Does anyone known how mounting works? We don't have mounting per see, when you tell settrans to start an active translator then the translator starts at once and get attached to the node. For passive translators settrans just writes a bit of meta-data into the file-system so that when the next time you access it (using stat(), open(), etc) it gets started.
Is there a central namespace or something like that? The file-system is our name-space. If the ext2fs translator is stored on a ext2 partition, how does the kernel load it? Using magic. :) This bit is quite complex, to put it simply, GRUB loads a bunch of modules (ext2fs and exec for example) so that the rest of the system can bootstrap. This is the _really really simple_ description. Does the kernel have any concept of "file system"? No. or interface that a fs translator has to implement? How you implement a file-system translator is up to the person implementing it. One can for example use libdiskfs, but if your thing is to rewrite everything just for the hell of it then you can do that. Does is the ftpfs finished and working? Yes, and no. On a good day it works, on a bad day it doesn't work. Hope this answers some of the questions. _______________________________________________ Help-hurd mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/help-hurd
