Bug 100801 - iwlwifi: 7265: TFD queue hang
Summary: iwlwifi: 7265: TFD queue hang
Status: CLOSED DUPLICATE of bug 95941
Alias: None
Product: Drivers
Classification: Unclassified
Component: network-wireless (show other bugs)
Hardware: Intel Linux
: P1 normal
Assignee: drivers_network-wireless@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-07-02 09:12 UTC by mick.saunders+kernel
Modified: 2015-07-12 07:09 UTC (History)
3 users (show)

See Also:
Kernel Version: 4.0.7-2-ARCH
Tree: Mainline
Regression: No


Attachments
dmesg.log showing iwlwifi errors (15.54 KB, text/x-log)
2015-07-02 09:12 UTC, mick.saunders+kernel
Details
lspci/lsmod/modinfo/early dmesg info for iwlwifi (40.14 KB, text/plain)
2015-07-02 09:22 UTC, mick.saunders+kernel
Details

Description mick.saunders+kernel 2015-07-02 09:12:18 UTC
Created attachment 181621 [details]
dmesg.log showing iwlwifi errors

First time noticing WiFi errors on this laptop.

* Noticed long periods of inactivity while web browsing.
* Started pings to 8.8.8.8 and noticed packet loss.
* Web browsing ultimately stopped, and I then noticed 'ping: sendmsg: No buffer space available'
* At this point I started collecting debugging information. (see attached files)
* I then disconnected from the current 2.4GHz network, and instead connected to a 5GHz network, the connection picked back up and seemed to work OK.

dmesg.log indicates a firmware/microcode bug, please see attached logging.

My hardware is a:
Metabox Clevo P650SG laptop
Specs:
Intel i7-4710HQ CPU
Intel Corporation Wireless 7265 (rev 48)
32GB RAM

Note: NVIDIA proprietary driver not loaded.

I seem to only be able to attach one file, so I'll have to add additional comments with the other debugging issue.
Comment 1 mick.saunders+kernel 2015-07-02 09:22:19 UTC
Created attachment 181631 [details]
lspci/lsmod/modinfo/early dmesg info for iwlwifi
Comment 2 Emmanuel Grumbach 2015-07-02 15:25:46 UTC
Can you try 4.1 with -13.ucode?
Comment 3 Emmanuel Grumbach 2015-07-02 15:31:10 UTC
Another easier thing to check. Please disable 40MHz operation on 2.4GHz with cfg80211 module parameter.

Thanks.
Comment 4 mick.saunders+kernel 2015-07-04 14:33:37 UTC
I used the laptop successfully on WiFi for a few hours yesterday even without applying any changes. I am having trouble reproducing the problem. If I can find a way to try and cause the issue again then I will definitely try your suggestions.

Do you happen to know a way I might be able to trigger the error or increase the chances of it happening again?
Comment 5 Emmanuel Grumbach 2015-07-04 19:40:07 UTC
Not really. Using 40MHz on 2.4GHz is usually a recipe for troubles.
Comment 6 mick.saunders+kernel 2015-07-07 12:50:14 UTC
OK the bug appeared again today while on the 2.4GHz network.

I tried using rmmod to remove the modules and reloaded with the cfg80211_disable_40mhz_24ghz=1 however after doing that and reloading iwlwifi again, I couldn't authenticate getting a lot of these in my dmesg:

[ 3227.707274] wlp4s0: authenticate with 10:0d:7f:74:f6:c2
[ 3227.710315] wlp4s0: send auth to 10:0d:7f:74:f6:c2 (try 1/3)
[ 3227.710981] wlp4s0: authenticated
[ 3228.017336] iwlwifi 0000:04:00.0: No association and the time event is over already...
[ 3228.017379] wlp4s0: Connection to AP 10:0d:7f:74:f6:c2 lost
[ 3229.560799] wlp4s0: aborting authentication with 10:0d:7f:74:f6:c2 by local choice (Reason: 3=DEAUTH_LEAVING)

However, I then made /etc/modprobe.d/cfg80211.conf with "options cfg80211 cfg80211_disable_40mhz_24ghz=1" and rebooted, and it seemed to authenticate fine, so I'm not sure if NetworkManager was playing up due to the module reload or if it was something else.

Anyway, I'll monitor it over the next week or so with that option on and see if occurs again.

Out of curiosity, is there a fix in 4.1 and in the newer -13 ucode for 40MHz on 2.4GHz operation? Just wondering why you recommended it first.
Comment 7 Emmanuel Grumbach 2015-07-07 12:56:51 UTC
the message: No association and time event is over has nothing to do with this bug.

There were a few fixes in -13.ucode that can impact the bug you are reporting.
Not specifically for 40Mhz.
Comment 8 Emmanuel Grumbach 2015-07-12 07:09:29 UTC
I found the bug that was about the exact same thing.

Please re-open if you see that -13 still causes issues.

*** This bug has been marked as a duplicate of bug 95941 ***

Note You need to log in before you can comment on or make changes to this bug.