[next,v2,3/5] locking/osq_lock: Use node->prev_cpu instead of saving node->prev.
Commit Message
node->prev is only used to update 'prev' in the unlikely case
of concurrent unqueues.
This can be replaced by a check for node->prev_cpu changing
and then calling decode_cpu() to get the changed 'prev' pointer.
node->cpu (or more particularly) prev->cpu is only used for the
osq_wait_next() call in the unqueue path.
Normally this is exactly the value that the initial xchg() read
from lock->tail (used to obtain 'prev'), but can get updated
by concurrent unqueues.
Both the 'prev' and 'cpu' members of optimistic_spin_node are
now unused and can be deleted.
Signed-off-by: David Laight <david.laight@aculab.com>
---
kernel/locking/osq_lock.c | 31 +++++++++++++++++--------------
1 file changed, 17 insertions(+), 14 deletions(-)
Comments
On 12/31/23 16:54, David Laight wrote:
> node->prev is only used to update 'prev' in the unlikely case
> of concurrent unqueues.
> This can be replaced by a check for node->prev_cpu changing
> and then calling decode_cpu() to get the changed 'prev' pointer.
>
> node->cpu (or more particularly) prev->cpu is only used for the
> osq_wait_next() call in the unqueue path.
> Normally this is exactly the value that the initial xchg() read
> from lock->tail (used to obtain 'prev'), but can get updated
> by concurrent unqueues.
>
> Both the 'prev' and 'cpu' members of optimistic_spin_node are
> now unused and can be deleted.
>
> Signed-off-by: David Laight <david.laight@aculab.com>
> ---
> kernel/locking/osq_lock.c | 31 +++++++++++++++++--------------
> 1 file changed, 17 insertions(+), 14 deletions(-)
>
> diff --git a/kernel/locking/osq_lock.c b/kernel/locking/osq_lock.c
> index eb8a6dfdb79d..27324b509f68 100644
> --- a/kernel/locking/osq_lock.c
> +++ b/kernel/locking/osq_lock.c
> @@ -13,9 +13,8 @@
> */
>
> struct optimistic_spin_node {
> - struct optimistic_spin_node *next, *prev;
> + struct optimistic_spin_node *next;
> int locked; /* 1 if lock acquired */
> - int cpu; /* encoded CPU # + 1 value */
> int prev_cpu; /* encoded CPU # + 1 value */
> };
>
> @@ -91,10 +90,9 @@ bool osq_lock(struct optimistic_spin_queue *lock)
> struct optimistic_spin_node *node = this_cpu_ptr(&osq_node);
> struct optimistic_spin_node *prev, *next;
> int curr = encode_cpu(smp_processor_id());
> - int old;
> + int prev_cpu;
>
> node->next = NULL;
> - node->cpu = curr;
>
> /*
> * We need both ACQUIRE (pairs with corresponding RELEASE in
> @@ -102,13 +100,12 @@ bool osq_lock(struct optimistic_spin_queue *lock)
> * the node fields we just initialised) semantics when updating
> * the lock tail.
> */
> - old = atomic_xchg(&lock->tail, curr);
> - if (old == OSQ_UNLOCKED_VAL)
> + prev_cpu = atomic_xchg(&lock->tail, curr);
> + if (prev_cpu == OSQ_UNLOCKED_VAL)
> return true;
>
> - node->prev_cpu = old;
> - prev = decode_cpu(old);
> - node->prev = prev;
> + node->prev_cpu = prev_cpu;
> + prev = decode_cpu(prev_cpu);
> node->locked = 0;
>
> /*
> @@ -174,9 +171,16 @@ bool osq_lock(struct optimistic_spin_queue *lock)
>
> /*
> * Or we race against a concurrent unqueue()'s step-B, in which
> - * case its step-C will write us a new @node->prev pointer.
> + * case its step-C will write us a new @node->prev_cpu value.
> */
> - prev = READ_ONCE(node->prev);
> + {
> + int new_prev_cpu = READ_ONCE(node->prev_cpu);
> +
> + if (new_prev_cpu == prev_cpu)
> + continue;
> + prev_cpu = new_prev_cpu;
> + prev = decode_cpu(prev_cpu);
> + }
Just a minor nit. It is not that common in the kernel to add another
nesting level just to reduce the scope of new_prev_cpu auto variable.
Anyway,
Reviewed-by: Waiman Long <longman@redhat.com>
@@ -13,9 +13,8 @@
*/
struct optimistic_spin_node {
- struct optimistic_spin_node *next, *prev;
+ struct optimistic_spin_node *next;
int locked; /* 1 if lock acquired */
- int cpu; /* encoded CPU # + 1 value */
int prev_cpu; /* encoded CPU # + 1 value */
};
@@ -91,10 +90,9 @@ bool osq_lock(struct optimistic_spin_queue *lock)
struct optimistic_spin_node *node = this_cpu_ptr(&osq_node);
struct optimistic_spin_node *prev, *next;
int curr = encode_cpu(smp_processor_id());
- int old;
+ int prev_cpu;
node->next = NULL;
- node->cpu = curr;
/*
* We need both ACQUIRE (pairs with corresponding RELEASE in
@@ -102,13 +100,12 @@ bool osq_lock(struct optimistic_spin_queue *lock)
* the node fields we just initialised) semantics when updating
* the lock tail.
*/
- old = atomic_xchg(&lock->tail, curr);
- if (old == OSQ_UNLOCKED_VAL)
+ prev_cpu = atomic_xchg(&lock->tail, curr);
+ if (prev_cpu == OSQ_UNLOCKED_VAL)
return true;
- node->prev_cpu = old;
- prev = decode_cpu(old);
- node->prev = prev;
+ node->prev_cpu = prev_cpu;
+ prev = decode_cpu(prev_cpu);
node->locked = 0;
/*
@@ -174,9 +171,16 @@ bool osq_lock(struct optimistic_spin_queue *lock)
/*
* Or we race against a concurrent unqueue()'s step-B, in which
- * case its step-C will write us a new @node->prev pointer.
+ * case its step-C will write us a new @node->prev_cpu value.
*/
- prev = READ_ONCE(node->prev);
+ {
+ int new_prev_cpu = READ_ONCE(node->prev_cpu);
+
+ if (new_prev_cpu == prev_cpu)
+ continue;
+ prev_cpu = new_prev_cpu;
+ prev = decode_cpu(prev_cpu);
+ }
}
/*
@@ -186,7 +190,7 @@ bool osq_lock(struct optimistic_spin_queue *lock)
* back to @prev.
*/
- next = osq_wait_next(lock, node, prev->cpu);
+ next = osq_wait_next(lock, node, prev_cpu);
if (!next)
return false;
@@ -198,8 +202,7 @@ bool osq_lock(struct optimistic_spin_queue *lock)
* it will wait in Step-A.
*/
- WRITE_ONCE(next->prev_cpu, prev->cpu);
- WRITE_ONCE(next->prev, prev);
+ WRITE_ONCE(next->prev_cpu, prev_cpu);
WRITE_ONCE(prev->next, next);
return false;