Ricky Beam <jfbeam@bluetopia.net> said:
> On Wed, 12 Apr 2000, Michael H. Warfield wrote:
> > Are you sure about that? For some reason, I was under the impression
> >that a special file (block or char device) didn't occupy any FS blocks. I
> Sorry for the confusion... I'm refering to drive space in general. There
> are two types of space in a file system... inodes and data. Special files
> don't take any data space (with the possible exception of symlinks); they
> only waste an inode with is one FS block -- the '-b' part of mke2fs. They
> better not be using any data space -- they don't contain any data.
But they do! Device identity, owner, group, permissions; soon ACLs and
other stuff in the making. Here I count 116 bytes per inode for
ext2. Sure, the space for block pointers is wasted, but I don't want to
even think of the hair in the kernel and on disk special small device
inodes would bring (hair the sysadmin looses trying to set up all this, if
I may add).
[...]
> > Now if there were only a CLEAN solution to the non-volatile
> >attributes problem. Set permissions and ownership of a device. Device
> If you define "clean solution" as '/usr/bin/vi'... either alter the
> settings in the source or reset them everytime it's loaded.
I wouldn't want to have to set permissions on my files by editing a file,
which can (and Murphy assures us, _will_) get lost or corrupted. And reset
it to what, by whom? What if the machine is running low on space, so the
chown/chmod can't run (or is plain overloaded, so it runs slowly), and some
attacker has a field day with leftover default permissions?
-- Horst von Brand vonbrand@sleipnir.valparaiso.cl Casilla 9G, Viņa del Mar, Chile +56 32 672616- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Apr 15 2000 - 21:00:20 EST