Messages in this thread Patch in this message | | | Subject | [PATCH v3] ip: fix truesize mismatch in ip fragmentation | From | Eric Dumazet <> | Date | Tue, 21 Sep 2010 17:58:25 +0200 |
| |
Le mardi 21 septembre 2010 à 16:16 +0200, Eric Dumazet a écrit : > Le mardi 21 septembre 2010 à 10:05 -0400, Nick Bowler a écrit : > > > This hunk introduces some whitespace damage. > > > > Anyway, I tried this with ESP on both IPv4 and IPv6 and it appears to > > correct the issue. Thanks! > > > > Indeed good catch. > > Here is an updated patch, I added your Tested-by > > Thanks for testing ! > > [PATCH] ip : fix truesize mismatch in ip fragmentation > > We should not set frag->destructor to sock_wkfree() until we are sure we > dont hit slow path in ip_fragment(). Or we risk uncharging > frag->truesize twice, and in the end, having negative socket > sk_wmem_alloc counter, or even freeing socket sooner than expected. > > Many thanks to Nick Bowler, who provided a very clean bug report and > test program. > > While Nick bisection pointed to commit 2b85a34e911bf483 (net: No more > expensive sock_hold()/sock_put() on each tx), underlying bug is older. >
Hmm, while looking at git history, I found commit from Patrick b2722b1c3a893e (ip_fragment: also adjust skb->truesize for packets not owned by a socket) As my patch reverts it, we probably want a more polished patch.
(Also port Patrick work to ipv6)
So here is a V3
[PATCH v3] ip: fix truesize mismatch in ip fragmentation
Special care should be taken when slow path is hit in ip_fragment() :
When walking through frags, we transfert truesize ownership from skb to frags. Then if we hit a slow_path condition, we must undo this or risk uncharging frags->truesize twice, and in the end, having negative socket sk_wmem_alloc counter, or even freeing socket sooner than expected.
Many thanks to Nick Bowler, who provided a very clean bug report and test program.
Thanks to Jarek for reviewing my first patch and providing a V2
While Nick bisection pointed to commit 2b85a34e911 (net: No more expensive sock_hold()/sock_put() on each tx), underlying bug is older (2.6.12-rc5)
A side effect is to extend work done in commit b2722b1c3a893e (ip_fragment: also adjust skb->truesize for packets not owned by a socket) to ipv6 as well.
Reported-and-bisected-by: Nick Bowler <nbowler@elliptictech.com> Tested-by: Nick Bowler <nbowler@elliptictech.com> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> CC: Jarek Poplawski <jarkao2@gmail.com> CC: Patrick McHardy <kaber@trash.net> --- net/ipv4/ip_output.c | 19 +++++++++++++------ net/ipv6/ip6_output.c | 18 +++++++++++++----- 2 files changed, 26 insertions(+), 11 deletions(-)
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c index 04b6989..a643f7a 100644 --- a/net/ipv4/ip_output.c +++ b/net/ipv4/ip_output.c @@ -488,9 +488,8 @@ int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *)) * we can switch to copy when see the first bad fragment. */ if (skb_has_frags(skb)) { - struct sk_buff *frag; + struct sk_buff *frag, *frag2; int first_len = skb_pagelen(skb); - int truesizes = 0; if (first_len - hlen > mtu || ((first_len - hlen) & 7) || @@ -503,18 +502,18 @@ int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *)) if (frag->len > mtu || ((frag->len & 7) && frag->next) || skb_headroom(frag) < hlen) - goto slow_path; + goto slow_path_undo; /* Partially cloned skb? */ if (skb_shared(frag)) - goto slow_path; + goto slow_path_undo; BUG_ON(frag->sk); if (skb->sk) { frag->sk = skb->sk; frag->destructor = sock_wfree; } - truesizes += frag->truesize; + skb->truesize -= frag->truesize; } /* Everything is OK. Generate! */ @@ -524,7 +523,6 @@ int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *)) frag = skb_shinfo(skb)->frag_list; skb_frag_list_init(skb); skb->data_len = first_len - skb_headlen(skb); - skb->truesize -= truesizes; skb->len = first_len; iph->tot_len = htons(first_len); iph->frag_off = htons(IP_MF); @@ -576,6 +574,15 @@ int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *)) } IP_INC_STATS(dev_net(dev), IPSTATS_MIB_FRAGFAILS); return err; + +slow_path_undo: + skb_walk_frags(skb, frag2) { + if (frag2 == frag) + break; + frag2->sk = NULL; + frag2->destructor = NULL; + skb->truesize += frag2->truesize; + } } slow_path: diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index d40b330..ca7ba44 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -639,7 +639,7 @@ static int ip6_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *)) if (skb_has_frags(skb)) { int first_len = skb_pagelen(skb); - int truesizes = 0; + struct sk_buff *frag2; if (first_len - hlen > mtu || ((first_len - hlen) & 7) || @@ -651,18 +651,18 @@ static int ip6_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *)) if (frag->len > mtu || ((frag->len & 7) && frag->next) || skb_headroom(frag) < hlen) - goto slow_path; + goto slow_path_undo; /* Partially cloned skb? */ if (skb_shared(frag)) - goto slow_path; + goto slow_path_undo; BUG_ON(frag->sk); if (skb->sk) { frag->sk = skb->sk; frag->destructor = sock_wfree; - truesizes += frag->truesize; } + skb->truesize -= frag->truesize; } err = 0; @@ -693,7 +693,6 @@ static int ip6_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *)) first_len = skb_pagelen(skb); skb->data_len = first_len - skb_headlen(skb); - skb->truesize -= truesizes; skb->len = first_len; ipv6_hdr(skb)->payload_len = htons(first_len - sizeof(struct ipv6hdr)); @@ -756,6 +755,15 @@ static int ip6_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *)) IPSTATS_MIB_FRAGFAILS); dst_release(&rt->dst); return err; + +slow_path_undo: + skb_walk_frags(skb, frag2) { + if (frag2 == frag) + break; + frag2->sk = NULL; + frag2->destructor = NULL; + skb->truesize += frag2->truesize; + } } slow_path:
-- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |