[PATCH v3 2/6] fs: add optional iostats counters to struct super_block

[Amir Goldstein <amir73il@gmail.com>]

With CONFIG_FS_IOSTATS, filesystem can attach an array of counters to
struct super_block by calling sb_iostats_init().

These counters will be used to collect per-sb I/O statistics and display
them in /proc/<pid>/mountstats.

Signed-off-by: Amir Goldstein <amir73il@gmail.com>
---
 fs/Kconfig                 |   8 +++
 fs/super.c                 |   2 +
 include/linux/fs.h         |  10 ++-
 include/linux/fs_iostats.h | 127 +++++++++++++++++++++++++++++++++++++
 4 files changed, 145 insertions(+), 2 deletions(-)
 create mode 100644 include/linux/fs_iostats.h

diff --git a/fs/Kconfig b/fs/Kconfig
index 6c7dc1387beb..394d9da6bda9 100644
--- a/fs/Kconfig
+++ b/fs/Kconfig
@@ -15,6 +15,14 @@ config VALIDATE_FS_PARSER
 	  Enable this to perform validation of the parameter description for a
 	  filesystem when it is registered.
 
+config FS_IOSTATS
+	bool "Enable generic filesystem I/O statistics"
+	help
+	  Enable this to allow collecting filesystem I/O statistics and display
+	  them in /proc/<pid>/mountstats.
+
+	  Say N if unsure.
+
 config FS_IOMAP
 	bool
 
diff --git a/fs/super.c b/fs/super.c
index f1d4a193602d..c447cadb402b 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -36,6 +36,7 @@
 #include <linux/lockdep.h>
 #include <linux/user_namespace.h>
 #include <linux/fs_context.h>
+#include <linux/fs_iostats.h>
 #include <uapi/linux/mount.h>
 #include "internal.h"
 
@@ -290,6 +291,7 @@ static void __put_super(struct super_block *s)
 		WARN_ON(s->s_dentry_lru.node);
 		WARN_ON(s->s_inode_lru.node);
 		WARN_ON(!list_empty(&s->s_mounts));
+		sb_iostats_destroy(s);
 		security_sb_free(s);
 		fscrypt_sb_free(s);
 		put_user_ns(s->s_user_ns);
diff --git a/include/linux/fs.h b/include/linux/fs.h
index e2d892b201b0..a71c94cbb6c1 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -1454,6 +1454,8 @@ struct sb_writers {
 	struct percpu_rw_semaphore	rw_sem[SB_FREEZE_LEVELS];
 };
 
+struct sb_iostats;
+
 struct super_block {
 	struct list_head	s_list;		/* Keep this first */
 	dev_t			s_dev;		/* search index; _not_ kdev_t */
@@ -1508,8 +1510,12 @@ struct super_block {
 	/* Granularity of c/m/atime in ns (cannot be worse than a second) */
 	u32			s_time_gran;
 	/* Time limits for c/m/atime in seconds */
-	time64_t		   s_time_min;
-	time64_t		   s_time_max;
+	time64_t		s_time_min;
+	time64_t		s_time_max;
+#ifdef CONFIG_FS_IOSTATS
+	/* Optional per-sb I/O stats */
+	struct sb_iostats	*s_iostats;
+#endif
 #ifdef CONFIG_FSNOTIFY
 	__u32			s_fsnotify_mask;
 	struct fsnotify_mark_connector __rcu	*s_fsnotify_marks;
diff --git a/include/linux/fs_iostats.h b/include/linux/fs_iostats.h
new file mode 100644
index 000000000000..60d1efbea7d9
--- /dev/null
+++ b/include/linux/fs_iostats.h
@@ -0,0 +1,127 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _LINUX_FS_IOSTATS_H
+#define _LINUX_FS_IOSTATS_H
+
+#include <linux/fs.h>
+#include <linux/percpu_counter.h>
+#include <linux/slab.h>
+
+/* Similar to task_io_accounting members */
+enum {
+	SB_IOSTATS_CHARS_RD,	/* bytes read via syscalls */
+	SB_IOSTATS_CHARS_WR,	/* bytes written via syscalls */
+	SB_IOSTATS_SYSCALLS_RD,	/* # of read syscalls */
+	SB_IOSTATS_SYSCALLS_WR,	/* # of write syscalls */
+	SB_IOSTATS_COUNTERS_NUM
+};
+
+struct sb_iostats {
+	time64_t		start_time;
+	struct percpu_counter	counter[SB_IOSTATS_COUNTERS_NUM];
+};
+
+#ifdef CONFIG_FS_IOSTATS
+static inline struct sb_iostats *sb_iostats(struct super_block *sb)
+{
+	return sb->s_iostats;
+}
+
+static inline bool sb_has_iostats(struct super_block *sb)
+{
+	return !!sb->s_iostats;
+}
+
+static inline void sb_iostats_destroy(struct super_block *sb)
+{
+	if (!sb->s_iostats)
+		return;
+
+	percpu_counters_destroy(sb->s_iostats->counter,
+				SB_IOSTATS_COUNTERS_NUM);
+	kfree(sb->s_iostats);
+	sb->s_iostats = NULL;
+}
+
+/* Initialize per-sb I/O stats */
+static inline int sb_iostats_init(struct super_block *sb)
+{
+	int err;
+
+	sb->s_iostats = kmalloc(sizeof(struct sb_iostats), GFP_KERNEL);
+	if (!sb->s_iostats)
+		return -ENOMEM;
+
+	err = percpu_counters_init(sb->s_iostats->counter,
+				   SB_IOSTATS_COUNTERS_NUM, 0, GFP_KERNEL);
+	if (err) {
+		kfree(sb->s_iostats);
+		sb->s_iostats = NULL;
+		return err;
+	}
+
+	sb->s_iostats->start_time = ktime_get_seconds();
+	return 0;
+}
+
+static inline void sb_iostats_counter_inc(struct super_block *sb, int id)
+{
+	if (!sb->s_iostats)
+		return;
+
+	percpu_counter_inc(&sb->s_iostats->counter[id]);
+}
+
+static inline void sb_iostats_counter_add(struct super_block *sb, int id,
+					  s64 amt)
+{
+	if (!sb->s_iostats)
+		return;
+
+	percpu_counter_add(&sb->s_iostats->counter[id], amt);
+}
+
+static inline s64 sb_iostats_counter_read(struct super_block *sb, int id)
+{
+	if (!sb->s_iostats)
+		return 0;
+
+	return percpu_counter_read_positive(&sb->s_iostats->counter[id]);
+}
+
+#else /* !CONFIG_FS_IOSTATS */
+
+static inline struct sb_iostats *sb_iostats(struct super_block *sb)
+{
+	return NULL;
+}
+
+static inline bool sb_has_iostats(struct super_block *sb)
+{
+	return false;
+}
+
+static inline int sb_iostats_init(struct super_block *sb)
+{
+	return -EOPNOTSUPP;
+}
+
+static inline void sb_iostats_destroy(struct super_block *sb)
+{
+}
+
+static inline void sb_iostats_counter_inc(struct super_block *sb, int id)
+{
+}
+
+static inline void sb_iostats_counter_add(struct super_block *sb, int id,
+					  s64 amt)
+{
+}
+
+static inline s64 sb_iostats_counter_read(struct super_block *sb, int id)
+{
+	return 0;
+}
+#endif
+
+#endif /* _LINUX_FS_IOSTATS_H */
-- 
2.25.1