From patchwork Thu Mar 11 20:16:49 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Don Brace X-Patchwork-Id: 12132875 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-13.6 required=3.0 tests=BAYES_00,DKIM_INVALID, DKIM_SIGNED,HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id AD113C433DB for ; Thu, 11 Mar 2021 20:17:58 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 699A664F8B for ; Thu, 11 Mar 2021 20:17:58 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230404AbhCKUR0 (ORCPT ); Thu, 11 Mar 2021 15:17:26 -0500 Received: from esa.microchip.iphmx.com ([68.232.153.233]:49147 "EHLO esa.microchip.iphmx.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230047AbhCKURA (ORCPT ); Thu, 11 Mar 2021 15:17:00 -0500 DKIM-Signature: v=1; a=rsa-sha256; c=simple/simple; d=microchip.com; i=@microchip.com; q=dns/txt; s=mchp; t=1615493820; x=1647029820; h=subject:from:to:cc:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=epyQf8EFvMmCBsaj+DAuo4vXd50N/tMp3YhH5iftAx4=; b=RnNU6kdJ9lQ78CMoF8BomhwZpZqPl+2Q6+lL90PCYD9TqSXwij7KDioV w7OjRz5rGPOCtwNGJQF8CmU5JQ45NFeMjWeDh9GP6fECw0pMrDrQA7whv MqTwhORy92byVq6v96cbg06pOUO7g86y5Qew3i4ZAwLK9IBV2DDfb9b6f a6efp4lyT/eNZl0LtlgiJrmm7zU9ufXvieXAHGcEcoN+JG3mHDpizq37F SzbB6rbI/6iMcZbsUYNEwVGPjAr5HdL+N0G8kX+G64qoqgONMJxCwnmi3 8c7mYHcPvmwpFXDEkk98yIHS6g4uKllT69s5icMBNy2YmUpzrEJkQznE7 g==; IronPort-SDR: FbvQ+dY9qxYmX/8SAXIOFf4DUAYQicbUEDL4cerxEReJdqDdSvXoRfsWLMs+6aNATeusj2sVSd ZOHmlCA6zp+PeiWOHMBNiFBz/dFQVOzccw/Ea2PS+1GsTSKufWokud8W59CkIcPzf6UTFcz+/u 3p8qaG9JEqvPUs9Kz0CfNOoSublzFOirrPo/zAR9dHv84g6nyyPMMd3QzZbL+HPOwUwhZ5jj9p RKUP6X9hTZciayf1HuCXozl97WZR6XACeGq/dXp/GA8CECBYwljDV244RXXFNrBmdcN2DmSKlw heg= X-IronPort-AV: E=Sophos;i="5.81,241,1610434800"; d="scan'208";a="118559093" Received: from smtpout.microchip.com (HELO email.microchip.com) ([198.175.253.82]) by esa1.microchip.iphmx.com with ESMTP/TLS/AES256-SHA256; 11 Mar 2021 13:17:00 -0700 Received: from chn-vm-ex02.mchp-main.com (10.10.85.144) by chn-vm-ex01.mchp-main.com (10.10.85.143) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2176.2; Thu, 11 Mar 2021 13:16:50 -0700 Received: from [127.0.1.1] (10.10.115.15) by chn-vm-ex02.mchp-main.com (10.10.85.144) with Microsoft SMTP Server id 15.1.2176.2 via Frontend Transport; Thu, 11 Mar 2021 13:16:49 -0700 Subject: [PATCH V5 20/31] smartpqi: update raid bypass handling From: Don Brace To: , , , , , , , , , , , CC: Date: Thu, 11 Mar 2021 14:16:49 -0600 Message-ID: <161549380976.25025.11776487034357231156.stgit@brunhilda> In-Reply-To: <161549045434.25025.17473629602756431540.stgit@brunhilda> References: <161549045434.25025.17473629602756431540.stgit@brunhilda> User-Agent: StGit/0.23-dirty MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-scsi@vger.kernel.org From: Kevin Barnett Simplify aio retry management by removing retry list and list management. Need to retry is already set in the response status. Also remove the bypass worker thread. AIO - Accelerated I/O I/O requests bypass the RAID engine and go directly to either an HBA disk or to a physical component of a RAID volume. Reviewed-by: Scott Benesh Reviewed-by: Scott Teel Signed-off-by: Kevin Barnett Signed-off-by: Don Brace --- drivers/scsi/smartpqi/smartpqi.h | 4 - drivers/scsi/smartpqi/smartpqi_init.c | 172 ++++----------------------------- 2 files changed, 19 insertions(+), 157 deletions(-) diff --git a/drivers/scsi/smartpqi/smartpqi.h b/drivers/scsi/smartpqi/smartpqi.h index fa1ebeea777d..1b9aa6e9e04c 100644 --- a/drivers/scsi/smartpqi/smartpqi.h +++ b/drivers/scsi/smartpqi/smartpqi.h @@ -1347,10 +1347,6 @@ struct pqi_ctrl_info { atomic_t num_blocked_threads; wait_queue_head_t block_requests_wait; - struct list_head raid_bypass_retry_list; - spinlock_t raid_bypass_retry_list_lock; - struct work_struct raid_bypass_retry_work; - struct pqi_ofa_memory *pqi_ofa_mem_virt_addr; dma_addr_t pqi_ofa_mem_dma_handle; void **pqi_ofa_chunk_virt_addr; diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index 4f7a38b96356..b44de3e25541 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -5280,12 +5280,6 @@ static inline int pqi_raid_submit_scsi_cmd(struct pqi_ctrl_info *ctrl_info, device, scmd, queue_group); } -static inline void pqi_schedule_bypass_retry(struct pqi_ctrl_info *ctrl_info) -{ - if (!pqi_ctrl_blocked(ctrl_info)) - schedule_work(&ctrl_info->raid_bypass_retry_work); -} - static bool pqi_raid_bypass_retry_needed(struct pqi_io_request *io_request) { struct scsi_cmnd *scmd; @@ -5302,7 +5296,7 @@ static bool pqi_raid_bypass_retry_needed(struct pqi_io_request *io_request) return false; device = scmd->device->hostdata; - if (pqi_device_offline(device)) + if (pqi_device_offline(device) || pqi_device_in_remove(device)) return false; ctrl_info = shost_to_hba(scmd->device->host); @@ -5312,132 +5306,6 @@ static bool pqi_raid_bypass_retry_needed(struct pqi_io_request *io_request) return true; } -static inline void pqi_add_to_raid_bypass_retry_list( - struct pqi_ctrl_info *ctrl_info, - struct pqi_io_request *io_request, bool at_head) -{ - unsigned long flags; - - spin_lock_irqsave(&ctrl_info->raid_bypass_retry_list_lock, flags); - if (at_head) - list_add(&io_request->request_list_entry, - &ctrl_info->raid_bypass_retry_list); - else - list_add_tail(&io_request->request_list_entry, - &ctrl_info->raid_bypass_retry_list); - spin_unlock_irqrestore(&ctrl_info->raid_bypass_retry_list_lock, flags); -} - -static void pqi_queued_raid_bypass_complete(struct pqi_io_request *io_request, - void *context) -{ - struct scsi_cmnd *scmd; - - scmd = io_request->scmd; - pqi_free_io_request(io_request); - pqi_scsi_done(scmd); -} - -static void pqi_queue_raid_bypass_retry(struct pqi_io_request *io_request) -{ - struct scsi_cmnd *scmd; - struct pqi_ctrl_info *ctrl_info; - - io_request->io_complete_callback = pqi_queued_raid_bypass_complete; - scmd = io_request->scmd; - scmd->result = 0; - ctrl_info = shost_to_hba(scmd->device->host); - - pqi_add_to_raid_bypass_retry_list(ctrl_info, io_request, false); - pqi_schedule_bypass_retry(ctrl_info); -} - -static int pqi_retry_raid_bypass(struct pqi_io_request *io_request) -{ - struct scsi_cmnd *scmd; - struct pqi_scsi_dev *device; - struct pqi_ctrl_info *ctrl_info; - struct pqi_queue_group *queue_group; - - scmd = io_request->scmd; - device = scmd->device->hostdata; - if (pqi_device_in_reset(device)) { - pqi_free_io_request(io_request); - set_host_byte(scmd, DID_RESET); - pqi_scsi_done(scmd); - return 0; - } - - ctrl_info = shost_to_hba(scmd->device->host); - queue_group = io_request->queue_group; - - pqi_reinit_io_request(io_request); - - return pqi_raid_submit_scsi_cmd_with_io_request(ctrl_info, io_request, - device, scmd, queue_group); -} - -static inline struct pqi_io_request *pqi_next_queued_raid_bypass_request( - struct pqi_ctrl_info *ctrl_info) -{ - unsigned long flags; - struct pqi_io_request *io_request; - - spin_lock_irqsave(&ctrl_info->raid_bypass_retry_list_lock, flags); - io_request = list_first_entry_or_null( - &ctrl_info->raid_bypass_retry_list, - struct pqi_io_request, request_list_entry); - if (io_request) - list_del(&io_request->request_list_entry); - spin_unlock_irqrestore(&ctrl_info->raid_bypass_retry_list_lock, flags); - - return io_request; -} - -static void pqi_retry_raid_bypass_requests(struct pqi_ctrl_info *ctrl_info) -{ - int rc; - struct pqi_io_request *io_request; - - pqi_ctrl_busy(ctrl_info); - - while (1) { - if (pqi_ctrl_blocked(ctrl_info)) - break; - io_request = pqi_next_queued_raid_bypass_request(ctrl_info); - if (!io_request) - break; - rc = pqi_retry_raid_bypass(io_request); - if (rc) { - pqi_add_to_raid_bypass_retry_list(ctrl_info, io_request, - true); - pqi_schedule_bypass_retry(ctrl_info); - break; - } - } - - pqi_ctrl_unbusy(ctrl_info); -} - -static void pqi_raid_bypass_retry_worker(struct work_struct *work) -{ - struct pqi_ctrl_info *ctrl_info; - - ctrl_info = container_of(work, struct pqi_ctrl_info, - raid_bypass_retry_work); - pqi_retry_raid_bypass_requests(ctrl_info); -} - -static void pqi_clear_all_queued_raid_bypass_retries( - struct pqi_ctrl_info *ctrl_info) -{ - unsigned long flags; - - spin_lock_irqsave(&ctrl_info->raid_bypass_retry_list_lock, flags); - INIT_LIST_HEAD(&ctrl_info->raid_bypass_retry_list); - spin_unlock_irqrestore(&ctrl_info->raid_bypass_retry_list_lock, flags); -} - static void pqi_aio_io_complete(struct pqi_io_request *io_request, void *context) { @@ -5445,12 +5313,11 @@ static void pqi_aio_io_complete(struct pqi_io_request *io_request, scmd = io_request->scmd; scsi_dma_unmap(scmd); - if (io_request->status == -EAGAIN) + if (io_request->status == -EAGAIN || pqi_raid_bypass_retry_needed(io_request)) { set_host_byte(scmd, DID_IMM_RETRY); - else if (pqi_raid_bypass_retry_needed(io_request)) { - pqi_queue_raid_bypass_retry(io_request); - return; + scmd->SCp.this_residual++; } + pqi_free_io_request(io_request); pqi_scsi_done(scmd); } @@ -5667,6 +5534,14 @@ static inline u16 pqi_get_hw_queue(struct pqi_ctrl_info *ctrl_info, return hw_queue; } +static inline bool pqi_is_bypass_eligible_request(struct scsi_cmnd *scmd) +{ + if (blk_rq_is_passthrough(scmd->request)) + return false; + + return scmd->SCp.this_residual == 0; +} + /* * This function gets called just before we hand the completed SCSI request * back to the SML. @@ -5792,9 +5667,7 @@ static int pqi_scsi_queue_command(struct Scsi_Host *shost, struct scsi_cmnd *scm return 0; } - pqi_ctrl_busy(ctrl_info); - if (pqi_ctrl_blocked(ctrl_info) || pqi_device_in_reset(device) || - pqi_ctrl_in_ofa(ctrl_info) || pqi_ctrl_in_shutdown(ctrl_info)) { + if (pqi_ctrl_blocked(ctrl_info)) { rc = SCSI_MLQUEUE_HOST_BUSY; goto out; } @@ -5811,13 +5684,12 @@ static int pqi_scsi_queue_command(struct Scsi_Host *shost, struct scsi_cmnd *scm if (pqi_is_logical_device(device)) { raid_bypassed = false; if (device->raid_bypass_enabled && - !blk_rq_is_passthrough(scmd->request)) { - if (!pqi_is_parity_write_stream(ctrl_info, scmd)) { - rc = pqi_raid_bypass_submit_scsi_cmd(ctrl_info, device, scmd, queue_group); - if (rc == 0 || rc == SCSI_MLQUEUE_HOST_BUSY) { - raid_bypassed = true; - atomic_inc(&device->raid_bypass_cnt); - } + pqi_is_bypass_eligible_request(scmd) && + !pqi_is_parity_write_stream(ctrl_info, scmd)) { + rc = pqi_raid_bypass_submit_scsi_cmd(ctrl_info, device, scmd, queue_group); + if (rc == 0 || rc == SCSI_MLQUEUE_HOST_BUSY) { + raid_bypassed = true; + atomic_inc(&device->raid_bypass_cnt); } } if (!raid_bypassed) @@ -8255,11 +8127,6 @@ static struct pqi_ctrl_info *pqi_alloc_ctrl_info(int numa_node) PQI_RESERVED_IO_SLOTS_SYNCHRONOUS_REQUESTS); init_waitqueue_head(&ctrl_info->block_requests_wait); - INIT_LIST_HEAD(&ctrl_info->raid_bypass_retry_list); - spin_lock_init(&ctrl_info->raid_bypass_retry_list_lock); - INIT_WORK(&ctrl_info->raid_bypass_retry_work, - pqi_raid_bypass_retry_worker); - ctrl_info->ctrl_id = atomic_inc_return(&pqi_controller_count) - 1; ctrl_info->irq_mode = IRQ_MODE_NONE; ctrl_info->max_msix_vectors = PQI_MAX_MSIX_VECTORS; @@ -8585,7 +8452,6 @@ static void pqi_take_ctrl_offline_deferred(struct pqi_ctrl_info *ctrl_info) pqi_cancel_update_time_worker(ctrl_info); pqi_ctrl_wait_until_quiesced(ctrl_info); pqi_fail_all_outstanding_requests(ctrl_info); - pqi_clear_all_queued_raid_bypass_retries(ctrl_info); pqi_ctrl_unblock_requests(ctrl_info); }