>Number:         5483
>Category:       port-sun3
>Synopsis:       le(4) does not document LANCE_REVC_BUG
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    gnats-admin (GNATS administrator)
>State:          open
>Class:          doc-bug
>Submitter-Id:   net
>Arrival-Date:   Fri May 22 03:35:00 1998
>Last-Modified:
>Originator:     Erik E. Fair
>Organization:
International Organization of Internet Clock Watchers
>Release:        NetBSD-current 19980522
>Environment:
	
System: NetBSD digital.clock.org 1.3 NetBSD 1.3 (GENERIC) #0: Thu Jan 1 18:26:28 MET 1998 pk@flambard:/usr/src1/sys/arch/sparc/compile/GENERIC sparc


>Description:
	The AMD 7990 revision C Ethernet controller chip has a bug; it does
	not reliably identify packets bound for the local Ethernet station.

	This is worked around in software by double checking the Ethernet
	destination station address in the driver when the kernel is compiled
	withe LANCE_REVC_BUG option. This option is on by default in GENERIC
	for NetBSD/sun3

	This is costly in cycles on a per-packet basis, and has implications
	for multicast support.

	This should be explicitly documented in the driver's manual page.

	Further, it should be documented in the manual page how one might try
	to identify the Revision "C" chip, so that a user can choose to go
	look, and see if s/he has the buggy chip or not, and tweak the
	LANCE_REVC_BUG kernel option appropriately. How do you identify the
	revision level of this chip?

>How-To-Repeat:
	man 4 le on sun3
>Fix:
	Text to be written.
>Audit-Trail:
>Unformatted: