From patchwork Mon Nov 26 16:54:28 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Keith Busch X-Patchwork-Id: 10698765 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 9985215A8 for ; Mon, 26 Nov 2018 16:57:38 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 8A7762A02C for ; Mon, 26 Nov 2018 16:57:38 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 7ECE32A080; Mon, 26 Nov 2018 16:57:38 +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=-7.9 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI, RCVD_IN_DNSWL_HI autolearn=unavailable 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 378FD2A06D for ; Mon, 26 Nov 2018 16:57:38 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726367AbeK0DwQ (ORCPT ); Mon, 26 Nov 2018 22:52:16 -0500 Received: from mga18.intel.com ([134.134.136.126]:52438 "EHLO mga18.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726280AbeK0DwO (ORCPT ); Mon, 26 Nov 2018 22:52:14 -0500 X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga008.jf.intel.com ([10.7.209.65]) by orsmga106.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 26 Nov 2018 08:57:33 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.56,283,1539673200"; d="scan'208";a="95155560" Received: from unknown (HELO localhost.lm.intel.com) ([10.232.112.69]) by orsmga008.jf.intel.com with ESMTP; 26 Nov 2018 08:57:32 -0800 From: Keith Busch To: linux-scsi@vger.kernel.org, linux-block@vger.kernel.org Cc: Jens Axboe , Martin Petersen , Bart Van Assche , Christoph Hellwig , Keith Busch Subject: [PATCHv4 1/3] blk-mq: Return true if request was completed Date: Mon, 26 Nov 2018 09:54:28 -0700 Message-Id: <20181126165430.4519-2-keith.busch@intel.com> X-Mailer: git-send-email 2.13.6 In-Reply-To: <20181126165430.4519-1-keith.busch@intel.com> References: <20181126165430.4519-1-keith.busch@intel.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 A driver may have internal state to cleanup if we're pretending a request didn't complete. Return 'false' if the command wasn't actually completed due to the timeout error injection, and true otherwise. Signed-off-by: Keith Busch Reviewed-by: Christoph Hellwig --- block/blk-mq.c | 5 +++-- include/linux/blk-mq.h | 2 +- 2 files changed, 4 insertions(+), 3 deletions(-) diff --git a/block/blk-mq.c b/block/blk-mq.c index 37674c1766a7..7c8cfa0cd420 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -638,11 +638,12 @@ static void hctx_lock(struct blk_mq_hw_ctx *hctx, int *srcu_idx) * Ends all I/O on a request. It does not handle partial completions. * The actual completion happens out-of-order, through a IPI handler. **/ -void blk_mq_complete_request(struct request *rq) +bool blk_mq_complete_request(struct request *rq) { if (unlikely(blk_should_fake_timeout(rq->q))) - return; + return false; __blk_mq_complete_request(rq); + return true; } EXPORT_SYMBOL(blk_mq_complete_request); diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h index ca0520ca6437..6e3da356a8eb 100644 --- a/include/linux/blk-mq.h +++ b/include/linux/blk-mq.h @@ -298,7 +298,7 @@ void blk_mq_add_to_requeue_list(struct request *rq, bool at_head, bool kick_requeue_list); void blk_mq_kick_requeue_list(struct request_queue *q); void blk_mq_delay_kick_requeue_list(struct request_queue *q, unsigned long msecs); -void blk_mq_complete_request(struct request *rq); +bool blk_mq_complete_request(struct request *rq); bool blk_mq_bio_list_merge(struct request_queue *q, struct list_head *list, struct bio *bio); bool blk_mq_queue_stopped(struct request_queue *q); From patchwork Mon Nov 26 16:54:29 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Keith Busch X-Patchwork-Id: 10698759 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 3DB4617D5 for ; Mon, 26 Nov 2018 16:57:37 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 2D6EA2A02C for ; Mon, 26 Nov 2018 16:57:37 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 220892A079; Mon, 26 Nov 2018 16:57:37 +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=-7.9 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI, RCVD_IN_DNSWL_HI 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 A56792A02C for ; Mon, 26 Nov 2018 16:57:36 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726375AbeK0DwQ (ORCPT ); Mon, 26 Nov 2018 22:52:16 -0500 Received: from mga18.intel.com ([134.134.136.126]:52441 "EHLO mga18.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726206AbeK0DwP (ORCPT ); Mon, 26 Nov 2018 22:52:15 -0500 X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga008.jf.intel.com ([10.7.209.65]) by orsmga106.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 26 Nov 2018 08:57:34 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.56,283,1539673200"; d="scan'208";a="95155565" Received: from unknown (HELO localhost.lm.intel.com) ([10.232.112.69]) by orsmga008.jf.intel.com with ESMTP; 26 Nov 2018 08:57:33 -0800 From: Keith Busch To: linux-scsi@vger.kernel.org, linux-block@vger.kernel.org Cc: Jens Axboe , Martin Petersen , Bart Van Assche , Christoph Hellwig , Keith Busch Subject: [PATCHv4 2/3] scsi: Do not rely on blk-mq for double completions Date: Mon, 26 Nov 2018 09:54:29 -0700 Message-Id: <20181126165430.4519-3-keith.busch@intel.com> X-Mailer: git-send-email 2.13.6 In-Reply-To: <20181126165430.4519-1-keith.busch@intel.com> References: <20181126165430.4519-1-keith.busch@intel.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 The scsi timeout error handling had been directly updating the block layer's request state to prevent a error handling and a natural completion from completing the same request twice. Fix this layering violation by having scsi control the fate of its commands with scsi owned flags rather than use blk-mq's. Signed-off-by: Keith Busch Reviewed-by: Christoph Hellwig --- drivers/scsi/scsi_error.c | 22 +++++++++++----------- drivers/scsi/scsi_lib.c | 13 ++++++++++++- include/scsi/scsi_cmnd.h | 4 ++++ 3 files changed, 27 insertions(+), 12 deletions(-) diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c index dd338a8cd275..16eef068e9e9 100644 --- a/drivers/scsi/scsi_error.c +++ b/drivers/scsi/scsi_error.c @@ -297,19 +297,19 @@ enum blk_eh_timer_return scsi_times_out(struct request *req) if (rtn == BLK_EH_DONE) { /* - * For blk-mq, we must set the request state to complete now - * before sending the request to the scsi error handler. This - * will prevent a use-after-free in the event the LLD manages - * to complete the request before the error handler finishes - * processing this timed out request. + * Set the command to complete first in order to prevent a real + * completion from releasing the command while error handling + * is using it. If the command was already completed, then the + * lower level driver beat the timeout handler, and it is safe + * to return without escalating error recovery. * - * If the request was already completed, then the LLD beat the - * time out handler from transferring the request to the scsi - * error handler. In that case we can return immediately as no - * further action is required. + * If timeout handling lost the race to a real completion, the + * block layer may ignore that due to a fake timeout injection, + * so return RESET_TIMER to allow error handling another shot + * at this command. */ - if (!blk_mq_mark_complete(req)) - return rtn; + if (test_and_set_bit(SCMD_STATE_COMPLETE, &scmd->state)) + return BLK_EH_RESET_TIMER; if (scsi_abort_command(scmd) != SUCCESS) { set_host_byte(scmd, DID_TIME_OUT); scsi_eh_scmd_add(scmd); diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c index 0df15cb738d2..0dbf25512778 100644 --- a/drivers/scsi/scsi_lib.c +++ b/drivers/scsi/scsi_lib.c @@ -1642,8 +1642,18 @@ static blk_status_t scsi_mq_prep_fn(struct request *req) static void scsi_mq_done(struct scsi_cmnd *cmd) { + if (unlikely(test_and_set_bit(SCMD_STATE_COMPLETE, &cmd->state))) + return; trace_scsi_dispatch_cmd_done(cmd); - blk_mq_complete_request(cmd->request); + + /* + * If the block layer didn't complete the request due to a timeout + * injection, scsi must clear its internal completed state so that the + * timeout handler will see it needs to escalate its own error + * recovery. + */ + if (unlikely(!blk_mq_complete_request(cmd->request))) + clear_bit(SCMD_STATE_COMPLETE, &cmd->state); } static void scsi_mq_put_budget(struct blk_mq_hw_ctx *hctx) @@ -1702,6 +1712,7 @@ static blk_status_t scsi_queue_rq(struct blk_mq_hw_ctx *hctx, if (!scsi_host_queue_ready(q, shost, sdev)) goto out_dec_target_busy; + clear_bit(SCMD_STATE_COMPLETE, &cmd->state); if (!(req->rq_flags & RQF_DONTPREP)) { ret = scsi_mq_prep_fn(req); if (ret != BLK_STS_OK) diff --git a/include/scsi/scsi_cmnd.h b/include/scsi/scsi_cmnd.h index d6fd2aba0380..3de905e205ce 100644 --- a/include/scsi/scsi_cmnd.h +++ b/include/scsi/scsi_cmnd.h @@ -61,6 +61,9 @@ struct scsi_pointer { /* flags preserved across unprep / reprep */ #define SCMD_PRESERVED_FLAGS (SCMD_UNCHECKED_ISA_DMA | SCMD_INITIALIZED) +/* for scmd->state */ +#define SCMD_STATE_COMPLETE (1 << 0) + struct scsi_cmnd { struct scsi_request req; struct scsi_device *device; @@ -145,6 +148,7 @@ struct scsi_cmnd { int result; /* Status code from lower level driver */ int flags; /* Command flags */ + unsigned long state; /* Command completion state */ unsigned char tag; /* SCSI-II queued command tag */ }; From patchwork Mon Nov 26 16:54:30 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Keith Busch X-Patchwork-Id: 10698761 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 63CA313AD for ; Mon, 26 Nov 2018 16:57:37 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 52BEA2A02C for ; Mon, 26 Nov 2018 16:57:37 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 470832A06D; Mon, 26 Nov 2018 16:57:37 +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=-7.9 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI, RCVD_IN_DNSWL_HI 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 DE8FB2A054 for ; Mon, 26 Nov 2018 16:57:36 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726380AbeK0DwR (ORCPT ); Mon, 26 Nov 2018 22:52:17 -0500 Received: from mga18.intel.com ([134.134.136.126]:52438 "EHLO mga18.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726318AbeK0DwQ (ORCPT ); Mon, 26 Nov 2018 22:52:16 -0500 X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga008.jf.intel.com ([10.7.209.65]) by orsmga106.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 26 Nov 2018 08:57:34 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.56,283,1539673200"; d="scan'208";a="95155572" Received: from unknown (HELO localhost.lm.intel.com) ([10.232.112.69]) by orsmga008.jf.intel.com with ESMTP; 26 Nov 2018 08:57:34 -0800 From: Keith Busch To: linux-scsi@vger.kernel.org, linux-block@vger.kernel.org Cc: Jens Axboe , Martin Petersen , Bart Van Assche , Christoph Hellwig , Keith Busch Subject: [PATCHv4 3/3] blk-mq: Simplify request completion state Date: Mon, 26 Nov 2018 09:54:30 -0700 Message-Id: <20181126165430.4519-4-keith.busch@intel.com> X-Mailer: git-send-email 2.13.6 In-Reply-To: <20181126165430.4519-1-keith.busch@intel.com> References: <20181126165430.4519-1-keith.busch@intel.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 There are no more users relying on blk-mq request states to prevent double completions, so replace the relatively expensive cmpxchg operation with WRITE_ONCE. Signed-off-by: Keith Busch Reviewed-by: Christoph Hellwig --- block/blk-mq.c | 4 +--- include/linux/blk-mq.h | 14 -------------- 2 files changed, 1 insertion(+), 17 deletions(-) diff --git a/block/blk-mq.c b/block/blk-mq.c index 7c8cfa0cd420..cda698804422 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -568,9 +568,7 @@ static void __blk_mq_complete_request(struct request *rq) bool shared = false; int cpu; - if (!blk_mq_mark_complete(rq)) - return; - + WRITE_ONCE(rq->state, MQ_RQ_COMPLETE); /* * Most of single queue controllers, there is only one irq vector * for handling IO completion, and the only irq's affinity is set diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h index 6e3da356a8eb..b8de11e0603b 100644 --- a/include/linux/blk-mq.h +++ b/include/linux/blk-mq.h @@ -329,20 +329,6 @@ void blk_mq_quiesce_queue_nowait(struct request_queue *q); unsigned int blk_mq_rq_cpu(struct request *rq); -/** - * blk_mq_mark_complete() - Set request state to complete - * @rq: request to set to complete state - * - * Returns true if request state was successfully set to complete. If - * successful, the caller is responsibile for seeing this request is ended, as - * blk_mq_complete_request will not work again. - */ -static inline bool blk_mq_mark_complete(struct request *rq) -{ - return cmpxchg(&rq->state, MQ_RQ_IN_FLIGHT, MQ_RQ_COMPLETE) == - MQ_RQ_IN_FLIGHT; -} - /* * Driver command data is immediately after the request. So subtract request * size to get back to the original request, add request size to get the PDU.