home *** CD-ROM | disk | FTP | other *** search
- Date: Wed, 20 Jul 94 18:33 MET DST
- From: knarf@nasim.cube.net (Frank Bartels)
- To: mint@atari.archive.umich.edu
- Subject: Re: /dev/kmem
- Newsgroups: nasim.lists.mint
- Organization: The Sunsite for ATARI-Friends
- X-Newsreader: TIN [version 1.2 PL2]
-
- In article <9407200826.AA02175@hpbeo79.bbn.hp.com> you wrote:
-
- > I think one advantage is that the kernel has more control over what it
- > publishes in /proc, and that the /proc device is visible to the user
- > for simple manipulations such as in the MultiTOS desktop.
-
- I think there is no reason to remove /proc after /dev/kmem is implemented.
-
- > I'm not sure, however, what a /dev/kmem offers. If it opens up the
- > complete kernel address space to other processes, I would definitely
- > prefer the /proc approach since it seems safer to me.
-
- /dev/kmem is usually not world readable, it's only readable bye the
- kmem-group. Thinks like ps and top are sgid-kmem (SunOS 4):
-
- -rwxr-sr-x 1 root kmem 36040 Oct 14 1990 /bin/ps*
- crw-r----- 1 root kmem 3, 1 Jan 25 23:42 /dev/kmem
- -rwxr-sr-x 1 root kmem 49152 Feb 2 02:08 /usr/local/bin/top*
-
- BTW: The same feature is used here for df and the group operator:
-
- -rwxr-sr-x 1 root operator 7904 Oct 14 1990 /bin/df*
- crw-r----- 1 root operator 17, 2 Jan 25 23:42 /dev/rsd0c
-
- Bye,
- Knarf
- --
- Frank Bartels | UUCP/ZModem: + 49 89 5469593 | MiNT is
- knarf@nasim.cube.net | Login: nuucp Index: /pub/ls-lR.nasim.gz | Now TOS!
-