With the goal of reducing the need for drivers to touch (and dereference)
fence->ops, we move the 64-bit seqnos flag from struct dma_fence_ops to
the fence->flags.
Drivers which were setting this flag are changed to use new
dma_fence_init64() instead of dma_fence_init().
v2:
* Streamlined init and added kerneldoc.
* Rebase for amdgpu userq which landed since.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Reviewed-by: Christian König <christian.koenig@amd.com> # v1
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20250515095004.28318-3-tvrtko.ursulin@igalia.com
Currently the TTM shrinker aborts shrinking as soon as it frees pages from
any of the page order pools and by doing so it can fail to respect the
freeing target which was configured by the shrinker core.
We use the wording "can fail" because the number of freed pages will
depend on the presence of pages in the pools and the order of the pools on
the LRU list. For example if there are no free pages in the high order
pools the shrinker core may require multiple passes over the TTM shrinker
before it will free the default target of 128 pages (assuming there are
free pages in the low order pools). This inefficiency can be compounded by
the pool LRU where multiple further calls into the TTM shrinker are
required to end up looking at the pool with pages.
Improve this by never freeing less than the shrinker core has requested.
At the same time we start reporting the number of scanned pages (freed in
this case), which prevents the core shrinker from giving up on the TTM
shrinker too soon and moving on.
v2:
* Simplify loop logic. (Christian)
* Improve commit message.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20250603112750.34997-2-tvrtko.ursulin@igalia.com
[WHY]
1. Drivers using DRM_GEM_SHADOW_PLANE_HELPER_FUNCS and
DRM_GEM_SHMEM_DRIVER_OPS (e.g., udl, ast) do not require
sg_table import.
They only need dma_buf_vmap() to access the shared buffer's
kernel virtual address.
2. On certain Aspeed-based boards, a dma_mask of 0xffff_ffff may
trigger SWIOTLB during dmabuf import. However, IO_TLB_SEGSIZE
restricts the maximum DMA streaming mapping memory, resulting in
errors like:
ast 0000:07:00.0: swiotlb buffer is full (sz: 3145728 bytes), total 32768 (slots), used 0 (slots)
[HOW]
Provide a gem_prime_import implementation without sg_table mapping
to avoid issues (e.g., "swiotlb buffer is full"). Drivers that do not
require sg_table can adopt this.
Signed-off-by: Shixiong Ou <oushixiong@kylinos.cn>
Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://lore.kernel.org/r/20250522070714.439824-1-oushixiong1025@163.com
This change is essentially a Panfrost port of commit a3707f53eb
("drm/panthor: show device-wide list of DRM GEM objects over DebugFS").
The DebugFS file is almost the same as in Panthor, minus the GEM object
usage flags, since Panfrost has no kernel-only BO's.
Two additional GEM state flags which are displayed but aren't relevant
to Panthor are 'Purged' and 'Purgeable', since Panfrost implements an
explicit shrinker and a madvise ioctl to flag objects as reclaimable.
Signed-off-by: Adrián Larumbe <adrian.larumbe@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Signed-off-by: Steven Price <steven.price@arm.com>
Link: https://lore.kernel.org/r/20250520174634.353267-5-adrian.larumbe@collabora.com
Perfcnt samples buffer is not exposed to UM, but we would like to keep
a tag on it as a potential debug aid.
PRIME imported GEM buffers are UM exposed, but since the usual Panfrost
UM driver code path is not followed in their creation, they might remain
unlabelled for their entire lifetime, so a generic tag was deemed
preferable. The tag is assigned before a UM handle is created so it
doesn't contradict the logic about labelling internal BOs.
Signed-off-by: Adrián Larumbe <adrian.larumbe@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Signed-off-by: Steven Price <steven.price@arm.com>
Link: https://lore.kernel.org/r/20250520174634.353267-3-adrian.larumbe@collabora.com
There is no need for separate locks for single jobs and the entire
scheduler. The dma_fence context can be protected by the scheduler lock,
allowing for removing the jobs' locks. This simplifies things and
reduces the likelyhood of deadlocks etc.
Replace the jobs' locks with the mock scheduler lock.
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250527101029.56491-2-phasta@kernel.org
Replace the calls to drm_atomic_get_connector_state() with
drm_atomic_get_new_connector_state() for cases which do not require
allocating the connector state, e.g. after drm_atomic_check_only() when
the intent is to only read the new connector state.
The rational is to avoid the need to handle the potential EDEADLK error
returned by the former helper, which would require restarting the entire
atomic sequence.
Acked-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Link: https://lore.kernel.org/r/20250527-hdmi-conn-yuv-v5-13-74c9c4a8ac0c@collabora.com
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Factor out the HDMI connector initialization from
drm_kunit_helper_connector_hdmi_init_funcs() into a common
__connector_hdmi_init() function, while extending its functionality to
allow setting custom (i.e. non-default) EDID data.
Introduce a macro as a wrapper over the new helper to allow dropping the
open coded EDID setup from all test cases.
The actual conversion will be handled separately; for now just apply it
to drm_kunit_helper_connector_hdmi_init() helper.
Reviewed-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Link: https://lore.kernel.org/r/20250527-hdmi-conn-yuv-v5-9-74c9c4a8ac0c@collabora.com
Signed-off-by: Maxime Ripard <mripard@kernel.org>
In preparation to support fallback to an alternative output format, e.g.
YUV420, when RGB cannot be used for any of the available color depths,
move the bpc try loop out of hdmi_compute_config() and, instead, make it
part of hdmi_compute_format(), while adding a new parameter to the
latter holding the output format to be checked and eventually set.
Since this helper now also changes hdmi.output_bpc in addition to
hdmi.output_format, highlight the extended functionality by renaming it
to hdmi_compute_format_bpc().
This improves code reusability and further extensibility, without
introducing any functional changes.
Reviewed-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Link: https://lore.kernel.org/r/20250527-hdmi-conn-yuv-v5-5-74c9c4a8ac0c@collabora.com
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Evaluating the requirement to use a limited RGB quantization range
involves a verification of the output format, among others, but this is
currently performed before actually computing the format, hence relying
on the old connector state.
Move the call to hdmi_is_limited_range() after hdmi_compute_config() to
ensure the verification is done on the updated output format.
Fixes: 027d435906 ("drm/connector: hdmi: Add RGB Quantization Range to the connector state")
Reviewed-by: Dmitry Baryshkov <lumag@kernel.org>
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20250527-hdmi-conn-yuv-v5-1-74c9c4a8ac0c@collabora.com
Signed-off-by: Maxime Ripard <mripard@kernel.org>
The DE33 is a newer version of the Allwinner Display Engine IP block,
found in the H616, H618, H700 and T507 SoCs. DE2 and DE3 are already
supported by the mainline driver.
Notable features (from the H616 datasheet and implemented):
- 4096 x 2048 (4K) output support
Other features (implemented but not in this patchset):
- AFBC ARM Frame Buffer Compression support
- YUV pipeline support
The DE2 and DE3 engines have a blender register range within the
mixer engine register map, whereas the DE33 separates this out into
a separate display group, and adds a top register map.
The DE33 also appears to remove the global double buffer control
register, present in the DE2 and DE3.
Extend the mixer to support the DE33.
Signed-off-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Signed-off-by: Ryan Walklin <ryan@testtoast.com>
Signed-off-by: Chris Morgan <macromorgan@hotmail.com>
Acked-by: Maxime Ripard <mripard@kernel.org>
Reviewed-by: Chen-Yu Tsai <wens@csie.org>
Link: https://lore.kernel.org/r/20250528092431.28825-7-ryan@testtoast.com
Signed-off-by: Maxime Ripard <mripard@kernel.org>
