[Bug middle-end/53826] New: [4.8 Regression] [alpha]: ICE in fold_convert_loc, at fold-const.c:2008

[Bug middle-end/53826] New: [4.8 Regression] [alpha]: ICE in fold_convert_loc, at fold-const.c:2008

[ubizjak at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53826

             Bug #: 53826
           Summary: [4.8 Regression] [alpha]: ICE in fold_convert_loc, at
                    fold-const.c:2008
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: ubizjak@gmail.com
              Host: x86_64-pc-linux-gnu
            Target: alpha-linux-gnu


Created attachment 27728
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=27728
Preprocessed source

Recent regression, triggered with libstdc++ testcase:

crosscompiler from x86_64-pc-linux-gnu to alpha-linux-gnu:

~/gcc-build-alpha/gcc/cc1plus -O2 -std=gnu++0x -quiet tuple_cat.ii
/home/uros/gcc-svn/trunk/libstdc++-v3/testsuite/20_util/tuple/creation_functions/tuple_cat.cc:
In function ‘int main()’:
/home/uros/gcc-svn/trunk/libstdc++-v3/testsuite/20_util/tuple/creation_functions/tuple_cat.cc:131:1:
internal compiler error: in fold_convert_loc, at fold-const.c:2008
 }
 ^
Please submit a full bug report,

(gdb) bt
#0  internal_error (gmsgid=gmsgid@entry=0xee36c5 "in %s, at %s:%d") at
../../gcc-svn/trunk/gcc/diagnostic.c:951
#1  0x0000000000d3e614 in fancy_abort (file=file@entry=0xe0cbc8
"../../gcc-svn/trunk/gcc/fold-const.c", line=line@entry=2008,
function=function@entry=0xe0bba0 "fold_convert_loc")
    at ../../gcc-svn/trunk/gcc/diagnostic.c:1011
#2  0x000000000082d6be in fold_convert_loc (loc=loc@entry=0,
type=0x7ffff146d348, arg=<optimized out>) at
../../gcc-svn/trunk/gcc/fold-const.c:2008
#3  0x0000000000bbc2e6 in build_zero_cst (type=<optimized out>) at
../../gcc-svn/trunk/gcc/tree.c:1668
#4  0x0000000000a783c9 in init_subtree_with_zero
(access=access@entry=0x1579060, gsi=gsi@entry=0x7fffffffda50,
insert_after=insert_after@entry=0 '\000', loc=loc@entry=2329653)
    at ../../gcc-svn/trunk/gcc/tree-sra.c:2521
#5  0x0000000000a783a9 in init_subtree_with_zero (access=<optimized out>,
gsi=gsi@entry=0x7fffffffda50, insert_after=insert_after@entry=0 '\000',
loc=2329653)
    at ../../gcc-svn/trunk/gcc/tree-sra.c:2532
#6  0x0000000000a80af1 in sra_modify_constructor_assign (gsi=0x7fffffffda50,
stmt=<synthetic pointer>) at ../../gcc-svn/trunk/gcc/tree-sra.c:2815
#7  sra_modify_assign (gsi=0x7fffffffda50, stmt=<synthetic pointer>) at
../../gcc-svn/trunk/gcc/tree-sra.c:2906
#8  sra_modify_function_body () at ../../gcc-svn/trunk/gcc/tree-sra.c:3141
#9  perform_intra_sra () at ../../gcc-svn/trunk/gcc/tree-sra.c:3247

(gdb) f 2
#2  0x000000000082d6be in fold_convert_loc (loc=loc@entry=0,
type=0x7ffff146d348, arg=<optimized out>) at
../../gcc-svn/trunk/gcc/fold-const.c:2008
2008          gcc_unreachable ();

(gdb) p debug_tree (type)
 <nullptr_type 0x7ffff146d348 decltype(nullptr) sizes-gimplified public
unsigned type_6 DI
    size <integer_cst 0x7ffff1345700 type <integer_type 0x7ffff13530a8
bitsizetype> constant 64>
    unit size <integer_cst 0x7ffff1345720 type <integer_type 0x7ffff1353000
sizetype> constant 8>
    align 8 symtab 0 alias set 20 canonical type 0x7ffff146d348
    pointer_to_this <pointer_type 0x7fffeeda4348> reference_to_this
<reference_type 0x7fffef114b28>>
$3 = void

(gdb) p debug_tree (orig)
 <integer_type 0x7ffff13535e8 int sizes-gimplified public type_6 SI
    size <integer_cst 0x7ffff1345a80 type <integer_type 0x7ffff13530a8
bitsizetype> constant 32>
    unit size <integer_cst 0x7ffff1345aa0 type <integer_type 0x7ffff1353000
sizetype> constant 4>
    align 32 symtab 0 alias set 2 canonical type 0x7ffff13535e8 precision 32
min <integer_cst 0x7ffff1345a20 -2147483648> max <integer_cst 0x7ffff1345a40
2147483647>
    pointer_to_this <pointer_type 0x7ffff13602a0> reference_to_this
<reference_type 0x7ffff04bf2a0>>
$4 = void

gcc needs to handle NULLPTR_TYPE here.

[Bug c++/53826] [4.8 Regression] [alpha]: ICE in fold_convert_loc, at fold-const.c:2008

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53826

Richard Guenther <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jason at gcc dot gnu.org
          Component|middle-end                  |c++
   Target Milestone|---                         |4.8.0

--- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-07-02 09:03:51 UTC ---
NULLPTR_TYPE should not prevail or POINTER_TYPE_P should include it.  Jason?

[Bug c++/53826] [4.8 Regression] [alpha]: ICE in fold_convert_loc, at fold-const.c:2008

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53826

--- Comment #2 from Jason Merrill <jason at gcc dot gnu.org> 2012-07-03 18:10:46 UTC ---
Author: jason
Date: Tue Jul  3 18:10:39 2012
New Revision: 189238

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=189238
Log:
    PR c++/53826
    * tree.c (build_zero_cst): Handle NULLPTR_TYPE.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/tree.c

[Bug c++/53826] [4.8 Regression] [alpha]: ICE in fold_convert_loc, at fold-const.c:2008

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53826

Jason Merrill <jason at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING
   Last reconfirmed|                            |2012-07-03
     Ever Confirmed|0                           |1

--- Comment #3 from Jason Merrill <jason at gcc dot gnu.org> 2012-07-03 18:21:12 UTC ---
Should be fixed now.

[Bug c++/53826] [4.8 Regression] [alpha]: ICE in fold_convert_loc, at fold-const.c:2008

[ubizjak at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53826

Uros Bizjak <ubizjak at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
                URL|                            |http://gcc.gnu.org/ml/gcc-p
                   |                            |atches/2012-07/msg00105.htm
                   |                            |l
         Resolution|                            |FIXED

--- Comment #4 from Uros Bizjak <ubizjak at gmail dot com> 2012-07-04 07:29:44 UTC ---
(In reply to comment #3)
> Should be fixed now.

Yes, the patch fixes the testsuite failure [1].

[1] http://gcc.gnu.org/ml/gcc-testresults/2012-07/msg00319.html

[Bug c++/53826] [4.8 Regression] [alpha]: ICE in fold_convert_loc, at fold-const.c:2008

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53826

--- Comment #5 from Jason Merrill <jason at gcc dot gnu.org> 2012-07-09 14:11:26 UTC ---
Author: jason
Date: Mon Jul  9 14:11:16 2012
New Revision: 189387

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=189387
Log:
    PR c++/53826
    * tree.c (build_zero_cst): Handle NULLPTR_TYPE.

Modified:
    branches/gcc-4_7-branch/gcc/ChangeLog
    branches/gcc-4_7-branch/gcc/tree.c

[Bug c++/53826] [4.8 Regression] [alpha]: ICE in fold_convert_loc, at fold-const.c:2008

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53826

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> 2012-09-10 09:52:19 UTC ---
*** Bug 53813 has been marked as a duplicate of this bug. ***