Unverified Commit fc2db2a2 authored by openeuler-ci-bot's avatar openeuler-ci-bot Committed by Gitee
Browse files

!2855 Support SW stats with debugfs 

Merge Pull Request from: @stinft 
 
Delete the code for querying hns RoCE SW stats with rdmatool from openEuler and add supporting SW stats with debugfs.

Juan Zhou (2):
   Revert "RDMA/hns: Fix missing dealloc_dfx_cnt() during device unregister"
   Revert "RDMA/hns: Add dfx cnt stats"
https://gitee.com/openeuler/kernel/issues/I8EX1J
 
Junxian Huang (4):
   RDMA/hns: Fix an inappropriate err code for unsupported operations
   RDMA/hns: Support SW stats with debugfs
   RDMA/hns: Remove return value checks of debugfs functions
   RDMA/hns: Don't set the HW stats ops for VF and HIP08
https://gitee.com/openeuler/kernel/issues/I8DJVZ
 
 
Link:https://gitee.com/openeuler/kernel/pulls/2855

 

Reviewed-by: default avatarChengchang Tang <tangchengchang@huawei.com>
Signed-off-by: default avatarJialin Zhang <zhangjialin11@huawei.com>
parents 289e92a8 34b53116

drivers/infiniband/hw/hns/hns_roce_cq.c

+2 −4
Original line number Diff line number Diff line
@@ -464,12 +464,10 @@ int hns_roce_create_cq(struct ib_cq *ib_cq, const struct ib_cq_init_attr *attr, 
	struct hns_roce_cq *hr_cq = to_hr_cq(ib_cq);

	struct ib_device *ibdev = &hr_dev->ib_dev;

	struct hns_roce_ib_create_cq ucmd = {};

	int ret;

	int ret = -EOPNOTSUPP;



	if (attr->flags) {

		ret = -EOPNOTSUPP;

	if (attr->flags)

		goto err_out;

	}



	ret = verify_cq_create_attr(hr_dev, attr);

	if (ret)

drivers/infiniband/hw/hns/hns_roce_debugfs.c

+71 −24
Original line number Diff line number Diff line
@@ -46,8 +46,6 @@ static void init_debugfs_seqfile(struct hns_debugfs_seqfile *seq, 


	entry = debugfs_create_file(name, 0400, parent, seq,

				    &hns_debugfs_seqfile_fops);

	if (IS_ERR(entry))

		return;



	seq->read = read_fn;

	seq->data = data;
@@ -85,11 +83,17 @@ struct hns_poe_debugfs { 
	struct dentry *root; /* dev debugfs entry */

};



struct hns_sw_stat_debugfs {

	struct dentry *root;

	struct hns_debugfs_seqfile sw_stat;

};



/* Debugfs for device */

struct hns_roce_dev_debugfs {

	struct dentry *root;

	struct hns_dca_debugfs *dca_root;

	struct hns_poe_debugfs *poe_root;

	struct hns_sw_stat_debugfs *sw_stat_root;

};



struct dca_mem_stats {
@@ -444,13 +448,8 @@ static void init_dca_ctx_debugfs(struct hns_dca_ctx_debugfs *dbgfs, 
{

	char name[DCA_CTX_PID_LEN];



	if (IS_ERR_OR_NULL(parent))

		return;



	dca_setup_pool_name(uctx ? uctx->pid : 0, !uctx, name, sizeof(name));

	dbgfs->root = debugfs_create_dir(name, parent);

	if (IS_ERR_OR_NULL(dbgfs->root))

		return;



	if (uctx) {

		init_debugfs_seqfile(&dbgfs->mem, "mstats", dbgfs->root,
@@ -477,18 +476,11 @@ create_dca_debugfs(struct hns_roce_dev *hr_dev, struct dentry *parent) 
{

	struct hns_dca_debugfs *dbgfs;



	if (IS_ERR(parent))

		return NULL;



	dbgfs = kzalloc(sizeof(*dbgfs), GFP_KERNEL);

	if (!dbgfs)

		return NULL;



	dbgfs->root = debugfs_create_dir("dca", parent);

	if (IS_ERR_OR_NULL(dbgfs->root)) {

		kfree(dbgfs);

		return NULL;

	}



	init_debugfs_seqfile(&dbgfs->pool, "pool", dbgfs->root,

			     dca_debugfs_pool_show, hr_dev);
@@ -634,24 +626,75 @@ void hns_roce_unregister_uctx_debugfs(struct hns_roce_dev *hr_dev, 
	}

}



static const char * const sw_stat_info[] = {

	[HNS_ROCE_DFX_AEQE_CNT] = "aeqe",

	[HNS_ROCE_DFX_CEQE_CNT] = "ceqe",

	[HNS_ROCE_DFX_CMDS_CNT] = "cmds",

	[HNS_ROCE_DFX_CMDS_ERR_CNT] = "cmds_err",

	[HNS_ROCE_DFX_MBX_POSTED_CNT] = "posted_mbx",

	[HNS_ROCE_DFX_MBX_POLLED_CNT] = "polled_mbx",

	[HNS_ROCE_DFX_MBX_EVENT_CNT] = "mbx_event",

	[HNS_ROCE_DFX_QP_CREATE_ERR_CNT] = "qp_create_err",

	[HNS_ROCE_DFX_QP_MODIFY_ERR_CNT] = "qp_modify_err",

	[HNS_ROCE_DFX_CQ_CREATE_ERR_CNT] = "cq_create_err",

	[HNS_ROCE_DFX_CQ_MODIFY_ERR_CNT] = "cq_modify_err",

	[HNS_ROCE_DFX_SRQ_CREATE_ERR_CNT] = "srq_create_err",

	[HNS_ROCE_DFX_SRQ_MODIFY_ERR_CNT] = "srq_modify_err",

	[HNS_ROCE_DFX_XRCD_ALLOC_ERR_CNT] = "xrcd_alloc_err",

	[HNS_ROCE_DFX_MR_REG_ERR_CNT] = "mr_reg_err",

	[HNS_ROCE_DFX_MR_REREG_ERR_CNT] = "mr_rereg_err",

	[HNS_ROCE_DFX_AH_CREATE_ERR_CNT] = "ah_create_err",

	[HNS_ROCE_DFX_MMAP_ERR_CNT] = "mmap_err",

	[HNS_ROCE_DFX_UCTX_ALLOC_ERR_CNT] = "uctx_alloc_err",

};



static int sw_stat_debugfs_show(struct seq_file *file, void *offset)

{

	struct hns_roce_dev *hr_dev = file->private;

	int i;



	for (i = 0; i < HNS_ROCE_DFX_CNT_TOTAL; i++)

		seq_printf(file, "%-20s --- %lld\n", sw_stat_info[i],

			   atomic64_read(&hr_dev->dfx_cnt[i]));



	return 0;

}



static struct hns_sw_stat_debugfs

	*create_sw_stat_debugfs(struct hns_roce_dev *hr_dev,

				struct dentry *parent)

{

	struct hns_sw_stat_debugfs *dbgfs;



	dbgfs = kvzalloc(sizeof(*dbgfs), GFP_KERNEL);

	if (!dbgfs)

		return NULL;



	dbgfs->root = debugfs_create_dir("sw_stat", parent);



	init_debugfs_seqfile(&dbgfs->sw_stat, "sw_stat", dbgfs->root,

			     sw_stat_debugfs_show, hr_dev);

	return dbgfs;

}



static void destroy_sw_stat_debugfs(struct hns_sw_stat_debugfs *sw_stat_dbgfs)

{

	cleanup_debugfs_seqfile(&sw_stat_dbgfs->sw_stat);

	debugfs_remove_recursive(sw_stat_dbgfs->root);

	kvfree(sw_stat_dbgfs);

}



/* debugfs for device */

void hns_roce_register_debugfs(struct hns_roce_dev *hr_dev)

{

	struct hns_roce_dev_debugfs *dbgfs;



	if (IS_ERR_OR_NULL(hns_roce_dbgfs_root))

		return;



	dbgfs = kzalloc(sizeof(*dbgfs), GFP_KERNEL);

	if (!dbgfs)

		return;



	dbgfs->root = debugfs_create_dir(dev_name(&hr_dev->ib_dev.dev),

					 hns_roce_dbgfs_root);

	if (IS_ERR(dbgfs->root)) {

		kfree(dbgfs);

		return;

	}



	if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_DCA_MODE)

		dbgfs->dca_root = create_dca_debugfs(hr_dev, dbgfs->root);
@@ -659,6 +702,8 @@ void hns_roce_register_debugfs(struct hns_roce_dev *hr_dev) 
	if (poe_is_supported(hr_dev))

		dbgfs->poe_root = create_poe_debugfs(hr_dev, dbgfs->root);



	dbgfs->sw_stat_root = create_sw_stat_debugfs(hr_dev, dbgfs->root);



	hr_dev->dbgfs = dbgfs;

}
@@ -666,9 +711,6 @@ void hns_roce_unregister_debugfs(struct hns_roce_dev *hr_dev) 
{

	struct hns_roce_dev_debugfs *dbgfs;



	if (IS_ERR_OR_NULL(hns_roce_dbgfs_root))

		return;



	dbgfs = hr_dev->dbgfs;

	if (!dbgfs)

		return;
@@ -683,6 +725,11 @@ void hns_roce_unregister_debugfs(struct hns_roce_dev *hr_dev) 
		dbgfs->poe_root = NULL;

	}



	if (dbgfs->sw_stat_root) {

		destroy_sw_stat_debugfs(dbgfs->sw_stat_root);

		dbgfs->sw_stat_root = NULL;

	}



	debugfs_remove_recursive(dbgfs->root);

	hr_dev->dbgfs = NULL;

	kfree(dbgfs);

drivers/infiniband/hw/hns/hns_roce_device.h

+3 −1
Original line number Diff line number Diff line
@@ -994,7 +994,7 @@ enum hns_roce_hw_pkt_stat_index { 
	HNS_ROCE_HW_CNT_TOTAL,

};



enum hns_roce_hw_dfx_stat_index {

enum hns_roce_sw_dfx_stat_index {

	HNS_ROCE_DFX_AEQE_CNT,

	HNS_ROCE_DFX_CEQE_CNT,

	HNS_ROCE_DFX_CMDS_CNT,
@@ -1005,7 +1005,9 @@ enum hns_roce_hw_dfx_stat_index { 
	HNS_ROCE_DFX_QP_CREATE_ERR_CNT,

	HNS_ROCE_DFX_QP_MODIFY_ERR_CNT,

	HNS_ROCE_DFX_CQ_CREATE_ERR_CNT,

	HNS_ROCE_DFX_CQ_MODIFY_ERR_CNT,

	HNS_ROCE_DFX_SRQ_CREATE_ERR_CNT,

	HNS_ROCE_DFX_SRQ_MODIFY_ERR_CNT,

	HNS_ROCE_DFX_XRCD_ALLOC_ERR_CNT,

	HNS_ROCE_DFX_MR_REG_ERR_CNT,

	HNS_ROCE_DFX_MR_REREG_ERR_CNT,

drivers/infiniband/hw/hns/hns_roce_hw_v2.c

+23 −12
Original line number Diff line number Diff line
@@ -6186,19 +6186,23 @@ static int hns_roce_v2_modify_srq(struct ib_srq *ibsrq, 
	struct hns_roce_srq_context *srq_context;

	struct hns_roce_srq_context *srqc_mask;

	struct hns_roce_cmd_mailbox *mailbox;

	int ret;

	int ret = -EOPNOTSUPP;



	/* Resizing SRQs is not supported yet */

	if (srq_attr_mask & IB_SRQ_MAX_WR)

		return -EINVAL;

		goto out;



	if (srq_attr_mask & IB_SRQ_LIMIT) {

		if (srq_attr->srq_limit > srq->wqe_cnt)

			return -EINVAL;

		if (srq_attr->srq_limit > srq->wqe_cnt) {

			ret = -EINVAL;

			goto out;

		}



		mailbox = hns_roce_alloc_cmd_mailbox(hr_dev);

		if (IS_ERR(mailbox))

			return PTR_ERR(mailbox);

		if (IS_ERR(mailbox)) {

			ret = PTR_ERR(mailbox);

			goto out;

		}



		srq_context = mailbox->buf;

		srqc_mask = (struct hns_roce_srq_context *)mailbox->buf + 1;
@@ -6211,15 +6215,17 @@ static int hns_roce_v2_modify_srq(struct ib_srq *ibsrq, 
		ret = hns_roce_cmd_mbox(hr_dev, mailbox->dma, 0,

					HNS_ROCE_CMD_MODIFY_SRQC, srq->srqn);

		hns_roce_free_cmd_mailbox(hr_dev, mailbox);

		if (ret) {

		if (ret)

			ibdev_err(&hr_dev->ib_dev,

				  "failed to handle cmd of modifying SRQ, ret = %d.\n",

				  ret);

			return ret;

		}

	}



	return 0;

out:

	if (ret)

		atomic64_inc(&hr_dev->dfx_cnt[HNS_ROCE_DFX_SRQ_MODIFY_ERR_CNT]);



	return ret;

}



static int hns_roce_v2_query_srq(struct ib_srq *ibsrq, struct ib_srq_attr *attr)
@@ -6263,8 +6269,9 @@ static int hns_roce_v2_modify_cq(struct ib_cq *cq, u16 cq_count, u16 cq_period) 
	int ret;



	mailbox = hns_roce_alloc_cmd_mailbox(hr_dev);

	if (IS_ERR(mailbox))

		return PTR_ERR(mailbox);

	ret = PTR_ERR_OR_ZERO(mailbox);

	if (ret)

		goto err_out;



	cq_context = mailbox->buf;

	cqc_mask = (struct hns_roce_v2_cq_context *)mailbox->buf + 1;
@@ -6294,6 +6301,10 @@ static int hns_roce_v2_modify_cq(struct ib_cq *cq, u16 cq_count, u16 cq_period) 
			  "failed to process cmd when modifying CQ, ret = %d.\n",

			  ret);



err_out:

	if (ret)

		atomic64_inc(&hr_dev->dfx_cnt[HNS_ROCE_DFX_CQ_MODIFY_ERR_CNT]);



	return ret;

}

drivers/infiniband/hw/hns/hns_roce_main.c

+21 −49
Original line number Diff line number Diff line
@@ -786,30 +786,10 @@ static void hns_roce_get_fw_ver(struct ib_device *device, char *str) 
		 sub_minor);

}



#define HNS_ROCE_DFX_STATS(ename, cname) \

	[HNS_ROCE_DFX_##ename##_CNT] = cname



#define HNS_ROCE_HW_CNT(ename, cname) \

	[HNS_ROCE_DFX_CNT_TOTAL + HNS_ROCE_HW_##ename##_CNT] = cname

	[HNS_ROCE_HW_##ename##_CNT] = cname



static const char *const hns_roce_port_stats_descs[] = {

	HNS_ROCE_DFX_STATS(AEQE, "aeqe"),

	HNS_ROCE_DFX_STATS(CEQE, "ceqe"),

	HNS_ROCE_DFX_STATS(CMDS, "cmds"),

	HNS_ROCE_DFX_STATS(CMDS_ERR, "cmds_err"),

	HNS_ROCE_DFX_STATS(MBX_POSTED, "posted_mbx"),

	HNS_ROCE_DFX_STATS(MBX_POLLED, "polled_mbx"),

	HNS_ROCE_DFX_STATS(MBX_EVENT, "mbx_event"),

	HNS_ROCE_DFX_STATS(QP_CREATE_ERR, "qp_create_err"),

	HNS_ROCE_DFX_STATS(QP_MODIFY_ERR, "qp_modify_err"),

	HNS_ROCE_DFX_STATS(CQ_CREATE_ERR, "cq_create_err"),

	HNS_ROCE_DFX_STATS(SRQ_CREATE_ERR, "srq_create_err"),

	HNS_ROCE_DFX_STATS(XRCD_ALLOC_ERR, "xrcd_alloc_err"),

	HNS_ROCE_DFX_STATS(MR_REG_ERR, "mr_reg_err"),

	HNS_ROCE_DFX_STATS(MR_REREG_ERR, "mr_rereg_err"),

	HNS_ROCE_DFX_STATS(AH_CREATE_ERR, "ah_create_err"),

	HNS_ROCE_DFX_STATS(MMAP_ERR, "mmap_err"),

	HNS_ROCE_DFX_STATS(UCTX_ALLOC_ERR, "uctx_alloc_err"),

	HNS_ROCE_HW_CNT(RX_RC_PKT, "rx_rc_pkt"),

	HNS_ROCE_HW_CNT(RX_UC_PKT, "rx_uc_pkt"),

	HNS_ROCE_HW_CNT(RX_UD_PKT, "rx_ud_pkt"),
@@ -838,21 +818,14 @@ static struct rdma_hw_stats *hns_roce_alloc_hw_port_stats(struct ib_device *devi 
							  u8 port_num)

{

	struct hns_roce_dev *hr_dev = to_hr_dev(device);

	int num_counters;



	if (port_num > hr_dev->caps.num_ports) {

		ibdev_err(device, "invalid port num.\n");

		return NULL;

	}



	if (hr_dev->pci_dev->revision >= PCI_REVISION_ID_HIP09 &&

	    !hr_dev->is_vf)

		num_counters = ARRAY_SIZE(hns_roce_port_stats_descs);

	else

		num_counters = HNS_ROCE_DFX_CNT_TOTAL;



	return rdma_alloc_hw_stats_struct(hns_roce_port_stats_descs,

					  num_counters,

					  ARRAY_SIZE(hns_roce_port_stats_descs),

					  RDMA_HW_STATS_DEFAULT_LIFESPAN);

}
@@ -861,9 +834,8 @@ static int hns_roce_get_hw_stats(struct ib_device *device, 
				 u8 port, int index)

{

	struct hns_roce_dev *hr_dev = to_hr_dev(device);

	int hw_counters = HNS_ROCE_HW_CNT_TOTAL;

	int num_counters = HNS_ROCE_HW_CNT_TOTAL;

	int ret;

	int i;



	if (port == 0)

		return 0;
@@ -871,24 +843,15 @@ static int hns_roce_get_hw_stats(struct ib_device *device, 
	if (port > hr_dev->caps.num_ports)

		return -EINVAL;



	for (i = 0; i < HNS_ROCE_DFX_CNT_TOTAL; i++)

		stats->value[i] = atomic64_read(&hr_dev->dfx_cnt[i]);



	if (hr_dev->pci_dev->revision <= PCI_REVISION_ID_HIP08 ||

	    hr_dev->is_vf)

		return HNS_ROCE_DFX_CNT_TOTAL;



	hw_counters = HNS_ROCE_HW_CNT_TOTAL;

	ret = hr_dev->hw->query_hw_counter(hr_dev,

					&stats->value[HNS_ROCE_DFX_CNT_TOTAL],

					port, &hw_counters);

	ret = hr_dev->hw->query_hw_counter(hr_dev, stats->value, port,

					   &num_counters);

	if (ret) {

		ibdev_err(device, "failed to query hw counter, ret = %d.\n",

			  ret);

		return ret;

	}



	return hw_counters + HNS_ROCE_DFX_CNT_TOTAL;

	return num_counters;

}



static void hns_roce_unregister_device(struct hns_roce_dev *hr_dev,
@@ -970,8 +933,6 @@ static const struct ib_device_ops hns_roce_dev_ops = { 
	.query_pkey = hns_roce_query_pkey,

	.query_port = hns_roce_query_port,

	.reg_user_mr = hns_roce_reg_user_mr,

	.alloc_hw_stats = hns_roce_alloc_hw_port_stats,

	.get_hw_stats = hns_roce_get_hw_stats,

	.init_port = hns_roce_create_port_files,



	INIT_RDMA_OBJ_SIZE(ib_ah, hns_roce_ah, ibah),
@@ -980,6 +941,11 @@ static const struct ib_device_ops hns_roce_dev_ops = { 
	INIT_RDMA_OBJ_SIZE(ib_ucontext, hns_roce_ucontext, ibucontext),

};



static const struct ib_device_ops hns_roce_dev_hw_stats_ops = {

	.alloc_hw_stats = hns_roce_alloc_hw_port_stats,

	.get_hw_stats = hns_roce_get_hw_stats,

};



static const struct ib_device_ops hns_roce_dev_mr_ops = {

	.rereg_user_mr = hns_roce_rereg_user_mr,

};
@@ -1102,6 +1068,10 @@ static int hns_roce_register_device(struct hns_roce_dev *hr_dev) 
		ib_set_device_ops(ib_dev, &hns_roce_dev_xrcd_ops);

	}



	if (hr_dev->pci_dev->revision >= PCI_REVISION_ID_HIP09 &&

	    !hr_dev->is_vf)

		ib_set_device_ops(ib_dev, &hns_roce_dev_hw_stats_ops);



	ib_set_device_ops(ib_dev, hr_dev->hw->hns_roce_dev_ops);

	ib_set_device_ops(ib_dev, &hns_roce_dev_ops);

	ib_set_device_ops(ib_dev, &hns_roce_dev_restrack_ops);
@@ -1459,7 +1429,7 @@ static void hns_roce_unregister_poe_ch(struct hns_roce_dev *hr_dev) 


static int hns_roce_alloc_dfx_cnt(struct hns_roce_dev *hr_dev)

{

	hr_dev->dfx_cnt = kcalloc(HNS_ROCE_DFX_CNT_TOTAL, sizeof(atomic64_t),

	hr_dev->dfx_cnt = kvcalloc(HNS_ROCE_DFX_CNT_TOTAL, sizeof(atomic64_t),

				   GFP_KERNEL);

	if (!hr_dev->dfx_cnt)

		return -ENOMEM;
@@ -1469,7 +1439,7 @@ static int hns_roce_alloc_dfx_cnt(struct hns_roce_dev *hr_dev) 


static void hns_roce_dealloc_dfx_cnt(struct hns_roce_dev *hr_dev)

{

	kfree(hr_dev->dfx_cnt);

	kvfree(hr_dev->dfx_cnt);

}



int hns_roce_init(struct hns_roce_dev *hr_dev)
@@ -1480,8 +1450,10 @@ int hns_roce_init(struct hns_roce_dev *hr_dev) 
	hr_dev->is_reset = false;



	ret = hns_roce_alloc_dfx_cnt(hr_dev);

	if (ret)

	if (ret) {

		dev_err(dev, "Alloc dfx_cnt failed!\n");

		return ret;

	}



	if (hr_dev->hw->cmq_init) {

		ret = hr_dev->hw->cmq_init(hr_dev);