From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757481Ab1FUXja (ORCPT ); Tue, 21 Jun 2011 19:39:30 -0400 Received: from mail-bw0-f46.google.com ([209.85.214.46]:65291 "EHLO mail-bw0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756665Ab1FUXj2 (ORCPT ); Tue, 21 Jun 2011 19:39:28 -0400 From: Per Forlin To: linaro-dev@lists.linaro.org, Nicolas Pitre , linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, linux-mmc@vger.kernel.org, Venkatraman S Cc: Chris Ball , Per Forlin Subject: [PATCH v7 00/11] use nonblock mmc requests to minimize latency Date: Wed, 22 Jun 2011 01:38:30 +0200 Message-Id: <1308699521-20556-1-git-send-email-per.forlin@linaro.org> X-Mailer: git-send-email 1.7.4.1 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org 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 v6: * minor update of doc for mmc_start_req and code clean up. * Indentifed a bug running tests on ext4 with discard enable. The test procedure is documented here: https://wiki.linaro.org/WorkingGroups/Kernel/Specs/StoragePerfMMC-async-req#Liability_test * Resolved bug by preventing mmc async request run in parallel to discard (mmc_erase). Per Forlin (11): mmc: 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: add member in mmc queue struct to hold request data mmc: add a block request prepare function mmc: move error code in mmc_block_issue_rw_rq to a separate function. mmc: add a second mmc queue request member mmc: test: add random fault injection in core.c mmc: add handling for two parallel block requests in issue_rw_rq drivers/mmc/card/block.c | 537 ++++++++++++++++++++++++----------------- drivers/mmc/card/mmc_test.c | 361 +++++++++++++++++++++++++++- drivers/mmc/card/queue.c | 184 +++++++++----- drivers/mmc/card/queue.h | 33 ++- drivers/mmc/core/core.c | 164 ++++++++++++- drivers/mmc/core/debugfs.c | 5 + drivers/mmc/host/mmci.c | 146 ++++++++++- 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, 1237 insertions(+), 329 deletions(-) -- 1.7.4.1 From mboxrd@z Thu Jan 1 00:00:00 1970 From: Per Forlin Subject: [PATCH v7 00/11] use nonblock mmc requests to minimize latency Date: Wed, 22 Jun 2011 01:38:30 +0200 Message-ID: <1308699521-20556-1-git-send-email-per.forlin@linaro.org> Return-path: Received: from mail-bw0-f46.google.com ([209.85.214.46]:65291 "EHLO mail-bw0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756665Ab1FUXj2 (ORCPT ); Tue, 21 Jun 2011 19:39:28 -0400 Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: linaro-dev@lists.linaro.org, Nicolas Pitre , linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, linux-mmc@vger.kernel.org, Venkatraman S Cc: Chris Ball , Per Forlin 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 v6: * minor update of doc for mmc_start_req and code clean up. * Indentifed a bug running tests on ext4 with discard enable. The test procedure is documented here: https://wiki.linaro.org/WorkingGroups/Kernel/Specs/StoragePerfMMC-async-req#Liability_test * Resolved bug by preventing mmc async request run in parallel to discard (mmc_erase). Per Forlin (11): mmc: 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: add member in mmc queue struct to hold request data mmc: add a block request prepare function mmc: move error code in mmc_block_issue_rw_rq to a separate function. mmc: add a second mmc queue request member mmc: test: add random fault injection in core.c mmc: add handling for two parallel block requests in issue_rw_rq drivers/mmc/card/block.c | 537 ++++++++++++++++++++++++----------------- drivers/mmc/card/mmc_test.c | 361 +++++++++++++++++++++++++++- drivers/mmc/card/queue.c | 184 +++++++++----- drivers/mmc/card/queue.h | 33 ++- drivers/mmc/core/core.c | 164 ++++++++++++- drivers/mmc/core/debugfs.c | 5 + drivers/mmc/host/mmci.c | 146 ++++++++++- 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, 1237 insertions(+), 329 deletions(-) -- 1.7.4.1 From mboxrd@z Thu Jan 1 00:00:00 1970 From: per.forlin@linaro.org (Per Forlin) Date: Wed, 22 Jun 2011 01:38:30 +0200 Subject: [PATCH v7 00/11] use nonblock mmc requests to minimize latency Message-ID: <1308699521-20556-1-git-send-email-per.forlin@linaro.org> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org 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 v6: * minor update of doc for mmc_start_req and code clean up. * Indentifed a bug running tests on ext4 with discard enable. The test procedure is documented here: https://wiki.linaro.org/WorkingGroups/Kernel/Specs/StoragePerfMMC-async-req#Liability_test * Resolved bug by preventing mmc async request run in parallel to discard (mmc_erase). Per Forlin (11): mmc: 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: add member in mmc queue struct to hold request data mmc: add a block request prepare function mmc: move error code in mmc_block_issue_rw_rq to a separate function. mmc: add a second mmc queue request member mmc: test: add random fault injection in core.c mmc: add handling for two parallel block requests in issue_rw_rq drivers/mmc/card/block.c | 537 ++++++++++++++++++++++++----------------- drivers/mmc/card/mmc_test.c | 361 +++++++++++++++++++++++++++- drivers/mmc/card/queue.c | 184 +++++++++----- drivers/mmc/card/queue.h | 33 ++- drivers/mmc/core/core.c | 164 ++++++++++++- drivers/mmc/core/debugfs.c | 5 + drivers/mmc/host/mmci.c | 146 ++++++++++- 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, 1237 insertions(+), 329 deletions(-) -- 1.7.4.1