Created attachment 735721 [details] dmesg output Description of problem: The system keeps disconnecting from the wifi network repeatedly. Version-Release number of selected component (if applicable): How reproducible: Frequently. Cannot be sure if this always happens. Steps to Reproduce: 1. Connect to my university network: UTS WPA 2. Use system normally 3. Actual results: System disconnects from network randomly. At this time, network manager forgets the network's password also for some reason. I must open network settings, re-enter the password and then reconnect to the network. I'm not sure if this is a kernel issue. Could it be a network manager issue? Reloading the brcmsmac module and restarting networkmanager does not work around this issue. Expected results: System should stay connected to the network and work normally. Additional info: dmesg output attached
Linux localhost 3.8.7-201.fc18.x86_64 #1 SMP Fri Apr 12 19:46:19 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux
My android phone connects just fine to the wifi network. I have to resort to USB tethering rather than using wifi because it keeps hanging up.
Please check the firmware version you have for me: latest release in linux-firmware is in lib/firmware/brcm/ and post here. Latest I have is: md5sum /lib/firmware/brcm/bcm43xx*.fw b0736e3590b05d27284fbb8a3efd50e1 brcm/bcm43xx-0.fw 5e51778ee011badcb42f1e2cb4ab3956 brcm/bcm43xx_hdr-0.fw
Hi John [root@localhost brcm]# ls bcm4329-fullmac-4.bin bcm43xx-0.fw bcm43xx_hdr-0.fw brcmfmac43236b.bin brcmfmac4329.bin brcmfmac4330.bin brcmfmac4334.bin [root@localhost brcm]# pwd /lib/firmware/brcm [root@localhost brcm]# md5sum * c53608f5818b702c46a012c57b4196ee bcm4329-fullmac-4.bin b0736e3590b05d27284fbb8a3efd50e1 bcm43xx-0.fw 5e51778ee011badcb42f1e2cb4ab3956 bcm43xx_hdr-0.fw f579673b5dc45640c814a9c68abcaf55 brcmfmac43236b.bin ff610cee869375a2d0c0be6b97b107fb brcmfmac4329.bin 4ec6341cbe351f13d787aaea99141bea brcmfmac4330.bin 0e5d2b9bfaaf3b6c43077405cfe92632 brcmfmac4334.bin [root@localhost brcm]# Thanks, Warm regards, Ankur
So latest firmware, that's cool. I see a combination of things, reason 3 (timeout) and reason 6 (occuring well after association completes, say 20 seconds.) Is this the same location you had issues with before in previous BZ? I think so. Do you see the same problems at say a starbucks or other less busy network? I'm wondering if this is a location with a lot of other traffic or noise that is creating a lot of the disconnects. You this just started, wasn't like this earlier? Can you describe the location for me: a lot of other people on the network? If you can get it for me, the supplicant log during this time might help some too /var/log/wpa_supplicant.log. You may need to add a -d (debug logging) to the command line to get sufficient detail. Can you upload that, please?
Yes, this is the same location. The network is university provided, and has the same ID: UTS-WPA everywhere in the university. The same university credentials work throughtout the university campus. It generally don't have issues using this network in university buildings. The disconnections etc behavioiur occurs in the hostel. The network here also has the same ID: UTS-WPA, but I'm not sure if the setup is the same. It's a hostel, with a lot of students, so yes, there would be quite a lot of traffic and multiple users and routers too. It wasn't this bad from the start. The patch in the other bug did seem to improve the stability somewhat. The notable point is that none of my cell phones (2 androids and 1 blackberry) experience this constant disconnection. They seem to connect and remain on the network just fine. It's only this laptop that has the issues. I don't have the disconnect issues on other networks, say at my sisters place. It is limited to this network currently. The wpa_supplicant.log is attached. Could you tell me where to add the -d flag please? I use network manager, not sure where wpa_supplicant is called. Thanks, Warm regards, Ankur
Created attachment 736834 [details] wpa supplicant log
Created attachment 736835 [details] fresh dmesg showing more wifi errors
(In reply to comment #7) > Created attachment 736834 [details] > wpa supplicant log interesting. The connection is using PEAP certificate based auth with MS-CHAPv2 inner auth. It does succeed, but then disconnection with great variety of reasons (that's not as easy to pin point.) Here's a partial list of disconnect reasons: Association reject for status 17: Information element in 4Way Handshake different from (Re)Association Request/Probe Response/Beacon Also..these more common. But a great number of reasons is unusual.. 4 inactivity 3 leaving or lost ap 1 unspec 6 class 2 from non-auth'ed Can you get the version of these two please? rpm -q NetworkManager rpm -q wpa_supplicant Wondering if an update is needed here. If you see an update or either or both, try that..This stuff is likely above the driver level.
Is the ampdu tx phy error message what you refer to? Checked a little, these transmits appear to be retried, I think. Need to look at this a bit more, certainly if it's an issue, a driver fix here. Thanks for the details. Good stuff. Oh, don't worry about the supplicant -d flag, info level is fine.
Hi John, Since 3.9.0-0.rc7.git3.1.fc19.x86_64, wifi's much better now. Since it's mid semester holidays here, the traffic has really gone down though. Only 2 of us are here out of a total 6 in my apartment. Similarly in the other apartments too. Dmesg is still littered with: [ 4705.129661] brcmsmac bcma0:0: phyerr 0x1, rate 0x14 [ 4705.129673] brcmsmac bcma0:0: brcms_c_ampdu_dotxstatus_complete: ampdu tx phy error (0x1) [ 4711.126072] brcmsmac bcma0:0: phyerr 0x1, rate 0x14 [ 4711.126082] brcmsmac bcma0:0: brcms_c_ampdu_dotxstatus_complete: ampdu tx phy error (0x1) [ 4711.126549] brcmsmac bcma0:0: phyerr 0x1, rate 0x14 [ 4711.126555] brcmsmac bcma0:0: brcms_c_ampdu_dotxstatus_complete: ampdu tx phy error (0x1) [ 4717.148470] brcmsmac bcma0:0: phyerr 0x10, rate 0x37 [ 4717.148480] brcmsmac bcma0:0: brcms_c_ampdu_dotxstatus_complete: ampdu tx phy error (0x10) [ 4717.148902] brcmsmac bcma0:0: phyerr 0x10, rate 0x37 [ 4717.148907] brcmsmac bcma0:0: brcms_c_ampdu_dotxstatus_complete: ampdu tx phy error (0x10) [ 4717.149352] brcmsmac bcma0:0: phyerr 0x10, rate 0x37 [ 4717.149358] brcmsmac bcma0:0: brcms_c_ampdu_dotxstatus_complete: ampdu tx phy error (0x10) [ 4766.575361] brcmsmac bcma0:0: phyerr 0x1, rate 0x14 [ 4766.575402] brcmsmac bcma0:0: brcms_c_ampdu_dotxstatus_complete: ampdu tx phy error (0x1) [ 4766.576185] brcmsmac bcma0:0: phyerr 0x1, rate 0x14 [ 4766.576216] brcmsmac bcma0:0: brcms_c_ampdu_dotxstatus_complete: ampdu tx phy error (0x1) [ 4798.126578] brcmsmac bcma0:0: phyerr 0x1, rate 0x14 [ 4798.126586] brcmsmac bcma0:0: brcms_c_ampdu_dotxstatus_complete: ampdu tx phy error (0x1) [ 4862.076515] brcmsmac bcma0:0: brcmsmac: brcms_ops_bss_info_changed: disassociated Since wifi is working decently (no frequent hang ups yet), these messages are probably just warnings or debugging, and not errors. Thanks, Ankur
Created attachment 739201 [details] wpa supplicant log for wifi working decently
Have looked at the above: I don't think this is functional issue, although it's certainly annoying if it creates a lot of "log noise". If radio condition deteriorate to the point of disconnection (which is normal part of wireless networking), re-connection will be initiated, this is normal wireless in noisy radio worlds. If you haven't already, update the two packages to help with connectivity issues. It will likely help when network is busy.. NetworkManager wpa_supplicant Since this issue seems much improved, fix is available, lets close this issue out? Fix is validate by code inspection and usage in v3.9 rc7 and beyond. If it reappears, we can open a new issue or revisit.
Hi, Yep. rc8 is good too. There aren't any updates for wpa_supplicant or NetworkManager but I'll keep an eye out. Closing both these bugs. Thanks for your help John :) Warm regards, Ankur
Terrific.. Just in time for the weekend. Thanks.