[bugzilla-daemon@xxxxxxxxxxxxx]

Comment # 2 on bug 9762 from David Howells

(In reply to comment #1)

> Do you have a link to a spec ?

Yes...  But it's not very clear:

    http://web.mit.edu/kolya/afs/rx/rx-spec

since that talks in terms of "Maximum Packet Size" and "Recommended Packet
Size".

>From line 329:

       The trailing fields after the variable-length acknowledgements
       section are not always 32-bit aligned with respect to the packet,
       and aren't always present.  (Their presence depends on the Rx
       version of the peer.)  The maximum and recommended packet sizes
       are, respectively, the largest possible packet size that the peer
       is willing to accept from us, and the size of the packet they
       would prefer to receive.  In absence of these fields, it should
       be assumed that the maximum allowed packet size is 1444 bytes.

There's also a diagram at line 867.

OpenAFS's librx puts the adjusted maximum MTU size first, then the actual
interface MTU:

http://git.openafs.org/?p=openafs.git;a=blob;f=src/rx/rx.c;h=0c05763755ae7725e243f003daef82d555e844cd;hb=b7326e487f6387033282cc5d1128b00a0456673a#l5649

---

      You are receiving this mail because:

• You are watching all bug changes.