Messages in this thread | | | Date | Wed, 10 Jun 2015 13:20:40 -0700 | Subject | Re: [PATCH] x86/asm/entry/32, selftests: Add test_syscall_vdso test | From | Kees Cook <> |
| |
On Wed, Jun 10, 2015 at 1:15 PM, Denys Vlasenko <dvlasenk@redhat.com> wrote: > On 06/10/2015 10:00 PM, Kees Cook wrote: >>> + printf("[SKIP]\tAT_SYSINFO not supplied, can't test\n"); >>> + exit(0); /* this is not a test failure */ >> >> Why is that not a test failure? It would mean it didn't actually test >> anything, which seems like a failure to me. > > Are you objecting to comment wording, or to exiting with 0? > > I exit with 0 because no bug was detected.
It seemed like a test failure to me: you're failing open ("couldn't configure test, I guess everything is okay") instead of failing closed ("couldn't configure test, something is terribly wrong").
If you can't locate how to make a syscall, then the test should fail, IMO, since it was not possible to perform the test, so you don't know if flags are being correctly handled across syscalls.
-Kees
-- Kees Cook Chrome OS Security
| |