This bug was not completely fixed but there is need also this patch[1].
Further testing has revealed a couple more problems
with the packet reordering/buffering code.
1) Some clients (notably the PPTP client) start their sequence
numbers at 1 instead of 0 as the RFC mandates. My previous
fix caused problems with these clients.
2) Duplicate packets were causing corruption when they were
placed on the queue but never used -or- when they were
placed on the queue but already existed on the queue (i.e.
they previously arrived out of order).
This bug was not completely fixed but there is need also this patch[1].
Further testing has revealed a couple more problems buffering code.
with the packet reordering/
1) Some clients (notably the PPTP client) start their sequence
numbers at 1 instead of 0 as the RFC mandates. My previous
fix caused problems with these clients.
2) Duplicate packets were causing corruption when they were
placed on the queue but never used -or- when they were
placed on the queue but already existed on the queue (i.e.
they previously arrived out of order).
It was already applied to cvs[2].
[1]http:// marc.info/ ?l=poptop- server& m=1177374534005 88&w=2 marc.info/ ?l=poptop- server& m=1177384424174 75&w=2
[2]http://