.ds TYPE C
.\"
.\" See the file LICENSE for redistribution information.
.\"
.\" Copyright (c) 1996, 1997
.\"	Sleepycat Software.  All rights reserved.
.\"
.\"	@(#)db_mpool.so	10.35 (Sleepycat) 12/3/97
.\"
.\"
.\" See the file LICENSE for redistribution information.
.\"
.\" Copyright (c) 1997
.\"	Sleepycat Software.  All rights reserved.
.\"
.\"	@(#)macros.so	10.38 (Sleepycat) 1/18/98
.\"
.\" We don't want hyphenation for any HTML documents.
.ie '\*[HTML]'YES'\{\
.nh
\}
.el\{\
.ds Hy
.hy
..
.ds Nh
.nh
..
\}
.\" The alternative text macro
.\" This macro takes two arguments:
.\"	+ the text produced if this is a "C" manpage
.\"	+ the text produced if this is a "CXX" or "JAVA" manpage
.\"
.de Al
.ie '\*[TYPE]'C'\{\\$1
\}
.el\{\\$2
\}
..
.\" Scoped name macro.
.\" Produces a_b, a::b, a.b depending on language
.\" This macro takes two arguments:
.\"	+ the class or prefix (without underscore)
.\"	+ the name within the class or following the prefix
.de Sc
.ie '\*[TYPE]'C'\{\\$1_\\$2
\}
.el\{\
.ie '\*[TYPE]'CXX'\{\\$1::\\$2
\}
.el\{\\$1.\\$2
\}
\}
..
.\" Scoped name for Java.
.\" Produces Db.b, for Java, otherwise just b.  This macro is used for
.\" constants that must be scoped in Java, but are global otherwise.
.\" This macro takes two arguments:
.\"	+ the class
.\"	+ the name within the class or following the prefix
.de Sj
.ie '\*[TYPE]'JAVA'\{\
.TP 5
Db.\\$1\}
.el\{\
.TP 5
\\$2\}
..
.\" The general information text macro.
.de Gn
.ie '\*[TYPE]'C'\{The DB library is a family of groups of functions that provides a modular
programming interface to transactions and record-oriented file access.
The library includes support for transactions, locking, logging and file
page caching, as well as various indexed access methods.
Many of the functional groups (e.g., the file page caching functions)
are useful independent of the other DB functions,
although some functional groups are explicitly based on other functional
groups (e.g., transactions and logging).
\}
.el\{The DB library is a family of classes that provides a modular
programming interface to transactions and record-oriented file access.
The library includes support for transactions, locking, logging and file
page caching, as well as various indexed access methods.
Many of the classes (e.g., the file page caching class)
are useful independent of the other DB classes,
although some classes are explicitly based on other classes
(e.g., transactions and logging).
\}
For a general description of the DB package, see
.IR db_intro (3).
..
.\" The library error macro, the local error macro.
.\" These macros take one argument:
.\"	+ the function name.
.de Ee
The
.I \\$1
.ie '\*[TYPE]'C'\{function may fail and return
.I errno
\}
.el\{method may fail and throw a
.IR DbException (3)
.if '\*[TYPE]'CXX'\{
or return
.I errno
\}
\}
for any of the errors specified for the following DB and library functions:
..
.de Ec
In addition, the
.I \\$1
.ie '\*[TYPE]'C'\{function may fail and return
.I errno
\}
.el\{method may fail and throw a
.IR DbException (3)
.ie '\*[TYPE]'CXX'\{or return
.I errno
\}
.el\{encapsulating an
.I errno
\}
\}
for the following conditions:
..
.de Ea
[EAGAIN]
A lock was unavailable.
..
.de Eb
[EBUSY]
The shared memory region was in use and the force flag was not set.
..
.de Em
[EAGAIN]
The shared memory region was locked and (repeatedly) unavailable.
..
.de Ei
[EINVAL]
An invalid flag value or parameter was specified.
..
.de Es
[EACCES]
An attempt was made to modify a read-only database.
..
.de Et
The DB_THREAD flag was specified and spinlocks are not implemented for
this architecture.
..
.de Ep
[EPERM]
Database corruption was detected.
All subsequent database calls (other than
.ie '\*[TYPE]'C'\{\
.IR DB->close )
\}
.el\{\
.IR Db::close )
\}
will return EPERM.
..
.de Ek
.if '\*[TYPE]'CXX'\{\
Methods marked as returning
.I errno
will, by default, throw an exception that encapsulates the error information.
The default error behavior can be changed, see
.IR DbException (3).
\}
..
.\" The SEE ALSO text macro
.de Sa
.\" make the line long for nroff.
.if n .ll 72
.nh
.na
.IR db_archive (1),
.IR db_checkpoint (1),
.IR db_deadlock (1),
.IR db_dump (1),
.IR db_load (1),
.IR db_recover (1),
.IR db_stat (1),
.IR db_intro (3),
.ie '\*[TYPE]'C'\{\
.IR db_appinit (3),
.IR db_cursor (3),
.IR db_dbm (3),
.IR db_internal (3),
.IR db_lock (3),
.IR db_log (3),
.IR db_mpool (3),
.IR db_open (3),
.IR db_thread (3),
.IR db_txn (3)
\}
.el\{\
.IR db_internal (3),
.IR db_thread (3),
.IR Db (3),
.IR Dbc (3),
.IR DbEnv (3),
.IR DbException (3),
.IR DbInfo (3),
.IR DbLock (3),
.IR DbLockTab (3),
.IR DbLog (3),
.IR DbLsn (3),
.IR DbMpool (3),
.IR DbMpoolFile (3),
.IR Dbt (3),
.IR DbTxn (3),
.IR DbTxnMgr (3)
\}
.ad
.Hy
..
.\" The function header macro.
.\" This macro takes one argument:
.\"	+ the function name.
.de Fn
.in 2
.I \\$1
.in
..
.\" The XXX_open function text macro, for merged create/open calls.
.\" This macro takes two arguments:
.\"	+ the interface, e.g., "transaction region"
.\"	+ the prefix, e.g., "txn" (or the class name for C++, e.g., "DbTxn")
.de Co
.ie '\*[TYPE]'C'\{\
.Fn \\$2_open
The
.I \\$2_open
function copies a pointer, to the \\$1 identified by the
.B directory
.IR dir ,
into the memory location referenced by
.IR regionp .
.PP
If the
.I dbenv
argument to
.I \\$2_open
was initialized using
.IR db_appinit ,
.I dir
is interpreted as described by
.IR db_appinit (3).
\}
.el\{\
.Fn \\$2::open
The
.I \\$2::open
.ie '\*[TYPE]'CXX'\{\
method copies a pointer, to the \\$1 identified by the
.B directory
.IR dir ,
into the memory location referenced by
.IR regionp .
\}
.el\{\
method returns a \\$1 identified by the
.B directory
.IR dir .
\}
.PP
If the
.I dbenv
argument to
.I \\$2::open
was initialized using
.IR DbEnv::appinit ,
.I dir
is interpreted as described by
.IR DbEnv (3).
\}
.PP
Otherwise,
if
.I dir
is not NULL,
it is interpreted relative to the current working directory of the process.
If
.I dir
is NULL,
the following environment variables are checked in order:
``TMPDIR'', ``TEMP'', and ``TMP''.
If one of them is set,
\\$1 files are created relative to the directory it specifies.
If none of them are set, the first possible one of the following
directories is used:
.IR /var/tmp ,
.IR /usr/tmp ,
.IR /temp ,
.IR /tmp ,
.I C:/temp
and
.IR C:/tmp .
.PP
All files associated with the \\$1 are created in this directory.
This directory must already exist when
.ie '\*[TYPE]'C'\{
\\$1_open
\}
.el\{\
\\$2::open
\}
is called.
If the \\$1 already exists,
the process must have permission to read and write the existing files.
If the \\$1 does not already exist,
it is optionally created and initialized.
..
.\" The common close language macro, for discarding created regions
.\" This macro takes one argument:
.\"	+ the function prefix, e.g., txn (the class name for C++, e.g., DbTxn)
.de Cc
In addition, if the
.I dir
argument to
.ie '\*[TYPE]'C'\{\
.ds Va db_appinit
.ds Vo \\$1_open
.ds Vu \\$1_unlink
\}
.el\{\
.ds Va DbEnv::appinit
.ds Vo \\$1::open
.ds Vu \\$1::unlink
\}
.I \\*(Vo
was NULL
and
.I dbenv
was not initialized using
.IR \\*(Va ,
.if '\\$1'memp'\{\
or the DB_MPOOL_PRIVATE flag was set,
\}
all files created for this shared region will be removed,
as if
.I \\*(Vu
were called.
.rm Va
.rm Vo
.rm Vu
..
.\" The DB_ENV information macro.
.\" This macro takes two arguments:
.\"	+ the function called to open, e.g., "txn_open"
.\"	+ the function called to close, e.g., "txn_close"
.de En
.ie '\*[TYPE]'C'\{\
based on the
.I dbenv
argument to
.IR \\$1 ,
which is a pointer to a structure of type DB_ENV (typedef'd in <db2/db.h>).
It is expected that applications will use a single DB_ENV structure as the
argument to all of the subsystems in the DB package.
In order to ensure compatibility with future releases of DB, all fields of
the DB_ENV structure that are not explicitly set should be initialized to 0
before the first time the structure is used.
Do this by declaring the structure external or static, or by calling the C
library routine
.IR bzero (3)
or
.IR memset (3).
.PP
The fields of the DB_ENV structure used by
.I \\$1
are described below.
.if '\*[TYPE]'CXX'\{\
As references to the DB_ENV structure may be maintained by
.IR \\$1 ,
it is necessary that the DB_ENV structure and memory it references be valid
until the
.I \\$2
function is called.
\}
.ie '\\$1'db_appinit'\{The
.I dbenv
argument may not be NULL.
If any of the fields of the
.I dbenv
are set to 0,
defaults appropriate for the system are used where possible.
\}
.el\{If
.I dbenv
is NULL
or any of its fields are set to 0,
defaults appropriate for the system are used where possible.
\}
.PP
The following fields in the DB_ENV structure may be initialized before calling
.IR \\$1 :
\}
.el\{\
based on which set methods have been used.
It is expected that applications will use a single DbEnv object as the
argument to all of the subsystems in the DB package.
The fields of the DbEnv object used by
.I \\$1
are described below.
As references to the DbEnv object may be maintained by
.IR \\$1 ,
it is necessary that the DbEnv object and memory it references be valid
until the object is destroyed.
.ie '\\$1'appinit'\{\
The
.I dbenv
argument may not be NULL.
If any of the fields of the
.I dbenv
are set to 0,
defaults appropriate for the system are used where possible.
\}
.el\{\
Any of the DbEnv fields that are not explicitly set will default to
appropriate values.
\}
.PP
The following fields in the DbEnv object may be initialized, using the
appropriate set method, before calling
.IR \\$1 :
\}
..
.\" The DB_ENV common fields macros.
.de Se
.ie '\*[TYPE]'JAVA'\{\
.TP 5
DbErrcall db_errcall;
.ns
.TP 5
String db_errpfx;
.ns
.TP 5
int db_verbose;
The error fields of the DbEnv behave as described for
.IR DbEnv (3).
\}
.el\{\
.ie '\*[TYPE]'CXX'\{\
.TP 5
void *(*db_errcall)(char *db_errpfx, char *buffer);
.ns
.TP 5
FILE *db_errfile;
.ns
.TP 5
const char *db_errpfx;
.ns
.TP 5
class ostream *db_error_stream;
.ns
.TP 5
int db_verbose;
The error fields of the DbEnv behave as described for
.IR DbEnv (3).
\}
.el\{\
void *(*db_errcall)(char *db_errpfx, char *buffer);
.ns
.TP 5
FILE *db_errfile;
.ns
.TP 5
const char *db_errpfx;
.ns
.TP 5
int db_verbose;
The error fields of the DB_ENV behave as described for
.IR db_appinit (3).
\}
\}
..
.\" The open flags.
.de Fm
The
.I flags
and
.I mode
arguments specify how files will be opened and/or created when they
don't already exist.
The flags value is specified by
.BR or 'ing
together one or more of the following values:
.Sj DB_CREATE
Create any underlying files, as necessary.
If the files do not already exist and the DB_CREATE flag is not specified,
the call will fail.
..
.\" DB_THREAD open flag macro.
.\" This macro takes two arguments:
.\"	+ the open function name
.\"	+ the object it returns.
.de Ft
.TP 5
.Sj DB_THREAD
Cause the \\$2 handle returned by the
.I \\$1
.Al function method
to be useable by multiple threads within a single address space,
i.e., to be ``free-threaded''.
.if '\*[TYPE]'JAVA'\{\
Threading is assumed in the Java API,
so no special flags are required,
and DB functions will always behave as if the DB_THREAD flag was specified.
\}
..
.\" The mode macro.
.\" This macro takes one argument:
.\"	+ the subsystem name.
.de Mo
All files created by the \\$1 are created with mode
.I mode
(as described in
.IR chmod (2))
and modified by the process' umask value at the time of creation (see
.IR umask (2)).
The group ownership of created files is based on the system and directory
defaults, and is not further specified by DB.
..
.\" The application exits macro.
.\" This macro takes one argument:
.\"	+ the application name.
.de Ex
The
.I \\$1
utility exits 0 on success, and >0 if an error occurs.
..
.\" The application -h section.
.\" This macro takes one argument:
.\"	+ the application name
.de Dh
DB_HOME
If the
.B \-h
option is not specified and the environment variable
.I DB_HOME
is set, it is used as the path of the database home, as described in
.IR db_appinit (3).
..
.\" The function DB_HOME ENVIRONMENT VARIABLES section.
.\" This macro takes one argument:
.\"	+ the open function name
.de Eh
DB_HOME
If the
.I dbenv
argument to
.I \\$1
was initialized using
.IR db_appinit ,
the environment variable DB_HOME may be used as the path of the database
home for the interpretation of the
.I dir
argument to
.IR \\$1 ,
as described in
.IR db_appinit (3).
.if \\n(.$>1 \{Specifically,
.I \\$1
is affected by the configuration string value of \\$2.\}
..
.\" The function TMPDIR ENVIRONMENT VARIABLES section.
.\" This macro takes two arguments:
.\"	+ the interface, e.g., "transaction region"
.\"	+ the prefix, e.g., "txn" (or the class name for C++, e.g., "DbTxn")
.de Ev
TMPDIR
If the
.I dbenv
argument to
.ie '\*[TYPE]'C'\{\
.ds Vo \\$2_open
\}
.el\{\
.ds Vo \\$2::open
\}
.I \\*(Vo
was NULL or not initialized using
.IR db_appinit ,
the environment variable TMPDIR may be used as the directory in which to
create the \\$1,
as described in the
.I \\*(Vo
section above.
.rm Vo
..
.\" The unused flags macro.
.de Fl
The
.I flags
parameter is currently unused, and must be set to 0.
..
.\" The no-space TP macro.
.de Nt
.br
.ns
.TP 5
..
.\" The return values of the functions macros.
.\" Rc is the standard two-value return with a suffix for more values.
.\" Ro is the standard two-value return but there were previous values.
.\" Rt is the standard two-value return, returning errno, 0, or < 0.
.\" These macros take one argument:
.\"	+ the routine name
.de Rc
The
.I \\$1
.ie '\*[TYPE]'C'\{function returns the value of
.I errno
on failure,
0 on success,
\}
.el\{method throws a
.IR DbException (3)
.ie '\*[TYPE]'CXX'\{or returns the value of
.I errno
on failure,
0 on success,
\}
.el\{that encapsulates an
.I errno
on failure,
\}
\}
..
.de Ro
Otherwise, the
.I \\$1
.ie '\*[TYPE]'C'\{function returns the value of
.I errno
on failure and 0 on success.
\}
.el\{method throws a
.IR DbException (3)
.ie '\*[TYPE]'CXX'\{or returns the value of
.I errno
on failure and 0 on success.
\}
.el\{that encapsulates an
.I errno
on failure,
\}
\}
..
.de Rt
The
.I \\$1
.ie '\*[TYPE]'C'\{function returns the value of
.I errno
on failure and 0 on success.
\}
.el\{method throws a
.IR DbException (3)
.ie '\*[TYPE]'CXX'\{or returns the value of
.I errno
on failure and 0 on success.
\}
.el\{that encapsulates an
.I errno
on failure.
\}
\}
..
.\" The TXN id macro.
.de Tx
.IP
If the file is being accessed under transaction protection,
the
.I txnid
parameter is a transaction ID returned from
.IR txn_begin ,
otherwise, NULL.
..
.\" The XXX_unlink function text macro.
.\" This macro takes two arguments:
.\"	+ the interface, e.g., "transaction region"
.\"	+ the prefix (for C++, this is the class name)
.de Un
.ie '\*[TYPE]'C'\{\
.ds Va db_appinit
.ds Vc \\$2_close
.ds Vo \\$2_open
.ds Vu \\$2_unlink
\}
.el\{\
.ds Va DbEnv::appinit
.ds Vc \\$2::close
.ds Vo \\$2::open
.ds Vu \\$2::unlink
\}
.Fn \\*(Vu
The
.I \\*(Vu
.Al function method
destroys the \\$1 identified by the directory
.IR dir ,
removing all files used to implement the \\$1.
.ie '\\$2'log' \{(The log files themselves and the directory
.I dir
are not removed.)\}
.el \{(The directory
.I dir
is not removed.)\}
If there are processes that have called
.I \\*(Vo
without calling
.I \\*(Vc
(i.e., there are processes currently using the \\$1),
.I \\*(Vu
will fail without further action,
unless the force flag is set,
in which case
.I \\*(Vu
will attempt to remove the \\$1 files regardless of any processes
still using the \\$1.
.PP
The result of attempting to forcibly destroy the region when a process
has the region open is unspecified.
Processes using a shared memory region maintain an open file descriptor
for it.
On UNIX systems, the region removal should succeed
and processes that have already joined the region should continue to
run in the region without change,
however processes attempting to join the \\$1 will either fail or
attempt to create a new region.
On other systems, e.g., WNT, where the
.IR unlink (2)
system call will fail if any process has an open file descriptor
for the file,
the region removal will fail.
.PP
In the case of catastrophic or system failure,
database recovery must be performed (see
.IR db_recover (1)
or the DB_RECOVER flags to
.IR \\*(Va (3)).
Alternatively, if recovery is not required because no database state is
maintained across failures,
it is possible to clean up a \\$1 by removing all of the
files in the directory specified to the
.I \\*(Vo
.Al function, method,
as \\$1 files are never created in any directory other than the one
specified to
.IR \\*(Vo .
Note, however,
that this has the potential to remove files created by the other DB
subsystems in this database environment.
.PP
.Rt \\*(Vu
.rm Va
.rm Vo
.rm Vu
.rm Vc
..
.\" Signal paragraph for standard utilities.
.\" This macro takes one argument:
.\"	+ the utility name.
.de Si
The
.I \\$1
utility attaches to DB shared memory regions.
In order to avoid region corruption,
it should always be given the chance to detach and exit gracefully.
To cause
.I \\$1
to clean up after itself and exit,
send it an interrupt signal (SIGINT).
..
.\" Logging paragraph for standard utilities.
.\" This macro takes one argument:
.\"	+ the utility name.
.de Pi
.B \-L
Log the execution of the \\$1 utility to the specified file in the
following format, where ``###'' is the process ID, and the date is
the time the utility starting running.
.sp
\\$1: ### Wed Jun 15 01:23:45 EDT 1995
.sp
This file will be removed if the \\$1 utility exits gracefully.
..
.\" Malloc paragraph.
.\" This macro takes one argument:
.\"	+ the allocated object
.de Ma
.if !'\*[TYPE]'JAVA'\{\
\\$1 are created in allocated memory.
If
.I db_malloc
is non-NULL,
it is called to allocate the memory,
otherwise,
the library function
.IR malloc (3)
is used.
The function
.I db_malloc
must match the calling conventions of the
.IR malloc (3)
library routine.
Regardless,
the caller is responsible for deallocating the returned memory.
To deallocate the returned memory,
free each returned memory pointer;
pointers inside the memory do not need to be individually freed.
\}
..
.\" Underlying function paragraph.
.\" This macro takes two arguments:
.\"	+ the function name
.\"	+ the utility name
.de Uf
The
.I \\$1
.Al function method
is the underlying function used by the
.IR \\$2 (1)
utility.
See the source code for the
.I \\$2
utility for an example of using
.I \\$1
in a UNIX environment.
..
.\" Underlying function paragraph, for C++.
.\" This macro takes three arguments:
.\"	+ the C++ method name
.\"	+ the function name for C
.\"	+ the utility name
.de Ux
The
.I \\$1
method is based on the C
.I \\$2
function, which
is the underlying function used by the
.IR \\$3 (1)
utility.
See the source code for the
.I \\$3
utility for an example of using
.I \\$2
in a UNIX environment.
..
.TH DB_MPOOL 3 "December 3, 1997"
.UC 7
.SH NAME
db_mpool \- shared memory buffer pool
.SH SYNOPSIS
.nf
.ft B
#include <db2/db.h>

int
memp_open(char *dir,
.ti +5
int flags, int mode, DB_ENV *dbenv, DB_MPOOL **regionp);

int
memp_close(DB_MPOOL *mp);

int
memp_fopen(DB_MPOOL *mp, char *file, int ftype, int flags,
.ti +5
int mode, size_t pagesize, int lsn_offset, DBT *pgcookie,
.ti +5
u_int8_t *uid, DB_MPOOLFILE **mpf);

int
memp_fclose(DB_MPOOLFILE *mpf);

int
memp_fget(DB_MPOOLFILE *mpf,
.ti +5
db_pgno_t *pgnoaddr, int flags, void **pagep);

int
memp_fput(DB_MPOOLFILE *mpf, void *pgaddr, int flags);

int
memp_fset(DB_MPOOLFILE *mpf, void *pgaddr, int flags);

int
memp_fsync(DB_MPOOLFILE *mpf);

int
memp_unlink(const char *dir, int force, DB_ENV *);

int
memp_register(DB_MPOOL *mp, int ftype,
.ti +5
int (*pgin)(db_pgno_t pgno, void *pgaddr, DBT *pgcookie),
.ti +5
int (*pgout)(db_pgno_t pgno, void *pgaddr, DBT *pgcookie));

int
memp_trickle(DB_MPOOL *mp, int pct, int *nwrotep);

int
memp_sync(DB_MPOOL *mp, LSN *lsn);

int
memp_stat(DB_MPOOL *mp, DB_MPOOL_STAT **gsp,
.ti +5
DB_MPOOL_FSTAT *(*fsp)[], void *(*db_malloc)(size_t));
.ft R
.fi
.SH DESCRIPTION
.Gn
.PP
This manual page describes the specific details of the memory pool interface.
.PP
The
.I db_mpool
functions are the library interface intended to provide general-purpose,
page-oriented buffer management of one or more files.
While designed to work with the other DB functions, these functions are
also useful for more general purposes.
The memory pools (DB_MPOOL's) are referred to in this document as
simply ``pools''.
Pools may be shared between processes.
Pools are usually filled by pages from one or more files (DB_MPOOLFILE's).
Pages in the pool are replaced in LRU (least-recently-used) order,
with each new page replacing the page that has been unused the longest.
Pages retrieved from the pool using
.I memp_fget
are ``pinned'' in the pool, by default,
until they are returned to the pool's control using the
.I memp_fput
function.
.PP
.Co "memory pool" mpool
.PP
.Fm
.TP 5
DB_MPOOL_PRIVATE
Create a private MPOOL that is not shared with any other process (although
it may be shared with other threads).
.TP 5
DB_NOMMAP
Always copy files in this memory pool into the local cache instead of mapping
them into process memory (see the description of the
.I mp_mmapsize
field of the DB_ENV structure for further information).
.Ft memp_open DB_MPOOL
.PP
.Mo "memory pool subsystem (other than files created by the \fImemp_fopen\fP function, which are separately specified)"
.PP
The memory pool subsystem is configured
.En "memp_open" "memp_closed"
.TP 5
.Se
.TP 5
size_t mp_mmapsize;
Files that are opened read-only in the pool (and that satisfy a few other
criteria) are, by default,
mapped into the process address space instead of being copied into the local
cache.
This can result in better-than-usual performance,
as available virtual memory is normally much larger than the local cache,
and page faults are faster than page copying on many systems.
However,
in the presence of limited virtual memory it can cause resource starvation,
and in the presence of large databases,
it can result in immense process sizes.
If
.I mp_mmapsize
is non-zero, 
it specifies the maximum file size, in bytes,
for a file to be mapped into the process address space.
By default,
it is set to 10Mb.
.TP 5
size_t mp_size;
The suggested size of the pool, in bytes.
This should be the size of the normal working data set of the application,
with some small amount of additional memory for unusual situations.
(Note,
the working set is not the same as the number of simultaneously referenced
pages,
and should be quite a bit larger!)
The default cache size is 128K bytes (16 8K byte pages),
and may not be less than 20K bytes.
.PP
.Rt memp_open
.PP
.Fn memp_close
The
.I memp_close
function closes the pool indicated by the DB_MPOOL pointer
.IR mp ,
as returned by
.IR memp_open .
This function does not imply a call to
.I memp_fsync
(or to
.IR memp_fclose )
i.e. no pages are written to the source file as as a result of calling
.IR memp_close .
.PP
.Cc memp
.PP
When multiple threads are using the DB_MPOOL handle concurrently,
only a single thread may call the
.I memp_close
function.
.PP
.Rt memp_close
.PP
.Fn memp_fopen
The
.I memp_fopen
function opens a file in the pool specified by the DB_MPOOL argument,
copying the DB_MPOOLFILE pointer representing it into the memory
location referenced by
.IR mpf .
.PP
The
.I file
argument is the name of the file to be opened.
If
.I file
is NULL,
a private file is created that cannot be shared with any other process
(although it may be shared with other threads).
.PP
The
.I ftype
argument should be the same as a
.I ftype
argument previously specified to the
.I memp_register
function,
unless no input or output processing of the file's pages are necessary,
in which case it should be 0.
(See the description of the
.I memp_register
function for more information.)
.PP
.Fm
.TP 5
DB_NOMMAP
Always copy this file into the local cache instead of mapping it into
process memory (see the description of the
.I mp_mmapsize
field of the DB_ENV structure for further information).
.TP 5
DB_RDONLY
Open any underlying files for reading only.
Any attempt to write the file using the pool functions will fail,
regardless of the actual permissions of the file.
.PP
.Mo "function \fImemp_fopen\fP"
.PP
The
.I pagesize
argument is the size, in bytes,
of the unit of transfer between the application and the pool,
although it is not necessarily the unit of transfer between the pool and
the source file.
.PP
The
.I lsn_offset
argument is the zero-based byte offset in the page of the page's log sequence
number (LSN),
or \-1 if no LSN offset is specified.
(See the description of the
.I memp_sync
function for more information.)
.PP
The
.I pgcookie
argument contains the byte string that is passed to the
.I pgin
and
.I pgout
functions for this file, if any.
(See the description of the
.I memp_register
function for more information.)
.PP
The
.I uid
argument is a unique identifier for the file.
The mpool
functions must be able to uniquely identify files in order that multiple
processes sharing a file will correctly share its underlying pages.
Normally, the
.I uid
argument should be NULL and the mpool functions will use the file's
device and inode numbers (see
.IR stat (2))
for this purpose.
On some filesystems, (e.g., FAT or NFS) file device and inode numbers are
not necessarily unique across system reboots.
.ft B
Applications wanting to maintain a shared memory buffer pool across system
reboots, where the pool contains pages from files stored on such filesystems,
must specify a unique file identifier to the memp_fopen call and each process
opening or registering the file must provide the same unique identifier.
.ft R
If the
.I uid
argument is non-NULL,
it must reference a DB_FILE_ID_LEN (as defined in <db2/db.h>) length array of
bytes that will be used to uniquely identify the file.
This should not be necessary for most applications.
Specifically, it is not necessary if the memory pool is re-instantiated after
each system reboot, the application is using the DB access methods instead of
calling the pool functions explicitly, or the files in the memory pool are
stored on filesystems where the file device and inode numbers do not change
across system reboots.
.PP
.Rt memp_fopen
.PP
.Fn memp_fclose
The
.I memp_fclose
function closes the source file indicated by the DB_MPOOLFILE pointer
.IR mpf .
This function does not imply a call to
.IR memp_fsync ,
i.e. no pages are written to the source file as as a result of calling
.IR memp_fclose .
.PP
In addition,
if the
.I file
argument to
.I memp_fopen
was NULL,
any underlying files created for this DB_MPOOLFILE will be removed.
.PP
.Rt memp_fclose
.PP
.Fn memp_fget
The
.I memp_fget
function copies a pointer to the page with the page number specified by
.IR pgnoaddr ,
from the source file specified by the DB_MPOOLFILE pointer
.IR mpf ,
into the memory location referenced by
.IR pagep .
If the page does not exist or cannot be retrieved,
.I memp_fget
will fail.
.PP
The returned page is size_t type aligned.
.PP
.ft B
Page numbers begin at 0, e.g., the first page in the file is page number 0,
not page number 1.
.ft R
.PP
The
.I flags
argument is specified by
.BR or 'ing
together one or more of the following values:
.TP 5
DB_MPOOL_CREATE
If the specified page does not exist, create it.
In this case, the
.I pgin
function, if specified, is called.
.TP 5
DB_MPOOL_LAST
Return the last page of the source file and copy its page number
to the location referenced by
.IR pgnoaddr .
.TP 5
DB_MPOOL_NEW
Create a new page in the file and copy its page number to the location
referenced by
.IR pgnoaddr .
In this case, the
.I pgin
function, if specified, is not called.
.PP
The DB_MPOOL_CREATE, DB_MPOOL_LAST and DB_MPOOL_NEW flags are mutually
exclusive.
.PP
Created pages have all their bytes set to 0.
.PP
All pages returned by
.I memp_fget
will be retained (i.e. ``pinned'') in the pool until a subsequent call to
.IR memp_fput .
.PP
.Rt memp_fget
.PP
.Fn memp_fput
The
.I memp_fput
function indicates that the page referenced by
.I pgaddr
can be evicted from the pool.
.I Pgaddr
must be an address previously returned by
.IR memp_fget .
.PP
The
.I flags
argument is specified by
.BR or 'ing
together one or more of the following values:
.TP 5
DB_MPOOL_CLEAN
Clear any previously set modification information (i.e.,
don't bother writing the page back to the source file).
.TP 5
DB_MPOOL_DIRTY
The page has been modified and must be written to the source file
before being evicted from the pool.
.TP 5
DB_MPOOL_DISCARD
The page is unlikely to be useful in the near future,
and should be discarded before other pages in the pool.
.PP
The DB_MPOOL_CLEAN and DB_MPOOL_DIRTY flags are mutually exclusive.
.PP
.Rt memp_fput
.PP
.Fn memp_fset
The
.I memp_fset
function sets the flags associated with the page referenced by
.I pgaddr
without unpinning it from the pool.
.I Pgaddr
must be an address previously returned by
.IR memp_fget .
The
.I flags
argument to
.I memp_fset
is specified by
.BR or 'ing
together one or more of the values specified as flags for the
.I memp_fput
call.
.PP
.Rt memp_fset
.PP
.Fn memp_fsync
The
.I memp_fsync
function writes all pages associated with the DB_MPOOLFILE pointer
.IR mpf ,
that were marked as modified using
.I memp_fput
or
.IR memp_fset ,
back to the source file.
If any of the modified pages are also pinned (i.e.,
currently referenced by this or another process)
.I memp_fsync
will ignore them.
.PP
.Rc memp_fsync
and DB_INCOMPLETE if there were pages which were modified but which
.I memp_fsync
was unable to write.
.PP
.Un "memory pool" mpool
.PP
.Fn memp_register
The
.I memp_register
function registers page-in and page-out functions for files of type
.I ftype
in the specified pool.
.PP
If the
.I pgin
function is non-NULL,
it is called each time a page is read into the memory pool from a file
of type
.IR ftype ,
or a page is created for a file of type
.I ftype
(see the DB_MPOOL_CREATE flag for the
.I memp_fget
function).
If the
.I pgout
function is non-NULL,
it is called each time a page is written to a file of type
.IR ftype .
.PP
Both the
.I pgin
and
.I pgout
functions are called with the page number,
a pointer to the page being read or written,
and any argument
.I pgcookie
that was specified to the
.I memp_fopen
function when the file was opened.
The
.I pgin
and
.I pgout
functions should return 0 on success,
and an applicable non-zero 
.I errno
value on failure,
in which case the
.I db_mpool
function calling it will also fail,
returning that
.I errno
value.
.PP
The purpose of the
.I memp_register
function is to support processing when pages are entered into,
or flushed from,
the pool.
A file type must be specified to make it possible for unrelated
threads or processes,
that are sharing a pool,
to evict each other's pages from the pool.
Applications should call
.IR memp_register ,
during initialization,
for each type of file requiring input or output processing that will be
sharing the underlying pool.
(No registry is necessary for the standard access method types,
btree, hash and recno, as
.IR db_open (3)
registers them separately.)
.PP
If a thread or process does not call
.I memp_register
for a file type,
it is impossible for it to evict pages for any file requiring input or
output processing from the pool.
For this reason,
.I memp_register
should always be called by each application sharing a pool for each type of
file included in the pool,
regardless of whether or not the application itself uses files of that type.
.PP
There are no standard values for
.IR ftype ,
.IR pgin ,
.I pgout
and
.IR pgcookie ,
except that the
.I ftype
value for a file must be a non-zero positive number,
as negative numbers are reserved for internal use by the DB library.
For this reason,
applications sharing a pool must coordinate their values amongst themselves.
.PP
.Rt memp_register
.PP
.Fn memp_trickle
The
.I memp_trickle
function ensures that at least
.I pct
percent of the pages in the shared memory pool are clean by writing dirty
pages to their backing files.
If the
.I nwrotep
argument is non-NULL,
the number of pages that were written to reach the correct percentage is
returned in the memory location it references.
.PP
The purpose of the
.I memp_trickle
function is to enable a memory pool manager to ensure that a page is
always available for reading in new information without having to wait
for a write.
.PP
.Rt memp_trickle
.PP
.Fn memp_sync
The
.I memp_sync
function ensures that all the modified pages in the pool with log sequence
numbers (LSNs) less than the
.I lsn
argument are written to disk.
.PP
.Rc memp_sync
and DB_INCOMPLETE if there were pages which need to be written but which
.I memp_sync
was unable to write immediately.
In addition,
if
.I memp_sync
returns success,
the value of
.I lsn
will be overwritten with the largest LSN from any page which was written by
.I memp_sync
to satisfy this request.
.PP
The purpose of the
.I memp_sync
function is to enable a transaction manager to ensure,
as part of a checkpoint,
that all pages modified by a certain time have been written to disk.
Pages in the pool which cannot be written back to disk immediately (e.g.,
are currently pinned) are written to disk as soon as it is possible to do
so.
The expected behavior of the transaction manager is to call the
.I memp_sync
function and then,
if the return indicates that some pages could not be written immediately,
to wait briefly and retry again with the same LSN until the
.I memp_sync
function returns that all pages have been written.
.PP
To support the
.I memp_sync
functionality,
it is necessary that the pool functions know the location of the LSN on
the page for each file type.
This location should be specified when the file is opened using the
.I memp_fopen
function.
(Note, it is not required that the LSN be aligned on the page in any way.)
.PP
.Fn memp_stat
The
.I memp_stat
function creates statistical structures and copies pointers to them into
user-specified memory locations.
The statistics include the number of files participating in the pool,
the active pages in the pool,
and information as to how effective the cache has been.
.PP
.Ma "Statistical structures"
.PP
If
.I gsp
is non-NULL, the global statistics for the memory pool
.I mp
are copied into the memory location it references.
The global statistics are stored in a structure of type
DB_MPOOL_STAT (typedef'd in <db2/db.h>).
.PP
The following DB_MPOOL_STAT fields will be filled in:
.TP 5
size_t st_cachesize;
Cache size in bytes.
.Nt
u_int32_t st_cache_hit;
Requested pages found in the cache.
.Nt
u_int32_t st_cache_miss;
Requested pages not found in the cache.
.Nt
u_int32_t st_map;
Requested pages mapped into the process' address space (there is no
available information as to whether or not this request caused disk I/O,
although examining the application page fault rate may be helpful).
.Nt
u_int32_t st_page_create;
Pages created in the cache.
.Nt
u_int32_t st_page_in;
Pages read into the cache.
.Nt
u_int32_t st_page_out;
Pages written from the cache to the backing file.
.Nt
u_int32_t st_ro_evict;
Clean pages forced from the cache.
.Nt
u_int32_t st_rw_evict;
Dirty pages forced from the cache.
.Nt
u_int32_t st_hash_buckets;
Number of hash buckets in buffer hash table.
.Nt
u_int32_t st_hash_searches;
Total number of buffer hash table lookups.
.Nt
u_int32_t st_hash_longest;
The longest chain ever encountered in buffer hash table lookups.
.Nt
u_int32_t st_hash_examined;
Total number of hash elements traversed during hash table lookups.
.Nt
u_int32_t st_page_clean;
Clean pages currently in the cache.
.Nt
u_int32_t st_page_dirty;
Dirty pages currently in the cache.
.Nt
u_int32_t st_page_trickle;
Dirty pages written using the
.I memp_trickle
interface.
.Nt
u_int32_t st_region_wait;
The number of times that a thread of control was forced to wait before
obtaining the region lock.
.Nt
u_int32_t st_region_nowait;
The number of times that a thread of control was able to obtain
the region lock without waiting.
.PP
If
.I fsp
is non-NULL,
a pointer to a NULL-terminated variable length array of statistics for
individual files,
in the memory pool
.IR mp ,
is copied into the memory location it references.
If no individual files currently exist in the memory pool,
.I fsp
will be set to NULL.
.PP
The per-file statistics are stored in structures of type
DB_MPOOL_FSTAT (typedef'd in <db2/db.h>).
The following DB_MPOOL_FSTAT fields will be filled in for each file in the
pool, i.e., each element of the array:
.TP 5
char *file_name;
The name of the file.
.Nt
size_t st_pagesize;
Page size in bytes.
.Nt
u_int32_t st_cache_hit;
Requested pages found in the cache.
.Nt
u_int32_t st_cache_miss;
Requested pages not found in the cache.
.Nt
u_int32_t st_map;
Requested pages mapped into the process' address space.
.Nt
u_int32_t st_page_create;
Pages created in the cache.
.Nt
u_int32_t st_page_in;
Pages read into the cache.
.Nt
u_int32_t st_page_out;
Pages written from the cache to the backing file.
.PP
.Rt memp_stat
.SH "ENVIRONMENT VARIABLES"
The following environment variables affect the execution of
.IR db_mpool :
.TP 5
.Eh memp_open
.TP 5
.Ev "memory pool" mpool
.SH ERRORS
.Ee memp_open
.na
.Nh
DBmemp->pgin(3), 
DBmemp->pgout(3), 
close(2), 
db_version(3), 
fcntl(2), 
fflush(3), 
fsync(2), 
log_compare(3), 
log_flush(3), 
lseek(2), 
malloc(3), 
memcmp(3), 
memcpy(3), 
memp_close(3), 
memp_unlink(3), 
memset(3), 
mmap(2), 
munmap(2), 
open(2), 
sigfillset(3), 
sigprocmask(2), 
stat(2), 
strcpy(3), 
strdup(3), 
strerror(3), 
strlen(3), 
time(3), 
unlink(2), 
and
write(2). 
.Hy
.ad
.PP
.Ec memp_open
.TP 5
.Em
.TP 5
.Ei
.sp
.Et
.sp
A NULL pathname was specified without the DB_MPOOL_PRIVATE flag.
.sp
The specified cache size was impossibly small.
.PP
.Ee memp_close
.na
.Nh
close(2), 
fcntl(2), 
fflush(3), 
memp_fclose(3), 
munmap(2), 
and
strerror(3). 
.Hy
.ad
.PP
.Ee memp_fopen
.na
.Nh
DBmemp->pgin(3), 
DBmemp->pgout(3), 
close(2), 
fcntl(2), 
fflush(3), 
fsync(2), 
log_compare(3), 
log_flush(3), 
lseek(2), 
malloc(3), 
memcmp(3), 
memcpy(3), 
memset(3), 
mmap(2), 
open(2), 
sigfillset(3), 
sigprocmask(2), 
stat(2), 
strcpy(3), 
strdup(3), 
strerror(3), 
strlen(3), 
time(3), 
unlink(2), 
and
write(2). 
.Hy
.ad
.PP
.Ec memp_fopen
.TP 5
.Ei
.sp
The file has already been entered into the pool,
and the
.I pagesize
value is not the same as when the file was entered into the pool,
or the length of the file is not zero or a multiple of the
.IR pagesize .
.sp
The DB_RDONLY flag was specified for an in-memory pool.
.PP
.Ee memp_fclose
.na
.Nh
close(2), 
fcntl(2), 
fflush(3), 
munmap(2), 
and
strerror(3). 
.Hy
.ad
.PP
.Ee memp_fget
.na
.Nh
DBmemp->pgin(3), 
DBmemp->pgout(3), 
close(2), 
fcntl(2), 
fflush(3), 
fsync(2), 
log_compare(3), 
log_flush(3), 
lseek(2), 
malloc(3), 
memcmp(3), 
memcpy(3), 
memset(3), 
mmap(2), 
open(2), 
read(2), 
sigfillset(3), 
sigprocmask(2), 
stat(2), 
strcpy(3), 
strdup(3), 
strerror(3), 
strlen(3), 
time(3), 
unlink(2), 
and
write(2). 
.Hy
.ad
.PP
.Ec memp_fget
.TP 5
[EAGAIN]
The page reference count has overflowed.
(This should never happen unless there's a bug in the application.)
.TP 5
.Ei
.sp
The DB_MPOOL_NEW flag was set and the source file was not opened for writing.
.sp
The requested page does not exist and DB_MPOOL_CREATE was not set.
.sp
More than one of DB_MPOOL_CREATE, DB_MPOOL_LAST and DB_MPOOL_NEW was set.
.TP 5
[ENOMEM]
The cache is full and no more pages will fit in the pool.
.PP
.Ee memp_fput
.na
.Nh
DBmemp->pgin(3), 
DBmemp->pgout(3), 
close(2), 
fcntl(2), 
fflush(3), 
fsync(2), 
log_compare(3), 
log_flush(3), 
lseek(2), 
malloc(3), 
memcmp(3), 
memcpy(3), 
memset(3), 
mmap(2), 
open(2), 
sigfillset(3), 
sigprocmask(2), 
stat(2), 
strcpy(3), 
strdup(3), 
strerror(3), 
strlen(3), 
time(3), 
unlink(2), 
and
write(2). 
.Hy
.ad
.PP
.Ec memp_fput
.TP 5
[EACCES]
The DB_MPOOL_DIRTY flag was set and the source file was not opened for
writing.
.TP 5
.Ei
.sp
The
.I pgaddr
parameter does not reference a page returned by
.IR memp_fget .
.sp
More than one of DB_MPOOL_CLEAN and DB_MPOOL_DIRTY was set.
.PP
.Ee memp_fset
.na
.Nh
fcntl(2), 
and
fflush(3). 
.Hy
.ad
.PP
.Ec memp_fset
.TP 5
.Ei
.PP
.Ee memp_fsync
.na
.Nh
DBmemp->pgin(3), 
DBmemp->pgout(3), 
close(2), 
fcntl(2), 
fflush(3), 
fsync(2), 
log_compare(3), 
log_flush(3), 
lseek(2), 
malloc(3), 
memcpy(3), 
memset(3), 
open(2), 
qsort(3), 
realloc(3), 
sigfillset(3), 
sigprocmask(2), 
stat(2), 
strcpy(3), 
strdup(3), 
strerror(3), 
strlen(3), 
unlink(2), 
and
write(2). 
.Hy
.ad
.PP
.Ee memp_unlink
.na
.Nh
close(2), 
fcntl(2), 
fflush(3), 
malloc(3), 
memcpy(3), 
memset(3), 
mmap(2), 
munmap(2), 
open(2), 
sigfillset(3), 
sigprocmask(2), 
stat(2), 
strcpy(3), 
strdup(3), 
strerror(3), 
strlen(3), 
and
unlink(2). 
.Hy
.ad
.PP
.Ec memp_unlink
.TP 5
.Eb
.PP
.Ee memp_register
.na
.Nh
fcntl(2), 
and
malloc(3). 
.Hy
.ad
.PP
.Ee memp_trickle
.na
.Nh
DBmemp->pgin(3), 
DBmemp->pgout(3), 
close(2), 
fcntl(2), 
fflush(3), 
fsync(2), 
log_compare(3), 
log_flush(3), 
lseek(2), 
malloc(3), 
memcmp(3), 
memcpy(3), 
memset(3), 
mmap(2), 
open(2), 
sigfillset(3), 
sigprocmask(2), 
stat(2), 
strcpy(3), 
strdup(3), 
strerror(3), 
strlen(3), 
time(3), 
unlink(2), 
and
write(2). 
.Hy
.ad
.PP
.Ec memp_trickle
.TP 5
.Ei
.PP
.Ee memp_sync
.na
.Nh
DBmemp->pgin(3), 
DBmemp->pgout(3), 
close(2), 
fcntl(2), 
fflush(3), 
fsync(2), 
log_compare(3), 
log_flush(3), 
lseek(2), 
malloc(3), 
memcmp(3), 
memcpy(3), 
memset(3), 
mmap(2), 
open(2), 
qsort(3), 
realloc(3), 
sigfillset(3), 
sigprocmask(2), 
stat(2), 
strcpy(3), 
strdup(3), 
strerror(3), 
strlen(3), 
time(3), 
unlink(2), 
and
write(2). 
.Hy
.ad
.PP
.Ec memp_sync
.TP 5
.Ei
.sp
The
.I memp_sync
function was called without logging having been initialized in the environment.
.PP
.Ee memp_stat
.na
.Nh
fcntl(2), 
malloc(3), 
memcpy(3), 
and
strlen(3). 
.Hy
.ad
.SH "SEE ALSO"
.Sa