[Bug tree-optimization/90020] [7/8/9 regression] -O2 -Os x86-64 wrong code generated for GNU Emacs

2019-04-11 Thread rguenth at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90020 --- Comment #12 from Richard Biener --- Author: rguenth Date: Thu Apr 11 07:34:20 2019 New Revision: 270275 URL: https://gcc.gnu.org/viewcvs?rev=270275&root=gcc&view=rev Log: 2019-04-11 Richard Biener PR tree-optimization/90020

[Bug tree-optimization/90020] [7/8/9 regression] -O2 -Os x86-64 wrong code generated for GNU Emacs

2019-04-09 Thread rguenth at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90020 --- Comment #11 from Richard Biener --- For the RTL issue there's compute_hash_table_work (struct gcse_hash_table_d *table) { ... /* First pass over the instructions records information used to determine when registers and memory

[Bug tree-optimization/90020] [7/8/9 regression] -O2 -Os x86-64 wrong code generated for GNU Emacs

2019-04-09 Thread rguenth at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90020 --- Comment #10 from Richard Biener --- (In reply to Martin Liška from comment #6) > I bisected GCC 4.9.x branch and it started with r215059, which is a backport > of 3 patches. I reverted changes in: > patching file gcc/recog.c > patching file g

[Bug tree-optimization/90020] [7/8/9 regression] -O2 -Os x86-64 wrong code generated for GNU Emacs

2019-04-09 Thread rguenth at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90020 --- Comment #9 from Richard Biener --- /* { dg-do run } */ /* { dg-require-weak "" } */ void __attribute__((noinline,noclone)) check (int i) { if (i == 0) __builtin_exit (0); } int i; extern int x __attribute__((weak)); int main(int argc

[Bug tree-optimization/90020] [7/8/9 regression] -O2 -Os x86-64 wrong code generated for GNU Emacs

2019-04-09 Thread rguenth at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90020 Richard Biener changed: What|Removed |Added Status|NEW |ASSIGNED Assignee|unassigned

[Bug tree-optimization/90020] [7/8/9 regression] -O2 -Os x86-64 wrong code generated for GNU Emacs

2019-04-09 Thread rguenth at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90020 --- Comment #7 from Richard Biener --- So looking at one issue I can see is code-hoisting hoisting MEM[(struct window *)window_6(D) + -5B].contents across a call that might not return. This can only happen for calls we can alias-disambiguate aga

[Bug tree-optimization/90020] [7/8/9 regression] -O2 -Os x86-64 wrong code generated for GNU Emacs

2019-04-09 Thread rguenth at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90020 Richard Biener changed: What|Removed |Added Priority|P3 |P2 Target Milestone|---

[Bug tree-optimization/90020] [7/8/9 regression] -O2 -Os x86-64 wrong code generated for GNU Emacs

2019-04-09 Thread marxin at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90020 --- Comment #6 from Martin Liška --- I bisected GCC 4.9.x branch and it started with r215059, which is a backport of 3 patches. I reverted changes in: patching file gcc/recog.c patching file gcc/tree-ssa-loop-niter.c patching file gcc/tree-vect-s

[Bug tree-optimization/90020] [7/8/9 regression] -O2 -Os x86-64 wrong code generated for GNU Emacs

2019-04-09 Thread marxin at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90020 Martin Liška changed: What|Removed |Added Status|ASSIGNED|NEW

[Bug tree-optimization/90020] [7/8/9 regression] -O2 -Os x86-64 wrong code generated for GNU Emacs

2019-04-09 Thread marxin at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90020 --- Comment #5 from Martin Liška --- Created attachment 46109 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=46109&action=edit Reduced test-case #1

[Bug tree-optimization/90020] [7/8/9 regression] -O2 -Os x86-64 wrong code generated for GNU Emacs

2019-04-09 Thread marxin at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90020 Martin Liška changed: What|Removed |Added Keywords||wrong-code CC|

[Bug tree-optimization/90020] [7/8/9 regression] -O2 -Os x86-64 wrong code generated for GNU Emacs

2019-04-09 Thread marxin at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90020 --- Comment #4 from Martin Liška --- Created attachment 46108 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=46108&action=edit Reduced test-case #0