hdr_recl calls zthr_wakeup() on destroyed zthr

There exists a race condition were hdr_recl() calls
zthr_wakeup() on a destroyed zthr. The timeline is the
following:

[1] hdr_recl() runs first and goes intro zthr_wakeup()
    because arc_initialized is set.
[2] arc_fini() is called by another thread, zeroes
    that flag, destroying the zthr, and goes into
    buf_init().
[3] hdr_recl() tries to enter the destroyed mutex
    and we blow up.

This patch ensures that the ARC's zthrs are not offloaded
any new work once arc_initialized is set and then destroys
them after all of the ARC state has been deleted.

Reviewed by: Matt Ahrens <[email protected]>
Reviewed by: Brian Behlendorf <[email protected]>
Signed-off-by: Serapheim Dimitropoulos <[email protected]>
Closes #9047 

1 files changed, 16 insertions, 4 deletions

diff --git a/module/zfs/arc.c b/module/zfs/arc.c
index 98bafeee2..90a731bff 100644
--- a/module/zfs/arc.c
+++ b/module/zfs/arc.c
@@ -21,7 +21,7 @@
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
  * Copyright (c) 2018, Joyent, Inc.
- * Copyright (c) 2011, 2018 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2019 by Delphix. All rights reserved.
  * Copyright (c) 2014 by Saso Kiselkov. All rights reserved.
  * Copyright 2017 Nexenta Systems, Inc.  All rights reserved.
  */
@@ -5086,6 +5086,9 @@ arc_kmem_reap_soon(void)
 static boolean_t
 arc_adjust_cb_check(void *arg, zthr_t *zthr)
 {
+	if (!arc_initialized)
+		return (B_FALSE);
+
 	/*
 	 * This is necessary so that any changes which may have been made to
 	 * many of the zfs_arc_* module parameters will be propagated to
@@ -5173,6 +5176,9 @@ arc_adjust_cb(void *arg, zthr_t *zthr)
 static boolean_t
 arc_reap_cb_check(void *arg, zthr_t *zthr)
 {
+	if (!arc_initialized)
+		return (B_FALSE);
+
 	int64_t free_memory = arc_available_memory();
 
 	/*
@@ -7933,11 +7939,9 @@ arc_fini(void)
 
 	list_destroy(&arc_prune_list);
 	mutex_destroy(&arc_prune_mtx);
-	(void) zthr_cancel(arc_adjust_zthr);
-	zthr_destroy(arc_adjust_zthr);
 
+	(void) zthr_cancel(arc_adjust_zthr);
 	(void) zthr_cancel(arc_reap_zthr);
-	zthr_destroy(arc_reap_zthr);
 
 	mutex_destroy(&arc_adjust_lock);
 	cv_destroy(&arc_adjust_waiters_cv);
@@ -7950,6 +7954,14 @@ arc_fini(void)
 	buf_fini();
 	arc_state_fini();
 
+	/*
+	 * We destroy the zthrs after all the ARC state has been
+	 * torn down to avoid the case of them receiving any
+	 * wakeup() signals after they are destroyed.
+	 */
+	zthr_destroy(arc_adjust_zthr);
+	zthr_destroy(arc_reap_zthr);
+
 	ASSERT0(arc_loaned_bytes);
 }