Re: RE24 call for testing

[Aaron Richton <richton@nbcs.rutgers.edu>]

On Tue, 29 Sep 2009, Quanah Gibson-Mount wrote:

> Please test RE24 for 2.4.19 preparation. Thanks!

Looks like mt_mutex might have been hit out of order? Under test058:

current thread: t@5
  [1] lwp_mutex_lock(0x1005d3578, 0x7fffffff7e712d70, 0x0, 0x1, 0x0, 0x7fffffff75ffe6d9), at 0x7fffffff7e717e6c
  [2] mutex_lock_kernel(0x1005d3578, 0x0, 0x0, 0x10533c, 0x7fffffff7e713b38, 0x0), at 0x7fffffff7e712e10
  [3] mutex_lock_internal(0x0, 0x10, 0x7fffffff7e601000, 0x1002542b0, 0x100cb07a2, 0x100cb0851), at 0x7fffffff7e713e58 
=>[4] ldap_pvt_thread_mutex_lock(mutex = 0x1005d3578), line 296 in "thr_posix.c"
  [5] syncprov_op_mod(op = 0x7fffffff75fff400, rs = 0x7fffffff75ffeec8), line 1965 in "syncprov.c"
  [6] overlay_op_walk(op = 0x7fffffff75fff400, rs = 0x7fffffff75ffeec8, which = op_modify, oi = 0x100719820, on = 0x10071b490), line 659 in "backover.c"
  [7] over_op_func(op = 0x7fffffff75fff400, rs = 0x7fffffff75ffeec8, which = op_modify), line 721 in "backover.c"
  [8] over_op_modify(op = 0x7fffffff75fff400, rs = 0x7fffffff75ffeec8), line 760 in "backover.c"
  [9] syncrepl_updateCookie(si = 0x1007188d0, op = 0x7fffffff75fff400, pdn = 0x1007172e0, syncCookie = 0x7fffffff75fff178), line 3059 in "syncrepl.c"
  [10] do_syncrep2(op = 0x7fffffff75fff400, si = 0x1007188d0), line 1177 in "syncrepl.c"
  [11] do_syncrepl(ctx = 0x7fffffff75fffc20, arg = 0x1005548f0), line 1358 in "syncrepl.c"
  [12] connection_read_thread(ctx = 0x7fffffff75fffc20, argv = 0x1e), line 1261 in "connection.c"
  [13] ldap_int_thread_pool_wrapper(xpool = 0x1005365e0), line 685 in "tpool.c"

(dbx) thread -blockedby t@5
Thread t@5 is blocked by:
0x00000001005d3578 (0x1005d3578): usync_? mutex(locked)
Lock is unowned

and the slapd is deadlocked.