virtio_net: get build_skb() buf by data ptr

In the case of merge, the page passed into page_to_skb() may be a head
page, not the page where the current data is located. So when trying to
get the buf where the data is located, we should get buf based on
headroom instead of offset.

This patch solves this problem. But if you don't use this patch, the
original code can also run, because if the page is not the page of the
current data, the calculated tailroom will be less than 0, and will not
enter the logic of build_skb() . The significance of this patch is to
modify this logical problem, allowing more situations to use
build_skb().

Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/virtio_net.c

@@ -401,18 +401,13 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
 	/* If headroom is not 0, there is an offset between the beginning of the
 	 * data and the allocated space, otherwise the data and the allocated
 	 * space are aligned.
+	 *
+	 * Buffers with headroom use PAGE_SIZE as alloc size, see
+	 * add_recvbuf_mergeable() + get_mergeable_buf_len()
 	 */
-	if (headroom) {
-		/* Buffers with headroom use PAGE_SIZE as alloc size,
-		 * see add_recvbuf_mergeable() + get_mergeable_buf_len()
-		 */
-		truesize = PAGE_SIZE;
-		tailroom = truesize - len - offset;
-		buf = page_address(page);
-	} else {
-		tailroom = truesize - len;
-		buf = p;
-	}
+	truesize = headroom ? PAGE_SIZE : truesize;
+	tailroom = truesize - len - headroom;
+	buf = p - headroom;
 
 	len -= hdr_len;
 	offset += hdr_padded_len;