> -----Original Message----- > From: Colin Ian King [mailto:colin.king@xxxxxxxxxxxxx] > Sent: Tuesday, April 21, 2020 5:00 PM > To: Raed Salem <raeds@xxxxxxxxxxxx> > Cc: Boris Pismenny <borisp@xxxxxxxxxxxx>; Saeed Mahameed > <saeedm@xxxxxxxxxxxx>; Leon Romanovsky <leon@xxxxxxxxxx>; Tariq > Toukan <tariqt@xxxxxxxxxxxx>; netdev@xxxxxxxxxxxxxxx; linux- > rdma@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx > Subject: re: net/mlx5: IPsec, Refactor SA handle creation and destruction > > Hi, > > Static analysis with Coverity has detected a potential issue with the following > commit: > > commit 7dfee4b1d79e1800818abcfb47747b162c9a2d31 > Author: Raed Salem <raeds@xxxxxxxxxxxx> > Date: Wed Oct 23 17:04:13 2019 +0300 > > net/mlx5: IPsec, Refactor SA handle creation and destruction > > The issue is in mlx5_fpga_is_ipsec_device() in > drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.c as follows: > > > 710 Bitwise-and with zero > bit_and_with_zero: accel_xfrm->attrs.action & > MLX5_ACCEL_ESP_ACTION_DECRYPT is always 0. This occurs as the logical > operand of if. > > 711 if (accel_xfrm->attrs.action & MLX5_ACCEL_ESP_ACTION_DECRYPT) { > Logically dead code (DEADCODE) > > 712 err = ida_simple_get(&fipsec->halloc, 1, 0, GFP_KERNEL); > 713 if (err < 0) { > 714 context = ERR_PTR(err); > 715 goto exists; > 716 } > 717 > 718 sa_ctx->sa_handle = err; > 719 if (sa_handle) > 720 *sa_handle = sa_ctx->sa_handle; > 721 } > > in include/linux/mlx5/accel.h MLX5_ACCEL_ESP_ACTION_DECRYPT is defined > as zero: > > 50 enum mlx5_accel_esp_action { > 51 MLX5_ACCEL_ESP_ACTION_DECRYPT, > 52 MLX5_ACCEL_ESP_ACTION_ENCRYPT, > 53 }; > > > I believe there are some other instances of this bit-wise and-ing with zero, > e.g. in mlx5_fpga_ipsec_release_sa_ctx() we have: > > 855 if (sa_ctx->fpga_xfrm->accel_xfrm.attrs.action & > 856 MLX5_ACCEL_ESP_ACTION_DECRYPT) > > Colin Thanks for the catch ... will Post a fix
