[PATCH 1/2] drm/i915: clear up backlight inversion confusion on gen4

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 04/23/2012 03:39 PM, Daniel Vetter wrote:
> On Mon, Apr 23, 2012 at 03:15:02PM +0200, Carsten Emde wrote:
>> On 04/23/2012 02:36 PM, Daniel Vetter wrote:
>>> On Mon, Apr 23, 2012 at 02:32:57PM +0200, Daniel Vetter wrote:
>>>> On Mon, Apr 23, 2012 at 01:54:23PM +0200, Carsten Emde wrote:
>>>>> On 04/23/2012 11:32 AM, Daniel Vetter wrote:
>>>>>> There's a bit in the docs for gen4 only that says whether the
>>>>>> backlight control is inverted. And both the quirk we have and
>>>>>> all bugs only concern i965gm and gm45 (and mostly Acer) afaics.
>>>>>>
>>>>>> So lets drop the quirk and use the bit instead.
>>>>>>
>>>>>> Also clean up the BLC register definitions a bit by correctly
>>>>>> grouping the CTL and CTL2 definitions together.
>>>>>> [..]
>>>>>> diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
>>>>>> index 4c844c6..5d215f0 100644
>>>>>> --- a/drivers/gpu/drm/i915/intel_display.c
>>>>>> +++ b/drivers/gpu/drm/i915/intel_display.c
>>>>>> @@ -6482,9 +6482,6 @@ static struct intel_quirk intel_quirks[] = {
>>>>>>
>>>>>>   	/* Sony Vaio Y cannot use SSC on LVDS */
>>>>>>   	{ 0x0046, 0x104d, 0x9076, quirk_ssc_force_disable },
>>>>>> -
>>>>>> -	/* Acer Aspire 5734Z must invert backlight brightness */
>>>>>> -	{ 0x2a42, 0x1025, 0x0459, quirk_invert_brightness },
>>>>> Made a first test with these lines removed and the new mechanism
>>>>> added.  Unfortunately, the screen is now dark again. Works after
>>>>> re-adding the quirk.
>>>>>
>>>>>> +	/* gen4 has a polarity bit */
>>>>>> +	if (IS_GEN4(dev)&&  (I915_READ(BLC_PWM_CTL2)&  BLM_POLARITY_I965))
>>>>>> +			return intel_panel_get_max_backlight(dev) - val;
>>>>>> +
>>>>> Apparently, these conditions are not met on the Acer Aspire 5734Z.
>>>>> Will add some debug output and evaluate further. For the time being,
>>>>> please leave the quirk in place.
>>>>
>>>> Hm, I've tried this and when I set this bit, panel brightness is inverted
>>>> on my gm45. Can you install intel-gpu-tools and read out the 2 backlight
>>>> control registers with
>>>>
>>>> # intel_reg_read 0x61254
>>>> # intel_reg_read 0x61250
>>>
>>> I've forgotten to add: Also please grab these register values when booting
>>> with nomodeset, so that we can compare the values the bios sets with
>>> whatever i915 sets.
>>
>> Without nomodeset:
>> # intel_reg_read 0x61254
>> 0x61254 : 0xB4A0B4A
>> # intel_reg_read 0x61250
>> 0x61250 : 0xE0000000
>>
>> With nomodeset:
>> # intel_reg_read 0x61254
>> 0x61254 : 0xB4A0B4A
>> # intel_reg_read 0x61250
>> 0x61250 : 0xE0000000
>>
>> Thus, (I915_READ(BLC_PWM_CTL2)&  BLM_POLARITY_I965) is never set.
>> But bit 29 would work, so we could add another conditional such as:
>> #define BLM_POLARITY_PCH_I965 (1<<  29) /* PCH only */
>> if (PCH)
>>    try (I915_READ(BLC_PWM_CTL2)&  BLM_POLARITY_PCH_I965)
>
> Ok, let's try something funny. Please boot with kms enabled and try out
> what the following values for CTL1 do:
>
> # intel_reg_write 0x61250 0x80000000
> # intel_reg_write 0x61250 0xa0000000
> # intel_reg_write 0x61250 0x90000000
> # intel_reg_write 0x61250 0xb0000000

# intel_reg_write 0x61250 0x80000000
Value before: 0xE0000000
Value after: 0x80000000
# intel_reg_read 0x61254
0x61254 : 0xB4A0B4A

# intel_reg_write 0x61250 0xa0000000
Value before: 0x80000000
Value after: 0xA0000000
# intel_reg_read 0x61254
0x61254 : 0xB4A0B4A

# intel_reg_write 0x61250 0x90000000
Value before: 0xA0000000
Value after: 0x90000000
# intel_reg_read 0x61254
0x61254 : 0xB4A0B4A

# intel_reg_write 0x61250 0xb0000000
Value before: 0x90000000
Value after: 0xB0000000
# intel_reg_read 0x61254
0x61254 : 0xB4A0B4A

Not much.

	-Carsten.


[Index of Archives]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]
  Powered by Linux