Pavel Machek ha scritto:
> Hi!
Hi!
>>> Try disabling acpi embedded controller.
>>>   
>>>       
>> How can I accomplish this? Are you referring to the i8042?
>>     
>
> rmmod acpi_ec or how is it called. But I'm not sure how easy this is.
>   
My lsmod doesn't list any acpi module - but I have kacpid, kacpi_listen
and hald-addon-acpi processes running.
I have compiled the kernel with embedded ACPI support, not modular -
maybe that's what you are talking about?

>   
>>> Try watching keyboard interrupts. Are they lost?
>>>   
>>>       
>> I am pretty sure they are. I think that ACPI pauses interrupts for a
>> while (100ms?) and so some keyboard interrupts are lost.
>>     
> input layer can poll keyboard on its own, perhaps that helps?
>   
I am not yet very good at kernel hacking - however yes, it gives me more
clues about identifying the issue (see also below link).

> Can you try to measure the interrupt latencies?
>   
I will try to do it myself, however somebody else the same problem of
mine has already made the tests, please see
http://dev.laptop.org/ticket/2401

> Does this happen in init=/bin/bash?
>   
I will perform a test, however the problem rarely verifies in a VT
console. I have tried both the kbd and keyboard drivers of X, without
any difference. The keyboard problem happens a lot more when plugging in
the battery (more ACPI traffic?).

I have emerged lm_sensors but can't get it running - it keeps saying "No
sensors found!" and complaining about kernel drivers not properly setup.
I have attached the output of sensors-detect, from which it seems that
the kernel is OK.

Thank you,
--
  Daniele

> 							Pavel
>
>