block: Separating discards from writes in Linux IO statistics
Subsystems (10)
cover block: Separating discards from writes in Linux IO statistics
This patch set separates block layer statistics for discards from writes. Discards are currently bundled with writes in the various /sys/block/*/stat files as well as in /proc/diskstats. However discards are nearly always used to mark storage that is no longer in use. There are many reasons having discard not counted with writes is useful:
1) For many non volatile memory devices it is just nice to know that discards are enabled and working properly.
2) Discards have different performance characteristics than writes. They are generally much faster and larger and bundling them makes performance statistics less meaningful.
3) Discards are not writes in terms of tracking device lifetime. If a device supports six device writes per day it is nice to know how many writes have actually been written to the device as discards do not count against that total.
Separation of discard statistics is accomplished by expanding the struct diskstat arrays to 3 entries for STAT_READ, STAT_WRITE, and STAT_DISCARD. A new rw_stat_group function is then used to convert from rw_flags (cmd_flags from requests, bi_rw from bios) into the appropriate stat group which is then tracked as before. Lastly the new statistics are appended to the current /sys/bloc/*/stat and /proc/diskstats on output such that they are the last four entries of each. These are analogous to the four read and write statistics.
* Number of discard ios completed
* Number of discard ios merged
* Number of discard sectors completed
* Milliseconds spent on discard requests.
[before ~]# cat sys/block/nvme0/stat
296550701 0 2372405688 67317193 19672752 0 7972237312
9375167 0 2787238 79718726
[after ~]# cat sys/block/nvme0/stat
296550701 0 2372405688 67317193 18034352 0 4616794112
9125902 0 2787238 79718726 1638400 0 3355443200
249265
Note that the discards have moved out of the write fields to the end and that the write fields are now smaller by the difference.
Adding the new statistics to the end of /sys/block/*/stat and /proc/diskstats is backwards compatible with both iostat and vmstat which pick up just the old fields:
[root@after ~]# iostat -x
Linux 4.5.0_68319_ge5065f4-dirty (##hostname###) 05/17/2016
_x86_64_ (48 CPU)
avg-cpu: %user %nice %system %iowait %steal %idle
0.41 0.00 0.23 0.01 0.00 99.35
Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s
avgrq-sz avgqu-sz await svctm %util
sda 0.01 7.50 0.20 3.61 16.65 708.41 190.03 0.09 22.53 1.16 0.44 nvme0n1 0.00 0.00 587.03 35.70 4696.25 9139.08 22.22 0.16 0.24 0.01 0.55
[root@after ~]# vmstat -d
disk- ------------reads------------ ------------writes----------- -----IO------
total merged sectors ms total merged sectors ms cur sec
ram0 0 0 0 0 0 0 0 0 0 0
ram1 0 0 0 0 0 0 0 0 0 0
ram2 0 0 0 0 0 0 0 0 0 0
ram3 0 0 0 0 0 0 0 0 0 0
ram4 0 0 0 0 0 0 0 0 0 0
ram5 0 0 0 0 0 0 0 0 0 0
ram6 0 0 0 0 0 0 0 0 0 0
ram7 0 0 0 0 0 0 0 0 0 0
ram8 0 0 0 0 0 0 0 0 0 0
ram9 0 0 0 0 0 0 0 0 0 0
ram10 0 0 0 0 0 0 0 0 0 0
ram11 0 0 0 0 0 0 0 0 0 0
ram12 0 0 0 0 0 0 0 0 0 0
ram13 0 0 0 0 0 0 0 0 0 0
ram14 0 0 0 0 0 0 0 0 0 0
ram15 0 0 0 0 0 0 0 0 0 0
sda 102903 5247 8408420 47424 1820306 3782041 357153613 43320121
0 2228
nvme0n1 145633279 0 1165066312 18981333 13107200 0
3355443200 6663655 0 1796
loop0 0 0 0 0 0 0 0 0 0 0
loop1 0 0 0 0 0 0 0 0 0 0[chop rest of loop devices]
Comments and suggestions welcome.
Michael Callahan (4):
block: Add part_stat_read_accum to read across field entries.
block: Define and use STAT_READ and STAT_WRITE
block: Add and use a rw_stat_group function for indexing disk_stat
fields.
block: Track DISCARD statistics and output them in stat and diskstat.
Documentation/ABI/testing/procfs-diskstats | 10 ++++++++++ Documentation/block/stat.txt | 28 ++++++++++++++++------------ Documentation/iostats.txt | 15 ++++++++++++++- block/bio.c | 14 +++++++++----- block/blk-core.c | 9 +++++---- block/genhd.c | 30 +++++++++++++++++++----------- block/partition-generic.c | 25 +++++++++++++++---------- drivers/block/drbd/drbd_receiver.c | 3 +-- drivers/block/drbd/drbd_req.c | 4 ++-- drivers/block/drbd/drbd_worker.c | 4 +--- drivers/block/rsxx/dev.c | 5 ++--- drivers/block/zram/zram_drv.c | 14 +++++++------- drivers/md/bcache/request.c | 10 +++++----- drivers/md/dm.c | 7 +++---- drivers/md/md.c | 8 ++++---- drivers/nvdimm/core.c | 8 +++++--- fs/ext4/super.c | 5 +++-- fs/ext4/sysfs.c | 6 ++++-- fs/f2fs/f2fs.h | 2 +- fs/f2fs/super.c | 3 ++- include/linux/bio.h | 4 ++-- include/linux/genhd.h | 24 ++++++++++++++++++++---- 22 files changed, 150 insertions(+), 88 deletions(-)
Signature · 1 line
1.9.3
PATCH 1/4 block: Add part_stat_read_accum to read across field entries.
Add a part_stat_read_accum macro to genhd.h to read and sum across field entries. For example to sum up the number read and write sectors completed. In addition to being ar reasonable cleanup by itself this will make it easier to add new stat fields in the future.
Signed-off-by Michael Callahan <[email protected]>
--- drivers/block/drbd/drbd_receiver.c | 3 +-- drivers/block/drbd/drbd_worker.c | 4 +--- drivers/md/md.c | 3 +-- include/linux/genhd.h | 4 ++++ 4 files changed, 7 insertions(+), 7 deletions(-)
drivers/block/drbd/drbd_receiver.c +1 -2
@@ -2504,8 +2504,7 @@ bool drbd_rs_c_min_rate_throttle(struct drbd_device *device)if (c_min_rate ==0 )return false; - curr_events = (int )part_stat_read (&disk->part0, sectors[0 ]) + - (int )part_stat_read (&disk->part0, sectors[1 ]) - + curr_events = (int )part_stat_read_accum (&disk->part0, sectors) -atomic_read (&device->rs_sect_ev); if (atomic_read(&device->ap_actlog_cnt)
drivers/block/drbd/drbd_worker.c +1 -3
@@ -1591,9 +1591,7 @@ void drbd_rs_controller_reset(struct drbd_device *device)atomic_set (&device->rs_sect_in,0 );atomic_set (&device->rs_sect_ev,0 ); device->rs_in_flight =0 ; - device->rs_last_events = - (int )part_stat_read (&disk->part0, sectors[0 ]) + - (int )part_stat_read (&disk->part0, sectors[1 ]); + device->rs_last_events = (int )part_stat_read_accum (&disk->part0, sectors);/* Updating the RCU protected object in place is necessary since this function gets called from atomic context.
drivers/md/md.c +1 -2
@@ -7635,8 +7635,7 @@ static int is_mddev_idle(struct mddev *mddev, int init)rcu_read_lock ();rdev_for_each_rcu (rdev, mddev) {struct gendisk *disk = rdev->bdev->bd_contains->bd_disk; - curr_events = (int )part_stat_read (&disk->part0, sectors[0 ]) + - (int )part_stat_read (&disk->part0, sectors[1 ]) - + curr_events = (int )part_stat_read_accum (&disk->part0, sectors) -atomic_read (&disk->sync_io);/* sync IO will cause sync_io to increase before the disk_stats * as sync_io is counted when a request starts, and
include/linux/genhd.h +4 -0
@@ -364,6 +364,10 @@ static inline void free_part_stats(struct hd_struct *part)#endif /* CONFIG_SMP */ +#define part_stat_read_accum (part, field) \ + (part_stat_read(part, field[0]) + \ + part_stat_read(part, field[1])) +#define part_stat_add (cpu, part, field, addnd) do { \ __part_stat_add((cpu), (part), field, addnd); \ if ((part)->partno) \
Signature · 1 line
1.9.3
PATCH 2/4 block: Define and use STAT_READ and STAT_WRITE
In reply to Michael CallahanAdd defines for STAT_READ and STAT_WRITE for indexing the partition stat entries. This clarifies some fs/ code which has hardcoded 1 for STAT_WRITE and will make it easier to extend the stats with additional fields.
Signed-off-by Michael Callahan <[email protected]>
--- block/genhd.c | 16 ++++++++-------- block/partition-generic.c | 16 ++++++++-------- fs/ext4/super.c | 5 +++-- fs/ext4/sysfs.c | 6 ++++-- fs/f2fs/f2fs.h | 2 +- fs/f2fs/super.c | 3 ++- include/linux/genhd.h | 9 ++++++--- 7 files changed, 32 insertions(+), 25 deletions(-)
block/genhd.c +8 -8
@@ -1199,14 +1199,14 @@ static int diskstats_show(struct seq_file *seqf, void *v)"%u %lu %lu %lu %u %u %u %u\n" ,MAJOR (part_devt (hd)),MINOR (part_devt (hd)),disk_name (gp, hd->partno, buf), -part_stat_read (hd, ios[READ ]), -part_stat_read (hd, merges[READ ]), -part_stat_read (hd, sectors[READ ]), -jiffies_to_msecs (part_stat_read (hd, ticks[READ ])), -part_stat_read (hd, ios[WRITE ]), -part_stat_read (hd, merges[WRITE ]), -part_stat_read (hd, sectors[WRITE ]), -jiffies_to_msecs (part_stat_read (hd, ticks[WRITE ])), +part_stat_read (hd, ios[STAT_READ ]), +part_stat_read (hd, merges[STAT_READ ]), +part_stat_read (hd, sectors[STAT_READ ]), +jiffies_to_msecs (part_stat_read (hd, ticks[STAT_READ ])), +part_stat_read (hd, ios[STAT_WRITE ]), +part_stat_read (hd, merges[STAT_WRITE ]), +part_stat_read (hd, sectors[STAT_WRITE ]), +jiffies_to_msecs (part_stat_read (hd, ticks[STAT_WRITE ])),part_in_flight (hd),jiffies_to_msecs (part_stat_read (hd, io_ticks)),jiffies_to_msecs (part_stat_read (hd, time_in_queue))
block/partition-generic.c +8 -8
@@ -123,14 +123,14 @@ ssize_t part_stat_show(struct device *dev,"%8lu %8lu %8llu %8u " "%8u %8u %8u" "\n" , -part_stat_read (p, ios[READ ]), -part_stat_read (p, merges[READ ]), - (unsigned long long )part_stat_read (p, sectors[READ ]), -jiffies_to_msecs (part_stat_read (p, ticks[READ ])), -part_stat_read (p, ios[WRITE ]), -part_stat_read (p, merges[WRITE ]), - (unsigned long long )part_stat_read (p, sectors[WRITE ]), -jiffies_to_msecs (part_stat_read (p, ticks[WRITE ])), +part_stat_read (p, ios[STAT_READ ]), +part_stat_read (p, merges[STAT_READ ]), + (unsigned long long )part_stat_read (p, sectors[STAT_READ ]), +jiffies_to_msecs (part_stat_read (p, ticks[STAT_READ ])), +part_stat_read (p, ios[STAT_WRITE ]), +part_stat_read (p, merges[STAT_WRITE ]), + (unsigned long long )part_stat_read (p, sectors[STAT_WRITE ]), +jiffies_to_msecs (part_stat_read (p, ticks[STAT_WRITE ])),part_in_flight (p),jiffies_to_msecs (part_stat_read (p, io_ticks)), jiffies_to_msecs(part_stat_read(p, time_in_queue)));
fs/ext4/super.c +3 -2
@@ -3183,7 +3183,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent) sbi->s_sb_block = sb_block;if (sb->s_bdev->bd_part) sbi->s_sectors_written_start = -part_stat_read (sb->s_bdev->bd_part, sectors[1 ]); +part_stat_read (sb->s_bdev->bd_part, sectors[STAT_WRITE ]);/* Cleanup superblock name */ strreplace (sb->s_id,'/' ,'!' );
@@ -4356,7 +4356,8 @@ static int ext4_commit_super(struct super_block *sb, int sync)if (sb->s_bdev->bd_part) es->s_kbytes_written =cpu_to_le64 (EXT4_SB (sb)->s_kbytes_written + - ((part_stat_read (sb->s_bdev->bd_part, sectors[1 ]) - + ((part_stat_read (sb->s_bdev->bd_part, + sectors[STAT_WRITE ]) -EXT4_SB (sb)->s_sectors_written_start) >>1 ));else es->s_kbytes_written =
fs/ext4/sysfs.c +4 -2
@@ -55,7 +55,8 @@ static ssize_t session_write_kbytes_show(struct ext4_attr *a,if (!sb->s_bdev->bd_part)return snprintf (buf,PAGE_SIZE ,"0\n" );return snprintf (buf,PAGE_SIZE ,"%lu\n" , - (part_stat_read (sb->s_bdev->bd_part, sectors[1 ]) - + (part_stat_read (sb->s_bdev->bd_part, + sectors[STAT_WRITE ]) - sbi->s_sectors_written_start) >>1 ); }
@@ -68,7 +69,8 @@ static ssize_t lifetime_write_kbytes_show(struct ext4_attr *a,return snprintf (buf,PAGE_SIZE ,"0\n" );return snprintf (buf,PAGE_SIZE ,"%llu\n" , (unsigned long long )(sbi->s_kbytes_written + - ((part_stat_read (sb->s_bdev->bd_part, sectors[1 ]) - + ((part_stat_read (sb->s_bdev->bd_part, + sectors[STAT_WRITE ]) -EXT4_SB (sb)->s_sectors_written_start) >>1 ))); }
fs/f2fs/f2fs.h +1 -1
@@ -873,7 +873,7 @@ struct f2fs_sb_info {* and the return value is in kbytes. s is of struct f2fs_sb_info. */ #define BD_PART_WRITTEN (s) \ -(((u64)part_stat_read(s->sb->s_bdev->bd_part, sectors[1]) - \ +(((u64)part_stat_read(s->sb->s_bdev->bd_part, sectors[STAT_WRITE]) - \ s->sectors_written_start) >> 1)static inline void f2fs_update_time (struct f2fs_sb_info *sbi,int type)
fs/f2fs/super.c +2 -1
@@ -1636,7 +1636,8 @@ try_onemore:/* For write statistics */ if (sb->s_bdev->bd_part) sbi->sectors_written_start = - (u64 )part_stat_read (sb->s_bdev->bd_part, sectors[1 ]); + (u64 )part_stat_read (sb->s_bdev->bd_part, + sectors[STAT_WRITE ]);/* Read accumulated write IO statistics if exists */ seg_i =CURSEG_I (sbi,CURSEG_HOT_NODE );
include/linux/genhd.h +6 -3
@@ -60,6 +60,9 @@ enum {#define DISK_MAX_PARTS 256#define DISK_NAME_LEN 32 +#define STAT_READ 0 +#define STAT_WRITE 1 +#include <linux/major.h> #include <linux/device.h> #include <linux/smp.h>
@@ -81,7 +84,7 @@ struct partition { } __attribute__((packed));struct disk_stats { -unsigned long sectors[2 ];/* READs and WRITEs */ +unsigned long sectors[2 ];/* STAT_READs and STAT_WRITEs */ unsigned long ios[2 ];unsigned long merges[2 ];unsigned long ticks[2 ];
@@ -365,8 +368,8 @@ static inline void free_part_stats(struct hd_struct *part)#endif /* CONFIG_SMP */ #define part_stat_read_accum (part, field) \ - (part_stat_read(part, field[0]) + \ - part_stat_read(part, field[1])) + (part_stat_read(part, field[STAT_READ]) + \ + part_stat_read(part, field[STAT_WRITE]))#define part_stat_add (cpu, part, field, addnd) do { \ __part_stat_add((cpu), (part), field, addnd); \
Signature · 1 line
1.9.3
PATCH 3/4 block: Add and use a rw_stat_group function for indexing disk_stat fields.
In reply to Michael CallahanAdd and use a new rw_stat_group function for indexing partition stat fields rather than indexing them by rq_data_dir or bio_data_dir. This function works similarly to rw_is_sync in that it takes the request::cmd_flags or bio::bi_rw flags and determines which stats should et updated.
In addition the first parameter to generic_start_io_acct and generic_end_io_acct is now a stat group rather than simply a read or write bit. STAT_READ and STAT_WRITE have been defined such that this is backwards compatible but callers who want to track new stats should be updated to pass the extended parameter. This has been done by this patch for all current consumers of these two functions.
Note that the partition in_flight counts are not part of the per-cpu statistics and as such are not indexed via this function. An additional stat_group_to_rw helper macro has been added to more cleanly handle that case.
Signed-off-by Michael Callahan <[email protected]>
--- block/bio.c | 14 +++++++++----- block/blk-core.c | 9 +++++---- drivers/block/drbd/drbd_req.c | 4 ++-- drivers/block/rsxx/dev.c | 5 ++--- drivers/block/zram/zram_drv.c | 14 +++++++------- drivers/md/bcache/request.c | 10 +++++----- drivers/md/dm.c | 7 +++---- drivers/md/md.c | 5 +++-- drivers/nvdimm/core.c | 8 +++++--- include/linux/bio.h | 4 ++-- include/linux/genhd.h | 5 +++++ 11 files changed, 48 insertions(+), 37 deletions(-)
block/bio.c +9 -5
@@ -1667,27 +1667,31 @@ void bio_check_pages_dirty(struct bio *bio) } } -void generic_start_io_acct (int rw,unsigned long sectors, +void generic_start_io_acct (int rw_flags,unsigned long sectors,struct hd_struct *part) { +const int sgrp =rw_stat_group (rw_flags); +const int rw = (rw_flags &REQ_WRITE ) !=0 ;int cpu =part_stat_lock ();part_round_stats (cpu, part); -part_stat_inc (cpu, part, ios[rw]); -part_stat_add (cpu, part, sectors[rw], sectors); +part_stat_inc (cpu, part, ios[sgrp]); +part_stat_add (cpu, part, sectors[sgrp], sectors);part_inc_in_flight (part, rw);part_stat_unlock (); }EXPORT_SYMBOL (generic_start_io_acct); -void generic_end_io_acct (int rw,struct hd_struct *part, +void generic_end_io_acct (int rw_flags,struct hd_struct *part,unsigned long start_time) {unsigned long duration = jiffies - start_time; +const int sgrp =rw_stat_group (rw_flags); +const int rw = (rw_flags &REQ_WRITE ) !=0 ;int cpu =part_stat_lock (); -part_stat_add (cpu, part, ticks[rw], duration); +part_stat_add (cpu, part, ticks[sgrp], duration);part_round_stats (cpu, part);part_dec_in_flight (part, rw);
block/blk-core.c +5 -4
@@ -2275,13 +2275,13 @@ EXPORT_SYMBOL_GPL(blk_rq_err_bytes);void blk_account_io_completion (struct request *req,unsigned int bytes) {if (blk_do_io_stat (req)) { -const int rw =rq_data_dir (req); +const int sgrp =rw_stat_group (req->cmd_flags);struct hd_struct *part;int cpu; cpu =part_stat_lock (); part = req->part; -part_stat_add (cpu, part, sectors[rw], bytes >>9 ); +part_stat_add (cpu, part, sectors[sgrp], bytes >>9 );part_stat_unlock (); } }
@@ -2295,6 +2295,7 @@ void blk_account_io_done(struct request *req)*/ if (blk_do_io_stat (req) && !(req->cmd_flags &REQ_FLUSH_SEQ )) {unsigned long duration = jiffies - req->start_time; +const int sgrp =rw_stat_group (req->cmd_flags);const int rw =rq_data_dir (req);struct hd_struct *part;int cpu;
@@ -2302,8 +2303,8 @@ void blk_account_io_done(struct request *req) cpu =part_stat_lock (); part = req->part; -part_stat_inc (cpu, part, ios[rw]); -part_stat_add (cpu, part, ticks[rw], duration); +part_stat_inc (cpu, part, ios[sgrp]); +part_stat_add (cpu, part, ticks[sgrp], duration);part_round_stats (cpu, part);part_dec_in_flight (part, rw);
drivers/block/drbd/drbd_req.c +2 -2
@@ -36,14 +36,14 @@ static bool drbd_may_do_local_read(struct drbd_device *device, sector_t sector,/* Update disk stats at start of I/O request */ static void _drbd_start_io_acct (struct drbd_device *device,struct drbd_request *req) { -generic_start_io_acct (bio_data_dir (req->master_bio), req->i.size >>9 , +generic_start_io_acct (req->master_bio->bi_rw, req->i.size >>9 , &device->vdisk->part0); }/* Update disk stats when completing request upwards */ static void _drbd_end_io_acct (struct drbd_device *device,struct drbd_request *req) { -generic_end_io_acct (bio_data_dir (req->master_bio), +generic_end_io_acct (req->master_bio->bi_rw, &device->vdisk->part0, req->start_jif); }
drivers/block/rsxx/dev.c +2 -3
@@ -112,7 +112,7 @@ static const struct block_device_operations rsxx_fops = {static void disk_stats_start (struct rsxx_cardinfo *card,struct bio *bio) { -generic_start_io_acct (bio_data_dir (bio),bio_sectors (bio), +generic_start_io_acct (bio->bi_rw,bio_sectors (bio), &card->gendisk->part0); }
@@ -120,8 +120,7 @@ static void disk_stats_complete(struct rsxx_cardinfo *card,struct bio *bio, unsigned long start_time) { - generic_end_io_acct(bio_data_dir(bio), &card->gendisk->part0, -start_time ); + generic_end_io_acct(bio->bi_rw, &card->gendisk->part0, start_time); } static void bio_dma_done_cb(struct rsxx_cardinfo *card,
drivers/block/zram/zram_drv.c +7 -7
@@ -835,12 +835,12 @@ static void zram_bio_discard(struct zram *zram, u32 index, }static int zram_bvec_rw (struct zram *zram,struct bio_vec *bvec,u32 index, -int offset,int rw) +int offset,int rw,int rw_flags) {unsigned long start_time = jiffies;int ret; -generic_start_io_acct (rw, bvec->bv_len >>SECTOR_SHIFT , +generic_start_io_acct (rw_flags, bvec->bv_len >>SECTOR_SHIFT , &zram->disk->part0);if (rw ==READ ) {
@@ -851,7 +851,7 @@ static int zram_bvec_rw(struct zram *zram, struct bio_vec *bvec, u32 index, ret =zram_bvec_write (zram, bvec, index, offset); } -generic_end_io_acct (rw, &zram->disk->part0, start_time); +generic_end_io_acct (rw_flags, &zram->disk->part0, start_time);if (unlikely (ret)) {if (rw ==READ )
@@ -895,15 +895,15 @@ static void __zram_make_request(struct zram *zram, struct bio *bio) bv.bv_len = max_transfer_size; bv.bv_offset = bvec.bv_offset; -if (zram_bvec_rw (zram, &bv, index, offset, rw) <0 ) +if (zram_bvec_rw (zram, &bv, index, offset, rw, bio->bi_rw) <0 ) goto out; bv.bv_len = bvec.bv_len - max_transfer_size; bv.bv_offset += max_transfer_size; -if (zram_bvec_rw (zram, &bv, index +1 ,0 , rw) <0 ) +if (zram_bvec_rw (zram, &bv, index +1 ,0 , rw, bio->bi_rw) <0 ) goto out; }else -if (zram_bvec_rw (zram, &bvec, index, offset, rw) <0 ) +if (zram_bvec_rw (zram, &bvec, index, offset, rw, bio->bi_rw) <0 ) goto out;update_position (&index, &offset, &bvec);
@@ -984,7 +984,7 @@ static int zram_rw_page(struct block_device *bdev, sector_t sector, bv.bv_len =PAGE_SIZE ; bv.bv_offset =0 ; - err =zram_bvec_rw (zram, &bv, index, offset, rw); + err =zram_bvec_rw (zram, &bv, index, offset, rw, rw ?REQ_WRITE :0 ); put_zram:zram_meta_put (zram); out:
drivers/md/bcache/request.c +5 -5
@@ -609,7 +609,7 @@ static void request_endio(struct bio *bio)static void bio_complete (struct search *s) {if (s->orig_bio) { -generic_end_io_acct (bio_data_dir (s->orig_bio), +generic_end_io_acct (s->orig_bio->bi_rw, &s->d->disk->part0, s->start_time);trace_bcache_request_end (s->d, s->orig_bio);
@@ -964,9 +964,9 @@ static blk_qc_t cached_dev_make_request(struct request_queue *q,struct search *s;struct bcache_device *d = bio->bi_bdev->bd_disk->private_data; struct cached_dev *dc = container_of(d, struct cached_dev, disk); -int rw =bio_data_dir (bio); +const int rw =bio_data_dir (bio); -generic_start_io_acct (rw,bio_sectors (bio), &d->disk->part0); +generic_start_io_acct (bio->bi_rw,bio_sectors (bio), &d->disk->part0); bio->bi_bdev = dc->bdev; bio->bi_iter.bi_sector += dc->sb.data_offset;
@@ -1079,9 +1079,9 @@ static blk_qc_t flash_dev_make_request(struct request_queue *q,struct search *s;struct closure *cl;struct bcache_device *d = bio->bi_bdev->bd_disk->private_data; -int rw =bio_data_dir (bio); +const int rw =bio_data_dir (bio); -generic_start_io_acct (rw,bio_sectors (bio), &d->disk->part0); +generic_start_io_acct (bio->bi_rw,bio_sectors (bio), &d->disk->part0); s =search_alloc (bio, d); cl = &s->cl;
drivers/md/dm.c +3 -4
@@ -733,9 +733,9 @@ static void end_io_acct(struct dm_io *io)struct bio *bio = io->bio;unsigned long duration = jiffies - io->start_time;int pending; -int rw =bio_data_dir (bio); +const int rw =bio_data_dir (bio); -generic_end_io_acct (rw, &dm_disk (md)->part0, io->start_time); +generic_end_io_acct (bio->bi_rw, &dm_disk (md)->part0, io->start_time);if (unlikely (dm_stats_used (&md->stats))) dm_stats_account_io(&md->stats, bio->bi_rw, bio->bi_iter.bi_sector,
@@ -1818,14 +1818,13 @@ static void __split_and_process_bio(struct mapped_device *md,*/ static blk_qc_t dm_make_request (struct request_queue *q,struct bio *bio) { -int rw =bio_data_dir (bio);struct mapped_device *md = q->queuedata;int srcu_idx;struct dm_table *map; map =dm_get_live_table (md, &srcu_idx); -generic_start_io_acct (rw,bio_sectors (bio), &dm_disk (md)->part0); +generic_start_io_acct (bio->bi_rw,bio_sectors (bio), &dm_disk (md)->part0);/* if we're suspended, we have to queue this io for later */ if (unlikely (test_bit (DMF_BLOCK_IO_FOR_SUSPEND , &md->flags))) {
drivers/md/md.c +3 -2
@@ -245,6 +245,7 @@ static DEFINE_SPINLOCK(all_mddevs_lock);static blk_qc_t md_make_request (struct request_queue *q,struct bio *bio) {const int rw =bio_data_dir (bio); +const int sgrp =rw_stat_group (bio->bi_rw);struct mddev *mddev = q->queuedata;unsigned int sectors;int cpu;
@@ -289,8 +290,8 @@ static blk_qc_t md_make_request(struct request_queue *q, struct bio *bio) mddev->pers->make_request (mddev, bio); cpu =part_stat_lock (); -part_stat_inc (cpu, &mddev->gendisk->part0, ios[rw]); -part_stat_add (cpu, &mddev->gendisk->part0, sectors[rw], sectors); +part_stat_inc (cpu, &mddev->gendisk->part0, ios[sgrp]); +part_stat_add (cpu, &mddev->gendisk->part0, sectors[sgrp], sectors);part_stat_unlock ();if (atomic_dec_and_test (&mddev->active_io) && mddev->suspended)
drivers/nvdimm/core.c +5 -3
@@ -219,12 +219,13 @@ void __nd_iostat_start(struct bio *bio, unsigned long *start) {struct gendisk *disk = bio->bi_bdev->bd_disk;const int rw =bio_data_dir (bio); +const int sgrp =rw_stat_group (bio->bi_rw);int cpu =part_stat_lock (); *start = jiffies;part_round_stats (cpu, &disk->part0); -part_stat_inc (cpu, &disk->part0, ios[rw]); -part_stat_add (cpu, &disk->part0, sectors[rw],bio_sectors (bio)); +part_stat_inc (cpu, &disk->part0, ios[sgrp]); +part_stat_add (cpu, &disk->part0, sectors[sgrp],bio_sectors (bio));part_inc_in_flight (&disk->part0, rw);part_stat_unlock (); }
@@ -234,10 +235,11 @@ void nd_iostat_end(struct bio *bio, unsigned long start) {struct gendisk *disk = bio->bi_bdev->bd_disk;unsigned long duration = jiffies - start; +const int sgrp =rw_stat_group (bio->bi_rw);const int rw =bio_data_dir (bio);int cpu =part_stat_lock (); -part_stat_add (cpu, &disk->part0, ticks[rw], duration); +part_stat_add (cpu, &disk->part0, ticks[sgrp], duration);part_round_stats (cpu, &disk->part0);part_dec_in_flight (&disk->part0, rw);part_stat_unlock ();
include/linux/bio.h +2 -2
@@ -494,9 +494,9 @@ extern struct bio *bio_copy_kern(struct request_queue *, void *, unsigned int,extern void bio_set_pages_dirty (struct bio *bio);extern void bio_check_pages_dirty (struct bio *bio); -void generic_start_io_acct (int rw,unsigned long sectors, +void generic_start_io_acct (int rw_flags,unsigned long sectors,struct hd_struct *part); -void generic_end_io_acct (int rw,struct hd_struct *part, +void generic_end_io_acct (int rw_flags,struct hd_struct *part,unsigned long start_time);#ifndef ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
include/linux/genhd.h +5 -0
@@ -417,6 +417,11 @@ static inline void free_part_info(struct hd_struct *part)kfree (part->info); } +static inline int rw_stat_group (unsigned int rw_flags) +{ +return (rw_flags &REQ_WRITE ) !=0 ; +} +/* block/blk-core.c */ extern void part_round_stats (int cpu,struct hd_struct *part);
Signature · 1 line
1.9.3
PATCH 4/4 block: Track DISCARD statistics and output them in stat and diskstat.
In reply to Michael CallahanAdd tracking of REQ_DISCARD ios to the partition statistics and append them to the various stat files in /sys as well as /proc/diskstats. These are tracked with the same four stats as reads and writes:
Number of discard ios completed.
Number of discard ios merged
Number of discard sectors completed
Milliseconds spent on discard requests
This is done via adding a new STAT_DISCARD define to genhd.h and then using it to index that stat field for discard requests.
Signed-off-by Michael Callahan <[email protected]>
--- Documentation/ABI/testing/procfs-diskstats | 10 ++++++++++ Documentation/block/stat.txt | 28 ++++++++++++++++------------ Documentation/iostats.txt | 15 ++++++++++++++- block/genhd.c | 14 +++++++++++--- block/partition-generic.c | 9 +++++++-- include/linux/genhd.h | 14 +++++++++----- 6 files changed, 67 insertions(+), 23 deletions(-)
Documentation/ABI/testing/procfs-diskstats +10 -0
@@ -5,6 +5,7 @@ Description: The /proc/diskstats file displays the I/O statistics of block devices. Each line contains the following 14 fields: + 1 - major number 2 - minor mumber 3 - device name
@@ -19,4 +20,13 @@ Description: 12 - I/Os currently in progress 13 - time spent doing I/Os (ms) 14 - weighted time spent doing I/Os (ms) + + Kernel 4.8 appends four more fields for discard tracking + putting the total at 18: + + 15 - discards completed successfully + 16 - discards merged + 17 - sectors discarded + 18 - time spent discarding + For more details refer to Documentation/iostats.txt
Documentation/block/stat.txt +16 -12
@@ -31,28 +31,32 @@ write ticks milliseconds total wait time for write requests in_flight requests number of I/Os currently in flight io_ticks milliseconds total time this block device has been active time_in_queue milliseconds total wait time for all requests +discard I/Os requests number of discard I/Os processed +discard merges requests number of discard I/Os merged with in-queue I/O +discard sectors sectors number of sectors discarded +discard ticks milliseconds total wait time for discard requests -read I/Os, write I/Os -===================== +read I/Os, write I/Os, discard I/0s +=================================== These values increment when an I/O request completes. -read merges, write merges -========================= +read merges, write merges, discard merges +========================================= These values increment when an I/O request is merged with an already-queued I/O request. -read sectors, write sectors -=========================== +read sectors, write sectors, discard_sectors +============================================ -These values count the number of sectors read from or written to this -block device. The "sectors" in question are the standard UNIX 512-byte -sectors, not any device- or filesystem-specific block size. The -counters are incremented when the I/O completes. +These values count the number of sectors read from, written to, or +discarded from this block device. The "sectors" in question are the +standard UNIX 512-byte sectors, not any device- or filesystem-specific +block size. The counters are incremented when the I/O completes. -read ticks, write ticks -======================= +read ticks, write ticks, discard ticks +====================================== These values count the number of milliseconds that I/O requests have waited on this block device. If there are multiple I/O requests waiting,
Documentation/iostats.txt +14 -1
@@ -22,7 +22,6 @@ Here are examples of these different formats: 3 0 39082680 hda 446216 784926 9550688 4382310 424847 312726 5922052 19310380 0 3376340 23705160 3 1 9221278 hda1 35486 0 35496 38030 0 0 0 0 0 38030 38030 - 2.6 sysfs: 446216 784926 9550688 4382310 424847 312726 5922052 19310380 0 3376340 23705160 35486 38030 38030 38030
@@ -31,6 +30,9 @@ Here are examples of these different formats: 3 0 hda 446216 784926 9550688 4382310 424847 312726 5922052 19310380 0 3376340 23705160 3 1 hda1 35486 38030 38030 38030 +4.8 diskstats: + 3 0 hda 446216 784926 9550688 4382310 424847 312726 5922052 19310380 0 3376340 23705160 0 0 0 0 + On 2.4 you might execute "grep 'hda ' /proc/partitions". On 2.6, you have a choice of "cat /sys/block/hda/stat" or "grep 'hda ' /proc/diskstats". The advantage of one over the other is that the sysfs choice works well
@@ -90,6 +92,17 @@ Field 11 -- weighted # of milliseconds spent doing I/Os last update of this field. This can provide an easy measure of both I/O completion time and the backlog that may be accumulating. +In 4.8 the following four fields are appended for discard tracking: + +Field 12 -- # of discards completed + This is the total number of discards completed successfully. +Field 13 -- # of discards merged + See the description of field 2 +Field 14 -- # of sectors discarded + This is the total number of sectors discarded successfully. +Field 15 -- # of milliseconds spent discarding + This is the total number of milliseconds spent by all discards (as + measured from __make_request() to end_that_request_last()). To avoid introducing performance bottlenecks, no locks are held while modifying these counters. This implies that minor inaccuracies may be
block/genhd.c +11 -3
@@ -1195,8 +1195,12 @@ static int diskstats_show(struct seq_file *seqf, void *v) cpu =part_stat_lock ();part_round_stats (cpu, hd);part_stat_unlock (); - seq_printf(seqf, "%4d %7d %s %lu %lu %lu " - "%u %lu %lu %lu %u %u %u %u\n", + seq_printf(seqf, "%4d %7d %s " + "%lu %lu %lu %u " + "%lu %lu %lu %u " + "%u %u %u " + "%lu %lu %lu %u" + "\n", MAJOR(part_devt(hd)), MINOR(part_devt(hd)), disk_name(gp, hd->partno, buf), part_stat_read(hd, ios[STAT_READ]),
@@ -1209,7 +1213,11 @@ static int diskstats_show(struct seq_file *seqf, void *v)jiffies_to_msecs (part_stat_read (hd, ticks[STAT_WRITE ])),part_in_flight (hd),jiffies_to_msecs (part_stat_read (hd, io_ticks)), -jiffies_to_msecs (part_stat_read (hd, time_in_queue)) +jiffies_to_msecs (part_stat_read (hd, time_in_queue)), +part_stat_read (hd, ios[STAT_DISCARD ]), +part_stat_read (hd, merges[STAT_DISCARD ]), +part_stat_read (hd, sectors[STAT_DISCARD ]), +jiffies_to_msecs (part_stat_read (hd, ticks[STAT_DISCARD ])) ); }disk_part_iter_exit (&piter);
block/partition-generic.c +7 -2
@@ -121,7 +121,8 @@ ssize_t part_stat_show(struct device *dev, return sprintf(buf, "%8lu %8lu %8llu %8u " "%8lu %8lu %8llu %8u " - "%8u %8u %8u" + "%8u %8u %8u " + "%8lu %8lu %8llu %8u" "\n", part_stat_read(p, ios[STAT_READ]), part_stat_read(p, merges[STAT_READ]),
@@ -133,7 +134,11 @@ ssize_t part_stat_show(struct device *dev,jiffies_to_msecs (part_stat_read (p, ticks[STAT_WRITE ])),part_in_flight (p),jiffies_to_msecs (part_stat_read (p, io_ticks)), - jiffies_to_msecs(part_stat_read(p, time_in_queue))); +jiffies_to_msecs (part_stat_read (p, time_in_queue)), +part_stat_read (p, ios[STAT_DISCARD ]), +part_stat_read (p, merges[STAT_DISCARD ]), + (unsigned long long )part_stat_read (p, sectors[STAT_DISCARD ]), + jiffies_to_msecs(part_stat_read(p, ticks[STAT_DISCARD]))); } ssize_t part_inflight_show(struct device *dev,
include/linux/genhd.h +9 -5
@@ -62,6 +62,7 @@ enum {#define STAT_READ 0#define STAT_WRITE 1 +#define STAT_DISCARD 2#include <linux/major.h> #include <linux/device.h>
@@ -84,10 +85,10 @@ struct partition { } __attribute__((packed));struct disk_stats { -unsigned long sectors[2 ];/* STAT_READs and STAT_WRITEs */ -unsigned long ios[2 ]; -unsigned long merges[2 ]; -unsigned long ticks[2 ]; +unsigned long sectors[3 ];/* STAT_READ, STAT_WRITE, STAT_DISCARD */ +unsigned long ios[3 ]; +unsigned long merges[3 ]; +unsigned long ticks[3 ];unsigned long io_ticks;unsigned long time_in_queue; };
@@ -369,7 +370,8 @@ static inline void free_part_stats(struct hd_struct *part)#define part_stat_read_accum (part, field) \ (part_stat_read(part, field[STAT_READ]) + \ - part_stat_read(part, field[STAT_WRITE])) + part_stat_read(part, field[STAT_WRITE]) + \ + part_stat_read(part, field[STAT_DISCARD]))#define part_stat_add (cpu, part, field, addnd) do { \ __part_stat_add((cpu), (part), field, addnd); \
@@ -419,6 +421,8 @@ static inline void free_part_info(struct hd_struct *part)static inline int rw_stat_group (unsigned int rw_flags) { +if (rw_flags &REQ_DISCARD ) +return STAT_DISCARD ;return (rw_flags &REQ_WRITE ) !=0 ; }
Signature · 1 line
1.9.3