Prev: soc/sh: Fix I2C dependency for SND_FSI_AK4642
Next: New issues in 2.6.35-rc{1,2}
From: Norbert Preining on 14 Jun 2010 12:20
Dear all,

(please cc)

Since AFAIR 2.6.32-rc the iwlagn driver seems to be completely hosed
with some AP. Permanently dropping connections, hanging, BUGing.
Unfortunately the only reply I get from iwl maintainers is "do git bisect"
which is a bit a pain with someting that is not stable reproducible.

It seems that since git is used *some* developers are relying on git
bisect instead of thinking what their code might do. I remember faintly
that Linus opposed the idea of kernel-debug for long time due to that
reasons, now we have it, answer is: We cannot reproduce, please bisect.

Here is a log of the most recent, I sent already many many different ones
to the ipw devel list, all without reasonable answer.

I would expect that all these warning messages and WARNINGS would at
least trigger some action like adding debug statements or whatsoever,
but alas, it didn't, and that since 32-pre.


[ 5131.444627] wlan0: authenticate with 00:0a:79:eb:56:10 (try 1)
[ 5131.447142] wlan0: authenticated
[ 5131.447187] wlan0: associate with 00:0a:79:eb:56:10 (try 1)
[ 5131.450903] wlan0: RX AssocResp from 00:0a:79:eb:56:10 (capab=0x411 status=0 aid=1)
[ 5131.450911] wlan0: associated
[ 5187.206459] iwlagn 0000:06:00.0: Fail finding valid aggregation tid: 0
[ 5191.499485] iwlagn 0000:06:00.0: Fail finding valid aggregation tid: 0
[ 5195.340108] iwlagn 0000:06:00.0: Fail finding valid aggregation tid: 0
[ 5205.513047] iwlagn 0000:06:00.0: Fail finding valid aggregation tid: 0
[ 5209.816242] iwlagn 0000:06:00.0: Fail finding valid aggregation tid: 0
[ 5212.914687] iwlagn 0000:06:00.0: iwlagn_tx_agg_start on ra = 00:0a:79:eb:56:10 tid = 0
[ 5328.057684] iwlagn 0000:06:00.0: Fail finding valid aggregation tid: 6
[ 5463.057686] iwlagn 0000:06:00.0: Fail finding valid aggregation tid: 6
[11296.057706] iwlagn 0000:06:00.0: Fail finding valid aggregation tid: 6
[12530.400033] Monitor-Mwait will be used to enter C-2 state
[12530.401175] Monitor-Mwait will be used to enter C-3 state
[13190.324128] No probe response from AP 00:0a:79:eb:56:10 after 500ms, disconnecting.
[13190.376076] cfg80211: Calling CRDA to update world regulatory domain
[13201.736859] wlan0: authenticate with 00:0a:79:eb:56:10 (try 1)
[13201.739331] wlan0: authenticated
[13201.739384] wlan0: associate with 00:0a:79:eb:56:10 (try 1)
[13201.743127] wlan0: RX AssocResp from 00:0a:79:eb:56:10 (capab=0x411 status=0 aid=1)
[13201.743136] wlan0: associated
[13206.311797] wlan0: deauthenticating from 00:0a:79:eb:56:10 by local choice (reason=3)
[13206.322205] cfg80211: Calling CRDA to update world regulatory domain
[13211.304596] wlan0: authenticate with 00:0a:79:eb:56:10 (try 1)
[13211.306947] wlan0: authenticated
[13211.306997] wlan0: associate with 00:0a:79:eb:56:10 (try 1)
[13211.310680] wlan0: RX AssocResp from 00:0a:79:eb:56:10 (capab=0x411 status=0 aid=1)
[13211.310688] wlan0: associated
[13224.320636] iwlagn 0000:06:00.0: iwlagn_tx_agg_start on ra = 00:0a:79:eb:56:10 tid = 0
[13227.926481] iwlagn 0000:06:00.0: low ack count detected, restart firmware
[13227.926493] iwlagn 0000:06:00.0: On demand firmware reload
[13227.972437] iwlagn 0000:06:00.0: Stopping AGG while state not ON or starting
[13227.972449] iwlagn 0000:06:00.0: queue number out of range: 0, must be 10 to 19
[13239.312276] wlan0: deauthenticating from 00:0a:79:eb:56:10 by local choice (reason=3)
[13239.340079] cfg80211: Calling CRDA to update world regulatory domain
[13311.462143] ------------[ cut here ]------------
[13311.462183] WARNING: at net/wireless/core.c:633 wdev_cleanup_work+0x52/0xbb [cfg80211]()
[13311.462189] Hardware name: VGN-Z11VN_B
[13311.462193] Modules linked in: vboxnetadp vboxnetflt sco bnep rfcomm l2cap crc16 hso binfmt_misc dm_crypt dm_mod isofs btrfs zlib_deflate crc32c libcrc32c vfat fat fuse vboxdrv loop uinput snd_hda_codec_realtek snd_hda_intel snd_hda_codec snd_hwdep snd_pcm_oss snd_mixer_oss snd_pcm snd_seq_dummy snd_seq_oss snd_seq_midi snd_rawmidi snd_seq_midi_event snd_seq arc4 btusb sony_laptop bluetooth tpm_infineon snd_timer snd_seq_device snd firewire_ohci firewire_core crc_itu_t joydev iwlagn(-) iwlcore mac80211 cfg80211 rfkill soundcore snd_page_alloc
[13311.462294] Pid: 1052, comm: cfg80211 Not tainted 2.6.35-rc3 #29
[13311.462299] Call Trace:
[13311.462315] [<ffffffff81033d29>] warn_slowpath_common+0x80/0x98
[13311.462325] [<ffffffff81033d56>] warn_slowpath_null+0x15/0x17
[13311.462339] [<ffffffffa001958e>] wdev_cleanup_work+0x52/0xbb [cfg80211]
[13311.462350] [<ffffffff81044103>] worker_thread+0x15d/0x1f3
[13311.462364] [<ffffffffa001953c>] ? wdev_cleanup_work+0x0/0xbb [cfg80211]
[13311.462373] [<ffffffff810479b9>] ? autoremove_wake_function+0x0/0x34
[13311.462385] [<ffffffff8136d181>] ? _raw_spin_unlock_irqrestore+0x20/0x2b
[13311.462394] [<ffffffff81043fa6>] ? worker_thread+0x0/0x1f3
[13311.462401] [<ffffffff810475d7>] kthread+0x7a/0x82
[13311.462410] [<ffffffff81002d94>] kernel_thread_helper+0x4/0x10
[13311.462418] [<ffffffff8104755d>] ? kthread+0x0/0x82
[13311.462426] [<ffffffff81002d90>] ? kernel_thread_helper+0x0/0x10
[13311.462432] ---[ end trace b1cb7d121e67bde3 ]---
[13311.521571] iwlagn 0000:06:00.0: PCI INT A disabled
[13316.830454] iwlagn: Intel(R) Wireless WiFi Link AGN driver for Linux, in-tree:
[13316.830457] iwlagn: Copyright(c) 2003-2010 Intel Corporation
[13316.830552] iwlagn 0000:06:00.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17
[13316.830583] iwlagn 0000:06:00.0: setting latency timer to 64
[13316.831777] iwlagn 0000:06:00.0: Detected Intel(R) WiFi Link 5100 AGN, REV=0x54
[13316.854135] iwlagn 0000:06:00.0: Tunable channels: 13 802.11bg, 24 802.11a channels
[13316.854251] iwlagn 0000:06:00.0: irq 46 for MSI/MSI-X
[13316.858767] iwlagn 0000:06:00.0: loaded firmware version 8.24.2.12
[13316.859883] phy1: Selected rate control algorithm 'iwl-agn-rs'
[13328.516652] wlan0: authenticate with 00:0a:79:eb:56:10 (try 1)
[13328.519137] wlan0: authenticated
[13328.519181] wlan0: associate with 00:0a:79:eb:56:10 (try 1)
[13328.522937] wlan0: RX AssocResp from 00:0a:79:eb:56:10 (capab=0x411 status=0 aid=1)
[13328.522945] wlan0: associated
[13341.695599] iwlagn 0000:06:00.0: iwlagn_tx_agg_start on ra = 00:0a:79:eb:56:10 tid = 0
[13733.621786] iwlagn 0000:06:00.0: Received BA when not expected
[13733.621800] iwlagn 0000:06:00.0: Read index for DMA queue txq id (0), index 15, is out of range [0-256] 17 17.
[13734.188691] iwlagn 0000:06:00.0: Received BA when not expected
[13734.188704] iwlagn 0000:06:00.0: Read index for DMA queue txq id (0), index 16, is out of range [0-256] 18 18.

and here it continues with many many many of those two lines. The only
way to resurrect the wifi card is unloading iwlagn and reloading it.

I have no idea where to go from here, I always thought that WARNINGS and
serious errors like that (complete drop of connection in irregular
intervals, sometimes several times in few minutes, sometimesonce an
hour) would trigger some action.

Hope someone is picking that up, thanks.

Best wishes

Norbert

----------------------------------------------------------------------------
Norbert Preining preining@{jaist.ac.jp, logic.at, debian.org}
JAIST, Japan TeX Live and Debian Developer
gpg DSA: 0x09C5B094 fp: 14DF 2E6C 0307 BE6D AD76 A9C0 D2BF 4AA3 09C5 B094
----------------------------------------------------------------------------
AGGLETHORPE (n.)
A dispute between two pooves in a boutique.
--- Douglas Adams, The Meaning of Liff
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Norbert Preining on 14 Jun 2010 12:30
On Mo, 14 Jun 2010, Christoph Hellwig wrote:
> I see the same issue, and I even tried the git bisect, but due to
> various issues in early boot / ACPI code in that merge window it's

Me, too, but as you said, hard to finish, it is getting weird when we
come down to the merge window.

Is there a way to restrict git bisect to consider *only* changes in
a specific sub-directory (directories) ? That would help.

> impossible to finish it. The ugly workaround for me is to always keep
> a ping running in the background, which keeps it from dropping the

I had the feeling that this helped a bit, but didn't explore it
in an organized way. Thanks for the tip.

Let's hope for the best

Norbert
------------------------------------------------------------------------
Norbert Preining preining@{jaist.ac.jp, logic.at, debian.org}
JAIST, Japan TeX Live & Debian Developer
DSA: 0x09C5B094 fp: 14DF 2E6C 0307 BE6D AD76 A9C0 D2BF 4AA3 09C5 B094
------------------------------------------------------------------------
`Er, hey Earthman...'
`Arthur,' said Arthur.
`Yeah, could you just sort of keep this robot with you and
guard this end of the passageway. OK?'
What from? You just said there's no
one here.'
`Yeah, well, just for safety, OK?' said Zaphod.
`Whose? Yours or mine?'
--- Arthur drawing the short straw on Magrathea.
--- Douglas Adams, The Hitchhikers Guide to the Galaxy
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Christoph Hellwig on 14 Jun 2010 12:30
On Tue, Jun 15, 2010 at 01:09:10AM +0900, Norbert Preining wrote:
> Dear all,
>
> (please cc)
>
> Since AFAIR 2.6.32-rc the iwlagn driver seems to be completely hosed
> with some AP. Permanently dropping connections, hanging, BUGing.
> Unfortunately the only reply I get from iwl maintainers is "do git bisect"
> which is a bit a pain with someting that is not stable reproducible.

I see the same issue, and I even tried the git bisect, but due to
various issues in early boot / ACPI code in that merge window it's
impossible to finish it. The ugly workaround for me is to always keep
a ping running in the background, which keeps it from dropping the
connection. Fortunately the BUGs only happen very rarely for me (and
unfortunately I'm not able to capture them, but I suspect it's iwlagn)

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Nick Bowler on 14 Jun 2010 12:40
On 01:24 Tue 15 Jun , Norbert Preining wrote:
> Is there a way to restrict git bisect to consider *only* changes in
> a specific sub-directory (directories) ? That would help.

Yes, there is an optional path argument to git bisect start which does
this, covered in the git-bisect man page.

--
Nick Bowler, Elliptic Technologies (http://www.elliptictech.com/)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: reinette chatre on 14 Jun 2010 13:00
Hi Norbert,

On Mon, 2010-06-14 at 09:09 -0700, Norbert Preining wrote:
> Dear all,
>
> (please cc)

Since you are a regular poster to the wireless lists and you also
submitted a bug report
(http://bugzilla.intellinuxwireless.org/show_bug.cgi?id=2214 ) for this
issue it does not seem as though you need help in sending this to the
appropriate people.

> [13341.695599] iwlagn 0000:06:00.0: iwlagn_tx_agg_start on ra = 00:0a:79:eb:56:10 tid = 0
> [13733.621786] iwlagn 0000:06:00.0: Received BA when not expected
> [13733.621800] iwlagn 0000:06:00.0: Read index for DMA queue txq id (0), index 15, is out of range [0-256] 17 17.
> [13734.188691] iwlagn 0000:06:00.0: Received BA when not expected
> [13734.188704] iwlagn 0000:06:00.0: Read index for DMA queue txq id (0), index 16, is out of range [0-256] 18 18.
>
> and here it continues with many many many of those two lines. The only
> way to resurrect the wifi card is unloading iwlagn and reloading it.

.... as I mention above you already submitted a bug report for this issue
and we directed you to a patch that addresses this. You have not replied
to that suggestion with any testing, instead you posted this message on
lkml.

>
> I have no idea where to go from here,

How about you test the patch we proposed and work with the developers
working on your bug report?

Reinette


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
 |  Next  |  Last
Pages: 1 2
Prev: soc/sh: Fix I2C dependency for SND_FSI_AK4642
Next: New issues in 2.6.35-rc{1,2}