On 4/19/20 11:01 AM, Dmitry Torokhov wrote:
> On Sun, Apr 19, 2020 at 09:46:23AM +0200, LuK1337 wrote:
>> From: Łukasz Patron <priv.luk@xxxxxxxxx>
>>
>> Appending { 0x05, 0x20, 0x02, 0x0f, 0x06 } to
>> xboxone_fw2015_init fixes an issue where the
>> controller is somewhat stuck in bluetooth
>> mode until you plug it into Windows PC.
>>
>> Signed-off-by: Łukasz Patron <priv.luk@xxxxxxxxx>
>
> Thank you Łukasz. Could you please tell me what device(s) have you
> observed the issue with? I am a bit worried if this may interfere with
> other devices that currently work fine with the driver. Cameron, Mark,
> do you have any concerns here?
>
Yes, I gave it a try here, and it looks like there's definitely some
breakages for non-updated controllers. It breaks the Covert Forces
edition Xbox One controller (0x045e, 0x02dd) and also another non-upgraded
Xbox One S controller of the same model as the affected controller
(0x045e, 0x02ea, firmware 3.1.1221.0).
On the plus side, it _does_ seem to fix the issue.
>> ---
>> drivers/input/joystick/xpad.c | 3 ++-
>> 1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/input/joystick/xpad.c b/drivers/input/joystick/xpad.c
>> index 6b40a1c68f9f..894fa81f717a 100644
>> --- a/drivers/input/joystick/xpad.c
>> +++ b/drivers/input/joystick/xpad.c
>> @@ -455,7 +455,8 @@ struct xboxone_init_packet {
>> * or later firmware installed (or present from the factory).
>> */
>> static const u8 xboxone_fw2015_init[] = {
>> - 0x05, 0x20, 0x00, 0x01, 0x00
>> + 0x05, 0x20, 0x00, 0x01, 0x00,
>> + 0x05, 0x20, 0x02, 0x0f, 0x06
>> };
One thing that strikes me as odd about this init packet is that it looks
like 2 packets strung together. Most of the other init packets have 0x20
as their second byte, and there's even something resembling a sequence
number in the third byte (0x02). Maybe these are supposed to be separate
init packets?
Hopefully we can find some init sequence that at least the One S controller
can tolerate on all known firmware versions. If not, the firmware rev is
available to us in bcdDevice, but I'd only like to use an explicit version
check as a last resort.
>> /*
>> --
>> 2.26.0
>>
>
> Thanks.
>
Regards,
Cameron
