mlx5 fixes 2022-05-17
From: Saeed Mahameed <[email protected]>
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 23dd4581350d4ffa23d58976ec46408f8f4c1e16:
NFC: nci: fix sleep in atomic context bugs caused by nci_skb_alloc (2022-05-17 17:55:53 -0700)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-fixes-2022-05-17
for you to fetch changes up to 16d42d313350946f4b9a8b74a13c99f0461a6572:
net/mlx5: Drain fw_reset when removing device (2022-05-17 23:03:57 -0700)
----------------------------------------------------------------
mlx5-fixes-2022-05-17
----------------------------------------------------------------
Aya Levin (1):
net/mlx5e: Block rx-gro-hw feature in switchdev mode
Gal Pressman (1):
net/mlx5e: Remove HW-GRO from reported features
Maor Dickman (1):
net/mlx5: DR, Fix missing flow_source when creating multi-destination FW table
Maxim Mikityanskiy (3):
net/mlx5e: Wrap mlx5e_trap_napi_poll into rcu_read_lock net/mlx5e: Properly block LRO when XDP is enabled net/mlx5e: Properly block HW GRO when XDP is enabled
Paul Blakey (2):
net/mlx5e: CT: Fix support for GRE tuples net/mlx5e: CT: Fix setting flow_source for smfs ct tuples
Shay Drory (2):
net/mlx5: Initialize flow steering during driver probe net/mlx5: Drain fw_reset when removing device
Yevgeny Kliteynik (1):
net/mlx5: DR, Ignore modify TTL on RX if device doesn't support it
.../ethernet/mellanox/mlx5/core/en/tc/ct_fs_smfs.c | 58 +++++---- drivers/net/ethernet/mellanox/mlx5/core/en/trap.c | 13 +- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 27 ++++- drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 131 ++++++++++++--------- drivers/net/ethernet/mellanox/mlx5/core/fs_core.h | 6 +- drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c | 25 +++- drivers/net/ethernet/mellanox/mlx5/core/fw_reset.h | 1 + drivers/net/ethernet/mellanox/mlx5/core/main.c | 19 ++- .../mellanox/mlx5/core/steering/dr_action.c | 71 +++++++---- .../ethernet/mellanox/mlx5/core/steering/dr_fw.c | 4 +- .../mellanox/mlx5/core/steering/dr_ste_v0.c | 4 +- .../mellanox/mlx5/core/steering/dr_types.h | 3 +- .../ethernet/mellanox/mlx5/core/steering/fs_dr.c | 4 +- .../ethernet/mellanox/mlx5/core/steering/mlx5dr.h | 3 +- 14 files changed, 246 insertions(+), 123 deletions(-)
PATCH 1/11 net/mlx5: DR, Fix missing flow_source when creating multi-destination FW table
From: Maor Dickman <[email protected]>
In order to support multiple destination FTEs with SW steering FW table is created with single FTE with multiple actions and SW steering rule forward to it. When creating this table, flow source isn't set according to the original FTE.
Fix this by passing the original FTE flow source to the created FW table.
--- .../net/ethernet/mellanox/mlx5/core/steering/dr_action.c | 6 ++++-- drivers/net/ethernet/mellanox/mlx5/core/steering/dr_fw.c | 4 +++- drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h | 3 ++- drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c | 4 +++- drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h | 3 ++- 5 files changed, 14 insertions(+), 6 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c +4 -2
@@ -842,7 +842,8 @@ struct mlx5dr_action * mlx5dr_action_create_mult_dest_tbl(struct mlx5dr_domain *dmn, struct mlx5dr_action_dest *dests, u32 num_of_dests, - bool ignore_flow_level) + bool ignore_flow_level, + u32 flow_source) { struct mlx5dr_cmd_flow_destination_hw_info *hw_dests;struct mlx5dr_action **ref_actions;
@@ -914,7 +915,8 @@ mlx5dr_action_create_mult_dest_tbl(struct mlx5dr_domain *dmn, reformat_req, &action->dest_tbl->fw_tbl.id, &action->dest_tbl->fw_tbl.group_id, - ignore_flow_level); + ignore_flow_level, + flow_source);if (ret) goto free_action;
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_fw.c +3 -1
@@ -104,7 +104,8 @@ int mlx5dr_fw_create_md_tbl(struct mlx5dr_domain *dmn,bool reformat_req, u32 *tbl_id, u32 *group_id, - bool ignore_flow_level) + bool ignore_flow_level, + u32 flow_source) { struct mlx5dr_cmd_create_flow_table_attr ft_attr = {};struct mlx5dr_cmd_fte_info fte_info = {};
@@ -139,6 +140,7 @@ int mlx5dr_fw_create_md_tbl(struct mlx5dr_domain *dmn, fte_info.val = val; fte_info.dest_arr = dest; fte_info.ignore_flow_level = ignore_flow_level; + fte_info.flow_context.flow_source = flow_source; ret =mlx5dr_cmd_set_fte (dmn->mdev,0 ,0 , &ft_info, *group_id, &fte_info);if (ret) {
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h +2 -1
@@ -1461,7 +1461,8 @@ int mlx5dr_fw_create_md_tbl(struct mlx5dr_domain *dmn,bool reformat_req, u32 *tbl_id, u32 *group_id, - bool ignore_flow_level); + bool ignore_flow_level, + u32 flow_source);void mlx5dr_fw_destroy_md_tbl (struct mlx5dr_domain *dmn,u32 tbl_id,u32 group_id);#endif /* _DR_TYPES_H_ */
drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c +3 -1
@@ -520,6 +520,7 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns, }else if (num_term_actions >1 ) {bool ignore_flow_level = !!(fte->action.flags &FLOW_ACT_IGNORE_FLOW_LEVEL ); +u32 flow_source = fte->flow_context.flow_source;if (num_actions ==MLX5_FLOW_CONTEXT_ACTION_MAX || fs_dr_num_actions ==MLX5_FLOW_CONTEXT_ACTION_MAX ) {
@@ -529,7 +530,8 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns, tmp_action =mlx5dr_action_create_mult_dest_tbl (domain, term_actions, num_term_actions, - ignore_flow_level); + ignore_flow_level, + flow_source);if (!tmp_action) { err = -EOPNOTSUPP ; goto free_actions;
drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h +2 -1
@@ -99,7 +99,8 @@ struct mlx5dr_action * mlx5dr_action_create_mult_dest_tbl(struct mlx5dr_domain *dmn, struct mlx5dr_action_dest *dests, u32 num_of_dests, - bool ignore_flow_level); + bool ignore_flow_level, + u32 flow_source);struct mlx5dr_action *mlx5dr_action_create_drop (void );
Signature · 1 line
2.36.1
Re: [net 01/11] net/mlx5: DR, Fix missing flow_source when creating multi-destination FW table
This series was applied to netdev/net.git (master) by Saeed Mahameed <[email protected]>:
Hello:
This series was applied to netdev/net.git (master) by Saeed Mahameed <[email protected]>:
Quoted 8 lines · On Tue, 17 May 2022 23:34:17 -0700 you wrote:
From: Maor Dickman <[email protected]>
In order to support multiple destination FTEs with SW steering FW table is created with single FTE with multiple actions and SW steering rule forward to it. When creating this table, flow source isn't set according to the original FTE.
[...]
Here is the summary with links:
- [net,01/11] net/mlx5: DR, Fix missing flow_source when creating multi-destination FW table
https://git.kernel.org/netdev/net/c/2c5fc6cd269a - [net,02/11] net/mlx5: Initialize flow steering during driver probe https://git.kernel.org/netdev/net/c/b33886971dbc - [net,03/11] net/mlx5: DR, Ignore modify TTL on RX if device doesn't support it https://git.kernel.org/netdev/net/c/785d7ed29551 - [net,04/11] net/mlx5e: Wrap mlx5e_trap_napi_poll into rcu_read_lock https://git.kernel.org/netdev/net/c/379169740b0a - [net,05/11] net/mlx5e: Block rx-gro-hw feature in switchdev mode https://git.kernel.org/netdev/net/c/15a5078cab30 - [net,06/11] net/mlx5e: Properly block LRO when XDP is enabled https://git.kernel.org/netdev/net/c/cf6e34c8c22f - [net,07/11] net/mlx5e: Properly block HW GRO when XDP is enabled https://git.kernel.org/netdev/net/c/b0617e7b3500 - [net,08/11] net/mlx5e: Remove HW-GRO from reported features https://git.kernel.org/netdev/net/c/6bbd723035ba - [net,09/11] net/mlx5e: CT: Fix support for GRE tuples https://git.kernel.org/netdev/net/c/8e1dcf499a67 - [net,10/11] net/mlx5e: CT: Fix setting flow_source for smfs ct tuples https://git.kernel.org/netdev/net/c/04c551bad371 - [net,11/11] net/mlx5: Drain fw_reset when removing device https://git.kernel.org/netdev/net/c/16d42d313350
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html
PATCH 2/11 net/mlx5: Initialize flow steering during driver probe
In reply to Saeed MahameedFrom: Shay Drory <[email protected]>
Currently, software objects of flow steering are created and destroyed during reload flow. In case a device is unloaded, the following error is printed during grace period:
mlx5_core 0000:00:0b.0: mlx5_fw_fatal_reporter_err_work:690:(pid 95):
Driver is in error state. Unloading
As a solution to fix use-after-free bugs, where we try to access these objects, when reading the value of flow_steering_mode devlink param[1], let's split flow steering creation and destruction into two routines:
* init and cleanup: memory, cache, and pools allocation/free.
* create and destroy: namespaces initialization and cleanup.
While at it, re-order the cleanup function to mirror the init function.
[1]
Kasan trace:
[ 385.119849 ] BUG: KASAN: use-after-free in mlx5_devlink_fs_mode_get+0x3b/0xa0 [ 385.119849 ] Read of size 4 at addr ffff888104b79308 by task bash/291 [ 385.119849 ] [ 385.119849 ] CPU: 1 PID: 291 Comm: bash Not tainted 5.17.0-rc1+ #2 [ 385.119849 ] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-2.fc32 04/01/2014 [ 385.119849 ] Call Trace: [ 385.119849 ] <TASK> [ 385.119849 ] dump_stack_lvl+0x6e/0x91 [ 385.119849 ] print_address_description.constprop.0+0x1f/0x160 [ 385.119849 ] ? mlx5_devlink_fs_mode_get+0x3b/0xa0 [ 385.119849 ] ? mlx5_devlink_fs_mode_get+0x3b/0xa0 [ 385.119849 ] kasan_report.cold+0x83/0xdf [ 385.119849 ] ? devlink_param_notify+0x20/0x190 [ 385.119849 ] ? mlx5_devlink_fs_mode_get+0x3b/0xa0 [ 385.119849 ] mlx5_devlink_fs_mode_get+0x3b/0xa0 [ 385.119849 ] devlink_nl_param_fill+0x18a/0xa50 [ 385.119849 ] ? _raw_spin_lock_irqsave+0x8d/0xe0 [ 385.119849 ] ? devlink_flash_update_timeout_notify+0xf0/0xf0 [ 385.119849 ] ? __wake_up_common+0x4b/0x1e0 [ 385.119849 ] ? preempt_count_sub+0x14/0xc0 [ 385.119849 ] ? _raw_spin_unlock_irqrestore+0x28/0x40 [ 385.119849 ] ? __wake_up_common_lock+0xe3/0x140 [ 385.119849 ] ? __wake_up_common+0x1e0/0x1e0 [ 385.119849 ] ? __sanitizer_cov_trace_const_cmp8+0x27/0x80 [ 385.119849 ] ? __rcu_read_unlock+0x48/0x70 [ 385.119849 ] ? kasan_unpoison+0x23/0x50 [ 385.119849 ] ? __kasan_slab_alloc+0x2c/0x80 [ 385.119849 ] ? memset+0x20/0x40 [ 385.119849 ] ? __sanitizer_cov_trace_const_cmp4+0x25/0x80 [ 385.119849 ] devlink_param_notify+0xce/0x190 [ 385.119849 ] devlink_unregister+0x92/0x2b0 [ 385.119849 ] remove_one+0x41/0x140 [ 385.119849 ] pci_device_remove+0x68/0x140 [ 385.119849 ] ? pcibios_free_irq+0x10/0x10 [ 385.119849 ] __device_release_driver+0x294/0x3f0 [ 385.119849 ] device_driver_detach+0x82/0x130 [ 385.119849 ] unbind_store+0x193/0x1b0 [ 385.119849 ] ? subsys_interface_unregister+0x270/0x270 [ 385.119849 ] drv_attr_store+0x4e/0x70 [ 385.119849 ] ? drv_attr_show+0x60/0x60 [ 385.119849 ] sysfs_kf_write+0xa7/0xc0 [ 385.119849 ] kernfs_fop_write_iter+0x23a/0x2f0 [ 385.119849 ] ? sysfs_kf_bin_read+0x160/0x160 [ 385.119849 ] new_sync_write+0x311/0x430 [ 385.119849 ] ? new_sync_read+0x480/0x480 [ 385.119849 ] ? _raw_spin_lock+0x87/0xe0 [ 385.119849 ] ? __sanitizer_cov_trace_cmp4+0x25/0x80 [ 385.119849 ] ? security_file_permission+0x94/0xa0 [ 385.119849 ] vfs_write+0x4c7/0x590 [ 385.119849 ] ksys_write+0xf6/0x1e0 [ 385.119849 ] ? __x64_sys_read+0x50/0x50 [ 385.119849 ] ? fpregs_assert_state_consistent+0x99/0xa0 [ 385.119849 ] do_syscall_64+0x3d/0x90 [ 385.119849 ] entry_SYSCALL_64_after_hwframe+0x44/0xae [ 385.119849 ] RIP: 0033:0x7fc36ef38504 [ 385.119849 ] Code: 00 f7 d8 64 89 02 48 c7 c0 ff ff ff ff eb b3 0f 1f 80 00 00 00 00 48 8d 05 f9 61 0d 00 8b 00 85 c0 75 13 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 54 c3 0f 1f 00 41 54 49 89 d4 55 48 89 f5 53 [ 385.119849 ] RSP: 002b:00007ffde0ff3d08 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 [ 385.119849 ] RAX: ffffffffffffffda RBX: 000000000000000c RCX: 00007fc36ef38504 [ 385.119849 ] RDX: 000000000000000c RSI: 00007fc370521040 RDI: 0000000000000001 [ 385.119849 ] RBP: 00007fc370521040 R08: 00007fc36f00b8c0 R09: 00007fc36ee4b740 [ 385.119849 ] R10: 0000000000000000 R11: 0000000000000246 R12: 00007fc36f00a760 [ 385.119849 ] R13: 000000000000000c R14: 00007fc36f005760 R15: 000000000000000c [ 385.119849 ] </TASK> [ 385.119849 ] [ 385.119849 ] Allocated by task 65: [ 385.119849 ] kasan_save_stack+0x1e/0x40 [ 385.119849 ] __kasan_kmalloc+0x81/0xa0 [ 385.119849 ] mlx5_init_fs+0x11b/0x1160 [ 385.119849 ] mlx5_load+0x13c/0x220 [ 385.119849 ] mlx5_load_one+0xda/0x160 [ 385.119849 ] mlx5_recover_device+0xb8/0x100 [ 385.119849 ] mlx5_health_try_recover+0x2f9/0x3a1 [ 385.119849 ] devlink_health_reporter_recover+0x75/0x100 [ 385.119849 ] devlink_health_report+0x26c/0x4b0 [ 385.275909 ] mlx5_fw_fatal_reporter_err_work+0x11e/0x1b0 [ 385.275909 ] process_one_work+0x520/0x970 [ 385.275909 ] worker_thread+0x378/0x950 [ 385.275909 ] kthread+0x1bb/0x200 [ 385.275909 ] ret_from_fork+0x1f/0x30 [ 385.275909 ] [ 385.275909 ] Freed by task 65: [ 385.275909 ] kasan_save_stack+0x1e/0x40 [ 385.275909 ] kasan_set_track+0x21/0x30 [ 385.275909 ] kasan_set_free_info+0x20/0x30 [ 385.275909 ] __kasan_slab_free+0xfc/0x140 [ 385.275909 ] kfree+0xa5/0x3b0 [ 385.275909 ] mlx5_unload+0x2e/0xb0 [ 385.275909 ] mlx5_unload_one+0x86/0xb0 [ 385.275909 ] mlx5_fw_fatal_reporter_err_work.cold+0xca/0xcf [ 385.275909 ] process_one_work+0x520/0x970 [ 385.275909 ] worker_thread+0x378/0x950 [ 385.275909 ] kthread+0x1bb/0x200 [ 385.275909 ] ret_from_fork+0x1f/0x30 [ 385.275909 ] [ 385.275909 ] The buggy address belongs to the object at ffff888104b79300 [ 385.275909 ] which belongs to the cache kmalloc-128 of size 128 [ 385.275909 ] The buggy address is located 8 bytes inside of [ 385.275909 ] 128-byte region [ffff888104b79300, ffff888104b79380) [ 385.275909 ] The buggy address belongs to the page: [ 385.275909 ] page:00000000de44dd39 refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x104b78 [ 385.275909 ] head:00000000de44dd39 order:1 compound_mapcount:0 [ 385.275909 ] flags: 0x8000000000010200(slab|head|zone=2) [ 385.275909 ] raw: 8000000000010200 0000000000000000 dead000000000122 ffff8881000428c0 [ 385.275909 ] raw: 0000000000000000 0000000080200020 00000001ffffffff 0000000000000000 [ 385.275909 ] page dumped because: kasan: bad access detected [ 385.275909 ] [ 385.275909 ] Memory state around the buggy address: [ 385.275909 ] ffff888104b79200: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 fc fc [ 385.275909 ] ffff888104b79280: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc [ 385.275909 ] >ffff888104b79300: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb [ 385.275909 ] ^ [ 385.275909 ] ffff888104b79380: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc [ 385.275909 ] ffff888104b79400: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb [ 385.275909 ]]
--- .../net/ethernet/mellanox/mlx5/core/fs_core.c | 131 ++++++++++-------- .../net/ethernet/mellanox/mlx5/core/fs_core.h | 6 +- .../net/ethernet/mellanox/mlx5/core/main.c | 15 +- 3 files changed, 91 insertions(+), 61 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/fs_core.c +75 -56
@@ -2663,28 +2663,6 @@ static void cleanup_root_ns(struct mlx5_flow_root_namespace *root_ns)clean_tree (&root_ns->ns.node); } -void mlx5_cleanup_fs (struct mlx5_core_dev *dev) -{ -struct mlx5_flow_steering *steering = dev->priv.steering; - -cleanup_root_ns (steering->root_ns); -cleanup_root_ns (steering->fdb_root_ns); - steering->fdb_root_ns =NULL ; -kfree (steering->fdb_sub_ns); - steering->fdb_sub_ns =NULL ; -cleanup_root_ns (steering->port_sel_root_ns); -cleanup_root_ns (steering->sniffer_rx_root_ns); -cleanup_root_ns (steering->sniffer_tx_root_ns); -cleanup_root_ns (steering->rdma_rx_root_ns); -cleanup_root_ns (steering->rdma_tx_root_ns); -cleanup_root_ns (steering->egress_root_ns); -mlx5_cleanup_fc_stats (dev); -kmem_cache_destroy (steering->ftes_cache); -kmem_cache_destroy (steering->fgs_cache); -mlx5_ft_pool_destroy (dev); -kfree (steering); -} -static int init_sniffer_tx_root_ns (struct mlx5_flow_steering *steering) {struct fs_prio *prio;
@@ -3086,42 +3064,27 @@ static int init_egress_root_ns(struct mlx5_flow_steering *steering)return err; } -int mlx5_init_fs (struct mlx5_core_dev *dev) +void mlx5_fs_core_cleanup (struct mlx5_core_dev *dev) { -struct mlx5_flow_steering *steering; -int err =0 ; - - err =mlx5_init_fc_stats (dev); -if (err) -return err; - - err =mlx5_ft_pool_init (dev); -if (err) -return err; - - steering =kzalloc (sizeof (*steering),GFP_KERNEL ); -if (!steering) { - err = -ENOMEM ; - goto err; - } - - steering->dev = dev; - dev->priv.steering = steering; +struct mlx5_flow_steering *steering = dev->priv.steering; -if (mlx5_fs_dr_is_supported (dev)) - steering->mode =MLX5_FLOW_STEERING_MODE_SMFS ; -else - steering->mode =MLX5_FLOW_STEERING_MODE_DMFS ; +cleanup_root_ns (steering->root_ns); +cleanup_root_ns (steering->fdb_root_ns); + steering->fdb_root_ns =NULL ; +kfree (steering->fdb_sub_ns); + steering->fdb_sub_ns =NULL ; +cleanup_root_ns (steering->port_sel_root_ns); +cleanup_root_ns (steering->sniffer_rx_root_ns); +cleanup_root_ns (steering->sniffer_tx_root_ns); +cleanup_root_ns (steering->rdma_rx_root_ns); +cleanup_root_ns (steering->rdma_tx_root_ns); +cleanup_root_ns (steering->egress_root_ns); +} - steering->fgs_cache =kmem_cache_create ("mlx5_fs_fgs" , -sizeof (struct mlx5_flow_group ),0 , -0 ,NULL ); - steering->ftes_cache =kmem_cache_create ("mlx5_fs_ftes" ,sizeof (struct fs_fte ),0 , -0 ,NULL ); -if (!steering->ftes_cache || !steering->fgs_cache) { - err = -ENOMEM ; - goto err; - } +int mlx5_fs_core_init (struct mlx5_core_dev *dev) +{ +struct mlx5_flow_steering *steering = dev->priv.steering; +int err =0 ; if ((((MLX5_CAP_GEN(dev, port_type) == MLX5_CAP_PORT_TYPE_ETH) && (MLX5_CAP_GEN(dev, nic_flow_table))) ||
@@ -3180,8 +3143,64 @@ int mlx5_init_fs(struct mlx5_core_dev *dev) }return 0 ; + +err: +mlx5_fs_core_cleanup (dev); +return err; +} + +void mlx5_fs_core_free (struct mlx5_core_dev *dev) +{ +struct mlx5_flow_steering *steering = dev->priv.steering; + +kmem_cache_destroy (steering->ftes_cache); +kmem_cache_destroy (steering->fgs_cache); +kfree (steering); +mlx5_ft_pool_destroy (dev); +mlx5_cleanup_fc_stats (dev); +} + +int mlx5_fs_core_alloc (struct mlx5_core_dev *dev) +{ +struct mlx5_flow_steering *steering; +int err =0 ; + + err =mlx5_init_fc_stats (dev); +if (err) +return err; + + err =mlx5_ft_pool_init (dev); +if (err) + goto err; + + steering =kzalloc (sizeof (*steering),GFP_KERNEL ); +if (!steering) { + err = -ENOMEM ; + goto err; + } + + steering->dev = dev; + dev->priv.steering = steering; + +if (mlx5_fs_dr_is_supported (dev)) + steering->mode =MLX5_FLOW_STEERING_MODE_SMFS ; +else + steering->mode =MLX5_FLOW_STEERING_MODE_DMFS ; + + steering->fgs_cache =kmem_cache_create ("mlx5_fs_fgs" , +sizeof (struct mlx5_flow_group ),0 , +0 ,NULL ); + steering->ftes_cache =kmem_cache_create ("mlx5_fs_ftes" ,sizeof (struct fs_fte ),0 , +0 ,NULL ); +if (!steering->ftes_cache || !steering->fgs_cache) { + err = -ENOMEM ; + goto err; + } + +return 0 ; + err: -mlx5_cleanup_fs (dev); +mlx5_fs_core_free (dev);return err; }
drivers/net/ethernet/mellanox/mlx5/core/fs_core.h +4 -2
@@ -298,8 +298,10 @@ int mlx5_flow_namespace_set_peer(struct mlx5_flow_root_namespace *ns,int mlx5_flow_namespace_set_mode (struct mlx5_flow_namespace *ns,enum mlx5_flow_steering_mode mode); -int mlx5_init_fs (struct mlx5_core_dev *dev); -void mlx5_cleanup_fs (struct mlx5_core_dev *dev); +int mlx5_fs_core_alloc (struct mlx5_core_dev *dev); +void mlx5_fs_core_free (struct mlx5_core_dev *dev); +int mlx5_fs_core_init (struct mlx5_core_dev *dev); +void mlx5_fs_core_cleanup (struct mlx5_core_dev *dev);int mlx5_fs_egress_acls_init (struct mlx5_core_dev *dev,int total_vports);void mlx5_fs_egress_acls_cleanup (struct mlx5_core_dev *dev);
drivers/net/ethernet/mellanox/mlx5/core/main.c +12 -3
@@ -938,6 +938,12 @@ static int mlx5_init_once(struct mlx5_core_dev *dev) goto err_sf_table_cleanup; } + err =mlx5_fs_core_alloc (dev); +if (err) { +mlx5_core_err (dev,"Failed to alloc flow steering\n" ); + goto err_fs; + } + dev->dm =mlx5_dm_create (dev);if (IS_ERR (dev->dm))mlx5_core_warn (dev,"Failed to init device memory%d\n" , err);
@@ -948,6 +954,8 @@ static int mlx5_init_once(struct mlx5_core_dev *dev)return 0 ; +err_fs: +mlx5_sf_table_cleanup (dev); err_sf_table_cleanup:mlx5_sf_hw_table_cleanup (dev); err_sf_hw_table_cleanup:
@@ -985,6 +993,7 @@ static void mlx5_cleanup_once(struct mlx5_core_dev *dev)mlx5_hv_vhca_destroy (dev->hv_vhca);mlx5_fw_tracer_destroy (dev->tracer);mlx5_dm_cleanup (dev); +mlx5_fs_core_free (dev);mlx5_sf_table_cleanup (dev);mlx5_sf_hw_table_cleanup (dev);mlx5_vhca_event_cleanup (dev);
@@ -1191,7 +1200,7 @@ static int mlx5_load(struct mlx5_core_dev *dev) goto err_tls_start; } - err =mlx5_init_fs (dev); + err =mlx5_fs_core_init (dev);if (err) {mlx5_core_err (dev,"Failed to init flow steering\n" ); goto err_fs;
@@ -1236,7 +1245,7 @@ static int mlx5_load(struct mlx5_core_dev *dev) err_vhca:mlx5_vhca_event_stop (dev); err_set_hca: -mlx5_cleanup_fs (dev); +mlx5_fs_core_cleanup (dev); err_fs:mlx5_accel_tls_cleanup (dev); err_tls_start:
@@ -1265,7 +1274,7 @@ static void mlx5_unload(struct mlx5_core_dev *dev)mlx5_ec_cleanup (dev);mlx5_sf_hw_table_destroy (dev);mlx5_vhca_event_stop (dev); -mlx5_cleanup_fs (dev); +mlx5_fs_core_cleanup (dev);mlx5_accel_ipsec_cleanup (dev);mlx5_accel_tls_cleanup (dev);mlx5_fpga_device_stop (dev);
Signature · 1 line
2.36.1
PATCH 3/11 net/mlx5: DR, Ignore modify TTL on RX if device doesn't support it
In reply to Saeed MahameedFrom: Yevgeny Kliteynik <[email protected]>
When modifying TTL, packet's csum has to be recalculated. Due to HW issue in ConnectX-5, csum recalculation for modify TTL on RX is supported through a work-around that is specifically enabled by configuration.
If the work-around isn't enabled, rather than adding an unsupported action the modify TTL action on RX should be ignored. Ignoring modify TTL action might result in zero actions, so in such cases we will not convert the match STE to modify STE, as it is done by FW in DMFS.
This patch fixes an issue where modify TTL action was ignored both on RX and TX instead of only on RX.
--- .../mellanox/mlx5/core/steering/dr_action.c | 65 +++++++++++++------ .../mellanox/mlx5/core/steering/dr_ste_v0.c | 4 +- 2 files changed, 48 insertions(+), 21 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c +46 -19
@@ -530,6 +530,37 @@ static int dr_action_handle_cs_recalc(struct mlx5dr_domain *dmn,return 0 ; } +static void dr_action_modify_ttl_adjust (struct mlx5dr_domain *dmn, +struct mlx5dr_ste_actions_attr *attr, +bool rx_rule, +bool *recalc_cs_required) +{ + *recalc_cs_required = false; + +/* if device supports csum recalculation - no adjustment needed */ +if (mlx5dr_ste_supp_ttl_cs_recalc (&dmn->info.caps)) +return ; + +/* no adjustment needed on TX rules */ +if (!rx_rule) +return ; + +if (!MLX5_CAP_ESW_FLOWTABLE (dmn->mdev, fdb_ipv4_ttl_modify)) { +/* Ignore the modify TTL action. +* It is always kept as last HW action. +*/ + attr->modify_actions--; +return ; + } + +if (dmn->type ==MLX5DR_DOMAIN_TYPE_FDB ) +/* Due to a HW bug on some devices, modifying TTL on RX flows +* will cause an incorrect checksum calculation. In such cases +* we will use a FW table to recalculate the checksum. +*/ + *recalc_cs_required = true; +} +static void dr_action_print_sequence (struct mlx5dr_domain *dmn,struct mlx5dr_action *actions[],int last_idx)
@@ -650,8 +681,9 @@ int mlx5dr_actions_build_ste_arr(struct mlx5dr_matcher *matcher,case DR_ACTION_TYP_MODIFY_HDR : attr.modify_index = action->rewrite->index; attr.modify_actions = action->rewrite->num_of_actions; - recalc_cs_required = action->rewrite->modify_ttl && - !mlx5dr_ste_supp_ttl_cs_recalc (&dmn->info.caps); +if (action->rewrite->modify_ttl) +dr_action_modify_ttl_adjust (dmn, &attr, rx_rule, + &recalc_cs_required);break ;case DR_ACTION_TYP_L2_TO_TNL_L2 :case DR_ACTION_TYP_L2_TO_TNL_L3 :
@@ -732,12 +764,7 @@ int mlx5dr_actions_build_ste_arr(struct mlx5dr_matcher *matcher, *new_hw_ste_arr_sz = nic_matcher->num_of_builders; last_ste = ste_arr +DR_STE_SIZE * (nic_matcher->num_of_builders -1 ); -/* Due to a HW bug in some devices, modifying TTL on RX flows will -* cause an incorrect checksum calculation. In this case we will -* use a FW table to recalculate. -*/ -if (dmn->type ==MLX5DR_DOMAIN_TYPE_FDB && - rx_rule && recalc_cs_required && dest_action) { +if (recalc_cs_required && dest_action) { ret =dr_action_handle_cs_recalc (dmn, dest_action, &attr.final_icm_addr);if (ret) { mlx5dr_err(dmn,
@@ -1558,12 +1585,6 @@ dr_action_modify_check_is_ttl_modify(const void *sw_action)return sw_field ==MLX5_ACTION_IN_FIELD_OUT_IP_TTL ; } -static bool dr_action_modify_ttl_ignore (struct mlx5dr_domain *dmn) -{ -return !mlx5dr_ste_supp_ttl_cs_recalc (&dmn->info.caps) && - !MLX5_CAP_ESW_FLOWTABLE (dmn->mdev, fdb_ipv4_ttl_modify); -} - static int dr_actions_convert_modify_header(struct mlx5dr_action *action, u32 max_hw_actions, u32 num_sw_actions,
@@ -1575,6 +1596,7 @@ static int dr_actions_convert_modify_header(struct mlx5dr_action *action,const struct mlx5dr_ste_action_modify_field *hw_dst_action_info;const struct mlx5dr_ste_action_modify_field *hw_src_action_info;struct mlx5dr_domain *dmn = action->rewrite->dmn; +__be64 *modify_ttl_sw_action =NULL ;int ret, i, hw_idx =0 ;__be64 *sw_action;__be64 hw_action;
@@ -1587,8 +1609,14 @@ static int dr_actions_convert_modify_header(struct mlx5dr_action *action, action->rewrite->allow_rx =1 ; action->rewrite->allow_tx =1 ; -for (i =0 ; i < num_sw_actions; i++) { - sw_action = &sw_actions[i]; +for (i =0 ; i < num_sw_actions || modify_ttl_sw_action; i++) { +/* modify TTL is handled separately, as a last action */ +if (i == num_sw_actions) { + sw_action = modify_ttl_sw_action; + modify_ttl_sw_action =NULL ; + }else { + sw_action = &sw_actions[i]; + } ret =dr_action_modify_check_field_limitation (action, sw_action);
@@ -1597,10 +1625,9 @@ static int dr_actions_convert_modify_header(struct mlx5dr_action *action,if (!(*modify_ttl) &&dr_action_modify_check_is_ttl_modify (sw_action)) { -if (dr_action_modify_ttl_ignore (dmn)) -continue ; - + modify_ttl_sw_action = sw_action; *modify_ttl = true; +continue ; }/* Convert SW action to HW action */
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v0.c +2 -2
@@ -420,7 +420,7 @@ dr_ste_v0_set_actions_tx(struct mlx5dr_domain *dmn,* encapsulation. The reason for that is that we support * modify headers for outer headers only */ -if (action_type_set[DR_ACTION_TYP_MODIFY_HDR ]) { +if (action_type_set[DR_ACTION_TYP_MODIFY_HDR ] && attr->modify_actions) {dr_ste_v0_set_entry_type (last_ste,DR_STE_TYPE_MODIFY_PKT ); dr_ste_v0_set_rewrite_actions(last_ste, attr->modify_actions,
@@ -513,7 +513,7 @@ dr_ste_v0_set_actions_rx(struct mlx5dr_domain *dmn, } } -if (action_type_set[DR_ACTION_TYP_MODIFY_HDR ]) { +if (action_type_set[DR_ACTION_TYP_MODIFY_HDR ] && attr->modify_actions) {if (dr_ste_v0_get_entry_type (last_ste) ==DR_STE_TYPE_MODIFY_PKT ) dr_ste_v0_arr_init_next(&last_ste, added_stes,
Signature · 1 line
2.36.1
PATCH 4/11 net/mlx5e: Wrap mlx5e_trap_napi_poll into rcu_read_lock
In reply to Saeed MahameedFrom: Maxim Mikityanskiy <[email protected]>
The body of mlx5e_napi_poll is wrapped into rcu_read_lock to be able to read the XDP program pointer using rcu_dereference. However, the trap RQ NAPI doesn't use rcu_read_lock, because the trap RQ works only in the non-linear mode, and mlx5e_skb_from_cqe_nonlinear, until recently, didn't support XDP and didn't call rcu_dereference.
Starting from the cited commit, mlx5e_skb_from_cqe_nonlinear supports XDP and calls rcu_dereference, but mlx5e_trap_napi_poll doesn't wrap it into rcu_read_lock. It leads to RCU-lockdep warnings like this:
WARNING: suspicious RCU usage
This commit fixes the issue by adding an rcu_read_lock to mlx5e_trap_napi_poll, similarly to mlx5e_napi_poll.
--- drivers/net/ethernet/mellanox/mlx5/core/en/trap.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/trap.c +10 -3
@@ -14,19 +14,26 @@ static int mlx5e_trap_napi_poll(struct napi_struct *napi, int budget)bool busy = false;int work_done =0 ; +rcu_read_lock (); + ch_stats->poll++; work_done =mlx5e_poll_rx_cq (&rq->cq, budget); busy |= work_done == budget; busy |= rq->post_wqes (rq); -if (busy) -return budget; +if (busy) { + work_done = budget; + goto out; + }if (unlikely (!napi_complete_done (napi, work_done))) -return work_done; + goto out;mlx5e_cq_arm (&rq->cq); + +out: +rcu_read_unlock ();return work_done; }
Signature · 1 line
2.36.1
PATCH 5/11 net/mlx5e: Block rx-gro-hw feature in switchdev mode
In reply to Saeed MahameedFrom: Aya Levin <[email protected]>
When the driver is in switchdev mode and rx-gro-hw is set, the RQ needs special CQE handling. Till then, block setting of rx-gro-hw feature in switchdev mode, to avoid failure while setting the feature due to failure while opening the RQ.
--- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 4 ++++ 1 file changed, 4 insertions(+)
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +4 -0
@@ -3864,6 +3864,10 @@ static netdev_features_t mlx5e_fix_uplink_rep_features(struct net_device *netdevif (netdev->features &NETIF_F_NTUPLE )netdev_warn (netdev,"Disabling ntuple, not supported in switchdev mode\n" ); + features &= ~NETIF_F_GRO_HW ; +if (netdev->features &NETIF_F_GRO_HW ) +netdev_warn (netdev,"Disabling HW_GRO, not supported in switchdev mode\n" ); +return features; }
Signature · 1 line
2.36.1
PATCH 6/11 net/mlx5e: Properly block LRO when XDP is enabled
In reply to Saeed MahameedFrom: Maxim Mikityanskiy <[email protected]>
LRO is incompatible and mutually exclusive with XDP. However, the needed checks are only made when enabling XDP. If LRO is enabled when XDP is already active, the command will succeed, and XDP will be skipped in the data path, although still enabled.
This commit fixes the bug by checking the XDP status in mlx5e_fix_features and disabling LRO if XDP is enabled.
--- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 7 +++++++ 1 file changed, 7 insertions(+)
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +7 -0
@@ -3900,6 +3900,13 @@ static netdev_features_t mlx5e_fix_features(struct net_device *netdev, } } +if (params->xdp_prog) { +if (features &NETIF_F_LRO ) { +netdev_warn (netdev,"LRO is incompatible with XDP\n" ); + features &= ~NETIF_F_LRO ; + } + } +if (MLX5E_GET_PFLAG (params,MLX5E_PFLAG_RX_CQE_COMPRESS )) { features &= ~NETIF_F_RXHASH ;if (netdev->features &NETIF_F_RXHASH )
Signature · 1 line
2.36.1
PATCH 7/11 net/mlx5e: Properly block HW GRO when XDP is enabled
In reply to Saeed MahameedFrom: Maxim Mikityanskiy <[email protected]>
HW GRO is incompatible and mutually exclusive with XDP and XSK. However, the needed checks are only made when enabling XDP. If HW GRO is enabled when XDP is already active, the command will succeed, and XDP will be skipped in the data path, although still enabled.
This commit fixes the bug by checking the XDP and XSK status in mlx5e_fix_features and disabling HW GRO if XDP is enabled.
--- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 12 ++++++++++++ 1 file changed, 12 insertions(+)
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +12 -0
@@ -3905,6 +3905,18 @@ static netdev_features_t mlx5e_fix_features(struct net_device *netdev,netdev_warn (netdev,"LRO is incompatible with XDP\n" ); features &= ~NETIF_F_LRO ; } +if (features &NETIF_F_GRO_HW ) { +netdev_warn (netdev,"HW GRO is incompatible with XDP\n" ); + features &= ~NETIF_F_GRO_HW ; + } + } + +if (priv->xsk.refcnt) { +if (features &NETIF_F_GRO_HW ) { +netdev_warn (netdev,"HW GRO is incompatible with AF_XDP (%u XSKs are active)\n" , + priv->xsk.refcnt); + features &= ~NETIF_F_GRO_HW ; + } }if (MLX5E_GET_PFLAG (params,MLX5E_PFLAG_RX_CQE_COMPRESS )) {
Signature · 1 line
2.36.1
PATCH 8/11 net/mlx5e: Remove HW-GRO from reported features
In reply to Saeed MahameedFrom: Gal Pressman <[email protected]>
We got reports of certain HW-GRO flows causing kernel call traces, which might be related to firmware. To be on the safe side, disable the feature for now and re-enable it once a driver/firmware fix is found.
--- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 4 ---- 1 file changed, 4 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +0 -4
@@ -4873,10 +4873,6 @@ static void mlx5e_build_nic_netdev(struct net_device *netdev) netdev->hw_features |=NETIF_F_HW_VLAN_CTAG_FILTER ; netdev->hw_features |=NETIF_F_HW_VLAN_STAG_TX ; -if (!!MLX5_CAP_GEN (mdev, shampo) && -mlx5e_check_fragmented_striding_rq_cap (mdev)) - netdev->hw_features |=NETIF_F_GRO_HW ; -if (mlx5e_tunnel_any_tx_proto_supported (mdev)) { netdev->hw_enc_features |=NETIF_F_HW_CSUM ; netdev->hw_enc_features |=NETIF_F_TSO ;
Signature · 1 line
2.36.1
PATCH 9/11 net/mlx5e: CT: Fix support for GRE tuples
In reply to Saeed MahameedFrom: Paul Blakey <[email protected]>
cited commit removed support for GRE tuples when software steering was enabled.
To bring back support for GRE tuples, add GRE ipv4/ipv6 matchers.
--- .../mellanox/mlx5/core/en/tc/ct_fs_smfs.c | 56 +++++++++++-------- 1 file changed, 32 insertions(+), 24 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/tc/ct_fs_smfs.c +32 -24
@@ -23,7 +23,7 @@ struct mlx5_ct_fs_smfs_matcher { };struct mlx5_ct_fs_smfs_matchers { -struct mlx5_ct_fs_smfs_matcher smfs_matchers[4 ]; +struct mlx5_ct_fs_smfs_matcher smfs_matchers[6 ];struct list_head used; };
@@ -44,7 +44,8 @@ struct mlx5_ct_fs_smfs_rule { };static inline void -mlx5_ct_fs_smfs_fill_mask (struct mlx5_ct_fs *fs,struct mlx5_flow_spec *spec,bool ipv4,bool tcp) +mlx5_ct_fs_smfs_fill_mask (struct mlx5_ct_fs *fs,struct mlx5_flow_spec *spec,bool ipv4,bool tcp, +bool gre) {void *headers_c =MLX5_ADDR_OF (fte_match_param, spec->match_criteria, outer_headers);
@@ -77,7 +78,7 @@ mlx5_ct_fs_smfs_fill_mask(struct mlx5_ct_fs *fs, struct mlx5_flow_spec *spec, boMLX5_SET_TO_ONES (fte_match_set_lyr_2_4, headers_c, tcp_dport);MLX5_SET (fte_match_set_lyr_2_4, headers_c, tcp_flags,ntohs (MLX5_CT_TCP_FLAGS_MASK )); - }else { + }else if (!gre) {MLX5_SET_TO_ONES (fte_match_set_lyr_2_4, headers_c, udp_sport);MLX5_SET_TO_ONES (fte_match_set_lyr_2_4, headers_c, udp_dport); }
@@ -87,7 +88,7 @@ mlx5_ct_fs_smfs_fill_mask(struct mlx5_ct_fs *fs, struct mlx5_flow_spec *spec, bostatic struct mlx5dr_matcher *mlx5_ct_fs_smfs_matcher_create (struct mlx5_ct_fs *fs,struct mlx5dr_table *tbl,bool ipv4, -bool tcp,u32 priority) +bool tcp,bool gre,u32 priority) {struct mlx5dr_matcher *dr_matcher;struct mlx5_flow_spec *spec;
@@ -96,7 +97,7 @@ mlx5_ct_fs_smfs_matcher_create(struct mlx5_ct_fs *fs, struct mlx5dr_table *tbl,if (!spec)return ERR_PTR (-ENOMEM ); -mlx5_ct_fs_smfs_fill_mask (fs, spec, ipv4, tcp); +mlx5_ct_fs_smfs_fill_mask (fs, spec, ipv4, tcp, gre); spec->match_criteria_enable =MLX5_MATCH_MISC_PARAMETERS_2 |MLX5_MATCH_OUTER_HEADERS ; dr_matcher =mlx5_smfs_matcher_create (tbl, priority, spec);
@@ -108,7 +109,7 @@ mlx5_ct_fs_smfs_matcher_create(struct mlx5_ct_fs *fs, struct mlx5dr_table *tbl, }static struct mlx5_ct_fs_smfs_matcher * -mlx5_ct_fs_smfs_matcher_get (struct mlx5_ct_fs *fs,bool nat,bool ipv4,bool tcp) +mlx5_ct_fs_smfs_matcher_get (struct mlx5_ct_fs *fs,bool nat,bool ipv4,bool tcp,bool gre) {struct mlx5_ct_fs_smfs *fs_smfs =mlx5_ct_fs_priv (fs);struct mlx5_ct_fs_smfs_matcher *m, *smfs_matcher;
@@ -119,7 +120,7 @@ mlx5_ct_fs_smfs_matcher_get(struct mlx5_ct_fs *fs, bool nat, bool ipv4, bool tcpint prio; matchers = nat ? &fs_smfs->matchers_nat : &fs_smfs->matchers; - smfs_matcher = &matchers->smfs_matchers[ipv4 *2 + tcp]; + smfs_matcher = &matchers->smfs_matchers[ipv4 *3 + tcp *2 + gre];if (refcount_inc_not_zero (&smfs_matcher->ref))return smfs_matcher;
@@ -145,11 +146,11 @@ mlx5_ct_fs_smfs_matcher_get(struct mlx5_ct_fs *fs, bool nat, bool ipv4, bool tcp } tbl = nat ? fs_smfs->ct_nat_tbl : fs_smfs->ct_tbl; - dr_matcher =mlx5_ct_fs_smfs_matcher_create (fs, tbl, ipv4, tcp, prio); + dr_matcher =mlx5_ct_fs_smfs_matcher_create (fs, tbl, ipv4, tcp, gre, prio);if (IS_ERR (dr_matcher)) {netdev_warn (fs->netdev, -"ct_fs_smfs: failed to create matcher (nat %d, ipv4 %d, tcp %d), err: %ld\n" , - nat, ipv4, tcp,PTR_ERR (dr_matcher)); +"ct_fs_smfs: failed to create matcher (nat %d, ipv4 %d, tcp %d, gre %d), err: %ld\n" , + nat, ipv4, tcp, gre,PTR_ERR (dr_matcher)); smfs_matcher =ERR_CAST (dr_matcher); goto out_unlock;
@@ -222,16 +223,17 @@ mlx5_ct_fs_smfs_destroy(struct mlx5_ct_fs *fs)static inline bool mlx5_tc_ct_valid_used_dissector_keys (const u32 used_keys) { -#define DISSECTOR_BIT (name) BIT(FLOW_DISSECTOR_KEY_ ## name) -const u32 basic_keys =DISSECTOR_BIT (BASIC ) |DISSECTOR_BIT (CONTROL ) | -DISSECTOR_BIT (PORTS ) |DISSECTOR_BIT (META ); -const u32 ipv4_tcp = basic_keys |DISSECTOR_BIT (IPV4_ADDRS ) |DISSECTOR_BIT (TCP ); -const u32 ipv4_udp = basic_keys |DISSECTOR_BIT (IPV4_ADDRS ); -const u32 ipv6_tcp = basic_keys |DISSECTOR_BIT (IPV6_ADDRS ) |DISSECTOR_BIT (TCP ); -const u32 ipv6_udp = basic_keys |DISSECTOR_BIT (IPV6_ADDRS ); +#define DISS_BIT (name) BIT(FLOW_DISSECTOR_KEY_ ## name) +const u32 basic_keys =DISS_BIT (BASIC ) |DISS_BIT (CONTROL ) |DISS_BIT (META ); +const u32 ipv4_tcp = basic_keys |DISS_BIT (IPV4_ADDRS ) |DISS_BIT (PORTS ) |DISS_BIT (TCP ); +const u32 ipv6_tcp = basic_keys |DISS_BIT (IPV6_ADDRS ) |DISS_BIT (PORTS ) |DISS_BIT (TCP ); +const u32 ipv4_udp = basic_keys |DISS_BIT (IPV4_ADDRS ) |DISS_BIT (PORTS ); +const u32 ipv6_udp = basic_keys |DISS_BIT (IPV6_ADDRS ) |DISS_BIT (PORTS ); +const u32 ipv4_gre = basic_keys |DISS_BIT (IPV4_ADDRS ); +const u32 ipv6_gre = basic_keys |DISS_BIT (IPV6_ADDRS );return (used_keys == ipv4_tcp || used_keys == ipv4_udp || used_keys == ipv6_tcp || - used_keys == ipv6_udp); + used_keys == ipv6_udp || used_keys == ipv4_gre || used_keys == ipv6_gre); } static bool
@@ -254,20 +256,24 @@ mlx5_ct_fs_smfs_ct_validate_flow_rule(struct mlx5_ct_fs *fs, struct flow_rule *fflow_rule_match_control (flow_rule, &control);flow_rule_match_ipv4_addrs (flow_rule, &ipv4_addrs);flow_rule_match_ipv6_addrs (flow_rule, &ipv6_addrs); -flow_rule_match_ports (flow_rule, &ports); -flow_rule_match_tcp (flow_rule, &tcp); +if (basic.key->ip_proto !=IPPROTO_GRE ) +flow_rule_match_ports (flow_rule, &ports); +if (basic.key->ip_proto ==IPPROTO_TCP ) +flow_rule_match_tcp (flow_rule, &tcp);if (basic.mask->n_proto !=htons (0xFFFF ) || (basic.key->n_proto !=htons (ETH_P_IP ) && basic.key->n_proto !=htons (ETH_P_IPV6 )) || basic.mask->ip_proto !=0xFF || - (basic.key->ip_proto !=IPPROTO_UDP && basic.key->ip_proto !=IPPROTO_TCP )) { + (basic.key->ip_proto !=IPPROTO_UDP && basic.key->ip_proto !=IPPROTO_TCP && + basic.key->ip_proto !=IPPROTO_GRE )) {ct_dbg ("rule uses unexpected basic match (n_proto 0x%04x/0x%04x, ip_proto 0x%02x/0x%02x)" ,ntohs (basic.key->n_proto),ntohs (basic.mask->n_proto), basic.key->ip_proto, basic.mask->ip_proto);return false; } -if (ports.mask->src !=htons (0xFFFF ) || ports.mask->dst !=htons (0xFFFF )) { +if (basic.key->ip_proto !=IPPROTO_GRE && + (ports.mask->src !=htons (0xFFFF ) || ports.mask->dst !=htons (0xFFFF ))) {ct_dbg ("rule uses ports match (src 0x%04x, dst 0x%04x)" , ports.mask->src, ports.mask->dst);return false;
@@ -291,7 +297,7 @@ mlx5_ct_fs_smfs_ct_rule_add(struct mlx5_ct_fs *fs, struct mlx5_flow_spec *spec,struct mlx5dr_action *actions[5 ];struct mlx5dr_rule *rule;int num_actions =0 , err; -bool nat, tcp, ipv4; +bool nat, tcp, ipv4, gre;if (!mlx5_ct_fs_smfs_ct_validate_flow_rule (fs, flow_rule))return ERR_PTR (-EOPNOTSUPP );
@@ -314,8 +320,10 @@ mlx5_ct_fs_smfs_ct_rule_add(struct mlx5_ct_fs *fs, struct mlx5_flow_spec *spec, ipv4 =mlx5e_tc_get_ip_version (spec, true) ==4 ; tcp =MLX5_GET (fte_match_param, spec->match_value, outer_headers.ip_protocol) ==IPPROTO_TCP ; + gre =MLX5_GET (fte_match_param, spec->match_value, + outer_headers.ip_protocol) ==IPPROTO_GRE ; - smfs_matcher =mlx5_ct_fs_smfs_matcher_get (fs, nat, ipv4, tcp); + smfs_matcher =mlx5_ct_fs_smfs_matcher_get (fs, nat, ipv4, tcp, gre);if (IS_ERR (smfs_matcher)) { err =PTR_ERR (smfs_matcher); goto err_matcher;
Signature · 1 line
2.36.1
PATCH 10/11 net/mlx5e: CT: Fix setting flow_source for smfs ct tuples
In reply to Saeed MahameedFrom: Paul Blakey <[email protected]>
Cited patch sets flow_source to ANY overriding the provided spec flow_source, avoiding the optimization done by commit c9c079b4deaa ("net/mlx5: CT: Set flow source hint from provided tuple device").
To fix the above, set the dr_rule flow_source from provided flow spec.
--- drivers/net/ethernet/mellanox/mlx5/core/en/tc/ct_fs_smfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/mellanox/mlx5/core/en/tc/ct_fs_smfs.c +1 -1
@@ -330,7 +330,7 @@ mlx5_ct_fs_smfs_ct_rule_add(struct mlx5_ct_fs *fs, struct mlx5_flow_spec *spec, } rule =mlx5_smfs_rule_create (smfs_matcher->dr_matcher, spec, num_actions, actions, -MLX5_FLOW_CONTEXT_FLOW_SOURCE_ANY_VPORT ); + spec->flow_context.flow_source);if (!rule) { err = -EINVAL ; goto err_create;
Signature · 1 line
2.36.1
PATCH 11/11 net/mlx5: Drain fw_reset when removing device
In reply to Saeed MahameedFrom: Shay Drory <[email protected]>
In case fw sync reset is called in parallel to device removal, device might stuck in the following deadlock:
CPU 0 CPU 1
----- -----
remove_one
uninit_one (locks intf_state_mutex)mlx5_sync_reset_now_event()
work in fw_reset->wq.
mlx5_enter_error_state()
mutex_lock (intf_state_mutex)
cleanup_once
fw_reset_cleanup()
destroy_workqueue(fw_reset->wq)Drain the fw_reset WQ, and make sure no new work is being queued, before entering uninit_one().
The Drain is done before devlink_unregister() since fw_reset, in some flows, is using devlink API devlink_remote_reload_actions_performed().
--- .../ethernet/mellanox/mlx5/core/fw_reset.c | 25 ++++++++++++++++--- .../ethernet/mellanox/mlx5/core/fw_reset.h | 1 + .../net/ethernet/mellanox/mlx5/core/main.c | 4 +++ 3 files changed, 27 insertions(+), 3 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/fw_reset.c +22 -3
@@ -8,7 +8,8 @@enum {MLX5_FW_RESET_FLAGS_RESET_REQUESTED ,MLX5_FW_RESET_FLAGS_NACK_RESET_REQUEST , -MLX5_FW_RESET_FLAGS_PENDING_COMP +MLX5_FW_RESET_FLAGS_PENDING_COMP , +MLX5_FW_RESET_FLAGS_DROP_NEW_REQUESTS };struct mlx5_fw_reset {
@@ -208,7 +209,10 @@ static void poll_sync_reset(struct timer_list *t)if (fatal_error) {mlx5_core_warn (dev,"Got Device Reset\n" ); -queue_work (fw_reset->wq, &fw_reset->reset_reload_work); +if (!test_bit (MLX5_FW_RESET_FLAGS_DROP_NEW_REQUESTS , &fw_reset->reset_flags)) +queue_work (fw_reset->wq, &fw_reset->reset_reload_work); +else +mlx5_core_err (dev,"Device is being removed, Drop new reset work\n" );return ; }
@@ -433,9 +437,12 @@ static int fw_reset_event_notifier(struct notifier_block *nb, unsigned long acti struct mlx5_fw_reset *fw_reset = mlx5_nb_cof(nb, struct mlx5_fw_reset, nb);struct mlx5_eqe *eqe = data; +if (test_bit (MLX5_FW_RESET_FLAGS_DROP_NEW_REQUESTS , &fw_reset->reset_flags)) +return NOTIFY_DONE ; +switch (eqe->sub_type) {case MLX5_GENERAL_SUBTYPE_FW_LIVE_PATCH_EVENT : -queue_work (fw_reset->wq, &fw_reset->fw_live_patch_work); +queue_work (fw_reset->wq, &fw_reset->fw_live_patch_work);break ;case MLX5_GENERAL_SUBTYPE_PCI_SYNC_FOR_FW_UPDATE_EVENT :mlx5_sync_reset_events_handle (fw_reset, eqe);
@@ -479,6 +486,18 @@ void mlx5_fw_reset_events_stop(struct mlx5_core_dev *dev)mlx5_eq_notifier_unregister (dev, &dev->priv.fw_reset->nb); } +void mlx5_drain_fw_reset (struct mlx5_core_dev *dev) +{ +struct mlx5_fw_reset *fw_reset = dev->priv.fw_reset; + +set_bit (MLX5_FW_RESET_FLAGS_DROP_NEW_REQUESTS , &fw_reset->reset_flags); +cancel_work_sync (&fw_reset->fw_live_patch_work); +cancel_work_sync (&fw_reset->reset_request_work); +cancel_work_sync (&fw_reset->reset_reload_work); +cancel_work_sync (&fw_reset->reset_now_work); +cancel_work_sync (&fw_reset->reset_abort_work); +} +int mlx5_fw_reset_init (struct mlx5_core_dev *dev) {struct mlx5_fw_reset *fw_reset =kzalloc (sizeof (*fw_reset),GFP_KERNEL );
drivers/net/ethernet/mellanox/mlx5/core/fw_reset.h +1 -0
@@ -16,6 +16,7 @@ int mlx5_fw_reset_set_live_patch(struct mlx5_core_dev *dev);int mlx5_fw_reset_wait_reset_done (struct mlx5_core_dev *dev);void mlx5_fw_reset_events_start (struct mlx5_core_dev *dev);void mlx5_fw_reset_events_stop (struct mlx5_core_dev *dev); +void mlx5_drain_fw_reset (struct mlx5_core_dev *dev);int mlx5_fw_reset_init (struct mlx5_core_dev *dev);void mlx5_fw_reset_cleanup (struct mlx5_core_dev *dev);
drivers/net/ethernet/mellanox/mlx5/core/main.c +4 -0
@@ -1627,6 +1627,10 @@ static void remove_one(struct pci_dev *pdev)struct mlx5_core_dev *dev =pci_get_drvdata (pdev);struct devlink *devlink =priv_to_devlink (dev); +/* mlx5_drain_fw_reset() is using devlink APIs. Hence, we must drain +* fw_reset before unregistering the devlink. +*/ +mlx5_drain_fw_reset (dev);devlink_unregister (devlink);mlx5_sriov_disable (pdev);mlx5_crdump_disable (dev);
Signature · 1 line
2.36.1