[PATCH] Revert "PCI/PM: Remove pci_pm_complete()"

[PATCH] Revert "PCI/PM: Remove pci_pm_complete()"

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
    
Revert commit 84822b158fd3 "PCI/PM: Remove pci_pm_complete()" that
removed the execution of the device driver's ->complete callback
from the PCI bus type.
    
That was a mistake, because those callbacks are not executed at all
now even if present.
    
Fixes: 84822b158fd3 "PCI/PM: Remove pci_pm_complete()"
Cc: 3.13+ <stable@vger.kernel.org> # 3.13+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c
index 3cb2210de553..75c6aaec5c9b 100644
--- a/drivers/pci/pci-driver.c
+++ b/drivers/pci/pci-driver.c
@@ -669,10 +669,18 @@ static int pci_pm_prepare(struct device *dev)
 	return pci_dev_keep_suspended(to_pci_dev(dev));
 }
 
+static void pci_pm_complete(struct device *dev)
+{
+	struct device_driver *drv = dev->driver;
+
+	if (drv && drv->pm && drv->pm->complete)
+		drv->pm->complete(dev);
+}
 
 #else /* !CONFIG_PM_SLEEP */
 
 #define pci_pm_prepare	NULL
+#define pci_pm_complete	NULL
 
 #endif /* !CONFIG_PM_SLEEP */
 
@@ -1203,6 +1211,7 @@ static int pci_pm_runtime_idle(struct device *dev)
 
 static const struct dev_pm_ops pci_dev_pm_ops = {
 	.prepare = pci_pm_prepare,
+	.complete = pci_pm_complete,
 	.suspend = pci_pm_suspend,
 	.resume = pci_pm_resume,
 	.freeze = pci_pm_freeze,

Re: [PATCH] Revert "PCI/PM: Remove pci_pm_complete()"

[Rafael J. Wysocki]

On Thu, May 28, 2015 at 4:12 AM, Rafael J. Wysocki <rjw@rjwysocki.net> wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
>
> Revert commit 84822b158fd3 "PCI/PM: Remove pci_pm_complete()" that
> removed the execution of the device driver's ->complete callback
> from the PCI bus type.
>
> That was a mistake, because those callbacks are not executed at all
> now even if present.
>
> Fixes: 84822b158fd3 "PCI/PM: Remove pci_pm_complete()"
> Cc: 3.13+ <stable@vger.kernel.org> # 3.13+
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

Scratch this, sorry.

The code is correct as is, because the bus type doesn't provide
->complete at all in which case the core will run the driver's
callback directly.

We'll probably need ->complete for PCI at one point, but then I'll
resubmit this one with a different changelog.

Thanks,
Rafael