Functional testing of mainline vchiq driver

[Stefan Wahren]

> Phil Elwell <phil at raspberrypi.org> hat am 31. Oktober 2016 um 13:15
> geschrieben:
> 
> 
> On 31/10/2016 12:10, Stefan Wahren wrote:
> >> Phil Elwell <phil at raspberrypi.org> hat am 31. Oktober 2016 um 12:05
> >> geschrieben:
> >>
> >>
> >> On 31/10/2016 10:42, Stefan Wahren wrote:
> >>> Phil Elwell <phil at raspberrypi.org> hat am 31. Oktober 2016 um
> >>> 11:28geschrieben:
> >>>> That's because the firmware doesn't read the cache line size value from
> >>>> the DTB - it knows which chip it is running on, and hence the correct
> >>>> value - so forcing an incorrect value for the ARM causes a discrepancy.
> >>> In case VPU and ARM core need the same value i suggest to request the
> >>> firmware
> >>> about that.
> >>> I assume there is no possibility to request the cache line size directly.
> >>> So we need to decide based on something like RPI_FIRMWARE_GET_BOARD_MODEL?
> >> The firmware already tells the ARM the correct value by ensuring it has
> >> an appropriate DTB file. Are you really suggesting adding a method to
> >> query at runtime a property of the hardware which cannot change?
> >>
> >> I thought Device Tree was meant to contain static device configuration
> >> information like that. The fact that there isn't already a
> >> cache-line-size property on the CPU nodes seems like an oversight.
> >>
> > Okay now the penny dropped. Forget about my suggestion.
> >
> > Sorry, one more question is it guaranteed that the VPU firmware finished
> > writing
> > to the Device Tree BEFORE the ARM core starts?
> Yes.
>

Thanks. I think we should keep cache-line-size property until discussing the
binding with the DT guys. But i suggest to change the property as mandatory with
a proper error message in case it's missing.