Re: [PATCH v2 0/3] PCI/PM: Do not save the config space of an inaccessible device

["Francisco Beltrán Millalén" <fbeltranmillalen@gmail.com>]

Hi Bjorn,

Thanks for reviewing the series.

On Thu, Oct 08, 2026 at 05:34:11PM -0500, Bjorn Helgaas wrote:
> If it would be safe to apply patches 2 and 3 without patch 1, I could
> go ahead and do that.
>
> Patch 2 returns errors from pci_save_state() in more cases, but
> hcd_pci_suspend_noirq() doesn't check for errors anyway, so I think
> it's would be no worse off it we applied patch 2 without patch 1.

I tested that combination on the MacBookPro14,3: patch 2 as you suggest
in your other reply, patch 1 disabled, the Alpine Ridge quirk applied,
and nothing plugged into one of the two Thunderbolt controllers.  The
system resumed normally.  The only difference was this warning, once
per boot:

  xhci_hcd 0000:07:00.0: Unable to change power state from D0 to D3hot, device inaccessible
  xhci_hcd 0000:07:00.0: PCI PM: State of device not saved by hcd_pci_suspend_noirq+0x0/0x1c0
  WARNING: CPU: 2 PID: 4344 at drivers/pci/pci-driver.c:888 pci_pm_suspend_noirq+0x2f4/0x300

hcd_pci_suspend_noirq() goes on to pci_prepare_to_sleep(), which fails
on the inaccessible controller and leaves it marked D3cold, and
pci_pm_suspend_noirq() then warns.  With patch 1 there is no warning.

So 2 and 3 can go in first, with patch 1 following through your tree
with an ack from Alan or Greg, or through the USB tree.  Patch 1 builds
on its own and has no effect in practice until patch 2 is in.

I'll send a v3 with your changes to patch 2.

Francisco