[PATCH v4 00/12] mmc: use nonblock mmc requests to minimize latency

S, Venkatraman svenkatr at ti.com
Thu Jun 16 09:39:25 EDT 2011

On Thu, May 26, 2011 at 3:27 AM, Per Forlin <per.forlin at linaro.org> 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 none 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 none blocking mmc request makes it
> possible to prepare the caches for next job parallel with an active
> mmc request.
> This is done by making the issue_rw_rq() none 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 v3:
>  * Based on 2.6.39-rc7
>  * Add error check for testlist in mmc_test.c
>  * Resolve in mmc-queue-thread that caused the mmc-thread to miss a wakeup.
>  * Move parallel request handling to core.c. This simplifies the interface
>   from 4 public functions to 1. This also gives access for SDIO to use the
>   same functionallity, even though the function is not tuned for the SDIO
>   execution flow yet.
> Per Forlin (12):
>  mmc: add none blocking mmc request function
>  omap_hsmmc: use original sg_len for dma_unmap_sg
>  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 none 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      |  452 +++++++++++++++++++++++++----------------
>  drivers/mmc/card/mmc_test.c   |  361 ++++++++++++++++++++++++++++++++-
>  drivers/mmc/card/queue.c      |  184 +++++++++++------
>  drivers/mmc/card/queue.h      |   32 +++-
>  drivers/mmc/core/core.c       |  165 ++++++++++++++-
>  drivers/mmc/core/debugfs.c    |    5 +
>  drivers/mmc/host/mmci.c       |  146 ++++++++++++--
>  drivers/mmc/host/mmci.h       |    8 +
>  drivers/mmc/host/omap_hsmmc.c |   90 ++++++++-
>  include/linux/mmc/core.h      |    6 +-
>  include/linux/mmc/host.h      |   19 ++
>  lib/Kconfig.debug             |   11 +
>  12 files changed, 1187 insertions(+), 292 deletions(-)

Nitpick.. The mmc_test.c changes should be at the end of the series,
after the async feature is available.


More information about the linux-arm-kernel mailing list