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

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

[Kuhanh Murugasen Krishnan]

From: "Murugasen Krishnan, Kuhanh" <kuhanh.murugasen.krishnan@altera.com>

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.

Reviewed-by: Dinh Nguyen <dinguyen@kernel.org>
Signed-off-by: Ang Tien Sung <tien.sung.ang@altera.com>
Signed-off-by: Murugasen Krishnan, Kuhanh <kuhanh.murugasen.krishnan@altera.com>
---
 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

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

[Xu Yilun]

On Thu, Oct 30, 2025 at 03:05:44AM +0800, Kuhanh Murugasen Krishnan wrote:
> From: "Murugasen Krishnan, Kuhanh" <kuhanh.murugasen.krishnan@altera.com>

Is this your first post?

https://lore.kernel.org/all/20250212223553.2717304-1-kuhanh.murugasen.krishnan@intel.com/

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.

> 
> 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?

> 
> Reviewed-by: Dinh Nguyen <dinguyen@kernel.org>

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

> Signed-off-by: Ang Tien Sung <tien.sung.ang@altera.com>
> Signed-off-by: Murugasen Krishnan, Kuhanh <kuhanh.murugasen.krishnan@altera.com>
> ---
>  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
> 
> 

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@altera.com>
> 
> Is this your first post?
> 
> https://lore.kernel.org/all/20250212223553.2717304-1-kuhanh.murugasen.krishnan@intel.com/
> 
> 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@kernel.org>
> 
> 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@altera.com>
>> Signed-off-by: Murugasen Krishnan, Kuhanh <kuhanh.murugasen.krishnan@altera.com>
>> ---
>>   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
>>
>>

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

[Xu Yilun]

On Wed, Nov 05, 2025 at 06:29:06AM +0000, Murugasen Krishnan, Kuhanh wrote:
> 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@altera.com>
> > 
> > Is this your first post?
> > 
> > https://lore.kernel.org/all/20250212223553.2717304-1-kuhanh.murugasen.krishnan@intel.com/
> > 
> > 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, 

Because your change is trying to reduce the scope of devices the driver
could support. I want to be cautious and ask for Public Documentation
for reference. I don't want to see someone later yells "Oh, my device is
broken!".

Please also add the Link of the Documentation in changelog.

> there was a bug previously which caused incorrect driver association 

So this is a bug, please tag with "fixes:" and Cc stable kernel.

> 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@kernel.org>
> > 
> > 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?

Yes, fix previous comments in your v2, and removes the Reviewed-by.
Reviewed-by is welcomed but should be publicly originated from the
named person.

Thanks,
Yilun

> 
> >> Signed-off-by: Ang Tien Sung <tien.sung.ang@altera.com>
> >> Signed-off-by: Murugasen Krishnan, Kuhanh <kuhanh.murugasen.krishnan@altera.com>
> >> ---
> >>   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
> >>
> >>
>