From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S1752305AbdJKXLl (ORCPT <rfc822;w@1wt.eu>);
        Wed, 11 Oct 2017 19:11:41 -0400
Received: from mga11.intel.com ([192.55.52.93]:11265 "EHLO mga11.intel.com"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S1751946AbdJKXLh (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
        Wed, 11 Oct 2017 19:11:37 -0400
X-ExtLoop1: 1
X-IronPort-AV: E=Sophos;i="5.43,363,1503385200"; 
   d="scan'208";a="1024255912"
Date: Wed, 11 Oct 2017 16:11:37 -0700
From: Andi Kleen <ak@linux.intel.com>
To: Borislav Petkov <bp@alien8.de>
Cc: "Luck, Tony" <tony.luck@intel.com>, Jeremy Cline <jcline@redhat.com>,
        Thomas Gleixner <tglx@linutronix.de>, Ingo Molnar <mingo@redhat.com>,
        "H. Peter Anvin" <hpa@zytor.com>, x86@kernel.org,
        linux-edac@vger.kernel.org, linux-kernel@vger.kernel.org,
        Laura Abbott <labbott@redhat.com>
Subject: Re: x86/mce: suspicious RCU usage in 4.13.4
Message-ID: <20171011231137.GI5109@tassilo.jf.intel.com>
References: <9840bf0d-6756-75eb-1a2d-d3aace235244@redhat.com>
 <20171010194426.s7keveirclglx6vh@pd.tnic>
 <20171010200841.uhniuwc7pulqgmic@intel.com>
 <871smasqux.fsf@linux.intel.com>
 <20171011115020.i4gkcy5pjifh2zsm@pd.tnic>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20171011115020.i4gkcy5pjifh2zsm@pd.tnic>
User-Agent: Mutt/1.9.0 (2017-09-02)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

> -	next = mce_log_get_idx_check(mcelog.next);
> +	next = mcelog.next;
>  
>  	/* Only supports full reads right now */
>  	err = -EINVAL;
> @@ -281,8 +266,6 @@ static ssize_t mce_chrdev_read(struct file *filp, char __user *ubuf,
>  		next = cmpxchg(&mcelog.next, prev, 0);
>  	} while (next != prev);
>  
> -	synchronize_sched();

Sorry I take back what I wrote earlier. This RCU is actually still needed,
otherwise the reader could see partially written entries.

So rather have to keep that, and change the read code to run with rcu_read_lock()

-Andi