home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.vxworks
- Path: sparky!uunet!wupost!csus.edu!netcom.com!hjb
- From: hjb@netcom.com (H. J. Bae)
- Subject: Re: Forwarding of MAC broadcasts
- Message-ID: <1993Jan10.140415.1003@netcom.com>
- Organization: Peaceful Star, Inc.
- References: <9301071938.AA09078@kuwait.whizkids> <C0MBsq.EBL@waterloo.hp.com>
- Date: Sun, 10 Jan 1993 14:04:15 GMT
- Lines: 30
-
- In article <C0MBsq.EBL@waterloo.hp.com> rypma@waterloo.hp.com (Ted Rypma) writes:
- >I just read Hwa Jin's response to this and I think I have a glimmering of
- >an explaination.
- >
- >Note that the _MAC_ address is a broadcast and the _IP_ address is not.
- >Now, one of the (many) failings of BSD 4.3 network code is the fact that
- >the "broadcastness" of the incoming frame is lost once you get into IP
- >protocol code. If the IP address is not broadcast, whether legitimately so
- >or as a result of a misunderstood subnet broadcast (due to incorrect or no
- >subnet mask), IP will forward if so configured. Of course, it REALLY should
- >NOT be forwarded because the MAC address was broadcast - information that
- >BSD 4.3 Tahoe quietly throws away.
- >
- >BSD Networking Release 2 code nicely fixes this (I'm not sure whether Reno
- >code does or not, but it may).
-
-
- as always, ted is quite correct about this. [ah... so that's what
- that datalink bcast flags/stuff was all about in net2 code, silly me]
- of course, if things are done "correctly" (correct subnet masks,
- correct ip broadcast addr derived from subnet mask, etc.) the problem
- should not occur. thanks.
-
-
- hwajin
-
- --
- Peaceful Star Project hjb@netcom.com
- 2899 Ford Street Available for peaceful, responsible work only.
- Oakland, CA 94601 (510) 536-7607
-