mirrors/madwifi
1
0
Fork 0
mirror of https://github.com/proski/madwifi synced 2024-11-22 06:21:47 +03:00
Code Issues Packages Projects Releases Wiki Activity

This is another case where we are getting clones of SKB and then modifying the header or the data. In a bridge, this is bad behavior. This fixes more bridge anomalies where incorrect header/data are sent to wired ports when packets are previously forwarded over madwifi port(s).

Browse Source 
git-svn-id: http://madwifi-project.org/svn/madwifi/trunk@3176 0192ed92-7a03-0410-a25b-9323aeb14dbd
This commit is contained in:
mtaylor 2008-01-16 22:17:53 +00:00
parent 00a96b9d09
commit 4183b417a2
1 changed files with 25 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
ath/if_ath.c
View File

@ -3080,9 +3080,34 @@ ath_hardstart(struct sk_buff *skb, struct net_device *dev)
/* queue is full, let the kernel backlog the skb */
netif_stop_queue(dev);
requeue = 1;
goto hardstart_fail;
}
#endif
/* If the skb data is shared, we will copy it so we can strip padding
* without affecting any other bridge ports. */
if (skb_cloned(skb)) {
/* Remember the original SKB so we can free up our references */
struct sk_buff *skb_orig = skb;
skb = skb_copy(skb, GFP_ATOMIC);
if (skb == NULL) {
requeue = 1;
netif_stop_queue(dev);
goto hardstart_fail;
}
/* If the clone works, bump the reference count for our copy. */
SKB_CB(skb)->ni = ieee80211_ref_node(SKB_CB(skb_orig)->ni);
ieee80211_dev_kfree_skb(&skb_orig);
} else {
if (SKB_CB(skb)->ni != NULL)
ieee80211_unref_node(&SKB_CB(skb)->ni);
skb_orphan(skb);
}
eh = (struct ether_header *) skb->data;
#ifdef ATH_SUPERG_FF
/* NB: use this lock to protect an->an_tx_ffbuf (and txq->axq_stageq)
* in athff_can_aggregate() call too. */
ATH_TXQ_LOCK_IRQ(txq);