[PATCH] input: ALPS - Add signature for HP Pavilion dm3 laptops

[PATCH] input: ALPS - Add signature for HP Pavilion dm3 laptops

[Chase Douglas]

Tested by a user running Ubuntu 9.10 in the following bug report.

BugLink: http://bugs.launchpad.net/bugs/545307

Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
Cc: stable@kernel.org
---
 drivers/input/mouse/alps.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/drivers/input/mouse/alps.c b/drivers/input/mouse/alps.c
index f93c2c0..432e3ef 100644
--- a/drivers/input/mouse/alps.c
+++ b/drivers/input/mouse/alps.c
@@ -63,6 +63,7 @@ static const struct alps_model_info alps_model_data[] = {
 	{ { 0x62, 0x02, 0x14 }, 0xcf, 0xcf,
 		ALPS_PASS | ALPS_DUALPOINT | ALPS_PS2_INTERLEAVED },
 	{ { 0x73, 0x02, 0x50 }, 0xcf, 0xcf, ALPS_FOUR_BUTTONS },	  /* Dell Vostro 1400 */
+	{ { 0x73, 0x02, 0x64 }, 0xf8, 0xf8, 0 },			  /* HP Pavilion dm3 */
 };
 
 /*
-- 
1.7.0

Re: [PATCH] input: ALPS - Add signature for HP Pavilion dm3 laptops

[Dmitry Torokhov]

On Sat, Mar 27, 2010 at 01:51:51PM -0400, Chase Douglas wrote:
> Tested by a user running Ubuntu 9.10 in the following bug report.
> 
> BugLink: http://bugs.launchpad.net/bugs/545307
> 
> Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
> Cc: stable@kernel.org

Will apply, thank you Chase.

> ---
>  drivers/input/mouse/alps.c |    1 +
>  1 files changed, 1 insertions(+), 0 deletions(-)
> 
> diff --git a/drivers/input/mouse/alps.c b/drivers/input/mouse/alps.c
> index f93c2c0..432e3ef 100644
> --- a/drivers/input/mouse/alps.c
> +++ b/drivers/input/mouse/alps.c
> @@ -63,6 +63,7 @@ static const struct alps_model_info alps_model_data[] = {
>  	{ { 0x62, 0x02, 0x14 }, 0xcf, 0xcf,
>  		ALPS_PASS | ALPS_DUALPOINT | ALPS_PS2_INTERLEAVED },
>  	{ { 0x73, 0x02, 0x50 }, 0xcf, 0xcf, ALPS_FOUR_BUTTONS },	  /* Dell Vostro 1400 */
> +	{ { 0x73, 0x02, 0x64 }, 0xf8, 0xf8, 0 },			  /* HP Pavilion dm3 */
>  };
>  
>  /*
> -- 
> 1.7.0
> 

-- 
Dmitry

Re: [PATCH] input: ALPS - Add signature for HP Pavilion dm3 laptops

[Chase Douglas]

On Sat, Mar 27, 2010 at 12:11 PM, Dmitry Torokhov
<dmitry.torokhov@gmail.com> wrote:
> On Sat, Mar 27, 2010 at 01:51:51PM -0400, Chase Douglas wrote:
>> Tested by a user running Ubuntu 9.10 in the following bug report.
>>
>> BugLink: http://bugs.launchpad.net/bugs/545307
>>
>> Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
>> Cc: stable@kernel.org
>
> Will apply, thank you Chase.

Dmitry,

After more testing, some users found that they no longer could use the
vertical edge scroll of their touchpad. I assumed everything was
working correctly because I had read a response to another user on
this list that if you add a new ID, things work, and you don't see any
protocol error messages in dmesg, then things should be correct. The
testers reported no errors, so I assumed the problem was in the X
synaptics driver. However, after receiving some event reports from a
tester, it may be that the normal ps2 driver is still sending events,
while the alps/synaptics kernel driver is not sending any. Do you have
any insight on what might be going on?

Thanks

-- Chase

Re: [PATCH] input: ALPS - Add signature for HP Pavilion dm3 laptops

[Dmitry Torokhov]

On Mon, Apr 19, 2010 at 01:36:33PM -0700, Chase Douglas wrote:
> On Sat, Mar 27, 2010 at 12:11 PM, Dmitry Torokhov
> <dmitry.torokhov@gmail.com> wrote:
> > On Sat, Mar 27, 2010 at 01:51:51PM -0400, Chase Douglas wrote:
> >> Tested by a user running Ubuntu 9.10 in the following bug report.
> >>
> >> BugLink: http://bugs.launchpad.net/bugs/545307
> >>
> >> Signed-off-by: Chase Douglas <chase.douglas@canonical.com>
> >> Cc: stable@kernel.org
> >
> > Will apply, thank you Chase.
> 
> Dmitry,
> 
> After more testing, some users found that they no longer could use the
> vertical edge scroll of their touchpad. I assumed everything was
> working correctly because I had read a response to another user on
> this list that if you add a new ID, things work, and you don't see any
> protocol error messages in dmesg, then things should be correct. The
> testers reported no errors, so I assumed the problem was in the X
> synaptics driver. However, after receiving some event reports from a
> tester, it may be that the normal ps2 driver is still sending events,
> while the alps/synaptics kernel driver is not sending any. Do you have
> any insight on what might be going on?
> 

That means that our standard magic knock did not work. It seems that
newer version of ALPS are using a different "magic knock" sequence for
them.

-- 
Dmitry

Re: [PATCH] input: ALPS - Add signature for HP Pavilion dm3 laptops

[Chase Douglas]

On Tue, Apr 20, 2010 at 12:29 AM, Dmitry Torokhov
<dmitry.torokhov@gmail.com> wrote:
> On Mon, Apr 19, 2010 at 01:36:33PM -0700, Chase Douglas wrote:
>> After more testing, some users found that they no longer could use the
>> vertical edge scroll of their touchpad. I assumed everything was
>> working correctly because I had read a response to another user on
>> this list that if you add a new ID, things work, and you don't see any
>> protocol error messages in dmesg, then things should be correct. The
>> testers reported no errors, so I assumed the problem was in the X
>> synaptics driver. However, after receiving some event reports from a
>> tester, it may be that the normal ps2 driver is still sending events,
>> while the alps/synaptics kernel driver is not sending any. Do you have
>> any insight on what might be going on?
>>
>
> That means that our standard magic knock did not work. It seems that
> newer version of ALPS are using a different "magic knock" sequence for
> them.

How can we find out what the new knock sequence is?

-- Chase

Re: [PATCH] input: ALPS - Add signature for HP Pavilion dm3 laptops

[Chase Douglas]

On Tue, Apr 20, 2010 at 9:01 AM, Chase Douglas
<chase.douglas@canonical.com> wrote:
> On Tue, Apr 20, 2010 at 12:29 AM, Dmitry Torokhov
> <dmitry.torokhov@gmail.com> wrote:
>> On Mon, Apr 19, 2010 at 01:36:33PM -0700, Chase Douglas wrote:
>>> After more testing, some users found that they no longer could use the
>>> vertical edge scroll of their touchpad. I assumed everything was
>>> working correctly because I had read a response to another user on
>>> this list that if you add a new ID, things work, and you don't see any
>>> protocol error messages in dmesg, then things should be correct. The
>>> testers reported no errors, so I assumed the problem was in the X
>>> synaptics driver. However, after receiving some event reports from a
>>> tester, it may be that the normal ps2 driver is still sending events,
>>> while the alps/synaptics kernel driver is not sending any. Do you have
>>> any insight on what might be going on?
>>>
>>
>> That means that our standard magic knock did not work. It seems that
>> newer version of ALPS are using a different "magic knock" sequence for
>> them.
>
> How can we find out what the new knock sequence is?

I've had more testing done by others with hardware. It's clear to me
that the magic knock sequence in alps.c does not work for this device.
I have tried both with ALPS_PASS as well, with no change. That means
that we have two potential ways to go from here:

a. Revert the signature and use the ImPS/2 Generic Wheel Mouse driver,
which provides wheel events automatically but without any absolute
event support
b. Leave the signature as is, which disables wheel events and enables
no extra functionality

I would call b a regression. Maybe we can figure out the new knock
protocol before the change propagates from tip to distros, but I think
it's better to revert the signature and reapply it later when we find
the right sequence. What do you think?

-- Chase

Re: [PATCH] input: ALPS - Add signature for HP Pavilion dm3 laptops

[Dmitry Torokhov]

On Tue, Apr 20, 2010 at 11:47:20PM -0400, Chase Douglas wrote:
> On Tue, Apr 20, 2010 at 9:01 AM, Chase Douglas
> <chase.douglas@canonical.com> wrote:
> > On Tue, Apr 20, 2010 at 12:29 AM, Dmitry Torokhov
> > <dmitry.torokhov@gmail.com> wrote:
> >> On Mon, Apr 19, 2010 at 01:36:33PM -0700, Chase Douglas wrote:
> >>> After more testing, some users found that they no longer could use the
> >>> vertical edge scroll of their touchpad. I assumed everything was
> >>> working correctly because I had read a response to another user on
> >>> this list that if you add a new ID, things work, and you don't see any
> >>> protocol error messages in dmesg, then things should be correct. The
> >>> testers reported no errors, so I assumed the problem was in the X
> >>> synaptics driver. However, after receiving some event reports from a
> >>> tester, it may be that the normal ps2 driver is still sending events,
> >>> while the alps/synaptics kernel driver is not sending any. Do you have
> >>> any insight on what might be going on?
> >>>
> >>
> >> That means that our standard magic knock did not work. It seems that
> >> newer version of ALPS are using a different "magic knock" sequence for
> >> them.
> >
> > How can we find out what the new knock sequence is?
> 

One way would be to run the $OTHER_OS with proper driver inside a VM and
pass-through all commands from that OS to real PS/2 hardware recodring
the sequence in process.

> I've had more testing done by others with hardware. It's clear to me
> that the magic knock sequence in alps.c does not work for this device.
> I have tried both with ALPS_PASS as well, with no change. That means
> that we have two potential ways to go from here:
> 
> a. Revert the signature and use the ImPS/2 Generic Wheel Mouse driver,
> which provides wheel events automatically but without any absolute
> event support
> b. Leave the signature as is, which disables wheel events and enables
> no extra functionality
> 
> I would call b a regression. Maybe we can figure out the new knock
> protocol before the change propagates from tip to distros, but I think
> it's better to revert the signature and reapply it later when we find
> the right sequence. What do you think?
> 

The change has been reverted in my tree and will be reverted in mainline
when I request next pull (before .34 is released).

Thanks.

-- 
Dmitry

Re: [PATCH] input: ALPS - Add signature for HP Pavilion dm3 laptops

[Chase Douglas]

On Wed, Apr 21, 2010 at 12:10 AM, Dmitry Torokhov
<dmitry.torokhov@gmail.com> wrote:
> On Tue, Apr 20, 2010 at 11:47:20PM -0400, Chase Douglas wrote:
>> On Tue, Apr 20, 2010 at 9:01 AM, Chase Douglas
>> <chase.douglas@canonical.com> wrote:
>> > On Tue, Apr 20, 2010 at 12:29 AM, Dmitry Torokhov
>> > <dmitry.torokhov@gmail.com> wrote:
>> >> That means that our standard magic knock did not work. It seems that
>> >> newer version of ALPS are using a different "magic knock" sequence for
>> >> them.
>> >
>> > How can we find out what the new knock sequence is?
>>
>
> One way would be to run the $OTHER_OS with proper driver inside a VM and
> pass-through all commands from that OS to real PS/2 hardware recodring
> the sequence in process.

Do you know of any documentation on how to do this? I've tried various
google searches, but I'm having difficulty finding information on
sniffing the ps/2 protocol through a vm.

Thanks

-- Chase

Re: [PATCH] input: ALPS - Add signature for HP Pavilion dm3 laptops

[Dmitry Torokhov]

On Wed, Apr 21, 2010 at 08:15:17AM -0400, Chase Douglas wrote:
> On Wed, Apr 21, 2010 at 12:10 AM, Dmitry Torokhov
> <dmitry.torokhov@gmail.com> wrote:
> > On Tue, Apr 20, 2010 at 11:47:20PM -0400, Chase Douglas wrote:
> >> On Tue, Apr 20, 2010 at 9:01 AM, Chase Douglas
> >> <chase.douglas@canonical.com> wrote:
> >> > On Tue, Apr 20, 2010 at 12:29 AM, Dmitry Torokhov
> >> > <dmitry.torokhov@gmail.com> wrote:
> >> >> That means that our standard magic knock did not work. It seems that
> >> >> newer version of ALPS are using a different "magic knock" sequence for
> >> >> them.
> >> >
> >> > How can we find out what the new knock sequence is?
> >>
> >
> > One way would be to run the $OTHER_OS with proper driver inside a VM and
> > pass-through all commands from that OS to real PS/2 hardware recodring
> > the sequence in process.
> 
> Do you know of any documentation on how to do this? I've tried various
> google searches, but I'm having difficulty finding information on
> sniffing the ps/2 protocol through a vm.
> 

I do not know of existing setups. I'd expect you'll have to pick in
QEMU/KVM guts and make it talk to psaux interface (the old, 2.4-style
PS/2 port). Then, on the host side, set up serio_raw driver instead of
psmouse to that VM can talk to real hardware.

-- 
Dmitry