From patchwork Fri Feb 6 12:00:15 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Luis Henriques X-Patchwork-Id: 437244 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Received: from huckleberry.canonical.com (huckleberry.canonical.com [91.189.94.19]) by ozlabs.org (Postfix) with ESMTP id 20794140271; Fri, 6 Feb 2015 23:07:40 +1100 (AEDT) Received: from localhost ([127.0.0.1] helo=huckleberry.canonical.com) by huckleberry.canonical.com with esmtp (Exim 4.76) (envelope-from ) id 1YJhhA-00057M-4L; Fri, 06 Feb 2015 12:07:36 +0000 Received: from youngberry.canonical.com ([91.189.89.112]) by huckleberry.canonical.com with esmtp (Exim 4.76) (envelope-from ) id 1YJhbd-0001d5-Vi for kernel-team@lists.ubuntu.com; Fri, 06 Feb 2015 12:01:53 +0000 Received: from av-217-129-142-138.netvisao.pt ([217.129.142.138] helo=localhost) by youngberry.canonical.com with esmtpsa (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1YJhbd-0002ja-Je; Fri, 06 Feb 2015 12:01:53 +0000 From: Luis Henriques To: linux-kernel@vger.kernel.org, stable@vger.kernel.org, kernel-team@lists.ubuntu.com Subject: [PATCH 3.16.y-ckt 037/135] ipr: wait for aborted command responses Date: Fri, 6 Feb 2015 12:00:15 +0000 Message-Id: <1423224113-10958-38-git-send-email-luis.henriques@canonical.com> X-Mailer: git-send-email 2.1.4 In-Reply-To: <1423224113-10958-1-git-send-email-luis.henriques@canonical.com> References: <1423224113-10958-1-git-send-email-luis.henriques@canonical.com> X-Extended-Stable: 3.16 Cc: Brian King , Christoph Hellwig X-BeenThere: kernel-team@lists.ubuntu.com X-Mailman-Version: 2.1.14 Precedence: list List-Id: Kernel team discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Errors-To: kernel-team-bounces@lists.ubuntu.com Sender: kernel-team-bounces@lists.ubuntu.com 3.16.7-ckt6 -stable review patch. If anyone has any objections, please let me know. ------------------ From: Brian King commit 6cdb08172bc89f0a39e1643c5e7eab362692fd1b upstream. Fixes a race condition in abort handling that was injected when multiple interrupt support was added. When only a single interrupt is present, the adapter guarantees it will send responses for aborted commands prior to the response for the abort command itself. With multiple interrupts, these responses generally come back on different interrupts, so we need to ensure the abort thread waits until the aborted command is complete so we don't perform a double completion. This race condition was being hit frequently in environments which were triggering command timeouts, which was resulting in a double completion causing a kernel oops. Signed-off-by: Brian King Reviewed-by: Wendy Xiong Tested-by: Wendy Xiong Signed-off-by: Christoph Hellwig Signed-off-by: Luis Henriques --- drivers/scsi/ipr.c | 92 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ drivers/scsi/ipr.h | 1 + 2 files changed, 93 insertions(+) diff --git a/drivers/scsi/ipr.c b/drivers/scsi/ipr.c index 924b0ba74dfe..748ed71dd92b 100644 --- a/drivers/scsi/ipr.c +++ b/drivers/scsi/ipr.c @@ -683,6 +683,7 @@ static void ipr_init_ipr_cmnd(struct ipr_cmnd *ipr_cmd, ipr_reinit_ipr_cmnd(ipr_cmd); ipr_cmd->u.scratch = 0; ipr_cmd->sibling = NULL; + ipr_cmd->eh_comp = NULL; ipr_cmd->fast_done = fast_done; init_timer(&ipr_cmd->timer); } @@ -848,6 +849,8 @@ static void ipr_scsi_eh_done(struct ipr_cmnd *ipr_cmd) scsi_dma_unmap(ipr_cmd->scsi_cmd); scsi_cmd->scsi_done(scsi_cmd); + if (ipr_cmd->eh_comp) + complete(ipr_cmd->eh_comp); list_add_tail(&ipr_cmd->queue, &ipr_cmd->hrrq->hrrq_free_q); } @@ -4843,6 +4846,84 @@ static int ipr_slave_alloc(struct scsi_device *sdev) return rc; } +/** + * ipr_match_lun - Match function for specified LUN + * @ipr_cmd: ipr command struct + * @device: device to match (sdev) + * + * Returns: + * 1 if command matches sdev / 0 if command does not match sdev + **/ +static int ipr_match_lun(struct ipr_cmnd *ipr_cmd, void *device) +{ + if (ipr_cmd->scsi_cmd && ipr_cmd->scsi_cmd->device == device) + return 1; + return 0; +} + +/** + * ipr_wait_for_ops - Wait for matching commands to complete + * @ipr_cmd: ipr command struct + * @device: device to match (sdev) + * @match: match function to use + * + * Returns: + * SUCCESS / FAILED + **/ +static int ipr_wait_for_ops(struct ipr_ioa_cfg *ioa_cfg, void *device, + int (*match)(struct ipr_cmnd *, void *)) +{ + struct ipr_cmnd *ipr_cmd; + int wait; + unsigned long flags; + struct ipr_hrr_queue *hrrq; + signed long timeout = IPR_ABORT_TASK_TIMEOUT; + DECLARE_COMPLETION_ONSTACK(comp); + + ENTER; + do { + wait = 0; + + for_each_hrrq(hrrq, ioa_cfg) { + spin_lock_irqsave(hrrq->lock, flags); + list_for_each_entry(ipr_cmd, &hrrq->hrrq_pending_q, queue) { + if (match(ipr_cmd, device)) { + ipr_cmd->eh_comp = ∁ + wait++; + } + } + spin_unlock_irqrestore(hrrq->lock, flags); + } + + if (wait) { + timeout = wait_for_completion_timeout(&comp, timeout); + + if (!timeout) { + wait = 0; + + for_each_hrrq(hrrq, ioa_cfg) { + spin_lock_irqsave(hrrq->lock, flags); + list_for_each_entry(ipr_cmd, &hrrq->hrrq_pending_q, queue) { + if (match(ipr_cmd, device)) { + ipr_cmd->eh_comp = NULL; + wait++; + } + } + spin_unlock_irqrestore(hrrq->lock, flags); + } + + if (wait) + dev_err(&ioa_cfg->pdev->dev, "Timed out waiting for aborted commands\n"); + LEAVE; + return wait ? FAILED : SUCCESS; + } + } + } while (wait); + + LEAVE; + return SUCCESS; +} + static int ipr_eh_host_reset(struct scsi_cmnd *cmd) { struct ipr_ioa_cfg *ioa_cfg; @@ -5062,11 +5143,17 @@ static int __ipr_eh_dev_reset(struct scsi_cmnd *scsi_cmd) static int ipr_eh_dev_reset(struct scsi_cmnd *cmd) { int rc; + struct ipr_ioa_cfg *ioa_cfg; + + ioa_cfg = (struct ipr_ioa_cfg *) cmd->device->host->hostdata; spin_lock_irq(cmd->device->host->host_lock); rc = __ipr_eh_dev_reset(cmd); spin_unlock_irq(cmd->device->host->host_lock); + if (rc == SUCCESS) + rc = ipr_wait_for_ops(ioa_cfg, cmd->device, ipr_match_lun); + return rc; } @@ -5244,13 +5331,18 @@ static int ipr_eh_abort(struct scsi_cmnd *scsi_cmd) { unsigned long flags; int rc; + struct ipr_ioa_cfg *ioa_cfg; ENTER; + ioa_cfg = (struct ipr_ioa_cfg *) scsi_cmd->device->host->hostdata; + spin_lock_irqsave(scsi_cmd->device->host->host_lock, flags); rc = ipr_cancel_op(scsi_cmd); spin_unlock_irqrestore(scsi_cmd->device->host->host_lock, flags); + if (rc == SUCCESS) + rc = ipr_wait_for_ops(ioa_cfg, scsi_cmd->device, ipr_match_lun); LEAVE; return rc; } diff --git a/drivers/scsi/ipr.h b/drivers/scsi/ipr.h index 31ed126f7143..c24c747595ba 100644 --- a/drivers/scsi/ipr.h +++ b/drivers/scsi/ipr.h @@ -1607,6 +1607,7 @@ struct ipr_cmnd { struct scsi_device *sdev; } u; + struct completion *eh_comp; struct ipr_hrr_queue *hrrq; struct ipr_ioa_cfg *ioa_cfg; };