Kent, Ian I. wrote: > "H. Peter Anvin" wrote: > > >>It's not the browsing that's the big problem (assuming the key database >>is enumerable, which not all of them are.) Rather, it's the "ls -l >>mounts everything" issue, which requires something called stat spoofing >>or inode spoofing. This is an extremely awkward operation, because it >>requires creating a fake VFS object that is valid for some operations >>(lookup, stat) but not for others. > > > I am implementing browsing in autofs working from a suggestion you made some time >ago. I have had fairly good results by implementing the vfs readdir function in the > autofs4 kernel module and the corresponding changes in the daemon. This has been >based on autofs 4.0.0 pre10 against the 2.4 kernel. >
Then you'll have the "ls -l mounts everything" problem. > Any other suggestions? > Any thoughts on how (problem) maps might be enumerated? > > Any idea what has happened to Jeremy? > There have been no updates to autofs 4.0.0 for maybe a year and a half. > How can we get rid of the pre label on autofs 4.0.0? > No idea.
