From patchwork Mon Oct 30 03:44:37 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Mike Christie X-Patchwork-Id: 10031993 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 0CF93603B4 for ; Mon, 30 Oct 2017 03:45:11 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id F1C4E286BD for ; Mon, 30 Oct 2017 03:45:10 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id E6B1428763; Mon, 30 Oct 2017 03:45:10 +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=-6.9 required=2.0 tests=BAYES_00,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 523DB2875C for ; Mon, 30 Oct 2017 03:45:10 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752272AbdJ3DpJ (ORCPT ); Sun, 29 Oct 2017 23:45:09 -0400 Received: from mx1.redhat.com ([209.132.183.28]:48438 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752266AbdJ3DpH (ORCPT ); Sun, 29 Oct 2017 23:45:07 -0400 Received: from smtp.corp.redhat.com (int-mx04.intmail.prod.int.phx2.redhat.com [10.5.11.14]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 6F2F9C04AC46; Mon, 30 Oct 2017 03:45:07 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mx1.redhat.com 6F2F9C04AC46 Authentication-Results: ext-mx07.extmail.prod.ext.phx2.redhat.com; dmarc=none (p=none dis=none) header.from=redhat.com Authentication-Results: ext-mx07.extmail.prod.ext.phx2.redhat.com; spf=fail smtp.mailfrom=mchristi@redhat.com Received: from rh2.redhat.com (ovpn-120-93.rdu2.redhat.com [10.10.120.93]) by smtp.corp.redhat.com (Postfix) with ESMTP id 373CE5D9C9; Mon, 30 Oct 2017 03:45:06 +0000 (UTC) From: Mike Christie To: martin.petersen@oracle.com, jejb@linux.vnet.ibm.com, linux-scsi@vger.kernel.org, target-devel@vger.kernel.org, nab@linux-iscsi.org Cc: Mike Christie Subject: [PATCH 17/19] tcmu: make ring buffer timer configurable Date: Sun, 29 Oct 2017 22:44:37 -0500 Message-Id: <1509335079-5276-18-git-send-email-mchristi@redhat.com> In-Reply-To: <1509335079-5276-1-git-send-email-mchristi@redhat.com> References: <1509335079-5276-1-git-send-email-mchristi@redhat.com> X-Scanned-By: MIMEDefang 2.79 on 10.5.11.14 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.31]); Mon, 30 Oct 2017 03:45:07 +0000 (UTC) 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 This adds a timer, qfull_time_out, that controls how long a device will wait for ring buffer space to open before failing the commands in the queue. It is useful to separate this timer from the cmd_time_out and default 30 sec one, because for HA setups cmd_time_out may be disbled and 30 seconds is too long to wait when some OSs like ESX will timeout commands after as little as 8 - 15 seconds. Signed-off-by: Mike Christie --- drivers/target/target_core_user.c | 131 +++++++++++++++++++++++++++++--------- 1 file changed, 101 insertions(+), 30 deletions(-) diff --git a/drivers/target/target_core_user.c b/drivers/target/target_core_user.c index 80dd17a..92eb918 100644 --- a/drivers/target/target_core_user.c +++ b/drivers/target/target_core_user.c @@ -142,8 +142,12 @@ struct tcmu_dev { struct idr commands; - struct timer_list timeout; + struct timer_list cmd_timer; unsigned int cmd_time_out; + + struct timer_list qfull_timer; + int qfull_time_out; + struct list_head timedout_entry; spinlock_t nl_cmd_lock; @@ -740,18 +744,14 @@ static inline size_t tcmu_cmd_get_cmd_size(struct tcmu_cmd *tcmu_cmd, return command_size; } -static int tcmu_setup_cmd_timer(struct tcmu_cmd *tcmu_cmd) +static int tcmu_setup_cmd_timer(struct tcmu_cmd *tcmu_cmd, unsigned int tmo, + struct timer_list *timer) { struct tcmu_dev *udev = tcmu_cmd->tcmu_dev; - unsigned long tmo = udev->cmd_time_out; int cmd_id; - /* - * If it was on the cmdr queue waiting we do not reset the timer - * for requeues and when it is finally sent to userspace. - */ if (tcmu_cmd->cmd_id) - return 0; + goto setup_timer; cmd_id = idr_alloc(&udev->commands, tcmu_cmd, 1, USHRT_MAX, GFP_NOWAIT); if (cmd_id < 0) { @@ -760,23 +760,38 @@ static int tcmu_setup_cmd_timer(struct tcmu_cmd *tcmu_cmd) } tcmu_cmd->cmd_id = cmd_id; - if (!tmo) - tmo = TCMU_TIME_OUT; - pr_debug("allocated cmd %u for dev %s tmo %lu\n", tcmu_cmd->cmd_id, udev->name, tmo / MSEC_PER_SEC); +setup_timer: + if (!tmo) + return 0; + tcmu_cmd->deadline = round_jiffies_up(jiffies + msecs_to_jiffies(tmo)); - mod_timer(&udev->timeout, tcmu_cmd->deadline); + mod_timer(timer, tcmu_cmd->deadline); return 0; } static int add_to_cmdr_queue(struct tcmu_cmd *tcmu_cmd) { struct tcmu_dev *udev = tcmu_cmd->tcmu_dev; + unsigned int tmo; int ret; - ret = tcmu_setup_cmd_timer(tcmu_cmd); + /* + * For backwards compat if qfull_time_out is not set use + * cmd_time_out and if that's not set use the default time out. + */ + if (!udev->qfull_time_out) + return -ETIMEDOUT; + else if (udev->qfull_time_out > 0) + tmo = udev->qfull_time_out; + else if (udev->cmd_time_out) + tmo = udev->cmd_time_out; + else + tmo = TCMU_TIME_OUT; + + ret = tcmu_setup_cmd_timer(tcmu_cmd, tmo, &udev->qfull_timer); if (ret) return ret; @@ -900,7 +915,8 @@ static sense_reason_t queue_cmd_ring(struct tcmu_cmd *tcmu_cmd, int *scsi_err) } entry->req.iov_bidi_cnt = iov_cnt; - ret = tcmu_setup_cmd_timer(tcmu_cmd); + ret = tcmu_setup_cmd_timer(tcmu_cmd, udev->cmd_time_out, + &udev->cmd_timer); if (ret) { tcmu_cmd_free_data(tcmu_cmd, tcmu_cmd->dbi_cnt); *scsi_err = TCM_OUT_OF_RESOURCES; @@ -1046,14 +1062,19 @@ static unsigned int tcmu_handle_completions(struct tcmu_dev *udev) handled++; } - if (mb->cmd_tail == mb->cmd_head && list_empty(&udev->cmdr_queue)) { - del_timer(&udev->timeout); - /* - * not more pending or waiting commands so try to reclaim - * blocks if needed. - */ - if (atomic_read(&global_db_count) > TCMU_GLOBAL_MAX_BLOCKS) - schedule_delayed_work(&tcmu_unmap_work, 0); + if (mb->cmd_tail == mb->cmd_head) { + /* no more pending commands */ + del_timer(&udev->cmd_timer); + + if (list_empty(&udev->cmdr_queue)) { + /* + * no more pending or waiting commands so try to + * reclaim blocks if needed. + */ + if (atomic_read(&global_db_count) > + TCMU_GLOBAL_MAX_BLOCKS) + schedule_delayed_work(&tcmu_unmap_work, 0); + } } return handled; @@ -1074,13 +1095,15 @@ static int tcmu_check_expired_cmd(int id, void *p, void *data) return 0; is_running = list_empty(&cmd->cmdr_queue_entry); - pr_debug("Timing out cmd %u on dev %s that is %s.\n", - id, udev->name, is_running ? "inflight" : "queued"); - - se_cmd = cmd->se_cmd; - cmd->se_cmd = NULL; if (is_running) { + /* + * If cmd_time_out is disabled but qfull is set deadline + * will only reflect the qfull timeout. Ignore it. + */ + if (!udev->cmd_time_out) + return 0; + set_bit(TCMU_CMD_BIT_EXPIRED, &cmd->flags); /* * target_complete_cmd will translate this to LUN COMM FAILURE @@ -1093,6 +1116,12 @@ static int tcmu_check_expired_cmd(int id, void *p, void *data) tcmu_free_cmd(cmd); scsi_status = SAM_STAT_TASK_SET_FULL; } + + pr_debug("Timing out cmd %u on dev %s that is %s.\n", + id, udev->name, is_running ? "inflight" : "queued"); + + se_cmd = cmd->se_cmd; + cmd->se_cmd = NULL; target_complete_cmd(se_cmd, scsi_status); return 0; } @@ -1101,7 +1130,7 @@ static void tcmu_device_timedout(unsigned long data) { struct tcmu_dev *udev = (struct tcmu_dev *)data; - pr_debug("%s cmd timeout has expired\n", udev->name); + pr_debug("%s cmd/qfull timeout has expired\n", udev->name); spin_lock(&timed_out_udevs_lock); if (list_empty(&udev->timedout_entry)) @@ -1148,6 +1177,7 @@ static struct se_device *tcmu_alloc_device(struct se_hba *hba, const char *name) udev->hba = hba; udev->cmd_time_out = TCMU_TIME_OUT; + udev->qfull_time_out = -1; mutex_init(&udev->cmdr_lock); @@ -1155,7 +1185,10 @@ static struct se_device *tcmu_alloc_device(struct se_hba *hba, const char *name) INIT_LIST_HEAD(&udev->cmdr_queue); idr_init(&udev->commands); - setup_timer(&udev->timeout, tcmu_device_timedout, + setup_timer(&udev->qfull_timer, tcmu_device_timedout, + (unsigned long)udev); + + setup_timer(&udev->cmd_timer, tcmu_device_timedout, (unsigned long)udev); init_waitqueue_head(&udev->nl_cmd_wq); @@ -1211,6 +1244,8 @@ static bool run_cmdr_queue(struct tcmu_dev *udev) goto done; } } + if (list_empty(&udev->cmdr_queue)) + del_timer(&udev->qfull_timer); done: return drained; } @@ -1710,7 +1745,8 @@ static void tcmu_destroy_device(struct se_device *dev) { struct tcmu_dev *udev = TCMU_DEV(dev); - del_timer_sync(&udev->timeout); + del_timer_sync(&udev->cmd_timer); + del_timer_sync(&udev->qfull_timer); mutex_lock(&root_udev_mutex); list_del(&udev->node); @@ -1890,6 +1926,40 @@ static ssize_t tcmu_cmd_time_out_store(struct config_item *item, const char *pag } CONFIGFS_ATTR(tcmu_, cmd_time_out); +static ssize_t tcmu_qfull_time_out_show(struct config_item *item, char *page) +{ + struct se_dev_attrib *da = container_of(to_config_group(item), + struct se_dev_attrib, da_group); + struct tcmu_dev *udev = TCMU_DEV(da->da_dev); + + return snprintf(page, PAGE_SIZE, "%ld\n", udev->qfull_time_out <= 0 ? + udev->qfull_time_out : + udev->qfull_time_out / MSEC_PER_SEC); +} + +static ssize_t tcmu_qfull_time_out_store(struct config_item *item, + const char *page, size_t count) +{ + struct se_dev_attrib *da = container_of(to_config_group(item), + struct se_dev_attrib, da_group); + struct tcmu_dev *udev = TCMU_DEV(da->da_dev); + s32 val; + int ret; + + ret = kstrtos32(page, 0, &val); + if (ret < 0) + return ret; + + if (val >= 0) { + udev->qfull_time_out = val * MSEC_PER_SEC; + } else { + printk(KERN_ERR "Invalid qfull timeout value %d\n", val); + return -EINVAL; + } + return count; +} +CONFIGFS_ATTR(tcmu_, qfull_time_out); + static ssize_t tcmu_dev_config_show(struct config_item *item, char *page) { struct se_dev_attrib *da = container_of(to_config_group(item), @@ -2035,6 +2105,7 @@ static ssize_t tcmu_emulate_write_cache_store(struct config_item *item, static struct configfs_attribute *tcmu_attrib_attrs[] = { &tcmu_attr_cmd_time_out, + &tcmu_attr_qfull_time_out, &tcmu_attr_dev_config, &tcmu_attr_dev_size, &tcmu_attr_emulate_write_cache,