net/mlx5: Fix tc max supported prio for nic mode
authorChris Mi <cmi@nvidia.com>
Tue, 14 Dec 2021 01:52:53 +0000 (03:52 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 5 Jan 2022 11:42:34 +0000 (12:42 +0100)
[ Upstream commit d671e109bd8548d067b27e39e183a484430bf102 ]

Only prio 1 is supported if firmware doesn't support ignore flow
level for nic mode. The offending commit removed the check wrongly.
Add it back.

Fixes: 9a99c8f1253a ("net/mlx5e: E-Switch, Offload all chain 0 priorities when modify header and forward action is not supported")
Signed-off-by: Chris Mi <cmi@nvidia.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
drivers/net/ethernet/mellanox/mlx5/core/lib/fs_chains.c

index 97e5845..d5e4763 100644 (file)
@@ -121,6 +121,9 @@ u32 mlx5_chains_get_nf_ft_chain(struct mlx5_fs_chains *chains)
 
 u32 mlx5_chains_get_prio_range(struct mlx5_fs_chains *chains)
 {
+       if (!mlx5_chains_prios_supported(chains))
+               return 1;
+
        if (mlx5_chains_ignore_flow_level_supported(chains))
                return UINT_MAX;