Re: [PATCH net] xsk: Fix error return code in __xp_assign_dev()

[Magnus Karlsson]

On Fri, Dec 4, 2020 at 9:49 AM Zhang Changzhong
<zhangchangzhong@xxxxxxxxxx> wrote:
>
> Fix to return a negative error code from the error handling
> case instead of 0, as done elsewhere in this function.
>
> Fixes: 921b68692abb ("xsk: Enable sharing of dma mappings")
> Reported-by: Hulk Robot <hulkci@xxxxxxxxxx>
> Signed-off-by: Zhang Changzhong <zhangchangzhong@xxxxxxxxxx>
> ---
>  net/xdp/xsk_buff_pool.c | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/net/xdp/xsk_buff_pool.c b/net/xdp/xsk_buff_pool.c
> index 9287edd..d5adeee 100644
> --- a/net/xdp/xsk_buff_pool.c
> +++ b/net/xdp/xsk_buff_pool.c
> @@ -175,6 +175,7 @@ static int __xp_assign_dev(struct xsk_buff_pool *pool,
>
>         if (!pool->dma_pages) {
>                 WARN(1, "Driver did not DMA map zero-copy buffers");
> +               err = -EINVAL;

Good catch! My intention here by not setting err is that it should
fall back to copy mode, which it does. The problem is that the
force_zc flag is disregarded when err is not set (see exit code below)
and your patch fixes that. If force_zc is set, we should exit out with
an error, not fall back. Could you please write about this in your
cover letter and send a v2?

BTW, what is the "Hulk Robot" that is in your Reported-by tag?

Thank you: Magnus

err_unreg_xsk:
        xp_disable_drv_zc(pool);
err_unreg_pool:
        if (!force_zc)
                err = 0; /* fallback to copy mode */
        if (err)
                xsk_clear_pool_at_qid(netdev, queue_id);
        return err;

>                 goto err_unreg_xsk;
>         }
>         pool->umem->zc = true;
> --
> 2.9.5
>