Re: [PATCH 1/2] Input: atmel_mxt_ts - add support for Google Pixel 2

From: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
To: Benjamin Tissoires <benjamin.tissoires@gmail.com>,
	Dmitry Torokhov <dmitry.torokhov@gmail.com>
Cc: linux-input <linux-input@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	Nick Dyer <nick.dyer@itdev.co.uk>,
	Yufeng Shen <miletus@chromium.org>,
	Benson Leung <bleung@chromium.org>,
	Daniel Kurtz <djkurtz@chromium.org>,
	Sjoerd Simons <sjoerd.simons@collabora.co.uk>,
	Olof Johansson <olof@lixom.net>
Subject: Re: [PATCH 1/2] Input: atmel_mxt_ts - add support for Google Pixel 2
Date: Thu, 16 Apr 2015 00:23:03 +0200	[thread overview]
Message-ID: <552EE4C7.4010001@collabora.co.uk> (raw)
In-Reply-To: <552ED615.5020301@gmail.com>

Hello Benjamin,

On 04/15/2015 11:20 PM, Benjamin Tissoires wrote:
> Hi guys,
> 
> On Wed, Apr 15, 2015 at 11:58 AM, Javier Martinez Canillas <javier.martinez@collabora.co.uk> wrote:
>> Hello Dmitry,
>>
>> On 04/08/2015 02:26 AM, Dmitry Torokhov wrote:
>>> This change allows atmel_mxt_ts to bind to ACPI-enumerated devices in
>>> Google Pixel 2 (2015).
>>>
>>> While newer version of ACPI standard allow use of device-tree-like
>>> properties in device descriptions, the version of ACPI implemented in
>>> Google BIOS does not support them, and we have to resort to DMI data to
>>> specify exact characteristics of the devices (touchpad vs. touchscreen,
>>> GPIO to button mapping, etc).
>>>
>>> Pixel 1 continues to use i2c devices and platform data created by
>>> chromeos-laptop driver, since ACPI does not enumerate them.
>>>
>>> Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
>>> ---
> [snipped]
>>
>> Patch looks good to me and I've tested it in a Pixel2 Chromebook using
>> evtest:
>>
>> Reviewed-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
>> Tested-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
>>
> 
> Hmm, this is weird. I tried to apply the 2 patches of this series to a 
> 4.0 fedora kernel[1], and the touch{pad|screen} are desperately muted.

Yes, is not going to work with just 4.0, I tested with today's -next.

> The abs_X/Y max shows up as 0, so there is something wrong with either 
> my touchpad (which works fine on ChromeOS) or with the driver.
> There are no differences between 4.0-rcX and the final 4.0 so I suspect 
> there must be something else.
> I also copied the 2 samus-* config files from the ChromeOS root to
> /lib/firmware but this does not change anything.
> 
> In [3], I enabled the debug output of atmel_mxt, and it seems that the 
> table of functions is missing the T9 one, which is the multitouch one... :(
> 

The atmel_mxt error message is somehow misleading. The problem is not
that the T9 multitouch object is missing (in fact that's the only one
supported by the driver in 4.0) but that the atmel chip in the pixel2
uses another multitouch object (T100).

So the driver tries to initialize the input device assuming that is a
T9 one and fails showing the "Invalid object type T9" error.

> Any help would be greatly appreciated so we can fix [2] and support 
> those laptops in Fedora directly.
>

Patches to add proper T100 support are already in Linus' tree but did
not make it to 4.0. So you need to cherry-pick commits:

b23157dc7427 ("Input: atmel_mxt_ts - implement support for T100 touch object") [0]
b6d2d3289f84 ("Input: atmel_mxt_ts - split out touchpad initialisation logic") [1].

Dmitry mentioned that he used a 4.0+ kernel with Fedora's config so I
guess he also was testing with linux-next or latest Linus' master.

> Cheers,
> Benjamin
> 
> 

Best regards,
Javier

[0]: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=b23157dc74272ac8ebffd1a566e3e822dbc3e65f
[1]: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=b6d2d3289f84e9c7449dff04fb959e29207acd80