Skip to contentSkip to conversation

fscache,erofs: fscache-based on-demand read semantics

PATCH Superseded Jeffle Xu 3 participants 38 messages Mar 25, 2022 – Mar 29, 2022

Subsystems (7)
Documentation VFS NFS client Tracing +3 more subsystems

v6 of 11 Superseded Latest is v11 →

lore
Share & tools
lore

Jeffle Xu

author
#0 lore raw

v6 cover fscache,erofs: fscache-based on-demand read semantics

changes since v5:
- cachefiles: Move the enabling of on-demand read mode to the end of the cachefiles subset of the patchset (David Howells) (patch 6)
- cachefiles: avoid the duplicate kstrdup() when handling cinit command. Also polish the commist message with the suggestion from David Howells. (David Howells) (patch 3)
- cachefiles: reuse the spinlock inside xarray to prevent the race condition, which also fixes GFP_KERNEL allocation while holding rw_lock (Matthew Wilcox) (patch 3)
- cachefiles: completion of READ request is done through CACHEFILES_IOC_CREAD ioctl on anon_fd (David Howells) (patch 5)
- erofs: rename erofs_bdev_mode() to erofs_is_nodev_mode() (Gao Xiang) (patch 10)
- erofs: expand the existing "struct erofs_map_blocks" rather than create a new "struct erofs_fscache_map" (Gao Xiang) (patch 17)
- erofs: fold functions handling readahead for inline/non-inline/hole into one function, which also omits use of "struct erofs_fscache_ra_ctx" (Gao Xiang) (patch 21)
- erofs: use folio APIs, though there's assumption that folio size equals PAGE_SIZE (Gao Xiang)
- erofs: rename "-o uuid=" mount option to "-o tag=" (Gao Xiang) (patch
22)

Kernel Patchset
---------------
Git tree:

[email protected]:lostjeffle/linux.git jingbo/dev-erofs-fscache-v6

Gitweb:

https://github.com/lostjeffle/linux/commits/jingbo/dev-erofs-fscache-v6

User Daemon for Quick Test
--------------------------
Git tree:

[email protected]:lostjeffle/demand-read-cachefilesd.git main

Gitweb:

https://github.com/lostjeffle/demand-read-cachefilesd

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/
v4: https://lore.kernel.org/lkml/[email protected]/T/#t
v5: https://lore.kernel.org/lkml/[email protected]/T/

[Background]
============
Nydus [1] is an image distribution service especially optimized for distribution over network. Nydus is an excellent container image acceleration solution, since it only pulls data from remote when needed, a.k.a. on-demand reading and it also supports chunk-based deduplication, compression, etc.

erofs (Enhanced Read-Only File System) is a filesystem designed for read-only scenarios. (Documentation/filesystem/erofs.rst)

Over the past months we've been focusing on supporting Nydus image service with in-kernel erofs format[2]. In that case, each container image will be organized in one bootstrap (metadata) and (optional) multiple data blobs in erofs format. Massive container images will be stored on one machine.

To accelerate the container startup (fetching container images from remote and then start the container), we do hope that the bootstrap & blob files could support on-demand read. That is, erofs can be mounted and accessed even when the bootstrap/data blob files have not been fully downloaded. Then it'll have native performance after data is available locally.

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 instead.

The fscache on-demand read feature aims to be implemented in a generic way so that it can benefit other use cases and/or filesystems if it's implemented in the fscache subsystem.

[1] https://nydus.dev
[2] https://sched.co/pcdL

[Overall Design]
================
Please refer to patch 7 ("cachefiles: document on-demand read mode") for more details.

When working in the original mode, cachefiles mainly serves as a local cache for remote networking fs, while in on-demand read mode, cachefiles can work in 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 feeds to the kernel fscache side.

The on-demand read mode relies on a simple protocol used for communication between kernel and user daemon.

The proposed implementation relies on the anonymous fd mechanism to avoid the dependence on the format of cache file. When a fscache cachefile is opened for the first time, an anon_fd associated with the cache file is sent to the 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 with 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 on-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, netfs (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.

Jeffle Xu (22):
fscache: export fscache_end_operation()
cachefiles: extract write routine
cachefiles: notify user daemon with anon_fd when looking up cookie
cachefiles: notify user daemon when withdrawing cookie
cachefiles: implement on-demand read
cachefiles: enable on-demand read mode
cachefiles: document on-demand read mode
erofs: use meta buffers for erofs_read_superblock()
erofs: make erofs_map_blocks() generally available
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_folios() 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 'tag' mount option

 .../filesystems/caching/cachefiles.rst        | 178 +++++++
 fs/cachefiles/Kconfig                         |  11 +
 fs/cachefiles/Makefile                        |   1 +
 fs/cachefiles/daemon.c                        |  89 +++-
 fs/cachefiles/interface.c                     |   2 +
 fs/cachefiles/internal.h                      |  64 +++
 fs/cachefiles/io.c                            |  72 ++-
 fs/cachefiles/namei.c                         |  16 +-
 fs/cachefiles/ondemand.c                      | 456 ++++++++++++++++++
 fs/erofs/Kconfig                              |  10 +
 fs/erofs/Makefile                             |   1 +
 fs/erofs/data.c                               |  24 +-
 fs/erofs/fscache.c                            | 444 +++++++++++++++++
 fs/erofs/inode.c                              |   8 +-
 fs/erofs/internal.h                           |  51 ++
 fs/erofs/super.c                              | 115 ++++-
 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               |  55 +++
 22 files changed, 1544 insertions(+), 90 deletions(-)
 create mode 100644 fs/cachefiles/ondemand.c
 create mode 100644 fs/erofs/fscache.c
 create mode 100644 include/uapi/linux/cachefiles.h
Signature · 1 line
2.27.0
author
#1 lore raw

PATCH v6 1/22 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
index f121c21590dc..ed1c9ed737f2 100644
@@ -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
index cfe901650ab0..39654ca72d3d 100644
@@ -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
index 296c5f1d9f35..d2430da8aa67 100644
@@ -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
author
#2 lore raw

PATCH v6 2/22 cachefiles: extract write routine

In reply to Jeffle Xu

Extract the generic routine of writing data to cache files, and make it generally available.

This will be used by the following patch implementing on-demand read mode. Since it's called inside cachefiles module in this case, 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
index c793d33b0224..e80673d0ab97 100644
@@ -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
author
#3 lore raw

PATCH v6 3/22 cachefiles: notify user daemon with anon_fd when looking up cookie

In reply to Jeffle Xu

Fscache/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.

As the first step, notify user daemon with anon_fd when looking up cookie.

Send the anonymous fd to user daemon when looking up cookie, 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 has no idea the cache file size, whilst user daemon needs to offer the hint on the cache file size.

---
 fs/cachefiles/Kconfig             |  11 +
 fs/cachefiles/Makefile            |   1 +
 fs/cachefiles/daemon.c            |  76 ++++++-
 fs/cachefiles/internal.h          |  44 ++++
 fs/cachefiles/namei.c             |  16 +-
 fs/cachefiles/ondemand.c          | 348 ++++++++++++++++++++++++++++++
 include/linux/fscache.h           |   1 +
 include/trace/events/cachefiles.h |   2 +
 include/uapi/linux/cachefiles.h   |  43 ++++
 9 files changed, 529 insertions(+), 13 deletions(-)
 create mode 100644 fs/cachefiles/ondemand.c
 create mode 100644 include/uapi/linux/cachefiles.h
fs/cachefiles/Kconfig +11 -0
index 719faeeda168..58aad1fb4c5c 100644
@@ -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/Makefile +1 -0
index 16d811f1a2fa..c37a7a9af10b 100644
@@ -16,5 +16,6 @@ cachefiles-y := \
 	xattr.o
 
 cachefiles-$(CONFIG_CACHEFILES_ERROR_INJECTION) += error_inject.o
+cachefiles-$(CONFIG_CACHEFILES_ONDEMAND) += ondemand.o
 
 obj-$(CONFIG_CACHEFILES) := cachefiles.o
fs/cachefiles/daemon.c +65 -11

126 lines not on this page · raw message

fs/cachefiles/internal.h +44 -0

73 lines not on this page · raw message

fs/cachefiles/namei.c +14 -2
index f256c8aff7bb..22aba4c6a762 100644
@@ -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;
fs/cachefiles/ondemand.c new file +348 -0

348 lines not on this page · raw message

include/linux/fscache.h +1 -0
index d2430da8aa67..a330354f33ca 100644
@@ -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
index c6f5aa74db89..371e5816e98c 100644
@@ -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 +43 -0

43 lines not on this page · raw message

Signature · 1 line
2.27.0
#4 lore raw

Re: [PATCH v6 03/22] cachefiles: notify user daemon with anon_fd when looking up cookie

Thank you for the patch! Yet something to improve:

Hi Jeffle,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on trondmy-nfs/linux-next]
[also build test ERROR on rostedt-trace/for-next linus/master v5.17] [cannot apply to xiang-erofs/dev-test dhowells-fs/fscache-next next-20220325] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Jeffle-Xu/fscache-erofs-fscache-based-on-demand-read-semantics/20220325-203555
base:   git://git.linux-nfs.org/projects/trondmy/linux-nfs.git linux-next
config: csky-defconfig (https://download.01.org/0day-ci/archive/20220326/[email protected]/config)
compiler: csky-linux-gcc (GCC) 11.2.0

reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# https://github.com/0day-ci/linux/commit/ec8aa2f84eb47244377e4b822dd77d82ee54714a
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Jeffle-Xu/fscache-erofs-fscache-based-on-demand-read-semantics/20220325-203555 git checkout ec8aa2f84eb47244377e4b822dd77d82ee54714a
# save the config file to linux build tree
mkdir build_dir
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-11.2.0 make.cross O=build_dir ARCH=csky SHELL=/bin/bash

If you fix the issue, kindly add following tag as appropriate

All errors (new ones prefixed by >>):

csky-linux-ld: fs/cachefiles/daemon.o: in function `cachefiles_ondemand_daemon_read':

>> daemon.c:(.text+0x97c): multiple definition of `cachefiles_ondemand_daemon_read'; fs/cachefiles/cache.o:cache.c:(.text+0x18): first defined here

csky-linux-ld: fs/cachefiles/interface.o: in function `cachefiles_ondemand_daemon_read':
interface.c:(.text+0x1ec): multiple definition of `cachefiles_ondemand_daemon_read'; fs/cachefiles/cache.o:cache.c:(.text+0x18): first defined here
csky-linux-ld: fs/cachefiles/io.o: in function `cachefiles_ondemand_daemon_read':
io.c:(.text+0x720): multiple definition of `cachefiles_ondemand_daemon_read'; fs/cachefiles/cache.o:cache.c:(.text+0x18): first defined here
csky-linux-ld: fs/cachefiles/key.o: in function `cachefiles_ondemand_daemon_read':
key.c:(.text+0x0): multiple definition of `cachefiles_ondemand_daemon_read'; fs/cachefiles/cache.o:cache.c:(.text+0x18): first defined here
csky-linux-ld: fs/cachefiles/main.o: in function `cachefiles_ondemand_daemon_read':
main.c:(.text+0x0): multiple definition of `cachefiles_ondemand_daemon_read'; fs/cachefiles/cache.o:cache.c:(.text+0x18): first defined here
csky-linux-ld: fs/cachefiles/namei.o: in function `cachefiles_ondemand_daemon_read':
namei.c:(.text+0xf8): multiple definition of `cachefiles_ondemand_daemon_read'; fs/cachefiles/cache.o:cache.c:(.text+0x18): first defined here
csky-linux-ld: fs/cachefiles/security.o: in function `cachefiles_ondemand_daemon_read':
security.c:(.text+0x24): multiple definition of `cachefiles_ondemand_daemon_read'; fs/cachefiles/cache.o:cache.c:(.text+0x18): first defined here
csky-linux-ld: fs/cachefiles/volume.o: in function `cachefiles_ondemand_daemon_read':
volume.c:(.text+0x0): multiple definition of `cachefiles_ondemand_daemon_read'; fs/cachefiles/cache.o:cache.c:(.text+0x18): first defined here
csky-linux-ld: fs/cachefiles/xattr.o: in function `cachefiles_ondemand_daemon_read':
xattr.c:(.text+0x0): multiple definition of `cachefiles_ondemand_daemon_read'; fs/cachefiles/cache.o:cache.c:(.text+0x18): first defined here

Signature · 2 lines
0-DAY CI Kernel Test Service
https://01.org/lkp
#5 lore raw

Re: [PATCH v6 03/22] cachefiles: notify user daemon with anon_fd when looking up cookie

In reply to Jeffle Xu
Thank you for the patch! Yet something to improve:

Hi Jeffle,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on trondmy-nfs/linux-next]
[also build test ERROR on rostedt-trace/for-next linus/master v5.17] [cannot apply to xiang-erofs/dev-test dhowells-fs/fscache-next next-20220325] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Jeffle-Xu/fscache-erofs-fscache-based-on-demand-read-semantics/20220325-203555
base:   git://git.linux-nfs.org/projects/trondmy/linux-nfs.git linux-next
config: i386-randconfig-a002 (https://download.01.org/0day-ci/archive/20220326/[email protected]/config)
compiler: clang version 15.0.0 (https://github.com/llvm/llvm-project 0f6d9501cf49ce02937099350d08f20c4af86f3d)

reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# https://github.com/0day-ci/linux/commit/ec8aa2f84eb47244377e4b822dd77d82ee54714a
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review Jeffle-Xu/fscache-erofs-fscache-based-on-demand-read-semantics/20220325-203555 git checkout ec8aa2f84eb47244377e4b822dd77d82ee54714a
# save the config file to linux build tree
mkdir build_dir
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=i386 SHELL=/bin/bash fs/

If you fix the issue, kindly add following tag as appropriate

All error/warnings (new ones prefixed by >>):

In file included from fs/cachefiles/cache.c:11:

>> fs/cachefiles/internal.h:285:9: warning: no previous prototype for function 'cachefiles_ondemand_daemon_read' [-Wmissing-prototypes]
   ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache,
           ^

fs/cachefiles/internal.h:285:1: note: declare 'static' if the function is not intended to be used outside of this translation unit ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache,
^
static
1 warning generated.
--

>> ld.lld: error: duplicate symbol: cachefiles_ondemand_daemon_read
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/cache.o:(cachefiles_ondemand_daemon_read)
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/daemon.o:(.text+0x0)
--
>> ld.lld: error: duplicate symbol: cachefiles_ondemand_daemon_read
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/cache.o:(cachefiles_ondemand_daemon_read)
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/interface.o:(.text+0x0)
--
>> ld.lld: error: duplicate symbol: cachefiles_ondemand_daemon_read
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/cache.o:(cachefiles_ondemand_daemon_read)
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/io.o:(.text+0x0)
--
>> ld.lld: error: duplicate symbol: cachefiles_ondemand_daemon_read
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/cache.o:(cachefiles_ondemand_daemon_read)
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/key.o:(.text+0x0)
--
>> ld.lld: error: duplicate symbol: cachefiles_ondemand_daemon_read
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/cache.o:(cachefiles_ondemand_daemon_read)
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/main.o:(.text+0x38C0)
--
>> ld.lld: error: duplicate symbol: cachefiles_ondemand_daemon_read
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/cache.o:(cachefiles_ondemand_daemon_read)
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/namei.o:(.text+0x0)
--
>> ld.lld: error: duplicate symbol: cachefiles_ondemand_daemon_read
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/cache.o:(cachefiles_ondemand_daemon_read)
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/security.o:(.text+0x0)
--
>> ld.lld: error: duplicate symbol: cachefiles_ondemand_daemon_read
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/cache.o:(cachefiles_ondemand_daemon_read)
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/volume.o:(.text+0x0)
--
>> ld.lld: error: duplicate symbol: cachefiles_ondemand_daemon_read
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/cache.o:(cachefiles_ondemand_daemon_read)
   >>> defined at internal.h:287 (fs/cachefiles/internal.h:287)
   >>> fs/cachefiles/xattr.o:(.text+0x0)
Signature · 2 lines
0-DAY CI Kernel Test Service
https://01.org/lkp
author
#6 lore raw

Re: [Linux-cachefs] [PATCH v6 03/22] cachefiles: notify user daemon with anon_fd when looking up cookie

In reply to Jeffle Xu
Quoted 13 lines · Jeffle Xu
fs/cachefiles/internal.h +7 -0
index e80673d0ab97..8a0f1b691aca 100644
@@ -15,6 +15,8 @@
 
+/*
+ * ondemand.c
+ */
+#ifdef CONFIG_CACHEFILES_ONDEMAND
+extern ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache,
+					       char __user *_buffer,
+					       size_t buflen);

fs/cachefiles/internal.h extern struct file *cachefiles_create_tmpfile(struct cachefiles_object *object); · line 286

Quoted 10 lines
fs/cachefiles/internal.h +9 -0
+
+extern int cachefiles_ondemand_cinit(struct cachefiles_cache *cache,
+				     char *args);
+
+extern int cachefiles_ondemand_init_object(struct cachefiles_object *object);
+
+#else
 
+ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache,
+					char __user *_buffer, size_t buflen)

Needs to be declared as static inline ...

Quoted 9 lines
fs/cachefiles/internal.h +9 -0
+{
+	return -EOPNOTSUPP;
+}
+
+static inline int cachefiles_ondemand_init_object(struct cachefiles_object *object)
+{
+	return 0;
+}
+#endif
Signature · 2 lines
Thanks,
Jeffle
author
#7 lore raw

PATCH v6 4/22 cachefiles: notify user daemon when withdrawing cookie

In reply to Jeffle Xu

Notify user daemon that cookie is going to be withdrawn, providing a hint that the associated anon_fd can be closed. The anon_fd attached in the CLOSE request shall be same with that in the previous OPEN request.

Be noted that this is only a hint. User daemon can close the anon_fd when receiving the CLOSE request, then it will receive another anon_fd if the cookie gets looked up. Or it can also ignore the CLOSE request, and keep writing data into the anon_fd. However the next time cookie gets looked up, the user daemon will still receive another anon_fd.

---
 fs/cachefiles/interface.c       |  2 ++
 fs/cachefiles/internal.h        |  3 +++
 fs/cachefiles/ondemand.c        | 27 +++++++++++++++++++++++++++
 include/uapi/linux/cachefiles.h |  5 +++++
 4 files changed, 37 insertions(+)
fs/cachefiles/interface.c +2 -0
index ae93cee9d25d..c5b8fefd4ccc 100644
@@ -322,6 +322,8 @@ static void cachefiles_commit_object(struct cachefiles_object *object,
 static void cachefiles_clean_up_object(struct cachefiles_object *object,
 				       struct cachefiles_cache *cache)
 {
+	cachefiles_ondemand_cleanup_object(object);
+
 	if (test_bit(FSCACHE_COOKIE_RETIRED, &object->cookie->flags)) {
 		if (!test_bit(CACHEFILES_OBJECT_USING_TMPFILE, &object->flags)) {
 			cachefiles_see_object(object, cachefiles_obj_see_clean_delete);
fs/cachefiles/internal.h +3 -0
index 8a0f1b691aca..c80b519a887b 100644
@@ -280,6 +280,7 @@ extern int cachefiles_ondemand_cinit(struct cachefiles_cache *cache,
 				     char *args);
 
 extern int cachefiles_ondemand_init_object(struct cachefiles_object *object);
+extern void cachefiles_ondemand_cleanup_object(struct cachefiles_object *object);
 
 #else
 ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache,
@@ -292,6 +293,8 @@ static inline int cachefiles_ondemand_init_object(struct cachefiles_object *obje
 {
 	return 0;
 }
+
+static inline void cachefiles_ondemand_cleanup_object(struct cachefiles_object *object) {}
 #endif
 
 /*
fs/cachefiles/ondemand.c +27 -0
index 0742c4a7797a..7fd518e01e5a 100644
@@ -199,6 +199,12 @@ ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache,
 		goto err_put_fd;
 	}
 
+	/* CLOSE request has no reply */
+	if (msg->opcode == CACHEFILES_OP_CLOSE) {
+		xa_erase(&cache->reqs, id);
+		complete(&req->done);
+	}
+
 	return n;
 
 err_put_fd:
@@ -322,6 +328,19 @@ static int init_open_req(struct cachefiles_req *req, void *private)
 	return 0;
 }
 
+static int init_close_req(struct cachefiles_req *req, void *private)
+{
+	struct cachefiles_object *object = req->object;
+	struct cachefiles_close *load = (void *)req->msg.data;
+	int fd = object->fd;
+
+	if (WARN_ON_ONCE(fd == -1))
+		return -EIO;
+
+	load->fd = fd;
+	return 0;
+}
+
 int cachefiles_ondemand_init_object(struct cachefiles_object *object)
 {
 	struct fscache_cookie *cookie = object->cookie;
@@ -346,3 +365,11 @@ int cachefiles_ondemand_init_object(struct cachefiles_object *object)
 					    CACHEFILES_OP_OPEN, data_len,
 					    init_open_req, NULL);
 }
+
+void cachefiles_ondemand_cleanup_object(struct cachefiles_object *object)
+{
+	cachefiles_ondemand_send_req(object,
+				     CACHEFILES_OP_CLOSE,
+				     sizeof(struct cachefiles_close),
+				     init_close_req, NULL);
+}
include/uapi/linux/cachefiles.h +5 -0
index 0c44d68be6bd..03047e4b7df2 100644
@@ -12,6 +12,7 @@
 
 enum cachefiles_opcode {
 	CACHEFILES_OP_OPEN,
+	CACHEFILES_OP_CLOSE,
 };
 
 /*
@@ -40,4 +41,8 @@ enum cachefiles_open_flags {
 	CACHEFILES_OPEN_WANT_CACHE_SIZE,
 };
 
+struct cachefiles_close {
+	__u32 fd;
+};
+
 #endif
Signature · 1 line
2.27.0
author
#8 lore raw

PATCH v6 5/22 cachefiles: implement on-demand read

In reply to Jeffle Xu

Implement 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/internal.h        |  7 +++
 fs/cachefiles/io.c              | 11 +++++
 fs/cachefiles/ondemand.c        | 81 +++++++++++++++++++++++++++++++++
 include/linux/netfs.h           |  1 +
 include/uapi/linux/cachefiles.h | 13 ++++++
 5 files changed, 113 insertions(+)
fs/cachefiles/internal.h +7 -0
index c80b519a887b..686f25097681 100644
@@ -281,6 +281,8 @@ extern int cachefiles_ondemand_cinit(struct cachefiles_cache *cache,
 
 extern int cachefiles_ondemand_init_object(struct cachefiles_object *object);
 extern void cachefiles_ondemand_cleanup_object(struct cachefiles_object *object);
+extern int cachefiles_ondemand_read(struct cachefiles_object *object,
+				    loff_t pos, size_t len);
 
 #else
 ssize_t cachefiles_ondemand_daemon_read(struct cachefiles_cache *cache,
@@ -295,6 +297,11 @@ static inline int cachefiles_ondemand_init_object(struct cachefiles_object *obje
 }
 
 static inline void cachefiles_ondemand_cleanup_object(struct cachefiles_object *object) {}
+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
index 8dbc1eb254a3..ee1283ba7a2c 100644
@@ -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;
fs/cachefiles/ondemand.c +81 -0

120 lines not on this page · raw message

include/linux/netfs.h +1 -0
index 614f22213e21..2a9c50d3a928 100644
@@ -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 +13 -0
index 03047e4b7df2..004335d44e16 100644
@@ -3,6 +3,7 @@
 #define _LINUX_CACHEFILES_H
 
 #include <linux/types.h>
+#include <linux/ioctl.h>
 
 /*
  * Fscache ensures that the maximum length of cookie key is 255. The volume key
@@ -13,6 +14,7 @@
 enum cachefiles_opcode {
 	CACHEFILES_OP_OPEN,
 	CACHEFILES_OP_CLOSE,
+	CACHEFILES_OP_READ,
 };
 
 /*
@@ -45,4 +47,15 @@ struct cachefiles_close {
 	__u32 fd;
 };
 
+struct cachefiles_read {
+	__u64 off;
+	__u64 len;
+	__u32 fd;
+};
+
+/*
+ * For CACHEFILES_IOC_CREAD, arg is the @id field of corresponding READ request.
+ */
+#define CACHEFILES_IOC_CREAD	_IOW(0x98, 1, long)
+
 #endif
Signature · 1 line
2.27.0
author
#9 lore raw

PATCH v6 6/22 cachefiles: enable on-demand read mode

In reply to Jeffle Xu

Enable on-demand read mode by adding an optional parameter to the "bind" command.

On-demand mode will be turned on when this parameter is "ondemand", i.e. "bind ondemand". Otherwise cachefiles will work in the original mode.

---
 fs/cachefiles/daemon.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)
fs/cachefiles/daemon.c +8 -5
index 91b88a499737..2c38c5361bda 100644
@@ -741,11 +741,6 @@ static int cachefiles_daemon_bind(struct cachefiles_cache *cache, char *args)
 	    cache->brun_percent  >= 100)
 		return -ERANGE;
 
-	if (*args) {
-		pr_err("'bind' command doesn't take an argument\n");
-		return -EINVAL;
-	}
-
 	if (!cache->rootdirname) {
 		pr_err("No cache directory specified\n");
 		return -EINVAL;
@@ -757,6 +752,14 @@ static int cachefiles_daemon_bind(struct cachefiles_cache *cache, char *args)
 		return -EBUSY;
 	}
 
+	if (IS_ENABLED(CONFIG_CACHEFILES_ONDEMAND) &&
+	    !strcmp(args, "ondemand")) {
+		set_bit(CACHEFILES_ONDEMAND_MODE, &cache->flags);
+	} else if (*args) {
+		pr_err("'bind' command doesn't take an argument\n");
+		return -EINVAL;
+	}
+
 	/* Make sure we have copies of the tag string */
 	if (!cache->tag) {
 		/*
Signature · 1 line
2.27.0
author
#10 lore raw

PATCH v6 7/22 cachefiles: document on-demand read mode

In reply to Jeffle Xu

Document new user interface introduced by on-demand read mode.

---
 .../filesystems/caching/cachefiles.rst        | 178 ++++++++++++++++++
 1 file changed, 178 insertions(+)
Documentation/filesystems/caching/cachefiles.rst +178 -0
index 8bf396b76359..46420fca6677 100644
@@ -28,6 +28,8 @@ Cache on Already Mounted Filesystem
 
  (*) Debugging.
 
+ (*) On-demand Read.
+
 
 
 Overview
@@ -482,3 +484,179 @@ 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.
+
+
+OPEN Request
+------------
+
+When netfs opens a cache file for the first time, a request with
+CACHEFILES_OP_OPEN opcode, a.k.a OPEN request will be sent to user daemon. The
+payload format is like::
+
+	struct cachefiles_open {
+		__u32 volume_key_len;
+		__u32 cookie_key_len;
+		__u32 fd;
+		__u32 flags;
+		__u8  data[];
+	};
+
+	* ``data`` contains volume_key and cookie_key in sequence.
+
+	* ``volume_key_len`` identifies the length of the volume key of the
+	  cache file, in bytes. volume_key is of string format, with a suffix
+	  '\0'.
+
+	* ``cookie_key_len`` identifies the length of the cookie key of the
+	  cache file, in bytes. The format of cookie_key is netfs specific. It
+	  can be of binary format.
+
+	* ``fd`` identifies the anonymous fd of the cache file, with which user
+	  daemon can perform write/llseek file operations on the cache file.
+
+
+OPEN 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 OPEN request.
+
+
+Besides, CACHEFILES_OPEN_WANT_CACHE_SIZE flag may be set in flags field of
+OPEN 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 has no idea the cache file size, whilst user daemon needs to offer the
+hint on the cache file size.
+
+Thus when receiving an OPEN request with CACHEFILES_OPEN_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 OPEN request.
+
+	* ``cache_size`` identifies the size of the cache file.
+
+
+CLOSE Request
+-------------
+When cookie withdrawed, a request with CACHEFILES_OP_CLOSE opcode, a.k.a CLOSE
+request, will be sent to user daemon. It will notify user daemon to close the
+attached anon_fd. The payload format is like::
+
+	struct cachefiles_close {
+		__u32 fd;
+	};
+
+	* ``fd`` identifies the anon_fd to be closed, which is exactly the same
+	  with that in OPEN request.
+
+
+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
+	  OPEN 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
+CACHEFILES_IOC_CREAD ioctl on the corresponding anon_fd::
+
+	ioctl(fd, CACHEFILES_IOC_CREAD, id);
+
+	* ``fd`` is exactly the fd field of the previous READ request.
+
+	* ``id`` is exactly the id field of the previous READ request.
Signature · 1 line
2.27.0
author
#11 lore raw

PATCH v6 8/22 erofs: use meta buffers for erofs_read_superblock()

In reply to Jeffle Xu

The 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
index 915eefe0d7e2..12755217631f 100644
@@ -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
author
#12 lore raw

PATCH v6 9/22 erofs: make erofs_map_blocks() generally available

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
index 226a57c57ee6..6e2a28242453 100644
@@ -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
index 5aa2cf2c2f80..e424293f47a2 100644
@@ -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
author
#13 lore raw

PATCH v6 10/22 erofs: add mode checking helper

In reply to Jeffle Xu

Until 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
index e424293f47a2..1486e2573667 100644
@@ -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_is_nodev_mode(struct super_block *sb)
+{
+	return !sb->s_bdev;
+}
+
 enum {
 	EROFS_ZIP_CACHE_DISABLED,
 	EROFS_ZIP_CACHE_READAHEAD,
Signature · 1 line
2.27.0
#14 lore raw

Re: [PATCH v6 10/22] erofs: add mode checking helper

Quoted 19 lines · Jeffle Xu

Until 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 +0 -0
index e424293f47a2..1486e2573667 100644

fs/erofs/internal.h struct erofs_sb_info { · line 164

Quoted 4 lines
fs/erofs/internal.h +1 -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_is_nodev_mode(struct super_block *sb)

I've seen a lot of such

+ if (IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) &&
+ erofs_is_nodev_mode(sb)) {

usages in the followup patches, which makes me wonder if the configuration can be checked in the helper as well. Also maybe rename it as erofs_is_fscache_mode()?

Thanks,
Gao Xiang

Quoted 9 lines
fs/erofs/internal.h +4 -0
+{
+	return !sb->s_bdev;
+}
+
 enum {
 	EROFS_ZIP_CACHE_DISABLED,
 	EROFS_ZIP_CACHE_READAHEAD,
Signature · 1 line
2.27.0
author
#15 lore raw

Re: [PATCH v6 10/22] erofs: add mode checking helper

Quoted 30 lines · Gao Xiang
Quoted 19 lines · Jeffle Xu

Until 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 +0 -0
index e424293f47a2..1486e2573667 100644

fs/erofs/internal.h struct erofs_sb_info { · line 164

Quoted 4 lines
fs/erofs/internal.h +1 -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_is_nodev_mode(struct super_block *sb)

I've seen a lot of such

(quoted diff) +2 -0
+		if (IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) &&
+		    erofs_is_nodev_mode(sb)) {
Quoted 4 lines

usages in the followup patches, which makes me wonder if the configuration can be checked in the helper as well. Also maybe rename it as erofs_is_fscache_mode()?

Sure. Will be done in the next version.

Signature · 2 lines
Thanks,
Jeffle
author
#16 lore raw

PATCH v6 11/22 erofs: register global fscache volume

In reply to Jeffle Xu

All 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. 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/Kconfig    | 10 ++++++++++
 fs/erofs/Makefile   |  1 +
 fs/erofs/fscache.c  | 22 ++++++++++++++++++++++
 fs/erofs/internal.h |  9 +++++++++
 fs/erofs/super.c    |  7 +++++++
 5 files changed, 49 insertions(+)
 create mode 100644 fs/erofs/fscache.c
fs/erofs/Kconfig +10 -0
index f57255ab88ed..a6f5d61bdc49 100644
@@ -98,3 +98,13 @@ config EROFS_FS_ZIP_LZMA
 	  systems will be readable without selecting this option.
 
 	  If unsure, say N.
+
+config EROFS_FS_ONDEMAND
+	bool "EROFS fscache-based ondeand-read"
+	depends on CACHEFILES_ONDEMAND && (EROFS_FS=m && FSCACHE && CACHEFILES || EROFS_FS=y && FSCACHE=y && CACHEFILES=y)
+	default n
+	help
+	  EROFS is mounted from data blob files and implement on-demand read
+	  semantics through fscache.
+
+	  If unsure, say N.
fs/erofs/Makefile +1 -0
index 8a3317e38e5a..99bbc597a3e9 100644
@@ -5,3 +5,4 @@ erofs-objs := super.o inode.o data.o namei.o dir.o utils.o pcpubuf.o sysfs.o
 erofs-$(CONFIG_EROFS_FS_XATTR) += xattr.o
 erofs-$(CONFIG_EROFS_FS_ZIP) += decompressor.o zmap.o zdata.o
 erofs-$(CONFIG_EROFS_FS_ZIP_LZMA) += decompressor_lzma.o
+erofs-$(CONFIG_EROFS_FS_ONDEMAND) += fscache.o
fs/erofs/fscache.c new file +22 -0
new file mode 100644
index 000000000000..08cf570a0810
@@ -0,0 +1,22 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (C) 2022, Alibaba Cloud
+ */
+#include <linux/fscache.h>
+#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 +9 -0
index 1486e2573667..45b8b0dd8a27 100644
@@ -616,6 +616,15 @@ static inline int z_erofs_load_lzma_config(struct super_block *sb,
 }
 #endif	/* !CONFIG_EROFS_FS_ZIP */
 
+/* fscache.c */
+#ifdef CONFIG_EROFS_FS_ONDEMAND
+int erofs_init_fscache(void);
+void erofs_exit_fscache(void);
+#else
+static inline int erofs_init_fscache(void) { return 0; }
+static inline void erofs_exit_fscache(void) {}
+#endif
+
 #define EFSCORRUPTED    EUCLEAN         /* Filesystem is corrupted */
 
 #endif	/* __EROFS_INTERNAL_H */
fs/erofs/super.c +7 -0
index 12755217631f..798f0c379e35 100644
@@ -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
author
#17 lore raw

PATCH v6 12/22 erofs: add cookie context helper functions

In reply to Jeffle Xu

Introduce "struct erofs_fscache" for managing cookie for backinig file, and helper functions for initializing and cleaning up this context structure.

---
 fs/erofs/fscache.c  | 61 +++++++++++++++++++++++++++++++++++++++++++++
 fs/erofs/internal.h | 14 +++++++++++
 2 files changed, 75 insertions(+)
fs/erofs/fscache.c +61 -0
index 08cf570a0810..73235fd43bf6 100644
@@ -7,6 +7,67 @@
 
 static struct fscache_volume *volume;
 
+static int erofs_fscache_init_cookie(struct erofs_fscache *ctx, char *path)
+{
+	struct fscache_cookie *cookie;
+
+	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 *ctx)
+{
+	struct fscache_cookie *cookie = ctx->cookie;
+
+	fscache_unuse_cookie(cookie, NULL, NULL);
+	fscache_relinquish_cookie(cookie, false);
+	ctx->cookie = NULL;
+}
+
+/*
+ * erofs_fscache_get - create an fscache context for blob file
+ * @sb:		superblock
+ * @path:	name of blob file
+ *
+ * Return: fscache context on success, ERR_PTR() on failure.
+ */
+struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path)
+{
+	struct erofs_fscache *ctx;
+	int ret;
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return ERR_PTR(-ENOMEM);
+
+	ret = erofs_fscache_init_cookie(ctx, path);
+	if (ret) {
+		erofs_err(sb, "failed to init cookie");
+		goto err;
+	}
+
+	return ctx;
+err:
+	kfree(ctx);
+	return ERR_PTR(ret);
+}
+
+void erofs_fscache_put(struct erofs_fscache *ctx)
+{
+	if (!ctx)
+		return;
+
+	erofs_fscache_cleanup_cookie(ctx);
+	kfree(ctx);
+}
+
 int __init erofs_init_fscache(void)
 {
 	volume = fscache_acquire_volume("erofs", NULL, NULL, 0);
fs/erofs/internal.h +14 -0
index 45b8b0dd8a27..d4f2b43cedae 100644
@@ -96,6 +96,10 @@ struct erofs_sb_lz4_info {
 	u16 max_pclusterblks;
 };
 
+struct erofs_fscache {
+	struct fscache_cookie *cookie;
+};
+
 struct erofs_sb_info {
 	struct erofs_mount_opts opt;	/* options */
 #ifdef CONFIG_EROFS_FS_ZIP
@@ -620,9 +624,19 @@ static inline int z_erofs_load_lzma_config(struct super_block *sb,
 #ifdef CONFIG_EROFS_FS_ONDEMAND
 int erofs_init_fscache(void);
 void erofs_exit_fscache(void);
+
+struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path);
+void erofs_fscache_put(struct erofs_fscache *ctx);
 #else
 static inline int erofs_init_fscache(void) { return 0; }
 static inline void erofs_exit_fscache(void) {}
+
+static inline struct erofs_fscache *erofs_fscache_get(struct super_block *sb,
+						      char *path)
+{
+	return ERR_PTR(-EOPNOTSUPP);
+}
+static inline void erofs_fscache_put(struct erofs_fscache *ctx) {}
 #endif
 
 #define EFSCORRUPTED    EUCLEAN         /* Filesystem is corrupted */
Signature · 1 line
2.27.0
#18 lore raw

Re: [PATCH v6 12/22] erofs: add cookie context helper functions

Hi Jeffle,

Quoted 15 lines · Jeffle Xu

Introduce "struct erofs_fscache" for managing cookie for backinig file, and helper functions for initializing and cleaning up this context structure.

---
 fs/erofs/fscache.c  | 61 +++++++++++++++++++++++++++++++++++++++++++++
 fs/erofs/internal.h | 14 +++++++++++
 2 files changed, 75 insertions(+)
fs/erofs/fscache.c +0 -0
index 08cf570a0810..73235fd43bf6 100644

fs/erofs/fscache.c line 16

Quoted 10 lines
fs/erofs/fscache.c +7 -0
@@ -7,6 +7,67 @@
 
 static struct fscache_volume *volume;
 
+static int erofs_fscache_init_cookie(struct erofs_fscache *ctx, char *path)
+{
+	struct fscache_cookie *cookie;
+
+	cookie = fscache_acquire_cookie(volume, FSCACHE_ADV_WANT_CACHE_SIZE,
+					path, strlen(path),
+					NULL, 0, 0);

It'd be better to rearrange in one line?

					path, strlen(path), NULL, 0, 0);
Quoted 15 lines
fs/erofs/fscache.c +15 -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 *ctx)
+{
+	struct fscache_cookie *cookie = ctx->cookie;
+
+	fscache_unuse_cookie(cookie, NULL, NULL);
+	fscache_relinquish_cookie(cookie, false);
+	ctx->cookie = NULL;

fs/erofs/fscache.c line 41

Quoted 10 lines
fs/erofs/fscache.c +10 -0
+}
+
+/*
+ * erofs_fscache_get - create an fscache context for blob file
+ * @sb:		superblock
+ * @path:	name of blob file
+ *
+ * Return: fscache context on success, ERR_PTR() on failure.
+ */
+struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path)

erofs_fscache_register?

fs/erofs/fscache.c line 50

Quoted 9 lines
fs/erofs/fscache.c +9 -0
+{
+	struct erofs_fscache *ctx;
+	int ret;
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return ERR_PTR(-ENOMEM);
+
+	ret = erofs_fscache_init_cookie(ctx, path);

Can we fold it here? No need to introduce such simple wrapper..

fs/erofs/fscache.c line 52

Quoted 2 lines
fs/erofs/fscache.c +2 -0
+	if (ret) {
+		erofs_err(sb, "failed to init cookie");

It would be better to print the path?

fs/erofs/fscache.c line 53

Quoted 1 line
fs/erofs/fscache.c +1 -0
+		goto err;
		kfree(ctx);
		return ERR_PTR(ret);

At least for now.

fs/erofs/fscache.c line 62

Quoted 9 lines
fs/erofs/fscache.c +9 -0
+	}
+
+	return ctx;
+err:
+	kfree(ctx);
+	return ERR_PTR(ret);
+}
+
+void erofs_fscache_put(struct erofs_fscache *ctx)

erofs_fscache_unregister?

fs/erofs/fscache.c line 67

Quoted 5 lines
fs/erofs/fscache.c +5 -0
+{
+	if (!ctx)
+		return;
+
+	erofs_fscache_cleanup_cookie(ctx);

Fold it too, since such helper doesn't simplify code a lot but need to take one more time to redirect..

Thanks,
Gao Xiang

Quoted 43 lines
fs/erofs/fscache.c +3 -0
+	kfree(ctx);
+}
+
 int __init erofs_init_fscache(void)
 {
 	volume = fscache_acquire_volume("erofs", NULL, NULL, 0);
fs/erofs/internal.h +14 -0
index 45b8b0dd8a27..d4f2b43cedae 100644
@@ -96,6 +96,10 @@ struct erofs_sb_lz4_info {
 	u16 max_pclusterblks;
 };
 
+struct erofs_fscache {
+	struct fscache_cookie *cookie;
+};
+
 struct erofs_sb_info {
 	struct erofs_mount_opts opt;	/* options */
 #ifdef CONFIG_EROFS_FS_ZIP
@@ -620,9 +624,19 @@ static inline int z_erofs_load_lzma_config(struct super_block *sb,
 #ifdef CONFIG_EROFS_FS_ONDEMAND
 int erofs_init_fscache(void);
 void erofs_exit_fscache(void);
+
+struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path);
+void erofs_fscache_put(struct erofs_fscache *ctx);
 #else
 static inline int erofs_init_fscache(void) { return 0; }
 static inline void erofs_exit_fscache(void) {}
+
+static inline struct erofs_fscache *erofs_fscache_get(struct super_block *sb,
+						      char *path)
+{
+	return ERR_PTR(-EOPNOTSUPP);
+}
+static inline void erofs_fscache_put(struct erofs_fscache *ctx) {}
 #endif
 
 #define EFSCORRUPTED    EUCLEAN         /* Filesystem is corrupted */
Signature · 1 line
2.27.0
author
#19 lore raw

Re: [PATCH v6 12/22] erofs: add cookie context helper functions

Quoted 21 lines · Gao Xiang

Hi Jeffle,

Quoted 18 lines · Jeffle Xu

Introduce "struct erofs_fscache" for managing cookie for backinig file, and helper functions for initializing and cleaning up this context structure.

---
 fs/erofs/fscache.c  | 61 +++++++++++++++++++++++++++++++++++++++++++++
 fs/erofs/internal.h | 14 +++++++++++
 2 files changed, 75 insertions(+)
fs/erofs/fscache.c +0 -0
index 08cf570a0810..73235fd43bf6 100644
@@ -7,6 +7,67 @@
 
 static struct fscache_volume *volume;
 
Quoted 9 lines

fs/erofs/fscache.c line 16

Quoted 7 lines
fs/erofs/fscache.c +7 -0
+static int erofs_fscache_init_cookie(struct erofs_fscache *ctx, char *path)
+{
+	struct fscache_cookie *cookie;
+
+	cookie = fscache_acquire_cookie(volume, FSCACHE_ADV_WANT_CACHE_SIZE,
+					path, strlen(path),
+					NULL, 0, 0);

It'd be better to rearrange in one line?

Sure.

Quoted 22 lines
					path, strlen(path), NULL, 0, 0);
Quoted 19 lines
fs/erofs/fscache.c +19 -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 *ctx)
+{
+	struct fscache_cookie *cookie = ctx->cookie;
+
+	fscache_unuse_cookie(cookie, NULL, NULL);
+	fscache_relinquish_cookie(cookie, false);
+	ctx->cookie = NULL;
+}
+
+/*
+ * erofs_fscache_get - create an fscache context for blob file
Quoted 8 lines

fs/erofs/fscache.c line 41

Quoted 6 lines
fs/erofs/fscache.c +6 -0
+ * @sb:		superblock
+ * @path:	name of blob file
+ *
+ * Return: fscache context on success, ERR_PTR() on failure.
+ */
+struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path)

erofs_fscache_register?

OK.

Quoted 13 lines

fs/erofs/fscache.c line 50

Quoted 10 lines
fs/erofs/fscache.c +9 -0
+{
+	struct erofs_fscache *ctx;
+	int ret;
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return ERR_PTR(-ENOMEM);
+
+	ret = erofs_fscache_init_cookie(ctx, path);

Can we fold it here? No need to introduce such simple wrapper..

Quoted 4 lines

fs/erofs/fscache.c line 52

Quoted 2 lines
fs/erofs/fscache.c +2 -0
+	if (ret) {
+		erofs_err(sb, "failed to init cookie");

It would be better to print the path?

OK.

Quoted 7 lines

fs/erofs/fscache.c line 53

Quoted 2 lines
fs/erofs/fscache.c +1 -0
+		goto err;
		kfree(ctx);
		return ERR_PTR(ret);

At least for now.

Yeah, it's better.

Quoted 12 lines

fs/erofs/fscache.c line 62

Quoted 10 lines
fs/erofs/fscache.c +9 -0
+	}
+
+	return ctx;
+err:
+	kfree(ctx);
+	return ERR_PTR(ret);
+}
+
+void erofs_fscache_put(struct erofs_fscache *ctx)

erofs_fscache_unregister?

OK.

Quoted 9 lines

fs/erofs/fscache.c line 67

Quoted 6 lines
fs/erofs/fscache.c +5 -0
+{
+	if (!ctx)
+		return;
+
+	erofs_fscache_cleanup_cookie(ctx);

Fold it too, since such helper doesn't simplify code a lot but need to take one more time to redirect..

OK.

Signature · 2 lines
Thanks,
Jeffle
author
#20 lore raw

PATCH v6 13/22 erofs: add anonymous inode managing page cache of blob file

In reply to Jeffle Xu

Introduce 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  | 41 ++++++++++++++++++++++++++++++++++++++++-
 fs/erofs/internal.h |  7 +++++--
 2 files changed, 45 insertions(+), 3 deletions(-)
fs/erofs/fscache.c +40 -1
index 73235fd43bf6..30383d9adb62 100644
@@ -7,6 +7,9 @@
 
 static struct fscache_volume *volume;
 
+static const struct address_space_operations erofs_fscache_blob_aops = {
+};
+
 static int erofs_fscache_init_cookie(struct erofs_fscache *ctx, char *path)
 {
 	struct fscache_cookie *cookie;
@@ -31,6 +34,29 @@ static inline void erofs_fscache_cleanup_cookie(struct erofs_fscache *ctx)
 	ctx->cookie = NULL;
 }
 
+static int erofs_fscache_get_inode(struct erofs_fscache *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);
+
+	ctx->inode = inode;
+	return 0;
+}
+
+static inline void erofs_fscache_put_inode(struct erofs_fscache *ctx)
+{
+	iput(ctx->inode);
+	ctx->inode = NULL;
+}
+
 /*
  * erofs_fscache_get - create an fscache context for blob file
  * @sb:		superblock
@@ -38,7 +64,8 @@ static inline void erofs_fscache_cleanup_cookie(struct erofs_fscache *ctx)
  *
  * Return: fscache context on success, ERR_PTR() on failure.
  */
-struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path)
+struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path,
+					bool need_inode)
 {
 	struct erofs_fscache *ctx;
 	int ret;
@@ -53,7 +80,18 @@ struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path)
 		goto err;
 	}
 
+	if (need_inode) {
+		ret = erofs_fscache_get_inode(ctx, sb);
+		if (ret) {
+			erofs_err(sb, "failed to get anonymous inode");
+			goto err_cookie;
+		}
+	}
+
 	return ctx;
+
+err_cookie:
+	erofs_fscache_cleanup_cookie(ctx);
 err:
 	kfree(ctx);
 	return ERR_PTR(ret);
@@ -65,6 +103,7 @@ void erofs_fscache_put(struct erofs_fscache *ctx)
 		return;
 
 	erofs_fscache_cleanup_cookie(ctx);
+	erofs_fscache_put_inode(ctx);
 	kfree(ctx);
 }
 
fs/erofs/internal.h +5 -2
index d4f2b43cedae..459f31803c3b 100644
@@ -98,6 +98,7 @@ struct erofs_sb_lz4_info {
 
 struct erofs_fscache {
 	struct fscache_cookie *cookie;
+	struct inode *inode;
 };
 
 struct erofs_sb_info {
@@ -625,14 +626,16 @@ 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 *erofs_fscache_get(struct super_block *sb, char *path);
+struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path,
+					bool need_inode);
 void erofs_fscache_put(struct erofs_fscache *ctx);
 #else
 static inline int erofs_init_fscache(void) { return 0; }
 static inline void erofs_exit_fscache(void) {}
 
 static inline struct erofs_fscache *erofs_fscache_get(struct super_block *sb,
-						      char *path)
+						      char *path,
+						      bool need_inode)
 {
 	return ERR_PTR(-EOPNOTSUPP);
 }
Signature · 1 line
2.27.0
#21 lore raw

Re: [Linux-cachefs] [PATCH v6 13/22] erofs: add anonymous inode managing page cache of blob file

Quoted 25 lines · Jeffle Xu

Introduce 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  | 41 ++++++++++++++++++++++++++++++++++++++++-
 fs/erofs/internal.h |  7 +++++--
 2 files changed, 45 insertions(+), 3 deletions(-)
fs/erofs/fscache.c +3 -0
index 73235fd43bf6..30383d9adb62 100644
@@ -7,6 +7,9 @@
 
 static struct fscache_volume *volume;
 
+static const struct address_space_operations erofs_fscache_blob_aops = {
+};
+
 static int erofs_fscache_init_cookie(struct erofs_fscache *ctx, char *path)
 {
 	struct fscache_cookie *cookie;

fs/erofs/fscache.c static inline void erofs_fscache_cleanup_cookie(struct erofs_fscache *ctx) · line 38

Quoted 5 lines
fs/erofs/fscache.c +2 -0
@@ -31,6 +34,29 @@ static inline void erofs_fscache_cleanup_cookie(struct erofs_fscache *ctx)
 	ctx->cookie = NULL;
 }
 
+static int erofs_fscache_get_inode(struct erofs_fscache *ctx,
+				   struct super_block *sb)

I think it can be folded as well.

Quoted 6 lines
fs/erofs/fscache.c +6 -0
+{
+	struct inode *const inode = new_inode(sb);
+
+	if (!inode)
+		return -ENOMEM;
+

fs/erofs/fscache.c static inline void erofs_fscache_cleanup_cookie(struct erofs_fscache *ctx) · line 54

Quoted 10 lines
fs/erofs/fscache.c +10 -0
+	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);
+
+	ctx->inode = inode;
+	return 0;
+}
+
+static inline void erofs_fscache_put_inode(struct erofs_fscache *ctx)

Ditto.

Quoted 19 lines
fs/erofs/fscache.c +7 -1
+{
+	iput(ctx->inode);
+	ctx->inode = NULL;
+}
+
 /*
  * erofs_fscache_get - create an fscache context for blob file
  * @sb:		superblock
@@ -38,7 +64,8 @@ static inline void erofs_fscache_cleanup_cookie(struct erofs_fscache *ctx)
  *
  * Return: fscache context on success, ERR_PTR() on failure.
  */
-struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path)
+struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path,
+					bool need_inode)
 {
 	struct erofs_fscache *ctx;
 	int ret;

fs/erofs/fscache.c struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path) · line 86

Quoted 7 lines
fs/erofs/fscache.c +4 -0
@@ -53,7 +80,18 @@ struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path)
 		goto err;
 	}
 
+	if (need_inode) {
+		ret = erofs_fscache_get_inode(ctx, sb);
+		if (ret) {
+			erofs_err(sb, "failed to get anonymous inode");

failed to get fscache inode of [path].

Thanks,
Gao Xiang

author
#22 lore raw

PATCH v6 14/22 erofs: add erofs_fscache_read_folios() helper

In reply to Jeffle Xu

Add erofs_fscache_read_folios() 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_folio() wrapper helper.

---
 fs/erofs/fscache.c | 39 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 39 insertions(+)
fs/erofs/fscache.c +39 -0
index 30383d9adb62..6a55f7b5f883 100644
@@ -7,6 +7,45 @@
 
 static struct fscache_volume *volume;
 
+/*
+ * erofs_fscache_read_folios - Read data from fscache.
+ *
+ * Fill the read 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_folios(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_folio(struct fscache_cookie *cookie,
+					   struct folio *folio, loff_t pstart)
+{
+	return erofs_fscache_read_folios(cookie, folio_file_mapping(folio),
+					 folio_pos(folio), folio_size(folio),
+					 pstart);
+}
+
 static const struct address_space_operations erofs_fscache_blob_aops = {
 };
 
Signature · 1 line
2.27.0
#23 lore raw

Re: [Linux-cachefs] [PATCH v6 14/22] erofs: add erofs_fscache_read_folios() helper

Quoted 46 lines · Jeffle Xu

Add erofs_fscache_read_folios() 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_folio() wrapper helper.

---
 fs/erofs/fscache.c | 39 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 39 insertions(+)
fs/erofs/fscache.c +23 -0
index 30383d9adb62..6a55f7b5f883 100644
@@ -7,6 +7,45 @@
 
 static struct fscache_volume *volume;
 
+/*
+ * erofs_fscache_read_folios - Read data from fscache.
+ *
+ * Fill the read 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_folios(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);

fs/erofs/fscache.c line 42

Quoted 10 lines
fs/erofs/fscache.c +10 -0
+
+	ret = fscache_read(&cres, pstart, &iter,
+			   NETFS_READ_HOLE_ONDEMAND, NULL, NULL);
+
+	fscache_end_operation(&cres);
+	return ret;
+}
+
+static inline int erofs_fscache_read_folio(struct fscache_cookie *cookie,
+					   struct folio *folio, loff_t pstart)

I've seen two users of this helper,

1) erofs_fscache_readpage_noinline
2) erofs_fscache_readpage_blob

Actually, we could fold such logic in these users since this additional helper doesn't enhance the readability but increase the overhead when reading the code.

It'd be better to get rid of it instead.

Thanks,
Gao Xiang

Quoted 16 lines
fs/erofs/fscache.c +6 -0
+{
+	return erofs_fscache_read_folios(cookie, folio_file_mapping(folio),
+					 folio_pos(folio), folio_size(folio),
+					 pstart);
+}
+
 static const struct address_space_operations erofs_fscache_blob_aops = {
 };
 
Signature · 6 lines
2.27.0

--
Linux-cachefs mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/linux-cachefs
author
#24 lore raw

PATCH v6 15/22 erofs: register cookie context for bootstrap blob

In reply to Jeffle Xu

Registers 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 inside .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    | 17 +++++++++++++++++
 2 files changed, 20 insertions(+)
fs/erofs/internal.h +3 -0
index 459f31803c3b..d8c886a7491e 100644
@@ -73,6 +73,7 @@ struct erofs_mount_opts {
 	/* threshold for decompression synchronously */
 	unsigned int max_sync_decompress_pages;
 #endif
+	char *tag;
 	unsigned int mount_opt;
 };
 
@@ -151,6 +152,8 @@ struct erofs_sb_info {
 	/* sysfs support */
 	struct kobject s_kobj;		/* /sys/fs/erofs/<devname> */
 	struct completion s_kobj_unregister;
+
+	struct erofs_fscache *bootstrap;
 };
 
 #define EROFS_SB(sb) ((struct erofs_sb_info *)(sb)->s_fs_info)
fs/erofs/super.c +17 -0
index 798f0c379e35..de5aeda4aea0 100644
@@ -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 (IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) && erofs_is_nodev_mode(sb)) {
+		struct erofs_fscache *bootstrap;
+
+		bootstrap = erofs_fscache_get(sb, ctx->opt.tag, 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(sbi->bootstrap);
 	fs_put_dax(sbi->dax_dev);
 	kfree(sbi);
 	sb->s_fs_info = NULL;
@@ -771,6 +782,12 @@ static void erofs_put_super(struct super_block *sb)
 	iput(sbi->managed_cache);
 	sbi->managed_cache = NULL;
 #endif
+	erofs_fscache_put(sbi->bootstrap);
+	/*
+	 * Set sbi->bootstrap to NULL, so that the following cleanup routine
+	 * inside .kill_sb() could be skipped then.
+	 */
+	sbi->bootstrap = NULL;
 }
 
 static struct file_system_type erofs_fs_type = {
Signature · 1 line
2.27.0
#25 lore raw

Re: [PATCH v6 15/22] erofs: register cookie context for bootstrap blob

Quoted 33 lines · Jeffle Xu

Registers 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 inside .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    | 17 +++++++++++++++++
 2 files changed, 20 insertions(+)
fs/erofs/internal.h +1 -0
index 459f31803c3b..d8c886a7491e 100644
@@ -73,6 +73,7 @@ struct erofs_mount_opts {
 	/* threshold for decompression synchronously */
 	unsigned int max_sync_decompress_pages;
 #endif
+	char *tag;
 	unsigned int mount_opt;
 };
 

fs/erofs/internal.h struct erofs_sb_info { · line 156

Quoted 5 lines
fs/erofs/internal.h +2 -0
@@ -151,6 +152,8 @@ struct erofs_sb_info {
 	/* sysfs support */
 	struct kobject s_kobj;		/* /sys/fs/erofs/<devname> */
 	struct completion s_kobj_unregister;
+
+	struct erofs_fscache *bootstrap;

the concept of bootstrap is nydus-specific. Actually here we need a fscache context of the primary device.

So I prefer struct erofs_fscache *s_fscache;

Also please help revise the subject and commit message about bootstrap.

Thanks,
Gao Xiang

author
#26 lore raw

Re: [PATCH v6 15/22] erofs: register cookie context for bootstrap blob

Quoted 43 lines · Gao Xiang
Quoted 33 lines · Jeffle Xu

Registers 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 inside .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    | 17 +++++++++++++++++
 2 files changed, 20 insertions(+)
fs/erofs/internal.h +1 -0
index 459f31803c3b..d8c886a7491e 100644
@@ -73,6 +73,7 @@ struct erofs_mount_opts {
 	/* threshold for decompression synchronously */
 	unsigned int max_sync_decompress_pages;
 #endif
+	char *tag;
 	unsigned int mount_opt;
 };
 

fs/erofs/internal.h struct erofs_sb_info { · line 156

Quoted 5 lines
fs/erofs/internal.h +2 -0
@@ -151,6 +152,8 @@ struct erofs_sb_info {
 	/* sysfs support */
 	struct kobject s_kobj;		/* /sys/fs/erofs/<devname> */
 	struct completion s_kobj_unregister;
+
+	struct erofs_fscache *bootstrap;

the concept of bootstrap is nydus-specific. Actually here we need a fscache context of the primary device.

Quoted 5 lines

So I prefer struct erofs_fscache *s_fscache;

Also please help revise the subject and commit message about bootstrap.

OK, will be done in the next version.

Signature · 2 lines
Thanks,
Jeffle
author
#27 lore raw

PATCH v6 16/22 erofs: implement fscache-based metadata read

In reply to Jeffle Xu

Implements the data plane of reading metadata from bootstrap blob file over fscache.

---
 fs/erofs/data.c     | 17 +++++++++++++++--
 fs/erofs/fscache.c  | 34 ++++++++++++++++++++++++++++++++++
 fs/erofs/internal.h |  8 ++++++++
 3 files changed, 57 insertions(+), 2 deletions(-)
fs/erofs/data.c +15 -2
index 6e2a28242453..b4571bea93d5 100644
@@ -31,15 +31,28 @@ 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 (IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) &&
+		    erofs_is_nodev_mode(sb)) {
+			struct folio *folio;
+
+			folio = erofs_fscache_get_folio(sbi->bootstrap, index);
+			if (IS_ERR(folio))
+				page = (struct page *)folio;
+			else
+				page = folio_page(folio, 0);
+		} else {
+			mapping = sb->s_bdev->bd_inode->i_mapping;
+			page = read_cache_page_gfp(mapping, index,
 				mapping_gfp_constraint(mapping, ~__GFP_FS));
+		}
 		if (IS_ERR(page))
 			return page;
 		/* should already be PageUptodate, no need to lock page */
fs/erofs/fscache.c +34 -0
index 6a55f7b5f883..91377939b4f7 100644
@@ -46,9 +46,43 @@ static inline int erofs_fscache_read_folio(struct fscache_cookie *cookie,
 					 pstart);
 }
 
+static int erofs_fscache_readpage_blob(struct file *data, struct page *page)
+{
+	int ret;
+	struct erofs_fscache *ctx = (struct erofs_fscache *)data;
+	struct folio *folio = page_folio(page);
+
+	ret = erofs_fscache_read_folio(ctx->cookie, folio, folio_pos(folio));
+	if (!ret)
+		folio_mark_uptodate(folio);
+
+	folio_unlock(folio);
+	return ret;
+}
+
 static const struct address_space_operations erofs_fscache_blob_aops = {
+	.readpage = erofs_fscache_readpage_blob,
 };
 
+/*
+ * erofs_fscache_get_folio - find and read page cache of blob file
+ * @ctx:	the context of the blob file
+ * @index:	the page index
+ *
+ * Get the page cache of the blob file at the index offset. It will find the
+ * page through the address space of the anonymous inode. This function is only
+ * used to read page cache of bootstrap blob file (metadata), since currently
+ * only bootstrap blob file manages an anonymous inode inside the fscache
+ * context.
+ *
+ * Return: up to date page on success, ERR_PTR() on failure.
+ */
+struct folio *erofs_fscache_get_folio(struct erofs_fscache *ctx, pgoff_t index)
+{
+	DBG_BUGON(!ctx->inode);
+	return read_mapping_folio(ctx->inode->i_mapping, index, ctx);
+}
+
 static int erofs_fscache_init_cookie(struct erofs_fscache *ctx, char *path)
 {
 	struct fscache_cookie *cookie;
fs/erofs/internal.h +8 -0
index d8c886a7491e..fa89a1e3012f 100644
@@ -632,6 +632,8 @@ void erofs_exit_fscache(void);
 struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path,
 					bool need_inode);
 void erofs_fscache_put(struct erofs_fscache *ctx);
+
+struct folio *erofs_fscache_get_folio(struct erofs_fscache *ctx, pgoff_t index);
 #else
 static inline int erofs_init_fscache(void) { return 0; }
 static inline void erofs_exit_fscache(void) {}
@@ -643,6 +645,12 @@ static inline struct erofs_fscache *erofs_fscache_get(struct super_block *sb,
 	return ERR_PTR(-EOPNOTSUPP);
 }
 static inline void erofs_fscache_put(struct erofs_fscache *ctx) {}
+
+static inline struct folio *erofs_fscache_get_folio(struct erofs_fscache *ctx,
+						    pgoff_t index)
+{
+	return ERR_PTR(-EOPNOTSUPP);
+}
 #endif
 
 #define EFSCORRUPTED    EUCLEAN         /* Filesystem is corrupted */
Signature · 1 line
2.27.0
author
#28 lore raw

PATCH v6 17/22 erofs: implement fscache-based data read for non-inline layout

In reply to Jeffle Xu

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  | 83 +++++++++++++++++++++++++++++++++++++++++++++
 fs/erofs/inode.c    |  8 ++++-
 fs/erofs/internal.h |  5 +++
 3 files changed, 95 insertions(+), 1 deletion(-)
fs/erofs/fscache.c +83 -0
index 91377939b4f7..4a9a4e60c15d 100644
@@ -60,10 +60,93 @@ static int erofs_fscache_readpage_blob(struct file *data, struct page *page)
 	return ret;
 }
 
+static inline int erofs_fscache_get_map(struct erofs_map_blocks *map,
+					struct super_block *sb)
+{
+	struct erofs_sb_info *sbi = EROFS_SB(sb);
+
+	map->m_fscache	= sbi->bootstrap;
+	return 0;
+}
+
+static int erofs_fscache_readpage_noinline(struct folio *folio,
+					   struct erofs_map_blocks *map)
+{
+	struct fscache_cookie *cookie = map->m_fscache->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 = map->m_pa + (map->o_la - map->m_la);
+
+	return erofs_fscache_read_folio(cookie, folio, start);
+}
+
+static int erofs_fscache_do_readpage(struct folio *folio)
+{
+	struct inode *inode = folio_file_mapping(folio)->host;
+	struct erofs_inode *vi = EROFS_I(inode);
+	struct super_block *sb = inode->i_sb;
+	struct erofs_map_blocks map;
+	int ret;
+
+	if (erofs_inode_is_data_compressed(vi->datalayout)) {
+		erofs_info(sb, "compressed layout not supported yet");
+		return -EOPNOTSUPP;
+	}
+
+	DBG_BUGON(folio_size(folio) != EROFS_BLKSIZ);
+
+	map.m_la = map.o_la = folio_pos(folio);
+
+	ret = erofs_map_blocks(inode, &map, EROFS_GET_BLOCKS_RAW);
+	if (ret)
+		return ret;
+
+	if (!(map.m_flags & EROFS_MAP_MAPPED)) {
+		folio_zero_range(folio, 0, folio_size(folio));
+		return 0;
+	}
+
+	ret = erofs_fscache_get_map(&map, sb);
+	if (ret)
+		return ret;
+
+	switch (vi->datalayout) {
+	case EROFS_INODE_FLAT_PLAIN:
+	case EROFS_INODE_CHUNK_BASED:
+		return erofs_fscache_readpage_noinline(folio, &map);
+	default:
+		DBG_BUGON(1);
+		return -EOPNOTSUPP;
+	}
+}
+
+static int erofs_fscache_readpage(struct file *file, struct page *page)
+{
+	struct folio *folio = page_folio(page);
+	int ret;
+
+	ret = erofs_fscache_do_readpage(folio);
+	if (!ret)
+		folio_mark_uptodate(folio);
+
+	folio_unlock(folio);
+	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,
+};
+
 /*
  * erofs_fscache_get_folio - find and read page cache of blob file
  * @ctx:	the context of the blob file
fs/erofs/inode.c +7 -1
index ff62f84f47d3..744faf3ef9f4 100644
@@ -296,7 +296,13 @@ 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;
+
+#ifdef CONFIG_EROFS_FS_ONDEMAND
+	if (erofs_is_nodev_mode(inode->i_sb))
+		inode->i_mapping->a_ops = &erofs_fscache_access_aops;
+#endif
+	if (!erofs_is_nodev_mode(inode->i_sb))
+		inode->i_mapping->a_ops = &erofs_raw_access_aops;
 
 out_unlock:
 	erofs_put_metabuf(&buf);
fs/erofs/internal.h +5 -0
index fa89a1e3012f..6537ededed51 100644
@@ -442,6 +442,9 @@ struct erofs_map_blocks {
 	unsigned short m_deviceid;
 	char m_algorithmformat;
 	unsigned int m_flags;
+
+	struct erofs_fscache *m_fscache;
+	erofs_off_t o_la;
 };
 
 /* Flags used by erofs_map_blocks_flatmode() */
@@ -634,6 +637,8 @@ struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path,
 void erofs_fscache_put(struct erofs_fscache *ctx);
 
 struct folio *erofs_fscache_get_folio(struct erofs_fscache *ctx, pgoff_t index);
+
+extern const struct address_space_operations erofs_fscache_access_aops;
 #else
 static inline int erofs_init_fscache(void) { return 0; }
 static inline void erofs_exit_fscache(void) {}
Signature · 1 line
2.27.0
#29 lore raw

Re: [PATCH v6 17/22] erofs: implement fscache-based data read for non-inline layout

Quoted 17 lines · Jeffle Xu

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  | 83 +++++++++++++++++++++++++++++++++++++++++++++
 fs/erofs/inode.c    |  8 ++++-
 fs/erofs/internal.h |  5 +++
 3 files changed, 95 insertions(+), 1 deletion(-)
fs/erofs/fscache.c +0 -0
index 91377939b4f7..4a9a4e60c15d 100644

fs/erofs/fscache.c static int erofs_fscache_readpage_blob(struct file *data, struct page *page) · line 64

Quoted 5 lines
fs/erofs/fscache.c +2 -0
@@ -60,10 +60,93 @@ static int erofs_fscache_readpage_blob(struct file *data, struct page *page)
 	return ret;
 }
 
+static inline int erofs_fscache_get_map(struct erofs_map_blocks *map,
+					struct super_block *sb)

I wonder if m_fscache should be settled in struct erofs_map_dev

And such helper can be merged into erofs_map_dev() as well.

Quoted 10 lines
fs/erofs/fscache.c +10 -0
+{
+	struct erofs_sb_info *sbi = EROFS_SB(sb);
+
+	map->m_fscache	= sbi->bootstrap;
+	return 0;
+}
+
+static int erofs_fscache_readpage_noinline(struct folio *folio,
+					   struct erofs_map_blocks *map)
+{

fs/erofs/fscache.c static int erofs_fscache_readpage_blob(struct file *data, struct page *page) · line 84

Quoted 10 lines
fs/erofs/fscache.c +10 -0
+	struct fscache_cookie *cookie = map->m_fscache->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 = map->m_pa + (map->o_la - map->m_la);

I think o_la can be directly replaced with "folio_pos(folio)".

Also such helper might be unneeded...

fs/erofs/fscache.c static int erofs_fscache_readpage_blob(struct file *data, struct page *page) · line 89

Quoted 5 lines
fs/erofs/fscache.c +5 -0
+
+	return erofs_fscache_read_folio(cookie, folio, start);
+}
+
+static int erofs_fscache_do_readpage(struct folio *folio)

Can it fold into erofs_fscache_readpage?
Another unneeded helper...

fs/erofs/fscache.c static int erofs_fscache_readpage_blob(struct file *data, struct page *page) · line 97

Quoted 8 lines
fs/erofs/fscache.c +8 -0
+{
+	struct inode *inode = folio_file_mapping(folio)->host;
+	struct erofs_inode *vi = EROFS_I(inode);
+	struct super_block *sb = inode->i_sb;
+	struct erofs_map_blocks map;
+	int ret;
+
+	if (erofs_inode_is_data_compressed(vi->datalayout)) {

It's impossible for now. So the check above is redundant.

Thanks,
Gao Xiang

Quoted 99 lines
fs/erofs/fscache.c +48 -0
+		erofs_info(sb, "compressed layout not supported yet");
+		return -EOPNOTSUPP;
+	}
+
+	DBG_BUGON(folio_size(folio) != EROFS_BLKSIZ);
+
+	map.m_la = map.o_la = folio_pos(folio);
+
+	ret = erofs_map_blocks(inode, &map, EROFS_GET_BLOCKS_RAW);
+	if (ret)
+		return ret;
+
+	if (!(map.m_flags & EROFS_MAP_MAPPED)) {
+		folio_zero_range(folio, 0, folio_size(folio));
+		return 0;
+	}
+
+	ret = erofs_fscache_get_map(&map, sb);
+	if (ret)
+		return ret;
+
+	switch (vi->datalayout) {
+	case EROFS_INODE_FLAT_PLAIN:
+	case EROFS_INODE_CHUNK_BASED:
+		return erofs_fscache_readpage_noinline(folio, &map);
+	default:
+		DBG_BUGON(1);
+		return -EOPNOTSUPP;
+	}
+}
+
+static int erofs_fscache_readpage(struct file *file, struct page *page)
+{
+	struct folio *folio = page_folio(page);
+	int ret;
+
+	ret = erofs_fscache_do_readpage(folio);
+	if (!ret)
+		folio_mark_uptodate(folio);
+
+	folio_unlock(folio);
+	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,
+};
+
 /*
  * erofs_fscache_get_folio - find and read page cache of blob file
  * @ctx:	the context of the blob file
fs/erofs/inode.c +7 -1
index ff62f84f47d3..744faf3ef9f4 100644
@@ -296,7 +296,13 @@ 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;
+
+#ifdef CONFIG_EROFS_FS_ONDEMAND
+	if (erofs_is_nodev_mode(inode->i_sb))
+		inode->i_mapping->a_ops = &erofs_fscache_access_aops;
+#endif
+	if (!erofs_is_nodev_mode(inode->i_sb))
+		inode->i_mapping->a_ops = &erofs_raw_access_aops;
 
 out_unlock:
 	erofs_put_metabuf(&buf);
fs/erofs/internal.h +5 -0
index fa89a1e3012f..6537ededed51 100644
@@ -442,6 +442,9 @@ struct erofs_map_blocks {
 	unsigned short m_deviceid;
 	char m_algorithmformat;
 	unsigned int m_flags;
+
+	struct erofs_fscache *m_fscache;
+	erofs_off_t o_la;
 };
 
 /* Flags used by erofs_map_blocks_flatmode() */
@@ -634,6 +637,8 @@ struct erofs_fscache *erofs_fscache_get(struct super_block *sb, char *path,
 void erofs_fscache_put(struct erofs_fscache *ctx);
 
 struct folio *erofs_fscache_get_folio(struct erofs_fscache *ctx, pgoff_t index);
+
+extern const struct address_space_operations erofs_fscache_access_aops;
 #else
 static inline int erofs_init_fscache(void) { return 0; }
 static inline void erofs_exit_fscache(void) {}
Signature · 1 line
2.27.0
author
#30 lore raw

PATCH v6 18/22 erofs: implement fscache-based data read for inline layout

In reply to Jeffle Xu

This 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 | 45 +++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 43 insertions(+), 2 deletions(-)
fs/erofs/fscache.c +43 -2
index 4a9a4e60c15d..d75958470645 100644
@@ -74,8 +74,9 @@ static int erofs_fscache_readpage_noinline(struct folio *folio,
 {
 	struct fscache_cookie *cookie = map->m_fscache->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
@@ -86,6 +87,42 @@ static int erofs_fscache_readpage_noinline(struct folio *folio,
 	return erofs_fscache_read_folio(cookie, folio, start);
 }
 
+static int erofs_fscache_readpage_inline(struct folio *folio,
+					 struct erofs_map_blocks *map)
+{
+	struct inode *inode = folio_file_mapping(folio)->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(map->m_pa);
+	blknr = erofs_blknr(map->m_pa);
+	len = map->m_llen;
+
+	src = erofs_read_metabuf(&buf, sb, blknr, EROFS_KMAP);
+	if (IS_ERR(src))
+		return PTR_ERR(src);
+
+	DBG_BUGON(folio_size(folio) != PAGE_SIZE);
+
+	dst = kmap(folio_page(folio, 0));
+	memcpy(dst, src + offset, len);
+	memset(dst + len, 0, PAGE_SIZE - len);
+	kunmap(folio_page(folio, 0));
+
+	erofs_put_metabuf(&buf);
+
+	return 0;
+}
+
 static int erofs_fscache_do_readpage(struct folio *folio)
 {
 	struct inode *inode = folio_file_mapping(folio)->host;
@@ -116,8 +153,12 @@ static int erofs_fscache_do_readpage(struct folio *folio)
 	if (ret)
 		return ret;
 
+	if (map.m_flags & EROFS_MAP_META)
+		return erofs_fscache_readpage_inline(folio, &map);
+
 	switch (vi->datalayout) {
 	case EROFS_INODE_FLAT_PLAIN:
+	case EROFS_INODE_FLAT_INLINE:
 	case EROFS_INODE_CHUNK_BASED:
 		return erofs_fscache_readpage_noinline(folio, &map);
 	default:
Signature · 1 line
2.27.0
author
#31 lore raw

PATCH v6 19/22 erofs: register cookie context for data blobs

In reply to Jeffle Xu

Similar 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.

Registers corresponding cookie context for each data blob file.

---
 fs/erofs/internal.h |  1 +
 fs/erofs/super.c    | 30 ++++++++++++++++++++++--------
 2 files changed, 23 insertions(+), 8 deletions(-)
fs/erofs/internal.h +1 -0
index 6537ededed51..94a118caf580 100644
@@ -52,6 +52,7 @@ struct erofs_device_info {
 	struct block_device *bdev;
 	struct dax_device *dax_dev;
 	u64 dax_part_off;
+	struct erofs_fscache *blob;
 
 	u32 blocks;
 	u32 mapped_blkaddr;
fs/erofs/super.c +22 -8
index de5aeda4aea0..9a6f35e0c22b 100644
@@ -259,15 +259,28 @@ 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 (IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) &&
+		    erofs_is_nodev_mode(sb)) {
+			struct erofs_fscache *blob;
+
+			blob = erofs_fscache_get(sb, dif->path, false);
+			if (IS_ERR(blob)) {
+				err = PTR_ERR(blob);
+				break;
+			}
+			dif->blob = blob;
+		} else {
+			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);
 		}
-		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 +707,7 @@ static int erofs_release_device_info(int id, void *ptr, void *data)
 {
 	struct erofs_device_info *dif = ptr;
 
+	erofs_fscache_put(dif->blob);
 	fs_put_dax(dif->dax_dev);
 	if (dif->bdev)
 		blkdev_put(dif->bdev, FMODE_READ | FMODE_EXCL);
Signature · 1 line
2.27.0
#32 lore raw

Re: [PATCH v6 19/22] erofs: register cookie context for data blobs

Quoted 18 lines · Jeffle Xu

Similar 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.

Registers corresponding cookie context for each data blob file.

---
 fs/erofs/internal.h |  1 +
 fs/erofs/super.c    | 30 ++++++++++++++++++++++--------
 2 files changed, 23 insertions(+), 8 deletions(-)
fs/erofs/internal.h +0 -0
index 6537ededed51..94a118caf580 100644

fs/erofs/internal.h struct erofs_device_info { · line 53

Quoted 2 lines
fs/erofs/internal.h +0 -0
@@ -52,6 +52,7 @@ struct erofs_device_info {
 	struct block_device *bdev;
 	struct dax_device *dax_dev;

Place

	struct erofs_fscache *fscache;

fs/erofs/internal.h struct erofs_device_info { · line 55

Quoted 2 lines
fs/erofs/internal.h +1 -0
 	u64 dax_part_off;
+	struct erofs_fscache *blob;

Instead here since `blob' is also nydus-specific.

Need to update the subject and commit message too.

Quoted 10 lines
fs/erofs/internal.h +0 -0
 	u32 blocks;
 	u32 mapped_blkaddr;
fs/erofs/super.c +0 -0
index de5aeda4aea0..9a6f35e0c22b 100644
@@ -259,15 +259,28 @@ static int erofs_init_devices(struct super_block *sb,
 		}
 		dis = ptr + erofs_blkoff(pos);

fs/erofs/super.c static int erofs_init_devices(struct super_block *sb, · line 264

Quoted 10 lines
fs/erofs/super.c +3 -6
 
-		bdev = blkdev_get_by_path(dif->path,
-					  FMODE_READ | FMODE_EXCL,
-					  sb->s_type);
-		if (IS_ERR(bdev)) {
-			err = PTR_ERR(bdev);
-			break;
+		if (IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) &&
+		    erofs_is_nodev_mode(sb)) {
+			struct erofs_fscache *blob;

Same here.

Thanks,
Gao Xiang

author
#33 lore raw

PATCH v6 20/22 erofs: implement fscache-based data read for data blobs

In reply to Jeffle Xu

Implements the data plane of reading data from data blob file over fscache.

---
 fs/erofs/data.c     |  3 +++
 fs/erofs/fscache.c  | 15 +++++++++++++--
 fs/erofs/internal.h |  1 +
 3 files changed, 17 insertions(+), 2 deletions(-)
fs/erofs/data.c +3 -0
index b4571bea93d5..b9a05de3c3b2 100644
@@ -206,6 +206,7 @@ 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_fscache = EROFS_SB(sb)->bootstrap;
 
 	if (map->m_deviceid) {
 		down_read(&devs->rwsem);
@@ -217,6 +218,7 @@ 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_fscache = dif->blob;
 		up_read(&devs->rwsem);
 	} else if (devs->extra_devices) {
 		down_read(&devs->rwsem);
@@ -234,6 +236,7 @@ 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_fscache = dif->blob;
 				break;
 			}
 		}
fs/erofs/fscache.c +13 -2
index d75958470645..cbb39657615e 100644
@@ -63,9 +63,20 @@ static int erofs_fscache_readpage_blob(struct file *data, struct page *page)
 static inline int erofs_fscache_get_map(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;
 
-	map->m_fscache	= sbi->bootstrap;
+	map->m_fscache	= mdev.m_fscache;
+	map->m_pa	= mdev.m_pa;
 	return 0;
 }
 
fs/erofs/internal.h +1 -0
index 94a118caf580..cea08f12a2c3 100644
@@ -487,6 +487,7 @@ struct erofs_map_dev {
 	struct block_device *m_bdev;
 	struct dax_device *m_daxdev;
 	u64 m_dax_part_off;
+	struct erofs_fscache *m_fscache;
 
 	erofs_off_t m_pa;
 	unsigned int m_deviceid;
Signature · 1 line
2.27.0
#34 lore raw

Re: [PATCH v6 20/22] erofs: implement fscache-based data read for data blobs

Quoted 43 lines · Jeffle Xu

Implements the data plane of reading data from data blob file over fscache.

---
 fs/erofs/data.c     |  3 +++
 fs/erofs/fscache.c  | 15 +++++++++++++--
 fs/erofs/internal.h |  1 +
 3 files changed, 17 insertions(+), 2 deletions(-)
fs/erofs/data.c +3 -0
index b4571bea93d5..b9a05de3c3b2 100644
@@ -206,6 +206,7 @@ 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_fscache = EROFS_SB(sb)->bootstrap;
 
 	if (map->m_deviceid) {
 		down_read(&devs->rwsem);
@@ -217,6 +218,7 @@ 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_fscache = dif->blob;
 		up_read(&devs->rwsem);
 	} else if (devs->extra_devices) {
 		down_read(&devs->rwsem);
@@ -234,6 +236,7 @@ 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_fscache = dif->blob;
 				break;
 			}
 		}
fs/erofs/fscache.c +0 -0
index d75958470645..cbb39657615e 100644

fs/erofs/fscache.c static int erofs_fscache_readpage_blob(struct file *data, struct page *page) · line 65

Quoted 3 lines
fs/erofs/fscache.c +0 -0
@@ -63,9 +63,20 @@ static int erofs_fscache_readpage_blob(struct file *data, struct page *page)
 static inline int erofs_fscache_get_map(struct erofs_map_blocks *map,
 					struct super_block *sb)
 {

So erofs_fscache_get_map seems really unneeded... erofs_map_dev can be used directly so we can avoid this patch.

Thanks,
Gao Xiang

Quoted 33 lines
fs/erofs/fscache.c +13 -2
-	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;
 
-	map->m_fscache	= sbi->bootstrap;
+	map->m_fscache	= mdev.m_fscache;
+	map->m_pa	= mdev.m_pa;
 	return 0;
 }
 
fs/erofs/internal.h +1 -0
index 94a118caf580..cea08f12a2c3 100644
@@ -487,6 +487,7 @@ struct erofs_map_dev {
 	struct block_device *m_bdev;
 	struct dax_device *m_daxdev;
 	u64 m_dax_part_off;
+	struct erofs_fscache *m_fscache;
 
 	erofs_off_t m_pa;
 	unsigned int m_deviceid;
Signature · 1 line
2.27.0
author
#35 lore raw

PATCH v6 21/22 erofs: implement fscache-based data readahead

In reply to Jeffle Xu

Implements fscache-based data readahead. Also registers an individual bdi for each erofs instance to enable readahead.

---
 fs/erofs/fscache.c | 114 +++++++++++++++++++++++++++++++++++++++++++++
 fs/erofs/super.c   |   4 ++
 2 files changed, 118 insertions(+)
fs/erofs/fscache.c +114 -0
index cbb39657615e..589d1e7c2b1b 100644
@@ -191,12 +191,126 @@ static int erofs_fscache_readpage(struct file *file, struct page *page)
 	return ret;
 }
 
+static inline size_t erofs_fscache_calc_len(struct erofs_map_blocks *map,
+					    size_t len, size_t done)
+{
+	/*
+	 * 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 length = map->m_llen - (map->o_la - map->m_la);
+
+	return min_t(size_t, length, len - done);
+}
+
+static inline void erofs_fscache_unlock_folios(struct readahead_control *rac,
+					       size_t len)
+{
+	while (len) {
+		struct folio *folio = readahead_folio(rac);
+
+		len -= folio_size(folio);
+		folio_mark_uptodate(folio);
+		folio_unlock(folio);
+	}
+}
+
+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;
+	size_t len, done = 0;
+	loff_t start;
+	int ret;
+
+	if (erofs_inode_is_data_compressed(vi->datalayout)) {
+		erofs_info(sb, "compressed layout not supported yet");
+		return;
+	}
+
+	if (!readahead_count(rac))
+		return;
+
+	start = readahead_pos(rac);
+	len = readahead_length(rac);
+
+	do {
+		struct erofs_map_blocks map;
+
+		map.m_la = map.o_la = start + done;
+
+		ret = erofs_map_blocks(inode, &map, EROFS_GET_BLOCKS_RAW);
+		if (ret)
+			return;
+
+		/* Read-ahead Hole
+		 * Two cases will hit this:
+		 * 1) EOF. Imposibble in readahead routine;
+		 * 2) hole. Only CHUNK_BASED layout supports hole.
+		 */
+		if (!(map.m_flags & EROFS_MAP_MAPPED)) {
+			struct iov_iter iter;
+			loff_t offset = start + done;
+			size_t count = erofs_fscache_calc_len(&map, len, done);
+
+			iov_iter_xarray(&iter, READ, &rac->mapping->i_pages, offset, count);
+			iov_iter_zero(count, &iter);
+
+			erofs_fscache_unlock_folios(rac, count);
+			ret = count;
+			continue;
+		}
+
+		ret = erofs_fscache_get_map(&map, sb);
+		if (ret)
+			return;
+
+		/* Read-ahead Inline */
+		if (map.m_flags & EROFS_MAP_META) {
+			struct folio *folio = readahead_folio(rac);
+
+			ret = erofs_fscache_readpage_inline(folio, &map);
+			if (!ret) {
+				folio_mark_uptodate(folio);
+				ret = folio_size(folio);
+			}
+
+			folio_unlock(folio);
+			continue;
+		}
+
+		/* Read-ahead No-inline */
+		if (vi->datalayout == EROFS_INODE_FLAT_PLAIN ||
+		    vi->datalayout == EROFS_INODE_FLAT_INLINE ||
+		    vi->datalayout == EROFS_INODE_CHUNK_BASED) {
+			struct fscache_cookie *cookie = map.m_fscache->cookie;
+			loff_t offset = start + done;
+			size_t count = erofs_fscache_calc_len(&map, len, done);
+			loff_t pstart = map.m_pa + (map.o_la - map.m_la);
+
+			ret = erofs_fscache_read_folios(cookie, rac->mapping,
+					offset, count, pstart);
+			if (!ret) {
+				erofs_fscache_unlock_folios(rac, count);
+				ret = count;
+			}
+		} else {
+			DBG_BUGON(1);
+			return;
+		}
+	} while (ret > 0 && ((done += ret) < 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,
 };
 
 /*
fs/erofs/super.c +4 -0
index 9a6f35e0c22b..8ac400581784 100644
@@ -619,6 +619,10 @@ 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
#36 lore raw

Re: [PATCH v6 21/22] erofs: implement fscache-based data readahead

Quoted 19 lines · Jeffle Xu

Implements fscache-based data readahead. Also registers an individual bdi for each erofs instance to enable readahead.

---
 fs/erofs/fscache.c | 114 +++++++++++++++++++++++++++++++++++++++++++++
 fs/erofs/super.c   |   4 ++
 2 files changed, 118 insertions(+)
fs/erofs/fscache.c +2 -0
index cbb39657615e..589d1e7c2b1b 100644
@@ -191,12 +191,126 @@ static int erofs_fscache_readpage(struct file *file, struct page *page)
 	return ret;
 }
 
+static inline size_t erofs_fscache_calc_len(struct erofs_map_blocks *map,
+					    size_t len, size_t done)

fs/erofs/fscache.c static int erofs_fscache_readpage(struct file *file, struct page *page) · line 205

Quoted 10 lines
fs/erofs/fscache.c +10 -0
+{
+	/*
+	 * 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 length = map->m_llen - (map->o_la - map->m_la);
+
+	return min_t(size_t, length, len - done);

This helper can be folded too.

Quoted 17 lines
fs/erofs/fscache.c +17 -0
+}
+
+static inline void erofs_fscache_unlock_folios(struct readahead_control *rac,
+					       size_t len)
+{
+	while (len) {
+		struct folio *folio = readahead_folio(rac);
+
+		len -= folio_size(folio);
+		folio_mark_uptodate(folio);
+		folio_unlock(folio);
+	}
+}
+
+static void erofs_fscache_readahead(struct readahead_control *rac)
+{
+	struct inode *inode = rac->mapping->host;

fs/erofs/fscache.c static int erofs_fscache_readpage(struct file *file, struct page *page) · line 232

Quoted 10 lines
fs/erofs/fscache.c +10 -0
+	struct erofs_inode *vi = EROFS_I(inode);
+	struct super_block *sb = inode->i_sb;
+	size_t len, done = 0;
+	loff_t start;
+	int ret;
+
+	if (erofs_inode_is_data_compressed(vi->datalayout)) {
+		erofs_info(sb, "compressed layout not supported yet");
+		return;
+	}

Redundant check.

Quoted 11 lines
fs/erofs/fscache.c +11 -0
+
+	if (!readahead_count(rac))
+		return;
+
+	start = readahead_pos(rac);
+	len = readahead_length(rac);
+
+	do {
+		struct erofs_map_blocks map;
+
+		map.m_la = map.o_la = start + done;

fs/erofs/fscache.c static int erofs_fscache_readpage(struct file *file, struct page *page) · line 253

Quoted 10 lines
fs/erofs/fscache.c +10 -0
+
+		ret = erofs_map_blocks(inode, &map, EROFS_GET_BLOCKS_RAW);
+		if (ret)
+			return;
+
+		/* Read-ahead Hole
+		 * Two cases will hit this:
+		 * 1) EOF. Imposibble in readahead routine;
+		 * 2) hole. Only CHUNK_BASED layout supports hole.
+		 */

/*
*
*/

and typo `Imposibble'. Also I think this comment may not be useful though.

Quoted 28 lines
fs/erofs/fscache.c +28 -0
+		if (!(map.m_flags & EROFS_MAP_MAPPED)) {
+			struct iov_iter iter;
+			loff_t offset = start + done;
+			size_t count = erofs_fscache_calc_len(&map, len, done);
+
+			iov_iter_xarray(&iter, READ, &rac->mapping->i_pages, offset, count);
+			iov_iter_zero(count, &iter);
+
+			erofs_fscache_unlock_folios(rac, count);
+			ret = count;
+			continue;
+		}
+
+		ret = erofs_fscache_get_map(&map, sb);
+		if (ret)
+			return;
+
+		/* Read-ahead Inline */
+		if (map.m_flags & EROFS_MAP_META) {
+			struct folio *folio = readahead_folio(rac);
+
+			ret = erofs_fscache_readpage_inline(folio, &map);
+			if (!ret) {
+				folio_mark_uptodate(folio);
+				ret = folio_size(folio);
+			}
+
+			folio_unlock(folio);

fs/erofs/fscache.c static int erofs_fscache_readpage(struct file *file, struct page *page) · line 291

Quoted 10 lines
fs/erofs/fscache.c +10 -0
+			continue;
+		}
+
+		/* Read-ahead No-inline */
+		if (vi->datalayout == EROFS_INODE_FLAT_PLAIN ||
+		    vi->datalayout == EROFS_INODE_FLAT_INLINE ||
+		    vi->datalayout == EROFS_INODE_CHUNK_BASED) {
+			struct fscache_cookie *cookie = map.m_fscache->cookie;
+			loff_t offset = start + done;
+			size_t count = erofs_fscache_calc_len(&map, len, done);

You could promote offset' and count' to the outer block. So another offset' and count' in !EROFS_MAP_MAPPED can be dropped then.

Thanks,
Gao Xiang

author
#37 lore raw

PATCH v6 22/22 erofs: add 'tag' mount option

In reply to Jeffle Xu

Introduce 'tag' 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 name of bootstrap blob file containing the complete erofs image.

---
 fs/erofs/super.c | 44 ++++++++++++++++++++++++++++++++++++++------
 1 file changed, 38 insertions(+), 6 deletions(-)
fs/erofs/super.c +38 -6
index 8ac400581784..6ea83f36842c 100644
@@ -403,6 +403,7 @@ enum {
 	Opt_dax,
 	Opt_dax_enum,
 	Opt_device,
+	Opt_tag,
 	Opt_err
 };
 
@@ -427,6 +428,7 @@ 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("tag",		Opt_tag),
 	{}
 };
 
@@ -522,6 +524,16 @@ static int erofs_fc_parse_param(struct fs_context *fc,
 		}
 		++ctx->devs->extra_devices;
 		break;
+	case Opt_tag:
+#ifdef CONFIG_EROFS_FS_ONDEMAND
+		kfree(ctx->opt.tag);
+		ctx->opt.tag = kstrdup(param->string, GFP_KERNEL);
+		if (!ctx->opt.tag)
+			return -ENOMEM;
+#else
+		errorfc(fc, "tag option not supported");
+#endif
+		break;
 	default:
 		return -ENOPARAM;
 	}
@@ -596,9 +608,14 @@ 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 (IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) && erofs_is_nodev_mode(sb)) {
+		sb->s_blocksize = EROFS_BLKSIZ;
+		sb->s_blocksize_bits = LOG_BLOCK_SIZE;
+	} else {
+		if (!sb_set_blocksize(sb, EROFS_BLKSIZ)) {
+			erofs_err(sb, "failed to set erofs blksize");
+			return -EINVAL;
+		}
 	}
 
 	sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
@@ -607,7 +624,6 @@ 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;
 
@@ -623,6 +639,10 @@ 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;
+	} else {
+		sbi->dax_dev = fs_dax_get_by_bdev(sb->s_bdev, &sbi->dax_part_off);
 	}
 
 	err = erofs_read_superblock(sb);
@@ -685,6 +705,11 @@ 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 (IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) && ctx->opt.tag)
+		return get_tree_nodev(fc, erofs_fc_fill_super);
+
 	return get_tree_bdev(fc, erofs_fc_fill_super);
 }
 
@@ -734,6 +759,7 @@ 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.tag);
 	kfree(ctx);
 }
 
@@ -774,7 +800,10 @@ static void erofs_kill_sb(struct super_block *sb)
 
 	WARN_ON(sb->s_magic != EROFS_SUPER_MAGIC);
 
-	kill_block_super(sb);
+	if (IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) && erofs_is_nodev_mode(sb))
+		generic_shutdown_super(sb);
+	else
+		kill_block_super(sb);
 
 	sbi = EROFS_SB(sb);
 	if (!sbi)
@@ -896,7 +925,10 @@ 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 = 0;
+
+	if (!IS_ENABLED(CONFIG_EROFS_FS_ONDEMAND) || !erofs_is_nodev_mode(sb))
+		id = huge_encode_dev(sb->s_bdev->bd_dev);
 
 	buf->f_type = sb->s_magic;
 	buf->f_bsize = EROFS_BLKSIZ;
Signature · 1 line
2.27.0