Re: BUG: CVS egcs-19980608 fails to bootstrap on hppa1.1-hp-hpux10.20

[Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>]

Hmm, this sounds a lot like the bug in the combiner triggered by
gcc.c-torture/compile/980511-1.c (try it) on powerpc-unknown-linux-gnu. You
could ask Jim Wilson if he has a fix available already.

Franz.


At 08:51 09.06.98 , Manfred Hollstein wrote:
>
>Sources:	- egcs from CVS last updated at 1998/06/08 13:25:45 MEST
>		- libg++-2.8.1.1a
>Patches:	None.
>Environment:	- GNU binutils 2.9.1
>		- dejagnu-971028
>		- gperf-2.7
>
>configure flags: --host=hppa1.1-hp-hpux10.20 --srcdir=../egcs-19980608
--prefix=/u/b60/manfred/gnu/hppa1.1-hp-hpux10.20 --with-gnu-as
--with-gxx-include-dir='/u/b60/manfred/gnu/hppa1.1-hp-hpux10.20/lib/gcc-lib/
hppa1.1-hp-hpux10.20/egcs-2.91.35/include/g++'
--with-local-prefix=/u/b60/manfred/gnu/hppa1.1-hp-hpux10.20 --verbose
--norecursion 
>BOOT_CFLAGS:	-O9 -funroll-all-loops -g -DUSE_GAS
>BOOT_CC:	gcc version 2.8.1
>
>This happens when the just built stage2 compiler tries to build libgcc2.a:
>
>  _divdi3
>  ../../egcs-19980608/gcc/haifa-sched.c:7593: Internal compiler error in
function schedule_region
>
>A quick look in gdb shows:
>
>(gdb) r libgcc2.i -quiet -dumpbase libgcc2.c -g -g1 -O2 -O9 -W -Wall
-version -funroll-all-loops -fPIC -fexceptions -o libgcc2.s
>Starting program:
/local/manfred/work/GNU/egcs-19980608-hppa1.1-hp-hpux10.20/gcc/cc1
libgcc2.i -quiet -dumpbase libgcc2.c -g -g1 -O2 -O9 -W -Wall -version
-funroll-all-loops -fPIC -fexceptions -o libgcc2.s
>warning: Unable to find __d_pid symbol in object file.
>warning: Suggest linking with /usr/lib/end.o.
>warning: GDB will be unable to track shl_load/shl_unload calls
>Breakpoint 1 at 0x7af496e8
>Breakpoint 2 at 0x7af528fc
>GNU C version egcs-2.91.35 19980608 (gcc2 ss-980502 experimental)
(hppa1.1-hp-hpux10.20) compiled by GNU C version egcs-2.91.35 19980608
(gcc2 ss-980502 experimental).
>../../egcs-19980608/gcc/haifa-sched.c:7593: Internal compiler error in
function schedule_region
>
>Breakpoint 2, 0x7af528fc in ___exit ()
>(gdb) up
>#1  0x1d3844 in schedule_region (rgn=1) at
../../egcs-19980608/gcc/haifa-sched.c:7593
>7593              abort ();
>(gdb) l
>7588            /* Sanity check.  There should be no REG_DEAD notes
leftover at the end.
>7589               In practice, this can occur as the result of bugs in
flow, combine.c,
>7590               and/or sched.c.  The values of the REG_DEAD notes
remaining are
>7591               meaningless, because dead_notes is just used as a free
list.  */
>7592            if (dead_notes != 0)
>7593              abort ();
>7594        }
>7595
>7596      /* restore line notes.  */
>7597      if (write_symbols != NO_DEBUG)
>(gdb) p lineno
>$1 = 680
>(gdb) p dead_notes
>$2 = (struct rtx_def *) 0x400f6be0
>(gdb) pr
>(expr_list:REG_DEAD (reg/v:DI 102)
>    (expr_list:REG_DEAD (reg:SI 103)
>        (expr_list:REG_DEAD (reg/v:DI 96)
>            (nil))))
>(gdb) 
>
>manfred
>