| Message ID | 20240924115001.916112-1-Jason@zx2c4.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | selftests: vDSO: align stack for O2-optimized memcpy | expand |
diff --git a/tools/testing/selftests/vDSO/vdso_standalone_test_x86.c b/tools/testing/selftests/vDSO/vdso_standalone_test_x86.c index 27f6fdf11969..644915862af8 100644 --- a/tools/testing/selftests/vDSO/vdso_standalone_test_x86.c +++ b/tools/testing/selftests/vDSO/vdso_standalone_test_x86.c @@ -131,6 +131,8 @@ asm ( "_start:\n\t" #ifdef __x86_64__ "mov %rsp,%rdi\n\t" + "and $-16,%rsp\n\t" + "sub $8,%rsp\n\t" "jmp c_main" #else "push %esp\n\t"
When switching on -O2, gcc generates SSE2 instructions that assume a 16-byte aligned stack, which the standalone test's start point wasn't aligning. Fix this with the usual alignnent sequence. Fixes: ecb8bd70d51 ("selftests: vDSO: build tests with O2 optimization") Reported-by: kernel test robot <oliver.sang@intel.com> Closes: https://lore.kernel.org/oe-lkp/202409241558.98e13f6f-oliver.sang@intel.com Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com> --- tools/testing/selftests/vDSO/vdso_standalone_test_x86.c | 2 ++ 1 file changed, 2 insertions(+)