Merge branches 'core', 'ipoib', 'iser', 'mlx4', 'ocrdma' and 'qib' into for-next
diff --git a/drivers/infiniband/core/umem.c b/drivers/infiniband/core/umem.c
index a3a2e9c..df0c4f6 100644
--- a/drivers/infiniband/core/umem.c
+++ b/drivers/infiniband/core/umem.c
@@ -105,6 +105,7 @@
umem->length = size;
umem->offset = addr & ~PAGE_MASK;
umem->page_size = PAGE_SIZE;
+ umem->pid = get_task_pid(current, PIDTYPE_PID);
/*
* We ask for writable memory if any access flags other than
* "remote read" are set. "Local write" and "remote write"
@@ -198,6 +199,7 @@
if (ret < 0) {
if (need_release)
__ib_umem_release(context->device, umem, 0);
+ put_pid(umem->pid);
kfree(umem);
} else
current->mm->pinned_vm = locked;
@@ -230,15 +232,19 @@
{
struct ib_ucontext *context = umem->context;
struct mm_struct *mm;
+ struct task_struct *task;
unsigned long diff;
__ib_umem_release(umem->context->device, umem, 1);
- mm = get_task_mm(current);
- if (!mm) {
- kfree(umem);
- return;
- }
+ task = get_pid_task(umem->pid, PIDTYPE_PID);
+ put_pid(umem->pid);
+ if (!task)
+ goto out;
+ mm = get_task_mm(task);
+ put_task_struct(task);
+ if (!mm)
+ goto out;
diff = PAGE_ALIGN(umem->length + umem->offset) >> PAGE_SHIFT;
@@ -262,9 +268,10 @@
} else
down_write(&mm->mmap_sem);
- current->mm->pinned_vm -= diff;
+ mm->pinned_vm -= diff;
up_write(&mm->mmap_sem);
mmput(mm);
+out:
kfree(umem);
}
EXPORT_SYMBOL(ib_umem_release);
diff --git a/drivers/infiniband/core/uverbs_marshall.c b/drivers/infiniband/core/uverbs_marshall.c
index e7bee46..abd9724 100644
--- a/drivers/infiniband/core/uverbs_marshall.c
+++ b/drivers/infiniband/core/uverbs_marshall.c
@@ -140,5 +140,9 @@
dst->packet_life_time = src->packet_life_time;
dst->preference = src->preference;
dst->packet_life_time_selector = src->packet_life_time_selector;
+
+ memset(dst->smac, 0, sizeof(dst->smac));
+ memset(dst->dmac, 0, sizeof(dst->dmac));
+ dst->vlan_id = 0xffff;
}
EXPORT_SYMBOL(ib_copy_path_rec_from_user);
diff --git a/drivers/infiniband/hw/ipath/ipath_user_pages.c b/drivers/infiniband/hw/ipath/ipath_user_pages.c
index dc66c45..1da1252 100644
--- a/drivers/infiniband/hw/ipath/ipath_user_pages.c
+++ b/drivers/infiniband/hw/ipath/ipath_user_pages.c
@@ -54,7 +54,7 @@
/* call with current->mm->mmap_sem held */
static int __ipath_get_user_pages(unsigned long start_page, size_t num_pages,
- struct page **p, struct vm_area_struct **vma)
+ struct page **p)
{
unsigned long lock_limit;
size_t got;
@@ -74,7 +74,7 @@
ret = get_user_pages(current, current->mm,
start_page + got * PAGE_SIZE,
num_pages - got, 1, 1,
- p + got, vma);
+ p + got, NULL);
if (ret < 0)
goto bail_release;
}
@@ -165,7 +165,7 @@
down_write(¤t->mm->mmap_sem);
- ret = __ipath_get_user_pages(start_page, num_pages, p, NULL);
+ ret = __ipath_get_user_pages(start_page, num_pages, p);
up_write(¤t->mm->mmap_sem);
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index e1e558a..c7586a1 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -59,6 +59,7 @@
#define MLX4_IB_FLOW_MAX_PRIO 0xFFF
#define MLX4_IB_FLOW_QPN_MASK 0xFFFFFF
+#define MLX4_IB_CARD_REV_A0 0xA0
MODULE_AUTHOR("Roland Dreier");
MODULE_DESCRIPTION("Mellanox ConnectX HCA InfiniBand driver");
@@ -119,6 +120,17 @@
return dmfs;
}
+static int num_ib_ports(struct mlx4_dev *dev)
+{
+ int ib_ports = 0;
+ int i;
+
+ mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_IB)
+ ib_ports++;
+
+ return ib_ports;
+}
+
static int mlx4_ib_query_device(struct ib_device *ibdev,
struct ib_device_attr *props)
{
@@ -126,6 +138,7 @@
struct ib_smp *in_mad = NULL;
struct ib_smp *out_mad = NULL;
int err = -ENOMEM;
+ int have_ib_ports;
in_mad = kzalloc(sizeof *in_mad, GFP_KERNEL);
out_mad = kmalloc(sizeof *out_mad, GFP_KERNEL);
@@ -142,6 +155,8 @@
memset(props, 0, sizeof *props);
+ have_ib_ports = num_ib_ports(dev->dev);
+
props->fw_ver = dev->dev->caps.fw_ver;
props->device_cap_flags = IB_DEVICE_CHANGE_PHY_PORT |
IB_DEVICE_PORT_ACTIVE_EVENT |
@@ -152,13 +167,15 @@
props->device_cap_flags |= IB_DEVICE_BAD_PKEY_CNTR;
if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_BAD_QKEY_CNTR)
props->device_cap_flags |= IB_DEVICE_BAD_QKEY_CNTR;
- if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_APM)
+ if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_APM && have_ib_ports)
props->device_cap_flags |= IB_DEVICE_AUTO_PATH_MIG;
if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_UD_AV_PORT)
props->device_cap_flags |= IB_DEVICE_UD_AV_PORT_ENFORCE;
if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_IPOIB_CSUM)
props->device_cap_flags |= IB_DEVICE_UD_IP_CSUM;
- if (dev->dev->caps.max_gso_sz && dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_BLH)
+ if (dev->dev->caps.max_gso_sz &&
+ (dev->dev->rev_id != MLX4_IB_CARD_REV_A0) &&
+ (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_BLH))
props->device_cap_flags |= IB_DEVICE_UD_TSO;
if (dev->dev->caps.bmme_flags & MLX4_BMME_FLAG_RESERVED_LKEY)
props->device_cap_flags |= IB_DEVICE_LOCAL_DMA_LKEY;
@@ -357,7 +374,7 @@
props->state = IB_PORT_DOWN;
props->phys_state = state_to_phys_state(props->state);
props->active_mtu = IB_MTU_256;
- spin_lock(&iboe->lock);
+ spin_lock_bh(&iboe->lock);
ndev = iboe->netdevs[port - 1];
if (!ndev)
goto out_unlock;
@@ -369,7 +386,7 @@
IB_PORT_ACTIVE : IB_PORT_DOWN;
props->phys_state = state_to_phys_state(props->state);
out_unlock:
- spin_unlock(&iboe->lock);
+ spin_unlock_bh(&iboe->lock);
out:
mlx4_free_cmd_mailbox(mdev->dev, mailbox);
return err;
@@ -811,11 +828,11 @@
if (!mqp->port)
return 0;
- spin_lock(&mdev->iboe.lock);
+ spin_lock_bh(&mdev->iboe.lock);
ndev = mdev->iboe.netdevs[mqp->port - 1];
if (ndev)
dev_hold(ndev);
- spin_unlock(&mdev->iboe.lock);
+ spin_unlock_bh(&mdev->iboe.lock);
if (ndev) {
ret = 1;
@@ -1262,11 +1279,11 @@
mutex_lock(&mqp->mutex);
ge = find_gid_entry(mqp, gid->raw);
if (ge) {
- spin_lock(&mdev->iboe.lock);
+ spin_lock_bh(&mdev->iboe.lock);
ndev = ge->added ? mdev->iboe.netdevs[ge->port - 1] : NULL;
if (ndev)
dev_hold(ndev);
- spin_unlock(&mdev->iboe.lock);
+ spin_unlock_bh(&mdev->iboe.lock);
if (ndev)
dev_put(ndev);
list_del(&ge->list);
@@ -1387,6 +1404,9 @@
int err;
struct mlx4_dev *dev = gw->dev->dev;
+ if (!gw->dev->ib_active)
+ return;
+
mailbox = mlx4_alloc_cmd_mailbox(dev);
if (IS_ERR(mailbox)) {
pr_warn("update gid table failed %ld\n", PTR_ERR(mailbox));
@@ -1417,6 +1437,9 @@
int err;
struct mlx4_dev *dev = gw->dev->dev;
+ if (!gw->dev->ib_active)
+ return;
+
mailbox = mlx4_alloc_cmd_mailbox(dev);
if (IS_ERR(mailbox)) {
pr_warn("reset gid table failed\n");
@@ -1551,7 +1574,7 @@
return 0;
iboe = &ibdev->iboe;
- spin_lock(&iboe->lock);
+ spin_lock_bh(&iboe->lock);
for (port = 1; port <= ibdev->dev->caps.num_ports; ++port)
if ((netif_is_bond_master(real_dev) &&
@@ -1561,7 +1584,7 @@
update_gid_table(ibdev, port, gid,
event == NETDEV_DOWN, 0);
- spin_unlock(&iboe->lock);
+ spin_unlock_bh(&iboe->lock);
return 0;
}
@@ -1634,13 +1657,21 @@
new_smac = mlx4_mac_to_u64(dev->dev_addr);
read_unlock(&dev_base_lock);
+ atomic64_set(&ibdev->iboe.mac[port - 1], new_smac);
+
+ /* no need for update QP1 and mac registration in non-SRIOV */
+ if (!mlx4_is_mfunc(ibdev->dev))
+ return;
+
mutex_lock(&ibdev->qp1_proxy_lock[port - 1]);
qp = ibdev->qp1_proxy[port - 1];
if (qp) {
int new_smac_index;
- u64 old_smac = qp->pri.smac;
+ u64 old_smac;
struct mlx4_update_qp_params update_params;
+ mutex_lock(&qp->mutex);
+ old_smac = qp->pri.smac;
if (new_smac == old_smac)
goto unlock;
@@ -1655,17 +1686,20 @@
release_mac = new_smac;
goto unlock;
}
-
+ /* if old port was zero, no mac was yet registered for this QP */
+ if (qp->pri.smac_port)
+ release_mac = old_smac;
qp->pri.smac = new_smac;
+ qp->pri.smac_port = port;
qp->pri.smac_index = new_smac_index;
-
- release_mac = old_smac;
}
unlock:
- mutex_unlock(&ibdev->qp1_proxy_lock[port - 1]);
if (release_mac != MLX4_IB_INVALID_MAC)
mlx4_unregister_mac(ibdev->dev, port, release_mac);
+ if (qp)
+ mutex_unlock(&qp->mutex);
+ mutex_unlock(&ibdev->qp1_proxy_lock[port - 1]);
}
static void mlx4_ib_get_dev_addr(struct net_device *dev,
@@ -1676,6 +1710,7 @@
struct inet6_dev *in6_dev;
union ib_gid *pgid;
struct inet6_ifaddr *ifp;
+ union ib_gid default_gid;
#endif
union ib_gid gid;
@@ -1696,12 +1731,15 @@
in_dev_put(in_dev);
}
#if IS_ENABLED(CONFIG_IPV6)
+ mlx4_make_default_gid(dev, &default_gid);
/* IPv6 gids */
in6_dev = in6_dev_get(dev);
if (in6_dev) {
read_lock_bh(&in6_dev->lock);
list_for_each_entry(ifp, &in6_dev->addr_list, if_list) {
pgid = (union ib_gid *)&ifp->addr;
+ if (!memcmp(pgid, &default_gid, sizeof(*pgid)))
+ continue;
update_gid_table(ibdev, port, pgid, 0, 0);
}
read_unlock_bh(&in6_dev->lock);
@@ -1723,24 +1761,33 @@
struct net_device *dev;
struct mlx4_ib_iboe *iboe = &ibdev->iboe;
int i;
+ int err = 0;
- for (i = 1; i <= ibdev->num_ports; ++i)
- if (reset_gid_table(ibdev, i))
- return -1;
+ for (i = 1; i <= ibdev->num_ports; ++i) {
+ if (rdma_port_get_link_layer(&ibdev->ib_dev, i) ==
+ IB_LINK_LAYER_ETHERNET) {
+ err = reset_gid_table(ibdev, i);
+ if (err)
+ goto out;
+ }
+ }
read_lock(&dev_base_lock);
- spin_lock(&iboe->lock);
+ spin_lock_bh(&iboe->lock);
for_each_netdev(&init_net, dev) {
u8 port = mlx4_ib_get_dev_port(dev, ibdev);
- if (port)
+ /* port will be non-zero only for ETH ports */
+ if (port) {
+ mlx4_ib_set_default_gid(ibdev, dev, port);
mlx4_ib_get_dev_addr(dev, ibdev, port);
+ }
}
- spin_unlock(&iboe->lock);
+ spin_unlock_bh(&iboe->lock);
read_unlock(&dev_base_lock);
-
- return 0;
+out:
+ return err;
}
static void mlx4_ib_scan_netdevs(struct mlx4_ib_dev *ibdev,
@@ -1754,7 +1801,7 @@
iboe = &ibdev->iboe;
- spin_lock(&iboe->lock);
+ spin_lock_bh(&iboe->lock);
mlx4_foreach_ib_transport_port(port, ibdev->dev) {
enum ib_port_state port_state = IB_PORT_NOP;
struct net_device *old_master = iboe->masters[port - 1];
@@ -1786,35 +1833,47 @@
port_state = (netif_running(curr_netdev) && netif_carrier_ok(curr_netdev)) ?
IB_PORT_ACTIVE : IB_PORT_DOWN;
mlx4_ib_set_default_gid(ibdev, curr_netdev, port);
+ if (curr_master) {
+ /* if using bonding/team and a slave port is down, we
+ * don't want the bond IP based gids in the table since
+ * flows that select port by gid may get the down port.
+ */
+ if (port_state == IB_PORT_DOWN) {
+ reset_gid_table(ibdev, port);
+ mlx4_ib_set_default_gid(ibdev,
+ curr_netdev,
+ port);
+ } else {
+ /* gids from the upper dev (bond/team)
+ * should appear in port's gid table
+ */
+ mlx4_ib_get_dev_addr(curr_master,
+ ibdev, port);
+ }
+ }
+ /* if bonding is used it is possible that we add it to
+ * masters only after IP address is assigned to the
+ * net bonding interface.
+ */
+ if (curr_master && (old_master != curr_master)) {
+ reset_gid_table(ibdev, port);
+ mlx4_ib_set_default_gid(ibdev,
+ curr_netdev, port);
+ mlx4_ib_get_dev_addr(curr_master, ibdev, port);
+ }
+
+ if (!curr_master && (old_master != curr_master)) {
+ reset_gid_table(ibdev, port);
+ mlx4_ib_set_default_gid(ibdev,
+ curr_netdev, port);
+ mlx4_ib_get_dev_addr(curr_netdev, ibdev, port);
+ }
} else {
reset_gid_table(ibdev, port);
}
- /* if using bonding/team and a slave port is down, we don't the bond IP
- * based gids in the table since flows that select port by gid may get
- * the down port.
- */
- if (curr_master && (port_state == IB_PORT_DOWN)) {
- reset_gid_table(ibdev, port);
- mlx4_ib_set_default_gid(ibdev, curr_netdev, port);
- }
- /* if bonding is used it is possible that we add it to masters
- * only after IP address is assigned to the net bonding
- * interface.
- */
- if (curr_master && (old_master != curr_master)) {
- reset_gid_table(ibdev, port);
- mlx4_ib_set_default_gid(ibdev, curr_netdev, port);
- mlx4_ib_get_dev_addr(curr_master, ibdev, port);
- }
-
- if (!curr_master && (old_master != curr_master)) {
- reset_gid_table(ibdev, port);
- mlx4_ib_set_default_gid(ibdev, curr_netdev, port);
- mlx4_ib_get_dev_addr(curr_netdev, ibdev, port);
- }
}
- spin_unlock(&iboe->lock);
+ spin_unlock_bh(&iboe->lock);
if (update_qps_port > 0)
mlx4_ib_update_qps(ibdev, dev, update_qps_port);
@@ -2156,6 +2215,9 @@
goto err_steer_free_bitmap;
}
+ for (j = 1; j <= ibdev->dev->caps.num_ports; j++)
+ atomic64_set(&iboe->mac[j - 1], ibdev->dev->caps.def_mac[j]);
+
if (ib_register_device(&ibdev->ib_dev, NULL))
goto err_steer_free_bitmap;
@@ -2192,12 +2254,8 @@
}
}
#endif
- for (i = 1 ; i <= ibdev->num_ports ; ++i)
- reset_gid_table(ibdev, i);
- rtnl_lock();
- mlx4_ib_scan_netdevs(ibdev, NULL, 0);
- rtnl_unlock();
- mlx4_ib_init_gid_table(ibdev);
+ if (mlx4_ib_init_gid_table(ibdev))
+ goto err_notif;
}
for (j = 0; j < ARRAY_SIZE(mlx4_class_attributes); ++j) {
@@ -2345,6 +2403,9 @@
struct mlx4_ib_dev *ibdev = ibdev_ptr;
int p;
+ ibdev->ib_active = false;
+ flush_workqueue(wq);
+
mlx4_ib_close_sriov(ibdev);
mlx4_ib_mad_cleanup(ibdev);
ib_unregister_device(&ibdev->ib_dev);
diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h
index e8cad39..6eb743f 100644
--- a/drivers/infiniband/hw/mlx4/mlx4_ib.h
+++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h
@@ -451,6 +451,7 @@
spinlock_t lock;
struct net_device *netdevs[MLX4_MAX_PORTS];
struct net_device *masters[MLX4_MAX_PORTS];
+ atomic64_t mac[MLX4_MAX_PORTS];
struct notifier_block nb;
struct notifier_block nb_inet;
struct notifier_block nb_inet6;
diff --git a/drivers/infiniband/hw/mlx4/mr.c b/drivers/infiniband/hw/mlx4/mr.c
index 9b0e80e..8f9325c 100644
--- a/drivers/infiniband/hw/mlx4/mr.c
+++ b/drivers/infiniband/hw/mlx4/mr.c
@@ -234,14 +234,13 @@
0);
if (IS_ERR(mmr->umem)) {
err = PTR_ERR(mmr->umem);
+ /* Prevent mlx4_ib_dereg_mr from free'ing invalid pointer */
mmr->umem = NULL;
goto release_mpt_entry;
}
n = ib_umem_page_count(mmr->umem);
shift = ilog2(mmr->umem->page_size);
- mmr->mmr.iova = virt_addr;
- mmr->mmr.size = length;
err = mlx4_mr_rereg_mem_write(dev->dev, &mmr->mmr,
virt_addr, length, n, shift,
*pmpt_entry);
@@ -249,6 +248,8 @@
ib_umem_release(mmr->umem);
goto release_mpt_entry;
}
+ mmr->mmr.iova = virt_addr;
+ mmr->mmr.size = length;
err = mlx4_ib_umem_write_mtt(dev, &mmr->mmr.mtt, mmr->umem);
if (err) {
@@ -262,6 +263,8 @@
* return a failure. But dereg_mr will free the resources.
*/
err = mlx4_mr_hw_write_mpt(dev->dev, &mmr->mmr, pmpt_entry);
+ if (!err && flags & IB_MR_REREG_ACCESS)
+ mmr->mmr.access = mr_access_flags;
release_mpt_entry:
mlx4_mr_hw_put_mpt(dev->dev, pmpt_entry);
diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c
index 6778045..577b477 100644
--- a/drivers/infiniband/hw/mlx4/qp.c
+++ b/drivers/infiniband/hw/mlx4/qp.c
@@ -964,9 +964,10 @@
MLX4_QP_STATE_RST, NULL, 0, 0, &qp->mqp))
pr_warn("modify QP %06x to RESET failed.\n",
qp->mqp.qpn);
- if (qp->pri.smac) {
+ if (qp->pri.smac || (!qp->pri.smac && qp->pri.smac_port)) {
mlx4_unregister_mac(dev->dev, qp->pri.smac_port, qp->pri.smac);
qp->pri.smac = 0;
+ qp->pri.smac_port = 0;
}
if (qp->alt.smac) {
mlx4_unregister_mac(dev->dev, qp->alt.smac_port, qp->alt.smac);
@@ -1325,7 +1326,8 @@
* If one was already assigned, but the new mac differs,
* unregister the old one and register the new one.
*/
- if (!smac_info->smac || smac_info->smac != smac) {
+ if ((!smac_info->smac && !smac_info->smac_port) ||
+ smac_info->smac != smac) {
/* register candidate now, unreg if needed, after success */
smac_index = mlx4_register_mac(dev->dev, port, smac);
if (smac_index >= 0) {
@@ -1390,21 +1392,13 @@
static int handle_eth_ud_smac_index(struct mlx4_ib_dev *dev, struct mlx4_ib_qp *qp, u8 *smac,
struct mlx4_qp_context *context)
{
- struct net_device *ndev;
u64 u64_mac;
int smac_index;
-
- ndev = dev->iboe.netdevs[qp->port - 1];
- if (ndev) {
- smac = ndev->dev_addr;
- u64_mac = mlx4_mac_to_u64(smac);
- } else {
- u64_mac = dev->dev->caps.def_mac[qp->port];
- }
+ u64_mac = atomic64_read(&dev->iboe.mac[qp->port - 1]);
context->pri_path.sched_queue = MLX4_IB_DEFAULT_SCHED_QUEUE | ((qp->port - 1) << 6);
- if (!qp->pri.smac) {
+ if (!qp->pri.smac && !qp->pri.smac_port) {
smac_index = mlx4_register_mac(dev->dev, qp->port, u64_mac);
if (smac_index >= 0) {
qp->pri.candidate_smac_index = smac_index;
@@ -1432,6 +1426,12 @@
int steer_qp = 0;
int err = -EINVAL;
+ /* APM is not supported under RoCE */
+ if (attr_mask & IB_QP_ALT_PATH &&
+ rdma_port_get_link_layer(&dev->ib_dev, qp->port) ==
+ IB_LINK_LAYER_ETHERNET)
+ return -ENOTSUPP;
+
context = kzalloc(sizeof *context, GFP_KERNEL);
if (!context)
return -ENOMEM;
@@ -1780,9 +1780,10 @@
if (qp->flags & MLX4_IB_QP_NETIF)
mlx4_ib_steer_qp_reg(dev, qp, 0);
}
- if (qp->pri.smac) {
+ if (qp->pri.smac || (!qp->pri.smac && qp->pri.smac_port)) {
mlx4_unregister_mac(dev->dev, qp->pri.smac_port, qp->pri.smac);
qp->pri.smac = 0;
+ qp->pri.smac_port = 0;
}
if (qp->alt.smac) {
mlx4_unregister_mac(dev->dev, qp->alt.smac_port, qp->alt.smac);
@@ -1806,11 +1807,12 @@
if (err && steer_qp)
mlx4_ib_steer_qp_reg(dev, qp, 0);
kfree(context);
- if (qp->pri.candidate_smac) {
+ if (qp->pri.candidate_smac ||
+ (!qp->pri.candidate_smac && qp->pri.candidate_smac_port)) {
if (err) {
mlx4_unregister_mac(dev->dev, qp->pri.candidate_smac_port, qp->pri.candidate_smac);
} else {
- if (qp->pri.smac)
+ if (qp->pri.smac || (!qp->pri.smac && qp->pri.smac_port))
mlx4_unregister_mac(dev->dev, qp->pri.smac_port, qp->pri.smac);
qp->pri.smac = qp->pri.candidate_smac;
qp->pri.smac_index = qp->pri.candidate_smac_index;
@@ -2083,6 +2085,16 @@
return 0;
}
+static void mlx4_u64_to_smac(u8 *dst_mac, u64 src_mac)
+{
+ int i;
+
+ for (i = ETH_ALEN; i; i--) {
+ dst_mac[i - 1] = src_mac & 0xff;
+ src_mac >>= 8;
+ }
+}
+
static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_send_wr *wr,
void *wqe, unsigned *mlx_seg_len)
{
@@ -2197,7 +2209,6 @@
}
if (is_eth) {
- u8 *smac;
struct in6_addr in6;
u16 pcp = (be32_to_cpu(ah->av.ib.sl_tclass_flowlabel) >> 29) << 13;
@@ -2210,12 +2221,17 @@
memcpy(&ctrl->imm, ah->av.eth.mac + 2, 4);
memcpy(&in6, sgid.raw, sizeof(in6));
- if (!mlx4_is_mfunc(to_mdev(ib_dev)->dev))
- smac = to_mdev(sqp->qp.ibqp.device)->
- iboe.netdevs[sqp->qp.port - 1]->dev_addr;
- else /* use the src mac of the tunnel */
- smac = ah->av.eth.s_mac;
- memcpy(sqp->ud_header.eth.smac_h, smac, 6);
+ if (!mlx4_is_mfunc(to_mdev(ib_dev)->dev)) {
+ u64 mac = atomic64_read(&to_mdev(ib_dev)->iboe.mac[sqp->qp.port - 1]);
+ u8 smac[ETH_ALEN];
+
+ mlx4_u64_to_smac(smac, mac);
+ memcpy(sqp->ud_header.eth.smac_h, smac, ETH_ALEN);
+ } else {
+ /* use the src mac of the tunnel */
+ memcpy(sqp->ud_header.eth.smac_h, ah->av.eth.s_mac, ETH_ALEN);
+ }
+
if (!memcmp(sqp->ud_header.eth.smac_h, sqp->ud_header.eth.dmac_h, 6))
mlx->flags |= cpu_to_be32(MLX4_WQE_CTRL_FORCE_LOOPBACK);
if (!is_vlan) {
diff --git a/drivers/infiniband/hw/qib/qib_debugfs.c b/drivers/infiniband/hw/qib/qib_debugfs.c
index 799a0c3..6abd3ed 100644
--- a/drivers/infiniband/hw/qib/qib_debugfs.c
+++ b/drivers/infiniband/hw/qib/qib_debugfs.c
@@ -193,6 +193,7 @@
struct qib_qp_iter *iter;
loff_t n = *pos;
+ rcu_read_lock();
iter = qib_qp_iter_init(s->private);
if (!iter)
return NULL;
@@ -224,7 +225,7 @@
static void _qp_stats_seq_stop(struct seq_file *s, void *iter_ptr)
{
- /* nothing for now */
+ rcu_read_unlock();
}
static int _qp_stats_seq_show(struct seq_file *s, void *iter_ptr)
diff --git a/drivers/infiniband/hw/qib/qib_qp.c b/drivers/infiniband/hw/qib/qib_qp.c
index 7fcc150..6ddc026 100644
--- a/drivers/infiniband/hw/qib/qib_qp.c
+++ b/drivers/infiniband/hw/qib/qib_qp.c
@@ -1325,7 +1325,6 @@
struct qib_qp *pqp = iter->qp;
struct qib_qp *qp;
- rcu_read_lock();
for (; n < dev->qp_table_size; n++) {
if (pqp)
qp = rcu_dereference(pqp->next);
@@ -1333,18 +1332,11 @@
qp = rcu_dereference(dev->qp_table[n]);
pqp = qp;
if (qp) {
- if (iter->qp)
- atomic_dec(&iter->qp->refcount);
- atomic_inc(&qp->refcount);
- rcu_read_unlock();
iter->qp = qp;
iter->n = n;
return 0;
}
}
- rcu_read_unlock();
- if (iter->qp)
- atomic_dec(&iter->qp->refcount);
return ret;
}
diff --git a/drivers/infiniband/hw/qib/qib_user_pages.c b/drivers/infiniband/hw/qib/qib_user_pages.c
index 2bc1d2b..74f90b2 100644
--- a/drivers/infiniband/hw/qib/qib_user_pages.c
+++ b/drivers/infiniband/hw/qib/qib_user_pages.c
@@ -52,7 +52,7 @@
* Call with current->mm->mmap_sem held.
*/
static int __qib_get_user_pages(unsigned long start_page, size_t num_pages,
- struct page **p, struct vm_area_struct **vma)
+ struct page **p)
{
unsigned long lock_limit;
size_t got;
@@ -69,7 +69,7 @@
ret = get_user_pages(current, current->mm,
start_page + got * PAGE_SIZE,
num_pages - got, 1, 1,
- p + got, vma);
+ p + got, NULL);
if (ret < 0)
goto bail_release;
}
@@ -136,7 +136,7 @@
down_write(¤t->mm->mmap_sem);
- ret = __qib_get_user_pages(start_page, num_pages, p, NULL);
+ ret = __qib_get_user_pages(start_page, num_pages, p);
up_write(¤t->mm->mmap_sem);
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
index d4e0057..ffb83b5 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
@@ -529,21 +529,13 @@
port_attr.state);
return;
}
+ priv->local_lid = port_attr.lid;
if (ib_query_gid(priv->ca, priv->port, 0, &priv->local_gid))
ipoib_warn(priv, "ib_query_gid() failed\n");
else
memcpy(priv->dev->dev_addr + 4, priv->local_gid.raw, sizeof (union ib_gid));
- {
- struct ib_port_attr attr;
-
- if (!ib_query_port(priv->ca, priv->port, &attr))
- priv->local_lid = attr.lid;
- else
- ipoib_warn(priv, "ib_query_port failed\n");
- }
-
if (!priv->broadcast) {
struct ipoib_mcast *broadcast;
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.c b/drivers/infiniband/ulp/iser/iscsi_iser.c
index 61ee91d..93ce62f 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.c
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.c
@@ -344,7 +344,6 @@
int is_leading)
{
struct iscsi_conn *conn = cls_conn->dd_data;
- struct iscsi_session *session;
struct iser_conn *ib_conn;
struct iscsi_endpoint *ep;
int error;
@@ -363,9 +362,17 @@
}
ib_conn = ep->dd_data;
- session = conn->session;
- if (iser_alloc_rx_descriptors(ib_conn, session))
- return -ENOMEM;
+ mutex_lock(&ib_conn->state_mutex);
+ if (ib_conn->state != ISER_CONN_UP) {
+ error = -EINVAL;
+ iser_err("iser_conn %p state is %d, teardown started\n",
+ ib_conn, ib_conn->state);
+ goto out;
+ }
+
+ error = iser_alloc_rx_descriptors(ib_conn, conn->session);
+ if (error)
+ goto out;
/* binds the iSER connection retrieved from the previously
* connected ep_handle to the iSCSI layer connection. exchanges
@@ -375,7 +382,9 @@
conn->dd_data = ib_conn;
ib_conn->iscsi_conn = conn;
- return 0;
+out:
+ mutex_unlock(&ib_conn->state_mutex);
+ return error;
}
static int
diff --git a/drivers/infiniband/ulp/iser/iscsi_iser.h b/drivers/infiniband/ulp/iser/iscsi_iser.h
index c877dad..9f0e0e3 100644
--- a/drivers/infiniband/ulp/iser/iscsi_iser.h
+++ b/drivers/infiniband/ulp/iser/iscsi_iser.h
@@ -69,7 +69,7 @@
#define DRV_NAME "iser"
#define PFX DRV_NAME ": "
-#define DRV_VER "1.4"
+#define DRV_VER "1.4.1"
#define iser_dbg(fmt, arg...) \
do { \
diff --git a/drivers/infiniband/ulp/iser/iser_verbs.c b/drivers/infiniband/ulp/iser/iser_verbs.c
index 3ef167f..3bfec4b 100644
--- a/drivers/infiniband/ulp/iser/iser_verbs.c
+++ b/drivers/infiniband/ulp/iser/iser_verbs.c
@@ -73,7 +73,7 @@
{
struct iser_cq_desc *cq_desc;
struct ib_device_attr *dev_attr = &device->dev_attr;
- int ret, i, j;
+ int ret, i;
ret = ib_query_device(device->ib_device, dev_attr);
if (ret) {
@@ -125,16 +125,20 @@
iser_cq_event_callback,
(void *)&cq_desc[i],
ISER_MAX_RX_CQ_LEN, i);
- if (IS_ERR(device->rx_cq[i]))
+ if (IS_ERR(device->rx_cq[i])) {
+ device->rx_cq[i] = NULL;
goto cq_err;
+ }
device->tx_cq[i] = ib_create_cq(device->ib_device,
NULL, iser_cq_event_callback,
(void *)&cq_desc[i],
ISER_MAX_TX_CQ_LEN, i);
- if (IS_ERR(device->tx_cq[i]))
+ if (IS_ERR(device->tx_cq[i])) {
+ device->tx_cq[i] = NULL;
goto cq_err;
+ }
if (ib_req_notify_cq(device->rx_cq[i], IB_CQ_NEXT_COMP))
goto cq_err;
@@ -160,14 +164,14 @@
handler_err:
ib_dereg_mr(device->mr);
dma_mr_err:
- for (j = 0; j < device->cqs_used; j++)
- tasklet_kill(&device->cq_tasklet[j]);
+ for (i = 0; i < device->cqs_used; i++)
+ tasklet_kill(&device->cq_tasklet[i]);
cq_err:
- for (j = 0; j < i; j++) {
- if (device->tx_cq[j])
- ib_destroy_cq(device->tx_cq[j]);
- if (device->rx_cq[j])
- ib_destroy_cq(device->rx_cq[j]);
+ for (i = 0; i < device->cqs_used; i++) {
+ if (device->tx_cq[i])
+ ib_destroy_cq(device->tx_cq[i]);
+ if (device->rx_cq[i])
+ ib_destroy_cq(device->rx_cq[i]);
}
ib_dealloc_pd(device->pd);
pd_err:
diff --git a/drivers/net/ethernet/mellanox/mlx4/mr.c b/drivers/net/ethernet/mellanox/mlx4/mr.c
index 7d717ec..193a6ad 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mr.c
+++ b/drivers/net/ethernet/mellanox/mlx4/mr.c
@@ -298,6 +298,7 @@
MLX4_CMD_TIME_CLASS_B, MLX4_CMD_WRAPPED);
}
+/* Must protect against concurrent access */
int mlx4_mr_hw_get_mpt(struct mlx4_dev *dev, struct mlx4_mr *mmr,
struct mlx4_mpt_entry ***mpt_entry)
{
@@ -305,13 +306,10 @@
int key = key_to_hw_index(mmr->key) & (dev->caps.num_mpts - 1);
struct mlx4_cmd_mailbox *mailbox = NULL;
- /* Make sure that at this point we have single-threaded access only */
-
if (mmr->enabled != MLX4_MPT_EN_HW)
return -EINVAL;
err = mlx4_HW2SW_MPT(dev, NULL, key);
-
if (err) {
mlx4_warn(dev, "HW2SW_MPT failed (%d).", err);
mlx4_warn(dev, "Most likely the MR has MWs bound to it.\n");
@@ -333,7 +331,6 @@
0, MLX4_CMD_QUERY_MPT,
MLX4_CMD_TIME_CLASS_B,
MLX4_CMD_WRAPPED);
-
if (err)
goto free_mailbox;
@@ -378,9 +375,10 @@
err = mlx4_SW2HW_MPT(dev, mailbox, key);
}
- mmr->pd = be32_to_cpu((*mpt_entry)->pd_flags) & MLX4_MPT_PD_MASK;
- if (!err)
+ if (!err) {
+ mmr->pd = be32_to_cpu((*mpt_entry)->pd_flags) & MLX4_MPT_PD_MASK;
mmr->enabled = MLX4_MPT_EN_HW;
+ }
return err;
}
EXPORT_SYMBOL_GPL(mlx4_mr_hw_write_mpt);
@@ -400,11 +398,12 @@
int mlx4_mr_hw_change_pd(struct mlx4_dev *dev, struct mlx4_mpt_entry *mpt_entry,
u32 pdn)
{
- u32 pd_flags = be32_to_cpu(mpt_entry->pd_flags);
+ u32 pd_flags = be32_to_cpu(mpt_entry->pd_flags) & ~MLX4_MPT_PD_MASK;
/* The wrapper function will put the slave's id here */
if (mlx4_is_mfunc(dev))
pd_flags &= ~MLX4_MPT_PD_VF_MASK;
- mpt_entry->pd_flags = cpu_to_be32((pd_flags & ~MLX4_MPT_PD_MASK) |
+
+ mpt_entry->pd_flags = cpu_to_be32(pd_flags |
(pdn & MLX4_MPT_PD_MASK)
| MLX4_MPT_PD_FLAG_EN_INV);
return 0;
@@ -600,14 +599,18 @@
{
int err;
- mpt_entry->start = cpu_to_be64(mr->iova);
- mpt_entry->length = cpu_to_be64(mr->size);
- mpt_entry->entity_size = cpu_to_be32(mr->mtt.page_shift);
+ mpt_entry->start = cpu_to_be64(iova);
+ mpt_entry->length = cpu_to_be64(size);
+ mpt_entry->entity_size = cpu_to_be32(page_shift);
err = mlx4_mtt_init(dev, npages, page_shift, &mr->mtt);
if (err)
return err;
+ mpt_entry->pd_flags &= cpu_to_be32(MLX4_MPT_PD_MASK |
+ MLX4_MPT_PD_FLAG_EN_INV);
+ mpt_entry->flags &= cpu_to_be32(MLX4_MPT_FLAG_FREE |
+ MLX4_MPT_FLAG_SW_OWNS);
if (mr->mtt.order < 0) {
mpt_entry->flags |= cpu_to_be32(MLX4_MPT_FLAG_PHYSICAL);
mpt_entry->mtt_addr = 0;
@@ -617,6 +620,14 @@
if (mr->mtt.page_shift == 0)
mpt_entry->mtt_sz = cpu_to_be32(1 << mr->mtt.order);
}
+ if (mr->mtt.order >= 0 && mr->mtt.page_shift == 0) {
+ /* fast register MR in free state */
+ mpt_entry->flags |= cpu_to_be32(MLX4_MPT_FLAG_FREE);
+ mpt_entry->pd_flags |= cpu_to_be32(MLX4_MPT_PD_FLAG_FAST_REG |
+ MLX4_MPT_PD_FLAG_RAE);
+ } else {
+ mpt_entry->flags |= cpu_to_be32(MLX4_MPT_FLAG_SW_OWNS);
+ }
mr->enabled = MLX4_MPT_EN_SW;
return 0;
diff --git a/drivers/net/ethernet/mellanox/mlx4/port.c b/drivers/net/ethernet/mellanox/mlx4/port.c
index 9ba0c1c..94eeb2c 100644
--- a/drivers/net/ethernet/mellanox/mlx4/port.c
+++ b/drivers/net/ethernet/mellanox/mlx4/port.c
@@ -103,7 +103,8 @@
int i;
for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
- if ((mac & MLX4_MAC_MASK) ==
+ if (table->refs[i] &&
+ (MLX4_MAC_MASK & mac) ==
(MLX4_MAC_MASK & be64_to_cpu(table->entries[i])))
return i;
}
@@ -165,12 +166,14 @@
mutex_lock(&table->mutex);
for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
- if (free < 0 && !table->entries[i]) {
- free = i;
+ if (!table->refs[i]) {
+ if (free < 0)
+ free = i;
continue;
}
- if (mac == (MLX4_MAC_MASK & be64_to_cpu(table->entries[i]))) {
+ if ((MLX4_MAC_MASK & mac) ==
+ (MLX4_MAC_MASK & be64_to_cpu(table->entries[i]))) {
/* MAC already registered, increment ref count */
err = i;
++table->refs[i];
diff --git a/include/rdma/ib_umem.h b/include/rdma/ib_umem.h
index 1ea0b65..a2bf41e 100644
--- a/include/rdma/ib_umem.h
+++ b/include/rdma/ib_umem.h
@@ -47,6 +47,7 @@
int writable;
int hugetlb;
struct work_struct work;
+ struct pid *pid;
struct mm_struct *mm;
unsigned long diff;
struct sg_table sg_head;