Re: earmhf issues on Beaglebone Black

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Wed, Jul 23, 2014 at 02:45:35PM -0700, Matt Thomas wrote:
> 
> On Jul 23, 2014, at 2:44 PM, Manuel Bouyer <bouyer%antioche.eu.org@localhost> 
> wrote:
> 
> > One difference with the linux code is that we're doing lots of
> > things before reading far and fsr. Linux reads these 2 registers right after
> > the clrex.
> > I wonder if something between clrex and reading these registers
> > (ENABLE_ALIGNMENT_FAULTS maybe ?) could cause fsr to be modified ?
> 
> Don?t think so.  One thing I wonder is whether this is an A8 problem or more
> generic.

In the "Cortex-A8 Technical Reference Manual Revision: r3p2"
from infocenter.arm.com, in
"3.2.28. c1, Secure Configuration Register"
there is a table, which says:
"For Nonsecure state, avoid this because the core can unexpectedly enter an
abort mode" when AW and EA are both 0, which is the default reset value.

But I'm not sure if this is relevant or not.


-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--