lkml.org 
[lkml]   [2018]   [Jun]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Subject[PATCH tip/core/rcu 19/40] rcu: Make rcu_nocb_wait_gp() check if GP already requested
Date
This commit makes rcu_nocb_wait_gp() check rdp->gp_seq_needed to see
if the current CPU already knows about the needed grace period having
already been requested. If so, it avoids acquiring the corresponding
leaf rcu_node structure's ->lock, thus decreasing contention. This
optimization is intended for cases where either multiple leader rcuo
kthreads are running on the same CPU or these kthreads are running on
a non-offloaded (e.g., housekeeping) CPU.

Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
[ paulmck: Move lock release past "if" as suggested by Joel Fernandes. ]
[ paulmck: Fix caching of furthest-future requested grace period. ]
---
kernel/rcu/tree.c | 5 +++++
kernel/rcu/tree_plugin.h | 15 ++++++++++-----
2 files changed, 15 insertions(+), 5 deletions(-)

diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
index e5d1369e099f..0da0db1128d0 100644
--- a/kernel/rcu/tree.c
+++ b/kernel/rcu/tree.c
@@ -1618,6 +1618,11 @@ static bool rcu_start_this_gp(struct rcu_node *rnp, struct rcu_data *rdp,
trace_rcu_grace_period(rsp->name, READ_ONCE(rsp->gpnum), TPS("newreq"));
ret = true; /* Caller must wake GP kthread. */
unlock_out:
+ /* Push furthest requested GP to leaf node and rcu_data structure. */
+ if (ULONG_CMP_LT(c, rnp_root->gp_seq_needed)) {
+ rnp->gp_seq_needed = rnp_root->gp_seq_needed;
+ rdp->gp_seq_needed = rnp_root->gp_seq_needed;
+ }
if (rnp != rnp_root)
raw_spin_unlock_rcu_node(rnp_root);
return ret;
diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
index f4a88e3c388d..ca73931f7b30 100644
--- a/kernel/rcu/tree_plugin.h
+++ b/kernel/rcu/tree_plugin.h
@@ -2104,12 +2104,17 @@ static void rcu_nocb_wait_gp(struct rcu_data *rdp)
bool needwake;
struct rcu_node *rnp = rdp->mynode;

- raw_spin_lock_irqsave_rcu_node(rnp, flags);
+ local_irq_save(flags);
c = rcu_seq_snap(&rdp->rsp->gp_seq);
- needwake = rcu_start_this_gp(rnp, rdp, c);
- raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
- if (needwake)
- rcu_gp_kthread_wake(rdp->rsp);
+ if (!rdp->gpwrap && ULONG_CMP_GE(rdp->gp_seq_needed, c)) {
+ local_irq_restore(flags);
+ } else {
+ raw_spin_lock_rcu_node(rnp); /* irqs already disabled. */
+ needwake = rcu_start_this_gp(rnp, rdp, c);
+ raw_spin_unlock_irqrestore_rcu_node(rnp, flags);
+ if (needwake)
+ rcu_gp_kthread_wake(rdp->rsp);
+ }

/*
* Wait for the grace period. Do so interruptibly to avoid messing
--
2.17.1
\
 
 \ /
  Last update: 2018-06-26 02:15    [W:0.223 / U:1.548 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site