[prev in list] [next in list] [prev in thread] [next in thread] 

List:       hpux-devtools
Subject:    RE: HPUX-DEVTOOLS: libclntsh.sl dumping core
From:       "Gagan Puri" <gaganpuri () infosys ! com>
Date:       2002-11-26 8:54:53
[Download RAW message or body]

Hi Mahesha,

Our code already conatins the MemRegisterTask call & this is the first call in
main program. Problem is somewhere else. Any pointres????

Regards,
Gagan

-----Original Message-----
From: mahesha [mailto:mahesha@india.hp.com]
Sent: Tuesday, November 26, 2002 2:11 PM
To: Gagan Puri; Dennis Handly; hpux-devtools@cxx.cup.hp.com
Subject: Re: HPUX-DEVTOOLS: libclntsh.sl dumping core


As I understand, some special initializations need to be done on HPUX
platform while using SmartHeap with multithreaded applications. You may need
to ensure that you have called MemRegisterTask before creating threads.
Otherwise its heap mutexes will not get initialized, and there may be heap
corruptions. You may want to refer to the smartheap documentation for
HPUX platform before proceding.

-mahesha

On Tuesday 26 November 2002 11:41 am, Gagan Puri wrote:
> Hi Dennis
>
> Thanks for the reply. It does give some insight into the problem.
> "abort" goes away as soon as I remove the smartheap calls.
>
> I have one more question : it seems that "free" is internally being called
> from libclntsh. now the backtrace shows that "free" is in libsmobj. this is
> my application's shared lib. and I have checked ( & checked _again_), code
> does not have any "free" call. does it imply that there is a thirdparty
> library corruption or could it be some pointer corruption in my code which
> occurred earlier but is causing problems now. (core occurs 6/10 times)
>
> -Gagan
>
> -----Original Message-----
> From: Dennis Handly [mailto:dhandly@cup.hp.com]
> Sent: Tuesday, November 26, 2002 11:16 AM
> To: hpux-devtools@cxx.cup.hp.com
> Subject: Re: HPUX-DEVTOOLS: libclntsh.sl dumping core
>
>
> #0  0xc382d00c in _shi_removeFromFreeList .../libsmobj.sl
> #3  0xc382cc04 in free
> /home/hpux/gagan/project/installDrop3/abc/lib/libsmobj.sl
> #4  0xc65f1958 in nlepedstr /3rdparty/oracle/OraHome1/lib32/libclntsh.sl
> #5  0xc0a673d0 in __specific_data_thread_exit /usr/lib/libpthread.1
> #6  0xc0a6a3f0 in pthread_exit /usr/lib/libpthread.1
>
> >It seems that the core is generated from libclntsh.sl. What could be the
> >problem?????
> >We are also using smartheap libraries in our code.
>
> Gagan
>
> You are aborting in your smartheap code.
> Either that doesn't work with threads, or you have corrupted your heap.
>
> The thread has exited and is spinning through the destructor functions for
> thread specific data allocated with pthread_key_create.
>
> I suppose there could be something wrong with nlepedstr calling free?
> I suggest you remove smartheap to see if the abort goes away.
>  _________________________________________________________________
>  To leave this mailing list, send mail to majordomo@cxx.cup.hp.com
>     with the message UNSUBSCRIBE hpux-devtools
>  _________________________________________________________________
>  _________________________________________________________________
>  To leave this mailing list, send mail to majordomo@cxx.cup.hp.com
>     with the message UNSUBSCRIBE hpux-devtools
>  _________________________________________________________________
 _________________________________________________________________
 To leave this mailing list, send mail to majordomo@cxx.cup.hp.com
    with the message UNSUBSCRIBE hpux-devtools
 _________________________________________________________________
 _________________________________________________________________
 To leave this mailing list, send mail to majordomo@cxx.cup.hp.com
    with the message UNSUBSCRIBE hpux-devtools
 _________________________________________________________________
[prev in list] [next in list] [prev in thread] [next in thread]