>Number:         1663
>Category:       kern
>Synopsis:       times returned by bpf are sometimes wrong
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people (Kernel Bug People)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Oct 20 16:05:01 1995
>Last-Modified:
>Originator:     dave archer
>Organization:
"	LIFO"
>Release:        1.0
>Environment:
System: NetBSD sync.cs.buffalo.edu 1.0A NetBSD 1.0A (SYNC) #0: Wed Mar 22 19:13:40 EST 1995 dmarcher@sync.cs.buffalo.edu:/u/src/netbsd-current/sys/arch/i386/compile/SYNC i386


>Description:
	sometimes the times reported by bpf are off by 10,000us.  
>How-To-Repeat:
	run tcpdump while lots of back-to-back packets are on the wire,
	watch long enough and time will go backwards.
>Fix:
	none known, although i suspect the optimizations in the i386's 
	microtime.s are at fault.  time does not seem to ever go backwards
	from a user-land program calling gettimeofday, however.
>Audit-Trail:
>Unformatted: