Re: [PATCH] PCI: pci-bridge-emul: Set position of PCI capabilities to real HW value

From: Bjorn Helgaas
Date: Thu Aug 18 2022 - 19:03:04 EST


On Fri, Aug 19, 2022 at 12:36:36AM +0200, Pali Rohár wrote:
> On Thursday 18 August 2022 17:31:36 Bjorn Helgaas wrote:
> > On Thu, Aug 18, 2022 at 03:50:54PM +0200, Pali Rohár wrote:
> > > On Sunday 03 July 2022 12:46:27 Pali Rohár wrote:
> > > > mvebu and aardvark HW have PCIe capabilities on different offset in PCI
> > > > config space. Extend pci-bridge-emul.c code to allow setting custom driver
> > > > custom value where PCIe capabilities starts.
> > > >
> > > > With this change PCIe capabilities of both drivers are reported at the same
> > > > location as where they are reported by U-Boot - in their real HW offset.
> >
> > Just curious since I haven't read the patch, and Lorenzo will take
> > care of this anyway, but does this fix a bug, i.e., does something
> > work when it didn't work before? Or does everything *work* without
> > this patch, but lspci reports capabilities at different offsets than
> > U-Boot?
>
> The last sentence is correct. Everything works with and also without
> this patch. Just without this patch lspci reports capabilities at
> different offsets than what HW reports and what U-Boot reports (U-Boot
> already reports offsets same as in HW).
>
> So lets say, that with this patch, it is easier to compare pci config
> space dump from u-boot and linux. And this simplify debugging.

I agree, that's a really annoying difference. Seems like a good thing
to fix (again, not having looked at the patch, I have no opinion on
the implementation :)).

Bjorn