Skip to content

Commit

Permalink
{NET, IB}/mlx4: Add device managed flow steering firmware API
Browse files Browse the repository at this point in the history 
The driver is modified to support three operation modes.

If supported by firmware use the device managed flow steering
API, that which we call device managed steering mode. Else, if
the firmware supports the B0 steering mode use it, and finally,
if none of the above, use the A0 steering mode.

When the steering mode is device managed, the code is modified
such that L2 based rules set by the mlx4_en driver for Ethernet
unicast and multicast, and the IB stack multicast attach calls
done through the mlx4_ib driver are all routed to use the device
managed API.

When attaching rule using device managed flow steering API,
the firmware returns a 64 bit registration id, which is to be
provided during detach.

Currently the firmware is always programmed during HCA initialization
to use standard L2 hashing. Future work should be done to allow
configuring the flow-steering hash function with common, non
proprietary means.

Signed-off-by: Hadar Hen Zion <hadarh@mellanox.co.il>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
  • Loading branch information
Hadar Hen Zion authored and David S. Miller committed Jul 7, 2012
1 parent 8fcfb4d commit 0ff1fb6
Show file tree
Hide file tree
Showing 14 changed files with 758 additions and 81 deletions.
62 changes: 55 additions & 7 deletions drivers/infiniband/hw/mlx4/main.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -718,26 +718,53 @@ int mlx4_ib_add_mc(struct mlx4_ib_dev *mdev, struct mlx4_ib_qp *mqp,
return ret;
}

struct mlx4_ib_steering {
struct list_head list;
u64 reg_id;
union ib_gid gid;
};

static int mlx4_ib_mcg_attach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid)
{
int err;
struct mlx4_ib_dev *mdev = to_mdev(ibqp->device);
struct mlx4_ib_qp *mqp = to_mqp(ibqp);
u64 reg_id;
struct mlx4_ib_steering *ib_steering = NULL;

if (mdev->dev->caps.steering_mode ==
MLX4_STEERING_MODE_DEVICE_MANAGED) {
ib_steering = kmalloc(sizeof(*ib_steering), GFP_KERNEL);
if (!ib_steering)
return -ENOMEM;
}

err = mlx4_multicast_attach(mdev->dev, &mqp->mqp, gid->raw,
!!(mqp->flags & MLX4_IB_QP_BLOCK_MULTICAST_LOOPBACK),
MLX4_PROT_IB_IPV6);
err = mlx4_multicast_attach(mdev->dev, &mqp->mqp, gid->raw, mqp->port,
!!(mqp->flags &
MLX4_IB_QP_BLOCK_MULTICAST_LOOPBACK),
MLX4_PROT_IB_IPV6, &reg_id);
if (err)
return err;
goto err_malloc;

err = add_gid_entry(ibqp, gid);
if (err)
goto err_add;

if (ib_steering) {
memcpy(ib_steering->gid.raw, gid->raw, 16);
ib_steering->reg_id = reg_id;
mutex_lock(&mqp->mutex);
list_add(&ib_steering->list, &mqp->steering_rules);
mutex_unlock(&mqp->mutex);
}
return 0;

err_add:
mlx4_multicast_detach(mdev->dev, &mqp->mqp, gid->raw, MLX4_PROT_IB_IPV6);
mlx4_multicast_detach(mdev->dev, &mqp->mqp, gid->raw,
MLX4_PROT_IB_IPV6, reg_id);
err_malloc:
kfree(ib_steering);

return err;
}

Expand Down Expand Up @@ -765,9 +792,30 @@ static int mlx4_ib_mcg_detach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid)
u8 mac[6];
struct net_device *ndev;
struct mlx4_ib_gid_entry *ge;
u64 reg_id = 0;

if (mdev->dev->caps.steering_mode ==
MLX4_STEERING_MODE_DEVICE_MANAGED) {
struct mlx4_ib_steering *ib_steering;

mutex_lock(&mqp->mutex);
list_for_each_entry(ib_steering, &mqp->steering_rules, list) {
if (!memcmp(ib_steering->gid.raw, gid->raw, 16)) {
list_del(&ib_steering->list);
break;
}
}
mutex_unlock(&mqp->mutex);
if (&ib_steering->list == &mqp->steering_rules) {
pr_err("Couldn't find reg_id for mgid. Steering rule is left attached\n");
return -EINVAL;
}
reg_id = ib_steering->reg_id;
kfree(ib_steering);
}

err = mlx4_multicast_detach(mdev->dev,
&mqp->mqp, gid->raw, MLX4_PROT_IB_IPV6);
err = mlx4_multicast_detach(mdev->dev, &mqp->mqp, gid->raw,
MLX4_PROT_IB_IPV6, reg_id);
if (err)
return err;

Expand Down
1 change: 1 addition & 0 deletions drivers/infiniband/hw/mlx4/mlx4_ib.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -163,6 +163,7 @@ struct mlx4_ib_qp {
u8 state;
int mlx_type;
struct list_head gid_list;
struct list_head steering_rules;
};

struct mlx4_ib_srq {
Expand Down
1 change: 1 addition & 0 deletions drivers/infiniband/hw/mlx4/qp.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -495,6 +495,7 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
spin_lock_init(&qp->sq.lock);
spin_lock_init(&qp->rq.lock);
INIT_LIST_HEAD(&qp->gid_list);
INIT_LIST_HEAD(&qp->steering_rules);

qp->state = IB_QPS_RESET;
if (init_attr->sq_sig_type == IB_SIGNAL_ALL_WR)
Expand Down
21 changes: 13 additions & 8 deletions drivers/net/ethernet/mellanox/mlx4/en_netdev.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -463,7 +463,8 @@ static void mlx4_en_do_set_multicast(struct work_struct *work)
err = mlx4_multicast_detach(mdev->dev,
&priv->rss_map.indir_qp,
mc_list,
MLX4_PROT_ETH);
MLX4_PROT_ETH,
mclist->reg_id);
if (err)
en_err(priv, "Fail to detach multicast address\n");

Expand All @@ -475,11 +476,14 @@ static void mlx4_en_do_set_multicast(struct work_struct *work)
if (mclist->action == MCLIST_ADD) {
/* attach the address */
memcpy(&mc_list[10], mclist->addr, ETH_ALEN);
/* needed for B0 steering support */
mc_list[5] = priv->port;
err = mlx4_multicast_attach(mdev->dev,
&priv->rss_map.indir_qp,
mc_list, 0,
MLX4_PROT_ETH);
mc_list,
priv->port, 0,
MLX4_PROT_ETH,
&mclist->reg_id);
if (err)
en_err(priv, "Fail to attach multicast address\n");

Expand Down Expand Up @@ -827,9 +831,10 @@ int mlx4_en_start_port(struct net_device *dev)

/* Attach rx QP to bradcast address */
memset(&mc_list[10], 0xff, ETH_ALEN);
mc_list[5] = priv->port;
mc_list[5] = priv->port; /* needed for B0 steering support */
if (mlx4_multicast_attach(mdev->dev, &priv->rss_map.indir_qp, mc_list,
0, MLX4_PROT_ETH))
priv->port, 0, MLX4_PROT_ETH,
&priv->broadcast_id))
mlx4_warn(mdev, "Failed Attaching Broadcast\n");

/* Must redo promiscuous mode setup. */
Expand Down Expand Up @@ -886,14 +891,14 @@ void mlx4_en_stop_port(struct net_device *dev)

/* Detach All multicasts */
memset(&mc_list[10], 0xff, ETH_ALEN);
mc_list[5] = priv->port;
mc_list[5] = priv->port; /* needed for B0 steering support */
mlx4_multicast_detach(mdev->dev, &priv->rss_map.indir_qp, mc_list,
MLX4_PROT_ETH);
MLX4_PROT_ETH, priv->broadcast_id);
list_for_each_entry(mclist, &priv->curr_list, list) {
memcpy(&mc_list[10], mclist->addr, ETH_ALEN);
mc_list[5] = priv->port;
mlx4_multicast_detach(mdev->dev, &priv->rss_map.indir_qp,
mc_list, MLX4_PROT_ETH);
mc_list, MLX4_PROT_ETH, mclist->reg_id);
}
mlx4_en_clear_list(dev);
list_for_each_entry_safe(mclist, tmp, &priv->curr_list, list) {
Expand Down
91 changes: 74 additions & 17 deletions drivers/net/ethernet/mellanox/mlx4/fw.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -123,7 +123,8 @@ static void dump_dev_cap_flags2(struct mlx4_dev *dev, u64 flags)
static const char * const fname[] = {
[0] = "RSS support",
[1] = "RSS Toeplitz Hash Function support",
[2] = "RSS XOR Hash Function support"
[2] = "RSS XOR Hash Function support",
[3] = "Device manage flow steering support"
};
int i;

Expand Down Expand Up @@ -391,6 +392,8 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
#define QUERY_DEV_CAP_RSVD_XRC_OFFSET 0x66
#define QUERY_DEV_CAP_MAX_XRC_OFFSET 0x67
#define QUERY_DEV_CAP_MAX_COUNTERS_OFFSET 0x68
#define QUERY_DEV_CAP_FLOW_STEERING_RANGE_EN_OFFSET 0x76
#define QUERY_DEV_CAP_FLOW_STEERING_MAX_QP_OFFSET 0x77
#define QUERY_DEV_CAP_RDMARC_ENTRY_SZ_OFFSET 0x80
#define QUERY_DEV_CAP_QPC_ENTRY_SZ_OFFSET 0x82
#define QUERY_DEV_CAP_AUX_ENTRY_SZ_OFFSET 0x84
Expand Down Expand Up @@ -474,6 +477,12 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
dev_cap->num_ports = field & 0xf;
MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_MSG_SZ_OFFSET);
dev_cap->max_msg_sz = 1 << (field & 0x1f);
MLX4_GET(field, outbox, QUERY_DEV_CAP_FLOW_STEERING_RANGE_EN_OFFSET);
if (field & 0x80)
dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_FS_EN;
dev_cap->fs_log_max_ucast_qp_range_size = field & 0x1f;
MLX4_GET(field, outbox, QUERY_DEV_CAP_FLOW_STEERING_MAX_QP_OFFSET);
dev_cap->fs_max_num_qp_per_entry = field;
MLX4_GET(stat_rate, outbox, QUERY_DEV_CAP_RATE_SUPPORT_OFFSET);
dev_cap->stat_rate_support = stat_rate;
MLX4_GET(ext_flags, outbox, QUERY_DEV_CAP_EXT_FLAGS_OFFSET);
Expand Down Expand Up @@ -1061,6 +1070,15 @@ int mlx4_INIT_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param)
#define INIT_HCA_LOG_MC_HASH_SZ_OFFSET (INIT_HCA_MCAST_OFFSET + 0x16)
#define INIT_HCA_UC_STEERING_OFFSET (INIT_HCA_MCAST_OFFSET + 0x18)
#define INIT_HCA_LOG_MC_TABLE_SZ_OFFSET (INIT_HCA_MCAST_OFFSET + 0x1b)
#define INIT_HCA_DEVICE_MANAGED_FLOW_STEERING_EN 0x6
#define INIT_HCA_FS_PARAM_OFFSET 0x1d0
#define INIT_HCA_FS_BASE_OFFSET (INIT_HCA_FS_PARAM_OFFSET + 0x00)
#define INIT_HCA_FS_LOG_ENTRY_SZ_OFFSET (INIT_HCA_FS_PARAM_OFFSET + 0x12)
#define INIT_HCA_FS_LOG_TABLE_SZ_OFFSET (INIT_HCA_FS_PARAM_OFFSET + 0x1b)
#define INIT_HCA_FS_ETH_BITS_OFFSET (INIT_HCA_FS_PARAM_OFFSET + 0x21)
#define INIT_HCA_FS_ETH_NUM_ADDRS_OFFSET (INIT_HCA_FS_PARAM_OFFSET + 0x22)
#define INIT_HCA_FS_IB_BITS_OFFSET (INIT_HCA_FS_PARAM_OFFSET + 0x25)
#define INIT_HCA_FS_IB_NUM_ADDRS_OFFSET (INIT_HCA_FS_PARAM_OFFSET + 0x26)
#define INIT_HCA_TPT_OFFSET 0x0f0
#define INIT_HCA_DMPT_BASE_OFFSET (INIT_HCA_TPT_OFFSET + 0x00)
#define INIT_HCA_LOG_MPT_SZ_OFFSET (INIT_HCA_TPT_OFFSET + 0x0b)
Expand Down Expand Up @@ -1119,14 +1137,44 @@ int mlx4_INIT_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param)
MLX4_PUT(inbox, param->rdmarc_base, INIT_HCA_RDMARC_BASE_OFFSET);
MLX4_PUT(inbox, param->log_rd_per_qp, INIT_HCA_LOG_RD_OFFSET);

/* multicast attributes */

MLX4_PUT(inbox, param->mc_base, INIT_HCA_MC_BASE_OFFSET);
MLX4_PUT(inbox, param->log_mc_entry_sz, INIT_HCA_LOG_MC_ENTRY_SZ_OFFSET);
MLX4_PUT(inbox, param->log_mc_hash_sz, INIT_HCA_LOG_MC_HASH_SZ_OFFSET);
if (dev->caps.steering_mode == MLX4_STEERING_MODE_B0)
MLX4_PUT(inbox, (u8) (1 << 3), INIT_HCA_UC_STEERING_OFFSET);
MLX4_PUT(inbox, param->log_mc_table_sz, INIT_HCA_LOG_MC_TABLE_SZ_OFFSET);
/* steering attributes */
if (dev->caps.steering_mode ==
MLX4_STEERING_MODE_DEVICE_MANAGED) {
*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |=
cpu_to_be32(1 <<
INIT_HCA_DEVICE_MANAGED_FLOW_STEERING_EN);

MLX4_PUT(inbox, param->mc_base, INIT_HCA_FS_BASE_OFFSET);
MLX4_PUT(inbox, param->log_mc_entry_sz,
INIT_HCA_FS_LOG_ENTRY_SZ_OFFSET);
MLX4_PUT(inbox, param->log_mc_table_sz,
INIT_HCA_FS_LOG_TABLE_SZ_OFFSET);
/* Enable Ethernet flow steering
* with udp unicast and tcp unicast
*/
MLX4_PUT(inbox, param->fs_hash_enable_bits,
INIT_HCA_FS_ETH_BITS_OFFSET);
MLX4_PUT(inbox, (u16) MLX4_FS_NUM_OF_L2_ADDR,
INIT_HCA_FS_ETH_NUM_ADDRS_OFFSET);
/* Enable IPoIB flow steering
* with udp unicast and tcp unicast
*/
MLX4_PUT(inbox, param->fs_hash_enable_bits,
INIT_HCA_FS_IB_BITS_OFFSET);
MLX4_PUT(inbox, (u16) MLX4_FS_NUM_OF_L2_ADDR,
INIT_HCA_FS_IB_NUM_ADDRS_OFFSET);
} else {
MLX4_PUT(inbox, param->mc_base, INIT_HCA_MC_BASE_OFFSET);
MLX4_PUT(inbox, param->log_mc_entry_sz,
INIT_HCA_LOG_MC_ENTRY_SZ_OFFSET);
MLX4_PUT(inbox, param->log_mc_hash_sz,
INIT_HCA_LOG_MC_HASH_SZ_OFFSET);
MLX4_PUT(inbox, param->log_mc_table_sz,
INIT_HCA_LOG_MC_TABLE_SZ_OFFSET);
if (dev->caps.steering_mode == MLX4_STEERING_MODE_B0)
MLX4_PUT(inbox, (u8) (1 << 3),
INIT_HCA_UC_STEERING_OFFSET);
}

/* TPT attributes */

Expand Down Expand Up @@ -1188,15 +1236,24 @@ int mlx4_QUERY_HCA(struct mlx4_dev *dev,
MLX4_GET(param->rdmarc_base, outbox, INIT_HCA_RDMARC_BASE_OFFSET);
MLX4_GET(param->log_rd_per_qp, outbox, INIT_HCA_LOG_RD_OFFSET);

/* multicast attributes */
/* steering attributes */
if (dev->caps.steering_mode ==
MLX4_STEERING_MODE_DEVICE_MANAGED) {

MLX4_GET(param->mc_base, outbox, INIT_HCA_MC_BASE_OFFSET);
MLX4_GET(param->log_mc_entry_sz, outbox,
INIT_HCA_LOG_MC_ENTRY_SZ_OFFSET);
MLX4_GET(param->log_mc_hash_sz, outbox,
INIT_HCA_LOG_MC_HASH_SZ_OFFSET);
MLX4_GET(param->log_mc_table_sz, outbox,
INIT_HCA_LOG_MC_TABLE_SZ_OFFSET);
MLX4_GET(param->mc_base, outbox, INIT_HCA_FS_BASE_OFFSET);
MLX4_GET(param->log_mc_entry_sz, outbox,
INIT_HCA_FS_LOG_ENTRY_SZ_OFFSET);
MLX4_GET(param->log_mc_table_sz, outbox,
INIT_HCA_FS_LOG_TABLE_SZ_OFFSET);
} else {
MLX4_GET(param->mc_base, outbox, INIT_HCA_MC_BASE_OFFSET);
MLX4_GET(param->log_mc_entry_sz, outbox,
INIT_HCA_LOG_MC_ENTRY_SZ_OFFSET);
MLX4_GET(param->log_mc_hash_sz, outbox,
INIT_HCA_LOG_MC_HASH_SZ_OFFSET);
MLX4_GET(param->log_mc_table_sz, outbox,
INIT_HCA_LOG_MC_TABLE_SZ_OFFSET);
}

/* TPT attributes */

Expand Down
3 changes: 3 additions & 0 deletions drivers/net/ethernet/mellanox/mlx4/fw.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -78,6 +78,8 @@ struct mlx4_dev_cap {
u16 wavelength[MLX4_MAX_PORTS + 1];
u64 trans_code[MLX4_MAX_PORTS + 1];
u16 stat_rate_support;
int fs_log_max_ucast_qp_range_size;
int fs_max_num_qp_per_entry;
u64 flags;
u64 flags2;
int reserved_uars;
Expand Down Expand Up @@ -165,6 +167,7 @@ struct mlx4_init_hca_param {
u8 log_mpt_sz;
u8 log_uar_sz;
u8 uar_page_sz; /* log pg sz in 4k chunks */
u8 fs_hash_enable_bits;
};

struct mlx4_init_ib_param {
Expand Down
Loading

0 comments on commit 0ff1fb6

Please sign in to comment.