!191 libhns: Fixed several bugs in libhns to openEuler-24.03-LTS

From: @cxh269 
Reviewed-by: @hellotcc 
Signed-off-by: @hellotcc

0037-libhns-Fix-out-of-order-issue-of-requester-when-sett.patch

@@ -0,0 +1,50 @@
+From 1e168586452dc8ee4b41384d4ba5030408ba28f7 Mon Sep 17 00:00:00 2001
+From: Junxian Huang <huangjunxian6@hisilicon.com>
+Date: Fri, 8 Nov 2024 17:04:09 +0800
+Subject: [PATCH] libhns: Fix out-of-order issue of requester when
+ setting FENCE
+
+mainline inclusion
+from mainline-master
+commit c4119911c212aaa552c9cb928fba0a696640c9b5
+category: bugfix
+bugzilla: https://gitee.com/src-openeuler/rdma-core/issues/IB3ZHQ
+CVE: NA
+Reference: https://github.com/linux-rdma/rdma-core/pull/1513/commits/c4119911c212aaa552c9cb928fba0a696640c9b5
+
+----------------------------------------------------------------------
+
+The FENCE indicator in hns WQE doesn't ensure that response data from
+a previous Read/Atomic operation has been written to the requester's
+memory before the subsequent Send/Write operation is processed. This
+may result in the subsequent Send/Write operation accessing the original
+data in memory instead of the expected response data.
+
+Unlike FENCE, the SO (Strong Order) indicator blocks the subsequent
+operation until the previous response data is written to memory and a
+bresp is returned. Set the SO indicator instead of FENCE to maintain
+strict order.
+
+Fixes: cbdf5e32a855 ("libhns: Reimplement verbs of post_send and post_recv for hip08 RoCE")
+Signed-off-by: Junxian Huang <huangjunxian6@hisilicon.com>
+Signed-off-by: Xinghai Cen <cenxinghai@h-partners.com>
+---
+ providers/hns/hns_roce_u_hw_v2.c | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/providers/hns/hns_roce_u_hw_v2.c b/providers/hns/hns_roce_u_hw_v2.c
+index 9371150..2debcb3 100644
+--- a/providers/hns/hns_roce_u_hw_v2.c
++++ b/providers/hns/hns_roce_u_hw_v2.c
+@@ -1527,7 +1527,7 @@ static int set_rc_wqe(void *wqe, struct hns_roce_qp *qp, struct ibv_send_wr *wr,
+ 
+ 	hr_reg_write_bool(wqe, RCWQE_CQE,
+ 			  !!(wr->send_flags & IBV_SEND_SIGNALED));
+-	hr_reg_write_bool(wqe, RCWQE_FENCE,
++	hr_reg_write_bool(wqe, RCWQE_SO,
+ 			  !!(wr->send_flags & IBV_SEND_FENCE));
+ 	hr_reg_write_bool(wqe, RCWQE_SE,
+ 			  !!(wr->send_flags & IBV_SEND_SOLICITED));
+-- 
+2.25.1
+

0038-libhns-Fix-reference-to-uninitialized-cq-pointer.patch

@@ -0,0 +1,68 @@
+From 18ec396f9e56062a97207643a4c0c453f24e07b1 Mon Sep 17 00:00:00 2001
+From: Chengchang Tang <tangchengchang@huawei.com>
+Date: Fri, 8 Nov 2024 17:04:08 +0800
+Subject: [PATCH] libhns: Fix reference to uninitialized cq pointer
+MIME-Version: 1.0
+Content-Type: text/plain; charset=utf-8
+Content-Transfer-Encoding: 8bit
+
+mainline inclusion
+from mainline-master
+commit 18e3117cdd161a3f40b8a917f24cfb5227a1d75a
+category: bugfix
+bugzilla: https://gitee.com/src-openeuler/rdma-core/issues/IB3ZHQ
+CVE: NA
+Reference: https://github.com/linux-rdma/rdma-core/pull/1513/commits/18e3117cdd161a3f40b8a917f24cfb5227a1d75a
+
+----------------------------------------------------------------------
+
+For QPs which do not have an SQ, such as XRC TGT，the send_cq
+pointer will not be initailized. Since the supported max_gs
+will be 0 in this case, check it and return before referencing
+the send_cq pointer.
+
+Fixes: cbdf5e32a855 ("libhns: Reimplement verbs of post_send and post_recv for hip08 RoCE")
+Signed-off-by: Chengchang Tang <tangchengchang@huawei.com>
+Signed-off-by: Junxian Huang <huangjunxian6@hisilicon.com>
+Signed-off-by: Xinghai Cen <cenxinghai@h-partners.com>
+---
+ providers/hns/hns_roce_u_hw_v2.c | 12 ++++++------
+ 1 file changed, 6 insertions(+), 6 deletions(-)
+
+diff --git a/providers/hns/hns_roce_u_hw_v2.c b/providers/hns/hns_roce_u_hw_v2.c
+index 2debcb3..465ef1e 100644
+--- a/providers/hns/hns_roce_u_hw_v2.c
++++ b/providers/hns/hns_roce_u_hw_v2.c
+@@ -1579,7 +1579,7 @@ int hns_roce_u_v2_post_send(struct ibv_qp *ibvqp, struct ibv_send_wr *wr,
+ 	struct hns_roce_context *ctx = to_hr_ctx(ibvqp->context);
+ 	struct hns_roce_qp *qp = to_hr_qp(ibvqp);
+ 	struct hns_roce_sge_info sge_info = {};
+-	struct hns_roce_rc_sq_wqe *wqe;
++	struct hns_roce_rc_sq_wqe *wqe = NULL;
+ 	struct ibv_qp_attr attr = {};
+ 	unsigned int wqe_idx, nreq;
+ 	int ret;
+@@ -1595,15 +1595,15 @@ int hns_roce_u_v2_post_send(struct ibv_qp *ibvqp, struct ibv_send_wr *wr,
+ 	sge_info.start_idx = qp->next_sge; /* start index of extend sge */
+ 
+ 	for (nreq = 0; wr; ++nreq, wr = wr->next) {
+-		if (hns_roce_v2_wq_overflow(&qp->sq, nreq,
+-					    to_hr_cq(qp->verbs_qp.qp.send_cq))) {
+-			ret = ENOMEM;
++		if (wr->num_sge > (int)qp->sq.max_gs) {
++			ret = qp->sq.max_gs > 0 ? EINVAL : EOPNOTSUPP;
+ 			*bad_wr = wr;
+ 			goto out;
+ 		}
+ 
+-		if (wr->num_sge > qp->sq.max_gs) {
+-			ret = EINVAL;
++		if (hns_roce_v2_wq_overflow(&qp->sq, nreq,
++					    to_hr_cq(qp->verbs_qp.qp.send_cq))) {
++			ret = ENOMEM;
+ 			*bad_wr = wr;
+ 			goto out;
+ 		}
+-- 
+2.25.1
+

0039-libhns-Fix-the-exception-branch-of-wr_start-is-not-l.patch

@@ -0,0 +1,46 @@
+From 816341b19c5e7138d44f3366afcc92d6edf8f434 Mon Sep 17 00:00:00 2001
+From: wenglianfa <wenglianfa@huawei.com>
+Date: Wed, 12 Jun 2024 17:11:13 +0800
+Subject: [PATCH] libhns: Fix the exception branch of wr_start() is not locked
+
+mainline inclusion
+from mainline-master
+commit d1d0ac518e20dc1eca4fcc62d091c69e20b114da
+category: feature
+bugzilla: https://gitee.com/src-openeuler/rdma-core/issues/IB66RT
+CVE: NA
+Reference:
+https://github.com/linux-rdma/rdma-core/pull/commits/d1d0ac518e20dc1eca4fcc62d091c69e20b114da
+
+------------------------------------------------------------------
+
+The provider should provide locking to ensure that ibv_wr_start()
+and ibv_wr_complete()/abort() form a per-QP critical section
+where no other threads can enter.
+
+The exception branch of wr_start() is not locked, fix it here.
+Because check_qp_send () does not require lock protection,
+hns_roce_spin_lock () is placed after check_qp_send ().
+
+Fixes: 36446a56eea5 ("libhns: Extended QP supports the new post send mechanism")
+Signed-off-by: wenglianfa <wenglianfa@huawei.com>
+Signed-off-by: Xinghai Cen <cenxinghai@h-partners.com>
+---
+ providers/hns/hns_roce_u_hw_v2.c | 1 +
+ 1 file changed, 1 insertion(+)
+
+diff --git a/providers/hns/hns_roce_u_hw_v2.c b/providers/hns/hns_roce_u_hw_v2.c
+index 465ef1e..e4232ea 100644
+--- a/providers/hns/hns_roce_u_hw_v2.c
++++ b/providers/hns/hns_roce_u_hw_v2.c
+@@ -2930,6 +2930,7 @@ static void wr_start(struct ibv_qp_ex *ibv_qp)
+ 
+ 	ret = check_qp_send(qp, ctx);
+ 	if (ret) {
++		hns_roce_spin_lock(&qp->sq.hr_lock);
+ 		qp->err = ret;
+ 		return;
+ 	}
+-- 
+2.25.1
+

0040-libhns-Fix-memory-leakage-when-DCA-is-enabled.patch

@@ -0,0 +1,39 @@
+From a73b31041073a45b04535c4183b39195bbebe514 Mon Sep 17 00:00:00 2001
+From: wenglianfa <wenglianfa@huawei.com>
+Date: Thu, 25 Jul 2024 11:06:01 +0800
+Subject: [PATCH] libhns: Fix memory leakage when DCA is enabled
+
+driver inclusion
+category: feature
+bugzilla: https://gitee.com/src-openeuler/rdma-core/issues/IB66RT
+
+------------------------------------------------------------------
+
+After DCA is enabled and a QP is created, the memory block
+applied for DCA is not free when the QP is destroyed. Here
+fix it.
+
+Fixes:26e79aa7eb8d ("libhns: Add support for attaching QP's WQE buffer")
+Signed-off-by: wenglianfa <wenglianfa@huawei.com>
+Signed-off-by: Xinghai Cen <cenxinghai@h-partners.com>
+---
+ providers/hns/hns_roce_u_verbs.c | 3 ++-
+ 1 file changed, 2 insertions(+), 1 deletion(-)
+
+diff --git a/providers/hns/hns_roce_u_verbs.c b/providers/hns/hns_roce_u_verbs.c
+index e30880c..c733b21 100644
+--- a/providers/hns/hns_roce_u_verbs.c
++++ b/providers/hns/hns_roce_u_verbs.c
+@@ -1357,7 +1357,8 @@ static void qp_free_wqe(struct hns_roce_qp *qp)
+ 
+ 	if (qp->rq.wqe_cnt)
+ 		free(qp->rq.wrid);
+-	hns_roce_free_buf(&qp->buf);
++	else
++		hns_roce_free_buf(&qp->buf);
+ }
+ 
+ static int qp_alloc_wqe(struct ibv_qp_init_attr_ex *attr,
+-- 
+2.25.1
+

0041-libhns-Fix-coredump-during-QP-destruction-when-send_.patch

@@ -0,0 +1,54 @@
+From acd954b18ca6075ae13f7a10f65bcf7eb9ac107e Mon Sep 17 00:00:00 2001
+From: Yuyu Li <liyuyu6@huawei.com>
+Date: Mon, 25 Nov 2024 16:13:48 +0800
+Subject: [PATCH] libhns: Fix coredump during QP destruction when send_cq
+ == recv_cq
+
+driver inclusion
+category: feature
+bugzilla: https://gitee.com/src-openeuler/rdma-core/issues/IB7JZL
+
+------------------------------------------------------------------
+
+If the specified send CQ and recv CQ are both
+the same CQ, the QP node in SCQ is not deleted.
+which causes a segfault to occur when recreating
+the QP. Here fix it.
+
+coredump info:
+0x0000ffff8fbc37d4 in list_add_before_
+0x0000ffff8fbc381c in list_add_tail_
+0x0000ffff8fbc9d9c in add_qp_to_cq_list
+0x0000ffff8fbca008 in create_qp
+0x0000ffff8fbca110 in hns_roce_u_create_qp
+0x0000ffff8feae39c in __ibv_create_qp_1_1
+0x0000000000401420 in test_ctrl_path
+
+Fixes:e95b5da96721 ("libhns: Support reporting wc as software mode")
+Signed-off-by: Yuyu Li <liyuyu6@huawei.com>
+Signed-off-by: Xinghai Cen <cenxinghai@h-partners.com>
+---
+ providers/hns/hns_roce_u_hw_v2.c | 7 ++++---
+ 1 file changed, 4 insertions(+), 3 deletions(-)
+
+diff --git a/providers/hns/hns_roce_u_hw_v2.c b/providers/hns/hns_roce_u_hw_v2.c
+index e4232ea..c746e03 100644
+--- a/providers/hns/hns_roce_u_hw_v2.c
++++ b/providers/hns/hns_roce_u_hw_v2.c
+@@ -2006,9 +2006,10 @@ static int hns_roce_u_v2_destroy_qp(struct ibv_qp *ibqp)
+ 		list_del(&qp->rcq_node);
+ 	}
+ 
+-	if (ibqp->send_cq && ibqp->send_cq != ibqp->recv_cq) {
+-		__hns_roce_v2_cq_clean(to_hr_cq(ibqp->send_cq), ibqp->qp_num,
+-				       NULL);
++	if (ibqp->send_cq) {
++		if (ibqp->send_cq != ibqp->recv_cq)
++			__hns_roce_v2_cq_clean(to_hr_cq(ibqp->send_cq), ibqp->qp_num,
++						NULL);
+ 		list_del(&qp->scq_node);
+ 	}
+ 
+-- 
+2.25.1
+

0042-libhns-Add-error-logs-to-help-diagnosis.patch

@@ -0,0 +1,243 @@
+From 71c49ce25111c8e853ad509430f14d22dbe94303 Mon Sep 17 00:00:00 2001
+From: Junxian Huang <huangjunxian6@hisilicon.com>
+Date: Fri, 27 Dec 2024 14:02:29 +0800
+Subject: [PATCH] libhns: Add error logs to help diagnosis
+
+mainline inclusion
+from mainline-master
+commit 7849f1b17f89b8baa0065adaf9cd04204698ea82
+category: feature
+bugzilla: https://gitee.com/src-openeuler/rdma-core/issues/IBFGPH
+CVE: NA
+
+Reference: https://github.com/linux-rdma/rdma-core/pull/1533/commits/7849f1b17f89b8baa0065adaf9cd04204698ea82
+
+----------------------------------------------------------------------
+
+Add error logs to help diagnosis.
+
+Signed-off-by: Junxian Huang <huangjunxian6@hisilicon.com>
+Signed-off-by: Xinghai Cen <cenxinghai@h-partners.com>
+---
+ providers/hns/hns_roce_u.c       |  4 +-
+ providers/hns/hns_roce_u_hw_v2.c |  3 ++
+ providers/hns/hns_roce_u_verbs.c | 79 ++++++++++++++++++++++++++------
+ 3 files changed, 70 insertions(+), 16 deletions(-)
+
+diff --git a/providers/hns/hns_roce_u.c b/providers/hns/hns_roce_u.c
+index e219b9e..ec995e7 100644
+--- a/providers/hns/hns_roce_u.c
++++ b/providers/hns/hns_roce_u.c
+@@ -424,8 +424,10 @@ static struct verbs_context *hns_roce_alloc_context(struct ibv_device *ibdev,
+ 
+ 	context->uar = mmap(NULL, hr_dev->page_size, PROT_READ | PROT_WRITE,
+ 			    MAP_SHARED, cmd_fd, 0);
+-	if (context->uar == MAP_FAILED)
++	if (context->uar == MAP_FAILED) {
++		verbs_err(&context->ibv_ctx, "failed to mmap uar page.\n");
+ 		goto err_set_attr;
++	}
+ 
+ 	if (init_dca_context(context, cmd_fd,
+ 			     &resp, ctx_attr, hr_dev->page_size))
+diff --git a/providers/hns/hns_roce_u_hw_v2.c b/providers/hns/hns_roce_u_hw_v2.c
+index c746e03..0628646 100644
+--- a/providers/hns/hns_roce_u_hw_v2.c
++++ b/providers/hns/hns_roce_u_hw_v2.c
+@@ -3057,6 +3057,9 @@ static int fill_send_wr_ops(const struct ibv_qp_init_attr_ex *attr,
+ 		fill_send_wr_ops_ud(qp_ex);
+ 		break;
+ 	default:
++		verbs_err(verbs_get_ctx(qp_ex->qp_base.context),
++			  "QP type %d not supported for qp_ex send ops.\n",
++			  attr->qp_type);
+ 		return -EOPNOTSUPP;
+ 	}
+ 
+diff --git a/providers/hns/hns_roce_u_verbs.c b/providers/hns/hns_roce_u_verbs.c
+index c733b21..e9acfab 100644
+--- a/providers/hns/hns_roce_u_verbs.c
++++ b/providers/hns/hns_roce_u_verbs.c
+@@ -422,8 +422,11 @@ static int verify_cq_create_attr(struct ibv_cq_init_attr_ex *attr,
+ {
+ 	struct hns_roce_pad *pad = to_hr_pad(attr->parent_domain);
+ 
+-	if (!attr->cqe || attr->cqe > context->max_cqe)
++	if (!attr->cqe || attr->cqe > context->max_cqe) {
++		verbs_err(&context->ibv_ctx, "unsupported cq depth %u.\n",
++			  attr->cqe);
+ 		return EINVAL;
++	}
+ 
+ 	if (!check_comp_mask(attr->comp_mask, CREATE_CQ_SUPPORTED_COMP_MASK)) {
+ 		verbs_err(&context->ibv_ctx, "unsupported cq comps 0x%x\n",
+@@ -431,8 +434,11 @@ static int verify_cq_create_attr(struct ibv_cq_init_attr_ex *attr,
+ 		return EOPNOTSUPP;
+ 	}
+ 
+-	if (!check_comp_mask(attr->wc_flags, CREATE_CQ_SUPPORTED_WC_FLAGS))
++	if (!check_comp_mask(attr->wc_flags, CREATE_CQ_SUPPORTED_WC_FLAGS)) {
++		verbs_err(&context->ibv_ctx, "unsupported wc flags 0x%llx.\n",
++			  attr->wc_flags);
+ 		return EOPNOTSUPP;
++	}
+ 
+ 	if (attr->comp_mask & IBV_CQ_INIT_ATTR_MASK_PD) {
+ 		if (!pad) {
+@@ -504,8 +510,11 @@ static int exec_cq_create_cmd(struct ibv_context *context,
+ 	ret = ibv_cmd_create_cq_ex(context, attr, &cq->verbs_cq,
+ 				   &cmd_ex.ibv_cmd, sizeof(cmd_ex),
+ 				   &resp_ex.ibv_resp, sizeof(resp_ex), 0);
+-	if (ret)
++	if (ret) {
++		verbs_err(verbs_get_ctx(context),
++			  "failed to exec create cq cmd, ret = %d.\n", ret);
+ 		return ret;
++	}
+ 
+ 	cq->cqn = resp_drv->cqn;
+ 	cq->flags = resp_drv->cap_flags;
+@@ -724,13 +733,20 @@ static int verify_srq_create_attr(struct hns_roce_context *context,
+ 				  struct ibv_srq_init_attr_ex *attr)
+ {
+ 	if (attr->srq_type != IBV_SRQT_BASIC &&
+-	    attr->srq_type != IBV_SRQT_XRC)
++	    attr->srq_type != IBV_SRQT_XRC) {
++		verbs_err(&context->ibv_ctx,
++			  "unsupported srq type, type = %d.\n", attr->srq_type);
+ 		return -EINVAL;
++	}
+ 
+ 	if (!attr->attr.max_sge ||
+ 	    attr->attr.max_wr > context->max_srq_wr ||
+-	    attr->attr.max_sge > context->max_srq_sge)
++	    attr->attr.max_sge > context->max_srq_sge) {
++		verbs_err(&context->ibv_ctx,
++			  "invalid srq attr size, max_wr = %u, max_sge = %u.\n",
++			  attr->attr.max_wr, attr->attr.max_sge);
+ 		return -EINVAL;
++	}
+ 
+ 	attr->attr.max_wr = max_t(uint32_t, attr->attr.max_wr,
+ 				  HNS_ROCE_MIN_SRQ_WQE_NUM);
+@@ -862,8 +878,12 @@ static int exec_srq_create_cmd(struct ibv_context *context,
+ 	ret = ibv_cmd_create_srq_ex(context, &srq->verbs_srq, init_attr,
+ 				    &cmd_ex.ibv_cmd, sizeof(cmd_ex),
+ 				    &resp_ex.ibv_resp, sizeof(resp_ex));
+-	if (ret)
++	if (ret) {
++		verbs_err(verbs_get_ctx(context),
++			  "failed to exec create srq cmd, ret = %d.\n",
++			  ret);
+ 		return ret;
++	}
+ 
+ 	srq->srqn = resp_ex.srqn;
+ 	srq->cap_flags = resp_ex.cap_flags;
+@@ -1086,9 +1106,12 @@ static int check_qp_create_mask(struct hns_roce_context *ctx,
+ 				struct ibv_qp_init_attr_ex *attr)
+ {
+ 	struct hns_roce_device *hr_dev = to_hr_dev(ctx->ibv_ctx.context.device);
++	int ret = 0;
+ 
+-	if (!check_comp_mask(attr->comp_mask, CREATE_QP_SUP_COMP_MASK))
+-		return EOPNOTSUPP;
++	if (!check_comp_mask(attr->comp_mask, CREATE_QP_SUP_COMP_MASK)) {
++		ret = EOPNOTSUPP;
++		goto out;
++	}
+ 
+ 	if (attr->comp_mask & IBV_QP_INIT_ATTR_SEND_OPS_FLAGS &&
+ 	    !check_comp_mask(attr->send_ops_flags, SEND_OPS_FLAG_MASK))
+@@ -1102,17 +1125,21 @@ static int check_qp_create_mask(struct hns_roce_context *ctx,
+ 	case IBV_QPT_RC:
+ 	case IBV_QPT_XRC_SEND:
+ 		if (!(attr->comp_mask & IBV_QP_INIT_ATTR_PD))
+-			return EINVAL;
++			ret = EINVAL;
+ 		break;
+ 	case IBV_QPT_XRC_RECV:
+ 		if (!(attr->comp_mask & IBV_QP_INIT_ATTR_XRCD))
+-			return EINVAL;
++			ret = EINVAL;
+ 		break;
+ 	default:
+ 		return EOPNOTSUPP;
+ 	}
+ 
+-	return 0;
++out:
++	if (ret)
++		verbs_err(&ctx->ibv_ctx, "invalid comp_mask 0x%x.\n",
++			  attr->comp_mask);
++	return ret;
+ }
+ 
+ static int hns_roce_qp_has_rq(struct ibv_qp_init_attr_ex *attr)
+@@ -1137,8 +1164,13 @@ static int verify_qp_create_cap(struct hns_roce_context *ctx,
+ 	if (cap->max_send_wr > ctx->max_qp_wr ||
+ 	    cap->max_recv_wr > ctx->max_qp_wr ||
+ 	    cap->max_send_sge > ctx->max_sge  ||
+-	    cap->max_recv_sge > ctx->max_sge)
++	    cap->max_recv_sge > ctx->max_sge) {
++		verbs_err(&ctx->ibv_ctx,
++			  "invalid qp cap size, max_send/recv_wr = {%u, %u}, max_send/recv_sge = {%u, %u}.\n",
++			  cap->max_send_wr, cap->max_recv_wr,
++			  cap->max_send_sge, cap->max_recv_sge);
+ 		return -EINVAL;
++	}
+ 
+ 	has_rq = hns_roce_qp_has_rq(attr);
+ 	if (!has_rq) {
+@@ -1147,12 +1179,20 @@ static int verify_qp_create_cap(struct hns_roce_context *ctx,
+ 	}
+ 
+ 	min_wqe_num = HNS_ROCE_V2_MIN_WQE_NUM;
+-	if (cap->max_send_wr < min_wqe_num)
++	if (cap->max_send_wr < min_wqe_num) {
++		verbs_debug(&ctx->ibv_ctx,
++			    "change sq depth from %u to minimum %u.\n",
++			    cap->max_send_wr, min_wqe_num);
+ 		cap->max_send_wr = min_wqe_num;
++	}
+ 
+ 	if (cap->max_recv_wr) {
+-		if (cap->max_recv_wr < min_wqe_num)
++		if (cap->max_recv_wr < min_wqe_num) {
++			verbs_debug(&ctx->ibv_ctx,
++				    "change rq depth from %u to minimum %u.\n",
++				    cap->max_recv_wr, min_wqe_num);
+ 			cap->max_recv_wr = min_wqe_num;
++		}
+ 
+ 		if (!cap->max_recv_sge)
+ 			return -EINVAL;
+@@ -1646,6 +1686,11 @@ static int qp_exec_create_cmd(struct ibv_qp_init_attr_ex *attr,
+ 	ret = ibv_cmd_create_qp_ex2(&ctx->ibv_ctx.context, &qp->verbs_qp, attr,
+ 				    &cmd_ex.ibv_cmd, sizeof(cmd_ex),
+ 				    &resp_ex.ibv_resp, sizeof(resp_ex));
++	if (ret) {
++		verbs_err(&ctx->ibv_ctx,
++			  "failed to exec create qp cmd, ret = %d.\n", ret);
++		return ret;
++	}
+ 
+ 	qp->flags = resp_ex.drv_payload.cap_flags;
+ 	*dwqe_mmap_key = resp_ex.drv_payload.dwqe_mmap_key;
+@@ -1707,8 +1752,12 @@ static int mmap_dwqe(struct ibv_context *ibv_ctx, struct hns_roce_qp *qp,
+ {
+ 	qp->dwqe_page = mmap(NULL, HNS_ROCE_DWQE_PAGE_SIZE, PROT_WRITE,
+ 			     MAP_SHARED, ibv_ctx->cmd_fd, dwqe_mmap_key);
+-	if (qp->dwqe_page == MAP_FAILED)
++	if (qp->dwqe_page == MAP_FAILED) {
++		verbs_err(verbs_get_ctx(ibv_ctx),
++			  "failed to mmap direct wqe page, QPN = %u.\n",
++			  qp->verbs_qp.qp.qp_num);
+ 		return -EINVAL;
++	}
+ 
+ 	return 0;
+ }
+-- 
+2.25.1
+

rdma-core.spec

@@ -1,8 +1,8 @@
 Name:           rdma-core
 Version:        50.0
-Release:        12
+Release:        18
 Summary:        RDMA core userspace libraries and daemons
-License:        GPLv2 or BSD
+License:        GPL-2.0-only OR BSD-2-Clause AND BSD-3-Clause
 Url:            https://github.com/linux-rdma/rdma-core
 Source:         https://github.com/linux-rdma/rdma-core/releases/download/v%{version}/%{name}-%{version}.tar.gz
 
@@ -41,6 +41,12 @@ patch32: 0032-libhns-Clean-up-signed-unsigned-mix-with-relational-.patch
 patch33: 0033-libhns-Fix-missing-flag-when-creating-qp-by-hnsdv_cr.patch
 patch34: 0034-librdmacm-Fix-an-overflow-bug-in-qsort-comparison-function.patch
 patch35: 0035-Fix-the-stride-calculation-for-MSN-PSN-area.patch
+patch37: 0037-libhns-Fix-out-of-order-issue-of-requester-when-sett.patch
+patch38: 0038-libhns-Fix-reference-to-uninitialized-cq-pointer.patch
+patch39: 0039-libhns-Fix-the-exception-branch-of-wr_start-is-not-l.patch
+patch40: 0040-libhns-Fix-memory-leakage-when-DCA-is-enabled.patch
+patch41: 0041-libhns-Fix-coredump-during-QP-destruction-when-send_.patch
+patch42: 0042-libhns-Add-error-logs-to-help-diagnosis.patch
 
 BuildRequires:  binutils cmake >= 2.8.11 gcc libudev-devel pkgconfig pkgconfig(libnl-3.0)
 BuildRequires:  pkgconfig(libnl-route-3.0) systemd systemd-devel
@@ -251,7 +257,6 @@ Obsoletes: infiniband-diags-help < %{version}-%{release}
 Man pages and other related documents for %{name}.
 
 %prep
-%setup
 %autosetup -v -p1
 
 %build
@@ -313,12 +318,6 @@ install -D -m 0644 ibacm_opts.cfg %{buildroot}%{_sysconfdir}/rdma/
 rm -rf %{buildroot}/%{_initrddir}/
 rm -f %{buildroot}/%{_sbindir}/srp_daemon.sh
 
-%ldconfig_scriptlets -n libibverbs
-
-%ldconfig_scriptlets -n libibumad
-
-%ldconfig_scriptlets -n librdmacm
-
 %post
 if [ -x /sbin/udevadm ];then
 /sbin/udevadm trigger --subsystem-match=infiniband --action=change || true
@@ -623,6 +622,43 @@ fi
 %doc %{_docdir}/%{name}-%{version}/70-persistent-ipoib.rules
 
 %changelog
+* Fri Jan 3 2025 Xinghai Cen <cenxinghai@h-partners.com> - 50.0-18
+- Type: requirement
+- ID: NA
+- SUG: NA
+- DESC: Add error logs to help diagnosis
+
+* Thu Nov 28 2024 Xinghai Cen <cenxinghai@h-partners.com> - 50.0-17
+- Type: bugfix
+- ID: NA
+- SUG: NA
+- DESC: Fix coredump during QP destruction when send_cq == recv_cq
+
+* Mon Nov 25 2024 Xinghai Cen <cenxinghai@h-partners.com> - 50.0-16
+- Type: bugfix
+- ID: NA
+- SUG: NA
+- DESC: Fixes several bugs for libhns
+
+* Fri Nov 15 2024 Xinghai Cen <cenxinghai@h-partners.com> - 50.0-15
+- Type: bugfix
+- ID: NA
+- SUG: NA
+- DESC: Two bugfixes in post_send flow
+
+* Fri Nov 08 2024 Funda Wang <fundawang@yeah.net> - 50.0-14
+- adopt to new cmake macro
+- migrated to SPDX license
+- drop useless setup macro, cause it duplicates with autosetup
+- drop useless ldconfig_scriptlets, it has been done through glibc's
+  filetriggers since openeuler 1.0
+
+* Tue Jan 7 2025 Xinghai Cen <cenxinghai@h-partners.com> - 50.0-13
+- Type: requirment
+- ID: NA
+- SUG: NA
+- DESC: No code modification, just a pile
+
 * Wed Jul 17 2024 dfh <dufuhang@kylinos.cn> - 50.0-12
 - Type: bugfix
 - ID: NA