[Bug target/37812] [4.4 Regression] Invalid mnemonic 'lvlx'
--- Comment #3 from jakub at gcc dot gnu dot org 2008-11-03 08:14 --- Isn't this just a user error though? I mean, for -mavx we also need an assembler that supports AVX instructions, we don't have a workaround for every AVX instruction. Even if you build with -march=native on SSSE3 capable CPU (i.e. core2 or later) GCC assumes your assembler supports SSSE3. Similarly for many other recently added instructions. lvlx insn is guarded with rs6000_cpu == PROCESSOR_CELL, how that got set? Are you configuring gcc with a cell arch eventhough your assembler lacks that support? Or is testing done with those flags? -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37812
[Bug target/37812] [4.4 Regression] Invalid mnemonic 'lvlx'
--- Comment #4 from dominiq at lps dot ens dot fr 2008-11-03 08:21 --- Isn't this just a user error though? I am just running the tests. Are you configuring gcc with a cell arch eventhough your assembler lacks that support? I use: ../gcc-4.4-work/configure --prefix=/opt/gcc/gcc4.4w --mandir=/opt/gcc/gcc4.4w/share/man --infodir=/opt/gcc/gcc4.4w/share/info --build=powerpc-apple-darwin9 --enable-languages=c,c++,fortran,objc,obj-c++,java --with-gmp=/sw --with-libiconv-prefix=/usr --with-system-zlib --x-includes=/usr/X11R6/include --x-libraries=/usr/X11R6/lib If there is something about cell, it is not visible here. Or is testing done with those flags? I don't think I have anything special for these tests. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37812
[Bug target/37812] [4.4 Regression] Invalid mnemonic 'lvlx'
--- Comment #5 from jakub at gcc dot gnu dot org 2008-11-03 09:07 --- Created an attachment (id=16615) -- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=16615action=view) gcc44-pr37812.patch Ah, I see, it is a testsuite bug then. altivec_check.h uses the lvlx insns inside of inline assembly unconditionally. There is no reason for that if not -mcpu=cell and when -mcpu=cell is used, the user and/or testcase should better make sure the assembler supports cell. Which leads to another bug, we need a new target-supports tcl function and guard all the -mcpu=cell tests with it (say powerpc_cell_ok or powerpc_ppu_ok, which would imply powerpc_altivec_ok and that these insns can be assembled). -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37812
[Bug target/37812] [4.4 Regression] Invalid mnemonic 'lvlx'
--- Comment #6 from jakub at gcc dot gnu dot org 2008-11-03 09:49 --- Updated patch with also target-supports posted. -- jakub at gcc dot gnu dot org changed: What|Removed |Added AssignedTo|unassigned at gcc dot gnu |jakub at gcc dot gnu dot org |dot org | URL||http://gcc.gnu.org/ml/gcc- ||patches/2008- ||11/msg00054.html Status|UNCONFIRMED |ASSIGNED Ever Confirmed|0 |1 Last reconfirmed|-00-00 00:00:00 |2008-11-03 09:49:49 date|| http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37812
[Bug target/37812] [4.4 Regression] Invalid mnemonic 'lvlx'
--- Comment #7 from dominiq at lps dot ens dot fr 2008-11-03 17:31 --- Updated patch with also target-supports posted. With the updated patch I now get only three failures for the powerpc directory FAIL: gcc.target/powerpc/ppc64-abi-1.c (test for excess errors) -- 64 bit mode only Excess errors: /var/folders/FK/FKCVPmNbH5SNynFQmqGomk+++TI/-Tmp-//cc3szi8j.s:23:Parameter syntax error (parameter 2) /var/folders/FK/FKCVPmNbH5SNynFQmqGomk+++TI/-Tmp-//cc3szi8j.s:23:Invalid mnemonic 'got(r2)' /var/folders/FK/FKCVPmNbH5SNynFQmqGomk+++TI/-Tmp-//cc3szi8j.s:95:Parameter syntax error (parameter 2) /var/folders/FK/FKCVPmNbH5SNynFQmqGomk+++TI/-Tmp-//cc3szi8j.s:95:Invalid mnemonic 'got(r2)' /var/folders/FK/FKCVPmNbH5SNynFQmqGomk+++TI/-Tmp-//cc3szi8j.s:165:Parameter syntax error (parameter 2) /var/folders/FK/FKCVPmNbH5SNynFQmqGomk+++TI/-Tmp-//cc3szi8j.s:165:Invalid mnemonic 'got(r2)' /var/folders/FK/FKCVPmNbH5SNynFQmqGomk+++TI/-Tmp-//cc3szi8j.s:251:Parameter syntax error (parameter 2) /var/folders/FK/FKCVPmNbH5SNynFQmqGomk+++TI/-Tmp-//cc3szi8j.s:251:Invalid mnemonic 'got(r2)' /var/folders/FK/FKCVPmNbH5SNynFQmqGomk+++TI/-Tmp-//cc3szi8j.s:366:Parameter syntax error (parameter 2) /var/folders/FK/FKCVPmNbH5SNynFQmqGomk+++TI/-Tmp-//cc3szi8j.s:366:Invalid mnemonic 'got(r2)' /var/folders/FK/FKCVPmNbH5SNynFQmqGomk+++TI/-Tmp-//cc3szi8j.s:443:Parameter syntax error (parameter 2) /var/folders/FK/FKCVPmNbH5SNynFQmqGomk+++TI/-Tmp-//cc3szi8j.s:443:Invalid mnemonic 'got(r2)' FAIL: gcc.target/powerpc/ppc64-double-1.c scan-assembler-not stfd -- 32 and 64 bit modes FAIL: gcc.target/powerpc/pr18096-1.c stack frame too large (test for errors, line 8) -- 32 bit mode FAIL: gcc.target/powerpc/pr18096-1.c (test for excess errors) /opt/gcc/gcc-4.4-work/gcc/testsuite/gcc.target/powerpc/pr18096-1.c: In function 'mkcatdefs': /opt/gcc/gcc-4.4-work/gcc/testsuite/gcc.target/powerpc/pr18096-1.c:7: error: total size of local objects too large Are these failures related to this PR or should I open a new one? -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37812
[Bug target/37812] [4.4 Regression] Invalid mnemonic 'lvlx'
--- Comment #8 from jakub at gcc dot gnu dot org 2008-11-03 17:43 --- Unrelated. For ppc64-double-1.c I think there is another PR open already. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37812
[Bug target/37812] [4.4 Regression] Invalid mnemonic 'lvlx'
--- Comment #9 from jakub at gcc dot gnu dot org 2008-11-03 22:01 --- Subject: Bug 37812 Author: jakub Date: Mon Nov 3 21:59:36 2008 New Revision: 141564 URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=141564 Log: PR target/37812 * gcc.target/powerpc/altivec_check.h (altivec_cell_check): Abort if __PPU__ isn't defined. * lib/target-supports.exp (check_effective_target_powerpc_ppu_ok): New. * gcc.target/powerpc/altivec-cell-8.c: Use powerpc_ppu_ok. Modified: trunk/gcc/testsuite/ChangeLog trunk/gcc/testsuite/gcc.target/powerpc/altivec-cell-8.c trunk/gcc/testsuite/gcc.target/powerpc/altivec_check.h trunk/gcc/testsuite/lib/target-supports.exp -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37812
[Bug target/37812] [4.4 Regression] Invalid mnemonic 'lvlx'
--- Comment #10 from jakub at gcc dot gnu dot org 2008-11-03 22:08 --- Fixed. -- jakub at gcc dot gnu dot org changed: What|Removed |Added Status|ASSIGNED|RESOLVED Resolution||FIXED http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37812
[Bug target/37812] [4.4 Regression] Invalid mnemonic 'lvlx'
-- rguenth at gcc dot gnu dot org changed: What|Removed |Added Priority|P3 |P2 Target Milestone|--- |4.4.0 Version|unknown |4.4.0 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37812
[Bug target/37812] [4.4 Regression] Invalid mnemonic 'lvlx'
--- Comment #2 from ubizjak at gmail dot com 2008-11-01 21:04 --- There was similar problem on x86 target with sun as and ffreep mnemonic. This was fixed by a configure check and conditional generation of either ffreep or .word\t0x. Please grep for HAVE_AS_IX86_FFREEP. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37812
[Bug target/37812] [4.4 Regression] Invalid mnemonic 'lvlx'
--- Comment #1 from pinskia at gcc dot gnu dot org 2008-10-12 23:05 --- Basically Apple's as does not have support for this instruction, I could use .long if really needed for this test. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37812