On Wed, Nov 09, 2011 at 10:53:34AM -0700, Jeff Law wrote:
> > which is only different on undefined paths. But I'm not sure that
> > more complicated cases, where there are other instructions between
> > the "if" and "*p = 0", would also be allowed by the C standard.
> > For example, I think a function call in the "else" branch, or
> > between the PHI and the dereference should prevent the
> > optimization, because the function call might never return for what
> > we know. Probably a volatile asm too. Does your patch do that?
> > (Testcases! :)).
> My patch totally ignores the other code on the unexecutable path. So
> we can miss externally visible side effects, if we were to somehow get
> on the unexecutable path. But that's the whole point, in a conforming
> program we can't ever get on the unexecutable path.
Well, side-effects that could be validly scheduled after the NULL pointer
dereference are just fine to be omitted. But if there is a function call
that might call exit (0); in it, or throw an exception and thus never
reach the trap point, then omitting those would be an invalid optimization.
Jakub
