Skip to contentSkip to conversation

drm/i915: Allow minimum brightness upon backlight enable

PATCH In review [email protected] 3 participants 4 messages Oct 1, 2015 – Oct 20, 2015

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/i915: Allow minimum brightness upon backlight enable

From: Clint Taylor <[email protected]>

backlight minimum is a valid value so you don't need to set maximum.

---
 drivers/gpu/drm/i915/intel_panel.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_panel.c +1 -1
index 4d28c7b..d509904 100644
@@ -1069,7 +1069,7 @@ void intel_panel_enable_backlight(struct intel_connector *connector)
 
 	WARN_ON(panel->backlight.max == 0);
 
-	if (panel->backlight.level <= panel->backlight.min) {
+	if (panel->backlight.level < panel->backlight.min) {
 		panel->backlight.level = panel->backlight.max;
 		if (panel->backlight.device)
 			panel->backlight.device->props.brightness =
Signature · 6 lines
1.7.9.5

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
maintainer
#1 lore raw
Quoted 14 lines · On Fri, 02 Oct 2015, [email protected] wrote:

From: Clint Taylor <[email protected]>

backlight minimum is a valid value so you don't need to set maximum.

---
 drivers/gpu/drm/i915/intel_panel.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_panel.c +0 -0
index 4d28c7b..d509904 100644

drivers/gpu/drm/i915/intel_panel.c void intel_panel_enable_backlight(struct intel_connector *connector) · line 1075

Quoted 8 lines
drivers/gpu/drm/i915/intel_panel.c +1 -1
@@ -1069,7 +1069,7 @@ void intel_panel_enable_backlight(struct intel_connector *connector)
 
 	WARN_ON(panel->backlight.max == 0);
 
-	if (panel->backlight.level <= panel->backlight.min) {
+	if (panel->backlight.level < panel->backlight.min) {
 		panel->backlight.level = panel->backlight.max;
 		if (panel->backlight.device)
 			panel->backlight.device->props.brightness =

This is policy in action in the kernel. We have this whole if block here because some, uh, misguided userspace sets brightness to minimum before switching the backlight off, and assumes enabling backlight cranks up the brightness as well. (*)

The condition used to be (level == 0), but since the minimum can now also be non-zero, we check (level <= min). If we changed this to (level < min), we'd regress the misguided userspaces, especially if min == 0.

See

commit 13f3fbe827d09e3182023c8c54058cbf97aa146e

Author: Jeremiah Mahler <[email protected]>
Date:   Mon Jan 12 11:01:03 2015 -0800

drm/i915: fix inconsistent brightness after resume

I'd really love to to just nuke the if block completely, but no regressions and all.

BR,
Jani.

(*) Imagine an amplifier that would set the volume to max if it was powered on with the volume knob turned to zero...

Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#2 lore raw
Quoted 18 lines · Jani Nikula
Quoted 17 lines · On Fri, 02 Oct 2015, [email protected] wrote:

From: Clint Taylor <[email protected]>

backlight minimum is a valid value so you don't need to set maximum.

---
 drivers/gpu/drm/i915/intel_panel.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_panel.c +0 -0
index 4d28c7b..d509904 100644
@@ -1069,7 +1069,7 @@ void intel_panel_enable_backlight(struct intel_connector *connector)
 
 	WARN_ON(panel->backlight.max == 0);
 
Quoted 10 lines

drivers/gpu/drm/i915/intel_panel.c void intel_panel_enable_backlight(struct intel_connector *connector) · line 1075

Quoted 5 lines
drivers/gpu/drm/i915/intel_panel.c +1 -1
-	if (panel->backlight.level <= panel->backlight.min) {
+	if (panel->backlight.level < panel->backlight.min) {
 		panel->backlight.level = panel->backlight.max;
 		if (panel->backlight.device)
 			panel->backlight.device->props.brightness =

This is policy in action in the kernel. We have this whole if block here because some, uh, misguided userspace sets brightness to minimum before switching the backlight off, and assumes enabling backlight cranks up the brightness as well. (*)

If you mean X, no. It has to set the backlight because on some intel models the backlight is not routed through i915 and so it needs to explicitly control it. On resume, it sets the previous value. This is just a broken kernel.
-Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
maintainer
#3 lore raw
Quoted 30 lines · Chris Wilson
Quoted 18 lines · Jani Nikula
Quoted 17 lines · On Fri, 02 Oct 2015, [email protected] wrote:

From: Clint Taylor <[email protected]>

backlight minimum is a valid value so you don't need to set maximum.

---
 drivers/gpu/drm/i915/intel_panel.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_panel.c +0 -0
index 4d28c7b..d509904 100644
@@ -1069,7 +1069,7 @@ void intel_panel_enable_backlight(struct intel_connector *connector)
 
 	WARN_ON(panel->backlight.max == 0);
 
Quoted 10 lines

drivers/gpu/drm/i915/intel_panel.c void intel_panel_enable_backlight(struct intel_connector *connector) · line 1075

Quoted 5 lines
drivers/gpu/drm/i915/intel_panel.c +1 -1
-	if (panel->backlight.level <= panel->backlight.min) {
+	if (panel->backlight.level < panel->backlight.min) {
 		panel->backlight.level = panel->backlight.max;
 		if (panel->backlight.device)
 			panel->backlight.device->props.brightness =

This is policy in action in the kernel. We have this whole if block here because some, uh, misguided userspace sets brightness to minimum before switching the backlight off, and assumes enabling backlight cranks up the brightness as well. (*)

Quoted 4 lines

If you mean X, no. It has to set the backlight because on some intel models the backlight is not routed through i915 and so it needs to explicitly control it. On resume, it sets the previous value. This is just a broken kernel.

Okay, I guess I'm willing to give this a shot. This may result in regression reports, but let's see.

However, the meaning of the whole check changes. What Clint now checks (level below minimum) shouldn't happen, something failed already. Maybe we should set the level to minimum instead, with debug logging?

BR,
Jani.

Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx