On Tue, Feb 13, 2018 at 12:21:55PM +0100, Jakub Jelinek wrote:
> On Tue, Feb 13, 2018 at 12:15:36PM +0100, Paolo Bonzini wrote:
> > On 13/02/2018 10:32, Martin Liška wrote:
> > > Hello.
> > >
> > > It caused PR84340, I'm suggesting following fix:
> > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84340#c3
> >
> > I don't think EAF_DIRECT is the issue. You could think of ASAN_MARK as
> > writing a global variable, which it can do because it's not const.
> >
> > The issue is that the ASAN_CHECK doesn't exist at early DSE time, and
> > thus causes the store to disappear.
>
> That doesn't make sense to me, because the testcases regressed with the change
> of "fn spec" attribute on ASAN_{CHECK,MARK}.
> If it was DSE removing the stores before asan pass, then it would FAIL
> before as well.
Sorry, while ASAN_CHECK is introduced late, ASAN_MARK is present there
already from the gimplification.
For use after scope, I guess a lot of the stores after end of scope
(rather than reads) are something DSE could consider removing.
So, shall we just disable DSE on vars where their address "escapes"
through ASAN_MARK when -fsanitize-address-use-after-scope?
Generally, dead stores could be eliminable when stored before the
corresponding ASAN_MARK poison (but even ASAN_MARK with "..W.." will
prevent those) and uneliminable when stored after ASAN_MARK poison.
For the "fn spec" for now, I'd just go with "..R.." for ASAN_CHECK and
NULL for ASAN_MARK for now.
Jakub
