Re: legacy PCI device behind a bridge not getting a valid IRQ on imx host controller

[Bjorn Helgaas <helgaas@kernel.org>]

On Fri, Aug 30, 2024 at 11:49:39AM -0400, Frank Li wrote:
> On Thu, Aug 29, 2024 at 05:00:05PM -0500, Bjorn Helgaas wrote:
> > On Thu, Aug 29, 2024 at 05:43:42PM -0400, Frank Li wrote:
> > > On Thu, Aug 29, 2024 at 04:22:35PM -0500, Bjorn Helgaas wrote:
> > > > On Wed, Aug 28, 2024 at 02:40:33PM -0700, Tim Harvey wrote:
> > > > > Greetings,
> > > > >
> > > > > I have a user that is using an IMX8MM SoC (dwc controller) with a
> > > > > miniPCIe card that has a PEX8112 PCI-to-PCIe bridge to a legacy PCI
> > > > > device and the device is not getting a valid interrupt.
> > > >
> > > > Does pci-imx6.c support INTx at all?
> > >
> > > Yes, dwc controller map INTx message to 4 irq lines, which connect to GIC.
> > > we tested it by add nomsi in kernel command line.
> >
> > Thanks, Frank.  Can you point me to the dwc code where this happens?
> > Maybe I can remember this for next time or add a comment to help
> > people find it.
> 
> I think it needn't special code to handle this. in dts
> 
>  interrupt-map = <0 0 0 1 &gic GIC_SPI 125 IRQ_TYPE_LEVEL_HIGH>,
>                  <0 0 0 2 &gic GIC_SPI 124 IRQ_TYPE_LEVEL_HIGH>,
>                  <0 0 0 3 &gic GIC_SPI 123 IRQ_TYPE_LEVEL_HIGH>,
>                  <0 0 0 4 &gic GIC_SPI 122 IRQ_TYPE_LEVEL_HIGH>;
> 
> It map INT(A,B,C,D) to 4 GIC irq.

OK, thanks!  I guess this happens in the of_irq_parse_pci() path, e.g.:

  imx6_pcie_probe
    dw_pcie_host_init
      devm_pci_alloc_host_bridge
        devm_of_pci_bridge_init
          bridge->map_irq = of_irq_parse_and_map_pci

  pci_device_probe                    # pci_bus_type.probe
    pci_assign_irq
      bridge->map_irq()
        of_irq_parse_and_map_pci
          of_irq_parse_pci
            of_irq_parse_one
              of_irq_parse_raw
                of_get_property(ipar, "interrupt-map", &imaplen)

So there really isn't any mention of INTx directly in imx6 or the dwc
core.

I suppose something like:

  - Set CONFIG_DYNAMIC_DEBUG=y

  - Boot with kernel parameters:

      ignore_loglevel dyndbg="file drivers/pci/* +p; file drivers/of/* +p"

should enable some debug output related to this.

Bjorn