On Tuesday, October 15, 2019 1:00:13 AM CEST Bjorn Helgaas wrote:
> From: Bjorn Helgaas
>
> pci_pm_resume() and pci_pm_restore() call pci_pm_default_resume(), which
> runs resume fixups before disabling wakeup events:
>
> static void pci_pm_default_resume(struct pci_dev *pci_dev)
> {
> pci_fixup_device(pci_fixup_resume, pci_dev);
> pci_enable_wake(pci_dev, PCI_D0, false);
> }
>
> pci_pm_runtime_resume() does both of these, but in the opposite order:
>
> pci_enable_wake(pci_dev, PCI_D0, false);
> pci_fixup_device(pci_fixup_resume, pci_dev);
>
> We should always use the same ordering unless there's a reason to do
> otherwise.
Right.
> Change pci_pm_runtime_resume() to call pci_pm_default_resume()
> instead of open-coding this, so the fixups are always done before disabling
> wakeup events.
>
> Signed-off-by: Bjorn Helgaas
No concerns about this change, so
Reviewed-by: Rafael J. Wysocki
> ---
> drivers/pci/pci-driver.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c
> index 0c3086793e4e..55acb658273f 100644
> --- a/drivers/pci/pci-driver.c
> +++ b/drivers/pci/pci-driver.c
> @@ -1345,8 +1345,7 @@ static int pci_pm_runtime_resume(struct device *dev)
> return 0;
>
> pci_fixup_device(pci_fixup_resume_early, pci_dev);
> - pci_enable_wake(pci_dev, PCI_D0, false);
> - pci_fixup_device(pci_fixup_resume, pci_dev);
> + pci_pm_default_resume(pci_dev);
>
> if (pm && pm->runtime_resume)
> rc = pm->runtime_resume(dev);
>
___
devel mailing list
de...@linuxdriverproject.org
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
