kernel-packages team mailing list archive

[Seth Forshee <seth.forshee+lp@xxxxxxxxxxxxx>]

On Thu, Jun 04, 2015 at 03:45:32PM -0000, Jason Gerard DeRose wrote:
> @sforshee - I'm trying to better understand the consequences of Ubuntu
> currently having an "an old regulatory database version", as some
> System76 customers are having WiFi problems when connecting to AC
> routers with an Intel 7265 and `sudo iw reg set US` seems to fix their
> issues.
> 
> In particular, should a regulatory domain be set automatically once the
> database is updated? Currently I'm seeing the world-regulatory domain
> (ie, not set), which seems to be an intersection of all domains and
> means some band that could be used aren't (in the US anyway).

The answer is a bit complicated. iwlwifi will actually read out a set of
custom regulatory rules from the firmware and supply those to the
wireless core. This becomes the "default" set of rules rather than the
generic world domain when no specific regulatory domain has been
supplied. This could be a similarly conservative set of rules meant to
be "safe" around the world, or it could reflect the rules of the country
where the device was intended to be sold (I'm not sure off the top of my
head which it is for this card, but my best recollection is that it's
the former).

To emphasize the point - even though the kernel has retrieved the world
domain from crda, prints them in dmesg, and outputs them in response to
'iw reg get', these rules are *not* being used for an Intel wireless
card unless the user has manually loaded those rules (e.g. 'iw reg set
00'). Instead the kernel is using the rules it was given by the device's
firmware.

When a specific regulatory domain is set a process in userspace (crda)
will read the rules for that domain from the regulatory database and
load them into the kernel, and those rules will be used instead of the
ones the driver supplied. Sometimes the access point will broadcast
information about the regulatory domain in its beacon frames, in which
case the regulatory rules for the country will be loaded automatically
when connecting to that AP. More often (in my experience at least) the
AP does not provide this information.

The update to the database will change nothing about this process. The
same rules are going to be supplied by iwlwifi, and country-specific
rules will still be loaded or not loaded automatically based on what the
AP is broadcasting exactly as before. All it will change is the rules
which get loaded into the kernel for a given regulatory domain.

You aren't specific about what kind of problems your customers are
experiencing. One possibility is that the AP is using a channel that the
default rules says is not allowed but the US rules do allow, and that
loading the US rules opens up use on that channel. If this is the
problem then it will not be fixed by updating the database.

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to wireless-regdb in Ubuntu.
https://bugs.launchpad.net/bugs/1437913

Title:
  8086:095b [Lenovo ThinkPad X250] Wifi unstable on vivid with Intel
  7265

Status in The Linux Kernel:
  Unknown
Status in linux package in Ubuntu:
  Invalid
Status in wireless-regdb package in Ubuntu:
  Confirmed
Status in linux source package in Trusty:
  Invalid
Status in wireless-regdb source package in Trusty:
  Confirmed
Status in linux source package in Vivid:
  Invalid
Status in wireless-regdb source package in Vivid:
  Confirmed
Status in linux source package in Wily:
  Invalid
Status in wireless-regdb source package in Wily:
  Confirmed

Bug description:
  After upgrading from utopic to vivid beta, wifi on the Intel 7265 on my Thinkpad X250 has become unstable (linux-image 3.19.0-10.10, linux-firmware 1.143):
  03:00.0 Network controller [0280]: Intel Corporation Wireless 7265 [8086:095b] (rev 59)
  	Subsystem: Intel Corporation Dual Band Wireless-AC 7265 [8086:5210]

  I'm seeing two primary problems:
   * On 5GHz 802.11ac, I'll get DEAUTH_LEAVING almost exactly every minute, followed by an immediate reconnect.
   * On both 2.4GHz and 5GHz, all network activity will cease after 10-20 minutes.  A manual reconnect will temporarily fix network activity.

  This is complicated by the fact that the newer firmware requires
  kernel 3.19 and kernel 3.19 requires the iwlwifi-7265D*.ucode firmware
  instead of the old iwlwifi-7265*.ucode firmware, so it's not possible
  to mix/match older/newer kernel/firmware combinations, nor do a kernel
  bisect.

  FAOD, the latest utopic kernel/firmware combination (linux-image
  3.16.0-33.44, linux-firmware 1.138.1) works fine, which I can use in
  vivid if I manually install them.

  I have tested ~kernel-ppa/mainline 3.19.3-031903.201503261036 and
  4.0.0-040000rc5.201503230035, which exhibit the same problems.

  WORKAROUND: iw reg set US

  ProblemType: Bug
  DistroRelease: Ubuntu 15.04
  Package: linux-image-3.19.0-10-generic 3.19.0-10.10
  ProcVersionSignature: Ubuntu 3.19.0-10.10-generic 3.19.2
  Uname: Linux 3.19.0-10-generic x86_64
  ApportVersion: 2.16.2-0ubuntu5
  Architecture: amd64
  AudioDevicesInUse:
   USER        PID ACCESS COMMAND
   /dev/snd/controlC0:  ryan       1862 F.... pulseaudio
   /dev/snd/controlC1:  ryan       1862 F.... pulseaudio
  CurrentDesktop: Unity
  Date: Sun Mar 29 11:42:13 2015
  HibernationDevice: RESUME=UUID=c21a0227-27e2-4c4f-8bd9-30793932acbc
  InstallationDate: Installed on 2015-03-07 (22 days ago)
  InstallationMedia: Ubuntu 14.10 "Utopic Unicorn" - Release amd64 (20141022.1)
  MachineType: LENOVO 20CMCTO1WW
  ProcFB: 0 inteldrmfb
  ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-3.19.0-10-generic root=/dev/mapper/ubuntu--vg-root ro acpi_osi=Linux acpi_backlight=vendor psmouse.proto=imps quiet splash vt.handoff=7
  RelatedPackageVersions:
   linux-restricted-modules-3.19.0-10-generic N/A
   linux-backports-modules-3.19.0-10-generic  N/A
   linux-firmware                             1.143
  SourcePackage: linux
  UpgradeStatus: Upgraded to vivid on 2015-03-29 (0 days ago)
  dmi.bios.date: 01/23/2015
  dmi.bios.vendor: LENOVO
  dmi.bios.version: N10ET28W (1.05 )
  dmi.board.asset.tag: Not Available
  dmi.board.name: 20CMCTO1WW
  dmi.board.vendor: LENOVO
  dmi.board.version: SDK0E50512 STD
  dmi.chassis.asset.tag: No Asset Information
  dmi.chassis.type: 10
  dmi.chassis.vendor: LENOVO
  dmi.chassis.version: None
  dmi.modalias: dmi:bvnLENOVO:bvrN10ET28W(1.05):bd01/23/2015:svnLENOVO:pn20CMCTO1WW:pvrThinkPadX250:rvnLENOVO:rn20CMCTO1WW:rvrSDK0E50512STD:cvnLENOVO:ct10:cvrNone:
  dmi.product.name: 20CMCTO1WW
  dmi.product.version: ThinkPad X250
  dmi.sys.vendor: LENOVO

To manage notifications about this bug go to:
https://bugs.launchpad.net/linux/+bug/1437913/+subscriptions