* [Bug tree-optimization/29446] [4.2 Regression] VRP ICE in compare_names
2006-10-12 22:37 [Bug fortran/29446] New: VRP ICE in compare_names kargl at gcc dot gnu dot org
@ 2006-10-13 3:19 ` pinskia at gcc dot gnu dot org
2006-10-13 3:34 ` pinskia at gcc dot gnu dot org
` (8 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2006-10-13 3:19 UTC (permalink / raw)
To: gcc-bugs
------- Comment #1 from pinskia at gcc dot gnu dot org 2006-10-13 03:19 -------
C testcase:
void f(int *n, int *m)
{
int i;
int ubound0;
int ubound1;
int stride2;
int offset3;
int size4;
int j;
int ubound5;
int size6;
int D919;
__SIZE_TYPE__ D920;
int D921;
unsigned int D922;
__SIZE_TYPE__ D923;
unsigned int D924;
ubound0 = *m;
stride2 = ubound0;
ubound1 = *n;
size4 = stride2 * ubound1;
D922 = size4 - 1;
int (*abc)[D922];
abc = _gfortran_internal_malloc (size4 * 4);
offset3 = ~stride2;
ubound5 = *n;
size6 = ubound5;
D919 = size6 - 1;
int (*xsd)[D919];
xsd = _gfortran_internal_malloc (size6 * 4);
i = 1;
if (i <= 2)
{
while (1)
{
{
_Bool D918;
j = 1;
if (j <= 3)
{
while (1)
{
{
_Bool D917;
if (i != j)
{
if (__builtin_expect (ubound0 <= 0, 0))
{
__builtin_abort ();
}
else
{
(void) 0;
}
if (__builtin_expect (ubound1 <= 0, 0))
{
__builtin_abort ();
}
else
{
(void) 0;
}
(*abc)[stride2 + offset3 + 1] = 0;
}
else
{
{
int D916;
if (__builtin_expect (ubound0 <= 0, 0))
{
__builtin_abort ();
}
else
{
(void) 0;
}
D916 = j;
if (__builtin_expect (D916 <= 0, 0))
{
__builtin_abort ();
}
else
{
(void) 0;
}
if (__builtin_expect (D916 > ubound1, 0))
{
__builtin_abort ();
}
else
{
(void) 0;
}
if (__builtin_expect (ubound5 <= 0, 0))
{
__builtin_abort ();
}
else
{
(void) 0;
}
(*abc)[D916 * stride2 + offset3 + 1] = (*xsd)[0];
}
}
D917 = j == 3;
j = j + 1;
if (D917) break; else (void) 0;
}
}
}
else
{
(void) 0;
}
// L4:;
D918 = i == 2;
i = i + 1;
if (D918) break; else (void) 0;
}
}
}
else
{
(void) 0;
}
//L2:;
_gfortran_internal_free ((void *) xsd);
_gfortran_internal_free ((void *) abc);
}
--
pinskia at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Severity|normal |critical
Status|UNCONFIRMED |NEW
Ever Confirmed|0 |1
Known to fail| |4.2.0
Known to work| |4.1.2
Last reconfirmed|0000-00-00 00:00:00 |2006-10-13 03:19:26
date| |
Summary|VRP ICE in compare_names |[4.2 Regression] VRP ICE in
| |compare_names
Target Milestone|--- |4.2.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29446
^ permalink raw reply [flat|nested] 11+ messages in thread* [Bug tree-optimization/29446] [4.2 Regression] VRP ICE in compare_names
2006-10-12 22:37 [Bug fortran/29446] New: VRP ICE in compare_names kargl at gcc dot gnu dot org
2006-10-13 3:19 ` [Bug tree-optimization/29446] [4.2 Regression] " pinskia at gcc dot gnu dot org
@ 2006-10-13 3:34 ` pinskia at gcc dot gnu dot org
2006-10-13 8:22 ` rguenth at gcc dot gnu dot org
` (7 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2006-10-13 3:34 UTC (permalink / raw)
To: gcc-bugs
------- Comment #2 from pinskia at gcc dot gnu dot org 2006-10-13 03:34 -------
here is a shorter testcase:
void f(void)
{
int i, ubound1, j, ubound5;
int (*abc)[3];
i = 1;
while (1)
{
if (j <= 3)
while (1)
{
_Bool D917;
if (i != j)
{
if (__builtin_expect (ubound1 <= 0, 0))
__builtin_abort ();
(*abc)[1] = 0;
}
else
{
int D916;
D916 = j;
if (__builtin_expect (D916 > ubound1, 0))
__builtin_abort ();
if (__builtin_expect (ubound5 <= 0, 0))
__builtin_abort ();
}
j = j + 1;
if (D917)
break;
}
i = i + 1;
}
}
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29446
^ permalink raw reply [flat|nested] 11+ messages in thread* [Bug tree-optimization/29446] [4.2 Regression] VRP ICE in compare_names
2006-10-12 22:37 [Bug fortran/29446] New: VRP ICE in compare_names kargl at gcc dot gnu dot org
2006-10-13 3:19 ` [Bug tree-optimization/29446] [4.2 Regression] " pinskia at gcc dot gnu dot org
2006-10-13 3:34 ` pinskia at gcc dot gnu dot org
@ 2006-10-13 8:22 ` rguenth at gcc dot gnu dot org
2006-10-13 9:19 ` rguenth at gcc dot gnu dot org
` (6 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: rguenth at gcc dot gnu dot org @ 2006-10-13 8:22 UTC (permalink / raw)
To: gcc-bugs
------- Comment #3 from rguenth at gcc dot gnu dot org 2006-10-13 08:22 -------
Slightly less undefined, ICEs with -O -ftree-vrp -funswitch-loops:
void f(_Bool D917, int j0, int ubound1, int ubound5)
{
int i, j = j0;
int (*abc)[3];
i = 1;
while (1)
{
if (j <= 3)
while (1)
{
if (i != j)
{
if (ubound1 <= 0)
return;
(*abc)[1] = 0;
}
else
{
if (j > ubound1)
return;
if (ubound5 <= 0)
return;
}
j = j + 1;
if (D917)
break;
}
i = i + 1;
}
}
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29446
^ permalink raw reply [flat|nested] 11+ messages in thread* [Bug tree-optimization/29446] [4.2 Regression] VRP ICE in compare_names
2006-10-12 22:37 [Bug fortran/29446] New: VRP ICE in compare_names kargl at gcc dot gnu dot org
` (2 preceding siblings ...)
2006-10-13 8:22 ` rguenth at gcc dot gnu dot org
@ 2006-10-13 9:19 ` rguenth at gcc dot gnu dot org
2006-10-13 10:07 ` rguenth at gcc dot gnu dot org
` (5 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: rguenth at gcc dot gnu dot org @ 2006-10-13 9:19 UTC (permalink / raw)
To: gcc-bugs
------- Comment #4 from rguenth at gcc dot gnu dot org 2006-10-13 09:18 -------
We end up comparing (GT_EXPR) i_92 with ubound1_111 which have the following
value ranges and equivalences:
i_92: [1, 3] EQUIVALENCES: { i_1 j_2 i_67 j_70 } (4 elements)
i_1: ~[0, 0] EQUIVALENCES: { } (0 elements)
j_2: VARYING
i_67: [1, ubound1_8] EQUIVALENCES: { i_1 } (1 elements)
j_70: [-INF, 3] EQUIVALENCES: { j_2 } (1 elements)
ubound1_111: [-INF, 0] EQUIVALENCES: { ubound1_8 ubound1_112 } (2 elements)
ubound1_8: VARYING
ubound1_112: [i_67, +INF] EQUIVALENCES: { ubound1_8 } (1 elements)
i_67 > ubound1_8 ([1, ubound1_8] > [ubound1_8, ubound1_8]) is false
i_67 > ubound1_111 ([1, ubound1_8] > [-INF, 0]) is true
i_92 > ubound1_111 ([1, 3] > [-INF, 0]) is true
So we have a cross-equivalency set inconsistency.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29446
^ permalink raw reply [flat|nested] 11+ messages in thread* [Bug tree-optimization/29446] [4.2 Regression] VRP ICE in compare_names
2006-10-12 22:37 [Bug fortran/29446] New: VRP ICE in compare_names kargl at gcc dot gnu dot org
` (3 preceding siblings ...)
2006-10-13 9:19 ` rguenth at gcc dot gnu dot org
@ 2006-10-13 10:07 ` rguenth at gcc dot gnu dot org
2006-10-13 10:39 ` rguenth at gcc dot gnu dot org
` (4 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: rguenth at gcc dot gnu dot org @ 2006-10-13 10:07 UTC (permalink / raw)
To: gcc-bugs
------- Comment #5 from rguenth at gcc dot gnu dot org 2006-10-13 10:07 -------
It looks like we would need to recurse down into symbolic ranges in
fix_equivalence_set, which would be far too costly. So a conservative
fix for 4.2 is to just not record symbolic equivalencies at all.
I wonder if there is a better way to deal with these issues than
fix_equivalence_set though.
--
rguenth at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |dnovillo at redhat dot com
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29446
^ permalink raw reply [flat|nested] 11+ messages in thread* [Bug tree-optimization/29446] [4.2 Regression] VRP ICE in compare_names
2006-10-12 22:37 [Bug fortran/29446] New: VRP ICE in compare_names kargl at gcc dot gnu dot org
` (4 preceding siblings ...)
2006-10-13 10:07 ` rguenth at gcc dot gnu dot org
@ 2006-10-13 10:39 ` rguenth at gcc dot gnu dot org
2006-10-13 12:45 ` patchapp at dberlin dot org
` (3 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: rguenth at gcc dot gnu dot org @ 2006-10-13 10:39 UTC (permalink / raw)
To: gcc-bugs
------- Comment #6 from rguenth at gcc dot gnu dot org 2006-10-13 10:38 -------
I have a patch to get rid of that beast completely.
--
rguenth at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
AssignedTo|unassigned at gcc dot gnu |rguenth at gcc dot gnu dot
|dot org |org
Status|NEW |ASSIGNED
Last reconfirmed|2006-10-13 03:19:26 |2006-10-13 10:38:52
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29446
^ permalink raw reply [flat|nested] 11+ messages in thread* [Bug tree-optimization/29446] [4.2 Regression] VRP ICE in compare_names
2006-10-12 22:37 [Bug fortran/29446] New: VRP ICE in compare_names kargl at gcc dot gnu dot org
` (5 preceding siblings ...)
2006-10-13 10:39 ` rguenth at gcc dot gnu dot org
@ 2006-10-13 12:45 ` patchapp at dberlin dot org
2006-10-13 20:09 ` rguenth at gcc dot gnu dot org
` (2 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: patchapp at dberlin dot org @ 2006-10-13 12:45 UTC (permalink / raw)
To: gcc-bugs
------- Comment #7 from patchapp at dberlin dot org 2006-10-13 12:45 -------
Subject: Bug number PR29446
A patch for this bug has been added to the patch tracker.
The mailing list url for the patch is
http://gcc.gnu.org/ml/gcc-patches/2006-10/msg00698.html
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29446
^ permalink raw reply [flat|nested] 11+ messages in thread* [Bug tree-optimization/29446] [4.2 Regression] VRP ICE in compare_names
2006-10-12 22:37 [Bug fortran/29446] New: VRP ICE in compare_names kargl at gcc dot gnu dot org
` (6 preceding siblings ...)
2006-10-13 12:45 ` patchapp at dberlin dot org
@ 2006-10-13 20:09 ` rguenth at gcc dot gnu dot org
2006-10-13 20:09 ` rguenth at gcc dot gnu dot org
2007-04-25 16:10 ` rguenth at gcc dot gnu dot org
9 siblings, 0 replies; 11+ messages in thread
From: rguenth at gcc dot gnu dot org @ 2006-10-13 20:09 UTC (permalink / raw)
To: gcc-bugs
------- Comment #8 from rguenth at gcc dot gnu dot org 2006-10-13 20:09 -------
Subject: Bug 29446
Author: rguenth
Date: Fri Oct 13 20:09:10 2006
New Revision: 117705
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=117705
Log:
2006-10-13 Richard Guenther <rguenther@suse.de>
PR tree-optimization/29446
* tree-vrp.c (fix_equivalence_set): Remove.
(extract_range_from_assert): Do not call fix_equivalence_set.
(debug_value_range): Print a newline.
(compare_name_with_value): For equivalence sets with
inconsistent value ranges conservatively bail out.
(compare_names): Likewise.
* gcc.dg/torture/pr29446.c: New testcase.
Added:
trunk/gcc/testsuite/gcc.dg/torture/pr29446.c
Modified:
trunk/gcc/ChangeLog
trunk/gcc/testsuite/ChangeLog
trunk/gcc/tree-vrp.c
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29446
^ permalink raw reply [flat|nested] 11+ messages in thread* [Bug tree-optimization/29446] [4.2 Regression] VRP ICE in compare_names
2006-10-12 22:37 [Bug fortran/29446] New: VRP ICE in compare_names kargl at gcc dot gnu dot org
` (7 preceding siblings ...)
2006-10-13 20:09 ` rguenth at gcc dot gnu dot org
@ 2006-10-13 20:09 ` rguenth at gcc dot gnu dot org
2007-04-25 16:10 ` rguenth at gcc dot gnu dot org
9 siblings, 0 replies; 11+ messages in thread
From: rguenth at gcc dot gnu dot org @ 2006-10-13 20:09 UTC (permalink / raw)
To: gcc-bugs
------- Comment #9 from rguenth at gcc dot gnu dot org 2006-10-13 20:09 -------
Fixed.
--
rguenth at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|ASSIGNED |RESOLVED
Resolution| |FIXED
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29446
^ permalink raw reply [flat|nested] 11+ messages in thread* [Bug tree-optimization/29446] [4.2 Regression] VRP ICE in compare_names
2006-10-12 22:37 [Bug fortran/29446] New: VRP ICE in compare_names kargl at gcc dot gnu dot org
` (8 preceding siblings ...)
2006-10-13 20:09 ` rguenth at gcc dot gnu dot org
@ 2007-04-25 16:10 ` rguenth at gcc dot gnu dot org
9 siblings, 0 replies; 11+ messages in thread
From: rguenth at gcc dot gnu dot org @ 2007-04-25 16:10 UTC (permalink / raw)
To: gcc-bugs
------- Comment #10 from rguenth at gcc dot gnu dot org 2007-04-25 17:10 -------
Subject: Bug 29446
Author: rguenth
Date: Wed Apr 25 17:10:31 2007
New Revision: 124158
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=124158
Log:
2007-04-25 Richard Guenther <rguenther@suse.de>
PR tree-optimization/31698
* g++.dg/other/pr31698.C: New testcase.
Backport from mainline:
2006-10-13 Richard Guenther <rguenther@suse.de>
PR tree-optimization/29446
* tree-vrp.c (fix_equivalence_set): Remove.
(extract_range_from_assert): Do not call fix_equivalence_set.
(debug_value_range): Print a newline.
(compare_name_with_value): For equivalence sets with
inconsistent value ranges conservatively bail out.
(compare_names): Likewise.
* gcc.dg/torture/pr29446.c: New testcase.
Added:
branches/gcc-4_1-branch/gcc/testsuite/g++.dg/other/pr31698.C
branches/gcc-4_1-branch/gcc/testsuite/gcc.dg/torture/pr29446.c
- copied unchanged from r117705,
trunk/gcc/testsuite/gcc.dg/torture/pr29446.c
Modified:
branches/gcc-4_1-branch/gcc/ChangeLog
branches/gcc-4_1-branch/gcc/testsuite/ChangeLog
branches/gcc-4_1-branch/gcc/tree-vrp.c
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29446
^ permalink raw reply [flat|nested] 11+ messages in thread