RE: [PATCH 3.18 132/134] rcu: Do RCU GP kthread self-wakeup from softirq and interrupt

[He, Bo]

Hi, Greg:
	Can you hold on the 3.18-stable branch, it seems there are some issue, please see the comments from Paul:

Comments from Paul:
I subjected all of the others to light rcutorture testing, which they passed.  This v3.18 patch hung, however.  Trying it again with stock
v3.18 got the same hang, so I believe we can exonerate the patch and give it a good firm "maybe" on 3.18.

Worth paying special attention to further test results from 3.18.x, though!

						

-----Original Message-----
From: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> 
Sent: Friday, March 22, 2019 7:16 PM
To: linux-kernel@xxxxxxxxxxxxxxx
Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>; stable@xxxxxxxxxxxxxxx; He, Bo <bo.he@xxxxxxxxx>; Zhang, Jun <jun.zhang@xxxxxxxxx>; Paul E. McKenney <paulmck@xxxxxxxxxxxxx>; Xiao, Jin <jin.xiao@xxxxxxxxx>; Bai, Jie A <jie.a.bai@xxxxxxxxx>
Subject: [PATCH 3.18 132/134] rcu: Do RCU GP kthread self-wakeup from softirq and interrupt

3.18-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Zhang, Jun <jun.zhang@xxxxxxxxx>

commit 1d1f898df6586c5ea9aeaf349f13089c6fa37903 upstream.

The rcu_gp_kthread_wake() function is invoked when it might be necessary to wake the RCU grace-period kthread.  Because self-wakeups are normally a useless waste of CPU cycles, if rcu_gp_kthread_wake() is invoked from this kthread, it naturally refuses to do the wakeup.

Unfortunately, natural though it might be, this heuristic fails when
rcu_gp_kthread_wake() is invoked from an interrupt or softirq handler that interrupted the grace-period kthread just after the final check of the wait-event condition but just before the schedule() call.  In this case, a wakeup is required, even though the call to rcu_gp_kthread_wake() is within the RCU grace-period kthread's context.  Failing to provide this wakeup can result in grace periods failing to start, which in turn results in out-of-memory conditions.

This race window is quite narrow, but it actually did happen during real testing.  It would of course need to be fixed even if it was strictly theoretical in nature.

This patch does not Cc stable because it does not apply cleanly to earlier kernel versions.

Fixes: 48a7639ce80c ("rcu: Make callers awaken grace-period kthread")
Reported-by: "He, Bo" <bo.he@xxxxxxxxx>
Co-developed-by: "Zhang, Jun" <jun.zhang@xxxxxxxxx>
Co-developed-by: "He, Bo" <bo.he@xxxxxxxxx>
Co-developed-by: "xiao, jin" <jin.xiao@xxxxxxxxx>
Co-developed-by: Bai, Jie A <jie.a.bai@xxxxxxxxx>
Signed-off: "Zhang, Jun" <jun.zhang@xxxxxxxxx>
Signed-off: "He, Bo" <bo.he@xxxxxxxxx>
Signed-off: "xiao, jin" <jin.xiao@xxxxxxxxx>
Signed-off: Bai, Jie A <jie.a.bai@xxxxxxxxx>
Signed-off-by: "Zhang, Jun" <jun.zhang@xxxxxxxxx> [ paulmck: Switch from !in_softirq() to "!in_interrupt() &&
  !in_serving_softirq() to avoid redundant wakeups and to also handle the
  interrupt-handler scenario as well as the softirq-handler scenario that
  actually occurred in testing. ]
Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxxxxx>
Link: https://lkml.kernel.org/r/CD6925E8781EFD4D8E11882D20FC406D52A11F61@xxxxxxxxxxxxxxxxxxxxxxxxxxxx
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
 kernel/rcu/tree.c |   20 ++++++++++++++------
 1 file changed, 14 insertions(+), 6 deletions(-)

--- a/kernel/rcu/tree.c
+++ b/kernel/rcu/tree.c
@@ -1399,15 +1399,23 @@ static int rcu_future_gp_cleanup(struct  }
 
 /*
- * Awaken the grace-period kthread for the specified flavor of RCU.
- * Don't do a self-awaken, and don't bother awakening when there is
- * nothing for the grace-period kthread to do (as in several CPUs
- * raced to awaken, and we lost), and finally don't try to awaken
- * a kthread that has not yet been created.
+ * Awaken the grace-period kthread.  Don't do a self-awaken (unless in
+ * an interrupt or softirq handler), and don't bother awakening when 
+ there
+ * is nothing for the grace-period kthread to do (as in several CPUs 
+ raced
+ * to awaken, and we lost), and finally don't try to awaken a kthread 
+ that
+ * has not yet been created.  If all those checks are passed, track 
+ some
+ * debug information and awaken.
+ *
+ * So why do the self-wakeup when in an interrupt or softirq handler
+ * in the grace-period kthread's context?  Because the kthread might 
+ have
+ * been interrupted just as it was going to sleep, and just after the 
+ final
+ * pre-sleep check of the awaken condition.  In this case, a wakeup 
+ really
+ * is required, and is therefore supplied.
  */
 static void rcu_gp_kthread_wake(struct rcu_state *rsp)  {
-	if (current == rsp->gp_kthread ||
+	if ((current == rsp->gp_kthread &&
+	     !in_interrupt() && !in_serving_softirq()) ||
 	    !ACCESS_ONCE(rsp->gp_flags) ||
 	    !rsp->gp_kthread)
 		return;