Skip to content

L
linux
Commit ee5dc049 authored by Tobin C. Harding's avatar Tobin C. Harding Committed by Jonathan Corbet
Browse files
Options

docs: filesystems: vfs: Render method descriptions 

Currently vfs.rst does not render well into HTML the method descriptions
for VFS data structures.  We can improve the HTML output by putting the
description string on a new line following the method name.
Suggested-by: default avatarJonathan Corbet <corbet@lwn.net>
Signed-off-by: default avatarTobin C. Harding <tobin@kernel.org>
Signed-off-by: default avatarJonathan Corbet <corbet@lwn.net>
parent af96c1e3
Show whitespace changes
Inline Side-by-side
Showing with 642 additions and 505 deletions
Documentation/filesystems/vfs.rst
View file @ ee5dc049
...@@ -125,35 +125,46 @@ members are defined: ...@@ -125,35 +125,46 @@ members are defined:
struct lock_class_key s_umount_key; struct lock_class_key s_umount_key;
}; };
``name``: the name of the filesystem type, such as "ext2", "iso9660", ``name``
the name of the filesystem type, such as "ext2", "iso9660",
"msdos" and so on "msdos" and so on
``fs_flags``: various flags (i.e. FS_REQUIRES_DEV, FS_NO_DCACHE, etc.) ``fs_flags``
various flags (i.e. FS_REQUIRES_DEV, FS_NO_DCACHE, etc.)
``mount``: the method to call when a new instance of this filesystem should ``mount``
be mounted the method to call when a new instance of this filesystem should
be mounted
``kill_sb``: the method to call when an instance of this filesystem ``kill_sb``
should be shut down the method to call when an instance of this filesystem should be
shut down
``owner``: for internal VFS use: you should initialize this to THIS_MODULE in
most cases.
``next``: for internal VFS use: you should initialize this to NULL ``owner``
for internal VFS use: you should initialize this to THIS_MODULE
in most cases.
``next``
for internal VFS use: you should initialize this to NULL
s_lock_key, s_umount_key: lockdep-specific s_lock_key, s_umount_key: lockdep-specific
The mount() method has the following arguments: The mount() method has the following arguments:
``struct file_system_type *fs_type``: describes the filesystem, partly initialized ``struct file_system_type *fs_type``
by the specific filesystem code describes the filesystem, partly initialized by the specific
filesystem code
``int flags``: mount flags ``int flags``
mount flags
``const char *dev_name``: the device name we are mounting. ``const char *dev_name``
the device name we are mounting.
``void *data``: arbitrary mount options, usually comes as an ASCII ``void *data``
string (see "Mount Options" section) arbitrary mount options, usually comes as an ASCII string (see
"Mount Options" section)
The mount() method must return the root dentry of the tree requested by The mount() method must return the root dentry of the tree requested by
caller. An active reference to its superblock must be grabbed and the caller. An active reference to its superblock must be grabbed and the
...@@ -178,22 +189,27 @@ implementation. ...@@ -178,22 +189,27 @@ implementation.
Usually, a filesystem uses one of the generic mount() implementations Usually, a filesystem uses one of the generic mount() implementations
and provides a fill_super() callback instead. The generic variants are: and provides a fill_super() callback instead. The generic variants are:
``mount_bdev``: mount a filesystem residing on a block device ``mount_bdev``
mount a filesystem residing on a block device
``mount_nodev``: mount a filesystem that is not backed by a device ``mount_nodev``
mount a filesystem that is not backed by a device
``mount_single``: mount a filesystem which shares the instance between ``mount_single``
all mounts mount a filesystem which shares the instance between all mounts
A fill_super() callback implementation has the following arguments: A fill_super() callback implementation has the following arguments:
``struct super_block *sb``: the superblock structure. The callback ``struct super_block *sb``
must initialize this properly. the superblock structure. The callback must initialize this
properly.
``void *data``: arbitrary mount options, usually comes as an ASCII ``void *data``
string (see "Mount Options" section) arbitrary mount options, usually comes as an ASCII string (see
"Mount Options" section)
``int silent``: whether or not to be silent on error ``int silent``
whether or not to be silent on error
The Superblock Object The Superblock Object
...@@ -240,87 +256,106 @@ noted. This means that most methods can block safely. All methods are ...@@ -240,87 +256,106 @@ noted. This means that most methods can block safely. All methods are
only called from a process context (i.e. not from an interrupt handler only called from a process context (i.e. not from an interrupt handler
or bottom half). or bottom half).
``alloc_inode``: this method is called by alloc_inode() to allocate memory ``alloc_inode``
for struct inode and initialize it. If this function is not this method is called by alloc_inode() to allocate memory for
struct inode and initialize it. If this function is not
defined, a simple 'struct inode' is allocated. Normally defined, a simple 'struct inode' is allocated. Normally
alloc_inode will be used to allocate a larger structure which alloc_inode will be used to allocate a larger structure which
contains a 'struct inode' embedded within it. contains a 'struct inode' embedded within it.
``destroy_inode``: this method is called by destroy_inode() to release ``destroy_inode``
resources allocated for struct inode. It is only required if this method is called by destroy_inode() to release resources
allocated for struct inode. It is only required if
->alloc_inode was defined and simply undoes anything done by ->alloc_inode was defined and simply undoes anything done by
->alloc_inode. ->alloc_inode.
``dirty_inode``: this method is called by the VFS to mark an inode dirty. ``dirty_inode``
this method is called by the VFS to mark an inode dirty.
``write_inode``: this method is called when the VFS needs to write an ``write_inode``
inode to disc. The second parameter indicates whether the write this method is called when the VFS needs to write an inode to
should be synchronous or not, not all filesystems check this flag. disc. The second parameter indicates whether the write should
be synchronous or not, not all filesystems check this flag.
``drop_inode``: called when the last access to the inode is dropped, ``drop_inode``
with the inode->i_lock spinlock held. called when the last access to the inode is dropped, with the
inode->i_lock spinlock held.
This method should be either NULL (normal UNIX filesystem This method should be either NULL (normal UNIX filesystem
semantics) or "generic_delete_inode" (for filesystems that do not semantics) or "generic_delete_inode" (for filesystems that do
want to cache inodes - causing "delete_inode" to always be not want to cache inodes - causing "delete_inode" to always be
called regardless of the value of i_nlink) called regardless of the value of i_nlink)
The "generic_delete_inode()" behavior is equivalent to the The "generic_delete_inode()" behavior is equivalent to the old
old practice of using "force_delete" in the put_inode() case, practice of using "force_delete" in the put_inode() case, but
but does not have the races that the "force_delete()" approach does not have the races that the "force_delete()" approach had.
had.
``delete_inode``: called when the VFS wants to delete an inode ``delete_inode``
called when the VFS wants to delete an inode
``put_super``: called when the VFS wishes to free the superblock ``put_super``
called when the VFS wishes to free the superblock
(i.e. unmount). This is called with the superblock lock held (i.e. unmount). This is called with the superblock lock held
``sync_fs``: called when VFS is writing out all dirty data associated with ``sync_fs``
a superblock. The second parameter indicates whether the method called when VFS is writing out all dirty data associated with a
superblock. The second parameter indicates whether the method
should wait until the write out has been completed. Optional. should wait until the write out has been completed. Optional.
``freeze_fs``: called when VFS is locking a filesystem and ``freeze_fs``
forcing it into a consistent state. This method is currently called when VFS is locking a filesystem and forcing it into a
used by the Logical Volume Manager (LVM). consistent state. This method is currently used by the Logical
Volume Manager (LVM).
``unfreeze_fs``: called when VFS is unlocking a filesystem and making it writable ``unfreeze_fs``
called when VFS is unlocking a filesystem and making it writable
again. again.
``statfs``: called when the VFS needs to get filesystem statistics. ``statfs``
called when the VFS needs to get filesystem statistics.
``remount_fs``: called when the filesystem is remounted. This is called ``remount_fs``
with the kernel lock held called when the filesystem is remounted. This is called with
the kernel lock held
``clear_inode``: called then the VFS clears the inode. Optional ``clear_inode``
called then the VFS clears the inode. Optional
``umount_begin``: called when the VFS is unmounting a filesystem. ``umount_begin``
called when the VFS is unmounting a filesystem.
``show_options``: called by the VFS to show mount options for ``show_options``
/proc/<pid>/mounts. (see "Mount Options" section) called by the VFS to show mount options for /proc/<pid>/mounts.
(see "Mount Options" section)
``quota_read``: called by the VFS to read from filesystem quota file. ``quota_read``
called by the VFS to read from filesystem quota file.
``quota_write``: called by the VFS to write to filesystem quota file. ``quota_write``
called by the VFS to write to filesystem quota file.
``nr_cached_objects``: called by the sb cache shrinking function for the ``nr_cached_objects``
filesystem to return the number of freeable cached objects it contains. called by the sb cache shrinking function for the filesystem to
return the number of freeable cached objects it contains.
Optional. Optional.
``free_cache_objects``: called by the sb cache shrinking function for the ``free_cache_objects``
filesystem to scan the number of objects indicated to try to free them. called by the sb cache shrinking function for the filesystem to
Optional, but any filesystem implementing this method needs to also scan the number of objects indicated to try to free them.
implement ->nr_cached_objects for it to be called correctly. Optional, but any filesystem implementing this method needs to
also implement ->nr_cached_objects for it to be called
correctly.
We can't do anything with any errors that the filesystem might We can't do anything with any errors that the filesystem might
encountered, hence the void return type. This will never be called if encountered, hence the void return type. This will never be
the VM is trying to reclaim under GFP_NOFS conditions, hence this called if the VM is trying to reclaim under GFP_NOFS conditions,
method does not need to handle that situation itself. hence this method does not need to handle that situation itself.
Implementations must include conditional reschedule calls inside any Implementations must include conditional reschedule calls inside
scanning loop that is done. This allows the VFS to determine any scanning loop that is done. This allows the VFS to
appropriate scan batch sizes without having to worry about whether determine appropriate scan batch sizes without having to worry
implementations will cause holdoff problems due to large scan batch about whether implementations will cause holdoff problems due to
sizes. large scan batch sizes.
Whoever sets up the inode is responsible for filling in the "i_op" Whoever sets up the inode is responsible for filling in the "i_op"
field. This is a pointer to a "struct inode_operations" which describes field. This is a pointer to a "struct inode_operations" which describes
...@@ -334,23 +369,31 @@ On filesystems that support extended attributes (xattrs), the s_xattr ...@@ -334,23 +369,31 @@ On filesystems that support extended attributes (xattrs), the s_xattr
superblock field points to a NULL-terminated array of xattr handlers. superblock field points to a NULL-terminated array of xattr handlers.
Extended attributes are name:value pairs. Extended attributes are name:value pairs.
``name``: Indicates that the handler matches attributes with the specified name ``name``
(such as "system.posix_acl_access"); the prefix field must be NULL. Indicates that the handler matches attributes with the specified
name (such as "system.posix_acl_access"); the prefix field must
be NULL.
``prefix``: Indicates that the handler matches all attributes with the specified ``prefix``
name prefix (such as "user."); the name field must be NULL. Indicates that the handler matches all attributes with the
specified name prefix (such as "user."); the name field must be
NULL.
``list``: Determine if attributes matching this xattr handler should be listed ``list``
for a particular dentry. Used by some listxattr implementations like Determine if attributes matching this xattr handler should be
generic_listxattr. listed for a particular dentry. Used by some listxattr
implementations like generic_listxattr.
``get``: Called by the VFS to get the value of a particular extended attribute. ``get``
This method is called by the getxattr(2) system call. Called by the VFS to get the value of a particular extended
attribute. This method is called by the getxattr(2) system
call.
``set``: Called by the VFS to set the value of a particular extended attribute. ``set``
When the new value is NULL, called to remove a particular extended Called by the VFS to set the value of a particular extended
attribute. This method is called by the the setxattr(2) and attribute. When the new value is NULL, called to remove a
removexattr(2) system calls. particular extended attribute. This method is called by the the
setxattr(2) and removexattr(2) system calls.
When none of the xattr handlers of a filesystem match the specified When none of the xattr handlers of a filesystem match the specified
attribute name or when a filesystem doesn't support extended attributes, attribute name or when a filesystem doesn't support extended attributes,
...@@ -399,128 +442,147 @@ As of kernel 2.6.22, the following members are defined: ...@@ -399,128 +442,147 @@ As of kernel 2.6.22, the following members are defined:
Again, all methods are called without any locks being held, unless Again, all methods are called without any locks being held, unless
otherwise noted. otherwise noted.
``create``: called by the open(2) and creat(2) system calls. Only ``create``
required if you want to support regular files. The dentry you called by the open(2) and creat(2) system calls. Only required
get should not have an inode (i.e. it should be a negative if you want to support regular files. The dentry you get should
dentry). Here you will probably call d_instantiate() with the not have an inode (i.e. it should be a negative dentry). Here
dentry and the newly created inode you will probably call d_instantiate() with the dentry and the
newly created inode
``lookup``: called when the VFS needs to look up an inode in a parent ``lookup``
called when the VFS needs to look up an inode in a parent
directory. The name to look for is found in the dentry. This directory. The name to look for is found in the dentry. This
method must call d_add() to insert the found inode into the method must call d_add() to insert the found inode into the
dentry. The "i_count" field in the inode structure should be dentry. The "i_count" field in the inode structure should be
incremented. If the named inode does not exist a NULL inode incremented. If the named inode does not exist a NULL inode
should be inserted into the dentry (this is called a negative should be inserted into the dentry (this is called a negative
dentry). Returning an error code from this routine must only dentry). Returning an error code from this routine must only be
be done on a real error, otherwise creating inodes with system done on a real error, otherwise creating inodes with system
calls like create(2), mknod(2), mkdir(2) and so on will fail. calls like create(2), mknod(2), mkdir(2) and so on will fail.
If you wish to overload the dentry methods then you should If you wish to overload the dentry methods then you should
initialise the "d_dop" field in the dentry; this is a pointer initialise the "d_dop" field in the dentry; this is a pointer to
to a struct "dentry_operations". a struct "dentry_operations". This method is called with the
This method is called with the directory inode semaphore held directory inode semaphore held
``link``: called by the link(2) system call. Only required if you want ``link``
to support hard links. You will probably need to call called by the link(2) system call. Only required if you want to
support hard links. You will probably need to call
d_instantiate() just as you would in the create() method d_instantiate() just as you would in the create() method
``unlink``: called by the unlink(2) system call. Only required if you ``unlink``
want to support deleting inodes called by the unlink(2) system call. Only required if you want
to support deleting inodes
``symlink``: called by the symlink(2) system call. Only required if you ``symlink``
want to support symlinks. You will probably need to call called by the symlink(2) system call. Only required if you want
to support symlinks. You will probably need to call
d_instantiate() just as you would in the create() method d_instantiate() just as you would in the create() method
``mkdir``: called by the mkdir(2) system call. Only required if you want ``mkdir``
called by the mkdir(2) system call. Only required if you want
to support creating subdirectories. You will probably need to to support creating subdirectories. You will probably need to
call d_instantiate() just as you would in the create() method call d_instantiate() just as you would in the create() method
``rmdir``: called by the rmdir(2) system call. Only required if you want ``rmdir``
called by the rmdir(2) system call. Only required if you want
to support deleting subdirectories to support deleting subdirectories
``mknod``: called by the mknod(2) system call to create a device (char, ``mknod``
block) inode or a named pipe (FIFO) or socket. Only required called by the mknod(2) system call to create a device (char,
if you want to support creating these types of inodes. You block) inode or a named pipe (FIFO) or socket. Only required if
will probably need to call d_instantiate() just as you would you want to support creating these types of inodes. You will
in the create() method probably need to call d_instantiate() just as you would in the
create() method
``rename``: called by the rename(2) system call to rename the object to ``rename``
have the parent and name given by the second inode and dentry. called by the rename(2) system call to rename the object to have
the parent and name given by the second inode and dentry.
The filesystem must return -EINVAL for any unsupported or The filesystem must return -EINVAL for any unsupported or
unknown flags. Currently the following flags are implemented: unknown flags. Currently the following flags are implemented:
(1) RENAME_NOREPLACE: this flag indicates that if the target (1) RENAME_NOREPLACE: this flag indicates that if the target of
of the rename exists the rename should fail with -EEXIST the rename exists the rename should fail with -EEXIST instead of
instead of replacing the target. The VFS already checks for replacing the target. The VFS already checks for existence, so
existence, so for local filesystems the RENAME_NOREPLACE for local filesystems the RENAME_NOREPLACE implementation is
implementation is equivalent to plain rename. equivalent to plain rename.
(2) RENAME_EXCHANGE: exchange source and target. Both must (2) RENAME_EXCHANGE: exchange source and target. Both must
exist; this is checked by the VFS. Unlike plain rename, exist; this is checked by the VFS. Unlike plain rename, source
source and target may be of different type. and target may be of different type.
``get_link``: called by the VFS to follow a symbolic link to the ``get_link``
inode it points to. Only required if you want to support called by the VFS to follow a symbolic link to the inode it
symbolic links. This method returns the symlink body points to. Only required if you want to support symbolic links.
to traverse (and possibly resets the current position with This method returns the symlink body to traverse (and possibly
nd_jump_link()). If the body won't go away until the inode resets the current position with nd_jump_link()). If the body
is gone, nothing else is needed; if it needs to be otherwise won't go away until the inode is gone, nothing else is needed;
pinned, arrange for its release by having get_link(..., ..., done) if it needs to be otherwise pinned, arrange for its release by
do set_delayed_call(done, destructor, argument). having get_link(..., ..., done) do set_delayed_call(done,
In that case destructor(argument) will be called once VFS is destructor, argument). In that case destructor(argument) will
done with the body you've returned. be called once VFS is done with the body you've returned. May
May be called in RCU mode; that is indicated by NULL dentry be called in RCU mode; that is indicated by NULL dentry
argument. If request can't be handled without leaving RCU mode, argument. If request can't be handled without leaving RCU mode,
have it return ERR_PTR(-ECHILD). have it return ERR_PTR(-ECHILD).
If the filesystem stores the symlink target in ->i_link, the If the filesystem stores the symlink target in ->i_link, the
VFS may use it directly without calling ->get_link(); however, VFS may use it directly without calling ->get_link(); however,
->get_link() must still be provided. ->i_link must not be ->get_link() must still be provided. ->i_link must not be
freed until after an RCU grace period. Writing to ->i_link freed until after an RCU grace period. Writing to ->i_link
post-iget() time requires a 'release' memory barrier. post-iget() time requires a 'release' memory barrier.
``readlink``: this is now just an override for use by readlink(2) for the ``readlink``
this is now just an override for use by readlink(2) for the
cases when ->get_link uses nd_jump_link() or object is not in cases when ->get_link uses nd_jump_link() or object is not in
fact a symlink. Normally filesystems should only implement fact a symlink. Normally filesystems should only implement
->get_link for symlinks and readlink(2) will automatically use ->get_link for symlinks and readlink(2) will automatically use
that. that.
``permission``: called by the VFS to check for access rights on a POSIX-like ``permission``
called by the VFS to check for access rights on a POSIX-like
filesystem. filesystem.
May be called in rcu-walk mode (mask & MAY_NOT_BLOCK). If in rcu-walk May be called in rcu-walk mode (mask & MAY_NOT_BLOCK). If in
mode, the filesystem must check the permission without blocking or rcu-walk mode, the filesystem must check the permission without
storing to the inode. blocking or storing to the inode.
If a situation is encountered that rcu-walk cannot handle, return If a situation is encountered that rcu-walk cannot handle,
return
-ECHILD and it will be called again in ref-walk mode. -ECHILD and it will be called again in ref-walk mode.
``setattr``: called by the VFS to set attributes for a file. This method ``setattr``
is called by chmod(2) and related system calls. called by the VFS to set attributes for a file. This method is
called by chmod(2) and related system calls.
``getattr``: called by the VFS to get attributes of a file. This method
is called by stat(2) and related system calls. ``getattr``
called by the VFS to get attributes of a file. This method is
``listxattr``: called by the VFS to list all extended attributes for a called by stat(2) and related system calls.
given file. This method is called by the listxattr(2) system call.
``listxattr``
``update_time``: called by the VFS to update a specific time or the i_version of called by the VFS to list all extended attributes for a given
an inode. If this is not defined the VFS will update the inode itself file. This method is called by the listxattr(2) system call.
and call mark_inode_dirty_sync.
``update_time``
``atomic_open``: called on the last component of an open. Using this optional called by the VFS to update a specific time or the i_version of
method the filesystem can look up, possibly create and open the file in an inode. If this is not defined the VFS will update the inode
one atomic operation. If it wants to leave actual opening to the itself and call mark_inode_dirty_sync.
caller (e.g. if the file turned out to be a symlink, device, or just
something filesystem won't do atomic open for), it may signal this by ``atomic_open``
returning finish_no_open(file, dentry). This method is only called if called on the last component of an open. Using this optional
the last component is negative or needs lookup. Cached positive dentries method the filesystem can look up, possibly create and open the
are still handled by f_op->open(). If the file was created, file in one atomic operation. If it wants to leave actual
FMODE_CREATED flag should be set in file->f_mode. In case of O_EXCL opening to the caller (e.g. if the file turned out to be a
the method must only succeed if the file didn't exist and hence FMODE_CREATED symlink, device, or just something filesystem won't do atomic
shall always be set on success. open for), it may signal this by returning finish_no_open(file,
dentry). This method is only called if the last component is
``tmpfile``: called in the end of O_TMPFILE open(). Optional, equivalent to negative or needs lookup. Cached positive dentries are still
atomically creating, opening and unlinking a file in given directory. handled by f_op->open(). If the file was created, FMODE_CREATED
flag should be set in file->f_mode. In case of O_EXCL the
method must only succeed if the file didn't exist and hence
FMODE_CREATED shall always be set on success.
``tmpfile``
called in the end of O_TMPFILE open(). Optional, equivalent to
atomically creating, opening and unlinking a file in given
directory.
The Address Space Object The Address Space Object
...@@ -673,70 +735,75 @@ cache in your filesystem. The following members are defined: ...@@ -673,70 +735,75 @@ cache in your filesystem. The following members are defined:
int (*swap_deactivate)(struct file *); int (*swap_deactivate)(struct file *);
}; };
``writepage``: called by the VM to write a dirty page to backing store. ``writepage``
This may happen for data integrity reasons (i.e. 'sync'), or called by the VM to write a dirty page to backing store. This
to free up memory (flush). The difference can be seen in may happen for data integrity reasons (i.e. 'sync'), or to free
wbc->sync_mode. up memory (flush). The difference can be seen in
The PG_Dirty flag has been cleared and PageLocked is true. wbc->sync_mode. The PG_Dirty flag has been cleared and
writepage should start writeout, should set PG_Writeback, PageLocked is true. writepage should start writeout, should set
and should make sure the page is unlocked, either synchronously PG_Writeback, and should make sure the page is unlocked, either
or asynchronously when the write operation completes. synchronously or asynchronously when the write operation
completes.
If wbc->sync_mode is WB_SYNC_NONE, ->writepage doesn't have to If wbc->sync_mode is WB_SYNC_NONE, ->writepage doesn't have to
try too hard if there are problems, and may choose to write out try too hard if there are problems, and may choose to write out
other pages from the mapping if that is easier (e.g. due to other pages from the mapping if that is easier (e.g. due to
internal dependencies). If it chooses not to start writeout, it internal dependencies). If it chooses not to start writeout, it
should return AOP_WRITEPAGE_ACTIVATE so that the VM will not keep should return AOP_WRITEPAGE_ACTIVATE so that the VM will not
calling ->writepage on that page. keep calling ->writepage on that page.
See the file "Locking" for more details. See the file "Locking" for more details.
``readpage``: called by the VM to read a page from backing store. ``readpage``
The page will be Locked when readpage is called, and should be called by the VM to read a page from backing store. The page
unlocked and marked uptodate once the read completes. will be Locked when readpage is called, and should be unlocked
If ->readpage discovers that it needs to unlock the page for and marked uptodate once the read completes. If ->readpage
some reason, it can do so, and then return AOP_TRUNCATED_PAGE. discovers that it needs to unlock the page for some reason, it
In this case, the page will be relocated, relocked and if can do so, and then return AOP_TRUNCATED_PAGE. In this case,
that all succeeds, ->readpage will be called again. the page will be relocated, relocked and if that all succeeds,
->readpage will be called again.
``writepages``: called by the VM to write out pages associated with the
``writepages``
called by the VM to write out pages associated with the
address_space object. If wbc->sync_mode is WBC_SYNC_ALL, then address_space object. If wbc->sync_mode is WBC_SYNC_ALL, then
the writeback_control will specify a range of pages that must be the writeback_control will specify a range of pages that must be
written out. If it is WBC_SYNC_NONE, then a nr_to_write is given written out. If it is WBC_SYNC_NONE, then a nr_to_write is
and that many pages should be written if possible. given and that many pages should be written if possible. If no
If no ->writepages is given, then mpage_writepages is used ->writepages is given, then mpage_writepages is used instead.
instead. This will choose pages from the address space that are This will choose pages from the address space that are tagged as
tagged as DIRTY and will pass them to ->writepage. DIRTY and will pass them to ->writepage.
``set_page_dirty``: called by the VM to set a page dirty. ``set_page_dirty``
This is particularly needed if an address space attaches called by the VM to set a page dirty. This is particularly
private data to a page, and that data needs to be updated when needed if an address space attaches private data to a page, and
a page is dirtied. This is called, for example, when a memory that data needs to be updated when a page is dirtied. This is
mapped page gets modified. called, for example, when a memory mapped page gets modified.
If defined, it should set the PageDirty flag, and the If defined, it should set the PageDirty flag, and the
PAGECACHE_TAG_DIRTY tag in the radix tree. PAGECACHE_TAG_DIRTY tag in the radix tree.
``readpages``: called by the VM to read pages associated with the address_space ``readpages``
object. This is essentially just a vector version of called by the VM to read pages associated with the address_space
readpage. Instead of just one page, several pages are object. This is essentially just a vector version of readpage.
requested. Instead of just one page, several pages are requested.
readpages is only used for read-ahead, so read errors are readpages is only used for read-ahead, so read errors are
ignored. If anything goes wrong, feel free to give up. ignored. If anything goes wrong, feel free to give up.
``write_begin``: ``write_begin``
Called by the generic buffered write code to ask the filesystem to Called by the generic buffered write code to ask the filesystem
prepare to write len bytes at the given offset in the file. The to prepare to write len bytes at the given offset in the file.
address_space should check that the write will be able to complete, The address_space should check that the write will be able to
by allocating space if necessary and doing any other internal complete, by allocating space if necessary and doing any other
housekeeping. If the write will update parts of any basic-blocks on internal housekeeping. If the write will update parts of any
storage, then those blocks should be pre-read (if they haven't been basic-blocks on storage, then those blocks should be pre-read
read already) so that the updated blocks can be written out properly. (if they haven't been read already) so that the updated blocks
can be written out properly.
The filesystem must return the locked pagecache page for the specified The filesystem must return the locked pagecache page for the
offset, in ``*pagep``, for the caller to write into. specified offset, in ``*pagep``, for the caller to write into.
It must be able to cope with short writes (where the length passed to It must be able to cope with short writes (where the length
write_begin is greater than the number of bytes copied into the page). passed to write_begin is greater than the number of bytes copied
into the page).
flags is a field for AOP_FLAG_xxx flags, described in flags is a field for AOP_FLAG_xxx flags, described in
include/linux/fs.h. include/linux/fs.h.
...@@ -744,114 +811,128 @@ cache in your filesystem. The following members are defined: ...@@ -744,114 +811,128 @@ cache in your filesystem. The following members are defined:
A void * may be returned in fsdata, which then gets passed into A void * may be returned in fsdata, which then gets passed into
write_end. write_end.
Returns 0 on success; < 0 on failure (which is the error code), in Returns 0 on success; < 0 on failure (which is the error code),
which case write_end is not called. in which case write_end is not called.
``write_end``: After a successful write_begin, and data copy, write_end must ``write_end``
be called. len is the original len passed to write_begin, and copied After a successful write_begin, and data copy, write_end must be
is the amount that was able to be copied. called. len is the original len passed to write_begin, and
copied is the amount that was able to be copied.
The filesystem must take care of unlocking the page and releasing it
refcount, and updating i_size. The filesystem must take care of unlocking the page and
releasing it refcount, and updating i_size.
Returns < 0 on failure, otherwise the number of bytes (<= 'copied')
that were able to be copied into pagecache. Returns < 0 on failure, otherwise the number of bytes (<=
'copied') that were able to be copied into pagecache.
``bmap``: called by the VFS to map a logical block offset within object to
physical block number. This method is used by the FIBMAP ``bmap``
ioctl and for working with swap-files. To be able to swap to called by the VFS to map a logical block offset within object to
a file, the file must have a stable mapping to a block physical block number. This method is used by the FIBMAP ioctl
device. The swap system does not go through the filesystem and for working with swap-files. To be able to swap to a file,
but instead uses bmap to find out where the blocks in the file the file must have a stable mapping to a block device. The swap
are and uses those addresses directly. system does not go through the filesystem but instead uses bmap
to find out where the blocks in the file are and uses those
``invalidatepage``: If a page has PagePrivate set, then invalidatepage addresses directly.
will be called when part or all of the page is to be removed
from the address space. This generally corresponds to either a ``invalidatepage``
If a page has PagePrivate set, then invalidatepage will be
called when part or all of the page is to be removed from the
address space. This generally corresponds to either a
truncation, punch hole or a complete invalidation of the address truncation, punch hole or a complete invalidation of the address
space (in the latter case 'offset' will always be 0 and 'length' space (in the latter case 'offset' will always be 0 and 'length'
will be PAGE_SIZE). Any private data associated with the page will be PAGE_SIZE). Any private data associated with the page
should be updated to reflect this truncation. If offset is 0 and should be updated to reflect this truncation. If offset is 0
length is PAGE_SIZE, then the private data should be released, and length is PAGE_SIZE, then the private data should be
because the page must be able to be completely discarded. This may released, because the page must be able to be completely
be done by calling the ->releasepage function, but in this case the discarded. This may be done by calling the ->releasepage
release MUST succeed. function, but in this case the release MUST succeed.
``releasepage``: releasepage is called on PagePrivate pages to indicate ``releasepage``
that the page should be freed if possible. ->releasepage releasepage is called on PagePrivate pages to indicate that the
should remove any private data from the page and clear the page should be freed if possible. ->releasepage should remove
PagePrivate flag. If releasepage() fails for some reason, it must any private data from the page and clear the PagePrivate flag.
indicate failure with a 0 return value. If releasepage() fails for some reason, it must indicate failure
releasepage() is used in two distinct though related cases. The with a 0 return value. releasepage() is used in two distinct
first is when the VM finds a clean page with no active users and though related cases. The first is when the VM finds a clean
wants to make it a free page. If ->releasepage succeeds, the page with no active users and wants to make it a free page. If
page will be removed from the address_space and become free. ->releasepage succeeds, the page will be removed from the
address_space and become free.
The second case is when a request has been made to invalidate The second case is when a request has been made to invalidate
some or all pages in an address_space. This can happen some or all pages in an address_space. This can happen through
through the fadvise(POSIX_FADV_DONTNEED) system call or by the the fadvise(POSIX_FADV_DONTNEED) system call or by the
filesystem explicitly requesting it as nfs and 9fs do (when filesystem explicitly requesting it as nfs and 9fs do (when they
they believe the cache may be out of date with storage) by believe the cache may be out of date with storage) by calling
calling invalidate_inode_pages2(). invalidate_inode_pages2(). If the filesystem makes such a call,
If the filesystem makes such a call, and needs to be certain and needs to be certain that all pages are invalidated, then its
that all pages are invalidated, then its releasepage will releasepage will need to ensure this. Possibly it can clear the
need to ensure this. Possibly it can clear the PageUptodate PageUptodate bit if it cannot free private data yet.
bit if it cannot free private data yet.
``freepage``
``freepage``: freepage is called once the page is no longer visible in freepage is called once the page is no longer visible in the
the page cache in order to allow the cleanup of any private page cache in order to allow the cleanup of any private data.
data. Since it may be called by the memory reclaimer, it Since it may be called by the memory reclaimer, it should not
should not assume that the original address_space mapping still assume that the original address_space mapping still exists, and
exists, and it should not block. it should not block.
``direct_IO``: called by the generic read/write routines to perform ``direct_IO``
direct_IO - that is IO requests which bypass the page cache called by the generic read/write routines to perform direct_IO -
and transfer data directly between the storage and the that is IO requests which bypass the page cache and transfer
application's address space. data directly between the storage and the application's address
space.
``isolate_page``: Called by the VM when isolating a movable non-lru page.
If page is successfully isolated, VM marks the page as PG_isolated ``isolate_page``
via __SetPageIsolated. Called by the VM when isolating a movable non-lru page. If page
is successfully isolated, VM marks the page as PG_isolated via
``migrate_page``: This is used to compact the physical memory usage. __SetPageIsolated.
If the VM wants to relocate a page (maybe off a memory card
that is signalling imminent failure) it will pass a new page ``migrate_page``
and an old page to this function. migrate_page should This is used to compact the physical memory usage. If the VM
transfer any private data across and update any references wants to relocate a page (maybe off a memory card that is
that it has to the page. signalling imminent failure) it will pass a new page and an old
page to this function. migrate_page should transfer any private
``putback_page``: Called by the VM when isolated page's migration fails. data across and update any references that it has to the page.
``launder_page``: Called before freeing a page - it writes back the dirty page. To ``putback_page``
prevent redirtying the page, it is kept locked during the whole Called by the VM when isolated page's migration fails.
operation.
``launder_page``
``is_partially_uptodate``: Called by the VM when reading a file through the Called before freeing a page - it writes back the dirty page.
pagecache when the underlying blocksize != pagesize. If the required To prevent redirtying the page, it is kept locked during the
block is up to date then the read can complete without needing the IO whole operation.
to bring the whole page up to date.
``is_partially_uptodate``
``is_dirty_writeback``: Called by the VM when attempting to reclaim a page. Called by the VM when reading a file through the pagecache when
The VM uses dirty and writeback information to determine if it needs the underlying blocksize != pagesize. If the required block is
to stall to allow flushers a chance to complete some IO. Ordinarily up to date then the read can complete without needing the IO to
it can use PageDirty and PageWriteback but some filesystems have bring the whole page up to date.
more complex state (unstable pages in NFS prevent reclaim) or
do not set those flags due to locking problems. This callback ``is_dirty_writeback``
allows a filesystem to indicate to the VM if a page should be Called by the VM when attempting to reclaim a page. The VM uses
treated as dirty or writeback for the purposes of stalling. dirty and writeback information to determine if it needs to
stall to allow flushers a chance to complete some IO.
``error_remove_page``: normally set to generic_error_remove_page if truncation Ordinarily it can use PageDirty and PageWriteback but some
is ok for this address space. Used for memory failure handling. filesystems have more complex state (unstable pages in NFS
prevent reclaim) or do not set those flags due to locking
problems. This callback allows a filesystem to indicate to the
VM if a page should be treated as dirty or writeback for the
purposes of stalling.
``error_remove_page``
normally set to generic_error_remove_page if truncation is ok
for this address space. Used for memory failure handling.
Setting this implies you deal with pages going away under you, Setting this implies you deal with pages going away under you,
unless you have them locked or reference counts increased. unless you have them locked or reference counts increased.
``swap_activate``: Called when swapon is used on a file to allocate ``swap_activate``
space if necessary and pin the block lookup information in Called when swapon is used on a file to allocate space if
memory. A return value of zero indicates success, necessary and pin the block lookup information in memory. A
in which case this file can be used to back swapspace. return value of zero indicates success, in which case this file
can be used to back swapspace.
``swap_deactivate``: Called during swapoff on files where swap_activate ``swap_deactivate``
was successful. Called during swapoff on files where swap_activate was
successful.
The File Object The File Object
...@@ -912,91 +993,120 @@ This describes how the VFS can manipulate an open file. As of kernel ...@@ -912,91 +993,120 @@ This describes how the VFS can manipulate an open file. As of kernel
Again, all methods are called without any locks being held, unless Again, all methods are called without any locks being held, unless
otherwise noted. otherwise noted.
``llseek``: called when the VFS needs to move the file position index ``llseek``
called when the VFS needs to move the file position index
``read``: called by read(2) and related system calls ``read``
called by read(2) and related system calls
``read_iter``: possibly asynchronous read with iov_iter as destination ``read_iter``
possibly asynchronous read with iov_iter as destination
``write``: called by write(2) and related system calls ``write``
called by write(2) and related system calls
``write_iter``: possibly asynchronous write with iov_iter as source ``write_iter``
possibly asynchronous write with iov_iter as source
``iopoll``: called when aio wants to poll for completions on HIPRI iocbs ``iopoll``
called when aio wants to poll for completions on HIPRI iocbs
``iterate``: called when the VFS needs to read the directory contents ``iterate``
called when the VFS needs to read the directory contents
``iterate_shared``: called when the VFS needs to read the directory contents ``iterate_shared``
when filesystem supports concurrent dir iterators called when the VFS needs to read the directory contents when
filesystem supports concurrent dir iterators
``poll``: called by the VFS when a process wants to check if there is ``poll``
called by the VFS when a process wants to check if there is
activity on this file and (optionally) go to sleep until there activity on this file and (optionally) go to sleep until there
is activity. Called by the select(2) and poll(2) system calls is activity. Called by the select(2) and poll(2) system calls
``unlocked_ioctl``: called by the ioctl(2) system call. ``unlocked_ioctl``
called by the ioctl(2) system call.
``compat_ioctl``: called by the ioctl(2) system call when 32 bit system calls ``compat_ioctl``
are used on 64 bit kernels. called by the ioctl(2) system call when 32 bit system calls are
used on 64 bit kernels.
``mmap``: called by the mmap(2) system call ``mmap``
called by the mmap(2) system call
``open``: called by the VFS when an inode should be opened. When the VFS ``open``
called by the VFS when an inode should be opened. When the VFS
opens a file, it creates a new "struct file". It then calls the opens a file, it creates a new "struct file". It then calls the
open method for the newly allocated file structure. You might open method for the newly allocated file structure. You might
think that the open method really belongs in think that the open method really belongs in "struct
"struct inode_operations", and you may be right. I think it's inode_operations", and you may be right. I think it's done the
done the way it is because it makes filesystems simpler to way it is because it makes filesystems simpler to implement.
implement. The open() method is a good place to initialize the The open() method is a good place to initialize the
"private_data" member in the file structure if you want to point "private_data" member in the file structure if you want to point
to a device structure to a device structure
``flush``: called by the close(2) system call to flush a file ``flush``
called by the close(2) system call to flush a file
``release``: called when the last reference to an open file is closed ``release``
called when the last reference to an open file is closed
``fsync``: called by the fsync(2) system call. Also see the section above ``fsync``
called by the fsync(2) system call. Also see the section above
entitled "Handling errors during writeback". entitled "Handling errors during writeback".
``fasync``: called by the fcntl(2) system call when asynchronous ``fasync``
called by the fcntl(2) system call when asynchronous
(non-blocking) mode is enabled for a file (non-blocking) mode is enabled for a file
``lock``: called by the fcntl(2) system call for F_GETLK, F_SETLK, and F_SETLKW ``lock``
commands called by the fcntl(2) system call for F_GETLK, F_SETLK, and
F_SETLKW commands
``get_unmapped_area``: called by the mmap(2) system call ``get_unmapped_area``
called by the mmap(2) system call
``check_flags``: called by the fcntl(2) system call for F_SETFL command ``check_flags``
called by the fcntl(2) system call for F_SETFL command
``flock``: called by the flock(2) system call ``flock``
called by the flock(2) system call
``splice_write``: called by the VFS to splice data from a pipe to a file. This ``splice_write``
called by the VFS to splice data from a pipe to a file. This
method is used by the splice(2) system call method is used by the splice(2) system call
``splice_read``: called by the VFS to splice data from file to a pipe. This ``splice_read``
called by the VFS to splice data from file to a pipe. This
method is used by the splice(2) system call method is used by the splice(2) system call
``setlease``: called by the VFS to set or release a file lock lease. setlease ``setlease``
called by the VFS to set or release a file lock lease. setlease
implementations should call generic_setlease to record or remove implementations should call generic_setlease to record or remove
the lease in the inode after setting it. the lease in the inode after setting it.
``fallocate``: called by the VFS to preallocate blocks or punch a hole. ``fallocate``
called by the VFS to preallocate blocks or punch a hole.
``copy_file_range``: called by the copy_file_range(2) system call.
``copy_file_range``
``remap_file_range``: called by the ioctl(2) system call for FICLONERANGE and called by the copy_file_range(2) system call.
FICLONE and FIDEDUPERANGE commands to remap file ranges. An
implementation should remap len bytes at pos_in of the source file into ``remap_file_range``
the dest file at pos_out. Implementations must handle callers passing called by the ioctl(2) system call for FICLONERANGE and FICLONE
in len == 0; this means "remap to the end of the source file". The and FIDEDUPERANGE commands to remap file ranges. An
return value should the number of bytes remapped, or the usual implementation should remap len bytes at pos_in of the source
negative error code if errors occurred before any bytes were remapped. file into the dest file at pos_out. Implementations must handle
The remap_flags parameter accepts REMAP_FILE_* flags. If callers passing in len == 0; this means "remap to the end of the
REMAP_FILE_DEDUP is set then the implementation must only remap if the source file". The return value should the number of bytes
requested file ranges have identical contents. If REMAP_CAN_SHORTEN is remapped, or the usual negative error code if errors occurred
set, the caller is ok with the implementation shortening the request before any bytes were remapped. The remap_flags parameter
length to satisfy alignment or EOF requirements (or any other reason). accepts REMAP_FILE_* flags. If REMAP_FILE_DEDUP is set then the
implementation must only remap if the requested file ranges have
``fadvise``: possibly called by the fadvise64() system call. identical contents. If REMAP_CAN_SHORTEN is set, the caller is
ok with the implementation shortening the request length to
satisfy alignment or EOF requirements (or any other reason).
``fadvise``
possibly called by the fadvise64() system call.
Note that the file operations are implemented by the specific Note that the file operations are implemented by the specific
filesystem in which the inode resides. When opening a device node filesystem in which the inode resides. When opening a device node
...@@ -1041,89 +1151,104 @@ defined: ...@@ -1041,89 +1151,104 @@ defined:
struct dentry *(*d_real)(struct dentry *, const struct inode *); struct dentry *(*d_real)(struct dentry *, const struct inode *);
}; };
``d_revalidate``: called when the VFS needs to revalidate a dentry. This ``d_revalidate``
is called whenever a name look-up finds a dentry in the called when the VFS needs to revalidate a dentry. This is
dcache. Most local filesystems leave this as NULL, because all their called whenever a name look-up finds a dentry in the dcache.
dentries in the dcache are valid. Network filesystems are different Most local filesystems leave this as NULL, because all their
since things can change on the server without the client necessarily dentries in the dcache are valid. Network filesystems are
being aware of it. different since things can change on the server without the
client necessarily being aware of it.
This function should return a positive value if the dentry is still
valid, and zero or a negative error code if it isn't. This function should return a positive value if the dentry is
still valid, and zero or a negative error code if it isn't.
d_revalidate may be called in rcu-walk mode (flags & LOOKUP_RCU).
If in rcu-walk mode, the filesystem must revalidate the dentry without d_revalidate may be called in rcu-walk mode (flags &
blocking or storing to the dentry, d_parent and d_inode should not be LOOKUP_RCU). If in rcu-walk mode, the filesystem must
used without care (because they can change and, in d_inode case, even revalidate the dentry without blocking or storing to the dentry,
become NULL under us). d_parent and d_inode should not be used without care (because
they can change and, in d_inode case, even become NULL under
If a situation is encountered that rcu-walk cannot handle, return us).
If a situation is encountered that rcu-walk cannot handle,
return
-ECHILD and it will be called again in ref-walk mode. -ECHILD and it will be called again in ref-walk mode.
``_weak_revalidate``: called when the VFS needs to revalidate a "jumped" dentry. ``_weak_revalidate``
This is called when a path-walk ends at dentry that was not acquired by called when the VFS needs to revalidate a "jumped" dentry. This
doing a lookup in the parent directory. This includes "/", "." and "..", is called when a path-walk ends at dentry that was not acquired
as well as procfs-style symlinks and mountpoint traversal. by doing a lookup in the parent directory. This includes "/",
"." and "..", as well as procfs-style symlinks and mountpoint
traversal.
In this case, we are less concerned with whether the dentry is still In this case, we are less concerned with whether the dentry is
fully correct, but rather that the inode is still valid. As with still fully correct, but rather that the inode is still valid.
d_revalidate, most local filesystems will set this to NULL since their As with d_revalidate, most local filesystems will set this to
dcache entries are always valid. NULL since their dcache entries are always valid.
This function has the same return code semantics as d_revalidate. This function has the same return code semantics as
d_revalidate.
d_weak_revalidate is only called after leaving rcu-walk mode. d_weak_revalidate is only called after leaving rcu-walk mode.
``d_hash``: called when the VFS adds a dentry to the hash table. The first ``d_hash``
called when the VFS adds a dentry to the hash table. The first
dentry passed to d_hash is the parent directory that the name is dentry passed to d_hash is the parent directory that the name is
to be hashed into. to be hashed into.
Same locking and synchronisation rules as d_compare regarding Same locking and synchronisation rules as d_compare regarding
what is safe to dereference etc. what is safe to dereference etc.
``d_compare``: called to compare a dentry name with a given name. The first ``d_compare``
called to compare a dentry name with a given name. The first
dentry is the parent of the dentry to be compared, the second is dentry is the parent of the dentry to be compared, the second is
the child dentry. len and name string are properties of the dentry the child dentry. len and name string are properties of the
to be compared. qstr is the name to compare it with. dentry to be compared. qstr is the name to compare it with.
Must be constant and idempotent, and should not take locks if Must be constant and idempotent, and should not take locks if
possible, and should not or store into the dentry. possible, and should not or store into the dentry. Should not
Should not dereference pointers outside the dentry without dereference pointers outside the dentry without lots of care
lots of care (eg. d_parent, d_inode, d_name should not be used). (eg. d_parent, d_inode, d_name should not be used).
However, our vfsmount is pinned, and RCU held, so the dentries and However, our vfsmount is pinned, and RCU held, so the dentries
inodes won't disappear, neither will our sb or filesystem module. and inodes won't disappear, neither will our sb or filesystem
->d_sb may be used. module. ->d_sb may be used.
It is a tricky calling convention because it needs to be called under It is a tricky calling convention because it needs to be called
"rcu-walk", ie. without any locks or references on things. under "rcu-walk", ie. without any locks or references on things.
``d_delete``: called when the last reference to a dentry is dropped and the ``d_delete``
dcache is deciding whether or not to cache it. Return 1 to delete called when the last reference to a dentry is dropped and the
immediately, or 0 to cache the dentry. Default is NULL which means to dcache is deciding whether or not to cache it. Return 1 to
always cache a reachable dentry. d_delete must be constant and delete immediately, or 0 to cache the dentry. Default is NULL
idempotent. which means to always cache a reachable dentry. d_delete must
be constant and idempotent.
``d_init``: called when a dentry is allocated
``d_init``
``d_release``: called when a dentry is really deallocated called when a dentry is allocated
``d_iput``: called when a dentry loses its inode (just prior to its ``d_release``
being deallocated). The default when this is NULL is that the called when a dentry is really deallocated
VFS calls iput(). If you define this method, you must call
iput() yourself ``d_iput``
called when a dentry loses its inode (just prior to its being
``d_dname``: called when the pathname of a dentry should be generated. deallocated). The default when this is NULL is that the VFS
Useful for some pseudo filesystems (sockfs, pipefs, ...) to delay calls iput(). If you define this method, you must call iput()
pathname generation. (Instead of doing it when dentry is created, yourself
it's done only when the path is needed.). Real filesystems probably
dont want to use it, because their dentries are present in global ``d_dname``
dcache hash, so their hash should be an invariant. As no lock is called when the pathname of a dentry should be generated.
held, d_dname() should not try to modify the dentry itself, unless Useful for some pseudo filesystems (sockfs, pipefs, ...) to
appropriate SMP safety is used. CAUTION : d_path() logic is quite delay pathname generation. (Instead of doing it when dentry is
tricky. The correct way to return for example "Hello" is to put it created, it's done only when the path is needed.). Real
at the end of the buffer, and returns a pointer to the first char. filesystems probably dont want to use it, because their dentries
dynamic_dname() helper function is provided to take care of this. are present in global dcache hash, so their hash should be an
invariant. As no lock is held, d_dname() should not try to
modify the dentry itself, unless appropriate SMP safety is used.
CAUTION : d_path() logic is quite tricky. The correct way to
return for example "Hello" is to put it at the end of the
buffer, and returns a pointer to the first char.
dynamic_dname() helper function is provided to take care of
this.
Example : Example :
...@@ -1135,52 +1260,57 @@ defined: ...@@ -1135,52 +1260,57 @@ defined:
dentry->d_inode->i_ino); dentry->d_inode->i_ino);
} }
``d_automount``: called when an automount dentry is to be traversed (optional). ``d_automount``
This should create a new VFS mount record and return the record to the called when an automount dentry is to be traversed (optional).
caller. The caller is supplied with a path parameter giving the This should create a new VFS mount record and return the record
automount directory to describe the automount target and the parent to the caller. The caller is supplied with a path parameter
VFS mount record to provide inheritable mount parameters. NULL should giving the automount directory to describe the automount target
be returned if someone else managed to make the automount first. If and the parent VFS mount record to provide inheritable mount
the vfsmount creation failed, then an error code should be returned. parameters. NULL should be returned if someone else managed to
If -EISDIR is returned, then the directory will be treated as an make the automount first. If the vfsmount creation failed, then
ordinary directory and returned to pathwalk to continue walking. an error code should be returned. If -EISDIR is returned, then
the directory will be treated as an ordinary directory and
If a vfsmount is returned, the caller will attempt to mount it on the returned to pathwalk to continue walking.
mountpoint and will remove the vfsmount from its expiration list in
the case of failure. The vfsmount should be returned with 2 refs on If a vfsmount is returned, the caller will attempt to mount it
it to prevent automatic expiration - the caller will clean up the on the mountpoint and will remove the vfsmount from its
additional ref. expiration list in the case of failure. The vfsmount should be
returned with 2 refs on it to prevent automatic expiration - the
This function is only used if DCACHE_NEED_AUTOMOUNT is set on the caller will clean up the additional ref.
dentry. This is set by __d_instantiate() if S_AUTOMOUNT is set on the
inode being added. This function is only used if DCACHE_NEED_AUTOMOUNT is set on
the dentry. This is set by __d_instantiate() if S_AUTOMOUNT is
``d_manage``: called to allow the filesystem to manage the transition from a set on the inode being added.
dentry (optional). This allows autofs, for example, to hold up clients
waiting to explore behind a 'mountpoint' while letting the daemon go ``d_manage``
past and construct the subtree there. 0 should be returned to let the called to allow the filesystem to manage the transition from a
calling process continue. -EISDIR can be returned to tell pathwalk to dentry (optional). This allows autofs, for example, to hold up
use this directory as an ordinary directory and to ignore anything clients waiting to explore behind a 'mountpoint' while letting
mounted on it and not to check the automount flag. Any other error the daemon go past and construct the subtree there. 0 should be
code will abort pathwalk completely. returned to let the calling process continue. -EISDIR can be
returned to tell pathwalk to use this directory as an ordinary
directory and to ignore anything mounted on it and not to check
the automount flag. Any other error code will abort pathwalk
completely.
If the 'rcu_walk' parameter is true, then the caller is doing a If the 'rcu_walk' parameter is true, then the caller is doing a
pathwalk in RCU-walk mode. Sleeping is not permitted in this mode, pathwalk in RCU-walk mode. Sleeping is not permitted in this
and the caller can be asked to leave it and call again by returning mode, and the caller can be asked to leave it and call again by
-ECHILD. -EISDIR may also be returned to tell pathwalk to returning -ECHILD. -EISDIR may also be returned to tell
ignore d_automount or any mounts. pathwalk to ignore d_automount or any mounts.
This function is only used if DCACHE_MANAGE_TRANSIT is set on the This function is only used if DCACHE_MANAGE_TRANSIT is set on
dentry being transited from. the dentry being transited from.
``d_real``: overlay/union type filesystems implement this method to return one of ``d_real``
the underlying dentries hidden by the overlay. It is used in two overlay/union type filesystems implement this method to return
different modes: one of the underlying dentries hidden by the overlay. It is
used in two different modes:
Called from file_dentry() it returns the real dentry matching the inode Called from file_dentry() it returns the real dentry matching
argument. The real dentry may be from a lower layer already copied up, the inode argument. The real dentry may be from a lower layer
but still referenced from the file. This mode is selected with a already copied up, but still referenced from the file. This
non-NULL inode argument. mode is selected with a non-NULL inode argument.
With NULL inode the topmost real underlying dentry is returned. With NULL inode the topmost real underlying dentry is returned.
...@@ -1195,40 +1325,47 @@ Directory Entry Cache API ...@@ -1195,40 +1325,47 @@ Directory Entry Cache API
There are a number of functions defined which permit a filesystem to There are a number of functions defined which permit a filesystem to
manipulate dentries: manipulate dentries:
``dget``: open a new handle for an existing dentry (this just increments ``dget``
open a new handle for an existing dentry (this just increments
the usage count) the usage count)
``dput``: close a handle for a dentry (decrements the usage count). If ``dput``
close a handle for a dentry (decrements the usage count). If
the usage count drops to 0, and the dentry is still in its the usage count drops to 0, and the dentry is still in its
parent's hash, the "d_delete" method is called to check whether parent's hash, the "d_delete" method is called to check whether
it should be cached. If it should not be cached, or if the dentry it should be cached. If it should not be cached, or if the
is not hashed, it is deleted. Otherwise cached dentries are put dentry is not hashed, it is deleted. Otherwise cached dentries
into an LRU list to be reclaimed on memory shortage. are put into an LRU list to be reclaimed on memory shortage.
``d_drop``: this unhashes a dentry from its parents hash list. A ``d_drop``
subsequent call to dput() will deallocate the dentry if its this unhashes a dentry from its parents hash list. A subsequent
usage count drops to 0 call to dput() will deallocate the dentry if its usage count
drops to 0
``d_delete``: delete a dentry. If there are no other open references to
the dentry then the dentry is turned into a negative dentry ``d_delete``
(the d_iput() method is called). If there are other delete a dentry. If there are no other open references to the
references, then d_drop() is called instead dentry then the dentry is turned into a negative dentry (the
d_iput() method is called). If there are other references, then
``d_add``: add a dentry to its parents hash list and then calls d_drop() is called instead
``d_add``
add a dentry to its parents hash list and then calls
d_instantiate() d_instantiate()
``d_instantiate``: add a dentry to the alias hash list for the inode and ``d_instantiate``
updates the "d_inode" member. The "i_count" member in the add a dentry to the alias hash list for the inode and updates
inode structure should be set/incremented. If the inode the "d_inode" member. The "i_count" member in the inode
pointer is NULL, the dentry is called a "negative structure should be set/incremented. If the inode pointer is
dentry". This function is commonly called when an inode is NULL, the dentry is called a "negative dentry". This function
created for an existing negative dentry is commonly called when an inode is created for an existing
negative dentry
``d_lookup``: look up a dentry given its parent and path name component
It looks up the child of that given name from the dcache ``d_lookup``
hash table. If it is found, the reference count is incremented look up a dentry given its parent and path name component It
and the dentry is returned. The caller must use dput() looks up the child of that given name from the dcache hash
to free the dentry when it finishes using it. table. If it is found, the reference count is incremented and
the dentry is returned. The caller must use dput() to free the
dentry when it finishes using it.
Mount Options Mount Options
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7