Re: [PATCH] rfs: Receive Flow Steering

From: Eric Dumazet <eric.dumazet@gmail.com>
To: Tom Herbert <therbert@google.com>
Cc: davem@davemloft.net, netdev@vger.kernel.org
Subject: Re: [PATCH] rfs: Receive Flow Steering
Date: Fri, 02 Apr 2010 14:37:51 +0200	[thread overview]
Message-ID: <1270211871.11099.2.camel@edumazet-laptop> (raw)
In-Reply-To: <1270197304.1936.179.camel@edumazet-laptop>

> > +static inline void rps_record_sock_flow(struct rps_sock_flow_table *table,
> > +					u32 hash)
> > +{
> 
> Hmm, so rps_record_sock_flow() is always called from non preemptable
> contextes ?
> 
> > +	rps_set_sock_flow(table, hash, smp_processor_id());
> > +}
> > +

I had to change this to :

static inline void rps_record_sock_flow(struct rps_sock_flow_table *table, u32 hash)
{
	/* we only give a hint, preemption can change our cpu under us */
	rps_set_sock_flow(table, hash, raw_smp_processor_id());
}