linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Tejun Heo <tj@kernel.org>
To: mingo@elte.hu, awalls@radix.net, linux-kernel@vger.kernel.org,
	jeff@garzik.org, akpm@linux-foundation.org,
	rusty@rustcorp.com.au, cl@linux-foundation.org,
	dhowells@redhat.com, arjan@linux.intel.com,
	johannes@sipsolutions.net, oleg@redhat.com, axboe@kernel.dk
Cc: Tejun Heo <tj@kernel.org>
Subject: [PATCH 22/30] workqueue: implement WQ_NON_REENTRANT
Date: Mon, 14 Jun 2010 23:37:39 +0200	[thread overview]
Message-ID: <1276551467-21246-23-git-send-email-tj@kernel.org> (raw)
In-Reply-To: <1276551467-21246-1-git-send-email-tj@kernel.org>

With gcwq managing all the workers and work->data pointing to the last
gcwq it was on, non-reentrance can be easily implemented by checking
whether the work is still running on the previous gcwq on queueing.
Implement it.

Signed-off-by: Tejun Heo <tj@kernel.org>
---
 include/linux/workqueue.h |    1 +
 kernel/workqueue.c        |   32 +++++++++++++++++++++++++++++---
 2 files changed, 30 insertions(+), 3 deletions(-)

diff --git a/include/linux/workqueue.h b/include/linux/workqueue.h
index 0a78141..07cf5e5 100644
--- a/include/linux/workqueue.h
+++ b/include/linux/workqueue.h
@@ -225,6 +225,7 @@ static inline unsigned int work_static(struct work_struct *work) { return 0; }
 enum {
 	WQ_FREEZEABLE		= 1 << 0, /* freeze during suspend */
 	WQ_SINGLE_CPU		= 1 << 1, /* only single cpu at a time */
+	WQ_NON_REENTRANT	= 1 << 2, /* guarantee non-reentrance */
 };
 
 extern struct workqueue_struct *
diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index f606c44..7994edb 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -534,11 +534,37 @@ static void __queue_work(unsigned int cpu, struct workqueue_struct *wq,
 
 	debug_work_activate(work);
 
-	/* determine gcwq to use */
+	/*
+	 * Determine gcwq to use.  SINGLE_CPU is inherently
+	 * NON_REENTRANT, so test it first.
+	 */
 	if (!(wq->flags & WQ_SINGLE_CPU)) {
-		/* just use the requested cpu for multicpu workqueues */
+		struct global_cwq *last_gcwq;
+
+		/*
+		 * It's multi cpu.  If @wq is non-reentrant and @work
+		 * was previously on a different cpu, it might still
+		 * be running there, in which case the work needs to
+		 * be queued on that cpu to guarantee non-reentrance.
+		 */
 		gcwq = get_gcwq(cpu);
-		spin_lock_irqsave(&gcwq->lock, flags);
+		if (wq->flags & WQ_NON_REENTRANT &&
+		    (last_gcwq = get_work_gcwq(work)) && last_gcwq != gcwq) {
+			struct worker *worker;
+
+			spin_lock_irqsave(&last_gcwq->lock, flags);
+
+			worker = find_worker_executing_work(last_gcwq, work);
+
+			if (worker && worker->current_cwq->wq == wq)
+				gcwq = last_gcwq;
+			else {
+				/* meh... not running there, queue here */
+				spin_unlock_irqrestore(&last_gcwq->lock, flags);
+				spin_lock_irqsave(&gcwq->lock, flags);
+			}
+		} else
+			spin_lock_irqsave(&gcwq->lock, flags);
 	} else {
 		unsigned int req_cpu = cpu;
 
-- 
1.6.4.2


  parent reply	other threads:[~2010-06-14 21:46 UTC|newest]

Thread overview: 129+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-06-14 21:37 [PATCHSET] workqueue: concurrency managed workqueue, take#5 Tejun Heo
2010-06-14 21:37 ` [PATCH 01/30] kthread: implement kthread_data() Tejun Heo
2010-06-14 21:37 ` [PATCH 02/30] acpi: use queue_work_on() instead of binding workqueue worker to cpu0 Tejun Heo
2010-06-14 21:37 ` [PATCH 03/30] workqueue: kill RT workqueue Tejun Heo
2010-06-14 21:37 ` [PATCH 04/30] workqueue: misc/cosmetic updates Tejun Heo
2010-06-14 21:37 ` [PATCH 05/30] workqueue: merge feature parameters into flags Tejun Heo
2010-06-14 21:37 ` [PATCH 06/30] workqueue: define masks for work flags and conditionalize STATIC flags Tejun Heo
2010-06-14 21:37 ` [PATCH 07/30] workqueue: separate out process_one_work() Tejun Heo
2010-06-14 21:37 ` [PATCH 08/30] workqueue: temporarily disable workqueue tracing Tejun Heo
2010-06-15 13:29   ` Frederic Weisbecker
2010-06-15 16:37     ` Tejun Heo
2010-06-14 21:37 ` [PATCH 09/30] workqueue: kill cpu_populated_map Tejun Heo
2010-06-14 21:37 ` [PATCH 10/30] workqueue: update cwq alignement Tejun Heo
2010-06-14 21:37 ` [PATCH 11/30] workqueue: reimplement workqueue flushing using color coded works Tejun Heo
2010-06-14 21:37 ` [PATCH 12/30] workqueue: introduce worker Tejun Heo
2010-06-14 21:37 ` [PATCH 13/30] workqueue: reimplement work flushing using linked works Tejun Heo
2010-06-14 21:37 ` [PATCH 14/30] workqueue: implement per-cwq active work limit Tejun Heo
2010-06-14 21:37 ` [PATCH 15/30] workqueue: reimplement workqueue freeze using max_active Tejun Heo
2010-06-14 21:37 ` [PATCH 16/30] workqueue: introduce global cwq and unify cwq locks Tejun Heo
2010-06-14 21:37 ` [PATCH 17/30] workqueue: implement worker states Tejun Heo
2010-06-14 21:37 ` [PATCH 18/30] workqueue: reimplement CPU hotplugging support using trustee Tejun Heo
2010-06-14 21:37 ` [PATCH 19/30] workqueue: make single thread workqueue shared worker pool friendly Tejun Heo
2010-06-14 21:37 ` [PATCH 20/30] workqueue: add find_worker_executing_work() and track current_cwq Tejun Heo
2010-06-14 21:37 ` [PATCH 21/30] workqueue: carry cpu number in work data once execution starts Tejun Heo
2010-06-14 21:37 ` Tejun Heo [this message]
2010-06-14 21:37 ` [PATCH 23/30] workqueue: use shared worklist and pool all workers per cpu Tejun Heo
2010-06-14 21:37 ` [PATCH 24/30] workqueue: implement concurrency managed dynamic worker pool Tejun Heo
2010-06-14 21:37 ` [PATCH 25/30] workqueue: increase max_active of keventd and kill current_is_keventd() Tejun Heo
2010-06-14 21:37 ` [PATCH 26/30] workqueue: add system_wq, system_long_wq and system_nrt_wq Tejun Heo
2010-06-14 21:37 ` [PATCH 27/30] workqueue: implement DEBUGFS/workqueue Tejun Heo
2010-06-15 13:54   ` Frederic Weisbecker
2010-06-15 16:42     ` Tejun Heo
2010-06-14 21:37 ` [PATCH 28/30] workqueue: implement several utility APIs Tejun Heo
2010-06-14 21:37 ` [PATCH 29/30] libata: take advantage of cmwq and remove concurrency limitations Tejun Heo
2010-06-14 21:37 ` [PATCH 30/30] async: use workqueue for worker pool Tejun Heo
2010-06-14 21:58 ` [PATCHSET] workqueue: concurrency managed workqueue, take#5 Andrew Morton
2010-06-14 22:17   ` Tejun Heo
2010-06-14 22:31     ` Daniel Walker
2010-06-14 22:33       ` Tejun Heo
2010-06-14 22:35         ` Daniel Walker
2010-06-14 22:44           ` Tejun Heo
2010-06-14 22:49             ` Daniel Walker
2010-06-14 22:52               ` Tejun Heo
2010-06-14 22:35     ` Andrew Morton
2010-06-14 22:43       ` Tejun Heo
2010-06-14 23:06         ` Andrew Morton
2010-06-15 12:53         ` tytso
2010-06-15 16:15           ` [PATCH] SubmittingPatches: add more about patch descriptions Randy Dunlap
2010-06-15 16:33             ` Christoph Lameter
2010-06-15 18:15   ` [PATCHSET] workqueue: concurrency managed workqueue, take#5 Stefan Richter
2010-06-15 19:39     ` Tejun Heo
2010-06-15  1:20 ` Jeff Garzik
2010-06-15 18:25 ` Overview of concurrency managed workqueue Tejun Heo
2010-06-15 18:40   ` Christoph Lameter
2010-06-15 18:44     ` Tejun Heo
2010-06-15 19:43   ` Daniel Walker
2010-06-16 12:10     ` Tejun Heo
2010-06-16 13:27       ` Daniel Walker
2010-06-16 13:30         ` Tejun Heo
2010-06-16 13:41           ` Daniel Walker
2010-06-16 13:45             ` Tejun Heo
2010-06-16 14:05               ` Daniel Walker
2010-06-16 14:15                 ` Tejun Heo
2010-06-16 14:34                   ` Daniel Walker
2010-06-16 14:50                     ` Tejun Heo
2010-06-16 15:11                       ` Daniel Walker
2010-06-16 15:50                         ` Tejun Heo
2010-06-16 16:30                           ` Daniel Walker
2010-06-16 16:55                             ` Tejun Heo
2010-06-16 18:22                               ` Daniel Walker
2010-06-16 18:46                                 ` Tejun Heo
2010-06-16 19:20                                   ` Tejun Heo
2010-06-16 19:46                                     ` Daniel Walker
2010-06-16 19:58                                       ` Tejun Heo
2010-06-17  5:29                                     ` Florian Mickler
2010-06-17  6:21                                       ` Florian Mickler
2010-06-17  8:28                                       ` Tejun Heo
2010-06-17 18:03                                       ` Daniel Walker
2010-06-18  6:36                                         ` Florian Mickler
2010-06-18 16:38                                           ` Daniel Walker
2010-06-16 19:36                                   ` Daniel Walker
2010-06-16 19:52                                     ` Tejun Heo
2010-06-16 20:19                                       ` Daniel Walker
2010-06-16 20:24                                         ` Tejun Heo
2010-06-16 20:40                                           ` Daniel Walker
2010-06-16 21:41                                             ` Tejun Heo
2010-06-17 23:15                               ` Andrew Morton
2010-06-18  8:03                                 ` Tejun Heo
2010-06-18  8:22                                   ` Tejun Heo
2010-06-18 17:29                                   ` Daniel Walker
2010-06-16 18:31                     ` Stefan Richter
2010-06-16 18:41                       ` Daniel Walker
2010-06-17 12:01                 ` Andy Walls
2010-06-17 16:56                   ` Daniel Walker
2010-06-17 23:16                   ` Andrew Morton
2010-06-18  7:16                     ` Tejun Heo
2010-06-18  7:31                       ` Andrew Morton
2010-06-18  8:09                         ` Tejun Heo
2010-06-18 17:02                           ` Andrew Morton
2010-06-18 17:28                             ` Tejun Heo
2010-06-19 15:53                               ` [PATCH] kthread: implement kthread_worker Tejun Heo
2010-06-21 20:33                                 ` Randy Dunlap
2010-06-22  7:31                                   ` Tejun Heo
2010-06-19  8:38                   ` Overview of concurrency managed workqueue Andi Kleen
2010-06-19  8:40                     ` Tejun Heo
2010-06-19  8:55                       ` Andi Kleen
2010-06-19  9:01                         ` Tejun Heo
2010-06-19  9:08                           ` Andi Kleen
2010-06-19  9:12                             ` Tejun Heo
2010-06-19  9:15                               ` Andi Kleen
2010-06-19  9:17                                 ` Tejun Heo
2010-06-19  9:27                                   ` Andi Kleen
2010-06-19  9:42                                     ` Tejun Heo
2010-06-19 12:20                                       ` Andi Kleen
2010-06-19 12:48                                         ` Tejun Heo
2010-06-17 22:28               ` Daniel Walker
2010-06-16  6:55   ` Florian Mickler
2010-06-16 12:22     ` Tejun Heo
2010-06-16 13:37   ` Johannes Berg
2010-06-16 13:39     ` Tejun Heo
2010-06-16 13:42       ` Johannes Berg
2010-06-17 23:14   ` Andrew Morton
2010-06-17 23:25     ` Joel Becker
2010-06-17 23:56       ` Andrew Morton
2010-06-18  7:15         ` Tejun Heo
2010-06-18  7:31     ` Tejun Heo
2010-06-15 18:29 ` [PATCHSET] workqueue: concurrency managed workqueue, take#5 Stefan Richter
2010-06-15 18:40   ` Tejun Heo
2010-06-15 20:29   ` Stefan Richter

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=1276551467-21246-23-git-send-email-tj@kernel.org \
    --to=tj@kernel.org \
    --cc=akpm@linux-foundation.org \
    --cc=arjan@linux.intel.com \
    --cc=awalls@radix.net \
    --cc=axboe@kernel.dk \
    --cc=cl@linux-foundation.org \
    --cc=dhowells@redhat.com \
    --cc=jeff@garzik.org \
    --cc=johannes@sipsolutions.net \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=oleg@redhat.com \
    --cc=rusty@rustcorp.com.au \
    /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).