[Vortex] Question regarding sequence number calculations

Phil Black-Knight pblack at tresys.com
Wed Nov 28 19:24:37 CET 2012


I'm trying to understand something in the vortex code, namely how sequence numbers are calculated when they overflow.

>From the BEEP RFC, the max seq number is 4294967295 (2^32 - 1), and that arithmetic needs to be done modulo 4294967296 (2^32).

I see in the vortex code that there are defines for both of these (MAX_SEQ_NO & MAX_SEQ_MOD) in vortex_types.h, but MAX_SEQ_MOD is never used.

In the code, there are a number of places where MAX_SEQ_NO is used and  a few places where it's used to perform modulus arithmetic.
Of particular concern (though admittedly, I'm no BEEP expect and I've only traced a little of the vortex code) is around line 2230 of vortex_channel.c, from the 1.1.12 release:

channel->last_seq_no = ((channel->last_seq_no + frame_size) % (MAX_SEQ_NO));

I think this incorrectly calculate the next seq number.  i.e. if the last_seq_no was 4294967295, and the frame size was 10, then (according to the RFC) last_seq_no should end up as 9, while this code would assign it to 10.

Similarly at line 2299 of the same file:

channel->last_seq_no_expected   = ((channel->last_seq_no_expected + frame_size) % (MAX_SEQ_NO)

Like I said, I don't really know the codebase or protocol all too well, so I'm a bit wary of making any modifications to the code base.

-Phil
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.aspl.es/pipermail/vortex/attachments/20121128/14e5509d/attachment.html>


More information about the Vortex mailing list