ithreads coredumps on join of joined thread #5850

p5pRT · 2002-08-21T10:46:03Z

Migrated from rt.perl.org#16681 (status was 'resolved')

Searchable as RT16681$

p5pRT · 2002-08-21T10:46:03Z

From davem@fdgroup.co.uk

Created by davem@gizmo.fdgroup.co.uk

the following simple bit of code coredumps for me on Solaris:

#!/usr/bin/perl -w
use threads;
my $a = 'x'x 1; #just a hook for the debugger

my $t = threads->new(sub {});
$t->join;
$t->join;

What appears to be happing is that during the first call to join,
Perl_ithread_join() does

sv_unmagic(SvRV(obj),PERL_MAGIC_shared_scalar);

which appears to the free up the ithread structure associated
with the thread. So the next call to join, rather than seeing
a PERL_ITHR_JOINED flag, just randomly coredumps.

This gdb stacktrace shows a watchpoint where thread->state changes
back from 6 to 0 during the first join.

Hardware watchpoint 3: ((ithread *) 2319872)->state

Old value = 6 '\006'
New value = 0 '\000'
0xff1e18e8 in memset () from /usr/platform/SUNW,Ultra-1/lib/libc_psr.so.1
(gdb) bt
#0 0xff1e18e8 in memset () from /usr/platform/SUNW,Ultra-1/lib/libc_psr.so.1
#1 0x33db0 in Perl_gv_init (my_perl=0x1ff358, gv=0x2e512c, stash=0x28dbe0,
name=0x1c9178 "DESTROY", len=7, multi=1) at gv.c:111
#2 0x34434 in Perl_gv_fetchmeth (my_perl=0x1ff358, stash=0x28dbe0,
name=0x1c9178 "DESTROY", len=7, level=0) at gv.c:215
#3 0x34c48 in Perl_gv_fetchmeth_autoload (my_perl=0x1ff358, stash=0x28dbe0,
name=0x1c9178 "DESTROY", len=7, level=0) at gv.c:328
#4 0x39188 in Perl_Gv_AMupdate (my_perl=0x1ff358, stash=0x28dbe0) at gv.c:1334
#5 0x396d8 in Perl_gv_handler (my_perl=0x1ff358, stash=0x28dbe0, id=65)
at gv.c:1406
#6 0xf0cf0 in Perl_sv_clear (my_perl=0x1ff358, sv=0x2e5138) at sv.c:4952
#7 0xf1e1c in Perl_sv_free (my_perl=0x1ff358, sv=0x2e5138) at sv.c:5192
#8 0x38bcc in Perl_gp_free (my_perl=0x1ff358, gv=0x2e5120) at gv.c:1248
#9 0xf15c8 in Perl_sv_clear (my_perl=0x1ff358, sv=0x2e5120) at sv.c:5026
#10 0xf1e1c in Perl_sv_free (my_perl=0x1ff358, sv=0x2e5120) at sv.c:5192
#11 0xe223c in do_clean_named_objs (my_perl=0x1ff358, sv=0x2e5120) at sv.c:367
#12 0xe1d9c in S_visit (my_perl=0x1ff358, f=0xe1fac <do_clean_named_objs>)
at sv.c:292
#13 0xe2280 in Perl_sv_clean_objs (my_perl=0x1ff358) at sv.c:388
#14 0x283a0 in perl_destruct (my_perl=0x1ff358) at perl.c:469
#15 0xff0922ec in Perl_ithread_destruct (my_perl=0x1e6dc0, thread=0x236600,
why=0xff095ae0 "no reference") at threads.xs:138
#16 0xff0927ac in ithread_mg_free (my_perl=0x1e6dc0, sv=0x1e7cd0, mg=0x2f93b0)
at threads.xs:204
#17 0xefc6c in Perl_sv_unmagic (my_perl=0x1e6dc0, sv=0x1e7cd0, type=110)
at sv.c:4687
#18 0xff093dfc in Perl_ithread_join (my_perl=0x1e6dc0, obj=0x1f78ac)
at threads.xs:524
#19 0xff094b04 in XS_threads_join (my_perl=0x1e6dc0, cv=0x26c514)
at threads.xs:598
#20 0xdf7b4 in Perl_pp_entersub (my_perl=0x1e6dc0) at pp_hot.c:2773
#21 0xad384 in Perl_runops_debug (my_perl=0x1e6dc0) at dump.c:1398
#22 0x2c244 in S_run_body (my_perl=0x1e6dc0, oldscope=1) at perl.c:1681
#23 0x2bb2c in perl_run (my_perl=0x1e6dc0) at perl.c:1600
#24 0x25ac8 in main (argc=2, argv=0xffbef5a4, env=0xffbef5b0) at perlmain.c:85

Perl Info


Flags:
    category=core
    severity=medium

Site configuration information for perl v5.8.0:

Configured by davem at Mon Aug  5 12:46:42 BST 2002.

Summary of my perl5 (revision 5.0 version 8 subversion 0) configuration:
  Platform:
    osname=solaris, osvers=2.8, archname=sun4-solaris-thread-multi
    uname='sunos gizmo 5.8 generic_108528-03 sun4u sparc sunw,ultra-1 '
    config_args='-des -Dcc=gcc -Dprefix=/home/davem/perl5/perl-5.8.0t.out -Duseithreads -Uinstallusrbinperl -Doptimize=-g'
    hint=recommended, useposix=true, d_sigaction=define
    usethreads=define use5005threads=undef useithreads=define usemultiplicity=define
    useperlio=define d_sfio=undef uselargefiles=define usesocks=undef
    use64bitint=undef use64bitall=undef uselongdouble=undef
    usemymalloc=n, bincompat5005=undef
  Compiler:
    cc='gcc', ccflags ='-D_REENTRANT -DDEBUGGING -fno-strict-aliasing -I/usr/local/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64',
    optimize='-g',
    cppflags='-D_REENTRANT -DDEBUGGING -fno-strict-aliasing -I/usr/local/include'
    ccversion='', gccversion='2.95.2 19991024 (release)', gccosandvers='solaris2.8'
    intsize=4, longsize=4, ptrsize=4, doublesize=8, byteorder=4321
    d_longlong=define, longlongsize=8, d_longdbl=define, longdblsize=16
    ivtype='long', ivsize=4, nvtype='double', nvsize=8, Off_t='off_t', lseeksize=8
    alignbytes=8, prototype=define
  Linker and Libraries:
    ld='gcc', ldflags =' -L/usr/local/lib '
    libpth=/usr/local/lib /usr/lib /usr/ccs/lib
    libs=-lsocket -lnsl -ldl -lm -lrt -lpthread -lc
    perllibs=-lsocket -lnsl -ldl -lm -lrt -lpthread -lc
    libc=/lib/libc.so, so=so, useshrplib=false, libperl=libperl.a
    gnulibc_version=''
  Dynamic Linking:
    dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=undef, ccdlflags=' '
    cccdlflags='-fPIC', lddlflags='-G -L/usr/local/lib'

Locally applied patches:
    


@INC for perl v5.8.0:
    /home/davem/perl5/perl-5.8.0t.out/lib/5.8.0/sun4-solaris-thread-multi
    /home/davem/perl5/perl-5.8.0t.out/lib/5.8.0
    /home/davem/perl5/perl-5.8.0t.out/lib/site_perl/5.8.0/sun4-solaris-thread-multi
    /home/davem/perl5/perl-5.8.0t.out/lib/site_perl/5.8.0
    /home/davem/perl5/perl-5.8.0t.out/lib/site_perl
    .


Environment for perl v5.8.0:
    HOME=/home/davem
    LANG=C
    LANGUAGE (unset)
    LD_LIBRARY_PATH (unset)
    LOGDIR (unset)
    PATH=/usr/ccs/bin:/usr/local/bin:/home/davem/bin:/usr/bin::/usr/dt/bin:/usr/openwin/bin:/bin:/usr/ucb
    PERL_BADLANG (unset)
    SHELL=/bin/bash

p5pRT · 2002-08-21T14:59:09Z

From @lizmat

At 10:46 AM 8/21/02 +0000, via RT wrote:

use threads;
my $a = 'x'x 1; #just a hook for the debugger
my $t = threads\->new$sub \{\}$;
$t\->join;
$t\->join;

Under Linux I get:

$ perl -Mthreads -e 'my $t = threads->new( sub {} ); $t->join; $t->join'
Thread already joined at -e line 1.

Liz

p5pRT · 2003-01-04T01:08:14Z

arthur@contiller.se - Status changed from 'new' to 'open'

p5pRT · 2003-01-04T01:09:08Z

From arthur@contiller.se

Hi,

Can you see if this still applies? I think I might have fixed it before
5.8 came out. I can't reproduce it on linux.

Arthur

p5pRT · 2003-01-04T01:25:40Z

From @iabyn

On Sat, Jan 04, 2003 at 01:09:09AM -0000, Arthur Bergman wrote:

Hi,

Can you see if this still applies? I think I might have fixed it before
5.8 came out. I can't reproduce it on linux.

On a copy of 18378 I have lying around, I can get strange effects
on Linux by executing some code that reallocs stuff between the joins:

use threads;
my $a = 'x'x 1; #just a hook for the debugger

my $t = threads->new(sub {});
$t->join;
for (1..100) {
$a[$_] = "\0xff" x $_;
}

$t->join;

[davem@percy 18378t]$ ./perl -Ilib /tmp/p
Name "main::a" used only once: possible typo at /tmp/p line 9.
panic: MUTEX_LOCK at /tmp/p line 12.

So I guess the bug is still active :-(

Summary of my perl5 (revision 5.0 version 9 subversion 0 patch 18378) configuration:
Platform:
osname=linux, osvers=2.4.7-10, archname=i686-linux-thread-multi
uname='linux percy 2.4.7-10 #1 thu sep 6 16:46:36 edt 2001 i686 unknown '
config_args='-des -Dusedevel -Dprefix=/home/davem/perl5/18378t.out -Uinstallusrbinperl -Duseithreads -Doptimize=-g'
hint=recommended, useposix=true, d_sigaction=define
usethreads=define useithreads=define usemultiplicity=define
useperlio=define d_sfio=undef uselargefiles=define usesocks=undef
use64bitint=undef use64bitall=undef uselongdouble=undef
usemymalloc=n, bincompat5005=undef
Compiler:
cc='cc', ccflags ='-D_REENTRANT -D_GNU_SOURCE -DTHREADS_HAVE_PIDS -DDEBUGGING -fno-strict-aliasing -I/usr/local/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -I/usr/include/gdbm',
optimize='-g',
cppflags='-D_REENTRANT -D_GNU_SOURCE -DTHREADS_HAVE_PIDS -DDEBUGGING -fno-strict-aliasing -I/usr/local/include -I/usr/include/gdbm'
ccversion='', gccversion='2.96 20000731 (Red Hat Linux 7.1 2.96-98)', gccosandvers=''
intsize=4, longsize=4, ptrsize=4, doublesize=8, byteorder=1234
d_longlong=define, longlongsize=8, d_longdbl=define, longdblsize=12
ivtype='long', ivsize=4, nvtype='double', nvsize=8, Off_t='off_t', lseeksize=8
alignbytes=4, prototype=define
Linker and Libraries:
ld='cc', ldflags =' -L/usr/local/lib'
libpth=/usr/local/lib /lib /usr/lib
libs=-lnsl -lndbm -lgdbm -ldl -lm -lpthread -lc -lcrypt -lutil -lrt
perllibs=-lnsl -ldl -lm -lpthread -lc -lcrypt -lutil -lrt
libc=/lib/libc-2.2.4.so, so=so, useshrplib=false, libperl=libperl.a
gnulibc_version='2.2.4'
Dynamic Linking:
dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=undef, ccdlflags='-rdynamic'
cccdlflags='-fpic', lddlflags='-shared -L/usr/local/lib'

Characteristics of this binary (from libperl):
Compile-time options: DEBUGGING MULTIPLICITY USE_ITHREADS USE_LARGE_FILES PERL_IMPLICIT_CONTEXT
Locally applied patches:
DEVEL18374
Built under linux
Compiled at Jan 1 2003 15:50:28
@INC:
lib
/home/davem/perl5/18378t.out/lib/5.9.0/i686-linux-thread-multi
/home/davem/perl5/18378t.out/lib/5.9.0
/home/davem/perl5/18378t.out/lib/site_perl/5.9.0/i686-linux-thread-multi
/home/davem/perl5/18378t.out/lib/site_perl/5.9.0
/home/davem/perl5/18378t.out/lib/site_perl
.

--
A walk of a thousand miles begins with a single step...
then continues for another 1,999,999 or so.

p5pRT · 2003-01-04T01:29:31Z

From arthur@contiller.se

On lördag, jan 4, 2003, at 02:22 Europe/Stockholm, Dave Mitchell wrote:

On Sat, Jan 04, 2003 at 01:09:09AM -0000, Arthur Bergman wrote:

Hi,

Can you see if this still applies? I think I might have fixed it
before
5.8 came out. I can't reproduce it on linux.

On a copy of 18378 I have lying around, I can get strange effects
on Linux by executing some code that reallocs stuff between the joins:
use threads;
my $a = 'x'x 1; \#just a hook for the debugger

my $t = threads\->new$sub \{\}$;
$t\->join;
for $1\.\.100$ \{
$a\[$\_\] = "\\0xff" x $\_;
\}

$t\->join;
[davem@percy 18378t]$ ./perl -Ilib /tmp/p
Name "main::a" used only once: possible typo at /tmp/p line 9.
panic: MUTEX_LOCK at /tmp/p line 12.

So I guess the bug is still active :-(

Gah, how annoying. Some seems very prematurely freed.

Can someone run a memory verifier on this and see what is going on
please?

Thanks

Arthur

p5pRT · 2003-04-16T17:44:05Z

From arthur@contiller.se

Hi,

Seems like substandard operating systems that actually free freed memory had a
problem :-)

This patch fixes this since destroy works.

Arthur

p5pRT · 2003-04-16T17:44:05Z

From arthur@contiller.se

19240.patch

--- ext/threads/t/join.t-19239	Wed Apr 16 19:37:42 2003
+++ ext/threads/t/join.t	Wed Apr 16 19:37:45 2003
@@ -10,7 +10,7 @@
 
 use ExtUtils::testlib;
 use strict;
-BEGIN { print "1..11\n" };
+BEGIN { print "1..12\n" };
 use threads;
 use threads::shared;
 
@@ -117,4 +117,17 @@
   }
 } else {
   skip("\$0 check: only on Linux");
+}
+
+{
+    my $t = threads->new(sub {});
+    $t->join;
+    my $x = threads->new(sub {});
+    $x->join;
+    eval {
+      $t->join;
+    };
+    my $ok = 0;
+    $ok++ if($@ =~/Thread already joined/);
+    ok($ok, "Double join works");
 }
--- ext/threads/threads.xs-19239	Wed Apr 16 19:37:51 2003
+++ ext/threads/threads.xs	Wed Apr 16 19:37:53 2003
@@ -601,7 +601,7 @@
 	/* We have finished with it */
 	thread->state |= PERL_ITHR_JOINED;
 	MUTEX_UNLOCK(&thread->mutex);
-    	sv_unmagic(SvRV(obj),PERL_MAGIC_shared_scalar);
+    	
 	return retparam;
     }
     return (AV*)NULL;

p5pRT · 2003-04-16T17:44:06Z

arthur@contiller.se - Status changed from 'open' to 'resolved'

p5pRT closed this as completed Apr 16, 2003

p5pRT added Severity Medium distro-Solaris type-core labels Oct 18, 2019

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

ithreads coredumps on join of joined thread #5850

ithreads coredumps on join of joined thread #5850

p5pRT commented Aug 21, 2002

p5pRT commented Aug 21, 2002

p5pRT commented Aug 21, 2002

p5pRT commented Jan 4, 2003

p5pRT commented Jan 4, 2003

p5pRT commented Jan 4, 2003

p5pRT commented Jan 4, 2003

p5pRT commented Apr 16, 2003

p5pRT commented Apr 16, 2003

p5pRT commented Apr 16, 2003

ithreads coredumps on join of joined thread #5850

ithreads coredumps on join of joined thread #5850

Comments

p5pRT commented Aug 21, 2002

p5pRT commented Aug 21, 2002

From davem@fdgroup.co.uk

Created by davem@gizmo.fdgroup.co.uk

p5pRT commented Aug 21, 2002

From @lizmat

p5pRT commented Jan 4, 2003

p5pRT commented Jan 4, 2003

From arthur@contiller.se

p5pRT commented Jan 4, 2003

From @iabyn

p5pRT commented Jan 4, 2003

From arthur@contiller.se

p5pRT commented Apr 16, 2003

From arthur@contiller.se

p5pRT commented Apr 16, 2003

From arthur@contiller.se

p5pRT commented Apr 16, 2003