From patchwork Mon Dec 4 18:06:23 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Bart Van Assche X-Patchwork-Id: 10091233 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id CBC99600C5 for ; Mon, 4 Dec 2017 18:06:39 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id C5005289F6 for ; Mon, 4 Dec 2017 18:06:39 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id B9EF728A61; Mon, 4 Dec 2017 18:06:39 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-6.8 required=2.0 tests=BAYES_00,DKIM_SIGNED, RCVD_IN_DNSWL_HI,T_DKIM_INVALID autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id E0DA4289F6 for ; Mon, 4 Dec 2017 18:06:38 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752415AbdLDSGg (ORCPT ); Mon, 4 Dec 2017 13:06:36 -0500 Received: from esa4.hgst.iphmx.com ([216.71.154.42]:29886 "EHLO esa4.hgst.iphmx.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752326AbdLDSGd (ORCPT ); Mon, 4 Dec 2017 13:06:33 -0500 DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=wdc.com; i=@wdc.com; q=dns/txt; s=dkim.wdc.com; t=1512410793; x=1543946793; h=from:to:cc:subject:date:message-id:in-reply-to: references; bh=d0D76ZrDMtOlxFDE2Ra+qpLQ5iY3TfE8cHZ9+LMB/AA=; b=mU8UKkeL28McxTd6VZeoMJWO7rfgcpRmpOtJNjgSh3qZ0PYSgEfosUUX MGjwD2Wdf6hdHXMBOS3EVMkBO1jWKWUcfrnh4hl2lcxSOF9Rfd2xAve72 V37bxvUBfRxIrH1GPGwCK8Vo972w0eJW9gyrcIehYRKDZ2c4TFGCrh6r7 qbhWztgZPOgK49Ksps6Yu8w9nLObna4rhDB9iYbtkXUlUVTId/sX/ah9x lcEkei/CMFsyybsNR5tXzi3uNFBgUiTsrFqdUW2nedTRKMzeS5zQNy0gI WWvKKnabruMRIyWBPRzfQVp6iqFA7VYsSmriitzEHsMV2+nu2woOH6KQp g==; X-IronPort-AV: E=Sophos;i="5.45,359,1508774400"; d="scan'208";a="63934868" Received: from h199-255-45-15.hgst.com (HELO uls-op-cesaep02.wdc.com) ([199.255.45.15]) by ob1.hgst.iphmx.com with ESMTP; 05 Dec 2017 02:06:24 +0800 Received: from uls-op-cesaip01.wdc.com ([10.248.3.36]) by uls-op-cesaep02.wdc.com with ESMTP; 04 Dec 2017 10:03:04 -0800 Received: from thinkpad-bart.sdcorp.global.sandisk.com (HELO thinkpad-bart.int.fusionio.com) ([10.11.166.51]) by uls-op-cesaip01.wdc.com with ESMTP; 04 Dec 2017 10:06:24 -0800 From: Bart Van Assche To: "Martin K . Petersen" , "James E . J . Bottomley" Cc: linux-scsi@vger.kernel.org, Bart Van Assche , Konstantin Khorenko , Stuart Hayes , Pavel Tikhomirov , Christoph Hellwig , Hannes Reinecke , Johannes Thumshirn , stable@vger.kernel.org Subject: [PATCH v3 1/2] Ensure that the SCSI error handler gets woken up Date: Mon, 4 Dec 2017 10:06:23 -0800 Message-Id: <20171204180624.18722-2-bart.vanassche@wdc.com> X-Mailer: git-send-email 2.15.0 In-Reply-To: <20171204180624.18722-1-bart.vanassche@wdc.com> References: <20171204180624.18722-1-bart.vanassche@wdc.com> Sender: linux-scsi-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-scsi@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP If scsi_eh_scmd_add() is called concurrently with scsi_host_queue_ready() while shost->host_blocked > 0 then it can happen that neither function wakes up the SCSI error handler. Fix this by making every function that decreases the host_busy counter wake up the error handler if necessary and by protecting the host_failed checks with the SCSI host lock. Reported-by: Pavel Tikhomirov References: https://marc.info/?l=linux-kernel&m=150461610630736 Fixes: commit 746650160866 ("scsi: convert host_busy to atomic_t") Signed-off-by: Bart Van Assche Cc: Konstantin Khorenko Cc: Stuart Hayes Cc: Pavel Tikhomirov Cc: Christoph Hellwig Cc: Hannes Reinecke Cc: Johannes Thumshirn Cc: Reviewed-by: Pavel Tikhomirov Tested-by: Stuart Hayes --- drivers/scsi/hosts.c | 6 ++++++ drivers/scsi/scsi_error.c | 18 ++++++++++++++++-- drivers/scsi/scsi_lib.c | 39 ++++++++++++++++++++++++++++----------- include/scsi/scsi_host.h | 2 ++ 4 files changed, 52 insertions(+), 13 deletions(-) diff --git a/drivers/scsi/hosts.c b/drivers/scsi/hosts.c index a306af6a5ea7..a0a7e4ff255c 100644 --- a/drivers/scsi/hosts.c +++ b/drivers/scsi/hosts.c @@ -324,6 +324,9 @@ static void scsi_host_dev_release(struct device *dev) scsi_proc_hostdir_rm(shost->hostt); + /* Wait for functions invoked through call_rcu(&shost->rcu, ...) */ + rcu_barrier(); + if (shost->tmf_work_q) destroy_workqueue(shost->tmf_work_q); if (shost->ehandler) @@ -331,6 +334,8 @@ static void scsi_host_dev_release(struct device *dev) if (shost->work_q) destroy_workqueue(shost->work_q); + destroy_rcu_head(&shost->rcu); + if (shost->shost_state == SHOST_CREATED) { /* * Free the shost_dev device name here if scsi_host_alloc() @@ -399,6 +404,7 @@ struct Scsi_Host *scsi_host_alloc(struct scsi_host_template *sht, int privsize) INIT_LIST_HEAD(&shost->starved_list); init_waitqueue_head(&shost->host_wait); mutex_init(&shost->scan_mutex); + init_rcu_head(&shost->rcu); index = ida_simple_get(&host_index_ida, 0, 0, GFP_KERNEL); if (index < 0) diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c index 5e89049e9b4e..258b8a741992 100644 --- a/drivers/scsi/scsi_error.c +++ b/drivers/scsi/scsi_error.c @@ -226,6 +226,17 @@ static void scsi_eh_reset(struct scsi_cmnd *scmd) } } +static void scsi_eh_inc_host_failed(struct rcu_head *head) +{ + struct Scsi_Host *shost = container_of(head, typeof(*shost), rcu); + unsigned long flags; + + spin_lock_irqsave(shost->host_lock, flags); + shost->host_failed++; + scsi_eh_wakeup(shost); + spin_unlock_irqrestore(shost->host_lock, flags); +} + /** * scsi_eh_scmd_add - add scsi cmd to error handling. * @scmd: scmd to run eh on. @@ -248,9 +259,12 @@ void scsi_eh_scmd_add(struct scsi_cmnd *scmd) scsi_eh_reset(scmd); list_add_tail(&scmd->eh_entry, &shost->eh_cmd_q); - shost->host_failed++; - scsi_eh_wakeup(shost); spin_unlock_irqrestore(shost->host_lock, flags); + /* + * Ensure that all tasks observe the host state change before the + * host_failed change. + */ + call_rcu(&shost->rcu, scsi_eh_inc_host_failed); } /** diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c index b6d3842b6809..5cbc69b2b1ae 100644 --- a/drivers/scsi/scsi_lib.c +++ b/drivers/scsi/scsi_lib.c @@ -318,22 +318,39 @@ static void scsi_init_cmd_errh(struct scsi_cmnd *cmd) cmd->cmd_len = scsi_command_size(cmd->cmnd); } -void scsi_device_unbusy(struct scsi_device *sdev) +/* + * Decrement the host_busy counter and wake up the error handler if necessary. + * Avoid as follows that the error handler is not woken up if shost->host_busy + * == shost->host_failed: use call_rcu() in scsi_eh_scmd_add() in combination + * with an RCU read lock in this function to ensure that this function in its + * entirety either finishes before scsi_eh_scmd_add() increases the + * host_failed counter or that it notices the shost state change made by + * scsi_eh_scmd_add(). + */ +static void scsi_dec_host_busy(struct Scsi_Host *shost) { - struct Scsi_Host *shost = sdev->host; - struct scsi_target *starget = scsi_target(sdev); unsigned long flags; + rcu_read_lock(); atomic_dec(&shost->host_busy); - if (starget->can_queue > 0) - atomic_dec(&starget->target_busy); - - if (unlikely(scsi_host_in_recovery(shost) && - (shost->host_failed || shost->host_eh_scheduled))) { + if (unlikely(scsi_host_in_recovery(shost))) { spin_lock_irqsave(shost->host_lock, flags); - scsi_eh_wakeup(shost); + if (shost->host_failed || shost->host_eh_scheduled) + scsi_eh_wakeup(shost); spin_unlock_irqrestore(shost->host_lock, flags); } + rcu_read_unlock(); +} + +void scsi_device_unbusy(struct scsi_device *sdev) +{ + struct Scsi_Host *shost = sdev->host; + struct scsi_target *starget = scsi_target(sdev); + + scsi_dec_host_busy(shost); + + if (starget->can_queue > 0) + atomic_dec(&starget->target_busy); atomic_dec(&sdev->device_busy); } @@ -1531,7 +1548,7 @@ static inline int scsi_host_queue_ready(struct request_queue *q, list_add_tail(&sdev->starved_entry, &shost->starved_list); spin_unlock_irq(shost->host_lock); out_dec: - atomic_dec(&shost->host_busy); + scsi_dec_host_busy(shost); return 0; } @@ -2017,7 +2034,7 @@ static blk_status_t scsi_queue_rq(struct blk_mq_hw_ctx *hctx, return BLK_STS_OK; out_dec_host_busy: - atomic_dec(&shost->host_busy); + scsi_dec_host_busy(shost); out_dec_target_busy: if (scsi_target(sdev)->can_queue > 0) atomic_dec(&scsi_target(sdev)->target_busy); diff --git a/include/scsi/scsi_host.h b/include/scsi/scsi_host.h index a8b7bf879ced..1a1df0d21ee3 100644 --- a/include/scsi/scsi_host.h +++ b/include/scsi/scsi_host.h @@ -571,6 +571,8 @@ struct Scsi_Host { struct blk_mq_tag_set tag_set; }; + struct rcu_head rcu; + atomic_t host_busy; /* commands actually active on low-level */ atomic_t host_blocked;