mlx4: Add alias_guid mechanism
For IB ports, we paravirtualize the GUID at index 0 on slaves. The
GUID at index 0 seen by a slave is the actual GUID occupying the GUID
table at the slave-id index.
The driver, by default, requests at startup time that subnet manager
populate its entire guid table with GUIDs. These guids are then mapped
(paravirtualized) to the slaves, and appear for each slave as its GUID
at index 0.
Until each slave has such a guid, its port status is DOWN.
The guid table is cached to support special QP paravirtualization, and
event propagation to slaves on guid change (we test to see if the guid
really changed before propagating an event to the slave).
To support this caching, add capability to __mlx4_ib_query_gid() to
obtain the network view (i.e., physical view) gid at index X, not just
the host (paravirtualized) view.
Based on a patch from Erez Shitrit <erezsh@mellanox.com>
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <roland@purestorage.com>
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index b959fe4..7d97578 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -59,6 +59,10 @@
MODULE_LICENSE("Dual BSD/GPL");
MODULE_VERSION(DRV_VERSION);
+int mlx4_ib_sm_guid_assign = 1;
+module_param_named(sm_guid_assign, mlx4_ib_sm_guid_assign, int, 0444);
+MODULE_PARM_DESC(sm_guid_assign, "Enable SM alias_GUID assignment if sm_guid_assign > 0 (Default: 1)");
+
static const char mlx4_ib_version[] =
DRV_NAME ": Mellanox ConnectX InfiniBand driver v"
DRV_VERSION " (" DRV_RELDATE ")\n";
@@ -349,12 +353,15 @@
return __mlx4_ib_query_port(ibdev, port, props, 0);
}
-static int __mlx4_ib_query_gid(struct ib_device *ibdev, u8 port, int index,
- union ib_gid *gid)
+int __mlx4_ib_query_gid(struct ib_device *ibdev, u8 port, int index,
+ union ib_gid *gid, int netw_view)
{
struct ib_smp *in_mad = NULL;
struct ib_smp *out_mad = NULL;
int err = -ENOMEM;
+ struct mlx4_ib_dev *dev = to_mdev(ibdev);
+ int clear = 0;
+ int mad_ifc_flags = MLX4_MAD_IFC_IGNORE_KEYS;
in_mad = kzalloc(sizeof *in_mad, GFP_KERNEL);
out_mad = kmalloc(sizeof *out_mad, GFP_KERNEL);
@@ -365,18 +372,29 @@
in_mad->attr_id = IB_SMP_ATTR_PORT_INFO;
in_mad->attr_mod = cpu_to_be32(port);
- err = mlx4_MAD_IFC(to_mdev(ibdev), MLX4_MAD_IFC_IGNORE_KEYS, port,
- NULL, NULL, in_mad, out_mad);
+ if (mlx4_is_mfunc(dev->dev) && netw_view)
+ mad_ifc_flags |= MLX4_MAD_IFC_NET_VIEW;
+
+ err = mlx4_MAD_IFC(dev, mad_ifc_flags, port, NULL, NULL, in_mad, out_mad);
if (err)
goto out;
memcpy(gid->raw, out_mad->data + 8, 8);
+ if (mlx4_is_mfunc(dev->dev) && !netw_view) {
+ if (index) {
+ /* For any index > 0, return the null guid */
+ err = 0;
+ clear = 1;
+ goto out;
+ }
+ }
+
init_query_mad(in_mad);
in_mad->attr_id = IB_SMP_ATTR_GUID_INFO;
in_mad->attr_mod = cpu_to_be32(index / 8);
- err = mlx4_MAD_IFC(to_mdev(ibdev), MLX4_MAD_IFC_IGNORE_KEYS, port,
+ err = mlx4_MAD_IFC(dev, mad_ifc_flags, port,
NULL, NULL, in_mad, out_mad);
if (err)
goto out;
@@ -384,6 +402,8 @@
memcpy(gid->raw + 8, out_mad->data + (index % 8) * 8, 8);
out:
+ if (clear)
+ memset(gid->raw + 8, 0, 8);
kfree(in_mad);
kfree(out_mad);
return err;
@@ -403,7 +423,7 @@
union ib_gid *gid)
{
if (rdma_port_get_link_layer(ibdev, port) == IB_LINK_LAYER_INFINIBAND)
- return __mlx4_ib_query_gid(ibdev, port, index, gid);
+ return __mlx4_ib_query_gid(ibdev, port, index, gid, 0);
else
return iboe_query_gid(ibdev, port, index, gid);
}
@@ -1566,6 +1586,11 @@
case MLX4_DEV_EVENT_PORT_UP:
if (p > ibdev->num_ports)
return;
+ if (mlx4_is_master(dev) &&
+ rdma_port_get_link_layer(&ibdev->ib_dev, p) ==
+ IB_LINK_LAYER_INFINIBAND) {
+ mlx4_ib_invalidate_all_guid_record(ibdev, p);
+ }
ibev.event = IB_EVENT_PORT_ACTIVE;
break;