Daniel P. Berrangé wrote:
On Fri, Feb 06, 2026 at 02:17:40PM +0100, Marek Marczykowski-Górecki wrote:
On Thu, Feb 05, 2026 at 10:38:27PM -0000, bertrandlec--- via Devel wrote:
Bertrand, can you share `lspci -t` output, and also `lspci -vvs 80:00.0` ? I don't have any device at address 80:00.0. Oh, that's why it fails... This feels like a firmware (if not hardware) issue, I think root complex should be visible normally... But maybe somebody more familiar with PCIe spec can help? Yeah, checking if there is any firmware update available would be a good idea, as to me it is a bit odd to see the separate hierarchy without a host bridge at 0000:80:00.0. Admittedly I'm not a PCIE expert though
I am using the lastest version of the Laptop firmware. The last version was published yesterday and it does not change that behaviour. I redid my test on that version. From my reading on PCI/PCIe (but I am definitively not an expert), a PCI Host Bridge or a PCIe Root Complex does not necessarely need to be addressable on the PCI bus, specially on a PCH when it is connected to the CPU SoC by a proprietary bus (Intel DMI Direct Media Interface in my case). On my laptop, the bus 0000:80 is recognized by Linux, as shown in the "lspci -t" output. My understanding is that it's a configuration that works and all devices are accessible. The second PCIe Root Complex is correctly described in ACPI tables enough to be accepted by Linux Kernel. Can I do something to help? Thanks Bertrand