[Bug lto/44891] New: non-trivial conversion ICE when comingling with -fwhopr

[Bug lto/44891] New: non-trivial conversion ICE when comingling with -fwhopr

[hubicka at gcc dot gnu dot org]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1113 bytes --]

jh@evans:/abuild/jh/mozilla-central/build8/ipc/ipdl>
/abuild/jh/trunk-install/bin/g++ ~/PTestDataStructuresParent.ii -O2 -r -fwhopr
-nostdlib
In file included from :244:0:
PTestDataStructuresParent.cpp: In member function ‘Read’:
PTestDataStructuresParent.cpp:1810:1: error: non-trivial conversion at
assignment
<float:32>
const ProcessHandle
# .MEM_303 = VDEF <.MEM_258>
MEM[(struct BarAttrs *)D.21979_35 + 20B].dummy_ = SR.1200_250(D);

PTestDataStructuresParent.cpp:1810:1: 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.


-- 
           Summary: non-trivial conversion ICE when comingling with -fwhopr
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: lto
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: hubicka at gcc dot gnu dot org
 GCC build triplet: x86_64-linux
GCC target triplet: x86_64-linux


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891

[Bug lto/44891] non-trivial conversion ICE when comingling with -fwhopr

[hubicka at gcc dot gnu dot org]

------- Comment #1 from hubicka at gcc dot gnu dot org  2010-07-09 14:52 -------
Created an attachment (id=21160)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=21160&action=view)
testcase


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891

[Bug lto/44891] non-trivial conversion ICE when comingling with -fwhopr

[rguenth at gcc dot gnu dot org]

------- Comment #2 from rguenth at gcc dot gnu dot org  2010-07-09 19:15 -------
Created an attachment (id=21168)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=21168&action=view)
reduced testcase


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891

[Bug lto/44891] non-trivial conversion ICE when comingling with -fwhopr

[rguenth at gcc dot gnu dot org]

-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
           Keywords|                            |ice-checking, ice-on-valid-
                   |                            |code, lto
   Last reconfirmed|0000-00-00 00:00:00         |2010-07-09 19:15:19
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891

[Bug tree-optimization/44891] non-trivial conversion ICE from early SRA

[rguenth at gcc dot gnu dot org]

------- Comment #3 from rguenth at gcc dot gnu dot org  2010-07-09 19:18 -------
Fails w/o LTO, at -O1, after early-SRA.

gcc> /abuild/rguenther/trunk-g/gcc/g++ -B /abuild/rguenther/trunk-g/gcc -O -S
PTestDataStructuresParent.3.ii 
PTestDataStructuresParent.3.ii: In member function 'bool
PTestDataStructuresParent::Read(PTestDataStructuresParent::Op*)':
PTestDataStructuresParent.3.ii:98:1: error: non-trivial conversion at
assignment
int
const float
# .MEM_12 = VDEF <.MEM_3(D)>
D.2561_9(D)->dummy_ = SR.2_11(D);

PTestDataStructuresParent.3.ii:98:1: 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.


-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rguenth at gcc dot gnu dot
                   |                            |org, jamborm at gcc dot gnu
                   |                            |dot org
          Component|lto                         |tree-optimization
           Keywords|lto                         |
            Summary|non-trivial conversion ICE  |non-trivial conversion ICE
                   |when comingling with -fwhopr|from early SRA


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891

[Bug tree-optimization/44891] non-trivial conversion ICE from early SRA

[hjl dot tools at gmail dot com]

------- Comment #4 from hjl dot tools at gmail dot com  2010-07-09 20:39 -------
It is caused by revision 161655:

http://gcc.gnu.org/ml/gcc-cvs/2010-07/msg00006.html


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891

[Bug tree-optimization/44891] non-trivial conversion ICE from early SRA

[rguenth at gcc dot gnu dot org]

------- Comment #5 from rguenth at gcc dot gnu dot org  2010-07-09 20:50 -------
Then it might be actually mine.


-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |rguenth at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED
   Last reconfirmed|2010-07-09 19:15:19         |2010-07-09 20:50:04
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891

[Bug tree-optimization/44891] non-trivial conversion ICE from early SRA

[rguenth at gcc dot gnu dot org]

------- Comment #6 from rguenth at gcc dot gnu dot org  2010-07-09 21:00 -------
Hm, somehow we loose the float typed access:

<L8>:
  MEM[(struct SpecificAttrs *)&D.1966].mType = 0;
  t_6 = MEM[(const struct SpecificAttrs *)&D.1966].mType;
  t.0_7 = (int) t_6;
  switch (t.0_7) <default: <L1>, case 1: <L3>>

<L3>:
  D.2156_8 = MEM[(const int &)&D.1966];
  D.2158_9(D)->dummy_ = D.2156_8;
  D.2155_10 = MEM[(const float &)&D.1966];
  MEM[(struct BarAttrs *)__v_2(D)].dummy_ = D.2155_10;

but the access tree just has

Access trees for D.1966 (UID: 1966):
access { base = (1966)'D.1966', offset = 0, size = 32, expr = MEM[(const int
&)&D.1966], type = const int, grp_write = 0, total_scalarization = 0, grp_read
= 1, grp_hint = 1, grp_assignment_read = 1,grp_covered = 0,
grp_unscalarizable_region = 0, grp_unscalarized_data = 0, grp_partial_lhs = 0,
grp_to_be_replaced = 0, grp_maybe_modified = 0,
grp_not_necessarilly_dereferenced = 0
access { base = (1966)'D.1966', offset = 32, size = 32, expr = MEM[(struct
SpecificAttrs *)&D.1966].mType, type = Type, grp_write = 1, total_scalarization
= 0, grp_read = 1, grp_hint = 0, grp_assignment_read = 1,grp_covered = 1,
grp_unscalarizable_region = 0, grp_unscalarized_data = 0, grp_partial_lhs = 0,
grp_to_be_replaced = 1, grp_maybe_modified = 0,
grp_not_necessarilly_dereferenced = 0

Hm.  Martin, maybe you are faster spotting the problem - the code is quite
convoluted.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891

[Bug tree-optimization/44891] [4.6 Regression] non-trivial conversion ICE from early SRA

[rguenth at gcc dot gnu dot org]

-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|non-trivial conversion ICE  |[4.6 Regression] non-trivial
                   |from early SRA              |conversion ICE from early
                   |                            |SRA
   Target Milestone|---                         |4.6.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891

[Bug tree-optimization/44891] [4.6 Regression] non-trivial conversion ICE from early SRA

[jamborm at gcc dot gnu dot org]

------- Comment #7 from jamborm at gcc dot gnu dot org  2010-07-21 12:51 -------
With MEM_REFs we now have fewer explicit typecasts and this has made
replace_uses_with_default_def_ssa_name create invalid statements.

FYI, this is not the ordinary replacing code path but rather the one
removing loads from known-to-be uninitialized memory.  I'll try to
reorganize the place where replace_uses_with_default_def_ssa_name is
called to both fix and not make it excessively ugly but it will be
ugly, I'm afraid.


-- 

jamborm at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|rguenth at gcc dot gnu dot  |jamborm at gcc dot gnu dot
                   |org                         |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891

[Bug tree-optimization/44891] [4.6 Regression] non-trivial conversion ICE from early SRA

[jamborm at gcc dot gnu dot org]

------- Comment #8 from jamborm at gcc dot gnu dot org  2010-07-21 17:29 -------
Created an attachment (id=21279)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=21279&action=view)
Proposed fix

I'm testing this patch as a fix.  Will submit it tomorrow if everything goes
fine.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891

[Bug tree-optimization/44891] [4.6 Regression] non-trivial conversion ICE from early SRA

[jamborm at gcc dot gnu dot org]

------- Comment #9 from jamborm at gcc dot gnu dot org  2010-07-22 12:52 -------
Subject: Bug 44891

Author: jamborm
Date: Thu Jul 22 12:52:14 2010
New Revision: 162413

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=162413
Log:
2010-07-22  Martin Jambor  <mjambor@suse.cz>

        PR tree-optimization/44891
        * tree-sra.c: Include gimple-pretty-print.h.
        (replace_uses_with_default_def_ssa_name): Renamed to
        get_repl_default_def_ssa_name, return the new SSA name instead of
        replacing the old one.
        (sra_modify_assign): Dump a message when removing a load, if the LHS
        is an SSA_NAME, do not do any propagation, just set the RHS to a
        default definition SSA NAME, type convert if necessary.
        * Makefile.in (tree-sra.o): Add gimple-pretty-print.h to dependencies.

        * testsuite/gcc.c-torture/compile/pr44891.c: New test.


Added:
    trunk/gcc/testsuite/gcc.c-torture/compile/pr44891.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/Makefile.in
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-sra.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891

[Bug tree-optimization/44891] [4.6 Regression] non-trivial conversion ICE from early SRA

[jamborm at gcc dot gnu dot org]

------- Comment #10 from jamborm at gcc dot gnu dot org  2010-07-22 12:54 -------
Fixed.


-- 

jamborm at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44891