From patchwork Mon Apr 23 16:12:37 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Raju Rangoju X-Patchwork-Id: 10357399 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 4A6F960225 for ; Mon, 23 Apr 2018 16:12:56 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 3AE9226E76 for ; Mon, 23 Apr 2018 16:12:56 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 2E32A237A5; Mon, 23 Apr 2018 16:12:56 +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 385B3237A5 for ; Mon, 23 Apr 2018 16:12:55 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755836AbeDWQMx (ORCPT ); Mon, 23 Apr 2018 12:12:53 -0400 Received: from stargate.chelsio.com ([12.32.117.8]:31879 "EHLO stargate.chelsio.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755772AbeDWQMx (ORCPT ); Mon, 23 Apr 2018 12:12:53 -0400 Received: from localhost ([10.193.185.238]) by stargate.chelsio.com (8.13.8/8.13.8) with ESMTP id w3NGCeid024061; Mon, 23 Apr 2018 09:12:41 -0700 From: Raju Rangoju To: jgg@mellanox.com, dledford@redhat.com, linux-rdma@vger.kernel.org Cc: swise@opengridcomputing.com, rajur@chelsio.com, bharat@chelsio.com, stable@vger.kernel.org Subject: [PATCH] RDMA/cxgb4: release hw resources on device removal Date: Mon, 23 Apr 2018 21:42:37 +0530 Message-Id: <20180423161237.3459-1-rajur@chelsio.com> X-Mailer: git-send-email 2.12.0 Sender: linux-rdma-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-rdma@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP The c4iw_rdev_close() logic was not releasing all the hw resources (PBL and RQT memory) during the device removal event (driver unload / system reboot). This can cause panic in gen_pool_destroy(). The module remove function will wait for all the hw resources to be released during the device removal event. Fixes c12a67fe(iw_cxgb4: free EQ queue memory on last deref) Signed-off-by: Raju Rangoju Reviewed-by: Steve Wise Cc: stable@vger.kernel.org --- drivers/infiniband/hw/cxgb4/device.c | 9 ++++++++- drivers/infiniband/hw/cxgb4/iw_cxgb4.h | 4 ++++ drivers/infiniband/hw/cxgb4/resource.c | 26 ++++++++++++++++++++++++-- 3 files changed, 36 insertions(+), 3 deletions(-) diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c index feeb8ee6f4a2..44161ca4d2a8 100644 --- a/drivers/infiniband/hw/cxgb4/device.c +++ b/drivers/infiniband/hw/cxgb4/device.c @@ -875,6 +875,11 @@ static int c4iw_rdev_open(struct c4iw_rdev *rdev) rdev->status_page->db_off = 0; + init_completion(&rdev->rqt_compl); + init_completion(&rdev->pbl_compl); + kref_init(&rdev->rqt_kref); + kref_init(&rdev->pbl_kref); + return 0; err_free_status_page_and_wr_log: if (c4iw_wr_log && rdev->wr_log) @@ -893,13 +898,15 @@ static int c4iw_rdev_open(struct c4iw_rdev *rdev) static void c4iw_rdev_close(struct c4iw_rdev *rdev) { - destroy_workqueue(rdev->free_workq); kfree(rdev->wr_log); c4iw_release_dev_ucontext(rdev, &rdev->uctx); free_page((unsigned long)rdev->status_page); c4iw_pblpool_destroy(rdev); c4iw_rqtpool_destroy(rdev); + wait_for_completion(&rdev->pbl_compl); + wait_for_completion(&rdev->rqt_compl); c4iw_ocqp_pool_destroy(rdev); + destroy_workqueue(rdev->free_workq); c4iw_destroy_resource(&rdev->resource); } diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h index cc929002c05e..a60def23e9ef 100644 --- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h +++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h @@ -185,6 +185,10 @@ struct c4iw_rdev { struct wr_log_entry *wr_log; int wr_log_size; struct workqueue_struct *free_workq; + struct completion rqt_compl; + struct completion pbl_compl; + struct kref rqt_kref; + struct kref pbl_kref; }; static inline int c4iw_fatal_error(struct c4iw_rdev *rdev) diff --git a/drivers/infiniband/hw/cxgb4/resource.c b/drivers/infiniband/hw/cxgb4/resource.c index 3cf25997ed2b..0ef25ae05e6f 100644 --- a/drivers/infiniband/hw/cxgb4/resource.c +++ b/drivers/infiniband/hw/cxgb4/resource.c @@ -260,12 +260,22 @@ u32 c4iw_pblpool_alloc(struct c4iw_rdev *rdev, int size) rdev->stats.pbl.cur += roundup(size, 1 << MIN_PBL_SHIFT); if (rdev->stats.pbl.cur > rdev->stats.pbl.max) rdev->stats.pbl.max = rdev->stats.pbl.cur; + kref_get(&rdev->pbl_kref); } else rdev->stats.pbl.fail++; mutex_unlock(&rdev->stats.lock); return (u32)addr; } +static void destroy_pblpool(struct kref *kref) +{ + struct c4iw_rdev *rdev; + + rdev = container_of(kref, struct c4iw_rdev, pbl_kref); + gen_pool_destroy(rdev->pbl_pool); + complete(&rdev->pbl_compl); +} + void c4iw_pblpool_free(struct c4iw_rdev *rdev, u32 addr, int size) { pr_debug("addr 0x%x size %d\n", addr, size); @@ -273,6 +283,7 @@ void c4iw_pblpool_free(struct c4iw_rdev *rdev, u32 addr, int size) rdev->stats.pbl.cur -= roundup(size, 1 << MIN_PBL_SHIFT); mutex_unlock(&rdev->stats.lock); gen_pool_free(rdev->pbl_pool, (unsigned long)addr, size); + kref_put(&rdev->pbl_kref, destroy_pblpool); } int c4iw_pblpool_create(struct c4iw_rdev *rdev) @@ -310,7 +321,7 @@ int c4iw_pblpool_create(struct c4iw_rdev *rdev) void c4iw_pblpool_destroy(struct c4iw_rdev *rdev) { - gen_pool_destroy(rdev->pbl_pool); + kref_put(&rdev->pbl_kref, destroy_pblpool); } /* @@ -331,12 +342,22 @@ u32 c4iw_rqtpool_alloc(struct c4iw_rdev *rdev, int size) rdev->stats.rqt.cur += roundup(size << 6, 1 << MIN_RQT_SHIFT); if (rdev->stats.rqt.cur > rdev->stats.rqt.max) rdev->stats.rqt.max = rdev->stats.rqt.cur; + kref_get(&rdev->rqt_kref); } else rdev->stats.rqt.fail++; mutex_unlock(&rdev->stats.lock); return (u32)addr; } +static void destroy_rqtpool(struct kref *kref) +{ + struct c4iw_rdev *rdev; + + rdev = container_of(kref, struct c4iw_rdev, rqt_kref); + gen_pool_destroy(rdev->rqt_pool); + complete(&rdev->rqt_compl); +} + void c4iw_rqtpool_free(struct c4iw_rdev *rdev, u32 addr, int size) { pr_debug("addr 0x%x size %d\n", addr, size << 6); @@ -344,6 +365,7 @@ void c4iw_rqtpool_free(struct c4iw_rdev *rdev, u32 addr, int size) rdev->stats.rqt.cur -= roundup(size << 6, 1 << MIN_RQT_SHIFT); mutex_unlock(&rdev->stats.lock); gen_pool_free(rdev->rqt_pool, (unsigned long)addr, size << 6); + kref_put(&rdev->rqt_kref, destroy_rqtpool); } int c4iw_rqtpool_create(struct c4iw_rdev *rdev) @@ -380,7 +402,7 @@ int c4iw_rqtpool_create(struct c4iw_rdev *rdev) void c4iw_rqtpool_destroy(struct c4iw_rdev *rdev) { - gen_pool_destroy(rdev->rqt_pool); + kref_put(&rdev->rqt_kref, destroy_rqtpool); } /*