Created attachment 641873 [details] data from abrt Description of problem: From time to time, while being connected to my home network, I loose the connection and I have to reboot to have working wifi again. I was trying to collect the necessary logs with abrt, but it failed repeatedly. The only thing I was able to do with it was save the attached log. Version-Release number of selected component (if applicable): Linux wycliff.ceplovi.cz 3.6.0-0.28.el7.x86_64 #1 SMP Thu Oct 4 11:18:08 EDT 2012 x86_64 x86_64 x86_64 GNU/Linux How reproducible: from time to time (like once a week or something like that)
Can we get full dmesg when this problem happen?
I would need the full log and debug flag TX_QUEUES: please add debug=0x80000000 as a module parameter when you load iwlwifi.
(In reply to comment #3) > I would need the full log and debug flag TX_QUEUES: > please add debug=0x80000000 as a module parameter when you load iwlwifi. Ehm, how would I do that? Otherwise, yes, when I know how I will start collecting the data, and let's see whether I catch some.
modprobe -r iwlwifi modprobe iwlwifi debug=0x80000000
Not sure if debug option is enabled in production kernel variant for RHEL7, if not kernel-debug should be installed. Also RHEL7 kernel does not include fixes from 3.6 -stable (i.e. fixes from 3.6.1 up to 3.6.6). Is possible that one of iwlwifi or mac80211 fixes from there address this issue too, Emmanuel ?
I don't know... We have a print there, it is probably a fw bug, but I need to get some prints to be sure... I also need IWLWIFI_DEBUG. BTW - this fw bug is handled in driver, but apparently, we still WARN_ON when it happens... Will have to check.
Created attachment 645693 [details] dmesg output (In reply to comment #5) > modprobe -r iwlwifi > modprobe iwlwifi debug=0x80000000 Even with Linux dhcppc5 3.6.0-0.29.el7.x86_64.debug #1 SMP Fri Nov 9 10:10:40 EST 2012 x86_64 x86_64 x86_64 GNU/Linux I get [ 17.841948] iwlwifi: Unknown parameter `debug' in dmesg.
Well, no wonder: dhcppc5:~# grep IWLWIFI /boot/config-3.6.0-0.29.el7.x86_64.debug CONFIG_IWLWIFI=m # CONFIG_IWLWIFI_DEBUG is not set # CONFIG_IWLWIFI_DEVICE_TRACING is not set # CONFIG_IWLWIFI_P2P is not set # CONFIG_IWLWIFI_EXPERIMENTAL_MFP is not set dhcppc5:~#
I just provided debug kernel to Matěj, waiting for debug logs ...
I opened bug 877409 - request for compile iwlwifi debug facilities, when that will be solved we can proceed here.
Created attachment 675866 [details] fix rate scaling can you please report back with this patch applied? Thanks
ping?
RHEL7 will update kernel to 3.8 or maybe 3.9 or even some newer. So it's not worth to test the patches now. Once RHEL7 kernel version will establish, we will see if bug is still there. Emmanuel , for now please ignore RHEL7 bug reports.
Matěj, is this bug still reproducible on newer RHEL7 kernel (i.e kernel-3.10.0-53.el7 or later) ?
Created attachment 826743 [details] output of dmesg (In reply to Stanislaw Gruszka from comment #16) > Matěj, is this bug still reproducible on newer RHEL7 kernel (i.e > kernel-3.10.0-53.el7 or later) ? a) Nightly gives me kernel-3.10.0-52.el7.x86_64 b) see attached dmesg output and journalctl one.
Created attachment 826745 [details] output of journalctl
-52.el7 kernel is fine for testing. I do not see any wireless problems in the logs. There are frequent DHCP renewals, but that is caused by small lease time parameter (600 seconds). I recommend you to increase that value in AP settings.
Matěj, is this problem still reproducible as described in comment 0 i.e. "From time to time, while being connected to my home network, I loose the connection and I have to reboot to have working wifi again" ? On log you provided in comment 18, there are only frequent DHCP renewals, but there is no sign of connection lost ...
(In reply to Stanislaw Gruszka from comment #21) > Matěj, is this problem still reproducible as described in comment 0 i.e. > "From time to time, while being connected to my home network, I loose the > connection and I have to reboot to have working wifi again" ? > > On log you provided in comment 18, there are only frequent DHCP renewals, > but there is no sign of connection lost ... I am still loosing connection, but now it is usually sufficient to restart NetworkManager.service.
I do not see any bad symptoms on updated logs, at least not on kernel/drivers side. Perhaps this bug should reassigned to NetworkManager ...
Removing Intel as there is no evidence that there is a driver bug.
I do not see possibly to fix this bug , sorry.