Commit 0254bc82 authored by Matan Barak's avatar Matan Barak Committed by David S. Miller

net/mlx4_core: Fix slave id computation for single port VF

The code that deals with computing the slave id based on a given GID
gave wrong results when the number of single port VFs wasn't the
same for port 1 vs. port 2 and the relevant VF is single ported on
port 2. As a result, incoming CM MADs were dispatched to the wrong VF.
Fixed that and added documentation to clarify the computation steps.

Fixes: 449fc488 ('net/mlx4: Adapt code for N-Port VF')
Signed-off-by: default avatarMatan Barak <matanb@mellanox.com>
Signed-off-by: default avatarOr Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parent 531d9014
...@@ -1106,6 +1106,9 @@ int mlx4_get_slave_from_roce_gid(struct mlx4_dev *dev, int port, u8 *gid, ...@@ -1106,6 +1106,9 @@ int mlx4_get_slave_from_roce_gid(struct mlx4_dev *dev, int port, u8 *gid,
} }
if (found_ix >= 0) { if (found_ix >= 0) {
/* Calculate a slave_gid which is the slave number in the gid
* table and not a globally unique slave number.
*/
if (found_ix < MLX4_ROCE_PF_GIDS) if (found_ix < MLX4_ROCE_PF_GIDS)
slave_gid = 0; slave_gid = 0;
else if (found_ix < MLX4_ROCE_PF_GIDS + (vf_gids % num_vfs) * else if (found_ix < MLX4_ROCE_PF_GIDS + (vf_gids % num_vfs) *
...@@ -1118,41 +1121,43 @@ int mlx4_get_slave_from_roce_gid(struct mlx4_dev *dev, int port, u8 *gid, ...@@ -1118,41 +1121,43 @@ int mlx4_get_slave_from_roce_gid(struct mlx4_dev *dev, int port, u8 *gid,
((vf_gids % num_vfs) * ((vf_gids / num_vfs + 1)))) / ((vf_gids % num_vfs) * ((vf_gids / num_vfs + 1)))) /
(vf_gids / num_vfs)) + vf_gids % num_vfs + 1; (vf_gids / num_vfs)) + vf_gids % num_vfs + 1;
/* Calculate the globally unique slave id */
if (slave_gid) { if (slave_gid) {
struct mlx4_active_ports exclusive_ports; struct mlx4_active_ports exclusive_ports;
struct mlx4_active_ports actv_ports; struct mlx4_active_ports actv_ports;
struct mlx4_slaves_pport slaves_pport_actv; struct mlx4_slaves_pport slaves_pport_actv;
unsigned max_port_p_one; unsigned max_port_p_one;
int num_slaves_before = 1; int num_vfs_before = 0;
int candidate_slave_gid;
/* Calculate how many VFs are on the previous port, if exists */
for (i = 1; i < port; i++) { for (i = 1; i < port; i++) {
bitmap_zero(exclusive_ports.ports, dev->caps.num_ports); bitmap_zero(exclusive_ports.ports, dev->caps.num_ports);
set_bit(i, exclusive_ports.ports); set_bit(i - 1, exclusive_ports.ports);
slaves_pport_actv = slaves_pport_actv =
mlx4_phys_to_slaves_pport_actv( mlx4_phys_to_slaves_pport_actv(
dev, &exclusive_ports); dev, &exclusive_ports);
num_slaves_before += bitmap_weight( num_vfs_before += bitmap_weight(
slaves_pport_actv.slaves, slaves_pport_actv.slaves,
dev->num_vfs + 1); dev->num_vfs + 1);
} }
if (slave_gid < num_slaves_before) { /* candidate_slave_gid isn't necessarily the correct slave, but
bitmap_zero(exclusive_ports.ports, dev->caps.num_ports); * it has the same number of ports and is assigned to the same
set_bit(port - 1, exclusive_ports.ports); * ports as the real slave we're looking for. On dual port VF,
slaves_pport_actv = * slave_gid = [single port VFs on port <port>] +
mlx4_phys_to_slaves_pport_actv( * [offset of the current slave from the first dual port VF] +
dev, &exclusive_ports); * 1 (for the PF).
slave_gid += bitmap_weight( */
slaves_pport_actv.slaves, candidate_slave_gid = slave_gid + num_vfs_before;
dev->num_vfs + 1) -
num_slaves_before; actv_ports = mlx4_get_active_ports(dev, candidate_slave_gid);
}
actv_ports = mlx4_get_active_ports(dev, slave_gid);
max_port_p_one = find_first_bit( max_port_p_one = find_first_bit(
actv_ports.ports, dev->caps.num_ports) + actv_ports.ports, dev->caps.num_ports) +
bitmap_weight(actv_ports.ports, bitmap_weight(actv_ports.ports,
dev->caps.num_ports) + 1; dev->caps.num_ports) + 1;
/* Calculate the real slave number */
for (i = 1; i < max_port_p_one; i++) { for (i = 1; i < max_port_p_one; i++) {
if (i == port) if (i == port)
continue; continue;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment