[LTP] [PATCH] syscalls/clone3: New tests

[Cyril Hrubis]

Hi!
> I tried this but here is the problem ..
> 
> - pidfd: Kernel doesn't check this pointer and we get EFAULT.
> - child_tid: Kernel checks for valid pointer and doesn't use the
>   variable if NULL, so test passes unexpectedly.

We do have tst_get_bad_addr() that produces PROT_NONE mapped page that
should get you EFAULT in both cases here.

> - parent_tid: Kernel doesn't verify the pointer, but doesn't return
>   the error code properly and so test passes unexpectedly.

Huh? That sounds like a bug. Or does it skip only NULL silently? In that
case tst_get_bad_addr() should fix this case as well.

-- 
Cyril Hrubis
chrubis@suse.cz