Re: [PATCH RFC 8/8] rcu: Fix cpustart tracepoint gp_seq number

From: Paul E. McKenney
Date: Mon May 14 2018 - 16:32:00 EST


On Sun, May 13, 2018 at 08:15:41PM -0700, Joel Fernandes (Google) wrote:
> cpustart shows a stale gp_seq. This is because rdp->gp_seq is updated
> only at the end of the __note_gp_changes function. For this reason, use
> rnp->gp_seq instead. I believe we can't update rdp->gp_seq too early so
> lets just use the gp_seq from rnp instead.
>
> Signed-off-by: Joel Fernandes (Google) <joel@xxxxxxxxxxxxxxxxx>

I took this one with the usual commit-log update. Please take a look!

Thanx, Paul

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

commit 3e14c6a5225d201776edd9a96d9c9b4435855446
Author: Joel Fernandes (Google) <joel@xxxxxxxxxxxxxxxxx>
Date: Sun May 13 20:15:41 2018 -0700

rcu: Fix cpustart tracepoint gp_seq number

The "cpustart" trace event shows a stale gp_seq. This is because it uses
rdp->gp_seq, which is updated only at the end of the __note_gp_changes()
function. This commit therefore instead uses rnp->gp_seq.

An alternative fix would be to update rdp->gp_seq earlier, but this would
break RCU's detection of the beginning of a new-to-this-CPU grace period.

Signed-off-by: Joel Fernandes (Google) <joel@xxxxxxxxxxxxxxxxx>
Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxxxxxxxxxx>

diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
index de6447dd73de..23b855f5c5cb 100644
--- a/kernel/rcu/tree.c
+++ b/kernel/rcu/tree.c
@@ -1796,7 +1796,7 @@ static bool __note_gp_changes(struct rcu_state *rsp, struct rcu_node *rnp,
* set up to detect a quiescent state, otherwise don't
* go looking for one.
*/
- trace_rcu_grace_period(rsp->name, rdp->gp_seq, TPS("cpustart"));
+ trace_rcu_grace_period(rsp->name, rnp->gp_seq, TPS("cpustart"));
need_gp = !!(rnp->qsmask & rdp->grpmask);
rdp->cpu_no_qs.b.norm = need_gp;
rdp->rcu_qs_ctr_snap = __this_cpu_read(rcu_dynticks.rcu_qs_ctr);