Re: Data corruption issue with splice() on 2.6.27.10

[Jarek Poplawski]

On Wed, Jan 07, 2009 at 11:24:07PM +1100, Herbert Xu wrote:
> On Wed, Jan 07, 2009 at 01:22:05PM +0100, Willy Tarreau wrote:
> >
> > > diff --git a/net/core/skbuff.c b/net/core/skbuff.c
> > > index 5110b35..4c080cd 100644
> > > --- a/net/core/skbuff.c
> > > +++ b/net/core/skbuff.c
> > > @@ -73,17 +73,13 @@ static struct kmem_cache *skbuff_fclone_cache __read_mostly;
> > >  static void sock_pipe_buf_release(struct pipe_inode_info *pipe,
> > >  				  struct pipe_buffer *buf)
> > >  {
> > > -	struct sk_buff *skb = (struct sk_buff *) buf->private;
> > > -
> > > -	kfree_skb(skb);
> > > +	put_page(buf->page);
> > >  }
> > >  
> > >  static void sock_pipe_buf_get(struct pipe_inode_info *pipe,
> > >  				struct pipe_buffer *buf)
> > >  {
> > > -	struct sk_buff *skb = (struct sk_buff *) buf->private;
> > > -
> > > -	skb_get(skb);
> > > +	get_page(buf->page);
> > >  }
> 
> Well this patch can only make it worse because not only are you
> still ref counting skb->head with get_page, but you've also
> completely removed the skb ref count which means that the corruption
> can only occur sooner.

But we don't need this skb... except its ->frags[]  pages, which are
get_paged?! (The rest is copied to new pages.)

Jarek P.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/