NetBSD/lib/libpuffs/puffs_ops.3
2007-05-18 14:38:15 +00:00

691 lines
22 KiB
Groff

.\" $NetBSD: puffs_ops.3,v 1.6 2007/05/18 14:38:15 pooka Exp $
.\"
.\" Copyright (c) 2007 Antti Kantee. All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.Dd May 18, 2007
.Dt PUFFS_OPS 3
.Os
.Sh NAME
.Nm puffs_ops
.Nd puffs callback operations
.Sh LIBRARY
.Lb libpuffs
.Sh SYNOPSIS
.In puffs.h
.Ft int
.Fo puffs_fs_statvfs
.Fa "struct puffs_cc *pcc" "struct statvfs *sbp" "pid_t pid"
.Fc
.Ft int
.Fo puffs_fs_sync
.Fa "struct puffs_cc *pcc" "int waitfor" "const struct puffs_cred *pcr"
.Fa "pid_t pid"
.Fc
.Ft int
.Fo puffs_fs_fhtonode
.Fa "struct puffs_cc *pcc" "void *fid" "size_t fidsize"
.Fa "void **fcookie" "enum vtype *ftype" "voff_t *fsize" "dev_t *fdev"
.Fc
.Ft int
.Fo puffs_fs_nodetofh
.Fa "struct puffs_cc *pcc" "void *cookie" "void *fid" "size_t *fidsize"
.Fc
.Ft void
.Fn puffs_fs_suspend "struct puffs_cc *pcc" "int status"
.Ft int
.Fn puffs_fs_unmount "struct puffs_cc *pcc" "int flags" "pid_t pid"
.Ft int
.Fo puffs_node_lookup
.Fa "struct puffs_cc *pcc" "void *opc" "void **newnode"
.Fa "enum vtype *newtype" "voff_t *newsize" "dev_t *newrdev"
.Fa "const struct puffs_cn *pcn"
.Fc
.Ft int
.Fo puffs_node_create
.Fa "struct puffs_cc *pcc" "void *opc" "void **newnode"
.Fa "const struct puffs_cn *pcn" "const struct vattr *va"
.Fc
.Ft int
.Fo puffs_node_mknod
.Fa "struct puffs_cc *pcc" "void *opc" "void **newnode"
.Fa "const struct puffs_cn *pcn" "const struct vattr *va"
.Fc
.Ft int
.Fo puffs_node_open
.Fa "struct puffs_cc *pcc" "void *opc" "int mode"
.Fa "const struct puffs_cred *pcr" "pid_t pid"
.Fc
.Ft int
.Fo puffs_node_close
.Fa "struct puffs_cc *pcc" "void *opc" "int flags"
.Fa "const struct puffs_cred *pcr" "pid_t pid"
.Fc
.Ft int
.Fo puffs_node_access
.Fa "struct puffs_cc *pcc" "void *opc" "int mode"
.Fa "const struct puffs_cred *pcr" "pid_t pid"
.Fc
.Ft int
.Fo puffs_node_getattr
.Fa "struct puffs_cc *pcc" "void *opc" "struct vattr *va"
.Fa "const struct puffs_cred *pcr" "pid_t pid"
.Fc
.Ft int
.Fo puffs_node_setattr
.Fa "struct puffs_cc *pcc" "void *opc" "const struct vattr *va"
.Fa "const struct puffs_cred *pcr" "pid_t pid"
.Fc
.Ft int
.Fo puffs_node_poll
.Fa "struct puffs_cc *pcc" "void *opc" "int *events" "pid_t pid"
.Fc
.Ft int
.Fo puffs_node_mmap
.Fa "struct puffs_cc *pcc" "void *opc" "int flags"
.Fa "const struct puffs_cred *pcr" "pid_t pid"
.Fc
.Ft int
.Fo puffs_node_fsync
.Fa "struct puffs_cc *pcc" "void *opc" "int flags" "off_t offlo"
.Fa "off_t offhi" "pid_t pid"
.Fc
.\".Ft int
.\".Fo puffs_node_seek
.\".Fa "struct puffs_cc *pcc" "void *opc" "off_t oldoff" "off_t "newoff"
.\".Fa "pid_t pid"
.\".Fc
.Ft int
.Fo puffs_node_remove
.Fa "struct puffs_cc *pcc" "void *opc" "void *targ"
.Fa "const struct puffs_cn *pcn"
.Fc
.Ft int
.Fo puffs_node_link
.Fa "struct puffs_cc *pcc" "void *opc" "void *targ"
.Fa "const struct puffs_cn *pcn"
.Fc
.Ft int
.Fo puffs_node_rename
.Fa "struct puffs_cc *pcc" "void *opc" "void *src"
.Fa "const struct puffs_cn *pcn_src" "void *targ_dir" "void *targ"
.Fa "const struct puffs_cn *pcn_targ"
.Fc
.Ft int
.Fo puffs_node_mkdir
.Fa "struct puffs_cc *pcc" "void *opc" "void **newnode"
.Fa "const struct puffs_cn *pcn" "const struct vattr *va"
.Fc
.Ft int
.Fo puffs_node_rmdir
.Fa "struct puffs_cc *pcc" "void *opc" "void *targ"
.Fa "const struct puffs_cn *pcn"
.Fc
.Ft int
.Fo puffs_node_symlink
.Fa "struct puffs_cc *pcc" "void *opc" "void **newnode"
.Fa "const struct puffs_cn *pcn_src" "const struct vattr *va"
.Fa "const char *link_target"
.Fc
.Ft int
.Fo puffs_node_readdir
.Fa "struct puffs_cc *pcc" "void *opc" "struct dirent *dent"
.Fa "off_t *readoff" "size_t *reslen" "const struct puffs_cred *pcr"
.Fa "int *eofflag" "off_t *cookies" "size_t *ncookies"
.Fc
.Ft int
.Fo puffs_node_readlink
.Fa "struct puffs_cc *pcc" "void *opc" "const struct puffs_cred *pcr"
.Fa "char *link" "size_t *linklen"
.Fc
.Ft int
.Fo puffs_node_read
.Fa "struct puffs_cc *pcc" "void *opc" "uint8_t *buf"
.Fa "off_t offset" "size_t *resid" "const struct puffs_cred *pcr" "int ioflag"
.Fc
.Ft int
.Fo puffs_node_write
.Fa "struct puffs_cc *pcc" "void *opc" "uint8_t *buf"
.Fa "off_t offset" "size_t *resid" "const struct puffs_cred *pcr" "int ioflag"
.Fc
.Ft int
.Fn puffs_node_print "struct puffs_cc *pcc" "void *opc"
.Ft int
.Fn puffs_node_reclaim "struct puffs_cc *pcc" "void *opc" "pid_t pid"
.Ft int
.Fo puffs_node_inactive
.Fa "struct puffs_cc *pcc" "void *opc" "pid_t pid" "int *refcount"
.Fc
.Ft void
.Fn puffs_setback "struct puffs_cc *pcc" "int op"
.Sh DESCRIPTION
.Em IMPORTANT NOTE!
This document describes interfaces which are not yet guaranteed to be
stable.
In case you update your system sources, please recompile everything
and fix complation errors.
If your sources are out-of-sync, incorrect operation may result.
The interfaces in this document will most likely be hugely simplified
in later versions or made transparent to the implementation.
.Pp
The operations
.Nm puffs
requires to function can be divided into two categories: file system
callbacks and node callbacks.
The former affect the entire file system while the latter are targeted
at a file or a directory and a file.
They are roughly equivalent to the vfs and vnode operations in the
kernel.
.Pp
All callbacks can be prototyped with the file system name and operation
name using the macro
.Fn PUFFSOP_PROTOS fsname .
.Ss File system callbacks (puffs_fs)
.Bl -tag -width xxxx
.It Fn puffs_fs_statvfs "pcc" "sbp" "pid"
The following fields of the argument
.Fa sbp
need to be filled:
.Bd -literal
* unsigned long f_bsize; file system block size
* unsigned long f_frsize; fundamental file system block size
* fsblkcnt_t f_blocks; number of blocks in file system,
* (in units of f_frsize)
*
* fsblkcnt_t f_bfree; free blocks avail in file system
* fsblkcnt_t f_bavail; free blocks avail to non-root
* fsblkcnt_t f_bresvd; blocks reserved for root
* fsfilcnt_t f_files; total file nodes in file system
* fsfilcnt_t f_ffree; free file nodes in file system
* fsfilcnt_t f_favail; free file nodes avail to non-root
* fsfilcnt_t f_fresvd; file nodes reserved for root
.Ed
The process requiring this information is given by
.Fa pid .
.It Fn puffs_fs_sync "pcc" "waitfor" "pcr" "pid"
All the dirty buffers that have been cached at the file server
level including metadata should be committed to stable storage.
The
.Fa waitfor
parameter affects the operation.
Possible values are:
.Bl -tag -width XMNT_NOWAITX
.It Dv MNT_WAIT
Wait for all I/O for complete until returning.
.It Dv MNT_NOWAIT
Initiate I/O, but do not wait for completion.
.It Dv MNT_LAZY
Synchorize data not synchoronized by the file system syncer,
i.e. data not written when
.Fn node_fsync
is called with
.Dv FSYNC_LAZY .
.El
.Pp
The credentials for the initiator of the sync operation are present in
.Fa pcr
and will usually be either file system or kernel credentials, but might
also be user credentials.
However, most of the time it is advisable to sync regardless of the
credentials of the caller.
.It Fn puffs_fs_fhtonode "pcc" "fid" "fidsize" "fcookie" "ftype" "fsize" "fdev"
Translates a file handle
.Fa fid
to a node.
The parameter
.Fa fidsize
indicates how large the file handle is.
In case the file system's handles are static length, this parameter can
be ignored as the kernel guarantees all file handles passed to the file
server are of correct length.
For dynamic length handles the field should be examined and
.Er EINVAL
returned in case the file handle length is not correct.
.Pp
This function provides essentially the same information to the kernel as
.Fn puffs_node_lookup .
The information is necessary for creating a new vnode corresponding to
the file handle.
.It Fn puffs_fs_nodetofh "pcc" "cookie" "fid" "fidsize"
Create a file handle from the node described by
.Fa cookie .
The file handle should contain enough information to reliably identify
the node even after reboots and the pathname/inode being replaced by
another file.
If this is not possible, it is up to the author of the file system to
act responsibly and decide if the file system can support file handles
at all.
.Pp
For file systems which want dynamic length file handles, this function
must check if the file handle space indicated by
.Fa fidsize
is large enough to accommodate the file handle for the node.
If not, it must fill in the correct size and return
.Er E2BIG .
In either case, the handle length should be supplied to the kernel in
.Fa fidsize .
File systems with static length handles can ignore the size parameter as
the kernel always supplies the correct size buffer.
.It Fn puffs_fs_suspend "pcc" "status"
Called when file system suspension reaches various phases.
See
.Xr puffs_suspend 3
for more information.
.It Fn puffs_fs_unmount "pcc" "flags" "pid"
Unmount the file system.
The kernel has assumedly flushed all cached data when this callback
is executed.
If the file system cannot currently be safely be unmounted, for whatever
reason, the kernel will honor an error value and not forcibly unmount.
However, if the flag
.Dv MNT_FORCE
is not honored by the file server, the kernel will forcibly unmount
the file system.
.El
.Ss Node callbacks
These operations operate in the level of individual files.
The file cookie is always provided as the second argument
.Fa opc .
If the operation is for a file, it will be the cookie of the file.
The case the operation involves a directory (such as
.Dq create file in directory ) ,
the cookie will be for the directory.
Some operations take additional cookies to describe the rest of
the operands.
The return value 0 signals success, else an appropriate errno value
should be returned.
Please note that neither this list nor the descriptions are complete.
.Bl -tag -width xxxx
.It Fn puffs_node_lookup "pcc" "opc" "newnode" "newtype" "newsize" "newrdev" "pcn"
This function is used to locate nodes, or in other words translate
pathname components to file system data structures.
The implementation should match the name in
.Fa pcn
against the existing entries in the directory provided by the cookie
.Fa opc .
If found, the cookie for the located node should be returned in
.Fa newnode .
Additionally, the type and size (latter applicable to regular files only)
should be returned in
.Fa newtype
and
.Fa newsize ,
respectively.
If the located entry is a block device or character device file,
the dev_t for the entry should be returned in
.Fa newrdev .
.Pp
The type of operation is found from
.Va pcn->pcn_nameiop :
.Bl -tag -width XPUFFSLOOKUP_LOOKUPX
.It Dv PUFFSLOOKUP_LOOKUP
Normal lookup operation.
.It Dv PUFFSLOOKUP_CREATE
Lookup to create a node.
.It Dv PUFFSLOOKUP_DELETE
Lookup for node deletion.
.It Dv PUFFSLOOKUP_RENAME
Lookup for the target of a rename operation (source will be looked
up using
.Dv PUFFSLOOKUP_DELETE ).
.El
.Pp
The final component from a pathname lookup usually requires special
treatment.
It can be identified by looking at the
.Va pcn->pcn_flags
fields for the flag
.Dv PUFFSLOOKUP_ISLASTCN .
For example, in most cases the lookup operation will want to check if
a delete, rename or create operation has enough credentials to perform
the operation.
.Pp
The return value 0 signals a found node and a nonzero value signals
an errno.
As a special case,
.Er ENOENT
signals "success" for cases where the lookup operation is
.Dv PUFFSLOOKUP_CREATE
or
.Dv PUFFSLOOKUP_RENAME .
Failure in these cases can be signalled by returning another appropriate
error code, for example
.Er EACCESS .
.It Fn puffs_node_create "pcc" "opc" "newnode" "pcn" "va"
.It Fn puffs_node_mkdir "pcc" "opc" "newnode" "pcn" "va"
.It Fn puffs_node_mknod "pcc" "opc" "newnode" "pcn" "va"
A file node is created in the directory denoted by the cookie
.Fa opc
by any of the above callbacks.
The name of the new file can be found from
.Fa pcn
and the attributes are specified by
.Fa va
and the cookie for the newly created node should be returned in
.Fa newnode .
The only difference between these three is that they create a regular
file, directory and device special file, respectively.
.Pp
In case of mknod, the device identifier can be found in
.Fa va-\*[Gt]va_rdev .
.It Fn puffs_node_open "pcc" "opc" "mode" "pcr" "pid"
Open the node denoted by the cookie
.Fa opc .
The parameter
.Fa mode
specifies the flags that
.Xr open 2
was called with, e.g.
.Dv O_APPEND
and
.Dv O_NONBLOCK .
.It Fn puffs_node_close "pcc" "opc" "flags" "pcr" "pid"
Close a node.
The parameter
.Fa flags
parameter describes the flags that the file was opened with.
.It Fn puffs_node_access "pcc" "opc" "mode" "pcr" "pid"
Check if the credentials of
.Fa pcr
have the right to perform the operation specified by
.Fa mode
onto the node
.Fa opc .
The argument
.Fa mode
can specify read, write or execute by
.Dv PUFFS_VREAD ,
.Dv PUFFS_VWRITE
and
.Dv PUFFS_VEXEC ,
respectively.
.It Fn puffs_node_getattr "pcc" "opc" "va" "pcr" "pid"
The attributes of the node specified by
.Fa opc
must be copied to the space pointed by
.Fa va .
.It Fn puffs_node_setattr "pcc" "opc" "va" "pcr" "pid"
The attributes for the node specified by
.Fa opc
must be set to those contained in
.Fa va .
Only fields of
.Fa va
which contain a value different from
.Dv PUFFS_VNOVAL
(typecast to the field's type!) contain a valid value.
.It Fn puffs_node_poll pcc opc events pid
Poll for events on node
.Ar opc .
If
.Xr poll 2
events specified in
.Ar events
are available, the function should set the bitmask to match available
events and return immediately.
Otherwise, the function should block (yield) until some events in
.Ar events
become available and only then set the
.Ar events
bitmask and return.
.Pp
In case this function returns an error,
.Dv POLLERR
(or it's
.Xr select 2
equivalent) will be delivered to the calling process.
.It Fn puffs_node_mmap "pcc" "opc" "flags" "pcr" "pid"
Called when a regular file is being memory mapped by
.Xr mmap 2 .
.Fa flags
is currently always 0.
.It Fn puffs_node_fsync "pcc" "opc" "flags" "offlo" "offhi" "pid"
Sychronize a node's contents onto stable storage.
This is necessary only if the file server caches some information
before committing it.
The parameter
.Fa flags
specifies the minimum level of sychronization required (XXX: they are
not yet available).
The parameters
.Fa offlo
and
.Fa offhi
specify the data offsets requiring to be synced.
A high offset of 0 means sync from
.Fa offlo
to the end of the file.
.It Fn puffs_node_remove "pcc" "opc" "targ" "pcn"
.It Fn puffs_node_rmdir "pcc" "opc" "targ" "pcn"
Remove the node
.Fa targ
from the directory indicated by
.Fa opc .
The directory entry name to be removed is provided by
.Fa pcn .
The rmdir operation removes only directories, while the remove
operation removes all other types except directories.
.Pp
It is paramount to note that the file system may not remove the
node data structures at this point, only the directory entry and prevent
lookups from finding the node again.
This is to retain the
.Ux
open file semantics.
The data may be removed only when
.Fn puffs_node_reclaim
is called for the node, as this assures there are no further users.
.It Fn puffs_node_link "pcc" "opc" "targ" "pcn"
Create a hard link for the node
.Fa targ
into the directory
.Fa opc .
The argument
.Fa pcn
provides the directory entry name for the new link.
.It Fn puffs_node_rename "pcc" "src_dir" "src" "pcn_src" "targ_dir" "targ" "pcn_targ"
Rename the node
.Fa src
with the name specified by
.Fa pcn_src
from the directory
.Fa src_dir .
The target directory and target name are given by
.Fa targ_dir
and
.Fa pcn_targ ,
respectively.
.B If
the target node already exists, it is specified by
.Fa targ
and must be replaced atomically.
Otherwise
.Fa targ
is gives as
.Dv NULL .
.Pp
It is legal to replace a directory node by another directory node with
the means of rename if the target directory is empty, otherwise
.Er ENOTEMPTY
should be returned.
All other types can replace all other types.
In case a rename between incompatible types is attempted, the errors
.Er ENOTDIR
or
.Er EISDIR
should be returned, depending on the target type.
.It Fn puffs_node_symlink "pcc" "opc" "newnode" "pcn_src" "va" "link_target"
Create a symbolic link into the directory
.Fa opc
with the name in
.Fa pcn_src
and the initial attributes in
.Fa va .
The link's target is
.Fa link_target
and the created node cookie should be returned in
.Fa newnode .
.It Fn puffs_node_readdir "pcc" "opc" "dent" "readoff" "reslen" "pcr" "eofflag" "cookies" "ncookies"
To read directory entries,
.Fn puffs_node_readdir
is called.
It should store directories as
.Va struct dirent
in the space pointed to by
.Fa dent .
The amount of space available is given by
.Fa reslen
and before returning it should be set to the amount of space
.Em remaining
in the buffer.
The argument
.Fa offset
is used to specify the offset to the directory.
Its intepretation is up to the file system and it should be set to
signal the continuation point when there is no more room for the next
entry in
.Fa dent .
It is most performant to return the maximal amount of directory
entries each call.
It is easiest to generate directory entries by using
.Fn puffs_nextdent ,
which also automatically advances the necessary pointers.
.Pp
In case end-of-directory is reached,
.Fa eofflag
should be set to one.
Note that even a new call to readdir may start where
.Fa readoff
points to end-of-directory.
.Pp
If the file system supports file handles, the arguments
.Fa cookies
and
.Fa ncookies
must be filled out.
.Fa cookies
is a vector for offsets corresponding to the read offsets.
One cookie should be filled out for each directory entry.
.Fa ncookies
is the number of slots for cookies in the cookie vector upon entry to
the function and must be set to the amount of cookies stored in the
vector (i.e. amount of directory entries read) upon exit.
There is always enough space in the cookie vector for the maximal number
of entries that will fit into the directory entry buffer.
For filling out the vector, the helper function
.Fn PUFFS_STORE_DCOOKIE cookies ncookies offset
can be used.
This properly checks against
.Fa cookies
being
.Dv NULL .
Note that
.Fa ncookies
must be initialized to zero before the first call to
.Fn PUFFS_STORE_DCOOKIE .
.It Fn puffs_node_readlink "pcc" "opc" "pcr" "link" "linklen"
Read the target of a symbolic link
.Fa opc .
The result is placed in the buffer pointed to by
.Fa link .
This buffer's length is given in
.Fa linklen
and it must be updated to reflect the real link length.
A terminating nul character should not be put into the buffer and
.Em "must not"
be included in the link length.
.It Fn puffs_node_read "pcc" "opc" "buf" "offset" "resid" "pcr" "ioflag"
Read the contents of a file
.Fa opc .
It will gather the data from
.Fa offset
in the file and read the number
.Fa resid
octets.
The buffer is guaranteed to have this much space.
The amount of data requested by
.Fa resid
should be read, except in the case of eof-of-file or an error.
The parameter
.Fa resid
should be set to indicate the amount of request NOT completed.
In the normal case this should be 0.
.It Fn puffs_node_write "pcc" "opc" "buf" "offset" "resid" "pcr" "ioflag"
.Fn puffs_node_write
Write data to a file
.Fa opc
at
.Fa offset
and extend the file if necessary.
The number of octets written is indicated by
.Fa resid ;
everything must be written or an error will be generated.
The parameter must be set to indicate the amount of data NOT written.
In case the flag
.Dv PUFFS_IO_APPEND
is specified, the data should be appended to the end of the file.
.It Fn puffs_node_print "pcc" "opc"
Print information about node. This is used only for kernel-initiated
diagnostic purposes.
.It Fn puffs_node_reclaim "pcc" "opc" "pid"
The kernel will no longer reference the cookie and resources associated
with it may be freed.
In case the file
.Fa opc
has a link count of zero, it may be safely removed now.
.It Fn puffs_node_inactive "pcc" "opc" "pid" "refcount"
The node
.Fa opc
has lost its last reference in the kernel.
However, the cookie must still remain valid until
.Fn puffs_node_reclaim
is called.
The file system should return its internal reference count on the file
(usually number of links to the file) in
.Fa refcount .
If this is zero, the kernel will call reclaim immediately.
.It Fn puffs_setback pcc op
Issue a "setback" operation which will be handled when the request response
is returned to the kernel.
Currently this can be only called from mmap, open, remove and rmdir.
The valid parameters for
.Ar op
are
.Dv PUFFS_SETBACK_INACT_N1
and
.Dv PUFFS_SETBACK_INACT_N2 .
These signal that a file system mounted with
.Dv PUFFS_KFLAG_IAONDEMAND
should call the file system inactive method for the specified node.
The node number 1 always means the operation cookie
.Ar opc ,
while the node number 2 can be used to specify the second node argument
present in some methods, e.g. remove.
.El
.Sh SEE ALSO
.Xr puffs 3 ,
.Xr vfsops 9 ,
.Xr vnodeops 9