Skip to contentSkip to conversation

ath10k hits warning in sta_info.c:839.

DISCUSSION Ben Greear 3 participants 5 messages Oct 22, 2013 – Oct 28, 2013 Qualcomm Atheros ATH10K wireless driver

lore
Share & tools
lore

Ben Greear

author
#0 lore raw

Kernel is stock 'ath' tree, with small printk to debug an ath10k crash.

This is FYI for now...will be looking at other ath10k crash bugs before digging into this tone.

Setup is 2 stations trying to associate to same AP, which causes endless failures and firmware crashes. Good for chasing bugs :)

DMAR:[fault reason 05] PTE Write access is not set
dmar: DRHD: handling fault status reg 3
dmar: DMAR:[DMA Write] Request device [05:00.0] fault addr ffd52000 DMAR:[fault reason 05] PTE Write access is not set
dmar: DRHD: handling fault status reg 3
dmar: DMAR:[DMA Write] Request device [05:00.0] fault addr ffd52000 DMAR:[fault reason 05] PTE Write access is not set

sta300: authentication with 00:03:83:3d:30:aa timed out [root@ct523-9292 ~]# ath10k: Failed to delete peer: 00:03:83:3d:30:aa for VDEV: 1
ath10k: WMI vdev stop failed: ret -108

------------[ cut here ]------------
WARNING: CPU: 1 PID: 6 at /mnt/sda/home/greearb/git/ath/net/mac80211/sta_info.c:839 __sta_info_destroy+0x12)
Modules linked in: nf_nat_ipv4 nf_nat veth 8021q garp stp mrp llc macvlan pktgen lockd f71882fg coretemp hw]
CPU: 1 PID: 6 Comm: kworker/u8:0 Tainted: G        WC   3.12.0-rc5-wl+ #1
Hardware name: To be filled by O.E.M. To be filled by O.E.M./HURONRIVER, BIOS 4.6.5 05/02/2012
Workqueue: phy0 ieee80211_iface_work [mac80211]
 0000000000000009 ffff8802160d3c28 ffffffff81566b46 ffff88021fa8ec78
 0000000000000000 ffff8802160d3c68 ffffffff8109e0a2 ffff8802160d3c78
 ffffffffa03ceff2 ffff880215b63800 ffff8800d81505c0 ffff88020dc647c0
Call Trace:
 [<ffffffff81566b46>] dump_stack+0x55/0x86
 [<ffffffff8109e0a2>] warn_slowpath_common+0x77/0x91
 [<ffffffffa03ceff2>] ? __sta_info_destroy+0x122/0x1d6 [mac80211]
 [<ffffffff8109e0d1>] warn_slowpath_null+0x15/0x17

dmar: DRHD: handling fault status reg 3
dmar: DMAR:[DMA Write] Request device [05:00.0] fault addr ffd51000 DMAR:[fault reason 05] PTE Write access is not set
dmar: DRHD: handling fault status reg 3
dmar: DMAR:[DMA Write] Request device [05:00.0] fault addr ffd51000 DMAR:[fault reason 05] PTE Write access is not set

 [<ffffffffa03ceff2>] __sta_info_destroy+0x122/0x1d6 [mac80211]
 [<ffffffffa03cf27a>] sta_info_destroy_addr+0x37/0x59 [mac80211]
 [<ffffffffa03edbcc>] ieee80211_destroy_auth_data+0x2a/0x8c [mac80211]
 [<ffffffffa03f1543>] ieee80211_sta_work+0x159/0xb2c [mac80211]
 [<ffffffffa03ea936>] ? ieee80211_wake_queues_by_reason+0x78/0x87 [mac80211]
 [<ffffffff8156cc88>] ? _raw_spin_unlock_irqrestore+0x27/0x32
 [<ffffffffa03d86bf>] ieee80211_iface_work+0x29f/0x2c2 [mac80211]
 [<ffffffffa03edc3b>] ? sdata_unlock+0xd/0xf [mac80211]
 [<ffffffff810afdc1>] ? pwq_activate_delayed_work+0x23/0x31
 [<ffffffff810b1846>] process_one_work+0x162/0x217
 [<ffffffff810b1cbb>] worker_thread+0x12e/0x1fb
 [<ffffffff810b1b8d>] ? rescuer_thread+0x268/0x268
 [<ffffffff810b6a8e>] kthread+0x88/0x90
 [<ffffffff810b6a06>] ? __kthread_parkme+0x60/0x60
 [<ffffffff8157160c>] ret_from_fork+0x7c/0xb0
 [<ffffffff810b6a06>] ? __kthread_parkme+0x60/0x60
---[ end trace 402f96cbcf4a3d3e ]---

sta300: deauthenticating from 00:03:83:3d:30:aa by local choice (reason=3)
ath10k: MSI-X interrupt handling (8 intrs)
ath10k: UART prints disabled
ath10k: firmware 999.999.0.636 booted
ath10k: htt target version 2.1
ath10k: device successfully recovered
ath10k: firmware crashed!

Signature · 8 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com


_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
author
#1 lore raw
Quoted 18 lines · Ben Greear

Kernel is stock 'ath' tree, with small printk to debug an ath10k crash.

This is FYI for now...will be looking at other ath10k crash bugs before digging into this tone.

Setup is 2 stations trying to associate to same AP, which causes endless failures and firmware crashes. Good for chasing bugs :)

DMAR:[fault reason 05] PTE Write access is not set
dmar: DRHD: handling fault status reg 3
dmar: DMAR:[DMA Write] Request device [05:00.0] fault addr ffd52000 DMAR:[fault reason 05] PTE Write access is not set
dmar: DRHD: handling fault status reg 3
dmar: DMAR:[DMA Write] Request device [05:00.0] fault addr ffd52000 DMAR:[fault reason 05] PTE Write access is not set

Quoted 7 lines

sta300: authentication with 00:03:83:3d:30:aa timed out [root@ct523-9292 ~]# ath10k: Failed to delete peer: 00:03:83:3d:30:aa for VDEV: 1
ath10k: WMI vdev stop failed: ret -108

------------[ cut here ]------------
WARNING: CPU: 1 PID: 6 at /mnt/sda/home/greearb/git/ath/net/mac80211/sta_info.c:839 __sta_info_destroy+0x12)
Modules linked in: nf_nat_ipv4 nf_nat veth 8021q garp stp mrp llc macvlan pktgen lockd f71882fg coretemp hw]

I think this may be a result of whatever bug or limitation caused the firmware to error and/or crash when adding a second station VIF and trying to associate it to the same AP.

Probably not a problem with the rest of the wifi stacks.

Thanks,
Ben

Signature · 8 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com


_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#2 lore raw
Quoted 27 lines · Ben Greear
Quoted 18 lines · Ben Greear

Kernel is stock 'ath' tree, with small printk to debug an ath10k crash.

This is FYI for now...will be looking at other ath10k crash bugs before digging into this tone.

Setup is 2 stations trying to associate to same AP, which causes endless failures and firmware crashes. Good for chasing bugs :)

DMAR:[fault reason 05] PTE Write access is not set
dmar: DRHD: handling fault status reg 3
dmar: DMAR:[DMA Write] Request device [05:00.0] fault addr ffd52000 DMAR:[fault reason 05] PTE Write access is not set
dmar: DRHD: handling fault status reg 3
dmar: DMAR:[DMA Write] Request device [05:00.0] fault addr ffd52000 DMAR:[fault reason 05] PTE Write access is not set

Quoted 7 lines

sta300: authentication with 00:03:83:3d:30:aa timed out [root@ct523-9292 ~]# ath10k: Failed to delete peer: 00:03:83:3d:30:aa for VDEV: 1
ath10k: WMI vdev stop failed: ret -108

------------[ cut here ]------------
WARNING: CPU: 1 PID: 6 at /mnt/sda/home/greearb/git/ath/net/mac80211/sta_info.c:839 __sta_info_destroy+0x12)
Modules linked in: nf_nat_ipv4 nf_nat veth 8021q garp stp mrp llc macvlan pktgen lockd f71882fg coretemp hw]
Quoted 5 lines

I think this may be a result of whatever bug or limitation caused the firmware to error and/or crash when adding a second station VIF and trying to associate it to the same AP.

Probably not a problem with the rest of the wifi stacks.

Yup. This originates from the firmware limitation. It's unable to handle two peer nodes with same mac addr on different vdevs. That's why it's impossible to associate 2 (or more) station interfaces to the same BSS.

Michał

author
#3 lore raw
Quoted 4 lines · Michal Kazior

Yup. This originates from the firmware limitation. It's unable to handle two peer nodes with same mac addr on different vdevs. That's why it's impossible to associate 2 (or more) station interfaces to the same BSS.

I'm hoping to fix that soon. From what I can tell, the firmware is actually crashing, not just asserting. I can find the 'program counter' in the firmware crash dump, but I am still working on how to get some sort of stack trace or source code location out of that.

In case it helps someone else, I believe the register dump lines:

ath10k: target Register Dump
ath10k: [00]: 0x4100016C 0x00000000 0x0098B7DA 0x00000000

decode as: Target, Line (if assert), ProgramCounter(PC), bad-virtual-addr

Thanks,
Ben

Signature · 8 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com


_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#4 lore raw
In reply to Michal Kazior
Quoted 32 lines · Michal Kazior
Quoted 31 lines · Ben Greear
Quoted 18 lines · Ben Greear

Kernel is stock 'ath' tree, with small printk to debug an ath10k crash.

This is FYI for now...will be looking at other ath10k crash bugs before digging into this tone.

Setup is 2 stations trying to associate to same AP, which causes endless failures and firmware crashes. Good for chasing bugs :)

DMAR:[fault reason 05] PTE Write access is not set
dmar: DRHD: handling fault status reg 3
dmar: DMAR:[DMA Write] Request device [05:00.0] fault addr ffd52000 DMAR:[fault reason 05] PTE Write access is not set
dmar: DRHD: handling fault status reg 3
dmar: DMAR:[DMA Write] Request device [05:00.0] fault addr ffd52000 DMAR:[fault reason 05] PTE Write access is not set

Quoted 7 lines

sta300: authentication with 00:03:83:3d:30:aa timed out [root@ct523-9292 ~]# ath10k: Failed to delete peer: 00:03:83:3d:30:aa for VDEV: 1
ath10k: WMI vdev stop failed: ret -108

------------[ cut here ]------------
WARNING: CPU: 1 PID: 6 at /mnt/sda/home/greearb/git/ath/net/mac80211/sta_info.c:839 __sta_info_destroy+0x12)
Modules linked in: nf_nat_ipv4 nf_nat veth 8021q garp stp mrp llc macvlan pktgen lockd f71882fg coretemp hw]

I think this may be a result of whatever bug or limitation caused the firmware to error and/or crash when adding a second station VIF and trying to associate it to the same AP.

Quoted 6 lines
Quoted 1 line

Probably not a problem with the rest of the wifi stacks.

Yup. This originates from the firmware limitation. It's unable to handle two peer nodes with same mac addr on different vdevs. That's why it's impossible to associate 2 (or more) station interfaces to the same BSS.

I think we should add some sort of check to ath10k to detect that and not crash firmware.

Signature · 1 line
Kalle Valo