lkml.org 
[lkml]   [2018]   [Jul]   [3]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
Subject[PATCH] sched/deadline: Add update_rq_clock(later_rq) in push_dl_task()
From: Steven Rostedt (VMware) <rostedt@goodmis.org>

Running tests on the deadline scheduler I triggered a WARN_ON() in
assert_clock_updated(). Unfortunately, this caused printk() to try
to wake up the klogd which caused a nasty deadlock. All I got out of
the output was:

"------------[ cut here ]------------"

But a bit of tweaking of where the 'cut here' prints were, I was able
to find the source. The warning is triggered by:

push_dl_task {
add_running_bw() {
__add_running_bw() {
cpufreq_update_util() {
data->func(data, rq_clock(rq), flags);
rq_clock() {
assert_clock_updated()

The rq_clock() is called on the rq of the CPU that is being pushed to.
There's no guarantee that it has been updated at this time. Since we
have the rq lock of that rq, the CPU doing the push can update the
clock to make sure that it is up-to-date before proceeding to do the
push.

Link: http://lkml.kernel.org/r/20180703105449.2a211fb2@gandalf.local.home
Suggested-by: Juri Lelli <juri.lelli@gmail.com>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
---
diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index fbfc3f1d368a..90b6e9df79b2 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -2085,6 +2085,8 @@ static int push_dl_task(struct rq *rq)
goto retry;
}

+ update_rq_clock(later_rq);
+
deactivate_task(rq, next_task, 0);
sub_running_bw(&next_task->dl, &rq->dl);
sub_rq_bw(&next_task->dl, &rq->dl);
\
 
 \ /
  Last update: 2018-07-03 18:04    [W:2.245 / U:0.204 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site