public inbox for gcc-bugs@sourceware.org help / color / mirror / Atom feed
From: "rguenth at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org> To: gcc-bugs@gcc.gnu.org Subject: [Bug tree-optimization/57359] store motion causes wrong code for union access at -O3 Date: Mon, 27 Apr 2020 09:35:48 +0000 [thread overview] Message-ID: <bug-57359-4-8mvUaEfxpI@http.gcc.gnu.org/bugzilla/> (raw) In-Reply-To: <bug-57359-4@http.gcc.gnu.org/bugzilla/> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=57359 --- Comment #29 from Richard Biener <rguenth at gcc dot gnu.org> --- And another testcase showing that with a conditional invariant store we may _never_ apply store-motion since conditional means we do not know the original order of stores in the loop. *sigh* typedef int A; typedef float B; float x[256]; void __attribute__((noinline,noclone)) foo(long unk, long ack) { B *q = x; for (long i = 0; i < unk; ++i) { q[i] = 42; if (ack & i) *((A*)q + 4) = 1; } } int main(void) { foo(5, 3); if (x[4] != 42) __builtin_abort(); return 0; } implementation-wise it should still work, in hoist_memory_references, when we computed the whole set of refs to apply store-motion to in a loop, track the sequencing of stores from entry to the latch edge and at merge points reject (parts of?) the sequence when there are mismatches in ordering. For the above there'd be { q[i], *((A*)q + 4) } and { q[i] } and thus a mismatch. If we reject only parts of the sequence those parts would need to be disambiguated against the previous stores in the sequence w/o TBAA. Trying to code that up now.
next prev parent reply other threads:[~2020-04-27 9:35 UTC|newest] Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top 2013-05-21 18:20 [Bug rtl-optimization/57359] New: wrong code for union access at -O3 on x86_64-linux dhazeghi at yahoo dot com 2013-05-21 18:26 ` [Bug rtl-optimization/57359] " pinskia at gcc dot gnu.org 2013-05-21 18:33 ` jakub at gcc dot gnu.org 2013-05-23 19:51 ` dhazeghi at yahoo dot com 2013-05-29 21:14 ` dhazeghi at yahoo dot com 2013-05-30 7:13 ` jakub at gcc dot gnu.org 2013-05-31 8:48 ` rguenth at gcc dot gnu.org 2013-05-31 16:32 ` dhazeghi at yahoo dot com 2013-05-31 21:07 ` joseph at codesourcery dot com 2013-06-03 19:15 ` dhazeghi at yahoo dot com 2013-06-04 12:37 ` [Bug tree-optimization/57359] " rguenth at gcc dot gnu.org 2020-04-20 7:15 ` [Bug tree-optimization/57359] store motion causes wrong code for union access at -O3 rguenth at gcc dot gnu.org 2020-04-20 8:57 ` rguenth at gcc dot gnu.org 2020-04-20 10:50 ` rguenth at gcc dot gnu.org 2020-04-20 12:03 ` rguenth at gcc dot gnu.org 2020-04-20 12:34 ` pascal_cuoq at hotmail dot com 2020-04-20 12:39 ` rguenther at suse dot de 2020-04-21 7:22 ` rguenth at gcc dot gnu.org 2020-04-27 9:17 ` rguenth at gcc dot gnu.org 2020-04-27 9:35 ` rguenth at gcc dot gnu.org [this message] 2020-04-27 12:48 ` rguenth at gcc dot gnu.org 2020-04-30 12:25 ` rguenth at gcc dot gnu.org 2020-05-07 10:12 ` cvs-commit at gcc dot gnu.org 2020-05-07 10:13 ` rguenth at gcc dot gnu.org 2020-05-11 12:56 ` rguenth at gcc dot gnu.org 2020-05-11 12:56 ` rguenth at gcc dot gnu.org 2020-05-11 14:53 ` rguenth at gcc dot gnu.org 2020-05-11 14:57 ` rguenth at gcc dot gnu.org 2020-05-12 6:26 ` rguenth at gcc dot gnu.org 2020-05-12 11:53 ` ro at gcc dot gnu.org 2020-05-12 12:13 ` rguenth at gcc dot gnu.org 2020-05-18 9:50 ` rguenth at gcc dot gnu.org 2020-06-11 15:55 ` pinskia at gcc dot gnu.org
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=bug-57359-4-8mvUaEfxpI@http.gcc.gnu.org/bugzilla/ \ --to=gcc-bugzilla@gcc.gnu.org \ --cc=gcc-bugs@gcc.gnu.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).