[Pkg-openldap-devel] Strange segfault with 2.4.44+dfsg-1 from git

[Ryan Tandy]

On Mon, Oct 24, 2016 at 10:25:04AM +0200, Ralf Mattes wrote:
>Hello list,

Hi,

> Program received signal SIGSEGV, Segmentation fault.
>[Switching to Thread 0x7f60dd7e1700 (LWP 32510)]
>0x00007f60dd7e0aa0 in ?? ()
>(gdb) bt
>#0  0x00007f60dd7e0aa0 in ?? ()
>#1  0x00007f61650bca41 in slap_writewait_play (op=0x7f60c8002550) at ../../../../servers/slapd/result.c:294
>#2  send_ldap_ber (op=0x7f60c8002550, ber=0x7f60dd64f250) at ../../../../servers/slapd/result.c:367
>#3  0x00007f61650bf651 in slap_send_search_entry (op=0x7f60c8002550, rs=0x7f60dd7e0aa0) at ../../../../servers/slapd/result.c:1430
>#4  0x00007f616003590b in mdb_search (op=0x7f60c8002550, rs=0x7f60dd7e0aa0) at ../../../../../servers/slapd/back-mdb/search.c:1086
>#5  0x00007f615f5f8cd6 in relay_back_op (op=0x7f60c8002550, rs=0x7f60dd7e0aa0, which=<optimized out>)
>    at ../../../../../servers/slapd/back-relay/op.c:210
>#6  0x00007f616511aeea in overlay_op_walk (op=op at entry=0x7f60c8002550, rs=0x7f60dd7e0aa0, which=op_search, oi=0x7f6165d2c020,
>    on=<optimized out>) at ../../../../servers/slapd/backover.c:677
>#7  0x00007f616511b044 in over_op_func (op=0x7f60c8002550, rs=<optimized out>, which=<optimized out>)
>    at ../../../../servers/slapd/backover.c:730
>#8  0x00007f61650af071 in fe_op_search (op=0x7f60c8002550, rs=0x7f60dd7e0aa0) at ../../../../servers/slapd/search.c:402
>#9  0x00007f61650ae9ee in do_search (op=0x7f60c8002550, rs=0x7f60dd7e0aa0) at ../../../../servers/slapd/search.c:247
>#10 0x00007f61650ac57c in connection_operation (ctx=0x7f60dd7e0c10, arg_v=0x7f60c8002550)
>    at ../../../../servers/slapd/connection.c:1158
>#11 0x00007f61650ac867 in connection_read_thread (ctx=0x7f60c8002550, argv=0x7f60dd7df710)
>    at ../../../../servers/slapd/connection.c:1294
>#12 0x00007f6164c0df22 in ldap_int_thread_pool_wrapper (xpool=0x7f6165c8ffa0) at ../../../../libraries/libldap_r/tpool.c:696
>#13 0x00007f61631f20a4 in start_thread (arg=0x7f60dd7e1700) at pthread_create.c:309
>#14 0x00007f6162f2762d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:111
>
>It seems that queries that return large amounts of data (i.e. including jpegPhoto attributes) will trigger this segfault.

This looks like ITS#8512.

http://www.openldap.org/its/?findid=8512

There doesn't seem to be a fix at this time.

I have not been able to reproduce the problem. Any information leading 
to being able to reproduce it would be very helpful, especially since 
the team have begun preparing for a 2.4.45 release. This sounds like an 
important bug to fix.

The ITS claims that this might have been introduced in 2.4.40. I have 
not been able to reproduce this in that version. Would it be possible 
for you to check whether this occurs in the version in jessie?

Thanks for your help!