[Bug c++/24049] New: compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[Bug c++/24049] New: compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[harold at alum dot mit dot edu]

The attached preprocessed file segfaults during compilation with the compile line:

gcc-4.1 -DHAVE_CONFIG_H -D__NO_CTYPE -O3 -DPROTOTYPES=0 -DUSE_ANSI=1 -DAMD64 -O3
-ffast-math -funroll-all-loops -fpeel-loops -ftracer -funswitch-loops
-ftree-vectorize -march=k8 -g -pthread -MT desc.o -MD -MP -MF ".deps/desc.Tpo"
-c -o desc.o desc.i

-- 
           Summary: compiler error: Segmentation fault In function
                    'DESX_CBCUpdate'
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: harold at alum dot mit dot edu
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: x86_64-linux-gnu
  GCC host triplet: x86_64-linux-gnu
GCC target triplet: x86_64-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug c++/24049] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[harold at alum dot mit dot edu]

------- Additional Comments From harold at alum dot mit dot edu  2005-09-24 14:04 -------
Created an attachment (id=9799)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=9799&action=view)
Segfaults on compilation.gcc-4.1 -DHAVE_CONFIG_H -D__NO_CTYPE -O3
-DPROTOTYPES=0 -DUSE_ANSI=1 -DAMD64 -O3 -ffast-math -funroll-all-loops
-fpeel-loops -ftracer -funswitch-loops -ftree-vectorize -march=


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug c++/24049] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-09-24 15:49 -------
Can you give the output of "gcc-4.1 -v"?

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug c++/24049] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[harold at alum dot mit dot edu]

------- Additional Comments From harold at alum dot mit dot edu  2005-09-24 19:06 -------
harold@cimarron /tmp $ gcc-4.1 -v
Using built-in specs.
Target: x86_64-unknown-linux-gnu
Configured with: ../gcc-4.1/configure --program-suffix=-4.1 --disable-multilib
--enable-languages=c,c++
Thread model: posix
gcc version 4.1.0 20050919 (experimental)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug middle-end/24049] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c++                         |middle-end
           Keywords|                            |GC, ice-on-valid-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug middle-end/24049] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-09-24 20:00 -------
Oh, I had missed the -g but that really should not matter.  Anyways it does not with "--param ggc-
min-heapsize=0 --param ggc-min-expand=0" which means this is a GC problem.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug middle-end/24049] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[harold at alum dot mit dot edu]

------- Additional Comments From harold at alum dot mit dot edu  2005-09-24 20:07 -------

I know that I should really only compile with such aggressive options for the
most compute-intensive modules, in order to avoid these sorts of issues, but I
thought you would appreciate the bugs in advance of the release.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug middle-end/24049] [4.1 regression] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-09-24 20:09 -------
Confirmed, reduced testcase:
int DES_CBCUpdate(unsigned char * output, int len)
{
  int work[2];
  unsigned int i;
  for(i = 0;i < len/8;i++)
    unscrunch (&output[8*i], work);
}

Only "-O1  -ftree-vectorize --param ggc-min-heapsize=0 --param ggc-min-expand=0"  is needed to 
reproduce this bug.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pinskia at gcc dot gnu dot
                   |                            |org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2005-09-24 20:09:10
               date|                            |
            Summary|compiler error: Segmentation|[4.1 regression] compiler
                   |fault In function           |error: Segmentation fault In
                   |'DESX_CBCUpdate'            |function 'DESX_CBCUpdate'
   Target Milestone|---                         |4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug middle-end/24049] [4.1 regression] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-09-24 20:12 -------
This is a very recent regression as it worked with 20050908.
I think it was caused by:
2005-09-17  Richard Henderson  <rth@redhat.com>

        * tree-pass.h, tree-flow.h, tree-ssa-dce.c: Revert last change.
        * tree-ssa-loop.c (pass_vect_dce): Remove.
        * passes.c (init_optimization_passes): Add pass_dce as a sub-pass
        of pass_vectorize.

The issue is that the scev is not very GC friendly.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rth at gcc dot gnu dot org
  GCC build triplet|x86_64-linux-gnu            |
   GCC host triplet|x86_64-linux-gnu            |
 GCC target triplet|x86_64-linux-gnu            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug middle-end/24049] [4.1 regression] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-09-27 23:29 -------
*** Bug 24095 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |janis187 at us dot ibm dot
                   |                            |com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug middle-end/24049] [4.1 regression] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[janis187 at us dot ibm dot com]

------- Additional Comments From janis187 at us dot ibm dot com  2005-09-27 23:52 -------
The reghunt for 24095 (now a duplicate of this one) identified this patch:

  http://gcc.gnu.org/ml/gcc-cvs/2005-09/msg00635.html

which is the one prior to the one that Andrew suspected.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug middle-end/24049] [4.1 regression] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[pinskia at physics dot uc dot edu]

------- Additional Comments From pinskia at physics dot uc dot edu  2005-09-27 23:57 -------
Subject: Re:  [4.1 regression] compiler error: Segmentation fault In function 'DESX_CBCUpdate'


On Sep 27, 2005, at 7:52 PM, janis187 at us dot ibm dot com wrote:

>
> ------- Additional Comments From janis187 at us dot ibm dot com  
> 2005-09-27 23:52 -------
> The reghunt for 24095 (now a duplicate of this one) identified this 
> patch:
>
>   http://gcc.gnu.org/ml/gcc-cvs/2005-09/msg00635.html
>
> which is the one prior to the one that Andrew suspected.

But it was reverted in the one I pointed out and changed to
invoked DCE a different way.

-- Pinski



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug middle-end/24049] [4.1 regression] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-09-28 00:04 -------
Note the older bug for this was PR 21639.  I wonder why the testcase was not applied.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049

[Bug middle-end/24049] [4.1 regression] compiler error: Segmentation fault In function 'DESX_CBCUpdate'

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-09-28 00:06 -------
PR 21805 was filed after PR 21639 was worked around.  But the weird part is that nobdoy applied the 
testcase for PR 21639.

One more thing, it seems like the vector_lower pass should be dependent on tree-vectorizer running.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|                            |21805
           Severity|normal                      |critical


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24049