From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bart Van Assche Subject: Re: [PATCH 5/5] IB/srp: Optimize completion queue polling Date: Fri, 04 Jul 2014 11:59:42 +0200 Message-ID: <53B67B0E.5070004@acm.org> References: <53B55E55.5040907@acm.org> <53B55F1F.6000704@acm.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Or Gerlitz Cc: Roland Dreier , Sagi Grimberg , Sebastian Parschauer , David Dillow , linux-rdma List-Id: linux-rdma@vger.kernel.org On 07/03/14 18:46, Or Gerlitz wrote: > On Thu, Jul 3, 2014 at 4:48 PM, Bart Van Assche wrote: > [...] >> >> Limit the number of poll cycles per completion notification to preserve fairness > > [...] > > Any reason not to make use of the block layer NAPI like API for that > (blk-iopoll http://lwn.net/Articles/346187/) ? Hello Or, As you maybe remember about four years ago I have been experimenting with adding blk-iopoll support in the SRP initiator. What I learned at that time is that the blk-iopoll framework defers work to softirq context. This means that a context switch from interrupt to softirq context has to occur before SRP completion processing can start. Recent measurements on current hardware have shown that such a context switch takes about 0.5 microseconds. Since I prefer to keep the latency of SRP I/O as low as possible I haven't looked further into using blk-iopoll for the SRP initiator driver. Bart. -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html