netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* skb_try_coalesce and fraglists (bug?)
@ 2015-02-04 16:33 Erik Hugne
  2015-02-04 17:28 ` Eric Dumazet
  0 siblings, 1 reply; 2+ messages in thread
From: Erik Hugne @ 2015-02-04 16:33 UTC (permalink / raw)
  To: netdev; +Cc: mkubecek, ying.xue

skb_try_coalesce should bail out for a number of reasons, if the target skb is
cloned, doesn't have enough room, or if either source or target skb have
fraglists.

However, it seems that the skb_has_frag_list check is done too late, and a small
skb may be copied into the tailroom of the head, even if it has a fraglist.

Wouldn't this be more correct?

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 56db472..8d02140 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -4056,6 +4056,9 @@ bool skb_try_coalesce(struct sk_buff *to, struct sk_buff *from,
        if (skb_cloned(to))
                return false;
 
+       if (skb_has_frag_list(to) || skb_has_frag_list(from))
+               return false;
+
        if (len <= skb_tailroom(to)) {
                if (len)
                        BUG_ON(skb_copy_bits(from, 0, skb_put(to, len), len));
@@ -4063,9 +4066,6 @@ bool skb_try_coalesce(struct sk_buff *to, struct sk_buff *from,
                return true;
        }
 
-       if (skb_has_frag_list(to) || skb_has_frag_list(from))
-               return false;
-
        if (skb_headlen(from) != 0) {
                struct page *page;
                unsigned int offset;

^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: skb_try_coalesce and fraglists (bug?)
  2015-02-04 16:33 skb_try_coalesce and fraglists (bug?) Erik Hugne
@ 2015-02-04 17:28 ` Eric Dumazet
  0 siblings, 0 replies; 2+ messages in thread
From: Eric Dumazet @ 2015-02-04 17:28 UTC (permalink / raw)
  To: Erik Hugne; +Cc: netdev, mkubecek, ying.xue

On Wed, 2015-02-04 at 17:33 +0100, Erik Hugne wrote:
> skb_try_coalesce should bail out for a number of reasons, if the target skb is
> cloned, doesn't have enough room, or if either source or target skb have
> fraglists.
> 
> However, it seems that the skb_has_frag_list check is done too late, and a small
> skb may be copied into the tailroom of the head, even if it has a fraglist.
> 
> Wouldn't this be more correct?
> 
> diff --git a/net/core/skbuff.c b/net/core/skbuff.c
> index 56db472..8d02140 100644
> --- a/net/core/skbuff.c
> +++ b/net/core/skbuff.c
> @@ -4056,6 +4056,9 @@ bool skb_try_coalesce(struct sk_buff *to, struct sk_buff *from,
>         if (skb_cloned(to))
>                 return false;
>  
> +       if (skb_has_frag_list(to) || skb_has_frag_list(from))
> +               return false;
> +
>         if (len <= skb_tailroom(to)) {
>                 if (len)
>                         BUG_ON(skb_copy_bits(from, 0, skb_put(to, len), len));
> @@ -4063,9 +4066,6 @@ bool skb_try_coalesce(struct sk_buff *to, struct sk_buff *from,
>                 return true;
>         }
>  
> -       if (skb_has_frag_list(to) || skb_has_frag_list(from))
> -               return false;
> -
>         if (skb_headlen(from) != 0) {
>                 struct page *page;
>                 unsigned int offset;
> 

Patch is not needed.

If to had a fraglist, skb_tailroom(to) would be 0

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2015-02-04 17:28 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-02-04 16:33 skb_try_coalesce and fraglists (bug?) Erik Hugne
2015-02-04 17:28 ` Eric Dumazet

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).