[PATCH 4.2 127/134] IB/mlx4: Fix potential deadlock when sending mad to wire

From: Greg Kroah-Hartman
Date: Sat Sep 26 2015 - 17:59:37 EST

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


From: Jack Morgenstein <jackm@xxxxxxxxxxxxxxxxxx>

commit 90c1d8b6350cca9d8a234f03c77a317a7613bcee upstream.

send_mad_to_wire takes the same spinlock that is taken in
the interrupt context. Therefore, it needs irqsave/restore.

Fixes: b9c5d6a64358 ('IB/mlx4: Add multicast group (MCG) paravirtualization for SR-IOV')
Signed-off-by: Jack Morgenstein <jackm@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Or Gerlitz <ogerlitz@xxxxxxxxxxxx>
Signed-off-by: Doug Ledford <dledford@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

drivers/infiniband/hw/mlx4/mcg.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)

--- a/drivers/infiniband/hw/mlx4/mcg.c
+++ b/drivers/infiniband/hw/mlx4/mcg.c
@@ -206,15 +206,16 @@ static int send_mad_to_wire(struct mlx4_
struct mlx4_ib_dev *dev = ctx->dev;
struct ib_ah_attr ah_attr;
+ unsigned long flags;

- spin_lock(&dev->sm_lock);
+ spin_lock_irqsave(&dev->sm_lock, flags);
if (!dev->sm_ah[ctx->port - 1]) {
/* port is not yet Active, sm_ah not ready */
- spin_unlock(&dev->sm_lock);
+ spin_unlock_irqrestore(&dev->sm_lock, flags);
return -EAGAIN;
mlx4_ib_query_ah(dev->sm_ah[ctx->port - 1], &ah_attr);
- spin_unlock(&dev->sm_lock);
+ spin_unlock_irqrestore(&dev->sm_lock, flags);
return mlx4_ib_send_to_wire(dev, mlx4_master_func_num(dev->dev),
ctx->port, IB_QPT_GSI, 0, 1, IB_QP1_QKEY,
&ah_attr, NULL, mad);

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/