[RESEND,v3,4/8] qdev: Add memory hot unplug support for bus-less devices.

From: Hu Tao <hutao@cn.fujitsu.com>

Implement bus-less device hot-remove in qdev_unplug(). It will call PCMachine
callback introduced in previous patch.

Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
---
 hw/core/qdev.c | 8 ++++++++
 1 file changed, 8 insertions(+)

On Wed, 27 Aug 2014 16:08:35 +0800
Tang Chen <tangchen@cn.fujitsu.com> wrote:

> From: Hu Tao <hutao@cn.fujitsu.com>
> 
> Implement bus-less device hot-remove in qdev_unplug(). It will call PCMachine
> callback introduced in previous patch.
> 
subject/commit message doesn't need to mention memory hotplug/PCMachine,
it's generic handling that applies to other devices even though
pc-dimm is the only bus-less device handled here.


> Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
> Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
> ---
>  hw/core/qdev.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/hw/core/qdev.c b/hw/core/qdev.c
> index da1ba48..e365a74 100644
> --- a/hw/core/qdev.c
> +++ b/hw/core/qdev.c
> @@ -228,6 +228,14 @@ void qdev_unplug(DeviceState *dev, Error **errp)
>  
>      if (dev->parent_bus && dev->parent_bus->hotplug_handler) {
>          hotplug_handler_unplug(dev->parent_bus->hotplug_handler, dev, errp);
> +    } else if (*errp == NULL) {
what's the reason for ^^^ condition here?

> +        HotplugHandler *hotplug_ctrl;
> +        MachineState *machine = MACHINE(qdev_get_machine());
> +        MachineClass *mc = MACHINE_GET_CLASS(machine);
> +
> +        hotplug_ctrl = mc->get_hotplug_handler(machine, dev);
> +        if (hotplug_ctrl)
> +            hotplug_handler_unplug(hotplug_ctrl, dev, errp);
>      } else {
>          assert(dc->unplug != NULL);
>          if (dc->unplug(dev) < 0) { /* legacy handler */

Hi Igor,

On 09/04/2014 09:22 PM, Igor Mammedov wrote:
> On Wed, 27 Aug 2014 16:08:35 +0800
> Tang Chen <tangchen@cn.fujitsu.com> wrote:
>
>> From: Hu Tao <hutao@cn.fujitsu.com>
>>
>> Implement bus-less device hot-remove in qdev_unplug(). It will call PCMachine
>> callback introduced in previous patch.
>>
> subject/commit message doesn't need to mention memory hotplug/PCMachine,
> it's generic handling that applies to other devices even though
> pc-dimm is the only bus-less device handled here.

Followed.

>
>> Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
>> Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
>> ---
>>   hw/core/qdev.c | 8 ++++++++
>>   1 file changed, 8 insertions(+)
>>
>> diff --git a/hw/core/qdev.c b/hw/core/qdev.c
>> index da1ba48..e365a74 100644
>> --- a/hw/core/qdev.c
>> +++ b/hw/core/qdev.c
>> @@ -228,6 +228,14 @@ void qdev_unplug(DeviceState *dev, Error **errp)
>>   
>>       if (dev->parent_bus && dev->parent_bus->hotplug_handler) {
>>           hotplug_handler_unplug(dev->parent_bus->hotplug_handler, dev, errp);
>> +    } else if (*errp == NULL) {
> what's the reason for ^^^ condition here?
It should be "else if (object_dynamic_cast(qdev_get_machine(), 
TYPE_MACHINE)) "

Just like we do in device_set_realized().

Thanks.
>> +        HotplugHandler *hotplug_ctrl;
>> +        MachineState *machine = MACHINE(qdev_get_machine());
>> +        MachineClass *mc = MACHINE_GET_CLASS(machine);
>> +
>> +        hotplug_ctrl = mc->get_hotplug_handler(machine, dev);
>> +        if (hotplug_ctrl)
>> +            hotplug_handler_unplug(hotplug_ctrl, dev, errp);
>>       } else {
>>           assert(dc->unplug != NULL);
>>           if (dc->unplug(dev) < 0) { /* legacy handler */
>
> .
>