PCI: PM: Skip devices in D0 for suspend-to-idle

Commit d491f2b75237 ("PCI: PM: Avoid possible suspend-to-idle issue")
attempted to avoid a problem with devices whose drivers want them to
stay in D0 over suspend-to-idle and resume, but it did not go as far
as it should with that.

Namely, first of all, the power state of a PCI bridge with a
downstream device in D0 must be D0 (based on the PCI PM spec r1.2,
sec 6, table 6-1, if the bridge is not in D0, there can be no PCI
transactions on its secondary bus), but that is not actively enforced
during system-wide PM transitions, so use the skip_bus_pm flag
introduced by commit d491f2b75237 for that.

Second, the configuration of devices left in D0 (whatever the reason)
during suspend-to-idle need not be changed and attempting to put them
into D0 again by force is pointless, so explicitly avoid doing that.

Fixes: d491f2b75237 ("PCI: PM: Avoid possible suspend-to-idle issue")
Reported-by: Kai-Heng Feng <[e-mail]>
Signed-off-by: Rafael J. Wysocki <[e-mail]>
Reviewed-by: Mika Westerberg <[e-mail]>
Tested-by: Kai-Heng Feng <[e-mail]>

drivers/pci/pci-driver.c

···
 	pci_power_up(pci_dev);
 	pci_restore_state(pci_dev);
 	pci_pme_restore(pci_dev);
-	pci_fixup_device(pci_fixup_resume_early, pci_dev);
 }
 
 /*
···
 
 	if (pci_dev->skip_bus_pm) {
 		/*
-		 * The function is running for the second time in a row without
+		 * Either the device is a bridge with a child in D0 below it, or
+		 * the function is running for the second time in a row without
 		 * going through full resume, which is possible only during
-		 * suspend-to-idle in a spurious wakeup case.  Moreover, the
-		 * device was originally left in D0, so its power state should
-		 * not be changed here and the device register values saved
-		 * originally should be restored on resume again.
+		 * suspend-to-idle in a spurious wakeup case.  The device should
+		 * be in D0 at this point, but if it is a bridge, it may be
+		 * necessary to save its state.
 		 */
-		pci_dev->state_saved = true;
-	} else if (pci_dev->state_saved) {
-		if (pci_dev->current_state == PCI_D0)
-			pci_dev->skip_bus_pm = true;
-	} else {
+		if (!pci_dev->state_saved)
+			pci_save_state(pci_dev);
+	} else if (!pci_dev->state_saved) {
 		pci_save_state(pci_dev);
 		if (pci_power_manageable(pci_dev))
 			pci_prepare_to_sleep(pci_dev);
···
 
 	dev_dbg(dev, "PCI PM: Suspend power state: %s\n",
 		pci_power_name(pci_dev->current_state));
+
+	if (pci_dev->current_state == PCI_D0) {
+		pci_dev->skip_bus_pm = true;
+		/*
+		 * Per PCI PM r1.2, table 6-1, a bridge must be in D0 if any
+		 * downstream device is in D0, so avoid changing the power state
+		 * of the parent bridge by setting the skip_bus_pm flag for it.
+		 */
+		if (pci_dev->bus->self)
+			pci_dev->bus->self->skip_bus_pm = true;
+	}
+
+	if (pci_dev->skip_bus_pm && !pm_suspend_via_firmware()) {
+		dev_dbg(dev, "PCI PM: Skipped\n");
+		goto Fixup;
+	}
 
 	pci_pm_set_unknown_state(pci_dev);
 
···
 	if (dev_pm_smart_suspend_and_suspended(dev))
 		pm_runtime_set_active(dev);
 
-	pci_pm_default_resume_early(pci_dev);
+	/*
+	 * In the suspend-to-idle case, devices left in D0 during suspend will
+	 * stay in D0, so it is not necessary to restore or update their
+	 * configuration here and attempting to put them into D0 again may
+	 * confuse some firmware, so avoid doing that.
+	 */
+	if (!pci_dev->skip_bus_pm || pm_suspend_via_firmware())
+		pci_pm_default_resume_early(pci_dev);
+
+	pci_fixup_device(pci_fixup_resume_early, pci_dev);
 
 	if (pci_has_legacy_pm_support(pci_dev))
 		return pci_legacy_resume_early(dev);
···
 	}
 
 	pci_pm_default_resume_early(pci_dev);
+	pci_fixup_device(pci_fixup_resume_early, pci_dev);
 
 	if (pci_has_legacy_pm_support(pci_dev))
 		return pci_legacy_resume_early(dev);