Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <c7aa9034-59f8-11aa-7170-86c0db012dda@redhat.com>
Date:   Wed, 10 May 2023 12:04:32 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.10.0
Subject: Re: [REGRESSION] wifi: brcmfmac: add firmware vendor info in driver
 info
Content-Language: en-US, nl
To:     Felix <nimrod4garoa@gmail.com>,
        Arend van Spriel <aspriel@gmail.com>,
        Franky Lin <franky.lin@broadcom.com>,
        Hante Meuleman <hante.meuleman@broadcom.com>
Cc:     regressions@lists.linux.dev,
        Arend van Spriel <arend.vanspriel@broadcom.com>,
        Kalle Valo <kvalo@kernel.org>, linux-kernel@vger.kernel.org,
        linux-wireless@vger.kernel.org,
        brcm80211-dev-list.pdl@broadcom.com,
        SHA-cyfmac-dev-list@infineon.com
References: <4ef3f252ff530cbfa336f5a0d80710020fc5cb1e.camel@gmail.com>
From:   Hans de Goede <hdegoede@redhat.com>
In-Reply-To: <4ef3f252ff530cbfa336f5a0d80710020fc5cb1e.camel@gmail.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

Hi Felix,

On 5/10/23 10:25, Felix wrote:
> Changes made by the mentioned commit lead to Oops when waking up after
> suspend to RAM.
> 
> When waking up a Dell XPS 13 9350 with BCM4350 wifi card after suspend
> to RAM, the kernel hangs with a NULL pointer dereference and Oops. The
> issues was introduced by commit
> da6d9c8ecd00e20218461007948f2b0a8e7fa242 and only happens when the
> brcmfmac module is loaded. This issue is the reason we could not
> upgrade to newer kernels than the 6.1 series on the affected machine.
> 
> 
> Details:
> 
> On a Dell XPS 13 9350 laptop with Broadcom BCM4350 wifi card (according
> to lspci), every kernel starting with 6.2 hangs when waking up after
> suspend to RAM. The issue persisted as of May 2nd (last tested commit
> was 348551ddaf311c76b01cdcbaf61b6fef06a49144). Passing the debug and
> no_console_suspend parameters to the kernel show that a NULL pointer
> dereference and an Oops happen at wakeup. Please find the kernel
> messages readable on the screen at this point attached
> (kernel_log_after_suspend.txt).
> 
> A bisection was successful and produced the first bad commit
> [da6d9c8ecd00e20218461007948f2b0a8e7fa242] wifi: brcmfmac: add firmware
> vendor info in driver info
> I've attached the bisection log as well (bisect.txt).
> 
> Suspend to RAM works on this machine on newer kernels once the brcmfmac
> module is unloaded.
> 
> Distribution: Arch Linux
> Architecture: x86_64
> Device: Dell XPS 13 9350 with BCM4350 wifi card
> Kernel config used at compilation (created using the Arch Linux default
> config and make olddefconfig):
> https://gist.githubusercontent.com/maclomhair/e33fa7eece8f8c77e5a88c4eede2f523/raw/429dfb022498c026225865dbb3ab2f75d5030a54/config
> Dmesg output after boot:
> https://gist.githubusercontent.com/maclomhair/47c75fa759d93800e7fb5e937fabef3e/raw/10750996a1ea9dcb3fa14d9b94e11fedd3abf280/dmesg
> 
> #regzbot introduced da6d9c8ecd00e20218461007948f2b0a8e7fa242

Thank you for reporting this and thank you for bisecting it too!

As a result of your excellent bug report I've been able to write
a fix for this. I'm confident enough in the fix that I've submitted
it upstream right away.

But it would still be good if you can test the fix and let us know
if it fixes this issue for you. You should have received a copy
of the patch when I submitted it upstream.

Regards,

Hans