Re: [Qemu-devel] [PATCH 04/61] virtio-pci : refactor virtio-pci device.

["KONRAD Frédéric" <fred.konrad@greensocs.com>]

On 08/01/2013 18:54, Peter Maydell wrote:
> On 7 January 2013 18:40,  <fred.konrad@greensocs.com> wrote:
>> From: KONRAD Frederic <fred.konrad@greensocs.com>
>>
>> Create the virtio-pci device. This transport device will create a
>> virtio-pci-bus, so one VirtIODevice can be connected.
>>
>> Signed-off-by: KONRAD Frederic <fred.konrad@greensocs.com>
>> ---
>>   hw/virtio-pci.c | 127 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>   hw/virtio-pci.h |  18 ++++++++
>>   2 files changed, 145 insertions(+)
>>
>> diff --git a/hw/virtio-pci.c b/hw/virtio-pci.c
>> index 5ff03e8..1f0ecbe 100644
>> --- a/hw/virtio-pci.c
>> +++ b/hw/virtio-pci.c
>> @@ -1163,6 +1163,130 @@ static TypeInfo virtio_scsi_info = {
>>       .class_init    = virtio_scsi_class_init,
>>   };
>>
>> +/*
>> + * virtio-pci : This is the PCIDevice which have a virtio-pci-bus.
> "has"
>
>> + */
>> +
>> +/* This is called by virtio-bus just after the device is plugged. */
>> +static void virtio_pci_device_plugged(DeviceState *d)
>> +{
>> +    VirtIOPCIProxy *proxy = VIRTIO_PCI(d);
>> +    VirtioBusState *bus = proxy->bus;
>> +    uint8_t *config;
>> +    uint32_t size;
>> +
>> +    /* Put the PCI IDs */
>> +    switch (virtio_device_get_id(proxy->bus)) {
>> +
>> +
>> +    default:
>> +        error_report("unknown device id\n");
>> +        break;
>> +
>> +    }
> This doesn't have any code in it yet (it gets added in later patches) but
> the final result looks very repetitive. I think you'd be better off just having
> some arrays:
>
> uint16_t virtio_pci_device_id[] = {
>       [VIRTIO_ID_BLOCK] = PCI_DEVICE_ID_VIRTIO_BLOCK,
>       [VIRTIO_ID_NET] = PCI_DEVICE_ID_VIRTIO_NET,
> (etc)
> };
>
> similarly for the class. Then you can just drop the switch statement.
>
> In fact I think you might as well put in the array entries for all
> the virtio devices in this patch rather than adding one in each of the
> "add virtio-foo device" patches; it will do no harm for them to be
> there early and it makes the later patches a little smaller.
And what happen when it is a bad ID ?

Fred

>
>> +
>> +    proxy->vdev = proxy->bus->vdev;
>> +
>> +    config = proxy->pci_dev.config;
>> +    if (proxy->class_code) {
>> +        pci_config_set_class(config, proxy->class_code);
>> +    }
>> +    pci_set_word(config + PCI_SUBSYSTEM_VENDOR_ID,
>> +                 pci_get_word(config + PCI_VENDOR_ID));
>> +    pci_set_word(config + PCI_SUBSYSTEM_ID, virtio_device_get_id(proxy->bus));
>> +    config[PCI_INTERRUPT_PIN] = 1;
>> --- a/hw/virtio-pci.h
>> +++ b/hw/virtio-pci.h
>> @@ -45,6 +45,22 @@ typedef struct {
>>       unsigned int users;
>>   } VirtIOIRQFD;
>>
>> +/*
>> + * virtio-pci : This is the PCIDevice which have a virtio-pci-bus.
> "has"
>
>> + */
> -- PMM
>