Implementing a make command

Simon Fowler simon at
Sun Feb 17 15:36:20 EST 2002

On Sun, Feb 17, 2002 at 03:04:57PM +1100, Bob Edwards wrote:
> Might be better off looking at the minix mount command's source code and
> also at the kernel code that implements the minix filesystem - it is about
> as simple a file system as you can get.
The filesystem itself is simple, but the implementation has gotten
rather more complex - it's been redone to use the gory details of
the block layer, which makes working out what's going on with the
VFS code a lot harder. 

As for minix's mount, I believe it uses the normal mount . . .

> There is also a effort underway to simplify user-land file systems (I'm
> not sure how this is implemented, but it has something to do with the
> kernel calling back into a user-land daemon to do the FS work - would be
> useful for an FTP file system, or a HTTP file system etc.).
There are three projects working on userspace filesystem interfaces
that I know of, as well as one that I'm writing myself: the
granddaddy of them all is the coda kernel module, which one of the
projects is based on (uservfs, used to be called podfuk). The other
one is called fuse, which seems cleaner to me, but less portable -
you need a kernel module, rather than using the coda interface,
which has been ported just about everywhere . . .


PGP public key Id 0x144A991C, or
(crappy) Homepage:
doe #237 (see 
My DeCSS mirror: 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 232 bytes
Desc: not available
Url :

More information about the linux mailing list