home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.linux
- Path: sparky!uunet!cis.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!gatech!destroyer!ubc-cs!news.UVic.CA!sanjuan!pmacdona
- From: pmacdona@sanjuan (Peter MacDonald)
- Subject: SLS typo, and an antecedent to device drivers
- Message-ID: <1992Aug15.223104.18515@sol.UVic.CA>
- Sender: news@sol.UVic.CA
- Nntp-Posting-Host: sanjuan.uvic.ca
- Organization: University of Victoria, Victoria, BC, CANADA
- Date: Sat, 15 Aug 92 22:31:04 GMT
- Lines: 39
-
- Oops. Forgot to point out the typo in my original SLS post.
- The disk sizes should have read
-
- Minimal: 6 Meg.
- Base: 20 Meg.
- Base+X 30 Meg.
-
- On another note, the problem with device drivers, is that
- (as someone else pointed out) the extra complexity they pose
- is not offset by an attendant gain in general functionality,
- at least in some of our opinions. Of course if someone were
- to go ahead and do it (like how about those who most vocally
- proclaim them), and adapted existing devices to them, I doubt
- Linus would turn the difs away :-)
-
- But hold on here, I am not just rambling, and yes there is a
- point. A much easier task would be to modify the kernel
- slightly to allow some of the compile time options to be
- selected at run time instead, ala rootdev. Like for instance:
-
- keyboard type
- startup video mode
- ramdisk or no
- # of pty's
- # of VC's
- max buffer size
- swap ?
-
- Perhaps, the word or two after the rootdev byte, could contain
- an offset to a static data table in the kernel (or is there
- enough room to just put it down there (Linus?). Then a
- rootdev like program could be used to set these options
- and one kernel could work for all (except those wising to
- turf SCSI code etc. They would still have to recompile).
-
- I think I know already how to do most of this, but with SLS
- am just to busy now. The line has been drawn :-).
-
- Peter.
-