From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751617AbdBMFEl (ORCPT ); Mon, 13 Feb 2017 00:04:41 -0500 Received: from mx1.redhat.com ([209.132.183.28]:60076 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750753AbdBMFEk (ORCPT ); Mon, 13 Feb 2017 00:04:40 -0500 Subject: Re: [PATCH RFC] ptr_ring: fix race conditions when resizing To: "Michael S. Tsirkin" , linux-kernel@vger.kernel.org References: <1486765845-26587-1-git-send-email-mst@redhat.com> Cc: Dmitry Vyukov , stable@vger.kernel.org, "David S. Miller" From: Jason Wang Message-ID: <0aa7f7fc-3c10-2927-15d6-b061ea1add1f@redhat.com> Date: Mon, 13 Feb 2017 13:04:32 +0800 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.7.0 MIME-Version: 1.0 In-Reply-To: <1486765845-26587-1-git-send-email-mst@redhat.com> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.39]); Mon, 13 Feb 2017 05:04:40 +0000 (UTC) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 2017年02月11日 06:31, Michael S. Tsirkin wrote: > Resizing currently drops consumer lock. This can cause entries to be > reordered, which isn't good in itself. More importantly, consumer can > detect a false ring empty condition and block forever. > > Further, nesting of consumer within producer lock is problematic for > tun, since it produces entries in an interrupt, which causes a In an bh actually? Except for this. Acked-by: Jason Wang > lock order reversal: > > CPU0 CPU1 > ---- ---- > consume: > lock(&(&r->consumer_lock)->rlock); > resize: > local_irq_disable(); > lock(&(&r->producer_lock)->rlock); > lock(&(&r->consumer_lock)->rlock); > > produce: > lock(&(&r->producer_lock)->rlock); > > To fix, nest producer lock within consumer lock during resize, > and keep consumer lock during the whole swap operation. > > Reported-by: Dmitry Vyukov > Cc: stable@vger.kernel.org > Cc: "David S. Miller" > Cc: Jason Wang > Signed-off-by: Michael S. Tsirkin > --- > include/linux/ptr_ring.h | 36 +++++++++++++++++++++++++++++++----- > 1 file changed, 31 insertions(+), 5 deletions(-) > > diff --git a/include/linux/ptr_ring.h b/include/linux/ptr_ring.h > index 2052011..6c70444 100644 > --- a/include/linux/ptr_ring.h > +++ b/include/linux/ptr_ring.h > @@ -111,6 +111,11 @@ static inline int __ptr_ring_produce(struct ptr_ring *r, void *ptr) > return 0; > } > > +/* > + * Note: resize (below) nests producer lock within consumer lock, so if you > + * consume in interrupt or BH context, you must disable interrupts/BH when > + * calling this. > + */ > static inline int ptr_ring_produce(struct ptr_ring *r, void *ptr) > { > int ret; > @@ -242,6 +247,11 @@ static inline void *__ptr_ring_consume(struct ptr_ring *r) > return ptr; > } > > +/* > + * Note: resize (below) nests producer lock within consumer lock, so if you > + * call this in interrupt or BH context, you must disable interrupts/BH when > + * producing. > + */ > static inline void *ptr_ring_consume(struct ptr_ring *r) > { > void *ptr; > @@ -357,7 +367,7 @@ static inline void **__ptr_ring_swap_queue(struct ptr_ring *r, void **queue, > void **old; > void *ptr; > > - while ((ptr = ptr_ring_consume(r))) > + while ((ptr = __ptr_ring_consume(r))) > if (producer < size) > queue[producer++] = ptr; > else if (destroy) > @@ -372,6 +382,12 @@ static inline void **__ptr_ring_swap_queue(struct ptr_ring *r, void **queue, > return old; > } > > +/* > + * Note: producer lock is nested within consumer lock, so if you > + * resize you must make sure all uses nest correctly. > + * In particular if you consume ring in interrupt or BH context, you must > + * disable interrupts/BH when doing so. > + */ > static inline int ptr_ring_resize(struct ptr_ring *r, int size, gfp_t gfp, > void (*destroy)(void *)) > { > @@ -382,17 +398,25 @@ static inline int ptr_ring_resize(struct ptr_ring *r, int size, gfp_t gfp, > if (!queue) > return -ENOMEM; > > - spin_lock_irqsave(&(r)->producer_lock, flags); > + spin_lock_irqsave(&(r)->consumer_lock, flags); > + spin_lock(&(r)->producer_lock); > > old = __ptr_ring_swap_queue(r, queue, size, gfp, destroy); > > - spin_unlock_irqrestore(&(r)->producer_lock, flags); > + spin_unlock(&(r)->producer_lock); > + spin_unlock_irqrestore(&(r)->consumer_lock, flags); > > kfree(old); > > return 0; > } > > +/* > + * Note: producer lock is nested within consumer lock, so if you > + * resize you must make sure all uses nest correctly. > + * In particular if you consume ring in interrupt or BH context, you must > + * disable interrupts/BH when doing so. > + */ > static inline int ptr_ring_resize_multiple(struct ptr_ring **rings, int nrings, > int size, > gfp_t gfp, void (*destroy)(void *)) > @@ -412,10 +436,12 @@ static inline int ptr_ring_resize_multiple(struct ptr_ring **rings, int nrings, > } > > for (i = 0; i < nrings; ++i) { > - spin_lock_irqsave(&(rings[i])->producer_lock, flags); > + spin_lock_irqsave(&(rings[i])->consumer_lock, flags); > + spin_lock(&(rings[i])->producer_lock); > queues[i] = __ptr_ring_swap_queue(rings[i], queues[i], > size, gfp, destroy); > - spin_unlock_irqrestore(&(rings[i])->producer_lock, flags); > + spin_unlock(&(rings[i])->producer_lock); > + spin_unlock_irqrestore(&(rings[i])->consumer_lock, flags); > } > > for (i = 0; i < nrings; ++i)