<feed xmlns='http://www.w3.org/2005/Atom'>
<title>linux-toradex.git/drivers/gpu/drm/amd/pm, 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/pm: send smu8 sclk level indices, not the level mask</title>
<updated>2026-11-18T06:45:05+00:00</updated>
<author>
<name>Joshua Horne</name>
<email>hornejd@etsu.edu</email>
</author>
<published>2026-11-10T02:01:45+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=535f6852652e3239c8688491ccd0ed4b6e09d14f'/>
<id>535f6852652e3239c8688491ccd0ed4b6e09d14f</id>
<content type='text'>
smu8_force_clock_level() passes the pp_dpm_sclk level mask straight to
PPSMC_MSG_SetSclkSoftMin and PPSMC_MSG_SetSclkSoftMax, which take a
level index, as the other smu8 callers do through smu8_get_sclk_level().
In manual mode, "echo 0 &gt; pp_dpm_sclk" thus selects level 1, "echo 2"
selects level 4, and masks above the highest level index are out of
range for the SMU, so e.g. "echo 3", "echo 7" or "echo 0 1 2 3 4 5 6 7"
do not select the requested levels. The original single level version
sent 1 &lt;&lt; level and had the same problem.

Send the lowest and highest level in the mask, as smu10 does, and reject
levels above those the SMU has enabled, as smu10 and vega20 do.

Fixes: d7341ef66870 ("drm/amd/powerplay: add some hwmgr functions for sysfs interface on Carrizo")
Reviewed-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Signed-off-by: Joshua Horne &lt;hornejd@etsu.edu&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 8ff6b368360adb24ccf145047b0261dfae77f28a)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
smu8_force_clock_level() passes the pp_dpm_sclk level mask straight to
PPSMC_MSG_SetSclkSoftMin and PPSMC_MSG_SetSclkSoftMax, which take a
level index, as the other smu8 callers do through smu8_get_sclk_level().
In manual mode, "echo 0 &gt; pp_dpm_sclk" thus selects level 1, "echo 2"
selects level 4, and masks above the highest level index are out of
range for the SMU, so e.g. "echo 3", "echo 7" or "echo 0 1 2 3 4 5 6 7"
do not select the requested levels. The original single level version
sent 1 &lt;&lt; level and had the same problem.

Send the lowest and highest level in the mask, as smu10 does, and reject
levels above those the SMU has enabled, as smu10 and vega20 do.

Fixes: d7341ef66870 ("drm/amd/powerplay: add some hwmgr functions for sysfs interface on Carrizo")
Reviewed-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Signed-off-by: Joshua Horne &lt;hornejd@etsu.edu&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 8ff6b368360adb24ccf145047b0261dfae77f28a)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amdgpu: use supported SMU clock messages on SMU 15.0.0</title>
<updated>2026-11-11T20:56:57+00:00</updated>
<author>
<name>Kanala Ramalingeswara Reddy</name>
<email>Kanala.RamalingeswaraReddy@amd.com</email>
</author>
<published>2026-10-28T16:24:53+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=6c9ad0fc63930830f12a4ccc9555121cc7db0aee'/>
<id>6c9ad0fc63930830f12a4ccc9555121cc7db0aee</id>
<content type='text'>
SMU 15.0.0 no longer implements SMU_MSG_SetHardMinGfxClk, or the
SMU_MSG_SetHardMin and SMU_MSG_SetHardMax messages. Sending them
makes the clock request fail.

Program the GFX hard-min clock with SMU_MSG_SetSoftMinGfxclk, which
this SMU firmware still supports.

Fixes: c7fc0f372374 ("drm/amd: Enable SMU 15_0_0 support")
Signed-off-by: Kanala Ramalingeswara Reddy &lt;Kanala.RamalingeswaraReddy@amd.com&gt;
Acked-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit a6ae245cea297fae613145d5e4c74145f81f2ee1)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
SMU 15.0.0 no longer implements SMU_MSG_SetHardMinGfxClk, or the
SMU_MSG_SetHardMin and SMU_MSG_SetHardMax messages. Sending them
makes the clock request fail.

Program the GFX hard-min clock with SMU_MSG_SetSoftMinGfxclk, which
this SMU firmware still supports.

Fixes: c7fc0f372374 ("drm/amd: Enable SMU 15_0_0 support")
Signed-off-by: Kanala Ramalingeswara Reddy &lt;Kanala.RamalingeswaraReddy@amd.com&gt;
Acked-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit a6ae245cea297fae613145d5e4c74145f81f2ee1)
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>
<entry>
<title>drm/amd/pm: track BACO state in the SMU v11 ArmD3 sequence</title>
<updated>2026-11-04T17:46:40+00:00</updated>
<author>
<name>Mario Limonciello</name>
<email>mario.limonciello@amd.com</email>
</author>
<published>2026-11-03T14:14:57+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=a26ec190925733dd205507493e936e59dbcd2b19'/>
<id>a26ec190925733dd205507493e936e59dbcd2b19</id>
<content type='text'>
On Navi1x/Sienna Cichlid with runtime PM, when the audio function is
enabled, BACO entry and exit are delegated to PMFW by sending
SMU_MSG_ArmD3 instead of going through smu_v11_0_baco_set_state().
Unlike the v13/v14/v15 helpers, smu_v11_0_baco_set_armd3_sequence()
does not update smu_baco-&gt;state, so the driver keeps reporting
SMU_BACO_STATE_EXIT while the ASIC is armed for / sitting in BACO.

At runtime resume amdgpu_device_baco_exit() checks
amdgpu_device_supports_baco(), which ends up in
smu_v11_0_get_bamaco_support().  Since the state never reads ENTER, the
"already in BACO" shortcut is skipped and the live BACO feature bit is
consulted instead.  PMFW clears that bit once ArmD3(BACO) is armed, so
the check reports "supported but not enabled", the exit fails with
-ENOTSUPP and the ULPS ArmD3 message is never sent.  The runtime resume
caller ignores the error, and a later secondary bus reset leaves the
device unreachable.  This shows up on a Navi 23 being handed from amdgpu
to vfio-pci:

  baco_enter: in_runpm=1 audio_enabled=1 armd3=1 baco_state=1 ret=0
  bamaco: state=1 ... feat supported=1 enabled=0 in_runpm=1
  baco_exit: supports=0 ...
  runtime resume: baco_exit ret=-524

Update smu_baco-&gt;state in the helper, mirroring the v13 implementation,
so the capability check short-circuits while in BACO.

Reported-by: Nathan Blomquist &lt;nblomquist@gmail.com&gt;
Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/6076
Tested-by: Nathan Blomquist &lt;nblomquist@gmail.com&gt;
Fixes: 13d75eadfa1a ("drm/amd/pm: correct the runpm handling for BACO supported ASIC")
Acked-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Link: https://patch.msgid.link/20261103141457.2558235-1-mario.limonciello@amd.com
Signed-off-by: Mario Limonciello &lt;mario.limonciello@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit d515b2694eab5b35533c7fb7af5bddff53ff60d3)
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
On Navi1x/Sienna Cichlid with runtime PM, when the audio function is
enabled, BACO entry and exit are delegated to PMFW by sending
SMU_MSG_ArmD3 instead of going through smu_v11_0_baco_set_state().
Unlike the v13/v14/v15 helpers, smu_v11_0_baco_set_armd3_sequence()
does not update smu_baco-&gt;state, so the driver keeps reporting
SMU_BACO_STATE_EXIT while the ASIC is armed for / sitting in BACO.

At runtime resume amdgpu_device_baco_exit() checks
amdgpu_device_supports_baco(), which ends up in
smu_v11_0_get_bamaco_support().  Since the state never reads ENTER, the
"already in BACO" shortcut is skipped and the live BACO feature bit is
consulted instead.  PMFW clears that bit once ArmD3(BACO) is armed, so
the check reports "supported but not enabled", the exit fails with
-ENOTSUPP and the ULPS ArmD3 message is never sent.  The runtime resume
caller ignores the error, and a later secondary bus reset leaves the
device unreachable.  This shows up on a Navi 23 being handed from amdgpu
to vfio-pci:

  baco_enter: in_runpm=1 audio_enabled=1 armd3=1 baco_state=1 ret=0
  bamaco: state=1 ... feat supported=1 enabled=0 in_runpm=1
  baco_exit: supports=0 ...
  runtime resume: baco_exit ret=-524

Update smu_baco-&gt;state in the helper, mirroring the v13 implementation,
so the capability check short-circuits while in BACO.

Reported-by: Nathan Blomquist &lt;nblomquist@gmail.com&gt;
Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/6076
Tested-by: Nathan Blomquist &lt;nblomquist@gmail.com&gt;
Fixes: 13d75eadfa1a ("drm/amd/pm: correct the runpm handling for BACO supported ASIC")
Acked-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
Link: https://patch.msgid.link/20261103141457.2558235-1-mario.limonciello@amd.com
Signed-off-by: Mario Limonciello &lt;mario.limonciello@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit d515b2694eab5b35533c7fb7af5bddff53ff60d3)
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amd/pm: populate unique ID for smu 13.0.7</title>
<updated>2026-11-04T17:39:37+00:00</updated>
<author>
<name>Kevin Wang</name>
<email>kevin.wang@amd.com</email>
</author>
<published>2026-10-29T03:50:31+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=079f9103d9cde68fb2ccc414e94520f43ee102de'/>
<id>079f9103d9cde68fb2ccc414e94520f43ee102de</id>
<content type='text'>
The smu 13.0.7 backend does not register get_unique_id, leaving
adev-&gt;unique_id at zero and KFD topology without the stable GPU
identifier consumed by HIP.

The failure is observed at:
    auto ierr = hipDeviceGetUuid(&amp;uuid, 0);

Fixes: 113cc31dde66 ("drm/amd/pm: add SMU_13_0_7 ppt_funcs for SMU_13_0_7")
Signed-off-by: Kevin Wang &lt;kevin.wang@amd.com&gt;
Reviewed-by: Kenneth Feng &lt;kenneth.feng@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 5e2585daaf8336d9ce1bf52e55a9b7a8cdbf625b)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
The smu 13.0.7 backend does not register get_unique_id, leaving
adev-&gt;unique_id at zero and KFD topology without the stable GPU
identifier consumed by HIP.

The failure is observed at:
    auto ierr = hipDeviceGetUuid(&amp;uuid, 0);

Fixes: 113cc31dde66 ("drm/amd/pm: add SMU_13_0_7 ppt_funcs for SMU_13_0_7")
Signed-off-by: Kevin Wang &lt;kevin.wang@amd.com&gt;
Reviewed-by: Kenneth Feng &lt;kenneth.feng@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 5e2585daaf8336d9ce1bf52e55a9b7a8cdbf625b)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amd/pm: preserve committed OD features for SMU 14.0.2</title>
<updated>2026-10-27T21:04:09+00:00</updated>
<author>
<name>Kevin Wang</name>
<email>kevin.wang@amd.com</email>
</author>
<published>2026-10-15T06:02:31+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=9b0890e72d19b0c430fe62bad290c7c79baa78e4'/>
<id>9b0890e72d19b0c430fe62bad290c7c79baa78e4</id>
<content type='text'>
PMFW treats FeatureCtrlMask as the complete set of active OD features.
Uploads containing only newly edited bits can drop earlier settings.

Fix the OD state handling to:

- preserve committed features across later commits
- keep fan-curve and legacy-fan bits mutually exclusive
- keep staged OD settings out of power-cap writes
- update cached state only after successful uploads
- clear restored features only after they return to defaults
- restore PPT for the current AC/DC source after resume

Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/3625
Signed-off-by: Kevin Wang &lt;kevin.wang@amd.com&gt;
Reviewed-by: Kenneth Feng &lt;kenneth.feng@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 8d4ce79aff1cd1ff4feea4967c30e04eb6be8c2d)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
PMFW treats FeatureCtrlMask as the complete set of active OD features.
Uploads containing only newly edited bits can drop earlier settings.

Fix the OD state handling to:

- preserve committed features across later commits
- keep fan-curve and legacy-fan bits mutually exclusive
- keep staged OD settings out of power-cap writes
- update cached state only after successful uploads
- clear restored features only after they return to defaults
- restore PPT for the current AC/DC source after resume

Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/3625
Signed-off-by: Kevin Wang &lt;kevin.wang@amd.com&gt;
Reviewed-by: Kenneth Feng &lt;kenneth.feng@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 8d4ce79aff1cd1ff4feea4967c30e04eb6be8c2d)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amd/pm: add zero fan stop temperature support on smu 14.0.2</title>
<updated>2026-10-27T21:04:01+00:00</updated>
<author>
<name>Kenneth Feng</name>
<email>kenneth.feng@amd.com</email>
</author>
<published>2026-10-08T12:28:09+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=af52c8fd45fbd5e3b7a329858733ac674abe539c'/>
<id>af52c8fd45fbd5e3b7a329858733ac674abe539c</id>
<content type='text'>
add zero fan stop temperature support on smu 14.0.2

Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/5905
Signed-off-by: Kenneth Feng &lt;kenneth.feng@amd.com&gt;
Reviewed-by: Kevin Wang &lt;kevin.wang@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit f990c94a2633d0ead477eb1b174e254a6d292e66)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
add zero fan stop temperature support on smu 14.0.2

Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/5905
Signed-off-by: Kenneth Feng &lt;kenneth.feng@amd.com&gt;
Reviewed-by: Kevin Wang &lt;kevin.wang@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit f990c94a2633d0ead477eb1b174e254a6d292e66)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amd/pm: preserve committed OD features for SMU 13.0.0/13.0.7</title>
<updated>2026-10-27T21:03:52+00:00</updated>
<author>
<name>Kevin Wang</name>
<email>kevin.wang@amd.com</email>
</author>
<published>2026-10-15T06:02:09+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=6027db7db773714f2e05df82093276f2a52a1fba'/>
<id>6027db7db773714f2e05df82093276f2a52a1fba</id>
<content type='text'>
PMFW treats FeatureCtrlMask as the complete set of active OD features.
Uploads containing only newly edited bits can drop earlier settings.

Fix the OD state handling to:

- preserve committed features across later commits
- keep staged OD settings out of power-cap writes
- update cached state only after successful uploads
- clear restored features only after they return to defaults
- restore PPT for the current AC/DC source after resume

Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/3625
Signed-off-by: Kevin Wang &lt;kevin.wang@amd.com&gt;
Reviewed-by: Kenneth Feng &lt;kenneth.feng@amd.com&gt;
Tested-by: José Miguel Sarasola &lt;alosarjos@gmail.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit adf63026e1fa706de48d986ac6c1a7cb64cbf927)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
PMFW treats FeatureCtrlMask as the complete set of active OD features.
Uploads containing only newly edited bits can drop earlier settings.

Fix the OD state handling to:

- preserve committed features across later commits
- keep staged OD settings out of power-cap writes
- update cached state only after successful uploads
- clear restored features only after they return to defaults
- restore PPT for the current AC/DC source after resume

Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/3625
Signed-off-by: Kevin Wang &lt;kevin.wang@amd.com&gt;
Reviewed-by: Kenneth Feng &lt;kenneth.feng@amd.com&gt;
Tested-by: José Miguel Sarasola &lt;alosarjos@gmail.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit adf63026e1fa706de48d986ac6c1a7cb64cbf927)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
<entry>
<title>drm/amd/pm:restate all customized OD features on every table upload for smu 13.0.0/smu 14.0.2</title>
<updated>2026-10-20T16:09:44+00:00</updated>
<author>
<name>Kenneth Feng</name>
<email>kenneth.feng@amd.com</email>
</author>
<published>2026-10-16T07:17:43+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=44d0d3031e47973ed65fd327669d5e2eaaf6c565'/>
<id>44d0d3031e47973ed65fd327669d5e2eaaf6c565</id>
<content type='text'>
PMFW interprets OverDriveTable.FeatureCtrlMask as the complete set of
overdrive features that are currently overridden, not as a delta
describing the operation at hand. On import it resets every feature
whose bit is clear back to its default value - for the VF curve that
means zeroing VoltageOffsetPerZoneBoundary[] outright.

The driver assumed the opposite. smu_v14_0_2_od_edit_dpm_table() clears
the mask after a successful commit, so a subsequent above-default
power1_cap write goes through smu_v14_0_2_set_ppt_limit() and uploads a
table carrying only PP_OD_FEATURE_PPT_BIT. PMFW then drops the user's
voltage offset and reverts to the stock V/F curve, while the driver's
cached user table still reports the offset through pp_od_clk_voltage.

Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/5935
Signed-off-by: Kenneth Feng &lt;kenneth.feng@amd.com&gt;
Reviewed-by: Kevin Wang &lt;kevin.wang@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 1b63bb397db208a8f84778a9d9cafc93da9373eb)
Cc: stable@vger.kernel.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
PMFW interprets OverDriveTable.FeatureCtrlMask as the complete set of
overdrive features that are currently overridden, not as a delta
describing the operation at hand. On import it resets every feature
whose bit is clear back to its default value - for the VF curve that
means zeroing VoltageOffsetPerZoneBoundary[] outright.

The driver assumed the opposite. smu_v14_0_2_od_edit_dpm_table() clears
the mask after a successful commit, so a subsequent above-default
power1_cap write goes through smu_v14_0_2_set_ppt_limit() and uploads a
table carrying only PP_OD_FEATURE_PPT_BIT. PMFW then drops the user's
voltage offset and reverts to the stock V/F curve, while the driver's
cached user table still reports the offset through pp_od_clk_voltage.

Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/5935
Signed-off-by: Kenneth Feng &lt;kenneth.feng@amd.com&gt;
Reviewed-by: Kevin Wang &lt;kevin.wang@amd.com&gt;
Signed-off-by: Alex Deucher &lt;alexander.deucher@amd.com&gt;
(cherry picked from commit 1b63bb397db208a8f84778a9d9cafc93da9373eb)
Cc: stable@vger.kernel.org
</pre>
</div>
</content>
</entry>
</feed>
