WSL2-Linux-Kernel/fs/erofs/utils.c

294 строки
6.8 KiB
C
Исходник Обычный вид История

// SPDX-License-Identifier: GPL-2.0-only
/*
* Copyright (C) 2018 HUAWEI, Inc.
* https://www.huawei.com/
* Created by Gao Xiang <gaoxiang25@huawei.com>
*/
#include "internal.h"
staging: erofs: introduce VLE decompression support This patch introduces the basic in-place VLE decompression implementation for the erofs file system. Compared with fixed-sized input compression, it implements what we call 'the variable-length extent compression' which specifies the same output size for each compression block to make the full use of IO bandwidth (which means almost all data from block device can be directly used for decomp- ression), improve the real (rather than just via data caching, which costs more memory) random read and keep the relatively lower compression ratios (it saves more storage space than fixed-sized input compression which is also configured with the same input block size), as illustrated below: |--- variable-length extent ---|------ VLE ------|--- VLE ---| /> clusterofs /> clusterofs /> clusterofs /> clusterofs ++---|-------++-----------++---------|-++-----------++-|---------++-| ...|| | || || | || || | || | ... original data ++---|-------++-----------++---------|-++-----------++-|---------++-| ++->cluster<-++->cluster<-++->cluster<-++->cluster<-++->cluster<-++ size size size size size \ / / / \ / / / \ / / / ++-----------++-----------++-----------++ ... || || || || ... compressed clusters ++-----------++-----------++-----------++ ++->cluster<-++->cluster<-++->cluster<-++ size size size The main point of 'in-place' refers to the decompression mode: Instead of allocating independent compressed pages and data structures, it reuses the allocated file cache pages at most to store its compressed data and the corresponding pagevec in a time-sharing approach by default, which will be useful for low memory scenario. In the end, unlike the other filesystems with (de)compression support using a relatively large compression block size, which reads and decompresses >= 128KB at once, and gains a more good-looking random read (In fact it collects small random reads into large sequential reads and caches all decompressed data in memory, but it is unacceptable especially for embedded devices with limited memory, and it is not the real random read), we select a universal small-sized 4KB compressed cluster, which is the smallest page size for most architectures, and all compressed clusters can be read and decompressed independently, which ensures random read number for all use cases. Signed-off-by: Gao Xiang <gaoxiang25@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2018-07-26 15:22:06 +03:00
#include <linux/pagevec.h>
struct page *erofs_allocpage(struct list_head *pool, gfp_t gfp)
{
struct page *page;
if (!list_empty(pool)) {
page = lru_to_page(pool);
DBG_BUGON(page_ref_count(page) != 1);
list_del(&page->lru);
} else {
page = alloc_page(gfp);
}
return page;
}
#if (EROFS_PCPUBUF_NR_PAGES > 0)
static struct {
u8 data[PAGE_SIZE * EROFS_PCPUBUF_NR_PAGES];
} ____cacheline_aligned_in_smp erofs_pcpubuf[NR_CPUS];
void *erofs_get_pcpubuf(unsigned int pagenr)
{
preempt_disable();
return &erofs_pcpubuf[smp_processor_id()].data[pagenr * PAGE_SIZE];
}
#endif
#ifdef CONFIG_EROFS_FS_ZIP
/* global shrink count (for all mounted EROFS instances) */
static atomic_long_t erofs_global_shrink_cnt;
static int erofs_workgroup_get(struct erofs_workgroup *grp)
{
int o;
repeat:
o = erofs_wait_on_workgroup_freezed(grp);
if (o <= 0)
return -1;
if (atomic_cmpxchg(&grp->refcount, o, o + 1) != o)
goto repeat;
/* decrease refcount paired by erofs_workgroup_put */
if (o == 1)
atomic_long_dec(&erofs_global_shrink_cnt);
return 0;
}
struct erofs_workgroup *erofs_find_workgroup(struct super_block *sb,
pgoff_t index)
{
struct erofs_sb_info *sbi = EROFS_SB(sb);
struct erofs_workgroup *grp;
repeat:
rcu_read_lock();
grp = xa_load(&sbi->managed_pslots, index);
if (grp) {
if (erofs_workgroup_get(grp)) {
/* prefer to relax rcu read side */
rcu_read_unlock();
goto repeat;
}
DBG_BUGON(index != grp->index);
}
rcu_read_unlock();
return grp;
}
struct erofs_workgroup *erofs_insert_workgroup(struct super_block *sb,
struct erofs_workgroup *grp)
{
struct erofs_sb_info *const sbi = EROFS_SB(sb);
struct erofs_workgroup *pre;
/*
* Bump up a reference count before making this visible
* to others for the XArray in order to avoid potential
* UAF without serialized by xa_lock.
*/
atomic_inc(&grp->refcount);
repeat:
xa_lock(&sbi->managed_pslots);
pre = __xa_cmpxchg(&sbi->managed_pslots, grp->index,
NULL, grp, GFP_NOFS);
if (pre) {
if (xa_is_err(pre)) {
pre = ERR_PTR(xa_err(pre));
} else if (erofs_workgroup_get(pre)) {
/* try to legitimize the current in-tree one */
xa_unlock(&sbi->managed_pslots);
cond_resched();
goto repeat;
}
atomic_dec(&grp->refcount);
grp = pre;
}
xa_unlock(&sbi->managed_pslots);
return grp;
}
static void __erofs_workgroup_free(struct erofs_workgroup *grp)
{
atomic_long_dec(&erofs_global_shrink_cnt);
erofs_workgroup_free_rcu(grp);
}
staging: erofs: introduce VLE decompression support This patch introduces the basic in-place VLE decompression implementation for the erofs file system. Compared with fixed-sized input compression, it implements what we call 'the variable-length extent compression' which specifies the same output size for each compression block to make the full use of IO bandwidth (which means almost all data from block device can be directly used for decomp- ression), improve the real (rather than just via data caching, which costs more memory) random read and keep the relatively lower compression ratios (it saves more storage space than fixed-sized input compression which is also configured with the same input block size), as illustrated below: |--- variable-length extent ---|------ VLE ------|--- VLE ---| /> clusterofs /> clusterofs /> clusterofs /> clusterofs ++---|-------++-----------++---------|-++-----------++-|---------++-| ...|| | || || | || || | || | ... original data ++---|-------++-----------++---------|-++-----------++-|---------++-| ++->cluster<-++->cluster<-++->cluster<-++->cluster<-++->cluster<-++ size size size size size \ / / / \ / / / \ / / / ++-----------++-----------++-----------++ ... || || || || ... compressed clusters ++-----------++-----------++-----------++ ++->cluster<-++->cluster<-++->cluster<-++ size size size The main point of 'in-place' refers to the decompression mode: Instead of allocating independent compressed pages and data structures, it reuses the allocated file cache pages at most to store its compressed data and the corresponding pagevec in a time-sharing approach by default, which will be useful for low memory scenario. In the end, unlike the other filesystems with (de)compression support using a relatively large compression block size, which reads and decompresses >= 128KB at once, and gains a more good-looking random read (In fact it collects small random reads into large sequential reads and caches all decompressed data in memory, but it is unacceptable especially for embedded devices with limited memory, and it is not the real random read), we select a universal small-sized 4KB compressed cluster, which is the smallest page size for most architectures, and all compressed clusters can be read and decompressed independently, which ensures random read number for all use cases. Signed-off-by: Gao Xiang <gaoxiang25@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2018-07-26 15:22:06 +03:00
int erofs_workgroup_put(struct erofs_workgroup *grp)
{
int count = atomic_dec_return(&grp->refcount);
if (count == 1)
atomic_long_inc(&erofs_global_shrink_cnt);
else if (!count)
__erofs_workgroup_free(grp);
staging: erofs: introduce VLE decompression support This patch introduces the basic in-place VLE decompression implementation for the erofs file system. Compared with fixed-sized input compression, it implements what we call 'the variable-length extent compression' which specifies the same output size for each compression block to make the full use of IO bandwidth (which means almost all data from block device can be directly used for decomp- ression), improve the real (rather than just via data caching, which costs more memory) random read and keep the relatively lower compression ratios (it saves more storage space than fixed-sized input compression which is also configured with the same input block size), as illustrated below: |--- variable-length extent ---|------ VLE ------|--- VLE ---| /> clusterofs /> clusterofs /> clusterofs /> clusterofs ++---|-------++-----------++---------|-++-----------++-|---------++-| ...|| | || || | || || | || | ... original data ++---|-------++-----------++---------|-++-----------++-|---------++-| ++->cluster<-++->cluster<-++->cluster<-++->cluster<-++->cluster<-++ size size size size size \ / / / \ / / / \ / / / ++-----------++-----------++-----------++ ... || || || || ... compressed clusters ++-----------++-----------++-----------++ ++->cluster<-++->cluster<-++->cluster<-++ size size size The main point of 'in-place' refers to the decompression mode: Instead of allocating independent compressed pages and data structures, it reuses the allocated file cache pages at most to store its compressed data and the corresponding pagevec in a time-sharing approach by default, which will be useful for low memory scenario. In the end, unlike the other filesystems with (de)compression support using a relatively large compression block size, which reads and decompresses >= 128KB at once, and gains a more good-looking random read (In fact it collects small random reads into large sequential reads and caches all decompressed data in memory, but it is unacceptable especially for embedded devices with limited memory, and it is not the real random read), we select a universal small-sized 4KB compressed cluster, which is the smallest page size for most architectures, and all compressed clusters can be read and decompressed independently, which ensures random read number for all use cases. Signed-off-by: Gao Xiang <gaoxiang25@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2018-07-26 15:22:06 +03:00
return count;
}
static bool erofs_try_to_release_workgroup(struct erofs_sb_info *sbi,
struct erofs_workgroup *grp)
{
/*
* If managed cache is on, refcount of workgroups
* themselves could be < 0 (freezed). In other words,
* there is no guarantee that all refcounts > 0.
*/
if (!erofs_workgroup_try_to_freeze(grp, 1))
return false;
/*
* Note that all cached pages should be unattached
* before deleted from the XArray. Otherwise some
* cached pages could be still attached to the orphan
* old workgroup when the new one is available in the tree.
*/
if (erofs_try_to_free_all_cached_pages(sbi, grp)) {
erofs_workgroup_unfreeze(grp, 1);
return false;
}
/*
* It's impossible to fail after the workgroup is freezed,
* however in order to avoid some race conditions, add a
* DBG_BUGON to observe this in advance.
*/
DBG_BUGON(xa_erase(&sbi->managed_pslots, grp->index) != grp);
/* last refcount should be connected with its managed pslot. */
erofs_workgroup_unfreeze(grp, 0);
__erofs_workgroup_free(grp);
return true;
}
static unsigned long erofs_shrink_workstation(struct erofs_sb_info *sbi,
unsigned long nr_shrink)
{
struct erofs_workgroup *grp;
unsigned int freed = 0;
unsigned long index;
staging: erofs: introduce VLE decompression support This patch introduces the basic in-place VLE decompression implementation for the erofs file system. Compared with fixed-sized input compression, it implements what we call 'the variable-length extent compression' which specifies the same output size for each compression block to make the full use of IO bandwidth (which means almost all data from block device can be directly used for decomp- ression), improve the real (rather than just via data caching, which costs more memory) random read and keep the relatively lower compression ratios (it saves more storage space than fixed-sized input compression which is also configured with the same input block size), as illustrated below: |--- variable-length extent ---|------ VLE ------|--- VLE ---| /> clusterofs /> clusterofs /> clusterofs /> clusterofs ++---|-------++-----------++---------|-++-----------++-|---------++-| ...|| | || || | || || | || | ... original data ++---|-------++-----------++---------|-++-----------++-|---------++-| ++->cluster<-++->cluster<-++->cluster<-++->cluster<-++->cluster<-++ size size size size size \ / / / \ / / / \ / / / ++-----------++-----------++-----------++ ... || || || || ... compressed clusters ++-----------++-----------++-----------++ ++->cluster<-++->cluster<-++->cluster<-++ size size size The main point of 'in-place' refers to the decompression mode: Instead of allocating independent compressed pages and data structures, it reuses the allocated file cache pages at most to store its compressed data and the corresponding pagevec in a time-sharing approach by default, which will be useful for low memory scenario. In the end, unlike the other filesystems with (de)compression support using a relatively large compression block size, which reads and decompresses >= 128KB at once, and gains a more good-looking random read (In fact it collects small random reads into large sequential reads and caches all decompressed data in memory, but it is unacceptable especially for embedded devices with limited memory, and it is not the real random read), we select a universal small-sized 4KB compressed cluster, which is the smallest page size for most architectures, and all compressed clusters can be read and decompressed independently, which ensures random read number for all use cases. Signed-off-by: Gao Xiang <gaoxiang25@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2018-07-26 15:22:06 +03:00
xa_for_each(&sbi->managed_pslots, index, grp) {
/* try to shrink each valid workgroup */
if (!erofs_try_to_release_workgroup(sbi, grp))
staging: erofs: introduce VLE decompression support This patch introduces the basic in-place VLE decompression implementation for the erofs file system. Compared with fixed-sized input compression, it implements what we call 'the variable-length extent compression' which specifies the same output size for each compression block to make the full use of IO bandwidth (which means almost all data from block device can be directly used for decomp- ression), improve the real (rather than just via data caching, which costs more memory) random read and keep the relatively lower compression ratios (it saves more storage space than fixed-sized input compression which is also configured with the same input block size), as illustrated below: |--- variable-length extent ---|------ VLE ------|--- VLE ---| /> clusterofs /> clusterofs /> clusterofs /> clusterofs ++---|-------++-----------++---------|-++-----------++-|---------++-| ...|| | || || | || || | || | ... original data ++---|-------++-----------++---------|-++-----------++-|---------++-| ++->cluster<-++->cluster<-++->cluster<-++->cluster<-++->cluster<-++ size size size size size \ / / / \ / / / \ / / / ++-----------++-----------++-----------++ ... || || || || ... compressed clusters ++-----------++-----------++-----------++ ++->cluster<-++->cluster<-++->cluster<-++ size size size The main point of 'in-place' refers to the decompression mode: Instead of allocating independent compressed pages and data structures, it reuses the allocated file cache pages at most to store its compressed data and the corresponding pagevec in a time-sharing approach by default, which will be useful for low memory scenario. In the end, unlike the other filesystems with (de)compression support using a relatively large compression block size, which reads and decompresses >= 128KB at once, and gains a more good-looking random read (In fact it collects small random reads into large sequential reads and caches all decompressed data in memory, but it is unacceptable especially for embedded devices with limited memory, and it is not the real random read), we select a universal small-sized 4KB compressed cluster, which is the smallest page size for most architectures, and all compressed clusters can be read and decompressed independently, which ensures random read number for all use cases. Signed-off-by: Gao Xiang <gaoxiang25@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2018-07-26 15:22:06 +03:00
continue;
++freed;
if (!--nr_shrink)
staging: erofs: introduce VLE decompression support This patch introduces the basic in-place VLE decompression implementation for the erofs file system. Compared with fixed-sized input compression, it implements what we call 'the variable-length extent compression' which specifies the same output size for each compression block to make the full use of IO bandwidth (which means almost all data from block device can be directly used for decomp- ression), improve the real (rather than just via data caching, which costs more memory) random read and keep the relatively lower compression ratios (it saves more storage space than fixed-sized input compression which is also configured with the same input block size), as illustrated below: |--- variable-length extent ---|------ VLE ------|--- VLE ---| /> clusterofs /> clusterofs /> clusterofs /> clusterofs ++---|-------++-----------++---------|-++-----------++-|---------++-| ...|| | || || | || || | || | ... original data ++---|-------++-----------++---------|-++-----------++-|---------++-| ++->cluster<-++->cluster<-++->cluster<-++->cluster<-++->cluster<-++ size size size size size \ / / / \ / / / \ / / / ++-----------++-----------++-----------++ ... || || || || ... compressed clusters ++-----------++-----------++-----------++ ++->cluster<-++->cluster<-++->cluster<-++ size size size The main point of 'in-place' refers to the decompression mode: Instead of allocating independent compressed pages and data structures, it reuses the allocated file cache pages at most to store its compressed data and the corresponding pagevec in a time-sharing approach by default, which will be useful for low memory scenario. In the end, unlike the other filesystems with (de)compression support using a relatively large compression block size, which reads and decompresses >= 128KB at once, and gains a more good-looking random read (In fact it collects small random reads into large sequential reads and caches all decompressed data in memory, but it is unacceptable especially for embedded devices with limited memory, and it is not the real random read), we select a universal small-sized 4KB compressed cluster, which is the smallest page size for most architectures, and all compressed clusters can be read and decompressed independently, which ensures random read number for all use cases. Signed-off-by: Gao Xiang <gaoxiang25@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2018-07-26 15:22:06 +03:00
break;
}
return freed;
}
/* protected by 'erofs_sb_list_lock' */
static unsigned int shrinker_run_no;
/* protects the mounted 'erofs_sb_list' */
static DEFINE_SPINLOCK(erofs_sb_list_lock);
static LIST_HEAD(erofs_sb_list);
void erofs_shrinker_register(struct super_block *sb)
{
struct erofs_sb_info *sbi = EROFS_SB(sb);
mutex_init(&sbi->umount_mutex);
spin_lock(&erofs_sb_list_lock);
list_add(&sbi->list, &erofs_sb_list);
spin_unlock(&erofs_sb_list_lock);
}
void erofs_shrinker_unregister(struct super_block *sb)
{
struct erofs_sb_info *const sbi = EROFS_SB(sb);
mutex_lock(&sbi->umount_mutex);
/* clean up all remaining workgroups in memory */
erofs_shrink_workstation(sbi, ~0UL);
spin_lock(&erofs_sb_list_lock);
list_del(&sbi->list);
spin_unlock(&erofs_sb_list_lock);
mutex_unlock(&sbi->umount_mutex);
}
static unsigned long erofs_shrink_count(struct shrinker *shrink,
struct shrink_control *sc)
{
return atomic_long_read(&erofs_global_shrink_cnt);
}
static unsigned long erofs_shrink_scan(struct shrinker *shrink,
struct shrink_control *sc)
{
struct erofs_sb_info *sbi;
struct list_head *p;
unsigned long nr = sc->nr_to_scan;
unsigned int run_no;
unsigned long freed = 0;
spin_lock(&erofs_sb_list_lock);
do {
run_no = ++shrinker_run_no;
} while (run_no == 0);
/* Iterate over all mounted superblocks and try to shrink them */
p = erofs_sb_list.next;
while (p != &erofs_sb_list) {
sbi = list_entry(p, struct erofs_sb_info, list);
/*
* We move the ones we do to the end of the list, so we stop
* when we see one we have already done.
*/
if (sbi->shrinker_run_no == run_no)
break;
if (!mutex_trylock(&sbi->umount_mutex)) {
p = p->next;
continue;
}
spin_unlock(&erofs_sb_list_lock);
sbi->shrinker_run_no = run_no;
freed += erofs_shrink_workstation(sbi, nr - freed);
spin_lock(&erofs_sb_list_lock);
/* Get the next list element before we move this one */
p = p->next;
/*
* Move this one to the end of the list to provide some
* fairness.
*/
list_move_tail(&sbi->list, &erofs_sb_list);
mutex_unlock(&sbi->umount_mutex);
if (freed >= nr)
break;
}
spin_unlock(&erofs_sb_list_lock);
return freed;
}
static struct shrinker erofs_shrinker_info = {
.scan_objects = erofs_shrink_scan,
.count_objects = erofs_shrink_count,
.seeks = DEFAULT_SEEKS,
};
int __init erofs_init_shrinker(void)
{
return register_shrinker(&erofs_shrinker_info);
}
void erofs_exit_shrinker(void)
{
unregister_shrinker(&erofs_shrinker_info);
}
#endif /* !CONFIG_EROFS_FS_ZIP */