[Bug tree-optimization/55260] New: [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[Bug tree-optimization/55260] New: [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[zsojka at seznam dot cz]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

             Bug #: 55260
           Summary: [4.8 Regression] ICE: in ipa_get_parm_lattices, at
                    ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: zsojka@seznam.cz


Created attachment 28650
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=28650
reduced testcase

Compiler output:
$ gcc -O2 -std=gnu++0x -fno-inline -fipa-cp-clone testcase.C 
testcase.C:20:1: internal compiler error: in ipa_get_parm_lattices, at
ipa-cp.c:263
 }
 ^
0x59033c ipa_get_parm_lattices
        /mnt/svn/gcc-trunk/gcc/ipa-cp.c:263
0x129c1e5 ipa_get_parm_lattices
        /mnt/svn/gcc-trunk/gcc/ipa-cp.c:2908
0x129c1e5 cgraph_edge_brings_all_agg_vals_for_node
        /mnt/svn/gcc-trunk/gcc/ipa-cp.c:3082
0x129c1e5 perhaps_add_new_callers
        /mnt/svn/gcc-trunk/gcc/ipa-cp.c:3134
0x129c1e5 decide_about_value
        /mnt/svn/gcc-trunk/gcc/ipa-cp.c:3207
0x129dfd0 decide_whether_version_node
        /mnt/svn/gcc-trunk/gcc/ipa-cp.c:3315
0x129dfd0 ipcp_decision_stage
        /mnt/svn/gcc-trunk/gcc/ipa-cp.c:3441
0x129dfd0 ipcp_driver
        /mnt/svn/gcc-trunk/gcc/ipa-cp.c:3483
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <http://gcc.gnu.org/bugs.html> for instructions.

Tested revisions:
r193368 - fail
4.7 r191640 - OK

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[dcb314 at hotmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

David Binderman <dcb314 at hotmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dcb314 at hotmail dot com

--- Comment #1 from David Binderman <dcb314 at hotmail dot com> 2012-11-11 10:32:16 UTC ---
This seems to have broken sometime between 7 Nov and 11 Nov.

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[markus at trippelsdorf dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

Markus Trippelsdorf <markus at trippelsdorf dot de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jamborm at gcc dot gnu.org,
                   |                            |markus at trippelsdorf dot
                   |                            |de

--- Comment #2 from Markus Trippelsdorf <markus at trippelsdorf dot de> 2012-11-16 10:04:56 UTC ---
This ICE also hits in the boost testsuite:

markus@x4 more % cat test.ii
struct B
{
    virtual void test_suite_finish ();
};
void
fn1 (B & p2)
{
    fn1 (p2);
    B & a = p2;
    a.test_suite_finish ();
    B b;
    fn1 (b);
}

markus@x4 more % c++ -O3 -c test.ii
test.ii:13:1: internal compiler error: in ipa_get_parm_lattices, at
ipa-cp.c:263
 }
 ^
0xeb2b52 ipa_get_parm_lattices
        /home/markus/gcc/gcc/ipa-cp.c:263
0xeb059a ipa_get_parm_lattices
        /home/markus/gcc/gcc/ipa-cp.c:2930
0xeb059a cgraph_edge_brings_all_agg_vals_for_node
        /home/markus/gcc/gcc/ipa-cp.c:3105
0xeb059a perhaps_add_new_callers
        /home/markus/gcc/gcc/ipa-cp.c:3157
0xeb059a decide_about_value
        /home/markus/gcc/gcc/ipa-cp.c:3230
0xeb2693 decide_whether_version_node
        /home/markus/gcc/gcc/ipa-cp.c:3324
0xeb2693 ipcp_decision_stage
        /home/markus/gcc/gcc/ipa-cp.c:3464
0xeb2693 ipcp_driver
        /home/markus/gcc/gcc/ipa-cp.c:3506
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <http://gcc.gnu.org/bugs.html> for instructions.

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

Martin Jambor <jamborm at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2012-11-16
         AssignedTo|unassigned at gcc dot       |jamborm at gcc dot gnu.org
                   |gnu.org                     |
     Ever Confirmed|0                           |1

--- Comment #3 from Martin Jambor <jamborm at gcc dot gnu.org> 2012-11-16 16:07:31 UTC ---
Mine.

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

--- Comment #4 from Martin Jambor <jamborm at gcc dot gnu.org> 2012-11-17 12:45:05 UTC ---
Even though they both lead to the same ICE, the testcase in the
summary and the one from comment #2 are actually caused by different
bugs in ipa-cp.c.

I have just submitted the summary testcase fix to the mailing list:
http://gcc.gnu.org/ml/gcc-patches/2012-11/msg01443.html

I'll deal with the other issue next week.

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

--- Comment #5 from Martin Jambor <jamborm at gcc dot gnu.org> 2012-11-20 11:20:47 UTC ---
Author: jamborm
Date: Tue Nov 20 11:20:41 2012
New Revision: 193657

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=193657
Log:
2012-11-20  Martin Jambor  <mjambor@suse.cz>

    PR tree-optimization/55260
    * ipa-cp.c (find_aggregate_values_for_callers_subset): Rename info to
    dest_info, use caller_info instead of info when determining whether
    callee is a clone.

    * testsuite/g++.dg/torture/pr55260-1.C: New test.


Added:
    trunk/gcc/testsuite/g++.dg/torture/pr55260-1.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ipa-cp.c
    trunk/gcc/testsuite/ChangeLog

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

--- Comment #6 from Martin Jambor <jamborm at gcc dot gnu.org> 2012-11-21 17:03:46 UTC ---
This is the second fix, as posted to the mailing list:

http://gcc.gnu.org/ml/gcc-patches/2012-11/msg01724.html

also slightly related is:

http://gcc.gnu.org/ml/gcc-patches/2012-11/msg01758.html

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

--- Comment #7 from Martin Jambor <jamborm at gcc dot gnu.org> 2012-11-21 17:04:54 UTC ---
Author: jamborm
Date: Wed Nov 21 17:04:45 2012
New Revision: 193700

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=193700
Log:
2012-11-21  Martin Jambor  <mjambor@suse.cz>

    PR tree-optimization/55260
    * ipa-cp.c (intersect_aggregates_with_edge): New function.
    (find_aggregate_values_for_callers_subset): Part moved to the function
    above.  Call it.
    (cgraph_edge_brings_all_agg_vals_for_node): Reimplemented using
    intersect_aggregates_with_edge.

    * testsuite/g++.dg/torture/pr55260-2.C: New test.


Added:
    trunk/gcc/testsuite/g++.dg/torture/pr55260-2.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ipa-cp.c
    trunk/gcc/testsuite/ChangeLog

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

Martin Jambor <jamborm at gcc dot gnu.org> changed:

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

--- Comment #8 from Martin Jambor <jamborm at gcc dot gnu.org> 2012-11-21 17:10:05 UTC ---
Fixed.

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[shenhanc at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

Han Shen <shenhanc at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |shenhanc at gmail dot com

--- Comment #9 from Han Shen <shenhanc at gmail dot com> ---
Created attachment 31806
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31806&action=edit
Test case for gcc 4.8 and trunk

Hi I found this same error in gcc 4_8 and trunk, to reproduce - 
gcc -c -std=c89 -m64 -O3 ~/preprocessor.pp.c 

preprocessor.pp.c: At top level:
preprocessor.pp.c:108:1: internal compiler error: in ipa_get_parm_lattices, at
ipa-cp.c:261
 }
 ^
0xedeba7 ipa_get_parm_lattices
    ../../trunk/gcc/ipa-cp.c:261
0xedeba7 find_aggregate_values_for_callers_subset
    ../../trunk/gcc/ipa-cp.c:3051
0xedeba7 decide_about_value
    ../../trunk/gcc/ipa-cp.c:3365
0xee0a6c decide_whether_version_node
    ../../trunk/gcc/ipa-cp.c:3411
0xee0a6c ipcp_decision_stage
    ../../trunk/gcc/ipa-cp.c:3556
0xee0a6c ipcp_driver
    ../../trunk/gcc/ipa-cp.c:3598
0xee0a6c execute
    ../../trunk/gcc/ipa-cp.c:3694
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <http://gcc.gnu.org/bugs.html> for instructions.

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[trippels at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

Markus Trippelsdorf <trippels at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
                 CC|                            |trippels at gcc dot gnu.org
         Resolution|FIXED                       |---
      Known to fail|                            |4.9.0

--- Comment #10 from Markus Trippelsdorf <trippels at gcc dot gnu.org> ---
Confirmed.

A bit further reduced:

markus@x4 tmp % cat test.i
typedef struct {
  int *ptr;
  int len;
} string;
typedef struct {
  string nantstr;
  int *nant;
} malv;
typedef struct {
  int *nor;
} list_heads;
int b;
list_heads *fn1(string, int *, unsigned);
void fn2(malv *p1, list_heads *p2, unsigned p3) {
  string a = p1->nantstr;
  fn1(a, p1->nant, p3);
}

void fn3(unsigned p1) { fn2(0, 0, p1); }

list_heads *fn1(string p1, int *p2, unsigned p3) {
  while (1) {
    if (p3)
      fn3(1);
    if (b)
      return 0;
    fn3(1);
  }
}

void fn5() {
  list_heads c;
  c.nor = 0;
  fn2(0, &c, 1);
}

markus@x4 tmp % /var/tmp/gcc_test/usr/local/bin/gcc -c -O2 -fno-inline
-fipa-cp-clone test.i
test.i:35:1: internal compiler error: in ipa_get_parm_lattices, at ipa-cp.c:261
 }
 ^
0xe99411 ipa_get_parm_lattices
        ../../gcc/gcc/ipa-cp.c:261
0xe99411 cgraph_edge_brings_all_agg_vals_for_node
        ../../gcc/gcc/ipa-cp.c:3187
0xe99411 perhaps_add_new_callers
        ../../gcc/gcc/ipa-cp.c:3245
0xe99411 decide_about_value
        ../../gcc/gcc/ipa-cp.c:3318
0xe9c2c3 decide_whether_version_node
        ../../gcc/gcc/ipa-cp.c:3411
0xe9c2c3 ipcp_decision_stage
        ../../gcc/gcc/ipa-cp.c:3556
0xe9c2c3 ipcp_driver
        ../../gcc/gcc/ipa-cp.c:3598
0xe9c2c3 execute
        ../../gcc/gcc/ipa-cp.c:3694
Please submit a full bug report,

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[trippels at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

--- Comment #11 from Markus Trippelsdorf <trippels at gcc dot gnu.org> ---
Martin, can you take a look?

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

--- Comment #13 from Martin Jambor <jamborm at gcc dot gnu.org> ---
Fix to this second issue (which is indeed very similar to the initial
one) has been sent to the mailing list:

http://gcc.gnu.org/ml/gcc-patches/2014-02/msg00201.html

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

--- Comment #14 from Martin Jambor <jamborm at gcc dot gnu.org> ---
Author: jamborm
Date: Thu Feb 20 13:28:34 2014
New Revision: 207941

URL: http://gcc.gnu.org/viewcvs?rev=207941&root=gcc&view=rev
Log:
2014-02-20  Martin Jambor  <mjambor@suse.cz>

    PR ipa/55260
    * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
    info when checking whether lattices are bottom.

testsuite/
    * gcc.dg/ipa/pr55260.c: New test.


Added:
    trunk/gcc/testsuite/gcc.dg/ipa/pr55260.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ipa-cp.c
    trunk/gcc/testsuite/ChangeLog

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[trippels at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

Markus Trippelsdorf <trippels at gcc dot gnu.org> changed:

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

--- Comment #15 from Markus Trippelsdorf <trippels at gcc dot gnu.org> ---
Fixed again. Thanks.

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

--- Comment #16 from Martin Jambor <jamborm at gcc dot gnu.org> ---
Not yet on the 4.8 branch, but I'm re-running the tests there now and hopefully
will commit it also there today or tomorrow.

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[trippels at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

--- Comment #17 from Markus Trippelsdorf <trippels at gcc dot gnu.org> ---
(In reply to Martin Jambor from comment #16)
> Not yet on the 4.8 branch, but I'm re-running the tests there now and
> hopefully will commit it also there today or tomorrow.

Ah, sorry.

[Bug tree-optimization/55260] [4.8 Regression] ICE: in ipa_get_parm_lattices, at ipa-cp.c:263 with -O2 -fno-inline -fipa-cp-clone

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55260

--- Comment #18 from Martin Jambor <jamborm at gcc dot gnu.org> ---
Author: jamborm
Date: Fri Feb 21 13:05:40 2014
New Revision: 207990

URL: http://gcc.gnu.org/viewcvs?rev=207990&root=gcc&view=rev
Log:
2014-02-21  Martin Jambor  <mjambor@suse.cz>

    PR ipa/55260
    * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
    info when checking whether lattices are bottom.

testsuite/
    * gcc.dg/ipa/pr55260.c: New test.


Added:
    branches/gcc-4_8-branch/gcc/testsuite/gcc.dg/ipa/pr55260.c
Modified:
    branches/gcc-4_8-branch/gcc/ChangeLog
    branches/gcc-4_8-branch/gcc/ipa-cp.c
    branches/gcc-4_8-branch/gcc/testsuite/ChangeLog