Created attachment 218231 [details] dmesg output I have a computer (intel based) with the Intel Dual band 3160, which has booth wireless and bluetooth capabilities. Excerpt from dmesg: ``` [ 612.459879] iwlwifi 0000:01:00.0: Detected Intel(R) Dual Band Wireless AC 3160, REV=0x164 ``` With the previous kernel I was able to use the Bluetooth, and the received signal for the wireless was always very strong. Starting with the latest kernel (I've tried 4.2, 4.4.0, and 4.4.11 mainline) I couldn't get the bluetooth working, and the degradation of the wifi signal is significant (sometimes not only the transfer is very slow, but the card cannot even associate with an Access Point). Kernel version: ``` Linux version 4.4.11-040411-generic (kernel@tangerine) (gcc version 5.3.1 20160413 (Ubuntu 5.3.1-14ubuntu2) ) #201605182255 SMP Thu May 19 02:57:22 UTC 2016 ``` I think this should be considered a regression, because it was working very well with previous kernels.
I have also updated to the latest mainline kernel (4.6.0-040600-generic) and I can still confirm the problem
please check with different wifi firmware version. ethtool -i wlan0 will tell you what version you are currently running.
This is the output: ``` mattions@apollo:~$ ethtool -i wlp1s0 driver: iwlwifi version: 4.6.0-040600-generic firmware-version: 17.311016.0 expansion-rom-version: bus-info: 0000:01:00.0 supports-statistics: yes supports-test: no supports-eeprom-access: no supports-register-dump: no supports-priv-flags: no ``` The device is named wlp1s0 on my system. I've tested the firmware from 13 to 17 (the range accepted by the kernel 4.6), but no difference could be detected.
Do you still have the kernel that was working well? What was that version? Can you please check that it still works well and the issue is not caused by an environmental change? We have a backport tree that can be easily bisectable. If you can go back to an old kernel, you can use that backport tree to try to bisect the problem. Please check the bug I just linked as well. Commit 923a8c1d8069104726bde55c37cec66324ccc328 and 34c8b24ff284 are related probably. But you have the fix in 4.6...
I knew it was working with Ubuntu 14.04 (https://wiki.ubuntu.com/TrustyTahr/ReleaseNotes/14.04), which was using the kernel: 3.13.0-24.46 Ubuntu Linux kernel I remember that I was doing the upgrade normally, and then one day the bluettooth was not working anymore, and the wireless performance degraded. Unfortunately I can not pick the exact kernel, however the latest ubuntu 14.04 has the 3.19 according to the releases, so maybe the problem could be between the two. I've tried to install very old kernels (via debian), however I have some dependencies conflict and they do not get install properly. I can't find the linked bug, where is it?
Look for See Also.
I was talking about this: https://bugzilla.kernel.org/attachment.cgi?id=178441&action=diff If you can't find the configuration that worked, we'd need to go for firmware debugging, that typically takes a lot of time. Can you please install our Core19 release (with Core14 firmware since your device is EOL'ed) and let us know? In your case, you can install the driver and the firmware only. No need to update the wpa_supplicant, but please make sure you disable VENDOR_CMDS as written in this page: https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi/core_release thank you.
@Emmanuel: so you say that the only firmware that should work properly in my case is the `/lib/firmware/iwlwifi-3160-14.ucode`? So basically the steps I have to do: 1) git clone https://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/backport-iwlwifi.git -b release/LinuxCore19 and then follow the steps to compile it. 2) How do I get Core14 firmware? Then I should just reboot? If this completely kills my wireless, can I remove it/uninstall it?
I just say it is worth trying. Actually, I just checked, you should be able to skip step 1. Your kernel is new enough to use the -17.ucode. Just check the version you use with ethtool -i wlp1s0 You can get the Core14 firmware from here https://git.kernel.org/cgit/linux/kernel/git/iwlwifi/linux-firmware.git/plain/iwlwifi-3160-17.ucode Just drop the file into /lib/firwmare and reboot, then check with ethtool -i wlp1s0
This is the result: ``` mattions@apollo:~$ ethtool -i wlp1s0 driver: iwlwifi version: 4.4.0-28-generic firmware-version: 17.352738.0 expansion-rom-version: bus-info: 0000:01:00.0 supports-statistics: yes supports-test: no supports-eeprom-access: no supports-register-dump: no supports-priv-flags: no ``` However bluetooth is dead, and the wireless signal did not improve. (I skipped the reboot, I did modprobe -r. Let me know if this change a lot the situation and I will do a reboot).
Ok thanks. can you please record tracing? https://wireless.wiki.kernel.org/en/users/drivers/iwlwifi/debugging#tracing This may give us an insight on why traffic doesn't flow. If that doesn't shed more light, we'd need to debug the firmware.
Created attachment 222931 [details] trace.data compressed Trace.dat obtained with sudo trace-cmd record -e iwlwifi
What signal do you see? From the data you sent, you have around -80 which is low, but possible. To further debug this we'd need to debug the firmware, but the problem is that our firmware team and especially the BT Coexistence team is completely overloaded and it can take a long time until we get inputs from them. Let me know if you choose to provide the firmware data and I'll send the instructions on how to proceed.
Another thing we can try is to disable 11n. 11n can have bandwidth and rate settings that can make the interference with BT even worse. please add: options iwlwifi 11n_disable=1 to /etc/modprobe.d/iwlwifi.conf and reload the module or reboot. Thanks!
Hi Emmanuel, I've added the option to the /etc/modprobe.d/iwlwifi.conf but no real improvement: before: ``` mattions@apollo:~$ sudo iwconfig lo no wireless extensions. wlp1s0 IEEE 802.11abgn ESSID:"palmaland" Mode:Managed Frequency:2.412 GHz Access Point: C0:3E:0F:E7:6B:49 Bit Rate=57.8 Mb/s Tx-Power=20 dBm Retry short limit:7 RTS thr:off Fragment thr:off Encryption key:off Power Management:off Link Quality=42/70 Signal level=-68 dBm Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0 Tx excessive retries:0 Invalid misc:577 Missed beacon:0 docker0 no wireless extensions. ``` After: ``` mattions@apollo:~$ sudo iwconfig lo no wireless extensions. wlp1s0 IEEE 802.11abg ESSID:"palmaland" Mode:Managed Frequency:2.412 GHz Access Point: C0:3E:0F:E7:6B:49 Bit Rate=48 Mb/s Tx-Power=20 dBm Retry short limit:7 RTS thr:off Fragment thr:off Encryption key:off Power Management:on Link Quality=44/70 Signal level=-66 dBm Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0 Tx excessive retries:1 Invalid misc:101 Missed beacon:0 docker0 no wireless extensions. ``` If I can offer my help on debugging the firmware. Thanks!
In any case, I don't really see any bug here. The bit rate is high, the signal level is fair (-66)...
Well the bluetooth is broken (and it was working) and I can't get signal anymore in a room in the office where before I was able to do it. From my point of view the regression is there.
I understand, but I am sorry to say that the symptoms aren't obvious and precise enough to be "debuggable". I am moving the bug to BT and removing the intel wifi team.
Hi, how can I provide the info needed for this bug? My problem is still the same: the bluetooth is not working and the wireless card picks up only a very weak signal. Thanks.