[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 Bill Schmidt changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #17 from Bill Schmidt --- GCC 5 has closed, so this can also.
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 --- Comment #16 from Bill Schmidt --- Ping...
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 Bill Schmidt changed: What|Removed |Added Status|UNCONFIRMED |NEW Last reconfirmed||2015-12-15 CC||wschmidt at gcc dot gnu.org Ever confirmed|0 |1 --- Comment #15 from Bill Schmidt --- Obviously confirmed at this point. Vlad, do you plan to backport this to GCC 5? We should get this closed if this is fixed.
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 Segher Boessenkool changed: What|Removed |Added CC||segher at gcc dot gnu.org --- Comment #14 from Segher Boessenkool --- Does this need a backport to GCC 5?
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 --- Comment #13 from Vladimir Makarov vmakarov at gcc dot gnu.org --- Author: vmakarov Date: Thu Mar 19 19:59:38 2015 New Revision: 221522 URL: https://gcc.gnu.org/viewcvs?rev=221522root=gccview=rev Log: 2015-03-19 Vladimir Makarov vmaka...@redhat.com PR rtl-optimization/63491 * lra-constraints.c (check_and_process_move): Use src instead of sreg. Remove some dead code. 2015-03-19 Vladimir Makarov vmaka...@redhat.com PR rtl-optimization/63491 * gcc.target/powerpc/pr63491.c: New. Added: trunk/gcc/testsuite/gcc.target/powerpc/pr63491.c Modified: trunk/gcc/lra-constraints.c trunk/gcc/testsuite/ChangeLog
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 --- Comment #12 from Vladimir Makarov vmakarov at gcc dot gnu.org --- (In reply to Peter Bergner from comment #11) Ok, I found the difference. The problem on the external gcc112 system (ie, the big-endian system) is that the system binutils doesn't support POWER8, so during gcc config/build, the power8 gcc support was disabled. I built/installed a new binutils and used that for my build and now my external build also ICEs. Vlad can you try using my build on gcc112? /home/bergner/gcc/build/gcc-fsf-mainline-r221324-2/gcc/xgcc -B/home/bergner/gcc/build/gcc-fsf-mainline-r221324-2/gcc -O1 -m64 -mcpu=power8 -mlra -S pr63491.i It uses /home/bergner/binutils/install/binutils-power8/bin/{as,ld}. Yes, I reproduced it with you binaries. Thanks, Peter. The crash is in a code concerning secondary reloads. It might take a few days to fix it.
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 --- Comment #11 from Peter Bergner bergner at gcc dot gnu.org --- Ok, I found the difference. The problem on the external gcc112 system (ie, the big-endian system) is that the system binutils doesn't support POWER8, so during gcc config/build, the power8 gcc support was disabled. I built/installed a new binutils and used that for my build and now my external build also ICEs. Vlad can you try using my build on gcc112? /home/bergner/gcc/build/gcc-fsf-mainline-r221324-2/gcc/xgcc -B/home/bergner/gcc/build/gcc-fsf-mainline-r221324-2/gcc -O1 -m64 -mcpu=power8 -mlra -S pr63491.i It uses /home/bergner/binutils/install/binutils-power8/bin/{as,ld}.
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 --- Comment #10 from Peter Bergner bergner at gcc dot gnu.org --- Strange, this fails with my builds on our internal POWER8 box, but doesn't with my build on the external gcc112 POWER8 compile farm system. I'll try and track down what the difference is and see if I can recreate it on the external system.
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 --- Comment #9 from Peter Bergner bergner at gcc dot gnu.org --- (In reply to Peter Bergner from comment #8) (In reply to Vladimir Makarov from comment #7) I tried again the test on gcc rev. 221324 (Mar 10) with the mentioned options and I've failed to reproduce the crash. Maybe a configure option thing? Maybe --disable-bootstrap which I have been using? I just tried using the same revision you did (221324) and I still see the same error: Well I did a normal bootstrap build and also a --enable-checking=release build and they both ICE with the test case, so I'm not sure how you can't be seeing this.
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 --- Comment #7 from Vladimir Makarov vmakarov at gcc dot gnu.org --- (In reply to Peter Bergner from comment #6) (In reply to Vladimir Makarov from comment #5) Sorry, I can not reproduce the bug on the today trunk. Probably it was fixed by numerous changes in LRA since Oct. This still fails for me today on my big endian system using revision 220792. Did you use the exact same compiler options I used? [bergner@makalu-lp1 LRA]$ cat pr63491.i typedef __int128_t __attribute__((__vector_size__(16))) vector_128_t; typedef unsigned long long scalar_64_t; vector_128_t foo (void) { union { scalar_64_t i64[2]; vector_128_t v128; } u; u.i64[0] = 1; u.i64[1] = 2; return u.v128; } [bergner@makalu-lp1 LRA]$ /home/bergner/gcc/build/gcc-fsf-mainline-pr63491/gcc/xgcc -B/home/bergner/gcc/build/gcc-fsf-mainline-pr63491/gcc -O1 -m64 -mcpu=power8 -mlra -S pr63491.i pr63491.i: In function ‘foo’: pr63491.i:14:1: internal compiler error: in check_and_process_move, at lra-constraints.c:1170 } ^ 0x10994a7b check_and_process_move /home/bergner/gcc/gcc-fsf-mainline-pr63491/gcc/lra-constraints.c:1167 0x1099d45b curr_insn_transform /home/bergner/gcc/gcc-fsf-mainline-pr63491/gcc/lra-constraints.c:3463 0x109a1cdb lra_constraints(bool) /home/bergner/gcc/gcc-fsf-mainline-pr63491/gcc/lra-constraints.c:4451 0x1098221b lra(_IO_FILE*) /home/bergner/gcc/gcc-fsf-mainline-pr63491/gcc/lra.c:2292 0x109050d7 do_reload /home/bergner/gcc/gcc-fsf-mainline-pr63491/gcc/ira.c:5418 0x1090576f execute /home/bergner/gcc/gcc-fsf-mainline-pr63491/gcc/ira.c:5589 Please submit a full bug report, with preprocessed source if appropriate. I tried again the test on gcc rev. 221324 (Mar 10) with the mentioned options and I've failed to reproduce the crash.
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 --- Comment #8 from Peter Bergner bergner at gcc dot gnu.org --- (In reply to Vladimir Makarov from comment #7) I tried again the test on gcc rev. 221324 (Mar 10) with the mentioned options and I've failed to reproduce the crash. Maybe a configure option thing? Maybe --disable-bootstrap which I have been using? I just tried using the same revision you did (221324) and I still see the same error: Target: powerpc64-linux Configured with: /home/bergner/gcc/gcc-fsf-mainline-vlad-lra/configure --target=powerpc64-linux --host=powerpc64-linux --build=powerpc64-linux --enable-threads=posix --enable-shared --enable-__cxa_atexit --enable-languages=c,c++ --enable-secureplt --enable-checking=yes --with-gmp=/home/bergner/tools/meissner --with-mpfr=/home/bergner/tools/meissner --with-mpc=/home/bergner/tools/meissner --with-long-double-128 --enable-decimal-float --disable-bootstrap --disable-alsa --prefix=/home/bergner/gcc/install/gcc-fsf-mainline-vlad-lra-r221324 Thread model: posix gcc version 5.0.0 20150310 (experimental) [trunk revision 221324] (GCC) [bergner@makalu-lp1 LRA]$ /home/bergner/gcc/build/gcc-fsf-mainline-vlad-lra-r221324/gcc/xgcc -B/home/bergner/gcc/build/gcc-fsf-mainline-vlad-lra-r221324/gcc -O1 -m64 -mcpu=power8 -mlra -S pr63491.i pr63491.i: In function ‘foo’: pr63491.i:14:1: internal compiler error: in check_and_process_move, at lra-constraints.c:1171 } ^ 0x1099c933 check_and_process_move /home/bergner/gcc/gcc-fsf-mainline-vlad-lra/gcc/lra-constraints.c:1168 0x109a5313 curr_insn_transform /home/bergner/gcc/gcc-fsf-mainline-vlad-lra/gcc/lra-constraints.c:3464 0x109a9b93 lra_constraints(bool) /home/bergner/gcc/gcc-fsf-mainline-vlad-lra/gcc/lra-constraints.c:4452 0x10989da3 lra(_IO_FILE*) /home/bergner/gcc/gcc-fsf-mainline-vlad-lra/gcc/lra.c:2301 0x1090cb97 do_reload /home/bergner/gcc/gcc-fsf-mainline-vlad-lra/gcc/ira.c:5418 0x1090d22f execute /home/bergner/gcc/gcc-fsf-mainline-vlad-lra/gcc/ira.c:5589 Please submit a full bug report, with preprocessed source if appropriate. Please include the complete backtrace with any bug report. See http://gcc.gnu.org/bugs.html for instructions.
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 --- Comment #6 from Peter Bergner bergner at gcc dot gnu.org --- (In reply to Vladimir Makarov from comment #5) Sorry, I can not reproduce the bug on the today trunk. Probably it was fixed by numerous changes in LRA since Oct. This still fails for me today on my big endian system using revision 220792. Did you use the exact same compiler options I used? [bergner@makalu-lp1 LRA]$ cat pr63491.i typedef __int128_t __attribute__((__vector_size__(16))) vector_128_t; typedef unsigned long long scalar_64_t; vector_128_t foo (void) { union { scalar_64_t i64[2]; vector_128_t v128; } u; u.i64[0] = 1; u.i64[1] = 2; return u.v128; } [bergner@makalu-lp1 LRA]$ /home/bergner/gcc/build/gcc-fsf-mainline-pr63491/gcc/xgcc -B/home/bergner/gcc/build/gcc-fsf-mainline-pr63491/gcc -O1 -m64 -mcpu=power8 -mlra -S pr63491.i pr63491.i: In function ‘foo’: pr63491.i:14:1: internal compiler error: in check_and_process_move, at lra-constraints.c:1170 } ^ 0x10994a7b check_and_process_move /home/bergner/gcc/gcc-fsf-mainline-pr63491/gcc/lra-constraints.c:1167 0x1099d45b curr_insn_transform /home/bergner/gcc/gcc-fsf-mainline-pr63491/gcc/lra-constraints.c:3463 0x109a1cdb lra_constraints(bool) /home/bergner/gcc/gcc-fsf-mainline-pr63491/gcc/lra-constraints.c:4451 0x1098221b lra(_IO_FILE*) /home/bergner/gcc/gcc-fsf-mainline-pr63491/gcc/lra.c:2292 0x109050d7 do_reload /home/bergner/gcc/gcc-fsf-mainline-pr63491/gcc/ira.c:5418 0x1090576f execute /home/bergner/gcc/gcc-fsf-mainline-pr63491/gcc/ira.c:5589 Please submit a full bug report, with preprocessed source if appropriate.
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 --- Comment #5 from Vladimir Makarov vmakarov at gcc dot gnu.org --- Sorry, I can not reproduce the bug on the today trunk. Probably it was fixed by numerous changes in LRA since Oct.
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 --- Comment #4 from Peter Bergner bergner at gcc dot gnu.org --- I'll add that we hit this as part of our work on trying to move POWER over to using LRA by default.
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 --- Comment #2 from Peter Bergner bergner at gcc dot gnu.org --- Slightly simplified test case that still fails without a function call: typedef __int128_t __attribute__((__vector_size__(16))) vector_128_t; typedef unsigned long long scalar_64_t; vector_128_t foo (void) { union { scalar_64_t i64[2]; vector_128_t v128; } u; u.i64[0] = 1; u.i64[1] = 2; return u.v128; }
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 David Edelsohn dje at gcc dot gnu.org changed: What|Removed |Added CC||saugustine at google dot com --- Comment #3 from David Edelsohn dje at gcc dot gnu.org --- *** Bug 63501 has been marked as a duplicate of this bug. ***
[Bug rtl-optimization/63491] Ice in LRA with simple vector test case on power
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63491 Peter Bergner bergner at gcc dot gnu.org changed: What|Removed |Added Target||powerpc64-linux, ||powerpc64le-linux CC||meissner at gcc dot gnu.org, ||vmakarov at gcc dot gnu.org --- Comment #1 from Peter Bergner bergner at gcc dot gnu.org --- CCing Vlad and Mike for their input.