Bug 100621

tried that too :

# cat /etc/modprobe.d/iwlwifi.conf 
options iwlwifi 11n_disable=1

and also:
options iwlwifi swcrypto=1

Please share your syslog output.
What is the kernel / supplicant version you have on the USB stick?

Note: Nothing has really changed in the kernel in that driver for a long time.

You mean syslog from the working fedora ?
I'll reboot into it now...

No, the syslog of the failing system

Working fedora liveUSB

$ cat /etc/fedora-release 
Fedora release 21 (Twenty One)

$ uname -a
Linux localhost 3.17.4-301.fc21.x86_64 #1 SMP Thu Nov 27 19:09:10 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux

Name        : wpa_supplicant
Version     : 2.0
Release     : 12.fc21

Created attachment 181161 [details]
dmesg from working fedora

Firmware from fedora

Name        : iwl5000-firmware
Version     : 8.83.5.1_1
Release     : 41.fc21

Created attachment 181171 [details]
dmesg from failing debian

I start now to understand that the Debian is working and not the Fedora... Not very clear from your first report.... Oh well...


I asked for the syslog and not for the dmesg.
Why supplicant version do you have on your Debian system?

Ah ok - got on my mail

Oh, please add -dd to your wpa_supplicant options.

can you record a tracing?

sudo trace-cmd record -e iwlwifi -e iwlwifi_data?

Debian sid is my installed system, which stopped working some times ago.
I tried booting a fedora 21 liveusb to test, and this one is working.

I sent the dmesgs before seeing that you wanted the full syslogs

I'll get you the other requested infos

On the failing debian system, I tried various versions of wpasupplicant
2.3-2
2.2-1
1.1-1
1.0-3.1
1.0-3
1.0-2

I also tried older libssl version because of this https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=788567

Created attachment 181191 [details]
dmesg from failing debian, new kernel, with iwlwifi debug options

Created attachment 181201 [details]
trace-cmd

gzipped output from

trace-cmd record -e iwlwifi -e iwlwifi_data

when trying to connect to AP (through NM-applet)

where do I put the -dd wpasupplicant option ?

From the driver point of view, I don't see any failure besides the fact that I can only see the EAPOL 1/4 and 4/4 which is strange but very unlikely to be related to the kernel which is what this bugzilla is all about.

Note that as I said, very little has changed in the driver for many years now. So if it works in Fedora and not in Debian, I'd check the supplicant / libssl and whatnot.

You mentioned that wpa_supplicant 2.0 works for you on Fedora, but didn't test that one on Debian.

To add -dd to the supplicant, you can launch it manually:
/sbin/wpa_supplicant -s -d -P /run/wpa_supplicant.wlan0.pid -i wlan0 -D nl80211 -c /etc/wpa_supplicant/wpa_supplicant.conf

You'll need to create a wpa_supplicant.conf file.

Yeah, I'm fairly confident this is not a kernel problem, as I keep a lot of old kernels around (that I build myself from linus's repo) and going back to an older kernel that was used when the wifi was working did not fix the problem.

I'm not so sure about the other components: firmware, wpa_supplicant, libssl...
BTW, are there *other* packages that make the wifi stack for this HW ?

I didn't test wpa_supplicant 2.0 since I couldn't find a binary package of this version, but going before 2.0 does not fix it anyways...

The strange thing, is that I'm currently on :

wpa_supplicant 1.0-2
firmware-iwlwifi 0.42
libssl 1.0.1k-3

Which is fairly far in the past, and nothing changed...

I'd need to double check with the working fedora, but I'm wondering if I'm not missing a piece of the puzzle...

The components that are needed are:
kernel
firmware
supplicant

now the supplicant itself has a few dependencies:
$ ldd /sbin/wpa_supplicant 
	linux-vdso.so.1 =>  (0x00007fff5d72e000)
	librt.so.1 => /lib/x86_64-linux-gnu/librt.so.1 (0x00007fd00db26000)
	libnl-3.so.200 => /lib/x86_64-linux-gnu/libnl-3.so.200 (0x00007fd00d90a000)
	libnl-genl-3.so.200 => /lib/x86_64-linux-gnu/libnl-genl-3.so.200 (0x00007fd00d704000)
	libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007fd00d500000)
	libpcsclite.so.1 => /lib/x86_64-linux-gnu/libpcsclite.so.1 (0x00007fd00d2f5000)
	libssl.so.1.0.0 => /lib/x86_64-linux-gnu/libssl.so.1.0.0 (0x00007fd00d094000)
	libcrypto.so.1.0.0 => /lib/x86_64-linux-gnu/libcrypto.so.1.0.0 (0x00007fd00ccb0000)
	libdbus-1.so.3 => /lib/x86_64-linux-gnu/libdbus-1.so.3 (0x00007fd00ca68000)
	libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007fd00c69e000)
	libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007fd00c480000)
	libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007fd00c178000)
	/lib64/ld-linux-x86-64.so.2 (0x00007fd00dd2e000)

The kernel and the firmware haven't changed for years for that device and you have the latest firmware.

I finally found something after stumbling upon that other bug report:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=790326

by creating a very minimal wpa.conf file:

network={
  ssid="NETWORK_NAME"
  scan_ssid=1
  key_mgmt=WPA-PSK
  psk="NETWORK_PASSWORD"
}

and then running:
wpa_supplicant -i wlan0 -ddd -Dwext -cwpa.conf

So I think that some update made wpa_supplicant configuration change from wext driver to another not working one...

That may be a NetworkManager change, or some other random change somewhere...

:-/

BTW thanks for all the help and time spent on this...

I hope, other people that fall into this trap, can find some help reading this bug report...

And for completeness sake, I also had to:

rfkill unblock wlan

Ok - so I guess we can close this issue and smash the wpa_s / NM packager at Debian :)

No, I wouldn't want to smash anyone, at least not until I fully
understand where the problem came from...

:-)

On Mon, Jun 29, 2015 at 8:40 PM,  <bugzilla-daemon@bugzilla.kernel.org> wrote:
> https://bugzilla.kernel.org/show_bug.cgi?id=100621
>
> --- Comment #24 from Emmanuel Grumbach <emmanuel.grumbach@intel.com> ---
> Ok - so I guess we can close this issue and smash the wpa_s / NM packager at
> Debian :)
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
> You reported the bug.

But yes, I'm OK to close this bug report...

On Mon, Jun 29, 2015 at 8:52 PM, Vincent Legoll
<vincent.legoll@gmail.com> wrote:
> No, I wouldn't want to smash anyone, at least not until I fully
> understand where the problem came from...
>
> :-)
>
> On Mon, Jun 29, 2015 at 8:40 PM,  <bugzilla-daemon@bugzilla.kernel.org>
> wrote:
>> https://bugzilla.kernel.org/show_bug.cgi?id=100621
>>
>> --- Comment #24 from Emmanuel Grumbach <emmanuel.grumbach@intel.com> ---
>> Ok - so I guess we can close this issue and smash the wpa_s / NM packager at
>> Debian :)
>>
>> --
>> You are receiving this mail because:
>> You are on the CC list for the bug.
>> You reported the bug.
>
>
>
> --
> Vincent Legoll