padata: make the sequence counter an atomic_t

Using a spinlock to atomically increase a counter sounds wrong -- we've
atomic_t for this!

Also move 'seq_nr' to a different cache line than 'lock' to reduce cache
line trashing. This has the nice side effect of decreasing the size of
struct parallel_data from 192 to 128 bytes for a x86-64 build, e.g.
occupying only two instead of three cache lines.

Those changes results in a 5% performance increase on an IPsec test run
using pcrypt.

Btw. the seq_lock spinlock was never explicitly initialized -- one more
reason to get rid of it.

Signed-off-by: Mathias Krause <mathias.krause@secunet.com>
Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

1 files changed, 1 insertions, 2 deletions

diff --git a/include/linux/padata.h b/include/linux/padata.h
index 86292be..4386946 100644
--- a/include/linux/padata.h
+++ b/include/linux/padata.h
@@ -129,10 +129,9 @@ struct parallel_data {
 	struct padata_serial_queue	__percpu *squeue;
 	atomic_t			reorder_objects;
 	atomic_t			refcnt;
+	atomic_t			seq_nr;
 	struct padata_cpumask		cpumask;
 	spinlock_t                      lock ____cacheline_aligned;
-	spinlock_t                      seq_lock;
-	unsigned int			seq_nr;
 	unsigned int			processed;
 	struct timer_list		timer;
 };