[PATCH 4.9 048/171] rxe: Fix a sleep-in-atomic bug in post_one_send

From: Greg Kroah-Hartman
Date: Thu Nov 08 2018 - 17:24:43 EST


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

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

[ Upstream commit 07d432bb97f19dd5e784175152f9fce3b2646133 ]

The driver may sleep under a spin lock, and the function call path is:
post_one_send (acquire the lock by spin_lock_irqsave)
init_send_wqe
copy_from_user --> may sleep

There is no flow that makes "qp->is_user" true, and copy_from_user may
cause bug when a non-user pointer is used. So the lines of copy_from_user
and check of "qp->is_user" are removed.

Signed-off-by: Jia-Ju Bai <baijiaju1990@xxxxxxx>
Reviewed-by: Leon Romanovsky <leonro@xxxxxxxxxxxx>
Acked-by: Moni Shoua <monis@xxxxxxxxxxxx>
Signed-off-by: Doug Ledford <dledford@xxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
drivers/infiniband/sw/rxe/rxe_verbs.c | 9 ++-------
1 file changed, 2 insertions(+), 7 deletions(-)

diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.c b/drivers/infiniband/sw/rxe/rxe_verbs.c
index ced416f5dffb..ef13082d6ca1 100644
--- a/drivers/infiniband/sw/rxe/rxe_verbs.c
+++ b/drivers/infiniband/sw/rxe/rxe_verbs.c
@@ -729,13 +729,8 @@ static int init_send_wqe(struct rxe_qp *qp, struct ib_send_wr *ibwr,

sge = ibwr->sg_list;
for (i = 0; i < num_sge; i++, sge++) {
- if (qp->is_user && copy_from_user(p, (__user void *)
- (uintptr_t)sge->addr, sge->length))
- return -EFAULT;
-
- else if (!qp->is_user)
- memcpy(p, (void *)(uintptr_t)sge->addr,
- sge->length);
+ memcpy(p, (void *)(uintptr_t)sge->addr,
+ sge->length);

p += sge->length;
}
--
2.17.1