3.16.37-rc1 review patch. If anyone has any objections, please let me know.
------------------
From: Eli Cohen <e...@mellanox.com>
commit c9b254955b9f8814966f5dabd34c39d0e0a2b437 upstream.
If the caller specified IB_SEND_FENCE in the send flags of the work
request and no previous work request stated that the successive one
should be fenced, the work request would be executed without a fence.
This could result in RDMA read or atomic operations failure due to a MR
being invalidated. Fix this by adding the mlx5 enumeration for fencing
RDMA/atomic operations and fix the logic to apply this.
Fixes: e126ba97dba9 ('mlx5: Add driver for Mellanox Connect-IB adapters')
Signed-off-by: Eli Cohen <e...@mellanox.com>
Signed-off-by: Leon Romanovsky <l...@kernel.org>
Signed-off-by: Doug Ledford <dledf...@redhat.com>
Signed-off-by: Ben Hutchings <b...@decadent.org.uk>
---
drivers/infiniband/hw/mlx5/qp.c | 7 ++++---
include/linux/mlx5/qp.h | 1 +
2 files changed, 5 insertions(+), 3 deletions(-)
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -2419,10 +2419,11 @@ static u8 get_fence(u8 fence, struct ib_
return MLX5_FENCE_MODE_SMALL_AND_FENCE;
else
return fence;
-
- } else {
- return 0;
+ } else if (unlikely(wr->send_flags & IB_SEND_FENCE)) {
+ return MLX5_FENCE_MODE_FENCE;
}
+
+ return 0;
}
static int begin_wqe(struct mlx5_ib_qp *qp, void **seg,
--- a/include/linux/mlx5/qp.h
+++ b/include/linux/mlx5/qp.h
@@ -140,6 +140,7 @@ enum {
enum {
MLX5_FENCE_MODE_NONE = 0 << 5,
MLX5_FENCE_MODE_INITIATOR_SMALL = 1 << 5,
+ MLX5_FENCE_MODE_FENCE = 2 << 5,
MLX5_FENCE_MODE_STRONG_ORDERING = 3 << 5,
MLX5_FENCE_MODE_SMALL_AND_FENCE = 4 << 5,
};
