Bug 1169309 - Tumbleweed Intel 7260 dual-band wifi/bluetooth
Tumbleweed Intel 7260 dual-band wifi/bluetooth
Status: NEW
Classification: openSUSE
Product: openSUSE Tumbleweed
Classification: openSUSE
Component: Kernel
Current
Macintosh SUSE Other
: P5 - None : Normal (vote)
: ---
Assigned To: openSUSE Kernel Bugs
E-mail List
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2020-04-13 15:12 UTC by Paul Cee
Modified: 2021-12-31 14:20 UTC (History)
4 users (show)

See Also:
Found By: ---
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---
jslaby: needinfo? (paulcee33)


Attachments
hwinfo output from 27" iMac (758.31 KB, text/plain)
2020-04-14 16:59 UTC, Paul Cee
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Paul Cee 2020-04-13 15:12:08 UTC
User-Agent:       Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/81.0.4044.92 Safari/537.36
Build Identifier: 

Prior to update in February 2020, my Intel 7260 PCIe card supplied both bluetooth and wifi on my iMac. Both the internal Apple (Broadcom based, I believe) bluetooth card and Intel would show up (hci0 and 1 respectively). After an update, the hci1 adapter vanished, and the 7260 bluetooth capabilities are gone. Wifi functions perfectly.

Reproducible: Always

Steps to Reproduce:
1.Boot system
2.Look at bluetooth adapters
3.Adapter hci1 not found

Dmesg and system logs do not show the "Intel read..." and even ATTEMPT to load bluetooth firmware. 
Actual Results:  
No changes after several updates.

Expected Results:  
Bluetooth adapter would have shown back up and functioned

See thread on suse forums.

https://forums.opensuse.org/showthread.php/539218-Tumbleweed-No-bluetooth-after-update/page2
Comment 1 Jiri Slaby 2020-04-14 08:55:51 UTC
5.5 was accepted to factory on 7th Feb. Could you try 5.4 kernel? Or 5.3 from leap?

https://download.opensuse.org/repositories/Kernel:/openSUSE-15.2/standard/
Comment 2 Paul Cee 2020-04-14 13:26:14 UTC
I'm using Tumbleweed, not Leap.  Will download/burn to USB, and see what happens, though.
Comment 3 Takashi Iwai 2020-04-14 13:34:58 UTC
In anyway, it'd be helpful to give the dmesg output and hwinfo output from the latest (non-working) 5.6.y kernel.
Comment 4 Paul Cee 2020-04-14 16:59:00 UTC
Created attachment 835745 [details]
hwinfo output from 27" iMac

Requested hwinfo data.
Comment 5 Paul Cee 2020-04-14 17:00:16 UTC
dmesg | egrep -i "blue|firm|intel|read"    
[    0.152796] process: using mwait in idle threads
[    0.152804] Spectre V2 : Enabling Restricted Speculation for firmware calls
[    0.262637] smpboot: CPU0: Intel(R) Core(TM) i7 CPU         860  @ 2.80GHz (family: 0x6, model: 0x1e, stepping: 0x5)
[    0.262746] Performance Events: PEBS fmt1+, Nehalem events, 16-deep LBR, Intel PMU driver.
[    0.328180] ACPI: [Firmware Bug]: BIOS _OSI(Linux) query ignored
[    1.711778] intel_idle: MWAIT substates: 0x1120
[    1.711779] intel_idle: v0.4.1 model 0x1E
[    1.712262] intel_idle: lapic_timer_reliable_states 0x2
[    1.727999] intel_pstate: CPU model not supported
[    2.041384] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[    2.108246] Write protecting the kernel read-only data: 20480k
[    2.457506] random: systemd: uninitialized urandom read (16 bytes read)
[    2.459321] random: systemd: uninitialized urandom read (16 bytes read)
[    2.459376] random: systemd: uninitialized urandom read (16 bytes read)
[    3.632989] [drm] radeon: 512M of VRAM memory ready
[    3.632990] [drm] radeon: 1024M of GTT memory ready.
[    3.687193] [drm] PCIE gen 2 link speeds already enabled
[    3.934129] Btrfs loaded, crc32c=crc32c-intel, assert=on
[    4.031153] usb 1-1.2: Product: Card Reader
[    4.335075] usb 1-1.1.1: Product: Bluetooth USB Host Controller
[    5.694469] scsi 6:0:0:0: Direct-Access     APPLE    SD Card Reader   1.00 PQ: 0 ANSI: 0
[   10.084160] intel ips 0000:00:1f.6: Non-IPS CPU detected.
[   10.084163] intel ips 0000:00:1f.6: IPS not supported on this CPU
[   10.198318] snd_hda_intel 0000:00:1b.0: enabling device (0000 -> 0002)
[   10.198745] snd_hda_intel 0000:01:00.1: enabling device (0000 -> 0002)
[   10.243545] Intel(R) Wireless WiFi driver for Linux
[   10.243546] Copyright(c) 2003- 2015 Intel Corporation
[   10.305305] iwlwifi 0000:03:00.0: loaded firmware version 17.3216344376.0 7260-17.ucode op_mode iwlmvm
[   10.427068] input: HDA Intel MID Line as /devices/pci0000:00/0000:00:1b.0/sound/card0/input10
[   10.458218] input: HDA Intel MID Headphone as /devices/pci0000:00/0000:00:1b.0/sound/card0/input11
[   10.458276] input: HDA Intel MID SPDIF In as /devices/pci0000:00/0000:00:1b.0/sound/card0/input12
[   10.491526] Bluetooth: Core ver 2.22
[   10.491543] Bluetooth: HCI device and connection manager initialized
[   10.491547] Bluetooth: HCI socket layer initialized
[   10.491549] Bluetooth: L2CAP socket layer initialized
[   10.491552] Bluetooth: SCO socket layer initialized
[   10.549474] iTCO_wdt: Intel TCO WatchDog Timer Driver v1.11
[   10.549539] iwlwifi 0000:03:00.0: Detected Intel(R) Dual Band Wireless AC 7260, REV=0x144
[   10.691009] Bluetooth: hci0: BCM: chip id 254 build 0518
[   10.692010] Bluetooth: hci0: BCM: product 05ac:8215
[   10.693002] Bluetooth: hci0: BCM: features 0x00
[   10.710016] Bluetooth: hci0: Bluetooth USB Host Controller
[   12.529335] Bluetooth: BNEP (Ethernet Emulation) ver 1.3
[   12.529336] Bluetooth: BNEP filters: protocol multicast
[   12.529341] Bluetooth: BNEP socket layer initialized
[   19.825532] Bluetooth: RFCOMM TTY layer initialized
[   19.825545] Bluetooth: RFCOMM socket layer initialized
[   19.825551] Bluetooth: RFCOMM ver 1.11
[   21.257053] IPv6: ADDRCONF(NETDEV_CHANGE): wlp3s0: link becomes ready
[  193.834984] Bluetooth: HIDP (Human Interface Emulation) ver 1.2
[  193.834989] Bluetooth: HIDP socket layer initialized


Let me know if more dmesg output is needed...tried to include relevant details
Comment 6 Jiri Slaby 2020-04-15 06:06:09 UTC
(In reply to Paul Cee from comment #2)
> I'm using Tumbleweed, not Leap.  Will download/burn to USB, and see what
> happens, though.

I asked to use kernel from Leap, not using Leap.
Comment 7 Paul Cee 2020-04-20 19:40:44 UTC
Difference being what? And how would I go about installing an old kernel for Leap on my current, up-to-date Tumbleweed?
Comment 8 Takashi Iwai 2020-04-20 20:00:58 UTC
You can install the older kernel on the TW system.  It usually works fine.

Simply grab kernel-default-5.3*.rpm from OBS download URL,
  http://download.opensuse.org/distribution/leap/15.2/repo/oss/x86_64/

then install it via zypper
  zypper in --oldpackage kernel-default-5.3*.rpm

Then boot with that kernel and see whether the problem still happens.

The difference between testing Leap system and testing Leap kernel-on-the-TW is that, by installing only the old kernel on TW system, we can conclude whether it's a *pure* kernel problem or not.  That is, if the problem appears / solved by booting a different kernel while keeping the rest, it assures that the cause is the kernel package.  OTOH, if we switch the whole system, it might be another culprit such as kernel-firmware package, the user-space network stack, systemd or whatever else.
Comment 9 Paul Cee 2020-04-21 14:14:14 UTC
That makes perfect sense. Installed and tried kernel-default-5.3.18-lp152.10.4.x86_64.rpm, and no joy. Adapter still doesn't show up. My thread on the openSUSE forums has an interesting (to me) tidbit. That is, Malcom has the same adapter as I do (Intel 7260), yet his dmesg output shows an "Intel Read...", then firmware gets loaded. I do NOT see that line in my output.
Comment 10 Takashi Iwai 2020-04-21 14:57:13 UTC
OK, thanks.  Could you give dmesg output from both 5.3 and TW kernels?  Even though it doesn't work, it can be a good hint.

Meanwhile, we can try to downgrade the kernel-firmware package.  Grab kernel-firmware*.rpm from
  http://download.opensuse.org/distribution/leap/15.2/repo/oss/noarch/

Unlike TW package, it's a big single package.  You can install it (forcibly) via rpm
  rpm -ivh kernel-firmware*.rpm --nodeps --force

This raw firmware file should have precedence over the split kernel-firmware-xxx packages of TW, so just try to boot with 5.3 kernel after this installation.
Comment 11 Paul Cee 2020-04-21 17:15:19 UTC
Done...loaded the firmware as directed, and booted with both the newly updated (2020-04-21) kernel in Tumbleweed, and with the 5.3 leap kernel/firmware.

Output of "dmesg | egrep -i "blue|firm|intel" for both.  

Leap
----
[    0.329480] Spectre V2 : Enabling Restricted Speculation for firmware calls
[    0.439295] smpboot: CPU0: Intel(R) Core(TM) i7 CPU         860  @ 2.80GHz (family: 0x6, model: 0x1e, stepping: 0x5)
[    0.439406] Performance Events: PEBS fmt1+, Nehalem events, 16-deep LBR, Intel PMU driver.
[    0.501794] ACPI: [Firmware Bug]: BIOS _OSI(Linux) query ignored
[    2.156937] intel_idle: MWAIT substates: 0x1120
[    2.156938] intel_idle: v0.4.1 model 0x1E
[    2.157349] intel_idle: lapic_timer_reliable_states 0x2
[    2.180156] intel_pstate: CPU model not supported
[    4.188382] Btrfs loaded, crc32c=crc32c-intel, assert=on
[    4.764737] usb 1-1.1.1: Product: Bluetooth USB Host Controller
[   10.369237] intel ips 0000:00:1f.6: Non-IPS CPU detected.
[   10.369239] intel ips 0000:00:1f.6: IPS not supported on this CPU
[   10.572609] Bluetooth: Core ver 2.22
[   10.572627] Bluetooth: HCI device and connection manager initialized
[   10.572631] Bluetooth: HCI socket layer initialized
[   10.572633] Bluetooth: L2CAP socket layer initialized
[   10.572636] Bluetooth: SCO socket layer initialized
[   10.635475] Intel(R) Wireless WiFi driver for Linux
[   10.635476] Copyright(c) 2003- 2015 Intel Corporation
[   10.644343] iwlwifi 0000:03:00.0: loaded firmware version 17.3216344376.0 op_mode iwlmvm
[   10.669383] iTCO_wdt: Intel TCO WatchDog Timer Driver v1.11
[   10.688796] snd_hda_intel 0000:00:1b.0: enabling device (0000 -> 0002)
[   10.689211] snd_hda_intel 0000:01:00.1: enabling device (0000 -> 0002)
[   10.812707] Bluetooth: hci0: BCM: chip id 254 build 0518
[   10.813715] Bluetooth: hci0: BCM: product 05ac:8215
[   10.814702] Bluetooth: hci0: BCM: features 0x00
[   10.832227] Bluetooth: hci0: Bluetooth USB Host Controller
[   10.879719] iwlwifi 0000:03:00.0: Detected Intel(R) Dual Band Wireless AC 7260, REV=0x144
[   10.894954] input: HDA Intel MID Line as /devices/pci0000:00/0000:00:1b.0/sound/card0/input11
[   10.925972] input: HDA Intel MID Headphone as /devices/pci0000:00/0000:00:1b.0/sound/card0/input12
[   10.926440] input: HDA Intel MID SPDIF In as /devices/pci0000:00/0000:00:1b.0/sound/card0/input13
[   12.664124] Bluetooth: BNEP (Ethernet Emulation) ver 1.3
[   12.664126] Bluetooth: BNEP filters: protocol multicast
[   12.664129] Bluetooth: BNEP socket layer initialized
[   50.721525] Bluetooth: RFCOMM TTY layer initialized
[   50.721531] Bluetooth: RFCOMM socket layer initialized
[   50.721537] Bluetooth: RFCOMM ver 1.11
[   57.927980] Bluetooth: HIDP (Human Interface Emulation) ver 1.2
[   57.927984] Bluetooth: HIDP socket layer initialized



Tumbleweed
----------[    0.150333] Spectre V2 : Enabling Restricted Speculation for firmware calls
[    0.260050] smpboot: CPU0: Intel(R) Core(TM) i7 CPU         860  @ 2.80GHz (family: 0x6, model: 0x1e, stepping: 0x5)
[    0.260153] Performance Events: PEBS fmt1+, Nehalem events, 16-deep LBR, Intel PMU driver.
[    0.322374] ACPI: [Firmware Bug]: BIOS _OSI(Linux) query ignored
[    1.709618] intel_idle: MWAIT substates: 0x1120
[    1.709619] intel_idle: v0.4.1 model 0x1E
[    1.709942] intel_idle: lapic_timer_reliable_states 0x2
[    1.733163] intel_pstate: CPU model not supported
[    3.713969] Btrfs loaded, crc32c=crc32c-intel, assert=on
[    4.261828] usb 1-1.1.1: Product: Bluetooth USB Host Controller
[    9.305787] intel ips 0000:00:1f.6: Non-IPS CPU detected.
[    9.305789] intel ips 0000:00:1f.6: IPS not supported on this CPU
[    9.349227] snd_hda_intel 0000:00:1b.0: enabling device (0000 -> 0002)
[    9.349588] snd_hda_intel 0000:01:00.1: enabling device (0000 -> 0002)
[    9.414512] Bluetooth: Core ver 2.22
[    9.414531] Bluetooth: HCI device and connection manager initialized
[    9.414535] Bluetooth: HCI socket layer initialized
[    9.414538] Bluetooth: L2CAP socket layer initialized
[    9.414542] Bluetooth: SCO socket layer initialized
[    9.465416] Intel(R) Wireless WiFi driver for Linux
[    9.465418] Copyright(c) 2003- 2015 Intel Corporation
[    9.497911] iwlwifi 0000:03:00.0: loaded firmware version 17.3216344376.0 7260-17.ucode op_mode iwlmvm
[    9.551906] input: HDA Intel MID Line as /devices/pci0000:00/0000:00:1b.0/sound/card0/input10
[    9.551967] input: HDA Intel MID Headphone as /devices/pci0000:00/0000:00:1b.0/sound/card0/input11
[    9.552024] input: HDA Intel MID SPDIF In as /devices/pci0000:00/0000:00:1b.0/sound/card0/input12
[    9.585303] iTCO_wdt: Intel TCO WatchDog Timer Driver v1.11
[    9.704805] Bluetooth: hci0: BCM: chip id 254 build 0518
[    9.717893] Bluetooth: hci0: BCM: product 05ac:8215
[    9.718882] Bluetooth: hci0: BCM: features 0x00
[    9.735900] Bluetooth: hci0: Bluetooth USB Host Controller
[    9.819906] iwlwifi 0000:03:00.0: Detected Intel(R) Dual Band Wireless AC 7260, REV=0x144
[   11.815976] Bluetooth: BNEP (Ethernet Emulation) ver 1.3
[   11.815977] Bluetooth: BNEP filters: protocol multicast
[   11.815980] Bluetooth: BNEP socket layer initialized
[   18.746034] Bluetooth: RFCOMM TTY layer initialized
[   18.746087] Bluetooth: RFCOMM socket layer initialized
[   18.746092] Bluetooth: RFCOMM ver 1.11
[   22.817245] Bluetooth: HIDP (Human Interface Emulation) ver 1.2
[   22.817250] Bluetooth: HIDP socket layer initialized
Comment 12 Paul Cee 2020-04-23 15:06:49 UTC
Any thoughts on perhaps trying an old version of kernel-firmware-iwlwifi? Can't seem to find an older one to load.
Comment 13 Takashi Iwai 2020-04-23 15:32:20 UTC
The old kernel-firmware.rpm from Leap 15.2 contains all firmware files, so installing it actually was enough.

Now looking at the kernel messages and also the hwinfo output you uploaded, the bluetooth devices are present.

So I wonder what actually tested and doesn't work any longer, more exactly?
Comment 14 Paul Cee 2020-04-23 17:46:00 UTC
The internal, built-in Apple bluetooth adapter loads and functions. The one on the Intel 7260 stopped. I used to have hci0 and hci1, and the 7260 was Bluetooth 4, so I used it to connect higher-speed things (speakers, etc.), while using the low-speed adapter for HID stuff. 

I realize that the kernel-firmware-iwlwifi package has all the versions in it, but thinking more along the lines of /lib/firmware/iwlwifi-7260-17.ucode.xz  It will automatically load the latest version, if I remember correctly, which in this case is the *17 file. Tried moving that file elsewhere, but then things complained about not being able to find it, and my wifi didn't work, until I moved it back and rebooted. 

Would love to try a previous version of that file, like the *10/etc., to see if it makes a difference.
Comment 15 Takashi Iwai 2020-04-23 17:53:23 UTC
(In reply to Paul Cee from comment #14)
> The internal, built-in Apple bluetooth adapter loads and functions. The one
> on the Intel 7260 stopped. I used to have hci0 and hci1, and the 7260 was
> Bluetooth 4, so I used it to connect higher-speed things (speakers, etc.),
> while using the low-speed adapter for HID stuff. 

OK thanks, now it's clear.

> I realize that the kernel-firmware-iwlwifi package has all the versions in
> it, but thinking more along the lines of
> /lib/firmware/iwlwifi-7260-17.ucode.xz  It will automatically load the
> latest version, if I remember correctly, which in this case is the *17 file.
> Tried moving that file elsewhere, but then things complained about not being
> able to find it, and my wifi didn't work, until I moved it back and
> rebooted. 

Oh right.  Simply uninstall kernel-firmware-* subpackages once when kernel-firmware.rpm is installed.

That said,
  rpm -ivh kernel-firmware.rpm --oldpackage --force
  rpm -e kernel-firmware-iwlwifi

BTW, did you try openSUSE Leap 15.2 Live image?  You can try it without installation.  If that works, you can get hwinfo output from there, as well as dmesg output.
Comment 16 Paul Cee 2020-04-27 13:17:55 UTC
(In reply to Takashi Iwai from comment #15)
> (In reply to Paul Cee from comment #14)
> > The internal, built-in Apple bluetooth adapter loads and functions. The one
> > on the Intel 7260 stopped. I used to have hci0 and hci1, and the 7260 was
> > Bluetooth 4, so I used it to connect higher-speed things (speakers, etc.),
> > while using the low-speed adapter for HID stuff. 
> 
> OK thanks, now it's clear.
> 
> > I realize that the kernel-firmware-iwlwifi package has all the versions in
> > it, but thinking more along the lines of
> > /lib/firmware/iwlwifi-7260-17.ucode.xz  It will automatically load the
> > latest version, if I remember correctly, which in this case is the *17 file.
> > Tried moving that file elsewhere, but then things complained about not being
> > able to find it, and my wifi didn't work, until I moved it back and
> > rebooted. 
> 
> Oh right.  Simply uninstall kernel-firmware-* subpackages once when
> kernel-firmware.rpm is installed.
> 
> That said,
>   rpm -ivh kernel-firmware.rpm --oldpackage --force
>   rpm -e kernel-firmware-iwlwifi

I will give that a shot as soon as I can, and advise. Thanks.

> BTW, did you try openSUSE Leap 15.2 Live image?  You can try it without
> installation.  If that works, you can get hwinfo output from there, as well
> as dmesg output.

Yep, no difference. Didn't capture the hwinfo, though, but did boot with the 15.2 image via USB. And I now KNOW that it is the Intel 7260, because I plugged in an Asus USB bluetooth adapter that I remembered I had laying around, and it came right up. Hci1 was there, and I could configure/use it. No problems.
Comment 17 Paul Cee 2020-04-30 17:24:53 UTC
I *THINK* I loaded the right kernel-firmware file, and rebooted...no luck. Not 100% sure that I grabbed the right file, though, so cannot be 100% sure of the results.
Comment 18 Takashi Iwai 2020-05-01 13:38:38 UTC
Unfortunately there is no proper way to see which firmware files have been already loaded.  You can add the boot option

  firmware_class.dyndbg=+p

then you'll see the debug messages from the firmware loader and see which files are accessed and loaded.
Comment 19 Miroslav Beneš 2020-09-02 10:46:54 UTC
Paul, any luck with Takashi's debugging proposal?

There is also 5.8.x kernel now in TW, which may work again, and more updates to TW as a whole, I am sure. Could you try that, please?
Comment 20 Paul Cee 2020-09-09 15:56:32 UTC
(In reply to Miroslav Beneš from comment #19)
> Paul, any luck with Takashi's debugging proposal?
> 
> There is also 5.8.x kernel now in TW, which may work again, and more updates
> to TW as a whole, I am sure. Could you try that, please?

Thanks, Miroslav, and everyone. I've been playing with this off and on for a while now, and I'm of the opinion that this is too much of an edge-case for Opensuse to deal with. I *THINK* the real problem is the Apple bluetooth hardware, and has nothing to do with the Intel adapter. Most folks probably aren't running on iMac hardware and probably don't have dual-bluetooth adapters to worry about.

I'm using the latest and greatest Tumbleweed (new SSD recently), with all updates applied. Still won't even see it. HOWEVER, I have put the same card into my Vaio laptop (again, running Tumbleweed), and it works fine...even get two bluetooth channels. 

Not sure this is worth pursuing, although I'm more than willing to offer up any dumps or info, if others want to pursue this.
Comment 21 Miroslav Beneš 2021-12-31 14:20:26 UTC
Some time has passed and there is much newer kernel in TW now? Has it started working just by chance, or is the issue still present?