OpenE2K
/
linux
6
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

net: Delete NETDEVICES_MULTIQUEUE kconfig option. 

Multiple TX queue support is a core networking feature.

Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
David S. Miller 2008-07-08 23:14:24 -07:00
parent c773e847ea
commit b19fa1fa91
9 changed files with 9 additions and 212 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

79
Documentation/networking/multiqueue.txt
View File

@ -3,19 +3,11 @@
===========================================
Section 1: Base driver requirements for implementing multiqueue support
Section 2: Qdisc support for multiqueue devices
Section 3: Brief howto using PRIO or RR for multiqueue devices
Intro: Kernel support for multiqueue devices
---------------------------------------------------------
Kernel support for multiqueue devices is only an API that is presented to the
netdevice layer for base drivers to implement. This feature is part of the
core networking stack, and all network devices will be running on the
multiqueue-aware stack. If a base driver only has one queue, then these
changes are transparent to that driver.
Kernel support for multiqueue devices is always present.
Section 1: Base driver requirements for implementing multiqueue support
-----------------------------------------------------------------------
@ -43,73 +35,4 @@ bitmap on device initialization. Below is an example from e1000:
netdev->features |= NETIF_F_MULTI_QUEUE;
#endif
Section 2: Qdisc support for multiqueue devices
-----------------------------------------------
Currently two qdiscs support multiqueue devices. A new round-robin qdisc,
sch_rr, and sch_prio. The qdisc is responsible for classifying the skb's to
bands and queues, and will store the queue mapping into skb->queue_mapping.
Use this field in the base driver to determine which queue to send the skb
to.
sch_rr has been added for hardware that doesn't want scheduling policies from
software, so it's a straight round-robin qdisc. It uses the same syntax and
classification priomap that sch_prio uses, so it should be intuitive to
configure for people who've used sch_prio.
In order to utilitize the multiqueue features of the qdiscs, the network
device layer needs to enable multiple queue support. This can be done by
selecting NETDEVICES_MULTIQUEUE under Drivers.
The PRIO qdisc naturally plugs into a multiqueue device. If
NETDEVICES_MULTIQUEUE is selected, then on qdisc load, the number of
bands requested is compared to the number of queues on the hardware. If they
are equal, it sets a one-to-one mapping up between the queues and bands. If
they're not equal, it will not load the qdisc. This is the same behavior
for RR. Once the association is made, any skb that is classified will have
skb->queue_mapping set, which will allow the driver to properly queue skb's
to multiple queues.
Section 3: Brief howto using PRIO and RR for multiqueue devices
---------------------------------------------------------------
The userspace command 'tc,' part of the iproute2 package, is used to configure
qdiscs. To add the PRIO qdisc to your network device, assuming the device is
called eth0, run the following command:
# tc qdisc add dev eth0 root handle 1: prio bands 4 multiqueue
This will create 4 bands, 0 being highest priority, and associate those bands
to the queues on your NIC. Assuming eth0 has 4 Tx queues, the band mapping
would look like:
band 0 => queue 0
band 1 => queue 1
band 2 => queue 2
band 3 => queue 3
Traffic will begin flowing through each queue if your TOS values are assigning
traffic across the various bands. For example, ssh traffic will always try to
go out band 0 based on TOS -> Linux priority conversion (realtime traffic),
so it will be sent out queue 0. ICMP traffic (pings) fall into the "normal"
traffic classification, which is band 1. Therefore pings will be send out
queue 1 on the NIC.
Note the use of the multiqueue keyword. This is only in versions of iproute2
that support multiqueue networking devices; if this is omitted when loading
a qdisc onto a multiqueue device, the qdisc will load and operate the same
if it were loaded onto a single-queue device (i.e. - sends all traffic to
queue 0).
Another alternative to multiqueue band allocation can be done by using the
multiqueue option and specify 0 bands. If this is the case, the qdisc will
allocate the number of bands to equal the number of queues that the device
reports, and bring the qdisc online.
The behavior of tc filters remains the same, where it will override TOS priority
classification.
Author: Peter P. Waskiewicz Jr. <peter.p.waskiewicz.jr@intel.com>

8
drivers/net/Kconfig
View File

@ -26,14 +26,6 @@ menuconfig NETDEVICES
# that for each of the symbols.
if NETDEVICES
config NETDEVICES_MULTIQUEUE
bool "Netdevice multiple hardware queue support"
---help---
Say Y here if you want to allow the network stack to use multiple
hardware TX queues on an ethernet device.
Most people will say N here.
config IFB
tristate "Intermediate Functional Block support"
depends on NET_CLS_ACT

14
drivers/net/cpmac.c
View File

@ -569,11 +569,7 @@ static int cpmac_start_xmit(struct sk_buff *skb, struct net_device *dev)
len = max(skb->len, ETH_ZLEN);
queue = skb_get_queue_mapping(skb);
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
netif_stop_subqueue(dev, queue);
#else
netif_stop_queue(dev);
#endif
desc = &priv->desc_ring[queue];
if (unlikely(desc->dataflags & CPMAC_OWN)) {
@ -626,24 +622,14 @@ static void cpmac_end_xmit(struct net_device *dev, int queue)
dev_kfree_skb_irq(desc->skb);
desc->skb = NULL;
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
if (netif_subqueue_stopped(dev, queue))
netif_wake_subqueue(dev, queue);
#else
if (netif_queue_stopped(dev))
netif_wake_queue(dev);
#endif
} else {
if (netif_msg_tx_err(priv) && net_ratelimit())
printk(KERN_WARNING
"%s: end_xmit: spurious interrupt\n", dev->name);
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
if (netif_subqueue_stopped(dev, queue))
netif_wake_subqueue(dev, queue);
#else
if (netif_queue_stopped(dev))
netif_wake_queue(dev);
#endif
}
}

6
drivers/net/ixgbe/ixgbe_ethtool.c
View File

@ -252,21 +252,15 @@ static int ixgbe_set_tso(struct net_device *netdev, u32 data)
netdev->features |= NETIF_F_TSO;
netdev->features |= NETIF_F_TSO6;
} else {
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
struct ixgbe_adapter *adapter = netdev_priv(netdev);
int i;
#endif
netif_stop_queue(netdev);
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
for (i = 0; i < adapter->num_tx_queues; i++)
netif_stop_subqueue(netdev, i);
#endif
netdev->features &= ~NETIF_F_TSO;
netdev->features &= ~NETIF_F_TSO6;
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
for (i = 0; i < adapter->num_tx_queues; i++)
netif_start_subqueue(netdev, i);
#endif
netif_start_queue(netdev);
}
return 0;

40
drivers/net/ixgbe/ixgbe_main.c
View File

@ -266,28 +266,16 @@ static bool ixgbe_clean_tx_irq(struct ixgbe_adapter *adapter,
* sees the new next_to_clean.
*/
smp_mb();
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
if (__netif_subqueue_stopped(netdev, tx_ring->queue_index) &&
!test_bit(__IXGBE_DOWN, &adapter->state)) {
netif_wake_subqueue(netdev, tx_ring->queue_index);
adapter->restart_queue++;
}
#else
if (netif_queue_stopped(netdev) &&
!test_bit(__IXGBE_DOWN, &adapter->state)) {
netif_wake_queue(netdev);
adapter->restart_queue++;
}
#endif
}
if (adapter->detect_tx_hung)
if (ixgbe_check_tx_hang(adapter, tx_ring, eop, eop_desc))
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
netif_stop_subqueue(netdev, tx_ring->queue_index);
#else
netif_stop_queue(netdev);
#endif
if (total_tx_packets >= tx_ring->work_limit)
IXGBE_WRITE_REG(&adapter->hw, IXGBE_EICS, tx_ring->eims_value);
@ -2192,11 +2180,7 @@ static void __devinit ixgbe_set_num_queues(struct ixgbe_adapter *adapter)
case (IXGBE_FLAG_RSS_ENABLED):
rss_m = 0xF;
nrq = rss_i;
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
ntq = rss_i;
#else
ntq = 1;
#endif
break;
case 0:
default:
@ -2370,10 +2354,8 @@ try_msi:
}
out:
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
/* Notify the stack of the (possibly) reduced Tx Queue count. */
adapter->netdev->egress_subqueue_count = adapter->num_tx_queues;
#endif
return err;
}
@ -2910,9 +2892,7 @@ static void ixgbe_watchdog(unsigned long data)
struct net_device *netdev = adapter->netdev;
bool link_up;
u32 link_speed = 0;
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
int i;
#endif
adapter->hw.mac.ops.check_link(&adapter->hw, &(link_speed), &link_up);
@ -2934,10 +2914,8 @@ static void ixgbe_watchdog(unsigned long data)
netif_carrier_on(netdev);
netif_wake_queue(netdev);
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
for (i = 0; i < adapter->num_tx_queues; i++)
netif_wake_subqueue(netdev, i);
#endif
} else {
/* Force detection of hung controller */
adapter->detect_tx_hung = true;
@ -3264,11 +3242,7 @@ static int __ixgbe_maybe_stop_tx(struct net_device *netdev,
{
struct ixgbe_adapter *adapter = netdev_priv(netdev);
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
netif_stop_subqueue(netdev, tx_ring->queue_index);
#else
netif_stop_queue(netdev);
#endif
/* Herbert's original patch had:
* smp_mb__after_netif_stop_queue();
* but since that doesn't exist yet, just open code it. */
@ -3280,11 +3254,7 @@ static int __ixgbe_maybe_stop_tx(struct net_device *netdev,
return -EBUSY;
/* A reprieve! - use start_queue because it doesn't call schedule */
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
netif_wake_subqueue(netdev, tx_ring->queue_index);
#else
netif_wake_queue(netdev);
#endif
++adapter->restart_queue;
return 0;
}
@ -3312,9 +3282,7 @@ static int ixgbe_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
unsigned int f;
unsigned int nr_frags = skb_shinfo(skb)->nr_frags;
len -= skb->data_len;
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
r_idx = (adapter->num_tx_queues - 1) & skb->queue_mapping;
#endif
tx_ring = &adapter->tx_ring[r_idx];
@ -3502,11 +3470,7 @@ static int __devinit ixgbe_probe(struct pci_dev *pdev,
pci_set_master(pdev);
pci_save_state(pdev);
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
netdev = alloc_etherdev_mq(sizeof(struct ixgbe_adapter), MAX_TX_QUEUES);
#else
netdev = alloc_etherdev(sizeof(struct ixgbe_adapter));
#endif
if (!netdev) {
err = -ENOMEM;
goto err_alloc_etherdev;
@ -3598,9 +3562,7 @@ static int __devinit ixgbe_probe(struct pci_dev *pdev,
if (pci_using_dac)
netdev->features |= NETIF_F_HIGHDMA;
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
netdev->features |= NETIF_F_MULTI_QUEUE;
#endif
/* make sure the EEPROM is good */
if (ixgbe_validate_eeprom_checksum(hw, NULL) < 0) {
@ -3668,10 +3630,8 @@ static int __devinit ixgbe_probe(struct pci_dev *pdev,
netif_carrier_off(netdev);
netif_stop_queue(netdev);
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
for (i = 0; i < adapter->num_tx_queues; i++)
netif_stop_subqueue(netdev, i);
#endif
ixgbe_napi_add_all(adapter);

47
drivers/net/s2io.c
View File

@ -546,13 +546,10 @@ static struct pci_driver s2io_driver = {
static inline void s2io_stop_all_tx_queue(struct s2io_nic *sp)
{
int i;
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
if (sp->config.multiq) {
for (i = 0; i < sp->config.tx_fifo_num; i++)
netif_stop_subqueue(sp->dev, i);
} else
#endif
{
} else {
for (i = 0; i < sp->config.tx_fifo_num; i++)
sp->mac_control.fifos[i].queue_state = FIFO_QUEUE_STOP;
netif_stop_queue(sp->dev);
@ -561,12 +558,9 @@ static inline void s2io_stop_all_tx_queue(struct s2io_nic *sp)
static inline void s2io_stop_tx_queue(struct s2io_nic *sp, int fifo_no)
{
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
if (sp->config.multiq)
netif_stop_subqueue(sp->dev, fifo_no);
else
#endif
{
else {
sp->mac_control.fifos[fifo_no].queue_state =
FIFO_QUEUE_STOP;
netif_stop_queue(sp->dev);
@ -576,13 +570,10 @@ static inline void s2io_stop_tx_queue(struct s2io_nic *sp, int fifo_no)
static inline void s2io_start_all_tx_queue(struct s2io_nic *sp)
{
int i;
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
if (sp->config.multiq) {
for (i = 0; i < sp->config.tx_fifo_num; i++)
netif_start_subqueue(sp->dev, i);
} else
#endif
{
} else {
for (i = 0; i < sp->config.tx_fifo_num; i++)
sp->mac_control.fifos[i].queue_state = FIFO_QUEUE_START;
netif_start_queue(sp->dev);
@ -591,12 +582,9 @@ static inline void s2io_start_all_tx_queue(struct s2io_nic *sp)
static inline void s2io_start_tx_queue(struct s2io_nic *sp, int fifo_no)
{
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
if (sp->config.multiq)
netif_start_subqueue(sp->dev, fifo_no);
else
#endif
{
else {
sp->mac_control.fifos[fifo_no].queue_state =
FIFO_QUEUE_START;
netif_start_queue(sp->dev);
@ -606,13 +594,10 @@ static inline void s2io_start_tx_queue(struct s2io_nic *sp, int fifo_no)
static inline void s2io_wake_all_tx_queue(struct s2io_nic *sp)
{
int i;
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
if (sp->config.multiq) {
for (i = 0; i < sp->config.tx_fifo_num; i++)
netif_wake_subqueue(sp->dev, i);
} else
#endif
{
} else {
for (i = 0; i < sp->config.tx_fifo_num; i++)
sp->mac_control.fifos[i].queue_state = FIFO_QUEUE_START;
netif_wake_queue(sp->dev);
@ -623,13 +608,10 @@ static inline void s2io_wake_tx_queue(
struct fifo_info *fifo, int cnt, u8 multiq)
{
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
if (multiq) {
if (cnt && __netif_subqueue_stopped(fifo->dev, fifo->fifo_no))
netif_wake_subqueue(fifo->dev, fifo->fifo_no);
} else
#endif
if (cnt && (fifo->queue_state == FIFO_QUEUE_STOP)) {
} else if (cnt && (fifo->queue_state == FIFO_QUEUE_STOP)) {
if (netif_queue_stopped(fifo->dev)) {
fifo->queue_state = FIFO_QUEUE_START;
netif_wake_queue(fifo->dev);
@ -4189,15 +4171,12 @@ static int s2io_xmit(struct sk_buff *skb, struct net_device *dev)
return NETDEV_TX_LOCKED;
}
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
if (sp->config.multiq) {
if (__netif_subqueue_stopped(dev, fifo->fifo_no)) {
spin_unlock_irqrestore(&fifo->tx_lock, flags);
return NETDEV_TX_BUSY;
}
} else
#endif
if (unlikely(fifo->queue_state == FIFO_QUEUE_STOP)) {
} else if (unlikely(fifo->queue_state == FIFO_QUEUE_STOP)) {
if (netif_queue_stopped(dev)) {
spin_unlock_irqrestore(&fifo->tx_lock, flags);
return NETDEV_TX_BUSY;
@ -7633,12 +7612,6 @@ static int s2io_verify_parm(struct pci_dev *pdev, u8 *dev_intr_type,
DBG_PRINT(ERR_DBG, "tx fifos\n");
}
#ifndef CONFIG_NETDEVICES_MULTIQUEUE
if (multiq) {
DBG_PRINT(ERR_DBG, "s2io: Multiqueue support not enabled\n");
multiq = 0;
}
#endif
if (multiq)
*dev_multiq = multiq;
@ -7783,12 +7756,10 @@ s2io_init_nic(struct pci_dev *pdev, const struct pci_device_id *pre)
pci_disable_device(pdev);
return -ENODEV;
}
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
if (dev_multiq)
dev = alloc_etherdev_mq(sizeof(struct s2io_nic), tx_fifo_num);
else
#endif
dev = alloc_etherdev(sizeof(struct s2io_nic));
dev = alloc_etherdev(sizeof(struct s2io_nic));
if (dev == NULL) {
DBG_PRINT(ERR_DBG, "Device allocation failed\n");
pci_disable_device(pdev);
@ -7979,10 +7950,8 @@ s2io_init_nic(struct pci_dev *pdev, const struct pci_device_id *pre)
dev->features |= NETIF_F_UFO;
dev->features |= NETIF_F_HW_CSUM;
}
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
if (config->multiq)
dev->features |= NETIF_F_MULTI_QUEUE;
#endif
dev->tx_timeout = &s2io_tx_watchdog;
dev->watchdog_timeo = WATCH_DOG_TIMEOUT;
INIT_WORK(&sp->rst_timer_task, s2io_restart_nic);

14
include/linux/netdevice.h
View File

@ -1043,9 +1043,7 @@ static inline int netif_running(const struct net_device *dev)
*/
static inline void netif_start_subqueue(struct net_device *dev, u16 queue_index)
{
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
clear_bit(__LINK_STATE_XOFF, &dev->egress_subqueue[queue_index].state);
#endif
}
/**
@ -1057,13 +1055,11 @@ static inline void netif_start_subqueue(struct net_device *dev, u16 queue_index)
*/
static inline void netif_stop_subqueue(struct net_device *dev, u16 queue_index)
{
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
#ifdef CONFIG_NETPOLL_TRAP
if (netpoll_trap())
return;
#endif
set_bit(__LINK_STATE_XOFF, &dev->egress_subqueue[queue_index].state);
#endif
}
/**
@ -1076,12 +1072,8 @@ static inline void netif_stop_subqueue(struct net_device *dev, u16 queue_index)
static inline int __netif_subqueue_stopped(const struct net_device *dev,
u16 queue_index)
{
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
return test_bit(__LINK_STATE_XOFF,
&dev->egress_subqueue[queue_index].state);
#else
return 0;
#endif
}
static inline int netif_subqueue_stopped(const struct net_device *dev,
@ -1099,7 +1091,6 @@ static inline int netif_subqueue_stopped(const struct net_device *dev,
*/
static inline void netif_wake_subqueue(struct net_device *dev, u16 queue_index)
{
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
#ifdef CONFIG_NETPOLL_TRAP
if (netpoll_trap())
return;
@ -1107,7 +1098,6 @@ static inline void netif_wake_subqueue(struct net_device *dev, u16 queue_index)
if (test_and_clear_bit(__LINK_STATE_XOFF,
&dev->egress_subqueue[queue_index].state))
__netif_schedule(&dev->tx_queue);
#endif
}
/**
@ -1119,11 +1109,7 @@ static inline void netif_wake_subqueue(struct net_device *dev, u16 queue_index)
*/
static inline int netif_is_multiqueue(const struct net_device *dev)
{
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
return (!!(NETIF_F_MULTI_QUEUE & dev->features));
#else
return 0;
#endif
}
/* Use this variant when it is known for sure that it

10
include/linux/skbuff.h
View File

@ -305,9 +305,7 @@ struct sk_buff {
#endif
int iif;
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
__u16 queue_mapping;
#endif
#ifdef CONFIG_NET_SCHED
__u16 tc_index; /* traffic control index */
#ifdef CONFIG_NET_CLS_ACT
@ -1671,25 +1669,17 @@ static inline void skb_init_secmark(struct sk_buff *skb)
static inline void skb_set_queue_mapping(struct sk_buff *skb, u16 queue_mapping)
{
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
skb->queue_mapping = queue_mapping;
#endif
}
static inline u16 skb_get_queue_mapping(struct sk_buff *skb)
{
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
return skb->queue_mapping;
#else
return 0;
#endif
}
static inline void skb_copy_queue_mapping(struct sk_buff *to, const struct sk_buff *from)
{
#ifdef CONFIG_NETDEVICES_MULTIQUEUE
to->queue_mapping = from->queue_mapping;
#endif
}
static inline int skb_is_gso(const struct sk_buff *skb)

3
net/mac80211/Kconfig
View File

@ -15,14 +15,11 @@ config MAC80211_QOS
def_bool y
depends on MAC80211
depends on NET_SCHED
depends on NETDEVICES_MULTIQUEUE
comment "QoS/HT support disabled"
depends on MAC80211 && !MAC80211_QOS
comment "QoS/HT support needs CONFIG_NET_SCHED"
depends on MAC80211 && !NET_SCHED
comment "QoS/HT support needs CONFIG_NETDEVICES_MULTIQUEUE"
depends on MAC80211 && !NETDEVICES_MULTIQUEUE
menu "Rate control algorithm selection"
depends on MAC80211 != n