vfs: hot data tracking
Subsystems (4)
cover vfs: hot data tracking
From: Zhi Yong Wu <[email protected]>
NOTE:
The patchset is currently post out mainly to make sure
it is going in the correct direction and hope to get some helpful comments from other guys.
TODO List:
1.) Need to do scalability or performance tests.
2.) Turn some Micro into tunables
TIME_TO_KICK, and HEAT_UPDATE_DELAY
3.) Rafactor hot_hash_is_aging()
If you just made the timeout value a timespec and compared
the _timespecs_, you would be doing a lot fewer conversions.
4.) Cleanup some unnecessary lock protect
5.) Add more comments to explain how to calc temperature
Ben Chociej, Matt Lupfer and Conor Scott originally wrote this code to
be very btrfs-specific. I've taken their code and attempted to
make it more generic and integrate it at the VFS level.
INTRODUCTION:
Essentially, this means maintaining some key stats
(like number of reads/writes, last read/write time, frequency of reads/writes), then distilling those numbers down to a single "temperature" value that reflects what data is "hot," and using that temperature to move data to SSDs.
The long-term goal of these patches is to allow some FSs,
e.g. Btrfs to intelligently utilize SSDs in a heterogenous volume. Incidentally, this project has been motivated by
the Project Ideas page on the Btrfs wiki.
Of course, users are warned not to run this code outside of development
environments. These patches are EXPERIMENTAL, and as such they might eat your data and/or memory. That said, the code should be relatively safe when the hottrack mount option are disabled.
MOTIVATION:
The overall goal of enabling hot data relocation to SSD has been
motivated by the Project Ideas page on the Btrfs wiki at
<https://btrfs.wiki.kernel.org/index.php/Project_ideas>.
It will divide into two steps. VFS provide hot data tracking function while specific FS will provide hot data relocation function. So as the first step of this goal, it is hoped that the patchset for hot data tracking will eventually mature into VFS.
This is essentially the traditional cache argument: SSD is fast and
expensive; HDD is cheap but slow. ZFS, for example, can already take advantage of SSD caching. Btrfs should also be able to take advantage of hybrid storage without many broad, sweeping changes to existing code.
SUMMARY:
- Hooks in existing vfs functions to track data access frequency
- New rbtrees for tracking access frequency of inodes and sub-file ranges (hot_rb.c)
The relationship between super_block and rbtree is as below:
super_block->s_hotinfo.hot_inode_tree
In include/linux/fs.h, one struct hot_info s_hotinfo is added to
super_block struct. Each FS instance can find hot tracking info s_hotinfo via its super_block. In this hot_info, it store a lot of hot tracking info such as hot_inode_tree, inode and range hash list, etc.
- A hash list for indexing data by its temperature (hot_hash.c)
- A debugfs interface for dumping data from the rbtrees (hot_debugfs.c)
- A background kthread for updating inode heat info
- Mount options for enabling temperature tracking(-o hottrack, default mean disabled) (hot_track.c)
- An ioctl to retrieve the frequency information collected for a certain file
- Ioctls to enable/disable frequency tracking per inode.
Usage syntax:
root@debian-i386:~# mount -o hottrack /dev/sdb /mnt [ 1505.894078] device label test devid 1 transid 29 /dev/sdb [ 1505.952977] btrfs: disk space caching is enabled [ 1506.069678] vfs: turning on hot data tracking root@debian-i386:~# mount -t debugfs none /sys/kernel/debug root@debian-i386:~# ls -l /sys/kernel/debug/vfs_hotdata/ total 0 drwxr-xr-x 2 root root 0 Aug 8 04:40 sdb root@debian-i386:~# ls -l /sys/kernel/debug/vfs_hotdata/sdb total 0 -rw-r--r-- 1 root root 0 Aug 8 04:40 inode_data -rw-r--r-- 1 root root 0 Aug 8 04:40 range_data root@debian-i386:~# vi /mnt/file root@debian-i386:~# cat /sys/kernel/debug/hot_track/sdb/inode_data inode #279, reads 0, writes 1, avg read time 18446744073709551615, avg write time 5251566408153596, temp 109 root@debian-i386:~# cat /sys/kernel/debug/hot_track/sdb/range_data inode #279, range start 0 (range len 1048576) reads 0, writes 1, avg read time 18446744073709551615, avg write time 1128690176623144209, temp 64 root@debian-i386:~# echo "hot data tracking test" >> /mnt/file root@debian-i386:~# cat /sys/kernel/debug/hot_track/sdb/inode_data inode #279, reads 0, writes 2, avg read time 18446744073709551615, avg write time 4923343766042451, temp 109 root@debian-i386:~# cat /sys/kernel/debug/hot_track/sdb/range_data inode #279, range start 0 (range len 1048576) reads 0, writes 2, avg read time 18446744073709551615, avg write time 1058147040842596150, temp 64 root@debian-i386:~#
Changelog from v2:
1.) Adjust the code based on Ram's comments. [Ram Pai]
v2:
1.) refactor the code based on Dave, Mingming and Ram's sugestions.
[Dave Hansen, Mingming Cao, and Ram Pai]
v1:
1.) refactor the code based on Dave's sugestions. [Dave Hansen] 2.) fix up some bugs.
Zhi Yong Wu (11):
vfs: introduce one structure hot_info
vfs: introduce one rb tree - hot_inode_tree
vfs: introduce 2 rb tree items - inode and range
vfs: add support for updating access frequency
vfs: add one new mount option '-o hottrack'
vfs: add init and exit support
vfs: introduce one hash table
vfs: enable hot data tracking
vfs: fork one private kthread to update temperature info
vfs: add 3 new ioctl interfaces
vfs: add debugfs support
fs/Makefile | 3 +- fs/compat_ioctl.c | 8 + fs/dcache.c | 2 + fs/direct-io.c | 10 + fs/hot_debugfs.c | 487 ++++++++++++++++++++++++++++++++++ fs/hot_debugfs.h | 60 +++++ fs/hot_hash.c | 369 ++++++++++++++++++++++++++ fs/hot_hash.h | 108 ++++++++ fs/hot_rb.c | 648 +++++++++++++++++++++++++++++++++++++++++++++ fs/hot_rb.h | 70 +++++ fs/hot_track.c | 113 ++++++++ fs/hot_track.h | 23 ++ fs/ioctl.c | 132 +++++++++ fs/namespace.c | 10 + fs/super.c | 11 + include/linux/fs.h | 15 + include/linux/hot_track.h | 169 ++++++++++++ mm/filemap.c | 8 + mm/page-writeback.c | 21 ++ mm/readahead.c | 9 + 20 files changed, 2275 insertions(+), 1 deletions(-) create mode 100644 fs/hot_debugfs.c create mode 100644 fs/hot_debugfs.h create mode 100644 fs/hot_hash.c create mode 100644 fs/hot_hash.h create mode 100644 fs/hot_rb.c create mode 100644 fs/hot_rb.h create mode 100644 fs/hot_track.c create mode 100644 fs/hot_track.h create mode 100644 include/linux/hot_track.h
Signature · 1 line
1.7.6.5
RFC 1/11 vfs: introduce one structure hot_info
From: Zhi Yong Wu <[email protected]>
One root structure hot_info is defined, is hooked
up in super_block, and will be used to hold rb trees root, hash list root and some other information, etc.
--- include/linux/fs.h | 4 ++++ include/linux/hot_track.h | 26 ++++++++++++++++++++++++++ 2 files changed, 30 insertions(+), 0 deletions(-) create mode 100644 include/linux/hot_track.h
include/linux/fs.h +4 -0
@@ -415,6 +415,7 @@ struct inodes_stat_t {#include <linux/migrate_mode.h> #include <linux/uidgid.h> #include <linux/lockdep.h> +#include <linux/hot_track.h> #include <asm/byteorder.h>
@@ -1578,6 +1579,9 @@ struct super_block {/* Being remounted read-only */ int s_readonly_remount; + +/* Hot data tracking info*/ +struct hot_info s_hotinfo; };/* superblock cache pruning functions */
include/linux/hot_track.h new file +26 -0
@@ -0,0 +1,26 @@ +/* +* include/linux/hot_track.h +* +* This file has definitions for VFS hot data tracking +* structures etc. +* +* Copyright (C) 2012 IBM Corp. All rights reserved. +* Written by Zhi Yong Wu <[email protected]> +* Ben Chociej <[email protected]> +* +* This program is free software; you can redistribute it and/or +* modify it under the terms of the GNU General Public +* License v2 as published by the Free Software Foundation. +*/ + +#ifndef _LINUX_HOTTRACK_H +#define _LINUX_HOTTRACK_H + +#include <linux/types.h> +#include <linux/rbtree.h> +#include <linux/kref.h> + +struct hot_info { +}; + +#endif /* _LINUX_HOTTRACK_H */
Signature · 1 line
1.7.6.5
RFC 2/11 vfs: introduce one rb tree - hot_inode_tree
In reply to [email protected]From: Zhi Yong Wu <[email protected]>
Adds hot_inode_tree struct to keep track of
frequently accessed files, and be keyed by {inode, offset}. Trees contain hot_inode_items representing those files and ranges.
Having these trees means that vfs can quickly determine the
temperature of some data by doing some calculations on the hot_freq_data struct that hangs off of the tree item.
--- fs/Makefile | 3 ++- fs/hot_rb.c | 30 ++++++++++++++++++++++++++++++ fs/hot_rb.h | 21 +++++++++++++++++++++ include/linux/hot_track.h | 9 +++++++++ 4 files changed, 62 insertions(+), 1 deletions(-) create mode 100644 fs/hot_rb.c create mode 100644 fs/hot_rb.h
fs/Makefile +2 -1
@@ -11,7 +11,8 @@ obj-y := open.o read_write.o file_table.o super.o \ attr.o bad_inode.o file.o filesystems.o namespace.o \ seq_file.o xattr.o libfs.o fs-writeback.o \ pnode.o drop_caches.o splice.o sync.o utimes.o \ - stack.o fs_struct.o statfs.o + stack.o fs_struct.o statfs.o \ + hot_rb.o ifeq ($(CONFIG_BLOCK),y) obj-y += buffer.o bio.o block_dev.o direct-io.o mpage.o ioprio.o
fs/hot_rb.c new file +30 -0
@@ -0,0 +1,30 @@ +/* +* fs/hot_rb.c +* +* Copyright (C) 2012 IBM Corp. All rights reserved. +* Written by Zhi Yong Wu <[email protected]> +* Ben Chociej <[email protected]> +* +* This program is free software; you can redistribute it and/or +* modify it under the terms of the GNU General Public +* License v2 as published by the Free Software Foundation. +*/ + +#include <linux/err.h> +#include <linux/slab.h> +#include <linux/module.h> +#include <linux/spinlock.h> +#include <linux/hardirq.h> +#include <linux/blkdev.h> +#include "hot_rb.h" +#include "hot_hash.h" + +/* +* Initialize the inode tree. Should be called for each new inode +* access or other user of the hot_inode interface. +*/ +void hot_rb_inode_tree_init (struct hot_inode_tree *tree) +{ + tree->map =RB_ROOT ; +rwlock_init (&tree->lock); +}
fs/hot_rb.h new file +21 -0
@@ -0,0 +1,21 @@ +/* +* fs/hot_rb.h +* +* Copyright (C) 2012 IBM Corp. All rights reserved. +* Written by Zhi Yong Wu <[email protected]> +* Ben Chociej <[email protected]> +* +* This program is free software; you can redistribute it and/or +* modify it under the terms of the GNU General Public +* License v2 as published by the Free Software Foundation. +*/ + +#ifndef __HOT_MAP__ +#define __HOT_MAP__ + +#include <linux/rbtree.h> +#include <linux/hot_track.h> + +void hot_rb_inode_tree_init (struct hot_inode_tree *tree); + +#endif /* __HOT_MAP__ */
include/linux/hot_track.h +9 -0
@@ -20,7 +20,16 @@#include <linux/rbtree.h> #include <linux/kref.h> +/* A tree that sits on the hot_info */ +struct hot_inode_tree { +struct rb_root map; +rwlock_t lock; +}; +struct hot_info { + +/* red-black tree that keeps track of fs-wide hot data */ +struct hot_inode_tree hot_inode_tree; };#endif /* _LINUX_HOTTRACK_H */
Signature · 1 line
1.7.6.5
RFC 3/11 vfs: introduce 2 rb tree items - inode and range
In reply to [email protected]From: Zhi Yong Wu <[email protected]>
Define two items hot_inode_item and hot_range_item,
one of them represents one tracked file
to keep track of its access frequency and the tree of ranges in this file, while the latter represents a file range of one inode.
Each of the two structures contains a hot_freq_data
struct with its frequency of access metrics (number of {reads, writes}, last {read,write} time, frequency of {reads,writes}).
Also, each hot_inode_item contains one hot_range_tree
struct which is keyed by {inode, offset, length} and used to keep track of all the ranges in this file.
--- fs/Makefile | 2 +- fs/dcache.c | 2 + fs/hot_rb.c | 74 +++++++++++++++++++++++++++++++++++++++++++++ fs/hot_rb.h | 9 +++++ fs/hot_track.c | 26 ++++++++++++++++ fs/hot_track.h | 20 ++++++++++++ include/linux/hot_track.h | 62 +++++++++++++++++++++++++++++++++++++ 7 files changed, 194 insertions(+), 1 deletions(-) create mode 100644 fs/hot_track.c create mode 100644 fs/hot_track.h
fs/Makefile +1 -1
@@ -12,7 +12,7 @@ obj-y := open.o read_write.o file_table.o super.o \ seq_file.o xattr.o libfs.o fs-writeback.o \ pnode.o drop_caches.o splice.o sync.o utimes.o \ stack.o fs_struct.o statfs.o \ - hot_rb.o + hot_rb.o hot_track.o ifeq ($(CONFIG_BLOCK),y) obj-y += buffer.o bio.o block_dev.o direct-io.o mpage.o ioprio.o
fs/dcache.c +2 -0
@@ -39,6 +39,7 @@#include <linux/ratelimit.h> #include "internal.h" #include "mount.h" +#include "hot_track.h" /* * Usage:
@@ -3164,6 +3165,7 @@ void __init vfs_caches_init(unsigned long mempages)inode_init ();files_init (mempages);mnt_init (); +hot_track_kmem_cache_init ();bdev_cache_init ();chrdev_init (); }
fs/hot_rb.c +74 -0
83 lines not on this page · raw message
fs/hot_rb.h +9 -0
@@ -16,6 +16,15 @@#include <linux/rbtree.h> #include <linux/hot_track.h> +/* values for hot_freq_data flags */ +#define FREQ_DATA_TYPE_INODE (1 << 0)/* freq data struct is for an inode */ +#define FREQ_DATA_TYPE_RANGE (1 << 1)/* freq data struct is for a range */ +void hot_rb_inode_tree_init (struct hot_inode_tree *tree); +void hot_rb_inode_item_init (void *_item); +void hot_rb_range_item_init (void *_item); + +int __init hot_rb_item_cache_init (void ); +#endif /* __HOT_MAP__ */
fs/hot_track.c new file +26 -0
@@ -0,0 +1,26 @@ +/* +* fs/hot_track.c +* +* Copyright (C) 2012 IBM Corp. All rights reserved. +* Written by Zhi Yong Wu <[email protected]> +* Ben Chociej <[email protected]> +* +* This program is free software; you can redistribute it and/or +* modify it under the terms of the GNU General Public +* License v2 as published by the Free Software Foundation. +*/ + +#include <linux/slab.h> +#include <linux/gfp.h> +#include <linux/parser.h> +#include "hot_track.h" + +/* +* Initialize kmem cache for hot_inode_item +* and hot_range_item +*/ +void __init hot_track_kmem_cache_init (void ) +{ +if (hot_rb_item_cache_init ()) +return ; +}
fs/hot_track.h new file +20 -0
@@ -0,0 +1,20 @@ +/* +* fs/hot_track.h +* +* Copyright (C) 2012 IBM Corp. All rights reserved. +* Written by Zhi Yong Wu <[email protected]> +* Ben Chociej <[email protected]> +* +* This program is free software; you can redistribute it and/or +* modify it under the terms of the GNU General Public +* License v2 as published by the Free Software Foundation. +*/ + +#ifndef __HOT_TRACK__ +#define __HOT_TRACK__ + +#include "hot_rb.h" + +void __init hot_track_kmem_cache_init (void ); + +#endif /* __HOT_TRACK__ */
include/linux/hot_track.h +62 -0
68 lines not on this page · raw message
Signature · 1 line
1.7.6.5
RFC 4/11 vfs: add support for updating access frequency
In reply to [email protected]From: Zhi Yong Wu <[email protected]>
Add some utils helpers to update access frequencies
for one file or its range.
--- fs/hot_rb.c | 359 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ fs/hot_rb.h | 22 ++++ 2 files changed, 381 insertions(+), 0 deletions(-)
fs/hot_rb.c +359 -0
@@ -102,3 +102,362 @@ void hot_rb_range_item_init(void *_item) hr->hot_freq_data.avg_delta_writes = (u64) -1 ; hr->hot_freq_data.flags =FREQ_DATA_TYPE_RANGE ; } + +/* +* Drops the reference out on hot_inode_item by one and free the structure +* if the reference count hits zero +*/ +void hot_rb_free_hot_inode_item (struct hot_inode_item *he) +{ +if (!he) +return ; + +if (atomic_dec_and_test (&he->refs.refcount)) { +WARN_ON (he->in_tree); +kmem_cache_free (hot_inode_item_cache, he); + } +} + +/* +* Drops the reference out on hot_range_item by one and free the structure +* if the reference count hits zero +*/ +static void hot_rb_free_hot_range_item (struct hot_range_item *hr) +{ +if (!hr) +return ; + +if (atomic_dec_and_test (&hr->refs.refcount)) { +WARN_ON (hr->in_tree); +kmem_cache_free (hot_range_item_cache, hr); + } +} + +static struct rb_node *hot_rb_insert_hot_inode_item (struct rb_root *root, +unsigned long inode_num, +struct rb_node *node) +{ +struct rb_node **p = &root->rb_node; +struct rb_node *parent =NULL ; +struct hot_inode_item *entry; + +/* walk tree to find insertion point */ +while (*p) { + parent = *p; + entry = rb_entry(parent, struct hot_inode_item, rb_node); + +if (inode_num < entry->i_ino) + p = &(*p)->rb_left; +else if (inode_num > entry->i_ino) + p = &(*p)->rb_right; +else +return parent; + } + + entry = rb_entry(node, struct hot_inode_item, rb_node); + entry->in_tree =1 ; +rb_link_node (node, parent, p); +rb_insert_color (node, root); + +return NULL ; +} + +static u64 hot_rb_range_end (struct hot_range_item *hr) +{ +if (hr->start + hr->len < hr->start) +return (u64)-1 ; + +return hr->start + hr->len -1 ; +} + +static struct rb_node *hot_rb_insert_hot_range_item (struct rb_root *root, +u64 start, +struct rb_node *node) +{ +struct rb_node **p = &root->rb_node; +struct rb_node *parent =NULL ; +struct hot_range_item *entry; + +/* ensure start is on a range boundary */ + start = start &RANGE_SIZE_MASK ; +/* walk tree to find insertion point */ +while (*p) { + parent = *p; + entry = rb_entry(parent, struct hot_range_item, rb_node); + +if (start < entry->start) + p = &(*p)->rb_left; +else if (start >=hot_rb_range_end (entry)) + p = &(*p)->rb_right; +else +return parent; + } + + entry = rb_entry(node, struct hot_range_item, rb_node); + entry->in_tree =1 ; +rb_link_node (node, parent, p); +rb_insert_color (node, root); + +return NULL ; +} + +/* +* Add a hot_inode_item to a hot_inode_tree. If the tree already contains +* an item with the index given, return -EEXIST +*/ +static int hot_rb_add_hot_inode_item (struct hot_inode_tree *tree, +struct hot_inode_item *he) +{ +int ret =0 ; +struct rb_node *rb; + + rb =hot_rb_insert_hot_inode_item ( + &tree->map, he->i_ino, &he->rb_node); +if (rb) { + ret = -EEXIST ; + goto out; + } + +kref_get (&he->refs); + +out: +return ret; +} + +/* +* Add a hot_range_item to a hot_range_tree. If the tree already contains +* an item with the index given, return -EEXIST +* +* Also optionally aggresively merge ranges (currently disabled) +*/ +static int hot_rb_add_hot_range_item (struct hot_range_tree *tree, +struct hot_range_item *hr) +{ +int ret =0 ; +struct rb_node *rb; + + rb =hot_rb_insert_hot_range_item ( + &tree->map, hr->start, &hr->rb_node); +if (rb) { + ret = -EEXIST ; + goto out; + } + +kref_get (&hr->refs); + +out: +return ret; +} + +/* +* Lookup a hot_inode_item in the hot_inode_tree with the given index +* (inode_num) +*/ +struct hot_inode_item +*hot_rb_lookup_hot_inode_item (struct hot_inode_tree *tree, +unsigned long inode_num) +{ +struct rb_node **p = &(tree->map.rb_node); +struct rb_node *parent =NULL ; +struct hot_inode_item *entry; + +while (*p) { + parent = *p; + entry = rb_entry(parent, struct hot_inode_item, rb_node); + +if (inode_num < entry->i_ino) + p = &(*p)->rb_left; +else if (inode_num > entry->i_ino) + p = &(*p)->rb_right; +else { +kref_get (&entry->refs); +return entry; + } + } + +return NULL ; +} + +/* +* Lookup a hot_range_item in a hot_range_tree with the given index +* (start, offset) +*/ +struct hot_range_item +*hot_rb_lookup_hot_range_item (struct hot_range_tree *tree, +u64 start) +{ +struct rb_node **p = &(tree->map.rb_node); +struct rb_node *parent =NULL ; +struct hot_range_item *entry; + +/* ensure start is on a range boundary */ + start = start &RANGE_SIZE_MASK ; +while (*p) { + parent = *p; + entry = rb_entry(parent, struct hot_range_item, rb_node); + +if (start < entry->start) + p = &(*p)->rb_left; +else if (start >hot_rb_range_end (entry)) + p = &(*p)->rb_right; +else { +kref_get (&entry->refs); +return entry; + } + } + +return NULL ; +} + +/* Update inode frequency struct */ +static struct hot_inode_item *hot_rb_update_inode_freq (struct inode *inode, +int rw) +{ +struct hot_info *root = &(inode->i_sb->s_hotinfo); +struct hot_inode_tree *hitree = &(root->hot_inode_tree); +struct hot_inode_item *he; + +read_lock (&hitree->lock); + he =hot_rb_lookup_hot_inode_item (hitree, inode->i_ino); +read_unlock (&hitree->lock); + +if (!he) { + he =kmem_cache_alloc (hot_inode_item_cache, +GFP_KERNEL |GFP_NOFS ); +if (!he) + goto out; + +write_lock (&hitree->lock); +hot_rb_inode_item_init (he); + he->i_ino = inode->i_ino; +hot_rb_add_hot_inode_item (hitree, he); +write_unlock (&hitree->lock); + } + +spin_lock (&he->lock); +hot_rb_update_freq (&he->hot_freq_data, rw); +spin_unlock (&he->lock); + +out: +return he; +} + +/* Update range frequency struct */ +static bool hot_rb_update_range_freq (struct hot_inode_item *he, +u64 off,u64 len,int rw, +struct hot_info *root) +{ +struct hot_range_tree *hrtree = &(he->hot_range_tree); +struct hot_range_item *hr =NULL ; +u64 start_off = off &RANGE_SIZE_MASK ; +u64 end_off = (off + len -1 ) &RANGE_SIZE_MASK ; +u64 cur; +int ret = true; + +if (len ==0 ) +return false; + +/* +* Align ranges on RANGE_SIZE boundary to prevent proliferation +* of range structs +*/ +for (cur = start_off; cur <= end_off; cur +=RANGE_SIZE ) { +read_lock (&hrtree->lock); + hr =hot_rb_lookup_hot_range_item (hrtree, cur); +read_unlock (&hrtree->lock); + +if (!hr) { + hr =kmem_cache_alloc (hot_range_item_cache, +GFP_KERNEL |GFP_NOFS ); +if (!hr) { + ret = false; + goto out; + } + +write_lock (&hrtree->lock); +hot_rb_range_item_init (hr); + hr->start = cur &RANGE_SIZE_MASK ; + hr->len =RANGE_SIZE ; + hr->hot_inode = he; +hot_rb_add_hot_range_item (hrtree, hr); +write_unlock (&hrtree->lock); + } + +spin_lock (&hr->lock); +hot_rb_update_freq (&hr->hot_freq_data, rw); +spin_unlock (&hr->lock); +hot_rb_free_hot_range_item (hr); + } + +out: +return ret; +} + +/* +* This function does the actual work of updating the frequency numbers, +* whatever they turn out to be. FREQ_POWER determines how many atime +* deltas we keep track of (as a power of 2). So, setting it to anything above +* 16ish is probably overkill. Also, the higher the power, the more bits get +* right shifted out of the timestamp, reducing precision, so take note of that +* as well. +* +* The caller should have already locked freq_data's parent's spinlock. +* +* FREQ_POWER, defined immediately below, determines how heavily to weight +* the current frequency numbers against the newest access. For example, a value +* of 4 means that the new access information will be weighted 1/16th (ie 2^-4) +* as heavily as the existing frequency info. In essence, this is a kludged- +* together version of a weighted average, since we can't afford to keep all of +* the information that it would take to get a _real_ weighted average. +*/ +void hot_rb_update_freq (struct hot_freq_data *freq_data,int rw) +{ +struct timespec old_atime; +struct timespec current_time; +struct timespec delta_ts; +u64 new_avg; +u64 new_delta; + +if (unlikely (rw)) { + old_atime = freq_data->last_write_time; + freq_data->nr_writes +=1 ; + new_avg = freq_data->avg_delta_writes; + }else { + old_atime = freq_data->last_read_time; + freq_data->nr_reads +=1 ; + new_avg = freq_data->avg_delta_reads; + } + + current_time =current_kernel_time (); + delta_ts =timespec_sub (current_time, old_atime); + new_delta =timespec_to_ns (&delta_ts) >>FREQ_POWER ; + + new_avg = (new_avg <<FREQ_POWER ) - new_avg + new_delta; + new_avg = new_avg >>FREQ_POWER ; + +if (unlikely (rw)) { + freq_data->last_write_time = current_time; + freq_data->avg_delta_writes = new_avg; + }else { + freq_data->last_read_time = current_time; + freq_data->avg_delta_reads = new_avg; + } +} + +/* main function to update access frequency from read/writepage(s) hooks */ +void hot_rb_update_freqs (struct inode *inode,u64 start, +u64 len,int rw) +{ +struct hot_inode_item *he; + + he =hot_rb_update_inode_freq (inode, rw); + +WARN_ON (!he); + +if (he) { +hot_rb_update_range_freq (he, start, len, + rw, &(inode->i_sb->s_hotinfo)); + +hot_rb_free_hot_inode_item (he); + } +}
fs/hot_rb.h +22 -0
@@ -19,12 +19,34 @@/* values for hot_freq_data flags */ #define FREQ_DATA_TYPE_INODE (1 << 0)/* freq data struct is for an inode */ #define FREQ_DATA_TYPE_RANGE (1 << 1)/* freq data struct is for a range */ +/* size of sub-file ranges */ +#define RANGE_SIZE (1<<20) +#define RANGE_SIZE_MASK (~((u64)(RANGE_SIZE - 1))) + +#define FREQ_POWER 4 + +struct hot_info ; +struct inode ;void hot_rb_inode_tree_init (struct hot_inode_tree *tree);void hot_rb_inode_item_init (void *_item);void hot_rb_range_item_init (void *_item); +struct hot_range_item +*hot_rb_lookup_hot_range_item (struct hot_range_tree *tree, +u64 start); + +struct hot_inode_item +*hot_rb_lookup_hot_inode_item (struct hot_inode_tree *tree, +unsigned long inode_num); + +void hot_rb_free_hot_inode_item (struct hot_inode_item *he); +int __init hot_rb_item_cache_init (void ); +void hot_rb_update_freq (struct hot_freq_data *freq_data,int rw); +void hot_rb_update_freqs (struct inode *inode,u64 start,u64 len, +int rw); +#endif /* __HOT_MAP__ */
Signature · 1 line
1.7.6.5
RFC 5/11 vfs: add one new mount option '-o hottrack'
In reply to [email protected]From: Zhi Yong Wu <[email protected]>
Introduce one new mount option '-o hottrack',
and add its parsing support.
Its usage looks like:
mount -o hottrack
mount -o nouser,hottrack
mount -o nouser,hottrack,loop
mount -o hottrack,nouser
--- fs/hot_track.c | 34 ++++++++++++++++++++++++++++++++++ fs/hot_track.h | 1 + fs/super.c | 5 +++++ include/linux/hot_track.h | 6 ++++++ 4 files changed, 46 insertions(+), 0 deletions(-)
fs/hot_track.c +34 -0
@@ -16,6 +16,40 @@#include "hot_track.h" /* +* Regular mount options parser for -hottrack option. +* return false if no -hottrack is specified; +* otherwise return true. And the -hottrack will be +* removed from options. +*/ +bool hot_track_parse_options (char *options) +{ +long len; +char *p; +static char opts_hot[] ="hottrack" ; + +if (!options) +return false; + + p =strstr (options, opts_hot); +if (!p) +return false; + +while (p) { + len = options +strlen (options) - (p +strlen (opts_hot)); +if (len ==0 ) { + options[0 ] ='\0' ; +break ; + } + +memmove (p, p +strlen (opts_hot) +1 , len); + p =strstr (options, opts_hot); + } + +printk (KERN_INFO "vfs: turning on hot data tracking\n" ); +return true; +} + +/* * Initialize kmem cache for hot_inode_item * and hot_range_item */
fs/hot_track.h +1 -0
@@ -15,6 +15,7 @@#include "hot_rb.h" +bool hot_track_parse_options (char *options);void __init hot_track_kmem_cache_init (void );#endif /* __HOT_TRACK__ */
fs/super.c +5 -0
@@ -35,6 +35,7 @@#include <linux/fsnotify.h> #include <linux/lockdep.h> #include "internal.h" +#include "hot_track.h" LIST_HEAD (super_blocks);
@@ -1125,6 +1126,7 @@ mount_fs(struct file_system_type *type, int flags, const char *name, void *data)struct dentry *root;struct super_block *sb;char *secdata =NULL ; +bool hottrack = false;int error = -ENOMEM ;if (data && !(type->fs_flags &FS_BINARY_MOUNTDATA )) {
@@ -1137,6 +1139,9 @@ mount_fs(struct file_system_type *type, int flags, const char *name, void *data) goto out_free_secdata; } +if (data &&hot_track_parse_options (data)) + hottrack = true; + root = type->mount (type, flags, name, data);if (IS_ERR (root)) { error =PTR_ERR (root);
include/linux/hot_track.h +6 -0
@@ -20,6 +20,11 @@#include <linux/rbtree.h> #include <linux/kref.h> +/* +* Flags for hot data tracking mount options. +*/ +#define HOT_MOUNT_HOT_TRACK (1 << 0) +/* A tree that sits on the hot_info */ struct hot_inode_tree {struct rb_root map;
@@ -89,6 +94,7 @@ struct hot_range_item { };struct hot_info { +unsigned long mount_opt;/* red-black tree that keeps track of fs-wide hot data */ struct hot_inode_tree hot_inode_tree;
Signature · 1 line
1.7.6.5
RFC 6/11 vfs: add init and exit support
In reply to [email protected]From: Zhi Yong Wu <[email protected]>
Add initialization function to create some
key data structures when hot tracking is enabled;
Clean up them when hot tracking is disabled.
--- fs/hot_rb.c | 45 +++++++++++++++++++++++++++++++++++++++++++++ fs/hot_rb.h | 1 + fs/hot_track.c | 15 +++++++++++++++ fs/hot_track.h | 2 ++ fs/namespace.c | 4 ++++ fs/super.c | 6 ++++++ 6 files changed, 73 insertions(+), 0 deletions(-)
fs/hot_rb.c +45 -0
@@ -133,6 +133,51 @@ static void hot_rb_free_hot_range_item(struct hot_range_item *hr) } } +static int hot_rb_remove_hot_inode_item (struct hot_inode_tree *tree, +struct hot_inode_item *he) +{ +int ret =0 ; +rb_erase (&he->rb_node, &tree->map); + he->in_tree =0 ; +return ret; +} + +static int hot_rb_remove_hot_range_item (struct hot_range_tree *tree, +struct hot_range_item *hr) +{ +int ret =0 ; +rb_erase (&hr->rb_node, &tree->map); + hr->in_tree =0 ; +return ret; +} + +/* Frees the entire hot_inode_tree. */ +void hot_rb_inode_tree_free (struct hot_info *root) +{ +struct rb_node *node, *node2; +struct hot_inode_item *he; +struct hot_range_item *hr; + +/* Free hot inode and range trees on fs root */ + node =rb_first (&root->hot_inode_tree.map); + +while (node) { + he = rb_entry(node, struct hot_inode_item, rb_node); + + node2 =rb_first (&he->hot_range_tree.map); +while (node2) { + hr = rb_entry(node2, struct hot_range_item, rb_node); +hot_rb_remove_hot_range_item (&he->hot_range_tree, hr); +hot_rb_free_hot_range_item (hr); + node2 =rb_first (&he->hot_range_tree.map); + } + +hot_rb_remove_hot_inode_item (&root->hot_inode_tree, he); +hot_rb_free_hot_inode_item (he); + node =rb_first (&root->hot_inode_tree.map); + } +} +static struct rb_node *hot_rb_insert_hot_inode_item (struct rb_root *root,unsigned long inode_num,struct rb_node *node)
fs/hot_rb.h +1 -0
@@ -42,6 +42,7 @@ struct hot_inode_item unsigned long inode_num);void hot_rb_free_hot_inode_item (struct hot_inode_item *he); +void hot_rb_inode_tree_free (struct hot_info *root);int __init hot_rb_item_cache_init (void );
fs/hot_track.c +15 -0
@@ -58,3 +58,18 @@ void __init hot_track_kmem_cache_init(void)if (hot_rb_item_cache_init ())return ; } + +/* +* Initialize the data structures for hot data tracking. +*/ +void hot_track_init (struct super_block *sb,const char *name) +{ + sb->s_hotinfo.mount_opt |=HOT_MOUNT_HOT_TRACK ; +hot_rb_inode_tree_init (&sb->s_hotinfo.hot_inode_tree); +} + +void hot_track_exit (struct super_block *sb) +{ + sb->s_hotinfo.mount_opt &= ~HOT_MOUNT_HOT_TRACK ; +hot_rb_inode_tree_free (&sb->s_hotinfo); +}
fs/hot_track.h +2 -0
@@ -17,5 +17,7 @@bool hot_track_parse_options (char *options);void __init hot_track_kmem_cache_init (void ); +void hot_track_init (struct super_block *sb,const char *name); +void hot_track_exit (struct super_block *sb);#endif /* __HOT_TRACK__ */
fs/namespace.c +4 -0
@@ -22,6 +22,7 @@#include <linux/uaccess.h> #include "pnode.h" #include "internal.h" +#include "hot_track.h" #define HASH_SHIFT ilog2(PAGE_SIZE / sizeof(struct list_head))#define HASH_SIZE (1UL << HASH_SHIFT)
@@ -1215,6 +1216,9 @@ static int do_umount(struct mount *mnt, int flags)return retval; } +if (sb->s_hotinfo.mount_opt &HOT_MOUNT_HOT_TRACK ) +hot_track_exit (sb); +down_write (&namespace_sem);br_write_lock (&vfsmount_lock); event++;
fs/super.c +6 -0
@@ -1153,6 +1153,9 @@ mount_fs(struct file_system_type *type, int flags, const char *name, void *data)WARN_ON (sb->s_bdi == &default_backing_dev_info); sb->s_flags |=MS_BORN ; +if (hottrack) +hot_track_init (sb, name); + error =security_sb_kern_mount (sb, flags, secdata);if (error) goto out_sb;
@@ -1170,6 +1173,9 @@ mount_fs(struct file_system_type *type, int flags, const char *name, void *data)free_secdata (secdata);return root; out_sb: +if (hottrack) +hot_track_exit (sb); +dput (root);deactivate_locked_super (sb); out_free_secdata:
Signature · 1 line
1.7.6.5
RFC 7/11 vfs: introduce one hash table
In reply to [email protected]From: Zhi Yong Wu <[email protected]>
Adds a hash table structure which contains
a lot of hash list and is used to efficiently look up the data temperature of a file or its ranges.
In each hash list of hash table, the hash node
will keep track of temperature info.
--- fs/Makefile | 2 +- fs/hot_hash.c | 67 +++++++++++++++++++++++++++++++++++++++++++++ fs/hot_hash.h | 26 +++++++++++++++++ fs/hot_rb.c | 20 +++++++++++++ fs/hot_rb.h | 2 + fs/hot_track.c | 10 +++++- fs/hot_track.h | 1 + include/linux/hot_track.h | 38 +++++++++++++++++++++++++ 8 files changed, 163 insertions(+), 3 deletions(-) create mode 100644 fs/hot_hash.c create mode 100644 fs/hot_hash.h
fs/Makefile +1 -1
@@ -12,7 +12,7 @@ obj-y := open.o read_write.o file_table.o super.o \ seq_file.o xattr.o libfs.o fs-writeback.o \ pnode.o drop_caches.o splice.o sync.o utimes.o \ stack.o fs_struct.o statfs.o \ - hot_rb.o hot_track.o + hot_rb.o hot_track.o hot_hash.o ifeq ($(CONFIG_BLOCK),y) obj-y += buffer.o bio.o block_dev.o direct-io.o mpage.o ioprio.o
fs/hot_hash.c new file +67 -0
67 lines not on this page · raw message
fs/hot_hash.h new file +26 -0
@@ -0,0 +1,26 @@ +/* +* fs/hot_hash.h +* +* Copyright (C) 2012 IBM Corp. All rights reserved. +* Written by Zhi Yong Wu <[email protected]> +* Ben Chociej <[email protected]> +* +* This program is free software; you can redistribute it and/or +* modify it under the terms of the GNU General Public +* License v2 as published by the Free Software Foundation. +*/ + +#ifndef __HOT_HASH__ +#define __HOT_HASH__ + +#include <linux/list.h> +#include <linux/hash.h> +#include <linux/hot_track.h> + +void hot_hash_node_init (void *_node); + +int __init hot_hash_node_cache_init (void ); + +void hot_hash_table_init (struct hot_info *root); + +#endif /* __HOT_HASH__ */
fs/hot_rb.c +20 -0
@@ -68,6 +68,18 @@ inode_err:return -ENOMEM ; } +void hot_rb_inode_item_exit (void ) +{ +if (hot_inode_item_cache) +kmem_cache_destroy (hot_inode_item_cache); +} + +void hot_rb_range_item_exit (void ) +{ +if (hot_range_item_cache) +kmem_cache_destroy (hot_range_item_cache); +} +/* * Initialize a new hot_inode_item structure. The new structure is * returned with a reference count of one and needs to be
@@ -80,6 +92,10 @@ void hot_rb_inode_item_init(void *_item)memset (he,0 ,sizeof (*he));kref_init (&he->refs);spin_lock_init (&he->lock); + he->heat_node =kmem_cache_alloc (hot_hash_node_cache, +GFP_KERNEL |GFP_NOFS ); +hot_hash_node_init (he->heat_node); + he->heat_node->hot_freq_data = &he->hot_freq_data; he->hot_freq_data.avg_delta_reads = (u64) -1 ; he->hot_freq_data.avg_delta_writes = (u64) -1 ; he->hot_freq_data.flags =FREQ_DATA_TYPE_INODE ;
@@ -98,6 +114,10 @@ void hot_rb_range_item_init(void *_item)memset (hr,0 ,sizeof (*hr));kref_init (&hr->refs);spin_lock_init (&hr->lock); + hr->heat_node =kmem_cache_alloc (hot_hash_node_cache, +GFP_KERNEL |GFP_NOFS ); +hot_hash_node_init (hr->heat_node); + hr->heat_node->hot_freq_data = &hr->hot_freq_data; hr->hot_freq_data.avg_delta_reads = (u64) -1 ; hr->hot_freq_data.avg_delta_writes = (u64) -1 ; hr->hot_freq_data.flags =FREQ_DATA_TYPE_RANGE ;
fs/hot_rb.h +2 -0
@@ -46,6 +46,8 @@ void hot_rb_inode_tree_free(struct hot_info *root);int __init hot_rb_item_cache_init (void ); +void hot_rb_inode_item_exit (void ); +void hot_rb_range_item_exit (void );void hot_rb_update_freq (struct hot_freq_data *freq_data,int rw);void hot_rb_update_freqs (struct inode *inode,u64 start,u64 len,int rw);
fs/hot_track.c +8 -2
@@ -50,13 +50,18 @@ bool hot_track_parse_options(char *options) }/* -* Initialize kmem cache for hot_inode_item -* and hot_range_item +* Initialize kmem cache for hot_inode_item, +* hot_range_item and hot_hash_node */ void __init hot_track_kmem_cache_init (void ) {if (hot_rb_item_cache_init ())return ; + +if (hot_hash_node_cache_init ()) { +hot_rb_inode_item_exit (); +hot_rb_range_item_exit (); + } }/*
@@ -66,6 +71,7 @@ void hot_track_init(struct super_block *sb, const char *name) { sb->s_hotinfo.mount_opt |=HOT_MOUNT_HOT_TRACK ;hot_rb_inode_tree_init (&sb->s_hotinfo.hot_inode_tree); +hot_hash_table_init (&sb->s_hotinfo); }void hot_track_exit (struct super_block *sb)
fs/hot_track.h +1 -0
@@ -14,6 +14,7 @@#define __HOT_TRACK__#include "hot_rb.h" +#include "hot_hash.h" bool hot_track_parse_options (char *options);void __init hot_track_kmem_cache_init (void );
include/linux/hot_track.h +38 -0
@@ -20,11 +20,17 @@#include <linux/rbtree.h> #include <linux/kref.h> +#define HEAT_HASH_BITS 8 +#define HEAT_HASH_SIZE (1 << HEAT_HASH_BITS) +/* * Flags for hot data tracking mount options. */ #define HOT_MOUNT_HOT_TRACK (1 << 0) +/* kmem_cache pointers for slab caches */ +extern struct kmem_cache *hot_hash_node_cache; +/* A tree that sits on the hot_info */ struct hot_inode_tree {struct rb_root map;
@@ -52,6 +58,28 @@ struct hot_freq_data {u32 last_temperature; }; +/* Hash list heads for hot hash table */ +struct hot_hash_head { +struct hlist_head hashhead; +rwlock_t rwlock; +u32 temperature; +}; + +/* Nodes stored in each hash list of hash table */ +struct hot_hash_node { +struct hlist_node hashnode; +struct list_head node; +struct hot_freq_data *hot_freq_data; +struct hot_hash_head *hlist; +spinlock_t lock;/* protects hlist */ + +/* +* number of references to this node +* equals 1 (hashlist entry) +*/ +struct kref refs; +}; +/* An item representing an inode and its access frequency */ struct hot_inode_item {/* node for hot_inode_tree rb_tree */
@@ -68,6 +96,8 @@ struct hot_inode_item {spinlock_t lock;/* prevents kfree */ struct kref refs; +/* hashlist node for this inode */ +struct hot_hash_node *heat_node; };/*
@@ -91,6 +121,8 @@ struct hot_range_item {spinlock_t lock;/* prevents kfree */ struct kref refs; +/* hashlist node for this range */ +struct hot_hash_node *heat_node; };struct hot_info {
@@ -98,6 +130,12 @@ struct hot_info {/* red-black tree that keeps track of fs-wide hot data */ struct hot_inode_tree hot_inode_tree; + +/* hash map of inode temperature */ +struct hot_hash_head heat_inode_hl[HEAT_HASH_SIZE ]; + +/* hash map of range temperature */ + struct hot_hash_head heat_range_hl[HEAT_HASH_SIZE]; };#endif /* _LINUX_HOTTRACK_H */
Signature · 1 line
1.7.6.5
RFC 8/11 vfs: enable hot data tracking
In reply to [email protected]From: Zhi Yong Wu <[email protected]>
Miscellaneous features that implement hot data tracking
and generally make the hot data functions a bit more friendly.
--- fs/direct-io.c | 10 ++++++++++ include/linux/hot_track.h | 11 +++++++++++ mm/filemap.c | 8 ++++++++ mm/page-writeback.c | 21 +++++++++++++++++++++ mm/readahead.c | 9 +++++++++ 5 files changed, 59 insertions(+), 0 deletions(-)
fs/direct-io.c +10 -0
@@ -37,6 +37,7 @@#include <linux/uio.h> #include <linux/atomic.h> #include <linux/prefetch.h> +#include <linux/hot_track.h> /* * How many user pages to map in one call to get_user_pages(). This determines
@@ -1297,6 +1298,15 @@ __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,prefetch (bdev->bd_queue);prefetch ((char *)bdev->bd_queue +SMP_CACHE_BYTES ); +/* Hot data tracking */ +if (TRACK_THIS_INODE (iocb->ki_filp->f_mapping->host) + &&iov_length (iov, nr_segs) >0 ) { +hot_rb_update_freqs (iocb->ki_filp->f_mapping->host, + (u64 )offset, + (u64 )iov_length (iov, nr_segs), + rw &WRITE ); + } +return do_blockdev_direct_IO (rw, iocb, inode, bdev, iov, offset, nr_segs, get_block, end_io, submit_io, flags);
include/linux/hot_track.h +11 -0
@@ -28,6 +28,14 @@*/ #define HOT_MOUNT_HOT_TRACK (1 << 0) +/* Hot data tracking -- guard macros */ +#define TRACKING_HOT_TRACK (root) \ + (root->s_hotinfo.mount_opt & HOT_MOUNT_HOT_TRACK) + +#define TRACK_THIS_INODE (inode) \ + ((TRACKING_HOT_TRACK(inode->i_sb)) && \ + !(inode->i_flags & S_NOHOTDATATRACK)) +/* kmem_cache pointers for slab caches */ extern struct kmem_cache *hot_hash_node_cache;
@@ -138,4 +146,7 @@ struct hot_info { struct hot_hash_head heat_range_hl[HEAT_HASH_SIZE]; }; +extern void hot_rb_update_freqs (struct inode *inode, +u64 start,u64 len,int rw); +#endif /* _LINUX_HOTTRACK_H */
mm/filemap.c +8 -0
@@ -33,6 +33,7 @@#include <linux/hardirq.h> /* for BUG_ON(!in_atomic()) only */ #include <linux/memcontrol.h> #include <linux/cleancache.h> +#include <linux/hot_track.h> #include "internal.h" /*
@@ -1224,6 +1225,13 @@ readpage:* PG_error will be set again if readpage fails. */ ClearPageError (page); + +/* Hot data tracking */ +if (TRACK_THIS_INODE (filp->f_mapping->host)) +hot_rb_update_freqs (filp->f_mapping->host, + (u64 )page->index <<PAGE_CACHE_SHIFT , +PAGE_CACHE_SIZE ,0 ); +/* Start the actual read. The read will unlock the page. */ error = mapping->a_ops->readpage (filp, page);
mm/page-writeback.c +21 -0
@@ -35,6 +35,7 @@#include <linux/buffer_head.h> /* __set_page_dirty_buffers */ #include <linux/pagevec.h> #include <linux/timer.h> +#include <linux/hot_track.h> #include <trace/events/writeback.h> /*
@@ -1895,13 +1896,33 @@ EXPORT_SYMBOL(generic_writepages);int do_writepages (struct address_space *mapping,struct writeback_control *wbc) {int ret; +pgoff_t start =0 ; +u64 prev_count =0 , count =0 ;if (wbc->nr_to_write <=0 )return 0 ; + +/* Hot data tracking */ +if (TRACK_THIS_INODE (mapping->host) + && wbc->range_cyclic) { + start = mapping->writeback_index <<PAGE_CACHE_SHIFT ; + prev_count = (u64 )wbc->nr_to_write; + } +if (mapping->a_ops->writepages) ret = mapping->a_ops->writepages (mapping, wbc);else ret =generic_writepages (mapping, wbc); + +/* Hot data tracking */ +if (TRACK_THIS_INODE (mapping->host) + && wbc->range_cyclic) { + count = prev_count - (u64 )wbc->nr_to_write; +if (count) +hot_rb_update_freqs (mapping->host, (u64 )start, + count *PAGE_CACHE_SIZE ,1 ); + } +return ret; }
mm/readahead.c +9 -0
@@ -19,6 +19,7 @@#include <linux/pagemap.h> #include <linux/syscalls.h> #include <linux/file.h> +#include <linux/hot_track.h> /* * Initialise a struct file's readahead state. Assumes that the caller has
@@ -138,6 +139,14 @@ static int read_pages(struct address_space *mapping, struct file *filp, out:blk_finish_plug (&plug); +/* Hot data tracking */ +if (TRACK_THIS_INODE (mapping->host) && nr_pages >0 ) { +u64 start = (u64)(list_entry (pages->prev, + struct page, lru)->index) << PAGE_CACHE_SHIFT; +hot_rb_update_freqs (mapping->host, start, + (u64 )nr_pages *PAGE_CACHE_SIZE ,0 ); + } +return ret; }
Signature · 1 line
1.7.6.5
RFC 9/11 vfs: fork one private kthread to update temperature info
In reply to [email protected]From: Zhi Yong Wu <[email protected]>
Fork and run one kernel kthread to calculate
that temperature based on some metrics kept
in custom frequency data structs, and store
the info in the hash table.
--- fs/hot_hash.c | 302 +++++++++++++++++++++++++++++++++++++++++++++ fs/hot_hash.h | 82 ++++++++++++ fs/hot_rb.c | 132 +++++++++++++++++++- fs/hot_rb.h | 15 +++ fs/hot_track.c | 31 +++++ include/linux/hot_track.h | 5 + 6 files changed, 561 insertions(+), 6 deletions(-)
fs/hot_hash.c +302 -0
@@ -18,6 +18,7 @@#include <linux/hardirq.h> #include <linux/hash.h> #include <linux/types.h> +#include <linux/kthread.h> #include <linux/freezer.h> #include <linux/fs.h> #include "hot_rb.h"
@@ -65,3 +66,304 @@ void hot_hash_table_init(struct hot_info *root) } } +static void hot_hash_node_free (struct hlist_head *head) +{ +struct hlist_node *pos =NULL , *pos2 =NULL ; +struct hot_hash_node *heatnode =NULL ; + +hlist_for_each_safe (pos, pos2, head) { + heatnode =hlist_entry (pos, + struct hot_hash_node, + hashnode); +hlist_del (pos); +kfree (heatnode); + } + +} + +void hot_hash_table_free (struct hot_info *root) +{ +int i; + +/* Free node/range heat hash lists */ +for (i =0 ; i <HEAT_HASH_SIZE ; i++) { +hot_hash_node_free (&root->heat_inode_hl[i].hashhead); +hot_hash_node_free (&root->heat_range_hl[i].hashhead); + } +} + +static u64 hot_hash_shift (u64 counter,u32 bits,bool shift_dir) +{ +if (shift_dir) +return counter << bits; +else +return counter >> bits; +} + +/* +* hot_hash_calc_temperature() is responsible for distilling the six heat +* criteria, which are described in detail in hot_hash.h) down into a single +* temperature value for the data, which is an integer between 0 +* and HEAT_MAX_VALUE. +* +* To accomplish this, the raw values from the hot_freq_data structure +* are shifted various ways in order to make the temperature calculation more +* or less sensitive to each value. +* +* Once this calibration has happened, we do some additional normalization and +* make sure that everything fits nicely in a u32. From there, we take a very +* rudimentary kind of "average" of each of the values, where the *_COEFF_POWER +* values act as weights for the average. +* +* Finally, we use the HEAT_HASH_BITS value, which determines the size of the +* heat hash list, to normalize the temperature to the proper granularity. +*/ +int hot_hash_calc_temperature (struct hot_freq_data *freq_data) +{ +u64 result =0 ; + +struct timespec ckt =current_kernel_time (); +u64 cur_time =timespec_to_ns (&ckt); + +u32 nrr_heat = (u32 )hot_hash_shift ((u64 )freq_data->nr_reads, +NRR_MULTIPLIER_POWER , true); +u32 nrw_heat = (u32 )hot_hash_shift ((u64 )freq_data->nr_writes, +NRW_MULTIPLIER_POWER , true); + +u64 ltr_heat = +hot_hash_shift ((cur_time -timespec_to_ns (&freq_data->last_read_time)), +LTR_DIVIDER_POWER , false); +u64 ltw_heat = +hot_hash_shift ((cur_time -timespec_to_ns (&freq_data->last_write_time)), +LTW_DIVIDER_POWER , false); + +u64 avr_heat = +hot_hash_shift ((((u64) -1 ) - freq_data->avg_delta_reads), +AVR_DIVIDER_POWER , false); +u64 avw_heat = +hot_hash_shift ((((u64) -1 ) - freq_data->avg_delta_writes), +AVW_DIVIDER_POWER , false); + +/* ltr_heat is now guaranteed to be u32 safe */ +if (ltr_heat >=hot_hash_shift ((u64 )1 ,32 , true)) + ltr_heat =0 ; +else + ltr_heat =hot_hash_shift ((u64 )1 ,32 , true) - ltr_heat; + +/* ltw_heat is now guaranteed to be u32 safe */ +if (ltw_heat >=hot_hash_shift ((u64 )1 ,32 , true)) + ltw_heat =0 ; +else + ltw_heat =hot_hash_shift ((u64 )1 ,32 , true) - ltw_heat; + +/* avr_heat is now guaranteed to be u32 safe */ +if (avr_heat >=hot_hash_shift ((u64 )1 ,32 , true)) + avr_heat = (u32) -1 ; + +/* avw_heat is now guaranteed to be u32 safe */ +if (avw_heat >=hot_hash_shift ((u64 )1 ,32 , true)) + avw_heat = (u32) -1 ; + + nrr_heat = (u32 )hot_hash_shift ((u64 )nrr_heat, + (3 -NRR_COEFF_POWER ), false); + nrw_heat = (u32 )hot_hash_shift ((u64 )nrw_heat, + (3 -NRW_COEFF_POWER ), false); + ltr_heat =hot_hash_shift (ltr_heat, (3 -LTR_COEFF_POWER ), false); + ltw_heat =hot_hash_shift (ltw_heat, (3 -LTW_COEFF_POWER ), false); + avr_heat =hot_hash_shift (avr_heat, (3 -AVR_COEFF_POWER ), false); + avw_heat =hot_hash_shift (avw_heat, (3 -AVW_COEFF_POWER ), false); + + result = nrr_heat + nrw_heat + (u32 ) ltr_heat + + (u32 ) ltw_heat + (u32 ) avr_heat + (u32 ) avw_heat; + +return result >> (32 -HEAT_HASH_BITS ); +} + +int hot_hash_is_aging (struct hot_freq_data *freq_data) +{ +int ret =0 ; +struct timespec ckt =current_kernel_time (); + +u64 cur_time =timespec_to_ns (&ckt); +u64 last_read_ns = + (cur_time -timespec_to_ns (&freq_data->last_read_time)); +u64 last_write_ns = + (cur_time -timespec_to_ns (&freq_data->last_write_time)); +u64 kick_ns =TIME_TO_KICK * (u64 )1000000000 ; + +if ((last_read_ns > kick_ns) && (last_write_ns > kick_ns)) + ret =1 ; + +return ret; +} + +/* +* Calc a new temperature and, if necessary, move the heat_node corresponding +* to this inode or range to the proper hashlist with the new temperature +*/ +void hot_hash_update_hash_table (struct hot_freq_data *freq_data, +struct hot_info *root) +{ +int temperature =0 ; +int moved =0 ; +struct hot_hash_head *buckets, *current_bucket =NULL ; +struct hot_inode_item *he; +struct hot_range_item *hr; + +if (freq_data->flags &FREQ_DATA_TYPE_INODE ) { + he =hot_freq_data_get_he (freq_data); + buckets = root->heat_inode_hl; + +spin_lock (&he->lock); + temperature =hot_hash_calc_temperature (freq_data); + freq_data->last_temperature = temperature; +spin_unlock (&he->lock); + +if (he ==NULL ) +return ; + +spin_lock (&he->heat_node->lock); +if (he->heat_node->hlist ==NULL ) { + current_bucket = buckets + temperature; + moved =1 ; + }else { +write_lock (&he->heat_node->hlist->rwlock); + current_bucket = he->heat_node->hlist; +if (current_bucket->temperature != temperature) { +hlist_del (&he->heat_node->hashnode); + current_bucket = buckets + temperature; + moved =1 ; + } +write_unlock (&he->heat_node->hlist->rwlock); + } + +if (moved) { +write_lock (¤t_bucket->rwlock); +hlist_add_head (&he->heat_node->hashnode, + ¤t_bucket->hashhead); + he->heat_node->hlist = current_bucket; +write_unlock (¤t_bucket->rwlock); + } +spin_unlock (&he->heat_node->lock); + }else if (freq_data->flags &FREQ_DATA_TYPE_RANGE ) { + hr =hot_freq_data_get_hr (freq_data); + buckets = root->heat_range_hl; + +spin_lock (&hr->lock); + temperature =hot_hash_calc_temperature (freq_data); + freq_data->last_temperature = temperature; +spin_unlock (&hr->lock); + +if (hr ==NULL ) +return ; + +spin_lock (&hr->heat_node->lock); +if (hr->heat_node->hlist ==NULL ) { + current_bucket = buckets + temperature; + moved =1 ; + }else { +write_lock (&hr->heat_node->hlist->rwlock); + current_bucket = hr->heat_node->hlist; +if (current_bucket->temperature != temperature) { +hlist_del (&hr->heat_node->hashnode); + current_bucket = buckets + temperature; + moved =1 ; + } +write_unlock (&hr->heat_node->hlist->rwlock); + } + +if (moved) { +write_lock (¤t_bucket->rwlock); +hlist_add_head (&hr->heat_node->hashnode, + ¤t_bucket->hashhead); + hr->heat_node->hlist = current_bucket; +write_unlock (¤t_bucket->rwlock); + } +spin_unlock (&hr->heat_node->lock); + } +} + +/* +* Update temperatures for each hot inode item and +* hot range item for aging purposes +*/ +static void hot_hash_temperature_update_iterator (struct hot_info *root) +{ +struct hot_inode_item *current_hot_inode; +struct hot_inode_tree *hot_inode_tree; +unsigned long inode_num; + + hot_inode_tree = &root->hot_inode_tree; + +/* walk the inode tree */ + current_hot_inode =hot_rb_find_next_hot_inode (root,0 ); +while (current_hot_inode) { +hot_hash_update_hash_table ( + ¤t_hot_inode->hot_freq_data, root); +hot_rb_update_range_data (current_hot_inode, root); + inode_num = current_hot_inode->i_ino; +hot_rb_free_hot_inode_item (current_hot_inode); + current_hot_inode =hot_rb_find_next_hot_inode (root, + inode_num +1 ); + } +} + +/* Determine if there is hot data tracking to be enabled */ +static bool hot_hash_global_hot_track (void ) +{ +struct super_block *sb; +bool ret = false; + +spin_lock (&sb_lock); +list_for_each_entry (sb, &super_blocks, s_list) { +if (hlist_unhashed (&sb->s_instances)) +continue ; +if (sb->s_hotinfo.mount_opt &HOT_MOUNT_HOT_TRACK ) + ret = true; + } +spin_unlock (&sb_lock); + +return ret; +} + +/* +* kthread iterates each hot_inode_item and hot_range_item +* and update temperatures to be shifted in heat hash table +* for purposes of relocation and such hot file detection +*/ +int hot_hash_temperature_update_kthread (void *arg) +{ +struct super_block *sb; +struct hot_info *root; +unsigned long delay; + +do { +spin_lock (&sb_lock); +list_for_each_entry (sb, &super_blocks, s_list) { +if (hlist_unhashed (&sb->s_instances)) +continue ; + delay =HZ *HEAT_UPDATE_DELAY ; + root = &sb->s_hotinfo; +if (mutex_trylock ( + &root->hot_data_update_kthread_mutex)) { +hot_hash_temperature_update_iterator (root); +mutex_unlock ( + &root->hot_data_update_kthread_mutex); + } +if (unlikely (freezing (current))) { +__refrigerator (true); + }else { +set_current_state (TASK_INTERRUPTIBLE ); +if (!kthread_should_stop ()) { +spin_unlock (&sb_lock); +schedule_timeout (delay); +spin_lock (&sb_lock); + } +__set_current_state (TASK_RUNNING ); + } + } +spin_unlock (&sb_lock); + }while (!kthread_should_stop () || !hot_hash_global_hot_track ()); + +return 0 ; +}
fs/hot_hash.h +82 -0
92 lines not on this page · raw message
fs/hot_rb.c +126 -6
153 lines not on this page · raw message
fs/hot_rb.h +15 -0
@@ -48,8 +48,23 @@ int __init hot_rb_item_cache_init(void);void hot_rb_inode_item_exit (void );void hot_rb_range_item_exit (void ); + +/* +* recalculates temperatures for inode or range +* and moves around in heat hash table based on temp +*/ +void hot_rb_update_heat_hash_list (struct hot_freq_data *freq_data, +struct hot_info *root); + +struct hot_inode_item +*hot_rb_find_next_hot_inode (struct hot_info *root, +u64 objectid);void hot_rb_update_freq (struct hot_freq_data *freq_data,int rw);void hot_rb_update_freqs (struct inode *inode,u64 start,u64 len,int rw); +/* Update temperatures for each range item for aging purposes */ +void hot_rb_update_range_data (struct hot_inode_item *hot_inode, +struct hot_info *root); +#endif /* __HOT_MAP__ */
fs/hot_track.c +31 -0
@@ -13,8 +13,11 @@#include <linux/slab.h> #include <linux/gfp.h> #include <linux/parser.h> +#include <linux/kthread.h> #include "hot_track.h" +struct task_struct *hot_track_temperature_update_kthread; +/* * Regular mount options parser for -hottrack option. * return false if no -hottrack is specified;
@@ -65,6 +68,30 @@ void __init hot_track_kmem_cache_init(void) }/* +* Fork and initialize kthread for each new mount point that +* periodically goes through hot inodes and hot ranges and ages them +* based on frequency of access +*/ +static void hot_track_fork_temperature_update_kthread (void ) +{ +if (hot_track_temperature_update_kthread) +return ; + + hot_track_temperature_update_kthread = +kthread_run (hot_hash_temperature_update_kthread,NULL , +"hot_temperature_update_kthread" ); +if (IS_ERR (hot_track_temperature_update_kthread)) +kthread_stop (hot_track_temperature_update_kthread); +} + +/* Stop the kthread to do temperature updates for all filesystems */ +static void hot_track_stop_temperature_update_kthread (void ) +{ +if (hot_track_temperature_update_kthread) +kthread_stop (hot_track_temperature_update_kthread); +} + +/* * Initialize the data structures for hot data tracking. */ void hot_track_init (struct super_block *sb,const char *name)
@@ -72,10 +99,14 @@ void hot_track_init(struct super_block *sb, const char *name) sb->s_hotinfo.mount_opt |=HOT_MOUNT_HOT_TRACK ;hot_rb_inode_tree_init (&sb->s_hotinfo.hot_inode_tree);hot_hash_table_init (&sb->s_hotinfo); +hot_track_fork_temperature_update_kthread (); +hot_debugfs_volume_init (name, sb); }void hot_track_exit (struct super_block *sb) { sb->s_hotinfo.mount_opt &= ~HOT_MOUNT_HOT_TRACK ; +hot_track_stop_temperature_update_kthread (); +hot_hash_table_free (&sb->s_hotinfo);hot_rb_inode_tree_free (&sb->s_hotinfo); }
include/linux/hot_track.h +5 -0
@@ -36,6 +36,8 @@ ((TRACKING_HOT_TRACK (inode->i_sb)) && \ !(inode->i_flags & S_NOHOTDATATRACK)) +extern struct task_struct *hot_track_temperature_update_kthread; +/* kmem_cache pointers for slab caches */ extern struct kmem_cache *hot_hash_node_cache;
@@ -144,6 +146,9 @@ struct hot_info {/* hash map of range temperature */ struct hot_hash_head heat_range_hl[HEAT_HASH_SIZE ]; + +/* protects hot data items while being iterated and updated */ +struct mutex hot_data_update_kthread_mutex; }; extern void hot_rb_update_freqs(struct inode *inode,
Signature · 1 line
1.7.6.5
RFC 10/11 vfs: add 3 new ioctl interfaces
In reply to [email protected]From: Zhi Yong Wu <[email protected]>
FS_IOC_GET_HEAT_INFO: return a struct containing the various
metrics collected in btrfs_freq_data structs, and also return a calculated data temperature based on those metrics. Optionally, retrieve the temperature from the hot data hash list instead of recalculating it.
FS_IOC_GET_HEAT_OPTS: return an integer representing the current
state of hot data tracking and migration:
0 = do nothing
1 = track frequency of access
FS_IOC_SET_HEAT_OPTS: change the state of hot data tracking and migration, as described above.
--- fs/compat_ioctl.c | 8 +++ fs/ioctl.c | 132 +++++++++++++++++++++++++++++++++++++++++++++ include/linux/fs.h | 11 ++++ include/linux/hot_track.h | 12 ++++ 4 files changed, 163 insertions(+), 0 deletions(-)
fs/compat_ioctl.c +8 -0
@@ -1390,6 +1390,11 @@ COMPATIBLE_IOCTL(TIOCSTART) COMPATIBLE_IOCTL(TIOCSTOP) #endif +/*Hot data tracking*/ +COMPATIBLE_IOCTL(FS_IOC_GET_HEAT_INFO) +COMPATIBLE_IOCTL(FS_IOC_SET_HEAT_OPTS) +COMPATIBLE_IOCTL(FS_IOC_GET_HEAT_OPTS) + /* fat 'r' ioctls. These are handled by fat with ->compat_ioctl, but we don't want warnings on other file systems. So declare them as compatible here. */
@@ -1572,6 +1577,9 @@ asmlinkage long compat_sys_ioctl(unsigned int fd, unsigned int cmd,case FIBMAP :case FIGETBSZ :case FIONREAD : +case FS_IOC_GET_HEAT_INFO : +case FS_IOC_SET_HEAT_OPTS : +case FS_IOC_GET_HEAT_OPTS :if (S_ISREG (filp->f_path.dentry->d_inode->i_mode))break ;/*FALL THROUGH*/
fs/ioctl.c +132 -0
150 lines not on this page · raw message
include/linux/fs.h +11 -0
@@ -256,6 +256,7 @@ struct inodes_stat_t {#define S_IMA 1024/* Inode has an associated IMA struct */ #define S_AUTOMOUNT 2048/* Automount/referral quasi-directory */ #define S_NOSEC 4096/* no suid or xattr security attributes */ +#define S_NOHOTDATATRACK (1 << 13)/* hot data tracking */ /* * Note that nosuid etc flags are inode-specific: setting some file-system
@@ -354,6 +355,16 @@ struct inodes_stat_t {#define FS_IOC32_SETVERSION _IOW('v', 2, int)/* +* Hot data tracking ioctls: +* +* HOT_INFO - retrieve info on frequency of access +*/ +#define FS_IOC_GET_HEAT_INFO _IOR('f', 17, \ + struct heat_info) +#define FS_IOC_SET_HEAT_OPTS _IOW('f', 18, int) +#define FS_IOC_GET_HEAT_OPTS _IOR('f', 19, int) + +/* * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS) */ #define FS_SECRM_FL 0x00000001/* Secure deletion */
include/linux/hot_track.h +12 -0
@@ -68,6 +68,18 @@ struct hot_freq_data {u32 last_temperature; }; +struct heat_info { +__u64 avg_delta_reads; +__u64 avg_delta_writes; +__u64 last_read_time; +__u64 last_write_time; +__u32 num_reads; +__u32 num_writes; +__u32 temperature; +__u8 live; +char filename[PATH_MAX ]; +}; +/* Hash list heads for hot hash table */ struct hot_hash_head {struct hlist_head hashhead;
Signature · 1 line
1.7.6.5
RFC 11/11 vfs: add debugfs support
In reply to [email protected]From: Zhi Yong Wu <[email protected]>
Add a /sys/kernel/debug/hot_track/<device_name>/ directory for each
volume that contains two files. The first, inode_data', contains the heat information for inodes that have been brought into the hot data map structures. The second, range_data', contains similar information for subfile ranges.
--- fs/Makefile | 2 +- fs/hot_debugfs.c | 487 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ fs/hot_debugfs.h | 60 +++++++ fs/hot_track.c | 1 + fs/hot_track.h | 3 +- fs/namespace.c | 6 + 6 files changed, 556 insertions(+), 3 deletions(-) create mode 100644 fs/hot_debugfs.c create mode 100644 fs/hot_debugfs.h
fs/Makefile +1 -1
@@ -12,7 +12,7 @@ obj-y := open.o read_write.o file_table.o super.o \ seq_file.o xattr.o libfs.o fs-writeback.o \ pnode.o drop_caches.o splice.o sync.o utimes.o \ stack.o fs_struct.o statfs.o \ - hot_rb.o hot_track.o hot_hash.o + hot_rb.o hot_track.o hot_hash.o hot_debugfs.o ifeq ($(CONFIG_BLOCK),y) obj-y += buffer.o bio.o block_dev.o direct-io.o mpage.o ioprio.o
fs/hot_debugfs.c new file +487 -0
487 lines not on this page · raw message
fs/hot_debugfs.h new file +60 -0
60 lines not on this page · raw message
fs/hot_track.c +1 -0
@@ -109,4 +109,5 @@ void hot_track_exit(struct super_block *sb)hot_track_stop_temperature_update_kthread ();hot_hash_table_free (&sb->s_hotinfo);hot_rb_inode_tree_free (&sb->s_hotinfo); +hot_debugfs_volume_exit (sb); }
fs/hot_track.h +1 -2
@@ -13,8 +13,7 @@#ifndef __HOT_TRACK__#define __HOT_TRACK__ -#include "hot_rb.h" -#include "hot_hash.h" +#include "hot_debugfs.h" bool hot_track_parse_options (char *options);void __init hot_track_kmem_cache_init (void );
fs/namespace.c +6 -0
@@ -2629,6 +2629,12 @@ void __init mnt_init(void) fs_kobj =kobject_create_and_add ("fs" ,NULL );if (!fs_kobj)printk (KERN_WARNING "%s: kobj create error\n" , __func__); + + err =hot_debugfs_init (); +if (err) +printk (KERN_WARNING "%s: sysfs_init error: %d\n" , + __func__, err); +init_rootfs ();init_mount_tree (); }
Signature · 1 line
1.7.6.5
2012/9/17 <[email protected]>:
Quoted 104 lines
From: Zhi Yong Wu <[email protected]>
NOTE:
The patchset is currently post out mainly to make sure
it is going in the correct direction and hope to get some helpful comments from other guys.TODO List:
1.) Need to do scalability or performance tests.
2.) Turn some Micro into tunables
TIME_TO_KICK, andHEAT_UPDATE_DELAY
3.) Rafactorhot_hash_is_aging()
If you just made the timeout value a timespec and compared
the _timespecs_, you would be doing a lot fewer conversions.
4.) Cleanup some unnecessary lock protect
5.) Add more comments to explain how to calc temperatureBen Chociej, Matt Lupfer and Conor Scott originally wrote this code to
be very btrfs-specific. I've taken their code and attempted to
make it more generic and integrate it at the VFS level.INTRODUCTION:
Essentially, this means maintaining some key stats
(like number of reads/writes, last read/write time, frequency of reads/writes), then distilling those numbers down to a single "temperature" value that reflects what data is "hot," and using that temperature to move data to SSDs.The long-term goal of these patches is to allow some FSs,
e.g. Btrfs to intelligently utilize SSDs in a heterogenous volume. Incidentally, this project has been motivated by
the Project Ideas page on the Btrfs wiki.Of course, users are warned not to run this code outside of development
environments. These patches are EXPERIMENTAL, and as such they might eat your data and/or memory. That said, the code should be relatively safe when the hottrack mount option are disabled.MOTIVATION:
The overall goal of enabling hot data relocation to SSD has been
motivated by the Project Ideas page on the Btrfs wiki at
<https://btrfs.wiki.kernel.org/index.php/Project_ideas>.
It will divide into two steps. VFS provide hot data tracking function while specific FS will provide hot data relocation function. So as the first step of this goal, it is hoped that the patchset for hot data tracking will eventually mature into VFS.This is essentially the traditional cache argument: SSD is fast and
expensive; HDD is cheap but slow. ZFS, for example, can already take advantage of SSD caching. Btrfs should also be able to take advantage of hybrid storage without many broad, sweeping changes to existing code.SUMMARY:
- Hooks in existing vfs functions to track data access frequency
- New rbtrees for tracking access frequency of inodes and sub-file ranges (
hot_rb.c)
The relationship betweensuper_blockand rbtree is as below:
super_block->s_hotinfo.hot_inode_tree
Ininclude/linux/fs.h, one structhot_infos_hotinfois added to
super_blockstruct. Each FS instance can find hot tracking infos_hotinfovia itssuper_block. In thishot_info, it store a lot of hot tracking info such ashot_inode_tree, inode and range hash list, etc.- A hash list for indexing data by its temperature (
hot_hash.c)- A debugfs interface for dumping data from the rbtrees (
hot_debugfs.c)- A background kthread for updating inode heat info
- Mount options for enabling temperature
tracking(-o hottrack, default mean disabled)(hot_track.c)- An ioctl to retrieve the frequency information collected for a certain file
- Ioctls to enable/disable frequency tracking per inode.
Usage syntax:
root@debian-i386:~# mount -o hottrack /dev/sdb /mnt [ 1505.894078] device label test devid 1 transid 29 /dev/sdb [ 1505.952977] btrfs: disk space caching is enabled [ 1506.069678] vfs: turning on hot data tracking root@debian-i386:~# mount -t debugfs none /sys/kernel/debug root@debian-i386:~# ls -l /sys/kernel/debug/vfs_hotdata/ total 0 drwxr-xr-x 2 root root 0 Aug 8 04:40 sdb root@debian-i386:~# ls -l /sys/kernel/debug/vfs_hotdata/sdb total 0(quoted diff) +0
-2-rw-r--r-- 1 root root 0 Aug 8 04:40 inode_data -rw-r--r-- 1 root root 0 Aug 8 04:40 range_dataroot@debian-i386:~# vi /mnt/file root@debian-i386:~# cat /sys/kernel/debug/hot_track/sdb/inode_data inode #279, reads 0, writes 1, avg read time 18446744073709551615, avg write time 5251566408153596, temp 109 root@debian-i386:~# cat /sys/kernel/debug/hot_track/sdb/range_data inode #279, range start 0 (range len 1048576) reads 0, writes 1, avg read time 18446744073709551615, avg write time 1128690176623144209, temp 64
Quoted 9 lines
root@debian-i386:~# echo "hot data tracking test" >> /mnt/file root@debian-i386:~# cat /sys/kernel/debug/hot_track/sdb/inode_data inode #279, reads 0, writes 2, avg read time 18446744073709551615, avg write time 4923343766042451, temp 109 root@debian-i386:~# cat /sys/kernel/debug/hot_track/sdb/range_data inode #279, range start 0 (range len 1048576) reads 0, writes 2, avg read time 18446744073709551615, avg write time 1058147040842596150, temp 64 root@debian-i386:~#
It's a good idea to add a new file under documentation and include this kind of information. For example what temp means, how it's worked out and how to "read" the avg read/write time (nanoseconds, microseconds, jiffies....??)
Marco
Quoted 111 lines · Marco Stornelli
2012/9/17 <[email protected]>:
Quoted 110 lines
From: Zhi Yong Wu <[email protected]>
NOTE:
The patchset is currently post out mainly to make sure
it is going in the correct direction and hope to get some helpful comments from other guys.TODO List:
1.) Need to do scalability or performance tests.
2.) Turn some Micro into tunables
TIME_TO_KICK, andHEAT_UPDATE_DELAY
3.) Rafactorhot_hash_is_aging()
If you just made the timeout value a timespec and compared
the _timespecs_, you would be doing a lot fewer conversions.
4.) Cleanup some unnecessary lock protect
5.) Add more comments to explain how to calc temperatureBen Chociej, Matt Lupfer and Conor Scott originally wrote this code to
be very btrfs-specific. I've taken their code and attempted to
make it more generic and integrate it at the VFS level.INTRODUCTION:
Essentially, this means maintaining some key stats
(like number of reads/writes, last read/write time, frequency of reads/writes), then distilling those numbers down to a single "temperature" value that reflects what data is "hot," and using that temperature to move data to SSDs.The long-term goal of these patches is to allow some FSs,
e.g. Btrfs to intelligently utilize SSDs in a heterogenous volume. Incidentally, this project has been motivated by
the Project Ideas page on the Btrfs wiki.Of course, users are warned not to run this code outside of development
environments. These patches are EXPERIMENTAL, and as such they might eat your data and/or memory. That said, the code should be relatively safe when the hottrack mount option are disabled.MOTIVATION:
The overall goal of enabling hot data relocation to SSD has been
motivated by the Project Ideas page on the Btrfs wiki at
<https://btrfs.wiki.kernel.org/index.php/Project_ideas>.
It will divide into two steps. VFS provide hot data tracking function while specific FS will provide hot data relocation function. So as the first step of this goal, it is hoped that the patchset for hot data tracking will eventually mature into VFS.This is essentially the traditional cache argument: SSD is fast and
expensive; HDD is cheap but slow. ZFS, for example, can already take advantage of SSD caching. Btrfs should also be able to take advantage of hybrid storage without many broad, sweeping changes to existing code.SUMMARY:
- Hooks in existing vfs functions to track data access frequency
- New rbtrees for tracking access frequency of inodes and sub-file ranges (
hot_rb.c)
The relationship betweensuper_blockand rbtree is as below:
super_block->s_hotinfo.hot_inode_tree
Ininclude/linux/fs.h, one structhot_infos_hotinfois added to
super_blockstruct. Each FS instance can find hot tracking infos_hotinfovia itssuper_block. In thishot_info, it store a lot of hot tracking info such ashot_inode_tree, inode and range hash list, etc.- A hash list for indexing data by its temperature (
hot_hash.c)- A debugfs interface for dumping data from the rbtrees (
hot_debugfs.c)- A background kthread for updating inode heat info
- Mount options for enabling temperature
tracking(-o hottrack, default mean disabled)(hot_track.c)- An ioctl to retrieve the frequency information collected for a certain file
- Ioctls to enable/disable frequency tracking per inode.
Usage syntax:
root@debian-i386:~# mount -o hottrack /dev/sdb /mnt [ 1505.894078] device label test devid 1 transid 29 /dev/sdb [ 1505.952977] btrfs: disk space caching is enabled [ 1506.069678] vfs: turning on hot data tracking root@debian-i386:~# mount -t debugfs none /sys/kernel/debug root@debian-i386:~# ls -l /sys/kernel/debug/vfs_hotdata/ total 0 drwxr-xr-x 2 root root 0 Aug 8 04:40 sdb root@debian-i386:~# ls -l /sys/kernel/debug/vfs_hotdata/sdb total 0(quoted diff) +0
-2-rw-r--r-- 1 root root 0 Aug 8 04:40 inode_data -rw-r--r-- 1 root root 0 Aug 8 04:40 range_dataroot@debian-i386:~# vi /mnt/file root@debian-i386:~# cat /sys/kernel/debug/hot_track/sdb/inode_data inode #279, reads 0, writes 1, avg read time 18446744073709551615, avg write time 5251566408153596, temp 109 root@debian-i386:~# cat /sys/kernel/debug/hot_track/sdb/range_data inode #279, range start 0 (range len 1048576) reads 0, writes 1, avg read time 18446744073709551615, avg write time 1128690176623144209, temp 64 root@debian-i386:~# echo "hot data tracking test" >> /mnt/file root@debian-i386:~# cat /sys/kernel/debug/hot_track/sdb/inode_data inode #279, reads 0, writes 2, avg read time 18446744073709551615, avg write time 4923343766042451, temp 109 root@debian-i386:~# cat /sys/kernel/debug/hot_track/sdb/range_data inode #279, range start 0 (range len 1048576) reads 0, writes 2,
Quoted 8 lines
Quoted 3 lines
avg read time 18446744073709551615, avg write time 1058147040842596150, temp 64 root@debian-i386:~#
It's a good idea to add a new file under documentation and include this kind of information. For example what temp means, how it's worked out and how to "read" the avg read/write time (nanoseconds, microseconds, jiffies....??)
Good suggestion, i will do, thanks. Do you have comments for other patches?
Quoted 2 lines
Marco
Signature · 3 lines
Regards, Zhi Yong Wu
Quoted 10 lines · On Mon, Sep 17, 2012 at 03:18:34PM +0800, [email protected] wrote:
From: Zhi Yong Wu <[email protected]>
NOTE:
The patchset is currently post out mainly to make sure
it is going in the correct direction and hope to get some helpful comments from other guys.TODO List:
Quoted 8 lines
1.) Need to do scalability or performance tests.
2.) Turn some Micro into tunables
TIME_TO_KICK, andHEAT_UPDATE_DELAY
3.) Rafactorhot_hash_is_aging()
If you just made the timeout value a timespec and compared
the _timespecs_, you would be doing a lot fewer conversions.
4.) Cleanup some unnecessary lock protect
5.) Add more comments to explain how to calc temperature
0) Documentation.
Quoted 4 lines
Zhi Yong Wu (11):
vfs: introduce one structurehot_info
vfs: introduce one rb tree -hot_inode_tree
vfs: introduce 2 rb tree items - inode and range
These three patches can probably just be flattened into one. Splitting out the first two doesn't add to the clarity of the series - add the structures in the patch that actually uses them so we don't ahve to jump between patches to see how they are used.
Quoted 31 lines
vfs: add support for updating access frequency
vfs: add one new mount option '-o hottrack'
vfs: add init and exit support
vfs: introduce one hash table
vfs: enable hot data tracking
vfs: fork one private kthread to update temperature info
vfs: add 3 new ioctl interfaces
vfs: add debugfs supportfs/Makefile | 3 +- fs/compat_ioctl.c | 8 + fs/dcache.c | 2 + fs/direct-io.c | 10 + fs/hot_debugfs.c | 487 ++++++++++++++++++++++++++++++++++ fs/hot_debugfs.h | 60 +++++ fs/hot_hash.c | 369 ++++++++++++++++++++++++++ fs/hot_hash.h | 108 ++++++++ fs/hot_rb.c | 648 +++++++++++++++++++++++++++++++++++++++++++++ fs/hot_rb.h | 70 +++++ fs/hot_track.c | 113 ++++++++ fs/hot_track.h | 23 ++ fs/ioctl.c | 132 +++++++++ fs/namespace.c | 10 + fs/super.c | 11 + include/linux/fs.h | 15 + include/linux/hot_track.h | 169 ++++++++++++ mm/filemap.c | 8 + mm/page-writeback.c | 21 ++ mm/readahead.c | 9 + 20 files changed, 2275 insertions(+), 1 deletions(-) create mode 100644 fs/hot_debugfs.c
Quoted 8 lines
create mode 100644 fs/hot_debugfs.h create mode 100644 fs/hot_hash.c create mode 100644 fs/hot_hash.h create mode 100644 fs/hot_rb.c create mode 100644 fs/hot_rb.h create mode 100644 fs/hot_track.c create mode 100644 fs/hot_track.h create mode 100644 include/linux/hot_track.h
So, 9 new files for tracking all of this? I'm not sure that so many new files are needed - putting it all in fs/hot_tracking.[ch] might make more sense, or if all 8 fs/hot* files remain, putting them in their own subdirectory might be an idea (like quota).
I'll comment on the code when I get a bit of time to look at it.
Cheers,
Dave.
Signature · 2 lines
Dave Chinner [email protected]
Quoted 27 lines · Dave Chinner
Quoted 10 lines · On Mon, Sep 17, 2012 at 03:18:34PM +0800, [email protected] wrote:
From: Zhi Yong Wu <[email protected]>
NOTE:
The patchset is currently post out mainly to make sure
it is going in the correct direction and hope to get some helpful comments from other guys.TODO List:
Quoted 8 lines
1.) Need to do scalability or performance tests.
2.) Turn some Micro into tunables
TIME_TO_KICK, andHEAT_UPDATE_DELAY
3.) Rafactorhot_hash_is_aging()
If you just made the timeout value a timespec and compared
the _timespecs_, you would be doing a lot fewer conversions.
4.) Cleanup some unnecessary lock protect
5.) Add more comments to explain how to calc temperature0) Documentation.
Quoted 4 lines
Zhi Yong Wu (11):
vfs: introduce one structurehot_info
vfs: introduce one rb tree -hot_inode_tree
vfs: introduce 2 rb tree items - inode and range
Quoted 4 lines
These three patches can probably just be flattened into one. Splitting out the first two doesn't add to the clarity of the series - add the structures in the patch that actually uses them so we don't ahve to jump between patches to see how they are used.
It make sense to me, OK, thanks.
Quoted 41 lines
Quoted 32 lines
vfs: add support for updating access frequency
vfs: add one new mount option '-o hottrack'
vfs: add init and exit support
vfs: introduce one hash table
vfs: enable hot data tracking
vfs: fork one private kthread to update temperature info
vfs: add 3 new ioctl interfaces
vfs: add debugfs supportfs/Makefile | 3 +- fs/compat_ioctl.c | 8 + fs/dcache.c | 2 + fs/direct-io.c | 10 + fs/hot_debugfs.c | 487 ++++++++++++++++++++++++++++++++++ fs/hot_debugfs.h | 60 +++++ fs/hot_hash.c | 369 ++++++++++++++++++++++++++ fs/hot_hash.h | 108 ++++++++ fs/hot_rb.c | 648 +++++++++++++++++++++++++++++++++++++++++++++ fs/hot_rb.h | 70 +++++ fs/hot_track.c | 113 ++++++++ fs/hot_track.h | 23 ++ fs/ioctl.c | 132 +++++++++ fs/namespace.c | 10 + fs/super.c | 11 + include/linux/fs.h | 15 + include/linux/hot_track.h | 169 ++++++++++++ mm/filemap.c | 8 + mm/page-writeback.c | 21 ++ mm/readahead.c | 9 + 20 files changed, 2275 insertions(+), 1 deletions(-) create mode 100644 fs/hot_debugfs.cQuoted 8 lines
create mode 100644 fs/hot_debugfs.h create mode 100644 fs/hot_hash.c create mode 100644 fs/hot_hash.h create mode 100644 fs/hot_rb.c create mode 100644 fs/hot_rb.h create mode 100644 fs/hot_track.c create mode 100644 fs/hot_track.h create mode 100644 include/linux/hot_track.h
Quoted 4 lines
So, 9 new files for tracking all of this? I'm not sure that so many new files are needed - putting it all in fs/hot_tracking.[ch] might make more sense, or if all 8
fs/hot*files remain, putting them in their own subdirectory might be an idea (like quota).
If all functions are in two files, they will be large and the logic is not so clear. so i prefer the latter. thanks.
Quoted 2 lines
I'll comment on the code when I get a bit of time to look at it.
Great, very look forward to your more comments.
Quoted 7 lines
Cheers,
Dave.
--
Dave Chinner
[email protected]
Signature · 3 lines
Regards, Zhi Yong Wu
Quoted 13 lines · Zhi Yong Wu
Quoted 12 lines · Dave Chinner
Quoted 10 lines · On Mon, Sep 17, 2012 at 03:18:34PM +0800, [email protected] wrote:
20 files changed, 2275 insertions(+), 1 deletions(-) create mode 100644 fs/hot_debugfs.c create mode 100644 fs/hot_debugfs.h create mode 100644 fs/hot_hash.c create mode 100644 fs/hot_hash.h create mode 100644 fs/hot_rb.c create mode 100644 fs/hot_rb.h create mode 100644 fs/hot_track.c create mode 100644 fs/hot_track.h create mode 100644 include/linux/hot_track.h
Quoted 6 lines
Quoted 4 lines
So, 9 new files for tracking all of this? I'm not sure that so many new files are needed - putting it all in fs/hot_tracking.[ch] might make more sense, or if all 8
fs/hot*files remain, putting them in their own subdirectory might be an idea (like quota).If all functions are in two files, they will be large and the logic is not so clear. so i prefer the latter. thanks.
A single file is much easier to handle when searching and editting, though. And realistically:
$ wc -l fs/*.c |sort -nr -k 1 |head -10 62483 total 4000 fs/namei.c 3281 fs/buffer.c 3169 fs/dcache.c 2677 fs/namespace.c 2364 fs/locks.c 2321 fs/exec.c 2120 fs/binfmt_elf.c 2053 fs/splice.c 1934 fs/eventpoll.c $
There are plenty of files for a single function/subsystem of around the aggregate size of this code, so everyone is used to dealing with this amount of logic for a particular feature in a single file.
I much prefer it that way, to tell the truth, because I find it fairly common to have 5-10 files open at once when tracking something through, say, the buffered IO path. If I've now got to have another 5-10 files open just to track that same code path through this index, then that makes it much more difficult to manage than if were all in one .c file.
IOWs, splitting code up into multiple files is not a win when the resultant files are all small and interdependent and you need to look at multiple files at the same time to understand what the code does as a whole....
Cheers,
Dave.
Signature · 2 lines
Dave Chinner [email protected]
Quoted 48 lines · Dave Chinner
Quoted 13 lines · Zhi Yong Wu
Quoted 12 lines · Dave Chinner
Quoted 10 lines · On Mon, Sep 17, 2012 at 03:18:34PM +0800, [email protected] wrote:
20 files changed, 2275 insertions(+), 1 deletions(-) create mode 100644 fs/hot_debugfs.c create mode 100644 fs/hot_debugfs.h create mode 100644 fs/hot_hash.c create mode 100644 fs/hot_hash.h create mode 100644 fs/hot_rb.c create mode 100644 fs/hot_rb.h create mode 100644 fs/hot_track.c create mode 100644 fs/hot_track.h create mode 100644 include/linux/hot_track.hQuoted 6 lines
Quoted 4 lines
So, 9 new files for tracking all of this? I'm not sure that so many new files are needed - putting it all in fs/hot_tracking.[ch] might make more sense, or if all 8
fs/hot*files remain, putting them in their own subdirectory might be an idea (like quota).If all functions are in two files, they will be large and the logic is not so clear. so i prefer the latter. thanks.
A single file is much easier to handle when searching and editting, though. And realistically:
$ wc -l fs/*.c |sort -nr -k 1 |head -10 62483 total 4000 fs/namei.c 3281 fs/buffer.c 3169 fs/dcache.c 2677 fs/namespace.c 2364 fs/locks.c 2321 fs/exec.c 2120 fs/binfmt_elf.c 2053 fs/splice.c 1934 fs/eventpoll.c $There are plenty of files for a single function/subsystem of around the aggregate size of this code, so everyone is used to dealing with this amount of logic for a particular feature in a single file.
I much prefer it that way, to tell the truth, because I find it fairly common to have 5-10 files open at once when tracking something through, say, the buffered IO path. If I've now got to have another 5-10 files open just to track that same code path through this index, then that makes it much more difficult to manage than if were all in one .c file.
Quoted 4 lines
IOWs, splitting code up into multiple files is not a win when the resultant files are all small and interdependent and you need to look at multiple files at the same time to understand what the code does as a whole....
Just completed move them to one separate directory fs/hot_tracking/. But what you said makes sense to me, OK, i will adopt your suggestion, thanks.
Quoted 7 lines
Cheers,
Dave.
--
Dave Chinner
[email protected]
Signature · 3 lines
Regards, Zhi Yong Wu