Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: Kernel crash when scanning for APs in huge WiFi environment -
 Ath10K - ARMV7 - Linux 4.14.73
To:     Vincent RUBY <vruby@cube-technologies.net>,
        "linux-wireless@vger.kernel.org" <linux-wireless@vger.kernel.org>
References: <1cef8054e4e447f2ba02d1de1e6afdf8@Cube-mail.cube.local>
 <e23f3f12dcab42ea9ab4b7384080cdf2@Cube-mail.cube.local>
 <c2f66f16-3800-b127-4c58-704a29ec693d@candelatech.com>
 <2eb86fb560cc482fa63bf42fafb97922@Cube-mail.cube.local>
From:   Ben Greear <greearb@candelatech.com>
Organization: Candela Technologies
Message-ID: <2b3885f7-c48a-a3a8-92a9-780405527bf3@candelatech.com>
Date:   Fri, 9 Jul 2021 09:34:01 -0700
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.2.2
MIME-Version: 1.0
In-Reply-To: <2eb86fb560cc482fa63bf42fafb97922@Cube-mail.cube.local>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 7/9/21 9:15 AM, Vincent RUBY wrote:
> Thank you Ben for your quick reply.
> 
> Unfortunately, when the system crashes, there is no detail. No trace or log can be retrieved, even in dmesg.
> It is simply rebooting.
> 
> Trying a newer kernel is not easy since we are hardware driver dependent.

Yep, is sucks to develop on ancient cruft.

You should be able to wire in a serial console somehow, that is first step to
debugging such kernel crashes.  But, assuming this is some hacked up SDK
or other non-standard kernel, the fix will be between you and your vendor.

You could also put a similar chipset NIC into a PC and try same test on latest
upstream kernel and normal OS like Ubunto or Fedora.
If it crashes there, that would likely be easier to debug and folks on the list
might could help.

Thanks,
Ben

> 
> VinceCube
> 
> -----Message d'origine-----
> De : Ben Greear <greearb@candelatech.com>
> Envoyé : vendredi 9 juillet 2021 17:59
> À : Vincent RUBY <vruby@cube-technologies.net>; linux-wireless@vger.kernel.org
> Objet : Re: Kernel crash when scanning for APs in huge WiFi environment - Ath10K - ARMV7 - Linux 4.14.73
> 
> On 7/9/21 8:48 AM, Vincent RUBY wrote:
>> Hi,
>>
>> I apologize in advance if I am writing at the wrong place.
>> We have developed a custom embedded board based on ARMV7. On the board there is a Telit WE866C3-P Wifi module (based on ath10k).
>> Linux : 4.14.73 (YOCTO/poky)
>> Microcontroller : SAMA5D27
>>
>> The board is working perfectly in a normal Wifi environment (e.g., at office).
>> But when the board is put in a huge Wifi environment (e.g., 350 APs with 2 antennas each with more than 600 visible BSSIDS) the Linux kernel crashes when attempting to connect to a Wifi network using wpa_supplicant.
>> When the Wifi antenna is attenuated with some electromagnetic obstacle (e.g., hand or metallic shield), the board connects without any problem.
>>
>> The problem is the same when running a scan from wpa_cli .
>>
>> Do you have any idea where I can find some clue to resolve this issue?
>>
>> Many thanks in advance !
>>
>> VinceCube
>>
> 
> Please show the details of the kernel crash.  And either way, answer is probably to try more recent kernel and see if problem is already fixed.
> 
> Thanks,
> Ben
> 
> --
> Ben Greear <greearb@candelatech.com>
> Candela Technologies Inc  http://www.candelatech.com
> 


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com