Apple Bluetooth Mouse and Keyboard pairing broken

[hungarianhc]

Hi there,

This is my first bug I've ever submitted so please let me know what I can do to make it more useful.

Hardware: iBook G4 1.25ghz 12", 1 gb RAM
Apple Bluetooth Keyboard
Apple Bluetooth Mouse

Something breaks Bluetooth somewhere along the bootup sequence. When LILO starts up, i can use my bluetooth keyboard to select whether to boot in linux, os x, or CD.

However, once GDM comes up, my bluetooth devices are no longer functional. This problem does not exist in Flight 3 so it must be new to flight 4. I re-tested flight 3 to confirm.

Upon running the command, "hidd --search", "no devices in range or visible" is returned to me. Please let me know if there are any tests I can conduct to make this more informative. Thank you.

-Kevin

[sam tygier (samtygier)]

this problem was also discussed at http://ubuntuforums.org/showthread.php?t=142987 and http://ubuntuforums.org/showthread.php?t=143388

people are removing bluez-cups bluez-pcmcia-support bluez-pin bluez-utils as a work around.

[sam tygier (samtygier)]

several discussions and a dupe bug report seems like evidence enough to confirm.

[hungarianhc (kev-hanson)]

I can confirm that this bug still exists as of March 20 in the daily build of Ubuntu. I can also confirm that removing bluez-* solved my problem.

[Arjan Nieuwenhuis (nieuwenhuis-a)]

i disabled the use of bluez-utils with bum (bootup manager). that worked for me. I got a Logitech dinovo set.

[mtron (mtron)]

in this thread http://ubuntuforums.org/showthread.php?t=159189, dapper beta useres (fresh install & upgrade from flight 6) reported that the init.d script "sudo /etc/init.d/bluez-utils restart" does not work. Also after pluggin' in the transreciever, a user had to manually "up" the subsystem.

[VinceLe (legoll)]

I tried Ubuntu 6.06 beta 2 live ppc on my macmini
and have the same problem, my bluetooth keyboard does not work
I managed to cut'n'paste some commands with the mouse but this is painful :-(
Maybe there is a way to get a sshd running on the live distro, I'll try that...

hcitool search said it found my keyboard
dmesg showed the device as being recognised
lsusb -v showed it connected to an internal usb port
/etc/init.d/bluez-utils restart didn't manage to get key presses passed to Xorg...

I didn't find a way to try the "workaround" I read in a forum post, that is to remove bluez-* packages

[VinceLe (legoll)]

Changed status to confirmed

[Zhares (launchpad1)]

Hi
I have a PowerMac G5 bi 2GHz with internal bluetooth. I've got apple bluetooth keyboard and logitech MX900.
I've just updated to Dapper and I've download bluez 2.25 source code to do some research on this bug.
The problem occurs in ./hidd/main.c, line 313, in function int create_device:
   err = ioctl(ctl, HIDPCONNADD, &req);
This call returns -1 and set errno to 14. An error is displayed in stderr (sometimes also in syslog) when I try to connect manualy with:
   ./hidd --connect <mac-address>
With breezy, I used to keep the bluetooth dongle in HID mode (hardware mode used in bootloader) to avoid the problem, but I had a similar error when I tried to connect using bluez.
I'm not confident enough with ioctl() to understand the error. Maybe someone can help with this call.
There seems to be a duplicate in #41786

[Mark (azovap002)]

After following all advice about removing bluez and turning sound off: our test unit is still completely unusable.

We apparently have the worst of all possible worlds - a G5 system with the known sound and power (fan) problems, but also the bluetooth problems.

The root prompt from "Linux single" occasionally just hangs, particularly inside the vim editor. We managed to edit gdm.conf to shut sound off and also use apt to remove bluez, but only with multiple reboots between editing attempts.

Regular bootup produces the GDM login, which is now barely functional We can type the user login data, but then GDM simply hangs with a stationary mouse cursor on screen and nothing else. We have original Apple wireless mouse and keyboard (no USB, sorry).

DAPPER FINAL DOES NOT FUNCTION ON OUR SYSTEM AT ALL EVEN WITH THE KNOWN WORKAROUNDS.

We're puzzled by the priority assignments of these dramatic hardware bugs. Bugs which completely prevent use of the OS merit maximum priority. Some of these bugs have been lying around a long time, too.

[Lothar (lothar-tradescape)]

I installed Kubuntu (final) and I have the same problems. My Targus Bluetooth mouse only works if I disable the bluez service and restart. Can anybody find how to fix this without disabling the whole bluez stuff, please.

[AJ Slater (aj-slater)]

Confirmed with Logitech diNovo setup. The issue is a GDM interaction. Perhapse a bug should be filed against GDM as well?

[Zhares (launchpad1)]

This is not a GDM bug but a bluez bug. I've tried with bluez 3 as well and got the same error. Or maybe it is a bug in the bluetooth module. It's probably a 64-bit or bigendian error since this code has been writen for x86.

There are 2 ways of having the internal bluetooth connect to keyboard and mouse to work with a G5:
- the HID mode: this is the default mode at boot time, the one wich is used in yaboot and openfirmware. It is also the mode used when bluez is not installed. In this mode, the bluetooth is totaly transparent to linux, which only sees a standard keyboard and mouce. This mode is NOT the good one: The mouse wheel is not supported, neither is the sound keys or the eject key. And it is impossible to connect any other bluetooth device (phone, headset, ...) to the computer.
- the HCI mode, which is the normal bluetooth mode. Mice and keyboards are detected using bluez utils, and linked to the computer like any other device. This is supposed to happen at boot time so that the input device are ready to talk to GDM. The connection and pairing works fine but when it comes to get live data, the hidd program failes a ioctl call.
I've posted this bug in the bluez mailing list but nobody seems to know what it is about.
I would be very happy to track that bug but I am no idea where to start. Maybe I'll try bluez-dev soon.

[pdecat (pdecat)]

Problem confirmed on stock Ubuntu 6.06 and fully updated as of 2006-08-05 (Linux myhost 2.6.15-26-k7 #1 SMP PREEMPT Thu Aug 3 03:40:32 UTC 2006 i686 GNU/Linux)

Fixed by commenting out hid2hci commands in /etc/init.d/bluez-utils :

$ diff /etc/init.d/bluez-utils /etc/init.d/bluez-utils.bak
202c202
< #enable_hci_input || true
---
> enable_hci_input || true
211c211
< #disable_hci_input || true
---
> disable_hci_input || true

[Zhares (launchpad1)]

Again, this is not a fix but a workaround that will disable bluetooth at the OS level and therefore prevent the bluetooth dongle to be used to connect to phones, PDAs, headset.

[Onkar Shinde (onkarshinde)]

Can somebody attach /etc/default/bluez-utils file?
Can someone confirm that HIDD_ENABLED is 1 in /etc/default/bluez-utils

Is this still issue with Ubuntu 6.06 (Dapper final)?

[pdecat (pdecat)]

Here you are.

It has "HIDD_ENABLED=0" in it.

[pdecat (pdecat)]

Forgot to specify it comes from Ubuntu 6.06.1.

[Onkar Shinde (onkarshinde)]

Wrong package.

[Onkar Shinde (onkarshinde)]

Rejecting as the problem is with configuration.
Please set HIDD_ENABLED=1 in /etc/default/bluez-utils and restart the bluetooth services. (sudo /etc/init.d/bluez-utils restart)

Reopen the bug is you face problems after changing configuration.

[Lothar (lothar-tradescape)]

changing HIDD_ENABLED=1 (was recommended in some other thread, too) did not make my mouse work on Dapper (I already changed it months ago).

[Lothar (lothar-tradescape)]

I forgot to say that it doesn't work with Edgy either (I downloaded the ISO and ran it from CD only).

[Onkar Shinde (onkarshinde)]

@Lothar

Can you please tell me if following instructions on https://help.ubuntu.com/community/BluetoothSetup helps you?

If it doesn't then please paste/attach outputs of following commands.
* hcitool dev
* lsmod |grep hidp
* grep hidp /etc/modules
* lsusb (assuming that you are using bluetooth dongle)
* dmesg | grep -i blue

[AJ Slater (aj-slater)]

This is still an issue on Dapper. Onkar's rejection reasons are wrong. Changing the HIDD_ENABLED setting does not work.
There is a problem with bluetooth and GDM interaction.
IIRC the problem is dbus permissions related.

This bug makes Ubunutu look very bad.
If you boot stock Dapper with a bluetooth keyboard (such as my Logitech DiNovo) The keyboard and mouse cease functioning as soon as GDM starts.

I'll try to get a daemon.log for you later today.

[Lothar (lothar-tradescape)]

 Onkar Shinde wrote:
>If it doesn't then please paste/attach outputs of following commands.
> * hcitool dev
> * lsmod |grep hidp
> * grep hidp /etc/modules
> * lsusb (assuming that you are using bluetooth dongle)
> * dmesg | grep -i blue

here they are. Let me know if you need anything else

lothar@janus$ hcitool dev
Devices:
        hci1 00:0C:55:FD:16:37

lothar@janus$ lsmod |grep hidp
hidp 36096 0
l2cap 30464 6 rfcomm,hidp
bluetooth 59396 8 rfcomm,hidp,l2cap,hci_usb

lothar@janus$ grep hidp /etc/modules

lothar@janus$ lsusb
Bus 003 Device 001: ID 0000:0000
Bus 001 Device 003: ID 0a5c:2009 Broadcom Corp.
Bus 001 Device 002: ID 0a5c:3535 Broadcom Corp.
Bus 001 Device 001: ID 0000:0000
Bus 002 Device 004: ID 04d9:048e Holtek Semiconductor, Inc.
Bus 002 Device 001: ID 0000:0000

lothar@janus$ dmesg | grep -i blue
[ 32.151269] Bluetooth: Core ver 2.8
[ 32.151275] Bluetooth: HCI device and connection manager initialized
[ 32.151284] Bluetooth: HCI socket layer initialized
[ 32.218292] Bluetooth: HCI USB driver ver 2.9
[ 65.430917] Bluetooth: L2CAP ver 2.8
[ 65.430921] Bluetooth: L2CAP socket layer initialized
[ 65.439003] Bluetooth: HIDP (Human Interface Emulation) ver 1.1-mh1
[ 70.429286] Bluetooth: RFCOMM socket layer initialized
[ 70.429295] Bluetooth: RFCOMM TTY layer initialized
[ 70.429296] Bluetooth: RFCOMM ver 1.7

from /var/log/syslog
Oct 12 13:46:12 janus kernel: [ 5731.545033] usb 2-1: new low speed USB device using ohci_hcd and address 4
Oct 12 13:46:12 janus kernel: [ 5731.622586] input: HID 04d9:048e as /class/input/input5
Oct 12 13:46:12 janus kernel: [ 5731.622618] input: USB HID v1.10 Mouse [HID 04d9:048e] on usb-0000:00:02.1-1
Oct 12 13:46:19 janus kernel: [ 5733.894290] ohci_hcd 0000:00:02.0: wakeup
Oct 12 13:46:19 janus kernel: [ 5734.022062] usb 1-2: new full speed USB device using ohci_hcd and address 2
Oct 12 13:46:20 janus kernel: [ 5734.093948] hub 1-2:1.0: USB hub found
Oct 12 13:46:20 janus kernel: [ 5734.094921] hub 1-2:1.0: 3 ports detected
Oct 12 13:46:20 janus kernel: [ 5734.203728] usb 1-2.1: new full speed USB device using ohci_hcd and address 3
Oct 12 13:46:20 janus hcid[4667]: HCI dev 1 registered
Oct 12 13:46:20 janus kernel: [ 5734.315526] usb 1-2.2: new full speed USB device using ohci_hcd and address 4
Oct 12 13:46:21 janus hcid[4667]: HCI dev 1 up
Oct 12 13:46:21 janus hcid[4667]: Starting security manager 1
Oct 12 13:46:21 janus kernel: [ 5734.369476] input: HID 0a5c:3502 as /class/input/input6
Oct 12 13:46:21 janus kernel: [ 5734.369493] input: USB HID v1.11 Keyboard [HID 0a5c:3502] on usb-0000:00:02.0-2.2
Oct 12 13:46:21 janus kernel: [ 5734.439304] usb 1-2.3: new full speed USB device using ohci_hcd and address 5
Oct 12 13:46:21 janus kernel: [ 5734.482919] input: HID 0a5c:3503 as /class/input/input7
Oct 12 13:46:21 janus kernel: [ 5734.482954] input: USB HID v1.11 Mouse [HID 0a5c:3503] on usb-0000:00:02.0-2.3
Oct 12 13:46:26 janus kernel: [ 5736.239401] usb 1-2.2: USB disconnect, address 4
Oct 12 13:46:26 janus kernel: [ 5736.285984] usb 1-2.3: USB disconnect, address 5
Oct 12 13:48:35 janus kernel: [ 5782.374127] usb 2-2: USB d...

[Lothar (lothar-tradescape)]

Hi Onkar,

if you want you can buy a Targus AMB01 mouse to make it work with Dapper.

http://froogle.google.com/froogle?q=Targus+AMB01&hl=en&lr=&ie=UTF-8&oe=UTF-8&safe=off&sa=X&oi=froogle&ct=title

If you can make it work on Dapper+ (updated *deb packets via apt-get update/install, no manual hacking), I am willing to reimburse you for the purchase via paypal (max. $50, as most expensive one on froogle is $47.95).

[Lothar (lothar-tradescape)]

@ Onkar Shinde

can you please reopen the bug? I could not do it. Thanks!

[Onkar Shinde (onkarshinde)]

Reopening since many people are having problem.

[Onkar Shinde (onkarshinde)]

@Lothar, AJ Slater

I totally agree with you that this is bad bug and should be fixed for Dapper.
I am no developer. I am just another user. I was just trying to help. Looks like my assumptions proved wrong.

@Lothar
One more question. Is the bluetooth device you are using (dongle or on board) is based on broadcom chipset. I am asking this because I couldn't understand which one of your 3 USB devices was bluetooth?

[Daniel Holbach (dholbach)]

Can somebody with an Edgy of today (2006-10-13)? A bunch of changes went in yesterday and the day before and it'd interesting to hear back.

[Onkar Shinde (onkarshinde)]

@Daniel

Looks like most of the reporters want it fixed/working out of box on Dapper since it is LTS release. ;-)

Following are my observations. See if that helps. Also correct me if I am wrong.
1. The problem is mostly on Apple hardware.
2. AFAIK, Apple uses broadcom cheapset based devices.
3. It looks like mouse/keyboards are getting disabled when login screen shows up (GDM starts).

I have neither Apple hardware nor bluetooth keyboad/mouse. So I can't provide more inputs.

[Daniel Holbach (dholbach)]

Onkar: it'd help to know if it works in Edgy. If it does, we need to find out, which obvious and small fix we can apply to Dapper.

[Lothar (lothar-tradescape)]

Onkar, Daniel:

I can confirm that the Mouse I use (Targus AMB01) uses a Broadcom Chipset (see lsusb above)

I tested it with Dapper on two different laptops with 32 and 64 bit. I also tested Edgy (I think it was Knot3). None of them work out of the box.

As said above I am willing to sponsor one developer with a Targus AMB01. Take advantage of that ;-)

Good night for now ;-)

[Daniel Holbach (dholbach)]

Please somebody try with a recent Edgy. Thanks a lot.

[hungarianhc (kev-hanson)]

Everyone here knows that you can still get a full working bluetooth setup if
you remove all bluez* packages right? everything works if you do that. This
problem is not specific to apples either. It has existed for way too long.

-Kevin

On 10/13/06, Daniel Holbach <email address hidden> wrote:
>
> Please somebody try with a recent Edgy. Thanks a lot.
>
> --
> Bluetooth Mouse and Keyboard Broken in Dapper Flight 4 & 6.06 beta 2
> https://launchpad.net/bugs/32415
>

[Etienne Goyer (etienne-goyer-outlands)]

Daniel,

I have an MS BT mouse/kb combo at home and two different USB BT dongle, but the rig is at home and I am not. I will test on the very latest Edgy (I am updating daily), tomorrow or Sunday at the latest. Stay tuned !

[Lothar (lothar-tradescape)]

@Daniel,

I can test.
where can I download an iso to test what you consider "a recent Edgy" ?
I need a live CD, as I don't want to overwrite my Dapper installation.

[Lothar (lothar-tradescape)]

@hungarianhc

no, deactivating/removing the bluez* packages is NO help.
it disables to use any other bluetooth devices besides HID and the scroll whell for mice does NOT work in this basic mode either.

[hungarianhc (kev-hanson)]

I don't have a scroll function on my mouse but removing bluez still
leaves my bt mouse and keyboard functional

On 10/13/06, Lothar <email address hidden> wrote:
> @hungarianhc
>
> no, deactivating/removing the bluez* packages is NO help.
> it disables to use any other bluetooth devices besides HID and the scroll
> whell for mice does NOT work in this basic mode either.
>
> --
> Bluetooth Mouse and Keyboard Broken in Dapper Flight 4 & 6.06 beta 2
> https://launchpad.net/bugs/32415
>

[Daniel Holbach (dholbach)]

Lothar: http://cdimage.ubuntu.com/daily-live/current/

[Zhares (launchpad1)]

Hi,

Don't waste your time testing all dongles and mouses, it is not device related. The configuration files are fine as weel, since they work on other architectures.
It is a bug in the bluetooth module (or ioctl) on ppc and ppc-64 architecture, which is probably a big-endian problem.
Has you can see in my first post, there is an failure in a call to ioctl on the hidd deamon. That is were we have to trace the bug. We need a kernel hacker to fix that one.

[pdecat (pdecat)]

I just downloaded and burned the latest edgy live CD from http://cdimage.ubuntu.com/daily-live/current/edgy-desktop-i386.iso (dated 14-Oct-2006 07:43).

I have a Logitech DiNovo Media Desktop set : bluetooth adapter + bt keyboard + bt mouse + bt keypad.

When I boot from the freshly burned CD, the keyboard works at boot menu (can choose language).

Once gnome is started, none of the mouse/keyboard/keypad works (as for Dapper 6.06 and 6.06.1 livecd).

Unplugging/replugging the bluetooth adapter makes everything work.

I freed 3GB, launched the install, everything went fine until reboot : same as for the live CD, nothing works.

Again, unplugging/replugging the bluetooth adapter makes everything work.

Attached are the dmesg, lspci and lsusb outputs from the live CD. I can provide those from the installed system if needed.

Best regards,
Patrick.

[Lothar (lothar-tradescape)]

@ Zhares
It has nothing to do with PPC/Endianess, as it fails on My Dell (intel P4) and my HP (AMD64) Notebooks. I don't know why there is so much ignorance about this bug. There are many people with DIFFERENT architectures reporting problems and still people like you pretend that everything is OK. That's what really hurts Ubuntus reputation.

@Daniel
 I tested it with the latest ISO downloaded from
http://cdimage.ubuntu.com/daily-live/current/

It still behaves like before. The mouse works for a brief period and then
ceases to work. I attached the logs and the output of some programs to a mail. But it seems not to get to the forum. Did you get the cc'ed mail sent to you (with the logs)?

[Lothar (lothar-tradescape)]

I tested it with the latest ISO downloaded from
 http://cdimage.ubuntu.com/daily-live/current/

It still behaves like before. The mouse works for a brief period and then
 ceases to work.

here are the logs, too as a tar.gz file.

[Zhares (launchpad1)]

@Lothar

I don't pretend that everything is OK, I just notice that people seems to look at the wrong place. There may be a problem on the config files, but that's not what I've noticed. I've seen an error 14 after a call to ioctl in hidd (with different dongles and devices) and some people are talking of disabling bluetooth (unplug/re-plug the dongle) to make it work in hardware mode. That is what hurts Ubuntu's reputation, I think. I'm sorry i've been rude in my previous post and I apologize for that.

What you describe seems to be the problem that everybody has: there is nothing that automatically detect you mouse over bluetooth. You need to run or setup hidd like explained everywhere:
sudo hidd --connect aa:bb:cc:dd:ee:ff
I agree that is a bug, but that is far away from the problems we have on ppc.

[Lothar (lothar-tradescape)]

@ Zhares
sorry if I misunderstud you; but I also never had any change to get my mouse to work. not even with sudo hidd --connect aa:bb:cc:dd:ee:ff. Maybe it's due to the cause that it uses a broadcom BT chipset (dongle) and as far as I know most Macs use broadcom, too. It's however not a kernel problem, as the mouse and the keyboard works in the (simpler) emulation mode that the kernel provides. It seems to be a problem in the bluez package, as the mouse ceases to work as soon as it is switched on.
I have sent a couple of posts to the bluez newsgroup, but nobody answered back. Maybe it affects only a relatively small set of hardware (like broadcom) and they think we are nuts by claiming it doesn;t work, as it works for so many others.
The worst are these people who recommend getting the mouse work in the original kernel (emulation) mode.They make me mad as they just don't get the problem. Disabling bluez is not a fix, it makes the mouse wheel not working and it prohibits the use of any other BT services (like download pictures from a cell phone).

[Lothar (lothar-tradescape)]

here's my strace of hidd when I try to do a connect:
Of course I made sure that hci1 is the dongle of the mouse before I used strace by unplugging replugging.

here's a snipped of dmesg after the strace

[27824.700420] usb 1-1: new full speed USB device using ohci_hcd and address 11
[27824.772316] hub 1-1:1.0: USB hub found
[27824.773288] hub 1-1:1.0: 3 ports detected
[27824.883425] usb 1-1.1: new full speed USB device using ohci_hcd and address 12
[27824.995224] usb 1-1.2: new full speed USB device using ohci_hcd and address 13
[27825.050163] input: HID 0a5c:3502 as /class/input/input10
[27825.050183] input: USB HID v1.11 Keyboard [HID 0a5c:3502] on usb-0000:00:02.0-1.2
[27825.120332] usb 1-1.3: new full speed USB device using ohci_hcd and address 14
[27825.164285] input: HID 0a5c:3503 as /class/input/input11
[27825.164322] input: USB HID v1.11 Mouse [HID 0a5c:3503] on usb-0000:00:02.0-1.3
[27828.086982] usb 1-1.2: USB disconnect, address 13
[27828.132903] usb 1-1.3: USB disconnect, address 14
[27840.172076] eth1: no IPv6 routers present
[27872.649276] hci_cmd_task: hci0 command tx timeout

here's the strace

root@janus$ hcitool dev
Devices:
        hci0 00:10:C6:68:25:23
        hci1 00:0C:55:FD:16:37

root@janus$ strace hidd -n --connect 00:0C:55:FD:16:37
execve("/usr/bin/hidd", ["hidd", "-n", "--connect", "00:0C:55:FD:16:37"], [/* 33 vars */]) = 0
uname({sys="Linux", node="janus", ...}) = 0
brk(0) = 0x507000
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2aaaaaac1000
access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory)
mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2aaaaaac2000
access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory)
open("/etc/ld.so.cache", O_RDONLY) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=70256, ...}) = 0
mmap(NULL, 70256, PROT_READ, MAP_PRIVATE, 3, 0) = 0x2aaaaaac4000
close(3) = 0
access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory)
open("/usr/lib/libbluetooth.so.1", O_RDONLY) = 3
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0`]\0\0\0"..., 640) = 640
fstat(3, {st_mode=S_IFREG|0644, st_size=82600, ...}) = 0
mmap(NULL, 1129472, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x2aaaaabc2000
mprotect(0x2aaaaabd4000, 1055744, PROT_NONE) = 0
mmap(0x2aaaaacd4000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x12000) = 0x2aaaaacd4000
close(3) = 0
access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory)
open("/lib/libm.so.6", O_RDONLY) = 3
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0000>\0\0"..., 640) = 640
fstat(3, {st_mode=S_IFREG|0644, st_size=546000, ...}) = 0
mmap(NULL, 1591784, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x2aaaaacd6000
mprotect(0x2aaaaad5b000, 1047016, PROT_NONE) = 0
mmap(0x2aaaaae5a000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x84000) = 0x2aaaaae5a000
close(3) = 0
access("/etc/ld.so.nohwcap", F_OK) = -1 ENO...

[Etienne Goyer (etienne-goyer-outlands)]

I have the Microsoft Desktop Elite for Bluetooth keyboard/mice combo (described athttp://www.microsoft.com/hardware/mouseandkeyboard/productdetails.aspx?pid=033). I dot have the original Microsoft USB dongle, I use one that describe itself as "Cambridge Silicon Radio, Ltd Bluetooth Dongle (HCI mode)".

I have HIDD_ENABLED=1 in /etc/default/bluetooth, and I have added "--connect xx:xx:..." for both the kb and mouse MAC address in HIDD_OPTIONS.

After a cold boot, with the dongle plugged in, device do not work. If I get to the shell and do "sudo hidd --search", device are not detected and do not work. If I push the "Make/Break Connection" button on the keyboard and mice and do "sudo hidd --search" again, bith device are detected and become functionnal.

If I do not use either device for about 10 minutes, they stop being usable. In this state, they act like after a cold boot; just doing "sudo hidd --search" is not enough, I have to press the "Make/Break Connection" then do "sudo hidd --search" to make them work again.

In Microsoft Windows XP SP2, device are working on cold boot without using the devices "Make/Break Connection" and remain usable even when unused for a long period of time.

My assumption is that this behavior is specific to this pair of device (hardware-related), but I have no other Bluetooth HID device to confirm or infirm that assumption. If ever get my hands on one, I will confirm either way.

I will bring the mouse with me this week, so if you want me to test anything else, I should be able to do so in short order.

[Piotr Szczepanik (piter75)]

I have the same issue with Edgy on PC using 2.6.17-10-generic kernel and MX5000 / MX1000 Logitech Desktop Bundle.

It worked fine few days after Edgy Beta1 install, but stopped working after updates sometime around 16th of October 2006 (probably after bluez-utils upgrade from 3.1 to 3.7). It worked great last few months with Dapper.

Keyboard and mouse stops responding after a couple of seconds from gdm startup (right after bluetooth service starts). After I unplug/plug in my dongle (Logitech's one) I can use my keyboard but "hcitool dev" shows no devices.

I may attach parts of my /var/log/messages before and after bluetooth starts if you find it useful.

[Dutch (straykat)]

Have just installed 32bit & 64bit Kubuntu 6.10 with my Logitech MX 5000 USB Bluetooth keyboard & mouse using the alternate installer. Keyboard & mouse were detected up until login on first restart.
Kubuntu 5.10, 6.06 & 6.06.1 in both 32 & 64bit flavours had my MX 5000 working out of the box.
I originally reported this at bug #67321.
This bug seems to have been around for a while & now is a show stopper in Edgy. This bug needs to be squashed soon please.

StratKat

[lx73 (alexanderkain)]

Same thing here on a Dell Latitude D810 connecting to a Dell Bluetooth & Mouse Combo, running a stock, updated edgy. After removing and reinserting the dongle into the same USB port the input devices start working.

[Amedee Van Gasse (amedee)]

Same problem as everybody else here: works fine during installer, stops working a few seconds after gdm OR kdm. (Tested with Ubuntu AND Kubuntu).
Edgy 6.10.
AMD Athlon64 3000+
Logitech dongle + mouse + keyboard MX5000

Tested with a 6.06LTS Kubuntu cd: works perfectly. Even the mouse wheel works. I suspect a problem in bluetooth-related software that changed version between 6.06 and 6.10.

[Amedee Van Gasse (amedee)]

I should add that Kubuntu 6.06 does *not* solve the problem, it only provides the keyboard&mouse in "emulation" mode.
I tested this by letting my phone (Nokia 6680) search for the pc: no find.

I downloaded the current feisty dvd: still no fix.
I can consistently simulate the problem like this in all versions of Ubuntu:

1. Mouse&keyboard work after unplug/replug of the dongle. They work in emulation mode (hid mode).
2. Sending a photo from my phone to the pc doesn't work: the phone doesn't find the pc.
3. sudo hid2hci --> changes from hid mode to hci mode
4. mouse and keyboard stop working
5. phone can find the pc (haven't tried to send a file but I assume this is rather trivial and will work if a bluetooth connection can be established)
6. To re-enable keyboard & mouse, repeat from step 1.

This is a major showstopper!
I thought hardware problems always got priority in Ubuntu?
Who do we have to bribe/go to bed with to get this solved?

[Amedee Van Gasse (amedee)]

To help the developers a bit, to clean up some mess, and hopefully to raise the priority of this bug a bit, I dug up all bugs that seemed obvious duplicates of this one. I'm sure there might be more dupes...

Anyway, a bug with 8 obvious duplicates certainly confirms that there is a serious problem, right? I think we all want to use our BT input devices as HCI devices, not as legacy HID devices. Because if BT isn't in HCI mode, there cannot be pairing with any other BT devices like phones,...

[wls (willis-sarka)]

I've since sold the Logitech MX 900 Bluetooth mouse with corresponding
BT hub / base charger.

amedee wrote:
> To help the developers a bit, to clean up some mess, and hopefully to
> raise the priority of this bug a bit, I dug up all bugs that seemed
> obvious duplicates of this one. I'm sure there might be more dupes...
>
> Anyway, a bug with 8 obvious duplicates certainly confirms that there is
> a serious problem, right? I think we all want to use our BT input
> devices as HCI devices, not as legacy HID devices. Because if BT isn't
> in HCI mode, there cannot be pairing with any other BT devices like
> phones,...
>

[Amedee Van Gasse (amedee)]

Op maandag 25-12-2006 om 21:12 uur [tijdzone +0000], schreef wls:
> I've since sold the Logitech MX 900 Bluetooth mouse with corresponding
> BT hub / base charger.

That's not an option for me.
It was a gift from my wife, after my previous keyboard broke.

Anyway, Ubuntu should work with this hardware. Period.

Good luck with your new mouse, whatever it is.

--
Amedee

[wls (willis-sarka)]

Oh, I agree. :)

I just went back to the old wired stuff. I'd like to experiment with a
USB Bluetooth adapter and my Nokia phone here in the near future.

amedee wrote:
> Op maandag 25-12-2006 om 21:12 uur [tijdzone +0000], schreef wls:
>> I've since sold the Logitech MX 900 Bluetooth mouse with corresponding
>> BT hub / base charger.
>
> That's not an option for me.
> It was a gift from my wife, after my previous keyboard broke.
>
> Anyway, Ubuntu should work with this hardware. Period.
>
> Good luck with your new mouse, whatever it is.
>

[alex108 (alex01-malyshev)]

Finally, I made two of my computers work with my Logitech diNovo BT
keyboard/mouse/num pad work in 6.10. I was fallowing several guides like
this https://help.ubuntu.com/community/BluetoothSetup and perhaps some
other posts in Ubuntu forums. It was not easy when I started but it
looks very easy now that it woks. Can't confirm if this is working on
Mac. Note that when you unplug and plug back in your keyboard it will
probably disable bluetooth and it will say something like device not
recognized when you try to use bluez-utils. So have a spare keyboard and
mouse plugged in to edit your configs and search/connect your BT
keyboard/mouse.

Though it would be nice if BT was easier or more obvious way of
configuring, I think some of the problems here could be fixed by
following the Bluethooth Setup guide. Don't sell your bluetooth!

[kisch (kisch)]

@Lothar (2006-10-14/15):
thanks for you strace log.
It proves that there are actually two diffent bugs, both related to hidd and/or the hidp kernel module.
A/ you observe a "connect timeout" from "hidd --connect ..."
B/ Zhares sees a "HID create error 14" from "hidd --connect ..."

Bug A seems to be related to certain dongle/device combinations, as Etienne states. It's not related to the platform - you see it on an Athlon, I see it on a PowerPC G4.

Bug B ONLY happens on 64bit PowerPC Apple G5 hardware. The HID device doesn't matter.

These bugs are really different. Your strace (bug A) shows hidd failing BEFORE it reaches the point where hidd executes ioctl(HIDPCONNADD), which would trigger bug B.

Bug B seems to be caused by a 64bit problem, just as Zhares assumed.

The ioctl which fails is defined in /usr/include/bluetooth/hidp.h:
#define HIDPCONNADD _IOW('H', 200, int)

... but the caller is supposed to pass not an int, but instead a pointer to a struct hidp_connadd_req, which is declared as:
{
        int ctrl_sock; /* Connected control socket */
        int intr_sock; /* Connected interrupt socket */
        uint16_t parser; /* Parser version */
        uint16_t rd_size; /* Report descriptor size */
        uint8_t *rd_data; /* Report descriptor data */
        uint8_t country;
... [remaining parts are arch-independently sized]
};

In the application the ioctl has to be called in this way (taken from bluez-utils/hidd/main.c):
        struct hidp_connadd_req req;
...
        err = ioctl(ctl, HIDPCONNADD, &req);

On a 64bit PowerPC system like a PowerMac G5, ubuntu uses a 64bit kernel and a mixed 32/64bit userspace. Specifically, hidd is compiled as a 32bit binary.
In a 64bit compile (kernel), the pointer "uint8_t *rd_data" is 8 bytes long, and it will also be aligned to a multiple of 8. Therefore the size of struct hidp_connadd_req will be larger in a 64bit compile than in a 32bit compile (hidd). I've attached a small test program to illustrate the difference -- try it:
gcc -m32 hidp_connadd_bug.c && ./a.out
gcc -m64 hidp_connadd_bug.c && ./a.out
The sizes of the struct are 160 vs. 168 bytes.

The hidp kernel module has this code in Linux/net/bluetooth/hidp/sock.c:

static int hidp_sock_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
{
        void __user *argp = (void __user *) arg;
        struct hidp_connadd_req ca;
...
        switch (cmd) {
        case HIDPCONNADD:
...
                if (copy_from_user(&ca, argp, sizeof(ca)))
                        return -EFAULT;
...

EFAULT is error 14, "bad address" (as observed).

The hidp kernel module tries to copy a structure hidp_conadd_req of a larger size (due to 64bit pointer size) than the user provided (due to 32bit pointer size). I suspect that this fails, so that the ioctl returns the error code, causing hidd to terminate.
Even if the copy succeeded, the struct would then be interpreted incorrectly by the hidp module (resulting in different error codes from the ioctl).

So I think that the root cause seems to be the definition of the bluetooth hidp ioctl API HIDPCONADD.
I'll try to prepare a patch for the kernel to verify ...

[kisch (kisch)]

I confirm my suspicion about the bug subtype B (occuring on ppc64 machines only).

The problem can be worked around by building a 64bit version of bluez-utils and bluez-libs.

Currently, I've got 32bit and 64bit variants installed in parallel.
hidd --connect xx:yy:zz:....
ALWAYS prints "HID create error 14" in the 32bit variant,
but ALWAYS succeeds in the 64bit variant.

While this is a possible workaround for technical users (you also need a 64bit glibc, I'm afraid), it's no solution for Ubuntu, because it uses 32bit applications for both 32bit and 64bit PowerPC platforms. The root cause is the kernel API, but it's probably easier in the moment to patch the bluez-utils instead. I'll try to provide a platform-specific patch.

After researching all that 32/64bit stuff it's also clear to me why the bug doesn't bite the amd64 platform. Amd64 uses 64bit userspace on a 64bit kernel, so there's no discrepancy in the ioctl structure.

A happy new year to everyone,
- Bernd Kischnick

[kisch (kisch)]

Here comes the patch.

It applies to bluez-utils 2.25 and 3.7 (and all versions inbetween, presumably).

The patch can safely be applied to the source tree for all platforms.
It checks at compile time that it's being build for PowerPC 32bit, otherwise it doesn't change the implementation.
Then it checks at runtime that it's running on a ppc64 system - in this case the structs passed through 64bit-broken bluetooth HIDP ioctls are re-packed into the layout that's expected by the 64bit kernel.

I had to wrap ioctl HIDPCONNADD as well as HIDPGETCONNLIST, which also contains a pointer in its request struct.

The patch is tested on PowerPC 32bit and 64bit (Apple Powerbook G4 and PowerMac G5).
As a result hidd is working as expected now, my bluetooth keyboard and mouse are detected, connected, and appear as input devices.

It's of course possible that there are still other 64bit issues in the bluetooth API and -utils.

- Bernd

[Marcel Holtmann (holtmann)]

The kernel should have been fixed to handle that properly. It is part of the 2.6.19-rc3 kernel and later. It fixes similar issues in BNEP and CMTP support.

[kisch (kisch)]

Thanks, Marcel!
You're right, that's a fix for this problem.

For everyone to verify, here's the commit entry:
http://www.kernel.org/git/?p=linux/kernel/git/stable/linux-2.6.19.y.git;a=commit;h=e9c5702e3c5558dade169949abd730173e87ef9c

Marcel has fixed the issue entirely in the kernel, leaving the ioctl API as-is.
This means that the bug will go away as soon as Ubuntu uses kernel 2.6.19, without changing the bluez-utils package at all.

So this bug should probably be associated with the kernel package instead (at least as far as it concerns problems on ppc64 platforms).

[Amedee Van Gasse (amedee)]

Does this kernel bugfix also apply for x86 and/or amd64? Or is it only for powerpc?

[kisch (kisch)]

The kernel patch applies to all platforms (it changes arch-independent sources).

But the bug doesn't occur on the other platforms:
- x86 is 32bit userspace on a 32bit kernel
- amd64 is 64bit userspace on a 64bit kernel.

The bug only appears when a 32bit userspace is used on a 64bit kernel - this is the powerpc64 platform.

So the kernel patch applies to all platforms, but it only fixes breakage specific for powerpc64.

[LGM6230 (lgm6230)]

Cheers for that will give it a go.

alex108 <email address hidden> wrote: Finally, I made two of my computers work with my Logitech diNovo BT
keyboard/mouse/num pad work in 6.10. I was fallowing several guides like
this https://help.ubuntu.com/community/BluetoothSetup and perhaps some
other posts in Ubuntu forums. It was not easy when I started but it
looks very easy now that it woks. Can't confirm if this is working on
Mac. Note that when you unplug and plug back in your keyboard it will
probably disable bluetooth and it will say something like device not
recognized when you try to use bluez-utils. So have a spare keyboard and
mouse plugged in to edit your configs and search/connect your BT
keyboard/mouse.

Though it would be nice if BT was easier or more obvious way of
configuring, I think some of the problems here could be fixed by
following the Bluethooth Setup guide. Don't sell your bluetooth!

--
Bluetooth Mouse and Keyboard Broken in Dapper/Edgy/Feisty
https://launchpad.net/bugs/32415

 Send instant messages to your online friends http://au.messenger.yahoo.com

[Mark (azovap002)]

Bug remains unsolved in latest Feisty PPC build. We tested

iMac G5 rev B
Apple bluetooth keyboard A1016
Apple bluetooth mouse A1015
Feisty daily build 29 Jan 2007
Alt-powerpc disc : "install-powerpc64"
(kernel 2.6.20-5-powerpc64-smp)

So kernel >= 2.6.19 evidently does not solve the problem(s). The GDM lockup remains. To echo/confirm the original bug, we are saying that for this test, "once GDM comes up, my bluetooth devices are no longer functional."

Workarounds offered above gave no usability. Typing keys and moving the mouse fail. There is some initial activity with keys but then lockup. You can type about 3 letters into GDM. The mouse cursor never moves.

Installed bluetooth config files are (wrongly?) defaulted to HIDD_ENABLED=0 which was changed, but to no avail.

Removing bluez* makes the keyboard work and GDM login possible. The Apple bluetooth mouse never works in GNOME at all with or without bluez*. It works fine in Mac OS X 10.4.8. The mouse hardware is ok.

Booting "Linux single" gives a usable prompt and the keyboard works there, as it does with the alt installer.

[pdecat (pdecat)]

Thanks to LGM6230 advices,

I worked around the problem and made my Logitech DiNovo Media Desktop 1.0 (bluetooth keyboard+mediapad+mouse) work with Ubuntu Dapper 6.06.1.

The only thing I needed to change from a stock install is (requires remote shell access or a spare wired keyboard) :

$ diff /etc/default/bluez-utils /etc/default/bluez-utils.orig
13,14c13,14
< HIDD_ENABLED=1
< HIDD_OPTIONS="--connect 00:07:61:07:XX:XX --connect 00:07:61:06:YY:YY --connect 00:07:61:06:ZZ:ZZ --master --server"
---
> HIDD_ENABLED=0
> HIDD_OPTIONS="--master --server"

I now have full Bluetooth functionality and can finally access to my mobile phone via bluetooth !

Best regards,
Patrick.

[pdecat (pdecat)]

> I now have full Bluetooth functionality and can finally access to my mobile phone
> via bluetooth !

PS: That's because I previously had to disable Bluetooth altogether to use my Logitech set.

Regards,
Patrick.

[Marcel Holtmann (holtmann)]

I think all these problems on Apple based system are caused by a race condition in the startup and init routines for the Bluetooth adapter. I am pretty sure that I finally fixed this. At least my Quad G5 running Edgy now has no problems anymore.

The main problem was that the command for setting the scan mode failed. And this means that the adapter was available, but neither page scan nor page inquiry were set. This results in a non-connectable and non-discoverable Bluetooth adapter. However HID mice and keyboards wanna reconnect to the adapter and if it is in non-connectable mode (page scan off) then they simply can't.

This race condition was present on all architectures, but for some strange reason only the Apple machines with the HID proxy adapters suffered from it.

Anyhow this has been fixed with the bluez-utils-3.9 release and for Feisty it might be a good idea to simply sync with the new upstream source.

[Amedee Van Gasse (amedee)]

On Tue, January 30, 2007 11:20, Marcel Holtmann said:
> I think all these problems on Apple based system are caused by a race
> condition in the startup and init routines for the Bluetooth adapter. I
> am pretty sure that I finally fixed this. At least my Quad G5 running
> Edgy now has no problems anymore.
>
> The main problem was that the command for setting the scan mode failed.
> And this means that the adapter was available, but neither page scan nor
> page inquiry were set. This results in a non-connectable and non-
> discoverable Bluetooth adapter. However HID mice and keyboards wanna
> reconnect to the adapter and if it is in non-connectable mode (page scan
> off) then they simply can't.
>
> This race condition was present on all architectures, but for some
> strange reason only the Apple machines with the HID proxy adapters
> suffered from it.
>
> Anyhow this has been fixed with the bluez-utils-3.9 release and for
> Feisty it might be a good idea to simply sync with the new upstream
> source.

I disagree, the problem is not only on Apples! See my previous posts.
It's also on x86. Or else it's a related but different problem, but "for
the rest of us", we can't tell the difference.

--
Amedee

[AJ Slater (aj-slater)]

> I disagree, the problem is not only on Apples!

I concur. I can reproduce the problem on x86 and x86_64.

AJ

[Mark (azovap002)]

"fixed with...bluez-utils-3.9...for Feisty...sync with...upstream source."

Sure, and put more frequent updates in the repos. The bug title includes Dapper/Edgy/Feisty.

From a Dapper x86-32 machine we checked bluez status. Bluez sits at 2.24 versus the current 3.9 source. That's quite a lag. This is "long-term support" for Dapper. So we feel it could be better.

Thanks for your hard work, we're not trying to be over-critical here. It just seems these bluetooth bugs are hurting too many users for too long if simple package maintenance can solve them.

[Andrew Frank (frank-geoinfo)]

i run ubuntu 6.10 at the current update level (30.1.2007) on a fujitsu tablett (386)

i followed the instructions in the wiki to install bluetooth such that it works from login - and it worked for a few times, but then stopped (nothing done with bluetooth in the meantime). i can install bluetooth devices with hidd --search, but it does not keep it after shut-down.

i have HIDD_ENABLED=1
and added the --connect xxx
and have --master -- server to
HIDD_OPTIONS
(does the order matter?)
and added hipp to /etc/modules

so all should be fine?
does bluetooth write a log when it starts so one can see what is wrong? where is it?

how would i update to the new version of bluez-utils 3.9 (but keep ubuntu 6.10 for the rest?)

thanks for the help! (and teh clear instructions in the wiki!)
andrew

[Amedee Van Gasse (amedee)]

This bug is marked as "needs info".
Please, what info do you need? I will give all information you need, and even more.
Is this bug really going to celebrate its first anniversary in a couple of days?

[Amedee Van Gasse (amedee)]

Today Ubuntu Feisty pulled in bluez-utils version 3.9.
The situation remains identical: I either have connection with keyboard+mouse, or with the phone, but never both.

[Marcel Holtmann (holtmann)]

Don't put --connect in HIDD_OPTIONS. This will fully mess with your system and whoever proposed this strange settings is wrong. In general the option --server is fully enough and HIDD_ENABLE must be on.

The --search or --connect needs only to be issued once. After that the mouse or keyboard will reconnect once it is used. Use hcidump -X -V to check if there are any problems with the re-connect.

[Amedee Van Gasse (amedee)]

Thank you for your reply. Attached you will find my /etc/default/bluetooth
The most relevant section is:

HIDD_ENABLED=1
HIDD_OPTIONS="--master --server"

I will see what happens when I remove --master.

One question though. You write that --search or --connect only needs to be issued once. When and how? Not by using the keyboard, because that won't work... Same goes for hcidump -X -V: I cannot type that when there is no connection...

Anyway, this is my "normal" output of hcidump -X -V

amedee@saruman:~$ hcidump -X -V
HCI sniffer - Bluetooth packet analyzer ver 1.33
Can't open device: No such device

[Amedee Van Gasse (amedee)]

One reboot later...
The problem remains identical.
I can use the bluetooth keyboard&mouse in legacy HID mode, OR I can use the bluetooth dongle in HCI mode to connect to my phone, but not both at the same time.

[Amedee Van Gasse (amedee)]

Attached is the output of

   sudo watch hcidump -X -V >> ~/hcidump.log

If anyone knows a better way of logging, please tell me.

[Amedee Van Gasse (amedee)]

Perhaps it's not important, but I think I should mention it.
On the forums someone mentioned it's not possible to get a bluetooth mouse working in Ubuntu when it's working in Windows. You have to disable it in Windows too.
http://www.ubuntuforums.org/showthread.php?t=87919&page=9
I have a multiboot with Windows XP, for firmware upgrades and other things that can't be done in any other way (yet). The bluetooth set works in Windows: keyboard&mouse + phone at the same time.

[pdecat (pdecat)]

> Don't put --connect in HIDD_OPTIONS. This will fully mess with your system and
> whoever proposed this strange settings is wrong.

Marcel,

putting the following in my /etc/default/bluez-utils is the only thing that ever made my keyboard, mediapad and mouse (all 3 are bluetooth devices) work with Ubuntu Dapper 6.06.1 :

HIDD_OPTIONS="--connect 00:07:61:07:XX:XX --connect 00:07:61:06:YY:YY --connect 00:07:61:06:ZZ:ZZ --master --server"

My system has never been more stable and I never lost keyboard, mediapad and mouse functionality since then.

Why would that mess my system since all 3 devices are always present?

Best regards,
Patrick.

[Piotr Szczepanik (piter75)]

I must admit that setting in /etc/default/bluetooth:

HIDD_OPTIONS="--master --connect 00:07:61:34:xx:xx--connect 00:07:61:37:xx:xx --server"

makes my MX5000 set working in HCI mode on Edgy x86.
I loose it only after I boot to Windows and back to Edgy but then replugging the dongle makes it working again.

Best regards,
Piotr Szczepanik

[Amedee Van Gasse (amedee)]

On di, 2007-02-20 at 15:34 +0000, Piotr Szczepanik wrote:
> I must admit that setting in /etc/default/bluetooth:
>
> HIDD_OPTIONS="--master --connect 00:07:61:34:xx:xx--connect
> 00:07:61:37:xx:xx --server"
>
> makes my MX5000 set working in HCI mode on Edgy x86.
> I loose it only after I boot to Windows and back to Edgy but then replugging the dongle makes it working again.

OK. And when you replug the dongle, what mode are you in? HID or HCI?
This is very important! Really, this is what this bug report is all
about.

Also, do you have any other bluetooth device like a mobile phone?
Can you connect to the phone? Can you send files to/from the phone?

What is your output of lsusb?

When your keyboard and mouse work, what happens when you type the
following?
sudo hid2hci

--
Amedee Van Gasse
<email address hidden>

[SickFinga (sickfinga)]

I can also comfirm this bug on Ubuntu 6.10 with Logitech Elite keyboard and MX900 combo (both USB)

Here is how it works for me.

When I boot into Ubuntu with bluetooth dongle pluged in, mouse and keyboard stops working.

When I unplug USB from the bluetooth dongle Ubuntu loads fine, keyboard works, mouse doesnt (mouse uses USB, keyboard PS2).
Now, if i plug in USB dongle back into to my computer, my mouse starts to work.

I set "HDD_ENABLE = 1", didnt help.

Amedee, I did what you asked Piotr to do.

lsbus

Bus 002 Device 003: ID 046d:c7xx Logitec, Inc
Bus 002 Device 002: ID 0451:20xx Texas Instruments, Inc. TUSB2036 Hub

When I type hid2hci, I get this
Switching device 046d:c7xx to HCI mode failed (No such file or directory)

If I set BLUETOOTH_ENABLED=0
Mouse and keyboard work fine, but I guess bluetooth doesnt work.

[SickFinga (sickfinga)]

OK, new findings.

I plugged in regular keyboard/mouse with bluetooth ones still plug in. As usual bluetooth ones didnt work.
I tried using my cellphone and it found my bluetooth dongle. I also tried hcitool scan and it found my mouse/keyboard.

If I do unplug/plug back in routine, and I try to use "hcitool scan", I get this
Device not available: No such device.
And my cellphone doesn't find the bluetooth dongle.

[Amedee Van Gasse (amedee)]

On do, 2007-03-01 at 19:21 +0000, SickFinga wrote:
> OK, new findings.
>
> I plugged in regular keyboard/mouse with bluetooth ones still plug in. As usual bluetooth ones didnt work.
> I tried using my cellphone and it found my bluetooth dongle. I also tried hcitool scan and it found my mouse/keyboard.
>
> If I do unplug/plug back in routine, and I try to use "hcitool scan", I get this
> Device not available: No such device.
> And my cellphone doesn't find the bluetooth dongle.
>

That is exactly what I expected would happen, and this is an excellent
description of the problem.

Unfortunately it's quite clumsy to work with two keyboards and two mice.

Some questions.

You wrote that "hcitool scan" found your mouse/keyboard. But they did'n
work, right? (until you unplugged the bluetooth dongle).

Second question: after unplugging, when bluetooth doesn't work, do
hid2hci. What happens? If your problem is the same as mine, your
cellfone should find the dongle again, and your mouse/keyboard should
stop working.

--
Amedee Van Gasse
<email address hidden>

[Piotr Szczepanik (piter75)]

On 3/1/07, amedee <email address hidden> wrote:
>
> That is exactly what I expected would happen, and this is an excellent
> description of the problem.

Hey Amedee,
sorry for not coming back for so long.

I did:
piter@piter-home:~$ sudo hid2hci and the result is:
Switching device 046d:c70a to HCI mode was successful
Switching device 046d:c70e to HCI mode was successful

piter@piter-home:~$ lsusb
Bus 003 Device 003: ID 04da:0f43 Panasonic (Matsushita)
Bus 003 Device 001: ID 0000:0000
Bus 002 Device 005: ID 046d:c709 Logitech, Inc.
Bus 002 Device 004: ID 046d:c70a Logitech, Inc.
Bus 002 Device 002: ID 046d:0b02 Logitech, Inc.
Bus 002 Device 003: ID 046d:c70e Logitech, Inc.
Bus 002 Device 001: ID 0000:0000
Bus 001 Device 003: ID 06bd:0002 AGFA-Gevaert NV SnapScan 1236U
Bus 001 Device 001: ID 0000:0000

After replugging the dongle the keyboard and mouse goes back to HID mode.
Then I type sudo hid2hci and it tells me that switching was successful.

hcitool scan finds my mobile phone but it does not find my keyboard and
mouse (although they are working well)

--
Best regards,
Piotr Szczepanik

[SickFinga (sickfinga)]

Ok, tried it again.

When I type hcitool scan, it find my mouse and keyboard (Piotr, you have to put your keyboard/mouse into discoverable mode by pressing the button on they keyboard/mouse). When I try to connect my mouse using "hidd --connect MACADDRESS" I get this error.
Can't get device information: File descriptor in bad state".

hciscan tool only works if I keep my bluetooth dongle plug in during booting process. If i unplug it before booting and plug it in after its done, hcitool scan doesn't work.

I got the same error all the time when I type hid2hci, "Switching device MACADDRESS, to HCI mode failed (No such file or directory).

My bluetooth phone only finds USB dongle when I keep it pluged in while booting, if I unplug it and then plug it back in, my cellphone does not find it.

[Kyrlian (kyrlian)]

Hi all.
Had this problem since my adoption of ubuntu 6 month ago. Get the dinovo set to work by unsinstalling all bluez packages. More recently I came back to this problem, and solved it using instruction (in french) at http://forum.ubuntu-fr.org/viewtopic.php?id=70506

in a nutshell - on edgy, with all bluez packages installed:

1/ check /etc/default/bluetooth
HIDD_ENABLED=1
and add "-i hci0" in HIDD_OPTIONS :
HIDD_OPTIONS="-i hci0 --master --server"

2/
add in /etc/init.d/bluetooth, in "start)" and "restart|force-reload)" sections :
 enable_hci_input || true

3/ I had to force the very first pairing with this script (launch with sudo)
hid2hci --tohci
sleep 1
/etc/init.d/bluetooth restart
sleep 1
hidd -i hci0 --connect xx:xx:... mouse adress
hidd -i hci0 --connect xx:xx... kb adress

now works fine, no need to unplug any more. And I can have my wiimote connected via bluetooth :)
one glitch tough : it lags a bit, seems the kb signal cuts the mouse signal and vice-versa, a pain when playing fps.

[Amedee Van Gasse (amedee)]

On zo, 2007-03-04 at 16:36 +0000, Kyrlian wrote:
> Hi all.
> Had this problem since my adoption of ubuntu 6 month ago. Get the
> dinovo set to work by unsinstalling all bluez packages. More recently
> I came back to this problem, and solved it using instruction (in
> french) at http://forum.ubuntu-fr.org/viewtopic.php?id=70506
>
> in a nutshell - on edgy, with all bluez packages installed:
>
> 1/ check /etc/default/bluetooth
> HIDD_ENABLED=1
> and add "-i hci0" in HIDD_OPTIONS :
> HIDD_OPTIONS="-i hci0 --master --server"

OK, I did that.

> 2/
> add in /etc/init.d/bluetooth, in "start)"

It's already there.

> and "restart|force-reload)" sections :
> enable_hci_input || true

OK, did that.

> 3/ I had to force the very first pairing with this script (launch with sudo)
> hid2hci --tohci
> sleep 1
> /etc/init.d/bluetooth restart
> sleep 1
> hidd -i hci0 --connect xx:xx:... mouse adress
> hidd -i hci0 --connect xx:xx... kb adress

This is my script:

hid2hci --tohci
sleep 1
/etc/init.d/bluetooth restart
sleep 1
hidd -i hci0 --connect 00:07:61:48:71:DE
hidd -i hci0 --connect 00:07:61:3F:DA:41

This is the output:

amedee@saruman:~$ sudo ./enable_bluetooth.sh
Switching device 046d:c70a to HCI mode was successful
Switching device 046d:c70e to HCI mode was successful
 * Restarting Bluetooth services
[ OK ]
Can't get device information: Invalid exchange

> now works fine, no need to unplug any more. And I can have my wiimote connected via bluetooth :)

I was writing a whole story on the website, but Firefox crashed so I'll
try again by email. Shorter this time.

*Something* has changed.
Mouse works, sending files to the phone works, keyboard doesn't work.

I attached some log files (in .tar.gz). Some of them may even be
relevant.

In some of the log files I see something about a bug. I'm not a kernel
hacker, but if I see something about a bug in a kernel log, then I'm
worried...

--
Amedee Van Gasse
<email address hidden>

[Amedee Van Gasse (amedee)]

No attachments, I see.
Second try...

I should also mention that the keyboard is found, but that kbluetoothd is unable to pair with the keyboard.

[aliubi (ubuntu-aliubi)]

I just installed Edgy on a couple of machines last week. (Let me preface this by saying I'm already an Ubuntu convert...however...) I have a Logitech DiNovo keyboard and mouse, with dongle.

The behaviour I see seems slightly different than what is described herein. I use sudo hidd --connect, and it works fine both on a Gateway Solo 1400 (using the Logitech dongle) and on an HP/Compaq nc6000 using the internal Bluetooth. Unfortunately, if idle for too long (I'm guessing the KB/mouse enter snooze mode) both machines will lose their connection to the KB/mouse until I reset both and sudo hidd --connect again.

I may have missed it, but I don't see this exact behaviour described above. I have no problem connecting, except that it's inconvenient to have to do it on such a regular basis (I've already written an alias because I got tired of finding it in the history).

Any ideas here? Anyone else seeing this issue?

Aliubi

[T. Koehler (leavenworth)]

Confirmed on feisty (up-to-date as of posting) on i386:

I can only use my bluetooth keyboard and mouse (Microsoft Office Desktop Elite) in HIDP-mode.

Patching the kernel (2.6.20-10-lowlatency) with the latest patch from www.bluez.org (-mh1) resolves the issue.

I had the same problem in several other distros (most notably gentoo), since about kernel version 2.6.12 (I had no bluetooth set before that).

I am curious why the bluez.org site states that its patches are included in the more recent kernels if I had to apply the patch manually to all the kernels since then to resolve the issue I have.

Proposed solution: Include the -mh-patch in the standard ubuntu kernel by default, especially if this patch fixes the multitude of issues other users seem to be having (tests, anyone?).

As an afternote: The bluetooth set worked fine in Edgy until about 12/06.

[T. Koehler (leavenworth)]

What other information do you seek? Please specify.

I propose to change the urgency to high, since it disables people with bluetooth-based input devices from using their computer at all.

[T. Koehler (leavenworth)]

Since this seems in part to be a kernel issue, the bug should also be filed against the kernel package, correct?

[Amedee Van Gasse (amedee)]

Op donderdag 15-03-2007 om 18:09 uur [tijdzone +0000], schreef T.
Koehler:
> Since this seems in part to be a kernel issue, the bug should also be
> filed against the kernel package, correct?

I agree.

I have also sent an email to the linux kernel mailing list (you can't go
any more upstream than that), after I saw a description of a very
similar (or the same) bug that one of the kernel developers was
experiencing.

--
Amedee Van Gasse <email address hidden>

[Amedee Van Gasse (amedee)]

Op donderdag 15-03-2007 om 18:03 uur [tijdzone +0000], schreef T.
Koehler:
> What other information do you seek? Please specify.
>
> I propose to change the urgency to high, since it disables people with
> bluetooth-based input devices from using their computer at all.
>
I agree with the urgency - although to be honest: in most cases it is
possible to use EITHER input functionality OR bluetooth functionality
(pairing with phones, headsets,...) at the same time.

The problem is not so bad for people who only have a bluetooth mouse and
still have a wired keyboard. It looks like they can work around the bug.
However the problem is dramatic for people who have both mouse and
keyboard BT-based.

--
Amedee Van Gasse <email address hidden>

[Amedee Van Gasse (amedee)]

I now have a working keyboard+mouse+phone!

Wat I did:
* update to the latest bluetooth version in feisty
* in this version, the mouse and the phone work, but not the keyboard
* I attached a spare ps2 keyboard
* I pressed the connection button on the keyboard and the dongle, while typing (on the other keyboard): sudo hidd --connect 00:07:61:3F:DA:41
* I had to repeat this several times. First I got "Can't create HID control channel: Connection refused" but after repeating this a few times, I got no errors and behold! A functional keyboard. I had to be really, REALLY fast...

Now I'm afraid to reboot because I fear I will have to repeat this procedure with the spare keyboard every time...

[Amedee Van Gasse (amedee)]

It's (almost) solved!

I have added the following to /etc/default/bluetooth:

HIDD_OPTIONS="-i hci0 --connect 00:07:61:3F:DA:41 --connect 00:07:61:48:71:DE --master --server"

00:07:61 is Logitech, 3F:DA:41 is the keyboard; 48:71:DE is the mouse.

I rebooted. It's my impression that the keyboard doesn't work in grub. During the boot process, the LCD on the keyboard asks for a password. When kdm comes up, both mouse and keyboard are working. I can also send pictures to/from the phone.

This leaves only one minor issue to investigate/solve: why doesn't the keyboard work in grub?

I don't consider entering a password on the keyboard a problem. I kinda like the idea that you have to enter a password on the physical device you use to enter a password on the computer. That's extra security ;-)

[T. Koehler (leavenworth)]

The latest kernel (2.6.20-13) fixes the problem for me.

[Jeff Parsons (justjeff)]

I just updated my Feisty install's kernel from 2.6.20-13-generic to 2.6.20-14-generic, with all other packages up-to-date, too. The problem wasn't fixed by -13, nor by -14.

I have a Logitech MX5000 keyboard / MX1000 mouse combo. The unplug (wait...) plug back in workaround works for me. Output from a few of the diagnostics mentioned above:

jeff@meatpopsicle:~$ uname -a
Linux meatpopsicle 2.6.20-14-generic #2 SMP Mon Apr 2 20:37:49 UTC 2007 i686 GNU/Linux
jeff@meatpopsicle:~$ hcitool dev
Devices:
jeff@meatpopsicle:~$ lsmod |grep hidp
jeff@meatpopsicle:~$ grep hidp /etc/modules
jeff@meatpopsicle:~$ lsusb
Bus 002 Device 008: ID 046d:c70a Logitech, Inc.
Bus 002 Device 007: ID 046d:c70e Logitech, Inc.
Bus 002 Device 006: ID 046d:0b02 Logitech, Inc.
Bus 002 Device 001: ID 0000:0000
Bus 001 Device 001: ID 0000:0000
jeff@meatpopsicle:~$ dmesg | grep -i blue
[ 65.502647] Bluetooth: Core ver 2.11
[ 65.502689] Bluetooth: HCI device and connection manager initialized
[ 65.502692] Bluetooth: HCI socket layer initialized
[ 65.533724] Bluetooth: L2CAP ver 2.8
[ 65.533725] Bluetooth: L2CAP socket layer initialized
[ 65.582419] Bluetooth: RFCOMM socket layer initialized
[ 65.582428] Bluetooth: RFCOMM TTY layer initialized
[ 65.582430] Bluetooth: RFCOMM ver 1.8
[ 66.095853] Bluetooth: HCI USB driver ver 2.9

[Eugene Lazutkin (eugene-lazutkin)]

Same problem here. I have the latest feisty, and I have to "sudo hidd --search" after rebooting, and periodically after that to re-attach the mouse. I cannot unplug/re-plug the dongle because my notebook has the internal Bluetooth module.

This bug thread goes for so long! How come it was never fixed? A noteboook + a bluetooth mouse is a fairy common combination. Heck, I even use bluetooth mouse on all my computers --- I don't like hanging wires.

[Marcel Holtmann (holtmann)]

Maybe this is Logitech specific. However I only have the old diNovo desktop and that one works perfectly fine for me. Even the re-connect has no problems.

[Amedee Van Gasse (amedee)]

The status of this more than a year old bug is still needs info.
Who do we have to bribe / give beer to / let our sister sleep with / sacrifice our firstborn for to get some attention to this bug?
A few posts and weeks back I wrote a functional workaround, but there still is no confirmation that my workaround a) works for everybody else and if a) == true then b) the workaround is included in the standard software as a permanent solution.
Really, beer++ for the person who can do something constructive. Because now it looks like someone isn't respecting the "Step down considerately" part of the Ubuntu CoC: nobody is picking up where someone else left off...

[Jan Claeys (janc)]

I guess there is more than enough info for the developers to have a look at this now...

[Jan Claeys (janc)]

This looks like the same or a similar bug: http://groups.google.com/group/ubuntu-bluetooth/browse_frm/thread/eda5d4dfefe905db ?

[SLerman (smlerman)]

I have the exact same problem as aliubi mentioned above (https://bugs.launchpad.net/ubuntu/+source/bluez-utils/+bug/32415/comments/97). My Logitech MX5000 keyboard and mouse work fine for a while, but the connection gets dropped if either device is idle for too long. I'm not sure what the exact idle time is, but both devices have been disconnected when I come home at the end of the day or wake up in the morning, so it's less than 8 hours. I'd rather not have my nice (and expensive) keyboard and mouse be completely useless.

[Mark (azovap002)]

Bluetooth failure still present.
iMac G5 rev B Apple bluetooth k/b & bluetooth mouse.
Feisty final powerpc alternate install CD.
Installation goes through fine (install-powerpc64).
Reboot works fine up to GDM login screen.
This becomes a lockout screen; no mouse, no k/b functionality.

[Eugene Lazutkin (eugene-lazutkin)]

BTW, it is not Logitech specific --- I have this problem with Miscrosoft Bluetooth mouse.

[Amedee Van Gasse (amedee)]

Op zaterdag 05-05-2007 om 05:23 uur [tijdzone +0000], schreef Eugene
Lazutkin:
> BTW, it is not Logitech specific --- I have this problem with Miscrosoft
> Bluetooth mouse.
>
Does it have a broadcomm-based usb receiver?

--
Amedee Van Gasse <email address hidden>

[Eugene Lazutkin (eugene-lazutkin)]

>> BTW, it is not Logitech specific --- I have this problem with Miscrosoft
>> Bluetooth mouse.
>>
>Does it have a broadcomm-based usb receiver?

It uses the built-in Bluetooth module in my notebook.

[Amedee Van Gasse (amedee)]

On Sat, May 5, 2007 09:06, Eugene Lazutkin wrote:
>>> BTW, it is not Logitech specific --- I have this problem with
>>> Miscrosoft
>>> Bluetooth mouse.
>>>
>>Does it have a broadcomm-based usb receiver?
>
> It uses the built-in Bluetooth module in my notebook.

Yes, and which chipset does your built-in bluetooth module use?
Broadcomm?

--
Amedee

[Zoltan (zoltans)]

Hi,
Just want to add my name to the list of peeps affected by this bug :-)
I have the Logitech MX Bluetooth Kyb/Mouse, which was fine until 5.10, and after having shelved this expensive monster for a year, I see it's still not fine in 7.04 - but at least re-plugging it seems to re-awaken it.
I'll keep an eye on this thread, but email me if there is anything I can add or help with.
Rgds,
Zoltan

[Amedee Van Gasse (amedee)]

On Tue, May 15, 2007 07:58, Zoltan wrote:
> Hi,
> Just want to add my name to the list of peeps affected by this bug :-)
> I have the Logitech MX Bluetooth Kyb/Mouse, which was fine until 5.10, and
> after having shelved this expensive monster for a year, I see it's still
> not fine in 7.04 - but at least re-plugging it seems to re-awaken it.
> I'll keep an eye on this thread, but email me if there is anything I can
> add or help with.

Do you have any other bluetooth devices? Mobile phone, headset,...
Do they work when you re-plug? Before and/or after?

--
Amedee Van Gasse
<email address hidden>

[Zoltan (zoltans)]

amedee wrote:
> On Tue, May 15, 2007 07:58, Zoltan wrote:
>
>> Hi,
>> Just want to add my name to the list of peeps affected by this bug :-)
>> I have the Logitech MX Bluetooth Kyb/Mouse, which was fine until 5.10, and
>> after having shelved this expensive monster for a year, I see it's still
>> not fine in 7.04 - but at least re-plugging it seems to re-awaken it.
>> I'll keep an eye on this thread, but email me if there is anything I can
>> add or help with.
>>
>
> Do you have any other bluetooth devices? Mobile phone, headset,...
> Do they work when you re-plug? Before and/or after?
>
The unit only connects to it's own kyb & mouse - my Nokia E70 cannot see it.
Z

--

===========================================
Geograph (Pty) Ltd.
P.O. Box 31255, Tokai 7966, South Africa.

B3-106 Tokai Village, cnr Vans/Tokai Rds,
Tokai 7945, Western Cape, South Africa.

Tel: [phone] Mobile: [phone]
Fax: [phone]
===========================================

[Amedee Van Gasse (amedee)]

On Tue, May 15, 2007 15:40, Zoltan wrote:
> amedee wrote:
>> On Tue, May 15, 2007 07:58, Zoltan wrote:
>>
>>> Hi,
>>> Just want to add my name to the list of peeps affected by this bug :-)
>>> I have the Logitech MX Bluetooth Kyb/Mouse, which was fine until 5.10,
>>> and
>>> after having shelved this expensive monster for a year, I see it's
>>> still
>>> not fine in 7.04 - but at least re-plugging it seems to re-awaken it.
>>> I'll keep an eye on this thread, but email me if there is anything I
>>> can
>>> add or help with.
>>>
>>
>> Do you have any other bluetooth devices? Mobile phone, headset,...
>> Do they work when you re-plug? Before and/or after?
>>
> The unit only connects to it's own kyb & mouse - my Nokia E70 cannot see
> it.

That confirms you have the same bug.
Your E70 *should* be able to see the pc and vice versa.
Because it can't, that means your bluetooth unit is now in HID legacy
(USB) mode. When you type sudo hid2hci at console, you will loose keyb &
rodent but your E70 will be able to see the pc. (if hid2hci dosn't give
any other error)

--
Amedee

[Zoltan (zoltans)]

amedee wrote:
> On Tue, May 15, 2007 15:40, Zoltan wrote:
>
>> amedee wrote:
>>
>>> On Tue, May 15, 2007 07:58, Zoltan wrote:
>>>
>>>
>>>> Hi,
>>>> Just want to add my name to the list of peeps affected by this bug :-)
>>>> I have the Logitech MX Bluetooth Kyb/Mouse, which was fine until 5.10,
>>>> and
>>>> after having shelved this expensive monster for a year, I see it's
>>>> still
>>>> not fine in 7.04 - but at least re-plugging it seems to re-awaken it.
>>>> I'll keep an eye on this thread, but email me if there is anything I
>>>> can
>>>> add or help with.
>>>>
>>>>
>>> Do you have any other bluetooth devices? Mobile phone, headset,...
>>> Do they work when you re-plug? Before and/or after?
>>>
>>>
>> The unit only connects to it's own kyb & mouse - my Nokia E70 cannot see
>> it.
>>
>
> That confirms you have the same bug.
> Your E70 *should* be able to see the pc and vice versa.
> Because it can't, that means your bluetooth unit is now in HID legacy
> (USB) mode. When you type sudo hid2hci at console, you will loose keyb &
> rodent but your E70 will be able to see the pc. (if hid2hci dosn't give
> any other error)
>
>
that is 100% correct. (just tried it)
So, do I just sit back and wait for a fix?
Z

--

===========================================
Geograph (Pty) Ltd.
P.O. Box 31255, Tokai 7966, South Africa.

B3-106 Tokai Village, cnr Vans/Tokai Rds,
Tokai 7945, Western Cape, South Africa.

Tel: [phone] Mobile: [phone]
Fax: [phone]
===========================================

[Amedee Van Gasse (amedee)]

Op dinsdag 15-05-2007 om 15:04 uur [tijdzone +0000], schreef Zoltan:
> amedee wrote:
> > On Tue, May 15, 2007 15:40, Zoltan wrote:
> >
> >> amedee wrote:
> >>
> >>> On Tue, May 15, 2007 07:58, Zoltan wrote:
> >>>
> >>>
> >>>> Hi,
> >>>> Just want to add my name to the list of peeps affected by this bug :-)
> >>>> I have the Logitech MX Bluetooth Kyb/Mouse, which was fine until 5.10,
> >>>> and
> >>>> after having shelved this expensive monster for a year, I see it's
> >>>> still
> >>>> not fine in 7.04 - but at least re-plugging it seems to re-awaken it.
> >>>> I'll keep an eye on this thread, but email me if there is anything I
> >>>> can
> >>>> add or help with.
> >>>>
> >>>>
> >>> Do you have any other bluetooth devices? Mobile phone, headset,...
> >>> Do they work when you re-plug? Before and/or after?
> >>>
> >>>
> >> The unit only connects to it's own kyb & mouse - my Nokia E70 cannot see
> >> it.
> >>
> >
> > That confirms you have the same bug.
> > Your E70 *should* be able to see the pc and vice versa.
> > Because it can't, that means your bluetooth unit is now in HID legacy
> > (USB) mode. When you type sudo hid2hci at console, you will loose keyb &
> > rodent but your E70 will be able to see the pc. (if hid2hci dosn't give
> > any other error)
> >
> >
> that is 100% correct. (just tried it)
> So, do I just sit back and wait for a fix?

Yeah, you, me and a couple of tens, hundred or perhaps even thousand
Logitech Bluetooth users that also happen to use Ubuntu Linux have been
waiting for more than a year now. Join the club. Annual renewal of your
membership is free.

--
Amedee Van Gasse <email address hidden>

[Zoltan (zoltans)]

amedee wrote:
> Op dinsdag 15-05-2007 om 15:04 uur [tijdzone +0000], schreef Zoltan:
>
>> amedee wrote:
>>
>>> On Tue, May 15, 2007 15:40, Zoltan wrote:
>>>
>>>
>>>> amedee wrote:
>>>>
>>>>
>>>>> On Tue, May 15, 2007 07:58, Zoltan wrote:
>>>>>
>>>>>
>>>>>
>>>>>> Hi,
>>>>>> Just want to add my name to the list of peeps affected by this bug :-)
>>>>>> I have the Logitech MX Bluetooth Kyb/Mouse, which was fine until 5.10,
>>>>>> and
>>>>>> after having shelved this expensive monster for a year, I see it's
>>>>>> still
>>>>>> not fine in 7.04 - but at least re-plugging it seems to re-awaken it.
>>>>>> I'll keep an eye on this thread, but email me if there is anything I
>>>>>> can
>>>>>> add or help with.
>>>>>>
>>>>>>
>>>>>>
>>>>> Do you have any other bluetooth devices? Mobile phone, headset,...
>>>>> Do they work when you re-plug? Before and/or after?
>>>>>
>>>>>
>>>>>
>>>> The unit only connects to it's own kyb & mouse - my Nokia E70 cannot see
>>>> it.
>>>>
>>>>
>>> That confirms you have the same bug.
>>> Your E70 *should* be able to see the pc and vice versa.
>>> Because it can't, that means your bluetooth unit is now in HID legacy
>>> (USB) mode. When you type sudo hid2hci at console, you will loose keyb &
>>> rodent but your E70 will be able to see the pc. (if hid2hci dosn't give
>>> any other error)
>>>
>>>
>>>
>> that is 100% correct. (just tried it)
>> So, do I just sit back and wait for a fix?
>>
>
> Yeah, you, me and a couple of tens, hundred or perhaps even thousand
> Logitech Bluetooth users that also happen to use Ubuntu Linux have been
> waiting for more than a year now. Join the club. Annual renewal of your
> membership is free.
>
>
yay

--

===========================================
Geograph (Pty) Ltd.
P.O. Box 31255, Tokai 7966, South Africa.

B3-106 Tokai Village, cnr Vans/Tokai Rds,
Tokai 7945, Western Cape, South Africa.

Tel: [phone] Mobile: [phone]
Fax: [phone]
===========================================

[Alex Mayorga (alex-mayorga)]

@amedee, Zoltan
Can you see if this helps http://ubuntuforums.org/showpost.php?p=2334318&postcount=17 I have a fully functional Logitech diNovo Laser now, at least one fellow ubuntero has reported success with that guide.

My 2¢,
Alex

[Amedee Van Gasse (amedee)]

On Tue, May 22, 2007 16:53, alex_mayorga wrote:
> @amedee, Zoltan
> Can you see if this helps
> http://ubuntuforums.org/showpost.php?p=2334318&postcount=17 I have a fully
> functional Logitech diNovo Laser now, at least one fellow ubuntero has
> reported success with that guide.

I'll try it tonight (not at that pc right now) but anyway... this is a
highly specialized manual operation, not suitable for first time Linux
users.
The main issue of this bug is that the Logitech bluetooth sets don't work
out of the box.

--
Amedee

[neskweek (sebastiencapou)]

Hi

I've got a Dinovo Keyboard set (First Generation) installed on a PC (Intel Pentium III Prescot E)
After some times on Gentoo I have installed a Kubuntu feisty with the alternate CD.

My box is running the 2.6.20-16-generic #2 SMP kernel.

I have this problem too. Having spend some times fuguring out on my gentoo how to make work the dinovo set, I've tried looking why kubuntu didn't do it.

First it can work by doing by hand :

hidd --connect AA:BB:CC:DD:EE:FF (Mouse or whatever)
hidd --connect AA:BB:CC:DD:EE:GG (keybord or whatever)
hidd --connect AA:BB:CC:DD:EE:HH (Mediapad or whatever)

Like in other distros by the way. The first time you HAVE TO hidd --connect your device by hand. (don't know why)
The problem is then when you reboot it don't keep the config (other distros does or may)

putting
HIDD_ENABLED=1
in /etc/default/bluetooth doesn't change this
nor in /etc/init.d/bluettoth

putting
HIDD_OPTIONS=" --connect AA:BB:CC:DD:EE:FF --connect AA:BB:CC:DD:EE:GG --master --server"
doesn't make it either (it may by the way but since you make a connect for the first time you need, on dinovo devices, to press a connection button, it doesn't work for that reason : you need to press the buttons on your devices at each boot.)

The most strange thing comes here :

I've tried to make an hcitool scan ... and nothing came. So I tried an hciconfig on the /dev/hci0... it didn't exists :
Normally it shouldhave : when you boot /etc/init.d/bluetooth it launches hid2hci script which create an /dev/hci0 device from /dev/hiddev0.

so I tried to run hid2hci manually... it gives on the console :
Switching device 046d:c704 to HCI mode was successful
But here is what I get in /var/log/messages :

[neskweek (sebastiencapou)]

(:P Send the last report as I didn't finished it. Sorry Off course the attached file is to read after my thread )

Then in the kernel bug report we can see that the bug you're hunting is located on the hidp kernel module.

I'm currently searching if this patch was installed in the ubuntus kernel : http://www.bluez.org/patches.html (my guess is that it's not. That could explain why bluetooth mouses and bluetooth keyboards don't come up )

Here is my thread on gentoo forums related to Dinovo Keyboard set. Hope it Helps:
http://forums.gentoo.org/viewtopic-t-231506-highlight-dinovo.html

PS : the hidd --connect thing is to be done when you want to use your mouses keyboards in BT mode. When you desinstall bluez-packages mouses and keyboards work because then you use another kernel modules (HID if I remember well, in USB section) BUT using that modules restrict the radius of use of your devices (1 or 2 meters instaed of 10 meters using BT mode). Also you may not be able to use all buttons as intended (don't know why ... just seen it).

PS2: As I was searching I saw this in /etc/default/bluetooth :

############ HIDD
#
# To have Bluetooth mouse and keyboard support, get the
# Linux 2.6.6 patch or better from bluez.org, and set
# HIDD_ENABLED to 1.

Clear enough : the bluez patch isn't installed that's why mouses and keyboards doesn't work in BT mode.
No need to dig more
I install it and come back to you to tell you the results

[Amedee Van Gasse (amedee)]

Any results yet, neskweek?

[neskweek (sebastiencapou)]

Sorry :P
Depited I forgot to post my reults :
When I tryed patching the kernel with bluez patch, the patch process reports that almost all modifications have already been done to the kernel (which is weird... maybe that's due to some modification made by the ubuntu team but not exactly the same that bluez patch. That's only an assumption)

Anyway I tried to compile it, but there's no use: compilation breaks each time. I don't remember where (sorry). I tried several times with differents options but it fails. It doesn't compile with that patch. Without it does.

Seeing that , as I wrote earlyer I "think" that ubuntu team applyed a custom patch derived from bluez one (or maybe a older patch created for older kernel version , but I doubt that one) for any reason. But that patch isn't working. I think another patch conflict with that one.

As I work on another thing I just spend two days on that. For now, I just desinstall bluez packages... Making my Dinovo set working as an USB device :/

[Amedee Van Gasse (amedee)]

On Fri, August 3, 2007 10:02, neskweek wrote:
> Sorry :P
> Depited I forgot to post my reults :
> When I tryed patching the kernel with bluez patch, the patch process
> reports that almost all modifications have already been done to the kernel
> (which is weird... maybe that's due to some modification made by the
> ubuntu team but not exactly the same that bluez patch. That's only an
> assumption)
>
> Anyway I tried to compile it, but there's no use: compilation breaks
> each time. I don't remember where (sorry). I tried several times with
> differents options but it fails. It doesn't compile with that patch.
> Without it does.
>
> Seeing that , as I wrote earlyer I "think" that ubuntu team applyed a
> custom patch derived from bluez one (or maybe a older patch created for
> older kernel version , but I doubt that one) for any reason. But that
> patch isn't working. I think another patch conflict with that one.

What if you compile a vanilla kernel from kernel.org?
Copy .config to the vanilla kernel, then make oldconfig and fix config
issues (if any).

--
Amedee

[bubulle (rambuehl)]

Hi All,
I am new to this discussion, just tripped over it while looking for something totally different.
If I say something silly, please let me know. Just trying to help, but I may be totally wrong here :-)

My understanding of the symptoms:
- a bluetooth mouse or keyboard, regardless of the brand or chipset, will stop working when GDM comes up
- reconnecting explicitely the device(s) at that point restores the device functionality, but then it is not working anymore at next boot, until GDM is up and running again
- the problem seems system dependant

It looks like the devices stop working at the exact moment when the Bluetooth stack is loaded and the adapter is switched from HID to HCI mode. Despite knowing the device address, the stack fails in reconnecting the device.

A similar problem can occur if there is a discrepancy in Bluetooth link key inside the broadcom BT adapter in HID mode, and the link key stored by the host stack.

Again, I have no clue about the actual implementation of Dapper (actually I heard that name for the first time today :-), but my guess is that there must be some code / HCI commands missing to synchronize the link keys, or alternatively that depending on the silicon used that synchronisation does not work as expected, and requires some undocumented commands. If not done already, it would probably help to look at the HCI traffic on USB on the original OS supplied with the machine (MacOS X) and compare it with what Dapper does.

[Ron Adams (tohuw)]

I can confirm a variation on Gusty Beta, fully patched as of 9/29. I have an MX Revolution mouse, which works fine throughout (as expected; it's not bluetooth and has its own dongle), but my MX 5000 bluetooth keyboard does not work until it is replugged.

AMD 64-FX running x86 Gusty Beta.

If you need any logs, let me know.

[Alex Mayorga (alex-mayorga)]

@Ron Adams,

Might it be you're indeed seeing Bug #123920 ?

[DaveAbrahams (boostpro)]

This is still a problem in Gutsy. My Logitech mouse stops working if I leave the computer too long. The only workaround I've found is to turn the physical bluetooth switch on the laptop off and on again.

[ChaseNeveu (chaseneveu-gmail)]

After searching the internet for solutions to my particular bluetooth problem, I've come across dozens of these bug reports. It's the same for each one:

"My keyboard/mouse won't reconnect after reboot/suspend/idle."

And the response has been the same too:

"Try entering this into your configs and/or uninstall bluetooth entirely." ('this' being a reference to instructions posted in one particular thread I've been directed to about 12 separate times; http://ubuntuforums.org/showthread.php?t=227057)

I've read through all these bug reports for a solution, tried the workarounds--I'm still unable to get my keyboard or mouse to consistently connect on boot. It's not a viable option for me to find another keyboard every time I need to LOG IN. It wouldn't make me so upset if this was a relatively new occurrence. It's not. It's been around for a long while. Dapper, Feisty, Gusty, and now symptoms on Hardy too. Correct me if I'm wrong, but I thought these bug reports were supposed to fix bugs, not direct people to workarounds that don't work.

If there are so many reports, how is this problem not yet fixed? How is this on the back burner, while other negligible problems are being solved? Or maybe I'm the only one who needs a keyboard or mouse to operate Ubuntu--I don't think that's the case, considering the prevalence of laptops and wireless solutions touted in this modern age. These problems need to be fixed, especially if usability is of concern here.

It's not very welcoming to boot into Linux, only to be greeted by a login screen with no input method.

[Amedee Van Gasse (amedee)]

On Sun, March 23, 2008 01:42, ChaseNeveu wrote:
> After searching the internet for solutions to my particular bluetooth
> problem, I've come across dozens of these bug reports. It's the same for
> each one:
>
> "My keyboard/mouse won't reconnect after reboot/suspend/idle."
>
> And the response has been the same too:
>
> "Try entering this into your configs and/or uninstall bluetooth
> entirely." ('this' being a reference to instructions posted in one
> particular thread I've been directed to about 12 separate times;
> http://ubuntuforums.org/showthread.php?t=227057)
>
> I've read through all these bug reports for a solution, tried the
> workarounds--I'm still unable to get my keyboard or mouse to
> consistently connect on boot. It's not a viable option for me to find
> another keyboard every time I need to LOG IN. It wouldn't make me so
> upset if this was a relatively new occurrence. It's not. It's been
> around for a long while. Dapper, Feisty, Gusty, and now symptoms on
> Hardy too. Correct me if I'm wrong, but I thought these bug reports were
> supposed to fix bugs, not direct people to workarounds that don't work.
>
> If there are so many reports, how is this problem not yet fixed? How is
> this on the back burner, while other negligible problems are being
> solved? Or maybe I'm the only one who needs a keyboard or mouse to
> operate Ubuntu--I don't think that's the case, considering the
> prevalence of laptops and wireless solutions touted in this modern age.
> These problems need to be fixed, especially if usability is of concern
> here.
>
> It's not very welcoming to boot into Linux, only to be greeted by a
> login screen with no input method.
>
ACK all the way.

I will give a Logitech bluetooth keyboard+mouse to the developer who can
solve this problem, or I will donate the value of that keyb+mouse to a
charity of his/her choice. That is an incentive of about 100€ if I'm not
mistaken. If there are others who are also willing to donate, I'll set up
a pledge site to get this bug fixed.

[Zoltan (zoltans)]

My logitech cordless elite bluetooth hub connects with USB.
When you get the first login prompt, just pull out the USB plug, wait a second or two and plug it back in. Both your keyboard and mouse will then wake up. Tadaaa!

I fully agree that this is the most pathetically tedious bug I have ever seen. It probably also causes the most bad press for Linux as every bluetooth user will get caught by it.
Surely the relevant developers can see from the amount of threads and postings on this subject, that it is a critical bug?

Guys (that's the developers out there), if you want to rule the world so you dont' have to keep looking out a window, clean up the visible bugs.

No one cares about the esoteric and deep down inside security bugs (ok, until they get caught out) that the Linux brain-surgeons fix. It's about visibility, not what is good and correct - Just ask our president elect: Mr Zuma ;-)

my 2c

[Amedee Van Gasse (amedee)]

On Tue, March 25, 2008 06:58, Zoltan wrote:
> My logitech cordless elite bluetooth hub connects with USB.
> When you get the first login prompt, just pull out the USB plug, wait a
> second or two and plug it back in. Both your keyboard and mouse will then
> wake up. Tadaaa!

Nada tadaaa!

When you unplug your bluetooth dongle, and plug it back in, it goes into a
"legacy mode" or "bios mode" or "usb compatibility mode" or "hid mode".
This is not good. The _hid_ mode has reduced bluetooth functionality. Your
mouse and keyboard will work (but only the standard buttons, not any fancy
multimedia buttons on your keyboard, and only 3 buttons on the mouse, not
any of the 9 other buttons). Any other bluetooth device, like a
smartphone, will not be able to connect.

To have full bluetooth functionality, your dongle needs to be in _hci_
mode. In hci mode, you have *all* buttons on keyboard and mouse, and you
will be able to connect with your smartphone.

When you issue the hid2hci command, you will notice that you are able to
connect your smartphone. However, when it asks you to enter a pin code on
the pc to verify the bluetooth connection, you will curse. Why? Because
with hid2hci, you have shut out your keyboard again...

--
Amedee

[Fabián Rodríguez (magicfab)]

During the past few weeks a lot of updates have been published for bluez-utils and related packages. See the changelog:
http://changelogs.ubuntu.com/changelogs/pool/main/b/bluez-utils/bluez-utils_3.26-0ubuntu4/changelog

These are available in up-to-date Ubuntu Hardy 8.04 installs.

Feb. 15 updates include input modules (so, keyboard and mouse) updates. If anyone having BT input devices pairing problems could download a LiveCD Hardy, update it and try the pairing method described in the (recently updated) docs, it would help isolate this problem from 8.04 and focus on what the solution for 7.10 could be:
https://help.ubuntu.com/community/BluetoothInputDevices

I've also subscribed Canonical QA to this bug, hoping we can get their input (no pun intended) :)

Thank you.

[Jaba (jpylvanainen)]

General
I followed the howto on:

http://www.ubuntugeek.com/howto-setup-bluetooth-keyboard-and-mouse-in-ubuntu.html

I got MX5000 to work in Hardy in bluetooth mode by using:
sudo hidd --search to connect the keyboard
sudo hidd --connect MOUSE_ADDR, to connect the mouse
then i restarted bluetooth and gdm and they started working.

There is still work to be done to make use of all the extra keys
and buttons.

[Amedee Van Gasse (amedee)]

Jaba schreef:
> General
> I followed the howto on:
>
> http://www.ubuntugeek.com/howto-setup-bluetooth-keyboard-and-mouse-in-
> ubuntu.html
>
>
> I got MX5000 to work in Hardy in bluetooth mode by using:
> sudo hidd --search to connect the keyboard
> sudo hidd --connect MOUSE_ADDR, to connect the mouse
> then i restarted bluetooth and gdm and they started working.
>
> There is still work to be done to make use of all the extra keys
> and buttons.
>
>
Did you type all that with your bluetooth keyboard?
And did you test another device too, like a smartphone.
All 3 devices should work: keyboard, mouse, smartphone.
If at least one of them does not work properly, the issue is not solved.

--
Amedee

[Leann Ogasawara (leannogasawara)]

@Jaba, thanks for testing and the feedback.

@everyone, tt unfortunately seems this report has gone unnoticed by developers for quite some time and I'd like to bring this to their attention. However, we first need to have the proper debugging information against the latest release. As Fabian pointed out, it would be great if people could test with the latest Hardy Heron 8.04 release. You should be able to test via a LiveCD. Hardy Heron 8.04 release candidates are currently available and the daily images can be found here: http://cdimage.ubuntu.com/daily-live/current/ . If this issue still exists, please include the following information. Be sure to attach each file separately.

* lsusb > lsusb.log
* dmesg > dmesg.log

Thanks again and we appreciate your help and support.

[DaveAbrahams (boostpro)]

This is still a problem in Hardy with all the latest updates: My Logitech mouse stops working if I leave the computer too long. The only workaround I've found is to turn the physical bluetooth switch on the laptop off and on again.

[Leann Ogasawara (leannogasawara)]

Hi Dave,

Care to also maybe attach your /var/log/syslog. Thanks.

[DaveAbrahams (boostpro)]

Done

[Jamey R (jamesracer)]

I'm using the Logitech diNovo Media Desktop Laser, and Hardy/8.04 64 bit. I seem to have the issue that requires unplug, and plug of the bluetooth dongle. All the mouse buttons (except the tilt wheel) work after some tinkering, but I still can't get the bluetooth to connect at boot. Let me know if I can help by providing any info. (I'm new to Linux [this month!]) But the whole newbie thing, still, has given me less fits than Vista Ultimate 64.

[Charlotte Curtis (c-f-curtis)]

I've got part of this bug too, in 8.04. Most of the time, my MX5000 keyboard and MX1000 mouse connect on startup just fine, but after they go to sleep (~15 minutes away from the computer), I lose connection. I have another mouse handy so that I can open up the bluetooth preferences and reconnect the devices (simply clicking "disconnect" then wiggling the mouse/hitting a key seems to work). Then, I need to restart X so that all of my mouse buttons work properly.

It seems as though there are two parts to this problem:
1) bluez doesn't know when devices disconnect. Since they reconnect automatically if I disconnect them manually, it seems as though automatic detection of the disconnect would solve most of the problem.
2) When a device disconnects, Xorg unloads the driver, as seen in this portion of my Xorg.0.log:

(EE) Read error: No such device (19, -1 != 16)
(II) Logitech MX1000: Off
(II) Logitech MX1000: Off
(II) UnloadModule: "evdev"

Is there a way to stop X from unloading modules and switching off input devices?

[GG (giorgio-gilestro)]

So I think I found people with a similar problem finally.
I have this symptoms: everything works ok until the computer goes to a long idle time: usually the overnight break.
When I am back in the morning the keyboard won't work anymore and I am forced to restart (I can still use the mouse).
Sometimes I get some segfault in gdm during the shuting down procedure but I am not sure that is related.
Now I unplugged the bluetooth dongle from the PC last week and guess what? Problem seems to have disappeared.
The funny thing is: *my keyboard is not a bluetooth one! It's just a regular usb keyboard*!!!

I know lots of people out there are experiencing the same issue. I could not think it was somehow related to bluetooth!

[DaveAbrahams (boostpro)]

This problem seems to have magically disappeared, perhaps with the latest updates. It's too early to tell for certain; I will report back here if it happens again.

[Amedee Van Gasse (amedee)]

On Thu, June 12, 2008 01:27, DaveAbrahams wrote:
> This problem seems to have magically disappeared, perhaps with the
> latest updates. It's too early to tell for certain; I will report back
> here if it happens again.

I still have the problem with Ubuntu 8.04.

But with OpenSuse, the bluetooth mouse&keyboard work out of the box. I
didn't have to do anything at all!
So the solution for me personally is: switch to OpenSuse. As long as it's
Linux, I really don't care which distribution it is.

--
Amedee

[Mario Limonciello (superm1)]

Since this was confirmed to be fixed by a user, I'm closing out the bug. If anyone has these problems still with the latest Intrepid packages, please open a new bug.

[Amedee Van Gasse (amedee)]

I wrote a workaround (not a fix) in this bug report:
https://bugs.launchpad.net/ubuntu/+source/bluez/+bug/318465

[Amedee Van Gasse (amedee)]

I disagree, the bug isn't fixed. Not even in 11.10.
A workaround is not a fix. A fix is something that makes it work out of the box in the next version.

[Imfor Umman (imforumman)]

This is still not fixed in 12.04 LTS: Logitech MX5500 Revolution does not work after booting the system, rendering the system dead for the non-tech-savy user and is just plain and simple a game-killer (bug-importance: Undecided / Medium???). Replugging the BT receiver is not an honest option.

BTW: The title is misleading: Not only Apple is affected and I have no idea about "Dapper/Edgy/Feisty", if this has been going on for six years, "everything" seems more inclusive.

[Thomas Hotz (thotz-deactivatedaccount)]

Can someone please reopen the bluez task? Thank you!