[Qemu-devel] q35 chipset support

[Jason Baron <jbaron@redhat.com>]

Hi,

I recently updated Isaku Yamahata's q35 patches to work on the latest qemu and
seabios trees. On the qemu side, most of the changes revolved around updating
to use QOM and updates to the memory API. I was also able to drop quite a few
patches that had already been resolved by the current qemu tree.

The trees seem pretty stable and can be found here:

git://github.com/jibaron/q35-qemu.git
git://github.com/jibaron/q35-seabios.git

A basic qemu cmd to start q35, looks like:

$ /usr/local/bin/qemu-system-x86_64  -name "f16" -M pc_q35 -m 1G -smp 4  -hda
./f16.img --enable-kvm  -bios <path>/q35-seabios/out/bios.bin
-acpitable file=<path>/q35-seabios/out/q35-acpi-dsdt.aml  -monitor stdio

The main reason to bring in q35 chipset support is for pcie support - thus
allowing drivers to make use of the extended config space, aer, pcie hotplug.
So, I wanted to start a discussion around what is missing from this patchset,
and how I should structure this patchset (perhaps, I should post the patches
to the list?). For example, does the q35 chipset support need to have feature
parity with the current i440fx before it would be considered? Or, can we bring
these features in incrementally?

As far as what I know is currently missing:

-device hotplug support needs to be added. As Michael Tsirkin pointed out
devices behind pcie bridges use up 4K i/o space, so we might quickly run out
of i/o space vs. pci bridges...
-migration vmstate needs to be updated (I didn't test or fixup migration yet)

I'm sure there is a lot more...

Thanks,

-Jason