[Bug bootstrap/17591] New: Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[Bug bootstrap/17591] New: Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[laurent at guerby dot net]

Somewhere between:
LAST_UPDATED: Mon Sep 20 17:27:11 UTC 2004
LAST_UPDATED: Tue Sep 21 07:01:37 UTC 2004

Ada bootstrap started failing on x86 (fine on x86_64) with:

stage1/xgcc -Bstage1/
-B/home/guerby/work/gcc/install/install-20040921T124859/i686-pc-linux-gnu/bin/
-c -O2 -g -fomit-frame-pointer      -gnatpg -gnata -I- -I. -Iada
-I/home/guerby/work/gcc/version-head/gcc/ada
/home/guerby/work/gcc/version-head/gcc/ada/osint.adb -o ada/osint.o

Execution terminated by unhandled exception
Exception name: STORAGE_ERROR
Message: stack overflow (or erroneous memory access)
Call stack traceback locations:
0x80a9816
make[2]: *** [ada/osint.o] Error 1
make[2]: Leaving directory `/home/guerby/work/gcc/build/build-20040921T124859/gcc'
make[1]: *** [stage2_build] Error 2
make[1]: Leaving directory `/home/guerby/work/gcc/build/build-20040921T124859/gcc'
make: *** [bootstrap] Error 2

Here is a gdb session:

(gdb) r -I- -I. -Iada -I/home/guerby/work/gcc/version-head/gcc/ada -quiet
-dumpbase osint.adb -O2 -fomit-frame-pointer -g -gnatpg -gnata -mtune=pentiumpro
-gnatO ada/osint.o /home/guerby/work/gcc/version-head/gcc/ada/osint.adb -o
/tmp/ccTN4cek.s
Starting program:
/home/guerby/work/gcc/build/build-20040921T124859/gcc/stage1/gnat1 -I- -I. -Iada
-I/home/guerby/work/gcc/version-head/gcc/ada -quiet -dumpbase osint.adb -O2
-fomit-frame-pointer -g -gnatpg -gnata -mtune=pentiumpro -gnatO ada/osint.o
/home/guerby/work/gcc/version-head/gcc/ada/osint.adb -o /tmp/ccTN4cek.s

Program received signal SIGSEGV, Segmentation fault.
0x08480d93 in operand_equal_p (arg0=0x4071b15c, arg1=0x40a98d98, flags=0) at
/home/guerby/work/gcc/version-head/gcc/fold-const.c:2335
2335      if (TYPE_UNSIGNED (TREE_TYPE (arg0)) != TYPE_UNSIGNED (TREE_TYPE (arg1)))
(gdb) bt
#0  0x08480d93 in operand_equal_p (arg0=0x4071b15c, arg1=0x40a98d98, flags=0) at
/home/guerby/work/gcc/version-head/gcc/fold-const.c:2335
During symbol reading, Incomplete CFI data; unspecified registers at 0x084816c8.
During symbol reading, Incomplete CFI data; unspecified registers at 0x084816c8.
During symbol reading, Incomplete CFI data; unspecified registers at 0x084816c8.
During symbol reading, Incomplete CFI data; unspecified registers at 0x084816c8.
During symbol reading, Incomplete CFI data; unspecified registers at 0x084816c8.
During symbol reading, Incomplete CFI data; unspecified registers at 0x084816c8.
During symbol reading, Incomplete CFI data; unspecified registers at 0x084816c8.
#1  0x084a1c40 in fold (expr=0x406e2804) at
/home/guerby/work/gcc/version-head/gcc/fold-const.c:6955
During symbol reading, Incomplete CFI data; unspecified registers at 0x084a1cf9.
During symbol reading, Incomplete CFI data; unspecified registers at 0x084a1cf9.
During symbol reading, Incomplete CFI data; unspecified registers at 0x084a1cf9.
During symbol reading, Incomplete CFI data; unspecified registers at 0x084a1cf9.
During symbol reading, Incomplete CFI data; unspecified registers at 0x084a1cf9.
#2  0x0842b190 in set_mem_attributes_minus_bitpos (ref=0x40abb6e4, t=0x40ac416c,
objectp=0, bitpos=0)
    at /home/guerby/work/gcc/version-head/gcc/emit-rtl.c:1636
During symbol reading, Incomplete CFI data; unspecified registers at 0x0842b1ab.
During symbol reading, Incomplete CFI data; unspecified registers at 0x0842b1ab.
During symbol reading, Incomplete CFI data; unspecified registers at 0x0842b1ab.
During symbol reading, Incomplete CFI data; unspecified registers at 0x0842b1ab.
During symbol reading, Incomplete CFI data; unspecified registers at 0x0842b1ab.
During symbol reading, Incomplete CFI data; unspecified registers at 0x0842b1ab.
#3  0x0842b6fe in set_mem_attributes (ref=0x40abb6e4, t=0x40ac416c, objectp=0)
at /home/guerby/work/gcc/version-head/gcc/emit-rtl.c:1723
During symbol reading, Incomplete CFI data; unspecified registers at 0x0842b6f9.
During symbol reading, Incomplete CFI data; unspecified registers at 0x0842b6f9.
#4  0x0845eca8 in expand_expr_real_1 (exp=0x4065a300, target=0x40ab0620,
tmode=QImode, modifier=EXPAND_NORMAL, alt_rtl=0xbfffe910)
    at /home/guerby/work/gcc/version-head/gcc/expr.c:6769
#5  0x0845c533 in expand_expr_real (exp=0x4065a300, target=0x40ab0620,
tmode=QImode, modifier=EXPAND_NORMAL, alt_rtl=0xbfffe910)
    at /home/guerby/work/gcc/version-head/gcc/expr.c:6313
#6  0x084530bc in store_expr (exp=0x4065a300, target=0x40ab0620, want_value=0)
at /home/guerby/work/gcc/version-head/gcc/expr.c:3933
#7  0x084522c6 in expand_assignment (to=0x40ac7074, from=0x4065a300,
want_value=0) at /home/guerby/work/gcc/version-head/gcc/expr.c:3746
#8  0x0846a08f in expand_expr_real_1 (exp=0x40719558, target=0x0,
tmode=VOIDmode, modifier=EXPAND_NORMAL, alt_rtl=0x0)
    at /home/guerby/work/gcc/version-head/gcc/expr.c:8087
#9  0x0845c503 in expand_expr_real (exp=0x40719558, target=0x4001b210,
tmode=VOIDmode, modifier=EXPAND_NORMAL, alt_rtl=0x0)
    at /home/guerby/work/gcc/version-head/gcc/expr.c:6307
#10 0x0866ed66 in expand_expr (exp=0x40719558, target=0x4001b210, mode=VOIDmode,
modifier=EXPAND_NORMAL) at expr.h:493
#11 0x08668d7e in expand_expr_stmt (exp=0x40719558) at
/home/guerby/work/gcc/version-head/gcc/stmt.c:1353
#12 0x086bd910 in expand_gimple_basic_block (bb=0x40ab5828, dump_file=0x0) at
/home/guerby/work/gcc/version-head/gcc/cfgexpand.c:1056
#13 0x086bdeff in tree_expand_cfg () at
/home/guerby/work/gcc/version-head/gcc/cfgexpand.c:1223
#14 0x082f8df9 in execute_one_pass (pass=0x8867460) at
/home/guerby/work/gcc/version-head/gcc/tree-optimize.c:501
#15 0x082f8ee5 in execute_pass_list (pass=0x8867460) at
/home/guerby/work/gcc/version-head/gcc/tree-optimize.c:536
#16 0x082f91a5 in tree_rest_of_compilation (fndecl=0x403ced24, nested_p=0 '\0')
at /home/guerby/work/gcc/version-head/gcc/tree-optimize.c:636
#17 0x0806f690 in gnat_expand_body (gnu_decl=0x403ced24) at
/home/guerby/work/gcc/version-head/gcc/ada/misc.c:652
#18 0x086e9dbc in cgraph_expand_function (node=0x4065fef4) at
/home/guerby/work/gcc/version-head/gcc/cgraphunit.c:1044
#19 0x086ede22 in cgraph_expand_all_functions () at
/home/guerby/work/gcc/version-head/gcc/cgraphunit.c:2710
#20 0x086ee202 in cgraph_optimize () at
/home/guerby/work/gcc/version-head/gcc/cgraphunit.c:2821
#21 0x0806dbb7 in gnat_parse_file (set_yydebug=0) at
/home/guerby/work/gcc/version-head/gcc/ada/misc.c:240
#22 0x0867a989 in compile_file () at
/home/guerby/work/gcc/version-head/gcc/toplev.c:991
#23 0x0867c326 in do_compile () at
/home/guerby/work/gcc/version-head/gcc/toplev.c:2079
#24 0x0867c38d in toplev_main (argc=19, argv=0xbfffefd4) at
/home/guerby/work/gcc/version-head/gcc/toplev.c:2111
#25 0x082cf00a in main (argc=19, argv=0xbfffefd4) at
/home/guerby/work/gcc/version-head/gcc/main.c:35
(gdb) l
35        return toplev_main (argc, (const char **) argv);
36      }
(gdb) f 0
#0  0x08480d93 in operand_equal_p (arg0=0x4071b15c, arg1=0x40a98d98, flags=0) at
/home/guerby/work/gcc/version-head/gcc/fold-const.c:2335
2335      if (TYPE_UNSIGNED (TREE_TYPE (arg0)) != TYPE_UNSIGNED (TREE_TYPE (arg1)))
(gdb) l
2330        return 0;
2331
2332      /* If both types don't have the same signedness, then we can't consider
2333         them equal.  We must check this before the STRIP_NOPS calls
2334         because they may change the signedness of the arguments.  */
2335      if (TYPE_UNSIGNED (TREE_TYPE (arg0)) != TYPE_UNSIGNED (TREE_TYPE (arg1)))
2336        return 0;
2337
2338      STRIP_NOPS (arg0);
2339      STRIP_NOPS (arg1);
(gdb) p arg0
$1 = 0x4071b15c
(gdb) p debug_tree(arg0)
 <var_decl 0x4071b15c upper_bound.313
    type <integer_type 0x40021488 integer asm_written public visited SI
        size <integer_cst 0x4001e408 constant invariant 32>
        unit size <integer_cst 0x4001e138 constant invariant 4>
        align 32 symtab 1073890112 alias set 119 precision 32 min <integer_cst
0x4001e3d8 -2147483648> max <integer_cst 0x4001e3f0 2147483647>
        pointer_to_this <pointer_type 0x40021e80>>
    used ignored SI file /home/guerby/work/gcc/version-head/gcc/ada/osint.adb
line 1148 size <integer_cst 0x4001e408 32> unit size <integer_cst 0x4001e138 4>
    align 32 context <function_decl 0x403ced24 osint__get_next_dir_in_path>
    (reg:SI 82 [ upper_bound.313 ]) chain <var_decl 0x4071b1d0>>
$2 = void
(gdb) p debug_tree(arg1)
 <ssa_name 0x40a98d98 nothrow var <var_decl 0x40717e80> def_stmt <ssa_name
0x40ab8cd0>
    version 34 in-free-list>
$3 = void
(gdb) call print_generic_stmt(stderr,exp,0)
No symbol "exp" in current context.
(gdb) up
#1  0x084a1c40 in fold (expr=0x406e2804) at
/home/guerby/work/gcc/version-head/gcc/fold-const.c:6955
6955          if ((! FLOAT_TYPE_P (type) || flag_unsafe_math_optimizations)
(gdb) p exp
No symbol "exp" in current context.
(gdb) up
#2  0x0842b190 in set_mem_attributes_minus_bitpos (ref=0x40abb6e4, t=0x40ac416c,
objectp=0, bitpos=0)
    at /home/guerby/work/gcc/version-head/gcc/emit-rtl.c:1636
1636                    index = fold (build2 (MINUS_EXPR, TREE_TYPE (index),
(gdb) p exp
No symbol "exp" in current context.
(gdb) up
#3  0x0842b6fe in set_mem_attributes (ref=0x40abb6e4, t=0x40ac416c, objectp=0)
at /home/guerby/work/gcc/version-head/gcc/emit-rtl.c:1723
1723      set_mem_attributes_minus_bitpos (ref, t, objectp, 0);
(gdb) p exp
No symbol "exp" in current context.
(gdb) up
#4  0x0845eca8 in expand_expr_real_1 (exp=0x4065a300, target=0x40ab0620,
tmode=QImode, modifier=EXPAND_NORMAL, alt_rtl=0xbfffe910)
    at /home/guerby/work/gcc/version-head/gcc/expr.c:6769
6769            set_mem_attributes (temp, orig, 0);
(gdb) p exp
$4 = 0x4065a300
(gdb) call print_generic_stmt(stderr,exp,0)
*(&(*D.3846)[D.7937]{lb: temp.808 sz: 1} + (character *) ivtmp.789);

(gdb) r -I- -I. -Iada -I/home/guerby/work/gcc/version-head/gcc/ada -quiet
-dumpbase osint.adb -O2 -fomit-frame-pointer -g -gnatpg -gnata -mtune=pentiumpro
-fno-ivopts -gnatO ada/osint.o
/home/guerby/work/gcc/version-head/gcc/ada/osint.adb -o /tmp/ccTN4cek.s
Starting program:
/home/guerby/work/gcc/build/build-20040921T124859/gcc/stage1/gnat1 -I- -I. -Iada
-I/home/guerby/work/gcc/version-head/gcc/ada -quiet -dumpbase osint.adb -O2
-fomit-frame-pointer -g -gnatpg -gnata -mtune=pentiumpro -fno-ivopts -gnatO
ada/osint.o /home/guerby/work/gcc/version-head/gcc/ada/osint.adb -o /tmp/ccTN4cek.s

Program exited normally.

I'll try a full bootstrap with -fno-ivopts.

Laurent

-- 
           Summary: Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when
                    -fno-ivopts
           Product: gcc
           Version: 4.0.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: bootstrap
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: laurent at guerby dot net
                CC: gcc-bugs at gcc dot gnu dot org,rakdver at atrey dot
                    karlin dot mff dot cuni dot cz
  GCC host triplet: i686-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug bootstrap/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-21 18:49 -------
I am going to assume you tried to bootstrap with --disable-checking so this is a dup of bug 17399.

*** This bug has been marked as a duplicate of 17399 ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |DUPLICATE


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug bootstrap/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[laurent at guerby dot net]

------- Additional Comments From laurent at guerby dot net  2004-09-21 19:02 -------
Forgot to mention: I have the following patch from Zdenek in my tree:
http://gcc.gnu.org/ml/gcc-patches/2004-09/msg01863.html

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug bootstrap/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-21 19:06 -------
Lets reopen this one for now.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |UNCONFIRMED
         Resolution|DUPLICATE                   |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug tree-optimization/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-21 19:09 -------
Hmm, this is looks like an out of ssa problem (we would have got a different error if expand_expr_real 
did not understand SSA_NAME [which it should not unless we are expanding during ivopts] ).

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|bootstrap                   |tree-optimization
           Keywords|                            |build, ice-on-valid-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug bootstrap/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[laurent at guerby dot net]

------- Additional Comments From laurent at guerby dot net  2004-09-22 21:24 -------
I deactivated ivopts with the following patch:

Index: common.opt
===================================================================
RCS file: /cvs/gcc/gcc/gcc/common.opt,v
retrieving revision 1.55
diff -u -r1.55 common.opt
--- common.opt  14 Sep 2004 08:05:45 -0000      1.55
+++ common.opt  22 Sep 2004 20:57:57 -0000
@@ -473,7 +473,7 @@
 Instrument function entry and exit with profiling calls

 fivopts
-Common Report Var(flag_ivopts) Init(1)
+Common Report Var(flag_ivopts) Init(0)
 Optimize induction variables on trees

 fkeep-inline-functions

It cured the problem, and also fixed the following
ACATS tests on x86: c37003a c43204e c48009b cc3224a cxaa010

The 19 remaining ACATS failure are identical on x86 and x86_64
and do not seem related to ivopts.

Zdenek, what information I can provide would be helpful to you on this issue?

I can try reducing some of the five tests above, here are their exact failure mode:

* c37003a
c37003a.adb: In function 'C37003A':
c37003a.adb:40: error: Definition in block 8 does not dominate use in block 59
for SSA_NAME: D.619_265
in statement:
ivtmp.99D.1044_1133 = &r1.4D.614_73->a2D.424{off: D.619_265 * 4}[1]{lb: 1 sz: 4};

+===========================GNAT BUG DETECTED==============================+
| 4.0.0 20040920 (experimental) (i686-pc-linux-gnu) verify_ssa failed.     |
| Error detected at c37003a.adb:198:5                                      |

* c43204e 
c43204e.adb: In function 'C43204E':
c43204e.adb:35: error: Definition in block 184 does not dominate use in block 490
for SSA_NAME: D.2674_2039
in statement:
ivtmp.365D.3607_12228 = &_initD.2608_788->ra22D.655{off: D.2674_2039 *
4}[J159b.239D.2678_36]{lb: SR.299D.3291_250 sz: D.1169_662 *
4}[SR.297D.3289_256]{lb: SR.297D.3289_256 sz: 4};

+===========================GNAT BUG DETECTED==============================+
| 4.0.0 20040920 (experimental) (i686-pc-linux-gnu) verify_ssa failed.     |
| Error detected at c43204e.adb:179:5                                      |

* c48009b
c48009b.adb: In function 'C48009B':
c48009b.adb:45: error: Definition in block 125 does not dominate use in block 124
for SSA_NAME: D.1048_538
in statement:
ivtmp.149D.1490_1845 = &VIEW_CONVERT_EXPR<struct
c48009b__T71bD.795>(*va_t_rec_arrD.423_480).bD.790.eD.786{off: D.1048_538 *
4}[1]{lb: 1 sz: 4};

+===========================GNAT BUG DETECTED==============================+
| 4.0.0 20040920 (experimental) (i686-pc-linux-gnu) verify_ssa failed.     |
| Error detected at c48009b.adb:255:5                                      |

* cc3224a
compiles but fails at run time:

,.,. CC3224A ACATS 2.5 04-09-20 21:39:14
---- CC3224A CHECK THAT A FORMAL ARRAY TYPE DENOTES ITS ACTUAL
                PARAMETER, AND THAT OPERATIONS OF THE FORMAL TYPE ARE
                THOSE IDENTIFIED WITH THE CORRESPONDING OPERATIONS OF
                THE ACTUAL TYPE.
   * CC3224A AND REPORT.FAILED - BOOLEAN.
   * CC3224A OR REPORT.FAILED - BOOLEAN.
**** CC3224A FAILED ****************************.

* cxaa010
cxaa010.adb: In function 'CXAA010':
cxaa010.adb:78: error: Expected an SSA_NAME object
while verifying SSA_NAME tc_id in statement
#   tc_idD.942 = V_MAY_DEF <tc_idD.942>;
#   VUSE <neighborhood_taxesD.676_978>;
tc_idD.942 = *ruatmp.227D.1832_4516;

+===========================GNAT BUG DETECTED==============================+
| 4.0.0 20040920 (experimental) (i686-pc-linux-gnu) verify_ssa failed.     |
| No source file position information available                            |

Laurent


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |laurent at guerby dot net
          Component|tree-optimization           |bootstrap


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug tree-optimization/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|bootstrap                   |tree-optimization


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug tree-optimization/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[rakdver at atrey dot karlin dot mff dot cuni dot cz]

------- Additional Comments From rakdver at atrey dot karlin dot mff dot cuni dot cz  2004-09-23 12:12 -------
Subject: Re:  Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

Hello,

> It cured the problem, and also fixed the following
> ACATS tests on x86: c37003a c43204e c48009b cc3224a cxaa010
> 
> The 19 remaining ACATS failure are identical on x86 and x86_64
> and do not seem related to ivopts.
> 
> Zdenek, what information I can provide would be helpful to you on this issue?
> 
> I can try reducing some of the five tests above, here are their exact failure mode:
> 
> * c37003a
> c37003a.adb: In function 'C37003A':
> c37003a.adb:40: error: Definition in block 8 does not dominate use in block 59
> for SSA_NAME: D.619_265
> in statement:
> ivtmp.99D.1044_1133 = &r1.4D.614_73->a2D.424{off: D.619_265 * 4}[1]{lb: 1 sz: 4};

once someone explains me what kind of brand new ultra cool absolutely
necessary feature should this "{off: D.619_265 * 4}" stuff be, fixing
this (and folowing two failures) should be straightforward.

> +===========================GNAT BUG DETECTED==============================+
> | 4.0.0 20040920 (experimental) (i686-pc-linux-gnu) verify_ssa failed.     |
> | Error detected at c37003a.adb:198:5                                      |
> 
> +===========================GNAT BUG DETECTED==============================+
> | 4.0.0 20040920 (experimental) (i686-pc-linux-gnu) verify_ssa failed.     |
> | Error detected at c43204e.adb:179:5                                      |
> 
> +===========================GNAT BUG DETECTED==============================+
> | 4.0.0 20040920 (experimental) (i686-pc-linux-gnu) verify_ssa failed.     |
> | Error detected at c48009b.adb:255:5                                      |
> 
> * cc3224a
> compiles but fails at run time:
> 
> ,.,. CC3224A ACATS 2.5 04-09-20 21:39:14
> ---- CC3224A CHECK THAT A FORMAL ARRAY TYPE DENOTES ITS ACTUAL
>                 PARAMETER, AND THAT OPERATIONS OF THE FORMAL TYPE ARE
>                 THOSE IDENTIFIED WITH THE CORRESPONDING OPERATIONS OF
>                 THE ACTUAL TYPE.
>    * CC3224A AND REPORT.FAILED - BOOLEAN.
>    * CC3224A OR REPORT.FAILED - BOOLEAN.
> **** CC3224A FAILED ****************************.

No idea, will check once I get to computer with Ada installed.

> * cxaa010
> cxaa010.adb: In function 'CXAA010':
> cxaa010.adb:78: error: Expected an SSA_NAME object
> while verifying SSA_NAME tc_id in statement
> #   tc_idD.942 = V_MAY_DEF <tc_idD.942>;
> #   VUSE <neighborhood_taxesD.676_978>;
> tc_idD.942 = *ruatmp.227D.1832_4516;
> 
> +===========================GNAT BUG DETECTED==============================+
> | 4.0.0 20040920 (experimental) (i686-pc-linux-gnu) verify_ssa failed.     |
> | No source file position information available                            |

Ditto.

Zdenek


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug bootstrap/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[laurent at guerby dot net]

------- Additional Comments From laurent at guerby dot net  2004-09-24 22:31 -------
Zdenek, I confirm your patch (private email) fixes bootstrap and all the ACATS
I mentionned but cc3224a and cxaa010 (on x86 with ivopts on). I will try
to reduce them a bit further.

(Should I open a separate PR for them?)



-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|tree-optimization           |bootstrap


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug tree-optimization/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-24 22:40 -------
Confirmed.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
          Component|bootstrap                   |tree-optimization
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2004-09-24 22:40:22
               date|                            |
   Target Milestone|---                         |4.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug tree-optimization/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[rakdver at atrey dot karlin dot mff dot cuni dot cz]

------- Additional Comments From rakdver at atrey dot karlin dot mff dot cuni dot cz  2004-09-25 11:11 -------
Subject: Re:  Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

> ------- Additional Comments From laurent at guerby dot net  2004-09-24 22:31 -------
> Zdenek, I confirm your patch (private email) fixes bootstrap and all the ACATS
> I mentionned but cc3224a and cxaa010 (on x86 with ivopts on). I will try
> to reduce them a bit further.
> 
> (Should I open a separate PR for them?)

that would be nice, yes.

Zdenek


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug tree-optimization/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-25 15:18 -------
Patch here: <http://gcc.gnu.org/ml/gcc-patches/2004-09/msg02653.html>.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug tree-optimization/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-09-26 08:32 -------
Subject: Bug 17591

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	rakdver@gcc.gnu.org	2004-09-26 08:32:41

Modified files:
	gcc            : ChangeLog tree-ssa-loop-im.c 
	                 tree-ssa-loop-ivopts.c 

Log message:
	PR bootstrap/17591
	* tree-ssa-loop-im.c (for_each_index): Call callback for component_refs
	with varying offset.
	* tree-ssa-loop-ivopts.c (abnormal_ssa_name_p): New function.
	(idx_contains_abnormal_ssa_name_p): Scan step and lower bound for
	ARRAY_REFS.
	(expr_invariant_in_loop_p): New function.
	(idx_find_step): Handle step and lower bound for ARRAY_REFs.  Handle
	component_ref_field_offset for COMPONENT_REFs.  Do not allow
	ALIGN_INDIRECT_REFs and MISALIGNED_INDIRECT_REFs.
	(add_address_candidates): Do not handle ALIGN_INDIRECT_REFs and
	MISALIGNED_INDIRECT_REFs.
	(idx_remove_ssa_names): Handle step and lower bound for ARRAY_REFs.
	(rewrite_address_base): Do not handle ALIGN_INDIRECT_REFs and
	MISALIGNED_INDIRECT_REFs.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.5628&r2=2.5629
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-ssa-loop-im.c.diff?cvsroot=gcc&r1=2.14&r2=2.15
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-ssa-loop-ivopts.c.diff?cvsroot=gcc&r1=2.12&r2=2.13



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591

[Bug tree-optimization/17591] Ada bootstrap SEGV in stage2 on osint.adb, no SEGV when -fno-ivopts

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-26 13:44 -------
Fixed.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17591