Yes. I pull out my copy of Stevens last night and learned that 65535
is the total packet size, headers+payload.
> On Sun, Dec 13, 1998 at 01:50:25PM +1100, Richard Gooch wrote:
> > Linux 2.0.36 -> Solaris 2.5.1 65507
> >
> > Linux 2.1.126 -> Linux 2.1.126 65476
> > Linux 2.1.126 -> Solaris 2.5.1 65507
> >
> > Linux 2.1.131 -> Linux 2.1.131 65476
> > Linux 2.1.131 -> Solaris 2.5.1 65507
> >
> > Solaris 2.5.1 -> Linux 2.0.36 51792
> > Solaris 2.5.1 -> Linux 2.1.126 65476
> > Solaris 2.5.1 -> Linux 2.1.131 65476
> >
> > Solaris 2.5.1 -> Solaris 2.5.1 65535 (what you expect)
> >
> > From this it appears Linux 2.1.x has an output limit of 65507 bytes
> > and an input limit of 65476 bytes. Further, Linux 2.0.36 has an output
> > limit of 65507 bytes (same as 2.1.x) and an input limit of 51792
> > bytes.
>
> Looks like Linux gets the transmit size right, and the receive size
> wrong.
Yep.
> > Packets which exceed the maximum size go to the big bit bucket in the
> > sky. RIP. This is a problem for programmes which "know" that they can
> > send 65535 byte packets over UDP.
>
> 65535? Do you mean including the UDP and IP header, or 65535 bytes of
> application data.
I (incorrectly) meant 65535 bytes of application data.
Regards,
Richard....
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/