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;
Date:   Wed, 20 Apr 2022 20:16:37 +0200
From:   Heiko Carstens <hca@linux.ibm.com>
To:     Pingfan Liu <kernelfans@gmail.com>
Cc:     linux-s390@vger.kernel.org, Vasily Gorbik <gor@linux.ibm.com>,
        Alexander Gordeev <agordeev@linux.ibm.com>,
        Christian Borntraeger <borntraeger@linux.ibm.com>,
        Sven Schnelle <svens@linux.ibm.com>,
        Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH 4/9] s390/irq: utilize RCU instead of irq_lock_sparse()
 in show_msi_interrupt()
Message-ID: <YmBOBRMWlk4H11sc@osiris>
References: <20220420140521.45361-1-kernelfans@gmail.com>
 <20220420140521.45361-5-kernelfans@gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20220420140521.45361-5-kernelfans@gmail.com>
Precedence: bulk

On Wed, Apr 20, 2022 at 10:05:16PM +0800, Pingfan Liu wrote:
> irq_desc can be accessed safely in RCU read section as demonstrated by
> kstat_irqs_usr(). And raw_spin_lock_irqsave() context can provide a rcu
> read section, which can be utilized to get rid of irq_lock_sparse().
> 
> Signed-off-by: Pingfan Liu <kernelfans@gmail.com>
> Cc: Heiko Carstens <hca@linux.ibm.com>
> Cc: Vasily Gorbik <gor@linux.ibm.com>
> Cc: Alexander Gordeev <agordeev@linux.ibm.com>
> Cc: Christian Borntraeger <borntraeger@linux.ibm.com>
> Cc: Sven Schnelle <svens@linux.ibm.com>
> Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> Cc: linux-kernel@vger.kernel.org
> To: linux-s390@vger.kernel.org
> ---
>  arch/s390/kernel/irq.c | 11 +++++------
>  1 file changed, 5 insertions(+), 6 deletions(-)
> 
> diff --git a/arch/s390/kernel/irq.c b/arch/s390/kernel/irq.c
> index 3033f616e256..6302dc7874cf 100644
> --- a/arch/s390/kernel/irq.c
> +++ b/arch/s390/kernel/irq.c
> @@ -205,12 +205,13 @@ static void show_msi_interrupt(struct seq_file *p, int irq)
>  	unsigned long flags;
>  	int cpu;
>  
> -	irq_lock_sparse();
> +	raw_spin_lock_irqsave(&desc->lock, flags);
>  	desc = irq_to_desc(irq);

How is this supposed to work? desc get's initialized after its random
stack value has been used as a pointer to lock something...