mlx5 fixes 2023-04-19
From: Saeed Mahameed <[email protected]>
V1->v2:
- #3, Remove "inline" keyword from a function declaration in c file.
- #1, FYI Automation is reporting "author Signed-off-by missing" but it is not ! I manually checked, so nothing was changed in patch 1
This series provides bug fixes to mlx5 driver. Please pull and let me know if there is any problem.
Thanks,
Saeed.
The following changes since commit 0f2a4af27b649c13ba76431552fe49c60120d0f6:
wifi: ath9k: Don't mark channelmap stack variable read-only in ath9k_mci_update_wlan_channels() (2023-04-20 15:26:37 -0700)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-fixes-2023-04-20
for you to fetch changes up to 081abcacaf0a9505c0d4cc9780b12e6ce5d33630:
Revert "net/mlx5e: Don't use termination table when redundant" (2023-04-20 18:47:33 -0700)
----------------------------------------------------------------
mlx5-fixes-2023-04-20
----------------------------------------------------------------
Aya Levin (1):
net/mlx5e: Nullify table pointer when failing to create
Chris Mi (3):
net/mlx5: E-switch, Create per vport table based on devlink encap mode net/mlx5: E-switch, Don't destroy indirect table in split rule net/mlx5: Release tunnel device after tc update skb
Moshe Shemesh (2):
Revert "net/mlx5: Remove "recovery" arg from mlx5_load_one() function" net/mlx5: Use recovery timeout on sync reset flow
Roi Dayan (1):
net/mlx5e: Fix error flow in representor failing to add vport rx rule
Vlad Buslov (3):
net/mlx5e: Don't clone flow post action attributes second time net/mlx5e: Release the label when replacing existing ct entry Revert "net/mlx5e: Don't use termination table when redundant"
drivers/net/ethernet/mellanox/mlx5/core/devlink.c | 2 +- .../net/ethernet/mellanox/mlx5/core/en/rep/tc.c | 1 + .../ethernet/mellanox/mlx5/core/en/tc/post_act.c | 11 ++------ .../ethernet/mellanox/mlx5/core/en/tc/post_act.h | 2 +- .../net/ethernet/mellanox/mlx5/core/en/tc/sample.c | 4 +-- drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c | 1 + drivers/net/ethernet/mellanox/mlx5/core/en_fs.c | 5 +++- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 1 + drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 2 ++ .../net/ethernet/mellanox/mlx5/core/esw/vporttbl.c | 12 +++++++- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 2 +- .../ethernet/mellanox/mlx5/core/eswitch_offloads.c | 4 +-- .../mellanox/mlx5/core/eswitch_offloads_termtbl.c | 32 +++------------------- drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c | 4 +-- drivers/net/ethernet/mellanox/mlx5/core/main.c | 9 +++--- .../net/ethernet/mellanox/mlx5/core/mlx5_core.h | 2 +- 16 files changed, 40 insertions(+), 54 deletions(-)
PATCH v2 1/10 net/mlx5e: Don't clone flow post action attributes second time
From: Vlad Buslov <[email protected]>
The code already clones post action attributes in mlx5e_clone_flow_attr_for_post_act(). Creating another copy in mlx5e_tc_post_act_add() is a erroneous leftover from original implementation. Instead, assign handle->attribute to post_attr provided by the caller. Note that cloning the attribute second time is not just wasteful but also causes issues like second copy not being properly updated in neigh update code which leads to following use-after-free:
Feb 21 09:02:00 c-237-177-40-045 kernel: BUG: KASAN: use-after-free in mlx5_cmd_set_fte+0x200d/0x24c0 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: kasan_report+0xbb/0x1a0 Feb 21 09:02:00 c-237-177-40-045 kernel: kasan_save_stack+0x1e/0x40 Feb 21 09:02:00 c-237-177-40-045 kernel: kasan_set_track+0x21/0x30 Feb 21 09:02:00 c-237-177-40-045 kernel: __kasan_kmalloc+0x7a/0x90 Feb 21 09:02:00 c-237-177-40-045 kernel: kasan_save_stack+0x1e/0x40 Feb 21 09:02:00 c-237-177-40-045 kernel: kasan_set_track+0x21/0x30 Feb 21 09:02:00 c-237-177-40-045 kernel: kasan_save_free_info+0x2a/0x40 Feb 21 09:02:00 c-237-177-40-045 kernel: ____kasan_slab_free+0x11a/0x1b0 Feb 21 09:02:00 c-237-177-40-045 kernel: page dumped because: kasan: bad access detected Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5_core 0000:08:00.0: mlx5_cmd_out_err:803:(pid 8833): SET_FLOW_TABLE_ENTRY(0x936) op_mod(0x0) failed, status bad resource state(0x9), syndrome (0xf2ff71), err(-22) Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5_core 0000:08:00.0 enp8s0f0: Failed to add post action rule Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5_core 0000:08:00.0: mlx5e_tc_encap_flows_add:190:(pid 8833): Failed to update flow post acts, -22 Feb 21 09:02:00 c-237-177-40-045 kernel: Call Trace: Feb 21 09:02:00 c-237-177-40-045 kernel: <TASK> Feb 21 09:02:00 c-237-177-40-045 kernel: dump_stack_lvl+0x57/0x7d Feb 21 09:02:00 c-237-177-40-045 kernel: print_report+0x170/0x471 Feb 21 09:02:00 c-237-177-40-045 kernel: ? mlx5_cmd_set_fte+0x200d/0x24c0 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: kasan_report+0xbb/0x1a0 Feb 21 09:02:00 c-237-177-40-045 kernel: ? mlx5_cmd_set_fte+0x200d/0x24c0 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5_cmd_set_fte+0x200d/0x24c0 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: ? __module_address.part.0+0x62/0x200 Feb 21 09:02:00 c-237-177-40-045 kernel: ? mlx5_cmd_stub_create_flow_table+0xd0/0xd0 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: ? __raw_spin_lock_init+0x3b/0x110 Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5_cmd_create_fte+0x80/0xb0 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: add_rule_fg+0xe80/0x19c0 [mlx5_core] -- Feb 21 09:02:00 c-237-177-40-045 kernel: Allocated by task 13476: Feb 21 09:02:00 c-237-177-40-045 kernel: kasan_save_stack+0x1e/0x40 Feb 21 09:02:00 c-237-177-40-045 kernel: kasan_set_track+0x21/0x30 Feb 21 09:02:00 c-237-177-40-045 kernel: __kasan_kmalloc+0x7a/0x90 Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5_packet_reformat_alloc+0x7b/0x230 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5e_tc_tun_create_header_ipv4+0x977/0xf10 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5e_attach_encap+0x15b4/0x1e10 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: post_process_attr+0x305/0xa30 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5e_tc_add_fdb_flow+0x4c0/0xcf0 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: __mlx5e_add_fdb_flow+0x7cf/0xe90 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5e_configure_flower+0xcaa/0x4b90 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5e_rep_setup_tc_cls_flower+0x99/0x1b0 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5e_rep_setup_tc_cb+0x133/0x1e0 [mlx5_core] -- Feb 21 09:02:00 c-237-177-40-045 kernel: Freed by task 8833: Feb 21 09:02:00 c-237-177-40-045 kernel: kasan_save_stack+0x1e/0x40 Feb 21 09:02:00 c-237-177-40-045 kernel: kasan_set_track+0x21/0x30 Feb 21 09:02:00 c-237-177-40-045 kernel: kasan_save_free_info+0x2a/0x40 Feb 21 09:02:00 c-237-177-40-045 kernel: ____kasan_slab_free+0x11a/0x1b0 Feb 21 09:02:00 c-237-177-40-045 kernel: __kmem_cache_free+0x1de/0x400 Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5_packet_reformat_dealloc+0xad/0x100 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5e_tc_encap_flows_del+0x3c0/0x500 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5e_rep_update_flows+0x40c/0xa80 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: mlx5e_rep_neigh_update+0x473/0x7a0 [mlx5_core] Feb 21 09:02:00 c-237-177-40-045 kernel: process_one_work+0x7c2/0x1310 Feb 21 09:02:00 c-237-177-40-045 kernel: worker_thread+0x59d/0xec0 Feb 21 09:02:00 c-237-177-40-045 kernel: kthread+0x28f/0x330
--- .../net/ethernet/mellanox/mlx5/core/en/tc/post_act.c | 11 ++--------- .../net/ethernet/mellanox/mlx5/core/en/tc/post_act.h | 2 +- 2 files changed, 3 insertions(+), 10 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/tc/post_act.c +2 -9
@@ -106,22 +106,17 @@ mlx5e_tc_post_act_offload(struct mlx5e_post_act *post_act, }struct mlx5e_post_act_handle * -mlx5e_tc_post_act_add (struct mlx5e_post_act *post_act,struct mlx5_flow_attr *attr) +mlx5e_tc_post_act_add (struct mlx5e_post_act *post_act,struct mlx5_flow_attr *post_attr) { -u32 attr_sz =ns_to_attr_sz (post_act->ns_type);struct mlx5e_post_act_handle *handle; -struct mlx5_flow_attr *post_attr;int err; handle =kzalloc (sizeof (*handle),GFP_KERNEL ); - post_attr =mlx5_alloc_flow_attr (post_act->ns_type); -if (!handle || !post_attr) { -kfree (post_attr); +if (!handle) {kfree (handle);return ERR_PTR (-ENOMEM ); } -memcpy (post_attr, attr, attr_sz); post_attr->chain =0 ; post_attr->prio =0 ; post_attr->ft = post_act->ft;
@@ -145,7 +140,6 @@ mlx5e_tc_post_act_add(struct mlx5e_post_act *post_act, struct mlx5_flow_attr *atreturn handle; err_xarray: -kfree (post_attr);kfree (handle);return ERR_PTR (err); }
@@ -164,7 +158,6 @@ mlx5e_tc_post_act_del(struct mlx5e_post_act *post_act, struct mlx5e_post_act_hanif (!IS_ERR_OR_NULL (handle->rule))mlx5e_tc_post_act_unoffload (post_act, handle);xa_erase (&post_act->ids, handle->id); -kfree (handle->attr);kfree (handle); }
drivers/net/ethernet/mellanox/mlx5/core/en/tc/post_act.h +1 -1
@@ -19,7 +19,7 @@ void mlx5e_tc_post_act_destroy(struct mlx5e_post_act *post_act);struct mlx5e_post_act_handle * -mlx5e_tc_post_act_add (struct mlx5e_post_act *post_act,struct mlx5_flow_attr *attr); +mlx5e_tc_post_act_add (struct mlx5e_post_act *post_act,struct mlx5_flow_attr *post_attr);void mlx5e_tc_post_act_del (struct mlx5e_post_act *post_act,struct mlx5e_post_act_handle *handle);
Signature · 1 line
2.39.2
PATCH v2 2/10 net/mlx5e: Release the label when replacing existing ct entry
In reply to Saeed MahameedFrom: Vlad Buslov <[email protected]>
Cited commit doesn't release the label mapping when replacing existing ct entry which leads to following memleak report:
unreferenced object 0xffff8881854cf280 (size 96):
comm "kworker/u48:74", pid 23093, jiffies 4296664564 (age 175.944s) hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
backtrace:
[<000000002722d368>] __kmalloc+0x4b/0x1c0
[<00000000cc44e18f>] mapping_add+0x6e8/0xc90 [mlx5_core]
[<000000003ad942a7>] mlx5_get_label_mapping+0x66/0xe0 [mlx5_core]
[<00000000266308ac>] mlx5_tc_ct_entry_create_mod_hdr+0x1c4/0xf50 [mlx5_core]
[<000000009a768b4f>] mlx5_tc_ct_entry_add_rule+0x16f/0xaf0 [mlx5_core]
[<00000000a178f3e5>] mlx5_tc_ct_block_flow_offload_add+0x10cb/0x1f90 [mlx5_core]
[<000000007b46c496>] mlx5_tc_ct_block_flow_offload+0x14a/0x630 [mlx5_core]
[<00000000a9a18ac5>] nf_flow_offload_tuple+0x1a3/0x390 [nf_flow_table]
[<00000000d0881951>] flow_offload_work_handler+0x257/0xd30 [nf_flow_table]
[<000000009e4935a4>] process_one_work+0x7c2/0x13e0
[<00000000f5cd36a7>] worker_thread+0x59d/0xec0
[<00000000baed1daf>] kthread+0x28f/0x330
[<0000000063d282a4>] ret_from_fork+0x1f/0x30Fix the issue by correctly releasing the label mapping.
--- drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c | 1 + 1 file changed, 1 insertion(+)
drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c +1 -0
@@ -920,6 +920,7 @@ mlx5_tc_ct_entry_replace_rule(struct mlx5_tc_ct_priv *ct_priv, zone_rule->rule = rule;mlx5_tc_ct_entry_destroy_mod_hdr (ct_priv, old_attr, zone_rule->mh); zone_rule->mh = mh; +mlx5_put_label_mapping (ct_priv, old_attr->ct_attr.ct_labels_id);kfree (old_attr);kvfree (spec);
Signature · 1 line
2.39.2
PATCH v2 3/10 net/mlx5: E-switch, Create per vport table based on devlink encap mode
In reply to Saeed MahameedFrom: Chris Mi <[email protected]>
Currently when creating per vport table, create flags are hardcoded. Devlink encap mode is set based on user input and HW capability. Create per vport table based on devlink encap mode.
--- .../net/ethernet/mellanox/mlx5/core/en/tc/sample.c | 4 ++-- .../net/ethernet/mellanox/mlx5/core/esw/vporttbl.c | 12 +++++++++++- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 2 +- .../ethernet/mellanox/mlx5/core/eswitch_offloads.c | 2 +- 4 files changed, 15 insertions(+), 5 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/tc/sample.c +2 -2
@@ -14,10 +14,10 @@#define MLX5_ESW_VPORT_TBL_SIZE_SAMPLE (64 * 1024) -static const struct esw_vport_tbl_namespace mlx5_esw_vport_tbl_sample_ns = { +static struct esw_vport_tbl_namespace mlx5_esw_vport_tbl_sample_ns = { .max_fte =MLX5_ESW_VPORT_TBL_SIZE_SAMPLE , .max_num_groups =0 ,/* default num of groups */ - .flags =MLX5_FLOW_TABLE_TUNNEL_EN_REFORMAT |MLX5_FLOW_TABLE_TUNNEL_EN_DECAP , + .flags =0 , };struct mlx5e_tc_psample {
drivers/net/ethernet/mellanox/mlx5/core/esw/vporttbl.c +11 -1
@@ -11,7 +11,7 @@ struct mlx5_vport_key {u16 prio;u16 vport;u16 vhca_id; -const struct esw_vport_tbl_namespace *vport_ns; +struct esw_vport_tbl_namespace *vport_ns; } __packed;struct mlx5_vport_table {
@@ -21,6 +21,14 @@ struct mlx5_vport_table {struct mlx5_vport_key key; }; +static void +esw_vport_tbl_init (struct mlx5_eswitch *esw,struct esw_vport_tbl_namespace *ns) +{ +if (esw->offloads.encap !=DEVLINK_ESWITCH_ENCAP_MODE_NONE ) + ns->flags |= (MLX5_FLOW_TABLE_TUNNEL_EN_REFORMAT | +MLX5_FLOW_TABLE_TUNNEL_EN_DECAP ); +} +static struct mlx5_flow_table *esw_vport_tbl_create (struct mlx5_eswitch *esw,struct mlx5_flow_namespace *ns,const struct esw_vport_tbl_namespace *vport_ns)
@@ -80,6 +88,7 @@ mlx5_esw_vporttbl_get(struct mlx5_eswitch *esw, struct mlx5_vport_tbl_attr *attru32 hkey;mutex_lock (&esw->fdb_table.offloads.vports.lock); +esw_vport_tbl_init (esw, attr->vport_ns); hkey =flow_attr_to_vport_key (esw, attr, &skey); e =esw_vport_tbl_lookup (esw, &skey, hkey);if (e) {
@@ -127,6 +136,7 @@ mlx5_esw_vporttbl_put(struct mlx5_eswitch *esw, struct mlx5_vport_tbl_attr *attru32 hkey;mutex_lock (&esw->fdb_table.offloads.vports.lock); +esw_vport_tbl_init (esw, attr->vport_ns); hkey =flow_attr_to_vport_key (esw, attr, &key); e =esw_vport_tbl_lookup (esw, &key, hkey);if (!e || --e->num_rules)
drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +1 -1
@@ -674,7 +674,7 @@ struct mlx5_vport_tbl_attr {u32 chain;u16 prio;u16 vport; -const struct esw_vport_tbl_namespace *vport_ns; +struct esw_vport_tbl_namespace *vport_ns; }; struct mlx5_flow_table *
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +1 -1
@@ -73,7 +73,7 @@#define MLX5_ESW_FT_OFFLOADS_DROP_RULE (1) -static const struct esw_vport_tbl_namespace mlx5_esw_vport_tbl_mirror_ns = { +static struct esw_vport_tbl_namespace mlx5_esw_vport_tbl_mirror_ns = { .max_fte =MLX5_ESW_VPORT_TBL_SIZE , .max_num_groups =MLX5_ESW_VPORT_TBL_NUM_GROUPS , .flags =0 ,
Signature · 1 line
2.39.2
PATCH v2 4/10 net/mlx5: E-switch, Don't destroy indirect table in split rule
In reply to Saeed MahameedFrom: Chris Mi <[email protected]>
Source port rewrite (forward to ovs internal port or statck device) isn't supported in the rule of split action. So there is no indirect table in split rule. The cited commit destroyes indirect table in split rule. The indirect table for other rules will be destroyed wrongly. It will cause traffic loss.
Fix it by removing the destroy function in split rule. And also remove the destroy function in error flow.
--- drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 2 -- 1 file changed, 2 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +0 -2
@@ -760,7 +760,6 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,kfree (dest);return rule; err_chain_src_rewrite: -esw_put_dest_tables_loop (esw, attr,0 , i);mlx5_esw_vporttbl_put (esw, &fwd_attr); err_get_fwd:mlx5_chains_put_table (chains, attr->chain, attr->prio,0 );
@@ -803,7 +802,6 @@ __mlx5_eswitch_del_rule(struct mlx5_eswitch *esw,if (fwd_rule) {mlx5_esw_vporttbl_put (esw, &fwd_attr);mlx5_chains_put_table (chains, attr->chain, attr->prio,0 ); -esw_put_dest_tables_loop (esw, attr,0 , esw_attr->split_count); }else {if (split)mlx5_esw_vporttbl_put (esw, &fwd_attr);
Signature · 1 line
2.39.2
PATCH v2 5/10 net/mlx5: Release tunnel device after tc update skb
In reply to Saeed MahameedFrom: Chris Mi <[email protected]>
The cited commit causes a regression. Tunnel device is not released after tc update skb if skb needs to be freed. The following error message will be printed:
unregister_netdevice: waiting for vxlan1 to become free. Usage count = 11
Fix it by releasing tunnel device if skb needs to be freed.
--- drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c | 1 + 1 file changed, 1 insertion(+)
drivers/net/ethernet/mellanox/mlx5/core/en/rep/tc.c +1 -0
@@ -715,5 +715,6 @@ void mlx5e_rep_tc_receive(struct mlx5_cqe64 *cqe, struct mlx5e_rq *rq,return ; free_skb: +dev_put (tc_priv.fwd_dev);dev_kfree_skb_any (skb); }
Signature · 1 line
2.39.2
PATCH v2 6/10 net/mlx5e: Fix error flow in representor failing to add vport rx rule
In reply to Saeed MahameedFrom: Roi Dayan <[email protected]>
On representor init rx error flow the flow steering pointer is being released so mlx5e_attach_netdev() doesn't have a valid fs pointer in its error flow. Make sure the pointer is nullified when released and add a check in mlx5e_fs_cleanup() to verify fs is not null as representor cleanup callback would be called anyway.
--- drivers/net/ethernet/mellanox/mlx5/core/en_fs.c | 2 ++ drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 1 + drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 2 ++ 3 files changed, 5 insertions(+)
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +2 -0
@@ -1490,6 +1490,8 @@ struct mlx5e_flow_steering *mlx5e_fs_init(const struct mlx5e_profile *profile,void mlx5e_fs_cleanup (struct mlx5e_flow_steering *fs) { +if (!fs) +return ;debugfs_remove_recursive (fs->dfs_root);mlx5e_fs_ethtool_free (fs);mlx5e_fs_tc_free (fs);
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +1 -0
@@ -5270,6 +5270,7 @@ static void mlx5e_nic_cleanup(struct mlx5e_priv *priv)mlx5e_health_destroy_reporters (priv);mlx5e_ktls_cleanup (priv);mlx5e_fs_cleanup (priv->fs); + priv->fs =NULL ; }static int mlx5e_init_nic_rx (struct mlx5e_priv *priv)
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +2 -0
@@ -828,6 +828,7 @@ static int mlx5e_init_ul_rep(struct mlx5_core_dev *mdev,static void mlx5e_cleanup_rep (struct mlx5e_priv *priv) {mlx5e_fs_cleanup (priv->fs); + priv->fs =NULL ; }static int mlx5e_create_rep_ttc_table (struct mlx5e_priv *priv)
@@ -994,6 +995,7 @@ static int mlx5e_init_rep_rx(struct mlx5e_priv *priv) priv->rx_res =NULL ; err_free_fs:mlx5e_fs_cleanup (priv->fs); + priv->fs =NULL ;return err; }
Signature · 1 line
2.39.2
PATCH v2 7/10 Revert "net/mlx5: Remove "recovery" arg from mlx5_load_one() function"
In reply to Saeed MahameedFrom: Moshe Shemesh <[email protected]>
This reverts commit 5977ac3910f1cbaf44dca48179118b25c206ac29.
Revert this patch as we need the "recovery" arg back in mlx5_load_one() function. This arg will be used in the next patch for using recovery timeout during sync reset flow.
--- drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/main.c | 9 +++++---- drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h | 2 +- 3 files changed, 7 insertions(+), 6 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c +1 -1
@@ -167,7 +167,7 @@ static void mlx5_fw_reset_complete_reload(struct mlx5_core_dev *dev)if (mlx5_health_wait_pci_up (dev))mlx5_core_err (dev,"reset reload flow aborted, PCI reads still not working\n" );else -mlx5_load_one (dev); +mlx5_load_one (dev, false);devlink_remote_reload_actions_performed (priv_to_devlink (dev),0 ,BIT (DEVLINK_RELOAD_ACTION_DRIVER_REINIT ) |BIT (DEVLINK_RELOAD_ACTION_FW_ACTIVATE ));
drivers/net/ethernet/mellanox/mlx5/core/main.c +5 -4
@@ -1509,13 +1509,13 @@ int mlx5_load_one_devl_locked(struct mlx5_core_dev *dev, bool recovery)return err; } -int mlx5_load_one (struct mlx5_core_dev *dev) +int mlx5_load_one (struct mlx5_core_dev *dev,bool recovery) {struct devlink *devlink =priv_to_devlink (dev);int ret;devl_lock (devlink); - ret =mlx5_load_one_devl_locked (dev, false); + ret =mlx5_load_one_devl_locked (dev, recovery);devl_unlock (devlink);return ret; }
@@ -1912,7 +1912,8 @@ static void mlx5_pci_resume(struct pci_dev *pdev)mlx5_pci_trace (dev,"Enter, loading driver..\n" ); - err =mlx5_load_one (dev); + err =mlx5_load_one (dev, false); +if (!err)devlink_health_reporter_state_update (dev->priv.health.fw_fatal_reporter,DEVLINK_HEALTH_REPORTER_STATE_HEALTHY );
@@ -2003,7 +2004,7 @@ static int mlx5_resume(struct pci_dev *pdev) {struct mlx5_core_dev *dev =pci_get_drvdata (pdev); -return mlx5_load_one (dev); +return mlx5_load_one (dev, false); }static const struct pci_device_id mlx5_core_pci_table[] = {
drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +1 -1
@@ -321,7 +321,7 @@ int mlx5_init_one(struct mlx5_core_dev *dev);void mlx5_uninit_one (struct mlx5_core_dev *dev);void mlx5_unload_one (struct mlx5_core_dev *dev,bool suspend);void mlx5_unload_one_devl_locked (struct mlx5_core_dev *dev,bool suspend); -int mlx5_load_one (struct mlx5_core_dev *dev); +int mlx5_load_one (struct mlx5_core_dev *dev,bool recovery);int mlx5_load_one_devl_locked (struct mlx5_core_dev *dev,bool recovery); int mlx5_vport_set_other_func_cap(struct mlx5_core_dev *dev, const void *hca_cap, u16 function_id,
Signature · 1 line
2.39.2
PATCH v2 8/10 net/mlx5: Use recovery timeout on sync reset flow
In reply to Saeed MahameedFrom: Moshe Shemesh <[email protected]>
Use the same timeout for sync reset flow and health recovery flow, since the former involves driver's recovery from firmware reset, which is similar to health recovery. Otherwise, in some cases, such as a firmware upgrade on the DPU, the firmware pre-init bit may not be ready within current timeout and the driver will abort loading back after reset.
--- drivers/net/ethernet/mellanox/mlx5/core/devlink.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/devlink.c +1 -1
@@ -202,7 +202,7 @@ static int mlx5_devlink_reload_up(struct devlink *devlink, enum devlink_reload_abreak ;/* On fw_activate action, also driver is reloaded and reinit performed */ *actions_performed |=BIT (DEVLINK_RELOAD_ACTION_DRIVER_REINIT ); - ret =mlx5_load_one_devl_locked (dev, false); + ret =mlx5_load_one_devl_locked (dev, true);break ;default :/* Unsupported action should not get to this function */
drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c +2 -2
@@ -167,7 +167,7 @@ static void mlx5_fw_reset_complete_reload(struct mlx5_core_dev *dev)if (mlx5_health_wait_pci_up (dev))mlx5_core_err (dev,"reset reload flow aborted, PCI reads still not working\n" );else -mlx5_load_one (dev, false); +mlx5_load_one (dev, true);devlink_remote_reload_actions_performed (priv_to_devlink (dev),0 ,BIT (DEVLINK_RELOAD_ACTION_DRIVER_REINIT ) |BIT (DEVLINK_RELOAD_ACTION_FW_ACTIVATE ));
@@ -499,7 +499,7 @@ int mlx5_fw_reset_wait_reset_done(struct mlx5_core_dev *dev) err = fw_reset->ret;if (test_and_clear_bit (MLX5_FW_RESET_FLAGS_RELOAD_REQUIRED , &fw_reset->reset_flags)) {mlx5_unload_one_devl_locked (dev, false); -mlx5_load_one_devl_locked (dev, false); +mlx5_load_one_devl_locked (dev, true); } out:clear_bit (MLX5_FW_RESET_FLAGS_PENDING_COMP , &fw_reset->reset_flags);
Signature · 1 line
2.39.2
PATCH v2 9/10 net/mlx5e: Nullify table pointer when failing to create
In reply to Saeed MahameedFrom: Aya Levin <[email protected]>
On failing to create promisc flow steering table, the pointer is returned with an error. Nullify it so unloading the driver won't try to destroy a non existing table.
Failing to create promisc table may happen over BF devices when the ARM side is going through a firmware tear down. The host side start a reload flow. While the driver unloads, it tries to remove the promisc table. Remove WARN in this state as it is a valid error flow.
--- drivers/net/ethernet/mellanox/mlx5/core/en_fs.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +2 -1
@@ -783,6 +783,7 @@ static int mlx5e_create_promisc_table(struct mlx5e_flow_steering *fs) ft->t =mlx5_create_auto_grouped_flow_table (fs->ns, &ft_attr);if (IS_ERR (ft->t)) { err =PTR_ERR (ft->t); + ft->t =NULL ;fs_err (fs,"fail to create promisc table err=%d\n" , err);return err; }
@@ -810,7 +811,7 @@ static void mlx5e_del_promisc_rule(struct mlx5e_flow_steering *fs)static void mlx5e_destroy_promisc_table (struct mlx5e_flow_steering *fs) { -if (WARN (!fs->promisc.ft.t,"Trying to remove non-existing promiscuous table" )) +if (!fs->promisc.ft.t)return ;mlx5e_del_promisc_rule (fs);mlx5_destroy_flow_table (fs->promisc.ft.t);
Signature · 1 line
2.39.2
PATCH v2 10/10 Revert "net/mlx5e: Don't use termination table when redundant"
In reply to Saeed MahameedFrom: Vlad Buslov <[email protected]>
This reverts commit 14624d7247fcd0f3114a6f5f17b3c8d1020fbbb7.
The termination table usage is requires for DMFS steering mode as firmware doesn't support mixed table destinations list which causes following syndrome with hairpin rules:
[81922.283225] mlx5_core 0000:08:00.0: mlx5_cmd_out_err:803:(pid 25977): SET_FLOW_TABLE_ENTRY(0x936) op_mod(0x0) failed, status bad parameter(0x3), syndrome (0xaca205), err(-22)
--- .../mlx5/core/eswitch_offloads_termtbl.c | 32 +++---------------- 1 file changed, 4 insertions(+), 28 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_termtbl.c +4 -28
@@ -210,18 +210,6 @@ static bool mlx5_eswitch_offload_is_uplink_port(const struct mlx5_eswitch *esw,return (port_mask & port_value) ==MLX5_VPORT_UPLINK ; } -static bool -mlx5_eswitch_is_push_vlan_no_cap (struct mlx5_eswitch *esw, -struct mlx5_flow_act *flow_act) -{ -if (flow_act->action &MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH && - !(mlx5_fs_get_capabilities (esw->dev,MLX5_FLOW_NAMESPACE_FDB ) & -MLX5_FLOW_STEERING_CAP_VLAN_PUSH_ON_RX )) -return true; - -return false; -} - bool mlx5_eswitch_termtbl_required(struct mlx5_eswitch *esw, struct mlx5_flow_attr *attr,
@@ -237,7 +225,10 @@ mlx5_eswitch_termtbl_required(struct mlx5_eswitch *esw, (!mlx5_eswitch_offload_is_uplink_port(esw, spec) && !esw_attr->int_port))return false; -if (mlx5_eswitch_is_push_vlan_no_cap (esw, flow_act)) +/* push vlan on RX */ +if (flow_act->action &MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH && + !(mlx5_fs_get_capabilities (esw->dev,MLX5_FLOW_NAMESPACE_FDB ) & +MLX5_FLOW_STEERING_CAP_VLAN_PUSH_ON_RX ))return true;/* hairpin */
@@ -261,31 +252,19 @@ mlx5_eswitch_add_termtbl_rule(struct mlx5_eswitch *esw,struct mlx5_flow_act term_tbl_act = {};struct mlx5_flow_handle *rule =NULL ;bool term_table_created = false; -bool is_push_vlan_on_rx;int num_vport_dests =0 ;int i, curr_dest; - is_push_vlan_on_rx =mlx5_eswitch_is_push_vlan_no_cap (esw, flow_act);mlx5_eswitch_termtbl_actions_move (flow_act, &term_tbl_act); term_tbl_act.action |=MLX5_FLOW_CONTEXT_ACTION_FWD_DEST ;for (i =0 ; i < num_dest; i++) {struct mlx5_termtbl_handle *tt; -bool hairpin = false;/* only vport destinations can be terminated */ if (dest[i].type !=MLX5_FLOW_DESTINATION_TYPE_VPORT )continue ; -if (attr->dests[num_vport_dests].rep && - attr->dests[num_vport_dests].rep->vport ==MLX5_VPORT_UPLINK ) - hairpin = true; - -if (!is_push_vlan_on_rx && !hairpin) { - num_vport_dests++; -continue ; - } -if (attr->dests[num_vport_dests].flags &MLX5_ESW_DEST_ENCAP ) { term_tbl_act.action |=MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT ; term_tbl_act.pkt_reformat = attr->dests[num_vport_dests].pkt_reformat;
@@ -333,9 +312,6 @@ mlx5_eswitch_add_termtbl_rule(struct mlx5_eswitch *esw,for (curr_dest =0 ; curr_dest < num_vport_dests; curr_dest++) {struct mlx5_termtbl_handle *tt = attr->dests[curr_dest].termtbl; -if (!tt) -continue ; - attr->dests[curr_dest].termtbl =NULL ;/* search for the destination associated with the
Signature · 1 line
2.39.2
Re: [net V2 01/10] net/mlx5e: Don't clone flow post action attributes second time
In reply to Saeed MahameedThis series was applied to netdev/net.git (main) by Saeed Mahameed <[email protected]>:
Hello:
This series was applied to netdev/net.git (main)
by Saeed Mahameed <[email protected]>:
Quoted 11 lines · On Thu, 20 Apr 2023 18:50:48 -0700 you wrote:
From: Vlad Buslov <[email protected]>
The code already clones post action attributes in
mlx5e_clone_flow_attr_for_post_act(). Creating another copy inmlx5e_tc_post_act_add()is a erroneous leftover from original implementation. Instead, assignhandle->attributetopost_attrprovided by the caller. Note that cloning the attribute second time is not just wasteful but also causes issues like second copy not being properly updated in neigh update code which leads to following use-after-free:[...]
Here is the summary with links:
- [net,V2,01/10] net/mlx5e: Don't clone flow post action attributes second time
https://git.kernel.org/netdev/net/c/e9fce818fe00 - [net,V2,02/10] net/mlx5e: Release the label when replacing existing ct entry https://git.kernel.org/netdev/net/c/8ac04a28144c - [net,V2,03/10] net/mlx5: E-switch, Create per vport table based on devlink encap mode https://git.kernel.org/netdev/net/c/fd745f4c0abe - [net,V2,04/10] net/mlx5: E-switch, Don't destroy indirect table in split rule https://git.kernel.org/netdev/net/c/4c8189302567 - [net,V2,05/10] net/mlx5: Release tunnel device after tc update skb https://git.kernel.org/netdev/net/c/4fbef0f8ea63 - [net,V2,06/10] net/mlx5e: Fix error flow in representor failing to add vport rx rule https://git.kernel.org/netdev/net/c/0a6b069cc60d - [net,V2,07/10] Revert "net/mlx5: Remove "recovery" arg from mlx5_load_one() function" https://git.kernel.org/netdev/net/c/21608a2cf38e - [net,V2,08/10] net/mlx5: Use recovery timeout on sync reset flow https://git.kernel.org/netdev/net/c/dfad99750c0f - [net,V2,09/10] net/mlx5e: Nullify table pointer when failing to create https://git.kernel.org/netdev/net/c/1b540decd03a - [net,V2,10/10] Revert "net/mlx5e: Don't use termination table when redundant" https://git.kernel.org/netdev/net/c/081abcacaf0a
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html