[XFRM]: Kill excessive refcounting of xfrm_state objects.

The refcounting done for timers and hash table insertions
are just wasted cycles.  We can eliminate all of this
refcounting because:

1) The implicit refcount when the xfrm_state object is active
   will always be held while the object is in the hash tables.
   We never kfree() the xfrm_state until long after we've made
   sure that it has been unhashed.

2) Timers are even easier.  Once we mark that x->km.state as
   anything other than XFRM_STATE_VALID (__xfrm_state_delete
   sets it to XFRM_STATE_DEAD), any timer that fires will
   do nothing and return without rearming the timer.

   Therefore we can defer the del_timer calls until when the
   object is about to be freed up during GC.  We have to use
   del_timer_sync() and defer it to GC because we can't do
   a del_timer_sync() while holding x->lock which all callers
   of __xfrm_state_delete hold.

This makes SA changes even more light-weight.

Signed-off-by: David S. Miller <[EMAIL PROTECTED]>
---
 net/xfrm/xfrm_state.c |   54 +++++++++++++------------------------------------
 1 files changed, 14 insertions(+), 40 deletions(-)

diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
index b1c0ae7..814e72e 100644
--- a/net/xfrm/xfrm_state.c
+++ b/net/xfrm/xfrm_state.c
@@ -229,10 +229,8 @@ void km_state_expired(struct xfrm_state 
 
 static void xfrm_state_gc_destroy(struct xfrm_state *x)
 {
-       if (del_timer(&x->timer))
-               BUG();
-       if (del_timer(&x->rtimer))
-               BUG();
+       del_timer_sync(&x->timer);
+       del_timer_sync(&x->rtimer);
        kfree(x->aalg);
        kfree(x->ealg);
        kfree(x->calg);
@@ -323,9 +321,9 @@ static void xfrm_timer_handler(unsigned 
        if (warn)
                km_state_expired(x, 0, 0);
 resched:
-       if (next != LONG_MAX &&
-           !mod_timer(&x->timer, jiffies + make_jiffies(next)))
-               xfrm_state_hold(x);
+       if (next != LONG_MAX)
+               mod_timer(&x->timer, jiffies + make_jiffies(next));
+
        goto out;
 
 expired:
@@ -340,7 +338,6 @@ expired:
 
 out:
        spin_unlock(&x->lock);
-       xfrm_state_put(x);
 }
 
 static void xfrm_replay_timer_handler(unsigned long data);
@@ -394,17 +391,10 @@ int __xfrm_state_delete(struct xfrm_stat
                x->km.state = XFRM_STATE_DEAD;
                spin_lock(&xfrm_state_lock);
                hlist_del(&x->bydst);
-               __xfrm_state_put(x);
-               if (x->id.spi) {
+               if (x->id.spi)
                        hlist_del(&x->byspi);
-                       __xfrm_state_put(x);
-               }
                xfrm_state_num--;
                spin_unlock(&xfrm_state_lock);
-               if (del_timer(&x->timer))
-                       __xfrm_state_put(x);
-               if (del_timer(&x->rtimer))
-                       __xfrm_state_put(x);
 
                /* All xfrm_state objects are created by xfrm_state_alloc.
                 * The xfrm_state_alloc call gives a reference, and that
@@ -579,14 +569,11 @@ xfrm_state_find(xfrm_address_t *daddr, x
                if (km_query(x, tmpl, pol) == 0) {
                        x->km.state = XFRM_STATE_ACQ;
                        hlist_add_head(&x->bydst, xfrm_state_bydst+h);
-                       xfrm_state_hold(x);
                        if (x->id.spi) {
                                h = xfrm_spi_hash(&x->id.daddr, x->id.spi, 
x->id.proto, family);
                                hlist_add_head(&x->byspi, xfrm_state_byspi+h);
-                               xfrm_state_hold(x);
                        }
                        x->lft.hard_add_expires_seconds = XFRM_ACQ_EXPIRES;
-                       xfrm_state_hold(x);
                        x->timer.expires = jiffies + XFRM_ACQ_EXPIRES*HZ;
                        add_timer(&x->timer);
                } else {
@@ -609,24 +596,17 @@ static void __xfrm_state_insert(struct x
 {
        unsigned int h;
 
-       h = xfrm_dst_hash(&x->id.daddr, x->props.reqid, x->props.family);
-
        x->genid = ++xfrm_state_genid;
 
+       h = xfrm_dst_hash(&x->id.daddr, x->props.reqid, x->props.family);
        hlist_add_head(&x->bydst, xfrm_state_bydst+h);
-       xfrm_state_hold(x);
 
        h = xfrm_spi_hash(&x->id.daddr, x->id.spi, x->id.proto, 
x->props.family);
-
        hlist_add_head(&x->byspi, xfrm_state_byspi+h);
-       xfrm_state_hold(x);
-
-       if (!mod_timer(&x->timer, jiffies + HZ))
-               xfrm_state_hold(x);
 
-       if (x->replay_maxage &&
-           !mod_timer(&x->rtimer, jiffies + x->replay_maxage))
-               xfrm_state_hold(x);
+       mod_timer(&x->timer, jiffies + HZ);
+       if (x->replay_maxage)
+               mod_timer(&x->rtimer, jiffies + x->replay_maxage);
 
        wake_up(&km_waitq);
 
@@ -738,7 +718,6 @@ static struct xfrm_state *__find_acq_cor
                xfrm_state_hold(x);
                x->timer.expires = jiffies + XFRM_ACQ_EXPIRES*HZ;
                add_timer(&x->timer);
-               xfrm_state_hold(x);
                hlist_add_head(&x->bydst, xfrm_state_bydst+h);
                wake_up(&km_waitq);
        }
@@ -840,8 +819,7 @@ out:
                memcpy(&x1->lft, &x->lft, sizeof(x1->lft));
                x1->km.dying = 0;
 
-               if (!mod_timer(&x1->timer, jiffies + HZ))
-                       xfrm_state_hold(x1);
+               mod_timer(&x1->timer, jiffies + HZ);
                if (x1->curlft.use_time)
                        xfrm_state_check_expire(x1);
 
@@ -866,8 +844,7 @@ int xfrm_state_check_expire(struct xfrm_
        if (x->curlft.bytes >= x->lft.hard_byte_limit ||
            x->curlft.packets >= x->lft.hard_packet_limit) {
                x->km.state = XFRM_STATE_EXPIRED;
-               if (!mod_timer(&x->timer, jiffies))
-                       xfrm_state_hold(x);
+               mod_timer(&x->timer, jiffies);
                return -EINVAL;
        }
 
@@ -1010,7 +987,6 @@ xfrm_alloc_spi(struct xfrm_state *x, u32
                spin_lock_bh(&xfrm_state_lock);
                h = xfrm_spi_hash(&x->id.daddr, x->id.spi, x->id.proto, 
x->props.family);
                hlist_add_head(&x->byspi, xfrm_state_byspi+h);
-               xfrm_state_hold(x);
                spin_unlock_bh(&xfrm_state_lock);
                wake_up(&km_waitq);
        }
@@ -1096,9 +1072,8 @@ void xfrm_replay_notify(struct xfrm_stat
        c.data.aevent = event;
        km_state_notify(x, &c);
 
-       if (x->replay_maxage &&
-           !mod_timer(&x->rtimer, jiffies + x->replay_maxage)) {
-               xfrm_state_hold(x);
+       if (x->replay_maxage) {
+               mod_timer(&x->rtimer, jiffies + x->replay_maxage);
                x->xflags &= ~XFRM_TIME_DEFER;
        }
 }
@@ -1118,7 +1093,6 @@ static void xfrm_replay_timer_handler(un
        }
 
        spin_unlock(&x->lock);
-       xfrm_state_put(x);
 }
 
 int xfrm_replay_check(struct xfrm_state *x, u32 seq)
-- 
1.4.2.rc2.g3e042

-
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

Reply via email to