[Bug boehm-gc/48299] New: [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[Bug boehm-gc/48299] New: [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

           Summary: [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: boehm-gc
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: hjl.tools@gmail.com


On Linux/x86, I got

In file included from
/export/gnu/import/git/gcc-avx/boehm-gc/testsuite/../include/gc.h:56:0,^M
                 from
/export/gnu/import/git/gcc-avx/boehm-gc/testsuite/../include/leak_detector.h:2,^M
                 from
/export/gnu/import/git/gcc-avx/boehm-gc/testsuite/boehm-gc.c/thread_leak_test.c:2:^M
/export/build/gnu/gcc-avx/build-x86_64-linux/x86_64-unknown-linux-gnu/32/boehm-gc/include/gc_config.h:44:0:
warning: "GC_LINUX_THREADS" redefined [enabled by default]^M
/export/gnu/import/git/gcc-avx/boehm-gc/testsuite/boehm-gc.c/thread_leak_test.c:1:0:
note: this is the location of the previous definition^M

FAIL: boehm-gc.c/thread_leak_test.c -O2 (test for excess errors)

Revision 171513 is OK. It may be caused by revision 171514:

http://gcc.gnu.org/ml/gcc-cvs/2011-03/msg00938.html

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[ro at CeBiTec dot Uni-Bielefeld.DE]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #1 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> 2011-03-28 08:49:47 UTC ---
Could you please check if this test worked before my patch?  It may have
been that the failure simply went unnoticed.

Thanks.
    Rainer

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #2 from H.J. Lu <hjl.tools at gmail dot com> 2011-03-28 12:31:35 UTC ---
(In reply to comment #1)
> Could you please check if this test worked before my patch?  It may have
> been that the failure simply went unnoticed.
> 

I don't think thread_leak_test.c was tested before.

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[ro at CeBiTec dot Uni-Bielefeld.DE]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #3 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> 2011-03-28 12:40:06 UTC ---
> I don't think thread_leak_test.c was tested before.

In that case, please try to compile it with the same flags used for one
of the other tests (e.g, leak_test.c) to check if the problem is with my
dg conversion or elsewhere.

In any case, if this wasn't tested before, it's not a regression.

Thanks.
    Rainer

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #4 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2011-03-29 11:21:53 UTC ---
At revision 171632 the test also failed on x86_64-apple-darwin10.7.0:

...
Executing on host: ../libtool --silent --tag=CC --mode=link 
/opt/gcc/build_w/gcc/xgcc -B/opt/gcc/build_w/gcc/
/opt/gcc/work/boehm-gc/testsuite/boehm-
gc.c/thread_leak_test.c
/opt/gcc/build_w/x86_64-apple-darwin10.7.0/./boehm-gc/libgcjgc.la  -O2 
-I/opt/gcc/build_w/x86_64-apple-darwin10.7.0/./boehm-g
c/include -I/opt/gcc/work/boehm-gc/testsuite/../include   -Wc,-shared-libgcc
-lpthread -lm   -m64 -o ./thread_leak_test    (timeout = 300)
PASS: boehm-gc.c/thread_leak_test.c -O2 (test for excess errors)
Setting LD_LIBRARY_PATH to
.:/opt/gcc/build_w/gcc:/opt/gcc/build_w/x86_64-apple-darwin10.7.0/./boehm-gc/.libs:.libs:.:/opt/gcc/build_w/gcc:/opt/gcc/bu
ild_w/x86_64-apple-darwin10.7.0/./boehm-gc/.libs:.libs
Leaked composite object at 0x1000c0fe0
(/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/thread_leak_test.c:12, sz=4,
NORMAL)

Leaked composite object at 0x1000c0ec0
(/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/thread_leak_test.c:12, sz=4,
NORMAL)
Leaked composite object at 0x1000c0f20
(/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/thread_leak_test.c:12, sz=4,
NORMAL)

Leaked composite object at start: 0x1000c0f00, appr. length: 48
WARNING: program timed out.
FAIL: boehm-gc.c/thread_leak_test.c -O2 execution testExecuting on host:
../libtool --mode=clean rm -f  thread_leak_test    (timeout = 300)
libtool: clean: rm -f thread_leak_test .libs/thread_leak_test
.libs/thread_leak_testS.o^M
libtool: clean: rmdir .libs >/dev/null 2>&1
...

while it passed at revision 171578. Manual runs give

[macbook] boehm-gc/testsuite% ../libtool --silent --tag=CC --mode=link 
/opt/gcc/build_w/gcc/xgcc -B/opt/gcc/build_w/gcc/
/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/leak_test.c
/opt/gcc/build_w/x86_64-apple-darwin10.7.0/i386/boehm-gc/libgcjgc.la  -O2 
-I/opt/gcc/build_w/x86_64-apple-darwin10.7.0/i386/boehm-gc/include
-I/opt/gcc/work/boehm-gc/testsuite/../include   -Wc,-shared-libgcc -lpthread
-lm   -m32 -o ./leak_test
[macbook] boehm-gc/testsuite% time ./leak_test                                 
                                                               Leaked composite
object at 0x92fd0 (/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/leak_test.c:16,
sz=4, NORMAL)
0.012u 0.021s 0:00.06 50.0%    0+0k 0+1io 0pf+0w
[macbook] boehm-gc/testsuite% ../libtool --silent --tag=CC --mode=link
/opt/gcc/build_w/gcc/xgcc -B/opt/gcc/build_w/gcc/
/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/leak_test.c
/opt/gcc/build_w/x86_64-apple-darwin10.7.0/i386/boehm-gc/libgcjgc.la -O2
-I/opt/gcc/build_w/x86_64-apple-darwin10.7.0/i386/boehm-gc/include
-I/opt/gcc/work/boehm-gc/testsuite/../include -Wc,-shared-libgcc -lpthread -lm
-m32 -o ./thread_leak_test
[macbook] boehm-gc/testsuite% time ./thread_leak_test                          
                                                                      Leaked
composite object at 0x92fd0
(/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/leak_test.c:16, sz=4, NORMAL)
0.012u 0.022s 0:00.04 75.0%    0+0k 0+0io 0pf+0w
[macbook] boehm-gc/testsuite% ../libtool --silent --tag=CC --mode=link
/opt/gcc/build_w/gcc/xgcc -B/opt/gcc/build_w/gcc/
/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/leak_test.c
/opt/gcc/build_w/x86_64-apple-darwin10.7.0/./boehm-gc/libgcjgc.la -O2
-I/opt/gcc/build_w/x86_64-apple-darwin10.7.0/./boehm-gc/include
-I/opt/gcc/work/boehm-gc/testsuite/../include -Wc,-shared-libgcc -lpthread -lm
-m64 -o ./leak_test
[macbook] boehm-gc/testsuite% time ./leak_test                                 
                                                                      Leaked
composite object at 0x1000beef0
(/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/leak_test.c:16, sz=8, NORMAL)
0.013u 0.024s 0:00.06 50.0%    0+0k 0+2io 0pf+0w
[macbook] boehm-gc/testsuite% ../libtool --silent --tag=CC --mode=link
/opt/gcc/build_w/gcc/xgcc -B/opt/gcc/build_w/gcc/
/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/thread_leak_test.c
/opt/gcc/build_w/x86_64-apple-darwin10.7.0/./boehm-gc/libgcjgc.la -O2
-I/opt/gcc/build_w/x86_64-apple-darwin10.7.0/./boehm-gc/include
-I/opt/gcc/work/boehm-gc/testsuite/../include -Wc,-shared-libgcc -lpthread -lm
-m64 -o ./thread_leak_test
[macbook] boehm-gc/testsuite% time ./thread_leak_testLeaked composite object at
0x1000c0fe0 (/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/thread_leak_test.c:11,
sz=4, NORMAL)
Leaked composite object at 0x1000c0f80
(/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/thread_leak_test.c:11, sz=4,
NORMAL)

Leaked composite object at 0x1000c0ef0
(/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/thread_leak_test.c:11, sz=4,
NORMAL)

Leaked composite object at 0x1000c0d10
(/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/thread_leak_test.c:11, sz=4,
NORMAL)

Leaked composite object at 0x1000c0e30
(/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/thread_leak_test.c:11, sz=4,
NORMAL)

0.013u 0.023s 0:00.04 75.0%    0+0k 0+0io 0pf+0w

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[ro at CeBiTec dot Uni-Bielefeld.DE]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #5 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> 2011-03-29 11:30:31 UTC ---
> --- Comment #4 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2011-03-29 11:21:53 UTC ---
> At revision 171632 the test also failed on x86_64-apple-darwin10.7.0:
[...]
> while it passed at revision 171578. Manual runs give

Those are both after the dg-based testsuite went in.

This either suggests a different change being responsible or a timing
issue.  Some of those tests can be quite sensitive to load.

Could you try to rerun the testcase manually several times and see if
the outcome is consistent or differs?

Thanks.
        Rainer

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #6 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2011-03-29 11:39:58 UTC ---
> This either suggests a different change being responsible or a timing
> issue.  Some of those tests can be quite sensitive to load.
>
> Could you try to rerun the testcase manually several times and see if
> the outcome is consistent or differs?

The failing test occured on a "quiet" state: i.e., terminal, safari, and xchat
opened but not used.
The tests run in a fraction of a second far away of the 300s timeout.

BTW I did not find a way to run only the boehm test suite: if I make check in
x86_64-apple-darwin10.7.0/boehm-gc, I get

WARNING: could not find `runtest'

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[ro at CeBiTec dot Uni-Bielefeld.DE]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #7 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> 2011-03-29 11:46:29 UTC ---
> The failing test occured on a "quiet" state: i.e., terminal, safari, and xchat
> opened but not used.
> The tests run in a fraction of a second far away of the 300s timeout.

That's not what I meant: there are tests that fail only once in a
while.  Rerun it 50 times and you observe failures only during a
fraction of those runs.  You'd see timeouts in the testsuite logs if
they occured.

> BTW I did not find a way to run only the boehm test suite: if I make check in
> x86_64-apple-darwin10.7.0/boehm-gc, I get
>
> WARNING: could not find `runtest'

Run make check in boehm-gc/testsuite instead.  Better yet, just build
the failing test once and manually rerun it in a loop with
LD_LIBRARY_PATH or equivalent set.

    Rainer

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #8 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2011-03-29 15:12:00 UTC ---
> Run make check in boehm-gc/testsuite instead.  Better yet, just build
> the failing test once and manually rerun it in a loop with
> LD_LIBRARY_PATH or equivalent set.

I have run the following script

#!/bin/sh
i=0
thread_leak_test
while [ $? == 0 ]
do
  i=`expr $i + 1`
  echo $i
  thread_leak_test
done

and saw it fail for i between 8 and 153. The symptom is always

Leaked composite object at start: 0x1000c0f?0, appr. length: 48

and the test starts to consume 100% of the cpu untill I stop it. Sampling the
test gives

Sampling process 85239 for 3 seconds with 1 millisecond of run time between
samples
Sampling completed, processing symbols...
Analysis of sampling thread_leak_test (pid 85239) every 1 millisecond
Process:         thread_leak_test [85239]
Path:           
/opt/gcc/build_w/x86_64-apple-darwin10.7.0/boehm-gc/testsuite/.libs/thread_leak_test
Load Address:    0x100000000
Identifier:      thread_leak_test
Version:         ??? (???)
Code Type:       X86-64 (Native)
Parent Process:  tcsh [191]

Date/Time:       2011-03-29 16:31:27.577 +0200
OS Version:      Mac OS X 10.6.7 (10J869)
Report Version:  6

Call graph:
    2682 Thread_17437964   DispatchQueue_1: com.apple.main-thread  (serial)
      2682 GC_obj_kinds
        2682 GC_try_to_collect_inner
          2682 GC_finish_collection
            2682 GC_set_fl_marks

Total number in stack (recursive counted multiple, when >=5):

Sort by top of stack, same collapsed (when >= 5):
        GC_set_fl_marks        2682

Binary Images:
       0x100000000 -        0x100000ff7 +thread_leak_test ??? (???)
<C97DEDBE-04FF-EFEE-1C1B-68710038E8C2>
/opt/gcc/build_w/x86_64-apple-darwin10.7.0/boehm-gc/testsuite/.libs/thread_leak_test
       0x100003000 -        0x10001dff7 +libgcjgc.1.dylib 2.2.0 (compatibility
2.0.0) <7CAAF920-9603-D889-1D10-4250288EE1C8>
/opt/gcc/build_w/x86_64-apple-darwin10.7.0/boehm-gc/.libs/libgcjgc.1.dylib
       0x10004b000 -        0x10005ffe7 +libgcc_s.1.dylib ??? (???)
<559F9DAA-51E8-95F3-B24B-9DDA7CCC1341> /opt/gcc/gcc4.7w/lib/libgcc_s.1.dylib
    0x7fff5fc00000 -     0x7fff5fc3bdef  dyld 132.1 (???)
<B536F2F1-9DF1-3B6C-1C2C-9075EA219A06> /usr/lib/dyld
    0x7fff80415000 -     0x7fff80419ff7  libmathCommon.A.dylib 315.0.0
(compatibility 1.0.0) <95718673-FEEE-B6ED-B127-BCDBDB60D4E5>
/usr/lib/system/libmathCommon.A.dylib
    0x7fff80fc4000 -     0x7fff81185fff  libSystem.B.dylib 125.2.10
(compatibility 1.0.0) <9BAEB2F2-B485-6349-E1AB-637FE12EE770>
/usr/lib/libSystem.B.dylib
    0x7fffffe00000 -     0x7fffffe01fff  libSystem.B.dylib ??? (???)
<9BAEB2F2-B485-6349-E1AB-637FE12EE770> /usr/lib/libSystem.B.dylib
Sample analysis of process 85239 written to file /dev/stdout

Note that I have applied the patch in
http://gcc.gnu.org/ml/gcc-patches/2011-03/msg01903.html without any change.

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[ro at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

Rainer Orth <ro at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |Hans.Boehm at hp dot com,
                   |                            |ro at gcc dot gnu.org

--- Comment #9 from Rainer Orth <ro at gcc dot gnu.org> 2011-03-30 15:56:25 UTC ---
(In reply to comment #8)

I've not too much experience debugging boehm-gc issues myself.  I'd suggest
asking Hans Boehm for advice (Cc:ed).

> Note that I have applied the patch in
> http://gcc.gnu.org/ml/gcc-patches/2011-03/msg01903.html without any change.

No wonder: this would only affect Linux targets.

  Rainer

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[ro at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #10 from Rainer Orth <ro at gcc dot gnu.org> 2011-03-30 15:57:54 UTC ---
Btw., I forgot: could you try to compile that testcase on the 4.6 branch with
the same options used for leak_test.c, just to make sure this isn't caused
by my dg conversion?

Thanks.
  Rainer

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #11 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2011-03-30 20:25:24 UTC ---
> Btw., I forgot: could you try to compile that testcase on the 4.6 branch with
> the same options used for leak_test.c, just to make sure this isn't caused
> by my dg conversion?

I have compiled the test case with
gcc version 4.6.0 20101106 (experimental) [trunk revision 166408p6] (GCC) and
gcc version 4.5.3 20110311 (prerelease) [gcc-4_5-branch revision 170869] (GCC)
using the libtool from 4.7 (I did not find the right flags for their local
version) and I have stopped the test from comment #8 after more than 1000
iterations (I.e., no failure) for a loaded or unloaded system. Note that the
output of the test case was of the kind

Leaked composite object at 0x1000beef0
(/opt/gcc/work/boehm-gc/testsuite/boehm-gc.c/leak_test.c:16, sz=8, NORMAL)

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.7.0

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[Hans.Boehm at hp dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #12 from Hans Boehm <Hans.Boehm at hp dot com> 2011-04-09 00:22:01 UTC ---
The duplicate definition of LINUX_THREADS looks like an old problem.  The
upstream source now says:

#ifndef GC_THREADS
#  define GC_THREADS
#endif /* GC_THREADS */

GC_THREADS eventually results in a definition of GC_LINUX_THREADS .

The other failure worries me a bit.  It strongly suggests that something was
prematurely collected, and put on a free list twice, resulting in a cycle in
the free list.  This is a common failure mode for such problems.  It also looks
to me like the actual objects leaked by this benchmark are the ones
corresponding to p[0], which should be smaller than 48 bytes.  This also
suggests at first glance that the so-called leaked object is getting
prematurely reclaimed in the failing runs.

It might be possible to debug this case by checking (for debugging purposes
only) in GC_add_leaked if GC_size(leaked) > 16, and stopping if it is.

The checked-in version of threads_leak_test, unlike the upstream version, fails
to call GC_INIT(), which may matter on a few platforms.

> -----Original Message-----
> From: pinskia at gcc dot gnu.org [mailto:gcc-bugzilla@gcc.gnu.org]
> Sent: Tuesday, April 05, 2011 5:27 PM
> To: Boehm, Hans
> Subject: [Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-
> gc.c/thread_leak_test.c
> 
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299
> 
> Andrew Pinski <pinskia at gcc dot gnu.org> changed:
> 
>            What    |Removed                     |Added
> -----------------------------------------------------------------------
> -----
>    Target Milestone|---                         |4.7.0
> 
> --
> Configure bugmail: http://gcc.gnu.org/bugzilla/userprefs.cgi?tab=email
> ------- You are receiving this mail because: -------
> You are on the CC list for the bug.

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011.08.02 13:54:32
     Ever Confirmed|0                           |1

--- Comment #13 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-08-02 13:54:32 UTC ---
Confirmed.

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[krebbel at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

Andreas Krebbel <krebbel at gcc dot gnu.org> changed:

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

--- Comment #14 from Andreas Krebbel <krebbel at gcc dot gnu.org> 2012-01-27 08:58:33 UTC ---
Fails on s390x as well.

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[howarth at nitro dot med.uc.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

Jack Howarth <howarth at nitro dot med.uc.edu> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |howarth at nitro dot
                   |                            |med.uc.edu

--- Comment #15 from Jack Howarth <howarth at nitro dot med.uc.edu> 2012-02-24 02:49:28 UTC ---
If boehm-gc.c/thread_leak_test.c wasn't tested before r171514 (and thus isn't a
regression), shouldn't we at least put a timeout on that test so it doesn't
stall the test suite?

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[howarth at nitro dot med.uc.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #16 from Jack Howarth <howarth at nitro dot med.uc.edu> 2012-02-25 02:23:06 UTC ---
Created attachment 26749
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=26749
back port of thread_leak_test.c from ivmai-bdwgc-8b168d0

The random failures in testsuite/boehm-gc.c/thread_leak_test.c on darwin are
eliminated if the current structure of tests/thread_leak_test.c in
ivmai-bdwgc-8b168d0 is back ported to gcc's boehm-gc. Note that the MS windows
specific changes were left out. In particular...

#ifdef GC_PTHREADS
# include <pthread.h>
#else
# include <windows.h>
#endif

and

#ifdef GC_PTHREADS
  void * test(void * arg)
#else
  DWORD WINAPI test(LPVOID arg)
#endif

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[howarth at nitro dot med.uc.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #17 from Jack Howarth <howarth at nitro dot med.uc.edu> 2012-02-25 02:27:14 UTC ---
(In reply to comment #16)
Note that at r184560 with the back ported thread_leak_test.c changes applied, I
am able to run the 64-bit thread_leak_test test repeatedly without failures.

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2

--- Comment #18 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-02-27 10:32:28 UTC ---
It fails everywhere.  But we don't seem to be sure it is a regression
(thus, the test was run before).  -> P2.

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[ro at CeBiTec dot Uni-Bielefeld.DE]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #19 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> 2012-02-27 11:06:43 UTC ---
> --- Comment #18 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-02-27 10:32:28 UTC ---
> It fails everywhere.  But we don't seem to be sure it is a regression
> (thus, the test was run before).  -> P2.

It certainly doesn't fail everywhere: at least it passes on all of my
targets (Solaris/SPARC and x86, IRIX and Tru64 Unix).

The test isn't run on the 4.6 branch (before the conversion of the
boehm-gc testsuite to dejagnu), but failures would have gone unnoticed
since they don't show up in mail-report.log.

    Rainer

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[howarth at nitro dot med.uc.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #20 from Jack Howarth <howarth at nitro dot med.uc.edu> 2012-02-27 13:54:20 UTC ---
Can folks on other failing targets try the backport of the thread_leak_test.c
from upstream bdwgc? This seems to solve the issue on x86_64-apple-darwin11.
Note that the test case has been refactored upstream to have GC_find_leak = 1
to be set on only once in main rather than repeatedly in the test subroutine
(among other changes).

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[howarth at nitro dot med.uc.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #21 from Jack Howarth <howarth at nitro dot med.uc.edu> 2012-02-27 16:44:56 UTC ---
Patch posted at http://gcc.gnu.org/ml/gcc-patches/2012-02/msg01341.html.

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[pmarlier at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

--- Comment #22 from pmarlier at gcc dot gnu.org 2012-02-28 15:37:57 UTC ---
Author: pmarlier
Date: Tue Feb 28 15:37:41 2012
New Revision: 184628

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=184628
Log:
2012-02-27  Jack Howarth  <howarth@bromo.med.uc.edu>
        Patrick Marlier  <patrick.marlier@gmail.com>

    PR boehm-gc/48299
    testsuite/boehm-gc.c/thread_leak_test.c: Merge upstream changes.


Modified:
    trunk/boehm-gc/ChangeLog
    trunk/boehm-gc/testsuite/boehm-gc.c/thread_leak_test.c

[Bug boehm-gc/48299] [4.7 Regression] FAIL: boehm-gc.c/thread_leak_test.c

[ubizjak at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48299

Uros Bizjak <ubizjak at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                URL|                            |http://gcc.gnu.org/ml/gcc-p
                   |                            |atches/2012-02/msg01341.htm
                   |                            |l
         Resolution|                            |FIXED

--- Comment #23 from Uros Bizjak <ubizjak at gmail dot com> 2012-03-01 19:31:34 UTC ---
Fixed.