102492 - TestXPTCInvoke segfaults under solaris when build with gcc 3.0.1

Status: RESOLVED WORKSFORME

(SeaMonkey :: Build Config, defect)

Description

[Nils Lohner]

/sup/opt/mozilla-0.9.4/bin > ./run-mozilla.sh ./TestXPTCInvoke
MOZILLA_FIVE_HOME=.
  LD_LIBRARY_PATH=.:./plugins:/usr/openwin/lib:/usr/lib:/lib:/usr/ucblib:/sup/li
                                                                               
              b:/usr/openwin/lib:/usr/lib:/lib:/usr/ucblib:/sup/lib
     LIBRARY_PATH=.:./components
       SHLIB_PATH=.
          LIBPATH=.
       ADDON_PATH=.
      MOZ_PROGRAM=./TestXPTCInvoke
      MOZ_TOOLKIT=
        moz_debug=0
     moz_debugger=
Segmentation Fault - core dumped


Oh no!  ./TestXPTCInvoke just dumped a core file.

Do you want to debug this ? You need a lot of memory for this, so watch out ?
[y/n] n

Comment 1

[Nils Lohner]

/sup/build/mozilla/debugbuild-solaris/dist/bin > gdb TestXPTCInvoke core
GNU gdb 5.0
Copyright 2000 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "sparc-sun-solaris2.7"...
Core was generated by `./TestXPTCInvoke'.
Program terminated with signal 9, Killed.
Reading symbols from
/sup/build/mozilla/debugbuild-solaris/dist/bin/./libxpcom.so...done.
Loaded symbols for /sup/build/mozilla/debugbuild-solaris/dist/bin/./libxpcom.so
Reading symbols from
/sup/build/mozilla/debugbuild-solaris/dist/bin/./libplds4.so...done.
Loaded symbols for /sup/build/mozilla/debugbuild-solaris/dist/bin/./libplds4.so
Reading symbols from
/sup/build/mozilla/debugbuild-solaris/dist/bin/./libplc4.so...done.
Loaded symbols for /sup/build/mozilla/debugbuild-solaris/dist/bin/./libplc4.so
Reading symbols from
/sup/build/mozilla/debugbuild-solaris/dist/bin/./libnspr4.so...done.
Loaded symbols for /sup/build/mozilla/debugbuild-solaris/dist/bin/./libnspr4.so
Reading symbols from /usr/lib/librt.so.1...done.
Loaded symbols for /usr/lib/librt.so.1
Reading symbols from /usr/lib/libsocket.so.1...done.
Loaded symbols for /usr/lib/libsocket.so.1
Reading symbols from /usr/lib/libdl.so.1...done.
Loaded symbols for /usr/lib/libdl.so.1
Reading symbols from /sup/lib/libstdc++.so.3...done.
Loaded symbols for /sup/lib/libstdc++.so.3
Reading symbols from /usr/lib/libm.so.1...done.
Loaded symbols for /usr/lib/libm.so.1
Reading symbols from /sup/lib/libgcc_s.so.1...done.
Loaded symbols for /sup/lib/libgcc_s.so.1
Reading symbols from /usr/lib/libpthread.so.1...done.
Loaded symbols for /usr/lib/libpthread.so.1
Reading symbols from /usr/lib/libc.so.1...done.
Loaded symbols for /usr/lib/libc.so.1
Reading symbols from /usr/lib/libthread.so.1...done.
Loaded symbols for /usr/lib/libthread.so.1
Reading symbols from /usr/lib/libnsl.so.1...done.
Loaded symbols for /usr/lib/libnsl.so.1
Reading symbols from /usr/lib/libaio.so.1...done.
Loaded symbols for /usr/lib/libaio.so.1
Reading symbols from /usr/lib/libmp.so.2...done.
Loaded symbols for /usr/lib/libmp.so.2
Reading symbols from /usr/platform/SUNW,Ultra-5_10/lib/libc_psr.so.1...done.
Loaded symbols for /usr/platform/SUNW,Ultra-5_10/lib/libc_psr.so.1
#0  0xfef8b9e8 in _lock_try_adaptive () from /usr/lib/libthread.so.1
(gdb) where
#0  0xfef8b9e8 in _lock_try_adaptive () from /usr/lib/libthread.so.1
#1  0xfef7c704 in pthread_mutex_lock () from /usr/lib/libthread.so.1
#2  0xff10f534 in _MD_ATOMIC_INCREMENT (val=0xff152330) at
../../../../../nsprpub/pr/src/misc/pratom.c:166
#3  0xff10f924 in PR_AtomicIncrement (val=0xff152330) at
../../../../../nsprpub/pr/src/misc/pratom.c:301
#4  0xff10c8a0 in PR_NewThreadPrivateIndex (newIndex=0xff154e6c, dtor=0xff10c500
<_PR_RELEASE_LOCK_STACK>)
    at ../../../../../nsprpub/pr/src/threads/prtpd.c:137
#5  0xff10c398 in _PR_InitRWLocks () at
../../../../../nsprpub/pr/src/threads/prrwlock.c:396
#6  0xff117a3c in _PR_InitStuff () at
../../../../../nsprpub/pr/src/misc/prinit.c:239
#7  0xff117a68 in _PR_ImplicitInitialization () at
../../../../../nsprpub/pr/src/misc/prinit.c:252
#8  0xff129488 in PR_NewLock () at
../../../../../nsprpub/pr/src/pthreads/ptsynch.c:149
#9  0xff2a4f6c in _Z12InitTraceLogv () at ../../../xpcom/base/nsTraceRefcnt.cpp:829
#10 0xff2a6030 in _ZN13nsTraceRefcnt7LogCtorEPvPKcj (aPtr=0x25c5c,
aType=0xff319f30 "nsHashtable", aInstanceSize=44)
    at ../../../xpcom/base/nsTraceRefcnt.cpp:1825
#11 0xff21b730 in _ZN11nsHashtableC2Eji (this=0x25c5c, aInitSize=16, threadSafe=0)
    at ../../../xpcom/ds/nsHashtable.cpp:232
#12 0xff2f2420 in _ZN19nsSupportsHashtableC1Eji (this=0x25c5c, aSize=16,
threadSafe=0)
    at ../../dist/include/nsHashtable.h:166
#13 0xff29cec4 in _ZN16nsLoggingServiceC1Ev (this=0x25c50) at
../../../xpcom/base/nsLogging.cpp:60
#14 0xff29def8 in _Z20EnsureLoggingServicev () at
../../../xpcom/base/nsLogging.cpp:244
#15 0xff29e2e0 in _Z9NS_GetLogPKcj (name=0xff327ab8 "LogInfo", controlFlags=2)
    at ../../../xpcom/base/nsLogging.cpp:309
#16 0xff2a018c in _Z41__static_initialization_and_destruction_0ii
(__initialize_p=1, __priority=65535)
    at ../../dist/include/nsBufferHandle.h:44
#17 0xff2a01c0 in _GLOBAL__I__ZN16nsLoggingServiceC2Ev () at
../../dist/include/nsBufferHandle.h:363
#18 0xff2ee764 in __do_global_ctors_aux () at
../../../string/obsolete/nsString2.cpp:1928
#19 0xff20c534 in _init () from
/sup/build/mozilla/debugbuild-solaris/dist/bin/./libxpcom.so
#20 0xff3b9f3c in ?? ()
#21 0xff3b9df4 in ?? ()
#22 0xff3c15c0 in ?? ()
#23 0xff3b2d44 in ?? ()
(gdb) quit
/sup/build/mozilla/debugbuild-solaris/dist/bin > cat ../../mozbuild
/sup/build/mozilla/mybuild > ../configure --enable-tests --enable-debug
--disable-optimize --disable-xprint --without-gnu-ld --without-gnu-nm
--without-gnu-as --with-as=//usr/ccs/bin/as --with-ld=/usr/ccs/bin/ld
--with-nm=/usr/ccs/bin/nm --prefix=/sup


/sup/build/mozilla/debugbuild-solaris/dist/bin > gcc --version
3.0.1
/sup/build/mozilla/debugbuild-solaris/dist/bin > uname -a
SunOS topaze 5.7 Generic_106541-02 sun4u sparc
/sup/build/mozilla/debugbuild-solaris/dist/bin > 

Comment 2

[Nils Lohner]

THis bug is very repeatable.  I have a core file available if someone is
interested.  Even with optimizsation and GNU linkers etc. the crash is exactly
the same.  I still have the tree built, so I can try things if anyone has any
experimental patches etc. to fix the problem.

Comment 3

[Bradley Baetz (:bbaetz)]

It needs some ASM fixes for the new ABI - see bug 71627

->xpconnect, and confirming based on 71627.

Comment 4

[Doug Turner (:dougt)]

John/David, is this yours? 

Comment 5

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

I think TestXPTCInvoke would get farther than it did here if bug 71627 were the
only problem.

Comment 6

[David Bradley]

From the stack trace I don't see anything related to XPT call other than the
test name. It appears to have crashed in _lock_try_adaptive while trying to
create a new lock, during creating the settings hash table while logging
something. So unless something bad happened, stack not cleaned up, or something
during a method call before this happened, I doubt it's an XPT call issue.

Comment 7

[Doug Turner (:dougt)]

not enough time.  0.9.5 -> 0.9.6

Comment 8

[Doug Turner (:dougt)]

Mike, I am reassign this to you since you own bug 71627.  

Comment 9

[Mike Shaver (:shaver emeritus)]

This looks like an NSPR issue, maybe -- wtc, can you take a look?

Comment 10

[Wan-Teh Chang]

It is crashing in the Solaris pthread library.  Could
you do
   % ldd ./TestXPTCInvoke
and paste the output here?  Also, please paste the
command line with which you linked TestXPTCInvoke.

You can try building and running NSPR test programs
to see if NSPR works.
   % cvs co mozilla/nsprpub
   % cd mozilla/nsprpub
   % ./configure
   % gmake
   % cd pr/tests
   % gmake
   # run a few tests
   % ./cvar -d
   % ./cvar2
   % ./perf

Comment 11

[Nils Lohner]

/sup/build/mozilla/debugbuild-solaris/dist/bin > ldd ./TestXPTCInvoke

        libxpcom.so =>   (file not found)
        libplds4.so =>   (file not found)
        libplc4.so =>    (file not found)
        libnspr4.so =>   (file not found)
        librt.so.1 =>    /usr/lib/librt.so.1
        libsocket.so.1 =>        /usr/lib/libsocket.so.1
        libdl.so.1 =>    /usr/lib/libdl.so.1
        libstdc++.so.3 =>        /sup/lib/libstdc++.so.3
        libm.so.1 =>     /usr/lib/libm.so.1
        libgcc_s.so.1 =>         /sup/lib/libgcc_s.so.1
        libpthread.so.1 =>       /usr/lib/libpthread.so.1
        libc.so.1 =>     /usr/lib/libc.so.1
        libaio.so.1 =>   /usr/lib/libaio.so.1
        libnsl.so.1 =>   /usr/lib/libnsl.so.1
        libmp.so.2 =>    /usr/lib/libmp.so.2
        libthread.so.1 =>        /usr/lib/libthread.so.1
/sup/build/mozilla/debugbuild-solaris/dist/bin > 

Comment 12

[Nils Lohner]

Here's the linking info that you wanted.  I can't provide more info as I'm
moving and need to clean out my accounts, and don't have time to do the CVS
steps you requested.  Sorry...

/sup/build/mozilla/debugbuild-solaris/xpcom > rm reflect/xptcall/tests/TestXPTCI
                             nvoke
/sup/build/mozilla/debugbuild-solaris/xpcom > make
make[1]: Entering directory `/sup/build/mozilla/debugbuild-solaris/xpcom/typelib'
....... 
make[1]: Entering directory
`/sup/build/mozilla/debugbuild-solaris/xpcom/reflect/xptcall/tests'
c++  -I/usr/openwin/include -fno-rtti -fno-exceptions -pedantic -Wno-long-long
-pipe -fshort-wchar -pthreads  -DDEBUG -DDEBUG_lohner -DTRACING -g -o
TestXPTCInvoke TestXPTCInvoke.o  -L../../../../dist/bin -L../../../../dist/lib
-L../../../../dist/bin -lxpcom -liberty
-L/sup/build/mozilla/debugbuild-solaris/dist/lib -lplds4 -lplc4 -lnspr4  
-lposix4 -lsocket -ldl -lm    
true TestXPTCInvoke
../../../../config/nsinstall -R -m 755 TestXPTCInvoke ../../../../dist/bin
make[1]: Leaving directory
`/sup/build/mozilla/debugbuild-solaris/xpcom/reflect/xptcall/tests'

Comment 13

[Wan-Teh Chang]

Based on the information provided by Nils Lohner,
I believe this is a build config problem.

The test program TestXPTCInvoke is linked with
libc.so.1 before libthread.so.1.  This linking
order is problematic and is caused by not using
the -pthread flag of c++.

You can achieve the equivalent effect by using
the -D_REENTRANT compiler flag and the -lpthread
-lthread linker flags, but I recommend using the
-pthread flag of gcc and c++ for both compiling
and linking.  (You would use the -mt flag with
the Solaris WorkShop or Forte 6 compilers.)

cls, is this your area or should this bug stay
in XPConnect?

Comment 14

[Wan-Teh Chang]

Sorry, I noticed that we are using the -pthreads flag.
My mistake.

I don't know why the -pthreads flag of c++ doesn't do
the right thing.  In any case, you can fix the problem
by linking with -lpthread -lthread before any system
libraries, that is, replacing
    -lposix4 -lsocket -ldl -lm
by
    -lpthread -lthread -lposix4 -lsocket -ldl -lm

Comment 15

[Wan-Teh Chang]

Comparing the library linking orders of 'c++ -pthreads' and
'CC -mt', I see the difference in the order between libc.so.1
and libthread.so.1.

Note that although 'c++ -pthreads' links with libpthread.so.1
before libc.so.1, I've found that to be insufficient, and in
fact, it is the order between libc.so.1 and libthread.so.1
that matters.

Here is the output of my experiments.  (The contents of the
test program foo.cpp is not important, except that it shows
that both 'c++ -pthreads' and 'CC -mt' define the _REENTRANT
macro.)

bonn:/u/wtc/tmp 53% c++ -v
Reading specs from /builds/local/lib/gcc-lib/sparc-sun-solaris2.6/2.95.2/specs
gcc version 2.95.2 19991024 (release)
bonn:/u/wtc/tmp 54% c++ -pthreads foo.cpp
bonn:/u/wtc/tmp 55% ./a.out
Hello world!
_REENTRANT
bonn:/u/wtc/tmp 56% ldd ./a.out
        libm.so.1 =>     /usr/lib/libm.so.1
        libpthread.so.1 =>       /usr/lib/libpthread.so.1
        libc.so.1 =>     /usr/lib/libc.so.1
        libdl.so.1 =>    /usr/lib/libdl.so.1
        libthread.so.1 =>        /usr/lib/libthread.so.1
        /usr/platform/SUNW,Ultra-5_10/lib/libc_psr.so.1
bonn:/u/wtc/tmp 57% CC -V
CC: Sun WorkShop 6 update 1 C++ 5.2 Patch 109508-03 2001/04/07
bonn:/u/wtc/tmp 58% CC -mt foo.cpp
bonn:/u/wtc/tmp 59% ./a.out
Hello world!
_REENTRANT
bonn:/u/wtc/tmp 60% ldd ./a.out
        libCrun.so.1 =>  /usr/lib/libCrun.so.1
        libm.so.1 =>     /opt/SUNWspro/lib/libm.so.1
        libw.so.1 =>     /usr/lib/libw.so.1
        libthread.so.1 =>        /usr/lib/libthread.so.1
        libc.so.1 =>     /usr/lib/libc.so.1
        libdl.so.1 =>    /usr/lib/libdl.so.1
        /usr/platform/SUNW,Ultra-5_10/lib/libc_psr.so.1

Comment 16

[Mike Shaver (:shaver emeritus)]

I'm going to throw this over to cls for build-flag fixage, I think.

But it's _so_ not a blocker.

Comment 17

[David Bradley]

*** Bug 133684 has been marked as a duplicate of this bug. ***

Comment 18

[David Bradley]

*** Bug 149461 has been marked as a duplicate of this bug. ***

Comment 19

[cls]

Mass reassign to default build config owner

Comment 20

[Nicolas Pioch]

 - Mozilla compiles fine with all variants of gcc-3.x
on my Solaris 8 sparc box as long as you don't use GNU binutils (as, ld...)

as documented at ftp://depot.mcom.com/pub/pioch/mozilla-1.4a/ README

Read also ftp://depot.mcom.com/pub/pioch/mozilla-1.3/ README

(my build is an UltraSparc, albeit 32-bit, sparc executable)

Comment 21

[(default assignee for unassigned bugs. does not receive bugmail)]

Mass reassign of Build/Config bugs to Leaf.

Comment 22

[Phil Schwartau]

setting default QA -

Comment 23

[Chase Phillips]

Mass reassign of open bugs for [e-mail] to [e-mail].

Comment 24

[J. Paul Reed [:preed]]

Mass re-assign of bugs that aren't on the build team radar, so bugs assigned to [e-mail] reflects reality.

If there is a bug you really think we need to be looking at, please *email* [e-mail] with a bug number and explanation.