mem-hotplug: implement get/put_online_mems

kmem_cache_{create,destroy,shrink} need to get a stable value of
cpu/node online mask, because they init/destroy/access per-cpu/node
kmem_cache parts, which can be allocated or destroyed on cpu/mem
hotplug.  To protect against cpu hotplug, these functions use
{get,put}_online_cpus.  However, they do nothing to synchronize with
memory hotplug - taking the slab_mutex does not eliminate the
possibility of race as described in patch 2.

What we need there is something like get_online_cpus, but for memory.
We already have lock_memory_hotplug, which serves for the purpose, but
it's a bit of a hammer right now, because it's backed by a mutex.  As a
result, it imposes some limitations to locking order, which are not
desirable, and can't be used just like get_online_cpus.  That's why in
patch 1 I substitute it with get/put_online_mems, which work exactly
like get/put_online_cpus except they block not cpu, but memory hotplug.

[ v1 can be found at https://lkml.org/lkml/2014/4/6/68.  I NAK'ed it by
  myself, because it used an rw semaphore for get/put_online_mems,
  making them dead lock prune.  ]

This patch (of 2):

{un}lock_memory_hotplug, which is used to synchronize against memory
hotplug, is currently backed by a mutex, which makes it a bit of a
hammer - threads that only want to get a stable value of online nodes
mask won't be able to proceed concurrently.  Also, it imposes some
strong locking ordering rules on it, which narrows down the set of its
usage scenarios.

This patch introduces get/put_online_mems, which are the same as
get/put_online_cpus, but for memory hotplug, i.e.  executing a code
inside a get/put_online_mems section will guarantee a stable value of
online nodes, present pages, etc.

lock_memory_hotplug()/unlock_memory_hotplug() are removed altogether.

Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Tang Chen <tangchen@cn.fujitsu.com>
Cc: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Cc: Toshi Kani <toshi.kani@hp.com>
Cc: Xishi Qiu <qiuxishi@huawei.com>
Cc: Jiang Liu <liuj97@gmail.com>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

1 files changed, 101 insertions, 41 deletions

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index a650db2..2906873 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -46,19 +46,84 @@
 static void generic_online_page(struct page *page);
 
 static online_page_callback_t online_page_callback = generic_online_page;
+static DEFINE_MUTEX(online_page_callback_lock);
 
-DEFINE_MUTEX(mem_hotplug_mutex);
+/* The same as the cpu_hotplug lock, but for memory hotplug. */
+static struct {
+	struct task_struct *active_writer;
+	struct mutex lock; /* Synchronizes accesses to refcount, */
+	/*
+	 * Also blocks the new readers during
+	 * an ongoing mem hotplug operation.
+	 */
+	int refcount;
+
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+	struct lockdep_map dep_map;
+#endif
+} mem_hotplug = {
+	.active_writer = NULL,
+	.lock = __MUTEX_INITIALIZER(mem_hotplug.lock),
+	.refcount = 0,
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+	.dep_map = {.name = "mem_hotplug.lock" },
+#endif
+};
+
+/* Lockdep annotations for get/put_online_mems() and mem_hotplug_begin/end() */
+#define memhp_lock_acquire_read() lock_map_acquire_read(&mem_hotplug.dep_map)
+#define memhp_lock_acquire()      lock_map_acquire(&mem_hotplug.dep_map)
+#define memhp_lock_release()      lock_map_release(&mem_hotplug.dep_map)
+
+void get_online_mems(void)
+{
+	might_sleep();
+	if (mem_hotplug.active_writer == current)
+		return;
+	memhp_lock_acquire_read();
+	mutex_lock(&mem_hotplug.lock);
+	mem_hotplug.refcount++;
+	mutex_unlock(&mem_hotplug.lock);
+
+}
 
-void lock_memory_hotplug(void)
+void put_online_mems(void)
 {
-	mutex_lock(&mem_hotplug_mutex);
+	if (mem_hotplug.active_writer == current)
+		return;
+	mutex_lock(&mem_hotplug.lock);
+
+	if (WARN_ON(!mem_hotplug.refcount))
+		mem_hotplug.refcount++; /* try to fix things up */
+
+	if (!--mem_hotplug.refcount && unlikely(mem_hotplug.active_writer))
+		wake_up_process(mem_hotplug.active_writer);
+	mutex_unlock(&mem_hotplug.lock);
+	memhp_lock_release();
+
 }
 
-void unlock_memory_hotplug(void)
+static void mem_hotplug_begin(void)
 {
-	mutex_unlock(&mem_hotplug_mutex);
+	mem_hotplug.active_writer = current;
+
+	memhp_lock_acquire();
+	for (;;) {
+		mutex_lock(&mem_hotplug.lock);
+		if (likely(!mem_hotplug.refcount))
+			break;
+		__set_current_state(TASK_UNINTERRUPTIBLE);
+		mutex_unlock(&mem_hotplug.lock);
+		schedule();
+	}
 }
 
+static void mem_hotplug_done(void)
+{
+	mem_hotplug.active_writer = NULL;
+	mutex_unlock(&mem_hotplug.lock);
+	memhp_lock_release();
+}
 
 /* add this memory to iomem resource */
 static struct resource *register_memory_resource(u64 start, u64 size)
@@ -727,14 +792,16 @@ int set_online_page_callback(online_page_callback_t callback)
 {
 	int rc = -EINVAL;
 
-	lock_memory_hotplug();
+	get_online_mems();
+	mutex_lock(&online_page_callback_lock);
 
 	if (online_page_callback == generic_online_page) {
 		online_page_callback = callback;
 		rc = 0;
 	}
 
-	unlock_memory_hotplug();
+	mutex_unlock(&online_page_callback_lock);
+	put_online_mems();
 
 	return rc;
 }
@@ -744,14 +811,16 @@ int restore_online_page_callback(online_page_callback_t callback)
 {
 	int rc = -EINVAL;
 
-	lock_memory_hotplug();
+	get_online_mems();
+	mutex_lock(&online_page_callback_lock);
 
 	if (online_page_callback == callback) {
 		online_page_callback = generic_online_page;
 		rc = 0;
 	}
 
-	unlock_memory_hotplug();
+	mutex_unlock(&online_page_callback_lock);
+	put_online_mems();
 
 	return rc;
 }
@@ -899,7 +968,7 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages, int online_typ
 	int ret;
 	struct memory_notify arg;
 
-	lock_memory_hotplug();
+	mem_hotplug_begin();
 	/*
 	 * This doesn't need a lock to do pfn_to_page().
 	 * The section can't be removed here because of the
@@ -907,23 +976,18 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages, int online_typ
 	 */
 	zone = page_zone(pfn_to_page(pfn));
 
+	ret = -EINVAL;
 	if ((zone_idx(zone) > ZONE_NORMAL || online_type == ONLINE_MOVABLE) &&
-	    !can_online_high_movable(zone)) {
-		unlock_memory_hotplug();
-		return -EINVAL;
-	}
+	    !can_online_high_movable(zone))
+		goto out;
 
 	if (online_type == ONLINE_KERNEL && zone_idx(zone) == ZONE_MOVABLE) {
-		if (move_pfn_range_left(zone - 1, zone, pfn, pfn + nr_pages)) {
-			unlock_memory_hotplug();
-			return -EINVAL;
-		}
+		if (move_pfn_range_left(zone - 1, zone, pfn, pfn + nr_pages))
+			goto out;
 	}
 	if (online_type == ONLINE_MOVABLE && zone_idx(zone) == ZONE_MOVABLE - 1) {
-		if (move_pfn_range_right(zone, zone + 1, pfn, pfn + nr_pages)) {
-			unlock_memory_hotplug();
-			return -EINVAL;
-		}
+		if (move_pfn_range_right(zone, zone + 1, pfn, pfn + nr_pages))
+			goto out;
 	}
 
 	/* Previous code may changed the zone of the pfn range */
@@ -939,8 +1003,7 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages, int online_typ
 	ret = notifier_to_errno(ret);
 	if (ret) {
 		memory_notify(MEM_CANCEL_ONLINE, &arg);
-		unlock_memory_hotplug();
-		return ret;
+		goto out;
 	}
 	/*
 	 * If this zone is not populated, then it is not in zonelist.
@@ -964,8 +1027,7 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages, int online_typ
 		       (((unsigned long long) pfn + nr_pages)
 			    << PAGE_SHIFT) - 1);
 		memory_notify(MEM_CANCEL_ONLINE, &arg);
-		unlock_memory_hotplug();
-		return ret;
+		goto out;
 	}
 
 	zone->present_pages += onlined_pages;
@@ -995,9 +1057,9 @@ int __ref online_pages(unsigned long pfn, unsigned long nr_pages, int online_typ
 
 	if (onlined_pages)
 		memory_notify(MEM_ONLINE, &arg);
-	unlock_memory_hotplug();
-
-	return 0;
+out:
+	mem_hotplug_done();
+	return ret;
 }
 #endif /* CONFIG_MEMORY_HOTPLUG_SPARSE */
 
@@ -1055,7 +1117,7 @@ int try_online_node(int nid)
 	if (node_online(nid))
 		return 0;
 
-	lock_memory_hotplug();
+	mem_hotplug_begin();
 	pgdat = hotadd_new_pgdat(nid, 0);
 	if (!pgdat) {
 		pr_err("Cannot online node %d due to NULL pgdat\n", nid);
@@ -1073,7 +1135,7 @@ int try_online_node(int nid)
 	}
 
 out:
-	unlock_memory_hotplug();
+	mem_hotplug_done();
 	return ret;
 }
 
@@ -1117,7 +1179,7 @@ int __ref add_memory(int nid, u64 start, u64 size)
 		new_pgdat = !p;
 	}
 
-	lock_memory_hotplug();
+	mem_hotplug_begin();
 
 	new_node = !node_online(nid);
 	if (new_node) {
@@ -1158,7 +1220,7 @@ error:
 	release_memory_resource(res);
 
 out:
-	unlock_memory_hotplug();
+	mem_hotplug_done();
 	return ret;
 }
 EXPORT_SYMBOL_GPL(add_memory);
@@ -1565,7 +1627,7 @@ static int __ref __offline_pages(unsigned long start_pfn,
 	if (!test_pages_in_a_zone(start_pfn, end_pfn))
 		return -EINVAL;
 
-	lock_memory_hotplug();
+	mem_hotplug_begin();
 
 	zone = page_zone(pfn_to_page(start_pfn));
 	node = zone_to_nid(zone);
@@ -1672,7 +1734,7 @@ repeat:
 	writeback_set_ratelimit();
 
 	memory_notify(MEM_OFFLINE, &arg);
-	unlock_memory_hotplug();
+	mem_hotplug_done();
 	return 0;
 
 failed_removal:
@@ -1684,7 +1746,7 @@ failed_removal:
 	undo_isolate_page_range(start_pfn, end_pfn, MIGRATE_MOVABLE);
 
 out:
-	unlock_memory_hotplug();
+	mem_hotplug_done();
 	return ret;
 }
 
@@ -1888,7 +1950,7 @@ void __ref remove_memory(int nid, u64 start, u64 size)
 
 	BUG_ON(check_hotplug_memory_range(start, size));
 
-	lock_memory_hotplug();
+	mem_hotplug_begin();
 
 	/*
 	 * All memory blocks must be offlined before removing memory.  Check
@@ -1897,10 +1959,8 @@ void __ref remove_memory(int nid, u64 start, u64 size)
 	 */
 	ret = walk_memory_range(PFN_DOWN(start), PFN_UP(start + size - 1), NULL,
 				check_memblock_offlined_cb);
-	if (ret) {
-		unlock_memory_hotplug();
+	if (ret)
 		BUG();
-	}
 
 	/* remove memmap entry */
 	firmware_map_remove(start, start + size, "System RAM");
@@ -1909,7 +1969,7 @@ void __ref remove_memory(int nid, u64 start, u64 size)
 
 	try_offline_node(nid);
 
-	unlock_memory_hotplug();
+	mem_hotplug_done();
 }
 EXPORT_SYMBOL_GPL(remove_memory);
 #endif /* CONFIG_MEMORY_HOTREMOVE */