linux-kernel - Re: [PATCH 1/3] Input: goodix - add option to disable firmware loading

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <07a7e9e2-0c82-fdc0-2db0-5011b94816e2@redhat.com>
Date:   Mon, 23 Nov 2020 09:46:03 +0100
From:   Hans de Goede <hdegoede@...hat.com>
To:     Dmitry Torokhov <dmitry.torokhov@...il.com>
Cc:     Andrej Valek <andrej.valek@...mens.com>, nick@...anahar.org,
        hadess@...ess.net, linux-input@...r.kernel.org,
        linux-kernel@...r.kernel.org, devicetree@...r.kernel.org
Subject: Re: [PATCH 1/3] Input: goodix - add option to disable firmware
 loading

Hi,

On 11/23/20 7:53 AM, Dmitry Torokhov wrote:
> On Fri, Oct 30, 2020 at 10:56:20AM +0100, Hans de Goede wrote:
>> Hi,
>>
>> On 10/29/20 9:36 PM, Dmitry Torokhov wrote:
>>> Hi Andrej,
>>>
>>> On Thu, Oct 29, 2020 at 06:03:11PM +0100, Andrej Valek wrote:
>>>> Firmware file loadind for GT911 controller takes too much time (~60s).
>>>> There is no check that configuration is the same which is already present.
>>>> This happens always during boot, which makes touchscreen unusable.
>>>>
>>>> Add there an option to prevent firmware file loading, but keep it enabled
>>>> by default.
>>>
>>> I thought that Goodix was losing firmware loading at poweroff. Is this
>>> not the case with this model?
>>
>> So first of all there are 2 sorts of firmware involved with the
>> Goodix touchscreen controllers, the actual firmware and a block
>> of config data for that firmware which I presume adjusts it for
>> the specific (model of) the digitizer which is attached.
>>
>> ATM the mainline Linux driver does not support models where
>> the actual firmware itself needs to be loaded (because they
>> only have RAM, so they come up without firmware).
>>
>> I do have one model tablet with a ROM-less goodix touchpad
>> controller, so if I ever find the time I might add support
>> for loading the actual firmware.
>>
>> So what we are talking about here is just loading the config
>> data and I'm a bit surprised that this take so long.
> 
> So I am still confused about this: is the config stored in RAM or NVRAM?
> I.e. do we actually need to re-load it every time on boot, or it
> supposed to be flashed only when it is changed (or lost)?

I only know about these touchscreens on x86, where we have a BIOS
muddling the waters.

We have seen devices which loose the config over suspend/resume,
which suggests it is in RAM. I recently added a fix for these which
saves the config at boot, which suggests that at least on the device
model with the suspend/resume issue the config is loaded into the chip
by the BIOS.

But I'm not sure that this is the case everywhere. Most other models
likely have the config in NVRAM.

I guess this is the same as with the firmware, and it differs per
model. I know for sure that their are RAM only models which need
the firmware loaded at boot, these are mostly found on ARM devs,
but I have one X86 devices (which currently does not work) which
also has RAM only and thus needs Linux to load the firmware
(which is not supported atm). These RAM only models, presumably
also have only RAM for the config.

Regards,

Hans