Are ICEs par for the course on HEAD?

Are ICEs par for the course on HEAD?

[Bernd Jendrissek]

There's a PC on my desk that is really, really glad that gcc ICE'd while
bootstrapping.  The compile had been going for about 24 hours by then...

/mnt/disk/src/gcc-native/gcc/xgcc -shared-libgcc -B/mnt/disk/src/gcc-native/gcc/ -nostdinc++ -L/mnt/disk/src/gcc-native/i686-pc-linux-gnu/libstdc++-v3/src -L/mnt/disk/src/gcc-native/i686-pc-linux-gnu/libstdc++-v3/src/.libs -B/usr/local/i686-pc-linux-gnu/bin/ -B/usr/local/i686-pc-linux-gnu/lib/ -isystem /usr/local/i686-pc-linux-gnu/include -DHAVE_CONFIG_H -I. -I/mnt/disk/src/gcc/libjava -I./include -I./gcj -I/mnt/disk/src/gcc/libjava -Iinclude -I/mnt/disk/src/gcc/libjava/include -I/mnt/disk/src/gcc/libjava/../boehm-gc/include -DGC_LINUX_THREADS=1 -D_REENTRANT=1 -DTHREAD_LOCAL_ALLOC=1 -DSILENT=1 -DNO_SIGNALS=1 -DNO_EXECUTE_PERMISSION=1 -DALL_INTERIOR_POINTERS=1 -DJAVA_FINALIZATION=1 -DGC_GCJ_SUPPORT=1 -DATOMIC_UNCOLLECTABLE=1 -I/mnt/disk/src/gcc/libjava/libltdl -I/mnt/disk/src/gcc/libjava/libltdl -I/mnt/disk/src/gcc/libjava/.././libjava/../gcc -I/mnt/disk/src/gcc/libjava/../libffi/include -I../libffi/include -fno-rtti -fnon-call-exceptions -fdollars-in-identifiers -Wswitch-enum !
-D__NO_MATH_INLINES -ffloat-store -I/usr/local/X11R6/include -W -Wall -D_GNU_SOURCE -DPREFIX=\"/usr/local\" -g -O2 -D_GNU_SOURCE -Wp,-MD,.deps/interpret.pp -c /mnt/disk/src/gcc/libjava/interpret.cc  -fPIC -DPIC -o .libs/interpret.o
/mnt/disk/src/gcc/libjava/interpret.cc: In member function `void 
   _Jv_InterpMethod::run(void*, ffi_raw*)':
/mnt/disk/src/gcc/libjava/interpret.cc:3187: internal compiler error: RTL 
   check: expected code `reg', have `subreg' in split_2, at insn-recog.c:48229
Please submit a full bug report,
with preprocessed source if appropriate.

Oh, BTW, I *did* configure with --enable-checks.  Here's my configure
invokation:

/mnt/disk/src/gcc-native: ../gcc/configure --enable-checking=misc,tree,rtl,rtlflag,gc,gcac --srcdir=/mnt/disk/src/gcc \
  --with-slibdir=/lib \
  --with-gnu-as \
  --with-gnu-ld \
  --with-gxx-include-dir=/usr/include/g++-v3 \
  --enable-__cxa_atexit \
  --enable-shared \
  --enable-threads=posix \
  --with-dwarf3 \
  --with-system-zlib

Hmm, I wonder if it's related to my having binutils 2.12.90.0.4?

LAST_UPDATED:
Tue Sep 10 21:56:51 SAST 2002
Tue Sep 10 19:56:51 UTC 2002

bernd
-- 
berndfoobar@users.sourceforge.net is probably better to bookmark than any
employer-specific email address I may have appearing in the headers.
  Vanity page: http://www.tsct.co.za/~berndj/

Re: Are ICEs par for the course on HEAD?

[Zack Weinberg]

On Fri, Sep 13, 2002 at 02:11:20AM +0200, Bernd Jendrissek wrote:
> There's a PC on my desk that is really, really glad that gcc ICE'd while
> bootstrapping.  The compile had been going for about 24 hours by
> then...

You will find that it goes much, much faster if you omit 'gcac' from
the list of checks to enable.  That mode is only for finding a very
specific category of bugs, and you would know if you were looking for
one of them.

> /mnt/disk/src/gcc/libjava/interpret.cc: In member function `void 
>    _Jv_InterpMethod::run(void*, ffi_raw*)':
> /mnt/disk/src/gcc/libjava/interpret.cc:3187: internal compiler error: RTL 
>    check: expected code `reg', have `subreg' in split_2, at insn-recog.c:48229
> Please submit a full bug report,
> with preprocessed source if appropriate.

This is a real bug; thank you for reporting it.  (People trying to
reproduce it should make sure they have configured with at least
--enable-checking=rtl.)

> Hmm, I wonder if it's related to my having binutils 2.12.90.0.4?

I doubt it.  This is a case of an internal data structure being
constructed incorrectly.  Binutils bugs don't usually do that.

zw

Re: Are ICEs par for the course on HEAD?

[Roger Sayle]

Hi Bernd and Zack,
> > /mnt/disk/src/gcc/libjava/interpret.cc: In member function `void
> >    _Jv_InterpMethod::run(void*, ffi_raw*)':
> > /mnt/disk/src/gcc/libjava/interpret.cc:3187: internal compiler error:
> > RTL check: expected code `reg', have `subreg' in split_2, at
> > insn-recog.c:48229
> > Please submit a full bug report,
> > with preprocessed source if appropriate.
>
> This is a real bug; thank you for reporting it.  (People trying to
> reproduce it should make sure they have configured with at least
> --enable-checking=rtl.)

Interestingly, after being in the tree for nearly two months, this
problem was (hopefully) fixed on mainline just a few hours ago.

http://gcc.gnu.org/ml/gcc-patches/2002-09/msg00748.html

[Just in case people have trouble trying to reproduce it :>]

Roger
--

Re: Are ICEs par for the course on HEAD?

[Roger Sayle]

Hi Bernd and Zack,
> > /mnt/disk/src/gcc/libjava/interpret.cc: In member function `void
> >    _Jv_InterpMethod::run(void*, ffi_raw*)':
> > /mnt/disk/src/gcc/libjava/interpret.cc:3187: internal compiler error:
> > RTL check: expected code `reg', have `subreg' in split_2, at
> > insn-recog.c:48229
> > Please submit a full bug report,
> > with preprocessed source if appropriate.
>
> This is a real bug; thank you for reporting it.  (People trying to
> reproduce it should make sure they have configured with at least
> --enable-checking=rtl.)

Interestingly, after being in the tree for nearly two months, this
problem was (hopefully) fixed on mainline just a few hours ago.

http://gcc.gnu.org/ml/gcc-patches/2002-09/msg00748.html

[Just in case people have trouble trying to reproduce it :>]

Roger
--