lkml.org 
[lkml]   [2020]   [Apr]   [2]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    Patch in this message
    /
    From
    Subject[PATCH v2 2/2] dmaengine: ioat: Decreasing allocation chunk size 2M -> 512K
    Date
    From: Leonid Ravich <Leonid.Ravich@emc.com>

    current IOAT driver using big (2MB) allocations chunk for its descriptors
    therefore each ioat dma engine need 2 such chunks
    (64k entres in ring each entry 64B = 4MB)
    requiring 2 * 2M * dmaengine contiguies memory chunk
    might fail due to memory fragmention.

    so we decreasing chunk size and using more chunks.

    Acked-by: Dave Jiang <dave.jiang@intel.com>
    Signed-off-by: Leonid Ravich <Leonid.Ravich@emc.com>
    ---
    drivers/dma/ioat/dma.h | 2 +-
    1 file changed, 1 insertion(+), 1 deletion(-)

    diff --git a/drivers/dma/ioat/dma.h b/drivers/dma/ioat/dma.h
    index 5216c6b..e6b622e 100644
    --- a/drivers/dma/ioat/dma.h
    +++ b/drivers/dma/ioat/dma.h
    @@ -83,7 +83,7 @@ struct ioatdma_device {

    #define IOAT_MAX_ORDER 16
    #define IOAT_MAX_DESCS (1 << IOAT_MAX_ORDER)
    -#define IOAT_CHUNK_SIZE (SZ_2M)
    +#define IOAT_CHUNK_SIZE (SZ_512K)
    #define IOAT_DESCS_PER_CHUNK (IOAT_CHUNK_SIZE / IOAT_DESC_SZ)

    struct ioat_descs {
    --
    1.9.3
    \
     
     \ /
      Last update: 2020-04-02 18:36    [W:3.343 / U:0.012 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site