[Bug rtl-optimization/20527] New: [4.1, 4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[Bug rtl-optimization/20527] New: [4.1, 4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[hp at gcc dot gnu dot org]

Seen on 4.0 "Thu Mar 17 06:05:21 UTC 2005" and trunk of "Thu Mar 17 15:06:06 UTC
2005" at -O2.  The test-case in the first attachment has this code snippet:
    vec += (base[i+1] - base[i]);

The expression at the right hand of that line is compiled at -O2 into:
        move.d [$r11+],$r9
        sub.d [$r11-4],$r9
Tsk tsk.  Some part of GCC doesn't adjust expressions when postincrements are
involved.

I don't see this bug with 3.2.1, so I call it a regression, though local changes
might affect the actual regressionness.  It is unknown whether it's also a
regression for 3.4 and 3.3.

-- 
           Summary: [4.1, 4.0 regression] Mishandling of postincrement
                    causes bzip2 miscompilation.
           Product: gcc
           Version: 4.0.0
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P2
         Component: rtl-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: hp at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org
GCC target triplet: cris-*-*


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.1, 4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[hp at gcc dot gnu dot org]

------- Additional Comments From hp at gcc dot gnu dot org  2005-03-18 03:35 -------
Created an attachment (id=8413)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8413&action=view)
Test-case

Test-case, derived from bzip2-1.0.2/huffman.c:BZ2_hbCreateDecodeTables
which also examplifies the effects of requiring long unwieldy license
notices with your code and derivates.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.1, 4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[hp at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |hp at gcc dot gnu dot org
                   |dot org                     |
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2005-03-18 03:36:16
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.1, 4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[hp at gcc dot gnu dot org]

------- Additional Comments From hp at gcc dot gnu dot org  2005-03-18 04:29 -------
The rtl dump for combine shows the first unmistakable sign of miscompilation:

(insn 35 34 36 2 (set (reg:SI 40)
        (mem:SI (post_inc:SI (reg:SI 31 [ ivtmp.5 ])) [3 S4 A8])) 32
{*movsi_internal} (nil)
    (expr_list:REG_INC (reg:SI 31 [ ivtmp.5 ])
        (nil)))

(note 36 35 37 2 NOTE_INSN_DELETED)

(insn 37 36 38 2 (set (reg:SI 42)
        (minus:SI (reg:SI 40)
            (mem:SI (plus:SI (reg:SI 31 [ ivtmp.5 ])
                    (const_int -4 [0xfffffffc])) [3 S4 A8]))) 70 {subsi3}
(insn_list:REG_DEP_TRUE 35 (nil))
    (expr_list:REG_INC (reg:SI 31 [ ivtmp.5 ])
        (expr_list:REG_DEAD (reg:SI 40)
            (nil))))

In contrast, the life pass just before it, looks right around this point.
Note also the REG_INC note on insn 37.  Looks spurious.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.1/4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[pinskia at gcc dot gnu dot org]

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.0/4.1 regression] Mishandling of postincrement causes bzip2 miscompilation.

[hp at gcc dot gnu dot org]

------- Additional Comments From hp at gcc dot gnu dot org  2005-03-22 01:15 -------
Patch at <URL:http://gcc.gnu.org/ml/gcc-patches/2005-03/msg02037.html>.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.0/4.1 regression] Mishandling of postincrement causes bzip2 miscompilation.

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-03-22 03:43 -------
Subject: Bug 20527

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	hp@gcc.gnu.org	2005-03-22 03:43:39

Modified files:
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/gcc.c-torture/execute: pr20527-1.c 

Log message:
	PR rtl-optimization/20527
	* gcc.c-torture/execute/pr20527-1.c: New test.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&r1=1.5191&r2=1.5192
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.c-torture/execute/pr20527-1.c.diff?cvsroot=gcc&r1=NONE&r2=1.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.0/4.1 regression] Mishandling of postincrement causes bzip2 miscompilation.

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-03-22 03:48 -------
Subject: Bug 20527

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	hp@gcc.gnu.org	2005-03-22 03:48:45

Modified files:
	gcc            : ChangeLog combine.c 

Log message:
	PR rtl-optimization/20527
	* combine.c (can_combine_p) [AUTO_INC_DEC]: When INSN has an
	REG_INC note, test that the register also isn't mentioned in PRED
	or SUCC.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.7953&r2=2.7954
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/combine.c.diff?cvsroot=gcc&r1=1.483&r2=1.484



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.1.0                       |4.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[hp at gcc dot gnu dot org]

------- Additional Comments From hp at gcc dot gnu dot org  2005-03-22 04:00 -------
Fixed on trunk (4.1), now waiting as per
<URL:http://gcc.gnu.org/ml/gcc-patches/2005-03/msg02051.html> before committing
to 4.0 and 3.4.


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[4.0/4.1 regression]        |[4.0 regression] Mishandling
                   |Mishandling of postincrement|of postincrement causes
                   |causes bzip2 miscompilation.|bzip2 miscompilation.


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[mmitchel at gcc dot gnu dot org]

------- Additional Comments From mmitchel at gcc dot gnu dot org  2005-04-04 22:55 -------
HP, your waiting period is up.  May we please have this patch for 4.0?

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[hp at gcc dot gnu dot org]

------- Additional Comments From hp at gcc dot gnu dot org  2005-04-05 09:47 -------
I may need some special absolution to apply this to 4.0, because I get
regressions for frv-elf there.  Well, ok, it's actually a timeout, but on a
2.8GHz P4 host, nonetheless):
WARNING: program timed out.
FAIL: 21_strings/basic_string/append/char/3.cc execution test

Hm, it's likely I had something else running the time the test failed
(the current run, LAST_UPDATED "Thu Mar 31 10:19:49 UTC 2005" was
with the patch removed), so I'll check frv-elf again
with the patch applied.  Results at the end of the day; the current run
is checking arm-elf and then there's sh-elf.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[hp at gcc dot gnu dot org]

------- Additional Comments From hp at gcc dot gnu dot org  2005-04-05 14:47 -------
Without special approval, I'm not going to apply this to 3.4, because it
causes a regression for v850-unknown-elf there, doubly confirmed, i.e. by
reverting the patch and noticing the failure disappear:
FAIL: g77.f-torture/execute/10197.f execution,  -O0
(Regarding 4.0, a frv-elf re-test run is in progress.)

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-04-05 16:39 -------
Subject: Bug 20527

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-4_0-branch
Changes by:	hp@gcc.gnu.org	2005-04-05 16:39:14

Modified files:
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/gcc.c-torture/execute: pr20527-1.c 

Log message:
	PR rtl-optimization/20527
	* gcc.c-torture/execute/pr20527-1.c: New test.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=1.5084.2.102&r2=1.5084.2.103
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.c-torture/execute/pr20527-1.c.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=NONE&r2=1.1.10.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-04-05 16:46 -------
Subject: Bug 20527

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-4_0-branch
Changes by:	hp@gcc.gnu.org	2005-04-05 16:45:56

Modified files:
	gcc            : ChangeLog combine.c 

Log message:
	PR rtl-optimization/20527
	* combine.c (can_combine_p) [AUTO_INC_DEC]: When INSN has an
	REG_INC note, test that the register also isn't mentioned in PRED
	or SUCC.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=2.7592.2.127&r2=2.7592.2.128
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/combine.c.diff?cvsroot=gcc&only_with_tag=gcc-4_0-branch&r1=1.475.4.2&r2=1.475.4.3



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527

[Bug rtl-optimization/20527] [4.0 regression] Mishandling of postincrement causes bzip2 miscompilation.

[hp at gcc dot gnu dot org]

------- Additional Comments From hp at gcc dot gnu dot org  2005-04-05 16:51 -------
No 3.4 commit and now committed to 4.0 too (the frv-elf timeout failure was not
repeatable): closing.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20527