Re: [PATCH 0/2] selftests/x86: AMX-related test improvements

[Shuah Khan]

On 4/1/22 4:10 PM, Chang S. Bae wrote:
> A couple of test updates are included:
>
> * With this option [1,2], the kernel's altstack check becomes stringent.
>    The x86 sigaltstack test is ignorant about this. Adjust the test now.
>    This check was established [3] to ensure every AMX task's altstack is
>    sufficient (regardless of that option) [4].
>
> * The AMX test wrongly fails on non-AMX machines. Fix the code to skip the
>    test instead.
>
> The series is available in this repository:
>    git://github.com/intel/amx-linux.git selftest
>
> [1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/x86/Kconfig#n2432
> [2] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/admin-guide/kernel-parameters.txt#n5676
> [3] 3aac3ebea08f ("x86/signal: Implement sigaltstack size validation")
> [4] 4b7ca609a33d ("x86/signal: Use fpu::__state_user_size for sigalt stack validation")
>
> Chang S. Bae (2):
>    selftests/x86/signal: Adjust the test to the kernel's altstack check
>    selftests/x86/amx: Fix the test to avoid failure when AMX is
>      unavailable
>
>   tools/testing/selftests/x86/amx.c         | 42 +++++++++++++++++------
>   tools/testing/selftests/x86/sigaltstack.c | 12 ++++++-
>   2 files changed, 42 insertions(+), 12 deletions(-)
>
>
> base-commit: f443e374ae131c168a065ea1748feac6b2e76613

These are fixes and I can take them through kselftest once the review
comments are addressed and get an ack from x86 maintainers.

thanks,
-- Shuah