On Fri, 23 Oct 2020 14:53:08 +0200,
Kai-Heng Feng wrote:
>
>
>
> > On Oct 23, 2020, at 19:34, Takashi Iwai <tiwai@xxxxxxx> wrote:
> >
> > On Fri, 23 Oct 2020 12:23:36 +0200,
> > Kai-Heng Feng wrote:
> >>
> >> @@ -1038,14 +1036,6 @@ static int azx_suspend(struct device *dev)
> >> __azx_runtime_suspend(chip);
> >> else
> >> pm_runtime_force_suspend(dev);
> >> - if (bus->irq >= 0) {
> >> - free_irq(bus->irq, chip);
> >> - bus->irq = -1;
> >> - chip->card->sync_irq = -1;
> >> - }
> >
> > This release of irq has nothing to do with MSI. There has been PCI
> > controllers that assign to a different IRQ line after the resume.
>
> Can this issue happened before commit 41017f0cac925 ("[PATCH] PCI: MSI(X) save/restore for suspend/resume") was merged?
It's not about MSI. The IRQ number itself may change after the
resume.
But I guess it's hard to prove it; the system was tad old, and I don't
know who own it now.
Takashi
