Re: [Qemu-devel] [PATCH] eepro100: Frame Reception

[Jason Wang <jasowang@redhat.com>]

On 2016年08月24日 00:46, Mateus Krepsky Ludwich wrote:
> Changed E100 device so it updates EOF, F, and Actual Count fields of Receive
> Frame Descriptor (RFD).

Actual Count fields were updated before this patch I think?

> Assuming RFD.actual_count equals to size.
>
> Signed-off-by: Mateus Krepsky Ludwich <mkludwich@gmail.com>
> ---
>   hw/net/eepro100.c | 3 +++
>   1 file changed, 3 insertions(+)
>
> diff --git a/hw/net/eepro100.c b/hw/net/eepro100.c
> index bab4dbf..c0f3816 100644
> --- a/hw/net/eepro100.c
> +++ b/hw/net/eepro100.c
> @@ -1739,6 +1739,7 @@ static ssize_t nic_receive(NetClientState *nc,
> const uint8_t * buf, size_t size)
>                    &rx, sizeof(eepro100_rx_t));
>       uint16_t rfd_command = le16_to_cpu(rx.command);
>       uint16_t rfd_size = le16_to_cpu(rx.size);
> +    uint16_t rfd_count = size | 0xc000 ; /* Setting EOF and F bits */
>
>       if (size > rfd_size) {
>           logout("Receive buffer (%" PRId16 " bytes) too small for data "
> @@ -1756,6 +1757,8 @@ static ssize_t nic_receive(NetClientState *nc,
> const uint8_t * buf, size_t size)
>                   offsetof(eepro100_rx_t, status), rfd_status);
>       stw_le_pci_dma(&s->dev, s->ru_base + s->ru_offset +
>                   offsetof(eepro100_rx_t, count), size);

Then this line can be removed, or what's maybe even better, you can just 
reuse rfd_size.

> +    stw_le_pci_dma(&s->dev, s->ru_base + s->ru_offset +
> +                offsetof(eepro100_rx_t, count), rfd_count);
>       /* Early receive interrupt not supported. */
>   #if 0
>       eepro100_er_interrupt(s);

We need a better title, e.g "Fixing EOF and F filed when updating RFD"

Thanks