From: Steven Rostedt <rostedt@goodmis.org>
To: Peter Zijlstra <peterz@infradead.org>
Cc: Peng Hao <peng.hao2@zte.com.cn>,
mingo@redhat.com, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3] sched/rt : return accurate release rq lock info
Date: Mon, 15 Oct 2018 11:42:20 -0400 [thread overview]
Message-ID: <20181015114220.70c3598d@gandalf.local.home> (raw)
In-Reply-To: <20181015092032.GO9867@hirez.programming.kicks-ass.net>
On Mon, 15 Oct 2018 11:20:32 +0200
Peter Zijlstra <peterz@infradead.org> wrote:
> > index 2e2955a..be0fc43 100644
> > --- a/kernel/sched/rt.c
> > +++ b/kernel/sched/rt.c
> > @@ -1754,7 +1754,7 @@ static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
> > !task_on_rq_queued(task))) {
> >
> > double_unlock_balance(rq, lowest_rq);
> > - lowest_rq = NULL;
> > + lowest_rq = RETRY_TASK;
> > break;
> > }
> > }
>
> I'm confused.. should not:
>
> /* try again */
> double_unlock_balance(rq, lowest_rq);
> lowest_rq = NULL;
>
> also return RETRY_TASK? That also is in the double_lock_balance() path
> and will this have had rq->lock() released.
I thought the same thing at first, but this is in the loop path, where
it does everything again. But now looking closer, I think there's a bug
in the original code.
We only do the check if the immediate double_lock_balance() released
the current task rq lock, but we don't take into account if it was
released earlier, which means it could have migrated and we never
noticed!
I believe the code should look like this:
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index 2e2955a8cf8f..2c9128ce61e2 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -1718,6 +1718,7 @@ static int find_lowest_rq(struct task_struct *task)
static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
{
struct rq *lowest_rq = NULL;
+ bool released = false;
int tries;
int cpu;
@@ -1740,7 +1741,7 @@ static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
}
/* if the prio of this runqueue changed, try again */
- if (double_lock_balance(rq, lowest_rq)) {
+ if (double_lock_balance(rq, lowest_rq) || released) {
/*
* We had to unlock the run queue. In
* the mean time, task could have
@@ -1754,7 +1755,7 @@ static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
!task_on_rq_queued(task))) {
double_unlock_balance(rq, lowest_rq);
- lowest_rq = NULL;
+ lowest_rq = RETRY_TASK;
break;
}
}
@@ -1764,10 +1765,15 @@ static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
break;
/* try again */
- double_unlock_balance(rq, lowest_rq);
+ if (double_unlock_balance(rq, lowest_rq))
+ released = true;
+
lowest_rq = NULL;
}
+ if (!lowest_rq && released)
+ lowest_rq = RETRY_TASK;
+
return lowest_rq;
}
-- Steve
next prev parent reply other threads:[~2018-10-15 15:42 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-10-05 22:22 [PATCH v3] sched/rt : return accurate release rq lock info Peng Hao
2018-10-05 14:26 ` Steven Rostedt
2018-10-05 15:46 ` Andrea Parri
[not found] ` <201810060003580936940@zte.com.cn>
2018-10-05 16:09 ` Steven Rostedt
2018-10-15 9:20 ` Peter Zijlstra
2018-10-15 15:42 ` Steven Rostedt [this message]
[not found] ` <201810160009439217654@zte.com.cn>
2018-10-15 17:20 ` Steven Rostedt
2018-10-16 12:35 ` Peter Zijlstra
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=20181015114220.70c3598d@gandalf.local.home \
--to=rostedt@goodmis.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peng.hao2@zte.com.cn \
--cc=peterz@infradead.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;
as well as URLs for NNTP newsgroup(s).