Re: [PATCH v2] PCI: loongson: Use DECLARE_PCI_FIXUP_EARLY for bridge_class_quirk()

[Tiezhu Yang]

On 07/14/2020 11:08 PM, Lorenzo Pieralisi wrote:
> On Tue, Jul 14, 2020 at 03:37:51PM +0800, Tiezhu Yang wrote:
> > On 06/12/2020 09:30 AM, Tiezhu Yang wrote:
> > > Use DECLARE_PCI_FIXUP_EARLY instead of DECLARE_PCI_FIXUP_HEADER
> > > for bridge_class_quirk() in pci-loongson.c, otherwise the fixup
> > > has no effect.
> > >
> > > Fixes: 1f58cca5cf2b ("PCI: Add Loongson PCI Controller support")
> > > Signed-off-by: Tiezhu Yang <yangtiezhu@xxxxxxxxxxx>
> > > ---
> > >
> > > v2:
> > >     - modify the patch subject used with lower case "loongson"
> > >
> > > This patch is based on mips-next tree.
> > >
> > >    drivers/pci/controller/pci-loongson.c | 6 +++---
> > >    1 file changed, 3 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/drivers/pci/controller/pci-loongson.c b/drivers/pci/controller/pci-loongson.c
> > > index 459009c..58b862a 100644
> > > --- a/drivers/pci/controller/pci-loongson.c
> > > +++ b/drivers/pci/controller/pci-loongson.c
> > > @@ -37,11 +37,11 @@ static void bridge_class_quirk(struct pci_dev *dev)
> > >    {
> > >    	dev->class = PCI_CLASS_BRIDGE_PCI << 8;
> > >    }
> > > -DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_LOONGSON,
> > > +DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_LOONGSON,
> > >    			DEV_PCIE_PORT_0, bridge_class_quirk);
> > > -DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_LOONGSON,
> > > +DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_LOONGSON,
> > >    			DEV_PCIE_PORT_1, bridge_class_quirk);
> > > -DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_LOONGSON,
> > > +DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_LOONGSON,
> > >    			DEV_PCIE_PORT_2, bridge_class_quirk);
> > >    static void system_bus_quirk(struct pci_dev *pdev)
> > Hi,
> >
> > Any comments?

Hi Lorenzo,

Thanks for your reply.

> (1) how was this driver tested if this patch is required ? Is it because
>      you are testing on a different platform ?

According to the datasheet of Loongson LS7A bridge chip, the old
version of Loongson LS7A PCIE port has a hardware bug about PCI
class.

As far as I know, the latest version has already fixed this bug.
I test it on the old version which has hardware bug.

> (2) Please explain why it is needed (I mean describe what happens
>      in current code and how this fixes it) in the commit log, it is
>      useful for people who may need to tweak this code further

OK, this patch is to maintain downward compatibility, I will add
more info to the commit message and then send v3.

Thanks,
Tiezhu

> I will apply it then, thanks.
>
> Lorenzo
>
> > Could you please apply this patch?
> >
> > Thanks,
> > Tiezhu