Modify

Opened 7 years ago

Closed 7 years ago

#9460 closed defect (fixed)

ath5k on ar71xx crashes

Reported by: Nenad Opsenica <nenad@…> Owned by: developers
Priority: normal Milestone:
Component: packages Version: Backfire 10.03.1 RC4
Keywords: Cc: bernard@…

Description

ath5k (latest backfire svn) is not working stable on ar71xx (Ubiquiti Routerstation, Atheros AR7161 rev 2). In fact, it depends on miniPCI hardware model, for example:

  • Wistron CM9 as STA, after issuing iw wlan0 scan:
    wlan0: associate with 00:0b:6b:35:77:51 (try 1)
    wlan0: RX AssocResp from 00:0b:6b:35:77:51 (capab=0x1 status=0 aid=1)
    wlan0: associated
    ieee80211 phy0: WMM queue=2 aci=0 acm=0 aifs=3 cWmin=15 cWmax=1023 txop=0 uapsd=0
    ieee80211 phy0: WMM queue=3 aci=1 acm=0 aifs=7 cWmin=15 cWmax=1023 txop=0 uapsd=0
    ieee80211 phy0: WMM queue=1 aci=2 acm=0 aifs=2 cWmin=7 cWmax=15 txop=94 uapsd=0
    ieee80211 phy0: WMM queue=0 aci=3 acm=0 aifs=2 cWmin=3 cWmax=7 txop=47 uapsd=0
    br-lan: port 2(wlan0) entering forwarding state
    PCI error 1 at PCI addr 0x10009930
    Data bus error, epc == 8308af40, ra == 8308af40
    Oops[#1]:
    Cpu 0
    $ 0   : 00000000 80310000 deadc0de 00000002
    $ 4   : b0009930 0000000a 000000e2 19d60000
    $ 8   : 00005848 800643e4 00000000 802ce000
    $12   : 802d9f28 00000000 000000cf 00000000
    $16   : 83898000 00000002 83b70200 83898000
    $20   : 00000000 00000001 00000003 83b71284
    $24   : 00000010 8309659c                 
    $28   : 8396a000 8396bd48 00000000 8308af40
    Hi    : 0000000a
    Lo    : 17980000
    epc   : 8308af40 ath5k_hw_gainf_calibrate+0x50/0x4a4 [ath5k]
        Not tainted
    ra    : 8308af40 ath5k_hw_gainf_calibrate+0x50/0x4a4 [ath5k]
    Status: 1000f403    KERNEL EXL IE
    Cause : 1080001c
    PrId  : 00019374 (MIPS 24Kc)
    Modules linked in: ath5k ath mac80211 cfg80211 compat_firmware_class compat arc4 aes_generic deflate ecb cbc leds_gpio button_he
    Process phy0 (pid: 501, threadinfo=8396a000, task=838568c8, tls=00000000)
    Stack : 0000000a 83898000 00000000 00000001 00000003 83084e88 00000000 00000013
            83b71284 8396bd90 83b70200 83898000 00000000 8308d750 00000000 00000000
            83898000 a3bf4f60 00000a16 0000000b 00000000 83b73e0c 00000000 00000000
            00000000 00000000 8389ab88 83b73e18 00008000 00000001 00000054 00000000
            00000000 83020280 83b71284 00000001 83b70200 83b70d40 83898000 83b70d54
            ...
    Call Trace:
    [<8308af40>] ath5k_hw_gainf_calibrate+0x50/0x4a4 [ath5k]
    [<8308d750>] ath5k_hw_reset+0x43c/0x10e4 [ath5k]
    [<830919b4>] ath5k_update_bssid_mask_and_opmode+0x30c/0x5c8 [ath5k]
    
    
    Code: 34049930  0c05c21f  00442021 <00401821> 30428000  144000ff  00000000  7c620dc0  00031e42
    Disabling lock debugging due to kernel taint
    
  • Ubiquiti SR5 as STA: associates to AP, data transfer works, after turning off AP and issuing iw wlan0 scan:
    command failed: Device or resource busy (-16)
    
    

If STA is associated to AP, iw wlan0 scan works as expected

Attachments (0)

Change History (9)

comment:1 Changed 7 years ago by anonymous

iw wlan0 scan will start a new scan only when the hardware is not already scanning. If it is scanning, try iw wlan0 scan dump instead to see the cached scan results.

comment:2 Changed 7 years ago by Nenad Opsenica <nenad@…>

Update: after stumbling upon http://comments.gmane.org/gmane.linux.kernel.wireless.general/69032, I have applied patch /changeset/27024.html and now both kinds of CM9 MiniPCI are working correctly - older and newer ones (RoHS).

comment:3 Changed 7 years ago by Nenad Opsenica <nenad@…>

Update 2: After running for >24h, oops happened again. It seems that with patch commit 27024 platform is more stable, but problem is not solved.

PCI error 1 at PCI addr 0x10009930
Data bus error, epc == 8738af40, ra == 8738af40
Oops[#1]:
Cpu 0
$ 0   : 00000000 80310000 deadc0de 00000002
$ 4   : b0009930 0000000a 000000ef 354e0000
$ 8   : 00005d5c 800643e4 00000000 802ce000
$12   : 802d9f28 00000000 000000cf 00000000
$16   : 87862000 00000002 873b0200 87862000
$20   : 00000001 873b0d54 00000003 873b0f48
$24   : 00000003 87396470
$28   : 87252000 87253d48 00000002 8738af40
Hi    : 00000014
Lo    : d2780000
epc   : 8738af40 ath5k_hw_gainf_calibrate+0x50/0x4a4 [ath5k]
    Not tainted
ra    : 8738af40 ath5k_hw_gainf_calibrate+0x50/0x4a4 [ath5k]
Status: 1000fc03    KERNEL EXL IE
Cause : 1080001c
PrId  : 00019374 (MIPS 24Kc)
Modules linked in: ath5k ath mac80211 cfg80211 compat_firmware_class compat pl2303 ark3116 ums_usbat ums_sddr55 ums_sddr09 ume
Process phy0 (pid: 1085, threadinfo=87252000, task=872168c8, tls=00000000)
Stack : 0000000a 87862000 00000001 873b0d54 00000003 87384e88 00000000 00000013
        873b0f48 87253d90 873b0200 87862000 00000001 8738d624 0000000a 87394c94
        87862000 a7ad4500 000009d3 8738f9cc 00000000 873b3e0c 00000000 00000000
        00000000 00000000 878643c0 873b3e18 00008000 00000001 00000034 00000000
        00000000 00000018 873b0f48 00000001 873b0200 873b0d40 87862000 873b0d54
        ...
Call Trace:
[<8738af40>] ath5k_hw_gainf_calibrate+0x50/0x4a4 [ath5k]
[<8738d624>] ath5k_hw_reset+0x310/0xfb8 [ath5k]
[<87391888>] ath5k_update_bssid_mask_and_opmode+0x30c/0x5c8 [ath5k]


Code: 34049930  0c05c493  00442021 <00401821> 30428000  144000ff  00000000  7c620dc0  00031e42
Disabling lock debugging due to kernel taint

comment:4 Changed 7 years ago by Nenad Opsenica <nenad@…>

Crash from previous post happened with ath5k configured as STA, approximately 1.5 min after AP was shutted down:

Jun 27 10:20:07 OpenWrt user.debug kernel: wlan0: authentication with 00:0b:6b:4d:51:45 timed out
Jun 27 10:21:43 OpenWrt user.alert kernel: PCI error 1 at PCI addr 0x10009930
Jun 27 10:21:43 OpenWrt user.alert kernel: Data bus error, epc == 8710af40, ra == 8710af40
Jun 27 10:21:43 OpenWrt user.warn kernel: Oops[#1]:
Jun 27 10:21:43 OpenWrt user.warn kernel: Cpu 0
Jun 27 10:21:43 OpenWrt user.warn kernel: $ 0   : 00000000 80310000 deadc0de 00000002
Jun 27 10:21:43 OpenWrt user.warn kernel: $ 4   : b0009930 0000000a 000000ef 354e0000
Jun 27 10:21:43 OpenWrt user.warn kernel: $ 8   : 00005d5c 800643e4 00000000 802ce000
Jun 27 10:21:43 OpenWrt user.warn kernel: $12   : 802d9f28 00000000 000000cf 00000000
Jun 27 10:21:43 OpenWrt user.warn kernel: $16   : 87860000 00000002 87130200 87860000
Jun 27 10:21:43 OpenWrt user.warn kernel: $20   : 00000001 87130d54 00000003 87130e94
Jun 27 10:21:43 OpenWrt user.warn kernel: $24   : 00000003 87116470

comment:5 Changed 7 years ago by nbd

please try latest trunk svn

comment:6 Changed 7 years ago by nbd

Also, please post the part in dmesg where it shows the chip revisions of your card

comment:7 Changed 7 years ago by Nenad Opsenica <nenad@…>

It also happens with mac80211 package from latest trunk svn. Ath5k is also configured as STA. It sometimes happens immediately after device boot-up when STA is not able to associate to the AP, but happens always after STA is associated and then AP is turnded off.

Chip data:

PCI: Enabling device 0000:00:12.0 (0000 -> 0002)
ath5k 0000:00:12.0: registered as 'phy0'
ath5k phy0: Atheros AR5213A chip found (MAC: 0x59, PHY: 0x43)
ath5k phy0: RF5112B multiband radio found (0x36)
# lspci -vvv

00:12.0 Ethernet controller: Atheros Communications Inc. Atheros AR5001X+ Wireless Network Adapter (rev 01)
        Subsystem: Wistron NeWeb Corp. CM9 Wireless a/b/g MiniPCI Adapter
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 168 (2500ns min, 7000ns max), Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 49
        Region 0: Memory at 10000000 (32-bit, non-prefetchable) [size=64K]
        Capabilities: [44] Power Management version 2
                Flags: PMEClk- DSI- D1- D2- AuxCurrent=375mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=2 PME-
        Kernel driver in use: ath5k



Kernel oops is same as before.

comment:8 Changed 7 years ago by Nenad Opsenica <nenad@…>

With patch http://nbd.name/446-ath5k_disable_32khz_clock.patch from #9458 this device is now running for 5 hours without a crash.

comment:9 Changed 7 years ago by nbd

  • Resolution set to fixed
  • Status changed from new to closed

Add Comment

Modify Ticket

Action
as closed .
The resolution will be deleted. Next status will be 'reopened'.
Author


E-mail address and user name can be saved in the Preferences.

 
Note: See TracTickets for help on using tickets.