All the objects that need to implement some callbacks in KMS have a
pointer in there structure to the main drm_device.
However, it's not the case for drm_private_objs, which makes it harder
than it needs to be to implement some of its callbacks. Let's add that
pointer.
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Tested-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Link: https://patch.msgid.link/20251014-drm-private-obj-reset-v2-1-6dd60e985e9d@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Add a paragraph to the GEM Objects Creation section about the
drm_gem_huge_mnt_create() helper and to the GEM objects mapping
section explaining how transparent huge pages are handled by GEM.
v4:
- fix wording after huge_pages handler removal
v6:
- fix wording after map_pages handler removal
v11:
- mention drm_gem_huge_mnt_create() helper
- add Boris and Maíra R-bs
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Reviewed-by: Bagas Sanjaya <bagasdotme@gmail.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Maíra Canal <mcanal@igalia.com>
Link: https://patch.msgid.link/20251205182231.194072-11-loic.molinari@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Introduce the 'panfrost.transparent_hugepage' boolean module parameter
(false by default). When the parameter is set to true, a new tmpfs
mountpoint is created and mounted using the 'huge=within_size'
option. It's then used at GEM object creation instead of the default
'shm_mnt' mountpoint in order to enable Transparent Hugepage (THP) for
the object (without having to rely on a system wide parameter).
v3:
- use huge tmpfs mountpoint in drm_device
v4:
- fix builds with CONFIG_TRANSPARENT_HUGEPAGE=n
- clean up mountpoint creation error handling
- print negative error value
v5:
- use drm_gem_has_huge_tmp() helper
- get rid of CONFIG_TRANSPARENT_HUGEPAGE ifdefs
v9:
- replace drm_gem_has_huge_tmp() by drm_gem_get_huge_tmp()
v11:
- enable 'panfrost.transparent_hugepage' by default
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://patch.msgid.link/20251205182231.194072-10-loic.molinari@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Introduce the 'panthor.transparent_hugepage' boolean module parameter
(false by default). When the parameter is set to true, a new tmpfs
mountpoint is created and mounted using the 'huge=within_size'
option. It's then used at GEM object creation instead of the default
'shm_mnt' mountpoint in order to enable Transparent Hugepage (THP) for
the object (without having to rely on a system wide parameter).
v3:
- use huge tmpfs mountpoint in drm_device
v4:
- fix builds with CONFIG_TRANSPARENT_HUGEPAGE=n
- clean up mountpoint creation error handling
- print negative error value
v5:
- use drm_gem_has_huge_tmp() helper
- get rid of CONFIG_TRANSPARENT_HUGEPAGE ifdefs
v9:
- replace drm_gem_has_huge_tmp() by drm_gem_get_huge_tmp()
v11:
- enable 'panthor.transparent_hugepage' by default
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://patch.msgid.link/20251205182231.194072-9-loic.molinari@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
drm_gem_object_init_with_mnt() and drm_gem_shmem_create_with_mnt() can
be removed now that the drivers use the new drm_gem_huge_mnt_create()
and drm_gem_get_huge_mnt() helpers.
v5:
- use drm_gem_has_huge_mnt() helper
- compile out shmem_file_setup_with_mnt() call in builds with
CONFIG_TRANSPARENT_HUGEPAGE=n
v9:
- replace drm_gem_has_huge_mnt() with drm_gem_get_huge_mnt()
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Maíra Canal <mcanal@igalia.com>
Link: https://patch.msgid.link/20251205182231.194072-8-loic.molinari@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Make use of the new drm_gem_huge_mnt_create() and
drm_gem_get_huge_mnt() helpers to avoid code duplication. Now that
it's just a few lines long, the single function in v3d_gemfs.c is
moved into v3d_gem.c.
v3:
- use huge tmpfs mountpoint in drm_device
- move v3d_gemfs.c into v3d_gem.c
v4:
- clean up mountpoint creation error handling
v5:
- fix CONFIG_TRANSPARENT_HUGEPAGE check
- use drm_gem_has_huge_mnt() helper
v8:
- don't access huge_mnt field with CONFIG_TRANSPARENT_HUGEPAGE=n
v9:
- replace drm_gem_has_huge_mnt() by drm_gem_get_huge_mnt()
v10:
- get rid of CONFIG_TRANSPARENT_HUGEPAGE ifdefs
v11:
- remove superfluous comment
- add Maíra and Boris R-bs
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Reviewed-by: Maíra Canal <mcanal@igalia.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://patch.msgid.link/20251205182231.194072-7-loic.molinari@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Make use of the new drm_gem_huge_mnt_create() and
drm_gem_get_huge_mnt() helpers to avoid code duplication. Now that
it's just a few lines long, the single function in i915_gemfs.c is
moved into i915_gem_shmem.c.
v3:
- use huge tmpfs mountpoint in drm_device
- move i915_gemfs.c into i915_gem_shmem.c
v4:
- clean up mountpoint creation error handling
v5:
- use drm_gem_has_huge_mnt() helper
v7:
- include <drm/drm_print.h> in i915_gem_shmem.c
v8:
- keep logging notice message with CONFIG_TRANSPARENT_HUGEPAGE=n
- don't access huge_mnt field with CONFIG_TRANSPARENT_HUGEPAGE=n
v9:
- replace drm_gem_has_huge_mnt() by drm_gem_get_huge_mnt()
- remove useless ternary op test in selftests/huge_pages.c
v12:
- fix layering violation in selftests (Tvrtko)
- fix incorrect filename in commit message
v13:
- add Tvrtko A-b
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Acked-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://patch.msgid.link/20251205182231.194072-6-loic.molinari@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Add the drm_gem_huge_mnt_create() and drm_gem_get_huge_mnt() helpers
to avoid code duplication in the i915, V3D, Panfrost and Panthor
drivers. The former creates and mounts a dedicated huge tmpfs
mountpoint, for the lifetime of a DRM device, used at GEM object
initialization. The latter retrieves the dedicated huge tmpfs
mountpoint used by a DRM device.
The next commits will port drivers to these helpers.
v3:
- store huge tmpfs mountpoint in drm_device
v4:
- return 0 in builds with CONFIG_TRANSPARENT_HUGEPAGE=n
- return 0 when huge_mnt already exists
- use new vfs_parse_fs_string() helper
v5:
- remove warning on !dev->huge_mnt and reset to NULL on free
- inline drm_gem_huge_mnt_create() to remove func from text and avoid
calls in builds with CONFIG_TRANSPARENT_HUGEPAGE=n
- compile out drm_device's huge_mnt field in builds with
CONFIG_TRANSPARENT_HUGEPAGE=n
- add drm_gem_has_huge_mnt() helper
v6:
- move huge_mnt doc into ifdef'd section
- either inline or export drm_gem_huge_mnt_create()
v7:
- include <drm/drm_device.h> in drm_gem.h
v9:
- replace drm_gem_has_huge_mnt() by drm_gem_get_huge_mnt()
v11:
- doc fixes
- add Boris and Maíra R-bs
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Maíra Canal <mcanal@igalia.com>
Link: https://patch.msgid.link/20251205182231.194072-5-loic.molinari@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
mmap() calls on the DRM file pointer currently always end up using
mm_get_unmapped_area() to get a free mapping region. On builds with
CONFIG_TRANSPARENT_HUGEPAGE enabled, this isn't ideal for GEM objects
backed by shmem buffers on mountpoints setting the 'huge=' option
because it can't correctly figure out the potentially huge address
alignment required.
This commit introduces the drm_gem_get_unmapped_area() function which
is meant to be used as a get_unmapped_area file operation on the DRM
file pointer to lookup GEM objects based on their fake offsets and get
a properly aligned region by calling shmem_get_unmapped_area() with
the right file pointer. If a GEM object isn't available at the given
offset or if the caller isn't granted access to it, the function falls
back to mm_get_unmapped_area().
This also makes drm_gem_get_unmapped_area() part of the default GEM
file operations so that all the DRM drivers can benefit from more
efficient mappings thanks to the huge page fault handler introduced in
previous commit 'drm/shmem-helper: Add huge page fault handler'.
The shmem_get_unmapped_area() function needs to be exported so that
it can be used from the DRM subsystem.
v3:
- include <linux/sched/mm.h> in drm_gem.c
- forward to shmem layer in builds with CONFIG_TRANSPARENT_HUGEPAGE=n
v6:
- use GPL variant to export drm_gem_get_unmapped_area()
- don't export shmem_get_unmapped_area() anymore (use f_op instead)
v11:
- rename drm_gem_object_lookup_from_offset() to
drm_gem_object_lookup_at_offset()
- add Boris R-b
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://patch.msgid.link/20251205182231.194072-4-loic.molinari@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Attempt a PMD sized PFN insertion into the VMA if the faulty address
of the fault handler is part of a huge page.
On builds with CONFIG_TRANSPARENT_HUGEPAGE enabled, if the mmap() user
address is PMD size aligned, if the GEM object is backed by shmem
buffers on mountpoints setting the 'huge=' option and if the shmem
backing store manages to allocate a huge folio, CPU mapping would then
benefit from significantly increased memcpy() performance. When these
conditions are met on a system with 2 MiB huge pages, an aligned copy
of 2 MiB would raise a single page fault instead of 4096.
v4:
- implement map_pages instead of huge_fault
v6:
- get rid of map_pages handler for now (keep it for another series
along with arm64 contpte support)
v11:
- remove page fault validity check helper
- rename drm_gem_shmem_map_pmd() to drm_gem_shmem_try_map_pmd()
- add Boris R-b
v12:
- move up ret var decl in fault handler to minimize diff
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://patch.msgid.link/20251205182231.194072-3-loic.molinari@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
For a fault at address addr, the page offset is
page_offset = (vmf->address - vma->vm_start) >> PAGE_SHIFT
= ((addr & PAGE_MASK) - vma->vm_start) >> PAGE_SHIFT
= (addr - vma->vm_start) >> PAGE_SHIFT
Since the faulty logical page offset based on VMA is
vmf->pgoff = vma->vm_pgoff + ((addr - vma->vm_start) >> PAGE_SHIFT)
We can slightly simplify the calculation using
page_offset = vmf->pgoff - vma->vm_pgoff
v11:
- remove misleading comment
- add Boris R-b
Signed-off-by: Loïc Molinari <loic.molinari@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://patch.msgid.link/20251205182231.194072-2-loic.molinari@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Set the GLB_COUNTER_EN bit to enable coherent propagation of GPU
timestamp values to shader cores. This is a prerequisite for exposing
device-coherent timestamp queries.
Bump the version to 1.6 so userspace can detect support.
v2:
- GLB_COUNTER_EN is not a toggle bit move to panthor_fw_update_reqs
Signed-off-by: Ashley Smith <ashley.smith@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251127115019.2113040-1-ashley.smith@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
All functions used for signaling a fence return an error code whose sole
purpose is to tell whether a fence was already signaled.
This is racy and has been used by almost no party in the kernel, and the
few users have been removed in preceding cleanup commits.
Turn all signaling-functions into void-functions.
Suggested-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251201105011.19386-9-phasta@kernel.org
The return code of dma_fence_signal() is not useful and shall be removed
from the kernel. To do so, the few users who rely on the return code
must be ported.
Use dma_fence_check_and_signal() and mapp its boolean return code to
dma_fence_signal()'s former value for already-signaled fences.
Suggested-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251201105011.19386-8-phasta@kernel.org
The return code of dma_fence_signal() is not really useful as there is
nothing reasonable to do if a fence was already signaled. That return
code shall be removed from the kernel.
Moreover, dma_fence_signal() should not be used to check whether fences
are signaled. That's what dma_fence_is_signaled() and
dma_fence_test_signaled_flag() exist for.
Replace the non-canonical usage of dma_fence_signal().
Suggested-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251201105011.19386-7-phasta@kernel.org
Xe is one of the few users utilizing the return code of
dma_fence_signal() to check whether a fence had already been signaled by
someone else.
To clean up and simplify the dma_fence API, the few kernel users relying
on that behavior shall be ported to an alternative function.
Replace dma_fence_signal_locked() with
dma_fence_check_and_signal_locked().
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251201105011.19386-6-phasta@kernel.org
amdkfd is one of the few users which relies on the return code of
dma_fence_signal(), which, so far, informs the caller whether the fence
had already been signaled.
As there are barely any users, dma_fence signaling functions shall get
the return value void. To do so, the few users must be ported to a
function which preserves the old behavior.
Replace the call to dma_fence_signal() with one to
dma_fence_check_and_signal().
Suggested-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251201105011.19386-5-phasta@kernel.org
The overwhelming majority of users of dma_fence signaling functions
don't care about whether the fence had already been signaled by someone
else. Therefore, the return code shall be removed from those functions.
For the few users who rely on the check, a new, specialized function
shall be provided.
Add dma_fence_check_and_signal(), which signals a fence if it had not
yet been signaled, and informs the user about that.
Add a counter part, dma_fence_check_and_signal_locked(), which doesn't
take the spinlock.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251201105011.19386-4-phasta@kernel.org
The dma_fence framework checks at many places whether the signaled flag
of a fence is already set. The code can be simplified and made more
readable by providing a helper function for that.
Add dma_fence_test_signaled_flag(), which only checks whether a fence is
signaled. Use it internally.
Suggested-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251201105011.19386-3-phasta@kernel.org
Add an eDP panel entry for AUO B140HAN06.4 that is also used in
some variants of Lenovo Flex 5G with Qcom SC8180 SoC.
The raw edid of the panel is:
00 ff ff ff ff ff ff 00 06 af 3d 64 00 00 00 00
2b 1d 01 04 a5 1f 11 78 03 b8 1a a6 54 4a 9b 26
0e 52 55 00 00 00 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 14 37 80 b8 70 38 24 40 10 10
3e 00 35 ae 10 00 00 18 10 2c 80 b8 70 38 24 40
10 10 3e 00 35 ae 10 00 00 18 00 00 00 fe 00 41
55 4f 0a 20 20 20 20 20 20 20 20 20 00 00 00 fe
00 42 31 34 30 48 41 4e 30 36 2e 34 20 0a 00 eb
I do not have access to the datasheet and but it is tested on above
mentioned laptop for a few weeks and seems to work just fine with
timing info of similar panels.
Cc: Bjorn Andersson <andersson@kernel.org>
Cc: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Alexey Klimov <alexey.klimov@linaro.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://patch.msgid.link/20251203074555.690613-1-alexey.klimov@linaro.org
At the moment the TTM code has a few places which exibit sub-optimal
patterns regarding local variable usage:
* Having a local with some object cached but not always using it.
* Having a local for a single use object member access.
* Failed opportunities to use a local to cache a pointer.
Lets tidy this a little bit and apply some more consistency.
It is mostly for consistency and redability but I have also checked that
there are not negative code generation effects. In fact there are more
positives:
add/remove: 0/0 grow/shrink: 3/9 up/down: 12/-175 (-163)
Function old new delta
ttm_pool_restore_and_alloc 415 423 +8
ttm_bo_vunmap 147 149 +2
ttm_bo_evict 521 523 +2
ttm_bo_vm_fault_reserved 972 970 -2
ttm_bo_vm_dummy_page 155 152 -3
ttm_bo_vm_fault 203 196 -7
ttm_bo_populate 158 150 -8
ttm_bo_move_memcpy 600 592 -8
ttm_bo_kmap 667 644 -23
ttm_bo_shrink 333 305 -28
ttm_bo_release 750 720 -30
ttm_bo_swapout_cb 691 625 -66
Total: Before=42717, After=42554, chg -0.38%
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Reviewed-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Link: https://lore.kernel.org/r/20250919131530.91247-5-tvrtko.ursulin@igalia.com
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
[tursulin: fixup conflict in ttm_bo_move_pipeline_evict]
No need to initialize a subset of fields to zero.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Acked-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Link: https://lore.kernel.org/r/20250919131530.91247-4-tvrtko.ursulin@igalia.com
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
[tursulin: fixup conflict in ttm_resource_manager_evict_all]
Function has a single caller and the resource cannot be NULL therefore
remove the early return check.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Reviewed-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20250919131530.91247-3-tvrtko.ursulin@igalia.com
Struct ttm_operation_ctx initializer in ttm_bo_init_validate assumes the
order of the structure fields when it is configuring the interruptible
flag.
Fix it by using named initialization.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Acked-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20250919131530.91247-2-tvrtko.ursulin@igalia.com
queue_check_job_completion calls queue_reset_timeout_locked to reset the
timeout when progress is made. We want the reset to happen when the
timeout is running, not when it is suspended.
Fixes: 345c5b7cc0 ("drm/panthor: Make the timeout per-queue instead of per-job")
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Link: https://patch.msgid.link/20251202174028.1600218-1-olvaffe@gmail.com
This commit removes the redundant call to disable the MCU firmware
in the suspend path.
Fixes: 5140725498 ("drm/panthor: Support GLB_REQ.STATE field for Mali-G1 GPUs")
Signed-off-by: Akash Goel <akash.goel@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Link: https://patch.msgid.link/20251203091911.145623-1-akash.goel@arm.com
An AS can be disabled in the middle of a VM operation (VM being
evicted from an AS slot, for instance). In that case, we need the
locked section to be unlocked before releasing the slot.
v2:
- Add an lockdep_assert_held() in panthor_mmu_as_disable()
- Collect R-bs
v3:
- Don't reset the locked_region range in the as_disable() path
Fixes: 6e2d3b3e85 ("drm/panthor: Add support for atomic page table updates")
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Liviu Dudau <liviu.dudau@arm.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Link: https://patch.msgid.link/20251203121750.404340-4-boris.brezillon@collabora.com
When we re-assign a slot to a different VM, we need to make sure the
old VM caches are flushed before doing the switch. Specialize
panthor_mmu_as_disable() so we can skip the slot programmation while
still getting the cache flushing, and call this helper from
panthor_vm_active() when an idle slot is recycled.
v2:
- Collect R-bs
Fixes: 6e2d3b3e85 ("drm/panthor: Add support for atomic page table updates")
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Liviu Dudau <liviu.dudau@arm.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Link: https://patch.msgid.link/20251203121750.404340-3-boris.brezillon@collabora.com
It appears the timeout can still be enabled when we reach that point,
because of the asynchronous progress check done on queues that resets
the timer when jobs are still in-flight, but progress was made.
We could add more checks to make sure the timer is not re-enabled when
a group can't run anymore, but we don't have a group to pass to
queue_check_job_completion() in some context.
It's just as safe (we just want to be sure the timer is stopped before
we destroy the queue) and simpler to drop the WARN_ON() in
group_free_queue().
v2:
- Collect R-bs
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Liviu Dudau <liviu.dudau@arm.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Link: https://patch.msgid.link/20251203121750.404340-2-boris.brezillon@collabora.com
No need to conditional compile that code, let the compilers dead code
elimination handle it instead.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Link: https://lore.kernel.org/r/20251006134713.1846-1-christian.koenig@amd.com
After issuing a firmware suspend request, the driver must ensure that the
suspend operation has completed before proceeding. Add polling of the
MPNPU_PWAITMODE register to confirm that the firmware has fully entered
the suspended state. This prevents race conditions where subsequent
operations assume the firmware is idle before it has actually completed
its suspend sequence.
Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org>
Reviewed-by: Maciej Falkowski <maciej.falkowski@linux.intel.com>
Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
Link: https://patch.msgid.link/20251202165427.507414-1-lizhi.hou@amd.com
Runqueues are currently almost everywhere being read unlocked in
drm/sched. At XDC 2025, the assembled developers were unsure whether
that's legal and whether it can be fixed. Someone should find out.
Add a todo entry for the unlocked runqueue reader problem.
Acked-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251107135701.244659-4-phasta@kernel.org
The GPU scheduler has a great many problems and deserves its own TODO
section.
Add a section and a first task describing the problem of
drm_sched_resubmit_jobs() being deprecated without a successor.
Acked-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251107135701.244659-3-phasta@kernel.org
There are no implementations of fb_debug_enter and fb_debug_leave.
Remove the callbacks from struct fb_ops and clean up the caller.
The field save_graphics in fbcon_par is also no longer required.
Remove it as well.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Simona Vetter <simona.vetter@ffwll.ch>
Acked-by: Daniel Thompson (RISCstar) <danielt@kernel.org>
Link: https://patch.msgid.link/20251125130634.1080966-6-tzimmermann@suse.de
Remove the debug_enter/debug_leave helpers, as there are no DRM
drivers supporting debugging with kgdb. Remove code to keep track
of existing fbdev-emulation state. None of this required any longer.
Also remove mode_set_base_atomic from struct drm_crtc_helper_funcs,
which has no callers or implementations.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Simona Vetter <simona.vetter@ffwll.ch>
Acked-by: Daniel Thompson (RISCstar) <danielt@kernel.org>
Link: https://patch.msgid.link/20251125130634.1080966-5-tzimmermann@suse.de
Remove the implementation of the CRTC helper mode_set_base_atomic
from radeon. It pretends to provide mode setting for kdb debugging,
but has been broken for some time.
Kdb output has been supported only for non-atomic mode setting since
commit 9c79e0b1d0 ("drm/fb-helper: Give up on kgdb for atomic drivers")
from 2017.
While radeon provides non-atomic mode setting, kdb assumes that the GEM
buffer object is at a fixed location in video memory. This assumption
currently blocks radeon from converting to generic fbdev emulation.
Fbdev-ttm helpers use a shadow buffer with a movable GEM buffer object.
Triggering kdb does therefore not update the display.
Another problem is that the current implementation does not handle
USB keyboard input. Therefore a serial terminal is required. Then when
continuing from the debugger, radeon fails with an error:
[7]kdb> go
[ 40.345523][ C7] BUG: scheduling while atomic: bash/1580/0x00110003
[...]
[ 40.345613][ C7] schedule+0x27/0xd0
[ 40.345615][ C7] schedule_timeout+0x7b/0x100
[ 40.345617][ C7] ? __pfx_process_timeout+0x10/0x10
[ 40.345619][ C7] msleep+0x31/0x50
[ 40.345621][ C7] radeon_crtc_load_lut+0x2e4/0xcb0 [radeon 31c1ee785de120fcfd0babcc09babb3770252b4e]
[ 40.345698][ C7] radeon_crtc_gamma_set+0xe/0x20 [radeon 31c1ee785de120fcfd0babcc09babb3770252b4e]
[ 40.345760][ C7] drm_fb_helper_debug_leave+0xd8/0x130
[ 40.345763][ C7] kgdboc_post_exp_handler+0x54/0x70
[...]
and the system hangs.
Support for kdb feels pretty much broken. Hence remove the whole kdb
support from radeon.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Simona Vetter <simona.vetter@ffwll.ch>
Acked-by: Daniel Thompson (RISCstar) <danielt@kernel.org>
Link: https://patch.msgid.link/20251125130634.1080966-4-tzimmermann@suse.de
Remove the implementation of the CRTC helper mode_set_base_atomic
from nouveau. It pretends to provide mode setting for kdb debugging,
but has been broken for some time.
Kdb output has been supported only for non-atomic mode setting since
commit 9c79e0b1d0 ("drm/fb-helper: Give up on kgdb for atomic drivers")
from 2017.
While nouveau provides non-atomic mode setting for some devices, kdb
assumes that the GEM buffer object is at a fixed location in video
memory. This has not been the case since
commit 4a16dd9d18 ("drm/nouveau/kms: switch to drm fbdev helpers")
from 2022. Fbdev-ttm helpers use a shadow buffer with a movable GEM
buffer object. Triggering kdb does therefore not update the display.
Hence remove the whole kdb support from nouveau.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Acked-by: Simona Vetter <simona.vetter@ffwll.ch>
Acked-by: Daniel Thompson (RISCstar) <danielt@kernel.org>
Link: https://patch.msgid.link/20251125130634.1080966-3-tzimmermann@suse.de
Remove all implementations of the CRTC helper mode_set_base_atomic
from amdgpu. It pretends to provide mode setting for kdb debugging,
but has been broken for some time.
Kdb output has been supported only for non-atomic mode setting since
commit 9c79e0b1d0 ("drm/fb-helper: Give up on kgdb for atomic drivers")
from 2017.
While amdgpu provides non-atomic mode setting for some devices, kdb
assumes that the GEM buffer object is at a fixed location in video
memory. This has not been the case since commit 087451f372 ("drm/amdgpu:
use generic fb helpers instead of setting up AMD own's.") from 2021.
Fbdev-ttm helpers use a shadow buffer with a movable GEM buffer object.
Triggering kdb does not update the display.
Hence remove the whole kdb support from amdgpu.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Simona Vetter <simona.vetter@ffwll.ch>
Acked-by: Daniel Thompson (RISCstar) <danielt@kernel.org>
Link: https://patch.msgid.link/20251125130634.1080966-2-tzimmermann@suse.de
Call drm_gem_fb_begin_cpu_access() and drm_gem_fb_end_cpu_access()
around cursor image updates. Imported buffers might have to be
synchronized for CPU access before they can be used.
Ignore errors from drm_gem_fb_begin_cpu_access(). These errors can
often be transitory. The cursor image will be updated on the next
frame. Meanwhile display a white square where the cursor would be.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>>
Link: https://patch.msgid.link/20251126094626.41985-4-tzimmermann@suse.de
Copy the ARGB4444 cursor buffer to system memory if it is located in
I/O memory. While this cannot happen with ast's native GEM objects, an
imported buffer object might be on the external device's I/O memory.
If the cursor buffer is located in system memory continue to use it
directly.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>>
Link: https://patch.msgid.link/20251126094626.41985-3-tzimmermann@suse.de
Move the format conversion of the cursor framebuffer into the new
helper ast_cursor_plane_get_argb4444(). It returns a buffer in system
memory, which the atomic_update handler copies to video memory.
The returned buffer is either the GEM buffer itself, or a temporary
copy within the plane in ARGB4444 format.
As a small change, list supported formats explicitly in the switch
statement. Do not assume ARGB8888 input by default. The cursor
framebuffer knows its format, so should we.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patch.msgid.link/20251126094626.41985-2-tzimmermann@suse.de
Currently the user can write anything into the drm.panic_screen modparam,
either at runtime via sysfs, or as a kernel boot time argument. Invalid
strings will be silently accepted and ignored at use time by defaulting to
the 'user' panic mode.
Let instead add some validation in order to have immediate feedback when
something has been mistyped, or not compiled in.
For example during kernel boot:
Booting kernel: `bsod' invalid for parameter `drm.panic_screen'
Or at runtime:
# echo -n bsod > /sys/module/drm/parameters/panic_screen
-bash: echo: write error: Invalid argument
Change of behavior is that when invalid mode is attempted to be
configured, currently the code will default to the 'user' mode, while with
this change the code will ignore it, and default to the mode set at kernel
build time via CONFIG_DRM_PANIC_SCREEN.
While at it lets also fix the module parameter description to include all
compiled in modes.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Jocelyn Falempe <jfalempe@redhat.com>
Cc: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20251127090349.92717-1-tvrtko.ursulin@igalia.com
It's only used in a couple places and everyone else is just using
sched_queue_delayed_work(sched, tick, 0) directly, so let's make
this consistent.
v2:
- Add R-b
v3:
- Collect R-b
Reviewed-by: Steven Price <steven.price@arm.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Link: https://patch.msgid.link/20251128094839.3856402-9-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
If the group is already assigned a slot but was idle before this job
submission, we need to make sure the priority rotation happens in the
future. Extract the existing logic living in group_schedule_locked()
and call this new sched_resume_tick() helper from the "group is
assigned a slot" path.
v2:
- Add R-b
v3:
- Re-use queue_mask to clear the bit
- Collect R-b
Fixes: de85488138 ("drm/panthor: Add the scheduler logical block")
Reviewed-by: Steven Price <steven.price@arm.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Link: https://patch.msgid.link/20251128094839.3856402-8-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
When we have multiple active groups with the same priority, we need to
keep ticking for the priority rotation to take place. If we don't do
that, we might starve slots with lower priorities.
It's annoying to deal with that in tick_ctx_update_resched_target(),
so let's add a ::stop_tick field to the tick context which is
initialized to true, and downgraded to false as soon as we detect
something that requires to tick to happen. This way we can complement
the current logic with extra conditions if needed.
v2:
- Add R-b
v3:
- Drop panthor_sched_tick_ctx::min_priority (no longer relevant)
- Collect R-b
Fixes: de85488138 ("drm/panthor: Add the scheduler logical block")
Reviewed-by: Steven Price <steven.price@arm.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Link: https://patch.msgid.link/20251128094839.3856402-7-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
We have a few paths where we schedule the tick work immediately without
changing the resched_target. If the tick was stopped, this would lead
to a remaining_jiffies that's always > 0, and it wouldn't force a full
tick in that case. Add extra checks to cover that case properly.
v2:
- Fix typo
- Simplify the code as suggested by Steve
v3:
- Collect R-b
Fixes: de85488138 ("drm/panthor: Add the scheduler logical block")
Reviewed-by: Steven Price <steven.price@arm.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Link: https://patch.msgid.link/20251128094839.3856402-6-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
When rotating group priorities, we want the group with the
highest priority to go back to the end of the queue, and all
other active groups to get their priority bumped, otherwise
some groups will never get a chance to run with the highest
priority. This implies moving the rotation itself to
tick_work(), and only dealing with old group ordering in
tick_ctx_insert_old_group().
v2:
- Add R-b
- Fix the commit message
v3:
- Drop the full_tick argument in tick_ctx_init()
- Collect R-b
Fixes: de85488138 ("drm/panthor: Add the scheduler logical block")
Reviewed-by: Steven Price <steven.price@arm.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Link: https://patch.msgid.link/20251128094839.3856402-5-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
We have a full tick when the remaining time to the next tick is zero,
not the other way around. Declare a full_tick variable so we don't get
that test wrong in other places.
v2:
- Add R-b
v3:
- Collect R-b
Fixes: de85488138 ("drm/panthor: Add the scheduler logical block")
Reviewed-by: Steven Price <steven.price@arm.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Link: https://patch.msgid.link/20251128094839.3856402-4-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Not only this only works once, because of how extract events work
(event is enabled if the req and ack bit differ, and it's signalled
by the FW by setting identical req and ack, to re-enable the event,
we need to toggle the bit, which we never do). But more importantly,
we never do anything with this event, so we're better off dropping it
when programming the CS slot.
v2:
- Add R-b
v3:
- Collect R-b
Reviewed-by: Steven Price <steven.price@arm.com>
Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Link: https://patch.msgid.link/20251128094839.3856402-3-boris.brezillon@collabora.com
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
