Re: magicmouse: claimed by neither input, hiddev nor hidraw

[Michael Poole <mdpoole@troilus.org>]

Justin P. Mattock writes:

> On 05/10/2010 02:52 PM, Jiri Kosina wrote:
[snip]
>> This sounds a bit strange.
>>
>> hidraw shouldn't be making too much difference in the case you describe.
>> hidraw is basically just a mean of relaying HID events to userspace so
>> that any driver/application in userspace can access them. But magicmouse
>> driver is written completely in kernelspace.
>>
>> Does anything on your system have /dev/hidraw* nodes open? (you could
>> check by lsof).
>>
>
>
> right now I see
> /dev/hidraw0,1,2,3
>
> ./lsof | grep /dev
> (showing bluetooth)
>
> bluetooth 2020   root    0u      CHR                1,3       0t0 2551
> /dev/null
> bluetooth 2020   root    1u      CHR                1,3       0t0 2551
> /dev/null
> bluetooth 2020   root    2u      CHR                1,3       0t0 2551
> /dev/null
> bluetooth 2020   root   14u      CHR              10,62       0t0 3895
> /dev/rfkill
>
> I can try a bisect on this and see.

A list of which patches you have applied would also be helpful.  The
standard 2.6.33.* kernels predate the merge of the hid-magicmouse
driver, but the dmesg entry you pasted makes it look like the driver was
present.

Michael Poole