:07] ======== drm_test_check_hdmi_color_format_420_only ======== [05:42:07] [PASSED] RGB should fail [05:42:07] [PASSED] YUV444 should fail [05:42:07] [PASSED] YUV422 should fail [05:42:07] [PASSED] YUV420 should work [05:42:07] ==== [PASSED] drm_test_check_hdmi_color_format_420_only ==== [05:42:07] ===== [PASSED] drm_atomic_helper_connector_hdmi_check ====== [05:42:07] === drm_atomic_helper_connector_hdmi_reset (6 subtests) ==== [05:42:07] [PASSED] drm_test_check_broadcast_rgb_value [05:42:07] [PASSED] drm_test_check_bpc_8_value [05:42:07] [PASSED] drm_test_check_bpc_10_value [05:42:07] [PASSED] drm_test_check_bpc_12_value [05:42:07] [PASSED] drm_test_check_format_value [05:42:07] [PASSED] drm_test_check_tmds_char_value [05:42:07] ===== [PASSED] drm_atomic_helper_connector_hdmi_reset ====== [05:42:07] = drm_atomic_helper_connector_hdmi_mode_valid (7 subtests) = [05:42:07] [PASSED] drm_test_check_mode_valid [05:42:07] [PASSED] drm_test_check_mode_valid_reject [05:42:07] [PASSED] drm_test_check_mode_valid_reject_rate [05:42:07] [PASSED] drm_test_check_mode_valid_reject_max_clock [05:42:07] [PASSED] drm_test_check_mode_valid_yuv420_only_max_clock [05:42:07] [PASSED] drm_test_check_mode_valid_reject_yuv420_only_connector [05:42:07] [PASSED] drm_test_check_mode_valid_accept_yuv420_also_connector_rgb [05:42:07] === [PASSED] drm_atomic_helper_connector_hdmi_mode_valid === [05:42:07] = drm_atomic_helper_connector_hdmi_infoframes (5 subtests) = [05:42:07] [PASSED] drm_test_check_infoframes [05:42:07] [PASSED] drm_test_check_reject_avi_infoframe [05:42:07] [PASSED] drm_test_check_reject_hdr_infoframe_bpc_8 [05:42:07] [PASSED] drm_test_check_reject_hdr_infoframe_bpc_10 [05:42:07] [PASSED] drm_test_check_reject_audio_infoframe [05:42:07] === [PASSED] drm_atomic_helper_connector_hdmi_infoframes === [05:42:07] ================= drm_managed (2 subtests) ================= [05:42:07] [PASSED] drm_test_managed_release_action [05:42:07] [PASSED] drm_test_managed_run_action [05:42:07] =================== [PASSED] drm_managed =================== [05:42:07] =================== drm_mm (6 subtests) ==================== [05:42:07] [PASSED] drm_test_mm_init [05:42:07] [PASSED] drm_test_mm_debug [05:42:07] [PASSED] drm_test_mm_align32 [05:42:07] [PASSED] drm_test_mm_align64 [05:42:07] [PASSED] drm_test_mm_lowest [05:42:07] [PASSED] drm_test_mm_highest [05:42:07] ===================== [PASSED] drm_mm ====================== [05:42:07] ============= drm_modes_analog_tv (5 subtests) ============= [05:42:07] [PASSED] drm_test_modes_analog_tv_mono_576i [05:42:07] [PASSED] drm_test_modes_analog_tv_ntsc_480i [05:42:07] [PASSED] drm_test_modes_analog_tv_ntsc_480i_inlined [05:42:07] [PASSED] drm_test_modes_analog_tv_pal_576i [05:42:07] [PASSED] drm_test_modes_analog_tv_pal_576i_inlined [05:42:07] =============== [PASSED] drm_modes_analog_tv =============== [05:42:07] ============== drm_plane_helper (2 subtests) =============== [05:42:07] =============== drm_test_check_plane_state ================ [05:42:07] [PASSED] clipping_simple [05:42:07] [PASSED] clipping_rotate_reflect [05:42:07] [PASSED] positioning_simple [05:42:07] [PASSED] upscaling [05:42:07] [PASSED] downscaling [05:42:07] [PASSED] rounding1 [05:42:07] [PASSED] rounding2 [05:42:07] [PASSED] rounding3 [05:42:07] [PASSED] rounding4 [05:42:07] =========== [PASSED] drm_test_check_plane_state ============ [05:42:07] =========== drm_test_check_invalid_plane_state ============ [05:42:07] [PASSED] positioning_invalid [05:42:07] [PASSED] upscaling_invalid [05:42:07] [PASSED] downscaling_invalid [05:42:07] ======= [PASSED] drm_test_check_invalid_plane_state ======== [05:42:07] ================ [PASSED] drm_plane_helper ================= [05:42:07] ====== drm_connector_helper_tv_get_modes (1 subtest) ======= [05:42:07] ====== drm_test_connector_helper_tv_get_modes_check ======= [05:42:07] [PASSED] None [05:42:07] [PASSED] PAL [05:42:07] [PASSED] NTSC [05:42:07] [PASSED] Both, NTSC Default [05:42:07] [PASSED] Both, PAL Default [05:42:07] [PASSED] Both, NTSC Default, with PAL on command-line [05:42:07] [PASSED] Both, PAL Default, with NTSC on command-line [05:42:07] == [PASSED] drm_test_connector_helper_tv_get_modes_check === [05:42:07] ======== [PASSED] drm_connector_helper_tv_get_modes ======== [05:42:07] ================== drm_rect (9 subtests) =================== [05:42:07] [PASSED] drm_test_rect_clip_scaled_div_by_zero [05:42:07] [PASSED] drm_test_rect_clip_scaled_not_clipped [05:42:07] [PASSED] drm_test_rect_clip_scaled_clipped [05:42:07] [PASSED] drm_test_rect_clip_scaled_signed_vs_unsigned [05:42:07] ================= drm_test_rect_intersect ================= [05:42:07] [PASSED] top-left x bottom-right: 2x2+1+1 x 2x2+0+0 [05:42:07] [PASSED] top-right x bottom-left: 2x2+0+0 x 2x2+1-1 [05:42:07] [PASSED] bottom-left x top-right: 2x2+1-1 x 2x2+0+0 [05:42:07] [PASSED] bottom-right x top-left: 2x2+0+0 x 2x2+1+1 [05:42:07] [PASSED] right x left: 2x1+0+0 x 3x1+1+0 [05:42:07] [PASSED] left x right: 3x1+1+0 x 2x1+0+0 [05:42:07] [PASSED] up x bottom: 1x2+0+0 x 1x3+0-1 [05:42:07] [PASSED] bottom x up: 1x3+0-1 x 1x2+0+0 [05:42:07] [PASSED] touching corner: 1x1+0+0 x 2x2+1+1 [05:42:07] [PASSED] touching side: 1x1+0+0 x 1x1+1+0 [05:42:07] [PASSED] equal rects: 2x2+0+0 x 2x2+0+0 [05:42:07] [PASSED] inside another: 2x2+0+0 x 1x1+1+1 [05:42:07] [PASSED] far away: 1x1+0+0 x 1x1+3+6 [05:42:07] [PASSED] points intersecting: 0x0+5+10 x 0x0+5+10 [05:42:07] [PASSED] points not intersecting: 0x0+0+0 x 0x0+5+10 [05:42:07] ============= [PASSED] drm_test_rect_intersect ============= [05:42:07] ================ drm_test_rect_calc_hscale ================ [05:42:07] [PASSED] normal use [05:42:07] [PASSED] out of max range [05:42:07] [PASSED] out of min range [05:42:07] [PASSED] zero dst [05:42:07] [PASSED] negative src [05:42:07] [PASSED] negative dst [05:42:07] ============ [PASSED] drm_test_rect_calc_hscale ============ [05:42:07] ================ drm_test_rect_calc_vscale ================ [05:42:07] [PASSED] normal use [05:42:07] [PASSED] out of max range [05:42:07] [PASSED] out of min range [05:42:07] [PASSED] zero dst [05:42:07] [PASSED] negative src [05:42:07] [PASSED] negative dst [05:42:07] ============ [PASSED] drm_test_rect_calc_vscale ============ [05:42:07] ================== drm_test_rect_rotate =================== [05:42:07] [PASSED] reflect-x [05:42:07] [PASSED] reflect-y [05:42:07] [PASSED] rotate-0 [05:42:07] [PASSED] rotate-90 [05:42:07] [PASSED] rotate-180 [05:42:07] [PASSED] rotate-270 [05:42:07] ============== [PASSED] drm_test_rect_rotate =============== [05:42:07] ================ drm_test_rect_rotate_inv ================= [05:42:07] [PASSED] reflect-x [05:42:07] [PASSED] reflect-y [05:42:07] [PASSED] rotate-0 [05:42:07] [PASSED] rotate-90 [05:42:07] [PASSED] rotate-180 [05:42:07] [PASSED] rotate-270 [05:42:07] ============ [PASSED] drm_test_rect_rotate_inv ============= [05:42:07] ==================== [PASSED] drm_rect ===================== [05:42:07] ============ drm_sysfb_modeset_test (1 subtest) ============ [05:42:07] ============ drm_test_sysfb_build_fourcc_list ============= [05:42:07] [PASSED] no native formats [05:42:07] [PASSED] XRGB8888 as native format [05:42:07] [PASSED] remove duplicates [05:42:07] [PASSED] convert alpha formats [05:42:07] [PASSED] random formats [05:42:07] ======== [PASSED] drm_test_sysfb_build_fourcc_list ========= [05:42:07] ============= [PASSED] drm_sysfb_modeset_test ============== [05:42:07] ================== drm_fixp (2 subtests) =================== [05:42:07] [PASSED] drm_test_int2fixp [05:42:07] [PASSED] drm_test_sm2fixp [05:42:07] ==================== [PASSED] drm_fixp ===================== [05:42:07] ============================================================ [05:42:07] Testing complete. Ran 637 tests: passed: 637 [05:42:07] Elapsed time: 26.793s total, 1.800s configuring, 24.823s building, 0.150s running
+ /kernel/tools/testing/kunit/kunit.py run --kunitconfig /kernel/drivers/gpu/drm/ttm/tests/.kunitconfig [05:42:07] Configuring KUnit Kernel ...
Regenerating .config ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig [05:42:09] Building KUnit Kernel ... Populating config with: $ make ARCH=um O=.kunit olddefconfig Building with: $ make all compile_commands.json scripts_gdb ARCH=um O=.kunit --jobs=48 [05:42:18] Starting KUnit Kernel (1/1)... [05:42:18] ============================================================ Running tests with: $ .kunit/linux kunit.enable=1 mem=1G console=tty kunit_shutdown=halt [05:42:19] ================= ttm_device (5 subtests) ================== [05:42:19] [PASSED] ttm_device_init_basic [05:42:19] [PASSED] ttm_device_init_multiple [05:42:19] [PASSED] ttm_device_fini_basic [05:42:19] [PASSED] ttm_device_init_no_vma_man [05:42:19] ================== ttm_device_init_pools ================== [05:42:19] [PASSED] No DMA allocations, no DMA32 required [05:42:19] [PASSED] DMA allocations, DMA32 required [05:42:19] [PASSED] No DMA allocations, DMA32 required [05:42:19] [PASSED] DMA allocations, no DMA32 required [05:42:19] ============== [PASSED] ttm_device_init_pools ============== [05:42:19] =================== [PASSED] ttm_device ==================== [05:42:19] ================== ttm_pool (8 subtests) =================== [05:42:19] ================== ttm_pool_alloc_basic =================== [05:42:19] [PASSED] One page [05:42:19] [PASSED] More than one page [05:42:19] [PASSED] Above the allocation limit [05:42:19] [PASSED] One page, with coherent DMA mappings enabled [05:42:19] [PASSED] Above the allocation limit, with coherent DMA mappings enabled [05:42:19] ============== [PASSED] ttm_pool_alloc_basic =============== [05:42:19] ============== ttm_pool_alloc_basic_dma_addr ============== [05:42:19] [PASSED] One page [05:42:19] [PASSED] More than one page [05:42:19] [PASSED] Above the allocation limit [05:42:19] [PASSED] One page, with coherent DMA mappings enabled [05:42:19] [PASSED] Above the allocation limit, with coherent DMA mappings enabled [05:42:19] ========== [PASSED] ttm_pool_alloc_basic_dma_addr ========== [05:42:19] [PASSED] ttm_pool_alloc_order_caching_match [05:42:19] [PASSED] ttm_pool_alloc_caching_mismatch [05:42:19] [PASSED] ttm_pool_alloc_order_mismatch [05:42:19] [PASSED] ttm_pool_free_dma_alloc [05:42:19] [PASSED] ttm_pool_free_no_dma_alloc [05:42:19] [PASSED] ttm_pool_fini_basic [05:42:19] ==================== [PASSED] ttm_pool ===================== [05:42:19] ================ ttm_resource (8 subtests) ================= [05:42:19] ================= ttm_resource_init_basic ================= [05:42:19] [PASSED] Init resource in TTM_PL_SYSTEM [05:42:19] [PASSED] Init resource in TTM_PL_VRAM [05:42:19] [PASSED] Init resource in a private placement [05:42:19] [PASSED] Init resource in TTM_PL_SYSTEM, set placement flags [05:42:19] ============= [PASSED] ttm_resource_init_basic ============= [05:42:19] [PASSED] ttm_resource_init_pinned [05:42:19] [PASSED] ttm_resource_fini_basic [05:42:19] [PASSED] ttm_resource_manager_init_basic [05:42:19] [PASSED] ttm_resource_manager_usage_basic [05:42:19] [PASSED] ttm_resource_manager_set_used_basic [05:42:19] [PASSED] ttm_sys_man_alloc_basic [05:42:19] [PASSED] ttm_sys_man_free_basic [05:42:19] ================== [PASSED] ttm_resource =================== [05:42:19] =================== ttm_tt (15 subtests) =================== [05:42:19] ==================== ttm_tt_init_basic ==================== [05:42:19] [PASSED] Page-aligned size [05:42:19] [PASSED] Extra pages requested [05:42:19] ================ [PASSED] ttm_tt_init_basic ================ [05:42:19] [PASSED] ttm_tt_init_misaligned [05:42:19] [PASSED] ttm_tt_fini_basic [05:42:19] [PASSED] ttm_tt_fini_sg [05:42:19] [PASSED] ttm_tt_fini_shmem [05:42:19] [PASSED] ttm_tt_create_basic [05:42:19] [PASSED] ttm_tt_create_invalid_bo_type [05:42:19] [PASSED] ttm_tt_create_ttm_exists [05:42:19] [PASSED] ttm_tt_create_failed [05:42:19] [PASSED] ttm_tt_destroy_basic [05:42:19] [PASSED] ttm_tt_populate_null_ttm [05:42:19] [PASSED] ttm_tt_populate_populated_ttm [05:42:19] [PASSED] ttm_tt_unpopulate_basic [05:42:19] [PASSED] ttm_tt_unpopulate_empty_ttm [05:42:19] [PASSED] ttm_tt_swapin_basic [05:42:19] ===================== [PASSED] ttm_tt ====================== [05:42:19] =================== ttm_bo (14 subtests) =================== [05:42:19] =========== ttm_bo_reserve_optimistic_no_ticket =========== [05:42:19] [PASSED] Cannot be interrupted and sleeps [05:42:19] [PASSED] Cannot be interrupted, locks straight away [05:42:19] [PASSED] Can be interrupted, sleeps [05:42:19] ======= [PASSED] ttm_bo_reserve_optimistic_no_ticket ======= [05:42:19] [PASSED] ttm_bo_reserve_locked_no_sleep [05:42:19] [PASSED] ttm_bo_reserve_no_wait_ticket [05:42:19] [PASSED] ttm_bo_reserve_double_resv [05:42:19] [PASSED] ttm_bo_reserve_interrupted [05:42:19] [PASSED] ttm_bo_reserve_deadlock [05:42:19] [PASSED] ttm_bo_unreserve_basic [05:42:19] [PASSED] ttm_bo_unreserve_pinned [05:42:19] [PASSED] ttm_bo_unreserve_bulk [05:42:19] [PASSED] ttm_bo_fini_basic [05:42:19] [PASSED] ttm_bo_fini_shared_resv [05:42:19] [PASSED] ttm_bo_pin_basic [05:42:19] [PASSED] ttm_bo_pin_unpin_resource [05:42:19] [PASSED] ttm_bo_multiple_pin_one_unpin [05:42:19] ===================== [PASSED] ttm_bo ====================== [05:42:19] ============== ttm_bo_validate (22 subtests) =============== [05:42:19] ============== ttm_bo_init_reserved_sys_man =============== [05:42:19] [PASSED] Buffer object for userspace [05:42:19] [PASSED] Kernel buffer object [05:42:19] [PASSED] Shared buffer object [05:42:19] ========== [PASSED] ttm_bo_init_reserved_sys_man =========== [05:42:19] ============== ttm_bo_init_reserved_mock_man ============== [05:42:19] [PASSED] Buffer object for userspace [05:42:19] [PASSED] Kernel buffer object [05:42:19] [PASSED] Shared buffer object [05:42:19] ========== [PASSED] ttm_bo_init_reserved_mock_man ========== [05:42:19] [PASSED] ttm_bo_init_reserved_resv [05:42:19] ================== ttm_bo_validate_basic ================== [05:42:19] [PASSED] Buffer object for userspace [05:42:19] [PASSED] Kernel buffer object [05:42:19] [PASSED] Shared buffer object [05:42:19] ============== [PASSED] ttm_bo_validate_basic ============== [05:42:19] [PASSED] ttm_bo_validate_invalid_placement [05:42:19] ============= ttm_bo_validate_same_placement ============== [05:42:19] [PASSED] System manager [05:42:19] [PASSED] VRAM manager [05:42:19] ========= [PASSED] ttm_bo_validate_same_placement ========== [05:42:19] [PASSED] ttm_bo_validate_failed_alloc [05:42:19] [PASSED] ttm_bo_validate_pinned [05:42:19] [PASSED] ttm_bo_validate_busy_placement [05:42:19] ================ ttm_bo_validate_multihop ================= [05:42:19] [PASSED] Buffer object for userspace [05:42:19] [PASSED] Kernel buffer object [05:42:19] [PASSED] Shared buffer object [05:42:19] ============ [PASSED] ttm_bo_validate_multihop ============= [05:42:19] ========== ttm_bo_validate_no_placement_signaled ========== [05:42:19] [PASSED] Buffer object in system domain, no page vector [05:42:19] [PASSED] Buffer object in system domain with an existing page vector [05:42:19] ====== [PASSED] ttm_bo_validate_no_placement_signaled ====== [05:42:19] ======== ttm_bo_validate_no_placement_not_signaled ======== [05:42:19] [PASSED] Buffer object for userspace [05:42:19] [PASSED] Kernel buffer object [05:42:19] [PASSED] Shared buffer object [05:42:19] ==== [PASSED] ttm_bo_validate_no_placement_not_signaled ==== [05:42:19] [PASSED] ttm_bo_validate_move_fence_signaled [05:42:19] ========= ttm_bo_validate_move_fence_not_signaled ========= [05:42:19] [PASSED] Waits for GPU [05:42:19] [PASSED] Tries to lock straight away [05:42:19] ===== [PASSED] ttm_bo_validate_move_fence_not_signaled ===== [05:42:19] [PASSED] ttm_bo_validate_swapout [05:42:19] [PASSED] ttm_bo_validate_happy_evict [05:42:19] [PASSED] ttm_bo_validate_all_pinned_evict [05:42:19] [PASSED] ttm_bo_validate_allowed_only_evict [05:42:19] [PASSED] ttm_bo_validate_deleted_evict [05:42:19] [PASSED] ttm_bo_validate_busy_domain_evict [05:42:19] [PASSED] ttm_bo_validate_evict_gutting [05:42:19] [PASSED] ttm_bo_validate_recrusive_evict [05:42:19] ================= [PASSED] ttm_bo_validate ================= [05:42:19] ============================================================ [05:42:19] Testing complete. Ran 102 tests: passed: 102 [05:42:19] Elapsed time: 11.974s total, 1.760s configuring, 9.949s building, 0.225s running
+ cleanup
++ stat -c %u:%g /kernel
+ chown -R 1003:1003 /kernel
Patchwork
CI✓ Xe.CI.BAT: success for drm/xe/guc_ads: allocate UM queues in a separate UC BO
In reply to Jia Yao== Series Details ==
== Series Details ==
Series: drm/xe/guc_ads: allocate UM queues in a separate UC BO
URL : https://patchwork.freedesktop.org/series/170879/
State : success
== Summary ==
CI Bug Log - changes from xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c_BAT -> xe-pw-170879v1_BAT
====================================================
Summary
-------
**SUCCESS**
No regressions found.
Participating hosts (13 -> 13)
------------------------------
No changes in participating hosts
Changes
-------
No changes found
Build changes
-------------
* IGT: IGT_9017 -> IGT_9018
* Linux: xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c -> xe-pw-170879v1
IGT_9017: 9017
IGT_9018: f4345861cbd5d97fd50b73bf7af5fe3286fca056 @ https://gitlab.freedesktop.org/drm/igt-gpu-tools.git
xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c: 5726f9c4bde0290e5e60e636b9632bb03a57498c
xe-pw-170879v1: 170879v1
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/index.html
Patchwork
CI✓ Xe.CI.FULL: success for drm/xe/guc_ads: allocate UM queues in a separate UC BO
In reply to Jia Yao== Series Details ==
== Series Details ==
Series: drm/xe/guc_ads: allocate UM queues in a separate UC BO
URL : https://patchwork.freedesktop.org/series/170879/
State : success
== Summary ==
CI Bug Log - changes from xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c_FULL -> xe-pw-170879v1_FULL
====================================================
Summary
-------
**SUCCESS**
No regressions found.
Participating hosts (2 -> 2)
------------------------------
No changes in participating hosts
Known issues
------------
Here are the changes found in xe-pw-170879v1_FULL that come from known issues:
### IGT changes ###
#### Issues hit ####
* igt@kms_big_fb@linear-max-hw-stride-64bpp-rotate-0-hflip:
- shard-bmg: NOTRUN -> [SKIP][1] ([Intel XE#7059] / [Intel XE#7085])
[1]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-7/igt@[email protected]
* igt@kms_big_fb@y-tiled-16bpp-rotate-0:
- shard-bmg: NOTRUN -> [SKIP][2] ([Intel XE#1124]) +6 other tests skip
[2]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-10/igt@[email protected]
* igt@kms_bw@connected-linear-tiling-3-displays-target-2160x1440p:
- shard-bmg: NOTRUN -> [SKIP][3] ([Intel XE#7679])
[3]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-1/igt@[email protected]
* igt@kms_bw@linear-tiling-3-displays-target-2560x1440p:
- shard-bmg: NOTRUN -> [SKIP][4] ([Intel XE#367])
[4]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-6/igt@[email protected]
* igt@kms_ccs@bad-pixel-format-4-tiled-dg2-mc-ccs:
- shard-bmg: NOTRUN -> [SKIP][5] ([Intel XE#2887]) +7 other tests skip
[5]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-2/igt@[email protected]
* igt@kms_ccs@random-ccs-data-4-tiled-lnl-ccs@pipe-c-dp-2:
- shard-bmg: NOTRUN -> [SKIP][6] ([Intel XE#2652]) +8 other tests skip
[6]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-9/igt@kms_ccs@[email protected]
* igt@kms_chamelium_color_pipeline@plane-lut3d-green-only:
- shard-bmg: NOTRUN -> [SKIP][7] ([Intel XE#7358]) +1 other test skip
[7]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-10/igt@[email protected]
* igt@kms_chamelium_frames@hdmi-frame-dump:
- shard-bmg: NOTRUN -> [SKIP][8] ([Intel XE#2252]) +4 other tests skip
[8]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-10/igt@[email protected]
* igt@kms_content_protection@dp-mst-type-1-suspend-resume:
- shard-bmg: NOTRUN -> [SKIP][9] ([Intel XE#6974])
[9]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-6/igt@[email protected]
* igt@kms_content_protection@legacy-hdcp14@pipe-a-dp-2:
- shard-bmg: NOTRUN -> [FAIL][10] ([Intel XE#1178] / [Intel XE#3304] / [Intel XE#7374]) +3 other tests fail
[10]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-10/igt@kms_content_protection@[email protected]
* igt@kms_cursor_crc@cursor-offscreen-128x42:
- shard-bmg: NOTRUN -> [SKIP][11] ([Intel XE#2320]) +3 other tests skip
[11]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-2/igt@[email protected]
* igt@kms_cursor_legacy@short-busy-flip-before-cursor-atomic-transitions:
- shard-bmg: NOTRUN -> [SKIP][12] ([Intel XE#2286] / [Intel XE#6035])
[12]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-3/igt@kms_cursor_legacy@short-busy-flip-before-cursor-atomic-transitions.html
* igt@kms_dirtyfb@drrs-dirtyfb-ioctl:
- shard-bmg: NOTRUN -> [SKIP][13] ([Intel XE#1508])
[13]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-8/igt@[email protected]
* igt@kms_dp_link_training@non-uhbr-mst:
- shard-bmg: NOTRUN -> [SKIP][14] ([Intel XE#4354] / [Intel XE#5882])
[14]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-6/igt@[email protected]
* igt@kms_dsc@dsc-with-output-formats-with-bpc-ultrajoiner:
- shard-bmg: NOTRUN -> [SKIP][15] ([Intel XE#8265]) +1 other test skip
[15]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-7/igt@[email protected]
* igt@kms_flip@flip-vs-expired-vblank-interruptible:
- shard-lnl: [PASS][16] -> [FAIL][17] ([Intel XE#301]) +1 other test fail
[16]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-lnl-8/igt@[email protected]
[17]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-lnl-2/igt@[email protected]
* igt@kms_flip_scaled_crc@flip-32bpp-4tile-to-32bpp-4tiledg2rcccs-downscaling:
- shard-bmg: NOTRUN -> [SKIP][18] ([Intel XE#7178] / [Intel XE#7349])
[18]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-6/igt@kms_flip_scaled_crc@flip-32bpp-4tile-to-32bpp-4tiledg2rcccs-downscaling.html
* igt@kms_frontbuffer_tracking@drrs-argb161616f-draw-blt:
- shard-bmg: NOTRUN -> [SKIP][19] ([Intel XE#7061] / [Intel XE#7356]) +2 other tests skip
[19]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-10/igt@[email protected]
* igt@kms_frontbuffer_tracking@fbc-1p-primscrn-pri-shrfb-draw-mmap-wc:
- shard-bmg: NOTRUN -> [SKIP][20] ([Intel XE#4141]) +5 other tests skip
[20]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-6/igt@kms_frontbuffer_tracking@fbc-1p-primscrn-pri-shrfb-draw-mmap-wc.html
* igt@kms_frontbuffer_tracking@fbcdrrshdr-2p-primscrn-pri-shrfb-draw-mmap-wc:
- shard-bmg: NOTRUN -> [SKIP][21] ([Intel XE#2311]) +33 other tests skip
[21]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-8/igt@kms_frontbuffer_tracking@fbcdrrshdr-2p-primscrn-pri-shrfb-draw-mmap-wc.html
* igt@kms_frontbuffer_tracking@fbcdrrshdr-tiling-y:
- shard-bmg: NOTRUN -> [SKIP][22] ([Intel XE#7399]) +1 other test skip
[22]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-8/igt@[email protected]
* igt@kms_frontbuffer_tracking@hdr-abgr161616f-draw-render:
- shard-bmg: NOTRUN -> [SKIP][23] ([Intel XE#7061]) +2 other tests skip
[23]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-10/igt@[email protected]
* igt@kms_frontbuffer_tracking@psrhdr-1p-primscrn-pri-shrfb-draw-render:
- shard-bmg: NOTRUN -> [SKIP][24] ([Intel XE#2313]) +34 other tests skip
[24]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-1/igt@kms_frontbuffer_tracking@psrhdr-1p-primscrn-pri-shrfb-draw-render.html
* igt@kms_hdmi_inject@inject-audio:
- shard-bmg: NOTRUN -> [SKIP][25] ([Intel XE#7308])
[25]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-6/igt@[email protected]
* igt@kms_hdr@invalid-hdr:
- shard-bmg: [PASS][26] -> [SKIP][27] ([Intel XE#1503])
[26]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-7/igt@[email protected]
[27]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-3/igt@[email protected]
* igt@kms_plane@pixel-format-y-tiled-ccs-modifier-source-clamping:
- shard-bmg: NOTRUN -> [SKIP][28] ([Intel XE#7283]) +2 other tests skip
[28]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-6/igt@[email protected]
* igt@kms_pm_backlight@brightness-with-dpms:
- shard-bmg: NOTRUN -> [SKIP][29] ([Intel XE#2938] / [Intel XE#7376] / [Intel XE#7760])
[29]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-9/igt@[email protected]
* igt@kms_pm_dc@dc6-dpms:
- shard-lnl: [PASS][30] -> [FAIL][31] ([Intel XE#8399]) +1 other test fail
[30]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-lnl-2/igt@[email protected]
[31]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-lnl-3/igt@[email protected]
* igt@kms_pm_dc@dc6-psr:
- shard-bmg: NOTRUN -> [SKIP][32] ([Intel XE#7794])
[32]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-6/igt@[email protected]
* igt@kms_pm_dc@deep-pkgc:
- shard-bmg: NOTRUN -> [SKIP][33] ([Intel XE#2505] / [Intel XE#7447])
[33]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-7/igt@[email protected]
* igt@kms_psr2_sf@psr2-cursor-plane-move-continuous-exceed-fully-sf:
- shard-bmg: NOTRUN -> [SKIP][34] ([Intel XE#1489]) +5 other tests skip
[34]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-4/igt@kms_psr2_sf@psr2-cursor-plane-move-continuous-exceed-fully-sf.html
* igt@kms_psr@fbc-psr-cursor-plane-move:
- shard-bmg: NOTRUN -> [SKIP][35] ([Intel XE#2234] / [Intel XE#2850]) +5 other tests skip
[35]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-9/igt@[email protected]
* igt@kms_psr@psr2-primary-render:
- shard-bmg: NOTRUN -> [SKIP][36] ([Intel XE#2234])
[36]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-3/igt@[email protected]
* igt@kms_tv_load_detect@load-detect:
- shard-bmg: NOTRUN -> [SKIP][37] ([Intel XE#2450] / [Intel XE#5857])
[37]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-8/igt@[email protected]
* igt@xe_evict@evict-threads-small-multi-queue:
- shard-bmg: NOTRUN -> [SKIP][38] ([Intel XE#8370])
[38]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-4/igt@[email protected]
* igt@xe_exec_basic@multigpu-many-execqueues-many-vm-userptr:
- shard-bmg: NOTRUN -> [SKIP][39] ([Intel XE#2322] / [Intel XE#7372]) +2 other tests skip
[39]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-3/igt@[email protected]
* igt@xe_exec_fault_mode@many-execqueues-multi-queue-rebind-prefetch:
- shard-bmg: NOTRUN -> [SKIP][40] ([Intel XE#8374]) +6 other tests skip
[40]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-9/igt@xe_exec_fault_mode@many-execqueues-multi-queue-rebind-prefetch.html
* igt@xe_exec_multi_queue@few-execs-preempt-mode-fault-dyn-priority:
- shard-bmg: NOTRUN -> [SKIP][41] ([Intel XE#8364]) +14 other tests skip
[41]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-2/igt@xe_exec_multi_queue@few-execs-preempt-mode-fault-dyn-priority.html
* igt@xe_exec_reset@multi-queue-close-fd:
- shard-bmg: NOTRUN -> [SKIP][42] ([Intel XE#8369])
[42]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-9/igt@[email protected]
* igt@xe_exec_threads@threads-multi-queue-rebind-err:
- shard-bmg: NOTRUN -> [SKIP][43] ([Intel XE#8378]) +4 other tests skip
[43]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-10/igt@[email protected]
* igt@xe_multigpu_svm@mgpu-xgpu-access-basic:
- shard-bmg: NOTRUN -> [SKIP][44] ([Intel XE#6964])
[44]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-3/igt@[email protected]
* igt@xe_oa@non-zero-reason-all@oag-0:
- shard-bmg: [PASS][45] -> [FAIL][46] ([Intel XE#7334]) +1 other test fail
[45]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-1/igt@xe_oa@[email protected]
[46]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-9/igt@xe_oa@[email protected]
* igt@xe_pm@s3-d3cold-basic-exec:
- shard-bmg: NOTRUN -> [SKIP][47] ([Intel XE#2284] / [Intel XE#7370])
[47]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-4/igt@[email protected]
* igt@xe_pxp@pxp-stale-queue-post-termination-irq:
- shard-bmg: NOTRUN -> [SKIP][48] ([Intel XE#4733] / [Intel XE#7417]) +1 other test skip
[48]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-6/igt@[email protected]
* igt@xe_query@multigpu-query-gt-list:
- shard-bmg: NOTRUN -> [SKIP][49] ([Intel XE#944]) +1 other test skip
[49]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-8/igt@[email protected]
* igt@xe_sriov_flr@flr-twice:
- shard-bmg: [PASS][50] -> [FAIL][51] ([Intel XE#6569])
[50]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-9/igt@[email protected]
[51]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-1/igt@[email protected]
* igt@xe_sriov_vram@vf-access-after-resize-down:
- shard-bmg: [PASS][52] -> [FAIL][53] ([Intel XE#7992])
[52]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-3/igt@[email protected]
[53]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-4/igt@[email protected]
#### Possible fixes ####
* igt@kms_cursor_legacy@flip-vs-cursor-atomic:
- shard-bmg: [FAIL][54] ([Intel XE#7571]) -> [PASS][55]
[54]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-8/igt@[email protected]
[55]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-10/igt@[email protected]
* igt@kms_flip@plain-flip-fb-recreate:
- shard-bmg: [FAIL][56] ([Intel XE#3098]) -> [PASS][57] +1 other test pass
[56]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-1/igt@[email protected]
[57]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-6/igt@[email protected]
* igt@xe_evict@evict-beng-mixed-many-threads-small:
- shard-bmg: [INCOMPLETE][58] ([Intel XE#6321] / [Intel XE#8355]) -> [PASS][59]
[58]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-9/igt@[email protected]
[59]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-3/igt@[email protected]
* igt@xe_exec_fault_mode@atomic-many:
- shard-bmg: [FAIL][60] ([Intel XE#8578]) -> [PASS][61]
[60]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-6/igt@[email protected]
[61]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-4/igt@[email protected]
* igt@xe_fault_injection@inject-fault-probe-function-xe_guc_ads_init:
- shard-bmg: [ABORT][62] ([Intel XE#8007]) -> [PASS][63] +1 other test pass
[62]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-5/igt@xe_fault_injection@inject-fault-probe-function-xe_guc_ads_init.html
[63]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-8/igt@xe_fault_injection@inject-fault-probe-function-xe_guc_ads_init.html
* igt@xe_pmu@all-fn-engine-activity-load@engine-drm_xe_engine_class_video_enhance0:
- shard-bmg: [FAIL][64] ([Intel XE#8639]) -> [PASS][65] +1 other test pass
[64]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-2/igt@xe_pmu@all-fn-engine-activity-load@engine-drm_xe_engine_class_video_enhance0.html
[65]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-8/igt@xe_pmu@all-fn-engine-activity-load@engine-drm_xe_engine_class_video_enhance0.html
* igt@xe_pmu@all-fn-engine-activity-load@engine-drm_xe_engine_class_video_enhance1:
- shard-bmg: [ABORT][66] ([Intel XE#8637]) -> [PASS][67] +1 other test pass
[66]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-2/igt@xe_pmu@all-fn-engine-activity-load@engine-drm_xe_engine_class_video_enhance1.html
[67]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-8/igt@xe_pmu@all-fn-engine-activity-load@engine-drm_xe_engine_class_video_enhance1.html
#### Warnings ####
* igt@kms_cursor_legacy@cursorb-vs-flipb-atomic-transitions:
- shard-lnl: [SKIP][68] ([Intel XE#309] / [Intel XE#7343] / [Intel XE#7935]) -> [SKIP][69] ([Intel XE#309] / [Intel XE#7343])
[68]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-lnl-8/igt@[email protected]
[69]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-lnl-2/igt@[email protected]
* igt@kms_hdr@brightness-with-hdr:
- shard-bmg: [SKIP][70] ([Intel XE#3544]) -> [SKIP][71] ([Intel XE#3374] / [Intel XE#3544])
[70]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-5/igt@[email protected]
[71]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-6/igt@[email protected]
* igt@kms_tiled_display@basic-test-pattern:
- shard-bmg: [SKIP][72] ([Intel XE#2426] / [Intel XE#5848]) -> [FAIL][73] ([Intel XE#1729] / [Intel XE#7424])
[72]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-6/igt@[email protected]
[73]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-4/igt@[email protected]
* igt@kms_tiled_display@basic-test-pattern-with-chamelium:
- shard-bmg: [SKIP][74] ([Intel XE#2509] / [Intel XE#7437]) -> [SKIP][75] ([Intel XE#2426] / [Intel XE#5848])
[74]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c/shard-bmg-10/igt@[email protected]
[75]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/shard-bmg-9/igt@[email protected]
[Intel XE#1124]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/1124
[Intel XE#1178]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/1178
[Intel XE#1489]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/1489
[Intel XE#1503]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/1503
[Intel XE#1508]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/1508
[Intel XE#1729]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/1729
[Intel XE#2234]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2234
[Intel XE#2252]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2252
[Intel XE#2284]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2284
[Intel XE#2286]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2286
[Intel XE#2311]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2311
[Intel XE#2313]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2313
[Intel XE#2320]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2320
[Intel XE#2322]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2322
[Intel XE#2426]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2426
[Intel XE#2450]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2450
[Intel XE#2505]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2505
[Intel XE#2509]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2509
[Intel XE#2652]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2652
[Intel XE#2850]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2850
[Intel XE#2887]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2887
[Intel XE#2938]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2938
[Intel XE#301]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/301
[Intel XE#309]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/309
[Intel XE#3098]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/3098
[Intel XE#3304]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/3304
[Intel XE#3374]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/3374
[Intel XE#3544]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/3544
[Intel XE#367]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/367
[Intel XE#4141]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/4141
[Intel XE#4354]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/4354
[Intel XE#4733]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/4733
[Intel XE#5848]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/5848
[Intel XE#5857]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/5857
[Intel XE#5882]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/5882
[Intel XE#6035]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/6035
[Intel XE#6321]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/6321
[Intel XE#6569]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/6569
[Intel XE#6964]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/6964
[Intel XE#6974]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/6974
[Intel XE#7059]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7059
[Intel XE#7061]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7061
[Intel XE#7085]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7085
[Intel XE#7178]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7178
[Intel XE#7283]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7283
[Intel XE#7308]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7308
[Intel XE#7334]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7334
[Intel XE#7343]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7343
[Intel XE#7349]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7349
[Intel XE#7356]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7356
[Intel XE#7358]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7358
[Intel XE#7370]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7370
[Intel XE#7372]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7372
[Intel XE#7374]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7374
[Intel XE#7376]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7376
[Intel XE#7399]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7399
[Intel XE#7417]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7417
[Intel XE#7424]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7424
[Intel XE#7437]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7437
[Intel XE#7447]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7447
[Intel XE#7571]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7571
[Intel XE#7679]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7679
[Intel XE#7760]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7760
[Intel XE#7794]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7794
[Intel XE#7935]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7935
[Intel XE#7992]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7992
[Intel XE#8007]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/8007
[Intel XE#8265]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/8265
[Intel XE#8355]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/8355
[Intel XE#8364]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/8364
[Intel XE#8369]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/8369
[Intel XE#8370]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/8370
[Intel XE#8374]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/8374
[Intel XE#8378]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/8378
[Intel XE#8399]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/8399
[Intel XE#8578]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/8578
[Intel XE#8637]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/8637
[Intel XE#8639]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/8639
[Intel XE#944]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/944
Build changes
-------------
* IGT: IGT_9017 -> IGT_9018
* Linux: xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c -> xe-pw-170879v1
IGT_9017: 9017
IGT_9018: f4345861cbd5d97fd50b73bf7af5fe3286fca056 @ https://gitlab.freedesktop.org/drm/igt-gpu-tools.git
xe-5453-5726f9c4bde0290e5e60e636b9632bb03a57498c: 5726f9c4bde0290e5e60e636b9632bb03a57498c
xe-pw-170879v1: 170879v1
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-170879v1/index.html
v4 cover drm/xe/guc_ads: allocate UM queues in a separate UC BO
In reply to Jia Yao== Problem ==
On platforms with Unified Shared Memory (USM), the xe driver maintains three hardware page-fault queues (PAGE_FAULT, PAGE_FAULT_RESPONSE, ACCESS_COUNTER) inside the GuC Additional Data Structures (ADS) blob. The ADS blob is mapped WB on the CPU side.
xe_guc_ads_populate() is called on every GT reset to reprogram the ADS blob; it begins with a full memset() of the blob to zero. The zeroed cache lines for the UM queue region reside in the CPU cache (LLC, WB) but are not immediately written back to DRAM.
GAM hardware writes fault descriptors to the queue via DPA - a non-coherent path that bypasses the CPU cache and writes directly to memory. GuC reads the queue via GGTT (WB), which hits the CPU cache. As a result GuC sees the stale zeroed data from the CPU cache instead of the descriptor written to DRAM by GAM. The driver then receives an all-zero fault descriptor and returns -EINVAL, causing an unnecessary engine reset.
This is documented in HSD as:
"When Fault queue is updated by the SW (Buffer zeroed), it can reside
in the CPU cache (L4:WB). A non coherent write from GAM HW would not consult the CPU cache and it will eventually get written into the memory (depends upon eviction/flushes) and CPU cache has stale data. When GUC HW reads the fault queue it gets it from the CPU cache, which is stale data."
The window is narrow but reliably reproducible: run xe_exec_reset --run-subtest gt-reset followed immediately by xe_exec_system_allocator --run-subtest threads-many-execqueues-mmap-new-race.
== Reproduction ==
# Requires a platform with has_usm (e.g. Xe3 iGFX) sudo xe_exec_reset --run-subtest gt-reset sudo xe_exec_system_allocator --run-subtest threads-many-execqueues-mmap-new-race
Symptom in dmesg:
[xe] ASID: 0, Faulted Address: 0x0000000000000000, FaultType: 0
The all-zero fault descriptor is the tell-tale sign of the stale cache line race.
== Fix ==
Allocate the UM queues in a dedicated UC BO (ads->um_queue_bo). UC mapping on the CPU side ensures the memset bypasses the cache and the zeroed data lands in DRAM immediately. GuC's non-coherent read then goes to DRAM and sees the correct descriptor written by GAM.
A separate BO also eliminates the secondary race where a descriptor written by the GPU between GDRST and GuC restart could be zeroed by the next call to xe_guc_ads_populate().
On dGFX, VRAM placement is used so that xe_bo_main_addr() returns a real device-physical address (DPA). The original ADS BO is explicitly migrated to VRAM by xe_guc_realloc_post_hwconfig(), so its xe_bo_main_addr() already returns a correct DPA. A separate SYSTEM BO for the UM queues would not go through that migration path; on systems where physical pages are allocated above 4 GB the IOMMU assigns an IOVA that differs from the physical address, so base_dpa would be wrong and GuC/GAM would access the wrong memory.
Why not just flush the cache after memset?
An earlier approach called drm_clflush_virt_range() on the UM queue region immediately after the memset to evict the stale zero lines from LLC. This reduces the failure rate significantly but does not fully eliminate the race for two reasons.
First, the Linux kernel maintains a WB direct mapping (linear map) that covers all physical memory. Even after flushing the vmap alias, the CPU can speculatively prefetch the same physical page through the WB direct-map alias, silently repopulating the cache with zeros. LNL is known to be particularly susceptible to this speculative prefetch behaviour. XE_BO_FLAG_NEEDS_UC calls set_memory_uc() which updates both the vmap PTE and the direct-map PTE to UC, closing this aliasing window entirely.
Second, the flush approach leaves the memory mapped WB on the CPU side permanently. Any future SW access - a debug read, a tracing hook, an inadvertent touch - can re-dirty the cache line and reintroduce the race. UC mapping makes the correct behaviour unconditional and self-documenting.
Changes in v4:
- Split the single v3 patch into three logically independent patches:
1. Separate BO allocation (root fix for the memset-zeroing race)
2. VRAM placement on dGFX (fix incorrect DPA when pages are above 4 GB)
3. UC mapping (close the CPU cache aliasing window on iGFX)
- Corrected the dGFX description: the original ADS BO works because xe_guc_realloc_post_hwconfig() explicitly migrates it to VRAM; a separate SYSTEM BO would not be migrated and its IOVA would differ from the physical address for high-memory pages. The previous description incorrectly attributed this to IOVA aliasing a PCI MMIO reserved window.
- Added physical-contiguity verification for no-IOMMU iGFX systems.
- xe_guc.c: removed the post-hwconfig reinit_in_vram call for um_queue_bo now that dGFX allocates it directly in VRAM.
Jia Yao (3):
drm/xe/guc_ads: allocate UM queues in a separate BO drm/xe/guc_ads: allocate UM queues in VRAM on dGFX drm/xe/guc_ads: use uncached mapping for UM queue BO
drivers/gpu/drm/xe/xe_guc_ads.c | 100 +++++++++++++++++++------- drivers/gpu/drm/xe/xe_guc_ads_types.h | 5 ++ 2 files changed, 80 insertions(+), 25 deletions(-)
Signature · 1 line
2.43.0
PATCH v4 1/3 drm/xe/guc_ads: allocate UM queues in a separate BO
xe_guc_ads_populate() calls xe_map_memset() on the ADS BO on every GT reset. The UM queue ring buffers were embedded in that same BO, so the memset discards fault descriptors already written by the GPU, causing GuC to forward an all-zero descriptor (Faulted Address = 0, ASID = 0).
Move the UM queue ring buffers into a dedicated BO so they are not affected by the ADS memset.
--- drivers/gpu/drm/xe/xe_guc.c | 6 ++ drivers/gpu/drm/xe/xe_guc_ads.c | 98 ++++++++++++++++++++------- drivers/gpu/drm/xe/xe_guc_ads_types.h | 5 ++ 3 files changed, 84 insertions(+), 25 deletions(-)
drivers/gpu/drm/xe/xe_guc.c +6 -0
@@ -714,6 +714,12 @@ static int xe_guc_realloc_post_hwconfig(struct xe_guc *guc)if (ret)return ret; +if (guc->ads.um_queue_bo) { + ret =xe_managed_bo_reinit_in_vram (xe, tile, &guc->ads.um_queue_bo); +if (ret) +return ret; + } +return 0 ; }
drivers/gpu/drm/xe/xe_guc_ads.c +73 -25
138 lines not on this page · raw message
drivers/gpu/drm/xe/xe_guc_ads_types.h +5 -0
@@ -16,6 +16,11 @@ struct xe_bo;struct xe_guc_ads {/** @bo: Xe BO for GuC ads blob */ struct xe_bo *bo; +/** +* @um_queue_bo: Dedicated BO for the HW fault ring (UM queues). +* NULL if the platform does not support USM. +*/ +struct xe_bo *um_queue_bo;/** @golden_lrc_size: golden LRC size */ size_t golden_lrc_size;/** @regset_size: size of register set passed to GuC for save/restore */
Signature · 1 line
2.43.0
Re: [PATCH v4 1/3] drm/xe/guc_ads: allocate UM queues in a separate BO
Quoted 100 lines · Jia Yao
xe_guc_ads_populate()callsxe_map_memset()on the ADS BO on every GT reset. The UM queue ring buffers were embedded in that same BO, so the memset discards fault descriptors already written by the GPU, causing GuC to forward an all-zero descriptor (Faulted Address = 0, ASID = 0).Move the UM queue ring buffers into a dedicated BO so they are not affected by the ADS memset.
Fixes dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")Cc Gwan-gyeong Mun <[email protected]>Cc Matthew Auld <[email protected]>Signed-off-by Jia Yao <[email protected]>--- drivers/gpu/drm/xe/xe_guc.c | 6 ++ drivers/gpu/drm/xe/xe_guc_ads.c | 98 ++++++++++++++++++++------- drivers/gpu/drm/xe/xe_guc_ads_types.h | 5 ++ 3 files changed, 84 insertions(+), 25 deletions(-)drivers/gpu/drm/xe/xe_guc.c +6
-0@@ -714,6 +714,12 @@ static int xe_guc_realloc_post_hwconfig(struct xe_guc *guc)if (ret)return ret; +if (guc->ads.um_queue_bo) { + ret =xe_managed_bo_reinit_in_vram (xe, tile, &guc->ads.um_queue_bo); +if (ret) +return ret; + } +return 0 ; }drivers/gpu/drm/xe/xe_guc_ads.c +9
-22@@ -16,6 +16,7 @@#include "regs/xe_gt_regs.h" #include "regs/xe_guc_regs.h" #include "xe_bo.h" +#include "xe_res_cursor.h" #include "xe_gt.h" #include "xe_gt_ccs_mode.h" #include "xe_gt_mcr.h" @@ -155,16 +156,6 @@ static size_t guc_ads_capture_size(struct xe_guc_ads *ads)return PAGE_ALIGN (ads->capture_size); } -static size_t guc_ads_um_queues_size (struct xe_guc_ads *ads) -{ -struct xe_device *xe =ads_to_xe (ads); - -if (!xe->info.has_usm) -return 0 ; - -return GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX ; -} -static size_t guc_ads_private_data_size (struct xe_guc_ads *ads) {return PAGE_ALIGN (ads_to_guc (ads)->fw.private_data_size);@@ -205,22 +196,12 @@ static size_t guc_ads_capture_offset(struct xe_guc_ads *ads)return PAGE_ALIGN (offset); } -static size_t guc_ads_um_queues_offset (struct xe_guc_ads *ads) -{ -u32 offset; - - offset =guc_ads_capture_offset (ads) + -guc_ads_capture_size (ads); - -return PAGE_ALIGN (offset); -} -static size_t guc_ads_private_data_offset (struct xe_guc_ads *ads) {size_t offset; - offset =guc_ads_um_queues_offset (ads) + -guc_ads_um_queues_size (ads); + offset =guc_ads_capture_offset (ads) + +guc_ads_capture_size (ads);return PAGE_ALIGN (offset); }@@ -409,6 +390,67 @@ int xe_guc_ads_init(struct xe_guc_ads *ads) ads->bo = bo; +if (xe->info.has_usm) { +/* +* Allocate a separate BO for the HW fault ring (UM queues). +* +* xe_guc_ads_populate() clears the entire ADS BO via +* xe_map_memset() on every GT reset. Keeping the UM queues
drivers/gpu/drm/xe/xe_guc_ads.c int xe_guc_ads_init(struct xe_guc_ads *ads) · line 408
Quoted 10 lines
drivers/gpu/drm/xe/xe_guc_ads.c +10
-0+* in a dedicated BO prevents that memset from discarding fault +* descriptors the GPU has already written into the ring, +* which would cause GuC to forward an all-zero descriptor to +* the driver (Faulted Address = 0, ASID = 0). +* +* Round the size up to the next power of two so that on iGPU +* (system memory, no IOMMU) the TTM pool issues a single +* alloc_pages(order=N) call, maximising the chance of getting +* a physically contiguous block. GuC requires contiguous DPA. +*/
From our offline discussion I think is this is correct, however there is no mention of this fix/change in the commit message? It only mentions memset issue.
Quoted 10 lines
drivers/gpu/drm/xe/xe_guc_ads.c +10
-0+size_t um_size =roundup_pow_of_two (GUC_UM_QUEUE_SIZE * +GUC_UM_HW_QUEUE_MAX ); + + bo =xe_managed_bo_create_pin_map (xe, tile, um_size, +XE_BO_FLAG_SYSTEM | +XE_BO_FLAG_GGTT | +XE_BO_FLAG_GGTT_INVALIDATE | +XE_BO_FLAG_PINNED_NORESTORE ); +if (IS_ERR (bo)) +return PTR_ERR (bo);
drivers/gpu/drm/xe/xe_guc_ads.c int xe_guc_ads_init(struct xe_guc_ads *ads) · line 428
Quoted 10 lines
drivers/gpu/drm/xe/xe_guc_ads.c +10
-0+ +/* +* On pre-Xe3p platforms, GAM (not GuC) accesses the UM queue +* ring via base_dpa, which must be a contiguous device physical +* address range. Without IOMMU, DMA addresses equal physical +* addresses and the allocator may return non-contiguous pages; +* verify contiguity. With IOMMU, the DMA mapping is a single +* contiguous IOVA so no check is needed. +*/ + if (unlikely(!device_iommu_mapped(xe->drm.dev)) &&
I think maybe drop the iommu check? Not sure if we can safely assume it will always give single contig dma blob.
drivers/gpu/drm/xe/xe_guc_ads.c int xe_guc_ads_init(struct xe_guc_ads *ads) · line 438
Quoted 10 lines
drivers/gpu/drm/xe/xe_guc_ads.c +10
-0+ !xe_bo_is_vram(bo) && + !xe_guc_using_main_gamctrl_queues(ads_to_guc(ads))) { +struct xe_res_cursor cur; +int i; + +xe_res_first_sg (xe_bo_sg (bo),0 , +GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX , &cur); +for (i =1 ; i <GUC_UM_HW_QUEUE_MAX ; i++) { +dma_addr_t expected =xe_res_dma (&cur) + +GUC_UM_QUEUE_SIZE ;
I think we can maybe simplify with:
if (drm_prime_get_contiguous_size(xe_bo_sg(bo)) < queue size)
...
Quoted 60 lines
drivers/gpu/drm/xe/xe_guc_ads.c +24
-3+ +xe_res_next (&cur,GUC_UM_QUEUE_SIZE ); +if (xe_res_dma (&cur) != expected) { +drm_err (&xe->drm, +"UM fault queue %d is physically non-contiguous (got %pad, expected %pad); GAM requires contiguous DPA. Enable IOMMU (intel_iommu=on) to fix.\n" , + i, &(dma_addr_t ){xe_res_dma (&cur)}, + &expected); +return -ENOMEM ; + } + } + } + + ads->um_queue_bo = bo; + } +return 0 ; }ALLOW_ERROR_INJECTION (xe_guc_ads_init,ERRNO );/* See xe_pci_probe() */ @@ -820,7 +862,7 @@ static void guc_mmio_reg_state_init(struct xe_guc_ads *ads)static void guc_um_init_params (struct xe_guc_ads *ads) { -u32 um_queue_offset =guc_ads_um_queues_offset (ads); +struct xe_bo *um_bo = ads->um_queue_bo;struct xe_guc *guc =ads_to_guc (ads);struct xe_device *xe =ads_to_xe (ads);u64 base_dpa;@@ -830,8 +872,14 @@ static void guc_um_init_params(struct xe_guc_ads *ads) with_dpa = !xe_guc_using_main_gamctrl_queues (guc); - base_ggtt =xe_bo_ggtt_addr (ads->bo) + um_queue_offset; - base_dpa =xe_bo_main_addr (ads->bo,PAGE_SIZE ) + um_queue_offset; +if (um_bo) { +/* All USM platforms: UM queues in dedicated um_queue_bo */ + base_ggtt =xe_bo_ggtt_addr (um_bo); + base_dpa =xe_bo_main_addr (um_bo,PAGE_SIZE ); + }else { +/* Platform does not support USM: no UM queues, nothing to do */ +return ; + }for (i =0 ; i <GUC_UM_HW_QUEUE_MAX ; ++i) {/* drivers/gpu/drm/xe/xe_guc_ads_types.h +5
-0@@ -16,6 +16,11 @@ struct xe_bo;struct xe_guc_ads {/** @bo: Xe BO for GuC ads blob */ struct xe_bo *bo; +/** +* @um_queue_bo: Dedicated BO for the HW fault ring (UM queues). +* NULL if the platform does not support USM. +*/ +struct xe_bo *um_queue_bo;/** @golden_lrc_size: golden LRC size */ size_t golden_lrc_size;/** @regset_size: size of register set passed to GuC for save/restore */
PATCH v4 2/3 drm/xe/guc_ads: allocate UM queues in VRAM on dGFX
In reply to Jia YaoOn dGFX the original ADS BO is explicitly migrated to VRAM by xe_guc_realloc_post_hwconfig() so that xe_bo_main_addr() returns a real device-physical address (DPA) rather than an IOMMU IOVA. A separate SYSTEM BO for the UM queues would not go through that migration path, so its IOVA would differ from the physical address on systems where pages are allocated above 4 GB. GuC/GAM accesses the UM queue ring via base_dpa (with_dpa=true on pre-Xe3p), so an incorrect DPA causes all-zero fault descriptors to be forwarded to the driver.
Use XE_BO_FLAG_VRAM_IF_DGFX() for the UM queue BO so that on dGFX the BO is allocated directly in VRAM and xe_bo_main_addr() returns the correct DPA without needing a post-hwconfig migration step.
--- drivers/gpu/drm/xe/xe_guc.c | 6 ------ drivers/gpu/drm/xe/xe_guc_ads.c | 11 ++++++----- 2 files changed, 6 insertions(+), 11 deletions(-)
drivers/gpu/drm/xe/xe_guc.c +0 -6
@@ -714,12 +714,6 @@ static int xe_guc_realloc_post_hwconfig(struct xe_guc *guc)if (ret)return ret; -if (guc->ads.um_queue_bo) { - ret =xe_managed_bo_reinit_in_vram (xe, tile, &guc->ads.um_queue_bo); -if (ret) -return ret; - } -return 0 ; }
drivers/gpu/drm/xe/xe_guc_ads.c +6 -5
@@ -409,11 +409,12 @@ int xe_guc_ads_init(struct xe_guc_ads *ads)size_t um_size =roundup_pow_of_two (GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX ); - bo =xe_managed_bo_create_pin_map (xe, tile, um_size, -XE_BO_FLAG_SYSTEM | -XE_BO_FLAG_GGTT | -XE_BO_FLAG_GGTT_INVALIDATE | -XE_BO_FLAG_PINNED_NORESTORE ); +u32 um_flags =XE_BO_FLAG_VRAM_IF_DGFX (tile) | +XE_BO_FLAG_GGTT | +XE_BO_FLAG_GGTT_INVALIDATE | +XE_BO_FLAG_PINNED_NORESTORE ; + + bo =xe_managed_bo_create_pin_map (xe, tile, um_size, um_flags);if (IS_ERR (bo))return PTR_ERR (bo);
Signature · 1 line
2.43.0
Re: [PATCH v4 2/3] drm/xe/guc_ads: allocate UM queues in VRAM on dGFX
Quoted 8 lines · Jia Yao
On dGFX the original ADS BO is explicitly migrated to VRAM by
xe_guc_realloc_post_hwconfig()so thatxe_bo_main_addr()returns a real device-physical address (DPA) rather than an IOMMU IOVA. A separate SYSTEM BO for the UM queues would not go through that migration path, so its IOVA would differ from the physical address on systems where pages are allocated above 4 GB. GuC/GAM accesses the UM queue ring viabase_dpa(with_dpa=true on pre-Xe3p), so an incorrect DPA causes all-zero fault descriptors to be forwarded to the driver.
I feel like I don't fully understand the explanation here. Is the issue not just the contig requirement for base_dpa? We now use attempt a best effort approach for system memory, however for dgpu we can go further and use VRAM to guarantee that requirement. If that is all this is doing then commit message can be simplified with something like:
"To fully guarantee the contig requirement for base_dpa, we can instead just use VRAM for the um queue, on dgpu..."
Quoted 54 lines
Use
XE_BO_FLAG_VRAM_IF_DGFX()for the UM queue BO so that on dGFX the BO is allocated directly in VRAM andxe_bo_main_addr()returns the correct DPA without needing a post-hwconfig migration step.Fixes dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")Cc Gwan-gyeong Mun <[email protected]>Cc Matthew Auld <[email protected]>Signed-off-by Jia Yao <[email protected]>--- drivers/gpu/drm/xe/xe_guc.c | 6 ------ drivers/gpu/drm/xe/xe_guc_ads.c | 11 ++++++----- 2 files changed, 6 insertions(+), 11 deletions(-)drivers/gpu/drm/xe/xe_guc.c +0
-6@@ -714,12 +714,6 @@ static int xe_guc_realloc_post_hwconfig(struct xe_guc *guc)if (ret)return ret; -if (guc->ads.um_queue_bo) { - ret =xe_managed_bo_reinit_in_vram (xe, tile, &guc->ads.um_queue_bo); -if (ret) -return ret; - } -return 0 ; }drivers/gpu/drm/xe/xe_guc_ads.c +6
-5@@ -409,11 +409,12 @@ int xe_guc_ads_init(struct xe_guc_ads *ads)size_t um_size =roundup_pow_of_two (GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX ); - bo =xe_managed_bo_create_pin_map (xe, tile, um_size, -XE_BO_FLAG_SYSTEM | -XE_BO_FLAG_GGTT | -XE_BO_FLAG_GGTT_INVALIDATE | -XE_BO_FLAG_PINNED_NORESTORE ); +u32 um_flags =XE_BO_FLAG_VRAM_IF_DGFX (tile) | +XE_BO_FLAG_GGTT | +XE_BO_FLAG_GGTT_INVALIDATE | +XE_BO_FLAG_PINNED_NORESTORE ; + + bo =xe_managed_bo_create_pin_map (xe, tile, um_size, um_flags);if (IS_ERR (bo))return PTR_ERR (bo);
Re: [PATCH v4 2/3] drm/xe/guc_ads: allocate UM queues in VRAM on dGFX
Quoted 10 lines · Matthew Auld
Quoted 8 lines · Jia Yao
On dGFX the original ADS BO is explicitly migrated to VRAM by
xe_guc_realloc_post_hwconfig()so thatxe_bo_main_addr()returns a real device-physical address (DPA) rather than an IOMMU IOVA. A separate SYSTEM BO for the UM queues would not go through that migration path, so its IOVA would differ from the physical address on systems where pages are allocated above 4 GB. GuC/GAM accesses the UM queue ring viabase_dpa(with_dpa=true on pre-Xe3p), so an incorrect DPA causes all-zero fault descriptors to be forwarded to the driver.
Quoted 3 lines
I feel like I don't fully understand the explanation here. Is the issue not just the contig requirement for
base_dpa? We now use attempt a best effort approach for system memory, however for dgpu we can go further and use VRAM to guarantee that requirement. If that is all this is doing then commit message can be simplified with something like:"To fully guarantee the contig requirement for
base_dpa, we can instead just use VRAM for the um queue, on dgpu..."
Or is the issue that this has to always be in VRAM, since there is no bit to say whether the DPA is system vs VRAM, like we do for the GGTT pte? On dgpu, assumption is that this is then always VRAM, if using DPA? If so, please make that clearer in the commit message.
Quoted 54 lines
Quoted 53 lines
Use
XE_BO_FLAG_VRAM_IF_DGFX()for the UM queue BO so that on dGFX the BO is allocated directly in VRAM andxe_bo_main_addr()returns the correct DPA without needing a post-hwconfig migration step.Fixes dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")Cc Gwan-gyeong Mun <[email protected]>Cc Matthew Auld <[email protected]>Signed-off-by Jia Yao <[email protected]>---
drivers/gpu/drm/xe/xe_guc.c | 6 ------drivers/gpu/drm/xe/xe_guc_ads.c| 11 ++++++-----2 files changed, 6 insertions(+), 11 deletions(-)drivers/gpu/drm/xe/xe_guc.c +0
-6@@ -714,12 +714,6 @@ static int xe_guc_realloc_post_hwconfig(struct xe_guc *guc) if (ret) return ret; - if (guc->ads.um_queue_bo) { - ret = xe_managed_bo_reinit_in_vram(xe, tile, &guc- >ads.um_queue_bo); - if (ret) - return ret; - } - return 0; }drivers/gpu/drm/xe/xe_guc_ads.c +0
-0@@ -409,11 +409,12 @@ int xe_guc_ads_init(struct xe_guc_ads *ads)
size_tum_size=roundup_pow_of_two(GUC_UM_QUEUE_SIZE * GUC_UM_HW_QUEUE_MAX);
- bo = xe_managed_bo_create_pin_map(xe, tile,um_size, -XE_BO_FLAG_SYSTEM|
-XE_BO_FLAG_GGTT|
-XE_BO_FLAG_GGTT_INVALIDATE|
-XE_BO_FLAG_PINNED_NORESTORE);
+ u32um_flags=XE_BO_FLAG_VRAM_IF_DGFX(tile)|
+XE_BO_FLAG_GGTT|
+XE_BO_FLAG_GGTT_INVALIDATE|
+XE_BO_FLAG_PINNED_NORESTORE;
+
+ bo =xe_managed_bo_create_pin_map(xe, tile, um_size, um_flags); if (IS_ERR(bo))
returnPTR_ERR(bo);
Matthew Brost
maintainerRe: [PATCH v4 2/3] drm/xe/guc_ads: allocate UM queues in VRAM on dGFX
Quoted 17 lines · Matthew Auld
Quoted 16 lines · Matthew Auld
Quoted 8 lines · Jia Yao
On dGFX the original ADS BO is explicitly migrated to VRAM by
xe_guc_realloc_post_hwconfig()so thatxe_bo_main_addr()returns a real device-physical address (DPA) rather than an IOMMU IOVA. A separate SYSTEM BO for the UM queues would not go through that migration path, so its IOVA would differ from the physical address on systems where pages are allocated above 4 GB. GuC/GAM accesses the UM queue ring viabase_dpa(with_dpa=true on pre-Xe3p), so an incorrect DPA causes all-zero fault descriptors to be forwarded to the driver.I feel like I don't fully understand the explanation here. Is the issue not just the contig requirement for
base_dpa? We now use attempt a best effort approach for system memory, however for dgpu we can go further and use VRAM to guarantee that requirement. If that is all this is doing then commit message can be simplified with something like:
Quoted 8 lines
Quoted 2 lines
"To fully guarantee the contig requirement for
base_dpa, we can instead just use VRAM for the um queue, on dgpu..."Or is the issue that this has to always be in VRAM, since there is no bit to say whether the DPA is system vs VRAM, like we do for the GGTT pte? On dgpu, assumption is that this is then always VRAM, if using DPA? If so, please make that clearer in the commit message.
The register interface to program the page fault queues are DPA / VRAM on dGPU iirc.
Matt
Quoted 56 lines
Quoted 55 lines
Quoted 54 lines
Use
XE_BO_FLAG_VRAM_IF_DGFX()for the UM queue BO so that on dGFX the BO is allocated directly in VRAM andxe_bo_main_addr()returns the correct DPA without needing a post-hwconfig migration step.Fixes dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")Cc Gwan-gyeong Mun <[email protected]>Cc Matthew Auld <[email protected]>Signed-off-by Jia Yao <[email protected]>---
drivers/gpu/drm/xe/xe_guc.c | 6 ------drivers/gpu/drm/xe/xe_guc_ads.c| 11 ++++++-----2 files changed, 6 insertions(+), 11 deletions(-)drivers/gpu/drm/xe/xe_guc.c +0
-6@@ -714,12 +714,6 @@ static int xe_guc_realloc_post_hwconfig(struct xe_guc *guc) if (ret) return ret; - if (guc->ads.um_queue_bo) { - ret = xe_managed_bo_reinit_in_vram(xe, tile, &guc- >ads.um_queue_bo); - if (ret) - return ret; - } - return 0; }drivers/gpu/drm/xe/ renamed from drivers/gpu/drm/xe/xe_guc_ads.c +0
-0
xe_guc_ads.c
index7f371f9c7268..53636333fb45100644drivers/gpu/drm/xe/xe_guc_ads.c +0
-0@@ -409,11 +409,12 @@ int xe_guc_ads_init(struct xe_guc_ads *ads)
size_tum_size=roundup_pow_of_two(GUC_UM_QUEUE_SIZE * GUC_UM_HW_QUEUE_MAX);
- bo = xe_managed_bo_create_pin_map(xe, tile,um_size, -XE_BO_FLAG_SYSTEM|
-XE_BO_FLAG_GGTT|
-XE_BO_FLAG_GGTT_INVALIDATE|
-XE_BO_FLAG_PINNED_NORESTORE);
+ u32um_flags=XE_BO_FLAG_VRAM_IF_DGFX(tile)|
+XE_BO_FLAG_GGTT|
+XE_BO_FLAG_GGTT_INVALIDATE|
+XE_BO_FLAG_PINNED_NORESTORE;
+
+ bo =xe_managed_bo_create_pin_map(xe, tile, um_size, um_flags); if (IS_ERR(bo))
returnPTR_ERR(bo);
PATCH v4 3/3 drm/xe/guc_ads: use uncached mapping for UM queue BO
In reply to Jia YaoThe HW unit that writes fault descriptors into the UM queue ring buffer may not be coherent with the CPU cache. If a descriptor write is still sitting in a GPU L3 cache line when GuC reads the slot, GuC sees the stale (all-zero) content from DRAM and forwards an invalid descriptor to the driver (Faulted Address = 0, ASID = 0).
Mark the UM queue BO as uncached (XE_BO_FLAG_NEEDS_UC) so that writes bypass the cache and are immediately visible to GuC.
--- drivers/gpu/drm/xe/xe_guc_ads.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/xe/xe_guc_ads.c +2 -1
@@ -412,7 +412,8 @@ int xe_guc_ads_init(struct xe_guc_ads *ads)u32 um_flags =XE_BO_FLAG_VRAM_IF_DGFX (tile) |XE_BO_FLAG_GGTT |XE_BO_FLAG_GGTT_INVALIDATE | -XE_BO_FLAG_PINNED_NORESTORE ; +XE_BO_FLAG_PINNED_NORESTORE | +XE_BO_FLAG_NEEDS_UC ; bo =xe_managed_bo_create_pin_map (xe, tile, um_size, um_flags);if (IS_ERR (bo))
Signature · 1 line
2.43.0
Re: [PATCH v4 3/3] drm/xe/guc_ads: use uncached mapping for UM queue BO
Quoted 10 lines · Jia Yao
The HW unit that writes fault descriptors into the UM queue ring buffer may not be coherent with the CPU cache. If a descriptor write is still sitting in a GPU L3 cache line when GuC reads the slot, GuC sees the stale (all-zero) content from DRAM and forwards an invalid descriptor to the driver (Faulted Address = 0, ASID = 0).
Mark the UM queue BO as uncached (
XE_BO_FLAG_NEEDS_UC) so that writes bypass the cache and are immediately visible to GuC.Fixes dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
note that XE_BO_FLAG_NEEDS_UC flag is not available at this point
it was initially added as XE_BO_NEEDS_UC flag in a commit f15de1936f8d and then renamed to XE_BO_FLAG_NEEDS_UC in commit 62742d126631451
Quoted 22 lines
Cc Gwan-gyeong Mun <[email protected]>Cc Matthew Auld <[email protected]>Signed-off-by Jia Yao <[email protected]>--- drivers/gpu/drm/xe/xe_guc_ads.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/gpu/drm/xe/xe_guc_ads.c +2
-1@@ -412,7 +412,8 @@ int xe_guc_ads_init(struct xe_guc_ads *ads)u32 um_flags =XE_BO_FLAG_VRAM_IF_DGFX (tile) |XE_BO_FLAG_GGTT |XE_BO_FLAG_GGTT_INVALIDATE | -XE_BO_FLAG_PINNED_NORESTORE ; +XE_BO_FLAG_PINNED_NORESTORE | +XE_BO_FLAG_NEEDS_UC ; bo =xe_managed_bo_create_pin_map (xe, tile, um_size, um_flags);if (IS_ERR (bo))
Re: [PATCH v4 3/3] drm/xe/guc_ads: use uncached mapping for UM queue BO
In reply to Jia YaoQuoted 10 lines · Jia Yao
The HW unit that writes fault descriptors into the UM queue ring buffer may not be coherent with the CPU cache. If a descriptor write is still sitting in a GPU L3 cache line when GuC reads the slot, GuC sees the stale (all-zero) content from DRAM and forwards an invalid descriptor to the driver (Faulted Address = 0, ASID = 0).
Mark the UM queue BO as uncached (
XE_BO_FLAG_NEEDS_UC) so that writes bypass the cache and are immediately visible to GuC.Fixes dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
As per Michal this is too old. I think we can instead do:
Quoted 4 lines
Cc Gwan-gyeong Mun <[email protected]>Cc Matthew Auld <[email protected]>Signed-off-by Jia Yao <[email protected]>
Quoted 18 lines
--- drivers/gpu/drm/xe/xe_guc_ads.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/gpu/drm/xe/xe_guc_ads.c +2
-1@@ -412,7 +412,8 @@ int xe_guc_ads_init(struct xe_guc_ads *ads)u32 um_flags =XE_BO_FLAG_VRAM_IF_DGFX (tile) |XE_BO_FLAG_GGTT |XE_BO_FLAG_GGTT_INVALIDATE | -XE_BO_FLAG_PINNED_NORESTORE ; +XE_BO_FLAG_PINNED_NORESTORE | +XE_BO_FLAG_NEEDS_UC ; bo =xe_managed_bo_create_pin_map (xe, tile, um_size, um_flags);if (IS_ERR (bo))
v5 cover drm/xe/guc_ads: allocate UM queues in a separate UC BO
In reply to Jia Yao== Problem ==
On platforms with Unified Shared Memory (USM), the xe driver maintains three hardware page-fault queues (PAGE_FAULT, PAGE_FAULT_RESPONSE, ACCESS_COUNTER) inside the GuC Additional Data Structures (ADS) blob. The ADS blob is mapped WB on the CPU side.
xe_guc_ads_populate() is called on every GT reset to reprogram the ADS blob; it begins with a full memset() of the blob to zero. The zeroed cache lines for the UM queue region reside in the CPU cache (LLC, WB) but are not immediately written back to DRAM.
GAM hardware writes fault descriptors to the queue via DPA - a non-coherent path that bypasses the CPU cache and writes directly to memory. GuC reads the queue via GGTT (WB), which hits the CPU cache. As a result GuC sees the stale zeroed data from the CPU cache instead of the descriptor written to DRAM by GAM. The driver then receives an all-zero fault descriptor and returns -EINVAL, causing an unnecessary engine reset.
This is documented in HSD as:
"When Fault queue is updated by the SW (Buffer zeroed), it can reside
in the CPU cache (L4:WB). A non coherent write from GAM HW would not consult the CPU cache and it will eventually get written into the memory (depends upon eviction/flushes) and CPU cache has stale data. When GUC HW reads the fault queue it gets it from the CPU cache, which is stale data."
The window is narrow but reliably reproducible: run xe_exec_reset --run-subtest gt-reset followed immediately by xe_exec_system_allocator --run-subtest threads-many-execqueues-mmap-new-race.
== Reproduction ==
# Requires a platform with has_usm (e.g. Xe3 iGFX)
sudo xe_exec_reset --run-subtest gt-reset
sudo xe_exec_system_allocator \
--run-subtest threads-many-execqueues-mmap-new-raceSymptom in dmesg:
[xe] ASID: 0, Faulted Address: 0x0000000000000000, FaultType: 0
The all-zero fault descriptor is the tell-tale sign of the stale cache line race.
== Fix ==
Allocate the UM queues in a dedicated UC BO (ads->um_queue_bo). UC mapping on the CPU side ensures the memset bypasses the cache and the zeroed data lands in DRAM immediately. GuC's non-coherent read then goes to DRAM and sees the correct descriptor written by GAM.
A separate BO also eliminates the secondary race where a descriptor written by the GPU between GDRST and GuC restart could be zeroed by the next call to xe_guc_ads_populate().
On dGFX, VRAM placement is used so that xe_bo_main_addr() returns a real device-physical address (DPA). The original ADS BO is explicitly migrated to VRAM by xe_guc_realloc_post_hwconfig(), so its xe_bo_main_addr() already returns a correct DPA. A separate SYSTEM BO for the UM queues would not go through that migration path; on systems where physical pages are allocated above 4 GB the IOMMU assigns an IOVA that differs from the physical address, so base_dpa would be wrong and GuC/GAM would access the wrong memory.
Why not just flush the cache after memset?
An earlier approach called drm_clflush_virt_range() on the UM queue region immediately after the memset to evict the stale zero lines from LLC. This reduces the failure rate significantly but does not fully eliminate the race for two reasons.
First, the Linux kernel maintains a WB direct mapping (linear map) that covers all physical memory. Even after flushing the vmap alias, the CPU can speculatively prefetch the same physical page through the WB direct-map alias, silently repopulating the cache with zeros. LNL is known to be particularly susceptible to this speculative prefetch behaviour. XE_BO_FLAG_NEEDS_UC calls set_memory_uc() which updates both the vmap PTE and the direct-map PTE to UC, closing this aliasing window entirely.
Second, the flush approach leaves the memory mapped WB on the CPU side permanently. Any future SW access - a debug read, a tracing hook, an inadvertent touch - can re-dirty the cache line and reintroduce the race. UC mapping makes the correct behaviour unconditional and self-documenting.
Changes in v4:
- Split the single v3 patch into three logically independent patches:
1. Separate BO allocation (root fix for the memset-zeroing race)
2. VRAM placement on dGFX (fix incorrect DPA when pages are above 4 GB)
3. UC mapping (close the CPU cache aliasing window on iGFX)
- Corrected the dGFX description: the original ADS BO works because xe_guc_realloc_post_hwconfig() explicitly migrates it to VRAM; a separate SYSTEM BO would not be migrated and its IOVA would differ from the physical address for high-memory pages. The previous description incorrectly attributed this to IOVA aliasing a PCI MMIO reserved window.
- Added physical-contiguity verification for no-IOMMU iGFX systems.
- xe_guc.c: removed the post-hwconfig reinit_in_vram call for um_queue_bo now that dGFX allocates it directly in VRAM.
Changes in v5:
- Fix the first patch conflict
- Update the ASCII block diagram of ads
- Simplify the contiguous memory check
- Improve code comments
Jia Yao (3):
drm/xe/guc_ads: allocate UM queues in a separate BO drm/xe/guc_ads: allocate UM queues in VRAM on dGFX drm/xe/guc_ads: use uncached mapping for UM queue BO
drivers/gpu/drm/xe/xe_guc_ads.c | 100 +++++++++++++++++++------- drivers/gpu/drm/xe/xe_guc_ads_types.h | 5 ++ 2 files changed, 80 insertions(+), 25 deletions(-)
Signature · 1 line
2.43.0
PATCH v5 1/3 drm/xe/guc_ads: allocate UM queues in a separate BO
xe_guc_ads_populate() calls xe_map_memset() on the ADS BO on every GT reset. The UM queue ring buffers were embedded in that same BO, so the memset discards fault descriptors already written by the GPU, causing GuC to forward an all-zero descriptor (Faulted Address = 0, ASID = 0).
Move the UM queue ring buffers into a dedicated BO so they are not affected by the ADS memset.
On pre-Xe3p platforms GAM accesses the UM queue ring via base_dpa, which must be a contiguous DMA address range. Round the allocation size up to the next power of two so that the TTM pool issues a single order-N alloc_pages() call, maximising the chance of physical contiguity. Verify the result with drm_prime_get_contiguous_size() and fail gracefully if the pages are non-contiguous.
--- drivers/gpu/drm/xe/xe_guc.c | 6 ++ drivers/gpu/drm/xe/xe_guc_ads.c | 96 ++++++++++++++++++--------- drivers/gpu/drm/xe/xe_guc_ads_types.h | 5 ++ 3 files changed, 75 insertions(+), 32 deletions(-)
drivers/gpu/drm/xe/xe_guc.c +6 -0
@@ -732,6 +732,12 @@ static int xe_guc_realloc_post_hwconfig(struct xe_guc *guc)if (ret)return ret; +if (guc->ads.um_queue_bo) { + ret =xe_managed_bo_reinit_in_vram (xe, tile, &guc->ads.um_queue_bo); +if (ret) +return ret; + } +return 0 ; }
drivers/gpu/drm/xe/xe_guc_ads.c +64 -32
149 lines not on this page · raw message
drivers/gpu/drm/xe/xe_guc_ads_types.h +5 -0
@@ -16,6 +16,11 @@ struct xe_bo;struct xe_guc_ads {/** @bo: Xe BO for GuC ads blob */ struct xe_bo *bo; +/** +* @um_queue_bo: Dedicated BO for the HW fault ring (UM queues). +* NULL if the platform does not support USM. +*/ +struct xe_bo *um_queue_bo;/** @golden_lrc_size: golden LRC size */ size_t golden_lrc_size;/** @regset_size: size of register set passed to GuC for save/restore */
Signature · 1 line
2.43.0
Re: [PATCH v5 1/3] drm/xe/guc_ads: allocate UM queues in a separate BO
Quoted 8 lines · Jia Yao
xe_guc_ads_populate()callsxe_map_memset()on the ADS BO on every GT reset. The UM queue ring buffers were embedded in that same BO, so the memset discards fault descriptors already written by the GPU, causing GuC to forward an all-zero descriptor (Faulted Address = 0, ASID = 0).Move the UM queue ring buffers into a dedicated BO so they are not affected by the ADS memset.
Quoted 8 lines
On pre-Xe3p platforms GAM accesses the UM queue ring via
base_dpa, which must be a contiguous DMA address range. Round the allocation size up to the next power of two so that the TTM pool issues a single order-Nalloc_pages()call, maximising the chance of physical contiguity. Verify the result withdrm_prime_get_contiguous_size()and fail gracefully if the pages are non-contiguous.Fixes dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
As per the last revision, I think we can just do:
LNL was the first production part we supported. Also since the last patch will need a recent-ish tag for the UC feature, and I think best have all three commits with the same tag.
Quoted 4 lines
Cc Gwan-gyeong Mun <[email protected]>Cc Matthew Auld <[email protected]>Signed-off-by Jia Yao <[email protected]>
For next time, it's best to also put the changelog accross revisions in each commit, which makes things a bit easier for reviewers.
Quoted 200 lines
--- drivers/gpu/drm/xe/xe_guc.c | 6 ++ drivers/gpu/drm/xe/xe_guc_ads.c | 96 ++++++++++++++++++--------- drivers/gpu/drm/xe/xe_guc_ads_types.h | 5 ++ 3 files changed, 75 insertions(+), 32 deletions(-)drivers/gpu/drm/xe/xe_guc.c +6
-0@@ -732,6 +732,12 @@ static int xe_guc_realloc_post_hwconfig(struct xe_guc *guc)if (ret)return ret; +if (guc->ads.um_queue_bo) { + ret =xe_managed_bo_reinit_in_vram (xe, tile, &guc->ads.um_queue_bo); +if (ret) +return ret; + } +return 0 ; }drivers/gpu/drm/xe/xe_guc_ads.c +64
-32@@ -8,6 +8,7 @@#include <linux/fault-inject.h> #include <drm/drm_managed.h> +#include <drm/drm_prime.h> #include <generated/xe_wa_oob.h> @@ -64,10 +65,14 @@ ads_to_map(struct xe_guc_ads *ads)/* * The Additional Data Struct (ADS) has pointers for different buffers used by -* the GuC. One single gem object contains the ADS struct itself (guc_ads) and -* all the extra buffers indirectly linked via the ADS struct's entries. +* the GuC. One gem object (ads->bo) contains the ADS struct itself (guc_ads) +* and most of the extra buffers linked via the ADS struct's entries. The UM +* fault queues (PAGE_FAULT, PAGE_FAULT_RESPONSE, ACCESS_COUNTER rings) are +* kept in a separate BO (ads->um_queue_bo) so that the full memset of ads->bo +* performed on every GT reset does not discard fault descriptors already +* written into the rings by the GPU. * -* Layout of the ADS blob allocated for the GuC: +* Layout of the ADS blob (ads->bo): * * +---------------------------------------+ <== base * | guc_ads | @@ -99,10 +104,6 @@ ads_to_map(struct xe_guc_ads *ads)* +---------------------------------------+ * | padding | * +---------------------------------------+ <== 4K aligned -* | UM queues | -* +---------------------------------------+ -* | padding | -* +---------------------------------------+ <== 4K aligned * | private data | * +---------------------------------------+ * | padding | @@ -156,16 +157,6 @@ static size_t guc_ads_capture_size(struct xe_guc_ads *ads)return PAGE_ALIGN (ads->capture_size); } -static size_t guc_ads_um_queues_size (struct xe_guc_ads *ads) -{ -struct xe_device *xe =ads_to_xe (ads); - -if (!xe->info.has_usm) -return 0 ; - -return GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX ; -} -static size_t guc_ads_private_data_size (struct xe_guc_ads *ads) {return PAGE_ALIGN (ads_to_guc (ads)->fw.private_data_size);@@ -206,22 +197,12 @@ static size_t guc_ads_capture_offset(struct xe_guc_ads *ads)return PAGE_ALIGN (offset); } -static size_t guc_ads_um_queues_offset (struct xe_guc_ads *ads) -{ -u32 offset; - - offset =guc_ads_capture_offset (ads) + -guc_ads_capture_size (ads); - -return PAGE_ALIGN (offset); -} -static size_t guc_ads_private_data_offset (struct xe_guc_ads *ads) {size_t offset; - offset =guc_ads_um_queues_offset (ads) + -guc_ads_um_queues_size (ads); + offset =guc_ads_capture_offset (ads) + +guc_ads_capture_size (ads);return PAGE_ALIGN (offset); }@@ -461,6 +442,51 @@ int xe_guc_ads_init(struct xe_guc_ads *ads) ads->bo = bo; +if (xe->info.has_usm) { +/* +* Allocate a separate BO for the HW fault ring (UM queues). +* +* xe_guc_ads_populate() clears the entire ADS BO via +* xe_map_memset() on every GT reset. Keeping the UM queues +* in a dedicated BO prevents that memset from discarding fault +* descriptors the GPU has already written into the ring, +* which would cause GuC to forward an all-zero descriptor to +* the driver (Faulted Address = 0, ASID = 0). +* +* Round the size up to the next power of two so that on iGPU +* (system memory, no IOMMU) the TTM pool issues a single +* alloc_pages(order=N) call, maximising the chance of getting +* a physically contiguous block. GuC requires contiguous DPA. +*/ +size_t um_size =roundup_pow_of_two (GUC_UM_QUEUE_SIZE * +GUC_UM_HW_QUEUE_MAX ); + + bo =xe_managed_bo_create_pin_map (xe, tile, um_size, +XE_BO_FLAG_SYSTEM | +XE_BO_FLAG_GGTT | +XE_BO_FLAG_GGTT_INVALIDATE | +XE_BO_FLAG_PINNED_NORESTORE ); +if (IS_ERR (bo)) +return PTR_ERR (bo); + +/* +* On pre-Xe3p platforms, GAM (not GuC) accesses the UM queue +* ring via base_dpa, which must be a contiguous DMA address +* range. Verify that the allocated pages are contiguous in +* DMA address space. +*/ +if (xe_bo_sg (bo) && + !xe_guc_using_main_gamctrl_queues (ads_to_guc (ads)) && +unlikely (drm_prime_get_contiguous_size (xe_bo_sg (bo)) < +GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX )) { +drm_err (&xe->drm, +"UM fault queue memory is not contiguous in DMA address space; GAM requires contiguous DPA\n" ); +return -ENOMEM ; + } + + ads->um_queue_bo = bo; + } +return 0 ; }ALLOW_ERROR_INJECTION (xe_guc_ads_init,ERRNO );/* See xe_pci_probe() */ @@ -893,7 +919,7 @@ static void guc_mmio_reg_state_init(struct xe_guc_ads *ads)static void guc_um_init_params (struct xe_guc_ads *ads) { -u32 um_queue_offset =guc_ads_um_queues_offset (ads); +struct xe_bo *um_bo = ads->um_queue_bo;struct xe_guc *guc =ads_to_guc (ads);struct xe_device *xe =ads_to_xe (ads);u64 base_dpa;@@ -903,8 +929,14 @@ static void guc_um_init_params(struct xe_guc_ads *ads) with_dpa = !xe_guc_using_main_gamctrl_queues (guc); - base_ggtt =xe_bo_ggtt_addr (ads->bo) + um_queue_offset; - base_dpa =xe_bo_main_addr (ads->bo,PAGE_SIZE ) + um_queue_offset; +if (um_bo) { +/* All USM platforms: UM queues in dedicated um_queue_bo */ + base_ggtt =xe_bo_ggtt_addr (um_bo); + base_dpa =xe_bo_main_addr (um_bo,PAGE_SIZE ); + }else { +/* Platform does not support USM: no UM queues, nothing to do */ +return ; + }for (i =0 ; i <GUC_UM_HW_QUEUE_MAX ; ++i) {/* drivers/gpu/drm/xe/xe_guc_ads_types.h +5
-0@@ -16,6 +16,11 @@ struct xe_bo;struct xe_guc_ads {/** @bo: Xe BO for GuC ads blob */ struct xe_bo *bo; +/** +* @um_queue_bo: Dedicated BO for the HW fault ring (UM queues). +* NULL if the platform does not support USM. +*/ +struct xe_bo *um_queue_bo;/** @golden_lrc_size: golden LRC size */ size_t golden_lrc_size;/** @regset_size: size of register set passed to GuC for save/restore */
PATCH v5 2/3 drm/xe/guc_ads: allocate UM queues in VRAM on dGFX
In reply to Jia YaoOn pre-Xe3p dGFX, GuC/GAM accesses the UM queue ring buffer via base_dpa, which is interpreted as a raw device-physical address. Unlike the GGTT PTE, there is no flag to indicate whether base_dpa points to VRAM or system memory; the hardware assumes it is always a VRAM address.
A separate SYSTEM BO for the UM queues would cause xe_bo_main_addr() to return an IOMMU IOVA instead of a true VRAM device-physical address. GuC then accesses the wrong memory and forwards all-zero fault descriptors to the driver (Faulted Address = 0, ASID = 0).
Use XE_BO_FLAG_VRAM_IF_DGFX() so that on dGFX the UM queue BO is placed directly in VRAM and xe_bo_main_addr() returns the correct device-physical address.
--- drivers/gpu/drm/xe/xe_guc.c | 6 ------ drivers/gpu/drm/xe/xe_guc_ads.c | 11 ++++++----- 2 files changed, 6 insertions(+), 11 deletions(-)
drivers/gpu/drm/xe/xe_guc.c +0 -6
@@ -732,12 +732,6 @@ static int xe_guc_realloc_post_hwconfig(struct xe_guc *guc)if (ret)return ret; -if (guc->ads.um_queue_bo) { - ret =xe_managed_bo_reinit_in_vram (xe, tile, &guc->ads.um_queue_bo); -if (ret) -return ret; - } -return 0 ; }
drivers/gpu/drm/xe/xe_guc_ads.c +6 -5
@@ -461,11 +461,12 @@ int xe_guc_ads_init(struct xe_guc_ads *ads)size_t um_size =roundup_pow_of_two (GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX ); - bo =xe_managed_bo_create_pin_map (xe, tile, um_size, -XE_BO_FLAG_SYSTEM | -XE_BO_FLAG_GGTT | -XE_BO_FLAG_GGTT_INVALIDATE | -XE_BO_FLAG_PINNED_NORESTORE ); +u32 um_flags =XE_BO_FLAG_VRAM_IF_DGFX (tile) | +XE_BO_FLAG_GGTT | +XE_BO_FLAG_GGTT_INVALIDATE | +XE_BO_FLAG_PINNED_NORESTORE ; + + bo =xe_managed_bo_create_pin_map (xe, tile, um_size, um_flags);if (IS_ERR (bo))return PTR_ERR (bo);
Signature · 1 line
2.43.0
Re: [PATCH v5 2/3] drm/xe/guc_ads: allocate UM queues in VRAM on dGFX
Quoted 11 lines · Jia Yao
On pre-Xe3p dGFX, GuC/GAM accesses the UM queue ring buffer via
base_dpa, which is interpreted as a raw device-physical address. Unlike the GGTT PTE, there is no flag to indicate whetherbase_dpapoints to VRAM or system memory; the hardware assumes it is always a VRAM address.A separate SYSTEM BO for the UM queues would cause
xe_bo_main_addr()to return an IOMMU IOVA instead of a true VRAM device-physical address. GuC then accesses the wrong memory and forwards all-zero fault descriptors to the driver (Faulted Address = 0, ASID = 0).
Quoted 5 lines
Use
XE_BO_FLAG_VRAM_IF_DGFX()so that on dGFX the UM queue BO is placed directly in VRAM andxe_bo_main_addr()returns the correct device-physical address.Fixes dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
Quoted 31 lines
Cc Gwan-gyeong Mun <[email protected]>Cc Matthew Auld <[email protected]>Signed-off-by Jia Yao <[email protected]>--- drivers/gpu/drm/xe/xe_guc.c | 6 ------ drivers/gpu/drm/xe/xe_guc_ads.c | 11 ++++++----- 2 files changed, 6 insertions(+), 11 deletions(-)drivers/gpu/drm/xe/xe_guc.c +0
-6@@ -732,12 +732,6 @@ static int xe_guc_realloc_post_hwconfig(struct xe_guc *guc)if (ret)return ret; -if (guc->ads.um_queue_bo) { - ret =xe_managed_bo_reinit_in_vram (xe, tile, &guc->ads.um_queue_bo); -if (ret) -return ret; - } -return 0 ; }drivers/gpu/drm/xe/xe_guc_ads.c +0
-0
drivers/gpu/drm/xe/xe_guc_ads.c int xe_guc_ads_init(struct xe_guc_ads *ads) · line 462
Quoted 2 lines
drivers/gpu/drm/xe/xe_guc_ads.c +0
-0@@ -461,11 +461,12 @@ int xe_guc_ads_init(struct xe_guc_ads *ads)size_t um_size =roundup_pow_of_two (GUC_UM_QUEUE_SIZE *GUC_UM_HW_QUEUE_MAX );
If you want we could also make the roundup conditional on igpu, to save a small amount of VRAM.
Either way,
Quoted 15 lines
drivers/gpu/drm/xe/xe_guc_ads.c +6
-5- bo =xe_managed_bo_create_pin_map (xe, tile, um_size, -XE_BO_FLAG_SYSTEM | -XE_BO_FLAG_GGTT | -XE_BO_FLAG_GGTT_INVALIDATE | -XE_BO_FLAG_PINNED_NORESTORE ); +u32 um_flags =XE_BO_FLAG_VRAM_IF_DGFX (tile) | +XE_BO_FLAG_GGTT | +XE_BO_FLAG_GGTT_INVALIDATE | +XE_BO_FLAG_PINNED_NORESTORE ; + + bo =xe_managed_bo_create_pin_map (xe, tile, um_size, um_flags);if (IS_ERR (bo))return PTR_ERR (bo);
PATCH v5 3/3 drm/xe/guc_ads: use uncached mapping for UM queue BO
In reply to Jia YaoThe HW unit that writes fault descriptors into the UM queue ring buffer may not be coherent with the CPU cache. If a descriptor write is still sitting in a GPU L3 cache line when GuC reads the slot, GuC sees the stale (all-zero) content from DRAM and forwards an invalid descriptor to the driver (Faulted Address = 0, ASID = 0).
Mark the UM queue BO as uncached (XE_BO_FLAG_NEEDS_UC) so that writes bypass the cache and are immediately visible to GuC.
--- drivers/gpu/drm/xe/xe_guc_ads.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/xe/xe_guc_ads.c +2 -1
@@ -464,7 +464,8 @@ int xe_guc_ads_init(struct xe_guc_ads *ads)u32 um_flags =XE_BO_FLAG_VRAM_IF_DGFX (tile) |XE_BO_FLAG_GGTT |XE_BO_FLAG_GGTT_INVALIDATE | -XE_BO_FLAG_PINNED_NORESTORE ; +XE_BO_FLAG_PINNED_NORESTORE | +XE_BO_FLAG_NEEDS_UC ; bo =xe_managed_bo_create_pin_map (xe, tile, um_size, um_flags);if (IS_ERR (bo))
Signature · 1 line
2.43.0
Re: [PATCH v5 3/3] drm/xe/guc_ads: use uncached mapping for UM queue BO
Quoted 9 lines · Jia Yao
The HW unit that writes fault descriptors into the UM queue ring buffer may not be coherent with the CPU cache. If a descriptor write is still sitting in a GPU L3 cache line when GuC reads the slot, GuC sees the stale (all-zero) content from DRAM and forwards an invalid descriptor to the driver (Faulted Address = 0, ASID = 0).
Mark the UM queue BO as uncached (
XE_BO_FLAG_NEEDS_UC) so that writes bypass the cache and are immediately visible to GuC.
Did you mean to drop the fixes tag here?
Quoted 3 lines
Cc Gwan-gyeong Mun <[email protected]>Cc Matthew Auld <[email protected]>Signed-off-by Jia Yao <[email protected]>