[Bug tree-optimization/21562] New: Quiet bad codegen (unrolling + tail call interaction)

[Bug tree-optimization/21562] New: Quiet bad codegen (unrolling + tail call interaction)

[dalej at gcc dot gnu dot org]

Compile the following with -O2 -funroll-loops and you get bad codegen.  An uninitilaized value
is used as "pos" in computing the store address for mr[pos]=1 at the top of the function.  The
first dump where it's wrong seems to be "nrv".  If you unroll by hand it comes out correct.

extern int delta[8];
extern unsigned char board[421];

void
ping_recurse(int pos, int mr[400],  int color)
{
  int k;
  mr[pos] = 1;

  for (k = 0; k < 4; k++) {
    int apos = pos + delta[k];
    if (mr[apos] == 1 || board[apos] == color)
       ping_recurse(apos, mr, color);
  }
}

Beginning of codegen on darwin8:

_ping_recurse:
        mflr r0
        bcl 20,31,"L00000000001$pb"
"L00000000001$pb":
        stmw r28,-16(r1)
        mr r29,r4
        mflr r31
        mr r28,r5
        stw r0,8(r1)
        stwu r1,-80(r1)
L25:
        addis r11,r31,ha16(L_delta$non_lazy_ptr-"L00000000001$pb")
        slwi r2,r30,2   <<<< R30 uninitiailized

-- 
           Summary: Quiet bad codegen (unrolling + tail call interaction)
           Product: gcc
           Version: 4.0.0
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P2
         Component: tree-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: dalej at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: powerpc-apple-darwin8.0.0
  GCC host triplet: powerpc-apple-darwin8.0.0
GCC target triplet: powerpc-apple-darwin8.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0/4.1 Regression] Quiet bad codegen (unrolling + tail call interaction)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-14 00:24 -------
Confirmed, it is not slight with checking enabled:
pr21562.c: In function 'ping_recurse':
pr21562.c:15: error: Wrong amount of branch edges after unconditional jump 0
pr21562.c:15: error: verify_flow_info: Incorrect blocks for fallthru 0->18
pr21562.c:15: internal compiler error: verify_flow_info failed
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
           Keywords|                            |ice-checking, ice-on-valid-
                   |                            |code
   Last reconfirmed|0000-00-00 00:00:00         |2005-05-14 00:24:40
               date|                            |
            Summary|Quiet bad codegen (unrolling|[4.0/4.1 Regression] Quiet
                   |+ tail call interaction)    |bad codegen (unrolling +
                   |                            |tail call interaction)
   Target Milestone|---                         |4.0.1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0/4.1 Regression] Quiet bad codegen (unrolling + tail call interaction)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-14 00:41 -------
(In reply to comment #1)
> Confirmed, it is not slight with checking enabled:
Which is really weird because it is not until expand time we catch it.
Oh and this ICEs on i686-pc-linux-gnu also.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
  GCC build triplet|powerpc-apple-darwin8.0.0   |
   GCC host triplet|powerpc-apple-darwin8.0.0   |
 GCC target triplet|powerpc-apple-darwin8.0.0   |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0/4.1 Regression] Quiet bad codegen (unrolling + tail call interaction)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-14 00:47 -------
Note this now fails at -O3 -fno-inline (for 4.1.0).


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0/4.1 Regression] Quiet bad codegen (unrolling + tail call interaction)

[dalej at gcc dot gnu dot org]

------- Additional Comments From dalej at gcc dot gnu dot org  2005-05-16 21:44 -------
OK, what happens is that tail call creates a block 0 with multiple predecessors, and the
out-of-ssa pass mangles that into this form:

  # BLOCK 0
  # PRED: 9 [14.9%]  (true,exec) 1 [29.0%]  (true,exec)
<L37>:;
  posD.1132 = aposD.1138;
  # SUCC: 12 [100.0%]  (fallthru)

  # BLOCK 12
  # PRED: 0 [100.0%]  (fallthru) ENTRY [100.0%]  (fallthru,exec)
<L38>:;
  *((intD.0 *) ((unsigned intD.3) posD.1132 * 4) + mrD.1133) = 1;

where ENTRY no longer falls into block 0, but is expected to jump to block 12.  Expand has no
idea such a thing is possible and falls through anyway.  While it's not obvious in 4.0 branch, the
addition of a verify_flow_info failure in mainline makes it apparent this is not supposed to happen.

One way to fix this is change tailcall so it splits block 0 instead of creating multiple predecessors
for it.  I'll go do that unless somebody has a better idea.


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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0/4.1 Regression] Quiet bad codegen (unrolling + tail call interaction)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-16 22:54 -------
(In reply to comment #4)
> where ENTRY no longer falls into block 0, but is expected to jump to block 12.  Expand has no
> idea such a thing is possible and falls through anyway.  While it's not obvious in 4.0 branch, the
> addition of a verify_flow_info failure in mainline makes it apparent this is not supposed to happen.
If you turn on checking it will fail on the 4.0 branch.

-- Pinski

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0/4.1 Regression] Quiet bad codegen (unrolling + tail call interaction)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-26 23:33 -------
*** Bug 21774 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |uttamp at us dot ibm dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0/4.1 Regression] Quiet bad codegen (unrolling + tail call interaction)

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-05-28 22:27 -------
Subject: Bug 21562

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	hubicka@gcc.gnu.org	2005-05-28 22:27:04

Modified files:
	gcc            : ChangeLog cfgexpand.c 
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/gcc.c-torture/compile: pr21562.c 

Log message:
	PR tree-optimization/21562
	* cfgexpand.c (construct_init_block): Deal properly with the case
	of entry edge not pointing to very first basic block.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.8935&r2=2.8936
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cfgexpand.c.diff?cvsroot=gcc&r1=2.39&r2=2.40
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&r1=1.5550&r2=1.5551
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.c-torture/compile/pr21562.c.diff?cvsroot=gcc&r1=NONE&r2=1.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0 Regression] Quiet bad codegen (unrolling + tail call interaction)

[steven at gcc dot gnu dot org]

------- Additional Comments From steven at gcc dot gnu dot org  2005-06-07 07:51 -------
So is this fixed now?

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |WAITING


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0 Regression] Quiet bad codegen (unrolling + tail call interaction)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-06-07 11:42 -------
(In reply to comment #8)
> So is this fixed now?

On the mainline yes but not on the 4.0 branch.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |NEW
   Last reconfirmed|2005-05-14 00:24:40         |2005-06-07 11:42:17
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0 Regression] Quiet bad codegen (unrolling + tail call interaction)

[mmitchel at gcc dot gnu dot org]

------- Additional Comments From mmitchel at gcc dot gnu dot org  2005-06-26 01:53 -------
Jan, would you please see if this patch can be easily applied to the 4.0 branch?
 I'm trying to clear out the known wrong-code problems in advance of 4.0.1, and
as this is fixed on the mainline, it might be easy to fix it on the branch.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0 Regression] Quiet bad codegen (unrolling + tail call interaction)

[hubicka at ucw dot cz]

------- Additional Comments From hubicka at ucw dot cz  2005-06-28 01:29 -------
Subject: Re:  [4.0 Regression] Quiet bad codegen (unrolling + tail call interaction)

> 
> ------- Additional Comments From mmitchel at gcc dot gnu dot org  2005-06-26 01:53 -------
> Jan, would you please see if this patch can be easily applied to the 4.0 branch?
>  I'm trying to clear out the known wrong-code problems in advance of 4.0.1, and
> as this is fixed on the mainline, it might be easy to fix it on the branch.

I don't think anything important changed here since 4.0, so it should
apply w/o problems...

Honza
> 
> -- 
> 
> 
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562
> 
> ------- You are receiving this mail because: -------
> You are the assignee for the bug, or are watching the assignee.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0 Regression] Quiet bad codegen (unrolling + tail call interaction)

[mmitchel at gcc dot gnu dot org]

------- Additional Comments From mmitchel at gcc dot gnu dot org  2005-07-06 16:15 -------
Sadly, I didn't see the last comment in this PR until after 4.0.1 RC3.

If there is a 4.0.1 RC4, this patch will be included; otherwise, it will be in
4.0.2.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mark at codesourcery dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0 Regression] Quiet bad codegen (unrolling + tail call interaction)

[mmitchel at gcc dot gnu dot org]

------- Additional Comments From mmitchel at gcc dot gnu dot org  2005-07-06 17:03 -------
Postponed until 4.0.2.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.0.1                       |4.0.2


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0 Regression] Quiet bad codegen (unrolling + tail call interaction)

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-08-05 17:51 -------
Subject: Bug 21562

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	apple-local-200502-branch
Changes by:	dalej@gcc.gnu.org	2005-08-05 17:51:14

Modified files:
	gcc            : ChangeLog.apple-ppc cfgexpand.c 

Log message:
	2005-08-05  Dale Johannesen  <dalej@apple.com>
	
	* cfgexpand.c (construct_init_block):  Roll in fix
	for PR 21562 from mainline.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.apple-ppc.diff?cvsroot=gcc&only_with_tag=apple-local-200502-branch&r1=1.1.4.106&r2=1.1.4.107
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cfgexpand.c.diff?cvsroot=gcc&only_with_tag=apple-local-200502-branch&r1=2.32.6.1&r2=2.32.6.2



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0 Regression] Quiet bad codegen (unrolling + tail call interaction)

[bonzini at gcc dot gnu dot org]

------- Additional Comments From bonzini at gcc dot gnu dot org  2005-08-18 09:50 -------
Patch applied to branch for 4.0.2.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562

[Bug tree-optimization/21562] [4.0 Regression] Quiet bad codegen (unrolling + tail call interaction)

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-08-18 09:51 -------
Subject: Bug 21562

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-4_0-branch
Changes by:	bonzini@gcc.gnu.org	2005-08-18 09:51:11

Modified files:
	gcc            : ChangeLog cfgexpand.c 
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/gcc.c-torture/compile: pr21562.c 

Log message:
	gcc:
	2005-08-18  Paolo Bonzini  <bonzini@gnu.org>
	
	PR c/21562
	* cfgexpand.c (construct_init_block): Deal properly with the case
	of entry edge not pointing to very first basic block.
	
	testsuite:
	2005-08-18  Paolo Bonzini  <bonzini@gcc.gnu.org>
	
	PR c/21562
	* gcc.c-torture/compile/pr21562.c: New.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=2.7592.2.376&r2=2.7592.2.377
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cfgexpand.c.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=2.33&r2=2.33.2.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=1.5084.2.339&r2=1.5084.2.340
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.c-torture/compile/pr21562.c.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=NONE&r2=1.1.18.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21562