Re: [PATCH v5 04/21] net: Remove receive_raw()

[Jason Wang]

On Tue, Oct 17, 2023 at 12:09 PM Akihiko Odaki <akihiko.odaki@daynix.com> wrote:
>
> While netmap implements virtio-net header, it does not implement
> receive_raw().

The only user for raw is the announcing. Netmap probably doesn't it at all.

> Instead of implementing receive_raw for netmap, add
> virtio-net headers in the common code and use receive_iov()/receive()
> instead. This also fixes the buffer size for the virtio-net header.
>
> Fixes: fbbdbddec0 ("tap: allow extended virtio header with hash info")
> Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
> ---
>  include/net/net.h   |  1 -
>  include/net/queue.h |  1 -
>  net/net.c           | 17 +++++++++--------
>  net/queue.c         | 30 ++++++++++--------------------
>  net/tap.c           |  1 -
>  5 files changed, 19 insertions(+), 31 deletions(-)
>
> diff --git a/include/net/net.h b/include/net/net.h
> index 011031ef77..971dc54897 100644
> --- a/include/net/net.h
> +++ b/include/net/net.h
> @@ -71,7 +71,6 @@ typedef struct NetClientInfo {
>      NetClientDriver type;
>      size_t size;
>      NetReceive *receive;
> -    NetReceive *receive_raw;
>      NetReceiveIOV *receive_iov;
>      NetCanReceive *can_receive;
>      NetStart *start;
> diff --git a/include/net/queue.h b/include/net/queue.h
> index 9f2f289d77..7a43863be2 100644
> --- a/include/net/queue.h
> +++ b/include/net/queue.h
> @@ -31,7 +31,6 @@ typedef struct NetQueue NetQueue;
>  typedef void (NetPacketSent) (NetClientState *sender, ssize_t ret);
>
>  #define QEMU_NET_PACKET_FLAG_NONE  0
> -#define QEMU_NET_PACKET_FLAG_RAW  (1<<0)
>
>  /* Returns:
>   *   >0 - success
> diff --git a/net/net.c b/net/net.c
> index 6d2fa8d40f..2d94d23c07 100644
> --- a/net/net.c
> +++ b/net/net.c
> @@ -752,8 +752,13 @@ ssize_t qemu_receive_packet_iov(NetClientState *nc, 
> const struct iovec *iov,
>
>  ssize_t qemu_send_packet_raw(NetClientState *nc, const uint8_t *buf, int 
> size)
>  {
> -    return qemu_send_packet_async_with_flags(nc, QEMU_NET_PACKET_FLAG_RAW,
> -                                             buf, size, NULL);
> +    struct virtio_net_hdr_v1_hash vnet_hdr = { };
> +    struct iovec iov[] = {
> +        { .iov_base = &vnet_hdr, .iov_len = nc->vnet_hdr_len },
> +        { .iov_base = (void *)buf, .iov_len = size }
> +    };

Having virtio-net specific code in the net layer is a layer violation.
I'd leave it to tap.

Thanks