* [dpdk-dev] [PATCH 0/6] Ethernet driver enhancements
@ 2014-05-14 18:55 Stephen Hemminger
2014-05-14 18:55 ` [dpdk-dev] [PATCH 1/6] Subjec: ethdev: add macro to cover all checksum flags Stephen Hemminger
` (5 more replies)
0 siblings, 6 replies; 11+ messages in thread
From: Stephen Hemminger @ 2014-05-14 18:55 UTC (permalink / raw)
To: dev
Includes replacing existing VMXNET3 driver. Our driver was independently
developed based on the OpenSolaris driver (rather than FreeBSD).
After evaluation, it wasn't worth fixing the original code; wholesale
replacement was lower risk since the driver had been used through
the whole product and QA cycle.
^ permalink raw reply [flat|nested] 11+ messages in thread
* [dpdk-dev] [PATCH 1/6] Subjec: ethdev: add macro to cover all checksum flags
2014-05-14 18:55 [dpdk-dev] [PATCH 0/6] Ethernet driver enhancements Stephen Hemminger
@ 2014-05-14 18:55 ` Stephen Hemminger
2014-05-14 18:55 ` [dpdk-dev] [PATCH 2/6] Subjet: ethdev: add common code to atomicly access link Stephen Hemminger
` (4 subsequent siblings)
5 siblings, 0 replies; 11+ messages in thread
From: Stephen Hemminger @ 2014-05-14 18:55 UTC (permalink / raw)
To: dev; +Cc: Stephen Hemminger
[-- Attachment #1: txq_flags_nochecksum.patch --]
[-- Type: text/plain, Size: 1025 bytes --]
Convience to allow for devices that handle no checksum offload,
but do do VLAN offload.
Signed-off-by: Stephen Hemminger <shemming@brocade.com>
--- a/lib/librte_ether/rte_ethdev.h 2014-05-14 11:26:51.817382768 -0700
+++ b/lib/librte_ether/rte_ethdev.h 2014-05-14 11:26:51.817382768 -0700
@@ -510,9 +510,12 @@ struct rte_eth_rxconf {
#define ETH_TXQ_FLAGS_NOXSUMSCTP 0x0200 /**< disable SCTP checksum offload */
#define ETH_TXQ_FLAGS_NOXSUMUDP 0x0400 /**< disable UDP checksum offload */
#define ETH_TXQ_FLAGS_NOXSUMTCP 0x0800 /**< disable TCP checksum offload */
-#define ETH_TXQ_FLAGS_NOOFFLOADS \
- (ETH_TXQ_FLAGS_NOVLANOFFL | ETH_TXQ_FLAGS_NOXSUMSCTP | \
- ETH_TXQ_FLAGS_NOXSUMUDP | ETH_TXQ_FLAGS_NOXSUMTCP)
+#define ETH_TXQ_FLAGS_NOXSUMS (ETH_TXQ_FLAGS_NOXSUMSCTP | \
+ ETH_TXQ_FLAGS_NOXSUMUDP | \
+ ETH_TXQ_FLAGS_NOXSUMTCP)
+#define ETH_TXQ_FLAGS_NOOFFLOADS (ETH_TXQ_FLAGS_NOVLANOFFL | \
+ ETH_TXQ_FLAGS_NOXSUMS)
+
/**
* A structure used to configure a TX ring of an Ethernet port.
*/
^ permalink raw reply [flat|nested] 11+ messages in thread
* [dpdk-dev] [PATCH 2/6] Subjet: ethdev: add common code to atomicly access link
2014-05-14 18:55 [dpdk-dev] [PATCH 0/6] Ethernet driver enhancements Stephen Hemminger
2014-05-14 18:55 ` [dpdk-dev] [PATCH 1/6] Subjec: ethdev: add macro to cover all checksum flags Stephen Hemminger
@ 2014-05-14 18:55 ` Stephen Hemminger
2014-05-15 8:43 ` Ivan Boule
2014-05-14 18:55 ` [dpdk-dev] [PATCH 3/6] ixgbe: use eth_dev_{get,set}_link Stephen Hemminger
` (3 subsequent siblings)
5 siblings, 1 reply; 11+ messages in thread
From: Stephen Hemminger @ 2014-05-14 18:55 UTC (permalink / raw)
To: dev; +Cc: Stephen Hemminger
[-- Attachment #1: eth-dev-link.patch --]
[-- Type: text/plain, Size: 1419 bytes --]
Many drivers copy/paste same code to atomicly access link information.
Add functions to do this instead.
Signed-off-by: Stephen Hemminger <shemming@brocade.com>
--- a/lib/librte_ether/rte_ethdev.h 2014-05-14 11:27:07.409461720 -0700
+++ b/lib/librte_ether/rte_ethdev.h 2014-05-14 11:27:07.409461720 -0700
@@ -1704,6 +1704,45 @@ extern void rte_eth_link_get_nowait(uint
struct rte_eth_link *link);
/**
+ * @internal
+ * Retrieve the link status from device in atomic fashion.
+ */
+static inline void rte_eth_dev_get_link(struct rte_eth_dev *dev,
+ struct rte_eth_link *link)
+{
+ int64_t val;
+
+ val = rte_atomic64_read((rte_atomic64_t *)&dev->data->dev_link);
+
+ memcpy(link, &val, sizeof(struct rte_eth_link));
+}
+
+/**
+ * @internal
+ * Set the link status from device in atomic fashion.
+ * Returns non-zero on success; 0 on failure
+ * Callers should retry on failure
+ */
+static inline void rte_eth_dev_set_link(struct rte_eth_dev *dev,
+ struct rte_eth_link *link)
+{
+ rte_atomic64_t *dst = (rte_atomic64_t *)&dev->data->dev_link;
+
+ rte_atomic64_set(dst, *(int64_t *)link);
+}
+
+/**
+ * @internal
+ * Clear link status.
+ */
+static inline void rte_eth_dev_clear_link(struct rte_eth_dev *dev)
+{
+ rte_atomic64_t *dst = (rte_atomic64_t *)&dev->data->dev_link;
+
+ rte_atomic64_set(dst, 0);
+}
+
+/**
* Retrieve the general I/O statistics of an Ethernet device.
*
* @param port_id
^ permalink raw reply [flat|nested] 11+ messages in thread
* [dpdk-dev] [PATCH 3/6] ixgbe: use eth_dev_{get,set}_link
2014-05-14 18:55 [dpdk-dev] [PATCH 0/6] Ethernet driver enhancements Stephen Hemminger
2014-05-14 18:55 ` [dpdk-dev] [PATCH 1/6] Subjec: ethdev: add macro to cover all checksum flags Stephen Hemminger
2014-05-14 18:55 ` [dpdk-dev] [PATCH 2/6] Subjet: ethdev: add common code to atomicly access link Stephen Hemminger
@ 2014-05-14 18:55 ` Stephen Hemminger
2014-05-14 18:55 ` [dpdk-dev] [PATCH 4/6] e1000: " Stephen Hemminger
` (2 subsequent siblings)
5 siblings, 0 replies; 11+ messages in thread
From: Stephen Hemminger @ 2014-05-14 18:55 UTC (permalink / raw)
To: dev; +Cc: Stephen Hemminger
[-- Attachment #1: ixgbe-dev-link.patch --]
[-- Type: text/plain, Size: 4117 bytes --]
New common code can be used, eliminating driver code.
Signed-off-by: Stephen Hemminger <shemming@brocade.com>
--- a/lib/librte_pmd_ixgbe/ixgbe_ethdev.c 2014-05-14 11:26:30.381274255 -0700
+++ b/lib/librte_pmd_ixgbe/ixgbe_ethdev.c 2014-05-14 11:37:51.319893900 -0700
@@ -56,7 +56,6 @@
#include <rte_alarm.h>
#include <rte_ether.h>
#include <rte_ethdev.h>
-#include <rte_atomic.h>
#include <rte_malloc.h>
#include "ixgbe_logs.h"
@@ -325,58 +324,6 @@ static struct eth_dev_ops ixgbevf_eth_de
.tx_queue_release = ixgbe_dev_tx_queue_release,
};
-/**
- * Atomically reads the link status information from global
- * structure rte_eth_dev.
- *
- * @param dev
- * - Pointer to the structure rte_eth_dev to read from.
- * - Pointer to the buffer to be saved with the link status.
- *
- * @return
- * - On success, zero.
- * - On failure, negative value.
- */
-static inline int
-rte_ixgbe_dev_atomic_read_link_status(struct rte_eth_dev *dev,
- struct rte_eth_link *link)
-{
- struct rte_eth_link *dst = link;
- struct rte_eth_link *src = &(dev->data->dev_link);
-
- if (rte_atomic64_cmpset((uint64_t *)dst, *(uint64_t *)dst,
- *(uint64_t *)src) == 0)
- return -1;
-
- return 0;
-}
-
-/**
- * Atomically writes the link status information into global
- * structure rte_eth_dev.
- *
- * @param dev
- * - Pointer to the structure rte_eth_dev to read from.
- * - Pointer to the buffer to be saved with the link status.
- *
- * @return
- * - On success, zero.
- * - On failure, negative value.
- */
-static inline int
-rte_ixgbe_dev_atomic_write_link_status(struct rte_eth_dev *dev,
- struct rte_eth_link *link)
-{
- struct rte_eth_link *dst = &(dev->data->dev_link);
- struct rte_eth_link *src = link;
-
- if (rte_atomic64_cmpset((uint64_t *)dst, *(uint64_t *)dst,
- *(uint64_t *)src) == 0)
- return -1;
-
- return 0;
-}
-
/*
* This function is the same as ixgbe_is_sfp() in ixgbe/ixgbe.h.
*/
@@ -1453,7 +1400,7 @@ ixgbe_dev_stop(struct rte_eth_dev *dev)
/* Clear recorded link status */
memset(&link, 0, sizeof(link));
- rte_ixgbe_dev_atomic_write_link_status(dev, &link);
+ rte_eth_dev_set_link(dev, &link);
}
/*
@@ -1764,11 +1711,8 @@ ixgbe_dev_link_update(struct rte_eth_dev
int link_up;
int diag;
- link.link_status = 0;
- link.link_speed = 0;
- link.link_duplex = 0;
- memset(&old, 0, sizeof(old));
- rte_ixgbe_dev_atomic_read_link_status(dev, &old);
+ memset(&link, 0, sizeof(link));
+ rte_eth_dev_get_link(dev, &old);
/* check if it needs to wait to complete, if lsc interrupt is enabled */
if (wait_to_complete == 0 || dev->data->dev_conf.intr_conf.lsc != 0)
@@ -1778,14 +1722,14 @@ ixgbe_dev_link_update(struct rte_eth_dev
if (diag != 0) {
link.link_speed = ETH_LINK_SPEED_100;
link.link_duplex = ETH_LINK_HALF_DUPLEX;
- rte_ixgbe_dev_atomic_write_link_status(dev, &link);
+ rte_eth_dev_set_link(dev, &link);
if (link.link_status == old.link_status)
return -1;
return 0;
}
if (link_up == 0) {
- rte_ixgbe_dev_atomic_write_link_status(dev, &link);
+ rte_eth_dev_set_link(dev, &link);
if (link.link_status == old.link_status)
return -1;
return 0;
@@ -1812,7 +1756,7 @@ ixgbe_dev_link_update(struct rte_eth_dev
link.link_speed = ETH_LINK_SPEED_10000;
break;
}
- rte_ixgbe_dev_atomic_write_link_status(dev, &link);
+ rte_eth_dev_set_link(dev, &link);
if (link.link_status == old.link_status)
return -1;
@@ -1946,8 +1890,7 @@ ixgbe_dev_link_status_print(struct rte_e
{
struct rte_eth_link link;
- memset(&link, 0, sizeof(link));
- rte_ixgbe_dev_atomic_read_link_status(dev, &link);
+ rte_eth_dev_get_link(dev, &link);
if (link.link_status) {
PMD_INIT_LOG(INFO, "Port %d: Link Up - speed %u Mbps - %s",
(int)(dev->data->port_id),
@@ -1993,8 +1936,7 @@ ixgbe_dev_interrupt_action(struct rte_et
if (intr->flags & IXGBE_FLAG_NEED_LINK_UPDATE) {
/* get the link status before link update, for predicting later */
- memset(&link, 0, sizeof(link));
- rte_ixgbe_dev_atomic_read_link_status(dev, &link);
+ rte_eth_dev_get_link(dev, &link);
ixgbe_dev_link_update(dev, 0);
^ permalink raw reply [flat|nested] 11+ messages in thread
* [dpdk-dev] [PATCH 4/6] e1000: use eth_dev_{get,set}_link
2014-05-14 18:55 [dpdk-dev] [PATCH 0/6] Ethernet driver enhancements Stephen Hemminger
` (2 preceding siblings ...)
2014-05-14 18:55 ` [dpdk-dev] [PATCH 3/6] ixgbe: use eth_dev_{get,set}_link Stephen Hemminger
@ 2014-05-14 18:55 ` Stephen Hemminger
2014-05-14 18:55 ` [dpdk-dev] [PATCH 5/6] ether: allow setting mac address Stephen Hemminger
2014-05-14 18:55 ` [dpdk-dev] [PATCH 6/6] vmxnet3: replace Intel driver Stephen Hemminger
5 siblings, 0 replies; 11+ messages in thread
From: Stephen Hemminger @ 2014-05-14 18:55 UTC (permalink / raw)
To: dev; +Cc: Stephen Hemminger
[-- Attachment #1: igb-dev-link.patch --]
[-- Type: text/plain, Size: 5914 bytes --]
New common code can be used, eliminating driver code.
Signed-off-by: Stephen Hemminger <shemming@brocade.com>
--- a/lib/librte_pmd_e1000/em_ethdev.c 2014-05-14 11:26:30.381274255 -0700
+++ b/lib/librte_pmd_e1000/em_ethdev.c 2014-05-14 11:42:25.976657022 -0700
@@ -157,58 +157,6 @@ static struct eth_dev_ops eth_em_ops = {
.mac_addr_remove = eth_em_rar_clear,
};
-/**
- * Atomically reads the link status information from global
- * structure rte_eth_dev.
- *
- * @param dev
- * - Pointer to the structure rte_eth_dev to read from.
- * - Pointer to the buffer to be saved with the link status.
- *
- * @return
- * - On success, zero.
- * - On failure, negative value.
- */
-static inline int
-rte_em_dev_atomic_read_link_status(struct rte_eth_dev *dev,
- struct rte_eth_link *link)
-{
- struct rte_eth_link *dst = link;
- struct rte_eth_link *src = &(dev->data->dev_link);
-
- if (rte_atomic64_cmpset((uint64_t *)dst, *(uint64_t *)dst,
- *(uint64_t *)src) == 0)
- return -1;
-
- return 0;
-}
-
-/**
- * Atomically writes the link status information into global
- * structure rte_eth_dev.
- *
- * @param dev
- * - Pointer to the structure rte_eth_dev to read from.
- * - Pointer to the buffer to be saved with the link status.
- *
- * @return
- * - On success, zero.
- * - On failure, negative value.
- */
-static inline int
-rte_em_dev_atomic_write_link_status(struct rte_eth_dev *dev,
- struct rte_eth_link *link)
-{
- struct rte_eth_link *dst = &(dev->data->dev_link);
- struct rte_eth_link *src = link;
-
- if (rte_atomic64_cmpset((uint64_t *)dst, *(uint64_t *)dst,
- *(uint64_t *)src) == 0)
- return -1;
-
- return 0;
-}
-
static int
eth_em_dev_init(__attribute__((unused)) struct eth_driver *eth_drv,
struct rte_eth_dev *eth_dev)
@@ -597,7 +545,7 @@ eth_em_stop(struct rte_eth_dev *dev)
/* clear the recorded link status */
memset(&link, 0, sizeof(link));
- rte_em_dev_atomic_write_link_status(dev, &link);
+ rte_eth_dev_set_link(dev, &link);
}
static void
@@ -916,8 +864,8 @@ eth_em_link_update(struct rte_eth_dev *d
break;
rte_delay_ms(EM_LINK_UPDATE_CHECK_INTERVAL);
}
- memset(&link, 0, sizeof(link));
- rte_em_dev_atomic_read_link_status(dev, &link);
+
+ rte_eth_dev_get_link(dev, &link);
old = link;
/* Now we check if a transition has happened */
@@ -930,7 +878,7 @@ eth_em_link_update(struct rte_eth_dev *d
link.link_duplex = 0;
link.link_status = 0;
}
- rte_em_dev_atomic_write_link_status(dev, &link);
+ rte_eth_dev_set_link(dev, &link);
/* not changed */
if (old.link_status == link.link_status)
@@ -1281,8 +1229,7 @@ eth_em_interrupt_action(struct rte_eth_d
if (ret < 0)
return 0;
- memset(&link, 0, sizeof(link));
- rte_em_dev_atomic_read_link_status(dev, &link);
+ rte_eth_dev_get_link(dev, &link)
if (link.link_status) {
PMD_INIT_LOG(INFO,
" Port %d: Link Up - speed %u Mbps - %s\n",
--- a/lib/librte_pmd_e1000/igb_ethdev.c 2014-05-14 11:26:30.381274255 -0700
+++ b/lib/librte_pmd_e1000/igb_ethdev.c 2014-05-14 11:42:59.512754542 -0700
@@ -215,58 +215,6 @@ static struct eth_dev_ops igbvf_eth_dev_
.tx_queue_release = eth_igb_tx_queue_release,
};
-/**
- * Atomically reads the link status information from global
- * structure rte_eth_dev.
- *
- * @param dev
- * - Pointer to the structure rte_eth_dev to read from.
- * - Pointer to the buffer to be saved with the link status.
- *
- * @return
- * - On success, zero.
- * - On failure, negative value.
- */
-static inline int
-rte_igb_dev_atomic_read_link_status(struct rte_eth_dev *dev,
- struct rte_eth_link *link)
-{
- struct rte_eth_link *dst = link;
- struct rte_eth_link *src = &(dev->data->dev_link);
-
- if (rte_atomic64_cmpset((uint64_t *)dst, *(uint64_t *)dst,
- *(uint64_t *)src) == 0)
- return -1;
-
- return 0;
-}
-
-/**
- * Atomically writes the link status information into global
- * structure rte_eth_dev.
- *
- * @param dev
- * - Pointer to the structure rte_eth_dev to read from.
- * - Pointer to the buffer to be saved with the link status.
- *
- * @return
- * - On success, zero.
- * - On failure, negative value.
- */
-static inline int
-rte_igb_dev_atomic_write_link_status(struct rte_eth_dev *dev,
- struct rte_eth_link *link)
-{
- struct rte_eth_link *dst = &(dev->data->dev_link);
- struct rte_eth_link *src = link;
-
- if (rte_atomic64_cmpset((uint64_t *)dst, *(uint64_t *)dst,
- *(uint64_t *)src) == 0)
- return -1;
-
- return 0;
-}
-
static inline void
igb_intr_enable(struct rte_eth_dev *dev)
{
@@ -864,7 +812,7 @@ eth_igb_stop(struct rte_eth_dev *dev)
/* clear the recorded link status */
memset(&link, 0, sizeof(link));
- rte_igb_dev_atomic_write_link_status(dev, &link);
+ rte_eth_dev_set_link(dev, &link);
}
static void
@@ -890,7 +838,7 @@ eth_igb_close(struct rte_eth_dev *dev)
igb_dev_clear_queues(dev);
memset(&link, 0, sizeof(link));
- rte_igb_dev_atomic_write_link_status(dev, &link);
+ rte_eth_dev_set_link(dev, &link);
}
static int
@@ -1312,8 +1260,8 @@ eth_igb_link_update(struct rte_eth_dev *
break;
rte_delay_ms(IGB_LINK_UPDATE_CHECK_INTERVAL);
}
- memset(&link, 0, sizeof(link));
- rte_igb_dev_atomic_read_link_status(dev, &link);
+
+ rte_eth_dev_get_link(dev, &link);
old = link;
/* Now we check if a transition has happened */
@@ -1326,7 +1274,7 @@ eth_igb_link_update(struct rte_eth_dev *
link.link_duplex = 0;
link.link_status = 0;
}
- rte_igb_dev_atomic_write_link_status(dev, &link);
+ rte_eth_dev_set_link(dev, &link);
/* not changed */
if (old.link_status == link.link_status)
@@ -1718,8 +1666,7 @@ eth_igb_interrupt_action(struct rte_eth_
if (ret < 0)
return 0;
- memset(&link, 0, sizeof(link));
- rte_igb_dev_atomic_read_link_status(dev, &link);
+ rte_eth_dev_get_link(dev, &link);
if (link.link_status) {
PMD_INIT_LOG(INFO,
" Port %d: Link Up - speed %u Mbps - %s\n",
^ permalink raw reply [flat|nested] 11+ messages in thread
* [dpdk-dev] [PATCH 5/6] ether: allow setting mac address
2014-05-14 18:55 [dpdk-dev] [PATCH 0/6] Ethernet driver enhancements Stephen Hemminger
` (3 preceding siblings ...)
2014-05-14 18:55 ` [dpdk-dev] [PATCH 4/6] e1000: " Stephen Hemminger
@ 2014-05-14 18:55 ` Stephen Hemminger
2014-05-15 9:04 ` Ivan Boule
2014-05-14 18:55 ` [dpdk-dev] [PATCH 6/6] vmxnet3: replace Intel driver Stephen Hemminger
5 siblings, 1 reply; 11+ messages in thread
From: Stephen Hemminger @ 2014-05-14 18:55 UTC (permalink / raw)
To: dev; +Cc: Stephen Hemminger
[-- Attachment #1: mac-addr-set.patch --]
[-- Type: text/plain, Size: 3233 bytes --]
Allow setting the default Ethernet address used on device.
The underlying drivers allow it but the DPDK was blocking any
attempts to change Ethernet address on device.
For most devices, this is just the same as filling in address
in mac address table entry 0, but for some devices this requires
special override.
Signed-off-by: Stephen Hemminger <shemming@brocade.com>
--- a/lib/librte_ether/rte_ethdev.h 2014-05-14 11:44:27.373014227 -0700
+++ b/lib/librte_ether/rte_ethdev.h 2014-05-14 11:44:27.373014227 -0700
@@ -982,6 +982,11 @@ typedef void (*eth_mac_addr_add_t)(struc
struct ether_addr *mac_addr,
uint32_t index,
uint32_t vmdq);
+/**< @internal Change default MAC address */
+
+typedef void (*eth_mac_addr_set_t)(struct rte_eth_dev *dev,
+ struct ether_addr *mac_addr);
+
/**< @internal Set a MAC address into Receive Address Address Register */
typedef int (*eth_uc_hash_table_set_t)(struct rte_eth_dev *dev,
@@ -1114,6 +1119,7 @@ struct eth_dev_ops {
priority_flow_ctrl_set_t priority_flow_ctrl_set; /**< Setup priority flow control.*/
eth_mac_addr_remove_t mac_addr_remove; /**< Remove MAC address */
eth_mac_addr_add_t mac_addr_add; /**< Add a MAC address */
+ eth_mac_addr_set_t mac_addr_set; /**< Change default MAC address */
eth_uc_hash_table_set_t uc_hash_table_set; /**< Set Unicast Table Array */
eth_uc_all_hash_table_set_t uc_all_hash_table_set; /**< Set Unicast hash bitmap */
eth_mirror_rule_set_t mirror_rule_set; /**< Add a traffic mirror rule.*/
@@ -2538,6 +2544,21 @@ int rte_eth_dev_mac_addr_add(uint8_t por
int rte_eth_dev_mac_addr_remove(uint8_t port, struct ether_addr *mac_addr);
/**
+ * Change the default MAC address
+ *
+ * @param port
+ * The port identifier of the Ethernet device.
+ * @param mac_addr
+ * MAC address to remove.
+ * @return
+ * - (0) if successfully added or *mac_addr" was already added.
+ * - (-ENOTSUP) if hardware doesn't support this feature.
+ * - (-ENODEV) if *port* is invalid.
+ * - (-EINVAL) if MAC address is invalid.
+ */
+int rte_eth_dev_macaddr_set(uint8_t port, struct ether_addr *mac_addr);
+
+/**
* Update Redirection Table(RETA) of Receive Side Scaling of Ethernet device.
*
* @param port
--- a/lib/librte_ether/rte_ethdev.c 2014-05-14 11:44:27.373014227 -0700
+++ b/lib/librte_ether/rte_ethdev.c 2014-05-14 11:44:27.373014227 -0700
@@ -1059,6 +1059,33 @@ rte_eth_macaddr_get(uint8_t port_id, str
}
int
+rte_eth_dev_macaddr_set(uint8_t port_id, struct ether_addr *addr)
+{
+ struct rte_eth_dev *dev;
+
+ if (port_id >= nb_ports) {
+ PMD_DEBUG_TRACE("Invalid port_id=%d\n", port_id);
+ return (-ENODEV);
+ }
+ if (is_zero_ether_addr(addr)) {
+ PMD_DEBUG_TRACE("port %d: Cannot add NULL MAC address\n", port_id);
+ return (-EINVAL);
+ }
+ dev = &rte_eth_devices[port_id];
+
+ if (*dev->dev_ops->mac_addr_set)
+ (*dev->dev_ops->mac_addr_set)(dev, addr);
+ else {
+ FUNC_PTR_OR_ERR_RET(*dev->dev_ops->mac_addr_add, -ENOTSUP);
+ (*dev->dev_ops->mac_addr_add)(dev, addr, 0, 0);
+ }
+
+ ether_addr_copy(addr, &dev->data->mac_addrs[0]);
+
+ return 0;
+}
+
+int
rte_eth_dev_vlan_filter(uint8_t port_id, uint16_t vlan_id, int on)
{
struct rte_eth_dev *dev;
^ permalink raw reply [flat|nested] 11+ messages in thread
* [dpdk-dev] [PATCH 6/6] vmxnet3: replace Intel driver
2014-05-14 18:55 [dpdk-dev] [PATCH 0/6] Ethernet driver enhancements Stephen Hemminger
` (4 preceding siblings ...)
2014-05-14 18:55 ` [dpdk-dev] [PATCH 5/6] ether: allow setting mac address Stephen Hemminger
@ 2014-05-14 18:55 ` Stephen Hemminger
2014-05-15 17:14 ` Thomas Monjalon
5 siblings, 1 reply; 11+ messages in thread
From: Stephen Hemminger @ 2014-05-14 18:55 UTC (permalink / raw)
To: dev; +Cc: Stephen Hemminger
[-- Attachment #1: vmxnet3-replace.patch --]
[-- Type: text/plain, Size: 172047 bytes --]
Replace the Intel VMXNET3 driver with the one developed by Brocade.
The Intel code was a missing features, and had style and other issues.
Major changes:
- supports multi-segment
- supports vlan offload
- per queue stats
Fixes:
- link update reports link change correctly
- allows reconfiguration of # of queues
(original code would have too small memzone).
Signed-off-by: Stephen Hemminger <shemming@brocade.com>
---
lib/librte_pmd_vmxnet3/Makefile | 23
lib/librte_pmd_vmxnet3/README | 20
lib/librte_pmd_vmxnet3/vmxnet3/README | 50
lib/librte_pmd_vmxnet3/vmxnet3/includeCheck.h | 40
lib/librte_pmd_vmxnet3/vmxnet3/upt1_defs.h | 117 -
lib/librte_pmd_vmxnet3/vmxnet3/vmware_pack_begin.h | 32
lib/librte_pmd_vmxnet3/vmxnet3/vmware_pack_end.h | 32
lib/librte_pmd_vmxnet3/vmxnet3/vmxnet3_defs.h | 751 -----------
lib/librte_pmd_vmxnet3/vmxnet3/vmxnet3_osdep.h | 48
lib/librte_pmd_vmxnet3/vmxnet3_defs.h | 717 ++++++++++
lib/librte_pmd_vmxnet3/vmxnet3_dev.h | 148 ++
lib/librte_pmd_vmxnet3/vmxnet3_ethdev.c | 1087 +++++++---------
lib/librte_pmd_vmxnet3/vmxnet3_ethdev.h | 187 --
lib/librte_pmd_vmxnet3/vmxnet3_logs.h | 11
lib/librte_pmd_vmxnet3/vmxnet3_ring.h | 176 --
lib/librte_pmd_vmxnet3/vmxnet3_rxtx.c | 1387 +++++++++------------
16 files changed, 2024 insertions(+), 2802 deletions(-)
--- a/lib/librte_pmd_vmxnet3/Makefile 2014-05-14 11:46:47.129439301 -0700
+++ b/lib/librte_pmd_vmxnet3/Makefile 2014-05-14 11:46:47.125439289 -0700
@@ -1,8 +1,8 @@
# BSD LICENSE
#
-# Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+# Copyright(c) 2012-2014 Brocade Computer Systems
# All rights reserved.
-#
+#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
@@ -39,25 +39,6 @@ LIB = librte_pmd_vmxnet3_uio.a
CFLAGS += -O3
CFLAGS += $(WERROR_FLAGS)
-ifeq ($(CC), icc)
-#
-# CFLAGS for icc
-#
-CFLAGS_LAD = -wd174 -wd593 -wd869 -wd981 -wd2259
-else
-#
-# CFLAGS for gcc
-#
-ifneq ($(shell test $(GCC_MAJOR_VERSION) -le 4 -a $(GCC_MINOR_VERSION) -le 3 && echo 1), 1)
-CFLAGS += -Wno-deprecated
-endif
-CFLAGS_LAD = -Wno-unused-parameter -Wno-unused-value
-CFLAGS_LAD += -Wno-strict-aliasing -Wno-format-extra-args
-
-endif
-
-VPATH += $(RTE_SDK)/lib/librte_pmd_vmxnet3/vmxnet3
-
#
# all source are stored in SRCS-y
#
--- a/lib/librte_pmd_vmxnet3/vmxnet3/README 2014-05-14 11:46:47.129439301 -0700
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,50 +0,0 @@
-..
- BSD LICENSE
-
- Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
- All rights reserved.
-
- Redistribution and use in source and binary forms, with or without
- modification, are permitted provided that the following conditions
- are met:
-
- * Redistributions of source code must retain the above copyright
- notice, this list of conditions and the following disclaimer.
- * Redistributions in binary form must reproduce the above copyright
- notice, this list of conditions and the following disclaimer in
- the documentation and/or other materials provided with the
- distribution.
- * Neither the name of Intel Corporation nor the names of its
- contributors may be used to endorse or promote products derived
- from this software without specific prior written permission.
-
- THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-
-Intel VMXNET3 driver
-===================
-
-This directory contains source code of FreeBSD VMXNET3 driver released by VMware.
-In which, upt1_defs.h and vmxnet3_defs.h is introduced without any change.
-The other 4 files: includeCheck.h, vmware_pack_begin.h, vmware_pack_end.h and vmxnet3_osdep.h
-are crated to adapt to the needs from above 2 files.
-
-Updating driver
-===============
-
-The following modifications have been made to this code to integrate it with the
-Intel DPDK:
-
-
--------------
-
-
--- a/lib/librte_pmd_vmxnet3/vmxnet3/includeCheck.h 2014-05-14 11:46:47.129439301 -0700
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,40 +0,0 @@
-/*-
- * BSD LICENSE
- *
- * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * * Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- * * Redistributions in binary form must reproduce the above copyright
- * notice, this list of conditions and the following disclaimer in
- * the documentation and/or other materials provided with the
- * distribution.
- * * Neither the name of Intel Corporation nor the names of its
- * contributors may be used to endorse or promote products derived
- * from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#ifndef _INCLUDECHECK_H
-#define _INCLUDECHECK_H
-
-#include "vmxnet3_osdep.h"
-
-#endif /* _INCLUDECHECK_H */
-
--- a/lib/librte_pmd_vmxnet3/vmxnet3/upt1_defs.h 2014-05-14 11:46:47.129439301 -0700
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,117 +0,0 @@
-/*********************************************************
- * Copyright (C) 2007 VMware, Inc. All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- * notice, this list of conditions and the following disclaimer in the
- * documentation and/or other materials provided with the distribution.
- *
- * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- *
- *********************************************************/
-
-/* upt1_defs.h
- *
- * Definitions for UPTv1
- *
- * Some of the defs are duplicated in vmkapi_net_upt.h, because
- * vmkapi_net_upt.h cannot distribute with OSS yet and vmkapi headers can
- * only include vmkapi headers. Make sure they are kept in sync!
- */
-
-#ifndef _UPT1_DEFS_H
-#define _UPT1_DEFS_H
-
-#define UPT1_MAX_TX_QUEUES 64
-#define UPT1_MAX_RX_QUEUES 64
-
-#define UPT1_MAX_INTRS (UPT1_MAX_TX_QUEUES + UPT1_MAX_RX_QUEUES)
-
-typedef
-#include "vmware_pack_begin.h"
-struct UPT1_TxStats {
- uint64 TSOPktsTxOK; /* TSO pkts post-segmentation */
- uint64 TSOBytesTxOK;
- uint64 ucastPktsTxOK;
- uint64 ucastBytesTxOK;
- uint64 mcastPktsTxOK;
- uint64 mcastBytesTxOK;
- uint64 bcastPktsTxOK;
- uint64 bcastBytesTxOK;
- uint64 pktsTxError;
- uint64 pktsTxDiscard;
-}
-#include "vmware_pack_end.h"
-UPT1_TxStats;
-
-typedef
-#include "vmware_pack_begin.h"
-struct UPT1_RxStats {
- uint64 LROPktsRxOK; /* LRO pkts */
- uint64 LROBytesRxOK; /* bytes from LRO pkts */
- /* the following counters are for pkts from the wire, i.e., pre-LRO */
- uint64 ucastPktsRxOK;
- uint64 ucastBytesRxOK;
- uint64 mcastPktsRxOK;
- uint64 mcastBytesRxOK;
- uint64 bcastPktsRxOK;
- uint64 bcastBytesRxOK;
- uint64 pktsRxOutOfBuf;
- uint64 pktsRxError;
-}
-#include "vmware_pack_end.h"
-UPT1_RxStats;
-
-/* interrupt moderation level */
-#define UPT1_IML_NONE 0 /* no interrupt moderation */
-#define UPT1_IML_HIGHEST 7 /* least intr generated */
-#define UPT1_IML_ADAPTIVE 8 /* adpative intr moderation */
-
-/* values for UPT1_RSSConf.hashFunc */
-#define UPT1_RSS_HASH_TYPE_NONE 0x0
-#define UPT1_RSS_HASH_TYPE_IPV4 0x01
-#define UPT1_RSS_HASH_TYPE_TCP_IPV4 0x02
-#define UPT1_RSS_HASH_TYPE_IPV6 0x04
-#define UPT1_RSS_HASH_TYPE_TCP_IPV6 0x08
-
-#define UPT1_RSS_HASH_FUNC_NONE 0x0
-#define UPT1_RSS_HASH_FUNC_TOEPLITZ 0x01
-
-#define UPT1_RSS_MAX_KEY_SIZE 40
-#define UPT1_RSS_MAX_IND_TABLE_SIZE 128
-
-typedef
-#include "vmware_pack_begin.h"
-struct UPT1_RSSConf {
- uint16 hashType;
- uint16 hashFunc;
- uint16 hashKeySize;
- uint16 indTableSize;
- uint8 hashKey[UPT1_RSS_MAX_KEY_SIZE];
- uint8 indTable[UPT1_RSS_MAX_IND_TABLE_SIZE];
-}
-#include "vmware_pack_end.h"
-UPT1_RSSConf;
-
-/* features */
-#define UPT1_F_RXCSUM 0x0001 /* rx csum verification */
-#define UPT1_F_RSS 0x0002
-#define UPT1_F_RXVLAN 0x0004 /* VLAN tag stripping */
-#define UPT1_F_LRO 0x0008
-
-#endif
--- a/lib/librte_pmd_vmxnet3/vmxnet3/vmware_pack_begin.h 2014-05-14 11:46:47.129439301 -0700
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,32 +0,0 @@
-/*-
- * BSD LICENSE
- *
- * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * * Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- * * Redistributions in binary form must reproduce the above copyright
- * notice, this list of conditions and the following disclaimer in
- * the documentation and/or other materials provided with the
- * distribution.
- * * Neither the name of Intel Corporation nor the names of its
- * contributors may be used to endorse or promote products derived
- * from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
--- a/lib/librte_pmd_vmxnet3/vmxnet3/vmware_pack_end.h 2014-05-14 11:46:47.129439301 -0700
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,32 +0,0 @@
-/*-
- * BSD LICENSE
- *
- * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * * Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- * * Redistributions in binary form must reproduce the above copyright
- * notice, this list of conditions and the following disclaimer in
- * the documentation and/or other materials provided with the
- * distribution.
- * * Neither the name of Intel Corporation nor the names of its
- * contributors may be used to endorse or promote products derived
- * from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
--- a/lib/librte_pmd_vmxnet3/vmxnet3/vmxnet3_defs.h 2014-05-14 11:46:47.129439301 -0700
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,751 +0,0 @@
-/*********************************************************
- * Copyright (C) 2007 VMware, Inc. All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- * notice, this list of conditions and the following disclaimer in the
- * documentation and/or other materials provided with the distribution.
- *
- * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- *
- *********************************************************/
-
-/*
- * vmxnet3_defs.h --
- *
- * Definitions shared by device emulation and guest drivers for
- * VMXNET3 NIC
- */
-
-#ifndef _VMXNET3_DEFS_H_
-#define _VMXNET3_DEFS_H_
-
-#define INCLUDE_ALLOW_USERLEVEL
-#define INCLUDE_ALLOW_VMKERNEL
-#define INCLUDE_ALLOW_DISTRIBUTE
-#define INCLUDE_ALLOW_VMKDRIVERS
-#define INCLUDE_ALLOW_VMCORE
-#define INCLUDE_ALLOW_MODULE
-#include "includeCheck.h"
-
-#include "upt1_defs.h"
-
-/* all registers are 32 bit wide */
-/* BAR 1 */
-#define VMXNET3_REG_VRRS 0x0 /* Vmxnet3 Revision Report Selection */
-#define VMXNET3_REG_UVRS 0x8 /* UPT Version Report Selection */
-#define VMXNET3_REG_DSAL 0x10 /* Driver Shared Address Low */
-#define VMXNET3_REG_DSAH 0x18 /* Driver Shared Address High */
-#define VMXNET3_REG_CMD 0x20 /* Command */
-#define VMXNET3_REG_MACL 0x28 /* MAC Address Low */
-#define VMXNET3_REG_MACH 0x30 /* MAC Address High */
-#define VMXNET3_REG_ICR 0x38 /* Interrupt Cause Register */
-#define VMXNET3_REG_ECR 0x40 /* Event Cause Register */
-
-#define VMXNET3_REG_WSAL 0xF00 /* Wireless Shared Address Lo */
-#define VMXNET3_REG_WSAH 0xF08 /* Wireless Shared Address Hi */
-#define VMXNET3_REG_WCMD 0xF18 /* Wireless Command */
-
-/* BAR 0 */
-#define VMXNET3_REG_IMR 0x0 /* Interrupt Mask Register */
-#define VMXNET3_REG_TXPROD 0x600 /* Tx Producer Index */
-#define VMXNET3_REG_RXPROD 0x800 /* Rx Producer Index for ring 1 */
-#define VMXNET3_REG_RXPROD2 0xA00 /* Rx Producer Index for ring 2 */
-
-#define VMXNET3_PT_REG_SIZE 4096 /* BAR 0 */
-#define VMXNET3_VD_REG_SIZE 4096 /* BAR 1 */
-
-/*
- * The two Vmxnet3 MMIO Register PCI BARs (BAR 0 at offset 10h and BAR 1 at
- * offset 14h) as well as the MSI-X BAR are combined into one PhysMem region:
- * <-VMXNET3_PT_REG_SIZE-><-VMXNET3_VD_REG_SIZE-><-VMXNET3_MSIX_BAR_SIZE-->
- * -------------------------------------------------------------------------
- * |Pass Thru Registers | Virtual Dev Registers | MSI-X Vector/PBA Table |
- * -------------------------------------------------------------------------
- * VMXNET3_MSIX_BAR_SIZE is defined in "vmxnet3Int.h"
- */
-#define VMXNET3_PHYSMEM_PAGES 4
-
-#define VMXNET3_REG_ALIGN 8 /* All registers are 8-byte aligned. */
-#define VMXNET3_REG_ALIGN_MASK 0x7
-
-/* I/O Mapped access to registers */
-#define VMXNET3_IO_TYPE_PT 0
-#define VMXNET3_IO_TYPE_VD 1
-#define VMXNET3_IO_ADDR(type, reg) (((type) << 24) | ((reg) & 0xFFFFFF))
-#define VMXNET3_IO_TYPE(addr) ((addr) >> 24)
-#define VMXNET3_IO_REG(addr) ((addr) & 0xFFFFFF)
-
-#ifndef __le16
-#define __le16 uint16
-#endif
-#ifndef __le32
-#define __le32 uint32
-#endif
-#ifndef __le64
-#define __le64 uint64
-#endif
-
-typedef enum {
- VMXNET3_CMD_FIRST_SET = 0xCAFE0000,
- VMXNET3_CMD_ACTIVATE_DEV = VMXNET3_CMD_FIRST_SET,
- VMXNET3_CMD_QUIESCE_DEV,
- VMXNET3_CMD_RESET_DEV,
- VMXNET3_CMD_UPDATE_RX_MODE,
- VMXNET3_CMD_UPDATE_MAC_FILTERS,
- VMXNET3_CMD_UPDATE_VLAN_FILTERS,
- VMXNET3_CMD_UPDATE_RSSIDT,
- VMXNET3_CMD_UPDATE_IML,
- VMXNET3_CMD_UPDATE_PMCFG,
- VMXNET3_CMD_UPDATE_FEATURE,
- VMXNET3_CMD_STOP_EMULATION,
- VMXNET3_CMD_LOAD_PLUGIN,
- VMXNET3_CMD_ACTIVATE_VF,
-
- VMXNET3_CMD_FIRST_GET = 0xF00D0000,
- VMXNET3_CMD_GET_QUEUE_STATUS = VMXNET3_CMD_FIRST_GET,
- VMXNET3_CMD_GET_STATS,
- VMXNET3_CMD_GET_LINK,
- VMXNET3_CMD_GET_PERM_MAC_LO,
- VMXNET3_CMD_GET_PERM_MAC_HI,
- VMXNET3_CMD_GET_DID_LO,
- VMXNET3_CMD_GET_DID_HI,
- VMXNET3_CMD_GET_DEV_EXTRA_INFO,
- VMXNET3_CMD_GET_CONF_INTR,
- VMXNET3_CMD_GET_ADAPTIVE_RING_INFO
-} Vmxnet3_Cmd;
-
-/* Adaptive Ring Info Flags */
-#define VMXNET3_DISABLE_ADAPTIVE_RING 1
-
-/*
- * Little Endian layout of bitfields -
- * Byte 0 : 7.....len.....0
- * Byte 1 : rsvd gen 13.len.8
- * Byte 2 : 5.msscof.0 ext1 dtype
- * Byte 3 : 13...msscof...6
- *
- * Big Endian layout of bitfields -
- * Byte 0: 13...msscof...6
- * Byte 1 : 5.msscof.0 ext1 dtype
- * Byte 2 : rsvd gen 13.len.8
- * Byte 3 : 7.....len.....0
- *
- * Thus, le32_to_cpu on the dword will allow the big endian driver to read
- * the bit fields correctly. And cpu_to_le32 will convert bitfields
- * bit fields written by big endian driver to format required by device.
- */
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_TxDesc {
- __le64 addr;
-
-#ifdef __BIG_ENDIAN_BITFIELD
- uint32 msscof:14; /* MSS, checksum offset, flags */
- uint32 ext1:1;
- uint32 dtype:1; /* descriptor type */
- uint32 rsvd:1;
- uint32 gen:1; /* generation bit */
- uint32 len:14;
-#else
- uint32 len:14;
- uint32 gen:1; /* generation bit */
- uint32 rsvd:1;
- uint32 dtype:1; /* descriptor type */
- uint32 ext1:1;
- uint32 msscof:14; /* MSS, checksum offset, flags */
-#endif /* __BIG_ENDIAN_BITFIELD */
-
-#ifdef __BIG_ENDIAN_BITFIELD
- uint32 tci:16; /* Tag to Insert */
- uint32 ti:1; /* VLAN Tag Insertion */
- uint32 ext2:1;
- uint32 cq:1; /* completion request */
- uint32 eop:1; /* End Of Packet */
- uint32 om:2; /* offload mode */
- uint32 hlen:10; /* header len */
-#else
- uint32 hlen:10; /* header len */
- uint32 om:2; /* offload mode */
- uint32 eop:1; /* End Of Packet */
- uint32 cq:1; /* completion request */
- uint32 ext2:1;
- uint32 ti:1; /* VLAN Tag Insertion */
- uint32 tci:16; /* Tag to Insert */
-#endif /* __BIG_ENDIAN_BITFIELD */
-}
-#include "vmware_pack_end.h"
-Vmxnet3_TxDesc;
-
-/* TxDesc.OM values */
-#define VMXNET3_OM_NONE 0
-#define VMXNET3_OM_CSUM 2
-#define VMXNET3_OM_TSO 3
-
-/* fields in TxDesc we access w/o using bit fields */
-#define VMXNET3_TXD_EOP_SHIFT 12
-#define VMXNET3_TXD_CQ_SHIFT 13
-#define VMXNET3_TXD_GEN_SHIFT 14
-#define VMXNET3_TXD_EOP_DWORD_SHIFT 3
-#define VMXNET3_TXD_GEN_DWORD_SHIFT 2
-
-#define VMXNET3_TXD_CQ (1 << VMXNET3_TXD_CQ_SHIFT)
-#define VMXNET3_TXD_EOP (1 << VMXNET3_TXD_EOP_SHIFT)
-#define VMXNET3_TXD_GEN (1 << VMXNET3_TXD_GEN_SHIFT)
-
-#define VMXNET3_TXD_GEN_SIZE 1
-#define VMXNET3_TXD_EOP_SIZE 1
-
-#define VMXNET3_HDR_COPY_SIZE 128
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_TxDataDesc {
- uint8 data[VMXNET3_HDR_COPY_SIZE];
-}
-#include "vmware_pack_end.h"
-Vmxnet3_TxDataDesc;
-
-#define VMXNET3_TCD_GEN_SHIFT 31
-#define VMXNET3_TCD_GEN_SIZE 1
-#define VMXNET3_TCD_TXIDX_SHIFT 0
-#define VMXNET3_TCD_TXIDX_SIZE 12
-#define VMXNET3_TCD_GEN_DWORD_SHIFT 3
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_TxCompDesc {
- uint32 txdIdx:12; /* Index of the EOP TxDesc */
- uint32 ext1:20;
-
- __le32 ext2;
- __le32 ext3;
-
- uint32 rsvd:24;
- uint32 type:7; /* completion type */
- uint32 gen:1; /* generation bit */
-}
-#include "vmware_pack_end.h"
-Vmxnet3_TxCompDesc;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_RxDesc {
- __le64 addr;
-
-#ifdef __BIG_ENDIAN_BITFIELD
- uint32 gen:1; /* Generation bit */
- uint32 rsvd:15;
- uint32 dtype:1; /* Descriptor type */
- uint32 btype:1; /* Buffer Type */
- uint32 len:14;
-#else
- uint32 len:14;
- uint32 btype:1; /* Buffer Type */
- uint32 dtype:1; /* Descriptor type */
- uint32 rsvd:15;
- uint32 gen:1; /* Generation bit */
-#endif
- __le32 ext1;
-}
-#include "vmware_pack_end.h"
-Vmxnet3_RxDesc;
-
-/* values of RXD.BTYPE */
-#define VMXNET3_RXD_BTYPE_HEAD 0 /* head only */
-#define VMXNET3_RXD_BTYPE_BODY 1 /* body only */
-
-/* fields in RxDesc we access w/o using bit fields */
-#define VMXNET3_RXD_BTYPE_SHIFT 14
-#define VMXNET3_RXD_GEN_SHIFT 31
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_RxCompDesc {
-#ifdef __BIG_ENDIAN_BITFIELD
- uint32 ext2:1;
- uint32 cnc:1; /* Checksum Not Calculated */
- uint32 rssType:4; /* RSS hash type used */
- uint32 rqID:10; /* rx queue/ring ID */
- uint32 sop:1; /* Start of Packet */
- uint32 eop:1; /* End of Packet */
- uint32 ext1:2;
- uint32 rxdIdx:12; /* Index of the RxDesc */
-#else
- uint32 rxdIdx:12; /* Index of the RxDesc */
- uint32 ext1:2;
- uint32 eop:1; /* End of Packet */
- uint32 sop:1; /* Start of Packet */
- uint32 rqID:10; /* rx queue/ring ID */
- uint32 rssType:4; /* RSS hash type used */
- uint32 cnc:1; /* Checksum Not Calculated */
- uint32 ext2:1;
-#endif /* __BIG_ENDIAN_BITFIELD */
-
- __le32 rssHash; /* RSS hash value */
-
-#ifdef __BIG_ENDIAN_BITFIELD
- uint32 tci:16; /* Tag stripped */
- uint32 ts:1; /* Tag is stripped */
- uint32 err:1; /* Error */
- uint32 len:14; /* data length */
-#else
- uint32 len:14; /* data length */
- uint32 err:1; /* Error */
- uint32 ts:1; /* Tag is stripped */
- uint32 tci:16; /* Tag stripped */
-#endif /* __BIG_ENDIAN_BITFIELD */
-
-
-#ifdef __BIG_ENDIAN_BITFIELD
- uint32 gen:1; /* generation bit */
- uint32 type:7; /* completion type */
- uint32 fcs:1; /* Frame CRC correct */
- uint32 frg:1; /* IP Fragment */
- uint32 v4:1; /* IPv4 */
- uint32 v6:1; /* IPv6 */
- uint32 ipc:1; /* IP Checksum Correct */
- uint32 tcp:1; /* TCP packet */
- uint32 udp:1; /* UDP packet */
- uint32 tuc:1; /* TCP/UDP Checksum Correct */
- uint32 csum:16;
-#else
- uint32 csum:16;
- uint32 tuc:1; /* TCP/UDP Checksum Correct */
- uint32 udp:1; /* UDP packet */
- uint32 tcp:1; /* TCP packet */
- uint32 ipc:1; /* IP Checksum Correct */
- uint32 v6:1; /* IPv6 */
- uint32 v4:1; /* IPv4 */
- uint32 frg:1; /* IP Fragment */
- uint32 fcs:1; /* Frame CRC correct */
- uint32 type:7; /* completion type */
- uint32 gen:1; /* generation bit */
-#endif /* __BIG_ENDIAN_BITFIELD */
-}
-#include "vmware_pack_end.h"
-Vmxnet3_RxCompDesc;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_RxCompDescExt {
- __le32 dword1;
- uint8 segCnt; /* Number of aggregated packets */
- uint8 dupAckCnt; /* Number of duplicate Acks */
- __le16 tsDelta; /* TCP timestamp difference */
- __le32 dword2[2];
-}
-#include "vmware_pack_end.h"
-Vmxnet3_RxCompDescExt;
-
-/* fields in RxCompDesc we access via Vmxnet3_GenericDesc.dword[3] */
-#define VMXNET3_RCD_TUC_SHIFT 16
-#define VMXNET3_RCD_IPC_SHIFT 19
-
-/* fields in RxCompDesc we access via Vmxnet3_GenericDesc.qword[1] */
-#define VMXNET3_RCD_TYPE_SHIFT 56
-#define VMXNET3_RCD_GEN_SHIFT 63
-
-/* csum OK for TCP/UDP pkts over IP */
-#define VMXNET3_RCD_CSUM_OK (1 << VMXNET3_RCD_TUC_SHIFT | 1 << VMXNET3_RCD_IPC_SHIFT)
-
-/* value of RxCompDesc.rssType */
-#define VMXNET3_RCD_RSS_TYPE_NONE 0
-#define VMXNET3_RCD_RSS_TYPE_IPV4 1
-#define VMXNET3_RCD_RSS_TYPE_TCPIPV4 2
-#define VMXNET3_RCD_RSS_TYPE_IPV6 3
-#define VMXNET3_RCD_RSS_TYPE_TCPIPV6 4
-
-/* a union for accessing all cmd/completion descriptors */
-typedef union Vmxnet3_GenericDesc {
- __le64 qword[2];
- __le32 dword[4];
- __le16 word[8];
- Vmxnet3_TxDesc txd;
- Vmxnet3_RxDesc rxd;
- Vmxnet3_TxCompDesc tcd;
- Vmxnet3_RxCompDesc rcd;
- Vmxnet3_RxCompDescExt rcdExt;
-} Vmxnet3_GenericDesc;
-
-#define VMXNET3_INIT_GEN 1
-
-/* Max size of a single tx buffer */
-#define VMXNET3_MAX_TX_BUF_SIZE (1 << 14)
-
-/* # of tx desc needed for a tx buffer size */
-#define VMXNET3_TXD_NEEDED(size) (((size) + VMXNET3_MAX_TX_BUF_SIZE - 1) / VMXNET3_MAX_TX_BUF_SIZE)
-
-/* max # of tx descs for a non-tso pkt */
-#define VMXNET3_MAX_TXD_PER_PKT 16
-
-/* Max size of a single rx buffer */
-#define VMXNET3_MAX_RX_BUF_SIZE ((1 << 14) - 1)
-/* Minimum size of a type 0 buffer */
-#define VMXNET3_MIN_T0_BUF_SIZE 128
-#define VMXNET3_MAX_CSUM_OFFSET 1024
-
-/* Ring base address alignment */
-#define VMXNET3_RING_BA_ALIGN 512
-#define VMXNET3_RING_BA_MASK (VMXNET3_RING_BA_ALIGN - 1)
-
-/* Ring size must be a multiple of 32 */
-#define VMXNET3_RING_SIZE_ALIGN 32
-#define VMXNET3_RING_SIZE_MASK (VMXNET3_RING_SIZE_ALIGN - 1)
-
-/* Max ring size */
-#define VMXNET3_TX_RING_MAX_SIZE 4096
-#define VMXNET3_TC_RING_MAX_SIZE 4096
-#define VMXNET3_RX_RING_MAX_SIZE 4096
-#define VMXNET3_RC_RING_MAX_SIZE 8192
-
-/* a list of reasons for queue stop */
-
-#define VMXNET3_ERR_NOEOP 0x80000000 /* cannot find the EOP desc of a pkt */
-#define VMXNET3_ERR_TXD_REUSE 0x80000001 /* reuse a TxDesc before tx completion */
-#define VMXNET3_ERR_BIG_PKT 0x80000002 /* too many TxDesc for a pkt */
-#define VMXNET3_ERR_DESC_NOT_SPT 0x80000003 /* descriptor type not supported */
-#define VMXNET3_ERR_SMALL_BUF 0x80000004 /* type 0 buffer too small */
-#define VMXNET3_ERR_STRESS 0x80000005 /* stress option firing in vmkernel */
-#define VMXNET3_ERR_SWITCH 0x80000006 /* mode switch failure */
-#define VMXNET3_ERR_TXD_INVALID 0x80000007 /* invalid TxDesc */
-
-/* completion descriptor types */
-#define VMXNET3_CDTYPE_TXCOMP 0 /* Tx Completion Descriptor */
-#define VMXNET3_CDTYPE_RXCOMP 3 /* Rx Completion Descriptor */
-#define VMXNET3_CDTYPE_RXCOMP_LRO 4 /* Rx Completion Descriptor for LRO */
-
-#define VMXNET3_GOS_BITS_UNK 0 /* unknown */
-#define VMXNET3_GOS_BITS_32 1
-#define VMXNET3_GOS_BITS_64 2
-
-#define VMXNET3_GOS_TYPE_UNK 0 /* unknown */
-#define VMXNET3_GOS_TYPE_LINUX 1
-#define VMXNET3_GOS_TYPE_WIN 2
-#define VMXNET3_GOS_TYPE_SOLARIS 3
-#define VMXNET3_GOS_TYPE_FREEBSD 4
-#define VMXNET3_GOS_TYPE_PXE 5
-
-/* All structures in DriverShared are padded to multiples of 8 bytes */
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_GOSInfo {
-#ifdef __BIG_ENDIAN_BITFIELD
- uint32 gosMisc: 10; /* other info about gos */
- uint32 gosVer: 16; /* gos version */
- uint32 gosType: 4; /* which guest */
- uint32 gosBits: 2; /* 32-bit or 64-bit? */
-#else
- uint32 gosBits: 2; /* 32-bit or 64-bit? */
- uint32 gosType: 4; /* which guest */
- uint32 gosVer: 16; /* gos version */
- uint32 gosMisc: 10; /* other info about gos */
-#endif /* __BIG_ENDIAN_BITFIELD */
-}
-#include "vmware_pack_end.h"
-Vmxnet3_GOSInfo;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_DriverInfo {
- __le32 version; /* driver version */
- Vmxnet3_GOSInfo gos;
- __le32 vmxnet3RevSpt; /* vmxnet3 revision supported */
- __le32 uptVerSpt; /* upt version supported */
-}
-#include "vmware_pack_end.h"
-Vmxnet3_DriverInfo;
-
-#define VMXNET3_REV1_MAGIC 0xbabefee1
-
-/*
- * QueueDescPA must be 128 bytes aligned. It points to an array of
- * Vmxnet3_TxQueueDesc followed by an array of Vmxnet3_RxQueueDesc.
- * The number of Vmxnet3_TxQueueDesc/Vmxnet3_RxQueueDesc are specified by
- * Vmxnet3_MiscConf.numTxQueues/numRxQueues, respectively.
- */
-#define VMXNET3_QUEUE_DESC_ALIGN 128
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_MiscConf {
- Vmxnet3_DriverInfo driverInfo;
- __le64 uptFeatures;
- __le64 ddPA; /* driver data PA */
- __le64 queueDescPA; /* queue descriptor table PA */
- __le32 ddLen; /* driver data len */
- __le32 queueDescLen; /* queue descriptor table len, in bytes */
- __le32 mtu;
- __le16 maxNumRxSG;
- uint8 numTxQueues;
- uint8 numRxQueues;
- __le32 reserved[4];
-}
-#include "vmware_pack_end.h"
-Vmxnet3_MiscConf;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_TxQueueConf {
- __le64 txRingBasePA;
- __le64 dataRingBasePA;
- __le64 compRingBasePA;
- __le64 ddPA; /* driver data */
- __le64 reserved;
- __le32 txRingSize; /* # of tx desc */
- __le32 dataRingSize; /* # of data desc */
- __le32 compRingSize; /* # of comp desc */
- __le32 ddLen; /* size of driver data */
- uint8 intrIdx;
- uint8 _pad[7];
-}
-#include "vmware_pack_end.h"
-Vmxnet3_TxQueueConf;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_RxQueueConf {
- __le64 rxRingBasePA[2];
- __le64 compRingBasePA;
- __le64 ddPA; /* driver data */
- __le64 reserved;
- __le32 rxRingSize[2]; /* # of rx desc */
- __le32 compRingSize; /* # of rx comp desc */
- __le32 ddLen; /* size of driver data */
- uint8 intrIdx;
- uint8 _pad[7];
-}
-#include "vmware_pack_end.h"
-Vmxnet3_RxQueueConf;
-
-enum vmxnet3_intr_mask_mode {
- VMXNET3_IMM_AUTO = 0,
- VMXNET3_IMM_ACTIVE = 1,
- VMXNET3_IMM_LAZY = 2
-};
-
-enum vmxnet3_intr_type {
- VMXNET3_IT_AUTO = 0,
- VMXNET3_IT_INTX = 1,
- VMXNET3_IT_MSI = 2,
- VMXNET3_IT_MSIX = 3
-};
-
-#define VMXNET3_MAX_TX_QUEUES 8
-#define VMXNET3_MAX_RX_QUEUES 16
-/* addition 1 for events */
-#define VMXNET3_MAX_INTRS 25
-
-/* value of intrCtrl */
-#define VMXNET3_IC_DISABLE_ALL 0x1 /* bit 0 */
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_IntrConf {
- Bool autoMask;
- uint8 numIntrs; /* # of interrupts */
- uint8 eventIntrIdx;
- uint8 modLevels[VMXNET3_MAX_INTRS]; /* moderation level for each intr */
- __le32 intrCtrl;
- __le32 reserved[2];
-}
-#include "vmware_pack_end.h"
-Vmxnet3_IntrConf;
-
-/* one bit per VLAN ID, the size is in the units of uint32 */
-#define VMXNET3_VFT_SIZE (4096 / (sizeof(uint32) * 8))
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_QueueStatus {
- Bool stopped;
- uint8 _pad[3];
- __le32 error;
-}
-#include "vmware_pack_end.h"
-Vmxnet3_QueueStatus;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_TxQueueCtrl {
- __le32 txNumDeferred;
- __le32 txThreshold;
- __le64 reserved;
-}
-#include "vmware_pack_end.h"
-Vmxnet3_TxQueueCtrl;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_RxQueueCtrl {
- Bool updateRxProd;
- uint8 _pad[7];
- __le64 reserved;
-}
-#include "vmware_pack_end.h"
-Vmxnet3_RxQueueCtrl;
-
-#define VMXNET3_RXM_UCAST 0x01 /* unicast only */
-#define VMXNET3_RXM_MCAST 0x02 /* multicast passing the filters */
-#define VMXNET3_RXM_BCAST 0x04 /* broadcast only */
-#define VMXNET3_RXM_ALL_MULTI 0x08 /* all multicast */
-#define VMXNET3_RXM_PROMISC 0x10 /* promiscuous */
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_RxFilterConf {
- __le32 rxMode; /* VMXNET3_RXM_xxx */
- __le16 mfTableLen; /* size of the multicast filter table */
- __le16 _pad1;
- __le64 mfTablePA; /* PA of the multicast filters table */
- __le32 vfTable[VMXNET3_VFT_SIZE]; /* vlan filter */
-}
-#include "vmware_pack_end.h"
-Vmxnet3_RxFilterConf;
-
-#define VMXNET3_PM_MAX_FILTERS 6
-#define VMXNET3_PM_MAX_PATTERN_SIZE 128
-#define VMXNET3_PM_MAX_MASK_SIZE (VMXNET3_PM_MAX_PATTERN_SIZE / 8)
-
-#define VMXNET3_PM_WAKEUP_MAGIC 0x01 /* wake up on magic pkts */
-#define VMXNET3_PM_WAKEUP_FILTER 0x02 /* wake up on pkts matching filters */
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_PM_PktFilter {
- uint8 maskSize;
- uint8 patternSize;
- uint8 mask[VMXNET3_PM_MAX_MASK_SIZE];
- uint8 pattern[VMXNET3_PM_MAX_PATTERN_SIZE];
- uint8 pad[6];
-}
-#include "vmware_pack_end.h"
-Vmxnet3_PM_PktFilter;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_PMConf {
- __le16 wakeUpEvents; /* VMXNET3_PM_WAKEUP_xxx */
- uint8 numFilters;
- uint8 pad[5];
- Vmxnet3_PM_PktFilter filters[VMXNET3_PM_MAX_FILTERS];
-}
-#include "vmware_pack_end.h"
-Vmxnet3_PMConf;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_VariableLenConfDesc {
- __le32 confVer;
- __le32 confLen;
- __le64 confPA;
-}
-#include "vmware_pack_end.h"
-Vmxnet3_VariableLenConfDesc;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_DSDevRead {
- /* read-only region for device, read by dev in response to a SET cmd */
- Vmxnet3_MiscConf misc;
- Vmxnet3_IntrConf intrConf;
- Vmxnet3_RxFilterConf rxFilterConf;
- Vmxnet3_VariableLenConfDesc rssConfDesc;
- Vmxnet3_VariableLenConfDesc pmConfDesc;
- Vmxnet3_VariableLenConfDesc pluginConfDesc;
-}
-#include "vmware_pack_end.h"
-Vmxnet3_DSDevRead;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_TxQueueDesc {
- Vmxnet3_TxQueueCtrl ctrl;
- Vmxnet3_TxQueueConf conf;
- /* Driver read after a GET command */
- Vmxnet3_QueueStatus status;
- UPT1_TxStats stats;
- uint8 _pad[88]; /* 128 aligned */
-}
-#include "vmware_pack_end.h"
-Vmxnet3_TxQueueDesc;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_RxQueueDesc {
- Vmxnet3_RxQueueCtrl ctrl;
- Vmxnet3_RxQueueConf conf;
- /* Driver read after a GET command */
- Vmxnet3_QueueStatus status;
- UPT1_RxStats stats;
- uint8 _pad[88]; /* 128 aligned */
-}
-#include "vmware_pack_end.h"
-Vmxnet3_RxQueueDesc;
-
-typedef
-#include "vmware_pack_begin.h"
-struct Vmxnet3_DriverShared {
- __le32 magic;
- __le32 pad; /* make devRead start at 64-bit boundaries */
- Vmxnet3_DSDevRead devRead;
- __le32 ecr;
- __le32 reserved[5];
-}
-#include "vmware_pack_end.h"
-Vmxnet3_DriverShared;
-
-#define VMXNET3_ECR_RQERR (1 << 0)
-#define VMXNET3_ECR_TQERR (1 << 1)
-#define VMXNET3_ECR_LINK (1 << 2)
-#define VMXNET3_ECR_DIC (1 << 3)
-#define VMXNET3_ECR_DEBUG (1 << 4)
-
-/* flip the gen bit of a ring */
-#define VMXNET3_FLIP_RING_GEN(gen) ((gen) = (gen) ^ 0x1)
-
-/* only use this if moving the idx won't affect the gen bit */
-#define VMXNET3_INC_RING_IDX_ONLY(idx, ring_size) \
-do {\
- (idx)++;\
- if (UNLIKELY((idx) == (ring_size))) {\
- (idx) = 0;\
- }\
-} while (0)
-
-#define VMXNET3_SET_VFTABLE_ENTRY(vfTable, vid) \
- vfTable[vid >> 5] |= (1 << (vid & 31))
-#define VMXNET3_CLEAR_VFTABLE_ENTRY(vfTable, vid) \
- vfTable[vid >> 5] &= ~(1 << (vid & 31))
-
-#define VMXNET3_VFTABLE_ENTRY_IS_SET(vfTable, vid) \
- ((vfTable[vid >> 5] & (1 << (vid & 31))) != 0)
-
-#define VMXNET3_MAX_MTU 9000
-#define VMXNET3_MIN_MTU 60
-
-#define VMXNET3_LINK_UP (10000 << 16 | 1) // 10 Gbps, up
-#define VMXNET3_LINK_DOWN 0
-
-#define VMXWIFI_DRIVER_SHARED_LEN 8192
-
-#define VMXNET3_DID_PASSTHRU 0xFFFF
-
-#endif /* _VMXNET3_DEFS_H_ */
--- a/lib/librte_pmd_vmxnet3/vmxnet3/vmxnet3_osdep.h 2014-05-14 11:46:47.129439301 -0700
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,48 +0,0 @@
-/*-
- * BSD LICENSE
- *
- * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * * Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- * * Redistributions in binary form must reproduce the above copyright
- * notice, this list of conditions and the following disclaimer in
- * the documentation and/or other materials provided with the
- * distribution.
- * * Neither the name of Intel Corporation nor the names of its
- * contributors may be used to endorse or promote products derived
- * from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#ifndef _VMXNET3_OSDEP_H
-#define _VMXNET3_OSDEP_H
-
-typedef uint64_t uint64;
-typedef uint32_t uint32;
-typedef uint16_t uint16;
-typedef uint8_t uint8;
-typedef int bool;
-typedef char Bool;
-
-#ifndef UNLIKELY
-#define UNLIKELY(x) __builtin_expect((x),0)
-#endif /* unlikely */
-
-#endif /* _VMXNET3_OSDEP_H */
--- a/lib/librte_pmd_vmxnet3/vmxnet3_ethdev.c 2014-05-14 11:46:47.129439301 -0700
+++ b/lib/librte_pmd_vmxnet3/vmxnet3_ethdev.c 2014-05-14 11:48:07.617690469 -0700
@@ -1,7 +1,7 @@
/*-
* BSD LICENSE
- *
- * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+ *
+ * Copyright (c) 2012-2014 Brocade Communications Systems, Inc.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
@@ -35,731 +35,660 @@
#include <stdio.h>
#include <errno.h>
#include <stdint.h>
-#include <string.h>
-#include <unistd.h>
#include <stdarg.h>
-#include <fcntl.h>
-#include <inttypes.h>
-#include <rte_byteorder.h>
-#include <rte_common.h>
-#include <rte_cycles.h>
+#include <string.h>
+#include <rte_common.h>
#include <rte_interrupts.h>
+#include <rte_byteorder.h>
#include <rte_log.h>
#include <rte_debug.h>
#include <rte_pci.h>
-#include <rte_atomic.h>
-#include <rte_branch_prediction.h>
+#include <rte_ether.h>
+#include <rte_ethdev.h>
#include <rte_memory.h>
#include <rte_memzone.h>
#include <rte_tailq.h>
#include <rte_eal.h>
-#include <rte_alarm.h>
-#include <rte_ether.h>
-#include <rte_ethdev.h>
#include <rte_atomic.h>
-#include <rte_string_fns.h>
+#include <rte_spinlock.h>
#include <rte_malloc.h>
+#include <rte_string_fns.h>
-#include "vmxnet3/vmxnet3_defs.h"
+#include "vmxnet3_dev.h"
-#include "vmxnet3_ring.h"
-#include "vmxnet3_logs.h"
-#include "vmxnet3_ethdev.h"
-
-#define PROCESS_SYS_EVENTS 0
-
-static int eth_vmxnet3_dev_init(struct eth_driver *eth_drv,
- struct rte_eth_dev *eth_dev);
-static int vmxnet3_dev_configure(struct rte_eth_dev *dev);
-static int vmxnet3_dev_start(struct rte_eth_dev *dev);
-static void vmxnet3_dev_stop(struct rte_eth_dev *dev);
-static void vmxnet3_dev_close(struct rte_eth_dev *dev);
-static void vmxnet3_dev_set_rxmode(struct vmxnet3_hw *hw, uint32_t feature, int set);
-static void vmxnet3_dev_promiscuous_enable(struct rte_eth_dev *dev);
-static void vmxnet3_dev_promiscuous_disable(struct rte_eth_dev *dev);
-static void vmxnet3_dev_allmulticast_enable(struct rte_eth_dev *dev);
-static void vmxnet3_dev_allmulticast_disable(struct rte_eth_dev *dev);
-static int vmxnet3_dev_link_update(struct rte_eth_dev *dev,
- int wait_to_complete);
-static void vmxnet3_dev_stats_get(struct rte_eth_dev *dev,
- struct rte_eth_stats *stats);
-static void vmxnet3_dev_info_get(struct rte_eth_dev *dev,
- struct rte_eth_dev_info *dev_info);
-#if PROCESS_SYS_EVENTS == 1
-static void vmxnet3_process_events(struct vmxnet3_hw *);
-#endif
-/*
- * The set of PCI devices this driver supports
- */
-static struct rte_pci_id pci_id_vmxnet3_map[] = {
+/* PCI identifier */
+#define PCI_VENDOR_ID_VMWARE 0x15AD
+#define PCI_DEVICE_ID_VMWARE_VMXNET3 0x07B0
+#define VMXNET3_DRIVER_VERSION "0.0.1"
+#define VMXNET3_DRIVER_VERSION_NUM 0x00000100
+
+static struct rte_pci_id pci_id_vmxnet3_map[] = {
#define RTE_PCI_DEV_ID_DECL_VMXNET3(vend, dev) {RTE_PCI_DEVICE(vend, dev)},
#include "rte_pci_dev_ids.h"
-
-{ .vendor_id = 0, /* sentinel */ },
+ { .vendor_id = 0, }
};
-static struct eth_dev_ops vmxnet3_eth_dev_ops = {
- .dev_configure = vmxnet3_dev_configure,
- .dev_start = vmxnet3_dev_start,
- .dev_stop = vmxnet3_dev_stop,
- .dev_close = vmxnet3_dev_close,
- .promiscuous_enable = vmxnet3_dev_promiscuous_enable,
- .promiscuous_disable = vmxnet3_dev_promiscuous_disable,
- .allmulticast_enable = vmxnet3_dev_allmulticast_enable,
- .allmulticast_disable = vmxnet3_dev_allmulticast_disable,
- .link_update = vmxnet3_dev_link_update,
- .stats_get = vmxnet3_dev_stats_get,
- .dev_infos_get = vmxnet3_dev_info_get,
- .rx_queue_setup = vmxnet3_dev_rx_queue_setup,
- .rx_queue_release = vmxnet3_dev_rx_queue_release,
- .tx_queue_setup = vmxnet3_dev_tx_queue_setup,
- .tx_queue_release = vmxnet3_dev_tx_queue_release,
-};
+static int vmxnet3_link_update(struct rte_eth_dev *, int);
-static const struct rte_memzone *
-gpa_zone_reserve(struct rte_eth_dev *dev, uint32_t size,
- const char *post_string, int socket_id, uint16_t align)
+/* Reserve contiguous memory zone.
+ * Lookup first before allocating
+ */
+static const struct rte_memzone *dma_zone_reserve(const char *name,
+ uint32_t size, int socket_id,
+ uint32_t align)
{
- char z_name[RTE_MEMZONE_NAMESIZE];
const struct rte_memzone *mz;
- rte_snprintf(z_name, sizeof(z_name), "%s_%d_%s",
- dev->driver->pci_drv.name, dev->data->port_id, post_string);
-
- mz = rte_memzone_lookup(z_name);
+ mz = rte_memzone_lookup(name);
if (mz)
return mz;
- return rte_memzone_reserve_aligned(z_name, size,
- socket_id, 0, align);
+ return rte_memzone_reserve_aligned(name, size, socket_id,
+ 0, align);
}
-/**
- * Atomically reads the link status information from global
- * structure rte_eth_dev.
- *
- * @param dev
- * - Pointer to the structure rte_eth_dev to read from.
- * - Pointer to the buffer to be saved with the link status.
- *
- * @return
- * - On success, zero.
- * - On failure, negative value.
- */
-static inline int
-rte_vmxnet3_dev_atomic_write_link_status(struct rte_eth_dev *dev,
- struct rte_eth_link *link)
-{
- struct rte_eth_link *dst = &(dev->data->dev_link);
- struct rte_eth_link *src = link;
-
- if (rte_atomic64_cmpset((uint64_t *)dst, *(uint64_t *)dst,
- *(uint64_t *)src) == 0)
- return -1;
+static uint32_t
+vmxnet3_cmd(struct vmxnet3_dev *dp, uint32_t cmd)
+{
+ rte_spinlock_lock(&dp->cmd_lock);
+ vmxnet3_bar1_put32(dp, VMXNET3_REG_CMD, cmd);
+ cmd = vmxnet3_bar1_get32(dp, VMXNET3_REG_CMD);
+ rte_spinlock_unlock(&dp->cmd_lock);
- return 0;
+ return cmd;
}
-/*
- * This function is based on vmxnet3_disable_intr()
- */
static void
-vmxnet3_disable_intr(struct vmxnet3_hw *hw)
+vmxnet3_vlan_filter_disable(struct vmxnet3_dev *dp)
{
- int i;
-
- PMD_INIT_FUNC_TRACE();
+ Vmxnet3_DriverShared *ds = dp->shared->addr;
+ uint32_t *vfTable = ds->devRead.rxFilterConf.vfTable;
- hw->shared->devRead.intrConf.intrCtrl |= VMXNET3_IC_DISABLE_ALL;
- for (i = 0; i < VMXNET3_MAX_INTRS; i++)
- VMXNET3_WRITE_BAR0_REG(hw, VMXNET3_REG_IMR + i * 8, 1);
+ memset(vfTable, 0, VMXNET3_VFT_SIZE * sizeof(uint32_t));
+ vmxnet3_cmd(dp, VMXNET3_CMD_UPDATE_VLAN_FILTERS);
}
-/*
- * It returns 0 on success.
- */
-static int
-eth_vmxnet3_dev_init(__attribute__((unused)) struct eth_driver *eth_drv,
- struct rte_eth_dev *eth_dev)
+static void
+vmxnet3_vlan_filter_restore(struct vmxnet3_dev *dp)
{
- struct rte_pci_device *pci_dev;
- struct vmxnet3_hw *hw =
- VMXNET3_DEV_PRIVATE_TO_HW(eth_dev->data->dev_private);
- uint32_t mac_hi, mac_lo, ver;
+ Vmxnet3_DriverShared *ds = dp->shared->addr;
+ uint32_t *vfTable = ds->devRead.rxFilterConf.vfTable;
- PMD_INIT_FUNC_TRACE();
+ memcpy(vfTable, dp->shadow_vfta, VMXNET3_VFT_SIZE * sizeof(uint32_t));
+ vmxnet3_cmd(dp, VMXNET3_CMD_UPDATE_VLAN_FILTERS);
+}
- eth_dev->dev_ops = &vmxnet3_eth_dev_ops;
- eth_dev->rx_pkt_burst = &vmxnet3_recv_pkts;
- eth_dev->tx_pkt_burst = &vmxnet3_xmit_pkts;
- pci_dev = eth_dev->pci_dev;
+static void
+vmxnet3_refresh_rxfilter(struct vmxnet3_dev *dp)
+{
+ Vmxnet3_DriverShared *ds = dp->shared->addr;
- /*
- * for secondary processes, we don't initialise any further as primary
- * has already done this work.
- */
- if (rte_eal_process_type() != RTE_PROC_PRIMARY)
- return 0;
+ ds->devRead.rxFilterConf.rxMode = rte_cpu_to_le_32(dp->rx_mode);
+ vmxnet3_cmd(dp, VMXNET3_CMD_UPDATE_RX_MODE);
+}
- /* Vendor and Device ID need to be set before init of shared code */
- hw->device_id = pci_dev->id.device_id;
- hw->vendor_id = pci_dev->id.vendor_id;
- hw->hw_addr0 = (void *)pci_dev->mem_resource[0].addr;
- hw->hw_addr1 = (void *)pci_dev->mem_resource[1].addr;
-
- hw->num_rx_queues = 1;
- hw->num_tx_queues = 1;
- hw->cur_mtu = ETHER_MTU;
- hw->bufs_per_pkt = 1;
-
- /* Check h/w version compatibility with driver. */
- ver = VMXNET3_READ_BAR1_REG(hw, VMXNET3_REG_VRRS);
- PMD_INIT_LOG(DEBUG, "Harware version : %d\n", ver);
- if (ver & 0x1)
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_VRRS, 1);
- else {
- PMD_INIT_LOG(ERR, "Uncompatiable h/w version, should be 0x1\n");
- return -EIO;
- }
-
- /* Check UPT version compatibility with driver. */
- ver = VMXNET3_READ_BAR1_REG(hw, VMXNET3_REG_UVRS);
- PMD_INIT_LOG(DEBUG, "UPT harware version : %d\n", ver);
- if (ver & 0x1)
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_UVRS, 1);
- else {
- PMD_INIT_LOG(ERR, "Incompatiable UPT version.\n");
- return -EIO;
- }
-
- /* Getting MAC Address */
- mac_lo = VMXNET3_READ_BAR1_REG(hw, VMXNET3_REG_MACL);
- mac_hi = VMXNET3_READ_BAR1_REG(hw, VMXNET3_REG_MACH);
- memcpy(hw->perm_addr , &mac_lo, 4);
- memcpy(hw->perm_addr+4, &mac_hi, 2);
-
- /* Allocate memory for storing MAC addresses */
- eth_dev->data->mac_addrs = rte_zmalloc("vmxnet3", ETHER_ADDR_LEN *
- VMXNET3_MAX_MAC_ADDRS, 0);
- if (eth_dev->data->mac_addrs == NULL) {
- PMD_INIT_LOG(ERR,
- "Failed to allocate %d bytes needed to store MAC addresses",
- ETHER_ADDR_LEN * VMXNET3_MAX_MAC_ADDRS);
- return -ENOMEM;
- }
- /* Copy the permanent MAC address */
- ether_addr_copy((struct ether_addr *) hw->perm_addr,
- ð_dev->data->mac_addrs[0]);
-
- PMD_INIT_LOG(DEBUG, "MAC Address : %02x:%02x:%02x:%02x:%02x:%02x \n",
- hw->perm_addr[0], hw->perm_addr[1], hw->perm_addr[2],
- hw->perm_addr[3], hw->perm_addr[4], hw->perm_addr[5]);
+static void vmxnet3_read_mac(struct vmxnet3_dev *dp, uint8_t *mac_addr)
+{
+ uint32_t t;
- /* Put device in Quiesce Mode */
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_CMD, VMXNET3_CMD_QUIESCE_DEV);
+ t = vmxnet3_bar1_get32(dp, VMXNET3_REG_MACL);
+ *(uint32_t *) mac_addr = t;
- return 0;
-}
+ t = vmxnet3_bar1_get32(dp, VMXNET3_REG_MACH);
+ mac_addr[4] = t & 0xff;
+ mac_addr[5] = (t >> 8) & 0xff;
-static struct eth_driver rte_vmxnet3_pmd = {
- {
- .name = "rte_vmxnet3_pmd",
- .id_table = pci_id_vmxnet3_map,
- .drv_flags = RTE_PCI_DRV_NEED_IGB_UIO,
- },
- .eth_dev_init = eth_vmxnet3_dev_init,
- .dev_private_size = sizeof(struct vmxnet3_adapter),
-};
+ PMD_INIT_LOG(DEBUG, "mac address %x:%x:%x:%x:%x:%x",
+ mac_addr[0], mac_addr[1], mac_addr[2],
+ mac_addr[3], mac_addr[4], mac_addr[5]);
+}
-/*
- * Driver initialization routine.
- * Invoked once at EAL init time.
- * Register itself as the [Poll Mode] Driver of Virtual PCI VMXNET3 devices.
- */
-int
-rte_vmxnet3_pmd_init(void)
+static int
+vmxnet3_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vid, int on)
{
- PMD_INIT_FUNC_TRACE();
+ struct vmxnet3_dev *dp = dev->data->dev_private;
+ Vmxnet3_DriverShared *ds = dp->shared->addr;
+ uint32_t *vfTable = ds->devRead.rxFilterConf.vfTable;
+
+ /* save state for restore */
+ if (on)
+ VMXNET3_SET_VFTABLE_ENTRY(dp->shadow_vfta, vid);
+ else
+ VMXNET3_CLEAR_VFTABLE_ENTRY(dp->shadow_vfta, vid);
- rte_eth_driver_register(&rte_vmxnet3_pmd);
+ if (dp->rx_mode & VMXNET3_RXM_PROMISC)
+ return 0;
+
+ /* set in hardware */
+ if (on)
+ VMXNET3_SET_VFTABLE_ENTRY(vfTable, vid);
+ else
+ VMXNET3_CLEAR_VFTABLE_ENTRY(vfTable, vid);
+
+ vmxnet3_cmd(dp, VMXNET3_CMD_UPDATE_VLAN_FILTERS);
return 0;
}
-static int
-vmxnet3_dev_configure(struct rte_eth_dev *dev)
+static void
+vmxnet3_vlan_offload_set(struct rte_eth_dev *dev, int mask)
{
- const struct rte_memzone *mz;
- struct vmxnet3_hw *hw =
- VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
- size_t size;
+ struct vmxnet3_dev *dp = dev->data->dev_private;
+ Vmxnet3_DriverShared *ds = dp->shared->addr;
+ Vmxnet3_DSDevRead *devRead = &ds->devRead;
- PMD_INIT_FUNC_TRACE();
+ if (mask & ETH_VLAN_STRIP_MASK)
+ devRead->misc.uptFeatures |= UPT1_F_RXVLAN;
+ else
+ devRead->misc.uptFeatures &= ~UPT1_F_RXVLAN;
- if (dev->data->nb_rx_queues > UINT8_MAX ||
- dev->data->nb_tx_queues > UINT8_MAX)
- return (-EINVAL);
-
- size = dev->data->nb_rx_queues * sizeof(struct Vmxnet3_TxQueueDesc) +
- dev->data->nb_tx_queues * sizeof (struct Vmxnet3_RxQueueDesc);
-
- if (size > UINT16_MAX)
- return (-EINVAL);
-
- hw->num_rx_queues = (uint8_t)dev->data->nb_rx_queues;
- hw->num_tx_queues = (uint8_t)dev->data->nb_tx_queues;
-
- /*
- * Allocate a memzone for Vmxnet3_DriverShared - Vmxnet3_DSDevRead
- * on current socket
- */
- mz = gpa_zone_reserve(dev, sizeof (struct Vmxnet3_DriverShared),
- "shared", rte_socket_id(), 8);
-
- if (mz == NULL) {
- PMD_INIT_LOG(ERR, "ERROR: Creating shared zone\n");
- return (-ENOMEM);
- }
- memset(mz->addr, 0, mz->len);
+ if (mask & ETH_VLAN_FILTER_MASK)
+ vmxnet3_vlan_filter_restore(dp);
+ else
+ vmxnet3_vlan_filter_disable(dp);
+}
- hw->shared = mz->addr;
- hw->sharedPA = mz->phys_addr;
+static void
+vmxnet3_write_mac(struct vmxnet3_dev *dp, const uint8_t *macaddr)
+{
+ uint32_t val32;
- /*
- * Allocate a memzone for Vmxnet3_RxQueueDesc - Vmxnet3_TxQueueDesc
- * on current socket
- */
- mz = gpa_zone_reserve(dev, size, "queuedesc",
- rte_socket_id(), VMXNET3_QUEUE_DESC_ALIGN);
- if (mz == NULL) {
- PMD_INIT_LOG(ERR, "ERROR: Creating queue descriptors zone\n");
- return (-ENOMEM);
- }
- memset(mz->addr, 0, mz->len);
+ memcpy(&val32, macaddr, sizeof(uint32_t));
+ vmxnet3_bar1_put32(dp, VMXNET3_REG_MACL, val32);
+ val32 = macaddr[5] << 8 | macaddr[4];
+ vmxnet3_bar1_put32(dp, VMXNET3_REG_MACH, val32);
+}
- hw->tqd_start = (Vmxnet3_TxQueueDesc *)mz->addr;
- hw->rqd_start = (Vmxnet3_RxQueueDesc *)(hw->tqd_start + hw->num_tx_queues);
+static const uint8_t rss_default_key[UPT1_RSS_MAX_KEY_SIZE] = {
+ 0xd9, 0x20, 0x1d, 0xb5, 0x51, 0x90, 0xd1, 0x32, 0x2e, 0xde,
+ 0xec, 0x8e, 0xe7, 0xc2, 0x8e, 0xe8, 0xe9, 0x42, 0x9b, 0xe4,
+ 0x67, 0x12, 0x22, 0x2d, 0xf6, 0x18, 0x40, 0x64, 0xcc, 0xcd,
+ 0x6c, 0x95, 0xb9, 0x22, 0x4a, 0x30, 0x3c, 0x8d, 0x12, 0x3e,
+};
- hw->queueDescPA = mz->phys_addr;
- hw->queue_desc_len = (uint16_t)size;
+static int vmxnet3_rss_setup(struct rte_eth_dev *eth_dev)
+{
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
+ const struct rte_eth_conf *eth_conf = ð_dev->data->dev_conf;
+ Vmxnet3_DriverShared *ds = dp->shared->addr;
+ Vmxnet3_DSDevRead *devRead = &ds->devRead;
+ const struct rte_memzone *rz;
+ struct UPT1_RSSConf *rssConf;
+ const uint8_t *hash_key;
+ unsigned i;
+ char z_name[RTE_MEMZONE_NAMESIZE];
- if(dev->data->dev_conf.rxmode.mq_mode == ETH_MQ_RX_RSS) {
-
- /* Allocate memory structure for UPT1_RSSConf and configure */
- mz = gpa_zone_reserve(dev, sizeof (struct VMXNET3_RSSConf), "rss_conf",
- rte_socket_id(), CACHE_LINE_SIZE);
- if (mz == NULL) {
- PMD_INIT_LOG(ERR, "ERROR: Creating rss_conf structure zone\n");
- return (-ENOMEM);
- }
- memset(mz->addr, 0, mz->len);
+ /* If no RSS hash types then RSS unnecessary. */
+ if (eth_conf->rx_adv_conf.rss_conf.rss_hf == 0)
+ return 0;
- hw->rss_conf = mz->addr;
- hw->rss_confPA = mz->phys_addr;
- }
+ rte_snprintf(z_name, sizeof(z_name), "vmxnet3_rss_%u",
+ eth_dev->data->port_id);
+
+ rz = dma_zone_reserve(z_name, sizeof(*rssConf), SOCKET_ID_ANY,
+ CACHE_LINE_SIZE);
+ if (rz == NULL)
+ return -ENOMEM;
+
+ /* NB: RSS definitions from VMXNET3 are a subset of DPDK
+ probably because both inherited from Msft NDIS spec? */
+ rssConf = rz->addr;
+ rssConf->hashType = eth_conf->rx_adv_conf.rss_conf.rss_hf & 0xf;
+ rssConf->hashFunc = UPT1_RSS_HASH_FUNC_TOEPLITZ;
+ rssConf->hashKeySize = UPT1_RSS_MAX_KEY_SIZE; /* 40 */
+ hash_key = eth_conf->rx_adv_conf.rss_conf.rss_key;
+ if (hash_key == NULL)
+ hash_key = rss_default_key;
+
+ memcpy(rssConf->hashKey, hash_key, UPT1_RSS_MAX_KEY_SIZE);
+
+ rssConf->indTableSize = UPT1_RSS_MAX_IND_TABLE_SIZE;
+ for (i = 0; i < rssConf->indTableSize; i++)
+ rssConf->indTable[i] = i % eth_dev->data->nb_rx_queues;
+
+ devRead->misc.uptFeatures |= UPT1_F_RSS;
+ devRead->rssConfDesc.confVer = 1;
+ devRead->rssConfDesc.confLen = sizeof(*rssConf);
+ devRead->rssConfDesc.confPA = rz->phys_addr;
return 0;
}
-static int
-vmxnet3_setup_driver_shared(struct rte_eth_dev *dev)
+static void vmxnet3_shared_setup(struct rte_eth_dev *eth_dev)
{
- struct rte_eth_conf port_conf = dev->data->dev_conf;
- struct vmxnet3_hw *hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
- Vmxnet3_DriverShared *shared = hw->shared;
- Vmxnet3_DSDevRead *devRead = &shared->devRead;
- uint32_t *mac_ptr;
- uint32_t val, i;
- int ret;
-
- shared->magic = VMXNET3_REV1_MAGIC;
- devRead->misc.driverInfo.version = VMXNET3_DRIVER_VERSION_NUM;
-
- /* Setting up Guest OS information */
- devRead->misc.driverInfo.gos.gosBits = sizeof(void *) == 4 ?
- VMXNET3_GOS_BITS_32 :
- VMXNET3_GOS_BITS_64;
- devRead->misc.driverInfo.gos.gosType = VMXNET3_GOS_TYPE_LINUX;
- devRead->misc.driverInfo.vmxnet3RevSpt = 1;
- devRead->misc.driverInfo.uptVerSpt = 1;
-
- devRead->misc.queueDescPA = hw->queueDescPA;
- devRead->misc.queueDescLen = hw->queue_desc_len;
- devRead->misc.mtu = hw->cur_mtu;
- devRead->misc.numTxQueues = hw->num_tx_queues;
- devRead->misc.numRxQueues = hw->num_rx_queues;
-
- /*
- * Set number of interrupts to 1
- * PMD disables all the interrupts but this is MUST to activate device
- * It needs at least one interrupt for link events to handle
- * So we'll disable it later after device activation if needed
- */
- devRead->intrConf.numIntrs = 1;
- devRead->intrConf.intrCtrl |= VMXNET3_IC_DISABLE_ALL;
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
+ Vmxnet3_DriverShared *ds = dp->shared->addr;
+ Vmxnet3_DSDevRead *devRead = &ds->devRead;
+ unsigned num_txq = eth_dev->data->nb_tx_queues;
+ unsigned num_rxq = eth_dev->data->nb_rx_queues;
+ uint32_t mtu = eth_dev->data->max_frame_size;
+ unsigned i;
+ size_t size;
- for (i = 0; i < hw->num_tx_queues; i++) {
- Vmxnet3_TxQueueDesc *tqd = &hw->tqd_start[i];
- vmxnet3_tx_queue_t *txq = dev->data->tx_queues[i];
-
- tqd->ctrl.txNumDeferred = 0;
- tqd->ctrl.txThreshold = 1;
- tqd->conf.txRingBasePA = txq->cmd_ring.basePA;
- tqd->conf.compRingBasePA = txq->comp_ring.basePA;
-
- tqd->conf.txRingSize = txq->cmd_ring.size;
- tqd->conf.compRingSize = txq->comp_ring.size;
- tqd->conf.intrIdx = txq->comp_ring.intr_idx;
- tqd->status.stopped = TRUE;
- tqd->status.error = 0;
- memset(&tqd->stats, 0, sizeof(tqd->stats));
- }
+ /* Setup shared area */
+ memset(ds, 0, sizeof(*ds));
- for (i = 0; i < hw->num_rx_queues; i++) {
- Vmxnet3_RxQueueDesc *rqd = &hw->rqd_start[i];
- vmxnet3_rx_queue_t *rxq = dev->data->rx_queues[i];
-
- rqd->conf.rxRingBasePA[0] = rxq->cmd_ring[0].basePA;
- rqd->conf.rxRingBasePA[1] = rxq->cmd_ring[1].basePA;
- rqd->conf.compRingBasePA = rxq->comp_ring.basePA;
-
- rqd->conf.rxRingSize[0] = rxq->cmd_ring[0].size;
- rqd->conf.rxRingSize[1] = rxq->cmd_ring[1].size;
- rqd->conf.compRingSize = rxq->comp_ring.size;
- rqd->conf.intrIdx = rxq->comp_ring.intr_idx;
- rqd->status.stopped = TRUE;
- rqd->status.error = 0;
- memset(&rqd->stats, 0, sizeof(rqd->stats));
- }
+ ds->magic = rte_cpu_to_le_32(VMXNET3_REV1_MAGIC);
+ devRead->misc.driverInfo.version
+ = rte_cpu_to_le_32(VMXNET3_DRIVER_VERSION_NUM);
- /* RxMode set to 0 of VMXNET3_RXM_xxx */
- devRead->rxFilterConf.rxMode = 0;
+ if (sizeof(void *) == 4)
+ devRead->misc.driverInfo.gos.gosBits = VMXNET3_GOS_BITS_32;
+ else
+ devRead->misc.driverInfo.gos.gosBits = VMXNET3_GOS_BITS_64;
- /* Setting up feature flags */
- if(dev->data->dev_conf.rxmode.hw_ip_checksum) {
- devRead->misc.uptFeatures |= VMXNET3_F_RXCSUM;
- }
+ devRead->misc.driverInfo.gos.gosType = VMXNET3_GOS_TYPE_LINUX;
+ devRead->misc.driverInfo.vmxnet3RevSpt = 1;
+ devRead->misc.driverInfo.uptVerSpt = 1;
- if(dev->data->dev_conf.rxmode.hw_vlan_strip) {
- devRead->misc.uptFeatures |= VMXNET3_F_RXVLAN;
+ devRead->misc.uptFeatures = 0;
+ if (eth_dev->data->dev_conf.rxmode.hw_ip_checksum)
+ devRead->misc.uptFeatures |= UPT1_F_RXCSUM;
+ vmxnet3_vlan_offload_set(eth_dev,
+ ETH_VLAN_STRIP_MASK | ETH_VLAN_FILTER_MASK);
+ devRead->misc.mtu = rte_cpu_to_le_32(mtu);
+
+ /* Setup Tx/Rx queue descriptor area */
+ size = num_txq * sizeof(struct Vmxnet3_TxQueueDesc);
+ size += num_rxq * sizeof(struct Vmxnet3_RxQueueDesc);
+
+ devRead->misc.queueDescPA = rte_cpu_to_le_64(dp->qdesc->phys_addr);
+ devRead->misc.queueDescLen = rte_cpu_to_le_32(size);
+
+ /* tx queue setup */
+ devRead->misc.numTxQueues = num_txq;
+ dp->tqd_start = dp->qdesc->addr;
+ for (i = 0; i < num_txq; i++) {
+ struct vmxnet3_tx_queue *txq = dp->tx_queue + i;
+ Vmxnet3_TxQueueDesc *txd = dp->tqd_start + i;
+
+ txq->shared = &txd->ctrl;
+ txd->conf.txRingBasePA
+ = rte_cpu_to_le_64(txq->tx_ring.phys_addr);
+ txd->conf.txRingSize
+ = rte_cpu_to_le_32(txq->tx_ring.size);
+ txd->conf.compRingBasePA
+ = rte_cpu_to_le_64(txq->comp_ring.phys_addr);
+ txd->conf.compRingSize
+ = rte_cpu_to_le_32(txq->comp_ring.size);
+ txd->conf.dataRingBasePA = 0;
+ txd->conf.dataRingSize = 0;
+ }
+
+ devRead->misc.numRxQueues = num_rxq;
+ dp->rqd_start = (struct Vmxnet3_RxQueueDesc *)(dp->tqd_start + num_txq);
+ for (i = 0; i < num_rxq; i++) {
+ struct vmxnet3_rx_queue *rxq = dp->rx_queue + i;
+ Vmxnet3_RxQueueDesc *rxd = dp->rqd_start + i;
+
+ rxq->shared = &rxd->ctrl;
+ rxd->conf.rxRingBasePA[0]
+ = rte_cpu_to_le_64(rxq->rx_ring.phys_addr);
+ rxd->conf.rxRingSize[0]
+ = rte_cpu_to_le_32(rxq->rx_ring.size);
+
+ rxd->conf.rxRingBasePA[1] = 0;
+ rxd->conf.rxRingSize[1] = 0;
+
+ rxd->conf.compRingBasePA
+ = rte_cpu_to_le_64(rxq->comp_ring.phys_addr);
+ rxd->conf.compRingSize
+ = rte_cpu_to_le_32(rxq->comp_ring.size);
}
- if(port_conf.rxmode.mq_mode == ETH_MQ_RX_RSS) {
- ret = vmxnet3_rss_configure(dev);
- if(ret != VMXNET3_SUCCESS) {
- return ret;
- }
- devRead->misc.uptFeatures |= VMXNET3_F_RSS;
- devRead->rssConfDesc.confVer = 1;
- devRead->rssConfDesc.confLen = sizeof(struct VMXNET3_RSSConf);
- devRead->rssConfDesc.confPA = hw->rss_confPA;
- }
+ /* interrupt settings */
+ devRead->intrConf.autoMask = 1;
+ devRead->intrConf.numIntrs = 1;
+ devRead->intrConf.eventIntrIdx = 0;
+ devRead->intrConf.modLevels[0] = UPT1_IML_ADAPTIVE;
+ devRead->intrConf.intrCtrl |= rte_cpu_to_le_32(VMXNET3_IC_DISABLE_ALL);
+}
- if(dev->data->dev_conf.rxmode.hw_vlan_filter) {
- ret = vmxnet3_vlan_configure(dev);
- if(ret != VMXNET3_SUCCESS) {
- return ret;
- }
- }
+/* Called after queues are setup */
+static int
+vmxnet3_configure(struct rte_eth_dev *eth_dev)
+{
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
+ Vmxnet3_DriverShared *ds = dp->shared->addr;
+ uint32_t *vfTable = ds->devRead.rxFilterConf.vfTable;
- PMD_INIT_LOG(DEBUG, "Writing MAC Address : %02x:%02x:%02x:%02x:%02x:%02x \n",
- hw->perm_addr[0], hw->perm_addr[1], hw->perm_addr[2],
- hw->perm_addr[3], hw->perm_addr[4], hw->perm_addr[5]);
+ PMD_INIT_LOG(DEBUG, "configure");
- /* Write MAC Address back to device */
- mac_ptr = (uint32_t *)hw->perm_addr;
- val = *mac_ptr;
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_MACL, val);
+ /* Default vlan filter: allow untagged traffic */
+ VMXNET3_SET_VFTABLE_ENTRY(vfTable, 0);
+ dp->shadow_vfta[0] = 1;
- val = (hw->perm_addr[5] << 8) | hw->perm_addr[4];
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_MACH, val);
+ /* Initialize Link state */
+ vmxnet3_link_update(eth_dev, 0);
- return VMXNET3_SUCCESS;
+ return 0;
}
-/*
- * Configure device link speed and setup link.
- * Must be called after eth_vmxnet3_dev_init. Other wise it might fail
- * It returns 0 on success.
- */
static int
-vmxnet3_dev_start(struct rte_eth_dev *dev)
+vmxnet3_start(struct rte_eth_dev *eth_dev)
{
- int status, ret;
- struct vmxnet3_hw *hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
+ uint32_t ret;
+ int err;
PMD_INIT_FUNC_TRACE();
- ret = vmxnet3_setup_driver_shared(dev);
- if(ret != VMXNET3_SUCCESS) {
- return ret;
- }
+ vmxnet3_shared_setup(eth_dev);
- /* Exchange shared data with device */
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_DSAL, VMXNET3_GET_ADDR_LO(
- hw->sharedPA));
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_DSAH, VMXNET3_GET_ADDR_HI(
- hw->sharedPA));
-
- /* Activate device by register write */
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_CMD, VMXNET3_CMD_ACTIVATE_DEV);
- status = VMXNET3_READ_BAR1_REG(hw, VMXNET3_REG_CMD);
-
- if (status != 0) {
- PMD_INIT_LOG(ERR, "Device activation in %s(): UNSUCCESSFUL\n", __func__);
- return -1;
+ if (vmxnet3_rss_setup(eth_dev) < 0) {
+ PMD_INIT_LOG(ERR, "Failed to setup RSS");
+ return -ENOMEM;
}
- /* Disable interrupts */
- vmxnet3_disable_intr(hw);
+ err = vmxnet3_rx_init(eth_dev);
+ if (err)
+ return err;
+
+ /* Activate device */
+ ret = vmxnet3_cmd(dp, VMXNET3_CMD_ACTIVATE_DEV);
+ if (ret != 0) {
+ PMD_INIT_LOG(ERR, "Failed to activate port %u error %u",
+ eth_dev->data->port_id, ret);
+
+ return -EINVAL;
+ }
+
+ /* Turn on Rx queues */
+ vmxnet3_rx_start(eth_dev);
+
+ /* Reset receive filter */
+ vmxnet3_vlan_filter_restore(dp);
+ dp->rx_mode |= VMXNET3_RXM_UCAST | VMXNET3_RXM_BCAST;
+ vmxnet3_refresh_rxfilter(dp);
- /*
- * Load RX queues with blank mbufs and update next2fill index for device
- * Update RxMode of the device
- */
- ret = vmxnet3_dev_rxtx_init(dev);
- if(ret != VMXNET3_SUCCESS) {
- PMD_INIT_LOG(ERR, "Device receive init in %s: UNSUCCESSFUL\n", __func__);
- return ret;
- }
+ return 0;
+}
- /* Setting proper Rx Mode and issue Rx Mode Update command */
- vmxnet3_dev_set_rxmode(hw, VMXNET3_RXM_UCAST | VMXNET3_RXM_ALL_MULTI, 1);
+static void
+vmxnet3_stop(struct rte_eth_dev *eth_dev)
+{
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
- /*
- * Don't need to handle events for now
- */
-#if PROCESS_SYS_EVENTS == 1
- events = VMXNET3_READ_BAR1_REG(hw, VMXNET3_REG_ECR);
- PMD_INIT_LOG(DEBUG, "Reading events: 0x%X\n\n", events);
- vmxnet3_process_events(hw);
-#endif
- return status;
+ PMD_INIT_FUNC_TRACE();
+
+ vmxnet3_cmd(dp, VMXNET3_CMD_QUIESCE_DEV);
+
+ vmxnet3_rx_flush_all(eth_dev);
+ vmxnet3_tx_flush_all(eth_dev);
}
-/*
- * Stop device: disable rx and tx functions to allow for reconfiguring.
- */
static void
-vmxnet3_dev_stop(struct rte_eth_dev *dev)
+vmxnet3_close(struct rte_eth_dev *eth_dev)
{
- struct rte_eth_link link;
- struct vmxnet3_hw *hw =
- VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
PMD_INIT_FUNC_TRACE();
- if(hw->adapter_stopped == TRUE) {
- PMD_INIT_LOG(DEBUG, "Device already closed.\n");
- return;
- }
+ vmxnet3_cmd(dp, VMXNET3_CMD_RESET_DEV);
+}
- /* disable interrupts */
- vmxnet3_disable_intr(hw);
+static void
+vmxnet3_promiscuous_enable(struct rte_eth_dev *eth_dev)
+{
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
- /* quiesce the device first */
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_CMD, VMXNET3_CMD_QUIESCE_DEV);
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_DSAL, 0);
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_DSAH, 0);
-
- /* reset the device */
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_CMD, VMXNET3_CMD_RESET_DEV);
- PMD_INIT_LOG(DEBUG, "Device reset.\n");
- hw->adapter_stopped = FALSE;
+ PMD_INIT_FUNC_TRACE();
- vmxnet3_dev_clear_queues(dev);
+ vmxnet3_vlan_filter_disable(dp);
- /* Clear recorded link status */
- memset(&link, 0, sizeof(link));
- rte_vmxnet3_dev_atomic_write_link_status(dev, &link);
+ dp->rx_mode |= VMXNET3_RXM_PROMISC;
+ vmxnet3_refresh_rxfilter(dp);
}
-/*
- * Reset and stop device.
- */
static void
-vmxnet3_dev_close(struct rte_eth_dev *dev)
+vmxnet3_promiscuous_disable(struct rte_eth_dev *eth_dev)
{
- struct vmxnet3_hw *hw =
- VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
PMD_INIT_FUNC_TRACE();
- vmxnet3_dev_stop(dev);
- hw->adapter_stopped = TRUE;
+ vmxnet3_vlan_filter_restore(dp);
+ dp->rx_mode &= ~VMXNET3_RXM_PROMISC;
+ vmxnet3_refresh_rxfilter(dp);
}
static void
-vmxnet3_dev_stats_get( struct rte_eth_dev *dev, struct rte_eth_stats *stats)
+vmxnet3_allmulticast_enable(struct rte_eth_dev *eth_dev)
{
- unsigned int i;
- struct vmxnet3_hw *hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
-
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_CMD, VMXNET3_CMD_GET_STATS);
-
- stats->opackets = 0;
- stats->obytes = 0;
- stats->oerrors = 0;
- stats->ipackets = 0;
- stats->ibytes = 0;
- stats->rx_nombuf = 0;
- stats->ierrors = 0;
- stats->imcasts = 0;
- stats->fdirmatch = 0;
- stats->fdirmiss = 0;
-
- for (i = 0; i < hw->num_tx_queues; i++) {
- stats->opackets += hw->tqd_start[i].stats.ucastPktsTxOK +
- hw->tqd_start[i].stats.mcastPktsTxOK +
- hw->tqd_start[i].stats.bcastPktsTxOK;
- stats->obytes += hw->tqd_start[i].stats.ucastBytesTxOK +
- hw->tqd_start[i].stats.mcastBytesTxOK +
- hw->tqd_start[i].stats.bcastBytesTxOK;
- stats->oerrors += hw->tqd_start[i].stats.pktsTxError +
- hw->tqd_start[i].stats.pktsTxDiscard;
- }
-
- for (i = 0; i < hw->num_rx_queues; i++) {
- stats->ipackets += hw->rqd_start[i].stats.ucastPktsRxOK +
- hw->rqd_start[i].stats.mcastPktsRxOK +
- hw->rqd_start[i].stats.bcastPktsRxOK;
- stats->ibytes += hw->rqd_start[i].stats.ucastBytesRxOK +
- hw->rqd_start[i].stats.mcastBytesRxOK +
- hw->rqd_start[i].stats.bcastBytesRxOK;
- stats->rx_nombuf += hw->rqd_start[i].stats.pktsRxOutOfBuf;
- stats->ierrors += hw->rqd_start[i].stats.pktsRxError;
- }
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
+ dp->rx_mode |= VMXNET3_RXM_ALL_MULTI;
+ vmxnet3_refresh_rxfilter(dp);
}
static void
-vmxnet3_dev_info_get(__attribute__((unused))struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info)
+vmxnet3_allmulticast_disable(struct rte_eth_dev *eth_dev)
{
- dev_info->max_rx_queues = VMXNET3_MAX_RX_QUEUES;
- dev_info->max_tx_queues = VMXNET3_MAX_TX_QUEUES;
- dev_info->min_rx_bufsize = 1518 + RTE_PKTMBUF_HEADROOM;
- dev_info->max_rx_pktlen = 16384; /* includes CRC, cf MAXFRS register */
- dev_info->max_mac_addrs = VMXNET3_MAX_MAC_ADDRS;
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
+
+ dp->rx_mode &= ~VMXNET3_RXM_ALL_MULTI;
+ vmxnet3_refresh_rxfilter(dp);
}
-/* return 0 means link status changed, -1 means not changed */
+/* update link status value. */
static int
-vmxnet3_dev_link_update(struct rte_eth_dev *dev, __attribute__((unused)) int wait_to_complete)
+vmxnet3_link_update(struct rte_eth_dev *dev,
+ int wait_to_complete __rte_unused)
{
- struct vmxnet3_hw *hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
- struct rte_eth_link link;
- uint32_t ret;
+ struct vmxnet3_dev *dp = dev->data->dev_private;
+ struct rte_eth_link old, link;
+ uint32_t status;
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_CMD, VMXNET3_CMD_GET_LINK);
- ret = VMXNET3_READ_BAR1_REG(hw, VMXNET3_REG_CMD);
+ memset(&link, 0, sizeof(link));
- if (!ret) {
- PMD_INIT_LOG(ERR, "Link Status Negative : %s()\n", __func__);
- return -1;
- }
+ rte_eth_dev_get_link(dev, &old);
- if (ret & 0x1) {
- link.link_status = 1;
- link.link_duplex = ETH_LINK_FULL_DUPLEX;
- link.link_speed = ETH_LINK_SPEED_10000;
+ status = vmxnet3_cmd(dp, VMXNET3_CMD_GET_LINK);
- rte_vmxnet3_dev_atomic_write_link_status(dev, &link);
+ PMD_DRV_LOG(INFO, "link status %#x", status);
- return 0;
+ if (status & 1) {
+ link.link_duplex = ETH_LINK_FULL_DUPLEX;
+ link.link_speed = status >> 16;
+ link.link_status = 1;
}
- return -1;
+ rte_eth_dev_set_link(dev, &link);
+
+ return (old.link_status == link.link_status) ? -1 : 0;
}
-/* Updating rxmode through Vmxnet3_DriverShared structure in adapter */
static void
-vmxnet3_dev_set_rxmode(struct vmxnet3_hw *hw, uint32_t feature, int set) {
+vmxnet3_stats_get(struct rte_eth_dev *eth_dev, struct rte_eth_stats *stats)
+{
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
+ unsigned i;
- struct Vmxnet3_RxFilterConf *rxConf = &hw->shared->devRead.rxFilterConf;
- if(set)
- rxConf->rxMode = rxConf->rxMode | feature;
- else
- rxConf->rxMode = rxConf->rxMode & (~feature);
+ if (!dp->tqd_start || !dp->rqd_start)
+ return;
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_CMD, VMXNET3_CMD_UPDATE_RX_MODE);
-}
+ /* Collect stats */
+ vmxnet3_cmd(dp, VMXNET3_CMD_GET_STATS);
-/* Promiscuous supported only if Vmxnet3_DriverShared is initialized in adapter */
-static void
-vmxnet3_dev_promiscuous_enable(struct rte_eth_dev *dev)
-{
- struct vmxnet3_hw *hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
- vmxnet3_dev_set_rxmode(hw, VMXNET3_RXM_PROMISC, 1);
-}
+ /* NB: RTE_ETHDEV_QUEUE_STAT_CNTRS (16) > VMXNET3_MAX_TX_QUEUES (8)*/
+ for (i = 0; i < eth_dev->data->nb_tx_queues; i++) {
+ struct UPT1_TxStats *txStats = &dp->tqd_start[i].stats;
-/* Promiscuous supported only if Vmxnet3_DriverShared is initialized in adapter */
-static void
-vmxnet3_dev_promiscuous_disable(struct rte_eth_dev *dev)
-{
- struct vmxnet3_hw *hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
- vmxnet3_dev_set_rxmode(hw, VMXNET3_RXM_PROMISC, 0);
-}
+ stats->q_opackets[i] = txStats->ucastPktsTxOK +
+ txStats->mcastPktsTxOK +
+ txStats->bcastPktsTxOK;
+ stats->q_obytes[i] = txStats->ucastBytesTxOK +
+ txStats->mcastBytesTxOK +
+ txStats->bcastBytesTxOK;
-/* Allmulticast supported only if Vmxnet3_DriverShared is initialized in adapter */
-static void
-vmxnet3_dev_allmulticast_enable(struct rte_eth_dev *dev)
-{
- struct vmxnet3_hw *hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
- vmxnet3_dev_set_rxmode(hw, VMXNET3_RXM_PROMISC, 1);
+ stats->opackets += stats->q_opackets[i];
+ stats->obytes += stats->q_obytes[i];
+ stats->oerrors += txStats->pktsTxError;
+ }
+
+ for (i = 0; i < eth_dev->data->nb_rx_queues; i++) {
+ struct UPT1_RxStats *rxStats = &dp->rqd_start[i].stats;
+
+ stats->q_ipackets[i] = rxStats->ucastPktsRxOK +
+ rxStats->mcastPktsRxOK +
+ rxStats->bcastPktsRxOK;
+
+ stats->q_ibytes[i] = rxStats->ucastBytesRxOK +
+ rxStats->mcastBytesRxOK +
+ rxStats->bcastBytesRxOK;
+
+ stats->ipackets += stats->q_ipackets[i];
+ stats->ibytes += stats->q_ibytes[i];
+
+ stats->q_errors[i] = rxStats->pktsRxError;
+ stats->ierrors += rxStats->pktsRxError;
+ stats->imcasts += rxStats->mcastPktsRxOK;
+
+ struct vmxnet3_rx_queue *rxq = &dp->rx_queue[i];
+ stats->ierrors += rxq->fcs_errors + rxq->drop_errors;
+ }
}
-/* Allmulticast supported only if Vmxnet3_DriverShared is initialized in adapter */
static void
-vmxnet3_dev_allmulticast_disable(struct rte_eth_dev *dev)
+vmxnet3_infos_get(struct rte_eth_dev *eth_dev __rte_unused,
+ struct rte_eth_dev_info *dev_info)
{
- struct vmxnet3_hw *hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
- vmxnet3_dev_set_rxmode(hw, VMXNET3_RXM_PROMISC, 0);
+ dev_info->max_rx_queues = VMXNET3_MAX_RX_QUEUES;
+ dev_info->max_tx_queues = VMXNET3_MAX_TX_QUEUES;
+ dev_info->min_rx_bufsize = VMXNET3_MIN_T0_BUF_SIZE;
+ dev_info->max_rx_pktlen = ETHER_MAX_LEN;
+ // in future can be VMXNET3_MAX_MTU + VMXNET3_MAX_ETH_HDR
+ dev_info->max_mac_addrs = 1;
}
-#if PROCESS_SYS_EVENTS == 1
static void
-vmxnet3_process_events(struct vmxnet3_hw *hw)
+vmxnet3_mac_addr_set(struct rte_eth_dev *eth_dev, struct ether_addr *mac_addr)
+{
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
+
+ vmxnet3_write_mac(dp, mac_addr->addr_bytes);
+}
+
+static struct eth_dev_ops vmxnet3_dev_ops = {
+ .dev_configure = vmxnet3_configure,
+ .dev_start = vmxnet3_start,
+ .dev_stop = vmxnet3_stop,
+ .dev_close = vmxnet3_close,
+ .promiscuous_enable = vmxnet3_promiscuous_enable,
+ .promiscuous_disable = vmxnet3_promiscuous_disable,
+ .allmulticast_enable = vmxnet3_allmulticast_enable,
+ .allmulticast_disable = vmxnet3_allmulticast_disable,
+ .link_update = vmxnet3_link_update,
+ .stats_get = vmxnet3_stats_get,
+ .dev_infos_get = vmxnet3_infos_get,
+ .vlan_filter_set = vmxnet3_vlan_filter_set,
+ .vlan_offload_set = vmxnet3_vlan_offload_set,
+ .rx_queue_setup = vmxnet3_rx_queue_setup,
+ .rx_queue_release = vmxnet3_rx_queue_release,
+ .tx_queue_setup = vmxnet3_tx_queue_setup,
+ .tx_queue_release = vmxnet3_tx_queue_release,
+ .mac_addr_set = vmxnet3_mac_addr_set,
+};
+
+static int
+vmxnet3_dev_init(struct eth_driver *eth_drv __rte_unused,
+ struct rte_eth_dev *eth_dev)
{
- uint32_t events = hw->shared->ecr;
- if (!events){
- PMD_INIT_LOG(ERR, "No events to process in %s()\n", __func__);
- return;
- }
+ struct vmxnet3_dev *dp = eth_dev->data->dev_private;
+ struct rte_pci_device *pci_dev = eth_dev->pci_dev;
+ uint32_t ver;
+ char z_name[RTE_MEMZONE_NAMESIZE];
+ size_t size;
- /*
- * ECR bits when written with 1b are cleared. Hence write
- * events back to ECR so that the bits which were set will be reset.
- */
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_ECR, events);
-
- /* Check if link state has changed */
- if (events & VMXNET3_ECR_LINK){
- PMD_INIT_LOG(ERR, "Process events in %s(): VMXNET3_ECR_LINK event\n", __func__);
- }
-
- /* Check if there is an error on xmit/recv queues */
- if (events & (VMXNET3_ECR_TQERR | VMXNET3_ECR_RQERR)) {
- VMXNET3_WRITE_BAR1_REG(hw, VMXNET3_REG_CMD, VMXNET3_CMD_GET_QUEUE_STATUS);
-
- if (hw->tqd_start->status.stopped)
- PMD_INIT_LOG(ERR, "tq error 0x%x\n",
- hw->tqd_start->status.error);
-
- if (hw->rqd_start->status.stopped)
- PMD_INIT_LOG(ERR, "rq error 0x%x\n",
- hw->rqd_start->status.error);
+ PMD_INIT_FUNC_TRACE();
- /* Reset the device */
- /* Have to reset the device */
- }
+ eth_dev->dev_ops = &vmxnet3_dev_ops;
+ eth_dev->rx_pkt_burst = &vmxnet3_recv_pkts;
+ eth_dev->tx_pkt_burst = &vmxnet3_xmit_pkts;
- if (events & VMXNET3_ECR_DIC)
- PMD_INIT_LOG(ERR, "Device implementation change event.\n");
+ dp->bar0 = pci_dev->mem_resource[0].addr;
+ if (!dp->bar0)
+ rte_panic("missing bar0 resource 0\n");
+
+ dp->bar1 = pci_dev->mem_resource[1].addr;
+ if (!dp->bar1)
+ rte_panic("missing bar1 resource 0\n");
+
+ rte_spinlock_init(&dp->cmd_lock);
+
+ /* Check the version numbers of the virtual device */
+ ver = vmxnet3_bar1_get32(dp, VMXNET3_REG_VRRS);
+ if (ver & 1)
+ vmxnet3_bar1_put32(dp, VMXNET3_REG_VRRS, 1);
+ else {
+ PMD_INIT_LOG(ERR, "incompatible h/w version %#x",
+ ver);
+ return -EBUSY;
+ }
+
+ ver = vmxnet3_bar1_get32(dp, VMXNET3_REG_UVRS);
+ if (ver & 1)
+ vmxnet3_bar1_put32(dp, VMXNET3_REG_UVRS, 1);
+ else {
+ PMD_INIT_LOG(ERR, "incompatible upt version %#x", ver);
+ return -EBUSY;
+ }
+
+ /* Only single mac address */
+ vmxnet3_read_mac(dp, dp->mac_addr.addr_bytes);
+ if (!is_valid_assigned_ether_addr(&dp->mac_addr)) {
+ PMD_INIT_LOG(ERR, "invalid mac address");
+ return -EINVAL;
+ }
+
+ eth_dev->data->mac_addrs = &dp->mac_addr;
+
+ /* Get memory zone for shared */
+ rte_snprintf(z_name, sizeof(z_name), "vmxnet3_shared_%u",
+ eth_dev->data->port_id);
+ dp->shared = dma_zone_reserve(z_name,
+ sizeof(struct Vmxnet3_DriverShared),
+ SOCKET_ID_ANY, CACHE_LINE_SIZE);
+ if (!dp->shared)
+ return -ENOMEM;
- if (events & VMXNET3_ECR_DEBUG)
- PMD_INIT_LOG(ERR, "Debug event generated by device.\n");
+ uint64_t paddr = dp->shared->phys_addr;
+ vmxnet3_bar1_put32(dp, VMXNET3_REG_DSAL, paddr);
+ vmxnet3_bar1_put32(dp, VMXNET3_REG_DSAH, paddr >> 32);
+
+ /* Get worst case zone for queue descriptors,
+ since can't be resized later */
+ rte_snprintf(z_name, sizeof(z_name), "vmxnet3_queues_%u",
+ eth_dev->data->port_id);
+
+ size = VMXNET3_MAX_TX_QUEUES * sizeof(struct Vmxnet3_TxQueueDesc);
+ size += VMXNET3_MAX_RX_QUEUES * sizeof(struct Vmxnet3_RxQueueDesc);
+ dp->qdesc = dma_zone_reserve(z_name, size, SOCKET_ID_ANY,
+ VMXNET3_QUEUE_DESC_ALIGN);
+ if (!dp->qdesc)
+ return -ENOMEM;
+ return 0;
}
+
+static struct eth_driver rte_vmxnet3_pmd = {
+ {
+ .name = "rte_vmxnet3_pmd",
+ .id_table = pci_id_vmxnet3_map,
+#ifdef RTE_EAL_UNBIND_PORTS
+ .drv_flags = RTE_PCI_DRV_NEED_IGB_UIO,
#endif
+ },
+ .eth_dev_init = vmxnet3_dev_init,
+ .dev_private_size = sizeof(struct vmxnet3_dev),
+};
+
+int rte_vmxnet3_pmd_init(void)
+{
+ PMD_INIT_FUNC_TRACE();
+
+ rte_eth_driver_register(&rte_vmxnet3_pmd);
+ return 0;
+}
--- a/lib/librte_pmd_vmxnet3/vmxnet3_ethdev.h 2014-05-14 11:46:47.129439301 -0700
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,187 +0,0 @@
-/*-
- * BSD LICENSE
- *
- * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * * Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- * * Redistributions in binary form must reproduce the above copyright
- * notice, this list of conditions and the following disclaimer in
- * the documentation and/or other materials provided with the
- * distribution.
- * * Neither the name of Intel Corporation nor the names of its
- * contributors may be used to endorse or promote products derived
- * from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#ifndef _VMXNET3_ETHDEV_H_
-#define _VMXNET3_ETHDEV_H_
-
-#ifdef RTE_LIBRTE_VMXNET3_DEBUG_DRIVER
-#define VMXNET3_ASSERT(x) do { \
- if(!(x)) rte_panic("VMXNET3: x"); \
-} while(0)
-#endif
-
-#define VMXNET3_MAX_MAC_ADDRS 1
-
-/* UPT feature to negotiate */
-#define VMXNET3_F_RXCSUM 0x0001
-#define VMXNET3_F_RSS 0x0002
-#define VMXNET3_F_RXVLAN 0x0004
-#define VMXNET3_F_LRO 0x0008
-
-/* Hash Types supported by device */
-#define VMXNET3_RSS_HASH_TYPE_NONE 0x0
-#define VMXNET3_RSS_HASH_TYPE_IPV4 0x01
-#define VMXNET3_RSS_HASH_TYPE_TCP_IPV4 0x02
-#define VMXNET3_RSS_HASH_TYPE_IPV6 0x04
-#define VMXNET3_RSS_HASH_TYPE_TCP_IPV6 0x08
-
-#define VMXNET3_RSS_HASH_FUNC_NONE 0x0
-#define VMXNET3_RSS_HASH_FUNC_TOEPLITZ 0x01
-
-#define VMXNET3_RSS_MAX_KEY_SIZE 40
-#define VMXNET3_RSS_MAX_IND_TABLE_SIZE 128
-
-/* RSS configuration structure - shared with device through GPA */
-typedef
-struct VMXNET3_RSSConf {
- uint16_t hashType;
- uint16_t hashFunc;
- uint16_t hashKeySize;
- uint16_t indTableSize;
- uint8_t hashKey[VMXNET3_RSS_MAX_KEY_SIZE];
- /*
- * indTable is only element that can be changed without
- * device quiesce-reset-update-activation cycle
- */
- uint8_t indTable[VMXNET3_RSS_MAX_IND_TABLE_SIZE];
-} VMXNET3_RSSConf;
-
-typedef
-struct vmxnet3_mf_table {
- void *mfTableBase; /* Multicast addresses list */
- uint64_t mfTablePA; /* Physical address of the list */
- uint16_t num_addrs; /* number of multicast addrs */
-} vmxnet3_mf_table_t;
-
-struct vmxnet3_hw {
-
- uint8_t *hw_addr0; /* BAR0: PT-Passthrough Regs */
- uint8_t *hw_addr1; /* BAR1: VD-Virtual Device Regs */
- /* BAR2: MSI-X Regs */
- /* BAR3: Port IO */
- void *back;
-
- uint16_t device_id;
- uint16_t vendor_id;
- uint16_t subsystem_device_id;
- uint16_t subsystem_vendor_id;
- bool adapter_stopped;
-
- uint8_t perm_addr[ETHER_ADDR_LEN];
- uint8_t num_tx_queues;
- uint8_t num_rx_queues;
- uint8_t bufs_per_pkt;
- uint16_t cur_mtu;
-
- Vmxnet3_TxQueueDesc *tqd_start; /* start address of all tx queue desc */
- Vmxnet3_RxQueueDesc *rqd_start; /* start address of all rx queue desc */
-
- Vmxnet3_DriverShared *shared;
- uint64_t sharedPA;
-
- uint64_t queueDescPA;
- uint16_t queue_desc_len;
-
- VMXNET3_RSSConf *rss_conf;
- uint64_t rss_confPA;
- vmxnet3_mf_table_t *mf_table;
-};
-
-/*
- * Structure to store private data for each driver instance (for each port).
- */
-struct vmxnet3_adapter {
- struct vmxnet3_hw hw;
-};
-
-#define VMXNET3_DEV_PRIVATE_TO_HW(adapter)\
- (&((struct vmxnet3_adapter *)adapter)->hw)
-
-#define VMXNET3_GET_ADDR_LO(reg) ((uint32_t)(reg))
-#define VMXNET3_GET_ADDR_HI(reg) ((uint32_t)(((uint64_t)(reg)) >> 32))
-
-/* Config space read/writes */
-
-#define VMXNET3_PCI_REG(reg) (*((volatile uint32_t *)(reg)))
-
-static inline uint32_t vmxnet3_read_addr(volatile void* addr)
-{
- return VMXNET3_PCI_REG(addr);
-}
-
-#define VMXNET3_PCI_REG_WRITE(reg, value) do { \
- VMXNET3_PCI_REG((reg)) = (value); \
-} while(0)
-
-#define VMXNET3_PCI_BAR0_REG_ADDR(hw, reg) \
- ((volatile uint32_t *)((char *)(hw)->hw_addr0 + (reg)))
-#define VMXNET3_READ_BAR0_REG(hw, reg) \
- vmxnet3_read_addr(VMXNET3_PCI_BAR0_REG_ADDR((hw), (reg)))
-#define VMXNET3_WRITE_BAR0_REG(hw, reg, value) \
- VMXNET3_PCI_REG_WRITE(VMXNET3_PCI_BAR0_REG_ADDR((hw), (reg)), (value))
-
-#define VMXNET3_PCI_BAR1_REG_ADDR(hw, reg) \
- ((volatile uint32_t *)((char *)(hw)->hw_addr1 + (reg)))
-#define VMXNET3_READ_BAR1_REG(hw, reg) \
- vmxnet3_read_addr(VMXNET3_PCI_BAR1_REG_ADDR((hw), (reg)))
-#define VMXNET3_WRITE_BAR1_REG(hw, reg, value) \
- VMXNET3_PCI_REG_WRITE(VMXNET3_PCI_BAR1_REG_ADDR((hw), (reg)), (value))
-
-/*
- * RX/TX function prototypes
- */
-
-void vmxnet3_dev_clear_queues(struct rte_eth_dev *dev);
-
-void vmxnet3_dev_rx_queue_release(void *rxq);
-void vmxnet3_dev_tx_queue_release(void *txq);
-
-int vmxnet3_dev_rx_queue_setup(struct rte_eth_dev *dev, uint16_t rx_queue_id,
- uint16_t nb_rx_desc, unsigned int socket_id,
- const struct rte_eth_rxconf *rx_conf,
- struct rte_mempool *mb_pool);
-int vmxnet3_dev_tx_queue_setup(struct rte_eth_dev *dev, uint16_t tx_queue_id,
- uint16_t nb_tx_desc, unsigned int socket_id,
- const struct rte_eth_txconf *tx_conf);
-
-int vmxnet3_dev_rxtx_init(struct rte_eth_dev *dev);
-
-int vmxnet3_rss_configure(struct rte_eth_dev *dev);
-int vmxnet3_vlan_configure(struct rte_eth_dev *dev);
-
-uint16_t vmxnet3_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
- uint16_t nb_pkts);
-uint16_t vmxnet3_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
- uint16_t nb_pkts);
-
-#endif /* _VMXNET3_ETHDEV_H_ */
--- a/lib/librte_pmd_vmxnet3/vmxnet3_logs.h 2014-05-14 11:46:47.129439301 -0700
+++ b/lib/librte_pmd_vmxnet3/vmxnet3_logs.h 2014-05-14 11:46:47.125439289 -0700
@@ -1,13 +1,13 @@
/*-
* BSD LICENSE
- *
- * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+ *
+ * Copyright(c) 2010-2012 Intel Corporation. All rights reserved.
* All rights reserved.
- *
+ *
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
- *
+ *
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above copyright
@@ -17,7 +17,7 @@
* * Neither the name of Intel Corporation nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
- *
+ *
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
@@ -29,6 +29,7 @@
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
*/
#ifndef _VMXNET3_LOGS_H_
--- a/lib/librte_pmd_vmxnet3/vmxnet3_ring.h 2014-05-14 11:46:47.129439301 -0700
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,176 +0,0 @@
-/*-
- * BSD LICENSE
- *
- * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * * Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- * * Redistributions in binary form must reproduce the above copyright
- * notice, this list of conditions and the following disclaimer in
- * the documentation and/or other materials provided with the
- * distribution.
- * * Neither the name of Intel Corporation nor the names of its
- * contributors may be used to endorse or promote products derived
- * from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
- * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
- * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
- * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
- * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
- * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
- * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
- * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- */
-
-#ifndef _VMXNET3_RING_H_
-#define _VMXNET3_RING_H_
-
-#define VMXNET3_RX_CMDRING_SIZE 2
-
-#define VMXNET3_DRIVER_VERSION_NUM 0x01012000
-
-/* Default ring size */
-#define VMXNET3_DEF_TX_RING_SIZE 512
-#define VMXNET3_DEF_RX_RING_SIZE 128
-
-#define VMXNET3_SUCCESS 0
-#define VMXNET3_FAIL -1
-
-#define TRUE 1
-#define FALSE 0
-
-
-typedef struct vmxnet3_buf_info {
- uint16_t len;
- struct rte_mbuf *m;
- uint64_t bufPA;
-}vmxnet3_buf_info_t;
-
-typedef struct vmxnet3_cmd_ring {
- vmxnet3_buf_info_t *buf_info;
- uint32_t size;
- uint32_t next2fill;
- uint32_t next2comp;
- uint8_t gen;
- uint8_t rid;
- Vmxnet3_GenericDesc *base;
- uint64_t basePA;
-} vmxnet3_cmd_ring_t;
-
-static inline void
-vmxnet3_cmd_ring_adv_next2fill(struct vmxnet3_cmd_ring *ring)
-{
- ring->next2fill++;
- if (unlikely(ring->next2fill == ring->size)) {
- ring->next2fill = 0;
- ring->gen = (uint8_t)(ring->gen ^ 1);
- }
-}
-
-static inline void
-vmxnet3_cmd_ring_adv_next2comp(struct vmxnet3_cmd_ring *ring)
-{
- VMXNET3_INC_RING_IDX_ONLY(ring->next2comp, ring->size);
-}
-
-static inline uint32_t
-vmxnet3_cmd_ring_desc_avail(struct vmxnet3_cmd_ring *ring)
-{
- return (ring->next2comp > ring->next2fill ? 0 : ring->size) +
- ring->next2comp - ring->next2fill - 1;
-}
-
-static inline bool
-vmxnet3_cmd_ring_desc_empty(struct vmxnet3_cmd_ring *ring)
-{
- return (ring->next2comp == ring->next2fill);
-}
-
-typedef struct vmxnet3_comp_ring {
- uint32_t size;
- uint32_t next2proc;
- uint8_t gen;
- uint8_t intr_idx;
- Vmxnet3_GenericDesc *base;
- uint64_t basePA;
-} vmxnet3_comp_ring_t;
-
-static inline void
-vmxnet3_comp_ring_adv_next2proc(struct vmxnet3_comp_ring *ring)
-{
- ring->next2proc++;
- if (unlikely(ring->next2proc == ring->size)) {
- ring->next2proc = 0;
- ring->gen = (uint8_t)(ring->gen ^ 1);
- }
-}
-
-struct vmxnet3_txq_stats {
- uint64_t drop_total; /* # of pkts dropped by the driver, the
- * counters below track droppings due to
- * different reasons
- */
- uint64_t drop_oversized;
- uint64_t drop_hdr_inspect_err;
- uint64_t drop_tso;
- uint64_t deferred;
- uint64_t tx_ring_full;
- uint64_t linearized; /* # of pkts linearized */
-};
-
-typedef struct vmxnet3_tx_ctx {
- int ip_type;
- bool is_vlan;
- bool is_cso;
-
- uint16_t evl_tag; /* only valid when is_vlan == TRUE */
- uint32_t eth_hdr_size; /* only valid for pkts requesting tso or csum
- * offloading */
- uint32_t ip_hdr_size;
- uint32_t l4_hdr_size;
-} vmxnet3_tx_ctx_t;
-
-typedef struct vmxnet3_tx_queue {
- struct vmxnet3_hw *hw;
- struct vmxnet3_cmd_ring cmd_ring;
- struct vmxnet3_comp_ring comp_ring;
- uint32_t qid;
- struct Vmxnet3_TxQueueDesc *shared;
- struct vmxnet3_txq_stats stats;
- bool stopped;
- uint16_t queue_id; /**< Device TX queue index. */
- uint8_t port_id; /**< Device port identifier. */
-} vmxnet3_tx_queue_t;
-
-
-struct vmxnet3_rxq_stats {
- uint64_t drop_total;
- uint64_t drop_err;
- uint64_t drop_fcs;
- uint64_t rx_buf_alloc_failure;
-};
-
-typedef struct vmxnet3_rx_queue {
- struct rte_mempool *mp;
- struct vmxnet3_hw *hw;
- struct vmxnet3_cmd_ring cmd_ring[VMXNET3_RX_CMDRING_SIZE];
- struct vmxnet3_comp_ring comp_ring;
- uint32_t qid1;
- uint32_t qid2;
- Vmxnet3_RxQueueDesc *shared;
- struct vmxnet3_rxq_stats stats;
- bool stopped;
- uint16_t queue_id; /**< Device RX queue index. */
- uint8_t port_id; /**< Device port identifier. */
-} vmxnet3_rx_queue_t;
-
-#endif /* _VMXNET3_RING_H_ */
--- a/lib/librte_pmd_vmxnet3/vmxnet3_rxtx.c 2014-05-14 11:46:47.129439301 -0700
+++ b/lib/librte_pmd_vmxnet3/vmxnet3_rxtx.c 2014-05-14 11:47:23.369551831 -0700
@@ -1,9 +1,9 @@
/*-
* BSD LICENSE
*
- * Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+ * Copyright (c) 2012-2014 Brocade Communications Systems, Inc.
* All rights reserved.
- *
+ *
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
@@ -32,923 +32,782 @@
*/
#include <sys/queue.h>
-
-#include <endian.h>
#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
#include <errno.h>
#include <stdint.h>
#include <stdarg.h>
-#include <unistd.h>
-#include <inttypes.h>
+#include <string.h>
-#include <rte_byteorder.h>
#include <rte_common.h>
-#include <rte_cycles.h>
+#include <rte_interrupts.h>
+#include <rte_byteorder.h>
#include <rte_log.h>
#include <rte_debug.h>
-#include <rte_interrupts.h>
#include <rte_pci.h>
+#include <rte_ether.h>
+#include <rte_ethdev.h>
#include <rte_memory.h>
#include <rte_memzone.h>
-#include <rte_launch.h>
#include <rte_tailq.h>
#include <rte_eal.h>
-#include <rte_per_lcore.h>
-#include <rte_lcore.h>
#include <rte_atomic.h>
-#include <rte_branch_prediction.h>
-#include <rte_ring.h>
-#include <rte_mempool.h>
+#include <rte_spinlock.h>
#include <rte_malloc.h>
-#include <rte_mbuf.h>
-#include <rte_ether.h>
-#include <rte_ethdev.h>
-#include <rte_prefetch.h>
-#include <rte_udp.h>
-#include <rte_tcp.h>
-#include <rte_sctp.h>
#include <rte_string_fns.h>
-#include <rte_errno.h>
-
-#include "vmxnet3/vmxnet3_defs.h"
-#include "vmxnet3_ring.h"
-#include "vmxnet3_logs.h"
-#include "vmxnet3_ethdev.h"
+#include "vmxnet3_dev.h"
+#ifdef RTE_PMD_PACKET_PREFETCH
+#define rte_ring_prefetch(p) rte_prefetch0(p)
+#define rte_packet_prefetch(p) rte_prefetch1(p)
+#else
+#define rte_ring_prefetch(p) do {} while (0)
+#define rte_packet_prefetch(p) do {} while (0)
+#endif
-#define RTE_MBUF_DATA_DMA_ADDR(mb) \
- (uint64_t) ((mb)->buf_physaddr + (uint64_t)((char *)((mb)->pkt.data) - \
- (char *)(mb)->buf_addr))
+#define RTE_MBUF_DATA_DMA_ADDR(mb) \
+ (uint64_t) ((mb)->buf_physaddr + \
+ (uint64_t) ((char *)((mb)->pkt.data) - (char *)(mb)->buf_addr))
#define RTE_MBUF_DATA_DMA_ADDR_DEFAULT(mb) \
(uint64_t) ((mb)->buf_physaddr + RTE_PKTMBUF_HEADROOM)
-static uint32_t rxprod_reg[2] = {VMXNET3_REG_RXPROD, VMXNET3_REG_RXPROD2};
-
-static inline int vmxnet3_post_rx_bufs(vmxnet3_rx_queue_t* , uint8_t);
-static inline void vmxnet3_tq_tx_complete(vmxnet3_tx_queue_t *);
-#ifdef RTE_LIBRTE_VMXNET3_DEBUG_DRIVER
-static void vmxnet3_rxq_dump(struct vmxnet3_rx_queue *);
-static void vmxnet3_txq_dump(struct vmxnet3_tx_queue *);
-#endif
-
-static inline struct rte_mbuf *
+/* Allocate mbuf from pool and do sanity checks */
+static struct rte_mbuf *
rte_rxmbuf_alloc(struct rte_mempool *mp)
{
struct rte_mbuf *m;
m = __rte_mbuf_raw_alloc(mp);
__rte_mbuf_sanity_check_raw(m, RTE_MBUF_PKT, 0);
- return (m);
+ return m;
}
-#ifdef RTE_LIBRTE_VMXNET3_DEBUG_DRIVER
-static void
-vmxnet3_rxq_dump(struct vmxnet3_rx_queue *rxq)
+/* Determine max size for mbuf's allocated from pool */
+static uint16_t rte_mbuf_buf_size(const struct rte_mempool *pool)
{
- uint32_t avail = 0;
- if (rxq == NULL)
- return;
-
- PMD_RX_LOG(DEBUG, "RXQ: cmd0 base : 0x%p cmd1 base : 0x%p comp ring base : 0x%p.\n",
- rxq->cmd_ring[0].base, rxq->cmd_ring[1].base, rxq->comp_ring.base);
- PMD_RX_LOG(DEBUG, "RXQ: cmd0 basePA : 0x%lx cmd1 basePA : 0x%lx comp ring basePA : 0x%lx.\n",
- (unsigned long)rxq->cmd_ring[0].basePA, (unsigned long)rxq->cmd_ring[1].basePA,
- (unsigned long)rxq->comp_ring.basePA);
-
- avail = vmxnet3_cmd_ring_desc_avail(&rxq->cmd_ring[0]);
- PMD_RX_LOG(DEBUG, "RXQ:cmd0: size=%u; free=%u; next2proc=%u; queued=%u\n",
- (uint32_t)rxq->cmd_ring[0].size, avail, rxq->comp_ring.next2proc,
- rxq->cmd_ring[0].size - avail);
-
- avail = vmxnet3_cmd_ring_desc_avail(&rxq->cmd_ring[1]);
- PMD_RX_LOG(DEBUG, "RXQ:cmd1 size=%u; free=%u; next2proc=%u; queued=%u\n",
- (uint32_t)rxq->cmd_ring[1].size, avail, rxq->comp_ring.next2proc,
- rxq->cmd_ring[1].size - avail);
+ const struct rte_pktmbuf_pool_private *mbp_priv
+ = (const struct rte_pktmbuf_pool_private *)
+ ((const char *)pool + sizeof(struct rte_mempool));
+ return mbp_priv->mbuf_data_room_size - RTE_PKTMBUF_HEADROOM;
}
-static void
-vmxnet3_txq_dump(struct vmxnet3_tx_queue *txq)
+/* Since this is pure para-virtualized device, don't need a real hardware
+ * barrier. But do need keep compiler from doing silly optimizations
+ */
+#ifdef __GNUC__
+#define mem_barrier() asm volatile("": : :"memory")
+#else
+#define mem_barrier() __memory_barrier() /* Intel intrinsic */
+#endif
+
+/*
+ * Create memzone for HW rings. malloc can't be used as the physical address is
+ * needed. If the memzone is already created, then this function returns a ptr
+ * to the old one.
+ */
+static const struct rte_memzone *
+ring_dma_zone_reserve(struct rte_eth_dev *dev, const char *ring_name,
+ uint16_t queue_id, uint32_t ring_size, int socket_id)
{
- uint32_t avail = 0;
- if (txq == NULL)
- return;
+ char z_name[RTE_MEMZONE_NAMESIZE];
+ const struct rte_memzone *mz;
- PMD_TX_LOG(DEBUG, "TXQ: cmd base : 0x%p comp ring base : 0x%p.\n",
- txq->cmd_ring.base, txq->comp_ring.base);
- PMD_TX_LOG(DEBUG, "TXQ: cmd basePA : 0x%lx comp ring basePA : 0x%lx.\n",
- (unsigned long)txq->cmd_ring.basePA, (unsigned long)txq->comp_ring.basePA);
-
- avail = vmxnet3_cmd_ring_desc_avail(&txq->cmd_ring);
- PMD_TX_LOG(DEBUG, "TXQ: size=%u; free=%u; next2proc=%u; queued=%u\n",
- (uint32_t)txq->cmd_ring.size, avail,
- txq->comp_ring.next2proc, txq->cmd_ring.size - avail);
+ rte_snprintf(z_name, sizeof(z_name), "%s_%s_%d_%d",
+ dev->driver->pci_drv.name, ring_name,
+ dev->data->port_id, queue_id);
+
+ mz = rte_memzone_lookup(z_name);
+ if (mz)
+ return mz;
+
+ return rte_memzone_reserve_aligned(z_name, ring_size, socket_id, 0,
+ VMXNET3_RING_BA_ALIGN);
}
-#endif
-static inline void
-vmxnet3_cmd_ring_release(vmxnet3_cmd_ring_t *ring)
+/* (Re)set dynamic tx_queue fields to defaults */
+static void
+vmxnet3_reset_tx_queue(struct vmxnet3_tx_queue *txq)
{
- while (ring->next2comp != ring->next2fill) {
- /* No need to worry about tx desc ownership, device is quiesced by now. */
- vmxnet3_buf_info_t *buf_info = ring->buf_info + ring->next2comp;
- if(buf_info->m) {
- rte_pktmbuf_free(buf_info->m);
- buf_info->m = NULL;
- buf_info->bufPA = 0;
- buf_info->len = 0;
+ txq->tx_ring.next = 0;
+ txq->next_to_clean = 0;
+ txq->tx_ring.gen = VMXNET3_INIT_GEN;
+ memset(txq->tx_ring.base, 0,
+ sizeof(Vmxnet3_GenericDesc) * txq->nb_tx_desc);
+
+ txq->comp_ring.next = 0;
+ txq->comp_ring.gen = VMXNET3_INIT_GEN;
+ memset(txq->comp_ring.base, 0,
+ sizeof(Vmxnet3_GenericDesc) * txq->nb_tx_desc);
+}
+
+static void vmxnet3_tx_queue_flush(struct vmxnet3_tx_queue *txq)
+{
+ /* Free up outstanding transmits */
+ while (txq->next_to_clean != txq->tx_ring.next) {
+ struct rte_mbuf *mb = txq->sw_ring[txq->next_to_clean];
+ if (mb) {
+ rte_pktmbuf_free_seg(mb);
+ txq->sw_ring[txq->next_to_clean] = NULL;
}
- vmxnet3_cmd_ring_adv_next2comp(ring);
+ txq->next_to_clean = (txq->next_to_clean + 1)
+ & (txq->tx_ring.size - 1);
}
- rte_free(ring->buf_info);
-}
-void
-vmxnet3_dev_tx_queue_release(void *txq)
-{
- vmxnet3_tx_queue_t *tq = txq;
- if (txq != NULL) {
- /* Release the cmd_ring */
- vmxnet3_cmd_ring_release(&tq->cmd_ring);
- }
+ vmxnet3_reset_tx_queue(txq);
}
-void
-vmxnet3_dev_rx_queue_release(void *rxq)
+void vmxnet3_tx_flush_all(struct rte_eth_dev *dev)
{
- int i;
- vmxnet3_rx_queue_t *rq = rxq;
- if (rxq != NULL) {
- /* Release both the cmd_rings */
- for (i = 0; i < VMXNET3_RX_CMDRING_SIZE; i++)
- vmxnet3_cmd_ring_release(&rq->cmd_ring[i]);
- }
+ struct vmxnet3_dev *dp = dev->data->dev_private;
+ unsigned i;
+
+ for (i = 0; i < dev->data->nb_tx_queues; i++)
+ vmxnet3_tx_queue_flush(dp->tx_queue + i);
}
void
-vmxnet3_dev_clear_queues(struct rte_eth_dev *dev)
+vmxnet3_tx_queue_release(void *arg)
{
- unsigned i;
+ struct vmxnet3_tx_queue *txq = arg;
PMD_INIT_FUNC_TRACE();
- for (i = 0; i < dev->data->nb_tx_queues; i++) {
- struct vmxnet3_tx_queue *txq = dev->data->tx_queues[i];
- if (txq != NULL) {
- txq->stopped = TRUE;
- vmxnet3_dev_tx_queue_release(txq);
- }
- }
+ if (txq == NULL)
+ return;
- for (i = 0; i < dev->data->nb_rx_queues; i++) {
- struct vmxnet3_rx_queue *rxq = dev->data->rx_queues[i];
- if(rxq != NULL) {
- rxq->stopped = TRUE;
- vmxnet3_dev_rx_queue_release(rxq);
- }
- }
+ vmxnet3_tx_queue_flush(txq);
+ rte_free(txq->sw_ring);
}
-static inline void
-vmxnet3_tq_tx_complete(vmxnet3_tx_queue_t *txq)
-{
- int completed = 0;
- struct rte_mbuf *mbuf;
- vmxnet3_comp_ring_t *comp_ring = &txq->comp_ring;
- struct Vmxnet3_TxCompDesc *tcd = (struct Vmxnet3_TxCompDesc *)
- (comp_ring->base + comp_ring->next2proc);
-
- while (tcd->gen == comp_ring->gen) {
-
- /* Release cmd_ring descriptor and free mbuf */
-#ifdef RTE_LIBRTE_VMXNET3_DEBUG_DRIVER
- VMXNET3_ASSERT(txq->cmd_ring.base[tcd->txdIdx].txd.eop == 1);
-#endif
- mbuf = txq->cmd_ring.buf_info[tcd->txdIdx].m;
- if (unlikely(mbuf == NULL))
- rte_panic("EOP desc does not point to a valid mbuf");
- else
- rte_pktmbuf_free(mbuf);
-
-
- txq->cmd_ring.buf_info[tcd->txdIdx].m = NULL;
- /* Mark the txd for which tcd was generated as completed */
- vmxnet3_cmd_ring_adv_next2comp(&txq->cmd_ring);
+int
+vmxnet3_tx_queue_setup(struct rte_eth_dev *dev,
+ uint16_t queue_idx,
+ uint16_t nb_desc,
+ unsigned int socket_id,
+ const struct rte_eth_txconf *tx_conf)
+{
+ struct vmxnet3_dev *dp = dev->data->dev_private;
+ struct vmxnet3_tx_queue *txq = dp->tx_queue + queue_idx;
+ const struct rte_memzone *tz;
- vmxnet3_comp_ring_adv_next2proc(comp_ring);
- tcd = (struct Vmxnet3_TxCompDesc *)(comp_ring->base +
- comp_ring->next2proc);
- completed++;
- }
+ PMD_INIT_FUNC_TRACE();
- PMD_TX_LOG(DEBUG, "Processed %d tx comps & command descs.\n", completed);
-}
+ /* Device does not support checksum offload (yet) */
+ if ((tx_conf->txq_flags & ETH_TXQ_FLAGS_NOXSUMS)
+ != ETH_TXQ_FLAGS_NOXSUMS) {
+ PMD_INIT_LOG(ERR, "TX Multi segment not support yet\n");
+ return -EINVAL;
+ }
-uint16_t
-vmxnet3_xmit_pkts( void *tx_queue, struct rte_mbuf **tx_pkts,
- uint16_t nb_pkts)
-{
- uint16_t nb_tx;
- Vmxnet3_TxDesc *txd = NULL;
- vmxnet3_buf_info_t *tbi = NULL;
- struct vmxnet3_hw *hw;
- struct rte_mbuf *txm;
- vmxnet3_tx_queue_t *txq = tx_queue;
+ /* Validate number of transmit descriptors. */
+ if (nb_desc == 0 || nb_desc > VMXNET3_TX_RING_MAX_SIZE ||
+ (nb_desc & (nb_desc - 1)) != 0) { /* power of 2 */
+ PMD_INIT_LOG(ERR, "Invalid number of TX descriptors\n");
+ return -EINVAL;
+ }
- hw = txq->hw;
+ /* Round up to VMXNET3_RING_SIZE_ALIGN */
+ nb_desc = (nb_desc + VMXNET3_RING_SIZE_MASK) & ~VMXNET3_RING_SIZE_MASK;
- if(txq->stopped) {
- PMD_TX_LOG(DEBUG, "Tx queue is stopped.\n");
- return 0;
+ /* Free memory prior to re-allocation if needed... */
+ vmxnet3_tx_queue_release(dev->data->tx_queues[queue_idx]);
+ dev->data->tx_queues[queue_idx] = NULL;
+
+ /*
+ * Allocate TX ring hardware descriptors. A memzone large enough to
+ * handle the maximum ring size is allocated in order to allow for
+ * resizing in later calls to the queue setup function.
+ */
+ tz = ring_dma_zone_reserve(dev, "tx_ring", queue_idx,
+ sizeof(Vmxnet3_TxDesc) * VMXNET3_TX_RING_MAX_SIZE,
+ socket_id);
+ if (tz == NULL) {
+ vmxnet3_tx_queue_release(txq);
+ return -ENOMEM;
}
- /* Free up the comp_descriptors aggressively */
- vmxnet3_tq_tx_complete(txq);
-
- nb_tx = 0;
- while(nb_tx < nb_pkts) {
+ txq->queue_id = queue_idx;
+ txq->port_id = dev->data->port_id;
+ txq->nb_tx_desc = nb_desc;
+ txq->tx_free_thresh = tx_conf->tx_free_thresh ? : 32;
+ txq->tx_prod = dp->bar0 + VMXNET3_REG_TXPROD + txq->queue_id * 8;
+
+ txq->tx_ring.base = tz->addr;
+ txq->tx_ring.size = nb_desc;
+ txq->tx_ring.phys_addr = tz->phys_addr;
+
+ tz = ring_dma_zone_reserve(dev, "tx_comp", queue_idx,
+ sizeof(Vmxnet3_TxCompDesc) * VMXNET3_TX_RING_MAX_SIZE,
+ socket_id);
+ if (tz == NULL) {
+ vmxnet3_tx_queue_release(txq);
+ return -ENOMEM;
+ }
+ txq->comp_ring.base = tz->addr;
+ txq->comp_ring.size = nb_desc;
+ txq->comp_ring.phys_addr = tz->phys_addr;
+
+ /* Allocate software ring
+ * Only needs to be sized based on request, since it is in
+ * freeable memory.
+ */
+ txq->sw_ring = rte_zmalloc_socket("txq_sw_ring",
+ sizeof(struct rte_mbuf *) * nb_desc,
+ CACHE_LINE_SIZE, socket_id);
+ if (txq->sw_ring == NULL) {
+ vmxnet3_tx_queue_release(txq);
+ return -ENOMEM;
+ }
- if(vmxnet3_cmd_ring_desc_avail(&txq->cmd_ring)) {
+ PMD_INIT_LOG(DEBUG, "sw_ring=%p hw_ring=%p dma_addr=0x%"PRIx64,
+ txq->sw_ring, txq->tx_ring.base, txq->tx_ring.phys_addr);
- txm = tx_pkts[nb_tx];
- /* Don't support scatter packets yet, free them if met */
- if (txm->pkt.nb_segs != 1) {
- PMD_TX_LOG(DEBUG, "Don't support scatter packets yet, drop!\n");
- rte_pktmbuf_free(tx_pkts[nb_tx]);
- txq->stats.drop_total++;
+ vmxnet3_reset_tx_queue(txq);
+ dev->data->tx_queues[queue_idx] = txq;
- nb_tx++;
- continue;
- }
+ return 0;
+}
- /* Needs to minus ether header len */
- if(txm->pkt.data_len > (hw->cur_mtu + ETHER_HDR_LEN)) {
- PMD_TX_LOG(DEBUG, "Packet data_len higher than MTU\n");
- rte_pktmbuf_free(tx_pkts[nb_tx]);
- txq->stats.drop_total++;
+/*
+ * Read transmit completion ring and free up mbufs from completed
+ * transmits.
+ */
+static void vmxnet3_xmit_cleanup(struct vmxnet3_tx_queue *txq)
+{
+ const Vmxnet3_TxCompDesc *tcd;
+ uint16_t last_completed;
- nb_tx++;
- continue;
- }
+ tcd = (Vmxnet3_TxCompDesc *)
+ (txq->comp_ring.base + txq->comp_ring.next);
- txd = (Vmxnet3_TxDesc *)(txq->cmd_ring.base + txq->cmd_ring.next2fill);
+ if (tcd->gen != txq->comp_ring.gen)
+ return;
- /* Fill the tx descriptor */
- tbi = txq->cmd_ring.buf_info + txq->cmd_ring.next2fill;
- tbi->bufPA = RTE_MBUF_DATA_DMA_ADDR(txm);
- txd->addr = tbi->bufPA;
- txd->len = txm->pkt.data_len;
-
- /* Mark the last descriptor as End of Packet. */
- txd->cq = 1;
- txd->eop = 1;
-
- /* Record current mbuf for freeing it later in tx complete */
-#ifdef RTE_LIBRTE_VMXNET3_DEBUG_DRIVER
- VMXNET3_ASSERT(txm);
-#endif
- tbi->m = txm;
+ mem_barrier();
- /* Set the offloading mode to default */
- txd->hlen = 0;
- txd->om = VMXNET3_OM_NONE;
- txd->msscof = 0;
-
- /* finally flip the GEN bit of the SOP desc */
- txd->gen = txq->cmd_ring.gen;
- txq->shared->ctrl.txNumDeferred++;
-
- /* move to the next2fill descriptor */
- vmxnet3_cmd_ring_adv_next2fill(&txq->cmd_ring);
- nb_tx++;
+ do {
+ last_completed = tcd->txdIdx;
+ if (last_completed > txq->tx_ring.size)
+ rte_panic("%s(): index %u out of tx ring\n",
+ __func__, last_completed);
+
+ vmxnet3_ring_advance(&txq->comp_ring);
+ tcd = (Vmxnet3_TxCompDesc *)
+ (txq->comp_ring.base + txq->comp_ring.next);
+ } while (tcd->gen == txq->comp_ring.gen);
+
+ PMD_TX_FREE_LOG(DEBUG,
+ "Cleaning TX descriptors: %4u to %4u "
+ "(port=%d queue=%d)",
+ txq->next_to_clean, last_completed,
+ txq->port_id, txq->queue_id);
+
+ struct rte_mbuf **sw_ring = txq->sw_ring;
+ uint16_t i = txq->next_to_clean;
+ for (;;) {
+ rte_pktmbuf_free_seg(sw_ring[i]);
+ sw_ring[i] = NULL;
- } else {
- PMD_TX_LOG(DEBUG, "No free tx cmd desc(s)\n");
- txq->stats.drop_total += (nb_pkts - nb_tx);
+ if (i == last_completed)
break;
- }
- }
-
- PMD_TX_LOG(DEBUG, "vmxnet3 txThreshold: %u", txq->shared->ctrl.txThreshold);
- if (txq->shared->ctrl.txNumDeferred >= txq->shared->ctrl.txThreshold) {
-
- txq->shared->ctrl.txNumDeferred = 0;
- /* Notify vSwitch that packets are available. */
- VMXNET3_WRITE_BAR0_REG(hw, (VMXNET3_REG_TXPROD + txq->queue_id * VMXNET3_REG_ALIGN),
- txq->cmd_ring.next2fill);
+ i = (i + 1) & (txq->tx_ring.size - 1);
}
+ txq->next_to_clean = (i + 1) & (txq->tx_ring.size - 1);
+}
- return (nb_tx);
+static inline unsigned
+vmxnet3_tx_used(const struct vmxnet3_tx_queue *txq)
+{
+ return (txq->tx_ring.next - txq->next_to_clean)
+ & (txq->tx_ring.size - 1);
+}
+
+static inline unsigned
+vmxnet3_tx_avail(const struct vmxnet3_tx_queue *txq)
+{
+ return txq->tx_ring.size - vmxnet3_tx_used(txq) - 1;
}
/*
- * Allocates mbufs and clusters. Post rx descriptors with buffer details
- * so that device can receive packets in those buffers.
- * Ring layout:
- * Among the two rings, 1st ring contains buffers of type 0 and type1.
- * bufs_per_pkt is set such that for non-LRO cases all the buffers required
- * by a frame will fit in 1st ring (1st buf of type0 and rest of type1).
- * 2nd ring contains buffers of type 1 alone. Second ring mostly be used
- * only for LRO.
- *
+ * Transmit a block of packets
*/
-static inline int
-vmxnet3_post_rx_bufs(vmxnet3_rx_queue_t* rxq, uint8_t ring_id)
+uint16_t vmxnet3_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
+ uint16_t nb_pkts)
{
- int err = 0;
- uint32_t i = 0, val = 0;
- struct vmxnet3_cmd_ring *ring = &rxq->cmd_ring[ring_id];
-
- while (vmxnet3_cmd_ring_desc_avail(ring) > 0) {
-
- struct Vmxnet3_RxDesc *rxd;
- struct rte_mbuf *mbuf;
- vmxnet3_buf_info_t *buf_info = &ring->buf_info[ring->next2fill];
- rxd = (struct Vmxnet3_RxDesc *)(ring->base + ring->next2fill);
-
- if (ring->rid == 0) {
- /* Usually: One HEAD type buf per packet
- * val = (ring->next2fill % rxq->hw->bufs_per_pkt) ?
- * VMXNET3_RXD_BTYPE_BODY : VMXNET3_RXD_BTYPE_HEAD;
- */
-
- /* We use single packet buffer so all heads here */
- val = VMXNET3_RXD_BTYPE_HEAD;
- } else {
- /* All BODY type buffers for 2nd ring; which won't be used at all by ESXi */
- val = VMXNET3_RXD_BTYPE_BODY;
- }
+ struct vmxnet3_tx_queue *txq = tx_queue;
+ Vmxnet3_TxQueueCtrl *txqCtrl = txq->shared;
+ unsigned nb_tx, num_desc = 0;
+ uint32_t tx_id;
+
+ if (vmxnet3_tx_used(txq) >= txq->tx_free_thresh)
+ vmxnet3_xmit_cleanup(txq);
+
+ /* TX loop */
+ tx_id = txq->tx_ring.next;
+ for (nb_tx = 0; nb_tx < nb_pkts; nb_tx++) {
+ struct rte_mbuf *tx_pkt = *tx_pkts++;
+ uint32_t tx_first = tx_id;
+ Vmxnet3_GenericDesc *txd;
- /* Allocate blank mbuf for the current Rx Descriptor */
- mbuf = rte_rxmbuf_alloc(rxq->mp);
- if (mbuf == NULL) {
- PMD_RX_LOG(ERR, "Error allocating mbuf in %s\n", __func__);
- rxq->stats.rx_buf_alloc_failure++;
- err = ENOMEM;
+ /*
+ * The number of descriptors that must be allocated for a
+ * packet is the number of segments of that packet.
+ */
+ if (tx_pkt->pkt.nb_segs > vmxnet3_tx_avail(txq))
break;
+
+ PMD_TX_LOG(DEBUG,
+ "port_id=%u queue_id=%u pktlen=%u segs=%u first=%u",
+ txq->port_id, txq->queue_id,
+ tx_pkt->pkt.pkt_len, tx_pkt->pkt.nb_segs, tx_id);
+
+ struct rte_mbuf *m_seg = tx_pkt;
+ uint32_t cur_gen = !txq->tx_ring.gen;
+ do {
+ /* Remember the transmit buffer for cleanup */
+ txq->sw_ring[tx_id] = m_seg;
+
+ /* NB: the following assumes that VMXNET3 maximum
+ transmit buffer size (16K) is greater than
+ maximum sizeof mbuf segment (2K). */
+ uint32_t dw2;
+ dw2 = m_seg->pkt.data_len;
+ dw2 |= cur_gen << VMXNET3_TXD_GEN_SHIFT;
+
+ txd = &txq->tx_ring.base[tx_id];
+ txd->txd.addr = RTE_MBUF_DATA_DMA_ADDR(m_seg);
+ txd->dword[2] = rte_cpu_to_le_32(dw2);
+ txd->dword[3] = 0;
+
+ m_seg = m_seg->pkt.next;
+ tx_id = vmxnet3_ring_advance(&txq->tx_ring);
+ cur_gen = txq->tx_ring.gen;
+
+ ++num_desc;
+ } while (m_seg != NULL);
+
+ /* Update the EOP descriptor */
+ txd->dword[3] |= VMXNET3_TXD_EOP | VMXNET3_TXD_CQ;
+
+ /* Update the SOP descriptor. Must be done last */
+ txd = txq->tx_ring.base + tx_first;
+ if (tx_pkt->ol_flags & PKT_TX_VLAN_PKT) {
+ txd->txd.ti = 1;
+ txd->txd.tci = tx_pkt->pkt.vlan_macip.f.vlan_tci;
}
- /*
- * Load mbuf pointer into buf_info[ring_size]
- * buf_info structure is equivalent to cookie for virtio-virtqueue
- */
- buf_info->m = mbuf;
- buf_info->len = (uint16_t)(mbuf->buf_len -
- RTE_PKTMBUF_HEADROOM);
- buf_info->bufPA = RTE_MBUF_DATA_DMA_ADDR_DEFAULT(mbuf);
-
- /* Load Rx Descriptor with the buffer's GPA */
- rxd->addr = buf_info->bufPA;
-
- /* After this point rxd->addr MUST not be NULL */
- rxd->btype = val;
- rxd->len = buf_info->len;
- /* Flip gen bit at the end to change ownership */
- rxd->gen = ring->gen;
-
- vmxnet3_cmd_ring_adv_next2fill(ring);
- i++;
- }
-
- /* Return error only if no buffers are posted at present */
- if (vmxnet3_cmd_ring_desc_avail(ring) >= (ring->size -1))
- return -err;
- else
- return i;
-}
+ /* TODO: Add transmit checksum offload here */
-/*
- * Process the Rx Completion Ring of given vmxnet3_rx_queue
- * for nb_pkts burst and return the number of packets received
- */
-uint16_t
-vmxnet3_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
-{
- uint16_t nb_rx;
- uint32_t nb_rxd, idx;
- uint8_t ring_idx;
- vmxnet3_rx_queue_t *rxq;
- Vmxnet3_RxCompDesc *rcd;
- vmxnet3_buf_info_t *rbi;
- Vmxnet3_RxDesc *rxd;
- struct rte_mbuf *rxm = NULL;
- struct vmxnet3_hw *hw;
+ /* flip the GEN bit on the SOP */
+ mem_barrier();
+ txd->dword[2] ^= rte_cpu_to_le_32(VMXNET3_TXD_GEN);
+ }
- nb_rx = 0;
- ring_idx = 0;
- nb_rxd = 0;
- idx = 0;
+ /* avoid ring update if no packets sent */
+ if (likely(num_desc > 0)) {
+ /* Increment pending value */
+ uint32_t defer = rte_le_to_cpu_32(txqCtrl->txNumDeferred);
+ defer += num_desc;
+ if (defer >= rte_le_to_cpu_32(txqCtrl->txThreshold)) {
+ txqCtrl->txNumDeferred = 0;
+ vmxnet3_write_reg(txq->tx_prod, tx_id);
+ } else
+ txqCtrl->txNumDeferred = rte_cpu_to_le_32(defer);
+ }
- rxq = rx_queue;
- hw = rxq->hw;
+ return nb_tx;
+}
- rcd = &rxq->comp_ring.base[rxq->comp_ring.next2proc].rcd;
+static void vmxnet3_rx_queue_flush(struct vmxnet3_rx_queue *rxq)
+{
+ unsigned i;
- if(rxq->stopped) {
- PMD_RX_LOG(DEBUG, "Rx queue is stopped.\n");
- return 0;
+ for (i = 0; i < rxq->nb_rx_desc; i++) {
+ struct rte_mbuf *mb = rxq->sw_ring[i];
+ if (mb) {
+ rte_pktmbuf_free_seg(mb);
+ rxq->sw_ring[i] = NULL;
+ }
}
+}
- while (rcd->gen == rxq->comp_ring.gen) {
+void vmxnet3_rx_flush_all(struct rte_eth_dev *dev)
+{
+ struct vmxnet3_dev *dp = dev->data->dev_private;
+ unsigned i;
- if(nb_rx >= nb_pkts)
- break;
- idx = rcd->rxdIdx;
- ring_idx = (uint8_t)((rcd->rqID == rxq->qid1) ? 0 : 1);
- rxd = (Vmxnet3_RxDesc *)rxq->cmd_ring[ring_idx].base + idx;
- rbi = rxq->cmd_ring[ring_idx].buf_info + idx;
+ for (i = 0; i < dev->data->nb_rx_queues; i++)
+ vmxnet3_rx_queue_flush(dp->rx_queue + i);
+}
- if(rcd->sop !=1 || rcd->eop != 1) {
- rte_pktmbuf_free_seg(rbi->m);
+void
+vmxnet3_rx_queue_release(void *arg)
+{
+ struct vmxnet3_rx_queue *rxq = arg;
- PMD_RX_LOG(DEBUG, "Packet spread across multiple buffers\n)");
- goto rcd_done;
+ PMD_INIT_FUNC_TRACE();
- } else {
+ if (rxq == NULL)
+ return;
- PMD_RX_LOG(DEBUG, "rxd idx: %d ring idx: %d.\n", idx, ring_idx);
+ vmxnet3_rx_queue_flush(rxq);
+ rte_free(rxq->sw_ring);
+}
-#ifdef RTE_LIBRTE_VMXNET3_DEBUG_DRIVER
- VMXNET3_ASSERT(rcd->len <= rxd->len);
- VMXNET3_ASSERT(rbi->m);
-#endif
- if (rcd->len == 0) {
- PMD_RX_LOG(DEBUG, "Rx buf was skipped. rxring[%d][%d]\n)",
- ring_idx, idx);
-#ifdef RTE_LIBRTE_VMXNET3_DEBUG_DRIVER
- VMXNET3_ASSERT(rcd->sop && rcd->eop);
-#endif
- rte_pktmbuf_free_seg(rbi->m);
+int
+vmxnet3_rx_queue_setup(struct rte_eth_dev *dev,
+ uint16_t queue_idx,
+ uint16_t nb_desc,
+ unsigned int socket_id,
+ const struct rte_eth_rxconf *rx_conf __rte_unused,
+ struct rte_mempool *mp)
+{
+ struct vmxnet3_dev *dp = dev->data->dev_private;
+ struct vmxnet3_rx_queue *rxq = dp->rx_queue + queue_idx;
+ const struct rte_memzone *rz;
- goto rcd_done;
- }
+ PMD_INIT_FUNC_TRACE();
- /* Assuming a packet is coming in a single packet buffer */
- if (rxd->btype != VMXNET3_RXD_BTYPE_HEAD) {
- PMD_RX_LOG(DEBUG, "Alert : Misbehaving device, incorrect "
- " buffer type used. iPacket dropped.\n");
- rte_pktmbuf_free_seg(rbi->m);
- goto rcd_done;
- }
-#ifdef RTE_LIBRTE_VMXNET3_DEBUG_DRIVER
- VMXNET3_ASSERT(rxd->btype == VMXNET3_RXD_BTYPE_HEAD);
-#endif
- /* Get the packet buffer pointer from buf_info */
- rxm = rbi->m;
+ /* Validate number of receive descriptors. */
+ if (nb_desc == 0 || nb_desc > VMXNET3_RX_RING_MAX_SIZE ||
+ (nb_desc & (nb_desc - 1)) != 0) { /* power of 2 */
+ PMD_INIT_LOG(ERR, "Invalid number of RX descriptors\n");
+ return -EINVAL;
+ }
- /* Clear descriptor associated buf_info to be reused */
- rbi->m = NULL;
- rbi->bufPA = 0;
-
- /* Update the index that we received a packet */
- rxq->cmd_ring[ring_idx].next2comp = idx;
-
- /* For RCD with EOP set, check if there is frame error */
- if (rcd->err) {
- rxq->stats.drop_total++;
- rxq->stats.drop_err++;
-
- if(!rcd->fcs) {
- rxq->stats.drop_fcs++;
- PMD_RX_LOG(ERR, "Recv packet dropped due to frame err.\n");
- }
- PMD_RX_LOG(ERR, "Error in received packet rcd#:%d rxd:%d\n",
- (int)(rcd - (struct Vmxnet3_RxCompDesc *)
- rxq->comp_ring.base), rcd->rxdIdx);
- rte_pktmbuf_free_seg(rxm);
+ if (dev->data->dev_conf.rxmode.max_rx_pkt_len > rte_mbuf_buf_size(mp)) {
+ PMD_INIT_LOG(ERR, "VMXNET3 don't support scatter packets yet\n");
+ return -EINVAL;
+ }
- goto rcd_done;
- }
+ /* Round up to VMXNET3_RING_SIZE_ALIGN */
+ nb_desc = (nb_desc + VMXNET3_RING_SIZE_MASK) & ~VMXNET3_RING_SIZE_MASK;
- /* Check for hardware stripped VLAN tag */
- if (rcd->ts) {
+ /* Free memory from previous allocation */
+ vmxnet3_rx_queue_release(dev->data->rx_queues[queue_idx]);
+ dev->data->rx_queues[queue_idx] = NULL;
+
+ /* First allocate the rx queue data structure */
+ rxq->mb_pool = mp;
+ rxq->nb_rx_desc = nb_desc;
+ rxq->queue_id = queue_idx;
+ rxq->port_id = dev->data->port_id;
- PMD_RX_LOG(ERR, "Received packet with vlan ID: %d.\n",
- rcd->tci);
- rxm->ol_flags = PKT_RX_VLAN_PKT;
-
-#ifdef RTE_LIBRTE_VMXNET3_DEBUG_DRIVER
- VMXNET3_ASSERT(rxm &&
- rte_pktmbuf_mtod(rxm, void *));
-#endif
- //Copy vlan tag in packet buffer
- rxm->pkt.vlan_macip.f.vlan_tci =
- rte_le_to_cpu_16((uint16_t)rcd->tci);
-
- } else
- rxm->ol_flags = 0;
-
- /* Initialize newly received packet buffer */
- rxm->pkt.in_port = rxq->port_id;
- rxm->pkt.nb_segs = 1;
- rxm->pkt.next = NULL;
- rxm->pkt.pkt_len = (uint16_t)rcd->len;
- rxm->pkt.data_len = (uint16_t)rcd->len;
- rxm->pkt.in_port = rxq->port_id;
- rxm->pkt.vlan_macip.f.vlan_tci = 0;
- rxm->pkt.data = (char *)rxm->buf_addr + RTE_PKTMBUF_HEADROOM;
-
- rx_pkts[nb_rx++] = rxm;
-
-rcd_done:
- rxq->cmd_ring[ring_idx].next2comp = idx;
- VMXNET3_INC_RING_IDX_ONLY(rxq->cmd_ring[ring_idx].next2comp, rxq->cmd_ring[ring_idx].size);
-
- /* It's time to allocate some new buf and renew descriptors */
- vmxnet3_post_rx_bufs(rxq, ring_idx);
- if (unlikely(rxq->shared->ctrl.updateRxProd)) {
- VMXNET3_WRITE_BAR0_REG(hw, rxprod_reg[ring_idx] + (rxq->queue_id * VMXNET3_REG_ALIGN),
- rxq->cmd_ring[ring_idx].next2fill);
- }
+ /*
+ * Allocate RX ring hardware descriptors. A memzone large enough to
+ * handle the maximum ring size is allocated in order to allow for
+ * resizing in later calls to the queue setup function.
+ */
+ rz = ring_dma_zone_reserve(dev, "rx_ring", queue_idx,
+ VMXNET3_RX_RING_MAX_SIZE * sizeof(Vmxnet3_RxDesc),
+ socket_id);
+ if (rz == NULL) {
+ vmxnet3_rx_queue_release(rxq);
+ return -ENOMEM;
+ }
+
+ rxq->rx_ring.base = rz->addr;
+ rxq->rx_ring.size = nb_desc;
+ rxq->rx_ring.phys_addr = rz->phys_addr;
+ rxq->rx_prod = dp->bar0 + VMXNET3_REG_RXPROD
+ + rxq->queue_id * VMXNET3_REG_ALIGN;
+
+ rz = ring_dma_zone_reserve(dev, "rx_comp", queue_idx,
+ VMXNET3_RX_RING_MAX_SIZE * sizeof(Vmxnet3_RxCompDesc),
+ socket_id);
+ if (rz == NULL) {
+ vmxnet3_rx_queue_release(rxq);
+ return -ENOMEM;
+ }
+
+ rxq->comp_ring.base = rz->addr;
+ rxq->comp_ring.size = nb_desc;
+ rxq->comp_ring.phys_addr = rz->phys_addr;
+
+ /* Allocate software ring */
+ rxq->sw_ring = rte_zmalloc_socket("rxq_sw_ring",
+ sizeof(struct rte_mbuf *) * nb_desc,
+ CACHE_LINE_SIZE, socket_id);
+ if (rxq->sw_ring == NULL) {
+ vmxnet3_rx_queue_release(rxq);
+ return -ENOMEM;
+ }
- /* Advance to the next descriptor in comp_ring */
- vmxnet3_comp_ring_adv_next2proc(&rxq->comp_ring);
+ PMD_INIT_LOG(DEBUG, "sw_ring=%p hw_ring=%p dma_addr=0x%"PRIx64,
+ rxq->sw_ring, rxq->rx_ring.base, rxq->rx_ring.phys_addr);
- rcd = &rxq->comp_ring.base[rxq->comp_ring.next2proc].rcd;
- nb_rxd++;
- if (nb_rxd > rxq->cmd_ring[0].size) {
- PMD_RX_LOG(ERR, "Used up quota of receiving packets,"
- " relinquish control.\n");
- break;
- }
- }
- }
+ dev->data->rx_queues[queue_idx] = rxq;
- return (nb_rx);
+ return 0;
}
/*
- * Create memzone for device rings. malloc can't be used as the physical address is
- * needed. If the memzone is already created, then this function returns a ptr
- * to the old one.
+ * Setup receive queue ring.
+ * All the slots are filled, but the last one is not marked
+ * as available, to avoid the problem where full == empty.
*/
-static const struct rte_memzone *
-ring_dma_zone_reserve(struct rte_eth_dev *dev, const char *ring_name,
- uint16_t queue_id, uint32_t ring_size, int socket_id)
+static int vmxnet3_rxqueue_init(struct vmxnet3_rx_queue *rxq)
{
- char z_name[RTE_MEMZONE_NAMESIZE];
- const struct rte_memzone *mz;
+ unsigned i;
+ uint16_t buf_size = rte_mbuf_buf_size(rxq->mb_pool);
- rte_snprintf(z_name, sizeof(z_name), "%s_%s_%d_%d",
- dev->driver->pci_drv.name, ring_name,
- dev->data->port_id, queue_id);
+ for (i = 0; i < rxq->rx_ring.size; i++) {
+ struct rte_mbuf *m;
- mz = rte_memzone_lookup(z_name);
- if (mz)
- return mz;
+ m = rxq->sw_ring[i];
+ if (!m) {
+ m = rte_rxmbuf_alloc(rxq->mb_pool);
+ if (!m) {
+ PMD_INIT_LOG(ERR,
+ "RX mbuf alloc failed qid=%hu",
+ rxq->queue_id);
+ return -ENOMEM;
+ }
+ rxq->sw_ring[i] = m;
+ }
+
+ Vmxnet3_GenericDesc *gd = rxq->rx_ring.base + i;
+ uint64_t pa = RTE_MBUF_DATA_DMA_ADDR_DEFAULT(m);
+ gd->rxd.addr = rte_cpu_to_le_64(pa);
+
+ uint32_t dw2 = !VMXNET3_INIT_GEN << VMXNET3_RXD_GEN_SHIFT |
+ VMXNET3_RXD_BTYPE_HEAD << VMXNET3_RXD_BTYPE_SHIFT |
+ buf_size;
+ gd->dword[2] = rte_cpu_to_le_32(dw2);
+
+ if (i == rxq->rx_ring.size - 1)
+ break;
- return rte_memzone_reserve_aligned(z_name, ring_size,
- socket_id, 0, VMXNET3_RING_BA_ALIGN);
+ gd->rxd.gen = VMXNET3_INIT_GEN;
+ vmxnet3_ring_advance(&rxq->rx_ring);
+ }
+
+ return 0;
}
-int
-vmxnet3_dev_tx_queue_setup(struct rte_eth_dev *dev,
- uint16_t queue_idx,
- uint16_t nb_desc,
- unsigned int socket_id,
- __attribute__((unused)) const struct rte_eth_txconf *tx_conf)
+int vmxnet3_rx_init(struct rte_eth_dev *dev)
{
- const struct rte_memzone *mz;
- struct vmxnet3_tx_queue *txq;
- struct vmxnet3_hw *hw;
- struct vmxnet3_cmd_ring *ring;
- struct vmxnet3_comp_ring *comp_ring;
- int size;
+ struct vmxnet3_dev *dp = dev->data->dev_private;
+ unsigned i;
+ int ret;
PMD_INIT_FUNC_TRACE();
- hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
- if ((tx_conf->txq_flags & ETH_TXQ_FLAGS_NOMULTSEGS) !=
- ETH_TXQ_FLAGS_NOMULTSEGS) {
- PMD_INIT_LOG(ERR, "TX Multi segment not support yet\n");
- return (-EINVAL);
- }
+ /* Configure and enable each RX queue. */
+ for (i = 0; i < dev->data->nb_rx_queues; i++) {
+ struct vmxnet3_rx_queue *rxq = dp->rx_queue + i;
- if ((tx_conf->txq_flags & ETH_TXQ_FLAGS_NOOFFLOADS) !=
- ETH_TXQ_FLAGS_NOOFFLOADS) {
- PMD_INIT_LOG(ERR, "TX not support offload function yet\n");
- return (-EINVAL);
+ rxq->rx_ring.next = 0;
+ memset(rxq->rx_ring.base, 0,
+ rxq->rx_ring.size * sizeof(Vmxnet3_RxDesc));
+ rxq->rx_ring.gen = VMXNET3_INIT_GEN;
+ rxq->start_seg = NULL;
+ rxq->prev_seg = &rxq->start_seg;
+
+ /* Allocate buffers for descriptor rings and set up queue */
+ ret = vmxnet3_rxqueue_init(rxq);
+ if (ret)
+ return ret;
+
+ rxq->comp_ring.next = 0;
+ memset(rxq->comp_ring.base, 0,
+ rxq->comp_ring.size * sizeof(Vmxnet3_RxCompDesc));
+ rxq->comp_ring.gen = VMXNET3_INIT_GEN;
}
- txq = rte_zmalloc("ethdev_tx_queue", sizeof(struct vmxnet3_tx_queue), CACHE_LINE_SIZE);
- if (txq == NULL) {
- PMD_INIT_LOG(ERR, "Can not allocate tx queue structure\n");
- return (-ENOMEM);
- }
+ return 0;
+}
- txq->queue_id = queue_idx;
- txq->port_id = dev->data->port_id;
- txq->shared = &hw->tqd_start[queue_idx];
- txq->hw = hw;
- txq->qid = queue_idx;
- txq->stopped = TRUE;
-
- ring = &txq->cmd_ring;
- comp_ring = &txq->comp_ring;
-
- /* Tx vmxnet ring length should be between 512-4096 */
- if (nb_desc < VMXNET3_DEF_TX_RING_SIZE) {
- PMD_INIT_LOG(ERR, "VMXNET3 Tx Ring Size Min: %u\n",
- VMXNET3_DEF_TX_RING_SIZE);
- return -EINVAL;
- } else if (nb_desc > VMXNET3_TX_RING_MAX_SIZE) {
- PMD_INIT_LOG(ERR, "VMXNET3 Tx Ring Size Max: %u\n",
- VMXNET3_TX_RING_MAX_SIZE);
- return -EINVAL;
- } else {
- ring->size = nb_desc;
- ring->size &= ~VMXNET3_RING_SIZE_MASK;
- }
- comp_ring->size = ring->size;
-
- /* Tx vmxnet rings structure initialization*/
- ring->next2fill = 0;
- ring->next2comp = 0;
- ring->gen = VMXNET3_INIT_GEN;
- comp_ring->next2proc = 0;
- comp_ring->gen = VMXNET3_INIT_GEN;
-
- size = sizeof(struct Vmxnet3_TxDesc) * ring->size;
- size += sizeof(struct Vmxnet3_TxCompDesc) * comp_ring->size;
-
- mz = ring_dma_zone_reserve(dev, "txdesc", queue_idx, size, socket_id);
- if (mz == NULL) {
- PMD_INIT_LOG(ERR, "ERROR: Creating queue descriptors zone\n");
- return (-ENOMEM);
- }
- memset(mz->addr, 0, mz->len);
-
- /* cmd_ring initialization */
- ring->base = mz->addr;
- ring->basePA = mz->phys_addr;
-
- /* comp_ring initialization */
- comp_ring->base = ring->base + ring->size;
- comp_ring->basePA = ring->basePA +
- (sizeof(struct Vmxnet3_TxDesc) * ring->size);
-
- /* cmd_ring0 buf_info allocation */
- ring->buf_info = rte_zmalloc("tx_ring_buf_info",
- ring->size * sizeof(vmxnet3_buf_info_t), CACHE_LINE_SIZE);
- if (ring->buf_info == NULL) {
- PMD_INIT_LOG(ERR, "ERROR: Creating tx_buf_info structure\n");
- return (-ENOMEM);
+void vmxnet3_rx_start(struct rte_eth_dev *dev)
+{
+ struct vmxnet3_dev *dp = dev->data->dev_private;
+ unsigned i;
+
+ for (i = 0; i < dev->data->nb_rx_queues; i++) {
+ struct vmxnet3_rx_queue *rxq = dp->rx_queue + i;
+ vmxnet3_write_reg(rxq->rx_prod, rxq->rx_ring.next);
}
+}
- /* Update the data portion with txq */
- dev->data->tx_queues[queue_idx] = txq;
+/* Partial packets with error are dropped. */
+static void vmxnet3_rx_error(struct vmxnet3_rx_queue *rxq,
+ const Vmxnet3_RxCompDesc *rcd)
+{
+ ++rxq->drop_errors;
+ if (rcd->fcs)
+ ++rxq->fcs_errors;
- return 0;
+ if (rxq->start_seg) {
+ rte_pktmbuf_free(rxq->start_seg);
+ rxq->start_seg = NULL;
+ rxq->prev_seg = &rxq->start_seg;
+ }
}
-int
-vmxnet3_dev_rx_queue_setup(struct rte_eth_dev *dev,
- uint16_t queue_idx,
- uint16_t nb_desc,
- unsigned int socket_id,
- __attribute__((unused)) const struct rte_eth_rxconf *rx_conf,
- struct rte_mempool *mp)
+/* TODO optimize this with a table */
+static uint32_t vmxnet3_pkt_flags(const Vmxnet3_RxCompDesc *rcd)
{
- const struct rte_memzone *mz;
- struct vmxnet3_rx_queue *rxq;
- struct vmxnet3_hw *hw;
- struct vmxnet3_cmd_ring *ring0, *ring1, *ring;
- struct vmxnet3_comp_ring *comp_ring;
- int size;
- uint8_t i;
- char mem_name[32];
- uint16_t buf_size;
- struct rte_pktmbuf_pool_private *mbp_priv;
- PMD_INIT_FUNC_TRACE();
- hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+ uint16_t pkt_flags = 0;
- mbp_priv = (struct rte_pktmbuf_pool_private *)
- rte_mempool_get_priv(mp);
- buf_size = (uint16_t) (mbp_priv->mbuf_data_room_size -
- RTE_PKTMBUF_HEADROOM);
+ if (rcd->rssType != VMXNET3_RCD_RSS_TYPE_NONE)
+ pkt_flags |= PKT_RX_RSS_HASH;
+ if (rcd->ts)
+ pkt_flags |= PKT_RX_VLAN_PKT;
+ if (rcd->v4)
+ pkt_flags |= PKT_RX_IPV4_HDR;
+ if (rcd->v6)
+ pkt_flags |= PKT_RX_IPV6_HDR;
- if (dev->data->dev_conf.rxmode.max_rx_pkt_len > buf_size) {
- PMD_INIT_LOG(ERR, "buf_size = %u, max_pkt_len = %u, "
- "VMXNET3 don't support scatter packets yet\n",
- buf_size, dev->data->dev_conf.rxmode.max_rx_pkt_len);
- return (-EINVAL);
- }
+ /* if Checksum has been checked */
+ if (!rcd->cnc) {
+ /* is IP header good */
+ if (!rcd->ipc)
+ pkt_flags |= PKT_RX_IP_CKSUM_BAD;
- rxq = rte_zmalloc("ethdev_rx_queue", sizeof(struct vmxnet3_rx_queue), CACHE_LINE_SIZE);
- if (rxq == NULL) {
- PMD_INIT_LOG(ERR, "Can not allocate rx queue structure\n");
- return (-ENOMEM);
- }
-
- rxq->mp = mp;
- rxq->queue_id = queue_idx;
- rxq->port_id = dev->data->port_id;
- rxq->shared = &hw->rqd_start[queue_idx];
- rxq->hw = hw;
- rxq->qid1 = queue_idx;
- rxq->qid2 = queue_idx + hw->num_rx_queues;
- rxq->stopped = TRUE;
-
- ring0 = &rxq->cmd_ring[0];
- ring1 = &rxq->cmd_ring[1];
- comp_ring = &rxq->comp_ring;
-
- /* Rx vmxnet rings length should be between 256-4096 */
- if(nb_desc < VMXNET3_DEF_RX_RING_SIZE) {
- PMD_INIT_LOG(ERR, "VMXNET3 Rx Ring Size Min: 256\n");
- return -EINVAL;
- } else if(nb_desc > VMXNET3_RX_RING_MAX_SIZE) {
- PMD_INIT_LOG(ERR, "VMXNET3 Rx Ring Size Max: 4096\n");
- return -EINVAL;
- } else {
- ring0->size = nb_desc;
- ring0->size &= ~VMXNET3_RING_SIZE_MASK;
- ring1->size = ring0->size;
- }
-
- comp_ring->size = ring0->size + ring1->size;
-
- /* Rx vmxnet rings structure initialization */
- ring0->next2fill = 0;
- ring1->next2fill = 0;
- ring0->next2comp = 0;
- ring1->next2comp = 0;
- ring0->gen = VMXNET3_INIT_GEN;
- ring1->gen = VMXNET3_INIT_GEN;
- comp_ring->next2proc = 0;
- comp_ring->gen = VMXNET3_INIT_GEN;
-
- size = sizeof(struct Vmxnet3_RxDesc) * (ring0->size + ring1->size);
- size += sizeof(struct Vmxnet3_RxCompDesc) * comp_ring->size;
-
- mz = ring_dma_zone_reserve(dev, "rxdesc", queue_idx, size, socket_id);
- if (mz == NULL) {
- PMD_INIT_LOG(ERR, "ERROR: Creating queue descriptors zone\n");
- return (-ENOMEM);
- }
- memset(mz->addr, 0, mz->len);
-
- /* cmd_ring0 initialization */
- ring0->base = mz->addr;
- ring0->basePA = mz->phys_addr;
-
- /* cmd_ring1 initialization */
- ring1->base = ring0->base + ring0->size;
- ring1->basePA = ring0->basePA + sizeof(struct Vmxnet3_RxDesc) * ring0->size;
-
- /* comp_ring initialization */
- comp_ring->base = ring1->base + ring1->size;
- comp_ring->basePA = ring1->basePA + sizeof(struct Vmxnet3_RxDesc) *
- ring1->size;
-
- /* cmd_ring0-cmd_ring1 buf_info allocation */
- for(i = 0; i < VMXNET3_RX_CMDRING_SIZE; i++) {
-
- ring = &rxq->cmd_ring[i];
- ring->rid = i;
- rte_snprintf(mem_name, sizeof(mem_name), "rx_ring_%d_buf_info", i);
-
- ring->buf_info = rte_zmalloc(mem_name, ring->size * sizeof(vmxnet3_buf_info_t), CACHE_LINE_SIZE);
- if (ring->buf_info == NULL) {
- PMD_INIT_LOG(ERR, "ERROR: Creating rx_buf_info structure\n");
- return (-ENOMEM);
- }
+ /* is Transport header good? */
+ if ((rcd->v4 || rcd->v6) && !rcd->tuc)
+ pkt_flags |= PKT_RX_L4_CKSUM_BAD;
}
- /* Update the data portion with rxq */
- dev->data->rx_queues[queue_idx] = rxq;
-
- return 0;
+ return pkt_flags;
}
+static void vmxnet3_rx_ring_advance(struct vmxnet3_cmd_ring *rx_ring,
+ uint16_t idx)
+{
+ while (rx_ring->next != idx) {
+ Vmxnet3_GenericDesc *rxd = rx_ring->base + rx_ring->next;
+ rxd->rxd.gen = rx_ring->gen;
+ vmxnet3_ring_advance(rx_ring);
+ }
+}
/*
- * Initializes Receive Unit
- * Load mbufs in rx queue in advance
+ * Receive burst of packets
*/
-int
-vmxnet3_dev_rxtx_init(struct rte_eth_dev *dev)
+uint16_t vmxnet3_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
+ uint16_t nb_pkts)
{
- struct vmxnet3_hw *hw;
- int i, ret;
- uint8_t j;
+ struct vmxnet3_rx_queue *rxq = rx_queue;
+ struct rte_eth_dev *dev = &rte_eth_devices[rxq->port_id];
+ uint16_t idx = 0, num_rxd = 0;
+ Vmxnet3_GenericDesc *rxd;
+
+ while (num_rxd < nb_pkts) {
+ struct rte_mbuf *mb, *nmb;
+ const Vmxnet3_RxCompDesc *rcd;
- PMD_INIT_FUNC_TRACE();
- hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+ rcd = (Vmxnet3_RxCompDesc *)
+ (rxq->comp_ring.base + rxq->comp_ring.next);
- for (i = 0; i < hw->num_rx_queues; i++) {
+ /* If buffer is not filled yet */
+ if (rcd->gen != rxq->comp_ring.gen)
+ break;
- vmxnet3_rx_queue_t *rxq = dev->data->rx_queues[i];
- for(j = 0;j < VMXNET3_RX_CMDRING_SIZE;j++) {
- /* Passing 0 as alloc_num will allocate full ring */
- ret = vmxnet3_post_rx_bufs(rxq, j);
- if (ret <= 0) {
- PMD_INIT_LOG(ERR, "ERROR: Posting Rxq: %d buffers ring: %d\n", i, j);
- return (-ret);
- }
- /* Updating device with the index:next2fill to fill the mbufs for coming packets */
- if (unlikely(rxq->shared->ctrl.updateRxProd)) {
- VMXNET3_WRITE_BAR0_REG(hw, rxprod_reg[j] + (rxq->queue_id * VMXNET3_REG_ALIGN),
- rxq->cmd_ring[j].next2fill);
- }
- }
- rxq->stopped = FALSE;
- }
+ mem_barrier();
- for (i = 0; i < dev->data->nb_tx_queues; i++) {
- struct vmxnet3_tx_queue *txq = dev->data->tx_queues[i];
- txq->stopped = FALSE;
- }
+ idx = rcd->rxdIdx;
+ PMD_RX_LOG(DEBUG,
+ "port=%u queue=%u id=%u len=%u%s%s%s",
+ rxq->port_id, rxq->queue_id, idx, rcd->len,
+ rcd->sop ? " sop" : "",
+ rcd->eop ? " eop" : "",
+ rcd->err ? " err" : "");
+
+ /* error detected? */
+ if (unlikely(rcd->eop && rcd->err)) {
+ PMD_DRV_LOG(ERR, "receive error");
+ vmxnet3_rx_error(rxq, rcd);
+ goto next_completion;
+ }
- return 0;
-}
+ /* Bogus zero length packet?? */
+ if (unlikely(rcd->len == 0)) {
+ PMD_DRV_LOG(ERR, "zero length packet");
+ goto next_completion;
+ }
-static uint8_t rss_intel_key[40] = {
- 0x6D, 0x5A, 0x56, 0xDA, 0x25, 0x5B, 0x0E, 0xC2,
- 0x41, 0x67, 0x25, 0x3D, 0x43, 0xA3, 0x8F, 0xB0,
- 0xD0, 0xCA, 0x2B, 0xCB, 0xAE, 0x7B, 0x30, 0xB4,
- 0x77, 0xCB, 0x2D, 0xA3, 0x80, 0x30, 0xF2, 0x0C,
- 0x6A, 0x42, 0xB7, 0x3B, 0xBE, 0xAC, 0x01, 0xFA,
-};
+ /* Check host data */
+ if (unlikely(rcd->rqID != rxq->queue_id))
+ rte_panic("%s(): completion qid %u != %u\n",
+ __func__, rcd->rqID, rxq->queue_id);
+
+ if (unlikely(idx > rxq->rx_ring.size))
+ rte_panic("%s(): id %u out of range %u\n",
+ __func__, idx, rxq->rx_ring.size);
+
+ nmb = rte_rxmbuf_alloc(rxq->mb_pool);
+ if (unlikely(nmb == NULL)) {
+ PMD_DRV_LOG(ERR,
+ "RX alloc failed port=%u queue=%u",
+ rxq->port_id, rxq->queue_id);
+ dev->data->rx_mbuf_alloc_failed++;
+ goto next_completion;
+ }
-/*
- * Configure RSS feature
- */
-int
-vmxnet3_rss_configure(struct rte_eth_dev *dev)
-{
- struct vmxnet3_hw *hw;
- struct VMXNET3_RSSConf *dev_rss_conf;
- struct rte_eth_rss_conf *port_rss_conf;
- uint8_t i, j;
+ /* Refill receive descriptor
+ * Length of new mbuf is same as old mbuf
+ */
+ rxd = rxq->rx_ring.base + idx;
+ uint64_t pa = RTE_MBUF_DATA_DMA_ADDR_DEFAULT(nmb);
+ rxd->rxd.addr = rte_cpu_to_le_64(pa);
+
+ /* Update mbuf */
+ mb = rxq->sw_ring[idx];
+ if (unlikely(mb == NULL))
+ rte_panic("%s(): no mbuf in ring[%u]\n",
+ __func__, idx);
+
+ rxq->sw_ring[idx] = nmb;
+
+ /* assume pkt.data is correct from pktmbuf_reset */
+ mb->pkt.data_len = rcd->len;
+ *rxq->prev_seg = mb;
+ rxq->prev_seg = &mb->pkt.next;
+
+ /* First part of packet of possibly fragmented */
+ if (rcd->sop) {
+ rxq->start_seg = mb;
+
+ rte_packet_prefetch(mb->pkt.data);
+ mb->pkt.pkt_len = rcd->len;
+ mb->pkt.nb_segs = 1;
+ mb->pkt.next = NULL;
+
+ mb->pkt.in_port = rxq->port_id;
+
+ /* Set offload values, only valid if flag is set. */
+ mb->pkt.vlan_macip.f.vlan_tci = rcd->tci;
+ mb->pkt.hash.rss = rcd->rssHash;
+ mb->ol_flags = vmxnet3_pkt_flags(rcd);
+ } else {
+ rxq->start_seg->pkt.pkt_len += rcd->len;
+ rxq->start_seg->pkt.nb_segs++;
+ }
- PMD_INIT_FUNC_TRACE();
- hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
- dev_rss_conf = hw->rss_conf;
- port_rss_conf = &dev->data->dev_conf.rx_adv_conf.rss_conf;
-
- /* loading hashFunc */
- dev_rss_conf->hashFunc = VMXNET3_RSS_HASH_FUNC_TOEPLITZ;
- /* loading hashKeySize */
- dev_rss_conf->hashKeySize = VMXNET3_RSS_MAX_KEY_SIZE;
- /* loading indTableSize : Must not exceed VMXNET3_RSS_MAX_IND_TABLE_SIZE (128)*/
- dev_rss_conf->indTableSize = (uint16_t)(hw->num_rx_queues * 4);
-
- if (port_rss_conf->rss_key == NULL) {
- /* Default hash key */
- port_rss_conf->rss_key = rss_intel_key;
- }
-
- /* loading hashKey */
- memcpy(&dev_rss_conf->hashKey[0], port_rss_conf->rss_key, dev_rss_conf->hashKeySize);
-
- /* loading indTable */
- for (i = 0, j = 0; i < dev_rss_conf->indTableSize; i++, j++) {
- if (j == dev->data->nb_rx_queues)
- j = 0;
- dev_rss_conf->indTable[i] = j;
- }
-
- /* loading hashType */
- dev_rss_conf->hashType = 0;
- if (port_rss_conf->rss_hf & ETH_RSS_IPV4)
- dev_rss_conf->hashType |= VMXNET3_RSS_HASH_TYPE_IPV4;
- if (port_rss_conf->rss_hf & ETH_RSS_IPV4_TCP)
- dev_rss_conf->hashType |= VMXNET3_RSS_HASH_TYPE_TCP_IPV4;
- if (port_rss_conf->rss_hf & ETH_RSS_IPV6)
- dev_rss_conf->hashType |= VMXNET3_RSS_HASH_TYPE_IPV6;
- if (port_rss_conf->rss_hf & ETH_RSS_IPV6_TCP)
- dev_rss_conf->hashType |= VMXNET3_RSS_HASH_TYPE_TCP_IPV6;
+ if (rcd->eop) {
+ rx_pkts[num_rxd++] = rxq->start_seg;
- return VMXNET3_SUCCESS;
-}
+ rxq->start_seg = NULL;
+ rxq->prev_seg = &rxq->start_seg;
+ }
-/*
- * Configure VLAN Filter feature
- */
-int
-vmxnet3_vlan_configure(struct rte_eth_dev *dev)
-{
- uint8_t i;
- struct vmxnet3_hw *hw = VMXNET3_DEV_PRIVATE_TO_HW(dev->data->dev_private);
- uint32_t *vf_table = hw->shared->devRead.rxFilterConf.vfTable;
+ next_completion:
+ /* force updates to rxd before advancing */
+ mem_barrier();
- PMD_INIT_FUNC_TRACE();
+ /* Advance receive ring (re-enable rx descriptor) */
+ vmxnet3_rx_ring_advance(&rxq->rx_ring, idx);
- /* Verify if this tag is already set */
- for (i = 0; i < VMXNET3_VFT_SIZE; i++) {
- /* Filter all vlan tags out by default */
- vf_table[i] = 0;
- /* To-Do: Provide another routine in dev_ops for user config */
+ if (unlikely(rxq->shared->updateRxProd))
+ vmxnet3_write_reg(rxq->rx_prod, rxq->rx_ring.next);
- PMD_INIT_LOG(DEBUG, "Registering VLAN portid: %"PRIu8" tag %u\n",
- dev->data->port_id, vf_table[i]);
+ /* advance to next completion */
+ vmxnet3_ring_advance(&rxq->comp_ring);
}
- return VMXNET3_SUCCESS;
+ return num_rxd;
}
--- /dev/null 1970-01-01 00:00:00.000000000 +0000
+++ b/lib/librte_pmd_vmxnet3/README 2014-05-14 11:46:47.129439301 -0700
@@ -0,0 +1,20 @@
+Brocade VMXNET3 driver
+=====================
+
+This directory contains the source code written based off of the
+driver for OpenSolaris.
+
+Some design notes about Vmxnet3 driver.
+
+1. Link state interrupt is not possible since irq is shared between
+ TX/RX and event. Might be possible to use MSI-X but then what if
+ MSI-X is not available.
+
+2. Transmit completion is done in lazy fashion by looking back at transmit
+ ring (no completion interrupt)
+
+3. Receiver support RSS, but Vmware does not support extensions
+ that Intel has for UDP.
+
+4. Receiver does not want LRO since LRO is broken for routing
+ and forwarding
--- /dev/null 1970-01-01 00:00:00.000000000 +0000
+++ b/lib/librte_pmd_vmxnet3/vmxnet3_defs.h 2014-05-14 11:46:47.129439301 -0700
@@ -0,0 +1,717 @@
+/*
+ * Copyright (C) 2007, 2011 VMware, Inc. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation version 2 and no later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
+ * or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ * for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+ *
+ */
+
+/*
+ * The contents of this file are subject to the terms of the Common
+ * Development and Distribution License (the "License") version 1.0
+ * and no later version. You may not use this file except in
+ * compliance with the License.
+ *
+ * You can obtain a copy of the License at
+ * http://www.opensource.org/licenses/cddl1.php
+ *
+ * See the License for the specific language governing permissions
+ * and limitations under the License.
+ *
+ */
+
+/*
+ * vmxnet3_defs.h --
+ *
+ * Definitions shared by device emulation and guest drivers for
+ * VMXNET3 NIC
+ */
+
+#ifndef _VMXNET3_DEFS_H_
+#define _VMXNET3_DEFS_H_
+
+/*
+ * Definitions for UPTv1
+ *
+ * Some of the defs are duplicated in vmkapi_net_upt.h, because
+ * vmkapi_net_upt.h cannot distribute with OSS yet and vmkapi headers can
+ * only include vmkapi headers. Make sure they are kept in sync!
+ */
+
+#define UPT1_MAX_TX_QUEUES 64
+#define UPT1_MAX_RX_QUEUES 64
+
+#define UPT1_MAX_INTRS (UPT1_MAX_TX_QUEUES + UPT1_MAX_RX_QUEUES)
+
+typedef struct UPT1_TxStats {
+ uint64 TSOPktsTxOK; /* TSO pkts post-segmentation */
+ uint64 TSOBytesTxOK;
+ uint64 ucastPktsTxOK;
+ uint64 ucastBytesTxOK;
+ uint64 mcastPktsTxOK;
+ uint64 mcastBytesTxOK;
+ uint64 bcastPktsTxOK;
+ uint64 bcastBytesTxOK;
+ uint64 pktsTxError;
+ uint64 pktsTxDiscard;
+} __attribute__((__packed__)) UPT1_TxStats;
+
+typedef struct UPT1_RxStats {
+ uint64 LROPktsRxOK; /* LRO pkts */
+ uint64 LROBytesRxOK; /* bytes from LRO pkts */
+ /* the following counters are for pkts from the wire, i.e., pre-LRO */
+ uint64 ucastPktsRxOK;
+ uint64 ucastBytesRxOK;
+ uint64 mcastPktsRxOK;
+ uint64 mcastBytesRxOK;
+ uint64 bcastPktsRxOK;
+ uint64 bcastBytesRxOK;
+ uint64 pktsRxOutOfBuf;
+ uint64 pktsRxError;
+} __attribute__((__packed__)) UPT1_RxStats;
+
+/* interrupt moderation level */
+#define UPT1_IML_NONE 0 /* no interrupt moderation */
+#define UPT1_IML_HIGHEST 7 /* least intr generated */
+#define UPT1_IML_ADAPTIVE 8 /* adpative intr moderation */
+
+/* values for UPT1_RSSConf.hashFunc */
+#define UPT1_RSS_HASH_TYPE_NONE 0x0
+#define UPT1_RSS_HASH_TYPE_IPV4 0x01
+#define UPT1_RSS_HASH_TYPE_TCP_IPV4 0x02
+#define UPT1_RSS_HASH_TYPE_IPV6 0x04
+#define UPT1_RSS_HASH_TYPE_TCP_IPV6 0x08
+
+#define UPT1_RSS_HASH_FUNC_NONE 0x0
+#define UPT1_RSS_HASH_FUNC_TOEPLITZ 0x01
+
+#define UPT1_RSS_MAX_KEY_SIZE 40
+#define UPT1_RSS_MAX_IND_TABLE_SIZE 128
+
+typedef struct UPT1_RSSConf {
+ uint16 hashType;
+ uint16 hashFunc;
+ uint16 hashKeySize;
+ uint16 indTableSize;
+ uint8 hashKey[UPT1_RSS_MAX_KEY_SIZE];
+ uint8 indTable[UPT1_RSS_MAX_IND_TABLE_SIZE];
+} __attribute__((__packed__)) UPT1_RSSConf;
+
+/* features */
+#define UPT1_F_RXCSUM 0x0001 /* rx csum verification */
+#define UPT1_F_RSS 0x0002
+#define UPT1_F_RXVLAN 0x0004 /* VLAN tag stripping */
+#define UPT1_F_LRO 0x0008
+
+/* all registers are 32 bit wide */
+/* BAR 1 */
+#define VMXNET3_REG_VRRS 0x0 /* Vmxnet3 Revision Report Selection */
+#define VMXNET3_REG_UVRS 0x8 /* UPT Version Report Selection */
+#define VMXNET3_REG_DSAL 0x10 /* Driver Shared Address Low */
+#define VMXNET3_REG_DSAH 0x18 /* Driver Shared Address High */
+#define VMXNET3_REG_CMD 0x20 /* Command */
+#define VMXNET3_REG_MACL 0x28 /* MAC Address Low */
+#define VMXNET3_REG_MACH 0x30 /* MAC Address High */
+#define VMXNET3_REG_ICR 0x38 /* Interrupt Cause Register */
+#define VMXNET3_REG_ECR 0x40 /* Event Cause Register */
+
+#define VMXNET3_REG_WSAL 0xF00 /* Wireless Shared Address Lo */
+#define VMXNET3_REG_WSAH 0xF08 /* Wireless Shared Address Hi */
+#define VMXNET3_REG_WCMD 0xF18 /* Wireless Command */
+
+/* BAR 0 */
+#define VMXNET3_REG_IMR 0x0 /* Interrupt Mask Register */
+#define VMXNET3_REG_TXPROD 0x600 /* Tx Producer Index */
+#define VMXNET3_REG_RXPROD 0x800 /* Rx Producer Index for ring 1 */
+#define VMXNET3_REG_RXPROD2 0xA00 /* Rx Producer Index for ring 2 */
+
+#define VMXNET3_PT_REG_SIZE 4096 /* BAR 0 */
+#define VMXNET3_VD_REG_SIZE 4096 /* BAR 1 */
+
+/*
+ * The two Vmxnet3 MMIO Register PCI BARs (BAR 0 at offset 10h and BAR 1 at
+ * offset 14h) as well as the MSI-X BAR are combined into one PhysMem region:
+ * <-VMXNET3_PT_REG_SIZE-><-VMXNET3_VD_REG_SIZE-><-VMXNET3_MSIX_BAR_SIZE-->
+ * -------------------------------------------------------------------------
+ * |Pass Thru Registers | Virtual Dev Registers | MSI-X Vector/PBA Table |
+ * -------------------------------------------------------------------------
+ * VMXNET3_MSIX_BAR_SIZE is defined in "vmxnet3Int.h"
+ */
+#define VMXNET3_PHYSMEM_PAGES 4
+
+#define VMXNET3_REG_ALIGN 8 /* All registers are 8-byte aligned. */
+#define VMXNET3_REG_ALIGN_MASK 0x7
+
+/* I/O Mapped access to registers */
+#define VMXNET3_IO_TYPE_PT 0
+#define VMXNET3_IO_TYPE_VD 1
+#define VMXNET3_IO_ADDR(type, reg) (((type) << 24) | ((reg) & 0xFFFFFF))
+#define VMXNET3_IO_TYPE(addr) ((addr) >> 24)
+#define VMXNET3_IO_REG(addr) ((addr) & 0xFFFFFF)
+
+#ifndef __le16
+#define __le16 uint16
+#endif
+#ifndef __le32
+#define __le32 uint32
+#endif
+#ifndef __le64
+#define __le64 uint64
+#endif
+
+#define VMXNET3_CMD_FIRST_SET 0xCAFE0000u
+#define VMXNET3_CMD_ACTIVATE_DEV VMXNET3_CMD_FIRST_SET
+#define VMXNET3_CMD_QUIESCE_DEV (VMXNET3_CMD_FIRST_SET + 1)
+#define VMXNET3_CMD_RESET_DEV (VMXNET3_CMD_FIRST_SET + 2)
+#define VMXNET3_CMD_UPDATE_RX_MODE (VMXNET3_CMD_FIRST_SET + 3)
+#define VMXNET3_CMD_UPDATE_MAC_FILTERS (VMXNET3_CMD_FIRST_SET + 4)
+#define VMXNET3_CMD_UPDATE_VLAN_FILTERS (VMXNET3_CMD_FIRST_SET + 5)
+#define VMXNET3_CMD_UPDATE_RSSIDT (VMXNET3_CMD_FIRST_SET + 6)
+#define VMXNET3_CMD_UPDATE_IML (VMXNET3_CMD_FIRST_SET + 7)
+#define VMXNET3_CMD_UPDATE_PMCFG (VMXNET3_CMD_FIRST_SET + 8)
+#define VMXNET3_CMD_UPDATE_FEATURE (VMXNET3_CMD_FIRST_SET + 9)
+#define VMXNET3_CMD_STOP_EMULATION (VMXNET3_CMD_FIRST_SET + 10)
+#define VMXNET3_CMD_LOAD_PLUGIN (VMXNET3_CMD_FIRST_SET + 11)
+#define VMXNET3_CMD_ACTIVATE_VF (VMXNET3_CMD_FIRST_SET + 12)
+
+#define VMXNET3_CMD_FIRST_GET 0xF00D0000u
+#define VMXNET3_CMD_GET_QUEUE_STATUS VMXNET3_CMD_FIRST_GET
+#define VMXNET3_CMD_GET_STATS (VMXNET3_CMD_FIRST_GET + 1)
+#define VMXNET3_CMD_GET_LINK (VMXNET3_CMD_FIRST_GET + 2)
+#define VMXNET3_CMD_GET_PERM_MAC_LO (VMXNET3_CMD_FIRST_GET + 3)
+#define VMXNET3_CMD_GET_PERM_MAC_HI (VMXNET3_CMD_FIRST_GET + 4)
+#define VMXNET3_CMD_GET_DID_LO (VMXNET3_CMD_FIRST_GET + 5)
+#define VMXNET3_CMD_GET_DID_HI (VMXNET3_CMD_FIRST_GET + 6)
+#define VMXNET3_CMD_GET_DEV_EXTRA_INFO (VMXNET3_CMD_FIRST_GET + 7)
+#define VMXNET3_CMD_GET_CONF_INTR (VMXNET3_CMD_FIRST_GET + 8)
+#define VMXNET3_CMD_GET_ADAPTIVE_RING_INFO (VMXNET3_CMD_FIRST_GET + 9)
+
+/* Adaptive Ring Info Flags */
+#define VMXNET3_DISABLE_ADAPTIVE_RING 1
+
+/*
+ * Little Endian layout of bitfields -
+ * Byte 0 : 7.....len.....0
+ * Byte 1 : rsvd gen 13.len.8
+ * Byte 2 : 5.msscof.0 ext1 dtype
+ * Byte 3 : 13...msscof...6
+ *
+ * Big Endian layout of bitfields -
+ * Byte 0: 13...msscof...6
+ * Byte 1 : 5.msscof.0 ext1 dtype
+ * Byte 2 : rsvd gen 13.len.8
+ * Byte 3 : 7.....len.....0
+ *
+ * Thus, le32_to_cpu on the dword will allow the big endian driver to read
+ * the bit fields correctly. And cpu_to_le32 will convert bitfields
+ * bit fields written by big endian driver to format required by device.
+ */
+
+typedef struct Vmxnet3_TxDesc {
+ __le64 addr;
+
+#ifdef __BIG_ENDIAN_BITFIELD
+ uint32 msscof:14; /* MSS, checksum offset, flags */
+ uint32 ext1:1;
+ uint32 dtype:1; /* descriptor type */
+ uint32 rsvd:1;
+ uint32 gen:1; /* generation bit */
+ uint32 len:14;
+#else
+ uint32 len:14;
+ uint32 gen:1; /* generation bit */
+ uint32 rsvd:1;
+ uint32 dtype:1; /* descriptor type */
+ uint32 ext1:1;
+ uint32 msscof:14; /* MSS, checksum offset, flags */
+#endif /* __BIG_ENDIAN_BITFIELD */
+
+#ifdef __BIG_ENDIAN_BITFIELD
+ uint32 tci:16; /* Tag to Insert */
+ uint32 ti:1; /* VLAN Tag Insertion */
+ uint32 ext2:1;
+ uint32 cq:1; /* completion request */
+ uint32 eop:1; /* End Of Packet */
+ uint32 om:2; /* offload mode */
+ uint32 hlen:10; /* header len */
+#else
+ uint32 hlen:10; /* header len */
+ uint32 om:2; /* offload mode */
+ uint32 eop:1; /* End Of Packet */
+ uint32 cq:1; /* completion request */
+ uint32 ext2:1;
+ uint32 ti:1; /* VLAN Tag Insertion */
+ uint32 tci:16; /* Tag to Insert */
+#endif /* __BIG_ENDIAN_BITFIELD */
+} __attribute__((__packed__)) Vmxnet3_TxDesc;
+
+/* TxDesc.OM values */
+#define VMXNET3_OM_NONE 0
+#define VMXNET3_OM_CSUM 2
+#define VMXNET3_OM_TSO 3
+
+/* fields in TxDesc we access w/o using bit fields */
+#define VMXNET3_TXD_EOP_SHIFT 12
+#define VMXNET3_TXD_CQ_SHIFT 13
+#define VMXNET3_TXD_GEN_SHIFT 14
+#define VMXNET3_TXD_EOP_DWORD_SHIFT 3
+#define VMXNET3_TXD_GEN_DWORD_SHIFT 2
+
+#define VMXNET3_TXD_CQ (1 << VMXNET3_TXD_CQ_SHIFT)
+#define VMXNET3_TXD_EOP (1 << VMXNET3_TXD_EOP_SHIFT)
+#define VMXNET3_TXD_GEN (1 << VMXNET3_TXD_GEN_SHIFT)
+
+#define VMXNET3_TXD_GEN_SIZE 1
+#define VMXNET3_TXD_EOP_SIZE 1
+
+#define VMXNET3_HDR_COPY_SIZE 128
+
+typedef struct Vmxnet3_TxDataDesc {
+ uint8 data[VMXNET3_HDR_COPY_SIZE];
+} __attribute__((__packed__)) Vmxnet3_TxDataDesc;
+
+#define VMXNET3_TCD_GEN_SHIFT 31
+#define VMXNET3_TCD_GEN_SIZE 1
+#define VMXNET3_TCD_TXIDX_SHIFT 0
+#define VMXNET3_TCD_TXIDX_SIZE 12
+#define VMXNET3_TCD_GEN_DWORD_SHIFT 3
+
+typedef struct Vmxnet3_TxCompDesc {
+ uint32 txdIdx:12; /* Index of the EOP TxDesc */
+ uint32 ext1:20;
+
+ __le32 ext2;
+ __le32 ext3;
+
+ uint32 rsvd:24;
+ uint32 type:7; /* completion type */
+ uint32 gen:1; /* generation bit */
+} __attribute__((__packed__)) Vmxnet3_TxCompDesc;
+
+typedef struct Vmxnet3_RxDesc {
+ __le64 addr;
+
+#ifdef __BIG_ENDIAN_BITFIELD
+ uint32 gen:1; /* Generation bit */
+ uint32 rsvd:15;
+ uint32 dtype:1; /* Descriptor type */
+ uint32 btype:1; /* Buffer Type */
+ uint32 len:14;
+#else
+ uint32 len:14;
+ uint32 btype:1; /* Buffer Type */
+ uint32 dtype:1; /* Descriptor type */
+ uint32 rsvd:15;
+ uint32 gen:1; /* Generation bit */
+#endif
+ __le32 ext1;
+} __attribute__((__packed__)) Vmxnet3_RxDesc;
+
+/* values of RXD.BTYPE */
+#define VMXNET3_RXD_BTYPE_HEAD 0 /* head only */
+#define VMXNET3_RXD_BTYPE_BODY 1 /* body only */
+
+/* fields in RxDesc we access w/o using bit fields */
+#define VMXNET3_RXD_BTYPE_SHIFT 14
+#define VMXNET3_RXD_GEN_SHIFT 31
+
+typedef struct Vmxnet3_RxCompDesc {
+#ifdef __BIG_ENDIAN_BITFIELD
+ uint32 ext2:1;
+ uint32 cnc:1; /* Checksum Not Calculated */
+ uint32 rssType:4; /* RSS hash type used */
+ uint32 rqID:10; /* rx queue/ring ID */
+ uint32 sop:1; /* Start of Packet */
+ uint32 eop:1; /* End of Packet */
+ uint32 ext1:2;
+ uint32 rxdIdx:12; /* Index of the RxDesc */
+#else
+ uint32 rxdIdx:12; /* Index of the RxDesc */
+ uint32 ext1:2;
+ uint32 eop:1; /* End of Packet */
+ uint32 sop:1; /* Start of Packet */
+ uint32 rqID:10; /* rx queue/ring ID */
+ uint32 rssType:4; /* RSS hash type used */
+ uint32 cnc:1; /* Checksum Not Calculated */
+ uint32 ext2:1;
+#endif /* __BIG_ENDIAN_BITFIELD */
+
+ __le32 rssHash; /* RSS hash value */
+
+#ifdef __BIG_ENDIAN_BITFIELD
+ uint32 tci:16; /* Tag stripped */
+ uint32 ts:1; /* Tag is stripped */
+ uint32 err:1; /* Error */
+ uint32 len:14; /* data length */
+#else
+ uint32 len:14; /* data length */
+ uint32 err:1; /* Error */
+ uint32 ts:1; /* Tag is stripped */
+ uint32 tci:16; /* Tag stripped */
+#endif /* __BIG_ENDIAN_BITFIELD */
+
+
+#ifdef __BIG_ENDIAN_BITFIELD
+ uint32 gen:1; /* generation bit */
+ uint32 type:7; /* completion type */
+ uint32 fcs:1; /* Frame CRC correct */
+ uint32 frg:1; /* IP Fragment */
+ uint32 v4:1; /* IPv4 */
+ uint32 v6:1; /* IPv6 */
+ uint32 ipc:1; /* IP Checksum Correct */
+ uint32 tcp:1; /* TCP packet */
+ uint32 udp:1; /* UDP packet */
+ uint32 tuc:1; /* TCP/UDP Checksum Correct */
+ uint32 csum:16;
+#else
+ uint32 csum:16;
+ uint32 tuc:1; /* TCP/UDP Checksum Correct */
+ uint32 udp:1; /* UDP packet */
+ uint32 tcp:1; /* TCP packet */
+ uint32 ipc:1; /* IP Checksum Correct */
+ uint32 v6:1; /* IPv6 */
+ uint32 v4:1; /* IPv4 */
+ uint32 frg:1; /* IP Fragment */
+ uint32 fcs:1; /* Frame CRC correct */
+ uint32 type:7; /* completion type */
+ uint32 gen:1; /* generation bit */
+#endif /* __BIG_ENDIAN_BITFIELD */
+} __attribute__((__packed__)) Vmxnet3_RxCompDesc;
+
+/* fields in RxCompDesc we access via Vmxnet3_GenericDesc.dword[3] */
+#define VMXNET3_RCD_TUC_SHIFT 16
+#define VMXNET3_RCD_IPC_SHIFT 19
+
+/* fields in RxCompDesc we access via Vmxnet3_GenericDesc.qword[1] */
+#define VMXNET3_RCD_TYPE_SHIFT 56
+#define VMXNET3_RCD_GEN_SHIFT 63
+
+/* csum OK for TCP/UDP pkts over IP */
+#define VMXNET3_RCD_CSUM_OK (1 << VMXNET3_RCD_TUC_SHIFT | \
+ 1 << VMXNET3_RCD_IPC_SHIFT)
+
+/* value of RxCompDesc.rssType */
+#define VMXNET3_RCD_RSS_TYPE_NONE 0
+#define VMXNET3_RCD_RSS_TYPE_IPV4 1
+#define VMXNET3_RCD_RSS_TYPE_TCPIPV4 2
+#define VMXNET3_RCD_RSS_TYPE_IPV6 3
+#define VMXNET3_RCD_RSS_TYPE_TCPIPV6 4
+
+/* a union for accessing all cmd/completion descriptors */
+typedef union Vmxnet3_GenericDesc {
+ __le64 qword[2];
+ __le32 dword[4];
+ __le16 word[8];
+ Vmxnet3_TxDesc txd;
+ Vmxnet3_RxDesc rxd;
+ Vmxnet3_TxCompDesc tcd;
+ Vmxnet3_RxCompDesc rcd;
+} Vmxnet3_GenericDesc;
+
+#define VMXNET3_INIT_GEN 1
+
+/* Max size of a single tx buffer */
+#define VMXNET3_MAX_TX_BUF_SIZE (1 << 14)
+
+/* # of tx desc needed for a tx buffer size */
+#define VMXNET3_TXD_NEEDED(size) \
+ (((size) + VMXNET3_MAX_TX_BUF_SIZE - 1) / VMXNET3_MAX_TX_BUF_SIZE)
+
+/* max # of tx descs for a non-tso pkt */
+#define VMXNET3_MAX_TXD_PER_PKT 16
+
+/* Max size of a single rx buffer */
+#define VMXNET3_MAX_RX_BUF_SIZE ((1 << 14) - 1)
+/* Minimum size of a type 0 buffer */
+#define VMXNET3_MIN_T0_BUF_SIZE 128
+#define VMXNET3_MAX_CSUM_OFFSET 1024
+
+/* Ring base address alignment */
+#define VMXNET3_RING_BA_ALIGN 512
+#define VMXNET3_RING_BA_MASK (VMXNET3_RING_BA_ALIGN - 1)
+
+/* Ring size must be a multiple of 32 */
+#define VMXNET3_RING_SIZE_ALIGN 32
+#define VMXNET3_RING_SIZE_MASK (VMXNET3_RING_SIZE_ALIGN - 1)
+
+/* Max ring size */
+#define VMXNET3_TX_RING_MAX_SIZE 4096
+#define VMXNET3_TC_RING_MAX_SIZE 4096
+#define VMXNET3_RX_RING_MAX_SIZE 4096
+#define VMXNET3_RC_RING_MAX_SIZE 8192
+
+/* a list of reasons for queue stop */
+
+#define VMXNET3_ERR_NOEOP 0x80000000 /* cannot find the EOP desc of a pkt */
+#define VMXNET3_ERR_TXD_REUSE 0x80000001 /* reuse TxDesc before tx completion */
+#define VMXNET3_ERR_BIG_PKT 0x80000002 /* too many TxDesc for a pkt */
+#define VMXNET3_ERR_DESC_NOT_SPT 0x80000003 /* descriptor type not supported */
+#define VMXNET3_ERR_SMALL_BUF 0x80000004 /* type 0 buffer too small */
+#define VMXNET3_ERR_STRESS 0x80000005 /* stress option firing in vmkernel */
+#define VMXNET3_ERR_SWITCH 0x80000006 /* mode switch failure */
+#define VMXNET3_ERR_TXD_INVALID 0x80000007 /* invalid TxDesc */
+
+/* completion descriptor types */
+#define VMXNET3_CDTYPE_TXCOMP 0 /* Tx Completion Descriptor */
+#define VMXNET3_CDTYPE_RXCOMP 3 /* Rx Completion Descriptor */
+
+#define VMXNET3_GOS_BITS_UNK 0 /* unknown */
+#define VMXNET3_GOS_BITS_32 1
+#define VMXNET3_GOS_BITS_64 2
+
+#define VMXNET3_GOS_TYPE_UNK 0 /* unknown */
+#define VMXNET3_GOS_TYPE_LINUX 1
+#define VMXNET3_GOS_TYPE_WIN 2
+#define VMXNET3_GOS_TYPE_SOLARIS 3
+#define VMXNET3_GOS_TYPE_FREEBSD 4
+#define VMXNET3_GOS_TYPE_PXE 5
+
+/* All structures in DriverShared are padded to multiples of 8 bytes */
+
+typedef struct Vmxnet3_GOSInfo {
+#ifdef __BIG_ENDIAN_BITFIELD
+ uint32 gosMisc: 10; /* other info about gos */
+ uint32 gosVer: 16; /* gos version */
+ uint32 gosType: 4; /* which guest */
+ uint32 gosBits: 2; /* 32-bit or 64-bit? */
+#else
+ uint32 gosBits: 2; /* 32-bit or 64-bit? */
+ uint32 gosType: 4; /* which guest */
+ uint32 gosVer: 16; /* gos version */
+ uint32 gosMisc: 10; /* other info about gos */
+#endif /* __BIG_ENDIAN_BITFIELD */
+} __attribute__((__packed__)) Vmxnet3_GOSInfo;
+
+typedef struct Vmxnet3_DriverInfo {
+ __le32 version; /* driver version */
+ Vmxnet3_GOSInfo gos;
+ __le32 vmxnet3RevSpt; /* vmxnet3 revision supported */
+ __le32 uptVerSpt; /* upt version supported */
+} __attribute__((__packed__)) Vmxnet3_DriverInfo;
+
+#define VMXNET3_REV1_MAGIC 0xbabefee1
+
+/*
+ * QueueDescPA must be 128 bytes aligned. It points to an array of
+ * Vmxnet3_TxQueueDesc followed by an array of Vmxnet3_RxQueueDesc.
+ * The number of Vmxnet3_TxQueueDesc/Vmxnet3_RxQueueDesc are specified by
+ * Vmxnet3_MiscConf.numTxQueues/numRxQueues, respectively.
+ */
+#define VMXNET3_QUEUE_DESC_ALIGN 128
+
+typedef struct Vmxnet3_MiscConf {
+ Vmxnet3_DriverInfo driverInfo;
+ __le64 uptFeatures;
+ __le64 ddPA; /* driver data PA */
+ __le64 queueDescPA; /* queue descriptor table PA */
+ __le32 ddLen; /* driver data len */
+ __le32 queueDescLen; /* queue descriptor table len, in bytes */
+ __le32 mtu;
+ __le16 maxNumRxSG;
+ uint8 numTxQueues;
+ uint8 numRxQueues;
+ __le32 reserved[4];
+} __attribute__((__packed__)) Vmxnet3_MiscConf;
+
+typedef struct Vmxnet3_TxQueueConf {
+ __le64 txRingBasePA;
+ __le64 dataRingBasePA;
+ __le64 compRingBasePA;
+ __le64 ddPA; /* driver data */
+ __le64 reserved;
+ __le32 txRingSize; /* # of tx desc */
+ __le32 dataRingSize; /* # of data desc */
+ __le32 compRingSize; /* # of comp desc */
+ __le32 ddLen; /* size of driver data */
+ uint8 intrIdx;
+ uint8 _pad[7];
+} __attribute__((__packed__)) Vmxnet3_TxQueueConf;
+
+typedef struct Vmxnet3_RxQueueConf {
+ __le64 rxRingBasePA[2];
+ __le64 compRingBasePA;
+ __le64 ddPA; /* driver data */
+ __le64 reserved;
+ __le32 rxRingSize[2]; /* # of rx desc */
+ __le32 compRingSize; /* # of rx comp desc */
+ __le32 ddLen; /* size of driver data */
+ uint8 intrIdx;
+ uint8 _pad[7];
+} __attribute__((__packed__)) Vmxnet3_RxQueueConf;
+
+enum vmxnet3_intr_mask_mode {
+ VMXNET3_IMM_AUTO = 0,
+ VMXNET3_IMM_ACTIVE = 1,
+ VMXNET3_IMM_LAZY = 2
+};
+
+enum vmxnet3_intr_type {
+ VMXNET3_IT_AUTO = 0,
+ VMXNET3_IT_INTX = 1,
+ VMXNET3_IT_MSI = 2,
+ VMXNET3_IT_MSIX = 3
+};
+
+#define VMXNET3_MAX_TX_QUEUES 8
+#define VMXNET3_MAX_RX_QUEUES 16
+/* addition 1 for events */
+#define VMXNET3_MAX_INTRS 25
+
+/* value of intrCtrl */
+#define VMXNET3_IC_DISABLE_ALL 0x1 /* bit 0 */
+
+typedef struct Vmxnet3_IntrConf {
+ Bool autoMask;
+ uint8 numIntrs; /* # of interrupts */
+ uint8 eventIntrIdx;
+ uint8 modLevels[VMXNET3_MAX_INTRS]; /* moderation level for each intr */
+ __le32 intrCtrl;
+ __le32 reserved[2];
+} __attribute__((__packed__)) Vmxnet3_IntrConf;
+
+/* one bit per VLAN ID, the size is in the units of uint32 */
+#define VMXNET3_VFT_SIZE (4096 / (sizeof (uint32) * 8))
+
+typedef struct Vmxnet3_QueueStatus {
+ Bool stopped;
+ uint8 _pad[3];
+ __le32 error;
+} __attribute__((__packed__)) Vmxnet3_QueueStatus;
+
+typedef struct Vmxnet3_TxQueueCtrl {
+ __le32 txNumDeferred;
+ __le32 txThreshold;
+ __le64 reserved;
+} __attribute__((__packed__)) Vmxnet3_TxQueueCtrl;
+
+typedef struct Vmxnet3_RxQueueCtrl {
+ Bool updateRxProd;
+ uint8 _pad[7];
+ __le64 reserved;
+} __attribute__((__packed__)) Vmxnet3_RxQueueCtrl;
+
+#define VMXNET3_RXM_UCAST 0x01 /* unicast only */
+#define VMXNET3_RXM_MCAST 0x02 /* multicast passing the filters */
+#define VMXNET3_RXM_BCAST 0x04 /* broadcast only */
+#define VMXNET3_RXM_ALL_MULTI 0x08 /* all multicast */
+#define VMXNET3_RXM_PROMISC 0x10 /* promiscuous */
+
+typedef struct Vmxnet3_RxFilterConf {
+ __le32 rxMode; /* VMXNET3_RXM_xxx */
+ __le16 mfTableLen; /* size of the multicast filter table */
+ __le16 _pad1;
+ __le64 mfTablePA; /* PA of the multicast filters table */
+ __le32 vfTable[VMXNET3_VFT_SIZE]; /* vlan filter */
+} __attribute__((__packed__)) Vmxnet3_RxFilterConf;
+
+#define VMXNET3_PM_MAX_FILTERS 6
+#define VMXNET3_PM_MAX_PATTERN_SIZE 128
+#define VMXNET3_PM_MAX_MASK_SIZE (VMXNET3_PM_MAX_PATTERN_SIZE / 8)
+
+#define VMXNET3_PM_WAKEUP_MAGIC 0x01 /* wake up on magic pkts */
+#define VMXNET3_PM_WAKEUP_FILTER 0x02 /* wake up on pkts matching filters */
+
+typedef struct Vmxnet3_PM_PktFilter {
+ uint8 maskSize;
+ uint8 patternSize;
+ uint8 mask[VMXNET3_PM_MAX_MASK_SIZE];
+ uint8 pattern[VMXNET3_PM_MAX_PATTERN_SIZE];
+ uint8 pad[6];
+} __attribute__((__packed__)) Vmxnet3_PM_PktFilter;
+
+typedef struct Vmxnet3_PMConf {
+ __le16 wakeUpEvents; /* VMXNET3_PM_WAKEUP_xxx */
+ uint8 numFilters;
+ uint8 pad[5];
+ Vmxnet3_PM_PktFilter filters[VMXNET3_PM_MAX_FILTERS];
+} __attribute__((__packed__)) Vmxnet3_PMConf;
+
+typedef struct Vmxnet3_VariableLenConfDesc {
+ __le32 confVer;
+ __le32 confLen;
+ __le64 confPA;
+} __attribute__((__packed__)) Vmxnet3_VariableLenConfDesc;
+
+typedef struct Vmxnet3_DSDevRead {
+ /* read-only region for device, read by dev in response to a SET cmd */
+ Vmxnet3_MiscConf misc;
+ Vmxnet3_IntrConf intrConf;
+ Vmxnet3_RxFilterConf rxFilterConf;
+ Vmxnet3_VariableLenConfDesc rssConfDesc;
+ Vmxnet3_VariableLenConfDesc pmConfDesc;
+ Vmxnet3_VariableLenConfDesc pluginConfDesc;
+} __attribute__((__packed__)) Vmxnet3_DSDevRead;
+
+typedef struct Vmxnet3_TxQueueDesc {
+ Vmxnet3_TxQueueCtrl ctrl;
+ Vmxnet3_TxQueueConf conf;
+ /* Driver read after a GET command */
+ Vmxnet3_QueueStatus status;
+ UPT1_TxStats stats;
+ uint8 _pad[88]; /* 128 aligned */
+} __attribute__((__packed__)) Vmxnet3_TxQueueDesc;
+
+typedef struct Vmxnet3_RxQueueDesc {
+ Vmxnet3_RxQueueCtrl ctrl;
+ Vmxnet3_RxQueueConf conf;
+ /* Driver read after a GET command */
+ Vmxnet3_QueueStatus status;
+ UPT1_RxStats stats;
+ uint8 _pad[88]; /* 128 aligned */
+} __attribute__((__packed__)) Vmxnet3_RxQueueDesc;
+
+typedef struct Vmxnet3_DriverShared {
+ __le32 magic;
+ __le32 pad; /* make devRead start at 64-bit boundaries */
+ Vmxnet3_DSDevRead devRead;
+ __le32 ecr;
+ __le32 reserved[5];
+} __attribute__((__packed__)) Vmxnet3_DriverShared;
+
+#define VMXNET3_ECR_RQERR (1 << 0)
+#define VMXNET3_ECR_TQERR (1 << 1)
+#define VMXNET3_ECR_LINK (1 << 2)
+#define VMXNET3_ECR_DIC (1 << 3)
+#define VMXNET3_ECR_DEBUG (1 << 4)
+
+/* flip the gen bit of a ring */
+#define VMXNET3_FLIP_RING_GEN(gen) ((gen) = (gen) ^ 0x1)
+
+/* only use this if moving the idx won't affect the gen bit */
+#define VMXNET3_INC_RING_IDX_ONLY(idx, ring_size) \
+do {\
+ (idx)++; \
+ if (UNLIKELY((idx) == (ring_size))) { \
+ (idx) = 0; \
+ } \
+} while (0)
+
+#define VMXNET3_SET_VFTABLE_ENTRY(vfTable, vid) \
+ vfTable[vid >> 5] |= (1 << (vid & 31))
+#define VMXNET3_CLEAR_VFTABLE_ENTRY(vfTable, vid) \
+ vfTable[vid >> 5] &= ~(1 << (vid & 31))
+
+#define VMXNET3_VFTABLE_ENTRY_IS_SET(vfTable, vid) \
+ ((vfTable[vid >> 5] & (1 << (vid & 31))) != 0)
+
+#define VMXNET3_MAX_MTU 9000
+#define VMXNET3_MIN_MTU 60
+
+#define VMXNET3_LINK_UP (10000 << 16 | 1) // 10 Gbps, up
+#define VMXNET3_LINK_DOWN 0
+
+#define VMXWIFI_DRIVER_SHARED_LEN 8192
+
+#define VMXNET3_DID_PASSTHRU 0xFFFF
+
+#endif /* _VMXNET3_DEFS_H_ */
--- /dev/null 1970-01-01 00:00:00.000000000 +0000
+++ b/lib/librte_pmd_vmxnet3/vmxnet3_dev.h 2014-05-14 11:46:47.129439301 -0700
@@ -0,0 +1,148 @@
+/*
+ * Copyright (c) 2012-2014 Brocade Communications Systems, Inc.
+ * All rights reserved.
+ */
+
+/* typedef's to make CDDL header happy */
+typedef uint8_t Bool;
+typedef uint8_t uint8;
+typedef uint16_t uint16;
+typedef uint32_t uint32;
+typedef uint64_t uint64;
+
+/* header file from OpenSolaris */
+#include "vmxnet3_defs.h"
+#include "vmxnet3_logs.h"
+
+#define VMXNET3_MAX_ETH_HDR 22
+
+/*
+ * Device specific data and access functions
+ */
+struct vmxnet3_cmd_ring {
+ union Vmxnet3_GenericDesc *base;
+ uint32_t size;
+ uint32_t next;
+ uint8_t gen;
+ uint64_t phys_addr;
+};
+
+static inline uint32_t
+vmxnet3_ring_advance(struct vmxnet3_cmd_ring *ring)
+{
+ if (++ring->next == ring->size) {
+ ring->next = 0;
+ VMXNET3_FLIP_RING_GEN(ring->gen);
+ }
+ return ring->next;
+}
+
+/**
+ * Structure associated with each TX queue.
+ */
+struct vmxnet3_tx_queue {
+ struct vmxnet3_cmd_ring tx_ring; /* Transmit pending */
+ struct vmxnet3_cmd_ring comp_ring; /* Transmit complete */
+ Vmxnet3_TxQueueCtrl *shared; /* Queue thresholds */
+ void *tx_prod; /* Tx producer register */
+ struct rte_mbuf **sw_ring; /* Associated buffers */
+ uint16_t nb_tx_desc; /* Max Tx descriptors */
+ uint16_t tx_free_thresh; /* Transmit cleanup level */
+ uint16_t next_to_clean; /* Next ring entry to clean */
+ uint16_t queue_id;
+ uint8_t port_id;
+} __rte_cache_aligned;
+
+/**
+ * Structure associated with each RX queue.
+ */
+struct vmxnet3_rx_queue {
+ struct vmxnet3_cmd_ring rx_ring; /* Receive pending */
+ struct vmxnet3_cmd_ring comp_ring; /* Receive completed */
+ struct rte_mempool *mb_pool; /* Receive buffer pool */
+ struct rte_mbuf **sw_ring; /* Associated buffers */
+ struct rte_mbuf *start_seg; /* Start of current Rx packet */
+ struct rte_mbuf **prev_seg; /* P2p to last segment */
+ Vmxnet3_RxQueueCtrl *shared; /* Queue thesholds */
+ void *rx_prod; /* Rx producer register */
+
+ uint64_t drop_errors; /* Rx packet errors */
+ uint64_t fcs_errors;
+
+ uint16_t nb_rx_desc; /* Max Rx descriptors */
+ uint16_t queue_id;
+ uint8_t port_id;
+} __rte_cache_aligned;
+
+/**
+ * Structure associated with each interface
+ */
+struct vmxnet3_dev {
+ uint8_t *bar0, *bar1;
+ uint32_t rx_mode;
+ struct ether_addr mac_addr;
+ rte_spinlock_t cmd_lock;
+ const struct rte_memzone *shared;
+ const struct rte_memzone *qdesc;
+ Vmxnet3_TxQueueDesc *tqd_start;
+ Vmxnet3_RxQueueDesc *rqd_start;
+ uint32_t shadow_vfta[VMXNET3_VFT_SIZE];
+
+ struct vmxnet3_tx_queue tx_queue[VMXNET3_MAX_TX_QUEUES];
+ struct vmxnet3_rx_queue rx_queue[VMXNET3_MAX_RX_QUEUES];
+};
+
+static inline uint32_t vmxnet3_read_reg(void *addr)
+{
+ return *((volatile uint32_t *)addr);
+}
+
+static inline void vmxnet3_write_reg(void *addr, uint32_t val)
+{
+ *((volatile uint32_t *)addr) = val;
+}
+
+static inline uint32_t vmxnet3_bar0_get32(struct vmxnet3_dev *dp, unsigned reg)
+{
+ return vmxnet3_read_reg(dp->bar0 + reg);
+}
+
+static inline void vmxnet3_bar0_put32(struct vmxnet3_dev *dp,
+ unsigned reg, uint32_t value)
+{
+ vmxnet3_write_reg(dp->bar0 + reg, value);
+}
+
+static inline uint32_t vmxnet3_bar1_get32(struct vmxnet3_dev *dp, unsigned reg)
+{
+ return vmxnet3_read_reg(dp->bar1 + reg);
+}
+
+static inline void vmxnet3_bar1_put32(struct vmxnet3_dev *dp,
+ unsigned reg, uint32_t value)
+{
+ vmxnet3_write_reg(dp->bar1 + reg, value);
+}
+
+extern int
+vmxnet3_tx_queue_setup(struct rte_eth_dev *dev,
+ uint16_t queue_idx,
+ uint16_t nb_desc,
+ unsigned int socket_id,
+ const struct rte_eth_txconf *tx_conf);
+extern void vmxnet3_tx_queue_release(void *rxq);
+extern void vmxnet3_tx_flush_all(struct rte_eth_dev *dev);
+extern uint16_t vmxnet3_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
+ uint16_t nb_pkts);
+
+extern int
+vmxnet3_rx_queue_setup(struct rte_eth_dev *dev, uint16_t queue_idx,
+ uint16_t nb_desc, unsigned int socket_id,
+ const struct rte_eth_rxconf *rx_conf,
+ struct rte_mempool *mp);
+extern void vmxnet3_rx_queue_release(void *rxq);
+extern int vmxnet3_rx_init(struct rte_eth_dev *dev);
+extern void vmxnet3_rx_start(struct rte_eth_dev *dev);
+extern void vmxnet3_rx_flush_all(struct rte_eth_dev *dev);
+extern uint16_t vmxnet3_recv_pkts(void *rxq, struct rte_mbuf **rx_pkts,
+ uint16_t nb_pkts);
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [dpdk-dev] [PATCH 2/6] Subjet: ethdev: add common code to atomicly access link
2014-05-14 18:55 ` [dpdk-dev] [PATCH 2/6] Subjet: ethdev: add common code to atomicly access link Stephen Hemminger
@ 2014-05-15 8:43 ` Ivan Boule
0 siblings, 0 replies; 11+ messages in thread
From: Ivan Boule @ 2014-05-15 8:43 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: dev, Stephen Hemminger
Hi Stephen,
The new function rte_eth_dev_set_link() below is declared as void.
Thus, the comments about its return value are wrong, and should be removed,
unless they still refer to a desired behaviour, in which case it is the
function itself that should be reworked to invoke rte_atomic64_cmpset and
to return its result.
Regards,
Ivan
2014-05-14 20:55 GMT+02:00 Stephen Hemminger <stephen@networkplumber.org>:
> Many drivers copy/paste same code to atomicly access link information.
> Add functions to do this instead.
>
> Signed-off-by: Stephen Hemminger <shemming@brocade.com>
>
> --- a/lib/librte_ether/rte_ethdev.h 2014-05-14 11:27:07.409461720 -0700
> +++ b/lib/librte_ether/rte_ethdev.h 2014-05-14 11:27:07.409461720 -0700
> @@ -1704,6 +1704,45 @@ extern void rte_eth_link_get_nowait(uint
> struct rte_eth_link *link);
>
> +
> +/**
> + * @internal
> + * Set the link status from device in atomic fashion.
> + * Returns non-zero on success; 0 on failure
> + * Callers should retry on failure
> + */
> +static inline void rte_eth_dev_set_link(struct rte_eth_dev *dev,
> + struct rte_eth_link *link)
> +{
> + rte_atomic64_t *dst = (rte_atomic64_t *)&dev->data->dev_link;
> +
> + rte_atomic64_set(dst, *(int64_t *)link);
> +}
> +
>
>
--
Ivan BOULE
6WIND
Software Engineer
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [dpdk-dev] [PATCH 5/6] ether: allow setting mac address
2014-05-14 18:55 ` [dpdk-dev] [PATCH 5/6] ether: allow setting mac address Stephen Hemminger
@ 2014-05-15 9:04 ` Ivan Boule
2014-05-21 15:24 ` Ivan Boule
0 siblings, 1 reply; 11+ messages in thread
From: Ivan Boule @ 2014-05-15 9:04 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: dev
Hi Stephen,
By default, changing the [current] default MAC address of a device is not
equivalent to adding a MAC address to it ! At least, the minimum to do
should consist in:
1. checking that the PMD exports both mac_addr_remove()
and mac_addr_add() functions, returning a NON_SUPPORTED error code
otherwise,
2. removing the current default MAC address by invoking the PMD's
mac_addr_remove() function,
3. adding the new MAC address by invoking the PMD's mac_addr_add()
function.
Regards,
Ivan
2014-05-14 20:55 GMT+02:00 Stephen Hemminger <stephen@networkplumber.org>:
> Allow setting the default Ethernet address used on device.
> The underlying drivers allow it but the DPDK was blocking any
> attempts to change Ethernet address on device.
>
> For most devices, this is just the same as filling in address
> in mac address table entry 0, but for some devices this requires
> special override.
>
> Signed-off-by: Stephen Hemminger <shemming@brocade.com>
>
>
> --- a/lib/librte_ether/rte_ethdev.h 2014-05-14 11:44:27.373014227 -0700
> +++ b/lib/librte_ether/rte_ethdev.h 2014-05-14 11:44:27.373014227 -0700
> @@ -982,6 +982,11 @@ typedef void (*eth_mac_addr_add_t)(struc
> struct ether_addr *mac_addr,
> uint32_t index,
> uint32_t vmdq);
> +/**< @internal Change default MAC address */
> +
> +typedef void (*eth_mac_addr_set_t)(struct rte_eth_dev *dev,
> + struct ether_addr *mac_addr);
> +
> /**< @internal Set a MAC address into Receive Address Address Register */
>
> typedef int (*eth_uc_hash_table_set_t)(struct rte_eth_dev *dev,
> @@ -1114,6 +1119,7 @@ struct eth_dev_ops {
> priority_flow_ctrl_set_t priority_flow_ctrl_set; /**< Setup
> priority flow control.*/
> eth_mac_addr_remove_t mac_addr_remove; /**< Remove MAC
> address */
> eth_mac_addr_add_t mac_addr_add; /**< Add a MAC address */
> + eth_mac_addr_set_t mac_addr_set; /**< Change default MAC
> address */
> eth_uc_hash_table_set_t uc_hash_table_set; /**< Set Unicast
> Table Array */
> eth_uc_all_hash_table_set_t uc_all_hash_table_set; /**< Set
> Unicast hash bitmap */
> eth_mirror_rule_set_t mirror_rule_set; /**< Add a traffic
> mirror rule.*/
> @@ -2538,6 +2544,21 @@ int rte_eth_dev_mac_addr_add(uint8_t por
> int rte_eth_dev_mac_addr_remove(uint8_t port, struct ether_addr
> *mac_addr);
>
> /**
> + * Change the default MAC address
> + *
> + * @param port
> + * The port identifier of the Ethernet device.
> + * @param mac_addr
> + * MAC address to remove.
> + * @return
> + * - (0) if successfully added or *mac_addr" was already added.
> + * - (-ENOTSUP) if hardware doesn't support this feature.
> + * - (-ENODEV) if *port* is invalid.
> + * - (-EINVAL) if MAC address is invalid.
> + */
> +int rte_eth_dev_macaddr_set(uint8_t port, struct ether_addr *mac_addr);
> +
> +/**
> * Update Redirection Table(RETA) of Receive Side Scaling of Ethernet
> device.
> *
> * @param port
> --- a/lib/librte_ether/rte_ethdev.c 2014-05-14 11:44:27.373014227 -0700
> +++ b/lib/librte_ether/rte_ethdev.c 2014-05-14 11:44:27.373014227 -0700
> @@ -1059,6 +1059,33 @@ rte_eth_macaddr_get(uint8_t port_id, str
> }
>
> int
> +rte_eth_dev_macaddr_set(uint8_t port_id, struct ether_addr *addr)
> +{
> + struct rte_eth_dev *dev;
> +
> + if (port_id >= nb_ports) {
> + PMD_DEBUG_TRACE("Invalid port_id=%d\n", port_id);
> + return (-ENODEV);
> + }
> + if (is_zero_ether_addr(addr)) {
> + PMD_DEBUG_TRACE("port %d: Cannot add NULL MAC address\n",
> port_id);
> + return (-EINVAL);
> + }
> + dev = &rte_eth_devices[port_id];
> +
> + if (*dev->dev_ops->mac_addr_set)
> + (*dev->dev_ops->mac_addr_set)(dev, addr);
> + else {
> + FUNC_PTR_OR_ERR_RET(*dev->dev_ops->mac_addr_add, -ENOTSUP);
> + (*dev->dev_ops->mac_addr_add)(dev, addr, 0, 0);
> + }
> +
> + ether_addr_copy(addr, &dev->data->mac_addrs[0]);
> +
> + return 0;
> +}
> +
> +int
> rte_eth_dev_vlan_filter(uint8_t port_id, uint16_t vlan_id, int on)
> {
> struct rte_eth_dev *dev;
>
>
--
Ivan BOULE
6WIND
Software Engineer
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [dpdk-dev] [PATCH 6/6] vmxnet3: replace Intel driver
2014-05-14 18:55 ` [dpdk-dev] [PATCH 6/6] vmxnet3: replace Intel driver Stephen Hemminger
@ 2014-05-15 17:14 ` Thomas Monjalon
0 siblings, 0 replies; 11+ messages in thread
From: Thomas Monjalon @ 2014-05-15 17:14 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: dev, Stephen Hemminger
Hi Stephen,
2014-05-14 11:55, Stephen Hemminger:
> Replace the Intel VMXNET3 driver with the one developed by Brocade.
> The Intel code was a missing features, and had style and other issues.
>
> Major changes:
> - supports multi-segment
> - supports vlan offload
> - per queue stats
>
> Fixes:
> - link update reports link change correctly
> - allows reconfiguration of # of queues
> (original code would have too small memzone).
>
> Signed-off-by: Stephen Hemminger <shemming@brocade.com>
We cannot replace a driver by another one without comparing performances.
Actually there is no good reason to remove an implementation if some people
are willing to maintain it.
Please send a patch to import your vmxnet3 PMD as extension in a separated
repository like these ones:
http://dpdk.org/browse
This way, we can have a first integration step and discuss/test this
implementation during the next release cycle.
Thank you for contributing
--
Thomas
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [dpdk-dev] [PATCH 5/6] ether: allow setting mac address
2014-05-15 9:04 ` Ivan Boule
@ 2014-05-21 15:24 ` Ivan Boule
0 siblings, 0 replies; 11+ messages in thread
From: Ivan Boule @ 2014-05-21 15:24 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: dev
Hi Stephen,
Looking more precisely to your patch, I think that your requirement can
beaddressed without adding
a new `mac_addr_set` function into the [already overloaded] set of
operations exported by a PMD.
In fact, changing the default MAC address used by a port consists in
executingthe following sequence of operations:
1- rte_eth_macaddr_get(port_id, &cur_mac_addr) to obtain the
default MACaddress currently used by the port,
2- rte_eth_dev_mac_addr_remove(port_id, &cur_mac_addr) to remove
thedefault MAC address currently used by the port,
3- rte_eth_dev_mac_addr_add(port_id, &new_mac_addr) to add the new
defaultMAC address to be used by the port.
This method takes advantage of the fact that the allocation of a
freeentry into the array `data->mac_addrs` that
contains the MAC addresses of a port searches the pool from the entry at
index 0, which contains the default MAC
address. This implementation simply needs to be mandatory from now.
The actual implementation might look like this.
Add the prototype of the function `rte_eth_dev_macaddr_set` in the
file`rte_ethdev.h`:
extern int rte_eth_dev_macaddr_set((uint8_t port_id, struct
ether_addr *addr);
And add the following generic implementation of the function
`rte_eth_dev_macaddr_set`in the file `rte_ethdev.c`:
int
rte_eth_dev_macaddr_set(uint8_t port_id, struct ether_addr *addr)
{
struct rte_eth_dev *dev;
struct ether_addr cur_addr;
int diag;
if (port_id >= nb_ports) {
PMD_DEBUG_TRACE("Invalid port_id=%d\n", port_id);
return (-ENODEV);
}
if (is_zero_ether_addr(addr)) {
PMD_DEBUG_TRACE("port %d: Cannot add NULL MAC address\n",
port_id);
return (-EINVAL);
}
dev = &rte_eth_devices[port_id];
/*
* Check that both mac_addr_remove and mac_addr_add operations are
* supported by the PMD of the port.
*/
FUNC_PTR_OR_ERR_RET(*dev->dev_ops->mac_addr_remove, -ENOTSUP);
FUNC_PTR_OR_ERR_RET(*dev->dev_ops->mac_addr_add, -ENOTSUP);
rte_eth_macaddr_get(port_id, &cur_addr);
diag = rte_eth_dev_mac_addr_remove(port_id, &cur_addr);
if (diag < 0)
return diag;
return rte_eth_dev_mac_addr_add(port_id, addr);
}
Regards,
Ivan
On 05/15/2014 11:04 AM, Ivan Boule wrote:
> Hi Stephen,
>
> By default, changing the [current] default MAC address of a device is
> not equivalent to adding a MAC address to it ! At least, the minimum
> to do should consist in:
> 1. checking that the PMD exports both mac_addr_remove()
> and mac_addr_add() functions, returning a NON_SUPPORTED error code
> otherwise,
> 2. removing the current default MAC address by invoking the PMD's
> mac_addr_remove() function,
> 3. adding the new MAC address by invoking the PMD's mac_addr_add()
> function.
>
> Regards,
> Ivan
>
> 2014-05-14 20:55 GMT+02:00 Stephen Hemminger
> <stephen@networkplumber.org <mailto:stephen@networkplumber.org>>:
>
> Allow setting the default Ethernet address used on device.
> The underlying drivers allow it but the DPDK was blocking any
> attempts to change Ethernet address on device.
>
> For most devices, this is just the same as filling in address
> in mac address table entry 0, but for some devices this requires
> special override.
>
> Signed-off-by: Stephen Hemminger <shemming@brocade.com
> <mailto:shemming@brocade.com>>
>
>
> --- a/lib/librte_ether/rte_ethdev.h 2014-05-14
> 11:44:27.373014227 -0700
> +++ b/lib/librte_ether/rte_ethdev.h 2014-05-14
> 11:44:27.373014227 -0700
> @@ -982,6 +982,11 @@ typedef void (*eth_mac_addr_add_t)(struc
> struct ether_addr *mac_addr,
> uint32_t index,
> uint32_t vmdq);
> +/**< @internal Change default MAC address */
> +
> +typedef void (*eth_mac_addr_set_t)(struct rte_eth_dev *dev,
> + struct ether_addr *mac_addr);
> +
> /**< @internal Set a MAC address into Receive Address Address
> Register */
>
> typedef int (*eth_uc_hash_table_set_t)(struct rte_eth_dev *dev,
> @@ -1114,6 +1119,7 @@ struct eth_dev_ops {
> priority_flow_ctrl_set_t priority_flow_ctrl_set; /**<
> Setup priority flow control.*/
> eth_mac_addr_remove_t mac_addr_remove; /**< Remove
> MAC address */
> eth_mac_addr_add_t mac_addr_add; /**< Add a MAC
> address */
> + eth_mac_addr_set_t mac_addr_set; /**< Change
> default MAC address */
> eth_uc_hash_table_set_t uc_hash_table_set; /**< Set
> Unicast Table Array */
> eth_uc_all_hash_table_set_t uc_all_hash_table_set; /**<
> Set Unicast hash bitmap */
> eth_mirror_rule_set_t mirror_rule_set; /**< Add a
> traffic mirror rule.*/
> @@ -2538,6 +2544,21 @@ int rte_eth_dev_mac_addr_add(uint8_t por
> int rte_eth_dev_mac_addr_remove(uint8_t port, struct ether_addr
> *mac_addr);
>
> /**
> + * Change the default MAC address
> + *
> + * @param port
> + * The port identifier of the Ethernet device.
> + * @param mac_addr
> + * MAC address to remove.
> + * @return
> + * - (0) if successfully added or *mac_addr" was already added.
> + * - (-ENOTSUP) if hardware doesn't support this feature.
> + * - (-ENODEV) if *port* is invalid.
> + * - (-EINVAL) if MAC address is invalid.
> + */
> +int rte_eth_dev_macaddr_set(uint8_t port, struct ether_addr
> *mac_addr);
> +
> +/**
> * Update Redirection Table(RETA) of Receive Side Scaling of
> Ethernet device.
> *
> * @param port
> --- a/lib/librte_ether/rte_ethdev.c 2014-05-14
> 11:44:27.373014227 -0700
> +++ b/lib/librte_ether/rte_ethdev.c 2014-05-14
> 11:44:27.373014227 -0700
> @@ -1059,6 +1059,33 @@ rte_eth_macaddr_get(uint8_t port_id, str
> }
>
> int
> +rte_eth_dev_macaddr_set(uint8_t port_id, struct ether_addr *addr)
> +{
> + struct rte_eth_dev *dev;
> +
> + if (port_id >= nb_ports) {
> + PMD_DEBUG_TRACE("Invalid port_id=%d\n", port_id);
> + return (-ENODEV);
> + }
> + if (is_zero_ether_addr(addr)) {
> + PMD_DEBUG_TRACE("port %d: Cannot add NULL MAC
> address\n", port_id);
> + return (-EINVAL);
> + }
> + dev = &rte_eth_devices[port_id];
> +
> + if (*dev->dev_ops->mac_addr_set)
> + (*dev->dev_ops->mac_addr_set)(dev, addr);
> + else {
> + FUNC_PTR_OR_ERR_RET(*dev->dev_ops->mac_addr_add, -ENOTSUP);
> + (*dev->dev_ops->mac_addr_add)(dev, addr, 0, 0);
> + }
> +
> + ether_addr_copy(addr, &dev->data->mac_addrs[0]);
> +
> + return 0;
> +}
> +
> +int
> rte_eth_dev_vlan_filter(uint8_t port_id, uint16_t vlan_id, int on)
> {
> struct rte_eth_dev *dev;
>
>
>
> --
> Ivan BOULE
> 6WIND
> Software Engineer
>
--
Ivan Boule
6WIND Development Engineer
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2014-05-21 15:23 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-05-14 18:55 [dpdk-dev] [PATCH 0/6] Ethernet driver enhancements Stephen Hemminger
2014-05-14 18:55 ` [dpdk-dev] [PATCH 1/6] Subjec: ethdev: add macro to cover all checksum flags Stephen Hemminger
2014-05-14 18:55 ` [dpdk-dev] [PATCH 2/6] Subjet: ethdev: add common code to atomicly access link Stephen Hemminger
2014-05-15 8:43 ` Ivan Boule
2014-05-14 18:55 ` [dpdk-dev] [PATCH 3/6] ixgbe: use eth_dev_{get,set}_link Stephen Hemminger
2014-05-14 18:55 ` [dpdk-dev] [PATCH 4/6] e1000: " Stephen Hemminger
2014-05-14 18:55 ` [dpdk-dev] [PATCH 5/6] ether: allow setting mac address Stephen Hemminger
2014-05-15 9:04 ` Ivan Boule
2014-05-21 15:24 ` Ivan Boule
2014-05-14 18:55 ` [dpdk-dev] [PATCH 6/6] vmxnet3: replace Intel driver Stephen Hemminger
2014-05-15 17:14 ` Thomas Monjalon
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).