在 2021/5/14 下午11:16, Xuan Zhuo 写道:
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.
I don't get how this can happen?
Maybe you can explain a little bit more?
receive_mergeable() call page_to_skb() in two places:
1) XDP_PASS for linearized page , in this case we use xdp_page
2) page_to_skb() for "normal" page, in this case the page contains the data
Thanks
So when trying to
get the buf where the data is located, you should directly use the
pointer(p) to get the address corresponding to the page.
At the same time, the offset of the data in the page should also be
obtained using offset_in_page().
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@xxxxxxxxxxxxxxxxx>
Acked-by: Michael S. Tsirkin <mst@xxxxxxxxxx>
---
drivers/net/virtio_net.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 3e46c12dde08..073fec4c0df1 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -407,8 +407,12 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
* see add_recvbuf_mergeable() + get_mergeable_buf_len()
*/
truesize = PAGE_SIZE;
- tailroom = truesize - len - offset;
- buf = page_address(page);
+
+ /* page maybe head page, so we should get the buf by p, not the
+ * page
+ */
+ tailroom = truesize - len - offset_in_page(p);
+ buf = (char *)((unsigned long)p & PAGE_MASK);
} else {
tailroom = truesize - len;
buf = p;
_______________________________________________
Virtualization mailing list
Virtualization@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
