Joern Rennecke wrote:
Andrew Haley wrote:
> > So it seems I have to abandom the system compiler for doing
bootstraps.
Can you make a simple testcase for this?
Andrew.
Attached.
This fails at -O0..-O2 for gcc 3.2.2 and gcc 3.2.3 on
i686-pc-linux-gnu. For gcc 3.4.3,
it still fails at
> Hmm, as far as I can tell, stage1 i386.c:ix86_split_ashr has been
> miscompiled by gcc (GCC) 3.2.3 20030502 (Red Hat Linux 3.2.3-54).
Thanks a bunch for sorting this out!
--
Eric Botcazou
Andrew Haley wrote:
>
> So it seems I have to abandom the system compiler for doing bootstraps.
Can you make a simple testcase for this?
Andrew.
Attached.
This fails at -O0..-O2 for gcc 3.2.2 and gcc 3.2.3 on
i686-pc-linux-gnu. For gcc 3.4.3,
it still fails at -O1.
4.2.0 appears to b
Joern RENNECKE writes:
> Joern Rennecke wrote:
>
> > Eric Botcazou wrote:
> >
> >>> make[3]: Leaving directory `/mnt/scratch/nightly/2006-07-04/i686'
> >>> Comparing stages 2 and 3
> >>> warning: ./cc1-checksum.o differs
> >>> warning: ./cc1plus-checksum.o differs
> >>> warning: ./cc1obj-
Joern Rennecke wrote:
Eric Botcazou wrote:
make[3]: Leaving directory `/mnt/scratch/nightly/2006-07-04/i686'
Comparing stages 2 and 3
warning: ./cc1-checksum.o differs
warning: ./cc1plus-checksum.o differs
warning: ./cc1obj-checksum.o differs
Bootstrap comparison failure!
Does the attach
> No, I still get the same set of .o files that differ, and the -fdump-noaddr
> (See PR other/28251) peephole2 dump for cfg.c compiled with stage1 /
> stage2 cc1 is still the same.
Then I'm puzzled. The original patch is supposed to be essentially a nop for
languages that do not use ARRAY_RANGE
Eric Botcazou wrote:
make[3]: Leaving directory `/mnt/scratch/nightly/2006-07-04/i686'
Comparing stages 2 and 3
warning: ./cc1-checksum.o differs
warning: ./cc1plus-checksum.o differs
warning: ./cc1obj-checksum.o differs
Bootstrap comparison failure!
Does the attached patch make any diffe
> make[3]: Leaving directory `/mnt/scratch/nightly/2006-07-04/i686'
> Comparing stages 2 and 3
> warning: ./cc1-checksum.o differs
> warning: ./cc1plus-checksum.o differs
> warning: ./cc1obj-checksum.o differs
> Bootstrap comparison failure!
Does the attached patch make any difference?
--
Eric B
On Jul 4, 2006, at 8:35 PM, Eric Botcazou wrote:
On the other hand, I now get tons of failures in libjava on x86-64
while
Andreas' testsuite is clean:
http://gcc.gnu.org/ml/gcc-testresults/2006-07/msg00164.html
... I get the tons of failures in libjava at rev 115159 too!
Those should be
> That might be my SRA patch (revision 115160) although it bootstraps fine
> for me on i586 and for Joe on i686:
> http://gcc.gnu.org/ml/gcc-testresults/2006-07/msg00157.html
Or might not because...
> On the other hand, I now get tons of failures in libjava on x86-64 while
> Andreas' testsuite is
> Are you using --with-arch=i686 ?
Yes, I cannot reproduce the bootstrap failure with
[EMAIL PROTECTED]:~/build/gcc/native32> gcc/xgcc -v
Using built-in specs.
Target: i686-pc-linux-gnu
Configured with: /home/eric/svn/gcc/configure i686-pc-linux-gnu
--prefix=/home/eric/install/gcc --with-as=/usr
Eric Botcazou wrote:
Not sure what's going on exactly...
Using -fdump-unnumbered dumps, I find the first significant difference
in 144r.peephole2:
13868a13869,13880
> (set (reg:SI 2 cx)
> (ior:SI (ashiftrt:SI (reg:SI 2 cx)
> (const_int
Joern Rennecke wrote:
Eric Botcazou wrote:
Not sure what's going on exactly...
The machines I use have the 'security enhancement' enabled which makes
addresses vary
between program invocations. So if code generation depends on pointer
values at any
point, this will cause varying beha
Eric Botcazou wrote:
Not sure what's going on exactly...
The machines I use have the 'security enhancement' enabled which makes
addresses vary
between program invocations. So if code generation depends on pointer
values at any
point, this will cause varying behaviour.
A common mistake
> Revision: 115174
>
> build/host/target: i686-pc-linux-gnu
>
> make[3]: Leaving directory `/mnt/scratch/nightly/2006-07-04/i686'
> Comparing stages 2 and 3
> warning: ./cc1-checksum.o differs
> warning: ./cc1plus-checksum.o differs
> warning: ./cc1obj-checksum.o differs
> Bootstrap comparison fail
Revision: 115174
build/host/target: i686-pc-linux-gnu
# This directory was configured as follows,
# on host linsvr6:
#
# ../srcw/configure --with-arch=i686 --disable-gdb
--enable-languages=c,c++,ja
va,objc
make[3]: Leaving directory `/mnt/scratch/nightly/2006-07-04/i686'
Comparing stages 2
16 matches
Mail list logo