mlx5e: add add missing BH locking around napi_schdule()
authorJakub Kicinski <kuba@kernel.org>
Wed, 5 May 2021 20:20:26 +0000 (13:20 -0700)
committerSaeed Mahameed <saeedm@nvidia.com>
Wed, 19 May 2021 06:01:55 +0000 (23:01 -0700)
commite63052a5dd3ce7979bff727a8f4bb6d6b3d1317b
treec11ae251b8bccc6cc9789315acbb293131cd81e4
parent6ff51ab8aa8fcbcddeeefce8ca705b575805d12b
mlx5e: add add missing BH locking around napi_schdule()

It's not correct to call napi_schedule() in pure process
context. Because we use __raise_softirq_irqoff() we require
callers to be in a context which will eventually lead to
softirq handling (hardirq, bh disabled, etc.).

With code as is users will see:

 NOHZ tick-stop error: Non-RCU local softirq work is pending, handler #08!!!

Fixes: a8dd7ac12fc3 ("net/mlx5e: Generalize RQ activation")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
drivers/net/ethernet/mellanox/mlx5/core/en_main.c