Date: Thu, 10 Feb 2005 16:59:28 +0100
From: Ingo Molnar <mingo@elte.hu>
To: Kevin Hilman <kevin@hilman.org>
Cc: linux-kernel@vger.kernel.org
Subject: Re: Real-Time Preemption and GFP_ATOMIC
Message-ID: <20050210155928.GA8263@elte.hu>
References: <83r7jyiyqx.fsf@www2.muking.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <83r7jyiyqx.fsf@www2.muking.org>
User-Agent: Mutt/1.4.1i
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 965
Lines: 24


* Kevin Hilman <kevin@hilman.org> wrote:

> To produce the following trace, I wrote a simple moudle which just has
> this as its init_module routine:
> 
>         local_irq_disable();
>         p = __get_free_page(GFP_ATOMIC);
>         local_irq_enable();

in the PREEMPT_RT kernel almost everything might sleep, so the general
rule is to not call anything 'complex' from an IRQs-off section. 
Depending on which is easier in your code, if you want to fix it up for
PREEMPT_RT then either move the GFP_ATOMIC allocation from under the
irqs-off section, or introduce a spinlock for the irqs-off section and
use spin_lock_irqsave(). (that is almost always needed anyway if you
really needed the irqs-off section.)

	Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/