public inbox for gcc-bugs@sourceware.org help / color / mirror / Atom feed
* [Bug tree-optimization/23449] New: vortex fails without -fno-strict-aliasing @ 2005-08-17 23:53 janis at gcc dot gnu dot org 2005-09-02 20:34 ` [Bug tree-optimization/23449] " janis at gcc dot gnu dot org 2005-09-07 21:07 ` janis at gcc dot gnu dot org 0 siblings, 2 replies; 5+ messages in thread From: janis at gcc dot gnu dot org @ 2005-08-17 23:53 UTC (permalink / raw) To: gcc-bugs The SPEC CPU2000 test vortex segfaults on powerpc64-linux with current mainline for either -m32 or -m64 with -O2. It succeeds when compiled with -fno-strict-aliasing. It's quite likely that this is due to invalid code in vortex itself, but nothing jumped out at me. I'm about to go on a two-week vacation, so I'm recording this in case someone else would like to continue the investigation. If mem00.c is compiled with "-O2 -fno-strict-aliasing" and the rest of vortex is compiled with "-O2", the test passes. Variables ChunkAddrChunk and NumChunkChunks, defined in mem00.h, are static and their values never change. If either one is made global instead, the test passes. If either variable is replaced with a global variable with the same value at lines 240-248, the test passes. If the for loop at line 240 is changed to increment ChunkChunk as the third expression in for loop instead of in the second expression, the test passes. -- Summary: vortex fails without -fno-strict-aliasing Product: gcc Version: 4.1.0 Status: UNCONFIRMED Severity: normal Priority: P2 Component: tree-optimization AssignedTo: unassigned at gcc dot gnu dot org ReportedBy: janis at gcc dot gnu dot org CC: gcc-bugs at gcc dot gnu dot org GCC target triplet: powerpc-linux http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23449 ^ permalink raw reply [flat|nested] 5+ messages in thread
* [Bug tree-optimization/23449] vortex fails without -fno-strict-aliasing 2005-08-17 23:53 [Bug tree-optimization/23449] New: vortex fails without -fno-strict-aliasing janis at gcc dot gnu dot org @ 2005-09-02 20:34 ` janis at gcc dot gnu dot org 2005-09-07 21:07 ` janis at gcc dot gnu dot org 1 sibling, 0 replies; 5+ messages in thread From: janis at gcc dot gnu dot org @ 2005-09-02 20:34 UTC (permalink / raw) To: gcc-bugs ------- Additional Comments From janis at gcc dot gnu dot org 2005-09-02 20:34 ------- The function Mem_NewOry in vortex's mem00.c uses lots of casts to void **, so the failure is due to invalid code rather than a GCC bug. I'll investigate further and report this to SPEC. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23449 ^ permalink raw reply [flat|nested] 5+ messages in thread
* [Bug tree-optimization/23449] vortex fails without -fno-strict-aliasing 2005-08-17 23:53 [Bug tree-optimization/23449] New: vortex fails without -fno-strict-aliasing janis at gcc dot gnu dot org 2005-09-02 20:34 ` [Bug tree-optimization/23449] " janis at gcc dot gnu dot org @ 2005-09-07 21:07 ` janis at gcc dot gnu dot org 1 sibling, 0 replies; 5+ messages in thread From: janis at gcc dot gnu dot org @ 2005-09-07 21:07 UTC (permalink / raw) To: gcc-bugs ------- Additional Comments From janis at gcc dot gnu dot org 2005-09-07 21:07 ------- The code in mem*.[ch] is much messier than I originally thought. There are lots of casts in assignments of pointer variables. The macros in mem00.h starting with Unit_Size, which are used in both lvalues and rvalues, dereference through pointer casts. I'm no longer planning to suggest a fix to SPEC. The code in vortex violates aliasing rules and any fix would not be portable C; we'll need to work around it by compiling with -fno-strict-aliasing until someone comes up with a better solution. Closing this PR as INVALID. -- What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |INVALID http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23449 ^ permalink raw reply [flat|nested] 5+ messages in thread
[parent not found: <bug-23449-4503@http.gcc.gnu.org/bugzilla/>]
* [Bug tree-optimization/23449] vortex fails without -fno-strict-aliasing [not found] <bug-23449-4503@http.gcc.gnu.org/bugzilla/> @ 2007-10-12 10:16 ` ubizjak at gmail dot com 2009-01-15 19:32 ` raksit at google dot com 1 sibling, 0 replies; 5+ messages in thread From: ubizjak at gmail dot com @ 2007-10-12 10:16 UTC (permalink / raw) To: gcc-bugs ------- Comment #3 from ubizjak at gmail dot com 2007-10-12 10:16 ------- *** Bug 31024 has been marked as a duplicate of this bug. *** -- ubizjak at gmail dot com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |kenneth dot hoste at elis | |dot ugent dot be http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23449 ^ permalink raw reply [flat|nested] 5+ messages in thread
* [Bug tree-optimization/23449] vortex fails without -fno-strict-aliasing [not found] <bug-23449-4503@http.gcc.gnu.org/bugzilla/> 2007-10-12 10:16 ` ubizjak at gmail dot com @ 2009-01-15 19:32 ` raksit at google dot com 1 sibling, 0 replies; 5+ messages in thread From: raksit at google dot com @ 2009-01-15 19:32 UTC (permalink / raw) To: gcc-bugs ------- Comment #4 from raksit at google dot com 2009-01-15 19:32 ------- *** Bug 38858 has been marked as a duplicate of this bug. *** -- raksit at google dot com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |raksit at google dot com http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23449 ^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2009-01-15 19:32 UTC | newest] Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2005-08-17 23:53 [Bug tree-optimization/23449] New: vortex fails without -fno-strict-aliasing janis at gcc dot gnu dot org 2005-09-02 20:34 ` [Bug tree-optimization/23449] " janis at gcc dot gnu dot org 2005-09-07 21:07 ` janis at gcc dot gnu dot org [not found] <bug-23449-4503@http.gcc.gnu.org/bugzilla/> 2007-10-12 10:16 ` ubizjak at gmail dot com 2009-01-15 19:32 ` raksit at google dot com
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).