Re: [Qemu-devel] [PATCH qom-next for-next 0/6] QOM realize for IndustryPack

[Andreas Färber]

Am 06.08.2013 10:48, schrieb Alberto Garcia:
> On Fri, Aug 02, 2013 at 11:16:55PM +0200, Andreas Färber wrote:
> 
>> This series converts IndustryPack devices to QOM realize/unrealize.
>>
>> It goes on to clean up file placement, promoting IndustryPack to use
>> its own subdirectory like PCI, ISA, virtio, etc.
> 
>> Andreas Färber (6):
>>   ipack: Convert to QOM realize
>>   ipack: QOM parent field cleanup for IPackBus
>>   ipack: QOM parent field cleanup for IPackDevice

I've now applied these three to qom-next.

>>   ipack: Simplify VMSTATE_IPACK_DEVICE() macro

mst disliked the according change for a PCI VMState macro, so I've not
applied this; but my RFC VMState series has not received a lot of review
either since then, in particular not from Juan, and I've just uncovered
an issue with armv7m_nvic requiring a v2. I'll extend v2 to drop this
macro completely.

>>   ipoctal232: QOM parent field cleanup

Rebased on lack of preceding patch as follows:

diff --git a/hw/char/ipoctal232.c b/hw/char/ipoctal232.c
index 6fe79da..99bab4d 100644
--- a/hw/char/ipoctal232.c
+++ b/hw/char/ipoctal232.c
@@ -155,7 +155,7 @@ static const VMStateDescription vmstate_ipoctal = {
     .minimum_version_id = 1,
     .minimum_version_id_old = 1,
     .fields      = (VMStateField[]) {
-        VMSTATE_IPACK_DEVICE(dev, IPOctalState),
+        VMSTATE_IPACK_DEVICE(parent_obj, IPOctalState),
         VMSTATE_STRUCT_ARRAY(ch, IPOctalState, N_CHANNELS, 1,
                              vmstate_scc2698_channel, SCC2698Channel),
         VMSTATE_STRUCT_ARRAY(blk, IPOctalState, N_BLOCKS, 1,

>>   ipack: Move IndustryPack out of hw/char/

There were unresolved IRC discussions with Paolo where exactly to place
which IndustryPack files, so that this series has been lying around.
Patch 03/06 makes it obvious that the header should live somewhere in
include/ for documenting the base QOM type though.

A minimally invasive alternative would be to move it from hw/char/ to
include/hw/char/ - would that be acceptable for everyone as first step?

The other issue was that while the ipoctal232 device is right in
hw/char/, tpci200.c and ipack.[hc] have nothing to do with char devices
- therefore this patch proposed hw/ipack/, matching drivers/ipack/ in Linux.

> Acked-by: Alberto Garcia <address@hidden>

https://github.com/afaerber/qemu-cpu/commits/qom-next

Thanks,
Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg