This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Move Asan instrumentation to sanopt pass
- From: Yuri Gribov <tetra2005 at gmail dot com>
- To: Jakub Jelinek <jakub at redhat dot com>
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>, Konstantin Serebryany <kcc at gcc dot gnu dot org>, Dmitry Vyukov <dvyukov at google dot com>, Viacheslav Garbuzov <v dot garbuzov at samsung dot com>, Marek Polacek <polacek at redhat dot com>
- Date: Sat, 19 Jul 2014 21:56:34 +0400
- Subject: Re: [PATCH] Move Asan instrumentation to sanopt pass
- Authentication-results: sourceware.org; auth=none
- References: <CAJOtW+6o6bFKDef5X81qL-Y2dw5fHT5Zoi-eYmsnveeoBd2jJA at mail dot gmail dot com> <20140718172323 dot GL3003 at laptop dot redhat dot com> <CAJOtW+5y5ziz4jQMbaOjUhAU1AyouKwDss4zN_f0dqU1QGLeMg at mail dot gmail dot com> <20140719093112 dot GO3003 at laptop dot redhat dot com>
On Sat, Jul 19, 2014 at 1:31 PM, Jakub Jelinek <jakub@redhat.com> wrote:
> On Fri, Jul 18, 2014 at 10:21:16PM +0400, Yuri Gribov wrote:
> Sounds like a bug, most likely in the lim pass. If it for some reason
> decides to load from possibly uninitialized memory before the loop, it
> better should make sure TREE_NO_WARNING is set. Can you please
> file a PR for this, with preprocessed asan.c and minimal gcc options to
> reproduce it?
Sure, I'll also check if I can submit a patch for this.
> As for the workaround, my strong preference would be to do something like
> following incrmeental patch instead, just use different gsi in the two
> loops.
Yeah, I also used this as a WAR originally.
-Y