Hi Thiery, Rob, Robin, and Joerg, On 26 Aug 2025 09:36, Jason Gunthorpe wrote: > On Mon, Aug 25, 2025 at 06:15:57PM -0700, Shyam Saini wrote: > > Hi Jason, > > > > On Mon, Aug 18, 2025 at 12:53:10PM -0300, Jason Gunthorpe wrote: > > > On Thu, Aug 14, 2025 at 07:32:36PM -0700, Shyam Saini wrote: > > > > On Thu, Aug 14, 2025 at 09:39:58PM -0300, Jason Gunthorpe wrote: > > > > > On Thu, Aug 14, 2025 at 04:30:18PM -0700, Shyam Saini wrote: > > > > > > or were you referring to [2]? > > > > > > > > > > > > In that case, the PCI child node data needs to be parsed, which is > > > > > > currently handled individually by each host controller driver. > > > > > > > > > > Yes, this looks like it may be what I was thinking of, the pci@1,0 > > > > > specifes the BDF effectively > > > > > > > > In that case, we'll need to parse the child DTS nodes properly > > > > within of_iommu_get_resv_regions(). I'll include this in v4. > > > > > > Kinda surprised this isn't happening already? It would be good to > > > refer to the original specs and describe how whatetever you propose is > > > aligned there. > > > > Just to confirm, does the v3 version of this series look good to you? > > If so, I’ll go ahead and respin the series with the iommu_set_sw_msi() > > change and address the other review comments from Jacob. > > > > Otherwise having pci devices nodes in the fdt or dts needs additional > > handling, let me know your preference :) > > I think you need to consult with the DT/OF owners on these questions.. > > I've forgotton so many of the details here. > > But once you settle on whatever the DT representation is it looks like > the right direction to me. >
The of_iommu driver currently does not correctly handle parsing of the memory-region property for individual PCI devices. This patch addresses the issue by explicitly setting dev->of_node for PCI devices, enabling proper parsing of the memory-region property within of_iommu_get_resv_regions(). Could you please share your feedback on this fix? Thanks, Shyam
