net/mlx5: fs, fix UAF in flow counter release

Fix a kernel trace [1] caused by releasing an HWS action of a local flow
counter in mlx5_cmd_hws_delete_fte(), where the HWS action refcount and
mutex were not initialized and the counter struct could already be freed
when deleting the rule.

Fix it by adding the missing initializations and adding refcount for the
local flow counter struct.

[1] Kernel log:
 Call Trace:
  <TASK>
  dump_stack_lvl+0x34/0x48
  mlx5_fs_put_hws_action.part.0.cold+0x21/0x94 [mlx5_core]
  mlx5_fc_put_hws_action+0x96/0xad [mlx5_core]
  mlx5_fs_destroy_fs_actions+0x8b/0x152 [mlx5_core]
  mlx5_cmd_hws_delete_fte+0x5a/0xa0 [mlx5_core]
  del_hw_fte+0x1ce/0x260 [mlx5_core]
  mlx5_del_flow_rules+0x12d/0x240 [mlx5_core]
  ? ttwu_queue_wakelist+0xf4/0x110
  mlx5_ib_destroy_flow+0x103/0x1b0 [mlx5_ib]
  uverbs_free_flow+0x20/0x50 [ib_uverbs]
  destroy_hw_idr_uobject+0x1b/0x50 [ib_uverbs]
  uverbs_destroy_uobject+0x34/0x1a0 [ib_uverbs]
  uobj_destroy+0x3c/0x80 [ib_uverbs]
  ib_uverbs_run_method+0x23e/0x360 [ib_uverbs]
  ? uverbs_finalize_object+0x60/0x60 [ib_uverbs]
  ib_uverbs_cmd_verbs+0x14f/0x2c0 [ib_uverbs]
  ? do_tty_write+0x1a9/0x270
  ? file_tty_write.constprop.0+0x98/0xc0
  ? new_sync_write+0xfc/0x190
  ib_uverbs_ioctl+0xd7/0x160 [ib_uverbs]
  __x64_sys_ioctl+0x87/0xc0
  do_syscall_64+0x59/0x90

Fixes: b581f42669 ("net/mlx5: fs, manage flow counters HWS action sharing by refcount")
Signed-off-by: Moshe Shemesh <moshe@nvidia.com>
Reviewed-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://patch.msgid.link/1758525094-816583-2-git-send-email-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This commit is contained in:
Moshe Shemesh
2025-09-22 10:11:32 +03:00
committed by Jakub Kicinski
parent 0efdfbba2e
commit 6043819e70
5 changed files with 33 additions and 5 deletions

View File

@@ -308,6 +308,8 @@ struct mlx5_fc *mlx5_fc_create(struct mlx5_core_dev *dev, bool aging);
void mlx5_fc_destroy(struct mlx5_core_dev *dev, struct mlx5_fc *counter);
struct mlx5_fc *mlx5_fc_local_create(u32 counter_id, u32 offset, u32 bulk_size);
void mlx5_fc_local_destroy(struct mlx5_fc *counter);
void mlx5_fc_local_get(struct mlx5_fc *counter);
void mlx5_fc_local_put(struct mlx5_fc *counter);
u64 mlx5_fc_query_lastuse(struct mlx5_fc *counter);
void mlx5_fc_query_cached(struct mlx5_fc *counter,
u64 *bytes, u64 *packets, u64 *lastuse);