[PATCH] Re: PTRACE_EVENT_CLONE patch causing deadlock?

[Ulrich Weigand]

Daniel Jacobowitz wrote:

> Probably it's never been tested.  Resuming the parent seems right to
> me - see lin_lwp_handle_extended.  Does just adding a PTRACE_CONT in
> there work?

Indeed, the following patch fixes the deadlock.
Tested on s390-ibm-linux and s390x-ibm-linux.
OK?

ChangeLog:

	* lin-lwp.c (child_wait): Continue inferior after processing
	PTRACE_EVENT_CLONE event.

Index: gdb/lin-lwp.c
===================================================================
RCS file: /cvs/src/src/gdb/lin-lwp.c,v
retrieving revision 1.56
diff -c -p -r1.56 lin-lwp.c
*** gdb/lin-lwp.c	26 Jul 2004 14:53:02 -0000	1.56
--- gdb/lin-lwp.c	29 Jul 2004 13:16:34 -0000
*************** child_wait (ptid_t ptid, struct target_w
*** 1213,1218 ****
--- 1213,1219 ----
  	    {
  	      ptrace (PTRACE_DETACH, ourstatus->value.related_pid, 0, 0);
  	      ourstatus->kind = TARGET_WAITKIND_IGNORE;
+ 	      ptrace (PTRACE_CONT, pid, 0, 0);
  	      pid = -1;
  	      save_errno = EINTR;
  	    }

-- 
  Dr. Ulrich Weigand
  weigand@informatik.uni-erlangen.de