[Bug c++/21085] New: Virtual memory exhausted with g++

[Bug c++/21085] New: Virtual memory exhausted with g++

[strieder at informatik dot uni-kl dot de]

Hello,   
   
there  must be a serious problem in   
   
 GNU C++ version 4.0.0 20050410 (prerelease) (i686-pc-linux-gnu)   
        compiled by GNU C version 3.3.4 (pre 3.3.5 20040809) (SuSE Linux 9.2).   
  
The same happens with  GNU C++ version 4.0.0 20050410 when built with itself.  
  
On a Pentium III 1GHz with enough RAM to avoid swapping, cc1plus will eat memory at  
about 200MB/sec until the OS kills it, as observable with top at a sub-second interval.  
   
The .ii file produced by the issued command has bzipped 131128 bytes. I have not found   
a sensible way to include that file here, you will find it under   
   
http://www-madlener.informatik.uni-kl.de/staff/strieder/triedriver.t.ii.bz2   
   
or I will send it on request.   
   
The code used to compile under gcc-2.95, gcc-3.3.x, on i386 ans SPARCs and gcc-3.4.[23] on  
i386.  
  
How the .ii was produced:  
  
g++ -v --save-temps -gdwarf-2   -Wall  -W -Winline -Woverloaded-virtual -I.    
-I/home/strieder/xssr/sys -I/home/strieder/xssr/sys/xssr-gui   -DINLINE='inline' -DDO_INLINE       
-c -o triedriver.t.o /home/strieder/xssr/sys/testdrivers/triedriver.t.cc    
Using built-in specs.    
Target: i686-pc-linux-gnu    
Configured with: ../gcc-4.0.0-20050410/configure --enable-threads=posix    
--prefix=/net/opt/gcc-4.0.0-20050410 --enable-languages=c,c++ --with-system-zlib    
--enable-shared --enable-__cxa_atexit    
Thread model: posix    
gcc version 4.0.0 20050410 (prerelease)    
 /net/opt/gcc-4.0.0-20050410/libexec/gcc/i686-pc-linux-gnu/4.0.0/cc1plus -E -quiet -v -I.    
-I/home/strieder/xssr/sys -I/home/strieder/xssr/sys/xssr-gui -D_GNU_SOURCE    
-DINLINE=inline -DDO_INLINE /home/strieder/xssr/sys/testdrivers/triedriver.t.cc    
-mtune=pentiumpro -Wall -W -Winline -Woverloaded-virtual -fworking-directory    
-fpch-preprocess -o triedriver.t.ii    
ignoring nonexistent directory    
"/net/opt/gcc-4.0.0-20050410/lib/gcc/i686-pc-linux-gnu/4.0.0/../../../../i686-pc-linux-gnu/include"    
#include "..." search starts here:    
#include <...> search starts here:    
 .    
 /home/strieder/xssr/sys    
 /home/strieder/xssr/sys/xssr-gui    
 /net/opt/gcc-4.0.0-20050410/lib/gcc/i686-pc-linux-gnu/4.0.0/../../../../include/c++/4.0.0    
 /net/opt/gcc-4.0.0-20050410/lib/gcc/i686-pc-linux-gnu/4.0.0/../../../../include/c++/4.0.0/i686-pc-linux-gnu    
 /net/opt/gcc-4.0.0-20050410/lib/gcc/i686-pc-linux-gnu/4.0.0/../../../../include/c++/4.0.0/backward    
 /usr/local/include    
 /net/opt/gcc-4.0.0-20050410/include    
 /net/opt/gcc-4.0.0-20050410/lib/gcc/i686-pc-linux-gnu/4.0.0/include    
 /usr/include    
End of search list.    
 /net/opt/gcc-4.0.0-20050410/libexec/gcc/i686-pc-linux-gnu/4.0.0/cc1plus -fpreprocessed    
triedriver.t.ii -quiet -dumpbase triedriver.t.cc -mtune=pentiumpro -auxbase-strip triedriver.t.o    
-gdwarf-2 -Wall -W -Winline -Woverloaded-virtual -version -o triedriver.t.s    
GNU C++ version 4.0.0 20050410 (prerelease) (i686-pc-linux-gnu)    
        compiled by GNU C version 3.3.4 (pre 3.3.5 20040809).    
GGC heuristics: --param ggc-min-expand=64 --param ggc-min-heapsize=64454    
g++: Internal error: Segmentation fault (program cc1plus)    
Please submit a full bug report.    
    
How to reproduce:  
  
g++ -Wall  -W -Winline -Woverloaded-virtual -I. -I/home/strieder/xssr/sys  
-I/home/strieder/xssr/sys/xssr-gui   -DINLINE='inline' -DDO_INLINE    -c -o triedriver.t.o  
triedriver.t.ii  
 
or 
 
g++ -c triedriver.t.ii  
  
or 
 
g++ -O2 -c triedriver.t.ii  
  
Regards,   
   
Bernd Strieder

-- 
           Summary: Virtual memory exhausted with g++
           Product: gcc
           Version: 4.0.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: strieder at informatik dot uni-kl dot de
                CC: gcc-bugs at gcc dot gnu dot org
  GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0/4.1 Regression] Virtual memory exhausted with g++

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-18 19:18 -------
On the mainline we don't get virtual memory exhausted but that is because there has been a patch or 
two to fix the memory usage but instead we get stack overflow:
#1207 0x002fc6e0 in fold_binary (code=TRUNC_MOD_EXPR, type=0x41d0c780, op0=0x26c9600, 
op1=0x26b95e0) at /Users/pinskia/src/alias/gcc/gcc/fold-const.c:8509
#1208 0x002fe288 in fold_build2 (code=TRUNC_MOD_EXPR, type=0x26c9600, op0=0x17, op1=0x7) 
at /Users/pinskia/src/alias/gcc/gcc/fold-const.c:10253
#1209 0x002fc6e0 in fold_binary (code=TRUNC_MOD_EXPR, type=0x41d0c780, op0=0x26c9600, 
op1=0x26b95c0) at /Users/pinskia/src/alias/gcc/gcc/fold-const.c:8509
#1210 0x002fe288 in fold_build2 (code=TRUNC_MOD_EXPR, type=0x26c9600, op0=0x17, op1=0x7) 
at /Users/pinskia/src/alias/gcc/gcc/fold-const.c:10253
#1211 0x002fc6e0 in fold_binary (code=TRUNC_MOD_EXPR, type=0x41d0c780, op0=0x26c9600, 
op1=0x26b95a0) at /Users/pinskia/src/alias/gcc/gcc/fold-const.c:8509


Let me try to get a reduced testcase.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c++                         |middle-end
           Keywords|                            |ice-on-valid-code
            Summary|Virtual memory exhausted    |[4.0/4.1 Regression] Virtual
                   |with g++                    |memory exhausted with g++
   Target Milestone|---                         |4.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0/4.1 Regression] Virtual memory exhausted with g++

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|                            |4.0.0 4.1.0
      Known to work|                            |3.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0/4.1 Regression] Virtual memory exhausted with g++

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-18 19:24 -------
Reduced testcase:
int main( int argc, char * argv[ ] )
{
  int maxstringlen = 1;
  int limit = 0, maxblock = 0, maxblockrem = 0;
  maxblockrem = (maxstringlen) % (2147483647 + 1);
}


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2005-04-18 19:24:03
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0/4.1 Regression] Virtual memory exhausted with g++

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-18 19:25 -------
(In reply to comment #2)
> Reduced testcase:

Even though there is an overflow which the C front-end says the code was in an if (0) block so it should 
not matter about that as it is only undefined then but we still should not be crashing.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0/4.1 Regression] Virtual memory exhausted with g++

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-18 19:38 -------
Most likely caused by:
2004-07-05  Roger Sayle  <roger@eyesopen.com>

        * fold-const.c (fold) <TRUNC_MOD_EXPR>: Optimize unsigned modulus
        by a power of two into a bit-wise AND, i.e. "X % C" as "X & (C-1)".
        Normalize "X % C" as "X % -C" for signed modulus and negative C.
        Optimize "X % -Y" as "X % Y" for signed modulus.
        <EQ_EXPR>: Recursively call "fold" when transforming "(X % Y) == 0"
        into "((unsigned) X % Y) == 0".

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |roger at eyesopen dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0/4.1 Regression] Virtual memory exhausted with g++

[phython at gcc dot gnu dot org]

------- Additional Comments From phython at gcc dot gnu dot org  2005-04-18 20:15 -------
Created an attachment (id=8678)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8678&action=view)
Don't alter overflowed constants

 sign_bit_p returns false for overflowed constants, this is bad for this
testcase.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |phython at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0/4.1 Regression] Virtual memory exhausted with g++

[roger at eyesopen dot com]

------- Additional Comments From roger at eyesopen dot com  2005-04-18 20:47 -------
Wow, that was fast.  I'd just started bootstrap and regression test on an
*identical* patch, and pulled up the PR to assign it to myself.  You win!
I'll preapprove yours, together with the testcase from the PR, if it passes
bootstrap and regression test without problems.

Thanks.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0/4.1 Regression] Virtual memory exhausted with g++

[ja2morri at csclub dot uwaterloo dot ca]

------- Additional Comments From ja2morri at csclub dot uwaterloo dot ca  2005-04-18 22:02 -------
Subject: Re:  [4.0/4.1 Regression] Virtual memory exhausted with g++


 Ahh, I posted the patch before starting the regtesting.  I only checked to
make sure it worked on the testcase.  I'll commit this patch after bootstrap
and regtesting finishes successfully on ia64-linux.



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0/4.1 Regression] Virtual memory exhausted with g++

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-04-19 02:36 -------
Subject: Bug 21085

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	phython@gcc.gnu.org	2005-04-19 02:36:04

Modified files:
	gcc            : ChangeLog fold-const.c 
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/gcc.dg: pr21085.c 

Log message:
	2005-04-18  James A. Morrison  <phython@gcc.gnu.org>
	
	PR tree-optimization/21085
	* fold-const (fold_binary): Don't change X % -C to X % C if C has
	overflowed.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.8355&r2=2.8356
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/fold-const.c.diff?cvsroot=gcc&r1=1.563&r2=1.564
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&r1=1.5368&r2=1.5369
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.dg/pr21085.c.diff?cvsroot=gcc&r1=NONE&r2=1.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0 only Regression] Virtual memory exhausted with g++

[phython at gcc dot gnu dot org]

------- Additional Comments From phython at gcc dot gnu dot org  2005-04-19 02:43 -------
Fixed in the previous commit.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED
            Summary|[4.0/4.1 Regression] Virtual|[4.0 only Regression]
                   |memory exhausted with g++   |Virtual memory exhausted
                   |                            |with g++
   Target Milestone|4.0.0                       |4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0 Regression] Virtual memory exhausted with g++

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-19 06:10 -------
This is still a bad bug which should be fixed for 4.0.1 at least.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |
            Summary|[4.0 only Regression]       |[4.0 Regression] Virtual
                   |Virtual memory exhausted    |memory exhausted with g++
                   |with g++                    |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0 Regression] Virtual memory exhausted with g++

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0 Regression] Virtual memory exhausted with g++

[phython at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|phython at gcc dot gnu dot  |unassigned at gcc dot gnu
                   |org                         |dot org
             Status|ASSIGNED                    |NEW
      Known to fail|4.0.0 4.1.0                 |4.0.0
      Known to work|3.4.0                       |3.4.0 4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0 Regression] Virtual memory exhausted with g++

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-04-20 13:52 -------
I will take this and will try to get it approved for 4.0.1.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |pinskia at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0 Regression] Virtual memory exhausted with g++

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-05-11 12:46 -------
Subject: Bug 21085

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-4_0-branch
Changes by:	pinskia@gcc.gnu.org	2005-05-11 12:46:09

Modified files:
	gcc            : ChangeLog fold-const.c tree-ssa-dse.c 

Log message:
	2005-05-11  James A. Morrison  <phython@gcc.gnu.org>
	
	PR tree-optimization/21085
	* fold-const (fold): Don't change X % -C to X % C if C has
	overflowed.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=2.7592.2.240&r2=2.7592.2.241
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/fold-const.c.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=1.517.2.7&r2=1.517.2.8
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-ssa-dse.c.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=2.17&r2=2.17.4.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0 Regression] Virtual memory exhausted with g++

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-05-11 12:50 -------
Subject: Bug 21085

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-4_0-branch
Changes by:	pinskia@gcc.gnu.org	2005-05-11 12:49:54

Modified files:
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/gcc.dg: pr21085.c 

Log message:
	2005-05-11  James A. Morrison  <phython@gcc.gnu.org>
	
	PR middle-end/21085
	* gcc.dg/pr21085.c: New test.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=1.5084.2.172&r2=1.5084.2.173
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.dg/pr21085.c.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=NONE&r2=1.1.6.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085

[Bug middle-end/21085] [4.0 Regression] Virtual memory exhausted with g++

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-11 12:49 -------
Fixed also for 4.0.1.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|4.1.0                       |4.0.1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21085