[mythtv-users] TiVo Slide Remote, Linux and Scan Codes

Jarod Wilson jarod at wilsonet.com
Tue Sep 7 14:19:15 UTC 2010


On Tue, Sep 7, 2010 at 9:38 AM, Jarod Wilson <jarod at wilsonet.com> wrote:
> On Tue, Sep 7, 2010 at 8:48 AM, Wayne Thomas <waynethomas69 at gmail.com> wrote:
>> On Tue, Sep 7, 2010 at 7:01 AM, Wayne Thomas <waynethomas69 at gmail.com> wrote:
>>>> Live TV:
>>>> drivers/hid/hid-input.c: HID Event: 0xffff 0x0
>>>> drivers/hid/hid-input.c: HID Event: 0xffff 0x3e
>>>> drivers/hid/hid-input.c: HID Event: 0xffff 0x3e
>>>> drivers/hid/hid-input.c: HID Event: 0xffff 0x0
>>>> drivers/hid/hid-input.c: HID Event: 0xc 0x20
>>>
>>> Can you confirm what that is the output of?
>
> That's from pressing the live tv button. And I can now confirm that
> indeed, most key presses are followed by a trailing 0xc 0x20.
>
>> > I have nothing to compare
>>> to, but that could be the result of a combination key press (i.e
>>> Shift+"+10").  I am also suspicious that the 0xffff events might be
>>> the result of the hid-driver not supporting a particular UP as that
>>> value doesn't make sense (nor does the sequence).
>
> 0xffff is part of the range in the hid spec listed as "vendor-specific".
>
>> ...and looking at jmk's hexdump for the select key...
>>
>>> hexdump /dev/usb/hiddev0
>>> 0000000 0000 000c 0000 0000 0041 000c 0001 0000
>>> 0000010 0041 000c 0000 0000 0000 000c 0001 0000
>>> 0000020 0020 000c 0053 0000
>>
>> ...there does appear to be more than a single 'key-press' occurring.
>> It looks to my like the output represents a key with output 0x000
>> being pressed (which is ignored by hid-input by the way), 0x41 being
>> pressed then released, 0x000 being released, and then, well, I'm not
>> sure what the that last bit is.  Everything for 'select' is on the
>> consumer page and the 0x000 is just ignored anyway which explains why
>> only adding 0x41 to hid-input.c worked.
>> With the output for Live TV the equivalent outputs appear to be on a
>> usage page 0xffff which isn't defined as a relevant page (yet).
>> HID_USAGE_PAGE is defined as 0xffff0000, but that is actually a mask
>> for extracting the usage page as opposed to being the usage page.
>
> Yes, its the mask, but in looking at the spec, its also a valid
> vendor-specific usage page.
>
>> Not
>> sure what the last bit is but it is the same as that for the select
>> key.  I will have a look at the output of some keys that represent
>> multiple key-presses tonight to see if I can work out what this last
>> bit is.
>>
>> If the usage page is in fact 0xffff I would recommend expanding the
>> topseed driver to deal with this as the Input Devs are actually trying
>> to prevent a proliferation of specific quirky drivers.  I would also
>> be tempted to rename the driver to something more appropriate like
>> hid-remotes but that is another story all together...
>
> At the moment, I have all keys working after an 8-line addition to
> hid-input.c. Basically, just mimic the HID_UP_FOOVENDOR cases. Going
> to clean things up a bit and send in an rfc patch, see what Jiri and
> Dmitry and company think.

Haven't posted it just yet, going to do one last sanity build and
test, but here's the patch to date:

http://wilsonet.com/jarod/junk/hid-support-tivo-slide-remote.patch


-- 
Jarod Wilson
jarod at wilsonet.com


More information about the mythtv-users mailing list