ath10k: firmware crash in station mode & problem DFS

Tue May 20 10:59:00 PDT 2014

On 20 May 2014 10:15, Vu Hai NGUYEN <vh.nguyen at actiasodielec.fr> wrote:
> I put my program on another device (same hardware as my actual device), 1 in AP using DFS canal and another in station mode.
> I got this message in station:
> wlan0: CTRL-EVENT-SCAN-STARTED
> wlan0: SME: Trying to authenticate with 04:f0:21:0e:38:be (SSID='ath10k' freq=52wlan0: direct probe to
> 04:f0:21:0e:38:be (try 1/3) 60 MHz)
> wlan0: authenticate with 04:f0:21:0e:38:be
> wlan0: direct probe to 04:f0:21:0e:38:be (try 2/3)
> wlan0: direct probe to 04:f0:21:0e:38:be (try 3/3)
> wlan0: authentication with 04:f0:21:0e:38:be timed out
> wlan0: CTRL-EVENT-SSID-TEMP-DISABLED id=0 ssid="ath10k" auth_failures=1 duration=10
>
> So the problem relied on authentication timed out, I wonder which parameter in the DFS mode can create a problem to establish  the connection?
>
Best if you can use 3rd device as a sniffer and capture airlogs.
Could you also send hostapd logs (captured with with -ddt)?
We will see if hostapd get this direct probe requests

BR
Janusz

>
>>> But I wonder if there will be any progress of the firmware in the future for promiscuous so that I can
>>> set up bridge mode?
>
>>Beats me. You should ask QCA about the future of ath10k firmware.
>
> LOL, sorry I thought that the firmware is developped by our group too. So the driver should adapt to the firmware then?
>
>>If you generate backports with kvalo/ath master branch then DFS won't
>>work as it is now (due to the bug I've mentioned earlier). Kalle
>>doesn't have the patch that deals with it, yet. You'll need to
>>cherry-pick the following 2 patches before generating backports from kvalo/ath:
>> https://git.kernel.org/cgit/linux/kernel/git/jberg/mac80211-next.git/commit/?id=67ae07a109f3d518085e3b81aa48740e8c5cc3f7
>>https://git.kernel.org/cgit/linux/kernel/git/jberg/mac80211-next.git/commit/?id=00ec75fc5a6499d8fdeb6ec9f8f5df68b9291c74
>
>>BTW, I am not sure this will work correctly if you will get "some"
>>backport and exchange only ath10k sources.
>>cfg80211/mac80211 could be incompatible with driver in such case.
>>Michal suggest to generate own backport form ath10k.git - this for
>>sure will work fine. In any other cases, while you will mix sources,
>>will be very hard to track your problems.
>
> I follow the instruction on this page http://wireless.kernel.org/en/users/Drivers/ath10k/backports#Compiling_custom_ath10k_backports to generate my own backport from master branch of ath10k. I also cherry-pick the 2 patches that you gave me. Then I see these 2 new options below cfg80211 certification onus in "make menuconfig" :
> [*]   cfg80211 certification onus
>   │ │                                   [*]     cfg80211 regulatory support for cellular base station hints
>   │ │                                   [*]     cfg80211 support for NO_IR relaxation
>
> I don't get the backtrace like before anymore, thing seems to work (Now I also get the cac_time=60, before it was 0) :
>
> ADDRCONF(NETDEV_UP): wlan0: link is not ready
> wlan0: interface state UNINITIALIZED->COUNTRY_UPDATE
> wlan0: interface state COUNTRY_UPDATE->DFS
> wlan0: DFS-CAC-START freq=5260 chan=52 sec_chan=0, width=0, seg0=0, seg1=0, cac_time=60s
> ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
> br0: port 2(wlan0) entering forwarding state
>
> The out put of iw dev show me:
>                 ssid ath10k
>                 type AP
>                 channel 52 (5260 MHz), width: 20 MHz, center1: 5260 MHz
>
> But still, I can not connect another station to my access point and I have no idea why. I don't see any warning or error from the output of dmesg (attached file).
> If I put my device into station mode in a canal DFS (for expl: 52) and set another device as an access point, my device can connect without any problem but it can not function
> as AP mode in a canal that employs DFS. (It can work in the others canal and other station can connect to it normally).
>
> I tried to return to the origin version of backport 3.15 and it's worse. First I don't get the 2 options below cfg80211 certification onus,
> then I got the same backtrace msg like before:
>
> wlan0: interface------------[ cut here ]------------
>  state UNINITIALWARNING: at /home/alberix2/Desktop/Wilinks_ath10k/ACCESS_POINT_A
> PPLI/Source/Driver_ath10k/backports-3.15-rc1-1/net/wireless/nl80211.c:5824 nl802
> 11_start_radar_detection+0xf8/0x16c [cfg80211]()
> IZED->COUNTRY_UPModules linked in:DATE
> wlan0: int ath10k_pci(O)erface state COU ath10k_core(O)NTRY_UPDATE->DFS ath(O)
> wlan0: DFS-CAC mac80211(O)-START freq=5260 cfg80211(O) chan=52 sec_cha compat(O)
> n=0, width=0, se arc4g0=0, seg1=0, ca ubifsc_time=0s
>  ubi [last unloaded: compat]
> Backtrace:
> [<c0011938>] (dump_backtrace+0x0/0x118) from [<c050205c>] (dump_stack+0x18/0x20)
>  r6:bf36be48 r5:000016c0 r4:00000000 r3:c10bc7ac
> [<c0502044>] (dump_stack+0x0/0x20) from [<c0066c38>] (warn_slowpath_common+0x54/
> 0x70)
> [<c0066be4>] (warn_slowpath_common+0x0/0x70) from [<c0066c78>] (warn_slowpath_nu
> ll+0x24/0x2c)
>  r8:00000000 r7:00000000 r6:c0ed2000 r5:d2ddc000 r4:c0ed23c8
> r3:00000009
> [<c0066c54>] (warn_slowpath_null+0x0/0x2c) from [<bf32fdac>] (nl80211_start_rada
> r_detection+0xf8/0x16c [cfg80211])
> [<bf32fcb4>] (nl80211_start_radar_detection+0x0/0x16c [cfg80211]) from [<c045348
> 8>] (genl_rcv_msg+0x1b4/0x1f4)
> [<c04532d4>] (genl_rcv_msg+0x0/0x1f4) from [<c0452904>] (netlink_rcv_skb+0x58/0x
> b8)
> [<c04528ac>] (netlink_rcv_skb+0x0/0xb8) from [<c04532c4>] (genl_rcv+0x20/0x30)
>  r6:d2a8f000 r5:d2b762e0 r4:d2b762e0 r3:c04532a4
> [<c04532a4>] (genl_rcv+0x0/0x30) from [<c04522d0>] (netlink_unicast+0x224/0x2c4)
>  r4:d0074400 r3:c04532a4
> [<c04520ac>] (netlink_unicast+0x0/0x2c4) from [<c04526a4>] (netlink_sendmsg+0x29
> 0/0x320)
> [<c0452414>] (netlink_sendmsg+0x0/0x320) from [<c0428eac>] (sock_sendmsg+0x98/0x
> b4)
> [<c0428e14>] (sock_sendmsg+0x0/0xb4) from [<c0429150>] (__sys_sendmsg+0x1d0/0x26
> c)
>  r4:d2843f64
> [<c0428f80>] (__sys_sendmsg+0x0/0x26c) from [<c042ab30>] (sys_sendmsg+0x44/0x68)
> [<c042aaec>] (sys_sendmsg+0x0/0x68) from [<c000ddc0>] (ret_fast_syscall+0x0/0x30
> )
>  r6:00134870 r5:4007b9f0 r4:00136660
> ---[ end trace 29f0f32871a48aa3 ]---
>
>
> NGUYEN Vu Hai
> Acita-Sodielec
> Route de Mayres - B.P. 9
> 12100 St GEORGES DE LUZENCON
> FRANCE
>
>
>