Hmmm, David, I think you missed something (or maybe Richard doesn't
mean what this really says):
] Correct! Now, is a directory access a file access? I don't think it
] should be. If it is, the new directory caching code was a lot of work
] that will be wasted in real-world applications such as a compile using
] `make`, etc. I think that only the directory "files" have been accessed,
] not every file. 'ls -la --time=atime` shows every file on my disk as
] having been accessed after `ls -R /`, which, in fact, has not happened.
] Only the directory files have been accessed.
What that appears to say to me is that stat(2) is dirtying the inodes for
files (not just directories) unless "ls -R /" is opening each file ?!
Darren