[OpenWrt-Devel] [PATCH] ath79: ar71xx create a separate driver for ar71xx pci interrupt controller.

[Dmitry Tunin]

> > This should fix a bug with one IRQ for all PCI devices and also handle the PCI_CORE interrupt.
> No it doesn't. Assigning it in dts doesn't mean that we "handled" it.
> But I still like this patch because it's doing a cleanup of the messy code.

I am still not very competent in the IRQ handling. I thought that if we set
interrupts = <4>; and the new partent, your "case 4:" may work.

> >
> > +                       pci_intc: interrupt-controller at 18060018 {
> > +                               compatible = "qca,ar7100-pci-intc";
> > +                               reg = <0x18060018 0x4>;
> > +                               interrupt-parent = <&cpuintc>;
> > +                               interrupts = <2>;
> > +                               interrupt-controller;
> > +                               #interrupt-cells = <1>;
> > +                       };
> > +
> For other chips this node is a subnode of reset-controller. But in
> this case here we can just put this above reset-controller at 18060024.
> Nodes in dts are supposed to be ordered by register address.

No problem with that, I thought about a child node. But it doesn't
make any difference IMHO.
> > +
> >                         pcie0: pcie-controller at 180c0000 {
> >                                 compatible = "qca,ar7100-pci";
> >                                 #address-cells = <3>;
> > @@ -105,14 +115,16 @@
> >                                 reg-names = "cfg_base";
> >                                 ranges = <0x2000000 0 0x10000000 0x10000000 0 0x07000000        /* pci memory */
> >                                           0x1000000 0 0x00000000 0x0000000 0 0x000001>;         /* io space */
> > -                               interrupt-parent = <&cpuintc>;
> > -                               interrupts = <2>;
> > +                               interrupt-parent = <&pci_intc>;
> > +                               interrupts = <4>;
> Do we really need this? We don't even have a handler that actually do
> anything for it. I think the above two lines can simply be dropped.

See above. I thought the existing handler could work. I must be wrong.

> > diff --git a/target/linux/ath79/patches-4.14/0020-MIPS-ath79-turn-pci-ar71xx-driver-into-a-pure-OF-dri.patch b/target/linux/ath79/patches-4.14/0020-MIPS-ath79-turn-pci-ar71xx-driver-into-a-pure-OF-dri.patch
> > index ea3514a..95ca6d1 100644
> > --- a/target/linux/ath79/patches-4.14/0020-MIPS-ath79-turn-pci-ar71xx-driver-into-a-pure-OF-dri.patch
> > +++ b/target/linux/ath79/patches-4.14/0020-MIPS-ath79-turn-pci-ar71xx-driver-into-a-pure-OF-dri.patch
> It's not a good idea to mix the removal of IRQ part into this patch. I
> suggest using a separated patch to do the removing.

It is a minor issue if everything else looks good. The idea was to
keep number of patches smaller.


> I haven't read the code thoroghly but it seemed that irq-ath79-misc
> works in a similar way. Is it possible to use the misc intc driver for
> PCI?

They are similar. I used that code as an example, but the
irq-ath79-pci has your mask/unmask for pci with reading the registers.
It is possible to make one driver for both, but is it really needed?

_______________________________________________
openwrt-devel mailing list
openwrt-devel at lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel