drm/xe/pm: Add pci d3cold_capable support

Adding pci d3cold_capable check in order to initialize
d3cold_allowed as false statically.
It avoids vram save/restore latency during runtime
suspend/resume

v2:
- Added else block to xe_pci_runtime_idle. [Rodrigo]

Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Anshuman Gupta <anshuman.gupta@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230718080703.239343-2-anshuman.gupta@intel.comSigned-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drivers/gpu/drm/xe/xe_device_types.h

@@ -327,6 +327,9 @@ struct xe_device {
 		bool hold_rpm;
 	} mem_access;
 
+	/** d3cold_capable: Indicates if root port is d3cold capable */
+	bool d3cold_capable;
+
 	/** @d3cold_allowed: Indicates if d3cold is a valid device state */
 	bool d3cold_allowed;
 

drivers/gpu/drm/xe/xe_pci.c

@@ -665,6 +665,7 @@ static int xe_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		goto err_pci_disable;
 
 	xe_pm_runtime_init(xe);
+	xe_pm_init(xe);
 
 	return 0;
 
@@ -777,18 +778,22 @@ static int xe_pci_runtime_idle(struct device *dev)
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct xe_device *xe = pdev_to_xe_device(pdev);
 
-	/*
-	 * TODO: d3cold should be allowed (true) if
-	 * (IS_DGFX(xe) && !xe_device_mem_access_ongoing(xe))
-	 * but maybe include some other conditions. So, before
-	 * we can re-enable the D3cold, we need to:
-	 * 1. rewrite the VRAM save / restore to avoid buffer object locks
-	 * 2. block D3cold if we have a big amount of device memory in use
-	 *    in order to reduce the latency.
-	 * 3. at resume, detect if we really lost power and avoid memory
-	 *    restoration if we were only up to d3cold
-	 */
-	 xe->d3cold_allowed = false;
+	if (!xe->d3cold_capable) {
+		xe->d3cold_allowed = false;
+	} else {
+		/*
+		 * TODO: d3cold should be allowed (true) if
+		 * (IS_DGFX(xe) && !xe_device_mem_access_ongoing(xe))
+		 * but maybe include some other conditions. So, before
+		 * we can re-enable the D3cold, we need to:
+		 * 1. rewrite the VRAM save / restore to avoid buffer object locks
+		 * 2. block D3cold if we have a big amount of device memory in use
+		 *    in order to reduce the latency.
+		 * 3. at resume, detect if we really lost power and avoid memory
+		 *    restoration if we were only up to d3cold
+		 */
+		xe->d3cold_allowed = false;
+	}
 
 	return 0;
 }

drivers/gpu/drm/xe/xe_pm.c

@@ -106,6 +106,21 @@ int xe_pm_resume(struct xe_device *xe)
 	return 0;
 }
 
+static bool xe_pm_pci_d3cold_capable(struct pci_dev *pdev)
+{
+	struct pci_dev *root_pdev;
+
+	root_pdev = pcie_find_root_port(pdev);
+	if (!root_pdev)
+		return false;
+
+	/* D3Cold requires PME capability and _PR3 power resource */
+	if (!pci_pme_capable(root_pdev, PCI_D3cold) || !pci_pr3_present(root_pdev))
+		return false;
+
+	return true;
+}
+
 void xe_pm_runtime_init(struct xe_device *xe)
 {
 	struct device *dev = xe->drm.dev;
@@ -118,6 +133,13 @@ void xe_pm_runtime_init(struct xe_device *xe)
 	pm_runtime_put_autosuspend(dev);
 }
 
+void xe_pm_init(struct xe_device *xe)
+{
+	struct pci_dev *pdev = to_pci_dev(xe->drm.dev);
+
+	xe->d3cold_capable = xe_pm_pci_d3cold_capable(pdev);
+}
+
 void xe_pm_runtime_fini(struct xe_device *xe)
 {
 	struct device *dev = xe->drm.dev;

drivers/gpu/drm/xe/xe_pm.h

@@ -14,6 +14,7 @@ int xe_pm_suspend(struct xe_device *xe);
 int xe_pm_resume(struct xe_device *xe);
 
 void xe_pm_runtime_init(struct xe_device *xe);
+void xe_pm_init(struct xe_device *xe);
 void xe_pm_runtime_fini(struct xe_device *xe);
 int xe_pm_runtime_suspend(struct xe_device *xe);
 int xe_pm_runtime_resume(struct xe_device *xe);