mm: turn vmap_purge_lock into a mutex

The purge_lock spinlock causes high latencies with non RT kernel.  This
has been reported multiple times on lkml [1] [2] and affects
applications like audio.

This patch replaces it with a mutex to allow preemption while holding
the lock.

Thanks to Joel Fernandes for the detailed report and analysis as well as
an earlier attempt at fixing this issue.

[1] http://lists.openwall.net/linux-kernel/2016/03/23/29
[2] https://lkml.org/lkml/2016/10/9/59

Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Christoph Hellwig <[email protected]>
Tested-by: Jisheng Zhang <[email protected]>
Cc: Andrey Ryabinin <[email protected]>
Cc: Joel Fernandes <[email protected]>
Cc: Chris Wilson <[email protected]>
Cc: John Dias <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: "H. Peter Anvin" <[email protected]>
Cc: Ingo Molnar <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

mm/vmalloc.c

@@ -606,7 +606,7 @@ static atomic_t vmap_lazy_nr = ATOMIC_INIT(0);
  * by this look, but we want to avoid concurrent calls for performance
  * reasons and to make the pcpu_get_vm_areas more deterministic.
  */
-static DEFINE_SPINLOCK(vmap_purge_lock);
+static DEFINE_MUTEX(vmap_purge_lock);
 
 /* for per-CPU blocks */
 static void purge_fragmented_blocks_allcpus(void);
@@ -660,9 +660,9 @@ static bool __purge_vmap_area_lazy(unsigned long start, unsigned long end)
  */
 static void try_purge_vmap_area_lazy(void)
 {
-	if (spin_trylock(&vmap_purge_lock)) {
+	if (mutex_trylock(&vmap_purge_lock)) {
 		__purge_vmap_area_lazy(ULONG_MAX, 0);
-		spin_unlock(&vmap_purge_lock);
+		mutex_unlock(&vmap_purge_lock);
 	}
 }
 
@@ -671,10 +671,10 @@ static void try_purge_vmap_area_lazy(void)
  */
 static void purge_vmap_area_lazy(void)
 {
-	spin_lock(&vmap_purge_lock);
+	mutex_lock(&vmap_purge_lock);
 	purge_fragmented_blocks_allcpus();
 	__purge_vmap_area_lazy(ULONG_MAX, 0);
-	spin_unlock(&vmap_purge_lock);
+	mutex_unlock(&vmap_purge_lock);
 }
 
 /*
@@ -1063,11 +1063,11 @@ void vm_unmap_aliases(void)
 		rcu_read_unlock();
 	}
 
-	spin_lock(&vmap_purge_lock);
+	mutex_lock(&vmap_purge_lock);
 	purge_fragmented_blocks_allcpus();
 	if (!__purge_vmap_area_lazy(start, end) && flush)
 		flush_tlb_kernel_range(start, end);
-	spin_unlock(&vmap_purge_lock);
+	mutex_unlock(&vmap_purge_lock);
 }
 EXPORT_SYMBOL_GPL(vm_unmap_aliases);