Linux Virtualization - Jun 2020 - [PATCH 5/6] vdpa: introduce virtio pci driver

On Fri, May 29, 2020 at 04:03:02PM +0800, Jason Wang
wrote:
> +static void vp_vdpa_set_vq_ready(struct vdpa_device *vdpa,
> +				 u16 qid, bool ready)
> +{
> +	struct vp_vdpa *vp_vdpa = vdpa_to_vp(vdpa);
> +
> +	vp_iowrite16(qid, &vp_vdpa->common->queue_select);
> +	vp_iowrite16(ready, &vp_vdpa->common->queue_enable);
> +}
> +
Looks like this needs to check and just skip the write if
ready == 0, right? Of course vdpa core then insists on calling
vp_vdpa_get_vq_ready which will warn. Maybe just drop the
check from core, move it to drivers which need it?

...

> +static const struct pci_device_id vp_vdpa_id_table[] = {
> +	{ PCI_DEVICE(PCI_VENDOR_ID_REDHAT_QUMRANET, PCI_ANY_ID) },
> +	{ 0 }
> +};
This looks like it'll create a mess with either virtio pci
or vdpa being loaded at random. Maybe just don't specify
any IDs for now. Down the road we could get a
distinct vendor ID or a range of device IDs for this.
> +MODULE_DEVICE_TABLE(pci, vp_vdpa_id_table);
> +
> +static struct pci_driver vp_vdpa_driver = {
> +	.name		= "vp-vdpa",
> +	.id_table	= vp_vdpa_id_table,
> +	.probe		= vp_vdpa_probe,
> +	.remove		= vp_vdpa_remove,
> +};
> +
> +module_pci_driver(vp_vdpa_driver);
> +
> +MODULE_AUTHOR("Jason Wang <jasowang at redhat.com>");
> +MODULE_DESCRIPTION("vp-vdpa");
> +MODULE_LICENSE("GPL");
> +MODULE_VERSION("1");
> -- 
> 2.20.1

On 2020/6/2 ??1:08, Michael S. Tsirkin wrote:
> On Fri, May 29, 2020 at 04:03:02PM +0800, Jason Wang wrote:
>> +static void vp_vdpa_set_vq_ready(struct vdpa_device *vdpa,
>> +				 u16 qid, bool ready)
>> +{
>> +	struct vp_vdpa *vp_vdpa = vdpa_to_vp(vdpa);
>> +
>> +	vp_iowrite16(qid, &vp_vdpa->common->queue_select);
>> +	vp_iowrite16(ready, &vp_vdpa->common->queue_enable);
>> +}
>> +
> Looks like this needs to check and just skip the write if
> ready == 0, right? Of course vdpa core then insists on calling
> vp_vdpa_get_vq_ready which will warn. Maybe just drop the
> check from core, move it to drivers which need it?
>
> ...

That may work, but it may cause inconsistent semantic for set_vq_ready 
if we leave it to the driver.

>
>
>> +static const struct pci_device_id vp_vdpa_id_table[] = {
>> +	{ PCI_DEVICE(PCI_VENDOR_ID_REDHAT_QUMRANET, PCI_ANY_ID) },
>> +	{ 0 }
>> +};
> This looks like it'll create a mess with either virtio pci
> or vdpa being loaded at random. Maybe just don't specify
> any IDs for now. Down the road we could get a
> distinct vendor ID or a range of device IDs for this.

Right, will do.

Thanks

>
>> +MODULE_DEVICE_TABLE(pci, vp_vdpa_id_table);
>> +
>> +static struct pci_driver vp_vdpa_driver = {
>> +	.name		= "vp-vdpa",
>> +	.id_table	= vp_vdpa_id_table,
>> +	.probe		= vp_vdpa_probe,
>> +	.remove		= vp_vdpa_remove,
>> +};
>> +
>> +module_pci_driver(vp_vdpa_driver);
>> +
>> +MODULE_AUTHOR("Jason Wang <jasowang at redhat.com>");
>> +MODULE_DESCRIPTION("vp-vdpa");
>> +MODULE_LICENSE("GPL");
>> +MODULE_VERSION("1");
>> -- 
>> 2.20.1

On 2020/6/2 ??3:08, Jason Wang wrote:
>>
>>> +static const struct pci_device_id vp_vdpa_id_table[] = {
>>> +??? { PCI_DEVICE(PCI_VENDOR_ID_REDHAT_QUMRANET, PCI_ANY_ID) },
>>> +??? { 0 }
>>> +};
>> This looks like it'll create a mess with either virtio pci
>> or vdpa being loaded at random. Maybe just don't specify
>> any IDs for now. Down the road we could get a
>> distinct vendor ID or a range of device IDs for this.
>
>
> Right, will do.
>
> Thanks 

Rethink about this. If we don't specify any ID, the binding won't work.

How about using a dedicated subsystem vendor id for this?

Thanks