Re: #540: "jumbo" frames

In message <CAP+FsNfKBazaKL4a8PE8mMtCdFhKj1tyXQ-6twAF=rW_X9Ehzg@mail.gmail.com>
, Roberto Peon writes:

>I don't see how CONTINUATION makes a hardware implementation more
>difficult, though?
>It doesn't at all change the buffer management.

It sure does:  The "magic" blocking of CONTINUATION frames into one
unit where the END_* bits don't actually mean END_* at all will
seriously complicate a hardware transmit-side implementation.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

Received on Wednesday, 25 June 2014 20:28:02 UTC