From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bruce Richardson Subject: Re: [PATCH 00/10] eal cleanup and new options Date: Tue, 25 Nov 2014 15:06:24 +0000 Message-ID: <20141125150624.GC6800@bricha3-MOBL3> References: <1416692622-28886-1-git-send-email-thomas.monjalon@6wind.com> <4091968.6vrJ5rTTlz@xps13> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: dev-VfR2kkLFssw@public.gmane.org To: Thomas Monjalon Return-path: Content-Disposition: inline In-Reply-To: <4091968.6vrJ5rTTlz@xps13> List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces-VfR2kkLFssw@public.gmane.org Sender: "dev" On Tue, Nov 25, 2014 at 03:55:19PM +0100, Thomas Monjalon wrote: > > There are some pending patches which requires to factorize some EAL parts > > in order to be correctly implemented. > > This patchset do the required clean-up and rework these patches to improve > > lcore handling: > > > > Didier Pallard (2): > > eal: add core list input format > > config: support 128 cores > > > > Patrick Lu (1): > > eal: get relative core index > > > > Simon Kuenzer (1): > > eal: add option --master-lcore > > > > Thomas Monjalon (6): > > eal: move internal headers in source directory > > eal: factorize common headers > > eal: fix header guards > > eal: factorize internal config reset > > eal: factorize options sanity check > > eal: factorize configuration adjustment > > Applied with last comments integrated. > > The conclusions to the vote about the -c/-l options are: > - people don't vote (don't care or don't read) > - a few votes give the majority to creating the -l option. > > Neil, I'm really not sure what is the best solution. Maybe that applying > this patch will make more voices raising. We'll see, all can be changed. > > -- My additional 2c here: While I don't see much wrong with having core lists passed using a long option, more and more cores are being added to the latest CPUs, so it's likely that DPDK users are going to have systems with >64 cores and will want to move from the coremask specifier to the core list option. In that case, having it as a short option makes a lot of sense. /Bruce