Nouveau - Aug 2016 - [PATCH] drm/nouveau/acpi: use DSM if bridge does not support D3cold

Even if PR3 support is available on the bridge, it will not be used if
the PCI layer considers it unavailable (i.e. on all laptops from 2013
and 2014). Ensure that this condition is checked to allow a fallback to
the Optimus DSM for device poweroff.

Initially I wanted to call pci_d3cold_enable before checking bridge_d3
(in case the user changed d3cold_allowed), but that is such an unlikely
case and likely fragile anyway. The current patch is suggested by Mika
in http://www.spinics.net/lists/linux-pci/msg52599.html

Cc: Mika Westerberg <mika.westerberg at linux.intel.com>
Signed-off-by: Peter Wu <peter at lekensteyn.nl>
---
Hi,

This idea is floating since July, but was blocked by the PCI D3cold patches.
Since these got into -rc1, here is the promised follow-up patch for v4.8.

Without this patch, some laptops from 2013 and 2014 will regress in v4.8 and
suck more power.

Kind regards,
Peter
---
 drivers/gpu/drm/nouveau/nouveau_acpi.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.c
b/drivers/gpu/drm/nouveau/nouveau_acpi.c
index f2ad17a..dc57b62 100644
--- a/drivers/gpu/drm/nouveau/nouveau_acpi.c
+++ b/drivers/gpu/drm/nouveau/nouveau_acpi.c
@@ -225,6 +225,17 @@ static bool nouveau_pr3_present(struct pci_dev *pdev)
 	if (!parent_pdev)
 		return false;
 
+	if (!parent_pdev->bridge_d3) {
+		/*
+		 * Parent PCI bridge is currently not power managed.
+		 * Since userspace can change these afterwards to be on
+		 * the safe side we stick with _DSM and prevent usage of
+		 * _PR3 from the bridge.
+		 */
+		pci_d3cold_disable(pdev);
+		return false;
+	}
+
 	parent_adev = ACPI_COMPANION(&parent_pdev->dev);
 	if (!parent_adev)
 		return false;
-- 
2.9.3

On Fri, Aug 26, 2016 at 01:00:54AM +0200, Peter Wu
wrote:
> Even if PR3 support is available on the bridge, it will not be used if
> the PCI layer considers it unavailable (i.e. on all laptops from 2013
> and 2014). Ensure that this condition is checked to allow a fallback to
> the Optimus DSM for device poweroff.
> 
> Initially I wanted to call pci_d3cold_enable before checking bridge_d3
> (in case the user changed d3cold_allowed), but that is such an unlikely
> case and likely fragile anyway. The current patch is suggested by Mika
> in http://www.spinics.net/lists/linux-pci/msg52599.html
> 
> Cc: Mika Westerberg <mika.westerberg at linux.intel.com>
Reviewed-by: Mika Westerberg <mika.westerberg at linux.intel.com>