(Illustration by Gaich Muramatsu)
Thanks! That will probably cure the problem rapidly :) !!! - Peter - On Tue, 7 Jul 1998, Bruce Janson wrote: > .. > Date: Tue, 7 Jul 1998 09:27:28 -0400 (EDT) > From: "Peter J. Braam" <braam_at_cs.cmu.edu> > .. > To: linux-coda_at_TELEMANN.CODA.CS.CMU.EDU > .. > Yes it is Coda and it is bad: this is one of the remaining kernel bugs > that we are trying to track down. These alias messages lead to wild > pointers, since dentry's continue to exist, but the inodes they point to > are being freed. > > If anyone can easily produce alias messages, I'd love to know how to nail > the bug! > .. > > Hi Peter, > I have had some success with the following script: > > $ cat aliases > #!/bin/sh > > >/tmp/a > > while : > do > ln /tmp/a /tmp/b > rm /tmp/b > done > $ > > Here, /tmp is a file system which uses the Linux Coda kernel > interface (Linux 2.1.107, Coda 4.6.0). > The above script usually produces an `alias' message > immediately, followed by a cascade of random errors. > > Hope it crashes your filesystem too (and I mean that in the > nicest possible way). > > Regards, > bruce. >Received on 1998-07-07 10:10:01