Subject: Re: lfs
To: Konrad Schroder <perseant@hhhh.org>
From: Thor Lancelot Simon <tls@rek.tjls.com>
List: tech-kern
Date: 02/27/2003 13:18:32
On Thu, Feb 27, 2003 at 10:10:03AM -0800, Konrad Schroder wrote:
> On Wed, Feb 26, 2003 at 04:32:07PM -0500, Thor Lancelot Simon wrote:
> 
> > Clearly the underlying problems here should be fixed, but in the interim,
> > how about simply reserving some more space so that even root can't write
> > into it?  It seems to me that max(5%, segsz * 10) should suffice to avoid
> > the kind of deadlock that gets us into this mess...
> 
> Yes, we could do that.  My guess is that it would have to be more like
> 20%, though.  Lowering lfs_bfree or raising lfs_minfreeseg should have

I don't understand this.  How can we have demand for 20% of the segments
in the filesystem, even if there is simultaneous cleaning and user I/O
going on?  How can we require 20% "slop" in order to avoid deadlock?  The
Sprite code, IIRC, reserved two segments, and noted that reserving one
should theoretically be sufficient.

Thor