skbuff: don't corrupt mac_header on skb expansion
Subsystems (2)
PATCH 1/1 skbuff: don't corrupt mac_header on skb expansion
The skb mac_header field is sometimes NULL (or ~0u) as a sentinel value. The places where skb is expanded add an offset which would change this flag into an invalid pointer (or offset).
net/core/skbuff.c +6 -3
@@ -657,7 +657,8 @@ static void copy_skb_header(struct sk_bu/* {transport,network,mac}_header are relative to skb->head */ new->transport_header += offset; new->network_header += offset; - new->mac_header += offset; +if (skb_mac_header_was_set (new)) + new->mac_header += offset;#endif skb_shinfo (new)->gso_size =skb_shinfo (old)->gso_size;skb_shinfo (new)->gso_segs =skb_shinfo (old)->gso_segs;
@@ -839,7 +840,8 @@ int pskb_expand_head(struct sk_buff *skb skb->tail += off; skb->transport_header += off; skb->network_header += off; - skb->mac_header += off; +if (skb_mac_header_was_set (skb)) + skb->mac_header += off; skb->csum_start += nhead; skb->cloned =0 ; skb->hdr_len =0 ;
@@ -931,7 +933,8 @@ struct sk_buff *skb_copy_expand(const st#ifdef NET_SKBUFF_DATA_USES_OFFSET n->transport_header += off; n->network_header += off; - n->mac_header += off; +if (skb_mac_header_was_set (skb)) + n->mac_header += off;#endif return n;
From: Stephen Hemminger <[email protected]>
Date: Wed, 17 Jun 2009 15:17:34 -0700
Quoted 5 lines
The skb
mac_headerfield is sometimes NULL (or ~0u) as a sentinel value. The places where skb is expanded add an offset which would change this flag into an invalid pointer (or offset).Signed-off-by Stephen Hemminger <[email protected]>
Applied.
Quoted 5 lines · Stephen Hemminger
The skb
mac_headerfield is sometimes NULL (or ~0u) as a sentinel value. The places where skb is expanded add an offset which would change this flag into an invalid pointer (or offset).Signed-off-by Stephen Hemminger <[email protected]>
What uses skb_mac_header_was_set apart from a debugging printk in route.c and nfnetlink_log (which is also essentially a debugging printk)?
This seems to be a bit much just to make a couple printks work?
What about using mac_len instead to figure out whether the MAC header is there?
Cheers,
Signature · 4 lines
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txtQuoted 11 lines · On Sun, 5 Jul 2009 11:55:21 +0800 Herbert Xu <[email protected]>…
Quoted 5 lines · Stephen Hemminger
The skb
mac_headerfield is sometimes NULL (or ~0u) as a sentinel value. The places where skb is expanded add an offset which would change this flag into an invalid pointer (or offset).Signed-off-by Stephen Hemminger <[email protected]>What uses
skb_mac_header_was_setapart from a debugging printk inroute.candnfnetlink_log(which is also essentially a debugging printk)?
Quoted 6 lines
This seems to be a bit much just to make a couple printks work?
What about using
mac_leninstead to figure out whether the MAC header is there?Cheers,
No worries, either way works, I was just trying to keep acme's original code