On 18.07.2025 13:52, Mykyta Poturai wrote: > On 16.07.25 11:36, Jan Beulich wrote: >> On 16.07.2025 09:43, Mykyta Poturai wrote: >>> Without pci-passthrough=on Xen does not know anything about present PCI >>> devices due to PHYSDEVOP_pci_device_add not executing. >> >> While the latter half of the sentence is true, Xen may know of PCI devices >> be other means. It scans the bus itself after all on x86, under certain >> conditions. Furthermore "pci-passthrough" is an Arm-only option, when the >> change here affects x86 as much. The description therefore will need some >> improvement; the code change itself is okay (on the assumption that it can >> be justified properly). > > Will adding an additional note that all of this only relates to Arm and > will not change x86 functionality be sufficient?
Probably. Jan > As far as I understand, > this sentence is completely true for Arm, even the yet-to-be-upstreamed > bus scanning routines in Xen are guarded by is_pci_passthroug_enabled in > the latest patches I've seen them in. >
