* [PATCH 07/25] mlx4_core: add port para-virtualization
From: Yevgeny Petrilin @ 2009-11-04 15:30 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
Ports are a shared resource among functions, so special behavior is needed here:
- Bring up ports if at least one function has done so.
- Bring down ports if all functions have done so.
- Aggregate IB port capabilities
Signed-off-by: Liran Liss <liranl-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/net/mlx4/cmd.c | 5 +++
drivers/net/mlx4/fw.c | 55 ++++++++++++++++++++++++++++++++++
drivers/net/mlx4/mlx4.h | 16 ++++++++++
drivers/net/mlx4/port.c | 76 +++++++++++++++++++++++++++++++++++++++++++++++
4 files changed, 152 insertions(+), 0 deletions(-)
diff --git a/drivers/net/mlx4/cmd.c b/drivers/net/mlx4/cmd.c
index dd23aba..d2646d1 100644
--- a/drivers/net/mlx4/cmd.c
+++ b/drivers/net/mlx4/cmd.c
@@ -526,6 +526,11 @@ static struct mlx4_cmd_info {
{MLX4_CMD_QUERY_FW, 0, 1, 0, NULL, NULL},
{MLX4_CMD_QUERY_ADAPTER, 0, 1, 0, NULL, NULL},
+ {MLX4_CMD_INIT_PORT, 0, 0, 0, NULL, mlx4_INIT_PORT_wrapper},
+ {MLX4_CMD_CLOSE_PORT, 0, 0, 0, NULL, mlx4_CLOSE_PORT_wrapper},
+ {MLX4_CMD_QUERY_PORT, 0, 1, 0, NULL, mlx4_QUERY_PORT_wrapper},
+ {MLX4_CMD_SET_PORT, 1, 0, 0, NULL, mlx4_SET_PORT_wrapper},
+
{MLX4_CMD_SW2HW_EQ, 1, 0, 0, NULL, NULL}, /* need verifier */
{MLX4_CMD_NOP, 0, 0, 0, NULL, NULL},
{MLX4_CMD_ALLOC_RES, 0, 0, 1, NULL, mlx4_RESOURCE_wrapper},
diff --git a/drivers/net/mlx4/fw.c b/drivers/net/mlx4/fw.c
index 4ca8060..a63f84a 100644
--- a/drivers/net/mlx4/fw.c
+++ b/drivers/net/mlx4/fw.c
@@ -135,6 +135,14 @@ int mlx4_MOD_STAT_CFG(struct mlx4_dev *dev, struct mlx4_mod_stat_cfg *cfg)
return err;
}
+int mlx4_QUERY_PORT_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox)
+{
+ return mlx4_cmd_box(dev, 0, outbox->dma, vhcr->in_modifier, 0, MLX4_CMD_QUERY_PORT,
+ MLX4_CMD_TIME_CLASS_B);
+}
+
int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
{
struct mlx4_cmd_mailbox *mailbox;
@@ -809,6 +817,29 @@ int mlx4_INIT_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param)
return err;
}
+int mlx4_INIT_PORT_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox)
+{
+ struct mlx4_priv *priv = mlx4_priv(dev);
+ int port;
+ int err;
+
+ port = vhcr->in_modifier;
+ if (priv->mfunc.master.slave_state[slave].init_port_mask & (1 << port))
+ return 0;
+
+ /* Enable port only if it was previously disabled */
+ if (!priv->mfunc.master.init_port_ref[port]) {
+ err = mlx4_INIT_PORT(dev, port);
+ if (err)
+ return err;
+ }
+ ++priv->mfunc.master.init_port_ref[port];
+ priv->mfunc.master.slave_state[slave].init_port_mask |= (1 << port);
+ return 0;
+}
+
int mlx4_INIT_PORT(struct mlx4_dev *dev, int port)
{
struct mlx4_cmd_mailbox *mailbox;
@@ -863,6 +894,30 @@ int mlx4_INIT_PORT(struct mlx4_dev *dev, int port)
}
EXPORT_SYMBOL_GPL(mlx4_INIT_PORT);
+int mlx4_CLOSE_PORT_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox)
+{
+ struct mlx4_priv *priv = mlx4_priv(dev);
+ int port;
+ int err;
+
+ port = vhcr->in_modifier;
+ if (!(priv->mfunc.master.slave_state[slave].init_port_mask & (1 << port)))
+ return 0;
+
+ /* CX1: master doesn't have interfaces - close port if this slave is
+ * the last user */
+ if (priv->mfunc.master.init_port_ref[port] == 1) {
+ err = mlx4_CLOSE_PORT(dev, port);
+ if (err)
+ return err;
+ }
+ --priv->mfunc.master.init_port_ref[port];
+ priv->mfunc.master.slave_state[slave].init_port_mask &= ~(1 << port);
+ return 0;
+}
+
int mlx4_CLOSE_PORT(struct mlx4_dev *dev, int port)
{
return mlx4_cmd(dev, 0, port, 0, MLX4_CMD_CLOSE_PORT, 1000);
diff --git a/drivers/net/mlx4/mlx4.h b/drivers/net/mlx4/mlx4.h
index fafa293..8a7a041 100644
--- a/drivers/net/mlx4/mlx4.h
+++ b/drivers/net/mlx4/mlx4.h
@@ -46,6 +46,7 @@
#include <linux/mlx4/driver.h>
#include <linux/mlx4/doorbell.h>
#include <linux/mlx4/cmd.h>
+#include <rdma/ib_verbs.h>
#define DRV_NAME "mlx4_core"
#define PFX DRV_NAME ": "
@@ -212,11 +213,14 @@ struct mlx4_slave_eqe {
struct mlx4_slave_state {
u8 comm_toggle;
u8 last_cmd;
+ u8 init_port_mask;
dma_addr_t vhcr_dma;
+ __be32 ib_cap_mask[MLX4_MAX_PORTS + 1];
};
struct mlx4_mfunc_master_ctx {
struct mlx4_slave_state *slave_state;
+ int init_port_ref[MLX4_MAX_PORTS + 1];
};
struct mlx4_vhcr {
@@ -505,6 +509,18 @@ void mlx4_init_mac_table(struct mlx4_dev *dev, struct mlx4_mac_table *table);
void mlx4_init_vlan_table(struct mlx4_dev *dev, struct mlx4_vlan_table *table);
int mlx4_SET_PORT(struct mlx4_dev *dev, u8 port);
+int mlx4_SET_PORT_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox);
+int mlx4_INIT_PORT_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox);
+int mlx4_CLOSE_PORT_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox);
+int mlx4_QUERY_PORT_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox);
int mlx4_get_port_ib_caps(struct mlx4_dev *dev, u8 port, __be32 *caps);
int mlx4_MCAST_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
diff --git a/drivers/net/mlx4/port.c b/drivers/net/mlx4/port.c
index 606aa58..67f0751 100644
--- a/drivers/net/mlx4/port.c
+++ b/drivers/net/mlx4/port.c
@@ -294,6 +294,82 @@ int mlx4_get_port_ib_caps(struct mlx4_dev *dev, u8 port, __be32 *caps)
return err;
}
+int mlx4_SET_PORT_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox)
+{
+ struct mlx4_priv *priv = mlx4_priv(dev);
+ int reset_qkey_viols;
+ int port;
+ int is_eth;
+ int err;
+ int i;
+ __be32 agg_cap_mask;
+ __be32 slave_cap_mask;
+ __be32 new_cap_mask;
+
+ port = vhcr->in_modifier & 0xff;
+ is_eth = vhcr->op_modifier;
+
+ /* For Ethernet, we currently support only slave0.
+ * TODO: add multi-vf support */
+ if (is_eth) {
+ if (slave)
+ return -EINVAL;
+ return mlx4_cmd(dev, inbox->dma, vhcr->in_modifier,
+ vhcr->op_modifier,
+ MLX4_CMD_SET_PORT,
+ MLX4_CMD_TIME_CLASS_B);
+ }
+
+ /* For IB, we only consider:
+ * - The capability mask, which is set to the aggregate of all slave frunction
+ * capabilities
+ * - The QKey violatin counter - reset according to each request.
+ */
+
+ if (dev->flags & MLX4_FLAG_OLD_PORT_CMDS) {
+ reset_qkey_viols = (*(u8 *) inbox->buf) & 0x40;
+ new_cap_mask = ((__be32 *) inbox->buf)[2];
+ } else {
+ reset_qkey_viols = ((u8 *) inbox->buf)[3] & 0x1;
+ new_cap_mask = ((__be32 *) inbox->buf)[1];
+ }
+
+ /* CX1: only slave0 has access to qp0 */
+ if (slave && (new_cap_mask & cpu_to_be32(IB_PORT_SM))) {
+ mlx4_warn(dev, "denying sm port capability for slave:%d\n", slave);
+ return -EINVAL;
+ }
+
+ agg_cap_mask = 0;
+ slave_cap_mask = priv->mfunc.master.slave_state[slave].ib_cap_mask[port];
+ priv->mfunc.master.slave_state[slave].ib_cap_mask[port] = new_cap_mask;
+ for (i = 0; i < dev->num_slaves; i++)
+ agg_cap_mask |= priv->mfunc.master.slave_state[slave].ib_cap_mask[port];
+
+#if 0
+ mlx4_warn(dev, "old_slave_cap:0x%x slave_cap:0x%x cap:0x%x qkey_reset:%d\n",
+ slave_cap_mask, priv->mfunc.master.slave_state[slave].ib_cap_mask[port],
+ agg_cap_mask, reset_qkey_viols);
+#endif
+
+ memset(inbox->buf, 0, 256);
+ if (dev->flags & MLX4_FLAG_OLD_PORT_CMDS) {
+ *(u8 *) inbox->buf = !!reset_qkey_viols << 6;
+ ((__be32 *) inbox->buf)[2] = agg_cap_mask;
+ } else {
+ ((u8 *) inbox->buf)[3] = !!reset_qkey_viols;
+ ((__be32 *) inbox->buf)[1] = agg_cap_mask;
+ }
+
+ err = mlx4_cmd(dev, inbox->dma, port, is_eth, MLX4_CMD_SET_PORT,
+ MLX4_CMD_TIME_CLASS_B);
+ if (err)
+ priv->mfunc.master.slave_state[slave].ib_cap_mask[port] = slave_cap_mask;
+ return err;
+}
+
int mlx4_SET_PORT(struct mlx4_dev *dev, u8 port)
{
struct mlx4_cmd_mailbox *mailbox;
--
1.6.1.3
--
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
^ permalink raw reply related
* [PATCH 08/25] mlx4_core: fix buggy parsing of reserved eq cap
From: Yevgeny Petrilin @ 2009-11-04 15:30 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
Signed-off-by: Liran Liss <liranl-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/net/mlx4/fw.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/net/mlx4/fw.c b/drivers/net/mlx4/fw.c
index a63f84a..87eb2d4 100644
--- a/drivers/net/mlx4/fw.c
+++ b/drivers/net/mlx4/fw.c
@@ -243,7 +243,7 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_MPT_OFFSET);
dev_cap->max_mpts = 1 << (field & 0x3f);
MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_EQ_OFFSET);
- dev_cap->reserved_eqs = 1 << (field & 0xf);
+ dev_cap->reserved_eqs = field & 0xf;
MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_EQ_OFFSET);
dev_cap->max_eqs = 1 << (field & 0xf);
MLX4_GET(field, outbox, QUERY_DEV_CAP_RSVD_MTT_OFFSET);
--
1.6.1.3
--
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
^ permalink raw reply related
* [PATCH 10/25] mlx4_core: track slave special qps
From: Yevgeny Petrilin @ 2009-11-04 15:31 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
A single function can own the actual Infiniband hw special qps.
This function demuxes inbound mads and protects outbound mads from spoofing.
To achieve this, it needs to know the special qps used by other functions.
This patch tracks special qp registration and notifies the owning function
on any changes.
Signed-off-by: Liran Liss <liranl-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/net/mlx4/cmd.c | 2 +
drivers/net/mlx4/eq.c | 15 +++++++++
drivers/net/mlx4/mlx4.h | 11 +++++++
drivers/net/mlx4/qp.c | 67 +++++++++++++++++++++++++++++++++++++++++++
include/linux/mlx4/cmd.h | 1 +
include/linux/mlx4/device.h | 3 ++
6 files changed, 99 insertions(+), 0 deletions(-)
diff --git a/drivers/net/mlx4/cmd.c b/drivers/net/mlx4/cmd.c
index 0a7c9c0..29074a0 100644
--- a/drivers/net/mlx4/cmd.c
+++ b/drivers/net/mlx4/cmd.c
@@ -525,6 +525,7 @@ static struct mlx4_cmd_info {
} cmd_info[] = {
{MLX4_CMD_QUERY_FW, 0, 1, 0, NULL, NULL},
{MLX4_CMD_QUERY_ADAPTER, 0, 1, 0, NULL, NULL},
+ {MLX4_CMD_GET_SLAVE_SQP, 0, 1, 0, NULL, mlx4_GET_SLAVE_SQP_wrapper},
{MLX4_CMD_INIT_PORT, 0, 0, 0, NULL, mlx4_INIT_PORT_wrapper},
{MLX4_CMD_CLOSE_PORT, 0, 0, 0, NULL, mlx4_CLOSE_PORT_wrapper},
@@ -568,6 +569,7 @@ static struct mlx4_cmd_info {
{MLX4_CMD_INIT2INIT_QP, 1, 0, 0, NULL, NULL}, /* need verifier */
{MLX4_CMD_SUSPEND_QP, 0, 0, 0, NULL, NULL}, /* need verifier */
{MLX4_CMD_UNSUSPEND_QP, 0, 0, 0, NULL, NULL}, /* need verifier */
+ {MLX4_CMD_CONF_SPECIAL_QP, 0, 0, 0, NULL, mlx4_CONF_SPECIAL_QP_wrapper},
{MLX4_CMD_MAD_IFC, 1, 1, 0, NULL, NULL}, /* need verifier */
/* Native multicast commands are not available for guests */
diff --git a/drivers/net/mlx4/eq.c b/drivers/net/mlx4/eq.c
index 1e8b62d..a27e1c4 100644
--- a/drivers/net/mlx4/eq.c
+++ b/drivers/net/mlx4/eq.c
@@ -358,6 +358,17 @@ static int mlx4_eq_int(struct mlx4_dev *dev, struct mlx4_eq *eq)
return eqes_found;
}
+static void mlx4_update_sqp(struct mlx4_dev *dev)
+{
+ if (!dev->caps.sqp_demux) {
+ mlx4_warn(dev, "unexpected update_sqp event\n");
+ return;
+ }
+ if (mlx4_GET_SLAVE_SQP(dev, mlx4_priv(dev)->mfunc.demux_sqp,
+ dev->caps.sqp_demux))
+ mlx4_warn(dev, "couldn't update sqp\n");
+}
+
void mlx4_slave_async_eq_poll(struct work_struct *work)
{
struct delayed_work *delay = container_of(work, struct delayed_work, work);
@@ -402,6 +413,10 @@ void mlx4_slave_async_eq_poll(struct work_struct *work)
mlx4_warn(dev, "slave async EQ overrun\n");
break;
+ case MLX4_EVENT_TYPE_SQP_UPDATE:
+ mlx4_update_sqp(dev);
+ break;
+
default:
mlx4_warn(dev, "Unhandled event:%02x\n", eqe.type);
}
diff --git a/drivers/net/mlx4/mlx4.h b/drivers/net/mlx4/mlx4.h
index fb1c6f0..b01820f 100644
--- a/drivers/net/mlx4/mlx4.h
+++ b/drivers/net/mlx4/mlx4.h
@@ -111,6 +111,7 @@ enum mlx4_alloc_mode {
};
enum {
+ MLX4_MFUNC_MAX = 64,
MLX4_MFUNC_MAX_EQES = 8,
MLX4_MFUNC_EQE_MASK = (MLX4_MFUNC_MAX_EQES - 1)
};
@@ -219,6 +220,7 @@ struct mlx4_slave_state {
struct mlx4_slave_eqe eq[MLX4_MFUNC_MAX_EQES];
u16 eq_pi;
u16 eq_ci;
+ int sqp_start;
spinlock_t lock;
};
@@ -246,6 +248,7 @@ struct mlx4_mfunc {
dma_addr_t vhcr_dma;
struct mlx4_mfunc_master_ctx master;
+ u32 demux_sqp[MLX4_MFUNC_MAX];
};
struct mlx4_cmd {
@@ -531,6 +534,14 @@ int mlx4_QUERY_PORT_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *v
struct mlx4_cmd_mailbox *outbox);
int mlx4_get_port_ib_caps(struct mlx4_dev *dev, u8 port, __be32 *caps);
+int mlx4_CONF_SPECIAL_QP_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox);
+int mlx4_GET_SLAVE_SQP(struct mlx4_dev *dev, u32 *sqp, int num);
+int mlx4_GET_SLAVE_SQP_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox);
+
int mlx4_MCAST_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
struct mlx4_cmd_mailbox *inbox,
struct mlx4_cmd_mailbox *outbox);
diff --git a/drivers/net/mlx4/qp.c b/drivers/net/mlx4/qp.c
index 065c7fc..99b9ded 100644
--- a/drivers/net/mlx4/qp.c
+++ b/drivers/net/mlx4/qp.c
@@ -145,6 +145,54 @@ int mlx4_qp_modify(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
}
EXPORT_SYMBOL_GPL(mlx4_qp_modify);
+u32 mlx4_get_slave_sqp(struct mlx4_dev *dev, int slave)
+{
+ if (mlx4_is_master(dev) && slave < dev->num_slaves) {
+ return mlx4_priv(dev)->mfunc.master.slave_state[slave].sqp_start;
+ }
+ if (mlx4_is_slave(dev) && slave < dev->caps.sqp_demux) {
+ return mlx4_priv(dev)->mfunc.demux_sqp[slave];
+ }
+ return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_slave_sqp);
+
+int mlx4_GET_SLAVE_SQP_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox)
+{
+ u32 *slave_sqp = outbox->buf;
+ int i;
+
+ /* CX1: special qp demultiplexing is done by slave0 */
+ if (slave) {
+ mlx4_warn(dev, "Denying slave_sqp request from slave:%d\n", slave);
+ return -EINVAL;
+ }
+ for (i = 0; i < 64; i++)
+ slave_sqp[i] = mlx4_get_slave_sqp(dev, i);
+ return 0;
+}
+
+int mlx4_GET_SLAVE_SQP(struct mlx4_dev *dev, u32 *sqp, int num)
+{
+ struct mlx4_cmd_mailbox *mailbox;
+ int err;
+
+ mailbox = mlx4_alloc_cmd_mailbox(dev);
+ if (IS_ERR(mailbox))
+ return PTR_ERR(mailbox);
+
+ err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 0, MLX4_CMD_GET_SLAVE_SQP,
+ MLX4_CMD_TIME_CLASS_A);
+ if (!err)
+ memcpy(sqp, mailbox->buf, sizeof (u32) * num);
+
+ mlx4_free_cmd_mailbox(dev, mailbox);
+ return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_GET_SLAVE_SQP);
+
int mlx4_qp_reserve_range(struct mlx4_dev *dev, int cnt, int align, int *base)
{
struct mlx4_priv *priv = mlx4_priv(dev);
@@ -331,6 +379,25 @@ static int mlx4_CONF_SPECIAL_QP(struct mlx4_dev *dev, u32 base_qpn)
MLX4_CMD_TIME_CLASS_B);
}
+int mlx4_CONF_SPECIAL_QP_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox)
+{
+ struct mlx4_priv *priv = mlx4_priv(dev);
+ int ret;
+
+ priv->mfunc.master.slave_state[slave].sqp_start = vhcr->in_modifier & 0xffffff;
+ if (!slave) {
+ /* CX1: slave0 owns real special QPs */
+ ret = mlx4_CONF_SPECIAL_QP(dev, priv->mfunc.master.slave_state[slave].sqp_start);
+ if (ret)
+ return ret;
+ }
+ /* Notify slave0 that an SQP change occured */
+ mlx4_slave_event(dev, 0, MLX4_EVENT_TYPE_SQP_UPDATE, 0, 0);
+ return 0;
+}
+
int mlx4_init_qp_table(struct mlx4_dev *dev)
{
struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table;
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index 8265843..f1659f2 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -124,6 +124,7 @@ enum {
MLX4_CMD_FREE_RES = 0x51,
MLX4_CMD_GET_EVENT = 0x52,
MLX4_CMD_MCAST_ATTACH = 0x54,
+ MLX4_CMD_GET_SLAVE_SQP = 0x55,
/* debug commands */
MLX4_CMD_QUERY_DEBUG_MSG = 0x2a,
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 9735f40..b20c8d8 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -99,6 +99,7 @@ enum mlx4_event {
MLX4_EVENT_TYPE_EQ_OVERFLOW = 0x0f,
MLX4_EVENT_TYPE_ECC_DETECT = 0x0e,
MLX4_EVENT_TYPE_CMD = 0x0a,
+ MLX4_EVENT_TYPE_SQP_UPDATE = 0xfe,
MLX4_EVENT_TYPE_NONE = 0xff,
};
@@ -242,6 +243,7 @@ struct mlx4_caps {
int log_num_prios;
enum mlx4_port_type port_type[MLX4_MAX_PORTS + 1];
u8 supported_type[MLX4_MAX_PORTS + 1];
+ u8 sqp_demux;
u32 port_mask;
enum mlx4_port_type possible_type[MLX4_MAX_PORTS + 1];
};
@@ -470,6 +472,7 @@ void mlx4_qp_release_range(struct mlx4_dev *dev, int base_qpn, int cnt);
int mlx4_qp_alloc(struct mlx4_dev *dev, int qpn, struct mlx4_qp *qp);
void mlx4_qp_free(struct mlx4_dev *dev, struct mlx4_qp *qp);
+u32 mlx4_get_slave_sqp(struct mlx4_dev *dev, int vf);
int mlx4_srq_alloc(struct mlx4_dev *dev, u32 pdn, struct mlx4_mtt *mtt,
u64 db_rec, struct mlx4_srq *srq);
--
1.6.1.3
--
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
^ permalink raw reply related
* [PATCH 11/25] mlx4_core: per-function capabilities
From: Yevgeny Petrilin @ 2009-11-04 15:31 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
The master function builds a HW profile, and manages all resources.
Other functions query the master for function-specific capabilities.
EQs, MSI-X vectors, and UARs are statically divided among all functions,
while other resources are dynamically assigned later upon request.
Signed-off-by: Liran Liss <liranl-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/net/mlx4/cmd.c | 2 +
drivers/net/mlx4/fw.c | 78 +++++++++++++++++++++++++++++++++++++++++-
drivers/net/mlx4/fw.h | 4 ++
drivers/net/mlx4/main.c | 70 ++++++++++++++++++++++++++++++++++++++-
drivers/net/mlx4/mlx4.h | 4 ++-
drivers/net/mlx4/profile.c | 26 ++++++++++++--
include/linux/mlx4/cmd.h | 1 +
include/linux/mlx4/device.h | 3 ++
8 files changed, 180 insertions(+), 8 deletions(-)
diff --git a/drivers/net/mlx4/cmd.c b/drivers/net/mlx4/cmd.c
index 29074a0..7afbfef 100644
--- a/drivers/net/mlx4/cmd.c
+++ b/drivers/net/mlx4/cmd.c
@@ -39,6 +39,7 @@
#include <asm/io.h>
#include "mlx4.h"
+#include "fw.h"
#include "en_port.h"
#define CMD_POLL_TOKEN 0xffff
@@ -524,6 +525,7 @@ static struct mlx4_cmd_info {
struct mlx4_cmd_mailbox *outbox);
} cmd_info[] = {
{MLX4_CMD_QUERY_FW, 0, 1, 0, NULL, NULL},
+ {MLX4_CMD_QUERY_SLAVE_CAP, 0, 1, 0, NULL, mlx4_QUERY_SLAVE_CAP_wrapper},
{MLX4_CMD_QUERY_ADAPTER, 0, 1, 0, NULL, NULL},
{MLX4_CMD_GET_SLAVE_SQP, 0, 1, 0, NULL, mlx4_GET_SLAVE_SQP_wrapper},
diff --git a/drivers/net/mlx4/fw.c b/drivers/net/mlx4/fw.c
index 87eb2d4..60889d3 100644
--- a/drivers/net/mlx4/fw.c
+++ b/drivers/net/mlx4/fw.c
@@ -143,6 +143,76 @@ int mlx4_QUERY_PORT_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *v
MLX4_CMD_TIME_CLASS_B);
}
+int mlx4_QUERY_SLAVE_CAP_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox)
+{
+ struct mlx4_caps *caps = outbox->buf;
+ int i;
+
+ memcpy(caps, &dev->caps, sizeof *caps);
+
+ /* CX1: blue-flame not supported in vfs */
+ caps->bf_reg_size = 0;
+ caps->bf_regs_per_page = 0;
+
+ /* CX1:
+ * On ingress, slave0 gets full ownership over qp1, and demultiplexes GSI mads
+ * on behalf of other slaves.
+ * On egress, slave>0 must tunnel their GSI mads for validation by the qp1 owner */
+ if (slave) {
+ caps->sqp_demux = 0;
+ for (i = 1; i <= dev->caps.num_ports; ++i) {
+ caps->gid_table_len[i] = 1;
+ caps->pkey_table_len[i] = 1;
+ }
+ } else {
+ caps->sqp_demux = dev->num_slaves;
+ for (i = 1; i <= dev->caps.num_ports; ++i) {
+ caps->gid_table_len[i] = dev->num_slaves;
+ caps->pkey_table_len[i] = 1;
+ }
+ }
+
+ /* Slave functions allocate themselves EQs, UARs, and PDs.
+ * - num is the maximum resource index
+ * - reserved is the minimum resource index */
+ caps->num_eqs = dev->caps.reserved_eqs + MLX4_MFUNC_EQ_NUM * (slave + 2);
+ caps->reserved_eqs = dev->caps.reserved_eqs + MLX4_MFUNC_EQ_NUM * (slave + 1);
+ caps->num_uars = dev->caps.num_uars * (slave + 1);
+ caps->reserved_uars = max_t(int, dev->caps.num_uars * slave, dev->caps.reserved_uars);
+
+ /* PDs have the same range in every guest; the distinction is in the msbs,
+ * which contains the guest ID (vf + 1) */
+ caps->pd_base = slave + 1;
+
+ /* All other resources are allocated by the master, but we still report
+ * 'num' and 'reserved' capabilities as follows:
+ * - num remains the maximum resource index
+ * - 'num - reserved' is the total available objects of a resource, but
+ * resource indices may be less than 'reserved'
+ * TODO: set per-resource quotas */
+ return 0;
+}
+
+int mlx4_QUERY_SLAVE_CAP(struct mlx4_dev *dev, struct mlx4_caps *caps)
+{
+ struct mlx4_cmd_mailbox *mailbox;
+ int err;
+
+ mailbox = mlx4_alloc_cmd_mailbox(dev);
+ if (IS_ERR(mailbox))
+ return PTR_ERR(mailbox);
+
+ err = mlx4_cmd_box(dev, 0, mailbox->dma, 0, 0, MLX4_CMD_QUERY_SLAVE_CAP,
+ MLX4_CMD_TIME_CLASS_A);
+ if (!err)
+ memcpy(caps, mailbox->buf, sizeof *caps);
+
+ mlx4_free_cmd_mailbox(dev, mailbox);
+ return err;
+}
+
int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
{
struct mlx4_cmd_mailbox *mailbox;
@@ -405,9 +475,13 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
* Each UAR has 4 EQ doorbells; so if a UAR is reserved, then
* we can't use any EQs whose doorbell falls on that page,
* even if the EQ itself isn't reserved.
+ * CX1: in multi-function mode, EQ doorbells are virtualized so the whole EQ
+ * range is available: the master and each of the slaves get 4 EQ doorbells
+ * starting from the first non-reserved EQ.
*/
- dev_cap->reserved_eqs = max(dev_cap->reserved_uars * 4,
- dev_cap->reserved_eqs);
+ if (!mlx4_is_mfunc(dev))
+ dev_cap->reserved_eqs = max(dev_cap->reserved_uars * 4,
+ dev_cap->reserved_eqs);
mlx4_dbg(dev, "Max ICM size %lld MB\n",
(unsigned long long) dev_cap->max_icm_sz >> 20);
diff --git a/drivers/net/mlx4/fw.h b/drivers/net/mlx4/fw.h
index 526d7f3..d066c69 100644
--- a/drivers/net/mlx4/fw.h
+++ b/drivers/net/mlx4/fw.h
@@ -160,6 +160,10 @@ struct mlx4_set_ib_param {
};
int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap);
+int mlx4_QUERY_SLAVE_CAP(struct mlx4_dev *dev, struct mlx4_caps *caps);
+int mlx4_QUERY_SLAVE_CAP_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
+ struct mlx4_cmd_mailbox *inbox,
+ struct mlx4_cmd_mailbox *outbox);
int mlx4_MAP_FA(struct mlx4_dev *dev, struct mlx4_icm *icm);
int mlx4_UNMAP_FA(struct mlx4_dev *dev);
int mlx4_RUN_FW(struct mlx4_dev *dev);
diff --git a/drivers/net/mlx4/main.c b/drivers/net/mlx4/main.c
index 291a505..4edc8d7 100644
--- a/drivers/net/mlx4/main.c
+++ b/drivers/net/mlx4/main.c
@@ -141,6 +141,7 @@ static void mlx4_set_port_mask(struct mlx4_dev *dev)
if (dev->caps.port_type[i] == MLX4_PORT_TYPE_IB)
dev->caps.port_mask |= 1 << (i - 1);
}
+
static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
{
int err;
@@ -185,6 +186,7 @@ static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
dev->caps.supported_type[i] = dev_cap->supported_port_types[i];
}
+ dev->caps.uar_page_size = PAGE_SIZE;
dev->caps.num_uars = dev_cap->uar_size / PAGE_SIZE;
dev->caps.local_ca_ack_delay = dev_cap->local_ca_ack_delay;
dev->caps.bf_reg_size = dev_cap->bf_reg_size;
@@ -211,7 +213,9 @@ static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
dev->caps.reserved_mtts = DIV_ROUND_UP(dev_cap->reserved_mtts,
dev->caps.mtts_per_seg);
dev->caps.reserved_mrws = dev_cap->reserved_mrws;
- dev->caps.reserved_uars = dev_cap->reserved_uars;
+
+ /* The first 128 UARs are used for EQ doorbells */
+ dev->caps.reserved_uars = max_t(int, 128, dev_cap->reserved_uars);
dev->caps.reserved_pds = dev_cap->reserved_pds;
dev->caps.mtt_entry_sz = dev->caps.mtts_per_seg * dev_cap->mtt_entry_sz;
dev->caps.max_msg_sz = dev_cap->max_msg_sz;
@@ -265,6 +269,70 @@ static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_ADDR] +
dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FC_EXCH];
+ /* CX1: vf0 takes role of GSI multiplexing instead of master */
+ dev->caps.sqp_demux = 0;
+ return 0;
+}
+
+static int mlx4_slave_cap(struct mlx4_dev *dev)
+{
+ int err;
+ u32 page_size;
+
+ err = mlx4_QUERY_SLAVE_CAP(dev, &dev->caps);
+ if (err)
+ return err;
+
+ page_size = ~dev->caps.page_size_cap + 1;
+ mlx4_warn(dev, "HCA minimum page size:%d\n", page_size);
+ if (page_size > PAGE_SIZE) {
+ mlx4_err(dev, "HCA minimum page size of %d bigger than "
+ "kernel PAGE_SIZE of %ld, aborting.\n",
+ page_size, PAGE_SIZE);
+ return -ENODEV;
+ }
+
+ /* TODO: relax this assumption */
+ if (dev->caps.uar_page_size != PAGE_SIZE) {
+ mlx4_err(dev, "UAR size:%d != kernel PAGE_SIZE of %ld\n",
+ dev->caps.uar_page_size, PAGE_SIZE);
+ return -ENODEV;
+ }
+
+ if (dev->caps.num_ports > MLX4_MAX_PORTS) {
+ mlx4_err(dev, "HCA has %d ports, but we only support %d, "
+ "aborting.\n", dev->caps.num_ports, MLX4_MAX_PORTS);
+ return -ENODEV;
+ }
+
+ if (dev->caps.uar_page_size * (dev->caps.num_uars -
+ dev->caps.reserved_uars) >
+ pci_resource_len(dev->pdev, 2)) {
+ mlx4_err(dev, "HCA reported UAR region size of 0x%x bigger than "
+ "PCI resource 2 size of 0x%llx, aborting.\n",
+ dev->caps.uar_page_size * dev->caps.num_uars,
+ (unsigned long long) pci_resource_len(dev->pdev, 2));
+ return -ENODEV;
+ }
+
+ /* Adjust eq number */
+ if (dev->caps.num_eqs - dev->caps.reserved_eqs > num_possible_cpus() + 1)
+ dev->caps.num_eqs = dev->caps.reserved_eqs + num_possible_cpus() + 1;
+
+#if 0
+ mlx4_warn(dev, "sqp_demux:%d\n", dev->caps.sqp_demux);
+ mlx4_warn(dev, "num_uars:%d reserved_uars:%d uar region:0x%x bar2:0x%llx\n",
+ dev->caps.num_uars, dev->caps.reserved_uars,
+ dev->caps.uar_page_size * dev->caps.num_uars,
+ pci_resource_len(dev->pdev, 2));
+ mlx4_warn(dev, "num_eqs:%d reserved_eqs:%d\n", dev->caps.num_eqs,
+ dev->caps.reserved_eqs);
+ mlx4_warn(dev, "num_pds:%d reserved_pds:%d slave_pd_shift:%d pd_base:%d\n",
+ dev->caps.num_pds,
+ dev->caps.reserved_pds,
+ dev->caps.slave_pd_shift,
+ dev->caps.pd_base);
+#endif
return 0;
}
diff --git a/drivers/net/mlx4/mlx4.h b/drivers/net/mlx4/mlx4.h
index b01820f..b8b30ca 100644
--- a/drivers/net/mlx4/mlx4.h
+++ b/drivers/net/mlx4/mlx4.h
@@ -69,7 +69,8 @@ enum {
};
enum {
- MLX4_NUM_PDS = 1 << 15
+ MLX4_NUM_PDS = 1 << 15,
+ MLX4_SLAVE_PD_SHIFT = 17, /* the 7 msbs encode the slave id */
};
enum {
@@ -112,6 +113,7 @@ enum mlx4_alloc_mode {
enum {
MLX4_MFUNC_MAX = 64,
+ MLX4_MFUNC_EQ_NUM = 4,
MLX4_MFUNC_MAX_EQES = 8,
MLX4_MFUNC_EQE_MASK = (MLX4_MFUNC_MAX_EQES - 1)
};
diff --git a/drivers/net/mlx4/profile.c b/drivers/net/mlx4/profile.c
index ca25b9d..f6197e9 100644
--- a/drivers/net/mlx4/profile.c
+++ b/drivers/net/mlx4/profile.c
@@ -105,9 +105,19 @@ u64 mlx4_make_profile(struct mlx4_dev *dev,
profile[MLX4_RES_AUXC].num = request->num_qp;
profile[MLX4_RES_SRQ].num = request->num_srq;
profile[MLX4_RES_CQ].num = request->num_cq;
- profile[MLX4_RES_EQ].num = min_t(unsigned, dev_cap->max_eqs,
- dev_cap->reserved_eqs +
- num_possible_cpus() + 1);
+ if (mlx4_is_master(dev)) {
+ profile[MLX4_RES_EQ].num = dev_cap->reserved_eqs +
+ MLX4_MFUNC_EQ_NUM *
+ (dev->num_slaves + 1);
+ if (profile[MLX4_RES_EQ].num > dev_cap->max_eqs) {
+ mlx4_warn(dev, "Not enough eqs for:%ld slave functions\n", dev->num_slaves);
+ kfree(profile);
+ return -ENOMEM;
+ }
+ } else
+ profile[MLX4_RES_EQ].num = min_t(unsigned, dev_cap->max_eqs,
+ dev_cap->reserved_eqs +
+ num_possible_cpus() + 1);
profile[MLX4_RES_DMPT].num = request->num_mpt;
profile[MLX4_RES_CMPT].num = MLX4_NUM_CMPTS;
profile[MLX4_RES_MTT].num = request->num_mtt;
@@ -196,7 +206,13 @@ u64 mlx4_make_profile(struct mlx4_dev *dev,
init_hca->log_num_cqs = profile[i].log_num;
break;
case MLX4_RES_EQ:
- dev->caps.num_eqs = profile[i].num;
+ if (mlx4_is_master(dev)) {
+ dev->caps.num_eqs = dev_cap->reserved_eqs +
+ min_t(unsigned,
+ MLX4_MFUNC_EQ_NUM,
+ num_possible_cpus() + 1);
+ } else
+ dev->caps.num_eqs = profile[i].num;
init_hca->eqc_base = profile[i].start;
init_hca->log_num_eqs = profile[i].log_num;
break;
@@ -232,6 +248,8 @@ u64 mlx4_make_profile(struct mlx4_dev *dev,
* of the HCA profile anyway.
*/
dev->caps.num_pds = MLX4_NUM_PDS;
+ dev->caps.slave_pd_shift = MLX4_SLAVE_PD_SHIFT;
+ dev->caps.pd_base = 0;
kfree(profile);
return total_size;
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index f1659f2..f265091 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -123,6 +123,7 @@ enum {
MLX4_CMD_ALLOC_RES = 0x50,
MLX4_CMD_FREE_RES = 0x51,
MLX4_CMD_GET_EVENT = 0x52,
+ MLX4_CMD_QUERY_SLAVE_CAP = 0x53,
MLX4_CMD_MCAST_ATTACH = 0x54,
MLX4_CMD_GET_SLAVE_SQP = 0x55,
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index b20c8d8..4b9091a 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -192,6 +192,7 @@ struct mlx4_caps {
int pkey_table_len[MLX4_MAX_PORTS + 1];
int local_ca_ack_delay;
int num_uars;
+ int uar_page_size;
int bf_reg_size;
int bf_regs_per_page;
int max_sq_sg;
@@ -226,6 +227,8 @@ struct mlx4_caps {
int num_qp_per_mgm;
int num_pds;
int reserved_pds;
+ int slave_pd_shift;
+ int pd_base;
int mtt_entry_sz;
u32 max_msg_sz;
u32 page_size_cap;
--
1.6.1.3
--
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
^ permalink raw reply related
* [PATCH 12/25] mlx4_core: associate resources with specific functions
From: Yevgeny Petrilin @ 2009-11-04 15:31 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
Resources are associated with functions by means of pds - the upper 7 msbs
specify the function id. Since cqs do not refer to a specific pd, the function
id is passed explicitly to hw in the cq context.
For uars, which are statically partitioned between slaves, we only need to adjust
their offsets with respect to the associated function bar.
Signed-off-by: Liran Liss <liranl-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/net/mlx4/cq.c | 4 +++-
drivers/net/mlx4/eq.c | 13 ++++++++++++-
drivers/net/mlx4/pd.c | 12 +++++++++---
3 files changed, 24 insertions(+), 5 deletions(-)
diff --git a/drivers/net/mlx4/cq.c b/drivers/net/mlx4/cq.c
index 3fb9f7f..9f9f246 100644
--- a/drivers/net/mlx4/cq.c
+++ b/drivers/net/mlx4/cq.c
@@ -59,7 +59,8 @@ struct mlx4_cq_context {
__be32 solicit_producer_index;
__be32 consumer_index;
__be32 producer_index;
- u32 reserved4[2];
+ __be32 pd; /* for sriov guest id */
+ u32 reserved4;
__be64 db_rec_addr;
};
@@ -293,6 +294,7 @@ int mlx4_cq_alloc(struct mlx4_dev *dev, int nent, struct mlx4_mtt *mtt,
mtt_addr = mlx4_mtt_addr(dev, mtt);
cq_context->mtt_base_addr_h = mtt_addr >> 32;
cq_context->mtt_base_addr_l = cpu_to_be32(mtt_addr & 0xffffffff);
+ cq_context->pd = cpu_to_be32(dev->caps.pd_base << dev->caps.slave_pd_shift);
cq_context->db_rec_addr = cpu_to_be64(db_rec);
err = mlx4_SW2HW_CQ(dev, mailbox, cq->cqn);
diff --git a/drivers/net/mlx4/eq.c b/drivers/net/mlx4/eq.c
index a27e1c4..46064cc 100644
--- a/drivers/net/mlx4/eq.c
+++ b/drivers/net/mlx4/eq.c
@@ -486,9 +486,20 @@ static void __iomem *mlx4_get_eq_uar(struct mlx4_dev *dev, struct mlx4_eq *eq)
{
struct mlx4_priv *priv = mlx4_priv(dev);
int index;
+ int offset;
- index = eq->eqn / 4 - dev->caps.reserved_eqs / 4;
+ /* CX1: slave EQ DBs are located in the comm channel page */
+ if (mlx4_is_slave(dev) || mlx4_is_master(dev)) {
+ if (eq->eqn - dev->caps.reserved_eqs >= MLX4_MFUNC_EQ_NUM) {
+ mlx4_err(dev, "eqn:%d doorbell out of range (reserved:%d)\n",
+ eq->eqn, dev->caps.reserved_eqs);
+ return NULL;
+ }
+ offset = 0x800 + (eq->eqn - dev->caps.reserved_eqs) * 8;
+ return ((void *) priv->mfunc.comm) + offset;
+ }
+ index = eq->eqn / 4 - dev->caps.reserved_eqs / 4;
if (!priv->eq_table.uar_map[index]) {
priv->eq_table.uar_map[index] =
ioremap(pci_resource_start(dev->pdev, 2) +
diff --git a/drivers/net/mlx4/pd.c b/drivers/net/mlx4/pd.c
index c4988d6..08eaf08 100644
--- a/drivers/net/mlx4/pd.c
+++ b/drivers/net/mlx4/pd.c
@@ -45,7 +45,7 @@ int mlx4_pd_alloc(struct mlx4_dev *dev, u32 *pdn)
*pdn = mlx4_bitmap_alloc(&priv->pd_bitmap);
if (*pdn == -1)
return -ENOMEM;
-
+ *pdn |= dev->caps.pd_base << dev->caps.slave_pd_shift;
return 0;
}
EXPORT_SYMBOL_GPL(mlx4_pd_alloc);
@@ -72,12 +72,18 @@ void mlx4_cleanup_pd_table(struct mlx4_dev *dev)
int mlx4_uar_alloc(struct mlx4_dev *dev, struct mlx4_uar *uar)
{
+ int offset;
+
uar->index = mlx4_bitmap_alloc(&mlx4_priv(dev)->uar_table.bitmap);
if (uar->index == -1)
return -ENOMEM;
- uar->pfn = (pci_resource_start(dev->pdev, 2) >> PAGE_SHIFT) + uar->index;
-
+ if (mlx4_is_slave(dev))
+ offset = uar->index % ((int) pci_resource_len(dev->pdev, 2) /
+ dev->caps.uar_page_size);
+ else
+ offset = uar->index;
+ uar->pfn = (pci_resource_start(dev->pdev, 2) >> PAGE_SHIFT) + offset;
return 0;
}
EXPORT_SYMBOL_GPL(mlx4_uar_alloc);
--
1.6.1.3
--
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
^ permalink raw reply related
* [PATCH 13/25] mlx4_core: remove unused driver/kernel uar (kar)
From: Yevgeny Petrilin @ 2009-11-04 15:31 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
Signed-off-by: Liran Liss <liranl-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/net/mlx4/main.c | 25 +------------------------
drivers/net/mlx4/mlx4.h | 2 --
2 files changed, 1 insertions(+), 26 deletions(-)
diff --git a/drivers/net/mlx4/main.c b/drivers/net/mlx4/main.c
index 4edc8d7..d8143c6 100644
--- a/drivers/net/mlx4/main.c
+++ b/drivers/net/mlx4/main.c
@@ -883,26 +883,11 @@ static int mlx4_setup_hca(struct mlx4_dev *dev)
return err;
}
- err = mlx4_uar_alloc(dev, &priv->driver_uar);
- if (err) {
- mlx4_err(dev, "Failed to allocate driver access region, "
- "aborting.\n");
- goto err_uar_table_free;
- }
-
- priv->kar = ioremap(priv->driver_uar.pfn << PAGE_SHIFT, PAGE_SIZE);
- if (!priv->kar) {
- mlx4_err(dev, "Couldn't map kernel access region, "
- "aborting.\n");
- err = -ENOMEM;
- goto err_uar_free;
- }
-
err = mlx4_init_pd_table(dev);
if (err) {
mlx4_err(dev, "Failed to initialize "
"protection domain table, aborting.\n");
- goto err_kar_unmap;
+ goto err_uar_table_free;
}
err = mlx4_init_mr_table(dev);
@@ -1015,12 +1000,6 @@ err_mr_table_free:
err_pd_table_free:
mlx4_cleanup_pd_table(dev);
-err_kar_unmap:
- iounmap(priv->kar);
-
-err_uar_free:
- mlx4_uar_free(dev, &priv->driver_uar);
-
err_uar_table_free:
mlx4_cleanup_uar_table(dev);
return err;
@@ -1316,8 +1295,6 @@ static void mlx4_remove_one(struct pci_dev *pdev)
mlx4_cleanup_mr_table(dev);
mlx4_cleanup_pd_table(dev);
- iounmap(priv->kar);
- mlx4_uar_free(dev, &priv->driver_uar);
mlx4_cleanup_uar_table(dev);
mlx4_free_eq_table(dev);
mlx4_close_hca(dev);
diff --git a/drivers/net/mlx4/mlx4.h b/drivers/net/mlx4/mlx4.h
index b8b30ca..516e375 100644
--- a/drivers/net/mlx4/mlx4.h
+++ b/drivers/net/mlx4/mlx4.h
@@ -401,8 +401,6 @@ struct mlx4_priv {
void __iomem *clr_base;
- struct mlx4_uar driver_uar;
- void __iomem *kar;
struct mlx4_port_info port[MLX4_MAX_PORTS + 1];
struct mlx4_sense sense;
struct mutex port_mutex;
--
1.6.1.3
--
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
^ permalink raw reply related
* [PATCH 15/25] mlx4_core: multi-function interface setup
From: Yevgeny Petrilin @ 2009-11-04 15:31 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
Master verifies hw healthiness.
Interface drivers are only supported on slaves.
Signed-off-by: Liran Liss <liranl-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/net/mlx4/intf.c | 21 ++++++++++++++-------
1 files changed, 14 insertions(+), 7 deletions(-)
diff --git a/drivers/net/mlx4/intf.c b/drivers/net/mlx4/intf.c
index 0e7eb10..f5f4523 100644
--- a/drivers/net/mlx4/intf.c
+++ b/drivers/net/mlx4/intf.c
@@ -88,8 +88,11 @@ int mlx4_register_interface(struct mlx4_interface *intf)
mutex_lock(&intf_mutex);
list_add_tail(&intf->list, &intf_list);
- list_for_each_entry(priv, &dev_list, dev_list)
- mlx4_add_device(intf, priv);
+ list_for_each_entry(priv, &dev_list, dev_list) {
+ /* CX1: master cannot run interfaces */
+ if (!mlx4_is_master(&priv->dev))
+ mlx4_add_device(intf, priv);
+ }
mutex_unlock(&intf_mutex);
@@ -135,11 +138,14 @@ int mlx4_register_device(struct mlx4_dev *dev)
mutex_lock(&intf_mutex);
list_add_tail(&priv->dev_list, &dev_list);
- list_for_each_entry(intf, &intf_list, list)
- mlx4_add_device(intf, priv);
-
+ /* CX1: master cannot run interfaces */
+ if (!mlx4_is_master(dev)) {
+ list_for_each_entry(intf, &intf_list, list)
+ mlx4_add_device(intf, priv);
+ }
mutex_unlock(&intf_mutex);
- mlx4_start_catas_poll(dev);
+ if (!mlx4_is_slave(dev))
+ mlx4_start_catas_poll(dev);
return 0;
}
@@ -149,7 +155,8 @@ void mlx4_unregister_device(struct mlx4_dev *dev)
struct mlx4_priv *priv = mlx4_priv(dev);
struct mlx4_interface *intf;
- mlx4_stop_catas_poll(dev);
+ if (!mlx4_is_slave(dev))
+ mlx4_stop_catas_poll(dev);
mutex_lock(&intf_mutex);
list_for_each_entry(intf, &intf_list, list)
--
1.6.1.3
--
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
^ permalink raw reply related
* [PATCH 16/25] mlx4_core: boot sriov
From: Yevgeny Petrilin @ 2009-11-04 15:31 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
Add virtual function device ids.
Establish comm channel when sriov is enabled, and boot vfs through it.
Each slave gets one msi-X for completions, the master also gets one for
async events.
Signed-off-by: Liran Liss <liranl-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
Signed-off-by: Yevgeny Petrilin <yevgenyp-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/net/mlx4/eq.c | 21 ++--
drivers/net/mlx4/main.c | 315 +++++++++++++++++++++++++++++++++++++----------
2 files changed, 260 insertions(+), 76 deletions(-)
diff --git a/drivers/net/mlx4/eq.c b/drivers/net/mlx4/eq.c
index b0c8022..29b2cd6 100644
--- a/drivers/net/mlx4/eq.c
+++ b/drivers/net/mlx4/eq.c
@@ -770,16 +770,18 @@ int mlx4_init_eq_table(struct mlx4_dev *dev)
goto err_out_unmap;
}
- err = mlx4_create_eq(dev, MLX4_NUM_ASYNC_EQE + MLX4_NUM_SPARE_EQE,
- (dev->flags & MLX4_FLAG_MSI_X) ? dev->caps.num_comp_vectors : 0,
- &priv->eq_table.eq[dev->caps.num_comp_vectors]);
- if (err)
- goto err_out_comp;
+ if (!mlx4_is_slave(dev)) {
+ err = mlx4_create_eq(dev, MLX4_NUM_ASYNC_EQE + MLX4_NUM_SPARE_EQE,
+ (dev->flags & MLX4_FLAG_MSI_X) ? dev->caps.num_comp_vectors : 0,
+ &priv->eq_table.eq[dev->caps.num_comp_vectors]);
+ if (err)
+ goto err_out_comp;
+ }
if (dev->flags & MLX4_FLAG_MSI_X) {
const char *eq_name;
- for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i) {
+ for (i = 0; i < dev->caps.num_comp_vectors + !(mlx4_is_slave(dev)); ++i) {
if (i < dev->caps.num_comp_vectors) {
snprintf(priv->eq_table.irq_names +
i * MLX4_IRQNAME_SIZE,
@@ -825,13 +827,14 @@ int mlx4_init_eq_table(struct mlx4_dev *dev)
priv->eq_table.eq[dev->caps.num_comp_vectors].eqn, err);
}
- for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i)
+ for (i = 0; i < dev->caps.num_comp_vectors + !(mlx4_is_slave(dev)); ++i)
eq_set_ci(&priv->eq_table.eq[i], 1);
return 0;
err_out_async:
- mlx4_free_eq(dev, &priv->eq_table.eq[dev->caps.num_comp_vectors]);
+ if (!mlx4_is_slave(dev))
+ mlx4_free_eq(dev, &priv->eq_table.eq[dev->caps.num_comp_vectors]);
err_out_comp:
i = dev->caps.num_comp_vectors;
@@ -866,7 +869,7 @@ void mlx4_cleanup_eq_table(struct mlx4_dev *dev)
mlx4_free_irqs(dev);
- for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i)
+ for (i = 0; i < dev->caps.num_comp_vectors + !mlx4_is_slave(dev); ++i)
mlx4_free_eq(dev, &priv->eq_table.eq[i]);
if (!mlx4_is_slave(dev))
diff --git a/drivers/net/mlx4/main.c b/drivers/net/mlx4/main.c
index 2852445..c2b6b35 100644
--- a/drivers/net/mlx4/main.c
+++ b/drivers/net/mlx4/main.c
@@ -73,6 +73,18 @@ MODULE_PARM_DESC(msi_x, "attempt to use MSI-X if nonzero");
#endif /* CONFIG_PCI_MSI */
+#ifdef CONFIG_PCI_IOV
+
+static int sr_iov;
+module_param(sr_iov, int, 0444);
+MODULE_PARM_DESC(sr_iov, "enable #sr_iov functions if sr_iov > 0");
+
+static int probe_vf;
+module_param(probe_vf, int, 0444);
+MODULE_PARM_DESC(probe_vf, "number of vfs to probe by pf driver (sr_iov > 0)");
+
+#endif /* CONFIG_PCI_IOV */
+
static char mlx4_version[] __devinitdata =
DRV_NAME ": Mellanox ConnectX core driver v"
DRV_VERSION " (" DRV_RELDATE ")\n";
@@ -779,12 +791,56 @@ static void mlx4_free_icms(struct mlx4_dev *dev)
mlx4_free_icm(dev, priv->fw.aux_icm, 0);
}
+static void mlx4_slave_exit(struct mlx4_dev *dev)
+{
+ struct mlx4_priv *priv = mlx4_priv(dev);
+
+ down(&priv->cmd.poll_sem);
+ if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0, MLX4_COMM_TIME))
+ mlx4_warn(dev, "Failed to close slave function.\n");
+ up(&priv->cmd.poll_sem);
+}
+
static void mlx4_close_hca(struct mlx4_dev *dev)
{
- mlx4_CLOSE_HCA(dev, 0);
- mlx4_free_icms(dev);
- mlx4_UNMAP_FA(dev);
- mlx4_free_icm(dev, mlx4_priv(dev)->fw.fw_icm, 0);
+ if (mlx4_is_slave(dev))
+ mlx4_slave_exit(dev);
+ else {
+ mlx4_CLOSE_HCA(dev, 0);
+ mlx4_free_icms(dev);
+ mlx4_UNMAP_FA(dev);
+ mlx4_free_icm(dev, mlx4_priv(dev)->fw.fw_icm, 0);
+ }
+}
+
+static int mlx4_init_slave(struct mlx4_dev *dev)
+{
+ struct mlx4_priv *priv = mlx4_priv(dev);
+ u64 dma = (u64) priv->mfunc.vhcr_dma;
+
+ down(&priv->cmd.poll_sem);
+ mlx4_warn(dev, "Sending reset\n");
+ if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0, MLX4_COMM_TIME))
+ goto err;
+ mlx4_warn(dev, "Sending vhcr0\n");
+ if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR0, dma >> 48,
+ MLX4_COMM_TIME))
+ goto err;
+ if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR1, dma >> 32,
+ MLX4_COMM_TIME))
+ goto err;
+ if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR2, dma >> 16,
+ MLX4_COMM_TIME))
+ goto err;
+ if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR_EN, dma, MLX4_COMM_TIME))
+ goto err;
+ up(&priv->cmd.poll_sem);
+ return 0;
+
+err:
+ mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0, 0);
+ up(&priv->cmd.poll_sem);
+ return -EIO;
}
static int mlx4_init_hca(struct mlx4_dev *dev)
@@ -798,51 +854,67 @@ static int mlx4_init_hca(struct mlx4_dev *dev)
u64 icm_size;
int err;
- err = mlx4_QUERY_FW(dev);
- if (err) {
- if (err == -EACCES)
- mlx4_info(dev, "non-primary physical function, skipping.\n");
- else
- mlx4_err(dev, "QUERY_FW command failed, aborting.\n");
- return err;
- }
+ if (!mlx4_is_slave(dev)) {
+ err = mlx4_QUERY_FW(dev); /* TODO: verify FW version in slaves as well */
+ if (err) {
+ if (err == -EACCES)
+ mlx4_info(dev, "non-primary physical function, skipping.\n");
+ else {
+ mlx4_err(dev, "QUERY_FW command failed, aborting.\n");
+ if (mlx4_is_master(dev))
+ mlx4_err(dev, "Are you using SRIOV-enabled firmware?\n");
+ }
+ return err;
+ }
- err = mlx4_load_fw(dev);
- if (err) {
- mlx4_err(dev, "Failed to start FW, aborting.\n");
- return err;
- }
+ err = mlx4_load_fw(dev);
+ if (err) {
+ mlx4_err(dev, "Failed to start FW, aborting.\n");
+ return err;
+ }
- mlx4_cfg.log_pg_sz_m = 1;
- mlx4_cfg.log_pg_sz = 0;
- err = mlx4_MOD_STAT_CFG(dev, &mlx4_cfg);
- if (err)
- mlx4_warn(dev, "Failed to override log_pg_sz parameter\n");
+ mlx4_cfg.log_pg_sz_m = 1;
+ mlx4_cfg.log_pg_sz = 0;
+ err = mlx4_MOD_STAT_CFG(dev, &mlx4_cfg);
+ if (err)
+ mlx4_warn(dev, "Failed to override log_pg_sz parameter\n");
- err = mlx4_dev_cap(dev, &dev_cap);
- if (err) {
- mlx4_err(dev, "QUERY_DEV_CAP command failed, aborting.\n");
- goto err_stop_fw;
- }
+ err = mlx4_dev_cap(dev, &dev_cap);
+ if (err) {
+ mlx4_err(dev, "QUERY_DEV_CAP command failed, aborting.\n");
+ goto err_stop_fw;
+ }
- profile = default_profile;
+ profile = default_profile;
+ icm_size = mlx4_make_profile(dev, &profile, &dev_cap, &init_hca);
+ if ((long long) icm_size < 0) {
+ err = icm_size;
+ goto err_stop_fw;
+ }
- icm_size = mlx4_make_profile(dev, &profile, &dev_cap, &init_hca);
- if ((long long) icm_size < 0) {
- err = icm_size;
- goto err_stop_fw;
- }
+ init_hca.log_uar_sz = ilog2(dev->caps.num_uars);
- init_hca.log_uar_sz = ilog2(dev->caps.num_uars);
+ err = mlx4_init_icm(dev, &dev_cap, &init_hca, icm_size);
+ if (err)
+ goto err_stop_fw;
- err = mlx4_init_icm(dev, &dev_cap, &init_hca, icm_size);
- if (err)
- goto err_stop_fw;
+ err = mlx4_INIT_HCA(dev, &init_hca);
+ if (err) {
+ mlx4_err(dev, "INIT_HCA command failed, aborting.\n");
+ goto err_free_icm;
+ }
+ } else {
+ err = mlx4_init_slave(dev);
+ if (err) {
+ mlx4_err(dev, "Failed to initialize slave\n");
+ return err;
+ }
- err = mlx4_INIT_HCA(dev, &init_hca);
- if (err) {
- mlx4_err(dev, "INIT_HCA command failed, aborting.\n");
- goto err_free_icm;
+ err = mlx4_slave_cap(dev);
+ if (err) {
+ mlx4_err(dev, "Failed to obtain slave caps\n");
+ goto err_close;
+ }
}
err = mlx4_QUERY_ADAPTER(dev, &adapter);
@@ -857,15 +929,17 @@ static int mlx4_init_hca(struct mlx4_dev *dev)
return 0;
err_close:
- mlx4_CLOSE_HCA(dev, 0);
+ mlx4_close_hca(dev);
err_free_icm:
- mlx4_free_icms(dev);
+ if (!mlx4_is_slave(dev))
+ mlx4_free_icms(dev);
err_stop_fw:
- mlx4_UNMAP_FA(dev);
- mlx4_free_icm(dev, priv->fw.fw_icm, 0);
-
+ if (!mlx4_is_slave(dev)) {
+ mlx4_UNMAP_FA(dev);
+ mlx4_free_icm(dev, priv->fw.fw_icm, 0);
+ }
return err;
}
@@ -1019,8 +1093,13 @@ static void mlx4_enable_msi_x(struct mlx4_dev *dev)
int i;
if (msi_x) {
- nreq = min_t(int, dev->caps.num_eqs - dev->caps.reserved_eqs,
- num_possible_cpus() + 1);
+ /* The master only uses en event EQ,
+ * Each one of the slaves have 1 completion eq */
+ if (mlx4_is_mfunc(dev))
+ nreq = 1;
+ else
+ nreq = min_t(int, dev->caps.num_eqs - dev->caps.reserved_eqs,
+ num_possible_cpus() + 1);
entries = kcalloc(nreq, sizeof *entries, GFP_KERNEL);
if (!entries)
goto no_msi;
@@ -1029,8 +1108,10 @@ static void mlx4_enable_msi_x(struct mlx4_dev *dev)
entries[i].entry = i;
retry:
+ printk("Requesting %d MSIX vectors\n", nreq);
err = pci_enable_msix(dev->pdev, entries, nreq);
if (err) {
+ printk("pci_enable_msix failed\n");
/* Try again if at least 2 vectors are available */
if (err > 1) {
mlx4_info(dev, "Requested %d vectors, "
@@ -1043,7 +1124,7 @@ static void mlx4_enable_msi_x(struct mlx4_dev *dev)
goto no_msi;
}
- dev->caps.num_comp_vectors = nreq - 1;
+ dev->caps.num_comp_vectors = nreq - !mlx4_is_slave(dev);
for (i = 0; i < nreq; ++i)
priv->eq_table.eq[i].irq = entries[i].vector;
@@ -1054,7 +1135,9 @@ static void mlx4_enable_msi_x(struct mlx4_dev *dev)
}
no_msi:
- dev->caps.num_comp_vectors = 1;
+ /* CX1: no completion vectors needed for the master since it doesn't
+ * have interfaces */
+ dev->caps.num_comp_vectors = mlx4_is_master(dev) ? 0 : 1;
for (i = 0; i < 2; ++i)
priv->eq_table.eq[i].irq = dev->pdev->irq;
@@ -1111,10 +1194,9 @@ static int __mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
}
/*
- * Check for BARs. We expect 0: 1MB
+ * Check for BARs.
*/
- if (!(pci_resource_flags(pdev, 0) & IORESOURCE_MEM) ||
- pci_resource_len(pdev, 0) != 1 << 20) {
+ if (!(pci_resource_flags(pdev, 0) & IORESOURCE_MEM)) {
dev_err(&pdev->dev, "Missing DCS, aborting.\n");
err = -ENODEV;
goto err_disable_pdev;
@@ -1172,34 +1254,98 @@ static int __mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
INIT_LIST_HEAD(&priv->pgdir_list);
mutex_init(&priv->pgdir_mutex);
- /*
- * Now reset the HCA before we touch the PCI capabilities or
- * attempt a firmware command, since a boot ROM may have left
- * the HCA in an undefined state.
- */
- err = mlx4_reset(dev);
- if (err) {
- mlx4_err(dev, "Failed to reset HCA, aborting.\n");
- goto err_free_dev;
+ /* Detect if this device is a virtual function */
+ switch (id->device) {
+ case 0x6341:
+ case 0x634b:
+ case 0x6733:
+ case 0x673d:
+ case 0x6369:
+ case 0x6751:
+ case 0x6765:
+#ifdef CONFIG_PCI_IOV
+ /* When acting as pf, we normally skip vfs unless explicitly
+ * requested to probe them.
+ * TODO: add ARI support */
+ if (sr_iov && PCI_FUNC(pdev->devfn) > probe_vf) {
+ mlx4_warn(dev, "Skipping virtual function:%d\n",
+ PCI_FUNC(pdev->devfn));
+ err = -ENODEV;
+ goto err_free_dev;
+ }
+#endif /* CONFIG_PCI_IOV */
+ mlx4_warn(dev, "Detected virtual function - running in slave mode\n");
+ dev->flags |= MLX4_FLAG_SLAVE;
+ break;
+ default:
+ ;
+ }
+
+ /* We reset the device and enable SRIOV only for physical devices */
+ if (!mlx4_is_slave(dev)) {
+ /*
+ * Now reset the HCA before we touch the PCI capabilities or
+ * attempt a firmware command, since a boot ROM may have left
+ * the HCA in an undefined state.
+ */
+ err = mlx4_reset(dev);
+ if (err) {
+ mlx4_err(dev, "Failed to reset HCA, aborting.\n");
+ goto err_free_dev;
+ }
+#ifdef CONFIG_PCI_IOV
+ if (sr_iov) {
+ mlx4_warn(dev, "Enabling sriov with:%d vfs\n", sr_iov);
+ if (pci_enable_sriov(pdev, sr_iov)) {
+ mlx4_err(dev, "Failed to enable sriov, aborting.\n");
+ goto err_free_dev;
+ }
+ mlx4_warn(dev, "Running in master mode\n");
+ dev->flags |= MLX4_FLAG_SRIOV | MLX4_FLAG_MASTER;
+ dev->num_slaves = sr_iov;
+ }
+#endif /* CONFIG_PCI_IOV */
}
if (mlx4_cmd_init(dev)) {
mlx4_err(dev, "Failed to init command interface, aborting.\n");
- goto err_free_dev;
+ goto err_sriov;
+ }
+
+ /* In slave functions, the communication channel must be initialized before
+ * posting commands */
+ if (mlx4_is_slave(dev)) {
+ if (mlx4_multi_func_init(dev)) {
+ mlx4_err(dev, "Failed to init slave mfunc interface, aborting.\n");
+ goto err_cmd;
+ }
}
err = mlx4_init_hca(dev);
if (err)
goto err_cmd;
+ /* In master functions, the communication channel must be initialized after obtaining
+ * its address from fw */
+ if (mlx4_is_master(dev)) {
+ if (mlx4_multi_func_init(dev)) {
+ mlx4_err(dev, "Failed to init master mfunc interface, aborting.\n");
+ goto err_close;
+ }
+ }
+
err = mlx4_alloc_eq_table(dev);
if (err)
goto err_close;
mlx4_enable_msi_x(dev);
+ if (mlx4_is_slave(dev) && !(dev->flags & MLX4_FLAG_MSI_X)) {
+ mlx4_err(dev, "INTx is not supported in slave mode, aborting.\n");
+ goto err_free_eq;
+ }
err = mlx4_setup_hca(dev);
- if (err == -EBUSY && (dev->flags & MLX4_FLAG_MSI_X)) {
+ if (err == -EBUSY && (dev->flags & MLX4_FLAG_MSI_X) && !mlx4_is_slave(dev)) {
dev->flags &= ~MLX4_FLAG_MSI_X;
pci_disable_msix(pdev);
err = mlx4_setup_hca(dev);
@@ -1223,6 +1369,12 @@ static int __mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
pci_set_drvdata(pdev, dev);
+ /* Start serving comm channel:
+ * - In master function: poll for commands
+ * - in slave functions: poll for events
+ * TODO - enable comm channel interrupts */
+ if (mlx4_is_master(dev) || mlx4_is_slave(dev))
+ queue_delayed_work(priv->mfunc.comm_wq, &priv->mfunc.comm_work, 0);
return 0;
err_port:
@@ -1233,7 +1385,8 @@ err_port:
mlx4_cleanup_qp_table(dev);
mlx4_cleanup_srq_table(dev);
mlx4_cleanup_cq_table(dev);
- mlx4_cmd_use_polling(dev);
+ if (!mlx4_is_master(dev) && !mlx4_is_slave(dev))
+ mlx4_cmd_use_polling(dev);
mlx4_cleanup_eq_table(dev);
mlx4_cleanup_mr_table(dev);
mlx4_cleanup_pd_table(dev);
@@ -1251,6 +1404,14 @@ err_close:
err_cmd:
mlx4_cmd_cleanup(dev);
+err_sriov:
+ if (mlx4_is_master(dev) || mlx4_is_slave(dev))
+ mlx4_multi_func_cleanup(dev);
+#ifdef CONFIG_PCI_IOV
+ if (dev->flags & MLX4_FLAG_SRIOV)
+ pci_disable_sriov(pdev);
+#endif /* CONFIG_PCI_IOV */
+
err_free_dev:
kfree(priv);
@@ -1283,19 +1444,24 @@ static void mlx4_remove_one(struct pci_dev *pdev)
int p;
if (dev) {
+ /* Stop serving commands and events over comm channel */
+ if (mlx4_is_master(dev) || mlx4_is_slave(dev))
+ cancel_delayed_work_sync(&priv->mfunc.comm_work);
mlx4_stop_sense(dev);
mlx4_unregister_device(dev);
for (p = 1; p <= dev->caps.num_ports; p++) {
mlx4_cleanup_port_info(&priv->port[p]);
- mlx4_CLOSE_PORT(dev, p);
+ if (!mlx4_is_slave(dev))
+ mlx4_CLOSE_PORT(dev, p);
}
mlx4_cleanup_mcg_table(dev);
mlx4_cleanup_qp_table(dev);
mlx4_cleanup_srq_table(dev);
mlx4_cleanup_cq_table(dev);
- mlx4_cmd_use_polling(dev);
+ if (!mlx4_is_master(dev) && !mlx4_is_slave(dev))
+ mlx4_cmd_use_polling(dev);
mlx4_cleanup_eq_table(dev);
mlx4_cleanup_mr_table(dev);
mlx4_cleanup_pd_table(dev);
@@ -1303,10 +1469,18 @@ static void mlx4_remove_one(struct pci_dev *pdev)
mlx4_cleanup_uar_table(dev);
mlx4_free_eq_table(dev);
mlx4_close_hca(dev);
+ if (mlx4_is_master(dev) || mlx4_is_slave(dev))
+ mlx4_multi_func_cleanup(dev);
mlx4_cmd_cleanup(dev);
if (dev->flags & MLX4_FLAG_MSI_X)
pci_disable_msix(pdev);
+#ifdef CONFIG_PCI_IOV
+ if (dev->flags & MLX4_FLAG_SRIOV) {
+ mlx4_warn(dev, "Disabling sriov\n");
+ pci_disable_sriov(pdev);
+ }
+#endif /* CONFIG_PCI_IOV */
kfree(priv);
pci_release_regions(pdev);
@@ -1321,14 +1495,21 @@ int mlx4_restart_one(struct pci_dev *pdev)
return __mlx4_init_one(pdev, NULL);
}
+
static struct pci_device_id mlx4_pci_table[] = {
{ PCI_VDEVICE(MELLANOX, 0x6340) }, /* MT25408 "Hermon" SDR */
+ { PCI_VDEVICE(MELLANOX, 0x6341) }, /* MT25408 "Hermon" SDR VF */
{ PCI_VDEVICE(MELLANOX, 0x634a) }, /* MT25408 "Hermon" DDR */
+ { PCI_VDEVICE(MELLANOX, 0x634b) }, /* MT25408 "Hermon" DDR VF */
{ PCI_VDEVICE(MELLANOX, 0x6354) }, /* MT25408 "Hermon" QDR */
{ PCI_VDEVICE(MELLANOX, 0x6732) }, /* MT25408 "Hermon" DDR PCIe gen2 */
+ { PCI_VDEVICE(MELLANOX, 0x6733) }, /* MT25408 "Hermon" DDR PCIe gen2 VF */
{ PCI_VDEVICE(MELLANOX, 0x673c) }, /* MT25408 "Hermon" QDR PCIe gen2 */
+ { PCI_VDEVICE(MELLANOX, 0x673d) }, /* MT25408 "Hermon" QDR PCIe gen2 VF */
{ PCI_VDEVICE(MELLANOX, 0x6368) }, /* MT25408 "Hermon" EN 10GigE */
+ { PCI_VDEVICE(MELLANOX, 0x6369) }, /* MT25408 "Hermon" EN 10GigE VF */
{ PCI_VDEVICE(MELLANOX, 0x6750) }, /* MT25408 "Hermon" EN 10GigE PCIe gen2 */
+ { PCI_VDEVICE(MELLANOX, 0x6751) }, /* MT25408 "Hermon" EN 10GigE PCIe gen2 VF */
{ PCI_VDEVICE(MELLANOX, 0x6372) }, /* MT25458 ConnectX EN 10GBASE-T 10GigE */
{ PCI_VDEVICE(MELLANOX, 0x675a) }, /* MT25458 ConnectX EN 10GBASE-T+Gen2 10GigE */
{ PCI_VDEVICE(MELLANOX, 0x6764) }, /* MT26468 ConnectX EN 10GigE PCIe gen2*/
--
1.6.1.3
--
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
^ permalink raw reply related
* [PATCH 17/25] mlx4: Changed Mac management
From: Yevgeny Petrilin @ 2009-11-04 15:31 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
RX qp is given from a pool of previously reserved range of QPs.
Activating Mac steering now.
Giving an option to change mac without the need to change the RX QP.
All the mac management is done by the master, added wrappers for the
mac management.
Signed-off-by: Yevgeny Petrilin <yevgenyp-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/net/mlx4/cmd.c | 19 ++++++++
drivers/net/mlx4/en_netdev.c | 32 +++++++-------
drivers/net/mlx4/en_port.c | 1 +
drivers/net/mlx4/en_port.h | 8 +++-
drivers/net/mlx4/en_rx.c | 11 +----
drivers/net/mlx4/main.c | 11 +++--
drivers/net/mlx4/mlx4.h | 7 ++-
drivers/net/mlx4/port.c | 96 ++++++++++++++++++++++++++++++++----------
include/linux/mlx4/cmd.h | 9 ++--
include/linux/mlx4/device.h | 5 +-
10 files changed, 137 insertions(+), 62 deletions(-)
diff --git a/drivers/net/mlx4/cmd.c b/drivers/net/mlx4/cmd.c
index 7afbfef..183ba8d 100644
--- a/drivers/net/mlx4/cmd.c
+++ b/drivers/net/mlx4/cmd.c
@@ -507,6 +507,24 @@ static int mlx4_RESOURCE_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vh
} else
mlx4_free_mtt_range(dev, param1 /* first */, param2 /* order */);
break;
+ case RES_MAC:
+ switch (vhcr->op) {
+ case MLX4_CMD_ALLOC_RES:
+ ret = mlx4_register_mac(dev, vhcr->op_modifier,
+ vhcr->in_param, (int *) &vhcr->out_param);
+ vhcr->errno = ret;
+ break;
+ case MLX4_CMD_FREE_RES:
+ mlx4_unregister_mac(dev, vhcr->op_modifier, vhcr->in_param);
+ break;
+ case MLX4_CMD_REPLACE_RES:
+ ret = mlx4_replace_mac(dev, vhcr->op_modifier,
+ vhcr->out_param, vhcr->in_param);
+ vhcr->errno = ret;
+ break;
+ default:
+ vhcr->errno = -EINVAL;
+ }
default:
vhcr->errno = -EINVAL;
}
@@ -538,6 +556,7 @@ static struct mlx4_cmd_info {
{MLX4_CMD_NOP, 0, 0, 0, NULL, NULL},
{MLX4_CMD_ALLOC_RES, 0, 0, 1, NULL, mlx4_RESOURCE_wrapper},
{MLX4_CMD_FREE_RES, 0, 0, 0, NULL, mlx4_RESOURCE_wrapper},
+ {MLX4_CMD_REPLACE_RES, 0, 0, 1, NULL, mlx4_RESOURCE_wrapper},
{MLX4_CMD_GET_EVENT, 0, 0, 1, NULL, mlx4_GET_EVENT_wrapper},
{MLX4_CMD_SW2HW_MPT, 1, 0, 0, NULL, NULL}, /* need verifier */
diff --git a/drivers/net/mlx4/en_netdev.c b/drivers/net/mlx4/en_netdev.c
index c48b0f4..fd96078 100644
--- a/drivers/net/mlx4/en_netdev.c
+++ b/drivers/net/mlx4/en_netdev.c
@@ -145,9 +145,8 @@ static void mlx4_en_do_set_mac(struct work_struct *work)
mutex_lock(&mdev->state_lock);
if (priv->port_up) {
/* Remove old MAC and insert the new one */
- mlx4_unregister_mac(mdev->dev, priv->port, priv->mac_index);
- err = mlx4_register_mac(mdev->dev, priv->port,
- priv->mac, &priv->mac_index);
+ err = mlx4_replace_mac(mdev->dev, priv->port,
+ priv->base_qpn, priv->mac);
if (err)
en_err(priv, "Failed changing HW MAC address\n");
} else
@@ -596,10 +595,19 @@ int mlx4_en_start_port(struct net_device *dev)
++rx_index;
}
+ /* Set port mac number */
+ en_dbg(DRV, priv, "Setting mac for port %d\n", priv->port);
+ err = mlx4_register_mac(mdev->dev, priv->port,
+ priv->mac, &priv->base_qpn);
+ if (err) {
+ en_err(priv, "Failed setting port mac\n");
+ goto cq_err;
+ }
+
err = mlx4_en_config_rss_steer(priv);
if (err) {
en_err(priv, "Failed configuring rss steering\n");
- goto cq_err;
+ goto mac_err;
}
/* Configure tx cq's and rings */
@@ -652,21 +660,13 @@ int mlx4_en_start_port(struct net_device *dev)
en_err(priv, "Failed setting default qp numbers\n");
goto tx_err;
}
- /* Set port mac number */
- en_dbg(DRV, priv, "Setting mac for port %d\n", priv->port);
- err = mlx4_register_mac(mdev->dev, priv->port,
- priv->mac, &priv->mac_index);
- if (err) {
- en_err(priv, "Failed setting port mac\n");
- goto tx_err;
- }
/* Init port */
en_dbg(HW, priv, "Initializing port\n");
err = mlx4_INIT_PORT(mdev->dev, priv->port);
if (err) {
en_err(priv, "Failed Initializing port\n");
- goto mac_err;
+ goto tx_err;
}
/* Schedule multicast task to populate multicast list */
@@ -676,8 +676,6 @@ int mlx4_en_start_port(struct net_device *dev)
netif_tx_start_all_queues(dev);
return 0;
-mac_err:
- mlx4_unregister_mac(mdev->dev, priv->port, priv->mac_index);
tx_err:
while (tx_index--) {
mlx4_en_deactivate_tx_ring(priv, &priv->tx_ring[tx_index]);
@@ -685,6 +683,8 @@ tx_err:
}
mlx4_en_release_rss_steer(priv);
+mac_err:
+ mlx4_unregister_mac(mdev->dev, priv->port, priv->base_qpn);
cq_err:
while (rx_index--)
mlx4_en_deactivate_cq(priv, &priv->rx_cq[rx_index]);
@@ -716,7 +716,7 @@ void mlx4_en_stop_port(struct net_device *dev)
mlx4_CLOSE_PORT(mdev->dev, priv->port);
/* Unregister Mac address for the port */
- mlx4_unregister_mac(mdev->dev, priv->port, priv->mac_index);
+ mlx4_unregister_mac(mdev->dev, priv->port, priv->base_qpn);
/* Free TX Rings */
for (i = 0; i < priv->tx_ring_num; i++) {
diff --git a/drivers/net/mlx4/en_port.c b/drivers/net/mlx4/en_port.c
index a29abe8..c099cb4 100644
--- a/drivers/net/mlx4/en_port.c
+++ b/drivers/net/mlx4/en_port.c
@@ -127,6 +127,7 @@ int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
memset(context, 0, sizeof *context);
context->base_qpn = cpu_to_be32(base_qpn);
+ context->n_mac = 0x7;
context->promisc = cpu_to_be32(promisc << SET_PORT_PROMISC_SHIFT | base_qpn);
context->mcast = cpu_to_be32(1 << SET_PORT_PROMISC_SHIFT | base_qpn);
context->intra_no_vlan = 0;
diff --git a/drivers/net/mlx4/en_port.h b/drivers/net/mlx4/en_port.h
index e6477f1..674a804 100644
--- a/drivers/net/mlx4/en_port.h
+++ b/drivers/net/mlx4/en_port.h
@@ -59,14 +59,18 @@ struct mlx4_set_port_general_context {
struct mlx4_set_port_rqp_calc_context {
__be32 base_qpn;
+ u8 rererved;
+ u8 n_mac;
+ u8 n_vlan;
+ u8 n_prio;
__be32 flags;
- u8 reserved[3];
+ u8 reserved2[3];
u8 mac_miss;
u8 intra_no_vlan;
u8 no_vlan;
u8 intra_vlan_miss;
u8 vlan_miss;
- u8 reserved2[3];
+ u8 reserved3[3];
u8 no_vlan_prio;
__be32 promisc;
__be32 mcast;
diff --git a/drivers/net/mlx4/en_rx.c b/drivers/net/mlx4/en_rx.c
index 03b781a..d3884ce 100644
--- a/drivers/net/mlx4/en_rx.c
+++ b/drivers/net/mlx4/en_rx.c
@@ -865,16 +865,10 @@ int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv)
}
/* Configure RSS indirection qp */
- err = mlx4_qp_reserve_range(mdev->dev, 1, 1, &priv->base_qpn);
- if (err) {
- en_err(priv, "Failed to reserve range for RSS "
- "indirection qp\n");
- goto rss_err;
- }
err = mlx4_qp_alloc(mdev->dev, priv->base_qpn, &rss_map->indir_qp);
if (err) {
en_err(priv, "Failed to allocate RSS indirection QP\n");
- goto reserve_err;
+ goto rss_err;
}
rss_map->indir_qp.event = mlx4_en_sqp_event;
mlx4_en_fill_qp_context(priv, 0, 0, 0, 1, priv->base_qpn,
@@ -900,8 +894,6 @@ indir_err:
MLX4_QP_STATE_RST, NULL, 0, 0, &rss_map->indir_qp);
mlx4_qp_remove(mdev->dev, &rss_map->indir_qp);
mlx4_qp_free(mdev->dev, &rss_map->indir_qp);
-reserve_err:
- mlx4_qp_release_range(mdev->dev, priv->base_qpn, 1);
rss_err:
for (i = 0; i < good_qps; i++) {
mlx4_qp_modify(mdev->dev, NULL, rss_map->state[i],
@@ -923,7 +915,6 @@ void mlx4_en_release_rss_steer(struct mlx4_en_priv *priv)
MLX4_QP_STATE_RST, NULL, 0, 0, &rss_map->indir_qp);
mlx4_qp_remove(mdev->dev, &rss_map->indir_qp);
mlx4_qp_free(mdev->dev, &rss_map->indir_qp);
- mlx4_qp_release_range(mdev->dev, priv->base_qpn, 1);
for (i = 0; i < priv->rx_ring_num; i++) {
mlx4_qp_modify(mdev->dev, NULL, rss_map->state[i],
diff --git a/drivers/net/mlx4/main.c b/drivers/net/mlx4/main.c
index c2b6b35..c5ee288 100644
--- a/drivers/net/mlx4/main.c
+++ b/drivers/net/mlx4/main.c
@@ -99,7 +99,7 @@ static struct mlx4_profile default_profile = {
.num_mtt = 1 << 20,
};
-static int log_num_mac = 2;
+static int log_num_mac = 7;
module_param_named(log_num_mac, log_num_mac, int, 0444);
MODULE_PARM_DESC(log_num_mac, "Log2 max number of MACs per ETH port (1-7)");
@@ -1150,9 +1150,12 @@ static int mlx4_init_port_info(struct mlx4_dev *dev, int port)
info->dev = dev;
info->port = port;
- mlx4_init_mac_table(dev, &info->mac_table);
- mlx4_init_vlan_table(dev, &info->vlan_table);
-
+ if (!mlx4_is_slave(dev)) {
+ mlx4_init_mac_table(dev, &info->mac_table);
+ mlx4_init_vlan_table(dev, &info->vlan_table);
+ info->base_qpn = dev->caps.reserved_qps_base[MLX4_QP_REGION_ETH_ADDR] +
+ (port - 1) * (1 << log_num_mac);
+ }
sprintf(info->dev_name, "mlx4_port%d", port);
info->port_attr.attr.name = info->dev_name;
info->port_attr.attr.mode = S_IRUGO | S_IWUSR;
diff --git a/drivers/net/mlx4/mlx4.h b/drivers/net/mlx4/mlx4.h
index 516e375..3e434e4 100644
--- a/drivers/net/mlx4/mlx4.h
+++ b/drivers/net/mlx4/mlx4.h
@@ -101,7 +101,10 @@ enum mlx4_resource {
RES_CQ,
RES_SRQ,
RES_MPT,
- RES_MTT
+ RES_MTT,
+ RES_MAC,
+ RES_VLAN,
+ RES_MCAST
};
enum mlx4_alloc_mode {
@@ -340,7 +343,6 @@ struct mlx4_catas_err {
struct mlx4_mac_table {
__be64 entries[MLX4_MAX_MAC_NUM];
- int refs[MLX4_MAX_MAC_NUM];
struct mutex mutex;
int total;
int max;
@@ -365,6 +367,7 @@ struct mlx4_port_info {
enum mlx4_port_type tmp_type;
struct mlx4_mac_table mac_table;
struct mlx4_vlan_table vlan_table;
+ int base_qpn;
};
struct mlx4_sense {
diff --git a/drivers/net/mlx4/port.c b/drivers/net/mlx4/port.c
index 67f0751..6e729dd 100644
--- a/drivers/net/mlx4/port.c
+++ b/drivers/net/mlx4/port.c
@@ -48,10 +48,8 @@ void mlx4_init_mac_table(struct mlx4_dev *dev, struct mlx4_mac_table *table)
int i;
mutex_init(&table->mutex);
- for (i = 0; i < MLX4_MAX_MAC_NUM; i++) {
+ for (i = 0; i < MLX4_MAX_MAC_NUM; i++)
table->entries[i] = 0;
- table->refs[i] = 0;
- }
table->max = 1 << dev->caps.log_num_macs;
table->total = 0;
}
@@ -90,24 +88,33 @@ static int mlx4_set_port_mac_table(struct mlx4_dev *dev, u8 port,
return err;
}
-int mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *index)
+int mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *qpn)
{
- struct mlx4_mac_table *table = &mlx4_priv(dev)->port[port].mac_table;
+ struct mlx4_port_info *info = &mlx4_priv(dev)->port[port];
+ struct mlx4_mac_table *table = &info->mac_table;
+ u64 out_param;
int i, err = 0;
int free = -1;
+ if (mlx4_is_slave(dev)) {
+ err = mlx4_cmd_imm(dev, mac, &out_param, RES_MAC, port,
+ MLX4_CMD_ALLOC_RES, MLX4_CMD_TIME_CLASS_A);
+ if (!err)
+ *qpn = out_param;
+ return err;
+ }
+
mlx4_dbg(dev, "Registering MAC: 0x%llx\n", (unsigned long long) mac);
mutex_lock(&table->mutex);
for (i = 0; i < MLX4_MAX_MAC_NUM - 1; i++) {
- if (free < 0 && !table->refs[i]) {
+ if (free < 0 && !table->entries[i]) {
free = i;
continue;
}
if (mac == (MLX4_MAC_MASK & be64_to_cpu(table->entries[i]))) {
- /* MAC already registered, increase refernce count */
- *index = i;
- ++table->refs[i];
+ /* MAC already registered, Must not have duplicates */
+ err = -EEXIST;
goto out;
}
}
@@ -120,18 +127,16 @@ int mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *index)
}
/* Register new MAC */
- table->refs[free] = 1;
table->entries[free] = cpu_to_be64(mac | MLX4_MAC_VALID);
err = mlx4_set_port_mac_table(dev, port, table->entries);
if (unlikely(err)) {
mlx4_err(dev, "Failed adding MAC: 0x%llx\n", (unsigned long long) mac);
- table->refs[free] = 0;
table->entries[free] = 0;
goto out;
}
- *index = free;
+ *qpn = info->base_qpn + free;
++table->total;
out:
mutex_unlock(&table->mutex);
@@ -139,20 +144,35 @@ out:
}
EXPORT_SYMBOL_GPL(mlx4_register_mac);
-void mlx4_unregister_mac(struct mlx4_dev *dev, u8 port, int index)
+static int validate_index(struct mlx4_dev *dev,
+ struct mlx4_mac_table *table, int index)
{
- struct mlx4_mac_table *table = &mlx4_priv(dev)->port[port].mac_table;
+ int err = 0;
- mutex_lock(&table->mutex);
- if (!table->refs[index]) {
- mlx4_warn(dev, "No MAC entry for index %d\n", index);
- goto out;
+ if (index < 0 || index >= table->max || !table->entries[index]) {
+ mlx4_warn(dev, "No valid Mac entry for the given index\n");
+ err = -EINVAL;
}
- if (--table->refs[index]) {
- mlx4_warn(dev, "Have more references for index %d,"
- "no need to modify MAC table\n", index);
- goto out;
+ return err;
+}
+
+void mlx4_unregister_mac(struct mlx4_dev *dev, u8 port, int qpn)
+{
+ struct mlx4_port_info *info = &mlx4_priv(dev)->port[port];
+ struct mlx4_mac_table *table = &info->mac_table;
+ int index = qpn - info->base_qpn;
+
+ if (mlx4_is_slave(dev)) {
+ mlx4_cmd(dev, qpn, RES_MAC, port,
+ MLX4_CMD_FREE_RES, MLX4_CMD_TIME_CLASS_A);
+ return;
}
+
+ mutex_lock(&table->mutex);
+
+ if (validate_index(dev, table, index))
+ goto out;
+
table->entries[index] = 0;
mlx4_set_port_mac_table(dev, port, table->entries);
--table->total;
@@ -161,6 +181,38 @@ out:
}
EXPORT_SYMBOL_GPL(mlx4_unregister_mac);
+int mlx4_replace_mac(struct mlx4_dev *dev, u8 port, int qpn, u64 new_mac)
+{
+ struct mlx4_port_info *info = &mlx4_priv(dev)->port[port];
+ struct mlx4_mac_table *table = &info->mac_table;
+ int index = qpn - info->base_qpn;
+ int err;
+
+ if (mlx4_is_slave(dev)) {
+ err = mlx4_cmd_imm(dev, new_mac, (u64 *) &qpn, RES_MAC, port,
+ MLX4_CMD_REPLACE_RES, MLX4_CMD_TIME_CLASS_A);
+ return err;
+ }
+
+ mutex_lock(&table->mutex);
+
+ err = validate_index(dev, table, index);
+ if (err)
+ goto out;
+
+ table->entries[index] = cpu_to_be64(new_mac | MLX4_MAC_VALID);
+
+ err = mlx4_set_port_mac_table(dev, port, table->entries);
+ if (unlikely(err)) {
+ mlx4_err(dev, "Failed adding MAC: 0x%llx\n", (unsigned long long) new_mac);
+ table->entries[index] = 0;
+ }
+out:
+ mutex_unlock(&table->mutex);
+ return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_replace_mac);
+
static int mlx4_set_port_vlan_table(struct mlx4_dev *dev, u8 port,
__be32 *entries)
{
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index f265091..f729c82 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -122,10 +122,11 @@ enum {
/* virtual commands */
MLX4_CMD_ALLOC_RES = 0x50,
MLX4_CMD_FREE_RES = 0x51,
- MLX4_CMD_GET_EVENT = 0x52,
- MLX4_CMD_QUERY_SLAVE_CAP = 0x53,
- MLX4_CMD_MCAST_ATTACH = 0x54,
- MLX4_CMD_GET_SLAVE_SQP = 0x55,
+ MLX4_CMD_REPLACE_RES = 0x52,
+ MLX4_CMD_GET_EVENT = 0x53,
+ MLX4_CMD_QUERY_SLAVE_CAP = 0x54,
+ MLX4_CMD_MCAST_ATTACH = 0x55,
+ MLX4_CMD_GET_SLAVE_SQP = 0x56,
/* debug commands */
MLX4_CMD_QUERY_DEBUG_MSG = 0x2a,
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index bbd398b..246b7bc 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -491,8 +491,9 @@ int mlx4_multicast_attach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
int block_mcast_loopback);
int mlx4_multicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16]);
-int mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *index);
-void mlx4_unregister_mac(struct mlx4_dev *dev, u8 port, int index);
+int mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *qpn);
+void mlx4_unregister_mac(struct mlx4_dev *dev, u8 port, int qpn);
+int mlx4_replace_mac(struct mlx4_dev *dev, u8 port, int qpn, u64 new_mac);
int mlx4_register_vlan(struct mlx4_dev *dev, u8 port, u16 vlan, int *index);
void mlx4_unregister_vlan(struct mlx4_dev *dev, u8 port, int index);
--
1.6.1.3
--
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
^ permalink raw reply related
* [PATCH 18/25] mlx4_en: Use reasonable resources for slaves.
From: Yevgeny Petrilin @ 2009-11-04 15:31 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
Each slave gets a single MSI-X vector when in multi function mode.
In this case each slave would allocate less resources (which
normally depend on number of completion vectors).
We set a minimum number of RX and TX queues to achieve an
acceptable performance in multi function mode.
Signed-off-by: Yevgeny Petrilin <yevgenyp-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/net/mlx4/en_main.c | 12 +++++++-----
drivers/net/mlx4/en_netdev.c | 3 ++-
drivers/net/mlx4/en_tx.c | 2 +-
drivers/net/mlx4/mlx4_en.h | 4 +++-
4 files changed, 13 insertions(+), 8 deletions(-)
diff --git a/drivers/net/mlx4/en_main.c b/drivers/net/mlx4/en_main.c
index 507e11f..3883071 100644
--- a/drivers/net/mlx4/en_main.c
+++ b/drivers/net/mlx4/en_main.c
@@ -93,8 +93,6 @@ static int mlx4_en_get_profile(struct mlx4_en_dev *mdev)
params->prof[i].tx_ppp = pfctx;
params->prof[i].tx_ring_size = MLX4_EN_DEF_TX_RING_SIZE;
params->prof[i].rx_ring_size = MLX4_EN_DEF_RX_RING_SIZE;
- params->prof[i].tx_ring_num = MLX4_EN_NUM_TX_RINGS +
- (!!pfcrx) * MLX4_EN_NUM_PPP_RINGS;
}
return 0;
@@ -216,11 +214,15 @@ static void *mlx4_en_add(struct mlx4_dev *dev)
/* If we did not receive an explicit number of Rx rings, default to
* the number of completion vectors populated by the mlx4_core */
mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) {
+ mdev->profile.prof[i].tx_ring_num = max_t(int,
+ min_t(int, roundup_pow_of_two(dev->caps.num_comp_vectors),
+ NUM_TX_RINGS), MIN_TX_RINGS) +
+ (!!pfcrx) * MLX4_EN_NUM_PPP_RINGS;
mlx4_info(mdev, "Using %d tx rings for port:%d\n",
mdev->profile.prof[i].tx_ring_num, i);
- mdev->profile.prof[i].rx_ring_num = min_t(int,
- roundup_pow_of_two(dev->caps.num_comp_vectors),
- MAX_RX_RINGS);
+ mdev->profile.prof[i].rx_ring_num = max_t(int,
+ min_t(int, roundup_pow_of_two(dev->caps.num_comp_vectors),
+ MAX_RX_RINGS), MIN_RX_RINGS);
mlx4_info(mdev, "Defaulting to %d rx rings for port:%d\n",
mdev->profile.prof[i].rx_ring_num, i);
}
diff --git a/drivers/net/mlx4/en_netdev.c b/drivers/net/mlx4/en_netdev.c
index fd96078..488f967 100644
--- a/drivers/net/mlx4/en_netdev.c
+++ b/drivers/net/mlx4/en_netdev.c
@@ -1024,7 +1024,8 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
*/
dev->netdev_ops = &mlx4_netdev_ops;
dev->watchdog_timeo = MLX4_EN_WATCHDOG_TIMEOUT;
- dev->real_num_tx_queues = MLX4_EN_NUM_TX_RINGS;
+ dev->real_num_tx_queues = priv->tx_ring_num -
+ !!(priv->prof->rx_ppp) * MLX4_EN_NUM_PPP_RINGS;
SET_ETHTOOL_OPS(dev, &mlx4_en_ethtool_ops);
diff --git a/drivers/net/mlx4/en_tx.c b/drivers/net/mlx4/en_tx.c
index 8c72799..a893733 100644
--- a/drivers/net/mlx4/en_tx.c
+++ b/drivers/net/mlx4/en_tx.c
@@ -583,7 +583,7 @@ u16 mlx4_en_select_queue(struct net_device *dev, struct sk_buff *skb)
*/
if (priv->prof->rx_ppp && priv->vlgrp && vlan_tx_tag_present(skb)) {
vlan_tag = vlan_tx_tag_get(skb);
- return MLX4_EN_NUM_TX_RINGS + (vlan_tag >> 13);
+ return priv->tx_ring_num - MLX4_EN_NUM_PPP_RINGS + (vlan_tag >> 13);
}
return skb_tx_hash(dev, skb);
diff --git a/drivers/net/mlx4/mlx4_en.h b/drivers/net/mlx4/mlx4_en.h
index 4376147..6e0c9ed 100644
--- a/drivers/net/mlx4/mlx4_en.h
+++ b/drivers/net/mlx4/mlx4_en.h
@@ -95,6 +95,9 @@
#define MLX4_EN_PAGE_SIZE (1 << MLX4_EN_PAGE_SHIFT)
#define MAX_TX_RINGS 16
#define MAX_RX_RINGS 16
+#define MIN_RX_RINGS 4
+#define NUM_TX_RINGS 8
+#define MIN_TX_RINGS 4
#define TXBB_SIZE 64
#define HEADROOM (2048 / TXBB_SIZE + 1)
#define STAMP_STRIDE 64
@@ -137,7 +140,6 @@ enum {
#define MLX4_EN_MIN_TX_SIZE (4096 / TXBB_SIZE)
#define MLX4_EN_SMALL_PKT_SIZE 64
-#define MLX4_EN_NUM_TX_RINGS 8
#define MLX4_EN_NUM_PPP_RINGS 8
#define MLX4_EN_DEF_TX_RING_SIZE 512
#define MLX4_EN_DEF_RX_RING_SIZE 1024
--
1.6.1.3
--
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
^ permalink raw reply related
* [PATCH 19/25] mlx4: Randomizing mac addresses for slaves
From: Yevgeny Petrilin @ 2009-11-04 15:31 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
Signed-off-by: Yevgeny Petrilin <yevgenyp-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/net/mlx4/fw.c | 6 ++++++
1 files changed, 6 insertions(+), 0 deletions(-)
diff --git a/drivers/net/mlx4/fw.c b/drivers/net/mlx4/fw.c
index 60889d3..9028592 100644
--- a/drivers/net/mlx4/fw.c
+++ b/drivers/net/mlx4/fw.c
@@ -32,6 +32,7 @@
* SOFTWARE.
*/
+#include <linux/etherdevice.h>
#include <linux/mlx4/cmd.h>
#include <linux/cache.h>
@@ -148,6 +149,7 @@ int mlx4_QUERY_SLAVE_CAP_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vh
struct mlx4_cmd_mailbox *outbox)
{
struct mlx4_caps *caps = outbox->buf;
+ u8 rand_mac[6];
int i;
memcpy(caps, &dev->caps, sizeof *caps);
@@ -165,6 +167,10 @@ int mlx4_QUERY_SLAVE_CAP_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vh
for (i = 1; i <= dev->caps.num_ports; ++i) {
caps->gid_table_len[i] = 1;
caps->pkey_table_len[i] = 1;
+#define MLX4_MAC_HEAD 0x2c9000000ULL
+ random_ether_addr(rand_mac);
+ caps->def_mac[i] = MLX4_MAC_HEAD | rand_mac[0] |
+ ((u64)(rand_mac[1]) << 8) | ((u64)(rand_mac[2]) << 16);
}
} else {
caps->sqp_demux = dev->num_slaves;
--
1.6.1.3
--
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
^ permalink raw reply related
* [PATCH 21/25] mlx4: Adding protocol fields to multicast group
From: Yevgeny Petrilin @ 2009-11-04 15:32 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
The multicast attachment mechanism will be used both by
IB and Ethernet, so we need to specify for each multicast
address (whether it is gid or mac) its protocol.
Signed-off-by: Yevgeny Petrilin <yevgenyp-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/infiniband/hw/mlx4/main.c | 6 +++-
drivers/net/mlx4/mcg.c | 41 +++++++++++++++++++++---------------
include/linux/mlx4/device.h | 12 +++++++++-
3 files changed, 38 insertions(+), 21 deletions(-)
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index 3cb3f47..5b67a31 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -450,13 +450,15 @@ static int mlx4_ib_mcg_attach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid)
return mlx4_multicast_attach(to_mdev(ibqp->device)->dev,
&to_mqp(ibqp)->mqp, gid->raw,
!!(to_mqp(ibqp)->flags &
- MLX4_IB_QP_BLOCK_MULTICAST_LOOPBACK));
+ MLX4_IB_QP_BLOCK_MULTICAST_LOOPBACK),
+ MLX4_PROT_IB_IPV6);
}
static int mlx4_ib_mcg_detach(struct ib_qp *ibqp, union ib_gid *gid, u16 lid)
{
return mlx4_multicast_detach(to_mdev(ibqp->device)->dev,
- &to_mqp(ibqp)->mqp, gid->raw);
+ &to_mqp(ibqp)->mqp, gid->raw,
+ MLX4_PROT_IB_IPV6);
}
static int init_node_data(struct mlx4_ib_dev *dev)
diff --git a/drivers/net/mlx4/mcg.c b/drivers/net/mlx4/mcg.c
index daa08f1..c4f928a 100644
--- a/drivers/net/mlx4/mcg.c
+++ b/drivers/net/mlx4/mcg.c
@@ -96,7 +96,8 @@ static int mlx4_MGID_HASH(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *mailbox
* entry in hash chain and *mgm holds end of hash chain.
*/
static int find_mgm(struct mlx4_dev *dev,
- u8 *gid, struct mlx4_cmd_mailbox *mgm_mailbox,
+ u8 *gid, enum mlx4_protocol prot,
+ struct mlx4_cmd_mailbox *mgm_mailbox,
u16 *hash, int *prev, int *index)
{
struct mlx4_cmd_mailbox *mailbox;
@@ -135,8 +136,9 @@ static int find_mgm(struct mlx4_dev *dev,
return err;
}
- if (!memcmp(mgm->gid, gid, 16))
- return err;
+ if (!memcmp(mgm->gid, gid, 16) &&
+ (prot == be32_to_cpu(mgm->members_count) >> 30))
+ return err;
*prev = *index;
*index = be32_to_cpu(mgm->next_gid_index) >> 6;
@@ -154,14 +156,17 @@ int mlx4_MCAST_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr,
qp.qpn = vhcr->in_modifier & 0xffffff;
if (vhcr->op_modifier)
- return mlx4_multicast_attach(dev, &qp, inbox->buf, vhcr->in_modifier >> 31);
+ return mlx4_multicast_attach(dev, &qp, inbox->buf,
+ vhcr->in_modifier >> 31,
+ (vhcr->in_modifier >> 28) & 0x7);
else
- return mlx4_multicast_detach(dev, &qp, inbox->buf);
+ return mlx4_multicast_detach(dev, &qp, inbox->buf,
+ (vhcr->in_modifier >> 28) & 0x7);
}
static int mlx4_MCAST(struct mlx4_dev *dev, struct mlx4_qp *qp,
- u8 gid[16], u8 attach,
- u8 block_loopback)
+ u8 gid[16], u8 attach, u8 block_loopback,
+ enum mlx4_protocol prot)
{
struct mlx4_cmd_mailbox *mailbox;
int err;
@@ -176,6 +181,7 @@ static int mlx4_MCAST(struct mlx4_dev *dev, struct mlx4_qp *qp,
memcpy(mailbox->buf, gid, 16);
qpn = qp->qpn;
+ qpn |= (prot << 28);
if (attach && block_loopback)
qpn |= (1 << 31);
@@ -186,7 +192,7 @@ static int mlx4_MCAST(struct mlx4_dev *dev, struct mlx4_qp *qp,
}
int mlx4_multicast_attach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
- int block_mcast_loopback)
+ int block_mcast_loopback, enum mlx4_protocol prot)
{
struct mlx4_priv *priv = mlx4_priv(dev);
struct mlx4_cmd_mailbox *mailbox;
@@ -199,7 +205,7 @@ int mlx4_multicast_attach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
int err;
if (mlx4_is_slave(dev))
- return mlx4_MCAST(dev, qp, gid, 1, block_mcast_loopback);
+ return mlx4_MCAST(dev, qp, gid, 1, block_mcast_loopback, prot);
mailbox = mlx4_alloc_cmd_mailbox(dev);
if (IS_ERR(mailbox))
@@ -208,7 +214,7 @@ int mlx4_multicast_attach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
mutex_lock(&priv->mcg_table.mutex);
- err = find_mgm(dev, gid, mailbox, &hash, &prev, &index);
+ err = find_mgm(dev, gid, prot, mailbox, &hash, &prev, &index);
if (err)
goto out;
@@ -230,7 +236,7 @@ int mlx4_multicast_attach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
memcpy(mgm->gid, gid, 16);
}
- members_count = be32_to_cpu(mgm->members_count);
+ members_count = be32_to_cpu(mgm->members_count) & 0xffffff;
if (members_count == MLX4_QP_PER_MGM) {
mlx4_err(dev, "MGM at index %x is full.\n", index);
err = -ENOMEM;
@@ -250,7 +256,7 @@ int mlx4_multicast_attach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
else
mgm->qp[members_count++] = cpu_to_be32(qp->qpn & MGM_QPN_MASK);
- mgm->members_count = cpu_to_be32(members_count);
+ mgm->members_count = cpu_to_be32(members_count | ((u32) prot << 30));
err = mlx4_WRITE_MCG(dev, index, mailbox);
if (err)
@@ -285,7 +291,8 @@ out:
}
EXPORT_SYMBOL_GPL(mlx4_multicast_attach);
-int mlx4_multicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16])
+int mlx4_multicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
+ enum mlx4_protocol prot)
{
struct mlx4_priv *priv = mlx4_priv(dev);
struct mlx4_cmd_mailbox *mailbox;
@@ -297,7 +304,7 @@ int mlx4_multicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16])
int err;
if (mlx4_is_slave(dev))
- return mlx4_MCAST(dev, qp, gid, 0, 0);
+ return mlx4_MCAST(dev, qp, gid, 0, 0, prot);
mailbox = mlx4_alloc_cmd_mailbox(dev);
if (IS_ERR(mailbox))
@@ -306,7 +313,7 @@ int mlx4_multicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16])
mutex_lock(&priv->mcg_table.mutex);
- err = find_mgm(dev, gid, mailbox, &hash, &prev, &index);
+ err = find_mgm(dev, gid, prot, mailbox, &hash, &prev, &index);
if (err)
goto out;
@@ -316,7 +323,7 @@ int mlx4_multicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16])
goto out;
}
- members_count = be32_to_cpu(mgm->members_count);
+ members_count = be32_to_cpu(mgm->members_count) & 0xffffff;
for (loc = -1, i = 0; i < members_count; ++i)
if ((be32_to_cpu(mgm->qp[i]) & MGM_QPN_MASK) == qp->qpn)
loc = i;
@@ -328,7 +335,7 @@ int mlx4_multicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16])
}
- mgm->members_count = cpu_to_be32(--members_count);
+ mgm->members_count = cpu_to_be32(--members_count | ((u32) prot << 30));
mgm->qp[loc] = mgm->qp[i - 1];
mgm->qp[i - 1] = 0;
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 246b7bc..3d74198 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -171,6 +171,13 @@ enum mlx4_special_vlan_idx {
MLX4_VLAN_REGULAR
};
+enum mlx4_protocol {
+ MLX4_PROT_IB_IPV6 = 0,
+ MLX4_PROT_ETH,
+ MLX4_PROT_IB_IPV4,
+ MLX4_PROT_FCOE
+};
+
enum {
MLX4_NUM_FEXCH = 64 * 1024,
};
@@ -488,8 +495,9 @@ int mlx4_INIT_PORT(struct mlx4_dev *dev, int port);
int mlx4_CLOSE_PORT(struct mlx4_dev *dev, int port);
int mlx4_multicast_attach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
- int block_mcast_loopback);
-int mlx4_multicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16]);
+ int block_mcast_loopback, enum mlx4_protocol prot);
+int mlx4_multicast_detach(struct mlx4_dev *dev, struct mlx4_qp *qp, u8 gid[16],
+ enum mlx4_protocol prot);
int mlx4_register_mac(struct mlx4_dev *dev, u8 port, u64 mac, int *qpn);
void mlx4_unregister_mac(struct mlx4_dev *dev, u8 port, int qpn);
--
1.6.1.3
--
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
^ permalink raw reply related
* [PATCH 25/25] mlx4_en: Loopback enable
From: Yevgeny Petrilin @ 2009-11-04 15:32 UTC (permalink / raw)
To: rdreier-FYB4Gu1CFyUAvxtiuMwx3w
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA, netdev-u79uwXL29TY76Z2rM5mHXA,
liranl-VPRAkNaXOzVS1MOuV/RT9w, tziporet-VPRAkNaXOzVS1MOuV/RT9w,
yevgenyp-VPRAkNaXOzVS1MOuV/RT9w
All packets are subject to loopback.
To enable loopback we write the dest mac to the wqe.
Signed-off-by: Yevgeny Petrilin <yevgenyp-VPRAkNaXOzVS1MOuV/RT9w@public.gmane.org>
---
drivers/net/mlx4/en_netdev.c | 2 +-
drivers/net/mlx4/en_tx.c | 14 ++++++++++++++
drivers/net/mlx4/mlx4_en.h | 1 +
3 files changed, 16 insertions(+), 1 deletions(-)
diff --git a/drivers/net/mlx4/en_netdev.c b/drivers/net/mlx4/en_netdev.c
index 8406fbb..947fb2f 100644
--- a/drivers/net/mlx4/en_netdev.c
+++ b/drivers/net/mlx4/en_netdev.c
@@ -108,7 +108,7 @@ static void mlx4_en_vlan_rx_kill_vid(struct net_device *dev, unsigned short vid)
mutex_unlock(&mdev->state_lock);
}
-static u64 mlx4_en_mac_to_u64(u8 *addr)
+u64 mlx4_en_mac_to_u64(u8 *addr)
{
u64 mac = 0;
int i;
diff --git a/drivers/net/mlx4/en_tx.c b/drivers/net/mlx4/en_tx.c
index a893733..cfbe361 100644
--- a/drivers/net/mlx4/en_tx.c
+++ b/drivers/net/mlx4/en_tx.c
@@ -599,6 +599,7 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
struct mlx4_wqe_data_seg *data;
struct skb_frag_struct *frag;
struct mlx4_en_tx_info *tx_info;
+ struct ethhdr *ethh;
int tx_ind = 0;
int nr_txbb;
int desc_size;
@@ -610,6 +611,8 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
int i;
int lso_header_size;
void *fragptr;
+ u64 mac;
+ u32 mac_l, mac_h;
real_size = get_real_size(skb, dev, &lso_header_size);
if (unlikely(!real_size))
@@ -675,6 +678,17 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
priv->port_stats.tx_chksum_offload++;
}
+ /* Copy dst mac address to wqe */
+ skb_reset_mac_header(skb);
+ ethh = eth_hdr(skb);
+ if (ethh && ethh->h_dest) {
+ mac = mlx4_en_mac_to_u64(ethh->h_dest);
+ mac_h = (u32) ((mac & 0xffff00000000) >> 16);
+ mac_l = (u32) (mac & 0xffffffff);
+ tx_desc->ctrl.srcrb_flags |= cpu_to_be32(mac_h);
+ tx_desc->ctrl.imm = cpu_to_be32(mac_l);
+ }
+
/* Handle LSO (TSO) packets */
if (lso_header_size) {
/* Mark opcode as LSO */
diff --git a/drivers/net/mlx4/mlx4_en.h b/drivers/net/mlx4/mlx4_en.h
index b4731d5..02d07e9 100644
--- a/drivers/net/mlx4/mlx4_en.h
+++ b/drivers/net/mlx4/mlx4_en.h
@@ -564,6 +564,7 @@ int mlx4_SET_PORT_qpn_calc(struct mlx4_dev *dev, u8 port, u32 base_qpn,
int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset);
+u64 mlx4_en_mac_to_u64(u8 *addr);
/*
* Globals
*/
--
1.6.1.3
--
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
^ permalink raw reply related
* Re: [RFC,PATCH] mutex: mutex_is_owner() helper
From: Ingo Molnar @ 2009-11-04 15:40 UTC (permalink / raw)
To: Eric Dumazet, Peter Zijlstra, Linus Torvalds
Cc: David S. Miller, Linux Netdev List, linux kernel
In-Reply-To: <4AF19D06.3060401@gmail.com>
* Eric Dumazet <eric.dumazet@gmail.com> wrote:
> mutex_is_locked() is called most of the time to check if mutex is locked by current
> thread. But it's a lazy check, because mutex might be locked by another thread.
>
> Adds a new mutex_is_owned_by() helper, that can check ownership if CONFIG_SMP or
> CONFIG_DEBUG_MUTEXES are set.
>
> Returns are
> 0 if mutex is unlocked.
> 1 if locked
> -1 if not locked by designated thread.
>
> Last return value is possible only if CONFIG_SMP=y or CONFIG_DEBUG_MUTEXES=y
>
> Example of use :
>
> int rtnl_is_locked(void)
> {
> return mutex_is_locked(&rtnl_mutex);
> }
> ->
> int rtnl_is_locked(void)
> {
> return mutex_is_owned_by(&rtnl_mutex, current_thread_info()) == 1;
> }
To make sure this does not extend mutexes to be used a recursive
mutexes, mind naming it more clearly, like debug_mutex_is_owned(), and
adding a comment that says that this shouldnt be taken?
Also, it's somewhat imprecise: on !SMP && !DEBUG_MUTEXES we might return
a false '1'. Which happens to work for the rtnl usecase - but might not
in other cases.
Ingo
^ permalink raw reply
* [PATCH] ems_usb: Fix byte order issues on big endian machines
From: Sebastian Haas @ 2009-11-04 15:48 UTC (permalink / raw)
To: netdev-u79uwXL29TY76Z2rM5mHXA; +Cc: socketcan-core-0fE9KPoRgkgATYTw5x5z8w
CPC-USB is using a ARM7 core with little endian byte order. The "id" field
in can_msg needs byte order conversion from/to CPU byte order.
Signed-off-by: Sebastian Haas <haas-zsNKPWJ8Pib6hrUXjxyGrA@public.gmane.org>
---
drivers/net/can/usb/ems_usb.c | 5 ++++-
1 files changed, 4 insertions(+), 1 deletions(-)
diff --git a/drivers/net/can/usb/ems_usb.c b/drivers/net/can/usb/ems_usb.c
index 9012e0a..abdbd9c 100644
--- a/drivers/net/can/usb/ems_usb.c
+++ b/drivers/net/can/usb/ems_usb.c
@@ -319,7 +319,7 @@ static void ems_usb_rx_can_msg(struct ems_usb *dev, struct ems_cpc_msg *msg)
cf = (struct can_frame *)skb_put(skb, sizeof(struct can_frame));
- cf->can_id = msg->msg.can_msg.id;
+ cf->can_id = le32_to_cpu(msg->msg.can_msg.id);
cf->can_dlc = min_t(u8, msg->msg.can_msg.length, 8);
if (msg->type == CPC_MSG_TYPE_EXT_CAN_FRAME
@@ -813,6 +813,9 @@ static netdev_tx_t ems_usb_start_xmit(struct sk_buff *skb, struct net_device *ne
msg->length = CPC_CAN_MSG_MIN_SIZE + cf->can_dlc;
}
+ /* Respect byte order */
+ msg->msg.can_msg.id = cpu_to_le32(msg->msg.can_msg.id);
+
for (i = 0; i < MAX_TX_URBS; i++) {
if (dev->tx_contexts[i].echo_index == MAX_TX_URBS) {
context = &dev->tx_contexts[i];
--
EMS Dr. Thomas Wuensche e.K.
Sonnenhang 3
85304 Ilmmuenster
HRA Neuburg a.d. Donau, HR-Nr. 70.106
Phone: +49-8441-490260
Fax : +49-8441-81860
http://www.ems-wuensche.com
^ permalink raw reply related
* [PATCH] ems_usb: Fix byte order issues on big endian machines
From: Sebastian Haas @ 2009-11-04 15:48 UTC (permalink / raw)
To: netdev; +Cc: socketcan-core
CPC-USB is using a ARM7 core with little endian byte order. The "id" field
in can_msg needs byte order conversion from/to CPU byte order.
Signed-off-by: Sebastian Haas <haas@ems-wuensche.com>
---
drivers/net/can/usb/ems_usb.c | 5 ++++-
1 files changed, 4 insertions(+), 1 deletions(-)
diff --git a/drivers/net/can/usb/ems_usb.c b/drivers/net/can/usb/ems_usb.c
index 9012e0a..abdbd9c 100644
--- a/drivers/net/can/usb/ems_usb.c
+++ b/drivers/net/can/usb/ems_usb.c
@@ -319,7 +319,7 @@ static void ems_usb_rx_can_msg(struct ems_usb *dev, struct ems_cpc_msg *msg)
cf = (struct can_frame *)skb_put(skb, sizeof(struct can_frame));
- cf->can_id = msg->msg.can_msg.id;
+ cf->can_id = le32_to_cpu(msg->msg.can_msg.id);
cf->can_dlc = min_t(u8, msg->msg.can_msg.length, 8);
if (msg->type == CPC_MSG_TYPE_EXT_CAN_FRAME
@@ -813,6 +813,9 @@ static netdev_tx_t ems_usb_start_xmit(struct sk_buff *skb, struct net_device *ne
msg->length = CPC_CAN_MSG_MIN_SIZE + cf->can_dlc;
}
+ /* Respect byte order */
+ msg->msg.can_msg.id = cpu_to_le32(msg->msg.can_msg.id);
+
for (i = 0; i < MAX_TX_URBS; i++) {
if (dev->tx_contexts[i].echo_index == MAX_TX_URBS) {
context = &dev->tx_contexts[i];
--
EMS Dr. Thomas Wuensche e.K.
Sonnenhang 3
85304 Ilmmuenster
HRA Neuburg a.d. Donau, HR-Nr. 70.106
Phone: +49-8441-490260
Fax : +49-8441-81860
http://www.ems-wuensche.com
^ permalink raw reply related
* [PATCHv8 0/3] vhost: a kernel-level virtio server
From: Michael S. Tsirkin @ 2009-11-04 15:52 UTC (permalink / raw)
To: netdev, virtualization, kvm, linux-kernel, mingo, linux-mm, akpm
Ok, I think I've addressed all comments so far here.
Rusty, I'd like this to go into linux-next, through your tree, and
hopefully 2.6.33. What do you think?
---
This implements vhost: a kernel-level backend for virtio,
The main motivation for this work is to reduce virtualization
overhead for virtio by removing system calls on data path,
without guest changes. For virtio-net, this removes up to
4 system calls per packet: vm exit for kick, reentry for kick,
iothread wakeup for packet, interrupt injection for packet.
This driver is pretty minimal, but it's fully functional (including
migration support interfaces), and already shows performance (especially
latency) improvement over userspace.
Some more detailed description attached to the patch itself.
The patches apply to both 2.6.32-rc6 and kvm.git. I'd like them to go
into linux-next if possible. Please comment.
Changelog from v7:
- Add note on RCU usage, mirroring this in vhost/vhost.h
- Fix locking typo noted by Eric Dumazet
- Fix warnings on 32 bit
Changelog from v6:
- review comments by Daniel Walker addressed
- checkpatch cleanup
- fix build on 32 bit
- maintainers entry corrected
Changelog from v5:
- tun support
- backends with virtio net header support (enables GSO, checksum etc)
- 32 bit compat fixed
- support indirect buffers, tx exit mitigation,
tx interrupt mitigation
- support write logging (allows migration without virtio ring code in userspace)
Changelog from v4:
- disable rx notification when have rx buffers
- addressed all comments from Rusty's review
- copy bugfixes from lguest commits:
ebf9a5a99c1a464afe0b4dfa64416fc8b273bc5c
e606490c440900e50ccf73a54f6fc6150ff40815
Changelog from v3:
- checkpatch fixes
Changelog from v2:
- Comments on RCU usage
- Compat ioctl support
- Make variable static
- Copied more idiomatic english from Rusty
Changes from v1:
- Move use_mm/unuse_mm from fs/aio.c to mm instead of copying.
- Reorder code to avoid need for forward declarations
- Kill a couple of debugging printks
Michael S. Tsirkin (3):
tun: export underlying socket
mm: export use_mm/unuse_mm to modules
vhost_net: a kernel-level virtio server
MAINTAINERS | 9 +
arch/x86/kvm/Kconfig | 1 +
drivers/Makefile | 1 +
drivers/net/tun.c | 101 ++++-
drivers/vhost/Kconfig | 11 +
drivers/vhost/Makefile | 2 +
drivers/vhost/net.c | 633 +++++++++++++++++++++++++++++
drivers/vhost/vhost.c | 970 ++++++++++++++++++++++++++++++++++++++++++++
drivers/vhost/vhost.h | 158 +++++++
include/linux/Kbuild | 1 +
include/linux/if_tun.h | 14 +
include/linux/miscdevice.h | 1 +
include/linux/vhost.h | 126 ++++++
mm/mmu_context.c | 3 +
14 files changed, 2012 insertions(+), 19 deletions(-)
create mode 100644 drivers/vhost/Kconfig
create mode 100644 drivers/vhost/Makefile
create mode 100644 drivers/vhost/net.c
create mode 100644 drivers/vhost/vhost.c
create mode 100644 drivers/vhost/vhost.h
create mode 100644 include/linux/vhost.h
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply
* [PATCHv8 1/3] tun: export underlying socket
From: Michael S. Tsirkin @ 2009-11-04 15:55 UTC (permalink / raw)
To: netdev, virtualization, kvm, linux-kernel, mingo, linux-mm, akpm
In-Reply-To: <cover.1257349249.git.mst@redhat.com>
Tun device looks similar to a packet socket
in that both pass complete frames from/to userspace.
This patch fills in enough fields in the socket underlying tun driver
to support sendmsg/recvmsg operations, and message flags
MSG_TRUNC and MSG_DONTWAIT, and exports access to this socket
to modules. Regular read/write behaviour is unchanged.
This way, code using raw sockets to inject packets
into a physical device, can support injecting
packets into host network stack almost without modification.
First user of this interface will be vhost virtualization
accelerator.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: David S. Miller <davem@davemloft.net>
---
DaveM agreed that this will go in through Rusty's tree together
with vhost patch, because vhost is the first user of the new
interface.
drivers/net/tun.c | 101 +++++++++++++++++++++++++++++++++++++++---------
include/linux/if_tun.h | 14 +++++++
2 files changed, 96 insertions(+), 19 deletions(-)
diff --git a/drivers/net/tun.c b/drivers/net/tun.c
index 4fdfa2a..18f8876 100644
--- a/drivers/net/tun.c
+++ b/drivers/net/tun.c
@@ -144,6 +144,7 @@ static int tun_attach(struct tun_struct *tun, struct file *file)
err = 0;
tfile->tun = tun;
tun->tfile = tfile;
+ tun->socket.file = file;
dev_hold(tun->dev);
sock_hold(tun->socket.sk);
atomic_inc(&tfile->count);
@@ -158,6 +159,7 @@ static void __tun_detach(struct tun_struct *tun)
/* Detach from net device */
netif_tx_lock_bh(tun->dev);
tun->tfile = NULL;
+ tun->socket.file = NULL;
netif_tx_unlock_bh(tun->dev);
/* Drop read queue */
@@ -387,7 +389,8 @@ static netdev_tx_t tun_net_xmit(struct sk_buff *skb, struct net_device *dev)
/* Notify and wake up reader process */
if (tun->flags & TUN_FASYNC)
kill_fasync(&tun->fasync, SIGIO, POLL_IN);
- wake_up_interruptible(&tun->socket.wait);
+ wake_up_interruptible_poll(&tun->socket.wait, POLLIN |
+ POLLRDNORM | POLLRDBAND);
return NETDEV_TX_OK;
drop:
@@ -743,7 +746,7 @@ static __inline__ ssize_t tun_put_user(struct tun_struct *tun,
len = min_t(int, skb->len, len);
skb_copy_datagram_const_iovec(skb, 0, iv, total, len);
- total += len;
+ total += skb->len;
tun->dev->stats.tx_packets++;
tun->dev->stats.tx_bytes += len;
@@ -751,34 +754,23 @@ static __inline__ ssize_t tun_put_user(struct tun_struct *tun,
return total;
}
-static ssize_t tun_chr_aio_read(struct kiocb *iocb, const struct iovec *iv,
- unsigned long count, loff_t pos)
+static ssize_t tun_do_read(struct tun_struct *tun,
+ struct kiocb *iocb, const struct iovec *iv,
+ ssize_t len, int noblock)
{
- struct file *file = iocb->ki_filp;
- struct tun_file *tfile = file->private_data;
- struct tun_struct *tun = __tun_get(tfile);
DECLARE_WAITQUEUE(wait, current);
struct sk_buff *skb;
- ssize_t len, ret = 0;
-
- if (!tun)
- return -EBADFD;
+ ssize_t ret = 0;
DBG(KERN_INFO "%s: tun_chr_read\n", tun->dev->name);
- len = iov_length(iv, count);
- if (len < 0) {
- ret = -EINVAL;
- goto out;
- }
-
add_wait_queue(&tun->socket.wait, &wait);
while (len) {
current->state = TASK_INTERRUPTIBLE;
/* Read frames from the queue */
if (!(skb=skb_dequeue(&tun->socket.sk->sk_receive_queue))) {
- if (file->f_flags & O_NONBLOCK) {
+ if (noblock) {
ret = -EAGAIN;
break;
}
@@ -805,6 +797,27 @@ static ssize_t tun_chr_aio_read(struct kiocb *iocb, const struct iovec *iv,
current->state = TASK_RUNNING;
remove_wait_queue(&tun->socket.wait, &wait);
+ return ret;
+}
+
+static ssize_t tun_chr_aio_read(struct kiocb *iocb, const struct iovec *iv,
+ unsigned long count, loff_t pos)
+{
+ struct file *file = iocb->ki_filp;
+ struct tun_file *tfile = file->private_data;
+ struct tun_struct *tun = __tun_get(tfile);
+ ssize_t len, ret;
+
+ if (!tun)
+ return -EBADFD;
+ len = iov_length(iv, count);
+ if (len < 0) {
+ ret = -EINVAL;
+ goto out;
+ }
+
+ ret = tun_do_read(tun, iocb, iv, len, file->f_flags & O_NONBLOCK);
+ ret = min_t(ssize_t, ret, len);
out:
tun_put(tun);
return ret;
@@ -847,7 +860,8 @@ static void tun_sock_write_space(struct sock *sk)
return;
if (sk->sk_sleep && waitqueue_active(sk->sk_sleep))
- wake_up_interruptible_sync(sk->sk_sleep);
+ wake_up_interruptible_sync_poll(sk->sk_sleep, POLLOUT |
+ POLLWRNORM | POLLWRBAND);
tun = container_of(sk, struct tun_sock, sk)->tun;
kill_fasync(&tun->fasync, SIGIO, POLL_OUT);
@@ -858,6 +872,37 @@ static void tun_sock_destruct(struct sock *sk)
free_netdev(container_of(sk, struct tun_sock, sk)->tun->dev);
}
+static int tun_sendmsg(struct kiocb *iocb, struct socket *sock,
+ struct msghdr *m, size_t total_len)
+{
+ struct tun_struct *tun = container_of(sock, struct tun_struct, socket);
+ return tun_get_user(tun, m->msg_iov, total_len,
+ m->msg_flags & MSG_DONTWAIT);
+}
+
+static int tun_recvmsg(struct kiocb *iocb, struct socket *sock,
+ struct msghdr *m, size_t total_len,
+ int flags)
+{
+ struct tun_struct *tun = container_of(sock, struct tun_struct, socket);
+ int ret;
+ if (flags & ~(MSG_DONTWAIT|MSG_TRUNC))
+ return -EINVAL;
+ ret = tun_do_read(tun, iocb, m->msg_iov, total_len,
+ flags & MSG_DONTWAIT);
+ if (ret > total_len) {
+ m->msg_flags |= MSG_TRUNC;
+ ret = flags & MSG_TRUNC ? ret : total_len;
+ }
+ return ret;
+}
+
+/* Ops structure to mimic raw sockets with tun */
+static const struct proto_ops tun_socket_ops = {
+ .sendmsg = tun_sendmsg,
+ .recvmsg = tun_recvmsg,
+};
+
static struct proto tun_proto = {
.name = "tun",
.owner = THIS_MODULE,
@@ -986,6 +1031,7 @@ static int tun_set_iff(struct net *net, struct file *file, struct ifreq *ifr)
goto err_free_dev;
init_waitqueue_head(&tun->socket.wait);
+ tun->socket.ops = &tun_socket_ops;
sock_init_data(&tun->socket, sk);
sk->sk_write_space = tun_sock_write_space;
sk->sk_sndbuf = INT_MAX;
@@ -1489,6 +1535,23 @@ static void tun_cleanup(void)
rtnl_link_unregister(&tun_link_ops);
}
+/* Get an underlying socket object from tun file. Returns error unless file is
+ * attached to a device. The returned object works like a packet socket, it
+ * can be used for sock_sendmsg/sock_recvmsg. The caller is responsible for
+ * holding a reference to the file for as long as the socket is in use. */
+struct socket *tun_get_socket(struct file *file)
+{
+ struct tun_struct *tun;
+ if (file->f_op != &tun_fops)
+ return ERR_PTR(-EINVAL);
+ tun = tun_get(file);
+ if (!tun)
+ return ERR_PTR(-EBADFD);
+ tun_put(tun);
+ return &tun->socket;
+}
+EXPORT_SYMBOL_GPL(tun_get_socket);
+
module_init(tun_init);
module_exit(tun_cleanup);
MODULE_DESCRIPTION(DRV_DESCRIPTION);
diff --git a/include/linux/if_tun.h b/include/linux/if_tun.h
index 3f5fd52..404abe0 100644
--- a/include/linux/if_tun.h
+++ b/include/linux/if_tun.h
@@ -86,4 +86,18 @@ struct tun_filter {
__u8 addr[0][ETH_ALEN];
};
+#ifdef __KERNEL__
+#if defined(CONFIG_TUN) || defined(CONFIG_TUN_MODULE)
+struct socket *tun_get_socket(struct file *);
+#else
+#include <linux/err.h>
+#include <linux/errno.h>
+struct file;
+struct socket;
+static inline struct socket *tun_get_socket(struct file *f)
+{
+ return ERR_PTR(-EINVAL);
+}
+#endif /* CONFIG_TUN */
+#endif /* __KERNEL__ */
#endif /* __IF_TUN_H */
--
1.6.5.2.143.g8cc62
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply related
* [PATCHv8 2/3] mm: export use_mm/unuse_mm to modules
From: Michael S. Tsirkin @ 2009-11-04 15:55 UTC (permalink / raw)
To: netdev, virtualization, kvm, linux-kernel, mingo, linux-mm, akpm
In-Reply-To: <cover.1257349249.git.mst@redhat.com>
vhost net module wants to do copy to/from user from a kernel thread,
which needs use_mm. Export it to modules.
Acked-by: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
mm/mmu_context.c | 3 +++
1 files changed, 3 insertions(+), 0 deletions(-)
diff --git a/mm/mmu_context.c b/mm/mmu_context.c
index ded9081..0777654 100644
--- a/mm/mmu_context.c
+++ b/mm/mmu_context.c
@@ -5,6 +5,7 @@
#include <linux/mm.h>
#include <linux/mmu_context.h>
+#include <linux/module.h>
#include <linux/sched.h>
#include <asm/mmu_context.h>
@@ -37,6 +38,7 @@ void use_mm(struct mm_struct *mm)
if (active_mm != mm)
mmdrop(active_mm);
}
+EXPORT_SYMBOL_GPL(use_mm);
/*
* unuse_mm
@@ -56,3 +58,4 @@ void unuse_mm(struct mm_struct *mm)
enter_lazy_tlb(mm, tsk);
task_unlock(tsk);
}
+EXPORT_SYMBOL_GPL(unuse_mm);
--
1.6.5.2.143.g8cc62
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply related
* [PATCHv8 3/3] vhost_net: a kernel-level virtio server
From: Michael S. Tsirkin @ 2009-11-04 15:57 UTC (permalink / raw)
To: netdev, virtualization, kvm, linux-kernel, mingo, linux-mm, akpm
In-Reply-To: <cover.1257349249.git.mst@redhat.com>
What it is: vhost net is a character device that can be used to reduce
the number of system calls involved in virtio networking.
Existing virtio net code is used in the guest without modification.
There's similarity with vringfd, with some differences and reduced scope
- uses eventfd for signalling
- structures can be moved around in memory at any time (good for
migration, bug work-arounds in userspace)
- write logging is supported (good for migration)
- support memory table and not just an offset (needed for kvm)
common virtio related code has been put in a separate file vhost.c and
can be made into a separate module if/when more backends appear. I used
Rusty's lguest.c as the source for developing this part : this supplied
me with witty comments I wouldn't be able to write myself.
What it is not: vhost net is not a bus, and not a generic new system
call. No assumptions are made on how guest performs hypercalls.
Userspace hypervisors are supported as well as kvm.
How it works: Basically, we connect virtio frontend (configured by
userspace) to a backend. The backend could be a network device, or a tap
device. Backend is also configured by userspace, including vlan/mac
etc.
Status: This works for me, and I haven't see any crashes.
Compared to userspace, people reported improved latency (as I save up to
4 system calls per packet), as well as better bandwidth and CPU
utilization.
Features that I plan to look at in the future:
- mergeable buffers
- zero copy
- scalability tuning: figure out the best threading model to use
Note on RCU usage (this is also documented in vhost.h, near
private_pointer which is the value protected by this variant of RCU):
what is happening is that the rcu_dereference() is being used
in a workqueue item. The role of rcu_read_lock() is taken on by the
start of execution of the workqueue item, of rcu_read_unlock() by the
end of execution of the workqueue item, and of synchronize_rcu() by
flush_workqueue()/flush_work().
Acked-by: Arnd Bergmann <arnd@arndb.de>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
Paul, I'm Cc you so that you can verify that nothing significant changed
in vhost/net.c since you approved it the previous time. Could you
please Ack use of rcu_dereference/rcu_assign_pointer there? Eric
Dumazet proposed that I open-code these, he suggested that this is an
abuse of these macros and not enough like RCU. What is your opinion?
MAINTAINERS | 9 +
arch/x86/kvm/Kconfig | 1 +
drivers/Makefile | 1 +
drivers/vhost/Kconfig | 11 +
drivers/vhost/Makefile | 2 +
drivers/vhost/net.c | 633 +++++++++++++++++++++++++++++
drivers/vhost/vhost.c | 970 ++++++++++++++++++++++++++++++++++++++++++++
drivers/vhost/vhost.h | 158 +++++++
include/linux/Kbuild | 1 +
include/linux/miscdevice.h | 1 +
include/linux/vhost.h | 126 ++++++
11 files changed, 1913 insertions(+), 0 deletions(-)
create mode 100644 drivers/vhost/Kconfig
create mode 100644 drivers/vhost/Makefile
create mode 100644 drivers/vhost/net.c
create mode 100644 drivers/vhost/vhost.c
create mode 100644 drivers/vhost/vhost.h
create mode 100644 include/linux/vhost.h
diff --git a/MAINTAINERS b/MAINTAINERS
index 8824115..980a69b 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -5619,6 +5619,15 @@ S: Maintained
F: Documentation/filesystems/vfat.txt
F: fs/fat/
+VIRTIO HOST (VHOST)
+M: "Michael S. Tsirkin" <mst@redhat.com>
+L: kvm@vger.kernel.org
+L: virtualization@lists.osdl.org
+L: netdev@vger.kernel.org
+S: Maintained
+F: drivers/vhost/
+F: include/linux/vhost.h
+
VIA RHINE NETWORK DRIVER
M: Roger Luethi <rl@hellgate.ch>
S: Maintained
diff --git a/arch/x86/kvm/Kconfig b/arch/x86/kvm/Kconfig
index b84e571..94f44d9 100644
--- a/arch/x86/kvm/Kconfig
+++ b/arch/x86/kvm/Kconfig
@@ -64,6 +64,7 @@ config KVM_AMD
# OK, it's a little counter-intuitive to do this, but it puts it neatly under
# the virtualization menu.
+source drivers/vhost/Kconfig
source drivers/lguest/Kconfig
source drivers/virtio/Kconfig
diff --git a/drivers/Makefile b/drivers/Makefile
index 6ee53c7..81e3659 100644
--- a/drivers/Makefile
+++ b/drivers/Makefile
@@ -106,6 +106,7 @@ obj-$(CONFIG_HID) += hid/
obj-$(CONFIG_PPC_PS3) += ps3/
obj-$(CONFIG_OF) += of/
obj-$(CONFIG_SSB) += ssb/
+obj-$(CONFIG_VHOST_NET) += vhost/
obj-$(CONFIG_VIRTIO) += virtio/
obj-$(CONFIG_VLYNQ) += vlynq/
obj-$(CONFIG_STAGING) += staging/
diff --git a/drivers/vhost/Kconfig b/drivers/vhost/Kconfig
new file mode 100644
index 0000000..9f409f4
--- /dev/null
+++ b/drivers/vhost/Kconfig
@@ -0,0 +1,11 @@
+config VHOST_NET
+ tristate "Host kernel accelerator for virtio net (EXPERIMENTAL)"
+ depends on NET && EVENTFD && EXPERIMENTAL
+ ---help---
+ This kernel module can be loaded in host kernel to accelerate
+ guest networking with virtio_net. Not to be confused with virtio_net
+ module itself which needs to be loaded in guest kernel.
+
+ To compile this driver as a module, choose M here: the module will
+ be called vhost_net.
+
diff --git a/drivers/vhost/Makefile b/drivers/vhost/Makefile
new file mode 100644
index 0000000..72dd020
--- /dev/null
+++ b/drivers/vhost/Makefile
@@ -0,0 +1,2 @@
+obj-$(CONFIG_VHOST_NET) += vhost_net.o
+vhost_net-y := vhost.o net.o
diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
new file mode 100644
index 0000000..aeafaf3
--- /dev/null
+++ b/drivers/vhost/net.c
@@ -0,0 +1,633 @@
+/* Copyright (C) 2009 Red Hat, Inc.
+ * Author: Michael S. Tsirkin <mst@redhat.com>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2.
+ *
+ * virtio-net server in host kernel.
+ */
+
+#include <linux/compat.h>
+#include <linux/eventfd.h>
+#include <linux/vhost.h>
+#include <linux/virtio_net.h>
+#include <linux/mmu_context.h>
+#include <linux/miscdevice.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <linux/workqueue.h>
+#include <linux/rcupdate.h>
+#include <linux/file.h>
+
+#include <linux/net.h>
+#include <linux/if_packet.h>
+#include <linux/if_arp.h>
+#include <linux/if_tun.h>
+
+#include <net/sock.h>
+
+#include "vhost.h"
+
+/* Max number of bytes transferred before requeueing the job.
+ * Using this limit prevents one virtqueue from starving others. */
+#define VHOST_NET_WEIGHT 0x80000
+
+enum {
+ VHOST_NET_VQ_RX = 0,
+ VHOST_NET_VQ_TX = 1,
+ VHOST_NET_VQ_MAX = 2,
+};
+
+enum vhost_net_poll_state {
+ VHOST_NET_POLL_DISABLED = 0,
+ VHOST_NET_POLL_STARTED = 1,
+ VHOST_NET_POLL_STOPPED = 2,
+};
+
+struct vhost_net {
+ struct vhost_dev dev;
+ struct vhost_virtqueue vqs[VHOST_NET_VQ_MAX];
+ struct vhost_poll poll[VHOST_NET_VQ_MAX];
+ /* Tells us whether we are polling a socket for TX.
+ * We only do this when socket buffer fills up.
+ * Protected by tx vq lock. */
+ enum vhost_net_poll_state tx_poll_state;
+};
+
+/* Pop first len bytes from iovec. Return number of segments used. */
+static int move_iovec_hdr(struct iovec *from, struct iovec *to,
+ size_t len, int iov_count)
+{
+ int seg = 0;
+ size_t size;
+ while (len && seg < iov_count) {
+ size = min(from->iov_len, len);
+ to->iov_base = from->iov_base;
+ to->iov_len = size;
+ from->iov_len -= size;
+ from->iov_base += size;
+ len -= size;
+ ++from;
+ ++to;
+ ++seg;
+ }
+ return seg;
+}
+
+/* Caller must have TX VQ lock */
+static void tx_poll_stop(struct vhost_net *net)
+{
+ if (likely(net->tx_poll_state != VHOST_NET_POLL_STARTED))
+ return;
+ vhost_poll_stop(net->poll + VHOST_NET_VQ_TX);
+ net->tx_poll_state = VHOST_NET_POLL_STOPPED;
+}
+
+/* Caller must have TX VQ lock */
+static void tx_poll_start(struct vhost_net *net, struct socket *sock)
+{
+ if (unlikely(net->tx_poll_state != VHOST_NET_POLL_STOPPED))
+ return;
+ vhost_poll_start(net->poll + VHOST_NET_VQ_TX, sock->file);
+ net->tx_poll_state = VHOST_NET_POLL_STARTED;
+}
+
+/* Expects to be always run from workqueue - which acts as
+ * read-size critical section for our kind of RCU. */
+static void handle_tx(struct vhost_net *net)
+{
+ struct vhost_virtqueue *vq = &net->dev.vqs[VHOST_NET_VQ_TX];
+ unsigned head, out, in, s;
+ struct msghdr msg = {
+ .msg_name = NULL,
+ .msg_namelen = 0,
+ .msg_control = NULL,
+ .msg_controllen = 0,
+ .msg_iov = vq->iov,
+ .msg_flags = MSG_DONTWAIT,
+ };
+ size_t len, total_len = 0;
+ int err, wmem;
+ size_t hdr_size;
+ struct socket *sock = rcu_dereference(vq->private_data);
+ if (!sock)
+ return;
+
+ wmem = atomic_read(&sock->sk->sk_wmem_alloc);
+ if (wmem >= sock->sk->sk_sndbuf)
+ return;
+
+ use_mm(net->dev.mm);
+ mutex_lock(&vq->mutex);
+ vhost_no_notify(vq);
+
+ if (wmem < sock->sk->sk_sndbuf * 2)
+ tx_poll_stop(net);
+ hdr_size = vq->hdr_size;
+
+ for (;;) {
+ head = vhost_get_vq_desc(&net->dev, vq, vq->iov, &out, &in,
+ NULL, NULL);
+ /* Nothing new? Wait for eventfd to tell us they refilled. */
+ if (head == vq->num) {
+ wmem = atomic_read(&sock->sk->sk_wmem_alloc);
+ if (wmem >= sock->sk->sk_sndbuf * 3 / 4) {
+ tx_poll_start(net, sock);
+ set_bit(SOCK_ASYNC_NOSPACE, &sock->flags);
+ break;
+ }
+ if (vhost_notify(vq))
+ continue;
+ break;
+ }
+ if (in) {
+ vq_err(vq, "Unexpected descriptor format for TX: "
+ "out %d, int %d\n", out, in);
+ break;
+ }
+ /* Skip header. TODO: support TSO. */
+ s = move_iovec_hdr(vq->iov, vq->hdr, hdr_size, out);
+ msg.msg_iovlen = out;
+ len = iov_length(vq->iov, out);
+ /* Sanity check */
+ if (!len) {
+ vq_err(vq, "Unexpected header len for TX: "
+ "%zd expected %zd\n",
+ iov_length(vq->hdr, s), hdr_size);
+ break;
+ }
+ /* TODO: Check specific error and bomb out unless ENOBUFS? */
+ err = sock->ops->sendmsg(NULL, sock, &msg, len);
+ if (unlikely(err < 0)) {
+ vhost_discard_vq_desc(vq);
+ tx_poll_start(net, sock);
+ break;
+ }
+ if (err != len)
+ pr_err("Truncated TX packet: "
+ " len %d != %zd\n", err, len);
+ vhost_add_used_and_trigger(&net->dev, vq, head, 0);
+ total_len += len;
+ if (unlikely(total_len >= VHOST_NET_WEIGHT)) {
+ vhost_poll_queue(&vq->poll);
+ break;
+ }
+ }
+
+ mutex_unlock(&vq->mutex);
+ unuse_mm(net->dev.mm);
+}
+
+/* Expects to be always run from workqueue - which acts as
+ * read-size critical section for our kind of RCU. */
+static void handle_rx(struct vhost_net *net)
+{
+ struct vhost_virtqueue *vq = &net->dev.vqs[VHOST_NET_VQ_RX];
+ unsigned head, out, in, log, s;
+ struct vhost_log *vq_log;
+ struct msghdr msg = {
+ .msg_name = NULL,
+ .msg_namelen = 0,
+ .msg_control = NULL, /* FIXME: get and handle RX aux data. */
+ .msg_controllen = 0,
+ .msg_iov = vq->iov,
+ .msg_flags = MSG_DONTWAIT,
+ };
+
+ struct virtio_net_hdr hdr = {
+ .flags = 0,
+ .gso_type = VIRTIO_NET_HDR_GSO_NONE
+ };
+
+ size_t len, total_len = 0;
+ int err;
+ size_t hdr_size;
+ struct socket *sock = rcu_dereference(vq->private_data);
+ if (!sock || skb_queue_empty(&sock->sk->sk_receive_queue))
+ return;
+
+ use_mm(net->dev.mm);
+ mutex_lock(&vq->mutex);
+ vhost_no_notify(vq);
+ hdr_size = vq->hdr_size;
+
+ vq_log = unlikely(vhost_has_feature(&net->dev, VHOST_F_LOG_ALL)) ?
+ vq->log : NULL;
+
+ for (;;) {
+ head = vhost_get_vq_desc(&net->dev, vq, vq->iov, &out, &in,
+ vq_log, &log);
+ /* OK, now we need to know about added descriptors. */
+ if (head == vq->num && vhost_notify(vq))
+ /* They could have slipped one in as we were doing that:
+ * check again. */
+ continue;
+ /* Nothing new? Wait for eventfd to tell us they refilled. */
+ if (head == vq->num)
+ break;
+ /* We don't need to be notified again. */
+ vhost_no_notify(vq);
+ if (out) {
+ vq_err(vq, "Unexpected descriptor format for RX: "
+ "out %d, int %d\n",
+ out, in);
+ break;
+ }
+ /* Skip header. TODO: support TSO/mergeable rx buffers. */
+ s = move_iovec_hdr(vq->iov, vq->hdr, hdr_size, in);
+ msg.msg_iovlen = in;
+ len = iov_length(vq->iov, in);
+ /* Sanity check */
+ if (!len) {
+ vq_err(vq, "Unexpected header len for RX: "
+ "%zd expected %zd\n",
+ iov_length(vq->hdr, s), hdr_size);
+ break;
+ }
+ err = sock->ops->recvmsg(NULL, sock, &msg,
+ len, MSG_DONTWAIT | MSG_TRUNC);
+ /* TODO: Check specific error and bomb out unless EAGAIN? */
+ if (err < 0) {
+ vhost_discard_vq_desc(vq);
+ break;
+ }
+ /* TODO: Should check and handle checksum. */
+ if (err > len) {
+ pr_err("Discarded truncated rx packet: "
+ " len %d > %zd\n", err, len);
+ vhost_discard_vq_desc(vq);
+ continue;
+ }
+ len = err;
+ err = memcpy_toiovec(vq->hdr, (unsigned char *)&hdr, hdr_size);
+ if (err) {
+ vq_err(vq, "Unable to write vnet_hdr at addr %p: %d\n",
+ vq->iov->iov_base, err);
+ break;
+ }
+ len += hdr_size;
+ vhost_add_used_and_trigger(&net->dev, vq, head, len);
+ if (unlikely(vq_log))
+ vhost_log_write(vq, vq_log, log, len);
+ total_len += len;
+ if (unlikely(total_len >= VHOST_NET_WEIGHT)) {
+ vhost_poll_queue(&vq->poll);
+ break;
+ }
+ }
+
+ mutex_unlock(&vq->mutex);
+ unuse_mm(net->dev.mm);
+}
+
+static void handle_tx_kick(struct work_struct *work)
+{
+ struct vhost_virtqueue *vq;
+ struct vhost_net *net;
+ vq = container_of(work, struct vhost_virtqueue, poll.work);
+ net = container_of(vq->dev, struct vhost_net, dev);
+ handle_tx(net);
+}
+
+static void handle_rx_kick(struct work_struct *work)
+{
+ struct vhost_virtqueue *vq;
+ struct vhost_net *net;
+ vq = container_of(work, struct vhost_virtqueue, poll.work);
+ net = container_of(vq->dev, struct vhost_net, dev);
+ handle_rx(net);
+}
+
+static void handle_tx_net(struct work_struct *work)
+{
+ struct vhost_net *net;
+ net = container_of(work, struct vhost_net, poll[VHOST_NET_VQ_TX].work);
+ handle_tx(net);
+}
+
+static void handle_rx_net(struct work_struct *work)
+{
+ struct vhost_net *net;
+ net = container_of(work, struct vhost_net, poll[VHOST_NET_VQ_RX].work);
+ handle_rx(net);
+}
+
+static int vhost_net_open(struct inode *inode, struct file *f)
+{
+ struct vhost_net *n = kzalloc(sizeof *n, GFP_KERNEL);
+ int r;
+ if (!n)
+ return -ENOMEM;
+ f->private_data = n;
+ n->vqs[VHOST_NET_VQ_TX].handle_kick = handle_tx_kick;
+ n->vqs[VHOST_NET_VQ_RX].handle_kick = handle_rx_kick;
+ r = vhost_dev_init(&n->dev, n->vqs, VHOST_NET_VQ_MAX);
+ if (r < 0) {
+ kfree(n);
+ return r;
+ }
+
+ vhost_poll_init(n->poll + VHOST_NET_VQ_TX, handle_tx_net, POLLOUT);
+ vhost_poll_init(n->poll + VHOST_NET_VQ_RX, handle_rx_net, POLLIN);
+ n->tx_poll_state = VHOST_NET_POLL_DISABLED;
+ return 0;
+}
+
+static void vhost_net_disable_vq(struct vhost_net *n, int index)
+{
+ if (!n->vqs[index].private_data)
+ return;
+ if (index == VHOST_NET_VQ_TX) {
+ tx_poll_stop(n);
+ n->tx_poll_state = VHOST_NET_POLL_DISABLED;
+ } else
+ vhost_poll_stop(n->poll + VHOST_NET_VQ_RX);
+}
+
+static void vhost_net_enable_vq(struct vhost_net *n, int index)
+{
+ struct socket *sock = n->vqs[index].private_data;
+ if (!sock)
+ return;
+ if (index == VHOST_NET_VQ_TX) {
+ n->tx_poll_state = VHOST_NET_POLL_STOPPED;
+ tx_poll_start(n, sock);
+ } else
+ vhost_poll_start(n->poll + VHOST_NET_VQ_RX, sock->file);
+}
+
+static struct socket *vhost_net_stop_vq(struct vhost_net *n, int index)
+{
+ struct socket *sock;
+
+ mutex_lock(&n->vqs[index].mutex);
+ sock = n->vqs[index].private_data;
+ vhost_net_disable_vq(n, index);
+ rcu_assign_pointer(n->vqs[index].private_data, NULL);
+ mutex_unlock(&n->vqs[index].mutex);
+ return sock;
+}
+
+static void vhost_net_stop(struct vhost_net *n, struct socket **tx_sock,
+ struct socket **rx_sock)
+{
+ *tx_sock = vhost_net_stop_vq(n, VHOST_NET_VQ_TX);
+ *rx_sock = vhost_net_stop_vq(n, VHOST_NET_VQ_RX);
+}
+
+static void vhost_net_flush_vq(struct vhost_net *n, int index)
+{
+ vhost_poll_flush(n->poll + index);
+ vhost_poll_flush(&n->dev.vqs[index].poll);
+}
+
+static void vhost_net_flush(struct vhost_net *n)
+{
+ vhost_net_flush_vq(n, VHOST_NET_VQ_TX);
+ vhost_net_flush_vq(n, VHOST_NET_VQ_RX);
+}
+
+static int vhost_net_release(struct inode *inode, struct file *f)
+{
+ struct vhost_net *n = f->private_data;
+ struct socket *tx_sock;
+ struct socket *rx_sock;
+
+ vhost_net_stop(n, &tx_sock, &rx_sock);
+ vhost_net_flush(n);
+ vhost_dev_cleanup(&n->dev);
+ if (tx_sock)
+ fput(tx_sock->file);
+ if (rx_sock)
+ fput(rx_sock->file);
+ /* We do an extra flush before freeing memory,
+ * since jobs can re-queue themselves. */
+ vhost_net_flush(n);
+ kfree(n);
+ return 0;
+}
+
+static struct socket *get_raw_socket(int fd)
+{
+ struct {
+ struct sockaddr_ll sa;
+ char buf[MAX_ADDR_LEN];
+ } uaddr;
+ int uaddr_len = sizeof uaddr, r;
+ struct socket *sock = sockfd_lookup(fd, &r);
+ if (!sock)
+ return ERR_PTR(-ENOTSOCK);
+
+ /* Parameter checking */
+ if (sock->sk->sk_type != SOCK_RAW) {
+ r = -ESOCKTNOSUPPORT;
+ goto err;
+ }
+
+ r = sock->ops->getname(sock, (struct sockaddr *)&uaddr.sa,
+ &uaddr_len, 0);
+ if (r)
+ goto err;
+
+ if (uaddr.sa.sll_family != AF_PACKET) {
+ r = -EPFNOSUPPORT;
+ goto err;
+ }
+ return sock;
+err:
+ fput(sock->file);
+ return ERR_PTR(r);
+}
+
+static struct socket *get_tun_socket(int fd)
+{
+ struct file *file = fget(fd);
+ struct socket *sock;
+ if (!file)
+ return ERR_PTR(-EBADF);
+ sock = tun_get_socket(file);
+ if (IS_ERR(sock))
+ fput(file);
+ return sock;
+}
+
+static struct socket *get_socket(int fd)
+{
+ struct socket *sock;
+ if (fd == -1)
+ return NULL;
+ sock = get_raw_socket(fd);
+ if (!IS_ERR(sock))
+ return sock;
+ sock = get_tun_socket(fd);
+ if (!IS_ERR(sock))
+ return sock;
+ return ERR_PTR(-ENOTSOCK);
+}
+
+static long vhost_net_set_backend(struct vhost_net *n, unsigned index, int fd)
+{
+ struct socket *sock, *oldsock = NULL;
+ struct vhost_virtqueue *vq;
+ int r;
+
+ mutex_lock(&n->dev.mutex);
+ r = vhost_dev_check_owner(&n->dev);
+ if (r)
+ goto done;
+
+ if (index >= VHOST_NET_VQ_MAX) {
+ r = -ENOBUFS;
+ goto done;
+ }
+ vq = n->vqs + index;
+ mutex_lock(&vq->mutex);
+ sock = get_socket(fd);
+ if (IS_ERR(sock)) {
+ r = PTR_ERR(sock);
+ goto done;
+ }
+
+ /* start polling new socket */
+ oldsock = vq->private_data;
+ if (sock == oldsock)
+ goto done;
+
+ vhost_net_disable_vq(n, index);
+ rcu_assign_pointer(vq->private_data, sock);
+ vhost_net_enable_vq(n, index);
+ mutex_unlock(&vq->mutex);
+done:
+ mutex_unlock(&n->dev.mutex);
+ if (oldsock) {
+ vhost_net_flush_vq(n, index);
+ fput(oldsock->file);
+ }
+ return r;
+}
+
+static long vhost_net_reset_owner(struct vhost_net *n)
+{
+ struct socket *tx_sock = NULL;
+ struct socket *rx_sock = NULL;
+ long r;
+ mutex_lock(&n->dev.mutex);
+ r = vhost_dev_check_owner(&n->dev);
+ if (r)
+ goto done;
+ vhost_net_stop(n, &tx_sock, &rx_sock);
+ vhost_net_flush(n);
+ r = vhost_dev_reset_owner(&n->dev);
+done:
+ mutex_unlock(&n->dev.mutex);
+ if (tx_sock)
+ fput(tx_sock->file);
+ if (rx_sock)
+ fput(rx_sock->file);
+ return r;
+}
+
+static void vhost_net_set_features(struct vhost_net *n, u64 features)
+{
+ size_t hdr_size = features & (1 << VHOST_NET_F_VIRTIO_NET_HDR) ?
+ sizeof(struct virtio_net_hdr) : 0;
+ int i;
+ mutex_lock(&n->dev.mutex);
+ n->dev.acked_features = features;
+ smp_wmb();
+ for (i = 0; i < VHOST_NET_VQ_MAX; ++i) {
+ mutex_lock(&n->vqs[i].mutex);
+ n->vqs[i].hdr_size = hdr_size;
+ mutex_unlock(&n->vqs[i].mutex);
+ }
+ mutex_unlock(&n->dev.mutex);
+ vhost_net_flush(n);
+}
+
+static long vhost_net_ioctl(struct file *f, unsigned int ioctl,
+ unsigned long arg)
+{
+ struct vhost_net *n = f->private_data;
+ void __user *argp = (void __user *)arg;
+ u32 __user *featurep = argp;
+ struct vhost_vring_file backend;
+ u64 features;
+ int r;
+ switch (ioctl) {
+ case VHOST_NET_SET_BACKEND:
+ r = copy_from_user(&backend, argp, sizeof backend);
+ if (r < 0)
+ return r;
+ return vhost_net_set_backend(n, backend.index, backend.fd);
+ case VHOST_GET_FEATURES:
+ features = VHOST_FEATURES;
+ return put_user(features, featurep);
+ case VHOST_ACK_FEATURES:
+ r = get_user(features, featurep);
+ /* No features for now */
+ if (r < 0)
+ return r;
+ if (features & ~VHOST_FEATURES)
+ return -EOPNOTSUPP;
+ vhost_net_set_features(n, features);
+ return 0;
+ case VHOST_RESET_OWNER:
+ return vhost_net_reset_owner(n);
+ default:
+ r = vhost_dev_ioctl(&n->dev, ioctl, arg);
+ vhost_net_flush(n);
+ return r;
+ }
+}
+
+#ifdef CONFIG_COMPAT
+static long vhost_net_compat_ioctl(struct file *f, unsigned int ioctl,
+ unsigned long arg)
+{
+ return vhost_net_ioctl(f, ioctl, (unsigned long)compat_ptr(arg));
+}
+#endif
+
+const static struct file_operations vhost_net_fops = {
+ .owner = THIS_MODULE,
+ .release = vhost_net_release,
+ .unlocked_ioctl = vhost_net_ioctl,
+#ifdef CONFIG_COMPAT
+ .compat_ioctl = vhost_net_compat_ioctl,
+#endif
+ .open = vhost_net_open,
+};
+
+static struct miscdevice vhost_net_misc = {
+ VHOST_NET_MINOR,
+ "vhost-net",
+ &vhost_net_fops,
+};
+
+int vhost_net_init(void)
+{
+ int r = vhost_init();
+ if (r)
+ goto err_init;
+ r = misc_register(&vhost_net_misc);
+ if (r)
+ goto err_reg;
+ return 0;
+err_reg:
+ vhost_cleanup();
+err_init:
+ return r;
+
+}
+module_init(vhost_net_init);
+
+void vhost_net_exit(void)
+{
+ misc_deregister(&vhost_net_misc);
+ vhost_cleanup();
+}
+module_exit(vhost_net_exit);
+
+MODULE_VERSION("0.0.1");
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Michael S. Tsirkin");
+MODULE_DESCRIPTION("Host kernel accelerator for virtio net");
diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c
new file mode 100644
index 0000000..985f7f8
--- /dev/null
+++ b/drivers/vhost/vhost.c
@@ -0,0 +1,970 @@
+/* Copyright (C) 2009 Red Hat, Inc.
+ * Copyright (C) 2006 Rusty Russell IBM Corporation
+ *
+ * Author: Michael S. Tsirkin <mst@redhat.com>
+ *
+ * Inspiration, some code, and most witty comments come from
+ * Documentation/lguest/lguest.c, by Rusty Russell
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2.
+ *
+ * Generic code for virtio server in host kernel.
+ */
+
+#include <linux/eventfd.h>
+#include <linux/vhost.h>
+#include <linux/virtio_net.h>
+#include <linux/mm.h>
+#include <linux/miscdevice.h>
+#include <linux/mutex.h>
+#include <linux/workqueue.h>
+#include <linux/rcupdate.h>
+#include <linux/poll.h>
+#include <linux/file.h>
+#include <linux/highmem.h>
+
+#include <linux/net.h>
+#include <linux/if_packet.h>
+#include <linux/if_arp.h>
+
+#include <net/sock.h>
+
+#include "vhost.h"
+
+enum {
+ VHOST_MEMORY_MAX_NREGIONS = 64,
+ VHOST_MEMORY_F_LOG = 0x1,
+};
+
+static struct workqueue_struct *vhost_workqueue;
+
+static void vhost_poll_func(struct file *file, wait_queue_head_t *wqh,
+ poll_table *pt)
+{
+ struct vhost_poll *poll;
+ poll = container_of(pt, struct vhost_poll, table);
+
+ poll->wqh = wqh;
+ add_wait_queue(wqh, &poll->wait);
+}
+
+static int vhost_poll_wakeup(wait_queue_t *wait, unsigned mode, int sync,
+ void *key)
+{
+ struct vhost_poll *poll;
+ poll = container_of(wait, struct vhost_poll, wait);
+ if (!((unsigned long)key & poll->mask))
+ return 0;
+
+ queue_work(vhost_workqueue, &poll->work);
+ return 0;
+}
+
+/* Init poll structure */
+void vhost_poll_init(struct vhost_poll *poll, work_func_t func,
+ unsigned long mask)
+{
+ INIT_WORK(&poll->work, func);
+ init_waitqueue_func_entry(&poll->wait, vhost_poll_wakeup);
+ init_poll_funcptr(&poll->table, vhost_poll_func);
+ poll->mask = mask;
+}
+
+/* Start polling a file. We add ourselves to file's wait queue. The caller must
+ * keep a reference to a file until after vhost_poll_stop is called. */
+void vhost_poll_start(struct vhost_poll *poll, struct file *file)
+{
+ unsigned long mask;
+ mask = file->f_op->poll(file, &poll->table);
+ if (mask)
+ vhost_poll_wakeup(&poll->wait, 0, 0, (void *)mask);
+}
+
+/* Stop polling a file. After this function returns, it becomes safe to drop the
+ * file reference. You must also flush afterwards. */
+void vhost_poll_stop(struct vhost_poll *poll)
+{
+ remove_wait_queue(poll->wqh, &poll->wait);
+}
+
+/* Flush any work that has been scheduled. When calling this, don't hold any
+ * locks that are also used by the callback. */
+void vhost_poll_flush(struct vhost_poll *poll)
+{
+ flush_work(&poll->work);
+}
+
+void vhost_poll_queue(struct vhost_poll *poll)
+{
+ queue_work(vhost_workqueue, &poll->work);
+}
+
+long vhost_dev_init(struct vhost_dev *dev,
+ struct vhost_virtqueue *vqs, int nvqs)
+{
+ int i;
+ dev->vqs = vqs;
+ dev->nvqs = nvqs;
+ mutex_init(&dev->mutex);
+
+ for (i = 0; i < dev->nvqs; ++i) {
+ dev->vqs[i].dev = dev;
+ mutex_init(&dev->vqs[i].mutex);
+ if (dev->vqs[i].handle_kick)
+ vhost_poll_init(&dev->vqs[i].poll,
+ dev->vqs[i].handle_kick,
+ POLLIN);
+ }
+ return 0;
+}
+
+/* Caller should have device mutex */
+long vhost_dev_check_owner(struct vhost_dev *dev)
+{
+ /* Are you the owner? If not, I don't think you mean to do that */
+ return dev->mm == current->mm ? 0 : -EPERM;
+}
+
+/* Caller should have device mutex */
+static long vhost_dev_set_owner(struct vhost_dev *dev)
+{
+ /* Is there an owner already? */
+ if (dev->mm)
+ return -EBUSY;
+ /* No owner, become one */
+ dev->mm = get_task_mm(current);
+ return 0;
+}
+
+/* Caller should have device mutex */
+long vhost_dev_reset_owner(struct vhost_dev *dev)
+{
+ struct vhost_memory *memory;
+
+ /* Restore memory to default 1:1 mapping. */
+ memory = kzalloc(offsetof(struct vhost_memory, regions) +
+ 2 * sizeof *memory->regions, GFP_KERNEL);
+ if (!memory)
+ return -ENOMEM;
+
+ vhost_dev_cleanup(dev);
+
+ memory->nregions = 2;
+ memory->regions[0].guest_phys_addr = 1;
+ memory->regions[0].userspace_addr = 1;
+ memory->regions[0].memory_size = ~0ULL;
+ memory->regions[1].guest_phys_addr = 0;
+ memory->regions[1].userspace_addr = 0;
+ memory->regions[1].memory_size = 1;
+ dev->memory = memory;
+ return 0;
+}
+
+/* Caller should have device mutex */
+void vhost_dev_cleanup(struct vhost_dev *dev)
+{
+ int i;
+ for (i = 0; i < dev->nvqs; ++i) {
+ if (dev->vqs[i].kick && dev->vqs[i].handle_kick) {
+ vhost_poll_stop(&dev->vqs[i].poll);
+ vhost_poll_flush(&dev->vqs[i].poll);
+ }
+ if (dev->vqs[i].error_ctx)
+ eventfd_ctx_put(dev->vqs[i].error_ctx);
+ if (dev->vqs[i].error)
+ fput(dev->vqs[i].error);
+ if (dev->vqs[i].kick)
+ fput(dev->vqs[i].kick);
+ if (dev->vqs[i].call_ctx)
+ eventfd_ctx_put(dev->vqs[i].call_ctx);
+ if (dev->vqs[i].call)
+ fput(dev->vqs[i].call);
+ dev->vqs[i].error_ctx = NULL;
+ dev->vqs[i].error = NULL;
+ dev->vqs[i].kick = NULL;
+ dev->vqs[i].call_ctx = NULL;
+ dev->vqs[i].call = NULL;
+ }
+ if (dev->log_ctx)
+ eventfd_ctx_put(dev->log_ctx);
+ dev->log_ctx = NULL;
+ if (dev->log_file)
+ fput(dev->log_file);
+ dev->log_file = NULL;
+ /* No one will access memory at this point */
+ kfree(dev->memory);
+ dev->memory = NULL;
+ if (dev->mm)
+ mmput(dev->mm);
+ dev->mm = NULL;
+}
+
+static long vhost_set_memory(struct vhost_dev *d, struct vhost_memory __user *m)
+{
+ struct vhost_memory mem, *newmem, *oldmem;
+ unsigned long size = offsetof(struct vhost_memory, regions);
+ long r;
+ r = copy_from_user(&mem, m, size);
+ if (r)
+ return r;
+ if (mem.padding)
+ return -EOPNOTSUPP;
+ if (mem.nregions > VHOST_MEMORY_MAX_NREGIONS)
+ return -E2BIG;
+ newmem = kmalloc(size + mem.nregions * sizeof *m->regions, GFP_KERNEL);
+ if (!newmem)
+ return -ENOMEM;
+
+ memcpy(newmem, &mem, size);
+ r = copy_from_user(newmem->regions, m->regions,
+ mem.nregions * sizeof *m->regions);
+ if (r) {
+ kfree(newmem);
+ return r;
+ }
+ oldmem = d->memory;
+ rcu_assign_pointer(d->memory, newmem);
+ synchronize_rcu();
+ kfree(oldmem);
+ return 0;
+}
+
+static int init_used(struct vhost_virtqueue *vq)
+{
+ int r = put_user(vq->used_flags, &vq->used->flags);
+ if (r)
+ return r;
+ return get_user(vq->last_used_idx, &vq->used->idx);
+}
+
+static long vhost_set_vring(struct vhost_dev *d, int ioctl, void __user *argp)
+{
+ struct file *eventfp, *filep = NULL,
+ *pollstart = NULL, *pollstop = NULL;
+ struct eventfd_ctx *ctx = NULL;
+ u32 __user *idxp = argp;
+ struct vhost_virtqueue *vq;
+ struct vhost_vring_state s;
+ struct vhost_vring_file f;
+ struct vhost_vring_addr a;
+ u32 idx;
+ long r;
+
+ r = get_user(idx, idxp);
+ if (r < 0)
+ return r;
+ if (idx > d->nvqs)
+ return -ENOBUFS;
+
+ vq = d->vqs + idx;
+
+ mutex_lock(&vq->mutex);
+
+ switch (ioctl) {
+ case VHOST_SET_VRING_NUM:
+ r = copy_from_user(&s, argp, sizeof s);
+ if (r < 0)
+ break;
+ if (s.num > 0xffff) {
+ r = -EINVAL;
+ break;
+ }
+ vq->num = s.num;
+ break;
+ case VHOST_SET_VRING_BASE:
+ r = copy_from_user(&s, argp, sizeof s);
+ if (r < 0)
+ break;
+ if (s.num > 0xffff) {
+ r = -EINVAL;
+ break;
+ }
+ vq->avail_idx = vq->last_avail_idx = s.num;
+ break;
+ case VHOST_GET_VRING_BASE:
+ s.index = idx;
+ s.num = vq->last_avail_idx;
+ r = copy_to_user(argp, &s, sizeof s);
+ break;
+ case VHOST_SET_VRING_DESC:
+ r = copy_from_user(&a, argp, sizeof a);
+ if (r < 0)
+ break;
+ if (a.padding) {
+ r = -EOPNOTSUPP;
+ break;
+ }
+ if ((u64)(unsigned long)a.user_addr != a.user_addr) {
+ r = -EFAULT;
+ break;
+ }
+ vq->desc = (void __user *)(unsigned long)a.user_addr;
+ break;
+ case VHOST_SET_VRING_AVAIL:
+ r = copy_from_user(&a, argp, sizeof a);
+ if (r < 0)
+ break;
+ if (a.padding) {
+ r = -EOPNOTSUPP;
+ break;
+ }
+ if ((u64)(unsigned long)a.user_addr != a.user_addr) {
+ r = -EFAULT;
+ break;
+ }
+ if (a.user_addr & (sizeof *vq->avail->ring - 1)) {
+ r = -EINVAL;
+ break;
+ }
+ vq->avail = (void __user *)(unsigned long)a.user_addr;
+ /* Forget the cached index value. */
+ vq->avail_idx = vq->last_avail_idx;
+ break;
+ case VHOST_SET_VRING_USED:
+ r = copy_from_user(&a, argp, sizeof a);
+ if (r < 0)
+ break;
+ if (a.padding) {
+ r = -EOPNOTSUPP;
+ break;
+ }
+ if ((u64)(unsigned long)a.user_addr != a.user_addr) {
+ r = -EFAULT;
+ break;
+ }
+ if (a.user_addr & (sizeof *vq->used->ring - 1)) {
+ r = -EINVAL;
+ break;
+ }
+ vq->used = (void __user *)(unsigned long)a.user_addr;
+ r = init_used(vq);
+ if (r)
+ break;
+ break;
+ case VHOST_SET_VRING_LOG:
+ r = copy_from_user(&a, argp, sizeof a);
+ if (r < 0)
+ break;
+ if (a.padding) {
+ r = -EOPNOTSUPP;
+ break;
+ }
+ if (a.user_addr == VHOST_VRING_LOG_DISABLE) {
+ vq->log_used = false;
+ break;
+ }
+ if (a.user_addr & (sizeof *vq->used->ring - 1)) {
+ r = -EINVAL;
+ break;
+ }
+ vq->log_used = true;
+ vq->log_addr = a.user_addr;
+ break;
+ case VHOST_SET_VRING_KICK:
+ r = copy_from_user(&f, argp, sizeof f);
+ if (r < 0)
+ break;
+ eventfp = f.fd == -1 ? NULL : eventfd_fget(f.fd);
+ if (IS_ERR(eventfp))
+ return PTR_ERR(eventfp);
+ if (eventfp != vq->kick) {
+ pollstop = filep = vq->kick;
+ pollstart = vq->kick = eventfp;
+ } else
+ filep = eventfp;
+ break;
+ case VHOST_SET_VRING_CALL:
+ r = copy_from_user(&f, argp, sizeof f);
+ if (r < 0)
+ break;
+ eventfp = f.fd == -1 ? NULL : eventfd_fget(f.fd);
+ if (IS_ERR(eventfp))
+ return PTR_ERR(eventfp);
+ if (eventfp != vq->call) {
+ filep = vq->call;
+ ctx = vq->call_ctx;
+ vq->call = eventfp;
+ vq->call_ctx = eventfp ?
+ eventfd_ctx_fileget(eventfp) : NULL;
+ } else
+ filep = eventfp;
+ break;
+ case VHOST_SET_VRING_ERR:
+ r = copy_from_user(&f, argp, sizeof f);
+ if (r < 0)
+ break;
+ eventfp = f.fd == -1 ? NULL : eventfd_fget(f.fd);
+ if (IS_ERR(eventfp))
+ return PTR_ERR(eventfp);
+ if (eventfp != vq->error) {
+ filep = vq->error;
+ vq->error = eventfp;
+ ctx = vq->error_ctx;
+ vq->error_ctx = eventfp ?
+ eventfd_ctx_fileget(eventfp) : NULL;
+ } else
+ filep = eventfp;
+ break;
+ default:
+ r = -ENOIOCTLCMD;
+ }
+
+ if (pollstop && vq->handle_kick)
+ vhost_poll_stop(&vq->poll);
+
+ if (ctx)
+ eventfd_ctx_put(ctx);
+ if (filep)
+ fput(filep);
+
+ if (pollstart && vq->handle_kick)
+ vhost_poll_start(&vq->poll, vq->kick);
+
+ mutex_unlock(&vq->mutex);
+
+ if (pollstop && vq->handle_kick)
+ vhost_poll_flush(&vq->poll);
+ return r;
+}
+
+long vhost_dev_ioctl(struct vhost_dev *d, unsigned int ioctl, unsigned long arg)
+{
+ void __user *argp = (void __user *)arg;
+ struct file *eventfp, *filep = NULL;
+ struct eventfd_ctx *ctx = NULL;
+ u64 p;
+ long r;
+ int i, fd;
+
+ mutex_lock(&d->mutex);
+ /* If you are not the owner, you can become one */
+ if (ioctl == VHOST_SET_OWNER) {
+ r = vhost_dev_set_owner(d);
+ goto done;
+ }
+
+ /* You must be the owner to do anything else */
+ r = vhost_dev_check_owner(d);
+ if (r)
+ goto done;
+
+ switch (ioctl) {
+ case VHOST_SET_MEM_TABLE:
+ r = vhost_set_memory(d, argp);
+ break;
+ case VHOST_SET_LOG_BASE:
+ r = copy_from_user(&p, argp, sizeof p);
+ if (r < 0)
+ break;
+ if ((u64)(unsigned long)p != p) {
+ r = -EFAULT;
+ break;
+ }
+ for (i = 0; i < d->nvqs; ++i) {
+ mutex_lock(&d->vqs[i].mutex);
+ d->vqs[i].log_base = (void __user *)(unsigned long)p;
+ mutex_unlock(&d->vqs[i].mutex);
+ }
+ break;
+ case VHOST_SET_LOG_FD:
+ r = get_user(fd, (int __user *)argp);
+ if (r < 0)
+ break;
+ eventfp = fd == -1 ? NULL : eventfd_fget(fd);
+ if (IS_ERR(eventfp)) {
+ r = PTR_ERR(eventfp);
+ break;
+ }
+ if (eventfp != d->log_file) {
+ filep = d->log_file;
+ ctx = d->log_ctx;
+ d->log_ctx = eventfp ?
+ eventfd_ctx_fileget(eventfp) : NULL;
+ } else
+ filep = eventfp;
+ for (i = 0; i < d->nvqs; ++i) {
+ mutex_lock(&d->vqs[i].mutex);
+ d->vqs[i].log_ctx = d->log_ctx;
+ mutex_unlock(&d->vqs[i].mutex);
+ }
+ if (ctx)
+ eventfd_ctx_put(ctx);
+ if (filep)
+ fput(filep);
+ break;
+ default:
+ r = vhost_set_vring(d, ioctl, argp);
+ break;
+ }
+done:
+ mutex_unlock(&d->mutex);
+ return r;
+}
+
+static const struct vhost_memory_region *find_region(struct vhost_memory *mem,
+ __u64 addr, __u32 len)
+{
+ struct vhost_memory_region *reg;
+ int i;
+ /* linear search is not brilliant, but we really have on the order of 6
+ * regions in practice */
+ for (i = 0; i < mem->nregions; ++i) {
+ reg = mem->regions + i;
+ if (reg->guest_phys_addr <= addr &&
+ reg->guest_phys_addr + reg->memory_size - 1 >= addr)
+ return reg;
+ }
+ return NULL;
+}
+
+/* TODO: This is really inefficient. We need something like get_user()
+ * (instruction directly accesses the data, with an exception table entry
+ * returning -EFAULT). See Documentation/x86/exception-tables.txt.
+ */
+static int set_bit_to_user(int nr, void __user *addr)
+{
+ unsigned long log = (unsigned long)addr;
+ struct page *page;
+ void *base;
+ int bit = nr + (log % PAGE_SIZE) * 8;
+ int r;
+ r = get_user_pages_fast(log, 1, 1, &page);
+ if (r)
+ return r;
+ base = kmap_atomic(page, KM_USER0);
+ set_bit(bit, base);
+ kunmap_atomic(base, KM_USER0);
+ set_page_dirty_lock(page);
+ put_page(page);
+ return 0;
+}
+
+static int log_write(void __user *log_base,
+ u64 write_address, u64 write_length)
+{
+ int r;
+ if (!write_length)
+ return 0;
+ write_address /= VHOST_PAGE_SIZE;
+ for (;;) {
+ u64 base = (u64)(unsigned long)log_base;
+ u64 log = base + write_address / 8;
+ int bit = write_address % 8;
+ if ((u64)(unsigned long)log != log)
+ return -EFAULT;
+ r = set_bit_to_user(bit, (void __user *)(unsigned long)log);
+ if (r < 0)
+ return r;
+ if (write_length <= VHOST_PAGE_SIZE)
+ break;
+ write_length -= VHOST_PAGE_SIZE;
+ write_address += VHOST_PAGE_SIZE;
+ }
+ return r;
+}
+
+int vhost_log_write(struct vhost_virtqueue *vq, struct vhost_log *log,
+ unsigned int log_num, u64 len)
+{
+ int i, r;
+
+ /* Make sure data written is seen before log. */
+ wmb();
+ for (i = 0; i < log_num; ++i) {
+ u64 l = min(log[i].len, len);
+ r = log_write(vq->log_base, log[i].addr, l);
+ if (r < 0)
+ return r;
+ len -= l;
+ if (!len)
+ return 0;
+ }
+ if (vq->log_ctx)
+ eventfd_signal(vq->log_ctx, 1);
+ /* Length written exceeds what we have stored. This is a bug. */
+ BUG();
+ return 0;
+}
+
+int translate_desc(struct vhost_dev *dev, u64 addr, u32 len,
+ struct iovec iov[], int iov_size)
+{
+ const struct vhost_memory_region *reg;
+ struct vhost_memory *mem;
+ struct iovec *_iov;
+ u64 s = 0;
+ int ret = 0;
+
+ rcu_read_lock();
+
+ mem = rcu_dereference(dev->memory);
+ while ((u64)len > s) {
+ u64 size;
+ if (ret >= iov_size) {
+ ret = -ENOBUFS;
+ break;
+ }
+ reg = find_region(mem, addr, len);
+ if (!reg) {
+ ret = -EFAULT;
+ break;
+ }
+ _iov = iov + ret;
+ size = reg->memory_size - addr + reg->guest_phys_addr;
+ _iov->iov_len = min((u64)len, size);
+ _iov->iov_base = (void *)(unsigned long)
+ (reg->userspace_addr + addr - reg->guest_phys_addr);
+ s += size;
+ addr += size;
+ ++ret;
+ }
+
+ rcu_read_unlock();
+ return ret;
+}
+
+/* Each buffer in the virtqueues is actually a chain of descriptors. This
+ * function returns the next descriptor in the chain,
+ * or -1 if we're at the end. */
+static unsigned next_desc(struct vring_desc *desc)
+{
+ unsigned int next;
+
+ /* If this descriptor says it doesn't chain, we're done. */
+ if (!(desc->flags & VRING_DESC_F_NEXT))
+ return -1;
+
+ /* Check they're not leading us off end of descriptors. */
+ next = desc->next;
+ /* Make sure compiler knows to grab that: we don't want it changing! */
+ /* We will use the result as an index in an array, so most
+ * architectures only need a compiler barrier here. */
+ read_barrier_depends();
+
+ return next;
+}
+
+static unsigned get_indirect(struct vhost_dev *dev, struct vhost_virtqueue *vq,
+ struct iovec iov[],
+ unsigned int *out_num, unsigned int *in_num,
+ struct vhost_log *log, unsigned int *log_num,
+ struct vring_desc *indirect)
+{
+ struct vring_desc desc;
+ unsigned int i = 0, count, found = 0;
+ int ret;
+
+ /* Sanity check */
+ if (indirect->len % sizeof desc) {
+ vq_err(vq, "Invalid length in indirect descriptor: "
+ "len 0x%llx not multiple of 0x%zx\n",
+ (unsigned long long)indirect->len,
+ sizeof desc);
+ return -EINVAL;
+ }
+
+ ret = translate_desc(dev, indirect->addr, indirect->len, vq->indirect,
+ ARRAY_SIZE(vq->indirect));
+ if (ret < 0) {
+ vq_err(vq, "Translation failure %d in indirect.\n", ret);
+ return ret;
+ }
+
+ /* We will use the result as an address to read from, so most
+ * architectures only need a compiler barrier here. */
+ read_barrier_depends();
+
+ count = indirect->len / sizeof desc;
+ /* Buffers are chained via a 16 bit next field, so
+ * we can have at most 2^16 of these. */
+ if (count > USHORT_MAX + 1) {
+ vq_err(vq, "Indirect buffer length too big: %d\n",
+ indirect->len);
+ return -E2BIG;
+ }
+
+ do {
+ unsigned iov_count = *in_num + *out_num;
+ if (++found > count) {
+ vq_err(vq, "Loop detected: last one at %u "
+ "indirect size %u\n",
+ i, count);
+ return -EINVAL;
+ }
+ if (memcpy_fromiovec((unsigned char *)&desc, vq->indirect,
+ sizeof desc)) {
+ vq_err(vq, "Failed indirect descriptor: idx %d, %zx\n",
+ i, (size_t)indirect->addr + i * sizeof desc);
+ return -EINVAL;
+ }
+ if (desc.flags & VRING_DESC_F_INDIRECT) {
+ vq_err(vq, "Nested indirect descriptor: idx %d, %zx\n",
+ i, (size_t)indirect->addr + i * sizeof desc);
+ return -EINVAL;
+ }
+
+ ret = translate_desc(dev, desc.addr, desc.len, iov + iov_count,
+ VHOST_NET_MAX_SG - iov_count);
+ if (ret < 0) {
+ vq_err(vq, "Translation failure %d indirect idx %d\n",
+ ret, i);
+ return ret;
+ }
+ /* If this is an input descriptor, increment that count. */
+ if (desc.flags & VRING_DESC_F_WRITE) {
+ *in_num += ret;
+ if (unlikely(log)) {
+ log[*log_num].addr = desc.addr;
+ log[*log_num].len = desc.len;
+ ++*log_num;
+ }
+ } else {
+ /* If it's an output descriptor, they're all supposed
+ * to come before any input descriptors. */
+ if (*in_num) {
+ vq_err(vq, "Indirect descriptor "
+ "has out after in: idx %d\n", i);
+ return -EINVAL;
+ }
+ *out_num += ret;
+ }
+ } while ((i = next_desc(&desc)) != -1);
+ return 0;
+}
+
+/* This looks in the virtqueue and for the first available buffer, and converts
+ * it to an iovec for convenient access. Since descriptors consist of some
+ * number of output then some number of input descriptors, it's actually two
+ * iovecs, but we pack them into one and note how many of each there were.
+ *
+ * This function returns the descriptor number found, or vq->num (which
+ * is never a valid descriptor number) if none was found. */
+unsigned vhost_get_vq_desc(struct vhost_dev *dev, struct vhost_virtqueue *vq,
+ struct iovec iov[],
+ unsigned int *out_num, unsigned int *in_num,
+ struct vhost_log *log, unsigned int *log_num)
+{
+ struct vring_desc desc;
+ unsigned int i, head, found = 0;
+ u16 last_avail_idx;
+ int ret;
+
+ /* Check it isn't doing very strange things with descriptor numbers. */
+ last_avail_idx = vq->last_avail_idx;
+ if (get_user(vq->avail_idx, &vq->avail->idx)) {
+ vq_err(vq, "Failed to access avail idx at %p\n",
+ &vq->avail->idx);
+ return vq->num;
+ }
+
+ if ((u16)(vq->avail_idx - last_avail_idx) > vq->num) {
+ vq_err(vq, "Guest moved used index from %u to %u",
+ last_avail_idx, vq->avail_idx);
+ return vq->num;
+ }
+
+ /* If there's nothing new since last we looked, return invalid. */
+ if (vq->avail_idx == last_avail_idx)
+ return vq->num;
+
+ /* Only get avail ring entries after they have been exposed by guest. */
+ rmb();
+
+ /* Grab the next descriptor number they're advertising, and increment
+ * the index we've seen. */
+ if (get_user(head, &vq->avail->ring[last_avail_idx % vq->num])) {
+ vq_err(vq, "Failed to read head: idx %d address %p\n",
+ last_avail_idx,
+ &vq->avail->ring[last_avail_idx % vq->num]);
+ return vq->num;
+ }
+
+ /* If their number is silly, that's an error. */
+ if (head >= vq->num) {
+ vq_err(vq, "Guest says index %u > %u is available",
+ head, vq->num);
+ return vq->num;
+ }
+
+ /* When we start there are none of either input nor output. */
+ *out_num = *in_num = 0;
+ if (unlikely(log))
+ *log_num = 0;
+
+ i = head;
+ do {
+ unsigned iov_count = *in_num + *out_num;
+ if (i >= vq->num) {
+ vq_err(vq, "Desc index is %u > %u, head = %u",
+ i, vq->num, head);
+ return vq->num;
+ }
+ if (++found > vq->num) {
+ vq_err(vq, "Loop detected: last one at %u "
+ "vq size %u head %u\n",
+ i, vq->num, head);
+ return vq->num;
+ }
+ ret = copy_from_user(&desc, vq->desc + i, sizeof desc);
+ if (ret) {
+ vq_err(vq, "Failed to get descriptor: idx %d addr %p\n",
+ i, vq->desc + i);
+ return vq->num;
+ }
+ if (desc.flags & VRING_DESC_F_INDIRECT) {
+ ret = get_indirect(dev, vq, iov, out_num, in_num,
+ log, log_num, &desc);
+ if (ret < 0) {
+ vq_err(vq, "Failure detected "
+ "in indirect descriptor at idx %d\n", i);
+ return vq->num;
+ }
+ continue;
+ }
+
+ ret = translate_desc(dev, desc.addr, desc.len, iov + iov_count,
+ VHOST_NET_MAX_SG - iov_count);
+ if (ret < 0) {
+ vq_err(vq, "Translation failure %d descriptor idx %d\n",
+ ret, i);
+ return vq->num;
+ }
+ if (desc.flags & VRING_DESC_F_WRITE) {
+ /* If this is an input descriptor,
+ * increment that count. */
+ *in_num += ret;
+ if (unlikely(log)) {
+ log[*log_num].addr = desc.addr;
+ log[*log_num].len = desc.len;
+ ++*log_num;
+ }
+ } else {
+ /* If it's an output descriptor, they're all supposed
+ * to come before any input descriptors. */
+ if (*in_num) {
+ vq_err(vq, "Descriptor has out after in: "
+ "idx %d\n", i);
+ return vq->num;
+ }
+ *out_num += ret;
+ }
+ } while ((i = next_desc(&desc)) != -1);
+
+ /* On success, increment avail index. */
+ vq->last_avail_idx++;
+ return head;
+}
+
+/* Reverse the effect of vhost_get_vq_desc. Useful for error handling. */
+void vhost_discard_vq_desc(struct vhost_virtqueue *vq)
+{
+ vq->last_avail_idx--;
+}
+
+/* After we've used one of their buffers, we tell them about it. We'll then
+ * want to send them an interrupt, using vq->call. */
+int vhost_add_used(struct vhost_virtqueue *vq,
+ unsigned int head, int len)
+{
+ struct vring_used_elem *used;
+
+ /* The virtqueue contains a ring of used buffers. Get a pointer to the
+ * next entry in that used ring. */
+ used = &vq->used->ring[vq->last_used_idx % vq->num];
+ if (put_user(head, &used->id)) {
+ vq_err(vq, "Failed to write used id");
+ return -EFAULT;
+ }
+ if (put_user(len, &used->len)) {
+ vq_err(vq, "Failed to write used len");
+ return -EFAULT;
+ }
+ /* Make sure buffer is written before we update index. */
+ wmb();
+ if (put_user(vq->last_used_idx + 1, &vq->used->idx)) {
+ vq_err(vq, "Failed to increment used idx");
+ return -EFAULT;
+ }
+ if (unlikely(vq->log_used)) {
+ /* Make sure data is seen before log. */
+ wmb();
+ log_write(vq->log_base, vq->log_addr + sizeof *vq->used->ring *
+ (vq->last_used_idx % vq->num),
+ sizeof *vq->used->ring);
+ log_write(vq->log_base, vq->log_addr, sizeof *vq->used->ring);
+ if (vq->log_ctx)
+ eventfd_signal(vq->log_ctx, 1);
+ }
+ vq->last_used_idx++;
+ return 0;
+}
+
+/* This actually sends the interrupt for this virtqueue */
+void vhost_trigger_irq(struct vhost_dev *dev, struct vhost_virtqueue *vq)
+{
+ __u16 flags = 0;
+ if (get_user(flags, &vq->avail->flags)) {
+ vq_err(vq, "Failed to get flags");
+ return;
+ }
+
+ /* If they don't want an interrupt, don't send one, unless empty. */
+ if ((flags & VRING_AVAIL_F_NO_INTERRUPT) &&
+ (vq->avail_idx != vq->last_avail_idx ||
+ !vhost_has_feature(dev, VIRTIO_F_NOTIFY_ON_EMPTY)))
+ return;
+
+ /* Send the Guest an interrupt tell them we used something up. */
+ if (vq->call_ctx)
+ eventfd_signal(vq->call_ctx, 1);
+}
+
+/* And here's the combo meal deal. Supersize me! */
+void vhost_add_used_and_trigger(struct vhost_dev *dev,
+ struct vhost_virtqueue *vq,
+ unsigned int head, int len)
+{
+ vhost_add_used(vq, head, len);
+ vhost_trigger_irq(dev, vq);
+}
+
+/* OK, now we need to know about added descriptors. */
+bool vhost_notify(struct vhost_virtqueue *vq)
+{
+ int r;
+ if (!(vq->used_flags & VRING_USED_F_NO_NOTIFY))
+ return false;
+ vq->used_flags &= ~VRING_USED_F_NO_NOTIFY;
+ r = put_user(vq->used_flags, &vq->used->flags);
+ if (r)
+ vq_err(vq, "Failed to disable notification: %d\n", r);
+ /* They could have slipped one in as we were doing that: make
+ * sure it's written, tell caller it needs to check again. */
+ mb();
+ return true;
+}
+
+/* We don't need to be notified again. */
+void vhost_no_notify(struct vhost_virtqueue *vq)
+{
+ int r;
+ if (vq->used_flags & VRING_USED_F_NO_NOTIFY)
+ return;
+ vq->used_flags |= VRING_USED_F_NO_NOTIFY;
+ r = put_user(vq->used_flags, &vq->used->flags);
+ if (r)
+ vq_err(vq, "Failed to enable notification: %d\n", r);
+}
+
+int vhost_init(void)
+{
+ vhost_workqueue = create_singlethread_workqueue("vhost");
+ if (!vhost_workqueue)
+ return -ENOMEM;
+ return 0;
+}
+
+void vhost_cleanup(void)
+{
+ destroy_workqueue(vhost_workqueue);
+}
diff --git a/drivers/vhost/vhost.h b/drivers/vhost/vhost.h
new file mode 100644
index 0000000..d3900de
--- /dev/null
+++ b/drivers/vhost/vhost.h
@@ -0,0 +1,158 @@
+#ifndef _VHOST_H
+#define _VHOST_H
+
+#include <linux/eventfd.h>
+#include <linux/vhost.h>
+#include <linux/mm.h>
+#include <linux/mutex.h>
+#include <linux/workqueue.h>
+#include <linux/poll.h>
+#include <linux/file.h>
+#include <linux/skbuff.h>
+#include <linux/uio.h>
+#include <linux/virtio_config.h>
+#include <linux/virtio_ring.h>
+
+struct vhost_device;
+
+enum {
+ VHOST_NET_MAX_SG = MAX_SKB_FRAGS + 2,
+};
+
+/* Poll a file (eventfd or socket) */
+/* Note: there's nothing vhost specific about this structure. */
+struct vhost_poll {
+ poll_table table;
+ wait_queue_head_t *wqh;
+ wait_queue_t wait;
+ /* struct which will handle all actual work. */
+ struct work_struct work;
+ unsigned long mask;
+};
+
+void vhost_poll_init(struct vhost_poll *poll, work_func_t func,
+ unsigned long mask);
+void vhost_poll_start(struct vhost_poll *poll, struct file *file);
+void vhost_poll_stop(struct vhost_poll *poll);
+void vhost_poll_flush(struct vhost_poll *poll);
+void vhost_poll_queue(struct vhost_poll *poll);
+
+struct vhost_log {
+ u64 addr;
+ u64 len;
+};
+
+/* The virtqueue structure describes a queue attached to a device. */
+struct vhost_virtqueue {
+ struct vhost_dev *dev;
+
+ /* The actual ring of buffers. */
+ struct mutex mutex;
+ unsigned int num;
+ struct vring_desc __user *desc;
+ struct vring_avail __user *avail;
+ struct vring_used __user *used;
+ struct file *kick;
+ struct file *call;
+ struct file *error;
+ struct eventfd_ctx *call_ctx;
+ struct eventfd_ctx *error_ctx;
+ struct eventfd_ctx *log_ctx;
+
+ struct vhost_poll poll;
+
+ /* The routine to call when the Guest pings us, or timeout. */
+ work_func_t handle_kick;
+
+ /* Last available index we saw. */
+ u16 last_avail_idx;
+
+ /* Caches available index value from user. */
+ u16 avail_idx;
+
+ /* Last index we used. */
+ u16 last_used_idx;
+
+ /* Used flags */
+ u16 used_flags;
+
+ /* Log writes to used structure. */
+ bool log_used;
+ u64 log_addr;
+
+ struct iovec indirect[VHOST_NET_MAX_SG];
+ struct iovec iov[VHOST_NET_MAX_SG];
+ struct iovec hdr[VHOST_NET_MAX_SG];
+ size_t hdr_size;
+ /* We use a kind of RCU to access private pointer.
+ * All readers access it from workqueue, which makes it possible to
+ * flush the workqueue instead of synchronize_rcu. Therefore readers do
+ * not need to call rcu_read_lock/rcu_read_unlock: the beginning of
+ * work item execution acts instead of rcu_read_lock() and the end of
+ * work item execution acts instead of rcu_read_lock().
+ * Writers use virtqueue mutex. */
+ void *private_data;
+ /* Log write descriptors */
+ void __user *log_base;
+ struct vhost_log log[VHOST_NET_MAX_SG];
+};
+
+struct vhost_dev {
+ /* Readers use RCU to access memory table pointer
+ * log base pointer and features.
+ * Writers use mutex below.*/
+ struct vhost_memory *memory;
+ struct mm_struct *mm;
+ struct mutex mutex;
+ unsigned acked_features;
+ struct vhost_virtqueue *vqs;
+ int nvqs;
+ struct file *log_file;
+ struct eventfd_ctx *log_ctx;
+};
+
+long vhost_dev_init(struct vhost_dev *, struct vhost_virtqueue *vqs, int nvqs);
+long vhost_dev_check_owner(struct vhost_dev *);
+long vhost_dev_reset_owner(struct vhost_dev *);
+void vhost_dev_cleanup(struct vhost_dev *);
+long vhost_dev_ioctl(struct vhost_dev *, unsigned int ioctl, unsigned long arg);
+
+unsigned vhost_get_vq_desc(struct vhost_dev *, struct vhost_virtqueue *,
+ struct iovec iov[],
+ unsigned int *out_num, unsigned int *in_num,
+ struct vhost_log *log, unsigned int *log_num);
+void vhost_discard_vq_desc(struct vhost_virtqueue *);
+
+int vhost_add_used(struct vhost_virtqueue *, unsigned int head, int len);
+void vhost_trigger_irq(struct vhost_dev *, struct vhost_virtqueue *);
+void vhost_add_used_and_trigger(struct vhost_dev *, struct vhost_virtqueue *,
+ unsigned int head, int len);
+void vhost_no_notify(struct vhost_virtqueue *);
+bool vhost_notify(struct vhost_virtqueue *);
+
+int vhost_log_write(struct vhost_virtqueue *vq, struct vhost_log *log,
+ unsigned int log_num, u64 len);
+
+int vhost_init(void);
+void vhost_cleanup(void);
+
+#define vq_err(vq, fmt, ...) do { \
+ pr_debug(pr_fmt(fmt), ##__VA_ARGS__); \
+ if ((vq)->error_ctx) \
+ eventfd_signal((vq)->error_ctx, 1);\
+ } while (0)
+
+enum {
+ VHOST_FEATURES = (1 << VIRTIO_F_NOTIFY_ON_EMPTY) |
+ (1 << VIRTIO_RING_F_INDIRECT_DESC) |
+ (1 << VHOST_F_LOG_ALL) |
+ (1 << VHOST_NET_F_VIRTIO_NET_HDR),
+};
+
+static inline int vhost_has_feature(struct vhost_dev *dev, int bit)
+{
+ unsigned acked_features = rcu_dereference(dev->acked_features);
+ return acked_features & (1 << bit);
+}
+
+#endif
diff --git a/include/linux/Kbuild b/include/linux/Kbuild
index 3f384d4..b6335c6 100644
--- a/include/linux/Kbuild
+++ b/include/linux/Kbuild
@@ -361,6 +361,7 @@ unifdef-y += uio.h
unifdef-y += unistd.h
unifdef-y += usbdevice_fs.h
unifdef-y += utsname.h
+unifdef-y += vhost.h
unifdef-y += videodev2.h
unifdef-y += videodev.h
unifdef-y += virtio_config.h
diff --git a/include/linux/miscdevice.h b/include/linux/miscdevice.h
index adaf3c1..8b5f7cc 100644
--- a/include/linux/miscdevice.h
+++ b/include/linux/miscdevice.h
@@ -30,6 +30,7 @@
#define HPET_MINOR 228
#define FUSE_MINOR 229
#define KVM_MINOR 232
+#define VHOST_NET_MINOR 233
#define MISC_DYNAMIC_MINOR 255
struct device;
diff --git a/include/linux/vhost.h b/include/linux/vhost.h
new file mode 100644
index 0000000..f66142f
--- /dev/null
+++ b/include/linux/vhost.h
@@ -0,0 +1,126 @@
+#ifndef _LINUX_VHOST_H
+#define _LINUX_VHOST_H
+/* Userspace interface for in-kernel virtio accelerators. */
+
+/* vhost is used to reduce the number of system calls involved in virtio.
+ *
+ * Existing virtio net code is used in the guest without modification.
+ *
+ * This header includes interface used by userspace hypervisor for
+ * device configuration.
+ */
+
+#include <linux/types.h>
+#include <linux/compiler.h>
+#include <linux/ioctl.h>
+#include <linux/virtio_config.h>
+#include <linux/virtio_ring.h>
+
+struct vhost_vring_state {
+ unsigned int index;
+ unsigned int num;
+};
+
+struct vhost_vring_file {
+ unsigned int index;
+ int fd; /* Pass -1 to unbind from file. */
+
+};
+
+struct vhost_vring_addr {
+ unsigned int index;
+ unsigned int padding;
+ __u64 user_addr;
+};
+
+struct vhost_memory_region {
+ __u64 guest_phys_addr;
+ __u64 memory_size; /* bytes */
+ __u64 userspace_addr;
+ __u64 flags_padding; /* No flags are currently specified. */
+};
+
+/* All region addresses and sizes must be 4K aligned. */
+#define VHOST_PAGE_SIZE 0x1000
+
+struct vhost_memory {
+ __u32 nregions;
+ __u32 padding;
+ struct vhost_memory_region regions[0];
+};
+
+/* ioctls */
+
+#define VHOST_VIRTIO 0xAF
+
+/* Features bitmask for forward compatibility. Transport bits are used for
+ * vhost specific features. */
+#define VHOST_GET_FEATURES _IOR(VHOST_VIRTIO, 0x00, __u64)
+#define VHOST_ACK_FEATURES _IOW(VHOST_VIRTIO, 0x00, __u64)
+
+/* Set current process as the (exclusive) owner of this file descriptor. This
+ * must be called before any other vhost command. Further calls to
+ * VHOST_OWNER_SET fail until VHOST_OWNER_RESET is called. */
+#define VHOST_SET_OWNER _IO(VHOST_VIRTIO, 0x01)
+/* Give up ownership, and reset the device to default values.
+ * Allows subsequent call to VHOST_OWNER_SET to succeed. */
+#define VHOST_RESET_OWNER _IO(VHOST_VIRTIO, 0x02)
+
+/* Set up/modify memory layout */
+#define VHOST_SET_MEM_TABLE _IOW(VHOST_VIRTIO, 0x03, struct vhost_memory)
+
+/* Write logging setup. */
+/* Memory writes can optionally be logged by setting bit at an offset
+ * (calculated from the physical address) from specified log base.
+ * The bit is set using an atomic 32 bit operation. */
+/* Set base address for logging. */
+#define VHOST_SET_LOG_BASE _IOW(VHOST_VIRTIO, 0x04, __u64)
+/* Specify an eventfd file descriptor to signal on log write. */
+#define VHOST_SET_LOG_FD _IOW(VHOST_VIRTIO, 0x07, int)
+
+/* Ring setup. These parameters can not be modified while ring is running
+ * (bound to a device). */
+/* Set number of descriptors in ring */
+#define VHOST_SET_VRING_NUM _IOW(VHOST_VIRTIO, 0x10, struct vhost_vring_state)
+/* Start of array of descriptors (virtually contiguous) */
+#define VHOST_SET_VRING_DESC _IOW(VHOST_VIRTIO, 0x11, struct vhost_vring_addr)
+/* Used structure address. Must be 32 bit aligned */
+#define VHOST_SET_VRING_USED _IOW(VHOST_VIRTIO, 0x12, struct vhost_vring_addr)
+/* Available structure address. Must be 16 bit aligned */
+#define VHOST_SET_VRING_AVAIL _IOW(VHOST_VIRTIO, 0x13, struct vhost_vring_addr)
+/* Base value where queue looks for available descriptors */
+#define VHOST_SET_VRING_BASE _IOW(VHOST_VIRTIO, 0x14, struct vhost_vring_state)
+/* Get accessor: reads index, writes value in num */
+#define VHOST_GET_VRING_BASE _IOWR(VHOST_VIRTIO, 0x14, struct vhost_vring_state)
+
+/* Logging support. Can be modified while ring is running. */
+/* Log writes to used structure, at offset calculated from specified address.
+ * Address must be 32 bit aligned. Pass 0x1 to disable logging. */
+#define VHOST_SET_VRING_LOG _IOW(VHOST_VIRTIO, 0x18, struct vhost_vring_addr)
+#define VHOST_VRING_LOG_DISABLE (0x1)
+
+/* The following ioctls use eventfd file descriptors to signal and poll
+ * for events. */
+
+/* Set eventfd to poll for added buffers */
+#define VHOST_SET_VRING_KICK _IOW(VHOST_VIRTIO, 0x20, struct vhost_vring_file)
+/* Set eventfd to signal when buffers have beed used */
+#define VHOST_SET_VRING_CALL _IOW(VHOST_VIRTIO, 0x21, struct vhost_vring_file)
+/* Set eventfd to signal an error */
+#define VHOST_SET_VRING_ERR _IOW(VHOST_VIRTIO, 0x22, struct vhost_vring_file)
+
+/* VHOST_NET specific defines */
+
+/* Attach virtio net ring to a raw socket, or tap device.
+ * The socket must be already bound to an ethernet device, this device will be
+ * used for transmit. Pass fd -1 to unbind from the socket and the transmit
+ * device. This can be used to stop the ring (e.g. for migration). */
+#define VHOST_NET_SET_BACKEND _IOW(VHOST_VIRTIO, 0x30, struct vhost_vring_file)
+
+/* Feature bits */
+/* Log all write descriptors. Can be changed while device is active. */
+#define VHOST_F_LOG_ALL 26
+/* vhost-net should add virtio_net_hdr for RX, and strip for TX packets. */
+#define VHOST_NET_F_VIRTIO_NET_HDR 27
+
+#endif
--
1.6.5.2.143.g8cc62
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply related
* Re: [PATCHv8 0/3] vhost: a kernel-level virtio server
From: Gregory Haskins @ 2009-11-04 16:02 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: netdev, virtualization, kvm, linux-kernel, mingo, linux-mm, akpm,
hpa, Rusty Russell, s.hetze, Daniel Walker, Eric Dumazet
In-Reply-To: <20091104155234.GA32673@redhat.com>
[-- Attachment #1: Type: text/plain, Size: 744 bytes --]
Michael S. Tsirkin wrote:
> Ok, I think I've addressed all comments so far here.
> Rusty, I'd like this to go into linux-next, through your tree, and
> hopefully 2.6.33. What do you think?
I think the benchmark data is a prerequisite for merge consideration, IMO.
Do you have anything for us to look at? I think comparison that show
the following are of interest:
throughput (e.g. netperf::TCP_STREAM): guest->host, guest->host->guest,
guest->host->remote, host->remote, remote->host->guest
latency (e.g. netperf::UDP_RR): same conditions as throughput
cpu-utilization
others?
Ideally, this should be at least between upstream virtio and vhost.
Bonus points if you include venet as well.
Kind regards,
-Greg
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 267 bytes --]
^ permalink raw reply
* [PATCH 1/1] NET: sungem, use spin_trylock_irqsave
From: Jiri Slaby @ 2009-11-04 16:06 UTC (permalink / raw)
To: davem; +Cc: netdev, linux-kernel, Jiri Slaby
Use spin_trylock_irqsave instead of open-coded
local_irq_save+spin_trylock.
Impact: cleanup
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
---
drivers/net/sungem.c | 4 +---
1 files changed, 1 insertions(+), 3 deletions(-)
diff --git a/drivers/net/sungem.c b/drivers/net/sungem.c
index 7019a0d..51bde03 100644
--- a/drivers/net/sungem.c
+++ b/drivers/net/sungem.c
@@ -1034,10 +1034,8 @@ static netdev_tx_t gem_start_xmit(struct sk_buff *skb,
(csum_stuff_off << 21));
}
- local_irq_save(flags);
- if (!spin_trylock(&gp->tx_lock)) {
+ if (!spin_trylock_irqsave(&gp->tx_lock, flags)) {
/* Tell upper layer to requeue */
- local_irq_restore(flags);
return NETDEV_TX_LOCKED;
}
/* We raced with gem_do_stop() */
--
1.6.4.2
^ permalink raw reply related
* Re: [PATCH] e1000: the power down when running ifdown command
From: Stephen Hemminger @ 2009-11-04 16:08 UTC (permalink / raw)
To: Naohiro Ooiwa
Cc: jeffrey.t.kirsher, jesse.brandeburg, peter.p.waskiewicz.jr,
john.ronciak, davem, Andrew Morton, netdev, svaidy, e1000-devel
In-Reply-To: <4AF1562F.8010401@miraclelinux.com>
On Wed, 04 Nov 2009 19:23:43 +0900
Naohiro Ooiwa <nooiwa@miraclelinux.com> wrote:
> Naohiro Ooiwa wrote:
> > Stephen Hemminger wrote:
> >> On Sat, 31 Oct 2009 18:39:52 +0900
> >> Naohiro Ooiwa <nooiwa@miraclelinux.com> wrote:
> >>
> >> Does this work with Wake On Lan?
> >
> > Yes, it works WOL.
>
> Sorry, I made a mistake.
> The WOL doesn't work when my patch applied to kernel.
> I wasn't myself.
>
> I consider the WOL and I will resent the patch.
> Thank you for your point.
>
>
> thanks,
> Naohiro Ooiwa
Good, thank you for checking. I like the idea of powering down the
PHY. Some of the drivers have shutdown hooks to power PHY back on
during shutdown to enable WOL.
--
^ permalink raw reply
* Re: [PATCHv8 0/3] vhost: a kernel-level virtio server
From: Michael S. Tsirkin @ 2009-11-04 16:23 UTC (permalink / raw)
To: Gregory Haskins
Cc: netdev, virtualization, kvm, linux-kernel, mingo, linux-mm, akpm,
hpa, Rusty Russell, s.hetze, Daniel Walker, Eric Dumazet
In-Reply-To: <4AF1A587.8000509@gmail.com>
On Wed, Nov 04, 2009 at 11:02:15AM -0500, Gregory Haskins wrote:
> Michael S. Tsirkin wrote:
> > Ok, I think I've addressed all comments so far here.
> > Rusty, I'd like this to go into linux-next, through your tree, and
> > hopefully 2.6.33. What do you think?
>
> I think the benchmark data is a prerequisite for merge consideration, IMO.
Shirley Ma was kind enough to send me some measurement results showing
how kernel level acceleration helps speed up you can find them here:
http://www.linux-kvm.org/page/VhostNet
Generally, I think that merging should happen *before* agressive
benchmarking/performance tuning: otherwise there is very substancial
risk that what is an optimization in one setup hurts performance in
another one. When code is upstream, people can bisect to debug
regressions. Another good reason is that I can stop spending time
rebasing and start profiling.
> Do you have anything for us to look at?
For guest to host, compared to latest qemu with userspace virtio
backend, latency drops by a factor of 6, bandwidth doubles, cpu
utilization drops slightly :)
> I think comparison that show the following are of interest:
>
> throughput (e.g. netperf::TCP_STREAM): guest->host, guest->host->guest,
> guest->host->remote, host->remote, remote->host->guest
>
> latency (e.g. netperf::UDP_RR): same conditions as throughput
>
> cpu-utilization
>
> others?
>
> Ideally, this should be at least between upstream virtio and vhost.
> Bonus points if you include venet as well.
And vmxnet3 :)
> Kind regards,
> -Greg
>
--
MST
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply
* Re: HTB accuracy on 10GbE
From: Ryousei Takano @ 2009-11-04 16:31 UTC (permalink / raw)
To: Eric Dumazet
Cc: Stephen Hemminger, Patrick McHardy, Linux Netdev List,
takano-ryousei
In-Reply-To: <4AF1660E.1080401@gmail.com>
Hi Eric,
Thanks for your suggestion.
On Wed, Nov 4, 2009 at 8:31 PM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> Ryousei Takano a écrit :
>
>>
>> I tried iperf with 60 seconds samples. I got the almost same result.
>>
>> Here is the result:
>> sender receiver
>> 1.000 1.00 1.00
>> 2.000 2.01 2.01
>> 3.000 3.03 3.02
>> 4.000 4.07 4.07
>> 5.000 5.05 5.05
>> 6.000 6.16 6.16
>> 7.000 7.22 7.22
>> 8.000 8.15 8.15
>> 9.000 9.23 9.23
>> 9.900 9.69 9.69
>>
>
> One thing to consider is the estimation error in qdisc_l2t(), rate table has only 256 slots
>
> static inline u32 qdisc_l2t(struct qdisc_rate_table* rtab, unsigned int pktlen)
> {
> int slot = pktlen + rtab->rate.cell_align + rtab->rate.overhead;
> if (slot < 0)
> slot = 0;
> slot >>= rtab->rate.cell_log;
> if (slot > 255)
> return (rtab->data[255]*(slot >> 8) + rtab->data[slot & 0xFF]);
> return rtab->data[slot];
> }
>
>
> Maybe you can try changing class mtu to 40000 instead of 9000, and quantum to 60000 too
>
> tc class add dev $DEV parent 1: classid 1:1 htb rate ${rate}mbit mtu 40000 quantum 60000
>
> (because your tcp stack sends large buffers ( ~ 60000 bytes) as your NIC can offload tcp segmentation)
>
>
You are right!
I am using TSO. The myri10ge driver is passing 64KB packets to the NIC.
I changed the class mtu parameter to 64000 instead of 9000.
Here is the result:
1.000 1.00
2.000 2.01
3.000 2.99
4.000 4.01
5.000 5.01
6.000 6.04
7.000 7.06
8.000 8.09
9.000 9.11
9.900 9.64
It's not so bad!
For more information, I updated the results on my page.
Best regards,
Ryousei
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox