Re: Are "skb->data" physically continuous?

[Nick Patavalis (npat@inaccessnetworks.com)]

On Mon, Sep 15, 2003 at 04:05:52PM +0300, Shmulik Hen wrote:
> 
> >     NETIF_F_FRAGLIST   <------------------- ??? WHAT IS THIS ???
> >         Scatter/gather IO.
> 
> That's for adapters that can follow a chain of skb's as well as scan 
> through the array of fragments in each skb (which is limited by 
> MAX_SKB_FRAGS). Not sure which level of the stack actually uses that 
> feature, but it may be usefull for passing a list of skb's, all 
> belonging to a single "packet".
> 

Af far as I was able to find-out by chaching pointers in the kernel
sources, "frag_list"s seem to be only used for upsteam packets
(received packets). Furthermore the only devices specifying the
NETIF_F_FRAGLIST feature are the loopback device and ieee1394. Both of
them don't seem to use the "frag_list"s that much!

Elsewhere (in a mailing list) I saw mentioned that frag-lists are used
when the maximum number of fragments in an skb is not enough for a
packet, so a second "skb" is allocated of which only "skb_shared_info"
is actually used to keep the extra fragments (not "skb->base"). This
sounds a bit strange, though (why waste a whole "skb" just to keep a
few extra fragment descriptors)!

Does anybody know what's the case with "frag_list"s?

/npat

-- 
When it is incorrect, it is, at least *authoritatively* incorrect
  -- Douglas Adams, The Hitchhiker's Guide to the Galaxy
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html