> So is "the current time" the time the packet was sent? If so, the best > that Ethereal can do is use the time the packet was *received* as the > time base. > > I was going to look to see if the rx header contains a timestamp. If so, > that would be the best base to use. Otherwise, the time the packet was > received would be fine. It's not critical.
the rx header does not contain a timestamp. The cache manager interprets callback expiration times as relative to the time the response (containing the attrs) was handled. That is one reason why the file server tacks on an extra slush period in its callback expiration logic. > An alternative would be to report it as a relative time rather than > convert it to an absolute time. > > This could be the best solution of all. Those of you who use afs, which > would you rather see? relative time more closely approximates its usage by the cache manager, and is probably a little bit more useful. _______________________________________________ OpenAFS-devel mailing list [EMAIL PROTECTED] https://lists.openafs.org/mailman/listinfo/openafs-devel
