From mboxrd@z Thu Jan 1 00:00:00 1970 From: merez@codeaurora.org Subject: Re: [PATCH v8 1/2] mmc: add packed command feature of eMMC4.5 Date: Tue, 10 Jul 2012 03:45:15 -0700 (PDT) Message-ID: <9177cd479f2dec231b65102cb35a3e6a.squirrel@www.codeaurora.org> References: <006801cd55b0$314960f0$93dc22d0$%jun@samsung.com> <87y5msqhxy.fsf@octavius.laptop.org> <87mx38qdcs.fsf@octavius.laptop.org> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Return-path: Received: from wolverine02.qualcomm.com ([199.106.114.251]:2455 "EHLO wolverine02.qualcomm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753224Ab2GJKpQ (ORCPT ); Tue, 10 Jul 2012 06:45:16 -0400 In-Reply-To: <87mx38qdcs.fsf@octavius.laptop.org> Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: Chris Ball Cc: merez@codeaurora.org, Seungwon Jeon , linux-mmc@vger.kernel.org, 'Subhash Jadavani' , "'S, Venkatraman'" , 'Saugata Das' , 'Namjae Jeon' On Tue, July 10, 2012 12:33 am, Chris Ball wrote: > Hi, > > On Tue, Jul 10 2012, merez@codeaurora.org wrote: >>> Sorry, I don't have a copy of the eMMC 4.5 spec -- is PACKED_EVENT_EN >>> a one-time programmable fuse on the eMMC, like BKOPS_ENABLE was? >> >> No, it's not. > > Excellent, thanks. > > I've pushed both v8 patches to mmc-next. Before pushing it up to Linus, > I'd like to get a better idea of the variety of performances differences > seen using packed writes -- has it been tried on devices from multiple > eMMC vendors, and multiple host controller vendors? Can we try to build > up a quick table of results, with a standard test like iozone, to check > that we've got decent test coverage? Seungwon suggests using: > > iozone -az -i0 -I -s 10m -f /target/test -e > > I'm also worried about how much code this is adding, but I agree that > it's worth it if the performance benefits are as described. > > Thanks, > > - Chris. > -- > Chris Ball > One Laptop Per Child > -- Hi Chris, I suggest to also use lmdd as a benchmark test for write packing, since write packing is mostly beneficial in long sequential operations. It can be used as follows: lmdd if=internal of=/data/file1 bs=128k count=3000 Thanks, Maya -- Sent by consultant of Qualcomm Innovation Center, Inc. Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum