mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation
Subsystems (10)
Matthew Brost
maintainerv4 cover mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation
TTM allocations at higher orders can drive Xe into a pathological reclaim loop when memory is fragmented:
kswapd → shrinker → eviction → rebind (exec ioctl) → repeat
In this state, reclaim is triggered despite substantial free memory, but fails to produce contiguous higher-order pages. The Xe shrinker then evicts active buffer objects, increasing faulting and rebind activity and further feeding the loop. The result is high CPU overhead and poor GPU forward progress.
This issue was first reported in [1] and independently observed internally and by Google.
A simple reproducer is:
- Boot an iGPU system with mem=8G
- Launch 10 Chrome tabs running the WebGL aquarium demo
- Configure each tab with ~5k fish
Under this workload, ftrace shows a continuous loop of:
xe_shrinker_scan (kswapd)xe_vma_rebind_exec
Performance degrades significantly, with each tab dropping to ~2 FPS on PTL (Ubuntu 24.04).
At the same time, /proc/buddyinfo shows substantial free memory but no higher-order availability. For example, the Normal zone:
Count: 4063 4595 3455 3400 3139 2762 2293 1655 643 0 0
This corresponds to ~2.8GB free memory, but no order-9 (2MB) blocks, indicating severe fragmentation.
This series addresses the issue in two ways:
TTM: Restrict direct reclaim to beneficial_order. Larger allocations use __GFP_NORETRY to fail quickly rather than triggering reclaim.
Xe: Introduce a heuristic in the shrinker to avoid eviction when running under kswapd and the system appears memory-rich but fragmented.
With these changes, the reclaim/eviction loop is eliminated. The same workload improves to ~10 FPS per tab (Ubuntu 24.04) or ~15 FPS per tab (Ubuntu 24.10), and kswapd activity subsides.
Buddyinfo after applying this series shows restored higher-order availability:
Count: 8526 7067 3092 1959 1292 660 194 28 20 13 1
Matt
v2:
- Layer with core MM / TTM helpers (Thomas)
v4:
- Fix build (CI)
[1] https://patchwork.freedesktop.org/patch/716404/?series=164353&rev=1
Matthew Brost (6):
mm: Wire up order in shrink_control
mm: Introduce zone_maybe_fragmented_in_shrinker() drm/ttm: Issue direct reclaim at beneficial_order drm/ttm: Introduce ttm_bo_shrink_kswap_maybe_fragmented() drm/xe: Set TTM device beneficial_order to 9 (2M) drm/xe: Avoid shrinker reclaim from kswapd under fragmentation
drivers/gpu/drm/ttm/ttm_bo_util.c | 38 +++++++++++++++++++++++++++++++ drivers/gpu/drm/ttm/ttm_pool.c | 4 ++-- drivers/gpu/drm/xe/xe_device.c | 3 ++- drivers/gpu/drm/xe/xe_shrinker.c | 3 +++ include/drm/ttm/ttm_bo.h | 2 ++ include/linux/shrinker.h | 3 +++ include/linux/vmstat.h | 12 ++++++++++ mm/internal.h | 4 ++-- mm/shrinker.c | 13 +++++++---- mm/vmscan.c | 7 +++--- 10 files changed, 76 insertions(+), 13 deletions(-)
Signature · 1 line
2.34.1
Matthew Brost
maintainerPATCH v4 1/6 mm: Wire up order in shrink_control
Pass the allocation order through shrink_control so shrinkers have visibility into the order that triggered reclaim.
This allows shrinkers to implement better heuristics, such as detecting high-order allocation pressure or fragmentation and avoiding eviction of working sets when reclaim is invoked from kswapd.
---
v4: Fix build without CONFIG_MEMCG (CI)
--- include/linux/shrinker.h | 3 +++ mm/internal.h | 4 ++-- mm/shrinker.c | 13 ++++++++----- mm/vmscan.c | 7 ++++--- 4 files changed, 17 insertions(+), 10 deletions(-)
include/linux/shrinker.h +3 -0
@@ -37,6 +37,9 @@ struct shrink_control {/* current node being shrunk (for NUMA aware shrinkers) */ int nid; +/* Allocation order we are currently trying to fulfil. */ +s8 order; +/* * How many objects scan_objects should scan and try to reclaim. * This is reset before every call, so it is safe for callees
mm/internal.h +2 -2
@@ -1759,8 +1759,8 @@ void __meminit __init_single_page(struct page *page, unsigned long pfn,void __meminit __init_page_from_nid (unsigned long pfn,int nid);/* shrinker related functions */ -unsigned long shrink_slab (gfp_t gfp_mask,int nid,struct mem_cgroup *memcg, -int priority); +unsigned long shrink_slab (gfp_t gfp_mask,int nid,s8 order, +struct mem_cgroup *memcg,int priority); int shmem_add_to_page_cache(struct folio *folio, struct address_space *mapping,
mm/shrinker.c +8 -5
@@ -466,7 +466,7 @@ static unsigned long do_shrink_slab(struct shrink_control *shrinkctl, }#ifdef CONFIG_MEMCG -static unsigned long shrink_slab_memcg (gfp_t gfp_mask,int nid, +static unsigned long shrink_slab_memcg (gfp_t gfp_mask,int nid,s8 order,struct mem_cgroup *memcg,int priority) {struct shrinker_info *info;
@@ -528,6 +528,7 @@ static unsigned long shrink_slab_memcg(gfp_t gfp_mask, int nid,struct shrink_control sc = { .gfp_mask = gfp_mask, .nid = nid, + .order = order, .memcg = memcg, };struct shrinker *shrinker;
@@ -587,7 +588,7 @@ static unsigned long shrink_slab_memcg(gfp_t gfp_mask, int nid, return freed; }#else /* !CONFIG_MEMCG */ -static unsigned long shrink_slab_memcg (gfp_t gfp_mask,int nid, +static unsigned long shrink_slab_memcg (gfp_t gfp_mask,int nid,s8 order,struct mem_cgroup *memcg,int priority) {return 0 ;
@@ -598,6 +599,7 @@ static unsigned long shrink_slab_memcg(gfp_t gfp_mask, int nid,* shrink_slab - shrink slab caches * @gfp_mask: allocation context * @nid: node whose slab caches to target +* @order: order of allocation * @memcg: memory cgroup whose slab caches to target * @priority: the reclaim priority *
@@ -614,8 +616,8 @@ static unsigned long shrink_slab_memcg(gfp_t gfp_mask, int nid,* * Returns the number of reclaimed slab objects. */ -unsigned long shrink_slab (gfp_t gfp_mask,int nid,struct mem_cgroup *memcg, -int priority) +unsigned long shrink_slab (gfp_t gfp_mask,int nid,s8 order, +struct mem_cgroup *memcg,int priority) {unsigned long ret, freed =0 ;struct shrinker *shrinker;
@@ -628,7 +630,7 @@ unsigned long shrink_slab(gfp_t gfp_mask, int nid, struct mem_cgroup *memcg,* oom. */ if (!mem_cgroup_disabled () && !mem_cgroup_is_root (memcg)) -return shrink_slab_memcg (gfp_mask, nid, memcg, priority); +return shrink_slab_memcg (gfp_mask, nid, order, memcg, priority);/* * lockless algorithm of global shrink.
@@ -656,6 +658,7 @@ unsigned long shrink_slab(gfp_t gfp_mask, int nid, struct mem_cgroup *memcg,struct shrink_control sc = { .gfp_mask = gfp_mask, .nid = nid, + .order = order, .memcg = memcg, };
mm/vmscan.c +4 -3
@@ -412,7 +412,7 @@ static unsigned long drop_slab_node(int nid) memcg =mem_cgroup_iter (NULL ,NULL ,NULL );do { - freed +=shrink_slab (GFP_KERNEL , nid, memcg,0 ); + freed +=shrink_slab (GFP_KERNEL , nid,0 , memcg,0 ); }while ((memcg =mem_cgroup_iter (NULL , memcg,NULL )) !=NULL );return freed;
@@ -5068,7 +5068,8 @@ static int shrink_one(struct lruvec *lruvec, struct scan_control *sc) success =try_to_shrink_lruvec (lruvec, sc); -shrink_slab (sc->gfp_mask, pgdat->node_id, memcg, sc->priority); +shrink_slab (sc->gfp_mask, pgdat->node_id, sc->order, memcg, + sc->priority);if (!sc->proactive) vmpressure(sc->gfp_mask, memcg, false, sc->nr_scanned - scanned,
@@ -6170,7 +6171,7 @@ static void shrink_node_memcgs(pg_data_t *pgdat, struct scan_control *sc)shrink_lruvec (lruvec, sc); -shrink_slab (sc->gfp_mask, pgdat->node_id, memcg, +shrink_slab (sc->gfp_mask, pgdat->node_id, sc->order, memcg, sc->priority);/* Record the group's reclaim efficiency */
Signature · 1 line
2.34.1
Matthew Brost
maintainerPATCH v4 2/6 mm: Introduce zone_maybe_fragmented_in_shrinker()
In reply to Matthew BrostIntroduce zone_maybe_fragmented_in_shrinker() as a lightweight helper to allow subsystems to make coarse decisions about reclaim behavior in the presence of likely fragmentation.
The helper implements a simple heuristic: if the number of free pages in a zone exceeds twice the high watermark, the zone is considered to have ample free memory and allocation failures are more likely due to fragmentation than overall memory pressure.
This is intentionally imprecise and is not meant to replace the core MM compaction or fragmentation accounting logic. Instead, it provides a cheap signal for callers (e.g., shrinkers) that wish to avoid overly aggressive reclaim when sufficient free memory exists but high-order allocations may still fail.
No functional changes; this is a preparatory helper for future users.
---
v3: s/zone_appear_fragmented/zone_maybe_fragmented_in_shrinker (David
Hildenbrand)
--- include/linux/vmstat.h | 12 ++++++++++++ 1 file changed, 12 insertions(+)
include/linux/vmstat.h +12 -0
@@ -483,6 +483,18 @@ static inline const char *zone_stat_name(enum zone_stat_item item)return vmstat_text[item]; } +static inline bool zone_maybe_fragmented_in_shrinker (struct zone *zone) +{ +/* +* Simple heuristic: if the number of free pages is more than twice the +* high watermark, this may suggest that the zone is heavily fragmented. +* When called from a shrinker, aggressively evicting memory in this case +* may do more harm to overall system performance than good. +*/ +return zone_page_state (zone,NR_FREE_PAGES ) > +high_wmark_pages (zone) *2 ; +} +#ifdef CONFIG_NUMA static inline const char *numa_stat_name (enum numa_stat_item item) {
Signature · 1 line
2.34.1
Re: [PATCH v4 2/6] mm: Introduce zone_maybe_fragmented_in_shrinker()
Quoted 48 lines · Matthew Brost
Introduce
zone_maybe_fragmented_in_shrinker()as a lightweight helper to
allow subsystems to make coarse decisions about reclaim behavior in the
presence of likely fragmentation.The helper implements a simple heuristic: if the number of free pages in a zone exceeds twice the high watermark, the zone is considered to have ample free memory and allocation failures are more likely due to fragmentation than overall memory pressure.
This is intentionally imprecise and is not meant to replace the core MM compaction or fragmentation accounting logic. Instead, it provides a cheap signal for callers (e.g., shrinkers) that wish to avoid overly aggressive reclaim when sufficient free memory exists but high-order allocations may still fail.
No functional changes; this is a preparatory helper for future users.
Cc Thomas Hellström <[email protected]>Cc Andrew Morton <[email protected]>Cc David Hildenbrand <[email protected]>Cc Lorenzo Stoakes <[email protected]>Cc "Liam R. Howlett" <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Mike Rapoport <[email protected]>Cc Suren Baghdasaryan <[email protected]>Cc Michal Hocko <[email protected]>Signed-off-by Matthew Brost <[email protected]>---
v3: s/zone_appear_fragmented/zone_maybe_fragmented_in_shrinker (David Hildenbrand)
---
include/linux/vmstat.h | 12 ++++++++++++1 file changed, 12 insertions(+)include/linux/vmstat.h +0
-0@@ -483,6 +483,18 @@ static inline const char *zone_stat_name(enum zone_stat_item item) return vmstat_text[item]; }
Quoted 1 line
on the below heuristic, I was thinking of the following case: a large memory system (say 16G, 32G), heavily fragmented (for whatever reason) but constraint by the IOMMU requiring large pages due to hw alignment, if I am not mistaken the below check will cause the shrinker to bail out too 'early' since the there's plenty of available memory but none of that is contiguous, then end result should be giving back small pages which should reduce performance, right?
below are some made up numbers:
Metric | 8GB | 16GB ----------------|-------------------|------------------- High Wmark | ~45MB (11k pgs) | ~90MB (23k pgs) Bail Gate (2x) | ~90MB (22k pgs) | ~180MB (46k pgs) Free RAM | 120MB | 7100MB (7.1GB) Shrinker | RUNS (Free<Gate) | BAILS (Free>Gate) Outcome | Merges 2MB blocks | 4KB pages
In other words, replacing the check with numbers:
System | Free RAM (Pages) | Gate (Pages) | Free < Gate? | Result -------------|------------------|--------------|--------------|------- 8GB | 20,480 (80MB) | 22,946 | 20480 < 22946| RUNS 16GB | 1,832,740 (7.1G) | 45,894 | 1.8M < 45k? | BAILS
Carlos
Quoted 19 lines
+static inline bool
zone_maybe_fragmented_in_shrinker(struct zone *zone)include/linux/vmstat.h +3
-0+{ +/* +* Simple heuristic: if the number of free pages is more than twice the
include/linux/vmstat.h +1
-0+* high watermark, this may suggest that the zone is heavily fragmented.
include/linux/vmstat.h +1
-0+* When called from a shrinker, aggressively evicting memory in this case
include/linux/vmstat.h +6
-0+* may do more harm to overall system performance than good. +*/ +return zone_page_state (zone,NR_FREE_PAGES ) > +high_wmark_pages (zone) *2 ; +} +#ifdef
CONFIG_NUMA
static inline const char *numa_stat_name(enumnuma_stat_itemitem) {
Re: PATCH v4 0/6] mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation
In reply to Matthew BrostQuoted 3 lines · Matthew Brost
Introduce
zone_maybe_fragmented_in_shrinker()as a lightweight helper to allow subsystems to make coarse decisions about reclaim behavior in the presence of likely fragmentation
I'm running Linus' master on my LunarLake (258v) laptop, and sometimes after compiling a kernel (of all things) I'd see kswapd0 thrash despite having quite a bit of free memory.
I finally traced it to the xe driver after seeing the "GPUActive" field in /proc/meminfo suddenly start rising, eventually growing larger than real memory by several times (see below).
This patchset fixes the issue, and I'm sure there'll be a fix going into Linus' master soon, but what I'M wondering is how could building a kernel (which is just in a KDE Konsole running on Wayland) make the GPActive grow from ~1.6G to > 30G (and continue to rise, RN I'm seeing 91839848 kBs and still growing).
-Kenny
----
SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 652640 kB GPUReclaim: 403988 kB
SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 651180 kB GPUReclaim: 406812 kB
SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 659004 kB GPUReclaim: 399396 kB
SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 666996 kB GPUReclaim: 392764 kB
<some hours later>
GPUActive: 91832468 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 488000 kB
GPUActive: 91832332 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 487988 kB
GPUActive: 91869376 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 486504 kB
----
Signature · 1 line
Kenneth R. Crudup / Sr. SW Engineer, Scott County Consulting, Orange County CA
Matthew Brost
maintainerRe: PATCH v4 0/6] mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation
Quoted 8 lines · Kenneth Crudup
Quoted 3 lines · Matthew Brost
Introduce
zone_maybe_fragmented_in_shrinker()as a lightweight helper to allow subsystems to make coarse decisions about reclaim behavior in the presence of likely fragmentationI'm running Linus' master on my LunarLake (258v) laptop, and sometimes after
+Dave
So is this 7.1-rc1? It looks like new feature to 7.1 added by Dave [1] and something look off here. Thanks for pointing this out.
I'm grabbing a machine now to see if I can recreate this...
Matt
[1] git format-patch -1 2232ba9c7931d
Quoted 57 lines
compiling a kernel (of all things) I'd see kswapd0 thrash despite having quite a bit of free memory.
I finally traced it to the xe driver after seeing the "GPUActive" field in
/proc/meminfosuddenly start rising, eventually growing larger than real memory by several times (see below).This patchset fixes the issue, and I'm sure there'll be a fix going into Linus' master soon, but what I'M wondering is how could building a kernel (which is just in a KDE Konsole running on Wayland) make the GPActive grow from ~1.6G to > 30G (and continue to rise, RN I'm seeing 91839848 kBs and still growing).
-Kenny
----
SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 652640 kB GPUReclaim: 403988 kBSwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 651180 kB GPUReclaim: 406812 kBSwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 659004 kB GPUReclaim: 399396 kBSwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 666996 kB GPUReclaim: 392764 kB<some hours later>
GPUActive: 91832468 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 488000 kBGPUActive: 91832332 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 487988 kBGPUActive: 91869376 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 486504 kB----
Signature · 2 lines
Kenneth R. Crudup / Sr. SW Engineer, Scott County Consulting, Orange County CA
Re: PATCH v4 0/6] mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation
Quoted 2 lines · Matthew Brost
So is this 7.1-rc1? It looks like new feature to 7.1 added by Dave [1] and something look off here. Thanks for pointing this out.
Yeah. I grab his master branch daily (as of 6fe0be6dc7fa RN).
Is this a "shoot the messenger" thing? IOW, is the reporting off, or is the memory usage really that high?
(BTW, those are in 30-second intervals)
Quoted 32 lines
Quoted 32 lines
----
SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 652640 kB GPUReclaim: 403988 kBSwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 651180 kB GPUReclaim: 406812 kBSwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 659004 kB GPUReclaim: 399396 kBSwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 666996 kB GPUReclaim: 392764 kB<some hours later>
GPUActive: 91832468 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 488000 kBGPUActive: 91832332 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 487988 kB
Quoted 5 lines
Quoted 5 lines
GPUActive: 91869376 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 486504 kB----
-K
Signature · 1 line
Kenneth R. Crudup / Sr. SW Engineer, Scott County Consulting, Orange County CA
Matthew Brost
maintainerRe: PATCH v4 0/6] mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation
Quoted 9 lines · Kenneth Crudup
Quoted 2 lines · Matthew Brost
So is this 7.1-rc1? It looks like new feature to 7.1 added by Dave [1] and something look off here. Thanks for pointing this out.
Yeah. I grab his master branch daily (as of
6fe0be6dc7faRN).Is this a "shoot the messenger" thing? IOW, is the reporting off, or is the
I don't think I'm firing any shots.
Quoted 1 line
memory usage really that high?
I've been able to recreate this. It looks like accounting is correct until the Xe shrinker runs - every time it kicks in GPUActive grows and will not reduce past some new floor value. It looks like an accounting bug in TTM or Xe (?).
Here is my output on a 8G PTL where I have intentionally triggered shrinker to evict at least 23875 BOs (most likey quite few more but this what I easily see in dmesg) after closing everything on desktop.
cat /proc/meminfo | grep GPU; cat /proc/buddyinfo;
GPUActive: 13100036 kB GPUReclaim: 152 kB Node 0, zone DMA 0 1 0 0 0 0 0 0 1 1 3 Node 0, zone DMA32 2320 1882 1523 1238 980 740 482 275 114 88 205 Node 0, zone Normal 9751 9343 6466 4237 2703 1162 805 420 191 145 289
Let me spend a bit of time here to see if I figure out where the accounting goes wrong.
Matt
Quoted 47 lines
(BTW, those are in 30-second intervals)
Quoted 32 lines
Quoted 32 lines
----
SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 652640 kB GPUReclaim: 403988 kBSwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 651180 kB GPUReclaim: 406812 kBSwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 659004 kB GPUReclaim: 399396 kBSwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 666996 kB GPUReclaim: 392764 kB<some hours later>
GPUActive: 91832468 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 488000 kBGPUActive: 91832332 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 487988 kBQuoted 5 lines
Quoted 5 lines
GPUActive: 91869376 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 486504 kB----
-K
Signature · 2 lines
Kenneth R. Crudup / Sr. SW Engineer, Scott County Consulting, Orange County CA
Matthew Brost
maintainerRe: PATCH v4 0/6] mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation
Quoted 25 lines · Matthew Brost
Quoted 9 lines · Kenneth Crudup
Quoted 2 lines · Matthew Brost
So is this 7.1-rc1? It looks like new feature to 7.1 added by Dave [1] and something look off here. Thanks for pointing this out.
Yeah. I grab his master branch daily (as of
6fe0be6dc7faRN).Is this a "shoot the messenger" thing? IOW, is the reporting off, or is the
I don't think I'm firing any shots.
Quoted 1 line
memory usage really that high?
I've been able to recreate this. It looks like accounting is correct until the Xe shrinker runs - every time it kicks in GPUActive grows and will not reduce past some new floor value. It looks like an accounting bug in TTM or Xe (?).
Here is my output on a 8G PTL where I have intentionally triggered shrinker to evict at least 23875 BOs (most likey quite few more but this what I easily see in dmesg) after closing everything on desktop.
cat
/proc/meminfo| grep GPU; cat/proc/buddyinfo;
Quoted 9 lines
GPUActive: 13100036 kB GPUReclaim: 152 kB Node 0, zone DMA 0 1 0 0 0 0 0 0 1 1 3 Node 0, zone DMA32 2320 1882 1523 1238 980 740 482 275 114 88 205 Node 0, zone Normal 9751 9343 6466 4237 2703 1162 805 420 191 145 289Let me spend a bit of time here to see if I figure out where the accounting goes wrong.
Looks like a simple accounting error in the shrinking path. Here is a fix
[1] that seems to work for me.
If you want to give a it try, that would be helpful.
Matt
[1] https://patchwork.freedesktop.org/series/165862/
Quoted 49 lines
Matt
Quoted 47 lines
(BTW, those are in 30-second intervals)
Quoted 32 lines
Quoted 32 lines
----
SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 652640 kB GPUReclaim: 403988 kBSwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 651180 kB GPUReclaim: 406812 kBSwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 659004 kB GPUReclaim: 399396 kBSwapTotal: 33554428 kB MemTotal: 32345672 kB GPUActive: 666996 kB GPUReclaim: 392764 kB<some hours later>
GPUActive: 91832468 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 488000 kBGPUActive: 91832332 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 487988 kBQuoted 5 lines
Quoted 5 lines
GPUActive: 91869376 kB SwapTotal: 33554428 kB MemTotal: 32345672 kB GPUReclaim: 486504 kB----
-K
Signature · 2 lines
Kenneth R. Crudup / Sr. SW Engineer, Scott County Consulting, Orange County CA
Re: PATCH v4 0/6] mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation
Quoted 2 lines · Matthew Brost
Looks like a simple accounting error in the shrinking path. Here is a fix
[1] that seems to work for me.
NGL, I lucked out with this error; I wouldn't have tracked down your patch series if I hadn't started looking at the Xe driver 'cause of that huge GPUActive value, rightly or wrongly- I was about to begin the potentially-painful process of un-merging the last DRM/Xe series from Linus' master (the thrashing doesn't happen in 7.0).
-Kenny "So ... thanks, Dave! :)" Crudup
Signature · 1 line
Kenneth R. Crudup / Sr. SW Engineer, Scott County Consulting, Orange County CA
Re: PATCH v4 0/6] mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation
In reply to Kenneth CrudupPerhaps this issue actually lies elsewhere?
I've been working with this patchset for a few days now and while it does seem to fix my "runaway kswapd" issue, every now and then I'd get a hard panic while seeking in large-format video files (using the "mpv" video player, FWIW). Suspecting it had to do with the shrinker path in this patch, I'd started looking at ways to burn thru memory to verify that hypothesis and found the following one-liner:
echo {1..1000000000}
It works perfectly- eats thru all it can, until it gets OOMed. Thing was, after this would run, kswapd would go runaway again, even if this was run on a freshly-booted system (I'd also have a few Chrome tabs open).
As I run Linus' master on my LunarLake box I tried running this again on the stock ((K)ubuntu 7.0.0-15-generic) kernel, and once the oom_reaper killed off the bash, the system settled down to normal again with no runaway kswapd as before. So I recompiled back to 7.0 with my preferred config options (minimal modules, a bigger log_buf, that kind of thing) and the runaway kswapd issue persisted after the bash was OOMed.
I then built 7.0.0 again, using the basic kernel.org source version (as I know sometimes Ubuntu special-sauces their kernels) BUT using the config from Ubuntu's stock generic 7.0, and huh- the problem didn't occur. I went all the way to 7.0.3 with the same result, so the problem had to have been in the config options I'd used.
It took some doing and some trial-and-error but I was able to generate a custom defconfig for my machine that has the right config options so that now after the bash gets killed off, kswapd quiets down. I've attached a diff of the savedefconfig'ed defconfig for reference (So is it CMA? PSI? CGROUP? ?? ... maybe someone here can tell me which option made the difference, I'm kind of over fine-graining it any longer :) )
But now here's the thing- since fixing the defconfig options, I'm now back to running Linus' master kernel (which has the drm-next Xe memory changes/fixes that presumably warrant the patchset in this thread) and no matter what I do to reproduce the reclaim/eviction loops (and I could easily before in the 7.1-rcX series- making a kernel, running bonnie++ , the bash one-liner above), I haven't been able to, and WITHOUT this patch.
-Kenny
Quoted 14 lines · Kenneth Crudup
Quoted 3 lines · Matthew Brost
Introduce
zone_maybe_fragmented_in_shrinker()as a lightweight helper to allow subsystems to make coarse decisions about reclaim behavior in the presence of likely fragmentationI'm running Linus' master on my LunarLake (258v) laptop, and sometimes after compiling a kernel (of all things) I'd see kswapd0 thrash despite having quite a bit of free memory.
I finally traced it to the xe driver after seeing the "GPUActive" field in
/proc/meminfosuddenly start rising, eventually growing larger than real memory by several times (see below).This patchset fixes the issue, and I'm sure there'll be a fix going into Linus' master soon, but what I'M wondering is how could building a kernel (which is just in a KDE Konsole running on Wayland) make the GPActive grow from ~1.6G to > 30G (and continue to rise, RN I'm seeing 91839848 kBs and still growing).
-Kenny
Signature · 1 line
Kenneth R. Crudup / Sr. SW Engineer, Scott County Consulting, Orange County CA
From 8da53fa9d1741a3fcfd9a54fae9ca5cdcb5c0d0c Mon Sep 17 00:00:00 2001
From: "Kenneth R. Crudup" <[email protected]>
Date: Thu, 7 May 2026 03:08:29 -0700
Subject: [PATCH] x86: Yoga9i: Fix memory-pressure issues, refine defconfig
further
--- arch/x86/configs/yoga9i_defconfig | 91 +++++++++++++------------------ 1 file changed, 38 insertions(+), 53 deletions(-)
arch/x86/configs/yoga9i_defconfig +38 -53
@@ -1,13 +1,15 @@ CONFIG_LOCALVERSION="-kenny" # CONFIG_LOCALVERSION_AUTO is not set CONFIG_KERNEL_ZSTD=y -CONFIG_DEFAULT_HOSTNAME="xps-9320" +CONFIG_DEFAULT_HOSTNAME="yoga-9i" CONFIG_SYSVIPC=y CONFIG_POSIX_MQUEUE=y CONFIG_WATCH_QUEUE=y CONFIG_AUDIT=y +CONFIG_NO_HZ_FULL=y CONFIG_NO_HZ=y CONFIG_HIGH_RES_TIMERS=y +CONFIG_POSIX_AUX_CLOCKS=y CONFIG_BPF_SYSCALL=y CONFIG_BPF_JIT=y CONFIG_BPF_JIT_ALWAYS_ON=y
@@ -20,6 +22,7 @@ CONFIG_TASKSTATS=y CONFIG_TASK_DELAY_ACCT=y CONFIG_TASK_XACCT=y CONFIG_TASK_IO_ACCOUNTING=y +CONFIG_PSI=y CONFIG_IKCONFIG=y CONFIG_IKCONFIG_PROC=y CONFIG_IKHEADERS=y
@@ -31,6 +34,7 @@ CONFIG_CFS_BANDWIDTH=y CONFIG_UCLAMP_TASK_GROUP=y CONFIG_CGROUP_PIDS=y CONFIG_CGROUP_RDMA=y +CONFIG_CGROUP_DMEM=y CONFIG_CGROUP_FREEZER=y CONFIG_CGROUP_HUGETLB=y CONFIG_CPUSETS=y
@@ -75,12 +79,12 @@ CONFIG_EFI=y CONFIG_EFI_STUB=y CONFIG_EFI_MIXED=y CONFIG_HZ_1000=y -CONFIG_RANDOMIZE_MEMORY_PHYSICAL_PADDING=0xa # CONFIG_CPU_MITIGATIONS is not set CONFIG_HIBERNATION=y # CONFIG_HIBERNATION_SNAPSHOT_DEV is not set CONFIG_PM_STD_PARTITION="/swapfile" CONFIG_PM_WAKELOCKS=y +CONFIG_PM_QOS_CPU_SYSTEM_WAKEUP=y CONFIG_PM_DEBUG=y CONFIG_PM_ADVANCED_DEBUG=y CONFIG_PM_TRACE_RTC=y
@@ -144,21 +148,29 @@ CONFIG_ZSWAP_DEFAULT_ON=y CONFIG_ZSWAP_SHRINKER_DEFAULT_ON=y CONFIG_ZSWAP_COMPRESSOR_DEFAULT_LZ4=y CONFIG_SLAB_FREELIST_RANDOM=y -CONFIG_SLAB_FREELIST_HARDENED=y +CONFIG_SLAB_BUCKETS=y +CONFIG_RANDOM_KMALLOC_CACHES=y CONFIG_SHUFFLE_PAGE_ALLOCATOR=y # CONFIG_COMPAT_BRK is not set +CONFIG_MEMORY_HOTPLUG=y +CONFIG_MHP_DEFAULT_ONLINE_TYPE_ONLINE_AUTO=y +CONFIG_MEMORY_HOTREMOVE=y CONFIG_PAGE_REPORTING=y CONFIG_KSM=y CONFIG_DEFAULT_MMAP_MIN_ADDR=65536 CONFIG_MEMORY_FAILURE=y -CONFIG_PERSISTENT_HUGE_ZERO_FOLIO=y CONFIG_TRANSPARENT_HUGEPAGE=y CONFIG_TRANSPARENT_HUGEPAGE_MADVISE=y -CONFIG_TRANSPARENT_HUGEPAGE_SHMEM_HUGE_ADVISE=y -CONFIG_TRANSPARENT_HUGEPAGE_TMPFS_HUGE_ADVISE=y +CONFIG_CMA=y +CONFIG_CMA_DEBUGFS=y +CONFIG_CMA_SYSFS=y CONFIG_MEM_SOFT_DIRTY=y CONFIG_IDLE_PAGE_TRACKING=y +CONFIG_ZONE_DEVICE=y +CONFIG_ANON_VMA_NAME=y CONFIG_USERFAULTFD=y +CONFIG_LRU_GEN=y +CONFIG_LRU_GEN_ENABLED=y CONFIG_NET=y CONFIG_PACKET=y CONFIG_PACKET_DIAG=y
@@ -387,10 +399,6 @@ CONFIG_BT_HIDP=y CONFIG_BT_LEDS=y CONFIG_BT_MSFTEXT=y CONFIG_BT_AOSPEXT=y -CONFIG_BT_HCIBTUSB=m -CONFIG_BT_HCIBTUSB_AUTOSUSPEND=y -# CONFIG_BT_HCIBTUSB_BCM is not set -# CONFIG_BT_HCIBTUSB_RTL is not set CONFIG_BT_INTEL_PCIE=m CONFIG_CFG80211=m CONFIG_CFG80211_DEBUGFS=y
@@ -400,6 +408,7 @@ CONFIG_MAC80211_MESH=y CONFIG_MAC80211_DEBUGFS=y CONFIG_RFKILL=y CONFIG_RFKILL_INPUT=y +CONFIG_PAGE_POOL_STATS=y CONFIG_PCI=y CONFIG_PCIEAER=y CONFIG_PCIE_DPC=y
@@ -416,10 +425,12 @@ CONFIG_DEVTMPFS_MOUNT=y # CONFIG_STANDALONE is not set CONFIG_FW_LOADER_USER_HELPER=y CONFIG_FW_LOADER_COMPRESS=y +CONFIG_FW_LOADER_COMPRESS_ZSTD=y CONFIG_CONNECTOR=y CONFIG_EDD=y CONFIG_EDD_OFF=y CONFIG_DMI_SYSFS=y +CONFIG_SYSFB_SIMPLEFB=y CONFIG_EFI_BOOTLOADER_CONTROL=y CONFIG_RESET_ATTACK_MITIGATION=y CONFIG_EFI_RCI2_TABLE=y
@@ -438,8 +449,9 @@ CONFIG_SRAM=y CONFIG_EEPROM_AT24=y CONFIG_EEPROM_EE1004=y CONFIG_INTEL_MEI_GSC=y +CONFIG_INTEL_MEI_CSC=y CONFIG_INTEL_MEI_GSC_PROXY=y -CONFIG_VMWARE_VMCI=m +CONFIG_VMWARE_VMCI=y CONFIG_SCSI=y CONFIG_BLK_DEV_SD=y CONFIG_CHR_DEV_ST=y
@@ -537,6 +549,7 @@ CONFIG_IGC=y # CONFIG_NET_VENDOR_XILINX is not set CONFIG_PHYLIB=y CONFIG_LED_TRIGGER_PHY=y +CONFIG_FIXED_PHY=y CONFIG_AX88796B_PHY=y CONFIG_SMSC_PHY=y CONFIG_PPP=y
@@ -550,15 +563,15 @@ CONFIG_PPPOL2TP=y CONFIG_PPP_ASYNC=y CONFIG_USB_RTL8150=m CONFIG_USB_RTL8152=m -CONFIG_USB_USBNET=m +CONFIG_USB_USBNET=y # CONFIG_USB_NET_AX8817X is not set -CONFIG_USB_NET_CDC_EEM=m -CONFIG_USB_NET_CDC_MBIM=m +CONFIG_USB_NET_AX88179_178A=m +CONFIG_USB_NET_CDC_EEM=y +CONFIG_USB_NET_CDC_MBIM=y # CONFIG_USB_NET_NET1080 is not set # CONFIG_USB_NET_CDC_SUBSET is not set # CONFIG_USB_NET_ZAURUS is not set CONFIG_USB_HSO=y -CONFIG_USB_SIERRA_NET=m CONFIG_USB_NET_AQC111=m # CONFIG_WLAN_VENDOR_ADMTEK is not set # CONFIG_WLAN_VENDOR_ATH is not set
@@ -633,10 +646,10 @@ CONFIG_I2C_DESIGNWARE_PCI=y CONFIG_SPI=y CONFIG_SPI_MEM=y CONFIG_SPI_CH341=y -CONFIG_SPI_CS42L43=m +CONFIG_SPI_CS42L43=y CONFIG_SPI_INTEL_PCI=y # CONFIG_PTP_1588_CLOCK is not set -CONFIG_PINCTRL_CS42L43=m +CONFIG_PINCTRL_CS42L43=y CONFIG_PINCTRL_INTEL_PLATFORM=y CONFIG_PINCTRL_TIGERLAKE=y CONFIG_GPIO_SYSFS=y
@@ -670,7 +683,6 @@ CONFIG_WATCHDOG_SYSFS=y CONFIG_WDAT_WDT=y CONFIG_INTEL_OC_WATCHDOG=y CONFIG_ITCO_WDT=y -CONFIG_MFD_CS42L43_SDW=y CONFIG_LPC_ICH=y CONFIG_LPC_SCH=y CONFIG_MFD_INTEL_LPSS_ACPI=y
@@ -687,8 +699,6 @@ CONFIG_MEDIA_USB_SUPPORT=y CONFIG_USB_GSPCA=y CONFIG_USB_VIDEO_CLASS=y CONFIG_MEDIA_PCI_SUPPORT=y -CONFIG_VIDEO_INTEL_IPU6=m -CONFIG_IPU_BRIDGE=y CONFIG_V4L_PLATFORM_DRIVERS=y CONFIG_V4L_MEM2MEM_DRIVERS=y CONFIG_VIDEO_MEM2MEM_DEINTERLACE=y
@@ -697,7 +707,7 @@ CONFIG_VIDEO_OV01A10=y CONFIG_DRM=y CONFIG_DRM_CLIENT_LOG=y CONFIG_DRM_DISPLAY_DP_AUX_CEC=y -CONFIG_DRM_EFIDRM=y +CONFIG_DRM_SIMPLEDRM=y CONFIG_DRM_UDL=y CONFIG_DRM_XE=m CONFIG_FB=y
@@ -741,7 +751,7 @@ CONFIG_SND_SOC=y # CONFIG_SND_SOC_INTEL_SST_TOPLEVEL is not set CONFIG_SND_SOC_INTEL_USER_FRIENDLY_LONG_NAMES=y CONFIG_SND_SOC_INTEL_SKL_HDA_DSP_GENERIC_MACH=y -CONFIG_SND_SOC_INTEL_SOUNDWIRE_SOF_MACH=m +CONFIG_SND_SOC_INTEL_SOUNDWIRE_SOF_MACH=y CONFIG_SND_SOC_SOF_TOPLEVEL=y CONFIG_SND_SOC_SOF_PCI=y CONFIG_SND_SOC_SOF_ACPI=y
@@ -760,23 +770,11 @@ CONFIG_SND_SOC_SOF_INTEL_TOPLEVEL=y # CONFIG_SND_SOC_SOF_ELKHARTLAKE is not set # CONFIG_SND_SOC_SOF_ALDERLAKE is not set # CONFIG_SND_SOC_SOF_METEORLAKE is not set -CONFIG_SND_SOC_SOF_LUNARLAKE=m +# CONFIG_SND_SOC_SOF_LUNARLAKE is not set # CONFIG_SND_SOC_SOF_PANTHERLAKE is not set CONFIG_SND_SOC_SOF_HDA_LINK=y CONFIG_SND_SOC_SOF_HDA_AUDIO_CODEC=y -CONFIG_SND_SOC_CS35L56_I2C=m CONFIG_SND_SOC_HDA=y -CONFIG_SND_SOC_MAX98363=y -CONFIG_SND_SOC_MAX98373_I2C=y -CONFIG_SND_SOC_MAX98373_SDW=y -CONFIG_SND_SOC_RT1308_SDW=y -CONFIG_SND_SOC_RT1316_SDW=y -CONFIG_SND_SOC_RT5682_SDW=y -CONFIG_SND_SOC_RT700_SDW=y -CONFIG_SND_SOC_RT711_SDW=y -CONFIG_SND_SOC_RT711_SDCA_SDW=y -CONFIG_SND_SOC_RT715_SDW=y -CONFIG_SND_SOC_RT715_SDCA_SDW=y # CONFIG_SND_SOC_SDW_MOCKUP is not set # CONFIG_SND_X86 is not set CONFIG_HID_BATTERY_STRENGTH=y
@@ -800,12 +798,9 @@ CONFIG_USB_LED_TRIG=y CONFIG_USB=y CONFIG_USB_ANNOUNCE_NEW_DEVICES=y CONFIG_USB_DYNAMIC_MINORS=y -CONFIG_USB_LEDS_TRIGGER_USBPORT=y -CONFIG_USB_MON=m CONFIG_USB_XHCI_HCD=y CONFIG_USB_ACM=y CONFIG_USB_PRINTER=y -CONFIG_USB_WDM=y CONFIG_USB_STORAGE=y CONFIG_USB_UAS=y CONFIG_USB_SERIAL=y
@@ -815,13 +810,12 @@ CONFIG_USB_SERIAL_SIMPLE=y CONFIG_USB_SERIAL_BELKIN=y CONFIG_USB_SERIAL_CH341=y CONFIG_USB_SERIAL_CP210X=y -CONFIG_USB_SERIAL_FTDI_SIO=m +CONFIG_USB_SERIAL_FTDI_SIO=y CONFIG_USB_SERIAL_KEYSPAN_PDA=y CONFIG_USB_SERIAL_KEYSPAN=y CONFIG_USB_SERIAL_MCT_U232=y CONFIG_USB_SERIAL_MXUPORT=y CONFIG_USB_SERIAL_PL2303=y -CONFIG_USB_SERIAL_SIERRAWIRELESS=y CONFIG_USB_SERIAL_OPTION=y CONFIG_USB_USBIO=y CONFIG_TYPEC=y
@@ -837,19 +831,9 @@ CONFIG_LEDS_BRIGHTNESS_HW_CHANGED=y CONFIG_LEDS_PWM=y CONFIG_LEDS_USER=y CONFIG_LEDS_PWM_MULTICOLOR=y -CONFIG_LEDS_TRIGGER_TIMER=y -CONFIG_LEDS_TRIGGER_ONESHOT=y -CONFIG_LEDS_TRIGGER_HEARTBEAT=y CONFIG_LEDS_TRIGGER_BACKLIGHT=y -CONFIG_LEDS_TRIGGER_CPU=y -CONFIG_LEDS_TRIGGER_ACTIVITY=y CONFIG_LEDS_TRIGGER_DEFAULT_ON=y -CONFIG_LEDS_TRIGGER_TRANSIENT=y -CONFIG_LEDS_TRIGGER_CAMERA=y CONFIG_LEDS_TRIGGER_PANIC=y -CONFIG_LEDS_TRIGGER_NETDEV=y -CONFIG_LEDS_TRIGGER_PATTERN=y -CONFIG_LEDS_TRIGGER_TTY=y CONFIG_EDAC=y CONFIG_EDAC_GHES=y CONFIG_EDAC_IGEN6=y
@@ -935,8 +919,7 @@ CONFIG_INTEL_RAPL_TPMI=y CONFIG_USB4=y # CONFIG_BLK_DEV_PMEM is not set # CONFIG_BTT is not set -CONFIG_COUNTER=y -CONFIG_INTERRUPT_CNT=y +# CONFIG_NVDIMM_DAX is not set CONFIG_VALIDATE_FS_PARSER=y CONFIG_EXT4_FS=y CONFIG_EXT4_FS_POSIX_ACL=y
@@ -959,7 +942,6 @@ CONFIG_FS_VERITY=y CONFIG_FS_VERITY_BUILTIN_SIGNATURES=y CONFIG_FANOTIFY=y CONFIG_FANOTIFY_ACCESS_PERMISSIONS=y -CONFIG_QUOTA=y CONFIG_QUOTA_NETLINK_INTERFACE=y CONFIG_AUTOFS_FS=y CONFIG_FUSE_FS=y
@@ -979,6 +961,7 @@ CONFIG_PROC_KCORE=y CONFIG_PROC_VMCORE_DEVICE_DUMP=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_INODE64=y +CONFIG_TMPFS_QUOTA=y CONFIG_HUGETLBFS=y CONFIG_EFIVAR_FS=y CONFIG_ECRYPT_FS=y
@@ -1138,6 +1121,7 @@ CONFIG_PACKING=y # CONFIG_XZ_DEC_SPARC is not set # CONFIG_XZ_DEC_RISCV is not set CONFIG_SWIOTLB_DYNAMIC=y +CONFIG_DMA_CMA=y CONFIG_IRQ_POLL=y CONFIG_FONTS=y CONFIG_FONT_8x8=y
@@ -1168,6 +1152,7 @@ CONFIG_BOOTPARAM_HARDLOCKUP_PANIC=y CONFIG_WQ_WATCHDOG=y CONFIG_WQ_CPU_INTENSIVE_REPORT=y CONFIG_SCHEDSTATS=y +CONFIG_NMI_CHECK_CPU=y CONFIG_STACKTRACE=y CONFIG_RCU_CPU_STALL_TIMEOUT=60 # CONFIG_RCU_TRACE is not set
Signature · 1 line
2.53.0
Matthew Brost
maintainerPATCH v4 3/6 drm/ttm: Issue direct reclaim at beneficial_order
In reply to Matthew BrostTriggering kswap at an order higher than beneficial_order makes little sense, as the driver has already indicated the optimal order at which reclaim is effective. Similarly, issuing direct reclaim or triggering kswap at a lower order than beneficial_order is ineffective, since the driver does not benefit from reclaiming lower-order pages.
As a result, direct reclaim should only be issued with __GFP_NORETRY at exactly beneficial_order, or as a fallback, direct reclaim without __GFP_NORETRY at order 0 when failure is not an option.
--- drivers/gpu/drm/ttm/ttm_pool.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/ttm/ttm_pool.c +2 -2
@@ -165,8 +165,8 @@ static struct page *ttm_pool_alloc_page(struct ttm_pool *pool, gfp_t gfp_flags,* Do not add latency to the allocation path for allocations orders * device tolds us do not bring them additional performance gains. */ -if (beneficial_order && order > beneficial_order) - gfp_flags &= ~__GFP_DIRECT_RECLAIM; +if (order && beneficial_order && order != beneficial_order) + gfp_flags &= ~__GFP_RECLAIM;if (!ttm_pool_uses_dma_alloc (pool)) { p =alloc_pages_node (pool->nid, gfp_flags, order);
Signature · 1 line
2.34.1
Re: [PATCH v4 3/6] drm/ttm: Issue direct reclaim at beneficial_order
Quoted 36 lines · Matthew Brost
Triggering kswap at an order higher than
beneficial_ordermakes little sense, as the driver has already indicated the optimal order at which reclaim is effective. Similarly, issuing direct reclaim or triggering kswap at a lower order thanbeneficial_orderis ineffective, since the driver does not benefit from reclaiming lower-order pages.As a result, direct reclaim should only be issued with
__GFP_NORETRYat exactlybeneficial_order, or as a fallback, direct reclaim without__GFP_NORETRYat order 0 when failure is not an option.Cc Tvrtko Ursulin <[email protected]>Cc Thomas Hellström <[email protected]>Cc Carlos Santa <[email protected]>Cc Christian Koenig <[email protected]>Cc Huang Rui <[email protected]>Cc Matthew Auld <[email protected]>Cc Matthew Brost <[email protected]>Cc Maarten Lankhorst <[email protected]>Cc Maxime Ripard <[email protected]>Cc Thomas Zimmermann <[email protected]>Cc David Airlie <[email protected]>Cc Simona Vetter <[email protected]>Cc Daniel Colascione <[email protected]>Signed-off-by Matthew Brost <[email protected]>Reviewed-by Christian Koenig <[email protected]>Reviewed-by Andi Shyti <[email protected]>--- drivers/gpu/drm/ttm/ttm_pool.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)drivers/gpu/drm/ttm/ttm_pool.c +0
-0
drivers/gpu/drm/ttm/ttm_pool.c static struct page *ttm_pool_alloc_page(struct ttm_pool *pool, gfp_t gfp_flags, · line 169
Quoted 7 lines
drivers/gpu/drm/ttm/ttm_pool.c +2
-2@@ -165,8 +165,8 @@ static struct page *ttm_pool_alloc_page(struct ttm_pool *pool, gfp_t gfp_flags,* Do not add latency to the allocation path for allocations orders * device tolds us do not bring them additional performance gains. */ -if (beneficial_order && order > beneficial_order) - gfp_flags &= ~__GFP_DIRECT_RECLAIM; +if (order && beneficial_order && order != beneficial_order) + gfp_flags &= ~__GFP_RECLAIM;
Just one additional nit: The subject line says "direct reclaim" but we disable both direct as well as kswapd based reclaim.
Potentially just update the subject line.
Regards,
Christian.
Quoted 3 lines
drivers/gpu/drm/ttm/ttm_pool.c +0
-0if (!ttm_pool_uses_dma_alloc (pool)) { p =alloc_pages_node (pool->nid, gfp_flags, order);
Matthew Brost
maintainerPATCH v4 4/6 drm/ttm: Introduce ttm_bo_shrink_kswap_maybe_fragmented()
In reply to Matthew BrostIntroduce ttm_bo_shrink_kswap_maybe_fragmented() to allow TTM users to distinguish background reclaim from kswapd that is likely driven by high-order allocation failures under fragmentation.
The helper returns true when:
- order of shrinker invocation is zero
- reclaim is running in kswapd, and
- the target node is valid, and
- one of the relevant zones reports free pages significantly above its high watermark (via zone_appears_fragmented()).
This provides a coarse signal that overall free memory is available, and that reclaim activity may be driven by fragmentation rather than true memory pressure.
The intent is to allow drivers to adjust shrinker behavior in this case, for example by preferring purgeable or low-value objects instead of aggressively evicting active working sets in the background reclaim path.
The heuristic is intentionally simple and conservative, and is not intended to replace core MM fragmentation or compaction decisions.
No functional change; this is a preparatory helper for TTM users.
---
v3:
- s/ttm_bo_shrink_kswap_fragmented/ttm_bo_shrink_kswap_maybe_fragmented (Andi)
- Wire in order (Thomas)
--- drivers/gpu/drm/ttm/ttm_bo_util.c | 38 +++++++++++++++++++++++++++++++ include/drm/ttm/ttm_bo.h | 2 ++ 2 files changed, 40 insertions(+)
drivers/gpu/drm/ttm/ttm_bo_util.c +38 -0
@@ -1169,3 +1169,41 @@ bool ttm_bo_shrink_avoid_wait(void)return !current_is_kswapd (); }EXPORT_SYMBOL (ttm_bo_shrink_avoid_wait); + +/** +* ttm_bo_shrink_kswap_maybe_fragmented() - Whether in kswap and memory might be +* fragmented +* @nid: current node being shrunk +* @order: order of shrinker invocation +* +* Return: true if in kswap and memory appears fragmented, false is not. +*/ +bool ttm_bo_shrink_kswap_maybe_fragmented (int nid,s8 order) +{ +enum zone_type zone_type; + +if (!order) +return false; + +if (!current_is_kswapd ()) +return false; + +if (!numa_valid_node (nid)) +return false; + +#if IS_ENABLED (CONFIG_ZONE_DMA32 ) + zone_type =ZONE_DMA32 ; +#else + zone_type =ZONE_NORMAL ; +#endif + +for (; zone_type <=ZONE_NORMAL ; ++zone_type) { +struct zone *zone = &NODE_DATA (nid)->node_zones[zone_type]; + +if (zone_maybe_fragmented_in_shrinker (zone)) +return true; + } + +return false; +} +EXPORT_SYMBOL (ttm_bo_shrink_kswap_maybe_fragmented);
include/drm/ttm/ttm_bo.h +2 -0
@@ -262,6 +262,8 @@ bool ttm_bo_shrink_suitable(struct ttm_buffer_object *bo, struct ttm_operation_cbool ttm_bo_shrink_avoid_wait (void ); +bool ttm_bo_shrink_kswap_maybe_fragmented (int nid,s8 order); +/** * ttm_bo_reserve: *
Signature · 1 line
2.34.1
Matthew Brost
maintainerPATCH v4 5/6 drm/xe: Set TTM device beneficial_order to 9 (2M)
In reply to Matthew BrostSet the TTM device beneficial_order to 9 (2M), which is the sweet spot for Xe when attempting reclaim on system memory BOs, as it matches the large GPU page size. This ensures reclaim is attempted at the most effective order for the driver.
--- drivers/gpu/drm/xe/xe_device.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/xe/xe_device.c +2 -1
@@ -500,7 +500,8 @@ struct xe_device *xe_device_create(struct pci_dev *pdev, err =ttm_device_init (&xe->ttm, &xe_ttm_funcs, xe->drm.dev, xe->drm.anon_inode->i_mapping, - xe->drm.vma_offset_manager,0 ); + xe->drm.vma_offset_manager, +TTM_ALLOCATION_POOL_BENEFICIAL_ORDER (get_order (SZ_2M )));if (WARN_ON (err))return ERR_PTR (err);
Signature · 1 line
2.34.1
Matthew Brost
maintainerPATCH v4 6/6 drm/xe: Avoid shrinker reclaim from kswapd under fragmentation
In reply to Matthew BrostWhen the Xe shrinker is invoked from kswapd, a large amount of free memory in usable zones relative to the high watermark is a strong signal that reclaim is being driven by fragmentation rather than true memory pressure.
In this case, shrinking Xe memory is unlikely to help kswapd make forward progress. Instead it can evict active GPU memory despite the system still having substantial free memory, increasing residency churn and reducing GPU forward progress.
Detect this case and bail out early from the Xe shrinker when running in kswapd, shrinker invocation is a higher order, and any usable zone has more than 2x its high watermark free.
---
v3:
- Wire in order to heuristic (Thomas)
--- drivers/gpu/drm/xe/xe_shrinker.c | 3 +++ 1 file changed, 3 insertions(+)
drivers/gpu/drm/xe/xe_shrinker.c +3 -0
@@ -236,6 +236,9 @@ static unsigned long xe_shrinker_scan(struct shrinker *shrink, struct shrink_conif (nr_scanned >= nr_to_scan || !can_backup) goto out; +if (ttm_bo_shrink_kswap_maybe_fragmented (sc->nid, sc->order)) + goto out; +/* If we didn't wake before, try to do it now if needed. */ if (!runtime_pm) runtime_pm =xe_shrinker_runtime_pm_get (shrinker, true,0 , can_backup);
Signature · 1 line
2.34.1
Patchwork
CI✗ CI.checkpatch: warning for mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation (rev3)
In reply to Matthew Brost== Series Details ==
== Series Details ==
Series: mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation (rev3)
URL : https://patchwork.freedesktop.org/series/165329/
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 c8c12e558adaef7a4d125d83b6e1f8824bc13b82
+ cd /kernel
+ git config --global --add safe.directory /kernel
+ git log -n1
commit baa0e85108cdb5402d65fe0838430a8e98ecf188
Author: Matthew Brost <[email protected]> Date: Thu Apr 30 12:18:09 2026 -0700
drm/xe: Avoid shrinker reclaim from kswapd under fragmentation
When the Xe shrinker is invoked from kswapd, a large amount of free memory in usable zones relative to the high watermark is a strong signal that reclaim is being driven by fragmentation rather than true memory pressure.
In this case, shrinking Xe memory is unlikely to help kswapd make forward progress. Instead it can evict active GPU memory despite the system still having substantial free memory, increasing residency churn and reducing GPU forward progress.
Detect this case and bail out early from the Xe shrinker when running in kswapd, shrinker invocation is a higher order, and any usable zone has more than 2x its high watermark free.
+ /mt/dim checkpatch 6472756e8384e3945f12b6726dd9dc1ac12f3bb4 drm-intel f0c13668a40c mm: Wire up order in shrink_control
-:79: CHECK:PARENTHESIS_ALIGNMENT: Alignment should match open parenthesis #79: FILE: mm/shrinker.c:470:
+static unsigned long shrink_slab_memcg(gfp_t gfp_mask, int nid, s8 order,
struct mem_cgroup *memcg, int priority)
-:96: CHECK:PARENTHESIS_ALIGNMENT: Alignment should match open parenthesis #96: FILE: mm/shrinker.c:592:
+static unsigned long shrink_slab_memcg(gfp_t gfp_mask, int nid, s8 order,
struct mem_cgroup *memcg, int priority)
total: 0 errors, 0 warnings, 2 checks, 99 lines checked 5377446e132a mm: Introduce zone_maybe_fragmented_in_shrinker() 5dc5b8f49f0b drm/ttm: Issue direct reclaim at beneficial_order b8ae9b5049f4 drm/ttm: Introduce ttm_bo_shrink_kswap_maybe_fragmented() 7150b7681ccd drm/xe: Set TTM device beneficial_order to 9 (2M) baa0e85108cd drm/xe: Avoid shrinker reclaim from kswapd under fragmentation
Patchwork
CI✓ CI.KUnit: success for mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation (rev3)
In reply to Matthew Brost== Series Details ==
== Series Details ==
Series: mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation (rev3)
URL : https://patchwork.freedesktop.org/series/165329/
State : success
== Summary ==
+ trap cleanup EXIT
+ /kernel/tools/testing/kunit/kunit.py run --kunitconfig /kernel/drivers/gpu/drm/xe/.kunitconfig [20:04:52] Configuring KUnit Kernel ...
Generating .config ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig [20:04:56] 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 [20:05:28] Starting KUnit Kernel (1/1)... [20:05:28] ============================================================ Running tests with: $ .kunit/linux kunit.enable=1 mem=1G console=tty kunit_shutdown=halt [20:05:28] ================== guc_buf (11 subtests) =================== [20:05:28] [PASSED] test_smallest [20:05:28] [PASSED] test_largest [20:05:28] [PASSED] test_granular [20:05:28] [PASSED] test_unique [20:05:28] [PASSED] test_overlap [20:05:28] [PASSED] test_reusable [20:05:28] [PASSED] test_too_big [20:05:28] [PASSED] test_flush [20:05:28] [PASSED] test_lookup [20:05:28] [PASSED] test_data [20:05:28] [PASSED] test_class [20:05:28] ===================== [PASSED] guc_buf ===================== [20:05:28] =================== guc_dbm (7 subtests) =================== [20:05:28] [PASSED] test_empty [20:05:28] [PASSED] test_default [20:05:28] ======================== test_size ======================== [20:05:28] [PASSED] 4 [20:05:28] [PASSED] 8 [20:05:28] [PASSED] 32 [20:05:28] [PASSED] 256 [20:05:28] ==================== [PASSED] test_size ==================== [20:05:28] ======================= test_reuse ======================== [20:05:28] [PASSED] 4 [20:05:28] [PASSED] 8 [20:05:28] [PASSED] 32 [20:05:28] [PASSED] 256 [20:05:28] =================== [PASSED] test_reuse ==================== [20:05:28] =================== test_range_overlap ==================== [20:05:28] [PASSED] 4 [20:05:28] [PASSED] 8 [20:05:28] [PASSED] 32 [20:05:28] [PASSED] 256 [20:05:28] =============== [PASSED] test_range_overlap ================ [20:05:28] =================== test_range_compact ==================== [20:05:28] [PASSED] 4 [20:05:28] [PASSED] 8 [20:05:28] [PASSED] 32 [20:05:28] [PASSED] 256 [20:05:28] =============== [PASSED] test_range_compact ================ [20:05:28] ==================== test_range_spare ===================== [20:05:28] [PASSED] 4 [20:05:28] [PASSED] 8 [20:05:28] [PASSED] 32 [20:05:28] [PASSED] 256 [20:05:28] ================ [PASSED] test_range_spare ================= [20:05:28] ===================== [PASSED] guc_dbm ===================== [20:05:28] =================== guc_idm (6 subtests) =================== [20:05:28] [PASSED] bad_init [20:05:28] [PASSED] no_init [20:05:28] [PASSED] init_fini [20:05:28] [PASSED] check_used [20:05:28] [PASSED] check_quota [20:05:28] [PASSED] check_all [20:05:28] ===================== [PASSED] guc_idm ===================== [20:05:28] ================== no_relay (3 subtests) =================== [20:05:28] [PASSED] xe_drops_guc2pf_if_not_ready [20:05:28] [PASSED] xe_drops_guc2vf_if_not_ready [20:05:28] [PASSED] xe_rejects_send_if_not_ready [20:05:28] ==================== [PASSED] no_relay ===================== [20:05:28] ================== pf_relay (14 subtests) ================== [20:05:28] [PASSED] pf_rejects_guc2pf_too_short [20:05:28] [PASSED] pf_rejects_guc2pf_too_long [20:05:28] [PASSED] pf_rejects_guc2pf_no_payload [20:05:28] [PASSED] pf_fails_no_payload [20:05:28] [PASSED] pf_fails_bad_origin [20:05:28] [PASSED] pf_fails_bad_type [20:05:28] [PASSED] pf_txn_reports_error [20:05:28] [PASSED] pf_txn_sends_pf2guc [20:05:28] [PASSED] pf_sends_pf2guc [20:05:28] [SKIPPED] pf_loopback_nop [20:05:28] [SKIPPED] pf_loopback_echo [20:05:28] [SKIPPED] pf_loopback_fail [20:05:28] [SKIPPED] pf_loopback_busy [20:05:28] [SKIPPED] pf_loopback_retry [20:05:28] ==================== [PASSED] pf_relay ===================== [20:05:28] ================== vf_relay (3 subtests) =================== [20:05:28] [PASSED] vf_rejects_guc2vf_too_short [20:05:28] [PASSED] vf_rejects_guc2vf_too_long [20:05:28] [PASSED] vf_rejects_guc2vf_no_payload [20:05:28] ==================== [PASSED] vf_relay ===================== [20:05:28] ================ pf_gt_config (9 subtests) ================= [20:05:28] [PASSED] fair_contexts_1vf [20:05:28] [PASSED] fair_doorbells_1vf [20:05:28] [PASSED] fair_ggtt_1vf [20:05:28] ====================== fair_vram_1vf ====================== [20:05:28] [PASSED] 3.50 GiB [20:05:28] [PASSED] 11.5 GiB [20:05:28] [PASSED] 15.5 GiB [20:05:28] [PASSED] 31.5 GiB [20:05:28] [PASSED] 63.5 GiB [20:05:28] [PASSED] 1.91 GiB [20:05:28] ================== [PASSED] fair_vram_1vf ================== [20:05:28] ================ fair_vram_1vf_admin_only ================= [20:05:28] [PASSED] 3.50 GiB [20:05:28] [PASSED] 11.5 GiB [20:05:28] [PASSED] 15.5 GiB [20:05:28] [PASSED] 31.5 GiB [20:05:28] [PASSED] 63.5 GiB [20:05:28] [PASSED] 1.91 GiB [20:05:28] ============ [PASSED] fair_vram_1vf_admin_only ============= [20:05:28] ====================== fair_contexts ====================== [20:05:28] [PASSED] 1 VF [20:05:28] [PASSED] 2 VFs [20:05:28] [PASSED] 3 VFs [20:05:28] [PASSED] 4 VFs [20:05:28] [PASSED] 5 VFs [20:05:28] [PASSED] 6 VFs [20:05:28] [PASSED] 7 VFs [20:05:28] [PASSED] 8 VFs [20:05:28] [PASSED] 9 VFs [20:05:28] [PASSED] 10 VFs [20:05:28] [PASSED] 11 VFs [20:05:28] [PASSED] 12 VFs [20:05:28] [PASSED] 13 VFs [20:05:28] [PASSED] 14 VFs [20:05:28] [PASSED] 15 VFs [20:05:28] [PASSED] 16 VFs [20:05:28] [PASSED] 17 VFs [20:05:28] [PASSED] 18 VFs [20:05:28] [PASSED] 19 VFs [20:05:28] [PASSED] 20 VFs [20:05:28] [PASSED] 21 VFs [20:05:28] [PASSED] 22 VFs [20:05:28] [PASSED] 23 VFs [20:05:28] [PASSED] 24 VFs [20:05:28] [PASSED] 25 VFs [20:05:28] [PASSED] 26 VFs [20:05:28] [PASSED] 27 VFs [20:05:28] [PASSED] 28 VFs [20:05:28] [PASSED] 29 VFs [20:05:28] [PASSED] 30 VFs [20:05:28] [PASSED] 31 VFs [20:05:28] [PASSED] 32 VFs [20:05:28] [PASSED] 33 VFs [20:05:28] [PASSED] 34 VFs [20:05:28] [PASSED] 35 VFs [20:05:28] [PASSED] 36 VFs [20:05:28] [PASSED] 37 VFs [20:05:28] [PASSED] 38 VFs [20:05:28] [PASSED] 39 VFs [20:05:28] [PASSED] 40 VFs [20:05:28] [PASSED] 41 VFs [20:05:28] [PASSED] 42 VFs [20:05:28] [PASSED] 43 VFs [20:05:28] [PASSED] 44 VFs [20:05:28] [PASSED] 45 VFs [20:05:28] [PASSED] 46 VFs [20:05:28] [PASSED] 47 VFs [20:05:28] [PASSED] 48 VFs [20:05:28] [PASSED] 49 VFs [20:05:28] [PASSED] 50 VFs [20:05:28] [PASSED] 51 VFs [20:05:28] [PASSED] 52 VFs [20:05:28] [PASSED] 53 VFs [20:05:28] [PASSED] 54 VFs [20:05:28] [PASSED] 55 VFs [20:05:28] [PASSED] 56 VFs [20:05:28] [PASSED] 57 VFs [20:05:28] [PASSED] 58 VFs [20:05:28] [PASSED] 59 VFs [20:05:28] [PASSED] 60 VFs [20:05:28] [PASSED] 61 VFs [20:05:28] [PASSED] 62 VFs [20:05:28] [PASSED] 63 VFs [20:05:28] ================== [PASSED] fair_contexts ================== [20:05:28] ===================== fair_doorbells ====================== [20:05:28] [PASSED] 1 VF [20:05:28] [PASSED] 2 VFs [20:05:28] [PASSED] 3 VFs [20:05:28] [PASSED] 4 VFs [20:05:28] [PASSED] 5 VFs [20:05:28] [PASSED] 6 VFs [20:05:28] [PASSED] 7 VFs [20:05:28] [PASSED] 8 VFs [20:05:28] [PASSED] 9 VFs [20:05:28] [PASSED] 10 VFs [20:05:28] [PASSED] 11 VFs [20:05:28] [PASSED] 12 VFs [20:05:28] [PASSED] 13 VFs [20:05:28] [PASSED] 14 VFs [20:05:28] [PASSED] 15 VFs [20:05:28] [PASSED] 16 VFs [20:05:28] [PASSED] 17 VFs [20:05:28] [PASSED] 18 VFs [20:05:28] [PASSED] 19 VFs [20:05:28] [PASSED] 20 VFs [20:05:28] [PASSED] 21 VFs [20:05:28] [PASSED] 22 VFs [20:05:28] [PASSED] 23 VFs [20:05:28] [PASSED] 24 VFs [20:05:28] [PASSED] 25 VFs [20:05:28] [PASSED] 26 VFs [20:05:28] [PASSED] 27 VFs [20:05:28] [PASSED] 28 VFs [20:05:28] [PASSED] 29 VFs [20:05:28] [PASSED] 30 VFs [20:05:28] [PASSED] 31 VFs [20:05:28] [PASSED] 32 VFs [20:05:28] [PASSED] 33 VFs [20:05:28] [PASSED] 34 VFs [20:05:28] [PASSED] 35 VFs [20:05:28] [PASSED] 36 VFs [20:05:28] [PASSED] 37 VFs [20:05:28] [PASSED] 38 VFs [20:05:28] [PASSED] 39 VFs [20:05:28] [PASSED] 40 VFs [20:05:28] [PASSED] 41 VFs [20:05:28] [PASSED] 42 VFs [20:05:28] [PASSED] 43 VFs [20:05:28] [PASSED] 44 VFs [20:05:28] [PASSED] 45 VFs [20:05:28] [PASSED] 46 VFs [20:05:28] [PASSED] 47 VFs [20:05:28] [PASSED] 48 VFs [20:05:28] [PASSED] 49 VFs [20:05:28] [PASSED] 50 VFs [20:05:28] [PASSED] 51 VFs [20:05:28] [PASSED] 52 VFs [20:05:28] [PASSED] 53 VFs [20:05:28] [PASSED] 54 VFs [20:05:28] [PASSED] 55 VFs [20:05:28] [PASSED] 56 VFs [20:05:28] [PASSED] 57 VFs [20:05:28] [PASSED] 58 VFs [20:05:28] [PASSED] 59 VFs [20:05:28] [PASSED] 60 VFs [20:05:28] [PASSED] 61 VFs [20:05:28] [PASSED] 62 VFs [20:05:28] [PASSED] 63 VFs [20:05:28] ================= [PASSED] fair_doorbells ================== [20:05:28] ======================== fair_ggtt ======================== [20:05:28] [PASSED] 1 VF [20:05:28] [PASSED] 2 VFs [20:05:28] [PASSED] 3 VFs [20:05:28] [PASSED] 4 VFs [20:05:28] [PASSED] 5 VFs [20:05:28] [PASSED] 6 VFs [20:05:28] [PASSED] 7 VFs [20:05:28] [PASSED] 8 VFs [20:05:28] [PASSED] 9 VFs [20:05:28] [PASSED] 10 VFs [20:05:28] [PASSED] 11 VFs [20:05:28] [PASSED] 12 VFs [20:05:28] [PASSED] 13 VFs [20:05:28] [PASSED] 14 VFs [20:05:28] [PASSED] 15 VFs [20:05:28] [PASSED] 16 VFs [20:05:28] [PASSED] 17 VFs [20:05:28] [PASSED] 18 VFs [20:05:28] [PASSED] 19 VFs [20:05:28] [PASSED] 20 VFs [20:05:28] [PASSED] 21 VFs [20:05:28] [PASSED] 22 VFs [20:05:28] [PASSED] 23 VFs [20:05:28] [PASSED] 24 VFs [20:05:28] [PASSED] 25 VFs [20:05:28] [PASSED] 26 VFs [20:05:28] [PASSED] 27 VFs [20:05:28] [PASSED] 28 VFs [20:05:28] [PASSED] 29 VFs [20:05:28] [PASSED] 30 VFs [20:05:28] [PASSED] 31 VFs [20:05:28] [PASSED] 32 VFs [20:05:28] [PASSED] 33 VFs [20:05:28] [PASSED] 34 VFs [20:05:28] [PASSED] 35 VFs [20:05:28] [PASSED] 36 VFs [20:05:28] [PASSED] 37 VFs [20:05:28] [PASSED] 38 VFs [20:05:28] [PASSED] 39 VFs [20:05:28] [PASSED] 40 VFs [20:05:28] [PASSED] 41 VFs [20:05:28] [PASSED] 42 VFs [20:05:28] [PASSED] 43 VFs [20:05:28] [PASSED] 44 VFs [20:05:28] [PASSED] 45 VFs [20:05:28] [PASSED] 46 VFs [20:05:28] [PASSED] 47 VFs [20:05:28] [PASSED] 48 VFs [20:05:28] [PASSED] 49 VFs [20:05:28] [PASSED] 50 VFs [20:05:28] [PASSED] 51 VFs [20:05:28] [PASSED] 52 VFs [20:05:28] [PASSED] 53 VFs [20:05:28] [PASSED] 54 VFs [20:05:28] [PASSED] 55 VFs [20:05:28] [PASSED] 56 VFs [20:05:28] [PASSED] 57 VFs [20:05:28] [PASSED] 58 VFs [20:05:28] [PASSED] 59 VFs [20:05:28] [PASSED] 60 VFs [20:05:28] [PASSED] 61 VFs [20:05:28] [PASSED] 62 VFs [20:05:28] [PASSED] 63 VFs [20:05:28] ==================== [PASSED] fair_ggtt ==================== [20:05:28] ======================== fair_vram ======================== [20:05:28] [PASSED] 1 VF [20:05:28] [PASSED] 2 VFs [20:05:28] [PASSED] 3 VFs [20:05:28] [PASSED] 4 VFs [20:05:28] [PASSED] 5 VFs [20:05:28] [PASSED] 6 VFs [20:05:28] [PASSED] 7 VFs [20:05:28] [PASSED] 8 VFs [20:05:28] [PASSED] 9 VFs [20:05:28] [PASSED] 10 VFs [20:05:28] [PASSED] 11 VFs [20:05:28] [PASSED] 12 VFs [20:05:28] [PASSED] 13 VFs [20:05:28] [PASSED] 14 VFs [20:05:28] [PASSED] 15 VFs [20:05:28] [PASSED] 16 VFs [20:05:28] [PASSED] 17 VFs [20:05:28] [PASSED] 18 VFs [20:05:28] [PASSED] 19 VFs [20:05:28] [PASSED] 20 VFs [20:05:28] [PASSED] 21 VFs [20:05:28] [PASSED] 22 VFs [20:05:28] [PASSED] 23 VFs [20:05:28] [PASSED] 24 VFs [20:05:28] [PASSED] 25 VFs [20:05:28] [PASSED] 26 VFs [20:05:28] [PASSED] 27 VFs [20:05:28] [PASSED] 28 VFs [20:05:28] [PASSED] 29 VFs [20:05:28] [PASSED] 30 VFs [20:05:28] [PASSED] 31 VFs [20:05:28] [PASSED] 32 VFs [20:05:28] [PASSED] 33 VFs [20:05:28] [PASSED] 34 VFs [20:05:28] [PASSED] 35 VFs [20:05:28] [PASSED] 36 VFs [20:05:28] [PASSED] 37 VFs [20:05:28] [PASSED] 38 VFs [20:05:28] [PASSED] 39 VFs [20:05:28] [PASSED] 40 VFs [20:05:28] [PASSED] 41 VFs [20:05:28] [PASSED] 42 VFs [20:05:28] [PASSED] 43 VFs [20:05:28] [PASSED] 44 VFs [20:05:28] [PASSED] 45 VFs [20:05:28] [PASSED] 46 VFs [20:05:28] [PASSED] 47 VFs [20:05:28] [PASSED] 48 VFs [20:05:28] [PASSED] 49 VFs [20:05:28] [PASSED] 50 VFs [20:05:28] [PASSED] 51 VFs [20:05:28] [PASSED] 52 VFs [20:05:28] [PASSED] 53 VFs [20:05:28] [PASSED] 54 VFs [20:05:28] [PASSED] 55 VFs [20:05:28] [PASSED] 56 VFs [20:05:28] [PASSED] 57 VFs [20:05:28] [PASSED] 58 VFs [20:05:28] [PASSED] 59 VFs [20:05:28] [PASSED] 60 VFs [20:05:28] [PASSED] 61 VFs [20:05:28] [PASSED] 62 VFs [20:05:28] [PASSED] 63 VFs [20:05:28] ==================== [PASSED] fair_vram ==================== [20:05:28] ================== [PASSED] pf_gt_config =================== [20:05:28] ===================== lmtt (1 subtest) ===================== [20:05:28] ======================== test_ops ========================= [20:05:28] [PASSED] 2-level [20:05:28] [PASSED] multi-level [20:05:28] ==================== [PASSED] test_ops ===================== [20:05:28] ====================== [PASSED] lmtt ======================= [20:05:28] ================= pf_service (11 subtests) ================= [20:05:28] [PASSED] pf_negotiate_any [20:05:28] [PASSED] pf_negotiate_base_match [20:05:28] [PASSED] pf_negotiate_base_newer [20:05:28] [PASSED] pf_negotiate_base_next [20:05:28] [SKIPPED] pf_negotiate_base_older [20:05:28] [PASSED] pf_negotiate_base_prev [20:05:28] [PASSED] pf_negotiate_latest_match [20:05:28] [PASSED] pf_negotiate_latest_newer [20:05:28] [PASSED] pf_negotiate_latest_next [20:05:28] [SKIPPED] pf_negotiate_latest_older [20:05:28] [SKIPPED] pf_negotiate_latest_prev [20:05:28] =================== [PASSED] pf_service ==================== [20:05:28] ================= xe_guc_g2g (2 subtests) ================== [20:05:28] ============== xe_live_guc_g2g_kunit_default ============== [20:05:28] ========= [SKIPPED] xe_live_guc_g2g_kunit_default ========== [20:05:28] ============== xe_live_guc_g2g_kunit_allmem =============== [20:05:28] ========== [SKIPPED] xe_live_guc_g2g_kunit_allmem ========== [20:05:28] =================== [SKIPPED] xe_guc_g2g =================== [20:05:28] =================== xe_mocs (2 subtests) =================== [20:05:28] ================ xe_live_mocs_kernel_kunit ================ [20:05:28] =========== [SKIPPED] xe_live_mocs_kernel_kunit ============ [20:05:28] ================ xe_live_mocs_reset_kunit ================= [20:05:28] ============ [SKIPPED] xe_live_mocs_reset_kunit ============ [20:05:28] ==================== [SKIPPED] xe_mocs ===================== [20:05:28] ================= xe_migrate (2 subtests) ================== [20:05:28] ================= xe_migrate_sanity_kunit ================= [20:05:28] ============ [SKIPPED] xe_migrate_sanity_kunit ============= [20:05:28] ================== xe_validate_ccs_kunit ================== [20:05:28] ============= [SKIPPED] xe_validate_ccs_kunit ============== [20:05:28] =================== [SKIPPED] xe_migrate =================== [20:05:28] ================== xe_dma_buf (1 subtest) ================== [20:05:28] ==================== xe_dma_buf_kunit ===================== [20:05:28] ================ [SKIPPED] xe_dma_buf_kunit ================ [20:05:28] =================== [SKIPPED] xe_dma_buf =================== [20:05:28] ================= xe_bo_shrink (1 subtest) ================= [20:05:28] =================== xe_bo_shrink_kunit ==================== [20:05:28] =============== [SKIPPED] xe_bo_shrink_kunit =============== [20:05:28] ================== [SKIPPED] xe_bo_shrink ================== [20:05:28] ==================== xe_bo (2 subtests) ==================== [20:05:28] ================== xe_ccs_migrate_kunit =================== [20:05:28] ============== [SKIPPED] xe_ccs_migrate_kunit ============== [20:05:28] ==================== xe_bo_evict_kunit ==================== [20:05:28] =============== [SKIPPED] xe_bo_evict_kunit ================ [20:05:28] ===================== [SKIPPED] xe_bo ====================== [20:05:28] ==================== args (13 subtests) ==================== [20:05:28] [PASSED] count_args_test [20:05:28] [PASSED] call_args_example [20:05:28] [PASSED] call_args_test [20:05:28] [PASSED] drop_first_arg_example [20:05:28] [PASSED] drop_first_arg_test [20:05:28] [PASSED] first_arg_example [20:05:28] [PASSED] first_arg_test [20:05:28] [PASSED] last_arg_example [20:05:28] [PASSED] last_arg_test [20:05:28] [PASSED] pick_arg_example [20:05:28] [PASSED] if_args_example [20:05:28] [PASSED] if_args_test [20:05:28] [PASSED] sep_comma_example [20:05:28] ====================== [PASSED] args ======================= [20:05:28] =================== xe_pci (3 subtests) ==================== [20:05:28] ==================== check_graphics_ip ==================== [20:05:28] [PASSED] 12.00 Xe_LP [20:05:28] [PASSED] 12.10 Xe_LP+ [20:05:28] [PASSED] 12.55 Xe_HPG [20:05:28] [PASSED] 12.60 Xe_HPC [20:05:28] [PASSED] 12.70 Xe_LPG [20:05:28] [PASSED] 12.71 Xe_LPG [20:05:28] [PASSED] 12.74 Xe_LPG+ [20:05:28] [PASSED] 20.01 Xe2_HPG [20:05:28] [PASSED] 20.02 Xe2_HPG [20:05:28] [PASSED] 20.04 Xe2_LPG [20:05:28] [PASSED] 30.00 Xe3_LPG [20:05:28] [PASSED] 30.01 Xe3_LPG [20:05:28] [PASSED] 30.03 Xe3_LPG [20:05:28] [PASSED] 30.04 Xe3_LPG [20:05:28] [PASSED] 30.05 Xe3_LPG [20:05:28] [PASSED] 35.10 Xe3p_LPG [20:05:28] [PASSED] 35.11 Xe3p_XPC [20:05:28] ================ [PASSED] check_graphics_ip ================ [20:05:28] ===================== check_media_ip ====================== [20:05:28] [PASSED] 12.00 Xe_M [20:05:28] [PASSED] 12.55 Xe_HPM [20:05:28] [PASSED] 13.00 Xe_LPM+ [20:05:28] [PASSED] 13.01 Xe2_HPM [20:05:28] [PASSED] 20.00 Xe2_LPM [20:05:28] [PASSED] 30.00 Xe3_LPM [20:05:28] [PASSED] 30.02 Xe3_LPM [20:05:28] [PASSED] 35.00 Xe3p_LPM [20:05:28] [PASSED] 35.03 Xe3p_HPM [20:05:28] ================= [PASSED] check_media_ip ================== [20:05:28] =================== check_platform_desc =================== [20:05:28] [PASSED] 0x9A60 (TIGERLAKE) [20:05:28] [PASSED] 0x9A68 (TIGERLAKE) [20:05:28] [PASSED] 0x9A70 (TIGERLAKE) [20:05:28] [PASSED] 0x9A40 (TIGERLAKE) [20:05:28] [PASSED] 0x9A49 (TIGERLAKE) [20:05:28] [PASSED] 0x9A59 (TIGERLAKE) [20:05:28] [PASSED] 0x9A78 (TIGERLAKE) [20:05:28] [PASSED] 0x9AC0 (TIGERLAKE) [20:05:28] [PASSED] 0x9AC9 (TIGERLAKE) [20:05:28] [PASSED] 0x9AD9 (TIGERLAKE) [20:05:28] [PASSED] 0x9AF8 (TIGERLAKE) [20:05:28] [PASSED] 0x4C80 (ROCKETLAKE) [20:05:28] [PASSED] 0x4C8A (ROCKETLAKE) [20:05:28] [PASSED] 0x4C8B (ROCKETLAKE) [20:05:28] [PASSED] 0x4C8C (ROCKETLAKE) [20:05:28] [PASSED] 0x4C90 (ROCKETLAKE) [20:05:28] [PASSED] 0x4C9A (ROCKETLAKE) [20:05:28] [PASSED] 0x4680 (ALDERLAKE_S) [20:05:28] [PASSED] 0x4682 (ALDERLAKE_S) [20:05:28] [PASSED] 0x4688 (ALDERLAKE_S) [20:05:28] [PASSED] 0x468A (ALDERLAKE_S) [20:05:28] [PASSED] 0x468B (ALDERLAKE_S) [20:05:28] [PASSED] 0x4690 (ALDERLAKE_S) [20:05:28] [PASSED] 0x4692 (ALDERLAKE_S) [20:05:28] [PASSED] 0x4693 (ALDERLAKE_S) [20:05:28] [PASSED] 0x46A0 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46A1 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46A2 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46A3 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46A6 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46A8 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46AA (ALDERLAKE_P) [20:05:28] [PASSED] 0x462A (ALDERLAKE_P) [20:05:28] [PASSED] 0x4626 (ALDERLAKE_P) [20:05:28] [PASSED] 0x4628 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46B0 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46B1 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46B2 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46B3 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46C0 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46C1 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46C2 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46C3 (ALDERLAKE_P) [20:05:28] [PASSED] 0x46D0 (ALDERLAKE_N) [20:05:28] [PASSED] 0x46D1 (ALDERLAKE_N) [20:05:28] [PASSED] 0x46D2 (ALDERLAKE_N) [20:05:28] [PASSED] 0x46D3 (ALDERLAKE_N) [20:05:28] [PASSED] 0x46D4 (ALDERLAKE_N) [20:05:28] [PASSED] 0xA721 (ALDERLAKE_P) [20:05:28] [PASSED] 0xA7A1 (ALDERLAKE_P) [20:05:28] [PASSED] 0xA7A9 (ALDERLAKE_P) [20:05:28] [PASSED] 0xA7AC (ALDERLAKE_P) [20:05:28] [PASSED] 0xA7AD (ALDERLAKE_P) [20:05:28] [PASSED] 0xA720 (ALDERLAKE_P) [20:05:28] [PASSED] 0xA7A0 (ALDERLAKE_P) [20:05:28] [PASSED] 0xA7A8 (ALDERLAKE_P) [20:05:28] [PASSED] 0xA7AA (ALDERLAKE_P) [20:05:28] [PASSED] 0xA7AB (ALDERLAKE_P) [20:05:28] [PASSED] 0xA780 (ALDERLAKE_S) [20:05:28] [PASSED] 0xA781 (ALDERLAKE_S) [20:05:28] [PASSED] 0xA782 (ALDERLAKE_S) [20:05:28] [PASSED] 0xA783 (ALDERLAKE_S) [20:05:28] [PASSED] 0xA788 (ALDERLAKE_S) [20:05:28] [PASSED] 0xA789 (ALDERLAKE_S) [20:05:28] [PASSED] 0xA78A (ALDERLAKE_S) [20:05:28] [PASSED] 0xA78B (ALDERLAKE_S) [20:05:28] [PASSED] 0x4905 (DG1) [20:05:28] [PASSED] 0x4906 (DG1) [20:05:28] [PASSED] 0x4907 (DG1) [20:05:28] [PASSED] 0x4908 (DG1) [20:05:28] [PASSED] 0x4909 (DG1) [20:05:28] [PASSED] 0x56C0 (DG2) [20:05:28] [PASSED] 0x56C2 (DG2) [20:05:28] [PASSED] 0x56C1 (DG2) [20:05:28] [PASSED] 0x7D51 (METEORLAKE) [20:05:28] [PASSED] 0x7DD1 (METEORLAKE) [20:05:28] [PASSED] 0x7D41 (METEORLAKE) [20:05:28] [PASSED] 0x7D67 (METEORLAKE) [20:05:28] [PASSED] 0xB640 (METEORLAKE) [20:05:28] [PASSED] 0x56A0 (DG2) [20:05:28] [PASSED] 0x56A1 (DG2) [20:05:28] [PASSED] 0x56A2 (DG2) [20:05:28] [PASSED] 0x56BE (DG2) [20:05:28] [PASSED] 0x56BF (DG2) [20:05:28] [PASSED] 0x5690 (DG2) [20:05:28] [PASSED] 0x5691 (DG2) [20:05:28] [PASSED] 0x5692 (DG2) [20:05:28] [PASSED] 0x56A5 (DG2) [20:05:28] [PASSED] 0x56A6 (DG2) [20:05:28] [PASSED] 0x56B0 (DG2) [20:05:28] [PASSED] 0x56B1 (DG2) [20:05:28] [PASSED] 0x56BA (DG2) [20:05:28] [PASSED] 0x56BB (DG2) [20:05:28] [PASSED] 0x56BC (DG2) [20:05:28] [PASSED] 0x56BD (DG2) [20:05:28] [PASSED] 0x5693 (DG2) [20:05:28] [PASSED] 0x5694 (DG2) [20:05:28] [PASSED] 0x5695 (DG2) [20:05:28] [PASSED] 0x56A3 (DG2) [20:05:28] [PASSED] 0x56A4 (DG2) [20:05:28] [PASSED] 0x56B2 (DG2) [20:05:28] [PASSED] 0x56B3 (DG2) [20:05:28] [PASSED] 0x5696 (DG2) [20:05:28] [PASSED] 0x5697 (DG2) [20:05:28] [PASSED] 0xB69 (PVC) [20:05:28] [PASSED] 0xB6E (PVC) [20:05:28] [PASSED] 0xBD4 (PVC) [20:05:28] [PASSED] 0xBD5 (PVC) [20:05:28] [PASSED] 0xBD6 (PVC) [20:05:28] [PASSED] 0xBD7 (PVC) [20:05:28] [PASSED] 0xBD8 (PVC) [20:05:28] [PASSED] 0xBD9 (PVC) [20:05:28] [PASSED] 0xBDA (PVC) [20:05:28] [PASSED] 0xBDB (PVC) [20:05:28] [PASSED] 0xBE0 (PVC) [20:05:28] [PASSED] 0xBE1 (PVC) [20:05:28] [PASSED] 0xBE5 (PVC) [20:05:28] [PASSED] 0x7D40 (METEORLAKE) [20:05:28] [PASSED] 0x7D45 (METEORLAKE) [20:05:28] [PASSED] 0x7D55 (METEORLAKE) [20:05:28] [PASSED] 0x7D60 (METEORLAKE) [20:05:28] [PASSED] 0x7DD5 (METEORLAKE) [20:05:28] [PASSED] 0x6420 (LUNARLAKE) [20:05:28] [PASSED] 0x64A0 (LUNARLAKE) [20:05:28] [PASSED] 0x64B0 (LUNARLAKE) [20:05:28] [PASSED] 0xE202 (BATTLEMAGE) [20:05:28] [PASSED] 0xE209 (BATTLEMAGE) [20:05:28] [PASSED] 0xE20B (BATTLEMAGE) [20:05:28] [PASSED] 0xE20C (BATTLEMAGE) [20:05:28] [PASSED] 0xE20D (BATTLEMAGE) [20:05:28] [PASSED] 0xE210 (BATTLEMAGE) [20:05:28] [PASSED] 0xE211 (BATTLEMAGE) [20:05:28] [PASSED] 0xE212 (BATTLEMAGE) [20:05:28] [PASSED] 0xE216 (BATTLEMAGE) [20:05:28] [PASSED] 0xE220 (BATTLEMAGE) [20:05:28] [PASSED] 0xE221 (BATTLEMAGE) [20:05:28] [PASSED] 0xE222 (BATTLEMAGE) [20:05:28] [PASSED] 0xE223 (BATTLEMAGE) [20:05:28] [PASSED] 0xB080 (PANTHERLAKE) [20:05:28] [PASSED] 0xB081 (PANTHERLAKE) [20:05:28] [PASSED] 0xB082 (PANTHERLAKE) [20:05:28] [PASSED] 0xB083 (PANTHERLAKE) [20:05:28] [PASSED] 0xB084 (PANTHERLAKE) [20:05:28] [PASSED] 0xB085 (PANTHERLAKE) [20:05:28] [PASSED] 0xB086 (PANTHERLAKE) [20:05:28] [PASSED] 0xB087 (PANTHERLAKE) [20:05:28] [PASSED] 0xB08F (PANTHERLAKE) [20:05:28] [PASSED] 0xB090 (PANTHERLAKE) [20:05:28] [PASSED] 0xB0A0 (PANTHERLAKE) [20:05:28] [PASSED] 0xB0B0 (PANTHERLAKE) [20:05:28] [PASSED] 0xFD80 (PANTHERLAKE) [20:05:28] [PASSED] 0xFD81 (PANTHERLAKE) [20:05:28] [PASSED] 0xD740 (NOVALAKE_S) [20:05:28] [PASSED] 0xD741 (NOVALAKE_S) [20:05:28] [PASSED] 0xD742 (NOVALAKE_S) [20:05:28] [PASSED] 0xD743 (NOVALAKE_S) [20:05:28] [PASSED] 0xD744 (NOVALAKE_S) [20:05:28] [PASSED] 0xD745 (NOVALAKE_S) [20:05:28] [PASSED] 0x674C (CRESCENTISLAND) [20:05:28] [PASSED] 0xD750 (NOVALAKE_P) [20:05:28] [PASSED] 0xD751 (NOVALAKE_P) [20:05:28] [PASSED] 0xD752 (NOVALAKE_P) [20:05:28] [PASSED] 0xD753 (NOVALAKE_P) [20:05:28] [PASSED] 0xD754 (NOVALAKE_P) [20:05:28] [PASSED] 0xD755 (NOVALAKE_P) [20:05:28] [PASSED] 0xD756 (NOVALAKE_P) [20:05:28] [PASSED] 0xD757 (NOVALAKE_P) [20:05:28] [PASSED] 0xD75F (NOVALAKE_P) [20:05:28] =============== [PASSED] check_platform_desc =============== [20:05:28] ===================== [PASSED] xe_pci ====================== [20:05:28] =================== xe_rtp (2 subtests) ==================== [20:05:28] =============== xe_rtp_process_to_sr_tests ================ [20:05:28] [PASSED] coalesce-same-reg [20:05:28] [PASSED] no-match-no-add [20:05:28] [PASSED] match-or [20:05:28] [PASSED] match-or-xfail [20:05:28] [PASSED] no-match-no-add-multiple-rules [20:05:28] [PASSED] two-regs-two-entries [20:05:28] [PASSED] clr-one-set-other [20:05:28] [PASSED] set-field [20:05:28] [PASSED] conflict-duplicate [20:05:28] [PASSED] conflict-not-disjoint [20:05:28] [PASSED] conflict-reg-type [20:05:28] =========== [PASSED] xe_rtp_process_to_sr_tests ============ [20:05:28] ================== xe_rtp_process_tests =================== [20:05:28] [PASSED] active1 [20:05:28] [PASSED] active2 [20:05:28] [PASSED] active-inactive [20:05:28] [PASSED] inactive-active [20:05:28] [PASSED] inactive-1st_or_active-inactive [20:05:28] [PASSED] inactive-2nd_or_active-inactive [20:05:28] [PASSED] inactive-last_or_active-inactive [20:05:28] [PASSED] inactive-no_or_active-inactive [20:05:28] ============== [PASSED] xe_rtp_process_tests =============== [20:05:28] ===================== [PASSED] xe_rtp ====================== [20:05:28] ==================== xe_wa (1 subtest) ===================== [20:05:28] ======================== xe_wa_gt ========================= [20:05:28] [PASSED] TIGERLAKE B0 [20:05:28] [PASSED] DG1 A0 [20:05:28] [PASSED] DG1 B0 [20:05:28] [PASSED] ALDERLAKE_S A0 [20:05:28] [PASSED] ALDERLAKE_S B0 [20:05:28] [PASSED] ALDERLAKE_S C0 [20:05:28] [PASSED] ALDERLAKE_S D0 [20:05:28] [PASSED] ALDERLAKE_P A0 [20:05:28] [PASSED] ALDERLAKE_P B0 [20:05:28] [PASSED] ALDERLAKE_P C0 [20:05:28] [PASSED] ALDERLAKE_S RPLS D0 [20:05:28] [PASSED] ALDERLAKE_P RPLU E0 [20:05:28] [PASSED] DG2 G10 C0 [20:05:28] [PASSED] DG2 G11 B1 [20:05:28] [PASSED] DG2 G12 A1 [20:05:28] [PASSED] METEORLAKE 12.70(Xe_LPG) A0 13.00(Xe_LPM+) A0 [20:05:28] [PASSED] METEORLAKE 12.71(Xe_LPG) A0 13.00(Xe_LPM+) A0 [20:05:28] [PASSED] METEORLAKE 12.74(Xe_LPG+) A0 13.00(Xe_LPM+) A0 [20:05:28] [PASSED] LUNARLAKE 20.04(Xe2_LPG) A0 20.00(Xe2_LPM) A0 [20:05:28] [PASSED] LUNARLAKE 20.04(Xe2_LPG) B0 20.00(Xe2_LPM) A0 [20:05:28] [PASSED] BATTLEMAGE 20.01(Xe2_HPG) A0 13.01(Xe2_HPM) A1 [20:05:28] [PASSED] PANTHERLAKE 30.00(Xe3_LPG) A0 30.00(Xe3_LPM) A0 [20:05:28] ==================== [PASSED] xe_wa_gt ===================== [20:05:28] ====================== [PASSED] xe_wa ====================== [20:05:28] ============================================================ [20:05:28] Testing complete. Ran 597 tests: passed: 579, skipped: 18 [20:05:28] Elapsed time: 35.990s total, 4.282s configuring, 31.092s building, 0.597s running
+ /kernel/tools/testing/kunit/kunit.py run --kunitconfig /kernel/drivers/gpu/drm/tests/.kunitconfig [20:05:28] Configuring KUnit Kernel ...
Regenerating .config ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig [20:05:30] 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 [20:05:54] Starting KUnit Kernel (1/1)... [20:05:54] ============================================================ Running tests with: $ .kunit/linux kunit.enable=1 mem=1G console=tty kunit_shutdown=halt [20:05:54] ============ drm_test_pick_cmdline (2 subtests) ============ [20:05:54] [PASSED] drm_test_pick_cmdline_res_1920_1080_60 [20:05:54] =============== drm_test_pick_cmdline_named =============== [20:05:54] [PASSED] NTSC [20:05:54] [PASSED] NTSC-J [20:05:54] [PASSED] PAL [20:05:54] [PASSED] PAL-M [20:05:54] =========== [PASSED] drm_test_pick_cmdline_named =========== [20:05:54] ============== [PASSED] drm_test_pick_cmdline ============== [20:05:54] == drm_test_atomic_get_connector_for_encoder (1 subtest) === [20:05:54] [PASSED] drm_test_drm_atomic_get_connector_for_encoder [20:05:54] ==== [PASSED] drm_test_atomic_get_connector_for_encoder ==== [20:05:54] =========== drm_validate_clone_mode (2 subtests) =========== [20:05:54] ============== drm_test_check_in_clone_mode =============== [20:05:54] [PASSED] in_clone_mode [20:05:54] [PASSED] not_in_clone_mode [20:05:54] ========== [PASSED] drm_test_check_in_clone_mode =========== [20:05:54] =============== drm_test_check_valid_clones =============== [20:05:54] [PASSED] not_in_clone_mode [20:05:54] [PASSED] valid_clone [20:05:54] [PASSED] invalid_clone [20:05:54] =========== [PASSED] drm_test_check_valid_clones =========== [20:05:54] ============= [PASSED] drm_validate_clone_mode ============= [20:05:54] ============= drm_validate_modeset (1 subtest) ============= [20:05:54] [PASSED] drm_test_check_connector_changed_modeset [20:05:54] ============== [PASSED] drm_validate_modeset =============== [20:05:54] ====== drm_test_bridge_get_current_state (2 subtests) ====== [20:05:54] [PASSED] drm_test_drm_bridge_get_current_state_atomic [20:05:54] [PASSED] drm_test_drm_bridge_get_current_state_legacy [20:05:54] ======== [PASSED] drm_test_bridge_get_current_state ======== [20:05:54] ====== drm_test_bridge_helper_reset_crtc (3 subtests) ====== [20:05:54] [PASSED] drm_test_drm_bridge_helper_reset_crtc_atomic [20:05:54] [PASSED] drm_test_drm_bridge_helper_reset_crtc_atomic_disabled [20:05:54] [PASSED] drm_test_drm_bridge_helper_reset_crtc_legacy [20:05:54] ======== [PASSED] drm_test_bridge_helper_reset_crtc ======== [20:05:54] ============== drm_bridge_alloc (2 subtests) =============== [20:05:54] [PASSED] drm_test_drm_bridge_alloc_basic [20:05:54] [PASSED] drm_test_drm_bridge_alloc_get_put [20:05:54] ================ [PASSED] drm_bridge_alloc ================= [20:05:54] ============= drm_cmdline_parser (40 subtests) ============= [20:05:54] [PASSED] drm_test_cmdline_force_d_only [20:05:54] [PASSED] drm_test_cmdline_force_D_only_dvi [20:05:54] [PASSED] drm_test_cmdline_force_D_only_hdmi [20:05:54] [PASSED] drm_test_cmdline_force_D_only_not_digital [20:05:54] [PASSED] drm_test_cmdline_force_e_only [20:05:54] [PASSED] drm_test_cmdline_res [20:05:54] [PASSED] drm_test_cmdline_res_vesa [20:05:54] [PASSED] drm_test_cmdline_res_vesa_rblank [20:05:54] [PASSED] drm_test_cmdline_res_rblank [20:05:54] [PASSED] drm_test_cmdline_res_bpp [20:05:54] [PASSED] drm_test_cmdline_res_refresh [20:05:54] [PASSED] drm_test_cmdline_res_bpp_refresh [20:05:54] [PASSED] drm_test_cmdline_res_bpp_refresh_interlaced [20:05:54] [PASSED] drm_test_cmdline_res_bpp_refresh_margins [20:05:54] [PASSED] drm_test_cmdline_res_bpp_refresh_force_off [20:05:54] [PASSED] drm_test_cmdline_res_bpp_refresh_force_on [20:05:54] [PASSED] drm_test_cmdline_res_bpp_refresh_force_on_analog [20:05:54] [PASSED] drm_test_cmdline_res_bpp_refresh_force_on_digital [20:05:54] [PASSED] drm_test_cmdline_res_bpp_refresh_interlaced_margins_force_on [20:05:54] [PASSED] drm_test_cmdline_res_margins_force_on [20:05:54] [PASSED] drm_test_cmdline_res_vesa_margins [20:05:54] [PASSED] drm_test_cmdline_name [20:05:54] [PASSED] drm_test_cmdline_name_bpp [20:05:54] [PASSED] drm_test_cmdline_name_option [20:05:54] [PASSED] drm_test_cmdline_name_bpp_option [20:05:54] [PASSED] drm_test_cmdline_rotate_0 [20:05:54] [PASSED] drm_test_cmdline_rotate_90 [20:05:54] [PASSED] drm_test_cmdline_rotate_180 [20:05:54] [PASSED] drm_test_cmdline_rotate_270 [20:05:54] [PASSED] drm_test_cmdline_hmirror [20:05:54] [PASSED] drm_test_cmdline_vmirror [20:05:54] [PASSED] drm_test_cmdline_margin_options [20:05:54] [PASSED] drm_test_cmdline_multiple_options [20:05:54] [PASSED] drm_test_cmdline_bpp_extra_and_option [20:05:54] [PASSED] drm_test_cmdline_extra_and_option [20:05:54] [PASSED] drm_test_cmdline_freestanding_options [20:05:54] [PASSED] drm_test_cmdline_freestanding_force_e_and_options [20:05:54] [PASSED] drm_test_cmdline_panel_orientation [20:05:54] ================ drm_test_cmdline_invalid ================= [20:05:54] [PASSED] margin_only [20:05:54] [PASSED] interlace_only [20:05:54] [PASSED] res_missing_x [20:05:54] [PASSED] res_missing_y [20:05:54] [PASSED] res_bad_y [20:05:54] [PASSED] res_missing_y_bpp [20:05:54] [PASSED] res_bad_bpp [20:05:54] [PASSED] res_bad_refresh [20:05:54] [PASSED] res_bpp_refresh_force_on_off [20:05:54] [PASSED] res_invalid_mode [20:05:54] [PASSED] res_bpp_wrong_place_mode [20:05:54] [PASSED] name_bpp_refresh [20:05:54] [PASSED] name_refresh [20:05:54] [PASSED] name_refresh_wrong_mode [20:05:54] [PASSED] name_refresh_invalid_mode [20:05:54] [PASSED] rotate_multiple [20:05:54] [PASSED] rotate_invalid_val [20:05:54] [PASSED] rotate_truncated [20:05:54] [PASSED] invalid_option [20:05:54] [PASSED] invalid_tv_option [20:05:54] [PASSED] truncated_tv_option [20:05:54] ============ [PASSED] drm_test_cmdline_invalid ============= [20:05:54] =============== drm_test_cmdline_tv_options =============== [20:05:54] [PASSED] NTSC [20:05:54] [PASSED] NTSC_443 [20:05:54] [PASSED] NTSC_J [20:05:54] [PASSED] PAL [20:05:54] [PASSED] PAL_M [20:05:54] [PASSED] PAL_N [20:05:54] [PASSED] SECAM [20:05:54] [PASSED] MONO_525 [20:05:54] [PASSED] MONO_625 [20:05:54] =========== [PASSED] drm_test_cmdline_tv_options =========== [20:05:54] =============== [PASSED] drm_cmdline_parser ================ [20:05:54] ========== drmm_connector_hdmi_init (20 subtests) ========== [20:05:54] [PASSED] drm_test_connector_hdmi_init_valid [20:05:54] [PASSED] drm_test_connector_hdmi_init_bpc_8 [20:05:54] [PASSED] drm_test_connector_hdmi_init_bpc_10 [20:05:54] [PASSED] drm_test_connector_hdmi_init_bpc_12 [20:05:54] [PASSED] drm_test_connector_hdmi_init_bpc_invalid [20:05:54] [PASSED] drm_test_connector_hdmi_init_bpc_null [20:05:54] [PASSED] drm_test_connector_hdmi_init_formats_empty [20:05:54] [PASSED] drm_test_connector_hdmi_init_formats_no_rgb [20:05:54] === drm_test_connector_hdmi_init_formats_yuv420_allowed === [20:05:54] [PASSED] supported_formats=0x9 yuv420_allowed=1 [20:05:54] [PASSED] supported_formats=0x9 yuv420_allowed=0 [20:05:54] [PASSED] supported_formats=0x5 yuv420_allowed=1 [20:05:54] [PASSED] supported_formats=0x5 yuv420_allowed=0 [20:05:54] === [PASSED] drm_test_connector_hdmi_init_formats_yuv420_allowed === [20:05:54] [PASSED] drm_test_connector_hdmi_init_null_ddc [20:05:54] [PASSED] drm_test_connector_hdmi_init_null_product [20:05:54] [PASSED] drm_test_connector_hdmi_init_null_vendor [20:05:54] [PASSED] drm_test_connector_hdmi_init_product_length_exact [20:05:54] [PASSED] drm_test_connector_hdmi_init_product_length_too_long [20:05:54] [PASSED] drm_test_connector_hdmi_init_product_valid [20:05:54] [PASSED] drm_test_connector_hdmi_init_vendor_length_exact [20:05:54] [PASSED] drm_test_connector_hdmi_init_vendor_length_too_long [20:05:54] [PASSED] drm_test_connector_hdmi_init_vendor_valid [20:05:54] ========= drm_test_connector_hdmi_init_type_valid ========= [20:05:54] [PASSED] HDMI-A [20:05:54] [PASSED] HDMI-B [20:05:54] ===== [PASSED] drm_test_connector_hdmi_init_type_valid ===== [20:05:54] ======== drm_test_connector_hdmi_init_type_invalid ======== [20:05:54] [PASSED] Unknown [20:05:54] [PASSED] VGA [20:05:54] [PASSED] DVI-I [20:05:54] [PASSED] DVI-D [20:05:54] [PASSED] DVI-A [20:05:54] [PASSED] Composite [20:05:54] [PASSED] SVIDEO [20:05:54] [PASSED] LVDS [20:05:54] [PASSED] Component [20:05:54] [PASSED] DIN [20:05:54] [PASSED] DP [20:05:54] [PASSED] TV [20:05:54] [PASSED] eDP [20:05:54] [PASSED] Virtual [20:05:54] [PASSED] DSI [20:05:54] [PASSED] DPI [20:05:54] [PASSED] Writeback [20:05:54] [PASSED] SPI [20:05:54] [PASSED] USB [20:05:54] ==== [PASSED] drm_test_connector_hdmi_init_type_invalid ==== [20:05:54] ============ [PASSED] drmm_connector_hdmi_init ============= [20:05:54] ============= drmm_connector_init (3 subtests) ============= [20:05:54] [PASSED] drm_test_drmm_connector_init [20:05:54] [PASSED] drm_test_drmm_connector_init_null_ddc [20:05:54] ========= drm_test_drmm_connector_init_type_valid ========= [20:05:54] [PASSED] Unknown [20:05:54] [PASSED] VGA [20:05:54] [PASSED] DVI-I [20:05:54] [PASSED] DVI-D [20:05:54] [PASSED] DVI-A [20:05:54] [PASSED] Composite [20:05:54] [PASSED] SVIDEO [20:05:54] [PASSED] LVDS [20:05:54] [PASSED] Component [20:05:54] [PASSED] DIN [20:05:54] [PASSED] DP [20:05:54] [PASSED] HDMI-A [20:05:54] [PASSED] HDMI-B [20:05:54] [PASSED] TV [20:05:54] [PASSED] eDP [20:05:54] [PASSED] Virtual [20:05:54] [PASSED] DSI [20:05:54] [PASSED] DPI [20:05:54] [PASSED] Writeback [20:05:54] [PASSED] SPI [20:05:54] [PASSED] USB [20:05:54] ===== [PASSED] drm_test_drmm_connector_init_type_valid ===== [20:05:54] =============== [PASSED] drmm_connector_init =============== [20:05:54] ========= drm_connector_dynamic_init (6 subtests) ========== [20:05:54] [PASSED] drm_test_drm_connector_dynamic_init [20:05:54] [PASSED] drm_test_drm_connector_dynamic_init_null_ddc [20:05:54] [PASSED] drm_test_drm_connector_dynamic_init_not_added [20:05:54] [PASSED] drm_test_drm_connector_dynamic_init_properties [20:05:54] ===== drm_test_drm_connector_dynamic_init_type_valid ====== [20:05:54] [PASSED] Unknown [20:05:54] [PASSED] VGA [20:05:54] [PASSED] DVI-I [20:05:54] [PASSED] DVI-D [20:05:54] [PASSED] DVI-A [20:05:54] [PASSED] Composite [20:05:54] [PASSED] SVIDEO [20:05:54] [PASSED] LVDS [20:05:54] [PASSED] Component [20:05:54] [PASSED] DIN [20:05:54] [PASSED] DP [20:05:54] [PASSED] HDMI-A [20:05:54] [PASSED] HDMI-B [20:05:54] [PASSED] TV [20:05:54] [PASSED] eDP [20:05:54] [PASSED] Virtual [20:05:54] [PASSED] DSI [20:05:54] [PASSED] DPI [20:05:54] [PASSED] Writeback [20:05:54] [PASSED] SPI [20:05:54] [PASSED] USB [20:05:54] = [PASSED] drm_test_drm_connector_dynamic_init_type_valid == [20:05:54] ======== drm_test_drm_connector_dynamic_init_name ========= [20:05:54] [PASSED] Unknown [20:05:54] [PASSED] VGA [20:05:54] [PASSED] DVI-I [20:05:54] [PASSED] DVI-D [20:05:54] [PASSED] DVI-A [20:05:54] [PASSED] Composite [20:05:54] [PASSED] SVIDEO [20:05:54] [PASSED] LVDS [20:05:54] [PASSED] Component [20:05:54] [PASSED] DIN [20:05:54] [PASSED] DP [20:05:54] [PASSED] HDMI-A [20:05:54] [PASSED] HDMI-B [20:05:54] [PASSED] TV [20:05:54] [PASSED] eDP [20:05:54] [PASSED] Virtual [20:05:54] [PASSED] DSI [20:05:54] [PASSED] DPI [20:05:54] [PASSED] Writeback [20:05:54] [PASSED] SPI [20:05:54] [PASSED] USB [20:05:54] ==== [PASSED] drm_test_drm_connector_dynamic_init_name ===== [20:05:54] =========== [PASSED] drm_connector_dynamic_init ============ [20:05:54] ==== drm_connector_dynamic_register_early (4 subtests) ===== [20:05:54] [PASSED] drm_test_drm_connector_dynamic_register_early_on_list [20:05:54] [PASSED] drm_test_drm_connector_dynamic_register_early_defer [20:05:54] [PASSED] drm_test_drm_connector_dynamic_register_early_no_init [20:05:54] [PASSED] drm_test_drm_connector_dynamic_register_early_no_mode_object [20:05:54] ====== [PASSED] drm_connector_dynamic_register_early ======= [20:05:54] ======= drm_connector_dynamic_register (7 subtests) ======== [20:05:54] [PASSED] drm_test_drm_connector_dynamic_register_on_list [20:05:54] [PASSED] drm_test_drm_connector_dynamic_register_no_defer [20:05:54] [PASSED] drm_test_drm_connector_dynamic_register_no_init [20:05:54] [PASSED] drm_test_drm_connector_dynamic_register_mode_object [20:05:54] [PASSED] drm_test_drm_connector_dynamic_register_sysfs [20:05:54] [PASSED] drm_test_drm_connector_dynamic_register_sysfs_name [20:05:54] [PASSED] drm_test_drm_connector_dynamic_register_debugfs [20:05:54] ========= [PASSED] drm_connector_dynamic_register ========== [20:05:54] = drm_connector_attach_broadcast_rgb_property (2 subtests) = [20:05:54] [PASSED] drm_test_drm_connector_attach_broadcast_rgb_property [20:05:54] [PASSED] drm_test_drm_connector_attach_broadcast_rgb_property_hdmi_connector [20:05:54] === [PASSED] drm_connector_attach_broadcast_rgb_property === [20:05:54] ========== drm_get_tv_mode_from_name (2 subtests) ========== [20:05:54] ========== drm_test_get_tv_mode_from_name_valid =========== [20:05:54] [PASSED] NTSC [20:05:54] [PASSED] NTSC-443 [20:05:54] [PASSED] NTSC-J [20:05:54] [PASSED] PAL [20:05:54] [PASSED] PAL-M [20:05:54] [PASSED] PAL-N [20:05:54] [PASSED] SECAM [20:05:54] [PASSED] Mono [20:05:54] ====== [PASSED] drm_test_get_tv_mode_from_name_valid ======= [20:05:54] [PASSED] drm_test_get_tv_mode_from_name_truncated [20:05:54] ============ [PASSED] drm_get_tv_mode_from_name ============ [20:05:54] = drm_test_connector_hdmi_compute_mode_clock (12 subtests) = [20:05:54] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb [20:05:54] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_10bpc [20:05:54] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_10bpc_vic_1 [20:05:54] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_12bpc [20:05:54] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_12bpc_vic_1 [20:05:54] [PASSED] drm_test_drm_hdmi_compute_mode_clock_rgb_double [20:05:54] = drm_test_connector_hdmi_compute_mode_clock_yuv420_valid = [20:05:54] [PASSED] VIC 96 [20:05:54] [PASSED] VIC 97 [20:05:54] [PASSED] VIC 101 [20:05:54] [PASSED] VIC 102 [20:05:54] [PASSED] VIC 106 [20:05:54] [PASSED] VIC 107 [20:05:54] === [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv420_valid === [20:05:54] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv420_10_bpc [20:05:54] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv420_12_bpc [20:05:54] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv422_8_bpc [20:05:54] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv422_10_bpc [20:05:54] [PASSED] drm_test_connector_hdmi_compute_mode_clock_yuv422_12_bpc [20:05:54] === [PASSED] drm_test_connector_hdmi_compute_mode_clock ==== [20:05:54] == drm_hdmi_connector_get_broadcast_rgb_name (2 subtests) == [20:05:54] === drm_test_drm_hdmi_connector_get_broadcast_rgb_name ==== [20:05:54] [PASSED] Automatic [20:05:54] [PASSED] Full [20:05:54] [PASSED] Limited 16:235 [20:05:54] === [PASSED] drm_test_drm_hdmi_connector_get_broadcast_rgb_name === [20:05:54] [PASSED] drm_test_drm_hdmi_connector_get_broadcast_rgb_name_invalid [20:05:54] ==== [PASSED] drm_hdmi_connector_get_broadcast_rgb_name ==== [20:05:54] == drm_hdmi_connector_get_output_format_name (2 subtests) == [20:05:54] === drm_test_drm_hdmi_connector_get_output_format_name ==== [20:05:54] [PASSED] RGB [20:05:54] [PASSED] YUV 4:2:0 [20:05:54] [PASSED] YUV 4:2:2 [20:05:54] [PASSED] YUV 4:4:4 [20:05:54] === [PASSED] drm_test_drm_hdmi_connector_get_output_format_name === [20:05:54] [PASSED] drm_test_drm_hdmi_connector_get_output_format_name_invalid [20:05:54] ==== [PASSED] drm_hdmi_connector_get_output_format_name ==== [20:05:54] ============= drm_damage_helper (21 subtests) ============== [20:05:54] [PASSED] drm_test_damage_iter_no_damage [20:05:54] [PASSED] drm_test_damage_iter_no_damage_fractional_src [20:05:54] [PASSED] drm_test_damage_iter_no_damage_src_moved [20:05:54] [PASSED] drm_test_damage_iter_no_damage_fractional_src_moved [20:05:54] [PASSED] drm_test_damage_iter_no_damage_not_visible [20:05:54] [PASSED] drm_test_damage_iter_no_damage_no_crtc [20:05:54] [PASSED] drm_test_damage_iter_no_damage_no_fb [20:05:54] [PASSED] drm_test_damage_iter_simple_damage [20:05:54] [PASSED] drm_test_damage_iter_single_damage [20:05:54] [PASSED] drm_test_damage_iter_single_damage_intersect_src [20:05:54] [PASSED] drm_test_damage_iter_single_damage_outside_src [20:05:54] [PASSED] drm_test_damage_iter_single_damage_fractional_src [20:05:54] [PASSED] drm_test_damage_iter_single_damage_intersect_fractional_src [20:05:54] [PASSED] drm_test_damage_iter_single_damage_outside_fractional_src [20:05:54] [PASSED] drm_test_damage_iter_single_damage_src_moved [20:05:54] [PASSED] drm_test_damage_iter_single_damage_fractional_src_moved [20:05:54] [PASSED] drm_test_damage_iter_damage [20:05:54] [PASSED] drm_test_damage_iter_damage_one_intersect [20:05:54] [PASSED] drm_test_damage_iter_damage_one_outside [20:05:54] [PASSED] drm_test_damage_iter_damage_src_moved [20:05:54] [PASSED] drm_test_damage_iter_damage_not_visible [20:05:54] ================ [PASSED] drm_damage_helper ================ [20:05:54] ============== drm_dp_mst_helper (3 subtests) ============== [20:05:54] ============== drm_test_dp_mst_calc_pbn_mode ============== [20:05:54] [PASSED] Clock 154000 BPP 30 DSC disabled [20:05:54] [PASSED] Clock 234000 BPP 30 DSC disabled [20:05:54] [PASSED] Clock 297000 BPP 24 DSC disabled [20:05:54] [PASSED] Clock 332880 BPP 24 DSC enabled [20:05:54] [PASSED] Clock 324540 BPP 24 DSC enabled [20:05:54] ========== [PASSED] drm_test_dp_mst_calc_pbn_mode ========== [20:05:54] ============== drm_test_dp_mst_calc_pbn_div =============== [20:05:54] [PASSED] Link rate 2000000 lane count 4 [20:05:54] [PASSED] Link rate 2000000 lane count 2 [20:05:54] [PASSED] Link rate 2000000 lane count 1 [20:05:54] [PASSED] Link rate 1350000 lane count 4 [20:05:54] [PASSED] Link rate 1350000 lane count 2 [20:05:54] [PASSED] Link rate 1350000 lane count 1 [20:05:54] [PASSED] Link rate 1000000 lane count 4 [20:05:54] [PASSED] Link rate 1000000 lane count 2 [20:05:54] [PASSED] Link rate 1000000 lane count 1 [20:05:54] [PASSED] Link rate 810000 lane count 4 [20:05:54] [PASSED] Link rate 810000 lane count 2 [20:05:54] [PASSED] Link rate 810000 lane count 1 [20:05:54] [PASSED] Link rate 540000 lane count 4 [20:05:54] [PASSED] Link rate 540000 lane count 2 [20:05:54] [PASSED] Link rate 540000 lane count 1 [20:05:54] [PASSED] Link rate 270000 lane count 4 [20:05:54] [PASSED] Link rate 270000 lane count 2 [20:05:54] [PASSED] Link rate 270000 lane count 1 [20:05:54] [PASSED] Link rate 162000 lane count 4 [20:05:54] [PASSED] Link rate 162000 lane count 2 [20:05:54] [PASSED] Link rate 162000 lane count 1 [20:05:54] ========== [PASSED] drm_test_dp_mst_calc_pbn_div =========== [20:05:54] ========= drm_test_dp_mst_sideband_msg_req_decode ========= [20:05:54] [PASSED] DP_ENUM_PATH_RESOURCES with port number [20:05:54] [PASSED] DP_POWER_UP_PHY with port number [20:05:54] [PASSED] DP_POWER_DOWN_PHY with port number [20:05:54] [PASSED] DP_ALLOCATE_PAYLOAD with SDP stream sinks [20:05:54] [PASSED] DP_ALLOCATE_PAYLOAD with port number [20:05:54] [PASSED] DP_ALLOCATE_PAYLOAD with VCPI [20:05:54] [PASSED] DP_ALLOCATE_PAYLOAD with PBN [20:05:54] [PASSED] DP_QUERY_PAYLOAD with port number [20:05:54] [PASSED] DP_QUERY_PAYLOAD with VCPI [20:05:54] [PASSED] DP_REMOTE_DPCD_READ with port number [20:05:54] [PASSED] DP_REMOTE_DPCD_READ with DPCD address [20:05:54] [PASSED] DP_REMOTE_DPCD_READ with max number of bytes [20:05:54] [PASSED] DP_REMOTE_DPCD_WRITE with port number [20:05:54] [PASSED] DP_REMOTE_DPCD_WRITE with DPCD address [20:05:54] [PASSED] DP_REMOTE_DPCD_WRITE with data array [20:05:54] [PASSED] DP_REMOTE_I2C_READ with port number [20:05:54] [PASSED] DP_REMOTE_I2C_READ with I2C device ID [20:05:54] [PASSED] DP_REMOTE_I2C_READ with transactions array [20:05:54] [PASSED] DP_REMOTE_I2C_WRITE with port number [20:05:54] [PASSED] DP_REMOTE_I2C_WRITE with I2C device ID [20:05:54] [PASSED] DP_REMOTE_I2C_WRITE with data array [20:05:54] [PASSED] DP_QUERY_STREAM_ENC_STATUS with stream ID [20:05:54] [PASSED] DP_QUERY_STREAM_ENC_STATUS with client ID [20:05:54] [PASSED] DP_QUERY_STREAM_ENC_STATUS with stream event [20:05:54] [PASSED] DP_QUERY_STREAM_ENC_STATUS with valid stream event [20:05:54] [PASSED] DP_QUERY_STREAM_ENC_STATUS with stream behavior [20:05:54] [PASSED] DP_QUERY_STREAM_ENC_STATUS with a valid stream behavior [20:05:54] ===== [PASSED] drm_test_dp_mst_sideband_msg_req_decode ===== [20:05:54] ================ [PASSED] drm_dp_mst_helper ================ [20:05:54] ================== drm_exec (7 subtests) =================== [20:05:54] [PASSED] sanitycheck [20:05:54] [PASSED] test_lock [20:05:54] [PASSED] test_lock_unlock [20:05:54] [PASSED] test_duplicates [20:05:54] [PASSED] test_prepare [20:05:54] [PASSED] test_prepare_array [20:05:54] [PASSED] test_multiple_loops [20:05:54] ==================== [PASSED] drm_exec ===================== [20:05:54] =========== drm_format_helper_test (17 subtests) =========== [20:05:54] ============== drm_test_fb_xrgb8888_to_gray8 ============== [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ========== [PASSED] drm_test_fb_xrgb8888_to_gray8 ========== [20:05:54] ============= drm_test_fb_xrgb8888_to_rgb332 ============== [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ========= [PASSED] drm_test_fb_xrgb8888_to_rgb332 ========== [20:05:54] ============= drm_test_fb_xrgb8888_to_rgb565 ============== [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ========= [PASSED] drm_test_fb_xrgb8888_to_rgb565 ========== [20:05:54] ============ drm_test_fb_xrgb8888_to_xrgb1555 ============= [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ======== [PASSED] drm_test_fb_xrgb8888_to_xrgb1555 ========= [20:05:54] ============ drm_test_fb_xrgb8888_to_argb1555 ============= [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ======== [PASSED] drm_test_fb_xrgb8888_to_argb1555 ========= [20:05:54] ============ drm_test_fb_xrgb8888_to_rgba5551 ============= [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ======== [PASSED] drm_test_fb_xrgb8888_to_rgba5551 ========= [20:05:54] ============= drm_test_fb_xrgb8888_to_rgb888 ============== [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ========= [PASSED] drm_test_fb_xrgb8888_to_rgb888 ========== [20:05:54] ============= drm_test_fb_xrgb8888_to_bgr888 ============== [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ========= [PASSED] drm_test_fb_xrgb8888_to_bgr888 ========== [20:05:54] ============ drm_test_fb_xrgb8888_to_argb8888 ============= [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ======== [PASSED] drm_test_fb_xrgb8888_to_argb8888 ========= [20:05:54] =========== drm_test_fb_xrgb8888_to_xrgb2101010 =========== [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ======= [PASSED] drm_test_fb_xrgb8888_to_xrgb2101010 ======= [20:05:54] =========== drm_test_fb_xrgb8888_to_argb2101010 =========== [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ======= [PASSED] drm_test_fb_xrgb8888_to_argb2101010 ======= [20:05:54] ============== drm_test_fb_xrgb8888_to_mono =============== [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ========== [PASSED] drm_test_fb_xrgb8888_to_mono =========== [20:05:54] ==================== drm_test_fb_swab ===================== [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ================ [PASSED] drm_test_fb_swab ================= [20:05:54] ============ drm_test_fb_xrgb8888_to_xbgr8888 ============= [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ======== [PASSED] drm_test_fb_xrgb8888_to_xbgr8888 ========= [20:05:54] ============ drm_test_fb_xrgb8888_to_abgr8888 ============= [20:05:54] [PASSED] single_pixel_source_buffer [20:05:54] [PASSED] single_pixel_clip_rectangle [20:05:54] [PASSED] well_known_colors [20:05:54] [PASSED] destination_pitch [20:05:54] ======== [PASSED] drm_test_fb_xrgb8888_to_abgr8888 ========= [20:05:54] ================= drm_test_fb_clip_offset ================= [20:05:54] [PASSED] pass through [20:05:54] [PASSED] horizontal offset [20:05:54] [PASSED] vertical offset [20:05:54] [PASSED] horizontal and vertical offset [20:05:54] [PASSED] horizontal offset (custom pitch) [20:05:54] [PASSED] vertical offset (custom pitch) [20:05:54] [PASSED] horizontal and vertical offset (custom pitch) [20:05:54] ============= [PASSED] drm_test_fb_clip_offset ============= [20:05:54] =================== drm_test_fb_memcpy ==================== [20:05:54] [PASSED] single_pixel_source_buffer: XR24 little-endian (0x34325258) [20:05:54] [PASSED] single_pixel_source_buffer: XRA8 little-endian (0x38415258) [20:05:54] [PASSED] single_pixel_source_buffer: YU24 little-endian (0x34325559) [20:05:54] [PASSED] single_pixel_clip_rectangle: XB24 little-endian (0x34324258) [20:05:54] [PASSED] single_pixel_clip_rectangle: XRA8 little-endian (0x38415258) [20:05:54] [PASSED] single_pixel_clip_rectangle: YU24 little-endian (0x34325559) [20:05:54] [PASSED] well_known_colors: XB24 little-endian (0x34324258) [20:05:54] [PASSED] well_known_colors: XRA8 little-endian (0x38415258) [20:05:54] [PASSED] well_known_colors: YU24 little-endian (0x34325559) [20:05:54] [PASSED] destination_pitch: XB24 little-endian (0x34324258) [20:05:54] [PASSED] destination_pitch: XRA8 little-endian (0x38415258) [20:05:54] [PASSED] destination_pitch: YU24 little-endian (0x34325559) [20:05:54] =============== [PASSED] drm_test_fb_memcpy ================ [20:05:54] ============= [PASSED] drm_format_helper_test ============== [20:05:54] ================= drm_format (18 subtests) ================= [20:05:54] [PASSED] drm_test_format_block_width_invalid [20:05:54] [PASSED] drm_test_format_block_width_one_plane [20:05:54] [PASSED] drm_test_format_block_width_two_plane [20:05:54] [PASSED] drm_test_format_block_width_three_plane [20:05:54] [PASSED] drm_test_format_block_width_tiled [20:05:54] [PASSED] drm_test_format_block_height_invalid [20:05:54] [PASSED] drm_test_format_block_height_one_plane [20:05:54] [PASSED] drm_test_format_block_height_two_plane [20:05:54] [PASSED] drm_test_format_block_height_three_plane [20:05:54] [PASSED] drm_test_format_block_height_tiled [20:05:54] [PASSED] drm_test_format_min_pitch_invalid [20:05:54] [PASSED] drm_test_format_min_pitch_one_plane_8bpp [20:05:54] [PASSED] drm_test_format_min_pitch_one_plane_16bpp [20:05:54] [PASSED] drm_test_format_min_pitch_one_plane_24bpp [20:05:54] [PASSED] drm_test_format_min_pitch_one_plane_32bpp [20:05:54] [PASSED] drm_test_format_min_pitch_two_plane [20:05:54] [PASSED] drm_test_format_min_pitch_three_plane_8bpp [20:05:54] [PASSED] drm_test_format_min_pitch_tiled [20:05:54] =================== [PASSED] drm_format ==================== [20:05:54] ============== drm_framebuffer (10 subtests) =============== [20:05:54] ========== drm_test_framebuffer_check_src_coords ========== [20:05:54] [PASSED] Success: source fits into fb [20:05:54] [PASSED] Fail: overflowing fb with x-axis coordinate [20:05:54] [PASSED] Fail: overflowing fb with y-axis coordinate [20:05:54] [PASSED] Fail: overflowing fb with source width [20:05:54] [PASSED] Fail: overflowing fb with source height [20:05:54] ====== [PASSED] drm_test_framebuffer_check_src_coords ====== [20:05:54] [PASSED] drm_test_framebuffer_cleanup [20:05:54] =============== drm_test_framebuffer_create =============== [20:05:54] [PASSED] ABGR8888 normal sizes [20:05:54] [PASSED] ABGR8888 max sizes [20:05:54] [PASSED] ABGR8888 pitch greater than min required [20:05:54] [PASSED] ABGR8888 pitch less than min required [20:05:54] [PASSED] ABGR8888 Invalid width [20:05:54] [PASSED] ABGR8888 Invalid buffer handle [20:05:54] [PASSED] No pixel format [20:05:54] [PASSED] ABGR8888 Width 0 [20:05:54] [PASSED] ABGR8888 Height 0 [20:05:54] [PASSED] ABGR8888 Out of bound height * pitch combination [20:05:54] [PASSED] ABGR8888 Large buffer offset [20:05:54] [PASSED] ABGR8888 Buffer offset for inexistent plane [20:05:54] [PASSED] ABGR8888 Invalid flag [20:05:54] [PASSED] ABGR8888 Set DRM_MODE_FB_MODIFIERS without modifiers [20:05:54] [PASSED] ABGR8888 Valid buffer modifier [20:05:54] [PASSED] ABGR8888 Invalid buffer modifier(DRM_FORMAT_MOD_SAMSUNG_64_32_TILE) [20:05:54] [PASSED] ABGR8888 Extra pitches without DRM_MODE_FB_MODIFIERS [20:05:54] [PASSED] ABGR8888 Extra pitches with DRM_MODE_FB_MODIFIERS [20:05:54] [PASSED] NV12 Normal sizes [20:05:54] [PASSED] NV12 Max sizes [20:05:54] [PASSED] NV12 Invalid pitch [20:05:54] [PASSED] NV12 Invalid modifier/missing DRM_MODE_FB_MODIFIERS flag [20:05:54] [PASSED] NV12 different modifier per-plane [20:05:54] [PASSED] NV12 with DRM_FORMAT_MOD_SAMSUNG_64_32_TILE [20:05:54] [PASSED] NV12 Valid modifiers without DRM_MODE_FB_MODIFIERS [20:05:54] [PASSED] NV12 Modifier for inexistent plane [20:05:54] [PASSED] NV12 Handle for inexistent plane [20:05:54] [PASSED] NV12 Handle for inexistent plane without DRM_MODE_FB_MODIFIERS [20:05:54] [PASSED] YVU420 DRM_MODE_FB_MODIFIERS set without modifier [20:05:54] [PASSED] YVU420 Normal sizes [20:05:54] [PASSED] YVU420 Max sizes [20:05:54] [PASSED] YVU420 Invalid pitch [20:05:54] [PASSED] YVU420 Different pitches [20:05:54] [PASSED] YVU420 Different buffer offsets/pitches [20:05:54] [PASSED] YVU420 Modifier set just for plane 0, without DRM_MODE_FB_MODIFIERS [20:05:54] [PASSED] YVU420 Modifier set just for planes 0, 1, without DRM_MODE_FB_MODIFIERS [20:05:54] [PASSED] YVU420 Modifier set just for plane 0, 1, with DRM_MODE_FB_MODIFIERS [20:05:54] [PASSED] YVU420 Valid modifier [20:05:54] [PASSED] YVU420 Different modifiers per plane [20:05:54] [PASSED] YVU420 Modifier for inexistent plane [20:05:54] [PASSED] YUV420_10BIT Invalid modifier(DRM_FORMAT_MOD_LINEAR) [20:05:54] [PASSED] X0L2 Normal sizes [20:05:54] [PASSED] X0L2 Max sizes [20:05:54] [PASSED] X0L2 Invalid pitch [20:05:54] [PASSED] X0L2 Pitch greater than minimum required [20:05:54] [PASSED] X0L2 Handle for inexistent plane [20:05:54] [PASSED] X0L2 Offset for inexistent plane, without DRM_MODE_FB_MODIFIERS set [20:05:54] [PASSED] X0L2 Modifier without DRM_MODE_FB_MODIFIERS set [20:05:54] [PASSED] X0L2 Valid modifier [20:05:54] [PASSED] X0L2 Modifier for inexistent plane [20:05:54] =========== [PASSED] drm_test_framebuffer_create =========== [20:05:54] [PASSED] drm_test_framebuffer_free [20:05:54] [PASSED] drm_test_framebuffer_init [20:05:54] [PASSED] drm_test_framebuffer_init_bad_format [20:05:54] [PASSED] drm_test_framebuffer_init_dev_mismatch [20:05:54] [PASSED] drm_test_framebuffer_lookup [20:05:54] [PASSED] drm_test_framebuffer_lookup_inexistent [20:05:54] [PASSED] drm_test_framebuffer_modifiers_not_supported [20:05:54] ================= [PASSED] drm_framebuffer ================= [20:05:54] ================ drm_gem_shmem (8 subtests) ================ [20:05:54] [PASSED] drm_gem_shmem_test_obj_create [20:05:54] [PASSED] drm_gem_shmem_test_obj_create_private [20:05:54] [PASSED] drm_gem_shmem_test_pin_pages [20:05:54] [PASSED] drm_gem_shmem_test_vmap [20:05:54] [PASSED] drm_gem_shmem_test_get_sg_table [20:05:54] [PASSED] drm_gem_shmem_test_get_pages_sgt [20:05:54] [PASSED] drm_gem_shmem_test_madvise [20:05:54] [PASSED] drm_gem_shmem_test_purge [20:05:54] ================== [PASSED] drm_gem_shmem ================== [20:05:54] === drm_atomic_helper_connector_hdmi_check (27 subtests) === [20:05:54] [PASSED] drm_test_check_broadcast_rgb_auto_cea_mode [20:05:54] [PASSED] drm_test_check_broadcast_rgb_auto_cea_mode_vic_1 [20:05:54] [PASSED] drm_test_check_broadcast_rgb_full_cea_mode [20:05:54] [PASSED] drm_test_check_broadcast_rgb_full_cea_mode_vic_1 [20:05:54] [PASSED] drm_test_check_broadcast_rgb_limited_cea_mode [20:05:54] [PASSED] drm_test_check_broadcast_rgb_limited_cea_mode_vic_1 [20:05:54] ====== drm_test_check_broadcast_rgb_cea_mode_yuv420 ======= [20:05:54] [PASSED] Automatic [20:05:54] [PASSED] Full [20:05:54] [PASSED] Limited 16:235 [20:05:54] == [PASSED] drm_test_check_broadcast_rgb_cea_mode_yuv420 === [20:05:54] [PASSED] drm_test_check_broadcast_rgb_crtc_mode_changed [20:05:54] [PASSED] drm_test_check_broadcast_rgb_crtc_mode_not_changed [20:05:54] [PASSED] drm_test_check_disable_connector [20:05:54] [PASSED] drm_test_check_hdmi_funcs_reject_rate [20:05:54] [PASSED] drm_test_check_max_tmds_rate_bpc_fallback_rgb [20:05:54] [PASSED] drm_test_check_max_tmds_rate_bpc_fallback_yuv420 [20:05:54] [PASSED] drm_test_check_max_tmds_rate_bpc_fallback_ignore_yuv422 [20:05:54] [PASSED] drm_test_check_max_tmds_rate_bpc_fallback_ignore_yuv420 [20:05:54] [PASSED] drm_test_check_driver_unsupported_fallback_yuv420 [20:05:54] [PASSED] drm_test_check_output_bpc_crtc_mode_changed [20:05:54] [PASSED] drm_test_check_output_bpc_crtc_mode_not_changed [20:05:54] [PASSED] drm_test_check_output_bpc_dvi [20:05:54] [PASSED] drm_test_check_output_bpc_format_vic_1 [20:05:54] [PASSED] drm_test_check_output_bpc_format_display_8bpc_only [20:05:54] [PASSED] drm_test_check_output_bpc_format_display_rgb_only [20:05:54] [PASSED] drm_test_check_output_bpc_format_driver_8bpc_only [20:05:54] [PASSED] drm_test_check_output_bpc_format_driver_rgb_only [20:05:54] [PASSED] drm_test_check_tmds_char_rate_rgb_8bpc [20:05:54] [PASSED] drm_test_check_tmds_char_rate_rgb_10bpc [20:05:54] [PASSED] drm_test_check_tmds_char_rate_rgb_12bpc [20:05:54] ===== [PASSED] drm_atomic_helper_connector_hdmi_check ====== [20:05:54] === drm_atomic_helper_connector_hdmi_reset (6 subtests) ==== [20:05:54] [PASSED] drm_test_check_broadcast_rgb_value [20:05:54] [PASSED] drm_test_check_bpc_8_value [20:05:54] [PASSED] drm_test_check_bpc_10_value [20:05:54] [PASSED] drm_test_check_bpc_12_value [20:05:54] [PASSED] drm_test_check_format_value [20:05:54] [PASSED] drm_test_check_tmds_char_value [20:05:54] ===== [PASSED] drm_atomic_helper_connector_hdmi_reset ====== [20:05:54] = drm_atomic_helper_connector_hdmi_mode_valid (4 subtests) = [20:05:54] [PASSED] drm_test_check_mode_valid [20:05:54] [PASSED] drm_test_check_mode_valid_reject [20:05:54] [PASSED] drm_test_check_mode_valid_reject_rate [20:05:54] [PASSED] drm_test_check_mode_valid_reject_max_clock [20:05:54] === [PASSED] drm_atomic_helper_connector_hdmi_mode_valid === [20:05:54] = drm_atomic_helper_connector_hdmi_infoframes (5 subtests) = [20:05:54] [PASSED] drm_test_check_infoframes [20:05:54] [PASSED] drm_test_check_reject_avi_infoframe [20:05:54] [PASSED] drm_test_check_reject_hdr_infoframe_bpc_8 [20:05:54] [PASSED] drm_test_check_reject_hdr_infoframe_bpc_10 [20:05:54] [PASSED] drm_test_check_reject_audio_infoframe [20:05:54] === [PASSED] drm_atomic_helper_connector_hdmi_infoframes === [20:05:54] ================= drm_managed (2 subtests) ================= [20:05:54] [PASSED] drm_test_managed_release_action [20:05:54] [PASSED] drm_test_managed_run_action [20:05:54] =================== [PASSED] drm_managed =================== [20:05:54] =================== drm_mm (6 subtests) ==================== [20:05:54] [PASSED] drm_test_mm_init [20:05:54] [PASSED] drm_test_mm_debug [20:05:54] [PASSED] drm_test_mm_align32 [20:05:54] [PASSED] drm_test_mm_align64 [20:05:54] [PASSED] drm_test_mm_lowest [20:05:54] [PASSED] drm_test_mm_highest [20:05:54] ===================== [PASSED] drm_mm ====================== [20:05:54] ============= drm_modes_analog_tv (5 subtests) ============= [20:05:54] [PASSED] drm_test_modes_analog_tv_mono_576i [20:05:54] [PASSED] drm_test_modes_analog_tv_ntsc_480i [20:05:54] [PASSED] drm_test_modes_analog_tv_ntsc_480i_inlined [20:05:54] [PASSED] drm_test_modes_analog_tv_pal_576i [20:05:54] [PASSED] drm_test_modes_analog_tv_pal_576i_inlined [20:05:54] =============== [PASSED] drm_modes_analog_tv =============== [20:05:54] ============== drm_plane_helper (2 subtests) =============== [20:05:54] =============== drm_test_check_plane_state ================ [20:05:54] [PASSED] clipping_simple [20:05:54] [PASSED] clipping_rotate_reflect [20:05:54] [PASSED] positioning_simple [20:05:54] [PASSED] upscaling [20:05:54] [PASSED] downscaling [20:05:54] [PASSED] rounding1 [20:05:54] [PASSED] rounding2 [20:05:54] [PASSED] rounding3 [20:05:54] [PASSED] rounding4 [20:05:54] =========== [PASSED] drm_test_check_plane_state ============ [20:05:54] =========== drm_test_check_invalid_plane_state ============ [20:05:54] [PASSED] positioning_invalid [20:05:54] [PASSED] upscaling_invalid [20:05:54] [PASSED] downscaling_invalid [20:05:54] ======= [PASSED] drm_test_check_invalid_plane_state ======== [20:05:54] ================ [PASSED] drm_plane_helper ================= [20:05:54] ====== drm_connector_helper_tv_get_modes (1 subtest) ======= [20:05:54] ====== drm_test_connector_helper_tv_get_modes_check ======= [20:05:54] [PASSED] None [20:05:54] [PASSED] PAL [20:05:54] [PASSED] NTSC [20:05:54] [PASSED] Both, NTSC Default [20:05:54] [PASSED] Both, PAL Default [20:05:54] [PASSED] Both, NTSC Default, with PAL on command-line [20:05:54] [PASSED] Both, PAL Default, with NTSC on command-line [20:05:54] == [PASSED] drm_test_connector_helper_tv_get_modes_check === [20:05:54] ======== [PASSED] drm_connector_helper_tv_get_modes ======== [20:05:54] ================== drm_rect (9 subtests) =================== [20:05:54] [PASSED] drm_test_rect_clip_scaled_div_by_zero [20:05:54] [PASSED] drm_test_rect_clip_scaled_not_clipped [20:05:54] [PASSED] drm_test_rect_clip_scaled_clipped [20:05:54] [PASSED] drm_test_rect_clip_scaled_signed_vs_unsigned [20:05:54] ================= drm_test_rect_intersect ================= [20:05:54] [PASSED] top-left x bottom-right: 2x2+1+1 x 2x2+0+0 [20:05:54] [PASSED] top-right x bottom-left: 2x2+0+0 x 2x2+1-1 [20:05:54] [PASSED] bottom-left x top-right: 2x2+1-1 x 2x2+0+0 [20:05:54] [PASSED] bottom-right x top-left: 2x2+0+0 x 2x2+1+1 [20:05:54] [PASSED] right x left: 2x1+0+0 x 3x1+1+0 [20:05:54] [PASSED] left x right: 3x1+1+0 x 2x1+0+0 [20:05:54] [PASSED] up x bottom: 1x2+0+0 x 1x3+0-1 [20:05:54] [PASSED] bottom x up: 1x3+0-1 x 1x2+0+0 [20:05:54] [PASSED] touching corner: 1x1+0+0 x 2x2+1+1 [20:05:54] [PASSED] touching side: 1x1+0+0 x 1x1+1+0 [20:05:54] [PASSED] equal rects: 2x2+0+0 x 2x2+0+0 [20:05:54] [PASSED] inside another: 2x2+0+0 x 1x1+1+1 [20:05:54] [PASSED] far away: 1x1+0+0 x 1x1+3+6 [20:05:54] [PASSED] points intersecting: 0x0+5+10 x 0x0+5+10 [20:05:54] [PASSED] points not intersecting: 0x0+0+0 x 0x0+5+10 [20:05:54] ============= [PASSED] drm_test_rect_intersect ============= [20:05:54] ================ drm_test_rect_calc_hscale ================ [20:05:54] [PASSED] normal use [20:05:54] [PASSED] out of max range [20:05:54] [PASSED] out of min range [20:05:54] [PASSED] zero dst [20:05:54] [PASSED] negative src [20:05:54] [PASSED] negative dst [20:05:54] ============ [PASSED] drm_test_rect_calc_hscale ============ [20:05:54] ================ drm_test_rect_calc_vscale ================ [20:05:54] [PASSED] normal use [20:05:54] [PASSED] out of max range [20:05:54] [PASSED] out of min range [20:05:54] [PASSED] zero dst [20:05:54] [PASSED] negative src [20:05:54] [PASSED] negative dst [20:05:54] ============ [PASSED] drm_test_rect_calc_vscale ============ [20:05:54] ================== drm_test_rect_rotate =================== [20:05:54] [PASSED] reflect-x [20:05:54] [PASSED] reflect-y [20:05:54] [PASSED] rotate-0 [20:05:54] [PASSED] rotate-90 [20:05:54] [PASSED] rotate-180 [20:05:54] [PASSED] rotate-270 [20:05:54] ============== [PASSED] drm_test_rect_rotate =============== [20:05:54] ================ drm_test_rect_rotate_inv ================= [20:05:54] [PASSED] reflect-x [20:05:54] [PASSED] reflect-y [20:05:54] [PASSED] rotate-0 [20:05:54] [PASSED] rotate-90 [20:05:54] [PASSED] rotate-180 [20:05:54] [PASSED] rotate-270 [20:05:54] ============ [PASSED] drm_test_rect_rotate_inv ============= [20:05:54] ==================== [PASSED] drm_rect ===================== [20:05:54] ============ drm_sysfb_modeset_test (1 subtest) ============ [20:05:54] ============ drm_test_sysfb_build_fourcc_list ============= [20:05:54] [PASSED] no native formats [20:05:54] [PASSED] XRGB8888 as native format [20:05:54] [PASSED] remove duplicates [20:05:54] [PASSED] convert alpha formats [20:05:54] [PASSED] random formats [20:05:54] ======== [PASSED] drm_test_sysfb_build_fourcc_list ========= [20:05:54] ============= [PASSED] drm_sysfb_modeset_test ============== [20:05:54] ================== drm_fixp (2 subtests) =================== [20:05:54] [PASSED] drm_test_int2fixp [20:05:54] [PASSED] drm_test_sm2fixp [20:05:54] ==================== [PASSED] drm_fixp ===================== [20:05:54] ============================================================ [20:05:54] Testing complete. Ran 621 tests: passed: 621 [20:05:54] Elapsed time: 25.906s total, 1.679s configuring, 24.059s building, 0.133s running
+ /kernel/tools/testing/kunit/kunit.py run --kunitconfig /kernel/drivers/gpu/drm/ttm/tests/.kunitconfig [20:05:54] Configuring KUnit Kernel ...
Regenerating .config ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig [20:05:56] 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 [20:06:06] Starting KUnit Kernel (1/1)... [20:06:06] ============================================================ Running tests with: $ .kunit/linux kunit.enable=1 mem=1G console=tty kunit_shutdown=halt [20:06:06] ================= ttm_device (5 subtests) ================== [20:06:06] [PASSED] ttm_device_init_basic [20:06:06] [PASSED] ttm_device_init_multiple [20:06:06] [PASSED] ttm_device_fini_basic [20:06:06] [PASSED] ttm_device_init_no_vma_man [20:06:06] ================== ttm_device_init_pools ================== [20:06:06] [PASSED] No DMA allocations, no DMA32 required [20:06:06] [PASSED] DMA allocations, DMA32 required [20:06:06] [PASSED] No DMA allocations, DMA32 required [20:06:06] [PASSED] DMA allocations, no DMA32 required [20:06:06] ============== [PASSED] ttm_device_init_pools ============== [20:06:06] =================== [PASSED] ttm_device ==================== [20:06:06] ================== ttm_pool (8 subtests) =================== [20:06:06] ================== ttm_pool_alloc_basic =================== [20:06:06] [PASSED] One page [20:06:06] [PASSED] More than one page [20:06:06] [PASSED] Above the allocation limit [20:06:06] [PASSED] One page, with coherent DMA mappings enabled [20:06:06] [PASSED] Above the allocation limit, with coherent DMA mappings enabled [20:06:06] ============== [PASSED] ttm_pool_alloc_basic =============== [20:06:06] ============== ttm_pool_alloc_basic_dma_addr ============== [20:06:06] [PASSED] One page [20:06:06] [PASSED] More than one page [20:06:06] [PASSED] Above the allocation limit [20:06:06] [PASSED] One page, with coherent DMA mappings enabled [20:06:06] [PASSED] Above the allocation limit, with coherent DMA mappings enabled [20:06:06] ========== [PASSED] ttm_pool_alloc_basic_dma_addr ========== [20:06:06] [PASSED] ttm_pool_alloc_order_caching_match [20:06:06] [PASSED] ttm_pool_alloc_caching_mismatch [20:06:06] [PASSED] ttm_pool_alloc_order_mismatch [20:06:06] [PASSED] ttm_pool_free_dma_alloc [20:06:06] [PASSED] ttm_pool_free_no_dma_alloc [20:06:06] [PASSED] ttm_pool_fini_basic [20:06:06] ==================== [PASSED] ttm_pool ===================== [20:06:06] ================ ttm_resource (8 subtests) ================= [20:06:06] ================= ttm_resource_init_basic ================= [20:06:06] [PASSED] Init resource in TTM_PL_SYSTEM [20:06:06] [PASSED] Init resource in TTM_PL_VRAM [20:06:06] [PASSED] Init resource in a private placement [20:06:06] [PASSED] Init resource in TTM_PL_SYSTEM, set placement flags [20:06:06] ============= [PASSED] ttm_resource_init_basic ============= [20:06:06] [PASSED] ttm_resource_init_pinned [20:06:06] [PASSED] ttm_resource_fini_basic [20:06:06] [PASSED] ttm_resource_manager_init_basic [20:06:06] [PASSED] ttm_resource_manager_usage_basic [20:06:06] [PASSED] ttm_resource_manager_set_used_basic [20:06:06] [PASSED] ttm_sys_man_alloc_basic [20:06:06] [PASSED] ttm_sys_man_free_basic [20:06:06] ================== [PASSED] ttm_resource =================== [20:06:06] =================== ttm_tt (15 subtests) =================== [20:06:06] ==================== ttm_tt_init_basic ==================== [20:06:06] [PASSED] Page-aligned size [20:06:06] [PASSED] Extra pages requested [20:06:06] ================ [PASSED] ttm_tt_init_basic ================ [20:06:06] [PASSED] ttm_tt_init_misaligned [20:06:06] [PASSED] ttm_tt_fini_basic [20:06:06] [PASSED] ttm_tt_fini_sg [20:06:06] [PASSED] ttm_tt_fini_shmem [20:06:06] [PASSED] ttm_tt_create_basic [20:06:06] [PASSED] ttm_tt_create_invalid_bo_type [20:06:06] [PASSED] ttm_tt_create_ttm_exists [20:06:06] [PASSED] ttm_tt_create_failed [20:06:06] [PASSED] ttm_tt_destroy_basic [20:06:06] [PASSED] ttm_tt_populate_null_ttm [20:06:06] [PASSED] ttm_tt_populate_populated_ttm [20:06:06] [PASSED] ttm_tt_unpopulate_basic [20:06:06] [PASSED] ttm_tt_unpopulate_empty_ttm [20:06:06] [PASSED] ttm_tt_swapin_basic [20:06:06] ===================== [PASSED] ttm_tt ====================== [20:06:06] =================== ttm_bo (14 subtests) =================== [20:06:06] =========== ttm_bo_reserve_optimistic_no_ticket =========== [20:06:06] [PASSED] Cannot be interrupted and sleeps [20:06:06] [PASSED] Cannot be interrupted, locks straight away [20:06:06] [PASSED] Can be interrupted, sleeps [20:06:06] ======= [PASSED] ttm_bo_reserve_optimistic_no_ticket ======= [20:06:06] [PASSED] ttm_bo_reserve_locked_no_sleep [20:06:06] [PASSED] ttm_bo_reserve_no_wait_ticket [20:06:06] [PASSED] ttm_bo_reserve_double_resv [20:06:06] [PASSED] ttm_bo_reserve_interrupted [20:06:06] [PASSED] ttm_bo_reserve_deadlock [20:06:06] [PASSED] ttm_bo_unreserve_basic [20:06:06] [PASSED] ttm_bo_unreserve_pinned [20:06:06] [PASSED] ttm_bo_unreserve_bulk [20:06:06] [PASSED] ttm_bo_fini_basic [20:06:06] [PASSED] ttm_bo_fini_shared_resv [20:06:06] [PASSED] ttm_bo_pin_basic [20:06:06] [PASSED] ttm_bo_pin_unpin_resource [20:06:06] [PASSED] ttm_bo_multiple_pin_one_unpin [20:06:06] ===================== [PASSED] ttm_bo ====================== [20:06:06] ============== ttm_bo_validate (22 subtests) =============== [20:06:06] ============== ttm_bo_init_reserved_sys_man =============== [20:06:06] [PASSED] Buffer object for userspace [20:06:06] [PASSED] Kernel buffer object [20:06:06] [PASSED] Shared buffer object [20:06:06] ========== [PASSED] ttm_bo_init_reserved_sys_man =========== [20:06:06] ============== ttm_bo_init_reserved_mock_man ============== [20:06:06] [PASSED] Buffer object for userspace [20:06:06] [PASSED] Kernel buffer object [20:06:06] [PASSED] Shared buffer object [20:06:06] ========== [PASSED] ttm_bo_init_reserved_mock_man ========== [20:06:06] [PASSED] ttm_bo_init_reserved_resv [20:06:06] ================== ttm_bo_validate_basic ================== [20:06:06] [PASSED] Buffer object for userspace [20:06:06] [PASSED] Kernel buffer object [20:06:06] [PASSED] Shared buffer object [20:06:06] ============== [PASSED] ttm_bo_validate_basic ============== [20:06:06] [PASSED] ttm_bo_validate_invalid_placement [20:06:06] ============= ttm_bo_validate_same_placement ============== [20:06:06] [PASSED] System manager [20:06:06] [PASSED] VRAM manager [20:06:06] ========= [PASSED] ttm_bo_validate_same_placement ========== [20:06:06] [PASSED] ttm_bo_validate_failed_alloc [20:06:06] [PASSED] ttm_bo_validate_pinned [20:06:06] [PASSED] ttm_bo_validate_busy_placement [20:06:06] ================ ttm_bo_validate_multihop ================= [20:06:06] [PASSED] Buffer object for userspace [20:06:06] [PASSED] Kernel buffer object [20:06:06] [PASSED] Shared buffer object [20:06:06] ============ [PASSED] ttm_bo_validate_multihop ============= [20:06:06] ========== ttm_bo_validate_no_placement_signaled ========== [20:06:06] [PASSED] Buffer object in system domain, no page vector [20:06:06] [PASSED] Buffer object in system domain with an existing page vector [20:06:06] ====== [PASSED] ttm_bo_validate_no_placement_signaled ====== [20:06:06] ======== ttm_bo_validate_no_placement_not_signaled ======== [20:06:06] [PASSED] Buffer object for userspace [20:06:06] [PASSED] Kernel buffer object [20:06:06] [PASSED] Shared buffer object [20:06:06] ==== [PASSED] ttm_bo_validate_no_placement_not_signaled ==== [20:06:06] [PASSED] ttm_bo_validate_move_fence_signaled [20:06:06] ========= ttm_bo_validate_move_fence_not_signaled ========= [20:06:06] [PASSED] Waits for GPU [20:06:06] [PASSED] Tries to lock straight away [20:06:06] ===== [PASSED] ttm_bo_validate_move_fence_not_signaled ===== [20:06:06] [PASSED] ttm_bo_validate_swapout [20:06:06] [PASSED] ttm_bo_validate_happy_evict [20:06:06] [PASSED] ttm_bo_validate_all_pinned_evict [20:06:06] [PASSED] ttm_bo_validate_allowed_only_evict [20:06:06] [PASSED] ttm_bo_validate_deleted_evict [20:06:06] [PASSED] ttm_bo_validate_busy_domain_evict [20:06:06] [PASSED] ttm_bo_validate_evict_gutting [20:06:06] [PASSED] ttm_bo_validate_recrusive_evict [20:06:06] ================= [PASSED] ttm_bo_validate ================= [20:06:06] ============================================================ [20:06:06] Testing complete. Ran 102 tests: passed: 102 [20:06:06] Elapsed time: 11.657s total, 1.681s configuring, 9.761s building, 0.184s running
+ cleanup
++ stat -c %u:%g /kernel
+ chown -R 1003:1003 /kernel
Patchwork
CI✓ Xe.CI.BAT: success for mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation (rev3)
In reply to Matthew Brost== Series Details ==
== Series Details ==
Series: mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation (rev3)
URL : https://patchwork.freedesktop.org/series/165329/
State : success
== Summary ==
CI Bug Log - changes from xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4_BAT -> xe-pw-165329v3_BAT
====================================================
Summary
-------
**SUCCESS**
No regressions found.
Participating hosts (13 -> 13)
------------------------------
No changes in participating hosts
Known issues
------------
Here are the changes found in xe-pw-165329v3_BAT that come from known issues:
### IGT changes ###
#### Issues hit ####
* igt@kms_flip@basic-flip-vs-wf_vblank@d-edp1:
- bat-adlp-7: [PASS][1] -> [DMESG-WARN][2] ([Intel XE#7483])
[1]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4/bat-adlp-7/igt@kms_flip@[email protected]
[2]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-165329v3/bat-adlp-7/igt@kms_flip@[email protected]
#### Possible fixes ####
* igt@kms_flip@basic-flip-vs-wf_vblank@c-edp1:
- bat-adlp-7: [DMESG-WARN][3] ([Intel XE#7483]) -> [PASS][4]
[3]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4/bat-adlp-7/igt@kms_flip@[email protected]
[4]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-165329v3/bat-adlp-7/igt@kms_flip@[email protected]
[Intel XE#7483]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7483
Build changes
-------------
* Linux: xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4 -> xe-pw-165329v3
IGT_8881: aa5853ef5b379b1e4558218c21ef4caeae112184 @ https://gitlab.freedesktop.org/drm/igt-gpu-tools.git
xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4: 6472756e8384e3945f12b6726dd9dc1ac12f3bb4
xe-pw-165329v3: 165329v3
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-165329v3/index.html
Andrew Morton
maintainerQuoted 13 lines · On Thu, 30 Apr 2026 12:18:03 -0700 Matthew Brost <[email protected]>…
TTM allocations at higher orders can drive Xe into a pathological reclaim loop when memory is fragmented:
kswapd → shrinker → eviction → rebind (exec ioctl) → repeat
In this state, reclaim is triggered despite substantial free memory, but fails to produce contiguous higher-order pages. The Xe shrinker then evicts active buffer objects, increasing faulting and rebind activity and further feeding the loop. The result is high CPU overhead and poor GPU forward progress.
...
Quoted 8 lines
This series addresses the issue in two ways:
TTM: Restrict direct reclaim to
beneficial_order. Larger allocations use__GFP_NORETRYto fail quickly rather than triggering reclaim.Xe: Introduce a heuristic in the shrinker to avoid eviction when running under kswapd and the system appears memory-rich but fragmented.
Please cc everyone on all the patches? It's kind of annoying to have to hunt around to find out how these proposed changes will be used. Personal preference, anyway.
AI review flagged a few possible issues:
https://sashiko.dev/#/patchset/[email protected]
Matthew Brost
maintainerQuoted 20 lines · Andrew Morton
Quoted 18 lines · On Thu, 30 Apr 2026 12:18:03 -0700 Matthew Brost <[email protected]>…
TTM allocations at higher orders can drive Xe into a pathological reclaim loop when memory is fragmented:
kswapd → shrinker → eviction → rebind (exec ioctl) → repeat
In this state, reclaim is triggered despite substantial free memory, but fails to produce contiguous higher-order pages. The Xe shrinker then evicts active buffer objects, increasing faulting and rebind activity and further feeding the loop. The result is high CPU overhead and poor GPU forward progress.
...
This series addresses the issue in two ways:
TTM: Restrict direct reclaim to
beneficial_order. Larger allocations use__GFP_NORETRYto fail quickly rather than triggering reclaim.
Quoted 8 lines
Quoted 3 lines
Xe: Introduce a heuristic in the shrinker to avoid eviction when running under kswapd and the system appears memory-rich but fragmented.
Please cc everyone on all the patches? It's kind of annoying to have to hunt around to find out how these proposed changes will be used. Personal preference, anyway.
Will do - we discussed this in the past and thought we landed on Cc everyone on the cover then individual patches but will blast everyone going forward.
Quoted 2 lines
AI review flagged a few possible issues:
https://sashiko.dev/#/patchset/[email protected]
Idk, who authors sashiko but what make it really nice if you could reply to it to talk things out.
Looking at replies...
- 'Could this global counter drift significantly'
this is looks right for multi-CPU which isn't really the target here, but will adjust
- 'Additionally, does NR_FREE_PAGES implicitly include CMA pages?' this is looks right, will adjust
- 'Can high_wmark_pages(zone) evaluate to zero during early boot' theoretically possible (?), but non-issue IMO, certainly a GPU shrinker which is current use case this is impossible but maybe add a warn_on if high_wmark_pages(zone) returns zero
- 'Is this description accurate?'
I inverted the TTM kernel doc vs the code, will fix
Matt
Andrew Morton
maintainerQuoted 5 lines · On Thu, 30 Apr 2026 23:28:08 -0700 Matthew Brost <[email protected]>…
Quoted 2 lines
AI review flagged a few possible issues:
https://sashiko.dev/#/patchset/[email protected]Idk, who authors sashiko but what make it really nice if you could reply to it to talk things out.
It's a gemini 3 thing, based on prompts developed by Roman Gushchin and Chris Mason and others. Google is making this available to kernel developers at a non-trivial expense.
And yes, it would be great if Sashiko were able to learn from our replies and to fine-tune its checking based on the human corrections. I've asked for this a few times but didn't really understand the reply ;)
Dave Chinner
NAKQuoted 33 lines · Matthew Brost
TTM allocations at higher orders can drive Xe into a pathological reclaim loop when memory is fragmented:
kswapd → shrinker → eviction → rebind (exec ioctl) → repeat
In this state, reclaim is triggered despite substantial free memory, but fails to produce contiguous higher-order pages. The Xe shrinker then evicts active buffer objects, increasing faulting and rebind activity and further feeding the loop. The result is high CPU overhead and poor GPU forward progress.
This issue was first reported in [1] and independently observed internally and by Google.
A simple reproducer is:
- Boot an iGPU system with mem=8G
- Launch 10 Chrome tabs running the WebGL aquarium demo
- Configure each tab with ~5k fishUnder this workload, ftrace shows a continuous loop of:
xe_shrinker_scan(kswapd)xe_vma_rebind_execPerformance degrades significantly, with each tab dropping to ~2 FPS on PTL (Ubuntu 24.04).
At the same time,
/proc/buddyinfoshows substantial free memory but no higher-order availability. For example, the Normal zone:Count: 4063 4595 3455 3400 3139 2762 2293 1655 643 0 0
Quoted 7 lines
This corresponds to ~2.8GB free memory, but no order-9 (2MB) blocks, indicating severe fragmentation.
This series addresses the issue in two ways:
TTM: Restrict direct reclaim to
beneficial_order. Larger allocations use__GFP_NORETRYto fail quickly rather than triggering reclaim.
NACK.
As I have said to the people trying to hack around direct reclaim for high order allocations being costly for the page cache, fix the problem with direct reclaim. (e.g.
https://lore.kernel.org/linux-xfs/adLlrSZ5oRAa_Hfd@dread/)
We should not be hacking around a problem in the mm infrastructure by changing allocation context flags every high order allocation call site that needs high order allocations. Understand and fix the infrastructure problem once and for all.
Quoted 3 lines
Xe: Introduce a heuristic in the shrinker to avoid eviction when running under kswapd and the system appears memory-rich but fragmented.
NACK on architectural grounds.
Custom heuristics in individual shrinkers to decide whether the should do what the mm subsystem has asked them to do has -always- been a mistake to allow. The mm subsystem makes the decision on how much cache shrinkage needs to occur, the shrinkers just do what they are told to do.
If we have a problem where a workload causes excessive shrinker reclaim, then we need to address the problem in the infrastructure because excessive reclaim affects the performance of -all- subsystems with shrinkable caches, not just the TTM subsystem.
As it is, I can't review what you've actually implemented because you only cc'd me on a single patch in the series. In future, please cc me on the whole patchset because shrinkers need to work as a coherent whole, not just in isolation....
-Dave.
Signature · 2 lines
Dave Chinner [email protected]
Matthew Brost
maintainerOn Fri, May 01, 2026 at 11:42:19AM +1000, Dave Chinner wrote:
Thanks for the feedback. I’m looking into this more, and it’s becoming clear that this is a hard problem—one that will likely require coordinated work between DRM and core MM to really sort out. That said, I do think what I have in place is a reasonable short-term fix.
More below.
Quoted 42 lines · Matthew Brost
Quoted 33 lines
TTM allocations at higher orders can drive Xe into a pathological reclaim loop when memory is fragmented:
kswapd → shrinker → eviction → rebind (exec ioctl) → repeat
In this state, reclaim is triggered despite substantial free memory, but fails to produce contiguous higher-order pages. The Xe shrinker then evicts active buffer objects, increasing faulting and rebind activity and further feeding the loop. The result is high CPU overhead and poor GPU forward progress.
This issue was first reported in [1] and independently observed internally and by Google.
A simple reproducer is:
- Boot an iGPU system with mem=8G
- Launch 10 Chrome tabs running the WebGL aquarium demo
- Configure each tab with ~5k fishUnder this workload, ftrace shows a continuous loop of:
xe_shrinker_scan(kswapd)xe_vma_rebind_execPerformance degrades significantly, with each tab dropping to ~2 FPS on PTL (Ubuntu 24.04).
At the same time,
/proc/buddyinfoshows substantial free memory but no higher-order availability. For example, the Normal zone:Count: 4063 4595 3455 3400 3139 2762 2293 1655 643 0 0
Quoted 7 lines
This corresponds to ~2.8GB free memory, but no order-9 (2MB) blocks, indicating severe fragmentation.
This series addresses the issue in two ways:
TTM: Restrict direct reclaim to
beneficial_order. Larger allocations use__GFP_NORETRYto fail quickly rather than triggering reclaim.
Quoted 7 lines
NACK.
As I have said to the people trying to hack around direct reclaim for high order allocations being costly for the page cache, fix the problem with direct reclaim. (e.g.
https://lore.kernel.org/linux-xfs/adLlrSZ5oRAa_Hfd@dread/)
I read your response. Maybe this isn't clear what is going here.
At beneficial_order: gfp == __GFP_RECLAIM | __GFP_NORETRY At order zero: gfp == __GFP_RECLAIM
This roughly existing behavior, the exact changes are here [1].
[1] https://patchwork.freedesktop.org/patch/722247/?series=165329&rev=3
If this is truly a NACK, then we can rethink it—likely by disabling reclaim at higher orders—but that has its own downsides for DRM and GPUs. Ideally, you want purgeable BOs to be evicted when a higher-order allocation fails; you really don’t want to end up in an insane kswap loop.
Quoted 5 lines
We should not be hacking around a problem in the mm infrastructure by changing allocation context flags every high order allocation call site that needs high order allocations. Understand and fix the infrastructure problem once and for all.
Well, I agree that we should aim to fix this in core MM, but as the saying goes, Rome wasn’t built in a day. The fact is that these GFP flags do exist, and suddenly drawing a line and declaring them no longer valid feels a bit unfair. I’ll also note that Intel—and I personally—have an interest in fixing shrinking, so you can expect follow-up work here.
Quoted 9 lines
Quoted 3 lines
Xe: Introduce a heuristic in the shrinker to avoid eviction when running under kswapd and the system appears memory-rich but fragmented.
NACK on architectural grounds.
Custom heuristics in individual shrinkers to decide whether the should do what the mm subsystem has asked them to do has -always- been a mistake to allow. The mm subsystem makes the decision on how
I’m not going to disagree with using custom heuristics in individual shrinkers, but I’d wager that most shrinkers sadly already implement custom heuristics.
Quoted 8 lines
much cache shrinkage needs to occur, the shrinkers just do what they are told to do.
If we have a problem where a workload causes excessive shrinker reclaim, then we need to address the problem in the infrastructure because excessive reclaim affects the performance of -all- subsystems with shrinkable caches, not just the TTM subsystem.
Yes, I agree, and I’ve thought about the implications of simply having TTM back off when a higher-order allocation fails, even when we actually have enough memory, and how that would affect everyone. This series at least fixes the “well, there goes my GUI” problem.
I do have another patch locally that prevents TTM from accidentally fragmenting memory and triggering the kswap loop, but under enough pressure I can still get the GUI to lock up for periods of time. With this series, however, I can’t reproduce that issue.
Quoted 5 lines
As it is, I can't review what you've actually implemented because you only cc'd me on a single patch in the series. In future, please cc me on the whole patchset because shrinkers need to work as a coherent whole, not just in isolation....
Sorry about this - Andrew just said the same thing. Here is PW link [2].
Or:
b4 mbox [email protected]
[2] https://patchwork.freedesktop.org/series/165329/
If you have any ideas on how to fix this in the core, let’s discuss. I have a bunch of ideas in my head, but core MM isn’t my native domain.
Matt
Quoted 4 lines
-Dave.
Signature · 2 lines
Dave Chinner [email protected]
Patchwork
CI✓ Xe.CI.FULL: success for mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation (rev3)
In reply to Matthew Brost== Series Details ==
== Series Details ==
Series: mm, drm/ttm, drm/xe: Avoid reclaim/eviction loops under fragmentation (rev3)
URL : https://patchwork.freedesktop.org/series/165329/
State : success
== Summary ==
CI Bug Log - changes from xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4_FULL -> xe-pw-165329v3_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-165329v3_FULL that come from known issues:
### IGT changes ###
#### Possible fixes ####
* igt@kms_cursor_edge_walk@256x256-top-edge@pipe-d-dp-2:
- shard-bmg: [FAIL][1] ([Intel XE#6841]) -> [PASS][2] +1 other test pass
[1]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4/shard-bmg-2/igt@kms_cursor_edge_walk@[email protected]
[2]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-165329v3/shard-bmg-1/igt@kms_cursor_edge_walk@[email protected]
* igt@kms_flip@flip-vs-expired-vblank-interruptible:
- shard-lnl: [FAIL][3] ([Intel XE#301] / [Intel XE#3149]) -> [PASS][4] +1 other test pass
[3]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4/shard-lnl-5/igt@[email protected]
[4]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-165329v3/shard-lnl-1/igt@[email protected]
* igt@kms_hdr@invalid-hdr:
- shard-bmg: [SKIP][5] ([Intel XE#1503]) -> [PASS][6]
[5]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4/shard-bmg-5/igt@[email protected]
[6]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-165329v3/shard-bmg-7/igt@[email protected]
* igt@xe_evict@evict-beng-mixed-many-threads-small:
- shard-bmg: [INCOMPLETE][7] ([Intel XE#6321]) -> [PASS][8]
[7]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4/shard-bmg-4/igt@[email protected]
[8]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-165329v3/shard-bmg-1/igt@[email protected]
#### Warnings ####
* igt@kms_hdr@brightness-with-hdr:
- shard-bmg: [SKIP][9] ([Intel XE#3544]) -> [SKIP][10] ([Intel XE#3374] / [Intel XE#3544])
[9]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4/shard-bmg-6/igt@[email protected]
[10]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-165329v3/shard-bmg-1/igt@[email protected]
* igt@kms_tiled_display@basic-test-pattern:
- shard-bmg: [FAIL][11] ([Intel XE#1729] / [Intel XE#7424]) -> [SKIP][12] ([Intel XE#2426] / [Intel XE#5848])
[11]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4/shard-bmg-5/igt@[email protected]
[12]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-165329v3/shard-bmg-7/igt@[email protected]
* igt@kms_tiled_display@basic-test-pattern-with-chamelium:
- shard-bmg: [SKIP][13] ([Intel XE#2426] / [Intel XE#5848]) -> [SKIP][14] ([Intel XE#2509] / [Intel XE#7437])
[13]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4/shard-bmg-9/igt@[email protected]
[14]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-165329v3/shard-bmg-5/igt@[email protected]
[Intel XE#1503]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/1503
[Intel XE#1729]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/1729
[Intel XE#2426]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2426
[Intel XE#2509]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/2509
[Intel XE#301]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/301
[Intel XE#3149]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/3149
[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#5848]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/5848
[Intel XE#6321]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/6321
[Intel XE#6841]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/6841
[Intel XE#7424]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7424
[Intel XE#7437]: https://gitlab.freedesktop.org/drm/xe/kernel/issues/7437
Build changes
-------------
* Linux: xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4 -> xe-pw-165329v3
IGT_8881: aa5853ef5b379b1e4558218c21ef4caeae112184 @ https://gitlab.freedesktop.org/drm/igt-gpu-tools.git
xe-4960-6472756e8384e3945f12b6726dd9dc1ac12f3bb4: 6472756e8384e3945f12b6726dd9dc1ac12f3bb4
xe-pw-165329v3: 165329v3
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-165329v3/index.html