[PATCH 0/4] drm/amd: Check whether secure display TA loaded successfully

[PATCH 0/4] drm/amd: Check whether secure display TA loaded successfully

[Adrian Yip]

Hi everyone,

This is a patch series of backports from a upstream commit:

    c760bcda8357 ("drm/amd: Check whether secure display TA loaded successfully")

to the following stable kernel trees:
  * 6.17.y
  * 6.12.y
  * 6.6.y
  * 6.1.y

Each patch applied without conflicts. 

Compiling tests will be done for patches as I send them in.
I have not tested backports personally, but Shuah khan has Kindly offered
  to test them.

This is my first patch, please do let me know if there are any corrections
  or criticisms, I will take them to heart.

Thank you so much Shuah for providing this opportunity! I'm very grateful for it!


Respectfully,
Adrian Yip


Mario Limonciello (1):
  drm/amd: Check whether secure display TA loaded successfully

 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

-- 
2.51.0

[PATCH 1/4 6.17.y] drm/amd: Check whether secure display TA loaded successfully

[Adrian Yip]

From: Mario Limonciello <mario.limonciello@amd.com>

commit c760bcda83571e07b72c10d9da175db5051ed971 upstream

[Why]
Not all renoir hardware supports secure display.  If the TA is present
but the feature isn't supported it will fail to load or send commands.
This shows ERR messages to the user that make it seems like there is
a problem.

[How]
Check the resp_status of the context to see if there was an error
before trying to send any secure display commands.

There were no code conflict when applying to 6.17.y.
This backport gets rid of below error messages on AMD GPUs (per Shuah
Khan's machine)

  kern  :err   :
  amdgpu 0000:0b:00.0: amdgpu: Secure display: Generic Failure.
  amdgpu 0000:0b:00.0: amdgpu: SECUREDISPLAY: query securedisplay TA
    failed. ret 0x0

Compile test was done.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/1415
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit c760bcda83571e07b72c10d9da175db5051ed971)
Cc: <stable@vger.kernel.org>
Signed-off-by: Adrian Yip <adrian.ytw@gmail.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 693357caa9a8..d9d7fc4c33cb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -2350,7 +2350,7 @@ static int psp_securedisplay_initialize(struct psp_context *psp)
 	}
 
 	ret = psp_ta_load(psp, &psp->securedisplay_context.context);
-	if (!ret) {
+	if (!ret && !psp->securedisplay_context.context.resp_status) {
 		psp->securedisplay_context.context.initialized = true;
 		mutex_init(&psp->securedisplay_context.mutex);
 	} else
-- 
2.51.0

[PATCH 2/4 6.12.y] drm/amd: Check whether secure display TA loaded successfully

[Adrian Yip]

From: Mario Limonciello <mario.limonciello@amd.com>

commit c760bcda83571e07b72c10d9da175db5051ed971 upstream

[Why]
Not all renoir hardware supports secure display.  If the TA is present
but the feature isn't supported it will fail to load or send commands.
This shows ERR messages to the user that make it seems like there is
a problem.

[How]
Check the resp_status of the context to see if there was an error
before trying to send any secure display commands.

There were no code conflict when applying to 6.12.y.
This backport gets rid of below error messages on AMD GPUs (per Shuah
Khan's machine)

  kern  :err   :
  amdgpu 0000:0b:00.0: amdgpu: Secure display: Generic Failure.
  amdgpu 0000:0b:00.0: amdgpu: SECUREDISPLAY: query securedisplay TA
    failed. ret 0x0

Compile test was done.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/1415
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit c760bcda83571e07b72c10d9da175db5051ed971)
Cc: <stable@vger.kernel.org>
Signed-off-by: Adrian Yip <adrian.ytw@gmail.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 8553ac4c0ad3..a8358d1d1acb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -2171,7 +2171,7 @@ static int psp_securedisplay_initialize(struct psp_context *psp)
 	}
 
 	ret = psp_ta_load(psp, &psp->securedisplay_context.context);
-	if (!ret) {
+	if (!ret && !psp->securedisplay_context.context.resp_status) {
 		psp->securedisplay_context.context.initialized = true;
 		mutex_init(&psp->securedisplay_context.mutex);
 	} else
-- 
2.51.0

[PATCH 3/4 6.6.y] drm/amd: Check whether secure display TA loaded successfully

[Adrian Yip]

From: Mario Limonciello <mario.limonciello@amd.com>

commit c760bcda83571e07b72c10d9da175db5051ed971 upstream

[Why]
Not all renoir hardware supports secure display.  If the TA is present
but the feature isn't supported it will fail to load or send commands.
This shows ERR messages to the user that make it seems like there is
a problem.

[How]
Check the resp_status of the context to see if there was an error
before trying to send any secure display commands.

There were no code conflict when applying to 6.6.y.
This backport gets rid of below error messages on AMD GPUs (per Shuah
Khan's machine)

  kern  :err   :
  amdgpu 0000:0b:00.0: amdgpu: Secure display: Generic Failure.
  amdgpu 0000:0b:00.0: amdgpu: SECUREDISPLAY: query securedisplay TA
    failed. ret 0x0

Compile test was done.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/1415
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit c760bcda83571e07b72c10d9da175db5051ed971)
Cc: <stable@vger.kernel.org>
Signed-off-by: Adrian Yip <adrian.ytw@gmail.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index c83445c2e37f..d358a08b5e00 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -2012,7 +2012,7 @@ static int psp_securedisplay_initialize(struct psp_context *psp)
 	}
 
 	ret = psp_ta_load(psp, &psp->securedisplay_context.context);
-	if (!ret) {
+	if (!ret && !psp->securedisplay_context.context.resp_status) {
 		psp->securedisplay_context.context.initialized = true;
 		mutex_init(&psp->securedisplay_context.mutex);
 	} else
-- 
2.51.0

[PATCH 4/4 6.1.y] drm/amd: Check whether secure display TA loaded successfully

[Adrian Yip]

From: Mario Limonciello <mario.limonciello@amd.com>

commit c760bcda83571e07b72c10d9da175db5051ed971 upstream

[Why]
Not all renoir hardware supports secure display.  If the TA is present
but the feature isn't supported it will fail to load or send commands.
This shows ERR messages to the user that make it seems like there is
a problem.

[How]
Check the resp_status of the context to see if there was an error
before trying to send any secure display commands.

There were no code conflict when applying to 6.1.y.
This backport gets rid of below error messages on AMD GPUs (per Shuah
Khan's machine)

  kern  :err   :
  amdgpu 0000:0b:00.0: amdgpu: Secure display: Generic Failure.
  amdgpu 0000:0b:00.0: amdgpu: SECUREDISPLAY: query securedisplay TA
    failed. ret 0x0

Compile test was done.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/1415
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit c760bcda83571e07b72c10d9da175db5051ed971)
Cc: <stable@vger.kernel.org>
Signed-off-by: Adrian Yip <adrian.ytw@gmail.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 0bc21106d9e8..e22eaf9d450d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -1959,7 +1959,7 @@ static int psp_securedisplay_initialize(struct psp_context *psp)
 	}
 
 	ret = psp_ta_load(psp, &psp->securedisplay_context.context);
-	if (!ret) {
+	if (!ret && !psp->securedisplay_context.context.resp_status) {
 		psp->securedisplay_context.context.initialized = true;
 		mutex_init(&psp->securedisplay_context.mutex);
 	} else
-- 
2.51.0

Re: [PATCH 0/4] drm/amd: Check whether secure display TA loaded successfully

[Greg KH]

On Sat, Oct 18, 2025 at 11:56:40AM -0500, Adrian Yip wrote:
> Hi everyone,
> 
> This is a patch series of backports from a upstream commit:
> 
>     c760bcda8357 ("drm/amd: Check whether secure display TA loaded successfully")
> 
> to the following stable kernel trees:
>   * 6.17.y
>   * 6.12.y
>   * 6.6.y
>   * 6.1.y
> 
> Each patch applied without conflicts. 
> 
> Compiling tests will be done for patches as I send them in.
> I have not tested backports personally, but Shuah khan has Kindly offered
>   to test them.
> 
> This is my first patch, please do let me know if there are any corrections
>   or criticisms, I will take them to heart.

Meta-comment, do not send a "patch series" that can not all be applied
to the same tree.  When applying a series of patches (1-4 as sent here),
that implies that all 4 go to the same place.  That's not the case here,
so I need to pick out the patches individually some how, and apply patch
1 to one tree, 2 to another, and so on.  It looks a bit odd to apply
patch 4/4 to only one tree, right?

I've picked these apart this time, but next time, just send 4 patches
separately, not as a series, as you can see others do on the stable
list, and all will be fine.

thanks,

greg k-h

Re: [PATCH 2/4 6.12.y] drm/amd: Check whether secure display TA loaded successfully

[Greg KH]

On Sat, Oct 18, 2025 at 11:56:42AM -0500, Adrian Yip wrote:
> From: Mario Limonciello <mario.limonciello@amd.com>
> 
> commit c760bcda83571e07b72c10d9da175db5051ed971 upstream
> 
> [Why]
> Not all renoir hardware supports secure display.  If the TA is present
> but the feature isn't supported it will fail to load or send commands.
> This shows ERR messages to the user that make it seems like there is
> a problem.
> 
> [How]
> Check the resp_status of the context to see if there was an error
> before trying to send any secure display commands.
> 
> There were no code conflict when applying to 6.12.y.
> This backport gets rid of below error messages on AMD GPUs (per Shuah
> Khan's machine)
> 
>   kern  :err   :
>   amdgpu 0000:0b:00.0: amdgpu: Secure display: Generic Failure.
>   amdgpu 0000:0b:00.0: amdgpu: SECUREDISPLAY: query securedisplay TA
>     failed. ret 0x0
> 
> Compile test was done.

Also, you added text here in the changelog, which isn't generally good
as we want to keep the changelog text as close as possible, if not
identical, to the original so that it makes it easier to track patches
across branches.

thanks,

greg k-h

Re: [PATCH 0/4] drm/amd: Check whether secure display TA loaded successfully

[Adrian Yip]

On 10/20/2025 4:06 AM, Greg KH wrote:
> On Sat, Oct 18, 2025 at 11:56:40AM -0500, Adrian Yip wrote:
>> Hi everyone,
>>
>> This is a patch series of backports from a upstream commit:
>>
>>      c760bcda8357 ("drm/amd: Check whether secure display TA loaded successfully")
>>
>> to the following stable kernel trees:
>>    * 6.17.y
>>    * 6.12.y
>>    * 6.6.y
>>    * 6.1.y
>>
>> Each patch applied without conflicts.
>>
>> Compiling tests will be done for patches as I send them in.
>> I have not tested backports personally, but Shuah khan has Kindly offered
>>    to test them.
>>
>> This is my first patch, please do let me know if there are any corrections
>>    or criticisms, I will take them to heart.
> 
> Meta-comment, do not send a "patch series" that can not all be applied
> to the same tree.  When applying a series of patches (1-4 as sent here),
> that implies that all 4 go to the same place.  That's not the case here,
> so I need to pick out the patches individually some how, and apply patch
> 1 to one tree, 2 to another, and so on.  It looks a bit odd to apply
> patch 4/4 to only one tree, right?
> 
> I've picked these apart this time, but next time, just send 4 patches
> separately, not as a series, as you can see others do on the stable
> list, and all will be fine.
> 
> thanks,
> 
> greg k-h

Thank you Greg, I appreciate your time and comments. I will be applying 
your feedback to future patches, and the explanation makes perfect 
sense. Sorry for the inconvenience.

There may also be some confusion with me saying Shuah offering to test 
them. For some context, Shuah told me about this fix after she noticed 
it on her machine and gave me the opportunity to make the patch. The 
offer to test them was given to me along with it, but I don't believe 
the tests were done on each of the stable trees yet.

I'm writing to make sure these patches are added responsibly.
Feel free to ignore if this isn't a problem. Thank you again.

Respectfully,

Adrian Yip