This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
RE: compiling CVS-latest egcs "locks" on DG/UX
- To: "'Jeff Garzik'" <jgarzik at pobox dot com>
- Subject: RE: compiling CVS-latest egcs "locks" on DG/UX
- From: "Eric Raskin" <ehr at listworks dot com>
- Date: Fri, 18 Jun 1999 10:49:04 -0400
- Cc: <egcs-bugs at egcs dot cygnus dot com>
- Reply-To: <ehr at listworks dot com>
Jeff:
Do a search on the egcs-patches archive for "recombine_givs". I'm pretty
sure there were some recent changes made to that section of code, but I
don't remember who made them 8-(. Whoever it was can probably help you
further now...
Eric Raskin
-----Original Message-----
From: Jeff Garzik [mailto:jgarzik@pobox.com]
Sent: Friday, June 18, 1999 10:44 AM
To: Eric Raskin
Cc: egcs-bugs@egcs.cygnus.com
Subject: RE: compiling CVS-latest egcs "locks" on DG/UX
Here is the dbx output, via gcore, of cc1 while it is in a
really-really-long/infinite loop compiling gcc/regclass.c in stage1.
Jeff
duet:/work/garzik/egcs> dbx -Igcc:gcc/cp:gcc/config:gcc/config/i386
gcc/stage1/cc1 /work/garzik/tmp/cc1.core.21641
Type "help" for a list of commands
ELF corefile process information:
UID GID PID PPID PGID SID SZ COMMAND
505 252 21641 21639 11082 12499 847 stage1/cc1
/var/tmp/ccbNfaaa.i -quiet -dumpbase regclass.c -mstandard -mlegend
Created on system: dgux duet R4.20MU04 generic AViiON
Processing object: /work/garzik/egcs/gcc/stage1/cc1
Updating linker symbol table...2648 symbols added. (Total: 2648)
Updating symbolic debugging information...81 modules added. (Total: 81)
Processing object: /usr/dglib/libc.so.1
Updating linker symbol table...1834 symbols added. (Total: 4482)
Updating symbolic debugging information...0 modules added. (Total: 81)
Stopped at line 7130, routine find_life_end(x=0x0834e998, stats=0x08040860,
insn=0x0834e9a4, biv=0x0834b4a0), file loop.c
* 7130 return find_life_end (SET_SRC (x), stats, insn, biv);
core-dump-requested
caught signal 5, SIGTRAP, trace trap - (core dumped)
(dbx) whe
Frame 0, line 7130, routine find_life_end(x=0x0834e998, stats=0x08040860,
insn=0x0834e9a4, biv=0x0834b4a0), file loop.c
Frame 1, line 7357, routine recombine_givs(bl=0x08373c58,
loop_start=0x0834e544, loop_end=0x0834ee68, unroll_p=0), file loop.c
Frame 2, line 4767, routine strength_reduce(scan_start=0x0834e57c,
end=0x0834ee68, loop_top=0x00000000, insn_count=19, loop_start=0x0834e544,
loop_end=0x0834ee68, loop_cont=0x0834ed58, unroll_p=0, bct_p=0), file loop.c
Frame 3, line 1174, routine scan_loop(loop_start=0x0834e544, end=0x0834ee68,
loop_cont=0x0834ed58, unroll_p=0, bct_p=0), file loop.c
Error: Attempt to access invalid register, $eax.
Frame 4, line 574, routine loop_optimize(f=0x0834b12c, dumpfile=,
unroll_p=0, bct_p=0), file loop.c
Frame 5, line 3943, routine rest_of_compilation(decl=0x083478c0), file
toplev.c
Frame 6, line 7270, routine finish_function(nested=0), file c-decl.c
Frame 7, line 350, routine yyparse(), file c-parse.y
Frame 8, line 3265, routine compile_file(name=0x0804701b ->
"/var/tmp/ccbNfaaa.i"), file toplev.c
Frame 9, line 5444, routine main(argc=18, argv=0x08046dec), file toplev.c
Frame 10, pc 0x8049129 (_start+105)
On Fri, 18 Jun 1999, Eric Raskin wrote:
> Jeff:
>
> Do a gcore on the running process to get a core dump, then use dbx to
look
> at it:
>
> dbx -I<your sources>/gcc:<your sources>/gcc/cp:<your
> sources>/gcc/config:<your sources>/gcc/config/i386 <your stage1 dir>/xgcc
> core
>
> A stack trace and/or a routine name where it is looping would be nice
:-).
> At least it will give the gcc gurus out there a chance to see what's
> happening.
>
> ------------------------------------------------------------------------
> Eric H. Raskin Voice: 914-769-7100 x321
> President, CSC Division Fax: 914-769-8070
> The Listworks Corp. E-Mail: ehr@listworks.com
> 1 Campus Drive
> Pleasantville, NY 10570
>
> -----Original Message-----
> From: owner-egcs-bugs@egcs.cygnus.com
> [mailto:owner-egcs-bugs@egcs.cygnus.com]On Behalf Of Jeff Garzik
> Sent: Friday, June 18, 1999 10:09 AM
> To: egcs-bugs@egcs.cygnus.com
> Subject: compiling CVS-latest egcs "locks" on DG/UX
>
>
> Platform: i586-dg-dguxR4.20MU04
> Platform compiler: gcc version 2.7.2 (ix86 DG/ux, DG-2.7.2.9, Feb 3 1998)
>
>
> I did a 'cvs up' yesterday afternoon, and attempted to compile egcs with
>
> ./configure --prefix=/usr/local/gcc.new --enable-languages=c++
> make bootstrap
>
> The compilation quite simply stops when stage1/xgcc compiles regclass.c!
>
> 'ps' shows that stage1/cc1 is using tons of CPU time, so my guess is
> that stage1/cc1 is in some sort of infinite, or really really long loop.
>
> Any help or suggestions are appreciated.
>
> Jeff
>
>
>
>
>
>
>
>