This patch-set includes the following major enhancement patches.
o introduce a new gloabl lock scheme
o add tracepoints on several major functions
o fix the overall cleaning process focused on victim selection
o apply the block plugging to merge IOs as much as possible
o enhance management of free nids and its list
o enhance the readahead mode for node pages
o address several cretical deadlock conditions
o reduce lock_page calls
The other minor bug fixes and enhancements are as follows.
o calculation mistakes: overflow
o bio types: READ, READA, and READ_SYNC
o fix the recovery flow, data races, and null pointer errors
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)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=Hqii
-----END PGP SIGNATURE-----
Merge tag 'f2fs-for-v3.10' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs
Pull f2fs updates from Jaegeuk Kim:
"This patch-set includes the following major enhancement patches.
- introduce a new gloabl lock scheme
- add tracepoints on several major functions
- fix the overall cleaning process focused on victim selection
- apply the block plugging to merge IOs as much as possible
- enhance management of free nids and its list
- enhance the readahead mode for node pages
- address several cretical deadlock conditions
- reduce lock_page calls
The other minor bug fixes and enhancements are as follows.
- calculation mistakes: overflow
- bio types: READ, READA, and READ_SYNC
- fix the recovery flow, data races, and null pointer errors"
* tag 'f2fs-for-v3.10' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (68 commits)
f2fs: cover free_nid management with spin_lock
f2fs: optimize scan_nat_page()
f2fs: code cleanup for scan_nat_page() and build_free_nids()
f2fs: bugfix for alloc_nid_failed()
f2fs: recover when journal contains deleted files
f2fs: continue to mount after failing recovery
f2fs: avoid deadlock during evict after f2fs_gc
f2fs: modify the number of issued pages to merge IOs
f2fs: remove useless #include <linux/proc_fs.h> as we're now using sysfs as debug entry.
f2fs: fix inconsistent using of NM_WOUT_THRESHOLD
f2fs: check truncation of mapping after lock_page
f2fs: enhance alloc_nid and build_free_nids flows
f2fs: add a tracepoint on f2fs_new_inode
f2fs: check nid == 0 in add_free_nid
f2fs: add REQ_META about metadata requests for submit
f2fs: give a chance to merge IOs by IO scheduler
f2fs: avoid frequent background GC
f2fs: add tracepoints to debug checkpoint request
f2fs: add tracepoints for write page operations
f2fs: add tracepoints to debug the block allocation
...
When unable to roll forward the journal, we shouldn't bail out and
not mount, we should continue to attempt the mount. Bad recovery data
is likely unrecoverable at this point, and requiring the user to try
to mount again doesn't solve any issues.
Signed-off-by: Chris Fries <C.Fries@motorola.com>
Reviewed-by: Russell Knize <rknize2@motorola.com>
Reviewed-by: Jason Hrycay <jason.hrycay@motorola.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
o Deadlock case #1
Thread 1:
- writeback_sb_inodes
- do_writepages
- f2fs_write_data_pages
- write_cache_pages
- f2fs_write_data_page
- f2fs_balance_fs
- wait mutex_lock(gc_mutex)
Thread 2:
- f2fs_balance_fs
- mutex_lock(gc_mutex)
- f2fs_gc
- f2fs_iget
- wait iget_locked(inode->i_lock)
Thread 3:
- do_unlinkat
- iput
- lock(inode->i_lock)
- evict
- inode_wait_for_writeback
o Deadlock case #2
Thread 1:
- __writeback_single_inode
: set I_SYNC
- do_writepages
- f2fs_write_data_page
- f2fs_balance_fs
- f2fs_gc
- iput
- evict
- inode_wait_for_writeback(I_SYNC)
In order to avoid this, even though iput is called with the zero-reference
count, we need to stop the eviction procedure if the inode is on writeback.
So this patch links f2fs_drop_inode which checks the I_SYNC flag.
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
Add tracepoints in f2fs for tracing the syncing
operations like filesystem sync, file sync enter/exit.
It will helf to trace the code under debugging scenarios.
Also add tracepoints for tracing the various inode operations
like building inode, eviction of inode, link/unlike of
inodes.
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Pankaj Kumar <pankaj.km@samsung.com>
Acked-by: Steven Rostedt <rostedt@goodmis.org>
[Jaegeuk: combine and modify the tracepoint structures]
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
Fix to return a negative error code from the error handling
case instead of 0, as returned elsewhere in this function.
Introduce by commit c0d39e(f2fs: fix return values from validate superblock)
Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Acked-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
In the previous version, f2fs uses global locks according to the usage types,
such as directory operations, block allocation, block write, and so on.
Reference the following lock types in f2fs.h.
enum lock_type {
RENAME, /* for renaming operations */
DENTRY_OPS, /* for directory operations */
DATA_WRITE, /* for data write */
DATA_NEW, /* for data allocation */
DATA_TRUNC, /* for data truncate */
NODE_NEW, /* for node allocation */
NODE_TRUNC, /* for node truncate */
NODE_WRITE, /* for node write */
NR_LOCK_TYPE,
};
In that case, we lose the performance under the multi-threading environment,
since every types of operations must be conducted one at a time.
In order to address the problem, let's share the locks globally with a mutex
array regardless of any types.
So, let users grab a mutex and perform their jobs in parallel as much as
possbile.
For this, I propose a new global lock scheme as follows.
0. Data structure
- f2fs_sb_info -> mutex_lock[NR_GLOBAL_LOCKS]
- f2fs_sb_info -> node_write
1. mutex_lock_op(sbi)
- try to get an avaiable lock from the array.
- returns the index of the gottern lock variable.
2. mutex_unlock_op(sbi, index of the lock)
- unlock the given index of the lock.
3. mutex_lock_all(sbi)
- grab all the locks in the array before the checkpoint.
4. mutex_unlock_all(sbi)
- release all the locks in the array after checkpoint.
5. block_operations()
- call mutex_lock_all()
- sync_dirty_dir_inodes()
- grab node_write
- sync_node_pages()
Note that,
the pairs of mutex_lock_op()/mutex_unlock_op() and
mutex_lock_all()/mutex_unlock_all() should be used together.
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
In order to do GC more reliably, I'd like to lock the vicitm summary page
until its GC is completed, and also prevent any checkpoint process.
Reviewed-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
This patch removes a bitmap for victim segments selected by foreground GC, and
modifies the other bitmap for victim segments selected by background GC.
1) foreground GC bitmap
: We don't need to manage this, since we just only one previous victim section
number instead of the whole victim history.
The f2fs uses the victim section number in order not to allocate currently
GC'ed section to current active logs.
2) background GC bitmap
: This bitmap is used to avoid selecting victims repeatedly by background GCs.
In addition, the victims are able to be selected by foreground GCs, since
there is no need to read victim blocks during foreground GCs.
By the fact that the foreground GC reclaims segments in a section unit, it'd
be better to manage this bitmap based on the section granularity.
Reviewed-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
We should handle errors during the recovery flow correctly.
For example, if we get -ENOMEM, we should report a mount failure instead of
conducting the remained mount procedure.
Reviewed-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
validate super block is not returning with proper values.
When failure from sb_bread it should reflect there is an EIO otherwise
it should return of EINVAL.
Returning, '1' is not conveying proper message as the return type.
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
Change f2fs so that a warning is emitted when an attempt is made to
mount a filesystem with the unsupported discard option.
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
The maximum filename length supported in linux is 255 characters.
So let's follow that.
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
Modify the request_module to prefix the file system type with "fs-"
and add aliases to all of the filesystems that can be built as modules
to match.
A common practice is to build all of the kernel code and leave code
that is not commonly needed as modules, with the result that many
users are exposed to any bug anywhere in the kernel.
Looking for filesystems with a fs- prefix limits the pool of possible
modules that can be loaded by mount to just filesystems trivially
making things safer with no real cost.
Using aliases means user space can control the policy of which
filesystem modules are auto-loaded by editing /etc/modprobe.d/*.conf
with blacklist and alias directives. Allowing simple, safe,
well understood work-arounds to known problematic software.
This also addresses a rare but unfortunate problem where the filesystem
name is not the same as it's module name and module auto-loading
would not work. While writing this patch I saw a handful of such
cases. The most significant being autofs that lives in the module
autofs4.
This is relevant to user namespaces because we can reach the request
module in get_fs_type() without having any special permissions, and
people get uncomfortable when a user specified string (in this case
the filesystem type) goes all of the way to request_module.
After having looked at this issue I don't think there is any
particular reason to perform any filtering or permission checks beyond
making it clear in the module request that we want a filesystem
module. The common pattern in the kernel is to call request_module()
without regards to the users permissions. In general all a filesystem
module does once loaded is call register_filesystem() and go to sleep.
Which means there is not much attack surface exposed by loading a
filesytem module unless the filesystem is mounted. In a user
namespace filesystems are not mounted unless .fs_flags = FS_USERNS_MOUNT,
which most filesystems do not set today.
Acked-by: Serge Hallyn <serge.hallyn@canonical.com>
Acked-by: Kees Cook <keescook@chromium.org>
Reported-by: Kees Cook <keescook@google.com>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
This patch makes clearer the ambiguous f2fs_gc flow as follows.
1. Remove intermediate checkpoint condition during f2fs_gc
(i.e., should_do_checkpoint() and GC_BLOCKED)
2. Remove unnecessary return values of f2fs_gc because of #1.
(i.e., GC_NODE, GC_OK, etc)
3. Simplify write_checkpoint() because of #2.
4. Clarify the main f2fs_gc flow.
o monitor how many freed sections during one iteration of do_garbage_collect().
o do GC more without checkpoints if we can't get enough free sections.
o do checkpoint once we've got enough free sections through forground GCs.
5. Adopt thread-logging (Slack-Space-Recycle) scheme more aggressively on data
log types. See. get_ssr_segement()
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
In f2fs, there are two superblocks. So when the first superblock was
invalidate, it should try to check another.
By Jaegeuk Kim:
o Remove a white space for coding style
o Clean up for code readability
o Fix a typo
Signed-off-by: Jianpeng Ma <majianpeng@gmail.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
In some system PAGE_CACHE_SIZE isn't 4K. So using F2FS_BLKSIZE to judge.
By Jaegeuk Kim:
o f2fs does not support no other 4KB page cache size.
Signed-off-by: Jianpeng Ma <majianpeng@gmail.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
This patch supports ioctl FIFREEZE and FITHAW to snapshot filesystem.
Before calling f2fs_freeze, all writers would be suspended and sync_fs
would be completed. So no f2fs has to do something.
Just background gc operation should be skipped due to generate dirty
nodes and data until unfreeze.
Signed-off-by: Changman Lee <cm224.lee@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
There is a typo in the ->show_options function for disable_ext_identify.
Fix it to match the spelling from the documentation.
Signed-off-by: Alejandro Martinez Ruiz <alex@nowcomputing.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
This patch enhances the checkpoint routine to cope with IO errors.
Basically f2fs detects IO errors from end_io_write, and the errors are able to
be occurred during one of data, node, and meta page writes.
In the previous code, when an IO error is occurred during writes, f2fs sets a
flag, CP_ERROR_FLAG, in the raw ckeckpoint buffer which will be written to disk.
Afterwards, write_checkpoint() will check the flag and remount f2fs as a
read-only (ro) mode.
However, even once f2fs is remounted as a ro mode, dirty checkpoint pages are
freely able to be written to disk by flusher or kswapd in background.
In such a case, after cold reboot, f2fs would restore the checkpoint data having
CP_ERROR_FLAG, resulting in disabling write_checkpoint and remounting f2fs as
a ro mode again.
Therefore, let's prevent any checkpoint page (meta) writes once an IO error is
occurred, and remount f2fs as a ro mode right away at that moment.
Reported-by: Oliver Winker <oliver@oli1170.net>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
Reviewed-by: Namjae Jeon <namjae.jeon@samsung.com>
Add __init to functions in init_f2fs_fs for code consistency.
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
As the "status" debugfs entry will be maintained for entire F2FS filesystem
irrespective of the number of partitions.
So, we can move the initialization to the init part of the f2fs and destroy will
be done from exit part. After making changes, for individual partition mount -
entry creation code will not be executed.
Signed-off-by: Jianpeng Ma <majianpeng@gmail.com>
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
Since, f2fs supports only 4KB blocksize, which is set at the beginning in
f2fs_fill_super. So, we do not need to again check this blocksize setting
in such case.
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
The f2fs_balance_fs() is to check the number of free sections and decide whether
it needs to conduct cleaning or not. If there are not enough free sections, the
cleaning job should be started.
In order to control an amount of free sections even under high utilization, f2fs
should call f2fs_balance_fs at all the VFS interfaces that are able to produce
dirty pages.
This patch adds the function calls in the missing interfaces as follows.
1. f2fs_setxattr()
The f2fs_setxattr() produces dirty node pages so that we should call
f2fs_balance_fs() either likewise doing in other VFS interfaces such as
f2fs_lookup(), f2fs_mkdir(), and so on.
2. f2fs_sync_file()
We should guarantee serving free sections for syncing metadata during fsync.
Previously, there is no space check before triggering checkpoint and
sync_node_pages.
Therefore, if a bunch of fsync calls are triggered under 100% of FS utilization,
f2fs is able to be faced with no free sections, resulting in BUG_ON().
3. f2fs_sync_fs()
Before calling write_checkpoint(), we should guarantee that there are minimum
free sections.
4. f2fs_write_inode()
f2fs_write_inode() is also able to produce dirty node pages.
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
Introduced f2fs_msg function to differentiate f2fs specific messages in
the log. And, added few informative prints in the mount path, to convey
proper error in case of mount failure.
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
We can directly return '0' from the function, instead of introducing a
'ret' variable.
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
This patch resolves a build warning reported by kbuild test robot.
"
fs/f2fs/segment.c: In function '__get_segment_type':
fs/f2fs/segment.c:806:1: warning: control reaches end of non-void
function [-Wreturn-type]
"
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
In f2fs_statfs(), f_files should be the total number of available inodes
instead of the currently allocated inodes.
So, this patch should resolve the reported bug below.
Note that, showing 10% usage is not a bug, since f2fs reveals whole volume size
as much as possible and shows the space overhead as *used*.
This policy is fair enough with respect to other file systems.
<Reported Bug>
(loop0 is backed by 1GiB file)
$ mkfs.f2fs /dev/loop0
F2FS-tools: Ver: 1.1.0 (2012-12-11)
Info: sector size = 512
Info: total sectors = 2097152 (in 512bytes)
Info: zone aligned segment0 blkaddr: 512
Info: format successful
$ mount /dev/loop0 mnt/
$ df mnt/
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/loop0 1046528 98312 929784 10%
/home/zeta/linux-devel/mtd-bench/mnt
$ df mnt/ -i
Filesystem Inodes IUsed IFree IUse% Mounted on
/dev/loop0 1 -465918 465919 - /home/zeta/linux-devel/mtd-bench/mnt
Notice IUsed is negative. Also, 10% usage on a fresh f2fs seems too
much to be correct.
Reported-and-Tested-by: Ezequiel Garcia <elezegarcia@gmail.com>
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
Since, __GFP_ZERO is used while f2fs inode allocation, so we do not
need memset for f2fs_inode_info, as this is already zeroed out.
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
print the invalid argument/value from parse_options in case of
mount failure.
Signed-off-by: Namjae Jeon <namjae.jeon@samsung.com>
Signed-off-by: Amit Sahrawat <a.sahrawat@samsung.com>
As pointed out by Randy Dunlap, this patch removes all usage of "/**" for comment
blocks. Instead, just use "/*".
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
This patch should resolve the bugs reported by the sparse tool.
Initial reports were written by "kbuild test robot" managed by fengguang.wu.
In my local machines, I've tested also by running:
> make C=2 CF="-D__CHECK_ENDIAN__"
Accordingly, I've found lots of warnings and bugs related to the endian
conversion. And I've fixed all at this moment.
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
This adds the implementation of superblock operations for f2fs, which includes
- init_f2fs_fs/exit_f2fs_fs
- f2fs_mount
- super_operations of f2fs
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
Linus Torvalds
Chris Fries
Jaegeuk Kim
Haicheng Li
Namjae Jeon
Wei Yongjun
Masanari Iida
Eric W. Biederman
majianpeng
Changman Lee
Alejandro Martinez Ruiz
Namjae Jeon