[Bug tree-optimization/33458] New: [4.3 regression] ICE "PHI def is not a GIMPLE value"

[Bug tree-optimization/33458] New: [4.3 regression] ICE "PHI def is not a GIMPLE value"

[reichelt at gcc dot gnu dot org]

The following valid testcase triggers an ICE on mainline
when compiled with "-O":

===================================
inline void foo(int* p, int n)
{
  for (; n > 0; --n, ++p)
    *p = 0;
}

struct A
{
  int x[2];
  A() { foo(x, 2); }
};

inline A getA() { return A(); }

struct B
{
  A a;
  B();
};

B::B() : a(getA()) {}
===================================

bug.cc: In constructor 'B::B()':
bug.cc:20: error: PHI def is not a GIMPLE value
p_2 = PHI <&this_1(D)->a.x[0](2), p_5(3)>

&this_1(D)->a.x[0];

bug.cc:20: internal compiler error: verify_stmts failed
Please submit a full bug report, [etc.]

The regression was introduced between 2007-06-23 an 2007-07-02.


-- 
           Summary: [4.3 regression] ICE "PHI def is not a GIMPLE value"
           Product: gcc
           Version: 4.3.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code, monitored
          Severity: normal
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: reichelt at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33458

[Bug tree-optimization/33458] [4.3 regression] ICE "PHI def is not a GIMPLE value"

[reichelt at gcc dot gnu dot org]

-- 

reichelt at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.3.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33458

[Bug tree-optimization/33458] [4.3 regression] ICE "PHI def is not a GIMPLE value"

[pinskia at gcc dot gnu dot org]

------- Comment #1 from pinskia at gcc dot gnu dot org  2007-09-17 21:56 -------
Confirmed, inlining and RVO and RSO are interacting together interestingly.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hubicka at gcc dot gnu dot
                   |                            |org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2007-09-17 21:56:36
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33458

[Bug tree-optimization/33458] [4.3 regression] ICE "PHI def is not a GIMPLE value"

[rguenth at gcc dot gnu dot org]

------- Comment #2 from rguenth at gcc dot gnu dot org  2007-09-18 12:56 -------
We substitute into the invariant address &<retval>.a[0] *this, which results
in a non-invariant address.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33458

[Bug tree-optimization/33458] [4.3 regression] ICE "PHI def is not a GIMPLE value"

[rwgk at yahoo dot com]

------- Comment #3 from rwgk at yahoo dot com  2007-09-22 04:59 -------
I'm getting an ICE that looks very similar (below).
However, I cannot reproduce the ICE with the test case posted when this
issue was opened. Therefore I'm wondering if my ICE is different.
What is the platform, and what is the exact g++ command to reproduce the
ICE with the original test case?

Reducing the sources leading to my ICE is probably serious work...

My platform is:
Fedora Core release 5 (Bordeaux)
% uname -r -m
2.6.15-1.2054_FC5 x86_64

g++ (GCC) 4.3.0 20070919 (experimental) SVN revision 128608
g++ (GCC) 4.3.0 20070922 (experimental) SVN revision 128661

ice_gimple.cpp: In constructor
'scitbx::iso_surface::triangulation<CoordinatesType,
ValueType>::triangulation(scitbx::af::const_ref<ValueType,
scitbx::af::c_grid_padded<3ul, long unsigned int> >, ValueType, const
scitbx::af::tiny<IndexValueType, 3ul>&, bool) [with CoordinatesType = double,
ValueType = double]':
ice_gimple.cpp:32: error: invalid operand to binary operator
&this_2(D)->n_cells.D.101998.elems[0];

ice_gimple.cpp:32: error: PHI def is not a GIMPLE value
result_34 = PHI <&this_2(D)->n_cells.D.101998.elems[0](2), result_36(3)>

&this_2(D)->n_cells.D.101998.elems[0];

ice_gimple.cpp:32: error: invalid operand to binary operator
&this_2(D)->cell_lengths.D.104936.elems[0];

ice_gimple.cpp:32: error: PHI def is not a GIMPLE value
result_48 = PHI <&this_2(D)->cell_lengths.D.104936.elems[0](5), result_51(6)>

&this_2(D)->cell_lengths.D.104936.elems[0];

ice_gimple.cpp:32: internal compiler error: verify_stmts failed
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33458

[Bug tree-optimization/33458] [4.3 regression] ICE "PHI def is not a GIMPLE value"

[mmitchel at gcc dot gnu dot org]

-- 

mmitchel at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33458

[Bug tree-optimization/33458] [4.3 regression] ICE "PHI def is not a GIMPLE value"

[janis at gcc dot gnu dot org]

------- Comment #4 from janis at gcc dot gnu dot org  2007-10-19 16:42 -------
A regression hunt on powerpc-linux identified the following patch:

    http://gcc.gnu.org/viewcvs?view=rev&rev=126198

    r126198 | rguenth | 2007-07-02 11:53:08 +0000 (Mon, 02 Jul 2007)

This results in the ICE reported by the submitter.


-- 

janis at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33458

[Bug tree-optimization/33458] [4.3 regression] ICE "PHI def is not a GIMPLE value"

[pinskia at gcc dot gnu dot org]

------- Comment #5 from pinskia at gcc dot gnu dot org  2007-11-03 19:54 -------
I can no longer reproduce this bug.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33458

[Bug tree-optimization/33458] [4.3 regression] ICE "PHI def is not a GIMPLE value"

[reichelt at gcc dot gnu dot org]

------- Comment #6 from reichelt at gcc dot gnu dot org  2007-11-03 20:21 -------
Still crashes for me with today's compiler.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33458

[Bug tree-optimization/33458] [4.3 regression] ICE "PHI def is not a GIMPLE value"

[jakub at gcc dot gnu dot org]

------- Comment #7 from jakub at gcc dot gnu dot org  2007-11-05 14:38 -------
Testing a fix.


-- 

jakub at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |jakub at gcc dot gnu dot org
                   |dot org                     |
             Status|NEW                         |ASSIGNED
   Last reconfirmed|2007-09-17 21:56:36         |2007-11-05 14:38:51
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33458

[Bug tree-optimization/33458] [4.3 regression] ICE "PHI def is not a GIMPLE value"

[jakub at gcc dot gnu dot org]

------- Comment #8 from jakub at gcc dot gnu dot org  2007-11-06 08:29 -------
Subject: Bug 33458

Author: jakub
Date: Tue Nov  6 08:29:48 2007
New Revision: 129921

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=129921
Log:
        PR tree-optimization/33458
        * tree-inline.c (copy_phis_for_bb): If PHI arg substitution creates
        !is_gimple_val PHI argument, gimplify it and insert it on edge.

        * g++.dg/opt/inline12.C: New test.

Added:
    trunk/gcc/testsuite/g++.dg/opt/inline12.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-inline.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33458

[Bug tree-optimization/33458] [4.3 regression] ICE "PHI def is not a GIMPLE value"

[jakub at gcc dot gnu dot org]

------- Comment #9 from jakub at gcc dot gnu dot org  2007-11-06 08:35 -------
Fixed.


-- 

jakub at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33458