keyboard dies during failed suspend attempt

keyboard dies during failed suspend attempt

[Pierre Ossman]

I discovered a problem with my laptop keyboard when the machine failed 
to suspend. Pavel Machek pointed me in your direction for guidance. :)

The original issue (swsusp failing) is in this thread:

http://marc.theaimsgroup.com/?t=113093802700002&r=1&w=2

The side issue is that the keyboard goes completely dead when the 
suspend fails like this. Not even hardware buttons that control the 
intensity of the TFT backlight work.

The problem doesn't happen every time, but it seems to be often enough 
to do some decent testing.

The problem seems to have appeared after 2.6.14 was released. Since the 
  problem is intermittent I can't be 100% sure of this, but it's fairly 
likely since none of the tests before 2.6.14 failed.

Rgds
Pierre

Re: keyboard dies during failed suspend attempt

[Dmitry Torokhov]

On Thursday 03 November 2005 03:54, Pierre Ossman wrote:
> I discovered a problem with my laptop keyboard when the machine failed 
> to suspend. Pavel Machek pointed me in your direction for guidance. :)
> 
> The original issue (swsusp failing) is in this thread:
> 
> http://marc.theaimsgroup.com/?t=113093802700002&r=1&w=2
> 
> The side issue is that the keyboard goes completely dead when the 
> suspend fails like this. Not even hardware buttons that control the 
> intensity of the TFT backlight work.
> 

Are these controlled by ACPI?

> The problem doesn't happen every time, but it seems to be often enough 
> to do some decent testing.
> 
> The problem seems to have appeared after 2.6.14 was released. Since the 
>   problem is intermittent I can't be 100% sure of this, but it's fairly 
> likely since none of the tests before 2.6.14 failed.
>

It feels like device_resume is not called somewhere when swsusp fails.

Could you try activating debug mode for i8042:

	echo 1 > /sys/modules/i8042/parameters/debug

...and then making it fail. Then we'll see if i8042 resume methods are
called and whether they succeed.

-- 
Dmitry

Re: keyboard dies during failed suspend attempt

[Pierre Ossman]

Dmitry Torokhov wrote:
> On Thursday 03 November 2005 03:54, Pierre Ossman wrote:
>   
>> I discovered a problem with my laptop keyboard when the machine failed 
>> to suspend. Pavel Machek pointed me in your direction for guidance. :)
>>
>> The original issue (swsusp failing) is in this thread:
>>
>> http://marc.theaimsgroup.com/?t=113093802700002&r=1&w=2
>>
>> The side issue is that the keyboard goes completely dead when the 
>> suspend fails like this. Not even hardware buttons that control the 
>> intensity of the TFT backlight work.
>>
>>     
>
> Are these controlled by ACPI?
>
>   

Haven't the slightest. There is nothing to control the backlight in 
/proc/acpi at least (all brightness say '<not supported>').

>> The problem doesn't happen every time, but it seems to be often enough 
>> to do some decent testing.
>>
>> The problem seems to have appeared after 2.6.14 was released. Since the 
>>   problem is intermittent I can't be 100% sure of this, but it's fairly 
>> likely since none of the tests before 2.6.14 failed.
>>
>>     
>
> It feels like device_resume is not called somewhere when swsusp fails.
>
> Could you try activating debug mode for i8042:
>
> 	echo 1 > /sys/modules/i8042/parameters/debug
>
> ...and then making it fail. Then we'll see if i8042 resume methods are
> called and whether they succeed.
>
>   

I've been suspending the machine all day now and I have yet to get it to 
fail. There must have been something different yesterday but I can't 
figure out what... I'll do some more testing during the weekend and see 
if I can provoke it again. Otherwise I suppose we can write it off as 
something temporary.

Rgds
Pierre