Ticket #292 (closed defect: fixed)

Opened 6 years ago

Last modified 6 years ago

Kernel panic madwifi-ng ad-hoc on high throughput

Reported by: dries.naudts@intec.ugent.be Assigned to:
Priority: blocker Milestone:
Component: madwifi: driver Version: trunk
Keywords: kernel panic, ad hoc Cc:
Patch is attached: 0 Pending:

Description

I'm using madwifi-ng from svn revision: 1364 OS: Linux 2.4.27 on mipsel architecture (AccessCube? from 4G systems). Two devices with each an atheros card configured in ad hoc mode 802.11a. When trying to run iperf between these two nodes I get a kernel panic.

Output:

bdc, ra == c01f01d09bc020 00000000 000
Oops in fault.c::do_page_fault, line 206:
                                         
$0 : 00000000 80000000 00000000 cccccccd 00000000 00000000 00000038 000050ebth_rate_sample: ath_rate_ctl_reset 00:00:00:00:00:00 no rates (fixed -1)(fix
$8 : 00000008 0000003c 00000000 00000002 33323130 100043f0 00000000 00000000                                                                            
$16: 83959160 803a0a40 8301f000 8301d020 c01dcbf8 00000000 83768000 00000008ath_rate_ctl_reset 00:00:00:00:00:00 no rates (fixed -1)s yetd -1)0 0       
$24: 00000000 2ae83650                   8312a000 8312bd30 000000ff c01f01d0
  
ath_beacon_setup: m 83728500 len 117                                    

Hi : 0000bd8cate_         
Lo : 00002f63             
epc   : c01bebdc    Tainted: P                              
Process iperf (pid: 1124, stackpage=8312a000)                                             
Stack:    000005da 83751740 8022d7d0 8312bdf8 8024fb40 8024fb7c 83430be0    

ath_rate_sample: 00:0e:35:b6:10:c7 ath_rate_tx_complete no rates y
 8020ac48 ffeb93d8 000000bc 8312be0c 7d1c114c 83818160 803a0a40 83778300                                                                        
 a381a300 8301f000 00000001 83768000 00000008 83818000 c01f01d0 00007f40mple: ath_rate_ctl_reset 00:00:00:00:00:00 no rates (fixed -1)(fixed -1)
 0000001c 00000038 000050eb 83818000 83818160 83768000 00000005 00000000                                                                        
 00000001 fffffff7 1000fc00 8312be58 10004388 00ff0000 ff000000 00000003m 83728500 len 117                                    
ath_rate_sample: 
 80116834 ...eset 00:00:00
Call Trace:   [<8022d7d0>] [<8024fb40>] [                                        
 [<80289ff0>] [<801b5a4c>] [<801b5660>] [<802898c8>] [<80115ba4>] [<80115be4>]0 no rates (fixed -1)tus 0                                                    
 [<80107640>] [<801046a8>]                          

Code: 31020003  14400484  310500f0 <8ea20048> 92630001  30420080  afa5002c  3077:00 no rates (fixed -1)s yetd -1): 00:0e:35:0b:e4:cc ath_rate_tx_complete no rat
0003  1440002d              
Kernel panic: Aiee, killing interrupt handler!                                              
In interrupt handler - not syncing    
ath_rate_sample: 00:0e:35:0b:

                                  

Any idea?

Change History

01/09/06 10:14:58 changed by mrenzmann

Please try again with the latest revision of madwifi-ng. If the oops is reproducable there, please use ksymoops on the oops, so that it contains more human-readable information. Thanks.

01/09/06 10:41:04 changed by mrenzmann

Probably related: #248.

03/02/06 22:33:34 changed by Vikram

  • priority changed from major to blocker.
  • patch_attached changed.

I have noticed this in the r1454 release. Not sure if it existed before that. I set up ad-hoc mode in three boxes and run netperf in some combinations. I did this two different times, and the following type of error occured (as extracted from /var/log/messages)

Notice the "ath_uapsd_processtriggers: no skbuff" message was then repeated numerous times; although it says that mysql was terminated. The other machine terminated more programs including mDNSResponser, xfs, X, sendmail, and others

Mar  2 13:12:04 ar83-176 kernel: oom-killer: gfp_mask=0xd2
Mar  2 13:12:04 ar83-176 kernel: Mem-info:
Mar  2 13:12:04 ar83-176 kernel: DMA per-cpu:
Mar  2 13:12:04 ar83-176 kernel: cpu 0 hot: low 2, high 6, batch 1
Mar  2 13:12:04 ar83-176 kernel: cpu 0 cold: low 0, high 2, batch 1
Mar  2 13:12:04 ar83-176 kernel: Normal per-cpu:
Mar  2 13:12:04 ar83-176 kernel: cpu 0 hot: low 62, high 186, batch 31
Mar  2 13:12:04 ar83-176 kernel: cpu 0 cold: low 0, high 62, batch 31
Mar  2 13:12:04 ar83-176 kernel: HighMem per-cpu: empty
Mar  2 13:12:04 ar83-176 kernel: 
Mar  2 13:12:04 ar83-176 kernel: Free pages:        5096kB (0kB HighMem)
Mar  2 13:12:04 ar83-176 kernel: Active:517 inactive:905 dirty:0 writeback:229 u
nstable:0 free:1274 slab:121087 mapped:23 pagetables:723
Mar  2 13:12:04 ar83-176 kernel: DMA free:2068kB min:88kB low:108kB high:132kB a
ctive:4kB inactive:0kB present:16384kB pages_scanned:16 all_unreclaimable? yes
Mar  2 13:12:07 ar83-176 kernel: lowmem_reserve[]: 0 495 495
Mar  2 13:12:09 ar83-176 kernel: Normal free:3028kB min:2800kB low:3500kB high:4
200kB active:2064kB inactive:3620kB present:507264kB pages_scanned:525 all_unrec
laimable? no
Mar  2 13:12:10 ar83-176 kernel: lowmem_reserve[]: 0 0 0
Mar  2 13:12:11 ar83-176 kernel: HighMem free:0kB min:128kB low:160kB high:192kB
 active:0kB inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
Mar  2 13:12:11 ar83-176 kernel: lowmem_reserve[]: 0 0 0
Mar  2 13:12:12 ar83-176 kernel: DMA: 1*4kB 0*8kB 1*16kB 0*32kB 0*64kB 0*128kB 0
*256kB 0*512kB 0*1024kB 1*2048kB 0*4096kB = 2068kB
Mar  2 13:12:12 ar83-176 kernel: Normal: 37*4kB 10*8kB 5*16kB 3*32kB 5*64kB 0*12
8kB 1*256kB 2*512kB 1*1024kB 0*2048kB 0*4096kB = 3028kB
Mar  2 13:12:13 ar83-176 kernel: HighMem: empty
Mar  2 13:12:14 ar83-176 kernel: Swap cache: add 23379, delete 23115, find 488/6
58, race 0+0
Mar  2 13:12:14 ar83-176 kernel: Free swap  = 959980kB
Mar  2 13:12:14 ar83-176 kernel: Total swap = 1048568kB
Mar  2 13:12:15 ar83-176 kernel: Free swap:       959980kB
Mar  2 13:12:15 ar83-176 kernel: 130912 pages of RAM
Mar  2 13:12:16 ar83-176 kernel: 0 pages of HIGHMEM
Mar  2 13:12:17 ar83-176 kernel: 2354 reserved pages
Mar  2 13:12:17 ar83-176 kernel: 4313 pages shared
Mar  2 13:12:18 ar83-176 kernel: 264 pages swap cached
Mar  2 13:12:19 ar83-176 kernel: Out of Memory: Killed process 2719 (mysqld).
Mar  2 13:12:20 ar83-176 kernel: ath_uapsd_processtriggers: no skbuff

I am using: Linux version 2.6.11-1.1369_FC4 and gcc version 4.0.0 20050525. The details of the card are

ath_hal: module license 'Proprietary' taints kernel.
ath_hal: 0.9.16.16 (AR5210, AR5211, AR5212, RF5111, RF5112, RF2413, RF5413)
wlan: 0.8.4.2 (svn 1454)
ath_rate_sample: 1.2 (svn 1454)
ath_pci: 0.9.4.5 (svn 1454)
PCI: Enabling device 0000:03:00.0 (0000 -> 0002)
ACPI: PCI Interrupt 0000:03:00.0[A] -> Link [LNKA] -> GSI 11 (level, low) -> IRQ 11
wifi0: 11a rates: 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps
wifi0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps
wifi0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps
wifi0: H/W encryption support: WEP AES AES_CCM TKIP
wifi0: mac 5.9 phy 4.3 radio 3.6
wifi0: Use hw queue 1 for WME_AC_BE traffic
wifi0: Use hw queue 0 for WME_AC_BK traffic
wifi0: Use hw queue 2 for WME_AC_VI traffic
wifi0: Use hw queue 3 for WME_AC_VO traffic
wifi0: Use hw queue 8 for CAB traffic
wifi0: Use hw queue 9 for beacons
wifi0: Atheros 5212: mem=0x20800000, irq=11

03/06/06 05:41:59 changed by anonymous

I have found the reson for the kernel panic is from a memory leak stated in ticket 343. Because of the high thouroughput the memory of the system will fill up quiet quickly and cause the memory to overflow causing the panic. Let me know if this is the same problem.

03/07/06 07:06:51 changed by mrenzmann

#343 is a ticket about a problem related to "wlanconfig: ioctl: No such device"?

But since this ticket is mentioning ad-hoc mode: can someone please check if the problem still exists in/after r1458?

03/10/06 06:38:25 changed by daniel@franke.name

Appears so -- I'm having a problem that I'm pretty sure is related. I'm running madwifi-ng on a garden-variety P4 desktop system under Debian Stable. The card is running in AP mode. When I start doing a large file transfer, I watch my free memory slowly disappear, and then after a while the system freezes up completely; not even SysRq? works.

03/25/06 17:17:23 changed by georg@boerde.de

There is another ad-hoc memory leak related fix in r1471 (see ticket #441 for details). Could you please recheck with that (or a later) revision?

04/20/06 16:13:42 changed by mrenzmann

  • status changed from new to closed.
  • resolution set to fixed.

No response from reporting parties for nearly one month, so this ticket will be closed. Feel free to reopen the ticket if the problem still exists in current revisions. Thanks.