linux-arch.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: Waiman.Long@hp.com, tglx@linutronix.de, mingo@kernel.org
Cc: linux-arch@vger.kernel.org, linux-kernel@vger.kernel.org,
	virtualization@lists.linux-foundation.org,
	xen-devel@lists.xenproject.org, kvm@vger.kernel.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, gleb@redhat.com, scott.norton@hp.com,
	chegu_vinod@hp.com, Peter Zijlstra <peterz@infradead.org>
Subject: [PATCH 04/11] qspinlock: Extract out the exchange of tail code word
Date: Sun, 15 Jun 2014 14:47:01 +0200	[thread overview]
Message-ID: <20140615130153.376621956@chello.nl> (raw)
In-Reply-To: 20140615124657.264658593@chello.nl

[-- Attachment #1: waiman_long-qspinlock-extract_out_the_exchange_of_tail_code_word.patch --]
[-- Type: text/plain, Size: 3002 bytes --]

From: Waiman Long <Waiman.Long@hp.com>

This patch extracts the logic for the exchange of new and previous tail
code words into a new xchg_tail() function which can be optimized in a
later patch.

Signed-off-by: Waiman Long <Waiman.Long@hp.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
---
 include/asm-generic/qspinlock_types.h |    2 +
 kernel/locking/qspinlock.c            |   58 +++++++++++++++++++++-------------
 2 files changed, 38 insertions(+), 22 deletions(-)

--- a/include/asm-generic/qspinlock_types.h
+++ b/include/asm-generic/qspinlock_types.h
@@ -61,6 +61,8 @@ 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_MASK		(_Q_TAIL_IDX_MASK | _Q_TAIL_CPU_MASK)
+
 #define _Q_LOCKED_VAL		(1U << _Q_LOCKED_OFFSET)
 #define _Q_PENDING_VAL		(1U << _Q_PENDING_OFFSET)
 
--- a/kernel/locking/qspinlock.c
+++ b/kernel/locking/qspinlock.c
@@ -86,6 +86,31 @@ static inline struct mcs_spinlock *decod
 #define _Q_LOCKED_PENDING_MASK	(_Q_LOCKED_MASK | _Q_PENDING_MASK)
 
 /**
+ * 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)
+{
+	u32 old, new, val = atomic_read(&lock->val);
+
+	for (;;) {
+		new = (val & _Q_LOCKED_PENDING_MASK) | tail;
+		old = atomic_cmpxchg(&lock->val, val, new);
+		if (old == val)
+			break;
+
+		val = old;
+	}
+	return old;
+}
+
+/**
  * queue_spin_lock_slowpath - acquire the queue spinlock
  * @lock: Pointer to queue spinlock structure
  * @val: Current value of the queue spinlock 32-bit word
@@ -182,36 +207,25 @@ void queue_spin_lock_slowpath(struct qsp
 	node->next = NULL;
 
 	/*
-	 * we already touched the queueing cacheline; don't bother with pending
-	 * stuff.
-	 *
-	 * trylock || xchg(lock, node)
-	 *
-	 * 0,0,0 -> 0,0,1 ; trylock
-	 * p,y,x -> n,y,x ; prev = xchg(lock, node)
+	 * We touched a (possibly) cold cacheline in the per-cpu queue node;
+	 * attempt the trylock once more in the hope someone let go while we
+	 * weren't watching.
 	 */
-	for (;;) {
-		new = _Q_LOCKED_VAL;
-		if (val)
-			new = tail | (val & _Q_LOCKED_PENDING_MASK);
-
-		old = atomic_cmpxchg(&lock->val, val, new);
-		if (old == val)
-			break;
-
-		val = old;
-	}
+	if (queue_spin_trylock(lock))
+		goto release;
 
 	/*
-	 * we won the trylock; forget about queueing.
+	 * we already touched the queueing cacheline; don't bother with pending
+	 * stuff.
+	 *
+	 * p,*,* -> n,*,*
 	 */
-	if (new == _Q_LOCKED_VAL)
-		goto release;
+	old = xchg_tail(lock, tail);
 
 	/*
 	 * if there was a previous node; link it and wait.
 	 */
-	if (old & ~_Q_LOCKED_PENDING_MASK) {
+	if (old & _Q_TAIL_MASK) {
 		prev = decode_tail(old);
 		ACCESS_ONCE(prev->next) = node;
 

  parent reply	other threads:[~2014-06-15 13:11 UTC|newest]

Thread overview: 120+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-06-15 12:46 [PATCH 00/11] qspinlock with paravirt support Peter Zijlstra
2014-06-15 12:46 ` Peter Zijlstra
2014-06-15 12:46 ` [PATCH 01/11] qspinlock: A simple generic 4-byte queue spinlock Peter Zijlstra
2014-06-15 12:46   ` Peter Zijlstra
2014-06-16 20:49   ` Konrad Rzeszutek Wilk
2014-06-16 20:49     ` Konrad Rzeszutek Wilk
2014-06-17 20:03     ` Konrad Rzeszutek Wilk
2014-06-17 20:03       ` Konrad Rzeszutek Wilk
2014-06-23 16:12       ` Peter Zijlstra
2014-06-23 16:12         ` Peter Zijlstra
2014-06-23 16:20         ` Konrad Rzeszutek Wilk
2014-06-23 16:20           ` Konrad Rzeszutek Wilk
2014-06-23 15:56     ` Peter Zijlstra
2014-06-23 16:16       ` Konrad Rzeszutek Wilk
2014-06-23 16:16         ` Konrad Rzeszutek Wilk
2014-06-17 20:05   ` Konrad Rzeszutek Wilk
2014-06-17 20:05     ` Konrad Rzeszutek Wilk
2014-06-23 16:26     ` Peter Zijlstra
2014-06-23 16:26       ` Peter Zijlstra
2014-06-23 16:45       ` Konrad Rzeszutek Wilk
2014-06-23 16:45         ` Konrad Rzeszutek Wilk
2014-06-15 12:46 ` [PATCH 02/11] qspinlock, x86: Enable x86-64 to use " Peter Zijlstra
2014-06-15 12:46   ` Peter Zijlstra
2014-06-15 12:47 ` [PATCH 03/11] qspinlock: Add pending bit Peter Zijlstra
2014-06-15 12:47   ` Peter Zijlstra
2014-06-17 20:36   ` Konrad Rzeszutek Wilk
2014-06-17 20:36     ` Konrad Rzeszutek Wilk
2014-06-17 20:51     ` Waiman Long
2014-06-17 20:51       ` Waiman Long
2014-06-17 21:07       ` Konrad Rzeszutek Wilk
2014-06-17 21:07         ` Konrad Rzeszutek Wilk
2014-06-17 21:10         ` Konrad Rzeszutek Wilk
2014-06-17 21:10           ` Konrad Rzeszutek Wilk
2014-06-17 22:25           ` Waiman Long
2014-06-17 22:25             ` Waiman Long
2014-06-24  8:24         ` Peter Zijlstra
2014-06-24  8:24           ` Peter Zijlstra
2014-06-18 11:29     ` Paolo Bonzini
2014-06-18 11:29       ` Paolo Bonzini
2014-06-18 13:36       ` Konrad Rzeszutek Wilk
2014-06-18 13:36         ` Konrad Rzeszutek Wilk
2014-06-23 16:35     ` Peter Zijlstra
2014-06-23 16:35       ` Peter Zijlstra
2014-06-15 12:47 ` Peter Zijlstra [this message]
2014-06-17 20:55   ` [PATCH 04/11] qspinlock: Extract out the exchange of tail code word Konrad Rzeszutek Wilk
2014-06-17 20:55     ` Konrad Rzeszutek Wilk
2014-06-18 11:37     ` Paolo Bonzini
2014-06-18 11:37       ` Paolo Bonzini
2014-06-18 13:50       ` Konrad Rzeszutek Wilk
2014-06-18 13:50         ` Konrad Rzeszutek Wilk
2014-06-18 15:46         ` Waiman Long
2014-06-18 15:46           ` Waiman Long
2014-06-18 15:49           ` Paolo Bonzini
2014-06-18 15:49             ` Paolo Bonzini
2014-06-18 16:02           ` Konrad Rzeszutek Wilk
2014-06-18 16:02             ` Konrad Rzeszutek Wilk
2014-06-24 10:47       ` Peter Zijlstra
2014-06-24 10:47         ` Peter Zijlstra
2014-06-15 12:47 ` [PATCH 05/11] qspinlock: Optimize for smaller NR_CPUS Peter Zijlstra
2014-06-15 12:47   ` Peter Zijlstra
2014-06-18 11:39   ` Paolo Bonzini
2014-06-18 11:39     ` Paolo Bonzini
2014-07-07 14:35     ` Peter Zijlstra
2014-07-07 14:35       ` Peter Zijlstra
2014-07-07 15:08       ` Paolo Bonzini
2014-07-07 15:08         ` Paolo Bonzini
2014-07-07 15:35         ` Peter Zijlstra
2014-07-07 15:35           ` Peter Zijlstra
2014-07-07 16:10           ` Paolo Bonzini
2014-07-07 16:10             ` Paolo Bonzini
2014-06-18 15:57   ` Konrad Rzeszutek Wilk
2014-06-18 15:57     ` Konrad Rzeszutek Wilk
2014-07-07 14:33     ` Peter Zijlstra
2014-07-07 14:33       ` Peter Zijlstra
2014-06-15 12:47 ` [PATCH 06/11] qspinlock: Optimize pending bit Peter Zijlstra
2014-06-15 12:47   ` Peter Zijlstra
2014-06-18 11:42   ` Paolo Bonzini
2014-06-18 11:42     ` Paolo Bonzini
2014-06-15 12:47 ` [PATCH 07/11] qspinlock: Use a simple write to grab the lock, if applicable Peter Zijlstra
2014-06-15 12:47   ` Peter Zijlstra
2014-06-18 16:36   ` Konrad Rzeszutek Wilk
2014-06-18 16:36     ` Konrad Rzeszutek Wilk
2014-07-07 14:51     ` Peter Zijlstra
2014-07-07 14:51       ` Peter Zijlstra
2014-06-15 12:47 ` [PATCH 08/11] qspinlock: Revert to test-and-set on hypervisors Peter Zijlstra
2014-06-15 12:47   ` Peter Zijlstra
2014-06-16 21:57   ` Waiman Long
2014-06-18 16:40   ` Konrad Rzeszutek Wilk
2014-06-18 16:40     ` Konrad Rzeszutek Wilk
2014-06-15 12:47 ` [PATCH 09/11] pvqspinlock, x86: Rename paravirt_ticketlocks_enabled Peter Zijlstra
2014-06-15 12:47   ` Peter Zijlstra
2014-06-18 16:43   ` Konrad Rzeszutek Wilk
2014-06-18 16:43     ` Konrad Rzeszutek Wilk
2014-06-15 12:47 ` [PATCH 10/11] qspinlock: Paravirt support Peter Zijlstra
2014-06-15 12:47   ` Peter Zijlstra
2014-06-16 22:08   ` Waiman Long
2014-06-18 12:03     ` Paolo Bonzini
2014-06-18 12:03       ` Paolo Bonzini
2014-06-18 15:26       ` Waiman Long
2014-06-18 15:26         ` Waiman Long
2014-07-07 15:20       ` Peter Zijlstra
2014-07-07 15:20         ` Peter Zijlstra
2014-07-07 15:20     ` Peter Zijlstra
2014-07-07 15:20       ` Peter Zijlstra
2014-06-17  0:53   ` Waiman Long
2014-06-17  0:53     ` Waiman Long
2014-06-18 12:04   ` Paolo Bonzini
2014-06-18 12:04     ` Paolo Bonzini
2014-06-20 13:46   ` Konrad Rzeszutek Wilk
2014-06-20 13:46     ` Konrad Rzeszutek Wilk
2014-07-07 15:27     ` Peter Zijlstra
2014-07-15 14:23       ` Konrad Rzeszutek Wilk
2014-07-15 14:23         ` Konrad Rzeszutek Wilk
2014-06-15 12:47 ` [PATCH 11/11] qspinlock, kvm: Add paravirt support Peter Zijlstra
2014-06-22 16:36   ` Raghavendra K T
2014-06-22 16:36     ` Raghavendra K T
2014-07-07 15:23     ` Peter Zijlstra
2014-07-07 15:23       ` Peter Zijlstra
2014-06-16 20:52 ` [PATCH 00/11] qspinlock with " Konrad Rzeszutek Wilk
2014-06-16 20:52   ` Konrad Rzeszutek Wilk

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=20140615130153.376621956@chello.nl \
    --to=a.p.zijlstra@chello.nl \
    --cc=Waiman.Long@hp.com \
    --cc=boris.ostrovsky@oracle.com \
    --cc=chegu_vinod@hp.com \
    --cc=david.vrabel@citrix.com \
    --cc=gleb@redhat.com \
    --cc=konrad.wilk@oracle.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-arch@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=oleg@redhat.com \
    --cc=paolo.bonzini@gmail.com \
    --cc=paulmck@linux.vnet.ibm.com \
    --cc=peterz@infradead.org \
    --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=xen-devel@lists.xenproject.org \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).