mlx5 updates 2023-04-20
Subsystems (6)
From: Saeed Mahameed <[email protected]>
This series provides misc updates.
For more information please see tag log below.
Please pull and let me know if there is any problem.
Thanks,
Saeed.
The following changes since commit e315e7b83a22043bffee450437d7089ef373cbf6:
net: libwx: fix memory leak in wx_setup_rx_resources (2023-04-20 15:39:15 +0200)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-updates-2023-04-20
for you to fetch changes up to f9c895a72a390656f9582e048fdcc3d2cec1dd7c:
net/mlx5: Update op_mode to op_mod for port selection (2023-04-20 18:35:50 -0700)
----------------------------------------------------------------
mlx5-updates-2023-04-20
1) Dragos Improves RX page pool, and provides some fixes to his previous series:
1.1) Fix releasing page_pool for striding RQ and legacy RQ nonlinear case 1.2) Hook NAPIs to page pools to gain more performance.
2) From Roi, Some cleanups to TC and eswitch modules.
3) Maher migrates vnic diagnostic counters reporting from debugfs to a dedicated devlink health reporter
Maher Says:
===========
net/mlx5: Expose vnic diagnostic counters using devlink
Currently, vnic diagnostic counters are exposed through the following debugfs:
$ ls /sys/kernel/debug/mlx5/0000:08:00.0/esw/vf_0/vnic_diag/ cq_overrun quota_exceeded_command total_q_under_processor_handle invalid_command send_queue_priority_update_flow nic_receive_steering_discard
The current design does not allow the hypervisor to view the diagnostic counters of its VFs, in case the VFs get bound to a VM. In other words, the counters are not exposed for representor interfaces. Furthermore, the debugfs design is inconvenient future-wise, in case more counters need to be reported by the driver in the future.
As these counters pertain to vNIC health, it is more appropriate to utilize the devlink health reporter to expose them.
Thus, this patchest includes the following changes:
* Drop the current vnic diagnostic counters debugfs interface.
* Add a vnic devlink health reporter for PFs/VFs core devices, which when diagnosed will dump vnic diagnostic counter values that are queried from FW.
* Add a vnic devlink health reporter for the representor interface, which serves the same purpose listed in the previous point, in addition to allowing the hypervisor to view its VFs diagnostic counters, even when the VFs are bounded to external VMs.
Example of devlink health reporter usage is:
$devlink health diagnose pci/0000:08:00.0 reporter vnic
vNIC env counters:
total_error_queues: 0 send_queue_priority_update_flow: 0
comp_eq_overrun: 0 async_eq_overrun: 0 cq_overrun: 0
invalid_command: 0 quota_exceeded_command: 0 nic_receive_steering_discard: 0
===========
4) SW steering fixes and improvements
Yevgeny Kliteynik Says:
=======================
These short patch series are just small fixes / improvements for SW steering:
- Patch 1: Fix dumping of legacy modify_hdr in debug dump to align to what is expected by parser
- Patch 2: Have separate threshold for ICM sync per ICM type
- Patch 3: Add more info to the steering debug dump - Linux version and device name
- Patch 4: Keep track of number of buddies that are currently in use per domain per buddy type
=======================
----------------------------------------------------------------
Dragos Tatulea (3):
net/mlx5e: RX, Fix releasing page_pool pages twice for striding RQ net/mlx5e: RX, Fix XDP_TX page release for legacy rq nonlinear case net/mlx5e: RX, Hook NAPIs to page pools
Eli Cohen (1):
net/mlx5: Include linux/pci.h for pci_msix_can_alloc_dyn()
Maher Sanalla (4):
Revert "net/mlx5: Expose steering dropped packets counter" Revert "net/mlx5: Expose vnic diagnostic counters for eswitch managed vports" net/mlx5: Add vnic devlink health reporter to PFs/VFs
net/mlx5e: Add vnic devlink health reporter to representors
Roi Dayan (3):
net/mlx5: E-Switch, Remove redundant dev arg from mlx5_esw_vport_alloc() net/mlx5: E-Switch, Remove unused mlx5_esw_offloads_vport_metadata_set() net/mlx5: Update op_mode to op_mod for port selection
Yevgeny Kliteynik (4):
net/mlx5: DR, Fix dumping of legacy modify_hdr in debug dump net/mlx5: DR, Calculate sync threshold of each pool according to its type net/mlx5: DR, Add more info in domain dbg dump
net/mlx5: DR, Add memory statistics for domain object
.../ethernet/mellanox/mlx5/devlink.rst | 33 ++++ drivers/net/ethernet/mellanox/mlx5/core/Makefile | 4 +- .../mellanox/mlx5/core/diag/reporter_vnic.c | 125 +++++++++++++ .../mellanox/mlx5/core/diag/reporter_vnic.h | 16 ++ drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 1 + drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 52 +++++- drivers/net/ethernet/mellanox/mlx5/core/en_rep.h | 1 + drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 11 +- .../net/ethernet/mellanox/mlx5/core/esw/debugfs.c | 198 --------------------- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 20 +-- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 6 - .../ethernet/mellanox/mlx5/core/eswitch_offloads.c | 25 --- drivers/net/ethernet/mellanox/mlx5/core/health.c | 4 + drivers/net/ethernet/mellanox/mlx5/core/main.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c | 1 + .../ethernet/mellanox/mlx5/core/steering/dr_dbg.c | 24 ++- .../mellanox/mlx5/core/steering/dr_icm_pool.c | 41 +++-- .../mellanox/mlx5/core/steering/dr_types.h | 3 + include/linux/mlx5/driver.h | 1 + include/linux/mlx5/mlx5_ifc.h | 2 +- 20 files changed, 297 insertions(+), 273 deletions(-) create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/diag/reporter_vnic.c create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/diag/reporter_vnic.h delete mode 100644 drivers/net/ethernet/mellanox/mlx5/core/esw/debugfs.c
PATCH 1/15 net/mlx5: DR, Fix dumping of legacy modify_hdr in debug dump
From: Yevgeny Kliteynik <[email protected]>
The steering dump parser expects to see 0 as rewrite num of actions in case pattern/args aren't supported - parsing of legacy modify header is based on this assumption.
Fix this to align to parser's expectation.
--- .../net/ethernet/mellanox/mlx5/core/steering/dr_dbg.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_dbg.c +6 -4
@@ -153,13 +153,15 @@ dr_dump_rule_action_mem(struct seq_file *file, const u64 rule_id,DR_DUMP_REC_TYPE_ACTION_MODIFY_HDR , action_id, rule_id, action->rewrite->index, action->rewrite->single_action_opt, - action->rewrite->num_of_actions, + ptrn_arg ? action->rewrite->num_of_actions :0 , ptrn_arg ? ptrn->index :0 , ptrn_arg ? mlx5dr_arg_get_obj_id(arg) : 0); -for (i =0 ; i < action->rewrite->num_of_actions; i++) { -seq_printf (file,",0x%016llx" , -be64_to_cpu (((__be64 *)rewrite_data)[i])); +if (ptrn_arg) { +for (i =0 ; i < action->rewrite->num_of_actions; i++) { +seq_printf (file,",0x%016llx" , +be64_to_cpu (((__be64 *)rewrite_data)[i])); + } }seq_puts (file,"\n" );
Signature · 1 line
2.39.2
PATCH 2/15 net/mlx5: DR, Calculate sync threshold of each pool according to its type
In reply to Saeed MahameedFrom: Yevgeny Kliteynik <[email protected]>
When certain ICM chunk is no longer needed, it needs to be freed. Fully freeing ICM memory involves issuing FW SYNC_STEERING command. This is very time consuming, and it is impractical to do it for every freed chunk.
Instead, we manage these 'freed' chunks in hot list (list of chunks that are not required by SW any more, but HW might still access them). When size of the hot list reaches certain threshold, we purge it and issue SYNC_STEERING FW command.
There is one threshold for all the different ICM types, which is not optimal, as different ICM types require different approach: STEs pool is very large, and it is very 'dynamic' in its nature, so letting hot list to become too large will result in a significant perf hiccup when purging the hot list. Modify action is much smaller and less dynamic, so we can let the hot list to grow to almost the size of the whole pool.
This patch fixes this problem: instead of having same hot memory threshold for all the pools, sync operation will be triggered in accordance with the ICM type.
--- .../mellanox/mlx5/core/steering/dr_icm_pool.c | 33 ++++++++++--------- 1 file changed, 18 insertions(+), 15 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c +18 -15
@@ -4,7 +4,9 @@#include "dr_types.h" #define DR_ICM_MODIFY_HDR_ALIGN_BASE 64 -#define DR_ICM_POOL_HOT_MEMORY_FRACTION 4 +#define DR_ICM_POOL_STE_HOT_MEM_PERCENT 25 +#define DR_ICM_POOL_MODIFY_HDR_PTRN_HOT_MEM_PERCENT 50 +#define DR_ICM_POOL_MODIFY_ACTION_HOT_MEM_PERCENT 90struct mlx5dr_icm_hot_chunk {struct mlx5dr_icm_buddy_mem *buddy_mem;
@@ -29,6 +31,8 @@ struct mlx5dr_icm_pool {struct mlx5dr_icm_hot_chunk *hot_chunks_arr;u32 hot_chunks_num;u64 hot_memory_size; +/* hot memory size threshold for triggering sync */ +u64 th; };struct mlx5dr_icm_dm {
@@ -330,15 +334,7 @@ dr_icm_chunk_init(struct mlx5dr_icm_chunk *chunk,static bool dr_icm_pool_is_sync_required (struct mlx5dr_icm_pool *pool) { -int allow_hot_size; - -/* sync when hot memory reaches a certain fraction of the pool size */ - allow_hot_size = -mlx5dr_icm_pool_chunk_size_to_byte (pool->max_log_chunk_sz, - pool->icm_type) / -DR_ICM_POOL_HOT_MEMORY_FRACTION ; - -return pool->hot_memory_size > allow_hot_size; +return pool->hot_memory_size > pool->th; }static void dr_icm_pool_clear_hot_chunks_arr (struct mlx5dr_icm_pool *pool)
@@ -503,8 +499,9 @@ void mlx5dr_icm_pool_free_htbl(struct mlx5dr_icm_pool *pool, struct mlx5dr_ste_hstruct mlx5dr_icm_pool *mlx5dr_icm_pool_create (struct mlx5dr_domain *dmn,enum mlx5dr_icm_type icm_type) { -u32 num_of_chunks, entry_size, max_hot_size; +u32 num_of_chunks, entry_size;struct mlx5dr_icm_pool *pool; +u32 max_hot_size =0 ; pool =kvzalloc (sizeof (*pool),GFP_KERNEL );if (!pool)
@@ -520,12 +517,21 @@ struct mlx5dr_icm_pool *mlx5dr_icm_pool_create(struct mlx5dr_domain *dmn,switch (icm_type) {case DR_ICM_TYPE_STE : pool->max_log_chunk_sz = dmn->info.max_log_sw_icm_sz; + max_hot_size =mlx5dr_icm_pool_chunk_size_to_byte (pool->max_log_chunk_sz, + pool->icm_type) * +DR_ICM_POOL_STE_HOT_MEM_PERCENT /100 ;break ;case DR_ICM_TYPE_MODIFY_ACTION : pool->max_log_chunk_sz = dmn->info.max_log_action_icm_sz; + max_hot_size =mlx5dr_icm_pool_chunk_size_to_byte (pool->max_log_chunk_sz, + pool->icm_type) * +DR_ICM_POOL_MODIFY_ACTION_HOT_MEM_PERCENT /100 ;break ;case DR_ICM_TYPE_MODIFY_HDR_PTRN : pool->max_log_chunk_sz = dmn->info.max_log_modify_hdr_pattern_icm_sz; + max_hot_size =mlx5dr_icm_pool_chunk_size_to_byte (pool->max_log_chunk_sz, + pool->icm_type) * +DR_ICM_POOL_MODIFY_HDR_PTRN_HOT_MEM_PERCENT /100 ;break ;default :WARN_ON (icm_type);
@@ -533,11 +539,8 @@ struct mlx5dr_icm_pool *mlx5dr_icm_pool_create(struct mlx5dr_domain *dmn, entry_size =mlx5dr_icm_pool_dm_type_to_entry_size (pool->icm_type); - max_hot_size =mlx5dr_icm_pool_chunk_size_to_byte (pool->max_log_chunk_sz, - pool->icm_type) / -DR_ICM_POOL_HOT_MEMORY_FRACTION ; - num_of_chunks =DIV_ROUND_UP (max_hot_size, entry_size) +1 ; + pool->th = max_hot_size; pool->hot_chunks_arr = kvcalloc(num_of_chunks, sizeof(struct mlx5dr_icm_hot_chunk),
Signature · 1 line
2.39.2
PATCH 3/15 net/mlx5: DR, Add more info in domain dbg dump
In reply to Saeed MahameedFrom: Yevgeny Kliteynik <[email protected]>
Add additinal items to domain info dump: Linux version and device name.
--- .../net/ethernet/mellanox/mlx5/core/steering/dr_dbg.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_dbg.c +9 -2
@@ -4,6 +4,7 @@#include <linux/debugfs.h> #include <linux/kernel.h> #include <linux/seq_file.h> +#include <linux/version.h> #include "dr_types.h" #define DR_DBG_PTR_TO_ID (p) ((u64)(uintptr_t)(p) & 0xFFFFFFFFULL)
@@ -632,9 +633,15 @@ dr_dump_domain(struct seq_file *file, struct mlx5dr_domain *dmn)u64 domain_id =DR_DBG_PTR_TO_ID (dmn);int ret; -seq_printf (file,"%d,0x%llx,%d,0%x,%d,%s\n" ,DR_DUMP_REC_TYPE_DOMAIN , +seq_printf (file,"%d,0x%llx,%d,0%x,%d,%u.%u.%u,%s,%d\n" , +DR_DUMP_REC_TYPE_DOMAIN , domain_id, dmn->type, dmn->info.caps.gvmi, - dmn->info.supp_sw_steering,pci_name (dmn->mdev->pdev)); + dmn->info.supp_sw_steering, +/* package version */ +LINUX_VERSION_MAJOR ,LINUX_VERSION_PATCHLEVEL , +LINUX_VERSION_SUBLEVEL , +pci_name (dmn->mdev->pdev), +0 );/* domain flags */ ret =dr_dump_domain_info (file, &dmn->info, domain_id);if (ret <0 )
Signature · 1 line
2.39.2
PATCH 4/15 net/mlx5: DR, Add memory statistics for domain object
In reply to Saeed MahameedFrom: Yevgeny Kliteynik <[email protected]>
Add counters for number of buddies that are currently in use per domain per buddy type (STE, MODIFY-HEADER, MODIFY-PATTERN).
--- drivers/net/ethernet/mellanox/mlx5/core/steering/dr_dbg.c | 7 +++++-- .../ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c | 8 +++++++- .../net/ethernet/mellanox/mlx5/core/steering/dr_types.h | 3 +++ 3 files changed, 15 insertions(+), 3 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_dbg.c +5 -2
@@ -633,7 +633,7 @@ dr_dump_domain(struct seq_file *file, struct mlx5dr_domain *dmn)u64 domain_id =DR_DBG_PTR_TO_ID (dmn);int ret; - seq_printf(file, "%d,0x%llx,%d,0%x,%d,%u.%u.%u,%s,%d\n", + seq_printf(file, "%d,0x%llx,%d,0%x,%d,%u.%u.%u,%s,%d,%u,%u,%u\n", DR_DUMP_REC_TYPE_DOMAIN, domain_id, dmn->type, dmn->info.caps.gvmi, dmn->info.supp_sw_steering,
@@ -641,7 +641,10 @@ dr_dump_domain(struct seq_file *file, struct mlx5dr_domain *dmn)LINUX_VERSION_MAJOR ,LINUX_VERSION_PATCHLEVEL ,LINUX_VERSION_SUBLEVEL ,pci_name (dmn->mdev->pdev), - 0); /* domain flags */ +0 ,/* domain flags */ + dmn->num_buddies[DR_ICM_TYPE_STE ], + dmn->num_buddies[DR_ICM_TYPE_MODIFY_ACTION ], + dmn->num_buddies[DR_ICM_TYPE_MODIFY_HDR_PTRN]); ret =dr_dump_domain_info (file, &dmn->info, domain_id);if (ret <0 )
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_icm_pool.c +7 -1
@@ -288,6 +288,8 @@ static int dr_icm_buddy_create(struct mlx5dr_icm_pool *pool)/* add it to the -start- of the list in order to search in it first */ list_add (&buddy->list_node, &pool->buddy_mem_list); + pool->dmn->num_buddies[pool->icm_type]++; +return 0 ; err_cleanup_buddy:
@@ -301,13 +303,17 @@ static int dr_icm_buddy_create(struct mlx5dr_icm_pool *pool)static void dr_icm_buddy_destroy (struct mlx5dr_icm_buddy_mem *buddy) { +enum mlx5dr_icm_type icm_type = buddy->pool->icm_type; +dr_icm_pool_mr_destroy (buddy->icm_mr);mlx5dr_buddy_cleanup (buddy); -if (buddy->pool->icm_type ==DR_ICM_TYPE_STE ) +if (icm_type ==DR_ICM_TYPE_STE )dr_icm_buddy_cleanup_ste_cache (buddy); + buddy->pool->dmn->num_buddies[icm_type]--; +kvfree (buddy); }
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h +3 -0
@@ -72,6 +72,7 @@ enum mlx5dr_icm_type {DR_ICM_TYPE_STE ,DR_ICM_TYPE_MODIFY_ACTION ,DR_ICM_TYPE_MODIFY_HDR_PTRN , + DR_ICM_TYPE_MAX, }; static inline enum mlx5dr_icm_chunk_size
@@ -955,6 +956,8 @@ struct mlx5dr_domain {struct list_head dbg_tbl_list;struct mlx5dr_dbg_dump_info dump_info;struct xarray definers_xa; +/* memory management statistics */ +u32 num_buddies[DR_ICM_TYPE_MAX ]; };struct mlx5dr_table_rx_tx {
Signature · 1 line
2.39.2
PATCH 5/15 Revert "net/mlx5: Expose steering dropped packets counter"
In reply to Saeed MahameedFrom: Maher Sanalla <[email protected]>
This reverts commit 4fe1b3a5f8fe2fdcedcaba9561e5b0ae5cb1d15b, which exposes the steering dropped packets counter via debugfs. The upcoming series will expose the counter via devlink health reporter instead of debugfs.
--- .../ethernet/mellanox/mlx5/core/esw/debugfs.c | 22 +++---------------- 1 file changed, 3 insertions(+), 19 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/esw/debugfs.c +3 -19
@@ -12,11 +12,10 @@ enum vnic_diag_counter {MLX5_VNIC_DIAG_CQ_OVERRUN ,MLX5_VNIC_DIAG_INVALID_COMMAND , MLX5_VNIC_DIAG_QOUTA_EXCEEDED_COMMAND, - MLX5_VNIC_DIAG_RX_STEERING_DISCARD, };static int mlx5_esw_query_vnic_diag (struct mlx5_vport *vport,enum vnic_diag_counter counter, -u64 *val) +u32 *val) {u32 out[MLX5_ST_SZ_DW (query_vnic_env_out)] = {};u32 in[MLX5_ST_SZ_DW (query_vnic_env_in)] = {};
@@ -58,10 +57,6 @@ static int mlx5_esw_query_vnic_diag(struct mlx5_vport *vport, enum vnic_diag_coucase MLX5_VNIC_DIAG_QOUTA_EXCEEDED_COMMAND : *val =MLX5_GET (vnic_diagnostic_statistics, vnic_diag_out, quota_exceeded_command);break ; -case MLX5_VNIC_DIAG_RX_STEERING_DISCARD : - *val =MLX5_GET64 (vnic_diagnostic_statistics, vnic_diag_out, - nic_receive_steering_discard); -break ; }return 0 ;
@@ -70,14 +65,14 @@ static int mlx5_esw_query_vnic_diag(struct mlx5_vport *vport, enum vnic_diag_coustatic int __show_vnic_diag (struct seq_file *file,struct mlx5_vport *vport,enum vnic_diag_counter type) { -u64 val =0 ; +u32 val =0 ;int ret; ret =mlx5_esw_query_vnic_diag (vport, type, &val);if (ret)return ret; -seq_printf (file,"%llu\n" , val); +seq_printf (file,"%d\n" , val);return 0 ; }
@@ -117,11 +112,6 @@ static int quota_exceeded_command_show(struct seq_file *file, void *priv)return __show_vnic_diag (file, file->private,MLX5_VNIC_DIAG_QOUTA_EXCEEDED_COMMAND ); } -static int rx_steering_discard_show (struct seq_file *file,void *priv) -{ -return __show_vnic_diag (file, file->private,MLX5_VNIC_DIAG_RX_STEERING_DISCARD ); -} -DEFINE_SHOW_ATTRIBUTE (total_q_under_processor_handle);DEFINE_SHOW_ATTRIBUTE (send_queue_priority_update_flow);DEFINE_SHOW_ATTRIBUTE (comp_eq_overrun);
@@ -129,7 +119,6 @@ DEFINE_SHOW_ATTRIBUTE(async_eq_overrun);DEFINE_SHOW_ATTRIBUTE (cq_overrun);DEFINE_SHOW_ATTRIBUTE (invalid_command);DEFINE_SHOW_ATTRIBUTE (quota_exceeded_command); -DEFINE_SHOW_ATTRIBUTE (rx_steering_discard);void mlx5_esw_vport_debugfs_destroy (struct mlx5_eswitch *esw,u16 vport_num) {
@@ -190,9 +179,4 @@ void mlx5_esw_vport_debugfs_create(struct mlx5_eswitch *esw, u16 vport_num, boolif (MLX5_CAP_GEN (esw->dev, quota_exceeded_count))debugfs_create_file ("quota_exceeded_command" ,0444 , vnic_diag, vport, "a_exceeded_command_fops); - -if (MLX5_CAP_GEN (esw->dev, nic_receive_steering_discard)) -debugfs_create_file ("rx_steering_discard" ,0444 , vnic_diag, vport, - &rx_steering_discard_fops); - }
Signature · 1 line
2.39.2
PATCH 6/15 Revert "net/mlx5: Expose vnic diagnostic counters for eswitch managed vports"
In reply to Saeed MahameedFrom: Maher Sanalla <[email protected]>
This reverts commit 606e6a72e29dff9e3341c4cc9b554420e4793f401 which exposes the vnic diagnostic counters via debugfs. Instead, The upcoming series will expose the same counters through devlink health reporter.
--- .../net/ethernet/mellanox/mlx5/core/Makefile | 2 +- .../ethernet/mellanox/mlx5/core/esw/debugfs.c | 182 ------------------ .../net/ethernet/mellanox/mlx5/core/eswitch.c | 6 - .../net/ethernet/mellanox/mlx5/core/eswitch.h | 5 - .../mellanox/mlx5/core/eswitch_offloads.c | 3 - 5 files changed, 1 insertion(+), 197 deletions(-) delete mode 100644 drivers/net/ethernet/mellanox/mlx5/core/esw/debugfs.c
drivers/net/ethernet/mellanox/mlx5/core/Makefile +1 -1
@@ -69,7 +69,7 @@ mlx5_core-$(CONFIG_MLX5_TC_SAMPLE) += en/tc/sample.o # mlx5_core-$(CONFIG_MLX5_ESWITCH) += eswitch.o eswitch_offloads.o eswitch_offloads_termtbl.o \ ecpf.o rdma.o esw/legacy.o \ - esw/debugfs.o esw/devlink_port.o esw/vporttbl.o esw/qos.o + esw/devlink_port.o esw/vporttbl.o esw/qos.o mlx5_core-$(CONFIG_MLX5_ESWITCH) += esw/acl/helper.o \ esw/acl/egress_lgcy.o esw/acl/egress_ofld.o \
drivers/net/ethernet/mellanox/mlx5/core/esw/debugfs.c deleted +0 -182
182 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +0 -6
@@ -36,7 +36,6 @@#include <linux/mlx5/vport.h> #include <linux/mlx5/fs.h> #include <linux/mlx5/mpfs.h> -#include <linux/debugfs.h> #include "esw/acl/lgcy.h" #include "esw/legacy.h" #include "esw/qos.h"
@@ -1056,7 +1055,6 @@ int mlx5_eswitch_load_vport(struct mlx5_eswitch *esw, u16 vport_num,if (err)return err; -mlx5_esw_vport_debugfs_create (esw, vport_num, false,0 ); err =esw_offloads_load_rep (esw, vport_num);if (err) goto err_rep;
@@ -1064,7 +1062,6 @@ int mlx5_eswitch_load_vport(struct mlx5_eswitch *esw, u16 vport_num,return err; err_rep: -mlx5_esw_vport_debugfs_destroy (esw, vport_num);mlx5_esw_vport_disable (esw, vport_num);return err; }
@@ -1072,7 +1069,6 @@ int mlx5_eswitch_load_vport(struct mlx5_eswitch *esw, u16 vport_num,void mlx5_eswitch_unload_vport (struct mlx5_eswitch *esw,u16 vport_num) {esw_offloads_unload_rep (esw, vport_num); -mlx5_esw_vport_debugfs_destroy (esw, vport_num);mlx5_esw_vport_disable (esw, vport_num); }
@@ -1672,7 +1668,6 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev) dev->priv.eswitch = esw;BLOCKING_INIT_NOTIFIER_HEAD (&esw->n_head); - esw->dbgfs =debugfs_create_dir ("esw" ,mlx5_debugfs_get_dev_root (esw->dev)); esw_info(dev, "Total vports %d, per vport: max uc(%d) max mc(%d)\n", esw->total_vports,
@@ -1696,7 +1691,6 @@ void mlx5_eswitch_cleanup(struct mlx5_eswitch *esw)esw_info (esw->dev,"cleanup\n" ); -debugfs_remove_recursive (esw->dbgfs); esw->dev->priv.eswitch =NULL ;destroy_workqueue (esw->work_queue);WARN_ON (refcount_read (&esw->qos.refcnt));
drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +0 -5
@@ -195,7 +195,6 @@ struct mlx5_vport {enum mlx5_eswitch_vport_event enabled_events;int index;struct devlink_port *dl_port; -struct dentry *dbgfs; };struct mlx5_esw_indir_table ;
@@ -343,7 +342,6 @@ struct mlx5_eswitch {u32 large_group_num; } params;struct blocking_notifier_head n_head; -struct dentry *dbgfs; };void esw_offloads_disable (struct mlx5_eswitch *esw);
@@ -704,9 +702,6 @@ int mlx5_esw_offloads_devlink_port_register(struct mlx5_eswitch *esw, u16 vport_void mlx5_esw_offloads_devlink_port_unregister (struct mlx5_eswitch *esw,u16 vport_num);struct devlink_port *mlx5_esw_offloads_devlink_port (struct mlx5_eswitch *esw,u16 vport_num); -void mlx5_esw_vport_debugfs_create (struct mlx5_eswitch *esw,u16 vport_num,bool is_sf,u16 sf_num); -void mlx5_esw_vport_debugfs_destroy (struct mlx5_eswitch *esw,u16 vport_num); -int mlx5_esw_devlink_sf_port_register (struct mlx5_eswitch *esw,struct devlink_port *dl_port,u16 vport_num,u32 controller,u32 sfnum);void mlx5_esw_devlink_sf_port_unregister (struct mlx5_eswitch *esw,u16 vport_num);
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +0 -3
@@ -3828,14 +3828,12 @@ int mlx5_esw_offloads_sf_vport_enable(struct mlx5_eswitch *esw, struct devlink_pif (err) goto devlink_err; -mlx5_esw_vport_debugfs_create (esw, vport_num, true, sfnum); err =mlx5_esw_offloads_rep_load (esw, vport_num);if (err) goto rep_err;return 0 ; rep_err: -mlx5_esw_vport_debugfs_destroy (esw, vport_num);mlx5_esw_devlink_sf_port_unregister (esw, vport_num); devlink_err:mlx5_esw_vport_disable (esw, vport_num);
@@ -3845,7 +3843,6 @@ int mlx5_esw_offloads_sf_vport_enable(struct mlx5_eswitch *esw, struct devlink_pvoid mlx5_esw_offloads_sf_vport_disable (struct mlx5_eswitch *esw,u16 vport_num) {mlx5_esw_offloads_rep_unload (esw, vport_num); -mlx5_esw_vport_debugfs_destroy (esw, vport_num);mlx5_esw_devlink_sf_port_unregister (esw, vport_num);mlx5_esw_vport_disable (esw, vport_num); }
Signature · 1 line
2.39.2
PATCH 7/15 net/mlx5: Add vnic devlink health reporter to PFs/VFs
In reply to Saeed MahameedFrom: Maher Sanalla <[email protected]>
Create a vnic devlink health reporter for PFs/VFs interfaces. The reporter's diagnose callback displays the values of vNIC/vport transport debug counters of PFs/VFs, as follows:
$ devlink health diagnose pci/0000:08:00.0 reporter vnic
vNIC env counters:
total_error_queues: 0 send_queue_priority_update_flow: 0
comp_eq_overrun: 0 async_eq_overrun: 0 cq_overrun: 0
invalid_command: 0 quota_exceeded_command: 0
nic_receive_steering_discard: 0Moreover, add documentation on the reporter functionality and the counters description.
While at it, expose the vNIC counters diagnose function to be used by the downstream patch, which will reveal the counters for representor interfaces.
--- .../ethernet/mellanox/mlx5/devlink.rst | 30 +++++ .../net/ethernet/mellanox/mlx5/core/Makefile | 2 +- .../mellanox/mlx5/core/diag/reporter_vnic.c | 125 ++++++++++++++++++ .../mellanox/mlx5/core/diag/reporter_vnic.h | 16 +++ .../net/ethernet/mellanox/mlx5/core/health.c | 4 + include/linux/mlx5/driver.h | 1 + 6 files changed, 177 insertions(+), 1 deletion(-) create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/diag/reporter_vnic.c create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/diag/reporter_vnic.h
Documentation/networking/device_drivers/ethernet/mellanox/mlx5/devlink.rst +30 -0
@@ -257,3 +257,33 @@ User commands examples: $ devlink health dump show pci/0000:82:00.1 reporter fw_fatal NOTE: This command can run only on PF. + +vnic reporter +------------- +The vnic reporter implements only the `diagnose` callback. +It is responsible for querying the vnic diagnostic counters from fw and displaying +them in realtime. + +Description of the vnic counters: +total_q_under_processor_handle: number of queues in an error state due to +an async error or errored command. +send_queue_priority_update_flow: number of QP/SQ priority/SL update +events. +cq_overrun: number of times CQ entered an error state due to an +overflow. +async_eq_overrun: number of times an EQ mapped to async events was +overrun. +comp_eq_overrun: number of times an EQ mapped to completion events was +overrun. +quota_exceeded_command: number of commands issued and failed due to quota +exceeded. +invalid_command: number of commands issued and failed dues to any reason +other than quota exceeded. +nic_receive_steering_discard: number of packets that completed RX flow +steering but were discarded due to a mismatch in flow table. + +User commands examples: +- Diagnose PF/VF vnic counters + $ devlink health diagnose pci/0000:82:00.1 reporter vnic + +NOTE: This command can run only on PF/VF ports.
drivers/net/ethernet/mellanox/mlx5/core/Makefile +1 -1
@@ -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 fs_ft_pool.o rl.o lag/debugfs.o lag/lag.o dev.o events.o wq.o lib/gid.o \ lib/devcom.o lib/pci_vsc.o lib/dm.o lib/fs_ttc.o diag/fs_tracepoint.o \ - diag/fw_tracer.o diag/crdump.o devlink.o diag/rsc_dump.o \ + diag/fw_tracer.o diag/crdump.o devlink.o diag/rsc_dump.o diag/reporter_vnic.o \ fw_reset.o qos.o lib/tout.o lib/aso.o #
drivers/net/ethernet/mellanox/mlx5/core/diag/reporter_vnic.c new file +125 -0
125 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/diag/reporter_vnic.h new file +16 -0
@@ -0,0 +1,16 @@ +/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB +* Copyright (c) 2023, NVIDIA CORPORATION & AFFILIATES. +*/ +#ifndef __MLX5_REPORTER_VNIC_H +#define __MLX5_REPORTER_VNIC_H + +#include "mlx5_core.h" + +void mlx5_reporter_vnic_create (struct mlx5_core_dev *dev); +void mlx5_reporter_vnic_destroy (struct mlx5_core_dev *dev); + +int mlx5_reporter_vnic_diagnose_counters (struct mlx5_core_dev *dev, +struct devlink_fmsg *fmsg, +u16 vport_num,bool other_vport); + +#endif /* __MLX5_REPORTER_VNIC_H */
drivers/net/ethernet/mellanox/mlx5/core/health.c +4 -0
@@ -42,6 +42,7 @@#include "lib/pci_vsc.h" #include "lib/tout.h" #include "diag/fw_tracer.h" +#include "diag/reporter_vnic.h" enum {MAX_MISSES =3 ,
@@ -898,6 +899,7 @@ void mlx5_health_cleanup(struct mlx5_core_dev *dev)cancel_delayed_work_sync (&health->update_fw_log_ts_work);destroy_workqueue (health->wq); +mlx5_reporter_vnic_destroy (dev);mlx5_fw_reporters_destroy (dev); }
@@ -907,6 +909,7 @@ int mlx5_health_init(struct mlx5_core_dev *dev)char *name;mlx5_fw_reporters_create (dev); +mlx5_reporter_vnic_create (dev); health = &dev->priv.health; name =kmalloc (64 ,GFP_KERNEL );
@@ -926,6 +929,7 @@ int mlx5_health_init(struct mlx5_core_dev *dev)return 0 ; out_err: +mlx5_reporter_vnic_destroy (dev);mlx5_fw_reporters_destroy (dev);return -ENOMEM ; }
include/linux/mlx5/driver.h +1 -0
@@ -439,6 +439,7 @@ struct mlx5_core_health {struct work_struct report_work;struct devlink_health_reporter *fw_reporter;struct devlink_health_reporter *fw_fatal_reporter; +struct devlink_health_reporter *vnic_reporter;struct delayed_work update_fw_log_ts_work; };
Signature · 1 line
2.39.2
PATCH 8/15 net/mlx5e: Add vnic devlink health reporter to representors
In reply to Saeed MahameedFrom: Maher Sanalla <[email protected]>
Create a new devlink health reporter for representor interface, which reports the values of representor vnic diagnostic counters when diagnosed.
This patch will allow admins to monitor VF diagnostic counters through the representor-interface vnic reporter.
Example of usage:
$ devlink health diagnose pci/0000:08:00.0/65537 reporter vnic
vNIC env counters:
total_error_queues: 0 send_queue_priority_update_flow: 0
comp_eq_overrun: 0 async_eq_overrun: 0 cq_overrun: 0
invalid_command: 0 quota_exceeded_command: 0
nic_receive_steering_discard: 0--- .../ethernet/mellanox/mlx5/devlink.rst | 5 +- .../net/ethernet/mellanox/mlx5/core/en_rep.c | 52 ++++++++++++++++++- .../net/ethernet/mellanox/mlx5/core/en_rep.h | 1 + 3 files changed, 55 insertions(+), 3 deletions(-)
Documentation/networking/device_drivers/ethernet/mellanox/mlx5/devlink.rst +4 -1
@@ -285,5 +285,8 @@ steering but were discarded due to a mismatch in flow table. User commands examples: - Diagnose PF/VF vnic counters $ devlink health diagnose pci/0000:82:00.1 reporter vnic +- Diagnose representor vnic counters (performed by supplying devlink port of the + representor, which can be obtained via devlink port command) + $ devlink health diagnose pci/0000:82:00.1/65537 reporter vnic -NOTE: This command can run only on PF/VF ports. +NOTE: This command can run over all interfaces such as PF/VF and representor ports.
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +50 -2
84 lines not on this page · raw message
drivers/net/ethernet/mellanox/mlx5/core/en_rep.h +1 -0
@@ -118,6 +118,7 @@ struct mlx5e_rep_priv {struct rtnl_link_stats64 prev_vf_vport_stats;struct mlx5_flow_handle *send_to_vport_meta_rule;struct rhashtable tc_ht; +struct devlink_health_reporter *rep_vnic_reporter; }; static inline
Signature · 1 line
2.39.2
PATCH 9/15 net/mlx5e: RX, Fix releasing page_pool pages twice for striding RQ
In reply to Saeed MahameedFrom: Dragos Tatulea <[email protected]>
mlx5e_free_rx_descs is responsible for calling the dealloc_wqe op which returns pages to the page_pool. This can happen during flush or close. For XSK, the regular RQ is flushed (when replaced by the XSK RQ) and also closed later. This is normally not a problem as the wqe list is empty on a second call to mlx5e_free_rx_descs. However, for striding RQ, the previously released wqes from the list will appear as missing and will be released a second time by mlx5e_free_rx_missing_descs.
This patch sets the no release bits on the striding RQ wqes in the dealloc_wqe op to prevent releasing the pages a second time.
Please note that the bits are set only in the control path during close and not in the data path.
--- drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 5 +++++ 1 file changed, 5 insertions(+)
drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +5 -0
@@ -861,6 +861,11 @@ static void mlx5e_dealloc_rx_mpwqe(struct mlx5e_rq *rq, u16 ix)struct mlx5e_mpw_info *wi =mlx5e_get_mpw_info (rq, ix);/* This function is called on rq/netdev close. */ mlx5e_free_rx_mpwqe (rq, wi); + +/* Avoid a second release of the wqe pages: dealloc is called also +* for missing wqes on an already flushed RQ. +*/ +bitmap_fill (wi->skip_release_bitmap, rq->mpwqe.pages_per_wqe); } INDIRECT_CALLABLE_SCOPE bool mlx5e_post_rx_wqes(struct mlx5e_rq *rq)
Signature · 1 line
2.39.2
PATCH 10/15 net/mlx5e: RX, Fix XDP_TX page release for legacy rq nonlinear case
In reply to Saeed MahameedFrom: Dragos Tatulea <[email protected]>
When the XDP handler marks the data for transmission (XDP_TX), it is incorrect to release the page fragment. Instead, the fragments should be marked as MLX5E_WQE_FRAG_SKIP_RELEASE because XDP will release the page directly to the page_pool (page_pool_put_defragged_page) after TX completion.
The linear case already does this. This patch fixes the nonlinear part as well.
Also, the looping over the fragments was incorrect: When handling pages after XDP_TX in the legacy rq nonlinear case, the loop was skipping the first wqe fragment.
--- drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +3 -3
@@ -1746,10 +1746,10 @@ mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi prog =rcu_dereference (rq->xdp_prog);if (prog &&mlx5e_xdp_handle (rq, prog, &mxbuf)) {if (test_bit (MLX5E_RQ_FLAG_XDP_XMIT , rq->flags)) { -int i; +struct mlx5e_wqe_frag_info *pwi; -for (i = wi - head_wi; i < rq->wqe.info.num_frags; i++) -mlx5e_put_rx_frag (rq, &head_wi[i]); +for (pwi = head_wi; pwi < wi; pwi++) + pwi->flags |=BIT (MLX5E_WQE_FRAG_SKIP_RELEASE ); }return NULL ;/* page/packet was consumed by XDP */ }
Signature · 1 line
2.39.2
PATCH 11/15 net/mlx5e: RX, Hook NAPIs to page pools
In reply to Saeed MahameedFrom: Dragos Tatulea <[email protected]>
Linking the NAPI to the rq page_pool to improve page_pool cache usage during skb recycling.
Here are the observed improvements for a iperf single stream test case:
- For 1500 MTU and legacy rq, seeing a 20% improvement of cache usage.
- For 9K MTU, seeing 33-40 % page_pool cache usage improvements for both striding and legacy rq (depending if the application is running on the same core as the rq or not).
--- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 1 + 1 file changed, 1 insertion(+)
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +1 -0
@@ -857,6 +857,7 @@ static int mlx5e_alloc_rq(struct mlx5e_params *params, pp_params.pool_size = pool_size; pp_params.nid = node; pp_params.dev = rq->pdev; + pp_params.napi = rq->cq.napi; pp_params.dma_dir = rq->buff.map_dir; pp_params.max_len =PAGE_SIZE ;
Signature · 1 line
2.39.2
PATCH 12/15 net/mlx5: Include linux/pci.h for pci_msix_can_alloc_dyn()
In reply to Saeed MahameedFrom: Eli Cohen <[email protected]>
Add include directive to assure pci_msix_can_alloc_dyn() prototype.
--- drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c | 1 + 1 file changed, 1 insertion(+)
drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c +1 -0
@@ -1,6 +1,7 @@// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB /* Copyright (c) 2019 Mellanox Technologies. */ +#include <linux/pci.h> #include <linux/interrupt.h> #include <linux/notifier.h> #include <linux/mlx5/driver.h>
Signature · 1 line
2.39.2
PATCH 13/15 net/mlx5: E-Switch, Remove redundant dev arg from mlx5_esw_vport_alloc()
In reply to Saeed MahameedFrom: Roi Dayan <[email protected]>
The passded esw->dev is redundant as esw being passed and esw->dev being used inside.
--- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +7 -7
@@ -1506,7 +1506,7 @@ int mlx5_esw_sf_max_hpf_functions(struct mlx5_core_dev *dev, u16 *max_sfs, u16 *return err; } -static int mlx5_esw_vport_alloc (struct mlx5_eswitch *esw,struct mlx5_core_dev *dev, +static int mlx5_esw_vport_alloc (struct mlx5_eswitch *esw,int index,u16 vport_num) {struct mlx5_vport *vport;
@@ -1560,7 +1560,7 @@ static int mlx5_esw_vports_init(struct mlx5_eswitch *esw)xa_init (&esw->vports); - err =mlx5_esw_vport_alloc (esw, dev, idx,MLX5_VPORT_PF ); + err =mlx5_esw_vport_alloc (esw, idx,MLX5_VPORT_PF );if (err) goto err;if (esw->first_host_vport ==MLX5_VPORT_PF )
@@ -1568,7 +1568,7 @@ static int mlx5_esw_vports_init(struct mlx5_eswitch *esw) idx++;for (i =0 ; i <mlx5_core_max_vfs (dev); i++) { - err =mlx5_esw_vport_alloc (esw, dev, idx, idx); + err =mlx5_esw_vport_alloc (esw, idx, idx);if (err) goto err;xa_set_mark (&esw->vports, idx,MLX5_ESW_VPT_VF );
@@ -1577,7 +1577,7 @@ static int mlx5_esw_vports_init(struct mlx5_eswitch *esw) } base_sf_num =mlx5_sf_start_function_id (dev);for (i =0 ; i <mlx5_sf_max_functions (dev); i++) { - err =mlx5_esw_vport_alloc (esw, dev, idx, base_sf_num + i); + err =mlx5_esw_vport_alloc (esw, idx, base_sf_num + i);if (err) goto err;xa_set_mark (&esw->vports, base_sf_num + i,MLX5_ESW_VPT_SF );
@@ -1588,7 +1588,7 @@ static int mlx5_esw_vports_init(struct mlx5_eswitch *esw)if (err) goto err;for (i =0 ; i < max_host_pf_sfs; i++) { - err =mlx5_esw_vport_alloc (esw, dev, idx, base_sf_num + i); + err =mlx5_esw_vport_alloc (esw, idx, base_sf_num + i);if (err) goto err;xa_set_mark (&esw->vports, base_sf_num + i,MLX5_ESW_VPT_SF );
@@ -1596,12 +1596,12 @@ static int mlx5_esw_vports_init(struct mlx5_eswitch *esw) }if (mlx5_ecpf_vport_exists (dev)) { - err =mlx5_esw_vport_alloc (esw, dev, idx,MLX5_VPORT_ECPF ); + err =mlx5_esw_vport_alloc (esw, idx,MLX5_VPORT_ECPF );if (err) goto err; idx++; } - err =mlx5_esw_vport_alloc (esw, dev, idx,MLX5_VPORT_UPLINK ); + err =mlx5_esw_vport_alloc (esw, idx,MLX5_VPORT_UPLINK );if (err) goto err;return 0 ;
Signature · 1 line
2.39.2
PATCH 14/15 net/mlx5: E-Switch, Remove unused mlx5_esw_offloads_vport_metadata_set()
In reply to Saeed MahameedFrom: Roi Dayan <[email protected]>
Remove unused function which also seems a duplicate of esw_port_metadata_set().
--- .../net/ethernet/mellanox/mlx5/core/eswitch.h | 1 - .../mellanox/mlx5/core/eswitch_offloads.c | 22 ------------------- 2 files changed, 23 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +0 -1
@@ -354,7 +354,6 @@ mlx5_eswitch_add_send_to_vport_meta_rule(struct mlx5_eswitch *esw, u16 vport_numvoid mlx5_eswitch_del_send_to_vport_meta_rule (struct mlx5_flow_handle *rule);bool mlx5_esw_vport_match_metadata_supported (const struct mlx5_eswitch *esw); -int mlx5_esw_offloads_vport_metadata_set (struct mlx5_eswitch *esw,bool enable);u32 mlx5_esw_match_metadata_alloc (struct mlx5_eswitch *esw);void mlx5_esw_match_metadata_free (struct mlx5_eswitch *esw,u32 metadata);
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +0 -22
@@ -2939,28 +2939,6 @@ static int esw_offloads_metadata_init(struct mlx5_eswitch *esw)return err; } -int mlx5_esw_offloads_vport_metadata_set (struct mlx5_eswitch *esw,bool enable) -{ -int err =0 ; - -down_write (&esw->mode_lock); -if (mlx5_esw_is_fdb_created (esw)) { - err = -EBUSY ; - goto done; - } -if (!mlx5_esw_vport_match_metadata_supported (esw)) { - err = -EOPNOTSUPP ; - goto done; - } -if (enable) - esw->flags |=MLX5_ESWITCH_VPORT_MATCH_METADATA ; -else - esw->flags &= ~MLX5_ESWITCH_VPORT_MATCH_METADATA ; -done: -up_write (&esw->mode_lock); -return err; -} -int esw_vport_create_offloads_acl_tables (struct mlx5_eswitch *esw,struct mlx5_vport *vport)
Signature · 1 line
2.39.2
PATCH 15/15 net/mlx5: Update op_mode to op_mod for port selection
In reply to Saeed MahameedFrom: Roi Dayan <[email protected]>
To be consistent with the other enum keys use OP_MOD instead of OP_MODE.
--- drivers/net/ethernet/mellanox/mlx5/core/main.c | 2 +- include/linux/mlx5/mlx5_ifc.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/main.c +1 -1
@@ -717,7 +717,7 @@ static int handle_hca_cap_port_selection(struct mlx5_core_dev *dev, MLX5_ST_SZ_BYTES(port_selection_cap));MLX5_SET (port_selection_cap, set_hca_cap, port_select_flow_table_bypass,1 ); - err =set_caps (dev, set_ctx,MLX5_SET_HCA_CAP_OP_MODE_PORT_SELECTION ); + err =set_caps (dev, set_ctx,MLX5_SET_HCA_CAP_OP_MOD_PORT_SELECTION );return err; }
include/linux/mlx5/mlx5_ifc.h +1 -1
@@ -69,7 +69,7 @@ enum {MLX5_SET_HCA_CAP_OP_MOD_ATOMIC =0x3 ,MLX5_SET_HCA_CAP_OP_MOD_ROCE =0x4 ,MLX5_SET_HCA_CAP_OP_MOD_GENERAL_DEVICE2 =0x20 , - MLX5_SET_HCA_CAP_OP_MODE_PORT_SELECTION = 0x25, + MLX5_SET_HCA_CAP_OP_MOD_PORT_SELECTION = 0x25, }; enum {
Signature · 1 line
2.39.2
Jakub Kicinski
maintainerRe: [net-next 11/15] net/mlx5e: RX, Hook NAPIs to page pools
In reply to Saeed MahameedQuoted 5 lines · On Thu, 20 Apr 2023 18:38:46 -0700 Saeed Mahameed wrote:
From: Dragos Tatulea <[email protected]>
Linking the NAPI to the rq
page_poolto improvepage_poolcache usage during skb recycling.
Quoted 8 lines
Here are the observed improvements for a iperf single stream test case:
- For 1500 MTU and legacy rq, seeing a 20% improvement of cache usage.
- For 9K MTU, seeing 33-40 %
page_poolcache usage improvements for both striding and legacy rq (depending if the application is running on the same core as the rq or not).
I think you'll need a strategically placed page_pool_unlink_napi() once
https://lore.kernel.org/all/[email protected]/
gets merged (which should me in minutes). Would you be able to follow up on this tomorrow?
Re: [net-next 01/15] net/mlx5: DR, Fix dumping of legacy modify_hdr in debug dump
In reply to Saeed MahameedThis series was applied to netdev/net-next.git (main) by Saeed Mahameed <[email protected]>:
Hello:
This series was applied to netdev/net-next.git (main) by Saeed Mahameed <[email protected]>:
Quoted 8 lines · On Thu, 20 Apr 2023 18:38:36 -0700 you wrote:
From: Yevgeny Kliteynik <[email protected]>
The steering dump parser expects to see 0 as rewrite num of actions in case pattern/args aren't supported - parsing of legacy modify header is based on this assumption.
Fix this to align to parser's expectation.[...]
Here is the summary with links:
- [net-next,01/15] net/mlx5: DR, Fix dumping of legacy modify_hdr in debug dump
https://git.kernel.org/netdev/net-next/c/075056005d8c - [net-next,02/15] net/mlx5: DR, Calculate sync threshold of each pool according to its type https://git.kernel.org/netdev/net-next/c/72b2cff68405 - [net-next,03/15] net/mlx5: DR, Add more info in domain dbg dump https://git.kernel.org/netdev/net-next/c/cedb6665bc33 - [net-next,04/15] net/mlx5: DR, Add memory statistics for domain object https://git.kernel.org/netdev/net-next/c/57295e069cd8 - [net-next,05/15] Revert "net/mlx5: Expose steering dropped packets counter" https://git.kernel.org/netdev/net-next/c/e267b8a52ca5 - [net-next,06/15] Revert "net/mlx5: Expose vnic diagnostic counters for eswitch managed vports" https://git.kernel.org/netdev/net-next/c/0a431418f685 - [net-next,07/15] net/mlx5: Add vnic devlink health reporter to PFs/VFs https://git.kernel.org/netdev/net-next/c/b0bc615df488 - [net-next,08/15] net/mlx5e: Add vnic devlink health reporter to representors https://git.kernel.org/netdev/net-next/c/cf14af140a5a - [net-next,09/15] net/mlx5e: RX, Fix releasing page_pool pages twice for striding RQ https://git.kernel.org/netdev/net-next/c/c8e9090233a7 - [net-next,10/15] net/mlx5e: RX, Fix XDP_TX page release for legacy rq nonlinear case https://git.kernel.org/netdev/net-next/c/40afb3b14496 - [net-next,11/15] net/mlx5e: RX, Hook NAPIs to page pools https://git.kernel.org/netdev/net-next/c/a880f814739c - [net-next,12/15] net/mlx5: Include linux/pci.h for pci_msix_can_alloc_dyn() https://git.kernel.org/netdev/net-next/c/45e261b7b821 - [net-next,13/15] net/mlx5: E-Switch, Remove redundant dev arg from mlx5_esw_vport_alloc() https://git.kernel.org/netdev/net-next/c/8ca52ada6267 - [net-next,14/15] net/mlx5: E-Switch, Remove unused mlx5_esw_offloads_vport_metadata_set() https://git.kernel.org/netdev/net-next/c/38d9a740f68d - [net-next,15/15] net/mlx5: Update op_mode to op_mod for port selection https://git.kernel.org/netdev/net-next/c/f9c895a72a39
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html
Re: [net-next 11/15] net/mlx5e: RX, Hook NAPIs to page pools
In reply to Jakub KicinskiQuoted 17 lines · Jakub Kicinski
Quoted 8 lines · On Thu, 20 Apr 2023 18:38:46 -0700 Saeed Mahameed wrote:
From: Dragos Tatulea <[email protected]>
Linking the NAPI to the rq
page_poolto improvepage_poolcache usage during skb recycling.Here are the observed improvements for a iperf single stream test case:
Quoted 7 lines
- For 1500 MTU and legacy rq, seeing a 20% improvement of cache usage.
- For 9K MTU, seeing 33-40 %
page_poolcache usage improvements for both striding and legacy rq (depending if the application is running on
the same core as the rq or not).
Quoted 5 lines
I think you'll need a strategically placed
page_pool_unlink_napi()once
https://lore.kernel.org/all/[email protected]/
gets merged (which should me in minutes). Would you be able to follow up on this tomorrow?
Thanks for the tip Jakub.
There's no "swap" stage in mlx5 (the page pool is destroyed while NAPI is still disabled). So I think the page_pool_unlink_napi that you added in page_pool_destroy is sufficient.
Thanks,
Dragos