On Mon, Oct 04, 2021 at 09:21:48AM +0200, Marek Behún wrote:
> On Sat, 2 Oct 2021 11:35:19 -0500
> Bjorn Helgaas <helgaas@xxxxxxxxxx> wrote:
>
> > On Fri, Oct 01, 2021 at 09:58:52PM +0200, Marek Behún wrote:
> > > From: Pali Rohár <pali@xxxxxxxxxx>
> > >
> > > Commit 43f5c77bcbd2 ("PCI: aardvark: Fix reporting CRS value") fixed
> > > handling of CRS response and when CRSSVE flag was not enabled it marked CRS
> > > response as failed transaction (due to simplicity).
> > >
> > > But pci-aardvark.c driver is already waiting up to the PIO_RETRY_CNT count
> > > for PIO config response and so we can with a small change implement
> > > re-issuing of config requests as described in PCIe base specification.
> > >
> > > This change implements re-issuing of config requests when response is CRS.
> > > Set upper bound of wait cycles to around PIO_RETRY_CNT, afterwards the
> > > transaction is marked as failed and an all-ones value is returned as
> > > before.
> >
> > Does this fix a problem?
>
> Hello Bjorn,
>
> Pali has suspisions that certain Marvell WiFi cards may need this [1]
> to work, but we do not have them so we cannot test this.
>
> I guess you think this should be considered a new feature instead of a
> fix, so that the Fixes tag should be removed, yes? Pali was of the
> opinion that this patch "fixes" the driver to conform more to the PCIe
> specification, that is why we added the Fixes tag.
>
> Anyway if you think this should be considered a new feature, this patch
> can be skipped. The following patches apply even without it.
I do not think we should apply to the mainline a fix that can't be
tested sorry, I will skip this patch.
Thanks,
Lorenzo
> Marek
>
> [1]
> https://lore.kernel.org/linux-wireless/CAHp75Vd5iCLELx8s+Zvcj8ufd2bN6CK26soDMkZyC1CwMO2Qeg@xxxxxxxxxxxxxx/
