NetBSD/lib/libpuffs/puffs.3

.\"	$NetBSD: puffs.3,v 1.13 2007/02/08 05:24:36 pooka Exp $
.\"
.\" Copyright (c) 2006, 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 February 8, 2007
.Dt PUFFS 3
.Os
.Sh NAME
.Nm puffs
.Nd Pass-to-Userspace Framework File System development interface
.Sh LIBRARY
.Lb libpuffs
.Sh SYNOPSIS
.In puffs.h
.Ft struct puffs_usermount *
.Fo puffs_mount
.Fa "struct puffs_ops *pops" "const char *dir" "int mntflags"
.Fa "const char *puffsname" "void *private"
.Fa "uint32_t pflags" "size_t maxreqlen"
.Fc
.Ft int
.Fo puffs_start
.Fa "struct puffs_usermount *pu" "void *rootcookie" "struct statvfs *svfsb"
.Fc
.Ft int
.Fn puffs_mainloop "struct puffs_usermount *pu" "int flags"
.Ft int
.Fn puffs_getselectable "struct puffs_usermount *pu"
.Ft int
.Fn puffs_setblockingmode "struct puffs_usermount *pu" "int mode"
.Ft int
.Fn puffs_getstate "struct puffs_usermount *pu"
.Ft int
.Fn puffs_setstacksize "struct puffs_usermount *pu" "size_t stacksize"
.Ft int
.Fn puffs_fs_unmount "struct puffs_cc *pcc" "int flags" "pid_t pid"
.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 *cred"
.Fa "pid_t pid"
.Fc
.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_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_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_remove
.Fa "struct puffs_cc *pcc" "void *opc" "void *targ"
.Fa "const struct puffs_cn *pcn"
.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_readdir
.Fa "struct puffs_cc *pcc" "void *opc" "struct dirent *dent"
.Fa "const struct puffs_cred *pcr" "off_t *readoff" "size_t *reslen"
.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_link
.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_readlink
.Fa "struct puffs_cc *pcc" "void *opc" "const struct puffs_cred *cred"
.Fa "char *link" "size_t *linklen"
.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 flags"
.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 "struct puffs_cred *pcr" "pid_t pid"
.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_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
.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
.Nm
provides a framework for creating file systems as userspace servers.
Operations are transported from the kernel virtual file system layer
to the concrete implementation behind
.Nm ,
where they are processed and results are sent back to the kernel.
.Pp
It is possible to use
.Nm
in two different ways.
Calling
.Fn puffs_mainloop
takes execution context away from the caller and automatically handles
all requests by using the callbacks.
Alternatively, control can be kept with the caller and operations
handled manually.
.Ss Library operation
The file system is mounted using
.Fn puffs_mount .
The callbacks are passed as the fields in the structure
.Fa pops .
They can be initialized using the macro
.Fn PUFFSOP_SET pops fsname type opname ,
which will initialize the operation
.Fn puffs_type_opname
in
.Fa pops
to
.Fn fsname_type_opname .
All operations are initialized to a default state with the call
.Fn PUFFSOP_INIT pops .
All of the VFS routines are mandatory, but all of the node operations
with the exception of
.Fn puffs_fs_lookup
are optional.
However, leaving operations blank will naturally have an effect on the
features available from the file system implementation.
The argument
.Fa dir
signifies the mount point,
.Fa mntflags
is the flagset given to
.Xr mount 2 ,
and
.Fa puffsname
is the name of the file system implementation.
Flags for
.Nm
can be given via
.Fa pflags .
Currently the following flags are supported:
.Bl -tag -width "XPUFFS_KFLAG_BUILDPATHX"
.It Dv PUFFS_KFLAG_NOCACHE
Do not cache anything in the kernel.
Currently this means not using the page cache for regular files or
the name cache (directory entry cache) for all files.
This is useful if the file system does not want to include flushing
and invalidation logic in case it is possible to modify the backend
from "under" our view of the file system (a distributed file system,
for example).
.It Dv PUFFS_KFLAG_ALLOPS
This flag requests that all operations are sent to userspace.
Normally the kernel shortcircuits unimplemented operations.
This flag is mostly useful for debugging purposes.
.It Dv PUFFS_FLAG_BUILDPATH
The framework will build a complete path name, which is supplied
with each operation and can be found from the
.Va pn_po.po_full_pcn
field in a
.Vt struct puffs_node .
The option assumes that the framework can map a cookie to a
.Vt struct puffs_node .
See
.Sx Cookies
for more information on cookie mapping.
See
.Xr puffs_path 3
for more information on library calls involving paths.
.It Dv PUFFS_FLAG_OPDUMP
This option makes the framework dump a textual representation of
each operation before executing it.
It is useful for debugging purposes.
.El
.Pp
Finally, the maximum operation buffer length is requested by
.Fa maxreqlen .
The field
.Va pu_maxreqlen
from the returned mount structure is the kernel sanity-checked value and
should always be consulted after the mount call returns.
Supplying 0 as this parameter will make the kernel choose the longest
possible buffer length.
In case of success,
.Fn puffs_fs_mount
returns the address of the user mount instance.
Otherwise,
.Dv NULL
is returned and errno is set to specify the error.
.Pp
To handle all requests automatically until the file system is
unmounted,
.Fn puffs_mainloop
should be used.
It returns 0 if the file system was succesfully unmounted or \-1 if it
was killed in action.
Unless
.Fa flags
is used to pass
.Dv PUFFSLOOP_NODAEMON ,
.Fn puffs_mainloop
will also detach from the terminal.
.Pp
.Fn puffs_getselectable
can be used to query a handle to do I/O multiplexing with:
.Xr select 2 ,
.Xr poll 2 ,
and
.Xr kqueue 2
are all examples of acceptable operations.
.Pp
The library can be set in blocking or non-blocking mode using
.Fn puffs_setblockingmode .
Acceptable values for the argument are
.Dv PUFFSDEV_BLOCK
and
.Dv PUFFSDEV_NONBLOCK .
.Pp
The state of the file system can be queried with
.Fn puffs_getstate .
It is maintained by the framework and is mostly useful for the framework
itself.
Possible values are
.Dv PUFFS_STATE_MOUNTING ,
.Dv PUFFS_STATE_RUNNING ,
.Dv PUFFS_STATE_UNMOUNTING
and
.Dv PUFFS_STATE_UNMOUNTED .
.Pp
The stack size for handling requests can be set with
.Fn puffs_setstacksize .
The default is one megabyte of stack space per request.
See
.Xr puffs_cc 3 .
.Pp
After the correct setup for the library has been established and the
backend has been initialized, the file system is made operational
by calling
.Fn puffs_start .
Immediately after the return of this function the file system must
be ready to process requests.
.Ss Cookies
Every file (regular file, directory, device node, ...) instance is
attached to the kernel using a cookie.
A cookie should uniquely map to a file during its lifetime.
If file instances are kept in memory, a simple strategy is to use
the virtual address of the structure describing the file.
The cookie can be recycled when
.Fn puffs_node_reclaim
is called for a node.
.Pp
For some operations (such as building paths) the framework needs to map
the cookie to the framework-level structure describing a file,
.Vt struct puffs_node .
It is advisable to simply use the
.Vt struct puffs_node
address as a cookie and store file system specific data in the private
portion of
.Vt struct puffs_node .
The library assumes this by default.
If it is not desireable, the file system implementation can call
.Fn puffs_set_cookiemap
to provide an alternative cookie-to-node mapping function.
.Ss File system callbacks
The callbacks do all the actual work in implementing the file system.
Currently they are fairly close to the vfs and vnode operations in
the kernel but with simplified operation.
This section describes the calls which relate to the file system
itself.
.Pp
All callbacks can be prototyped with the file system name and operation
name using the macro
.Fn PUFFSOP_PROTOS fsname .
.Pp
.Fn puffs_fs_statvfs
should fill in the following fields of
.Fa sbp :
.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 .
.Pp
The file system should be sychronized to storage when
.Fn puffs_fs_sync
is called.
The
.Fa waitfor
parameter should handled similarly as inside the kernel.
.Pp
The file system should be unmounted when
.Fn puffs_fs_unmount
is called.
If the flag
.Dv MNT_FORCE
is not honored, the kernel will proceed to forcibly unmount the file system
despite this.
.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.
.Pp
The
.Fn puffs_node_lookup
function is used to locate nodes.
The implementation should match the name in
.Fa pcn
against the existing entries in the directory provided by the cookie.
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 .
Otherwise, 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 .
.Pp
.Fn puffs_node_getattr
fills out a struct vattr pointed to by
.Fa va .
.Pp
.Fn puffs_node_setattr
sets the attributes in
.Fa va .
Instead of setting everything according to that file, only fields which
are not marked
.Dv VNOVAL
should be set.
.Pp
A file node is created in the directory specified by the cookie when
.Fn puffs_node_create
is called.
The attributes are specified by
.Fa va
and the cookie for the newly created node should be returned in
.Fa newnode .
Similarly,
.Fn puffs_node_mkdir
creates a directory.
.Pp
.Fn puffs_node_remove
removes the file
.Fa targ
from the directory indicated by the cookie.
Similarly,
.Fn puffs_node_rmdir
removes a directory.
The name of the directory entry to remove is described by
.Fa pcn .
.Pp
To read directory entries,
.Fn puffs_node_readdir
is called.
It should store directories as struct dirents 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 systme 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.
In case the directory was exhausted, the parameters should not be
modified to signal end-of-directory.
.Pp
A node rename is done by calling
.Fn puffs_node_rename .
If the destination file cookie is non-null, it must be removed
and the new entry overwritten atomically.
The directory entry names to be used are described by the
struct puffs_cn's (cf. create and remove).
.Pp
A hard link is created by
.Fn puffs_node_link .
In practice this means adding a directory entry described by
.Fa pcn
to the cookied directory and the entry pointing to the target node.
.Pp
A symbolic link in turn is created by
.Fn puffs_node_symlink .
It is similar to creating a regular file, except that
.Fa link_target
specifies the target of the link which should be set for the link.
.Pp
To read the target of a symbolic link,
.Fa puffs_node_readlink
is called.
The path in the link target should be copied to
.Fa link
and the length without the terminating nul set in
.Fa linklen .
.Pp
A device node is created using
.Fn puffs_node_mknod .
The only difference to creating a normal file is that the attribute
struct contains the device identifier in
.Fa va-\*[Gt]va_rdev .
.Pp
Files are opened with a call to
.Fn puffs_node_open .
Most of the time this can be left unimplemented, unless special
resource allocation is required.
.Pp
.Fn puffs_node_close
releases all the resources allocated by
.Fn puffs_node_open .
.Pp
To check if access of type
.Va mode
to a file is allowed,
.Fn puffs_node_access
is called.
This controls file access, not e.g.
.Fn puffs_node_open .
.Pp
.Fn puffs_node_read
reads the contents of a file.
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.
.Pp
.Fn puffs_node_write
writes data to a file at
.Fa offset
extending 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.
.Pp
.Fn puffs_node_reclaim
signals that the cookie will no longer be referenced without a further
call to
.Fn puffs_node_lookup .
This information can be used to free resources and specifically release
a file for which no directory entries remain.
.Pp
.Fn puffs_node_inactive
signals that the kernel has released its last reference to the node.
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.
.Sh SEE ALSO
.Xr puffs_cc 3 ,
.Xr puffs_flush 3 ,
.Xr puffs_node 3 ,
.Xr puffs_path 3 ,
.Xr puffs_req 3 ,
.Xr puffs_suspend 3 ,
.Xr puffs 4
.Sh HISTORY
An unsupported experimental version of
.Nm
first appeared in
.Nx 4.0 .
.Sh AUTHORS
.An Antti Kantee Aq pooka@iki.fi
.Sh BUGS
Under construction.