[PATCH 6/7] accel/qaic: Add config structs for supported cards

Lizhi Hou lizhi.hou at amd.com
Fri Dec 20 18:08:36 UTC 2024


On 12/20/24 09:15, Jeffrey Hugo wrote:
> On 12/13/2024 5:35 PM, Lizhi Hou wrote:
>>
>> On 12/13/24 13:33, Jeffrey Hugo wrote:
>>> -static struct qaic_device *create_qdev(struct pci_dev *pdev, const 
>>> struct pci_device_id *id)
>>> +static struct qaic_device *create_qdev(struct pci_dev *pdev,
>>> +                       const struct qaic_device_config *config)
>>>   {
>>>       struct device *dev = &pdev->dev;
>>>       struct qaic_drm_device *qddev;
>>> @@ -365,12 +391,10 @@ static struct qaic_device *create_qdev(struct 
>>> pci_dev *pdev, const struct pci_de
>>>           return NULL;
>>>       qdev->dev_state = QAIC_OFFLINE;
>>> -    if (id->device == PCI_DEV_AIC080 || id->device == 
>>> PCI_DEV_AIC100) {
>>> -        qdev->num_dbc = 16;
>>> -        qdev->dbc = devm_kcalloc(dev, qdev->num_dbc, 
>>> sizeof(*qdev->dbc), GFP_KERNEL);
>>> -        if (!qdev->dbc)
>>> -            return NULL;
>>> -    }
>>> +    qdev->num_dbc = 16;
>>
>> Is it better to put num_dbc in qaic_device_config?
>
> I think there is no clear "right answer".  All known devices use 16. 
> There may be a future device which has a different value, at which 
> point I think this needs to be in qaic_device_config.  For this patch, 
> I was conservative and only included items in qaic_device_config which 
> do vary between the known devices.
>
> I'll think in this a bit more.
Reviewed-by: Lizhi Hou <lizhi.hou at amd.com>
>
>>
>> Thanks,
>>
>> Lizhi
>>


More information about the dri-devel mailing list