home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.next.software
- Path: sparky!uunet!decwrl!access.usask.ca!kakwa.ucs.ualberta.ca!news
- From: sherwood@fenris.space.ualberta.ca (Sherwood Botsford)
- Subject: Re: Mach and nbuf.
- Message-ID: <1992Sep9.214538.1146@kakwa.ucs.ualberta.ca>
- Sender: news@kakwa.ucs.ualberta.ca
- Nntp-Posting-Host: fenris.space.ualberta.ca
- Organization: University Of Alberta, Edmonton Canada
- References: <1992Aug28.034938.13638@access.usask.ca>
- Distribution: USA
- Date: Wed, 9 Sep 1992 21:45:38 GMT
- Lines: 28
-
- jake@skatter.usask.ca writes
- > I followed the thread about changing the kernel parameter
- 'nbuf' with
- > some interest, and probably incomplete understanding. I
- decided to
- > alter the number of buffers just to see what sort of
- difference it would
- > make. I used the bsd() sdmach nbuf=NNN command, rather
- than hack the
- > kernel. I understood that the command above would have
- to be re-entered
- > at every reboot, while the purpose of the hack was to
- make the change
- > permanent. However, I rebooted the machine to undo the
- change in
- > nbuf and to my suprise it came back up with the altered
- number of buffers.
- >
- > I mention this 1) to find out more about this behavior,
- and 2) to say that
- > doing it this way seems a lot easier than hacking the
- kernel.
- >
-
- The change only has to be re-entered if you power down the
- machine. The EPROM remembers the boot command last used,
- but is limited to a fairly short string, and a change in
- nbuf makes the string too long.
-