jh@pcsbst.UUCP (10/30/87)
We at PCS would be interested to know about System V.3 or V.3.1 FSS (File System Switch) implementations which gives the ability to mount MS-DOS file systems. We have following questions: - Did someone try to do this yet? - Are changes necessary to UNIX, e.g. in namei(). - Did you encounter some severe problems? - Is the software available in source code ? I am interested in a cpio like program which may read and write MS-DOS file systems, as well. Is such a software available in source code? Johannes Heuft unido!pcsbst!jh
mikep@ism780c.UUCP (Michael A. Petonic) (11/03/87)
In article <4300002@pcsbst.UUCP> jh@pcsbst.UUCP writes: > > >We at PCS would be interested to know about System V.3 or V.3.1 >FSS (File System Switch) implementations which gives the ability >to mount MS-DOS file systems. > >We have following questions: > >- Did someone try to do this yet? Yes, it has been implemented in 386/ix, INTERACTIVE's port of UNIX System V.3. >- Are changes necessary to UNIX, e.g. in namei(). Yes, very briefly, there is a front end to namei() which is generic and calls FileSystem dependant namei() routines depending upon which filesystem you are working on. There are plenty of changes made to the kernel since FSS is implemented almost wholy in the kernel. From personal knowledge, all the modifications done to implement a DOS-FSS by INTERACTIVE were done in a manner that was set forward by AT&T in their guidlines for making FSS. I don't have the document with me, otherwise I would state the title, but I'm pretty sure you must have it if you know as much as you do. >- Did you encounter some severe problems? There are always compromises to be made, like determining what an inode is on an MS-DOS filesystem, file mode protection on DOS, etc. These are to be conquered on a one-by-one basis. >- Is the software available in source code ? I'm not a marketing type so I don't know. I assume that if you buy the source for 386/ix, you'd get the source for the FSS, also. >I am interested in a cpio like program which may read and write >MS-DOS file systems, as well. Is such a software available in source >code? There was a public domain posting of one to comp.sources about a couple months ago. Also, almost every UNIX related system on x86 machines that I've seen have some version of this program. For instance, INTERACTIVE has something called DOSSETTE, and INTEL's Xenix 3.x (sco's) has a similar program. > > Johannes Heuft > > unido!pcsbst!jh If you want more detailed information, you might want to contact the following person: {sdcrdcf|attunix|microsoft|sfmin}!ism780c!arun Since he did the kernel work for FSS here. I worked on one small teeny weeny aspect of FSS's impact on a command, so I don't really know the nitty gritty about it.
fred@rover.UUCP (Fred Christiansen) (11/09/87)
In article <7704@ism780c.UUCP> mikep@ism780c.UUCP (Michael A. Petonic) writes: >Yes, very briefly, there is a front end to namei() which is generic [...] > [...] >that was set forward by AT&T in their guidlines for >making FSS. I don't have the document with me, otherwise >I would state the title, but I'm pretty sure you must have it [...] Could someone e-mail me the name of this docucment? I'd appreciate having it for some work (OSI/FTAM) that I'm involved with. Thanks. -- Fred Christiansen ("Canajun, eh?") | hao!noao!mcdsun!nud!fred (ARPA gate) Motorola Microcomputer Div., Tempe, AZ | utzoo!mnetor!mot!fred "The greatest thing a father can do for | ihnp4!mot!fred his children is to love their mother." | hplabs!motsj1!ellymae!fred