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;
Message-ID: <1540f1d0-4fda-aaf4-acda-f5aaa6952a00@colorfullife.com>
Date:   Fri, 23 Dec 2022 11:54:29 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.5.1
Subject: Re: [PATCH 3/3] kernel/irq/manage.c: disable_irq() might sleep.
To:     "Sverdlin, Alexander" <alexander.sverdlin@siemens.com>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "akpm@linux-foundation.org" <akpm@linux-foundation.org>
Cc:     "1vier1@web.de" <1vier1@web.de>,
        "tglx@linutronix.de" <tglx@linutronix.de>
References: <20221216150155.200389-1-manfred@colorfullife.com>
 <20221216150441.200533-1-manfred@colorfullife.com>
 <20221216150441.200533-3-manfred@colorfullife.com>
 <ca1972cec9724189b5a97a09069f3ee4f8c3f97c.camel@siemens.com>
Content-Language: en-US
From:   Manfred Spraul <manfred@colorfullife.com>
In-Reply-To: <ca1972cec9724189b5a97a09069f3ee4f8c3f97c.camel@siemens.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk

Hi Alexander,

On 12/20/22 07:58, Sverdlin, Alexander wrote:
> Hi Manfred!
>
> On Fri, 2022-12-16 at 16:04 +0100, Manfred Spraul wrote:
>> diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c
>> index 5b7cf28df290..8ce75495e04f 100644
>> --- a/kernel/irq/manage.c
>> +++ b/kernel/irq/manage.c
>> @@ -723,10 +723,13 @@ EXPORT_SYMBOL(disable_irq_nosync);
>>    *     to complete before returning. If you use this function while
>>    *     holding a resource the IRQ handler may need you will
>> deadlock.
>>    *
>> - *     This function may be called - with care - from IRQ context.
>> + *     Can only be called from preemptible code as it might sleep
>> when
>> + *     an interrupt thread is associated to @irq.
>> + *
>>    */
>>   void disable_irq(unsigned int irq)
>>   {
>> +       might_sleep();
> I'm not sure about this, latest wait_event() inside synchronize_irq()
> has it already.
>
>>          if (!__disable_irq_nosync(irq))
>>                  synchronize_irq(irq);
>>   }

That is the whole point: might_sleep() should be always called. We are 
clarifying an API definition. Everyone who uses disable_irq() from 
non-sleeping context should get a warning, 100% of the time.

Not just within synchronize_irq() if there is an active threaded irq 
handler.

--

     Manfred