[PATCH 4.9 037/105] RDMA/uverbs: Fix the check for port number

[Greg Kroah-Hartman]

4.9-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Ismail, Mustafa <mustafa.ismail@xxxxxxxxx>

commit 5a7a88f1b488e4ee49eb3d5b82612d4d9ffdf2c3 upstream.

The port number is only valid if IB_QP_PORT is set in the mask.
So only check port number if it is valid to prevent modify_qp from
failing due to an invalid port number.

Fixes: 5ecce4c9b17b("Check port number supplied by user verbs cmds")
Reviewed-by: Steve Wise <swise@xxxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Mustafa Ismail <mustafa.ismail@xxxxxxxxx>
Tested-by: Mike Marciniszyn <mike.marciniszyn@xxxxxxxxx>
Signed-off-by: Doug Ledford <dledford@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>


---
 drivers/infiniband/core/uverbs_cmd.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

--- a/drivers/infiniband/core/uverbs_cmd.c
+++ b/drivers/infiniband/core/uverbs_cmd.c
@@ -2342,8 +2342,9 @@ ssize_t ib_uverbs_modify_qp(struct ib_uv
 	if (copy_from_user(&cmd, buf, sizeof cmd))
 		return -EFAULT;
 
-	if (cmd.port_num < rdma_start_port(ib_dev) ||
-	    cmd.port_num > rdma_end_port(ib_dev))
+	if ((cmd.attr_mask & IB_QP_PORT) &&
+	    (cmd.port_num < rdma_start_port(ib_dev) ||
+	     cmd.port_num > rdma_end_port(ib_dev)))
 		return -EINVAL;
 
 	INIT_UDATA(&udata, buf + sizeof cmd, NULL, in_len - sizeof cmd,