Thanks for your reply.

In my test, when warning_threshold is set to 100 as default, the block time 
indeed reduce to 100ms. So I think poll_timer_wait_until() call is to set the 
max block time for poll_bock()'s waiting time.

> -----Original Message-----
> From: Ben Pfaff [mailto:[email protected]]
> Sent: Thursday, December 06, 2018 10:58 PM
> To: Lilijun (Jerry, Cloud Networking) <[email protected]>
> Cc: [email protected]
> Subject: Re: [ovs-dev] [PATCH] [PATCH] rcu: reduce RCU block time
> 
> On Thu, Dec 06, 2018 at 08:28:19PM +0800, Lilijun wrote:
> > When calling ovsrcu_synchronize(), it always block 1000ms because we
> > poll block until elapsed time become greater than
> > warning_threshold(1000).That's too long for some configuration commands.
> > So this patch reduces warning_threshold's default value to 100 and
> > print logs after it have elapsed more than 1000ms.
> >
> > Change-Id: Ib698afb47a600de00f52d8a5aa63039f0097514f
> > Signed-off-by: Lilijun <[email protected]>
> 
> I think that this patch reflects a misunderstanding.  It is the
> seq_wait() call that makes ovsrcu_synchronize() wake up when
> synchronization is complete.  The poll_timer_wait_until() call is there just 
> to
> produce the warning messages.
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to