From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751587Ab1F3Mgs (ORCPT ); Thu, 30 Jun 2011 08:36:48 -0400 Received: from na3sys009aog109.obsmtp.com ([74.125.149.201]:44509 "EHLO na3sys009aog109.obsmtp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751031Ab1F3Mgl convert rfc822-to-8bit (ORCPT ); Thu, 30 Jun 2011 08:36:41 -0400 MIME-Version: 1.0 In-Reply-To: <1309248717-14606-1-git-send-email-per.forlin@linaro.org> References: <1309248717-14606-1-git-send-email-per.forlin@linaro.org> Date: Thu, 30 Jun 2011 18:06:39 +0530 Message-ID: Subject: Re: [PATCH v8 00/12] use nonblock mmc requests to minimize latency From: "Poddar, Sourav" To: Per Forlin Cc: linaro-dev@lists.linaro.org, Nicolas Pitre , linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, linux-mmc@vger.kernel.org, Nickolay Nickolaev , Venkatraman S , Linus Walleij , Chris Ball Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8BIT Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Jun 28, 2011 at 1:41 PM, Per Forlin wrote: > How significant is the cache maintenance over head? > It depends, the eMMC are much faster now > compared to a few years ago and cache maintenance cost more due to > multiple cache levels and speculative cache pre-fetch. In relation the > cost for handling the caches have increased and is now a bottle neck > dealing with fast eMMC together with DMA. > > The intention for introducing non-blocking mmc requests is to minimize the > time between a mmc request ends and another mmc request starts. In the > current implementation the MMC controller is idle when dma_map_sg and > dma_unmap_sg is processing. Introducing non-blocking mmc request makes it > possible to prepare the caches for next job in parallel to an active > mmc request. > > This is done by making the issue_rw_rq() non-blocking. > The increase in throughput is proportional to the time it takes to > prepare (major part of preparations is dma_map_sg and dma_unmap_sg) > a request and how fast the memory is. The faster the MMC/SD is > the more significant the prepare request time becomes. Measurements on U5500 > and Panda on eMMC and SD shows significant performance gain for large > reads when running DMA mode. In the PIO case the performance is unchanged. > > There are two optional hooks pre_req() and post_req() that the host driver > may implement in order to move work to before and after the actual mmc_request > function is called. In the DMA case pre_req() may do dma_map_sg() and prepare > the dma descriptor and post_req runs the dma_unmap_sg. > > Details on measurements from IOZone and mmc_test: > https://wiki.linaro.org/WorkingGroups/Kernel/Specs/StoragePerfMMC-async-req > > Changes since v7: >  * rebase on mmc-next, on top of Russell's updated error handling. >  * Clarify description of mmc_start_req() >  * Resolve compile without CONFIG_DMA_ENIGNE issue for mmci >  * Add mmc test to measure how performance is affected by sg length >  * Add missing wait_for_busy in mmc_test non-blocking test. This call got lost >   in v4 of this patchset when refactoring mmc_start_req. >  * Add sub-prefix (core block queue) to relevant patches. > > Per Forlin (12): >  mmc: core: add non-blocking mmc request function >  omap_hsmmc: add support for pre_req and post_req >  mmci: implement pre_req() and post_req() >  mmc: mmc_test: add debugfs file to list all tests >  mmc: mmc_test: add test for non-blocking transfers >  mmc: mmc_test: test to measure how sg_len affect performance >  mmc: block: add member in mmc queue struct to hold request data >  mmc: block: add a block request prepare function >  mmc: block: move error code in issue_rw_rq to a separate function. >  mmc: queue: add a second mmc queue request member >  mmc: core: add random fault injection >  mmc: block: add handling for two parallel block requests in >    issue_rw_rq > >  drivers/mmc/card/block.c      |  505 ++++++++++++++++++++++++----------------- >  drivers/mmc/card/mmc_test.c   |  491 ++++++++++++++++++++++++++++++++++++++-- >  drivers/mmc/card/queue.c      |  184 ++++++++++------ >  drivers/mmc/card/queue.h      |   33 ++- >  drivers/mmc/core/core.c       |  167 +++++++++++++- >  drivers/mmc/core/debugfs.c    |    5 + >  drivers/mmc/host/mmci.c       |  147 +++++++++++- >  drivers/mmc/host/mmci.h       |    8 + >  drivers/mmc/host/omap_hsmmc.c |   87 +++++++- >  include/linux/mmc/core.h      |    6 +- >  include/linux/mmc/host.h      |   24 ++ >  lib/Kconfig.debug             |   11 + >  12 files changed, 1345 insertions(+), 323 deletions(-) Boot tested on Omap4430 Blaze board. Tested-by: Sourav Poddar