Re: New feature

Alan Cox (
Wed, 24 Sep 1997 08:50:39 +0100 (BST)

> > The ATIME updates dirty all the inodes.
> That's true and ok, but it shouldn't be re-reading the media again
> should it? The dirty bit should be cleared, but the buffer shouldn't
> be marked invalid...

Its writing to the disk, to write all the new atime values into the
inode blocks