[Coco] Re: NitrOS9 on MESS?

[Lothan]

I doubt this is an issue with index hole timing because it also affects
writes to the virtual hard disk. My initial thought was to compare code in
various versions of NitrOS-9 to determine what changes have been made in the
RBF module that might lead to this bug in MESS, but I haven't been able to
find archives of the source yet. If someone can supply me with the code
revisions from a working version of RBF through to the current version, I'd
me be willing to try to help track this down. My other alternative is to
dump data at various points in MESS, but I suspect this route may prove to
be futile.

-----Original Message-----
From: coco-admin at maltedmedia.com [mailto:coco-admin at maltedmedia.com] On
Behalf Of Roger Taylor
Sent: Thursday, December 04, 2003 12:47 PM
To: coco at maltedmedia.com
Subject: Re: [Coco] Re: NitrOS9 on MESS?

At 08:11 AM 12/4/2003 -0500, you wrote:

>Disks will become corrupt after a certain amount of usage.  This affects 
>both real disks images and VHD virtual hard disks.  Apparently,
                         ------------------------------------------
>somewhere deep in the RBF module the cache structure somehow unloads to 
>disk incorrectly and the disk gets corrupted.
>Robert Gault and I probably put in at least 100+ hours between us trying 
>to track this one down, without any luck.  This bug affects both the 6809 
>and 6309 versions of NitrOS-9, but does not affect older versions of 
>NitrOS-9 nor OS-9.  What I would like to see is someone isolate what 
>specific change to NitrOS-9 caused this bug to manifest; that information 
>would probably unlock the secret of this bug.



At least you know that it's occuring while RBF is writing out to a storage 
device.  So if this does not happen on a real CoCo, and the code is exactly 
the same, then it's M.E.S.S... probably a timing issue or the index hole 
signal code.  ??



{Roger Taylor}


-- 
Coco mailing list
Coco at maltedmedia.com
http://five.pairlist.net/mailman/listinfo/coco