On Tue, 25 Feb 2003, Mr. Wong wrote:

Can someone give me some pointers on this? What should I be using? Has

anyone done this before? I am aware that there will be inconsistancies and

errors introduced but I'm willing to tolerate / fix / learn about the issues

involved.

There is a way to do this. What you want to do is get the source for the nullfs filesystem and modify it to cover the mount vnode instead of mapping a vnode from another part of the file system. Then *all* file system accesses go though your layer before hitting the underlying HFS or whatever file system. You can add code to VOP_WRITE, VOP_ACCESS, VOP_RMDIR, VOP_CREATE, etc to prevent writing to the underlying file system. nullfs already does this if the mount flags say the nullfs mount layer is a read-only file system. You should read up about the VFS stack and stacked file systems to undertsand whats going on. Also look at the nullfs code. If you're serious about doing this, I would highly recommend you setup a 2 computer Kernel debug setup as well so you can figure out the ref-count and locking kernel panics you'll probably see during development. I've also been using the URL http://www.gsp.com/support/man/ which has man page entries for the various VOP/VFS functions, even though its FreeBSD oriented the ideas still work. Also the file vnode_if.src in the open source tree (in xnu i think) is very helpful in the locking protocols. All of this is of course not offically supported by Apple as they are not promising not to break this in a future release. --- Marek Kozubal marek@portents.com _______________________________________________ darwin-kernel mailing list | darwin-kernel@lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.