[Bug tree-optimization/38748] New: [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[Bug tree-optimization/38748] New: [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[rguenth at gcc dot gnu dot org]

The load from *q should be CSEd but it isn't because tree forwprop introduces
a VIEW_CONVERT_EXPR from a load of p->f which conflicts with the store
through *r.

/* { dg-do "compile" } */
/* { dg-options "-O -fdump-tree-fre" } */

struct S { float f; };
int __attribute__((noinline))
foo (float *r, struct S *p)
{
  int *q = (int *)&p->f;
  int i = *q;
  *r = 0.0;
  return i + *q;
}

/* { dg-final { scan-tree-dump-times "\\\*q" 1 "fre" } } */
/* { dg-final { cleanup-tree-dump "fre" } } */


-- 
           Summary: [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR
           Product: gcc
           Version: 4.4.0
            Status: UNCONFIRMED
          Keywords: missed-optimization
          Severity: normal
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: rguenth at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38748

[Bug tree-optimization/38748] [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[rguenth at gcc dot gnu dot org]

------- Comment #1 from rguenth at gcc dot gnu dot org  2009-01-06 21:33 -------
Related to PR38747.


-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|                            |38747
      Known to work|                            |4.3.2
   Target Milestone|---                         |4.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38748

[Bug tree-optimization/38748] [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[rguenth at gcc dot gnu dot org]

------- Comment #2 from rguenth at gcc dot gnu dot org  2009-01-11 14:29 -------
This hits tramp3d a lot, we end up with

D.919534 = VIEW_CONVERT_EXPR<const struct
Loc>(D.919509.D.78313.D.78008).D.78313.D.78008.domain_m[0].D.77181.D.47055.D.46936.domain_m;

vs. D.919509.D.78313.D.78008...

the following fixes it:

Index: gcc/tree-ssa-forwprop.c
===================================================================
--- gcc/tree-ssa-forwprop.c     (revision 143251)
+++ gcc/tree-ssa-forwprop.c     (working copy)
@@ -775,10 +775,13 @@ forward_propagate_addr_expr_1 (tree name
       && !TYPE_VOLATILE (TREE_TYPE (rhs))
       && !TYPE_VOLATILE (TREE_TYPE (TREE_OPERAND (def_rhs, 0)))
       && operand_equal_p (TYPE_SIZE (TREE_TYPE (rhs)),
-                         TYPE_SIZE (TREE_TYPE (TREE_OPERAND (def_rhs, 0))),
0)) 
+                         TYPE_SIZE (TREE_TYPE (TREE_OPERAND (def_rhs, 0))), 0)
+      && (rhs2 = get_base_address (TREE_OPERAND (def_rhs, 0))) != NULL_TREE
+      && !INDIRECT_REF_P (rhs2))
    {
      tree new_rhs = unshare_expr (TREE_OPERAND (def_rhs, 0));
      new_rhs = fold_build1 (VIEW_CONVERT_EXPR, TREE_TYPE (rhs), new_rhs);
+     STRIP_USELESS_TYPE_CONVERSION (new_rhs);
      if (TREE_CODE (new_rhs) != VIEW_CONVERT_EXPR)
        {
         /* If we have folded the VIEW_CONVERT_EXPR then the result is only


-- 

rguenth at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38748

[Bug tree-optimization/38748] [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[rguenth at gcc dot gnu dot org]

------- Comment #3 from rguenth at gcc dot gnu dot org  2009-01-11 14:41 -------
I can even see things like

SR.20836 = VIEW_CONVERT_EXPR<const struct Interval>(VIEW_CONVERT_EXPR<const
struct
Interval>(this->domain_m.D.115040.D.114739.domain_m[1].D.111565.D.45708.D.45522).D.45708.D.45522).D.45708.D.45522.domain_m[0];

in final cleanup :/


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38748

[Bug tree-optimization/38748] [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[pinskia at gcc dot gnu dot org]

------- Comment #4 from pinskia at gcc dot gnu dot org  2009-01-13 01:09 -------
We should be able to fold:
VIEW_CONVERT_EXPR<const struct Loc>(D.919509.D.78313.D.78008)

Into just D.919509 .

The indirect issue is a different one which needs to be fixed anyways as it
causes aliasing issues which were not in the original code in some cases.


-- 

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|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2009-01-13 01:09:36
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38748

[Bug tree-optimization/38748] [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[rguenther at suse dot de]

------- Comment #5 from rguenther at suse dot de  2009-01-13 08:55 -------
Subject: Re:  [4.4 Regression] Missed FRE because
 of VIEW_CONVERT_EXPR

On Tue, 13 Jan 2009, pinskia at gcc dot gnu dot org wrote:

> ------- Comment #4 from pinskia at gcc dot gnu dot org  2009-01-13 01:09 -------
> We should be able to fold:
> VIEW_CONVERT_EXPR<const struct Loc>(D.919509.D.78313.D.78008)
> 
> Into just D.919509 .

True.  I tried a quick shot at it but it turns out to be slightly
complicated ;)

Richard.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38748

[Bug tree-optimization/38748] [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[rguenth at gcc dot gnu dot org]

-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38748

[Bug tree-optimization/38748] [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[pinskia at gcc dot gnu dot org]

------- Comment #6 from pinskia at gcc dot gnu dot org  2009-01-15 23:16 -------
The CSE of the load is only valid at -O2 and above (strict aliasing turned on).


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38748

[Bug tree-optimization/38748] [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[pinskia at gcc dot gnu dot org]

------- Comment #7 from pinskia at gcc dot gnu dot org  2009-01-15 23:30 -------
PR 38747 has the best patch which allows most optimizations still but only
disabling it for the invalid cases when the aliasing sets are not equal.
Take:
struct S { unsigned f; };
int __attribute__((noinline))
foo ( struct S *p)
{
  int *q = (int *)&p->f;
  int i = *q;
  return i + p->f;
}

We should be able to optimize this to just "tmp = p->f; return (int)(tmp +
tmp);" which we do with my patch unlike the one in comment #2 which disables
too many optimizations.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38748

[Bug tree-optimization/38748] [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[rguenth at gcc dot gnu dot org]

------- Comment #8 from rguenth at gcc dot gnu dot org  2009-01-20 17:10 -------
Subject: Bug 38748

Author: rguenth
Date: Tue Jan 20 17:10:40 2009
New Revision: 143523

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=143523
Log:
2009-01-20  Andrew Pinski  <andrew_pinski@playstation.sony.com>
        Richard Guenther  <rguenther@suse.de>

        PR tree-optimization/38747
        PR tree-optimization/38748
        * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Disable the VCE
        conversion if the base address is an indirect reference and the
        aliasing sets could cause issues.

        * gcc.dg/tree-ssa/struct-aliasing-1.c: New test.
        * gcc.dg/tree-ssa/struct-aliasing-2.c: Likewise.
        * gcc.c-torture/execute/struct-aliasing-1.c: Likewise.

Added:
    trunk/gcc/testsuite/gcc.c-torture/execute/struct-aliasing-1.c
    trunk/gcc/testsuite/gcc.dg/tree-ssa/struct-aliasing-1.c
    trunk/gcc/testsuite/gcc.dg/tree-ssa/struct-aliasing-2.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-ssa-forwprop.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38748

[Bug tree-optimization/38748] [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[rguenth at gcc dot gnu dot org]

------- Comment #9 from rguenth at gcc dot gnu dot org  2009-01-20 17:11 -------
Fixed.


-- 

rguenth at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38748

[Bug tree-optimization/38748] [4.4 Regression] Missed FRE because of VIEW_CONVERT_EXPR

[hjl at gcc dot gnu dot org]

------- Comment #10 from hjl at gcc dot gnu dot org  2009-01-30 17:32 -------
Subject: Bug 38748

Author: hjl
Date: Fri Jan 30 17:31:24 2009
New Revision: 143798

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=143798
Log:
2009-01-30  H.J. Lu  <hongjiu.lu@intel.com>

        2009-01-27  Richard Guenther  <rguenther@suse.de>

        PR tree-optimization/38503
        * g++.dg/warn/Wstrict-aliasing-bogus-placement-new.C: New testcase.

        2009-01-26  Richard Guenther  <rguenther@suse.de>

        PR tree-optimization/38745
        * g++.dg/torture/pr38745.C: New testcase.

        2009-01-26  Richard Guenther  <rguenther@suse.de>

        PR middle-end/38851
        * g++.dg/warn/Wuninitialized-1.C: New testcase.

        2009-01-20  Andrew Pinski  <andrew_pinski@playstation.sony.com>
                    Richard Guenther  <rguenther@suse.de>

        PR tree-optimization/38747
        PR tree-optimization/38748
        * gcc.dg/tree-ssa/struct-aliasing-1.c: New test.
        * gcc.c-torture/execute/struct-aliasing-1.c: Likewise.

Added:
    branches/gcc-4_3-branch/gcc/testsuite/g++.dg/torture/pr38745.C
      - copied unchanged from r143797,
trunk/gcc/testsuite/g++.dg/torture/pr38745.C
   
branches/gcc-4_3-branch/gcc/testsuite/g++.dg/warn/Wstrict-aliasing-bogus-placement-new.C
      - copied unchanged from r143797,
trunk/gcc/testsuite/g++.dg/warn/Wstrict-aliasing-bogus-placement-new.C
    branches/gcc-4_3-branch/gcc/testsuite/g++.dg/warn/Wuninitialized-1.C
      - copied unchanged from r143797,
trunk/gcc/testsuite/g++.dg/warn/Wuninitialized-1.C
   
branches/gcc-4_3-branch/gcc/testsuite/gcc.c-torture/execute/struct-aliasing-1.c
      - copied unchanged from r143797,
trunk/gcc/testsuite/gcc.c-torture/execute/struct-aliasing-1.c
    branches/gcc-4_3-branch/gcc/testsuite/gcc.dg/tree-ssa/struct-aliasing-1.c
      - copied unchanged from r143797,
trunk/gcc/testsuite/gcc.dg/tree-ssa/struct-aliasing-1.c
Modified:
    branches/gcc-4_3-branch/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38748