[Bug c/17560] New: Segmentation fault compiling prdtoa.c with -Os

[Bug c/17560] New: Segmentation fault compiling prdtoa.c with -Os

[kherron+gcc at fmailbox dot com]

Gcc was built from cvs this morning, configured as follows:

../gcc/configure --prefix=/home/kherron/gcc --program-prefix=cvs- \
        --with-cpu=athlon --with-arch=athlon --enable-threads \
        --enable-languages='c,c++'

prdtoa.c is part of NSPR, which is part of mozilla. Compiling with this command:

/home/kherron/gcc/bin/cvs-gcc -o prdtoa.o -c -save-temps -ansi -Wall -pthread
-Os -ffast-math -g -fno-inline -fPIC  -UNDEBUG -DDEBUG_kherron 
-DMOZILLA_CLIENT=1 -DDEBUG=1 -DXP_UNIX=1 -D_POSIX_SOURCE=1 -D_BSD_SOURCE=1
-D_SVID_SOURCE=1 -D_LARGEFILE64_SOURCE=1 -DHAVE_FCNTL_FILE_LOCKING=1 -DLINUX=1
-Di386=1 -DHAVE_LCHOWN=1 -DHAVE_STRERROR=1 -D_REENTRANT=1  -DFORCE_PR_LOG
-D_PR_PTHREADS -UHAVE_CVAR_BUILT_ON_SEM -D_NSPR_BUILD_
-I/extra/kherron/moz/gcc4/dist/include/nspr
-I/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/include
-I/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/include/private 
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c

produces a segfault:

/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c: In function
'Balloc':
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:588: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c: In function 'mult':
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:883: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:897: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c: In function
'pow5mult':
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:949: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c: In function
'lshift':
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1030: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c: In function 'b2d':
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1252: warning:
suggest parentheses around + or - inside shift
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1254: warning:
suggest parentheses around + or - inside shift
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1254: warning:
suggest parentheses around + or - inside shift
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1259: warning:
suggest parentheses around + or - inside shift
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1261: warning:
suggest parentheses around + or - inside shift
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c: In function 'd2b':
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1332: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1336: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1337: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1338: warning:
suggest parentheses around + or - inside shift
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c: In function
'PR_strtod':
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1904: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1964: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:1983: warning:
suggest parentheses around + or - inside shift
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c: In function
'nrv_alloc':
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:2647: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c: In function 'dtoa':
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:2827: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:2871: warning:
suggest parentheses around + or - inside shift
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:2872: warning:
suggest parentheses around + or - inside shift
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:2872: warning:
suggest parentheses around + or - inside shift
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:2981: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:3095: warning:
suggest parentheses around && within ||
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:3142: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:3180: warning:
suggest parentheses around assignment used as truth value
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:3269: warning:
suggest parentheses around && within ||
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:3286: warning:
suggest parentheses around && within ||
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:3346: warning:
suggest parentheses around && within ||
/extra/kherron/moz/gcc4/../mozilla/nsprpub/pr/src/misc/prdtoa.c:3357: warning:
label 'trimzeros' defined but not used
cvs-gcc: Internal error: Segmentation fault (program cc1)
Please submit a full bug report.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.

When I remove "-Os" from the command line, or change it to something else like
-O3, gcc compiles the file successfully.

-- 
           Summary: Segmentation fault compiling prdtoa.c with -Os
           Product: gcc
           Version: 4.0.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: kherron+gcc at fmailbox dot com
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: i686-pc-linux-gnu
  GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug c/17560] Segmentation fault compiling prdtoa.c with -Os

[kherron+gcc at fmailbox dot com]

------- Additional Comments From kherron+gcc at fmailbox dot com  2004-09-19 16:53 -------
Created an attachment (id=7172)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7172&action=view)
prdtoa.i


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug c/17560] Segmentation fault compiling prdtoa.c with -Os

[kherron+gcc at fmailbox dot com]

------- Additional Comments From kherron+gcc at fmailbox dot com  2004-09-19 16:55 -------
Created an attachment (id=7173)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7173&action=view)
prdtoa.s

This is the assembler produced before the crash.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Segmentation fault compiling prdtoa.c with -Os

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-19 19:19 -------
This is a bug in scalar evolution code:
#0  0x004d00e4 in find_var_scev_info (var=0x0) at /Volumes/UFS_Partition/pinskia/src/devel/
speedup/gcc/gcc/tree-scalar-evolution.c:338
#1  0x004d3340 in analyze_scalar_evolution (loop=0x41364940, var=0xaebf54) at /Volumes/
UFS_Partition/pinskia/src/devel/speedup/gcc/gcc/tree-scalar-evolution.c:1913
#2  0x004d2cb8 in interpret_rhs_modify_expr (loop=0x41364940, opnd1=0x9514ec, 
type=0x4127f64c) at /Volumes/UFS_Partition/pinskia/src/devel/speedup/gcc/gcc/tree-scalar-
evolution.c:1730


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |spop at gcc dot gnu dot org
           Severity|normal                      |critical
          Component|c                           |tree-optimization
           Keywords|                            |ice-on-valid-code
            Summary|Segmentation fault compiling|[4.0 Regression]
                   |prdtoa.c with -Os           |Segmentation fault compiling
                   |                            |prdtoa.c with -Os
   Target Milestone|---                         |4.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Segmentation fault compiling prdtoa.c with -Os

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-09-20 02:00 -------
Here's a reduced testcase:

================================
void foo()
{
    int i, j;

    if (i)
    {
        i++;
        goto L;
    }
    for (; i-1; )
    {
      L:
        for (j=0; j<i; j++) ;
    }
}
================================

Just compile with "gcc -Os" to see the crash.
We seem to run into an infinite loop. The compiler runs out of memory
rather quickly.


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |reichelt at gcc dot gnu dot
                   |                            |org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
           Keywords|                            |monitored
   Last reconfirmed|0000-00-00 00:00:00         |2004-09-20 02:00:53
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Segmentation fault compiling prdtoa.c with -Os

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-09-20 02:02 -------
It's a recent regression. Phils's regression hunter reports:
: Search converges between 2004-09-12-trunk (#538) and 2004-09-14-trunk (#539).

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Segmentation fault compiling prdtoa.c with -Os

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-20 02:08 -------
I should note that is the date when IV-OPTs come into which is what is causing the crash by calling 
scalar evolution code.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Segmentation fault compiling prdtoa.c with -Os

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-27 21:29 -------
This is not fixed with the patch for 17520 either.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Segmentation fault compiling prdtoa.c with -Os

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-10-07 18:55 -------
*** Bug 17883 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |echristo at redhat dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Segmentation fault compiling prdtoa.c with -Os

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-07 21:31 -------
*** Bug 17607 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |loki at inf dot u-szeged dot
                   |                            |hu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Segmentation fault compiling prdtoa.c with -Os

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-07 22:54 -------
*** Bug 17704 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tobi at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-10-08 01:16 -------
Jan, your patch
http://gcc.gnu.org/ml/gcc-cvs/2004-09/msg00629.html
seems to be responsible for the regression.

Could you please have a look?


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hubicka at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-08 03:27 -------
Even though Jan's patch introduces the failure, this is a latent bug in the scalar evolution code.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[sebastian dot pop at cri dot ensmp dot fr]

------- Additional Comments From sebastian dot pop at cri dot ensmp dot fr  2004-10-08 12:07 -------
Subject: Re:  [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

Here is what the code looks like at the level of my analyzer.

loop_0
{
  bb_0 (preds = {bb_-1}, succs = {bb_0bb_0})
  {
  <bb 0>:
    if (i_4 != 0) goto <L0>; else goto <L6>;

  }
  bb_1 (preds = {bb_0}, succs = {bb_1})
  {
  <L0>:;
    i_7 = i_4 + 1;

  }
  bb_2 (preds = {bb_5bb_1}, succs = {bb_2})
  {
    # i_1 = PHI <i_2(5), i_7(1)>;
  L:;
    goto <bb 4> (<L5>);

  }
  bb_5 (preds = {bb_0bb_4}, succs = {bb_5bb_5})
  {
    # i_2 = PHI <i_4(0), i_1(4)>;
  <L6>:;
    if (i_2 != 1) goto L; else goto <L8>;

  }
  bb_6 (preds = {bb_5}, succs = {bb_6})
  {
  <L8>:;
    return;

  }
  loop_1
  {
    bb_3 (preds = {bb_4}, succs = {bb_3})
    {
    <L4>:;
      j_6 = j_3 + 1;

    }
    bb_4 (preds = {bb_3bb_2}, succs = {bb_4bb_4})
    {
      # j_3 = PHI <0(2), j_6(3)>;
    <L5>:;
      if (j_3 < i_1) goto <L4>; else goto <L6>;

    }
  }
}

The problem is that we have a cycle "i_1 -> i_2 -> i_1" outside any
loop, and I'm not sure about the semantics of this representation.  
By the way, is this a valid SSA form?



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[hubicka at ucw dot cz]

------- Additional Comments From hubicka at ucw dot cz  2004-10-08 12:16 -------
Subject: Re:  [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

> 
> ------- Additional Comments From sebastian dot pop at cri dot ensmp dot fr  2004-10-08 12:07 -------
> Subject: Re:  [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os
> 
> Here is what the code looks like at the level of my analyzer.
> 
> loop_0
> {
>   bb_0 (preds = {bb_-1}, succs = {bb_0bb_0})
>   {
>   <bb 0>:
>     if (i_4 != 0) goto <L0>; else goto <L6>;
> 
>   }
>   bb_1 (preds = {bb_0}, succs = {bb_1})
>   {
>   <L0>:;
>     i_7 = i_4 + 1;
> 
>   }
>   bb_2 (preds = {bb_5bb_1}, succs = {bb_2})
>   {
>     # i_1 = PHI <i_2(5), i_7(1)>;
>   L:;
>     goto <bb 4> (<L5>);
> 
>   }
>   bb_5 (preds = {bb_0bb_4}, succs = {bb_5bb_5})
>   {
>     # i_2 = PHI <i_4(0), i_1(4)>;
>   <L6>:;
>     if (i_2 != 1) goto L; else goto <L8>;
> 
>   }
>   bb_6 (preds = {bb_5}, succs = {bb_6})
>   {
>   <L8>:;
>     return;
> 
>   }
>   loop_1
>   {
>     bb_3 (preds = {bb_4}, succs = {bb_3})
>     {
>     <L4>:;
>       j_6 = j_3 + 1;
> 
>     }
>     bb_4 (preds = {bb_3bb_2}, succs = {bb_4bb_4})
>     {
>       # j_3 = PHI <0(2), j_6(3)>;
>     <L5>:;
>       if (j_3 < i_1) goto <L4>; else goto <L6>;
> 
>     }
>   }
> }
> 
> The problem is that we have a cycle "i_1 -> i_2 -> i_1" outside any
> loop, and I'm not sure about the semantics of this representation.  
> By the way, is this a valid SSA form?

It seems to me that this can definitly happen in the case of irreducible
regions...

Honza
> 
> 
> 
> -- 
> 
> 
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560
> 
> ------- You are receiving this mail because: -------
> You are on the CC list for the bug, or are watching someone who is.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[dnovillo at redhat dot com]

------- Additional Comments From dnovillo at redhat dot com  2004-10-08 12:39 -------
Subject: Re:  [4.0 Regression] Infinite
	recursion in tree-scalar-evolution with -Os

On Fri, 2004-10-08 at 08:07, sebastian dot pop at cri dot ensmp dot fr
wrote:

> The problem is that we have a cycle "i_1 -> i_2 -> i_1" outside any
> loop, and I'm not sure about the semantics of this representation.  
> By the way, is this a valid SSA form?
> 
It is, but this graph is irreducible.  Block #3 is only reachable via a
back edge.  Or put it another way, block #4 is the entry block for two
different loops.


Diego.



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

Re: [Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[Diego Novillo]

On Fri, 2004-10-08 at 08:07, sebastian dot pop at cri dot ensmp dot fr
wrote:

> The problem is that we have a cycle "i_1 -> i_2 -> i_1" outside any
> loop, and I'm not sure about the semantics of this representation.  
> By the way, is this a valid SSA form?
> 
It is, but this graph is irreducible.  Block #3 is only reachable via a
back edge.  Or put it another way, block #4 is the entry block for two
different loops.


Diego.

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[sebastian dot pop at cri dot ensmp dot fr]

------- Additional Comments From sebastian dot pop at cri dot ensmp dot fr  2004-10-08 13:03 -------
Subject: Re:  [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

> It seems to me that this can definitly happen in the case of irreducible
> regions...
> 
> Honza

Then the following patch solves the problem by avoiding the analysis
of these cycles.


Index: tree-scalar-evolution.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/tree-scalar-evolution.c,v
retrieving revision 2.8
diff -c -c -3 -p -r2.8 tree-scalar-evolution.c
*** tree-scalar-evolution.c	1 Oct 2004 09:05:57 -0000	2.8
--- tree-scalar-evolution.c	8 Oct 2004 12:59:37 -0000
*************** tree chrec_dont_know;
*** 284,289 ****
--- 284,290 ----
  tree chrec_known;
  
  static bitmap already_instantiated;
+ static bitmap already_analyzed;
  
  static htab_t scalar_evolution_info;
  
*************** follow_ssa_edge_in_condition_phi (struct
*** 1398,1403 ****
--- 1399,1409 ----
  
    for (i = 1; i < PHI_NUM_ARGS (condition_phi); i++)
      {
+       /* Quickly give up when the evolution of one of the branches is
+ 	 not known.  */
+       if (*evolution_of_loop == chrec_dont_know)
+ 	return true;
+ 
        if (!follow_ssa_edge_in_condition_phi_branch (i, loop, condition_phi,
  						    halting_phi,
  						    &evolution_of_branch,
*************** analyze_scalar_evolution_1 (struct loop 
*** 1833,1838 ****
--- 1839,1853 ----
        goto set_and_end;
      }
  
+   /* Fail on irreducible CFGs.  */
+   if (bitmap_bit_p (already_analyzed, SSA_NAME_VERSION (var)))
+     {
+       res = chrec_dont_know;
+       goto set_and_end;
+     }
+   else
+     bitmap_set_bit (already_analyzed, SSA_NAME_VERSION (var));
+ 
    if (res != chrec_not_analyzed_yet)
      {
        if (loop != bb->loop_father)
*************** analyze_scalar_evolution_1 (struct loop 
*** 1870,1875 ****
--- 1885,1892 ----
  
   set_and_end:
  
+   bitmap_clear_bit (already_analyzed, SSA_NAME_VERSION (var));
+ 
    /* Keep the symbolic form.  */
    if (res == chrec_dont_know)
      res = var;
*************** scev_initialize (struct loops *loops)
*** 2450,2455 ****
--- 2467,2473 ----
    scalar_evolution_info = htab_create (100, hash_scev_info,
  				       eq_scev_info, del_scev_info);
    already_instantiated = BITMAP_XMALLOC ();
+   already_analyzed = BITMAP_XMALLOC ();
    
    initialize_scalar_evolutions_analyzer ();
  
*************** scev_finalize (void)
*** 2545,2549 ****
--- 2563,2568 ----
  {
    htab_delete (scalar_evolution_info);
    BITMAP_XFREE (already_instantiated);
+   BITMAP_XFREE (already_analyzed);
  }
  


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

Re: [Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[Diego Novillo]

On Fri, 2004-10-08 at 09:03, sebastian dot pop at cri dot ensmp dot fr
wrote:

> Then the following patch solves the problem by avoiding the analysis
> of these cycles.
> 
You could also try splitting blocks, but if there's a cheaper work
around for 4.0, that'd be safer.


Diego.

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[dnovillo at redhat dot com]

------- Additional Comments From dnovillo at redhat dot com  2004-10-08 13:09 -------
Subject: Re:  [4.0 Regression] Infinite
	recursion in tree-scalar-evolution with -Os

On Fri, 2004-10-08 at 09:03, sebastian dot pop at cri dot ensmp dot fr
wrote:

> Then the following patch solves the problem by avoiding the analysis
> of these cycles.
> 
You could also try splitting blocks, but if there's a cheaper work
around for 4.0, that'd be safer.


Diego.



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[rakdver at gcc dot gnu dot org]

------- Additional Comments From rakdver at gcc dot gnu dot org  2004-10-08 13:21 -------
This fix should not be necessary (in fact we had the code like this before, but
I removed it and replaced it by current solution; the problem with is that it
behaves really weird, for example the results are dependent on in which order we
process the variables) -- we check for irreducible regions in scev already. 
Could you check whether the edges are marked as irreducible? If not, this would
be the problem.

Zdenek

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[sebastian dot pop at cri dot ensmp dot fr]

------- Additional Comments From sebastian dot pop at cri dot ensmp dot fr  2004-10-08 13:49 -------
Subject: Re:  [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

> Could you check whether the edges are marked as irreducible? If not, this would
> be the problem.
> 

Indeed, both edges for the phi "i_2 = phi (i_4, i_1)" are not marked
by EDGE_IRREDUCIBLE_LOOP, and consequently backedge_phi_arg_p returns
always false.  Same for the first edge i_1 -> i_2 of "i_1 = phi (i_2,
i_7)" that is not marked as EDGE_IRREDUCIBLE_LOOP.



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[rakdver at gcc dot gnu dot org]

------- Additional Comments From rakdver at gcc dot gnu dot org  2004-10-08 20:17 -------
Then this should be the proper fix.  I will commit it as obvious once it
passes regtesting.

Index: predict.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/predict.c,v
retrieving revision 1.128
diff -c -3 -p -r1.128 predict.c
*** predict.c   28 Sep 2004 07:59:48 -0000      1.128
--- predict.c   8 Oct 2004 20:15:44 -0000
*************** tree_estimate_probability (void)
*** 1300,1305 ****
--- 1300,1306 ----

    tree_bb_level_predictions ();

+   mark_irreducible_loops (&loops_info);
    predict_loops (&loops_info, false);

    FOR_EACH_BB (bb)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[rakdver at gcc dot gnu dot org]

------- Additional Comments From rakdver at gcc dot gnu dot org  2004-10-09 08:00 -------
Patch:

http://gcc.gnu.org/ml/gcc-patches/2004-10/msg00810.html

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-10-11 10:28 -------
*** Bug 17928 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |bryner at brianryner dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-10-16 16:45 -------
Subject: Bug 17560

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	rakdver@gcc.gnu.org	2004-10-16 16:45:38

Modified files:
	gcc            : ChangeLog predict.c 

Log message:
	PR tree-optimization/17560
	* predict.c (tree_estimate_probability): Mark irreducible
	loops.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.5905&r2=2.5906
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/predict.c.diff?cvsroot=gcc&r1=1.128&r2=1.129



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[rakdver at gcc dot gnu dot org]

------- Additional Comments From rakdver at gcc dot gnu dot org  2004-10-16 16:46 -------
Fixed.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560

[Bug tree-optimization/17560] [4.0 Regression] Infinite recursion in tree-scalar-evolution with -Os

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-10-17 19:16 -------
*** Bug 17552 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|17552                       |
                 CC|                            |jakub at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17560