On Mon, Jan 21, 2008 at 02:40:43AM -0800, David Miller wrote:
> From: Joonwoo Park <[EMAIL PROTECTED]>
> Date: Tue, 22 Jan 2008 00:08:57 +0900
>
> > The rt_run_flush() can be stucked if it was called while netdev is on the
> > high load.
> > It's possible when pushing rtable to rt_hash is faster than pulling
> > from it.
> >
> > Signed-off-by: Joonwoo Park <[EMAIL PROTECTED]>
>
> I agree with the analysis of the problem, however not the solution.
>
> This will absolutely kill software interrupt latency.
>
> In fact, we have moved much of the flush work into a workqueue in
> net-2.6.25 because of how important that is
>
> We need to find some other way to solve this.
>
Dave, Eric,
Thanks so much for comments.
I did stress tests and I found that the real problem was not consumer & supplier
issue.
It was the problem for me to innumerable enabling & disabling the softirq.
But I'm still thinking need of considering issue 'faster caching than flush'.
:)
ifconfig up on heavy loaded interface.
Before patching:
time ifconfig eth1 up
BUG: soft lockup - CPU#0 stuck for 11s! [events/0:9]
...
After patching:
time ifconfig eth1 up
real 0m0.007s
user 0m0.000s
sys 0m0.004s
Thanks!
Joonwoo
>From 87c29506de967e811ad5b57cd2e1a002134e878f Mon Sep 17 00:00:00 2001
From: Joonwoo Park <[EMAIL PROTECTED]>
Date: Wed, 23 Jan 2008 15:16:54 +0900
Subject: [PATCH] [IPV4] route: reduce locking/unlocking in rt_run_flush
The rt_run_flush does spin_lock_bh/spin_unlock_bh for rt_hash_mask + 1
times.
The rt_hash_mask takes from 32767 to 65535, so it's big overhead.
In addition, disable_bh/enable_bh for many times in the rt_run_flush
can cause stuck on a machine with heavily pended softirqs.
This patch reduces locking/unlocking as doing it with jumping the lock
slots.
ifconfig up on heavy loaded interface.
Before:
time ifconfig eth1 up
BUG: soft lockup - CPU#0 stuck for 11s! [events/0:9]
...
After:
time ifconfig eth1 up
real 0m0.007s
user 0m0.000s
sys 0m0.004s
Signed-off-by: Joonwoo Park <[EMAIL PROTECTED]>
---
net/ipv4/route.c | 38 +++++++++++++++++++++++++++-----------
1 files changed, 27 insertions(+), 11 deletions(-)
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 28484f3..79a401f 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -239,9 +239,20 @@ static spinlock_t *rt_hash_locks;
for (i = 0; i < RT_HASH_LOCK_SZ; i++) \
spin_lock_init(&rt_hash_locks[i]); \
}
+# define rt_hash_lock(lock) &rt_hash_locks[lock]
+# define rt_hash_for_each_lock(lock) \
+ for (lock = 0; lock < RT_HASH_LOCK_SZ; lock++)
+# define rt_hash_for_each_slot(slot, lock, mask) \
+ for (slot = 0, mask = lock; \
+ slot < ((int)rt_hash_mask + 1) / RT_HASH_LOCK_SZ; \
+ slot++, mask = (slot * RT_HASH_LOCK_SZ) + lock)
#else
# define rt_hash_lock_addr(slot) NULL
# define rt_hash_lock_init()
+# define rt_hash_lock(lock) NULL
+# define rt_hash_for_each_lock(lock) do { lock = 0; } while (0);
+# define rt_hash_for_each_slot(slot, lock, mask) \
+ for (slot = 0, lock = 0, mask = 0; mask <= rt_hash_mask; mask++)
#endif
static struct rt_hash_bucket *rt_hash_table;
@@ -613,24 +624,29 @@ static void rt_check_expire(struct work_struct *work)
*/
static void rt_run_flush(unsigned long dummy)
{
- int i;
+ int slot, lock, mask;
struct rtable *rth, *next;
rt_deadline = 0;
get_random_bytes(&rt_hash_rnd, 4);
- for (i = rt_hash_mask; i >= 0; i--) {
- spin_lock_bh(rt_hash_lock_addr(i));
- rth = rt_hash_table[i].chain;
- if (rth)
- rt_hash_table[i].chain = NULL;
- spin_unlock_bh(rt_hash_lock_addr(i));
-
- for (; rth; rth = next) {
- next = rth->u.dst.rt_next;
- rt_free(rth);
+ rt_hash_for_each_lock(lock) {
+ spin_lock_bh(rt_hash_lock(lock));
+ rt_hash_for_each_slot(slot, lock, mask) {
+ rth = rt_hash_table[mask].chain;
+
+ if (rth) {
+ rt_hash_table[mask].chain = NULL;
+ spin_unlock_bh(rt_hash_lock(lock));
+ for (; rth; rth = next) {
+ next = rth->u.dst.rt_next;
+ rt_free(rth);
+ }
+ spin_lock_bh(rt_hash_lock(lock));
+ }
}
+ spin_unlock_bh(rt_hash_lock(lock));
}
}
--
1.5.3.rc5
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
