Re: [PATCH] HID: logitech-hidpp: Handle timeout differently from busy

[Linux regression tracking (Thorsten Leemhuis)]

On 21.06.23 17:10, Benjamin Tissoires wrote:
> On Wed, Jun 21, 2023 at 4:09 PM Linux regression tracking (Thorsten
> Leemhuis) <regressions@xxxxxxxxxxxxx> wrote:
>>
>> On 15.06.23 13:24, Mark Lord wrote:
>>> On 2023-06-15 03:24 AM, Linux regression tracking (Thorsten Leemhuis) wrote:
>>> ...
>>>> https://bugzilla.kernel.org/show_bug.cgi?id=217412
>>>>
>>>> --- Comment #47 from Mark Blakeney ---
>>>> @Juha, kernel 6.3.7 adds the 2 patches intended to fix this bug and the
>>>> startup delay is now gone. However, I have had 2 cases over the last 5
>>>> days in which I have been running 6.3.7 where my mouse fails to be
>>>> detected at all after startup. I have to pull the Logitech receiver
>>>> out/in to get the mouse working. Never seen this issue before so I
>>>> suspect the patches are not right.
>>>> ```
>>>
>>> I too have had that happen with recent kernels,
>>
>> Ahh, good to know!
>>
>>> but have not yet put
>>> a finger to a specific version or cause.
>>>
>>> Just toggling the power button on the wireless mouse is enough for
>>> it to "re-appear".
>>>
>>> The 5.4.xx kernels never had this issue.  I went straight from those
>>> to the 6.3.xx ones, where it does happen sometimes, both with and without
>>> the recent "delay" fixes.
>>
>> From Mark Blakeney it sounded a lot like this is something that started
>> with 6.3, but would be good to confirm. Which brings me to the reason
>> why I write this mail:
>>
>> Is anyone still working on this? There was radio silence already for a
>> week now. Okay, it's not really urgent, but I guess this should not fall
>> through the cracks.
> 
> Sorry for the radio silence.

No worries, we all have much to do. But I thought it was time for a
quick status inquiry.

> I worked on hidpp yesterday and submitted
> a new patch for a problem that was overlooked in the previous fixes:
> https://lore.kernel.org/linux-input/20230621-logitech-fixes-v1-1-32e70933c0b0@xxxxxxxxxx/

Great, many thx!

> The loop was not properly initializing all its local states, meaning
> that when we encountered a "please retry" from the device, we could
> never do the actual retry and returned a different error than in the
> 6.2 series.
> 
> Would anyone give it a shot?

Might be worth mentioning it in the bugzilla ticket, as I sadly can't CC
those people here. I'll do this for once (konstantin's bugbot will soon
solve this problem...)

Ciao, Thorsten (wearing his 'the Linux kernel's regression tracker' hat)
--
Everything you wanna know about Linux kernel regression tracking:
https://linux-regtracking.leemhuis.info/about/#tldr
If I did something stupid, please tell me, as explained on that page.