linux-kernel - Re: Question about x86/mm/gup.c's use of disabled interrupts

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20090320155730.GD6698@linux.vnet.ibm.com>
Date:	Fri, 20 Mar 2009 08:57:31 -0700
From:	"Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>
To:	Jeremy Fitzhardinge <jeremy@...p.org>
Cc:	Nick Piggin <nickpiggin@...oo.com.au>, Avi Kivity <avi@...hat.com>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	Linux Memory Management List <linux-mm@...ck.org>,
	Xen-devel <xen-devel@...ts.xensource.com>,
	Jan Beulich <jbeulich@...ell.com>, Ingo Molnar <mingo@...e.hu>
Subject: Re: Question about x86/mm/gup.c's use of disabled interrupts

On Fri, Mar 20, 2009 at 08:38:46AM -0700, Jeremy Fitzhardinge wrote:
> Paul E. McKenney wrote:
>>> Ah, interesting.  So disabling interrupts prevents the RCU free from 
>>> happening, and non-atomic pte fetching is a non-issue.  So it doesn't 
>>> address the PAE side of the problem.
>>
>> This would be rcu_sched, correct?
>
> I guess?  Whatever it is that ends up calling all the rcu callbacks after 
> the idle.  A cpu with disabled interrupts can't go through idle, right?  Or 
> is there an explicit way to hold off rcu?

For synchronize_rcu() and call_rcu(), the only guaranteed way to hold
off RCU is rcu_read_lock() and rcu_read_unlock().

For call_rcu_bh, the only guaranteed way to hold off RCU is
rcu_read_lock_bh() and rcu_read_unlock_bh().

For synchronize_srcu(), the only guaranteed way to hold off RCU is
srcu_read_lock() and srcu_read_unlock().

For synchronize_sched() and call_rcu_sched(), anything that disables
preemption, including disabling irqs, holds off RCU.

Although disabling irqs can indeed hold off RCU in some other cases,
the only guarantee is for synchronize_sched() and call_rcu_sched().

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