<feed xmlns='http://www.w3.org/2005/Atom'>
<title>linux-toradex.git/drivers/gpu/drm/amd/display/dc, branch master</title>
<subtitle>Linux kernel for Apalis and Colibri modules</subtitle>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/'/>
<entry>
<title>drm/amd/display: Don't warn about links without an HPD GPIO</title>
<updated>2026-11-18T06:50:45+00:00</updated>
<author>
<name>Thomas Lamprecht</name>
<email>t.lamprecht@proxmox.com</email>
</author>
<published>2026-11-02T11:29:43+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=fb3d0ee0f7cf41488aff58c45f1b31a8f54f3372'/>
<id>fb3d0ee0f7cf41488aff58c45f1b31a8f54f3372</id>
<content type='text'>
Since commit def3488eb0fd ("drm/amd/display: refactor HPD to increase
flexibility"), the link encoder programs the HPD filter, and
dce110_program_hpd_filter() and dcn10_program_hpd_filter() hit ASSERT(0)
if the encoder has no HPD GPIO. That is the case if link_get_hpd_gpio()
finds no usable HPD info for the connector in the VBIOS, where
program_hpd_filter() used to return early without a warning.

This logs a warning with a stack trace on every boot of some systems,
for example with Ryzen 7 PRO 4750G APUs on ASRock Rack X570D4U and
X470D4U2-2T boards:

  WARNING: drivers/gpu/drm/amd/amdgpu/../display/dc/dio/dcn10/dcn10_link_encoder.c:1514 at dcn10_program_hpd_filter+0x39/0x90 [amdgpu], CPU#7: (udev-worker)/836
  Call Trace:
   program_hpd_filter+0x5d/0xb0 [amdgpu]
   construct_phy+0x96f/0xe60 [amdgpu]
   link_create+0x1f0/0x240 [amdgpu]
   create_links+0x164/0x4f0 [amdgpu]
   dc_create+0x3ba/0x720 [amdgpu]
   amdgpu_dm_init+0x2c9/0x420 [amdgpu]

Return false without asserting if there is no HPD GPIO, as before, and
keep the assertion for failing to lock the pin.

Fixes: def3488eb0fd ("drm/amd/display: refactor HPD to increase flexibility")
Link: https://forum.proxmox.com/threads/kernel-7-0-14-24-and-7-0-14-25-bug-in-amdgpu.187181/
Reviewed-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Signed-off-by: Thomas Lamprecht &lt;t.lamprecht@proxmox.com&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 62ea146c23e14f958744507c37025f1a7f9c041b)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Since commit def3488eb0fd ("drm/amd/display: refactor HPD to increase
flexibility"), the link encoder programs the HPD filter, and
dce110_program_hpd_filter() and dcn10_program_hpd_filter() hit ASSERT(0)
if the encoder has no HPD GPIO. That is the case if link_get_hpd_gpio()
finds no usable HPD info for the connector in the VBIOS, where
program_hpd_filter() used to return early without a warning.

This logs a warning with a stack trace on every boot of some systems,
for example with Ryzen 7 PRO 4750G APUs on ASRock Rack X570D4U and
X470D4U2-2T boards:

  WARNING: drivers/gpu/drm/amd/amdgpu/../display/dc/dio/dcn10/dcn10_link_encoder.c:1514 at dcn10_program_hpd_filter+0x39/0x90 [amdgpu], CPU#7: (udev-worker)/836
  Call Trace:
   program_hpd_filter+0x5d/0xb0 [amdgpu]
   construct_phy+0x96f/0xe60 [amdgpu]
   link_create+0x1f0/0x240 [amdgpu]
   create_links+0x164/0x4f0 [amdgpu]
   dc_create+0x3ba/0x720 [amdgpu]
   amdgpu_dm_init+0x2c9/0x420 [amdgpu]

Return false without asserting if there is no HPD GPIO, as before, and
keep the assertion for failing to lock the pin.

Fixes: def3488eb0fd ("drm/amd/display: refactor HPD to increase flexibility")
Link: https://forum.proxmox.com/threads/kernel-7-0-14-24-and-7-0-14-25-bug-in-amdgpu.187181/
Reviewed-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Signed-off-by: Thomas Lamprecht &lt;t.lamprecht@proxmox.com&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 62ea146c23e14f958744507c37025f1a7f9c041b)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amd/display: Fix logger leak on perf trace allocation failure</title>
<updated>2026-11-18T06:49:59+00:00</updated>
<author>
<name>Jackie Liu</name>
<email>liuyun01@kylinos.cn</email>
</author>
<published>2026-10-26T11:10:47+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=5b148562d193fd63e5d1ae3c442d3d54403c7e51'/>
<id>5b148562d193fd63e5d1ae3c442d3d54403c7e51</id>
<content type='text'>
dc_construct_ctx() allocates a logger and stores an alias in dc-&gt;dml.logger
before allocating the performance trace object. If dc_perf_trace_create()
fails, the function frees dc_ctx without releasing the logger. Since
dc-&gt;ctx is assigned only after the trace allocation succeeds,
dc_destruct() cannot reclaim the logger on this error path.

Free the logger before discarding dc_ctx and clear dc-&gt;dml.logger to avoid
leaving a dangling alias after the failed construction.

Fixes: 5d72e247e58c ("drm/amd/display: switch DC over to the new DRM logging macros")
Reviewed-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Signed-off-by: Jackie Liu &lt;liuyun01@kylinos.cn&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit ac088505e74932f276ea6e3bbe19713e5770e7a1)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
dc_construct_ctx() allocates a logger and stores an alias in dc-&gt;dml.logger
before allocating the performance trace object. If dc_perf_trace_create()
fails, the function frees dc_ctx without releasing the logger. Since
dc-&gt;ctx is assigned only after the trace allocation succeeds,
dc_destruct() cannot reclaim the logger on this error path.

Free the logger before discarding dc_ctx and clear dc-&gt;dml.logger to avoid
leaving a dangling alias after the failed construction.

Fixes: 5d72e247e58c ("drm/amd/display: switch DC over to the new DRM logging macros")
Reviewed-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Signed-off-by: Jackie Liu &lt;liuyun01@kylinos.cn&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit ac088505e74932f276ea6e3bbe19713e5770e7a1)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amd/display: Abort DPMS ON when enabling the link fails</title>
<updated>2026-11-18T06:47:49+00:00</updated>
<author>
<name>Abdulla Bubshait</name>
<email>darkstego@gmail.com</email>
</author>
<published>2026-10-28T20:26:41+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=4aa51b904cac06fe2c82623ce24d80853af730f6'/>
<id>4aa51b904cac06fe2c82623ce24d80853af730f6</id>
<content type='text'>
Since commit 843049bac7c0 ("drm/amd/display: Split DPMS ON into
parts"), link_set_dpms_on_enable_link() returns
DC_DPMS_FAILED_HANDSHAKE when enable_link() fails and the stream
enable has to be aborted: any failure other than link training, or a
link training failure on DP MST. The link has already been disabled at
that point. The step loop in link_set_dpms_on() treats
DC_DPMS_FAILED_HANDSHAKE like DC_DPMS_SUCCESS and goes on to
link_set_dpms_on_post_enable_link().

Before the split, this path returned right after disable_link(),
because MST displays can't proceed without link training. Now, for
failed MST handshakes MST slots are still allocated.  When the link
trains again later, the new payload starts after the leaked slots so
ACT never completes, and the display stays black until
reboot. Replugging or power-cycling the monitor does not recover it.

Seen on Navi 33 (DCN 3.2.1) with a DP 2.1 MST monitor (Odyssey Neo G9)
at UHBR10 x4 with DSC. Link training sometimes fails on DPMS
wake. With this change, the failed attempt leaves no payload behind,
and the next modeset after the sink reconnects bring the monitor back
up.

Return early on DC_DPMS_FAILED_HANDSHAKE, as for
DC_DPMS_SKIPPED_HANDSHAKE. This restores the pre-split behaviour. Link
training failures on non-MST signals still return DC_DPMS_SUCCESS and
continue as before.

Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/6047
Fixes: 843049bac7c0 ("drm/amd/display: Split DPMS ON into parts")
Reviewed-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Signed-off-by: Abdulla Bubshait &lt;darkstego@gmail.com&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit ece3dae13fb106684f8248f88387b1fbcfef8206)
Cc: stable@vger.kernel.org # 7.3
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Since commit 843049bac7c0 ("drm/amd/display: Split DPMS ON into
parts"), link_set_dpms_on_enable_link() returns
DC_DPMS_FAILED_HANDSHAKE when enable_link() fails and the stream
enable has to be aborted: any failure other than link training, or a
link training failure on DP MST. The link has already been disabled at
that point. The step loop in link_set_dpms_on() treats
DC_DPMS_FAILED_HANDSHAKE like DC_DPMS_SUCCESS and goes on to
link_set_dpms_on_post_enable_link().

Before the split, this path returned right after disable_link(),
because MST displays can't proceed without link training. Now, for
failed MST handshakes MST slots are still allocated.  When the link
trains again later, the new payload starts after the leaked slots so
ACT never completes, and the display stays black until
reboot. Replugging or power-cycling the monitor does not recover it.

Seen on Navi 33 (DCN 3.2.1) with a DP 2.1 MST monitor (Odyssey Neo G9)
at UHBR10 x4 with DSC. Link training sometimes fails on DPMS
wake. With this change, the failed attempt leaves no payload behind,
and the next modeset after the sink reconnects bring the monitor back
up.

Return early on DC_DPMS_FAILED_HANDSHAKE, as for
DC_DPMS_SKIPPED_HANDSHAKE. This restores the pre-split behaviour. Link
training failures on non-MST signals still return DC_DPMS_SUCCESS and
continue as before.

Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/6047
Fixes: 843049bac7c0 ("drm/amd/display: Split DPMS ON into parts")
Reviewed-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Signed-off-by: Abdulla Bubshait &lt;darkstego@gmail.com&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit ece3dae13fb106684f8248f88387b1fbcfef8206)
Cc: stable@vger.kernel.org # 7.3
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amd/display: Seed boot clocks to avoid lowering pre-OS display clocks</title>
<updated>2026-11-18T06:46:56+00:00</updated>
<author>
<name>Jing Zhou</name>
<email>Jing.Zhou@amd.com</email>
</author>
<published>2026-11-05T15:20:58+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=eaf4ff850d97eb608482fa3497f1b89b2e628781'/>
<id>eaf4ff850d97eb608482fa3497f1b89b2e628781</id>
<content type='text'>
[WHY]
USB4 DPIA HPD can arrive early in boot, before the first stream commit.
Its link detection runs prepare_phy_clocks(), which goes through
clk_mgr_exit_optimized_pwr_state() and clk_mgr_optimize_pwr_state()
and calls update_clocks(dc-&gt;current_state).

At that point both dc-&gt;current_state-&gt;bw_ctx.bw.dcn.clk and
clk_mgr-&gt;clks are still zero, because init_clocks() memsets them. On the
safe_to_lower path, dcn42_update_clocks() clamps the requested DPPCLK to
the 100MHz floor and programs it. This drops DPPCLK below what the
GOP-lit eDP pipe needs, and the panel flashes white.

[HOW]
1. In dcn42_init_clocks() and dcn42b_init_clocks(), seed clk_mgr-&gt;clks
   dispclk/dppclk from the boot snapshot(10kHz units, x10).
2. Add a DC core helper, sync_current_state_clocks_from_clk_mgr().
   After every hwss.init_hw() call (legacy_hardware_init, D0 entry in
   dc_set_power_state, and dc_reinitialize_hardware), it copies
   clk_mgr-&gt;clks into dc-&gt;current_state-&gt;bw_ctx.bw.dcn.clk for DCN.
   A later update_clocks() before the first commit then sees the real
   pre-OS clocks and does not lower them. Raising clocks is still
   allowed and does not cause a flash.

Cc: Nicholas Kazlauskas &lt;nicholas.kazlauskas@amd.com&gt;
Cc: Charlene Liu &lt;charlene.liu@amd.com&gt;
Cc: Mario Limonciello &lt;mario.limonciello@amd.com&gt;
Cc: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Reviewed-by: ChiaHsuan Chung &lt;chiahsuan.chung@amd.com&gt;
Signed-off-by: Jing Zhou &lt;Jing.Zhou@amd.com&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Tested-by: Dan Wheeler &lt;daniel.wheeler@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 2f289740a04f08a5ce4b5e3c69af6e32ab4a540e)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
[WHY]
USB4 DPIA HPD can arrive early in boot, before the first stream commit.
Its link detection runs prepare_phy_clocks(), which goes through
clk_mgr_exit_optimized_pwr_state() and clk_mgr_optimize_pwr_state()
and calls update_clocks(dc-&gt;current_state).

At that point both dc-&gt;current_state-&gt;bw_ctx.bw.dcn.clk and
clk_mgr-&gt;clks are still zero, because init_clocks() memsets them. On the
safe_to_lower path, dcn42_update_clocks() clamps the requested DPPCLK to
the 100MHz floor and programs it. This drops DPPCLK below what the
GOP-lit eDP pipe needs, and the panel flashes white.

[HOW]
1. In dcn42_init_clocks() and dcn42b_init_clocks(), seed clk_mgr-&gt;clks
   dispclk/dppclk from the boot snapshot(10kHz units, x10).
2. Add a DC core helper, sync_current_state_clocks_from_clk_mgr().
   After every hwss.init_hw() call (legacy_hardware_init, D0 entry in
   dc_set_power_state, and dc_reinitialize_hardware), it copies
   clk_mgr-&gt;clks into dc-&gt;current_state-&gt;bw_ctx.bw.dcn.clk for DCN.
   A later update_clocks() before the first commit then sees the real
   pre-OS clocks and does not lower them. Raising clocks is still
   allowed and does not cause a flash.

Cc: Nicholas Kazlauskas &lt;nicholas.kazlauskas@amd.com&gt;
Cc: Charlene Liu &lt;charlene.liu@amd.com&gt;
Cc: Mario Limonciello &lt;mario.limonciello@amd.com&gt;
Cc: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Reviewed-by: ChiaHsuan Chung &lt;chiahsuan.chung@amd.com&gt;
Signed-off-by: Jing Zhou &lt;Jing.Zhou@amd.com&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Tested-by: Dan Wheeler &lt;daniel.wheeler@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 2f289740a04f08a5ce4b5e3c69af6e32ab4a540e)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amd/display: Update DCN42B stutter latency to match DCN42</title>
<updated>2026-11-18T06:46:33+00:00</updated>
<author>
<name>Allen Pan</name>
<email>Allen.Pan@amd.com</email>
</author>
<published>2026-11-02T20:52:40+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=007e7bf5692d93535adf943cd93c0e49b4e20e9d'/>
<id>007e7bf5692d93535adf943cd93c0e49b4e20e9d</id>
<content type='text'>
[Why]
The DCN42B stutter latencies do not match DCN42, which can lead to
underflow and stutter issues.

[How]
Update the DCN42B LPDDR5/LPCAMM2 stutter latencies to the same
values DCN42 already uses.

Cc: Dillon Varone &lt;dillon.varone@amd.com&gt;
Cc: Charlene Liu &lt;charlene.liu@amd.com&gt;
Cc: Mario Limonciello &lt;mario.limonciello@amd.com&gt;
Cc: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Reviewed-by: ChiaHsuan Chung &lt;chiahsuan.chung@amd.com&gt;
Signed-off-by: Allen Pan &lt;Allen.Pan@amd.com&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Tested-by: Dan Wheeler &lt;daniel.wheeler@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit c6b576ac64cedc990f3885efa85603fe5998da3f)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
[Why]
The DCN42B stutter latencies do not match DCN42, which can lead to
underflow and stutter issues.

[How]
Update the DCN42B LPDDR5/LPCAMM2 stutter latencies to the same
values DCN42 already uses.

Cc: Dillon Varone &lt;dillon.varone@amd.com&gt;
Cc: Charlene Liu &lt;charlene.liu@amd.com&gt;
Cc: Mario Limonciello &lt;mario.limonciello@amd.com&gt;
Cc: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Reviewed-by: ChiaHsuan Chung &lt;chiahsuan.chung@amd.com&gt;
Signed-off-by: Allen Pan &lt;Allen.Pan@amd.com&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Tested-by: Dan Wheeler &lt;daniel.wheeler@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit c6b576ac64cedc990f3885efa85603fe5998da3f)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amd/display: Read DRAM channel width from SMU init table</title>
<updated>2026-11-18T06:46:23+00:00</updated>
<author>
<name>Wenjing Liu</name>
<email>wenjing.liu@amd.com</email>
</author>
<published>2026-11-04T19:39:45+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=3846f4542b4ba5a95405527d1e68864a9e9f3fcf'/>
<id>3846f4542b4ba5a95405527d1e68864a9e9f3fcf</id>
<content type='text'>
[Why]
The channel count came from the SMU init table but the channel
width came from VBIOS vram info. Both values need to describe the
same memory type.

[How]
Add DramChannelWidthBytes to MemoryConfig_t, using a reserved slot
so the layout is unchanged. Populate the channel width in bw_params
from the init table instead of VBIOS vram info.

Cc: Dillon Varone &lt;dillon.varone@amd.com&gt;
Cc: Mario Limonciello &lt;mario.limonciello@amd.com&gt;
Cc: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Reviewed-by: ChiaHsuan Chung &lt;chiahsuan.chung@amd.com&gt;
Signed-off-by: Wenjing Liu &lt;wenjing.liu@amd.com&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Tested-by: Dan Wheeler &lt;daniel.wheeler@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 37d64c58f610d771b7ab8cfbf7d9f5606939947b)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
[Why]
The channel count came from the SMU init table but the channel
width came from VBIOS vram info. Both values need to describe the
same memory type.

[How]
Add DramChannelWidthBytes to MemoryConfig_t, using a reserved slot
so the layout is unchanged. Populate the channel width in bw_params
from the init table instead of VBIOS vram info.

Cc: Dillon Varone &lt;dillon.varone@amd.com&gt;
Cc: Mario Limonciello &lt;mario.limonciello@amd.com&gt;
Cc: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Reviewed-by: ChiaHsuan Chung &lt;chiahsuan.chung@amd.com&gt;
Signed-off-by: Wenjing Liu &lt;wenjing.liu@amd.com&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Tested-by: Dan Wheeler &lt;daniel.wheeler@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 37d64c58f610d771b7ab8cfbf7d9f5606939947b)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amd/display: Flip dml2_core_utils_get_segment_horizontal_contiguous() Inputs</title>
<updated>2026-11-18T06:46:13+00:00</updated>
<author>
<name>Austin Zheng</name>
<email>Austin.Zheng@amd.com</email>
</author>
<published>2026-09-23T18:58:32+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=8bcb6863ac318dfadf622598171c79a020fdd7ff'/>
<id>8bcb6863ac318dfadf622598171c79a020fdd7ff</id>
<content type='text'>
[Why/How]
Parameters for dml2_core_utils_get_segment_horizontal_contiguous() should be flipped.
Should be swizzle mode then bytes per pixel.

Cc: Dillon Varone &lt;dillon.varone@amd.com&gt;
Cc: Mario Limonciello &lt;mario.limonciello@amd.com&gt;
Cc: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Reviewed-by: ChiaHsuan Chung &lt;chiahsuan.chung@amd.com&gt;
Signed-off-by: Austin Zheng &lt;Austin.Zheng@amd.com&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Tested-by: Dan Wheeler &lt;daniel.wheeler@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 8e985b38672c3de9355dd8f166527f96e472706c)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
[Why/How]
Parameters for dml2_core_utils_get_segment_horizontal_contiguous() should be flipped.
Should be swizzle mode then bytes per pixel.

Cc: Dillon Varone &lt;dillon.varone@amd.com&gt;
Cc: Mario Limonciello &lt;mario.limonciello@amd.com&gt;
Cc: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Reviewed-by: ChiaHsuan Chung &lt;chiahsuan.chung@amd.com&gt;
Signed-off-by: Austin Zheng &lt;Austin.Zheng@amd.com&gt;
Signed-off-by: Wayne Lin &lt;Wayne.Lin@amd.com&gt;
Tested-by: Dan Wheeler &lt;daniel.wheeler@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 8e985b38672c3de9355dd8f166527f96e472706c)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amd/display: Fix __udivdi3 link error in DCN6 DML2</title>
<updated>2026-11-11T20:53:14+00:00</updated>
<author>
<name>Ivan Lipski</name>
<email>ivan.lipski@amd.com</email>
</author>
<published>2026-10-28T15:11:11+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=d489c42c9bd542b3fbbda97b97dc9a0b07010eb3'/>
<id>d489c42c9bd542b3fbbda97b97dc9a0b07010eb3</id>
<content type='text'>
[Why&amp;How]
dchub_v2_get_sop_total_available_bandwidth_KBps() and
dchub_v2_is_qos_bandwidth_supported_by_sop() divide 64-bit integers
by 100. On 32-bit builds this emits a call to __udivdi3, which the
kernel does not link.

Use double-cast division instead.

Fixes: 7f7d7ea1fa51 ("drm/amd/display: Add new sources for DCN6")
Cc: Linlin Yang &lt;yanglinlin@kylinos.cn&gt;
Cc: Mario Limonciello &lt;mario.limonciello@amd.com&gt;
Cc: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Cc: Dillon Varone &lt;dillon.varone@amd.com&gt;
Signed-off-by: Ivan Lipski &lt;ivan.lipski@amd.com&gt;
Reviewed-by: Alex Hung &lt;alex.hung@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 83e7c8cb33363f2a2f893758c630f9a805ac4573)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
[Why&amp;How]
dchub_v2_get_sop_total_available_bandwidth_KBps() and
dchub_v2_is_qos_bandwidth_supported_by_sop() divide 64-bit integers
by 100. On 32-bit builds this emits a call to __udivdi3, which the
kernel does not link.

Use double-cast division instead.

Fixes: 7f7d7ea1fa51 ("drm/amd/display: Add new sources for DCN6")
Cc: Linlin Yang &lt;yanglinlin@kylinos.cn&gt;
Cc: Mario Limonciello &lt;mario.limonciello@amd.com&gt;
Cc: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Cc: Dillon Varone &lt;dillon.varone@amd.com&gt;
Signed-off-by: Ivan Lipski &lt;ivan.lipski@amd.com&gt;
Reviewed-by: Alex Hung &lt;alex.hung@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 83e7c8cb33363f2a2f893758c630f9a805ac4573)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
<entry>
<title>Merge tag 'devm_kmalloc_obj-v7.4-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux</title>
<updated>2026-11-08T22:55:54+00:00</updated>
<author>
<name>Linus Torvalds</name>
<email>torvalds@linux-foundation.org</email>
</author>
<published>2026-11-08T22:55:54+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=e0d8e8de8bcb1ef7b749033b5af5518ba8df7bec'/>
<id>e0d8e8de8bcb1ef7b749033b5af5518ba8df7bec</id>
<content type='text'>
Pull devm_kmalloc_obj conversion from Kees Cook:

 - treewide refresh of kmalloc_obj() conversions

 - devm_kmalloc_obj conversion rules added to Coccinelle scripting

 - treewide conversion of devm_kmalloc() family to devm_kmalloc_obj()

* tag 'devm_kmalloc_obj-v7.4-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
  treewide: Convert devm_kmalloc() family to the devm_kmalloc_obj() family
  coccinelle: kmalloc_obj: Include devm_kmalloc_obj conversion rules
  treewide: refresh kmalloc_obj() conversions
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Pull devm_kmalloc_obj conversion from Kees Cook:

 - treewide refresh of kmalloc_obj() conversions

 - devm_kmalloc_obj conversion rules added to Coccinelle scripting

 - treewide conversion of devm_kmalloc() family to devm_kmalloc_obj()

* tag 'devm_kmalloc_obj-v7.4-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
  treewide: Convert devm_kmalloc() family to the devm_kmalloc_obj() family
  coccinelle: kmalloc_obj: Include devm_kmalloc_obj conversion rules
  treewide: refresh kmalloc_obj() conversions
</pre>
</div>
</content>
</entry>
<entry>
<title>treewide: refresh kmalloc_obj() conversions</title>
<updated>2026-11-06T05:55:37+00:00</updated>
<author>
<name>Kees Cook</name>
<email>kees+treewide@kernel.org</email>
</author>
<published>2026-10-29T06:08:12+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=0af298e62e5cff4ebf9bec47b8711177a7e2f115'/>
<id>0af298e62e5cff4ebf9bec47b8711177a7e2f115</id>
<content type='text'>
Refresh the tree with run of the existing Coccinelle rules in
scripts/coccinelle/api/kmalloc_objs.cocci, now that they also convert
arrays of multi-byte integers.

The conversions break down as:

	435	kzalloc_objs()
	169	kmalloc_objs()
	 52	kvzalloc_objs()
	 47	kzalloc_obj()
	 42	kvmalloc_objs()
	  8	kmalloc_obj()
	  2	kzalloc_flex()
	  2	kvmalloc_flex()

and 37 existing kmalloc_obj() family calls drop a redundant GFP_KERNEL.

Build tested with allmodconfig on x86_64 with GCC 16.2.0, and, together
with the devm_kmalloc_obj() conversion that follows, on x86_64, arm64,
i386, arm, riscv, s390, m68k, mips, parisc, sparc, loongarch, arc,
xtensa, and openrisc with GCC 16.2.0 with no new warnings.

Signed-off-by: Kees Cook &lt;kees+treewide@kernel.org&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Refresh the tree with run of the existing Coccinelle rules in
scripts/coccinelle/api/kmalloc_objs.cocci, now that they also convert
arrays of multi-byte integers.

The conversions break down as:

	435	kzalloc_objs()
	169	kmalloc_objs()
	 52	kvzalloc_objs()
	 47	kzalloc_obj()
	 42	kvmalloc_objs()
	  8	kmalloc_obj()
	  2	kzalloc_flex()
	  2	kvmalloc_flex()

and 37 existing kmalloc_obj() family calls drop a redundant GFP_KERNEL.

Build tested with allmodconfig on x86_64 with GCC 16.2.0, and, together
with the devm_kmalloc_obj() conversion that follows, on x86_64, arm64,
i386, arm, riscv, s390, m68k, mips, parisc, sparc, loongarch, arc,
xtensa, and openrisc with GCC 16.2.0 with no new warnings.

Signed-off-by: Kees Cook &lt;kees+treewide@kernel.org&gt;
</pre>
</div>
</content>
</entry>
</feed>
