diff mbox

ACPI / video: check _DOD list when creating backlight device

Message ID 5479C3BB.5080905@intel.com (mailing list archive)
State Not Applicable, archived
Headers show

Commit Message

Aaron Lu Nov. 29, 2014, 1:01 p.m. UTC 
On 11/29/2014 01:18 AM, Brian Norris wrote:
> On Fri, Nov 28, 2014 at 07:55:00PM +0800, Aaron Lu wrote:
>> On 11/28/2014 05:59 PM, Brian Norris wrote:
>>> And indeed, there is a regression! My Dell Latituded E6410's backlight
>>> control no longer works after this commit, and I get messages like this
>>> instead:
>>>
>>> [   57.214610] ACPI: Failed to switch the brightness
>>>
>>> If I revert this commit, my backlight controls work again. Also, I
>>> regain a cooling device (?) that was being ignored:
>>>
>>> [    1.332682] acpi device:02: registered as cooling_device0
>>>
>>> Do you need any additional info to handle the regression, or should we
>>> just revert the patch?
>>
>> Please attach acpidump, dmesg with video.dyndbg="module video +pft" in
>> kernel cmdline, list the /sys/class/backlight with and without this
>> commit, thanks.
> 
> Appending dmesg and attaching acpidump. Unfortunately, I'm using a few
> binary modules, since Wifi and GPU suspend/resume support are broken
> otherwise, but I did test briefly with Nouveau anyway. Results are
> essentially the same with Nouveau (although Nouveau seems to provide an
> extra backlight device, though it doesn't seem to do anything for me).
> 
> Before reverting:
> 
> $ ls -al /sys/class/backlight
> total 0
> drwxr-xr-x  2 root root 0 Nov 28 09:01 .
> drwxr-xr-x 58 root root 0 Nov 28 09:01 ..
> 
> After reverting:
> 
> $ ls -al /sys/class/backlight
> total 0
> drwxr-xr-x  2 root root 0 Nov 28 09:10 .
> drwxr-xr-x 50 root root 0 Nov 28 09:10 ..
> lrwxrwxrwx  1 root root 0 Nov 28 09:09 acpi_video0 -> ../../devices/pci0000:00/0000:00:01.0/0000:01:00.0/backlight/acpi_video0

The current logic to bind device doesn't work with your system, that
logic is there for a long time so I prefer not to change that, instead,
I can change the newly added function acpi_video_device_in_dod to let it
directly compare the 0-12 bits of the ID to decide if the video output
device is in the _DOD list. Can you please try the following patch?


Thanks,
Aaron
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Comments

Brian Norris Nov. 29, 2014, 5:34 p.m. UTC | #1 
On Sat, Nov 29, 2014 at 09:01:47PM +0800, Aaron Lu wrote:
> On 11/29/2014 01:18 AM, Brian Norris wrote:
> > On Fri, Nov 28, 2014 at 07:55:00PM +0800, Aaron Lu wrote:
> >> On 11/28/2014 05:59 PM, Brian Norris wrote:
> >>> Do you need any additional info to handle the regression, or should we
> >>> just revert the patch?
> >>
> >> Please attach acpidump, dmesg with video.dyndbg="module video +pft" in
> >> kernel cmdline, list the /sys/class/backlight with and without this
> >> commit, thanks.
[...]
> > Before reverting:
> > 
> > $ ls -al /sys/class/backlight
> > total 0
> > drwxr-xr-x  2 root root 0 Nov 28 09:01 .
> > drwxr-xr-x 58 root root 0 Nov 28 09:01 ..
> > 
> > After reverting:
> > 
> > $ ls -al /sys/class/backlight
> > total 0
> > drwxr-xr-x  2 root root 0 Nov 28 09:10 .
> > drwxr-xr-x 50 root root 0 Nov 28 09:10 ..
> > lrwxrwxrwx  1 root root 0 Nov 28 09:09 acpi_video0 -> ../../devices/pci0000:00/0000:00:01.0/0000:01:00.0/backlight/acpi_video0
> 
> The current logic to bind device doesn't work with your system, that
> logic is there for a long time so I prefer not to change that, instead,
> I can change the newly added function acpi_video_device_in_dod to let it
> directly compare the 0-12 bits of the ID to decide if the video output
> device is in the _DOD list. Can you please try the following patch?
> 
> diff --git a/drivers/acpi/video.c b/drivers/acpi/video.c
> index 41e6b977ceb2..185a57d13723 100644
> --- a/drivers/acpi/video.c
> +++ b/drivers/acpi/video.c
> @@ -1164,7 +1164,8 @@ static bool acpi_video_device_in_dod(struct acpi_video_device *device)
>  		return true;
>  
>  	for (i = 0; i < video->attached_count; i++) {
> -		if (video->attached_array[i].bind_info == device)
> +		if ((video->attached_array[i].value.int_val & 0xfff) ==
> +		    (device->device_id & 0xfff))
>  			return true;
>  	}

Thanks for the patch! This works better for me:

$ ls -al /sys/class/backlight/
total 0
drwxr-xr-x  2 root root 0 Nov 29 09:00 .
drwxr-xr-x 50 root root 0 Nov 29 09:00 ..
lrwxrwxrwx  1 root root 0 Nov 29 08:59 acpi_video0 -> ../../devices/pci0000:00/0000:00:01.0/0000:01:00.0/backlight/acpi_video0

And I see no dmesg dev_dbg() prints from the acpi/video module.

Tested-by: Brian Norris <computersforpeace@gmail.com>

Should we try to get this into 3.18, then?

Thanks,
Brian
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/drivers/acpi/video.c b/drivers/acpi/video.c
index 41e6b977ceb2..185a57d13723 100644
--- a/drivers/acpi/video.c
+++ b/drivers/acpi/video.c
@@ -1164,7 +1164,8 @@  static bool acpi_video_device_in_dod(struct acpi_video_device *device)
 		return true;
 
 	for (i = 0; i < video->attached_count; i++) {
-		if (video->attached_array[i].bind_info == device)
+		if ((video->attached_array[i].value.int_val & 0xfff) ==
+		    (device->device_id & 0xfff))
 			return true;
 	}