Trying to get back into the kernel networking game :)
"A.N.Kuznetsov" <kuznet@ms2.inr.ac.ru> writes:
>: The problem is (2) requires quite a bit of work and I never got around to
>: doing it. But I do think it's the Right Way.
>
>I believe it's right way. At least in 2.1.79 the attempt
>to invert this order (only for link layer headers) is made and
>I see no reasons why not to complete this inversion.
>
>The only problem is that we must allocate skb with necessary room
>at head now. SACKS (and extension headers) would require to overincrease
>MAX_HEADER space.
As I recall, changes to MAX_HEADER used imply nasty side effects on
broken network drivers. I hope this issue is dead for a while now though.
>: Or am I horribly wrong? Also is anybody working on the SACK?
>
>Dave said something about SACKS, but I do not know details.
>Andy, I suspect you already did it. I mean not SACKS, but inversion
>of header filling order. IPv6 exthdrs seems to be much more natural
>in this case.
About SACK, I was working on it before I disappeared for 5 months to
start my life over again. I'm back on-line now and hoping to contribute
to getting the TCP code up to speed again. As has been noted I had
started making room for SACK in the code but did not finish.
I have a little bit more code sitting around in the wings, but I'm
not sure how well it fits with the changes that have happened to the
kernel lately. I'd be happy to talk this over if anyone wants to pitch in
or take over on the SACK front.
About the header construction, Andi perhaps you could say if you did
invert the filling order (ala BSD) to construct the header just before
sending out the packet? If it's not to bad as far as the sockbuf
stuff goes this might give us the last little bit needed to be able to
do SACK somewhat cleanly.
Also, while I'm doing scatter shot questions. Alan, you mentioned some
problems with TCP window scaling and such. Can you be more specific?
I got the code in but never really got much testing feedback before I
disappeared, and haven't really heard anything since. I'll take a knife
to the code anyway and see what it's doing, but if you could point me
at a known problem, or something that should specifically be fixed it
would be a bit faster.
Cheers,
--
Eric Schenk www: http://www.loonie.net/~eschenk
email: eschenk@loonie.net, eschenk@rogers.wave.ca
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu