| Message ID | 20190320063653.22638-1-drake@endlessm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | mmc: alcor: fix DMA reads | expand |
Am 20.03.19 um 07:36 schrieb Daniel Drake: > Setting max_blk_count to 1 here was causing the mmc block layer > to always use the MMC_READ_SINGLE_BLOCK command here, which the > driver does not DMA-accelerate. > > Drop the max_blk_ settings here. The mmc host defaults suffice, > along with the max_segs and max_seg_size settings, which I have > now documented in more detail. > > Now each MMC command reads 4 512-byte blocks, using DMA instead of > PIO. On my SD card, this increases read performance (measured with dd) > from 167kb/sec to 4.6mb/sec. > > Link: http://lkml.kernel.org/r/CAD8Lp47L5T3jnAjBiPs1cQ+yFA3L6LJtgFvMETnBrY63-Zdi2g@mail.gmail.com > Signed-off-by: Daniel Drake <drake@endlessm.com> Reviewed-by: Oleksij Rempel <linux@rempel-privat.de> > --- > drivers/mmc/host/alcor.c | 25 +++++++++++++++++++------ > 1 file changed, 19 insertions(+), 6 deletions(-) > > diff --git a/drivers/mmc/host/alcor.c b/drivers/mmc/host/alcor.c > index c712b7deb3a9d..82a97866e0cf4 100644 > --- a/drivers/mmc/host/alcor.c > +++ b/drivers/mmc/host/alcor.c > @@ -1044,14 +1044,27 @@ static void alcor_init_mmc(struct alcor_sdmmc_host *host) > mmc->caps2 = MMC_CAP2_NO_SDIO; > mmc->ops = &alcor_sdc_ops; > > - /* Hardware cannot do scatter lists */ > + /* The hardware does DMA data transfer of 4096 bytes to/from a single > + * buffer address. Scatterlists are not supported, but upon DMA > + * completion (signalled via IRQ), the original vendor driver does > + * then immediately set up another DMA transfer of the next 4096 > + * bytes. > + * > + * This means that we need to handle the I/O in 4096 byte chunks. > + * Lacking a way to limit the sglist entries to 4096 bytes, we instead > + * impose that only one segment is provided, with maximum size 4096, > + * which also happens to be the minimum size. This means that the > + * single-entry sglist handled by this driver can be handed directly > + * to the hardware, nice and simple. > + * > + * Unfortunately though, that means we only do 4096 bytes I/O per > + * MMC command. A future improvement would be to make the driver > + * accept sg lists and entries of any size, and simply iterate > + * through them 4096 bytes at a time. > + */ > mmc->max_segs = AU6601_MAX_DMA_SEGMENTS; > mmc->max_seg_size = AU6601_MAX_DMA_BLOCK_SIZE; > - > - mmc->max_blk_size = mmc->max_seg_size; > - mmc->max_blk_count = mmc->max_segs; > - > - mmc->max_req_size = mmc->max_seg_size * mmc->max_segs; > + mmc->max_req_size = mmc->max_seg_size; > } > > static int alcor_pci_sdmmc_drv_probe(struct platform_device *pdev) >
On Wed, 20 Mar 2019 at 07:36, Daniel Drake <drake@endlessm.com> wrote: > > Setting max_blk_count to 1 here was causing the mmc block layer > to always use the MMC_READ_SINGLE_BLOCK command here, which the > driver does not DMA-accelerate. > > Drop the max_blk_ settings here. The mmc host defaults suffice, > along with the max_segs and max_seg_size settings, which I have > now documented in more detail. > > Now each MMC command reads 4 512-byte blocks, using DMA instead of > PIO. On my SD card, this increases read performance (measured with dd) > from 167kb/sec to 4.6mb/sec. > > Link: http://lkml.kernel.org/r/CAD8Lp47L5T3jnAjBiPs1cQ+yFA3L6LJtgFvMETnBrY63-Zdi2g@mail.gmail.com > Signed-off-by: Daniel Drake <drake@endlessm.com> Applies for fixes, thanks! I also added: Fixes: c5413ad815a6 ("mmc: add new Alcor Micro Cardreader SD/MMC driver") Cc: stable@vger.kernel.org Kind regards Uffe > --- > drivers/mmc/host/alcor.c | 25 +++++++++++++++++++------ > 1 file changed, 19 insertions(+), 6 deletions(-) > > diff --git a/drivers/mmc/host/alcor.c b/drivers/mmc/host/alcor.c > index c712b7deb3a9d..82a97866e0cf4 100644 > --- a/drivers/mmc/host/alcor.c > +++ b/drivers/mmc/host/alcor.c > @@ -1044,14 +1044,27 @@ static void alcor_init_mmc(struct alcor_sdmmc_host *host) > mmc->caps2 = MMC_CAP2_NO_SDIO; > mmc->ops = &alcor_sdc_ops; > > - /* Hardware cannot do scatter lists */ > + /* The hardware does DMA data transfer of 4096 bytes to/from a single > + * buffer address. Scatterlists are not supported, but upon DMA > + * completion (signalled via IRQ), the original vendor driver does > + * then immediately set up another DMA transfer of the next 4096 > + * bytes. > + * > + * This means that we need to handle the I/O in 4096 byte chunks. > + * Lacking a way to limit the sglist entries to 4096 bytes, we instead > + * impose that only one segment is provided, with maximum size 4096, > + * which also happens to be the minimum size. This means that the > + * single-entry sglist handled by this driver can be handed directly > + * to the hardware, nice and simple. > + * > + * Unfortunately though, that means we only do 4096 bytes I/O per > + * MMC command. A future improvement would be to make the driver > + * accept sg lists and entries of any size, and simply iterate > + * through them 4096 bytes at a time. > + */ > mmc->max_segs = AU6601_MAX_DMA_SEGMENTS; > mmc->max_seg_size = AU6601_MAX_DMA_BLOCK_SIZE; > - > - mmc->max_blk_size = mmc->max_seg_size; > - mmc->max_blk_count = mmc->max_segs; > - > - mmc->max_req_size = mmc->max_seg_size * mmc->max_segs; > + mmc->max_req_size = mmc->max_seg_size; > } > > static int alcor_pci_sdmmc_drv_probe(struct platform_device *pdev) > -- > 2.19.1 >
diff --git a/drivers/mmc/host/alcor.c b/drivers/mmc/host/alcor.c index c712b7deb3a9d..82a97866e0cf4 100644 --- a/drivers/mmc/host/alcor.c +++ b/drivers/mmc/host/alcor.c @@ -1044,14 +1044,27 @@ static void alcor_init_mmc(struct alcor_sdmmc_host *host) mmc->caps2 = MMC_CAP2_NO_SDIO; mmc->ops = &alcor_sdc_ops; - /* Hardware cannot do scatter lists */ + /* The hardware does DMA data transfer of 4096 bytes to/from a single + * buffer address. Scatterlists are not supported, but upon DMA + * completion (signalled via IRQ), the original vendor driver does + * then immediately set up another DMA transfer of the next 4096 + * bytes. + * + * This means that we need to handle the I/O in 4096 byte chunks. + * Lacking a way to limit the sglist entries to 4096 bytes, we instead + * impose that only one segment is provided, with maximum size 4096, + * which also happens to be the minimum size. This means that the + * single-entry sglist handled by this driver can be handed directly + * to the hardware, nice and simple. + * + * Unfortunately though, that means we only do 4096 bytes I/O per + * MMC command. A future improvement would be to make the driver + * accept sg lists and entries of any size, and simply iterate + * through them 4096 bytes at a time. + */ mmc->max_segs = AU6601_MAX_DMA_SEGMENTS; mmc->max_seg_size = AU6601_MAX_DMA_BLOCK_SIZE; - - mmc->max_blk_size = mmc->max_seg_size; - mmc->max_blk_count = mmc->max_segs; - - mmc->max_req_size = mmc->max_seg_size * mmc->max_segs; + mmc->max_req_size = mmc->max_seg_size; } static int alcor_pci_sdmmc_drv_probe(struct platform_device *pdev)
Setting max_blk_count to 1 here was causing the mmc block layer to always use the MMC_READ_SINGLE_BLOCK command here, which the driver does not DMA-accelerate. Drop the max_blk_ settings here. The mmc host defaults suffice, along with the max_segs and max_seg_size settings, which I have now documented in more detail. Now each MMC command reads 4 512-byte blocks, using DMA instead of PIO. On my SD card, this increases read performance (measured with dd) from 167kb/sec to 4.6mb/sec. Link: http://lkml.kernel.org/r/CAD8Lp47L5T3jnAjBiPs1cQ+yFA3L6LJtgFvMETnBrY63-Zdi2g@mail.gmail.com Signed-off-by: Daniel Drake <drake@endlessm.com> --- drivers/mmc/host/alcor.c | 25 +++++++++++++++++++------ 1 file changed, 19 insertions(+), 6 deletions(-)