On Thu, May 29, 2025 at 01:58:20PM -0700, Sean Christopherson wrote:
> Skip the test if its target kernel/test file isn't available so that
> skipping a test that isn't supported for a given config doesn't require
> manually flagging the testcase in unittests.cfg. This fixes "failures"
> on x86 with CONFIG_EFI=y due to some tests not being built for EFI, but
> not being annotated in x86/unittests.cfg.
>
> Alternatively, testcases could be marked noefi (or efi-only), but that'd
> require more manual effort, and there's no obvious advantage to doing so.
>
> Signed-off-by: Sean Christopherson <seanjc@xxxxxxxxxx>
> ---
> scripts/runtime.bash | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/scripts/runtime.bash b/scripts/runtime.bash
> index ee229631..a94d940d 100644
> --- a/scripts/runtime.bash
> +++ b/scripts/runtime.bash
> @@ -150,6 +150,11 @@ function run()
> done
> fi
>
> + if [ ! -f "$kernel" ]; then
> + print_result "SKIP" $testname "" "Test file '$kernel' not found";
> + return 2;
> + fi
> +
I see mkstandalone.sh already has something like this. There's still one
other place, though, which is print_testname(). Should we filter tests
from the listing that are missing their kernels?
Thanks,
drew
> log=$(premature_failure) && {
> skip=true
> if [ "${CONFIG_EFI}" == "y" ]; then
>
> base-commit: 72d110d8286baf1b355301cc8c8bdb42be2663fb
> --
> 2.49.0.1204.g71687c7c1d-goog
>
