HID blacklist for the Apple IR sensor

HID blacklist for the Apple IR sensor

[Tino Keitel]

Hi,

in commit a417a21e10831bca695b4ba9c74f4ddf5a95ac06, the Apple IR sensor
was added to a blacklist so that no HID device is created anymore.
However, as of lirc 0.8.2, there is a lirc driver called "macmini" that
requires that HID device to work. There is also a driver that creates
an input device for the Apple IR sensor. It can be found in the
mactel-linux patches[1], but it is not in the kernel yet. I also don't
know how both sides behave if they are used concurrently.

I think that the blacklist entry for USB_DEVICE_ID_APPLE_IR should be
removed at least as long as the appleir.patch is not part of the
mainline kernel.

Please CC me in replies.

Regards,
Tino

[1] http://mactel-linux.svn.sf.net/viewvc/mactel-linux/trunk/kernel/mactel-patches-2.6.22/

Re: HID blacklist for the Apple IR sensor

[Jiri Kosina]

On Tue, 10 Jul 2007, Tino Keitel wrote:

> in commit a417a21e10831bca695b4ba9c74f4ddf5a95ac06, the Apple IR sensor 
> was added to a blacklist so that no HID device is created anymore. 
> However, as of lirc 0.8.2, there is a lirc driver called "macmini" that 
> requires that HID device to work. There is also a driver that creates an 
> input device for the Apple IR sensor. It can be found in the 
> mactel-linux patches[1], but it is not in the kernel yet. I also don't 
> know how both sides behave if they are used concurrently.

Hi Tino,

so do I understand it correctly that there are two ways to use this device

- through lirc, which of course requires input device to be created by hid
- using appleir.patch, but this one is not yet upstream (what are the 
  obstacles?)

If it is so, and the device is not handled properly by any other in-tree 
driver, we should remove the ignore quirk.

> I think that the blacklist entry for USB_DEVICE_ID_APPLE_IR should be 
> removed at least as long as the appleir.patch is not part of the 
> mainline kernel.

Would you care to send me a patch with proper changelog and Signed-off-by 
line?

Thanks,

-- 
Jiri Kosina

Re: HID blacklist for the Apple IR sensor

[Tino Keitel]

On Tue, Jul 10, 2007 at 22:29:55 +0200, Jiri Kosina wrote:
> On Tue, 10 Jul 2007, Tino Keitel wrote:
> 
> > in commit a417a21e10831bca695b4ba9c74f4ddf5a95ac06, the Apple IR sensor 
> > was added to a blacklist so that no HID device is created anymore. 
> > However, as of lirc 0.8.2, there is a lirc driver called "macmini" that 
> > requires that HID device to work. There is also a driver that creates an 
> > input device for the Apple IR sensor. It can be found in the 
> > mactel-linux patches[1], but it is not in the kernel yet. I also don't 
> > know how both sides behave if they are used concurrently.
> 
> Hi Tino,
> 
> so do I understand it correctly that there are two ways to use this device
> 
> - through lirc, which of course requires input device to be created by hid
> - using appleir.patch, but this one is not yet upstream (what are the 
>   obstacles?)

Correct.

I've tried to enable both, and it looks like the appleir driver doesn't
work. The appleir driver was loaded as a module. Removing the device
from the blacklist would introduce a regression for all users of the
appleir driver with kernels >= 2.6.21 and CONFIG_USB_HIDDEV enabled.

Isn't there a convenient way to (1) use the HID device if desired and
(2) use the appleir and deacticate this particular HID device in this
case?

> 
> If it is so, and the device is not handled properly by any other in-tree 
> driver, we should remove the ignore quirk.

I don't know of any other in-tree driver.

> 
> > I think that the blacklist entry for USB_DEVICE_ID_APPLE_IR should be 
> > removed at least as long as the appleir.patch is not part of the 
> > mainline kernel.
> 
> Would you care to send me a patch with proper changelog and Signed-off-by 
> line?

See the attachment.

Regards,
Tino

Re: HID blacklist for the Apple IR sensor

[Tino Keitel]

[-- Attachment #1: Type: text/plain, Size: 114 bytes --]

On Wed, Jul 11, 2007 at 08:00:16 +0200, Tino Keitel wrote:

[...]

> See the attachment.

Now you really can. :-)

[-- Attachment #2: remove_Apple_IR_from_hiddev_blacklist.diff --]
[-- Type: text/x-diff, Size: 1113 bytes --]

Author: Tino Keitel <tino.keitel@tikei.de>
Date:   Wed Jul 11 07:52:11 +0200

    HID: remove the Applie IR sensor from the hid_blacklist
    
    The IR sensor in some newer Apple computers has no other
    driver in the kernel, yet. However, the macmini driver in lirc
    requires a HID device for the IR sensor.
    
    Signed-off-by: Tino Keitel <tino.keitel@tikei.de>

diff --git a/drivers/hid/usbhid/hid-quirks.c b/drivers/hid/usbhid/hid-quirks.c
index f6c4145..62569e7 100644
--- a/drivers/hid/usbhid/hid-quirks.c
+++ b/drivers/hid/usbhid/hid-quirks.c
@@ -300,7 +300,6 @@ static const struct hid_blacklist {
 	{ USB_VENDOR_ID_AIPTEK, USB_DEVICE_ID_AIPTEK_24, HID_QUIRK_IGNORE },
 	{ USB_VENDOR_ID_AIRCABLE, USB_DEVICE_ID_AIRCABLE1, HID_QUIRK_IGNORE },
 	{ USB_VENDOR_ID_ALCOR, USB_DEVICE_ID_ALCOR_USBRS232, HID_QUIRK_IGNORE },
-	{ USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_IR, HID_QUIRK_IGNORE },
 	{ USB_VENDOR_ID_BERKSHIRE, USB_DEVICE_ID_BERKSHIRE_PCWD, HID_QUIRK_IGNORE },
 	{ USB_VENDOR_ID_CIDC, 0x0103, HID_QUIRK_IGNORE },
 	{ USB_VENDOR_ID_CYPRESS, USB_DEVICE_ID_CYPRESS_HIDCOM, HID_QUIRK_IGNORE },

Re: HID blacklist for the Apple IR sensor

[Soeren Sonnenburg]

On Tue, 2007-07-10 at 20:26 +0200, Tino Keitel wrote:
> Hi,
> 
> in commit a417a21e10831bca695b4ba9c74f4ddf5a95ac06, the Apple IR sensor
> was added to a blacklist so that no HID device is created anymore.
> However, as of lirc 0.8.2, there is a lirc driver called "macmini" that
> requires that HID device to work. There is also a driver that creates
> an input device for the Apple IR sensor. It can be found in the
> mactel-linux patches[1], but it is not in the kernel yet. I also don't
> know how both sides behave if they are used concurrently.

Sorry for the `late' reply... I was travelling...

> I think that the blacklist entry for USB_DEVICE_ID_APPLE_IR should be
> removed at least as long as the appleir.patch is not part of the
> mainline kernel.

If there is indeed a lirc driver for this providing all the
functionality why would one ever want to have any appleir.patch in the
future? I meant to pave the way for appleir to go in, but now I don't
see a need for it as an in-kernel driver anymore.

So I am fine with reverting the blacklist entry... and I am CC'ing this
to Nicolas as I think the appleir.patch should be removed from the
mactel svn for future releases.

Soeren
-- 
Sometimes, there's a moment as you're waking, when you become aware of
the real world around you, but you're still dreaming.

Re: HID blacklist for the Apple IR sensor

[Tino Keitel]

On Wed, Jul 11, 2007 at 16:48:59 +0200, Soeren Sonnenburg wrote:

[...]

> If there is indeed a lirc driver for this providing all the
> functionality why would one ever want to have any appleir.patch in the
> future? I meant to pave the way for appleir to go in, but now I don't
> see a need for it as an in-kernel driver anymore.

When I didn't had lirc running using the input device, some Gnome
thingie seemed to react on key presses, so that the volume could be
adjusted. With the hiddev, this won't work anymore, but I think that
this was just generic functionality because the appleir driver
generated the proper key events for volume adjustments, so nothing
specific to the Apple IR.

Regards,
Tino

Re: HID blacklist for the Apple IR sensor

[Soeren Sonnenburg]

On Wed, 2007-07-11 at 23:37 +0200, Tino Keitel wrote:
> On Wed, Jul 11, 2007 at 16:48:59 +0200, Soeren Sonnenburg wrote:
> 
> [...]
> 
> > If there is indeed a lirc driver for this providing all the
> > functionality why would one ever want to have any appleir.patch in the
> > future? I meant to pave the way for appleir to go in, but now I don't
> > see a need for it as an in-kernel driver anymore.
> 
> When I didn't had lirc running using the input device, some Gnome
> thingie seemed to react on key presses, so that the volume could be
> adjusted. With the hiddev, this won't work anymore, but I think that
> this was just generic functionality because the appleir driver
> generated the proper key events for volume adjustments, so nothing
> specific to the Apple IR.

True, but I think everyone under linux expects infred stuff to work with
lirc and there is probably already some userspace daemon translating
remote control presses into keyboad presses...

Soeren
-- 
Sometimes, there's a moment as you're waking, when you become aware of
the real world around you, but you're still dreaming.