From mboxrd@z Thu Jan 1 00:00:00 1970 From: Benjamin Herrenschmidt Subject: Re: [PATCH 2/4] arch: Add lightweight memory barriers fast_rmb() and fast_wmb() Date: Tue, 18 Nov 2014 11:41:45 +1100 Message-ID: <1416271305.18381.19.camel@kernel.crashing.org> References: <20141117171005.22333.96544.stgit@ahduyck-server> <20141117171812.22333.90395.stgit@ahduyck-server> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: Alexander Duyck , "linux-arch@vger.kernel.org" , Network Development , Linux Kernel Mailing List , Mathieu Desnoyers , Peter Zijlstra , Heiko Carstens , Ingo Molnar , Michael Neuling , Russell King - ARM Linux , donald.c.skidmore@intel.com, matthew.vick@intel.com, Geert Uytterhoeven , Jeff Kirsher , Francois Romieu , Paul McKenney , nic_swsd@realtek.com, Will Deacon , Michael Ellerman , Tony Luck , Oleg Nesterov , Martin Schwidefsky Return-path: In-Reply-To: Sender: linux-arch-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Mon, 2014-11-17 at 12:52 -0800, Linus Torvalds wrote: > Maybe "dma_*mb()" would be acceptable, and ends up having the same > naming convention as "smb_*mb()", and explains what it's about. Yes, that was what I was about to suggest as well. > And yes, in the same spirit, it would probably be good to try to > eventually get rid of the plain "*mb()" functions, and perhaps call > them "mmio_*mb()" to clarify that they are about ordering memory wrt > mmio. It will always be somewhat unclear to users who don't read the doc anyway :) IE. the dma_* ones do only DMA vs DMA (or vs other processors) but the mmio_* one do anything vs anything. Not a huge deal tho. I still like dma_* for Alexander's new stuff but I wouldn't bother with changing the existing ones. Ben.