From: Rick Lindsley To access all the system's disk statitics we currently need to access one sysfs file per disk. This clearly will not be acceptable with thousands of disks. The patch aggregates the system-wide statistics in real time and exposes them via /proc/diskstats drivers/block/genhd.c | 86 ++++++++++++++++++++++++++++++++++++++++++++++++++ fs/proc/proc_misc.c | 13 +++++++ 2 files changed, 99 insertions(+) diff -puN drivers/block/genhd.c~aggregated-disk-stats drivers/block/genhd.c --- 25/drivers/block/genhd.c~aggregated-disk-stats 2003-04-09 00:32:19.000000000 -0700 +++ 25-akpm/drivers/block/genhd.c 2003-04-09 00:32:19.000000000 -0700 @@ -532,6 +532,92 @@ static struct kobj_type ktype_block = { static decl_subsys(block,&ktype_block); +/* + * aggregate disk stat collector. Uses the same stats that the sysfs + * entries do, above, but makes them available through one seq_file. + * Watching a few disks may be efficient through sysfs, but watching + * all of them will be more efficient through this interface. + * + * The output looks suspiciously like /proc/partitions with a bunch of + * extra fields. + */ + +/* iterator */ +static void *diskstats_start(struct seq_file *part, loff_t *pos) +{ + loff_t k = *pos; + struct list_head *p; + + down_read(&block_subsys.rwsem); + list_for_each(p, &block_subsys.kset.list) + if (!k--) + return list_entry(p, struct gendisk, kobj.entry); + return NULL; +} + +static void *diskstats_next(struct seq_file *part, void *v, loff_t *pos) +{ + struct list_head *p = ((struct gendisk *)v)->kobj.entry.next; + ++*pos; + return p==&block_subsys.kset.list ? NULL : + list_entry(p, struct gendisk, kobj.entry); +} + +static void diskstats_stop(struct seq_file *part, void *v) +{ + up_read(&block_subsys.rwsem); +} + +static int diskstats_show(struct seq_file *s, void *v) +{ + struct gendisk *gp = v; + char buf[64]; + int n = 0; + + /* + if (&sgp->kobj.entry == block_subsys.kset.list.next) + seq_puts(s, "major minor name" + " rio rmerge rsect ruse wio wmerge " + "wsect wuse running use aveq" + "\n\n"); + */ + + disk_round_stats(gp); + seq_printf(s, "%4d %4d %s %u %u %llu %u %u %u %llu %u %u %u %u\n", + gp->major, n + gp->first_minor, disk_name(gp, n, buf), + disk_stat_read(gp, reads), disk_stat_read(gp, read_merges), + (unsigned long long)disk_stat_read(gp, read_sectors), + jiffies_to_msec(disk_stat_read(gp, read_ticks)), + disk_stat_read(gp, writes), disk_stat_read(gp, write_merges), + (unsigned long long)disk_stat_read(gp, write_sectors), + jiffies_to_msec(disk_stat_read(gp, write_ticks)), + disk_stat_read(gp, in_flight), + jiffies_to_msec(disk_stat_read(gp, io_ticks)), + jiffies_to_msec(disk_stat_read(gp, time_in_queue))); + + /* now show all non-0 size partitions of it */ + for (n = 0; n < gp->minors - 1; n++) { + struct hd_struct *hd = gp->part[n]; + + if (hd && hd->nr_sects) + seq_printf(s, "%4d %4d %s %u %u %u %u\n", + gp->major, n + gp->first_minor + 1, + disk_name(gp, n + 1, buf), + hd->reads, hd->read_sectors, + hd->writes, hd->write_sectors); + } + + return 0; +} + +struct seq_operations diskstats_op = { + start: diskstats_start, + next: diskstats_next, + stop: diskstats_stop, + show: diskstats_show +}; + + struct gendisk *alloc_disk(int minors) { struct gendisk *disk = kmalloc(sizeof(struct gendisk), GFP_KERNEL); diff -puN fs/proc/proc_misc.c~aggregated-disk-stats fs/proc/proc_misc.c --- 25/fs/proc/proc_misc.c~aggregated-disk-stats 2003-04-09 00:32:19.000000000 -0700 +++ 25-akpm/fs/proc/proc_misc.c 2003-04-09 00:32:19.000000000 -0700 @@ -345,6 +345,18 @@ static struct file_operations proc_parti .release = seq_release, }; +extern struct seq_operations diskstats_op; +static int diskstats_open(struct inode *inode, struct file *file) +{ + return seq_open(file, &diskstats_op); +} +static struct file_operations proc_diskstats_operations = { + open: diskstats_open, + read: seq_read, + llseek: seq_lseek, + release: seq_release, +}; + #ifdef CONFIG_MODULES extern struct seq_operations modules_op; static int modules_open(struct inode *inode, struct file *file) @@ -657,6 +669,7 @@ void __init proc_misc_init(void) create_seq_entry("slabinfo",S_IWUSR|S_IRUGO,&proc_slabinfo_operations); create_seq_entry("buddyinfo",S_IRUGO, &fragmentation_file_operations); create_seq_entry("vmstat",S_IRUGO, &proc_vmstat_file_operations); + create_seq_entry("diskstats", 0, &proc_diskstats_operations); #ifdef CONFIG_MODULES create_seq_entry("modules", 0, &proc_modules_operations); #endif _