Re: [PATCH bpf-next v2 3/3] xsk: build skb by page

[Alexander Lobakin <alobakin@pm.me>]

From: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Date: Tue, 19 Jan 2021 17:45:12 +0800

> This patch is used to construct skb based on page to save memory copy
> overhead.
> 
> This function is implemented based on IFF_TX_SKB_NO_LINEAR. Only the
> network card priv_flags supports IFF_TX_SKB_NO_LINEAR will use page to
> directly construct skb. If this feature is not supported, it is still
> necessary to copy data to construct skb.
> 
> ---------------- Performance Testing ------------
> 
> The test environment is Aliyun ECS server.
> Test cmd:
> ```
> xdpsock -i eth0 -t  -S -s <msg size>
> ```
> 
> Test result data:
> 
> size    64      512     1024    1500
> copy    1916747 1775988 1600203 1440054
> page    1974058 1953655 1945463 1904478
> percent 3.0%    10.0%   21.58%  32.3%
> 
> Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
> Reviewed-by: Dust Li <dust.li@linux.alibaba.com>
> ---
>  net/xdp/xsk.c | 112 ++++++++++++++++++++++++++++++++++++++++++++++++----------
>  1 file changed, 94 insertions(+), 18 deletions(-)
> 
> diff --git a/net/xdp/xsk.c b/net/xdp/xsk.c
> index 8037b04..8c291f8 100644
> --- a/net/xdp/xsk.c
> +++ b/net/xdp/xsk.c
> @@ -430,6 +430,95 @@ static void xsk_destruct_skb(struct sk_buff *skb)
>  	sock_wfree(skb);
>  }
>  
> +static struct sk_buff *xsk_build_skb_zerocopy(struct xdp_sock *xs,
> +					      struct xdp_desc *desc)
> +{
> +	u32 len, offset, copy, copied;
> +	struct sk_buff *skb;
> +	struct page *page;
> +	char *buffer;
> +	int err = 0, i;
> +	u64 addr;
> +
> +	skb = sock_alloc_send_skb(&xs->sk, 0, 1, &err);
> +	if (unlikely(!skb))
> +		return NULL;

You can propagate err from here to the outer function:

	if (unlikely(!skb))
		return ERR_PTR(err);

> +	addr = desc->addr;
> +	len = desc->len;
> +
> +	buffer = xsk_buff_raw_get_data(xs->pool, addr);
> +	offset = offset_in_page(buffer);
> +	addr = buffer - (char *)xs->pool->addrs;
> +
> +	for (copied = 0, i = 0; copied < len; ++i) {

i++ would be less confusing here. You build skb frags from frag 0
anyway.

> +		page = xs->pool->umem->pgs[addr >> PAGE_SHIFT];
> +
> +		get_page(page);
> +
> +		copy = min((u32)(PAGE_SIZE - offset), len - copied);

It's better to use min_t(u32, ...) instead of manual casting.

> +
> +		skb_fill_page_desc(skb, i, page, offset, copy);
> +
> +		copied += copy;
> +		addr += copy;
> +		offset = 0;
> +	}
> +
> +	skb->len += len;
> +	skb->data_len += len;
> +	skb->truesize += len;
> +
> +	refcount_add(len, &xs->sk.sk_wmem_alloc);
> +
> +	return skb;
> +}
> +
> +static struct sk_buff *xsk_build_skb(struct xdp_sock *xs,
> +				     struct xdp_desc *desc)
> +{
> +	struct sk_buff *skb = NULL;
> +	int err = -ENOMEM;
> +
> +	if (xs->dev->priv_flags & IFF_TX_SKB_NO_LINEAR) {
> +		skb = xsk_build_skb_zerocopy(xs, desc);
> +		if (unlikely(!skb))
> +			goto err;

1. You should'n use goto err here, as skb == NULL, so kfree_skb(skb)
   is redundant.
2. If you would use ERR_PTR() in xsk_build_skb_zerocopy(),
   the condition should look like:

		if (IS_ERR(skb))
			return PTR_ERR(skb);

> +	} else {
> +		char *buffer;
> +		u64 addr;
> +		u32 len;
> +		int err;
> +
> +		len = desc->len;
> +		skb = sock_alloc_send_skb(&xs->sk, len, 1, &err);
> +		if (unlikely(!skb))
> +			goto err;

Same here, if skb == NULL, just return without calling kfree_skb().

> +		skb_put(skb, len);
> +		addr = desc->addr;
> +		buffer = xsk_buff_raw_get_data(xs->pool, desc->addr);
> +		err = skb_store_bits(skb, 0, buffer, len);
> +
> +		if (unlikely(err)) {
> +			err = -EINVAL;

You already have errno in err, no need to override it.

> +			goto err;
> +		}
> +	}
> +
> +	skb->dev = xs->dev;
> +	skb->priority = xs->sk.sk_priority;
> +	skb->mark = xs->sk.sk_mark;
> +	skb_shinfo(skb)->destructor_arg = (void *)(long)desc->addr;
> +	skb->destructor = xsk_destruct_skb;
> +
> +	return skb;
> +
> +err:
> +	kfree_skb(skb);
> +	return ERR_PTR(err);
> +}
> +
>  static int xsk_generic_xmit(struct sock *sk)
>  {
>  	struct xdp_sock *xs = xdp_sk(sk);
> @@ -446,43 +535,30 @@ static int xsk_generic_xmit(struct sock *sk)
>  		goto out;
>  
>  	while (xskq_cons_peek_desc(xs->tx, &desc, xs->pool)) {
> -		char *buffer;
> -		u64 addr;
> -		u32 len;
> -
>  		if (max_batch-- == 0) {
>  			err = -EAGAIN;
>  			goto out;
>  		}
>  
> -		len = desc.len;
> -		skb = sock_alloc_send_skb(sk, len, 1, &err);
> -		if (unlikely(!skb))
> +		skb = xsk_build_skb(xs, &desc);
> +		if (IS_ERR(skb)) {
> +			err = PTR_ERR(skb);
>  			goto out;
> +		}
>  
> -		skb_put(skb, len);
> -		addr = desc.addr;
> -		buffer = xsk_buff_raw_get_data(xs->pool, addr);
> -		err = skb_store_bits(skb, 0, buffer, len);
>  		/* This is the backpressure mechanism for the Tx path.
>  		 * Reserve space in the completion queue and only proceed
>  		 * if there is space in it. This avoids having to implement
>  		 * any buffering in the Tx path.
>  		 */
>  		spin_lock_irqsave(&xs->pool->cq_lock, flags);
> -		if (unlikely(err) || xskq_prod_reserve(xs->pool->cq)) {
> +		if (xskq_prod_reserve(xs->pool->cq)) {
>  			spin_unlock_irqrestore(&xs->pool->cq_lock, flags);
>  			kfree_skb(skb);
>  			goto out;
>  		}
>  		spin_unlock_irqrestore(&xs->pool->cq_lock, flags);
>  
> -		skb->dev = xs->dev;
> -		skb->priority = sk->sk_priority;
> -		skb->mark = sk->sk_mark;
> -		skb_shinfo(skb)->destructor_arg = (void *)(long)desc.addr;
> -		skb->destructor = xsk_destruct_skb;
> -
>  		err = __dev_direct_xmit(skb, xs->queue_id);
>  		if  (err == NETDEV_TX_BUSY) {
>  			/* Tell user-space to retry the send */

So please recheck the code and then retest it, especially error
paths (you can inject errors manually here to ensure they work).

Thanks,
Al