Re: [Bug 213873] New: Hotplug ethernet cable to runtime suspended Realtek NIC doesn't raise PCIe PME IRQ on Intel ADL

[Bjorn Helgaas <helgaas@kernel.org>]

On Tue, Jul 27, 2021 at 06:10:13AM +0000, bugzilla-daemon@bugzilla.kernel.org wrote:
> https://bugzilla.kernel.org/show_bug.cgi?id=213873
> 
>             Bug ID: 213873
>            Summary: Hotplug ethernet cable to runtime suspended Realtek
>                     NIC doesn't raise PCIe PME IRQ on Intel ADL
>            Product: Drivers
>            Version: 2.5
>     Kernel Version: mainline
>           Hardware: All
>                 OS: Linux
>               Tree: Mainline
>             Status: NEW
>           Severity: normal
>           Priority: P1
>          Component: PCI
>           Assignee: drivers_pci@kernel-bugs.osdl.org
>           Reporter: kai.heng.feng@canonical.com
>         Regression: No
> 
> Both PCIe bridge and Realtek NIC are runtime suspended. When ethernet cable is
> plugged, PMEStatus and PMEPending are marked:
> RootSta: PME ReqID 0200, PMEStatus+ PMEPending+
> 
> But the IRQ isn't raised:
> $ cat /proc/interrupts  | grep PME
>  145:          0          0          0          0          0          0        
>  0          0          0          0          0          0   VMD-MSI  126  PCIe
> PME, aerdrv, pcie-dpc
> 
> So the hotplug event isn't detected.
> 
> -- 

From the lspci after cable hotplug (comment #3):

  0000:00:1c.0 PCI bridge ...
    RootCtl: ErrCorrectable- ErrNon-Fatal- ErrFatal- PMEIntEna- CRSVisible-
    RootSta: PME ReqID 0200, PMEStatus+ PMEPending+

So it looks like the root port is not enabled to generate a PME
interrupt.  pcie_pme_interrupt_enable() looks like it *could* enable
that.  From a quick look, my guess is that pme.c doesn't claim this
root port because the platform says it doesn't support it:

  acpi PNP0A08:00: _OSC: platform does not support [PCIeHotplug SHPCHotplug PME]
  acpi PNP0A08:00: _OSC: OS now controls [AER PCIeCapability LTR]

Does booting with "pcie_ports=native" make it work?  That's not a very
good solution, of course, because then both the firmware and Linux
think they own this functionality.

Is there a BIOS update that would fix this?

Bjorn