Bug 190831 - Intel Bluetooth 8260 breaks with latest firmware
Summary: Intel Bluetooth 8260 breaks with latest firmware
Status: CLOSED MOVED
Alias: None
Product: Drivers
Classification: Unclassified
Component: network-wireless (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: DO NOT USE - assign "network-wireless-intel" component instead
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-12-21 17:54 UTC by Jan Steffens
Modified: 2017-11-25 10:39 UTC (History)
17 users (show)

See Also:
Kernel Version: 4.9
Subsystem:
Regression: Yes
Bisected commit-id:


Attachments
DDC file from release (38 bytes, patch)
2017-01-05 13:14 UTC, G Jaya Praveen
Details | Diff
Journal (Kernel & Userspace) from my Laptop (6.83 KB, text/x-log)
2017-01-05 13:59 UTC, Stefan Schmid
Details
Patches (569.35 KB, application/octet-stream)
2017-01-06 06:19 UTC, G Jaya Praveen
Details
DDC file from release (18 bytes, application/octet-stream)
2017-01-06 06:27 UTC, G Jaya Praveen
Details
2th Journal (Kernel & Userspace) from my Laptop (5.28 KB, text/plain)
2017-01-06 06:46 UTC, Stefan Schmid
Details
Latest patch (387.36 KB, application/x-zip-compressed)
2017-01-06 07:27 UTC, G Jaya Praveen
Details
SFI latest (573.09 KB, patch)
2017-01-17 17:07 UTC, G Jaya Praveen
Details | Diff
DDC latest (36 bytes, patch)
2017-01-17 17:08 UTC, G Jaya Praveen
Details | Diff
Patch SFI and DDC (389.69 KB, patch)
2017-01-19 06:54 UTC, G Jaya Praveen
Details | Diff
DDC file (18 bytes, application/octet-stream)
2017-01-19 11:16 UTC, G Jaya Praveen
Details
attachment-32063-0.html (1.41 KB, text/html)
2017-02-25 02:51 UTC, ahxcker
Details
Core24 with uSniffer (2.12 MB, application/octet-stream)
2017-02-26 10:31 UTC, Emmanuel Grumbach
Details
Debug-Log from iwlwifi on Kernel 4.10.1 (4.26 KB, application/octet-stream)
2017-03-01 11:28 UTC, Stefan Schmid
Details
attachment-14468-0.html (1.03 KB, text/html)
2017-03-01 15:34 UTC, ahxcker
Details
agurenko-dmesg.log (78.19 KB, text/plain)
2017-03-01 19:10 UTC, Gurenko Alex
Details
new driver from master branch, firmware 22 and 27 still not working (31.69 KB, application/gzip)
2017-03-01 19:30 UTC, cyberwolf
Details
agurenko-dmesg.log with -27 firmware (74.02 KB, text/plain)
2017-03-01 19:30 UTC, Gurenko Alex
Details
Core24 with uSniffer (2.12 MB, application/octet-stream)
2017-03-01 19:37 UTC, Emmanuel Grumbach
Details
agurenko-dmesg.log latest (70.51 KB, text/plain)
2017-03-01 19:54 UTC, Gurenko Alex
Details

Description Jan Steffens 2016-12-21 17:54:07 UTC
My Bluetooth mouse only works for a few seconds (the exact time appears random) after the Bluetooth controller is reset. At first, it seems only its "frame rate" is reduced drastically, then it completely stops working. It also makes pairing nearly impossible. "hciconfig hci0 reset" briefly gets the mouse working again.


Firmware revision "0.0 build 95 week 37 2016" from linux-firmware 91ddce4 (built 2016-12-05) is broken. The driver boots with these messages:

Bluetooth: hci0: Bootloader revision 0.0 build 2 week 52 2014
Bluetooth: hci0: Device revision is 5
Bluetooth: hci0: Secure boot is enabled
Bluetooth: hci0: OTP lock is enabled
Bluetooth: hci0: API lock is enabled
Bluetooth: hci0: Debug lock is disabled
Bluetooth: hci0: Minimum firmware build 1 week 10 2014
Bluetooth: hci0: Found device firmware: intel/ibt-11-5.sfi
Bluetooth: hci0: Waiting for firmware download to complete
Bluetooth: hci0: Firmware loaded in 1474407 usecs
Bluetooth: hci0: Waiting for device to boot
Bluetooth: hci0: Device booted in 10684 usecs
Bluetooth: hci0: Found Intel DDC parameters: intel/ibt-11-5.ddc
Bluetooth: hci0: Failed to send Intel_Write_DDC (-22)


Firmware revision "0.0 build 30 week 18 2016" from linux-firmware 9c71af9 (built 2016-10-05) works fine. The driver boots with these messages:

Bluetooth: hci0: Bootloader revision 0.0 build 2 week 52 2014
Bluetooth: hci0: Device revision is 5
Bluetooth: hci0: Secure boot is enabled
Bluetooth: hci0: OTP lock is enabled
Bluetooth: hci0: API lock is enabled
Bluetooth: hci0: Debug lock is disabled
Bluetooth: hci0: Minimum firmware build 1 week 10 2014
Bluetooth: hci0: Found device firmware: intel/ibt-11-5.sfi
Bluetooth: hci0: Waiting for firmware download to complete
Bluetooth: hci0: Firmware loaded in 1555030 usecs
Bluetooth: hci0: Waiting for device to boot
Bluetooth: hci0: Device booted in 16567 usecs
Bluetooth: hci0: Found Intel DDC parameters: intel/ibt-11-5.ddc
Bluetooth: hci0: Applying Intel DDC parameters completed
Bluetooth: hci0: Setting Intel event mask failed (-16)


This problem seemed similar to https://bugzilla.kernel.org/show_bug.cgi?id=189641 but the loss of function doesn't correlate with any messages. However, while it is broken I do see the occasional repeat of

Bluetooth: hci0 command 0x200c tx timeout

with varying command numbers.


As an aside, it appears the driver will only print the version if a firmware is already loaded. This is a bit irritating as after a complete power-off (to clear the firmware) btusb has to be unloaded and reloaded in order to print the version.


Kernel is linux-zen 4.8.14 x86_64.

Hardware:
Intel Wireless 8260
Logitech MX Master

Downstream bug: https://bugs.archlinux.org/task/52214
Comment 1 Jan Steffens 2016-12-21 17:55:33 UTC
I might try running the new firmware file with the old DDC parameter file. Maybe that'll work.
Comment 2 Jan Steffens 2016-12-21 18:08:39 UTC
New firmware with old DDC indeed works. The "Setting Intel event mask failed" message seen with the old firmware didn't reappear, either.

Bluetooth: hci0: Bootloader revision 0.0 build 2 week 52 2014
Bluetooth: hci0: Device revision is 5
Bluetooth: hci0: Secure boot is enabled
Bluetooth: hci0: OTP lock is enabled
Bluetooth: hci0: API lock is enabled
Bluetooth: hci0: Debug lock is disabled
Bluetooth: hci0: Minimum firmware build 1 week 10 2014
Bluetooth: hci0: Found device firmware: intel/ibt-11-5.sfi
Bluetooth: hci0: Waiting for firmware download to complete
Bluetooth: hci0: Firmware loaded in 1424291 usecs
Bluetooth: hci0: Waiting for device to boot
Bluetooth: hci0: Device booted in 15526 usecs
Bluetooth: hci0: Found Intel DDC parameters: intel/ibt-11-5.ddc
Bluetooth: hci0: Applying Intel DDC parameters completed
Comment 3 Jan Steffens 2016-12-21 19:56:50 UTC
Actually, that combination just got stuck in the "low framerate" state again, but it did not completely stop working. Again, a hciconfig reset restores it to working order only for a few seconds.
Comment 4 Jan Steffens 2016-12-21 20:23:45 UTC
Retested with 4.9. No change.
Comment 5 G Jaya Praveen 2016-12-22 06:49:55 UTC
Hi,

A Patch with the latest DDc is submitted. Still awaiting for merge.(8th December sent to merge).
Issue will be resolved for you once it got merged to Linux-firmware.
Comment 6 Jan Steffens 2016-12-22 20:35:47 UTC
Thank you. Is that patch already available in any public branch?
Comment 7 G Jaya Praveen 2016-12-23 09:16:25 UTC
Hi,

Currently requested the maintainers to merge it, but still not present in Linux-firmware public branch.
Comment 8 Stefan Schmid 2017-01-05 10:09:03 UTC
(In reply to G Jaya Praveen from comment #5)
> Hi,
> 
> A Patch with the latest DDc is submitted. Still awaiting for merge.(8th
> December sent to merge).
> Issue will be resolved for you once it got merged to Linux-firmware.

We have already January and a functioning firmware is still not available. Would it be possible to get an alternative link for the download of the new firmware?
Comment 9 G Jaya Praveen 2017-01-05 13:14:43 UTC
Created attachment 250381 [details]
DDC file from release

Adding the ddc file for you to verify till the actual fw get merged.
Comment 10 Stefan Schmid 2017-01-05 13:59:28 UTC
Created attachment 250391 [details]
Journal (Kernel & Userspace) from my Laptop

(In reply to G Jaya Praveen from comment #9)
> Created attachment 250381 [details]
> DDC file from release
> 
> Adding the ddc file for you to verify till the actual fw get merged.

With this file it does not work either.

My Device:
01:00.0 Network controller [0280]: Intel Corporation Wireless 8260 [8086:24f3] (rev 3a)
Bus 001 Device 006: ID 8087:0a2b Intel Corp.
Comment 11 Marcin Kurek 2017-01-05 14:44:20 UTC
I can observe same symptoms as reporter on Precision 5510 with 8260 [8086:24f3] (rev 3a) Bluetooth works with older version of firmware:

2010a4f3942b6df95ed928f92db3d000  ibt-11-5.ddc
2f06625d1cab8e6edb339dc7a0db4e91  ibt-11-5.sfi

With new version my mouse stops a seconds after connection. When I try to mix old sdi and old ddc it seems to works, but the mouse pointer is jumpy as hell and quite hard to use (but as mentioned the event mask error is not apear with mixed versions)

I will try with attached one in a second and see.
Comment 12 Marcin Kurek 2017-01-05 14:53:03 UTC
No good. New ddc file doesn't help here either. It more info is required please ask.
Comment 13 G Jaya Praveen 2017-01-06 06:19:34 UTC
Created attachment 250451 [details]
Patches
Comment 14 G Jaya Praveen 2017-01-06 06:27:52 UTC
Created attachment 250461 [details]
DDC file from release
Comment 15 G Jaya Praveen 2017-01-06 06:28:25 UTC
Please try with the latest sfi and ddc attached.
Comment 16 Stefan Schmid 2017-01-06 06:46:19 UTC
Created attachment 250471 [details]
2th Journal (Kernel & Userspace) from my Laptop

(In reply to G Jaya Praveen from comment #15)
> Please try with the latest sfi and ddc attached.

With these two files it works.
It is now even possible to connect several connections without the mouse problems.
Comment 17 G Jaya Praveen 2017-01-06 06:59:53 UTC
Ok.. With the latest release SFI and the  DDC,that I had attached yesterday. you still see the issues right?
Comment 18 Stefan Schmid 2017-01-06 07:06:20 UTC
(In reply to G Jaya Praveen from comment #17)
> Ok.. With the latest release SFI and the  DDC,that I had attached yesterday.
> you still see the issues right?

Attachment 250451 [details] & 250471 work properly.
Attachment 250381 [details] has been damaged during uploading.
Comment 19 G Jaya Praveen 2017-01-06 07:27:25 UTC
Created attachment 250481 [details]
Latest patch
Comment 20 G Jaya Praveen 2017-01-06 07:28:10 UTC
Please test again with this latest patch attached and update.

Thanks
Comment 21 Stefan Schmid 2017-01-06 08:00:11 UTC
(In reply to G Jaya Praveen from comment #19)
> Created attachment 250481 [details]
> Latest patch

With these two files it also works but if more than one connection is active then the mouse does not function properly.
Comment 22 Marcin Kurek 2017-01-06 12:25:12 UTC
Hmm, the #250451 and #250461 looks like current version of .sfi file and older .ddc file.

Anyway just tired with #250481:

873629534cf972411618c6c13aef80f2  ibt-11-5.ddc
413da432ba53076ed31a1dda17a1f43d  ibt-11-5.sfi

The .sfi file has same md5 sum as file currently in linux-firmware repository and only .ddc is different. And it loads fine:

Bluetooth: hci0: Bootloader revision 0.0 build 2 week 52 2014
Bluetooth: hci0: Device revision is 5
Bluetooth: hci0: Secure boot is enabled
Bluetooth: hci0: OTP lock is enabled
Bluetooth: hci0: API lock is enabled
Bluetooth: hci0: Debug lock is disabled
Bluetooth: hci0: Minimum firmware build 1 week 10 2014
Bluetooth: hci0: Found device firmware: intel/ibt-11-5.sfi
Bluetooth: hci0: Waiting for firmware download to complete
Bluetooth: hci0: Firmware loaded in 1312477 usecs
Bluetooth: hci0: Waiting for device to boot
Bluetooth: hci0: Device booted in 10720 usecs
Bluetooth: hci0: Found Intel DDC parameters: intel/ibt-11-5.ddc
Bluetooth: hci0: Applying Intel DDC parameters completed

I can not see any event mask errors which is nice and for now I had no problem with sudden mouse stops/disconect, but it still feels a bit jumpy and definitely not so smooth as it should be.
Comment 23 Stefan Schmid 2017-01-08 10:46:45 UTC
Excuse me, I have to correct my previous statement

With this two Files:
2010a4f3942b6df95ed928f92db3d000  ibt-11-5.ddc Attachment 250451 [details]
413da432ba53076ed31a1dda17a1f43d  ibt-11-5.sfi Attachment 250461 [details]

Multiple Bluetooth connections are not a problem but a high WLAN activity. And somehow neither works anymore.
Comment 24 Marcin Kurek 2017-01-13 16:45:59 UTC
Any news on that ? As I can see the non working version was merged in linux-firmware repository just now: https://git.kernel.org/cgit/linux/kernel/git/firmware/linux-firmware.git/commit/?id=581f24500138f5e410d51ab63b205be9a52f4c77
Comment 25 Jan Steffens 2017-01-15 16:43:47 UTC
Yeah, the fix that was merged into linux-firmware is just the combination I tested in comment 3. It still has issues.
Comment 26 G Jaya Praveen 2017-01-17 17:07:40 UTC
Created attachment 252121 [details]
SFI latest
Comment 27 G Jaya Praveen 2017-01-17 17:08:08 UTC
Created attachment 252131 [details]
DDC latest
Comment 28 G Jaya Praveen 2017-01-17 17:09:46 UTC
Check with the latest FW again, this is still not upstreamed.
Please check and update the status.
Comment 29 Stefan Schmid 2017-01-18 13:49:41 UTC
(In reply to G Jaya Praveen from comment #28)
> Check with the latest FW again, this is still not upstreamed.
> Please check and update the status.

You have uploaded these files as "patch", so they are now corrupted. Please upload them again as "application / octet-stream" or in a tgz archive.
Comment 30 G Jaya Praveen 2017-01-19 06:54:34 UTC
Created attachment 252421 [details]
Patch SFI and DDC
Comment 31 G Jaya Praveen 2017-01-19 06:54:59 UTC
Zipped and attached the latest. Please check now.
Comment 32 Marcin Kurek 2017-01-19 10:22:46 UTC
I give a quick try to 252421, but it gives me an error on boot:

Bluetooth: hci0: Bootloader revision 0.0 build 2 week 52 2014
Bluetooth: hci0: Device revision is 5
Bluetooth: hci0: Secure boot is enabled
Bluetooth: hci0: OTP lock is enabled
Bluetooth: hci0: API lock is enabled
Bluetooth: hci0: Debug lock is disabled
Bluetooth: hci0: Minimum firmware build 1 week 10 2014
Bluetooth: hci0: Found device firmware: intel/ibt-11-5.sfi
Bluetooth: hci0: Waiting for firmware download to complete
Bluetooth: hci0: Firmware loaded in 1333260 usecs
Bluetooth: hci0: Waiting for device to boot
Bluetooth: hci0: Device booted in 10661 usecs
Bluetooth: hci0: Found Intel DDC parameters: intel/ibt-11-5.ddc
Bluetooth: hci0: Failed to send Intel_Write_DDC (-95)

But I give a try to another combination with ddc file from 250481 and this new sfi:

873629534cf972411618c6c13aef80f2  ibt-11-5.ddc
9bddc765acb16165ae970e7b588b5ea3  ibt-11-5.sfi

This combination seems to be fine, but it takes looooong time for mouse to make initial connection (I was sure bluetooth doesn't work at all)
Comment 33 Stefan Schmid 2017-01-19 10:31:24 UTC
The DDC file from the last upload may not work. It is not a data file like the others.

ibt-11-5.ddc: ASCII text, with no line terminators
Comment 34 G Jaya Praveen 2017-01-19 11:16:03 UTC
Created attachment 252431 [details]
DDC file

Check with this DDC.
Comment 35 Stefan Schmid 2017-01-20 13:21:45 UTC
(In reply to G Jaya Praveen from comment #30)
> Created attachment 252421 [details]
> Patch SFI and DDC

(In reply to G Jaya Praveen from comment #34)
> Created attachment 252431 [details]
> DDC file
> 
> Check with this DDC.

Bluetooth and WLAN are stable even if both active data are transmitted, but the mouse is sometimes a bit laggy.

Is it normal that every data transmission via Bluetooth only reaches about 721 kbit/s (Bluetooth 1.2)?
Comment 36 Gurenko Alex 2017-01-29 13:35:31 UTC
 I'm having similar issue with Bluetooth, but with Microsoft Surface Ergonomic keyboard. I've posted problem description here: https://ask.fedoraproject.org/en/question/100782/fedora-25-bluetooth-keyboard-hangs-bluetooth-stack/

 If it's broken only in latest linux-firmware, downgrading should fix this particular issue, right? Because for me it's not.
Comment 37 Robert E. Yawn 2017-01-31 00:34:35 UTC
I see this issue when using my bluetooth mouse (MS Arc Touch).

Not everyone is mentioning the devices that they are using.  Note that the MS Arc Touch and MS Surface Ergonomic Keyboard (as reported by Gurenko Alex) are both BT 4.0 devices.

My Logitech K810, a BT 3.0 device, works fine when it is the only device connected.  It is only when the BT 4.0 device is connected that these issues are occurring.
Comment 38 Gurenko Alex 2017-01-31 07:14:08 UTC
 Yes, I probably should've mentioned, that my Bose QuiteComfort 35 still works just fine, although I cannot find which bluetooth version it utilizes.
 I've had to reinstall my Fedora 25 KDE yesterday and it again worked just fine until I've upgraded all packages and even then it worked one session until the first reboot with new firmware.
Comment 39 Robert E. Yawn 2017-01-31 18:31:50 UTC
Interesting.  According to the official specs, the Bose QuietComfort 35 supports BT 4.1: https://www.cnet.com/products/bose-quietcomfort-35/specs/

It's possible that the issue maybe cropping up with certain hardware combos, too.  For example, I'm running Dell Precision Workstation 5510.  It uses the Int 8260 wifi/BT 4.1 combo with a Skylake CPU.
Comment 40 Gurenko Alex 2017-02-01 06:41:40 UTC
(In reply to Robert E. Yawn from comment #39)
> Interesting.  According to the official specs, the Bose QuietComfort 35
> supports BT 4.1: https://www.cnet.com/products/bose-quietcomfort-35/specs/
> 
> It's possible that the issue maybe cropping up with certain hardware combos,
> too.  For example, I'm running Dell Precision Workstation 5510.  It uses the
> Int 8260 wifi/BT 4.1 combo with a Skylake CPU.

I'm using Lenovo T460p with same combo, but laptop-grade CPU (i7-6820HQ).
Comment 41 requies89 2017-02-01 13:24:08 UTC
Same problem here. 
Laptop Dell 7559 i7-6700HQ (upgraded with Intel 8260) and mouse Microsoft Designer (Bluetooth 4.0).

Addendum - If i turn off and on the bluetooth leaving the wifi off, the mouse works with no problems. If i turn on the wifi and i connect to a wifi network (2.4G or 5G) the mouse stops working after 5-10 seconds.

We need a fix. :/
Comment 42 Marc Collin 2017-02-04 03:11:38 UTC
similar issue but with intel 7260
kernel: iwlwifi 0000:03:00.0: Detected Intel(R) Dual Band Wireless AC 7260, REV=0x144

i was thinking to buy an intel 8265
Comment 43 grk86 2017-02-09 06:48:17 UTC
Same problem here.

First stand:
- ThinkPad T460s
- Intel 8260 (rev 3a)
- kernel 4.9.7-201
- logitech mx master

Case: mouse stops working after 1-2 minutes of active work. When i need to reconnect it to laptop with bluetooth settings, after that mouse also works about 1-2 minutes and drop the connection. With unifying receiver it works fine.

Second:
- Zotac id83 plus
- kernel 4.4
- logitech mx master (the same)

Case: mouse stops working after 5-10 seconds of active work, if wifi is turned off it works fine, but then wifi in ON its stop working. With unifying receiver it works fine.
Comment 44 Marc Collin 2017-02-09 20:57:15 UTC
when my mouse stop to work, i see
reset full-speed USB device number 3 using ehci-pci
Comment 45 cyberwolf 2017-02-11 15:24:56 UTC
same issue with intel 8265
Comment 46 Marc Collin 2017-02-11 22:59:37 UTC
tried to disable auto suspend of the usb port... same problem
Comment 47 Tony Houghton 2017-02-18 20:43:16 UTC
I've got a similar problem with Debian kernel 4.9.6-3, except my mouse, a Microsoft 3600, just seems to stop, without going into a low framerate first. However, it did seem to go into a low framerate when I tried to see if my BT speaker would work. The speaker would connect, but either not become available in sound settings, or attempting to play to it caused the player (mpv) to just pause at 00:00 indefinitely.

When trying to open GNOME's bluetooth settings when the mouse has stopped working, there's a delay of several seconds before it appears, and I see this in journalctl:

Feb 18 19:29:31 lenny bluetoothd[1140]: Failed to set mode: Failed (0x03)
Feb 18 19:29:31 lenny kernel: Bluetooth: hci0 command 0x0c1a tx timeout
Feb 18 19:29:33 lenny bluetoothd[1140]: Failed to set mode: Failed (0x03)
Feb 18 19:29:33 lenny gnome-control-c[1824]: Failed to set Discoverable to 1: GDBus.Error:org.bluez.Error.Failed: Failed
Feb 18 19:29:33 lenny kernel: Bluetooth: hci0 command 0x0c1a tx timeout
Feb 18 19:29:35 lenny kernel: Bluetooth: hci0 command 0x200c tx timeout
Feb 18 19:29:37 lenny bluetoothd[1140]: Failed to set mode: Failed (0x03)
Feb 18 19:29:37 lenny kernel: Bluetooth: hci0 command 0x0c1a tx timeout
Feb 18 19:29:39 lenny bluetoothd[1140]: Failed to set mode: Failed (0x03)
Feb 18 19:29:39 lenny kernel: Bluetooth: hci0 command 0x0c1a tx timeout
Feb 18 19:29:47 lenny kernel: Bluetooth: hci0 command 0x200c tx timeout
Feb 18 19:29:59 lenny kernel: Bluetooth: hci0 command 0x200c tx timeout
Feb 18 19:30:11 lenny kernel: Bluetooth: hci0 command 0x200c tx timeout
F

When the settings dialog eventually opens, it says the mouse is still Connected. Turning bluetooth off and on again makes the mouse reconnect, but only for a few seconds.
Comment 48 ahxcker 2017-02-23 16:28:46 UTC
I got same problem i checked some wikis and forums tried many ways like: disable NetworkManager, bluetoothctl pair device

Lenovo Thinkpad T460p
debian stretch (i used Arch Linux with latest kernel 2017-02-15, same issue)
kernel 4.9
Plasma for desktop env 
Intel dual band wireless 8260

Checked dmesg:
[ 656.245284] Bluetooth: hci0: Bootloader revision 0.0 build 2 week 52 2014
[ 656.252273] Bluetooth: hci0: Device revision is 5
[ 656.252274] Bluetooth: hci0: Secure boot is enabled
[ 656.252275] Bluetooth: hci0: OTP lock is enabled
[ 656.252275] Bluetooth: hci0: API lock is enabled
[ 656.252276] Bluetooth: hci0: Debug lock is disabled
[ 656.252276] Bluetooth: hci0: Minimum firmware build 1 week 10 2014
[ 656.252510] Bluetooth: hci0: Found device firmware: intel/ibt-11-5.sfi
[ 657.920245] Bluetooth: hci0: Waiting for firmware download to complete
[ 657.920247] Bluetooth: hci0: Firmware loaded in 1636664 usecs
[ 657.920294] Bluetooth: hci0: Waiting for device to boot
[ 657.931249] Bluetooth: hci0: Device booted in 10706 usecs
[ 657.931276] Bluetooth: hci0: Found Intel DDC parameters: intel/ibt-11-5.ddc
[ 657.932254] Bluetooth: hci0: Failed to send Intel_Write_DDC (-22)

First connection will keep for few minutes after it will lost connection.
When you check plasma Bluetooth-applet, it shows device is connected however, when you try to disconnect device, it will keep showing connected unless you disable Bluetooth,
It will automatically connect your mouse when you enable Bluetooth.
Comment 49 Tony Houghton 2017-02-23 16:39:58 UTC
FWIW... I still had the Debian package firmware-iwlwifi version 20160824-1 available in my cache so I installed that. With this version the situation is much improved, but still not perfect. Every so often (something like a couple of minutes) the mouse stops working for a few seconds, then the bluetooth icon in the status bar briefly flashes off, then the mouse connection is restored.
Comment 50 Emmanuel Grumbach 2017-02-24 08:47:41 UTC
OK, we have a wifi BT coex issue here. Taking this bug to wifi.

Pease be more precise with wifi firmware versions. Are we talking about -22.ucode versus -19.ucode?
Comment 51 Stefan Schmid 2017-02-24 09:08:58 UTC
(In reply to Emmanuel Grumbach from comment #50)
> Pease be more precise with wifi firmware versions. Are we talking about
> -22.ucode versus -19.ucode?

TERRA MOBILE 1542 i7-6700T
Linux 4.10.0-gentoo (at the Moment, 28.02.2017 Europe/Zurich)
pcie: Intel Corporation Wireless 8260 [8086:24f3] (rev 3a)
usb: ID 8087:0a2b Intel Corp.

md5sum 14428864561a970cac91b355e8cd9b61  /lib/firmware/iwlwifi-8000C-22.ucode
md5sum 873629534cf972411618c6c13aef80f2  /lib/firmware/intel/ibt-11-5.ddc
md5sum 9bddc765acb16165ae970e7b588b5ea3  /lib/firmware/intel/ibt-11-5.sfi

WLAN and Bluetooth are stable, but Bluetooth HIDs are sometimes a bit laggy.
Comment 52 Emmanuel Grumbach 2017-02-24 10:56:35 UTC
@Stefan:
Thanks, but then, no bug on this combo of Wi-Fi and BT firmware versions?
This bug seems to be about BT not working at all.
Comment 53 Stefan Schmid 2017-02-24 11:07:58 UTC
(In reply to Emmanuel Grumbach from comment #52)
> Thanks, but then, no bug on this combo of Wi-Fi and BT firmware versions?
> This bug seems to be about BT not working at all.

I'm not sure why the mouse is not working properly. I can only show you my kernel log:
https://drive.google.com/file/d/0B3pmY9R_R3jqWmtQb2szQ3V0dFU/view?usp=sharing
Comment 54 Tony Houghton 2017-02-24 12:12:58 UTC
With the older package (20160824-1) I described in comment 49:

14428864561a970cac91b355e8cd9b61  /lib/firmware/iwlwifi-8000C-22.ucode
2010a4f3942b6df95ed928f92db3d000  /lib/firmware/intel/ibt-11-5.ddc
2f06625d1cab8e6edb339dc7a0db4e91  /lib/firmware/intel/ibt-11-5.sfi

With Debian latest (20161130-2), as described in comment 47:

14428864561a970cac91b355e8cd9b61  /lib/firmware/iwlwifi-8000C-22.ucode
42ca857a390890dda668af3670367806  /lib/firmware/intel/ibt-11-5.ddc
413da432ba53076ed31a1dda17a1f43d  /lib/firmware/intel/ibt-11-5.sfi
Comment 55 Emmanuel Grumbach 2017-02-24 13:31:51 UTC
@Tony,

Oh, OK. So clearly, this a regression in  BT firmware. I am from the Wifi team. So I can't do much for this.

I'll send instructions on Sunday on how to collect debug data to try to see what happens.
Comment 56 Marc Collin 2017-02-25 02:33:41 UTC
i read on the web then if the people don't use the wifi.. bluetooth work fine...

when i use ethernet cable instead of wifi... bluetooth work alway fine... 
i got issue because is stoped to use an ethernet cable...


so i don't know if it's only a bluetooth problem, i think there is also a wifi issue.
Comment 57 ahxcker 2017-02-25 02:51:06 UTC
Created attachment 254911 [details]
attachment-32063-0.html

You can try disable NetworkManager to see if Bluetooth work.

On Fri, Feb 24, 2017, 9:33 PM <bugzilla-daemon@bugzilla.kernel.org> wrote:

> https://bugzilla.kernel.org/show_bug.cgi?id=190831
>
> --- Comment #56 from Marc Collin (collinm@laboiteaprog.com) ---
> i read on the web then if the people don't use the wifi.. bluetooth work
> fine...
>
> when i use ethernet cable instead of wifi... bluetooth work alway fine...
> i got issue because is stoped to use an ethernet cable...
>
>
> so i don't know if it's only a bluetooth problem, i think there is also a
> wifi
> issue.
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
Comment 58 Marcin Kurek 2017-02-25 07:28:22 UTC
In my case Bluetooth mouse works smooth as long as WiFi is not connected. When both Bluetooth and WiFi are used:

- With stock BT firmware. Mouse stops responding completely after few seconds or in best case a minute or so. Reloading btusb module fixes it for next few seconds.
- With recent BT firmware post here it works most of the time, but mouse is a bit jumpy (still better than no mouse at all)

Maybe it's worth to mention in my case I use BT-LE mouse.
Comment 59 Emmanuel Grumbach 2017-02-26 10:31:55 UTC
Created attachment 254925 [details]
Core24 with uSniffer

Hi all,

I attached a WiFi firmware with debug enabled. To use it, you'll first need to make sure your WiFi driver is new enough to load -27.ucode. To get our latest WiFi driver, please take the Core24 driver (or the master branch) from:
https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi/core_release

Then, please reproduce the problem and while the problem is happening, create a debug dump:

https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi/debugging#firmware_debugging

You'll need the 'echo 1 > /sys/kernel/debug/iwlwifi/0000\:0X\:00.0/iwlmvm/fw_dbg_collect' command mentioned there.

Please take the time to read our privacy notice: https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi/debugging#privacy_aspects

Thank you.
Comment 60 Emmanuel Grumbach 2017-02-26 10:34:17 UTC
I forgot to mention: when you provide the data, please open another bug so that it will be easier to track each user. Make sure to CC linuxwifi@intel.com to the bug you'll create - thanks.
Comment 61 Emmanuel Grumbach 2017-02-28 16:41:02 UTC
Anyone? :)
Comment 62 Gurenko Alex 2017-02-28 17:01:38 UTC
(In reply to Emmanuel Grumbach from comment #61)
> Anyone? :)

Happy to help, but I keep forgetting my BT keyboard at home and I have problem with my work laptop. Also I might need some help with WiFi firmware installation. Are you available on IRC or something?
Comment 63 Tony Houghton 2017-02-28 17:18:33 UTC
When I tried to compile backport-iwlwifi I got a load of errors about config_enabled(), which was apparently removed from kernel 4.9: <https://patchwork.kernel.org/patch/9378087/>. Is there another branch or something for the driver that will compile on kernel 4.9, or a simple fix for this one?
Comment 64 Emmanuel Grumbach 2017-02-28 17:32:02 UTC
(In reply to Gurenko Alex from comment #62)
> (In reply to Emmanuel Grumbach from comment #61)
> > Anyone? :)
> 
> Happy to help, but I keep forgetting my BT keyboard at home and I have
> problem with my work laptop. Also I might need some help with WiFi firmware
> installation. Are you available on IRC or something?

My wife forbids IRC (well, no. She doesn't. But you got my point :))
Installing the WiFi firmware is as easy as installing the BT FW.
Just drop the file under /lib/firwmare (and *not* /lib/firmware/intel/ which is the directory BT looks at), and then reload iwlwifi module (or reboot).

(In reply to Tony Houghton from comment #63)
> When I tried to compile backport-iwlwifi I got a load of errors about
> config_enabled(), which was apparently removed from kernel 4.9:
> <https://patchwork.kernel.org/patch/9378087/>. Is there another branch or
> something for the driver that will compile on kernel 4.9, or a simple fix
> for this one?

Can you please take the master branch of the backport project? I want to believe that the push I did earlier fixed that.
Comment 65 Gurenko Alex 2017-03-01 05:54:13 UTC
(In reply to Emmanuel Grumbach from comment #64)
> (In reply to Gurenko Alex from comment #62)
> > (In reply to Emmanuel Grumbach from comment #61)
> > > Anyone? :)
> > 
> > Happy to help, but I keep forgetting my BT keyboard at home and I have
> > problem with my work laptop. Also I might need some help with WiFi firmware
> > installation. Are you available on IRC or something?
> 
> My wife forbids IRC (well, no. She doesn't. But you got my point :))
> Installing the WiFi firmware is as easy as installing the BT FW.
> Just drop the file under /lib/firwmare (and *not* /lib/firmware/intel/ which
> is the directory BT looks at), and then reload iwlwifi module (or reboot).

 Right, I have not looked into that at all and I forgot keyboard at home again (sry, not enough sleep lately). I'll try it tonight and let you know
Comment 66 Stefan Schmid 2017-03-01 11:28:30 UTC
Created attachment 255005 [details]
Debug-Log from iwlwifi on Kernel 4.10.1

(In reply to Emmanuel Grumbach from comment #59)
> Created attachment 254925 [details]
> Core24 with uSniffer

(In reply to Emmanuel Grumbach from comment #61)
> Anyone? :)

This firmware is not loaded by my kernel, but I was able to activate and collect the debug output with Kernel 4.10.1 and my setup (comment #51).
But with this setup, it does not happen anymore.
Comment 67 Emmanuel Grumbach 2017-03-01 11:32:46 UTC
@Stefan: in the log you attached, you still load -22.ucode.
Comment 68 Tony Houghton 2017-03-01 14:53:04 UTC
(In reply to Emmanuel Grumbach from comment #64)

> Can you please take the master branch of the backport project? I want to
> believe that the push I did earlier fixed that.

Yes, that works, thanks. But it works too well, my mouse is now behaving perfectly! I used the udev rule for creating dump logs automatically, and so far none have appeared. I take it there would be no point in creating one manually while it's behaving correctly?

If you want to get to the bottom of this to make sure it doesn't regress again, and you can make a debugging version of -22.ucode I'll be happy to test that for you.
Comment 69 Emmanuel Grumbach 2017-03-01 14:59:32 UTC
If you are happy, I won't do anything on -22.ucode.
No point to create manual dump.

Anyone else before I close the bug?
Comment 70 Gurenko Alex 2017-03-01 15:13:38 UTC
(In reply to Emmanuel Grumbach from comment #69)
> If you are happy, I won't do anything on -22.ucode.
> No point to create manual dump.
> 
> Anyone else before I close the bug?

Can you please leave it open? I'll be able to test it in ~2-3 hours and I don't want to be left alone if the problem persists for me :)
Comment 71 Emmanuel Grumbach 2017-03-01 15:32:01 UTC
No worries. I am getting notifications on closed bugs anyway. But I'll wait until Sunday or so.
Comment 72 ahxcker 2017-03-01 15:34:50 UTC
Created attachment 255011 [details]
attachment-14468-0.html

Thanks for work in on our issue.

On Wed, Mar 1, 2017, 10:32 AM <bugzilla-daemon@bugzilla.kernel.org> wrote:

> https://bugzilla.kernel.org/show_bug.cgi?id=190831
>
> --- Comment #71 from Emmanuel Grumbach (emmanuel.grumbach@intel.com) ---
> No worries. I am getting notifications on closed bugs anyway. But I'll wait
> until Sunday or so.
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
Comment 73 Gurenko Alex 2017-03-01 19:00:00 UTC
(In reply to Emmanuel Grumbach from comment #71)
> No worries. I am getting notifications on closed bugs anyway. But I'll wait
> until Sunday or so.

So, I gave it a go, and assuming new firmware was used (couldn't see it anywhere). It...didn't work for me...

Do I understand correctly, that I only needed to copy iwlwifi-8000C-27.ucode to /lib/firmware and reboot, nothing else? It should be picked up automatically?

I've echoed 1 to the /sys/... so what kind of logs you want me to grab?

It's interesting that disabling WiFi keeps keyboard working, but enabling Wifi kills BT connection instantly.
Comment 74 Emmanuel Grumbach 2017-03-01 19:07:05 UTC
please attach the dmesg output.
Comment 75 Gurenko Alex 2017-03-01 19:10:48 UTC
Created attachment 255019 [details]
agurenko-dmesg.log

Here is my dmesg.log
Comment 76 Emmanuel Grumbach 2017-03-01 19:16:12 UTC
iwlwifi 0000:03:00.0: loaded firmware version 22.361476.0 op_mode iwlmvm

You are still using the old firmware.

As mentioned in comment 59, you also need to install the backport driver.
Check https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi/core_release#how_to_install_the_driver

But of course, don't take Core14, but rather the master branch:

git clone https://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/backport-iwlwifi.git


You can also check the firmware version with:

ethtool -i wlp3s0 | grep firmware
Comment 77 cyberwolf 2017-03-01 19:30:06 UTC
Created attachment 255021 [details]
new driver from master branch, firmware 22 and 27 still not working


I installed the new driver from master branch:

	git clone https://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/backport-iwlwifi.git -b master

	cd backport-iwlwifi

	make defconfig-iwlwifi-public
	sed -i 's/CPTCFG_IWLMVM_VENDOR_CMDS=y/# CPTCFG_IWLMVM_VENDOR_CMDS is not set/' .config
	make -j4

	make install

Tried with both iwlwifi-8000C-22.ucode and iwlwifi-8000C-27.ucode
The situation looks the same, BT is working as long as wifi is disconnected then it stop immediately.
Attached dmesg with both firmware
Comment 78 Gurenko Alex 2017-03-01 19:30:48 UTC
Created attachment 255023 [details]
agurenko-dmesg.log with -27 firmware

So, yeah, looks like that didn't work either and behaviour was really weird, some letters on a keyboard were working fine, other didn't at all, all hot keys worked though.
Comment 79 Emmanuel Grumbach 2017-03-01 19:32:23 UTC
@cyberwolf: ok - can you please record the data required as explain comment 59?
Comment 80 Emmanuel Grumbach 2017-03-01 19:37:44 UTC
Created attachment 255025 [details]
Core24 with uSniffer

Please use this firmware.
Somehow the debug wasn't enabled on the previous firmware. I hope this time it will work. I just don't have a 8260 to test right now.
Comment 81 Gurenko Alex 2017-03-01 19:54:04 UTC
Created attachment 255031 [details]
agurenko-dmesg.log latest

(In reply to Emmanuel Grumbach from comment #80)
> Created attachment 255025 [details]
> Core24 with uSniffer
> 
> Please use this firmware.
> Somehow the debug wasn't enabled on the previous firmware. I hope this time
> it will work. I just don't have a 8260 to test right now.

so, I've tried this file too and it did reconnect on the first pause (I give it 30 seconds), but hang on a second try
Comment 82 cyberwolf 2017-03-01 21:08:28 UTC
@Emmanuel I have a problem with gpg encryption:

:~# gpg -r 18D075865D915B5E386F660F2D0B96FE6E363201 --encrypt iwl.dump 
gpg: 18D075865D915B5E386F660F2D0B96FE6E363201: skipped: No public key
gpg: iwl.dump: encryption failed: No public key
Comment 83 Emmanuel Grumbach 2017-03-02 06:34:28 UTC
@cyberwolf
gpg --recv-keys 18D075865D915B5E386F660F2D0B96FE6E363201
Then you should be able to use it. Please open a new bug when you have the data.

@Gurenka Alex - I saw that the debugging started and that you triggered a dump. Very good. Now, please open a new bug, describe exactly the scenario (What BT profile you use e.g.) and attach the dump. Let me remind you to read the privacy notice.

Please remember to CC linuxwifi@intel.com to the bugs you open.
Comment 84 Gurenko Alex 2017-03-02 07:17:14 UTC
(In reply to Emmanuel Grumbach from comment #83)
> @Gurenka Alex - I saw that the debugging started and that you triggered a
> dump. Very good. Now, please open a new bug, describe exactly the scenario
> (What BT profile you use e.g.) and attach the dump. Let me remind you to
> read the privacy notice.
> 
> Please remember to CC linuxwifi@intel.com to the bugs you open.

Thank you, I've created a bug here: https://bugzilla.kernel.org/show_bug.cgi?id=194759
Comment 85 Emmanuel Grumbach 2017-03-07 17:10:51 UTC
I'd like to close this bug which is too generic. For now, only Alex opened another bug with the required data. Anyone else?
Comment 86 ahxcker 2017-03-07 18:56:59 UTC
(In reply to cyberwolf from comment #77)
> Created attachment 255021 [details]
> new driver from master branch, firmware 22 and 27 still not working
> 
> 
> I installed the new driver from master branch:
> 
>       git clone
> https://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/backport-iwlwifi.git
> -b master
> 
>       cd backport-iwlwifi
> 
>       make defconfig-iwlwifi-public
>       sed -i 's/CPTCFG_IWLMVM_VENDOR_CMDS=y/# CPTCFG_IWLMVM_VENDOR_CMDS is
> not
> set/' .config
>       make -j4
> 
>       make install
> 
> Tried with both iwlwifi-8000C-22.ucode and iwlwifi-8000C-27.ucode
> The situation looks the same, BT is working as long as wifi is disconnected
> then it stop immediately.
> Attached dmesg with both firmware

I have to use sudo make install because it is not working when I type make install.

I did something with you and also I have not fixed my problem yet. after installing latest driver my firmware version becomes to 27.455470.0
I think it is the latest version.

(In reply to Emmanuel Grumbach from comment #85)
> I'd like to close this bug which is too generic. For now, only Alex opened
> another bug with the required data. Anyone else?

Are we done here? 

I followed the link below:
https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi/debugging#firmware_debugging

set: 
  CONFIG_ALLOW_DEV_COREDUMP to Y. 

try to 
  echo 1 > /sys/kernel/debug/iwlwifi/0000\:0X\:00.0/iwlmvm/fw_dbg_collect

got an error code:
  bash: /sys/kernel/debug/iwlwifi/0000:0X:00.0/iwlmvm/fw_dbg_collect: No such file or directory

tried to touch a file
  touch: cannot touch '/sys/kernel/debug/iwlwifi/0000:0X:00.0/iwlmvm/fw_dbg_collect': No such file or directory

tried to use vim but doesn't allow me to write
Comment 87 Emmanuel Grumbach 2017-03-07 19:09:34 UTC
You need to replace X in the command with the right number in your system as explained in the wiki
Comment 88 ahxcker 2017-03-07 19:43:47 UTC
(In reply to Emmanuel Grumbach from comment #87)
> You need to replace X in the command with the right number in your system as
> explained in the wiki

thanks for help

https://bugzilla.kernel.org/show_bug.cgi?id=194815
this is my report and cc to linuxwifi@intel.com
Comment 89 Emmanuel Grumbach 2017-03-08 07:10:16 UTC
OK - close this now for now.

Note You need to log in before you can comment on or make changes to this bug.