Re: SASL causes segmentation fault (ITS#3172)

[Kurt@OpenLDAP.org]

At 02:44 PM 6/7/2004, quanah@stanford.edu wrote:


>--On Monday, June 07, 2004 8:49 PM +0000 Kurt Zeilenga 
><openldap-its@OpenLDAP.org> wrote:
>
>> What were the other threads doing?  Are you sure it was this
>> thread which caused the segfault?
>
>initial segfault shows:
>
>Program received signal SIGSEGV, Segmentation fault.
>[Switching to Thread 5 (LWP 1)]
>0xff2ffb80 in sb_debug_write (sbiod=0x342390, buf=0x10c2d58, len=4657)
>    at sockbuf.c:846
>846             ret = LBER_SBIOD_WRITE_NEXT( sbiod, buf, len );
>
>
>info threads
>  9 Thread 4 (LWP 3)  0xfee1d394 in _poll () from /usr/lib/libc.so.1
>  8 Thread 3          0xfed4d9b8 in _reap_wait () from 
>/usr/lib/libthread.so.1
>  7 Thread 2 (LWP 2)  0xfee1eb58 in _signotifywait () from 
>/usr/lib/libc.so.1
>  6 LWP    3          0xfee1d394 in _poll () from /usr/lib/libc.so.1
>* 5 Thread 5 (LWP 1)  0xff2ffb80 in sb_debug_write (sbiod=0x342390,
>    buf=0x10c2d58, len=4657) at sockbuf.c:846
>  4 LWP    4          0xfed591fc in private___lwp_cond_wait ()
>   from /usr/lib/libthread.so.1
>  3 LWP    2          0xfee1eb58 in _signotifywait () from 
>/usr/lib/libc.so.1
>  2 Thread 1          0xfed4da10 in _reap_wait_cancel ()
>   from /usr/lib/libthread.so.1
>  1 LWP    1          0xff2ffb80 in sb_debug_write (sbiod=0x342390,
>    buf=0x10c2d58, len=4657) at sockbuf.c:846


This looks odd to me.  Seems like thread 1 and 5 share
LWP 1 and the same thread stack.  But then again I am
not familiar with the Solaris thread/LWP model.
Anyways, I wonder if one crash is due to a thread stack
overrun.  You might try doubling LDAP_PVT_THREAD_STACK_SIZE
(include/ldap_pvt_thread.h).

Kurt