md: add blktrace event for writes to superblock

superblock write is an expensive operation. With raid5-cache, it can be called
regularly. Tracing to help performance debug.

Signed-off-by: Shaohua Li <shli@fb.com>
Cc: NeilBrown <neilb@suse.com>

diff --git a/drivers/md/md.c b/drivers/md/md.c
index 1f1c7f007b683ab6b50b8ca0bef505da8b04f23b..d3cef771e4225104e7f5df5ce30fab8135e0f985 100644 (file)
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -64,6 +64,7 @@
 #include <linux/raid/md_p.h>
 #include <linux/raid/md_u.h>
 #include <linux/slab.h>
+#include <trace/events/block.h>
 #include "md.h"
 #include "bitmap.h"
 #include "md-cluster.h"
@@ -2403,6 +2404,8 @@ repeat:
        pr_debug("md: updating %s RAID superblock on device (in sync %d)\n",
                 mdname(mddev), mddev->in_sync);
 
+       if (mddev->queue)
+               blk_add_trace_msg(mddev->queue, "md md_update_sb");
        bitmap_update_sb(mddev->bitmap);
        rdev_for_each(rdev, mddev) {
                char b[BDEVNAME_SIZE];