Re: [PATCH 2/2] fpga: altera-cvp: Propagate PCI config access errors

[Daisuke Matsuda]

On 2026/07/23 12:42, Xu Yilun wrote:
> On Wed, Jul 01, 2026 at 09:26:33AM +0000, Daisuke Matsuda wrote:
> > From: Daisuke Matsuda <matsuda@xxxxxxxxxxxx>
> >
> > The CvP control path accesses the VSEC registers through PCI config space,
> > but several reads and writes ignore the return value. If a config read
> > fails, the driver can make decisions from an undefined register value and
> > continue programming with bogus status or control bits.
>
> Searching from the kernel, most pci_read/write_config_dword() don't
> check return values. Could you help me understand why they don't but
> this driver has the specific need?

Hi Yilun,

Thank you for taking a look.

I agree that the vast majority of drivers do not check the return values, and I don't have a strong reason to make this driver behave differently.

I'll drop this patch and send a v2 with only patch 1/2.

Thanks,
Daisuke

> Thanks,
> Yilun