Wireless key on Fujitsu E753 not working since a57f7f9175b8ccbc9df83ac13860488913115de4

Wireless key on Fujitsu E753 not working since a57f7f9175b8ccbc9df83ac13860488913115de4

[Jan Hinnerk Stosch]

Hi,

since commit "a57f7f9175b8ccbc9df83ac13860488913115de4" (bisected) the
wireless key on my Fujitsu E753 is not working any more. If I boot up
with parameter "acpi_osi="!Windows 2012"" it works, though.
I filed an ACPI-bug
(https://bugzilla.kernel.org/show_bug.cgi?id=60161), but was told
(https://bugzilla.kernel.org/show_bug.cgi?id=60161#c5) that this is
caused by the fujitsu_laptop driver. So I report it here. Should I
file a bug anywhere or is it enough to write it to the list?

Thanks in advance,

jhs

Re: Wireless key on Fujitsu E753 not working since a57f7f9175b8ccbc9df83ac13860488913115de4

[Jan Hinnerk Stosch]

Hi,

as I have not received any answer I'd like to ask again, what to do
with that problem, because I don't have a clue at all...

Regards,

jhs

2013/6/28 Jan Hinnerk Stosch <janhinnerk.stosch@gmail.com>:
> Hi,
>
> since commit "a57f7f9175b8ccbc9df83ac13860488913115de4" (bisected) the
> wireless key on my Fujitsu E753 is not working any more. If I boot up
> with parameter "acpi_osi="!Windows 2012"" it works, though.
> I filed an ACPI-bug
> (https://bugzilla.kernel.org/show_bug.cgi?id=60161), but was told
> (https://bugzilla.kernel.org/show_bug.cgi?id=60161#c5) that this is
> caused by the fujitsu_laptop driver. So I report it here. Should I
> file a bug anywhere or is it enough to write it to the list?
>
> Thanks in advance,
>
> jhs

Re: Wireless key on Fujitsu E753 not working since a57f7f9175b8ccbc9df83ac13860488913115de4

[Jonathan Woithe]

> 2013/6/28 Jan Hinnerk Stosch <janhinnerk.stosch@gmail.com>:
> > since commit "a57f7f9175b8ccbc9df83ac13860488913115de4" (bisected) the
> > wireless key on my Fujitsu E753 is not working any more. If I boot up
> > with parameter "acpi_osi="!Windows 2012"" it works, though.
> > I filed an ACPI-bug
> > (https://bugzilla.kernel.org/show_bug.cgi?id=60161), but was told
> > (https://bugzilla.kernel.org/show_bug.cgi?id=60161#c5) that this is
> > caused by the fujitsu_laptop driver. So I report it here. Should I
> > file a bug anywhere or is it enough to write it to the list?

Does this "wireless key" do what you expect if the fujitsu_laptop module is
not loaded?

I know nothing about the E753.  Is the "wireless key" a dedicated key on the
keyboard, a separate switch, or something entirely different?  It sounds
like it's just an enable switch.

jonathan

Re: Wireless key on Fujitsu E753 not working since a57f7f9175b8ccbc9df83ac13860488913115de4

[Jan Hinnerk Stosch]

2013/8/12 Jonathan Woithe <jwoithe@just42.net>:
>> 2013/6/28 Jan Hinnerk Stosch <janhinnerk.stosch@gmail.com>:
>> > since commit "a57f7f9175b8ccbc9df83ac13860488913115de4" (bisected) the
>> > wireless key on my Fujitsu E753 is not working any more. If I boot up
>> > with parameter "acpi_osi="!Windows 2012"" it works, though.
>> > I filed an ACPI-bug
>> > (https://bugzilla.kernel.org/show_bug.cgi?id=60161), but was told
>> > (https://bugzilla.kernel.org/show_bug.cgi?id=60161#c5) that this is
>> > caused by the fujitsu_laptop driver. So I report it here. Should I
>> > file a bug anywhere or is it enough to write it to the list?
>
> Does this "wireless key" do what you expect if the fujitsu_laptop module is
> not loaded?
>
> I know nothing about the E753.  Is the "wireless key" a dedicated key on the
> keyboard, a separate switch, or something entirely different?  It sounds
> like it's just an enable switch.
>
> jonathan

No, the key also does not work without fujitsu_laptop module.

The key is one of two special keys - one for ECO mode (which I have
not tested or used at all) and the wireless key which switches of all
wireless functions (WLAN and bluetooth). They are not part of the
"main" keyboard. According to the output of "rfkill list" it works
indeed as a hardware switch (Hard blocked: yes/no).

Re: Wireless key on Fujitsu E753 not working since a57f7f9175b8ccbc9df83ac13860488913115de4

[Jonathan Woithe]

On Tue, Aug 13, 2013 at 10:27:53AM +0200, Jan Hinnerk Stosch wrote:
> 2013/8/12 Jonathan Woithe <jwoithe@just42.net>:
> >> 2013/6/28 Jan Hinnerk Stosch <janhinnerk.stosch@gmail.com>:
> >> > since commit "a57f7f9175b8ccbc9df83ac13860488913115de4" (bisected) the
> >> > wireless key on my Fujitsu E753 is not working any more. If I boot up
> >> > with parameter "acpi_osi="!Windows 2012"" it works, though.
> >> > I filed an ACPI-bug
> >> > (https://bugzilla.kernel.org/show_bug.cgi?id=60161), but was told
> >> > (https://bugzilla.kernel.org/show_bug.cgi?id=60161#c5) that this is
> >> > caused by the fujitsu_laptop driver. So I report it here. Should I
> >> > file a bug anywhere or is it enough to write it to the list?
> >
> > Does this "wireless key" do what you expect if the fujitsu_laptop module is
> > not loaded?
> >
> > I know nothing about the E753.  Is the "wireless key" a dedicated key on the
> > keyboard, a separate switch, or something entirely different?  It sounds
> > like it's just an enable switch.
> 
> No, the key also does not work without fujitsu_laptop module.

Ok, thanks for confirming.

> The key is one of two special keys - one for ECO mode (which I have
> not tested or used at all) and the wireless key which switches of all
> wireless functions (WLAN and bluetooth). They are not part of the
> "main" keyboard. According to the output of "rfkill list" it works
> indeed as a hardware switch (Hard blocked: yes/no).

I see.  On some Fujitsu laptops the RF kill switch is strictly a hardware
thing - it does not require any software assistance at all.  This is the
case on the unit I have (S7020).  However, there were some models which did
need software assistance, and this was added to the fujitsu-laptop driver by
someone with such a machine.

I've had a look at the commit you referenced, and all it seems to do is add
a new _OSI string to ACPI.  I am not overly familiar with how the APCI side
of thing works.  It seems odd that adding this string to the kernel should
make fujitsu-laptop (or any other driver) behave differently.  Does the
presence of this new string somehow cause the FUNC_RFKILL ACPI method
(called via call_fext_func()) to behave differently?

If someone could explain what the new _OSI string could be doing (or point
to documentation) I could try to work out what might be happening.  I am a
little restricted in what I can do though because I don't have fujitsu
hardware that supports the software RF kill switch.

My best guess is that the new OSI string activates an alternative ACPI
interface which does not include the custom Fujitsu RF kill interface, or
which implements a different non-standard RF kill interface which
fujitsu-laptop doesn't know about.  However, I don't have a great grasp of
the underlying ACPI functionality and would be very happy to be corrected on
this point by those more knowledgeable than I in this area.

Regards
  jonathan

Re: Wireless key on Fujitsu E753 not working since a57f7f9175b8ccbc9df83ac13860488913115de4

[Jan Hinnerk Stosch]

Hi,

I just wanted to ask whether there is something I could do about this?
Should I ask in another mailing list or sth?

Regards

jhs

2013/9/24 Jonathan Woithe <jwoithe@just42.net>:
> On Tue, Aug 13, 2013 at 10:27:53AM +0200, Jan Hinnerk Stosch wrote:
>> 2013/8/12 Jonathan Woithe <jwoithe@just42.net>:
>> >> 2013/6/28 Jan Hinnerk Stosch <janhinnerk.stosch@gmail.com>:
>> >> > since commit "a57f7f9175b8ccbc9df83ac13860488913115de4" (bisected) the
>> >> > wireless key on my Fujitsu E753 is not working any more. If I boot up
>> >> > with parameter "acpi_osi="!Windows 2012"" it works, though.
>> >> > I filed an ACPI-bug
>> >> > (https://bugzilla.kernel.org/show_bug.cgi?id=60161), but was told
>> >> > (https://bugzilla.kernel.org/show_bug.cgi?id=60161#c5) that this is
>> >> > caused by the fujitsu_laptop driver. So I report it here. Should I
>> >> > file a bug anywhere or is it enough to write it to the list?
>> >
>> > Does this "wireless key" do what you expect if the fujitsu_laptop module is
>> > not loaded?
>> >
>> > I know nothing about the E753.  Is the "wireless key" a dedicated key on the
>> > keyboard, a separate switch, or something entirely different?  It sounds
>> > like it's just an enable switch.
>>
>> No, the key also does not work without fujitsu_laptop module.
>
> Ok, thanks for confirming.
>
>> The key is one of two special keys - one for ECO mode (which I have
>> not tested or used at all) and the wireless key which switches of all
>> wireless functions (WLAN and bluetooth). They are not part of the
>> "main" keyboard. According to the output of "rfkill list" it works
>> indeed as a hardware switch (Hard blocked: yes/no).
>
> I see.  On some Fujitsu laptops the RF kill switch is strictly a hardware
> thing - it does not require any software assistance at all.  This is the
> case on the unit I have (S7020).  However, there were some models which did
> need software assistance, and this was added to the fujitsu-laptop driver by
> someone with such a machine.
>
> I've had a look at the commit you referenced, and all it seems to do is add
> a new _OSI string to ACPI.  I am not overly familiar with how the APCI side
> of thing works.  It seems odd that adding this string to the kernel should
> make fujitsu-laptop (or any other driver) behave differently.  Does the
> presence of this new string somehow cause the FUNC_RFKILL ACPI method
> (called via call_fext_func()) to behave differently?
>
> If someone could explain what the new _OSI string could be doing (or point
> to documentation) I could try to work out what might be happening.  I am a
> little restricted in what I can do though because I don't have fujitsu
> hardware that supports the software RF kill switch.
>
> My best guess is that the new OSI string activates an alternative ACPI
> interface which does not include the custom Fujitsu RF kill interface, or
> which implements a different non-standard RF kill interface which
> fujitsu-laptop doesn't know about.  However, I don't have a great grasp of
> the underlying ACPI functionality and would be very happy to be corrected on
> this point by those more knowledgeable than I in this area.
>
> Regards
>   jonathan

Re: Wireless key on Fujitsu E753 not working since a57f7f9175b8ccbc9df83ac13860488913115de4

[Jonathan Woithe]

Jo Jan

On Fri, Nov 22, 2013 at 08:36:23PM +0100, Jan Hinnerk Stosch wrote:
> I just wanted to ask whether there is something I could do about this?

I'm not sure.  To recap what you have found:

> since commit "a57f7f9175b8ccbc9df83ac13860488913115de4" (bisected) the
> wireless key on my Fujitsu E753 is not working any more. If I boot up
> with parameter "acpi_osi="!Windows 2012"" it works, though.
> I filed an ACPI-bug
> (https://bugzilla.kernel.org/show_bug.cgi?id=60161), but was told
> (https://bugzilla.kernel.org/show_bug.cgi?id=60161#c5) that this is
> caused by the fujitsu_laptop driver. So I report it here. Should I
> file a bug anywhere or is it enough to write it to the list?

You then answered a followup question which indicated that if fujitsu_laptop
module is *not* loaded the key still doesn't work.

Does the key work without fujitsu_laptop loaded on a kernel before the
commit in question?

As I have said previously I am not sure I am in a position to offer
assistance here.  Firstly I do not have access to any Fujitsu hardware other
than my S7020, and that hardware does not implement a wireless key like the
E753 does.  Secondly, I do not presently understand how the OSI changes in
commit a57f7f9175b8ccbc9df83ac13860488913115de4 could affect the
fujitsu_laptop module and break the handling of this "wireless key".

> Should I ask in another mailing list or sth?

I don't think so.  Platform-driver-x86 is the appropriate mailing list.

As I said in an earlier email:

> If someone could explain what the new _OSI string could be doing (or point
> to documentation) I could try to work out what might be happening.  I am a
> little restricted in what I can do though because I don't have fujitsu
> hardware that supports the software RF kill switch.

Understanding what this _OSI string does and how it could be causing
fujitsu_laptop to loose control of the "wireless" key is fundamental to
solving this problem.  I do not have this knowledge at this point in time,
and I do not have the time available right now to start looking into the
whole OSI thing from scratch.  Not having hardware affected by the problem
also makes it difficult for me to investigate.

To make progress with this issue, we need to work out what this OSI change
has done to the seemingly unrelated code in fujitsu_laptop.  This involves
working out what OSI does in the first place.

Sorry that I can't be more definitive than this.

Regards
  jonathan

Re: Wireless key on Fujitsu E753 not working since a57f7f9175b8ccbc9df83ac13860488913115de4

[Jan Hinnerk Stosch]

Thanks for your answer and sorry for the delay.
I finally got some time to look into this:
When I boot with the named boot option the key does work with and
without fujitsu_laptop being loaded.
Additionally I found a bug report at launchpad:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1159690 with some
more people affected by this, who are using notebooks from different
manufacturers, so this doesn't really seem to be related to
fujitsu_laptop module, does it?

Regards
jhs

2013-11-30 11:30 GMT+01:00 Jonathan Woithe <jwoithe@just42.net>:
> Jo Jan
>
> On Fri, Nov 22, 2013 at 08:36:23PM +0100, Jan Hinnerk Stosch wrote:
>> I just wanted to ask whether there is something I could do about this?
>
> I'm not sure.  To recap what you have found:
>
>> since commit "a57f7f9175b8ccbc9df83ac13860488913115de4" (bisected) the
>> wireless key on my Fujitsu E753 is not working any more. If I boot up
>> with parameter "acpi_osi="!Windows 2012"" it works, though.
>> I filed an ACPI-bug
>> (https://bugzilla.kernel.org/show_bug.cgi?id=60161), but was told
>> (https://bugzilla.kernel.org/show_bug.cgi?id=60161#c5) that this is
>> caused by the fujitsu_laptop driver. So I report it here. Should I
>> file a bug anywhere or is it enough to write it to the list?
>
> You then answered a followup question which indicated that if fujitsu_laptop
> module is *not* loaded the key still doesn't work.
>
> Does the key work without fujitsu_laptop loaded on a kernel before the
> commit in question?
>
> As I have said previously I am not sure I am in a position to offer
> assistance here.  Firstly I do not have access to any Fujitsu hardware other
> than my S7020, and that hardware does not implement a wireless key like the
> E753 does.  Secondly, I do not presently understand how the OSI changes in
> commit a57f7f9175b8ccbc9df83ac13860488913115de4 could affect the
> fujitsu_laptop module and break the handling of this "wireless key".
>
>> Should I ask in another mailing list or sth?
>
> I don't think so.  Platform-driver-x86 is the appropriate mailing list.
>
> As I said in an earlier email:
>
>> If someone could explain what the new _OSI string could be doing (or point
>> to documentation) I could try to work out what might be happening.  I am a
>> little restricted in what I can do though because I don't have fujitsu
>> hardware that supports the software RF kill switch.
>
> Understanding what this _OSI string does and how it could be causing
> fujitsu_laptop to loose control of the "wireless" key is fundamental to
> solving this problem.  I do not have this knowledge at this point in time,
> and I do not have the time available right now to start looking into the
> whole OSI thing from scratch.  Not having hardware affected by the problem
> also makes it difficult for me to investigate.
>
> To make progress with this issue, we need to work out what this OSI change
> has done to the seemingly unrelated code in fujitsu_laptop.  This involves
> working out what OSI does in the first place.
>
> Sorry that I can't be more definitive than this.
>
> Regards
>   jonathan

Re: Wireless key on Fujitsu E753 not working since a57f7f9175b8ccbc9df83ac13860488913115de4

[Jonathan Woithe]

On Mon, Feb 10, 2014 at 09:42:07AM +0100, Jan Hinnerk Stosch wrote:
> Thanks for your answer and sorry for the delay.

No problem.

> I finally got some time to look into this:
> When I boot with the named boot option the key does work with and
> without fujitsu_laptop being loaded.
> Additionally I found a bug report at launchpad:
> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1159690 with some
> more people affected by this, who are using notebooks from different
> manufacturers, so this doesn't really seem to be related to
> fujitsu_laptop module, does it?

I tend to agree.  The fact that the named boot option makes this work
regardless of whether fujitsu_laptop is loaded or not indicates to me that
the mechanism driving this key on your laptop does not involve
fujitsu_laptop.  My guess at present is that on your laptop the key is
implemented in a standard way and therefore the handling of it is taken care
of by generic code.  By and large the code in fujitsu_laptop exists to deal
with features on laptops which were implemented in non-standard ways. 
Affected laptops tend to be older (newer ones tend to comply more closely
with what we'll call standards for the purposes of this discussion).  This
would explain why fujitsu_laptop doesn't have any effect on the key.

As to why the named boot option seems to affect the key, I don't know and
can't offer any sugestions due to a lack of knowledge about the acpi_osi
option and how it affects the ACPI subsystem.

To summarise: based on the tests you have done I don't think this is related
to the fujitsu_laptop module since the key's behaviour is not affected by
the presence or absence of this module.

Regards
  jonathan

Re: Wireless key on Fujitsu E753 not working since a57f7f9175b8ccbc9df83ac13860488913115de4

[Jan Hinnerk Stosch]

Ok, thanks for your confirmation.

Regards,

jhs

2014-02-17 0:06 GMT+01:00 Jonathan Woithe <jwoithe@just42.net>:
> On Mon, Feb 10, 2014 at 09:42:07AM +0100, Jan Hinnerk Stosch wrote:
>> Thanks for your answer and sorry for the delay.
>
> No problem.
>
>> I finally got some time to look into this:
>> When I boot with the named boot option the key does work with and
>> without fujitsu_laptop being loaded.
>> Additionally I found a bug report at launchpad:
>> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1159690 with some
>> more people affected by this, who are using notebooks from different
>> manufacturers, so this doesn't really seem to be related to
>> fujitsu_laptop module, does it?
>
> I tend to agree.  The fact that the named boot option makes this work
> regardless of whether fujitsu_laptop is loaded or not indicates to me that
> the mechanism driving this key on your laptop does not involve
> fujitsu_laptop.  My guess at present is that on your laptop the key is
> implemented in a standard way and therefore the handling of it is taken care
> of by generic code.  By and large the code in fujitsu_laptop exists to deal
> with features on laptops which were implemented in non-standard ways.
> Affected laptops tend to be older (newer ones tend to comply more closely
> with what we'll call standards for the purposes of this discussion).  This
> would explain why fujitsu_laptop doesn't have any effect on the key.
>
> As to why the named boot option seems to affect the key, I don't know and
> can't offer any sugestions due to a lack of knowledge about the acpi_osi
> option and how it affects the ACPI subsystem.
>
> To summarise: based on the tests you have done I don't think this is related
> to the fujitsu_laptop module since the key's behaviour is not affected by
> the presence or absence of this module.
>
> Regards
>   jonathan