QEMU and mvme3100

Michael Davidsaver mdavidsaver at gmail.com
Sat Oct 28 14:58:12 UTC 2017


On 10/27/2017 02:13 PM, Joel Sherrill wrote:
> 
> 
> On Sun, Oct 22, 2017 at 10:17 PM, Michael Davidsaver
> <mdavidsaver at gmail.com <mailto:mdavidsaver at gmail.com>> wrote:
> 
>     For awhile now I've had a side project to add emulation of the mvme3100
>     board to QEMU with RTEMS as the guest.  I think I have it basically
>     working* and would be interesting to know if anyone else has a use for
>     it.  One of my goals is running the EPICS unittest suite (which passes).
>      It might be useful for RTEMS testing as well.
> 
> 
> It will be useful for RTEMS. 
> 
> Have you submitted the patch upstream to qemu? We like the patches in the
> RSB build of tools to come from upstream sources.

I am starting the process.  However, this isn't a trivial change.  Based
on my first experience with the QEMU project, I expect that this will be
a months long process assuming no major issues.

> Also it would be very much appreciated if you could modify the RSB Qemu
> configuration to include this patch from the upstream Qemu submission.

Ah, RSB...  I'll add this to my TODO list.

> And finally, we are trying to get RTEMS test results on as many BSPs as
> possible. The RTEMS Tester is ready for real users. Testing on a simulator
> moves a BSP from Tier 3 to Tier 2 and results on real hardware move it
> to Tier 1. 

This is what motivated me to bring this up.  I've had good results from
automated testing of EPICS Base on RTEMS.  We're coming up on a release
and have thus far not had our usual last minute RTEMS/vxWorks
regression.  Of course, the last last minute hasn't passed yet (and
there is still vxWorks).

> The addition of RTEMS Tester and adoption of tiers for BSPs and ports was 
> a decision the core developers made to increase confidence and give us
> guidance on when a BSP/port should be deprecated. 
> 
> Chris Johns has posted some instructions this week on how to configured
> the RTEMS Tester for your local setup. I am sure he would be very happy to
> help you get it running for mvme3100 qemu and real hardware.

Contacting Chris is on my TODO list.  I'm hoping that he already has a
recipe for qemu and pc386 (or some other BSP).  It should be
straightforward to adapt this for the mvme3100.

I no longer have access to a real mvme3100, which is part of what led me
to QEMU.

>     Anyone interested in testing will need to build QEMU from
> 
>     > git clone --branch vme https://github.com/mdavidsaver/qemu.git
>     <https://github.com/mdavidsaver/qemu.git>
>     and also build a stub bootloader
> 
>     > git clone --branch tomload
>     https://github.com/mdavidsaver/baremetal.git
>     <https://github.com/mdavidsaver/baremetal.git>
> 
>     Change PREFIX in Makefile as appropriate
> 
>     Then, to eg. run the RTEMS ticker demo
> 
>     > qemu-system-ppc -M mvme3100-1152 -serial stdio -no-reboot \
>     > -d guest_errors,unimp \
>     > -bios tomload.bin \
>     > -kernel powerpc/rtems4.10/ticker.bin
> 
> 
> 
>     * I see occasional 1 sec delays in network packet tx/rx which I haven't
>     tracked down yet.
> 
> 
> We still haven't automated network testing but that's on the wish list
> at least.
> 
> Thanks.
> 
> --joel 
> 
> 
>     _______________________________________________
>     users mailing list
>     users at rtems.org <mailto:users at rtems.org>
>     http://lists.rtems.org/mailman/listinfo/users
>     <http://lists.rtems.org/mailman/listinfo/users>
> 
> 




More information about the users mailing list