The XE_PL_TT watermark was set to 50% of system memory.
The idea behind that was unclear since the net effect is that
TT memory will be evicted to TTM_PL_SYSTEM memory if that
watermark is exceeded, requiring PPGTT rebinds and dma
remapping. But there is no similar watermark for TTM_PL_1SYSTEM
memory.
The TTM functionality that tries to swap out system memory to
shmem objects if a 50% limit of total system memory is reached
is orthogonal to this, and with the shrinker added, it's no
longer in effect.
Replace the 50% TTM_PL_TT limit with a 100% limit, in effect
allowing all graphics memory to be bound to the device unless it
has been swapped out by the shrinker.
Rather than relying on the TTM watermark accounting add a shrinker
for xe_bos in TT or system memory.
Leverage the newly added TTM per-page shrinking and shmem backup
support.
Although xe doesn't fully support WONTNEED (purgeable) bos yet,
introduce and add shrinker support for purgeable ttm_tts.
v2:
- Cleanups bugfixes and a KUNIT shrinker test.
- Add writeback support, and activate if kswapd.
v3:
- Move the try_shrink() helper to core TTM.
- Minor cleanups.
v4:
- Add runtime pm for the shrinker. Shrinking may require an active
device for CCS metadata copying.
v5:
- Separately purge ghost- and zombie objects in the shrinker.
- Fix a format specifier - type inconsistency. (Kernel test robot).
v7:
- s/long/s64/ (Christian König)
- s/sofar/progress/ (Matt Brost)
v8:
- Rebase on Xe KUNIT update.
- Add content verifying to the shrinker kunit test.
- Split out TTM changes to a separate patch.
- Get rid of multiple bool arguments for clarity (Matt Brost)
- Avoid an error pointer dereference (Matt Brost)
- Avoid an integer overflow (Matt Auld)
- Address misc review comments by Matt Brost.
v9:
- Fix a compliation error.
- Rebase.
v10:
- Update to new LRU walk interface.
- Rework ghost-, zombie and purged object shrinking.
- Rebase.
v11:
- Use additional TTM helpers.
- Honor __GFP_FS and __GFP_IO
- Rebase.
v13:
- Use ttm_tt_setup_backup().
v14:
- Don't set up backup on imported bos.
v15:
- Rebase on backup interface changes.
Cc: Christian König <christian.koenig@amd.com> Cc: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com> Cc: Matthew Brost <matthew.brost@intel.com> Cc: <dri-devel@lists.freedesktop.org> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> Reviewed-by: Matthew Brost <matthew.brost@intel.com> Acked-by: Christian König <christian.koenig@amd.com> Link: https://lore.kernel.org/intel-xe/20250305092220.123405-7-thomas.hellstrom@linux.intel.com
Add a number of helpers for shrinking that access core TTM and
core MM functionality in a way that make them unsuitable for
driver open-coding.
v11:
- New patch (split off from previous) and additional helpers.
v13:
- Adapt to ttm_backup interface change.
- Take resource off LRU when backed up.
Export a LRU walker for driver shrinker use. The walker
initially supports only trylocking, since that's the
method used by shrinkes. The walker makes use of
scoped_guard() to allow exiting from the LRU walk loop
without performing any explicit unlocking or
cleanup.
v8:
- Split out from another patch.
- Use a struct for bool arguments to increase readability (Matt Brost).
- Unmap user-space cpu-mappings before shrinking pages.
- Explain non-fatal error codes (Matt Brost)
v10:
- Instead of using the existing helper, Wrap the interface inside out and
provide a loop to de-midlayer things the LRU iteration (Christian König).
- Removing the R-B by Matt Brost since the patch was significantly changed.
v11:
- Split the patch up to include just the LRU walk helper.
Use fault-injection to test partial TTM swapout and interrupted swapin.
Return -EINTR for swapin to test the callers ability to handle and
restart the swapin, and on swapout perform a partial swapout to test that
the swapin and release_shrunken functionality.
v8:
- Use the core fault-injection system.
v9:
- Fix compliation failure for !CONFIG_FAULT_INJECTION
Cc: Christian König <christian.koenig@amd.com> Cc: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com> Cc: Matthew Brost <matthew.brost@intel.com> Cc: <dri-devel@lists.freedesktop.org> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> Reviewed-by: Matthew Brost <matthew.brost@intel.com> Reviewed-by: Christian König <christian.koenig@amd.com> Link: https://lore.kernel.org/intel-xe/20250305092220.123405-4-thomas.hellstrom@linux.intel.com
drm/ttm/pool, drm/ttm/tt: Provide a helper to shrink pages
Provide a helper to shrink ttm_tt page-vectors on a per-page
basis. A ttm_backup backend could then in theory get away with
allocating a single temporary page for each struct ttm_tt.
This is accomplished by splitting larger pages before trying to
back them up.
In the future we could allow ttm_backup to handle backing up
large pages as well, but currently there's no benefit in
doing that, since the shmem backup backend would have to
split those anyway to avoid allocating too much temporary
memory, and if the backend instead inserts pages into the
swap-cache, those are split on reclaim by the core.
Due to potential backup- and recover errors, allow partially swapped
out struct ttm_tt's, although mark them as swapped out stopping them
from being swapped out a second time. More details in the ttm_pool.c
DOC section.
v2:
- A couple of cleanups and error fixes in ttm_pool_back_up_tt.
- s/back_up/backup/
- Add a writeback parameter to the exported interface.
v8:
- Use a struct for flags for readability (Matt Brost)
- Address misc other review comments (Matt Brost)
v9:
- Update the kerneldoc for the ttm_tt::backup field.
v10:
- Rebase.
v13:
- Rebase on ttm_backup interface change. Update kerneldoc.
- Rebase and adjust ttm_tt_is_swapped().
v15:
- Rebase on ttm_backup return value change.
- Rebase on previous restructuring of ttm_pool_alloc()
- Rework the ttm_pool backup interface (Christian König)
- Remove cond_resched() (Christian König)
- Get rid of the need to allocate an intermediate page array
when restoring a multi-order page (Christian König)
- Update documentation.
Cc: Christian König <christian.koenig@amd.com> Cc: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com> Cc: Matthew Brost <matthew.brost@intel.com> Cc: <dri-devel@lists.freedesktop.org> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> Reviewed-by: Matthew Brost <matthew.brost@intel.com> Acked-by: Christian Koenig <christian.koenig@amd.com> Link: https://lore.kernel.org/intel-xe/20250305092220.123405-3-thomas.hellstrom@linux.intel.com
Provide a standalone shmem backup implementation.
Given the ttm_backup interface, this could
later on be extended to providing other backup
implementation than shmem, with one use-case being
GPU swapout to a user-provided fd.
v5:
- Fix a UAF. (kernel test robot, Dan Carptenter)
v6:
- Rename ttm_backup_shmem_copy_page() function argument
(Matthew Brost)
- Add some missing documentation
v8:
- Use folio_file_page to get to the page we want to writeback
instead of using the first page of the folio.
v13:
- Remove the base class abstraction (Christian König)
- Include ttm_backup_bytes_avail().
v14:
- Fix kerneldoc for ttm_backup_bytes_avail() (0-day)
- Work around casting of __randomize_layout struct pointer (0-day)
v15:
- Return negative error code from ttm_backup_backup_page()
(Christian König)
- Doc fixes. (Christian König).
Cc: Christian König <christian.koenig@amd.com> Cc: Somalapuram Amaranath <Amaranath.Somalapuram@amd.com> Cc: Matthew Brost <matthew.brost@intel.com> Cc: <dri-devel@lists.freedesktop.org> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> Reviewed-by: Matthew Brost <matthew.brost@intel.com> Reviewed-by: Christian König <christian.koenig@amd.com> Link: https://lore.kernel.org/intel-xe/20250305092220.123405-2-thomas.hellstrom@linux.intel.com
Steven Price [Thu, 13 Feb 2025 16:12:48 +0000 (16:12 +0000)]
drm/panthor: Clean up FW version information display
Assigning a string to an array which is too small to include the NUL
byte at the end causes a warning on some compilers. But this function
also has some other oddities like the 'header' array which is only ever
used within sizeof().
Tidy up the function by removing the 'header' array, allow the NUL byte
to be present in git_sha_header, and calculate the length directly from
git_sha_header.
Reported-by: Will Deacon <will@kernel.org> Closes: https://lore.kernel.org/all/20250213154237.GA11897@willie-the-truck/ Fixes: 9d443deb0441 ("drm/panthor: Display FW version information") Signed-off-by: Steven Price <steven.price@arm.com> Acked-by: Will Deacon <will@kernel.org> Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com> Link: https://patchwork.freedesktop.org/patch/msgid/20250213161248.1642392-1-steven.price@arm.com
Adrián Larumbe [Mon, 3 Mar 2025 19:08:46 +0000 (19:08 +0000)]
drm/panthor: Avoid sleep locking in the internal BO size path
Commit 434e5ca5b5d7 ("drm/panthor: Expose size of driver internal BO's over
fdinfo") locks the VMS xarray, to avoid UAF errors when the same VM is
being concurrently destroyed by another thread. However, that puts the
current thread in atomic context, which means taking the VMS' heap locks
will trigger a warning as the thread is no longer allowed to sleep.
Because in this case replacing the heap mutex with a spinlock isn't
feasible, the fdinfo handler no longer traverses the list of heaps for
every single VM associated with an open DRM file. Instead, when a new heap
chunk is allocated, its size is accumulated into a pool-wide tally, which
also makes the atomic context code path somewhat faster.
Signed-off-by: Adrián Larumbe <adrian.larumbe@collabora.com> Fixes: 434e5ca5b5d7 ("drm/panthor: Expose size of driver internal BO's over fdinfo") 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://patchwork.freedesktop.org/patch/msgid/20250303190923.1639985-2-adrian.larumbe@collabora.com
Adrián Larumbe [Mon, 3 Mar 2025 19:08:45 +0000 (19:08 +0000)]
drm/panthor: Replace sleep locks with spinlocks in fdinfo path
Commit 0590c94c3596 ("drm/panthor: Fix race condition when gathering fdinfo
group samples") introduced an xarray lock to deal with potential
use-after-free errors when accessing groups fdinfo figures. However, this
toggles the kernel's atomic context status, so the next nested mutex lock
will raise a warning when the kernel is compiled with mutex debug options:
CONFIG_DEBUG_RT_MUTEXES=y
CONFIG_DEBUG_MUTEXES=y
Replace Panthor's group fdinfo data mutex with a guarded spinlock.
Signed-off-by: Adrián Larumbe <adrian.larumbe@collabora.com> Fixes: 0590c94c3596 ("drm/panthor: Fix race condition when gathering fdinfo group samples") Reviewed-by: Liviu Dudau <liviu.dudau@arm.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://patchwork.freedesktop.org/patch/msgid/20250303190923.1639985-1-adrian.larumbe@collabora.com
Jiapeng Chong [Wed, 5 Mar 2025 02:05:46 +0000 (10:05 +0800)]
drm: adp: Remove unnecessary print function dev_err()
The print function dev_err() is redundant because platform_get_irq_byname()
already prints an error.
./drivers/gpu/drm/adp/adp_drv.c:470:2-9: line 470 is redundant because platform_get_irq() already prints an error.
./drivers/gpu/drm/adp/adp_drv.c:476:2-9: line 476 is redundant because platform_get_irq() already prints an error.
drm_sched_job_cleanup()'s documentation claims that calling
drm_sched_job_arm() is a "point of no return", implying that afterwards
a job cannot be cancelled anymore.
This is not correct, as proven by the function's code itself, which
takes a previous call to drm_sched_job_arm() into account. In truth, the
decisive factors are whether fences have been shared (e.g., with other
processes) and if the job has been submitted to an entity already.
Heiko Stuebner [Tue, 4 Mar 2025 12:44:18 +0000 (13:44 +0100)]
drm/rockchip: lvds: lower log severity for missing pinctrl settings
While missing lvds pinctrl is unexpected and is reported, we nevertheless
don't fail setting up the device and instead continue without explicit
pinctrl handling. So lower the log-level from error to warning to reflect
that.
Heiko Stuebner [Tue, 4 Mar 2025 12:44:17 +0000 (13:44 +0100)]
drm/rockchip: lvds: Hide scary error messages on probe deferral
Commit 52d11c863ac9 ("drm/rockchip: lvds: do not print scary message when
probing defer") already started hiding scary messages that are not relevant
if the requested supply just returned EPROBE_DEFER, but there are more
possible sources - like the phy.
So modernize the whole logging in the probe path by replacing the
remaining deprecated DRM_DEV_ERROR with appropriate dev_err(_probe)
and drm_err calls.
The distinction here is that all messages talking about mishaps of the
lvds element use dev_err(_probe) while messages caused by interaction
with the main Rockchip drm-device use drm_err.
Heiko Stuebner [Tue, 4 Mar 2025 12:44:16 +0000 (13:44 +0100)]
drm/rockchip: lvds: move pclk preparation in with clk_get
The LVDS block needs a separate pclk only on some socs, so currently
requests and prepares it in the soc-specific probe function, but common
code is required to unprepare it in the error path or on driver remove.
While this works because clk_unprepare just does nothing if clk is NULL,
this mismatch of who is responsible still is not very nice.
The clock-framework already has a helper for clk-get-and-prepare even
with devres support in devm_clk_get_prepared().
This will get and prepare the clock and also unprepare it on driver
removal, saving the driver from having to handle it "manually".
Jani Nikula [Thu, 23 Jan 2025 15:09:12 +0000 (17:09 +0200)]
drm/print: require struct drm_device for drm_err() and friends
The expectation is that the struct drm_device based logging helpers get
passed an actual struct drm_device pointer rather than some random
struct pointer where you can dereference the ->dev member.
Add a static inline helper to convert struct drm_device to struct
device, with the main benefit being the type checking of the macro
argument.
As a side effect, this also reduces macro argument double references.
Jani Nikula [Thu, 23 Jan 2025 15:09:10 +0000 (17:09 +0200)]
drm/sched: stop passing non struct drm_device to drm_err() and friends
The expectation is that the struct drm_device based logging helpers get
passed an actual struct drm_device pointer rather than some random
struct pointer where you can dereference the ->dev member.
Convert drm_err(sched, ...) to dev_err(sched->dev, ...) and
similar. This matches current usage, as struct drm_device is not
available, but drops "[drm]" or "[drm] *ERROR*" prefix from logging.
Unfortunately, there's no dev_WARN_ON(), so the conversion is not
exactly the same.
Jani Nikula [Thu, 23 Jan 2025 15:09:09 +0000 (17:09 +0200)]
drm/rockchip: stop passing non struct drm_device to drm_err() and friends
The expectation is that the struct drm_device based logging helpers get
passed an actual struct drm_device pointer rather than some random
struct pointer where you can dereference the ->dev member.
Convert drm_err(hdmi, ...) to dev_err(hdmi->dev, ...). This matches
current usage, but drops "[drm] *ERROR*" prefix from logging.
Jani Nikula [Thu, 23 Jan 2025 15:09:08 +0000 (17:09 +0200)]
drm/mipi-dsi: stop passing non struct drm_device to drm_err() and friends
The expectation is that the struct drm_device based logging helpers get
passed an actual struct drm_device pointer rather than some random
struct pointer where you can dereference the ->dev member.
Convert drm_err(host, ...) to dev_err(host->dev, ...). This matches
current usage, as struct drm_device is not available, but drops "[drm]
*ERROR*" from logs.
drm/appletbdrm: Fix format specifier for size_t variables
When building for a 32-bit platform, there are some warnings (or errors
with CONFIG_WERROR=y) due to an incorrect specifier for 'size_t'
variables, which is typedef'd as 'unsigned int' for these architectures:
drivers/gpu/drm/tiny/appletbdrm.c:171:17: error: format specifies type 'unsigned long' but the argument has type 'size_t' (aka 'unsigned int') [-Werror,-Wformat]
170 | drm_err(drm, "Actual size (%d) doesn't match expected size (%lu)\n",
| ~~~
| %zu
171 | actual_size, size);
| ^~~~
...
drivers/gpu/drm/tiny/appletbdrm.c:212:17: error: format specifies type 'unsigned long' but the argument has type 'size_t' (aka 'unsigned int') [-Werror,-Wformat]
211 | drm_err(drm, "Actual size (%d) doesn't match expected size (%lu)\n",
| ~~~
| %zu
212 | actual_size, size);
| ^~~~
Heiko Stuebner [Mon, 3 Mar 2025 18:22:56 +0000 (19:22 +0100)]
drm/rockchip: vop2: add missing bitfield.h include
Commit 328e6885996c ("drm/rockchip: vop2: Add platform specific callback")
moved per soc configuration code to the other per-soc data into
rockchip_vop2_reg.c, but forgot to also include bitfield.h for the used
FIELD_PREP macro. Add this missing include.
Fixes: 328e6885996c ("drm/rockchip: vop2: Add platform specific callback") Reported-by: kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202503040135.fgoyWdLB-lkp@intel.com/ Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Heiko Stuebner <heiko@sntech.de> Link: https://patchwork.freedesktop.org/patch/msgid/20250303182256.1727178-1-heiko@sntech.de
Andy Yan [Mon, 3 Mar 2025 03:44:21 +0000 (11:44 +0800)]
drm/rockchip: vop2: Add support for rk3576
VOP2 on rk3576:
Three video ports:
VP0 Max 4096x2160
VP1 Max 2560x1600
VP2 Max 1920x1080
2 4K Cluster windows with AFBC/RFBC, line RGB and YUV
4 Esmart windows with line RGB/YUV support:
Esmart0/1: 4K
Esmart2/3: 2k, or worked together as a single 4K plane at shared
line buffer mode.
Compared to the previous VOP, another difference is that each VP
has its own independent vsync interrupt number.
Andy Yan [Mon, 3 Mar 2025 03:44:19 +0000 (11:44 +0800)]
dt-bindings: display: vop2: Add missing rockchip,grf property for rk3566/8
The clock polarity of RGB signal output is controlled by GRF, this
property is already being used in the current device tree, but
forgot to describe it as a required property in the binding file.
Andy Yan [Mon, 3 Mar 2025 03:44:16 +0000 (11:44 +0800)]
drm/rockchip: vop2: Set plane possible crtcs by possible vp mask
In the upcoming VOP of rk3576, a window cannot attach to all Video
Ports, we introduce a possible_vp_mask for every window to indicate
which Video Ports this window can attach to.
Andy Yan [Mon, 3 Mar 2025 03:44:15 +0000 (11:44 +0800)]
drm/rockchip: vop2: Register the primary plane and overlay plane separately
In the upcoming VOP of rk3576, a Window cannot attach to all Video Ports,
so make sure all VP find it's suitable primary plane, then register the
remain windows as overlay plane will make code easier.
Kerem Karabay [Wed, 26 Feb 2025 16:04:36 +0000 (16:04 +0000)]
drm/tiny: add driver for Apple Touch Bars in x86 Macs
The Touch Bars found on x86 Macs support two USB configurations: one
where the device presents itself as a HID keyboard and can display
predefined sets of keys, and one where the operating system has full
control over what is displayed.
This commit adds support for the display functionality of the second
configuration. Functionality for the first configuration has been
merged in the HID tree.
Note that this driver has only been tested on T2 Macs, and only includes
the USB device ID for these devices. Testing on T1 Macs would be
appreciated.
Credit goes to Ben (Bingxing) Wang on GitHub for reverse engineering
most of the protocol.
Also, as requested by Andy, I would like to clarify the use of __packed
structs in this driver:
- All the packed structs are aligned except for appletbdrm_msg_information.
- We have to pack appletbdrm_msg_information since it is requirement of
the protocol.
- We compared binaries compiled by keeping the rest structs __packed and
not __packed using bloat-o-meter, and __packed was not affecting code
generation.
- To maintain consistency, rest structs have been kept __packed.
I would also like to point out that since the driver was reverse-engineered
the actual data types of the protocol might be different, including, but
not limited to, endianness.
Maxime Ripard [Thu, 20 Feb 2025 13:25:37 +0000 (14:25 +0100)]
drm/tests: Drop drm_kunit_helper_acquire_ctx_alloc()
lockdep complains when a lock is released in a separate thread the
lock is taken in, and it turns out that kunit does run its actions in a
separate thread than the test ran in.
This means that drm_kunit_helper_acquire_ctx_alloc() just cannot work as
it's supposed to, so let's just get rid of it.
Andy Yan [Tue, 18 Feb 2025 11:27:33 +0000 (19:27 +0800)]
drm/rockchip: vop2: Introduce vop hardware version
There is a version number hardcoded in the VOP VERSION_INFO
register, and the version number increments sequentially based
on the production order of the SoC.
So using this version number to distinguish different VOP features
will simplify the code.
Andy Yan [Tue, 18 Feb 2025 11:27:32 +0000 (19:27 +0800)]
drm/rockchip: vop2: Support for different layer select configuration between VPs
In the upcoming VOP for rk3576, every VP has it's own LAYER_SEL
register, and the configuration value of each VP for the same
window maybe different, so extend the layer_sel_id to array,
let it can descption the layer select configuration value for
different VP.
Andy Yan [Tue, 18 Feb 2025 11:27:30 +0000 (19:27 +0800)]
drm/rockchip: vop2: Add platform specific callback
The VOP interface mux, overlay, background delay cycle configuration
of different SOC are much different. Add platform specific callback
ops to let the core driver look cleaner and more refined.
Andy Yan [Tue, 18 Feb 2025 11:27:28 +0000 (19:27 +0800)]
drm/rockchip: vop2: use devm_regmap_field_alloc for cluster-regs
Right now vop2_cluster_init() copies the base vop2_cluster_regs
and adapts the reg value with the current window's offset before
adding the fields to the regmap.
This conflicts with the notion of reg_fields being const, see
https://lore.kernel.org/all/20240706-regmap-const-structs-v1-1-d08c776da787@weissschuh.net/
for reference, which now causes checkpatch to actually warn about that.
So instead of creating one big copy and changing it afterwards,
add the reg_fields individually using devm_regmap_field_alloc().
Functional it is the same, just that the reg_field we're handling
can stay const.
ipu_cpmem_set_yuv_interleaved() was added in 2012 by
commit 0125f21b2baf ("staging: drm/imx: Add
ipu_cpmem_set_yuv_interleaved()")
but has remained unused.
ipu_cpmem_get_burstsize() was added in 2016 by
commit 03085911d7bb ("gpu: ipu-cpmem: Add ipu_cpmem_get_burstsize()")
but has remained unused.
ipu_csi_get_window(), ipu_csi_is_interlaced() and
ipu_csi_set_test_generator() were added in 2014 by
commit 2ffd48f2e7ae ("gpu: ipu-v3: Add Camera Sensor Interface unit")
but have remained unused.
ipu_image_convert_enum_format() and ipu_image_convert_sync() were both
added in 2016 by
commit cd98e85a6b78 ("gpu: ipu-v3: Add queued image conversion support")
but have remained unused.
Remove them.
ipu_image_convert_sync() was the last user of
image_convert_sync_complete().
-Wflex-array-member-not-at-end was introduced in GCC-14, and we are
getting ready to enable it, globally.
So, in order to avoid ending up with flexible-array members in the
middle of other structs, we use the `struct_group_tagged()` helper
to separate the flexible arrays from the rest of the members in the
flexible structures. We then use the newly created tagged `struct
nvif_ioctl_v0_hdr` and `struct nvif_ioctl_mthd_v0_hdr` to replace the
type of the objects causing trouble in multiple structures.
We also want to ensure that when new members need to be added to the
flexible structures, they are always included within the newly created
tagged structs. For this, we use `static_assert()`. This ensures that the
memory layout for both the flexible structure and the new tagged struct
is the same after any changes.
So, with these changes, fix the following warnings:
drivers/gpu/drm/nouveau/nvif/object.c:60:38: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/gpu/drm/nouveau/nvif/object.c:233:38: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/gpu/drm/nouveau/nvif/object.c:214:38: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/gpu/drm/nouveau/nvif/object.c:152:38: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/gpu/drm/nouveau/nvif/object.c:138:38: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/gpu/drm/nouveau/nvif/object.c:104:38: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/gpu/drm/nouveau/nouveau_svm.c:83:35: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/gpu/drm/nouveau/nouveau_svm.c:82:30: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
Lizhi Hou [Wed, 26 Feb 2025 16:18:10 +0000 (08:18 -0800)]
accel/amdxdna: Check interrupt register before mailbox_rx_worker exits
There is a timeout failure been found during stress tests. If the firmware
generates a mailbox response right after driver clears the mailbox channel
interrupt register, the hardware will not generate an interrupt for the
response. This causes the unexpected mailbox command timeout.
To handle this failure, driver checks the interrupt register before
exiting mailbox_rx_worker(). If there is a new response, driver goes back
to process it.
Heiko Stuebner [Thu, 20 Feb 2025 23:41:41 +0000 (00:41 +0100)]
drm/rockchip: Fix shutdown when no drm-device is set up
When the drm-driver probes, it mainly creates the component device, where
all the sub-drivers (vops, hdmi, etc) hook into.
This will cause the shutdown handler to get called on shutdown, even
though the drm-device might not have been set up, or the component bind
might have failed.
So use the new component helper to check whether the drm-device is up
and only then call the drm-atomic helper to release all the drm magic.
This prevents failures when the drm-device is never set, or has been
freed up already for example by a probe-defer during the component bind.
Heiko Stuebner [Thu, 20 Feb 2025 23:41:40 +0000 (00:41 +0100)]
drivers: base: component: add function to query the bound status
The component helpers already expose the bound status in debugfs, but at
times it might be necessary to also check that state in the kernel and
act differently depending on the result.
For example the shutdown handler of a drm-driver might need to stop
a whole output pipeline if the drm device is up and running, but may
run into problems if that drm-device has never been set up before,
for example because the binding deferred.
So add a little helper that returns the bound status for a componet
device.
drm/rockchip: vop2: Consistently use dev_err_probe()
Replace drm_err() calls in vop2_bind() and vop2_create_crtcs() with
dev_err_probe(), to simplify error handling and improve consistency.
Additionally, ensure the already existing dev_err_probe() invocations
pass drm->dev instead of dev as their first argument, so that we get the
actual reason in case of -EPROBE_DEFER errors:
drm/rockchip: vop2: Improve display modes handling on RK3588 HDMI1
The RK3588 specific implementation is currently quite limited in terms
of handling the full range of display modes supported by the connected
screens, e.g. 2560x1440@75Hz, 2048x1152@60Hz, 1024x768@60Hz are just a
few of them.
Additionally, it doesn't cope well with non-integer refresh rates like
59.94, 29.97, 23.98, etc.
Make use of HDMI1 PHY PLL as a more accurate DCLK source to handle
all display modes up to 4K@60Hz.
Damon Ding [Mon, 24 Feb 2025 08:13:12 +0000 (16:13 +0800)]
drm/rockchip: analogix_dp: Expand device data to support multiple edp display
There are two main modifications: one is expanding struct
rockchip_dp_chip_data to an array, and the other is adding
&rockchip_dp_chip_data.reg to separate different edp devices.
The Rockchip W552793DBA-V10 display/touchscreen board contains a
Wanchanglong W552793BAA panel, which in turn is using a Raydium
RM67200 MIPI-DSI controller. Add a DSI panel driver for it.
The W552793BAA panel init sequence has been taken from the RK3588
EVB1 vendor kernel devicetree.
The Rockchip W552793DBA-V10 display/touchscreen board contains a
Wanchanglong W552793BAA panel, which in turn is using a Raydium
RM67200 MIPI-DSI controller. Add a DT binding for the DSI panel.
Anusha Srivatsa [Tue, 25 Feb 2025 22:20:45 +0000 (17:20 -0500)]
drm/mediatek: move to devm_platform_ioremap_resource() usage
Replace platform_get_resource + devm_ioremap_resource
with just devm_platform_ioremap_resource()
Used Coccinelle to do this change. SmPl patch:
@rule_1@
identifier res;
expression ioremap_res;
identifier pdev;
@@
-struct resource *res;
...
-res = platform_get_resource(pdev,...);
-ioremap_res = devm_ioremap_resource(...);
+ioremap_res = devm_platform_ioremap_resource(pdev,0);
Cc: Chun-Kuang Hu <chunkuang.hu@kernel.org> Cc: Philipp Zabel <p.zabel@pengutronix.de> Reviewed-by: Maxime Ripard <mripard@kernel.org> Reviewed-by: CK Hu <ck.hu@mediatek.com> Signed-off-by: Anusha Srivatsa <asrivats@redhat.com> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Link: https://lore.kernel.org/r/20250225-memory-drm-misc-next-v1-4-9d0e8761107a@redhat.com Signed-off-by: Maxime Ripard <mripard@kernel.org>
Abel Vesa [Mon, 3 Feb 2025 10:57:59 +0000 (12:57 +0200)]
drm/msm/dp: Add support for LTTPR handling
Link Training Tunable PHY Repeaters (LTTPRs) are defined in DisplayPort
1.4a specification. As the name suggests, these PHY repeaters are
capable of adjusting their output for link training purposes.
According to the DisplayPort standard, LTTPRs have two operating
modes:
- non-transparent - it replies to DPCD LTTPR field specific AUX
requests, while passes through all other AUX requests
- transparent - it passes through all AUX requests.
Switching between these two modes is done by the DPTX by issuing
an AUX write to the DPCD PHY_REPEATER_MODE register.
The msm DP driver is currently lacking any handling of LTTPRs.
This means that if at least one LTTPR is found between DPTX and DPRX,
the link training would fail if that LTTPR was not already configured
in transparent mode.
The section 3.6.6.1 from the DisplayPort v2.0 specification mandates
that before link training with the LTTPR is started, the DPTX may place
the LTTPR in non-transparent mode by first switching to transparent mode
and then to non-transparent mode. This operation seems to be needed only
on first link training and doesn't need to be done again until device is
unplugged.
It has been observed on a few X Elite-based platforms which have
such LTTPRs in their board design that the DPTX needs to follow the
procedure described above in order for the link training to be successful.
So add support for reading the LTTPR DPCD caps to figure out the number
of such LTTPRs first. Then, for platforms (or Type-C dongles) that have
at least one such an LTTPR, set its operation mode to transparent mode
first and then to non-transparent, just like the mentioned section of
the specification mandates.
Tested-by: Johan Hovold <johan+linaro@kernel.org> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Reviewed-by: Johan Hovold <johan+linaro@kernel.org> Signed-off-by: Abel Vesa <abel.vesa@linaro.org> Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com> Acked-by: Jani Nikula <jani.nikula@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20250203-drm-dp-msm-add-lttpr-transparent-mode-set-v5-4-c865d0e56d6e@linaro.org Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Abel Vesa [Mon, 3 Feb 2025 10:57:58 +0000 (12:57 +0200)]
drm/i915/dp: Use the generic helper to control LTTPR transparent mode
LTTPRs operating modes are defined by the DisplayPort standard and the
generic framework now provides a helper to switch between them, which
is handling the explicit disabling of non-transparent mode and its
disable->enable sequence mentioned in the DP Standard v2.0 section
3.6.6.1.
So use the new drm generic helper instead as it makes the code a bit
cleaner. Since the driver specific implementation holds the
lttrp_common_caps, if the call to the drm generic helper fails, the
lttrp_common_caps need to be updated as the helper has already rolled
back to transparent mode.
Abel Vesa [Mon, 3 Feb 2025 10:57:57 +0000 (12:57 +0200)]
drm/nouveau/dp: Use the generic helper to control LTTPR transparent mode
LTTPRs operating modes are defined by the DisplayPort standard and the
generic framework now provides a helper to switch between them, which
is handling the explicit disabling of non-transparent mode and its
disable->enable sequence mentioned in the DP Standard v2.0 section
3.6.6.1.
So use the new drm generic helper instead as it makes the code a bit
cleaner.
Abel Vesa [Mon, 3 Feb 2025 10:57:56 +0000 (12:57 +0200)]
drm/dp: Add helper to set LTTPRs in transparent mode
According to the DisplayPort standard, LTTPRs have two operating
modes:
- non-transparent - it replies to DPCD LTTPR field specific AUX
requests, while passes through all other AUX requests
- transparent - it passes through all AUX requests.
Switching between this two modes is done by the DPTX by issuing
an AUX write to the DPCD PHY_REPEATER_MODE register.
Add a generic helper that allows switching between these modes.
Also add a generic wrapper for the helper that handles the explicit
disabling of non-transparent mode and its disable->enable sequence
mentioned in the DP Standard v2.0 section 3.6.6.1. Do this in order
to move this handling out of the vendor specific driver implementation
into the generic framework.
Tested-by: Johan Hovold <johan+linaro@kernel.org> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Reviewed-by: Johan Hovold <johan+linaro@kernel.org> Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com> Signed-off-by: Abel Vesa <abel.vesa@linaro.org> Acked-by: Jani Nikula <jani.nikula@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20250203-drm-dp-msm-add-lttpr-transparent-mode-set-v5-1-c865d0e56d6e@linaro.org Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Tejas Vipin [Thu, 20 Feb 2025 04:57:21 +0000 (10:27 +0530)]
drm/mipi-dsi: extend "multi" functions and use them in sony-td4353-jdi
Removes mipi_dsi_dcs_set_tear_off and replaces it with a
multi version as after replacing it in sony-td4353-jdi, it doesn't
appear anywhere else. sony-td4353-jdi is converted to use multi style
functions, including mipi_dsi_dcs_set_tear_off_multi.
drm/repaper: fix integer overflows in repeat functions
There are conditions, albeit somewhat unlikely, under which right hand
expressions, calculating the end of time period in functions like
repaper_frame_fixed_repeat(), may overflow.
For instance, if 'factor10x' in repaper_get_temperature() is high
enough (170), as is 'epd->stage_time' in repaper_probe(), then the
resulting value of 'end' will not fit in unsigned int expression.
Mitigate this by casting 'epd->factored_stage_time' to wider type before
any multiplication is done.
Found by Linux Verification Center (linuxtesting.org) with static
analysis tool SVACE.
Uwe Kleine-König [Mon, 17 Feb 2025 17:49:36 +0000 (18:49 +0100)]
drm/bridge: ti-sn65dsi86: Check for CONFIG_PWM using IS_REACHABLE()
Currently CONFIG_PWM is a bool but I intend to change it to tristate. If
CONFIG_PWM=m in the configuration, the cpp symbol CONFIG_PWM isn't
defined and so the PWM code paths in the ti-sn65dsi86 driver are not
used.
The correct way to check for CONFIG_PWM is using IS_REACHABLE which does
the right thing for all cases
CONFIG_DRM_TI_SN65DSI86 ∈ { y, m } x CONFIG_PWM ∈ { y, m, n }.
There is no change until CONFIG_PWM actually becomes tristate.
Tvrtko Ursulin [Fri, 21 Feb 2025 10:50:37 +0000 (10:50 +0000)]
drm/sched: Move internal prototypes to internal header
Now that we have a header file for internal scheduler interfaces we can
move some more prototypes into it. By doing that we eliminate the chance
of drivers trying to use something which was not intended to be used.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com> Cc: Christian König <christian.koenig@amd.com> Cc: Danilo Krummrich <dakr@kernel.org> Cc: Matthew Brost <matthew.brost@intel.com> Cc: Philipp Stanner <phasta@kernel.org> Signed-off-by: Philipp Stanner <phasta@kernel.org> Link: https://patchwork.freedesktop.org/patch/msgid/20250221105038.79665-6-tvrtko.ursulin@igalia.com
Tvrtko Ursulin [Fri, 21 Feb 2025 10:50:35 +0000 (10:50 +0000)]
drm/sched: Remove a hole from struct drm_sched_job
We can re-order some struct members and take u32 credits outside of the
pointer sandwich and also for the last_dependency member we can get away
with an unsigned int since for dependency we use xa_limit_32b.
Pahole report before:
/* size: 160, cachelines: 3, members: 14 */
/* sum members: 156, holes: 1, sum holes: 4 */
/* last cacheline: 32 bytes */
And after:
/* size: 152, cachelines: 3, members: 14 */
/* last cacheline: 24 bytes */
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com> Cc: Christian König <christian.koenig@amd.com> Cc: Danilo Krummrich <dakr@kernel.org> Cc: Matthew Brost <matthew.brost@intel.com> Cc: Philipp Stanner <phasta@kernel.org> Acked-by: Danilo Krummrich <dakr@kernel.org> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Philipp Stanner <phasta@kernel.org> Link: https://patchwork.freedesktop.org/patch/msgid/20250221105038.79665-4-tvrtko.ursulin@igalia.com
Tvrtko Ursulin [Fri, 21 Feb 2025 10:50:34 +0000 (10:50 +0000)]
drm/amdgpu: Pop jobs from the queue more robustly
Replace a copy of DRM scheduler's to_drm_sched_job with a copy of a newly
added drm_sched_entity_queue_pop.
This allows breaking the hidden dependency that queue_node has to be the
first element in struct drm_sched_job.
A comment is also added with a reference to the mailing list discussion
explaining the copied helper will be removed when the whole broken
amdgpu_job_stop_all_jobs_on_sched is removed.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com> Cc: Christian König <christian.koenig@amd.com> Cc: Danilo Krummrich <dakr@kernel.org> Cc: Matthew Brost <matthew.brost@intel.com> Cc: Philipp Stanner <phasta@kernel.org> Cc: Zhang, Hawking <Hawking.Zhang@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Philipp Stanner <phasta@kernel.org> Link: https://patchwork.freedesktop.org/patch/msgid/20250221105038.79665-3-tvrtko.ursulin@igalia.com
Tvrtko Ursulin [Fri, 21 Feb 2025 10:50:33 +0000 (10:50 +0000)]
drm/sched: Add internal job peek/pop API
Idea is to add helpers for peeking and popping jobs from entities with
the goal of decoupling the hidden assumption in the code that queue_node
is the first element in struct drm_sched_job.
That assumption usually comes in the form of:
while ((job = to_drm_sched_job(spsc_queue_pop(&entity->job_queue))))
Which breaks if the queue_node is re-positioned due to_drm_sched_job
being implemented with a container_of.
This also allows us to remove duplicate definitions of to_drm_sched_job.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com> Cc: Christian König <christian.koenig@amd.com> Cc: Danilo Krummrich <dakr@kernel.org> Cc: Matthew Brost <matthew.brost@intel.com> Cc: Philipp Stanner <phasta@kernel.org> Signed-off-by: Philipp Stanner <phasta@kernel.org> Link: https://patchwork.freedesktop.org/patch/msgid/20250221105038.79665-2-tvrtko.ursulin@igalia.com
Linus Torvalds [Sun, 23 Feb 2025 01:32:00 +0000 (17:32 -0800)]
Merge tag 'v6.14-rc3-smb3-client-fix-part2' of git://git.samba.org/sfrench/cifs-2.6
Pull smb client fix from Steve French:
- Fix potential null pointer dereference
* tag 'v6.14-rc3-smb3-client-fix-part2' of git://git.samba.org/sfrench/cifs-2.6:
smb: client: Add check for next_buffer in receive_encrypted_standard()
projects / thirdparty / kernel / linux.git / log
