On Mon, Mar 21, 2022 at 07:22:45PM +0100, Benjamin Tissoires wrote:
Giving a little bit more context here (and quoting Peter).
""" The problem with [37ef4c19b4] is that it removes functionality - before a clickpad was falsely advertised but the button worked, now in the affected devices it simply no longer works because the button code gets filtered. And user-space can't work around this. ... So the main question remains: why are we doing this?
And the answer here is: because libinput can't handle clickpads with right buttons. But that's not really true either, libinput just doesn't want to, and for no other reason than that it's easier to handle it this way. """
So basically, we tried to fix a choice on libinput assuming that all devices are perfect, for the only sake of making it easy for libinput. But the solution prevents further tweaks, and we then need to manually quirk devices in the kernel which involves a slightly heavier difficulty for end users than just dropping a config file or changing a setting in their UI.
With that said, this patch is: Nacked-by: me
José, could you send a revert of 37ef4c19b4, and add "Cc: stable@vger.kernel.org" and all the other tags for the regression tracker bot?
Sure, for reference, this patch should be ignored in favor of: https://lore.kernel.org/linux-input/20220321184404.20025-1-jose.exposito89@g...
Jose