The patch titled
fix flush_workqueue() vs CPU_DEAD race
has been added to the -mm tree. Its filename is
fix-flush_workqueue-vs-cpu_dead-race.patch
See http://www.zip.com.au/~akpm/linux/patches/stuff/added-to-mm.txt to find
out what to do about this
------------------------------------------------------
Subject: fix flush_workqueue() vs CPU_DEAD race
From: Oleg Nesterov <oleg@xxxxxxxxxx>
"[PATCH 1/2] reimplement flush_workqueue()" fixed one race when CPU goes
down while flush_cpu_workqueue() plays with it. But there is another
problem, CPU can die before flush_workqueue() has a chance to call
flush_cpu_workqueue(). In that case pending work_structs can migrate to
CPU which was already checked, so we should redo the
"for_each_online_cpu(cpu)" loop.
Signed-off-by: Oleg Nesterov <oleg@xxxxxxxxxx>
Cc: Gautham R Shenoy <ego@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxx>
---
kernel/workqueue.c | 8 ++++++++
1 files changed, 8 insertions(+)
diff -puN kernel/workqueue.c~fix-flush_workqueue-vs-cpu_dead-race kernel/workqueue.c
--- a/kernel/workqueue.c~fix-flush_workqueue-vs-cpu_dead-race
+++ a/kernel/workqueue.c
@@ -65,6 +65,7 @@ struct workqueue_struct {
/* All the per-cpu workqueues on the system, for hotplug cpu to add/remove
threads to each one as cpus come/go. */
+static long hotplug_sequence __read_mostly;
static DEFINE_MUTEX(workqueue_mutex);
static LIST_HEAD(workqueues);
@@ -454,10 +455,16 @@ void fastcall flush_workqueue(struct wor
/* Always use first cpu's area. */
flush_cpu_workqueue(per_cpu_ptr(wq->cpu_wq, singlethread_cpu));
} else {
+ long sequence;
int cpu;
+again:
+ sequence = hotplug_sequence;
for_each_online_cpu(cpu)
flush_cpu_workqueue(per_cpu_ptr(wq->cpu_wq, cpu));
+
+ if (unlikely(sequence != hotplug_sequence))
+ goto again;
}
mutex_unlock(&workqueue_mutex);
}
@@ -880,6 +887,7 @@ static int __devinit workqueue_cpu_callb
cleanup_workqueue_thread(wq, hotcpu);
list_for_each_entry(wq, &workqueues, list)
take_over_work(wq, hotcpu);
+ hotplug_sequence++;
break;
case CPU_LOCK_RELEASE:
_
Patches currently in -mm which might be from oleg@xxxxxxxxxx are
clone-flag-clone_parent_tidptr-leaves-invalid-results-in-memory.patch
doc-atomic_add_unless-doesnt-imply-mb-on-failure.patch
vt-refactor-console-sak-processing.patch
procfs-fix-race-between-proc_readdir-and-remove_proc_entry.patch
procfs-fix-race-between-proc_readdir-and-remove_proc_entry-fix.patch
kill_pid_info-kill-acquired_tasklist_lock.patch
tty-make-__proc_set_tty-static.patch
tty-clarify-disassociate_ctty.patch
tty-fix-the-locking-for-signal-session-in-disassociate_ctty.patch
signal-use-kill_pgrp-not-kill_pg-in-the-sunos-compatibility-code.patch
signal-rewrite-kill_something_info-so-it-uses-newer-helpers.patch
pid-make-session_of_pgrp-use-struct-pid-instead-of-pid_t.patch
pid-use-struct-pid-for-talking-about-process-groups-in-exitc.patch
pid-replace-is_orphaned_pgrp-with-is_current_pgrp_orphaned.patch
tty-update-the-tty-layer-to-work-with-struct-pid.patch
pid-replace-do-while_each_task_pid-with-do-while_each_pid_task.patch
pid-remove-now-unused-do_each_task_pid-and-while_each_task_pid.patch
pid-remove-the-now-unused-kill_pg-kill_pg_info-and-__kill_pg_info.patch
reimplement-flush_workqueue.patch
implement-flush_work.patch
implement-flush_work-sanity.patch
implement-flush_work_keventd.patch
aio-use-flush_work.patch
kblockd-use-flush_work.patch
relayfs-use-flush_keventd_work.patch
tg3-use-flush_keventd_work.patch
e1000-use-flush_keventd_work.patch
libata-use-flush_work.patch
phy-use-flush_work.patch
fix-flush_workqueue-vs-cpu_dead-race.patch
-
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
