lkml.org 
[lkml]   [2020]   [Jan]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    Patch in this message
    /
    From
    Subject[PATCH 13/14] net: axienet: Allow DMA to beyond 4GB
    Date
    With all DMA address accesses wrapped, we can actually support 64-bit
    DMA if this option was chosen at IP integration time.
    Since there is no way to autodetect the actual address bus width, we
    make use of the existing "xlnx,addrwidth" DT property to let the driver
    know about the width. The value in this property should match the
    "Address Width" parameter used when synthesizing the IP.

    This increases the DMA mask to let the kernel choose buffers from
    memory at higher addresses.

    Cc: Rob Herring <robh+dt@kernel.org>
    Cc: Mark Rutland <mark.rutland@arm.com>
    Cc: devicetree@vger.kernel.org
    Signed-off-by: Andre Przywara <andre.przywara@arm.com>
    ---
    .../net/ethernet/xilinx/xilinx_axienet_main.c | 19 +++++++++++++++++++
    1 file changed, 19 insertions(+)

    diff --git a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
    index f7f593df0c11..e036834549b3 100644
    --- a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
    +++ b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c
    @@ -1786,6 +1786,7 @@ static int axienet_probe(struct platform_device *pdev)
    struct net_device *ndev;
    const void *mac_addr;
    struct resource *ethres;
    + int addr_width = 32;
    u32 value;

    ndev = alloc_etherdev(sizeof(*lp));
    @@ -1915,6 +1916,8 @@ static int axienet_probe(struct platform_device *pdev)
    &dmares);
    lp->rx_irq = irq_of_parse_and_map(np, 1);
    lp->tx_irq = irq_of_parse_and_map(np, 0);
    + of_property_read_u32(np, "xlnx,addrwidth", &addr_width);
    +
    of_node_put(np);
    lp->eth_irq = platform_get_irq(pdev, 0);
    } else {
    @@ -1944,6 +1947,9 @@ static int axienet_probe(struct platform_device *pdev)
    * We can detect this case by writing all 1's to one such register
    * and see if that sticks: when the IP is configured for 32 bits
    * only, those registers are RES0.
    + * We can't autodetect the actual width this way, so we still use
    + * a 32-bit DMA mask and rely on the xlnk,addrwidth DT property
    + * to set this properly.
    * Those MSB registers were introduced in IP v7.1, which we check first.
    */
    if ((axienet_ior(lp, XAE_ID_OFFSET) >> 24) >= 0x9) {
    @@ -1961,6 +1967,19 @@ static int axienet_probe(struct platform_device *pdev)
    }
    }

    + if (!(lp->features & XAE_FEATURE_DMA_64BIT)) {
    + if (addr_width > 32)
    + dev_warn(&pdev->dev, "trimming DMA width from %d to 32 bits\n",
    + addr_width);
    + addr_width = 32;
    + }
    +
    + ret = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(addr_width));
    + if (ret) {
    + dev_err(&pdev->dev, "No suitable DMA available\n");
    + goto free_netdev;
    + }
    +
    /* Check for Ethernet core IRQ (optional) */
    if (lp->eth_irq <= 0)
    dev_info(&pdev->dev, "Ethernet core IRQ not defined\n");
    --
    2.17.1
    \
     
     \ /
      Last update: 2020-01-10 13:09    [W:4.254 / U:0.048 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site