[libhid-discuss] usb_control_msg

[Charles Lepple]

On Jun 26, 2007, at 7:55 PM, Jorgen Lundman wrote:

> Charles Lepple wrote:
>>
>> Note that libhid can't claim it exclusively under OS X - it just
>> pretends it can. It works in some cases, but it looks like  
>> something is
>> actually using the native HID calls to read from your device  
>> (instead of
>> what I have seen, where the device can't be exclusively claimed, but
>> nothing is reading from it unless you fire up HID Explorer).
>
> Yeah, I did not realise this initially, but appears to be the case.  
> This
> is no good for me and this device (nor my camera but that's not  
> what I'm
> working on). So I would need to branch off and do OS specific code. I
> dislike writing code that only works on one platform, it was  
> attractive
> to keep using libusb, and use libhid as they are portable.

I should rephrase that: libhid (and libusb) can't claim a HID device  
exclusively under OS X *without the help of a kext*.

So you could have cross-platform code if you can get the kext working.

>>
>> You shouldn't have to reboot to use a codeless kext - the most that
>> might be needed is unplugging and replugging the device.
>>
>> That said, I haven't been successful in creating a codeless kext that
>> actually lets me claim the device.
>
> You know, I actually gave that a stab this morning. Added my codeless
> kext, and kextloaded it, claimed all was good, showed the Composite
> device in kextstat. But unplugging and re-plugging in the device, and
> HID manager steals it. (As seen by USB Prober). I can tinker with  
> it in
> HID explorer again. I suspect reboot is needed, but I really do not  
> like
> rebooting. Maybe I'll try over the weekend :)

Where did you install the kext? Or did you just load it in place?

> You know, I came a across a libhid source in SVN which had lots  
> more HID
> magic, is that a future version coming?

The trunk doesn't have much beyond 0.2.16 yet. What SVN revision and/ 
or branch are you referring to?

-- 
Charles Lepple