On 01/09/2015 04:26 AM, Takashi Iwai wrote:
At Fri, 09 Jan 2015 09:48:42 +0100, Clemens Ladisch wrote:
Paul Bonser wrote:
On 01/08/2015 03:56 PM, Clemens Ladisch wrote:
Paul Bonser wrote:
The Akai MPC Element incorrectly reports its bInterfaceClass as 255, but otherwise implements the USB MIDI spec correctly.
This adds a quirks-table.h entry which allows the device to be recognized as a standard USB MIDI device.
+++ b/sound/usb/quirks-table.h +{
- /* Akai MPC Element */
- USB_DEVICE(0x09e8, 0x0021),
- .driver_info = (unsigned long) & (const struct snd_usb_audio_quirk) {
.ifnum = QUIRK_ANY_INTERFACE,.type = QUIRK_COMPOSITE,.data = & (const struct snd_usb_audio_quirk[]) {{.ifnum = 0,.type = QUIRK_IGNORE_INTERFACE},{.ifnum = 1,.type = QUIRK_MIDI_STANDARD_INTERFACE},{.ifnum = -1}}- }
+},
Why a composite quirk? Does a single quirk entry not work?
A single quirk entry results in an error message in dmesg: "snd-usb-audio: probe of 3-2:1.0 failed with error -5"
That message is normal for unhandled interfaces.
It's still not good to show a message, so using QUIRK_IGNORE_INTERFACE is good enough as a workaround. So I applied it as is to for-next branch now. Though...
Does it actually prevent the device from working?
... it's still interesting why it needs a special handling. If we find out more and have a better fix, let's put on top of the current patch.
Interface 0 is the empty audio control interface (with no endpoints) that standard USB MIDI devices have, but as with interface 1, it has its bInterfaceClass set to 255 instead of 1.
As far as I can tell the error does not prevent the device itself from working. I tried using QUIRK_AUDIO_STANDARD_INTERFACE and I believe it had the same error. I'm guessing this is because it has no endpoints, so the driver doesn't know what to do with it?