[4356] in testers
Re: Possible serious problem w/ rev 1.0 R5000 Indy
daemon@ATHENA.MIT.EDU (Robert A Basch)
Fri Apr 28 17:06:46 2000
Message-Id: <200004282103.RAA503854@aupair.mit.edu>
To: Robert A Basch <rbasch@mit.edu>
cc: testers@mit.edu
In-Reply-To: Your message of "Wed, 26 Apr 2000 16:27:52 EDT."
<200004262027.QAA492028@aupair.mit.edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Fri, 28 Apr 2000 17:03:57 -0400
From: Robert A Basch <rbasch@MIT.EDU>
> I have found that forcing the PROM environment variable _R5000_CVT_WAR
> to 0 makes the problem go away, apparently by disabling the workaround
> for the rev 1.0 processor flaw.
This kind of variable can only be set via the PROM setenv command, not
via the IRIX nvram command, so doing this would be impractical. There
does seem to be a way to build the kernel with the workaround disabled;
again, though, this should only be considered as a last resort.
Bob
