xen_balloon: support memory auto onlining policy

Add support for the newly added kernel memory auto onlining policy to
Xen ballon driver.

Signed-off-by: Vitaly Kuznetsov <[email protected]>
Suggested-by: Daniel Kiper <[email protected]>
Reviewed-by: Daniel Kiper <[email protected]>
Acked-by: David Vrabel <[email protected]>
Cc: Jonathan Corbet <[email protected]>
Cc: Greg Kroah-Hartman <[email protected]>
Cc: Daniel Kiper <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Tang Chen <[email protected]>
Cc: David Rientjes <[email protected]>
Cc: Naoya Horiguchi <[email protected]>
Cc: Xishi Qiu <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: "K. Y. Srinivasan" <[email protected]>
Cc: Igor Mammedov <[email protected]>
Cc: Kay Sievers <[email protected]>
Cc: Konrad Rzeszutek Wilk <[email protected]>
Cc: Boris Ostrovsky <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

2 files changed, 25 insertions, 9 deletions

diff --git a/drivers/xen/Kconfig b/drivers/xen/Kconfig
index 73708acce3ca..979a8317204f 100644
--- a/drivers/xen/Kconfig
+++ b/drivers/xen/Kconfig
@@ -37,23 +37,30 @@ config XEN_BALLOON_MEMORY_HOTPLUG
 
 	  Memory could be hotplugged in following steps:
 
-	    1) dom0: xl mem-max <domU> <maxmem>
+	    1) target domain: ensure that memory auto online policy is in
+	       effect by checking /sys/devices/system/memory/auto_online_blocks
+	       file (should be 'online').
+
+	    2) control domain: xl mem-max <target-domain> <maxmem>
 	       where <maxmem> is >= requested memory size,
 
-	    2) dom0: xl mem-set <domU> <memory>
+	    3) control domain: xl mem-set <target-domain> <memory>
 	       where <memory> is requested memory size; alternatively memory
 	       could be added by writing proper value to
 	       /sys/devices/system/xen_memory/xen_memory0/target or
-	       /sys/devices/system/xen_memory/xen_memory0/target_kb on dumU,
+	       /sys/devices/system/xen_memory/xen_memory0/target_kb on the
+	       target domain.
 
-	    3) domU: for i in /sys/devices/system/memory/memory*/state; do \
-	               [ "`cat "$i"`" = offline ] && echo online > "$i"; done
+	  Alternatively, if memory auto onlining was not requested at step 1
+	  the newly added memory can be manually onlined in the target domain
+	  by doing the following:
 
-	  Memory could be onlined automatically on domU by adding following line to udev rules:
+		for i in /sys/devices/system/memory/memory*/state; do \
+		  [ "`cat "$i"`" = offline ] && echo online > "$i"; done
 
-	  SUBSYSTEM=="memory", ACTION=="add", RUN+="/bin/sh -c '[ -f /sys$devpath/state ] && echo online > /sys$devpath/state'"
+	  or by adding the following line to udev rules:
 
-	  In that case step 3 should be omitted.
+	  SUBSYSTEM=="memory", ACTION=="add", RUN+="/bin/sh -c '[ -f /sys$devpath/state ] && echo online > /sys$devpath/state'"
 
 config XEN_BALLOON_MEMORY_HOTPLUG_LIMIT
 	int "Hotplugged memory limit (in GiB) for a PV guest"
diff --git a/drivers/xen/balloon.c b/drivers/xen/balloon.c
index e6058debd01b..7c8a2cf16f58 100644
--- a/drivers/xen/balloon.c
+++ b/drivers/xen/balloon.c
@@ -338,7 +338,16 @@ static enum bp_state reserve_additional_memory(void)
 	}
 #endif
 
-	rc = add_memory_resource(nid, resource, false);
+	/*
+	 * add_memory_resource() will call online_pages() which in its turn
+	 * will call xen_online_page() callback causing deadlock if we don't
+	 * release balloon_mutex here. Unlocking here is safe because the
+	 * callers drop the mutex before trying again.
+	 */
+	mutex_unlock(&balloon_mutex);
+	rc = add_memory_resource(nid, resource, memhp_auto_online);
+	mutex_lock(&balloon_mutex);
+
 	if (rc) {
 		pr_warn("Cannot add additional memory (%i)\n", rc);
 		goto err;