Date: Fri, 6 Feb 2015 13:28:02 +0100
From: Joerg Roedel <joro@8bytes.org>
To: Jiang Liu <jiang.liu@linux.intel.com>
Cc: Joerg Roedel <jroedel@suse.de>, Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, "H. Peter Anvin" <hpa@zytor.com>,
        Prarit Bhargava <prarit@redhat.com>, Yinghai Lu <yinghai@kernel.org>,
        x86@kernel.org, linux-kernel@vger.kernel.org, alnovak@suse.com
Subject: Re: [PATCH] x86: irq: Check for valid irq descriptor in
 check_irq_vectors_for_cpu_disable
Message-ID: <20150206122802.GT3702@8bytes.org>
References: <20150204132754.GA10078@suse.de>
 <54D304DE.90506@linux.intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <54D304DE.90506@linux.intel.com>
User-Agent: Mutt/1.5.21 (2010-09-15)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 983
Lines: 30

Hi Jiang,

On Thu, Feb 05, 2015 at 01:51:26PM +0800, Jiang Liu wrote:
> Reviewed-by: Jiang Liu <jiang.liu@linux.intel.com>

Thanks for your review.

> Actually there's another racing pattern.
> for (irq = 0; irq < nr_irqs; irq++) {
> 	desc = irq_to_desc(irq);
> 	access desc->xxx
> }
> 
> When sparsing IRQ is enabled, there's no mechanism to protect
> desc returned by irq_to_desc(). Once I have considered a brute
> solution of disabling freeing of irq_desc:(

Hmm, how about wrapping the places that use irq_desc in rcu_read_lock()
and do a synchronize_rcu() before we free it (at least in the SPARSE_IRQ
case)? It wouldn't be a real RCU data structure, but we make at least
sure that we don't free an irq_desc thats in use.


	Joerg

--
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/