RTEMS on T2080 in hypervisor environment

[Matthew Dahl]

Sebastian,
Thank you for the quick reply.

I can confirm that the HV is supplying the device tree address (as specified in the HV DTB) to the RTEMS guest and the HV is populating that address in RTEMS guest memory with the 'guest' DTB.
I suspect there is some illegal memory access going on. Currently in the process of attempting to confirm this by utilizing the CW probe.

Being first is always fun or at least interesting :)
Will keep banging on it

Thanks,
Matthew Dahl
Senior Software Engineer
Astronautics Corporation of America
(414) 449-4000 x4684


-----Original Message-----
From: Sebastian Huber [mailto:sebastian.huber at embedded-brains.de] 
Sent: Friday, June 23, 2017 12:23 AM
To: Matthew Dahl; users at rtems.org
Subject: Re: RTEMS on T2080 in hypervisor environment

Hello Matthew,


On 22/06/17 19:47, Matthew Dahl wrote:
>
> Hello, I am currently working on developing a server based on the NXP 
> T2080.
>
> Our system will have the T2080 executing a hypervisor and two 
> partitions, one of these partitions is going to be utilizing RTEMS and 
> the other will be Linux.
>
> We have successfully built and tested on our board without the 
> hypervisor and the RTEMS application runs as expected.
>
> The issue is executing under the hypervisor, when this is attempted it 
> appears that the application 'stalls' early in the boot, specifically 
> in the call to the bsp_fdt_copy function which is called from .start.
>
> I can 'track' what instructions the application is getting to using 
> the hypervisor, but like I said above it gets to an if statement and 
> doesn't enter it nor does it jump past it.
>
> My question is: Are there currently plans for a creation/addition of 
> the QorIQ BSP such that the applications execute under the hypervisor?
> Additionally is there any recommended ways to attempt to debug this 
> issue deeper with a bit more than essentially 'print' statements to 
> determine if the code is reaching specific points, etc.
>

you are the first interested to run RTEMS under control of a hypervisor 
on this platform. The BSP needs an FDT which must be supplied by the 
boot loader (or the hypervisor in your case). For debugging I would use 
a Lauterbach debugger or tracer.

-- 
Sebastian Huber, embedded brains GmbH

Address : Dornierstr. 4, D-82178 Puchheim, Germany
Phone   : +49 89 189 47 41-16
Fax     : +49 89 189 47 41-09
E-Mail  : sebastian.huber at embedded-brains.de
PGP     : Public key available on request.

Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG.