* [net-next V3 01/13] devlink: Force enclosing array on binary fmsg data
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 02/13] net/mlx5: Add support for resource dump Saeed Mahameed
` (13 subsequent siblings)
14 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: netdev, Aya Levin, Jiri Pirko, Saeed Mahameed
From: Aya Levin <ayal@mellanox.com>
Add a new API for start/end binary array brackets [] to force array
around binary data as required from JSON. With this restriction, re-open
API to set binary fmsg data.
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
include/net/devlink.h | 5 +++
net/core/devlink.c | 94 +++++++++++++++++++++++++++++++++++++++----
2 files changed, 91 insertions(+), 8 deletions(-)
diff --git a/include/net/devlink.h b/include/net/devlink.h
index ce5cea428fdc..149c108be66f 100644
--- a/include/net/devlink.h
+++ b/include/net/devlink.h
@@ -981,12 +981,17 @@ int devlink_fmsg_pair_nest_end(struct devlink_fmsg *fmsg);
int devlink_fmsg_arr_pair_nest_start(struct devlink_fmsg *fmsg,
const char *name);
int devlink_fmsg_arr_pair_nest_end(struct devlink_fmsg *fmsg);
+int devlink_fmsg_binary_pair_nest_start(struct devlink_fmsg *fmsg,
+ const char *name);
+int devlink_fmsg_binary_pair_nest_end(struct devlink_fmsg *fmsg);
int devlink_fmsg_bool_put(struct devlink_fmsg *fmsg, bool value);
int devlink_fmsg_u8_put(struct devlink_fmsg *fmsg, u8 value);
int devlink_fmsg_u32_put(struct devlink_fmsg *fmsg, u32 value);
int devlink_fmsg_u64_put(struct devlink_fmsg *fmsg, u64 value);
int devlink_fmsg_string_put(struct devlink_fmsg *fmsg, const char *value);
+int devlink_fmsg_binary_put(struct devlink_fmsg *fmsg, const void *value,
+ u16 value_len);
int devlink_fmsg_bool_pair_put(struct devlink_fmsg *fmsg, const char *name,
bool value);
diff --git a/net/core/devlink.c b/net/core/devlink.c
index 549ee56b7a21..216bdd25ce39 100644
--- a/net/core/devlink.c
+++ b/net/core/devlink.c
@@ -4237,6 +4237,12 @@ struct devlink_fmsg_item {
struct devlink_fmsg {
struct list_head item_list;
+ bool putting_binary; /* This flag forces enclosing of binary data
+ * in an array brackets. It forces using
+ * of designated API:
+ * devlink_fmsg_binary_pair_nest_start()
+ * devlink_fmsg_binary_pair_nest_end()
+ */
};
static struct devlink_fmsg *devlink_fmsg_alloc(void)
@@ -4280,17 +4286,26 @@ static int devlink_fmsg_nest_common(struct devlink_fmsg *fmsg,
int devlink_fmsg_obj_nest_start(struct devlink_fmsg *fmsg)
{
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
return devlink_fmsg_nest_common(fmsg, DEVLINK_ATTR_FMSG_OBJ_NEST_START);
}
EXPORT_SYMBOL_GPL(devlink_fmsg_obj_nest_start);
static int devlink_fmsg_nest_end(struct devlink_fmsg *fmsg)
{
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
return devlink_fmsg_nest_common(fmsg, DEVLINK_ATTR_FMSG_NEST_END);
}
int devlink_fmsg_obj_nest_end(struct devlink_fmsg *fmsg)
{
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
return devlink_fmsg_nest_end(fmsg);
}
EXPORT_SYMBOL_GPL(devlink_fmsg_obj_nest_end);
@@ -4301,6 +4316,9 @@ static int devlink_fmsg_put_name(struct devlink_fmsg *fmsg, const char *name)
{
struct devlink_fmsg_item *item;
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
if (strlen(name) + 1 > DEVLINK_FMSG_MAX_SIZE)
return -EMSGSIZE;
@@ -4321,6 +4339,9 @@ int devlink_fmsg_pair_nest_start(struct devlink_fmsg *fmsg, const char *name)
{
int err;
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
err = devlink_fmsg_nest_common(fmsg, DEVLINK_ATTR_FMSG_PAIR_NEST_START);
if (err)
return err;
@@ -4335,6 +4356,9 @@ EXPORT_SYMBOL_GPL(devlink_fmsg_pair_nest_start);
int devlink_fmsg_pair_nest_end(struct devlink_fmsg *fmsg)
{
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
return devlink_fmsg_nest_end(fmsg);
}
EXPORT_SYMBOL_GPL(devlink_fmsg_pair_nest_end);
@@ -4344,6 +4368,9 @@ int devlink_fmsg_arr_pair_nest_start(struct devlink_fmsg *fmsg,
{
int err;
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
err = devlink_fmsg_pair_nest_start(fmsg, name);
if (err)
return err;
@@ -4360,6 +4387,9 @@ int devlink_fmsg_arr_pair_nest_end(struct devlink_fmsg *fmsg)
{
int err;
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
err = devlink_fmsg_nest_end(fmsg);
if (err)
return err;
@@ -4372,6 +4402,30 @@ int devlink_fmsg_arr_pair_nest_end(struct devlink_fmsg *fmsg)
}
EXPORT_SYMBOL_GPL(devlink_fmsg_arr_pair_nest_end);
+int devlink_fmsg_binary_pair_nest_start(struct devlink_fmsg *fmsg,
+ const char *name)
+{
+ int err;
+
+ err = devlink_fmsg_arr_pair_nest_start(fmsg, name);
+ if (err)
+ return err;
+
+ fmsg->putting_binary = true;
+ return err;
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_binary_pair_nest_start);
+
+int devlink_fmsg_binary_pair_nest_end(struct devlink_fmsg *fmsg)
+{
+ if (!fmsg->putting_binary)
+ return -EINVAL;
+
+ fmsg->putting_binary = false;
+ return devlink_fmsg_arr_pair_nest_end(fmsg);
+}
+EXPORT_SYMBOL_GPL(devlink_fmsg_binary_pair_nest_end);
+
static int devlink_fmsg_put_value(struct devlink_fmsg *fmsg,
const void *value, u16 value_len,
u8 value_nla_type)
@@ -4396,40 +4450,59 @@ static int devlink_fmsg_put_value(struct devlink_fmsg *fmsg,
int devlink_fmsg_bool_put(struct devlink_fmsg *fmsg, bool value)
{
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
return devlink_fmsg_put_value(fmsg, &value, sizeof(value), NLA_FLAG);
}
EXPORT_SYMBOL_GPL(devlink_fmsg_bool_put);
int devlink_fmsg_u8_put(struct devlink_fmsg *fmsg, u8 value)
{
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
return devlink_fmsg_put_value(fmsg, &value, sizeof(value), NLA_U8);
}
EXPORT_SYMBOL_GPL(devlink_fmsg_u8_put);
int devlink_fmsg_u32_put(struct devlink_fmsg *fmsg, u32 value)
{
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
return devlink_fmsg_put_value(fmsg, &value, sizeof(value), NLA_U32);
}
EXPORT_SYMBOL_GPL(devlink_fmsg_u32_put);
int devlink_fmsg_u64_put(struct devlink_fmsg *fmsg, u64 value)
{
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
return devlink_fmsg_put_value(fmsg, &value, sizeof(value), NLA_U64);
}
EXPORT_SYMBOL_GPL(devlink_fmsg_u64_put);
int devlink_fmsg_string_put(struct devlink_fmsg *fmsg, const char *value)
{
+ if (fmsg->putting_binary)
+ return -EINVAL;
+
return devlink_fmsg_put_value(fmsg, value, strlen(value) + 1,
NLA_NUL_STRING);
}
EXPORT_SYMBOL_GPL(devlink_fmsg_string_put);
-static int devlink_fmsg_binary_put(struct devlink_fmsg *fmsg, const void *value,
- u16 value_len)
+int devlink_fmsg_binary_put(struct devlink_fmsg *fmsg, const void *value,
+ u16 value_len)
{
+ if (!fmsg->putting_binary)
+ return -EINVAL;
+
return devlink_fmsg_put_value(fmsg, value, value_len, NLA_BINARY);
}
+EXPORT_SYMBOL_GPL(devlink_fmsg_binary_put);
int devlink_fmsg_bool_pair_put(struct devlink_fmsg *fmsg, const char *name,
bool value)
@@ -4540,10 +4613,11 @@ int devlink_fmsg_binary_pair_put(struct devlink_fmsg *fmsg, const char *name,
const void *value, u32 value_len)
{
u32 data_size;
+ int end_err;
u32 offset;
int err;
- err = devlink_fmsg_arr_pair_nest_start(fmsg, name);
+ err = devlink_fmsg_binary_pair_nest_start(fmsg, name);
if (err)
return err;
@@ -4553,14 +4627,18 @@ int devlink_fmsg_binary_pair_put(struct devlink_fmsg *fmsg, const char *name,
data_size = DEVLINK_FMSG_MAX_SIZE;
err = devlink_fmsg_binary_put(fmsg, value + offset, data_size);
if (err)
- return err;
+ break;
+ /* Exit from loop with a break (instead of
+ * return) to make sure putting_binary is turned off in
+ * devlink_fmsg_binary_pair_nest_end
+ */
}
- err = devlink_fmsg_arr_pair_nest_end(fmsg);
- if (err)
- return err;
+ end_err = devlink_fmsg_binary_pair_nest_end(fmsg);
+ if (end_err)
+ err = end_err;
- return 0;
+ return err;
}
EXPORT_SYMBOL_GPL(devlink_fmsg_binary_pair_put);
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* [net-next V3 02/13] net/mlx5: Add support for resource dump
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 01/13] devlink: Force enclosing array on binary fmsg data Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 03/13] net/mlx5e: Gather reporters APIs together Saeed Mahameed
` (12 subsequent siblings)
14 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: netdev, Aya Levin, Moshe Shemesh, Jiri Pirko, Saeed Mahameed
From: Aya Levin <ayal@mellanox.com>
On driver load:
- Initialize resource dump data structure and memory access tools (mkey
& pd).
- Read the resource dump's menu which contains the FW segment
identifier. Each record is identified by the segment name (ASCII).
During the driver's course of life, users (like reporters) may request
dumps per segment. The user should create a command providing the
segment identifier (SW enumeration) and command keys. In return, the
user receives a command context. In order to receive the dump, the user
should supply the command context and a memory (aligned to a PAGE) on
which the dump content will be written. Since the dump may be larger
than the given memory, the user may resubmit the command until received
an indication of end-of-dump. It is the user's responsibility to destroy
the command.
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/Makefile | 2 +-
.../mellanox/mlx5/core/diag/rsc_dump.c | 286 ++++++++++++++++++
.../mellanox/mlx5/core/diag/rsc_dump.h | 58 ++++
.../net/ethernet/mellanox/mlx5/core/main.c | 12 +
include/linux/mlx5/driver.h | 1 +
5 files changed, 358 insertions(+), 1 deletion(-)
create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/diag/rsc_dump.c
create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/diag/rsc_dump.h
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
index d3e06cec8317..e0bb8e12356e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile
+++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
@@ -16,7 +16,7 @@ mlx5_core-y := main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \
transobj.o vport.o sriov.o fs_cmd.o fs_core.o pci_irq.o \
fs_counters.o rl.o lag.o dev.o events.o wq.o lib/gid.o \
lib/devcom.o lib/pci_vsc.o lib/dm.o diag/fs_tracepoint.o \
- diag/fw_tracer.o diag/crdump.o devlink.o
+ diag/fw_tracer.o diag/crdump.o devlink.o diag/rsc_dump.o
#
# Netdev basic
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/rsc_dump.c b/drivers/net/ethernet/mellanox/mlx5/core/diag/rsc_dump.c
new file mode 100644
index 000000000000..17ab7efe693d
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/rsc_dump.c
@@ -0,0 +1,286 @@
+// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
+/* Copyright (c) 2019 Mellanox Technologies. */
+
+#include "rsc_dump.h"
+#include "lib/mlx5.h"
+
+#define MLX5_SGMT_TYPE(SGMT) MLX5_SGMT_TYPE_##SGMT
+#define MLX5_SGMT_STR_ASSING(SGMT)[MLX5_SGMT_TYPE(SGMT)] = #SGMT
+static const char *const mlx5_rsc_sgmt_name[] = {
+ MLX5_SGMT_STR_ASSING(HW_CQPC),
+ MLX5_SGMT_STR_ASSING(HW_SQPC),
+ MLX5_SGMT_STR_ASSING(HW_RQPC),
+ MLX5_SGMT_STR_ASSING(FULL_SRQC),
+ MLX5_SGMT_STR_ASSING(FULL_CQC),
+ MLX5_SGMT_STR_ASSING(FULL_EQC),
+ MLX5_SGMT_STR_ASSING(FULL_QPC),
+ MLX5_SGMT_STR_ASSING(SND_BUFF),
+ MLX5_SGMT_STR_ASSING(RCV_BUFF),
+ MLX5_SGMT_STR_ASSING(SRQ_BUFF),
+ MLX5_SGMT_STR_ASSING(CQ_BUFF),
+ MLX5_SGMT_STR_ASSING(EQ_BUFF),
+ MLX5_SGMT_STR_ASSING(SX_SLICE),
+ MLX5_SGMT_STR_ASSING(SX_SLICE_ALL),
+ MLX5_SGMT_STR_ASSING(RDB),
+ MLX5_SGMT_STR_ASSING(RX_SLICE_ALL),
+};
+
+struct mlx5_rsc_dump {
+ u32 pdn;
+ struct mlx5_core_mkey mkey;
+ u16 fw_segment_type[MLX5_SGMT_TYPE_NUM];
+};
+
+struct mlx5_rsc_dump_cmd {
+ u64 mem_size;
+ u8 cmd[MLX5_ST_SZ_BYTES(resource_dump)];
+};
+
+static int mlx5_rsc_dump_sgmt_get_by_name(char *name)
+{
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(mlx5_rsc_sgmt_name); i++)
+ if (!strcmp(name, mlx5_rsc_sgmt_name[i]))
+ return i;
+
+ return -EINVAL;
+}
+
+static void mlx5_rsc_dump_read_menu_sgmt(struct mlx5_rsc_dump *rsc_dump, struct page *page)
+{
+ void *data = page_address(page);
+ enum mlx5_sgmt_type sgmt_idx;
+ int num_of_items;
+ char *sgmt_name;
+ void *member;
+ void *menu;
+ int i;
+
+ menu = MLX5_ADDR_OF(menu_resource_dump_response, data, menu);
+ num_of_items = MLX5_GET(resource_dump_menu_segment, menu, num_of_records);
+
+ for (i = 0; i < num_of_items; i++) {
+ member = MLX5_ADDR_OF(resource_dump_menu_segment, menu, record[i]);
+ sgmt_name = MLX5_ADDR_OF(resource_dump_menu_record, member, segment_name);
+ sgmt_idx = mlx5_rsc_dump_sgmt_get_by_name(sgmt_name);
+ if (sgmt_idx == -EINVAL)
+ continue;
+ rsc_dump->fw_segment_type[sgmt_idx] = MLX5_GET(resource_dump_menu_record,
+ member, segment_type);
+ }
+}
+
+static int mlx5_rsc_dump_trigger(struct mlx5_core_dev *dev, struct mlx5_rsc_dump_cmd *cmd,
+ struct page *page)
+{
+ struct mlx5_rsc_dump *rsc_dump = dev->rsc_dump;
+ struct device *ddev = &dev->pdev->dev;
+ u32 out_seq_num;
+ u32 in_seq_num;
+ dma_addr_t dma;
+ int err;
+
+ dma = dma_map_page(ddev, page, 0, cmd->mem_size, DMA_FROM_DEVICE);
+ if (unlikely(dma_mapping_error(ddev, dma)))
+ return -ENOMEM;
+
+ in_seq_num = MLX5_GET(resource_dump, cmd->cmd, seq_num);
+ MLX5_SET(resource_dump, cmd->cmd, mkey, rsc_dump->mkey.key);
+ MLX5_SET64(resource_dump, cmd->cmd, address, dma);
+
+ err = mlx5_core_access_reg(dev, cmd->cmd, sizeof(cmd->cmd), cmd->cmd,
+ sizeof(cmd->cmd), MLX5_REG_RESOURCE_DUMP, 0, 1);
+ if (err) {
+ mlx5_core_err(dev, "Resource dump: Failed to access err %d\n", err);
+ goto out;
+ }
+ out_seq_num = MLX5_GET(resource_dump, cmd->cmd, seq_num);
+ if (out_seq_num && (in_seq_num + 1 != out_seq_num))
+ err = -EIO;
+out:
+ dma_unmap_page(ddev, dma, cmd->mem_size, DMA_FROM_DEVICE);
+ return err;
+}
+
+struct mlx5_rsc_dump_cmd *mlx5_rsc_dump_cmd_create(struct mlx5_core_dev *dev,
+ struct mlx5_rsc_key *key)
+{
+ struct mlx5_rsc_dump_cmd *cmd;
+ int sgmt_type;
+
+ if (IS_ERR_OR_NULL(dev->rsc_dump))
+ return ERR_PTR(-EOPNOTSUPP);
+
+ sgmt_type = dev->rsc_dump->fw_segment_type[key->rsc];
+ if (!sgmt_type && key->rsc != MLX5_SGMT_TYPE_MENU)
+ return ERR_PTR(-EOPNOTSUPP);
+
+ cmd = kzalloc(sizeof(*cmd), GFP_KERNEL);
+ if (!cmd) {
+ mlx5_core_err(dev, "Resource dump: Failed to allocate command\n");
+ return ERR_PTR(-ENOMEM);
+ }
+ MLX5_SET(resource_dump, cmd->cmd, segment_type, sgmt_type);
+ MLX5_SET(resource_dump, cmd->cmd, index1, key->index1);
+ MLX5_SET(resource_dump, cmd->cmd, index2, key->index2);
+ MLX5_SET(resource_dump, cmd->cmd, num_of_obj1, key->num_of_obj1);
+ MLX5_SET(resource_dump, cmd->cmd, num_of_obj2, key->num_of_obj2);
+ MLX5_SET(resource_dump, cmd->cmd, size, key->size);
+ cmd->mem_size = key->size;
+ return cmd;
+}
+
+void mlx5_rsc_dump_cmd_destroy(struct mlx5_rsc_dump_cmd *cmd)
+{
+ kfree(cmd);
+}
+
+int mlx5_rsc_dump_next(struct mlx5_core_dev *dev, struct mlx5_rsc_dump_cmd *cmd,
+ struct page *page, int *size)
+{
+ bool more_dump;
+ int err;
+
+ if (IS_ERR_OR_NULL(dev->rsc_dump))
+ return -EOPNOTSUPP;
+
+ err = mlx5_rsc_dump_trigger(dev, cmd, page);
+ if (err) {
+ mlx5_core_err(dev, "Resource dump: Failed to trigger dump, %d\n", err);
+ return err;
+ }
+ *size = MLX5_GET(resource_dump, cmd->cmd, size);
+ more_dump = MLX5_GET(resource_dump, cmd->cmd, more_dump);
+
+ return more_dump;
+}
+
+#define MLX5_RSC_DUMP_MENU_SEGMENT 0xffff
+static int mlx5_rsc_dump_menu(struct mlx5_core_dev *dev)
+{
+ struct mlx5_rsc_dump_cmd *cmd = NULL;
+ struct mlx5_rsc_key key = {};
+ struct page *page;
+ int size;
+ int err;
+
+ page = alloc_page(GFP_KERNEL);
+ if (!page)
+ return -ENOMEM;
+
+ key.rsc = MLX5_SGMT_TYPE_MENU;
+ key.size = PAGE_SIZE;
+ cmd = mlx5_rsc_dump_cmd_create(dev, &key);
+ if (IS_ERR(cmd)) {
+ err = PTR_ERR(cmd);
+ goto free_page;
+ }
+ MLX5_SET(resource_dump, cmd->cmd, segment_type, MLX5_RSC_DUMP_MENU_SEGMENT);
+
+ do {
+ err = mlx5_rsc_dump_next(dev, cmd, page, &size);
+ if (err < 0)
+ goto destroy_cmd;
+
+ mlx5_rsc_dump_read_menu_sgmt(dev->rsc_dump, page);
+
+ } while (err > 0);
+
+destroy_cmd:
+ mlx5_rsc_dump_cmd_destroy(cmd);
+free_page:
+ __free_page(page);
+
+ return err;
+}
+
+static int mlx5_rsc_dump_create_mkey(struct mlx5_core_dev *mdev, u32 pdn,
+ struct mlx5_core_mkey *mkey)
+{
+ int inlen = MLX5_ST_SZ_BYTES(create_mkey_in);
+ void *mkc;
+ u32 *in;
+ int err;
+
+ in = kvzalloc(inlen, GFP_KERNEL);
+ if (!in)
+ return -ENOMEM;
+
+ mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry);
+ MLX5_SET(mkc, mkc, access_mode_1_0, MLX5_MKC_ACCESS_MODE_PA);
+ MLX5_SET(mkc, mkc, lw, 1);
+ MLX5_SET(mkc, mkc, lr, 1);
+
+ MLX5_SET(mkc, mkc, pd, pdn);
+ MLX5_SET(mkc, mkc, length64, 1);
+ MLX5_SET(mkc, mkc, qpn, 0xffffff);
+
+ err = mlx5_core_create_mkey(mdev, mkey, in, inlen);
+
+ kvfree(in);
+ return err;
+}
+
+struct mlx5_rsc_dump *mlx5_rsc_dump_create(struct mlx5_core_dev *dev)
+{
+ struct mlx5_rsc_dump *rsc_dump;
+
+ if (!MLX5_CAP_DEBUG(dev, resource_dump)) {
+ mlx5_core_dbg(dev, "Resource dump: capability not present\n");
+ return NULL;
+ }
+ rsc_dump = kzalloc(sizeof(*rsc_dump), GFP_KERNEL);
+ if (!rsc_dump)
+ return ERR_PTR(-ENOMEM);
+
+ return rsc_dump;
+}
+
+void mlx5_rsc_dump_destroy(struct mlx5_core_dev *dev)
+{
+ if (IS_ERR_OR_NULL(dev->rsc_dump))
+ return;
+ kfree(dev->rsc_dump);
+}
+
+int mlx5_rsc_dump_init(struct mlx5_core_dev *dev)
+{
+ struct mlx5_rsc_dump *rsc_dump = dev->rsc_dump;
+ int err;
+
+ if (IS_ERR_OR_NULL(dev->rsc_dump))
+ return 0;
+
+ err = mlx5_core_alloc_pd(dev, &rsc_dump->pdn);
+ if (err) {
+ mlx5_core_warn(dev, "Resource dump: Failed to allocate PD %d\n", err);
+ return err;
+ }
+ err = mlx5_rsc_dump_create_mkey(dev, rsc_dump->pdn, &rsc_dump->mkey);
+ if (err) {
+ mlx5_core_err(dev, "Resource dump: Failed to create mkey, %d\n", err);
+ goto free_pd;
+ }
+ err = mlx5_rsc_dump_menu(dev);
+ if (err) {
+ mlx5_core_err(dev, "Resource dump: Failed to read menu, %d\n", err);
+ goto destroy_mkey;
+ }
+ return err;
+
+destroy_mkey:
+ mlx5_core_destroy_mkey(dev, &rsc_dump->mkey);
+free_pd:
+ mlx5_core_dealloc_pd(dev, rsc_dump->pdn);
+ return err;
+}
+
+void mlx5_rsc_dump_cleanup(struct mlx5_core_dev *dev)
+{
+ if (IS_ERR_OR_NULL(dev->rsc_dump))
+ return;
+
+ mlx5_core_destroy_mkey(dev, &dev->rsc_dump->mkey);
+ mlx5_core_dealloc_pd(dev, dev->rsc_dump->pdn);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/diag/rsc_dump.h b/drivers/net/ethernet/mellanox/mlx5/core/diag/rsc_dump.h
new file mode 100644
index 000000000000..3b7573461a45
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/diag/rsc_dump.h
@@ -0,0 +1,58 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2019 Mellanox Technologies. */
+
+#ifndef __MLX5_RSC_DUMP_H
+#define __MLX5_RSC_DUMP__H
+
+#include <linux/mlx5/driver.h>
+#include "mlx5_core.h"
+
+enum mlx5_sgmt_type {
+ MLX5_SGMT_TYPE_HW_CQPC,
+ MLX5_SGMT_TYPE_HW_SQPC,
+ MLX5_SGMT_TYPE_HW_RQPC,
+ MLX5_SGMT_TYPE_FULL_SRQC,
+ MLX5_SGMT_TYPE_FULL_CQC,
+ MLX5_SGMT_TYPE_FULL_EQC,
+ MLX5_SGMT_TYPE_FULL_QPC,
+ MLX5_SGMT_TYPE_SND_BUFF,
+ MLX5_SGMT_TYPE_RCV_BUFF,
+ MLX5_SGMT_TYPE_SRQ_BUFF,
+ MLX5_SGMT_TYPE_CQ_BUFF,
+ MLX5_SGMT_TYPE_EQ_BUFF,
+ MLX5_SGMT_TYPE_SX_SLICE,
+ MLX5_SGMT_TYPE_SX_SLICE_ALL,
+ MLX5_SGMT_TYPE_RDB,
+ MLX5_SGMT_TYPE_RX_SLICE_ALL,
+ MLX5_SGMT_TYPE_MENU,
+ MLX5_SGMT_TYPE_TERMINATE,
+
+ MLX5_SGMT_TYPE_NUM, /* Keep last */
+};
+
+struct mlx5_rsc_key {
+ enum mlx5_sgmt_type rsc;
+ int index1;
+ int index2;
+ int num_of_obj1;
+ int num_of_obj2;
+ int size;
+};
+
+#define MLX5_RSC_DUMP_ALL 0xFFFF
+struct mlx5_rsc_dump_cmd;
+struct mlx5_rsc_dump;
+
+struct mlx5_rsc_dump *mlx5_rsc_dump_create(struct mlx5_core_dev *dev);
+void mlx5_rsc_dump_destroy(struct mlx5_core_dev *dev);
+
+int mlx5_rsc_dump_init(struct mlx5_core_dev *dev);
+void mlx5_rsc_dump_cleanup(struct mlx5_core_dev *dev);
+
+struct mlx5_rsc_dump_cmd *mlx5_rsc_dump_cmd_create(struct mlx5_core_dev *dev,
+ struct mlx5_rsc_key *key);
+void mlx5_rsc_dump_cmd_destroy(struct mlx5_rsc_dump_cmd *cmd);
+
+int mlx5_rsc_dump_next(struct mlx5_core_dev *dev, struct mlx5_rsc_dump_cmd *cmd,
+ struct page *page, int *size);
+#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index f554cfddcf4e..204a26bf0a5f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -70,6 +70,7 @@
#include "diag/fw_tracer.h"
#include "ecpf.h"
#include "lib/hv_vhca.h"
+#include "diag/rsc_dump.h"
MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
MODULE_DESCRIPTION("Mellanox 5th generation network adapters (ConnectX series) core driver");
@@ -880,6 +881,7 @@ static int mlx5_init_once(struct mlx5_core_dev *dev)
dev->tracer = mlx5_fw_tracer_create(dev);
dev->hv_vhca = mlx5_hv_vhca_create(dev);
+ dev->rsc_dump = mlx5_rsc_dump_create(dev);
return 0;
@@ -909,6 +911,7 @@ static int mlx5_init_once(struct mlx5_core_dev *dev)
static void mlx5_cleanup_once(struct mlx5_core_dev *dev)
{
+ mlx5_rsc_dump_destroy(dev);
mlx5_hv_vhca_destroy(dev->hv_vhca);
mlx5_fw_tracer_destroy(dev->tracer);
mlx5_dm_cleanup(dev);
@@ -1079,6 +1082,12 @@ static int mlx5_load(struct mlx5_core_dev *dev)
mlx5_hv_vhca_init(dev->hv_vhca);
+ err = mlx5_rsc_dump_init(dev);
+ if (err) {
+ mlx5_core_err(dev, "Failed to init Resource dump\n");
+ goto err_rsc_dump;
+ }
+
err = mlx5_fpga_device_start(dev);
if (err) {
mlx5_core_err(dev, "fpga device start failed %d\n", err);
@@ -1134,6 +1143,8 @@ static int mlx5_load(struct mlx5_core_dev *dev)
err_ipsec_start:
mlx5_fpga_device_stop(dev);
err_fpga_start:
+ mlx5_rsc_dump_cleanup(dev);
+err_rsc_dump:
mlx5_hv_vhca_cleanup(dev->hv_vhca);
mlx5_fw_tracer_cleanup(dev->tracer);
err_fw_tracer:
@@ -1155,6 +1166,7 @@ static void mlx5_unload(struct mlx5_core_dev *dev)
mlx5_accel_ipsec_cleanup(dev);
mlx5_accel_tls_cleanup(dev);
mlx5_fpga_device_stop(dev);
+ mlx5_rsc_dump_cleanup(dev);
mlx5_hv_vhca_cleanup(dev->hv_vhca);
mlx5_fw_tracer_cleanup(dev->tracer);
mlx5_eq_table_destroy(dev);
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index 277a51d3ec40..f99cbe249425 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -722,6 +722,7 @@ struct mlx5_core_dev {
struct mlx5_clock clock;
struct mlx5_ib_clock_info *clock_info;
struct mlx5_fw_tracer *tracer;
+ struct mlx5_rsc_dump *rsc_dump;
u32 vsc_addr;
struct mlx5_hv_vhca *hv_vhca;
};
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* [net-next V3 03/13] net/mlx5e: Gather reporters APIs together
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 01/13] devlink: Force enclosing array on binary fmsg data Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 02/13] net/mlx5: Add support for resource dump Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 04/13] net/mlx5e: Support dump callback in TX reporter Saeed Mahameed
` (11 subsequent siblings)
14 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: netdev, Aya Levin, Moshe Shemesh, Jiri Pirko, Saeed Mahameed
From: Aya Levin <ayal@mellanox.com>
Assemble all the API's to ease insertion of dump callbacks in the
following patches in the set.
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../mellanox/mlx5/core/en/reporter_rx.c | 82 +++++++++----------
.../mellanox/mlx5/core/en/reporter_tx.c | 58 ++++++-------
2 files changed, 70 insertions(+), 70 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c
index 6c72b592315b..cfa6941fca6b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c
@@ -102,19 +102,6 @@ static int mlx5e_rx_reporter_err_icosq_cqe_recover(void *ctx)
return err;
}
-void mlx5e_reporter_icosq_cqe_err(struct mlx5e_icosq *icosq)
-{
- struct mlx5e_priv *priv = icosq->channel->priv;
- char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN];
- struct mlx5e_err_ctx err_ctx = {};
-
- err_ctx.ctx = icosq;
- err_ctx.recover = mlx5e_rx_reporter_err_icosq_cqe_recover;
- sprintf(err_str, "ERR CQE on ICOSQ: 0x%x", icosq->sqn);
-
- mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx);
-}
-
static int mlx5e_rq_to_ready(struct mlx5e_rq *rq, int curr_state)
{
struct net_device *dev = rq->netdev;
@@ -171,19 +158,6 @@ static int mlx5e_rx_reporter_err_rq_cqe_recover(void *ctx)
return err;
}
-void mlx5e_reporter_rq_cqe_err(struct mlx5e_rq *rq)
-{
- struct mlx5e_priv *priv = rq->channel->priv;
- char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN];
- struct mlx5e_err_ctx err_ctx = {};
-
- err_ctx.ctx = rq;
- err_ctx.recover = mlx5e_rx_reporter_err_rq_cqe_recover;
- sprintf(err_str, "ERR CQE on RQ: 0x%x", rq->rqn);
-
- mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx);
-}
-
static int mlx5e_rx_reporter_timeout_recover(void *ctx)
{
struct mlx5e_icosq *icosq;
@@ -201,21 +175,6 @@ static int mlx5e_rx_reporter_timeout_recover(void *ctx)
return err;
}
-void mlx5e_reporter_rx_timeout(struct mlx5e_rq *rq)
-{
- struct mlx5e_icosq *icosq = &rq->channel->icosq;
- struct mlx5e_priv *priv = rq->channel->priv;
- char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN];
- struct mlx5e_err_ctx err_ctx = {};
-
- err_ctx.ctx = rq;
- err_ctx.recover = mlx5e_rx_reporter_timeout_recover;
- sprintf(err_str, "RX timeout on channel: %d, ICOSQ: 0x%x RQ: 0x%x, CQ: 0x%x\n",
- icosq->channel->ix, icosq->sqn, rq->rqn, rq->cq.mcq.cqn);
-
- mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx);
-}
-
static int mlx5e_rx_reporter_recover_from_ctx(struct mlx5e_err_ctx *err_ctx)
{
return err_ctx->recover(err_ctx->ctx);
@@ -371,6 +330,47 @@ static int mlx5e_rx_reporter_diagnose(struct devlink_health_reporter *reporter,
return err;
}
+void mlx5e_reporter_rx_timeout(struct mlx5e_rq *rq)
+{
+ struct mlx5e_icosq *icosq = &rq->channel->icosq;
+ struct mlx5e_priv *priv = rq->channel->priv;
+ char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN];
+ struct mlx5e_err_ctx err_ctx = {};
+
+ err_ctx.ctx = rq;
+ err_ctx.recover = mlx5e_rx_reporter_timeout_recover;
+ sprintf(err_str, "RX timeout on channel: %d, ICOSQ: 0x%x RQ: 0x%x, CQ: 0x%x\n",
+ icosq->channel->ix, icosq->sqn, rq->rqn, rq->cq.mcq.cqn);
+
+ mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx);
+}
+
+void mlx5e_reporter_rq_cqe_err(struct mlx5e_rq *rq)
+{
+ struct mlx5e_priv *priv = rq->channel->priv;
+ char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN];
+ struct mlx5e_err_ctx err_ctx = {};
+
+ err_ctx.ctx = rq;
+ err_ctx.recover = mlx5e_rx_reporter_err_rq_cqe_recover;
+ sprintf(err_str, "ERR CQE on RQ: 0x%x", rq->rqn);
+
+ mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx);
+}
+
+void mlx5e_reporter_icosq_cqe_err(struct mlx5e_icosq *icosq)
+{
+ struct mlx5e_priv *priv = icosq->channel->priv;
+ char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN];
+ struct mlx5e_err_ctx err_ctx = {};
+
+ err_ctx.ctx = icosq;
+ err_ctx.recover = mlx5e_rx_reporter_err_icosq_cqe_recover;
+ sprintf(err_str, "ERR CQE on ICOSQ: 0x%x", icosq->sqn);
+
+ mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx);
+}
+
static const struct devlink_health_reporter_ops mlx5_rx_reporter_ops = {
.name = "rx",
.recover = mlx5e_rx_reporter_recover,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
index b468549e96ff..623c949db54c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
@@ -82,19 +82,6 @@ static int mlx5e_tx_reporter_err_cqe_recover(void *ctx)
return err;
}
-void mlx5e_reporter_tx_err_cqe(struct mlx5e_txqsq *sq)
-{
- struct mlx5e_priv *priv = sq->channel->priv;
- char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN];
- struct mlx5e_err_ctx err_ctx = {0};
-
- err_ctx.ctx = sq;
- err_ctx.recover = mlx5e_tx_reporter_err_cqe_recover;
- sprintf(err_str, "ERR CQE on SQ: 0x%x", sq->sqn);
-
- mlx5e_health_report(priv, priv->tx_reporter, err_str, &err_ctx);
-}
-
static int mlx5e_tx_reporter_timeout_recover(void *ctx)
{
struct mlx5_eq_comp *eq;
@@ -110,22 +97,6 @@ static int mlx5e_tx_reporter_timeout_recover(void *ctx)
return err;
}
-int mlx5e_reporter_tx_timeout(struct mlx5e_txqsq *sq)
-{
- struct mlx5e_priv *priv = sq->channel->priv;
- char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN];
- struct mlx5e_err_ctx err_ctx;
-
- err_ctx.ctx = sq;
- err_ctx.recover = mlx5e_tx_reporter_timeout_recover;
- sprintf(err_str,
- "TX timeout on queue: %d, SQ: 0x%x, CQ: 0x%x, SQ Cons: 0x%x SQ Prod: 0x%x, usecs since last trans: %u\n",
- sq->channel->ix, sq->sqn, sq->cq.mcq.cqn, sq->cc, sq->pc,
- jiffies_to_usecs(jiffies - sq->txq->trans_start));
-
- return mlx5e_health_report(priv, priv->tx_reporter, err_str, &err_ctx);
-}
-
/* state lock cannot be grabbed within this function.
* It can cause a dead lock or a read-after-free.
*/
@@ -275,6 +246,35 @@ static int mlx5e_tx_reporter_diagnose(struct devlink_health_reporter *reporter,
return err;
}
+void mlx5e_reporter_tx_err_cqe(struct mlx5e_txqsq *sq)
+{
+ struct mlx5e_priv *priv = sq->channel->priv;
+ char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN];
+ struct mlx5e_err_ctx err_ctx = {};
+
+ err_ctx.ctx = sq;
+ err_ctx.recover = mlx5e_tx_reporter_err_cqe_recover;
+ sprintf(err_str, "ERR CQE on SQ: 0x%x", sq->sqn);
+
+ mlx5e_health_report(priv, priv->tx_reporter, err_str, &err_ctx);
+}
+
+int mlx5e_reporter_tx_timeout(struct mlx5e_txqsq *sq)
+{
+ struct mlx5e_priv *priv = sq->channel->priv;
+ char err_str[MLX5E_REPORTER_PER_Q_MAX_LEN];
+ struct mlx5e_err_ctx err_ctx = {};
+
+ err_ctx.ctx = sq;
+ err_ctx.recover = mlx5e_tx_reporter_timeout_recover;
+ sprintf(err_str,
+ "TX timeout on queue: %d, SQ: 0x%x, CQ: 0x%x, SQ Cons: 0x%x SQ Prod: 0x%x, usecs since last trans: %u\n",
+ sq->channel->ix, sq->sqn, sq->cq.mcq.cqn, sq->cc, sq->pc,
+ jiffies_to_usecs(jiffies - sq->txq->trans_start));
+
+ return mlx5e_health_report(priv, priv->tx_reporter, err_str, &err_ctx);
+}
+
static const struct devlink_health_reporter_ops mlx5_tx_reporter_ops = {
.name = "tx",
.recover = mlx5e_tx_reporter_recover,
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* [net-next V3 04/13] net/mlx5e: Support dump callback in TX reporter
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
` (2 preceding siblings ...)
2020-02-11 22:32 ` [net-next V3 03/13] net/mlx5e: Gather reporters APIs together Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 05/13] net/mlx5e: Support dump callback in RX reporter Saeed Mahameed
` (10 subsequent siblings)
14 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: netdev, Aya Levin, Moshe Shemesh, Jiri Pirko, Saeed Mahameed
From: Aya Levin <ayal@mellanox.com>
Add support for SQ's FW dump on TX reporter's events. Use Resource dump
API to retrieve the relevant data: SX slice, SQ dump and SQ buffer. Wrap
it in formatted messages and store the binary output in devlink core.
Example:
$ devlink health dump show pci/0000:00:0b.0 reporter tx
SX Slice:
data:
00 00 00 00 00 00 00 80 00 01 00 00 00 00 ad de
22 01 00 00 00 00 ad de 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ff ff ff ff 01 00 00 00 00 00 00 00 00 00 00 00
00 02 01 00 00 00 00 80 00 01 00 00 00 00 ad de
22 01 00 00 00 00 ad de 00 20 40 90 81 88 ff ff
00 00 00 00 00 00 00 00 15 00 15 00 00 00 00 00
ff ff ff ff 01 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 80 81 ae 41 06 00 ea ff ff
SQs:
SQ:
index: 1511
data:
00 00 00 00 00 00 00 80 00 01 00 00 00 00 ad de
22 01 00 00 00 00 ad de 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ff ff ff ff 01 00 00 00 00 00 00 00 00 00 00 00
00 02 01 00 00 00 00 80 00 01 00 00 00 00 ad de
22 01 00 00 00 00 ad de 00 20 40 90 81 88 ff ff
00 00 00 00 00 00 00 00 15 00 15 00 00 00 00 00
ff ff ff ff 01 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 80 81 ae 41 06 00 ea ff ff
SQ:
index: 1516
data:
00 00 00 00 00 00 00 80 00 01 00 00 00 00 ad de
22 01 00 00 00 00 ad de 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ff ff ff ff 01 00 00 00 00 00 00 00 00 00 00 00
00 02 01 00 00 00 00 80 00 01 00 00 00 00 ad de
22 01 00 00 00 00 ad de 00 20 40 90 81 88 ff ff
00 00 00 00 00 00 00 00 15 00 15 00 00 00 00 00
ff ff ff ff 01 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 80 81 ae 41 06 00 ea ff ff
$ devlink health dump show pci/0000:00:0b.0 reporter tx -jp
{
"SX Slice": {
"data": [ 0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,2,1,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,32,64,144,129,136,255,255,0,0,0,0,0,0,0,0,21,0,21,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,129,174,65,6,0,234,255,255],
},
"SQs": [ {
"SQ": {
"index": 1511,
"data": [ 0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,2,1,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,32,64,144,129,136,255,255,0,0,0,0,0,0,0,0,21,0,21,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,129,174,65,6,0,234,255,255]
}
},{
"SQ": {
"index": 1516,
"data": [ 0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,2,1,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,32,64,144,129,136,255,255,0,0,0,0,0,0,0,0,21,0,21,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,129,174,65,6,0,234,255,255]
}
} ]
}
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../ethernet/mellanox/mlx5/core/en/health.c | 105 +++++++++++++++
.../ethernet/mellanox/mlx5/core/en/health.h | 8 +-
.../mellanox/mlx5/core/en/reporter_tx.c | 123 ++++++++++++++++++
3 files changed, 234 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/health.c b/drivers/net/ethernet/mellanox/mlx5/core/en/health.c
index 3a975641f902..7178f421d2cb 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/health.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/health.c
@@ -3,6 +3,7 @@
#include "health.h"
#include "lib/eq.h"
+#include "lib/mlx5.h"
int mlx5e_reporter_named_obj_nest_start(struct devlink_fmsg *fmsg, char *name)
{
@@ -204,3 +205,107 @@ int mlx5e_health_report(struct mlx5e_priv *priv,
return devlink_health_report(reporter, err_str, err_ctx);
}
+
+#define MLX5_HEALTH_DEVLINK_MAX_SIZE 1024
+static int mlx5e_health_rsc_fmsg_binary(struct devlink_fmsg *fmsg,
+ const void *value, u32 value_len)
+
+{
+ u32 data_size;
+ u32 offset;
+ int err;
+
+ for (offset = 0; offset < value_len; offset += data_size) {
+ data_size = value_len - offset;
+ if (data_size > MLX5_HEALTH_DEVLINK_MAX_SIZE)
+ data_size = MLX5_HEALTH_DEVLINK_MAX_SIZE;
+ err = devlink_fmsg_binary_put(fmsg, value + offset, data_size);
+ if (err)
+ break;
+ }
+ return err;
+}
+
+int mlx5e_health_rsc_fmsg_dump(struct mlx5e_priv *priv, struct mlx5_rsc_key *key,
+ struct devlink_fmsg *fmsg)
+{
+ struct mlx5_core_dev *mdev = priv->mdev;
+ struct mlx5_rsc_dump_cmd *cmd;
+ struct page *page;
+ int cmd_err, err;
+ int end_err;
+ int size;
+
+ if (IS_ERR_OR_NULL(mdev->rsc_dump))
+ return -EOPNOTSUPP;
+
+ page = alloc_page(GFP_KERNEL);
+ if (!page)
+ return -ENOMEM;
+
+ err = devlink_fmsg_binary_pair_nest_start(fmsg, "data");
+ if (err)
+ return err;
+
+ cmd = mlx5_rsc_dump_cmd_create(mdev, key);
+ if (IS_ERR(cmd)) {
+ err = PTR_ERR(cmd);
+ goto free_page;
+ }
+
+ do {
+ cmd_err = mlx5_rsc_dump_next(mdev, cmd, page, &size);
+ if (cmd_err < 0) {
+ err = cmd_err;
+ goto destroy_cmd;
+ }
+
+ err = mlx5e_health_rsc_fmsg_binary(fmsg, page_address(page), size);
+ if (err)
+ goto destroy_cmd;
+
+ } while (cmd_err > 0);
+
+destroy_cmd:
+ mlx5_rsc_dump_cmd_destroy(cmd);
+ end_err = devlink_fmsg_binary_pair_nest_end(fmsg);
+ if (end_err)
+ err = end_err;
+free_page:
+ __free_page(page);
+ return err;
+}
+
+int mlx5e_health_queue_dump(struct mlx5e_priv *priv, struct devlink_fmsg *fmsg,
+ int queue_idx, char *lbl)
+{
+ struct mlx5_rsc_key key = {};
+ int err;
+
+ key.rsc = MLX5_SGMT_TYPE_FULL_QPC;
+ key.index1 = queue_idx;
+ key.size = PAGE_SIZE;
+ key.num_of_obj1 = 1;
+
+ err = devlink_fmsg_obj_nest_start(fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, lbl);
+ if (err)
+ return err;
+
+ err = devlink_fmsg_u32_pair_put(fmsg, "index", queue_idx);
+ if (err)
+ return err;
+
+ err = mlx5e_health_rsc_fmsg_dump(priv, &key, fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_end(fmsg);
+ if (err)
+ return err;
+
+ return devlink_fmsg_obj_nest_end(fmsg);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/health.h b/drivers/net/ethernet/mellanox/mlx5/core/en/health.h
index d3693fa547ac..e90e3aec422f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/health.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/health.h
@@ -5,6 +5,7 @@
#define __MLX5E_EN_HEALTH_H
#include "en.h"
+#include "diag/rsc_dump.h"
#define MLX5E_RX_ERR_CQE(cqe) (get_cqe_opcode(cqe) != MLX5_CQE_RESP_SEND)
@@ -36,6 +37,7 @@ void mlx5e_reporter_rx_timeout(struct mlx5e_rq *rq);
struct mlx5e_err_ctx {
int (*recover)(void *ctx);
+ int (*dump)(struct mlx5e_priv *priv, struct devlink_fmsg *fmsg, void *ctx);
void *ctx;
};
@@ -48,6 +50,8 @@ int mlx5e_health_report(struct mlx5e_priv *priv,
int mlx5e_health_create_reporters(struct mlx5e_priv *priv);
void mlx5e_health_destroy_reporters(struct mlx5e_priv *priv);
void mlx5e_health_channels_update(struct mlx5e_priv *priv);
-
-
+int mlx5e_health_rsc_fmsg_dump(struct mlx5e_priv *priv, struct mlx5_rsc_key *key,
+ struct devlink_fmsg *fmsg);
+int mlx5e_health_queue_dump(struct mlx5e_priv *priv, struct devlink_fmsg *fmsg,
+ int queue_idx, char *lbl);
#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
index 623c949db54c..1772c9ce3938 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
@@ -246,6 +246,126 @@ static int mlx5e_tx_reporter_diagnose(struct devlink_health_reporter *reporter,
return err;
}
+static int mlx5e_tx_reporter_dump_sq(struct mlx5e_priv *priv, struct devlink_fmsg *fmsg,
+ void *ctx)
+{
+ struct mlx5_rsc_key key = {};
+ struct mlx5e_txqsq *sq = ctx;
+ int err;
+
+ if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+ return 0;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "SX Slice");
+ if (err)
+ return err;
+
+ key.size = PAGE_SIZE;
+ key.rsc = MLX5_SGMT_TYPE_SX_SLICE_ALL;
+ err = mlx5e_health_rsc_fmsg_dump(priv, &key, fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_end(fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "SQ");
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "QPC");
+ if (err)
+ return err;
+
+ key.rsc = MLX5_SGMT_TYPE_FULL_QPC;
+ key.index1 = sq->sqn;
+ key.num_of_obj1 = 1;
+
+ err = mlx5e_health_rsc_fmsg_dump(priv, &key, fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_end(fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "send_buff");
+ if (err)
+ return err;
+
+ key.rsc = MLX5_SGMT_TYPE_SND_BUFF;
+ key.num_of_obj2 = MLX5_RSC_DUMP_ALL;
+ err = mlx5e_health_rsc_fmsg_dump(priv, &key, fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_end(fmsg);
+ if (err)
+ return err;
+
+ return mlx5e_reporter_named_obj_nest_end(fmsg);
+}
+
+static int mlx5e_tx_reporter_dump_all_sqs(struct mlx5e_priv *priv,
+ struct devlink_fmsg *fmsg)
+{
+ struct mlx5_rsc_key key = {};
+ int i, tc, err;
+
+ if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+ return 0;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "SX Slice");
+ if (err)
+ return err;
+
+ key.size = PAGE_SIZE;
+ key.rsc = MLX5_SGMT_TYPE_SX_SLICE_ALL;
+ err = mlx5e_health_rsc_fmsg_dump(priv, &key, fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_end(fmsg);
+ if (err)
+ return err;
+
+ err = devlink_fmsg_arr_pair_nest_start(fmsg, "SQs");
+ if (err)
+ return err;
+
+ for (i = 0; i < priv->channels.num; i++) {
+ struct mlx5e_channel *c = priv->channels.c[i];
+
+ for (tc = 0; tc < priv->channels.params.num_tc; tc++) {
+ struct mlx5e_txqsq *sq = &c->sq[tc];
+
+ err = mlx5e_health_queue_dump(priv, fmsg, sq->sqn, "SQ");
+ if (err)
+ return err;
+ }
+ }
+ return devlink_fmsg_arr_pair_nest_end(fmsg);
+}
+
+static int mlx5e_tx_reporter_dump_from_ctx(struct mlx5e_priv *priv,
+ struct mlx5e_err_ctx *err_ctx,
+ struct devlink_fmsg *fmsg)
+{
+ return err_ctx->dump(priv, fmsg, err_ctx->ctx);
+}
+
+static int mlx5e_tx_reporter_dump(struct devlink_health_reporter *reporter,
+ struct devlink_fmsg *fmsg, void *context,
+ struct netlink_ext_ack *extack)
+{
+ struct mlx5e_priv *priv = devlink_health_reporter_priv(reporter);
+ struct mlx5e_err_ctx *err_ctx = context;
+
+ return err_ctx ? mlx5e_tx_reporter_dump_from_ctx(priv, err_ctx, fmsg) :
+ mlx5e_tx_reporter_dump_all_sqs(priv, fmsg);
+}
+
void mlx5e_reporter_tx_err_cqe(struct mlx5e_txqsq *sq)
{
struct mlx5e_priv *priv = sq->channel->priv;
@@ -254,6 +374,7 @@ void mlx5e_reporter_tx_err_cqe(struct mlx5e_txqsq *sq)
err_ctx.ctx = sq;
err_ctx.recover = mlx5e_tx_reporter_err_cqe_recover;
+ err_ctx.dump = mlx5e_tx_reporter_dump_sq;
sprintf(err_str, "ERR CQE on SQ: 0x%x", sq->sqn);
mlx5e_health_report(priv, priv->tx_reporter, err_str, &err_ctx);
@@ -267,6 +388,7 @@ int mlx5e_reporter_tx_timeout(struct mlx5e_txqsq *sq)
err_ctx.ctx = sq;
err_ctx.recover = mlx5e_tx_reporter_timeout_recover;
+ err_ctx.dump = mlx5e_tx_reporter_dump_sq;
sprintf(err_str,
"TX timeout on queue: %d, SQ: 0x%x, CQ: 0x%x, SQ Cons: 0x%x SQ Prod: 0x%x, usecs since last trans: %u\n",
sq->channel->ix, sq->sqn, sq->cq.mcq.cqn, sq->cc, sq->pc,
@@ -279,6 +401,7 @@ static const struct devlink_health_reporter_ops mlx5_tx_reporter_ops = {
.name = "tx",
.recover = mlx5e_tx_reporter_recover,
.diagnose = mlx5e_tx_reporter_diagnose,
+ .dump = mlx5e_tx_reporter_dump,
};
#define MLX5_REPORTER_TX_GRACEFUL_PERIOD 500
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* [net-next V3 05/13] net/mlx5e: Support dump callback in RX reporter
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
` (3 preceding siblings ...)
2020-02-11 22:32 ` [net-next V3 04/13] net/mlx5e: Support dump callback in TX reporter Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 06/13] mlx5: Use proper logging and tracing line terminations Saeed Mahameed
` (9 subsequent siblings)
14 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: netdev, Aya Levin, Moshe Shemesh, Jiri Pirko, Saeed Mahameed
From: Aya Levin <ayal@mellanox.com>
Add support for SQ's FW dump on RX reporter's events. Use Resource dump
API to retrieve the relevant data: RX slice, RQ dump, RX buffer and
ICOSQ dump (depends on the error). Wrap it in formatted messages and
store the binary output in devlink core.
Example:
$ devlink health dump show pci/0000:00:0b.0 reporter rx
RX Slice:
data:
00 00 00 00 00 00 00 80 00 01 00 00 00 00 ad de
22 01 00 00 00 00 ad de 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ff ff ff ff 01 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 80 00 01 00 00 00 00 ad de
22 01 00 00 00 00 ad de 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ff ff ff ff 01 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 80 00 01 00 00 00 00 ad de
RQs:
RQ:
rqn: 1512
data:
00 00 00 00 00 00 00 80 00 01 00 00 00 00 ad de
22 01 00 00 00 00 ad de 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ff ff ff ff 01 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 80 00 01 00 00 00 00 ad de
22 01 00 00 00 00 ad de 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ff ff ff ff 01 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 80 00 01 00 00 00 00 ad de
RQ:
rqn: 1517
data:
00 00 00 00 00 00 00 80 00 01 00 00 00 00 ad de
22 01 00 00 00 00 ad de 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ff ff ff ff 01 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 80 00 01 00 00 00 00 ad de
22 01 00 00 00 00 ad de 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
ff ff ff ff 01 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 80 00 01 00 00 00 00 ad de
$ devlink health dump show pci/0000:00:0b.0 reporter rx -jp
{
"RX Slice": {
"data":[ 0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,2,1,0,0,0,0,128,0,1,0,0,0,0,173,222]
},
"RQs": [ {
"RQ": {
"index": 1512,
"data": [ 0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,2,1,0,0,0,0,128,0,1,0,0,0,0,173,222]
}
},{
"RQ": {
"index": 1517,
"data": [ 0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,128,0,1,0,0,0,0,173,222,34,1,0,0,0,0,173,222,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,1,0,0,0,0,0,0,0,0,0,0,0,0,2,1,0,0,0,0,128,0,1,0,0,0,0,173]
}
} ]
}
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../mellanox/mlx5/core/en/reporter_rx.c | 183 ++++++++++++++++++
1 file changed, 183 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c
index cfa6941fca6b..9599fdd620a8 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c
@@ -330,6 +330,185 @@ static int mlx5e_rx_reporter_diagnose(struct devlink_health_reporter *reporter,
return err;
}
+static int mlx5e_rx_reporter_dump_icosq(struct mlx5e_priv *priv, struct devlink_fmsg *fmsg,
+ void *ctx)
+{
+ struct mlx5e_txqsq *icosq = ctx;
+ struct mlx5_rsc_key key = {};
+ int err;
+
+ if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+ return 0;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "SX Slice");
+ if (err)
+ return err;
+
+ key.size = PAGE_SIZE;
+ key.rsc = MLX5_SGMT_TYPE_SX_SLICE_ALL;
+ err = mlx5e_health_rsc_fmsg_dump(priv, &key, fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_end(fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "ICOSQ");
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "QPC");
+ if (err)
+ return err;
+
+ key.rsc = MLX5_SGMT_TYPE_FULL_QPC;
+ key.index1 = icosq->sqn;
+ key.num_of_obj1 = 1;
+
+ err = mlx5e_health_rsc_fmsg_dump(priv, &key, fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_end(fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "send_buff");
+ if (err)
+ return err;
+
+ key.rsc = MLX5_SGMT_TYPE_SND_BUFF;
+ key.num_of_obj2 = MLX5_RSC_DUMP_ALL;
+
+ err = mlx5e_health_rsc_fmsg_dump(priv, &key, fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_end(fmsg);
+ if (err)
+ return err;
+
+ return mlx5e_reporter_named_obj_nest_end(fmsg);
+}
+
+static int mlx5e_rx_reporter_dump_rq(struct mlx5e_priv *priv, struct devlink_fmsg *fmsg,
+ void *ctx)
+{
+ struct mlx5_rsc_key key = {};
+ struct mlx5e_rq *rq = ctx;
+ int err;
+
+ if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+ return 0;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "RX Slice");
+ if (err)
+ return err;
+
+ key.size = PAGE_SIZE;
+ key.rsc = MLX5_SGMT_TYPE_RX_SLICE_ALL;
+ err = mlx5e_health_rsc_fmsg_dump(priv, &key, fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_end(fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "RQ");
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "QPC");
+ if (err)
+ return err;
+
+ key.rsc = MLX5_SGMT_TYPE_FULL_QPC;
+ key.index1 = rq->rqn;
+ key.num_of_obj1 = 1;
+
+ err = mlx5e_health_rsc_fmsg_dump(priv, &key, fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_end(fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "receive_buff");
+ if (err)
+ return err;
+
+ key.rsc = MLX5_SGMT_TYPE_RCV_BUFF;
+ key.num_of_obj2 = MLX5_RSC_DUMP_ALL;
+ err = mlx5e_health_rsc_fmsg_dump(priv, &key, fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_end(fmsg);
+ if (err)
+ return err;
+
+ return mlx5e_reporter_named_obj_nest_end(fmsg);
+}
+
+static int mlx5e_rx_reporter_dump_all_rqs(struct mlx5e_priv *priv,
+ struct devlink_fmsg *fmsg)
+{
+ struct mlx5_rsc_key key = {};
+ int i, err;
+
+ if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+ return 0;
+
+ err = mlx5e_reporter_named_obj_nest_start(fmsg, "RX Slice");
+ if (err)
+ return err;
+
+ key.size = PAGE_SIZE;
+ key.rsc = MLX5_SGMT_TYPE_RX_SLICE_ALL;
+ err = mlx5e_health_rsc_fmsg_dump(priv, &key, fmsg);
+ if (err)
+ return err;
+
+ err = mlx5e_reporter_named_obj_nest_end(fmsg);
+ if (err)
+ return err;
+
+ err = devlink_fmsg_arr_pair_nest_start(fmsg, "RQs");
+ if (err)
+ return err;
+
+ for (i = 0; i < priv->channels.num; i++) {
+ struct mlx5e_rq *rq = &priv->channels.c[i]->rq;
+
+ err = mlx5e_health_queue_dump(priv, fmsg, rq->rqn, "RQ");
+ if (err)
+ return err;
+ }
+
+ return devlink_fmsg_arr_pair_nest_end(fmsg);
+}
+
+static int mlx5e_rx_reporter_dump_from_ctx(struct mlx5e_priv *priv,
+ struct mlx5e_err_ctx *err_ctx,
+ struct devlink_fmsg *fmsg)
+{
+ return err_ctx->dump(priv, fmsg, err_ctx->ctx);
+}
+
+static int mlx5e_rx_reporter_dump(struct devlink_health_reporter *reporter,
+ struct devlink_fmsg *fmsg, void *context,
+ struct netlink_ext_ack *extack)
+{
+ struct mlx5e_priv *priv = devlink_health_reporter_priv(reporter);
+ struct mlx5e_err_ctx *err_ctx = context;
+
+ return err_ctx ? mlx5e_rx_reporter_dump_from_ctx(priv, err_ctx, fmsg) :
+ mlx5e_rx_reporter_dump_all_rqs(priv, fmsg);
+}
+
void mlx5e_reporter_rx_timeout(struct mlx5e_rq *rq)
{
struct mlx5e_icosq *icosq = &rq->channel->icosq;
@@ -339,6 +518,7 @@ void mlx5e_reporter_rx_timeout(struct mlx5e_rq *rq)
err_ctx.ctx = rq;
err_ctx.recover = mlx5e_rx_reporter_timeout_recover;
+ err_ctx.dump = mlx5e_rx_reporter_dump_rq;
sprintf(err_str, "RX timeout on channel: %d, ICOSQ: 0x%x RQ: 0x%x, CQ: 0x%x\n",
icosq->channel->ix, icosq->sqn, rq->rqn, rq->cq.mcq.cqn);
@@ -353,6 +533,7 @@ void mlx5e_reporter_rq_cqe_err(struct mlx5e_rq *rq)
err_ctx.ctx = rq;
err_ctx.recover = mlx5e_rx_reporter_err_rq_cqe_recover;
+ err_ctx.dump = mlx5e_rx_reporter_dump_rq;
sprintf(err_str, "ERR CQE on RQ: 0x%x", rq->rqn);
mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx);
@@ -366,6 +547,7 @@ void mlx5e_reporter_icosq_cqe_err(struct mlx5e_icosq *icosq)
err_ctx.ctx = icosq;
err_ctx.recover = mlx5e_rx_reporter_err_icosq_cqe_recover;
+ err_ctx.dump = mlx5e_rx_reporter_dump_icosq;
sprintf(err_str, "ERR CQE on ICOSQ: 0x%x", icosq->sqn);
mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx);
@@ -375,6 +557,7 @@ static const struct devlink_health_reporter_ops mlx5_rx_reporter_ops = {
.name = "rx",
.recover = mlx5e_rx_reporter_recover,
.diagnose = mlx5e_rx_reporter_diagnose,
+ .dump = mlx5e_rx_reporter_dump,
};
#define MLX5E_REPORTER_RX_GRACEFUL_PERIOD 500
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* [net-next V3 06/13] mlx5: Use proper logging and tracing line terminations
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
` (4 preceding siblings ...)
2020-02-11 22:32 ` [net-next V3 05/13] net/mlx5e: Support dump callback in RX reporter Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 07/13] net/mlx5e: Set FEC to auto when configured mode is not supported Saeed Mahameed
` (8 subsequent siblings)
14 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski; +Cc: netdev, Joe Perches, Saeed Mahameed
From: Joe Perches <joe@perches.com>
netdev_err should use newline termination but mlx5_health_report
is used in a trace output function devlink_health_report where
no newline should be used.
Remove the newlines from a couple formats and add a format string
of "%s\n" to the netdev_err call to not directly output the
logging string.
Also use snprintf to avoid any possible output string overrun.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en/health.c | 2 +-
.../net/ethernet/mellanox/mlx5/core/en/reporter_rx.c | 9 +++++----
.../net/ethernet/mellanox/mlx5/core/en/reporter_tx.c | 10 +++++-----
3 files changed, 11 insertions(+), 10 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/health.c b/drivers/net/ethernet/mellanox/mlx5/core/en/health.c
index 7178f421d2cb..68d019b27642 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/health.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/health.c
@@ -198,7 +198,7 @@ int mlx5e_health_report(struct mlx5e_priv *priv,
struct devlink_health_reporter *reporter, char *err_str,
struct mlx5e_err_ctx *err_ctx)
{
- netdev_err(priv->netdev, err_str);
+ netdev_err(priv->netdev, "%s\n", err_str);
if (!reporter)
return err_ctx->recover(&err_ctx->ctx);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c
index 9599fdd620a8..ce2b41c61090 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c
@@ -519,8 +519,9 @@ void mlx5e_reporter_rx_timeout(struct mlx5e_rq *rq)
err_ctx.ctx = rq;
err_ctx.recover = mlx5e_rx_reporter_timeout_recover;
err_ctx.dump = mlx5e_rx_reporter_dump_rq;
- sprintf(err_str, "RX timeout on channel: %d, ICOSQ: 0x%x RQ: 0x%x, CQ: 0x%x\n",
- icosq->channel->ix, icosq->sqn, rq->rqn, rq->cq.mcq.cqn);
+ snprintf(err_str, sizeof(err_str),
+ "RX timeout on channel: %d, ICOSQ: 0x%x RQ: 0x%x, CQ: 0x%x",
+ icosq->channel->ix, icosq->sqn, rq->rqn, rq->cq.mcq.cqn);
mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx);
}
@@ -534,7 +535,7 @@ void mlx5e_reporter_rq_cqe_err(struct mlx5e_rq *rq)
err_ctx.ctx = rq;
err_ctx.recover = mlx5e_rx_reporter_err_rq_cqe_recover;
err_ctx.dump = mlx5e_rx_reporter_dump_rq;
- sprintf(err_str, "ERR CQE on RQ: 0x%x", rq->rqn);
+ snprintf(err_str, sizeof(err_str), "ERR CQE on RQ: 0x%x", rq->rqn);
mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx);
}
@@ -548,7 +549,7 @@ void mlx5e_reporter_icosq_cqe_err(struct mlx5e_icosq *icosq)
err_ctx.ctx = icosq;
err_ctx.recover = mlx5e_rx_reporter_err_icosq_cqe_recover;
err_ctx.dump = mlx5e_rx_reporter_dump_icosq;
- sprintf(err_str, "ERR CQE on ICOSQ: 0x%x", icosq->sqn);
+ snprintf(err_str, sizeof(err_str), "ERR CQE on ICOSQ: 0x%x", icosq->sqn);
mlx5e_health_report(priv, priv->rx_reporter, err_str, &err_ctx);
}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
index 1772c9ce3938..2028ce9b151f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/reporter_tx.c
@@ -375,7 +375,7 @@ void mlx5e_reporter_tx_err_cqe(struct mlx5e_txqsq *sq)
err_ctx.ctx = sq;
err_ctx.recover = mlx5e_tx_reporter_err_cqe_recover;
err_ctx.dump = mlx5e_tx_reporter_dump_sq;
- sprintf(err_str, "ERR CQE on SQ: 0x%x", sq->sqn);
+ snprintf(err_str, sizeof(err_str), "ERR CQE on SQ: 0x%x", sq->sqn);
mlx5e_health_report(priv, priv->tx_reporter, err_str, &err_ctx);
}
@@ -389,10 +389,10 @@ int mlx5e_reporter_tx_timeout(struct mlx5e_txqsq *sq)
err_ctx.ctx = sq;
err_ctx.recover = mlx5e_tx_reporter_timeout_recover;
err_ctx.dump = mlx5e_tx_reporter_dump_sq;
- sprintf(err_str,
- "TX timeout on queue: %d, SQ: 0x%x, CQ: 0x%x, SQ Cons: 0x%x SQ Prod: 0x%x, usecs since last trans: %u\n",
- sq->channel->ix, sq->sqn, sq->cq.mcq.cqn, sq->cc, sq->pc,
- jiffies_to_usecs(jiffies - sq->txq->trans_start));
+ snprintf(err_str, sizeof(err_str),
+ "TX timeout on queue: %d, SQ: 0x%x, CQ: 0x%x, SQ Cons: 0x%x SQ Prod: 0x%x, usecs since last trans: %u",
+ sq->channel->ix, sq->sqn, sq->cq.mcq.cqn, sq->cc, sq->pc,
+ jiffies_to_usecs(jiffies - sq->txq->trans_start));
return mlx5e_health_report(priv, priv->tx_reporter, err_str, &err_ctx);
}
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* [net-next V3 07/13] net/mlx5e: Set FEC to auto when configured mode is not supported
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
` (5 preceding siblings ...)
2020-02-11 22:32 ` [net-next V3 06/13] mlx5: Use proper logging and tracing line terminations Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 08/13] net/mlx5e: Enforce setting of a single FEC mode Saeed Mahameed
` (7 subsequent siblings)
14 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: netdev, Aya Levin, Eran Ben Elisha, Saeed Mahameed
From: Aya Levin <ayal@mellanox.com>
When configuring FEC mode, driver tries to set it for all available
link types. If a link type doesn't support a FEC mode, set this link
type to auto (FW best effort). Prior to this patch, when a link type
didn't support a FEC mode is was set to no FEC.
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/en/port.c | 22 +++++--------------
1 file changed, 6 insertions(+), 16 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/port.c b/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
index fce6eccdcf8b..f0dc0ca3ddc4 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
@@ -501,8 +501,6 @@ int mlx5e_get_fec_mode(struct mlx5_core_dev *dev, u32 *fec_mode_active,
int mlx5e_set_fec_mode(struct mlx5_core_dev *dev, u8 fec_policy)
{
- u8 fec_policy_nofec = BIT(MLX5E_FEC_NOFEC);
- bool fec_mode_not_supp_in_speed = false;
u32 out[MLX5_ST_SZ_DW(pplm_reg)] = {};
u32 in[MLX5_ST_SZ_DW(pplm_reg)] = {};
int sz = MLX5_ST_SZ_BYTES(pplm_reg);
@@ -526,23 +524,15 @@ int mlx5e_set_fec_mode(struct mlx5_core_dev *dev, u8 fec_policy)
for (i = 0; i < MLX5E_FEC_SUPPORTED_SPEEDS; i++) {
mlx5e_get_fec_cap_field(out, &fec_caps, fec_supported_speeds[i]);
- /* policy supported for link speed, or policy is auto */
- if (fec_caps & fec_policy || fec_policy == fec_policy_auto) {
+ /* policy supported for link speed */
+ if (fec_caps & fec_policy)
mlx5e_fec_admin_field(out, &fec_policy, 1,
fec_supported_speeds[i]);
- } else {
- /* turn off FEC if supported. Else, leave it the same */
- if (fec_caps & fec_policy_nofec)
- mlx5e_fec_admin_field(out, &fec_policy_nofec, 1,
- fec_supported_speeds[i]);
- fec_mode_not_supp_in_speed = true;
- }
+ else
+ /* set FEC to auto*/
+ mlx5e_fec_admin_field(out, &fec_policy_auto, 1,
+ fec_supported_speeds[i]);
}
- if (fec_mode_not_supp_in_speed)
- mlx5_core_dbg(dev,
- "FEC policy 0x%x is not supported for some speeds",
- fec_policy);
-
return mlx5_core_access_reg(dev, out, sz, out, sz, MLX5_REG_PPLM, 0, 1);
}
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* [net-next V3 08/13] net/mlx5e: Enforce setting of a single FEC mode
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
` (6 preceding siblings ...)
2020-02-11 22:32 ` [net-next V3 07/13] net/mlx5e: Set FEC to auto when configured mode is not supported Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 09/13] net/mlx5e: Advertise globaly supported FEC modes Saeed Mahameed
` (6 subsequent siblings)
14 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski; +Cc: netdev, Aya Levin, Saeed Mahameed
From: Aya Levin <ayal@mellanox.com>
Ethtool command allow setting of several FEC modes in a single set
command. The driver can only set a single FEC mode at a time. With this
patch driver will reply not-supported on setting several FEC modes.
Signed-off-by: Aya Levin <ayal@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index d674cb679895..d1664ff1772b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -1541,6 +1541,10 @@ static int mlx5e_set_fecparam(struct net_device *netdev,
int mode;
int err;
+ if (bitmap_weight((unsigned long *)&fecparam->fec,
+ ETHTOOL_FEC_BASER_BIT + 1) > 1)
+ return -EOPNOTSUPP;
+
for (mode = 0; mode < ARRAY_SIZE(pplm_fec_2_ethtool); mode++) {
if (!(pplm_fec_2_ethtool[mode] & fecparam->fec))
continue;
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* [net-next V3 09/13] net/mlx5e: Advertise globaly supported FEC modes
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
` (7 preceding siblings ...)
2020-02-11 22:32 ` [net-next V3 08/13] net/mlx5e: Enforce setting of a single FEC mode Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 10/13] net/mlxe5: Separate between FEC and current speed Saeed Mahameed
` (5 subsequent siblings)
14 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: netdev, Aya Levin, Eran Ben Elisha, Saeed Mahameed
From: Aya Levin <ayal@mellanox.com>
Ethtool advertise supported link modes on an interface. Per each FEC
mode, query if there is a link type which supports it. If so, add this
FEC mode to the supported FEC modes list. Prior to this patch, ethtool
advertised only the supported FEC modes on the current link type.
Add an explicit mapping between internal FEC modes and ethtool link mode
bits. With this change, adding new FEC modes in the downstream patch
would be easier.
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/en/port.c | 17 +++---
.../net/ethernet/mellanox/mlx5/core/en/port.h | 2 +-
.../ethernet/mellanox/mlx5/core/en_ethtool.c | 55 +++++++++----------
3 files changed, 38 insertions(+), 36 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/port.c b/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
index f0dc0ca3ddc4..26c7849eeb7c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
@@ -441,13 +441,13 @@ static int mlx5e_get_fec_cap_field(u32 *pplm,
return 0;
}
-int mlx5e_get_fec_caps(struct mlx5_core_dev *dev, u8 *fec_caps)
+bool mlx5e_fec_in_caps(struct mlx5_core_dev *dev, int fec_policy)
{
u32 out[MLX5_ST_SZ_DW(pplm_reg)] = {};
u32 in[MLX5_ST_SZ_DW(pplm_reg)] = {};
int sz = MLX5_ST_SZ_BYTES(pplm_reg);
- u32 current_fec_speed;
int err;
+ int i;
if (!MLX5_CAP_GEN(dev, pcam_reg))
return -EOPNOTSUPP;
@@ -458,13 +458,16 @@ int mlx5e_get_fec_caps(struct mlx5_core_dev *dev, u8 *fec_caps)
MLX5_SET(pplm_reg, in, local_port, 1);
err = mlx5_core_access_reg(dev, in, sz, out, sz, MLX5_REG_PPLM, 0, 0);
if (err)
- return err;
+ return false;
- err = mlx5e_port_linkspeed(dev, ¤t_fec_speed);
- if (err)
- return err;
+ for (i = 0; i < MLX5E_FEC_SUPPORTED_SPEEDS; i++) {
+ u8 fec_caps;
- return mlx5e_get_fec_cap_field(out, fec_caps, current_fec_speed);
+ mlx5e_get_fec_cap_field(out, &fec_caps, fec_supported_speeds[i]);
+ if (fec_caps & fec_policy)
+ return true;
+ }
+ return false;
}
int mlx5e_get_fec_mode(struct mlx5_core_dev *dev, u32 *fec_mode_active,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/port.h b/drivers/net/ethernet/mellanox/mlx5/core/en/port.h
index 4a7f4497692b..025d86577567 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/port.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/port.h
@@ -60,7 +60,7 @@ int mlx5e_port_set_pbmc(struct mlx5_core_dev *mdev, void *in);
int mlx5e_port_query_priority2buffer(struct mlx5_core_dev *mdev, u8 *buffer);
int mlx5e_port_set_priority2buffer(struct mlx5_core_dev *mdev, u8 *buffer);
-int mlx5e_get_fec_caps(struct mlx5_core_dev *dev, u8 *fec_caps);
+bool mlx5e_fec_in_caps(struct mlx5_core_dev *dev, int fec_policy);
int mlx5e_get_fec_mode(struct mlx5_core_dev *dev, u32 *fec_mode_active,
u8 *fec_configured_mode);
int mlx5e_set_fec_mode(struct mlx5_core_dev *dev, u8 fec_policy);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index d1664ff1772b..6624e0a82cd9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -650,45 +650,44 @@ static u32 pplm2ethtool_fec(u_long fec_mode, unsigned long size)
return 0;
}
-/* we use ETHTOOL_FEC_* offset and apply it to ETHTOOL_LINK_MODE_FEC_*_BIT */
-static u32 ethtool_fec2ethtool_caps(u_long ethtool_fec_code)
-{
- u32 offset;
-
- offset = find_first_bit(ðtool_fec_code, sizeof(u32));
- offset -= ETHTOOL_FEC_OFF_BIT;
- offset += ETHTOOL_LINK_MODE_FEC_NONE_BIT;
-
- return offset;
-}
+#define MLX5E_ADVERTISE_SUPPORTED_FEC(mlx5_fec, ethtool_fec) \
+ do { \
+ if (mlx5e_fec_in_caps(dev, 1 << (mlx5_fec))) \
+ __set_bit(ethtool_fec, \
+ link_ksettings->link_modes.supported);\
+ } while (0)
+
+static const u32 pplm_fec_2_ethtool_linkmodes[] = {
+ [MLX5E_FEC_NOFEC] = ETHTOOL_LINK_MODE_FEC_NONE_BIT,
+ [MLX5E_FEC_FIRECODE] = ETHTOOL_LINK_MODE_FEC_BASER_BIT,
+ [MLX5E_FEC_RS_528_514] = ETHTOOL_LINK_MODE_FEC_RS_BIT,
+};
static int get_fec_supported_advertised(struct mlx5_core_dev *dev,
struct ethtool_link_ksettings *link_ksettings)
{
- u_long fec_caps = 0;
- u32 active_fec = 0;
- u32 offset;
+ u_long active_fec = 0;
u32 bitn;
int err;
- err = mlx5e_get_fec_caps(dev, (u8 *)&fec_caps);
+ err = mlx5e_get_fec_mode(dev, (u32 *)&active_fec, NULL);
if (err)
return (err == -EOPNOTSUPP) ? 0 : err;
- err = mlx5e_get_fec_mode(dev, &active_fec, NULL);
- if (err)
- return err;
-
- for_each_set_bit(bitn, &fec_caps, ARRAY_SIZE(pplm_fec_2_ethtool)) {
- u_long ethtool_bitmask = pplm_fec_2_ethtool[bitn];
+ MLX5E_ADVERTISE_SUPPORTED_FEC(MLX5E_FEC_NOFEC,
+ ETHTOOL_LINK_MODE_FEC_NONE_BIT);
+ MLX5E_ADVERTISE_SUPPORTED_FEC(MLX5E_FEC_FIRECODE,
+ ETHTOOL_LINK_MODE_FEC_BASER_BIT);
+ MLX5E_ADVERTISE_SUPPORTED_FEC(MLX5E_FEC_RS_528_514,
+ ETHTOOL_LINK_MODE_FEC_RS_BIT);
- offset = ethtool_fec2ethtool_caps(ethtool_bitmask);
- __set_bit(offset, link_ksettings->link_modes.supported);
- }
-
- active_fec = pplm2ethtool_fec(active_fec, sizeof(u32) * BITS_PER_BYTE);
- offset = ethtool_fec2ethtool_caps(active_fec);
- __set_bit(offset, link_ksettings->link_modes.advertising);
+ /* active fec is a bit set, find out which bit is set and
+ * advertise the corresponding ethtool bit
+ */
+ bitn = find_first_bit(&active_fec, sizeof(u32) * BITS_PER_BYTE);
+ if (bitn < ARRAY_SIZE(pplm_fec_2_ethtool_linkmodes))
+ __set_bit(pplm_fec_2_ethtool_linkmodes[bitn],
+ link_ksettings->link_modes.advertising);
return 0;
}
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* [net-next V3 10/13] net/mlxe5: Separate between FEC and current speed
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
` (8 preceding siblings ...)
2020-02-11 22:32 ` [net-next V3 09/13] net/mlx5e: Advertise globaly supported FEC modes Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 11/13] ethtool: Add support for low latency RS FEC Saeed Mahameed
` (4 subsequent siblings)
14 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski; +Cc: netdev, Aya Levin, Saeed Mahameed
From: Aya Levin <ayal@mellanox.com>
FEC mode is per link type, not necessary per speed. This patch access
FEC register by link modes instead of speeds. This patch will allow
further enhacment of link modes supporting FEC with the same speed
(different lane type).
Signed-off-by: Aya Levin <ayal@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/en/port.c | 146 ++++++++----------
1 file changed, 62 insertions(+), 84 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/port.c b/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
index 26c7849eeb7c..16c94950d206 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
@@ -343,64 +343,45 @@ int mlx5e_port_set_priority2buffer(struct mlx5_core_dev *mdev, u8 *buffer)
return err;
}
-static u32 fec_supported_speeds[] = {
- 10000,
- 40000,
- 25000,
- 50000,
- 56000,
- 100000
+enum mlx5e_fec_supported_link_mode {
+ MLX5E_FEC_SUPPORTED_LINK_MODES_10G_40G,
+ MLX5E_FEC_SUPPORTED_LINK_MODES_25G,
+ MLX5E_FEC_SUPPORTED_LINK_MODES_50G,
+ MLX5E_FEC_SUPPORTED_LINK_MODES_56G,
+ MLX5E_FEC_SUPPORTED_LINK_MODES_100G,
+ MLX5E_MAX_FEC_SUPPORTED_LINK_MODE,
};
-#define MLX5E_FEC_SUPPORTED_SPEEDS ARRAY_SIZE(fec_supported_speeds)
+#define MLX5E_FEC_OVERRIDE_ADMIN_POLICY(buf, policy, write, link) \
+ do { \
+ u8 *_policy = &(policy); \
+ u32 *_buf = buf; \
+ \
+ if (write) \
+ MLX5_SET(pplm_reg, _buf, fec_override_admin_##link, *_policy); \
+ else \
+ *_policy = MLX5_GET(pplm_reg, _buf, fec_override_admin_##link); \
+ } while (0)
/* get/set FEC admin field for a given speed */
-static int mlx5e_fec_admin_field(u32 *pplm,
- u8 *fec_policy,
- bool write,
- u32 speed)
+static int mlx5e_fec_admin_field(u32 *pplm, u8 *fec_policy, bool write,
+ enum mlx5e_fec_supported_link_mode link_mode)
{
- switch (speed) {
- case 10000:
- case 40000:
- if (!write)
- *fec_policy = MLX5_GET(pplm_reg, pplm,
- fec_override_admin_10g_40g);
- else
- MLX5_SET(pplm_reg, pplm,
- fec_override_admin_10g_40g, *fec_policy);
+ switch (link_mode) {
+ case MLX5E_FEC_SUPPORTED_LINK_MODES_10G_40G:
+ MLX5E_FEC_OVERRIDE_ADMIN_POLICY(pplm, *fec_policy, write, 10g_40g);
break;
- case 25000:
- if (!write)
- *fec_policy = MLX5_GET(pplm_reg, pplm,
- fec_override_admin_25g);
- else
- MLX5_SET(pplm_reg, pplm,
- fec_override_admin_25g, *fec_policy);
+ case MLX5E_FEC_SUPPORTED_LINK_MODES_25G:
+ MLX5E_FEC_OVERRIDE_ADMIN_POLICY(pplm, *fec_policy, write, 25g);
break;
- case 50000:
- if (!write)
- *fec_policy = MLX5_GET(pplm_reg, pplm,
- fec_override_admin_50g);
- else
- MLX5_SET(pplm_reg, pplm,
- fec_override_admin_50g, *fec_policy);
+ case MLX5E_FEC_SUPPORTED_LINK_MODES_50G:
+ MLX5E_FEC_OVERRIDE_ADMIN_POLICY(pplm, *fec_policy, write, 50g);
break;
- case 56000:
- if (!write)
- *fec_policy = MLX5_GET(pplm_reg, pplm,
- fec_override_admin_56g);
- else
- MLX5_SET(pplm_reg, pplm,
- fec_override_admin_56g, *fec_policy);
+ case MLX5E_FEC_SUPPORTED_LINK_MODES_56G:
+ MLX5E_FEC_OVERRIDE_ADMIN_POLICY(pplm, *fec_policy, write, 56g);
break;
- case 100000:
- if (!write)
- *fec_policy = MLX5_GET(pplm_reg, pplm,
- fec_override_admin_100g);
- else
- MLX5_SET(pplm_reg, pplm,
- fec_override_admin_100g, *fec_policy);
+ case MLX5E_FEC_SUPPORTED_LINK_MODES_100G:
+ MLX5E_FEC_OVERRIDE_ADMIN_POLICY(pplm, *fec_policy, write, 100g);
break;
default:
return -EINVAL;
@@ -408,32 +389,28 @@ static int mlx5e_fec_admin_field(u32 *pplm,
return 0;
}
+#define MLX5E_GET_FEC_OVERRIDE_CAP(buf, link) \
+ MLX5_GET(pplm_reg, buf, fec_override_cap_##link)
+
/* returns FEC capabilities for a given speed */
-static int mlx5e_get_fec_cap_field(u32 *pplm,
- u8 *fec_cap,
- u32 speed)
+static int mlx5e_get_fec_cap_field(u32 *pplm, u8 *fec_cap,
+ enum mlx5e_fec_supported_link_mode link_mode)
{
- switch (speed) {
- case 10000:
- case 40000:
- *fec_cap = MLX5_GET(pplm_reg, pplm,
- fec_override_cap_10g_40g);
+ switch (link_mode) {
+ case MLX5E_FEC_SUPPORTED_LINK_MODES_10G_40G:
+ *fec_cap = MLX5E_GET_FEC_OVERRIDE_CAP(pplm, 10g_40g);
break;
- case 25000:
- *fec_cap = MLX5_GET(pplm_reg, pplm,
- fec_override_cap_25g);
+ case MLX5E_FEC_SUPPORTED_LINK_MODES_25G:
+ *fec_cap = MLX5E_GET_FEC_OVERRIDE_CAP(pplm, 25g);
break;
- case 50000:
- *fec_cap = MLX5_GET(pplm_reg, pplm,
- fec_override_cap_50g);
+ case MLX5E_FEC_SUPPORTED_LINK_MODES_50G:
+ *fec_cap = MLX5E_GET_FEC_OVERRIDE_CAP(pplm, 50g);
break;
- case 56000:
- *fec_cap = MLX5_GET(pplm_reg, pplm,
- fec_override_cap_56g);
+ case MLX5E_FEC_SUPPORTED_LINK_MODES_56G:
+ *fec_cap = MLX5E_GET_FEC_OVERRIDE_CAP(pplm, 56g);
break;
- case 100000:
- *fec_cap = MLX5_GET(pplm_reg, pplm,
- fec_override_cap_100g);
+ case MLX5E_FEC_SUPPORTED_LINK_MODES_100G:
+ *fec_cap = MLX5E_GET_FEC_OVERRIDE_CAP(pplm, 100g);
break;
default:
return -EINVAL;
@@ -460,10 +437,10 @@ bool mlx5e_fec_in_caps(struct mlx5_core_dev *dev, int fec_policy)
if (err)
return false;
- for (i = 0; i < MLX5E_FEC_SUPPORTED_SPEEDS; i++) {
+ for (i = 0; i < MLX5E_MAX_FEC_SUPPORTED_LINK_MODE; i++) {
u8 fec_caps;
- mlx5e_get_fec_cap_field(out, &fec_caps, fec_supported_speeds[i]);
+ mlx5e_get_fec_cap_field(out, &fec_caps, i);
if (fec_caps & fec_policy)
return true;
}
@@ -476,8 +453,8 @@ int mlx5e_get_fec_mode(struct mlx5_core_dev *dev, u32 *fec_mode_active,
u32 out[MLX5_ST_SZ_DW(pplm_reg)] = {};
u32 in[MLX5_ST_SZ_DW(pplm_reg)] = {};
int sz = MLX5_ST_SZ_BYTES(pplm_reg);
- u32 link_speed;
int err;
+ int i;
if (!MLX5_CAP_GEN(dev, pcam_reg))
return -EOPNOTSUPP;
@@ -493,13 +470,16 @@ int mlx5e_get_fec_mode(struct mlx5_core_dev *dev, u32 *fec_mode_active,
*fec_mode_active = MLX5_GET(pplm_reg, out, fec_mode_active);
if (!fec_configured_mode)
- return 0;
-
- err = mlx5e_port_linkspeed(dev, &link_speed);
- if (err)
- return err;
+ goto out;
- return mlx5e_fec_admin_field(out, fec_configured_mode, 0, link_speed);
+ *fec_configured_mode = 0;
+ for (i = 0; i < MLX5E_MAX_FEC_SUPPORTED_LINK_MODE; i++) {
+ mlx5e_fec_admin_field(out, fec_configured_mode, 0, i);
+ if (*fec_configured_mode != 0)
+ goto out;
+ }
+out:
+ return 0;
}
int mlx5e_set_fec_mode(struct mlx5_core_dev *dev, u8 fec_policy)
@@ -525,16 +505,14 @@ int mlx5e_set_fec_mode(struct mlx5_core_dev *dev, u8 fec_policy)
MLX5_SET(pplm_reg, out, local_port, 1);
- for (i = 0; i < MLX5E_FEC_SUPPORTED_SPEEDS; i++) {
- mlx5e_get_fec_cap_field(out, &fec_caps, fec_supported_speeds[i]);
+ for (i = 0; i < MLX5E_MAX_FEC_SUPPORTED_LINK_MODE; i++) {
+ mlx5e_get_fec_cap_field(out, &fec_caps, i);
/* policy supported for link speed */
if (fec_caps & fec_policy)
- mlx5e_fec_admin_field(out, &fec_policy, 1,
- fec_supported_speeds[i]);
+ mlx5e_fec_admin_field(out, &fec_policy, 1, i);
else
/* set FEC to auto*/
- mlx5e_fec_admin_field(out, &fec_policy_auto, 1,
- fec_supported_speeds[i]);
+ mlx5e_fec_admin_field(out, &fec_policy_auto, 1, i);
}
return mlx5_core_access_reg(dev, out, sz, out, sz, MLX5_REG_PPLM, 0, 1);
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* [net-next V3 11/13] ethtool: Add support for low latency RS FEC
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
` (9 preceding siblings ...)
2020-02-11 22:32 ` [net-next V3 10/13] net/mlxe5: Separate between FEC and current speed Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 23:10 ` Andrew Lunn
2020-02-11 22:32 ` [net-next V3 12/13] net/mlx5e: Add support for FEC modes based on 50G per lane links Saeed Mahameed
` (3 subsequent siblings)
14 siblings, 1 reply; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: netdev, Aya Levin, Eran Ben Elisha, Andrew Lunn, Saeed Mahameed
From: Aya Levin <ayal@mellanox.com>
Add support for low latency Reed Solomon FEC as LLRS.
The LL-FEC is defined by the 25G/50G ethernet consortium,
in the document titled "Low Latency Reed Solomon Forward Error Correction"
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
CC: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/phy/phy-core.c | 2 +-
include/uapi/linux/ethtool.h | 4 +++-
net/ethtool/common.c | 1 +
net/ethtool/linkmodes.c | 1 +
4 files changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/net/phy/phy-core.c b/drivers/net/phy/phy-core.c
index a4d2d59fceca..e083e7a76ada 100644
--- a/drivers/net/phy/phy-core.c
+++ b/drivers/net/phy/phy-core.c
@@ -8,7 +8,7 @@
const char *phy_speed_to_str(int speed)
{
- BUILD_BUG_ON_MSG(__ETHTOOL_LINK_MODE_MASK_NBITS != 74,
+ BUILD_BUG_ON_MSG(__ETHTOOL_LINK_MODE_MASK_NBITS != 75,
"Enum ethtool_link_mode_bit_indices and phylib are out of sync. "
"If a speed or mode has been added please update phy_speed_to_str "
"and the PHY settings array.\n");
diff --git a/include/uapi/linux/ethtool.h b/include/uapi/linux/ethtool.h
index 4295ebfa2f91..d586ee5e10a1 100644
--- a/include/uapi/linux/ethtool.h
+++ b/include/uapi/linux/ethtool.h
@@ -1330,6 +1330,7 @@ enum ethtool_fec_config_bits {
ETHTOOL_FEC_OFF_BIT,
ETHTOOL_FEC_RS_BIT,
ETHTOOL_FEC_BASER_BIT,
+ ETHTOOL_FEC_LLRS_BIT,
};
#define ETHTOOL_FEC_NONE (1 << ETHTOOL_FEC_NONE_BIT)
@@ -1337,6 +1338,7 @@ enum ethtool_fec_config_bits {
#define ETHTOOL_FEC_OFF (1 << ETHTOOL_FEC_OFF_BIT)
#define ETHTOOL_FEC_RS (1 << ETHTOOL_FEC_RS_BIT)
#define ETHTOOL_FEC_BASER (1 << ETHTOOL_FEC_BASER_BIT)
+#define ETHTOOL_FEC_LLRS (1 << ETHTOOL_FEC_LLRS_BIT)
/* CMDs currently supported */
#define ETHTOOL_GSET 0x00000001 /* DEPRECATED, Get settings.
@@ -1521,7 +1523,7 @@ enum ethtool_link_mode_bit_indices {
ETHTOOL_LINK_MODE_400000baseLR8_ER8_FR8_Full_BIT = 71,
ETHTOOL_LINK_MODE_400000baseDR8_Full_BIT = 72,
ETHTOOL_LINK_MODE_400000baseCR8_Full_BIT = 73,
-
+ ETHTOOL_LINK_MODE_FEC_LLRS_BIT = 74,
/* must be last entry */
__ETHTOOL_LINK_MODE_MASK_NBITS
};
diff --git a/net/ethtool/common.c b/net/ethtool/common.c
index 636ec6d5110e..7b6969af5ae7 100644
--- a/net/ethtool/common.c
+++ b/net/ethtool/common.c
@@ -168,6 +168,7 @@ const char link_mode_names[][ETH_GSTRING_LEN] = {
__DEFINE_LINK_MODE_NAME(400000, LR8_ER8_FR8, Full),
__DEFINE_LINK_MODE_NAME(400000, DR8, Full),
__DEFINE_LINK_MODE_NAME(400000, CR8, Full),
+ __DEFINE_SPECIAL_MODE_NAME(FEC_LLRS, "LLRS"),
};
static_assert(ARRAY_SIZE(link_mode_names) == __ETHTOOL_LINK_MODE_MASK_NBITS);
diff --git a/net/ethtool/linkmodes.c b/net/ethtool/linkmodes.c
index 96f20be64553..f049b97072fe 100644
--- a/net/ethtool/linkmodes.c
+++ b/net/ethtool/linkmodes.c
@@ -237,6 +237,7 @@ static const struct link_mode_info link_mode_params[] = {
__DEFINE_LINK_MODE_PARAMS(400000, LR8_ER8_FR8, Full),
__DEFINE_LINK_MODE_PARAMS(400000, DR8, Full),
__DEFINE_LINK_MODE_PARAMS(400000, CR8, Full),
+ __DEFINE_SPECIAL_MODE_PARAMS(FEC_LLRS),
};
static const struct nla_policy
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* Re: [net-next V3 11/13] ethtool: Add support for low latency RS FEC
2020-02-11 22:32 ` [net-next V3 11/13] ethtool: Add support for low latency RS FEC Saeed Mahameed
@ 2020-02-11 23:10 ` Andrew Lunn
0 siblings, 0 replies; 19+ messages in thread
From: Andrew Lunn @ 2020-02-11 23:10 UTC (permalink / raw)
To: Saeed Mahameed
Cc: David S. Miller, Jakub Kicinski, netdev, Aya Levin, Eran Ben Elisha
On Tue, Feb 11, 2020 at 02:32:52PM -0800, Saeed Mahameed wrote:
> From: Aya Levin <ayal@mellanox.com>
>
> Add support for low latency Reed Solomon FEC as LLRS.
>
> The LL-FEC is defined by the 25G/50G ethernet consortium,
> in the document titled "Low Latency Reed Solomon Forward Error Correction"
>
> Signed-off-by: Aya Levin <ayal@mellanox.com>
> Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
> CC: Andrew Lunn <andrew@lunn.ch>
> Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Andrew
^ permalink raw reply [flat|nested] 19+ messages in thread
* [net-next V3 12/13] net/mlx5e: Add support for FEC modes based on 50G per lane links
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
` (10 preceding siblings ...)
2020-02-11 22:32 ` [net-next V3 11/13] ethtool: Add support for low latency RS FEC Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 22:32 ` [net-next V3 13/13] net/mlx5: Remove a useless 'drain_workqueue()' call in 'mlx5e_ipsec_cleanup()' Saeed Mahameed
` (2 subsequent siblings)
14 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski; +Cc: netdev, Aya Levin, Saeed Mahameed
From: Aya Levin <ayal@mellanox.com>
Introduce new FEC modes:
- RS-FEC-(544,514)
- LL_RS-FEC-(272,257+1)
Add support in ethtool for set and get callbacks for the new modes
above. While RS-FEC-(544,514) is mapped to exsiting RS FEC mode,
LL_RS-FEC-(272,257+1) is mapped to a new ethtool link mode: LL-RS.
Add support for FEC on 50G per lane link modes up to 400G. The new link
modes uses a u16 fields instead of u8 fields for the legacy link modes.
Signed-off-by: Aya Levin <ayal@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/en/port.c | 90 ++++++++++++++++---
.../net/ethernet/mellanox/mlx5/core/en/port.h | 6 +-
.../ethernet/mellanox/mlx5/core/en_ethtool.c | 14 ++-
3 files changed, 94 insertions(+), 16 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/port.c b/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
index 16c94950d206..2c4a670c8ffd 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/port.c
@@ -349,12 +349,18 @@ enum mlx5e_fec_supported_link_mode {
MLX5E_FEC_SUPPORTED_LINK_MODES_50G,
MLX5E_FEC_SUPPORTED_LINK_MODES_56G,
MLX5E_FEC_SUPPORTED_LINK_MODES_100G,
+ MLX5E_FEC_SUPPORTED_LINK_MODE_50G_1X,
+ MLX5E_FEC_SUPPORTED_LINK_MODE_100G_2X,
+ MLX5E_FEC_SUPPORTED_LINK_MODE_200G_4X,
+ MLX5E_FEC_SUPPORTED_LINK_MODE_400G_8X,
MLX5E_MAX_FEC_SUPPORTED_LINK_MODE,
};
+#define MLX5E_FEC_FIRST_50G_PER_LANE_MODE MLX5E_FEC_SUPPORTED_LINK_MODE_50G_1X
+
#define MLX5E_FEC_OVERRIDE_ADMIN_POLICY(buf, policy, write, link) \
do { \
- u8 *_policy = &(policy); \
+ u16 *_policy = &(policy); \
u32 *_buf = buf; \
\
if (write) \
@@ -363,8 +369,21 @@ enum mlx5e_fec_supported_link_mode {
*_policy = MLX5_GET(pplm_reg, _buf, fec_override_admin_##link); \
} while (0)
+#define MLX5E_FEC_OVERRIDE_ADMIN_50G_POLICY(buf, policy, write, link) \
+ do { \
+ u16 *__policy = &(policy); \
+ bool _write = (write); \
+ \
+ if (_write && *__policy) \
+ *__policy = find_first_bit((u_long *)__policy, \
+ sizeof(u16) * BITS_PER_BYTE);\
+ MLX5E_FEC_OVERRIDE_ADMIN_POLICY(buf, *__policy, _write, link); \
+ if (!_write && *__policy) \
+ *__policy = 1 << *__policy; \
+ } while (0)
+
/* get/set FEC admin field for a given speed */
-static int mlx5e_fec_admin_field(u32 *pplm, u8 *fec_policy, bool write,
+static int mlx5e_fec_admin_field(u32 *pplm, u16 *fec_policy, bool write,
enum mlx5e_fec_supported_link_mode link_mode)
{
switch (link_mode) {
@@ -383,6 +402,18 @@ static int mlx5e_fec_admin_field(u32 *pplm, u8 *fec_policy, bool write,
case MLX5E_FEC_SUPPORTED_LINK_MODES_100G:
MLX5E_FEC_OVERRIDE_ADMIN_POLICY(pplm, *fec_policy, write, 100g);
break;
+ case MLX5E_FEC_SUPPORTED_LINK_MODE_50G_1X:
+ MLX5E_FEC_OVERRIDE_ADMIN_50G_POLICY(pplm, *fec_policy, write, 50g_1x);
+ break;
+ case MLX5E_FEC_SUPPORTED_LINK_MODE_100G_2X:
+ MLX5E_FEC_OVERRIDE_ADMIN_50G_POLICY(pplm, *fec_policy, write, 100g_2x);
+ break;
+ case MLX5E_FEC_SUPPORTED_LINK_MODE_200G_4X:
+ MLX5E_FEC_OVERRIDE_ADMIN_50G_POLICY(pplm, *fec_policy, write, 200g_4x);
+ break;
+ case MLX5E_FEC_SUPPORTED_LINK_MODE_400G_8X:
+ MLX5E_FEC_OVERRIDE_ADMIN_50G_POLICY(pplm, *fec_policy, write, 400g_8x);
+ break;
default:
return -EINVAL;
}
@@ -393,7 +424,7 @@ static int mlx5e_fec_admin_field(u32 *pplm, u8 *fec_policy, bool write,
MLX5_GET(pplm_reg, buf, fec_override_cap_##link)
/* returns FEC capabilities for a given speed */
-static int mlx5e_get_fec_cap_field(u32 *pplm, u8 *fec_cap,
+static int mlx5e_get_fec_cap_field(u32 *pplm, u16 *fec_cap,
enum mlx5e_fec_supported_link_mode link_mode)
{
switch (link_mode) {
@@ -412,6 +443,18 @@ static int mlx5e_get_fec_cap_field(u32 *pplm, u8 *fec_cap,
case MLX5E_FEC_SUPPORTED_LINK_MODES_100G:
*fec_cap = MLX5E_GET_FEC_OVERRIDE_CAP(pplm, 100g);
break;
+ case MLX5E_FEC_SUPPORTED_LINK_MODE_50G_1X:
+ *fec_cap = MLX5E_GET_FEC_OVERRIDE_CAP(pplm, 50g_1x);
+ break;
+ case MLX5E_FEC_SUPPORTED_LINK_MODE_100G_2X:
+ *fec_cap = MLX5E_GET_FEC_OVERRIDE_CAP(pplm, 100g_2x);
+ break;
+ case MLX5E_FEC_SUPPORTED_LINK_MODE_200G_4X:
+ *fec_cap = MLX5E_GET_FEC_OVERRIDE_CAP(pplm, 200g_4x);
+ break;
+ case MLX5E_FEC_SUPPORTED_LINK_MODE_400G_8X:
+ *fec_cap = MLX5E_GET_FEC_OVERRIDE_CAP(pplm, 400g_8x);
+ break;
default:
return -EINVAL;
}
@@ -420,6 +463,7 @@ static int mlx5e_get_fec_cap_field(u32 *pplm, u8 *fec_cap,
bool mlx5e_fec_in_caps(struct mlx5_core_dev *dev, int fec_policy)
{
+ bool fec_50g_per_lane = MLX5_CAP_PCAM_FEATURE(dev, fec_50G_per_lane_in_pplm);
u32 out[MLX5_ST_SZ_DW(pplm_reg)] = {};
u32 in[MLX5_ST_SZ_DW(pplm_reg)] = {};
int sz = MLX5_ST_SZ_BYTES(pplm_reg);
@@ -438,7 +482,10 @@ bool mlx5e_fec_in_caps(struct mlx5_core_dev *dev, int fec_policy)
return false;
for (i = 0; i < MLX5E_MAX_FEC_SUPPORTED_LINK_MODE; i++) {
- u8 fec_caps;
+ u16 fec_caps;
+
+ if (i >= MLX5E_FEC_FIRST_50G_PER_LANE_MODE && !fec_50g_per_lane)
+ break;
mlx5e_get_fec_cap_field(out, &fec_caps, i);
if (fec_caps & fec_policy)
@@ -448,8 +495,9 @@ bool mlx5e_fec_in_caps(struct mlx5_core_dev *dev, int fec_policy)
}
int mlx5e_get_fec_mode(struct mlx5_core_dev *dev, u32 *fec_mode_active,
- u8 *fec_configured_mode)
+ u16 *fec_configured_mode)
{
+ bool fec_50g_per_lane = MLX5_CAP_PCAM_FEATURE(dev, fec_50G_per_lane_in_pplm);
u32 out[MLX5_ST_SZ_DW(pplm_reg)] = {};
u32 in[MLX5_ST_SZ_DW(pplm_reg)] = {};
int sz = MLX5_ST_SZ_BYTES(pplm_reg);
@@ -474,6 +522,9 @@ int mlx5e_get_fec_mode(struct mlx5_core_dev *dev, u32 *fec_mode_active,
*fec_configured_mode = 0;
for (i = 0; i < MLX5E_MAX_FEC_SUPPORTED_LINK_MODE; i++) {
+ if (i >= MLX5E_FEC_FIRST_50G_PER_LANE_MODE && !fec_50g_per_lane)
+ break;
+
mlx5e_fec_admin_field(out, fec_configured_mode, 0, i);
if (*fec_configured_mode != 0)
goto out;
@@ -482,13 +533,13 @@ int mlx5e_get_fec_mode(struct mlx5_core_dev *dev, u32 *fec_mode_active,
return 0;
}
-int mlx5e_set_fec_mode(struct mlx5_core_dev *dev, u8 fec_policy)
+int mlx5e_set_fec_mode(struct mlx5_core_dev *dev, u16 fec_policy)
{
+ bool fec_50g_per_lane = MLX5_CAP_PCAM_FEATURE(dev, fec_50G_per_lane_in_pplm);
u32 out[MLX5_ST_SZ_DW(pplm_reg)] = {};
u32 in[MLX5_ST_SZ_DW(pplm_reg)] = {};
int sz = MLX5_ST_SZ_BYTES(pplm_reg);
- u8 fec_policy_auto = 0;
- u8 fec_caps = 0;
+ u16 fec_policy_auto = 0;
int err;
int i;
@@ -498,6 +549,9 @@ int mlx5e_set_fec_mode(struct mlx5_core_dev *dev, u8 fec_policy)
if (!MLX5_CAP_PCAM_REG(dev, pplm))
return -EOPNOTSUPP;
+ if (fec_policy >= (1 << MLX5E_FEC_LLRS_272_257_1) && !fec_50g_per_lane)
+ return -EOPNOTSUPP;
+
MLX5_SET(pplm_reg, in, local_port, 1);
err = mlx5_core_access_reg(dev, in, sz, out, sz, MLX5_REG_PPLM, 0, 0);
if (err)
@@ -506,10 +560,26 @@ int mlx5e_set_fec_mode(struct mlx5_core_dev *dev, u8 fec_policy)
MLX5_SET(pplm_reg, out, local_port, 1);
for (i = 0; i < MLX5E_MAX_FEC_SUPPORTED_LINK_MODE; i++) {
+ u16 conf_fec = fec_policy;
+ u16 fec_caps = 0;
+
+ if (i >= MLX5E_FEC_FIRST_50G_PER_LANE_MODE && !fec_50g_per_lane)
+ break;
+
+ /* RS fec in ethtool is mapped to MLX5E_FEC_RS_528_514
+ * to link modes up to 25G per lane and to
+ * MLX5E_FEC_RS_544_514 in the new link modes based on
+ * 50 G per lane
+ */
+ if (conf_fec == (1 << MLX5E_FEC_RS_528_514) &&
+ i >= MLX5E_FEC_FIRST_50G_PER_LANE_MODE)
+ conf_fec = (1 << MLX5E_FEC_RS_544_514);
+
mlx5e_get_fec_cap_field(out, &fec_caps, i);
+
/* policy supported for link speed */
- if (fec_caps & fec_policy)
- mlx5e_fec_admin_field(out, &fec_policy, 1, i);
+ if (fec_caps & conf_fec)
+ mlx5e_fec_admin_field(out, &conf_fec, 1, i);
else
/* set FEC to auto*/
mlx5e_fec_admin_field(out, &fec_policy_auto, 1, i);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/port.h b/drivers/net/ethernet/mellanox/mlx5/core/en/port.h
index 025d86577567..a2ddd446dd59 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/port.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/port.h
@@ -62,13 +62,15 @@ int mlx5e_port_set_priority2buffer(struct mlx5_core_dev *mdev, u8 *buffer);
bool mlx5e_fec_in_caps(struct mlx5_core_dev *dev, int fec_policy);
int mlx5e_get_fec_mode(struct mlx5_core_dev *dev, u32 *fec_mode_active,
- u8 *fec_configured_mode);
-int mlx5e_set_fec_mode(struct mlx5_core_dev *dev, u8 fec_policy);
+ u16 *fec_configured_mode);
+int mlx5e_set_fec_mode(struct mlx5_core_dev *dev, u16 fec_policy);
enum {
MLX5E_FEC_NOFEC,
MLX5E_FEC_FIRECODE,
MLX5E_FEC_RS_528_514,
+ MLX5E_FEC_RS_544_514 = 7,
+ MLX5E_FEC_LLRS_272_257_1 = 9,
};
#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index 6624e0a82cd9..68b520df07e4 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -633,6 +633,8 @@ static const u32 pplm_fec_2_ethtool[] = {
[MLX5E_FEC_NOFEC] = ETHTOOL_FEC_OFF,
[MLX5E_FEC_FIRECODE] = ETHTOOL_FEC_BASER,
[MLX5E_FEC_RS_528_514] = ETHTOOL_FEC_RS,
+ [MLX5E_FEC_RS_544_514] = ETHTOOL_FEC_RS,
+ [MLX5E_FEC_LLRS_272_257_1] = ETHTOOL_FEC_LLRS,
};
static u32 pplm2ethtool_fec(u_long fec_mode, unsigned long size)
@@ -661,6 +663,8 @@ static const u32 pplm_fec_2_ethtool_linkmodes[] = {
[MLX5E_FEC_NOFEC] = ETHTOOL_LINK_MODE_FEC_NONE_BIT,
[MLX5E_FEC_FIRECODE] = ETHTOOL_LINK_MODE_FEC_BASER_BIT,
[MLX5E_FEC_RS_528_514] = ETHTOOL_LINK_MODE_FEC_RS_BIT,
+ [MLX5E_FEC_RS_544_514] = ETHTOOL_LINK_MODE_FEC_RS_BIT,
+ [MLX5E_FEC_LLRS_272_257_1] = ETHTOOL_LINK_MODE_FEC_LLRS_BIT,
};
static int get_fec_supported_advertised(struct mlx5_core_dev *dev,
@@ -680,6 +684,8 @@ static int get_fec_supported_advertised(struct mlx5_core_dev *dev,
ETHTOOL_LINK_MODE_FEC_BASER_BIT);
MLX5E_ADVERTISE_SUPPORTED_FEC(MLX5E_FEC_RS_528_514,
ETHTOOL_LINK_MODE_FEC_RS_BIT);
+ MLX5E_ADVERTISE_SUPPORTED_FEC(MLX5E_FEC_LLRS_272_257_1,
+ ETHTOOL_LINK_MODE_FEC_LLRS_BIT);
/* active fec is a bit set, find out which bit is set and
* advertise the corresponding ethtool bit
@@ -1510,7 +1516,7 @@ static int mlx5e_get_fecparam(struct net_device *netdev,
{
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5_core_dev *mdev = priv->mdev;
- u8 fec_configured = 0;
+ u16 fec_configured = 0;
u32 fec_active = 0;
int err;
@@ -1526,7 +1532,7 @@ static int mlx5e_get_fecparam(struct net_device *netdev,
return -EOPNOTSUPP;
fecparam->fec = pplm2ethtool_fec((u_long)fec_configured,
- sizeof(u8) * BITS_PER_BYTE);
+ sizeof(u16) * BITS_PER_BYTE);
return 0;
}
@@ -1536,12 +1542,12 @@ static int mlx5e_set_fecparam(struct net_device *netdev,
{
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5_core_dev *mdev = priv->mdev;
- u8 fec_policy = 0;
+ u16 fec_policy = 0;
int mode;
int err;
if (bitmap_weight((unsigned long *)&fecparam->fec,
- ETHTOOL_FEC_BASER_BIT + 1) > 1)
+ ETHTOOL_FEC_LLRS_BIT + 1) > 1)
return -EOPNOTSUPP;
for (mode = 0; mode < ARRAY_SIZE(pplm_fec_2_ethtool); mode++) {
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* [net-next V3 13/13] net/mlx5: Remove a useless 'drain_workqueue()' call in 'mlx5e_ipsec_cleanup()'
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
` (11 preceding siblings ...)
2020-02-11 22:32 ` [net-next V3 12/13] net/mlx5e: Add support for FEC modes based on 50G per lane links Saeed Mahameed
@ 2020-02-11 22:32 ` Saeed Mahameed
2020-02-11 23:21 ` [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Jeff Kirsher
2020-02-12 0:54 ` David Miller
14 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-11 22:32 UTC (permalink / raw)
To: David S. Miller, Jakub Kicinski
Cc: netdev, Christophe JAILLET, Saeed Mahameed
From: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
'destroy_workqueue()' already calls 'drain_workqueue()', there is no need
to call it explicitly.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c
index cf58c9637904..29626c6c9c25 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c
@@ -433,7 +433,6 @@ void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv)
if (!ipsec)
return;
- drain_workqueue(ipsec->wq);
destroy_workqueue(ipsec->wq);
ida_destroy(&ipsec->halloc);
--
2.24.1
^ permalink raw reply related [flat|nested] 19+ messages in thread* Re: [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
` (12 preceding siblings ...)
2020-02-11 22:32 ` [net-next V3 13/13] net/mlx5: Remove a useless 'drain_workqueue()' call in 'mlx5e_ipsec_cleanup()' Saeed Mahameed
@ 2020-02-11 23:21 ` Jeff Kirsher
2020-02-12 19:29 ` Saeed Mahameed
2020-02-12 0:54 ` David Miller
14 siblings, 1 reply; 19+ messages in thread
From: Jeff Kirsher @ 2020-02-11 23:21 UTC (permalink / raw)
To: Saeed Mahameed; +Cc: David S. Miller, Jakub Kicinski, netdev
On Tue, Feb 11, 2020 at 2:36 PM Saeed Mahameed <saeedm@mellanox.com> wrote:
>
> Hi Dave,
>
> This series adds some updates to mlx5 driver
> 1) Devlink health dump support for both rx and tx health reporters.
> 2) FEC modes supports.
> 3) two misc small patches.
>
> V3:
> - Improve ethtool patch "FEC LLRS" commit message as requested by
> Andrew Lunn.
> - Since we've missed the last cycle, dropped two small fixes patches,
> as they should go to net now.
>
> V2:
> - Remove "\n" from snprintf, happened due to rebase with a conflicting
> feature, Thanks Joe for spotting this.
>
> For more information please see tag log below.
>
> Please pull and let me know if there is any problem.
>
> Please note that the series starts with a merge of mlx5-next branch,
> to resolve and avoid dependency with rdma tree.
>
> Note about non-mlx5 change:
> For the FEC link modes support, Aya added the define for
> low latency Reed Solomon FEC as LLRS, in: include/uapi/linux/ethtool.h
>
> Thanks,
> Saeed.
Dave has not opened up net-next yet, or do you know something I don't...
^ permalink raw reply [flat|nested] 19+ messages in thread* Re: [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24
2020-02-11 23:21 ` [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Jeff Kirsher
@ 2020-02-12 19:29 ` Saeed Mahameed
0 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-12 19:29 UTC (permalink / raw)
To: jeffrey.t.kirsher; +Cc: davem, kuba, netdev
On Tue, 2020-02-11 at 15:21 -0800, Jeff Kirsher wrote:
> On Tue, Feb 11, 2020 at 2:36 PM Saeed Mahameed <saeedm@mellanox.com>
> wrote:
> > Hi Dave,
> >
> > This series adds some updates to mlx5 driver
> > 1) Devlink health dump support for both rx and tx health reporters.
> > 2) FEC modes supports.
> > 3) two misc small patches.
> >
> > V3:
> > - Improve ethtool patch "FEC LLRS" commit message as requested by
> > Andrew Lunn.
> > - Since we've missed the last cycle, dropped two small fixes
> > patches,
> > as they should go to net now.
> >
> > V2:
> > - Remove "\n" from snprintf, happened due to rebase with a
> > conflicting
> > feature, Thanks Joe for spotting this.
> >
> > For more information please see tag log below.
> >
> > Please pull and let me know if there is any problem.
> >
> > Please note that the series starts with a merge of mlx5-next
> > branch,
> > to resolve and avoid dependency with rdma tree.
> >
> > Note about non-mlx5 change:
> > For the FEC link modes support, Aya added the define for
> > low latency Reed Solomon FEC as LLRS, in:
> > include/uapi/linux/ethtool.h
> >
> > Thanks,
> > Saeed.
>
> Dave has not opened up net-next yet, or do you know something I
> don't...
My bad !
^ permalink raw reply [flat|nested] 19+ messages in thread
* Re: [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24
2020-02-11 22:32 [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Saeed Mahameed
` (13 preceding siblings ...)
2020-02-11 23:21 ` [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24 Jeff Kirsher
@ 2020-02-12 0:54 ` David Miller
2020-02-12 19:27 ` Saeed Mahameed
14 siblings, 1 reply; 19+ messages in thread
From: David Miller @ 2020-02-12 0:54 UTC (permalink / raw)
To: saeedm; +Cc: kuba, netdev
Deferred until net-next opens back up, resubmit this then.
Thank you.
^ permalink raw reply [flat|nested] 19+ messages in thread* Re: [pull request][net-next V3 00/13] Mellanox, mlx5 updates 2020-01-24
2020-02-12 0:54 ` David Miller
@ 2020-02-12 19:27 ` Saeed Mahameed
0 siblings, 0 replies; 19+ messages in thread
From: Saeed Mahameed @ 2020-02-12 19:27 UTC (permalink / raw)
To: davem; +Cc: kuba, netdev
On Tue, 2020-02-11 at 16:54 -0800, David Miller wrote:
> Deferred until net-next opens back up, resubmit this then.
>
Ohh shoot ! and i wondered why net-next is still quiet and rc1 is not
merged yet ! I had enough clues not to submit and yet I did..
My bad, will do better next time.
Thanks!
^ permalink raw reply [flat|nested] 19+ messages in thread