Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 03 May 2012 15:33:38 -0700
From:      Sean Bruno <seanbru@yahoo-inc.com>
To:        John Baldwin <jhb@freebsd.org>
Cc:        "freebsd-net@freebsd.org" <freebsd-net@freebsd.org>, Jack Vogel <jfvogel@gmail.com>
Subject:   Re: igb(4) Pondering a bind to cpu patch
Message-ID:  <1336084418.3077.21.camel@powernoodle-l7.corp.yahoo.com>
In-Reply-To: <1335382225.2722.6.camel@powernoodle-l7.corp.yahoo.com>
References:  <1335312667.11564.13.camel@powernoodle-l7.corp.yahoo.com> <201204250932.21378.jhb@freebsd.org> <1335382225.2722.6.camel@powernoodle-l7.corp.yahoo.com>

next in thread | previous in thread | raw e-mail | index | archive | help


On Wed, 2012-04-25 at 12:30 -0700, Sean Bruno wrote:
> On Wed, 2012-04-25 at 06:32 -0700, John Baldwin wrote:
> > CPU IDs are not guaranteed to be dense.  However, you can use
> > CPU_FIRST() and 
> > CPU_NEXT() with your static global instead.
> > 
> Ah, does CPU_NEXT() reset to 0 when it reaches the end of its list of
> CPUs?
> 

Ah, I see.  So, yeah, here's a v2 of the patch that does "the right"
thing with non-sparse cpus, mulitple queues, and mulitple physical
interfaces.

http://people.freebsd.org/~sbruno/if_igb.c.txt

> 
> > OTOH, if igb were to just leave the interrupts alone instead of
> > binding them 
> > by hand, they would get round-robin assigned among available cores
> > already.  I 
> > think in this case the best approach might be to add a tunable to
> > disable 
> > igb's manual binding and instead let the default system round-robin
> > be 
> > preserved. 
> 
> also, yes.  Why *are* we binding to CPUs in the first place?  Are we
> afraid that the scheduler won't do the right thing and we're trying to
> work around some unknown performance issue ?
> 
> Sean
> 

Still haven't seen a good reason to bind the queues by default in the
first place.  

Sean




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1336084418.3077.21.camel>