[Bug middle-end/52306] ICE in CSE

2012-12-23 Thread tg at mirbsd dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52306 --- Comment #11 from Thorsten Glaser 2012-12-24 00:53:01 UTC --- Created attachment 29040 --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=29040 preprocessed source of festival occurrence Here's another one, from festival (some TTS engine th

[Bug middle-end/52306] ICE in CSE

2012-12-06 Thread jakub at gcc dot gnu.org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52306 --- Comment #10 from Jakub Jelinek 2012-12-06 09:52:46 UTC --- cselib.c ICEs if it sees an invalid insn, i.e. one that modifies the same register or memory more than once. So it is a bug in whatever pass created that insn.

[Bug middle-end/52306] ICE in CSE

2012-12-06 Thread mikpe at it dot uu.se
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52306 --- Comment #9 from Mikael Pettersson 2012-12-06 09:50:12 UTC --- The ICE with gcc 4.8 on the second kde4libs test case started with Richard Biener's "Add gimple load/store predicates, use them from stmt estimates" patch in r192987: http:

[Bug middle-end/52306] ICE in CSE

2012-12-04 Thread mikpe at it dot uu.se
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52306 --- Comment #8 from Mikael Pettersson 2012-12-04 22:05:19 UTC --- The new test case from kde4libs ICEs g++ 4.8-20121202 and 4.6-20121130, but not 4.7-20121201, when targeting x86-linux. The 4.8 ICE looks as follows: ../../kdeui/notifica

[Bug middle-end/52306] ICE in CSE

2012-12-04 Thread tg at mirbsd dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52306 --- Comment #7 from Thorsten Glaser 2012-12-04 12:58:35 UTC --- Created attachment 28876 --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=28876 preprocessed source of second kde4libs occurrence OK. I found another one in the meantime, which

[Bug middle-end/52306] ICE in CSE

2012-12-04 Thread mikpe at it dot uu.se
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52306 --- Comment #6 from Mikael Pettersson 2012-12-04 10:27:34 UTC --- (In reply to comment #5) > Also happens in src:kde4libs (= 4:4.8.4-4) on > ../../kdecore/util/kpluginfactory.cpp – I tested -O1, which helped. Just to clarify, compiling with -O1

[Bug middle-end/52306] ICE in CSE

2012-12-04 Thread tg at mirbsd dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52306 --- Comment #5 from Thorsten Glaser 2012-12-04 09:26:12 UTC --- Also happens in src:kde4libs (= 4:4.8.4-4) on ../../kdecore/util/kpluginfactory.cpp – I tested -O1, which helped. If there’s any need of _more_ preprocessed source… just shout, but

[Bug middle-end/52306] ICE in CSE

2012-05-06 Thread tg at mirbsd dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52306 Thorsten Glaser changed: What|Removed |Added Known to fail||4.6.3 --- Comment #4 from Thorsten Glas

[Bug middle-end/52306] ICE in CSE

2012-05-06 Thread tg at mirbsd dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52306 --- Comment #3 from Thorsten Glaser 2012-05-06 14:16:07 UTC --- Created attachment 27325 --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=27325 preprocessed source of libxslt occurrence The same thing happens in libxslt-1.1.26 but even with P

[Bug middle-end/52306] ICE in CSE

2012-03-25 Thread aoliva at gcc dot gnu.org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52306 Alexandre Oliva changed: What|Removed |Added Status|UNCONFIRMED |NEW Last reconfirmed|

[Bug middle-end/52306] ICE in CSE

2012-03-25 Thread aoliva at gcc dot gnu.org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52306 --- Comment #1 from Alexandre Oliva 2012-03-25 12:00:34 UTC --- The bug, duplicated by compiling attachment 26150, from bug 43437 comment 16, with a cross-compiler to m68k-elf with -O -c, exposes a reload problem. Before reload, we have this ins