Skip to contentSkip to conversation

Boot time Bluetooth BUG: warning: (value > m) at hid-core.c:793

DISCUSSION Amedee Van Gasse 2 participants 4 messages Mar 5, 2007 – Mar 23, 2007 1 message not in the archive

lore
Share & tools
lore
#0 lore raw
Quoted 7 lines · Jiri Kosina

What puzzles me a bit is a fact that both the bugreporters seem to state pretty clearly that this started happening after some update, am I right?

Or did the hardware before work all the time in the HID mode, without even being attempted to switch to HCI (which seems to cause the bug)?

It _appears_ that the bug reports started after an update of bluez-utils, but this remains to be confirmed. It could be coincidence/temporal correlation (which does not imply causation).

As for myself, I have a Logitech MX5000 keyboard/mouse with usb dongle, and I can confirm that the keyboard/mouse have always worked in HID. But in HID mode, no other bluetooth functions are possible (like pairing with a smartphone).
It's possible to switch from HID to HCI and have a working mouse +bluetooth receiver, but in that case the keyboard can't pair.

I used this script to get functional bluetooth (and unresponsive keyboard):

hid2hci --tohci
sleep 1
/etc/init.d/bluetooth restart
sleep 1

# connect with the mouse
hidd -i hci0 --connect 00:07:61:48:71:DE
# connect with the keyboard
hidd -i hci0 --connect 00:07:61:3F:DA:41

I think the "-i hci0" isn't even needed, but it can't hurt.

This triggers a lot of occurences of the following in my logs:

BUG: at drivers/hid/hid-core.c:785 implement()
 [<f8d59608>] hid_output_report+0x288/0x300 [hid]
 [<f8d7d4c8>] hid_submit_ctrl+0x58/0x200 [usbhid]
 [schedule_timeout+83/208] schedule_timeout+0x53/0xd0
 [finish_wait+45/112] finish_wait+0x2d/0x70
 [<f8d7d78b>] usbhid_submit_report+0x11b/0x1b0 [usbhid]
 [autoremove_wake_function+0/80] autoremove_wake_function+0x0/0x50
 [<f8d7f9d6>] hiddev_ioctl+0x446/0xb40 [usbhid]
 [filemap_nopage+753/928] filemap_nopage+0x2f1/0x3a0
 [kmap_atomic+134/160] kmap_atomic+0x86/0xa0
 [kunmap_atomic+107/112] kunmap_atomic+0x6b/0x70
 [__handle_mm_fault+633/2624] __handle_mm_fault+0x279/0xa40
 [nameidata_to_filp+53/64] nameidata_to_filp+0x35/0x40
 [do_ioctl+120/144] do_ioctl+0x78/0x90
 [vfs_ioctl+92/672] vfs_ioctl+0x5c/0x2a0
 [sys_ioctl+114/144] sys_ioctl+0x72/0x90
 [sysenter_past_esp+105/169] sysenter_past_esp+0x69/0xa9
 =======================

There are a few Ubuntu bug reports that describe the problem, with some log files and other useful info:
https://bugs.launchpad.net/ubuntu/+source/bluez-utils/+bug/32415
https://launchpad.net/ubuntu/+source/bluez-utils/+bug/66884/
I have found similar descriptions in other distros.

I have exactly the same experience as Vincent Fortier described (keyboard working at boot, some keys repeating sometimes, keyboard unresponsive when bluetooth comes up, keyboard works again after dongle unplug/replug, but no bluetooth)

Anyway, is this a kernel problem or is bluez-utils to blame?

PS: I bumped into this topic when googling for "046d:c70e". I'm not a developer (I'm what they call "a user"), but I'm not afraid to get my hands dirty. If you want me to run debug versions or patched kernels, I can do that. After all, I have the hardware and I might give you valuable information (even if I wouldn't recognise it when it spat me in the face).

Signature · 2 lines
Amedee Van Gasse
[email protected]
#1 lore raw

(added Marcel to CC)

Quoted 3 lines · Amedee Van Gasse

It _appears_ that the bug reports started after an update of bluez-utils, but this remains to be confirmed. It could be coincidence/temporal correlation (which does not imply causation).

This would be in fact very interesting to know. 1) does the BUG trigger only with newer versions of bluez-utils? 2) If so, does everything work with these older versions, or do the things fail in a same way, but silently?

Quoted 1 line

Anyway, is this a kernel problem or is bluez-utils to blame?

I would bet that the problem is that we don't know yet the correct vendor-specific report that is needed to be sent to the dongle to switch from HID to HCI mode.

Quoted 6 lines

PS: I bumped into this topic when googling for "046d:c70e". I'm not a developer (I'm what they call "a user"), but I'm not afraid to get my hands dirty. If you want me to run debug versions or patched kernels, I can do that. After all, I have the hardware and I might give you valuable information (even if I wouldn't recognise it when it spat me in the face).

Would you also be capable of catching the report which switches the dongle from HID to HCI in other operating system, for which there already is a support?

Thanks,

Signature · 1 line
Jiri Kosina

A message that is not in the archive (sent off-list or not yet received): 1173191358.9292.4.camel@saruman

#3 lore raw
Quoted 9 lines · Amedee Van Gasse
Quoted 4 lines

This would be in fact very interesting to know. 1) does the BUG trigger only with newer versions of bluez-utils? 2) If so, does everything work with these older versions, or do the things fail in a same way, but silently?

I could install an older version of my distribution and see what happens, but that means I'll have to reserve some disk space and a lot more time. I hope one of the many live cds also support bluetooth, that would speed things up a lot. Will try to do that, but it'll take some time.

I think that for the initial test just installing older version of bluez-utils might be enough to verify whether they do or don't cause the BUG to trigger.

Quoted 4 lines

I installed something called usbsnoop and waved a lot of dead chickens to convince the bluetooth receiver to pair again with the keyboard. Attached is the logfile. I hope this helps, if not just tell me what you need.

Thanks. Looks like there are the following reports that could potentially be used to switch the dongle from HID to HCI (assuming that Logitech didn't change behavior of its devices too much)

ff 80 00 00 03 00
ff 80 80 02 00 00
ff 80 80 01 00 00
(and maybe ff 80 09 01 00 00 ?)

This doesn't look like what is currently supported by hid2hci function switch_logitech(), but I would rather leave this to Marcel.

Thanks,

Signature · 1 line
Jiri Kosina
#4 lore raw
In reply to Amedee Van Gasse
Quoted 3 lines · Amedee Van Gasse

It _appears_ that the bug reports started after an update of bluez-utils, but this remains to be confirmed. It could be coincidence/temporal correlation (which does not imply causation).

Hi Amedee,

did you have time to confirm whether this started to happen after bluez-utils update, as you stated in the mail above some time ago?

Thanks,

Signature · 1 line
Jiri Kosina