2.6.21-rc2 radeon backlight

2.6.21-rc2 radeon backlight

[Alex Romosan]

the backlight on my thinkpad still (2.6.20 worked fine) doesn't come
on if i have the radeon backlight enabled. without it, i guess it's
the ibm acpi modules that controls the backlight and it seems to work
fine.

--alex--

-- 
| I believe the moment is at hand when, by a paranoiac and active |
|  advance of the mind, it will be possible (simultaneously with  |
|  automatism and other passive states) to systematize confusion  |
|  and thus to help to discredit completely the world of reality. |

Re: 2.6.21-rc2 radeon backlight

[Andrew Morton]

On Wed, 28 Feb 2007 08:32:43 -0800
Alex Romosan <romosan@sycorax.lbl.gov> wrote:

> the backlight on my thinkpad still (2.6.20 worked fine) doesn't come
> on if i have the radeon backlight enabled. without it, i guess it's
> the ibm acpi modules that controls the backlight and it seems to work
> fine.
> 

Unclear.  Are you saying that the backlight comes on OK if you use the IBM
acpi module?

Re: 2.6.21-rc2 radeon backlight

[Alex Romosan]

Andrew Morton <akpm@linux-foundation.org> writes:

> On Wed, 28 Feb 2007 08:32:43 -0800
> Alex Romosan <romosan@sycorax.lbl.gov> wrote:
>
>> the backlight on my thinkpad still (2.6.20 worked fine) doesn't come
>> on if i have the radeon backlight enabled. without it, i guess it's
>> the ibm acpi modules that controls the backlight and it seems to work
>> fine.
>> 
>
> Unclear. Are you saying that the backlight comes on OK if you use
> the IBM acpi module?

yes, if i disable the radeon backlight and use the ibm acpi module,
than the backlight works. if i enable the radeon backlight, the screen
stays dark and i can't turn it on (i tried using radeontool to control
it but nothing happened).

--alex--

-- 
| I believe the moment is at hand when, by a paranoiac and active |
|  advance of the mind, it will be possible (simultaneously with  |
|  automatism and other passive states) to systematize confusion  |
|  and thus to help to discredit completely the world of reality. |

Re: 2.6.21-rc2 radeon backlight

[Andrew Morton]

On Fri, 02 Mar 2007 09:24:03 -0800
Alex Romosan <romosan@sycorax.lbl.gov> wrote:

> Andrew Morton <akpm@linux-foundation.org> writes:
> 
> > On Wed, 28 Feb 2007 08:32:43 -0800
> > Alex Romosan <romosan@sycorax.lbl.gov> wrote:
> >
> >> the backlight on my thinkpad still (2.6.20 worked fine) doesn't come
> >> on if i have the radeon backlight enabled. without it, i guess it's
> >> the ibm acpi modules that controls the backlight and it seems to work
> >> fine.
> >> 
> >
> > Unclear. Are you saying that the backlight comes on OK if you use
> > the IBM acpi module?
> 
> yes, if i disable the radeon backlight and use the ibm acpi module,
> than the backlight works. if i enable the radeon backlight, the screen
> stays dark and i can't turn it on (i tried using radeontool to control
> it but nothing happened).

Richard, is this actually a bug, or is it a config error or something like that?

And should we track it as a post-2.6.20 regression?

Re: 2.6.21-rc2 radeon backlight

[David Miller]

From: Andrew Morton <akpm@linux-foundation.org>
Date: Fri, 2 Mar 2007 12:29:20 -0800

> On Fri, 02 Mar 2007 09:24:03 -0800
> Alex Romosan <romosan@sycorax.lbl.gov> wrote:
> 
> > Andrew Morton <akpm@linux-foundation.org> writes:
> > 
> > > On Wed, 28 Feb 2007 08:32:43 -0800
> > > Alex Romosan <romosan@sycorax.lbl.gov> wrote:
> > >
> > >> the backlight on my thinkpad still (2.6.20 worked fine) doesn't come
> > >> on if i have the radeon backlight enabled. without it, i guess it's
> > >> the ibm acpi modules that controls the backlight and it seems to work
> > >> fine.
> > >> 
> > >
> > > Unclear. Are you saying that the backlight comes on OK if you use
> > > the IBM acpi module?
> > 
> > yes, if i disable the radeon backlight and use the ibm acpi module,
> > than the backlight works. if i enable the radeon backlight, the screen
> > stays dark and i can't turn it on (i tried using radeontool to control
> > it but nothing happened).
> 
> Richard, is this actually a bug, or is it a config error or something like that?
> 
> And should we track it as a post-2.6.20 regression?

Why don't we compromise for 2.6.21 by marking the Radeon backlight
stuff EXPERIMENTAL until the cases where IBM ACPI works and the
Radeon backlight stuff does not are worked out?

Re: 2.6.21-rc2 radeon backlight

[Henrique de Moraes Holschuh]

On Fri, 02 Mar 2007, David Miller wrote:
> Why don't we compromise for 2.6.21 by marking the Radeon backlight
> stuff EXPERIMENTAL until the cases where IBM ACPI works and the
> Radeon backlight stuff does not are worked out?

What ibm-acpi backlight control does is *very* different from what the
radeon backlight control does.

ibm-acpi talks to the ThinkPad EC to change the brightness of the backlight,
it has *NO* on/off support.  Radeon talks to the video chipset to turn the
LVDS output on/off (or something to that effect), and thus to turn the
backlight on/off.

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh

Re: 2.6.21-rc2 radeon backlight

[Alex Romosan]

Henrique de Moraes Holschuh <hmh@hmh.eng.br> writes:

> On Fri, 02 Mar 2007, David Miller wrote:
>> Why don't we compromise for 2.6.21 by marking the Radeon backlight
>> stuff EXPERIMENTAL until the cases where IBM ACPI works and the
>> Radeon backlight stuff does not are worked out?
>
> What ibm-acpi backlight control does is *very* different from what the
> radeon backlight control does.
>
> ibm-acpi talks to the ThinkPad EC to change the brightness of the backlight,
> it has *NO* on/off support.  Radeon talks to the video chipset to turn the
> LVDS output on/off (or something to that effect), and thus to turn the
> backlight on/off.

if i don't enable the radeon backlight support the backlight works and
i can turn it on and off using radeontool. if i enable the radeon
backlight support the screen stays dark (although i can log on my
laptop remotely). if i try to use radeontool to turn the backlight on
nothing happens. so, at least on the thinkpad, the radeon backlight
support disables some other kind of backlight support that works
otherwise.

--alex--

-- 
| I believe the moment is at hand when, by a paranoiac and active |
|  advance of the mind, it will be possible (simultaneously with  |
|  automatism and other passive states) to systematize confusion  |
|  and thus to help to discredit completely the world of reality. |

Re: 2.6.21-rc2 radeon backlight

[David Miller]

From: Alex Romosan <romosan@sycorax.lbl.gov>
Date: Fri, 02 Mar 2007 13:22:27 -0800

> if i don't enable the radeon backlight support the backlight works and
> i can turn it on and off using radeontool. if i enable the radeon
> backlight support the screen stays dark (although i can log on my
> laptop remotely). if i try to use radeontool to turn the backlight on
> nothing happens. so, at least on the thinkpad, the radeon backlight
> support disables some other kind of backlight support that works
> otherwise.

Right, that's the crux of the problem.

The Radeon backlight support conflicts somehow with the
way the ACPI code programs the same hardware, and until
we figure out exactly what's going on, this is a regression.

Re: [Linux-fbdev-devel] 2.6.21-rc2 radeon backlight

[James Simmons]

> > > yes, if i disable the radeon backlight and use the ibm acpi module,
> > > than the backlight works. if i enable the radeon backlight, the screen
> > > stays dark and i can't turn it on (i tried using radeontool to control
> > > it but nothing happened).
> > 
> > Richard, is this actually a bug, or is it a config error or something like that?
> > 
> > And should we track it as a post-2.6.20 regression?
> 
> Why don't we compromise for 2.6.21 by marking the Radeon backlight
> stuff EXPERIMENTAL until the cases where IBM ACPI works and the
> Radeon backlight stuff does not are worked out?

How about this?

diff --git a/drivers/video/Kconfig b/drivers/video/Kconfig
index b8f0a11..eb54d1f 100644
--- a/drivers/video/Kconfig
+++ b/drivers/video/Kconfig
@@ -679,7 +679,6 @@ config FB_NVIDIA
 	depends on FB && PCI
 	select I2C_ALGOBIT if FB_NVIDIA_I2C
 	select I2C if FB_NVIDIA_I2C
-	select FB_BACKLIGHT if FB_NVIDIA_BACKLIGHT
 	select FB_MODE_HELPERS
 	select FB_CFB_FILLRECT
 	select FB_CFB_COPYAREA
@@ -708,7 +707,8 @@ config FB_NVIDIA_I2C
 
 config FB_NVIDIA_BACKLIGHT
 	bool "Support for backlight control"
-	depends on FB_NVIDIA
+	depends on FB_NVIDIA && EXPERIMENTAL
+	select FB_BACKLIGHT
 	default y
 	help
 	  Say Y here if you want to control the backlight of your display.
@@ -717,7 +717,6 @@ config FB_RIVA
 	tristate "nVidia Riva support"
 	depends on FB && PCI
 	select FB_DDC if FB_RIVA_I2C
-	select FB_BACKLIGHT if FB_RIVA_BACKLIGHT
 	select FB_MODE_HELPERS
 	select FB_CFB_FILLRECT
 	select FB_CFB_COPYAREA
@@ -754,7 +753,8 @@ config FB_RIVA_DEBUG
 
 config FB_RIVA_BACKLIGHT
 	bool "Support for backlight control"
-	depends on FB_RIVA
+	depends on FB_RIVA && EXPERIMENTAL
+	select FB_BACKLIGHT
 	default y
 	help
 	  Say Y here if you want to control the backlight of your display.
@@ -994,7 +994,6 @@ config FB_RADEON
 	tristate "ATI Radeon display support"
 	depends on FB && PCI
 	select FB_DDC if FB_RADEON_I2C
-	select FB_BACKLIGHT if FB_RADEON_BACKLIGHT
 	select FB_MODE_HELPERS
 	select FB_CFB_FILLRECT
 	select FB_CFB_COPYAREA
@@ -1024,7 +1023,8 @@ config FB_RADEON_I2C
 
 config FB_RADEON_BACKLIGHT
 	bool "Support for backlight control"
-	depends on FB_RADEON
+	depends on FB_RADEON && EXPERIMENTAL 
+	select FB_BACKLIGHT
 	default y
 	help
 	  Say Y here if you want to control the backlight of your display.
@@ -1044,7 +1044,6 @@ config FB_ATY128
 	select FB_CFB_FILLRECT
 	select FB_CFB_COPYAREA
 	select FB_CFB_IMAGEBLIT
-	select FB_BACKLIGHT if FB_ATY128_BACKLIGHT
 	select FB_MACMODES if PPC_PMAC
 	help
 	  This driver supports graphics boards with the ATI Rage128 chips.
@@ -1056,7 +1055,8 @@ config FB_ATY128
 
 config FB_ATY128_BACKLIGHT
 	bool "Support for backlight control"
-	depends on FB_ATY128
+	depends on FB_ATY128 && EXPERIMENTAL
+	select FB_BACKLIGHT
 	default y
 	help
 	  Say Y here if you want to control the backlight of your display.
@@ -1067,7 +1067,6 @@ config FB_ATY
 	select FB_CFB_FILLRECT
 	select FB_CFB_COPYAREA
 	select FB_CFB_IMAGEBLIT
-	select FB_BACKLIGHT if FB_ATY_BACKLIGHT
 	select FB_MACMODES if PPC
 	help
 	  This driver supports graphics boards with the ATI Mach64 chips.
@@ -1106,7 +1105,8 @@ config FB_ATY_GX
 
 config FB_ATY_BACKLIGHT
 	bool "Support for backlight control"
-	depends on FB_ATY
+	depends on FB_ATY && EXPERIMENTAL
+	select FB_BACKLIGHT	
 	default y
 	help
 	  Say Y here if you want to control the backlight of your display.

Re: [Linux-fbdev-devel] 2.6.21-rc2 radeon backlight

[Andrew Morton]

(davem cc restored)

On Fri, 2 Mar 2007 21:41:12 +0000 (GMT)
James Simmons <jsimmons@infradead.org> wrote:

> > > > yes, if i disable the radeon backlight and use the ibm acpi module,
> > > > than the backlight works. if i enable the radeon backlight, the screen
> > > > stays dark and i can't turn it on (i tried using radeontool to control
> > > > it but nothing happened).
> > > 
> > > Richard, is this actually a bug, or is it a config error or something like that?
> > > 
> > > And should we track it as a post-2.6.20 regression?
> > 
> > Why don't we compromise for 2.6.21 by marking the Radeon backlight
> > stuff EXPERIMENTAL until the cases where IBM ACPI works and the
> > Radeon backlight stuff does not are worked out?
> 
> How about this?
> 
> diff --git a/drivers/video/Kconfig b/drivers/video/Kconfig
> index b8f0a11..eb54d1f 100644
> --- a/drivers/video/Kconfig
> +++ b/drivers/video/Kconfig
> @@ -679,7 +679,6 @@ config FB_NVIDIA
>  	depends on FB && PCI
>  	select I2C_ALGOBIT if FB_NVIDIA_I2C
>  	select I2C if FB_NVIDIA_I2C
> -	select FB_BACKLIGHT if FB_NVIDIA_BACKLIGHT
>  	select FB_MODE_HELPERS
>  	select FB_CFB_FILLRECT
>  	select FB_CFB_COPYAREA
> @@ -708,7 +707,8 @@ config FB_NVIDIA_I2C
>  
>  config FB_NVIDIA_BACKLIGHT
>  	bool "Support for backlight control"
> -	depends on FB_NVIDIA
> +	depends on FB_NVIDIA && EXPERIMENTAL
> +	select FB_BACKLIGHT
>  	default y
>  	help
>  	  Say Y here if you want to control the backlight of your display.
> @@ -717,7 +717,6 @@ config FB_RIVA
>  	tristate "nVidia Riva support"
>  	depends on FB && PCI
>  	select FB_DDC if FB_RIVA_I2C
> -	select FB_BACKLIGHT if FB_RIVA_BACKLIGHT
>  	select FB_MODE_HELPERS
>  	select FB_CFB_FILLRECT
>  	select FB_CFB_COPYAREA
> @@ -754,7 +753,8 @@ config FB_RIVA_DEBUG
>  
>  config FB_RIVA_BACKLIGHT
>  	bool "Support for backlight control"
> -	depends on FB_RIVA
> +	depends on FB_RIVA && EXPERIMENTAL
> +	select FB_BACKLIGHT
>  	default y
>  	help
>  	  Say Y here if you want to control the backlight of your display.
> @@ -994,7 +994,6 @@ config FB_RADEON
>  	tristate "ATI Radeon display support"
>  	depends on FB && PCI
>  	select FB_DDC if FB_RADEON_I2C
> -	select FB_BACKLIGHT if FB_RADEON_BACKLIGHT
>  	select FB_MODE_HELPERS
>  	select FB_CFB_FILLRECT
>  	select FB_CFB_COPYAREA
> @@ -1024,7 +1023,8 @@ config FB_RADEON_I2C
>  
>  config FB_RADEON_BACKLIGHT
>  	bool "Support for backlight control"
> -	depends on FB_RADEON
> +	depends on FB_RADEON && EXPERIMENTAL 
> +	select FB_BACKLIGHT
>  	default y
>  	help
>  	  Say Y here if you want to control the backlight of your display.
> @@ -1044,7 +1044,6 @@ config FB_ATY128
>  	select FB_CFB_FILLRECT
>  	select FB_CFB_COPYAREA
>  	select FB_CFB_IMAGEBLIT
> -	select FB_BACKLIGHT if FB_ATY128_BACKLIGHT
>  	select FB_MACMODES if PPC_PMAC
>  	help
>  	  This driver supports graphics boards with the ATI Rage128 chips.
> @@ -1056,7 +1055,8 @@ config FB_ATY128
>  
>  config FB_ATY128_BACKLIGHT
>  	bool "Support for backlight control"
> -	depends on FB_ATY128
> +	depends on FB_ATY128 && EXPERIMENTAL
> +	select FB_BACKLIGHT
>  	default y
>  	help
>  	  Say Y here if you want to control the backlight of your display.
> @@ -1067,7 +1067,6 @@ config FB_ATY
>  	select FB_CFB_FILLRECT
>  	select FB_CFB_COPYAREA
>  	select FB_CFB_IMAGEBLIT
> -	select FB_BACKLIGHT if FB_ATY_BACKLIGHT
>  	select FB_MACMODES if PPC
>  	help
>  	  This driver supports graphics boards with the ATI Mach64 chips.
> @@ -1106,7 +1105,8 @@ config FB_ATY_GX
>  
>  config FB_ATY_BACKLIGHT
>  	bool "Support for backlight control"
> -	depends on FB_ATY
> +	depends on FB_ATY && EXPERIMENTAL
> +	select FB_BACKLIGHT	
>  	default y
>  	help
>  	  Say Y here if you want to control the backlight of your display.

Sorry, but my confidence level on this one is very low.  We've had heaps
and heaps of Kconfig-related build errors in exactly this area.  I'd prefer
that a patch like this one have a lot of testing and review (and a
changelog?) before we let it near a tree.

coz I've been bitten before.

Re: [Linux-fbdev-devel] 2.6.21-rc2 radeon backlight

[Richard Purdie]

On Fri, 2007-03-02 at 13:58 -0800, Andrew Morton wrote:
> > @@ -1106,7 +1105,8 @@ config FB_ATY_GX
> >  
> >  config FB_ATY_BACKLIGHT
> >  	bool "Support for backlight control"
> > -	depends on FB_ATY
> > +	depends on FB_ATY && EXPERIMENTAL
> > +	select FB_BACKLIGHT	
> >  	default y
> >  	help
> >  	  Say Y here if you want to control the backlight of your display.
> 
> Sorry, but my confidence level on this one is very low.  We've had heaps
> and heaps of Kconfig-related build errors in exactly this area.  I'd prefer
> that a patch like this one have a lot of testing and review (and a
> changelog?) before we let it near a tree.

Agreed, I can tell from what I've seen so far this will at least break
the PMAC users...

See my proposed patch, if that doesn't work, we should revert the
problematic change.

Richard

Re: [Linux-fbdev-devel] 2.6.21-rc2 radeon backlight

[James Simmons]

> > Sorry, but my confidence level on this one is very low.  We've had heaps
> > and heaps of Kconfig-related build errors in exactly this area.  I'd prefer
> > that a patch like this one have a lot of testing and review (and a
> > changelog?) before we let it near a tree.
> 
> Agreed, I can tell from what I've seen so far this will at least break
> the PMAC users...
> 
> See my proposed patch, if that doesn't work, we should revert the
> problematic change.

I will ask work if I can take home a LCD monitor with a backlight to 
find the real problem and fix. I wouldn't be able to do that until next 
weekend. So take what ever patch Rich sends.

Re: 2.6.21-rc2 radeon backlight

[Richard Purdie]

On Fri, 2007-03-02 at 12:29 -0800, Andrew Morton wrote:
> On Fri, 02 Mar 2007 09:24:03 -0800
> Alex Romosan <romosan@sycorax.lbl.gov> wrote:
> > > Unclear. Are you saying that the backlight comes on OK if you use
> > > the IBM acpi module?
> > 
> > yes, if i disable the radeon backlight and use the ibm acpi module,
> > than the backlight works. if i enable the radeon backlight, the screen
> > stays dark and i can't turn it on (i tried using radeontool to control
> > it but nothing happened).
> 
> Richard, is this actually a bug, or is it a config error or something like that?
> 
> And should we track it as a post-2.6.20 regression?

Its a regression IMO. Arguably its a Kconfig error but a nasty one as
the defaults cause the problems. Different people seem to have different
interpretations but to me it appears that the patch from James causes
backlights to fail to turn on for a variety of devices which worked
before.

I propose the following patch (I was previously waiting on James for
this). It avoids backing out the problematic Kconfig changes but means a
user has to explicitly enable the backlight via a kernel or module
parameter.

Can people with backlight problems try enabling them in Kconfig but
applying the following patch? Hopefully I have you all cc'd.

If it works I will add it to the other fix I have queued and pass to
Linus via the backlight tree. If it doesn't, I will revert the
problematic Kconfig changes for the next -rc.

----

Enabling the backlight by default appears to cause problems for mamny
users. Disable backlight controls unless explicitly enabled by users via
a module parameter. Since PMAC users are known to work, default to
enabled in that case.

Signed-off-by: Richard Purdie <rpurdie@rpsys.net>

--- a/drivers/video/aty/aty128fb.c
+++ b/drivers/video/aty/aty128fb.c
@@ -357,6 +357,12 @@ static int default_lcd_on __devinitdata = 1;
 static int mtrr = 1;
 #endif
 
+#ifdef CONFIG_PMAC_BACKLIGHT
+static int backlight __devinitdata = 1;
+#else
+static int backlight __devinitdata = 0;
+#endif
+
 /* PLL constants */
 struct aty128_constants {
 	u32 ref_clk;
@@ -1652,6 +1658,9 @@ static int __devinit aty128fb_setup(char *options)
 		} else if (!strncmp(this_opt, "crt:", 4)) {
 			default_crt_on = simple_strtoul(this_opt+4, NULL, 0);
 			continue;
+		} else if (!strncmp(this_opt, "backlight:", 10)) {
+			backlight = simple_strtoul(this_opt+10, NULL, 0);
+			continue;
 		}
 #ifdef CONFIG_MTRR
 		if(!strncmp(this_opt, "nomtrr", 6)) {
@@ -1985,7 +1994,8 @@ static int __devinit aty128_init(struct pci_dev *pdev, const struct pci_device_i
 	par->lock_blank = 0;
 
 #ifdef CONFIG_FB_ATY128_BACKLIGHT
-	aty128_bl_init(par);
+	if (backlight)
+		aty128_bl_init(par);
 #endif
 
 	if (register_framebuffer(info) < 0)
diff --git a/drivers/video/aty/atyfb_base.c b/drivers/video/aty/atyfb_base.c
index a7e0062..00a5183 100644
--- a/drivers/video/aty/atyfb_base.c
+++ b/drivers/video/aty/atyfb_base.c
@@ -308,6 +308,12 @@ static int xclk;
 static int comp_sync __devinitdata = -1;
 static char *mode;
 
+#ifdef CONFIG_PMAC_BACKLIGHT
+static int backlight __devinitdata = 1;
+#else
+static int backlight __devinitdata = 0;
+#endif
+
 #ifdef CONFIG_PPC
 static int default_vmode __devinitdata = VMODE_CHOOSE;
 static int default_cmode __devinitdata = CMODE_CHOOSE;
@@ -2575,7 +2581,7 @@ static int __devinit aty_init(struct fb_info *info)
 			   | (USE_F32KHZ | TRISTATE_MEM_EN), par);
 	} else
 #endif
-	if (M64_HAS(MOBIL_BUS)) {
+	if (M64_HAS(MOBIL_BUS) && backlight) {
 #ifdef CONFIG_FB_ATY_BACKLIGHT
 		aty_bl_init (par);
 #endif
@@ -3757,6 +3763,8 @@ static int __init atyfb_setup(char *options)
 			xclk = simple_strtoul(this_opt+5, NULL, 0);
 		else if (!strncmp(this_opt, "comp_sync:", 10))
 			comp_sync = simple_strtoul(this_opt+10, NULL, 0);
+		else if (!strncmp(this_opt, "backlight:", 10))
+			backlight = simple_strtoul(this_opt+10, NULL, 0);
 #ifdef CONFIG_PPC
 		else if (!strncmp(this_opt, "vmode:", 6)) {
 			unsigned int vmode =
diff --git a/drivers/video/aty/radeon_base.c b/drivers/video/aty/radeon_base.c
index 7e228ad..7de3ff9 100644
--- a/drivers/video/aty/radeon_base.c
+++ b/drivers/video/aty/radeon_base.c
@@ -268,6 +268,11 @@ static int nomtrr = 0;
 #endif
 static int force_sleep;
 static int ignore_devlist;
+#ifdef CONFIG_PMAC_BACKLIGHT
+static int backlight = 1;
+#else
+static int backlight = 0;
+#endif
 
 /*
  * prototypes
@@ -2349,7 +2354,8 @@ static int __devinit radeonfb_pci_register (struct pci_dev *pdev,
 						 MTRR_TYPE_WRCOMB, 1);
 #endif
 
-	radeonfb_bl_init(rinfo);
+	if (backlight)
+		radeonfb_bl_init(rinfo);
 
 	printk ("radeonfb (%s): %s\n", pci_name(rinfo->pdev), rinfo->name);
 
@@ -2470,6 +2476,8 @@ static int __init radeonfb_setup (char *options)
 			force_dfp = 1;
 		} else if (!strncmp(this_opt, "panel_yres:", 11)) {
 			panel_yres = simple_strtoul((this_opt+11), NULL, 0);
+		} else if (!strncmp(this_opt, "backlight:", 10)) {
+			backlight = simple_strtoul(this_opt+10, NULL, 0);
 #ifdef CONFIG_MTRR
 		} else if (!strncmp(this_opt, "nomtrr", 6)) {
 			nomtrr = 1;
diff --git a/drivers/video/nvidia/nvidia.c b/drivers/video/nvidia/nvidia.c
index c18e955..b97ec69 100644
--- a/drivers/video/nvidia/nvidia.c
+++ b/drivers/video/nvidia/nvidia.c
@@ -83,6 +83,11 @@ static int bpp __devinitdata = 8;
 #ifdef CONFIG_MTRR
 static int nomtrr __devinitdata = 0;
 #endif
+#ifdef CONFIG_PMAC_BACKLIGHT
+static int backlight __devinitdata = 1;
+#else
+static int backlight __devinitdata = 0;
+#endif
 
 static char *mode_option __devinitdata = NULL;
 
@@ -1311,7 +1316,10 @@ static int __devinit nvidiafb_probe(struct pci_dev *pd,
 	nvidia_save_vga(par, &par->SavedReg);
 
 	pci_set_drvdata(pd, info);
-	nvidia_bl_init(par);
+
+	if (backlight)
+		nvidia_bl_init(par);
+
 	if (register_framebuffer(info) < 0) {
 		printk(KERN_ERR PFX "error registering nVidia framebuffer\n");
 		goto err_out_iounmap_fb;
@@ -1408,6 +1416,8 @@ static int __devinit nvidiafb_setup(char *options)
 			paneltweak = simple_strtoul(this_opt+11, NULL, 0);
 		} else if (!strncmp(this_opt, "vram:", 5)) {
 			vram = simple_strtoul(this_opt+5, NULL, 0);
+		} else if (!strncmp(this_opt, "backlight:", 10)) {
+			backlight = simple_strtoul(this_opt+10, NULL, 0);
 #ifdef CONFIG_MTRR
 		} else if (!strncmp(this_opt, "nomtrr", 6)) {
 			nomtrr = 1;
diff --git a/drivers/video/riva/fbdev.c b/drivers/video/riva/fbdev.c
index f8a3d60..1d1c7c6 100644
--- a/drivers/video/riva/fbdev.c
+++ b/drivers/video/riva/fbdev.c
@@ -215,6 +215,11 @@ static int noaccel   __devinitdata = 0;
 #ifdef CONFIG_MTRR
 static int nomtrr __devinitdata = 0;
 #endif
+#ifdef CONFIG_PMAC_BACKLIGHT
+static int backlight __devinitdata = 1;
+#else
+static int backlight __devinitdata = 0;
+#endif
 
 static char *mode_option __devinitdata = NULL;
 static int  strictmode       = 0;
@@ -2059,7 +2064,10 @@ static int __devinit rivafb_probe(struct pci_dev *pd,
 	info->monspecs.modedb = NULL;
 
 	pci_set_drvdata(pd, info);
-	riva_bl_init(info->par);
+
+	if (backlight)
+		riva_bl_init(info->par);
+
 	ret = register_framebuffer(info);
 	if (ret < 0) {
 		printk(KERN_ERR PFX
@@ -2157,6 +2165,8 @@ static int __init rivafb_setup(char *options)
 				forceCRTC = -1;
 		} else if (!strncmp(this_opt, "flatpanel", 9)) {
 			flatpanel = 1;
+		} else if (!strncmp(this_opt, "backlight:", 10)) {
+			backlight = simple_strtoul(this_opt+10, NULL, 0);
 #ifdef CONFIG_MTRR
 		} else if (!strncmp(this_opt, "nomtrr", 6)) {
 			nomtrr = 1;

Re: 2.6.21-rc2 radeon backlight

[Alex Romosan]

Richard Purdie <rpurdie@rpsys.net> writes:

> I propose the following patch (I was previously waiting on James for
> this). It avoids backing out the problematic Kconfig changes but
> means a user has to explicitly enable the backlight via a kernel or
> module parameter.
>
> Can people with backlight problems try enabling them in Kconfig but
> applying the following patch? Hopefully I have you all cc'd.
>
> If it works I will add it to the other fix I have queued and pass to
> Linus via the backlight tree. If it doesn't, I will revert the
> problematic Kconfig changes for the next -rc.

with the patch applied i can boot into 2.6.21-rc2 and have a working
backlight when i enable radeon backlight support. so i guess the patch
successfully disables the new radeon backlight stuff on my thinkpad
and therefore it does what it says it does.

--alex--

-- 
| I believe the moment is at hand when, by a paranoiac and active |
|  advance of the mind, it will be possible (simultaneously with  |
|  automatism and other passive states) to systematize confusion  |
|  and thus to help to discredit completely the world of reality. |

Re: 2.6.21-rc2 radeon backlight

[James Simmons]

> > Richard, is this actually a bug, or is it a config error or something like that?
> > 
> > And should we track it as a post-2.6.20 regression?
> 
> Its a regression IMO. Arguably its a Kconfig error but a nasty one as
> the defaults cause the problems. Different people seem to have different
> interpretations but to me it appears that the patch from James causes
> backlights to fail to turn on for a variety of devices which worked
> before.

   It is NOT a Kconfig error. The problem is that for many fbdev drivers 
the backlight code is broken. For some magic reason it only works on 
pmacs. Most likely because the firmware properly sets up the backlight. 
Plus we have the conflict with acpi backlight. 
   Think about it. Enabling the backlight code breaks the backlight. 
Without the backlight driver the default behaviour of the backlight works.
Who sets up the default behavior? I don't have a LCD panel with a 
backlight othewise I would track the problem down.

Re: 2.6.21-rc2 radeon backlight

[Richard Purdie]

On Sat, 2007-03-03 at 16:06 +0000, James Simmons wrote:
> > > Richard, is this actually a bug, or is it a config error or something like that?
> > > 
> > > And should we track it as a post-2.6.20 regression?
> > 
> > Its a regression IMO. Arguably its a Kconfig error but a nasty one as
> > the defaults cause the problems. Different people seem to have different
> > interpretations but to me it appears that the patch from James causes
> > backlights to fail to turn on for a variety of devices which worked
> > before.
> 
>    It is NOT a Kconfig error. The problem is that for many fbdev drivers 
> the backlight code is broken. For some magic reason it only works on 
> pmacs. Most likely because the firmware properly sets up the backlight. 
> Plus we have the conflict with acpi backlight. 
>    Think about it. Enabling the backlight code breaks the backlight. 
> Without the backlight driver the default behaviour of the backlight works.
> Who sets up the default behavior? I don't have a LCD panel with a 
> backlight othewise I would track the problem down.

I not arguing several fb driver's backlight code isn't broken, it is. We
have a Kconfig problem though since we used to stop users selecting
things that didn't work and now we're allowing them. Worse still, the
defaults break for people. That is a regression.

Anyhow, the patch I proposed should let people enable/disable it at
runtime with defaults known to work which should address the problem
until someone figures out how to fix the backlight drivers themselves
properly.

Cheers,

Richard