[Bug boehm-gc/64042] New: FAIL: boehm-gc.c/gctest.c -O2 execution test

[Bug boehm-gc/64042] New: FAIL: boehm-gc.c/gctest.c -O2 execution test

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64042

            Bug ID: 64042
           Summary: FAIL: boehm-gc.c/gctest.c -O2 execution test
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: boehm-gc
          Assignee: unassigned at gcc dot gnu.org
          Reporter: vries at gcc dot gnu.org

I run into this failure once. This time, it's when testing a nonbootstrap build
from an AFAIU unrelated patch set on top of the gomp-4_0-branch.

boehm-gc.log:
...
PASS: boehm-gc.c/gctest.c -O2 (test for excess errors)
Setting LD_LIBRARY_PATH to
.:/home/vries/gcc_versions/data/test-devel-oacc-11/with/nobootstrap/build/gcc:/home/vries/gcc_versions/data/test-devel-oac\
c-11/with/nobootstrap/build/gcc/32:/home/vries/gcc_versions/data/test-devel-oacc-11/with/nobootstrap/build/x86_64-unknown-linux-gnu/./boehm-gc/.libs:\
.libs:.:/home/vries/gcc_versions/data/test-devel-oacc-11/with/nobootstrap/build/gcc:/home/vries/gcc_versions/data/test-devel-oacc-11/with/nobootstrap\
/build/gcc/32:/home/vries/gcc_versions/data/test-devel-oacc-11/with/nobootstrap/build/x86_64-unknown-linux-gnu/./boehm-gc/.libs:.libs:/home/vries/gcc\
_versions/infra/lib
spawn [open ...]^M
Switched to incremental mode
Emulating dirty bits with mprotect/signals
Segfault at 0x2b2d48936000
Unexpected bus error or segmentation fault
FAIL: boehm-gc.c/gctest.c -O2 execution test
...

core backtrace:
...
Core was generated by
`/home/vries/gcc_versions/data/test-devel-oacc-11/with/nobootstrap/build/x86_64-'.
Program terminated with signal SIGABRT, Aborted.
#0  0x00002b2d47c31bb9 in __GI_raise (sig=sig@entry=6) at
../nptl/sysdeps/unix/sysv/linux/raise.c:56
56    ../nptl/sysdeps/unix/sysv/linux/raise.c: No such file or directory.
(gdb) bt
#0  0x00002b2d47c31bb9 in __GI_raise (sig=sig@entry=6) at
../nptl/sysdeps/unix/sysv/linux/raise.c:56
#1  0x00002b2d47c34fc8 in __GI_abort () at abort.c:89
#2  0x00002b2d47093c59 in GC_abort (msg=msg@entry=0x2b2d4709bb48 "Unexpected
bus error or segmentation fault")
    at
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/boehm-gc/misc.c:1081
#3  0x00002b2d4709554e in GC_write_fault_handler (sig=<optimized out>,
si=0x2b2d481c06f0, scp=0x2b2d481c05c0)
    at
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/boehm-gc/os_dep.c:2643
#4  <signal handler called>
#5  GC_typed_mark_proc (addr=0x2b2d48936000, mark_stack_ptr=0x2b2d485e1a00,
mark_stack_limit=0x2b2d487dd000, env=4)
    at
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/boehm-gc/typd_mlc.c:415
#6  0x00002b2d47091aeb in GC_mark_from (mark_stack_top=<optimized out>,
mark_stack=0x2b2d485dd000, mark_stack_limit=0x2b2d487dd000)
    at
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/boehm-gc/mark.c:699
#7  0x00002b2d47092be5 in GC_mark_some
(cold_gc_frame=cold_gc_frame@entry=0x2b2d481c0c2c "")
    at
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/boehm-gc/mark.c:361
#8  0x00002b2d4708a648 in GC_stopped_mark
(stop_func=stop_func@entry=0x2b2d4708a3c0 <GC_timeout_stop_func>)
    at
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/boehm-gc/alloc.c:531
#9  0x00002b2d4708aeac in GC_maybe_gc () at
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/boehm-gc/alloc.c:307
#10 0x00002b2d4708b81d in GC_allocobj (sz=sz@entry=39, kind=kind@entry=6)
    at
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/boehm-gc/alloc.c:1073
#11 0x00002b2d4708fced in GC_generic_malloc_inner (lb=lb@entry=271,
k=k@entry=6)
    at
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/boehm-gc/malloc.c:136
#12 0x00002b2d4708fe24 in GC_generic_malloc (lb=271, k=6) at
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/boehm-gc/malloc.c:192
#13 0x00002b2d4709819e in GC_malloc_explicitly_typed (lb=<optimized out>,
d=18446462525718396929)
    at
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/boehm-gc/typd_mlc.c:653
#14 0x000000000040363e in typed_test ()
#15 0x0000000000000000 in ?? ()
...

Investigating the segfault site:
...
#4  <signal handler called>
(gdb) up
#5  GC_typed_mark_proc (addr=0x2b2d48936000, mark_stack_ptr=0x2b2d485e1a00,
mark_stack_limit=0x2b2d487dd000, env=4)
    at
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/boehm-gc/typd_mlc.c:415
415                current = *current_p;
(gdb) p current_p
$1 = (word *) 0x2b2d48936000
(gdb) p *current_p
$2 = 0
(gdb) 
...

[Bug boehm-gc/64042] FAIL: boehm-gc.c/gctest.c -O2 execution test

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64042

--- Comment #1 from vries at gcc dot gnu.org ---
configure line:
...
Target: x86_64-unknown-linux-gnu
Configured with:
/home/vries/gcc_versions/data/test-devel-oacc-11/with/src/configure
--prefix=/home/vries/gcc_versions/data/test-devel-oacc-11/with/nobootstrap/install
--with-cloog=/home/vries/gcc_versions/infra
--with-ppl=/home/vries/gcc_versions/infra
--with-gmp=/home/vries/gcc_versions/infra
--with-mpfr=/home/vries/gcc_versions/infra
--with-mpc=/home/vries/gcc_versions/infra --disable-bootstrap
--enable-checking=yes,rtl --enable-languages=c,fortran,ada,java,objc,c++
Thread model: posix
...

[Bug boehm-gc/64042] FAIL: boehm-gc.c/gctest.c -O2 execution test

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64042

--- Comment #2 from vries at gcc dot gnu.org ---
Googling the sigsegv site GC_typed_mark_proc finds these threads, which may be
related:
- http://www.hpl.hp.com/hosted/linux/mail-archives/gc/2010-March/003796.html
- http://www.hpl.hp.com/hosted/linux/mail-archives/gc/2010-March/003799.html

[Bug boehm-gc/64042] FAIL: boehm-gc.c/gctest.c -O2 execution test

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64042

--- Comment #3 from vries at gcc dot gnu.org ---
> I run into this failure once.

I meant: I run into this failure once in a while.

It's sort of reproducible. In running the testcase a thousand times, it
triggered 3 times:
...
Segfault at 0x7f088a81f000
Unexpected bus error or segmentation fault
Segfault at 0x7febfdf73000
Unexpected bus error or segmentation fault
Segfault at 0x7f0af403c000
Unexpected bus error or segmentation fault
...

[Bug boehm-gc/64042] FAIL: boehm-gc.c/gctest.c -O2 execution test

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64042

vries at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Version|unknown                     |5.0

--- Comment #4 from vries at gcc dot gnu.org ---
Reproduced with clean trunk bootstrap build, happened again 3 times in 1000
iterations:
...
Segfault at 0x7f0ef467b000
Unexpected bus error or segmentation fault
Segfault at 0x7fbb47d1f000
Unexpected bus error or segmentation fault
Segfault at 0x7ff71ee51000
Unexpected bus error or segmentation fault
...

Marking version 5.0

[Bug boehm-gc/64042] FAIL: boehm-gc.c/gctest.c -O2 execution test

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64042

--- Comment #5 from vries at gcc dot gnu.org ---
An easy way to reproduce, is to add this patch:
...
diff -u
/home/vries/gcc_versions/data/ref-master-14-12-02/src/boehm-gc/testsuite/boehm-gc.c/c.exp
./c.exp
---
/home/vries/gcc_versions/data/ref-master-14-12-02/src/boehm-gc/testsuite/boehm-gc.c/c.exp
   2014-12-02 02:51:25.000000000 +0100
+++ ./c.exp    2014-11-26 14:39:41.964962842 +0100
@@ -19,4 +19,8 @@

 boehm-gc-dg-runtest [lsort [glob -nocomplain $srcdir/$subdir/*.c]] "-O2" ""

+for {set i 0} {$i < 1000} {incr i} {
+    boehm-gc-dg-runtest [lsort [glob -nocomplain $srcdir/$subdir/gctest.c]]
"-O2" ""
+}
+
 dg-finish
...

This way (1000 * compile/run) it triggered twice for me.

[Bug boehm-gc/64042] FAIL: boehm-gc.c/gctest.c -O2 execution test

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64042

--- Comment #6 from vries at gcc dot gnu.org ---
Reproduced it once more (exactly as described in comment 5).

[Bug boehm-gc/64042] FAIL: boehm-gc.c/gctest.c -O2 execution test

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64042

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-01-13
     Ever confirmed|0                           |1

--- Comment #7 from H.J. Lu <hjl.tools at gmail dot com> ---
I have seen it on Linux/x86 at random:

https://gcc.gnu.org/ml/gcc-testresults/2014-11/msg00621.html

[Bug boehm-gc/64042] FAIL: boehm-gc.c/gctest.c -O2 execution test

[law at redhat dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64042

Jeffrey A. Law <law at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |law at redhat dot com

--- Comment #8 from Jeffrey A. Law <law at redhat dot com> ---
I was able to reproduce using the instructions Tom provided.  Looks like 4
failures across the 1000 iterations.

[Bug boehm-gc/64042] FAIL: boehm-gc.c/gctest.c -O2 execution test

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64042

vries at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |aph at redhat dot com,
                   |                            |pinskia at gmail dot com

--- Comment #9 from vries at gcc dot gnu.org ---
CC-ing libobjc, java maintainer

[Bug boehm-gc/64042] FAIL: boehm-gc.c/gctest.c -O2 execution test

[iains at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64042

Iain Sandoe <iains at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |iains at gcc dot gnu.org

--- Comment #10 from Iain Sandoe <iains at gcc dot gnu.org> ---
I got 1 fail in a 1000 on x86_64-darwin12, so it is not Linux-specific.

[Bug boehm-gc/64042] FAIL: boehm-gc.c/gctest.c -O2 execution test

[vries at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64042

--- Comment #11 from vries at gcc dot gnu.org ---
Reported upstream here:
https://lists.opendylan.org/pipermail/bdwgc/2015-January/006071.html