linux/net/core/dst.c
Eric Dumazet 86bba269d0 [PATCH] NET : convert IP route cache garbage collection from softirq processing to a workqueue
When the periodic IP route cache flush is done (every 600 seconds on
default configuration), some hosts suffer a lot and eventually trigger
the "soft lockup" message.

dst_run_gc() is doing a scan of a possibly huge list of dst_entries,
eventually freeing some (less than 1%) of them, while holding the
dst_lock spinlock for the whole scan.

Then it rearms a timer to redo the full thing 1/10 s later...
The slowdown can last one minute or so, depending on how active are
the tcp sessions.

This second version of the patch converts the processing from a softirq
based one to a workqueue.

Even if the list of entries in garbage_list is huge, host is still
responsive to softirqs and can make progress.

Instead of resetting gc timer to 0.1 second if one entry was freed in a
gc run, we do this if more than 10% of entries were freed.

Before patch :

Aug 16 06:21:37 SRV1 kernel: BUG: soft lockup detected on CPU#0!
Aug 16 06:21:37 SRV1 kernel:
Aug 16 06:21:37 SRV1 kernel: Call Trace:
Aug 16 06:21:37 SRV1 kernel:  <IRQ>  [<ffffffff802286f0>] wake_up_process+0x10/0x20
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff80251e09>] softlockup_tick+0xe9/0x110
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803cd380>] dst_run_gc+0x0/0x140
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff802376f3>] run_local_timers+0x13/0x20
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff802379c7>] update_process_times+0x57/0x90
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff80216034>] smp_local_timer_interrupt+0x34/0x60
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff802165cc>] smp_apic_timer_interrupt+0x5c/0x80
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff8020a816>] apic_timer_interrupt+0x66/0x70
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803cd3d3>] dst_run_gc+0x53/0x140
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803cd3c6>] dst_run_gc+0x46/0x140
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff80237148>] run_timer_softirq+0x148/0x1c0
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff8023340c>] __do_softirq+0x6c/0xe0
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff8020ad6c>] call_softirq+0x1c/0x30
Aug 16 06:21:37 SRV1 kernel:  <EOI>  [<ffffffff8020cb34>] do_softirq+0x34/0x90
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff802331cf>] local_bh_enable_ip+0x3f/0x60
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff80422913>] _spin_unlock_bh+0x13/0x20
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803dfde8>] rt_garbage_collect+0x1d8/0x320
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803cd4dd>] dst_alloc+0x1d/0xa0
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803e1433>] __ip_route_output_key+0x573/0x800
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803c02e2>] sock_common_recvmsg+0x32/0x50
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803e16dc>] ip_route_output_flow+0x1c/0x60
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff80400160>] tcp_v4_connect+0x150/0x610
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803ebf07>] inet_bind_bucket_create+0x17/0x60
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff8040cd16>] inet_stream_connect+0xa6/0x2c0
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff80422981>] _spin_lock_bh+0x11/0x30
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803c0bdf>] lock_sock_nested+0xcf/0xe0
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff80422981>] _spin_lock_bh+0x11/0x30
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803be551>] sys_connect+0x71/0xa0
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803eee3f>] tcp_setsockopt+0x1f/0x30
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803c030f>] sock_common_setsockopt+0xf/0x20
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff803be4bd>] sys_setsockopt+0x9d/0xc0
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff8028881e>] sys_ioctl+0x5e/0x80
Aug 16 06:21:37 SRV1 kernel:  [<ffffffff80209c4e>] system_call+0x7e/0x83

After patch : (RT_CACHE_DEBUG set to 2 to get following traces)

dst_total: 75469 delayed: 74109 work_perf: 141 expires: 150 elapsed: 8092 us
dst_total: 78725 delayed: 73366 work_perf: 743 expires: 400 elapsed: 8542 us
dst_total: 86126 delayed: 71844 work_perf: 1522 expires: 775 elapsed: 8849 us
dst_total: 100173 delayed: 68791 work_perf: 3053 expires: 1256 elapsed: 9748 us
dst_total: 121798 delayed: 64711 work_perf: 4080 expires: 1997 elapsed: 10146 us
dst_total: 154522 delayed: 58316 work_perf: 6395 expires: 25 elapsed: 11402 us
dst_total: 154957 delayed: 58252 work_perf: 64 expires: 150 elapsed: 6148 us
dst_total: 157377 delayed: 57843 work_perf: 409 expires: 400 elapsed: 6350 us
dst_total: 163745 delayed: 56679 work_perf: 1164 expires: 775 elapsed: 7051 us
dst_total: 176577 delayed: 53965 work_perf: 2714 expires: 1389 elapsed: 8120 us
dst_total: 198993 delayed: 49627 work_perf: 4338 expires: 1997 elapsed: 8909 us
dst_total: 226638 delayed: 46865 work_perf: 2762 expires: 2748 elapsed: 7351 us

I successfully reduced the IP route cache of many hosts by a four factor
thanks to this patch. Previously, I had to disable "ip route flush cache"
to avoid crashes.

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2007-10-10 16:49:15 -07:00

339 lines
7.8 KiB
C

/*
* net/core/dst.c Protocol independent destination cache.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*
*/
#include <linux/bitops.h>
#include <linux/errno.h>
#include <linux/init.h>
#include <linux/kernel.h>
#include <linux/workqueue.h>
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/string.h>
#include <linux/types.h>
#include <net/net_namespace.h>
#include <net/dst.h>
/*
* Theory of operations:
* 1) We use a list, protected by a spinlock, to add
* new entries from both BH and non-BH context.
* 2) In order to keep spinlock held for a small delay,
* we use a second list where are stored long lived
* entries, that are handled by the garbage collect thread
* fired by a workqueue.
* 3) This list is guarded by a mutex,
* so that the gc_task and dst_dev_event() can be synchronized.
*/
#if RT_CACHE_DEBUG >= 2
static atomic_t dst_total = ATOMIC_INIT(0);
#endif
/*
* We want to keep lock & list close together
* to dirty as few cache lines as possible in __dst_free().
* As this is not a very strong hint, we dont force an alignment on SMP.
*/
static struct {
spinlock_t lock;
struct dst_entry *list;
unsigned long timer_inc;
unsigned long timer_expires;
} dst_garbage = {
.lock = __SPIN_LOCK_UNLOCKED(dst_garbage.lock),
.timer_inc = DST_GC_MAX,
};
static void dst_gc_task(struct work_struct *work);
static void ___dst_free(struct dst_entry * dst);
static DECLARE_DELAYED_WORK(dst_gc_work, dst_gc_task);
static DEFINE_MUTEX(dst_gc_mutex);
/*
* long lived entries are maintained in this list, guarded by dst_gc_mutex
*/
static struct dst_entry *dst_busy_list;
static void dst_gc_task(struct work_struct *work)
{
int delayed = 0;
int work_performed = 0;
unsigned long expires = ~0L;
struct dst_entry *dst, *next, head;
struct dst_entry *last = &head;
#if RT_CACHE_DEBUG >= 2
ktime_t time_start = ktime_get();
struct timespec elapsed;
#endif
mutex_lock(&dst_gc_mutex);
next = dst_busy_list;
loop:
while ((dst = next) != NULL) {
next = dst->next;
prefetch(&next->next);
if (likely(atomic_read(&dst->__refcnt))) {
last->next = dst;
last = dst;
delayed++;
continue;
}
work_performed++;
dst = dst_destroy(dst);
if (dst) {
/* NOHASH and still referenced. Unless it is already
* on gc list, invalidate it and add to gc list.
*
* Note: this is temporary. Actually, NOHASH dst's
* must be obsoleted when parent is obsoleted.
* But we do not have state "obsoleted, but
* referenced by parent", so it is right.
*/
if (dst->obsolete > 1)
continue;
___dst_free(dst);
dst->next = next;
next = dst;
}
}
spin_lock_bh(&dst_garbage.lock);
next = dst_garbage.list;
if (next) {
dst_garbage.list = NULL;
spin_unlock_bh(&dst_garbage.lock);
goto loop;
}
last->next = NULL;
dst_busy_list = head.next;
if (!dst_busy_list)
dst_garbage.timer_inc = DST_GC_MAX;
else {
/*
* if we freed less than 1/10 of delayed entries,
* we can sleep longer.
*/
if (work_performed <= delayed/10) {
dst_garbage.timer_expires += dst_garbage.timer_inc;
if (dst_garbage.timer_expires > DST_GC_MAX)
dst_garbage.timer_expires = DST_GC_MAX;
dst_garbage.timer_inc += DST_GC_INC;
} else {
dst_garbage.timer_inc = DST_GC_INC;
dst_garbage.timer_expires = DST_GC_MIN;
}
expires = dst_garbage.timer_expires;
/*
* if the next desired timer is more than 4 seconds in the future
* then round the timer to whole seconds
*/
if (expires > 4*HZ)
expires = round_jiffies_relative(expires);
schedule_delayed_work(&dst_gc_work, expires);
}
spin_unlock_bh(&dst_garbage.lock);
mutex_unlock(&dst_gc_mutex);
#if RT_CACHE_DEBUG >= 2
elapsed = ktime_to_timespec(ktime_sub(ktime_get(), time_start));
printk(KERN_DEBUG "dst_total: %d delayed: %d work_perf: %d"
" expires: %lu elapsed: %lu us\n",
atomic_read(&dst_total), delayed, work_performed,
expires,
elapsed.tv_sec * USEC_PER_SEC + elapsed.tv_nsec / NSEC_PER_USEC);
#endif
}
static int dst_discard(struct sk_buff *skb)
{
kfree_skb(skb);
return 0;
}
void * dst_alloc(struct dst_ops * ops)
{
struct dst_entry * dst;
if (ops->gc && atomic_read(&ops->entries) > ops->gc_thresh) {
if (ops->gc())
return NULL;
}
dst = kmem_cache_zalloc(ops->kmem_cachep, GFP_ATOMIC);
if (!dst)
return NULL;
atomic_set(&dst->__refcnt, 0);
dst->ops = ops;
dst->lastuse = jiffies;
dst->path = dst;
dst->input = dst->output = dst_discard;
#if RT_CACHE_DEBUG >= 2
atomic_inc(&dst_total);
#endif
atomic_inc(&ops->entries);
return dst;
}
static void ___dst_free(struct dst_entry * dst)
{
/* The first case (dev==NULL) is required, when
protocol module is unloaded.
*/
if (dst->dev == NULL || !(dst->dev->flags&IFF_UP)) {
dst->input = dst->output = dst_discard;
}
dst->obsolete = 2;
}
void __dst_free(struct dst_entry * dst)
{
spin_lock_bh(&dst_garbage.lock);
___dst_free(dst);
dst->next = dst_garbage.list;
dst_garbage.list = dst;
if (dst_garbage.timer_inc > DST_GC_INC) {
dst_garbage.timer_inc = DST_GC_INC;
dst_garbage.timer_expires = DST_GC_MIN;
schedule_delayed_work(&dst_gc_work, dst_garbage.timer_expires);
}
spin_unlock_bh(&dst_garbage.lock);
}
struct dst_entry *dst_destroy(struct dst_entry * dst)
{
struct dst_entry *child;
struct neighbour *neigh;
struct hh_cache *hh;
smp_rmb();
again:
neigh = dst->neighbour;
hh = dst->hh;
child = dst->child;
dst->hh = NULL;
if (hh && atomic_dec_and_test(&hh->hh_refcnt))
kfree(hh);
if (neigh) {
dst->neighbour = NULL;
neigh_release(neigh);
}
atomic_dec(&dst->ops->entries);
if (dst->ops->destroy)
dst->ops->destroy(dst);
if (dst->dev)
dev_put(dst->dev);
#if RT_CACHE_DEBUG >= 2
atomic_dec(&dst_total);
#endif
kmem_cache_free(dst->ops->kmem_cachep, dst);
dst = child;
if (dst) {
int nohash = dst->flags & DST_NOHASH;
if (atomic_dec_and_test(&dst->__refcnt)) {
/* We were real parent of this dst, so kill child. */
if (nohash)
goto again;
} else {
/* Child is still referenced, return it for freeing. */
if (nohash)
return dst;
/* Child is still in his hash table */
}
}
return NULL;
}
/* Dirty hack. We did it in 2.2 (in __dst_free),
* we have _very_ good reasons not to repeat
* this mistake in 2.3, but we have no choice
* now. _It_ _is_ _explicit_ _deliberate_
* _race_ _condition_.
*
* Commented and originally written by Alexey.
*/
static inline void dst_ifdown(struct dst_entry *dst, struct net_device *dev,
int unregister)
{
if (dst->ops->ifdown)
dst->ops->ifdown(dst, dev, unregister);
if (dev != dst->dev)
return;
if (!unregister) {
dst->input = dst->output = dst_discard;
} else {
dst->dev = &loopback_dev;
dev_hold(&loopback_dev);
dev_put(dev);
if (dst->neighbour && dst->neighbour->dev == dev) {
dst->neighbour->dev = &loopback_dev;
dev_put(dev);
dev_hold(&loopback_dev);
}
}
}
static int dst_dev_event(struct notifier_block *this, unsigned long event, void *ptr)
{
struct net_device *dev = ptr;
struct dst_entry *dst, *last = NULL;
if (dev->nd_net != &init_net)
return NOTIFY_DONE;
switch (event) {
case NETDEV_UNREGISTER:
case NETDEV_DOWN:
mutex_lock(&dst_gc_mutex);
for (dst = dst_busy_list; dst; dst = dst->next) {
last = dst;
dst_ifdown(dst, dev, event != NETDEV_DOWN);
}
spin_lock_bh(&dst_garbage.lock);
dst = dst_garbage.list;
dst_garbage.list = NULL;
spin_unlock_bh(&dst_garbage.lock);
if (last)
last->next = dst;
else
dst_busy_list = dst;
for (; dst; dst = dst->next) {
dst_ifdown(dst, dev, event != NETDEV_DOWN);
}
mutex_unlock(&dst_gc_mutex);
break;
}
return NOTIFY_DONE;
}
static struct notifier_block dst_dev_notifier = {
.notifier_call = dst_dev_event,
};
void __init dst_init(void)
{
register_netdevice_notifier(&dst_dev_notifier);
}
EXPORT_SYMBOL(__dst_free);
EXPORT_SYMBOL(dst_alloc);
EXPORT_SYMBOL(dst_destroy);