[Bug regression/54084] New: Bunch of fails for x86

[Bug regression/54084] New: Bunch of fails for x86

[izamyatin at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54084

             Bug #: 54084
           Summary: Bunch of fails for x86
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: regression
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: izamyatin@gmail.com
            Target: x86


After r189803 (http://gcc.gnu.org/ml/gcc-cvs/2012-07/msg00708.html) regression
mailer shows bunch of errors for x86_64:
FAIL: g++.dg/opt/pr44919.C -std=gnu++11 (internal compiler error)
FAIL: g++.dg/opt/pr46640.C -std=gnu++11 (internal compiler error)
FAIL: g++.dg/opt/pr46640.C -std=gnu++11 (test for excess errors)
FAIL: g++.dg/opt/pr46640.C -std=gnu++98 (internal compiler error)
FAIL: gcc.dg/pr45352-3.c (internal compiler error)
FAIL: gcc.dg/pr45352.c (internal compiler error)
FAIL: gcc.dg/pr45570.c (internal compiler error)
FAIL: gcc.dg/pr46521.c (internal compiler error)
FAIL: gcc.dg/pr46522.c (internal compiler error)
FAIL: gcc.dg/pr48144.c (internal compiler error)
FAIL: gcc.dg/pr48235.c (internal compiler error)
FAIL: gcc.dg/pr51557.c (internal compiler error)
FAIL: gcc.target/i386/pr45352-1.c (internal compiler error)
FAIL: gcc.target/i386/pr45352.c (internal compiler error)
FAIL: gcc.target/i386/pr45352.c (test for excess errors)
FAIL: gfortran.dg/pr44691.f  -O  (internal compiler error)


and i686:
FAIL: gcc.dg/pr45570.c (internal compiler error)
FAIL: gcc.dg/pr46585.c (internal compiler error)
FAIL: gcc.dg/pr48144.c (internal compiler error)
FAIL: gcc.dg/pr48235.c (internal compiler error)
FAIL: gcc.dg/pr48442.c (internal compiler error)
FAIL: gcc.dg/pr50205.c (internal compiler error)
FAIL: gcc.dg/pr50340.c (internal compiler error)
FAIL: gcc.dg/pr51557.c (internal compiler error)
FAIL: gcc.target/i386/pr45352-1.c (internal compiler error)
FAIL: gcc.target/i386/pr45352-2.c (internal compiler error)
FAIL: gcc.target/i386/pr45352.c (internal compiler error)
FAIL: gcc.target/i386/pr45913.c (internal compiler error)
FAIL: gcc.target/i386/pr48237.c (internal compiler error)

Error message looks like
internal compiler error: in free_regset_pool, at sel-sched-ir.c:994^M
Please submit a full bug report,^M
with preprocessed source if appropriate.^M
See <http://gcc.gnu.org/bugs.html> for instructions.^M
 status 1

[Bug regression/54084] Bunch of fails for x86

[ubizjak at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54084

Uros Bizjak <ubizjak at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2012-07-24
                 CC|                            |stevenb.gcc at gmail dot
                   |                            |com
   Target Milestone|---                         |4.8.0
     Ever Confirmed|0                           |1

--- Comment #1 from Uros Bizjak <ubizjak at gmail dot com> 2012-07-24 14:05:48 UTC ---
Confirmed.

[Bug regression/54084] Bunch of fails for x86

[steven at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54084

Steven Bosscher <steven at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
                 CC|                            |steven at gcc dot gnu.org
         AssignedTo|unassigned at gcc dot       |steven at gcc dot gnu.org
                   |gnu.org                     |

--- Comment #2 from Steven Bosscher <steven at gcc dot gnu.org> 2012-07-24 15:31:50 UTC ---
I suspect this is a 64-bit pointer vs. 32-bit integer issue.

Could you please test with this patch? 

Index: sel-sched-ir.c
===================================================================
--- sel-sched-ir.c      (revision 189808)
+++ sel-sched-ir.c      (working copy)
@@ -954,7 +954,9 @@ return_regset_to_pool (regset rs)
 static int
 cmp_v_in_regset_pool (const void *x, const void *xx)
 {
-  return *((const regset *) x) - *((const regset *) xx);
+  intptr_t d = (intptr_t) *((const regset *) x);
+  intptr_t dd = (intptr_t) *((const regset *) xx);
+  return d - dd;
 }
 #endif

[Bug regression/54084] Bunch of fails for x86

[izamyatin at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54084

--- Comment #3 from Igor Zamyatin <izamyatin at gmail dot com> 2012-07-24 17:16:11 UTC ---
Seems ok now

[Bug regression/54084] Bunch of fails for x86

[wschmidt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54084

William J. Schmidt <wschmidt at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |wschmidt at gcc dot gnu.org

--- Comment #4 from William J. Schmidt <wschmidt at gcc dot gnu.org> 2012-07-25 15:55:07 UTC ---
I'm seeing this problem today on powerpc64 as well.

[Bug regression/54084] Bunch of fails for x86

[izamyatin at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54084

--- Comment #5 from Igor Zamyatin <izamyatin at gmail dot com> 2012-07-26 08:44:01 UTC ---
Looks like r189812 fixed some failures but not all of them.

Patch from comment 2 fixes all problems

[Bug regression/54084] Bunch of fails for x86

[steven at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54084

Steven Bosscher <steven at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |burnus at gcc dot gnu.org

--- Comment #6 from Steven Bosscher <steven at gcc dot gnu.org> 2012-07-26 11:59:39 UTC ---
*** Bug 54099 has been marked as a duplicate of this bug. ***

[Bug regression/54084] Bunch of fails for x86

[steven at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54084

--- Comment #7 from Steven Bosscher <steven at gcc dot gnu.org> 2012-07-26 12:08:31 UTC ---
This is the variant of the patch that I will commit after testing:

Index: sel-sched-ir.c
===================================================================
--- sel-sched-ir.c      (revision 189887)
+++ sel-sched-ir.c      (working copy)
@@ -954,7 +954,13 @@ return_regset_to_pool (regset rs)
 static int
 cmp_v_in_regset_pool (const void *x, const void *xx)
 {
-  return *((const regset *) x) - *((const regset *) xx);
+  uintptr_t r1 = (uintptr_t) *((const regset *) x);
+  uintptr_t r2 = (uintptr_t) *((const regset *) xx);
+  if (r1 > r2)
+    return 1;
+  else if (r1 < r2)
+    return -1;
+  gcc_unreachable ();
 }
 #endif

[Bug regression/54084] Bunch of fails for x86

[steven at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54084

--- Comment #8 from Steven Bosscher <steven at gcc dot gnu.org> 2012-07-26 13:21:28 UTC ---
Author: steven
Date: Thu Jul 26 13:21:21 2012
New Revision: 189891

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=189891
Log:
    PR regression/54084
    * sel-sched-ir.c (cmp_v_in_regset_pool): Clarify logic, fix
    pointer difference check.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/sel-sched-ir.c

[Bug regression/54084] Bunch of fails for x86

[steven at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54084

Steven Bosscher <steven at gcc dot gnu.org> changed:

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

--- Comment #9 from Steven Bosscher <steven at gcc dot gnu.org> 2012-07-26 13:24:03 UTC ---
.