Re: [Qemu-devel] [PATCH V6 for-2.3 00/26] hw/pc: implement multiple primary busses for pc machines

[Paolo Bonzini <pbonzini@redhat.com>]

On 20/03/2015 09:37, Marcel Apfelbaum wrote:
>> I think this should not be committed to 2.3 after -rc0.
> While I agree that looking at the above the series do not
> fit for 2.3, I still think the risk is not big because:
> - The biggest amount of modifications are in pci_bus/pci_host
>   and *is only*  a movement of code into a new file,
>   no changes at all (This cover almost half of the series changes)
> - The series does not intervene with usual process, meaning if
>   pxb is not used, it will work exactly like before
> - The bios already supports pxb for this release, is a pity to wait
>   for another QEMU version.
> - The series were in mailing list for more than a month before rc,
>   but the real review process started at rc time.
>   (begging the question what a non-maintainer must do to get its code in
> time)
> - The series is mature, rebased on the latest QEMU
>   and was reviewed in depth and tested by the community.

I understand, but you have to draw a line somewhere.  The easiest place
to draw the line is 2.3.0-rc0.

Maintainers can already prepare branches with patches destined to 2.4,
and these 26 can be included there.  As far as you are concerned, if the
code is ready you can stop worrying about it.

The problem isn't non-maintainer vs. maintainer.  There was a dependency
on Igor's series, which took some time to get in---partly because it was
included in a 100-patch pull request, which isn't exactly the preferred
way to send pull request in absence of a serious effort on continuous
integration of maintainer branches.

Paolo