Re: [PATCH net-next 21/33] virtio_net: add XDP frame size in two code paths

On 2020/4/27 下午10:32, Jesper Dangaard Brouer wrote:

On Mon, 27 Apr 2020 15:21:02 +0800
Jason Wang[off-list ref]  wrote:

quoted

On 2020/4/23 上午12:09, Jesper Dangaard Brouer wrote:

quoted

The virtio_net driver is running inside the guest-OS. There are two
XDP receive code-paths in virtio_net, namely receive_small() and
receive_mergeable(). The receive_big() function does not support XDP.

In receive_small() the frame size is available in buflen. The buffer
backing these frames are allocated in add_recvbuf_small() with same
size, except for the headroom, but tailroom have reserved room for
skb_shared_info. The headroom is encoded in ctx pointer as a value.

In receive_mergeable() the frame size is more dynamic. There are two
basic cases: (1) buffer size is based on a exponentially weighted
moving average (see DECLARE_EWMA) of packet length. Or (2) in case
virtnet_get_headroom() have any headroom then buffer size is
PAGE_SIZE. The ctx pointer is this time used for encoding two values;
the buffer len "truesize" and headroom. In case (1) if the rx buffer
size is underestimated, the packet will have been split over more
buffers (num_buf info in virtio_net_hdr_mrg_rxbuf placed in top of
buffer area). If that happens the XDP path does a xdp_linearize_page
operation.

Cc: Jason Wang<redacted>
Signed-off-by: Jesper Dangaard Brouer<redacted>
---
   drivers/net/virtio_net.c |   15 ++++++++++++---
   1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 11f722460513..1df3676da185 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c

@@ -689,6 +689,7 @@ static struct sk_buff *receive_small(struct net_device *dev,
   		xdp.data_end = xdp.data + len;
   		xdp.data_meta = xdp.data;
   		xdp.rxq = &rq->xdp_rxq;
+		xdp.frame_sz = buflen;
   		orig_data = xdp.data;
   		act = bpf_prog_run_xdp(xdp_prog, &xdp);
   		stats->xdp_packets++;

@@ -797,10 +798,11 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
   	int offset = buf - page_address(page);
   	struct sk_buff *head_skb, *curr_skb;
   	struct bpf_prog *xdp_prog;
-	unsigned int truesize;
+	unsigned int truesize = mergeable_ctx_to_truesize(ctx);
   	unsigned int headroom = mergeable_ctx_to_headroom(ctx);
-	int err;
   	unsigned int metasize = 0;
+	unsigned int frame_sz;
+	int err;
   
   	head_skb = NULL;
   	stats->bytes += len - vi->hdr_len;

@@ -821,6 +823,11 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
   		if (unlikely(hdr->hdr.gso_type))
   			goto err_xdp;
   
+		/* Buffers with headroom use PAGE_SIZE as alloc size,
+		 * see add_recvbuf_mergeable() + get_mergeable_buf_len()
+		 */
+		frame_sz = headroom ? PAGE_SIZE : truesize;
+
   		/* This happens when rx buffer size is underestimated
   		 * or headroom is not enough because of the buffer
   		 * was refilled before XDP is set. This should only

@@ -834,6 +841,8 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
   						      page, offset,
   						      VIRTIO_XDP_HEADROOM,
   						      &len);
+			frame_sz = PAGE_SIZE;

Should this be PAGE_SIZE -  SKB_DATA_ALIGN(sizeof(struct skb_shared_info))?

No, frame_sz include the SKB_DATA_ALIGN(sizeof(struct skb_shared_info)) length.

Ok, consider mergeable buffer path depends on the truesize which is encoded in ctx.

It looks to the the calculation in add_recvfbuf_mergeable() is wrong, we need count both headroom and tailroom there.

We probably need the attached 2 patches to fix this.

(untested, will test it tomorrow).

Thanks