[PATCH] lib/amd: add memleak functions

[vitaly prosyak]

On 2025-02-27 05:27, Peter Senna Tschudin wrote:
> Hello again Vitaly,
>
> [...]
>
>> Hi Peter,
>>
>> Thanks for the update! I really appreciate the effort you’ve put into making the approach more global and vendor-agnostic—great work!
>>
>> It would be incredibly useful to have both --kmemleak options available (for the entire test list and individual tests).
>>
>> Could you kindly provide guidance on how to add these options (-konce or -keach) to igt-runner? For example, how should I modify the following commands to include them?
>>
>> sudo ./scripts/run-tests.sh -t -v /home/vprosyak/src/igt-gpu-tools/build/tests/amd/amd_basic  
> I’ve sent another revision of the patch that updates scripts/run-tests.sh,
> adding the -K option. This allows you to run tests like this:
>
> sudo ./scripts/run-tests.sh -K each -t -v /home/vprosyak/src/igt-gpu-tools/build/tests/amd/amd_basic
>
>> or
>>
>> sudo ./scripts/run-tests.sh -v -T /home/infra/igt/custom.testlist  
> Same here:
>
> sudo ./scripts/run-tests.sh -v -K each -T /home/infra/igt/custom.testlist
>
> You can find the patch here: https://patchwork.freedesktop.org/series/143996/
Hi Peter,

Feel free to use:

Reviewed-by: Vitaly Prosyak <vitaly.prosyak at amd.com>

for your three patches. We truly appreciate your work, and we're eagerly awaiting your merge!

The key point is that it shouldn't introduce any regressions since the -keach or -konce options will be disabled by default. Once we enable it and put CI under pressure, we might uncover some surprises—potentially false-positive leaks. If that happens, we can try resolving them by issuing the scan and clear commands twice.

Let’s see how it plays out! Thanks again for your effort.

Thanks, Vitaly

>
> Would you be open to reviewing the new patch and sending your Reviewed-by
> once you think it’s ready? That would help speed up the merge process.
>
> Let me know if I can assist in any way.
>
> Cheers,
>
> Peter