Hi,
On Wed, Oct 10, 2012 at 07:17:58PM -0700, Kuninori Morimoto wrote:
> This patch enabled dma mapping, and used dma transfer handler
> on mod_host
>
> Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@xxxxxxxxxxx>
> ---
> Felipe
>
> I sent some patches before, and I guess it will go to v3.8 branch.
> please let me know if you want me to resend these.
>
> drivers/usb/renesas_usbhs/mod_host.c | 17 +++++++++++++++--
> 1 file changed, 15 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/usb/renesas_usbhs/mod_host.c b/drivers/usb/renesas_usbhs/mod_host.c
> index 9b69a13..e856b44 100644
> --- a/drivers/usb/renesas_usbhs/mod_host.c
> +++ b/drivers/usb/renesas_usbhs/mod_host.c
> @@ -681,9 +681,9 @@ static int usbhsh_queue_push(struct usb_hcd *hcd,
> }
>
> if (usb_pipein(urb->pipe))
> - pipe->handler = &usbhs_fifo_pio_pop_handler;
> + pipe->handler = &usbhs_fifo_dma_pop_handler;
> else
> - pipe->handler = &usbhs_fifo_pio_push_handler;
> + pipe->handler = &usbhs_fifo_dma_push_handler;
>
> buf = (void *)(urb->transfer_buffer + urb->actual_length);
> len = urb->transfer_buffer_length - urb->actual_length;
> @@ -916,6 +916,19 @@ static int usbhsh_dcp_queue_push(struct usb_hcd *hcd,
> */
> static int usbhsh_dma_map_ctrl(struct usbhs_pkt *pkt, int map)
> {
> + if (map) {
> + struct usbhsh_request *ureq = usbhsh_pkt_to_ureq(pkt);
> + struct urb *urb = ureq->urb;
> +
> + /* it can not use scatter/gather */
> + if (urb->num_sgs)
> + return -EINVAL;
I'm not the most knowledgeable on the host stack, but is this really
necessary ? Isn't host stack supposed to check sg_tablesize and handle
that properly ?
Alan ? Greg ?
> + pkt->dma = urb->transfer_dma;
> + if (!pkt->dma)
> + return -EINVAL;
> + }
> +
> return 0;
> }
>
> --
> 1.7.9.5
>
--
balbi
Attachment:
signature.asc
Description: Digital signature
