Re: [RESEND][PATCH 1/1] fpga: altera-cvp: Limit driver registration to specific device ID

[Murugasen Krishnan, Kuhanh]

On 3/11/2025 4:44 pm, Xu Yilun wrote:
> On Thu, Oct 30, 2025 at 03:05:44AM +0800, Kuhanh Murugasen Krishnan wrote:
>> From: "Murugasen Krishnan, Kuhanh" <kuhanh.murugasen.krishnan@xxxxxxxxxx>
> 
> Is this your first post?
> 
> https://lore.kernel.org/all/20250212223553.2717304-1-kuhanh.murugasen.krishnan@xxxxxxxxx/
> 
> Please mark the patch v2 if this patch is for the same issue. And please
> firstly response the talk and make clear all previous concerns, rather than
> just sent the patch and left.
> 
Thanks Yilun for your review. Yes that was the first post, however I do 
not have access to my @intel email address anymore and had to resend 
this with a better commit title and description for clearer explanation 
of this patch. Apologies for the inconvenience.

>>
>> The Altera CvP driver previously used PCI_ANY_ID, which caused it to
>> bind to all PCIe devices with the Altera vendor ID. This led to
>> incorrect driver association when multiple PCIe devices with different
>> device IDs were present on the same platform.
>>
>> Update the device ID table to use 0x00 instead of PCI_ANY_ID so that
>> the driver only attaches to the intended device.
> 
> So could you please answer the previous concern here?
> 
> Does dev_id 0x00 covers all supported devices? Do you have any DOC for
> this?
> 
Yes it will connect to all supported Altera FPGA devices correctly, 
there was a bug previously which caused incorrect driver association 
with the use of PCI_ANY_ID. Limiting the driver registration to 0x00 
allows the driver to attach to the intended Altera FPGA device correctly 
since the FPGA default address is 0x00. Using PCI_ANY_ID could 
potentially allow the CVP driver to associate to other PCI devices.

>>
>> Reviewed-by: Dinh Nguyen <dinguyen@xxxxxxxxxx>
> 
> I didn't see where the tag is from. Generally we don't prefer a
> Reviewed-by tag firstly appear from other than the person named.
> 
> Thanks,
> Yilun
> 
This patch was reviewed internally by Dinh and the tag was added. Should 
I send a v2 patch with this "Reviewed-by" removed?

>> Signed-off-by: Ang Tien Sung <tien.sung.ang@xxxxxxxxxx>
>> Signed-off-by: Murugasen Krishnan, Kuhanh <kuhanh.murugasen.krishnan@xxxxxxxxxx>
>> ---
>>   drivers/fpga/altera-cvp.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/fpga/altera-cvp.c b/drivers/fpga/altera-cvp.c
>> index 5af0bd33890c..97e9d4d981ad 100644
>> --- a/drivers/fpga/altera-cvp.c
>> +++ b/drivers/fpga/altera-cvp.c
>> @@ -560,7 +560,7 @@ static int altera_cvp_probe(struct pci_dev *pdev,
>>   static void altera_cvp_remove(struct pci_dev *pdev);
>>   
>>   static struct pci_device_id altera_cvp_id_tbl[] = {
>> -	{ PCI_VDEVICE(ALTERA, PCI_ANY_ID) },
>> +	{ PCI_VDEVICE(ALTERA, 0x00) },
>>   	{ }
>>   };
>>   MODULE_DEVICE_TABLE(pci, altera_cvp_id_tbl);
>> -- 
>> 2.25.1
>>
>>