Re: nslcd 0.9.8 Core Dump (SIGABRT) on Service Stop - Thread Cleanup Issue

[Kokila Koks <kokilavaradhan [at] gmail.com>]

Hi, 

Here's the gdb analysis:
$ gdb-multiarch nslcd core.nslcd.1626.*

Program terminated with signal SIGABRT, Aborted.
#0  0x76dab064 in ?? ()

(gdb) info threads
  Id   Target Id         Frame
* 1    LWP 1627          0x76dab064 in ?? ()
  2    LWP 1626          0x76e10214 in ?? ()
  3    LWP 1629          0x76e10218 in ?? ()

(gdb) info registers
r7  = 0xaf (175)     # ARM syscall: kill()
pc  = 0x76dab064
lr  = 0x76e09374

(gdb) bt
#0  0x76dab064 in ?? ()
#1  0x76e09374 in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)


The binary is stripped (production build), so function names aren't available. However, the register state shows the crashed thread (LWP 1627) was executing kill() syscall (r7=175) to send SIGABRT to itself, consistent with abort() being called after a failed assertion during shutdown.
System logs confirm: "thread 0 is still running, shutting down anyway" immediately before the crash.

Thanks ,

kokila V