Received-SPF: pass (google.com: domain of linux-wireless+bounces-5153-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Date: Fri, 22 Mar 2024 08:16:17 -0700
From: Breno Leitao <leitao@debian.org>
To: Kalle Valo <kvalo@kernel.org>
Cc: Jeff Johnson <jjohnson@kernel.org>, kuba@kernel.org,
	keescook@chromium.org,
	"open list:NETWORKING DRIVERS (WIRELESS)" <linux-wireless@vger.kernel.org>,
	"open list:QUALCOMM ATHEROS ATH11K WIRELESS DRIVER" <ath11k@lists.infradead.org>,
	open list <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] wifi: ath11k: allocate dummy net_device dynamically
Message-ID: <Zf2gwaEU4PQ+6GMe@gmail.com>
References: <20240319185735.1268980-1-leitao@debian.org>
 <871q85as8e.fsf@kernel.org>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <871q85as8e.fsf@kernel.org>

On Wed, Mar 20, 2024 at 04:32:49PM +0200, Kalle Valo wrote:
> Breno Leitao <leitao@debian.org> writes:
> 
> > Embedding net_device into structures prohibits the usage of flexible
> > arrays in the net_device structure. For more details, see the discussion
> > at [1].
> >
> > Un-embed the net_device from struct ath11k_ext_irq_grp by converting it
> > into a pointer. Then use the leverage alloc_netdev() to allocate the
> > net_device object at ath11k_ahb_config_ext_irq() for ahb, and
> > ath11k_pcic_ext_irq_config() for pcic.
> >
> >  The free of the device occurs at ath11k_ahb_free_ext_irq() for the ahb
> > case, and ath11k_pcic_free_ext_irq() for the pcic case.
> >
> > [1] https://lore.kernel.org/all/20240229225910.79e224cf@kernel.org/
> >
> > Signed-off-by: Breno Leitao <leitao@debian.org>
> 
> This crashes on my nuc x86 test box with WCN6855 hw2.0 when running
> rmmod, stacktrace below. I used tag ath-202403201010 from my ath.git
> tree as the baseline. Sorry that I'm not able to debug this further
> right now.

Thanks for the detailed log. I am trying to understood the issue, and I think
we are deferring a null pointer, which is likely net_device->dev_addr.

Here, dev->dev_addr seems to be NULL.

>   27:	83 e6 07             	and    $0x7,%esi
>   2a:*	42 0f b6 14 30       	movzbl (%rax,%r14,1),%edx		<-- trapping instruction

In this case, and I understand RAX seems to be dev->addr, and R14 is the
KASAN shadow:

	RAX = 0000000000000000
	R14 = dffffc0000000000 (CONFIG_KASAN_SHADOW_OFFSET)

So, we are clearly deferring a NULL pointer.

Now, why? Not sure yet.