Follow up on the previous devm_drm_bridge_add() conversion by
converting the remaining manual cleanup in chipone-icn6211 to the
managed helper.
Suggested-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Osama Abdelkader <osama.abdelkader@gmail.com>
Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Link: https://patch.msgid.link/20260521212529.151493-1-osama.abdelkader@gmail.com
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
As it turns out, Turing isn't the only architecture that needs this. On
this Dell Precision 7780 with an AD103 GPU, along with pretty much every
other laptop I tested, runtime PM is still somewhat unreliable. At first
glance it seems as if it's fixed, but lowering the autosuspend delay to
500ms and then doing a stress test of suspend/resume cycles on the GPU ends
up causing everything to start timing out.
After quite a lot of digging, I eventually landed back on this magic
timeout in r535_gsp_fini(). As it turns out, increasing the timeout ends up
fixing the runtime PM issues as far as I can tell, even during intense
stress testing.
Unfortunately after spending quite a bit of time trying to dig through
OpenRM to figure out what this magic sleep is actually doing, I've also
come up short with any reasonable explanation. In lieu of that, I'm going
to include the observations I did make while trying to figure this out in
hopes someone eventually does figure this out:
* The magic sleep has to occur after fbsr is initialized. Performing it at
any time before that doesn't appear to work.
* In situations where runtime PM starts getting flaky, some rather
interesting visual effects end up happening on occasion before the GPU
fully falls over. In particular, squares that look like the result of an
incomplete blitting operation to a tiled buffer end up showing up on
applications like vkcube. Interestingly enough, they remain in precisely
the same place between runtime PM cycles until the GPU falls over - even
when restarting vkcube multiple times, and even when vkcube is actively
updating the screen. Even more interestingly, they're not limited to a
specific framebuffer - you can see the squares changing as the cube
rotates around.
We cannot however, say that this is likely to be a incomplete fbsr
operation. The magic sleep happens before fbsr is actually saved (which
happens on the GSP unload), so it's something else.
* During a short bit of testing with a desktop that I have, the magic sleep
seemed to make no difference to whether or not suspend/resume works. It
seems to generally work almost always. So we can assume this is likely
exclusive to runtime PM, not S3.
As well, here's a list of the things I tried before settling on the magic
sleep:
* Hooking up NV2080_CTRL_CMD_INTERNAL_GCX_ENTRY_PREREQUISITE and then
blocking runtime PM until OpenRM signals that GC6/GCOFF is ready appears
to make no difference.
* Hooking up some (maybe not all, unsure about that part) bits of comptag
saving including:
* Fetching static memsys information from GSP
* Adding the size of the comptag storage to the fbsr data
* Adding a GA103+ workaround for disabling raw compression mode during
fbsr (it doesn't seem like it applies for any systems I tried it on
anyhow)
* Setting bPreserveVideoMemoryAllocations=1 in GspSystemInfo
So, until we can figure this out properly - just sleep for longer.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Fixes: 53dac06238 ("drm/nouveau/gsp: add support for 570.144")
Cc: <stable@vger.kernel.org> # v6.16+
Reviewed-by: Dave Airlie <airlied@redhat.com>
Link: https://patch.msgid.link/20260814194542.781955-5-lyude@redhat.com
There's two flags that we've never been setting when asking GSP to suspend
the GPU, which OpenRM does set:
GPU_STATE_FLAGS_PRESERVING
GPU_STATE_FLAGS_PM_TRANSITION
These flags aren't -supposed- to do much in GSP, they're mostly used by
OpenRM itself for state tracking. The only thing they do from GSP's side is
control whether or not a single display related workaround is applied
during suspend.
But as it turns out, that single workaround is actually quite crucial for
getting runtime PM working with nouveau - and without it set we end up
seeing a lot more failures with runtime PM resume. So, let's start setting
it.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Fixes: 53dac06238 ("drm/nouveau/gsp: add support for 570.144")
Cc: <stable@vger.kernel.org> # v6.16+
Reviewed-by: Dave Airlie <airlied@redhat.com>
Link: https://patch.msgid.link/20260814194542.781955-4-lyude@redhat.com
Previously, it looked as if we were able to fix suspend/resume on some
desktops by setting Gcoff based on whether or not we were entering runtime
PM. This was a mistake though - the only time suspend/resume would end up
actually working was if Gcoff = 0.
It seems like it's likely the main reason for this is the FBSR GcOff
argument actually controls GSP's behavior with regards to which buffers it
decides to save across suspend/resume. When GcOff = 1, RM reserved regions
are saved unless they are marked as LOST_ON_SUSPEND, and RM channel-context
and kernel-client buffers are also saved -including- when they are
LOST_ON_SUSPEND. This means with GcOff = 1, we end up having GSP save and
restore buffers that actually need to be reinitialized on resume - causing
the failures we're setting.
Thanks to John Hubbard from Nvidia for providing some background on what
these options do in the GSP firmware do!
Signed-off-by: Lyude Paul <lyude@redhat.com>
Fixes: 53dac06238 ("drm/nouveau/gsp: add support for 570.144")
Cc: <stable@vger.kernel.org> # v6.16+
Reviewed-by: Dave Airlie <airlied@redhat.com>
Link: https://patch.msgid.link/20260814194542.781955-3-lyude@redhat.com
This reverts commit 8302d0afea.
It turns out this looked like the right fix on some systems, but it's not -
as this causes runtime PM to actually fail on many a laptop.
Fixes: 8302d0afea ("nouveau/gsp: fix suspend/resume regression on r570 firmware")
Cc: <stable@vger.kernel.org> # v6.19+
Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Link: https://patch.msgid.link/20260814194542.781955-2-lyude@redhat.com
Nouveau is one of the very few modern hardware drivers in the kernel that
doesn't have atomic modesetting enabled by default, in part because when it
was originally written by Ben there wasn't much in the way of good atomic
modesetting clients to actually test things out with.
Nowadays however, atomic modesetting is very much the norm - and support in
userspace for non-atomic drivers is starting to bitrot a bit - leading to
its own set of issues. At the same time, many of those issues are fixed by
just turning on atomic in nouveau. Plus, I've been running nouveau with
atomic modesetting on by default for most of the machines I work on, and
I've already fixed quite a number of issues to the point where things seem
quite stable (excluding a single screen flashing bug on my desktop, which I
am not particularly convinced has anything to do with atomic modesetting).
Now that we've protected against breaking things for chipsets where atomic
isn't supported (<nv50) - let's enable it by default on generations of
hardware that support it. We'll leave the module parameter around for the
time being, as turning it off may be helpful in the possibility that we hit
regressions.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Link: https://patch.msgid.link/20260813204803.2097176-6-lyude@redhat.com
Atomic modesetting support was never added for pre-nv50 chipsets, so make
sure we don't allow it to be forced on. Additionally, print a small warning
when it's not supported.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Link: https://patch.msgid.link/20260813204803.2097176-5-lyude@redhat.com
The way we handled the nouveau.atomic module parameter before was fairly
broken, and had a number of issues:
- It was only ever actually parsed in the case of PCI devices.
- When nouveau.atomic was enabled, it would add the cap for atomic
modesetting to the global driver_pci structure. This meant that if one
GPU on a system supported atomic and another didn't, it would still get
enabled for both.
Looking into this exposed further silliness in the way that we actually
handle the drm_driver struct. We have one global structure for platform
devices, and another for PCI devices - both of which are literally
identical.
So before we start preparing to enable atomic modesetting by default, let's
fix this. Instead of sharing driver_pci and driver_platform, we instead
create driver_legacy_kms and driver_atomic_kms, each of which is identical
except for the DRIVER_ATOMIC capabilities flag, and then assign either
depending on the nouveau_atomic module parameter.
Doing this is also preferable, as the next step for enabling atomic
modesetting by default will be ensuring that we don't enable it for legacy
devices that still don't support it. This requires only checking the atomic
modesetting module parameter after the NVKM device is ready, as this allows
us to check the GPU family that nouveau is running on.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Link: https://patch.msgid.link/20260813204803.2097176-4-lyude@redhat.com
Seems like we never remembered to start printing the value for this, so
let's start to aid in troubleshooting in case we run into any issues with
atomic.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Link: https://patch.msgid.link/20260813204803.2097176-3-lyude@redhat.com
Sashiko caught this while reviewing the patches for enabling atomic by
default - if we fail to allocate the DRM device pointer, we'll attempt to
free the error pointer that it returns rather than the actual struct.
Let's fix this while we're at it.
Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Link: https://patch.msgid.link/20260813204803.2097176-2-lyude@redhat.com
create_sgt() receives the BO size as a u64 but stores the number of
pages in an int. qaic_create_bo_ioctl() page-aligns the user-supplied
size without an upper bound, so DIV_ROUND_UP(size, PAGE_SIZE) can exceed
INT_MAX and truncate when assigned to nr_pages.
When the page count wraps to a small positive value (for example a BO
larger than 16 TiB on a 4 KiB page kernel), create_sgt() succeeds after
allocating only a handful of pages while the GEM object still records
the full size, so the request never fails with -ENOMEM. The temporary
pages array and the allocation loop both use the truncated count, so the
backing SG table ends up smaller than obj->size. Later slice operations
validate against obj->size and then clone ranges from this short sgt.
Compute the page count as a u64 and reject sizes that do not fit in
nr_pages before the narrowing conversion.
Fixes: ff13be8303 ("accel/qaic: Add datapath")
Assisted-by: Claude:claude-fable-5
Signed-off-by: Hao-Qun Huang <alvinhuang0603@gmail.com>
Reviewed-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com>
Signed-off-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com>
Link: https://patch.msgid.link/20260708031959.966009-1-alvinhuang0603@gmail.com
qaic_ssr.c uses mhi_device_id implicitly. Include <linux/device-id/mhi.h>.
Fixes: 9675093ace ("accel/qaic: Implement basic SSR handling")
Signed-off-by: Youssef Samir <youssef.abdulrahman@oss.qualcomm.com>
Reviewd-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com>
Signed-off-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com>
Link: https://patch.msgid.link/20260727173326.2219191-6-youssef.abdulrahman@oss.qualcomm.com
qaic_drv.c uses mhi_device_id implicitly. Include <linux/device-id/mhi.h>.
Fixes: c501ca23a6 ("accel/qaic: Add uapi and core driver file")
Signed-off-by: Youssef Samir <youssef.abdulrahman@oss.qualcomm.com>
Reviewd-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com>
Signed-off-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com>
Link: https://patch.msgid.link/20260727173326.2219191-3-youssef.abdulrahman@oss.qualcomm.com
qaic_timesync.c and sahara.c use struct mhi_device_id implicitly since
commit 995832b2ce ("Replace <linux/mod_devicetable.h> by more specific
<linux/device-id/*.h> (c files)") removed <linux/mod_devicetable.h>.
Explicitly include <linux/device-id/mhi.h> in qaic_timesync.c and sahara.c.
Fixes: 995832b2ce ("Replace <linux/mod_devicetable.h> by more specific <linux/device-id/*.h> (c files)")
Signed-off-by: Youssef Samir <youssef.abdulrahman@oss.qualcomm.com>
Reviewed-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com>
Signed-off-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com>
Link: https://patch.msgid.link/20260727173326.2219191-2-youssef.abdulrahman@oss.qualcomm.com
The mipi_dsi_dcs_*() functions used by this driver are deprecated
in favour of their _multi() counterparts, as noted in
Documentation/gpu/todo.rst. The _multi() variants record the
first error in a context structure and skip subsequent calls once
an error is set, removing the need to check the return value
after each command. They also log failures internally, making the
per-call dev_err() calls redundant.
Convert prepare(), enable(), disable(), and unprepare() to use
mipi_dsi_dcs_*_multi().
unprepare() previously returned an error if
mipi_dsi_dcs_enter_sleep_mode() failed, skipping RESET assertion
and regulator_disable(). Because drm_panel_unprepare() does not
clear panel->prepared when the callback returns an error,
drm_panel_prepare() would then return early on the next call,
leaving the panel powered and unable to be re-initialised. The
converted code always asserts RESET, disables the regulator, and
returns 0.
Also fix a typo in a comment ("Enabe" -> "Enable").
Signed-off-by: Akash Sukhavasi <akash.sukhavasi@gmail.com>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Signed-off-by: Linus Walleij <linusw@kernel.org>
Link: https://patch.msgid.link/20260810-mipi-dsi-s6d16d0-multi-v2-1-8fcdfbf9bbd2@gmail.com
Initialise the SCDC debugfs file for connectors that call the HDMI state
helper's debugfs_init helper. This makes all connectors that use the
helper expose the scdc_status file in their connector debugfs.
Acked-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Tested-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Link: https://patch.msgid.link/20260724-scdc-link-health-v9-5-bdda406d016d@collabora.com
Signed-off-by: Daniel Stone <daniels@collabora.com>
The HDMI debugfs files are only useful for connectors that already make
use of the HDMI state helpers. In addition, relying on HDMI-specific
functions in drm_debugfs.c can lead to circular linking problems later
down the line.
Move hdmi_debugfs_add() and friends to a new helper in
drm_hdmi_state_helper.c, and make all current users (vc4, sun4i, bridge)
either use it as a debugfs_init func, or call it directly in its
debugfs_init func.
Suggested-by: Maxime Ripard <mripard@kernel.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Acked-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Tested-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Link: https://patch.msgid.link/20260724-scdc-link-health-v9-4-bdda406d016d@collabora.com
Signed-off-by: Daniel Stone <daniels@collabora.com>
HDMI 2.1 redefines previously reserved fields in SCDC for various new
uses. No version check needs to be performed, as an HDMI 2.0 sink's
reserved SCDC fields are well-defined to be 0, and any zero-ness of
these fields for an HDMI 2.0 sink is not a surprise for SCDC parsers for
HDMI 2.1.
Implement reading and outputting these fields over debugfs.
Reviewed-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Tested-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Link: https://patch.msgid.link/20260724-scdc-link-health-v9-3-bdda406d016d@collabora.com
Signed-off-by: Daniel Stone <daniels@collabora.com>
SCDC provides status information on the current display link. At the
very least, it may be useful to expose this info through debugfs.
Add a debugfs entry for it under the connector, which displays a few
more details parsed out of the SCDC registers. A new
drm_scdc_debugfs_init function can be called by the connector
implementation to initialise the debugfs file.
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Acked-by: Maxime Ripard <mripard@kernel.org>
Tested-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Link: https://patch.msgid.link/20260724-scdc-link-health-v9-2-bdda406d016d@collabora.com
Signed-off-by: Daniel Stone <daniels@collabora.com>
drm_scdc_read and drm_scdc_write, both of which are only used within
drm_scdc_helper (although exported), use a ssize_t as their return type.
This would make sense if they returned the number of bytes read/written
on success, and negative errno otherwise. However, they return 0 on
success.
Demote them to "int" as their return type, in order to avoid needlessly
using 64 bits when less suffices.
No functional change.
Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Reviewed-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Tested-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Link: https://patch.msgid.link/20260724-scdc-link-health-v9-1-bdda406d016d@collabora.com
Signed-off-by: Daniel Stone <daniels@collabora.com>
drm_log_draw_new_line() calls drm_log_clear_line() and
drm_log_draw_line(), which independently call
drm_client_buffer_vmap_local(), drm_client_buffer_vunmap_local(),
and drm_client_buffer_flush(). For each call to
drm_log_draw_new_line(), this results in 2 vmap/vunmap pairs and
2 flush calls, even though vmap_local maps the entire framebuffer
each time.
Refactor drm_log_clear_line() and drm_log_draw_line() to accept a
pre-mapped iosys_map by value, removing the per-line vmap/vunmap/flush
calls. Move the single vmap/vunmap pair and flush up to
drm_log_draw_new_line(), which now maps once before clearing and
drawing, then issues a single flush after unmapping.
Signed-off-by: Shixiong Ou <oushixiong@kylinos.cn>
Link: https://patch.msgid.link/20260729084823.693723-1-oushixiong1025@163.com
Signed-off-by: Jocelyn Falempe <jfalempe@redhat.com>
Simplify error condition logic in lvds device tree property parsing. It
makes it easier to add extra properties without handling complex goto
entries.
Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Tested-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Wojciech Dubowik <Wojciech.Dubowik@mt.com>
Link: https://patch.msgid.link/20260728122032.692482-2-wojciech.dubowik@mt.com
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
With HDMI FreeSync detection now sourced from the common DRM EDID
parser, the DMUB parsing for AMD VSDB is no longer used.
Remove all relevant fields and files for this flow, it is no longer
necessary to ever take firmware round-trip for this.
Signed-off-by: Alex Huang <Alex.Huang2@amd.com>
Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org>
Link: https://patch.msgid.link/20260804143339.714548-5-Alex.Huang2@amd.com
Signed-off-by: Mario Limonciello <superm1@kernel.org>
Many duplicate codes can be extracted to common code, fill
amdgpu_dm_connector fields outside of the DP/eDP, HDMI branches,
pull the VSDB data once using get_amd_vsdb() and then derive the refresh
rate ranges.
When a monitor advertises support for FreeSync, the refresh rate from
the AMD VSDB is used, otherwise, falls back on the monitor_range, either
EDID refresh rate range field or DisplayID VRR data.
Signed-off-by: Alex Huang <Alex.Huang2@amd.com>
Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org>
Link: https://patch.msgid.link/20260804143339.714548-4-Alex.Huang2@amd.com
(Adjust for movement to amdgpu_dm_connector.c)
Signed-off-by: Mario Limonciello <superm1@kernel.org>
HDMI FreeSync on amdgpu is detected using the AMD VSDB parser in the
DMUB firmware, now that the common DRM EDID parser can do the same, use
the common parser.
Extend get_amd_vsdb() to copy the relevant flags and data now being
extracted from the AMD VSDB. Additionally, parse_hdmi_amd_vsdb() and the
parse_edid_cea() firmware helpers now have no callers; remove them.
Signed-off-by: Alex Huang <Alex.Huang2@amd.com>
Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org>
Link: https://patch.msgid.link/20260804143339.714548-3-Alex.Huang2@amd.com
(Adjust for movement to amdgpu_dm_connector.c)
Signed-off-by: Mario Limonciello <superm1@kernel.org>
Commit 118362a962 ("drm/edid: Parse AMD Vendor-Specific Data Block")
added parsing of the AMD VSDB, but only for version 3 and only for the
replay/panel-type/luminance fields. Monitors that only advertise
FreeSync through a v1 or v2 AMD VSDB therefore get no variable refresh
rate range from the common EDID parser.
Restructure drm_parse_amd_vsdb() and implement parsing for the FreeSync
refresh range fields provided in the AMD VSDB. Additionally, version > 3
will be parsed as if they were version 3 (with a warning of this
occurring) as future versions will likely be backwards compatible.
Signed-off-by: Alex Huang <Alex.Huang2@amd.com>
Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org>
Link: https://patch.msgid.link/20260804143339.714548-2-Alex.Huang2@amd.com
Signed-off-by: Mario Limonciello <superm1@kernel.org>
Since commit 41b4b11da0 ("drm: Add valid clones check"),
all encoders attached to the same CRTC must advertise
each other through possible_clones.
Komeda creates one writeback connector per CRTC. Allow the writeback
encoder to be cloned only with encoders that can drive the same CRTC,
and update those encoders reciprocally.
Signed-off-by: Asad Malik <asad.malik@arm.com>
Signed-off-by: Vincenzo Frascino <vincenzo.frascino@arm.com>
Co-developed-by: Raveendra Talabattula <raveendra.talabattula@arm.com>
Signed-off-by: Raveendra Talabattula <raveendra.talabattula@arm.com>
Reviewed-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Link: https://patch.msgid.link/20260803142122.4127851-3-raveendra.talabattula@arm.com
GLB_CORE_ID contains the product and version information returned by the
hardware. The current macros parses VERSION_MINOR as a 4-bit field and
VERSION_STATUS as an 8-bit field, which does not match the register
layout. As a result, the driver reports an incorrect version number.
Fix the parsing macros to match the hardware specification:
- VERSION_MINOR is 8 bits at [11:4]
- VERSION_STATUS is 4 bits at [3:0]
Fixes: bd628c1bed ("drm/komeda: komeda_dev/pipeline/component definition and initialzation")
Signed-off-by: Raveendra Talabattula <raveendra.talabattula@arm.com>
Reviewed-by: Liviu Dudau <liviu.dudau@arm.com>
Signed-off-by: Liviu Dudau <liviu.dudau@arm.com>
Link: https://patch.msgid.link/20260803142122.4127851-2-raveendra.talabattula@arm.com
Register the VRAM manager with the dmem cgroup reclaim infrastructure
so that lowering dmem.max below current VRAM usage triggers TTM
eviction rather than failing with -EBUSY.
Guard place->flags in amdgpu_ttm_bo_eviction_valuable() against NULL,
as the TTM reclaim path passes a NULL place in cgroup drain mode.
Use drmm_cgroup_register_region() so that the region is automatically
unregistered at DRM device release, after drm_dev_unplug() has already
made drm_dev_enter() return false. The drm_dev_enter/exit guard in the
reclaim callback ensures no reclaim work touches the TTM manager after
driver unbind, closing the window between vram_mgr_fini() (called from
drm_driver.release) and the drmm cleanup that unregisters the region.
v3:
- Rebased on fix for uninitialized list and buddy allocator on the
drmm_cgroup_register_region() error path.
v5:
- Rebased on the introduction of struct dmem_cgroup_init.
- Clear the reclaim callback in amdgpu_vram_mgr_fini() to prevent
use-after-free if cgroup reclaim is triggered after driver unbind
while userspace holds an open DRM file descriptor. (Sashiko-bot)
- Switch from drmm_cgroup_register_region() to the raw
dmem_cgroup_register_region() and store the region in
amdgpu_vram_mgr.cg_region. Call dmem_cgroup_unregister_region()
in amdgpu_vram_mgr_fini() after ttm_resource_manager_evict_all()
to drain in-flight reclaim callbacks, and clear man->cg afterwards.
This is required because amdgpu's vram manager fini is called
explicitly during driver unbind, which may precede the DRM device
release and thus precede any drmm-based cleanup. (Sashiko-bot)
v6:
- Fix mgr->cg_region never being assigned, so
dmem_cgroup_unregister_region() in fini silently no-ops on NULL
and leaks the region. (Sashiko-bot)
- Reorder fini to call set_used(false) and evict_all() before
dmem_cgroup_unregister_region(), so ttm_resource_free() can
uncharge via man->cg during eviction; clear man->cg after
unregister. (Sashiko-bot)
v7:
- Move dmem_cgroup_unregister_region() before the early return on
evict_all() failure; not doing so leaves a dangling reclaim callback
pointing to the partially-torn-down VRAM manager, causing a
use-after-free when the cgroup later triggers reclaim. (Sashiko-bot)
- Switch back to drmm_cgroup_register_region() with a drm_dev_enter/
exit guard in the reclaim callback (matching xe), rather than manual
register/unregister. drm_dev_unplug() fires before vram_mgr_fini(),
so drm_dev_enter() returning false prevents any reclaim from touching
the manager during teardown. This also fixes the "vram" name
collision on multi-GPU systems, since drmm_cgroup_register_region()
automatically prefixes with "drm/<pci-addr>/". (Sashiko-bot)
v8:
- Move the dmem cgroup region teardown back into
amdgpu_vram_mgr_fini(): register the region with
dmem_cgroup_register_region() (keeping the "drm/<unique>/vram"
prefix), store it in amdgpu_vram_mgr.cg_region, and unregister it
explicitly after ttm_resource_manager_evict_all(). The v7
drmm-only approach left a use-after-free on the probe-error
teardown path: drm_dev_unplug() is not called there, so the
drm_dev_enter() guard stays open while drm_driver.release destroys
the manager before the drmm action unregisters the region. The
explicit unregister drains in-flight reclaim on every teardown
path; the drm_dev_enter() guard is kept as defense against reclaim
after unplug. (Sashiko-bot)
v9:
- Don't leak a gpu buddy manager on cgroup init failure. (Sashiko-bot)
- Acquire a runtime PM reference (pm_runtime_get_sync/put_autosuspend)
around the TTM reclaim call in amdgpu_vram_mgr_dmem_reclaim().
(Sashiko-bot)
Assisted-by: GitHub_Copilot:claude-sonnet-4.6
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Tested-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Link: https://patch.msgid.link/20260725100036.2372-7-thomas.hellstrom@linux.intel.com
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
Register the VRAM manager with the dmem cgroup reclaim infrastructure
so that lowering dmem.max below current VRAM usage triggers TTM
eviction rather than failing with -EBUSY.
v4:
- Rebased on drm-tip; dropped the XE_PL_STOLEN guard as stolen memory
uses a separate TTM manager and never calls __xe_ttm_vram_mgr_init().
v5:
- Rebased on the introduction of struct dmem_cgroup_init.
- Register the fini drmm action before drmm_cgroup_register_region() so
that devres LIFO teardown runs unregister_region() first (draining any
in-flight reclaim callbacks via the rwsem) and xe_ttm_vram_mgr_fini()
second, ensuring the manager is never accessed by a reclaim callback
after teardown. (Sashiko-bot)
- Wrap the reclaim callback in xe_ttm_vram_mgr_dmem_reclaim() using
drm_dev_enter()/drm_dev_exit() to prevent TTM reclaim from running
after driver unbind.
v9:
- Close ttm_resource_manager_set_used() races around vram manager
init and fini.
- Formatting fix.
- Acquire a runtime PM reference around the TTM reclaim call in
xe_ttm_vram_mgr_dmem_reclaim(). (Sashiko-bot)
Assisted-by: GitHub_Copilot:claude-sonnet-4.6
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> #v8
Tested-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Link: https://patch.msgid.link/20260725100036.2372-6-thomas.hellstrom@linux.intel.com
Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
Add ttm_bo_evict_cgroup() to evict buffer objects charged to a specific
dmem cgroup pool from a resource manager's LRU until a byte target is
met. Add ttm_resource_manager_set_dmem_region() to associate a dmem
cgroup region with a resource manager; drivers supply their own
dmem_cgroup_ops with ttm_resource_manager_dmem_reclaim as the reclaim
function and the manager pointer as reclaim_priv in the dmem_cgroup_init
to wire up TTM eviction as the reclaim callback.
The eviction context is interruptible; signals abort the operation and
propagate back through the write() syscall.
Introduce a new mode for the bo LRU walker so that sleeping locks
can be taken. This can be used when the caller doesn't hold any
previous dma_resv locks, and where it intends to hold at most
one lock at a time.
Like the rest of the TTM eviction this should sooner than later
be converted to full WW transactions.
v3:
- Fix ttm_resource_manager_set_dmem_region() storing an error pointer
in man->cg unconditionally. (Sashiko-bot)
- Fix kernel-doc function name format for ttm_bo_evict_cgroup() and
ttm_resource_manager_set_dmem_region().
v5:
- Rebased on the introduction of struct dmem_cgroup_init.
- Handle NULL region in ttm_resource_manager_set_dmem_region() to clear
the reclaim callback, preventing use-after-free when the manager is
torn down while the dmem region outlives it. (Sashiko-bot)
- Return 0 on any progress (even partial eviction), -ENOSPC only when
nothing was freed; fixes callers that expected 0 on partial success.
- Document that the reclaim callback should return 0 if some progress
was made, -ENOSPC if no progress at all, or another error for fatal
failures.
v8:
- Fix ttm_resource_manager_set_dmem_region() using IS_ERR_OR_NULL(),
which skipped the assignment for a NULL region and thus never
cleared man->cg. Use IS_ERR() so that a NULL region detaches the
region as the kernel-doc and the v5 changelog intended. (Sashiko-bot)
v9:
- Don't leak cgroup charges for bos that may have survived dmemcg
region fini.
- Drop the misleading "Capture size before eviction in case res is
cleared" comment in ttm_bo_evict_cb(). (Maarten Lankhorst)
Assisted-by: GitHub_Copilot:claude-sonnet-4.6
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> #v7
Tested-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Link: https://patch.msgid.link/20260725100036.2372-5-thomas.hellstrom@linux.intel.com
Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
Add an optional reclaim callback to struct dmem_cgroup_region. When
dmem.max is set below the current usage of a cgroup pool, the new limit
is applied immediately (so that concurrent allocations are throttled
while reclaim is in progress) and then the driver is asked to evict
memory to bring usage back below the limit.
Reclaim is attempted up to a bounded number of times. No error is
returned to userspace if usage remains above the limit after reclaim,
and a pending signal will abort the reclaim loop early. This matches
the behavior of memory.max in the memory cgroup controller.
Also honor O_NONBLOCK so that if that flag is set during the
max value write, no reclaim is initiated. The idea is to avoid
charging the reclaim cost to the writer of the max value.
v2:
- Write max before reclaim is attempted (Maarten)
- Let signals abort the reclaim without error (Maarten)
- If a new max value is written with the O_NONBLOCK flag,
reclaim is not attempted (Maarten)
- Extract region from the pool parameter rather than
passing it explicitly to set_resource_xxx().
v3:
- Use an rw_semaphore (unregister_sem) to protect reclaim callbacks
against concurrent region unregistration: readers (reclaim) hold the
read side; dmem_cgroup_unregister_region() takes the write side to
drain in-flight callbacks before returning. (Sashiko-bot)
v5:
- Rebased on the introduction of struct dmem_cgroup_init.
- Use nonblock=true in reset_all_resource_limits() to avoid sleeping
inside rcu_read_lock() in dmemcs_offline(). (Sashiko-bot)
- Compare usage against the truncated limit value stored in cnt.max,
not the original u64. (Sashiko-bot)
- Use a DMEM_MAX_RECLAIM_RETRIES (16) retry budget instead of 5, matching
the memcg controller's MAX_RECLAIM_RETRIES. Only -ENOSPC (no progress)
counts against the retry budget; other errors terminate the loop
immediately.
v6:
- Fix dmem_cgroup_ops->reclaim docstring: -ENOSPC does not stop reclaim
immediately but is retried up to DMEM_MAX_RECLAIM_RETRIES times; only
other negative errors terminate the loop. (Sashiko-bot)
v7:
- Replace the per-region rw_semaphore with a static SRCU domain
(dmemcg_srcu). SRCU is a better fit than rwsem for this use: it
avoids the per-region lock overhead on every reclaim call, and
synchronize_srcu() at unregister time is a rare operation. (Maarten)
- Trim in-function comments to focus on what rather than how.
Assisted-by: GitHub_Copilot:claude-sonnet-4.6
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Tested-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Link: https://patch.msgid.link/20260725100036.2372-4-thomas.hellstrom@linux.intel.com
Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
Replace the bare u64 size argument to dmem_cgroup_register_region() and
drmm_cgroup_register_region() with a const struct dmem_cgroup_init *
pointer. The struct currently carries only the size field, but using a
struct makes the API extensible: future callers can supply additional
initialization parameters without adding more positional arguments.
Update all in-tree callers (amdgpu, xe) to use a compound-literal
initializer.
v5:
- Commit introduced.
Assisted-by: GitHub_Copilot:claude-sonnet-4.6
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Tested-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Link: https://patch.msgid.link/20260725100036.2372-3-thomas.hellstrom@linux.intel.com
Acked-by: Dave Airlie <airlied@redhat.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
drmm_cgroup_register_region() is called before INIT_LIST_HEAD() and
gpu_buddy_init() in amdgpu_vram_mgr_init(). If it fails, the function
returns early and bypasses those initializations.
Since adev->mman.initialized is set to true before amdgpu_vram_mgr_init()
is called, a failure triggers amdgpu_ttm_fini(), which calls
amdgpu_vram_mgr_fini(), which then:
- Calls list_for_each_entry_safe() on reservations_pending and
reserved_pages, whose list_head::next pointers are zero-initialized
(NULL). The loop does not recognize them as empty and dereferences NULL.
- Calls gpu_buddy_fini(), which iterates free_trees[] unconditionally
via for_each_free_tree(). Since mm->free_trees is NULL
(never allocated), this dereferences NULL.
Both result in a kernel panic on the module load error path.
Fix by moving drmm_cgroup_register_region() to after the list and buddy
allocator are fully initialized, so the teardown path is safe to run.
Reported-by: Sashiko-bot <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260428073116.15687-1-thomas.hellstrom@linux.intel.com?part=4
Fixes: 2b624a2c18 ("drm/ttm: Handle cgroup based eviction in TTM")
Cc: Friedrich Vock <friedrich.vock@gmx.de>
Cc: Maarten Lankhorst <dev@lankhorst.se>
Cc: Tejun Heo <tj@kernel.org>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: amd-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Cc: stable@vger.kernel.org # v6.14+
Assisted-by: GitHub_Copilot:claude-sonnet-4.6
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
Reviewed-By: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Tested-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Link: https://patch.msgid.link/20260725100036.2372-2-thomas.hellstrom@linux.intel.com
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
The mipi_dsi_dcs_*() functions used by this driver are deprecated in
favor of their _multi() counterparts, as noted in
Documentation/gpu/todo.rst. The _multi() variants record the first
error in a context structure and skip every later call once an error
is set, so the return value no longer has to be checked after each
command. They also log their own failures, which makes the per-call
dev_err() calls redundant.
Convert prepare() and unprepare(). prepare() uses mipi_dsi_msleep()
for the delays between DSI commands. unprepare() uses plain
usleep_range() so the delays run unconditionally after the
accumulated error is cleared. The delays in the GPIO reset sequence
stay as plain msleep() and usleep_range(), since they run before any
DSI transaction.
unprepare() now disables the regulator unconditionally and returns 0.
Previously a failure of set_display_off() was logged and the sequence
continued, while a failure of enter_sleep_mode() returned early,
leaving the regulator enabled and the panel unable to be brought back
up, since drm_panel_unprepare() skips panel->prepared = false on
error. The accumulated error from set_display_off() is cleared so
that the delay and enter_sleep_mode() are still attempted, preserving
the original fall-through behavior. Both drm_panel_prepare() and
drm_panel_unprepare() return void, so the error was never propagated
to a caller in any case.
Signed-off-by: Akash Sukhavasi <akash.sukhavasi@gmail.com>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://patch.msgid.link/20260730-mipi-dsi-tl070wsh30-multi-v3-1-60592caef4f4@gmail.com
Fix sent -> send in the documentation for all variants of
pvr_kccb_send_cmd*().
Signed-off-by: Luigi Santivetti <luigi.santivetti@imgtec.com>
Reviewed-by: Alexandru Dadu <alexandru.dadu@imgtec.com>
Link: https://patch.msgid.link/20260804-staging-pvr-docs-fixes-v2-3-a5a9569a1c1d@imgtec.com
Signed-off-by: Alessio Belle <alessio.belle@imgtec.com>
Update and remove some old comment in the PVR GEM documentation.
Signed-off-by: Matt Coster <matt.coster@imgtec.com>
Signed-off-by: Luigi Santivetti <luigi.santivetti@imgtec.com>
Reviewed-by: Alessio Belle <alessio.belle@imgtec.com>
Link: https://patch.msgid.link/20260804-staging-pvr-docs-fixes-v2-1-a5a9569a1c1d@imgtec.com
Signed-off-by: Alessio Belle <alessio.belle@imgtec.com>
ssd130x_update_bl() runs for every SSD13xx panel, but it only works for
SSD130x and SSD132x: it writes the single global SSD13XX_CONTRAST (0x81)
command, which those two families expose. SSD133x has no such command --
it has three per-channel contrast registers (CONTRAST_A/B/C) that must be
scaled together -- so ssd130x_update_bl() has no effect on it.
Make backlight_ops.update_status a per-family choice. SSD130x and SSD132x
keep ssd130x_update_bl() because they share the SSD13XX_CONTRAST interface,
while SSD133x gets ssd133x_update_bl(), which drives the three channels
through ssd133x_set_contrast().
Signed-off-by: Amit Barzilai <amit.barzilai22@gmail.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patch.msgid.link/20260729053054.29374-3-amit.barzilai22@gmail.com
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
