Received-SPF: pass (google.com: domain of linux-wireless+bounces-1947-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Message-ID: <e87848a8-da4e-571b-2e07-7b7aa1bc8bed@candelatech.com>
Date: Mon, 15 Jan 2024 15:14:59 -0800
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.2.0
Subject: Re: [PATCH 5/8] wifi: mac80211: disallow drivers with HT wider than
 HE
Content-Language: en-MW
To: Johannes Berg <johannes@sipsolutions.net>, Kalle Valo <kvalo@kernel.org>,
 Miri Korenblit <miriam.rachel.korenblit@intel.com>
Cc: linux-wireless@vger.kernel.org,
 Gregory Greenman <gregory.greenman@intel.com>
References: <20240111161746.3978601-1-miriam.rachel.korenblit@intel.com>
 <20240111181514.da15fe3214d2.I4df51ad2f4c844615c168bf9bdb498925b3c77d4@changeid>
 <87wmsehf3d.fsf@kernel.org>
 <aae05a63171cf0f3c81dedc24d3b0a558ce530f5.camel@sipsolutions.net>
 <b1046a3b-8c29-aa64-2954-adec6c5d9bc9@candelatech.com>
 <88ad7e8849f7b29826ddf922734997d73539b21d.camel@sipsolutions.net>
From: Ben Greear <greearb@candelatech.com>
Organization: Candela Technologies
In-Reply-To: <88ad7e8849f7b29826ddf922734997d73539b21d.camel@sipsolutions.net>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

On 1/15/24 14:00, Johannes Berg wrote:
> On Fri, 2024-01-12 at 11:58 -0800, Ben Greear wrote:
>>
>> I tried backporting this patch into my 6.7 tree.  An mtk7915 radio system blows up badly
>> in this case.  Likely this is mt76 bug, but also...it used to work and the crash doesn't
>> make it very obvious that the above code is to blame.
> 
> Yeah, hence my comment about kernel developers hopefully being able to
> figure it out :-)
> 
>> mt7915e 0000:06:00.0: mt7915_register_device failed, ret: -22
>> mt7915e 0000:06:00.0: mt7915_pci_probe had error on try 3/3, ret: -22
> 
> Felix says this kind of retry logic doesn't exist upstream, maybe you
> have some delta in your tree that's making it crash?
> 
> Also, from what he says and looking at the code, it should register with
> HE 40 MHz capability set whenever has_he==true, so also here, do you
> have any non-upstream changes that could affect it?

Yes, I do.  Don't think I have anything that adjusts the 40Mhz, but possible.

Anyway, I know how to fix/work around it now, so if it works for Felix
then fine with me.

I still think you should have a pr_err or warn_on_once though, since if that clause
hits, at best case the radio suddenly became un-available.

Thanks,
Ben

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