korg/linux
0
0
Fork 0
mirror of https://mirrors.bfsu.edu.cn/git/linux.git synced 2024-11-25 13:14:07 +08:00
Code Issues Packages Projects Releases Wiki Activity
104439a887
linux/net/sched/sch_generic.c

620 lines
14 KiB
C
Raw Normal View History

Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* net/sched/sch_generic.c Generic packet scheduler routines.
*
* 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; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
* Jamal Hadi Salim, <hadi@cyberus.ca> 990601
* - Ingress support
*/
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/init.h>
#include <linux/rcupdate.h>
#include <linux/list.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
/* Main transmission queue. */
/* Main qdisc structure lock.
However, modifications
to data, participating in scheduling must be additionally
protected with dev->queue_lock spinlock.
The idea is the following:
- enqueue, dequeue are serialized via top level device
spinlock dev->queue_lock.
[NET_SCHED]: Fix fallout from dev->qdisc RCU change The move of qdisc destruction to a rcu callback broke locking in the entire qdisc layer by invalidating previously valid assumptions about the context in which changes to the qdisc tree occur. The two assumptions were: - since changes only happen in process context, read_lock doesn't need bottem half protection. Now invalid since destruction of inner qdiscs, classifiers, actions and estimators happens in the RCU callback unless they're manually deleted, resulting in dead-locks when read_lock in process context is interrupted by write_lock_bh in bottem half context. - since changes only happen under the RTNL, no additional locking is necessary for data not used during packet processing (f.e. u32_list). Again, since destruction now happens in the RCU callback, this assumption is not valid anymore, causing races while using this data, which can result in corruption or use-after-free. Instead of "fixing" this by disabling bottem halfs everywhere and adding new locks/refcounting, this patch makes these assumptions valid again by moving destruction back to process context. Since only the dev->qdisc pointer is protected by RCU, but ->enqueue and the qdisc tree are still protected by dev->qdisc_lock, destruction of the tree can be performed immediately and only the final free needs to happen in the rcu callback to make sure dev_queue_xmit doesn't access already freed memory. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 07:45:45 +08:00
- tree walking is protected by read_lock(qdisc_tree_lock)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
and this lock is used only in process context.
[NET_SCHED]: Fix fallout from dev->qdisc RCU change The move of qdisc destruction to a rcu callback broke locking in the entire qdisc layer by invalidating previously valid assumptions about the context in which changes to the qdisc tree occur. The two assumptions were: - since changes only happen in process context, read_lock doesn't need bottem half protection. Now invalid since destruction of inner qdiscs, classifiers, actions and estimators happens in the RCU callback unless they're manually deleted, resulting in dead-locks when read_lock in process context is interrupted by write_lock_bh in bottem half context. - since changes only happen under the RTNL, no additional locking is necessary for data not used during packet processing (f.e. u32_list). Again, since destruction now happens in the RCU callback, this assumption is not valid anymore, causing races while using this data, which can result in corruption or use-after-free. Instead of "fixing" this by disabling bottem halfs everywhere and adding new locks/refcounting, this patch makes these assumptions valid again by moving destruction back to process context. Since only the dev->qdisc pointer is protected by RCU, but ->enqueue and the qdisc tree are still protected by dev->qdisc_lock, destruction of the tree can be performed immediately and only the final free needs to happen in the rcu callback to make sure dev_queue_xmit doesn't access already freed memory. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 07:45:45 +08:00
- updates to tree are made only under rtnl semaphore,
hence this lock may be made without local bh disabling.
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
qdisc_tree_lock must be grabbed BEFORE dev->queue_lock!
*/
DEFINE_RWLOCK(qdisc_tree_lock);
void qdisc_lock_tree(struct net_device *dev)
{
[NET_SCHED]: Fix fallout from dev->qdisc RCU change The move of qdisc destruction to a rcu callback broke locking in the entire qdisc layer by invalidating previously valid assumptions about the context in which changes to the qdisc tree occur. The two assumptions were: - since changes only happen in process context, read_lock doesn't need bottem half protection. Now invalid since destruction of inner qdiscs, classifiers, actions and estimators happens in the RCU callback unless they're manually deleted, resulting in dead-locks when read_lock in process context is interrupted by write_lock_bh in bottem half context. - since changes only happen under the RTNL, no additional locking is necessary for data not used during packet processing (f.e. u32_list). Again, since destruction now happens in the RCU callback, this assumption is not valid anymore, causing races while using this data, which can result in corruption or use-after-free. Instead of "fixing" this by disabling bottem halfs everywhere and adding new locks/refcounting, this patch makes these assumptions valid again by moving destruction back to process context. Since only the dev->qdisc pointer is protected by RCU, but ->enqueue and the qdisc tree are still protected by dev->qdisc_lock, destruction of the tree can be performed immediately and only the final free needs to happen in the rcu callback to make sure dev_queue_xmit doesn't access already freed memory. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 07:45:45 +08:00
write_lock(&qdisc_tree_lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
spin_lock_bh(&dev->queue_lock);
}
void qdisc_unlock_tree(struct net_device *dev)
{
spin_unlock_bh(&dev->queue_lock);
[NET_SCHED]: Fix fallout from dev->qdisc RCU change The move of qdisc destruction to a rcu callback broke locking in the entire qdisc layer by invalidating previously valid assumptions about the context in which changes to the qdisc tree occur. The two assumptions were: - since changes only happen in process context, read_lock doesn't need bottem half protection. Now invalid since destruction of inner qdiscs, classifiers, actions and estimators happens in the RCU callback unless they're manually deleted, resulting in dead-locks when read_lock in process context is interrupted by write_lock_bh in bottem half context. - since changes only happen under the RTNL, no additional locking is necessary for data not used during packet processing (f.e. u32_list). Again, since destruction now happens in the RCU callback, this assumption is not valid anymore, causing races while using this data, which can result in corruption or use-after-free. Instead of "fixing" this by disabling bottem halfs everywhere and adding new locks/refcounting, this patch makes these assumptions valid again by moving destruction back to process context. Since only the dev->qdisc pointer is protected by RCU, but ->enqueue and the qdisc tree are still protected by dev->qdisc_lock, destruction of the tree can be performed immediately and only the final free needs to happen in the rcu callback to make sure dev_queue_xmit doesn't access already freed memory. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 07:45:45 +08:00
write_unlock(&qdisc_tree_lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
/*
dev->queue_lock serializes queue accesses for this device
AND dev->qdisc pointer itself.
[NET]: Add netif_tx_lock Various drivers use xmit_lock internally to synchronise with their transmission routines. They do so without setting xmit_lock_owner. This is fine as long as netpoll is not in use. With netpoll it is possible for deadlocks to occur if xmit_lock_owner isn't set. This is because if a printk occurs while xmit_lock is held and xmit_lock_owner is not set can cause netpoll to attempt to take xmit_lock recursively. While it is possible to resolve this by getting netpoll to use trylock, it is suboptimal because netpoll's sole objective is to maximise the chance of getting the printk out on the wire. So delaying or dropping the message is to be avoided as much as possible. So the only alternative is to always set xmit_lock_owner. The following patch does this by introducing the netif_tx_lock family of functions that take care of setting/unsetting xmit_lock_owner. I renamed xmit_lock to _xmit_lock to indicate that it should not be used directly. I didn't provide irq versions of the netif_tx_lock functions since xmit_lock is meant to be a BH-disabling lock. This is pretty much a straight text substitution except for a small bug fix in winbond. It currently uses netif_stop_queue/spin_unlock_wait to stop transmission. This is unsafe as an IRQ can potentially wake up the queue. So it is safer to use netif_tx_disable. The hamradio bits used spin_lock_irq but it is unnecessary as xmit_lock must never be taken in an IRQ handler. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-10 03:20:56 +08:00
netif_tx_lock serializes accesses to device driver.
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[NET]: Add netif_tx_lock Various drivers use xmit_lock internally to synchronise with their transmission routines. They do so without setting xmit_lock_owner. This is fine as long as netpoll is not in use. With netpoll it is possible for deadlocks to occur if xmit_lock_owner isn't set. This is because if a printk occurs while xmit_lock is held and xmit_lock_owner is not set can cause netpoll to attempt to take xmit_lock recursively. While it is possible to resolve this by getting netpoll to use trylock, it is suboptimal because netpoll's sole objective is to maximise the chance of getting the printk out on the wire. So delaying or dropping the message is to be avoided as much as possible. So the only alternative is to always set xmit_lock_owner. The following patch does this by introducing the netif_tx_lock family of functions that take care of setting/unsetting xmit_lock_owner. I renamed xmit_lock to _xmit_lock to indicate that it should not be used directly. I didn't provide irq versions of the netif_tx_lock functions since xmit_lock is meant to be a BH-disabling lock. This is pretty much a straight text substitution except for a small bug fix in winbond. It currently uses netif_stop_queue/spin_unlock_wait to stop transmission. This is unsafe as an IRQ can potentially wake up the queue. So it is safer to use netif_tx_disable. The hamradio bits used spin_lock_irq but it is unnecessary as xmit_lock must never be taken in an IRQ handler. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-10 03:20:56 +08:00
dev->queue_lock and netif_tx_lock are mutually exclusive,
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if one is grabbed, another must be free.
*/
/* Kick device.
Note, that this procedure can be called by a watchdog timer, so that
we do not check dev->tbusy flag here.
Returns: 0 - queue is empty.
>0 - queue is not empty, but throttled.
<0 - queue is not empty. Device is throttled, if dev->tbusy != 0.
NOTE: Called under dev->queue_lock with locally disabled BH.
*/
[NET]: Prevent multiple qdisc runs Having two or more qdisc_run's contend against each other is bad because it can induce packet reordering if the packets have to be requeued. It appears that this is an unintended consequence of relinquinshing the queue lock while transmitting. That in turn is needed for devices that spend a lot of time in their transmit routine. There are no advantages to be had as devices with queues are inherently single-threaded (the loopback device is not but then it doesn't have a queue). Even if you were to add a queue to a parallel virtual device (e.g., bolt a tbf filter in front of an ipip tunnel device), you would still want to process the queue in sequence to ensure that the packets are ordered correctly. The solution here is to steal a bit from net_device to prevent this. BTW, as qdisc_restart is no longer used by anyone as a module inside the kernel (IIRC it used to with netif_wake_queue), I have not exported the new __qdisc_run function. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-20 14:57:59 +08:00
static inline int qdisc_restart(struct net_device *dev)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
struct Qdisc *q = dev->qdisc;
struct sk_buff *skb;
/* Dequeue packet */
[NET]: Add generic segmentation offload This patch adds the infrastructure for generic segmentation offload. The idea is to tap into the potential savings of TSO without hardware support by postponing the allocation of segmented skb's until just before the entry point into the NIC driver. The same structure can be used to support software IPv6 TSO, as well as UFO and segmentation offload for other relevant protocols, e.g., DCCP. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 17:57:17 +08:00
if (((skb = dev->gso_skb)) || ((skb = q->dequeue(q)))) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
unsigned nolock = (dev->features & NETIF_F_LLTX);
[NET]: Add generic segmentation offload This patch adds the infrastructure for generic segmentation offload. The idea is to tap into the potential savings of TSO without hardware support by postponing the allocation of segmented skb's until just before the entry point into the NIC driver. The same structure can be used to support software IPv6 TSO, as well as UFO and segmentation offload for other relevant protocols, e.g., DCCP. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 17:57:17 +08:00
dev->gso_skb = NULL;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* When the driver has LLTX set it does its own locking
* in start_xmit. No need to add additional overhead by
* locking again. These checks are worth it because
* even uncongested locks can be quite expensive.
* The driver can do trylock like here too, in case
* of lock congestion it should return -1 and the packet
* will be requeued.
*/
if (!nolock) {
[NET]: Add netif_tx_lock Various drivers use xmit_lock internally to synchronise with their transmission routines. They do so without setting xmit_lock_owner. This is fine as long as netpoll is not in use. With netpoll it is possible for deadlocks to occur if xmit_lock_owner isn't set. This is because if a printk occurs while xmit_lock is held and xmit_lock_owner is not set can cause netpoll to attempt to take xmit_lock recursively. While it is possible to resolve this by getting netpoll to use trylock, it is suboptimal because netpoll's sole objective is to maximise the chance of getting the printk out on the wire. So delaying or dropping the message is to be avoided as much as possible. So the only alternative is to always set xmit_lock_owner. The following patch does this by introducing the netif_tx_lock family of functions that take care of setting/unsetting xmit_lock_owner. I renamed xmit_lock to _xmit_lock to indicate that it should not be used directly. I didn't provide irq versions of the netif_tx_lock functions since xmit_lock is meant to be a BH-disabling lock. This is pretty much a straight text substitution except for a small bug fix in winbond. It currently uses netif_stop_queue/spin_unlock_wait to stop transmission. This is unsafe as an IRQ can potentially wake up the queue. So it is safer to use netif_tx_disable. The hamradio bits used spin_lock_irq but it is unnecessary as xmit_lock must never be taken in an IRQ handler. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-10 03:20:56 +08:00
if (!netif_tx_trylock(dev)) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
collision:
/* So, someone grabbed the driver. */
/* It may be transient configuration error,
when hard_start_xmit() recurses. We detect
it by checking xmit owner and drop the
packet when deadloop is detected.
*/
if (dev->xmit_lock_owner == smp_processor_id()) {
kfree_skb(skb);
if (net_ratelimit())
printk(KERN_DEBUG "Dead loop on netdevice %s, fix it urgently!\n", dev->name);
return -1;
}
__get_cpu_var(netdev_rx_stat).cpu_collision++;
goto requeue;
}
}
{
/* And release queue */
spin_unlock(&dev->queue_lock);
if (!netif_queue_stopped(dev)) {
int ret;
[NET]: Add generic segmentation offload This patch adds the infrastructure for generic segmentation offload. The idea is to tap into the potential savings of TSO without hardware support by postponing the allocation of segmented skb's until just before the entry point into the NIC driver. The same structure can be used to support software IPv6 TSO, as well as UFO and segmentation offload for other relevant protocols, e.g., DCCP. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 17:57:17 +08:00
ret = dev_hard_start_xmit(skb, dev);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (ret == NETDEV_TX_OK) {
if (!nolock) {
[NET]: Add netif_tx_lock Various drivers use xmit_lock internally to synchronise with their transmission routines. They do so without setting xmit_lock_owner. This is fine as long as netpoll is not in use. With netpoll it is possible for deadlocks to occur if xmit_lock_owner isn't set. This is because if a printk occurs while xmit_lock is held and xmit_lock_owner is not set can cause netpoll to attempt to take xmit_lock recursively. While it is possible to resolve this by getting netpoll to use trylock, it is suboptimal because netpoll's sole objective is to maximise the chance of getting the printk out on the wire. So delaying or dropping the message is to be avoided as much as possible. So the only alternative is to always set xmit_lock_owner. The following patch does this by introducing the netif_tx_lock family of functions that take care of setting/unsetting xmit_lock_owner. I renamed xmit_lock to _xmit_lock to indicate that it should not be used directly. I didn't provide irq versions of the netif_tx_lock functions since xmit_lock is meant to be a BH-disabling lock. This is pretty much a straight text substitution except for a small bug fix in winbond. It currently uses netif_stop_queue/spin_unlock_wait to stop transmission. This is unsafe as an IRQ can potentially wake up the queue. So it is safer to use netif_tx_disable. The hamradio bits used spin_lock_irq but it is unnecessary as xmit_lock must never be taken in an IRQ handler. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-10 03:20:56 +08:00
netif_tx_unlock(dev);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
spin_lock(&dev->queue_lock);
return -1;
}
if (ret == NETDEV_TX_LOCKED && nolock) {
spin_lock(&dev->queue_lock);
goto collision;
}
}
/* NETDEV_TX_BUSY - we need to requeue */
/* Release the driver */
if (!nolock) {
[NET]: Add netif_tx_lock Various drivers use xmit_lock internally to synchronise with their transmission routines. They do so without setting xmit_lock_owner. This is fine as long as netpoll is not in use. With netpoll it is possible for deadlocks to occur if xmit_lock_owner isn't set. This is because if a printk occurs while xmit_lock is held and xmit_lock_owner is not set can cause netpoll to attempt to take xmit_lock recursively. While it is possible to resolve this by getting netpoll to use trylock, it is suboptimal because netpoll's sole objective is to maximise the chance of getting the printk out on the wire. So delaying or dropping the message is to be avoided as much as possible. So the only alternative is to always set xmit_lock_owner. The following patch does this by introducing the netif_tx_lock family of functions that take care of setting/unsetting xmit_lock_owner. I renamed xmit_lock to _xmit_lock to indicate that it should not be used directly. I didn't provide irq versions of the netif_tx_lock functions since xmit_lock is meant to be a BH-disabling lock. This is pretty much a straight text substitution except for a small bug fix in winbond. It currently uses netif_stop_queue/spin_unlock_wait to stop transmission. This is unsafe as an IRQ can potentially wake up the queue. So it is safer to use netif_tx_disable. The hamradio bits used spin_lock_irq but it is unnecessary as xmit_lock must never be taken in an IRQ handler. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-10 03:20:56 +08:00
netif_tx_unlock(dev);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
spin_lock(&dev->queue_lock);
q = dev->qdisc;
}
/* Device kicked us out :(
This is possible in three cases:
0. driver is locked
1. fastroute is enabled
2. device cannot determine busy state
before start of transmission (f.e. dialout)
3. device is buggy (ppp)
*/
requeue:
[NET]: Add generic segmentation offload This patch adds the infrastructure for generic segmentation offload. The idea is to tap into the potential savings of TSO without hardware support by postponing the allocation of segmented skb's until just before the entry point into the NIC driver. The same structure can be used to support software IPv6 TSO, as well as UFO and segmentation offload for other relevant protocols, e.g., DCCP. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 17:57:17 +08:00
if (skb->next)
dev->gso_skb = skb;
else
q->ops->requeue(skb, q);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
netif_schedule(dev);
return 1;
}
[PKT_SCHED]: netetm: trap infinite loop hange on qlen underflow Due to bugs in netem (fixed by later patches), it is possible to get qdisc qlen to go negative. If this happens the CPU ends up spinning forever in qdisc_run(). So add a BUG_ON() to trap it. Signed-off-by: Stephen Hemminger <shemminger@osdl.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-05-04 07:24:03 +08:00
BUG_ON((int) q->q.qlen < 0);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
return q->q.qlen;
}
[NET]: Prevent multiple qdisc runs Having two or more qdisc_run's contend against each other is bad because it can induce packet reordering if the packets have to be requeued. It appears that this is an unintended consequence of relinquinshing the queue lock while transmitting. That in turn is needed for devices that spend a lot of time in their transmit routine. There are no advantages to be had as devices with queues are inherently single-threaded (the loopback device is not but then it doesn't have a queue). Even if you were to add a queue to a parallel virtual device (e.g., bolt a tbf filter in front of an ipip tunnel device), you would still want to process the queue in sequence to ensure that the packets are ordered correctly. The solution here is to steal a bit from net_device to prevent this. BTW, as qdisc_restart is no longer used by anyone as a module inside the kernel (IIRC it used to with netif_wake_queue), I have not exported the new __qdisc_run function. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-20 14:57:59 +08:00
void __qdisc_run(struct net_device *dev)
{
[NET]: Prevent transmission after dev_deactivate The dev_deactivate function has bit-rotted since the introduction of lockless drivers. In particular, the spin_unlock_wait call at the end has no effect on the xmit routine of lockless drivers. With a little bit of work, we can make it much more useful by providing the guarantee that when it returns, no more calls to the xmit routine of the underlying driver will be made. The idea is simple. There are two entry points in to the xmit routine. The first comes from dev_queue_xmit. That one is easily stopped by using synchronize_rcu. This works because we set the qdisc to noop_qdisc before the synchronize_rcu call. That in turn causes all subsequent packets sent to dev_queue_xmit to be dropped. The synchronize_rcu call also ensures all outstanding calls leave their critical section. The other entry point is from qdisc_run. Since we now have a bit that indicates whether it's running, all we have to do is to wait until the bit is off. I've removed the loop to wait for __LINK_STATE_SCHED to clear. This is useless because netif_wake_queue can cause it to be set again. It is also harmless because we've disarmed qdisc_run. I've also removed the spin_unlock_wait on xmit_lock because its only purpose of making sure that all outstanding xmit_lock holders have exited is also given by dev_watchdog_down. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 17:28:18 +08:00
if (unlikely(dev->qdisc == &noop_qdisc))
goto out;
[NET]: Prevent multiple qdisc runs Having two or more qdisc_run's contend against each other is bad because it can induce packet reordering if the packets have to be requeued. It appears that this is an unintended consequence of relinquinshing the queue lock while transmitting. That in turn is needed for devices that spend a lot of time in their transmit routine. There are no advantages to be had as devices with queues are inherently single-threaded (the loopback device is not but then it doesn't have a queue). Even if you were to add a queue to a parallel virtual device (e.g., bolt a tbf filter in front of an ipip tunnel device), you would still want to process the queue in sequence to ensure that the packets are ordered correctly. The solution here is to steal a bit from net_device to prevent this. BTW, as qdisc_restart is no longer used by anyone as a module inside the kernel (IIRC it used to with netif_wake_queue), I have not exported the new __qdisc_run function. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-20 14:57:59 +08:00
while (qdisc_restart(dev) < 0 && !netif_queue_stopped(dev))
/* NOTHING */;
[NET]: Prevent transmission after dev_deactivate The dev_deactivate function has bit-rotted since the introduction of lockless drivers. In particular, the spin_unlock_wait call at the end has no effect on the xmit routine of lockless drivers. With a little bit of work, we can make it much more useful by providing the guarantee that when it returns, no more calls to the xmit routine of the underlying driver will be made. The idea is simple. There are two entry points in to the xmit routine. The first comes from dev_queue_xmit. That one is easily stopped by using synchronize_rcu. This works because we set the qdisc to noop_qdisc before the synchronize_rcu call. That in turn causes all subsequent packets sent to dev_queue_xmit to be dropped. The synchronize_rcu call also ensures all outstanding calls leave their critical section. The other entry point is from qdisc_run. Since we now have a bit that indicates whether it's running, all we have to do is to wait until the bit is off. I've removed the loop to wait for __LINK_STATE_SCHED to clear. This is useless because netif_wake_queue can cause it to be set again. It is also harmless because we've disarmed qdisc_run. I've also removed the spin_unlock_wait on xmit_lock because its only purpose of making sure that all outstanding xmit_lock holders have exited is also given by dev_watchdog_down. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 17:28:18 +08:00
out:
[NET]: Prevent multiple qdisc runs Having two or more qdisc_run's contend against each other is bad because it can induce packet reordering if the packets have to be requeued. It appears that this is an unintended consequence of relinquinshing the queue lock while transmitting. That in turn is needed for devices that spend a lot of time in their transmit routine. There are no advantages to be had as devices with queues are inherently single-threaded (the loopback device is not but then it doesn't have a queue). Even if you were to add a queue to a parallel virtual device (e.g., bolt a tbf filter in front of an ipip tunnel device), you would still want to process the queue in sequence to ensure that the packets are ordered correctly. The solution here is to steal a bit from net_device to prevent this. BTW, as qdisc_restart is no longer used by anyone as a module inside the kernel (IIRC it used to with netif_wake_queue), I have not exported the new __qdisc_run function. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-20 14:57:59 +08:00
clear_bit(__LINK_STATE_QDISC_RUNNING, &dev->state);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
static void dev_watchdog(unsigned long arg)
{
struct net_device *dev = (struct net_device *)arg;
[NET]: Add netif_tx_lock Various drivers use xmit_lock internally to synchronise with their transmission routines. They do so without setting xmit_lock_owner. This is fine as long as netpoll is not in use. With netpoll it is possible for deadlocks to occur if xmit_lock_owner isn't set. This is because if a printk occurs while xmit_lock is held and xmit_lock_owner is not set can cause netpoll to attempt to take xmit_lock recursively. While it is possible to resolve this by getting netpoll to use trylock, it is suboptimal because netpoll's sole objective is to maximise the chance of getting the printk out on the wire. So delaying or dropping the message is to be avoided as much as possible. So the only alternative is to always set xmit_lock_owner. The following patch does this by introducing the netif_tx_lock family of functions that take care of setting/unsetting xmit_lock_owner. I renamed xmit_lock to _xmit_lock to indicate that it should not be used directly. I didn't provide irq versions of the netif_tx_lock functions since xmit_lock is meant to be a BH-disabling lock. This is pretty much a straight text substitution except for a small bug fix in winbond. It currently uses netif_stop_queue/spin_unlock_wait to stop transmission. This is unsafe as an IRQ can potentially wake up the queue. So it is safer to use netif_tx_disable. The hamradio bits used spin_lock_irq but it is unnecessary as xmit_lock must never be taken in an IRQ handler. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-10 03:20:56 +08:00
netif_tx_lock(dev);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (dev->qdisc != &noop_qdisc) {
if (netif_device_present(dev) &&
netif_running(dev) &&
netif_carrier_ok(dev)) {
if (netif_queue_stopped(dev) &&
[PKT_SCHED]: Potential jiffy wrap bug in dev_watchdog(). There is a potential jiffy wraparound bug in the transmit watchdog that is easily avoided by using time_after(). Signed-off-by: Stephen Hemminger <shemminger@osdl.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-05-17 06:02:12 +08:00
time_after(jiffies, dev->trans_start + dev->watchdog_timeo)) {
printk(KERN_INFO "NETDEV WATCHDOG: %s: transmit timed out\n",
dev->name);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dev->tx_timeout(dev);
}
if (!mod_timer(&dev->watchdog_timer, jiffies + dev->watchdog_timeo))
dev_hold(dev);
}
}
[NET]: Add netif_tx_lock Various drivers use xmit_lock internally to synchronise with their transmission routines. They do so without setting xmit_lock_owner. This is fine as long as netpoll is not in use. With netpoll it is possible for deadlocks to occur if xmit_lock_owner isn't set. This is because if a printk occurs while xmit_lock is held and xmit_lock_owner is not set can cause netpoll to attempt to take xmit_lock recursively. While it is possible to resolve this by getting netpoll to use trylock, it is suboptimal because netpoll's sole objective is to maximise the chance of getting the printk out on the wire. So delaying or dropping the message is to be avoided as much as possible. So the only alternative is to always set xmit_lock_owner. The following patch does this by introducing the netif_tx_lock family of functions that take care of setting/unsetting xmit_lock_owner. I renamed xmit_lock to _xmit_lock to indicate that it should not be used directly. I didn't provide irq versions of the netif_tx_lock functions since xmit_lock is meant to be a BH-disabling lock. This is pretty much a straight text substitution except for a small bug fix in winbond. It currently uses netif_stop_queue/spin_unlock_wait to stop transmission. This is unsafe as an IRQ can potentially wake up the queue. So it is safer to use netif_tx_disable. The hamradio bits used spin_lock_irq but it is unnecessary as xmit_lock must never be taken in an IRQ handler. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-10 03:20:56 +08:00
netif_tx_unlock(dev);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dev_put(dev);
}
static void dev_watchdog_init(struct net_device *dev)
{
init_timer(&dev->watchdog_timer);
dev->watchdog_timer.data = (unsigned long)dev;
dev->watchdog_timer.function = dev_watchdog;
}
void __netdev_watchdog_up(struct net_device *dev)
{
if (dev->tx_timeout) {
if (dev->watchdog_timeo <= 0)
dev->watchdog_timeo = 5*HZ;
if (!mod_timer(&dev->watchdog_timer, jiffies + dev->watchdog_timeo))
dev_hold(dev);
}
}
static void dev_watchdog_up(struct net_device *dev)
{
__netdev_watchdog_up(dev);
}
static void dev_watchdog_down(struct net_device *dev)
{
[NET]: Add netif_tx_lock Various drivers use xmit_lock internally to synchronise with their transmission routines. They do so without setting xmit_lock_owner. This is fine as long as netpoll is not in use. With netpoll it is possible for deadlocks to occur if xmit_lock_owner isn't set. This is because if a printk occurs while xmit_lock is held and xmit_lock_owner is not set can cause netpoll to attempt to take xmit_lock recursively. While it is possible to resolve this by getting netpoll to use trylock, it is suboptimal because netpoll's sole objective is to maximise the chance of getting the printk out on the wire. So delaying or dropping the message is to be avoided as much as possible. So the only alternative is to always set xmit_lock_owner. The following patch does this by introducing the netif_tx_lock family of functions that take care of setting/unsetting xmit_lock_owner. I renamed xmit_lock to _xmit_lock to indicate that it should not be used directly. I didn't provide irq versions of the netif_tx_lock functions since xmit_lock is meant to be a BH-disabling lock. This is pretty much a straight text substitution except for a small bug fix in winbond. It currently uses netif_stop_queue/spin_unlock_wait to stop transmission. This is unsafe as an IRQ can potentially wake up the queue. So it is safer to use netif_tx_disable. The hamradio bits used spin_lock_irq but it is unnecessary as xmit_lock must never be taken in an IRQ handler. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-10 03:20:56 +08:00
netif_tx_lock_bh(dev);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (del_timer(&dev->watchdog_timer))
[NET]: dev_put/dev_hold cleanup Get rid of the old __dev_put macro that is just a hold over from pre 2.6 kernel. And turn dev_hold into an inline instead of a macro. Signed-off-by: Stephen Hemminger <shemminger@osdl.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-03-21 14:32:28 +08:00
dev_put(dev);
[NET]: Add netif_tx_lock Various drivers use xmit_lock internally to synchronise with their transmission routines. They do so without setting xmit_lock_owner. This is fine as long as netpoll is not in use. With netpoll it is possible for deadlocks to occur if xmit_lock_owner isn't set. This is because if a printk occurs while xmit_lock is held and xmit_lock_owner is not set can cause netpoll to attempt to take xmit_lock recursively. While it is possible to resolve this by getting netpoll to use trylock, it is suboptimal because netpoll's sole objective is to maximise the chance of getting the printk out on the wire. So delaying or dropping the message is to be avoided as much as possible. So the only alternative is to always set xmit_lock_owner. The following patch does this by introducing the netif_tx_lock family of functions that take care of setting/unsetting xmit_lock_owner. I renamed xmit_lock to _xmit_lock to indicate that it should not be used directly. I didn't provide irq versions of the netif_tx_lock functions since xmit_lock is meant to be a BH-disabling lock. This is pretty much a straight text substitution except for a small bug fix in winbond. It currently uses netif_stop_queue/spin_unlock_wait to stop transmission. This is unsafe as an IRQ can potentially wake up the queue. So it is safer to use netif_tx_disable. The hamradio bits used spin_lock_irq but it is unnecessary as xmit_lock must never be taken in an IRQ handler. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-10 03:20:56 +08:00
netif_tx_unlock_bh(dev);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[NET]: Deinline netif_carrier_{on,off}(). # grep -r 'netif_carrier_o[nf]' linux-2.6.12 | wc -l 246 # size vmlinux.org vmlinux.carrier text data bss dec hex filename 4339634 1054414 259296 5653344 564360 vmlinux.org 4337710 1054414 259296 5651420 563bdc vmlinux.carrier And this ain't an allyesconfig kernel! Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-12 06:32:53 +08:00
void netif_carrier_on(struct net_device *dev)
{
if (test_and_clear_bit(__LINK_STATE_NOCARRIER, &dev->state))
linkwatch_fire_event(dev);
if (netif_running(dev))
__netdev_watchdog_up(dev);
}
void netif_carrier_off(struct net_device *dev)
{
if (!test_and_set_bit(__LINK_STATE_NOCARRIER, &dev->state))
linkwatch_fire_event(dev);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/* "NOOP" scheduler: the best scheduler, recommended for all interfaces
under all circumstances. It is difficult to invent anything faster or
cheaper.
*/
[PKT_SCHED]: noop/noqueue qdisc style cleanups Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:59:08 +08:00
static int noop_enqueue(struct sk_buff *skb, struct Qdisc * qdisc)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
kfree_skb(skb);
return NET_XMIT_CN;
}
[PKT_SCHED]: noop/noqueue qdisc style cleanups Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:59:08 +08:00
static struct sk_buff *noop_dequeue(struct Qdisc * qdisc)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
return NULL;
}
[PKT_SCHED]: noop/noqueue qdisc style cleanups Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:59:08 +08:00
static int noop_requeue(struct sk_buff *skb, struct Qdisc* qdisc)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
if (net_ratelimit())
[PKT_SCHED]: noop/noqueue qdisc style cleanups Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:59:08 +08:00
printk(KERN_DEBUG "%s deferred output. It is buggy.\n",
skb->dev->name);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
kfree_skb(skb);
return NET_XMIT_CN;
}
struct Qdisc_ops noop_qdisc_ops = {
.id = "noop",
.priv_size = 0,
.enqueue = noop_enqueue,
.dequeue = noop_dequeue,
.requeue = noop_requeue,
.owner = THIS_MODULE,
};
struct Qdisc noop_qdisc = {
.enqueue = noop_enqueue,
.dequeue = noop_dequeue,
.flags = TCQ_F_BUILTIN,
.ops = &noop_qdisc_ops,
.list = LIST_HEAD_INIT(noop_qdisc.list),
};
static struct Qdisc_ops noqueue_qdisc_ops = {
.id = "noqueue",
.priv_size = 0,
.enqueue = noop_enqueue,
.dequeue = noop_dequeue,
.requeue = noop_requeue,
.owner = THIS_MODULE,
};
static struct Qdisc noqueue_qdisc = {
.enqueue = NULL,
.dequeue = noop_dequeue,
.flags = TCQ_F_BUILTIN,
.ops = &noqueue_qdisc_ops,
.list = LIST_HEAD_INIT(noqueue_qdisc.list),
};
static const u8 prio2band[TC_PRIO_MAX+1] =
{ 1, 2, 2, 2, 1, 2, 0, 0 , 1, 1, 1, 1, 1, 1, 1, 1 };
/* 3-band FIFO queue: old style, but should be a bit faster than
generic prio+fifo combination.
*/
[PKT_SCHED]: Cleanup pfifo_fast qdisc and remove unnecessary code Removes the skb trimming code which is not needed since we never touch the skb upon failure. Removes unnecessary initializers, and simplifies the code a bit. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:53 +08:00
#define PFIFO_FAST_BANDS 3
[PKT_SCHED]: Add and use prio2list() in the pfifo_fast qdisc prio2list() returns the relevant sk_buff_head for the band specified by the priority for a given skb. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:35 +08:00
static inline struct sk_buff_head *prio2list(struct sk_buff *skb,
struct Qdisc *qdisc)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
struct sk_buff_head *list = qdisc_priv(qdisc);
[PKT_SCHED]: Add and use prio2list() in the pfifo_fast qdisc prio2list() returns the relevant sk_buff_head for the band specified by the priority for a given skb. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:35 +08:00
return list + prio2band[skb->priority & TC_PRIO_MAX];
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[PKT_SCHED]: Cleanup pfifo_fast qdisc and remove unnecessary code Removes the skb trimming code which is not needed since we never touch the skb upon failure. Removes unnecessary initializers, and simplifies the code a bit. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:53 +08:00
static int pfifo_fast_enqueue(struct sk_buff *skb, struct Qdisc* qdisc)
[PKT_SCHED]: Add and use prio2list() in the pfifo_fast qdisc prio2list() returns the relevant sk_buff_head for the band specified by the priority for a given skb. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:35 +08:00
{
struct sk_buff_head *list = prio2list(skb, qdisc);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[PKT_SCHED]: Transform pfifo_fast to use generic queue management interface Gives pfifo_fast a byte based backlog. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:15 +08:00
if (skb_queue_len(list) < qdisc->dev->tx_queue_len) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
qdisc->q.qlen++;
[PKT_SCHED]: Transform pfifo_fast to use generic queue management interface Gives pfifo_fast a byte based backlog. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:15 +08:00
return __qdisc_enqueue_tail(skb, qdisc, list);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[PKT_SCHED]: Transform pfifo_fast to use generic queue management interface Gives pfifo_fast a byte based backlog. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:15 +08:00
return qdisc_drop(skb, qdisc);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[PKT_SCHED]: Cleanup pfifo_fast qdisc and remove unnecessary code Removes the skb trimming code which is not needed since we never touch the skb upon failure. Removes unnecessary initializers, and simplifies the code a bit. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:53 +08:00
static struct sk_buff *pfifo_fast_dequeue(struct Qdisc* qdisc)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
int prio;
struct sk_buff_head *list = qdisc_priv(qdisc);
[PKT_SCHED]: Reduce branch mispredictions in pfifo_fast_dequeue The current call to __qdisc_dequeue_head leads to a branch misprediction for every loop iteration, the fact that the most common priority is 2 makes this even worse. This issue has been brought up by Eric Dumazet <dada1@cosmosbay.com> but unlike his solution which was to manually unroll the loop, this approach preserves the possibility to increase the number of bands at compile time. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-19 04:30:53 +08:00
for (prio = 0; prio < PFIFO_FAST_BANDS; prio++) {
if (!skb_queue_empty(list + prio)) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
qdisc->q.qlen--;
[PKT_SCHED]: Reduce branch mispredictions in pfifo_fast_dequeue The current call to __qdisc_dequeue_head leads to a branch misprediction for every loop iteration, the fact that the most common priority is 2 makes this even worse. This issue has been brought up by Eric Dumazet <dada1@cosmosbay.com> but unlike his solution which was to manually unroll the loop, this approach preserves the possibility to increase the number of bands at compile time. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-19 04:30:53 +08:00
return __qdisc_dequeue_head(qdisc, list + prio);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
}
[PKT_SCHED]: Cleanup pfifo_fast qdisc and remove unnecessary code Removes the skb trimming code which is not needed since we never touch the skb upon failure. Removes unnecessary initializers, and simplifies the code a bit. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:53 +08:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
return NULL;
}
[PKT_SCHED]: Cleanup pfifo_fast qdisc and remove unnecessary code Removes the skb trimming code which is not needed since we never touch the skb upon failure. Removes unnecessary initializers, and simplifies the code a bit. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:53 +08:00
static int pfifo_fast_requeue(struct sk_buff *skb, struct Qdisc* qdisc)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
qdisc->q.qlen++;
[PKT_SCHED]: Add and use prio2list() in the pfifo_fast qdisc prio2list() returns the relevant sk_buff_head for the band specified by the priority for a given skb. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:35 +08:00
return __qdisc_requeue(skb, qdisc, prio2list(skb, qdisc));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[PKT_SCHED]: Cleanup pfifo_fast qdisc and remove unnecessary code Removes the skb trimming code which is not needed since we never touch the skb upon failure. Removes unnecessary initializers, and simplifies the code a bit. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:53 +08:00
static void pfifo_fast_reset(struct Qdisc* qdisc)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
int prio;
struct sk_buff_head *list = qdisc_priv(qdisc);
[PKT_SCHED]: Cleanup pfifo_fast qdisc and remove unnecessary code Removes the skb trimming code which is not needed since we never touch the skb upon failure. Removes unnecessary initializers, and simplifies the code a bit. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:53 +08:00
for (prio = 0; prio < PFIFO_FAST_BANDS; prio++)
[PKT_SCHED]: Transform pfifo_fast to use generic queue management interface Gives pfifo_fast a byte based backlog. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:15 +08:00
__qdisc_reset_queue(qdisc, list + prio);
qdisc->qstats.backlog = 0;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
qdisc->q.qlen = 0;
}
static int pfifo_fast_dump(struct Qdisc *qdisc, struct sk_buff *skb)
{
[PKT_SCHED]: Cleanup pfifo_fast qdisc and remove unnecessary code Removes the skb trimming code which is not needed since we never touch the skb upon failure. Removes unnecessary initializers, and simplifies the code a bit. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:53 +08:00
struct tc_prio_qopt opt = { .bands = PFIFO_FAST_BANDS };
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
memcpy(&opt.priomap, prio2band, TC_PRIO_MAX+1);
RTA_PUT(skb, TCA_OPTIONS, sizeof(opt), &opt);
return skb->len;
rtattr_failure:
return -1;
}
static int pfifo_fast_init(struct Qdisc *qdisc, struct rtattr *opt)
{
[PKT_SCHED]: Cleanup pfifo_fast qdisc and remove unnecessary code Removes the skb trimming code which is not needed since we never touch the skb upon failure. Removes unnecessary initializers, and simplifies the code a bit. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:53 +08:00
int prio;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
struct sk_buff_head *list = qdisc_priv(qdisc);
[PKT_SCHED]: Cleanup pfifo_fast qdisc and remove unnecessary code Removes the skb trimming code which is not needed since we never touch the skb upon failure. Removes unnecessary initializers, and simplifies the code a bit. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:53 +08:00
for (prio = 0; prio < PFIFO_FAST_BANDS; prio++)
skb_queue_head_init(list + prio);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
return 0;
}
static struct Qdisc_ops pfifo_fast_ops = {
.id = "pfifo_fast",
[PKT_SCHED]: Cleanup pfifo_fast qdisc and remove unnecessary code Removes the skb trimming code which is not needed since we never touch the skb upon failure. Removes unnecessary initializers, and simplifies the code a bit. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-06-19 13:58:53 +08:00
.priv_size = PFIFO_FAST_BANDS * sizeof(struct sk_buff_head),
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
.enqueue = pfifo_fast_enqueue,
.dequeue = pfifo_fast_dequeue,
.requeue = pfifo_fast_requeue,
.init = pfifo_fast_init,
.reset = pfifo_fast_reset,
.dump = pfifo_fast_dump,
.owner = THIS_MODULE,
};
[PKT_SCHED]: Cleanup qdisc creation and alignment macros Adds qdisc_alloc() to share code between qdisc_create() and qdisc_create_dflt(). Hides the qdisc alignment behind macros and makes use of them. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-06 05:15:09 +08:00
struct Qdisc *qdisc_alloc(struct net_device *dev, struct Qdisc_ops *ops)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
void *p;
struct Qdisc *sch;
[PKT_SCHED]: Cleanup qdisc creation and alignment macros Adds qdisc_alloc() to share code between qdisc_create() and qdisc_create_dflt(). Hides the qdisc alignment behind macros and makes use of them. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-06 05:15:09 +08:00
unsigned int size;
int err = -ENOBUFS;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/* ensure that the Qdisc and the private data are 32-byte aligned */
[PKT_SCHED]: Cleanup qdisc creation and alignment macros Adds qdisc_alloc() to share code between qdisc_create() and qdisc_create_dflt(). Hides the qdisc alignment behind macros and makes use of them. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-06 05:15:09 +08:00
size = QDISC_ALIGN(sizeof(*sch));
size += ops->priv_size + (QDISC_ALIGNTO - 1);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[NET]: Conversions from kmalloc+memset to k(z|c)alloc. Signed-off-by: Panagiotis Issaris <takis@issaris.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-07-22 05:51:30 +08:00
p = kzalloc(size, GFP_KERNEL);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (!p)
[PKT_SCHED]: Cleanup qdisc creation and alignment macros Adds qdisc_alloc() to share code between qdisc_create() and qdisc_create_dflt(). Hides the qdisc alignment behind macros and makes use of them. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-06 05:15:09 +08:00
goto errout;
sch = (struct Qdisc *) QDISC_ALIGN((unsigned long) p);
sch->padded = (char *) sch - (char *) p;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
INIT_LIST_HEAD(&sch->list);
skb_queue_head_init(&sch->q);
sch->ops = ops;
sch->enqueue = ops->enqueue;
sch->dequeue = ops->dequeue;
sch->dev = dev;
dev_hold(dev);
sch->stats_lock = &dev->queue_lock;
atomic_set(&sch->refcnt, 1);
[PKT_SCHED]: Cleanup qdisc creation and alignment macros Adds qdisc_alloc() to share code between qdisc_create() and qdisc_create_dflt(). Hides the qdisc alignment behind macros and makes use of them. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-06 05:15:09 +08:00
return sch;
errout:
return ERR_PTR(-err);
}
[NET_SCHED]: Set parent classid in default qdiscs Set parent classids in default qdiscs to allow walking up the tree from outside the qdiscs. This is needed by the next patch. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-11-30 09:35:18 +08:00
struct Qdisc * qdisc_create_dflt(struct net_device *dev, struct Qdisc_ops *ops,
unsigned int parentid)
[PKT_SCHED]: Cleanup qdisc creation and alignment macros Adds qdisc_alloc() to share code between qdisc_create() and qdisc_create_dflt(). Hides the qdisc alignment behind macros and makes use of them. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-06 05:15:09 +08:00
{
struct Qdisc *sch;
sch = qdisc_alloc(dev, ops);
if (IS_ERR(sch))
goto errout;
[NET_SCHED]: Set parent classid in default qdiscs Set parent classids in default qdiscs to allow walking up the tree from outside the qdiscs. This is needed by the next patch. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-11-30 09:35:18 +08:00
sch->parent = parentid;
[PKT_SCHED]: Cleanup qdisc creation and alignment macros Adds qdisc_alloc() to share code between qdisc_create() and qdisc_create_dflt(). Hides the qdisc alignment behind macros and makes use of them. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-06 05:15:09 +08:00
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (!ops->init || ops->init(sch, NULL) == 0)
return sch;
[PKT_SCHED]: Fix missing qdisc_destroy() in qdisc_create_dflt() qdisc_create_dflt() is missing to destroy the newly allocated default qdisc if the initialization fails resulting in leaks of all kinds. The only caller in mainline which may trigger this bug is sch_tbf.c in tbf_create_dflt_qdisc(). Note: qdisc_create_dflt() doesn't fulfill the official locking requirements of qdisc_destroy() but since the qdisc could never be seen by the outside world this doesn't matter and it can stay as-is until the locking of pkt_sched is cleaned up. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-24 01:12:44 +08:00
qdisc_destroy(sch);
[PKT_SCHED]: Cleanup qdisc creation and alignment macros Adds qdisc_alloc() to share code between qdisc_create() and qdisc_create_dflt(). Hides the qdisc alignment behind macros and makes use of them. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-07-06 05:15:09 +08:00
errout:
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
return NULL;
}
/* Under dev->queue_lock and BH! */
void qdisc_reset(struct Qdisc *qdisc)
{
struct Qdisc_ops *ops = qdisc->ops;
if (ops->reset)
ops->reset(qdisc);
}
/* this is the rcu callback function to clean up a qdisc when there
* are no further references to it */
static void __qdisc_destroy(struct rcu_head *head)
{
struct Qdisc *qdisc = container_of(head, struct Qdisc, q_rcu);
kfree((char *) qdisc - qdisc->padded);
}
/* Under dev->queue_lock and BH! */
void qdisc_destroy(struct Qdisc *qdisc)
{
[NET_SCHED]: Fix fallout from dev->qdisc RCU change The move of qdisc destruction to a rcu callback broke locking in the entire qdisc layer by invalidating previously valid assumptions about the context in which changes to the qdisc tree occur. The two assumptions were: - since changes only happen in process context, read_lock doesn't need bottem half protection. Now invalid since destruction of inner qdiscs, classifiers, actions and estimators happens in the RCU callback unless they're manually deleted, resulting in dead-locks when read_lock in process context is interrupted by write_lock_bh in bottem half context. - since changes only happen under the RTNL, no additional locking is necessary for data not used during packet processing (f.e. u32_list). Again, since destruction now happens in the RCU callback, this assumption is not valid anymore, causing races while using this data, which can result in corruption or use-after-free. Instead of "fixing" this by disabling bottem halfs everywhere and adding new locks/refcounting, this patch makes these assumptions valid again by moving destruction back to process context. Since only the dev->qdisc pointer is protected by RCU, but ->enqueue and the qdisc tree are still protected by dev->qdisc_lock, destruction of the tree can be performed immediately and only the final free needs to happen in the rcu callback to make sure dev_queue_xmit doesn't access already freed memory. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 07:45:45 +08:00
struct Qdisc_ops *ops = qdisc->ops;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (qdisc->flags & TCQ_F_BUILTIN ||
[NET_SCHED]: Fix fallout from dev->qdisc RCU change The move of qdisc destruction to a rcu callback broke locking in the entire qdisc layer by invalidating previously valid assumptions about the context in which changes to the qdisc tree occur. The two assumptions were: - since changes only happen in process context, read_lock doesn't need bottem half protection. Now invalid since destruction of inner qdiscs, classifiers, actions and estimators happens in the RCU callback unless they're manually deleted, resulting in dead-locks when read_lock in process context is interrupted by write_lock_bh in bottem half context. - since changes only happen under the RTNL, no additional locking is necessary for data not used during packet processing (f.e. u32_list). Again, since destruction now happens in the RCU callback, this assumption is not valid anymore, causing races while using this data, which can result in corruption or use-after-free. Instead of "fixing" this by disabling bottem halfs everywhere and adding new locks/refcounting, this patch makes these assumptions valid again by moving destruction back to process context. Since only the dev->qdisc pointer is protected by RCU, but ->enqueue and the qdisc tree are still protected by dev->qdisc_lock, destruction of the tree can be performed immediately and only the final free needs to happen in the rcu callback to make sure dev_queue_xmit doesn't access already freed memory. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 07:45:45 +08:00
!atomic_dec_and_test(&qdisc->refcnt))
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
return;
[NET_SCHED]: Fix fallout from dev->qdisc RCU change The move of qdisc destruction to a rcu callback broke locking in the entire qdisc layer by invalidating previously valid assumptions about the context in which changes to the qdisc tree occur. The two assumptions were: - since changes only happen in process context, read_lock doesn't need bottem half protection. Now invalid since destruction of inner qdiscs, classifiers, actions and estimators happens in the RCU callback unless they're manually deleted, resulting in dead-locks when read_lock in process context is interrupted by write_lock_bh in bottem half context. - since changes only happen under the RTNL, no additional locking is necessary for data not used during packet processing (f.e. u32_list). Again, since destruction now happens in the RCU callback, this assumption is not valid anymore, causing races while using this data, which can result in corruption or use-after-free. Instead of "fixing" this by disabling bottem halfs everywhere and adding new locks/refcounting, this patch makes these assumptions valid again by moving destruction back to process context. Since only the dev->qdisc pointer is protected by RCU, but ->enqueue and the qdisc tree are still protected by dev->qdisc_lock, destruction of the tree can be performed immediately and only the final free needs to happen in the rcu callback to make sure dev_queue_xmit doesn't access already freed memory. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 07:45:45 +08:00
list_del(&qdisc->list);
#ifdef CONFIG_NET_ESTIMATOR
gen_kill_estimator(&qdisc->bstats, &qdisc->rate_est);
#endif
if (ops->reset)
ops->reset(qdisc);
if (ops->destroy)
ops->destroy(qdisc);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[NET_SCHED]: Fix fallout from dev->qdisc RCU change The move of qdisc destruction to a rcu callback broke locking in the entire qdisc layer by invalidating previously valid assumptions about the context in which changes to the qdisc tree occur. The two assumptions were: - since changes only happen in process context, read_lock doesn't need bottem half protection. Now invalid since destruction of inner qdiscs, classifiers, actions and estimators happens in the RCU callback unless they're manually deleted, resulting in dead-locks when read_lock in process context is interrupted by write_lock_bh in bottem half context. - since changes only happen under the RTNL, no additional locking is necessary for data not used during packet processing (f.e. u32_list). Again, since destruction now happens in the RCU callback, this assumption is not valid anymore, causing races while using this data, which can result in corruption or use-after-free. Instead of "fixing" this by disabling bottem halfs everywhere and adding new locks/refcounting, this patch makes these assumptions valid again by moving destruction back to process context. Since only the dev->qdisc pointer is protected by RCU, but ->enqueue and the qdisc tree are still protected by dev->qdisc_lock, destruction of the tree can be performed immediately and only the final free needs to happen in the rcu callback to make sure dev_queue_xmit doesn't access already freed memory. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 07:45:45 +08:00
module_put(ops->owner);
dev_put(qdisc->dev);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
call_rcu(&qdisc->q_rcu, __qdisc_destroy);
}
void dev_activate(struct net_device *dev)
{
/* No queueing discipline is attached to device;
create default one i.e. pfifo_fast for devices,
which need queueing and noqueue_qdisc for
virtual interfaces
*/
if (dev->qdisc_sleeping == &noop_qdisc) {
struct Qdisc *qdisc;
if (dev->tx_queue_len) {
[NET_SCHED]: Set parent classid in default qdiscs Set parent classids in default qdiscs to allow walking up the tree from outside the qdiscs. This is needed by the next patch. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-11-30 09:35:18 +08:00
qdisc = qdisc_create_dflt(dev, &pfifo_fast_ops,
TC_H_ROOT);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (qdisc == NULL) {
printk(KERN_INFO "%s: activation failed\n", dev->name);
return;
}
[NET_SCHED]: Fix fallout from dev->qdisc RCU change The move of qdisc destruction to a rcu callback broke locking in the entire qdisc layer by invalidating previously valid assumptions about the context in which changes to the qdisc tree occur. The two assumptions were: - since changes only happen in process context, read_lock doesn't need bottem half protection. Now invalid since destruction of inner qdiscs, classifiers, actions and estimators happens in the RCU callback unless they're manually deleted, resulting in dead-locks when read_lock in process context is interrupted by write_lock_bh in bottem half context. - since changes only happen under the RTNL, no additional locking is necessary for data not used during packet processing (f.e. u32_list). Again, since destruction now happens in the RCU callback, this assumption is not valid anymore, causing races while using this data, which can result in corruption or use-after-free. Instead of "fixing" this by disabling bottem halfs everywhere and adding new locks/refcounting, this patch makes these assumptions valid again by moving destruction back to process context. Since only the dev->qdisc pointer is protected by RCU, but ->enqueue and the qdisc tree are still protected by dev->qdisc_lock, destruction of the tree can be performed immediately and only the final free needs to happen in the rcu callback to make sure dev_queue_xmit doesn't access already freed memory. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 07:45:45 +08:00
write_lock(&qdisc_tree_lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
list_add_tail(&qdisc->list, &dev->qdisc_list);
[NET_SCHED]: Fix fallout from dev->qdisc RCU change The move of qdisc destruction to a rcu callback broke locking in the entire qdisc layer by invalidating previously valid assumptions about the context in which changes to the qdisc tree occur. The two assumptions were: - since changes only happen in process context, read_lock doesn't need bottem half protection. Now invalid since destruction of inner qdiscs, classifiers, actions and estimators happens in the RCU callback unless they're manually deleted, resulting in dead-locks when read_lock in process context is interrupted by write_lock_bh in bottem half context. - since changes only happen under the RTNL, no additional locking is necessary for data not used during packet processing (f.e. u32_list). Again, since destruction now happens in the RCU callback, this assumption is not valid anymore, causing races while using this data, which can result in corruption or use-after-free. Instead of "fixing" this by disabling bottem halfs everywhere and adding new locks/refcounting, this patch makes these assumptions valid again by moving destruction back to process context. Since only the dev->qdisc pointer is protected by RCU, but ->enqueue and the qdisc tree are still protected by dev->qdisc_lock, destruction of the tree can be performed immediately and only the final free needs to happen in the rcu callback to make sure dev_queue_xmit doesn't access already freed memory. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 07:45:45 +08:00
write_unlock(&qdisc_tree_lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
} else {
qdisc = &noqueue_qdisc;
}
[NET_SCHED]: Fix fallout from dev->qdisc RCU change The move of qdisc destruction to a rcu callback broke locking in the entire qdisc layer by invalidating previously valid assumptions about the context in which changes to the qdisc tree occur. The two assumptions were: - since changes only happen in process context, read_lock doesn't need bottem half protection. Now invalid since destruction of inner qdiscs, classifiers, actions and estimators happens in the RCU callback unless they're manually deleted, resulting in dead-locks when read_lock in process context is interrupted by write_lock_bh in bottem half context. - since changes only happen under the RTNL, no additional locking is necessary for data not used during packet processing (f.e. u32_list). Again, since destruction now happens in the RCU callback, this assumption is not valid anymore, causing races while using this data, which can result in corruption or use-after-free. Instead of "fixing" this by disabling bottem halfs everywhere and adding new locks/refcounting, this patch makes these assumptions valid again by moving destruction back to process context. Since only the dev->qdisc pointer is protected by RCU, but ->enqueue and the qdisc tree are still protected by dev->qdisc_lock, destruction of the tree can be performed immediately and only the final free needs to happen in the rcu callback to make sure dev_queue_xmit doesn't access already freed memory. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 07:45:45 +08:00
write_lock(&qdisc_tree_lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
dev->qdisc_sleeping = qdisc;
[NET_SCHED]: Fix fallout from dev->qdisc RCU change The move of qdisc destruction to a rcu callback broke locking in the entire qdisc layer by invalidating previously valid assumptions about the context in which changes to the qdisc tree occur. The two assumptions were: - since changes only happen in process context, read_lock doesn't need bottem half protection. Now invalid since destruction of inner qdiscs, classifiers, actions and estimators happens in the RCU callback unless they're manually deleted, resulting in dead-locks when read_lock in process context is interrupted by write_lock_bh in bottem half context. - since changes only happen under the RTNL, no additional locking is necessary for data not used during packet processing (f.e. u32_list). Again, since destruction now happens in the RCU callback, this assumption is not valid anymore, causing races while using this data, which can result in corruption or use-after-free. Instead of "fixing" this by disabling bottem halfs everywhere and adding new locks/refcounting, this patch makes these assumptions valid again by moving destruction back to process context. Since only the dev->qdisc pointer is protected by RCU, but ->enqueue and the qdisc tree are still protected by dev->qdisc_lock, destruction of the tree can be performed immediately and only the final free needs to happen in the rcu callback to make sure dev_queue_xmit doesn't access already freed memory. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-09-28 07:45:45 +08:00
write_unlock(&qdisc_tree_lock);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[NET]: Disable queueing when carrier is lost. Some network drivers call netif_stop_queue() when detecting loss of carrier. This leads to packets being queued up at the qdisc level for an unbound period of time. In order to prevent this effect, the core networking stack will now cease to queue packets for any device, that is operationally down (i.e. the queue is flushed and disabled). Signed-off-by: Tommy S. Christensen <tommy.christensen@tpack.net> Acked-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2005-05-04 07:18:52 +08:00
if (!netif_carrier_ok(dev))
/* Delay activation until next carrier-on event */
return;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
spin_lock_bh(&dev->queue_lock);
rcu_assign_pointer(dev->qdisc, dev->qdisc_sleeping);
if (dev->qdisc != &noqueue_qdisc) {
dev->trans_start = jiffies;
dev_watchdog_up(dev);
}
spin_unlock_bh(&dev->queue_lock);
}
void dev_deactivate(struct net_device *dev)
{
struct Qdisc *qdisc;
spin_lock_bh(&dev->queue_lock);
qdisc = dev->qdisc;
dev->qdisc = &noop_qdisc;
qdisc_reset(qdisc);
spin_unlock_bh(&dev->queue_lock);
dev_watchdog_down(dev);
[NET]: Prevent transmission after dev_deactivate The dev_deactivate function has bit-rotted since the introduction of lockless drivers. In particular, the spin_unlock_wait call at the end has no effect on the xmit routine of lockless drivers. With a little bit of work, we can make it much more useful by providing the guarantee that when it returns, no more calls to the xmit routine of the underlying driver will be made. The idea is simple. There are two entry points in to the xmit routine. The first comes from dev_queue_xmit. That one is easily stopped by using synchronize_rcu. This works because we set the qdisc to noop_qdisc before the synchronize_rcu call. That in turn causes all subsequent packets sent to dev_queue_xmit to be dropped. The synchronize_rcu call also ensures all outstanding calls leave their critical section. The other entry point is from qdisc_run. Since we now have a bit that indicates whether it's running, all we have to do is to wait until the bit is off. I've removed the loop to wait for __LINK_STATE_SCHED to clear. This is useless because netif_wake_queue can cause it to be set again. It is also harmless because we've disarmed qdisc_run. I've also removed the spin_unlock_wait on xmit_lock because its only purpose of making sure that all outstanding xmit_lock holders have exited is also given by dev_watchdog_down. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 17:28:18 +08:00
/* Wait for outstanding dev_queue_xmit calls. */
synchronize_rcu();
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[NET]: Prevent transmission after dev_deactivate The dev_deactivate function has bit-rotted since the introduction of lockless drivers. In particular, the spin_unlock_wait call at the end has no effect on the xmit routine of lockless drivers. With a little bit of work, we can make it much more useful by providing the guarantee that when it returns, no more calls to the xmit routine of the underlying driver will be made. The idea is simple. There are two entry points in to the xmit routine. The first comes from dev_queue_xmit. That one is easily stopped by using synchronize_rcu. This works because we set the qdisc to noop_qdisc before the synchronize_rcu call. That in turn causes all subsequent packets sent to dev_queue_xmit to be dropped. The synchronize_rcu call also ensures all outstanding calls leave their critical section. The other entry point is from qdisc_run. Since we now have a bit that indicates whether it's running, all we have to do is to wait until the bit is off. I've removed the loop to wait for __LINK_STATE_SCHED to clear. This is useless because netif_wake_queue can cause it to be set again. It is also harmless because we've disarmed qdisc_run. I've also removed the spin_unlock_wait on xmit_lock because its only purpose of making sure that all outstanding xmit_lock holders have exited is also given by dev_watchdog_down. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 17:28:18 +08:00
/* Wait for outstanding qdisc_run calls. */
while (test_bit(__LINK_STATE_QDISC_RUNNING, &dev->state))
yield();
[NET]: Add generic segmentation offload This patch adds the infrastructure for generic segmentation offload. The idea is to tap into the potential savings of TSO without hardware support by postponing the allocation of segmented skb's until just before the entry point into the NIC driver. The same structure can be used to support software IPv6 TSO, as well as UFO and segmentation offload for other relevant protocols, e.g., DCCP. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-22 17:57:17 +08:00
if (dev->gso_skb) {
kfree_skb(dev->gso_skb);
dev->gso_skb = NULL;
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
void dev_init_scheduler(struct net_device *dev)
{
qdisc_lock_tree(dev);
dev->qdisc = &noop_qdisc;
dev->qdisc_sleeping = &noop_qdisc;
INIT_LIST_HEAD(&dev->qdisc_list);
qdisc_unlock_tree(dev);
dev_watchdog_init(dev);
}
void dev_shutdown(struct net_device *dev)
{
struct Qdisc *qdisc;
qdisc_lock_tree(dev);
qdisc = dev->qdisc_sleeping;
dev->qdisc = &noop_qdisc;
dev->qdisc_sleeping = &noop_qdisc;
qdisc_destroy(qdisc);
#if defined(CONFIG_NET_SCH_INGRESS) || defined(CONFIG_NET_SCH_INGRESS_MODULE)
if ((qdisc = dev->qdisc_ingress) != NULL) {
dev->qdisc_ingress = NULL;
qdisc_destroy(qdisc);
}
#endif
BUG_TRAP(!timer_pending(&dev->watchdog_timer));
qdisc_unlock_tree(dev);
}
[NET]: Deinline netif_carrier_{on,off}(). # grep -r 'netif_carrier_o[nf]' linux-2.6.12 | wc -l 246 # size vmlinux.org vmlinux.carrier text data bss dec hex filename 4339634 1054414 259296 5653344 564360 vmlinux.org 4337710 1054414 259296 5651420 563bdc vmlinux.carrier And this ain't an allyesconfig kernel! Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-12 06:32:53 +08:00
EXPORT_SYMBOL(netif_carrier_on);
EXPORT_SYMBOL(netif_carrier_off);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
EXPORT_SYMBOL(noop_qdisc);
EXPORT_SYMBOL(qdisc_create_dflt);
EXPORT_SYMBOL(qdisc_destroy);
EXPORT_SYMBOL(qdisc_reset);
EXPORT_SYMBOL(qdisc_lock_tree);
EXPORT_SYMBOL(qdisc_unlock_tree);
Reference in New Issue Copy Permalink