public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Rakib Mullick <rakib.mullick@gmail.com>
To: linux-kernel@vger.kernel.org
Cc: mingo@kernel.org, Andrew Morton <akpm@linux-foundation.org>,
	Oleg Nesterov <oleg@redhat.com>
Subject: Do we really need curr_target in signal_struct ?
Date: Tue, 28 Jan 2014 13:57:20 +0600	[thread overview]
Message-ID: <1390895840.8373.2.camel@beeld> (raw)

I was wondering what might be the possible use of curr_target in signal_struct (atleast,
it's not doing what it's comment says). Also, I'm not seeing any real use of it except
in kernel/signal.c::complete_signal() where it's use as loop breaking condition in thread's
list traversing. As an alternative of using curr_target we can use get_nr_thread() count
to get # of threads in a group and can remove curr_target completely. This will also help
us to get rid from overhead of maintaining ->curr_target at fork()/exit() path. Below is
rough patch, I can prepare a good one if everyone agrees.

Or we really need it?

---
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 485234d..1fd65b7 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -554,7 +554,7 @@ struct signal_struct {
 	wait_queue_head_t	wait_chldexit;	/* for wait4() */
 
 	/* current thread group signal load-balancing target: */
-	struct task_struct	*curr_target;
+	//struct task_struct	*curr_target;
 
 	/* shared signal handling: */
 	struct sigpending	shared_pending;
diff --git a/kernel/exit.c b/kernel/exit.c
index 1e77fc6..4a2cf37 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -113,8 +113,8 @@ static void __exit_signal(struct task_struct *tsk)
 		if (sig->notify_count > 0 && !--sig->notify_count)
 			wake_up_process(sig->group_exit_task);
 
-		if (tsk == sig->curr_target)
-			sig->curr_target = next_thread(tsk);
+		//if (tsk == sig->curr_target)
+		//	sig->curr_target = next_thread(tsk);
 		/*
 		 * Accumulate here the counters for all threads but the
 		 * group leader as they die, so they can be added into
diff --git a/kernel/fork.c b/kernel/fork.c
index 2f11bbe..8de4928 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -1041,7 +1041,7 @@ static int copy_signal(unsigned long clone_flags, struct task_struct *tsk)
 	tsk->thread_node = (struct list_head)LIST_HEAD_INIT(sig->thread_head);
 
 	init_waitqueue_head(&sig->wait_chldexit);
-	sig->curr_target = tsk;
+	//sig->curr_target = tsk;
 	init_sigpending(&sig->shared_pending);
 	INIT_LIST_HEAD(&sig->posix_timers);
 
diff --git a/kernel/signal.c b/kernel/signal.c
index 940b30e..1a1280a 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -945,6 +945,7 @@ static void complete_signal(int sig, struct task_struct *p, int group)
 {
 	struct signal_struct *signal = p->signal;
 	struct task_struct *t;
+	int i;
 
 	/*
 	 * Now find a thread we can wake up to take the signal off the queue.
@@ -961,21 +962,16 @@ static void complete_signal(int sig, struct task_struct *p, int group)
 		 */
 		return;
 	else {
-		/*
-		 * Otherwise try to find a suitable thread.
-		 */
-		t = signal->curr_target;
-		while (!wants_signal(sig, t)) {
+		i = get_nr_threads(p);
+		t = p;
+		do {
+			--i;
 			t = next_thread(t);
-			if (t == signal->curr_target)
-				/*
-				 * No thread needs to be woken.
-				 * Any eligible threads will see
-				 * the signal in the queue soon.
-				 */
+			if (!i)
 				return;
-		}
-		signal->curr_target = t;
+		} while (!wants_signal(sig, t));
+	
+		//signal->curr_target = t;
 	}
 
 	/*



             reply	other threads:[~2014-01-28  8:06 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-01-28  7:57 Rakib Mullick [this message]
2014-01-28 16:43 ` Do we really need curr_target in signal_struct ? Oleg Nesterov
2014-01-29  4:09   ` Rakib Mullick
2014-01-29  4:45     ` Rakib Mullick
2014-01-29 14:55     ` Oleg Nesterov
2014-01-29 16:07       ` Rakib Mullick
2014-01-29 18:32         ` Oleg Nesterov
2014-01-30  7:02           ` Rakib Mullick
2014-01-31 18:53             ` Rakib Mullick
2014-02-01 16:51               ` Oleg Nesterov
2014-02-02 16:50                 ` Rakib Mullick
2014-02-03 16:39                   ` Oleg Nesterov
2014-02-04  4:32                     ` Rakib Mullick
2014-02-04 17:34                       ` Oleg Nesterov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1390895840.8373.2.camel@beeld \
    --to=rakib.mullick@gmail.com \
    --cc=akpm@linux-foundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=oleg@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox