NodeMonitor: Added B_WATCH_CHILDREN flag.

* Added a directory argument for notify_{stat/attribute}_changed().
* This allows to watch only a directory, and get the notifications for
all of its files, not just add/remove entry notifications.

nfs4: Don't flush cache when closing read only share reservation

NFS4 specification requires that modified data has to be flushed before
sending CLOSE request. Our original implementation was stricter than that
by flushing all modified data at each file close. That caused problems
with e.g. Tracker, which likes to periodically open files for read only,
in a situation when there were still asynchronous writes pending
(e.g. large file copied).

nfs4: Use nothrow new when creating LockInfo

nfs4: fix few issues related with file caches
* update metadata cache when writing to cache
* do not limit size of a io request
* minor checks in Inode::Write

nfs4: use exponential backoff when the server ask to wait

nfs4: support hard links properly

The main purpose of this patch is to prevent VFS from removing a vnode to early
what might have happened if the NFS client knew it had more than one name but
then one of them was deleted. Moreover, all discovered and still valid names
are stored what may be useful in proper file handle recovery when they are
volatile.

This patch fixes both #9558 and #9561.

nfs4: update vnode objects and cached data when rename overrides a file
* this patch also fixes unnecessary cache invalidation when a file is renamed
without changing its parent directory

nfs4: Fix #9558: in some cases sending SETATTR without open owner is legal

nfs4: Fix CID 991494: memory leak if acquiring a lock fails

nfs4: Fix #9556: wrong error handling in CreateObject

nfs4: Remove check for presence of mandatory attributes

This fixes CID 991495.

nfs4: VFS uses signed offsets while NFS4 expects unsigned

nfs4: Fix write commit race condition

nfs4: Add not null assertions for Inode::fOpenState and fDelegation

nfs4: Remove CacheRevalidator

There is no point in periodically revalidating all existing directory caches.
Directory snapshot can be revalidated when readdir is invoked.

nfs4: Fix problems with reading directory entries
* Inode::ReadDir missed an entry when given buffer was too small
* Each OpenDirCookie now has its own copy of directory snapshot what
would prevent Inode::ReadDir from accessing freed memory when removing
files and reading directory entries simultaneously
* Several minor issues fixed

nfs4: Complete outstanding AIO requests before releasing open state

nfs4: Introduce VnodeToInode mapping

VnodeToInode is another abstraction layer between VFS vnodes and NFS nodes.
It simplifies dealing with race conditions as well as faking some guarantees
VFS expects but NFS does not provide.

nfs4: Fix timespec attribute allocation size

* copying the time attributes was fixed to copy the whole
st_?tim timespec struct but the allocation size was still only
that of st_?time subfield, which is only a time_t.

Signed-off-by: François Revol <revol@free.fr>
Signed-off-by: Pawel Dziepak <pdziepak@quarnos.org>

nfs4: Fix race condition between inode creation and removal

If the underlying file system reuses inode numbers it is possible that
an entry of inode with reused id in InodeIdMap will be removed.
This patch should also improve behavior when one of many hard links
is removed.

nfs4: Add numerous assertion checks

nfs4, dns_resolver: fix numerous style violations

nfs4: Fix some cache revalidator problems

nfs4: Update FileInfo at rename

nfs4: Fix node removal

nfs4: Fix bugs when metadata cache is disabled

nfs4: Fix several problems with emulated attributes

nfs4: Fix file handle recovery

nfs4: Make the client more configurable

nfs4: Fix server reboot recovery

nfs4: Fill struct stat st_blksize and st_blocks fields

nfs4: Add basic support for extended attributes

nfs4: Add attribute directory related hooks

nfs4: Implement node monitoring

nfs4: Check if user is allowed to use open state

nfs4: Return delegation before open upgrade

nfs4: Do not sync too often if delegation is held

nfs4: Reclaim delegations after server reboot

nfs4: Return delegations when asked to

nfs4: Move cache management inside Inode class

nfs4: Use one open state per inode

nfs4: Do not open too much files on server

nfs4: Add ACCESS cache

nfs4: Move low level NFS4 code to Inode's base class

nfs4: Revalidate cache before locks

nfs4: Use existing write cookie, do not send COMMIT if not necessary

nfs4: Basic data cache implementation

nfs4: Fix several issues introduced in previous commits

nfs4: Local creation of node should not invalidate the cache

nfs4: Do not let local changes invalidate cache

nfs4: Add directory and name basic implementation

nfs4: Make Inode::Access() use cached attributes

nfs4: Cache file attributes and filesystem info

nfs4: Rename Filehandle to FileHandle and Filesystem to FileSystem

nfs4: Remove several superfluous 'else'

nfs4: Add few lacking error checks after strdup() or malloc()

nfs4: Fix double free at unmounting

nfs4: Abandon owner and owner_group when id mapping causes problems

nfs4: Support legacy owner and owner_group values

nfs4: Allow WriteStat() to set UID and GID

nfs4: Do not attempt to set owner, owner_group when not supported

nfs4: Add id mapper

nfs4: Move file handle recovery code out of Inode class

nfs4: Use FileInfo instead of filehandle to identify a file

nfs4: Allow the server to release unused lock owners

nfs4: Add proper lock owners management

nfs4: Solve confusion between bigtime_t and seconds

nfs4: vfs makes sure that l_start + l_len <= OFF_MAX

nfs4: Be careful when translating unsigned lock length to signed

nfs4: Enforce checking lock type against open mode

nfs4: Use team_id instead of thread_id as lock owner

nfs4: Put file and dir specific code in separate files

nfs4: Add full support for lease migration

nfs4: Use correct mode in share reclaimation

nfs4: Filesystem is responsible for its opened files

nfs4: Use MutexLocker for locking

nfs4: Take advantage of supported_attrs attribute

nfs4: Check for reace conditions in filehandle recovery

nfs4: close() should also wake snoozing threads

When server returns error codes like LOCKED, DEALY, etc the client waits
some time before retrying (unless in O_NONBLOCK mode), close() should wake
snoozing threads.

nfs4: Verify inode number before opening a file

nfs4: Fix LookUp returning wrong error code if file doesn't exist

nfs4: Add support for file locking

nfs4: Add support for O_NONBLOCK mode

nfs4: Read from st_*tim instead of st_*time

nfs4: Wrong filehandle used when creating a file

nfs4: Add write_stat() hook

nfs4: Partial handling of NFS4ERR_LEASE_MOVED error

nfs4: Fix removing nodes other than NF4DIR or NF4REG

nfs4: Add create_dir() hook

nfs4: Improve handling errors in NFS replies

nfs4: Put all NFS error handling code in one function

nfs4: Add support for O_TRUNC mode

nfs4: Add support for O_APPEND mode

nfs4: Fix cookie and vnode creation in create() hook

nfs4: Add write() hook

nfs4: Add create() hook

nfs4: Add create_symlink() hook

nfs4: Add link() hook

nfs4: unlink() and remove_dir() should check node type

nfs4: Add unlink() and remove_dir() hooks

nfs4: Add rename() hook

nfs4: close*() should wake all blocked operations

nfs4: Add support for server migration

nfs4: Fix ReadDir() bug when skipping inodes

nfs4: Forbid accessing multiple NFS4 filesystems from one local mount

Haiku assumes that a pair dev_t:ino_t uniquely identifies a file. That may not be the case if we allow multiple filesystems share one dev_t.

nfs4: Fix OPEN and CLOSE sequence numbers

nfs4: Add support for volatile filehandles

nfs4: Let opened files survive server reboot

nfs4: Behave correctly if server is in grace period

nfs4: Add read_symlink() hook

nfs4: Add per server NFS4 objects

RPC::Server is now able to store RPC programs data.
NFS4Server objects are currently used mainly for managing client id.

nfs4: Let stat() retrieve access, change, etc dates

nfs4: Fix timeout handling code

nfs4: Use ACCESS if server does not provide file mode

nfs4: Partial support for servers not providing FileId

nfs4: Add Request class for better handling of NFS4 requests

nfs4: Add access() hook

nfs4: Add close() and free_cookie() hooks

nfs4: Add read() hook

nfs4: Add open() hook

nfs4: Inodes should know their name and parent's filehandle

nfs4: readdir() should not read more than requested

Additionally, file attribute decoding checks more strictly whether
received data are sane.

nfs4: Do not store pointer to parent in inodes

nfs4: Add lookup() and get_vnode() hooks

get_vnode() needs a way to retrieve filehandle knowing only inode number.
Since NFS4 sever does not provide such services client manages its own mapping
between ino_t and filehandles.

nfs4: Add read_dir() hook

nfs4: Add open_dir, close_dir and free_dir_cookie hooks

nfs4: Add basic nfs4_read_stat() procedure

Not all data are retrieved and there is a room for improvement in handling
situations when server does not provide necessary information.

nfs4: Implement nfs4_mount() procedure

nfs4: Add stub Inode class, fix and extend Filesystem class

Additionally PUTFH support is added to both RequestBuilder and
ReplyInterpreter.

nfs4: Don't flush cache when closing read only share reservation

NFS4 specification requires that modified data has to be flushed before
sending CLOSE request. Our original implementation was stricter than that
by flushing all modified data at each file close. That caused problems
with e.g. Tracker, which likes to periodically open files for read only,
in a situation when there were still asynchronous writes pending
(e.g. large file copied).

nfs4: Use nothrow new when creating LockInfo

nfs4: fix few issues related with file caches
* update metadata cache when writing to cache
* do not limit size of a io request
* minor checks in Inode::Write

nfs4: use exponential backoff when the server ask to wait

nfs4: support hard links properly

The main purpose of this patch is to prevent VFS from removing a vnode to early
what might have happened if the NFS client knew it had more than one name but
then one of them was deleted. Moreover, all discovered and still valid names
are stored what may be useful in proper file handle recovery when they are
volatile.

This patch fixes both #9558 and #9561.

nfs4: update vnode objects and cached data when rename overrides a file
* this patch also fixes unnecessary cache invalidation when a file is renamed
without changing its parent directory

nfs4: Fix #9558: in some cases sending SETATTR without open owner is legal

nfs4: Fix CID 991494: memory leak if acquiring a lock fails

nfs4: Fix #9556: wrong error handling in CreateObject

nfs4: Remove check for presence of mandatory attributes

This fixes CID 991495.

nfs4: VFS uses signed offsets while NFS4 expects unsigned

nfs4: Fix write commit race condition

nfs4: Add not null assertions for Inode::fOpenState and fDelegation

nfs4: Remove CacheRevalidator

There is no point in periodically revalidating all existing directory caches.
Directory snapshot can be revalidated when readdir is invoked.

nfs4: Fix problems with reading directory entries
* Inode::ReadDir missed an entry when given buffer was too small
* Each OpenDirCookie now has its own copy of directory snapshot what
would prevent Inode::ReadDir from accessing freed memory when removing
files and reading directory entries simultaneously
* Several minor issues fixed

nfs4: Complete outstanding AIO requests before releasing open state

nfs4: Introduce VnodeToInode mapping

VnodeToInode is another abstraction layer between VFS vnodes and NFS nodes.
It simplifies dealing with race conditions as well as faking some guarantees
VFS expects but NFS does not provide.

nfs4: Fix timespec attribute allocation size

* copying the time attributes was fixed to copy the whole
st_?tim timespec struct but the allocation size was still only
that of st_?time subfield, which is only a time_t.

Signed-off-by: François Revol <revol@free.fr>
Signed-off-by: Pawel Dziepak <pdziepak@quarnos.org>

nfs4: Fix race condition between inode creation and removal

If the underlying file system reuses inode numbers it is possible that
an entry of inode with reused id in InodeIdMap will be removed.
This patch should also improve behavior when one of many hard links
is removed.

nfs4: Add numerous assertion checks

nfs4, dns_resolver: fix numerous style violations

nfs4: Fix some cache revalidator problems

nfs4: Update FileInfo at rename

nfs4: Fix node removal

nfs4: Fix bugs when metadata cache is disabled

nfs4: Fix several problems with emulated attributes

nfs4: Fix file handle recovery

nfs4: Make the client more configurable

nfs4: Fix server reboot recovery

nfs4: Fill struct stat st_blksize and st_blocks fields

nfs4: Add basic support for extended attributes

nfs4: Add attribute directory related hooks

nfs4: Implement node monitoring

nfs4: Check if user is allowed to use open state

nfs4: Return delegation before open upgrade

nfs4: Do not sync too often if delegation is held

nfs4: Reclaim delegations after server reboot

nfs4: Return delegations when asked to

nfs4: Move cache management inside Inode class

nfs4: Use one open state per inode

nfs4: Do not open too much files on server

nfs4: Add ACCESS cache

nfs4: Move low level NFS4 code to Inode's base class

nfs4: Revalidate cache before locks

nfs4: Use existing write cookie, do not send COMMIT if not necessary

nfs4: Basic data cache implementation

nfs4: Fix several issues introduced in previous commits

nfs4: Local creation of node should not invalidate the cache

nfs4: Do not let local changes invalidate cache

nfs4: Add directory and name basic implementation

nfs4: Make Inode::Access() use cached attributes

nfs4: Cache file attributes and filesystem info

nfs4: Rename Filehandle to FileHandle and Filesystem to FileSystem

nfs4: Remove several superfluous 'else'

nfs4: Add few lacking error checks after strdup() or malloc()

nfs4: Fix double free at unmounting

nfs4: Abandon owner and owner_group when id mapping causes problems

nfs4: Support legacy owner and owner_group values

nfs4: Allow WriteStat() to set UID and GID

nfs4: Do not attempt to set owner, owner_group when not supported

nfs4: Add id mapper

nfs4: Move file handle recovery code out of Inode class

nfs4: Use FileInfo instead of filehandle to identify a file

nfs4: Allow the server to release unused lock owners

nfs4: Add proper lock owners management

nfs4: Solve confusion between bigtime_t and seconds

nfs4: vfs makes sure that l_start + l_len <= OFF_MAX

nfs4: Be careful when translating unsigned lock length to signed

nfs4: Enforce checking lock type against open mode

nfs4: Use team_id instead of thread_id as lock owner

nfs4: Put file and dir specific code in separate files

nfs4: Add full support for lease migration

nfs4: Use correct mode in share reclaimation

nfs4: Filesystem is responsible for its opened files

nfs4: Use MutexLocker for locking

nfs4: Take advantage of supported_attrs attribute

nfs4: Check for reace conditions in filehandle recovery

nfs4: close() should also wake snoozing threads

When server returns error codes like LOCKED, DEALY, etc the client waits
some time before retrying (unless in O_NONBLOCK mode), close() should wake
snoozing threads.

nfs4: Verify inode number before opening a file

nfs4: Fix LookUp returning wrong error code if file doesn't exist

nfs4: Add support for file locking

nfs4: Add support for O_NONBLOCK mode

nfs4: Read from st_*tim instead of st_*time

nfs4: Wrong filehandle used when creating a file

nfs4: Add write_stat() hook

nfs4: Partial handling of NFS4ERR_LEASE_MOVED error

nfs4: Fix removing nodes other than NF4DIR or NF4REG

nfs4: Add create_dir() hook

nfs4: Improve handling errors in NFS replies

nfs4: Put all NFS error handling code in one function

nfs4: Add support for O_TRUNC mode

nfs4: Add support for O_APPEND mode

nfs4: Fix cookie and vnode creation in create() hook

nfs4: Add write() hook

nfs4: Add create() hook

nfs4: Add create_symlink() hook

nfs4: Add link() hook

nfs4: unlink() and remove_dir() should check node type

nfs4: Add unlink() and remove_dir() hooks

nfs4: Add rename() hook

nfs4: close*() should wake all blocked operations

nfs4: Add support for server migration

nfs4: Fix ReadDir() bug when skipping inodes

nfs4: Forbid accessing multiple NFS4 filesystems from one local mount

Haiku assumes that a pair dev_t:ino_t uniquely identifies a file. That may not be the case if we allow multiple filesystems share one dev_t.

nfs4: Fix OPEN and CLOSE sequence numbers

nfs4: Add support for volatile filehandles

nfs4: Let opened files survive server reboot

nfs4: Behave correctly if server is in grace period

nfs4: Add read_symlink() hook

nfs4: Add per server NFS4 objects

RPC::Server is now able to store RPC programs data.
NFS4Server objects are currently used mainly for managing client id.

nfs4: Let stat() retrieve access, change, etc dates

nfs4: Fix timeout handling code

nfs4: Use ACCESS if server does not provide file mode

nfs4: Partial support for servers not providing FileId

nfs4: Add Request class for better handling of NFS4 requests

nfs4: Add access() hook

nfs4: Add close() and free_cookie() hooks

nfs4: Add read() hook

nfs4: Add open() hook

nfs4: Inodes should know their name and parent's filehandle

nfs4: readdir() should not read more than requested

Additionally, file attribute decoding checks more strictly whether
received data are sane.

nfs4: Do not store pointer to parent in inodes

nfs4: Add lookup() and get_vnode() hooks

get_vnode() needs a way to retrieve filehandle knowing only inode number.
Since NFS4 sever does not provide such services client manages its own mapping
between ino_t and filehandles.

nfs4: Add read_dir() hook

nfs4: Add open_dir, close_dir and free_dir_cookie hooks

nfs4: Add basic nfs4_read_stat() procedure

Not all data are retrieved and there is a room for improvement in handling
situations when server does not provide necessary information.

nfs4: Implement nfs4_mount() procedure

nfs4: Add stub Inode class, fix and extend Filesystem class

Additionally PUTFH support is added to both RequestBuilder and
ReplyInterpreter.