*BSD News Article 18949


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!agate!agate.berkeley.edu!cgd
From: cgd@erewhon.CS.Berkeley.EDU (Chris G. Demetriou)
Newsgroups: comp.os.386bsd.questions
Subject: Re: AUI with ed0 and 3c503 8-bit
Date: 27 Jul 93 16:39:39
Organization: Kernel Hackers 'r' Us
Lines: 34
Message-ID: <CGD.93Jul27163939@erewhon.CS.Berkeley.EDU>
References: <233rnr$8oh@terminator.rs.itd.umich.edu>
	<CGD.93Jul27115107@erewhon.cs.berkeley.edu>
	<1993Jul27.195636.23569@emba.uvm.edu>
NNTP-Posting-Host: erewhon.cs.berkeley.edu
In-reply-to: wollman@UVM.EDU's message of Tue, 27 Jul 1993 19:56:36 GMT

In article <1993Jul27.195636.23569@emba.uvm.edu> wollman@UVM.EDU (Garrett Wollman) writes:
>>                     so many bugs, so little time.
>
>And this is one of them.  The IFF_LLC0 flag should never have been
>stolen for this purpose, and should not be used in this manner.

if.h says:

#define IFF_LLC0        0x1000          /* interface driver control/status */
#define IFF_LLC1        0x2000          /* interface driver control/status */
#define IFF_LLC2        0x4000          /* interface driver control/status */

well, this sure seems like "interface driver control" or "interface
driver status" to me...

what do *you* think the LLC flags should be used for?

>The
>right thing to do is to define a new if_flags value (and while you're
>at it change if_flags to be a u_int, not a u_short) and test /that/
>in the drivers.

i don't agree; the LLC flags are a per-driver "link-level" control
flag.  if you don't want to use them for per-driver link-level control
well, then...




chris
--
Chris G. Demetriou                                   cgd@cs.berkeley.edu

                     so many bugs, so little time.