subject:"\[Bug rtl\-optimization\/58365\] \[4.7\/4.8\/4.9 Regression\] likely wrong code bug"

[Bug rtl-optimization/58365] [4.7/4.8/4.9 Regression] likely wrong code bug

2013-09-10 Thread jakub at gcc dot gnu.org

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58365 --- Comment #10 from Jakub Jelinek --- Author: jakub Date: Tue Sep 10 11:48:30 2013 New Revision: 202435 URL: http://gcc.gnu.org/viewcvs?rev=202435&root=gcc&view=rev Log: PR rtl-optimization/58365 * cfgcleanup.c (merge_memattrs): Also cle

[Bug rtl-optimization/58365] [4.7/4.8/4.9 Regression] likely wrong code bug

2013-09-10 Thread jakub at gcc dot gnu.org

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58365 --- Comment #9 from Jakub Jelinek --- Author: jakub Date: Tue Sep 10 11:47:19 2013 New Revision: 202434 URL: http://gcc.gnu.org/viewcvs?rev=202434&root=gcc&view=rev Log: PR rtl-optimization/58365 * cfgcleanup.c (merge_memattrs): Also clea

[Bug rtl-optimization/58365] [4.7/4.8/4.9 Regression] likely wrong code bug

2013-09-10 Thread jakub at gcc dot gnu.org

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58365 --- Comment #8 from Jakub Jelinek --- Created attachment 30781 --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=30781&action=edit gcc49-pr58365.patch Untested fix.

[Bug rtl-optimization/58365] [4.7/4.8/4.9 Regression] likely wrong code bug

2013-09-10 Thread mkuvyrkov at gcc dot gnu.org

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58365 --- Comment #6 from Maxim Kuvyrkov --- (In reply to Jakub Jelinek from comment #5) > That is actually the crossjumping pass after csa (jump2; though yes, > previously it has been part of csa). In any case, I don't see anything > wrong with the cr

[Bug rtl-optimization/58365] [4.7/4.8/4.9 Regression] likely wrong code bug

2013-09-10 Thread jakub at gcc dot gnu.org

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58365 Jakub Jelinek changed: What|Removed |Added Assignee|mkuvyrkov at gcc dot gnu.org |jakub at gcc dot gnu.org --- Comme

[Bug rtl-optimization/58365] [4.7/4.8/4.9 Regression] likely wrong code bug

2013-09-09 Thread jakub at gcc dot gnu.org

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58365 --- Comment #5 from Jakub Jelinek --- That is actually the crossjumping pass after csa (jump2; though yes, previously it has been part of csa). In any case, I don't see anything wrong with the cross jumping, it turns the two: (set (me

[Bug rtl-optimization/58365] [4.7/4.8/4.9 Regression] likely wrong code bug

2013-09-09 Thread mkuvyrkov at gcc dot gnu.org

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58365 Maxim Kuvyrkov changed: What|Removed |Added CC||rth at gcc dot gnu.org --- Comment #4 fr

[Bug rtl-optimization/58365] [4.7/4.8/4.9 Regression] likely wrong code bug

2013-09-09 Thread mkuvyrkov at gcc dot gnu.org

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58365 --- Comment #3 from Maxim Kuvyrkov --- Bug introduced in revision 162592. Investigating. commit 48d5e32fb40bb7973e596b36d12ad4041d8ef10a Author: mkuvyrkov Date: Tue Jul 27 19:38:10 2010 + PR target/42495 PR middle-end/425