Subject: Re: an in-kernel getcwd() implementation
To: Jonathan Stone <jonathan@DSG.Stanford.EDU>
From: Bill Sommerfeld <firstname.lastname@example.org>
Date: 03/07/1999 16:13:21
> systat vmstat keeps printing lines about a namei proc-cache...
> now you mention it maybe it never has nonzero hit rate.
No, it does show up non-zero in systat on occasion.
It has to do with the use of per-process pointers into directories to
prevent a loop stat'ing every file in a directory from taking O(n**2)
time. This can't help reverse lookups in any meaningful way.
> nullfs doesnt blow up on mount anymore, but it may still be horribly
> broken internally. i just wondered why it was absent.
I also haven't tested it with adosfs, coda, filecorefs, isofs,
msdosfs, nfs, and probably a few others.
Note also that mount -o union (which is what I think I referred to in
my mail) is different from mount -t unionfs.