mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Thomas Gleixner <tglx@linutronix.de>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: LKML <linux-kernel@vger.kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@kernel.org>,
	Lai Jiangshan <laijs@cn.fujitsu.com>,
	Jason Low <jason.low2@hp.com>, Brad Mouring <bmouring@ni.com>
Subject: Re: [patch V3 7/7] rtmutex: Avoid pointless requeueing in the deadlock detection chain walk
Date: Tue, 10 Jun 2014 19:41:39 +0200 (CEST)	[thread overview]
Message-ID: <alpine.DEB.2.10.1406101630350.31315@nanos> (raw)
In-Reply-To: <20140609212033.1cff0999@gandalf.local.home>

On Mon, 9 Jun 2014, Steven Rostedt wrote:
> On Mon, 09 Jun 2014 20:28:10 -0000
> Thomas Gleixner <tglx@linutronix.de> wrote:
> >  	/*
> > +	 * If we just follow the lock chain for deadlock detection, no
> > +	 * need to do all the requeue operations. We avoid a truckload
> 
> s/We/To/
> 
> 
> > +	 * of conditinals around the various places below and just do
> 
> s/ and/, /

Ok.
 
> > +	 * the minimum chain walk checks here.
> > +	 */
> > +	if (!requeue) {
> > +		/* Release the task */
> > +		raw_spin_unlock_irqrestore(&task->pi_lock, flags);
> > +		put_task_struct(task);
> > +
> > +		/* If there is no owner of the lock, end of chain. */
> > +		if (!rt_mutex_owner(lock)) {
> > +			raw_spin_unlock(&lock->wait_lock);
> > +			return 0;
> > +		}
> > +
> > +		/* Grab the next task, i.e. owner of @lock */
> > +		task = rt_mutex_owner(lock);
> > +		get_task_struct(task);
> > +		raw_spin_lock_irqsave(&task->pi_lock, flags);
> > +
> > +		/* Store whether owner is blocked itself and drop locks */
> > +		next_lock = task_blocked_on(task);
> > +		raw_spin_unlock_irqrestore(&task->pi_lock, flags);
> > +		raw_spin_unlock(&lock->wait_lock);
> > +
> > +		/* If owner is not blocked, end of chain. */
> > +		if (!next_lock)
> > +			goto out_put_task;
> 
> On the loop back around, have something like:
> 
> 	if (top_waiter) {
> 		if (!task_has_pi_waiters(task))
> 			goto out_unlock_pi;

	The task has at least one pi waiter.
 
> 		if (!requeue &&
> 		    top_waiter != task_top_pi_waiter(task)) {
> 			if (!detect_deadlock)
> 				goto out_unlock_pi;
> 			else
> 				requeue = false;
> 		}

	Errm? if requeue is off we are in deadlock detection chainwalk
	mode. So all we care about is whether task is blocked on
	next_lock or not.

What you actually missed is that we need to read out top_waiter for
the current lock. Fixed already.
    

Thanks,

	tglx

  parent reply	other threads:[~2014-06-10 17:41 UTC|newest]

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-06-09 20:28 [patch V3 0/7] rtmutex: Code clarification and optimization Thomas Gleixner
2014-06-09 20:28 ` [patch V3 1/7] rtmutex: Deobfuscate chain walk Thomas Gleixner
2014-06-09 20:59   ` Steven Rostedt
2014-06-10  3:52     ` Lai Jiangshan
2014-06-10  3:21   ` Jason Low
2014-06-10 13:57   ` Brad Mouring
2014-06-09 20:28 ` [patch V3 2/7] rtmutex: Clarify the boost/deboost part Thomas Gleixner
2014-06-10  0:37   ` Steven Rostedt
2014-06-10  3:22   ` Jason Low
2014-06-10 14:04   ` Brad Mouring
2014-06-09 20:28 ` [patch V3 3/7] rtmutex: Document pi chain walk Thomas Gleixner
2014-06-10  0:45   ` Steven Rostedt
2014-06-10  3:51     ` Lai Jiangshan
2014-06-10 14:21   ` Brad Mouring
2014-06-09 20:28 ` [patch V3 4/7] rtmutex: Siplify remove_waiter() Thomas Gleixner
2014-06-10  0:53   ` Steven Rostedt
2014-06-10  3:35     ` Lai Jiangshan
2014-06-10  3:44     ` Jason Low
2014-06-10 14:10   ` Brad Mouring
2014-06-09 20:28 ` [patch V3 5/7] rtmutex: Confine deadlock logic to futex Thomas Gleixner
2014-06-10  0:59   ` Steven Rostedt
2014-06-10  4:03     ` Lai Jiangshan
2014-06-10 17:39     ` Thomas Gleixner
2014-06-09 20:28 ` [patch V3 6/7] rtmutex: Cleanup deadlock detector debug logic Thomas Gleixner
2014-06-10  1:04   ` Steven Rostedt
2014-06-10 15:09   ` Brad Mouring
2014-06-09 20:28 ` [patch V3 7/7] rtmutex: Avoid pointless requeueing in the deadlock detection chain walk Thomas Gleixner
2014-06-10  1:20   ` Steven Rostedt
2014-06-10  3:48     ` Jason Low
2014-06-10 17:41     ` Thomas Gleixner [this message]
2014-06-10 17:47       ` Steven Rostedt
2014-06-10 20:45         ` Thomas Gleixner
2014-06-10 14:57   ` Brad Mouring
2014-06-10 15:19     ` Steven Rostedt
2014-06-10 17:43       ` Thomas Gleixner
2014-06-10 17:51         ` Steven Rostedt
2014-06-10 20:46           ` Thomas Gleixner
2014-06-10  0:27 ` [patch V3 0/7] rtmutex: Code clarification and optimization Steven Rostedt
2014-06-10  0:35   ` Steven Rostedt
2014-06-10  3:00     ` Lai Jiangshan

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=alpine.DEB.2.10.1406101630350.31315@nanos \
    --to=tglx@linutronix.de \
    --cc=bmouring@ni.com \
    --cc=jason.low2@hp.com \
    --cc=laijs@cn.fujitsu.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=peterz@infradead.org \
    --cc=rostedt@goodmis.org \
    /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

all inboxes | Powered by JetHome®