[PATCH 2/2] wcn36xx: don't process 'valid' descriptors

[Eugene Krasnikov]

Make sense, thanx!

2014-03-18 12:09 GMT+00:00 Bob Copeland <me at bobcopeland.com>:
> The DMA engine will reset the valid bit after a descriptor is
> complete; any with the valid bit still set may still be in
> use by the hardware, so check that before processing the
> descriptor.  We were checking this but only after the first
> iteration of the do-while loop.
>
> Signed-off-by: Bob Copeland <bob at cozybit.com>
> ---
>  dxe.c |    5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/dxe.c b/dxe.c
> index 61f398b..51a0467 100644
> --- a/dxe.c
> +++ b/dxe.c
> @@ -356,6 +356,8 @@ static void reap_tx_dxes(struct wcn36xx *wcn, struct wcn36xx_dxe_ch *ch)
>          * and while-do will not make any cycles.
>          */
>         do {
> +               if (ctl->desc->ctrl & WCN36XX_DXE_CTRL_VALID_MASK)
> +                       break;
>                 if (ctl->skb) {
>                         dma_unmap_single(NULL, ctl->desc->src_addr_l,
>                                          ctl->skb->len, DMA_TO_DEVICE);
> @@ -374,8 +376,7 @@ static void reap_tx_dxes(struct wcn36xx *wcn, struct wcn36xx_dxe_ch *ch)
>                         ctl->skb = NULL;
>                 }
>                 ctl = ctl->next;
> -       } while (ctl != ch->head_blk_ctl &&
> -              !(ctl->desc->ctrl & WCN36XX_DXE_CTRL_VALID_MASK));
> +       } while (ctl != ch->head_blk_ctl);
>
>         ch->tail_blk_ctl = ctl;
>  }
> --
> 1.7.10.4
>
>
> _______________________________________________
> wcn36xx mailing list
> wcn36xx at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/wcn36xx



-- 
Best regards,
Eugene