ipv6: ip6_fragment: fix headroom tests and skb leak
Subsystems (3)
PATCH 1/1 ipv6: ip6_fragment: fix headroom tests and skb leak
David Woodhouse reports skb_under_panic when we try to push ethernet header to fragmented ipv6 skbs:
skbuff: skb_under_panic: text:c1277f1e len:1294 put:14 head:dec98000 data:dec97ffc tail:0xdec9850a end:0xdec98f40 dev:br-lan
[..]ip6_finish_output2+0x196/0x4da
David further debugged this:
[..] offending fragments were arriving here with skb_headroom(skb)==10. Which is reasonable, being the Solos ADSL card's header of 8 bytes followed by 2 bytes of PPP frame type.
The problem is that if netfilter ipv6 defragmentation is used, skb_cow() in ip6_forward will only see reassembled skb.
Therefore, headroom is overestimated by 8 bytes (we pulled fragment header) and we don't check the skbs in the frag_list either.
We can't do these checks in netfilter defrag since outdev isn't known yet.
Furthermore, existing tests in ip6_fragment did not consider the fragment or ipv6 header size when checking headroom of the fraglist skbs.
While at it, also fix a skb leak on memory allocation -- ip6_fragment must consume the skb.
I tested this e1000 driver hacked to not allocate additional headroom (we end up in slowpath, since LL_RESERVED_SPACE is 16).
If 2 bytes of headroom are allocated, fastpath is taken (14 byte ethernet header was pulled, so 16 byte headroom available in all fragments).
--- net/ipv6/ip6_output.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-)
net/ipv6/ip6_output.c +8 -6
@@ -586,20 +586,22 @@ int ip6_fragment(struct sock *sk, struct sk_buff *skb, frag_id =ipv6_select_ident (net, &ipv6_hdr (skb)->daddr, &ipv6_hdr (skb)->saddr); + hroom =LL_RESERVED_SPACE (rt->dst.dev);if (skb_has_frag_list (skb)) {int first_len =skb_pagelen (skb);struct sk_buff *frag2;if (first_len - hlen > mtu || ((first_len - hlen) &7 ) || -skb_cloned (skb)) +skb_cloned (skb) || +skb_headroom (skb) < (hroom +sizeof (struct frag_hdr ))) goto slow_path; skb_walk_frags(skb, frag) {/* Correct geometry. */ if (frag->len > mtu || ((frag->len &7 ) && frag->next) || -skb_headroom (frag) < hlen) +skb_headroom (frag) < (hlen + hroom +sizeof (struct frag_hdr ))) goto slow_path_clean;/* Partially cloned skb? */
@@ -616,8 +618,6 @@ int ip6_fragment(struct sock *sk, struct sk_buff *skb, err =0 ; offset =0 ; - frag =skb_shinfo (skb)->frag_list; -skb_frag_list_init (skb);/* BUILD HEADER */ *prevhdr =NEXTHDR_FRAGMENT ;
@@ -625,8 +625,11 @@ int ip6_fragment(struct sock *sk, struct sk_buff *skb,if (!tmp_hdr) {IP6_INC_STATS (net,ip6_dst_idev (skb_dst (skb)),IPSTATS_MIB_FRAGFAILS ); -return -ENOMEM ; + err = -ENOMEM ; + goto fail; } + frag =skb_shinfo (skb)->frag_list; +skb_frag_list_init (skb);__skb_pull (skb, hlen); fh = (struct frag_hdr *)__skb_push (skb,sizeof (struct frag_hdr ));
@@ -723,7 +726,6 @@ slow_path:*/ *prevhdr =NEXTHDR_FRAGMENT ; - hroom =LL_RESERVED_SPACE (rt->dst.dev); troom = rt->dst.dev->needed_tailroom;/*
Signature · 1 line
2.0.5
Quoted 2 lines · Florian Westphal
I tested this e1000 driver hacked to not allocate additional headroom (we end up in slowpath, since
LL_RESERVED_SPACEis 16).
And it works on the originally-offending setup too; thanks.
Quoted 2 lines
Reported-by David Woodhouse <[email protected]>Diagnosed-by David Woodhouse <[email protected]>
They generally prefer me to use @intel.com for those too, if you would. I draw the line at using it for actual email communication though :)
Signature · 2 lines
David Woodhouse Open Source Technology Centre [email protected] Intel Corporation
From: Florian Westphal <[email protected]>
Date: Wed, 16 Sep 2015 17:26:14 +0200
Quoted 30 lines
David Woodhouse reports
skb_under_panicwhen we try to push ethernet header to fragmented ipv6 skbs:skbuff:
skb_under_panic: text:c1277f1e len:1294 put:14 head:dec98000 data:dec97ffc tail:0xdec9850a end:0xdec98f40 dev:br-lan
[..]ip6_finish_output2+0x196/0x4daDavid further debugged this:
[..] offending fragments were arriving here with skb_headroom(skb)==10. Which is reasonable, being the Solos ADSL card's header of 8 bytes followed by 2 bytes of PPP frame type.The problem is that if netfilter ipv6 defragmentation is used,
skb_cow()inip6_forwardwill only see reassembled skb.Therefore, headroom is overestimated by 8 bytes (we pulled fragment header) and we don't check the skbs in the
frag_listeither.We can't do these checks in netfilter defrag since outdev isn't known yet.
Furthermore, existing tests in
ip6_fragmentdid not consider the fragment or ipv6 header size when checking headroom of the fraglist skbs.While at it, also fix a skb leak on memory allocation --
ip6_fragmentmust consume the skb.I tested this e1000 driver hacked to not allocate additional headroom (we end up in slowpath, since
LL_RESERVED_SPACEis 16).
Quoted 7 lines
If 2 bytes of headroom are allocated, fastpath is taken (14 byte ethernet header was pulled, so 16 byte headroom available in all fragments).
Reported-by David Woodhouse <[email protected]>Diagnosed-by David Woodhouse <[email protected]>Signed-off-by Florian Westphal <[email protected]>
Applied, thanks.