[PATCH rdma-rc 0/2] IPoIB and mlx5 fixes

[PATCH rdma-rc 0/2] IPoIB and mlx5 fixes

[Leon Romanovsky]

There are two fixes in this series,

Fix from Alex who solved lockdep real complain about ABBA locks sequence.

Patch from Majd a little bit larger than I would like, but most of the
diffstat is simple code moves from one file to another.

The patches are available in the git repository at:
  git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git tags/rdma-rc-2017-12-21

	Thanks
---------------------------------------

Alex Vesker (1):
  IB/ipoib: Fix lockdep issue found on ipoib_ib_dev_heavy_flush

Majd Dibbiny (1):
  IB/mlx5: Fix congestion counters in LAG mode

 drivers/infiniband/hw/mlx5/cmd.c              | 11 ------
 drivers/infiniband/hw/mlx5/cmd.h              |  2 -
 drivers/infiniband/hw/mlx5/main.c             | 35 +++--------------
 drivers/infiniband/ulp/ipoib/ipoib_ib.c       |  7 ++--
 drivers/net/ethernet/mellanox/mlx5/core/lag.c | 56 +++++++++++++++++++++++++++
 include/linux/mlx5/driver.h                   |  4 ++
 6 files changed, 69 insertions(+), 46 deletions(-)

-- 
2.15.1

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH rdma-rc 1/2] IB/mlx5: Fix congestion counters in LAG mode

[Leon Romanovsky]

From: Majd Dibbiny <majd-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Congestion counters are counted and queried per physical function.
When working in LAG mode, CNP packets can be sent or received on both
of the functions, thus congestion counters should be aggregated from
the two physical functions.

Fixes: e1f24a79f424 ("IB/mlx5: Support congestion related counters")
Signed-off-by: Majd Dibbiny <majd-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Reviewed-by: Aviv Heller <avivh-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Signed-off-by: Leon Romanovsky <leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
---
 drivers/infiniband/hw/mlx5/cmd.c              | 11 ------
 drivers/infiniband/hw/mlx5/cmd.h              |  2 -
 drivers/infiniband/hw/mlx5/main.c             | 35 +++--------------
 drivers/net/ethernet/mellanox/mlx5/core/lag.c | 56 +++++++++++++++++++++++++++
 include/linux/mlx5/driver.h                   |  4 ++
 5 files changed, 66 insertions(+), 42 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/cmd.c b/drivers/infiniband/hw/mlx5/cmd.c
index 470995fa38d2..6f6712f87a73 100644
--- a/drivers/infiniband/hw/mlx5/cmd.c
+++ b/drivers/infiniband/hw/mlx5/cmd.c
@@ -47,17 +47,6 @@ int mlx5_cmd_null_mkey(struct mlx5_core_dev *dev, u32 *null_mkey)
 	return err;
 }
 
-int mlx5_cmd_query_cong_counter(struct mlx5_core_dev *dev,
-				bool reset, void *out, int out_size)
-{
-	u32 in[MLX5_ST_SZ_DW(query_cong_statistics_in)] = { };
-
-	MLX5_SET(query_cong_statistics_in, in, opcode,
-		 MLX5_CMD_OP_QUERY_CONG_STATISTICS);
-	MLX5_SET(query_cong_statistics_in, in, clear, reset);
-	return mlx5_cmd_exec(dev, in, sizeof(in), out, out_size);
-}
-
 int mlx5_cmd_query_cong_params(struct mlx5_core_dev *dev, int cong_point,
 			       void *out, int out_size)
 {
diff --git a/drivers/infiniband/hw/mlx5/cmd.h b/drivers/infiniband/hw/mlx5/cmd.h
index af4c24596274..78ffded7cc2c 100644
--- a/drivers/infiniband/hw/mlx5/cmd.h
+++ b/drivers/infiniband/hw/mlx5/cmd.h
@@ -37,8 +37,6 @@
 #include <linux/mlx5/driver.h>
 
 int mlx5_cmd_null_mkey(struct mlx5_core_dev *dev, u32 *null_mkey);
-int mlx5_cmd_query_cong_counter(struct mlx5_core_dev *dev,
-				bool reset, void *out, int out_size);
 int mlx5_cmd_query_cong_params(struct mlx5_core_dev *dev, int cong_point,
 			       void *out, int out_size);
 int mlx5_cmd_modify_cong_params(struct mlx5_core_dev *mdev,
diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index 543d0a4c8bf3..b4ef4d9b6ce5 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -3737,34 +3737,6 @@ static int mlx5_ib_query_q_counters(struct mlx5_ib_dev *dev,
 	return ret;
 }
 
-static int mlx5_ib_query_cong_counters(struct mlx5_ib_dev *dev,
-				       struct mlx5_ib_port *port,
-				       struct rdma_hw_stats *stats)
-{
-	int outlen = MLX5_ST_SZ_BYTES(query_cong_statistics_out);
-	void *out;
-	int ret, i;
-	int offset = port->cnts.num_q_counters;
-
-	out = kvzalloc(outlen, GFP_KERNEL);
-	if (!out)
-		return -ENOMEM;
-
-	ret = mlx5_cmd_query_cong_counter(dev->mdev, false, out, outlen);
-	if (ret)
-		goto free;
-
-	for (i = 0; i < port->cnts.num_cong_counters; i++) {
-		stats->value[i + offset] =
-			be64_to_cpup((__be64 *)(out +
-				     port->cnts.offsets[i + offset]));
-	}
-
-free:
-	kvfree(out);
-	return ret;
-}
-
 static int mlx5_ib_get_hw_stats(struct ib_device *ibdev,
 				struct rdma_hw_stats *stats,
 				u8 port_num, int index)
@@ -3782,7 +3754,12 @@ static int mlx5_ib_get_hw_stats(struct ib_device *ibdev,
 	num_counters = port->cnts.num_q_counters;
 
 	if (MLX5_CAP_GEN(dev->mdev, cc_query_allowed)) {
-		ret = mlx5_ib_query_cong_counters(dev, port, stats);
+		ret = mlx5_lag_query_cong_counters(dev->mdev,
+						   stats->value +
+						   port->cnts.num_q_counters,
+						   port->cnts.num_cong_counters,
+						   port->cnts.offsets +
+						   port->cnts.num_q_counters);
 		if (ret)
 			return ret;
 		num_counters += port->cnts.num_cong_counters;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
index f26f97fe4666..582b2f18010a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
@@ -137,6 +137,17 @@ int mlx5_cmd_destroy_vport_lag(struct mlx5_core_dev *dev)
 }
 EXPORT_SYMBOL(mlx5_cmd_destroy_vport_lag);
 
+static int mlx5_cmd_query_cong_counter(struct mlx5_core_dev *dev,
+				       bool reset, void *out, int out_size)
+{
+	u32 in[MLX5_ST_SZ_DW(query_cong_statistics_in)] = { };
+
+	MLX5_SET(query_cong_statistics_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_CONG_STATISTICS);
+	MLX5_SET(query_cong_statistics_in, in, clear, reset);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, out_size);
+}
+
 static struct mlx5_lag *mlx5_lag_dev_get(struct mlx5_core_dev *dev)
 {
 	return dev->priv.lag;
@@ -633,3 +644,48 @@ bool mlx5_lag_intf_add(struct mlx5_interface *intf, struct mlx5_priv *priv)
 	/* If bonded, we do not add an IB device for PF1. */
 	return false;
 }
+
+int mlx5_lag_query_cong_counters(struct mlx5_core_dev *dev,
+				 u64 *values,
+				 int num_counters,
+				 size_t *offsets)
+{
+	int outlen = MLX5_ST_SZ_BYTES(query_cong_statistics_out);
+	struct mlx5_core_dev *mdev[MLX5_MAX_PORTS];
+	struct mlx5_lag *ldev;
+	int num_ports;
+	int ret, i, j;
+	void *out;
+
+	out = kvzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	memset(values, 0, sizeof(*values) * num_counters);
+
+	mutex_lock(&lag_mutex);
+	ldev = mlx5_lag_dev_get(dev);
+	if (ldev && mlx5_lag_is_bonded(ldev)) {
+		num_ports = MLX5_MAX_PORTS;
+		mdev[0] = ldev->pf[0].dev;
+		mdev[1] = ldev->pf[1].dev;
+	} else {
+		num_ports = 1;
+		mdev[0] = dev;
+	}
+
+	for (i = 0; i < num_ports; ++i) {
+		ret = mlx5_cmd_query_cong_counter(mdev[i], false, out, outlen);
+		if (ret)
+			goto unlock;
+
+		for (j = 0; j < num_counters; ++j)
+			values[j] += be64_to_cpup((__be64 *)(out + offsets[j]));
+	}
+
+unlock:
+	mutex_unlock(&lag_mutex);
+	kvfree(out);
+	return ret;
+}
+EXPORT_SYMBOL(mlx5_lag_query_cong_counters);
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index a886b51511ab..8846919356ca 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -1164,6 +1164,10 @@ int mlx5_cmd_create_vport_lag(struct mlx5_core_dev *dev);
 int mlx5_cmd_destroy_vport_lag(struct mlx5_core_dev *dev);
 bool mlx5_lag_is_active(struct mlx5_core_dev *dev);
 struct net_device *mlx5_lag_get_roce_netdev(struct mlx5_core_dev *dev);
+int mlx5_lag_query_cong_counters(struct mlx5_core_dev *dev,
+				 u64 *values,
+				 int num_counters,
+				 size_t *offsets);
 struct mlx5_uars_page *mlx5_get_uars_page(struct mlx5_core_dev *mdev);
 void mlx5_put_uars_page(struct mlx5_core_dev *mdev, struct mlx5_uars_page *up);
 
-- 
2.15.1

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH rdma-rc 2/2] IB/ipoib: Fix lockdep issue found on ipoib_ib_dev_heavy_flush

[Leon Romanovsky]

From: Alex Vesker <valex-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

The locking order of vlan_rwsem (LOCK A) and then rtnl (LOCK B),
contradicts other flows such as ipoib_open possibly causing a deadlock.
To prevent this deadlock heavy flush is called with RTNL locked and
only then tries to acquire vlan_rwsem.
This deadlock is possible only when there are child interfaces.

[  140.941758] ======================================================
[  140.946276] WARNING: possible circular locking dependency detected
[  140.950950] 4.15.0-rc1+ #9 Tainted: G           O
[  140.954797] ------------------------------------------------------
[  140.959424] kworker/u32:1/146 is trying to acquire lock:
[  140.963450]  (rtnl_mutex){+.+.}, at: [<ffffffffc083516a>] __ipoib_ib_dev_flush+0x2da/0x4e0 [ib_ipoib]
[  140.970006]
but task is already holding lock:
[  140.975141]  (&priv->vlan_rwsem){++++}, at: [<ffffffffc0834ee1>] __ipoib_ib_dev_flush+0x51/0x4e0 [ib_ipoib]
[  140.982105]
which lock already depends on the new lock.
[  140.990023]
the existing dependency chain (in reverse order) is:
[  140.998650]
-> #1 (&priv->vlan_rwsem){++++}:
[  141.005276]        down_read+0x4d/0xb0
[  141.009560]        ipoib_open+0xad/0x120 [ib_ipoib]
[  141.014400]        __dev_open+0xcb/0x140
[  141.017919]        __dev_change_flags+0x1a4/0x1e0
[  141.022133]        dev_change_flags+0x23/0x60
[  141.025695]        devinet_ioctl+0x704/0x7d0
[  141.029156]        sock_do_ioctl+0x20/0x50
[  141.032526]        sock_ioctl+0x221/0x300
[  141.036079]        do_vfs_ioctl+0xa6/0x6d0
[  141.039656]        SyS_ioctl+0x74/0x80
[  141.042811]        entry_SYSCALL_64_fastpath+0x1f/0x96
[  141.046891]
-> #0 (rtnl_mutex){+.+.}:
[  141.051701]        lock_acquire+0xd4/0x220
[  141.055212]        __mutex_lock+0x88/0x970
[  141.058631]        __ipoib_ib_dev_flush+0x2da/0x4e0 [ib_ipoib]
[  141.063160]        __ipoib_ib_dev_flush+0x71/0x4e0 [ib_ipoib]
[  141.067648]        process_one_work+0x1f5/0x610
[  141.071429]        worker_thread+0x4a/0x3f0
[  141.074890]        kthread+0x141/0x180
[  141.078085]        ret_from_fork+0x24/0x30
[  141.081559]

other info that might help us debug this:
[  141.088967]  Possible unsafe locking scenario:
[  141.094280]        CPU0                    CPU1
[  141.097953]        ----                    ----
[  141.101640]   lock(&priv->vlan_rwsem);
[  141.104771]                                lock(rtnl_mutex);
[  141.109207]                                lock(&priv->vlan_rwsem);
[  141.114032]   lock(rtnl_mutex);
[  141.116800]
 *** DEADLOCK ***

Fixes: b4b678b06f6e ("IB/ipoib: Grab rtnl lock on heavy flush when calling ndo_open/stop")
Signed-off-by: Alex Vesker <valex-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Signed-off-by: Leon Romanovsky <leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
---
 drivers/infiniband/ulp/ipoib/ipoib_ib.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
index 3b96cdaf9a83..e6151a29c412 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
@@ -1236,13 +1236,10 @@ static void __ipoib_ib_dev_flush(struct ipoib_dev_priv *priv,
 		ipoib_ib_dev_down(dev);
 
 	if (level == IPOIB_FLUSH_HEAVY) {
-		rtnl_lock();
 		if (test_bit(IPOIB_FLAG_INITIALIZED, &priv->flags))
 			ipoib_ib_dev_stop(dev);
 
-		result = ipoib_ib_dev_open(dev);
-		rtnl_unlock();
-		if (result)
+		if (ipoib_ib_dev_open(dev))
 			return;
 
 		if (netif_queue_stopped(dev))
@@ -1282,7 +1279,9 @@ void ipoib_ib_dev_flush_heavy(struct work_struct *work)
 	struct ipoib_dev_priv *priv =
 		container_of(work, struct ipoib_dev_priv, flush_heavy);
 
+	rtnl_lock();
 	__ipoib_ib_dev_flush(priv, IPOIB_FLUSH_HEAVY, 0);
+	rtnl_unlock();
 }
 
 void ipoib_ib_dev_cleanup(struct net_device *dev)
-- 
2.15.1

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH rdma-rc 0/2] IPoIB and mlx5 fixes

[Jason Gunthorpe]

On Thu, Dec 21, 2017 at 05:38:25PM +0200, Leon Romanovsky wrote:
> There are two fixes in this series,
> 
> Fix from Alex who solved lockdep real complain about ABBA locks sequence.
> 
> Patch from Majd a little bit larger than I would like, but most of the
> diffstat is simple code moves from one file to another.
> 
> The patches are available in the git repository at:
>   git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git tags/rdma-rc-2017-12-21

Applied both to for-rc, thanks

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html