fscache,erofs: fscache-based on-demand read semantics
Subsystems (7)
v4 cover fscache,erofs: fscache-based on-demand read semantics
changes since v3:
- cachefiles: The current implementation relies on the anonymous fd mechanism to avoid the dependence on the format of cache file. When cache file is opened for the first time, an anon_fd associated with the cache file is sent to user daemon. User daemon could fetch and write data to cache file with the given anon_fd. The following write to the anon_fd will finally call to cachefiles kernel module, which will write data to cache file in the latest format of cache file. Thus the on-demand read mode can
keep working no matter how cache file format could change in the future. (patch 4)
- cachefiles: the on-demand read mode reuses the existing "/dev/cachefiles" devnode (patch 3)
- erofs: squash several commits implementing readahead into single commit (patch 20)
- erofs: refactor the readahead routine, so that it can read multiple pages each round (patch 20)
- patch 1 and 7 have already been cherry-picked by the maintainers, but have not been merged to the master. Keep them here for completeness.
RFC: https://lore.kernel.org/all/[email protected]/t/
v1: https://lore.kernel.org/lkml/[email protected]/T/
v2: https://lore.kernel.org/all/[email protected]/t/
v3: https://lore.kernel.org/lkml/[email protected]/T/
[Background]
============
Nydus [1] is a container image distribution service specially optimised for distribution over network. Nydus is an excellent container image acceleration solution, since it only pulls data from remote when it's really needed, a.k.a. on-demand reading.
erofs (Enhanced Read-Only File System) is a filesystem specially optimised for read-only scenarios. (Documentation/filesystem/erofs.rst)
Recently we are focusing on erofs in container images distribution scenario [2], trying to combine it with nydus. In this case, erofs can be mounted from one bootstrap file (metadata) with (optional) multiple data blob files (data) stored on another local filesystem. (All these files are actually image files in erofs disk format.)
To accelerate the container startup (fetching container image from remote and then start the container), we do hope that the bootstrap blob file could support demand read. That is, erofs can be mounted and accessed even when the bootstrap/data blob files have not been fully downloaded.
That means we have to manage the cache state of the bootstrap/data blob files (if cache hit, read directly from the local cache; if cache miss, fetch the data somehow). It would be painful and may be dumb for erofs to implement the cache management itself. Thus we prefer fscache/cachefiles to do the cache management. Besides, the demand-read feature shall be general and it can benefit other using scenarios if it can be implemented in fscache level.
[1] https://nydus.dev
[2] https://sched.co/pcdL
[Overall Design]
================
Please refer to patch 6 ("cachefiles: document on-demand read mode") for more details.
When working in original mode, cachefiles mainly serves as a local cache for remote networking fs, while in on-demand read mode, cachefiles can boost the scenario where on-demand read semantics is needed, e.g. container image distribution.
The essential difference between these two modes is that, in original mode, when cache miss, netfs itself will fetch data from remote, and then write the fetched data into cache file. While in on-demand read mode, a user daemon is responsible for fetching data and then writing to the cache file.
The on-demand read mode relies on a simple protocol used for communication between kernel and user daemon.
The current implementation relies on the anonymous fd mechanism to avoid the dependence on the format of cache file. When cache file is opened for the first time, an anon_fd associated with the cache file is sent to user daemon. With the given anon_fd, user daemon could fetch and write data into the cache file in the background, even when kernel has not triggered the cache miss. Besides, the write() syscall to the anon_fd will finally call cachefiles kernel module, which will write data to cache file in the latest format of cache file.
1. cache miss
When cache miss, cachefiles kernel module will notify user daemon the anon_fd, along with the requested file range. When notified, user dameon needs to fetch data of the requested file range, and then write the fetched data into cache file with the given anonymous fd. When finished processing the request, user daemon needs to notify the kernel.
After notifying the user daemon, the kernel read routine will hang there, until the request is handled by user daemon. When it's awaken by the notification from user daemon, i.e. the corresponding hole has been filled by the user daemon, it will retry to read from the same file range.
2. cache hit
Once data is already ready in cache file, netfs will read from cache file directly.
[Advantage of fscache-based demand-read]
========================================
1. Asynchronous Prefetch
In current mechanism, fscache is responsible for cache state management, while the data plane (fetch data from local/remote on cache miss) is done on the user daemon side.
If data has already been ready in the backing file, the upper fs (e.g. erofs) will read from the backing file directly and won't be trapped to user space anymore. Thus the user daemon could fetch data (from remote) asynchronously on the background, and thus accelerate the backing file accessing in some degree.
2. Support massive blob files
Besides this mechanism supports a large amount of backing files, and thus can benefit the densely employed scenario.
In our using scenario, one container image can correspond to one bootstrap file (required) and multiple data blob files (optional). For example, one container image for node.js will corresponds to ~20 files in total. In densely employed environment, there could be as many as hundreds of containers and thus thousands of backing files on one machine.
[Test]
==========
You could start a quick test by
https://github.com/lostjeffle/demand-read-cachefilesd
Jeffle Xu (21):
fscache: export fscache_end_operation()
cachefiles: export write routine
cachefiles: introduce on-demand read mode
cachefiles: notify user daemon with anon_fd when opening cache file
cachefiles: implement on-demand read
cachefiles: document on-demand read mode
erofs: use meta buffers for erofs_read_superblock()
erofs: export erofs_map_blocks()
erofs: add mode checking helper
erofs: register global fscache volume
erofs: add cookie context helper functions
erofs: add anonymous inode managing page cache of blob file
erofs: add erofs_fscache_read_pages() helper
erofs: register cookie context for bootstrap blob
erofs: implement fscache-based metadata read
erofs: implement fscache-based data read for non-inline layout
erofs: implement fscache-based data read for inline layout
erofs: register cookie context for data blobs
erofs: implement fscache-based data read for data blobs
erofs: implement fscache-based data readahead
erofs: add 'uuid' mount option
.../filesystems/caching/cachefiles.rst | 159 +++++ fs/cachefiles/Kconfig | 11 + fs/cachefiles/daemon.c | 576 +++++++++++++++++- fs/cachefiles/internal.h | 48 ++ fs/cachefiles/io.c | 72 ++- fs/cachefiles/namei.c | 16 +- fs/erofs/Makefile | 3 +- fs/erofs/data.c | 18 +- fs/erofs/fscache.c | 496 +++++++++++++++ fs/erofs/inode.c | 6 +- fs/erofs/internal.h | 30 + fs/erofs/super.c | 106 +++- fs/fscache/internal.h | 11 - fs/nfs/fscache.c | 8 - include/linux/fscache.h | 15 + include/linux/netfs.h | 1 + include/trace/events/cachefiles.h | 2 + include/uapi/linux/cachefiles.h | 48 ++ 18 files changed, 1526 insertions(+), 100 deletions(-) create mode 100644 fs/erofs/fscache.c create mode 100644 include/uapi/linux/cachefiles.h
Signature · 1 line
2.27.0
PATCH v4 1/21 fscache: export fscache_end_operation()
Export fscache_end_operation() to avoid code duplication.
Besides, considering the paired fscache_begin_read_operation() is already exported, it shall make sense to also export fscache_end_operation().
--- fs/fscache/internal.h | 11 ----------- fs/nfs/fscache.c | 8 -------- include/linux/fscache.h | 14 ++++++++++++++ 3 files changed, 14 insertions(+), 19 deletions(-)
fs/fscache/internal.h +0 -11
@@ -70,17 +70,6 @@ static inline void fscache_see_cookie(struct fscache_cookie *cookie, where); } -/* -* io.c -*/ -static inline void fscache_end_operation (struct netfs_cache_resources *cres) -{ -const struct netfs_cache_ops *ops =fscache_operation_valid (cres); - -if (ops) - ops->end_operation (cres); -} -/* * main.c */
fs/nfs/fscache.c +0 -8
@@ -249,14 +249,6 @@ void nfs_fscache_release_file(struct inode *inode, struct file *filp) } } -static inline void fscache_end_operation (struct netfs_cache_resources *cres) -{ -const struct netfs_cache_ops *ops =fscache_operation_valid (cres); - -if (ops) - ops->end_operation (cres); -} -/* * Fallback page reading interface. */
include/linux/fscache.h +14 -0
@@ -456,6 +456,20 @@ int fscache_begin_read_operation(struct netfs_cache_resources *cres,return -ENOBUFS ; } +/** +* fscache_end_operation - End the read operation for the netfs lib +* @cres: The cache resources for the read operation +* +* Clean up the resources at the end of the read request. +*/ +static inline void fscache_end_operation (struct netfs_cache_resources *cres) +{ +const struct netfs_cache_ops *ops =fscache_operation_valid (cres); + +if (ops) + ops->end_operation (cres); +} +/** * fscache_read - Start a read from the cache. * @cres: The cache resources to use
Signature · 1 line
2.27.0
PATCH v4 2/21 cachefiles: export write routine
In reply to Jeffle XuExtract and export the generic routine of writing data to cache files.
This is used by the following patch implementing on-demand read mode. Since it will be called inside cachefiles, make the interface generic and unrelated to netfs_cache_resources.
It is worth nothing that, ki->inval_counter is not initialized after this cleanup. It shall not make any visible difference, since inval_counter is no longer used in the write completion routine, i.e. cachefiles_write_complete().
--- fs/cachefiles/internal.h | 10 +++++++ fs/cachefiles/io.c | 61 +++++++++++++++++++++++----------------- 2 files changed, 45 insertions(+), 26 deletions(-)
fs/cachefiles/internal.h +10 -0
@@ -201,6 +201,16 @@ extern void cachefiles_put_object(struct cachefiles_object *object,*/ extern bool cachefiles_begin_operation (struct netfs_cache_resources *cres,enum fscache_want_state want_state); +extern int __cachefiles_prepare_write (struct cachefiles_object *object, +struct file *file, +loff_t *_start,size_t *_len, +bool no_space_allocated_yet); +extern int __cachefiles_write (struct cachefiles_object *object, +struct file *file, +loff_t start_pos, +struct iov_iter *iter, +netfs_io_terminated_t term_func, +void *term_func_priv);/* * key.c
fs/cachefiles/io.c +35 -26
109 lines not on this page · raw message
Signature · 1 line
2.27.0
PATCH v4 3/21 cachefiles: introduce on-demand read mode
In reply to Jeffle XuFscache/cachefiles used to serve as a local cache for remote fs. This patch, along with the following patches, introduces a new on-demand read mode for cachefiles, which can boost the scenario where on-demand read semantics is needed, e.g. container image distribution.
The essential difference between the original mode and on-demand read mode is that, in the original mode, when cache miss, netfs itself will fetch data from remote, and then write the fetched data into cache file. While in on-demand read mode, a user daemon is responsible for fetching data and then writing to the cache file.
This patch only adds the command to enable on-demand read mode. An optional parameter to "bind" command is added. On-demand mode will be turned on when this optional argument matches "ondemand" exactly, i.e. "bind ondemand". Otherwise cachefiles will keep working in the original mode.
The following patches will implement the data plane of on-demand read mode.
--- fs/cachefiles/Kconfig | 11 ++++ fs/cachefiles/daemon.c | 132 +++++++++++++++++++++++++++++++-------- fs/cachefiles/internal.h | 6 ++ 3 files changed, 124 insertions(+), 25 deletions(-)
fs/cachefiles/Kconfig +11 -0
@@ -26,3 +26,14 @@ config CACHEFILES_ERROR_INJECTION help This permits error injection to be enabled in cachefiles whilst a cache is in service. + +config CACHEFILES_ONDEMAND + bool "Support for on-demand read" + depends on CACHEFILES + default n + help + This permits on-demand read mode of cachefiles. In this mode, when + cache miss, the cachefiles backend instead of netfs, is responsible + for fetching data, e.g. through user daemon. + + If unsure, say N.
fs/cachefiles/daemon.c +107 -25
195 lines not on this page · raw message
fs/cachefiles/internal.h +6 -0
@@ -15,6 +15,7 @@#include <linux/fscache-cache.h> #include <linux/cred.h> #include <linux/security.h> +#include <linux/xarray.h> #define CACHEFILES_DIO_BLOCK_SIZE 4096
@@ -98,9 +99,14 @@ struct cachefiles_cache {#define CACHEFILES_DEAD 1/* T if cache dead */ #define CACHEFILES_CULLING 2/* T if cull engaged */ #define CACHEFILES_STATE_CHANGED 3/* T if state changed (poll trigger) */ +#define CACHEFILES_ONDEMAND_MODE 4/* T if in on-demand read mode */ char *rootdirname;/* name of cache root directory */ char *secctx;/* LSM security context */ char *tag;/* cache binding tag */ +#ifdef CONFIG_CACHEFILES_ONDEMAND +struct xarray reqs;/* xarray of pending on-demand requests */ +rwlock_t reqs_lock;/* Lock for reqs xarray */ +#endif };#include <trace/events/cachefiles.h>
Signature · 1 line
2.27.0
PATCH v4 4/21 cachefiles: notify user daemon with anon_fd when opening cache file
In reply to Jeffle XuSend the anonymous fd to user daemon when opening cache file for the first time, no matter whether the cache file exist there or not. With the given anonymous fd, user daemon can fetch and then write data into cache file in advance, even when cache miss has not happended yet.
Also add one advisory flag (FSCACHE_ADV_WANT_CACHE_SIZE) suggesting that cache file size shall be retrieved at runtime. This helps the scenario where one cache file can contain multiple netfs files for the purpose of deduplication, e.g. In this case, netfs itself may has no idea the cache file size, whilst user daemon needs to offer the hint on the cache file size.
--- fs/cachefiles/daemon.c | 348 +++++++++++++++++++++++++++++- fs/cachefiles/internal.h | 24 +++ fs/cachefiles/namei.c | 16 +- include/linux/fscache.h | 1 + include/trace/events/cachefiles.h | 2 + include/uapi/linux/cachefiles.h | 41 ++++ 6 files changed, 429 insertions(+), 3 deletions(-) create mode 100644 include/uapi/linux/cachefiles.h
fs/cachefiles/daemon.c +347 -1
@@ -19,6 +19,7 @@#include <linux/ctype.h> #include <linux/string.h> #include <linux/fs_struct.h> +#include <linux/anon_inodes.h> #include "internal.h" static int cachefiles_daemon_open (struct inode *,struct file *);
@@ -43,6 +44,9 @@ static int cachefiles_daemon_secctx(struct cachefiles_cache *, char *);static int cachefiles_daemon_tag (struct cachefiles_cache *,char *);static int cachefiles_daemon_bind (struct cachefiles_cache *,char *);static void cachefiles_daemon_unbind (struct cachefiles_cache *); +#ifdef CONFIG_CACHEFILES_ONDEMAND +static int cachefiles_ondemand_cinit (struct cachefiles_cache *,char *); +#endif static unsigned long cachefiles_open;
@@ -75,6 +79,9 @@ static const struct cachefiles_daemon_cmd cachefiles_daemon_cmds[] = { { "inuse", cachefiles_daemon_inuse }, { "secctx", cachefiles_daemon_secctx }, { "tag", cachefiles_daemon_tag }, +#ifdef CONFIG_CACHEFILES_ONDEMAND + { "cinit", cachefiles_ondemand_cinit }, +#endif { "", NULL } };
@@ -87,6 +94,21 @@ static inline void cachefiles_ondemand_open(struct cachefiles_cache *cache)static inline void cachefiles_ondemand_release (struct cachefiles_cache *cache) { +struct cachefiles_req *req; +unsigned long index; + +/* +* 1) Cache has been marked as dead state, and then 2) flush all pending +* requests in @reqs xarray. The barrier inside set_bit() will ensure +* that above two ops won't be reordered. +*/ +write_lock (&cache->reqs_lock); +xa_for_each (&cache->reqs, index, req) { + req->error = -EIO ; +complete (&req->done); + } +write_unlock (&cache->reqs_lock); +xa_destroy (&cache->reqs); }
@@ -114,6 +136,330 @@ bool cachefiles_ondemand_daemon_bind(struct cachefiles_cache *cache, char *args)return false; } +static int cachefiles_ondemand_fd_release (struct inode *inode,struct file *file) +{ +struct cachefiles_object *object = file->private_data; + +/* +* Uninstall anon_fd to the cachefiles object, so that no further +* associated requests will get enqueued. +*/ + object->fd =-1 ; + +cachefiles_put_object (object, cachefiles_obj_put_ondemand_fd); +return 0 ; +} + +static ssize_t cachefiles_ondemand_fd_write_iter (struct kiocb *kiocb, +struct iov_iter *iter) +{ +struct cachefiles_object *object = kiocb->ki_filp->private_data; +struct cachefiles_cache *cache = object->volume->cache; +struct file *file = object->file; +size_t len = iter->count; +loff_t pos = kiocb->ki_pos; +const struct cred *saved_cred; +int ret; + +if (!file) +return -ENOBUFS ; + +cachefiles_begin_secure (cache, &saved_cred); + ret =__cachefiles_prepare_write (object, file, &pos, &len, true); +cachefiles_end_secure (cache, saved_cred); +if (ret <0 ) +return ret; + + ret =__cachefiles_write (object, file, pos, iter,NULL ,NULL ); +if (!ret) + ret = len; + +return ret; +} + +static loff_t cachefiles_ondemand_fd_llseek (struct file *filp,loff_t pos,int whence) +{ +struct cachefiles_object *object = filp->private_data; +struct file *file = object->file; + +if (!file) +return -ENOBUFS ; + +return vfs_llseek (file, pos, whence); +} + +static const struct file_operations cachefiles_ondemand_fd_fops = { + .owner =THIS_MODULE , + .release = cachefiles_ondemand_fd_release, + .write_iter = cachefiles_ondemand_fd_write_iter, + .llseek = cachefiles_ondemand_fd_llseek, +}; + +/* +* Init request completion +* - command: "cinit <id>[,<cache_size>]" +*/ +static int cachefiles_ondemand_cinit (struct cachefiles_cache *cache,char *args) +{ +struct cachefiles_req *req; +struct cachefiles_init *init; +struct fscache_cookie *cookie; +char *tmp, *pid, *psize; +unsigned long id, size =0 ; +int ret; + +if (!test_bit (CACHEFILES_ONDEMAND_MODE , &cache->flags)) +return -EOPNOTSUPP ; + +if (!*args) { +pr_err ("Empty id specified\n" ); +return -EINVAL ; + } + + tmp =kstrdup (args,GFP_KERNEL ); +if (!tmp) +return -ENOMEM ; + + pid = tmp; + psize =strchr (tmp,',' ); +if (psize) { + *psize =0 ; + psize++; + + ret =kstrtoul (psize,0 , &size); +if (ret) + goto out; + } + + ret =kstrtoul (pid,0 , &id); +if (ret) + goto out; + + req =xa_erase (&cache->reqs, id); +if (!req) { + ret = -EINVAL ; + goto out; + } + + init = (void *)req->msg.data; +if (init->flags &CACHEFILES_INIT_FL_WANT_CACHE_SIZE ) { +if (WARN_ON_ONCE (!size)) { + req->error = -EINVAL ; + }else { + cookie = req->object->cookie; + cookie->object_size = size; +if (size) +set_bit (FSCACHE_COOKIE_NO_DATA_TO_READ , &cookie->flags); +else +clear_bit (FSCACHE_COOKIE_NO_DATA_TO_READ , &cookie->flags); + } + } + +complete (&req->done); +out: +kfree (tmp); +return ret; +} + +static int cachefiles_ondemand_get_fd (struct cachefiles_req *req) +{ +struct cachefiles_init *init; +struct cachefiles_object *object; +struct fd f; +int ret; + + object =cachefiles_grab_object (req->object, + cachefiles_obj_get_ondemand_fd); + + ret =anon_inode_getfd ("[cachefiles]" , &cachefiles_ondemand_fd_fops, + object,O_WRONLY ); +if (ret <0 ) { +cachefiles_put_object (object, cachefiles_obj_put_ondemand_fd); +return ret; + } + + f =fdget_pos (ret); +if (WARN_ON_ONCE (!f.file)) +return -EBADFD ; + + f.file->f_mode |=FMODE_PWRITE |FMODE_LSEEK ; +fdput_pos (f); + + init = (void *)req->msg.data; + init->fd = object->fd = ret; + +return 0 ; +} + +static ssize_t cachefiles_ondemand_daemon_read (struct cachefiles_cache *cache, +char __user *_buffer, +size_t buflen) +{ +struct cachefiles_req *req; +struct cachefiles_msg *msg; +unsigned long id =0 ; +size_t n; +int ret =0 ; + +/* +* Search for request that has not ever been processed, to prevent +* requests from being sent to user daemon repeatedly. +*/ + req =xa_find (&cache->reqs, &id,UINT_MAX ,CACHEFILES_REQ_NEW ); +if (!req) +return 0 ; + + msg = &req->msg; + msg->id = id; + + n = msg->len; +if (n > buflen) +return -EMSGSIZE ; + +if (msg->opcode ==CACHEFILES_OP_INIT ) { + ret =cachefiles_ondemand_get_fd (req); +if (ret) { + req =xa_erase (&cache->reqs, id); +if (WARN_ON_ONCE (!req)) +return ret; + + req->error = ret; +complete (&req->done); +return ret; + } + } + +if (copy_to_user (_buffer, msg, n) !=0 ) +return -EFAULT ; + +xa_clear_mark (&cache->reqs, id,CACHEFILES_REQ_NEW ); +return n; +} + +static inline int cachefiles_ondemand_check (struct cachefiles_cache *cache) +{ +if (!test_bit (CACHEFILES_ONDEMAND_MODE , &cache->flags)) +return -EOPNOTSUPP ; + +if (test_bit (CACHEFILES_DEAD , &cache->flags)) +return -EIO ; + +return 0 ; +} + +/* +* Stop enqueuing requests when daemon is dying. So we need to 1) check cache +* state, and 2) enqueue request only if cache is not in dead state. The above +* two ops need to be atomic as a whole. @reqs_lock is used here to ensure that. +* Otherwise, request may be enqueued after @reqs xarray has been flushed, in +* which case the orphan request will never be completed and thus netfs will +* hang there forever. +*/ +static int cachefiles_ondemand_send_req (struct cachefiles_cache *cache, +struct cachefiles_req *req) +{ +struct xarray *xa = &cache->reqs; +int ret; +u32 id; + +read_lock (&cache->reqs_lock); + +/* recheck dead state under lock */ +if (test_bit (CACHEFILES_DEAD , &cache->flags)) { +read_unlock (&cache->reqs_lock); +return -EIO ; + } + +xa_lock (xa); + ret =__xa_alloc (xa, &id, req, xa_limit_32b,GFP_KERNEL ); +if (!ret) +__xa_set_mark (xa, id,CACHEFILES_REQ_NEW ); +xa_unlock (xa); + +read_unlock (&cache->reqs_lock); + +if (!ret) { +wake_up_all (&cache->daemon_pollwq); +wait_for_completion (&req->done); + ret = req->error; + } + +return ret; +} + +static inline +struct cachefiles_req *cachefiles_alloc_req (struct cachefiles_object *object, +enum cachefiles_opcode opcode, +size_t data_len) +{ +struct cachefiles_req *req; + + req =kzalloc (sizeof (*req) + data_len,GFP_KERNEL ); +if (req) { + req->object = object; +init_completion (&req->done); + req->msg.opcode = opcode; + req->msg.len =sizeof (struct cachefiles_msg ) + data_len; + } + +return req; +} + +static struct cachefiles_req * +cachefiles_alloc_init_req (struct cachefiles_object *object) +{ +struct fscache_cookie *cookie = object->cookie; +struct fscache_volume *volume = object->volume->vcookie; +struct cachefiles_req *req; +struct cachefiles_init *init; +void *volume_key, *cookie_key; +size_t volume_key_len, cookie_key_len, data_len; + +/* volume key is of string format */ + volume_key_len = volume->key[0 ] +1 ; + volume_key = volume->key +1 ; + +/* cookie key is of binary format */ + cookie_key_len = cookie->key_len; + cookie_key =fscache_get_key (cookie); + + data_len =sizeof (*init) + volume_key_len + cookie_key_len; + req =cachefiles_alloc_req (object,CACHEFILES_OP_INIT , data_len); +if (!req) +return NULL ; + + init = (void *)req->msg.data; + init->volume_key_len = volume_key_len; + init->cookie_key_len = cookie_key_len; +memcpy (init->data, volume_key, volume_key_len); +memcpy (init->data + volume_key_len, cookie_key, cookie_key_len); + +if (object->cookie->advice &FSCACHE_ADV_WANT_CACHE_SIZE ) + init->flags |=CACHEFILES_INIT_FL_WANT_CACHE_SIZE ; + +return req; +} + +int cachefiles_ondemand_init_object (struct cachefiles_object *object) +{ +struct cachefiles_cache *cache = object->volume->cache; +struct cachefiles_req *req; +int ret; + + ret =cachefiles_ondemand_check (cache); +if (ret) +return ret; + + req =cachefiles_alloc_init_req (object); +if (!req) +return -ENOMEM ; + + ret =cachefiles_ondemand_send_req (cache, req); + +kfree (req); +return ret; +} +#else static inline void cachefiles_ondemand_open (struct cachefiles_cache *cache) {}static inline void cachefiles_ondemand_release (struct cachefiles_cache *cache) {}
@@ -129,7 +475,6 @@ bool cachefiles_ondemand_daemon_bind(struct cachefiles_cache *cache, char *args) {return false; } -#endif static inline ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache,
@@ -137,6 +482,7 @@ ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache, {return -EOPNOTSUPP ; } +#endif /* * Prepare a cache for caching.
fs/cachefiles/internal.h +24 -0
@@ -16,6 +16,7 @@#include <linux/cred.h> #include <linux/security.h> #include <linux/xarray.h> +#include <linux/cachefiles.h> #define CACHEFILES_DIO_BLOCK_SIZE 4096
@@ -59,6 +60,9 @@ struct cachefiles_object { enum cachefiles_content content_info:8; /* Info about content presence */unsigned long flags;#define CACHEFILES_OBJECT_USING_TMPFILE 0/* Have an unlinked tmpfile */ +#ifdef CONFIG_CACHEFILES_ONDEMAND +int fd;/* anonymous fd */ +#endif };/*
@@ -109,6 +113,15 @@ struct cachefiles_cache {#endif }; +struct cachefiles_req { +struct cachefiles_object *object; +struct completion done; +int error; +struct cachefiles_msg msg; +}; + +#define CACHEFILES_REQ_NEW XA_MARK_1 +#include <trace/events/cachefiles.h> static inline
@@ -152,6 +165,17 @@ extern int cachefiles_has_space(struct cachefiles_cache *cache,*/ extern const struct file_operations cachefiles_daemon_fops; +#ifdef CONFIG_CACHEFILES_ONDEMAND +extern int cachefiles_ondemand_init_object (struct cachefiles_object *object); + +#else +static inline +int cachefiles_ondemand_init_object (struct cachefiles_object *object) +{ +return 0 ; +} +#endif +/* * error_inject.c */
fs/cachefiles/namei.c +14 -2
@@ -444,10 +444,9 @@ struct file *cachefiles_create_tmpfile(struct cachefiles_object *object)struct dentry *fan = volume->fanout[(u8 )object->cookie->key_hash];struct file *file;struct path path; -uint64_t ni_size = object->cookie->object_size; +uint64_t ni_size;long ret; - ni_size =round_up (ni_size,CACHEFILES_DIO_BLOCK_SIZE );cachefiles_begin_secure (cache, &saved_cred);
@@ -473,6 +472,15 @@ struct file *cachefiles_create_tmpfile(struct cachefiles_object *object) goto out_dput; } + ret =cachefiles_ondemand_init_object (object); +if (ret <0 ) { + file =ERR_PTR (ret); + goto out_dput; + } + + ni_size = object->cookie->object_size; + ni_size =round_up (ni_size,CACHEFILES_DIO_BLOCK_SIZE ); +if (ni_size >0 ) {trace_cachefiles_trunc (object,d_backing_inode (path.dentry),0 , ni_size, cachefiles_trunc_expand_tmpfile);
@@ -573,6 +581,10 @@ static bool cachefiles_open_file(struct cachefiles_object *object, }_debug ("file -> %pd positive" , dentry); + ret =cachefiles_ondemand_init_object (object); +if (ret <0 ) + goto error_fput; + ret =cachefiles_check_auxdata (object, file);if (ret <0 ) goto check_failed;
include/linux/fscache.h +1 -0
@@ -39,6 +39,7 @@ struct fscache_cookie;#define FSCACHE_ADV_SINGLE_CHUNK 0x01/* The object is a single chunk of data */ #define FSCACHE_ADV_WRITE_CACHE 0x00/* Do cache if written to locally */ #define FSCACHE_ADV_WRITE_NOCACHE 0x02/* Don't cache if written to locally */ +#define FSCACHE_ADV_WANT_CACHE_SIZE 0x04/* Retrieve cache size at runtime */ #define FSCACHE_INVAL_DIO_WRITE 0x01/* Invalidate due to DIO write */
include/trace/events/cachefiles.h +2 -0
@@ -31,6 +31,8 @@ enum cachefiles_obj_ref_trace { cachefiles_obj_see_lookup_failed, cachefiles_obj_see_withdraw_cookie, cachefiles_obj_see_withdrawal, + cachefiles_obj_get_ondemand_fd, + cachefiles_obj_put_ondemand_fd, }; enum fscache_why_object_killed {
include/uapi/linux/cachefiles.h new file +41 -0
41 lines not on this page · raw message
Signature · 1 line
2.27.0
PATCH v4 5/21 cachefiles: implement on-demand read
In reply to Jeffle XuImplement the data plane of on-demand read mode.
A new NETFS_READ_HOLE_ONDEMAND flag is introduced to indicate that on-demand read should be done when a cache miss encountered. In this case, the read routine will send a READ request to user daemon, along with the anonymous fd and the file range that shall be read. Now user daemon is responsible for fetching data in the given file range, and then writing the fetched data into cache file with the given anonymous fd.
After sending the READ request, the read routine will hang there, until the READ request is handled by user daemon. Then it will retry to read from the same file range. If a cache miss is encountered again on the same file range, the read routine will fail then.
--- fs/cachefiles/daemon.c | 98 +++++++++++++++++++++++++++++++++ fs/cachefiles/internal.h | 8 +++ fs/cachefiles/io.c | 11 ++++ include/linux/netfs.h | 1 + include/uapi/linux/cachefiles.h | 7 +++ 5 files changed, 125 insertions(+)
fs/cachefiles/daemon.c +98 -0
@@ -46,6 +46,7 @@ static int cachefiles_daemon_bind(struct cachefiles_cache *, char *);static void cachefiles_daemon_unbind (struct cachefiles_cache *);#ifdef CONFIG_CACHEFILES_ONDEMAND static int cachefiles_ondemand_cinit (struct cachefiles_cache *,char *); +static int cachefiles_ondemand_cread (struct cachefiles_cache *,char *);#endif static unsigned long cachefiles_open;
@@ -81,6 +82,7 @@ static const struct cachefiles_daemon_cmd cachefiles_daemon_cmds[] = { { "tag", cachefiles_daemon_tag }, #ifdef CONFIG_CACHEFILES_ONDEMAND { "cinit", cachefiles_ondemand_cinit }, + { "cread", cachefiles_ondemand_cread }, #endif { "", NULL } };
@@ -139,6 +141,9 @@ bool cachefiles_ondemand_daemon_bind(struct cachefiles_cache *cache, char *args)static int cachefiles_ondemand_fd_release (struct inode *inode,struct file *file) {struct cachefiles_object *object = file->private_data; +struct cachefiles_cache *cache = object->volume->cache; +struct cachefiles_req *req; +unsigned long index;/* * Uninstall anon_fd to the cachefiles object, so that no further
@@ -146,6 +151,15 @@ static int cachefiles_ondemand_fd_release(struct inode *inode, struct file *file*/ object->fd =-1 ; +/* complete all associated pending requests */ +xa_for_each (&cache->reqs, index, req) { +if (req->object == object && + req->msg.opcode ==CACHEFILES_OP_READ ) { + req->error = -EIO ; +complete (&req->done); + } + } +cachefiles_put_object (object, cachefiles_obj_put_ondemand_fd);return 0 ; }
@@ -261,6 +275,36 @@ static int cachefiles_ondemand_cinit(struct cachefiles_cache *cache, char *args)return ret; } +/* +* Read request completion +* - command: "cread <id>" +*/ +static int cachefiles_ondemand_cread (struct cachefiles_cache *cache,char *args) +{ +struct cachefiles_req *req; +unsigned long id; +int ret; + +if (!test_bit (CACHEFILES_ONDEMAND_MODE , &cache->flags)) +return -EOPNOTSUPP ; + +if (!*args) { +pr_err ("Empty id specified\n" ); +return -EINVAL ; + } + + ret =kstrtoul (args,0 , &id); +if (ret) +return ret; + + req =xa_erase (&cache->reqs, id); +if (!req) +return -EINVAL ; + +complete (&req->done); +return 0 ; +} +static int cachefiles_ondemand_get_fd (struct cachefiles_req *req) {struct cachefiles_init *init;
@@ -460,6 +504,60 @@ int cachefiles_ondemand_init_object(struct cachefiles_object *object)return ret; } +static struct cachefiles_req * +cachefiles_alloc_read_req (struct cachefiles_object *object, +loff_t pos,size_t len) +{ +struct cachefiles_req *req; +struct cachefiles_read *read; +int fd = object->fd; + +/* Stop enqueuig request when daemon closes anon_fd prematurely. */ +if (WARN_ON_ONCE (fd ==-1 )) +return NULL ; + + req =cachefiles_alloc_req (object,CACHEFILES_OP_READ ,sizeof (*read)); +if (!req) +return NULL ; + + read = (void *)&req->msg.data; + read->off = pos; + read->len = len; + read->fd = fd; + +return req; +} + +int cachefiles_ondemand_read (struct cachefiles_object *object, +loff_t pos,size_t len) +{ +struct cachefiles_cache *cache = object->volume->cache; +struct cachefiles_req *req; +int ret; + + ret =cachefiles_ondemand_check (cache); +if (ret) +return ret; + + req =cachefiles_alloc_read_req (object, pos, len); +if (!req) +return -ENOMEM ; + +/* +* 1) Checking object->fd and 2) enqueuing request into xarray, is not +* atomic as a whole here. Thus similarly, when anon_fd is closed, it's +* possible that a new request may be enqueued into xarray, after +* associated requests in xarray have already been flushed. But it won't +* cause infinite hang since user daemon will still fetch and handle +* this request. And since the anon_fd has alrady been closed, any +* following file operation with this anon_fd will fail in this case. +*/ + ret =cachefiles_ondemand_send_req (cache, req); + +kfree (req); +return ret; +} +#else static inline void cachefiles_ondemand_open (struct cachefiles_cache *cache) {}static inline void cachefiles_ondemand_release (struct cachefiles_cache *cache) {}
fs/cachefiles/internal.h +8 -0
@@ -167,6 +167,8 @@ extern const struct file_operations cachefiles_daemon_fops;#ifdef CONFIG_CACHEFILES_ONDEMAND extern int cachefiles_ondemand_init_object (struct cachefiles_object *object); +extern int cachefiles_ondemand_read (struct cachefiles_object *object, +loff_t pos,size_t len);#else static inline
@@ -174,6 +176,12 @@ int cachefiles_ondemand_init_object(struct cachefiles_object *object) {return 0 ; } + +static inline int cachefiles_ondemand_read (struct cachefiles_object *object, +loff_t pos,size_t len) +{ +return -EOPNOTSUPP ; +}#endif /*
fs/cachefiles/io.c +11 -0
@@ -95,6 +95,7 @@ static int cachefiles_read(struct netfs_cache_resources *cres, file,file_inode (file)->i_ino, start_pos, len, i_size_read(file_inode(file))); +retry:/* If the caller asked us to seek for data before doing the read, then * we should do that now. If we find a gap, we fill it with zeros. */
@@ -119,6 +120,16 @@ static int cachefiles_read(struct netfs_cache_resources *cres,if (read_hole ==NETFS_READ_HOLE_FAIL ) goto presubmission_error; +if (read_hole ==NETFS_READ_HOLE_ONDEMAND ) { +if (!cachefiles_ondemand_read (object, off, len)) { +/* fail the read if no progress achieved */ + read_hole =NETFS_READ_HOLE_FAIL ; + goto retry; + } + + goto presubmission_error; + } +iov_iter_zero (len, iter); skipped = len; ret =0 ;
include/linux/netfs.h +1 -0
@@ -203,6 +203,7 @@ enum netfs_read_from_hole {NETFS_READ_HOLE_IGNORE ,NETFS_READ_HOLE_CLEAR ,NETFS_READ_HOLE_FAIL , + NETFS_READ_HOLE_ONDEMAND, };/*
include/uapi/linux/cachefiles.h +7 -0
@@ -8,6 +8,7 @@enum cachefiles_opcode {CACHEFILES_OP_INIT , +CACHEFILES_OP_READ , };/*
@@ -38,4 +39,10 @@ enum cachefiles_init_flags {#define CACHEFILES_INIT_FL_WANT_CACHE_SIZE (1 << CACHEFILES_INIT_WANT_CACHE_SIZE) +struct cachefiles_read { +__u64 off; +__u64 len; +__u32 fd; +}; +#endif
Signature · 1 line
2.27.0
PATCH v4 6/21 cachefiles: document on-demand read mode
In reply to Jeffle XuDocument new user interface introduced by on-demand read mode.
--- .../filesystems/caching/cachefiles.rst | 159 ++++++++++++++++++ 1 file changed, 159 insertions(+)
Documentation/filesystems/caching/cachefiles.rst +159 -0
@@ -28,6 +28,8 @@ Cache on Already Mounted Filesystem (*) Debugging. + (*) On-demand Read. + Overview
@@ -482,3 +484,160 @@ the control file. For example:: echo $((1|4|8)) >/sys/module/cachefiles/parameters/debug will turn on all function entry debugging. + + +On-demand Read +============== + +When working in original mode, cachefiles mainly serves as a local cache for +remote networking fs, while in on-demand read mode, cachefiles can boost the +scenario where on-demand read semantics is needed, e.g. container image +distribution. + +The essential difference between these two modes is that, in original mode, +when cache miss, netfs itself will fetch data from remote, and then write the +fetched data into cache file. While in on-demand read mode, a user daemon is +responsible for fetching data and then writing to the cache file. + +``CONFIG_CACHEFILES_ONDEMAND`` shall be enabled to support on-demand read mode. + + +Protocol Communication +---------------------- + +The on-demand read mode relies on a simple protocol used for communication +between kernel and user daemon. The model is like:: + + kernel --[request]--> user daemon --[reply]--> kernel + +The cachefiles kernel module will send requests to user daemon when needed. +User daemon needs to poll on the devnode ('/dev/cachefiles') to check if +there's pending request to be processed. A POLLIN event will be returned +when there's pending request. + +Then user daemon needs to read the devnode to fetch one request and process it +accordingly. It is worth nothing that each read only gets one request. When +finished processing the request, user dameon needs to write the reply to the +devnode. + +Each request is started with a message header like:: + + struct cachefiles_msg { + __u32 id; + __u32 opcode; + __u32 len; + __u8 data[]; + }; + + * ``id`` identifies the position of this request in an internal xarray + managing all pending requests. + + * ``opcode`` identifies the type of this request. + + * ``data`` identifies the payload of this request. + + * ``len`` identifies the whole length of this request, including the + header and following type specific payload. + + +Turn on On-demand Mode +---------------------- + +An optional parameter is added to "bind" command:: + + bind [ondemand] + +When "bind" command takes without argument, it defaults to the original mode. +When "bind" command takes with "ondemand" argument, i.e. "bind ondemand", +on-demand read mode will be enabled. + + +INIT Request +------------ + +When netfs opens a cache file for the first time, a request with +CACHEFILES_OP_INIT opcode, a.k.a INIT request will be sent to user daemon. The +payload format is like:: + + struct cachefiles_init { + __u32 volume_key_len; + __u32 cookie_key_len; + __u32 fd; + __u32 flags; + __u8 data[]; + }; + + * ``volume_key_len`` identifies the length of the volume key of the + cache file, in bytes. + + * ``cookie_key_len`` identifies the length of the cookie key of the + cache file, in bytes. + + * ``fd`` identifies the anonymous fd of the cache file, with which user + daemon can perform write/llseek file operations on the cache file. + + * ``data`` contains volume_key and cookie_key in sequence. + +INIT request contains (volume_key, cookie_key, anon_fd) triple for corresponding +cache file. With this triple, user daemon could fetch and write data into the +cache file in the background, even when kernel has not triggered the cache miss +yet. User daemon is able to distinguish the requested cache file with the given +(volume_key, cookie_key), and write the fetched data into cache file with the +given anon_fd. + +After recording the (volume_key, cookie_key, anon_fd) triple, user daemon shall +reply with "cinit" (complete init) command:: + + cinit <id> + + * ``id`` is exactly the id field of the previous INIT request. + + +Besides, CACHEFILES_INIT_WANT_CACHE_SIZE flag may be set in ``flags`` of INIT +request. This flag is used in the scenario where one cache file can contain +multiple netfs files for the purpose of deduplication, e.g. In this case, netfs +itself may has no idea the cache file size, whilst user daemon needs to offer +the hint on the cache file size. + +Thus when receiving an INIT request with CACHEFILES_INIT_WANT_CACHE_SIZE flag +set, user daemon must reply with the cache file size:: + + cinit <id>,<cache_size> + + * ``id`` is exactly the id field of the previous INIT request. + + * ``cache_size`` identifies the size of the cache file. + + +READ Request +------------ + +When on-demand read mode is turned on, and cache miss encountered, kernel will +send a request with CACHEFILES_OP_READ opcode, a.k.a READ request, to user +daemon. It will notify user daemon to fetch data in the requested file range. +The payload format is like:: + + struct cachefiles_read { + __u64 off; + __u64 len; + __u32 fd; + }; + + * ``off`` identifies the starting offset of the requested file range. + + * ``len`` identifies the length of the requested file range. + + * ``fd`` identifies the anonymous fd of the requested cache file. It is + guaranteed that it shall be the same with the fd field in the previous + INIT request. + +When receiving one READ request, user daemon needs to fetch data of the +requested file range, and then write the fetched data into cache file with the +given anonymous fd. + +When finished processing the READ request, user daemon needs to reply with +"cread" (complete read) command:: + + cread <id> + + * ``id`` is exactly the id field of the previous READ request.
Signature · 1 line
2.27.0
PATCH v4 7/21 erofs: use meta buffers for erofs_read_superblock()
In reply to Jeffle XuThe only change is that, meta buffers read cache page without __GFP_FS flag, which shall not matter.
--- fs/erofs/super.c | 13 +++++-------- 1 file changed, 5 insertions(+), 8 deletions(-)
fs/erofs/super.c +5 -8
@@ -281,21 +281,19 @@ static int erofs_init_devices(struct super_block *sb,static int erofs_read_superblock (struct super_block *sb) {struct erofs_sb_info *sbi; -struct page *page; +struct erofs_buf buf = __EROFS_BUF_INITIALIZER;struct erofs_super_block *dsb;unsigned int blkszbits;void *data;int ret; - page =read_mapping_page (sb->s_bdev->bd_inode->i_mapping,0 ,NULL ); -if (IS_ERR (page)) { + data =erofs_read_metabuf (&buf, sb,0 ,EROFS_KMAP ); +if (IS_ERR (data)) {erofs_err (sb,"cannot read erofs superblock" ); -return PTR_ERR (page); +return PTR_ERR (data); } sbi =EROFS_SB (sb); - - data =kmap (page); dsb = (struct erofs_super_block *)(data +EROFS_SUPER_OFFSET ); ret = -EINVAL ;
@@ -365,8 +363,7 @@ static int erofs_read_superblock(struct super_block *sb)if (erofs_sb_has_ztailpacking (sbi))erofs_info (sb,"EXPERIMENTAL compressed inline data feature in use. Use at your own risk!" ); out: -kunmap (page); -put_page (page); +erofs_put_metabuf (&buf);return ret; }
Signature · 1 line
2.27.0
Re: [PATCH v4 07/21] erofs: use meta buffers for erofs_read_superblock()
Quoted 4 lines · Jeffle Xu
The only change is that, meta buffers read cache page without
__GFP_FSflag, which shall not matter.Signed-off-by Jeffle Xu <[email protected]>
Thanks,
PATCH v4 8/21 erofs: export erofs_map_blocks()
In reply to Jeffle Xu... so that it can be used in the following introduced fs/erofs/fscache.c.
--- fs/erofs/data.c | 4 ++-- fs/erofs/internal.h | 2 ++ 2 files changed, 4 insertions(+), 2 deletions(-)
fs/erofs/data.c +2 -2
@@ -104,8 +104,8 @@ static int erofs_map_blocks_flatmode(struct inode *inode,return 0 ; } -static int erofs_map_blocks (struct inode *inode, -struct erofs_map_blocks *map,int flags) +int erofs_map_blocks (struct inode *inode, +struct erofs_map_blocks *map,int flags) {struct super_block *sb = inode->i_sb;struct erofs_inode *vi =EROFS_I (inode);
fs/erofs/internal.h +2 -0
@@ -484,6 +484,8 @@ void *erofs_read_metabuf(struct erofs_buf *buf, struct super_block *sb,int erofs_map_dev (struct super_block *sb,struct erofs_map_dev *dev);int erofs_fiemap (struct inode *inode,struct fiemap_extent_info *fieinfo,u64 start,u64 len); +int erofs_map_blocks (struct inode *inode, +struct erofs_map_blocks *map,int flags);/* inode.c */ static inline unsigned long erofs_inode_hash (erofs_nid_t nid)
Signature · 1 line
2.27.0
PATCH v4 9/21 erofs: add mode checking helper
In reply to Jeffle XuUntil then erofs is exactly blockdev based filesystem. In other using scenarios (e.g. container image), erofs needs to run upon files.
This patch set is going to introduces a new nodev mode, in which erofs could be mounted from a bootstrap blob file containing complete erofs image.
Add a helper checking which mode erofs works in.
--- fs/erofs/internal.h | 5 +++++ 1 file changed, 5 insertions(+)
fs/erofs/internal.h +5 -0
@@ -161,6 +161,11 @@ struct erofs_sb_info {#define set_opt (opt, option) ((opt)->mount_opt |= EROFS_MOUNT_##option)#define test_opt (opt, option) ((opt)->mount_opt & EROFS_MOUNT_##option) +static inline bool erofs_bdev_mode (struct super_block *sb) +{ +return sb->s_bdev; +} +enum {EROFS_ZIP_CACHE_DISABLED ,EROFS_ZIP_CACHE_READAHEAD ,
Signature · 1 line
2.27.0
PATCH v4 10/21 erofs: register global fscache volume
In reply to Jeffle XuAll erofs instances will share one global fscache volume.
In this using scenario, one erofs instance could be mounted from one (or multiple) blob files instead of blkdev. The number of blob files that each erofs instance could correspond to is limited, since these blob files are quite large in size. For example, when used for container image distribution, one erofs instance used for container image for node.js will correspond to ~20 blob files in total. Thus in densely employed environment, there could be as many as hundreds of containers and thus thousands of fscache cookies under one fscache volume.
Then as for cachefiles backend, the hash table managing all cookies under one volume contains 32K slots. Thus the hashing functionality shall scale well in this case. Besides, cachefiles backend will scatter backing files under 256 fan sub-directoris, and thus the scalability of looking up backing files shall also not be an issue.
--- fs/erofs/Makefile | 3 ++- fs/erofs/fscache.c | 21 +++++++++++++++++++++ fs/erofs/internal.h | 5 +++++ fs/erofs/super.c | 7 +++++++ 4 files changed, 35 insertions(+), 1 deletion(-) create mode 100644 fs/erofs/fscache.c
fs/erofs/Makefile +2 -1
@@ -1,7 +1,8 @@ # SPDX-License-Identifier: GPL-2.0-only obj-$(CONFIG_EROFS_FS) += erofs.o -erofs-objs := super.o inode.o data.o namei.o dir.o utils.o pcpubuf.o sysfs.o +erofs-objs := super.o inode.o data.o namei.o dir.o utils.o pcpubuf.o sysfs.o \ + fscache.o erofs-$(CONFIG_EROFS_FS_XATTR) += xattr.o erofs-$(CONFIG_EROFS_FS_ZIP) += decompressor.o zmap.o zdata.o erofs-$(CONFIG_EROFS_FS_ZIP_LZMA) += decompressor_lzma.o
fs/erofs/fscache.c new file +21 -0
@@ -0,0 +1,21 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* +* Copyright (C) 2021, Alibaba Cloud +*/ +#include "internal.h" + +static struct fscache_volume *volume; + +int __init erofs_init_fscache (void ) +{ + volume =fscache_acquire_volume ("erofs" ,NULL ,NULL ,0 ); +if (!volume) +return -EINVAL ; + +return 0 ; +} + +void erofs_exit_fscache (void ) +{ +fscache_relinquish_volume (volume,NULL , false); +}
fs/erofs/internal.h +5 -0
@@ -17,6 +17,7 @@#include <linux/slab.h> #include <linux/vmalloc.h> #include <linux/iomap.h> +#include <linux/fscache.h> #include "erofs_fs.h" /* redefine pr_fmt "erofs: " */
@@ -616,6 +617,10 @@ static inline int z_erofs_load_lzma_config(struct super_block *sb, }#endif /* !CONFIG_EROFS_FS_ZIP */ +/* fscache.c */ +int erofs_init_fscache (void ); +void erofs_exit_fscache (void ); +#define EFSCORRUPTED EUCLEAN/* Filesystem is corrupted */ #endif /* __EROFS_INTERNAL_H */
fs/erofs/super.c +7 -0
@@ -814,6 +814,10 @@ static int __init erofs_module_init(void)if (err) goto sysfs_err; + err =erofs_init_fscache (); +if (err) + goto fscache_err; + err =register_filesystem (&erofs_fs_type);if (err) goto fs_err;
@@ -821,6 +825,8 @@ static int __init erofs_module_init(void)return 0 ; fs_err: +erofs_exit_fscache (); +fscache_err:erofs_exit_sysfs (); sysfs_err:z_erofs_exit_zip_subsystem ();
@@ -841,6 +847,7 @@ static void __exit erofs_module_exit(void)/* Ensure all RCU free inodes / pclusters are safe to be destroyed. */ rcu_barrier (); +erofs_exit_fscache ();erofs_exit_sysfs ();z_erofs_exit_zip_subsystem ();z_erofs_lzma_exit ();
Signature · 1 line
2.27.0
PATCH v4 11/21 erofs: add cookie context helper functions
In reply to Jeffle XuIntroduce 'struct erofs_cookie_ctx' for managing cookie for backing file, and the following introduced API for reading from backing file.
Besides, introduce two helper functions for initializing and cleaning up erofs_cookie_ctx.
struct erofs_cookie_ctx * erofs_fscache_get_ctx(struct super_block *sb, char *path);
void erofs_fscache_put_ctx(struct erofs_cookie_ctx *ctx);
--- fs/erofs/fscache.c | 78 +++++++++++++++++++++++++++++++++++++++++++++ fs/erofs/internal.h | 8 +++++ 2 files changed, 86 insertions(+)
fs/erofs/fscache.c +78 -0
@@ -6,6 +6,84 @@static struct fscache_volume *volume; +static int erofs_fscache_init_cookie (struct erofs_fscache_context *ctx, +char *path) +{ +struct fscache_cookie *cookie; + +/* +* @object_size shall be non-zero to avoid +* FSCACHE_COOKIE_NO_DATA_TO_READ. +*/ + cookie =fscache_acquire_cookie (volume,FSCACHE_ADV_WANT_CACHE_SIZE , + path,strlen (path), +NULL ,0 ,0 ); +if (!cookie) +return -EINVAL ; + +fscache_use_cookie (cookie, false); + ctx->cookie = cookie; +return 0 ; +} + +static inline +void erofs_fscache_cleanup_cookie (struct erofs_fscache_context *ctx) +{ +struct fscache_cookie *cookie = ctx->cookie; + +fscache_unuse_cookie (cookie,NULL ,NULL ); +fscache_relinquish_cookie (cookie, false); + ctx->cookie =NULL ; +} + +static int erofs_fscache_init_ctx (struct erofs_fscache_context *ctx, +struct super_block *sb,char *path) +{ +int ret; + + ret =erofs_fscache_init_cookie (ctx, path); +if (ret) { +erofs_err (sb,"failed to init cookie" ); +return ret; + } + +return 0 ; +} + +static inline +void erofs_fscache_cleanup_ctx (struct erofs_fscache_context *ctx) +{ +erofs_fscache_cleanup_cookie (ctx); +} + +struct erofs_fscache_context *erofs_fscache_get_ctx (struct super_block *sb, +char *path) +{ +struct erofs_fscache_context *ctx; +int ret; + + ctx =kzalloc (sizeof (*ctx),GFP_KERNEL ); +if (!ctx) +return ERR_PTR (-ENOMEM ); + + ret =erofs_fscache_init_ctx (ctx, sb, path); +if (ret) { +kfree (ctx); +return ERR_PTR (ret); + } + +return ctx; +} + +void erofs_fscache_put_ctx (struct erofs_fscache_context *ctx) +{ +if (!ctx) +return ; + +erofs_fscache_cleanup_ctx (ctx); +kfree (ctx); +} +int __init erofs_init_fscache (void ) { volume =fscache_acquire_volume ("erofs" ,NULL ,NULL ,0 );
fs/erofs/internal.h +8 -0
@@ -97,6 +97,10 @@ struct erofs_sb_lz4_info {u16 max_pclusterblks; }; +struct erofs_fscache_context { +struct fscache_cookie *cookie; +}; +struct erofs_sb_info {struct erofs_mount_opts opt;/* options */ #ifdef CONFIG_EROFS_FS_ZIP
@@ -621,6 +625,10 @@ static inline int z_erofs_load_lzma_config(struct super_block *sb,int erofs_init_fscache (void );void erofs_exit_fscache (void ); +struct erofs_fscache_context *erofs_fscache_get_ctx (struct super_block *sb, +char *path); +void erofs_fscache_put_ctx (struct erofs_fscache_context *ctx); +#define EFSCORRUPTED EUCLEAN/* Filesystem is corrupted */ #endif /* __EROFS_INTERNAL_H */
Signature · 1 line
2.27.0
PATCH v4 12/21 erofs: add anonymous inode managing page cache of blob file
In reply to Jeffle XuIntroduce one anonymous inode for managing page cache of corresponding blob file. Then erofs could read directly from the address space of the anonymous inode when cache hit.
--- fs/erofs/fscache.c | 45 ++++++++++++++++++++++++++++++++++++++++++--- fs/erofs/internal.h | 3 ++- 2 files changed, 44 insertions(+), 4 deletions(-)
fs/erofs/fscache.c +42 -3
@@ -6,6 +6,9 @@static struct fscache_volume *volume; +static const struct address_space_operations erofs_fscache_blob_aops = { +}; +static int erofs_fscache_init_cookie (struct erofs_fscache_context *ctx,char *path) {
@@ -36,8 +39,34 @@ void erofs_fscache_cleanup_cookie(struct erofs_fscache_context *ctx) ctx->cookie =NULL ; } +static int erofs_fscache_get_inode (struct erofs_fscache_context *ctx, +struct super_block *sb) +{ +struct inode *const inode =new_inode (sb); + +if (!inode) +return -ENOMEM ; + +set_nlink (inode,1 ); + inode->i_size =OFFSET_MAX ; + + inode->i_mapping->a_ops = &erofs_fscache_blob_aops; +mapping_set_gfp_mask (inode->i_mapping, +GFP_NOFS | __GFP_HIGHMEM | __GFP_MOVABLE); + ctx->inode = inode; +return 0 ; +} + +static inline +void erofs_fscache_put_inode (struct erofs_fscache_context *ctx) +{ +iput (ctx->inode); + ctx->inode =NULL ; +} +static int erofs_fscache_init_ctx (struct erofs_fscache_context *ctx, -struct super_block *sb,char *path) +struct super_block *sb,char *path, +bool need_inode) {int ret;
@@ -47,6 +76,15 @@ static int erofs_fscache_init_ctx(struct erofs_fscache_context *ctx,return ret; } +if (need_inode) { + ret =erofs_fscache_get_inode (ctx, sb); +if (ret) { +erofs_err (sb,"failed to get anonymous inode" ); +erofs_fscache_cleanup_cookie (ctx); +return ret; + } + } +return 0 ; }
@@ -54,10 +92,11 @@ static inlinevoid erofs_fscache_cleanup_ctx (struct erofs_fscache_context *ctx) {erofs_fscache_cleanup_cookie (ctx); +erofs_fscache_put_inode (ctx); }struct erofs_fscache_context *erofs_fscache_get_ctx (struct super_block *sb, -char *path) +char *path,bool need_inode) {struct erofs_fscache_context *ctx;int ret;
@@ -66,7 +105,7 @@ struct erofs_fscache_context *erofs_fscache_get_ctx(struct super_block *sb,if (!ctx)return ERR_PTR (-ENOMEM ); - ret =erofs_fscache_init_ctx (ctx, sb, path); + ret =erofs_fscache_init_ctx (ctx, sb, path, need_inode);if (ret) {kfree (ctx);return ERR_PTR (ret);
fs/erofs/internal.h +2 -1
@@ -99,6 +99,7 @@ struct erofs_sb_lz4_info {struct erofs_fscache_context {struct fscache_cookie *cookie; +struct inode *inode; };struct erofs_sb_info {
@@ -626,7 +627,7 @@ int erofs_init_fscache(void);void erofs_exit_fscache (void );struct erofs_fscache_context *erofs_fscache_get_ctx (struct super_block *sb, -char *path); +char *path,bool need_inode);void erofs_fscache_put_ctx (struct erofs_fscache_context *ctx);#define EFSCORRUPTED EUCLEAN/* Filesystem is corrupted */
Signature · 1 line
2.27.0
PATCH v4 13/21 erofs: add erofs_fscache_read_pages() helper
In reply to Jeffle XuAdd erofs_fscache_read_pages() helper reading from fscache. It supports on-demand read semantics. That is, it will make the backend prepare for the data when cache miss. Once data ready, it will reinitiate a read from the cache.
This helper can then be used to implement .readpage()/.readahead() of on-demand read semantics.
Besides also add erofs_fscache_read_page() wrapper helper.
--- fs/erofs/fscache.c | 38 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 38 insertions(+)
fs/erofs/fscache.c +38 -0
@@ -6,6 +6,44 @@static struct fscache_volume *volume; +/* +* erofs_fscache_read_pages - Read data from fscache. +* +* Fill the readed data into page cache described by @start/len, which +* shall be both aligned with PAGE_SIZE. @pstart describes the corresponding +* physical start address in the cache file. +*/ +static int erofs_fscache_read_pages (struct fscache_cookie *cookie, +struct address_space *mapping, +loff_t start,size_t len, +loff_t pstart) +{ +struct netfs_cache_resources cres; +struct iov_iter iter; +int ret; + +memset (&cres,0 ,sizeof (cres)); + + ret =fscache_begin_read_operation (&cres, cookie); +if (ret) +return ret; + +iov_iter_xarray (&iter,READ , &mapping->i_pages, start, len); + + ret =fscache_read (&cres, pstart, &iter, +NETFS_READ_HOLE_ONDEMAND ,NULL ,NULL ); + +fscache_end_operation (&cres); +return ret; +} + +static inline int erofs_fscache_read_page (struct fscache_cookie *cookie, +struct page *page,loff_t pstart) +{ +return erofs_fscache_read_pages (cookie,page_file_mapping (page), +page_offset (page),PAGE_SIZE , pstart); +} +static const struct address_space_operations erofs_fscache_blob_aops = { };
Signature · 1 line
2.27.0
PATCH v4 14/21 erofs: register cookie context for bootstrap blob
In reply to Jeffle XuRegisters fscache_cookie for the bootstrap blob file. The bootstrap blob file can be specified by a new mount option, which is going to be introduced by a following patch.
Something worth mentioning about the cleanup routine.
1. The init routine is prior to when the root inode gets initialized, and thus the corresponding cleanup routine shall be placed under .kill_sb() callback.
2. The init routine will instantiate anonymous inodes under the super_block, and thus .put_super() callback shall also contain the cleanup routine. Or we'll get "VFS: Busy inodes after unmount." warning.
--- fs/erofs/internal.h | 3 +++ fs/erofs/super.c | 13 +++++++++++++ 2 files changed, 16 insertions(+)
fs/erofs/internal.h +3 -0
@@ -75,6 +75,7 @@ struct erofs_mount_opts {unsigned int max_sync_decompress_pages;#endif unsigned int mount_opt; +char *uuid; };struct erofs_dev_context {
@@ -152,6 +153,8 @@ struct erofs_sb_info {/* sysfs support */ struct kobject s_kobj;/* /sys/fs/erofs/<devname> */ struct completion s_kobj_unregister; + +struct erofs_fscache_context *bootstrap; };#define EROFS_SB (sb) ((struct erofs_sb_info *)(sb)->s_fs_info)
fs/erofs/super.c +13 -0
@@ -598,6 +598,16 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc) sbi->devs = ctx->devs; ctx->devs =NULL ; +if (!erofs_bdev_mode (sb)) { +struct erofs_fscache_context *bootstrap; + + bootstrap =erofs_fscache_get_ctx (sb, ctx->opt.uuid, true); +if (IS_ERR (bootstrap)) +return PTR_ERR (bootstrap); + + sbi->bootstrap = bootstrap; + } + err =erofs_read_superblock (sb);if (err)return err;
@@ -753,6 +763,7 @@ static void erofs_kill_sb(struct super_block *sb)return ;erofs_free_dev_context (sbi->devs); +erofs_fscache_put_ctx (sbi->bootstrap);fs_put_dax (sbi->dax_dev);kfree (sbi); sb->s_fs_info =NULL ;
@@ -771,6 +782,8 @@ static void erofs_put_super(struct super_block *sb)iput (sbi->managed_cache); sbi->managed_cache =NULL ;#endif +erofs_fscache_put_ctx (sbi->bootstrap); + sbi->bootstrap =NULL ; }static struct file_system_type erofs_fs_type = {
Signature · 1 line
2.27.0
PATCH v4 15/21 erofs: implement fscache-based metadata read
In reply to Jeffle XuThis patch implements the data plane of reading metadata from bootstrap blob file over fscache.
Be noted that currently it only supports the scenario where the backing file has no hole. Once it hits a hole of the backing file, erofs will fail the IO with -EOPNOTSUPP for now. The following patch will fix this issue, i.e. implementing the demand reading mode.
--- fs/erofs/data.c | 11 +++++++++-- fs/erofs/fscache.c | 24 ++++++++++++++++++++++++ fs/erofs/internal.h | 3 +++ 3 files changed, 36 insertions(+), 2 deletions(-)
fs/erofs/data.c +9 -2
@@ -31,15 +31,22 @@ void erofs_put_metabuf(struct erofs_buf *buf)void *erofs_read_metabuf (struct erofs_buf *buf,struct super_block *sb,erofs_blk_t blkaddr,enum erofs_kmap_type type) { -struct address_space *const mapping = sb->s_bdev->bd_inode->i_mapping; +struct address_space *mapping; +struct erofs_sb_info *sbi =EROFS_SB (sb);erofs_off_t offset =blknr_to_addr (blkaddr);pgoff_t index = offset >>PAGE_SHIFT ;struct page *page = buf->page;if (!page || page->index != index) {erofs_put_metabuf (buf); - page =read_cache_page_gfp (mapping, index, +if (erofs_bdev_mode (sb)) { + mapping = sb->s_bdev->bd_inode->i_mapping; + page =read_cache_page_gfp (mapping, index,mapping_gfp_constraint (mapping, ~__GFP_FS)); + }else { + page =erofs_fscache_read_cache_page (sbi->bootstrap, + index); + }if (IS_ERR (page))return page;/* should already be PageUptodate, no need to lock page */
fs/erofs/fscache.c +24 -0
@@ -44,9 +44,33 @@ static inline int erofs_fscache_read_page(struct fscache_cookie *cookie, page_offset(page), PAGE_SIZE, pstart); } +static int erofs_fscache_readpage_blob (struct file *data,struct page *page) +{ +int ret; +struct erofs_fscache_context *ctx = + (struct erofs_fscache_context *)data; + + ret =erofs_fscache_read_page (ctx->cookie, page,page_offset (page)); +if (!ret) +SetPageUptodate (page); +else +SetPageError (page); + +unlock_page (page); +return ret; +} +static const struct address_space_operations erofs_fscache_blob_aops = { + .readpage = erofs_fscache_readpage_blob, }; +struct page *erofs_fscache_read_cache_page (struct erofs_fscache_context *ctx, +pgoff_t index) +{ +DBG_BUGON (!ctx->inode); +return read_mapping_page (ctx->inode->i_mapping, index, ctx); +} +static int erofs_fscache_init_cookie (struct erofs_fscache_context *ctx,char *path) {
fs/erofs/internal.h +3 -0
@@ -633,6 +633,9 @@ struct erofs_fscache_context *erofs_fscache_get_ctx(struct super_block *sb, char *path, bool need_inode);void erofs_fscache_put_ctx (struct erofs_fscache_context *ctx); +struct page *erofs_fscache_read_cache_page (struct erofs_fscache_context *ctx, +pgoff_t index); +#define EFSCORRUPTED EUCLEAN/* Filesystem is corrupted */ #endif /* __EROFS_INTERNAL_H */
Signature · 1 line
2.27.0
PATCH v4 16/21 erofs: implement fscache-based data read for non-inline layout
In reply to Jeffle XuThis patch implements the data plane of reading data from bootstrap blob file over fscache for non-inline layout.
Be noted that compressed layout is not supported yet.
--- fs/erofs/fscache.c | 94 +++++++++++++++++++++++++++++++++++++++++++++ fs/erofs/inode.c | 6 ++- fs/erofs/internal.h | 1 + 3 files changed, 100 insertions(+), 1 deletion(-)
fs/erofs/fscache.c +94 -0
@@ -4,6 +4,12 @@*/ #include "internal.h" +struct erofs_fscache_map { +struct erofs_fscache_context *m_ctx; +erofs_off_t m_pa, m_la, o_la; +u64 m_llen; +}; +static struct fscache_volume *volume;/*
@@ -60,10 +66,98 @@ static int erofs_fscache_readpage_blob(struct file *data, struct page *page)return ret; } +static inline int erofs_fscache_get_map (struct erofs_fscache_map *fsmap, +struct erofs_map_blocks *map, +struct super_block *sb) +{ +struct erofs_sb_info *sbi =EROFS_SB (sb); + + fsmap->m_ctx = sbi->bootstrap; + fsmap->m_la = map->m_la; + fsmap->m_pa = map->m_pa; + fsmap->m_llen = map->m_llen; + +return 0 ; +} + +static int erofs_fscache_readpage_noinline (struct page *page, +struct erofs_fscache_map *fsmap) +{ +struct fscache_cookie *cookie = fsmap->m_ctx->cookie; +/* +* 1) For FLAT_PLAIN layout, the output map.m_la shall be equal to o_la, +* and the output map.m_pa is exactly the physical address of o_la. +* 2) For CHUNK_BASED layout, the output map.m_la is rounded down to the +* nearest chunk boundary, and the output map.m_pa is actually the +* physical address of this chunk boundary. So we need to recalculate +* the actual physical address of o_la. +*/ +loff_t start = fsmap->m_pa + (fsmap->o_la - fsmap->m_la); + +return erofs_fscache_read_page (cookie, page, start); +} + +static int erofs_fscache_do_readpage (struct page *page) +{ +struct inode *inode = page->mapping->host; +struct erofs_inode *vi =EROFS_I (inode); +struct super_block *sb = inode->i_sb; +struct erofs_map_blocks map; +struct erofs_fscache_map fsmap; +int ret; + +if (erofs_inode_is_data_compressed (vi->datalayout)) { +erofs_info (sb,"compressed layout not supported yet" ); +return -EOPNOTSUPP ; + } + + map.m_la = fsmap.o_la =page_offset (page); + + ret =erofs_map_blocks (inode, &map,EROFS_GET_BLOCKS_RAW ); +if (ret) +return ret; + +if (!(map.m_flags &EROFS_MAP_MAPPED )) { +zero_user (page,0 ,PAGE_SIZE ); +return 0 ; + } + + ret =erofs_fscache_get_map (&fsmap, &map, sb); +if (ret) +return ret; + +switch (vi->datalayout) { +case EROFS_INODE_FLAT_PLAIN : +case EROFS_INODE_CHUNK_BASED : +return erofs_fscache_readpage_noinline (page, &fsmap); +default : +DBG_BUGON (1 ); +return -EOPNOTSUPP ; + } +} + +static int erofs_fscache_readpage (struct file *file,struct page *page) +{ +int ret; + + ret =erofs_fscache_do_readpage (page); +if (!ret) +SetPageUptodate (page); +else +SetPageError (page); + +unlock_page (page); +return ret; +} +static const struct address_space_operations erofs_fscache_blob_aops = { .readpage = erofs_fscache_readpage_blob, }; +const struct address_space_operations erofs_fscache_access_aops = { + .readpage = erofs_fscache_readpage, +}; +struct page *erofs_fscache_read_cache_page (struct erofs_fscache_context *ctx,pgoff_t index) {
fs/erofs/inode.c +5 -1
@@ -296,7 +296,11 @@ static int erofs_fill_inode(struct inode *inode, int isdir) err =z_erofs_fill_inode (inode); goto out_unlock; } - inode->i_mapping->a_ops = &erofs_raw_access_aops; + +if (erofs_bdev_mode (inode->i_sb)) + inode->i_mapping->a_ops = &erofs_raw_access_aops; +else + inode->i_mapping->a_ops = &erofs_fscache_access_aops; out_unlock:erofs_put_metabuf (&buf);
fs/erofs/internal.h +1 -0
@@ -393,6 +393,7 @@ struct page *erofs_grab_cache_page_nowait(struct address_space *mapping,extern const struct super_operations erofs_sops;extern const struct address_space_operations erofs_raw_access_aops; +extern const struct address_space_operations erofs_fscache_access_aops;extern const struct address_space_operations z_erofs_aops;/*
Signature · 1 line
2.27.0
PATCH v4 17/21 erofs: implement fscache-based data read for inline layout
In reply to Jeffle XuThis patch implements the data plane of reading data from bootstrap blob file over fscache for inline layout.
For the heading non-inline part, the data plane for non-inline layout is resued, while only the tail packing part needs special handling.
--- fs/erofs/fscache.c | 43 +++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 41 insertions(+), 2 deletions(-)
fs/erofs/fscache.c +41 -2
@@ -85,8 +85,9 @@ static int erofs_fscache_readpage_noinline(struct page *page, {struct fscache_cookie *cookie = fsmap->m_ctx->cookie;/* -* 1) For FLAT_PLAIN layout, the output map.m_la shall be equal to o_la, -* and the output map.m_pa is exactly the physical address of o_la. +* 1) For FLAT_PLAIN and FLAT_INLINE (the heading non tail packing part) +* layout, the output map.m_la shall be equal to o_la, and the output +* map.m_pa is exactly the physical address of o_la. * 2) For CHUNK_BASED layout, the output map.m_la is rounded down to the * nearest chunk boundary, and the output map.m_pa is actually the * physical address of this chunk boundary. So we need to recalculate
@@ -97,6 +98,40 @@ static int erofs_fscache_readpage_noinline(struct page *page,return erofs_fscache_read_page (cookie, page, start); } +static int erofs_fscache_readpage_inline (struct page *page, +struct erofs_fscache_map *fsmap) +{ +struct inode *inode = page->mapping->host; +struct super_block *sb = inode->i_sb; +struct erofs_buf buf = __EROFS_BUF_INITIALIZER; +erofs_blk_t blknr; +size_t offset, len; +void *src, *dst; + +/* +* For inline (tail packing) layout, the offset may be non-zero, which +* can be calculated from corresponding physical address directly. +* Currently only flat layout supports inline (FLAT_INLINE), and the +* output map.m_pa is exactly the physical address of o_la in this case. +*/ + offset =erofs_blkoff (fsmap->m_pa); + blknr =erofs_blknr (fsmap->m_pa); + len = fsmap->m_llen; + + src =erofs_read_metabuf (&buf, sb, blknr,EROFS_KMAP ); +if (IS_ERR (src)) +return PTR_ERR (src); + + dst =kmap (page); +memcpy (dst, src + offset, len); +memset (dst + len,0 ,PAGE_SIZE - len); +kunmap (page); + +erofs_put_metabuf (&buf); + +return 0 ; +} +static int erofs_fscache_do_readpage (struct page *page) {struct inode *inode = page->mapping->host;
@@ -126,8 +161,12 @@ static int erofs_fscache_do_readpage(struct page *page)if (ret)return ret; +if (map.m_flags &EROFS_MAP_META ) +return erofs_fscache_readpage_inline (page, &fsmap); +switch (vi->datalayout) {case EROFS_INODE_FLAT_PLAIN : +case EROFS_INODE_FLAT_INLINE :case EROFS_INODE_CHUNK_BASED :return erofs_fscache_readpage_noinline (page, &fsmap);default :
Signature · 1 line
2.27.0
PATCH v4 18/21 erofs: register cookie context for data blobs
In reply to Jeffle XuSimilar to the multi device mode, erofs could be mounted from multiple blob files (one bootstrap blob file and optional multiple data blob files). In this case, each device slot contains the path of corresponding data blob file.
This patch registers corresponding cookie context for each data blob file.
--- fs/erofs/internal.h | 1 + fs/erofs/super.c | 27 +++++++++++++++++++-------- 2 files changed, 20 insertions(+), 8 deletions(-)
fs/erofs/internal.h +1 -0
@@ -53,6 +53,7 @@ struct erofs_device_info {struct block_device *bdev;struct dax_device *dax_dev;u64 dax_part_off; +struct erofs_fscache_context *ctx;u32 blocks;u32 mapped_blkaddr;
fs/erofs/super.c +19 -8
@@ -250,6 +250,7 @@ static int erofs_init_devices(struct super_block *sb,down_read (&sbi->devs->rwsem);idr_for_each_entry (&sbi->devs->tree, dif, id) {struct block_device *bdev; +struct erofs_fscache_context *ctx; ptr =erofs_read_metabuf (&buf, sb,erofs_blknr (pos),EROFS_KMAP );
@@ -259,15 +260,24 @@ static int erofs_init_devices(struct super_block *sb, } dis = ptr +erofs_blkoff (pos); - bdev =blkdev_get_by_path (dif->path, -FMODE_READ |FMODE_EXCL , - sb->s_type); -if (IS_ERR (bdev)) { - err =PTR_ERR (bdev); -break ; +if (erofs_bdev_mode (sb)) { + bdev =blkdev_get_by_path (dif->path, +FMODE_READ |FMODE_EXCL , + sb->s_type); +if (IS_ERR (bdev)) { + err =PTR_ERR (bdev); +break ; + } + dif->bdev = bdev; + dif->dax_dev =fs_dax_get_by_bdev (bdev, &dif->dax_part_off); + }else { + ctx =erofs_fscache_get_ctx (sb, dif->path, false); +if (IS_ERR (ctx)) { + err =PTR_ERR (ctx); +break ; + } + dif->ctx = ctx; } - dif->bdev = bdev; - dif->dax_dev =fs_dax_get_by_bdev (bdev, &dif->dax_part_off); dif->blocks =le32_to_cpu (dis->blocks); dif->mapped_blkaddr =le32_to_cpu (dis->mapped_blkaddr); sbi->total_blocks += dif->blocks;
@@ -694,6 +704,7 @@ static int erofs_release_device_info(int id, void *ptr, void *data) {struct erofs_device_info *dif = ptr; +erofs_fscache_put_ctx (dif->ctx);fs_put_dax (dif->dax_dev);if (dif->bdev)blkdev_put (dif->bdev,FMODE_READ |FMODE_EXCL );
Signature · 1 line
2.27.0
PATCH v4 19/21 erofs: implement fscache-based data read for data blobs
In reply to Jeffle XuThis patch implements the data plane of reading data from data blob file over fscache.
--- fs/erofs/data.c | 3 +++ fs/erofs/fscache.c | 16 +++++++++++++--- fs/erofs/internal.h | 1 + 3 files changed, 17 insertions(+), 3 deletions(-)
fs/erofs/data.c +3 -0
@@ -200,6 +200,7 @@ int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *map) map->m_bdev = sb->s_bdev; map->m_daxdev =EROFS_SB (sb)->dax_dev; map->m_dax_part_off =EROFS_SB (sb)->dax_part_off; + map->m_ctx =EROFS_SB (sb)->bootstrap;if (map->m_deviceid) {down_read (&devs->rwsem);
@@ -211,6 +212,7 @@ int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *map) map->m_bdev = dif->bdev; map->m_daxdev = dif->dax_dev; map->m_dax_part_off = dif->dax_part_off; + map->m_ctx = dif->ctx;up_read (&devs->rwsem); }else if (devs->extra_devices) {down_read (&devs->rwsem);
@@ -228,6 +230,7 @@ int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *map) map->m_bdev = dif->bdev; map->m_daxdev = dif->dax_dev; map->m_dax_part_off = dif->dax_part_off; + map->m_ctx = dif->ctx;break ; } }
fs/erofs/fscache.c +13 -3
@@ -70,11 +70,21 @@ static inline int erofs_fscache_get_map(struct erofs_fscache_map *fsmap,struct erofs_map_blocks *map, struct super_block *sb) { - struct erofs_sb_info *sbi = EROFS_SB(sb); + struct erofs_map_dev mdev; +int ret; + + mdev = (struct erofs_map_dev ) { + .m_deviceid = map->m_deviceid, + .m_pa = map->m_pa, + }; + + ret =erofs_map_dev (sb, &mdev); +if (ret) +return ret; - fsmap->m_ctx = sbi->bootstrap; + fsmap->m_ctx = mdev.m_ctx; + fsmap->m_pa = mdev.m_pa; fsmap->m_la = map->m_la; - fsmap->m_pa = map->m_pa; fsmap->m_llen = map->m_llen;return 0 ;
fs/erofs/internal.h +1 -0
@@ -486,6 +486,7 @@ struct erofs_map_dev {struct block_device *m_bdev;struct dax_device *m_daxdev;u64 m_dax_part_off; +struct erofs_fscache_context *m_ctx;erofs_off_t m_pa;unsigned int m_deviceid;
Signature · 1 line
2.27.0
PATCH v4 20/21 erofs: implement fscache-based data readahead
In reply to Jeffle XuThis patch implements fscache-based data readahead. Also registers an individual bdi for each erofs instance to enable readahead.
--- fs/erofs/fscache.c | 153 +++++++++++++++++++++++++++++++++++++++++++++ fs/erofs/super.c | 4 ++ 2 files changed, 157 insertions(+)
fs/erofs/fscache.c +153 -0
@@ -10,6 +10,13 @@ struct erofs_fscache_map {u64 m_llen; }; +struct erofs_fscahce_ra_ctx { +struct readahead_control *rac; +struct address_space *mapping; +loff_t start; +size_t len, done; +}; +static struct fscache_volume *volume;/*
@@ -199,12 +206,158 @@ static int erofs_fscache_readpage(struct file *file, struct page *page)return ret; } +static inline size_t erofs_fscache_calc_len (struct erofs_fscahce_ra_ctx *ractx, +struct erofs_fscache_map *fsmap) +{ +/* +* 1) For CHUNK_BASED layout, the output m_la is rounded down to the +* nearest chunk boundary, and the output m_llen actually starts from +* the start of the containing chunk. +* 2) For other cases, the output m_la is equal to o_la. +*/ +size_t len = fsmap->m_llen - (fsmap->o_la - fsmap->m_la); + +return min_t (size_t, len, ractx->len - ractx->done); +} + +static inline void erofs_fscache_unlock_pages (struct readahead_control *rac, +size_t len) +{ +while (len) { +struct page *page =readahead_page (rac); + +SetPageUptodate (page); +unlock_page (page); +put_page (page); + + len -=PAGE_SIZE ; + } +} + +static int erofs_fscache_ra_hole (struct erofs_fscahce_ra_ctx *ractx, +struct erofs_fscache_map *fsmap) +{ +struct iov_iter iter; +loff_t start = ractx->start + ractx->done; +size_t length =erofs_fscache_calc_len (ractx, fsmap); + +iov_iter_xarray (&iter,READ , &ractx->mapping->i_pages, start, length); +iov_iter_zero (length, &iter); + +erofs_fscache_unlock_pages (ractx->rac, length); +return length; +} + +static int erofs_fscache_ra_noinline (struct erofs_fscahce_ra_ctx *ractx, +struct erofs_fscache_map *fsmap) +{ +struct fscache_cookie *cookie = fsmap->m_ctx->cookie; +loff_t start = ractx->start + ractx->done; +size_t length =erofs_fscache_calc_len (ractx, fsmap); +loff_t pstart = fsmap->m_pa + (fsmap->o_la - fsmap->m_la); +int ret; + + ret =erofs_fscache_read_pages (cookie, ractx->mapping, + start, length, pstart); +if (!ret) { +erofs_fscache_unlock_pages (ractx->rac, length); + ret = length; + } + +return ret; +} + +static int erofs_fscache_ra_inline (struct erofs_fscahce_ra_ctx *ractx, +struct erofs_fscache_map *fsmap) +{ +struct page *page =readahead_page (ractx->rac); +int ret; + + ret =erofs_fscache_readpage_inline (page, fsmap); +if (!ret) { +SetPageUptodate (page); + ret =PAGE_SIZE ; + } + +unlock_page (page); +put_page (page); +return ret; +} + +static void erofs_fscache_readahead (struct readahead_control *rac) +{ +struct inode *inode = rac->mapping->host; +struct erofs_inode *vi =EROFS_I (inode); +struct super_block *sb = inode->i_sb; +struct erofs_fscahce_ra_ctx ractx; +int ret; + +if (erofs_inode_is_data_compressed (vi->datalayout)) { +erofs_info (sb,"compressed layout not supported yet" ); +return ; + } + +if (!readahead_count (rac)) +return ; + + ractx = (struct erofs_fscahce_ra_ctx ) { + .rac = rac, + .mapping = rac->mapping, + .start =readahead_pos (rac), + .len =readahead_length (rac), + }; + +do { +struct erofs_map_blocks map; +struct erofs_fscache_map fsmap; + + map.m_la = fsmap.o_la = ractx.start + ractx.done; + + ret =erofs_map_blocks (inode, &map,EROFS_GET_BLOCKS_RAW ); +if (ret) +return ; + +if (!(map.m_flags &EROFS_MAP_MAPPED )) { +/* +* Two cases will hit this: +* 1) EOF. Imposibble in readahead routine; +* 2) hole. Only CHUNK_BASED layout supports hole. +*/ + fsmap.m_la = map.m_la; + fsmap.m_llen = map.m_llen; + ret =erofs_fscache_ra_hole (&ractx, &fsmap); +continue ; + } + + ret =erofs_fscache_get_map (&fsmap, &map, sb); +if (ret) +return ; + +if (map.m_flags &EROFS_MAP_META ) { + ret =erofs_fscache_ra_inline (&ractx, &fsmap); +continue ; + } + +switch (vi->datalayout) { +case EROFS_INODE_FLAT_PLAIN : +case EROFS_INODE_FLAT_INLINE : +case EROFS_INODE_CHUNK_BASED : + ret =erofs_fscache_ra_noinline (&ractx, &fsmap); +break ; +default : +DBG_BUGON (1 ); +return ; + } + }while (ret >0 && ((ractx.done += ret) < ractx.len)); +} +static const struct address_space_operations erofs_fscache_blob_aops = { .readpage = erofs_fscache_readpage_blob, };const struct address_space_operations erofs_fscache_access_aops = { .readpage = erofs_fscache_readpage, + .readahead = erofs_fscache_readahead, }; struct page *erofs_fscache_read_cache_page(struct erofs_fscache_context *ctx,
fs/erofs/super.c +4 -0
@@ -616,6 +616,10 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)return PTR_ERR (bootstrap); sbi->bootstrap = bootstrap; + + err =super_setup_bdi (sb); +if (err) +return err; } err =erofs_read_superblock (sb);
Signature · 1 line
2.27.0
PATCH v4 21/21 erofs: add 'uuid' mount option
In reply to Jeffle XuIntroduce 'uuid' mount option to enable on-demand read sementics. In this case, erofs could be mounted from blob files instead of blkdev. By then users could specify the path of bootstrap blob file containing the complete erofs image.
--- fs/erofs/super.c | 44 +++++++++++++++++++++++++++++++++++++------- 1 file changed, 37 insertions(+), 7 deletions(-)
fs/erofs/super.c +37 -7
@@ -400,6 +400,7 @@ enum { Opt_dax, Opt_dax_enum, Opt_device, + Opt_uuid, Opt_err };
@@ -424,6 +425,7 @@ static const struct fs_parameter_spec erofs_fs_parameters[] = {fsparam_flag ("dax" , Opt_dax),fsparam_enum ("dax" , Opt_dax_enum, erofs_dax_param_enums),fsparam_string ("device" , Opt_device), + fsparam_string("uuid", Opt_uuid), {} };
@@ -519,6 +521,12 @@ static int erofs_fc_parse_param(struct fs_context *fc, } ++ctx->devs->extra_devices;break ; +case Opt_uuid: +kfree (ctx->opt.uuid); + ctx->opt.uuid =kstrdup (param->string,GFP_KERNEL ); +if (!ctx->opt.uuid) +return -ENOMEM ; +break ;default :return -ENOPARAM ; }
@@ -593,9 +601,14 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc) sb->s_magic =EROFS_SUPER_MAGIC ; -if (!sb_set_blocksize (sb,EROFS_BLKSIZ )) { -erofs_err (sb,"failed to set erofs blksize" ); -return -EINVAL ; +if (erofs_bdev_mode (sb)) { +if (!sb_set_blocksize (sb,EROFS_BLKSIZ )) { +erofs_err (sb,"failed to set erofs blksize" ); +return -EINVAL ; + } + }else { + sb->s_blocksize =EROFS_BLKSIZ ; + sb->s_blocksize_bits =LOG_BLOCK_SIZE ; } sbi =kzalloc (sizeof (*sbi),GFP_KERNEL );
@@ -604,11 +617,12 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc) sb->s_fs_info = sbi; sbi->opt = ctx->opt; - sbi->dax_dev =fs_dax_get_by_bdev (sb->s_bdev, &sbi->dax_part_off); sbi->devs = ctx->devs; ctx->devs =NULL ; -if (!erofs_bdev_mode (sb)) { +if (erofs_bdev_mode (sb)) { + sbi->dax_dev =fs_dax_get_by_bdev (sb->s_bdev, &sbi->dax_part_off); + }else {struct erofs_fscache_context *bootstrap; bootstrap =erofs_fscache_get_ctx (sb, ctx->opt.uuid, true);
@@ -620,6 +634,8 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc) err =super_setup_bdi (sb);if (err)return err; + + sbi->dax_dev =NULL ; } err =erofs_read_superblock (sb);
@@ -682,6 +698,11 @@ static int erofs_fc_fill_super(struct super_block *sb, struct fs_context *fc)static int erofs_fc_get_tree (struct fs_context *fc) { +struct erofs_fs_context *ctx = fc->fs_private; + +if (ctx->opt.uuid) +return get_tree_nodev (fc, erofs_fc_fill_super); +return get_tree_bdev (fc, erofs_fc_fill_super); }
@@ -731,6 +752,7 @@ static void erofs_fc_free(struct fs_context *fc)struct erofs_fs_context *ctx = fc->fs_private;erofs_free_dev_context (ctx->devs); +kfree (ctx->opt.uuid);kfree (ctx); }
@@ -771,7 +793,10 @@ static void erofs_kill_sb(struct super_block *sb)WARN_ON (sb->s_magic !=EROFS_SUPER_MAGIC ); -kill_block_super (sb); +if (erofs_bdev_mode (sb)) +kill_block_super (sb); +else +generic_shutdown_super (sb); sbi =EROFS_SB (sb);if (!sbi)
@@ -889,7 +914,12 @@ static int erofs_statfs(struct dentry *dentry, struct kstatfs *buf) {struct super_block *sb = dentry->d_sb;struct erofs_sb_info *sbi =EROFS_SB (sb); -u64 id =huge_encode_dev (sb->s_bdev->bd_dev); +u64 id; + +if (erofs_bdev_mode (sb)) + id =huge_encode_dev (sb->s_bdev->bd_dev); +else + id =0 ;/* TODO */ buf->f_type = sb->s_magic; buf->f_bsize =EROFS_BLKSIZ ;
Signature · 1 line
2.27.0
Hi,
We're also interested in this way, hoping for the formal solution upstream so we can make use of it as well.
Thanks,
daowen
Quoted 184 lines · On Mon, 7 Mar 2022 20:32:44 +0800 Jeffle Xu <[email protected]> wrote:
changes since v3:
- cachefiles: The current implementation relies on the anonymous fd mechanism to avoid
the dependence on the format of cache file. When cache file is opened for the first time, ananon_fdassociated with the cache file is sent to user daemon. User daemon could fetch and write data to cache file with the givenanon_fd. The following write to theanon_fdwill finally call to cachefiles kernel module, which will write data to cache file in the latest format of cache file. Thus the on-demand read mode can keep working no matter how cache file format could change in the future. (patch 4)
- cachefiles: the on-demand read mode reuses the existing "/dev/cachefiles" devnode (patch 3)
- erofs: squash several commits implementing readahead into single commit (patch 20)
- erofs: refactor the readahead routine, so that it can read multiple pages each round (patch 20)
- patch 1 and 7 have already been cherry-picked by the maintainers, but have not been merged to the master. Keep them here for completeness.RFC: https://lore.kernel.org/all/[email protected]/t/
v1: https://lore.kernel.org/lkml/[email protected]/T/
v2: https://lore.kernel.org/all/[email protected]/t/
v3: https://lore.kernel.org/lkml/[email protected]/T/[Background]
============
Nydus [1] is a container image distribution service specially optimised for distribution over network. Nydus is an excellent container image acceleration solution, since it only pulls data from remote when it's really needed, a.k.a. on-demand reading.erofs (Enhanced Read-Only File System) is a filesystem specially optimised for read-only scenarios. (
Documentation/filesystem/erofs.rst)Recently we are focusing on erofs in container images distribution scenario [2], trying to combine it with nydus. In this case, erofs can be mounted from one bootstrap file (metadata) with (optional) multiple data blob files (data) stored on another local filesystem. (All these files are actually image files in erofs disk format.)
To accelerate the container startup (fetching container image from remote and then start the container), we do hope that the bootstrap blob file could support demand read. That is, erofs can be mounted and accessed even when the bootstrap/data blob files have not been fully downloaded.
That means we have to manage the cache state of the bootstrap/data blob files (if cache hit, read directly from the local cache; if cache miss, fetch the data somehow). It would be painful and may be dumb for erofs to implement the cache management itself. Thus we prefer fscache/cachefiles to do the cache management. Besides, the demand-read feature shall be general and it can benefit other using scenarios if it can be implemented in fscache level.
[1] https://nydus.dev
[2] https://sched.co/pcdL[Overall Design]
================Please refer to patch 6 ("cachefiles: document on-demand read mode") for more details.
When working in original mode, cachefiles mainly serves as a local cache for remote networking fs, while in on-demand read mode, cachefiles can boost the scenario where on-demand read semantics is needed, e.g. container image distribution.
The essential difference between these two modes is that, in original mode, when cache miss, netfs itself will fetch data from remote, and then write the fetched data into cache file. While in on-demand read mode, a user daemon is responsible for fetching data and then writing to the cache file.
The on-demand read mode relies on a simple protocol used for communication between kernel and user daemon.
The current implementation relies on the anonymous fd mechanism to avoid the dependence on the format of cache file. When cache file is opened for the first time, an
anon_fdassociated with the cache file is sent to user daemon. With the givenanon_fd, user daemon could fetch and write data into the cache file in the background, even when kernel has not triggered the cache miss. Besides, thewrite()syscall to theanon_fdwill finally call cachefiles kernel module, which will write data to cache file in the latest format of cache file.1. cache miss
When cache miss, cachefiles kernel module will notify user daemon theanon_fd, along with the requested file range. When notified, user dameon needs to fetch data of the requested file range, and then write the fetched data into cache file with the given anonymous fd. When finished processing the request, user daemon needs to notify the kernel.After notifying the user daemon, the kernel read routine will hang there, until the request is handled by user daemon. When it's awaken by the notification from user daemon, i.e. the corresponding hole has been filled by the user daemon, it will retry to read from the same file range.
2. cache hit
Once data is already ready in cache file, netfs will read from cache file directly.[Advantage of fscache-based demand-read]
========================================
1. Asynchronous Prefetch
In current mechanism, fscache is responsible for cache state management, while the data plane (fetch data from local/remote on cache miss) is done on the user daemon side.If data has already been ready in the backing file, the upper fs (e.g. erofs) will read from the backing file directly and won't be trapped to user space anymore. Thus the user daemon could fetch data (from remote) asynchronously on the background, and thus accelerate the backing file accessing in some degree.
2. Support massive blob files
Besides this mechanism supports a large amount of backing files, and thus can benefit the densely employed scenario.In our using scenario, one container image can correspond to one bootstrap file (required) and multiple data blob files (optional). For example, one container image for node.js will corresponds to ~20 files in total. In densely employed environment, there could be as many as hundreds of containers and thus thousands of backing files on one machine.
[Test]
==========
You could start a quick test by
https://github.com/lostjeffle/demand-read-cachefilesdJeffle Xu (21):
fscache: exportfscache_end_operation()
cachefiles: export write routine
cachefiles: introduce on-demand read mode
cachefiles: notify user daemon withanon_fdwhen opening cache file
cachefiles: implement on-demand read
cachefiles: document on-demand read mode
erofs: use meta buffers forerofs_read_superblock()
erofs: exporterofs_map_blocks()
erofs: add mode checking helper
erofs: register global fscache volume
erofs: add cookie context helper functions
erofs: add anonymous inode managing page cache of blob file
erofs: adderofs_fscache_read_pages()helper
erofs: register cookie context for bootstrap blob
erofs: implement fscache-based metadata read
erofs: implement fscache-based data read for non-inline layout
erofs: implement fscache-based data read for inline layout
erofs: register cookie context for data blobs
erofs: implement fscache-based data read for data blobs
erofs: implement fscache-based data readahead
erofs: add 'uuid' mount option.../filesystems/caching/cachefiles.rst | 159 +++++
fs/cachefiles/Kconfig| 11 +fs/cachefiles/daemon.c| 576 +++++++++++++++++-fs/cachefiles/internal.h| 48 ++fs/cachefiles/io.c| 72 ++-fs/cachefiles/namei.c| 16 +-fs/erofs/Makefile| 3 +-fs/erofs/data.c| 18 +-fs/erofs/fscache.c| 496 +++++++++++++++fs/erofs/inode.c| 6 +-fs/erofs/internal.h| 30 +fs/erofs/super.c| 106 +++-fs/fscache/internal.h| 11 -fs/nfs/fscache.c| 8 -include/linux/fscache.h| 15 +include/linux/netfs.h| 1 +include/trace/events/cachefiles.h| 2 +include/uapi/linux/cachefiles.h| 48 ++18 files changed, 1526 insertions(+), 100 deletions(-)create mode 100644
fs/erofs/fscache.ccreate mode 100644include/uapi/linux/cachefiles.h--
2.27.0
This sounds great!
Hope it will come into use to speed up fscache.
Best wishes,
naihao
-------- Original Message --------
changes since v3:
- cachefiles: The current implementation relies on the anonymous fd mechanism to avoid the dependence on the format of cache file. When cache file is opened for the first time, an anon_fd associated with the cache file is sent to user daemon. User daemon could fetch and write data to cache file with the given anon_fd. The following write to the anon_fd will finally call to cachefiles kernel module, which will write data to cache file in the latest format of cache file. Thus the on-demand read mode can keep working no matter how cache file format could change in the future. (patch 4)
- cachefiles: the on-demand read mode reuses the existing "/dev/cachefiles" devnode (patch 3)
- erofs: squash several commits implementing readahead into single commit (patch 20)
- erofs: refactor the readahead routine, so that it can read multiple pages each round (patch 20)
- patch 1 and 7 have already been cherry-picked by the maintainers, but have not been merged to the master. Keep them here for completeness.
RFC: https://lore.kernel.org/all/[email protected]/t/
v1: https://lore.kernel.org/lkml/[email protected]/T/
v2: https://lore.kernel.org/all/[email protected]/t/
v3: https://lore.kernel.org/lkml/[email protected]/T/
[Background]
============
Nydus [1] is a container image distribution service specially optimised for distribution over network. Nydus is an excellent container image acceleration solution, since it only pulls data from remote when it's really needed, a.k.a. on-demand reading.
erofs (Enhanced Read-Only File System) is a filesystem specially optimised for read-only scenarios. (Documentation/filesystem/erofs.rst)
Recently we are focusing on erofs in container images distribution scenario [2], trying to combine it with nydus. In this case, erofs can be mounted from one bootstrap file (metadata) with (optional) multiple data blob files (data) stored on another local filesystem. (All these files are actually image files in erofs disk format.)
To accelerate the container startup (fetching container image from remote and then start the container), we do hope that the bootstrap blob file could support demand read. That is, erofs can be mounted and accessed even when the bootstrap/data blob files have not been fully downloaded.
That means we have to manage the cache state of the bootstrap/data blob files (if cache hit, read directly from the local cache; if cache miss, fetch the data somehow). It would be painful and may be dumb for erofs to implement the cache management itself. Thus we prefer fscache/cachefiles to do the cache management. Besides, the demand-read feature shall be general and it can benefit other using scenarios if it can be implemented in fscache level.
[1] https://nydus.dev
[2] https://sched.co/pcdL
[Overall Design]
================
Please refer to patch 6 ("cachefiles: document on-demand read mode") for more details.
When working in original mode, cachefiles mainly serves as a local cache for remote networking fs, while in on-demand read mode, cachefiles can boost the scenario where on-demand read semantics is needed, e.g. container image distribution.
The essential difference between these two modes is that, in original mode, when cache miss, netfs itself will fetch data from remote, and then write the fetched data into cache file. While in on-demand read mode, a user daemon is responsible for fetching data and then writing to the cache file.
The on-demand read mode relies on a simple protocol used for communication between kernel and user daemon.
The current implementation relies on the anonymous fd mechanism to avoid the dependence on the format of cache file. When cache file is opened for the first time, an anon_fd associated with the cache file is sent to user daemon. With the given anon_fd, user daemon could fetch and write data into the cache file in the background, even when kernel has not triggered the cache miss. Besides, the write() syscall to the anon_fd will finally call cachefiles kernel module, which will write data to cache file in the latest format of cache file.
1. cache miss
When cache miss, cachefiles kernel module will notify user daemon the anon_fd, along with the requested file range. When notified, user dameon needs to fetch data of the requested file range, and then write the fetched data into cache file with the given anonymous fd. When finished processing the request, user daemon needs to notify the kernel.
After notifying the user daemon, the kernel read routine will hang there, until the request is handled by user daemon. When it's awaken by the notification from user daemon, i.e. the corresponding hole has been filled by the user daemon, it will retry to read from the same file range.
2. cache hit
Once data is already ready in cache file, netfs will read from cache file directly.
[Advantage of fscache-based demand-read]
========================================
1. Asynchronous Prefetch
In current mechanism, fscache is responsible for cache state management, while the data plane (fetch data from local/remote on cache miss) is done on the user daemon side.
If data has already been ready in the backing file, the upper fs (e.g. erofs) will read from the backing file directly and won't be trapped to user space anymore. Thus the user daemon could fetch data (from remote) asynchronously on the background, and thus accelerate the backing file accessing in some degree.
2. Support massive blob files
Besides this mechanism supports a large amount of backing files, and thus can benefit the densely employed scenario.
In our using scenario, one container image can correspond to one bootstrap file (required) and multiple data blob files (optional). For example, one container image for node.js will corresponds to ~20 files in total. In densely employed environment, there could be as many as hundreds of containers and thus thousands of backing files on one machine.
[Test]
==========
You could start a quick test by
https://github.com/lostjeffle/demand-read-cachefilesd
Jeffle Xu (21):
fscache: export fscache_end_operation()
cachefiles: export write routine
cachefiles: introduce on-demand read mode
cachefiles: notify user daemon with anon_fd when opening cache file
cachefiles: implement on-demand read
cachefiles: document on-demand read mode
erofs: use meta buffers for erofs_read_superblock()
erofs: export erofs_map_blocks()
erofs: add mode checking helper
erofs: register global fscache volume
erofs: add cookie context helper functions
erofs: add anonymous inode managing page cache of blob file
erofs: add erofs_fscache_read_pages() helper
erofs: register cookie context for bootstrap blob
erofs: implement fscache-based metadata read
erofs: implement fscache-based data read for non-inline layout
erofs: implement fscache-based data read for inline layout
erofs: register cookie context for data blobs
erofs: implement fscache-based data read for data blobs
erofs: implement fscache-based data readahead
erofs: add 'uuid' mount option
.../filesystems/caching/cachefiles.rst | 159 +++++
fs/cachefiles/Kconfig | 11 +
fs/cachefiles/daemon.c | 576 +++++++++++++++++-
fs/cachefiles/internal.h | 48 ++
fs/cachefiles/io.c | 72 ++-
fs/cachefiles/namei.c | 16 +-
fs/erofs/Makefile | 3 +-
fs/erofs/data.c | 18 +-
fs/erofs/fscache.c | 496 +++++++++++++++
fs/erofs/inode.c | 6 +-
fs/erofs/internal.h | 30 +
fs/erofs/super.c | 106 +++-
fs/fscache/internal.h | 11 -
fs/nfs/fscache.c | 8 -
include/linux/fscache.h | 15 +
include/linux/netfs.h | 1 +
include/trace/events/cachefiles.h | 2 +
include/uapi/linux/cachefiles.h | 48 ++
18 files changed, 1526 insertions(+), 100 deletions(-) create mode 100644 fs/erofs/fscache.c
create mode 100644 include/uapi/linux/cachefiles.h
--
2.27.0
Hi David,
We indeed value the fscache based on-demand read feature, and we believe fscache will benefit more scenarios then. Our community partners are also quite interested in this feature.
Appreciate if you could take a look on it, and please let me know if you have any concern.
Thanks.
Jeffle
Quoted 179 lines · Jeffle Xu
changes since v3:
- cachefiles: The current implementation relies on the anonymous fd mechanism to avoid the dependence on the format of cache file. When cache file is opened for the first time, ananon_fdassociated with the cache file is sent to user daemon. User daemon could fetch and write data to cache file with the givenanon_fd. The following write to theanon_fdwill finally call to cachefiles kernel module, which will write data to cache file in the latest format of cache file. Thus the on-demand read mode can
keep working no matter how cache file format could change in the future. (patch 4)
- cachefiles: the on-demand read mode reuses the existing "/dev/cachefiles" devnode (patch 3)
- erofs: squash several commits implementing readahead into single commit (patch 20)
- erofs: refactor the readahead routine, so that it can read multiple pages each round (patch 20)
- patch 1 and 7 have already been cherry-picked by the maintainers, but have not been merged to the master. Keep them here for completeness.RFC: https://lore.kernel.org/all/[email protected]/t/
v1: https://lore.kernel.org/lkml/[email protected]/T/
v2: https://lore.kernel.org/all/[email protected]/t/
v3: https://lore.kernel.org/lkml/[email protected]/T/[Background]
============
Nydus [1] is a container image distribution service specially optimised for distribution over network. Nydus is an excellent container image acceleration solution, since it only pulls data from remote when it's really needed, a.k.a. on-demand reading.erofs (Enhanced Read-Only File System) is a filesystem specially optimised for read-only scenarios. (
Documentation/filesystem/erofs.rst)Recently we are focusing on erofs in container images distribution scenario [2], trying to combine it with nydus. In this case, erofs can be mounted from one bootstrap file (metadata) with (optional) multiple data blob files (data) stored on another local filesystem. (All these files are actually image files in erofs disk format.)
To accelerate the container startup (fetching container image from remote and then start the container), we do hope that the bootstrap blob file could support demand read. That is, erofs can be mounted and accessed even when the bootstrap/data blob files have not been fully downloaded.
That means we have to manage the cache state of the bootstrap/data blob files (if cache hit, read directly from the local cache; if cache miss, fetch the data somehow). It would be painful and may be dumb for erofs to implement the cache management itself. Thus we prefer fscache/cachefiles to do the cache management. Besides, the demand-read feature shall be general and it can benefit other using scenarios if it can be implemented in fscache level.
[1] https://nydus.dev
[2] https://sched.co/pcdL[Overall Design]
================Please refer to patch 6 ("cachefiles: document on-demand read mode") for more details.
When working in original mode, cachefiles mainly serves as a local cache for remote networking fs, while in on-demand read mode, cachefiles can boost the scenario where on-demand read semantics is needed, e.g. container image distribution.
The essential difference between these two modes is that, in original mode, when cache miss, netfs itself will fetch data from remote, and then write the fetched data into cache file. While in on-demand read mode, a user daemon is responsible for fetching data and then writing to the cache file.
The on-demand read mode relies on a simple protocol used for communication between kernel and user daemon.
The current implementation relies on the anonymous fd mechanism to avoid the dependence on the format of cache file. When cache file is opened for the first time, an
anon_fdassociated with the cache file is sent to user daemon. With the givenanon_fd, user daemon could fetch and write data into the cache file in the background, even when kernel has not triggered the cache miss. Besides, thewrite()syscall to theanon_fdwill finally call cachefiles kernel module, which will write data to cache file in the latest format of cache file.1. cache miss
When cache miss, cachefiles kernel module will notify user daemon theanon_fd, along with the requested file range. When notified, user dameon needs to fetch data of the requested file range, and then write the fetched data into cache file with the given anonymous fd. When finished processing the request, user daemon needs to notify the kernel.After notifying the user daemon, the kernel read routine will hang there, until the request is handled by user daemon. When it's awaken by the notification from user daemon, i.e. the corresponding hole has been filled by the user daemon, it will retry to read from the same file range.
2. cache hit
Once data is already ready in cache file, netfs will read from cache file directly.[Advantage of fscache-based demand-read]
========================================
1. Asynchronous Prefetch
In current mechanism, fscache is responsible for cache state management, while the data plane (fetch data from local/remote on cache miss) is done on the user daemon side.If data has already been ready in the backing file, the upper fs (e.g. erofs) will read from the backing file directly and won't be trapped to user space anymore. Thus the user daemon could fetch data (from remote) asynchronously on the background, and thus accelerate the backing file accessing in some degree.
2. Support massive blob files
Besides this mechanism supports a large amount of backing files, and thus can benefit the densely employed scenario.In our using scenario, one container image can correspond to one bootstrap file (required) and multiple data blob files (optional). For example, one container image for node.js will corresponds to ~20 files in total. In densely employed environment, there could be as many as hundreds of containers and thus thousands of backing files on one machine.
[Test]
==========
You could start a quick test by
https://github.com/lostjeffle/demand-read-cachefilesdJeffle Xu (21):
fscache: exportfscache_end_operation()
cachefiles: export write routine
cachefiles: introduce on-demand read mode
cachefiles: notify user daemon withanon_fdwhen opening cache file
cachefiles: implement on-demand read
cachefiles: document on-demand read mode
erofs: use meta buffers forerofs_read_superblock()
erofs: exporterofs_map_blocks()
erofs: add mode checking helper
erofs: register global fscache volume
erofs: add cookie context helper functions
erofs: add anonymous inode managing page cache of blob file
erofs: adderofs_fscache_read_pages()helper
erofs: register cookie context for bootstrap blob
erofs: implement fscache-based metadata read
erofs: implement fscache-based data read for non-inline layout
erofs: implement fscache-based data read for inline layout
erofs: register cookie context for data blobs
erofs: implement fscache-based data read for data blobs
erofs: implement fscache-based data readahead
erofs: add 'uuid' mount option.../filesystems/caching/cachefiles.rst | 159 +++++ fs/cachefiles/Kconfig | 11 + fs/cachefiles/daemon.c | 576 +++++++++++++++++- fs/cachefiles/internal.h | 48 ++ fs/cachefiles/io.c | 72 ++- fs/cachefiles/namei.c | 16 +- fs/erofs/Makefile | 3 +- fs/erofs/data.c | 18 +- fs/erofs/fscache.c | 496 +++++++++++++++ fs/erofs/inode.c | 6 +- fs/erofs/internal.h | 30 + fs/erofs/super.c | 106 +++- fs/fscache/internal.h | 11 - fs/nfs/fscache.c | 8 - include/linux/fscache.h | 15 + include/linux/netfs.h | 1 + include/trace/events/cachefiles.h | 2 + include/uapi/linux/cachefiles.h | 48 ++ 18 files changed, 1526 insertions(+), 100 deletions(-) create mode 100644 fs/erofs/fscache.c create mode 100644 include/uapi/linux/cachefiles.h
Signature · 2 lines
Thanks, Jeffle