skbuff: in skb_copy_ubufs unclone before releasing zerocopy
Subsystems (2)
PATCH 1/1 skbuff: in skb_copy_ubufs unclone before releasing zerocopy
From: Willem de Bruijn <[email protected]>
skb_copy_ubufs must unclone before it is safe to modify its skb_shared_info with skb_zcopy_clear.
Commit b90ddd568792 ("skbuff: skb_copy_ubufs must release uarg even without user frags") ensures that all skbs release their zerocopy state, even those without frags.
But I forgot an edge case where such an skb arrives that is cloned.
The stack does not build such packets. Vhost/tun skbs have their frags orphaned before cloning. TCP skbs only attach zerocopy state when a frag is added.
But if TCP packets can be trimmed or linearized, this might occur. Tracing the code I found no instance so far (e.g., skb_linearize ends up calling skb_zcopy_clear if !skb->data_len).
Still, it is non-obvious that no path exists. And it is fragile to rely on this.
---
I should have spotted this when preparing the above patch, of course. Apologies for the added work.
Since I did not spot a path that triggers the issue, this can also go to net-next, instead. It applies to both cleanly.
--- net/core/skbuff.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
net/core/skbuff.c +3 -3
@@ -1177,12 +1177,12 @@ int skb_copy_ubufs(struct sk_buff *skb, gfp_t gfp_mask)int i, new_frags;u32 d_off; -if (!num_frags) - goto release; -if (skb_shared (skb) ||skb_unclone (skb, gfp_mask))return -EINVAL ; +if (!num_frags) + goto release; + new_frags = (__skb_pagelen (skb) +PAGE_SIZE -1 ) >>PAGE_SHIFT ;for (i =0 ; i < new_frags; i++) { page =alloc_page (gfp_mask);
Signature · 1 line
2.15.1.620.gb9897f4670-goog
David Miller
maintainerFrom: Willem de Bruijn <[email protected]>
Date: Thu, 28 Dec 2017 12:38:13 -0500
Quoted 19 lines
From: Willem de Bruijn <[email protected]>
skb_copy_ubufsmust unclone before it is safe to modify itsskb_shared_infowithskb_zcopy_clear.Commit
b90ddd568792("skbuff:skb_copy_ubufsmust release uarg even without user frags") ensures that all skbs release their zerocopy state, even those without frags.But I forgot an edge case where such an skb arrives that is cloned.
The stack does not build such packets. Vhost/tun skbs have their frags orphaned before cloning. TCP skbs only attach zerocopy state when a frag is added.
But if TCP packets can be trimmed or linearized, this might occur. Tracing the code I found no instance so far (e.g.,
skb_linearizeends up callingskb_zcopy_clearif !skb->data_len).
Quoted 5 lines
Still, it is non-obvious that no path exists. And it is fragile to rely on this.
Fixes b90ddd568792 ("skbuff: skb_copy_ubufs must release uarg even without user frags")Signed-off-by Willem de Bruijn <[email protected]>
Applied and queued up for -stable.