Description of problem: Wifi card fails to operate on reboot from failed suspend/resume on laptop; these two commands fix instantly: sudo rm /var/lib/NetworkManager/NetworkManager.state sudo service NetworkManager restart Version-Release number of selected component (if applicable): NetworkManager-0.8.0-6.git20100408.fc12.i686 How reproducible: Reproducible every time Steps to Reproduce: 1. With wifi card (pcmcia) up and connection established, close laptop lid 2. After ~30s open lid and observe computer does not resume 3. Reboot and observe that wifi card (pcmcia) lights do not light Actual results: Wifi card fails to operate Expected results: Wifi card lights and operates normally Additional info: The laptop is a HP ze5250. The contents of the NetworkManager.state file before removal are: -------- start of file [main] NetworkingEnabled=false WirelessEnabled=true -------- end of file The contents after these two commands: rm /var/lib/NetworkManager/NetworkManager.state service NetworkManager restart are: -------- start of file [main] NetworkingEnabled=true WirelessEnabled=true WWANEnabled=true -------- end of file
The laptop's smolt ID is (taken without the wifi card): http://www.smolts.org/client/show/pub_8c1095ea-3e1e-451f-bfd7-e8609fbd20d9
Another report of what seems to be my exact problem, albeit on a ThinkPad, not a HP, and on Ubuntu, not Fedora, is at this URL: http://osdir.com/ml/ubuntu-users/2010-05/msg00156.html
Still another URL, https://bugs.launchpad.net/ubuntu/+bug/555571 This one shows a fix, exactly the fix I described above, and claims that "This bug affects 24 people", I don't know how they arrived at that number.
Another smolt of the laptop I reproduced this problem on. This one includes the Broadcom 4318 card: Broadcom Corporation BCM4318 [AirForce One 54g] 802.11g Wireless LAN Controller The smolt info is at: http://www.smolts.org/client/show/pub_8c1095ea-3e1e-451f-bfd7-e8609fbd20d9
Also see Bugid 589108 comment 1, exactly the same problem and circumvention as I described above. And note from that bug report that an alternative way to encounter this problem is with battery drain. This happened to me too, months ago.
Created attachment 414302 [details] Patch that demonstrates proof of concept described in Comment 6 See Comment 6. The patch is to src/main.c and adds option --force-enabled. When the option is used, parse_state_file() forces NetworkingEnabled to be true.
The real fix for this is to stop using sleep/wake as the "Networking Enabled" commands, which is how they are currently wired up. That's planned.
*** This bug has been marked as a duplicate of bug 589108 ***