Re: [PATCH] x86_64: another fix for canonical RIPs during signal handling

From: Andi Kleen
Date: Fri Jun 23 2006 - 08:32:41 EST



>
> that's not true. if the application expects to crash due to a bad
> signal handler then rip=0 may or may not achieve that, depending on
> what mapping exists at that address - this is inconsistent behaviour
> (from userland's point of view) created by the kernel itself, hence
> this is a kernel bug and should be fixed.

If it "wants" to crash it can just jump to 0 (or whatever unmapped address
it has) by itself. No need to involve the kernel here.

The only point of the patch was to not make the kernel/CPU crash due
to CPU bugs triggered by applications. But we really
don't care what happens to the application when it corrupts its stack frame.

-Andi
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/