NV50 backlight brightness

NV50 backlight brightness

[Aaron Sowry]

[-- Attachment #1.1: Type: text/plain, Size: 1117 bytes --]

Hi,

The NV50 codepath in nouveau_backlight.c invariably sets max_brightness to 1025, regardless of the chipset in question. The maximum brightness is not 1025 for all NV50 chipsets (for example, by setting the brightness via ACPI controls and peeking the brightness register, I find that the maximum brightness level for my a8 card is 0x4001df67 although I am unsure of the significance of the MSB).

This means when anything manipulates the backlight using nouveau controls (in my case, usually g-p-m), I get very low backlight brightnesses. 2 questions:

1) Is this a known problem, and has it been looked at before? My affected laptop is a ThinkPad and I know that some Macbooks are also affected, so I would be surprised if it isn't fairly common.

2) In the absence of a reliable way to determine a card's max_brightness by peeking registers or similar, is some kind of mapping table in the code an acceptable solution for assigning max_brightness values to certain chipsets?

This has been discussed at some length on the RedHat bugzilla here:
https://bugzilla.redhat.com/show_bug.cgi?id=625171

Cheers,
Aaron

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 489 bytes --]

[-- Attachment #2: Type: text/plain, Size: 181 bytes --]

_______________________________________________
Nouveau mailing list
Nouveau-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: NV50 backlight brightness

[Ben Skeggs]

On Tue, 2010-10-19 at 11:10 +0200, Aaron Sowry wrote:
> Hi,
> 
> The NV50 codepath in nouveau_backlight.c invariably sets max_brightness to 1025, regardless of the chipset in question. The maximum brightness is not 1025 for all NV50 chipsets (for example, by setting the brightness via ACPI controls and peeking the brightness register, I find that the maximum brightness level for my a8 card is 0x4001df67 although I am unsure of the significance of the MSB).
> 
> This means when anything manipulates the backlight using nouveau controls (in my case, usually g-p-m), I get very low backlight brightnesses. 2 questions:
> 
> 1) Is this a known problem, and has it been looked at before? My affected laptop is a ThinkPad and I know that some Macbooks are also affected, so I would be surprised if it isn't fairly common.
> 
> 2) In the absence of a reliable way to determine a card's max_brightness by peeking registers or similar, is some kind of mapping table in the code an acceptable solution for assigning max_brightness values to certain chipsets?
I have a question to you that's important first:  if you fix nouveau's
max_brightness level, can you actually properly change the brightness?

I have an eDP laptop that shows the same backlight level regardless of
what's in that register.

Ben.
> 
> This has been discussed at some length on the RedHat bugzilla here:
> https://bugzilla.redhat.com/show_bug.cgi?id=625171
> 
> Cheers,
> Aaron
> _______________________________________________
> Nouveau mailing list
> Nouveau-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
> http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: NV50 backlight brightness

[Aaron Sowry]

[-- Attachment #1.1.1: Type: text/plain, Size: 580 bytes --]

On Wed, Oct 20, 2010 at 09:17:39PM +1000, Ben Skeggs wrote:
> I have a question to you that's important first:  if you fix nouveau's
> max_brightness level, can you actually properly change the brightness?
> 
> I have an eDP laptop that shows the same backlight level regardless of
> what's in that register.
> 

By applying the attached patch I can properly change the brightness levels using g-p-m. The only question is what the 0x40000000 bit represents as set by BIOS/ACPI - both 0x40001fff and 0x00001fff seem to work fine as far as brightness levels are concerned.

[-- Attachment #1.1.2: linux-2.6.35-20100921.patch --]
[-- Type: text/x-diff, Size: 781 bytes --]

diff -uNrp kernel-2.6.35.fc14.orig/drivers/gpu/drm/nouveau/nouveau_backlight.c kernel-2.6.35.fc14.new/drivers/gpu/drm/nouveau/nouveau_backlight.c
--- kernel-2.6.35.fc14.orig/drivers/gpu/drm/nouveau/nouveau_backlight.c	2010-09-10 14:04:09.301972164 +0200
+++ kernel-2.6.35.fc14.new/drivers/gpu/drm/nouveau/nouveau_backlight.c	2010-10-21 10:04:03.709309995 +0200
@@ -120,7 +120,11 @@ static int nouveau_nv50_backlight_init(s
 		return 0;
 
 	memset(&props, 0, sizeof(struct backlight_properties));
-	props.max_brightness = 1025;
+	if (dev_priv->chipset >= 0xa8)  
+		props.max_brightness = 0x0001ffff;
+	else
+		props.max_brightness = 1025;
+
 	bd = backlight_device_register("nv_backlight", &dev->pdev->dev, dev,
 				       &nv50_bl_ops, &props);
 	if (IS_ERR(bd))

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 489 bytes --]

[-- Attachment #2: Type: text/plain, Size: 181 bytes --]

_______________________________________________
Nouveau mailing list
Nouveau-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: NV50 backlight brightness

[Ben Skeggs]

On Thu, 2010-10-21 at 11:16 +0200, Aaron Sowry wrote:
> On Wed, Oct 20, 2010 at 09:17:39PM +1000, Ben Skeggs wrote:
> > I have a question to you that's important first:  if you fix nouveau's
> > max_brightness level, can you actually properly change the brightness?
> > 
> > I have an eDP laptop that shows the same backlight level regardless of
> > what's in that register.
> > 
> 
> By applying the attached patch I can properly change the brightness levels using g-p-m. The only question is what the 0x40000000 bit represents as set by BIOS/ACPI - both 0x40001fff and 0x00001fff seem to work fine as far as brightness levels are concerned.
Ok.  I'll look at all the kmmio/vbios data I have tomorrow, and come up
with some kind of "solution", it'll likely require tweaking as we see
more data.

What value does 0x610084 contain on boot for you?

Ben.

Re: NV50 backlight brightness

[Aaron Sowry]

[-- Attachment #1.1.1: Type: text/plain, Size: 376 bytes --]

On Thu, Oct 21, 2010 at 08:45:20PM +1000, Ben Skeggs wrote:
> What value does 0x610084 contain on boot for you?
> 
Using the attached patch (hopefully early enough on in the initialisation):

[aaron@jules ~]$ dmesg | grep NV50
[drm] nouveau 0000:01:00.0: Detected an NV50 generation card (0x0a8600a2)
NV50 0x00610084: 0xa0000000

What's the relevance of this value?

[-- Attachment #1.1.2: reg_peek.patch --]
[-- Type: text/x-diff, Size: 572 bytes --]

--- /home/aaron/rpmbuild/BUILD/kernel-2.6.35.fc14.orig/drivers/gpu/drm/nouveau/nv50_display.c	2010-09-09 23:00:38.774029000 +0200
+++ /home/aaron/rpmbuild/BUILD/kernel-2.6.35.fc14/linux-2.6.35.x86_64/drivers/gpu/drm/nouveau/nv50_display.c	2010-10-22 11:36:32.252185209 +0200
@@ -205,6 +205,8 @@ nv50_display_init(struct drm_device *dev
 
 	NV_DEBUG_KMS(dev, "\n");
 
+	printk("NV50 0x00610084: 0x%08x\n", nv_rd32(dev, 0x00610084));
+
 	nv_wr32(dev, 0x00610184, nv_rd32(dev, 0x00614004));
 	/*
 	 * I think the 0x006101XX range is some kind of main control area

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 489 bytes --]

[-- Attachment #2: Type: text/plain, Size: 181 bytes --]

_______________________________________________
Nouveau mailing list
Nouveau-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
http://lists.freedesktop.org/mailman/listinfo/nouveau

Re: NV50 backlight brightness

[Ben Skeggs]

Sent from my iPhone

On 22/10/2010, at 19:57, Aaron Sowry <aaron+nv-+1tCnOwpXBmzQB+pC5nmwQ@public.gmane.org> wrote:

> On Thu, Oct 21, 2010 at 08:45:20PM +1000, Ben Skeggs wrote:
>> What value does 0x610084 contain on boot for you?
>> 
> Using the attached patch (hopefully early enough on in the initialisation):
> 
> [aaron@jules ~]$ dmesg | grep NV50
> [drm] nouveau 0000:01:00.0: Detected an NV50 generation card (0x0a8600a2)
> NV50 0x00610084: 0xa0000000
> 
> What's the relevance of this value?
Oops, nothing, that was a typo. 0x61c084 is what I meant :)

> <reg_peek.patch>

Re: NV50 backlight brightness

[Aaron Sowry]

[-- Attachment #1.1: Type: text/plain, Size: 214 bytes --]

On Fri, Oct 22, 2010 at 09:17:37PM +1000, Ben Skeggs wrote:
> > What's the relevance of this value?
> Oops, nothing, that was a typo. 0x61c084 is what I meant :)
> 

Heh. It's 0x40012ccd initially.

/Aaron

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 489 bytes --]

[-- Attachment #2: Type: text/plain, Size: 181 bytes --]

_______________________________________________
Nouveau mailing list
Nouveau-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
http://lists.freedesktop.org/mailman/listinfo/nouveau