packages/rdma-core
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

!189 [sync] PR-187: libhns: Add error logs to help diagnosis

Browse Source 
From: @openeuler-sync-bot 
Reviewed-by: @hellotcc 
Signed-off-by: @hellotcc
This commit is contained in:
openeuler-ci-bot 2025-01-06 07:22:09 +00:00 committed by Gitee
commit fc3e056495
No known key found for this signature in database
GPG Key ID: 173E9B9CA92EEF8F
2 changed files with 250 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

242
0042-libhns-Add-error-logs-to-help-diagnosis.patch Normal file
View File

@ -0,0 +1,242 @@
From 60c45b5f7c2cd0c2e7139d472406f071f327bb91 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>
---
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;
}
--
2.33.0

9
rdma-core.spec
View File

@ -1,6 +1,6 @@
Name: rdma-core
Version: 50.0
Release: 17
Release: 18
Summary: RDMA core userspace libraries and daemons
License: GPL-2.0-only OR BSD-2-Clause AND BSD-3-Clause
Url: https://github.com/linux-rdma/rdma-core
@ -47,6 +47,7 @@ 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
@ -626,6 +627,12 @@ 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