Description of problem: Saw this in /var/log/messages: Sep 27 16:55:38 tlondon kernel: iwl3945: Can not allocate SKB buffers Sep 27 16:55:42 tlondon kernel: iwl3945: Can not allocate SKB buffers Sep 27 16:55:44 tlondon kernel: iwl3945: Can not allocate SKB buffers I could detect no failures or unusual behavior, and could find no other messages. Version-Release number of selected component (if applicable): kernel-2.6.27-0.354.rc7.git3.fc10.i686 How reproducible: First time I noticed it.... Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Continue to see this on Thinkpad X61: Oct 18 10:06:42 tlondon kernel: iwlagn: Can not allocate SKB buffers Oct 18 10:06:44 tlondon kernel: iwlagn: Can not allocate SKB buffers Oct 18 10:06:46 tlondon kernel: iwlagn: Can not allocate SKB buffers Oct 18 10:07:03 tlondon kernel: iwlagn: Can not allocate SKB buffers Oct 18 10:08:09 tlondon kernel: iwlagn: Can not allocate SKB buffers Oct 18 10:08:16 tlondon kernel: iwlagn: Can not allocate SKB buffers Wireless is not associated, NetworkManager is running supporting "wired" connection. 03:00.0 Network controller: Intel Corporation PRO/Wireless 4965 AG or AGN Network Connection (rev 61)
Oops, forgot kernel version: kernel-2.6.27.2-23.rc1.fc10.x86_64 [Spin2]
I'm not sure this is really an iwlagn issue -- it is literally reporting a failure of alloc_skb. Is your system low on memory?
2GB ThinkPad X60 and X61. Don't think so.... Typical top lines from 'top': [root@tlondon ~]# top top - 13:24:49 up 4:34, 4 users, load average: 0.50, 0.63, 0.41 Tasks: 164 total, 2 running, 161 sleeping, 0 stopped, 1 zombie Cpu(s): 6.5%us, 5.0%sy, 0.0%ni, 86.6%id, 1.7%wa, 0.2%hi, 0.2%si, 0.0%st Mem: 2016320k total, 1970648k used, 45672k free, 31000k buffers Swap: 4063224k total, 100k used, 4063124k free, 605440k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 3749 tbl 20 0 966m 537m 22m S 12.0 27.3 38:53.84 qemu-kvm 3585 tbl 20 0 1265m 85m 22m S 3.6 4.3 8:03.07 rhythmbox
Not 100% sure, but seems to happen frequently around supplicant changes from 6 -> 7 Oct 24 10:47:40 tlondon NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7 Oct 24 10:57:40 tlondon NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6 Oct 24 10:57:40 tlondon NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7 Oct 24 10:59:39 tlondon kernel: iwlagn: Can not allocate SKB buffers Oct 24 10:59:40 tlondon kernel: iwlagn: Can not allocate SKB buffers Oct 24 11:07:40 tlondon NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6 Oct 24 11:07:40 tlondon NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7
This may be related to this bug Ubuntu is tracking: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/250139 It appears to interfere with WPA and, for me, DHCPDISCOVER.
I know is sounds strange, but I think I only get these messages when I am @work and associated with an 802.1X network. I don't think I see these when I am @home with my "vanilla WPA2" network.
#7 is wrong.... I do get it @home with standard WPA2: Nov 16 18:55:17 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:17 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:17 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:18 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:18 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:18 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:18 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:18 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:18 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:18 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:33 tlondon kernel: __ratelimit: 4228 callbacks suppressed Nov 16 18:55:33 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:35 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:36 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:39 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:39 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:40 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:41 tlondon kernel: iwlagn: Can not allocate SKB buffers Nov 16 18:55:43 tlondon kernel: iwlagn: Can not allocate SKB buffers Output from "top": top - 19:17:18 up 32 min, 3 users, load average: 0.15, 0.31, 0.53 Tasks: 167 total, 2 running, 165 sleeping, 0 stopped, 0 zombie Cpu(s): 3.4%us, 1.3%sy, 0.0%ni, 94.7%id, 0.6%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 2018336k total, 1897740k used, 120596k free, 411348k buffers Swap: 4063224k total, 152k used, 4063072k free, 763680k cached Something in particular to look for in kernel space? slabinfo?
I happened to notice these messages from F10 Preview LiveCD *not* connected to any wireless network. I'm doing block device copies between partitions on the hard disk using "dd" and noticed these messages scroll by the text console... kernel-2.6.27.4-68.fc10.i686 ThinkPad T61 iwl4965 Maybe this is caused by lots of block layer activity? I doesn't seem to happen until I start the "dd".
This bug appears to have been reported against 'rawhide' during the Fedora 10 development cycle. Changing version to '10'. More information and reason for this action is here: http://fedoraproject.org/wiki/BugZappers/HouseKeeping
Is this still happening with current F10 kernels?
I'm running rawhide (kernel-2.6.28-0.121.rc7.git5.fc11.x86_64), and yes, it is still happening. For example: Dec 10 07:28:21 tlondon kernel: iwlagn: Can not allocate SKB buffers Dec 10 07:28:49 tlondon kernel: iwlagn: Can not allocate SKB buffers Dec 10 07:30:16 tlondon kernel: iwlagn: Can not allocate SKB buffers Dec 10 07:30:48 tlondon kernel: iwlagn: Can not allocate SKB buffers Dec 10 07:30:50 tlondon NetworkManager: <info> (wlan0): supplicant connection state: completed -> group handshake Dec 10 07:30:50 tlondon NetworkManager: <info> (wlan0): supplicant connection state: group handshake -> completed Dec 10 07:30:50 tlondon kernel: iwlagn: Can not allocate SKB buffers Dec 10 07:34:21 tlondon kernel: iwlagn: Can not allocate SKB buffers Dec 10 07:34:33 tlondon kernel: iwlagn: Can not allocate SKB buffers Dec 10 07:34:35 tlondon kernel: iwlagn: Can not allocate SKB buffers Dec 10 07:34:36 tlondon kernel: iwlagn: Can not allocate SKB buffers Dec 10 07:34:41 tlondon kernel: iwlagn: Can not allocate SKB buffers Dec 10 07:34:42 tlondon kernel: iwlagn: Can not allocate SKB buffers I can get hundreds/thousands of these.
Should this bug be moved back to rawhide (from 10)? What is the process for that?
http://koji.fedoraproject.org/koji/buildinfo?buildID=75463 The kernels being built at the link above have a patch that is supposed to address this. Once they are build, can you confirm?
OK. I've downloaded/installed kernel-2.6.28-0.131.rc8.git4.fc11.x86_64.rpm. Will reboot and monitor.....
Have not seen "SKB" in /var/log/messages since upgrading to F10. Current kernel is kernel-2.6.27.7-134.fc10.
I've been running kernel-2.6.28-0.131.rc8.git4.fc11.x86_64.rpm for about 5 hours now in my "@work" wireless environment (802.1X, etc.) with NO recurrence of the "Can not allocate SKB buffers" message. Looks good so far! I'll continue to monitor .... Thanks for the fix!