f2fs: add tracepoint for f2fs iostat

Added a tracepoint to see iostat of f2fs. Default period of that
is 3 second. This tracepoint can be used to be monitoring
I/O statistics periodically.
Signed-off-by: Daeho Jeong <daehojeong@google.com>
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>

Documentation/ABI/testing/sysfs-fs-f2fs

@@ -332,3 +332,9 @@ Description:	Give a way to attach REQ_META|FUA to data writes
 		*      REQ_META     |      REQ_FUA      |
 		*    5 |    4 |   3 |    2 |    1 |   0 |
 		* Cold | Warm | Hot | Cold | Warm | Hot |
+
+What:		/sys/fs/f2fs/<disk>/iostat_period_ms
+Date:		April 2020
+Contact:	"Daeho Jeong" <daehojeong@google.com>
+Description:	Give a way to change iostat_period time. 3secs by default.
+		The new iostat trace gives stats gap given the period.

fs/f2fs/f2fs.h

@@ -1505,7 +1505,10 @@ struct f2fs_sb_info {
 	/* For app/fs IO statistics */
 	spinlock_t iostat_lock;
 	unsigned long long write_iostat[NR_IO_TYPE];
+	unsigned long long prev_write_iostat[NR_IO_TYPE];
 	bool iostat_enable;
+	unsigned long iostat_next_period;
+	unsigned int iostat_period_ms;
 
 	/* to attach REQ_META|REQ_FUA flags */
 	unsigned int data_io_flag;
@@ -2999,16 +3002,25 @@ static inline int get_inline_xattr_addrs(struct inode *inode)
 		sizeof((f2fs_inode)->field))			\
 		<= (F2FS_OLD_ATTRIBUTE_SIZE + (extra_isize)))	\
 
+#define DEFAULT_IOSTAT_PERIOD_MS	3000
+#define MIN_IOSTAT_PERIOD_MS		100
+/* maximum period of iostat tracing is 1 day */
+#define MAX_IOSTAT_PERIOD_MS		8640000
+
 static inline void f2fs_reset_iostat(struct f2fs_sb_info *sbi)
 {
 	int i;
 
 	spin_lock(&sbi->iostat_lock);
-	for (i = 0; i < NR_IO_TYPE; i++)
+	for (i = 0; i < NR_IO_TYPE; i++) {
 		sbi->write_iostat[i] = 0;
+		sbi->prev_write_iostat[i] = 0;
+	}
 	spin_unlock(&sbi->iostat_lock);
 }
 
+extern void f2fs_record_iostat(struct f2fs_sb_info *sbi);
+
 static inline void f2fs_update_iostat(struct f2fs_sb_info *sbi,
 			enum iostat_type type, unsigned long long io_bytes)
 {
@@ -3022,6 +3034,8 @@ static inline void f2fs_update_iostat(struct f2fs_sb_info *sbi,
 			sbi->write_iostat[APP_WRITE_IO] -
 			sbi->write_iostat[APP_DIRECT_IO];
 	spin_unlock(&sbi->iostat_lock);
+
+	f2fs_record_iostat(sbi);
 }
 
 #define __is_large_section(sbi)		((sbi)->segs_per_sec > 1)

fs/f2fs/super.c

@@ -3424,6 +3424,7 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
 	/* init iostat info */
 	spin_lock_init(&sbi->iostat_lock);
 	sbi->iostat_enable = false;
+	sbi->iostat_period_ms = DEFAULT_IOSTAT_PERIOD_MS;
 
 	for (i = 0; i < NR_PAGE_TYPE; i++) {
 		int n = (i == META) ? 1: NR_TEMP_TYPE;

fs/f2fs/sysfs.c

@@ -15,6 +15,7 @@
 #include "f2fs.h"
 #include "segment.h"
 #include "gc.h"
+#include <trace/events/f2fs.h>
 
 static struct proc_dir_entry *f2fs_proc_root;
 
@@ -379,6 +380,15 @@ static ssize_t __sbi_store(struct f2fs_attr *a,
 		return count;
 	}
 
+	if (!strcmp(a->attr.name, "iostat_period_ms")) {
+		if (t < MIN_IOSTAT_PERIOD_MS || t > MAX_IOSTAT_PERIOD_MS)
+			return -EINVAL;
+		spin_lock(&sbi->iostat_lock);
+		sbi->iostat_period_ms = (unsigned int)t;
+		spin_unlock(&sbi->iostat_lock);
+		return count;
+	}
+
 	*ui = (unsigned int)t;
 
 	return count;
@@ -535,6 +545,7 @@ F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, gc_idle_interval, interval_time[GC_TIME]);
 F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info,
 		umount_discard_timeout, interval_time[UMOUNT_DISCARD_TIMEOUT]);
 F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, iostat_enable, iostat_enable);
+F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, iostat_period_ms, iostat_period_ms);
 F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, readdir_ra, readdir_ra);
 F2FS_RW_ATTR(F2FS_SBI, f2fs_sb_info, gc_pin_file_thresh, gc_pin_file_threshold);
 F2FS_RW_ATTR(F2FS_SBI, f2fs_super_block, extension_list, extension_list);
@@ -615,6 +626,7 @@ static struct attribute *f2fs_attrs[] = {
 	ATTR_LIST(gc_idle_interval),
 	ATTR_LIST(umount_discard_timeout),
 	ATTR_LIST(iostat_enable),
+	ATTR_LIST(iostat_period_ms),
 	ATTR_LIST(readdir_ra),
 	ATTR_LIST(gc_pin_file_thresh),
 	ATTR_LIST(extension_list),
@@ -751,6 +763,33 @@ static int __maybe_unused segment_bits_seq_show(struct seq_file *seq,
 	return 0;
 }
 
+void f2fs_record_iostat(struct f2fs_sb_info *sbi)
+{
+	unsigned long long iostat_diff[NR_IO_TYPE];
+	int i;
+
+	if (time_is_after_jiffies(sbi->iostat_next_period))
+		return;
+
+	/* Need double check under the lock */
+	spin_lock(&sbi->iostat_lock);
+	if (time_is_after_jiffies(sbi->iostat_next_period)) {
+		spin_unlock(&sbi->iostat_lock);
+		return;
+	}
+	sbi->iostat_next_period = jiffies +
+				msecs_to_jiffies(sbi->iostat_period_ms);
+
+	for (i = 0; i < NR_IO_TYPE; i++) {
+		iostat_diff[i] = sbi->write_iostat[i] -
+				sbi->prev_write_iostat[i];
+		sbi->prev_write_iostat[i] = sbi->write_iostat[i];
+	}
+	spin_unlock(&sbi->iostat_lock);
+
+	trace_f2fs_iostat(sbi, iostat_diff);
+}
+
 static int __maybe_unused iostat_info_seq_show(struct seq_file *seq,
 					       void *offset)
 {

include/trace/events/f2fs.h

@@ -1812,6 +1812,58 @@ DEFINE_EVENT(f2fs_zip_end, f2fs_decompress_pages_end,
 	TP_ARGS(inode, cluster_idx, compressed_size, ret)
 );
 
+TRACE_EVENT(f2fs_iostat,
+
+	TP_PROTO(struct f2fs_sb_info *sbi, unsigned long long *iostat),
+
+	TP_ARGS(sbi, iostat),
+
+	TP_STRUCT__entry(
+		__field(dev_t,	dev)
+		__field(unsigned long long,	app_dio)
+		__field(unsigned long long,	app_bio)
+		__field(unsigned long long,	app_wio)
+		__field(unsigned long long,	app_mio)
+		__field(unsigned long long,	fs_dio)
+		__field(unsigned long long,	fs_nio)
+		__field(unsigned long long,	fs_mio)
+		__field(unsigned long long,	fs_gc_dio)
+		__field(unsigned long long,	fs_gc_nio)
+		__field(unsigned long long,	fs_cp_dio)
+		__field(unsigned long long,	fs_cp_nio)
+		__field(unsigned long long,	fs_cp_mio)
+		__field(unsigned long long,	fs_discard)
+	),
+
+	TP_fast_assign(
+		__entry->dev		= sbi->sb->s_dev;
+		__entry->app_dio	= iostat[APP_DIRECT_IO];
+		__entry->app_bio	= iostat[APP_BUFFERED_IO];
+		__entry->app_wio	= iostat[APP_WRITE_IO];
+		__entry->app_mio	= iostat[APP_MAPPED_IO];
+		__entry->fs_dio		= iostat[FS_DATA_IO];
+		__entry->fs_nio		= iostat[FS_NODE_IO];
+		__entry->fs_mio		= iostat[FS_META_IO];
+		__entry->fs_gc_dio	= iostat[FS_GC_DATA_IO];
+		__entry->fs_gc_nio	= iostat[FS_GC_NODE_IO];
+		__entry->fs_cp_dio	= iostat[FS_CP_DATA_IO];
+		__entry->fs_cp_nio	= iostat[FS_CP_NODE_IO];
+		__entry->fs_cp_mio	= iostat[FS_CP_META_IO];
+		__entry->fs_discard	= iostat[FS_DISCARD];
+	),
+
+	TP_printk("dev = (%d,%d), "
+		"app [write=%llu (direct=%llu, buffered=%llu), mapped=%llu], "
+		"fs [data=%llu, node=%llu, meta=%llu, discard=%llu], "
+		"gc [data=%llu, node=%llu], "
+		"cp [data=%llu, node=%llu, meta=%llu]",
+		show_dev(__entry->dev), __entry->app_wio, __entry->app_dio,
+		__entry->app_bio, __entry->app_mio, __entry->fs_dio,
+		__entry->fs_nio, __entry->fs_mio, __entry->fs_discard,
+		__entry->fs_gc_dio, __entry->fs_gc_nio, __entry->fs_cp_dio,
+		__entry->fs_cp_nio, __entry->fs_cp_mio)
+);
+
 #endif /* _TRACE_F2FS_H */
 
  /* This part must be outside protection */