Skip to content

Commit

Permalink
net: test nouarg before dereferencing zerocopy pointers
Browse files Browse the repository at this point in the history
Zerocopy skbs without completion notification were added for packet
sockets with PACKET_TX_RING user buffers. Those signal completion
through the TP_STATUS_USER bit in the ring. Zerocopy annotation was
added only to avoid premature notification after clone or orphan, by
triggering a copy on these paths for these packets.

The mechanism had to define a special "no-uarg" mode because packet
sockets already use skb_uarg(skb) == skb_shinfo(skb)->destructor_arg
for a different pointer.

Before deferencing skb_uarg(skb), verify that it is a real pointer.

Fixes: 5cd8d46 ("packet: copy user buffers before orphan or clone")
Signed-off-by: Willem de Bruijn <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
  • Loading branch information
wdebruij authored and davem330 committed May 16, 2019
1 parent b4e467c commit 185ce5c
Showing 1 changed file with 6 additions and 3 deletions.
9 changes: 6 additions & 3 deletions include/linux/skbuff.h
Original file line number Diff line number Diff line change
Expand Up @@ -1434,10 +1434,12 @@ static inline void skb_zcopy_clear(struct sk_buff *skb, bool zerocopy)
struct ubuf_info *uarg = skb_zcopy(skb);

if (uarg) {
if (uarg->callback == sock_zerocopy_callback) {
if (skb_zcopy_is_nouarg(skb)) {
/* no notification callback */
} else if (uarg->callback == sock_zerocopy_callback) {
uarg->zerocopy = uarg->zerocopy && zerocopy;
sock_zerocopy_put(uarg);
} else if (!skb_zcopy_is_nouarg(skb)) {
} else {
uarg->callback(uarg, zerocopy);
}

Expand Down Expand Up @@ -2691,7 +2693,8 @@ static inline int skb_orphan_frags(struct sk_buff *skb, gfp_t gfp_mask)
{
if (likely(!skb_zcopy(skb)))
return 0;
if (skb_uarg(skb)->callback == sock_zerocopy_callback)
if (!skb_zcopy_is_nouarg(skb) &&
skb_uarg(skb)->callback == sock_zerocopy_callback)
return 0;
return skb_copy_ubufs(skb, gfp_mask);
}
Expand Down

0 comments on commit 185ce5c

Please sign in to comment.