skbuff: skb_copy_ubufs must release uarg even without user frags
authorWillem de Bruijn <willemb@google.com>
Wed, 20 Dec 2017 22:37:50 +0000 (17:37 -0500)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 2 Jan 2018 19:31:13 +0000 (20:31 +0100)
[ Upstream commit b90ddd568792bcb0054eaf0f61785c8f80c3bd1c ]

skb_copy_ubufs creates a private copy of frags[] to release its hold
on user frags, then calls uarg->callback to notify the owner.

Call uarg->callback even when no frags exist. This edge case can
happen when zerocopy_sg_from_iter finds enough room in skb_headlen
to copy all the data.

Fixes: 3ece782693c4 ("sock: skb_copy_ubufs support for compound pages")
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
net/core/skbuff.c

index 3c2e1db..4a10e96 100644 (file)
@@ -1182,7 +1182,7 @@ int skb_copy_ubufs(struct sk_buff *skb, gfp_t gfp_mask)
        u32 d_off;
 
        if (!num_frags)
-               return 0;
+               goto release;
 
        if (skb_shared(skb) || skb_unclone(skb, gfp_mask))
                return -EINVAL;
@@ -1242,6 +1242,7 @@ int skb_copy_ubufs(struct sk_buff *skb, gfp_t gfp_mask)
        __skb_fill_page_desc(skb, new_frags - 1, head, 0, d_off);
        skb_shinfo(skb)->nr_frags = new_frags;
 
+release:
        skb_zcopy_clear(skb, false);
        return 0;
 }