Hi Shawn,
I'd suggest to change the designware by dwc on the patch title.
On 25/10/2018 03:47, Shawn Lin wrote:
> On some platform, external MSI domain is using instead of the one
> created by designware driver. For instance, if using GIC-V3-ITS
> as a MSI domain, we only need set msi-map in the devicetree but
> never need any bit in the designware driver to handle MSI stuff.
> So skip allocating its own MSI domain for that case.
>
> Signed-off-by: Shawn Lin <shawn.lin@xxxxxxxxxxxxxx>
>
> ---
>
> drivers/pci/controller/dwc/pcie-designware-host.c | 7 ++++++-
> 1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/pci/controller/dwc/pcie-designware-host.c b/drivers/pci/controller/dwc/pcie-designware-host.c
> index 29a0575..39254b1 100644
> --- a/drivers/pci/controller/dwc/pcie-designware-host.c
> +++ b/drivers/pci/controller/dwc/pcie-designware-host.c
> @@ -278,6 +278,10 @@ int dw_pcie_allocate_domains(struct pcie_port *pp)
> return -ENOMEM;
> }
>
> + /* Rely on the external MSI domain */
> + if (device_property_read_bool(dev, "msi-map"))
> + return 0;
> +
I would suggest to read this property only in the init of the host and
configuring it in a Boolean variable to be created in the pp structure and use
it to verify the feature presence or not.
> pp->msi_domain = pci_msi_create_irq_domain(fwnode,
> &dw_pcie_msi_domain_info,
> pp->irq_domain);
> @@ -433,7 +437,8 @@ int dw_pcie_host_init(struct pcie_port *pp)
> if (ret)
> pci->num_viewport = 2;
>
> - if (IS_ENABLED(CONFIG_PCI_MSI)) {
> + if (IS_ENABLED(CONFIG_PCI_MSI) &&
> + !device_property_read_bool(dev, "msi-map")) {
You could use the previous boolean variable here also.
> /*
> * If a specific SoC driver needs to change the
> * default number of vectors, it needs to implement
>
You should protect the dw_pcie_free_msi(), since it uses the pp->msi_domain and
pp->msi_irq also, maybe using the boolean variable also.
Regards,
Gustavo
