Bug 198357
Summary: | iwlwifi: failed to remove key (1, ff:ff:ff:ff:ff:ff) from hardware (-22) | ||
---|---|---|---|
Product: | Drivers | Reporter: | zzpyynzjbszkbtipim |
Component: | network-wireless | Assignee: | DO NOT USE - assign "network-wireless-intel" component instead (linuxwifi) |
Status: | CLOSED CODE_FIX | ||
Severity: | normal | CC: | bjo, bugzilla, dd.nexus, Esokrarkose, linuxwifi, mark.harfouche, zmeyski |
Priority: | P1 | ||
Hardware: | Intel | ||
OS: | Linux | ||
Kernel Version: | 4.15.0-rc6-00048-ge1915c8195b3 | Subsystem: | |
Regression: | Yes | Bisected commit-id: | |
Attachments: |
Hardware/software info, logs from wpa_supplicant and the kernel
Fix canditate Tweak of proposed candidate fix Fix canditate dmesg 4.16rc1 lspci -vvnn |
Description
zzpyynzjbszkbtipim
2018-01-05 03:06:56 UTC
Created attachment 273623 [details]
Fix canditate
please try the patch attached.
Created attachment 273641 [details]
Tweak of proposed candidate fix
Hi Emmanuel, thanks for the patch. With it, I tested booting a couple times, and suspending a couple times after that. The error message did not appear, so that is great.
However, I did note that on one of the boots, I got a failed ASSERT_RTNL() inside the new ieee80211_iter_keys() call while associating. So the kernel dump_stack()'d, then half a second later printed:
[ 9.588473] iwlwifi 0000:01:00.0: No association and the time event is over already...
[ 9.588562] wlp1s0: Connection to AP <redacted> lost
This happened a couple more times in succession. Only on the 4th time, it succeeded.
I looked at the patch, and it seems a likely culprit is the new ieee80211_iter_keys() call being done after the mvm->mutex is unlocked. So I moved it above the unlock -- into the pre-existing old_state/new_state checks. See attached. I have tested this and no failed ASSERT_RTNL so far (and still no failed-to-remove-key error). Thoughts?
Created attachment 273651 [details]
Fix canditate
Fix that was merged in our internal tree.
This is still showing up in 4.15.3 and 4.16rc1. Created attachment 274173 [details]
dmesg 4.16rc1
Created attachment 274175 [details]
lspci -vvnn
Code hasn't been sent yet. We have a backlog of patches that need to be sent upstream. For anyone following along, the commit is now in Linus's tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=e4f13ad07823b . It looks like it was merged on 2018/03/22 by https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=c4f4d2f91772 . I've patched the commit against 4.15.15 and still got the issue: [ 83.667043] wlp1s0: deauthenticating from e8:df:70:41:f2:43 by local choice (Reason: 3=DEAUTH_LEAVING) [ 83.679337] wlp1s0: failed to remove key (1, ff:ff:ff:ff:ff:ff) from hardware (-22) Sorry, my issue was caused initially by another issue with the sata-controller which brought the system to the hanging state. The message still appear in 4.15.0 Still happening in 4.16.0, is the fix still not applied? |