Re: [gentoo-user] trouble with wireless on new install

[Mick]

On Wednesday 29 Jul 2015 19:35:57 gottlieb@×××.edu wrote:
> On Wed, Jul 29 2015, Mick wrote:
> > On Wednesday 29 Jul 2015 01:36:22 gottlieb@×××.edu wrote:
> >> I am having trouble with wireless on a new install (gnome/systemd).
> >>
> >> lspci reports
> >>
> >>   02:00.0 Network controller: Intel Corporation Wireless 7265 (rev 59)
> >>
> >> I looked this up and it requires the iwlwifi driver and iwlmvm, which I
> >> enabled in the kernel (as modules). lsmod reports
> >>
> >>     Module                  Size  Used by
> >>     iwlmvm                142993  0
> >>     mac80211              425803  1 iwlmvm
> >>     x86_pkg_temp_thermal     4567  0
> >>     iwlwifi                88075  1 iwlmvm
> >>
> >> I have merged linux-firmware and ls /lib/firmware/*7265* reports
> >>
> >>     /lib/firmware/iwlwifi-7265-10.ucode
> >>     /lib/firmware/iwlwifi-7265-9.ucode
> >>     /lib/firmware/iwlwifi-7265-12.ucode
> >>
> >> /lib/firmware/iwlwifi-7265D-10.ucode /lib/firmware/iwlwifi-7265-8.ucode
> >> /lib/firmware/iwlwifi-7265D-12.ucode
> >>
> >> dmsg | grep iwl reports
> >>
> >>     [    2.819953] iwlwifi 0000:02:00.0: irq 48 for MSI/MSI-X
> >>     [    2.824130] iwlwifi 0000:02:00.0: loaded firmware version
> >>     23.11.10.0
> >>
> >> op_modeiwlmvm [    2.836969] iwlwifi 0000:02:00.0: Detected Intel(R)
> >> Dual Band Wireless AC 7265, REV=0x210 [    2.838620] iwlwifi
> >> 0000:02:00.0: L1 Disabled - LTR Enabled [    2.838775] iwlwifi
> >> 0000:02:00.0: L1 Disabled - LTR Enabled [    2.896168] ieee80211 phy0:
> >> Selected rate control algorithm 'iwl-mvm-rs' [    2.899325] iwlwifi
> >> 0000:02:00.0 wlp2s0: renamed from wlan0
> >>
> >> The complaint seems to be "L1 disabled"
> >>
> >> NetworkManager started the wired network but failed with the wireless
> >> E6430s ~ # systemctl -l status NetworkManager
> >> ● NetworkManager.service - Network Manager
> >>
> >>    Loaded: loaded (/usr/lib64/systemd/system/NetworkManager.service;
> >>
> >> enabled; vendor preset: enabled) Active: active (running) since Tue
> >> 2015-07-28 19:56:57 EDT; 9min ago Main PID: 232 (NetworkManager)
> >>
> >>    CGroup: /system.slice/NetworkManager.service
> >>
> >>            ├─232 /usr/sbin/NetworkManager --no-daemon
> >>            └─256 /sbin/dhclient -d -q -sf /usr/libexec/nm-dhcp-helper
> >>            -pf
> >>
> >> /var/run/dhclient-eno1.pid -lf
> >> /var/lib/NetworkManager/dhclient-0caab3d6-148f-416a-9906-547ed08596bf-en
> >> o1 .lease -cf /var/lib/NetworkManager/dhclient-eno1.conf eno1
> >>
> >> Jul 28 19:56:59 E6430s dhclient[256]: bound to 192.168.1.107 -- renewal
> >> in 43096 seconds. Jul 28 19:56:59 E6430s NetworkManager[232]: <info>
> >> NetworkManager state is now CONNECTED_GLOBAL Jul 28 19:56:59 E6430s
> >> NetworkManager[232]: <info>  Policy set 'Wired connection 1' (eno1) as
> >> default for IPv4 routing and DNS. Jul 28 19:56:59 E6430s
> >> NetworkManager[232]: <info>  (eno1): Activation: successful, device
> >> activated. Jul 28 19:57:01 E6430s NetworkManager[232]: <info>  (wlp2s0):
> >> supplicant interface state: disconnected -> inactive Jul 28 19:57:05
> >> E6430s NetworkManager[232]: <info>  startup complete Jul 28 19:57:31
> >> E6430s NetworkManager[232]: <info>  (eno1): Activation: Stage 4 of 5
> >> (IPv6 Configure Timeout) scheduled... Jul 28 19:57:31 E6430s
> >> NetworkManager[232]: <info>  (eno1): Activation: Stage 4 of 5 (IPv6
> >> Configure Timeout) started... Jul 28 19:57:31 E6430s
> >> NetworkManager[232]: <info>  (eno1): Activation: Stage 4 of 5 (IPv6
> >> Configure Timeout) complete. Jul 28 19:57:57 E6430s
> >> NetworkManager[232]: <info>  (wlp2s0): supplicant interface state:
> >> inactive -> scanning
> >>
> >> What step did I forget?
> >>
> >> thanks in advance,
> >> allan
> >
> > There are reports of driver problems, so you may not have forgotten
> > anything. Usual suggestion is to run the latest kernel and firmware, but
> > I don't have this card to know more about it.  What happens after the
> > 'inactive -> scanning' step above?  Have you tried to disconnect the
> > wired ethernet connection and wait a bit longer for the wireless?
>
> I just wrote an addendum that crossed in the mail with your msg.  This
> kernel with these drivers worked last week.  I just noticed that I
> didn't have an /etc/wpa_supplicant so added the same trivial one I have
> on my current machine
>     # The below line not be changed otherwise we refuse to work
>     ctrl_interface=/var/run/wpa_supplicant
>
>     network={
>             key_mgmt=NONE
>             priority=-9999999
>     }
> Unplugging the wired enet didn't help (I waited a minute).
> The trivial /etc/wpa_supplicant/wpa_supplicant.conf had no effect.

I assume that you had restarted your wireless interface?


> It looks strange.  dmesg shows the iwlwifi driver responding,
> recognizing the card, and loading the microcode using iwlmvm (see above)
> What does L1 disabled mean?  I googled and found other getting that msg
> but no explanation of what it means.
>
> thanks,
> allan

I think (but not sure) that L1 is a legacy power management feature of PCIe.  
LTR is a more dynamic, latency based, power management standard, which auto-
adjusts the power on the device depending on how long it takes to wake up.  L1 
on its own would consume more of your battery (if it is a laptop), with LTR it 
would switch off the power of parts of the circuit so as to avoid exceeding 
the latency requirement of the device (not all devices take the same time to 
wake up).

Could it be that MSWindows has set up on the hardware some aggressive power 
management setting, which Linux cannot wake up the device from?

Two things I would try:

1. In Linux - modinfo <module_name>

Check what options this gives and tweak the power settings accordingly as your 
modprobe it, or add it in /etc/modprobe.d/<module_name>.conf.  Also check the 
relevant kernel documentation in case it gives more details.

2. In MSWindows - Device Manager

Go into the Hardware/Device Manager and check the different tabs of the 
driver.  Make a note of the original settings and then tweak the power 
settings so that the device does not go to sleep.  Reboot into MSWindows (for 
good luck) and then boot into Linux.

Eventually, a more up to date driver ought to deal with this, if all my 
suggestions fail.

HTH.
-- 
Regards,
Mick