From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eliezer Tamir Subject: Re: [PATCH net-next] net: rename low latency sockets functions to busy poll Date: Wed, 10 Jul 2013 08:21:21 +0300 Message-ID: <51DCEF51.3080503@linux.intel.com> References: <20130709162514.062f1361@lwn.net> <20130709.160611.636749585237773459.davem@davemloft.net> <51DCD50C.6070606@linux.intel.com> <20130709.214156.1738814737075691376.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: corbet@lwn.net, linux-kernel@vger.kernel.org, netdev@vger.kernel.org, torvalds@linux-foundation.org, akpm@linux-foundation.org, dwmw2@infradead.org, eliezer@tamir.org.il To: David Miller Return-path: In-Reply-To: <20130709.214156.1738814737075691376.davem@davemloft.net> Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On 10/07/2013 07:41, David Miller wrote: > From: Eliezer Tamir > Date: Wed, 10 Jul 2013 06:29:16 +0300 > >> If the following names changes are acceptable I will try to send out >> a patch today. >> 2. ndo_ll_poll -> ndo_busy_poll >> >> - not technically accurate since the ndo callback does not itself busy >> poll, it's just used to implement it. > > I think this name change is accurate, it expresses the two elements of > what it does. It's busy waiting, in that it's doing a synchronous > scan of the device's RX queue, and it's polling just like NAPI polling > does. OK > Well... what would be great would be to come up with some single > interface that drivers can implement rather than having to have > both napi->poll and netdevice_ops->ndo_ll_poll(). But that's a task > for a later date. > > Therefore, ndo_busy_poll is probably best for now. I will think about this, maybe we could even unify ndo_poll_controller. It seems like said unified method would have to have an extra parameter that would indicate from which context it was called: 1. from napi poll (bh) 2. from poll controller (with interrupts disabled) 3. from busy poll (user context) -of course not for today. > Nope, looks complete. And also do the manpage update. Where do I find the repository for the manpages?