Web lists-archives.com

[PATCH 4.19 122/148] PCI / PM: Allow runtime PM without callback functions

4.19-stable review patch.  If anyone has any objections, please let me know.


From: Jarkko Nikula <jarkko.nikula@xxxxxxxxxxxxxxx>

commit c5eb1190074cfb14c5d9cac692f1912eecf1a5e4 upstream.

a9c8088c7988 ("i2c: i801: Don't restore config registers on runtime PM")
nullified the runtime PM suspend/resume callback pointers while keeping the
runtime PM enabled.

This caused the SMBus PCI device to stay in D0 with
/sys/devices/.../power/runtime_status showing "error" when the runtime PM
framework attempted to autosuspend the device.  This is due to PCI bus
runtime PM, which checks for driver runtime PM callbacks and returns
-ENOSYS if they are not set.

Since i2c-i801.c doesn't need to do anything device-specific for runtime
PM, Jean Delvare proposed this be fixed in the PCI core rather than adding
dummy runtime PM callback functions in the PCI drivers.

Change pci_pm_runtime_suspend()/pci_pm_runtime_resume() so they allow
changing the PCI device power state during runtime PM transitions even if
the driver supplies no runtime PM callbacks.

This fixes the runtime PM regression on i2c-i801.c.

It is not obvious why the code previously required the runtime PM
callbacks.  The test has been there since the code was introduced by
6cbf82148ff2 ("PCI PM: Run-time callbacks for PCI bus type").

On the other hand, a similar change was done to generic runtime PM
callbacks in 05aa55dddb9e ("PM / Runtime: Lenient generic runtime pm

Fixes: a9c8088c7988 ("i2c: i801: Don't restore config registers on runtime PM")
Reported-by: Mika Westerberg <mika.westerberg@xxxxxxxxxxxxxxx>
Signed-off-by: Jarkko Nikula <jarkko.nikula@xxxxxxxxxxxxxxx>
Signed-off-by: Bjorn Helgaas <bhelgaas@xxxxxxxxxx>
Reviewed-by: Jean Delvare <jdelvare@xxxxxxx>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
Cc: stable@xxxxxxxxxxxxxxx	# v4.18+
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

 drivers/pci/pci-driver.c |   27 ++++++++++++---------------
 1 file changed, 12 insertions(+), 15 deletions(-)

--- a/drivers/pci/pci-driver.c
+++ b/drivers/pci/pci-driver.c
@@ -1251,30 +1251,29 @@ static int pci_pm_runtime_suspend(struct
 		return 0;
-	if (!pm || !pm->runtime_suspend)
-		return -ENOSYS;
 	pci_dev->state_saved = false;
-	error = pm->runtime_suspend(dev);
-	if (error) {
+	if (pm && pm->runtime_suspend) {
+		error = pm->runtime_suspend(dev);
 		 * -EBUSY and -EAGAIN is used to request the runtime PM core
 		 * to schedule a new suspend, so log the event only with debug
 		 * log level.
-		if (error == -EBUSY || error == -EAGAIN)
+		if (error == -EBUSY || error == -EAGAIN) {
 			dev_dbg(dev, "can't suspend now (%pf returned %d)\n",
 				pm->runtime_suspend, error);
-		else
+			return error;
+		} else if (error) {
 			dev_err(dev, "can't suspend (%pf returned %d)\n",
 				pm->runtime_suspend, error);
-		return error;
+			return error;
+		}
 	pci_fixup_device(pci_fixup_suspend, pci_dev);
-	if (!pci_dev->state_saved && pci_dev->current_state != PCI_D0
+	if (pm && pm->runtime_suspend
+	    && !pci_dev->state_saved && pci_dev->current_state != PCI_D0
 	    && pci_dev->current_state != PCI_UNKNOWN) {
 		WARN_ONCE(pci_dev->current_state != prev,
 			"PCI PM: State of device not saved by %pF\n",
@@ -1292,7 +1291,7 @@ static int pci_pm_runtime_suspend(struct
 static int pci_pm_runtime_resume(struct device *dev)
-	int rc;
+	int rc = 0;
 	struct pci_dev *pci_dev = to_pci_dev(dev);
 	const struct dev_pm_ops *pm = dev->driver ? dev->driver->pm : NULL;
@@ -1306,14 +1305,12 @@ static int pci_pm_runtime_resume(struct
 	if (!pci_dev->driver)
 		return 0;
-	if (!pm || !pm->runtime_resume)
-		return -ENOSYS;
 	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);
-	rc = pm->runtime_resume(dev);
+	if (pm && pm->runtime_resume)
+		rc = pm->runtime_resume(dev);
 	pci_dev->runtime_d3cold = false;