IB/mlx5: avoid destroying a NULL mr in reg_user_mr error flow

commit 11d748045c6dadb279d1acdb6d2ea8f3f2ede85b upstream.

The mlx5_ib_reg_user_mr() function will attempt to call clean_mr() in
its error flow even though there is never a case where the error flow
occurs with a valid MR pointer to destroy.

Remove the clean_mr() call and the incorrect comment above it.

Fixes: b4cfe447d47b ("IB/mlx5: Implement on demand paging by adding
support for MMU notifiers")
Cc: Eli Cohen <eli@mellanox.com>
Signed-off-by: Haggai Eran <haggaie@mellanox.com>
Reviewed-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

diff --git a/drivers/infiniband/hw/mlx5/mr.c b/drivers/infiniband/hw/mlx5/mr.c
index bc9a0de897cb466d62d69dde0330ca96581953ca..dbb75c0de8486b43bf2d2fa88d5a4fca830d94cb 100644 (file)
--- a/drivers/infiniband/hw/mlx5/mr.c
+++ b/drivers/infiniband/hw/mlx5/mr.c
@@ -1118,19 +1118,7 @@ struct ib_mr *mlx5_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
        return &mr->ibmr;
 
 error:
-       /*
-        * Destroy the umem *before* destroying the MR, to ensure we
-        * will not have any in-flight notifiers when destroying the
-        * MR.
-        *
-        * As the MR is completely invalid to begin with, and this
-        * error path is only taken if we can't push the mr entry into
-        * the pagefault tree, this is safe.
-        */
-
        ib_umem_release(umem);
-       /* Kill the MR, and return an error code. */
-       clean_mr(mr);
        return ERR_PTR(err);
 }