Re: [Qemu-devel] SunOS support

[Paolo Bonzini]

On 22/09/2017 14:05, Peter Maydell wrote:
> On 20 September 2017 at 19:50, Peter Tribble <address@hidden> wrote:
>> To introduce myself: I'm a member of the illumos community (the successor
>> to OpenSolaris, to those unfamiliar with us), and I maintain my own illumos
>> distribution.
>>
>> Having seen the scary 'SUPPORT FOR THIS HOST OS WILL GO AWAY'
>> message, I'm reaching out to see what needs to be done so that support for
>> SunOS (not just illumos, I include Oracle's Solaris in the same family)
>> needs to be kept and, where possible, enhanced.
>>
>> I'm willing to act as a contact in this effort, and can work with others in
>> the illumos community to see if there are other resources we can bring to 
>> bear.
> 
> Hi; thanks for getting in touch with us. Kamil Rytarowski (who I've
> cc'd) is also interested in keeping Solaris-variant support working.
> 
> Essentially what we need as upstream is:
>  * access to a machine which we can use for our continuous
>    integration build testing, so we don't break compile
>    support for the platform. This is ideally a machine that
>    somebody else admins and we just use (because we don't
>    want to become solaris/illumos admins ;-)), but failing
>    that, instructions on how to get a VM running under
>    KVM on Linux would also be OK (that's how we've ended
>    up handling the BSDs)

I would even reverse the order since now we're handling the BSDs using
the VM test infrastructure.  Let's say having both would be best.

Paolo

>  * somebody to look at the places where 'make && make check'
>    currently fails and submit upstream patches for them. Kamil
>    has been doing a great job here but would probably
>    appreciate extra help :-)
>  * somebody who's willing to be listed in our MAINTAINERS file
>    as maintaining the port, so we can ask them questions if/when
>    any platform-specific issues come up in future
> 
> Overall we're happy to continue to support hosts that people
> are still using -- we just want to avoid blindly maintaining
> code for platforms we can't test and where we don't have
> any idea if anybody's even using it (for instance we just
> dropped the support for AIX and for Itanium CPUs...)
> 
> thanks
> -- PMM
>