Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Thomas Gleixner <tglx@linutronix.de>
To:     chengkaitao <pilgrimtao@gmail.com>
Cc:     linux-kernel@vger.kernel.org, smuchun@gmail.com,
        Kaitao Cheng <pilgrimtao@gmail.com>
Subject: Re: [RESEND v2] irq: Refactor irq_wait_for_interrupt info to simplify the code
In-Reply-To: <20200106154430.3413-1-pilgrimtao@gmail.com>
References: <20200106154430.3413-1-pilgrimtao@gmail.com>
Date:   Wed, 15 Jan 2020 14:43:14 +0100
Message-ID: <877e1sg7il.fsf@nanos.tec.linutronix.de>
MIME-Version: 1.0
Content-Type: text/plain
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

chengkaitao <pilgrimtao@gmail.com> writes:
> From: Kaitao Cheng <pilgrimtao@gmail.com>
>
> Cleanup extra if(test_and_clear_bit), and put the other one in front.

That simplifies the code but opens a race window:

 CPU 0                                CPU 1
                                      irq_wait_for_interrupt()
                                        has not yet reached schedule()
 free_irq()
   remove_action();
   synchronize_irq();                   

   #ifdef CONFIG_DEBUG_SHIRQ
    action->handler()                   if (test_and_clear_bit())
                            ---> bit is not set yet                                           
      --> SET thread running
   #endif                               

   kthread_stop()                       if (kthread_stop())

                   ---> Leave with bit set and thread active count != 0

That's just the most obvious example...

Thanks,

        tglx