Skip to content

Commit

Permalink
rcu: Rearrange locking in rcu_start_gp()
Browse files Browse the repository at this point in the history
If CPUs are to give prior notice of needed grace periods, it will be
necessary to invoke rcu_start_gp() without dropping the root rcu_node
structure's ->lock.  This commit takes a first step in this direction
by moving the release of this lock to the end of rcu_start_gp().

Signed-off-by: Paul E. McKenney <paul.mckenney@linaro.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
  • Loading branch information
Paul E. McKenney authored and Paul E. McKenney committed Mar 26, 2013
1 parent c0f4dfd commit b92db6c
Showing 1 changed file with 2 additions and 4 deletions.
6 changes: 2 additions & 4 deletions kernel/rcutree.c
Original file line number Diff line number Diff line change
Expand Up @@ -1521,16 +1521,14 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags)
raw_spin_unlock_irqrestore(&rnp->lock, flags);
return;
}

rsp->gp_flags = RCU_GP_FLAG_INIT;
raw_spin_unlock(&rnp->lock); /* Interrupts remain disabled. */

/* Ensure that CPU is aware of completion of last grace period. */
rcu_process_gp_end(rsp, rdp);
local_irq_restore(flags);
__rcu_process_gp_end(rsp, rdp->mynode, rdp);

/* Wake up rcu_gp_kthread() to start the grace period. */
wake_up(&rsp->gp_wq);
raw_spin_unlock_irqrestore(&rnp->lock, flags);
}

/*
Expand Down

0 comments on commit b92db6c

Please sign in to comment.