Bug 48131 - ieee80211 phy0: brcms_c_prec_enq_head: No where to go, prec == 4
Summary: ieee80211 phy0: brcms_c_prec_enq_head: No where to go, prec == 4
Status: CLOSED CODE_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: network-wireless (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: brcm80211 driver developers
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-10-01 16:29 UTC by Stanislaw Gruszka
Modified: 2013-01-24 14:28 UTC (History)
3 users (show)

See Also:
Kernel Version: 3.5
Subsystem:
Regression: No
Bisected commit-id:


Attachments
/my-v3.5.config (121.14 KB, text/plain)
2012-10-01 16:32 UTC, Stanislaw Gruszka
Details
dmesg-3.6-patched.txt (272.45 KB, text/plain)
2012-10-18 09:59 UTC, Stanislaw Gruszka
Details
dmesg-3.6-patched-timestamps.txt (208.70 KB, text/plain)
2012-10-22 11:39 UTC, Stanislaw Gruszka
Details
nowhere to go patch 1 (7.25 KB, patch)
2012-10-24 11:11 UTC, Piotr Haber
Details | Diff
dmesg.patched.txt (371.56 KB, text/plain)
2012-10-30 15:31 UTC, Stanislaw Gruszka
Details
full_dmesg.patched.txt.gz (684.58 KB, application/x-gzip)
2012-10-30 16:00 UTC, Stanislaw Gruszka
Details
Rework tx code to avoid internal buffering of packets (61.88 KB, patch)
2012-10-31 00:23 UTC, Aleksandr Mezin
Details | Diff

Description Stanislaw Gruszka 2012-10-01 16:29:42 UTC
I have tons of 'brcms_c_prec_enq_head: No where to go, prec == 4' messages in dmesg, practically they overwrite all other messages from the buffer:

> [root@dhcp-31-100 ~]# dmesg | grep -c 'brcms_c_prec_enq_head: No where to go,
> prec == 4'
> 4226

This happen when downloading file using wget through WPA2-enterprise 11g network.
Comment 1 Stanislaw Gruszka 2012-10-01 16:32:26 UTC
Created attachment 81791 [details]
/my-v3.5.config

Kernel config file.
Comment 2 Stanislaw Gruszka 2012-10-01 16:33:44 UTC
Adapter:

a0:00.0 Network controller [0280]: Broadcom Corporation BCM43224 802.11a/b/g/n [14e4:4353] (rev 01)
	Subsystem: Dell Wireless 1520 Half-size Mini PCIe Card [1028:000e]
	Physical Slot: 5
	Flags: bus master, fast devsel, latency 0, IRQ 26
	Memory at f3600000 (64-bit, non-prefetchable) [size=16K]
	Capabilities: <access denied>
	Kernel driver in use: bcma-pci-bridge
	Kernel modules: bcma
Comment 3 Stanislaw Gruszka 2012-10-18 09:59:49 UTC
Created attachment 83821 [details]
dmesg-3.6-patched.txt

Dmesg from 3.6 kernel with patch from
https://bugzilla.kernel.org/show_bug.cgi?id=48391#c16

There are still lots of "No where to go" messages and when NM periodic scan trigger there are additional WARNINGs.
Comment 4 Stanislaw Gruszka 2012-10-18 10:04:05 UTC
Some more detailed network parameters:

WPA2-Enterprise, authentication: PEAP v1, inner authentication: GTC.

Corresponding "iw dev wlan0 scan" output:

BSS 00:3a:98:a0:30:51 (on wlan8) -- associated
	TSF: 594862694797 usec (6d, 21:14:22)
	freq: 2412
	beacon interval: 100
	capability: ESS Privacy ShortPreamble ShortSlotTime (0x0431)
	signal: -49.00 dBm
	SSID: Red Hat
	Supported rates: 1.0* 2.0* 5.5* 6.0 9.0 11.0* 12.0 18.0 
	DS Parameter set: channel 1
	Country: CZ (in/outdoor), data: 01 0d 14
	ERP: <no flags>
	RSN:	 * Version: 1
		 * Group cipher: CCMP
		 * Pairwise ciphers: CCMP
		 * Authentication suites: IEEE 802.1X 00-40-96:0
		 * Capabilities: 4-PTKSA-RC 4-GTKSA-RC (0x0028)
	Extended supported rates: 24.0 36.0 48.0 54.0 
	WMM:	* Parameter version 1
		* u-APSD
		* BE: CW 15-1023, AIFSN 3
		* BK: CW 15-1023, AIFSN 7
		* VI: CW 7-15, AIFSN 2, TXOP 3008 usec
		* VO: acm CW 3-7, AIFSN 2, TXOP 1504 usec
Comment 5 Piotr Haber 2012-10-22 09:26:22 UTC
does this only happen when using WPA2-Enterprise?
what throughput are you reaching? 
could you post log with timestamps so we can judge how long this packet dropping lasts?
Comment 6 Stanislaw Gruszka 2012-10-22 11:38:24 UTC
"No where to go" messages occurs _sporadically_ on some other networks too. I did not get "WARNING at brcms_ops_flush" on other network.

Throughput is about 1MBytes/s , it decrease to about 500kBytes/s when errors start to show up in dmesg.
Comment 7 Stanislaw Gruszka 2012-10-22 11:39:32 UTC
Created attachment 84261 [details]
dmesg-3.6-patched-timestamps.txt
Comment 8 Piotr Haber 2012-10-24 11:11:50 UTC
Created attachment 84651 [details]
nowhere to go patch 1

i cannot really reproduce such a flood of messages
here is a patch that modifies packet dropping policy 
could you try it?
preferably with one from Bug #48391
Comment 9 Stanislaw Gruszka 2012-10-30 15:31:08 UTC
Created attachment 85311 [details]
dmesg.patched.txt

Dmesg from 3.7-rc3 kernel with nowhere.patch and ops_flush.patch .
Comment 10 Stanislaw Gruszka 2012-10-30 16:00:48 UTC
Created attachment 85321 [details]
full_dmesg.patched.txt.gz

Since dmesg wrapped out, I'm attaching full log started from brcmsmac module load.
Comment 11 Aleksandr Mezin 2012-10-31 00:23:34 UTC
Created attachment 85331 [details]
Rework tx code to avoid internal buffering of packets

I have:
0d:00.0 Network controller: Broadcom Corporation BCM4313 802.11b/g/n Wireless LAN Controller (rev 01)

When these "Nowhere to go" errors started to appear, not only performance decreased, but connection stopped to work at all.

It seems that this patch fixes the problem: https://patchwork.kernel.org/patch/1651801/
I applied it to kernel 3.6.4. Now I don't see these "Nowhere to go" errors. Downloaded 4GB iso image using Transmission and didn't notice any problems.
Comment 12 Stanislaw Gruszka 2013-01-24 14:28:11 UTC
I confirm the fix, I'm not able to reproduce the problem any longer.

Note You need to log in before you can comment on or make changes to this bug.