From: Robert S. <rs...@fr...> - 2002-10-23 02:28:47
|
On Tue, 22 Oct 2002 16:06:50 -0400 Jeff Dike <jd...@ka...> wrote: JD> ja...@st... said: JD> > Like the uml kernel is caching files in its own cache instead of JD> > reading them off the host kernel. JD> > JD> > this would appear to be completly normall to me :> JD> JD> It is normal Really? I'd think a cache should be invalidated when whatever is behind that cache has changed. If this isn't the way it works, it should be documented in great big red capital letters. If it is already documented, the lettters need to be redder, bigger and bolder, as I totally missed it. It is counter-intuitive for a file system not to stay in sync. JD> I'm considering adding a no-page-cache flag to hostfs which would cause it JD> to do everything synchronously. I'd say that this should be the default mode. -- Robert Story, NET-SNMP Junkie <irc://irc.openprojects.net/#net-snmp> <http://www.net-snmp.org/> You are lost in a twisty maze of little standards, all different. |