3.16.40-rc1 review patch. If anyone has any objections, please let me know.
------------------
From: Jack Morgenstein <ja...@dev.mellanox.co.il>
commit a7e1f04905e5b2b90251974dddde781301b6be37 upstream.
When switching from polling-based fw commands to event-based fw
commands, there is a race condition which could cause a fw command
in another task to hang: that task will keep waiting for the polling
sempahore, but may never be able to acquire it. This is due to
mlx4_cmd_use_events, which "down"s the sempahore back to 0.
During driver initialization, this is not a problem, since no other
tasks which invoke FW commands are active.
However, there is a problem if the driver switches to polling mode
and then back to event mode during normal operation.
The "test_interrupts" feature does exactly that.
Running "ethtool -t <eth device> offline" causes the PF driver to
temporarily switch to polling mode, and then back to event mode.
(Note that for VF drivers, such switching is not performed).
Fix this by adding a read-write semaphore for protection when
switching between modes.
Fixes: 225c7b1feef1 ("IB/mlx4: Add a driver Mellanox ConnectX InfiniBand
adapters")
Signed-off-by: Jack Morgenstein <ja...@dev.mellanox.co.il>
Signed-off-by: Matan Barak <mat...@mellanox.com>
Signed-off-by: Tariq Toukan <tar...@mellanox.com>
Signed-off-by: David S. Miller <da...@davemloft.net>
[bwh: Backported to 3.16: adjust context, indentation]
Signed-off-by: Ben Hutchings <b...@decadent.org.uk>
---
drivers/net/ethernet/mellanox/mlx4/cmd.c | 23 +++++++++++++++++------
drivers/net/ethernet/mellanox/mlx4/mlx4.h | 2 ++
2 files changed, 19 insertions(+), 6 deletions(-)
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -606,14 +606,20 @@ int __mlx4_cmd(struct mlx4_dev *dev, u64
return -EIO;
if (!mlx4_is_mfunc(dev) || (native && mlx4_is_master(dev))) {
+ int ret;
+
+ down_read(&mlx4_priv(dev)->cmd.switch_sem);
if (mlx4_priv(dev)->cmd.use_events)
- return mlx4_cmd_wait(dev, in_param, out_param,
- out_is_imm, in_modifier,
- op_modifier, op, timeout);
+ ret = mlx4_cmd_wait(dev, in_param, out_param,
+ out_is_imm, in_modifier,
+ op_modifier, op, timeout);
else
- return mlx4_cmd_poll(dev, in_param, out_param,
- out_is_imm, in_modifier,
- op_modifier, op, timeout);
+ ret = mlx4_cmd_poll(dev, in_param, out_param,
+ out_is_imm, in_modifier,
+ op_modifier, op, timeout);
+
+ up_read(&mlx4_priv(dev)->cmd.switch_sem);
+ return ret;
}
return mlx4_slave_cmd(dev, in_param, out_param, out_is_imm,
in_modifier, op_modifier, op, timeout);
@@ -2092,6 +2098,7 @@ int mlx4_cmd_init(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
+ init_rwsem(&priv->cmd.switch_sem);
mutex_init(&priv->cmd.hcr_mutex);
mutex_init(&priv->cmd.slave_cmd_mutex);
sema_init(&priv->cmd.poll_sem, 1);
@@ -2188,6 +2195,7 @@ int mlx4_cmd_use_events(struct mlx4_dev
if (!priv->cmd.context)
return -ENOMEM;
+ down_write(&priv->cmd.switch_sem);
for (i = 0; i < priv->cmd.max_cmds; ++i) {
priv->cmd.context[i].token = i;
priv->cmd.context[i].next = i + 1;
@@ -2207,6 +2215,7 @@ int mlx4_cmd_use_events(struct mlx4_dev
down(&priv->cmd.poll_sem);
priv->cmd.use_events = 1;
+ up_write(&priv->cmd.switch_sem);
return err;
}
@@ -2219,6 +2228,7 @@ void mlx4_cmd_use_polling(struct mlx4_de
struct mlx4_priv *priv = mlx4_priv(dev);
int i;
+ down_write(&priv->cmd.switch_sem);
priv->cmd.use_events = 0;
for (i = 0; i < priv->cmd.max_cmds; ++i)
@@ -2227,6 +2237,7 @@ void mlx4_cmd_use_polling(struct mlx4_de
kfree(priv->cmd.context);
up(&priv->cmd.poll_sem);
+ up_write(&priv->cmd.switch_sem);
}
struct mlx4_cmd_mailbox *mlx4_alloc_cmd_mailbox(struct mlx4_dev *dev)
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -43,6 +43,7 @@
#include <linux/timer.h>
#include <linux/semaphore.h>
#include <linux/workqueue.h>
+#include <linux/rwsem.h>
#include <linux/mlx4/device.h>
#include <linux/mlx4/driver.h>
@@ -598,6 +599,7 @@ struct mlx4_cmd {
struct mutex slave_cmd_mutex;
struct semaphore poll_sem;
struct semaphore event_sem;
+ struct rw_semaphore switch_sem;
int max_cmds;
spinlock_t context_lock;
int free_head;
