Re: [PATCH v11 8/9] iommu/arm-smmu-v3: Add in-kernel support for NVIDIA Tegra241 (Grace) CMDQV

From: Will Deacon
Date: Fri Aug 23 2024 - 11:39:42 EST


On Fri, Aug 16, 2024 at 10:41:06AM -0700, Nicolin Chen wrote:
> On Fri, Aug 16, 2024 at 03:19:26PM +0100, Will Deacon wrote:
>
> > [...]
> >
> > > +struct arm_smmu_device *
> > > +tegra241_cmdqv_acpi_dsdt_probe(struct arm_smmu_device *smmu,
> > > + struct acpi_iort_node *node)
> > > +{
> > > + struct resource *res;
> > > + int irq;
> > > +
> > > + /* Keep the pointer smmu intact if !res */
> > > + res = tegra241_cmdqv_find_acpi_resource(smmu, node, &irq);
> > > + if (!res)
> > > + return smmu;
> >
> > Given that this probing code will end up running on non-tegra hardware
> > when CONFIG_TEGRA241_CMDQV is enabled, please can you move the common
> > part into the main driver?
>
> You mean moving the tegra241_cmdqv_acpi_dsdt_probe() into the main
> driver? I think this v11 puts a function call int he header already.

Right. I basically want it so that !tegra folks don't have to open
the tegra-specific file to look at the code running on their SoC.

We should also running that code at all on implementations which _do_
advertise a distinct model number in IORT.

Will