[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <a308c0f3-ddab-49bc-8162-e98082576895@gmx.de>
Date: Wed, 30 Jul 2025 19:52:32 +0200
From: Armin Wolf <W_Armin@....de>
To: Gladyshev Ilya <foxido@...ido.dev>
Cc: linux-input@...r.kernel.org, nikita.nikita.krasnov@...il.com,
Hans de Goede <hansg@...nel.org>,
Ilpo Järvinen <ilpo.jarvinen@...ux.intel.com>,
linux-kernel@...r.kernel.org, platform-driver-x86@...r.kernel.org
Subject: Re: [PATCH v2] platform/x86: Add WMI driver for Redmibook keyboard.
Am 30.07.25 um 19:37 schrieb Gladyshev Ilya:
> Thx for review
>
> On 7/30/25 19:33, Armin Wolf wrote:
>>> +
>>> + if (obj->type != ACPI_TYPE_BUFFER) {
>>> + dev_err(&wdev->dev, "Bad response type %u\n", obj->type);
>>> + return;
>>> + }
>>> +
>>> + if (obj->buffer.length != 32) {
>>
>> Please also accept oversized buffers.
>>
> Sorry if this is a stupid question, but isn't any size other than 32 a
> sign of a firmware bug?
>
Theoretically yes, practically no because the Windows WMI-ACPI driver does accept oversized buffers :(.
In order to support ACPI implementations that rely on the Windows behavior your driver should also
accept oversized buffers.
>>> +
>>> + /* AI key quirk */
>>> + if (entry->keycode == KEY_ASSISTANT) {
>>> + value = !(payload & AI_KEY_VALUE_MASK);
>>
>> I would rather check the payload for 0x00011901 here.
> Personally I prefer to think about it as "some value encoded in
> payload" rather than "is it AI button release payload?", because in
> the latter case alternatives are more unclear. It's just a preference,
> btw, would fix if you insist.
>
Alright, in this case you can ignore my comment.
Thanks,
Armin Wolf
> --
> Gladyshev Ilya
>
Powered by blists - more mailing lists