Hello Ian, I think the 'ghosting' feature is vital to interoperability between Linux and Solaris systems. Sun calls this feature 'browsing' and Solaris has had it for some time. I don't believe that Irix, AIX and other Unices have this ability, so it would be a big plus if Linux could do it. I'm surprised that this feature hasn't made it into the autofs kernel module yet, it would seem to be a feature that would be widely used.
I can't speak for anybody else but I would very much like to keep up with current developments on the autofs list. (isn't that the point of a mailing list? :-) ) -A -----Original Message----- From: Ian Kent [mailto:[EMAIL PROTECTED]] Sent: Monday, February 10, 2003 8:43 AM To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: [autofs] 2.4.20 autofs4 changes Hi all, I have written a patch for the autofs4 kernel module (and autofs 4.0.0-pre10 daemon) to deal with ghosting of directories (and to a limited degree, direct automount maps). Refer to http://www.themaw.net/includehtml.php?sendit=ians_html/autofs.html for more information if you wish. There have been a few changes to the autofs4 module in kernel release 2.4.20 that break this patch. I have looked through the Changelog for 2.4.20 and 2.4.19 and don't see any reference to the changes. I have also had a brief look through the archives of this list without success. A first look indicates that the problem for me has been caused by the removal of the file_operations definition autofs4_dir_operations in root.c. Then the dcache default file_operations are used for inode initialization in inode.c instead. My initial impression is that if there is a specific reason for using the dcache default then I will need to work on an autofs4 implementation of dcache_dir_lseek. The rest may be OK. To save me some time and possible pain I am hoping to find out the reasoning behind this change from the implementor. Please copy any reply to my personal email address as I am not subscribed to this list. I think that the autofs list is closed to people who are not subscribed so I will forward any replies that don't make it there. Should notification of these changes be posted to the autofs list? -- Ian Kent <[EMAIL PROTECTED]> _______________________________________________ autofs mailing list [EMAIL PROTECTED] http://linux.kernel.org/mailman/listinfo/autofs _______________________________________________ autofs mailing list [EMAIL PROTECTED] http://linux.kernel.org/mailman/listinfo/autofs
