Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Kalle Valo <kvalo@kernel.org>
To:     Manivannan Sadhasivam <mani@kernel.org>
Cc:     Qiang Yu <quic_qianyu@quicinc.com>, quic_hemantk@quicinc.com,
        loic.poulain@linaro.org, quic_jhugo@quicinc.com,
        mhi@lists.linux.dev, linux-arm-msm@vger.kernel.org,
        linux-kernel@vger.kernel.org, quic_cang@quicinc.com,
        ath11k@lists.infradead.org
Subject: Re: [PATCH v3 1/1] bus: mhi: host: Fix up null pointer access in mhi_irq_handler
References: <1658459838-30802-1-git-send-email-quic_qianyu@quicinc.com>
        <20220726080636.GE5522@workstation>
Date:   Tue, 26 Jul 2022 20:53:58 +0300
In-Reply-To: <20220726080636.GE5522@workstation> (Manivannan Sadhasivam's
        message of "Tue, 26 Jul 2022 13:36:36 +0530")
Message-ID: <87czdrrc95.fsf@kernel.org>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk

Manivannan Sadhasivam <mani@kernel.org> writes:

> +ath11k, Kalle
>
> On Fri, Jul 22, 2022 at 11:17:18AM +0800, Qiang Yu wrote:
>> The irq handler for a shared IRQ ought to be prepared for running
>> even now it's being freed. So let's check the pointer used by
>> mhi_irq_handler to avoid null pointer access since it is probably
>> released before freeing IRQ.
>> 
>> Signed-off-by: Qiang Yu <quic_qianyu@quicinc.com>
>
> Reviewed-by: Manivannan Sadhasivam <mani@kernel.org>

This fixes the crash and my regression tests pass now, thanks. But
please see my question below.

Tested-by: Kalle Valo <kvalo@kernel.org>

>> +	/*
>> +	 * If CONFIG_DEBUG_SHIRQ is set, the IRQ handler will get invoked during __free_irq()
>> +	 * and by that time mhi_ctxt() would've freed. So check for the existence of mhi_ctxt
>> +	 * before handling the IRQs.
>> +	 */
>> +	if (!mhi_cntrl->mhi_ctxt) {
>> +		dev_dbg(&mhi_cntrl->mhi_dev->dev,
>> +			"mhi_ctxt has been freed\n");
>> +		return IRQ_HANDLED;
>> +	}

I don't see any protection accessing mhi_cntrl->mhi_ctxt, is this really
free of race conditions?

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches