From patchwork Thu May 17 08:02:53 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Wei Hu (Xavier)" X-Patchwork-Id: 10405715 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 EE20F60230 for ; Thu, 17 May 2018 07:30:39 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id D57E928879 for ; Thu, 17 May 2018 07:30:39 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id C852028877; Thu, 17 May 2018 07:30:39 +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 5690828877 for ; Thu, 17 May 2018 07:30:39 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752150AbeEQHaK (ORCPT ); Thu, 17 May 2018 03:30:10 -0400 Received: from szxga04-in.huawei.com ([45.249.212.190]:7679 "EHLO huawei.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1751427AbeEQHaJ (ORCPT ); Thu, 17 May 2018 03:30:09 -0400 Received: from DGGEMS405-HUB.china.huawei.com (unknown [172.30.72.58]) by Forcepoint Email with ESMTP id 5E4B6CFACA782; Thu, 17 May 2018 15:29:54 +0800 (CST) Received: from linux-ioko.site (10.71.200.31) by DGGEMS405-HUB.china.huawei.com (10.3.19.205) with Microsoft SMTP Server id 14.3.382.0; Thu, 17 May 2018 15:29:51 +0800 From: "Wei Hu (Xavier)" To: , CC: , , , Subject: [PATCH rdma-next 5/5] RDMA/hns: Fix the illegal memory operation when cross page Date: Thu, 17 May 2018 16:02:53 +0800 Message-ID: <1526544173-106587-6-git-send-email-xavier.huwei@huawei.com> X-Mailer: git-send-email 1.9.1 In-Reply-To: <1526544173-106587-1-git-send-email-xavier.huwei@huawei.com> References: <1526544173-106587-1-git-send-email-xavier.huwei@huawei.com> MIME-Version: 1.0 X-Originating-IP: [10.71.200.31] X-CFilter-Loop: Reflected 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 This patch fixed the potential illegal operation when using the extend sge buffer cross page in post send operation. The bug will cause the calltrace. Reported-by: Jie Chen Reported-by: Xiping Zhang (Francis) Fixes: b1c1583("RDMA/hns: Get rid of virt_to_page and vmap calls after dma_alloc_coherent") Signed-off-by: Wei Hu (Xavier) --- drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 70 +++++++++++++++++++++--------- drivers/infiniband/hw/hns/hns_roce_hw_v2.h | 1 + 2 files changed, 51 insertions(+), 20 deletions(-) diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c index e1c44a6..5393149 100644 --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c @@ -52,6 +52,52 @@ static void set_data_seg_v2(struct hns_roce_v2_wqe_data_seg *dseg, dseg->len = cpu_to_le32(sg->length); } +static void set_extend_sge(struct hns_roce_qp *qp, struct ib_send_wr *wr, + unsigned int *sge_ind) +{ + struct hns_roce_v2_wqe_data_seg *dseg; + struct ib_sge *sg; + int num_in_wqe = 0; + int extend_sge_num; + int fi_sge_num; + int se_sge_num; + int shift; + int i; + + if (qp->ibqp.qp_type == IB_QPT_RC || qp->ibqp.qp_type == IB_QPT_UC) + num_in_wqe = HNS_ROCE_V2_UC_RC_SGE_NUM_IN_WQE; + extend_sge_num = wr->num_sge - num_in_wqe; + sg = wr->sg_list + num_in_wqe; + shift = qp->hr_buf.page_shift; + + /* + * Check whether wr->num_sge sges are in the same page. If not, we + * should calculate how many sges in the first page and the second + * page. + */ + dseg = get_send_extend_sge(qp, (*sge_ind) & (qp->sge.sge_cnt - 1)); + fi_sge_num = (((((u64)dseg >> shift) + 1) << shift) - (u64)dseg) / + sizeof(struct hns_roce_v2_wqe_data_seg); + if (extend_sge_num > fi_sge_num) { + se_sge_num = extend_sge_num - fi_sge_num; + for (i = 0; i < fi_sge_num; i++) { + set_data_seg_v2(dseg++, sg + i); + (*sge_ind)++; + } + dseg = get_send_extend_sge(qp, + (*sge_ind) & (qp->sge.sge_cnt - 1)); + for (i = 0; i < se_sge_num; i++) { + set_data_seg_v2(dseg++, sg + fi_sge_num + i); + (*sge_ind)++; + } + } else { + for (i = 0; i < extend_sge_num; i++) { + set_data_seg_v2(dseg++, sg + i); + (*sge_ind)++; + } + } +} + static int set_rwqe_data_seg(struct ib_qp *ibqp, struct ib_send_wr *wr, struct hns_roce_v2_rc_send_wqe *rc_sq_wqe, void *wqe, unsigned int *sge_ind, @@ -85,7 +131,7 @@ static int set_rwqe_data_seg(struct ib_qp *ibqp, struct ib_send_wr *wr, roce_set_bit(rc_sq_wqe->byte_4, V2_RC_SEND_WQE_BYTE_4_INLINE_S, 1); } else { - if (wr->num_sge <= 2) { + if (wr->num_sge <= HNS_ROCE_V2_UC_RC_SGE_NUM_IN_WQE) { for (i = 0; i < wr->num_sge; i++) { if (likely(wr->sg_list[i].length)) { set_data_seg_v2(dseg, wr->sg_list + i); @@ -98,24 +144,14 @@ static int set_rwqe_data_seg(struct ib_qp *ibqp, struct ib_send_wr *wr, V2_RC_SEND_WQE_BYTE_20_MSG_START_SGE_IDX_S, (*sge_ind) & (qp->sge.sge_cnt - 1)); - for (i = 0; i < 2; i++) { + for (i = 0; i < HNS_ROCE_V2_UC_RC_SGE_NUM_IN_WQE; i++) { if (likely(wr->sg_list[i].length)) { set_data_seg_v2(dseg, wr->sg_list + i); dseg++; } } - dseg = get_send_extend_sge(qp, - (*sge_ind) & (qp->sge.sge_cnt - 1)); - - for (i = 0; i < wr->num_sge - 2; i++) { - if (likely(wr->sg_list[i + 2].length)) { - set_data_seg_v2(dseg, - wr->sg_list + 2 + i); - dseg++; - (*sge_ind)++; - } - } + set_extend_sge(qp, wr, sge_ind); } roce_set_field(rc_sq_wqe->byte_16, @@ -318,13 +354,7 @@ static int hns_roce_v2_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr, memcpy(&ud_sq_wqe->dgid[0], &ah->av.dgid[0], GID_LEN_V2); - dseg = get_send_extend_sge(qp, - sge_ind & (qp->sge.sge_cnt - 1)); - for (i = 0; i < wr->num_sge; i++) { - set_data_seg_v2(dseg + i, wr->sg_list + i); - sge_ind++; - } - + set_extend_sge(qp, wr, &sge_ind); ind++; } else if (ibqp->qp_type == IB_QPT_RC) { rc_sq_wqe = wqe; diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.h b/drivers/infiniband/hw/hns/hns_roce_hw_v2.h index f16df1b..97f9bef 100644 --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.h +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.h @@ -77,6 +77,7 @@ #define HNS_ROCE_V2_MAX_INNER_MTPT_NUM 2 #define HNS_ROCE_INVALID_LKEY 0x100 #define HNS_ROCE_CMQ_TX_TIMEOUT 30000 +#define HNS_ROCE_V2_UC_RC_SGE_NUM_IN_WQE 2 #define HNS_ROCE_CONTEXT_HOP_NUM 1 #define HNS_ROCE_MTT_HOP_NUM 1