[Bug tree-optimization/24365] New: [4.1 Regression]

[Bug tree-optimization/24365] New: [4.1 Regression]

[pinskia at gcc dot gnu dot org]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1616 bytes --]

Take the following C++ code:

      typedef __complex__ double cdouble;
  inline cdouble to_complex(const double &r, const double &i = 0.0) {
     cdouble z;
     __real__ z = r;
     return z;
 }
  template<class value_type, bool simple_copy = true, bool atomic = true> class
RefPointer {
  };
       template<typename elt_t>   class Vector : public RefPointer<elt_t> {
    public:     typedef elt_t number;
      static elt_t elt_zero();
    };
    template <typename t> inline t min(t a, t b) {
 }
    template<> inline cdouble Vector<cdouble>::elt_zero() {
 return to_complex(0.0);
 }
    template<typename elt_t>   class Tensor : public Vector<elt_t> {
   };
    cdouble trace(const Tensor<cdouble> &t) {
     cdouble output = t.elt_zero();
 }
---
Compile at -O1, we get:
t.cc: In function ‘double __complex__ trace(const Tensor<double
__complex__>&)’:
t.cc:20: error: statement makes a memory store, but has no V_MAY_DEFS nor
V_MUST_DEFS
REALPART_EXPR <D.1799_14> = D.1803_7;
t.cc:20: internal compiler error: verify_ssa failed
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.


-- 
           Summary: [4.1 Regression]
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: critical
          Priority: P2
         Component: tree-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: pinskia at gcc dot gnu dot org
GCC target triplet: i686-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1 Regression] statement makes a memory store with complex

[pinskia at gcc dot gnu dot org]

-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1 Regression] statement makes a memory store with complex

[pinskia at gcc dot gnu dot org]

------- Comment #1 from pinskia at gcc dot gnu dot org  2005-10-14 13:48 -------
Reduced testcase:
typedef __complex__ double cdouble;
inline cdouble to_complex(double r) {
  cdouble z;
  __real__ z = r;
  return z;
}
cdouble elt_zero() {
  return to_complex(0.0);
}
-----

The inliner is remapping z/retval wrong and forgetting to copy if the complex
is a aggregate or a scaler.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2005-10-14 13:48:27
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1 Regression] statement makes a memory store with complex

[pinskia at gcc dot gnu dot org]

------- Comment #2 from pinskia at gcc dot gnu dot org  2005-10-14 13:49 -------
One more thing, this only happens with the C++ front-end.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1 Regression] statement makes a memory store with complex

[pinskia at gcc dot gnu dot org]

-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|critical                    |normal


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1 Regression] statement makes a memory store with complex

[pinskia at gcc dot gnu dot org]

------- Comment #3 from pinskia at gcc dot gnu dot org  2005-10-14 14:00 -------
Oh, the reason why I could confirm this one is because this is a split off from
PR 24362.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
OtherBugsDependingO|                            |24362
              nThis|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1 Regression] statement makes a memory store with complex

[pinskia at gcc dot gnu dot org]

------- Comment #4 from pinskia at gcc dot gnu dot org  2005-10-26 22:38 -------
This is a combined return slot optimization and inliner bug.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jason at gcc dot gnu dot
                   |                            |org, hubicka at gcc dot gnu
                   |                            |dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1 Regression] statement makes a memory store with complex

[steven at gcc dot gnu dot org]

------- Comment #5 from steven at gcc dot gnu dot org  2005-10-29 16:28 -------
This does not fail for me...  Neither the original test case nor the reduced
one. The compiler I used was "GNU C++ version 4.1.0 20051029 (experimental)
(x86_64-unknown-linux-gnu)"


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1 Regression] statement makes a memory store with complex

[steven at gcc dot gnu dot org]

------- Comment #6 from steven at gcc dot gnu dot org  2005-10-29 16:41 -------
This does ICE for i686 or for AMD64 -m32.  For 64 bits it passes for me.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1 Regression] statement makes a memory store with complex

[pinskia at gcc dot gnu dot org]

------- Comment #7 from pinskia at gcc dot gnu dot org  2005-10-29 18:01 -------
I have a fix but it could cause missed optimizations but from the looks of it,
that is because we don't use SFT on complex variables (or it might be on
RESULT_DECL which is the real issue).


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |pinskia at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1 Regression] statement makes a memory store with complex

[mmitchel at gcc dot gnu dot org]

------- Comment #8 from mmitchel at gcc dot gnu dot org  2005-10-31 06:18 -------
ICE on reasonable valid code; showstopper.


-- 

mmitchel at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P2                          |P1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1 Regression] statement makes a memory store with complex

[janis at gcc dot gnu dot org]

------- Comment #9 from janis at gcc dot gnu dot org  2005-11-02 17:15 -------
A regression hunt on powerpc-linux identified this patch:

r100793 | rth | 2005-06-09 07:43:40 +0000 (Thu, 09 Jun 2005) | 38 lines

        PR tree-opt/20610


-- 

janis at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1/4.2 Regression] statement makes a memory store with complex

[pinskia at gcc dot gnu dot org]

------- Comment #10 from pinskia at gcc dot gnu dot org  2006-01-13 14:30 -------
Subject: Bug 24365

Author: pinskia
Date: Fri Jan 13 14:30:08 2006
New Revision: 109667

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=109667
Log:
2006-01-13  Andrew Pinski  <pinskia@physics.uc.edu>

        PR tree-opt/24365
        * tree-inline.c (declare_return_variable): If the return variable
        was marked as a non gimple complex, set the return slot variable
        to be a non gimple comple variable.

2006-01-13  Andrew Pinski  <pinskia@physics.uc.edu>

        PR tree-opt/24365
        * g++.dg/opt/complex5.C: New test.


Added:
    trunk/gcc/testsuite/g++.dg/opt/complex5.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-inline.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1/4.2 Regression] statement makes a memory store with complex

[pinskia at gcc dot gnu dot org]

------- Comment #11 from pinskia at gcc dot gnu dot org  2006-01-13 14:38 -------
Subject: Bug 24365

Author: pinskia
Date: Fri Jan 13 14:38:03 2006
New Revision: 109668

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=109668
Log:
2006-01-13  Andrew Pinski  <pinskia@physics.uc.edu>

        PR tree-opt/24365
        * tree-inline.c (declare_return_variable): If the return variable
        was marked as a non gimple complex, set the return slot variable
        to be a non gimple comple variable.

2006-01-13  Andrew Pinski  <pinskia@physics.uc.edu>

        PR tree-opt/24365
        * g++.dg/opt/complex5.C: New test.


Added:
    branches/gcc-4_1-branch/gcc/testsuite/g++.dg/opt/complex5.C
      - copied unchanged from r109667,
trunk/gcc/testsuite/g++.dg/opt/complex5.C
Modified:
    branches/gcc-4_1-branch/gcc/ChangeLog
    branches/gcc-4_1-branch/gcc/testsuite/ChangeLog
    branches/gcc-4_1-branch/gcc/tree-inline.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1/4.2 Regression] statement makes a memory store with complex

[pinskia at gcc dot gnu dot org]

------- Comment #12 from pinskia at gcc dot gnu dot org  2006-01-13 14:38 -------
Fixed.


-- 

pinskia at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1/4.2 Regression] statement makes a memory store with complex

[jason at gcc dot gnu dot org]

------- Comment #13 from jason at gcc dot gnu dot org  2006-02-09 09:54 -------
Subject: Bug 24365

Author: jason
Date: Thu Feb  9 09:54:36 2006
New Revision: 110789

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=110789
Log:
        PR c++/25979
        * tree.def: Elaborate on difference from MODIFY_EXPR.
        * doc/c-tree.texi (INIT_EXPR): Likewise.
        * gimplify.c (internal_get_tmp_var): Use INIT_EXPR.
        (gimplify_decl_expr, gimplify_init_ctor_eval): Likewise.
        (gimplify_target_expr): Likewise.
        (gimplify_cond_expr): Remove target handling.
        (gimplify_modify_expr): Don't clobber INIT_EXPR code here.
        (gimplify_expr): Clobber it here.
        (gimplify_modify_expr_rhs): Push assignment into COND_EXPR here.
        Do return slot optimization if we have an INIT_EXPR.

        PR tree-opt/24365
        * tree-inline.c (declare_return_variable): Also clear
        DECL_COMPLEX_GIMPLE_REG_P as needed in the modify_dest case.

        PR c++/16405
        * gimplify.c (gimplify_modify_expr_rhs): Re-enable *& handling.

        PR middle-end/22439
        * gimplify.c (gimplify_one_sizepos): Fix typo.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/cp-gimplify.c
    trunk/gcc/cp/typeck2.c
    trunk/gcc/doc/c-tree.texi
    trunk/gcc/gimplify.c
    trunk/gcc/testsuite/g++.dg/opt/temp1.C
    trunk/gcc/tree-inline.c
    trunk/gcc/tree.def


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1/4.2 Regression] statement makes a memory store with complex

[jason at gcc dot gnu dot org]

------- Comment #14 from jason at gcc dot gnu dot org  2006-02-10 17:32 -------
Subject: Bug 24365

Author: jason
Date: Fri Feb 10 17:32:10 2006
New Revision: 110838

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=110838
Log:
        PR c++/25979
        * tree.def: Elaborate on difference from MODIFY_EXPR.
        * doc/c-tree.texi (INIT_EXPR): Likewise.
        * gimplify.c (internal_get_tmp_var): Use INIT_EXPR.
        (gimplify_decl_expr, gimplify_init_ctor_eval): Likewise.
        (gimplify_target_expr): Likewise.
        (gimplify_cond_expr): Remove target handling.
        (gimplify_modify_expr): Don't clobber INIT_EXPR code here.
        (gimplify_expr): Clobber it here.
        (gimplify_modify_expr_rhs): Push assignment into COND_EXPR here.
        Do return slot optimization if we have an INIT_EXPR.

        PR tree-opt/24365
        * tree-inline.c (declare_return_variable): Also clear
        DECL_COMPLEX_GIMPLE_REG_P as needed in the modify_dest case.

        PR c++/16405
        * gimplify.c (gimplify_modify_expr_rhs): Re-enable *& handling.

        PR middle-end/22439
        * gimplify.c (gimplify_one_sizepos): Fix typo.

Modified:
    branches/gcc-4_1-branch/gcc/ChangeLog
    branches/gcc-4_1-branch/gcc/cp/ChangeLog
    branches/gcc-4_1-branch/gcc/cp/cp-gimplify.c
    branches/gcc-4_1-branch/gcc/cp/typeck2.c
    branches/gcc-4_1-branch/gcc/doc/c-tree.texi
    branches/gcc-4_1-branch/gcc/gimplify.c
    branches/gcc-4_1-branch/gcc/testsuite/g++.dg/opt/temp1.C
    branches/gcc-4_1-branch/gcc/tree-inline.c
    branches/gcc-4_1-branch/gcc/tree.def


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1/4.2 Regression] statement makes a memory store with complex

[jason at gcc dot gnu dot org]

------- Comment #16 from jason at gcc dot gnu dot org  2006-02-12 08:01 -------
Subject: Bug 24365

Author: jason
Date: Sun Feb 12 08:01:04 2006
New Revision: 110888

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=110888
Log:
        PR tree-opt/24365
        * tree-inline.c (declare_return_variable): Just don't use the
        modify target if it's a gimple complex reg and the return slot
        isn't.

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


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365

[Bug tree-optimization/24365] [4.1/4.2 Regression] statement makes a memory store with complex

[jason at gcc dot gnu dot org]

------- Comment #15 from jason at gcc dot gnu dot org  2006-02-12 08:00 -------
Subject: Bug 24365

Author: jason
Date: Sun Feb 12 08:00:51 2006
New Revision: 110887

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=110887
Log:
        PR tree-opt/24365
        * tree-inline.c (declare_return_variable): Just don't use the
        modify target if it's a gimple complex reg and the return slot
        isn't.

Modified:
    branches/gcc-4_1-branch/gcc/ChangeLog
    branches/gcc-4_1-branch/gcc/tree-inline.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24365