Intel-8260 not init'd by intel-microcode=3.20170707.1~ubuntu16.04.0

Bug #1713532 reported by yarko
12
This bug affects 1 person
Affects Status Importance Assigned to Milestone
intel-microcode (Ubuntu)
Invalid
Undecided
Unassigned

Bug Description

I'm running:

LSB Version: core-9.20160110ubuntu0.2-amd64:core-9.20160110ubuntu0.2-noarch:printing-9.20160110ubuntu0.2-amd64:printing-9.20160110ubuntu0.2-noarch:security-9.20160110ubuntu0.2-amd64:security-9.20160110ubuntu0.2-noarch
Distributor ID: Ubuntu
Description: Ubuntu 16.04.3 LTS
Release: 16.04
Codename: xenial

On a NUC6i5SYH, with 32G RAM, and BIOS 0061

Recently, the following package update happened:
intel-microcode=3.20170707.1~ubuntu16.04.0

Upon reboot, bluetooth was unavailable / uninitialized.
Poking around w/ various tools, I think the insightful bit came from:
`service bluetooth status -l` which showed multiple lines of
`not enough free handles to register service` (see attached image).

Rolling back to:
`intel-microcode=3.20151106.1`

and rebooting restored bluetooth function (I use a bluetooth keyboard).

I did try (and leave) changing /etc/bluetooth/main.conf as follows:
`Autoenable=true`

It didn't help under the new intel-micorcode, and seems to have no noticable deleterious effect under the old intel-microcode.

Let me know if there's any other / more information I can provide to help.

Revision history for this message
yarko (yarkot) wrote :
Revision history for this message
Robie Basak (racb) wrote :

Thank you for the report. For reference, bug 1700373 was the SRU tracking bug for the intel-microcode update.

tags: added: regression-update
Revision history for this message
Dimitri John Ledkov (xnox) wrote :

This may require escalating back to Intel.

Revision history for this message
Henrique de Moraes Holschuh (hmh) wrote : Re: [Bug 1700373][Bug 1713532] regression on intel-microcode 20170707

On Tue, 29 Aug 2017, Robie Basak wrote:
> Regression report in bug 1713532. Please could someone familiar with
> these microcode updates take a look?

First, we should check for the possibility of a kernel or userspace
issue that is getting in the way of the bluetooth firmware loading when
a microcode update is available to be applied...

Hopefully, this should be easy enough to do: intel-microcode package
3.20151106.1 is reported to work, so it is just a matter of testing
successive intel-microcode packages, and looking at the full logs of the
attempts.

Please report the contents of /proc/cpuinfo with the latest
intel-microcode package that works, and /proc/cpuinfo with the package
that doesn't (3.20170707.1~ubuntu16.04.0).

Also, please report the output of "dmesg" or from "journalctl -k -b"
(i.e. the kernel logs) in the failure case, so that we can search for
some sort of kernel message by the bluetooth stack, etc.

You could test when the breakage happened using the Debian packages,
They should be 100% compatible with the Ubuntu OS, as far as I know.

Packages to test (in order, you can stop when the breakage disappears):

http://snapshot.debian.org/package/intel-microcode/3.20170707.1/
http://snapshot.debian.org/package/intel-microcode/3.20170511.1/
http://snapshot.debian.org/package/intel-microcode/3.20161104.1/
http://snapshot.debian.org/package/intel-microcode/3.20160607.2/

Thanks.

--
  Henrique Holschuh

Revision history for this message
Dave Chiluk (chiluk) wrote :

According to the bios changelog
https://downloadmirror.intel.com/26841/eng/SY_0061_ReleaseNotes.pdf

0061 of your bios should already include the microcode that is also available in the intel-microcode package. As such, on boot I'd expect a no-op when checking for newer microcode version. We definitely need to know microcode version your machine is reporting with both packages installed.

Is there any chance that linux-firmware got updated as well? I suspect linux-firmware to be the culprit over intel-microcode.

Revision history for this message
yarko (yarkot) wrote :

A couple of quick responses:

@chiluk - linux-firmware installed is showing as:

(apt list --installed output):

linux-firmware/xenial-updates,xenial-updates,xenial-security,xenial-security,now 1.157.11 all [installed,automatic]

Revision history for this message
Simon Déziel (sdeziel) wrote :

Could this be a duplicate of LP: #1705633? @yarko, could you please provide which kernel you used?

Revision history for this message
yarko (yarkot) wrote :

Good point, Simon:

I'm running: 4.4.0-92-generic

As I'm just getting ready to sit down to this, and see:

linux-image-generic/xenial-updates,xenial-security 4.4.0.93.98 amd64 [upgradable from: 4.4.0.92.97]

Perhaps I'll (?):

- upgrade kernel (and related files queued);
- upgrade to latest intel-microcode;

and go from there...

Revision history for this message
Dave Chiluk (chiluk) wrote :

@yarko, Please start by gathering the information @hmh requested. Those include:
1. output of /proc/cpuinfo
2. /var/log/kern.log

Please provide both of these with both the working and newest microcode. Also please reverify bluetooth in both cases, as we need to make sure that the issue wasn't simply resolved with a reboot.

Thank you,

Revision history for this message
yarko (yarkot) wrote :

I'm glad, then, I waited until someone reacted to my "plan".

Checking /var/log/kern.log, I have the log from when I first booted w/ the failing situation.
I know this, because I shutdown (rare) in anticipation of a coming Hurricane Harvey, and tweeted the next day when weather was stable.

The (attached) log - the interesting starting point, first boot w/ intel-microcode 3.2017.07.07.1 is "Aug 26 12:52" (shutdown was "Aug 26 00:10").

In the meantime, I'll upgrade just the intel-microcode to the Ubuntu release, and shortly provide that associated kern.log and /proc/cpuinfo, then cycle thru the various debian releases @hmh listed, and (finally) end with Ubuntu upgrades, including intel-microcode and kernel.

Let's see what happens...

Revision history for this message
yarko (yarkot) wrote :

Ok - here's my starting point (now):

Bluetooth working.
-----------------
attached archive has:
- installed version info for:
  - intel-microcode
  - linux-firmware
  - linux-image
- cpuinfo
- kern.log (boot log: output of journctl -k -b)

Revision history for this message
yarko (yarkot) wrote :

With no more than intel-microcode (re-) upgraded to 3.20170707.1~ubuntu16.04.0, and booting - now bluetooth works.

I can say this: with initial dropout of bluetooth, I rebooted I think a total of 3 times:
Once just reboot; Once complete power-down / power-up. Then, only downgrading intel-microcode worked.

Just now, I upgraded intel-microcode (simply, ubuntu dist - "sudo apt install intel-microcode" to bring it current), powered- down/up, and bluetooth works.

The only other thing I can note:
- I did revert the setting in /etc/bluetooth/main.conf => Autoconf=false (actually, commented out, so default as distributed)
- I booted with a USB keyboard plugged in (also);
- The bluetooth icon has (unusually?) a "lock" icon (upper toolbar, right)

Looking at /var/log/apt/history.log -the only other thing changed / I tried (before downgrade - but not rebooted after) was an install of unity-control-center (no doubt from googling around, before I looked at install history for the root of my problem).

Here is (this time) the logs from this boot.

=> I'll wait for guidance / request from someone here on next steps.

If nothing else, I will do the "normal" next step of apt-upgrading, which will upgrade my kernel (but I'll wait to hear from anyone here, first).

Revision history for this message
Robie Basak (racb) wrote :

Thank you for the report.

Based on that I think it's safe to say that this isn't caused by the issued update of intel-microcode? So I'll mark the intel-microcode task Invalid.

If you determine that the regression was caused by the kernel update, then please add a bug task against the "linux" package. Use "Also affects distribution/package" and choose Ubuntu and the package linux.

Changed in intel-microcode (Ubuntu):
status: New → Invalid
Revision history for this message
yarko (yarkot) wrote :

Thanks Robie.

Only sense I can make of it (since I struggled for a while, before trying downgrade of intel-microcode) --- downgrade flushed / did something w/ some settings, somewhere (and that persisted).

I've upgraded to 4.4.0.93.98 kernel now, with latest intel-microcode, and all continues to work apparently well.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.