LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Peter Zijlstra <peterz@infradead.org>
To: Waiman.Long@hp.com
Cc: tglx@linutronix.de, mingo@redhat.com, hpa@zytor.com,
	peterz@infradead.org, paolo.bonzini@gmail.com,
	konrad.wilk@oracle.com, boris.ostrovsky@oracle.com,
	paulmck@linux.vnet.ibm.com, riel@redhat.com,
	torvalds@linux-foundation.org, raghavendra.kt@linux.vnet.ibm.com,
	david.vrabel@citrix.com, oleg@redhat.com, scott.norton@hp.com,
	doug.hatch@hp.com, linux-arch@vger.kernel.org, x86@kernel.org,
	linux-kernel@vger.kernel.org,
	virtualization@lists.linux-foundation.org,
	xen-devel@lists.xenproject.org, kvm@vger.kernel.org,
	luto@amacapital.net
Subject: [PATCH 5/9] qspinlock: Optimize for smaller NR_CPUS
Date: Mon, 16 Mar 2015 14:16:18 +0100	[thread overview]
Message-ID: <20150316133112.084664282@infradead.org> (raw)
In-Reply-To: <20150316131613.720617163@infradead.org>

[-- Attachment #1: peter_zijlstra-qspinlock-optimize_for_smaller_nr_cpus.patch --]
[-- Type: text/plain, Size: 5852 bytes --]

From: Peter Zijlstra <peterz@infradead.org>

When we allow for a max NR_CPUS < 2^14 we can optimize the pending
wait-acquire and the xchg_tail() operations.

By growing the pending bit to a byte, we reduce the tail to 16bit.
This means we can use xchg16 for the tail part and do away with all
the repeated compxchg() operations.

This in turn allows us to unconditionally acquire; the locked state
as observed by the wait loops cannot change. And because both locked
and pending are now a full byte we can use simple stores for the
state transition, obviating one atomic operation entirely.

This optimization is needed to make the qspinlock achieve performance
parity with ticket spinlock at light load.

All this is horribly broken on Alpha pre EV56 (and any other arch that
cannot do single-copy atomic byte stores).

Cc: Ingo Molnar <mingo@redhat.com>
Cc: David Vrabel <david.vrabel@citrix.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Scott J Norton <scott.norton@hp.com>
Cc: Paolo Bonzini <paolo.bonzini@gmail.com>
Cc: Douglas Hatch <doug.hatch@hp.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>
Signed-off-by: Waiman Long <Waiman.Long@hp.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1421784755-21945-6-git-send-email-Waiman.Long@hp.com
---
 include/asm-generic/qspinlock_types.h |   13 ++++++
 kernel/locking/qspinlock.c            |   71 +++++++++++++++++++++++++++++++++-
 2 files changed, 83 insertions(+), 1 deletion(-)

--- a/include/asm-generic/qspinlock_types.h
+++ b/include/asm-generic/qspinlock_types.h
@@ -35,6 +35,14 @@ typedef struct qspinlock {
 /*
  * Bitfields in the atomic value:
  *
+ * When NR_CPUS < 16K
+ *  0- 7: locked byte
+ *     8: pending
+ *  9-15: not used
+ * 16-17: tail index
+ * 18-31: tail cpu (+1)
+ *
+ * When NR_CPUS >= 16K
  *  0- 7: locked byte
  *     8: pending
  *  9-10: tail index
@@ -47,7 +55,11 @@ typedef struct qspinlock {
 #define _Q_LOCKED_MASK		_Q_SET_MASK(LOCKED)
 
 #define _Q_PENDING_OFFSET	(_Q_LOCKED_OFFSET + _Q_LOCKED_BITS)
+#if CONFIG_NR_CPUS < (1U << 14)
+#define _Q_PENDING_BITS		8
+#else
 #define _Q_PENDING_BITS		1
+#endif
 #define _Q_PENDING_MASK		_Q_SET_MASK(PENDING)
 
 #define _Q_TAIL_IDX_OFFSET	(_Q_PENDING_OFFSET + _Q_PENDING_BITS)
@@ -58,6 +70,7 @@ typedef struct qspinlock {
 #define _Q_TAIL_CPU_BITS	(32 - _Q_TAIL_CPU_OFFSET)
 #define _Q_TAIL_CPU_MASK	_Q_SET_MASK(TAIL_CPU)
 
+#define _Q_TAIL_OFFSET		_Q_TAIL_IDX_OFFSET
 #define _Q_TAIL_MASK		(_Q_TAIL_IDX_MASK | _Q_TAIL_CPU_MASK)
 
 #define _Q_LOCKED_VAL		(1U << _Q_LOCKED_OFFSET)
--- a/kernel/locking/qspinlock.c
+++ b/kernel/locking/qspinlock.c
@@ -24,6 +24,7 @@
 #include <linux/percpu.h>
 #include <linux/hardirq.h>
 #include <linux/mutex.h>
+#include <asm/byteorder.h>
 #include <asm/qspinlock.h>
 
 /*
@@ -56,6 +57,10 @@
  * node; whereby avoiding the need to carry a node from lock to unlock, and
  * preserving existing lock API. This also makes the unlock code simpler and
  * faster.
+ *
+ * N.B. The current implementation only supports architectures that allow
+ *      atomic operations on smaller 8-bit and 16-bit data types.
+ *
  */
 
 #include "mcs_spinlock.h"
@@ -96,6 +101,64 @@ static inline struct mcs_spinlock *decod
 
 #define _Q_LOCKED_PENDING_MASK (_Q_LOCKED_MASK | _Q_PENDING_MASK)
 
+/*
+ * By using the whole 2nd least significant byte for the pending bit, we
+ * can allow better optimization of the lock acquisition for the pending
+ * bit holder.
+ */
+#if _Q_PENDING_BITS == 8
+
+struct __qspinlock {
+	union {
+		atomic_t val;
+		struct {
+#ifdef __LITTLE_ENDIAN
+			u16	locked_pending;
+			u16	tail;
+#else
+			u16	tail;
+			u16	locked_pending;
+#endif
+		};
+	};
+};
+
+/**
+ * clear_pending_set_locked - take ownership and clear the pending bit.
+ * @lock: Pointer to queue spinlock structure
+ * @val : Current value of the queue spinlock 32-bit word
+ *
+ * *,1,0 -> *,0,1
+ *
+ * Lock stealing is not allowed if this function is used.
+ */
+static __always_inline void
+clear_pending_set_locked(struct qspinlock *lock, u32 val)
+{
+	struct __qspinlock *l = (void *)lock;
+
+	WRITE_ONCE(l->locked_pending, _Q_LOCKED_VAL);
+}
+
+/*
+ * xchg_tail - Put in the new queue tail code word & retrieve previous one
+ * @lock : Pointer to queue spinlock structure
+ * @tail : The new queue tail code word
+ * Return: The previous queue tail code word
+ *
+ * xchg(lock, tail)
+ *
+ * p,*,* -> n,*,* ; prev = xchg(lock, node)
+ */
+static __always_inline u32 xchg_tail(struct qspinlock *lock, u32 tail)
+{
+	struct __qspinlock *l = (void *)lock;
+
+	return (u32)xchg(&l->tail, tail >> _Q_TAIL_OFFSET) << _Q_TAIL_OFFSET;
+}
+
+#else /* _Q_PENDING_BITS == 8 */
+
 /**
  * clear_pending_set_locked - take ownership and clear the pending bit.
  * @lock: Pointer to queue spinlock structure
@@ -143,6 +206,7 @@ static __always_inline u32 xchg_tail(str
 	}
 	return old;
 }
+#endif /* _Q_PENDING_BITS == 8 */
 
 /**
  * queue_spin_lock_slowpath - acquire the queue spinlock
@@ -217,8 +281,13 @@ void queue_spin_lock_slowpath(struct qsp
 	 * we're pending, wait for the owner to go away.
 	 *
 	 * *,1,1 -> *,1,0
+	 *
+	 * this wait loop must be a load-acquire such that we match the
+	 * store-release that clears the locked bit and create lock
+	 * sequentiality; this is because not all clear_pending_set_locked()
+	 * implementations imply full barriers.
 	 */
-	while ((val = atomic_read(&lock->val)) & _Q_LOCKED_MASK)
+	while ((val = smp_load_acquire(&lock->val.counter)) & _Q_LOCKED_MASK)
 		cpu_relax();
 
 	/*



  parent reply	other threads:[~2015-03-16 13:36 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-03-16 13:16 [PATCH 0/9] qspinlock stuff -v15 Peter Zijlstra
2015-03-16 13:16 ` [PATCH 1/9] qspinlock: A simple generic 4-byte queue spinlock Peter Zijlstra
2015-03-16 13:16 ` [PATCH 2/9] qspinlock, x86: Enable x86-64 to use " Peter Zijlstra
2015-03-16 13:16 ` [PATCH 3/9] qspinlock: Add pending bit Peter Zijlstra
2015-03-16 13:16 ` [PATCH 4/9] qspinlock: Extract out code snippets for the next patch Peter Zijlstra
2015-03-16 13:16 ` Peter Zijlstra [this message]
2015-03-16 13:16 ` [PATCH 6/9] qspinlock: Use a simple write to grab the lock Peter Zijlstra
2015-03-16 13:16 ` [PATCH 7/9] qspinlock: Revert to test-and-set on hypervisors Peter Zijlstra
2015-03-16 13:16 ` [PATCH 8/9] qspinlock: Generic paravirt support Peter Zijlstra
     [not found]   ` <5509E51D.7040909@hp.com>
2015-03-19 10:12     ` Peter Zijlstra
2015-03-19 12:25       ` Peter Zijlstra
2015-03-19 13:43         ` Peter Zijlstra
2015-03-19 23:25         ` Waiman Long
2015-04-01 16:20         ` Waiman Long
2015-04-01 17:12           ` Peter Zijlstra
2015-04-01 17:42             ` Peter Zijlstra
2015-04-01 18:17               ` Peter Zijlstra
2015-04-01 18:54                 ` Waiman Long
2015-04-01 18:48                   ` Peter Zijlstra
2015-04-01 19:58                     ` Waiman Long
2015-04-01 21:03                       ` Peter Zijlstra
2015-04-02 16:28                         ` Waiman Long
2015-04-02 17:20                           ` Peter Zijlstra
2015-04-02 19:48                             ` Peter Zijlstra
2015-04-03  3:39                               ` Waiman Long
2015-04-03 13:43                               ` Peter Zijlstra
2015-04-01 20:10             ` Waiman Long
2015-03-16 13:16 ` [PATCH 9/9] qspinlock,x86,kvm: Implement KVM support for paravirt qspinlock Peter Zijlstra
     [not found]   ` <550A3863.2060808@hp.com>
2015-03-19 10:01     ` Peter Zijlstra
2015-03-19 21:08       ` Waiman Long
2015-03-20  7:43         ` Raghavendra K T
2015-03-16 14:08 ` [Xen-devel] [PATCH 0/9] qspinlock stuff -v15 David Vrabel
2015-03-18 20:36 ` Waiman Long
2015-03-19 18:01 ` [Xen-devel] " David Vrabel
2015-03-19 18:32   ` Peter Zijlstra
2015-03-25 19:47 ` Konrad Rzeszutek Wilk
2015-03-26 20:21   ` Peter Zijlstra
2015-03-27 14:07     ` Konrad Rzeszutek Wilk
2015-03-30 16:41       ` Waiman Long
2015-03-30 16:25   ` Waiman Long
2015-03-30 16:29     ` Peter Zijlstra
2015-03-30 16:43       ` Waiman Long
2015-03-27  6:40 ` Raghavendra K T

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20150316133112.084664282@infradead.org \
    --to=peterz@infradead.org \
    --cc=Waiman.Long@hp.com \
    --cc=boris.ostrovsky@oracle.com \
    --cc=david.vrabel@citrix.com \
    --cc=doug.hatch@hp.com \
    --cc=hpa@zytor.com \
    --cc=konrad.wilk@oracle.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-arch@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=luto@amacapital.net \
    --cc=mingo@redhat.com \
    --cc=oleg@redhat.com \
    --cc=paolo.bonzini@gmail.com \
    --cc=paulmck@linux.vnet.ibm.com \
    --cc=raghavendra.kt@linux.vnet.ibm.com \
    --cc=riel@redhat.com \
    --cc=scott.norton@hp.com \
    --cc=tglx@linutronix.de \
    --cc=torvalds@linux-foundation.org \
    --cc=virtualization@lists.linux-foundation.org \
    --cc=x86@kernel.org \
    --cc=xen-devel@lists.xenproject.org \
    --subject='Re: [PATCH 5/9] qspinlock: Optimize for smaller NR_CPUS' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).