Description of problem: iwlwifi is producing very strange log messages. Sometimes it hangs or the wireless connection gets extremely slow. A workaround is to rmmod and modprobe it. It has been somewhat flaky in kernel 3.2, but is now almost unusable in all of the 3.3 series Version-Release number of selected component (if applicable): 3.3.0-0.rc1.git5.1.fc17.x86_64 How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Jan 30 14:33:59 localhost kernel: [ 1.347 wwf 000:00 xageainealdo a=0:2b:a3:3td=0<3>[ 129.927312] DMA-API: debuggin oto eoy-dsbig<6>[ 282.314537] iwlwifi 0000:03:00.0: Tx aggregation enabled on ra = 02:0c:22 i Jan 30 14:38:30 localhost kernel: [ 552.857270] iwlwifi 0000:03:00.0: Tx aggregation enabled on ra = 00:22:b0:ca:32:23 tid = 6 Jan 30 14:40:08 localhost kernel: 0:2b:a3:3td=6<6>[ 650.715079] iwlwifi 0000:03:00.0: Tx aggregation enabled on ra = 00:22:b0:ca:32:23 tid = 0 Jan 30 14:40:28 localhost kernel: [ 671.212865] iwlwii00:30.:T grgto nbe nr 02:0c:22 i Jan 30 14:40:38 localhost kernel: f 000:00 xageainealdo a=0:2b:a3:3td=0<6>[ 681.151237] iwlwifi 0000:03:00.0: Tx aggregation enabled on ra = 00:22:b0:ca:32:23 tid = 0 Jan 30 14:42:10 localhost kernel: 02:0c:22 i Jan 30 14:42:41 localhost kernel: f 000:00 xageainealdo a2:0c:22 i Jan 30 14:43:20 localhost kernel: r 02:0c:22 i Jan 30 14:45:13 localhost kernel: iwlwifi 0000:03:00.0: Tx aggregation enabled on ra = 00:22:b0:ca:32:23 tid = 0 Jan 30 14:48:34 localhost kernel: iwlwifi 0000:03:00.0: Tx aggregation enabled on ra = 00:22:b0:ca:32:23 tid = 6 Jan 30 14:50:46 localhost kernel: [ 773.334937] iwlwifi 0000:03:00.0: Tx aggregation enabled on ra =0:2b:a3:3td=0<6>[ 803.674126] iwlwii00:30.:T grgto nbe nr = 00:2b:a3:3td=0<6>[ 842.372334] iwlwifi 0000:03:00.0: Tx aggregation enabled ona=0:2b:a3:3td=0:30.:T grgto nbe nr 02:0c:22 i Jan 30 14:55:11 localhost kernel: iwlwifi 00000:00 xageainealdo a=0:2b:a3:3td=6egto nbe nr 02:0c:22 i Jan 30 14:55:48 localhost kernel: iwlwifi 0000:03:00.0: Tx aggregation enabled on ra = 00:22:b0:ca:32:23 tid = 0 Jan 30 14:56:45 localhost kernel: [ 1552.361561] iwlwifi 0000:03:00.0: Tx aggrainealdo a=0:2b:a3:3td=6<6>[ 1646.951683] iwlwifi 0000:03:00.0: Tx aggregation enabled on ra = 00:22:b0:ca:32:23 tid = 0 Expected results: Additional info: The aggregation spew has been around since 3.2. The other stuff is new. It is so unintelligible that I also considered that it might be some of the systemd logging stuff, but only iwlwifi seems to produce it.
This has been reported heavily in f16, where we're using compat-wireless-3.3-rc1. There are some patches floating around to fix some issues, but I don't know which is needed to fix this nor when it will show up in rawhide.
3.3.0-0.rc2.git3.2.fc17.x86_64 wireless has gotten much worse to the point where wireless is mostly non-functional. The fixes for both the functionality and the log messages may be http://lkml.indiana.edu/hypermail/linux/kernel/1202.0/01789.html which haven't been pulled into the main tree yet.
Updated Rawide to use a later compat-wireless snapshot. Please attempt to recreate with kernel-3.3.0-0.rc2.git3.3.fc17 (or later) kernels, and report the results here.
Using 3.3.0-0.rc2.git4.1.fc17.x86_64 The strange and continuous error messages have gone away. More importantly, wireless has been rock solid. This is the first 3.3 version where wireless has been dependable.... thanks. (I'll leave this open for a couple of days to see if there are other me-too's)
I spoke too soon... it still looks like something is scrawling Feb 6 18:07:15 localhost kernel: [12857.804124] iwlwifi 0000:03:00.0: Queu1 tc o 00 s Feb 6 18:07:15 localhost kernel: 3[25.016 wwf 000:00 urn Wra_t 9wiepr17<>187846]ilii00:30.:CretH edpr5 rt_t 1 Feb 6 18:07:15 localhost kernel: 3[25.023 wwf 000:00 ndmn imaerla Feb 6 18:07:15 localhost kernel: 6[90.037 f821 5300Kz-5300Kz@400Kz,(0 B,30 B)derri:3d3e8s:ffe6c0err0i igi-..o031.[3d0001c0]e 1sukfr100m.<>187843]ilii00:30.:CretS edpr5 rt_t 1 Feb 6 18:07:15 localhost kernel: [12857.814737] ieee80211 phy0: Hardware restart was requested Feb 6 18:07:15 localhost kernel: 3[25.026 wwf 000:00 urn Wra_t 9wiepr17<>187847]ilii00:30.:O eadfrwr eod<6>[12857.816166] iwlwifi 0000:03:00.0: L1 Enabled; Disabling L0S Feb 6 18:07:15 localhost kernel: [12857.822752] iwlwifi 0000:03:00.0: Radio type=0x2-0x2-0x1
Hmmm...ok, I didn't read this correctly at first. So something is corrupting the output in /var/log/messages ? Thats...wierd... Could you attach your complete /var/log/messages file here?
Created attachment 559975 [details] /var/log/messages Only the iwlwifi messages are corrupt. You'll have to scan the log to see them since even the "iwlwifi" prefix is missing.
In the last day of using 3.3.0-0.rc3.git2.2.fc18.x86_64 I haven't seen any corrupt output. Still getting Feb 12 07:45:00 localhost kernel: [31190.893492] iwlwifi 0000:03:00.0: Queue 2 stuck for 10000 ms. Feb 12 07:45:00 localhost kernel: [31190.893497] iwlwifi 0000:03:00.0: Current SW read_ptr 14 write_ptr 31 Feb 12 07:45:00 localhost kernel: [31190.893577] iwlwifi 0000:03:00.0: Current HW read_ptr 14 write_ptr 31 Feb 12 07:45:00 localhost kernel: [31190.893579] iwlwifi 0000:03:00.0: On demand firmware reload Feb 12 07:45:00 localhost kernel: [31190.896283] ieee80211 phy2: Hardware restart was requested Feb 12 07:45:00 localhost kernel: [31190.896619] iwlwifi 0000:03:00.0: L1 Enabled; Disabling L0S Feb 12 07:45:00 localhost kernel: [31190.903179] iwlwifi 0000:03:00.0: Radio type=0x2-0x2-0x1 which results in wireless being dead until rmmod iwlwifi; modprobe iwlwifi
Using 3.3.0-0.rc3.git6.2.fc18.x86_64 The corrupt output problem is gone, though the firmware reload problem remains. Closing this since the original problem is gone. Either there is another bug related to the firmware reload, or someone can make a new bug.