[Bug middle-end/11414] New: [Regression 3.4] Segementation fault compiling csets.adb

[Bug middle-end/11414] New: [Regression 3.4] Segementation fault compiling csets.adb

[danglin at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11414

           Summary: [Regression 3.4] Segementation fault compiling csets.adb
           Product: gcc
           Version: 3.4
            Status: UNCONFIRMED
          Severity: critical
          Priority: P3
         Component: middle-end
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: danglin at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: hppa-unknown-linux-gnu
  GCC host triplet: hppa-unknown-linux-gnu
GCC target triplet: hppa-unknown-linux-gnu

stage2/xgcc -Bstage2/ -B/home/dave/opt/gnu/hppa-linux/bin/ -c -g -O2     
-gnatpg -gnata -I- -I. -Iada -I../../gcc/gcc/ada ../../gcc/gcc/ada/csets.adb -o
ada/csets.o
+===========================GNAT BUG DETECTED==============================+
| 3.4 20030521 (experimental) (hppa-unknown-linux-gnu) Segmentation fault  |
| Error detected at ../../gcc/gcc/ada/csets.adb:1092:17                    |
| Please submit a bug report; see http://gcc.gnu.org/bugs.html.            |
| Include the entire contents of this bug box in the report.               |
| Include the exact gcc or gnatmake command that you entered.              |
| Also include sources listed below in gnatchop format                     |
| (concatenated together with no headers between files).                   |
+==========================================================================+

See:
http://gcc.gnu.org/ml/gcc-patches/2003-05/msg01841.html

The problem has been tracked to the miscompilation of the file namet.adb
in stage2.  The error occurs because the label LABEL_NUSES count for
`label_ref 80' in the following instruction is corrupted either before
or during label replacement:

(jump_insn 66 65 70 4 0x40978750 (set (pc)
        (if_then_else (ne (reg:SI 105)
                (reg:SI 131))
            (label_ref 80)
            (pc))) 25 {*pa.md:1674} (nil)
    (expr_list:REG_EQUAL (if_then_else (ne (reg:SI 105)
                (const_int 112 [0x70]))
            (label_ref 80)
            (pc))
    (nil)))

The presence of the REG_EQUAL note causes the count for code_label 80
to be decrimented twice in a label replacement operation in load_mems.
Currently, LABEL_REFs in notes do not contribute to the LABEL_NUSES count
when notes are added and removed.

A possible fix has been posted here:
http://gcc.gnu.org/ml/gcc-patches/2003-07/msg00247.html

[Bug middle-end/11414] [Regression 3.4] Segementation fault compiling csets.adb

[danglin at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11414


danglin at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |zlomekj at suse dot cz, kazu
                   |                            |at cs dot umass dot edu,
                   |                            |roger at eyesopen dot com,
                   |                            |joern dot rennecke at superh
                   |                            |dot com


------- Additional Comments From danglin at gcc dot gnu dot org  2003-07-09 17:17 -------
I have added the authors of the following two patches to the CC list in the
hope that they can provide suggestions on how to resolve this PR.

The first patch added the notes which exposed the problem in label replacement.
The second patch revised the replacement of labels, although it looks as if
the problem predates this fix.
 
2003-05-20  Roger Sayle  <roger@eyesopen.com>
            Kazu Hirata  <kazu@cs.umass.edu>
            Joern Rennecke  <joern.rennecke@superh.com>

        * gcse.c (cprop_jump):  Make use of REG_EQUAL notes on both
        setcc and jump, if they exist.  If substituted instruction
        fails to validate, store current effort in a REG_EQUAL note.
        (cprop_insn): Don't attempt further substitutions if the
        current instruction has been deleted.
        (local_cprop_pass): Likewise.

        * jump.c (redirect_jump):  Also update REG_EQUAL note, if
        one is attached to the jump instruction.
        (invert_jump): Delete REG_EQUAL note on jump, if one exists.

2003-03-15  Josef Zlomek  <zlomekj@suse.cz>

        * rtl.h (subrtx_p): Renamed to rtx_referenced_p.
        (rtx_pair): Added new element update_label_nuses, renamed to
        replace_label_data.
        * cfgcleanup.c (outgoing_edges_match, try_crossjump_to_edge): Use
        replace_label_data instead of rtx_pair.
        * loop.c (load_mems): Likewise.
        * rtlanal.c (replace_label): Replace label in pool constants and in
        INSN_LIST (in REG_LABEL note).
        (subrtx_p): Renamed to rtx_referenced_p.
        (subrtx_p_1): Renamed to rtx_referenced_p_1, compare the interior of
        LABEL_REF with CODE_LABEL, traverse constants from pool.

[Bug middle-end/11414] [3.4 Regression] Segementation fault compiling csets.adb

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11414


pinskia at physics dot uc dot edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[Regression 3.4]            |[3.4 Regression]
                   |Segementation fault         |Segementation fault
                   |compiling csets.adb         |compiling csets.adb

[Bug middle-end/11414] [3.4 Regression] Segementation fault compiling csets.adb

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11414


pinskia at physics dot uc dot edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |debian-gcc at lists dot
                   |                            |debian dot org


------- Additional Comments From pinskia at physics dot uc dot edu  2003-07-16 00:59 -------
*** Bug 11524 has been marked as a duplicate of this bug. ***

[Bug middle-end/11414] [3.4 Regression] Segementation fault compiling csets.adb

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11414


pinskia at physics dot uc dot edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2003-07-20 16:43:58
               date|                            |


------- Additional Comments From pinskia at physics dot uc dot edu  2003-07-20 16:43 -------
Since there has been another report of this, then this can be confirmed and since the 
patch to fix this has not been approved.

[Bug middle-end/11414] [3.4 Regression] Segementation fault compiling csets.adb

[danglin at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11414


danglin at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1


------- Additional Comments From danglin at gcc dot gnu dot org  2003-09-22 15:19 -------
Change priority.

[Bug middle-end/11414] [3.4 Regression] Segementation fault compiling csets.adb

[pinskia at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11414


pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
           Keywords|                            |ice-on-valid-code
         Resolution|                            |FIXED


------- Additional Comments From pinskia at gcc dot gnu dot org  2003-10-24 00:57 -------
Fixed for 3.4 by:
	PR middle-end/11414
	* loop.c (load_mems): Use redirect_jump to forward jumps from
	the original loop end label to the new "loop sink" block's label.