kern/150138: [patch] signal sent to stopped, traced process not immediately handled on continue

David Xu davidxu at freebsd.org
Tue Aug 31 05:30:06 UTC 2010


The following reply was made to PR kern/150138; it has been noted by GNATS.

From: David Xu <davidxu at freebsd.org>
To: Dan McNulty <dkmcnulty at gmail.com>
Cc: FreeBSD-gnats-submit at freebsd.org
Subject: Re: kern/150138: [patch] signal sent to stopped, traced process not
 immediately handled on continue
Date: Tue, 31 Aug 2010 13:27:01 +0000

 Dan McNulty wrote:
 > On Tue, Aug 31, 2010 at 6:09 AM, David Xu <davidxu at freebsd.org> wrote:
 >> User & wrote:
 >>
 >>>> Fix:
 >>> The attached patch ( to CURRENT ) takes a stab at fixing the problem. The
 >>> patch
 >>> adds code so that even if the process is being traced, a signal sent to a
 >>> interruptibly sleeping thread wakes up the thread so it will handle the
 >>> signal
 >>> when the process leaves the stopped state. The added code was copied from
 >>> the
 >>> code that handled stopped, non-traced processes later in the same
 >>> function.
 >> No patch ?
 >>
 > 
 > Sorry, in trying to figure out send-pr, didn't actually attach the
 > patch and other files.
 > 
 > They should be attached now.
 > 
 
 I also propose a new patch, it seems if the process is being traced,
 we should hand off job control to debugger, all job control caused
 by STOP/CONT signals should be skipped, but normal signal should
 remove thread from sleep queue.
 
 
 Index: kern_sig.c
 ===================================================================
 --- kern_sig.c	(版本 211859)
 +++ kern_sig.c	(工作副本)
 @@ -2139,20 +2139,9 @@
   	 * We try do the per-process part here.
   	 */
   	if (P_SHOULDSTOP(p)) {
 -		/*
 -		 * The process is in stopped mode. All the threads should be
 -		 * either winding down or already on the suspended queue.
 -		 */
 -		if (p->p_flag & P_TRACED) {
 -			/*
 -			 * The traced process is already stopped,
 -			 * so no further action is necessary.
 -			 * No signal can restart us.
 -			 */
 -			goto out;
 -		}
 -
   		if (sig == SIGKILL) {
 +			if (p->p_flag & P_TRACED)
 +				goto out;
   			/*
   			 * SIGKILL sets process running.
   			 * It will die elsewhere.
 @@ -2163,6 +2152,8 @@
   		}
 
   		if (prop & SA_CONT) {
 +			if (p->p_flag & P_TRACED)
 +				goto out;
   			/*
   			 * If SIGCONT is default (or ignored), we continue the
   			 * process but don't leave the signal in sigqueue as
 @@ -2207,6 +2198,8 @@
   		}
 
   		if (prop & SA_STOP) {
 +			if (p->p_flag & P_TRACED)
 +				goto out;
   			/*
   			 * Already stopped, don't need to stop again
   			 * (If we did the shell could get confused).


More information about the freebsd-bugs mailing list