linux-stable.git/include/drm/ttm, branch master

drm/ttm: Fix ttm_bo_swapout() infinite LRU walk on swapout failure

2026-05-08T15:19:44+00:00

When ttm_tt_swapout() fails, the current code calls
ttm_resource_add_bulk_move() followed by ttm_resource_move_to_lru_tail()
to restore the resource's bulk_move membership.

However, ttm_resource_move_to_lru_tail() places the resource at the tail
of the LRU list which, relative to the walk cursor's hitch node (placed
immediately after the resource when it was yielded), puts the resource
*in front of the* the hitch. The next list_for_each_entry_continue() from
the hitch finds the same resource again, causing an infinite loop.

Fix by deferring del_bulk_move to the success path only.

On the success path, TTM_TT_FLAG_SWAPPED has just been set by
ttm_tt_swapout() but the resource is still tracked in the bulk_move range,
so ttm_resource_del_bulk_move()'s !ttm_resource_unevictable() guard would
incorrectly skip the removal. Introduce
ttm_resource_del_bulk_move_unevictable() which bypasses that guard.

Reported-by: Jatin Kataria 
Fixes: fc5d96670eb2 ("drm/ttm: Move swapped objects off the manager's LRU list")
Cc: Christian König 
Cc: Matthew Brost 
Cc: 
Cc:  # v6.13+
Assisted-by: GitHub_Copilot:claude-sonnet-4.6
Signed-off-by: Thomas Hellström 
Reviewed-by: Christian König 
Tested-by: Boqun Feng 
Link: https://patch.msgid.link/20260428094442.16985-1-thomas.hellstrom@linux.intel.com

ttm/pool: port to list_lru. (v2)

2026-04-07T20:52:47+00:00

This is an initial port of the TTM pools for
write combined and uncached pages to use the list_lru.

This makes the pool's more NUMA aware and avoids
needing separate NUMA pools (later commit enables this).

Cc: Christian Koenig 
Cc: Johannes Weiner 
Cc: Dave Chinner 
Reviewed-by: Christian König 
Signed-off-by: Dave Airlie

drm/ttm: Update the struct ttm_operation_ctx kerneldoc

2026-03-20T09:39:00+00:00

Update the kerneldoc with a more elaborate description of some members,
including the gfp_retry_mayfail member. Use inline kerneldoc.

Suggested-by: Simona Vetter 
Signed-off-by: Thomas Hellström 
Reviewed-by: Maarten Lankhorst 
Acked-by: Christian König 
Link: https://patch.msgid.link/20260317141856.237876-4-thomas.hellstrom@linux.intel.com

drm/ttm: Avoid invoking the OOM killer when reading back swapped content

2026-03-20T09:38:10+00:00

In situations where the system is very short on RAM, the shmem
readback from swap-space may invoke the OOM killer.

However, since this might be a recoverable situation where the caller
is indicating this by setting
struct ttm_operation_ctx::gfp_retry_mayfail to true, adjust the gfp
value used by the allocation accordingly.

Signed-off-by: Thomas Hellström 
Reviewed-by: Maarten Lankhorst 
Acked-by: Christian König 
Link: https://patch.msgid.link/20260317141856.237876-3-thomas.hellstrom@linux.intel.com

drm/ttm: rework pipelined eviction fence handling

2025-11-26T12:12:23+00:00

Until now ttm stored a single pipelined eviction fence which means
drivers had to use a single entity for these evictions.

To lift this requirement, this commit allows up to 8 entities to
be used.

Ideally a dma_resv object would have been used as a container of
the eviction fences, but the locking rules makes it complex.
dma_resv all have the same ww_class, which means "Attempting to
lock more mutexes after ww_acquire_done." is an error.

One alternative considered was to introduced a 2nd ww_class for
specific resv to hold a single "transient" lock (= the resv lock
would only be held for a short period, without taking any other
locks).

The other option, is to statically reserve a fence array, and
extend the existing code to deal with N fences, instead of 1.

The driver is still responsible to reserve the correct number
of fence slots.

Signed-off-by: Pierre-Eric Pelloux-Prayer 
Link: https://lore.kernel.org/r/20251121101315.3585-20-pierre-eric.pelloux-prayer@amd.com
Reviewed-by: Christian König 
Signed-off-by: Christian König

drm/ttm: Fix @alloc_flags description

2025-11-06T10:35:47+00:00

Stephen Rothwell reports htmldocs warnings when merging drm-misc tree:

Documentation/gpu/drm-mm:40: include/drm/ttm/ttm_device.h:225: ERROR: Unknown target name: "ttm_allocation". [docutils]
Documentation/gpu/drm-mm:43: drivers/gpu/drm/ttm/ttm_device.c:202: ERROR: Unknown target name: "ttm_allocation". [docutils]
Documentation/gpu/drm-mm:73: include/drm/ttm/ttm_pool.h:68: ERROR: Unknown target name: "ttm_allocation_pool". [docutils]
Documentation/gpu/drm-mm:76: drivers/gpu/drm/ttm/ttm_pool.c:1070: ERROR: Unknown target name: "ttm_allocation_pool". [docutils]

Fix these by adding missing wildcard on TTM_ALLOCATION_* and
TTM_ALLOCATION_POOL_* in @alloc_flags description.

Fixes: 0af5b6a8f8dd ("drm/ttm: Replace multiple booleans with flags in pool init")
Fixes: 77e19f8d3297 ("drm/ttm: Replace multiple booleans with flags in device init")
Fixes: 402b3a865090 ("drm/ttm: Add an allocation flag to propagate -ENOSPC on OOM")
Reported-by: Stephen Rothwell 
Closes: https://lore.kernel.org/linux-next/20251105161838.55b962a3@canb.auug.org.au/
Signed-off-by: Bagas Sanjaya 
Acked-by: Tvrtko Ursulin 
Reviewed-by: Christian König 
Signed-off-by: Tvrtko Ursulin 
Link: https://lore.kernel.org/r/20251106005217.14026-1-bagasdotme@gmail.com

drm/ttm: Add an allocation flag to propagate -ENOSPC on OOM

2025-10-31T09:27:28+00:00

Some graphics APIs differentiate between out-of-graphics-memory and
out-of-host-memory (system memory). Add a device init flag to have -ENOSPC
propagated from the resource managers instead of being converted to
-ENOMEM, to aid driver stacks in determining what error code to return or
whether corrective action can be taken at the driver level.

Co-developed-by: Thomas Hellström 
Cc: Christian König 
Cc: Matthew Brost 
Signed-off-by: Tvrtko Ursulin 
Reviewed-by: Thomas Hellström 
Reviewed-by: Christian König 
Signed-off-by: Tvrtko Ursulin 
Link: https://lore.kernel.org/r/20251020115411.36818-7-tvrtko.ursulin@igalia.com

drm/ttm: Allow drivers to specify maximum beneficial TTM pool size

2025-10-31T09:14:50+00:00

GPUs typically benefit from contiguous memory via reduced TLB pressure and
improved caching performance, where the maximum size of contiguous block
which adds a performance benefit is related to hardware design.

TTM pool allocator by default tries (hard) to allocate up to the system
MAX_PAGE_ORDER blocks. This varies by the CPU platform and can also be
configured via Kconfig.

If that limit was set to be higher than the GPU can make an extra use of,
lets allow the individual drivers to let TTM know over which allocation
order can the pool allocator afford to make a little bit less effort with.

We implement this by disabling direct reclaim for those allocations, which
reduces the allocation latency and lowers the demands on the page
allocator, in cases where expending this effort is not critical for the
GPU in question.

Signed-off-by: Tvrtko Ursulin 
Cc: Christian König 
Cc: Thadeu Lima de Souza Cascardo 
Reviewed-by: Christian König 
Signed-off-by: Tvrtko Ursulin 
Link: https://lore.kernel.org/r/20251020115411.36818-5-tvrtko.ursulin@igalia.com

drm/ttm: Replace multiple booleans with flags in device init

2025-10-31T09:14:35+00:00

Multiple consecutive boolean function arguments are usually not very
readable.

Replace the ones in ttm_device_init() with flags with the additional
benefit of soon being able to pass in more data with just a one off
code base churning cost.

Signed-off-by: Tvrtko Ursulin 
Cc: Alex Deucher 
Cc: Christian König 
Cc: Danilo Krummrich 
Cc: Dave Airlie 
Cc: Gerd Hoffmann 
Cc: Joonas Lahtinen 
Cc: Lucas De Marchi 
Cc: Lyude Paul 
Cc: Maarten Lankhorst 
Cc: Maxime Ripard 
Cc: Rodrigo Vivi 
Cc: Sui Jingfeng 
Cc: Thomas Hellström 
Cc: Thomas Zimmermann 
Cc: Zack Rusin 
Acked-by: Christian König 
Acked-by: Zack Rusin 
Acked-by: Thomas Hellström  # For xe
Reviewed-by: Christian König 
Signed-off-by: Tvrtko Ursulin 
Link: https://lore.kernel.org/r/20251020115411.36818-4-tvrtko.ursulin@igalia.com
[tursulin: fixup checkpatch while applying]

drm/ttm: Replace multiple booleans with flags in pool init

2025-10-31T09:01:08+00:00

Multiple consecutive boolean function arguments are usually not very
readable.

Replace the ones in ttm_pool_init() with flags with the additional
benefit of soon being able to pass in more data with just this one
code base churning cost.

Signed-off-by: Tvrtko Ursulin 
Cc: Alex Deucher 
Cc: Christian König 
Cc: Thomas Hellström 
Reviewed-by: Christian König 
Signed-off-by: Tvrtko Ursulin 
Link: https://lore.kernel.org/r/20251020115411.36818-3-tvrtko.ursulin@igalia.com