[Bug bootstrap/107193] New: [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[Bug bootstrap/107193] New: [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[crazylht at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

            Bug ID: 107193
           Summary: [13 regression] bootstrap error caused by
                    r13-3172-gf30e9fd33e56a5
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: bootstrap
          Assignee: unassigned at gcc dot gnu.org
          Reporter: crazylht at gmail dot com
  Target Milestone: ---

Created attachment 53683
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53683&action=edit
_sd_to_udi.i

export/users2/liuhongt/tools-build/build_git_trunk_master_native_bootstrap/./gcc/xgcc
-B/export/users2/liuhongt/tools-build/build_git_trunk_master_native_bootstrap/./gcc/
-B/export/users2/liuhongt/install/git_trunk_master_native_bootstrap/x86_64-pc-linux-gnu/bin/
-B/export/users2/liuhongt/install/git_trunk_master_native_bootstrap/x86_64-pc-linux-gnu/lib/
-isystem
/export/users2/liuhongt/install/git_trunk_master_native_bootstrap/x86_64-pc-linux-gnu/include
-isystem
/export/users2/liuhongt/install/git_trunk_master_native_bootstrap/x86_64-pc-linux-gnu/sys-include
  -fno-checking -g -O2 -m32 -O2  -g -O2 -DIN_GCC    -W -Wall -Wno-narrowing
-Wwrite-strings -Wcast-qual -Wstrict-prototypes -Wmissing-prototypes
-Wold-style-definition  -isystem ./include   -fpic -mlong-double-80
-DUSE_ELF_SYMVER -fcf-protection -mshstk -g -DIN_LIBGCC2 -fbuilding-libgcc
-fno-stack-protector   -fpic -mlong-double-80 -DUSE_ELF_SYMVER -fcf-protection
-mshstk -I. -I. -I../../.././gcc
-I/export/users2/liuhongt/gcc/git_trunk/master/libgcc
-I/export/users2/liuhongt/gcc/git_trunk/master/libgcc/.
-I/export/users2/liuhongt/gcc/git_trunk/master/libgcc/../gcc
-I/export/users2/liuhongt/gcc/git_trunk/master/libgcc/../include
-I/export/users2/liuhongt/gcc/git_trunk/master/libgcc/config/libbid
-DENABLE_DECIMAL_BID_FORMAT -DHAVE_CC_TLS  -DUSE_TLS  -o _sd_to_udi.o -MT
_sd_to_udi.o -MD -MP -MF _sd_to_udi.dep -DFINE_GRAINED_LIBRARIES -DL_sd_to_udi
-DWIDTH=32 -c
/export/users2/liuhongt/gcc/git_trunk/master/libgcc/config/libbid/_sd_to_udi.c
during GIMPLE pass: cfg
/export/users2/liuhongt/gcc/git_trunk/master/libgcc/config/libbid/_sd_to_udi.c:
In function ‘__bid_fixunssddi’:
/export/users2/liuhongt/gcc/git_trunk/master/libgcc/config/libbid/_sd_to_udi.c:29:1:
internal compiler error: Segmentation fault
   29 | __bid_fixunssddi (_Decimal32 x) {
      | ^~~~~~~~~~~~~~~~
Please submit a full bug report, with preprocessed source (by using
-freport-bug).
See <https://gcc.gnu.org/bugs/> for instructions.
make[5]: *** [Makefile:722: _sd_to_udi.o] Error 1
make[5]: Leaving directory
'/export/users2/liuhongt/tools-build/build_git_trunk_master_native_bootstrap/x86_64-pc-linux-gnu/32/libgcc'
make[4]: *** [Makefile:1214: multi-do] Error 1
make[4]: Leaving directory
'/export/users2/liuhongt/tools-build/build_git_trunk_master_native_bootstrap/x86_64-pc-linux-gnu/libgcc'
make[3]: *** [Makefile:127: all-multi] Error 2
make[3]: Leaving directory
'/export/users2/liuhongt/tools-build/build_git_trunk_master_native_bootstrap/x86_64-pc-linux-gnu/libgcc'
make[2]: *** [Makefile:24226: all-stage2-target-libgcc] Error 2
make[2]: Leaving directory
'/export/users2/liuhongt/tools-build/build_git_trunk_master_native_bootstrap'
make[1]: *** [Makefile:30117: stage2-bubble] Error 2
make[1]: Leaving directory
'/export/users2/liuhongt/tools-build/build_git_trunk_master_native_bootstrap'
make: *** [Makefile:1071: all] Error 2


reproduce with attached testcase

./gcc/xgcc -B ./gcc -O2 _sd_to_udi.i -g -m32

And It looks like there's some random when reproducing.(sometimes pass,
sometimes failed).

[Bug bootstrap/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |13.0

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |blocker

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

Richard Biener <rguenth at gcc dot gnu.org> changed:

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

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[crazylht at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

--- Comment #1 from Hongtao.liu <crazylht at gmail dot com> ---
Looks similar as
https://gcc.gnu.org/pipermail/gcc-regression/2022-October/077050.html

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[crazylht at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

--- Comment #2 from Hongtao.liu <crazylht at gmail dot com> ---

> 
> reproduce with attached testcase
> 
> ./gcc/xgcc -B ./gcc -O2 _sd_to_udi.i -g -m32
> 
> And It looks like there's some random when reproducing.(sometimes pass,
> sometimes failed).

configure of gcc:
 --with-cpu=native --with-arch=native --disable-libsanitizer
--enable-checking=yes,rtl,extra

On cascadelake.
Remove --with-cpu=native --with-arch=native seems hide the issue.

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[crazylht at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

--- Comment #3 from Hongtao.liu <crazylht at gmail dot com> ---
1212      /* Traverse the basic block, if two function calls within a basic
block
 1213        are mapped to the same line, assign a new discriminator because a
call
 1214        stmt could be a split point of a basic block.  */
 1215      for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi); gsi_next (&gsi))
 1216        {
 1217          gimple *stmt = gsi_stmt (gsi);
 1218          expanded_location curr_locus_e;


Shouldn't curr_locus_e be defined outside of the loop?
orelse, the second iteration, it will be uninitialized and pased to
same_line_p？
 1219          if (curr_locus == UNKNOWN_LOCATION)
 1220            {
 1221              curr_locus = gimple_location (stmt);
 1222              curr_locus_e = expand_location (curr_locus);
 1223            }
 1224          else if (!same_line_p (curr_locus, &curr_locus_e,
gimple_location (stmt)))
 1225            {
 1226              curr_locus = gimple_location (stmt);
 1227              curr_locus_e = expand_location (curr_locus);
 1228              curr_discr = 0;
 1229            }
 1230          else if (curr_discr != 0)
 1231            {
 1232              location_t loc = gimple_location (stmt);
 1233              location_t dloc = location_with_discriminator (loc,
curr_discr);
 1234              gimple_set_location (stmt, dloc);
 1235            }
 1236          /* Allocate a new discriminator for CALL stmt.  */
 1237          if (gimple_code (stmt) == GIMPLE_CALL)
 1238            curr_discr = next_discriminator_for_locus (curr_locus);
 1239        }

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2022-10-10
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
                 CC|                            |marxin at gcc dot gnu.org

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[erozen at microsoft dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

--- Comment #4 from Eugene Rozenfeld <erozen at microsoft dot com> ---
Yes, that's the problem. Sorry about that, will send a patch with the fix
shortly.

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hjl.tools at gmail dot com

--- Comment #5 from H.J. Lu <hjl.tools at gmail dot com> ---
*** Bug 107205 has been marked as a duplicate of this bug. ***

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

--- Comment #6 from H.J. Lu <hjl.tools at gmail dot com> ---
     gimple *last = last_stmt (bb);
      location_t locus = last ? gimple_location (last) : UNKNOWN_LOCATION;
      location_t curr_locus = UNKNOWN_LOCATION;
      int curr_discr = 0;

      /* Traverse the basic block, if two function calls within a basic block
        are mapped to the same line, assign a new discriminator because a call
        stmt could be a split point of a basic block.  */
      for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi); gsi_next (&gsi))
        {
          gimple *stmt = gsi_stmt (gsi);
          expanded_location curr_locus_e;
          if (curr_locus == UNKNOWN_LOCATION)
            {
              curr_locus = gimple_location (stmt);
              curr_locus_e = expand_location (curr_locus);
            }
          else if (!same_line_p (curr_locus, &curr_locus_e, gimple_location
(stmt)))
            {
              curr_locus = gimple_location (stmt);
              curr_locus_e = expand_location (curr_locus);
              curr_discr = 0;
            }
          else if (curr_discr != 0)
            {
              location_t loc = gimple_location (stmt);
              location_t dloc = location_with_discriminator (loc, curr_discr);
              gimple_set_location (stmt, dloc);
            }
          /* Allocate a new discriminator for CALL stmt.  */
          if (gimple_code (stmt) == GIMPLE_CALL)
            curr_discr = next_discriminator_for_locus (curr_locus);
        }

      if (locus == UNKNOWN_LOCATION)
          Will it be changed by the loop above?
        continue;

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[erozen at microsoft dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

--- Comment #7 from Eugene Rozenfeld <erozen at microsoft dot com> ---
No, locus won't be changed by the loop. But the purpose of the loop is to
change statement locations (by adding discriminators) in this line:

  gimple_set_location (stmt, dloc);

I think the code would more clear if this line
  location_t locus = last ? gimple_location (last) : UNKNOWN_LOCATION;
were moved to after the loop.

I'll include that change in the patch.

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[erozen at microsoft dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

--- Comment #8 from Eugene Rozenfeld <erozen at microsoft dot com> ---
Created attachment 53690
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53690&action=edit
Proposed patch

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

--- Comment #9 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Eugene Rozenfeld <erozen@gcc.gnu.org>:

https://gcc.gnu.org/g:80f414e6d73f9f1683f93d83ce63a6a482e54bee

commit r13-3213-g80f414e6d73f9f1683f93d83ce63a6a482e54bee
Author: Eugene Rozenfeld <erozen@microsoft.com>
Date:   Mon Oct 10 14:10:31 2022 -0700

    Fix PR107193.

    The bug was introduced in f30e9fd33e56a5a721346ea6140722e1b193db42.
    A variable (cur_locus_e) was incorrectly declared inside a loop.
    I also moved two other declarations (last and locus) down to make
    the code more clear.

    Tested on x86_64-pc-linux-gnu.

    gcc/ChangeLog:
            PR debug/107193
            * tree-cfg.cc (assign_discriminators): Move declaration of
cur_locus_e
            out of the loop.

[Bug debug/107193] [13 regression] bootstrap error caused by r13-3172-gf30e9fd33e56a5

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107193

Martin Liška <marxin at gcc dot gnu.org> changed:

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

--- Comment #10 from Martin Liška <marxin at gcc dot gnu.org> ---
I've just re-run ubsan-bootstrap and it's fixed now.
Thank you.