From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from out-183.mta1.migadu.com (out-183.mta1.migadu.com [95.215.58.183]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id C6A541F951 for ; Wed, 11 Sep 2024 05:21:10 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=95.215.58.183 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1726032073; cv=none; b=PXj8/BuH/J+ezYi0l50KBzvZsZ4VjpJHuyQzGY4JJuEH+a2WSbcjoP31oojXF2LIEqzabIUxurC/A/fD4E9a29wZe5652gC8bryrdKme4EWTEaQYnuV5tIZZZpNHsVmtfEcnnKw/34Nw98t71xS+R7V2or6D08f9c34UbCLm9f4= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1726032073; c=relaxed/simple; bh=HJlpz8eduYiyw8yyvJNhR+nu02+JfJVfh/68TFMzlPg=; h=Content-Type:From:Mime-Version:Subject:Date:Message-Id:References: Cc:In-Reply-To:To; b=cD91Z1wWLGCaBCR1JMBzW5EM5ZpqA50NTl4kSaOBeaElR7c127vy22BmtFlxxZ48trJXZlCW8/9kYE38/JmwqCzh+qqCx+uVxeiFxG4YsNrYiLPdXjCMGnSaQ+eb/bXjlmMAyV2bhG0HvlDAHhxxZn9InVirQyBQXm2M68ae+yE= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.dev; spf=pass smtp.mailfrom=linux.dev; dkim=pass (1024-bit key) header.d=linux.dev header.i=@linux.dev header.b=p/6TrfyO; arc=none smtp.client-ip=95.215.58.183 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.dev Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=linux.dev Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux.dev header.i=@linux.dev header.b="p/6TrfyO" Content-Type: text/plain; charset=utf-8 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.dev; s=key1; t=1726032068; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=1IbthqVd0Y1HGEHfUp7IhQWZyBV6AeAPAti9kFjOXTU=; b=p/6TrfyOdz5bPa6WQHxmu5CrSrzWuD58AHrK4GL6TaXF7zIyNDidvBfFMjFvw0fjEMFWbB 5D/70SbNDh5UHHQPaylSqV8cjcePsGHKqrL5P/jcn7NB2rinYtC5tzXKRKO7xUY764X/Te dYyL4fjw/KvIp2itN9TmQHps3CDGE8Q= Content-Transfer-Encoding: quoted-printable X-Report-Abuse: Please report any abuse attempt to abuse@migadu.com and include these headers. From: Muchun Song Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 (1.0) Subject: Re: [PATCH v2 2/3] block: fix ordering between checking QUEUE_FLAG_QUIESCED and adding requests Date: Wed, 11 Sep 2024 13:20:28 +0800 Message-Id: <12E43452-371C-495D-B06A-DC2DA92CE0E8@linux.dev> References: Cc: Jens Axboe , Muchun Song , Yu Kuai , "open list:BLOCK LAYER" , LKML , stable@vger.kernel.org In-Reply-To: To: Ming Lei X-Migadu-Flow: FLOW_OUT > On Sep 11, 2024, at 11:59, Muchun Song wrote: >=20 > =EF=BB=BF >=20 >> On Sep 11, 2024, at 11:54, Ming Lei wrote: >>=20 >>> On Tue, Sep 10, 2024 at 07:22:16AM -0600, Jens Axboe wrote: >>> On 9/3/24 2:16 AM, Muchun Song wrote: >>>> Supposing the following scenario. >>>>=20 >>>> CPU0 CPU1 >>>>=20 >>>> blk_mq_insert_request() 1) store blk_mq_unquiesce_queue() >>>> blk_mq_run_hw_queue() blk_queue_flag_clear(QUEUE_= FLAG_QUIESCED) 3) store >>>> if (blk_queue_quiesced()) 2) load blk_mq_run_hw_queues() >>>> return blk_mq_run_hw_queue()= >>>> blk_mq_sched_dispatch_requests() if (!blk_mq_hctx_= has_pending()) 4) load >>>> return >>>>=20 >>>> The full memory barrier should be inserted between 1) and 2), as well a= s >>>> between 3) and 4) to make sure that either CPU0 sees QUEUE_FLAG_QUIESCE= D is >>>> cleared or CPU1 sees dispatch list or setting of bitmap of software que= ue. >>>> Otherwise, either CPU will not re-run the hardware queue causing starva= tion. >>>>=20 >>>> So the first solution is to 1) add a pair of memory barrier to fix the >>>> problem, another solution is to 2) use hctx->queue->queue_lock to synch= ronize >>>> QUEUE_FLAG_QUIESCED. Here, we chose 2) to fix it since memory barrier i= s not >>>> easy to be maintained. >>>=20 >>> Same comment here, 72-74 chars wide please. >>>=20 >>>> diff --git a/block/blk-mq.c b/block/blk-mq.c >>>> index b2d0f22de0c7f..ac39f2a346a52 100644 >>>> --- a/block/blk-mq.c >>>> +++ b/block/blk-mq.c >>>> @@ -2202,6 +2202,24 @@ void blk_mq_delay_run_hw_queue(struct blk_mq_hw_= ctx *hctx, unsigned long msecs) >>>> } >>>> EXPORT_SYMBOL(blk_mq_delay_run_hw_queue); >>>>=20 >>>> +static inline bool blk_mq_hw_queue_need_run(struct blk_mq_hw_ctx *hctx= ) >>>> +{ >>>> + bool need_run; >>>> + >>>> + /* >>>> + * When queue is quiesced, we may be switching io scheduler, or >>>> + * updating nr_hw_queues, or other things, and we can't run queue >>>> + * any more, even blk_mq_hctx_has_pending() can't be called safely= . >>>> + * >>>> + * And queue will be rerun in blk_mq_unquiesce_queue() if it is >>>> + * quiesced. >>>> + */ >>>> + __blk_mq_run_dispatch_ops(hctx->queue, false, >>>> + need_run =3D !blk_queue_quiesced(hctx->queue) && >>>> + blk_mq_hctx_has_pending(hctx)); >>>> + return need_run; >>>> +} >>>=20 >>> This __blk_mq_run_dispatch_ops() is also way too wide, why didn't you >>> just break it like where you copied it from? >>>=20 >>>> + >>>> /** >>>> * blk_mq_run_hw_queue - Start to run a hardware queue. >>>> * @hctx: Pointer to the hardware queue to run. >>>> @@ -2222,20 +2240,23 @@ void blk_mq_run_hw_queue(struct blk_mq_hw_ctx *= hctx, bool async) >>>>=20 >>>> might_sleep_if(!async && hctx->flags & BLK_MQ_F_BLOCKING); >>>>=20 >>>> - /* >>>> - * When queue is quiesced, we may be switching io scheduler, or >>>> - * updating nr_hw_queues, or other things, and we can't run queue >>>> - * any more, even __blk_mq_hctx_has_pending() can't be called safe= ly. >>>> - * >>>> - * And queue will be rerun in blk_mq_unquiesce_queue() if it is >>>> - * quiesced. >>>> - */ >>>> - __blk_mq_run_dispatch_ops(hctx->queue, false, >>>> - need_run =3D !blk_queue_quiesced(hctx->queue) && >>>> - blk_mq_hctx_has_pending(hctx)); >>>> + need_run =3D blk_mq_hw_queue_need_run(hctx); >>>> + if (!need_run) { >>>> + unsigned long flags; >>>>=20 >>>> - if (!need_run) >>>> - return; >>>> + /* >>>> + * synchronize with blk_mq_unquiesce_queue(), becuase we check >>>> + * if hw queue is quiesced locklessly above, we need the use >>>> + * ->queue_lock to make sure we see the up-to-date status to >>>> + * not miss rerunning the hw queue. >>>> + */ >>>> + spin_lock_irqsave(&hctx->queue->queue_lock, flags); >>>> + need_run =3D blk_mq_hw_queue_need_run(hctx); >>>> + spin_unlock_irqrestore(&hctx->queue->queue_lock, flags); >>>> + >>>> + if (!need_run) >>>> + return; >>>> + } >>>=20 >>> Is this not solvable on the unquiesce side instead? It's rather a shame >>> to add overhead to the fast path to avoid a race with something that's >>> super unlikely, like quisce. >>=20 >> Yeah, it can be solved by adding synchronize_rcu()/srcu() in unquiesce >> side, but SCSI may call it in non-sleepable context via scsi_internal_dev= ice_unblock_nowait(). >=20 > Another approach will be like the fix for BLK_MQ_S_STOPPED (in patch 3), > we could add a pair of mb into blk_queue_quiesced() and > blk_mq_unquiesce_queue(). In which case, the fix will not affect any fast > path, only slow path need the barrier overhead. I misunderstood Jens=E2=80=99s question. I think Ming is right. This approach only tries to reduce the overhead as=20 much as possible even for slow path compared to=20 spinlock_based approach.=20 Not solving the problem only from the unquiesce side. Muchun, Thanks. >=20 > diff --git a/block/blk-mq.c b/block/blk-mq.c > index b2d0f22de0c7f..45588ddb08d6b 100644 > --- a/block/blk-mq.c > +++ b/block/blk-mq.c > @@ -264,6 +264,12 @@ void blk_mq_unquiesce_queue(struct request_queue *q) > ; > } else if (!--q->quiesce_depth) { > blk_queue_flag_clear(QUEUE_FLAG_QUIESCED, q); > + /* > + * Pairs with the smp_mb() in blk_queue_quiesced() to orde= r the > + * clearing of QUEUE_FLAG_QUIESCED above and the checking o= f > + * dispatch list in the subsequent routine. > + */ > + smp_mb__after_atomic(); > run_queue =3D true; > } > spin_unlock_irqrestore(&q->queue_lock, flags); > diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h > index b8196e219ac22..7a71462892b66 100644 > --- a/include/linux/blkdev.h > +++ b/include/linux/blkdev.h > @@ -628,7 +628,25 @@ void blk_queue_flag_clear(unsigned int flag, struct r= equest_queue *q); > #define blk_noretry_request(rq) \ > ((rq)->cmd_flags & (REQ_FAILFAST_DEV|REQ_FAILFAST_TRANSPORT| \ > REQ_FAILFAST_DRIVER)) > -#define blk_queue_quiesced(q) test_bit(QUEUE_FLAG_QUIESCED, &(q)->queue_= flags) > + > +static inline bool blk_queue_quiesced(struct request_queue *q) > +{ > + /* Fast path: hardware queue is unquiesced most of the time. */ > + if (likely(!test_bit(QUEUE_FLAG_QUIESCED, &q->queue_flags))) > + return false; > + > + /* > + * This barrier is used to order adding of dispatch list before an= d > + * the test of QUEUE_FLAG_QUIESCED below. Pairs with the memory ba= rrier > + * in blk_mq_unquiesce_queue() so that dispatch code could either s= ee > + * QUEUE_FLAG_QUIESCED is cleared or dispatch list is not empty t= o > + * avoid missing dispatching requests. > + */ > + smp_mb(); > + > + return test_bit(QUEUE_FLAG_QUIESCED, &q->queue_flags); > +} > + > #define blk_queue_pm_only(q) atomic_read(&(q)->pm_only) > #define blk_queue_registered(q) test_bit(QUEUE_FLAG_REGISTERED, &(q= )->queue_flags) > #define blk_queue_sq_sched(q) test_bit(QUEUE_FLAG_SQ_SCHED, &(q)->queue_f= lags) >=20 > Muchun, > Thanks. >=20 >>=20 >>=20 >> Thanks, >> Ming >=20 >=20