* [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11
@ 2020-12-04 10:40 Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 01/11] RDMA/hns: Limit the length of data copied between kernel and userspace Weihang Li
` (10 more replies)
0 siblings, 11 replies; 16+ messages in thread
From: Weihang Li @ 2020-12-04 10:40 UTC (permalink / raw)
To: dledford, jgg; +Cc: leon, linux-rdma, linuxarm
There are miscellaneous updates for hns driver:
* #1 fixes a potential length issue when copying udata.
* #2 fixes the unreasonable judgment when using HEM of SRQ and SCCC.
* #3 fixes wrong value of Traffic Class.
* #4 and #5 fix issues about Service Level.
* #6 ~ #11 are cleanups, including removing dead code, fixing coding style
issues and so on.
Changes since v1:
- Only do shift on tclass when using RoCEv2 in #3.
Previous version:
v1: https://patchwork.kernel.org/project/linux-rdma/cover/1606899553-54592-1-git-send-email-liweihang@huawei.com/
Lang Cheng (1):
RDMA/hns: Fix coding style issues
Weihang Li (3):
RDMA/hns: Do shift on traffic class when using RoCEv2
RDMA/hns: Avoid filling sl in high 3 bits of vlan_id
RDMA/hns: WARN_ON if get a reserved sl from users
Wenpeng Liang (3):
RDMA/hns: Limit the length of data copied between kernel and userspace
RDMA/hns: Normalization the judgment of some features
RDMA/hns: Fix incorrect symbol types
Xinhao Liu (1):
RDMA/hns: Clear redundant variable initialization
Yixian Liu (2):
RDMA/hns: Remove unnecessary access right set during INIT2INIT
RDMA/hns: Simplify AEQE process for different types of queue
Yixing Liu (1):
RDMA/hns: Fix inaccurate prints
drivers/infiniband/hw/hns/hns_roce_ah.c | 13 +--
drivers/infiniband/hw/hns/hns_roce_alloc.c | 4 +-
drivers/infiniband/hw/hns/hns_roce_cmd.c | 37 +++---
drivers/infiniband/hw/hns/hns_roce_cmd.h | 6 +-
drivers/infiniband/hw/hns/hns_roce_common.h | 14 +--
drivers/infiniband/hw/hns/hns_roce_cq.c | 42 +++----
drivers/infiniband/hw/hns/hns_roce_db.c | 8 +-
drivers/infiniband/hw/hns/hns_roce_device.h | 87 ++++++--------
drivers/infiniband/hw/hns/hns_roce_hem.c | 42 +++----
drivers/infiniband/hw/hns/hns_roce_hem.h | 2 +-
drivers/infiniband/hw/hns/hns_roce_hw_v1.c | 41 +++----
drivers/infiniband/hw/hns/hns_roce_hw_v1.h | 2 +-
drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 174 +++++++++-------------------
drivers/infiniband/hw/hns/hns_roce_hw_v2.h | 6 +-
drivers/infiniband/hw/hns/hns_roce_main.c | 19 +--
drivers/infiniband/hw/hns/hns_roce_mr.c | 25 ++--
drivers/infiniband/hw/hns/hns_roce_pd.c | 13 ++-
drivers/infiniband/hw/hns/hns_roce_qp.c | 82 +++++++------
drivers/infiniband/hw/hns/hns_roce_srq.c | 48 ++++----
19 files changed, 295 insertions(+), 370 deletions(-)
--
2.8.1
^ permalink raw reply [flat|nested] 16+ messages in thread
* [PATCH v2 for-next 01/11] RDMA/hns: Limit the length of data copied between kernel and userspace
2020-12-04 10:40 [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11 Weihang Li
@ 2020-12-04 10:40 ` Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 02/11] RDMA/hns: Normalization the judgment of some features Weihang Li
` (9 subsequent siblings)
10 siblings, 0 replies; 16+ messages in thread
From: Weihang Li @ 2020-12-04 10:40 UTC (permalink / raw)
To: dledford, jgg; +Cc: leon, linux-rdma, linuxarm
From: Wenpeng Liang <liangwenpeng@huawei.com>
For ib_copy_from_user(), the length of udata may not be the same as that
of cmd. For ib_copy_to_user(), the length of udata may not be the same as
that of resp. So limit the length to prevent out-of-bounds read and write
operations from ib_copy_from_user() and ib_copy_to_user().
Fixes: de77503a5940 ("RDMA/hns: RDMA/hns: Assign rq head pointer when enable rq record db")
Fixes: 633fb4d9fdaa ("RDMA/hns: Use structs to describe the uABI instead of opencoding")
Fixes: ae85bf92effc ("RDMA/hns: Optimize qp param setup flow")
Fixes: 6fd610c5733d ("RDMA/hns: Support 0 hop addressing for SRQ buffer")
Fixes: 9d9d4ff78884 ("RDMA/hns: Update the kernel header file of hns")
Signed-off-by: Wenpeng Liang <liangwenpeng@huawei.com>
Signed-off-by: Weihang Li <liweihang@huawei.com>
---
drivers/infiniband/hw/hns/hns_roce_cq.c | 5 +++--
drivers/infiniband/hw/hns/hns_roce_main.c | 3 ++-
drivers/infiniband/hw/hns/hns_roce_pd.c | 11 ++++++-----
drivers/infiniband/hw/hns/hns_roce_qp.c | 9 ++++++---
drivers/infiniband/hw/hns/hns_roce_srq.c | 10 +++++-----
5 files changed, 22 insertions(+), 16 deletions(-)
diff --git a/drivers/infiniband/hw/hns/hns_roce_cq.c b/drivers/infiniband/hw/hns/hns_roce_cq.c
index 68f355f..b94a560 100644
--- a/drivers/infiniband/hw/hns/hns_roce_cq.c
+++ b/drivers/infiniband/hw/hns/hns_roce_cq.c
@@ -277,7 +277,7 @@ int hns_roce_create_cq(struct ib_cq *ib_cq, const struct ib_cq_init_attr *attr,
if (udata) {
ret = ib_copy_from_udata(&ucmd, udata,
- min(sizeof(ucmd), udata->inlen));
+ min(udata->inlen, sizeof(ucmd)));
if (ret) {
ibdev_err(ibdev, "Failed to copy CQ udata, err %d\n",
ret);
@@ -316,7 +316,8 @@ int hns_roce_create_cq(struct ib_cq *ib_cq, const struct ib_cq_init_attr *attr,
if (udata) {
resp.cqn = hr_cq->cqn;
- ret = ib_copy_to_udata(udata, &resp, sizeof(resp));
+ ret = ib_copy_to_udata(udata, &resp,
+ min(udata->outlen, sizeof(resp)));
if (ret)
goto err_cqc;
}
diff --git a/drivers/infiniband/hw/hns/hns_roce_main.c b/drivers/infiniband/hw/hns/hns_roce_main.c
index f01590d..cab95fd 100644
--- a/drivers/infiniband/hw/hns/hns_roce_main.c
+++ b/drivers/infiniband/hw/hns/hns_roce_main.c
@@ -328,7 +328,8 @@ static int hns_roce_alloc_ucontext(struct ib_ucontext *uctx,
resp.cqe_size = hr_dev->caps.cqe_sz;
- ret = ib_copy_to_udata(udata, &resp, sizeof(resp));
+ ret = ib_copy_to_udata(udata, &resp,
+ min(udata->outlen, sizeof(resp)));
if (ret)
goto error_fail_copy_to_udata;
diff --git a/drivers/infiniband/hw/hns/hns_roce_pd.c b/drivers/infiniband/hw/hns/hns_roce_pd.c
index 98f6949..f78fa1d 100644
--- a/drivers/infiniband/hw/hns/hns_roce_pd.c
+++ b/drivers/infiniband/hw/hns/hns_roce_pd.c
@@ -70,16 +70,17 @@ int hns_roce_alloc_pd(struct ib_pd *ibpd, struct ib_udata *udata)
}
if (udata) {
- struct hns_roce_ib_alloc_pd_resp uresp = {.pdn = pd->pdn};
+ struct hns_roce_ib_alloc_pd_resp resp = {.pdn = pd->pdn};
- if (ib_copy_to_udata(udata, &uresp, sizeof(uresp))) {
+ ret = ib_copy_to_udata(udata, &resp,
+ min(udata->outlen, sizeof(resp)));
+ if (ret) {
hns_roce_pd_free(to_hr_dev(ib_dev), pd->pdn);
- ibdev_err(ib_dev, "failed to copy to udata\n");
- return -EFAULT;
+ ibdev_err(ib_dev, "failed to copy to udata, ret = %d\n", ret);
}
}
- return 0;
+ return ret;
}
int hns_roce_dealloc_pd(struct ib_pd *pd, struct ib_udata *udata)
diff --git a/drivers/infiniband/hw/hns/hns_roce_qp.c b/drivers/infiniband/hw/hns/hns_roce_qp.c
index 34aa086..ebf152f 100644
--- a/drivers/infiniband/hw/hns/hns_roce_qp.c
+++ b/drivers/infiniband/hw/hns/hns_roce_qp.c
@@ -925,9 +925,12 @@ static int set_qp_param(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp,
}
if (udata) {
- if (ib_copy_from_udata(ucmd, udata, sizeof(*ucmd))) {
- ibdev_err(ibdev, "Failed to copy QP ucmd\n");
- return -EFAULT;
+ ret = ib_copy_from_udata(ucmd, udata,
+ min(udata->inlen, sizeof(*ucmd)));
+ if (ret) {
+ ibdev_err(ibdev,
+ "failed to copy QP ucmd, ret = %d\n", ret);
+ return ret;
}
ret = set_user_sq_size(hr_dev, &init_attr->cap, hr_qp, ucmd);
diff --git a/drivers/infiniband/hw/hns/hns_roce_srq.c b/drivers/infiniband/hw/hns/hns_roce_srq.c
index 27646b9..6ce250e 100644
--- a/drivers/infiniband/hw/hns/hns_roce_srq.c
+++ b/drivers/infiniband/hw/hns/hns_roce_srq.c
@@ -304,7 +304,8 @@ int hns_roce_create_srq(struct ib_srq *ib_srq,
srq->max_gs = init_attr->attr.max_sge;
if (udata) {
- ret = ib_copy_from_udata(&ucmd, udata, sizeof(ucmd));
+ ret = ib_copy_from_udata(&ucmd, udata,
+ min(udata->inlen, sizeof(ucmd)));
if (ret) {
ibdev_err(ibdev, "Failed to copy SRQ udata, err %d\n",
ret);
@@ -347,11 +348,10 @@ int hns_roce_create_srq(struct ib_srq *ib_srq,
resp.srqn = srq->srqn;
if (udata) {
- if (ib_copy_to_udata(udata, &resp,
- min(udata->outlen, sizeof(resp)))) {
- ret = -EFAULT;
+ ret = ib_copy_to_udata(udata, &resp,
+ min(udata->outlen, sizeof(resp)));
+ if (ret)
goto err_srqc_alloc;
- }
}
return 0;
--
2.8.1
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH v2 for-next 02/11] RDMA/hns: Normalization the judgment of some features
2020-12-04 10:40 [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11 Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 01/11] RDMA/hns: Limit the length of data copied between kernel and userspace Weihang Li
@ 2020-12-04 10:40 ` Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 03/11] RDMA/hns: Do shift on traffic class when using RoCEv2 Weihang Li
` (8 subsequent siblings)
10 siblings, 0 replies; 16+ messages in thread
From: Weihang Li @ 2020-12-04 10:40 UTC (permalink / raw)
To: dledford, jgg; +Cc: leon, linux-rdma, linuxarm
From: Wenpeng Liang <liangwenpeng@huawei.com>
Whether to enable the these features should better depend on the enable
flags, not the value of related fields.
Fixes: 5c1f167af112 ("RDMA/hns: Init SRQ table for hip08")
Fixes: 3cb2c996c9dc ("RDMA/hns: Add support for SCCC in size of 64 Bytes")
Signed-off-by: Wenpeng Liang <liangwenpeng@huawei.com>
Signed-off-by: Weihang Li <liweihang@huawei.com>
---
drivers/infiniband/hw/hns/hns_roce_hem.c | 4 ++--
drivers/infiniband/hw/hns/hns_roce_main.c | 8 ++++----
drivers/infiniband/hw/hns/hns_roce_qp.c | 2 +-
3 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/infiniband/hw/hns/hns_roce_hem.c b/drivers/infiniband/hw/hns/hns_roce_hem.c
index 5c302ae..359e5dd 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hem.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hem.c
@@ -1028,7 +1028,7 @@ void hns_roce_cleanup_hem_table(struct hns_roce_dev *hr_dev,
void hns_roce_cleanup_hem(struct hns_roce_dev *hr_dev)
{
- if (hr_dev->caps.srqc_entry_sz)
+ if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_SRQ)
hns_roce_cleanup_hem_table(hr_dev,
&hr_dev->srq_table.table);
hns_roce_cleanup_hem_table(hr_dev, &hr_dev->cq_table.table);
@@ -1038,7 +1038,7 @@ void hns_roce_cleanup_hem(struct hns_roce_dev *hr_dev)
if (hr_dev->caps.cqc_timer_entry_sz)
hns_roce_cleanup_hem_table(hr_dev,
&hr_dev->cqc_timer_table);
- if (hr_dev->caps.sccc_sz)
+ if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_QP_FLOW_CTRL)
hns_roce_cleanup_hem_table(hr_dev,
&hr_dev->qp_table.sccc_table);
if (hr_dev->caps.trrl_entry_sz)
diff --git a/drivers/infiniband/hw/hns/hns_roce_main.c b/drivers/infiniband/hw/hns/hns_roce_main.c
index cab95fd..9ca4892 100644
--- a/drivers/infiniband/hw/hns/hns_roce_main.c
+++ b/drivers/infiniband/hw/hns/hns_roce_main.c
@@ -607,7 +607,7 @@ static int hns_roce_init_hem(struct hns_roce_dev *hr_dev)
goto err_unmap_trrl;
}
- if (hr_dev->caps.srqc_entry_sz) {
+ if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_SRQ) {
ret = hns_roce_init_hem_table(hr_dev, &hr_dev->srq_table.table,
HEM_TYPE_SRQC,
hr_dev->caps.srqc_entry_sz,
@@ -619,7 +619,7 @@ static int hns_roce_init_hem(struct hns_roce_dev *hr_dev)
}
}
- if (hr_dev->caps.sccc_sz) {
+ if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_QP_FLOW_CTRL) {
ret = hns_roce_init_hem_table(hr_dev,
&hr_dev->qp_table.sccc_table,
HEM_TYPE_SCCC,
@@ -680,11 +680,11 @@ static int hns_roce_init_hem(struct hns_roce_dev *hr_dev)
hns_roce_cleanup_hem_table(hr_dev, &hr_dev->qpc_timer_table);
err_unmap_ctx:
- if (hr_dev->caps.sccc_sz)
+ if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_QP_FLOW_CTRL)
hns_roce_cleanup_hem_table(hr_dev,
&hr_dev->qp_table.sccc_table);
err_unmap_srq:
- if (hr_dev->caps.srqc_entry_sz)
+ if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_SRQ)
hns_roce_cleanup_hem_table(hr_dev, &hr_dev->srq_table.table);
err_unmap_cq:
diff --git a/drivers/infiniband/hw/hns/hns_roce_qp.c b/drivers/infiniband/hw/hns/hns_roce_qp.c
index ebf152f..0e784c5 100644
--- a/drivers/infiniband/hw/hns/hns_roce_qp.c
+++ b/drivers/infiniband/hw/hns/hns_roce_qp.c
@@ -335,7 +335,7 @@ static int alloc_qpc(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp)
}
}
- if (hr_dev->caps.sccc_sz) {
+ if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_QP_FLOW_CTRL) {
/* Alloc memory for SCC CTX */
ret = hns_roce_table_get(hr_dev, &qp_table->sccc_table,
hr_qp->qpn);
--
2.8.1
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH v2 for-next 03/11] RDMA/hns: Do shift on traffic class when using RoCEv2
2020-12-04 10:40 [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11 Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 01/11] RDMA/hns: Limit the length of data copied between kernel and userspace Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 02/11] RDMA/hns: Normalization the judgment of some features Weihang Li
@ 2020-12-04 10:40 ` Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 04/11] RDMA/hns: Avoid filling sl in high 3 bits of vlan_id Weihang Li
` (7 subsequent siblings)
10 siblings, 0 replies; 16+ messages in thread
From: Weihang Li @ 2020-12-04 10:40 UTC (permalink / raw)
To: dledford, jgg; +Cc: leon, linux-rdma, linuxarm
The high 6 bits of traffic class in GRH is DSCP (Differentiated Services
Codepoint), the driver should shift it before the hardware gets it when
using RoCEv2.
Fixes: 606bf89e98ef ("RDMA/hns: Refactor for hns_roce_v2_modify_qp function")
Fixes: fba429fcf9a5 ("RDMA/hns: Fix missing fields in address vector")
Signed-off-by: Weihang Li <liweihang@huawei.com>
---
drivers/infiniband/hw/hns/hns_roce_ah.c | 2 +-
drivers/infiniband/hw/hns/hns_roce_device.h | 8 ++++++++
drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 10 +++-------
3 files changed, 12 insertions(+), 8 deletions(-)
diff --git a/drivers/infiniband/hw/hns/hns_roce_ah.c b/drivers/infiniband/hw/hns/hns_roce_ah.c
index b09ef33..c9a44db 100644
--- a/drivers/infiniband/hw/hns/hns_roce_ah.c
+++ b/drivers/infiniband/hw/hns/hns_roce_ah.c
@@ -77,7 +77,7 @@ int hns_roce_create_ah(struct ib_ah *ibah, struct rdma_ah_init_attr *init_attr,
ah->av.flowlabel = grh->flow_label;
ah->av.udp_sport = get_ah_udp_sport(ah_attr);
ah->av.sl = rdma_ah_get_sl(ah_attr);
- ah->av.tclass = grh->traffic_class;
+ ah->av.tclass = get_tclass(grh);
memcpy(ah->av.dgid, grh->dgid.raw, HNS_ROCE_GID_SIZE);
memcpy(ah->av.mac, ah_attr->roce.dmac, ETH_ALEN);
diff --git a/drivers/infiniband/hw/hns/hns_roce_device.h b/drivers/infiniband/hw/hns/hns_roce_device.h
index a5c6bb0..c7044f7 100644
--- a/drivers/infiniband/hw/hns/hns_roce_device.h
+++ b/drivers/infiniband/hw/hns/hns_roce_device.h
@@ -1157,6 +1157,14 @@ static inline u32 to_hr_hem_entries_shift(u32 count, u32 buf_shift)
return ilog2(to_hr_hem_entries_count(count, buf_shift));
}
+#define DSCP_SHIFT 2
+
+static inline u8 get_tclass(const struct ib_global_route *grh)
+{
+ return grh->sgid_attr->gid_type == IB_GID_TYPE_ROCE_UDP_ENCAP ?
+ grh->traffic_class >> DSCP_SHIFT : grh->traffic_class;
+}
+
int hns_roce_init_uar_table(struct hns_roce_dev *dev);
int hns_roce_uar_alloc(struct hns_roce_dev *dev, struct hns_roce_uar *uar);
void hns_roce_uar_free(struct hns_roce_dev *dev, struct hns_roce_uar *uar);
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
index 8d37067..7a0c1ab 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
@@ -4596,15 +4596,11 @@ static int hns_roce_v2_set_path(struct ib_qp *ibqp,
roce_set_field(qpc_mask->byte_24_mtu_tc, V2_QPC_BYTE_24_HOP_LIMIT_M,
V2_QPC_BYTE_24_HOP_LIMIT_S, 0);
- if (is_udp)
- roce_set_field(context->byte_24_mtu_tc, V2_QPC_BYTE_24_TC_M,
- V2_QPC_BYTE_24_TC_S, grh->traffic_class >> 2);
- else
- roce_set_field(context->byte_24_mtu_tc, V2_QPC_BYTE_24_TC_M,
- V2_QPC_BYTE_24_TC_S, grh->traffic_class);
-
+ roce_set_field(context->byte_24_mtu_tc, V2_QPC_BYTE_24_TC_M,
+ V2_QPC_BYTE_24_TC_S, get_tclass(&attr->ah_attr.grh));
roce_set_field(qpc_mask->byte_24_mtu_tc, V2_QPC_BYTE_24_TC_M,
V2_QPC_BYTE_24_TC_S, 0);
+
roce_set_field(context->byte_28_at_fl, V2_QPC_BYTE_28_FL_M,
V2_QPC_BYTE_28_FL_S, grh->flow_label);
roce_set_field(qpc_mask->byte_28_at_fl, V2_QPC_BYTE_28_FL_M,
--
2.8.1
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH v2 for-next 04/11] RDMA/hns: Avoid filling sl in high 3 bits of vlan_id
2020-12-04 10:40 [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11 Weihang Li
` (2 preceding siblings ...)
2020-12-04 10:40 ` [PATCH v2 for-next 03/11] RDMA/hns: Do shift on traffic class when using RoCEv2 Weihang Li
@ 2020-12-04 10:40 ` Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 05/11] RDMA/hns: WARN_ON if get a reserved sl from users Weihang Li
` (6 subsequent siblings)
10 siblings, 0 replies; 16+ messages in thread
From: Weihang Li @ 2020-12-04 10:40 UTC (permalink / raw)
To: dledford, jgg; +Cc: leon, linux-rdma, linuxarm
Only the low 12 bits of vlan_id is valid, and service level has been filled
in Address Vector. So there is no need to fill sl in vlan_id in Address
Vector.
Signed-off-by: Weihang Li <liweihang@huawei.com>
---
drivers/infiniband/hw/hns/hns_roce_ah.c | 11 +----------
1 file changed, 1 insertion(+), 10 deletions(-)
diff --git a/drivers/infiniband/hw/hns/hns_roce_ah.c b/drivers/infiniband/hw/hns/hns_roce_ah.c
index c9a44db..cc258ed 100644
--- a/drivers/infiniband/hw/hns/hns_roce_ah.c
+++ b/drivers/infiniband/hw/hns/hns_roce_ah.c
@@ -36,9 +36,6 @@
#include <rdma/ib_cache.h>
#include "hns_roce_device.h"
-#define VLAN_SL_MASK 7
-#define VLAN_SL_SHIFT 13
-
static inline u16 get_ah_udp_sport(const struct rdma_ah_attr *ah_attr)
{
u32 fl = ah_attr->grh.flow_label;
@@ -84,18 +81,12 @@ int hns_roce_create_ah(struct ib_ah *ibah, struct rdma_ah_init_attr *init_attr,
/* HIP08 needs to record vlan info in Address Vector */
if (hr_dev->pci_dev->revision <= PCI_REVISION_ID_HIP08) {
- ah->av.vlan_en = 0;
-
ret = rdma_read_gid_l2_fields(ah_attr->grh.sgid_attr,
&ah->av.vlan_id, NULL);
if (ret)
return ret;
- if (ah->av.vlan_id < VLAN_N_VID) {
- ah->av.vlan_en = 1;
- ah->av.vlan_id |= (rdma_ah_get_sl(ah_attr) & VLAN_SL_MASK) <<
- VLAN_SL_SHIFT;
- }
+ ah->av.vlan_en = ah->av.vlan_id < VLAN_N_VID;
}
return ret;
--
2.8.1
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH v2 for-next 05/11] RDMA/hns: WARN_ON if get a reserved sl from users
2020-12-04 10:40 [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11 Weihang Li
` (3 preceding siblings ...)
2020-12-04 10:40 ` [PATCH v2 for-next 04/11] RDMA/hns: Avoid filling sl in high 3 bits of vlan_id Weihang Li
@ 2020-12-04 10:40 ` Weihang Li
2020-12-09 21:09 ` Jason Gunthorpe
2020-12-04 10:40 ` [PATCH v2 for-next 06/11] RDMA/hns: Remove unnecessary access right set during INIT2INIT Weihang Li
` (5 subsequent siblings)
10 siblings, 1 reply; 16+ messages in thread
From: Weihang Li @ 2020-12-04 10:40 UTC (permalink / raw)
To: dledford, jgg; +Cc: leon, linux-rdma, linuxarm
According to the RoCE v1 specification, the sl (service level) 0-7 are
mapped directly to priorities 0-7 respectively, sl 8-15 are reserved. The
driver should verify whether the value of sl is larger than 7, if so, an
exception should be returned.
Fixes: 172505cfa3a8 ("RDMA/hns: Add check for the validity of sl configuration")
Fixes: d6a3627e311c ("RDMA/hns: Optimize wqe buffer set flow for post send")
Signed-off-by: Weihang Li <liweihang@huawei.com>
---
drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
index 7a0c1ab..15e1313 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
@@ -433,6 +433,10 @@ static int fill_ud_av(struct hns_roce_v2_ud_send_wqe *ud_sq_wqe,
V2_UD_SEND_WQE_BYTE_36_TCLASS_S, ah->av.tclass);
roce_set_field(ud_sq_wqe->byte_40, V2_UD_SEND_WQE_BYTE_40_FLOW_LABEL_M,
V2_UD_SEND_WQE_BYTE_40_FLOW_LABEL_S, ah->av.flowlabel);
+
+ if (WARN_ON(ah->av.sl > MAX_SERVICE_LEVEL))
+ return -EINVAL;
+
roce_set_field(ud_sq_wqe->byte_40, V2_UD_SEND_WQE_BYTE_40_SL_M,
V2_UD_SEND_WQE_BYTE_40_SL_S, ah->av.sl);
@@ -4609,12 +4613,8 @@ static int hns_roce_v2_set_path(struct ib_qp *ibqp,
memset(qpc_mask->dgid, 0, sizeof(grh->dgid.raw));
hr_qp->sl = rdma_ah_get_sl(&attr->ah_attr);
- if (unlikely(hr_qp->sl > MAX_SERVICE_LEVEL)) {
- ibdev_err(ibdev,
- "failed to fill QPC, sl (%d) shouldn't be larger than %d.\n",
- hr_qp->sl, MAX_SERVICE_LEVEL);
+ if (WARN_ON(hr_qp->sl > MAX_SERVICE_LEVEL))
return -EINVAL;
- }
roce_set_field(context->byte_28_at_fl, V2_QPC_BYTE_28_SL_M,
V2_QPC_BYTE_28_SL_S, hr_qp->sl);
--
2.8.1
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH v2 for-next 06/11] RDMA/hns: Remove unnecessary access right set during INIT2INIT
2020-12-04 10:40 [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11 Weihang Li
` (4 preceding siblings ...)
2020-12-04 10:40 ` [PATCH v2 for-next 05/11] RDMA/hns: WARN_ON if get a reserved sl from users Weihang Li
@ 2020-12-04 10:40 ` Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 07/11] RDMA/hns: Fix coding style issues Weihang Li
` (4 subsequent siblings)
10 siblings, 0 replies; 16+ messages in thread
From: Weihang Li @ 2020-12-04 10:40 UTC (permalink / raw)
To: dledford, jgg; +Cc: leon, linux-rdma, linuxarm
From: Yixian Liu <liuyixian@huawei.com>
As the qp access right is checked and setted in common function
hns_roce_v2_set_opt_fields(), there is no need to set again for a special
case INIT2INIT.
Fixes: 926a01dc000d ("RDMA/hns: Add QP operations support for hip08 SoC")
Fixes: 7db82697b8bf ("RDMA/hns: Add support for extended atomic in userspace")
Signed-off-by: Yixian Liu <liuyixian@huawei.com>
Signed-off-by: Weihang Li <liweihang@huawei.com>
---
drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 46 ------------------------------
1 file changed, 46 deletions(-)
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
index 15e1313..2d994d2 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
@@ -3975,7 +3975,6 @@ static void modify_qp_reset_to_init(struct ib_qp *ibqp,
roce_set_bit(context->byte_172_sq_psn, V2_QPC_BYTE_172_FRE_S, 1);
- hr_qp->access_flags = attr->qp_access_flags;
roce_set_field(context->byte_252_err_txcqn, V2_QPC_BYTE_252_TX_CQN_M,
V2_QPC_BYTE_252_TX_CQN_S, to_hr_cq(ibqp->send_cq)->cqn);
@@ -4004,51 +4003,6 @@ static void modify_qp_init_to_init(struct ib_qp *ibqp,
roce_set_field(qpc_mask->byte_4_sqpn_tst, V2_QPC_BYTE_4_TST_M,
V2_QPC_BYTE_4_TST_S, 0);
- if (attr_mask & IB_QP_ACCESS_FLAGS) {
- roce_set_bit(context->byte_76_srqn_op_en, V2_QPC_BYTE_76_RRE_S,
- !!(attr->qp_access_flags & IB_ACCESS_REMOTE_READ));
- roce_set_bit(qpc_mask->byte_76_srqn_op_en, V2_QPC_BYTE_76_RRE_S,
- 0);
-
- roce_set_bit(context->byte_76_srqn_op_en, V2_QPC_BYTE_76_RWE_S,
- !!(attr->qp_access_flags &
- IB_ACCESS_REMOTE_WRITE));
- roce_set_bit(qpc_mask->byte_76_srqn_op_en, V2_QPC_BYTE_76_RWE_S,
- 0);
-
- roce_set_bit(context->byte_76_srqn_op_en, V2_QPC_BYTE_76_ATE_S,
- !!(attr->qp_access_flags &
- IB_ACCESS_REMOTE_ATOMIC));
- roce_set_bit(qpc_mask->byte_76_srqn_op_en, V2_QPC_BYTE_76_ATE_S,
- 0);
- roce_set_bit(context->byte_76_srqn_op_en,
- V2_QPC_BYTE_76_EXT_ATE_S,
- !!(attr->qp_access_flags &
- IB_ACCESS_REMOTE_ATOMIC));
- roce_set_bit(qpc_mask->byte_76_srqn_op_en,
- V2_QPC_BYTE_76_EXT_ATE_S, 0);
- } else {
- roce_set_bit(context->byte_76_srqn_op_en, V2_QPC_BYTE_76_RRE_S,
- !!(hr_qp->access_flags & IB_ACCESS_REMOTE_READ));
- roce_set_bit(qpc_mask->byte_76_srqn_op_en, V2_QPC_BYTE_76_RRE_S,
- 0);
-
- roce_set_bit(context->byte_76_srqn_op_en, V2_QPC_BYTE_76_RWE_S,
- !!(hr_qp->access_flags & IB_ACCESS_REMOTE_WRITE));
- roce_set_bit(qpc_mask->byte_76_srqn_op_en, V2_QPC_BYTE_76_RWE_S,
- 0);
-
- roce_set_bit(context->byte_76_srqn_op_en, V2_QPC_BYTE_76_ATE_S,
- !!(hr_qp->access_flags & IB_ACCESS_REMOTE_ATOMIC));
- roce_set_bit(qpc_mask->byte_76_srqn_op_en, V2_QPC_BYTE_76_ATE_S,
- 0);
- roce_set_bit(context->byte_76_srqn_op_en,
- V2_QPC_BYTE_76_EXT_ATE_S,
- !!(hr_qp->access_flags & IB_ACCESS_REMOTE_ATOMIC));
- roce_set_bit(qpc_mask->byte_76_srqn_op_en,
- V2_QPC_BYTE_76_EXT_ATE_S, 0);
- }
-
roce_set_field(context->byte_16_buf_ba_pg_sz, V2_QPC_BYTE_16_PD_M,
V2_QPC_BYTE_16_PD_S, to_hr_pd(ibqp->pd)->pdn);
roce_set_field(qpc_mask->byte_16_buf_ba_pg_sz, V2_QPC_BYTE_16_PD_M,
--
2.8.1
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH v2 for-next 07/11] RDMA/hns: Fix coding style issues
2020-12-04 10:40 [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11 Weihang Li
` (5 preceding siblings ...)
2020-12-04 10:40 ` [PATCH v2 for-next 06/11] RDMA/hns: Remove unnecessary access right set during INIT2INIT Weihang Li
@ 2020-12-04 10:40 ` Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 08/11] RDMA/hns: Clear redundant variable initialization Weihang Li
` (3 subsequent siblings)
10 siblings, 0 replies; 16+ messages in thread
From: Weihang Li @ 2020-12-04 10:40 UTC (permalink / raw)
To: dledford, jgg; +Cc: leon, linux-rdma, linuxarm
From: Lang Cheng <chenglang@huawei.com>
Just format the code without modifying anything, including fixing some
redundant and missing blanks and spaces and changing the variable
definition order.
Signed-off-by: Lang Cheng <chenglang@huawei.com>
Signed-off-by: Weihang Li <liweihang@huawei.com>
---
drivers/infiniband/hw/hns/hns_roce_cmd.c | 27 +++++++++++++--------------
drivers/infiniband/hw/hns/hns_roce_cmd.h | 4 ++--
drivers/infiniband/hw/hns/hns_roce_cq.c | 2 +-
drivers/infiniband/hw/hns/hns_roce_hem.c | 20 ++++++++++----------
drivers/infiniband/hw/hns/hns_roce_hem.h | 2 +-
drivers/infiniband/hw/hns/hns_roce_hw_v1.c | 9 +++------
drivers/infiniband/hw/hns/hns_roce_hw_v1.h | 2 +-
drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 9 +++------
drivers/infiniband/hw/hns/hns_roce_hw_v2.h | 6 +++---
drivers/infiniband/hw/hns/hns_roce_main.c | 6 +++---
drivers/infiniband/hw/hns/hns_roce_mr.c | 4 ++--
drivers/infiniband/hw/hns/hns_roce_qp.c | 2 +-
drivers/infiniband/hw/hns/hns_roce_srq.c | 1 -
13 files changed, 43 insertions(+), 51 deletions(-)
diff --git a/drivers/infiniband/hw/hns/hns_roce_cmd.c b/drivers/infiniband/hw/hns/hns_roce_cmd.c
index 455d533..c493d76 100644
--- a/drivers/infiniband/hw/hns/hns_roce_cmd.c
+++ b/drivers/infiniband/hw/hns/hns_roce_cmd.c
@@ -36,9 +36,9 @@
#include "hns_roce_device.h"
#include "hns_roce_cmd.h"
-#define CMD_POLL_TOKEN 0xffff
-#define CMD_MAX_NUM 32
-#define CMD_TOKEN_MASK 0x1f
+#define CMD_POLL_TOKEN 0xffff
+#define CMD_MAX_NUM 32
+#define CMD_TOKEN_MASK 0x1f
static int hns_roce_cmd_mbox_post_hw(struct hns_roce_dev *hr_dev, u64 in_param,
u64 out_param, u32 in_modifier,
@@ -93,8 +93,8 @@ static int hns_roce_cmd_mbox_poll(struct hns_roce_dev *hr_dev, u64 in_param,
void hns_roce_cmd_event(struct hns_roce_dev *hr_dev, u16 token, u8 status,
u64 out_param)
{
- struct hns_roce_cmd_context
- *context = &hr_dev->cmd.context[token & hr_dev->cmd.token_mask];
+ struct hns_roce_cmd_context *context =
+ &hr_dev->cmd.context[token % hr_dev->cmd.max_cmds];
if (token != context->token)
return;
@@ -164,8 +164,8 @@ static int hns_roce_cmd_mbox_wait(struct hns_roce_dev *hr_dev, u64 in_param,
int ret;
down(&hr_dev->cmd.event_sem);
- ret = __hns_roce_cmd_mbox_wait(hr_dev, in_param, out_param,
- in_modifier, op_modifier, op, timeout);
+ ret = __hns_roce_cmd_mbox_wait(hr_dev, in_param, out_param, in_modifier,
+ op_modifier, op, timeout);
up(&hr_dev->cmd.event_sem);
return ret;
@@ -231,9 +231,8 @@ int hns_roce_cmd_use_events(struct hns_roce_dev *hr_dev)
struct hns_roce_cmdq *hr_cmd = &hr_dev->cmd;
int i;
- hr_cmd->context = kmalloc_array(hr_cmd->max_cmds,
- sizeof(*hr_cmd->context),
- GFP_KERNEL);
+ hr_cmd->context =
+ kcalloc(hr_cmd->max_cmds, sizeof(*hr_cmd->context), GFP_KERNEL);
if (!hr_cmd->context)
return -ENOMEM;
@@ -262,8 +261,8 @@ void hns_roce_cmd_use_polling(struct hns_roce_dev *hr_dev)
hr_cmd->use_events = 0;
}
-struct hns_roce_cmd_mailbox
- *hns_roce_alloc_cmd_mailbox(struct hns_roce_dev *hr_dev)
+struct hns_roce_cmd_mailbox *
+hns_roce_alloc_cmd_mailbox(struct hns_roce_dev *hr_dev)
{
struct hns_roce_cmd_mailbox *mailbox;
@@ -271,8 +270,8 @@ struct hns_roce_cmd_mailbox
if (!mailbox)
return ERR_PTR(-ENOMEM);
- mailbox->buf = dma_pool_alloc(hr_dev->cmd.pool, GFP_KERNEL,
- &mailbox->dma);
+ mailbox->buf =
+ dma_pool_alloc(hr_dev->cmd.pool, GFP_KERNEL, &mailbox->dma);
if (!mailbox->buf) {
kfree(mailbox);
return ERR_PTR(-ENOMEM);
diff --git a/drivers/infiniband/hw/hns/hns_roce_cmd.h b/drivers/infiniband/hw/hns/hns_roce_cmd.h
index 1915bac..8e63b82 100644
--- a/drivers/infiniband/hw/hns/hns_roce_cmd.h
+++ b/drivers/infiniband/hw/hns/hns_roce_cmd.h
@@ -143,8 +143,8 @@ int hns_roce_cmd_mbox(struct hns_roce_dev *hr_dev, u64 in_param, u64 out_param,
unsigned long in_modifier, u8 op_modifier, u16 op,
unsigned long timeout);
-struct hns_roce_cmd_mailbox
- *hns_roce_alloc_cmd_mailbox(struct hns_roce_dev *hr_dev);
+struct hns_roce_cmd_mailbox *
+hns_roce_alloc_cmd_mailbox(struct hns_roce_dev *hr_dev);
void hns_roce_free_cmd_mailbox(struct hns_roce_dev *hr_dev,
struct hns_roce_cmd_mailbox *mailbox);
diff --git a/drivers/infiniband/hw/hns/hns_roce_cq.c b/drivers/infiniband/hw/hns/hns_roce_cq.c
index b94a560..9b6c493 100644
--- a/drivers/infiniband/hw/hns/hns_roce_cq.c
+++ b/drivers/infiniband/hw/hns/hns_roce_cq.c
@@ -41,9 +41,9 @@
static int alloc_cqc(struct hns_roce_dev *hr_dev, struct hns_roce_cq *hr_cq)
{
+ struct ib_device *ibdev = &hr_dev->ib_dev;
struct hns_roce_cmd_mailbox *mailbox;
struct hns_roce_cq_table *cq_table;
- struct ib_device *ibdev = &hr_dev->ib_dev;
u64 mtts[MTT_MIN_COUNT] = { 0 };
dma_addr_t dma_handle;
int ret;
diff --git a/drivers/infiniband/hw/hns/hns_roce_hem.c b/drivers/infiniband/hw/hns/hns_roce_hem.c
index 359e5dd..303c8dd 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hem.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hem.c
@@ -209,9 +209,9 @@ int hns_roce_calc_hem_mhop(struct hns_roce_dev *hr_dev,
{
struct device *dev = hr_dev->dev;
u32 chunk_ba_num;
+ u32 chunk_size;
u32 table_idx;
u32 bt_num;
- u32 chunk_size;
if (get_hem_table_config(hr_dev, mhop, table->type))
return -EINVAL;
@@ -343,15 +343,15 @@ static int hns_roce_set_hem(struct hns_roce_dev *hr_dev,
{
spinlock_t *lock = &hr_dev->bt_cmd_lock;
struct device *dev = hr_dev->dev;
- long end;
- unsigned long flags;
struct hns_roce_hem_iter iter;
void __iomem *bt_cmd;
__le32 bt_cmd_val[2];
__le32 bt_cmd_h = 0;
+ unsigned long flags;
__le32 bt_cmd_l;
- u64 bt_ba;
int ret = 0;
+ u64 bt_ba;
+ long end;
/* Find the HEM(Hardware Entry Memory) entry */
unsigned long i = (obj & (table->num_obj - 1)) /
@@ -651,8 +651,8 @@ int hns_roce_table_get(struct hns_roce_dev *hr_dev,
struct hns_roce_hem_table *table, unsigned long obj)
{
struct device *dev = hr_dev->dev;
- int ret = 0;
unsigned long i;
+ int ret = 0;
if (hns_roce_check_whether_mhop(hr_dev, table->type))
return hns_roce_table_mhop_get(hr_dev, table, obj);
@@ -800,14 +800,14 @@ void *hns_roce_table_find(struct hns_roce_dev *hr_dev,
struct hns_roce_hem_chunk *chunk;
struct hns_roce_hem_mhop mhop;
struct hns_roce_hem *hem;
- void *addr = NULL;
unsigned long mhop_obj = obj;
unsigned long obj_per_chunk;
unsigned long idx_offset;
int offset, dma_offset;
+ void *addr = NULL;
+ u32 hem_idx = 0;
int length;
int i, j;
- u32 hem_idx = 0;
if (!table->lowmem)
return NULL;
@@ -977,8 +977,8 @@ static void hns_roce_cleanup_mhop_hem_table(struct hns_roce_dev *hr_dev,
{
struct hns_roce_hem_mhop mhop;
u32 buf_chunk_size;
- int i;
u64 obj;
+ int i;
if (hns_roce_calc_hem_mhop(hr_dev, table, NULL, &mhop))
return;
@@ -1313,8 +1313,8 @@ static int hem_list_alloc_root_bt(struct hns_roce_dev *hr_dev,
const struct hns_roce_buf_region *regions,
int region_cnt)
{
- struct roce_hem_item *hem, *temp_hem, *root_hem;
struct list_head temp_list[HNS_ROCE_MAX_BT_REGION];
+ struct roce_hem_item *hem, *temp_hem, *root_hem;
const struct hns_roce_buf_region *r;
struct list_head temp_root;
struct list_head temp_btm;
@@ -1419,8 +1419,8 @@ int hns_roce_hem_list_request(struct hns_roce_dev *hr_dev,
{
const struct hns_roce_buf_region *r;
int ofs, end;
- int ret;
int unit;
+ int ret;
int i;
if (region_cnt > HNS_ROCE_MAX_BT_REGION) {
diff --git a/drivers/infiniband/hw/hns/hns_roce_hem.h b/drivers/infiniband/hw/hns/hns_roce_hem.h
index c6bd982..13fdeb3 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hem.h
+++ b/drivers/infiniband/hw/hns/hns_roce_hem.h
@@ -175,4 +175,4 @@ static inline dma_addr_t hns_roce_hem_addr(struct hns_roce_hem_iter *iter)
return sg_dma_address(&iter->chunk->mem[iter->page_idx]);
}
-#endif /*_HNS_ROCE_HEM_H*/
+#endif /* _HNS_ROCE_HEM_H */
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v1.c b/drivers/infiniband/hw/hns/hns_roce_hw_v1.c
index f18380f..eb0fd72 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v1.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v1.c
@@ -239,7 +239,7 @@ static int hns_roce_v1_post_send(struct ib_qp *ibqp,
break;
}
- /*Ctrl field, ctrl set type: sig, solic, imm, fence */
+ /* Ctrl field, ctrl set type: sig, solic, imm, fence */
/* SO wait for conforming application scenarios */
ctrl->flag |= (wr->send_flags & IB_SEND_SIGNALED ?
cpu_to_le32(HNS_ROCE_WQE_CQ_NOTIFY) : 0) |
@@ -300,7 +300,7 @@ static int hns_roce_v1_post_send(struct ib_qp *ibqp,
}
ctrl->flag |= cpu_to_le32(HNS_ROCE_WQE_INLINE);
} else {
- /*sqe num is two */
+ /* sqe num is two */
for (i = 0; i < wr->num_sge; i++)
set_data_seg(dseg + i, wr->sg_list + i);
@@ -1165,7 +1165,7 @@ static int hns_roce_raq_init(struct hns_roce_dev *hr_dev)
}
raq->e_raq_buf->map = addr;
- /* Configure raq extended address. 48bit 4K align*/
+ /* Configure raq extended address. 48bit 4K align */
roce_write(hr_dev, ROCEE_EXT_RAQ_REG, raq->e_raq_buf->map >> 12);
/* Configure raq_shift */
@@ -2760,7 +2760,6 @@ static int hns_roce_v1_m_qp(struct ib_qp *ibqp, const struct ib_qp_attr *attr,
roce_set_field(context->qpc_bytes_16,
QP_CONTEXT_QPC_BYTES_16_QP_NUM_M,
QP_CONTEXT_QPC_BYTES_16_QP_NUM_S, hr_qp->qpn);
-
} else if (cur_state == IB_QPS_INIT && new_state == IB_QPS_INIT) {
roce_set_field(context->qpc_bytes_4,
QP_CONTEXT_QPC_BYTES_4_TRANSPORT_SERVICE_TYPE_M,
@@ -3795,7 +3794,6 @@ static int hns_roce_v1_aeq_int(struct hns_roce_dev *hr_dev,
int event_type;
while ((aeqe = next_aeqe_sw_v1(eq))) {
-
/* Make sure we read the AEQ entry after we have checked the
* ownership bit
*/
@@ -3900,7 +3898,6 @@ static int hns_roce_v1_ceq_int(struct hns_roce_dev *hr_dev,
u32 cqn;
while ((ceqe = next_ceqe_sw_v1(eq))) {
-
/* Make sure we read CEQ entry after we have checked the
* ownership bit
*/
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v1.h b/drivers/infiniband/hw/hns/hns_roce_hw_v1.h
index ffd0156..46ab0a3 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v1.h
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v1.h
@@ -419,7 +419,7 @@ struct hns_roce_wqe_data_seg {
struct hns_roce_wqe_raddr_seg {
__le32 rkey;
- __le32 len;/* reserved */
+ __le32 len; /* reserved */
__le64 raddr;
};
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
index 2d994d2..ce2eeb2 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
@@ -1025,8 +1025,8 @@ static int hns_roce_v2_rst_process_cmd(struct hns_roce_dev *hr_dev)
struct hns_roce_v2_priv *priv = hr_dev->priv;
struct hnae3_handle *handle = priv->handle;
const struct hnae3_ae_ops *ops = handle->ae_algo->ops;
- unsigned long instance_stage; /* the current instance stage */
- unsigned long reset_stage; /* the current reset stage */
+ unsigned long instance_stage; /* the current instance stage */
+ unsigned long reset_stage; /* the current reset stage */
unsigned long reset_cnt;
bool sw_resetting;
bool hw_resetting;
@@ -2451,7 +2451,6 @@ static int hns_roce_init_link_table(struct hns_roce_dev *hr_dev,
if (i < (pg_num - 1))
entry[i].blk_ba1_nxt_ptr |=
(i + 1) << HNS_ROCE_LINK_TABLE_NXT_PTR_S;
-
}
link_tbl->npages = pg_num;
link_tbl->pg_sz = buf_chk_sz;
@@ -5615,16 +5614,14 @@ static int hns_roce_v2_aeq_int(struct hns_roce_dev *hr_dev,
case HNS_ROCE_EVENT_TYPE_CQ_OVERFLOW:
hns_roce_cq_event(hr_dev, cqn, event_type);
break;
- case HNS_ROCE_EVENT_TYPE_DB_OVERFLOW:
- break;
case HNS_ROCE_EVENT_TYPE_MB:
hns_roce_cmd_event(hr_dev,
le16_to_cpu(aeqe->event.cmd.token),
aeqe->event.cmd.status,
le64_to_cpu(aeqe->event.cmd.out_param));
break;
+ case HNS_ROCE_EVENT_TYPE_DB_OVERFLOW:
case HNS_ROCE_EVENT_TYPE_CEQ_OVERFLOW:
- break;
case HNS_ROCE_EVENT_TYPE_FLR:
break;
default:
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.h b/drivers/infiniband/hw/hns/hns_roce_hw_v2.h
index fac8536..bdaccf8 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.h
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.h
@@ -449,7 +449,7 @@ struct hns_roce_srq_context {
#define SRQC_BYTE_60_SRQ_DB_RECORD_ADDR_S 1
#define SRQC_BYTE_60_SRQ_DB_RECORD_ADDR_M GENMASK(31, 1)
-enum{
+enum {
V2_MPT_ST_VALID = 0x1,
V2_MPT_ST_FREE = 0x2,
};
@@ -1094,9 +1094,9 @@ struct hns_roce_v2_ud_send_wqe {
u8 sgid_index;
u8 smac_index;
u8 dgid[GID_LEN_V2];
-
};
-#define V2_UD_SEND_WQE_BYTE_4_OPCODE_S 0
+
+#define V2_UD_SEND_WQE_BYTE_4_OPCODE_S 0
#define V2_UD_SEND_WQE_BYTE_4_OPCODE_M GENMASK(4, 0)
#define V2_UD_SEND_WQE_BYTE_4_OWNER_S 7
diff --git a/drivers/infiniband/hw/hns/hns_roce_main.c b/drivers/infiniband/hw/hns/hns_roce_main.c
index 9ca4892..b6f27f9 100644
--- a/drivers/infiniband/hw/hns/hns_roce_main.c
+++ b/drivers/infiniband/hw/hns/hns_roce_main.c
@@ -556,8 +556,8 @@ static int hns_roce_register_device(struct hns_roce_dev *hr_dev)
static int hns_roce_init_hem(struct hns_roce_dev *hr_dev)
{
- int ret;
struct device *dev = hr_dev->dev;
+ int ret;
ret = hns_roce_init_hem_table(hr_dev, &hr_dev->mr_table.mtpt_table,
HEM_TYPE_MTPT, hr_dev->caps.mtpt_entry_sz,
@@ -714,8 +714,8 @@ static int hns_roce_init_hem(struct hns_roce_dev *hr_dev)
*/
static int hns_roce_setup_hca(struct hns_roce_dev *hr_dev)
{
- int ret;
struct device *dev = hr_dev->dev;
+ int ret;
spin_lock_init(&hr_dev->sm_lock);
spin_lock_init(&hr_dev->bt_cmd_lock);
@@ -839,8 +839,8 @@ void hns_roce_handle_device_err(struct hns_roce_dev *hr_dev)
int hns_roce_init(struct hns_roce_dev *hr_dev)
{
- int ret;
struct device *dev = hr_dev->dev;
+ int ret;
if (hr_dev->hw->reset) {
ret = hr_dev->hw->reset(hr_dev, true);
diff --git a/drivers/infiniband/hw/hns/hns_roce_mr.c b/drivers/infiniband/hw/hns/hns_roce_mr.c
index 87e2e62..5c43089 100644
--- a/drivers/infiniband/hw/hns/hns_roce_mr.c
+++ b/drivers/infiniband/hw/hns/hns_roce_mr.c
@@ -167,10 +167,10 @@ static void hns_roce_mr_free(struct hns_roce_dev *hr_dev,
static int hns_roce_mr_enable(struct hns_roce_dev *hr_dev,
struct hns_roce_mr *mr)
{
- int ret;
unsigned long mtpt_idx = key_to_hw_index(mr->key);
- struct device *dev = hr_dev->dev;
struct hns_roce_cmd_mailbox *mailbox;
+ struct device *dev = hr_dev->dev;
+ int ret;
/* Allocate mailbox memory */
mailbox = hns_roce_alloc_cmd_mailbox(hr_dev);
diff --git a/drivers/infiniband/hw/hns/hns_roce_qp.c b/drivers/infiniband/hw/hns/hns_roce_qp.c
index 0e784c5..b59f3e0 100644
--- a/drivers/infiniband/hw/hns/hns_roce_qp.c
+++ b/drivers/infiniband/hw/hns/hns_roce_qp.c
@@ -114,8 +114,8 @@ void hns_roce_qp_event(struct hns_roce_dev *hr_dev, u32 qpn, int event_type)
static void hns_roce_ib_qp_event(struct hns_roce_qp *hr_qp,
enum hns_roce_event type)
{
- struct ib_event event;
struct ib_qp *ibqp = &hr_qp->ibqp;
+ struct ib_event event;
if (ibqp->event_handler) {
event.device = ibqp->device;
diff --git a/drivers/infiniband/hw/hns/hns_roce_srq.c b/drivers/infiniband/hw/hns/hns_roce_srq.c
index 6ce250e..bf0b51e 100644
--- a/drivers/infiniband/hw/hns/hns_roce_srq.c
+++ b/drivers/infiniband/hw/hns/hns_roce_srq.c
@@ -240,7 +240,6 @@ static int alloc_srq_idx(struct hns_roce_dev *hr_dev, struct hns_roce_srq *srq,
err = -ENOMEM;
goto err_idx_mtr;
}
-
}
return 0;
--
2.8.1
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH v2 for-next 08/11] RDMA/hns: Clear redundant variable initialization
2020-12-04 10:40 [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11 Weihang Li
` (6 preceding siblings ...)
2020-12-04 10:40 ` [PATCH v2 for-next 07/11] RDMA/hns: Fix coding style issues Weihang Li
@ 2020-12-04 10:40 ` Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 09/11] RDMA/hns: Fix incorrect symbol types Weihang Li
` (2 subsequent siblings)
10 siblings, 0 replies; 16+ messages in thread
From: Weihang Li @ 2020-12-04 10:40 UTC (permalink / raw)
To: dledford, jgg; +Cc: leon, linux-rdma, linuxarm
From: Xinhao Liu <liuxinhao5@hisilicon.com>
There is no need to initialize some variable because they will be assigned
with a value later.
Signed-off-by: Xinhao Liu <liuxinhao5@hisilicon.com>
Signed-off-by: Weihang Li <liweihang@huawei.com>
---
drivers/infiniband/hw/hns/hns_roce_hem.c | 2 +-
drivers/infiniband/hw/hns/hns_roce_hw_v1.c | 8 ++++----
2 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/infiniband/hw/hns/hns_roce_hem.c b/drivers/infiniband/hw/hns/hns_roce_hem.c
index 303c8dd..f19bbcc 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hem.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hem.c
@@ -887,7 +887,7 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev,
unsigned long buf_chunk_size;
unsigned long bt_chunk_size;
unsigned long bt_chunk_num;
- unsigned long num_bt_l0 = 0;
+ unsigned long num_bt_l0;
u32 hop_num;
if (get_hem_table_config(hr_dev, &mhop, type))
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v1.c b/drivers/infiniband/hw/hns/hns_roce_hw_v1.c
index eb0fd72..0f4273d 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v1.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v1.c
@@ -353,8 +353,8 @@ static int hns_roce_v1_post_recv(struct ib_qp *ibqp,
unsigned long flags = 0;
unsigned int wqe_idx;
int ret = 0;
- int nreq = 0;
- int i = 0;
+ int nreq;
+ int i;
u32 reg_val;
spin_lock_irqsave(&hr_qp->rq.lock, flags);
@@ -2300,7 +2300,7 @@ int hns_roce_v1_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *wc)
struct hns_roce_qp *cur_qp = NULL;
unsigned long flags;
int npolled;
- int ret = 0;
+ int ret;
spin_lock_irqsave(&hr_cq->lock, flags);
@@ -4123,7 +4123,7 @@ static int hns_roce_v1_create_eq(struct hns_roce_dev *hr_dev,
void __iomem *eqc = hr_dev->eq_table.eqc_base[eq->eqn];
struct device *dev = &hr_dev->pdev->dev;
dma_addr_t tmp_dma_addr;
- u32 eqcuridx_val = 0;
+ u32 eqcuridx_val;
u32 eqconsindx_val;
u32 eqshift_val;
__le32 tmp2 = 0;
--
2.8.1
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH v2 for-next 09/11] RDMA/hns: Fix incorrect symbol types
2020-12-04 10:40 [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11 Weihang Li
` (7 preceding siblings ...)
2020-12-04 10:40 ` [PATCH v2 for-next 08/11] RDMA/hns: Clear redundant variable initialization Weihang Li
@ 2020-12-04 10:40 ` Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 10/11] RDMA/hns: Fix inaccurate prints Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 11/11] RDMA/hns: Simplify AEQE process for different types of queue Weihang Li
10 siblings, 0 replies; 16+ messages in thread
From: Weihang Li @ 2020-12-04 10:40 UTC (permalink / raw)
To: dledford, jgg; +Cc: leon, linux-rdma, linuxarm
From: Wenpeng Liang <liangwenpeng@huawei.com>
Types of some fields, variables and parameters of some functions should be
unsigned.
Signed-off-by: Wenpeng Liang <liangwenpeng@huawei.com>
Signed-off-by: Weihang Li <liweihang@huawei.com>
---
drivers/infiniband/hw/hns/hns_roce_cmd.c | 10 +++---
drivers/infiniband/hw/hns/hns_roce_cmd.h | 2 +-
drivers/infiniband/hw/hns/hns_roce_common.h | 14 ++++----
drivers/infiniband/hw/hns/hns_roce_db.c | 8 ++---
drivers/infiniband/hw/hns/hns_roce_device.h | 53 +++++++++++++++--------------
drivers/infiniband/hw/hns/hns_roce_hw_v1.c | 8 ++---
drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 30 ++++++++--------
drivers/infiniband/hw/hns/hns_roce_main.c | 2 +-
drivers/infiniband/hw/hns/hns_roce_mr.c | 11 +++---
drivers/infiniband/hw/hns/hns_roce_qp.c | 8 ++---
10 files changed, 73 insertions(+), 73 deletions(-)
diff --git a/drivers/infiniband/hw/hns/hns_roce_cmd.c b/drivers/infiniband/hw/hns/hns_roce_cmd.c
index c493d76..339e3fd 100644
--- a/drivers/infiniband/hw/hns/hns_roce_cmd.c
+++ b/drivers/infiniband/hw/hns/hns_roce_cmd.c
@@ -60,7 +60,7 @@ static int hns_roce_cmd_mbox_post_hw(struct hns_roce_dev *hr_dev, u64 in_param,
static int __hns_roce_cmd_mbox_poll(struct hns_roce_dev *hr_dev, u64 in_param,
u64 out_param, unsigned long in_modifier,
u8 op_modifier, u16 op,
- unsigned long timeout)
+ unsigned int timeout)
{
struct device *dev = hr_dev->dev;
int ret;
@@ -78,7 +78,7 @@ static int __hns_roce_cmd_mbox_poll(struct hns_roce_dev *hr_dev, u64 in_param,
static int hns_roce_cmd_mbox_poll(struct hns_roce_dev *hr_dev, u64 in_param,
u64 out_param, unsigned long in_modifier,
- u8 op_modifier, u16 op, unsigned long timeout)
+ u8 op_modifier, u16 op, unsigned int timeout)
{
int ret;
@@ -108,7 +108,7 @@ void hns_roce_cmd_event(struct hns_roce_dev *hr_dev, u16 token, u8 status,
static int __hns_roce_cmd_mbox_wait(struct hns_roce_dev *hr_dev, u64 in_param,
u64 out_param, unsigned long in_modifier,
u8 op_modifier, u16 op,
- unsigned long timeout)
+ unsigned int timeout)
{
struct hns_roce_cmdq *cmd = &hr_dev->cmd;
struct hns_roce_cmd_context *context;
@@ -159,7 +159,7 @@ static int __hns_roce_cmd_mbox_wait(struct hns_roce_dev *hr_dev, u64 in_param,
static int hns_roce_cmd_mbox_wait(struct hns_roce_dev *hr_dev, u64 in_param,
u64 out_param, unsigned long in_modifier,
- u8 op_modifier, u16 op, unsigned long timeout)
+ u8 op_modifier, u16 op, unsigned int timeout)
{
int ret;
@@ -173,7 +173,7 @@ static int hns_roce_cmd_mbox_wait(struct hns_roce_dev *hr_dev, u64 in_param,
int hns_roce_cmd_mbox(struct hns_roce_dev *hr_dev, u64 in_param, u64 out_param,
unsigned long in_modifier, u8 op_modifier, u16 op,
- unsigned long timeout)
+ unsigned int timeout)
{
int ret;
diff --git a/drivers/infiniband/hw/hns/hns_roce_cmd.h b/drivers/infiniband/hw/hns/hns_roce_cmd.h
index 8e63b82..8025e7f 100644
--- a/drivers/infiniband/hw/hns/hns_roce_cmd.h
+++ b/drivers/infiniband/hw/hns/hns_roce_cmd.h
@@ -141,7 +141,7 @@ enum {
int hns_roce_cmd_mbox(struct hns_roce_dev *hr_dev, u64 in_param, u64 out_param,
unsigned long in_modifier, u8 op_modifier, u16 op,
- unsigned long timeout);
+ unsigned int timeout);
struct hns_roce_cmd_mailbox *
hns_roce_alloc_cmd_mailbox(struct hns_roce_dev *hr_dev);
diff --git a/drivers/infiniband/hw/hns/hns_roce_common.h b/drivers/infiniband/hw/hns/hns_roce_common.h
index 29469e1..5afee04 100644
--- a/drivers/infiniband/hw/hns/hns_roce_common.h
+++ b/drivers/infiniband/hw/hns/hns_roce_common.h
@@ -38,19 +38,19 @@
#define roce_raw_write(value, addr) \
__raw_writel((__force u32)cpu_to_le32(value), (addr))
-#define roce_get_field(origin, mask, shift) \
- (((le32_to_cpu(origin)) & (mask)) >> (shift))
+#define roce_get_field(origin, mask, shift) \
+ ((le32_to_cpu(origin) & (mask)) >> (u32)(shift))
#define roce_get_bit(origin, shift) \
roce_get_field((origin), (1ul << (shift)), (shift))
-#define roce_set_field(origin, mask, shift, val) \
- do { \
- (origin) &= ~cpu_to_le32(mask); \
- (origin) |= cpu_to_le32(((u32)(val) << (shift)) & (mask)); \
+#define roce_set_field(origin, mask, shift, val) \
+ do { \
+ (origin) &= ~cpu_to_le32(mask); \
+ (origin) |= cpu_to_le32(((u32)(val) << (u32)(shift)) & (mask)); \
} while (0)
-#define roce_set_bit(origin, shift, val) \
+#define roce_set_bit(origin, shift, val) \
roce_set_field((origin), (1ul << (shift)), (shift), (val))
#define FIELD_LOC(field_type, field_h, field_l) field_type, field_h, field_l
diff --git a/drivers/infiniband/hw/hns/hns_roce_db.c b/drivers/infiniband/hw/hns/hns_roce_db.c
index bff6abd..5cb7376c 100644
--- a/drivers/infiniband/hw/hns/hns_roce_db.c
+++ b/drivers/infiniband/hw/hns/hns_roce_db.c
@@ -95,8 +95,8 @@ static struct hns_roce_db_pgdir *hns_roce_alloc_db_pgdir(
static int hns_roce_alloc_db_from_pgdir(struct hns_roce_db_pgdir *pgdir,
struct hns_roce_db *db, int order)
{
- int o;
- int i;
+ unsigned long o;
+ unsigned long i;
for (o = order; o <= 1; ++o) {
i = find_first_bit(pgdir->bits[o], HNS_ROCE_DB_PER_PAGE >> o);
@@ -154,8 +154,8 @@ int hns_roce_alloc_db(struct hns_roce_dev *hr_dev, struct hns_roce_db *db,
void hns_roce_free_db(struct hns_roce_dev *hr_dev, struct hns_roce_db *db)
{
- int o;
- int i;
+ unsigned long o;
+ unsigned long i;
mutex_lock(&hr_dev->pgdir_mutex);
diff --git a/drivers/infiniband/hw/hns/hns_roce_device.h b/drivers/infiniband/hw/hns/hns_roce_device.h
index c7044f7..af73086 100644
--- a/drivers/infiniband/hw/hns/hns_roce_device.h
+++ b/drivers/infiniband/hw/hns/hns_roce_device.h
@@ -320,7 +320,7 @@ struct hns_roce_hem_table {
};
struct hns_roce_buf_region {
- int offset; /* page offset */
+ u32 offset; /* page offset */
u32 count; /* page count */
int hopnum; /* addressing hop num */
};
@@ -340,10 +340,10 @@ struct hns_roce_buf_attr {
size_t size; /* region size */
int hopnum; /* multi-hop addressing hop num */
} region[HNS_ROCE_MAX_BT_REGION];
- int region_count; /* valid region count */
+ unsigned int region_count; /* valid region count */
unsigned int page_shift; /* buffer page shift */
bool fixed_page; /* decide page shift is fixed-size or maximum size */
- int user_access; /* umem access flag */
+ unsigned int user_access; /* umem access flag */
bool mtt_only; /* only alloc buffer-required MTT memory */
};
@@ -354,7 +354,7 @@ struct hns_roce_hem_cfg {
unsigned int buf_pg_shift; /* buffer page shift */
unsigned int buf_pg_count; /* buffer page count */
struct hns_roce_buf_region region[HNS_ROCE_MAX_BT_REGION];
- int region_count;
+ unsigned int region_count;
};
/* memory translate region */
@@ -402,7 +402,7 @@ struct hns_roce_wq {
u64 *wrid; /* Work request ID */
spinlock_t lock;
u32 wqe_cnt; /* WQE num */
- int max_gs;
+ u32 max_gs;
int offset;
int wqe_shift; /* WQE size */
u32 head;
@@ -468,8 +468,8 @@ struct hns_roce_db {
} u;
dma_addr_t dma;
void *virt_addr;
- int index;
- int order;
+ unsigned long index;
+ unsigned long order;
};
struct hns_roce_cq {
@@ -517,8 +517,8 @@ struct hns_roce_srq {
u64 *wrid;
struct hns_roce_idx_que idx_que;
spinlock_t lock;
- int head;
- int tail;
+ u16 head;
+ u16 tail;
struct mutex mutex;
void (*event)(struct hns_roce_srq *srq, enum hns_roce_event event);
};
@@ -756,11 +756,11 @@ struct hns_roce_eq {
int type_flag; /* Aeq:1 ceq:0 */
int eqn;
u32 entries;
- int log_entries;
+ u32 log_entries;
int eqe_size;
int irq;
int log_page_size;
- int cons_index;
+ u32 cons_index;
struct hns_roce_buf_list *buf_list;
int over_ignore;
int coalesce;
@@ -768,7 +768,7 @@ struct hns_roce_eq {
int hop_num;
struct hns_roce_mtr mtr;
u16 eq_max_cnt;
- int eq_period;
+ u32 eq_period;
int shift;
int event_type;
int sub_type;
@@ -791,8 +791,8 @@ struct hns_roce_caps {
u32 max_sq_inline;
u32 max_rq_sg;
u32 max_extend_sg;
- int num_qps;
- u32 reserved_qps;
+ u32 num_qps;
+ u32 reserved_qps;
int num_qpc_timer;
int num_cqc_timer;
int num_srqs;
@@ -804,7 +804,7 @@ struct hns_roce_caps {
u32 max_srq_desc_sz;
int max_qp_init_rdma;
int max_qp_dest_rdma;
- int num_cqs;
+ u32 num_cqs;
u32 max_cqes;
u32 min_cqes;
u32 min_wqes;
@@ -813,7 +813,7 @@ struct hns_roce_caps {
int num_aeq_vectors;
int num_comp_vectors;
int num_other_vectors;
- int num_mtpts;
+ u32 num_mtpts;
u32 num_mtt_segs;
u32 num_cqe_segs;
u32 num_srqwqe_segs;
@@ -930,7 +930,7 @@ struct hns_roce_hw {
int (*post_mbox)(struct hns_roce_dev *hr_dev, u64 in_param,
u64 out_param, u32 in_modifier, u8 op_modifier, u16 op,
u16 token, int event);
- int (*chk_mbox)(struct hns_roce_dev *hr_dev, unsigned long timeout);
+ int (*chk_mbox)(struct hns_roce_dev *hr_dev, unsigned int timeout);
int (*rst_prc_mbox)(struct hns_roce_dev *hr_dev);
int (*set_gid)(struct hns_roce_dev *hr_dev, u8 port, int gid_index,
const union ib_gid *gid, const struct ib_gid_attr *attr);
@@ -1105,15 +1105,16 @@ static inline struct hns_roce_qp
return xa_load(&hr_dev->qp_table_xa, qpn & (hr_dev->caps.num_qps - 1));
}
-static inline void *hns_roce_buf_offset(struct hns_roce_buf *buf, int offset)
+static inline void *hns_roce_buf_offset(struct hns_roce_buf *buf,
+ unsigned int offset)
{
return (char *)(buf->trunk_list[offset >> buf->trunk_shift].buf) +
(offset & ((1 << buf->trunk_shift) - 1));
}
-static inline dma_addr_t hns_roce_buf_page(struct hns_roce_buf *buf, int idx)
+static inline dma_addr_t hns_roce_buf_page(struct hns_roce_buf *buf, u32 idx)
{
- int offset = idx << buf->page_shift;
+ unsigned int offset = idx << buf->page_shift;
return buf->trunk_list[offset >> buf->trunk_shift].map +
(offset & ((1 << buf->trunk_shift) - 1));
@@ -1188,7 +1189,7 @@ int hns_roce_mtr_create(struct hns_roce_dev *hr_dev, struct hns_roce_mtr *mtr,
void hns_roce_mtr_destroy(struct hns_roce_dev *hr_dev,
struct hns_roce_mtr *mtr);
int hns_roce_mtr_map(struct hns_roce_dev *hr_dev, struct hns_roce_mtr *mtr,
- dma_addr_t *pages, int page_cnt);
+ dma_addr_t *pages, unsigned int page_cnt);
int hns_roce_init_pd_table(struct hns_roce_dev *hr_dev);
int hns_roce_init_mr_table(struct hns_roce_dev *hr_dev);
@@ -1271,10 +1272,10 @@ struct ib_qp *hns_roce_create_qp(struct ib_pd *ib_pd,
int hns_roce_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
int attr_mask, struct ib_udata *udata);
void init_flush_work(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp);
-void *hns_roce_get_recv_wqe(struct hns_roce_qp *hr_qp, int n);
-void *hns_roce_get_send_wqe(struct hns_roce_qp *hr_qp, int n);
-void *hns_roce_get_extend_sge(struct hns_roce_qp *hr_qp, int n);
-bool hns_roce_wq_overflow(struct hns_roce_wq *hr_wq, int nreq,
+void *hns_roce_get_recv_wqe(struct hns_roce_qp *hr_qp, unsigned int n);
+void *hns_roce_get_send_wqe(struct hns_roce_qp *hr_qp, unsigned int n);
+void *hns_roce_get_extend_sge(struct hns_roce_qp *hr_qp, unsigned int n);
+bool hns_roce_wq_overflow(struct hns_roce_wq *hr_wq, u32 nreq,
struct ib_cq *ib_cq);
enum hns_roce_qp_state to_hns_roce_state(enum ib_qp_state state);
void hns_roce_lock_cqs(struct hns_roce_cq *send_cq,
@@ -1304,7 +1305,7 @@ void hns_roce_cq_completion(struct hns_roce_dev *hr_dev, u32 cqn);
void hns_roce_cq_event(struct hns_roce_dev *hr_dev, u32 cqn, int event_type);
void hns_roce_qp_event(struct hns_roce_dev *hr_dev, u32 qpn, int event_type);
void hns_roce_srq_event(struct hns_roce_dev *hr_dev, u32 srqn, int event_type);
-int hns_get_gid_index(struct hns_roce_dev *hr_dev, u8 port, int gid_index);
+u8 hns_get_gid_index(struct hns_roce_dev *hr_dev, u8 port, int gid_index);
void hns_roce_handle_device_err(struct hns_roce_dev *hr_dev);
int hns_roce_init(struct hns_roce_dev *hr_dev);
void hns_roce_exit(struct hns_roce_dev *hr_dev);
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v1.c b/drivers/infiniband/hw/hns/hns_roce_hw_v1.c
index 0f4273d..b7dd867 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v1.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v1.c
@@ -288,7 +288,7 @@ static int hns_roce_v1_post_send(struct ib_qp *ibqp,
ret = -EINVAL;
*bad_wr = wr;
dev_err(dev, "inline len(1-%d)=%d, illegal",
- ctrl->msg_length,
+ le32_to_cpu(ctrl->msg_length),
hr_dev->caps.max_sq_inline);
goto out;
}
@@ -1639,7 +1639,7 @@ static int hns_roce_v1_post_mbox(struct hns_roce_dev *hr_dev, u64 in_param,
}
static int hns_roce_v1_chk_mbox(struct hns_roce_dev *hr_dev,
- unsigned long timeout)
+ unsigned int timeout)
{
u8 __iomem *hcr = hr_dev->reg_base + ROCEE_MB1_REG;
unsigned long end;
@@ -3600,10 +3600,10 @@ static int hns_roce_v1_destroy_cq(struct ib_cq *ibcq, struct ib_udata *udata)
return 0;
}
-static void set_eq_cons_index_v1(struct hns_roce_eq *eq, int req_not)
+static void set_eq_cons_index_v1(struct hns_roce_eq *eq, u32 req_not)
{
roce_raw_write((eq->cons_index & HNS_ROCE_V1_CONS_IDX_M) |
- (req_not << eq->log_entries), eq->doorbell);
+ (req_not << eq->log_entries), eq->db_reg);
}
static void hns_roce_v1_wq_catas_err_handle(struct hns_roce_dev *hr_dev,
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
index ce2eeb2..6c006af 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
@@ -656,7 +656,7 @@ static int hns_roce_v2_post_send(struct ib_qp *ibqp,
unsigned int sge_idx;
unsigned int wqe_idx;
void *wqe = NULL;
- int nreq;
+ u32 nreq;
int ret;
spin_lock_irqsave(&qp->sq.lock, flags);
@@ -834,7 +834,7 @@ static void *get_srq_wqe(struct hns_roce_srq *srq, int n)
return hns_roce_buf_offset(srq->buf_mtr.kmem, n << srq->wqe_shift);
}
-static void *get_idx_buf(struct hns_roce_idx_que *idx_que, int n)
+static void *get_idx_buf(struct hns_roce_idx_que *idx_que, unsigned int n)
{
return hns_roce_buf_offset(idx_que->mtr.kmem,
n << idx_que->entry_shift);
@@ -875,12 +875,12 @@ static int hns_roce_v2_post_srq_recv(struct ib_srq *ibsrq,
struct hns_roce_v2_wqe_data_seg *dseg;
struct hns_roce_v2_db srq_db;
unsigned long flags;
+ unsigned int ind;
__le32 *srq_idx;
int ret = 0;
int wqe_idx;
void *wqe;
int nreq;
- int ind;
int i;
spin_lock_irqsave(&srq->lock, flags);
@@ -1125,7 +1125,7 @@ static void hns_roce_cmq_init_regs(struct hns_roce_dev *hr_dev, bool ring_type)
roce_write(hr_dev, ROCEE_TX_CMQ_BASEADDR_H_REG,
upper_32_bits(dma));
roce_write(hr_dev, ROCEE_TX_CMQ_DEPTH_REG,
- ring->desc_num >> HNS_ROCE_CMQ_DESC_NUM_S);
+ (u32)ring->desc_num >> HNS_ROCE_CMQ_DESC_NUM_S);
roce_write(hr_dev, ROCEE_TX_CMQ_HEAD_REG, 0);
roce_write(hr_dev, ROCEE_TX_CMQ_TAIL_REG, 0);
} else {
@@ -1133,7 +1133,7 @@ static void hns_roce_cmq_init_regs(struct hns_roce_dev *hr_dev, bool ring_type)
roce_write(hr_dev, ROCEE_RX_CMQ_BASEADDR_H_REG,
upper_32_bits(dma));
roce_write(hr_dev, ROCEE_RX_CMQ_DEPTH_REG,
- ring->desc_num >> HNS_ROCE_CMQ_DESC_NUM_S);
+ (u32)ring->desc_num >> HNS_ROCE_CMQ_DESC_NUM_S);
roce_write(hr_dev, ROCEE_RX_CMQ_HEAD_REG, 0);
roce_write(hr_dev, ROCEE_RX_CMQ_TAIL_REG, 0);
}
@@ -1919,8 +1919,8 @@ static void set_default_caps(struct hns_roce_dev *hr_dev)
}
}
-static void calc_pg_sz(int obj_num, int obj_size, int hop_num, int ctx_bt_num,
- int *buf_page_size, int *bt_page_size, u32 hem_type)
+static void calc_pg_sz(u32 obj_num, u32 obj_size, u32 hop_num, u32 ctx_bt_num,
+ u32 *buf_page_size, u32 *bt_page_size, u32 hem_type)
{
u64 obj_per_chunk;
u64 bt_chunk_size = PAGE_SIZE;
@@ -2399,10 +2399,10 @@ static int hns_roce_init_link_table(struct hns_roce_dev *hr_dev,
u32 buf_chk_sz;
dma_addr_t t;
int func_num = 1;
- int pg_num_a;
- int pg_num_b;
- int pg_num;
- int size;
+ u32 pg_num_a;
+ u32 pg_num_b;
+ u32 pg_num;
+ u32 size;
int i;
switch (type) {
@@ -2598,7 +2598,7 @@ static int hns_roce_query_mbox_status(struct hns_roce_dev *hr_dev)
struct hns_roce_cmq_desc desc;
struct hns_roce_mbox_status *mb_st =
(struct hns_roce_mbox_status *)desc.data;
- enum hns_roce_cmd_return_status status;
+ int status;
hns_roce_cmq_setup_basic_desc(&desc, HNS_ROCE_OPC_QUERY_MB_ST, true);
@@ -2669,7 +2669,7 @@ static int hns_roce_v2_post_mbox(struct hns_roce_dev *hr_dev, u64 in_param,
}
static int hns_roce_v2_chk_mbox(struct hns_roce_dev *hr_dev,
- unsigned long timeout)
+ unsigned int timeout)
{
struct device *dev = hr_dev->dev;
unsigned long end;
@@ -3067,7 +3067,7 @@ static void *get_cqe_v2(struct hns_roce_cq *hr_cq, int n)
return hns_roce_buf_offset(hr_cq->mtr.kmem, n * hr_cq->cqe_size);
}
-static void *get_sw_cqe_v2(struct hns_roce_cq *hr_cq, int n)
+static void *get_sw_cqe_v2(struct hns_roce_cq *hr_cq, unsigned int n)
{
struct hns_roce_v2_cqe *cqe = get_cqe_v2(hr_cq, n & hr_cq->ib_cq.cqe);
@@ -3414,7 +3414,7 @@ static int hns_roce_v2_poll_one(struct hns_roce_cq *hr_cq,
int is_send;
u16 wqe_ctr;
u32 opcode;
- int qpn;
+ u32 qpn;
int ret;
/* Find cqe according to consumer index */
diff --git a/drivers/infiniband/hw/hns/hns_roce_main.c b/drivers/infiniband/hw/hns/hns_roce_main.c
index b6f27f9..2c56552 100644
--- a/drivers/infiniband/hw/hns/hns_roce_main.c
+++ b/drivers/infiniband/hw/hns/hns_roce_main.c
@@ -54,7 +54,7 @@
* GID[0][0], GID[1][0],.....GID[N - 1][0],
* And so on
*/
-int hns_get_gid_index(struct hns_roce_dev *hr_dev, u8 port, int gid_index)
+u8 hns_get_gid_index(struct hns_roce_dev *hr_dev, u8 port, int gid_index)
{
return gid_index * hr_dev->caps.num_ports + port;
}
diff --git a/drivers/infiniband/hw/hns/hns_roce_mr.c b/drivers/infiniband/hw/hns/hns_roce_mr.c
index 5c43089..00544f2 100644
--- a/drivers/infiniband/hw/hns/hns_roce_mr.c
+++ b/drivers/infiniband/hw/hns/hns_roce_mr.c
@@ -509,7 +509,7 @@ int hns_roce_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, int sg_nents,
ibdev_err(ibdev, "failed to map sg mtr, ret = %d.\n", ret);
ret = 0;
} else {
- mr->pbl_mtr.hem_cfg.buf_pg_shift = ilog2(ibmr->page_size);
+ mr->pbl_mtr.hem_cfg.buf_pg_shift = (u32)ilog2(ibmr->page_size);
ret = mr->npages;
}
@@ -828,12 +828,12 @@ static int mtr_get_pages(struct hns_roce_dev *hr_dev, struct hns_roce_mtr *mtr,
}
int hns_roce_mtr_map(struct hns_roce_dev *hr_dev, struct hns_roce_mtr *mtr,
- dma_addr_t *pages, int page_cnt)
+ dma_addr_t *pages, unsigned int page_cnt)
{
struct ib_device *ibdev = &hr_dev->ib_dev;
struct hns_roce_buf_region *r;
+ unsigned int i;
int err;
- int i;
/*
* Only use the first page address as root ba when hopnum is 0, this
@@ -870,13 +870,12 @@ int hns_roce_mtr_find(struct hns_roce_dev *hr_dev, struct hns_roce_mtr *mtr,
int offset, u64 *mtt_buf, int mtt_max, u64 *base_addr)
{
struct hns_roce_hem_cfg *cfg = &mtr->hem_cfg;
+ int mtt_count, left;
int start_index;
- int mtt_count;
int total = 0;
__le64 *mtts;
- int npage;
+ u32 npage;
u64 addr;
- int left;
if (!mtt_buf || mtt_max < 1)
goto done;
diff --git a/drivers/infiniband/hw/hns/hns_roce_qp.c b/drivers/infiniband/hw/hns/hns_roce_qp.c
index b59f3e0..77b8a6e 100644
--- a/drivers/infiniband/hw/hns/hns_roce_qp.c
+++ b/drivers/infiniband/hw/hns/hns_roce_qp.c
@@ -1323,22 +1323,22 @@ static inline void *get_wqe(struct hns_roce_qp *hr_qp, int offset)
return hns_roce_buf_offset(hr_qp->mtr.kmem, offset);
}
-void *hns_roce_get_recv_wqe(struct hns_roce_qp *hr_qp, int n)
+void *hns_roce_get_recv_wqe(struct hns_roce_qp *hr_qp, unsigned int n)
{
return get_wqe(hr_qp, hr_qp->rq.offset + (n << hr_qp->rq.wqe_shift));
}
-void *hns_roce_get_send_wqe(struct hns_roce_qp *hr_qp, int n)
+void *hns_roce_get_send_wqe(struct hns_roce_qp *hr_qp, unsigned int n)
{
return get_wqe(hr_qp, hr_qp->sq.offset + (n << hr_qp->sq.wqe_shift));
}
-void *hns_roce_get_extend_sge(struct hns_roce_qp *hr_qp, int n)
+void *hns_roce_get_extend_sge(struct hns_roce_qp *hr_qp, unsigned int n)
{
return get_wqe(hr_qp, hr_qp->sge.offset + (n << hr_qp->sge.sge_shift));
}
-bool hns_roce_wq_overflow(struct hns_roce_wq *hr_wq, int nreq,
+bool hns_roce_wq_overflow(struct hns_roce_wq *hr_wq, u32 nreq,
struct ib_cq *ib_cq)
{
struct hns_roce_cq *hr_cq;
--
2.8.1
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH v2 for-next 10/11] RDMA/hns: Fix inaccurate prints
2020-12-04 10:40 [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11 Weihang Li
` (8 preceding siblings ...)
2020-12-04 10:40 ` [PATCH v2 for-next 09/11] RDMA/hns: Fix incorrect symbol types Weihang Li
@ 2020-12-04 10:40 ` Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 11/11] RDMA/hns: Simplify AEQE process for different types of queue Weihang Li
10 siblings, 0 replies; 16+ messages in thread
From: Weihang Li @ 2020-12-04 10:40 UTC (permalink / raw)
To: dledford, jgg; +Cc: leon, linux-rdma, linuxarm
From: Yixing Liu <liuyixing1@huawei.com>
Some %d in print format string should be %u, and some prints miss the
useful errno or are in nonstandard format. Just fix above issues.
Signed-off-by: Yixing Liu <liuyixing1@huawei.com>
Signed-off-by: Weihang Li <liweihang@huawei.com>
---
drivers/infiniband/hw/hns/hns_roce_alloc.c | 4 +-
drivers/infiniband/hw/hns/hns_roce_cq.c | 35 ++++++++---------
drivers/infiniband/hw/hns/hns_roce_hem.c | 16 ++++----
drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 26 ++++++-------
drivers/infiniband/hw/hns/hns_roce_mr.c | 10 ++---
drivers/infiniband/hw/hns/hns_roce_pd.c | 2 +-
drivers/infiniband/hw/hns/hns_roce_qp.c | 61 +++++++++++++++++-------------
drivers/infiniband/hw/hns/hns_roce_srq.c | 37 ++++++++++--------
8 files changed, 104 insertions(+), 87 deletions(-)
diff --git a/drivers/infiniband/hw/hns/hns_roce_alloc.c b/drivers/infiniband/hw/hns/hns_roce_alloc.c
index dad2b9b..4bcaaa05 100644
--- a/drivers/infiniband/hw/hns/hns_roce_alloc.c
+++ b/drivers/infiniband/hw/hns/hns_roce_alloc.c
@@ -260,7 +260,7 @@ int hns_roce_get_kmem_bufs(struct hns_roce_dev *hr_dev, dma_addr_t *bufs,
end = start + buf_cnt;
if (end > buf->npages) {
dev_err(hr_dev->dev,
- "Failed to check kmem bufs, end %d + %d total %d!\n",
+ "failed to check kmem bufs, end %d + %d total %u!\n",
start, buf_cnt, buf->npages);
return -EINVAL;
}
@@ -282,7 +282,7 @@ int hns_roce_get_umem_bufs(struct hns_roce_dev *hr_dev, dma_addr_t *bufs,
u64 addr;
if (page_shift < HNS_HW_PAGE_SHIFT) {
- dev_err(hr_dev->dev, "Failed to check umem page shift %d!\n",
+ dev_err(hr_dev->dev, "failed to check umem page shift %u!\n",
page_shift);
return -EINVAL;
}
diff --git a/drivers/infiniband/hw/hns/hns_roce_cq.c b/drivers/infiniband/hw/hns/hns_roce_cq.c
index 9b6c493..1edcbbe 100644
--- a/drivers/infiniband/hw/hns/hns_roce_cq.c
+++ b/drivers/infiniband/hw/hns/hns_roce_cq.c
@@ -50,29 +50,29 @@ static int alloc_cqc(struct hns_roce_dev *hr_dev, struct hns_roce_cq *hr_cq)
ret = hns_roce_mtr_find(hr_dev, &hr_cq->mtr, 0, mtts, ARRAY_SIZE(mtts),
&dma_handle);
- if (ret < 1) {
- ibdev_err(ibdev, "Failed to find CQ mtr\n");
+ if (!ret) {
+ ibdev_err(ibdev, "failed to find CQ mtr, ret = %d.\n", ret);
return -EINVAL;
}
cq_table = &hr_dev->cq_table;
ret = hns_roce_bitmap_alloc(&cq_table->bitmap, &hr_cq->cqn);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc CQ bitmap, err %d\n", ret);
+ ibdev_err(ibdev, "failed to alloc CQ bitmap, ret = %d.\n", ret);
return ret;
}
/* Get CQC memory HEM(Hardware Entry Memory) table */
ret = hns_roce_table_get(hr_dev, &cq_table->table, hr_cq->cqn);
if (ret) {
- ibdev_err(ibdev, "Failed to get CQ(0x%lx) context, err %d\n",
+ ibdev_err(ibdev, "failed to get CQ(0x%lx) context, ret = %d.\n",
hr_cq->cqn, ret);
goto err_out;
}
ret = xa_err(xa_store(&cq_table->array, hr_cq->cqn, hr_cq, GFP_KERNEL));
if (ret) {
- ibdev_err(ibdev, "Failed to xa_store CQ\n");
+ ibdev_err(ibdev, "failed to xa_store CQ, ret = %d.\n", ret);
goto err_put;
}
@@ -91,7 +91,7 @@ static int alloc_cqc(struct hns_roce_dev *hr_dev, struct hns_roce_cq *hr_cq)
hns_roce_free_cmd_mailbox(hr_dev, mailbox);
if (ret) {
ibdev_err(ibdev,
- "Failed to send create cmd for CQ(0x%lx), err %d\n",
+ "failed to send create cmd for CQ(0x%lx), ret = %d.\n",
hr_cq->cqn, ret);
goto err_xa;
}
@@ -147,7 +147,7 @@ static int alloc_cq_buf(struct hns_roce_dev *hr_dev, struct hns_roce_cq *hr_cq,
{
struct ib_device *ibdev = &hr_dev->ib_dev;
struct hns_roce_buf_attr buf_attr = {};
- int err;
+ int ret;
buf_attr.page_shift = hr_dev->caps.cqe_buf_pg_sz + HNS_HW_PAGE_SHIFT;
buf_attr.region[0].size = hr_cq->cq_depth * hr_cq->cqe_size;
@@ -155,13 +155,13 @@ static int alloc_cq_buf(struct hns_roce_dev *hr_dev, struct hns_roce_cq *hr_cq,
buf_attr.region_count = 1;
buf_attr.fixed_page = true;
- err = hns_roce_mtr_create(hr_dev, &hr_cq->mtr, &buf_attr,
+ ret = hns_roce_mtr_create(hr_dev, &hr_cq->mtr, &buf_attr,
hr_dev->caps.cqe_ba_pg_sz + HNS_HW_PAGE_SHIFT,
udata, addr);
- if (err)
- ibdev_err(ibdev, "Failed to alloc CQ mtr, err %d\n", err);
+ if (ret)
+ ibdev_err(ibdev, "failed to alloc CQ mtr, ret = %d.\n", ret);
- return err;
+ return ret;
}
static void free_cq_buf(struct hns_roce_dev *hr_dev, struct hns_roce_cq *hr_cq)
@@ -255,13 +255,13 @@ int hns_roce_create_cq(struct ib_cq *ib_cq, const struct ib_cq_init_attr *attr,
return -EOPNOTSUPP;
if (cq_entries < 1 || cq_entries > hr_dev->caps.max_cqes) {
- ibdev_err(ibdev, "Failed to check CQ count %d max=%d\n",
+ ibdev_err(ibdev, "failed to check CQ count %u, max = %u.\n",
cq_entries, hr_dev->caps.max_cqes);
return -EINVAL;
}
if (vector >= hr_dev->caps.num_comp_vectors) {
- ibdev_err(ibdev, "Failed to check CQ vector=%d max=%d\n",
+ ibdev_err(ibdev, "failed to check CQ vector = %d, max = %d.\n",
vector, hr_dev->caps.num_comp_vectors);
return -EINVAL;
}
@@ -279,7 +279,7 @@ int hns_roce_create_cq(struct ib_cq *ib_cq, const struct ib_cq_init_attr *attr,
ret = ib_copy_from_udata(&ucmd, udata,
min(udata->inlen, sizeof(ucmd)));
if (ret) {
- ibdev_err(ibdev, "Failed to copy CQ udata, err %d\n",
+ ibdev_err(ibdev, "failed to copy CQ udata, ret = %d.\n",
ret);
return ret;
}
@@ -289,19 +289,20 @@ int hns_roce_create_cq(struct ib_cq *ib_cq, const struct ib_cq_init_attr *attr,
ret = alloc_cq_buf(hr_dev, hr_cq, udata, ucmd.buf_addr);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc CQ buf, err %d\n", ret);
+ ibdev_err(ibdev, "failed to alloc CQ buf, ret = %d.\n", ret);
return ret;
}
ret = alloc_cq_db(hr_dev, hr_cq, udata, ucmd.db_addr, &resp);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc CQ db, err %d\n", ret);
+ ibdev_err(ibdev, "failed to alloc CQ db, ret = %d.\n", ret);
goto err_cq_buf;
}
ret = alloc_cqc(hr_dev, hr_cq);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc CQ context, err %d\n", ret);
+ ibdev_err(ibdev,
+ "failed to alloc CQ context, ret = %d.\n", ret);
goto err_cq_db;
}
diff --git a/drivers/infiniband/hw/hns/hns_roce_hem.c b/drivers/infiniband/hw/hns/hns_roce_hem.c
index f19bbcc..7e6e894 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hem.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hem.c
@@ -195,7 +195,7 @@ static int get_hem_table_config(struct hns_roce_dev *hr_dev,
mhop->hop_num = hr_dev->caps.gmv_hop_num;
break;
default:
- dev_err(dev, "Table %d not support multi-hop addressing!\n",
+ dev_err(dev, "table %u not support multi-hop addressing!\n",
type);
return -EINVAL;
}
@@ -243,8 +243,8 @@ int hns_roce_calc_hem_mhop(struct hns_roce_dev *hr_dev,
mhop->l0_idx = table_idx;
break;
default:
- dev_err(dev, "Table %d not support hop_num = %d!\n",
- table->type, mhop->hop_num);
+ dev_err(dev, "table %u not support hop_num = %u!\n",
+ table->type, mhop->hop_num);
return -EINVAL;
}
if (mhop->l0_idx >= mhop->ba_l0_num)
@@ -449,13 +449,13 @@ static int calc_hem_config(struct hns_roce_dev *hr_dev,
index->buf = l0_idx;
break;
default:
- ibdev_err(ibdev, "Table %d not support mhop.hop_num = %d!\n",
+ ibdev_err(ibdev, "table %u not support mhop.hop_num = %u!\n",
table->type, mhop->hop_num);
return -EINVAL;
}
if (unlikely(index->buf >= table->num_hem)) {
- ibdev_err(ibdev, "Table %d exceed hem limt idx %llu,max %lu!\n",
+ ibdev_err(ibdev, "table %u exceed hem limt idx %llu, max %lu!\n",
table->type, index->buf, table->num_hem);
return -EINVAL;
}
@@ -725,15 +725,15 @@ static void clear_mhop_hem(struct hns_roce_dev *hr_dev,
step_idx = hop_num;
if (hr_dev->hw->clear_hem(hr_dev, table, obj, step_idx))
- ibdev_warn(ibdev, "Clear hop%d HEM failed.\n", hop_num);
+ ibdev_warn(ibdev, "failed to clear hop%u HEM.\n", hop_num);
if (index->inited & HEM_INDEX_L1)
if (hr_dev->hw->clear_hem(hr_dev, table, obj, 1))
- ibdev_warn(ibdev, "Clear HEM step 1 failed.\n");
+ ibdev_warn(ibdev, "failed to clear HEM step 1.\n");
if (index->inited & HEM_INDEX_L0)
if (hr_dev->hw->clear_hem(hr_dev, table, obj, 0))
- ibdev_warn(ibdev, "Clear HEM step 0 failed.\n");
+ ibdev_warn(ibdev, "failed to clear HEM step 0.\n");
}
}
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
index 6c006af..7f4e725 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
@@ -765,7 +765,7 @@ static int hns_roce_v2_post_recv(struct ib_qp *ibqp,
wqe_idx = (hr_qp->rq.head + nreq) & (hr_qp->rq.wqe_cnt - 1);
if (unlikely(wr->num_sge > hr_qp->rq.max_gs)) {
- ibdev_err(ibdev, "rq:num_sge=%d >= qp->sq.max_gs=%d\n",
+ ibdev_err(ibdev, "num_sge = %d >= max_sge = %u.\n",
wr->num_sge, hr_qp->rq.max_gs);
ret = -EINVAL;
*bad_wr = wr;
@@ -1950,8 +1950,8 @@ static void calc_pg_sz(u32 obj_num, u32 obj_size, u32 hop_num, u32 ctx_bt_num,
obj_per_chunk = ctx_bt_num * obj_per_chunk_default;
break;
default:
- pr_err("Table %d not support hop_num = %d!\n", hem_type,
- hop_num);
+ pr_err("table %u not support hop_num = %u!\n", hem_type,
+ hop_num);
return;
}
@@ -3683,7 +3683,7 @@ static int get_op_for_set_hem(struct hns_roce_dev *hr_dev, u32 type,
break;
default:
dev_warn(hr_dev->dev,
- "Table %d not to be written by mailbox!\n", type);
+ "table %u not to be written by mailbox!\n", type);
return -EINVAL;
}
@@ -3819,7 +3819,7 @@ static int hns_roce_v2_clear_hem(struct hns_roce_dev *hr_dev,
case HEM_TYPE_GMV:
return 0;
default:
- dev_warn(dev, "Table %d not to be destroyed by mailbox!\n",
+ dev_warn(dev, "table %u not to be destroyed by mailbox!\n",
table->type);
return 0;
}
@@ -4411,7 +4411,7 @@ static int modify_qp_rtr_to_rts(struct ib_qp *ibqp,
ret = config_qp_sq_buf(hr_dev, hr_qp, context, qpc_mask);
if (ret) {
- ibdev_err(ibdev, "failed to config sq buf, ret %d\n", ret);
+ ibdev_err(ibdev, "failed to config sq buf, ret = %d.\n", ret);
return ret;
}
@@ -4898,7 +4898,7 @@ static int hns_roce_v2_modify_qp(struct ib_qp *ibqp,
/* SW pass context to HW */
ret = hns_roce_v2_qp_modify(hr_dev, context, qpc_mask, hr_qp);
if (ret) {
- ibdev_err(ibdev, "failed to modify QP, ret = %d\n", ret);
+ ibdev_err(ibdev, "failed to modify QP, ret = %d.\n", ret);
goto out;
}
@@ -4991,7 +4991,7 @@ static int hns_roce_v2_query_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr,
ret = hns_roce_v2_query_qpc(hr_dev, hr_qp, &context);
if (ret) {
- ibdev_err(ibdev, "failed to query QPC, ret = %d\n", ret);
+ ibdev_err(ibdev, "failed to query QPC, ret = %d.\n", ret);
ret = -EINVAL;
goto out;
}
@@ -5114,7 +5114,7 @@ static int hns_roce_v2_destroy_qp_common(struct hns_roce_dev *hr_dev,
hr_qp->state, IB_QPS_RESET);
if (ret)
ibdev_err(ibdev,
- "failed to modify QP to RST, ret = %d\n",
+ "failed to modify QP to RST, ret = %d.\n",
ret);
}
@@ -5153,7 +5153,7 @@ static int hns_roce_v2_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata)
ret = hns_roce_v2_destroy_qp_common(hr_dev, hr_qp, udata);
if (ret)
ibdev_err(&hr_dev->ib_dev,
- "failed to destroy QP 0x%06lx, ret = %d\n",
+ "failed to destroy QP, QPN = 0x%06lx, ret = %d.\n",
hr_qp->qpn, ret);
hns_roce_qp_destroy(hr_dev, hr_qp, udata);
@@ -5176,7 +5176,7 @@ static int hns_roce_v2_qp_flow_control_init(struct hns_roce_dev *hr_dev,
hns_roce_cmq_setup_basic_desc(&desc, HNS_ROCE_OPC_RESET_SCCC, false);
ret = hns_roce_cmq_send(hr_dev, &desc, 1);
if (ret) {
- ibdev_err(ibdev, "failed to reset SCC ctx, ret = %d\n", ret);
+ ibdev_err(ibdev, "failed to reset SCC ctx, ret = %d.\n", ret);
goto out;
}
@@ -5186,7 +5186,7 @@ static int hns_roce_v2_qp_flow_control_init(struct hns_roce_dev *hr_dev,
clr->qpn = cpu_to_le32(hr_qp->qpn);
ret = hns_roce_cmq_send(hr_dev, &desc, 1);
if (ret) {
- ibdev_err(ibdev, "failed to clear SCC ctx, ret = %d\n", ret);
+ ibdev_err(ibdev, "failed to clear SCC ctx, ret = %d.\n", ret);
goto out;
}
@@ -5435,7 +5435,7 @@ static int hns_roce_v2_modify_cq(struct ib_cq *cq, u16 cq_count, u16 cq_period)
hns_roce_free_cmd_mailbox(hr_dev, mailbox);
if (ret)
ibdev_err(&hr_dev->ib_dev,
- "failed to process cmd when modifying CQ, ret = %d\n",
+ "failed to process cmd when modifying CQ, ret = %d.\n",
ret);
return ret;
diff --git a/drivers/infiniband/hw/hns/hns_roce_mr.c b/drivers/infiniband/hw/hns/hns_roce_mr.c
index 00544f2..84029fd05 100644
--- a/drivers/infiniband/hw/hns/hns_roce_mr.c
+++ b/drivers/infiniband/hw/hns/hns_roce_mr.c
@@ -185,14 +185,14 @@ static int hns_roce_mr_enable(struct hns_roce_dev *hr_dev,
else
ret = hr_dev->hw->frmr_write_mtpt(hr_dev, mailbox->buf, mr);
if (ret) {
- dev_err(dev, "Write mtpt fail!\n");
+ dev_err(dev, "failed to write mtpt, ret = %d.\n", ret);
goto err_page;
}
ret = hns_roce_hw_create_mpt(hr_dev, mailbox,
mtpt_idx & (hr_dev->caps.num_mtpts - 1));
if (ret) {
- dev_err(dev, "CREATE_MPT failed (%d)\n", ret);
+ dev_err(dev, "failed to create mpt, ret = %d.\n", ret);
goto err_page;
}
@@ -495,7 +495,7 @@ int hns_roce_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, int sg_nents,
ret = ib_sg_to_pages(ibmr, sg, sg_nents, sg_offset, hns_roce_set_page);
if (ret < 1) {
- ibdev_err(ibdev, "failed to store sg pages %d %d, cnt = %d.\n",
+ ibdev_err(ibdev, "failed to store sg pages %u %u, cnt = %d.\n",
mr->npages, mr->pbl_mtr.hem_cfg.buf_pg_count, ret);
goto err_page_list;
}
@@ -849,7 +849,7 @@ int hns_roce_mtr_map(struct hns_roce_dev *hr_dev, struct hns_roce_mtr *mtr,
if (r->offset + r->count > page_cnt) {
err = -EINVAL;
ibdev_err(ibdev,
- "Failed to check mtr%d end %d + %d, max %d\n",
+ "failed to check mtr%u end %u + %u, max %u.\n",
i, r->offset, r->count, page_cnt);
return err;
}
@@ -857,7 +857,7 @@ int hns_roce_mtr_map(struct hns_roce_dev *hr_dev, struct hns_roce_mtr *mtr,
err = mtr_map_region(hr_dev, mtr, &pages[r->offset], r);
if (err) {
ibdev_err(ibdev,
- "Failed to map mtr%d offset %d, err %d\n",
+ "failed to map mtr%u offset %u, ret = %d.\n",
i, r->offset, err);
return err;
}
diff --git a/drivers/infiniband/hw/hns/hns_roce_pd.c b/drivers/infiniband/hw/hns/hns_roce_pd.c
index f78fa1d..012a769 100644
--- a/drivers/infiniband/hw/hns/hns_roce_pd.c
+++ b/drivers/infiniband/hw/hns/hns_roce_pd.c
@@ -65,7 +65,7 @@ int hns_roce_alloc_pd(struct ib_pd *ibpd, struct ib_udata *udata)
ret = hns_roce_pd_alloc(to_hr_dev(ib_dev), &pd->pdn);
if (ret) {
- ibdev_err(ib_dev, "failed to alloc pd, ret = %d\n", ret);
+ ibdev_err(ib_dev, "failed to alloc pd, ret = %d.\n", ret);
return ret;
}
diff --git a/drivers/infiniband/hw/hns/hns_roce_qp.c b/drivers/infiniband/hw/hns/hns_roce_qp.c
index 77b8a6e..192962d 100644
--- a/drivers/infiniband/hw/hns/hns_roce_qp.c
+++ b/drivers/infiniband/hw/hns/hns_roce_qp.c
@@ -514,12 +514,12 @@ static int check_sq_size_with_integrity(struct hns_roce_dev *hr_dev,
/* Sanity check SQ size before proceeding */
if (ucmd->log_sq_stride > max_sq_stride ||
ucmd->log_sq_stride < HNS_ROCE_IB_MIN_SQ_STRIDE) {
- ibdev_err(&hr_dev->ib_dev, "Failed to check SQ stride size\n");
+ ibdev_err(&hr_dev->ib_dev, "failed to check SQ stride size.\n");
return -EINVAL;
}
if (cap->max_send_sge > hr_dev->caps.max_sq_sg) {
- ibdev_err(&hr_dev->ib_dev, "Failed to check SQ SGE size %d\n",
+ ibdev_err(&hr_dev->ib_dev, "failed to check SQ SGE size %u.\n",
cap->max_send_sge);
return -EINVAL;
}
@@ -622,7 +622,7 @@ static int set_kernel_sq_size(struct hns_roce_dev *hr_dev,
cnt = roundup_pow_of_two(max(cap->max_send_wr, hr_dev->caps.min_wqes));
if (cnt > hr_dev->caps.max_wqes) {
- ibdev_err(ibdev, "failed to check WQE num, WQE num = %d.\n",
+ ibdev_err(ibdev, "failed to check WQE num, WQE num = %u.\n",
cnt);
return -EINVAL;
}
@@ -796,7 +796,8 @@ static int alloc_qp_db(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp,
&hr_qp->sdb);
if (ret) {
ibdev_err(ibdev,
- "Failed to map user SQ doorbell\n");
+ "failed to map user SQ doorbell, ret = %d.\n",
+ ret);
goto err_out;
}
hr_qp->en_flags |= HNS_ROCE_QP_CAP_SQ_RECORD_DB;
@@ -808,7 +809,8 @@ static int alloc_qp_db(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp,
&hr_qp->rdb);
if (ret) {
ibdev_err(ibdev,
- "Failed to map user RQ doorbell\n");
+ "failed to map user RQ doorbell, ret = %d.\n",
+ ret);
goto err_sdb;
}
hr_qp->en_flags |= HNS_ROCE_QP_CAP_RQ_RECORD_DB;
@@ -825,7 +827,8 @@ static int alloc_qp_db(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp,
ret = hns_roce_alloc_db(hr_dev, &hr_qp->rdb, 0);
if (ret) {
ibdev_err(ibdev,
- "Failed to alloc kernel RQ doorbell\n");
+ "failed to alloc kernel RQ doorbell, ret = %d.\n",
+ ret);
goto err_out;
}
*hr_qp->rdb.db_record = 0;
@@ -868,14 +871,14 @@ static int alloc_kernel_wrid(struct hns_roce_dev *hr_dev,
sq_wrid = kcalloc(hr_qp->sq.wqe_cnt, sizeof(u64), GFP_KERNEL);
if (ZERO_OR_NULL_PTR(sq_wrid)) {
- ibdev_err(ibdev, "Failed to alloc SQ wrid\n");
+ ibdev_err(ibdev, "failed to alloc SQ wrid.\n");
return -ENOMEM;
}
if (hr_qp->rq.wqe_cnt) {
rq_wrid = kcalloc(hr_qp->rq.wqe_cnt, sizeof(u64), GFP_KERNEL);
if (ZERO_OR_NULL_PTR(rq_wrid)) {
- ibdev_err(ibdev, "Failed to alloc RQ wrid\n");
+ ibdev_err(ibdev, "failed to alloc RQ wrid.\n");
ret = -ENOMEM;
goto err_sq;
}
@@ -935,11 +938,15 @@ static int set_qp_param(struct hns_roce_dev *hr_dev, struct hns_roce_qp *hr_qp,
ret = set_user_sq_size(hr_dev, &init_attr->cap, hr_qp, ucmd);
if (ret)
- ibdev_err(ibdev, "Failed to set user SQ size\n");
+ ibdev_err(ibdev,
+ "failed to set user SQ size, ret = %d.\n",
+ ret);
} else {
ret = set_kernel_sq_size(hr_dev, &init_attr->cap, hr_qp);
if (ret)
- ibdev_err(ibdev, "Failed to set kernel SQ size\n");
+ ibdev_err(ibdev,
+ "failed to set kernel SQ size, ret = %d.\n",
+ ret);
}
return ret;
@@ -968,45 +975,48 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev,
ret = set_qp_param(hr_dev, hr_qp, init_attr, udata, &ucmd);
if (ret) {
- ibdev_err(ibdev, "Failed to set QP param\n");
+ ibdev_err(ibdev, "failed to set QP param, ret = %d.\n", ret);
return ret;
}
if (!udata) {
ret = alloc_kernel_wrid(hr_dev, hr_qp);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc wrid\n");
+ ibdev_err(ibdev, "failed to alloc wrid, ret = %d.\n",
+ ret);
return ret;
}
}
ret = alloc_qp_db(hr_dev, hr_qp, init_attr, udata, &ucmd, &resp);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc QP doorbell\n");
+ ibdev_err(ibdev, "failed to alloc QP doorbell, ret = %d.\n",
+ ret);
goto err_wrid;
}
ret = alloc_qp_buf(hr_dev, hr_qp, init_attr, udata, ucmd.buf_addr);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc QP buffer\n");
+ ibdev_err(ibdev, "failed to alloc QP buffer, ret = %d.\n", ret);
goto err_db;
}
ret = alloc_qpn(hr_dev, hr_qp);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc QPN\n");
+ ibdev_err(ibdev, "failed to alloc QPN, ret = %d.\n", ret);
goto err_buf;
}
ret = alloc_qpc(hr_dev, hr_qp);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc QP context\n");
+ ibdev_err(ibdev, "failed to alloc QP context, ret = %d.\n",
+ ret);
goto err_qpn;
}
ret = hns_roce_qp_store(hr_dev, hr_qp, init_attr);
if (ret) {
- ibdev_err(ibdev, "Failed to store QP\n");
+ ibdev_err(ibdev, "failed to store QP, ret = %d.\n", ret);
goto err_qpc;
}
@@ -1170,9 +1180,8 @@ static int hns_roce_check_qp_attr(struct ib_qp *ibqp, struct ib_qp_attr *attr,
if ((attr_mask & IB_QP_PORT) &&
(attr->port_num == 0 || attr->port_num > hr_dev->caps.num_ports)) {
- ibdev_err(&hr_dev->ib_dev,
- "attr port_num invalid.attr->port_num=%d\n",
- attr->port_num);
+ ibdev_err(&hr_dev->ib_dev, "invalid attr, port_num = %u.\n",
+ attr->port_num);
return -EINVAL;
}
@@ -1180,8 +1189,8 @@ static int hns_roce_check_qp_attr(struct ib_qp *ibqp, struct ib_qp_attr *attr,
p = attr_mask & IB_QP_PORT ? (attr->port_num - 1) : hr_qp->port;
if (attr->pkey_index >= hr_dev->caps.pkey_table_len[p]) {
ibdev_err(&hr_dev->ib_dev,
- "attr pkey_index invalid.attr->pkey_index=%d\n",
- attr->pkey_index);
+ "invalid attr, pkey_index = %u.\n",
+ attr->pkey_index);
return -EINVAL;
}
}
@@ -1189,16 +1198,16 @@ static int hns_roce_check_qp_attr(struct ib_qp *ibqp, struct ib_qp_attr *attr,
if (attr_mask & IB_QP_MAX_QP_RD_ATOMIC &&
attr->max_rd_atomic > hr_dev->caps.max_qp_init_rdma) {
ibdev_err(&hr_dev->ib_dev,
- "attr max_rd_atomic invalid.attr->max_rd_atomic=%d\n",
- attr->max_rd_atomic);
+ "invalid attr, max_rd_atomic = %u.\n",
+ attr->max_rd_atomic);
return -EINVAL;
}
if (attr_mask & IB_QP_MAX_DEST_RD_ATOMIC &&
attr->max_dest_rd_atomic > hr_dev->caps.max_qp_dest_rdma) {
ibdev_err(&hr_dev->ib_dev,
- "attr max_dest_rd_atomic invalid.attr->max_dest_rd_atomic=%d\n",
- attr->max_dest_rd_atomic);
+ "invalid attr, max_dest_rd_atomic = %u.\n",
+ attr->max_dest_rd_atomic);
return -EINVAL;
}
diff --git a/drivers/infiniband/hw/hns/hns_roce_srq.c b/drivers/infiniband/hw/hns/hns_roce_srq.c
index bf0b51e..bf229c2 100644
--- a/drivers/infiniband/hw/hns/hns_roce_srq.c
+++ b/drivers/infiniband/hw/hns/hns_roce_srq.c
@@ -93,7 +93,8 @@ static int alloc_srqc(struct hns_roce_dev *hr_dev, struct hns_roce_srq *srq,
ret = hns_roce_mtr_find(hr_dev, &srq->buf_mtr, 0, mtts_wqe,
ARRAY_SIZE(mtts_wqe), &dma_handle_wqe);
if (ret < 1) {
- ibdev_err(ibdev, "Failed to find mtr for SRQ WQE\n");
+ ibdev_err(ibdev, "failed to find mtr for SRQ WQE, ret = %d.\n",
+ ret);
return -ENOBUFS;
}
@@ -101,32 +102,34 @@ static int alloc_srqc(struct hns_roce_dev *hr_dev, struct hns_roce_srq *srq,
ret = hns_roce_mtr_find(hr_dev, &srq->idx_que.mtr, 0, mtts_idx,
ARRAY_SIZE(mtts_idx), &dma_handle_idx);
if (ret < 1) {
- ibdev_err(ibdev, "Failed to find mtr for SRQ idx\n");
+ ibdev_err(ibdev, "failed to find mtr for SRQ idx, ret = %d.\n",
+ ret);
return -ENOBUFS;
}
ret = hns_roce_bitmap_alloc(&srq_table->bitmap, &srq->srqn);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc SRQ number, err %d\n", ret);
+ ibdev_err(ibdev,
+ "failed to alloc SRQ number, ret = %d.\n", ret);
return -ENOMEM;
}
ret = hns_roce_table_get(hr_dev, &srq_table->table, srq->srqn);
if (ret) {
- ibdev_err(ibdev, "Failed to get SRQC table, err %d\n", ret);
+ ibdev_err(ibdev, "failed to get SRQC table, ret = %d.\n", ret);
goto err_out;
}
ret = xa_err(xa_store(&srq_table->xa, srq->srqn, srq, GFP_KERNEL));
if (ret) {
- ibdev_err(ibdev, "Failed to store SRQC, err %d\n", ret);
+ ibdev_err(ibdev, "failed to store SRQC, ret = %d.\n", ret);
goto err_put;
}
mailbox = hns_roce_alloc_cmd_mailbox(hr_dev);
if (IS_ERR_OR_NULL(mailbox)) {
ret = -ENOMEM;
- ibdev_err(ibdev, "Failed to alloc mailbox for SRQC\n");
+ ibdev_err(ibdev, "failed to alloc mailbox for SRQC.\n");
goto err_xa;
}
@@ -137,7 +140,7 @@ static int alloc_srqc(struct hns_roce_dev *hr_dev, struct hns_roce_srq *srq,
ret = hns_roce_hw_create_srq(hr_dev, mailbox, srq->srqn);
hns_roce_free_cmd_mailbox(hr_dev, mailbox);
if (ret) {
- ibdev_err(ibdev, "Failed to config SRQC, err %d\n", ret);
+ ibdev_err(ibdev, "failed to config SRQC, ret = %d.\n", ret);
goto err_xa;
}
@@ -198,7 +201,8 @@ static int alloc_srq_buf(struct hns_roce_dev *hr_dev, struct hns_roce_srq *srq,
hr_dev->caps.srqwqe_ba_pg_sz +
HNS_HW_PAGE_SHIFT, udata, addr);
if (err)
- ibdev_err(ibdev, "Failed to alloc SRQ buf mtr, err %d\n", err);
+ ibdev_err(ibdev,
+ "failed to alloc SRQ buf mtr, ret = %d.\n", err);
return err;
}
@@ -229,14 +233,15 @@ static int alloc_srq_idx(struct hns_roce_dev *hr_dev, struct hns_roce_srq *srq,
hr_dev->caps.idx_ba_pg_sz + HNS_HW_PAGE_SHIFT,
udata, addr);
if (err) {
- ibdev_err(ibdev, "Failed to alloc SRQ idx mtr, err %d\n", err);
+ ibdev_err(ibdev,
+ "failed to alloc SRQ idx mtr, ret = %d.\n", err);
return err;
}
if (!udata) {
idx_que->bitmap = bitmap_zalloc(srq->wqe_cnt, GFP_KERNEL);
if (!idx_que->bitmap) {
- ibdev_err(ibdev, "Failed to alloc SRQ idx bitmap\n");
+ ibdev_err(ibdev, "failed to alloc SRQ idx bitmap.\n");
err = -ENOMEM;
goto err_idx_mtr;
}
@@ -306,7 +311,7 @@ int hns_roce_create_srq(struct ib_srq *ib_srq,
ret = ib_copy_from_udata(&ucmd, udata,
min(udata->inlen, sizeof(ucmd)));
if (ret) {
- ibdev_err(ibdev, "Failed to copy SRQ udata, err %d\n",
+ ibdev_err(ibdev, "failed to copy SRQ udata, ret = %d.\n",
ret);
return ret;
}
@@ -314,20 +319,21 @@ int hns_roce_create_srq(struct ib_srq *ib_srq,
ret = alloc_srq_buf(hr_dev, srq, udata, ucmd.buf_addr);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc SRQ buffer, err %d\n", ret);
+ ibdev_err(ibdev,
+ "failed to alloc SRQ buffer, ret = %d.\n", ret);
return ret;
}
ret = alloc_srq_idx(hr_dev, srq, udata, ucmd.que_addr);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc SRQ idx, err %d\n", ret);
+ ibdev_err(ibdev, "failed to alloc SRQ idx, ret = %d.\n", ret);
goto err_buf_alloc;
}
if (!udata) {
ret = alloc_srq_wrid(hr_dev, srq);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc SRQ wrid, err %d\n",
+ ibdev_err(ibdev, "failed to alloc SRQ wrid, ret = %d.\n",
ret);
goto err_idx_alloc;
}
@@ -339,7 +345,8 @@ int hns_roce_create_srq(struct ib_srq *ib_srq,
ret = alloc_srqc(hr_dev, srq, to_hr_pd(ib_srq->pd)->pdn, cqn, 0, 0);
if (ret) {
- ibdev_err(ibdev, "Failed to alloc SRQ context, err %d\n", ret);
+ ibdev_err(ibdev,
+ "failed to alloc SRQ context, ret = %d.\n", ret);
goto err_wrid_alloc;
}
--
2.8.1
^ permalink raw reply related [flat|nested] 16+ messages in thread
* [PATCH v2 for-next 11/11] RDMA/hns: Simplify AEQE process for different types of queue
2020-12-04 10:40 [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11 Weihang Li
` (9 preceding siblings ...)
2020-12-04 10:40 ` [PATCH v2 for-next 10/11] RDMA/hns: Fix inaccurate prints Weihang Li
@ 2020-12-04 10:40 ` Weihang Li
10 siblings, 0 replies; 16+ messages in thread
From: Weihang Li @ 2020-12-04 10:40 UTC (permalink / raw)
To: dledford, jgg; +Cc: leon, linux-rdma, linuxarm
From: Yixian Liu <liuyixian@huawei.com>
There is no need to get queue number repeatly for different queues from an
AEQE entity, as they are the same. Furthermore, redefine the AEQE structure
to make the codes more readable.
In addition, HNS_ROCE_EVENT_TYPE_CEQ_OVERFLOW is removed because the
hardware never reports this event.
Signed-off-by: Yixian Liu <liuyixian@huawei.com>
Signed-off-by: Wenpeng Liang <liangwenpeng@huawei.com>
Signed-off-by: Weihang Li <liweihang@huawei.com>
---
drivers/infiniband/hw/hns/hns_roce_device.h | 26 ++---------------
drivers/infiniband/hw/hns/hns_roce_hw_v1.c | 16 ++++-------
drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 43 ++++++++++-------------------
3 files changed, 23 insertions(+), 62 deletions(-)
diff --git a/drivers/infiniband/hw/hns/hns_roce_device.h b/drivers/infiniband/hw/hns/hns_roce_device.h
index af73086..42a93c8 100644
--- a/drivers/infiniband/hw/hns/hns_roce_device.h
+++ b/drivers/infiniband/hw/hns/hns_roce_device.h
@@ -169,7 +169,6 @@ enum hns_roce_event {
/* 0x10 and 0x11 is unused in currently application case */
HNS_ROCE_EVENT_TYPE_DB_OVERFLOW = 0x12,
HNS_ROCE_EVENT_TYPE_MB = 0x13,
- HNS_ROCE_EVENT_TYPE_CEQ_OVERFLOW = 0x14,
HNS_ROCE_EVENT_TYPE_FLR = 0x15,
};
@@ -645,10 +644,9 @@ enum {
struct hns_roce_work {
struct hns_roce_dev *hr_dev;
struct work_struct work;
- u32 qpn;
- u32 cqn;
int event_type;
int sub_type;
+ u32 queue_num;
};
struct hns_roce_qp {
@@ -716,28 +714,10 @@ struct hns_roce_aeqe {
__le32 asyn;
union {
struct {
- __le32 qp;
+ __le32 num;
u32 rsv0;
u32 rsv1;
- } qp_event;
-
- struct {
- __le32 srq;
- u32 rsv0;
- u32 rsv1;
- } srq_event;
-
- struct {
- __le32 cq;
- u32 rsv0;
- u32 rsv1;
- } cq_event;
-
- struct {
- __le32 ceqe;
- u32 rsv0;
- u32 rsv1;
- } ce_event;
+ } queue_event;
struct {
__le64 out_param;
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v1.c b/drivers/infiniband/hw/hns/hns_roce_hw_v1.c
index b7dd867..cc20231 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v1.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v1.c
@@ -3683,10 +3683,10 @@ static void hns_roce_v1_qp_err_handle(struct hns_roce_dev *hr_dev,
int phy_port;
int qpn;
- qpn = roce_get_field(aeqe->event.qp_event.qp,
+ qpn = roce_get_field(aeqe->event.queue_event.num,
HNS_ROCE_AEQE_EVENT_QP_EVENT_QP_QPN_M,
HNS_ROCE_AEQE_EVENT_QP_EVENT_QP_QPN_S);
- phy_port = roce_get_field(aeqe->event.qp_event.qp,
+ phy_port = roce_get_field(aeqe->event.queue_event.num,
HNS_ROCE_AEQE_EVENT_QP_EVENT_PORT_NUM_M,
HNS_ROCE_AEQE_EVENT_QP_EVENT_PORT_NUM_S);
if (qpn <= 1)
@@ -3717,9 +3717,9 @@ static void hns_roce_v1_cq_err_handle(struct hns_roce_dev *hr_dev,
struct device *dev = &hr_dev->pdev->dev;
u32 cqn;
- cqn = roce_get_field(aeqe->event.cq_event.cq,
- HNS_ROCE_AEQE_EVENT_CQ_EVENT_CQ_CQN_M,
- HNS_ROCE_AEQE_EVENT_CQ_EVENT_CQ_CQN_S);
+ cqn = roce_get_field(aeqe->event.queue_event.num,
+ HNS_ROCE_AEQE_EVENT_CQ_EVENT_CQ_CQN_M,
+ HNS_ROCE_AEQE_EVENT_CQ_EVENT_CQ_CQN_S);
switch (event_type) {
case HNS_ROCE_EVENT_TYPE_CQ_ACCESS_ERROR:
@@ -3848,12 +3848,6 @@ static int hns_roce_v1_aeq_int(struct hns_roce_dev *hr_dev,
case HNS_ROCE_EVENT_TYPE_DB_OVERFLOW:
hns_roce_v1_db_overflow_handle(hr_dev, aeqe);
break;
- case HNS_ROCE_EVENT_TYPE_CEQ_OVERFLOW:
- dev_warn(dev, "CEQ 0x%lx overflow.\n",
- roce_get_field(aeqe->event.ce_event.ceqe,
- HNS_ROCE_AEQE_EVENT_CE_EVENT_CEQE_CEQN_M,
- HNS_ROCE_AEQE_EVENT_CE_EVENT_CEQE_CEQN_S));
- break;
default:
dev_warn(dev, "Unhandled event %d on EQ %d at idx %u.\n",
event_type, eq->eqn, eq->cons_index);
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
index 7f4e725..dca0f60 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
@@ -5446,8 +5446,6 @@ static void hns_roce_irq_work_handle(struct work_struct *work)
struct hns_roce_work *irq_work =
container_of(work, struct hns_roce_work, work);
struct ib_device *ibdev = &irq_work->hr_dev->ib_dev;
- u32 qpn = irq_work->qpn;
- u32 cqn = irq_work->cqn;
switch (irq_work->event_type) {
case HNS_ROCE_EVENT_TYPE_PATH_MIG:
@@ -5463,15 +5461,15 @@ static void hns_roce_irq_work_handle(struct work_struct *work)
break;
case HNS_ROCE_EVENT_TYPE_WQ_CATAS_ERROR:
ibdev_err(ibdev, "Local work queue 0x%x catast error, sub_event type is: %d\n",
- qpn, irq_work->sub_type);
+ irq_work->queue_num, irq_work->sub_type);
break;
case HNS_ROCE_EVENT_TYPE_INV_REQ_LOCAL_WQ_ERROR:
ibdev_err(ibdev, "Invalid request local work queue 0x%x error.\n",
- qpn);
+ irq_work->queue_num);
break;
case HNS_ROCE_EVENT_TYPE_LOCAL_WQ_ACCESS_ERROR:
ibdev_err(ibdev, "Local access violation work queue 0x%x error, sub_event type is: %d\n",
- qpn, irq_work->sub_type);
+ irq_work->queue_num, irq_work->sub_type);
break;
case HNS_ROCE_EVENT_TYPE_SRQ_LIMIT_REACH:
ibdev_warn(ibdev, "SRQ limit reach.\n");
@@ -5483,10 +5481,10 @@ static void hns_roce_irq_work_handle(struct work_struct *work)
ibdev_err(ibdev, "SRQ catas error.\n");
break;
case HNS_ROCE_EVENT_TYPE_CQ_ACCESS_ERROR:
- ibdev_err(ibdev, "CQ 0x%x access err.\n", cqn);
+ ibdev_err(ibdev, "CQ 0x%x access err.\n", irq_work->queue_num);
break;
case HNS_ROCE_EVENT_TYPE_CQ_OVERFLOW:
- ibdev_warn(ibdev, "CQ 0x%x overflow\n", cqn);
+ ibdev_warn(ibdev, "CQ 0x%x overflow\n", irq_work->queue_num);
break;
case HNS_ROCE_EVENT_TYPE_DB_OVERFLOW:
ibdev_warn(ibdev, "DB overflow.\n");
@@ -5502,8 +5500,7 @@ static void hns_roce_irq_work_handle(struct work_struct *work)
}
static void hns_roce_v2_init_irq_work(struct hns_roce_dev *hr_dev,
- struct hns_roce_eq *eq,
- u32 qpn, u32 cqn)
+ struct hns_roce_eq *eq, u32 queue_num)
{
struct hns_roce_work *irq_work;
@@ -5513,10 +5510,9 @@ static void hns_roce_v2_init_irq_work(struct hns_roce_dev *hr_dev,
INIT_WORK(&(irq_work->work), hns_roce_irq_work_handle);
irq_work->hr_dev = hr_dev;
- irq_work->qpn = qpn;
- irq_work->cqn = cqn;
irq_work->event_type = eq->event_type;
irq_work->sub_type = eq->sub_type;
+ irq_work->queue_num = queue_num;
queue_work(hr_dev->irq_workq, &(irq_work->work));
}
@@ -5568,10 +5564,8 @@ static int hns_roce_v2_aeq_int(struct hns_roce_dev *hr_dev,
struct hns_roce_aeqe *aeqe = next_aeqe_sw_v2(eq);
int aeqe_found = 0;
int event_type;
+ u32 queue_num;
int sub_type;
- u32 srqn;
- u32 qpn;
- u32 cqn;
while (aeqe) {
/* Make sure we read AEQ entry after we have checked the
@@ -5585,15 +5579,9 @@ static int hns_roce_v2_aeq_int(struct hns_roce_dev *hr_dev,
sub_type = roce_get_field(aeqe->asyn,
HNS_ROCE_V2_AEQE_SUB_TYPE_M,
HNS_ROCE_V2_AEQE_SUB_TYPE_S);
- qpn = roce_get_field(aeqe->event.qp_event.qp,
- HNS_ROCE_V2_AEQE_EVENT_QUEUE_NUM_M,
- HNS_ROCE_V2_AEQE_EVENT_QUEUE_NUM_S);
- cqn = roce_get_field(aeqe->event.cq_event.cq,
- HNS_ROCE_V2_AEQE_EVENT_QUEUE_NUM_M,
- HNS_ROCE_V2_AEQE_EVENT_QUEUE_NUM_S);
- srqn = roce_get_field(aeqe->event.srq_event.srq,
- HNS_ROCE_V2_AEQE_EVENT_QUEUE_NUM_M,
- HNS_ROCE_V2_AEQE_EVENT_QUEUE_NUM_S);
+ queue_num = roce_get_field(aeqe->event.queue_event.num,
+ HNS_ROCE_V2_AEQE_EVENT_QUEUE_NUM_M,
+ HNS_ROCE_V2_AEQE_EVENT_QUEUE_NUM_S);
switch (event_type) {
case HNS_ROCE_EVENT_TYPE_PATH_MIG:
@@ -5604,15 +5592,15 @@ static int hns_roce_v2_aeq_int(struct hns_roce_dev *hr_dev,
case HNS_ROCE_EVENT_TYPE_SRQ_LAST_WQE_REACH:
case HNS_ROCE_EVENT_TYPE_INV_REQ_LOCAL_WQ_ERROR:
case HNS_ROCE_EVENT_TYPE_LOCAL_WQ_ACCESS_ERROR:
- hns_roce_qp_event(hr_dev, qpn, event_type);
+ hns_roce_qp_event(hr_dev, queue_num, event_type);
break;
case HNS_ROCE_EVENT_TYPE_SRQ_LIMIT_REACH:
case HNS_ROCE_EVENT_TYPE_SRQ_CATAS_ERROR:
- hns_roce_srq_event(hr_dev, srqn, event_type);
+ hns_roce_srq_event(hr_dev, queue_num, event_type);
break;
case HNS_ROCE_EVENT_TYPE_CQ_ACCESS_ERROR:
case HNS_ROCE_EVENT_TYPE_CQ_OVERFLOW:
- hns_roce_cq_event(hr_dev, cqn, event_type);
+ hns_roce_cq_event(hr_dev, queue_num, event_type);
break;
case HNS_ROCE_EVENT_TYPE_MB:
hns_roce_cmd_event(hr_dev,
@@ -5621,7 +5609,6 @@ static int hns_roce_v2_aeq_int(struct hns_roce_dev *hr_dev,
le64_to_cpu(aeqe->event.cmd.out_param));
break;
case HNS_ROCE_EVENT_TYPE_DB_OVERFLOW:
- case HNS_ROCE_EVENT_TYPE_CEQ_OVERFLOW:
case HNS_ROCE_EVENT_TYPE_FLR:
break;
default:
@@ -5638,7 +5625,7 @@ static int hns_roce_v2_aeq_int(struct hns_roce_dev *hr_dev,
if (eq->cons_index > (2 * eq->entries - 1))
eq->cons_index = 0;
- hns_roce_v2_init_irq_work(hr_dev, eq, qpn, cqn);
+ hns_roce_v2_init_irq_work(hr_dev, eq, queue_num);
aeqe = next_aeqe_sw_v2(eq);
}
--
2.8.1
^ permalink raw reply related [flat|nested] 16+ messages in thread
* Re: [PATCH v2 for-next 05/11] RDMA/hns: WARN_ON if get a reserved sl from users
2020-12-04 10:40 ` [PATCH v2 for-next 05/11] RDMA/hns: WARN_ON if get a reserved sl from users Weihang Li
@ 2020-12-09 21:09 ` Jason Gunthorpe
2020-12-10 4:00 ` liweihang
0 siblings, 1 reply; 16+ messages in thread
From: Jason Gunthorpe @ 2020-12-09 21:09 UTC (permalink / raw)
To: Weihang Li; +Cc: dledford, leon, linux-rdma, linuxarm
On Fri, Dec 04, 2020 at 06:40:30PM +0800, Weihang Li wrote:
> According to the RoCE v1 specification, the sl (service level) 0-7 are
> mapped directly to priorities 0-7 respectively, sl 8-15 are reserved. The
> driver should verify whether the value of sl is larger than 7, if so, an
> exception should be returned.
>
> Fixes: 172505cfa3a8 ("RDMA/hns: Add check for the validity of sl configuration")
> Fixes: d6a3627e311c ("RDMA/hns: Optimize wqe buffer set flow for post send")
> Signed-off-by: Weihang Li <liweihang@huawei.com>
> ---
> drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
> index 7a0c1ab..15e1313 100644
> --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
> +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
> @@ -433,6 +433,10 @@ static int fill_ud_av(struct hns_roce_v2_ud_send_wqe *ud_sq_wqe,
> V2_UD_SEND_WQE_BYTE_36_TCLASS_S, ah->av.tclass);
> roce_set_field(ud_sq_wqe->byte_40, V2_UD_SEND_WQE_BYTE_40_FLOW_LABEL_M,
> V2_UD_SEND_WQE_BYTE_40_FLOW_LABEL_S, ah->av.flowlabel);
> +
> + if (WARN_ON(ah->av.sl > MAX_SERVICE_LEVEL))
> + return -EINVAL;
> +
> roce_set_field(ud_sq_wqe->byte_40, V2_UD_SEND_WQE_BYTE_40_SL_M,
> V2_UD_SEND_WQE_BYTE_40_SL_S, ah->av.sl);
>
> @@ -4609,12 +4613,8 @@ static int hns_roce_v2_set_path(struct ib_qp *ibqp,
> memset(qpc_mask->dgid, 0, sizeof(grh->dgid.raw));
>
> hr_qp->sl = rdma_ah_get_sl(&attr->ah_attr);
> - if (unlikely(hr_qp->sl > MAX_SERVICE_LEVEL)) {
> - ibdev_err(ibdev,
> - "failed to fill QPC, sl (%d) shouldn't be larger than %d.\n",
> - hr_qp->sl, MAX_SERVICE_LEVEL);
> + if (WARN_ON(hr_qp->sl > MAX_SERVICE_LEVEL))
> return -EINVAL;
> - }
>
> roce_set_field(context->byte_28_at_fl, V2_QPC_BYTE_28_SL_M,
> V2_QPC_BYTE_28_SL_S, hr_qp->sl);
Can any of these warn_on's be triggered by user space? That would not
be OK
Jason
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v2 for-next 05/11] RDMA/hns: WARN_ON if get a reserved sl from users
2020-12-09 21:09 ` Jason Gunthorpe
@ 2020-12-10 4:00 ` liweihang
2020-12-10 13:45 ` Jason Gunthorpe
0 siblings, 1 reply; 16+ messages in thread
From: liweihang @ 2020-12-10 4:00 UTC (permalink / raw)
To: Jason Gunthorpe; +Cc: dledford, leon, linux-rdma, Linuxarm
On 2020/12/10 5:09, Jason Gunthorpe wrote:
> On Fri, Dec 04, 2020 at 06:40:30PM +0800, Weihang Li wrote:
>> According to the RoCE v1 specification, the sl (service level) 0-7 are
>> mapped directly to priorities 0-7 respectively, sl 8-15 are reserved. The
>> driver should verify whether the value of sl is larger than 7, if so, an
>> exception should be returned.
>>
>> Fixes: 172505cfa3a8 ("RDMA/hns: Add check for the validity of sl configuration")
>> Fixes: d6a3627e311c ("RDMA/hns: Optimize wqe buffer set flow for post send")
>> Signed-off-by: Weihang Li <liweihang@huawei.com>
>> ---
>> drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 10 +++++-----
>> 1 file changed, 5 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
>> index 7a0c1ab..15e1313 100644
>> --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
>> +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
>> @@ -433,6 +433,10 @@ static int fill_ud_av(struct hns_roce_v2_ud_send_wqe *ud_sq_wqe,
>> V2_UD_SEND_WQE_BYTE_36_TCLASS_S, ah->av.tclass);
>> roce_set_field(ud_sq_wqe->byte_40, V2_UD_SEND_WQE_BYTE_40_FLOW_LABEL_M,
>> V2_UD_SEND_WQE_BYTE_40_FLOW_LABEL_S, ah->av.flowlabel);
>> +
>> + if (WARN_ON(ah->av.sl > MAX_SERVICE_LEVEL))
>> + return -EINVAL;
>> +
>> roce_set_field(ud_sq_wqe->byte_40, V2_UD_SEND_WQE_BYTE_40_SL_M,
>> V2_UD_SEND_WQE_BYTE_40_SL_S, ah->av.sl);
>>
>> @@ -4609,12 +4613,8 @@ static int hns_roce_v2_set_path(struct ib_qp *ibqp,
>> memset(qpc_mask->dgid, 0, sizeof(grh->dgid.raw));
>>
>> hr_qp->sl = rdma_ah_get_sl(&attr->ah_attr);
>> - if (unlikely(hr_qp->sl > MAX_SERVICE_LEVEL)) {
>> - ibdev_err(ibdev,
>> - "failed to fill QPC, sl (%d) shouldn't be larger than %d.\n",
>> - hr_qp->sl, MAX_SERVICE_LEVEL);
>> + if (WARN_ON(hr_qp->sl > MAX_SERVICE_LEVEL))
>> return -EINVAL;
>> - }
>>
>> roce_set_field(context->byte_28_at_fl, V2_QPC_BYTE_28_SL_M,
>> V2_QPC_BYTE_28_SL_S, hr_qp->sl);
>
> Can any of these warn_on's be triggered by user space? That would not
> be OK
>
> Jason
>
Hi Jason,
Thanks for your comments, I understand that error that can be triggered by
userspace shouldn't use WARN_ON(). So I shouldn't use WARN_ON() in
hns_roce_v2_set_path().
As for the error in process of post_send, you suggested me to warn_on if
a kernel user try to pass in an illegal opcode. So I guess I should use
WARN_ON() too in sl's check when filling a UD WQE. Am I right?
Weihang
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v2 for-next 05/11] RDMA/hns: WARN_ON if get a reserved sl from users
2020-12-10 4:00 ` liweihang
@ 2020-12-10 13:45 ` Jason Gunthorpe
2020-12-10 13:55 ` liweihang
0 siblings, 1 reply; 16+ messages in thread
From: Jason Gunthorpe @ 2020-12-10 13:45 UTC (permalink / raw)
To: liweihang; +Cc: dledford, leon, linux-rdma, Linuxarm
On Thu, Dec 10, 2020 at 04:00:16AM +0000, liweihang wrote:
> On 2020/12/10 5:09, Jason Gunthorpe wrote:
> > On Fri, Dec 04, 2020 at 06:40:30PM +0800, Weihang Li wrote:
> >> According to the RoCE v1 specification, the sl (service level) 0-7 are
> >> mapped directly to priorities 0-7 respectively, sl 8-15 are reserved. The
> >> driver should verify whether the value of sl is larger than 7, if so, an
> >> exception should be returned.
> >>
> >> Fixes: 172505cfa3a8 ("RDMA/hns: Add check for the validity of sl configuration")
> >> Fixes: d6a3627e311c ("RDMA/hns: Optimize wqe buffer set flow for post send")
> >> Signed-off-by: Weihang Li <liweihang@huawei.com>
> >> drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 10 +++++-----
> >> 1 file changed, 5 insertions(+), 5 deletions(-)
> >>
> >> diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
> >> index 7a0c1ab..15e1313 100644
> >> +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
> >> @@ -433,6 +433,10 @@ static int fill_ud_av(struct hns_roce_v2_ud_send_wqe *ud_sq_wqe,
> >> V2_UD_SEND_WQE_BYTE_36_TCLASS_S, ah->av.tclass);
> >> roce_set_field(ud_sq_wqe->byte_40, V2_UD_SEND_WQE_BYTE_40_FLOW_LABEL_M,
> >> V2_UD_SEND_WQE_BYTE_40_FLOW_LABEL_S, ah->av.flowlabel);
> >> +
> >> + if (WARN_ON(ah->av.sl > MAX_SERVICE_LEVEL))
> >> + return -EINVAL;
> >> +
> >> roce_set_field(ud_sq_wqe->byte_40, V2_UD_SEND_WQE_BYTE_40_SL_M,
> >> V2_UD_SEND_WQE_BYTE_40_SL_S, ah->av.sl);
> >>
> >> @@ -4609,12 +4613,8 @@ static int hns_roce_v2_set_path(struct ib_qp *ibqp,
> >> memset(qpc_mask->dgid, 0, sizeof(grh->dgid.raw));
> >>
> >> hr_qp->sl = rdma_ah_get_sl(&attr->ah_attr);
> >> - if (unlikely(hr_qp->sl > MAX_SERVICE_LEVEL)) {
> >> - ibdev_err(ibdev,
> >> - "failed to fill QPC, sl (%d) shouldn't be larger than %d.\n",
> >> - hr_qp->sl, MAX_SERVICE_LEVEL);
> >> + if (WARN_ON(hr_qp->sl > MAX_SERVICE_LEVEL))
> >> return -EINVAL;
> >> - }
> >>
> >> roce_set_field(context->byte_28_at_fl, V2_QPC_BYTE_28_SL_M,
> >> V2_QPC_BYTE_28_SL_S, hr_qp->sl);
> >
> > Can any of these warn_on's be triggered by user space? That would not
> > be OK
> >
> > Jason
> >
>
> Hi Jason,
>
> Thanks for your comments, I understand that error that can be triggered by
> userspace shouldn't use WARN_ON(). So I shouldn't use WARN_ON() in
> hns_roce_v2_set_path().
>
> As for the error in process of post_send, you suggested me to warn_on if
> a kernel user try to pass in an illegal opcode. So I guess I should use
> WARN_ON() too in sl's check when filling a UD WQE. Am I right?
Userspace should not be able to trigger warn_on
Bad kernel ULPs are OK to trigger warn_on
Jason
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v2 for-next 05/11] RDMA/hns: WARN_ON if get a reserved sl from users
2020-12-10 13:45 ` Jason Gunthorpe
@ 2020-12-10 13:55 ` liweihang
0 siblings, 0 replies; 16+ messages in thread
From: liweihang @ 2020-12-10 13:55 UTC (permalink / raw)
To: Jason Gunthorpe; +Cc: dledford, leon, linux-rdma, Linuxarm
On 2020/12/10 21:45, Jason Gunthorpe wrote:
> On Thu, Dec 10, 2020 at 04:00:16AM +0000, liweihang wrote:
>> On 2020/12/10 5:09, Jason Gunthorpe wrote:
>>> On Fri, Dec 04, 2020 at 06:40:30PM +0800, Weihang Li wrote:
>>>> According to the RoCE v1 specification, the sl (service level) 0-7 are
>>>> mapped directly to priorities 0-7 respectively, sl 8-15 are reserved. The
>>>> driver should verify whether the value of sl is larger than 7, if so, an
>>>> exception should be returned.
>>>>
>>>> Fixes: 172505cfa3a8 ("RDMA/hns: Add check for the validity of sl configuration")
>>>> Fixes: d6a3627e311c ("RDMA/hns: Optimize wqe buffer set flow for post send")
>>>> Signed-off-by: Weihang Li <liweihang@huawei.com>
>>>> drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 10 +++++-----
>>>> 1 file changed, 5 insertions(+), 5 deletions(-)
>>>>
>>>> diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
>>>> index 7a0c1ab..15e1313 100644
>>>> +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
>>>> @@ -433,6 +433,10 @@ static int fill_ud_av(struct hns_roce_v2_ud_send_wqe *ud_sq_wqe,
>>>> V2_UD_SEND_WQE_BYTE_36_TCLASS_S, ah->av.tclass);
>>>> roce_set_field(ud_sq_wqe->byte_40, V2_UD_SEND_WQE_BYTE_40_FLOW_LABEL_M,
>>>> V2_UD_SEND_WQE_BYTE_40_FLOW_LABEL_S, ah->av.flowlabel);
>>>> +
>>>> + if (WARN_ON(ah->av.sl > MAX_SERVICE_LEVEL))
>>>> + return -EINVAL;
>>>> +
>>>> roce_set_field(ud_sq_wqe->byte_40, V2_UD_SEND_WQE_BYTE_40_SL_M,
>>>> V2_UD_SEND_WQE_BYTE_40_SL_S, ah->av.sl);
>>>>
>>>> @@ -4609,12 +4613,8 @@ static int hns_roce_v2_set_path(struct ib_qp *ibqp,
>>>> memset(qpc_mask->dgid, 0, sizeof(grh->dgid.raw));
>>>>
>>>> hr_qp->sl = rdma_ah_get_sl(&attr->ah_attr);
>>>> - if (unlikely(hr_qp->sl > MAX_SERVICE_LEVEL)) {
>>>> - ibdev_err(ibdev,
>>>> - "failed to fill QPC, sl (%d) shouldn't be larger than %d.\n",
>>>> - hr_qp->sl, MAX_SERVICE_LEVEL);
>>>> + if (WARN_ON(hr_qp->sl > MAX_SERVICE_LEVEL))
>>>> return -EINVAL;
>>>> - }
>>>>
>>>> roce_set_field(context->byte_28_at_fl, V2_QPC_BYTE_28_SL_M,
>>>> V2_QPC_BYTE_28_SL_S, hr_qp->sl);
>>>
>>> Can any of these warn_on's be triggered by user space? That would not
>>> be OK
>>>
>>> Jason
>>>
>>
>> Hi Jason,
>>
>> Thanks for your comments, I understand that error that can be triggered by
>> userspace shouldn't use WARN_ON(). So I shouldn't use WARN_ON() in
>> hns_roce_v2_set_path().
>>
>> As for the error in process of post_send, you suggested me to warn_on if
>> a kernel user try to pass in an illegal opcode. So I guess I should use
>> WARN_ON() too in sl's check when filling a UD WQE. Am I right?
>
> Userspace should not be able to trigger warn_on
>
> Bad kernel ULPs are OK to trigger warn_on
>
> Jason
>
I see, Thank you.
Weihang
^ permalink raw reply [flat|nested] 16+ messages in thread
end of thread, other threads:[~2020-12-10 13:56 UTC | newest]
Thread overview: 16+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-12-04 10:40 [PATCH v2 for-next 00/11] RDMA/hns: Updates for 5.11 Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 01/11] RDMA/hns: Limit the length of data copied between kernel and userspace Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 02/11] RDMA/hns: Normalization the judgment of some features Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 03/11] RDMA/hns: Do shift on traffic class when using RoCEv2 Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 04/11] RDMA/hns: Avoid filling sl in high 3 bits of vlan_id Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 05/11] RDMA/hns: WARN_ON if get a reserved sl from users Weihang Li
2020-12-09 21:09 ` Jason Gunthorpe
2020-12-10 4:00 ` liweihang
2020-12-10 13:45 ` Jason Gunthorpe
2020-12-10 13:55 ` liweihang
2020-12-04 10:40 ` [PATCH v2 for-next 06/11] RDMA/hns: Remove unnecessary access right set during INIT2INIT Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 07/11] RDMA/hns: Fix coding style issues Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 08/11] RDMA/hns: Clear redundant variable initialization Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 09/11] RDMA/hns: Fix incorrect symbol types Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 10/11] RDMA/hns: Fix inaccurate prints Weihang Li
2020-12-04 10:40 ` [PATCH v2 for-next 11/11] RDMA/hns: Simplify AEQE process for different types of queue Weihang Li
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.