diff --git a/libraries/libmdb/.gitignore b/libraries/libmdb/.gitignore
new file mode 100644
index 0000000000000000000000000000000000000000..134c457f7da637b19196ef906ea3ebe71db3c449
--- /dev/null
+++ b/libraries/libmdb/.gitignore
@@ -0,0 +1,15 @@
+mtest
+mtest[23456]
+testdb
+mdb_stat
+*.[ao]
+*.so
+*[~#]
+*.bak
+*.orig
+*.rej
+core
+core.*
+valgrind.*
+man/
+html/
diff --git a/libraries/libmdb/COPYRIGHT b/libraries/libmdb/COPYRIGHT
new file mode 100644
index 0000000000000000000000000000000000000000..28b83dd2d21aafa144aaa48e051bcec7d0fb3a51
--- /dev/null
+++ b/libraries/libmdb/COPYRIGHT
@@ -0,0 +1,20 @@
+Copyright 2011 Howard Chu, Symas Corp.
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted only as authorized by the OpenLDAP
+Public License.
+
+A copy of this license is available in the file LICENSE in the
+top-level directory of the distribution or, alternatively, at
+<http://www.OpenLDAP.org/license.html>.
+
+OpenLDAP is a registered trademark of the OpenLDAP Foundation.
+
+Individual files and/or contributed packages may be copyright by
+other parties and/or subject to additional restrictions.
+
+This work also contains materials derived from public sources.
+
+Additional information about OpenLDAP can be obtained at
+<http://www.openldap.org/>.
diff --git a/libraries/libmdb/Doxyfile b/libraries/libmdb/Doxyfile
new file mode 100644
index 0000000000000000000000000000000000000000..aa2c3058357d2f1dcae99bf31cf10ecf7dae5078
--- /dev/null
+++ b/libraries/libmdb/Doxyfile
@@ -0,0 +1,1631 @@
+# Doxyfile 1.7.1
+
+# This file describes the settings to be used by the documentation system
+# doxygen (www.doxygen.org) for a project
+#
+# All text after a hash (#) is considered a comment and will be ignored
+# The format is:
+# TAG = value [value, ...]
+# For lists items can also be appended using:
+# TAG += value [value, ...]
+# Values that contain spaces should be placed between quotes (" ")
+
+#---------------------------------------------------------------------------
+# Project related configuration options
+#---------------------------------------------------------------------------
+
+# This tag specifies the encoding used for all characters in the config file
+# that follow. The default is UTF-8 which is also the encoding used for all
+# text before the first occurrence of this tag. Doxygen uses libiconv (or the
+# iconv built into libc) for the transcoding. See
+# http://www.gnu.org/software/libiconv for the list of possible encodings.
+
+DOXYFILE_ENCODING = UTF-8
+
+# The PROJECT_NAME tag is a single word (or a sequence of words surrounded
+# by quotes) that should identify the project.
+
+PROJECT_NAME = MDB
+
+# The PROJECT_NUMBER tag can be used to enter a project or revision number.
+# This could be handy for archiving the generated documentation or
+# if some version control system is used.
+
+PROJECT_NUMBER =
+
+# The OUTPUT_DIRECTORY tag is used to specify the (relative or absolute)
+# base path where the generated documentation will be put.
+# If a relative path is entered, it will be relative to the location
+# where doxygen was started. If left blank the current directory will be used.
+
+OUTPUT_DIRECTORY =
+
+# If the CREATE_SUBDIRS tag is set to YES, then doxygen will create
+# 4096 sub-directories (in 2 levels) under the output directory of each output
+# format and will distribute the generated files over these directories.
+# Enabling this option can be useful when feeding doxygen a huge amount of
+# source files, where putting all generated files in the same directory would
+# otherwise cause performance problems for the file system.
+
+CREATE_SUBDIRS = NO
+
+# The OUTPUT_LANGUAGE tag is used to specify the language in which all
+# documentation generated by doxygen is written. Doxygen will use this
+# information to generate all constant output in the proper language.
+# The default language is English, other supported languages are:
+# Afrikaans, Arabic, Brazilian, Catalan, Chinese, Chinese-Traditional,
+# Croatian, Czech, Danish, Dutch, Esperanto, Farsi, Finnish, French, German,
+# Greek, Hungarian, Italian, Japanese, Japanese-en (Japanese with English
+# messages), Korean, Korean-en, Lithuanian, Norwegian, Macedonian, Persian,
+# Polish, Portuguese, Romanian, Russian, Serbian, Serbian-Cyrilic, Slovak,
+# Slovene, Spanish, Swedish, Ukrainian, and Vietnamese.
+
+OUTPUT_LANGUAGE = English
+
+# If the BRIEF_MEMBER_DESC tag is set to YES (the default) Doxygen will
+# include brief member descriptions after the members that are listed in
+# the file and class documentation (similar to JavaDoc).
+# Set to NO to disable this.
+
+BRIEF_MEMBER_DESC = YES
+
+# If the REPEAT_BRIEF tag is set to YES (the default) Doxygen will prepend
+# the brief description of a member or function before the detailed description.
+# Note: if both HIDE_UNDOC_MEMBERS and BRIEF_MEMBER_DESC are set to NO, the
+# brief descriptions will be completely suppressed.
+
+REPEAT_BRIEF = YES
+
+# This tag implements a quasi-intelligent brief description abbreviator
+# that is used to form the text in various listings. Each string
+# in this list, if found as the leading text of the brief description, will be
+# stripped from the text and the result after processing the whole list, is
+# used as the annotated text. Otherwise, the brief description is used as-is.
+# If left blank, the following values are used ("$name" is automatically
+# replaced with the name of the entity): "The $name class" "The $name widget"
+# "The $name file" "is" "provides" "specifies" "contains"
+# "represents" "a" "an" "the"
+
+ABBREVIATE_BRIEF =
+
+# If the ALWAYS_DETAILED_SEC and REPEAT_BRIEF tags are both set to YES then
+# Doxygen will generate a detailed section even if there is only a brief
+# description.
+
+ALWAYS_DETAILED_SEC = NO
+
+# If the INLINE_INHERITED_MEMB tag is set to YES, doxygen will show all
+# inherited members of a class in the documentation of that class as if those
+# members were ordinary class members. Constructors, destructors and assignment
+# operators of the base classes will not be shown.
+
+INLINE_INHERITED_MEMB = NO
+
+# If the FULL_PATH_NAMES tag is set to YES then Doxygen will prepend the full
+# path before files name in the file list and in the header files. If set
+# to NO the shortest path that makes the file name unique will be used.
+
+FULL_PATH_NAMES = YES
+
+# If the FULL_PATH_NAMES tag is set to YES then the STRIP_FROM_PATH tag
+# can be used to strip a user-defined part of the path. Stripping is
+# only done if one of the specified strings matches the left-hand part of
+# the path. The tag can be used to show relative paths in the file list.
+# If left blank the directory from which doxygen is run is used as the
+# path to strip.
+
+STRIP_FROM_PATH =
+
+# The STRIP_FROM_INC_PATH tag can be used to strip a user-defined part of
+# the path mentioned in the documentation of a class, which tells
+# the reader which header file to include in order to use a class.
+# If left blank only the name of the header file containing the class
+# definition is used. Otherwise one should specify the include paths that
+# are normally passed to the compiler using the -I flag.
+
+STRIP_FROM_INC_PATH =
+
+# If the SHORT_NAMES tag is set to YES, doxygen will generate much shorter
+# (but less readable) file names. This can be useful is your file systems
+# doesn't support long names like on DOS, Mac, or CD-ROM.
+
+SHORT_NAMES = NO
+
+# If the JAVADOC_AUTOBRIEF tag is set to YES then Doxygen
+# will interpret the first line (until the first dot) of a JavaDoc-style
+# comment as the brief description. If set to NO, the JavaDoc
+# comments will behave just like regular Qt-style comments
+# (thus requiring an explicit @brief command for a brief description.)
+
+JAVADOC_AUTOBRIEF = NO
+
+# If the QT_AUTOBRIEF tag is set to YES then Doxygen will
+# interpret the first line (until the first dot) of a Qt-style
+# comment as the brief description. If set to NO, the comments
+# will behave just like regular Qt-style comments (thus requiring
+# an explicit \brief command for a brief description.)
+
+QT_AUTOBRIEF = NO
+
+# The MULTILINE_CPP_IS_BRIEF tag can be set to YES to make Doxygen
+# treat a multi-line C++ special comment block (i.e. a block of //! or ///
+# comments) as a brief description. This used to be the default behaviour.
+# The new default is to treat a multi-line C++ comment block as a detailed
+# description. Set this tag to YES if you prefer the old behaviour instead.
+
+MULTILINE_CPP_IS_BRIEF = NO
+
+# If the INHERIT_DOCS tag is set to YES (the default) then an undocumented
+# member inherits the documentation from any documented member that it
+# re-implements.
+
+INHERIT_DOCS = YES
+
+# If the SEPARATE_MEMBER_PAGES tag is set to YES, then doxygen will produce
+# a new page for each member. If set to NO, the documentation of a member will
+# be part of the file/class/namespace that contains it.
+
+SEPARATE_MEMBER_PAGES = NO
+
+# The TAB_SIZE tag can be used to set the number of spaces in a tab.
+# Doxygen uses this value to replace tabs by spaces in code fragments.
+
+TAB_SIZE = 4
+
+# This tag can be used to specify a number of aliases that acts
+# as commands in the documentation. An alias has the form "name=value".
+# For example adding "sideeffect=\par Side Effects:\n" will allow you to
+# put the command \sideeffect (or @sideeffect) in the documentation, which
+# will result in a user-defined paragraph with heading "Side Effects:".
+# You can put \n's in the value part of an alias to insert newlines.
+
+ALIASES =
+
+# Set the OPTIMIZE_OUTPUT_FOR_C tag to YES if your project consists of C
+# sources only. Doxygen will then generate output that is more tailored for C.
+# For instance, some of the names that are used will be different. The list
+# of all members will be omitted, etc.
+
+OPTIMIZE_OUTPUT_FOR_C = YES
+
+# Set the OPTIMIZE_OUTPUT_JAVA tag to YES if your project consists of Java
+# sources only. Doxygen will then generate output that is more tailored for
+# Java. For instance, namespaces will be presented as packages, qualified
+# scopes will look different, etc.
+
+OPTIMIZE_OUTPUT_JAVA = NO
+
+# Set the OPTIMIZE_FOR_FORTRAN tag to YES if your project consists of Fortran
+# sources only. Doxygen will then generate output that is more tailored for
+# Fortran.
+
+OPTIMIZE_FOR_FORTRAN = NO
+
+# Set the OPTIMIZE_OUTPUT_VHDL tag to YES if your project consists of VHDL
+# sources. Doxygen will then generate output that is tailored for
+# VHDL.
+
+OPTIMIZE_OUTPUT_VHDL = NO
+
+# Doxygen selects the parser to use depending on the extension of the files it
+# parses. With this tag you can assign which parser to use for a given extension.
+# Doxygen has a built-in mapping, but you can override or extend it using this
+# tag. The format is ext=language, where ext is a file extension, and language
+# is one of the parsers supported by doxygen: IDL, Java, Javascript, CSharp, C,
+# C++, D, PHP, Objective-C, Python, Fortran, VHDL, C, C++. For instance to make
+# doxygen treat .inc files as Fortran files (default is PHP), and .f files as C
+# (default is Fortran), use: inc=Fortran f=C. Note that for custom extensions
+# you also need to set FILE_PATTERNS otherwise the files are not read by doxygen.
+
+EXTENSION_MAPPING =
+
+# If you use STL classes (i.e. std::string, std::vector, etc.) but do not want
+# to include (a tag file for) the STL sources as input, then you should
+# set this tag to YES in order to let doxygen match functions declarations and
+# definitions whose arguments contain STL classes (e.g. func(std::string); v.s.
+# func(std::string) {}). This also make the inheritance and collaboration
+# diagrams that involve STL classes more complete and accurate.
+
+BUILTIN_STL_SUPPORT = NO
+
+# If you use Microsoft's C++/CLI language, you should set this option to YES to
+# enable parsing support.
+
+CPP_CLI_SUPPORT = NO
+
+# Set the SIP_SUPPORT tag to YES if your project consists of sip sources only.
+# Doxygen will parse them like normal C++ but will assume all classes use public
+# instead of private inheritance when no explicit protection keyword is present.
+
+SIP_SUPPORT = NO
+
+# For Microsoft's IDL there are propget and propput attributes to indicate getter
+# and setter methods for a property. Setting this option to YES (the default)
+# will make doxygen to replace the get and set methods by a property in the
+# documentation. This will only work if the methods are indeed getting or
+# setting a simple type. If this is not the case, or you want to show the
+# methods anyway, you should set this option to NO.
+
+IDL_PROPERTY_SUPPORT = YES
+
+# If member grouping is used in the documentation and the DISTRIBUTE_GROUP_DOC
+# tag is set to YES, then doxygen will reuse the documentation of the first
+# member in the group (if any) for the other members of the group. By default
+# all members of a group must be documented explicitly.
+
+DISTRIBUTE_GROUP_DOC = NO
+
+# Set the SUBGROUPING tag to YES (the default) to allow class member groups of
+# the same type (for instance a group of public functions) to be put as a
+# subgroup of that type (e.g. under the Public Functions section). Set it to
+# NO to prevent subgrouping. Alternatively, this can be done per class using
+# the \nosubgrouping command.
+
+SUBGROUPING = YES
+
+INLINE_GROUPED_CLASSES = YES
+# When TYPEDEF_HIDES_STRUCT is enabled, a typedef of a struct, union, or enum
+# is documented as struct, union, or enum with the name of the typedef. So
+# typedef struct TypeS {} TypeT, will appear in the documentation as a struct
+# with name TypeT. When disabled the typedef will appear as a member of a file,
+# namespace, or class. And the struct will be named TypeS. This can typically
+# be useful for C code in case the coding convention dictates that all compound
+# types are typedef'ed and only the typedef is referenced, never the tag name.
+
+TYPEDEF_HIDES_STRUCT = YES
+
+# The SYMBOL_CACHE_SIZE determines the size of the internal cache use to
+# determine which symbols to keep in memory and which to flush to disk.
+# When the cache is full, less often used symbols will be written to disk.
+# For small to medium size projects (<1000 input files) the default value is
+# probably good enough. For larger projects a too small cache size can cause
+# doxygen to be busy swapping symbols to and from disk most of the time
+# causing a significant performance penality.
+# If the system has enough physical memory increasing the cache will improve the
+# performance by keeping more symbols in memory. Note that the value works on
+# a logarithmic scale so increasing the size by one will rougly double the
+# memory usage. The cache size is given by this formula:
+# 2^(16+SYMBOL_CACHE_SIZE). The valid range is 0..9, the default is 0,
+# corresponding to a cache size of 2^16 = 65536 symbols
+
+SYMBOL_CACHE_SIZE = 0
+
+#---------------------------------------------------------------------------
+# Build related configuration options
+#---------------------------------------------------------------------------
+
+# If the EXTRACT_ALL tag is set to YES doxygen will assume all entities in
+# documentation are documented, even if no documentation was available.
+# Private class members and static file members will be hidden unless
+# the EXTRACT_PRIVATE and EXTRACT_STATIC tags are set to YES
+
+EXTRACT_ALL = NO
+
+# If the EXTRACT_PRIVATE tag is set to YES all private members of a class
+# will be included in the documentation.
+
+EXTRACT_PRIVATE = NO
+
+# If the EXTRACT_STATIC tag is set to YES all static members of a file
+# will be included in the documentation.
+
+EXTRACT_STATIC = YES
+
+# If the EXTRACT_LOCAL_CLASSES tag is set to YES classes (and structs)
+# defined locally in source files will be included in the documentation.
+# If set to NO only classes defined in header files are included.
+
+EXTRACT_LOCAL_CLASSES = YES
+
+# This flag is only useful for Objective-C code. When set to YES local
+# methods, which are defined in the implementation section but not in
+# the interface are included in the documentation.
+# If set to NO (the default) only methods in the interface are included.
+
+EXTRACT_LOCAL_METHODS = NO
+
+# If this flag is set to YES, the members of anonymous namespaces will be
+# extracted and appear in the documentation as a namespace called
+# 'anonymous_namespace{file}', where file will be replaced with the base
+# name of the file that contains the anonymous namespace. By default
+# anonymous namespace are hidden.
+
+EXTRACT_ANON_NSPACES = NO
+
+# If the HIDE_UNDOC_MEMBERS tag is set to YES, Doxygen will hide all
+# undocumented members of documented classes, files or namespaces.
+# If set to NO (the default) these members will be included in the
+# various overviews, but no documentation section is generated.
+# This option has no effect if EXTRACT_ALL is enabled.
+
+HIDE_UNDOC_MEMBERS = NO
+
+# If the HIDE_UNDOC_CLASSES tag is set to YES, Doxygen will hide all
+# undocumented classes that are normally visible in the class hierarchy.
+# If set to NO (the default) these classes will be included in the various
+# overviews. This option has no effect if EXTRACT_ALL is enabled.
+
+HIDE_UNDOC_CLASSES = NO
+
+# If the HIDE_FRIEND_COMPOUNDS tag is set to YES, Doxygen will hide all
+# friend (class|struct|union) declarations.
+# If set to NO (the default) these declarations will be included in the
+# documentation.
+
+HIDE_FRIEND_COMPOUNDS = NO
+
+# If the HIDE_IN_BODY_DOCS tag is set to YES, Doxygen will hide any
+# documentation blocks found inside the body of a function.
+# If set to NO (the default) these blocks will be appended to the
+# function's detailed documentation block.
+
+HIDE_IN_BODY_DOCS = NO
+
+# The INTERNAL_DOCS tag determines if documentation
+# that is typed after a \internal command is included. If the tag is set
+# to NO (the default) then the documentation will be excluded.
+# Set it to YES to include the internal documentation.
+
+INTERNAL_DOCS = NO
+
+# If the CASE_SENSE_NAMES tag is set to NO then Doxygen will only generate
+# file names in lower-case letters. If set to YES upper-case letters are also
+# allowed. This is useful if you have classes or files whose names only differ
+# in case and if your file system supports case sensitive file names. Windows
+# and Mac users are advised to set this option to NO.
+
+CASE_SENSE_NAMES = YES
+
+# If the HIDE_SCOPE_NAMES tag is set to NO (the default) then Doxygen
+# will show members with their full class and namespace scopes in the
+# documentation. If set to YES the scope will be hidden.
+
+HIDE_SCOPE_NAMES = NO
+
+# If the SHOW_INCLUDE_FILES tag is set to YES (the default) then Doxygen
+# will put a list of the files that are included by a file in the documentation
+# of that file.
+
+SHOW_INCLUDE_FILES = YES
+
+# If the FORCE_LOCAL_INCLUDES tag is set to YES then Doxygen
+# will list include files with double quotes in the documentation
+# rather than with sharp brackets.
+
+FORCE_LOCAL_INCLUDES = NO
+
+# If the INLINE_INFO tag is set to YES (the default) then a tag [inline]
+# is inserted in the documentation for inline members.
+
+INLINE_INFO = YES
+
+# If the SORT_MEMBER_DOCS tag is set to YES (the default) then doxygen
+# will sort the (detailed) documentation of file and class members
+# alphabetically by member name. If set to NO the members will appear in
+# declaration order.
+
+SORT_MEMBER_DOCS = NO
+
+# If the SORT_BRIEF_DOCS tag is set to YES then doxygen will sort the
+# brief documentation of file, namespace and class members alphabetically
+# by member name. If set to NO (the default) the members will appear in
+# declaration order.
+
+SORT_BRIEF_DOCS = NO
+
+# If the SORT_MEMBERS_CTORS_1ST tag is set to YES then doxygen
+# will sort the (brief and detailed) documentation of class members so that
+# constructors and destructors are listed first. If set to NO (the default)
+# the constructors will appear in the respective orders defined by
+# SORT_MEMBER_DOCS and SORT_BRIEF_DOCS.
+# This tag will be ignored for brief docs if SORT_BRIEF_DOCS is set to NO
+# and ignored for detailed docs if SORT_MEMBER_DOCS is set to NO.
+
+SORT_MEMBERS_CTORS_1ST = NO
+
+# If the SORT_GROUP_NAMES tag is set to YES then doxygen will sort the
+# hierarchy of group names into alphabetical order. If set to NO (the default)
+# the group names will appear in their defined order.
+
+SORT_GROUP_NAMES = NO
+
+# If the SORT_BY_SCOPE_NAME tag is set to YES, the class list will be
+# sorted by fully-qualified names, including namespaces. If set to
+# NO (the default), the class list will be sorted only by class name,
+# not including the namespace part.
+# Note: This option is not very useful if HIDE_SCOPE_NAMES is set to YES.
+# Note: This option applies only to the class list, not to the
+# alphabetical list.
+
+SORT_BY_SCOPE_NAME = NO
+
+# The GENERATE_TODOLIST tag can be used to enable (YES) or
+# disable (NO) the todo list. This list is created by putting \todo
+# commands in the documentation.
+
+GENERATE_TODOLIST = YES
+
+# The GENERATE_TESTLIST tag can be used to enable (YES) or
+# disable (NO) the test list. This list is created by putting \test
+# commands in the documentation.
+
+GENERATE_TESTLIST = YES
+
+# The GENERATE_BUGLIST tag can be used to enable (YES) or
+# disable (NO) the bug list. This list is created by putting \bug
+# commands in the documentation.
+
+GENERATE_BUGLIST = YES
+
+# The GENERATE_DEPRECATEDLIST tag can be used to enable (YES) or
+# disable (NO) the deprecated list. This list is created by putting
+# \deprecated commands in the documentation.
+
+GENERATE_DEPRECATEDLIST= YES
+
+# The ENABLED_SECTIONS tag can be used to enable conditional
+# documentation sections, marked by \if sectionname ... \endif.
+
+ENABLED_SECTIONS =
+
+# The MAX_INITIALIZER_LINES tag determines the maximum number of lines
+# the initial value of a variable or define consists of for it to appear in
+# the documentation. If the initializer consists of more lines than specified
+# here it will be hidden. Use a value of 0 to hide initializers completely.
+# The appearance of the initializer of individual variables and defines in the
+# documentation can be controlled using \showinitializer or \hideinitializer
+# command in the documentation regardless of this setting.
+
+MAX_INITIALIZER_LINES = 30
+
+# Set the SHOW_USED_FILES tag to NO to disable the list of files generated
+# at the bottom of the documentation of classes and structs. If set to YES the
+# list will mention the files that were used to generate the documentation.
+
+SHOW_USED_FILES = YES
+
+# If the sources in your project are distributed over multiple directories
+# then setting the SHOW_DIRECTORIES tag to YES will show the directory hierarchy
+# in the documentation. The default is NO.
+
+SHOW_DIRECTORIES = NO
+
+# Set the SHOW_FILES tag to NO to disable the generation of the Files page.
+# This will remove the Files entry from the Quick Index and from the
+# Folder Tree View (if specified). The default is YES.
+
+SHOW_FILES = YES
+
+# Set the SHOW_NAMESPACES tag to NO to disable the generation of the
+# Namespaces page.
+# This will remove the Namespaces entry from the Quick Index
+# and from the Folder Tree View (if specified). The default is YES.
+
+SHOW_NAMESPACES = YES
+
+# The FILE_VERSION_FILTER tag can be used to specify a program or script that
+# doxygen should invoke to get the current version for each file (typically from
+# the version control system). Doxygen will invoke the program by executing (via
+# popen()) the command <command> <input-file>, where <command> is the value of
+# the FILE_VERSION_FILTER tag, and <input-file> is the name of an input file
+# provided by doxygen. Whatever the program writes to standard output
+# is used as the file version. See the manual for examples.
+
+FILE_VERSION_FILTER =
+
+# The LAYOUT_FILE tag can be used to specify a layout file which will be parsed
+# by doxygen. The layout file controls the global structure of the generated
+# output files in an output format independent way. The create the layout file
+# that represents doxygen's defaults, run doxygen with the -l option.
+# You can optionally specify a file name after the option, if omitted
+# DoxygenLayout.xml will be used as the name of the layout file.
+
+LAYOUT_FILE =
+
+#---------------------------------------------------------------------------
+# configuration options related to warning and progress messages
+#---------------------------------------------------------------------------
+
+# The QUIET tag can be used to turn on/off the messages that are generated
+# by doxygen. Possible values are YES and NO. If left blank NO is used.
+
+QUIET = NO
+
+# The WARNINGS tag can be used to turn on/off the warning messages that are
+# generated by doxygen. Possible values are YES and NO. If left blank
+# NO is used.
+
+WARNINGS = YES
+
+# If WARN_IF_UNDOCUMENTED is set to YES, then doxygen will generate warnings
+# for undocumented members. If EXTRACT_ALL is set to YES then this flag will
+# automatically be disabled.
+
+WARN_IF_UNDOCUMENTED = YES
+
+# If WARN_IF_DOC_ERROR is set to YES, doxygen will generate warnings for
+# potential errors in the documentation, such as not documenting some
+# parameters in a documented function, or documenting parameters that
+# don't exist or using markup commands wrongly.
+
+WARN_IF_DOC_ERROR = YES
+
+# This WARN_NO_PARAMDOC option can be abled to get warnings for
+# functions that are documented, but have no documentation for their parameters
+# or return value. If set to NO (the default) doxygen will only warn about
+# wrong or incomplete parameter documentation, but not about the absence of
+# documentation.
+
+WARN_NO_PARAMDOC = NO
+
+# The WARN_FORMAT tag determines the format of the warning messages that
+# doxygen can produce. The string should contain the $file, $line, and $text
+# tags, which will be replaced by the file and line number from which the
+# warning originated and the warning text. Optionally the format may contain
+# $version, which will be replaced by the version of the file (if it could
+# be obtained via FILE_VERSION_FILTER)
+
+WARN_FORMAT = "$file:$line: $text"
+
+# The WARN_LOGFILE tag can be used to specify a file to which warning
+# and error messages should be written. If left blank the output is written
+# to stderr.
+
+WARN_LOGFILE =
+
+#---------------------------------------------------------------------------
+# configuration options related to the input files
+#---------------------------------------------------------------------------
+
+# The INPUT tag can be used to specify the files and/or directories that contain
+# documented source files. You may enter file names like "myfile.cpp" or
+# directories like "/usr/src/myproject". Separate the files or directories
+# with spaces.
+
+INPUT = mdb.h midl.h mdb.c midl.c
+
+# This tag can be used to specify the character encoding of the source files
+# that doxygen parses. Internally doxygen uses the UTF-8 encoding, which is
+# also the default input encoding. Doxygen uses libiconv (or the iconv built
+# into libc) for the transcoding. See http://www.gnu.org/software/libiconv for
+# the list of possible encodings.
+
+INPUT_ENCODING = UTF-8
+
+# If the value of the INPUT tag contains directories, you can use the
+# FILE_PATTERNS tag to specify one or more wildcard pattern (like *.cpp
+# and *.h) to filter out the source-files in the directories. If left
+# blank the following patterns are tested:
+# *.c *.cc *.cxx *.cpp *.c++ *.java *.ii *.ixx *.ipp *.i++ *.inl *.h *.hh *.hxx
+# *.hpp *.h++ *.idl *.odl *.cs *.php *.php3 *.inc *.m *.mm *.py *.f90
+
+FILE_PATTERNS =
+
+# The RECURSIVE tag can be used to turn specify whether or not subdirectories
+# should be searched for input files as well. Possible values are YES and NO.
+# If left blank NO is used.
+
+RECURSIVE = NO
+
+# The EXCLUDE tag can be used to specify files and/or directories that should
+# excluded from the INPUT source files. This way you can easily exclude a
+# subdirectory from a directory tree whose root is specified with the INPUT tag.
+
+EXCLUDE =
+
+# The EXCLUDE_SYMLINKS tag can be used select whether or not files or
+# directories that are symbolic links (a Unix filesystem feature) are excluded
+# from the input.
+
+EXCLUDE_SYMLINKS = NO
+
+# If the value of the INPUT tag contains directories, you can use the
+# EXCLUDE_PATTERNS tag to specify one or more wildcard patterns to exclude
+# certain files from those directories. Note that the wildcards are matched
+# against the file with absolute path, so to exclude all test directories
+# for example use the pattern */test/*
+
+EXCLUDE_PATTERNS =
+
+# The EXCLUDE_SYMBOLS tag can be used to specify one or more symbol names
+# (namespaces, classes, functions, etc.) that should be excluded from the
+# output. The symbol name can be a fully qualified name, a word, or if the
+# wildcard * is used, a substring. Examples: ANamespace, AClass,
+# AClass::ANamespace, ANamespace::*Test
+
+EXCLUDE_SYMBOLS =
+
+# The EXAMPLE_PATH tag can be used to specify one or more files or
+# directories that contain example code fragments that are included (see
+# the \include command).
+
+EXAMPLE_PATH =
+
+# If the value of the EXAMPLE_PATH tag contains directories, you can use the
+# EXAMPLE_PATTERNS tag to specify one or more wildcard pattern (like *.cpp
+# and *.h) to filter out the source-files in the directories. If left
+# blank all files are included.
+
+EXAMPLE_PATTERNS =
+
+# If the EXAMPLE_RECURSIVE tag is set to YES then subdirectories will be
+# searched for input files to be used with the \include or \dontinclude
+# commands irrespective of the value of the RECURSIVE tag.
+# Possible values are YES and NO. If left blank NO is used.
+
+EXAMPLE_RECURSIVE = NO
+
+# The IMAGE_PATH tag can be used to specify one or more files or
+# directories that contain image that are included in the documentation (see
+# the \image command).
+
+IMAGE_PATH =
+
+# The INPUT_FILTER tag can be used to specify a program that doxygen should
+# invoke to filter for each input file. Doxygen will invoke the filter program
+# by executing (via popen()) the command <filter> <input-file>, where <filter>
+# is the value of the INPUT_FILTER tag, and <input-file> is the name of an
+# input file. Doxygen will then use the output that the filter program writes
+# to standard output.
+# If FILTER_PATTERNS is specified, this tag will be
+# ignored.
+
+INPUT_FILTER =
+
+# The FILTER_PATTERNS tag can be used to specify filters on a per file pattern
+# basis.
+# Doxygen will compare the file name with each pattern and apply the
+# filter if there is a match.
+# The filters are a list of the form:
+# pattern=filter (like *.cpp=my_cpp_filter). See INPUT_FILTER for further
+# info on how filters are used. If FILTER_PATTERNS is empty, INPUT_FILTER
+# is applied to all files.
+
+FILTER_PATTERNS =
+
+# If the FILTER_SOURCE_FILES tag is set to YES, the input filter (if set using
+# INPUT_FILTER) will be used to filter the input files when producing source
+# files to browse (i.e. when SOURCE_BROWSER is set to YES).
+
+FILTER_SOURCE_FILES = NO
+
+#---------------------------------------------------------------------------
+# configuration options related to source browsing
+#---------------------------------------------------------------------------
+
+# If the SOURCE_BROWSER tag is set to YES then a list of source files will
+# be generated. Documented entities will be cross-referenced with these sources.
+# Note: To get rid of all source code in the generated output, make sure also
+# VERBATIM_HEADERS is set to NO.
+
+SOURCE_BROWSER = NO
+
+# Setting the INLINE_SOURCES tag to YES will include the body
+# of functions and classes directly in the documentation.
+
+INLINE_SOURCES = NO
+
+# Setting the STRIP_CODE_COMMENTS tag to YES (the default) will instruct
+# doxygen to hide any special comment blocks from generated source code
+# fragments. Normal C and C++ comments will always remain visible.
+
+STRIP_CODE_COMMENTS = YES
+
+# If the REFERENCED_BY_RELATION tag is set to YES
+# then for each documented function all documented
+# functions referencing it will be listed.
+
+REFERENCED_BY_RELATION = NO
+
+# If the REFERENCES_RELATION tag is set to YES
+# then for each documented function all documented entities
+# called/used by that function will be listed.
+
+REFERENCES_RELATION = NO
+
+# If the REFERENCES_LINK_SOURCE tag is set to YES (the default)
+# and SOURCE_BROWSER tag is set to YES, then the hyperlinks from
+# functions in REFERENCES_RELATION and REFERENCED_BY_RELATION lists will
+# link to the source code.
+# Otherwise they will link to the documentation.
+
+REFERENCES_LINK_SOURCE = YES
+
+# If the USE_HTAGS tag is set to YES then the references to source code
+# will point to the HTML generated by the htags(1) tool instead of doxygen
+# built-in source browser. The htags tool is part of GNU's global source
+# tagging system (see http://www.gnu.org/software/global/global.html). You
+# will need version 4.8.6 or higher.
+
+USE_HTAGS = NO
+
+# If the VERBATIM_HEADERS tag is set to YES (the default) then Doxygen
+# will generate a verbatim copy of the header file for each class for
+# which an include is specified. Set to NO to disable this.
+
+VERBATIM_HEADERS = YES
+
+#---------------------------------------------------------------------------
+# configuration options related to the alphabetical class index
+#---------------------------------------------------------------------------
+
+# If the ALPHABETICAL_INDEX tag is set to YES, an alphabetical index
+# of all compounds will be generated. Enable this if the project
+# contains a lot of classes, structs, unions or interfaces.
+
+ALPHABETICAL_INDEX = YES
+
+# If the alphabetical index is enabled (see ALPHABETICAL_INDEX) then
+# the COLS_IN_ALPHA_INDEX tag can be used to specify the number of columns
+# in which this list will be split (can be a number in the range [1..20])
+
+COLS_IN_ALPHA_INDEX = 5
+
+# In case all classes in a project start with a common prefix, all
+# classes will be put under the same header in the alphabetical index.
+# The IGNORE_PREFIX tag can be used to specify one or more prefixes that
+# should be ignored while generating the index headers.
+
+IGNORE_PREFIX =
+
+#---------------------------------------------------------------------------
+# configuration options related to the HTML output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_HTML tag is set to YES (the default) Doxygen will
+# generate HTML output.
+
+GENERATE_HTML = YES
+
+# The HTML_OUTPUT tag is used to specify where the HTML docs will be put.
+# If a relative path is entered the value of OUTPUT_DIRECTORY will be
+# put in front of it. If left blank `html' will be used as the default path.
+
+HTML_OUTPUT = html
+
+# The HTML_FILE_EXTENSION tag can be used to specify the file extension for
+# each generated HTML page (for example: .htm,.php,.asp). If it is left blank
+# doxygen will generate files with .html extension.
+
+HTML_FILE_EXTENSION = .html
+
+# The HTML_HEADER tag can be used to specify a personal HTML header for
+# each generated HTML page. If it is left blank doxygen will generate a
+# standard header.
+
+HTML_HEADER =
+
+# The HTML_FOOTER tag can be used to specify a personal HTML footer for
+# each generated HTML page. If it is left blank doxygen will generate a
+# standard footer.
+
+HTML_FOOTER =
+
+# The HTML_STYLESHEET tag can be used to specify a user-defined cascading
+# style sheet that is used by each HTML page. It can be used to
+# fine-tune the look of the HTML output. If the tag is left blank doxygen
+# will generate a default style sheet. Note that doxygen will try to copy
+# the style sheet file to the HTML output directory, so don't put your own
+# stylesheet in the HTML output directory as well, or it will be erased!
+
+HTML_STYLESHEET =
+
+# The HTML_COLORSTYLE_HUE tag controls the color of the HTML output.
+# Doxygen will adjust the colors in the stylesheet and background images
+# according to this color. Hue is specified as an angle on a colorwheel,
+# see http://en.wikipedia.org/wiki/Hue for more information.
+# For instance the value 0 represents red, 60 is yellow, 120 is green,
+# 180 is cyan, 240 is blue, 300 purple, and 360 is red again.
+# The allowed range is 0 to 359.
+
+HTML_COLORSTYLE_HUE = 220
+
+# The HTML_COLORSTYLE_SAT tag controls the purity (or saturation) of
+# the colors in the HTML output. For a value of 0 the output will use
+# grayscales only. A value of 255 will produce the most vivid colors.
+
+HTML_COLORSTYLE_SAT = 100
+
+# The HTML_COLORSTYLE_GAMMA tag controls the gamma correction applied to
+# the luminance component of the colors in the HTML output. Values below
+# 100 gradually make the output lighter, whereas values above 100 make
+# the output darker. The value divided by 100 is the actual gamma applied,
+# so 80 represents a gamma of 0.8, The value 220 represents a gamma of 2.2,
+# and 100 does not change the gamma.
+
+HTML_COLORSTYLE_GAMMA = 80
+
+# If the HTML_TIMESTAMP tag is set to YES then the footer of each generated HTML
+# page will contain the date and time when the page was generated. Setting
+# this to NO can help when comparing the output of multiple runs.
+
+HTML_TIMESTAMP = YES
+
+# If the HTML_ALIGN_MEMBERS tag is set to YES, the members of classes,
+# files or namespaces will be aligned in HTML using tables. If set to
+# NO a bullet list will be used.
+
+HTML_ALIGN_MEMBERS = YES
+
+# If the HTML_DYNAMIC_SECTIONS tag is set to YES then the generated HTML
+# documentation will contain sections that can be hidden and shown after the
+# page has loaded. For this to work a browser that supports
+# JavaScript and DHTML is required (for instance Mozilla 1.0+, Firefox
+# Netscape 6.0+, Internet explorer 5.0+, Konqueror, or Safari).
+
+HTML_DYNAMIC_SECTIONS = NO
+
+# If the GENERATE_DOCSET tag is set to YES, additional index files
+# will be generated that can be used as input for Apple's Xcode 3
+# integrated development environment, introduced with OSX 10.5 (Leopard).
+# To create a documentation set, doxygen will generate a Makefile in the
+# HTML output directory. Running make will produce the docset in that
+# directory and running "make install" will install the docset in
+# ~/Library/Developer/Shared/Documentation/DocSets so that Xcode will find
+# it at startup.
+# See http://developer.apple.com/tools/creatingdocsetswithdoxygen.html
+# for more information.
+
+GENERATE_DOCSET = NO
+
+# When GENERATE_DOCSET tag is set to YES, this tag determines the name of the
+# feed. A documentation feed provides an umbrella under which multiple
+# documentation sets from a single provider (such as a company or product suite)
+# can be grouped.
+
+DOCSET_FEEDNAME = "Doxygen generated docs"
+
+# When GENERATE_DOCSET tag is set to YES, this tag specifies a string that
+# should uniquely identify the documentation set bundle. This should be a
+# reverse domain-name style string, e.g. com.mycompany.MyDocSet. Doxygen
+# will append .docset to the name.
+
+DOCSET_BUNDLE_ID = org.doxygen.Project
+
+# When GENERATE_PUBLISHER_ID tag specifies a string that should uniquely identify
+# the documentation publisher. This should be a reverse domain-name style
+# string, e.g. com.mycompany.MyDocSet.documentation.
+
+DOCSET_PUBLISHER_ID = org.doxygen.Publisher
+
+# The GENERATE_PUBLISHER_NAME tag identifies the documentation publisher.
+
+DOCSET_PUBLISHER_NAME = Publisher
+
+# If the GENERATE_HTMLHELP tag is set to YES, additional index files
+# will be generated that can be used as input for tools like the
+# Microsoft HTML help workshop to generate a compiled HTML help file (.chm)
+# of the generated HTML documentation.
+
+GENERATE_HTMLHELP = NO
+
+# If the GENERATE_HTMLHELP tag is set to YES, the CHM_FILE tag can
+# be used to specify the file name of the resulting .chm file. You
+# can add a path in front of the file if the result should not be
+# written to the html output directory.
+
+CHM_FILE =
+
+# If the GENERATE_HTMLHELP tag is set to YES, the HHC_LOCATION tag can
+# be used to specify the location (absolute path including file name) of
+# the HTML help compiler (hhc.exe). If non-empty doxygen will try to run
+# the HTML help compiler on the generated index.hhp.
+
+HHC_LOCATION =
+
+# If the GENERATE_HTMLHELP tag is set to YES, the GENERATE_CHI flag
+# controls if a separate .chi index file is generated (YES) or that
+# it should be included in the master .chm file (NO).
+
+GENERATE_CHI = NO
+
+# If the GENERATE_HTMLHELP tag is set to YES, the CHM_INDEX_ENCODING
+# is used to encode HtmlHelp index (hhk), content (hhc) and project file
+# content.
+
+CHM_INDEX_ENCODING =
+
+# If the GENERATE_HTMLHELP tag is set to YES, the BINARY_TOC flag
+# controls whether a binary table of contents is generated (YES) or a
+# normal table of contents (NO) in the .chm file.
+
+BINARY_TOC = NO
+
+# The TOC_EXPAND flag can be set to YES to add extra items for group members
+# to the contents of the HTML help documentation and to the tree view.
+
+TOC_EXPAND = NO
+
+# If the GENERATE_QHP tag is set to YES and both QHP_NAMESPACE and
+# QHP_VIRTUAL_FOLDER are set, an additional index file will be generated
+# that can be used as input for Qt's qhelpgenerator to generate a
+# Qt Compressed Help (.qch) of the generated HTML documentation.
+
+GENERATE_QHP = NO
+
+# If the QHG_LOCATION tag is specified, the QCH_FILE tag can
+# be used to specify the file name of the resulting .qch file.
+# The path specified is relative to the HTML output folder.
+
+QCH_FILE =
+
+# The QHP_NAMESPACE tag specifies the namespace to use when generating
+# Qt Help Project output. For more information please see
+# http://doc.trolltech.com/qthelpproject.html#namespace
+
+QHP_NAMESPACE = org.doxygen.Project
+
+# The QHP_VIRTUAL_FOLDER tag specifies the namespace to use when generating
+# Qt Help Project output. For more information please see
+# http://doc.trolltech.com/qthelpproject.html#virtual-folders
+
+QHP_VIRTUAL_FOLDER = doc
+
+# If QHP_CUST_FILTER_NAME is set, it specifies the name of a custom filter to
+# add. For more information please see
+# http://doc.trolltech.com/qthelpproject.html#custom-filters
+
+QHP_CUST_FILTER_NAME =
+
+# The QHP_CUST_FILT_ATTRS tag specifies the list of the attributes of the
+# custom filter to add. For more information please see
+# <a href="http://doc.trolltech.com/qthelpproject.html#custom-filters">
+# Qt Help Project / Custom Filters</a>.
+
+QHP_CUST_FILTER_ATTRS =
+
+# The QHP_SECT_FILTER_ATTRS tag specifies the list of the attributes this
+# project's
+# filter section matches.
+# <a href="http://doc.trolltech.com/qthelpproject.html#filter-attributes">
+# Qt Help Project / Filter Attributes</a>.
+
+QHP_SECT_FILTER_ATTRS =
+
+# If the GENERATE_QHP tag is set to YES, the QHG_LOCATION tag can
+# be used to specify the location of Qt's qhelpgenerator.
+# If non-empty doxygen will try to run qhelpgenerator on the generated
+# .qhp file.
+
+QHG_LOCATION =
+
+# If the GENERATE_ECLIPSEHELP tag is set to YES, additional index files
+# will be generated, which together with the HTML files, form an Eclipse help
+# plugin. To install this plugin and make it available under the help contents
+# menu in Eclipse, the contents of the directory containing the HTML and XML
+# files needs to be copied into the plugins directory of eclipse. The name of
+# the directory within the plugins directory should be the same as
+# the ECLIPSE_DOC_ID value. After copying Eclipse needs to be restarted before
+# the help appears.
+
+GENERATE_ECLIPSEHELP = NO
+
+# A unique identifier for the eclipse help plugin. When installing the plugin
+# the directory name containing the HTML and XML files should also have
+# this name.
+
+ECLIPSE_DOC_ID = org.doxygen.Project
+
+# The DISABLE_INDEX tag can be used to turn on/off the condensed index at
+# top of each HTML page. The value NO (the default) enables the index and
+# the value YES disables it.
+
+DISABLE_INDEX = NO
+
+# This tag can be used to set the number of enum values (range [1..20])
+# that doxygen will group on one line in the generated HTML documentation.
+
+ENUM_VALUES_PER_LINE = 4
+
+# The GENERATE_TREEVIEW tag is used to specify whether a tree-like index
+# structure should be generated to display hierarchical information.
+# If the tag value is set to YES, a side panel will be generated
+# containing a tree-like index structure (just like the one that
+# is generated for HTML Help). For this to work a browser that supports
+# JavaScript, DHTML, CSS and frames is required (i.e. any modern browser).
+# Windows users are probably better off using the HTML help feature.
+
+GENERATE_TREEVIEW = NO
+
+# By enabling USE_INLINE_TREES, doxygen will generate the Groups, Directories,
+# and Class Hierarchy pages using a tree view instead of an ordered list.
+
+USE_INLINE_TREES = NO
+
+# If the treeview is enabled (see GENERATE_TREEVIEW) then this tag can be
+# used to set the initial width (in pixels) of the frame in which the tree
+# is shown.
+
+TREEVIEW_WIDTH = 250
+
+# When the EXT_LINKS_IN_WINDOW option is set to YES doxygen will open
+# links to external symbols imported via tag files in a separate window.
+
+EXT_LINKS_IN_WINDOW = NO
+
+# Use this tag to change the font size of Latex formulas included
+# as images in the HTML documentation. The default is 10. Note that
+# when you change the font size after a successful doxygen run you need
+# to manually remove any form_*.png images from the HTML output directory
+# to force them to be regenerated.
+
+FORMULA_FONTSIZE = 10
+
+# Use the FORMULA_TRANPARENT tag to determine whether or not the images
+# generated for formulas are transparent PNGs. Transparent PNGs are
+# not supported properly for IE 6.0, but are supported on all modern browsers.
+# Note that when changing this option you need to delete any form_*.png files
+# in the HTML output before the changes have effect.
+
+FORMULA_TRANSPARENT = YES
+
+# When the SEARCHENGINE tag is enabled doxygen will generate a search box
+# for the HTML output. The underlying search engine uses javascript
+# and DHTML and should work on any modern browser. Note that when using
+# HTML help (GENERATE_HTMLHELP), Qt help (GENERATE_QHP), or docsets
+# (GENERATE_DOCSET) there is already a search function so this one should
+# typically be disabled. For large projects the javascript based search engine
+# can be slow, then enabling SERVER_BASED_SEARCH may provide a better solution.
+
+SEARCHENGINE = YES
+
+# When the SERVER_BASED_SEARCH tag is enabled the search engine will be
+# implemented using a PHP enabled web server instead of at the web client
+# using Javascript. Doxygen will generate the search PHP script and index
+# file to put on the web server. The advantage of the server
+# based approach is that it scales better to large projects and allows
+# full text search. The disadvances is that it is more difficult to setup
+# and does not have live searching capabilities.
+
+SERVER_BASED_SEARCH = NO
+
+#---------------------------------------------------------------------------
+# configuration options related to the LaTeX output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_LATEX tag is set to YES (the default) Doxygen will
+# generate Latex output.
+
+GENERATE_LATEX = NO
+
+# The LATEX_OUTPUT tag is used to specify where the LaTeX docs will be put.
+# If a relative path is entered the value of OUTPUT_DIRECTORY will be
+# put in front of it. If left blank `latex' will be used as the default path.
+
+LATEX_OUTPUT = latex
+
+# The LATEX_CMD_NAME tag can be used to specify the LaTeX command name to be
+# invoked. If left blank `latex' will be used as the default command name.
+# Note that when enabling USE_PDFLATEX this option is only used for
+# generating bitmaps for formulas in the HTML output, but not in the
+# Makefile that is written to the output directory.
+
+LATEX_CMD_NAME = latex
+
+# The MAKEINDEX_CMD_NAME tag can be used to specify the command name to
+# generate index for LaTeX. If left blank `makeindex' will be used as the
+# default command name.
+
+MAKEINDEX_CMD_NAME = makeindex
+
+# If the COMPACT_LATEX tag is set to YES Doxygen generates more compact
+# LaTeX documents. This may be useful for small projects and may help to
+# save some trees in general.
+
+COMPACT_LATEX = NO
+
+# The PAPER_TYPE tag can be used to set the paper type that is used
+# by the printer. Possible values are: a4, a4wide, letter, legal and
+# executive. If left blank a4wide will be used.
+
+PAPER_TYPE = a4wide
+
+# The EXTRA_PACKAGES tag can be to specify one or more names of LaTeX
+# packages that should be included in the LaTeX output.
+
+EXTRA_PACKAGES =
+
+# The LATEX_HEADER tag can be used to specify a personal LaTeX header for
+# the generated latex document. The header should contain everything until
+# the first chapter. If it is left blank doxygen will generate a
+# standard header. Notice: only use this tag if you know what you are doing!
+
+LATEX_HEADER =
+
+# If the PDF_HYPERLINKS tag is set to YES, the LaTeX that is generated
+# is prepared for conversion to pdf (using ps2pdf). The pdf file will
+# contain links (just like the HTML output) instead of page references
+# This makes the output suitable for online browsing using a pdf viewer.
+
+PDF_HYPERLINKS = YES
+
+# If the USE_PDFLATEX tag is set to YES, pdflatex will be used instead of
+# plain latex in the generated Makefile. Set this option to YES to get a
+# higher quality PDF documentation.
+
+USE_PDFLATEX = YES
+
+# If the LATEX_BATCHMODE tag is set to YES, doxygen will add the \\batchmode.
+# command to the generated LaTeX files. This will instruct LaTeX to keep
+# running if errors occur, instead of asking the user for help.
+# This option is also used when generating formulas in HTML.
+
+LATEX_BATCHMODE = NO
+
+# If LATEX_HIDE_INDICES is set to YES then doxygen will not
+# include the index chapters (such as File Index, Compound Index, etc.)
+# in the output.
+
+LATEX_HIDE_INDICES = NO
+
+# If LATEX_SOURCE_CODE is set to YES then doxygen will include
+# source code with syntax highlighting in the LaTeX output.
+# Note that which sources are shown also depends on other settings
+# such as SOURCE_BROWSER.
+
+LATEX_SOURCE_CODE = NO
+
+#---------------------------------------------------------------------------
+# configuration options related to the RTF output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_RTF tag is set to YES Doxygen will generate RTF output
+# The RTF output is optimized for Word 97 and may not look very pretty with
+# other RTF readers or editors.
+
+GENERATE_RTF = NO
+
+# The RTF_OUTPUT tag is used to specify where the RTF docs will be put.
+# If a relative path is entered the value of OUTPUT_DIRECTORY will be
+# put in front of it. If left blank `rtf' will be used as the default path.
+
+RTF_OUTPUT = rtf
+
+# If the COMPACT_RTF tag is set to YES Doxygen generates more compact
+# RTF documents. This may be useful for small projects and may help to
+# save some trees in general.
+
+COMPACT_RTF = NO
+
+# If the RTF_HYPERLINKS tag is set to YES, the RTF that is generated
+# will contain hyperlink fields. The RTF file will
+# contain links (just like the HTML output) instead of page references.
+# This makes the output suitable for online browsing using WORD or other
+# programs which support those fields.
+# Note: wordpad (write) and others do not support links.
+
+RTF_HYPERLINKS = NO
+
+# Load stylesheet definitions from file. Syntax is similar to doxygen's
+# config file, i.e. a series of assignments. You only have to provide
+# replacements, missing definitions are set to their default value.
+
+RTF_STYLESHEET_FILE =
+
+# Set optional variables used in the generation of an rtf document.
+# Syntax is similar to doxygen's config file.
+
+RTF_EXTENSIONS_FILE =
+
+#---------------------------------------------------------------------------
+# configuration options related to the man page output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_MAN tag is set to YES (the default) Doxygen will
+# generate man pages
+
+GENERATE_MAN = YES
+
+# The MAN_OUTPUT tag is used to specify where the man pages will be put.
+# If a relative path is entered the value of OUTPUT_DIRECTORY will be
+# put in front of it. If left blank `man' will be used as the default path.
+
+MAN_OUTPUT = man
+
+# The MAN_EXTENSION tag determines the extension that is added to
+# the generated man pages (default is the subroutine's section .3)
+
+MAN_EXTENSION = .3
+
+# If the MAN_LINKS tag is set to YES and Doxygen generates man output,
+# then it will generate one additional man file for each entity
+# documented in the real man page(s). These additional files
+# only source the real man page, but without them the man command
+# would be unable to find the correct page. The default is NO.
+
+MAN_LINKS = NO
+
+#---------------------------------------------------------------------------
+# configuration options related to the XML output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_XML tag is set to YES Doxygen will
+# generate an XML file that captures the structure of
+# the code including all documentation.
+
+GENERATE_XML = NO
+
+# The XML_OUTPUT tag is used to specify where the XML pages will be put.
+# If a relative path is entered the value of OUTPUT_DIRECTORY will be
+# put in front of it. If left blank `xml' will be used as the default path.
+
+XML_OUTPUT = xml
+
+# The XML_SCHEMA tag can be used to specify an XML schema,
+# which can be used by a validating XML parser to check the
+# syntax of the XML files.
+
+XML_SCHEMA =
+
+# The XML_DTD tag can be used to specify an XML DTD,
+# which can be used by a validating XML parser to check the
+# syntax of the XML files.
+
+XML_DTD =
+
+# If the XML_PROGRAMLISTING tag is set to YES Doxygen will
+# dump the program listings (including syntax highlighting
+# and cross-referencing information) to the XML output. Note that
+# enabling this will significantly increase the size of the XML output.
+
+XML_PROGRAMLISTING = YES
+
+#---------------------------------------------------------------------------
+# configuration options for the AutoGen Definitions output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_AUTOGEN_DEF tag is set to YES Doxygen will
+# generate an AutoGen Definitions (see autogen.sf.net) file
+# that captures the structure of the code including all
+# documentation. Note that this feature is still experimental
+# and incomplete at the moment.
+
+GENERATE_AUTOGEN_DEF = NO
+
+#---------------------------------------------------------------------------
+# configuration options related to the Perl module output
+#---------------------------------------------------------------------------
+
+# If the GENERATE_PERLMOD tag is set to YES Doxygen will
+# generate a Perl module file that captures the structure of
+# the code including all documentation. Note that this
+# feature is still experimental and incomplete at the
+# moment.
+
+GENERATE_PERLMOD = NO
+
+# If the PERLMOD_LATEX tag is set to YES Doxygen will generate
+# the necessary Makefile rules, Perl scripts and LaTeX code to be able
+# to generate PDF and DVI output from the Perl module output.
+
+PERLMOD_LATEX = NO
+
+# If the PERLMOD_PRETTY tag is set to YES the Perl module output will be
+# nicely formatted so it can be parsed by a human reader.
+# This is useful
+# if you want to understand what is going on.
+# On the other hand, if this
+# tag is set to NO the size of the Perl module output will be much smaller
+# and Perl will parse it just the same.
+
+PERLMOD_PRETTY = YES
+
+# The names of the make variables in the generated doxyrules.make file
+# are prefixed with the string contained in PERLMOD_MAKEVAR_PREFIX.
+# This is useful so different doxyrules.make files included by the same
+# Makefile don't overwrite each other's variables.
+
+PERLMOD_MAKEVAR_PREFIX =
+
+#---------------------------------------------------------------------------
+# Configuration options related to the preprocessor
+#---------------------------------------------------------------------------
+
+# If the ENABLE_PREPROCESSING tag is set to YES (the default) Doxygen will
+# evaluate all C-preprocessor directives found in the sources and include
+# files.
+
+ENABLE_PREPROCESSING = YES
+
+# If the MACRO_EXPANSION tag is set to YES Doxygen will expand all macro
+# names in the source code. If set to NO (the default) only conditional
+# compilation will be performed. Macro expansion can be done in a controlled
+# way by setting EXPAND_ONLY_PREDEF to YES.
+
+MACRO_EXPANSION = NO
+
+# If the EXPAND_ONLY_PREDEF and MACRO_EXPANSION tags are both set to YES
+# then the macro expansion is limited to the macros specified with the
+# PREDEFINED and EXPAND_AS_DEFINED tags.
+
+EXPAND_ONLY_PREDEF = NO
+
+# If the SEARCH_INCLUDES tag is set to YES (the default) the includes files
+# in the INCLUDE_PATH (see below) will be search if a #include is found.
+
+SEARCH_INCLUDES = YES
+
+# The INCLUDE_PATH tag can be used to specify one or more directories that
+# contain include files that are not input files but should be processed by
+# the preprocessor.
+
+INCLUDE_PATH =
+
+# You can use the INCLUDE_FILE_PATTERNS tag to specify one or more wildcard
+# patterns (like *.h and *.hpp) to filter out the header-files in the
+# directories. If left blank, the patterns specified with FILE_PATTERNS will
+# be used.
+
+INCLUDE_FILE_PATTERNS =
+
+# The PREDEFINED tag can be used to specify one or more macro names that
+# are defined before the preprocessor is started (similar to the -D option of
+# gcc). The argument of the tag is a list of macros of the form: name
+# or name=definition (no spaces). If the definition and the = are
+# omitted =1 is assumed. To prevent a macro definition from being
+# undefined via #undef or recursively expanded use the := operator
+# instead of the = operator.
+
+PREDEFINED = DEBUG=2 __GNUC__=1
+
+# If the MACRO_EXPANSION and EXPAND_ONLY_PREDEF tags are set to YES then
+# this tag can be used to specify a list of macro names that should be expanded.
+# The macro definition that is found in the sources will be used.
+# Use the PREDEFINED tag if you want to use a different macro definition.
+
+EXPAND_AS_DEFINED =
+
+# If the SKIP_FUNCTION_MACROS tag is set to YES (the default) then
+# doxygen's preprocessor will remove all function-like macros that are alone
+# on a line, have an all uppercase name, and do not end with a semicolon. Such
+# function macros are typically used for boiler-plate code, and will confuse
+# the parser if not removed.
+
+SKIP_FUNCTION_MACROS = YES
+
+#---------------------------------------------------------------------------
+# Configuration::additions related to external references
+#---------------------------------------------------------------------------
+
+# The TAGFILES option can be used to specify one or more tagfiles.
+# Optionally an initial location of the external documentation
+# can be added for each tagfile. The format of a tag file without
+# this location is as follows:
+#
+# TAGFILES = file1 file2 ...
+# Adding location for the tag files is done as follows:
+#
+# TAGFILES = file1=loc1 "file2 = loc2" ...
+# where "loc1" and "loc2" can be relative or absolute paths or
+# URLs. If a location is present for each tag, the installdox tool
+# does not have to be run to correct the links.
+# Note that each tag file must have a unique name
+# (where the name does NOT include the path)
+# If a tag file is not located in the directory in which doxygen
+# is run, you must also specify the path to the tagfile here.
+
+TAGFILES =
+
+# When a file name is specified after GENERATE_TAGFILE, doxygen will create
+# a tag file that is based on the input files it reads.
+
+GENERATE_TAGFILE =
+
+# If the ALLEXTERNALS tag is set to YES all external classes will be listed
+# in the class index. If set to NO only the inherited external classes
+# will be listed.
+
+ALLEXTERNALS = NO
+
+# If the EXTERNAL_GROUPS tag is set to YES all external groups will be listed
+# in the modules index. If set to NO, only the current project's groups will
+# be listed.
+
+EXTERNAL_GROUPS = YES
+
+# The PERL_PATH should be the absolute path and name of the perl script
+# interpreter (i.e. the result of `which perl').
+
+PERL_PATH = /usr/bin/perl
+
+#---------------------------------------------------------------------------
+# Configuration options related to the dot tool
+#---------------------------------------------------------------------------
+
+# If the CLASS_DIAGRAMS tag is set to YES (the default) Doxygen will
+# generate a inheritance diagram (in HTML, RTF and LaTeX) for classes with base
+# or super classes. Setting the tag to NO turns the diagrams off. Note that
+# this option is superseded by the HAVE_DOT option below. This is only a
+# fallback. It is recommended to install and use dot, since it yields more
+# powerful graphs.
+
+CLASS_DIAGRAMS = YES
+
+# You can define message sequence charts within doxygen comments using the \msc
+# command. Doxygen will then run the mscgen tool (see
+# http://www.mcternan.me.uk/mscgen/) to produce the chart and insert it in the
+# documentation. The MSCGEN_PATH tag allows you to specify the directory where
+# the mscgen tool resides. If left empty the tool is assumed to be found in the
+# default search path.
+
+MSCGEN_PATH =
+
+# If set to YES, the inheritance and collaboration graphs will hide
+# inheritance and usage relations if the target is undocumented
+# or is not a class.
+
+HIDE_UNDOC_RELATIONS = YES
+
+# If you set the HAVE_DOT tag to YES then doxygen will assume the dot tool is
+# available from the path. This tool is part of Graphviz, a graph visualization
+# toolkit from AT&T and Lucent Bell Labs. The other options in this section
+# have no effect if this option is set to NO (the default)
+
+HAVE_DOT = NO
+
+# The DOT_NUM_THREADS specifies the number of dot invocations doxygen is
+# allowed to run in parallel. When set to 0 (the default) doxygen will
+# base this on the number of processors available in the system. You can set it
+# explicitly to a value larger than 0 to get control over the balance
+# between CPU load and processing speed.
+
+DOT_NUM_THREADS = 0
+
+# By default doxygen will write a font called FreeSans.ttf to the output
+# directory and reference it in all dot files that doxygen generates. This
+# font does not include all possible unicode characters however, so when you need
+# these (or just want a differently looking font) you can specify the font name
+# using DOT_FONTNAME. You need need to make sure dot is able to find the font,
+# which can be done by putting it in a standard location or by setting the
+# DOTFONTPATH environment variable or by setting DOT_FONTPATH to the directory
+# containing the font.
+
+DOT_FONTNAME = FreeSans.ttf
+
+# The DOT_FONTSIZE tag can be used to set the size of the font of dot graphs.
+# The default size is 10pt.
+
+DOT_FONTSIZE = 10
+
+# By default doxygen will tell dot to use the output directory to look for the
+# FreeSans.ttf font (which doxygen will put there itself). If you specify a
+# different font using DOT_FONTNAME you can set the path where dot
+# can find it using this tag.
+
+DOT_FONTPATH =
+
+# If the CLASS_GRAPH and HAVE_DOT tags are set to YES then doxygen
+# will generate a graph for each documented class showing the direct and
+# indirect inheritance relations. Setting this tag to YES will force the
+# the CLASS_DIAGRAMS tag to NO.
+
+CLASS_GRAPH = YES
+
+# If the COLLABORATION_GRAPH and HAVE_DOT tags are set to YES then doxygen
+# will generate a graph for each documented class showing the direct and
+# indirect implementation dependencies (inheritance, containment, and
+# class references variables) of the class with other documented classes.
+
+COLLABORATION_GRAPH = YES
+
+# If the GROUP_GRAPHS and HAVE_DOT tags are set to YES then doxygen
+# will generate a graph for groups, showing the direct groups dependencies
+
+GROUP_GRAPHS = YES
+
+# If the UML_LOOK tag is set to YES doxygen will generate inheritance and
+# collaboration diagrams in a style similar to the OMG's Unified Modeling
+# Language.
+
+UML_LOOK = NO
+
+# If set to YES, the inheritance and collaboration graphs will show the
+# relations between templates and their instances.
+
+TEMPLATE_RELATIONS = NO
+
+# If the ENABLE_PREPROCESSING, SEARCH_INCLUDES, INCLUDE_GRAPH, and HAVE_DOT
+# tags are set to YES then doxygen will generate a graph for each documented
+# file showing the direct and indirect include dependencies of the file with
+# other documented files.
+
+INCLUDE_GRAPH = YES
+
+# If the ENABLE_PREPROCESSING, SEARCH_INCLUDES, INCLUDED_BY_GRAPH, and
+# HAVE_DOT tags are set to YES then doxygen will generate a graph for each
+# documented header file showing the documented files that directly or
+# indirectly include this file.
+
+INCLUDED_BY_GRAPH = YES
+
+# If the CALL_GRAPH and HAVE_DOT options are set to YES then
+# doxygen will generate a call dependency graph for every global function
+# or class method. Note that enabling this option will significantly increase
+# the time of a run. So in most cases it will be better to enable call graphs
+# for selected functions only using the \callgraph command.
+
+CALL_GRAPH = NO
+
+# If the CALLER_GRAPH and HAVE_DOT tags are set to YES then
+# doxygen will generate a caller dependency graph for every global function
+# or class method. Note that enabling this option will significantly increase
+# the time of a run. So in most cases it will be better to enable caller
+# graphs for selected functions only using the \callergraph command.
+
+CALLER_GRAPH = NO
+
+# If the GRAPHICAL_HIERARCHY and HAVE_DOT tags are set to YES then doxygen
+# will graphical hierarchy of all classes instead of a textual one.
+
+GRAPHICAL_HIERARCHY = YES
+
+# If the DIRECTORY_GRAPH, SHOW_DIRECTORIES and HAVE_DOT tags are set to YES
+# then doxygen will show the dependencies a directory has on other directories
+# in a graphical way. The dependency relations are determined by the #include
+# relations between the files in the directories.
+
+DIRECTORY_GRAPH = YES
+
+# The DOT_IMAGE_FORMAT tag can be used to set the image format of the images
+# generated by dot. Possible values are png, jpg, or gif
+# If left blank png will be used.
+
+DOT_IMAGE_FORMAT = png
+
+# The tag DOT_PATH can be used to specify the path where the dot tool can be
+# found. If left blank, it is assumed the dot tool can be found in the path.
+
+DOT_PATH =
+
+# The DOTFILE_DIRS tag can be used to specify one or more directories that
+# contain dot files that are included in the documentation (see the
+# \dotfile command).
+
+DOTFILE_DIRS =
+
+# The DOT_GRAPH_MAX_NODES tag can be used to set the maximum number of
+# nodes that will be shown in the graph. If the number of nodes in a graph
+# becomes larger than this value, doxygen will truncate the graph, which is
+# visualized by representing a node as a red box. Note that doxygen if the
+# number of direct children of the root node in a graph is already larger than
+# DOT_GRAPH_MAX_NODES then the graph will not be shown at all. Also note
+# that the size of a graph can be further restricted by MAX_DOT_GRAPH_DEPTH.
+
+DOT_GRAPH_MAX_NODES = 50
+
+# The MAX_DOT_GRAPH_DEPTH tag can be used to set the maximum depth of the
+# graphs generated by dot. A depth value of 3 means that only nodes reachable
+# from the root by following a path via at most 3 edges will be shown. Nodes
+# that lay further from the root node will be omitted. Note that setting this
+# option to 1 or 2 may greatly reduce the computation time needed for large
+# code bases. Also note that the size of a graph can be further restricted by
+# DOT_GRAPH_MAX_NODES. Using a depth of 0 means no depth restriction.
+
+MAX_DOT_GRAPH_DEPTH = 0
+
+# Set the DOT_TRANSPARENT tag to YES to generate images with a transparent
+# background. This is disabled by default, because dot on Windows does not
+# seem to support this out of the box. Warning: Depending on the platform used,
+# enabling this option may lead to badly anti-aliased labels on the edges of
+# a graph (i.e. they become hard to read).
+
+DOT_TRANSPARENT = NO
+
+# Set the DOT_MULTI_TARGETS tag to YES allow dot to generate multiple output
+# files in one run (i.e. multiple -o and -T options on the command line). This
+# makes dot run faster, but since only newer versions of dot (>1.8.10)
+# support this, this feature is disabled by default.
+
+DOT_MULTI_TARGETS = YES
+
+# If the GENERATE_LEGEND tag is set to YES (the default) Doxygen will
+# generate a legend page explaining the meaning of the various boxes and
+# arrows in the dot generated graphs.
+
+GENERATE_LEGEND = YES
+
+# If the DOT_CLEANUP tag is set to YES (the default) Doxygen will
+# remove the intermediate dot files that are used to generate
+# the various graphs.
+
+DOT_CLEANUP = YES
diff --git a/libraries/libmdb/LICENSE b/libraries/libmdb/LICENSE
new file mode 100644
index 0000000000000000000000000000000000000000..05ad7571e448b9d83ead5d4691274d9484574714
--- /dev/null
+++ b/libraries/libmdb/LICENSE
@@ -0,0 +1,47 @@
+The OpenLDAP Public License
+ Version 2.8, 17 August 2003
+
+Redistribution and use of this software and associated documentation
+("Software"), with or without modification, are permitted provided
+that the following conditions are met:
+
+1. Redistributions in source form must retain copyright statements
+ and notices,
+
+2. Redistributions in binary form must reproduce applicable copyright
+ statements and notices, this list of conditions, and the following
+ disclaimer in the documentation and/or other materials provided
+ with the distribution, and
+
+3. Redistributions must contain a verbatim copy of this document.
+
+The OpenLDAP Foundation may revise this license from time to time.
+Each revision is distinguished by a version number. You may use
+this Software under terms of this license revision or under the
+terms of any subsequent revision of the license.
+
+THIS SOFTWARE IS PROVIDED BY THE OPENLDAP FOUNDATION AND ITS
+CONTRIBUTORS ``AS IS'' AND ANY EXPRESSED 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 OPENLDAP FOUNDATION, ITS CONTRIBUTORS, OR THE AUTHOR(S)
+OR OWNER(S) OF THE SOFTWARE 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.
+
+The names of the authors and copyright holders must not be used in
+advertising or otherwise to promote the sale, use or other dealing
+in this Software without specific, written prior permission. Title
+to copyright in this Software shall at all times remain with copyright
+holders.
+
+OpenLDAP is a registered trademark of the OpenLDAP Foundation.
+
+Copyright 1999-2003 The OpenLDAP Foundation, Redwood City,
+California, USA. All Rights Reserved. Permission to copy and
+distribute verbatim copies of this document is granted.
diff --git a/libraries/libmdb/Makefile b/libraries/libmdb/Makefile
new file mode 100644
index 0000000000000000000000000000000000000000..e49e275638aabf2e14770057ca480092b569f814
--- /dev/null
+++ b/libraries/libmdb/Makefile
@@ -0,0 +1,41 @@
+CC = gcc
+W = -W -Wall -Wno-unused-parameter -Wbad-function-cast
+OPT = -O2 -g
+CFLAGS = -pthread $(OPT) $(W) $(XCFLAGS)
+LDLIBS =
+
+PROGS = mdb_stat mtest mtest2 mtest3 mtest4 mtest5
+all: libmdb.a libmdb.so $(PROGS)
+
+clean:
+ rm -rf $(PROGS) *.[ao] *.so *~ testdb
+
+test: all
+ mkdir testdb
+ ./mtest && ./mdb_stat testdb
+
+libmdb.a: mdb.o midl.o
+ ar rs $@ mdb.o midl.o
+
+libmdb.so: mdb.o midl.o
+ gcc -shared -o $@ mdb.o midl.o
+
+mdb_stat: mdb_stat.o libmdb.a
+mtest: mtest.o libmdb.a
+mtest2: mtest2.o libmdb.a
+mtest3: mtest3.o libmdb.a
+mtest4: mtest4.o libmdb.a
+mtest5: mtest5.o libmdb.a
+mtest6: mtest6.o libmdb.a
+
+mdb.o: mdb.c mdb.h midl.h
+ $(CC) $(CFLAGS) -fPIC $(CPPFLAGS) -c mdb.c
+
+midl.o: midl.c midl.h
+ $(CC) $(CFLAGS) -fPIC $(CPPFLAGS) -c midl.c
+
+%: %.o
+ $(CC) $(CFLAGS) $(LDFLAGS) $^ $(LDLIBS) -o $@
+
+%.o: %.c mdb.h
+ $(CC) $(CFLAGS) $(CPPFLAGS) -c $<
diff --git a/libraries/libmdb/make.sh b/libraries/libmdb/make.sh
new file mode 100755
index 0000000000000000000000000000000000000000..9b5aff213691844bf10a258f74fea80a4581838d
--- /dev/null
+++ b/libraries/libmdb/make.sh
@@ -0,0 +1,9 @@
+#!/bin/sh
+warning_filter() {
+ egrep -v ': In function .*:$|: warning: (ISO C99 requires rest arguments to be used|format .*%p.* has type .*struct MDB_.*\*)'
+}
+
+exit `{
+ { XCFLAGS="-std=c99 -pedantic" make "$@" 2>&1; echo $? >&3; } |
+ warning_filter >&2
+} 3>&1`
diff --git a/libraries/libmdb/mdb.c b/libraries/libmdb/mdb.c
new file mode 100644
index 0000000000000000000000000000000000000000..32f50fec1e7ce6225f86c737200cba86ea05d09e
--- /dev/null
+++ b/libraries/libmdb/mdb.c
@@ -0,0 +1,5970 @@
+/** @file mdb.c
+ * @brief memory-mapped database library
+ *
+ * A Btree-based database management library modeled loosely on the
+ * BerkeleyDB API, but much simplified.
+ */
+/*
+ * Copyright 2011 Howard Chu, Symas Corp.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted only as authorized by the OpenLDAP
+ * Public License.
+ *
+ * A copy of this license is available in the file LICENSE in the
+ * top-level directory of the distribution or, alternatively, at
+ * <http://www.OpenLDAP.org/license.html>.
+ *
+ * This code is derived from btree.c written by Martin Hedenfalk.
+ *
+ * Copyright (c) 2009, 2010 Martin Hedenfalk <martin@bzero.se>
+ *
+ * Permission to use, copy, modify, and distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/param.h>
+#ifdef _WIN32
+#include <windows.h>
+#else
+#include <sys/uio.h>
+#include <sys/mman.h>
+#ifdef HAVE_SYS_FILE_H
+#include <sys/file.h>
+#endif
+#include <fcntl.h>
+#endif
+
+#include <assert.h>
+#include <errno.h>
+#include <limits.h>
+#include <stddef.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+#include <unistd.h>
+
+#ifndef _WIN32
+#include <pthread.h>
+#ifdef __APPLE__
+#include <semaphore.h>
+#endif
+#endif
+
+#ifndef BYTE_ORDER
+#define BYTE_ORDER __BYTE_ORDER
+#endif
+#ifndef LITTLE_ENDIAN
+#define LITTLE_ENDIAN __LITTLE_ENDIAN
+#endif
+#ifndef BIG_ENDIAN
+#define BIG_ENDIAN __BIG_ENDIAN
+#endif
+
+#if defined(__i386) || defined(__x86_64)
+#define MISALIGNED_OK 1
+#endif
+
+#include "mdb.h"
+#include "midl.h"
+
+#if (BYTE_ORDER == LITTLE_ENDIAN) == (BYTE_ORDER == BIG_ENDIAN)
+# error "Unknown or unsupported endianness (BYTE_ORDER)"
+#elif (-6 & 5) || CHAR_BIT != 8 || UINT_MAX < 0xffffffff || ULONG_MAX % 0xFFFF
+# error "Two's complement, reasonably sized integer types, please"
+#endif
+
+/** @defgroup internal MDB Internals
+ * @{
+ */
+/** @defgroup compat Windows Compatibility Macros
+ * A bunch of macros to minimize the amount of platform-specific ifdefs
+ * needed throughout the rest of the code. When the features this library
+ * needs are similar enough to POSIX to be hidden in a one-or-two line
+ * replacement, this macro approach is used.
+ * @{
+ */
+#ifdef _WIN32
+#define pthread_t DWORD
+#define pthread_mutex_t HANDLE
+#define pthread_key_t DWORD
+#define pthread_self() GetCurrentThreadId()
+#define pthread_key_create(x,y) (*(x) = TlsAlloc())
+#define pthread_key_delete(x) TlsFree(x)
+#define pthread_getspecific(x) TlsGetValue(x)
+#define pthread_setspecific(x,y) TlsSetValue(x,y)
+#define pthread_mutex_unlock(x) ReleaseMutex(x)
+#define pthread_mutex_lock(x) WaitForSingleObject(x, INFINITE)
+#define LOCK_MUTEX_R(env) pthread_mutex_lock((env)->me_rmutex)
+#define UNLOCK_MUTEX_R(env) pthread_mutex_unlock((env)->me_rmutex)
+#define LOCK_MUTEX_W(env) pthread_mutex_lock((env)->me_wmutex)
+#define UNLOCK_MUTEX_W(env) pthread_mutex_unlock((env)->me_wmutex)
+#define getpid() GetCurrentProcessId()
+#define fdatasync(fd) (!FlushFileBuffers(fd))
+#define ErrCode() GetLastError()
+#define GET_PAGESIZE(x) {SYSTEM_INFO si; GetSystemInfo(&si); (x) = si.dwPageSize;}
+#define close(fd) CloseHandle(fd)
+#define munmap(ptr,len) UnmapViewOfFile(ptr)
+#else
+#ifdef __APPLE__
+#define LOCK_MUTEX_R(env) sem_wait((env)->me_rmutex)
+#define UNLOCK_MUTEX_R(env) sem_post((env)->me_rmutex)
+#define LOCK_MUTEX_W(env) sem_wait((env)->me_wmutex)
+#define UNLOCK_MUTEX_W(env) sem_post((env)->me_wmutex)
+#define fdatasync(fd) fsync(fd)
+#else
+#ifdef ANDROID
+#define fdatasync(fd) fsync(fd)
+#endif
+ /** Lock the reader mutex.
+ */
+#define LOCK_MUTEX_R(env) pthread_mutex_lock(&(env)->me_txns->mti_mutex)
+ /** Unlock the reader mutex.
+ */
+#define UNLOCK_MUTEX_R(env) pthread_mutex_unlock(&(env)->me_txns->mti_mutex)
+
+ /** Lock the writer mutex.
+ * Only a single write transaction is allowed at a time. Other writers
+ * will block waiting for this mutex.
+ */
+#define LOCK_MUTEX_W(env) pthread_mutex_lock(&(env)->me_txns->mti_wmutex)
+ /** Unlock the writer mutex.
+ */
+#define UNLOCK_MUTEX_W(env) pthread_mutex_unlock(&(env)->me_txns->mti_wmutex)
+#endif /* __APPLE__ */
+
+ /** Get the error code for the last failed system function.
+ */
+#define ErrCode() errno
+
+ /** An abstraction for a file handle.
+ * On POSIX systems file handles are small integers. On Windows
+ * they're opaque pointers.
+ */
+#define HANDLE int
+
+ /** A value for an invalid file handle.
+ * Mainly used to initialize file variables and signify that they are
+ * unused.
+ */
+#define INVALID_HANDLE_VALUE (-1)
+
+ /** Get the size of a memory page for the system.
+ * This is the basic size that the platform's memory manager uses, and is
+ * fundamental to the use of memory-mapped files.
+ */
+#define GET_PAGESIZE(x) ((x) = sysconf(_SC_PAGE_SIZE))
+#endif
+
+#if defined(_WIN32) || defined(__APPLE__)
+#define MNAME_LEN 32
+#endif
+
+/** @} */
+
+#ifndef _WIN32
+/** A flag for opening a file and requesting synchronous data writes.
+ * This is only used when writing a meta page. It's not strictly needed;
+ * we could just do a normal write and then immediately perform a flush.
+ * But if this flag is available it saves us an extra system call.
+ *
+ * @note If O_DSYNC is undefined but exists in /usr/include,
+ * preferably set some compiler flag to get the definition.
+ * Otherwise compile with the less efficient -DMDB_DSYNC=O_SYNC.
+ */
+#ifndef MDB_DSYNC
+# define MDB_DSYNC O_DSYNC
+#endif
+#endif
+
+ /** A page number in the database.
+ * Note that 64 bit page numbers are overkill, since pages themselves
+ * already represent 12-13 bits of addressable memory, and the OS will
+ * always limit applications to a maximum of 63 bits of address space.
+ *
+ * @note In the #MDB_node structure, we only store 48 bits of this value,
+ * which thus limits us to only 60 bits of addressable data.
+ */
+typedef ID pgno_t;
+
+ /** A transaction ID.
+ * See struct MDB_txn.mt_txnid for details.
+ */
+typedef ID txnid_t;
+
+/** @defgroup debug Debug Macros
+ * @{
+ */
+#ifndef DEBUG
+ /** Enable debug output.
+ * Set this to 1 for copious tracing. Set to 2 to add dumps of all IDLs
+ * read from and written to the database (used for free space management).
+ */
+#define DEBUG 0
+#endif
+
+#if !(__STDC_VERSION__ >= 199901L || defined(__GNUC__))
+# define DPRINTF (void) /* Vararg macros may be unsupported */
+#elif DEBUG
+ /** Print a debug message with printf formatting. */
+# define DPRINTF(fmt, ...) /**< Requires 2 or more args */ \
+ fprintf(stderr, "%s:%d " fmt "\n", __func__, __LINE__, __VA_ARGS__)
+#else
+# define DPRINTF(fmt, ...) ((void) 0)
+#endif
+ /** Print a debug string.
+ * The string is printed literally, with no format processing.
+ */
+#define DPUTS(arg) DPRINTF("%s", arg)
+/** @} */
+
+ /** A default memory page size.
+ * The actual size is platform-dependent, but we use this for
+ * boot-strapping. We probably should not be using this any more.
+ * The #GET_PAGESIZE() macro is used to get the actual size.
+ *
+ * Note that we don't currently support Huge pages. On Linux,
+ * regular data files cannot use Huge pages, and in general
+ * Huge pages aren't actually pageable. We rely on the OS
+ * demand-pager to read our data and page it out when memory
+ * pressure from other processes is high. So until OSs have
+ * actual paging support for Huge pages, they're not viable.
+ */
+#define PAGESIZE 4096
+
+ /** The minimum number of keys required in a database page.
+ * Setting this to a larger value will place a smaller bound on the
+ * maximum size of a data item. Data items larger than this size will
+ * be pushed into overflow pages instead of being stored directly in
+ * the B-tree node. This value used to default to 4. With a page size
+ * of 4096 bytes that meant that any item larger than 1024 bytes would
+ * go into an overflow page. That also meant that on average 2-3KB of
+ * each overflow page was wasted space. The value cannot be lower than
+ * 2 because then there would no longer be a tree structure. With this
+ * value, items larger than 2KB will go into overflow pages, and on
+ * average only 1KB will be wasted.
+ */
+#define MDB_MINKEYS 2
+
+ /** A stamp that identifies a file as an MDB file.
+ * There's nothing special about this value other than that it is easily
+ * recognizable, and it will reflect any byte order mismatches.
+ */
+#define MDB_MAGIC 0xBEEFC0DE
+
+ /** The version number for a database's file format. */
+#define MDB_VERSION 1
+
+ /** The maximum size of a key in the database.
+ * While data items have essentially unbounded size, we require that
+ * keys all fit onto a regular page. This limit could be raised a bit
+ * further if needed; to something just under #PAGESIZE / #MDB_MINKEYS.
+ */
+#define MAXKEYSIZE 511
+
+#if DEBUG
+ /** A key buffer.
+ * @ingroup debug
+ * This is used for printing a hex dump of a key's contents.
+ */
+#define DKBUF char kbuf[(MAXKEYSIZE*2+1)]
+ /** Display a key in hex.
+ * @ingroup debug
+ * Invoke a function to display a key in hex.
+ */
+#define DKEY(x) mdb_dkey(x, kbuf)
+#else
+#define DKBUF typedef int dummy_kbuf /* so we can put ';' after */
+#define DKEY(x)
+#endif
+
+/** @defgroup lazylock Lazy Locking
+ * Macros for locks that are't actually needed.
+ * The DB view is always consistent because all writes are wrapped in
+ * the wmutex. Finer-grained locks aren't necessary.
+ * @{
+ */
+#ifndef LAZY_LOCKS
+ /** Use lazy locking. I.e., don't lock these accesses at all. */
+#define LAZY_LOCKS 1
+#endif
+#if LAZY_LOCKS
+ /** Grab the reader lock */
+#define LAZY_MUTEX_LOCK(x)
+ /** Release the reader lock */
+#define LAZY_MUTEX_UNLOCK(x)
+ /** Release the DB table reader/writer lock */
+#define LAZY_RWLOCK_UNLOCK(x)
+ /** Grab the DB table write lock */
+#define LAZY_RWLOCK_WRLOCK(x)
+ /** Grab the DB table read lock */
+#define LAZY_RWLOCK_RDLOCK(x)
+ /** Declare the DB table rwlock. Should not be followed by ';'. */
+#define LAZY_RWLOCK_DEF(x)
+ /** Initialize the DB table rwlock */
+#define LAZY_RWLOCK_INIT(x,y)
+ /** Destroy the DB table rwlock */
+#define LAZY_RWLOCK_DESTROY(x)
+#else
+#define LAZY_MUTEX_LOCK(x) pthread_mutex_lock(x)
+#define LAZY_MUTEX_UNLOCK(x) pthread_mutex_unlock(x)
+#define LAZY_RWLOCK_UNLOCK(x) pthread_rwlock_unlock(x)
+#define LAZY_RWLOCK_WRLOCK(x) pthread_rwlock_wrlock(x)
+#define LAZY_RWLOCK_RDLOCK(x) pthread_rwlock_rdlock(x)
+#define LAZY_RWLOCK_DEF(x) pthread_rwlock_t x;
+#define LAZY_RWLOCK_INIT(x,y) pthread_rwlock_init(x,y)
+#define LAZY_RWLOCK_DESTROY(x) pthread_rwlock_destroy(x)
+#endif
+/** @} */
+
+ /** An invalid page number.
+ * Mainly used to denote an empty tree.
+ */
+#define P_INVALID (~0UL)
+
+ /** Test if a flag \b f is set in a flag word \b w. */
+#define F_ISSET(w, f) (((w) & (f)) == (f))
+
+ /** Used for offsets within a single page.
+ * Since memory pages are typically 4 or 8KB in size, 12-13 bits,
+ * this is plenty.
+ */
+typedef uint16_t indx_t;
+
+ /** Default size of memory map.
+ * This is certainly too small for any actual applications. Apps should always set
+ * the size explicitly using #mdb_env_set_mapsize().
+ */
+#define DEFAULT_MAPSIZE 1048576
+
+/** @defgroup readers Reader Lock Table
+ * Readers don't acquire any locks for their data access. Instead, they
+ * simply record their transaction ID in the reader table. The reader
+ * mutex is needed just to find an empty slot in the reader table. The
+ * slot's address is saved in thread-specific data so that subsequent read
+ * transactions started by the same thread need no further locking to proceed.
+ *
+ * Since the database uses multi-version concurrency control, readers don't
+ * actually need any locking. This table is used to keep track of which
+ * readers are using data from which old transactions, so that we'll know
+ * when a particular old transaction is no longer in use. Old transactions
+ * that have discarded any data pages can then have those pages reclaimed
+ * for use by a later write transaction.
+ *
+ * The lock table is constructed such that reader slots are aligned with the
+ * processor's cache line size. Any slot is only ever used by one thread.
+ * This alignment guarantees that there will be no contention or cache
+ * thrashing as threads update their own slot info, and also eliminates
+ * any need for locking when accessing a slot.
+ *
+ * A writer thread will scan every slot in the table to determine the oldest
+ * outstanding reader transaction. Any freed pages older than this will be
+ * reclaimed by the writer. The writer doesn't use any locks when scanning
+ * this table. This means that there's no guarantee that the writer will
+ * see the most up-to-date reader info, but that's not required for correct
+ * operation - all we need is to know the upper bound on the oldest reader,
+ * we don't care at all about the newest reader. So the only consequence of
+ * reading stale information here is that old pages might hang around a
+ * while longer before being reclaimed. That's actually good anyway, because
+ * the longer we delay reclaiming old pages, the more likely it is that a
+ * string of contiguous pages can be found after coalescing old pages from
+ * many old transactions together.
+ *
+ * @todo We don't actually do such coalescing yet, we grab pages from one
+ * old transaction at a time.
+ * @{
+ */
+ /** Number of slots in the reader table.
+ * This value was chosen somewhat arbitrarily. 126 readers plus a
+ * couple mutexes fit exactly into 8KB on my development machine.
+ * Applications should set the table size using #mdb_env_set_maxreaders().
+ */
+#define DEFAULT_READERS 126
+
+ /** The size of a CPU cache line in bytes. We want our lock structures
+ * aligned to this size to avoid false cache line sharing in the
+ * lock table.
+ * This value works for most CPUs. For Itanium this should be 128.
+ */
+#ifndef CACHELINE
+#define CACHELINE 64
+#endif
+
+ /** The information we store in a single slot of the reader table.
+ * In addition to a transaction ID, we also record the process and
+ * thread ID that owns a slot, so that we can detect stale information,
+ * e.g. threads or processes that went away without cleaning up.
+ * @note We currently don't check for stale records. We simply re-init
+ * the table when we know that we're the only process opening the
+ * lock file.
+ */
+typedef struct MDB_rxbody {
+ /** The current Transaction ID when this transaction began.
+ * Multiple readers that start at the same time will probably have the
+ * same ID here. Again, it's not important to exclude them from
+ * anything; all we need to know is which version of the DB they
+ * started from so we can avoid overwriting any data used in that
+ * particular version.
+ */
+ txnid_t mrb_txnid;
+ /** The process ID of the process owning this reader txn. */
+ pid_t mrb_pid;
+ /** The thread ID of the thread owning this txn. */
+ pthread_t mrb_tid;
+} MDB_rxbody;
+
+ /** The actual reader record, with cacheline padding. */
+typedef struct MDB_reader {
+ union {
+ MDB_rxbody mrx;
+ /** shorthand for mrb_txnid */
+#define mr_txnid mru.mrx.mrb_txnid
+#define mr_pid mru.mrx.mrb_pid
+#define mr_tid mru.mrx.mrb_tid
+ /** cache line alignment */
+ char pad[(sizeof(MDB_rxbody)+CACHELINE-1) & ~(CACHELINE-1)];
+ } mru;
+} MDB_reader;
+
+ /** The header for the reader table.
+ * The table resides in a memory-mapped file. (This is a different file
+ * than is used for the main database.)
+ *
+ * For POSIX the actual mutexes reside in the shared memory of this
+ * mapped file. On Windows, mutexes are named objects allocated by the
+ * kernel; we store the mutex names in this mapped file so that other
+ * processes can grab them. This same approach is also used on
+ * MacOSX/Darwin (using named semaphores) since MacOSX doesn't support
+ * process-shared POSIX mutexes. For these cases where a named object
+ * is used, the object name is derived from a 64 bit FNV hash of the
+ * environment pathname. As such, naming collisions are extremely
+ * unlikely. If a collision occurs, the results are unpredictable.
+ */
+typedef struct MDB_txbody {
+ /** Stamp identifying this as an MDB lock file. It must be set
+ * to #MDB_MAGIC. */
+ uint32_t mtb_magic;
+ /** Version number of this lock file. Must be set to #MDB_VERSION. */
+ uint32_t mtb_version;
+#if defined(_WIN32) || defined(__APPLE__)
+ char mtb_rmname[MNAME_LEN];
+#else
+ /** Mutex protecting access to this table.
+ * This is the reader lock that #LOCK_MUTEX_R acquires.
+ */
+ pthread_mutex_t mtb_mutex;
+#endif
+ /** The ID of the last transaction committed to the database.
+ * This is recorded here only for convenience; the value can always
+ * be determined by reading the main database meta pages.
+ */
+ txnid_t mtb_txnid;
+ /** The number of slots that have been used in the reader table.
+ * This always records the maximum count, it is not decremented
+ * when readers release their slots.
+ */
+ unsigned mtb_numreaders;
+ /** The ID of the most recent meta page in the database.
+ * This is recorded here only for convenience; the value can always
+ * be determined by reading the main database meta pages.
+ */
+ uint32_t mtb_me_toggle;
+} MDB_txbody;
+
+ /** The actual reader table definition. */
+typedef struct MDB_txninfo {
+ union {
+ MDB_txbody mtb;
+#define mti_magic mt1.mtb.mtb_magic
+#define mti_version mt1.mtb.mtb_version
+#define mti_mutex mt1.mtb.mtb_mutex
+#define mti_rmname mt1.mtb.mtb_rmname
+#define mti_txnid mt1.mtb.mtb_txnid
+#define mti_numreaders mt1.mtb.mtb_numreaders
+#define mti_me_toggle mt1.mtb.mtb_me_toggle
+ char pad[(sizeof(MDB_txbody)+CACHELINE-1) & ~(CACHELINE-1)];
+ } mt1;
+ union {
+#if defined(_WIN32) || defined(__APPLE__)
+ char mt2_wmname[MNAME_LEN];
+#define mti_wmname mt2.mt2_wmname
+#else
+ pthread_mutex_t mt2_wmutex;
+#define mti_wmutex mt2.mt2_wmutex
+#endif
+ char pad[(sizeof(pthread_mutex_t)+CACHELINE-1) & ~(CACHELINE-1)];
+ } mt2;
+ MDB_reader mti_readers[1];
+} MDB_txninfo;
+/** @} */
+
+/** Common header for all page types.
+ * Overflow records occupy a number of contiguous pages with no
+ * headers on any page after the first.
+ */
+typedef struct MDB_page {
+#define mp_pgno mp_p.p_pgno
+#define mp_next mp_p.p_next
+ union {
+ pgno_t p_pgno; /**< page number */
+ void * p_next; /**< for in-memory list of freed structs */
+ } mp_p;
+ uint16_t mp_pad;
+/** @defgroup mdb_page Page Flags
+ * @ingroup internal
+ * Flags for the page headers.
+ * @{
+ */
+#define P_BRANCH 0x01 /**< branch page */
+#define P_LEAF 0x02 /**< leaf page */
+#define P_OVERFLOW 0x04 /**< overflow page */
+#define P_META 0x08 /**< meta page */
+#define P_DIRTY 0x10 /**< dirty page */
+#define P_LEAF2 0x20 /**< for #MDB_DUPFIXED records */
+#define P_SUBP 0x40 /**< for #MDB_DUPSORT sub-pages */
+/** @} */
+ uint16_t mp_flags; /**< @ref mdb_page */
+#define mp_lower mp_pb.pb.pb_lower
+#define mp_upper mp_pb.pb.pb_upper
+#define mp_pages mp_pb.pb_pages
+ union {
+ struct {
+ indx_t pb_lower; /**< lower bound of free space */
+ indx_t pb_upper; /**< upper bound of free space */
+ } pb;
+ uint32_t pb_pages; /**< number of overflow pages */
+ } mp_pb;
+ indx_t mp_ptrs[1]; /**< dynamic size */
+} MDB_page;
+
+ /** Size of the page header, excluding dynamic data at the end */
+#define PAGEHDRSZ ((unsigned) offsetof(MDB_page, mp_ptrs))
+
+ /** Address of first usable data byte in a page, after the header */
+#define METADATA(p) ((void *)((char *)(p) + PAGEHDRSZ))
+
+ /** Number of nodes on a page */
+#define NUMKEYS(p) (((p)->mp_lower - PAGEHDRSZ) >> 1)
+
+ /** The amount of space remaining in the page */
+#define SIZELEFT(p) (indx_t)((p)->mp_upper - (p)->mp_lower)
+
+ /** The percentage of space used in the page, in tenths of a percent. */
+#define PAGEFILL(env, p) (1000L * ((env)->me_psize - PAGEHDRSZ - SIZELEFT(p)) / \
+ ((env)->me_psize - PAGEHDRSZ))
+ /** The minimum page fill factor, in tenths of a percent.
+ * Pages emptier than this are candidates for merging.
+ */
+#define FILL_THRESHOLD 250
+
+ /** Test if a page is a leaf page */
+#define IS_LEAF(p) F_ISSET((p)->mp_flags, P_LEAF)
+ /** Test if a page is a LEAF2 page */
+#define IS_LEAF2(p) F_ISSET((p)->mp_flags, P_LEAF2)
+ /** Test if a page is a branch page */
+#define IS_BRANCH(p) F_ISSET((p)->mp_flags, P_BRANCH)
+ /** Test if a page is an overflow page */
+#define IS_OVERFLOW(p) F_ISSET((p)->mp_flags, P_OVERFLOW)
+ /** Test if a page is a sub page */
+#define IS_SUBP(p) F_ISSET((p)->mp_flags, P_SUBP)
+
+ /** The number of overflow pages needed to store the given size. */
+#define OVPAGES(size, psize) ((PAGEHDRSZ-1 + (size)) / (psize) + 1)
+
+ /** Header for a single key/data pair within a page.
+ * We guarantee 2-byte alignment for nodes.
+ */
+typedef struct MDB_node {
+ /** lo and hi are used for data size on leaf nodes and for
+ * child pgno on branch nodes. On 64 bit platforms, flags
+ * is also used for pgno. (Branch nodes have no flags).
+ * They are in host byte order in case that lets some
+ * accesses be optimized into a 32-bit word access.
+ */
+#define mn_lo mn_offset[BYTE_ORDER!=LITTLE_ENDIAN]
+#define mn_hi mn_offset[BYTE_ORDER==LITTLE_ENDIAN] /**< part of dsize or pgno */
+ unsigned short mn_offset[2]; /**< storage for #mn_lo and #mn_hi */
+/** @defgroup mdb_node Node Flags
+ * @ingroup internal
+ * Flags for node headers.
+ * @{
+ */
+#define F_BIGDATA 0x01 /**< data put on overflow page */
+#define F_SUBDATA 0x02 /**< data is a sub-database */
+#define F_DUPDATA 0x04 /**< data has duplicates */
+
+/** valid flags for #mdb_node_add() */
+#define NODE_ADD_FLAGS (F_DUPDATA|F_SUBDATA|MDB_RESERVE|MDB_APPEND)
+
+/** @} */
+ unsigned short mn_flags; /**< @ref mdb_node */
+ unsigned short mn_ksize; /**< key size */
+ char mn_data[1]; /**< key and data are appended here */
+} MDB_node;
+
+ /** Size of the node header, excluding dynamic data at the end */
+#define NODESIZE offsetof(MDB_node, mn_data)
+
+ /** Bit position of top word in page number, for shifting mn_flags */
+#define PGNO_TOPWORD ((pgno_t)-1 > 0xffffffffu ? 32 : 0)
+
+ /** Size of a node in a branch page with a given key.
+ * This is just the node header plus the key, there is no data.
+ */
+#define INDXSIZE(k) (NODESIZE + ((k) == NULL ? 0 : (k)->mv_size))
+
+ /** Size of a node in a leaf page with a given key and data.
+ * This is node header plus key plus data size.
+ */
+#define LEAFSIZE(k, d) (NODESIZE + (k)->mv_size + (d)->mv_size)
+
+ /** Address of node \b i in page \b p */
+#define NODEPTR(p, i) ((MDB_node *)((char *)(p) + (p)->mp_ptrs[i]))
+
+ /** Address of the key for the node */
+#define NODEKEY(node) (void *)((node)->mn_data)
+
+ /** Address of the data for a node */
+#define NODEDATA(node) (void *)((char *)(node)->mn_data + (node)->mn_ksize)
+
+ /** Get the page number pointed to by a branch node */
+#define NODEPGNO(node) \
+ ((node)->mn_lo | ((pgno_t) (node)->mn_hi << 16) | \
+ (PGNO_TOPWORD ? ((pgno_t) (node)->mn_flags << PGNO_TOPWORD) : 0))
+ /** Set the page number in a branch node */
+#define SETPGNO(node,pgno) do { \
+ (node)->mn_lo = (pgno) & 0xffff; (node)->mn_hi = (pgno) >> 16; \
+ if (PGNO_TOPWORD) (node)->mn_flags = (pgno) >> PGNO_TOPWORD; } while(0)
+
+ /** Get the size of the data in a leaf node */
+#define NODEDSZ(node) ((node)->mn_lo | ((unsigned)(node)->mn_hi << 16))
+ /** Set the size of the data for a leaf node */
+#define SETDSZ(node,size) do { \
+ (node)->mn_lo = (size) & 0xffff; (node)->mn_hi = (size) >> 16;} while(0)
+ /** The size of a key in a node */
+#define NODEKSZ(node) ((node)->mn_ksize)
+
+ /** The address of a key in a LEAF2 page.
+ * LEAF2 pages are used for #MDB_DUPFIXED sorted-duplicate sub-DBs.
+ * There are no node headers, keys are stored contiguously.
+ */
+#define LEAF2KEY(p, i, ks) ((char *)(p) + PAGEHDRSZ + ((i)*(ks)))
+
+ /** Set the \b node's key into \b key, if requested. */
+#define MDB_SET_KEY(node, key) { if ((key) != NULL) { \
+ (key)->mv_size = NODEKSZ(node); (key)->mv_data = NODEKEY(node); } }
+
+ /** Information about a single database in the environment. */
+typedef struct MDB_db {
+ uint32_t md_pad; /**< also ksize for LEAF2 pages */
+ uint16_t md_flags; /**< @ref mdb_open */
+ uint16_t md_depth; /**< depth of this tree */
+ pgno_t md_branch_pages; /**< number of internal pages */
+ pgno_t md_leaf_pages; /**< number of leaf pages */
+ pgno_t md_overflow_pages; /**< number of overflow pages */
+ size_t md_entries; /**< number of data items */
+ pgno_t md_root; /**< the root page of this tree */
+} MDB_db;
+
+ /** Handle for the DB used to track free pages. */
+#define FREE_DBI 0
+ /** Handle for the default DB. */
+#define MAIN_DBI 1
+
+ /** Meta page content. */
+typedef struct MDB_meta {
+ /** Stamp identifying this as an MDB data file. It must be set
+ * to #MDB_MAGIC. */
+ uint32_t mm_magic;
+ /** Version number of this lock file. Must be set to #MDB_VERSION. */
+ uint32_t mm_version;
+ void *mm_address; /**< address for fixed mapping */
+ size_t mm_mapsize; /**< size of mmap region */
+ MDB_db mm_dbs[2]; /**< first is free space, 2nd is main db */
+ /** The size of pages used in this DB */
+#define mm_psize mm_dbs[0].md_pad
+ /** Any persistent environment flags. @ref mdb_env */
+#define mm_flags mm_dbs[0].md_flags
+ pgno_t mm_last_pg; /**< last used page in file */
+ txnid_t mm_txnid; /**< txnid that committed this page */
+} MDB_meta;
+
+ /** Auxiliary DB info.
+ * The information here is mostly static/read-only. There is
+ * only a single copy of this record in the environment.
+ */
+typedef struct MDB_dbx {
+ MDB_val md_name; /**< name of the database */
+ MDB_cmp_func *md_cmp; /**< function for comparing keys */
+ MDB_cmp_func *md_dcmp; /**< function for comparing data items */
+ MDB_rel_func *md_rel; /**< user relocate function */
+ void *md_relctx; /**< user-provided context for md_rel */
+} MDB_dbx;
+
+ /** A database transaction.
+ * Every operation requires a transaction handle.
+ */
+struct MDB_txn {
+ MDB_txn *mt_parent; /**< parent of a nested txn */
+ MDB_txn *mt_child; /**< nested txn under this txn */
+ pgno_t mt_next_pgno; /**< next unallocated page */
+ /** The ID of this transaction. IDs are integers incrementing from 1.
+ * Only committed write transactions increment the ID. If a transaction
+ * aborts, the ID may be re-used by the next writer.
+ */
+ txnid_t mt_txnid;
+ MDB_env *mt_env; /**< the DB environment */
+ /** The list of pages that became unused during this transaction.
+ */
+ IDL mt_free_pgs;
+ union {
+ ID2L dirty_list; /**< modified pages */
+ MDB_reader *reader; /**< this thread's slot in the reader table */
+ } mt_u;
+ /** Array of records for each DB known in the environment. */
+ MDB_dbx *mt_dbxs;
+ /** Array of MDB_db records for each known DB */
+ MDB_db *mt_dbs;
+/** @defgroup mt_dbflag Transaction DB Flags
+ * @ingroup internal
+ * @{
+ */
+#define DB_DIRTY 0x01 /**< DB was written in this txn */
+#define DB_STALE 0x02 /**< DB record is older than txnID */
+/** @} */
+ /** Array of cursors for each DB */
+ MDB_cursor **mt_cursors;
+ /** Array of flags for each DB */
+ unsigned char *mt_dbflags;
+ /** Number of DB records in use. This number only ever increments;
+ * we don't decrement it when individual DB handles are closed.
+ */
+ MDB_dbi mt_numdbs;
+
+/** @defgroup mdb_txn Transaction Flags
+ * @ingroup internal
+ * @{
+ */
+#define MDB_TXN_RDONLY 0x01 /**< read-only transaction */
+#define MDB_TXN_ERROR 0x02 /**< an error has occurred */
+/** @} */
+ unsigned int mt_flags; /**< @ref mdb_txn */
+ /** Tracks which of the two meta pages was used at the start
+ * of this transaction.
+ */
+ unsigned int mt_toggle;
+};
+
+/** Enough space for 2^32 nodes with minimum of 2 keys per node. I.e., plenty.
+ * At 4 keys per node, enough for 2^64 nodes, so there's probably no need to
+ * raise this on a 64 bit machine.
+ */
+#define CURSOR_STACK 32
+
+struct MDB_xcursor;
+
+ /** Cursors are used for all DB operations */
+struct MDB_cursor {
+ /** Next cursor on this DB in this txn */
+ MDB_cursor *mc_next;
+ /** Original cursor if this is a shadow */
+ MDB_cursor *mc_orig;
+ /** Context used for databases with #MDB_DUPSORT, otherwise NULL */
+ struct MDB_xcursor *mc_xcursor;
+ /** The transaction that owns this cursor */
+ MDB_txn *mc_txn;
+ /** The database handle this cursor operates on */
+ MDB_dbi mc_dbi;
+ /** The database record for this cursor */
+ MDB_db *mc_db;
+ /** The database auxiliary record for this cursor */
+ MDB_dbx *mc_dbx;
+ /** The @ref mt_dbflag for this database */
+ unsigned char *mc_dbflag;
+ unsigned short mc_snum; /**< number of pushed pages */
+ unsigned short mc_top; /**< index of top page, mc_snum-1 */
+/** @defgroup mdb_cursor Cursor Flags
+ * @ingroup internal
+ * Cursor state flags.
+ * @{
+ */
+#define C_INITIALIZED 0x01 /**< cursor has been initialized and is valid */
+#define C_EOF 0x02 /**< No more data */
+#define C_SUB 0x04 /**< Cursor is a sub-cursor */
+#define C_SHADOW 0x08 /**< Cursor is a dup from a parent txn */
+#define C_ALLOCD 0x10 /**< Cursor was malloc'd */
+/** @} */
+ unsigned int mc_flags; /**< @ref mdb_cursor */
+ MDB_page *mc_pg[CURSOR_STACK]; /**< stack of pushed pages */
+ indx_t mc_ki[CURSOR_STACK]; /**< stack of page indices */
+};
+
+ /** Context for sorted-dup records.
+ * We could have gone to a fully recursive design, with arbitrarily
+ * deep nesting of sub-databases. But for now we only handle these
+ * levels - main DB, optional sub-DB, sorted-duplicate DB.
+ */
+typedef struct MDB_xcursor {
+ /** A sub-cursor for traversing the Dup DB */
+ MDB_cursor mx_cursor;
+ /** The database record for this Dup DB */
+ MDB_db mx_db;
+ /** The auxiliary DB record for this Dup DB */
+ MDB_dbx mx_dbx;
+ /** The @ref mt_dbflag for this Dup DB */
+ unsigned char mx_dbflag;
+} MDB_xcursor;
+
+ /** A set of pages freed by an earlier transaction. */
+typedef struct MDB_oldpages {
+ /** Usually we only read one record from the FREEDB at a time, but
+ * in case we read more, this will chain them together.
+ */
+ struct MDB_oldpages *mo_next;
+ /** The ID of the transaction in which these pages were freed. */
+ txnid_t mo_txnid;
+ /** An #IDL of the pages */
+ pgno_t mo_pages[1]; /* dynamic */
+} MDB_oldpages;
+
+ /** The database environment. */
+struct MDB_env {
+ HANDLE me_fd; /**< The main data file */
+ HANDLE me_lfd; /**< The lock file */
+ HANDLE me_mfd; /**< just for writing the meta pages */
+ /** Failed to update the meta page. Probably an I/O error. */
+#define MDB_FATAL_ERROR 0x80000000U
+ uint32_t me_flags; /**< @ref mdb_env */
+ uint32_t me_extrapad; /**< unused for now */
+ unsigned int me_maxreaders; /**< size of the reader table */
+ MDB_dbi me_numdbs; /**< number of DBs opened */
+ MDB_dbi me_maxdbs; /**< size of the DB table */
+ char *me_path; /**< path to the DB files */
+ char *me_map; /**< the memory map of the data file */
+ MDB_txninfo *me_txns; /**< the memory map of the lock file */
+ MDB_meta *me_metas[2]; /**< pointers to the two meta pages */
+ MDB_txn *me_txn; /**< current write transaction */
+ size_t me_mapsize; /**< size of the data memory map */
+ off_t me_size; /**< current file size */
+ pgno_t me_maxpg; /**< me_mapsize / me_psize */
+ unsigned int me_psize; /**< size of a page, from #GET_PAGESIZE */
+ unsigned int me_db_toggle; /**< which DB table is current */
+ txnid_t me_wtxnid; /**< ID of last txn we committed */
+ MDB_dbx *me_dbxs; /**< array of static DB info */
+ MDB_db *me_dbs[2]; /**< two arrays of MDB_db info */
+ MDB_oldpages *me_pghead; /**< list of old page records */
+ pthread_key_t me_txkey; /**< thread-key for readers */
+ MDB_page *me_dpages; /**< list of malloc'd blocks for re-use */
+ /** IDL of pages that became unused in a write txn */
+ IDL me_free_pgs;
+ /** ID2L of pages that were written during a write txn */
+ ID2 me_dirty_list[MDB_IDL_UM_SIZE];
+ /** rwlock for the DB tables, if #LAZY_LOCKS is false */
+ LAZY_RWLOCK_DEF(me_dblock)
+#ifdef _WIN32
+ HANDLE me_rmutex; /* Windows mutexes don't reside in shared mem */
+ HANDLE me_wmutex;
+#endif
+#ifdef __APPLE__
+ sem_t *me_rmutex; /* Apple doesn't support shared mutexes */
+ sem_t *me_wmutex;
+#endif
+};
+ /** max number of pages to commit in one writev() call */
+#define MDB_COMMIT_PAGES 64
+
+static MDB_page *mdb_page_alloc(MDB_cursor *mc, int num);
+static MDB_page *mdb_page_new(MDB_cursor *mc, uint32_t flags, int num);
+static int mdb_page_touch(MDB_cursor *mc);
+
+static int mdb_page_get(MDB_txn *txn, pgno_t pgno, MDB_page **mp);
+static int mdb_page_search_root(MDB_cursor *mc,
+ MDB_val *key, int modify);
+static int mdb_page_search(MDB_cursor *mc,
+ MDB_val *key, int modify);
+static int mdb_page_merge(MDB_cursor *csrc, MDB_cursor *cdst);
+static int mdb_page_split(MDB_cursor *mc, MDB_val *newkey, MDB_val *newdata,
+ pgno_t newpgno, unsigned int nflags);
+
+static int mdb_env_read_header(MDB_env *env, MDB_meta *meta);
+static int mdb_env_read_meta(MDB_env *env, int *which);
+static int mdb_env_write_meta(MDB_txn *txn);
+
+static MDB_node *mdb_node_search(MDB_cursor *mc, MDB_val *key, int *exactp);
+static int mdb_node_add(MDB_cursor *mc, indx_t indx,
+ MDB_val *key, MDB_val *data, pgno_t pgno, unsigned int flags);
+static void mdb_node_del(MDB_page *mp, indx_t indx, int ksize);
+static void mdb_node_shrink(MDB_page *mp, indx_t indx);
+static int mdb_node_move(MDB_cursor *csrc, MDB_cursor *cdst);
+static int mdb_node_read(MDB_txn *txn, MDB_node *leaf, MDB_val *data);
+static size_t mdb_leaf_size(MDB_env *env, MDB_val *key, MDB_val *data);
+static size_t mdb_branch_size(MDB_env *env, MDB_val *key);
+
+static int mdb_rebalance(MDB_cursor *mc);
+static int mdb_update_key(MDB_page *mp, indx_t indx, MDB_val *key);
+
+static void mdb_cursor_pop(MDB_cursor *mc);
+static int mdb_cursor_push(MDB_cursor *mc, MDB_page *mp);
+
+static int mdb_cursor_del0(MDB_cursor *mc, MDB_node *leaf);
+static int mdb_cursor_sibling(MDB_cursor *mc, int move_right);
+static int mdb_cursor_next(MDB_cursor *mc, MDB_val *key, MDB_val *data, MDB_cursor_op op);
+static int mdb_cursor_prev(MDB_cursor *mc, MDB_val *key, MDB_val *data, MDB_cursor_op op);
+static int mdb_cursor_set(MDB_cursor *mc, MDB_val *key, MDB_val *data, MDB_cursor_op op,
+ int *exactp);
+static int mdb_cursor_first(MDB_cursor *mc, MDB_val *key, MDB_val *data);
+static int mdb_cursor_last(MDB_cursor *mc, MDB_val *key, MDB_val *data);
+
+static void mdb_cursor_init(MDB_cursor *mc, MDB_txn *txn, MDB_dbi dbi, MDB_xcursor *mx);
+static void mdb_xcursor_init0(MDB_cursor *mc);
+static void mdb_xcursor_init1(MDB_cursor *mc, MDB_node *node);
+
+static int mdb_drop0(MDB_cursor *mc, int subs);
+static void mdb_default_cmp(MDB_txn *txn, MDB_dbi dbi);
+
+/** @cond */
+static MDB_cmp_func mdb_cmp_memn, mdb_cmp_memnr, mdb_cmp_int, mdb_cmp_cint, mdb_cmp_long;
+/** @endcond */
+
+#ifdef _WIN32
+static SECURITY_DESCRIPTOR mdb_null_sd;
+static SECURITY_ATTRIBUTES mdb_all_sa;
+static int mdb_sec_inited;
+#endif
+
+/** Return the library version info. */
+char *
+mdb_version(int *major, int *minor, int *patch)
+{
+ if (major) *major = MDB_VERSION_MAJOR;
+ if (minor) *minor = MDB_VERSION_MINOR;
+ if (patch) *patch = MDB_VERSION_PATCH;
+ return MDB_VERSION_STRING;
+}
+
+/** Table of descriptions for MDB @ref errors */
+static char *const mdb_errstr[] = {
+ "MDB_KEYEXIST: Key/data pair already exists",
+ "MDB_NOTFOUND: No matching key/data pair found",
+ "MDB_PAGE_NOTFOUND: Requested page not found",
+ "MDB_CORRUPTED: Located page was wrong type",
+ "MDB_PANIC: Update of meta page failed",
+ "MDB_VERSION_MISMATCH: Database environment version mismatch"
+};
+
+char *
+mdb_strerror(int err)
+{
+ if (!err)
+ return ("Successful return: 0");
+
+ if (err >= MDB_KEYEXIST && err <= MDB_VERSION_MISMATCH)
+ return mdb_errstr[err - MDB_KEYEXIST];
+
+ return strerror(err);
+}
+
+#if DEBUG
+/** Display a key in hexadecimal and return the address of the result.
+ * @param[in] key the key to display
+ * @param[in] buf the buffer to write into. Should always be #DKBUF.
+ * @return The key in hexadecimal form.
+ */
+char *
+mdb_dkey(MDB_val *key, char *buf)
+{
+ char *ptr = buf;
+ unsigned char *c = key->mv_data;
+ unsigned int i;
+ if (key->mv_size > MAXKEYSIZE)
+ return "MAXKEYSIZE";
+ /* may want to make this a dynamic check: if the key is mostly
+ * printable characters, print it as-is instead of converting to hex.
+ */
+#if 1
+ for (i=0; i<key->mv_size; i++)
+ ptr += sprintf(ptr, "%02x", *c++);
+#else
+ sprintf(buf, "%.*s", key->mv_size, key->mv_data);
+#endif
+ return buf;
+}
+#endif
+
+int
+mdb_cmp(MDB_txn *txn, MDB_dbi dbi, const MDB_val *a, const MDB_val *b)
+{
+ return txn->mt_dbxs[dbi].md_cmp(a, b);
+}
+
+int
+mdb_dcmp(MDB_txn *txn, MDB_dbi dbi, const MDB_val *a, const MDB_val *b)
+{
+ if (txn->mt_dbxs[dbi].md_dcmp)
+ return txn->mt_dbxs[dbi].md_dcmp(a, b);
+ else
+ return EINVAL; /* too bad you can't distinguish this from a valid result */
+}
+
+/** Allocate a single page.
+ * Re-use old malloc'd pages first, otherwise just malloc.
+ */
+static MDB_page *
+mdb_page_malloc(MDB_cursor *mc) {
+ MDB_page *ret;
+ if (mc->mc_txn->mt_env->me_dpages) {
+ ret = mc->mc_txn->mt_env->me_dpages;
+ mc->mc_txn->mt_env->me_dpages = ret->mp_next;
+ } else {
+ ret = malloc(mc->mc_txn->mt_env->me_psize);
+ }
+ return ret;
+}
+
+/** Allocate pages for writing.
+ * If there are free pages available from older transactions, they
+ * will be re-used first. Otherwise a new page will be allocated.
+ * @param[in] mc cursor A cursor handle identifying the transaction and
+ * database for which we are allocating.
+ * @param[in] num the number of pages to allocate.
+ * @return Address of the allocated page(s). Requests for multiple pages
+ * will always be satisfied by a single contiguous chunk of memory.
+ */
+static MDB_page *
+mdb_page_alloc(MDB_cursor *mc, int num)
+{
+ MDB_txn *txn = mc->mc_txn;
+ MDB_page *np;
+ pgno_t pgno = P_INVALID;
+ ID2 mid;
+
+ if (txn->mt_txnid > 2) {
+
+ if (!txn->mt_env->me_pghead && mc->mc_dbi != FREE_DBI &&
+ txn->mt_dbs[FREE_DBI].md_root != P_INVALID) {
+ /* See if there's anything in the free DB */
+ MDB_cursor m2;
+ MDB_node *leaf;
+ txnid_t *kptr, oldest;
+
+ mdb_cursor_init(&m2, txn, FREE_DBI, NULL);
+ mdb_page_search(&m2, NULL, 0);
+ leaf = NODEPTR(m2.mc_pg[m2.mc_top], 0);
+ kptr = (txnid_t *)NODEKEY(leaf);
+
+ {
+ unsigned int i;
+ oldest = txn->mt_txnid - 1;
+ for (i=0; i<txn->mt_env->me_txns->mti_numreaders; i++) {
+ txnid_t mr = txn->mt_env->me_txns->mti_readers[i].mr_txnid;
+ if (mr && mr < oldest)
+ oldest = mr;
+ }
+ }
+
+ if (oldest > *kptr) {
+ /* It's usable, grab it.
+ */
+ MDB_oldpages *mop;
+ MDB_val data;
+ pgno_t *idl;
+
+ mdb_node_read(txn, leaf, &data);
+ idl = (ID *) data.mv_data;
+ mop = malloc(sizeof(MDB_oldpages) + MDB_IDL_SIZEOF(idl) - sizeof(pgno_t));
+ mop->mo_next = txn->mt_env->me_pghead;
+ mop->mo_txnid = *kptr;
+ txn->mt_env->me_pghead = mop;
+ memcpy(mop->mo_pages, idl, MDB_IDL_SIZEOF(idl));
+
+#if DEBUG > 1
+ {
+ unsigned int i;
+ DPRINTF("IDL read txn %zu root %zu num %zu",
+ mop->mo_txnid, txn->mt_dbs[FREE_DBI].md_root, idl[0]);
+ for (i=0; i<idl[0]; i++) {
+ DPRINTF("IDL %zu", idl[i+1]);
+ }
+ }
+#endif
+ /* drop this IDL from the DB */
+ m2.mc_ki[m2.mc_top] = 0;
+ m2.mc_flags = C_INITIALIZED;
+ mdb_cursor_del(&m2, 0);
+ }
+ }
+ if (txn->mt_env->me_pghead) {
+ MDB_oldpages *mop = txn->mt_env->me_pghead;
+ if (num > 1) {
+ /* FIXME: For now, always use fresh pages. We
+ * really ought to search the free list for a
+ * contiguous range.
+ */
+ ;
+ } else {
+ /* peel pages off tail, so we only have to truncate the list */
+ pgno = MDB_IDL_LAST(mop->mo_pages);
+ if (MDB_IDL_IS_RANGE(mop->mo_pages)) {
+ mop->mo_pages[2]++;
+ if (mop->mo_pages[2] > mop->mo_pages[1])
+ mop->mo_pages[0] = 0;
+ } else {
+ mop->mo_pages[0]--;
+ }
+ if (MDB_IDL_IS_ZERO(mop->mo_pages)) {
+ txn->mt_env->me_pghead = mop->mo_next;
+ free(mop);
+ }
+ }
+ }
+ }
+
+ if (pgno == P_INVALID) {
+ /* DB size is maxed out */
+ if (txn->mt_next_pgno + num >= txn->mt_env->me_maxpg) {
+ DPUTS("DB size maxed out");
+ return NULL;
+ }
+ }
+ if (txn->mt_env->me_dpages && num == 1) {
+ np = txn->mt_env->me_dpages;
+ txn->mt_env->me_dpages = np->mp_next;
+ } else {
+ if ((np = malloc(txn->mt_env->me_psize * num )) == NULL)
+ return NULL;
+ }
+ if (pgno == P_INVALID) {
+ np->mp_pgno = txn->mt_next_pgno;
+ txn->mt_next_pgno += num;
+ } else {
+ np->mp_pgno = pgno;
+ }
+ mid.mid = np->mp_pgno;
+ mid.mptr = np;
+ mdb_mid2l_insert(txn->mt_u.dirty_list, &mid);
+
+ return np;
+}
+
+/** Touch a page: make it dirty and re-insert into tree with updated pgno.
+ * @param[in] mc cursor pointing to the page to be touched
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_page_touch(MDB_cursor *mc)
+{
+ MDB_page *mp = mc->mc_pg[mc->mc_top];
+ pgno_t pgno;
+
+ if (!F_ISSET(mp->mp_flags, P_DIRTY)) {
+ MDB_page *np;
+ if ((np = mdb_page_alloc(mc, 1)) == NULL)
+ return ENOMEM;
+ DPRINTF("touched db %u page %zu -> %zu", mc->mc_dbi, mp->mp_pgno, np->mp_pgno);
+ assert(mp->mp_pgno != np->mp_pgno);
+ mdb_midl_append(&mc->mc_txn->mt_free_pgs, mp->mp_pgno);
+ pgno = np->mp_pgno;
+ memcpy(np, mp, mc->mc_txn->mt_env->me_psize);
+ mp = np;
+ mp->mp_pgno = pgno;
+ mp->mp_flags |= P_DIRTY;
+
+finish:
+ /* Adjust other cursors pointing to mp */
+ if (mc->mc_flags & C_SUB) {
+ MDB_cursor *m2, *m3;
+ MDB_dbi dbi = mc->mc_dbi-1;
+
+ for (m2 = mc->mc_txn->mt_cursors[dbi]; m2; m2=m2->mc_next) {
+ if (m2 == mc) continue;
+ m3 = &m2->mc_xcursor->mx_cursor;
+ if (m3->mc_pg[mc->mc_top] == mc->mc_pg[mc->mc_top]) {
+ m3->mc_pg[mc->mc_top] = mp;
+ }
+ }
+ } else {
+ MDB_cursor *m2;
+
+ for (m2 = mc->mc_txn->mt_cursors[mc->mc_dbi]; m2; m2=m2->mc_next) {
+ if (m2 == mc) continue;
+ if (m2->mc_pg[mc->mc_top] == mc->mc_pg[mc->mc_top]) {
+ m2->mc_pg[mc->mc_top] = mp;
+ }
+ }
+ }
+ mc->mc_pg[mc->mc_top] = mp;
+ /** If this page has a parent, update the parent to point to
+ * this new page.
+ */
+ if (mc->mc_top)
+ SETPGNO(NODEPTR(mc->mc_pg[mc->mc_top-1], mc->mc_ki[mc->mc_top-1]), mp->mp_pgno);
+ else
+ mc->mc_db->md_root = mp->mp_pgno;
+ } else if (mc->mc_txn->mt_parent) {
+ MDB_page *np;
+ ID2 mid;
+ /* If txn has a parent, make sure the page is in our
+ * dirty list.
+ */
+ if (mc->mc_txn->mt_u.dirty_list[0].mid) {
+ unsigned x = mdb_mid2l_search(mc->mc_txn->mt_u.dirty_list, mp->mp_pgno);
+ if (x <= mc->mc_txn->mt_u.dirty_list[0].mid &&
+ mc->mc_txn->mt_u.dirty_list[x].mid == mp->mp_pgno) {
+ if (mc->mc_txn->mt_u.dirty_list[x].mptr != mp) {
+ mp = mc->mc_txn->mt_u.dirty_list[x].mptr;
+ mc->mc_pg[mc->mc_top] = mp;
+ }
+ return 0;
+ }
+ }
+ /* No - copy it */
+ np = mdb_page_malloc(mc);
+ memcpy(np, mp, mc->mc_txn->mt_env->me_psize);
+ mid.mid = np->mp_pgno;
+ mid.mptr = np;
+ mdb_mid2l_insert(mc->mc_txn->mt_u.dirty_list, &mid);
+ mp = np;
+ goto finish;
+ }
+ return 0;
+}
+
+int
+mdb_env_sync(MDB_env *env, int force)
+{
+ int rc = 0;
+ if (force || !F_ISSET(env->me_flags, MDB_NOSYNC)) {
+ if (fdatasync(env->me_fd))
+ rc = ErrCode();
+ }
+ return rc;
+}
+
+/** Make shadow copies of all of parent txn's cursors */
+static int
+mdb_cursor_shadow(MDB_txn *src, MDB_txn *dst)
+{
+ MDB_cursor *mc, *m2;
+ unsigned int i, j, size;
+
+ for (i=0;i<src->mt_numdbs; i++) {
+ if (src->mt_cursors[i]) {
+ size = sizeof(MDB_cursor);
+ if (src->mt_cursors[i]->mc_xcursor)
+ size += sizeof(MDB_xcursor);
+ for (m2 = src->mt_cursors[i]; m2; m2=m2->mc_next) {
+ mc = malloc(size);
+ if (!mc)
+ return ENOMEM;
+ mc->mc_orig = m2;
+ mc->mc_txn = dst;
+ mc->mc_dbi = i;
+ mc->mc_db = &dst->mt_dbs[i];
+ mc->mc_dbx = m2->mc_dbx;
+ mc->mc_dbflag = &dst->mt_dbflags[i];
+ mc->mc_snum = m2->mc_snum;
+ mc->mc_top = m2->mc_top;
+ mc->mc_flags = m2->mc_flags | C_SHADOW;
+ for (j=0; j<mc->mc_snum; j++) {
+ mc->mc_pg[j] = m2->mc_pg[j];
+ mc->mc_ki[j] = m2->mc_ki[j];
+ }
+ if (m2->mc_xcursor) {
+ MDB_xcursor *mx, *mx2;
+ mx = (MDB_xcursor *)(mc+1);
+ mc->mc_xcursor = mx;
+ mx2 = m2->mc_xcursor;
+ mx->mx_db = mx2->mx_db;
+ mx->mx_dbx = mx2->mx_dbx;
+ mx->mx_dbflag = mx2->mx_dbflag;
+ mx->mx_cursor.mc_txn = dst;
+ mx->mx_cursor.mc_dbi = mx2->mx_cursor.mc_dbi;
+ mx->mx_cursor.mc_db = &mx->mx_db;
+ mx->mx_cursor.mc_dbx = &mx->mx_dbx;
+ mx->mx_cursor.mc_dbflag = &mx->mx_dbflag;
+ mx->mx_cursor.mc_snum = mx2->mx_cursor.mc_snum;
+ mx->mx_cursor.mc_top = mx2->mx_cursor.mc_top;
+ mx->mx_cursor.mc_flags = mx2->mx_cursor.mc_flags | C_SHADOW;
+ for (j=0; j<mx2->mx_cursor.mc_snum; j++) {
+ mx->mx_cursor.mc_pg[j] = mx2->mx_cursor.mc_pg[j];
+ mx->mx_cursor.mc_ki[j] = mx2->mx_cursor.mc_ki[j];
+ }
+ } else {
+ mc->mc_xcursor = NULL;
+ }
+ mc->mc_next = dst->mt_cursors[i];
+ dst->mt_cursors[i] = mc;
+ }
+ }
+ }
+ return MDB_SUCCESS;
+}
+
+/** Merge shadow cursors back into parent's */
+static void
+mdb_cursor_merge(MDB_txn *txn)
+{
+ MDB_dbi i;
+ for (i=0; i<txn->mt_numdbs; i++) {
+ if (txn->mt_cursors[i]) {
+ MDB_cursor *mc;
+ while ((mc = txn->mt_cursors[i])) {
+ txn->mt_cursors[i] = mc->mc_next;
+ if (mc->mc_flags & C_SHADOW) {
+ MDB_cursor *m2 = mc->mc_orig;
+ unsigned int j;
+ m2->mc_snum = mc->mc_snum;
+ m2->mc_top = mc->mc_top;
+ for (j=0; j<mc->mc_snum; j++) {
+ m2->mc_pg[j] = mc->mc_pg[j];
+ m2->mc_ki[j] = mc->mc_ki[j];
+ }
+ }
+ if (mc->mc_flags & C_ALLOCD)
+ free(mc);
+ }
+ }
+ }
+}
+
+static void
+mdb_txn_reset0(MDB_txn *txn);
+
+/** Common code for #mdb_txn_begin() and #mdb_txn_renew().
+ * @param[in] txn the transaction handle to initialize
+ * @return 0 on success, non-zero on failure. This can only
+ * fail for read-only transactions, and then only if the
+ * reader table is full.
+ */
+static int
+mdb_txn_renew0(MDB_txn *txn)
+{
+ MDB_env *env = txn->mt_env;
+ char mt_dbflag = 0;
+
+ if (txn->mt_flags & MDB_TXN_RDONLY) {
+ MDB_reader *r = pthread_getspecific(env->me_txkey);
+ if (!r) {
+ unsigned int i;
+ pid_t pid = getpid();
+ pthread_t tid = pthread_self();
+
+ LOCK_MUTEX_R(env);
+ for (i=0; i<env->me_txns->mti_numreaders; i++)
+ if (env->me_txns->mti_readers[i].mr_pid == 0)
+ break;
+ if (i == env->me_maxreaders) {
+ UNLOCK_MUTEX_R(env);
+ return ENOMEM;
+ }
+ env->me_txns->mti_readers[i].mr_pid = pid;
+ env->me_txns->mti_readers[i].mr_tid = tid;
+ if (i >= env->me_txns->mti_numreaders)
+ env->me_txns->mti_numreaders = i+1;
+ UNLOCK_MUTEX_R(env);
+ r = &env->me_txns->mti_readers[i];
+ pthread_setspecific(env->me_txkey, r);
+ }
+ txn->mt_toggle = env->me_txns->mti_me_toggle;
+ txn->mt_txnid = env->me_txns->mti_txnid;
+ /* This happens if a different process was the
+ * last writer to the DB.
+ */
+ if (env->me_wtxnid < txn->mt_txnid)
+ mt_dbflag = DB_STALE;
+ r->mr_txnid = txn->mt_txnid;
+ txn->mt_u.reader = r;
+ } else {
+ LOCK_MUTEX_W(env);
+
+ txn->mt_txnid = env->me_txns->mti_txnid;
+ if (env->me_wtxnid < txn->mt_txnid)
+ mt_dbflag = DB_STALE;
+ txn->mt_txnid++;
+ txn->mt_toggle = env->me_txns->mti_me_toggle;
+ txn->mt_u.dirty_list = env->me_dirty_list;
+ txn->mt_u.dirty_list[0].mid = 0;
+ txn->mt_free_pgs = env->me_free_pgs;
+ txn->mt_free_pgs[0] = 0;
+ txn->mt_next_pgno = env->me_metas[txn->mt_toggle]->mm_last_pg+1;
+ env->me_txn = txn;
+ }
+
+ /* Copy the DB arrays */
+ LAZY_RWLOCK_RDLOCK(&env->me_dblock);
+ txn->mt_numdbs = env->me_numdbs;
+ txn->mt_dbxs = env->me_dbxs; /* mostly static anyway */
+ memcpy(txn->mt_dbs, env->me_metas[txn->mt_toggle]->mm_dbs, 2 * sizeof(MDB_db));
+ if (txn->mt_numdbs > 2)
+ memcpy(txn->mt_dbs+2, env->me_dbs[env->me_db_toggle]+2,
+ (txn->mt_numdbs - 2) * sizeof(MDB_db));
+ LAZY_RWLOCK_UNLOCK(&env->me_dblock);
+
+ memset(txn->mt_dbflags, mt_dbflag, env->me_numdbs);
+
+ return MDB_SUCCESS;
+}
+
+int
+mdb_txn_renew(MDB_txn *txn)
+{
+ int rc;
+
+ if (!txn)
+ return EINVAL;
+
+ if (txn->mt_env->me_flags & MDB_FATAL_ERROR) {
+ DPUTS("environment had fatal error, must shutdown!");
+ return MDB_PANIC;
+ }
+
+ rc = mdb_txn_renew0(txn);
+ if (rc == MDB_SUCCESS) {
+ DPRINTF("renew txn %zu%c %p on mdbenv %p, root page %zu",
+ txn->mt_txnid, (txn->mt_flags & MDB_TXN_RDONLY) ? 'r' : 'w',
+ (void *)txn, (void *)txn->mt_env, txn->mt_dbs[MAIN_DBI].md_root);
+ }
+ return rc;
+}
+
+int
+mdb_txn_begin(MDB_env *env, MDB_txn *parent, unsigned int flags, MDB_txn **ret)
+{
+ MDB_txn *txn;
+ int rc, size;
+
+ if (env->me_flags & MDB_FATAL_ERROR) {
+ DPUTS("environment had fatal error, must shutdown!");
+ return MDB_PANIC;
+ }
+ if (parent) {
+ /* parent already has an active child txn */
+ if (parent->mt_child) {
+ return EINVAL;
+ }
+ }
+ size = sizeof(MDB_txn) + env->me_maxdbs * (sizeof(MDB_db)+1);
+ if (!(flags & MDB_RDONLY))
+ size += env->me_maxdbs * sizeof(MDB_cursor *);
+
+ if ((txn = calloc(1, size)) == NULL) {
+ DPRINTF("calloc: %s", strerror(ErrCode()));
+ return ENOMEM;
+ }
+ txn->mt_dbs = (MDB_db *)(txn+1);
+ if (flags & MDB_RDONLY) {
+ txn->mt_flags |= MDB_TXN_RDONLY;
+ txn->mt_dbflags = (unsigned char *)(txn->mt_dbs + env->me_maxdbs);
+ } else {
+ txn->mt_cursors = (MDB_cursor **)(txn->mt_dbs + env->me_maxdbs);
+ txn->mt_dbflags = (unsigned char *)(txn->mt_cursors + env->me_maxdbs);
+ }
+ txn->mt_env = env;
+
+ if (parent) {
+ txn->mt_free_pgs = mdb_midl_alloc();
+ if (!txn->mt_free_pgs) {
+ free(txn);
+ return ENOMEM;
+ }
+ txn->mt_u.dirty_list = malloc(sizeof(ID2)*MDB_IDL_UM_SIZE);
+ if (!txn->mt_u.dirty_list) {
+ free(txn->mt_free_pgs);
+ free(txn);
+ return ENOMEM;
+ }
+ txn->mt_txnid = parent->mt_txnid;
+ txn->mt_toggle = parent->mt_toggle;
+ txn->mt_u.dirty_list[0].mid = 0;
+ txn->mt_free_pgs[0] = 0;
+ txn->mt_next_pgno = parent->mt_next_pgno;
+ parent->mt_child = txn;
+ txn->mt_parent = parent;
+ txn->mt_numdbs = parent->mt_numdbs;
+ txn->mt_dbxs = parent->mt_dbxs;
+ memcpy(txn->mt_dbs, parent->mt_dbs, txn->mt_numdbs * sizeof(MDB_db));
+ memcpy(txn->mt_dbflags, parent->mt_dbflags, txn->mt_numdbs);
+ mdb_cursor_shadow(parent, txn);
+ rc = 0;
+ } else {
+ rc = mdb_txn_renew0(txn);
+ }
+ if (rc)
+ free(txn);
+ else {
+ *ret = txn;
+ DPRINTF("begin txn %zu%c %p on mdbenv %p, root page %zu",
+ txn->mt_txnid, (txn->mt_flags & MDB_TXN_RDONLY) ? 'r' : 'w',
+ (void *) txn, (void *) env, txn->mt_dbs[MAIN_DBI].md_root);
+ }
+
+ return rc;
+}
+
+/** Common code for #mdb_txn_reset() and #mdb_txn_abort().
+ * @param[in] txn the transaction handle to reset
+ */
+static void
+mdb_txn_reset0(MDB_txn *txn)
+{
+ MDB_env *env = txn->mt_env;
+
+ if (F_ISSET(txn->mt_flags, MDB_TXN_RDONLY)) {
+ txn->mt_u.reader->mr_txnid = 0;
+ } else {
+ MDB_oldpages *mop;
+ MDB_page *dp;
+ unsigned int i;
+
+ /* close(free) all cursors */
+ for (i=0; i<txn->mt_numdbs; i++) {
+ if (txn->mt_cursors[i]) {
+ MDB_cursor *mc;
+ while ((mc = txn->mt_cursors[i])) {
+ txn->mt_cursors[i] = mc->mc_next;
+ if (mc->mc_flags & C_ALLOCD)
+ free(mc);
+ }
+ }
+ }
+
+ /* return all dirty pages to dpage list */
+ for (i=1; i<=txn->mt_u.dirty_list[0].mid; i++) {
+ dp = txn->mt_u.dirty_list[i].mptr;
+ if (!IS_OVERFLOW(dp) || dp->mp_pages == 1) {
+ dp->mp_next = txn->mt_env->me_dpages;
+ txn->mt_env->me_dpages = dp;
+ } else {
+ /* large pages just get freed directly */
+ free(dp);
+ }
+ }
+
+ if (txn->mt_parent) {
+ txn->mt_parent->mt_child = NULL;
+ free(txn->mt_free_pgs);
+ free(txn->mt_u.dirty_list);
+ return;
+ } else {
+ if (mdb_midl_shrink(&txn->mt_free_pgs))
+ env->me_free_pgs = txn->mt_free_pgs;
+ }
+
+ while ((mop = txn->mt_env->me_pghead)) {
+ txn->mt_env->me_pghead = mop->mo_next;
+ free(mop);
+ }
+
+ env->me_txn = NULL;
+ /* The writer mutex was locked in mdb_txn_begin. */
+ UNLOCK_MUTEX_W(env);
+ }
+}
+
+void
+mdb_txn_reset(MDB_txn *txn)
+{
+ if (txn == NULL)
+ return;
+
+ DPRINTF("reset txn %zu%c %p on mdbenv %p, root page %zu",
+ txn->mt_txnid, (txn->mt_flags & MDB_TXN_RDONLY) ? 'r' : 'w',
+ (void *) txn, (void *)txn->mt_env, txn->mt_dbs[MAIN_DBI].md_root);
+
+ mdb_txn_reset0(txn);
+}
+
+void
+mdb_txn_abort(MDB_txn *txn)
+{
+ if (txn == NULL)
+ return;
+
+ DPRINTF("abort txn %zu%c %p on mdbenv %p, root page %zu",
+ txn->mt_txnid, (txn->mt_flags & MDB_TXN_RDONLY) ? 'r' : 'w',
+ (void *)txn, (void *)txn->mt_env, txn->mt_dbs[MAIN_DBI].md_root);
+
+ if (txn->mt_child)
+ mdb_txn_abort(txn->mt_child);
+
+ mdb_txn_reset0(txn);
+ free(txn);
+}
+
+int
+mdb_txn_commit(MDB_txn *txn)
+{
+ int n, done;
+ unsigned int i;
+ ssize_t rc;
+ off_t size;
+ MDB_page *dp;
+ MDB_env *env;
+ pgno_t next;
+ MDB_cursor mc;
+
+ assert(txn != NULL);
+ assert(txn->mt_env != NULL);
+
+ if (txn->mt_child) {
+ mdb_txn_commit(txn->mt_child);
+ txn->mt_child = NULL;
+ }
+
+ env = txn->mt_env;
+
+ if (F_ISSET(txn->mt_flags, MDB_TXN_RDONLY)) {
+ if (txn->mt_numdbs > env->me_numdbs) {
+ /* update the DB tables */
+ int toggle = !env->me_db_toggle;
+ MDB_db *ip, *jp;
+ MDB_dbi i;
+
+ ip = &env->me_dbs[toggle][env->me_numdbs];
+ jp = &txn->mt_dbs[env->me_numdbs];
+ LAZY_RWLOCK_WRLOCK(&env->me_dblock);
+ for (i = env->me_numdbs; i < txn->mt_numdbs; i++) {
+ *ip++ = *jp++;
+ }
+
+ env->me_db_toggle = toggle;
+ env->me_numdbs = txn->mt_numdbs;
+ LAZY_RWLOCK_UNLOCK(&env->me_dblock);
+ }
+ mdb_txn_abort(txn);
+ return MDB_SUCCESS;
+ }
+
+ if (F_ISSET(txn->mt_flags, MDB_TXN_ERROR)) {
+ DPUTS("error flag is set, can't commit");
+ if (txn->mt_parent)
+ txn->mt_parent->mt_flags |= MDB_TXN_ERROR;
+ mdb_txn_abort(txn);
+ return EINVAL;
+ }
+
+ /* Merge (and close) our cursors with parent's */
+ mdb_cursor_merge(txn);
+
+ if (txn->mt_parent) {
+ MDB_db *ip, *jp;
+ MDB_dbi i;
+ unsigned x, y;
+ ID2L dst, src;
+
+ /* Update parent's DB table */
+ ip = &txn->mt_parent->mt_dbs[2];
+ jp = &txn->mt_dbs[2];
+ for (i = 2; i < txn->mt_numdbs; i++) {
+ if (ip->md_root != jp->md_root)
+ *ip = *jp;
+ ip++; jp++;
+ }
+ txn->mt_parent->mt_numdbs = txn->mt_numdbs;
+
+ /* Append our free list to parent's */
+ mdb_midl_append_list(&txn->mt_parent->mt_free_pgs,
+ txn->mt_free_pgs);
+ mdb_midl_free(txn->mt_free_pgs);
+
+ /* Merge our dirty list with parent's */
+ dst = txn->mt_parent->mt_u.dirty_list;
+ src = txn->mt_u.dirty_list;
+ x = mdb_mid2l_search(dst, src[1].mid);
+ for (y=1; y<=src[0].mid; y++) {
+ while (x <= dst[0].mid && dst[x].mid != src[y].mid) x++;
+ if (x > dst[0].mid)
+ break;
+ free(dst[x].mptr);
+ dst[x].mptr = src[y].mptr;
+ }
+ x = dst[0].mid;
+ for (; y<=src[0].mid; y++) {
+ if (++x >= MDB_IDL_UM_MAX)
+ return ENOMEM;
+ dst[x] = src[y];
+ }
+ dst[0].mid = x;
+ free(txn->mt_u.dirty_list);
+ txn->mt_parent->mt_child = NULL;
+ free(txn);
+ return MDB_SUCCESS;
+ }
+
+ if (txn != env->me_txn) {
+ DPUTS("attempt to commit unknown transaction");
+ mdb_txn_abort(txn);
+ return EINVAL;
+ }
+
+ if (!txn->mt_u.dirty_list[0].mid)
+ goto done;
+
+ DPRINTF("committing txn %zu %p on mdbenv %p, root page %zu",
+ txn->mt_txnid, (void *)txn, (void *)env, txn->mt_dbs[MAIN_DBI].md_root);
+
+ mdb_cursor_init(&mc, txn, FREE_DBI, NULL);
+
+ /* should only be one record now */
+ if (env->me_pghead) {
+ /* make sure first page of freeDB is touched and on freelist */
+ mdb_page_search(&mc, NULL, 1);
+ }
+ /* save to free list */
+ if (!MDB_IDL_IS_ZERO(txn->mt_free_pgs)) {
+ MDB_val key, data;
+ pgno_t i;
+
+ /* make sure last page of freeDB is touched and on freelist */
+ key.mv_size = MAXKEYSIZE+1;
+ key.mv_data = NULL;
+ mdb_page_search(&mc, &key, 1);
+
+ mdb_midl_sort(txn->mt_free_pgs);
+#if DEBUG > 1
+ {
+ unsigned int i;
+ ID *idl = txn->mt_free_pgs;
+ DPRINTF("IDL write txn %zu root %zu num %zu",
+ txn->mt_txnid, txn->mt_dbs[FREE_DBI].md_root, idl[0]);
+ for (i=0; i<idl[0]; i++) {
+ DPRINTF("IDL %zu", idl[i+1]);
+ }
+ }
+#endif
+ /* write to last page of freeDB */
+ key.mv_size = sizeof(pgno_t);
+ key.mv_data = &txn->mt_txnid;
+ data.mv_data = txn->mt_free_pgs;
+ /* The free list can still grow during this call,
+ * despite the pre-emptive touches above. So check
+ * and make sure the entire thing got written.
+ */
+ do {
+ i = txn->mt_free_pgs[0];
+ data.mv_size = MDB_IDL_SIZEOF(txn->mt_free_pgs);
+ rc = mdb_cursor_put(&mc, &key, &data, 0);
+ if (rc) {
+ mdb_txn_abort(txn);
+ return rc;
+ }
+ } while (i != txn->mt_free_pgs[0]);
+ if (mdb_midl_shrink(&txn->mt_free_pgs))
+ env->me_free_pgs = txn->mt_free_pgs;
+ }
+ /* should only be one record now */
+ if (env->me_pghead) {
+ MDB_val key, data;
+ MDB_oldpages *mop;
+
+ mop = env->me_pghead;
+ env->me_pghead = NULL;
+ key.mv_size = sizeof(pgno_t);
+ key.mv_data = &mop->mo_txnid;
+ data.mv_size = MDB_IDL_SIZEOF(mop->mo_pages);
+ data.mv_data = mop->mo_pages;
+ mdb_cursor_put(&mc, &key, &data, 0);
+ free(mop);
+ }
+
+ /* Update DB root pointers. Their pages have already been
+ * touched so this is all in-place and cannot fail.
+ */
+ {
+ MDB_dbi i;
+ MDB_val data;
+ data.mv_size = sizeof(MDB_db);
+
+ mdb_cursor_init(&mc, txn, MAIN_DBI, NULL);
+ for (i = 2; i < txn->mt_numdbs; i++) {
+ if (txn->mt_dbflags[i] & DB_DIRTY) {
+ data.mv_data = &txn->mt_dbs[i];
+ mdb_cursor_put(&mc, &txn->mt_dbxs[i].md_name, &data, 0);
+ }
+ }
+ }
+
+ /* Commit up to MDB_COMMIT_PAGES dirty pages to disk until done.
+ */
+ next = 0;
+ i = 1;
+ do {
+#ifdef _WIN32
+ /* Windows actually supports scatter/gather I/O, but only on
+ * unbuffered file handles. Since we're relying on the OS page
+ * cache for all our data, that's self-defeating. So we just
+ * write pages one at a time. We use the ov structure to set
+ * the write offset, to at least save the overhead of a Seek
+ * system call.
+ */
+ OVERLAPPED ov;
+ memset(&ov, 0, sizeof(ov));
+ for (; i<=txn->mt_u.dirty_list[0].mid; i++) {
+ size_t wsize;
+ dp = txn->mt_u.dirty_list[i].mptr;
+ DPRINTF("committing page %zu", dp->mp_pgno);
+ size = dp->mp_pgno * env->me_psize;
+ ov.Offset = size & 0xffffffff;
+ ov.OffsetHigh = size >> 16;
+ ov.OffsetHigh >>= 16;
+ /* clear dirty flag */
+ dp->mp_flags &= ~P_DIRTY;
+ wsize = env->me_psize;
+ if (IS_OVERFLOW(dp)) wsize *= dp->mp_pages;
+ rc = WriteFile(env->me_fd, dp, wsize, NULL, &ov);
+ if (!rc) {
+ n = ErrCode();
+ DPRINTF("WriteFile: %d", n);
+ mdb_txn_abort(txn);
+ return n;
+ }
+ }
+ done = 1;
+#else
+ struct iovec iov[MDB_COMMIT_PAGES];
+ n = 0;
+ done = 1;
+ size = 0;
+ for (; i<=txn->mt_u.dirty_list[0].mid; i++) {
+ dp = txn->mt_u.dirty_list[i].mptr;
+ if (dp->mp_pgno != next) {
+ if (n) {
+ DPRINTF("committing %u dirty pages", n);
+ rc = writev(env->me_fd, iov, n);
+ if (rc != size) {
+ n = ErrCode();
+ if (rc > 0)
+ DPUTS("short write, filesystem full?");
+ else
+ DPRINTF("writev: %s", strerror(n));
+ mdb_txn_abort(txn);
+ return n;
+ }
+ n = 0;
+ size = 0;
+ }
+ lseek(env->me_fd, dp->mp_pgno * env->me_psize, SEEK_SET);
+ next = dp->mp_pgno;
+ }
+ DPRINTF("committing page %zu", dp->mp_pgno);
+ iov[n].iov_len = env->me_psize;
+ if (IS_OVERFLOW(dp)) iov[n].iov_len *= dp->mp_pages;
+ iov[n].iov_base = dp;
+ size += iov[n].iov_len;
+ next = dp->mp_pgno + (IS_OVERFLOW(dp) ? dp->mp_pages : 1);
+ /* clear dirty flag */
+ dp->mp_flags &= ~P_DIRTY;
+ if (++n >= MDB_COMMIT_PAGES) {
+ done = 0;
+ i++;
+ break;
+ }
+ }
+
+ if (n == 0)
+ break;
+
+ DPRINTF("committing %u dirty pages", n);
+ rc = writev(env->me_fd, iov, n);
+ if (rc != size) {
+ n = ErrCode();
+ if (rc > 0)
+ DPUTS("short write, filesystem full?");
+ else
+ DPRINTF("writev: %s", strerror(n));
+ mdb_txn_abort(txn);
+ return n;
+ }
+#endif
+ } while (!done);
+
+ /* Drop the dirty pages.
+ */
+ for (i=1; i<=txn->mt_u.dirty_list[0].mid; i++) {
+ dp = txn->mt_u.dirty_list[i].mptr;
+ if (!IS_OVERFLOW(dp) || dp->mp_pages == 1) {
+ dp->mp_next = txn->mt_env->me_dpages;
+ txn->mt_env->me_dpages = dp;
+ } else {
+ free(dp);
+ }
+ txn->mt_u.dirty_list[i].mid = 0;
+ }
+ txn->mt_u.dirty_list[0].mid = 0;
+
+ if ((n = mdb_env_sync(env, 0)) != 0 ||
+ (n = mdb_env_write_meta(txn)) != MDB_SUCCESS) {
+ mdb_txn_abort(txn);
+ return n;
+ }
+ env->me_wtxnid = txn->mt_txnid;
+
+done:
+ env->me_txn = NULL;
+ /* update the DB tables */
+ {
+ int toggle = !env->me_db_toggle;
+ MDB_db *ip, *jp;
+ MDB_dbi i;
+
+ ip = &env->me_dbs[toggle][2];
+ jp = &txn->mt_dbs[2];
+ LAZY_RWLOCK_WRLOCK(&env->me_dblock);
+ for (i = 2; i < txn->mt_numdbs; i++) {
+ if (ip->md_root != jp->md_root)
+ *ip = *jp;
+ ip++; jp++;
+ }
+
+ env->me_db_toggle = toggle;
+ env->me_numdbs = txn->mt_numdbs;
+ LAZY_RWLOCK_UNLOCK(&env->me_dblock);
+ }
+
+ UNLOCK_MUTEX_W(env);
+ free(txn);
+
+ return MDB_SUCCESS;
+}
+
+/** Read the environment parameters of a DB environment before
+ * mapping it into memory.
+ * @param[in] env the environment handle
+ * @param[out] meta address of where to store the meta information
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_env_read_header(MDB_env *env, MDB_meta *meta)
+{
+ char page[PAGESIZE];
+ MDB_page *p;
+ MDB_meta *m;
+ int rc, err;
+
+ /* We don't know the page size yet, so use a minimum value.
+ */
+
+#ifdef _WIN32
+ if (!ReadFile(env->me_fd, page, PAGESIZE, (DWORD *)&rc, NULL) || rc == 0)
+#else
+ if ((rc = read(env->me_fd, page, PAGESIZE)) == 0)
+#endif
+ {
+ return ENOENT;
+ }
+ else if (rc != PAGESIZE) {
+ err = ErrCode();
+ if (rc > 0)
+ err = EINVAL;
+ DPRINTF("read: %s", strerror(err));
+ return err;
+ }
+
+ p = (MDB_page *)page;
+
+ if (!F_ISSET(p->mp_flags, P_META)) {
+ DPRINTF("page %zu not a meta page", p->mp_pgno);
+ return EINVAL;
+ }
+
+ m = METADATA(p);
+ if (m->mm_magic != MDB_MAGIC) {
+ DPUTS("meta has invalid magic");
+ return EINVAL;
+ }
+
+ if (m->mm_version != MDB_VERSION) {
+ DPRINTF("database is version %u, expected version %u",
+ m->mm_version, MDB_VERSION);
+ return MDB_VERSION_MISMATCH;
+ }
+
+ memcpy(meta, m, sizeof(*m));
+ return 0;
+}
+
+/** Write the environment parameters of a freshly created DB environment.
+ * @param[in] env the environment handle
+ * @param[out] meta address of where to store the meta information
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_env_init_meta(MDB_env *env, MDB_meta *meta)
+{
+ MDB_page *p, *q;
+ MDB_meta *m;
+ int rc;
+ unsigned int psize;
+
+ DPUTS("writing new meta page");
+
+ GET_PAGESIZE(psize);
+
+ meta->mm_magic = MDB_MAGIC;
+ meta->mm_version = MDB_VERSION;
+ meta->mm_psize = psize;
+ meta->mm_last_pg = 1;
+ meta->mm_flags = env->me_flags & 0xffff;
+ meta->mm_flags |= MDB_INTEGERKEY;
+ meta->mm_dbs[0].md_root = P_INVALID;
+ meta->mm_dbs[1].md_root = P_INVALID;
+
+ p = calloc(2, psize);
+ p->mp_pgno = 0;
+ p->mp_flags = P_META;
+
+ m = METADATA(p);
+ memcpy(m, meta, sizeof(*meta));
+
+ q = (MDB_page *)((char *)p + psize);
+
+ q->mp_pgno = 1;
+ q->mp_flags = P_META;
+
+ m = METADATA(q);
+ memcpy(m, meta, sizeof(*meta));
+
+#ifdef _WIN32
+ {
+ DWORD len;
+ rc = WriteFile(env->me_fd, p, psize * 2, &len, NULL);
+ rc = (len == psize * 2) ? MDB_SUCCESS : ErrCode();
+ }
+#else
+ rc = write(env->me_fd, p, psize * 2);
+ rc = (rc == (int)psize * 2) ? MDB_SUCCESS : ErrCode();
+#endif
+ free(p);
+ return rc;
+}
+
+/** Update the environment info to commit a transaction.
+ * @param[in] txn the transaction that's being committed
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_env_write_meta(MDB_txn *txn)
+{
+ MDB_env *env;
+ MDB_meta meta, metab;
+ off_t off;
+ int rc, len, toggle;
+ char *ptr;
+#ifdef _WIN32
+ OVERLAPPED ov;
+#endif
+
+ assert(txn != NULL);
+ assert(txn->mt_env != NULL);
+
+ toggle = !txn->mt_toggle;
+ DPRINTF("writing meta page %d for root page %zu",
+ toggle, txn->mt_dbs[MAIN_DBI].md_root);
+
+ env = txn->mt_env;
+
+ metab.mm_txnid = env->me_metas[toggle]->mm_txnid;
+ metab.mm_last_pg = env->me_metas[toggle]->mm_last_pg;
+
+ ptr = (char *)&meta;
+ off = offsetof(MDB_meta, mm_dbs[0].md_depth);
+ len = sizeof(MDB_meta) - off;
+
+ ptr += off;
+ meta.mm_dbs[0] = txn->mt_dbs[0];
+ meta.mm_dbs[1] = txn->mt_dbs[1];
+ meta.mm_last_pg = txn->mt_next_pgno - 1;
+ meta.mm_txnid = txn->mt_txnid;
+
+ if (toggle)
+ off += env->me_psize;
+ off += PAGEHDRSZ;
+
+ /* Write to the SYNC fd */
+#ifdef _WIN32
+ {
+ memset(&ov, 0, sizeof(ov));
+ ov.Offset = off;
+ WriteFile(env->me_mfd, ptr, len, (DWORD *)&rc, &ov);
+ }
+#else
+ rc = pwrite(env->me_mfd, ptr, len, off);
+#endif
+ if (rc != len) {
+ int r2;
+ rc = ErrCode();
+ DPUTS("write failed, disk error?");
+ /* On a failure, the pagecache still contains the new data.
+ * Write some old data back, to prevent it from being used.
+ * Use the non-SYNC fd; we know it will fail anyway.
+ */
+ meta.mm_last_pg = metab.mm_last_pg;
+ meta.mm_txnid = metab.mm_txnid;
+#ifdef _WIN32
+ WriteFile(env->me_fd, ptr, len, NULL, &ov);
+#else
+ r2 = pwrite(env->me_fd, ptr, len, off);
+#endif
+ env->me_flags |= MDB_FATAL_ERROR;
+ return rc;
+ }
+ /* Memory ordering issues are irrelevant; since the entire writer
+ * is wrapped by wmutex, all of these changes will become visible
+ * after the wmutex is unlocked. Since the DB is multi-version,
+ * readers will get consistent data regardless of how fresh or
+ * how stale their view of these values is.
+ */
+ LAZY_MUTEX_LOCK(&env->me_txns->mti_mutex);
+ txn->mt_env->me_txns->mti_me_toggle = toggle;
+ txn->mt_env->me_txns->mti_txnid = txn->mt_txnid;
+ LAZY_MUTEX_UNLOCK(&env->me_txns->mti_mutex);
+
+ return MDB_SUCCESS;
+}
+
+/** Check both meta pages to see which one is newer.
+ * @param[in] env the environment handle
+ * @param[out] which address of where to store the meta toggle ID
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_env_read_meta(MDB_env *env, int *which)
+{
+ int toggle = 0;
+
+ assert(env != NULL);
+
+ if (env->me_metas[0]->mm_txnid < env->me_metas[1]->mm_txnid)
+ toggle = 1;
+
+ DPRINTF("Using meta page %d", toggle);
+ *which = toggle;
+
+ return MDB_SUCCESS;
+}
+
+int
+mdb_env_create(MDB_env **env)
+{
+ MDB_env *e;
+
+ e = calloc(1, sizeof(MDB_env));
+ if (!e)
+ return ENOMEM;
+
+ e->me_free_pgs = mdb_midl_alloc();
+ if (!e->me_free_pgs) {
+ free(e);
+ return ENOMEM;
+ }
+ e->me_maxreaders = DEFAULT_READERS;
+ e->me_maxdbs = 2;
+ e->me_fd = INVALID_HANDLE_VALUE;
+ e->me_lfd = INVALID_HANDLE_VALUE;
+ e->me_mfd = INVALID_HANDLE_VALUE;
+ *env = e;
+ return MDB_SUCCESS;
+}
+
+int
+mdb_env_set_mapsize(MDB_env *env, size_t size)
+{
+ if (env->me_map)
+ return EINVAL;
+ env->me_mapsize = size;
+ if (env->me_psize)
+ env->me_maxpg = env->me_mapsize / env->me_psize;
+ return MDB_SUCCESS;
+}
+
+int
+mdb_env_set_maxdbs(MDB_env *env, MDB_dbi dbs)
+{
+ if (env->me_map)
+ return EINVAL;
+ env->me_maxdbs = dbs;
+ return MDB_SUCCESS;
+}
+
+int
+mdb_env_set_maxreaders(MDB_env *env, unsigned int readers)
+{
+ if (env->me_map || readers < 1)
+ return EINVAL;
+ env->me_maxreaders = readers;
+ return MDB_SUCCESS;
+}
+
+int
+mdb_env_get_maxreaders(MDB_env *env, unsigned int *readers)
+{
+ if (!env || !readers)
+ return EINVAL;
+ *readers = env->me_maxreaders;
+ return MDB_SUCCESS;
+}
+
+/** Further setup required for opening an MDB environment
+ */
+static int
+mdb_env_open2(MDB_env *env, unsigned int flags)
+{
+ int i, newenv = 0, toggle;
+ MDB_meta meta;
+ MDB_page *p;
+
+ env->me_flags = flags;
+
+ memset(&meta, 0, sizeof(meta));
+
+ if ((i = mdb_env_read_header(env, &meta)) != 0) {
+ if (i != ENOENT)
+ return i;
+ DPUTS("new mdbenv");
+ newenv = 1;
+ }
+
+ if (!env->me_mapsize) {
+ env->me_mapsize = newenv ? DEFAULT_MAPSIZE : meta.mm_mapsize;
+ }
+
+#ifdef _WIN32
+ {
+ HANDLE mh;
+ LONG sizelo, sizehi;
+ sizelo = env->me_mapsize & 0xffffffff;
+ sizehi = env->me_mapsize >> 16; /* pointless on WIN32, only needed on W64 */
+ sizehi >>= 16;
+ /* Windows won't create mappings for zero length files.
+ * Just allocate the maxsize right now.
+ */
+ if (newenv) {
+ SetFilePointer(env->me_fd, sizelo, sizehi ? &sizehi : NULL, 0);
+ if (!SetEndOfFile(env->me_fd))
+ return ErrCode();
+ SetFilePointer(env->me_fd, 0, NULL, 0);
+ }
+ mh = CreateFileMapping(env->me_fd, NULL, PAGE_READONLY,
+ sizehi, sizelo, NULL);
+ if (!mh)
+ return ErrCode();
+ env->me_map = MapViewOfFileEx(mh, FILE_MAP_READ, 0, 0, env->me_mapsize,
+ meta.mm_address);
+ CloseHandle(mh);
+ if (!env->me_map)
+ return ErrCode();
+ }
+#else
+ i = MAP_SHARED;
+ if (meta.mm_address && (flags & MDB_FIXEDMAP))
+ i |= MAP_FIXED;
+ env->me_map = mmap(meta.mm_address, env->me_mapsize, PROT_READ, i,
+ env->me_fd, 0);
+ if (env->me_map == MAP_FAILED)
+ return ErrCode();
+#endif
+
+ if (newenv) {
+ meta.mm_mapsize = env->me_mapsize;
+ if (flags & MDB_FIXEDMAP)
+ meta.mm_address = env->me_map;
+ i = mdb_env_init_meta(env, &meta);
+ if (i != MDB_SUCCESS) {
+ munmap(env->me_map, env->me_mapsize);
+ return i;
+ }
+ }
+ env->me_psize = meta.mm_psize;
+
+ env->me_maxpg = env->me_mapsize / env->me_psize;
+
+ p = (MDB_page *)env->me_map;
+ env->me_metas[0] = METADATA(p);
+ env->me_metas[1] = (MDB_meta *)((char *)env->me_metas[0] + meta.mm_psize);
+
+ if ((i = mdb_env_read_meta(env, &toggle)) != 0)
+ return i;
+
+ DPRINTF("opened database version %u, pagesize %u",
+ env->me_metas[toggle]->mm_version, env->me_psize);
+ DPRINTF("depth: %u", env->me_metas[toggle]->mm_dbs[MAIN_DBI].md_depth);
+ DPRINTF("entries: %zu", env->me_metas[toggle]->mm_dbs[MAIN_DBI].md_entries);
+ DPRINTF("branch pages: %zu", env->me_metas[toggle]->mm_dbs[MAIN_DBI].md_branch_pages);
+ DPRINTF("leaf pages: %zu", env->me_metas[toggle]->mm_dbs[MAIN_DBI].md_leaf_pages);
+ DPRINTF("overflow pages: %zu", env->me_metas[toggle]->mm_dbs[MAIN_DBI].md_overflow_pages);
+ DPRINTF("root: %zu", env->me_metas[toggle]->mm_dbs[MAIN_DBI].md_root);
+
+ return MDB_SUCCESS;
+}
+
+#ifndef _WIN32
+/** Release a reader thread's slot in the reader lock table.
+ * This function is called automatically when a thread exits.
+ * Windows doesn't support destructor callbacks for thread-specific storage,
+ * so this function is not compiled there.
+ * @param[in] ptr This points to the slot in the reader lock table.
+ */
+static void
+mdb_env_reader_dest(void *ptr)
+{
+ MDB_reader *reader = ptr;
+
+ reader->mr_txnid = 0;
+ reader->mr_pid = 0;
+ reader->mr_tid = 0;
+}
+#endif
+
+/** Downgrade the exclusive lock on the region back to shared */
+static void
+mdb_env_share_locks(MDB_env *env)
+{
+ int toggle = 0;
+
+ if (env->me_metas[0]->mm_txnid < env->me_metas[1]->mm_txnid)
+ toggle = 1;
+ env->me_txns->mti_me_toggle = toggle;
+ env->me_txns->mti_txnid = env->me_metas[toggle]->mm_txnid;
+
+#ifdef _WIN32
+ {
+ OVERLAPPED ov;
+ /* First acquire a shared lock. The Unlock will
+ * then release the existing exclusive lock.
+ */
+ memset(&ov, 0, sizeof(ov));
+ LockFileEx(env->me_lfd, 0, 0, 1, 0, &ov);
+ UnlockFile(env->me_lfd, 0, 0, 1, 0);
+ }
+#else
+ {
+ struct flock lock_info;
+ /* The shared lock replaces the existing lock */
+ memset((void *)&lock_info, 0, sizeof(lock_info));
+ lock_info.l_type = F_RDLCK;
+ lock_info.l_whence = SEEK_SET;
+ lock_info.l_start = 0;
+ lock_info.l_len = 1;
+ fcntl(env->me_lfd, F_SETLK, &lock_info);
+ }
+#endif
+}
+#if defined(_WIN32) || defined(__APPLE__)
+/*
+ * hash_64 - 64 bit Fowler/Noll/Vo-0 FNV-1a hash code
+ *
+ * @(#) $Revision: 5.1 $
+ * @(#) $Id: hash_64a.c,v 5.1 2009/06/30 09:01:38 chongo Exp $
+ * @(#) $Source: /usr/local/src/cmd/fnv/RCS/hash_64a.c,v $
+ *
+ * http://www.isthe.com/chongo/tech/comp/fnv/index.html
+ *
+ ***
+ *
+ * Please do not copyright this code. This code is in the public domain.
+ *
+ * LANDON CURT NOLL DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
+ * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO
+ * EVENT SHALL LANDON CURT NOLL BE LIABLE FOR ANY SPECIAL, INDIRECT OR
+ * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF
+ * USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR
+ * OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
+ * PERFORMANCE OF THIS SOFTWARE.
+ *
+ * By:
+ * chongo <Landon Curt Noll> /\oo/\
+ * http://www.isthe.com/chongo/
+ *
+ * Share and Enjoy! :-)
+ */
+
+typedef unsigned long long mdb_hash_t;
+#define MDB_HASH_INIT ((mdb_hash_t)0xcbf29ce484222325ULL)
+
+/** perform a 64 bit Fowler/Noll/Vo FNV-1a hash on a buffer
+ * @param[in] str string to hash
+ * @param[in] hval initial value for hash
+ * @return 64 bit hash
+ *
+ * NOTE: To use the recommended 64 bit FNV-1a hash, use MDB_HASH_INIT as the
+ * hval arg on the first call.
+ */
+static mdb_hash_t
+mdb_hash_str(char *str, mdb_hash_t hval)
+{
+ unsigned char *s = (unsigned char *)str; /* unsigned string */
+ /*
+ * FNV-1a hash each octet of the string
+ */
+ while (*s) {
+ /* xor the bottom with the current octet */
+ hval ^= (mdb_hash_t)*s++;
+
+ /* multiply by the 64 bit FNV magic prime mod 2^64 */
+ hval += (hval << 1) + (hval << 4) + (hval << 5) +
+ (hval << 7) + (hval << 8) + (hval << 40);
+ }
+ /* return our new hash value */
+ return hval;
+}
+
+/** Hash the string and output the hash in hex.
+ * @param[in] str string to hash
+ * @param[out] hexbuf an array of 17 chars to hold the hash
+ */
+static void
+mdb_hash_hex(char *str, char *hexbuf)
+{
+ int i;
+ mdb_hash_t h = mdb_hash_str(str, MDB_HASH_INIT);
+ for (i=0; i<8; i++) {
+ hexbuf += sprintf(hexbuf, "%02x", (unsigned int)h & 0xff);
+ h >>= 8;
+ }
+}
+#endif
+
+/** Open and/or initialize the lock region for the environment.
+ * @param[in] env The MDB environment.
+ * @param[in] lpath The pathname of the file used for the lock region.
+ * @param[in] mode The Unix permissions for the file, if we create it.
+ * @param[out] excl Set to true if we got an exclusive lock on the region.
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_env_setup_locks(MDB_env *env, char *lpath, int mode, int *excl)
+{
+ int rc;
+ off_t size, rsize;
+
+ *excl = 0;
+
+#ifdef _WIN32
+ if ((env->me_lfd = CreateFile(lpath, GENERIC_READ|GENERIC_WRITE,
+ FILE_SHARE_READ|FILE_SHARE_WRITE, NULL, OPEN_ALWAYS,
+ FILE_ATTRIBUTE_NORMAL, NULL)) == INVALID_HANDLE_VALUE) {
+ rc = ErrCode();
+ return rc;
+ }
+ /* Try to get exclusive lock. If we succeed, then
+ * nobody is using the lock region and we should initialize it.
+ */
+ {
+ if (LockFile(env->me_lfd, 0, 0, 1, 0)) {
+ *excl = 1;
+ } else {
+ OVERLAPPED ov;
+ memset(&ov, 0, sizeof(ov));
+ if (!LockFileEx(env->me_lfd, 0, 0, 1, 0, &ov)) {
+ rc = ErrCode();
+ goto fail;
+ }
+ }
+ }
+ size = GetFileSize(env->me_lfd, NULL);
+#else
+ if ((env->me_lfd = open(lpath, O_RDWR|O_CREAT, mode)) == -1) {
+ rc = ErrCode();
+ return rc;
+ }
+ /* Try to get exclusive lock. If we succeed, then
+ * nobody is using the lock region and we should initialize it.
+ */
+ {
+ struct flock lock_info;
+ memset((void *)&lock_info, 0, sizeof(lock_info));
+ lock_info.l_type = F_WRLCK;
+ lock_info.l_whence = SEEK_SET;
+ lock_info.l_start = 0;
+ lock_info.l_len = 1;
+ rc = fcntl(env->me_lfd, F_SETLK, &lock_info);
+ if (rc == 0) {
+ *excl = 1;
+ } else {
+ lock_info.l_type = F_RDLCK;
+ rc = fcntl(env->me_lfd, F_SETLKW, &lock_info);
+ if (rc) {
+ rc = ErrCode();
+ goto fail;
+ }
+ }
+ }
+ size = lseek(env->me_lfd, 0, SEEK_END);
+#endif
+ rsize = (env->me_maxreaders-1) * sizeof(MDB_reader) + sizeof(MDB_txninfo);
+ if (size < rsize && *excl) {
+#ifdef _WIN32
+ SetFilePointer(env->me_lfd, rsize, NULL, 0);
+ if (!SetEndOfFile(env->me_lfd)) {
+ rc = ErrCode();
+ goto fail;
+ }
+#else
+ if (ftruncate(env->me_lfd, rsize) != 0) {
+ rc = ErrCode();
+ goto fail;
+ }
+#endif
+ } else {
+ rsize = size;
+ size = rsize - sizeof(MDB_txninfo);
+ env->me_maxreaders = size/sizeof(MDB_reader) + 1;
+ }
+#ifdef _WIN32
+ {
+ HANDLE mh;
+ mh = CreateFileMapping(env->me_lfd, NULL, PAGE_READWRITE,
+ 0, 0, NULL);
+ if (!mh) {
+ rc = ErrCode();
+ goto fail;
+ }
+ env->me_txns = MapViewOfFileEx(mh, FILE_MAP_WRITE, 0, 0, rsize, NULL);
+ CloseHandle(mh);
+ if (!env->me_txns) {
+ rc = ErrCode();
+ goto fail;
+ }
+ }
+#else
+ env->me_txns = mmap(0, rsize, PROT_READ|PROT_WRITE, MAP_SHARED,
+ env->me_lfd, 0);
+ if (env->me_txns == MAP_FAILED) {
+ rc = ErrCode();
+ goto fail;
+ }
+#endif
+ if (*excl) {
+#ifdef _WIN32
+ char hexbuf[17];
+ if (!mdb_sec_inited) {
+ InitializeSecurityDescriptor(&mdb_null_sd,
+ SECURITY_DESCRIPTOR_REVISION);
+ SetSecurityDescriptorDacl(&mdb_null_sd, TRUE, 0, FALSE);
+ mdb_all_sa.nLength = sizeof(SECURITY_ATTRIBUTES);
+ mdb_all_sa.bInheritHandle = FALSE;
+ mdb_all_sa.lpSecurityDescriptor = &mdb_null_sd;
+ mdb_sec_inited = 1;
+ }
+ mdb_hash_hex(lpath, hexbuf);
+ sprintf(env->me_txns->mti_rmname, "Global\\MDBr%s", hexbuf);
+ env->me_rmutex = CreateMutex(&mdb_all_sa, FALSE, env->me_txns->mti_rmname);
+ if (!env->me_rmutex) {
+ rc = ErrCode();
+ goto fail;
+ }
+ sprintf(env->me_txns->mti_wmname, "Global\\MDBw%s", hexbuf);
+ env->me_wmutex = CreateMutex(&mdb_all_sa, FALSE, env->me_txns->mti_wmname);
+ if (!env->me_wmutex) {
+ rc = ErrCode();
+ goto fail;
+ }
+#else /* _WIN32 */
+#ifdef __APPLE__
+ char hexbuf[17];
+ mdb_hash_hex(lpath, hexbuf);
+ sprintf(env->me_txns->mti_rmname, "MDBr%s", hexbuf);
+ if (sem_unlink(env->me_txns->mti_rmname)) {
+ rc = ErrCode();
+ if (rc != ENOENT && rc != EINVAL)
+ goto fail;
+ }
+ env->me_rmutex = sem_open(env->me_txns->mti_rmname, O_CREAT, mode, 1);
+ if (!env->me_rmutex) {
+ rc = ErrCode();
+ goto fail;
+ }
+ sprintf(env->me_txns->mti_wmname, "MDBw%s", hexbuf);
+ if (sem_unlink(env->me_txns->mti_wmname)) {
+ rc = ErrCode();
+ if (rc != ENOENT && rc != EINVAL)
+ goto fail;
+ }
+ env->me_wmutex = sem_open(env->me_txns->mti_wmname, O_CREAT, mode, 1);
+ if (!env->me_wmutex) {
+ rc = ErrCode();
+ goto fail;
+ }
+#else /* __APPLE__ */
+ pthread_mutexattr_t mattr;
+
+ pthread_mutexattr_init(&mattr);
+ rc = pthread_mutexattr_setpshared(&mattr, PTHREAD_PROCESS_SHARED);
+ if (rc) {
+ goto fail;
+ }
+ pthread_mutex_init(&env->me_txns->mti_mutex, &mattr);
+ pthread_mutex_init(&env->me_txns->mti_wmutex, &mattr);
+#endif /* __APPLE__ */
+#endif /* _WIN32 */
+ env->me_txns->mti_version = MDB_VERSION;
+ env->me_txns->mti_magic = MDB_MAGIC;
+ env->me_txns->mti_txnid = 0;
+ env->me_txns->mti_numreaders = 0;
+ env->me_txns->mti_me_toggle = 0;
+
+ } else {
+ if (env->me_txns->mti_magic != MDB_MAGIC) {
+ DPUTS("lock region has invalid magic");
+ rc = EINVAL;
+ goto fail;
+ }
+ if (env->me_txns->mti_version != MDB_VERSION) {
+ DPRINTF("lock region is version %u, expected version %u",
+ env->me_txns->mti_version, MDB_VERSION);
+ rc = MDB_VERSION_MISMATCH;
+ goto fail;
+ }
+ rc = ErrCode();
+ if (rc != EACCES && rc != EAGAIN) {
+ goto fail;
+ }
+#ifdef _WIN32
+ env->me_rmutex = OpenMutex(SYNCHRONIZE, FALSE, env->me_txns->mti_rmname);
+ if (!env->me_rmutex) {
+ rc = ErrCode();
+ goto fail;
+ }
+ env->me_wmutex = OpenMutex(SYNCHRONIZE, FALSE, env->me_txns->mti_wmname);
+ if (!env->me_wmutex) {
+ rc = ErrCode();
+ goto fail;
+ }
+#endif
+#ifdef __APPLE__
+ env->me_rmutex = sem_open(env->me_txns->mti_rmname, 0);
+ if (!env->me_rmutex) {
+ rc = ErrCode();
+ goto fail;
+ }
+ env->me_wmutex = sem_open(env->me_txns->mti_wmname, 0);
+ if (!env->me_wmutex) {
+ rc = ErrCode();
+ goto fail;
+ }
+#endif
+ }
+ return MDB_SUCCESS;
+
+fail:
+ close(env->me_lfd);
+ env->me_lfd = INVALID_HANDLE_VALUE;
+ return rc;
+
+}
+
+ /** The name of the lock file in the DB environment */
+#define LOCKNAME "/lock.mdb"
+ /** The name of the data file in the DB environment */
+#define DATANAME "/data.mdb"
+ /** The suffix of the lock file when no subdir is used */
+#define LOCKSUFF "-lock"
+
+int
+mdb_env_open(MDB_env *env, const char *path, unsigned int flags, mode_t mode)
+{
+ int oflags, rc, len, excl;
+ char *lpath, *dpath;
+
+ len = strlen(path);
+ if (flags & MDB_NOSUBDIR) {
+ rc = len + sizeof(LOCKSUFF) + len + 1;
+ } else {
+ rc = len + sizeof(LOCKNAME) + len + sizeof(DATANAME);
+ }
+ lpath = malloc(rc);
+ if (!lpath)
+ return ENOMEM;
+ if (flags & MDB_NOSUBDIR) {
+ dpath = lpath + len + sizeof(LOCKSUFF);
+ sprintf(lpath, "%s" LOCKSUFF, path);
+ strcpy(dpath, path);
+ } else {
+ dpath = lpath + len + sizeof(LOCKNAME);
+ sprintf(lpath, "%s" LOCKNAME, path);
+ sprintf(dpath, "%s" DATANAME, path);
+ }
+
+ rc = mdb_env_setup_locks(env, lpath, mode, &excl);
+ if (rc)
+ goto leave;
+
+#ifdef _WIN32
+ if (F_ISSET(flags, MDB_RDONLY)) {
+ oflags = GENERIC_READ;
+ len = OPEN_EXISTING;
+ } else {
+ oflags = GENERIC_READ|GENERIC_WRITE;
+ len = OPEN_ALWAYS;
+ }
+ mode = FILE_ATTRIBUTE_NORMAL;
+ if ((env->me_fd = CreateFile(dpath, oflags, FILE_SHARE_READ|FILE_SHARE_WRITE,
+ NULL, len, mode, NULL)) == INVALID_HANDLE_VALUE) {
+ rc = ErrCode();
+ goto leave;
+ }
+#else
+ if (F_ISSET(flags, MDB_RDONLY))
+ oflags = O_RDONLY;
+ else
+ oflags = O_RDWR | O_CREAT;
+
+ if ((env->me_fd = open(dpath, oflags, mode)) == -1) {
+ rc = ErrCode();
+ goto leave;
+ }
+#endif
+
+ if ((rc = mdb_env_open2(env, flags)) == MDB_SUCCESS) {
+ /* synchronous fd for meta writes */
+#ifdef _WIN32
+ if (!(flags & (MDB_RDONLY|MDB_NOSYNC)))
+ mode |= FILE_FLAG_WRITE_THROUGH;
+ if ((env->me_mfd = CreateFile(dpath, oflags, FILE_SHARE_READ|FILE_SHARE_WRITE,
+ NULL, len, mode, NULL)) == INVALID_HANDLE_VALUE) {
+ rc = ErrCode();
+ goto leave;
+ }
+#else
+ if (!(flags & (MDB_RDONLY|MDB_NOSYNC)))
+ oflags |= MDB_DSYNC;
+ if ((env->me_mfd = open(dpath, oflags, mode)) == -1) {
+ rc = ErrCode();
+ goto leave;
+ }
+#endif
+ env->me_path = strdup(path);
+ DPRINTF("opened dbenv %p", (void *) env);
+ pthread_key_create(&env->me_txkey, mdb_env_reader_dest);
+ LAZY_RWLOCK_INIT(&env->me_dblock, NULL);
+ if (excl)
+ mdb_env_share_locks(env);
+ env->me_dbxs = calloc(env->me_maxdbs, sizeof(MDB_dbx));
+ env->me_dbs[0] = calloc(env->me_maxdbs, sizeof(MDB_db));
+ env->me_dbs[1] = calloc(env->me_maxdbs, sizeof(MDB_db));
+ env->me_numdbs = 2;
+ }
+
+leave:
+ if (rc) {
+ if (env->me_fd != INVALID_HANDLE_VALUE) {
+ close(env->me_fd);
+ env->me_fd = INVALID_HANDLE_VALUE;
+ }
+ if (env->me_lfd != INVALID_HANDLE_VALUE) {
+ close(env->me_lfd);
+ env->me_lfd = INVALID_HANDLE_VALUE;
+ }
+ }
+ free(lpath);
+ return rc;
+}
+
+void
+mdb_env_close(MDB_env *env)
+{
+ MDB_page *dp;
+
+ if (env == NULL)
+ return;
+
+ while (env->me_dpages) {
+ dp = env->me_dpages;
+ env->me_dpages = dp->mp_next;
+ free(dp);
+ }
+
+ free(env->me_dbs[1]);
+ free(env->me_dbs[0]);
+ free(env->me_dbxs);
+ free(env->me_path);
+
+ LAZY_RWLOCK_DESTROY(&env->me_dblock);
+ pthread_key_delete(env->me_txkey);
+
+ if (env->me_map) {
+ munmap(env->me_map, env->me_mapsize);
+ }
+ close(env->me_mfd);
+ close(env->me_fd);
+ if (env->me_txns) {
+ pid_t pid = getpid();
+ unsigned int i;
+ for (i=0; i<env->me_txns->mti_numreaders; i++)
+ if (env->me_txns->mti_readers[i].mr_pid == pid)
+ env->me_txns->mti_readers[i].mr_pid = 0;
+ munmap(env->me_txns, (env->me_maxreaders-1)*sizeof(MDB_reader)+sizeof(MDB_txninfo));
+ }
+ close(env->me_lfd);
+ mdb_midl_free(env->me_free_pgs);
+ free(env);
+}
+
+/** Compare two items pointing at aligned size_t's */
+static int
+mdb_cmp_long(const MDB_val *a, const MDB_val *b)
+{
+ return (*(size_t *)a->mv_data < *(size_t *)b->mv_data) ? -1 :
+ *(size_t *)a->mv_data > *(size_t *)b->mv_data;
+}
+
+/** Compare two items pointing at aligned int's */
+static int
+mdb_cmp_int(const MDB_val *a, const MDB_val *b)
+{
+ return (*(unsigned int *)a->mv_data < *(unsigned int *)b->mv_data) ? -1 :
+ *(unsigned int *)a->mv_data > *(unsigned int *)b->mv_data;
+}
+
+/** Compare two items pointing at ints of unknown alignment.
+ * Nodes and keys are guaranteed to be 2-byte aligned.
+ */
+static int
+mdb_cmp_cint(const MDB_val *a, const MDB_val *b)
+{
+#if BYTE_ORDER == LITTLE_ENDIAN
+ unsigned short *u, *c;
+ int x;
+
+ u = (unsigned short *) ((char *) a->mv_data + a->mv_size);
+ c = (unsigned short *) ((char *) b->mv_data + a->mv_size);
+ do {
+ x = *--u - *--c;
+ } while(!x && u > (unsigned short *)a->mv_data);
+ return x;
+#else
+ return memcmp(a->mv_data, b->mv_data, a->mv_size);
+#endif
+}
+
+/** Compare two items lexically */
+static int
+mdb_cmp_memn(const MDB_val *a, const MDB_val *b)
+{
+ int diff;
+ ssize_t len_diff;
+ unsigned int len;
+
+ len = a->mv_size;
+ len_diff = (ssize_t) a->mv_size - (ssize_t) b->mv_size;
+ if (len_diff > 0) {
+ len = b->mv_size;
+ len_diff = 1;
+ }
+
+ diff = memcmp(a->mv_data, b->mv_data, len);
+ return diff ? diff : len_diff<0 ? -1 : len_diff;
+}
+
+/** Compare two items in reverse byte order */
+static int
+mdb_cmp_memnr(const MDB_val *a, const MDB_val *b)
+{
+ const unsigned char *p1, *p2, *p1_lim;
+ ssize_t len_diff;
+ int diff;
+
+ p1_lim = (const unsigned char *)a->mv_data;
+ p1 = (const unsigned char *)a->mv_data + a->mv_size;
+ p2 = (const unsigned char *)b->mv_data + b->mv_size;
+
+ len_diff = (ssize_t) a->mv_size - (ssize_t) b->mv_size;
+ if (len_diff > 0) {
+ p1_lim += len_diff;
+ len_diff = 1;
+ }
+
+ while (p1 > p1_lim) {
+ diff = *--p1 - *--p2;
+ if (diff)
+ return diff;
+ }
+ return len_diff<0 ? -1 : len_diff;
+}
+
+/** Search for key within a page, using binary search.
+ * Returns the smallest entry larger or equal to the key.
+ * If exactp is non-null, stores whether the found entry was an exact match
+ * in *exactp (1 or 0).
+ * Updates the cursor index with the index of the found entry.
+ * If no entry larger or equal to the key is found, returns NULL.
+ */
+static MDB_node *
+mdb_node_search(MDB_cursor *mc, MDB_val *key, int *exactp)
+{
+ unsigned int i = 0, nkeys;
+ int low, high;
+ int rc = 0;
+ MDB_page *mp = mc->mc_pg[mc->mc_top];
+ MDB_node *node = NULL;
+ MDB_val nodekey;
+ MDB_cmp_func *cmp;
+ DKBUF;
+
+ nkeys = NUMKEYS(mp);
+
+ DPRINTF("searching %u keys in %s %spage %zu",
+ nkeys, IS_LEAF(mp) ? "leaf" : "branch", IS_SUBP(mp) ? "sub-" : "",
+ mp->mp_pgno);
+
+ assert(nkeys > 0);
+
+ low = IS_LEAF(mp) ? 0 : 1;
+ high = nkeys - 1;
+ cmp = mc->mc_dbx->md_cmp;
+
+ /* Branch pages have no data, so if using integer keys,
+ * alignment is guaranteed. Use faster mdb_cmp_int.
+ */
+ if (cmp == mdb_cmp_cint && IS_BRANCH(mp)) {
+ if (NODEPTR(mp, 1)->mn_ksize == sizeof(size_t))
+ cmp = mdb_cmp_long;
+ else
+ cmp = mdb_cmp_int;
+ }
+
+ if (IS_LEAF2(mp)) {
+ nodekey.mv_size = mc->mc_db->md_pad;
+ node = NODEPTR(mp, 0); /* fake */
+ while (low <= high) {
+ i = (low + high) >> 1;
+ nodekey.mv_data = LEAF2KEY(mp, i, nodekey.mv_size);
+ rc = cmp(key, &nodekey);
+ DPRINTF("found leaf index %u [%s], rc = %i",
+ i, DKEY(&nodekey), rc);
+ if (rc == 0)
+ break;
+ if (rc > 0)
+ low = i + 1;
+ else
+ high = i - 1;
+ }
+ } else {
+ while (low <= high) {
+ i = (low + high) >> 1;
+
+ node = NODEPTR(mp, i);
+ nodekey.mv_size = NODEKSZ(node);
+ nodekey.mv_data = NODEKEY(node);
+
+ rc = cmp(key, &nodekey);
+#if DEBUG
+ if (IS_LEAF(mp))
+ DPRINTF("found leaf index %u [%s], rc = %i",
+ i, DKEY(&nodekey), rc);
+ else
+ DPRINTF("found branch index %u [%s -> %zu], rc = %i",
+ i, DKEY(&nodekey), NODEPGNO(node), rc);
+#endif
+ if (rc == 0)
+ break;
+ if (rc > 0)
+ low = i + 1;
+ else
+ high = i - 1;
+ }
+ }
+
+ if (rc > 0) { /* Found entry is less than the key. */
+ i++; /* Skip to get the smallest entry larger than key. */
+ if (!IS_LEAF2(mp))
+ node = NODEPTR(mp, i);
+ }
+ if (exactp)
+ *exactp = (rc == 0);
+ /* store the key index */
+ mc->mc_ki[mc->mc_top] = i;
+ if (i >= nkeys)
+ /* There is no entry larger or equal to the key. */
+ return NULL;
+
+ /* nodeptr is fake for LEAF2 */
+ return node;
+}
+
+#if 0
+static void
+mdb_cursor_adjust(MDB_cursor *mc, func)
+{
+ MDB_cursor *m2;
+
+ for (m2 = mc->mc_txn->mt_cursors[mc->mc_dbi]; m2; m2=m2->mc_next) {
+ if (m2->mc_pg[m2->mc_top] == mc->mc_pg[mc->mc_top]) {
+ func(mc, m2);
+ }
+ }
+}
+#endif
+
+/** Pop a page off the top of the cursor's stack. */
+static void
+mdb_cursor_pop(MDB_cursor *mc)
+{
+ MDB_page *top;
+
+ if (mc->mc_snum) {
+ top = mc->mc_pg[mc->mc_top];
+ mc->mc_snum--;
+ if (mc->mc_snum)
+ mc->mc_top--;
+
+ DPRINTF("popped page %zu off db %u cursor %p", top->mp_pgno,
+ mc->mc_dbi, (void *) mc);
+ }
+}
+
+/** Push a page onto the top of the cursor's stack. */
+static int
+mdb_cursor_push(MDB_cursor *mc, MDB_page *mp)
+{
+ DPRINTF("pushing page %zu on db %u cursor %p", mp->mp_pgno,
+ mc->mc_dbi, (void *) mc);
+
+ if (mc->mc_snum >= CURSOR_STACK) {
+ assert(mc->mc_snum < CURSOR_STACK);
+ return ENOMEM;
+ }
+
+ mc->mc_top = mc->mc_snum++;
+ mc->mc_pg[mc->mc_top] = mp;
+ mc->mc_ki[mc->mc_top] = 0;
+
+ return MDB_SUCCESS;
+}
+
+/** Find the address of the page corresponding to a given page number.
+ * @param[in] txn the transaction for this access.
+ * @param[in] pgno the page number for the page to retrieve.
+ * @param[out] ret address of a pointer where the page's address will be stored.
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_page_get(MDB_txn *txn, pgno_t pgno, MDB_page **ret)
+{
+ MDB_page *p = NULL;
+
+ if (!F_ISSET(txn->mt_flags, MDB_TXN_RDONLY) && txn->mt_u.dirty_list[0].mid) {
+ unsigned x;
+ x = mdb_mid2l_search(txn->mt_u.dirty_list, pgno);
+ if (x <= txn->mt_u.dirty_list[0].mid && txn->mt_u.dirty_list[x].mid == pgno) {
+ p = txn->mt_u.dirty_list[x].mptr;
+ }
+ }
+ if (!p) {
+ if (pgno <= txn->mt_env->me_metas[txn->mt_toggle]->mm_last_pg)
+ p = (MDB_page *)(txn->mt_env->me_map + txn->mt_env->me_psize * pgno);
+ }
+ *ret = p;
+ if (!p) {
+ DPRINTF("page %zu not found", pgno);
+ assert(p != NULL);
+ }
+ return (p != NULL) ? MDB_SUCCESS : MDB_PAGE_NOTFOUND;
+}
+
+/** Search for the page a given key should be in.
+ * Pushes parent pages on the cursor stack. This function continues a
+ * search on a cursor that has already been initialized. (Usually by
+ * #mdb_page_search() but also by #mdb_node_move().)
+ * @param[in,out] mc the cursor for this operation.
+ * @param[in] key the key to search for. If NULL, search for the lowest
+ * page. (This is used by #mdb_cursor_first().)
+ * @param[in] modify If true, visited pages are updated with new page numbers.
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_page_search_root(MDB_cursor *mc, MDB_val *key, int modify)
+{
+ MDB_page *mp = mc->mc_pg[mc->mc_top];
+ DKBUF;
+ int rc;
+
+
+ while (IS_BRANCH(mp)) {
+ MDB_node *node;
+ indx_t i;
+
+ DPRINTF("branch page %zu has %u keys", mp->mp_pgno, NUMKEYS(mp));
+ assert(NUMKEYS(mp) > 1);
+ DPRINTF("found index 0 to page %zu", NODEPGNO(NODEPTR(mp, 0)));
+
+ if (key == NULL) /* Initialize cursor to first page. */
+ i = 0;
+ else if (key->mv_size > MAXKEYSIZE && key->mv_data == NULL) {
+ /* cursor to last page */
+ i = NUMKEYS(mp)-1;
+ } else {
+ int exact;
+ node = mdb_node_search(mc, key, &exact);
+ if (node == NULL)
+ i = NUMKEYS(mp) - 1;
+ else {
+ i = mc->mc_ki[mc->mc_top];
+ if (!exact) {
+ assert(i > 0);
+ i--;
+ }
+ }
+ }
+
+ if (key)
+ DPRINTF("following index %u for key [%s]",
+ i, DKEY(key));
+ assert(i < NUMKEYS(mp));
+ node = NODEPTR(mp, i);
+
+ if ((rc = mdb_page_get(mc->mc_txn, NODEPGNO(node), &mp)))
+ return rc;
+
+ mc->mc_ki[mc->mc_top] = i;
+ if ((rc = mdb_cursor_push(mc, mp)))
+ return rc;
+
+ if (modify) {
+ if ((rc = mdb_page_touch(mc)) != 0)
+ return rc;
+ mp = mc->mc_pg[mc->mc_top];
+ }
+ }
+
+ if (!IS_LEAF(mp)) {
+ DPRINTF("internal error, index points to a %02X page!?",
+ mp->mp_flags);
+ return MDB_CORRUPTED;
+ }
+
+ DPRINTF("found leaf page %zu for key [%s]", mp->mp_pgno,
+ key ? DKEY(key) : NULL);
+
+ return MDB_SUCCESS;
+}
+
+/** Search for the page a given key should be in.
+ * Pushes parent pages on the cursor stack. This function just sets up
+ * the search; it finds the root page for \b mc's database and sets this
+ * as the root of the cursor's stack. Then #mdb_page_search_root() is
+ * called to complete the search.
+ * @param[in,out] mc the cursor for this operation.
+ * @param[in] key the key to search for. If NULL, search for the lowest
+ * page. (This is used by #mdb_cursor_first().)
+ * @param[in] modify If true, visited pages are updated with new page numbers.
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_page_search(MDB_cursor *mc, MDB_val *key, int modify)
+{
+ int rc;
+ pgno_t root;
+
+ /* Make sure the txn is still viable, then find the root from
+ * the txn's db table.
+ */
+ if (F_ISSET(mc->mc_txn->mt_flags, MDB_TXN_ERROR)) {
+ DPUTS("transaction has failed, must abort");
+ return EINVAL;
+ } else {
+ /* Make sure we're using an up-to-date root */
+ if (mc->mc_dbi > MAIN_DBI) {
+ if ((*mc->mc_dbflag & DB_STALE) ||
+ (modify && !(*mc->mc_dbflag & DB_DIRTY))) {
+ MDB_cursor mc2;
+ unsigned char dbflag = 0;
+ mdb_cursor_init(&mc2, mc->mc_txn, MAIN_DBI, NULL);
+ rc = mdb_page_search(&mc2, &mc->mc_dbx->md_name, modify);
+ if (rc)
+ return rc;
+ if (*mc->mc_dbflag & DB_STALE) {
+ MDB_val data;
+ int exact = 0;
+ MDB_node *leaf = mdb_node_search(&mc2,
+ &mc->mc_dbx->md_name, &exact);
+ if (!exact)
+ return MDB_NOTFOUND;
+ mdb_node_read(mc->mc_txn, leaf, &data);
+ memcpy(mc->mc_db, data.mv_data, sizeof(MDB_db));
+ }
+ if (modify)
+ dbflag = DB_DIRTY;
+ *mc->mc_dbflag = dbflag;
+ }
+ }
+ root = mc->mc_db->md_root;
+
+ if (root == P_INVALID) { /* Tree is empty. */
+ DPUTS("tree is empty");
+ return MDB_NOTFOUND;
+ }
+ }
+
+ assert(root > 1);
+ if ((rc = mdb_page_get(mc->mc_txn, root, &mc->mc_pg[0])))
+ return rc;
+
+ mc->mc_snum = 1;
+ mc->mc_top = 0;
+
+ DPRINTF("db %u root page %zu has flags 0x%X",
+ mc->mc_dbi, root, mc->mc_pg[0]->mp_flags);
+
+ if (modify) {
+ if ((rc = mdb_page_touch(mc)))
+ return rc;
+ }
+
+ return mdb_page_search_root(mc, key, modify);
+}
+
+/** Return the data associated with a given node.
+ * @param[in] txn The transaction for this operation.
+ * @param[in] leaf The node being read.
+ * @param[out] data Updated to point to the node's data.
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_node_read(MDB_txn *txn, MDB_node *leaf, MDB_val *data)
+{
+ MDB_page *omp; /* overflow page */
+ pgno_t pgno;
+ int rc;
+
+ if (!F_ISSET(leaf->mn_flags, F_BIGDATA)) {
+ data->mv_size = NODEDSZ(leaf);
+ data->mv_data = NODEDATA(leaf);
+ return MDB_SUCCESS;
+ }
+
+ /* Read overflow data.
+ */
+ data->mv_size = NODEDSZ(leaf);
+ memcpy(&pgno, NODEDATA(leaf), sizeof(pgno));
+ if ((rc = mdb_page_get(txn, pgno, &omp))) {
+ DPRINTF("read overflow page %zu failed", pgno);
+ return rc;
+ }
+ data->mv_data = METADATA(omp);
+
+ return MDB_SUCCESS;
+}
+
+int
+mdb_get(MDB_txn *txn, MDB_dbi dbi,
+ MDB_val *key, MDB_val *data)
+{
+ MDB_cursor mc;
+ MDB_xcursor mx;
+ int exact = 0;
+ DKBUF;
+
+ assert(key);
+ assert(data);
+ DPRINTF("===> get db %u key [%s]", dbi, DKEY(key));
+
+ if (txn == NULL || !dbi || dbi >= txn->mt_numdbs)
+ return EINVAL;
+
+ if (key->mv_size == 0 || key->mv_size > MAXKEYSIZE) {
+ return EINVAL;
+ }
+
+ mdb_cursor_init(&mc, txn, dbi, &mx);
+ return mdb_cursor_set(&mc, key, data, MDB_SET, &exact);
+}
+
+/** Find a sibling for a page.
+ * Replaces the page at the top of the cursor's stack with the
+ * specified sibling, if one exists.
+ * @param[in] mc The cursor for this operation.
+ * @param[in] move_right Non-zero if the right sibling is requested,
+ * otherwise the left sibling.
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_cursor_sibling(MDB_cursor *mc, int move_right)
+{
+ int rc;
+ MDB_node *indx;
+ MDB_page *mp;
+
+ if (mc->mc_snum < 2) {
+ return MDB_NOTFOUND; /* root has no siblings */
+ }
+
+ mdb_cursor_pop(mc);
+ DPRINTF("parent page is page %zu, index %u",
+ mc->mc_pg[mc->mc_top]->mp_pgno, mc->mc_ki[mc->mc_top]);
+
+ if (move_right ? (mc->mc_ki[mc->mc_top] + 1u >= NUMKEYS(mc->mc_pg[mc->mc_top]))
+ : (mc->mc_ki[mc->mc_top] == 0)) {
+ DPRINTF("no more keys left, moving to %s sibling",
+ move_right ? "right" : "left");
+ if ((rc = mdb_cursor_sibling(mc, move_right)) != MDB_SUCCESS)
+ return rc;
+ } else {
+ if (move_right)
+ mc->mc_ki[mc->mc_top]++;
+ else
+ mc->mc_ki[mc->mc_top]--;
+ DPRINTF("just moving to %s index key %u",
+ move_right ? "right" : "left", mc->mc_ki[mc->mc_top]);
+ }
+ assert(IS_BRANCH(mc->mc_pg[mc->mc_top]));
+
+ indx = NODEPTR(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top]);
+ if ((rc = mdb_page_get(mc->mc_txn, NODEPGNO(indx), &mp)))
+ return rc;;
+
+ mdb_cursor_push(mc, mp);
+
+ return MDB_SUCCESS;
+}
+
+/** Move the cursor to the next data item. */
+static int
+mdb_cursor_next(MDB_cursor *mc, MDB_val *key, MDB_val *data, MDB_cursor_op op)
+{
+ MDB_page *mp;
+ MDB_node *leaf;
+ int rc;
+
+ if (mc->mc_flags & C_EOF) {
+ return MDB_NOTFOUND;
+ }
+
+ assert(mc->mc_flags & C_INITIALIZED);
+
+ mp = mc->mc_pg[mc->mc_top];
+
+ if (mc->mc_db->md_flags & MDB_DUPSORT) {
+ leaf = NODEPTR(mp, mc->mc_ki[mc->mc_top]);
+ if (F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ if (op == MDB_NEXT || op == MDB_NEXT_DUP) {
+ rc = mdb_cursor_next(&mc->mc_xcursor->mx_cursor, data, NULL, MDB_NEXT);
+ if (op != MDB_NEXT || rc == MDB_SUCCESS)
+ return rc;
+ }
+ } else {
+ mc->mc_xcursor->mx_cursor.mc_flags &= ~C_INITIALIZED;
+ if (op == MDB_NEXT_DUP)
+ return MDB_NOTFOUND;
+ }
+ }
+
+ DPRINTF("cursor_next: top page is %zu in cursor %p", mp->mp_pgno, (void *) mc);
+
+ if (mc->mc_ki[mc->mc_top] + 1u >= NUMKEYS(mp)) {
+ DPUTS("=====> move to next sibling page");
+ if (mdb_cursor_sibling(mc, 1) != MDB_SUCCESS) {
+ mc->mc_flags |= C_EOF;
+ mc->mc_flags &= ~C_INITIALIZED;
+ return MDB_NOTFOUND;
+ }
+ mp = mc->mc_pg[mc->mc_top];
+ DPRINTF("next page is %zu, key index %u", mp->mp_pgno, mc->mc_ki[mc->mc_top]);
+ } else
+ mc->mc_ki[mc->mc_top]++;
+
+ DPRINTF("==> cursor points to page %zu with %u keys, key index %u",
+ mp->mp_pgno, NUMKEYS(mp), mc->mc_ki[mc->mc_top]);
+
+ if (IS_LEAF2(mp)) {
+ key->mv_size = mc->mc_db->md_pad;
+ key->mv_data = LEAF2KEY(mp, mc->mc_ki[mc->mc_top], key->mv_size);
+ return MDB_SUCCESS;
+ }
+
+ assert(IS_LEAF(mp));
+ leaf = NODEPTR(mp, mc->mc_ki[mc->mc_top]);
+
+ if (F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ mdb_xcursor_init1(mc, leaf);
+ }
+ if (data) {
+ if ((rc = mdb_node_read(mc->mc_txn, leaf, data) != MDB_SUCCESS))
+ return rc;
+
+ if (F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ rc = mdb_cursor_first(&mc->mc_xcursor->mx_cursor, data, NULL);
+ if (rc != MDB_SUCCESS)
+ return rc;
+ }
+ }
+
+ MDB_SET_KEY(leaf, key);
+ return MDB_SUCCESS;
+}
+
+/** Move the cursor to the previous data item. */
+static int
+mdb_cursor_prev(MDB_cursor *mc, MDB_val *key, MDB_val *data, MDB_cursor_op op)
+{
+ MDB_page *mp;
+ MDB_node *leaf;
+ int rc;
+
+ assert(mc->mc_flags & C_INITIALIZED);
+
+ mp = mc->mc_pg[mc->mc_top];
+
+ if (mc->mc_db->md_flags & MDB_DUPSORT) {
+ leaf = NODEPTR(mp, mc->mc_ki[mc->mc_top]);
+ if (op == MDB_PREV || op == MDB_PREV_DUP) {
+ if (F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ rc = mdb_cursor_prev(&mc->mc_xcursor->mx_cursor, data, NULL, MDB_PREV);
+ if (op != MDB_PREV || rc == MDB_SUCCESS)
+ return rc;
+ } else {
+ mc->mc_xcursor->mx_cursor.mc_flags &= ~C_INITIALIZED;
+ if (op == MDB_PREV_DUP)
+ return MDB_NOTFOUND;
+ }
+ }
+ }
+
+ DPRINTF("cursor_prev: top page is %zu in cursor %p", mp->mp_pgno, (void *) mc);
+
+ if (mc->mc_ki[mc->mc_top] == 0) {
+ DPUTS("=====> move to prev sibling page");
+ if (mdb_cursor_sibling(mc, 0) != MDB_SUCCESS) {
+ mc->mc_flags &= ~C_INITIALIZED;
+ return MDB_NOTFOUND;
+ }
+ mp = mc->mc_pg[mc->mc_top];
+ mc->mc_ki[mc->mc_top] = NUMKEYS(mp) - 1;
+ DPRINTF("prev page is %zu, key index %u", mp->mp_pgno, mc->mc_ki[mc->mc_top]);
+ } else
+ mc->mc_ki[mc->mc_top]--;
+
+ mc->mc_flags &= ~C_EOF;
+
+ DPRINTF("==> cursor points to page %zu with %u keys, key index %u",
+ mp->mp_pgno, NUMKEYS(mp), mc->mc_ki[mc->mc_top]);
+
+ if (IS_LEAF2(mp)) {
+ key->mv_size = mc->mc_db->md_pad;
+ key->mv_data = LEAF2KEY(mp, mc->mc_ki[mc->mc_top], key->mv_size);
+ return MDB_SUCCESS;
+ }
+
+ assert(IS_LEAF(mp));
+ leaf = NODEPTR(mp, mc->mc_ki[mc->mc_top]);
+
+ if (F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ mdb_xcursor_init1(mc, leaf);
+ }
+ if (data) {
+ if ((rc = mdb_node_read(mc->mc_txn, leaf, data) != MDB_SUCCESS))
+ return rc;
+
+ if (F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ rc = mdb_cursor_last(&mc->mc_xcursor->mx_cursor, data, NULL);
+ if (rc != MDB_SUCCESS)
+ return rc;
+ }
+ }
+
+ MDB_SET_KEY(leaf, key);
+ return MDB_SUCCESS;
+}
+
+/** Set the cursor on a specific data item. */
+static int
+mdb_cursor_set(MDB_cursor *mc, MDB_val *key, MDB_val *data,
+ MDB_cursor_op op, int *exactp)
+{
+ int rc;
+ MDB_page *mp;
+ MDB_node *leaf;
+ DKBUF;
+
+ assert(mc);
+ assert(key);
+ assert(key->mv_size > 0);
+
+ /* See if we're already on the right page */
+ if (mc->mc_flags & C_INITIALIZED) {
+ MDB_val nodekey;
+
+ mp = mc->mc_pg[mc->mc_top];
+ if (!NUMKEYS(mp)) {
+ mc->mc_ki[mc->mc_top] = 0;
+ return MDB_NOTFOUND;
+ }
+ if (mp->mp_flags & P_LEAF2) {
+ nodekey.mv_size = mc->mc_db->md_pad;
+ nodekey.mv_data = LEAF2KEY(mp, 0, nodekey.mv_size);
+ } else {
+ leaf = NODEPTR(mp, 0);
+ MDB_SET_KEY(leaf, &nodekey);
+ }
+ rc = mc->mc_dbx->md_cmp(key, &nodekey);
+ if (rc == 0) {
+ /* Probably happens rarely, but first node on the page
+ * was the one we wanted.
+ */
+ mc->mc_ki[mc->mc_top] = 0;
+ leaf = NODEPTR(mp, 0);
+ if (exactp)
+ *exactp = 1;
+ goto set1;
+ }
+ if (rc > 0) {
+ unsigned int i;
+ unsigned int nkeys = NUMKEYS(mp);
+ if (nkeys > 1) {
+ if (mp->mp_flags & P_LEAF2) {
+ nodekey.mv_data = LEAF2KEY(mp,
+ nkeys-1, nodekey.mv_size);
+ } else {
+ leaf = NODEPTR(mp, nkeys-1);
+ MDB_SET_KEY(leaf, &nodekey);
+ }
+ rc = mc->mc_dbx->md_cmp(key, &nodekey);
+ if (rc == 0) {
+ /* last node was the one we wanted */
+ mc->mc_ki[mc->mc_top] = nkeys-1;
+ leaf = NODEPTR(mp, nkeys-1);
+ if (exactp)
+ *exactp = 1;
+ goto set1;
+ }
+ if (rc < 0) {
+ /* This is definitely the right page, skip search_page */
+ rc = 0;
+ goto set2;
+ }
+ }
+ /* If any parents have right-sibs, search.
+ * Otherwise, there's nothing further.
+ */
+ for (i=0; i<mc->mc_top; i++)
+ if (mc->mc_ki[i] <
+ NUMKEYS(mc->mc_pg[i])-1)
+ break;
+ if (i == mc->mc_top) {
+ /* There are no other pages */
+ mc->mc_ki[mc->mc_top] = nkeys;
+ return MDB_NOTFOUND;
+ }
+ }
+ if (!mc->mc_top) {
+ /* There are no other pages */
+ mc->mc_ki[mc->mc_top] = 0;
+ return MDB_NOTFOUND;
+ }
+ }
+
+ rc = mdb_page_search(mc, key, 0);
+ if (rc != MDB_SUCCESS)
+ return rc;
+
+ mp = mc->mc_pg[mc->mc_top];
+ assert(IS_LEAF(mp));
+
+set2:
+ leaf = mdb_node_search(mc, key, exactp);
+ if (exactp != NULL && !*exactp) {
+ /* MDB_SET specified and not an exact match. */
+ return MDB_NOTFOUND;
+ }
+
+ if (leaf == NULL) {
+ DPUTS("===> inexact leaf not found, goto sibling");
+ if ((rc = mdb_cursor_sibling(mc, 1)) != MDB_SUCCESS)
+ return rc; /* no entries matched */
+ mp = mc->mc_pg[mc->mc_top];
+ assert(IS_LEAF(mp));
+ leaf = NODEPTR(mp, 0);
+ }
+
+set1:
+ mc->mc_flags |= C_INITIALIZED;
+ mc->mc_flags &= ~C_EOF;
+
+ if (IS_LEAF2(mp)) {
+ key->mv_size = mc->mc_db->md_pad;
+ key->mv_data = LEAF2KEY(mp, mc->mc_ki[mc->mc_top], key->mv_size);
+ return MDB_SUCCESS;
+ }
+
+ if (F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ mdb_xcursor_init1(mc, leaf);
+ }
+ if (data) {
+ if (F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ if (op == MDB_SET || op == MDB_SET_RANGE) {
+ rc = mdb_cursor_first(&mc->mc_xcursor->mx_cursor, data, NULL);
+ } else {
+ int ex2, *ex2p;
+ if (op == MDB_GET_BOTH) {
+ ex2p = &ex2;
+ ex2 = 0;
+ } else {
+ ex2p = NULL;
+ }
+ rc = mdb_cursor_set(&mc->mc_xcursor->mx_cursor, data, NULL, MDB_SET_RANGE, ex2p);
+ if (rc != MDB_SUCCESS)
+ return rc;
+ }
+ } else if (op == MDB_GET_BOTH || op == MDB_GET_BOTH_RANGE) {
+ MDB_val d2;
+ if ((rc = mdb_node_read(mc->mc_txn, leaf, &d2)) != MDB_SUCCESS)
+ return rc;
+ rc = mc->mc_dbx->md_dcmp(data, &d2);
+ if (rc) {
+ if (op == MDB_GET_BOTH || rc > 0)
+ return MDB_NOTFOUND;
+ }
+
+ } else {
+ if (mc->mc_xcursor)
+ mc->mc_xcursor->mx_cursor.mc_flags &= ~C_INITIALIZED;
+ if ((rc = mdb_node_read(mc->mc_txn, leaf, data)) != MDB_SUCCESS)
+ return rc;
+ }
+ }
+
+ /* The key already matches in all other cases */
+ if (op == MDB_SET_RANGE)
+ MDB_SET_KEY(leaf, key);
+ DPRINTF("==> cursor placed on key [%s]", DKEY(key));
+
+ return rc;
+}
+
+/** Move the cursor to the first item in the database. */
+static int
+mdb_cursor_first(MDB_cursor *mc, MDB_val *key, MDB_val *data)
+{
+ int rc;
+ MDB_node *leaf;
+
+ if (!(mc->mc_flags & C_INITIALIZED) || mc->mc_top) {
+ rc = mdb_page_search(mc, NULL, 0);
+ if (rc != MDB_SUCCESS)
+ return rc;
+ }
+ assert(IS_LEAF(mc->mc_pg[mc->mc_top]));
+
+ leaf = NODEPTR(mc->mc_pg[mc->mc_top], 0);
+ mc->mc_flags |= C_INITIALIZED;
+ mc->mc_flags &= ~C_EOF;
+
+ mc->mc_ki[mc->mc_top] = 0;
+
+ if (IS_LEAF2(mc->mc_pg[mc->mc_top])) {
+ key->mv_size = mc->mc_db->md_pad;
+ key->mv_data = LEAF2KEY(mc->mc_pg[mc->mc_top], 0, key->mv_size);
+ return MDB_SUCCESS;
+ }
+
+ if (data) {
+ if (F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ mdb_xcursor_init1(mc, leaf);
+ rc = mdb_cursor_first(&mc->mc_xcursor->mx_cursor, data, NULL);
+ if (rc)
+ return rc;
+ } else {
+ if (mc->mc_xcursor)
+ mc->mc_xcursor->mx_cursor.mc_flags &= ~C_INITIALIZED;
+ if ((rc = mdb_node_read(mc->mc_txn, leaf, data)) != MDB_SUCCESS)
+ return rc;
+ }
+ }
+ MDB_SET_KEY(leaf, key);
+ return MDB_SUCCESS;
+}
+
+/** Move the cursor to the last item in the database. */
+static int
+mdb_cursor_last(MDB_cursor *mc, MDB_val *key, MDB_val *data)
+{
+ int rc;
+ MDB_node *leaf;
+ MDB_val lkey;
+
+ lkey.mv_size = MAXKEYSIZE+1;
+ lkey.mv_data = NULL;
+
+ if (!(mc->mc_flags & C_INITIALIZED) || mc->mc_top) {
+ rc = mdb_page_search(mc, &lkey, 0);
+ if (rc != MDB_SUCCESS)
+ return rc;
+ }
+ assert(IS_LEAF(mc->mc_pg[mc->mc_top]));
+
+ leaf = NODEPTR(mc->mc_pg[mc->mc_top], NUMKEYS(mc->mc_pg[mc->mc_top])-1);
+ mc->mc_flags |= C_INITIALIZED;
+ mc->mc_flags &= ~C_EOF;
+
+ mc->mc_ki[mc->mc_top] = NUMKEYS(mc->mc_pg[mc->mc_top]) - 1;
+
+ if (IS_LEAF2(mc->mc_pg[mc->mc_top])) {
+ key->mv_size = mc->mc_db->md_pad;
+ key->mv_data = LEAF2KEY(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top], key->mv_size);
+ return MDB_SUCCESS;
+ }
+
+ if (data) {
+ if (F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ mdb_xcursor_init1(mc, leaf);
+ rc = mdb_cursor_last(&mc->mc_xcursor->mx_cursor, data, NULL);
+ if (rc)
+ return rc;
+ } else {
+ if (mc->mc_xcursor)
+ mc->mc_xcursor->mx_cursor.mc_flags &= ~C_INITIALIZED;
+ if ((rc = mdb_node_read(mc->mc_txn, leaf, data)) != MDB_SUCCESS)
+ return rc;
+ }
+ }
+
+ MDB_SET_KEY(leaf, key);
+ return MDB_SUCCESS;
+}
+
+int
+mdb_cursor_get(MDB_cursor *mc, MDB_val *key, MDB_val *data,
+ MDB_cursor_op op)
+{
+ int rc;
+ int exact = 0;
+
+ assert(mc);
+
+ switch (op) {
+ case MDB_GET_BOTH:
+ case MDB_GET_BOTH_RANGE:
+ if (data == NULL || mc->mc_xcursor == NULL) {
+ rc = EINVAL;
+ break;
+ }
+ /* FALLTHRU */
+ case MDB_SET:
+ case MDB_SET_RANGE:
+ if (key == NULL || key->mv_size == 0 || key->mv_size > MAXKEYSIZE) {
+ rc = EINVAL;
+ } else if (op == MDB_SET_RANGE)
+ rc = mdb_cursor_set(mc, key, data, op, NULL);
+ else
+ rc = mdb_cursor_set(mc, key, data, op, &exact);
+ break;
+ case MDB_GET_MULTIPLE:
+ if (data == NULL ||
+ !(mc->mc_db->md_flags & MDB_DUPFIXED) ||
+ !(mc->mc_flags & C_INITIALIZED)) {
+ rc = EINVAL;
+ break;
+ }
+ rc = MDB_SUCCESS;
+ if (!(mc->mc_xcursor->mx_cursor.mc_flags & C_INITIALIZED) ||
+ (mc->mc_xcursor->mx_cursor.mc_flags & C_EOF))
+ break;
+ goto fetchm;
+ case MDB_NEXT_MULTIPLE:
+ if (data == NULL ||
+ !(mc->mc_db->md_flags & MDB_DUPFIXED)) {
+ rc = EINVAL;
+ break;
+ }
+ if (!(mc->mc_flags & C_INITIALIZED))
+ rc = mdb_cursor_first(mc, key, data);
+ else
+ rc = mdb_cursor_next(mc, key, data, MDB_NEXT_DUP);
+ if (rc == MDB_SUCCESS) {
+ if (mc->mc_xcursor->mx_cursor.mc_flags & C_INITIALIZED) {
+ MDB_cursor *mx;
+fetchm:
+ mx = &mc->mc_xcursor->mx_cursor;
+ data->mv_size = NUMKEYS(mx->mc_pg[mx->mc_top]) *
+ mx->mc_db->md_pad;
+ data->mv_data = METADATA(mx->mc_pg[mx->mc_top]);
+ mx->mc_ki[mx->mc_top] = NUMKEYS(mx->mc_pg[mx->mc_top])-1;
+ } else {
+ rc = MDB_NOTFOUND;
+ }
+ }
+ break;
+ case MDB_NEXT:
+ case MDB_NEXT_DUP:
+ case MDB_NEXT_NODUP:
+ if (!(mc->mc_flags & C_INITIALIZED))
+ rc = mdb_cursor_first(mc, key, data);
+ else
+ rc = mdb_cursor_next(mc, key, data, op);
+ break;
+ case MDB_PREV:
+ case MDB_PREV_DUP:
+ case MDB_PREV_NODUP:
+ if (!(mc->mc_flags & C_INITIALIZED) || (mc->mc_flags & C_EOF))
+ rc = mdb_cursor_last(mc, key, data);
+ else
+ rc = mdb_cursor_prev(mc, key, data, op);
+ break;
+ case MDB_FIRST:
+ rc = mdb_cursor_first(mc, key, data);
+ break;
+ case MDB_FIRST_DUP:
+ if (data == NULL ||
+ !(mc->mc_db->md_flags & MDB_DUPSORT) ||
+ !(mc->mc_flags & C_INITIALIZED) ||
+ !(mc->mc_xcursor->mx_cursor.mc_flags & C_INITIALIZED)) {
+ rc = EINVAL;
+ break;
+ }
+ rc = mdb_cursor_first(&mc->mc_xcursor->mx_cursor, data, NULL);
+ break;
+ case MDB_LAST:
+ rc = mdb_cursor_last(mc, key, data);
+ break;
+ case MDB_LAST_DUP:
+ if (data == NULL ||
+ !(mc->mc_db->md_flags & MDB_DUPSORT) ||
+ !(mc->mc_flags & C_INITIALIZED) ||
+ !(mc->mc_xcursor->mx_cursor.mc_flags & C_INITIALIZED)) {
+ rc = EINVAL;
+ break;
+ }
+ rc = mdb_cursor_last(&mc->mc_xcursor->mx_cursor, data, NULL);
+ break;
+ default:
+ DPRINTF("unhandled/unimplemented cursor operation %u", op);
+ rc = EINVAL;
+ break;
+ }
+
+ return rc;
+}
+
+/** Touch all the pages in the cursor stack.
+ * Makes sure all the pages are writable, before attempting a write operation.
+ * @param[in] mc The cursor to operate on.
+ */
+static int
+mdb_cursor_touch(MDB_cursor *mc)
+{
+ int rc;
+
+ if (mc->mc_dbi > MAIN_DBI && !(*mc->mc_dbflag & DB_DIRTY)) {
+ MDB_cursor mc2;
+ mdb_cursor_init(&mc2, mc->mc_txn, MAIN_DBI, NULL);
+ rc = mdb_page_search(&mc2, &mc->mc_dbx->md_name, 1);
+ if (rc)
+ return rc;
+ *mc->mc_dbflag = DB_DIRTY;
+ }
+ for (mc->mc_top = 0; mc->mc_top < mc->mc_snum; mc->mc_top++) {
+ rc = mdb_page_touch(mc);
+ if (rc)
+ return rc;
+ }
+ mc->mc_top = mc->mc_snum-1;
+ return MDB_SUCCESS;
+}
+
+int
+mdb_cursor_put(MDB_cursor *mc, MDB_val *key, MDB_val *data,
+ unsigned int flags)
+{
+ MDB_node *leaf = NULL;
+ MDB_val xdata, *rdata, dkey;
+ MDB_page *fp;
+ MDB_db dummy;
+ int do_sub = 0;
+ unsigned int mcount = 0;
+ size_t nsize;
+ int rc, rc2;
+ char pbuf[PAGESIZE];
+ char dbuf[MAXKEYSIZE+1];
+ unsigned int nflags;
+ DKBUF;
+
+ if (F_ISSET(mc->mc_txn->mt_flags, MDB_TXN_RDONLY))
+ return EACCES;
+
+ DPRINTF("==> put db %u key [%s], size %zu, data size %zu",
+ mc->mc_dbi, DKEY(key), key ? key->mv_size:0, data->mv_size);
+
+ dkey.mv_size = 0;
+
+ if (flags == MDB_CURRENT) {
+ if (!(mc->mc_flags & C_INITIALIZED))
+ return EINVAL;
+ rc = MDB_SUCCESS;
+ } else if (mc->mc_db->md_root == P_INVALID) {
+ MDB_page *np;
+ /* new database, write a root leaf page */
+ DPUTS("allocating new root leaf page");
+ if ((np = mdb_page_new(mc, P_LEAF, 1)) == NULL) {
+ return ENOMEM;
+ }
+ mc->mc_snum = 0;
+ mdb_cursor_push(mc, np);
+ mc->mc_db->md_root = np->mp_pgno;
+ mc->mc_db->md_depth++;
+ *mc->mc_dbflag = DB_DIRTY;
+ if ((mc->mc_db->md_flags & (MDB_DUPSORT|MDB_DUPFIXED))
+ == MDB_DUPFIXED)
+ np->mp_flags |= P_LEAF2;
+ mc->mc_flags |= C_INITIALIZED;
+ rc = MDB_NOTFOUND;
+ goto top;
+ } else {
+ int exact = 0;
+ MDB_val d2;
+ rc = mdb_cursor_set(mc, key, &d2, MDB_SET, &exact);
+ if ((flags & MDB_NOOVERWRITE) && rc == 0) {
+ DPRINTF("duplicate key [%s]", DKEY(key));
+ *data = d2;
+ return MDB_KEYEXIST;
+ }
+ if (rc && rc != MDB_NOTFOUND)
+ return rc;
+ }
+
+ /* Cursor is positioned, now make sure all pages are writable */
+ rc2 = mdb_cursor_touch(mc);
+ if (rc2)
+ return rc2;
+
+top:
+ /* The key already exists */
+ if (rc == MDB_SUCCESS) {
+ /* there's only a key anyway, so this is a no-op */
+ if (IS_LEAF2(mc->mc_pg[mc->mc_top])) {
+ unsigned int ksize = mc->mc_db->md_pad;
+ if (key->mv_size != ksize)
+ return EINVAL;
+ if (flags == MDB_CURRENT) {
+ char *ptr = LEAF2KEY(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top], ksize);
+ memcpy(ptr, key->mv_data, ksize);
+ }
+ return MDB_SUCCESS;
+ }
+
+ leaf = NODEPTR(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top]);
+
+ /* DB has dups? */
+ if (F_ISSET(mc->mc_db->md_flags, MDB_DUPSORT)) {
+ /* Was a single item before, must convert now */
+more:
+ if (!F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ /* Just overwrite the current item */
+ if (flags == MDB_CURRENT)
+ goto current;
+
+ dkey.mv_size = NODEDSZ(leaf);
+ dkey.mv_data = NODEDATA(leaf);
+ /* data matches, ignore it */
+ if (!mc->mc_dbx->md_dcmp(data, &dkey))
+ return (flags == MDB_NODUPDATA) ? MDB_KEYEXIST : MDB_SUCCESS;
+
+ /* create a fake page for the dup items */
+ memcpy(dbuf, dkey.mv_data, dkey.mv_size);
+ dkey.mv_data = dbuf;
+ fp = (MDB_page *)pbuf;
+ fp->mp_pgno = mc->mc_pg[mc->mc_top]->mp_pgno;
+ fp->mp_flags = P_LEAF|P_DIRTY|P_SUBP;
+ fp->mp_lower = PAGEHDRSZ;
+ fp->mp_upper = PAGEHDRSZ + dkey.mv_size + data->mv_size;
+ if (mc->mc_db->md_flags & MDB_DUPFIXED) {
+ fp->mp_flags |= P_LEAF2;
+ fp->mp_pad = data->mv_size;
+ } else {
+ fp->mp_upper += 2 * sizeof(indx_t) + 2 * NODESIZE +
+ (dkey.mv_size & 1) + (data->mv_size & 1);
+ }
+ mdb_node_del(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top], 0);
+ do_sub = 1;
+ rdata = &xdata;
+ xdata.mv_size = fp->mp_upper;
+ xdata.mv_data = pbuf;
+ flags |= F_DUPDATA;
+ goto new_sub;
+ }
+ if (!F_ISSET(leaf->mn_flags, F_SUBDATA)) {
+ /* See if we need to convert from fake page to subDB */
+ MDB_page *mp;
+ unsigned int offset;
+ unsigned int i;
+
+ fp = NODEDATA(leaf);
+ if (flags == MDB_CURRENT) {
+ fp->mp_flags |= P_DIRTY;
+#ifdef MISALIGNED_OK
+ fp->mp_pgno = mc->mc_pg[mc->mc_top]->mp_pgno;
+#else
+ {
+ unsigned short *src, *dst;
+ int i;
+ dst = (unsigned short *)&fp->mp_pgno;
+ src = (unsigned short *)&mc->mc_pg[mc->mc_top]->mp_pgno;
+ for (i=0; i<sizeof(fp->mp_pgno)/sizeof(unsigned short); i++)
+ *dst++ = *src++;
+ }
+#endif
+ mc->mc_xcursor->mx_cursor.mc_pg[0] = fp;
+ flags |= F_DUPDATA;
+ goto put_sub;
+ }
+ if (mc->mc_db->md_flags & MDB_DUPFIXED) {
+ offset = fp->mp_pad;
+ } else {
+ offset = NODESIZE + sizeof(indx_t) + data->mv_size;
+ }
+ offset += offset & 1;
+ if (NODESIZE + sizeof(indx_t) + NODEKSZ(leaf) + NODEDSZ(leaf) +
+ offset >= (mc->mc_txn->mt_env->me_psize - PAGEHDRSZ) /
+ MDB_MINKEYS) {
+ /* yes, convert it */
+ dummy.md_flags = 0;
+ if (mc->mc_db->md_flags & MDB_DUPFIXED) {
+ dummy.md_pad = fp->mp_pad;
+ dummy.md_flags = MDB_DUPFIXED;
+ if (mc->mc_db->md_flags & MDB_INTEGERDUP)
+ dummy.md_flags |= MDB_INTEGERKEY;
+ }
+ dummy.md_depth = 1;
+ dummy.md_branch_pages = 0;
+ dummy.md_leaf_pages = 1;
+ dummy.md_overflow_pages = 0;
+ dummy.md_entries = NUMKEYS(fp);
+ rdata = &xdata;
+ xdata.mv_size = sizeof(MDB_db);
+ xdata.mv_data = &dummy;
+ mp = mdb_page_alloc(mc, 1);
+ if (!mp)
+ return ENOMEM;
+ offset = mc->mc_txn->mt_env->me_psize - NODEDSZ(leaf);
+ flags |= F_DUPDATA|F_SUBDATA;
+ dummy.md_root = mp->mp_pgno;
+ } else {
+ /* no, just grow it */
+ rdata = &xdata;
+ xdata.mv_size = NODEDSZ(leaf) + offset;
+ xdata.mv_data = pbuf;
+ mp = (MDB_page *)pbuf;
+ mp->mp_pgno = mc->mc_pg[mc->mc_top]->mp_pgno;
+ flags |= F_DUPDATA;
+ }
+ mp->mp_flags = fp->mp_flags | P_DIRTY;
+ mp->mp_pad = fp->mp_pad;
+ mp->mp_lower = fp->mp_lower;
+ mp->mp_upper = fp->mp_upper + offset;
+ if (IS_LEAF2(fp)) {
+ memcpy(METADATA(mp), METADATA(fp), NUMKEYS(fp) * fp->mp_pad);
+ } else {
+ nsize = NODEDSZ(leaf) - fp->mp_upper;
+ memcpy((char *)mp + mp->mp_upper, (char *)fp + fp->mp_upper, nsize);
+ for (i=0; i<NUMKEYS(fp); i++)
+ mp->mp_ptrs[i] = fp->mp_ptrs[i] + offset;
+ }
+ mdb_node_del(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top], 0);
+ do_sub = 1;
+ goto new_sub;
+ }
+ /* data is on sub-DB, just store it */
+ flags |= F_DUPDATA|F_SUBDATA;
+ goto put_sub;
+ }
+current:
+ /* same size, just replace it */
+ if (!F_ISSET(leaf->mn_flags, F_BIGDATA) &&
+ NODEDSZ(leaf) == data->mv_size) {
+ if (F_ISSET(flags, MDB_RESERVE))
+ data->mv_data = NODEDATA(leaf);
+ else
+ memcpy(NODEDATA(leaf), data->mv_data, data->mv_size);
+ goto done;
+ }
+ mdb_node_del(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top], 0);
+ } else {
+ DPRINTF("inserting key at index %i", mc->mc_ki[mc->mc_top]);
+ }
+
+ rdata = data;
+
+new_sub:
+ nflags = flags & NODE_ADD_FLAGS;
+ nsize = IS_LEAF2(mc->mc_pg[mc->mc_top]) ? key->mv_size : mdb_leaf_size(mc->mc_txn->mt_env, key, rdata);
+ if (SIZELEFT(mc->mc_pg[mc->mc_top]) < nsize) {
+ if (( flags & (F_DUPDATA|F_SUBDATA)) == F_DUPDATA )
+ nflags &= ~MDB_APPEND;
+ rc = mdb_page_split(mc, key, rdata, P_INVALID, nflags);
+ } else {
+ /* There is room already in this leaf page. */
+ rc = mdb_node_add(mc, mc->mc_ki[mc->mc_top], key, rdata, 0, nflags);
+ if (rc == 0 && !do_sub) {
+ /* Adjust other cursors pointing to mp */
+ MDB_cursor *m2, *m3;
+ MDB_dbi dbi = mc->mc_dbi;
+ unsigned i = mc->mc_top;
+ MDB_page *mp = mc->mc_pg[i];
+
+ if (mc->mc_flags & C_SUB)
+ dbi--;
+
+ for (m2 = mc->mc_txn->mt_cursors[dbi]; m2; m2=m2->mc_next) {
+ if (mc->mc_flags & C_SUB)
+ m3 = &m2->mc_xcursor->mx_cursor;
+ else
+ m3 = m2;
+ if (m3 == mc) continue;
+ if (m3->mc_pg[i] == mp && m3->mc_ki[i] >= mc->mc_ki[i]) {
+ m3->mc_ki[i]++;
+ }
+ }
+ }
+ }
+
+ if (rc != MDB_SUCCESS)
+ mc->mc_txn->mt_flags |= MDB_TXN_ERROR;
+ else {
+ /* Now store the actual data in the child DB. Note that we're
+ * storing the user data in the keys field, so there are strict
+ * size limits on dupdata. The actual data fields of the child
+ * DB are all zero size.
+ */
+ if (do_sub) {
+ MDB_db *db;
+ int xflags;
+put_sub:
+ xdata.mv_size = 0;
+ xdata.mv_data = "";
+ leaf = NODEPTR(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top]);
+ if (flags & MDB_CURRENT) {
+ xflags = MDB_CURRENT;
+ } else {
+ mdb_xcursor_init1(mc, leaf);
+ xflags = (flags & MDB_NODUPDATA) ? MDB_NOOVERWRITE : 0;
+ }
+ /* converted, write the original data first */
+ if (dkey.mv_size) {
+ rc = mdb_cursor_put(&mc->mc_xcursor->mx_cursor, &dkey, &xdata, xflags);
+ if (rc)
+ return rc;
+ {
+ /* Adjust other cursors pointing to mp */
+ MDB_cursor *m2;
+ unsigned i = mc->mc_top;
+ MDB_page *mp = mc->mc_pg[i];
+
+ for (m2 = mc->mc_txn->mt_cursors[mc->mc_dbi]; m2; m2=m2->mc_next) {
+ if (m2 == mc) continue;
+ if (m2->mc_pg[i] == mp && m2->mc_ki[i] == mc->mc_ki[i]) {
+ mdb_xcursor_init1(m2, leaf);
+ }
+ }
+ }
+ }
+ xflags |= (flags & MDB_APPEND);
+ rc = mdb_cursor_put(&mc->mc_xcursor->mx_cursor, data, &xdata, xflags);
+ if (flags & F_SUBDATA) {
+ db = NODEDATA(leaf);
+ memcpy(db, &mc->mc_xcursor->mx_db, sizeof(MDB_db));
+ }
+ }
+ /* sub-writes might have failed so check rc again.
+ * Don't increment count if we just replaced an existing item.
+ */
+ if (!rc && !(flags & MDB_CURRENT))
+ mc->mc_db->md_entries++;
+ if (flags & MDB_MULTIPLE) {
+ mcount++;
+ if (mcount < data[1].mv_size) {
+ data[0].mv_data = (char *)data[0].mv_data + data[0].mv_size;
+ leaf = NODEPTR(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top]);
+ goto more;
+ }
+ }
+ }
+done:
+ return rc;
+}
+
+int
+mdb_cursor_del(MDB_cursor *mc, unsigned int flags)
+{
+ MDB_node *leaf;
+ int rc;
+
+ if (F_ISSET(mc->mc_txn->mt_flags, MDB_TXN_RDONLY))
+ return EACCES;
+
+ if (!mc->mc_flags & C_INITIALIZED)
+ return EINVAL;
+
+ rc = mdb_cursor_touch(mc);
+ if (rc)
+ return rc;
+
+ leaf = NODEPTR(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top]);
+
+ if (!IS_LEAF2(mc->mc_pg[mc->mc_top]) && F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ if (flags != MDB_NODUPDATA) {
+ if (!F_ISSET(leaf->mn_flags, F_SUBDATA)) {
+ mc->mc_xcursor->mx_cursor.mc_pg[0] = NODEDATA(leaf);
+ }
+ rc = mdb_cursor_del(&mc->mc_xcursor->mx_cursor, 0);
+ /* If sub-DB still has entries, we're done */
+ if (mc->mc_xcursor->mx_db.md_entries) {
+ if (leaf->mn_flags & F_SUBDATA) {
+ /* update subDB info */
+ MDB_db *db = NODEDATA(leaf);
+ memcpy(db, &mc->mc_xcursor->mx_db, sizeof(MDB_db));
+ } else {
+ /* shrink fake page */
+ mdb_node_shrink(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top]);
+ }
+ mc->mc_db->md_entries--;
+ return rc;
+ }
+ /* otherwise fall thru and delete the sub-DB */
+ }
+
+ if (leaf->mn_flags & F_SUBDATA) {
+ /* add all the child DB's pages to the free list */
+ rc = mdb_drop0(&mc->mc_xcursor->mx_cursor, 0);
+ if (rc == MDB_SUCCESS) {
+ mc->mc_db->md_entries -=
+ mc->mc_xcursor->mx_db.md_entries;
+ }
+ }
+ }
+
+ return mdb_cursor_del0(mc, leaf);
+}
+
+/** Allocate and initialize new pages for a database.
+ * @param[in] mc a cursor on the database being added to.
+ * @param[in] flags flags defining what type of page is being allocated.
+ * @param[in] num the number of pages to allocate. This is usually 1,
+ * unless allocating overflow pages for a large record.
+ * @return Address of a page, or NULL on failure.
+ */
+static MDB_page *
+mdb_page_new(MDB_cursor *mc, uint32_t flags, int num)
+{
+ MDB_page *np;
+
+ if ((np = mdb_page_alloc(mc, num)) == NULL)
+ return NULL;
+ DPRINTF("allocated new mpage %zu, page size %u",
+ np->mp_pgno, mc->mc_txn->mt_env->me_psize);
+ np->mp_flags = flags | P_DIRTY;
+ np->mp_lower = PAGEHDRSZ;
+ np->mp_upper = mc->mc_txn->mt_env->me_psize;
+
+ if (IS_BRANCH(np))
+ mc->mc_db->md_branch_pages++;
+ else if (IS_LEAF(np))
+ mc->mc_db->md_leaf_pages++;
+ else if (IS_OVERFLOW(np)) {
+ mc->mc_db->md_overflow_pages += num;
+ np->mp_pages = num;
+ }
+
+ return np;
+}
+
+/** Calculate the size of a leaf node.
+ * The size depends on the environment's page size; if a data item
+ * is too large it will be put onto an overflow page and the node
+ * size will only include the key and not the data. Sizes are always
+ * rounded up to an even number of bytes, to guarantee 2-byte alignment
+ * of the #MDB_node headers.
+ * @param[in] env The environment handle.
+ * @param[in] key The key for the node.
+ * @param[in] data The data for the node.
+ * @return The number of bytes needed to store the node.
+ */
+static size_t
+mdb_leaf_size(MDB_env *env, MDB_val *key, MDB_val *data)
+{
+ size_t sz;
+
+ sz = LEAFSIZE(key, data);
+ if (data->mv_size >= env->me_psize / MDB_MINKEYS) {
+ /* put on overflow page */
+ sz -= data->mv_size - sizeof(pgno_t);
+ }
+ sz += sz & 1;
+
+ return sz + sizeof(indx_t);
+}
+
+/** Calculate the size of a branch node.
+ * The size should depend on the environment's page size but since
+ * we currently don't support spilling large keys onto overflow
+ * pages, it's simply the size of the #MDB_node header plus the
+ * size of the key. Sizes are always rounded up to an even number
+ * of bytes, to guarantee 2-byte alignment of the #MDB_node headers.
+ * @param[in] env The environment handle.
+ * @param[in] key The key for the node.
+ * @return The number of bytes needed to store the node.
+ */
+static size_t
+mdb_branch_size(MDB_env *env, MDB_val *key)
+{
+ size_t sz;
+
+ sz = INDXSIZE(key);
+ if (sz >= env->me_psize / MDB_MINKEYS) {
+ /* put on overflow page */
+ /* not implemented */
+ /* sz -= key->size - sizeof(pgno_t); */
+ }
+
+ return sz + sizeof(indx_t);
+}
+
+/** Add a node to the page pointed to by the cursor.
+ * @param[in] mc The cursor for this operation.
+ * @param[in] indx The index on the page where the new node should be added.
+ * @param[in] key The key for the new node.
+ * @param[in] data The data for the new node, if any.
+ * @param[in] pgno The page number, if adding a branch node.
+ * @param[in] flags Flags for the node.
+ * @return 0 on success, non-zero on failure. Possible errors are:
+ * <ul>
+ * <li>ENOMEM - failed to allocate overflow pages for the node.
+ * <li>ENOSPC - there is insufficient room in the page. This error
+ * should never happen since all callers already calculate the
+ * page's free space before calling this function.
+ * </ul>
+ */
+static int
+mdb_node_add(MDB_cursor *mc, indx_t indx,
+ MDB_val *key, MDB_val *data, pgno_t pgno, unsigned int flags)
+{
+ unsigned int i;
+ size_t node_size = NODESIZE;
+ indx_t ofs;
+ MDB_node *node;
+ MDB_page *mp = mc->mc_pg[mc->mc_top];
+ MDB_page *ofp = NULL; /* overflow page */
+ DKBUF;
+
+ assert(mp->mp_upper >= mp->mp_lower);
+
+ DPRINTF("add to %s %spage %zu index %i, data size %zu key size %zu [%s]",
+ IS_LEAF(mp) ? "leaf" : "branch",
+ IS_SUBP(mp) ? "sub-" : "",
+ mp->mp_pgno, indx, data ? data->mv_size : 0,
+ key ? key->mv_size : 0, key ? DKEY(key) : NULL);
+
+ if (IS_LEAF2(mp)) {
+ /* Move higher keys up one slot. */
+ int ksize = mc->mc_db->md_pad, dif;
+ char *ptr = LEAF2KEY(mp, indx, ksize);
+ dif = NUMKEYS(mp) - indx;
+ if (dif > 0)
+ memmove(ptr+ksize, ptr, dif*ksize);
+ /* insert new key */
+ memcpy(ptr, key->mv_data, ksize);
+
+ /* Just using these for counting */
+ mp->mp_lower += sizeof(indx_t);
+ mp->mp_upper -= ksize - sizeof(indx_t);
+ return MDB_SUCCESS;
+ }
+
+ if (key != NULL)
+ node_size += key->mv_size;
+
+ if (IS_LEAF(mp)) {
+ assert(data);
+ if (F_ISSET(flags, F_BIGDATA)) {
+ /* Data already on overflow page. */
+ node_size += sizeof(pgno_t);
+ } else if (data->mv_size >= mc->mc_txn->mt_env->me_psize / MDB_MINKEYS) {
+ int ovpages = OVPAGES(data->mv_size, mc->mc_txn->mt_env->me_psize);
+ /* Put data on overflow page. */
+ DPRINTF("data size is %zu, put on overflow page",
+ data->mv_size);
+ node_size += sizeof(pgno_t);
+ if ((ofp = mdb_page_new(mc, P_OVERFLOW, ovpages)) == NULL)
+ return ENOMEM;
+ DPRINTF("allocated overflow page %zu", ofp->mp_pgno);
+ flags |= F_BIGDATA;
+ } else {
+ node_size += data->mv_size;
+ }
+ }
+ node_size += node_size & 1;
+
+ if (node_size + sizeof(indx_t) > SIZELEFT(mp)) {
+ DPRINTF("not enough room in page %zu, got %u ptrs",
+ mp->mp_pgno, NUMKEYS(mp));
+ DPRINTF("upper - lower = %u - %u = %u", mp->mp_upper, mp->mp_lower,
+ mp->mp_upper - mp->mp_lower);
+ DPRINTF("node size = %zu", node_size);
+ return ENOSPC;
+ }
+
+ /* Move higher pointers up one slot. */
+ for (i = NUMKEYS(mp); i > indx; i--)
+ mp->mp_ptrs[i] = mp->mp_ptrs[i - 1];
+
+ /* Adjust free space offsets. */
+ ofs = mp->mp_upper - node_size;
+ assert(ofs >= mp->mp_lower + sizeof(indx_t));
+ mp->mp_ptrs[indx] = ofs;
+ mp->mp_upper = ofs;
+ mp->mp_lower += sizeof(indx_t);
+
+ /* Write the node data. */
+ node = NODEPTR(mp, indx);
+ node->mn_ksize = (key == NULL) ? 0 : key->mv_size;
+ node->mn_flags = flags;
+ if (IS_LEAF(mp))
+ SETDSZ(node,data->mv_size);
+ else
+ SETPGNO(node,pgno);
+
+ if (key)
+ memcpy(NODEKEY(node), key->mv_data, key->mv_size);
+
+ if (IS_LEAF(mp)) {
+ assert(key);
+ if (ofp == NULL) {
+ if (F_ISSET(flags, F_BIGDATA))
+ memcpy(node->mn_data + key->mv_size, data->mv_data,
+ sizeof(pgno_t));
+ else if (F_ISSET(flags, MDB_RESERVE))
+ data->mv_data = node->mn_data + key->mv_size;
+ else
+ memcpy(node->mn_data + key->mv_size, data->mv_data,
+ data->mv_size);
+ } else {
+ memcpy(node->mn_data + key->mv_size, &ofp->mp_pgno,
+ sizeof(pgno_t));
+ if (F_ISSET(flags, MDB_RESERVE))
+ data->mv_data = METADATA(ofp);
+ else
+ memcpy(METADATA(ofp), data->mv_data, data->mv_size);
+ }
+ }
+
+ return MDB_SUCCESS;
+}
+
+/** Delete the specified node from a page.
+ * @param[in] mp The page to operate on.
+ * @param[in] indx The index of the node to delete.
+ * @param[in] ksize The size of a node. Only used if the page is
+ * part of a #MDB_DUPFIXED database.
+ */
+static void
+mdb_node_del(MDB_page *mp, indx_t indx, int ksize)
+{
+ unsigned int sz;
+ indx_t i, j, numkeys, ptr;
+ MDB_node *node;
+ char *base;
+
+ DPRINTF("delete node %u on %s page %zu", indx,
+ IS_LEAF(mp) ? "leaf" : "branch", mp->mp_pgno);
+ assert(indx < NUMKEYS(mp));
+
+ if (IS_LEAF2(mp)) {
+ int x = NUMKEYS(mp) - 1 - indx;
+ base = LEAF2KEY(mp, indx, ksize);
+ if (x)
+ memmove(base, base + ksize, x * ksize);
+ mp->mp_lower -= sizeof(indx_t);
+ mp->mp_upper += ksize - sizeof(indx_t);
+ return;
+ }
+
+ node = NODEPTR(mp, indx);
+ sz = NODESIZE + node->mn_ksize;
+ if (IS_LEAF(mp)) {
+ if (F_ISSET(node->mn_flags, F_BIGDATA))
+ sz += sizeof(pgno_t);
+ else
+ sz += NODEDSZ(node);
+ }
+ sz += sz & 1;
+
+ ptr = mp->mp_ptrs[indx];
+ numkeys = NUMKEYS(mp);
+ for (i = j = 0; i < numkeys; i++) {
+ if (i != indx) {
+ mp->mp_ptrs[j] = mp->mp_ptrs[i];
+ if (mp->mp_ptrs[i] < ptr)
+ mp->mp_ptrs[j] += sz;
+ j++;
+ }
+ }
+
+ base = (char *)mp + mp->mp_upper;
+ memmove(base + sz, base, ptr - mp->mp_upper);
+
+ mp->mp_lower -= sizeof(indx_t);
+ mp->mp_upper += sz;
+}
+
+/** Compact the main page after deleting a node on a subpage.
+ * @param[in] mp The main page to operate on.
+ * @param[in] indx The index of the subpage on the main page.
+ */
+static void
+mdb_node_shrink(MDB_page *mp, indx_t indx)
+{
+ MDB_node *node;
+ MDB_page *sp, *xp;
+ char *base;
+ int osize, nsize;
+ int delta;
+ indx_t i, numkeys, ptr;
+
+ node = NODEPTR(mp, indx);
+ sp = (MDB_page *)NODEDATA(node);
+ osize = NODEDSZ(node);
+
+ delta = sp->mp_upper - sp->mp_lower;
+ SETDSZ(node, osize - delta);
+ xp = (MDB_page *)((char *)sp + delta);
+
+ /* shift subpage upward */
+ if (IS_LEAF2(sp)) {
+ nsize = NUMKEYS(sp) * sp->mp_pad;
+ memmove(METADATA(xp), METADATA(sp), nsize);
+ } else {
+ int i;
+ nsize = osize - sp->mp_upper;
+ numkeys = NUMKEYS(sp);
+ for (i=numkeys-1; i>=0; i--)
+ xp->mp_ptrs[i] = sp->mp_ptrs[i] - delta;
+ }
+ xp->mp_upper = sp->mp_lower;
+ xp->mp_lower = sp->mp_lower;
+ xp->mp_flags = sp->mp_flags;
+ xp->mp_pad = sp->mp_pad;
+ xp->mp_pgno = mp->mp_pgno;
+
+ /* shift lower nodes upward */
+ ptr = mp->mp_ptrs[indx];
+ numkeys = NUMKEYS(mp);
+ for (i = 0; i < numkeys; i++) {
+ if (mp->mp_ptrs[i] <= ptr)
+ mp->mp_ptrs[i] += delta;
+ }
+
+ base = (char *)mp + mp->mp_upper;
+ memmove(base + delta, base, ptr - mp->mp_upper + NODESIZE + NODEKSZ(node));
+ mp->mp_upper += delta;
+}
+
+/** Initial setup of a sorted-dups cursor.
+ * Sorted duplicates are implemented as a sub-database for the given key.
+ * The duplicate data items are actually keys of the sub-database.
+ * Operations on the duplicate data items are performed using a sub-cursor
+ * initialized when the sub-database is first accessed. This function does
+ * the preliminary setup of the sub-cursor, filling in the fields that
+ * depend only on the parent DB.
+ * @param[in] mc The main cursor whose sorted-dups cursor is to be initialized.
+ */
+static void
+mdb_xcursor_init0(MDB_cursor *mc)
+{
+ MDB_xcursor *mx = mc->mc_xcursor;
+
+ mx->mx_cursor.mc_xcursor = NULL;
+ mx->mx_cursor.mc_txn = mc->mc_txn;
+ mx->mx_cursor.mc_db = &mx->mx_db;
+ mx->mx_cursor.mc_dbx = &mx->mx_dbx;
+ mx->mx_cursor.mc_dbi = mc->mc_dbi+1;
+ mx->mx_cursor.mc_dbflag = &mx->mx_dbflag;
+ mx->mx_cursor.mc_snum = 0;
+ mx->mx_cursor.mc_flags = C_SUB;
+ mx->mx_dbx.md_cmp = mc->mc_dbx->md_dcmp;
+ mx->mx_dbx.md_dcmp = NULL;
+ mx->mx_dbx.md_rel = mc->mc_dbx->md_rel;
+}
+
+/** Final setup of a sorted-dups cursor.
+ * Sets up the fields that depend on the data from the main cursor.
+ * @param[in] mc The main cursor whose sorted-dups cursor is to be initialized.
+ * @param[in] node The data containing the #MDB_db record for the
+ * sorted-dup database.
+ */
+static void
+mdb_xcursor_init1(MDB_cursor *mc, MDB_node *node)
+{
+ MDB_xcursor *mx = mc->mc_xcursor;
+
+ if (node->mn_flags & F_SUBDATA) {
+ MDB_db *db = NODEDATA(node);
+ mx->mx_db = *db;
+ mx->mx_cursor.mc_snum = 0;
+ mx->mx_cursor.mc_flags = C_SUB;
+ } else {
+ MDB_page *fp = NODEDATA(node);
+ mx->mx_db.md_pad = mc->mc_pg[mc->mc_top]->mp_pad;
+ mx->mx_db.md_flags = 0;
+ mx->mx_db.md_depth = 1;
+ mx->mx_db.md_branch_pages = 0;
+ mx->mx_db.md_leaf_pages = 1;
+ mx->mx_db.md_overflow_pages = 0;
+ mx->mx_db.md_entries = NUMKEYS(fp);
+#ifdef MISALIGNED_OK
+ mx->mx_db.md_root = fp->mp_pgno;
+#else
+ {
+ unsigned short *src, *dst;
+ int i;
+ dst = (unsigned short *)&mx->mx_db.md_root;
+ src = (unsigned short *)&fp->mp_pgno;
+ for (i=0; i<sizeof(fp->mp_pgno)/sizeof(unsigned short); i++)
+ *dst++ = *src++;
+ }
+#endif
+ mx->mx_cursor.mc_snum = 1;
+ mx->mx_cursor.mc_flags = C_INITIALIZED|C_SUB;
+ mx->mx_cursor.mc_top = 0;
+ mx->mx_cursor.mc_pg[0] = fp;
+ mx->mx_cursor.mc_ki[0] = 0;
+ if (mc->mc_db->md_flags & MDB_DUPFIXED) {
+ mx->mx_db.md_flags = MDB_DUPFIXED;
+ mx->mx_db.md_pad = fp->mp_pad;
+ if (mc->mc_db->md_flags & MDB_INTEGERDUP)
+ mx->mx_db.md_flags |= MDB_INTEGERKEY;
+ }
+ }
+ DPRINTF("Sub-db %u for db %u root page %zu", mx->mx_cursor.mc_dbi, mc->mc_dbi,
+ mx->mx_db.md_root);
+ mx->mx_dbflag = (F_ISSET(mc->mc_pg[mc->mc_top]->mp_flags, P_DIRTY)) ?
+ DB_DIRTY : 0;
+ mx->mx_dbx.md_name.mv_data = NODEKEY(node);
+ mx->mx_dbx.md_name.mv_size = node->mn_ksize;
+ if (mx->mx_dbx.md_cmp == mdb_cmp_int && mx->mx_db.md_pad == sizeof(size_t))
+ mx->mx_dbx.md_cmp = mdb_cmp_long;
+}
+
+/** Initialize a cursor for a given transaction and database. */
+static void
+mdb_cursor_init(MDB_cursor *mc, MDB_txn *txn, MDB_dbi dbi, MDB_xcursor *mx)
+{
+ mc->mc_orig = NULL;
+ mc->mc_dbi = dbi;
+ mc->mc_txn = txn;
+ mc->mc_db = &txn->mt_dbs[dbi];
+ mc->mc_dbx = &txn->mt_dbxs[dbi];
+ mc->mc_dbflag = &txn->mt_dbflags[dbi];
+ mc->mc_snum = 0;
+ mc->mc_flags = 0;
+ if (txn->mt_dbs[dbi].md_flags & MDB_DUPSORT) {
+ assert(mx != NULL);
+ mc->mc_xcursor = mx;
+ mdb_xcursor_init0(mc);
+ } else {
+ mc->mc_xcursor = NULL;
+ }
+}
+
+int
+mdb_cursor_open(MDB_txn *txn, MDB_dbi dbi, MDB_cursor **ret)
+{
+ MDB_cursor *mc;
+ MDB_xcursor *mx = NULL;
+ size_t size = sizeof(MDB_cursor);
+
+ if (txn == NULL || ret == NULL || !dbi || dbi >= txn->mt_numdbs)
+ return EINVAL;
+
+ if (txn->mt_dbs[dbi].md_flags & MDB_DUPSORT)
+ size += sizeof(MDB_xcursor);
+
+ if ((mc = malloc(size)) != NULL) {
+ if (txn->mt_dbs[dbi].md_flags & MDB_DUPSORT) {
+ mx = (MDB_xcursor *)(mc + 1);
+ }
+ mdb_cursor_init(mc, txn, dbi, mx);
+ if (txn->mt_cursors) {
+ mc->mc_next = txn->mt_cursors[dbi];
+ txn->mt_cursors[dbi] = mc;
+ }
+ mc->mc_flags |= C_ALLOCD;
+ } else {
+ return ENOMEM;
+ }
+
+ *ret = mc;
+
+ return MDB_SUCCESS;
+}
+
+/* Return the count of duplicate data items for the current key */
+int
+mdb_cursor_count(MDB_cursor *mc, size_t *countp)
+{
+ MDB_node *leaf;
+
+ if (mc == NULL || countp == NULL)
+ return EINVAL;
+
+ if (!(mc->mc_db->md_flags & MDB_DUPSORT))
+ return EINVAL;
+
+ leaf = NODEPTR(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top]);
+ if (!F_ISSET(leaf->mn_flags, F_DUPDATA)) {
+ *countp = 1;
+ } else {
+ if (!(mc->mc_xcursor->mx_cursor.mc_flags & C_INITIALIZED))
+ return EINVAL;
+
+ *countp = mc->mc_xcursor->mx_db.md_entries;
+ }
+ return MDB_SUCCESS;
+}
+
+void
+mdb_cursor_close(MDB_cursor *mc)
+{
+ if (mc != NULL) {
+ /* remove from txn, if tracked */
+ if (mc->mc_txn->mt_cursors) {
+ MDB_cursor **prev = &mc->mc_txn->mt_cursors[mc->mc_dbi];
+ while (*prev && *prev != mc) prev = &(*prev)->mc_next;
+ if (*prev == mc)
+ *prev = mc->mc_next;
+ }
+ if (mc->mc_flags & C_ALLOCD)
+ free(mc);
+ }
+}
+
+MDB_txn *
+mdb_cursor_txn(MDB_cursor *mc)
+{
+ if (!mc) return NULL;
+ return mc->mc_txn;
+}
+
+MDB_dbi
+mdb_cursor_dbi(MDB_cursor *mc)
+{
+ if (!mc) return 0;
+ return mc->mc_dbi;
+}
+
+/** Replace the key for a node with a new key.
+ * @param[in] mp The page containing the node to operate on.
+ * @param[in] indx The index of the node to operate on.
+ * @param[in] key The new key to use.
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_update_key(MDB_page *mp, indx_t indx, MDB_val *key)
+{
+ indx_t ptr, i, numkeys;
+ int delta;
+ size_t len;
+ MDB_node *node;
+ char *base;
+ DKBUF;
+
+ node = NODEPTR(mp, indx);
+ ptr = mp->mp_ptrs[indx];
+ DPRINTF("update key %u (ofs %u) [%.*s] to [%s] on page %zu",
+ indx, ptr,
+ (int)node->mn_ksize, (char *)NODEKEY(node),
+ DKEY(key),
+ mp->mp_pgno);
+
+ delta = key->mv_size - node->mn_ksize;
+ if (delta) {
+ if (delta > 0 && SIZELEFT(mp) < delta) {
+ DPRINTF("OUCH! Not enough room, delta = %d", delta);
+ return ENOSPC;
+ }
+
+ numkeys = NUMKEYS(mp);
+ for (i = 0; i < numkeys; i++) {
+ if (mp->mp_ptrs[i] <= ptr)
+ mp->mp_ptrs[i] -= delta;
+ }
+
+ base = (char *)mp + mp->mp_upper;
+ len = ptr - mp->mp_upper + NODESIZE;
+ memmove(base - delta, base, len);
+ mp->mp_upper -= delta;
+
+ node = NODEPTR(mp, indx);
+ node->mn_ksize = key->mv_size;
+ }
+
+ memcpy(NODEKEY(node), key->mv_data, key->mv_size);
+
+ return MDB_SUCCESS;
+}
+
+/** Move a node from csrc to cdst.
+ */
+static int
+mdb_node_move(MDB_cursor *csrc, MDB_cursor *cdst)
+{
+ int rc;
+ MDB_node *srcnode;
+ MDB_val key, data;
+ DKBUF;
+
+ /* Mark src and dst as dirty. */
+ if ((rc = mdb_page_touch(csrc)) ||
+ (rc = mdb_page_touch(cdst)))
+ return rc;
+
+ if (IS_LEAF2(csrc->mc_pg[csrc->mc_top])) {
+ srcnode = NODEPTR(csrc->mc_pg[csrc->mc_top], 0); /* fake */
+ key.mv_size = csrc->mc_db->md_pad;
+ key.mv_data = LEAF2KEY(csrc->mc_pg[csrc->mc_top], csrc->mc_ki[csrc->mc_top], key.mv_size);
+ data.mv_size = 0;
+ data.mv_data = NULL;
+ } else {
+ srcnode = NODEPTR(csrc->mc_pg[csrc->mc_top], csrc->mc_ki[csrc->mc_top]);
+ if (csrc->mc_ki[csrc->mc_top] == 0 && IS_BRANCH(csrc->mc_pg[csrc->mc_top])) {
+ unsigned int snum = csrc->mc_snum;
+ MDB_node *s2;
+ /* must find the lowest key below src */
+ mdb_page_search_root(csrc, NULL, 0);
+ s2 = NODEPTR(csrc->mc_pg[csrc->mc_top], 0);
+ key.mv_size = NODEKSZ(s2);
+ key.mv_data = NODEKEY(s2);
+ csrc->mc_snum = snum--;
+ csrc->mc_top = snum;
+ } else {
+ key.mv_size = NODEKSZ(srcnode);
+ key.mv_data = NODEKEY(srcnode);
+ }
+ data.mv_size = NODEDSZ(srcnode);
+ data.mv_data = NODEDATA(srcnode);
+ }
+ DPRINTF("moving %s node %u [%s] on page %zu to node %u on page %zu",
+ IS_LEAF(csrc->mc_pg[csrc->mc_top]) ? "leaf" : "branch",
+ csrc->mc_ki[csrc->mc_top],
+ DKEY(&key),
+ csrc->mc_pg[csrc->mc_top]->mp_pgno,
+ cdst->mc_ki[cdst->mc_top], cdst->mc_pg[cdst->mc_top]->mp_pgno);
+
+ /* Add the node to the destination page.
+ */
+ rc = mdb_node_add(cdst, cdst->mc_ki[cdst->mc_top], &key, &data, NODEPGNO(srcnode),
+ srcnode->mn_flags);
+ if (rc != MDB_SUCCESS)
+ return rc;
+
+ /* Delete the node from the source page.
+ */
+ mdb_node_del(csrc->mc_pg[csrc->mc_top], csrc->mc_ki[csrc->mc_top], key.mv_size);
+
+ {
+ /* Adjust other cursors pointing to mp */
+ MDB_cursor *m2, *m3;
+ MDB_dbi dbi = csrc->mc_dbi;
+ MDB_page *mp = csrc->mc_pg[csrc->mc_top];
+
+ if (csrc->mc_flags & C_SUB)
+ dbi--;
+
+ for (m2 = csrc->mc_txn->mt_cursors[dbi]; m2; m2=m2->mc_next) {
+ if (m2 == csrc) continue;
+ if (csrc->mc_flags & C_SUB)
+ m3 = &m2->mc_xcursor->mx_cursor;
+ else
+ m3 = m2;
+ if (m3->mc_pg[csrc->mc_top] == mp && m3->mc_ki[csrc->mc_top] ==
+ csrc->mc_ki[csrc->mc_top]) {
+ m3->mc_pg[csrc->mc_top] = cdst->mc_pg[cdst->mc_top];
+ m3->mc_ki[csrc->mc_top] = cdst->mc_ki[cdst->mc_top];
+ }
+ }
+ }
+
+ /* Update the parent separators.
+ */
+ if (csrc->mc_ki[csrc->mc_top] == 0) {
+ if (csrc->mc_ki[csrc->mc_top-1] != 0) {
+ if (IS_LEAF2(csrc->mc_pg[csrc->mc_top])) {
+ key.mv_data = LEAF2KEY(csrc->mc_pg[csrc->mc_top], 0, key.mv_size);
+ } else {
+ srcnode = NODEPTR(csrc->mc_pg[csrc->mc_top], 0);
+ key.mv_size = NODEKSZ(srcnode);
+ key.mv_data = NODEKEY(srcnode);
+ }
+ DPRINTF("update separator for source page %zu to [%s]",
+ csrc->mc_pg[csrc->mc_top]->mp_pgno, DKEY(&key));
+ if ((rc = mdb_update_key(csrc->mc_pg[csrc->mc_top-1], csrc->mc_ki[csrc->mc_top-1],
+ &key)) != MDB_SUCCESS)
+ return rc;
+ }
+ if (IS_BRANCH(csrc->mc_pg[csrc->mc_top])) {
+ MDB_val nullkey;
+ nullkey.mv_size = 0;
+ rc = mdb_update_key(csrc->mc_pg[csrc->mc_top], 0, &nullkey);
+ assert(rc == MDB_SUCCESS);
+ }
+ }
+
+ if (cdst->mc_ki[cdst->mc_top] == 0) {
+ if (cdst->mc_ki[cdst->mc_top-1] != 0) {
+ if (IS_LEAF2(csrc->mc_pg[csrc->mc_top])) {
+ key.mv_data = LEAF2KEY(cdst->mc_pg[cdst->mc_top], 0, key.mv_size);
+ } else {
+ srcnode = NODEPTR(cdst->mc_pg[cdst->mc_top], 0);
+ key.mv_size = NODEKSZ(srcnode);
+ key.mv_data = NODEKEY(srcnode);
+ }
+ DPRINTF("update separator for destination page %zu to [%s]",
+ cdst->mc_pg[cdst->mc_top]->mp_pgno, DKEY(&key));
+ if ((rc = mdb_update_key(cdst->mc_pg[cdst->mc_top-1], cdst->mc_ki[cdst->mc_top-1],
+ &key)) != MDB_SUCCESS)
+ return rc;
+ }
+ if (IS_BRANCH(cdst->mc_pg[cdst->mc_top])) {
+ MDB_val nullkey;
+ nullkey.mv_size = 0;
+ rc = mdb_update_key(cdst->mc_pg[cdst->mc_top], 0, &nullkey);
+ assert(rc == MDB_SUCCESS);
+ }
+ }
+
+ return MDB_SUCCESS;
+}
+
+/** Merge one page into another.
+ * The nodes from the page pointed to by \b csrc will
+ * be copied to the page pointed to by \b cdst and then
+ * the \b csrc page will be freed.
+ * @param[in] csrc Cursor pointing to the source page.
+ * @param[in] cdst Cursor pointing to the destination page.
+ */
+static int
+mdb_page_merge(MDB_cursor *csrc, MDB_cursor *cdst)
+{
+ int rc;
+ indx_t i, j;
+ MDB_node *srcnode;
+ MDB_val key, data;
+ unsigned nkeys;
+
+ DPRINTF("merging page %zu into %zu", csrc->mc_pg[csrc->mc_top]->mp_pgno,
+ cdst->mc_pg[cdst->mc_top]->mp_pgno);
+
+ assert(csrc->mc_snum > 1); /* can't merge root page */
+ assert(cdst->mc_snum > 1);
+
+ /* Mark dst as dirty. */
+ if ((rc = mdb_page_touch(cdst)))
+ return rc;
+
+ /* Move all nodes from src to dst.
+ */
+ j = nkeys = NUMKEYS(cdst->mc_pg[cdst->mc_top]);
+ if (IS_LEAF2(csrc->mc_pg[csrc->mc_top])) {
+ key.mv_size = csrc->mc_db->md_pad;
+ key.mv_data = METADATA(csrc->mc_pg[csrc->mc_top]);
+ for (i = 0; i < NUMKEYS(csrc->mc_pg[csrc->mc_top]); i++, j++) {
+ rc = mdb_node_add(cdst, j, &key, NULL, 0, 0);
+ if (rc != MDB_SUCCESS)
+ return rc;
+ key.mv_data = (char *)key.mv_data + key.mv_size;
+ }
+ } else {
+ for (i = 0; i < NUMKEYS(csrc->mc_pg[csrc->mc_top]); i++, j++) {
+ srcnode = NODEPTR(csrc->mc_pg[csrc->mc_top], i);
+
+ key.mv_size = srcnode->mn_ksize;
+ key.mv_data = NODEKEY(srcnode);
+ data.mv_size = NODEDSZ(srcnode);
+ data.mv_data = NODEDATA(srcnode);
+ rc = mdb_node_add(cdst, j, &key, &data, NODEPGNO(srcnode), srcnode->mn_flags);
+ if (rc != MDB_SUCCESS)
+ return rc;
+ }
+ }
+
+ DPRINTF("dst page %zu now has %u keys (%.1f%% filled)",
+ cdst->mc_pg[cdst->mc_top]->mp_pgno, NUMKEYS(cdst->mc_pg[cdst->mc_top]), (float)PAGEFILL(cdst->mc_txn->mt_env, cdst->mc_pg[cdst->mc_top]) / 10);
+
+ /* Unlink the src page from parent and add to free list.
+ */
+ mdb_node_del(csrc->mc_pg[csrc->mc_top-1], csrc->mc_ki[csrc->mc_top-1], 0);
+ if (csrc->mc_ki[csrc->mc_top-1] == 0) {
+ key.mv_size = 0;
+ if ((rc = mdb_update_key(csrc->mc_pg[csrc->mc_top-1], 0, &key)) != MDB_SUCCESS)
+ return rc;
+ }
+
+ mdb_midl_append(&csrc->mc_txn->mt_free_pgs, csrc->mc_pg[csrc->mc_top]->mp_pgno);
+ if (IS_LEAF(csrc->mc_pg[csrc->mc_top]))
+ csrc->mc_db->md_leaf_pages--;
+ else
+ csrc->mc_db->md_branch_pages--;
+ {
+ /* Adjust other cursors pointing to mp */
+ MDB_cursor *m2, *m3;
+ MDB_dbi dbi = csrc->mc_dbi;
+ MDB_page *mp = cdst->mc_pg[cdst->mc_top];
+
+ if (csrc->mc_flags & C_SUB)
+ dbi--;
+
+ for (m2 = csrc->mc_txn->mt_cursors[dbi]; m2; m2=m2->mc_next) {
+ if (m2 == csrc) continue;
+ if (csrc->mc_flags & C_SUB)
+ m3 = &m2->mc_xcursor->mx_cursor;
+ else
+ m3 = m2;
+ if (m3->mc_pg[csrc->mc_top] == csrc->mc_pg[csrc->mc_top]) {
+ m3->mc_pg[csrc->mc_top] = mp;
+ m3->mc_ki[csrc->mc_top] += nkeys;
+ }
+ }
+ }
+ mdb_cursor_pop(csrc);
+
+ return mdb_rebalance(csrc);
+}
+
+/** Copy the contents of a cursor.
+ * @param[in] csrc The cursor to copy from.
+ * @param[out] cdst The cursor to copy to.
+ */
+static void
+mdb_cursor_copy(const MDB_cursor *csrc, MDB_cursor *cdst)
+{
+ unsigned int i;
+
+ cdst->mc_txn = csrc->mc_txn;
+ cdst->mc_dbi = csrc->mc_dbi;
+ cdst->mc_db = csrc->mc_db;
+ cdst->mc_dbx = csrc->mc_dbx;
+ cdst->mc_snum = csrc->mc_snum;
+ cdst->mc_top = csrc->mc_top;
+ cdst->mc_flags = csrc->mc_flags;
+
+ for (i=0; i<csrc->mc_snum; i++) {
+ cdst->mc_pg[i] = csrc->mc_pg[i];
+ cdst->mc_ki[i] = csrc->mc_ki[i];
+ }
+}
+
+/** Rebalance the tree after a delete operation.
+ * @param[in] mc Cursor pointing to the page where rebalancing
+ * should begin.
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_rebalance(MDB_cursor *mc)
+{
+ MDB_node *node;
+ int rc;
+ unsigned int ptop;
+ MDB_cursor mn;
+
+ DPRINTF("rebalancing %s page %zu (has %u keys, %.1f%% full)",
+ IS_LEAF(mc->mc_pg[mc->mc_top]) ? "leaf" : "branch",
+ mc->mc_pg[mc->mc_top]->mp_pgno, NUMKEYS(mc->mc_pg[mc->mc_top]), (float)PAGEFILL(mc->mc_txn->mt_env, mc->mc_pg[mc->mc_top]) / 10);
+
+ if (PAGEFILL(mc->mc_txn->mt_env, mc->mc_pg[mc->mc_top]) >= FILL_THRESHOLD) {
+ DPRINTF("no need to rebalance page %zu, above fill threshold",
+ mc->mc_pg[mc->mc_top]->mp_pgno);
+ return MDB_SUCCESS;
+ }
+
+ if (mc->mc_snum < 2) {
+ MDB_page *mp = mc->mc_pg[0];
+ if (NUMKEYS(mp) == 0) {
+ DPUTS("tree is completely empty");
+ mc->mc_db->md_root = P_INVALID;
+ mc->mc_db->md_depth = 0;
+ mc->mc_db->md_leaf_pages = 0;
+ mdb_midl_append(&mc->mc_txn->mt_free_pgs, mp->mp_pgno);
+ mc->mc_snum = 0;
+ mc->mc_top = 0;
+ {
+ /* Adjust other cursors pointing to mp */
+ MDB_cursor *m2, *m3;
+ MDB_dbi dbi = mc->mc_dbi;
+
+ if (mc->mc_flags & C_SUB)
+ dbi--;
+
+ for (m2 = mc->mc_txn->mt_cursors[dbi]; m2; m2=m2->mc_next) {
+ if (m2 == mc) continue;
+ if (mc->mc_flags & C_SUB)
+ m3 = &m2->mc_xcursor->mx_cursor;
+ else
+ m3 = m2;
+ if (m3->mc_pg[0] == mp) {
+ m3->mc_snum = 0;
+ m3->mc_top = 0;
+ }
+ }
+ }
+ } else if (IS_BRANCH(mp) && NUMKEYS(mp) == 1) {
+ DPUTS("collapsing root page!");
+ mdb_midl_append(&mc->mc_txn->mt_free_pgs, mp->mp_pgno);
+ mc->mc_db->md_root = NODEPGNO(NODEPTR(mp, 0));
+ if ((rc = mdb_page_get(mc->mc_txn, mc->mc_db->md_root,
+ &mc->mc_pg[0])))
+ return rc;
+ mc->mc_db->md_depth--;
+ mc->mc_db->md_branch_pages--;
+ {
+ /* Adjust other cursors pointing to mp */
+ MDB_cursor *m2, *m3;
+ MDB_dbi dbi = mc->mc_dbi;
+
+ if (mc->mc_flags & C_SUB)
+ dbi--;
+
+ for (m2 = mc->mc_txn->mt_cursors[dbi]; m2; m2=m2->mc_next) {
+ if (m2 == mc) continue;
+ if (mc->mc_flags & C_SUB)
+ m3 = &m2->mc_xcursor->mx_cursor;
+ else
+ m3 = m2;
+ if (m3->mc_pg[0] == mp) {
+ m3->mc_pg[0] = mc->mc_pg[0];
+ }
+ }
+ }
+ } else
+ DPUTS("root page doesn't need rebalancing");
+ return MDB_SUCCESS;
+ }
+
+ /* The parent (branch page) must have at least 2 pointers,
+ * otherwise the tree is invalid.
+ */
+ ptop = mc->mc_top-1;
+ assert(NUMKEYS(mc->mc_pg[ptop]) > 1);
+
+ /* Leaf page fill factor is below the threshold.
+ * Try to move keys from left or right neighbor, or
+ * merge with a neighbor page.
+ */
+
+ /* Find neighbors.
+ */
+ mdb_cursor_copy(mc, &mn);
+ mn.mc_xcursor = NULL;
+
+ if (mc->mc_ki[ptop] == 0) {
+ /* We're the leftmost leaf in our parent.
+ */
+ DPUTS("reading right neighbor");
+ mn.mc_ki[ptop]++;
+ node = NODEPTR(mc->mc_pg[ptop], mn.mc_ki[ptop]);
+ if ((rc = mdb_page_get(mc->mc_txn, NODEPGNO(node), &mn.mc_pg[mn.mc_top])))
+ return rc;
+ mn.mc_ki[mn.mc_top] = 0;
+ mc->mc_ki[mc->mc_top] = NUMKEYS(mc->mc_pg[mc->mc_top]);
+ } else {
+ /* There is at least one neighbor to the left.
+ */
+ DPUTS("reading left neighbor");
+ mn.mc_ki[ptop]--;
+ node = NODEPTR(mc->mc_pg[ptop], mn.mc_ki[ptop]);
+ if ((rc = mdb_page_get(mc->mc_txn, NODEPGNO(node), &mn.mc_pg[mn.mc_top])))
+ return rc;
+ mn.mc_ki[mn.mc_top] = NUMKEYS(mn.mc_pg[mn.mc_top]) - 1;
+ mc->mc_ki[mc->mc_top] = 0;
+ }
+
+ DPRINTF("found neighbor page %zu (%u keys, %.1f%% full)",
+ mn.mc_pg[mn.mc_top]->mp_pgno, NUMKEYS(mn.mc_pg[mn.mc_top]), (float)PAGEFILL(mc->mc_txn->mt_env, mn.mc_pg[mn.mc_top]) / 10);
+
+ /* If the neighbor page is above threshold and has at least two
+ * keys, move one key from it.
+ *
+ * Otherwise we should try to merge them.
+ */
+ if (PAGEFILL(mc->mc_txn->mt_env, mn.mc_pg[mn.mc_top]) >= FILL_THRESHOLD && NUMKEYS(mn.mc_pg[mn.mc_top]) >= 2)
+ return mdb_node_move(&mn, mc);
+ else { /* FIXME: if (has_enough_room()) */
+ mc->mc_flags &= ~C_INITIALIZED;
+ if (mc->mc_ki[ptop] == 0)
+ return mdb_page_merge(&mn, mc);
+ else
+ return mdb_page_merge(mc, &mn);
+ }
+}
+
+/** Complete a delete operation started by #mdb_cursor_del(). */
+static int
+mdb_cursor_del0(MDB_cursor *mc, MDB_node *leaf)
+{
+ int rc;
+
+ /* add overflow pages to free list */
+ if (!IS_LEAF2(mc->mc_pg[mc->mc_top]) && F_ISSET(leaf->mn_flags, F_BIGDATA)) {
+ int i, ovpages;
+ pgno_t pg;
+
+ memcpy(&pg, NODEDATA(leaf), sizeof(pg));
+ ovpages = OVPAGES(NODEDSZ(leaf), mc->mc_txn->mt_env->me_psize);
+ for (i=0; i<ovpages; i++) {
+ DPRINTF("freed ov page %zu", pg);
+ mdb_midl_append(&mc->mc_txn->mt_free_pgs, pg);
+ pg++;
+ }
+ }
+ mdb_node_del(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top], mc->mc_db->md_pad);
+ mc->mc_db->md_entries--;
+ rc = mdb_rebalance(mc);
+ if (rc != MDB_SUCCESS)
+ mc->mc_txn->mt_flags |= MDB_TXN_ERROR;
+
+ return rc;
+}
+
+int
+mdb_del(MDB_txn *txn, MDB_dbi dbi,
+ MDB_val *key, MDB_val *data)
+{
+ MDB_cursor mc;
+ MDB_xcursor mx;
+ MDB_cursor_op op;
+ MDB_val rdata, *xdata;
+ int rc, exact;
+ DKBUF;
+
+ assert(key != NULL);
+
+ DPRINTF("====> delete db %u key [%s]", dbi, DKEY(key));
+
+ if (txn == NULL || !dbi || dbi >= txn->mt_numdbs)
+ return EINVAL;
+
+ if (F_ISSET(txn->mt_flags, MDB_TXN_RDONLY)) {
+ return EACCES;
+ }
+
+ if (key->mv_size == 0 || key->mv_size > MAXKEYSIZE) {
+ return EINVAL;
+ }
+
+ mdb_cursor_init(&mc, txn, dbi, &mx);
+
+ exact = 0;
+ if (data) {
+ op = MDB_GET_BOTH;
+ rdata = *data;
+ xdata = &rdata;
+ } else {
+ op = MDB_SET;
+ xdata = NULL;
+ }
+ rc = mdb_cursor_set(&mc, key, xdata, op, &exact);
+ if (rc == 0)
+ rc = mdb_cursor_del(&mc, data ? 0 : MDB_NODUPDATA);
+ return rc;
+}
+
+/** Split a page and insert a new node.
+ * @param[in,out] mc Cursor pointing to the page and desired insertion index.
+ * The cursor will be updated to point to the actual page and index where
+ * the node got inserted after the split.
+ * @param[in] newkey The key for the newly inserted node.
+ * @param[in] newdata The data for the newly inserted node.
+ * @param[in] newpgno The page number, if the new node is a branch node.
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_page_split(MDB_cursor *mc, MDB_val *newkey, MDB_val *newdata, pgno_t newpgno,
+ unsigned int nflags)
+{
+ unsigned int flags;
+ int rc = MDB_SUCCESS, ins_new = 0, new_root = 0;
+ indx_t newindx;
+ pgno_t pgno = 0;
+ unsigned int i, j, split_indx, nkeys, pmax;
+ MDB_node *node;
+ MDB_val sepkey, rkey, xdata, *rdata = &xdata;
+ MDB_page *copy;
+ MDB_page *mp, *rp, *pp;
+ unsigned int ptop;
+ MDB_cursor mn;
+ DKBUF;
+
+ mp = mc->mc_pg[mc->mc_top];
+ newindx = mc->mc_ki[mc->mc_top];
+
+ DPRINTF("-----> splitting %s page %zu and adding [%s] at index %i",
+ IS_LEAF(mp) ? "leaf" : "branch", mp->mp_pgno,
+ DKEY(newkey), mc->mc_ki[mc->mc_top]);
+
+ if (mc->mc_snum < 2) {
+ if ((pp = mdb_page_new(mc, P_BRANCH, 1)) == NULL)
+ return ENOMEM;
+ /* shift current top to make room for new parent */
+ mc->mc_pg[1] = mc->mc_pg[0];
+ mc->mc_ki[1] = mc->mc_ki[0];
+ mc->mc_pg[0] = pp;
+ mc->mc_ki[0] = 0;
+ mc->mc_db->md_root = pp->mp_pgno;
+ DPRINTF("root split! new root = %zu", pp->mp_pgno);
+ mc->mc_db->md_depth++;
+ new_root = 1;
+
+ /* Add left (implicit) pointer. */
+ if ((rc = mdb_node_add(mc, 0, NULL, NULL, mp->mp_pgno, 0)) != MDB_SUCCESS) {
+ /* undo the pre-push */
+ mc->mc_pg[0] = mc->mc_pg[1];
+ mc->mc_ki[0] = mc->mc_ki[1];
+ mc->mc_db->md_root = mp->mp_pgno;
+ mc->mc_db->md_depth--;
+ return rc;
+ }
+ mc->mc_snum = 2;
+ mc->mc_top = 1;
+ ptop = 0;
+ } else {
+ ptop = mc->mc_top-1;
+ DPRINTF("parent branch page is %zu", mc->mc_pg[ptop]->mp_pgno);
+ }
+
+ /* Create a right sibling. */
+ if ((rp = mdb_page_new(mc, mp->mp_flags, 1)) == NULL)
+ return ENOMEM;
+ DPRINTF("new right sibling: page %zu", rp->mp_pgno);
+
+ mdb_cursor_copy(mc, &mn);
+ mn.mc_pg[mn.mc_top] = rp;
+ mn.mc_ki[ptop] = mc->mc_ki[ptop]+1;
+
+ if (nflags & MDB_APPEND) {
+ mn.mc_ki[mn.mc_top] = 0;
+ sepkey = *newkey;
+ nkeys = 0;
+ split_indx = 0;
+ goto newsep;
+ }
+
+ nkeys = NUMKEYS(mp);
+ split_indx = nkeys / 2 + 1;
+
+ if (IS_LEAF2(rp)) {
+ char *split, *ins;
+ int x;
+ unsigned int lsize, rsize, ksize;
+ /* Move half of the keys to the right sibling */
+ copy = NULL;
+ x = mc->mc_ki[mc->mc_top] - split_indx;
+ ksize = mc->mc_db->md_pad;
+ split = LEAF2KEY(mp, split_indx, ksize);
+ rsize = (nkeys - split_indx) * ksize;
+ lsize = (nkeys - split_indx) * sizeof(indx_t);
+ mp->mp_lower -= lsize;
+ rp->mp_lower += lsize;
+ mp->mp_upper += rsize - lsize;
+ rp->mp_upper -= rsize - lsize;
+ sepkey.mv_size = ksize;
+ if (newindx == split_indx) {
+ sepkey.mv_data = newkey->mv_data;
+ } else {
+ sepkey.mv_data = split;
+ }
+ if (x<0) {
+ ins = LEAF2KEY(mp, mc->mc_ki[mc->mc_top], ksize);
+ memcpy(rp->mp_ptrs, split, rsize);
+ sepkey.mv_data = rp->mp_ptrs;
+ memmove(ins+ksize, ins, (split_indx - mc->mc_ki[mc->mc_top]) * ksize);
+ memcpy(ins, newkey->mv_data, ksize);
+ mp->mp_lower += sizeof(indx_t);
+ mp->mp_upper -= ksize - sizeof(indx_t);
+ } else {
+ if (x)
+ memcpy(rp->mp_ptrs, split, x * ksize);
+ ins = LEAF2KEY(rp, x, ksize);
+ memcpy(ins, newkey->mv_data, ksize);
+ memcpy(ins+ksize, split + x * ksize, rsize - x * ksize);
+ rp->mp_lower += sizeof(indx_t);
+ rp->mp_upper -= ksize - sizeof(indx_t);
+ mc->mc_ki[mc->mc_top] = x;
+ mc->mc_pg[mc->mc_top] = rp;
+ }
+ goto newsep;
+ }
+
+ /* For leaf pages, check the split point based on what
+ * fits where, since otherwise add_node can fail.
+ */
+ if (IS_LEAF(mp)) {
+ unsigned int psize, nsize;
+ /* Maximum free space in an empty page */
+ pmax = mc->mc_txn->mt_env->me_psize - PAGEHDRSZ;
+ nsize = mdb_leaf_size(mc->mc_txn->mt_env, newkey, newdata);
+ if (newindx < split_indx) {
+ psize = nsize;
+ for (i=0; i<split_indx; i++) {
+ node = NODEPTR(mp, i);
+ psize += NODESIZE + NODEKSZ(node) + sizeof(indx_t);
+ if (F_ISSET(node->mn_flags, F_BIGDATA))
+ psize += sizeof(pgno_t);
+ else
+ psize += NODEDSZ(node);
+ psize += psize & 1;
+ if (psize > pmax) {
+ split_indx = i;
+ break;
+ }
+ }
+ } else {
+ psize = nsize;
+ for (i=nkeys-1; i>=split_indx; i--) {
+ node = NODEPTR(mp, i);
+ psize += NODESIZE + NODEKSZ(node) + sizeof(indx_t);
+ if (F_ISSET(node->mn_flags, F_BIGDATA))
+ psize += sizeof(pgno_t);
+ else
+ psize += NODEDSZ(node);
+ psize += psize & 1;
+ if (psize > pmax) {
+ split_indx = i+1;
+ break;
+ }
+ }
+ }
+ }
+
+ /* First find the separating key between the split pages.
+ */
+ if (newindx == split_indx) {
+ sepkey.mv_size = newkey->mv_size;
+ sepkey.mv_data = newkey->mv_data;
+ } else {
+ node = NODEPTR(mp, split_indx);
+ sepkey.mv_size = node->mn_ksize;
+ sepkey.mv_data = NODEKEY(node);
+ }
+
+newsep:
+ DPRINTF("separator is [%s]", DKEY(&sepkey));
+
+ /* Copy separator key to the parent.
+ */
+ if (SIZELEFT(mn.mc_pg[ptop]) < mdb_branch_size(mc->mc_txn->mt_env, &sepkey)) {
+ mn.mc_snum--;
+ mn.mc_top--;
+ rc = mdb_page_split(&mn, &sepkey, NULL, rp->mp_pgno, 0);
+
+ /* Right page might now have changed parent.
+ * Check if left page also changed parent.
+ */
+ if (mn.mc_pg[ptop] != mc->mc_pg[ptop] &&
+ mc->mc_ki[ptop] >= NUMKEYS(mc->mc_pg[ptop])) {
+ mc->mc_pg[ptop] = mn.mc_pg[ptop];
+ mc->mc_ki[ptop] = mn.mc_ki[ptop] - 1;
+ }
+ } else {
+ mn.mc_top--;
+ rc = mdb_node_add(&mn, mn.mc_ki[ptop], &sepkey, NULL, rp->mp_pgno, 0);
+ mn.mc_top++;
+ }
+ if (rc != MDB_SUCCESS) {
+ return rc;
+ }
+ if (nflags & MDB_APPEND) {
+ mc->mc_pg[mc->mc_top] = rp;
+ mc->mc_ki[mc->mc_top] = 0;
+ return mdb_node_add(mc, 0, newkey, newdata, newpgno, nflags);
+ }
+ if (IS_LEAF2(rp)) {
+ goto done;
+ }
+
+ /* Move half of the keys to the right sibling. */
+
+ /* grab a page to hold a temporary copy */
+ copy = mdb_page_malloc(mc);
+ if (copy == NULL)
+ return ENOMEM;
+
+ copy->mp_pgno = mp->mp_pgno;
+ copy->mp_flags = mp->mp_flags;
+ copy->mp_lower = PAGEHDRSZ;
+ copy->mp_upper = mc->mc_txn->mt_env->me_psize;
+ mc->mc_pg[mc->mc_top] = copy;
+ for (i = j = 0; i <= nkeys; j++) {
+ if (i == split_indx) {
+ /* Insert in right sibling. */
+ /* Reset insert index for right sibling. */
+ j = (i == newindx && ins_new);
+ mc->mc_pg[mc->mc_top] = rp;
+ }
+
+ if (i == newindx && !ins_new) {
+ /* Insert the original entry that caused the split. */
+ rkey.mv_data = newkey->mv_data;
+ rkey.mv_size = newkey->mv_size;
+ if (IS_LEAF(mp)) {
+ rdata = newdata;
+ } else
+ pgno = newpgno;
+ flags = nflags;
+
+ ins_new = 1;
+
+ /* Update page and index for the new key. */
+ mc->mc_ki[mc->mc_top] = j;
+ } else if (i == nkeys) {
+ break;
+ } else {
+ node = NODEPTR(mp, i);
+ rkey.mv_data = NODEKEY(node);
+ rkey.mv_size = node->mn_ksize;
+ if (IS_LEAF(mp)) {
+ xdata.mv_data = NODEDATA(node);
+ xdata.mv_size = NODEDSZ(node);
+ rdata = &xdata;
+ } else
+ pgno = NODEPGNO(node);
+ flags = node->mn_flags;
+
+ i++;
+ }
+
+ if (!IS_LEAF(mp) && j == 0) {
+ /* First branch index doesn't need key data. */
+ rkey.mv_size = 0;
+ }
+
+ rc = mdb_node_add(mc, j, &rkey, rdata, pgno, flags);
+ }
+
+ nkeys = NUMKEYS(copy);
+ for (i=0; i<nkeys; i++)
+ mp->mp_ptrs[i] = copy->mp_ptrs[i];
+ mp->mp_lower = copy->mp_lower;
+ mp->mp_upper = copy->mp_upper;
+ memcpy(NODEPTR(mp, nkeys-1), NODEPTR(copy, nkeys-1),
+ mc->mc_txn->mt_env->me_psize - copy->mp_upper);
+
+ /* reset back to original page */
+ if (newindx < split_indx) {
+ mc->mc_pg[mc->mc_top] = mp;
+ if (nflags & MDB_RESERVE) {
+ node = NODEPTR(mp, mc->mc_ki[mc->mc_top]);
+ if (!(node->mn_flags & F_BIGDATA))
+ newdata->mv_data = NODEDATA(node);
+ }
+ }
+
+ /* return tmp page to freelist */
+ copy->mp_next = mc->mc_txn->mt_env->me_dpages;
+ mc->mc_txn->mt_env->me_dpages = copy;
+done:
+ {
+ /* Adjust other cursors pointing to mp */
+ MDB_cursor *m2, *m3;
+ MDB_dbi dbi = mc->mc_dbi;
+
+ if (mc->mc_flags & C_SUB)
+ dbi--;
+
+ for (m2 = mc->mc_txn->mt_cursors[dbi]; m2; m2=m2->mc_next) {
+ if (m2 == mc) continue;
+ if (mc->mc_flags & C_SUB)
+ m3 = &m2->mc_xcursor->mx_cursor;
+ else
+ m3 = m2;
+ if (!(m3->mc_flags & C_INITIALIZED))
+ continue;
+ if (new_root) {
+ /* root split */
+ for (i=m3->mc_top; i>0; i--) {
+ m3->mc_ki[i+1] = m3->mc_ki[i];
+ m3->mc_pg[i+1] = m3->mc_pg[i];
+ }
+ m3->mc_ki[0] = mc->mc_ki[0];
+ m3->mc_pg[0] = mc->mc_pg[0];
+ m3->mc_snum++;
+ m3->mc_top++;
+ }
+ if (m3->mc_pg[mc->mc_top] == mp) {
+ if (m3->mc_ki[m3->mc_top] >= split_indx) {
+ m3->mc_pg[m3->mc_top] = rp;
+ m3->mc_ki[m3->mc_top] -= split_indx;
+ }
+ }
+ }
+ }
+ return rc;
+}
+
+int
+mdb_put(MDB_txn *txn, MDB_dbi dbi,
+ MDB_val *key, MDB_val *data, unsigned int flags)
+{
+ MDB_cursor mc;
+ MDB_xcursor mx;
+
+ assert(key != NULL);
+ assert(data != NULL);
+
+ if (txn == NULL || !dbi || dbi >= txn->mt_numdbs)
+ return EINVAL;
+
+ if (F_ISSET(txn->mt_flags, MDB_TXN_RDONLY)) {
+ return EACCES;
+ }
+
+ if (key->mv_size == 0 || key->mv_size > MAXKEYSIZE) {
+ return EINVAL;
+ }
+
+ if ((flags & (MDB_NOOVERWRITE|MDB_NODUPDATA|MDB_RESERVE|MDB_APPEND)) != flags)
+ return EINVAL;
+
+ mdb_cursor_init(&mc, txn, dbi, &mx);
+ return mdb_cursor_put(&mc, key, data, flags);
+}
+
+/** Only a subset of the @ref mdb_env flags can be changed
+ * at runtime. Changing other flags requires closing the environment
+ * and re-opening it with the new flags.
+ */
+#define CHANGEABLE (MDB_NOSYNC)
+int
+mdb_env_set_flags(MDB_env *env, unsigned int flag, int onoff)
+{
+ if ((flag & CHANGEABLE) != flag)
+ return EINVAL;
+ if (onoff)
+ env->me_flags |= flag;
+ else
+ env->me_flags &= ~flag;
+ return MDB_SUCCESS;
+}
+
+int
+mdb_env_get_flags(MDB_env *env, unsigned int *arg)
+{
+ if (!env || !arg)
+ return EINVAL;
+
+ *arg = env->me_flags;
+ return MDB_SUCCESS;
+}
+
+int
+mdb_env_get_path(MDB_env *env, const char **arg)
+{
+ if (!env || !arg)
+ return EINVAL;
+
+ *arg = env->me_path;
+ return MDB_SUCCESS;
+}
+
+/** Common code for #mdb_stat() and #mdb_env_stat().
+ * @param[in] env the environment to operate in.
+ * @param[in] db the #MDB_db record containing the stats to return.
+ * @param[out] arg the address of an #MDB_stat structure to receive the stats.
+ * @return 0, this function always succeeds.
+ */
+static int
+mdb_stat0(MDB_env *env, MDB_db *db, MDB_stat *arg)
+{
+ arg->ms_psize = env->me_psize;
+ arg->ms_depth = db->md_depth;
+ arg->ms_branch_pages = db->md_branch_pages;
+ arg->ms_leaf_pages = db->md_leaf_pages;
+ arg->ms_overflow_pages = db->md_overflow_pages;
+ arg->ms_entries = db->md_entries;
+
+ return MDB_SUCCESS;
+}
+int
+mdb_env_stat(MDB_env *env, MDB_stat *arg)
+{
+ int toggle;
+
+ if (env == NULL || arg == NULL)
+ return EINVAL;
+
+ mdb_env_read_meta(env, &toggle);
+
+ return mdb_stat0(env, &env->me_metas[toggle]->mm_dbs[MAIN_DBI], arg);
+}
+
+/** Set the default comparison functions for a database.
+ * Called immediately after a database is opened to set the defaults.
+ * The user can then override them with #mdb_set_compare() or
+ * #mdb_set_dupsort().
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ */
+static void
+mdb_default_cmp(MDB_txn *txn, MDB_dbi dbi)
+{
+ if (txn->mt_dbs[dbi].md_flags & MDB_REVERSEKEY)
+ txn->mt_dbxs[dbi].md_cmp = mdb_cmp_memnr;
+ else if (txn->mt_dbs[dbi].md_flags & MDB_INTEGERKEY)
+ txn->mt_dbxs[dbi].md_cmp = mdb_cmp_cint;
+ else
+ txn->mt_dbxs[dbi].md_cmp = mdb_cmp_memn;
+
+ if (txn->mt_dbs[dbi].md_flags & MDB_DUPSORT) {
+ if (txn->mt_dbs[dbi].md_flags & MDB_INTEGERDUP) {
+ if (txn->mt_dbs[dbi].md_flags & MDB_DUPFIXED)
+ txn->mt_dbxs[dbi].md_dcmp = mdb_cmp_int;
+ else
+ txn->mt_dbxs[dbi].md_dcmp = mdb_cmp_cint;
+ } else if (txn->mt_dbs[dbi].md_flags & MDB_REVERSEDUP) {
+ txn->mt_dbxs[dbi].md_dcmp = mdb_cmp_memnr;
+ } else {
+ txn->mt_dbxs[dbi].md_dcmp = mdb_cmp_memn;
+ }
+ } else {
+ txn->mt_dbxs[dbi].md_dcmp = NULL;
+ }
+}
+
+int mdb_open(MDB_txn *txn, const char *name, unsigned int flags, MDB_dbi *dbi)
+{
+ MDB_val key, data;
+ MDB_dbi i;
+ MDB_cursor mc;
+ int rc, dbflag, exact;
+ size_t len;
+
+ if (txn->mt_dbxs[FREE_DBI].md_cmp == NULL) {
+ mdb_default_cmp(txn, FREE_DBI);
+ }
+
+ /* main DB? */
+ if (!name) {
+ *dbi = MAIN_DBI;
+ if (flags & (MDB_DUPSORT|MDB_REVERSEKEY|MDB_INTEGERKEY))
+ txn->mt_dbs[MAIN_DBI].md_flags |= (flags & (MDB_DUPSORT|MDB_REVERSEKEY|MDB_INTEGERKEY));
+ mdb_default_cmp(txn, MAIN_DBI);
+ return MDB_SUCCESS;
+ }
+
+ if (txn->mt_dbxs[MAIN_DBI].md_cmp == NULL) {
+ mdb_default_cmp(txn, MAIN_DBI);
+ }
+
+ /* Is the DB already open? */
+ len = strlen(name);
+ for (i=2; i<txn->mt_numdbs; i++) {
+ if (len == txn->mt_dbxs[i].md_name.mv_size &&
+ !strncmp(name, txn->mt_dbxs[i].md_name.mv_data, len)) {
+ *dbi = i;
+ return MDB_SUCCESS;
+ }
+ }
+
+ if (txn->mt_numdbs >= txn->mt_env->me_maxdbs - 1)
+ return ENFILE;
+
+ /* Find the DB info */
+ dbflag = 0;
+ exact = 0;
+ key.mv_size = len;
+ key.mv_data = (void *)name;
+ mdb_cursor_init(&mc, txn, MAIN_DBI, NULL);
+ rc = mdb_cursor_set(&mc, &key, &data, MDB_SET, &exact);
+ if (rc == MDB_SUCCESS) {
+ /* make sure this is actually a DB */
+ MDB_node *node = NODEPTR(mc.mc_pg[mc.mc_top], mc.mc_ki[mc.mc_top]);
+ if (!(node->mn_flags & F_SUBDATA))
+ return EINVAL;
+ } else if (rc == MDB_NOTFOUND && (flags & MDB_CREATE)) {
+ /* Create if requested */
+ MDB_db dummy;
+ data.mv_size = sizeof(MDB_db);
+ data.mv_data = &dummy;
+ memset(&dummy, 0, sizeof(dummy));
+ dummy.md_root = P_INVALID;
+ dummy.md_flags = flags & 0xffff;
+ rc = mdb_cursor_put(&mc, &key, &data, F_SUBDATA);
+ dbflag = DB_DIRTY;
+ }
+
+ /* OK, got info, add to table */
+ if (rc == MDB_SUCCESS) {
+ txn->mt_dbxs[txn->mt_numdbs].md_name.mv_data = strdup(name);
+ txn->mt_dbxs[txn->mt_numdbs].md_name.mv_size = len;
+ txn->mt_dbxs[txn->mt_numdbs].md_rel = NULL;
+ txn->mt_dbflags[txn->mt_numdbs] = dbflag;
+ memcpy(&txn->mt_dbs[txn->mt_numdbs], data.mv_data, sizeof(MDB_db));
+ *dbi = txn->mt_numdbs;
+ txn->mt_env->me_dbs[0][txn->mt_numdbs] = txn->mt_dbs[txn->mt_numdbs];
+ txn->mt_env->me_dbs[1][txn->mt_numdbs] = txn->mt_dbs[txn->mt_numdbs];
+ mdb_default_cmp(txn, txn->mt_numdbs);
+ txn->mt_numdbs++;
+ }
+
+ return rc;
+}
+
+int mdb_stat(MDB_txn *txn, MDB_dbi dbi, MDB_stat *arg)
+{
+ if (txn == NULL || arg == NULL || dbi >= txn->mt_numdbs)
+ return EINVAL;
+
+ return mdb_stat0(txn->mt_env, &txn->mt_dbs[dbi], arg);
+}
+
+void mdb_close(MDB_env *env, MDB_dbi dbi)
+{
+ char *ptr;
+ if (dbi <= MAIN_DBI || dbi >= env->me_numdbs)
+ return;
+ ptr = env->me_dbxs[dbi].md_name.mv_data;
+ env->me_dbxs[dbi].md_name.mv_data = NULL;
+ env->me_dbxs[dbi].md_name.mv_size = 0;
+ free(ptr);
+}
+
+/** Add all the DB's pages to the free list.
+ * @param[in] mc Cursor on the DB to free.
+ * @param[in] subs non-Zero to check for sub-DBs in this DB.
+ * @return 0 on success, non-zero on failure.
+ */
+static int
+mdb_drop0(MDB_cursor *mc, int subs)
+{
+ int rc;
+
+ rc = mdb_page_search(mc, NULL, 0);
+ if (rc == MDB_SUCCESS) {
+ MDB_node *ni;
+ MDB_cursor mx;
+ unsigned int i;
+
+ /* LEAF2 pages have no nodes, cannot have sub-DBs */
+ if (!subs || IS_LEAF2(mc->mc_pg[mc->mc_top]))
+ mdb_cursor_pop(mc);
+
+ mdb_cursor_copy(mc, &mx);
+ while (mc->mc_snum > 0) {
+ if (IS_LEAF(mc->mc_pg[mc->mc_top])) {
+ for (i=0; i<NUMKEYS(mc->mc_pg[mc->mc_top]); i++) {
+ ni = NODEPTR(mc->mc_pg[mc->mc_top], i);
+ if (ni->mn_flags & F_SUBDATA) {
+ mdb_xcursor_init1(mc, ni);
+ rc = mdb_drop0(&mc->mc_xcursor->mx_cursor, 0);
+ if (rc)
+ return rc;
+ }
+ }
+ } else {
+ for (i=0; i<NUMKEYS(mc->mc_pg[mc->mc_top]); i++) {
+ pgno_t pg;
+ ni = NODEPTR(mc->mc_pg[mc->mc_top], i);
+ pg = NODEPGNO(ni);
+ /* free it */
+ mdb_midl_append(&mc->mc_txn->mt_free_pgs, pg);
+ }
+ }
+ if (!mc->mc_top)
+ break;
+ rc = mdb_cursor_sibling(mc, 1);
+ if (rc) {
+ /* no more siblings, go back to beginning
+ * of previous level. (stack was already popped
+ * by mdb_cursor_sibling)
+ */
+ for (i=1; i<mc->mc_top; i++)
+ mc->mc_pg[i] = mx.mc_pg[i];
+ }
+ }
+ /* free it */
+ mdb_midl_append(&mc->mc_txn->mt_free_pgs,
+ mc->mc_db->md_root);
+ }
+ return 0;
+}
+
+int mdb_drop(MDB_txn *txn, MDB_dbi dbi, int del)
+{
+ MDB_cursor *mc;
+ int rc;
+
+ if (!txn || !dbi || dbi >= txn->mt_numdbs)
+ return EINVAL;
+
+ rc = mdb_cursor_open(txn, dbi, &mc);
+ if (rc)
+ return rc;
+
+ rc = mdb_drop0(mc, mc->mc_db->md_flags & MDB_DUPSORT);
+ if (rc)
+ mdb_cursor_close(mc);
+ return rc;
+
+ /* Can't delete the main DB */
+ if (del && dbi > MAIN_DBI) {
+ rc = mdb_del(txn, MAIN_DBI, &mc->mc_dbx->md_name, NULL);
+ if (!rc)
+ mdb_close(txn->mt_env, dbi);
+ } else {
+ txn->mt_dbflags[dbi] |= DB_DIRTY;
+ txn->mt_dbs[dbi].md_depth = 0;
+ txn->mt_dbs[dbi].md_branch_pages = 0;
+ txn->mt_dbs[dbi].md_leaf_pages = 0;
+ txn->mt_dbs[dbi].md_overflow_pages = 0;
+ txn->mt_dbs[dbi].md_entries = 0;
+ txn->mt_dbs[dbi].md_root = P_INVALID;
+ }
+ mdb_cursor_close(mc);
+ return rc;
+}
+
+int mdb_set_compare(MDB_txn *txn, MDB_dbi dbi, MDB_cmp_func *cmp)
+{
+ if (txn == NULL || !dbi || dbi >= txn->mt_numdbs)
+ return EINVAL;
+
+ txn->mt_dbxs[dbi].md_cmp = cmp;
+ return MDB_SUCCESS;
+}
+
+int mdb_set_dupsort(MDB_txn *txn, MDB_dbi dbi, MDB_cmp_func *cmp)
+{
+ if (txn == NULL || !dbi || dbi >= txn->mt_numdbs)
+ return EINVAL;
+
+ txn->mt_dbxs[dbi].md_dcmp = cmp;
+ return MDB_SUCCESS;
+}
+
+int mdb_set_relfunc(MDB_txn *txn, MDB_dbi dbi, MDB_rel_func *rel)
+{
+ if (txn == NULL || !dbi || dbi >= txn->mt_numdbs)
+ return EINVAL;
+
+ txn->mt_dbxs[dbi].md_rel = rel;
+ return MDB_SUCCESS;
+}
+
+int mdb_set_relctx(MDB_txn *txn, MDB_dbi dbi, void *ctx)
+{
+ if (txn == NULL || !dbi || dbi >= txn->mt_numdbs)
+ return EINVAL;
+
+ txn->mt_dbxs[dbi].md_relctx = ctx;
+ return MDB_SUCCESS;
+}
+
+/** @} */
diff --git a/libraries/libmdb/mdb.h b/libraries/libmdb/mdb.h
new file mode 100644
index 0000000000000000000000000000000000000000..9075c41e0c6d929f0556fdca2aed1d9df7e348dd
--- /dev/null
+++ b/libraries/libmdb/mdb.h
@@ -0,0 +1,978 @@
+/** @file mdb.h
+ * @brief memory-mapped database library
+ *
+ * @mainpage MDB Memory-Mapped Database Manager
+ * MDB is a Btree-based database management library modeled loosely on the
+ * BerkeleyDB API, but much simplified. The entire database is exposed
+ * in a read-only memory map, and all data fetches return data directly
+ * from the mapped memory, so no malloc's or memcpy's occur during
+ * data fetches. As such, the library is extremely simple because it
+ * requires no page caching layer of its own, and it is extremely high
+ * performance and memory-efficient. It is also fully transactional with
+ * full ACID semantics, and because the memory map is read-only, the
+ * database integrity cannot be corrupted by stray pointer writes from
+ * application code.
+ *
+ * The library is fully thread-aware and supports concurrent read/write
+ * access from multiple processes and threads. Data pages use a copy-on-
+ * write strategy so no active data pages are ever overwritten, which
+ * also provides resistance to corruption and eliminates the need of any
+ * special recovery procedures after a system crash. Writes are fully
+ * serialized; only one write transaction may be active at a time, which
+ * guarantees that writers can never deadlock. The database structure is
+ * multi-versioned so readers run with no locks; writers cannot block
+ * readers, and readers don't block writers.
+ *
+ * Unlike other well-known database mechanisms which use either write-ahead
+ * transaction logs or append-only data writes, MDB requires no maintenance
+ * during operation. Both write-ahead loggers and append-only databases
+ * require periodic checkpointing and/or compaction of their log or database
+ * files otherwise they grow without bound. MDB tracks free pages within
+ * the database and re-uses them for new write operations, so the database
+ * size does not grow without bound in normal use.
+ *
+ * @author Howard Chu, Symas Corporation.
+ *
+ * @copyright Copyright 2011 Howard Chu, Symas Corp. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted only as authorized by the OpenLDAP
+ * Public License.
+ *
+ * A copy of this license is available in the file LICENSE in the
+ * top-level directory of the distribution or, alternatively, at
+ * <http://www.OpenLDAP.org/license.html>.
+ *
+ * @par Derived From:
+ * This code is derived from btree.c written by Martin Hedenfalk.
+ *
+ * Copyright (c) 2009, 2010 Martin Hedenfalk <martin@bzero.se>
+ *
+ * Permission to use, copy, modify, and distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+ * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+ * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+ */
+#ifndef _MDB_H_
+#define _MDB_H_
+
+#include <sys/types.h>
+
+/** @defgroup public Public API
+ * @{
+ */
+/** @defgroup Version Version Macros
+ * @{
+ */
+/** Library major version */
+#define MDB_VERSION_MAJOR 0
+/** Library minor version */
+#define MDB_VERSION_MINOR 9
+/** Library patch version */
+#define MDB_VERSION_PATCH 0
+
+/** Combine args a,b,c into a single integer for easy version comparisons */
+#define MDB_VERINT(a,b,c) (((a) << 24) | ((b) << 16) | (c))
+
+/** The full library version as a single integer */
+#define MDB_VERSION_FULL \
+ MDB_VERINT(MDB_VERSION_MAJOR,MDB_VERSION_MINOR,MDB_VERSION_PATCH)
+
+/** The release date of this library version */
+#define MDB_VERSION_DATE "September 1, 2011"
+
+/** A stringifier for the version info */
+#define MDB_VERSTR(a,b,c,d) "MDB " #a "." #b "." #c ": (" #d ")"
+
+/** A helper for the stringifier macro */
+#define MDB_VERFOO(a,b,c,d) MDB_VERSTR(a,b,c,d)
+
+/** The full library version as a C string */
+#define MDB_VERSION_STRING \
+ MDB_VERFOO(MDB_VERSION_MAJOR,MDB_VERSION_MINOR,MDB_VERSION_PATCH,MDB_VERSION_DATE)
+/** @} */
+
+/** @brief Opaque structure for a database environment.
+ *
+ * A DB environment supports multiple databases, all residing in the same
+ * shared-memory map.
+ */
+typedef struct MDB_env MDB_env;
+
+/** @brief Opaque structure for a transaction handle.
+ *
+ * All database operations require a transaction handle. Transactions may be
+ * read-only or read-write.
+ */
+typedef struct MDB_txn MDB_txn;
+
+/** @brief A handle for an individual database in the DB environment. */
+typedef unsigned int MDB_dbi;
+
+/** @brief Opaque structure for navigating through a database */
+typedef struct MDB_cursor MDB_cursor;
+
+/** @brief Generic structure used for passing keys and data in and out of the database. */
+typedef struct MDB_val {
+ size_t mv_size; /**< size of the data item */
+ void *mv_data; /**< address of the data item */
+} MDB_val;
+
+/** @brief A callback function used to compare two keys in a database */
+typedef int (MDB_cmp_func)(const MDB_val *a, const MDB_val *b);
+
+/** @brief A callback function used to relocate a position-dependent data item
+ * in a fixed-address database.
+ *
+ * The \b newptr gives the item's desired address in
+ * the memory map, and \b oldptr gives its previous address. The item's actual
+ * data resides at the address in \b item. This callback is expected to walk
+ * through the fields of the record in \b item and modify any
+ * values based at the \b oldptr address to be relative to the \b newptr address.
+ * @param[in,out] item The item that is to be relocated.
+ * @param[in] oldptr The previous address.
+ * @param[in] newptr The new address to relocate to.
+ * @param[in] relctx An application-provided context, set by #mdb_set_relctx().
+ * @todo This feature is currently unimplemented.
+ */
+typedef void (MDB_rel_func)(MDB_val *item, void *oldptr, void *newptr, void *relctx);
+
+/** @defgroup mdb_env Environment Flags
+ * @{
+ */
+ /** mmap at a fixed address */
+#define MDB_FIXEDMAP 0x01
+ /** no environment directory */
+#define MDB_NOSUBDIR 0x02
+ /** don't fsync after commit */
+#define MDB_NOSYNC 0x10000
+ /** read only */
+#define MDB_RDONLY 0x20000
+/** @} */
+
+/** @defgroup mdb_open Database Flags
+ * @{
+ */
+ /** use reverse string keys */
+#define MDB_REVERSEKEY 0x02
+ /** use sorted duplicates */
+#define MDB_DUPSORT 0x04
+ /** numeric keys in native byte order.
+ * The keys must all be of the same size. */
+#define MDB_INTEGERKEY 0x08
+ /** with #MDB_DUPSORT, sorted dup items have fixed size */
+#define MDB_DUPFIXED 0x10
+ /** with #MDB_DUPSORT, dups are numeric in native byte order */
+#define MDB_INTEGERDUP 0x20
+ /** with #MDB_DUPSORT, use reverse string dups */
+#define MDB_REVERSEDUP 0x40
+ /** create DB if not already existing */
+#define MDB_CREATE 0x40000
+/** @} */
+
+/** @defgroup mdb_put Write Flags
+ * @{
+ */
+/** For put: Don't write if the key already exists. */
+#define MDB_NOOVERWRITE 0x10
+/** Only for #MDB_DUPSORT<br>
+ * For put: don't write if the key and data pair already exist.<br>
+ * For mdb_cursor_del: remove all duplicate data items.
+ */
+#define MDB_NODUPDATA 0x20
+/** For mdb_cursor_put: overwrite the current key/data pair */
+#define MDB_CURRENT 0x40
+/** For put: Just reserve space for data, don't copy it. Return a
+ * pointer to the reserved space.
+ */
+#define MDB_RESERVE 0x10000
+/** Data is being appended, don't split full pages. */
+#define MDB_APPEND 0x20000
+/** Store multiple data items in one call. */
+#define MDB_MULTIPLE 0x40000
+/* @} */
+
+/** @brief Cursor Get operations.
+ *
+ * This is the set of all operations for retrieving data
+ * using a cursor.
+ */
+typedef enum MDB_cursor_op {
+ MDB_FIRST, /**< Position at first key/data item */
+ MDB_FIRST_DUP, /**< Position at first data item of current key.
+ Only for #MDB_DUPSORT */
+ MDB_GET_BOTH, /**< Position at key/data pair. Only for #MDB_DUPSORT */
+ MDB_GET_BOTH_RANGE, /**< position at key, nearest data. Only for #MDB_DUPSORT */
+ MDB_GET_MULTIPLE, /**< Return all the duplicate data items at the current
+ cursor position. Only for #MDB_DUPFIXED */
+ MDB_LAST, /**< Position at last key/data item */
+ MDB_LAST_DUP, /**< Position at last data item of current key.
+ Only for #MDB_DUPSORT */
+ MDB_NEXT, /**< Position at next data item */
+ MDB_NEXT_DUP, /**< Position at next data item of current key.
+ Only for #MDB_DUPSORT */
+ MDB_NEXT_MULTIPLE, /**< Return all duplicate data items at the next
+ cursor position. Only for #MDB_DUPFIXED */
+ MDB_NEXT_NODUP, /**< Position at first data item of next key.
+ Only for #MDB_DUPSORT */
+ MDB_PREV, /**< Position at previous data item */
+ MDB_PREV_DUP, /**< Position at previous data item of current key.
+ Only for #MDB_DUPSORT */
+ MDB_PREV_NODUP, /**< Position at last data item of previous key.
+ Only for #MDB_DUPSORT */
+ MDB_SET, /**< Position at specified key */
+ MDB_SET_RANGE /**< Position at first key greater than or equal to specified key. */
+} MDB_cursor_op;
+
+/** @defgroup errors Return Codes
+ *
+ * BerkeleyDB uses -30800 to -30999, we'll go under them
+ * @{
+ */
+ /** Successful result */
+#define MDB_SUCCESS 0
+ /** key/data pair already exists */
+#define MDB_KEYEXIST (-30799)
+ /** key/data pair not found (EOF) */
+#define MDB_NOTFOUND (-30798)
+ /** Requested page not found - this usually indicates corruption */
+#define MDB_PAGE_NOTFOUND (-30797)
+ /** Located page was wrong type */
+#define MDB_CORRUPTED (-30796)
+ /** Update of meta page failed, probably I/O error */
+#define MDB_PANIC (-30795)
+ /** Environment version mismatch */
+#define MDB_VERSION_MISMATCH (-30794)
+/** @} */
+
+/** @brief Statistics for a database in the environment */
+typedef struct MDB_stat {
+ unsigned int ms_psize; /**< Size of a database page.
+ This is currently the same for all databases. */
+ unsigned int ms_depth; /**< Depth (height) of the B-tree */
+ size_t ms_branch_pages; /**< Number of internal (non-leaf) pages */
+ size_t ms_leaf_pages; /**< Number of leaf pages */
+ size_t ms_overflow_pages; /**< Number of overflow pages */
+ size_t ms_entries; /**< Number of data items */
+} MDB_stat;
+
+ /** @brief Return the mdb library version information.
+ *
+ * @param[out] major if non-NULL, the library major version number is copied here
+ * @param[out] minor if non-NULL, the library minor version number is copied here
+ * @param[out] patch if non-NULL, the library patch version number is copied here
+ * @retval "version string" The library version as a string
+ */
+char *mdb_version(int *major, int *minor, int *patch);
+
+ /** @brief Return a string describing a given error code.
+ *
+ * This function is a superset of the ANSI C X3.159-1989 (ANSI C) strerror(3)
+ * function. If the error code is greater than or equal to 0, then the string
+ * returned by the system function strerror(3) is returned. If the error code
+ * is less than 0, an error string corresponding to the MDB library error is
+ * returned. See @ref errors for a list of MDB-specific error codes.
+ * @param[in] err The error code
+ * @retval "error message" The description of the error
+ */
+char *mdb_strerror(int err);
+
+ /** @brief Create an MDB environment handle.
+ *
+ * This function allocates memory for a #MDB_env structure. To release
+ * the allocated memory and discard the handle, call #mdb_env_close().
+ * Before the handle may be used, it must be opened using #mdb_env_open().
+ * Various other options may also need to be set before opening the handle,
+ * e.g. #mdb_env_set_mapsize(), #mdb_env_set_maxreaders(), #mdb_env_set_maxdbs(),
+ * depending on usage requirements.
+ * @param[out] env The address where the new handle will be stored
+ * @return A non-zero error value on failure and 0 on success.
+ */
+int mdb_env_create(MDB_env **env);
+
+ /** @brief Open an environment handle.
+ *
+ * If this function fails, #mdb_env_close() must be called to discard the #MDB_env handle.
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ * @param[in] path The directory in which the database files reside. This
+ * directory must already exist and be writable.
+ * @param[in] flags Special options for this environment. This parameter
+ * must be set to 0 or by bitwise OR'ing together one or more of the
+ * values described here.
+ * <ul>
+ * <li>#MDB_FIXEDMAP
+ * use a fixed address for the mmap region. This flag must be specified
+ * when creating the environment, and is stored persistently in the environment.
+ * If successful, the memory map will always reside at the same virtual address
+ * and pointers used to reference data items in the database will be constant
+ * across multiple invocations. This option may not always work, depending on
+ * how the operating system has allocated memory to shared libraries and other uses.
+ * The feature is highly experimental.
+ * <li>#MDB_NOSUBDIR
+ * By default, MDB creates its environment in a directory whose
+ * pathname is given in \b path, and creates its data and lock files
+ * under that directory. With this option, \b path is used as-is for
+ * the database main data file. The database lock file is the \b path
+ * with "-lock" appended.
+ * <li>#MDB_NOSYNC
+ * Don't perform a synchronous flush after committing a transaction. This means
+ * transactions will exhibit the ACI (atomicity, consistency, and isolation)
+ * properties, but not D (durability); that is database integrity will be
+ * maintained but it is possible some number of the most recently committed
+ * transactions may be undone after a system crash. The number of transactions
+ * at risk is governed by how often the system flushes dirty buffers to disk
+ * and how often #mdb_env_sync() is called. This flag may be changed
+ * at any time using #mdb_env_set_flags().
+ * <li>#MDB_RDONLY
+ * Open the environment in read-only mode. No write operations will be allowed.
+ * </ul>
+ * @param[in] mode The UNIX permissions to set on created files. This parameter
+ * is ignored on Windows.
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>#MDB_VERSION_MISMATCH - the version of the MDB library doesn't match the
+ * version that created the database environment.
+ * <li>EINVAL - the environment file headers are corrupted.
+ * <li>ENOENT - the directory specified by the path parameter doesn't exist.
+ * <li>EACCES - the user didn't have permission to access the environment files.
+ * <li>EAGAIN - the environment was locked by another process.
+ * </ul>
+ */
+int mdb_env_open(MDB_env *env, const char *path, unsigned int flags, mode_t mode);
+
+ /** @brief Return statistics about the MDB environment.
+ *
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ * @param[out] stat The address of an #MDB_stat structure
+ * where the statistics will be copied
+ */
+int mdb_env_stat(MDB_env *env, MDB_stat *stat);
+
+ /** @brief Flush the data buffers to disk.
+ *
+ * Data is always written to disk when #mdb_txn_commit() is called,
+ * but the operating system may keep it buffered. MDB always flushes
+ * the OS buffers upon commit as well, unless the environment was
+ * opened with #MDB_NOSYNC.
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ * @param[in] force If non-zero, force the flush to occur. Otherwise
+ * if the environment has the #MDB_NOSYNC flag set the flushes
+ * will be omitted.
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified.
+ * <li>EIO - an error occurred during synchronization.
+ * </ul>
+ */
+int mdb_env_sync(MDB_env *env, int force);
+
+ /** @brief Close the environment and release the memory map.
+ *
+ * Only a single thread may call this function. All transactions, databases,
+ * and cursors must already be closed before calling this function. Attempts to
+ * use any such handles after calling this function will cause a SIGSEGV.
+ * The environment handle will be freed and must not be used again after this call.
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ */
+void mdb_env_close(MDB_env *env);
+
+ /** @brief Set environment flags.
+ *
+ * This may be used to set some flags that weren't already set during
+ * #mdb_env_open(), or to unset these flags. Currently only the
+ * #MDB_NOSYNC flag setting may be changed with this function.
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ * @param[in] flags The flags to change, bitwise OR'ed together
+ * @param[in] onoff A non-zero value sets the flags, zero clears them.
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_env_set_flags(MDB_env *env, unsigned int flags, int onoff);
+
+ /** @brief Get environment flags.
+ *
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ * @param[out] flags The address of an integer to store the flags
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_env_get_flags(MDB_env *env, unsigned int *flags);
+
+ /** @brief Return the path that was used in #mdb_env_open().
+ *
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ * @param[out] path Address of a string pointer to contain the path. This
+ * is the actual string in the environment, not a copy. It should not be
+ * altered in any way.
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_env_get_path(MDB_env *env, const char **path);
+
+ /** @brief Set the size of the memory map to use for this environment.
+ *
+ * The size should be a multiple of the OS page size. The default is
+ * 10485760 bytes. The size of the memory map is also the maximum size
+ * of the database. The value should be chosen as large as possible,
+ * to accomodate future growth of the database.
+ * This function may only be called after #mdb_env_create() and before #mdb_env_open().
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ * @param[in] size The size in bytes
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified, or the environment is already open.
+ * </ul>
+ */
+int mdb_env_set_mapsize(MDB_env *env, size_t size);
+
+ /** @brief Set the maximum number of threads for the environment.
+ *
+ * This defines the number of slots in the lock table that is used to track readers in the
+ * the environment. The default is 126.
+ * This function may only be called after #mdb_env_create() and before #mdb_env_open().
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ * @param[in] readers The maximum number of threads
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified, or the environment is already open.
+ * </ul>
+ */
+int mdb_env_set_maxreaders(MDB_env *env, unsigned int readers);
+
+ /** @brief Get the maximum number of threads for the environment.
+ *
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ * @param[out] readers Address of an integer to store the number of readers
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_env_get_maxreaders(MDB_env *env, unsigned int *readers);
+
+ /** @brief Set the maximum number of databases for the environment.
+ *
+ * This function is only needed if multiple databases will be used in the
+ * environment. Simpler applications that only use a single database can ignore
+ * this option.
+ * This function may only be called after #mdb_env_create() and before #mdb_env_open().
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ * @param[in] dbs The maximum number of databases
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified, or the environment is already open.
+ * </ul>
+ */
+int mdb_env_set_maxdbs(MDB_env *env, MDB_dbi dbs);
+
+ /** @brief Create a transaction for use with the environment.
+ *
+ * The transaction handle may be discarded using #mdb_txn_abort() or #mdb_txn_commit().
+ * @note Transactions may not span threads, a transaction must only be used by a
+ * single thread.
+ * @note Cursors may not span transactions; each cursor must be opened and closed
+ * within a single transaction.
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ * @param[in] parent If this parameter is non-NULL, the new transaction
+ * will be a nested transaction, with the transaction indicated by \b parent
+ * as its parent. Transactions may be nested to any level. A parent
+ * transaction may not issue any other operations besides mdb_txn_begin,
+ * mdb_txn_abort, or mdb_txn_commit while it has active child transactions.
+ * @param[in] flags Special options for this transaction. This parameter
+ * must be set to 0 or by bitwise OR'ing together one or more of the
+ * values described here.
+ * <ul>
+ * <li>#MDB_RDONLY
+ * This transaction will not perform any write operations.
+ * </ul>
+ * @param[out] txn Address where the new #MDB_txn handle will be stored
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>#MDB_PANIC - a fatal error occurred earlier and the environment
+ * must be shut down.
+ * <li>ENOMEM - out of memory, or a read-only transaction was requested and
+ * the reader lock table is full. See #mdb_env_set_maxreaders().
+ * </ul>
+ */
+int mdb_txn_begin(MDB_env *env, MDB_txn *parent, unsigned int flags, MDB_txn **txn);
+
+ /** @brief Commit all the operations of a transaction into the database.
+ *
+ * All cursors opened within the transaction will be closed by this call. The cursors
+ * and transaction handle will be freed and must not be used again after this call.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified.
+ * <li>ENOSPC - no more disk space.
+ * <li>EIO - a low-level I/O error occurred while writing.
+ * </ul>
+ */
+int mdb_txn_commit(MDB_txn *txn);
+
+ /** @brief Abandon all the operations of the transaction instead of saving them.
+ *
+ * All cursors opened within the transaction will be closed by this call. The cursors
+ * and transaction handle will be freed and must not be used again after this call.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ */
+void mdb_txn_abort(MDB_txn *txn);
+
+ /** @brief Reset a read-only transaction.
+ *
+ * This releases the current reader lock but doesn't free the
+ * transaction handle, allowing it to be used again later by #mdb_txn_renew().
+ * It otherwise has the same effect as #mdb_txn_abort() but saves some memory
+ * allocation/deallocation overhead if a thread is going to start a new
+ * read-only transaction again soon.
+ * All cursors opened within the transaction must be closed before the transaction
+ * is reset.
+ * Reader locks generally don't interfere with writers, but they keep old
+ * versions of database pages allocated. Thus they prevent the old pages
+ * from being reused when writers commit new data, and so under heavy load
+ * the database size may grow much more rapidly than otherwise.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ */
+void mdb_txn_reset(MDB_txn *txn);
+
+ /** @brief Renew a read-only transaction.
+ *
+ * This acquires a new reader lock for a transaction handle that had been
+ * released by #mdb_txn_reset(). It must be called before a reset transaction
+ * may be used again.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>#MDB_PANIC - a fatal error occurred earlier and the environment
+ * must be shut down.
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_txn_renew(MDB_txn *txn);
+
+ /** @brief Open a database in the environment.
+ *
+ * The database handle may be discarded by calling #mdb_close(). Only
+ * one thread should call this function; it is not mutex-protected in
+ * a read-only transaction.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] name The name of the database to open. If only a single
+ * database is needed in the enviroment, this value may be NULL.
+ * @param[in] flags Special options for this database. This parameter
+ * must be set to 0 or by bitwise OR'ing together one or more of the
+ * values described here.
+ * <ul>
+ * <li>#MDB_REVERSEKEY
+ * Keys are strings to be compared in reverse order, from the end
+ * of the strings to the beginning. By default, Keys are treated as strings and
+ * compared from beginning to end.
+ * <li>#MDB_DUPSORT
+ * Duplicate keys may be used in the database. (Or, from another perspective,
+ * keys may have multiple data items, stored in sorted order.) By default
+ * keys must be unique and may have only a single data item.
+ * <li>#MDB_INTEGERKEY
+ * Keys are binary integers in native byte order. Setting this option
+ * requires all keys to be the same size, typically sizeof(int)
+ * or sizeof(size_t).
+ * <li>#MDB_DUPFIXED
+ * This flag may only be used in combination with #MDB_DUPSORT. This option
+ * tells the library that the data items for this database are all the same
+ * size, which allows further optimizations in storage and retrieval. When
+ * all data items are the same size, the #MDB_GET_MULTIPLE and #MDB_NEXT_MULTIPLE
+ * cursor operations may be used to retrieve multiple items at once.
+ * <li>#MDB_INTEGERDUP
+ * This option specifies that duplicate data items are also integers, and
+ * should be sorted as such.
+ * <li>#MDB_REVERSEDUP
+ * This option specifies that duplicate data items should be compared as
+ * strings in reverse order.
+ * <li>#MDB_CREATE
+ * Create the named database if it doesn't exist. This option is not
+ * allowed in a read-only transaction or a read-only environment.
+ * </ul>
+ * @param[out] dbi Address where the new #MDB_dbi handle will be stored
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>#MDB_NOTFOUND - the specified database doesn't exist in the environment
+ * and #MDB_CREATE was not specified.
+ * <li>ENFILE - too many databases have been opened. See #mdb_env_set_maxdbs().
+ * </ul>
+ */
+int mdb_open(MDB_txn *txn, const char *name, unsigned int flags, MDB_dbi *dbi);
+
+ /** @brief Retrieve statistics for a database.
+ *
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ * @param[out] stat The address of an #MDB_stat structure
+ * where the statistics will be copied
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_stat(MDB_txn *txn, MDB_dbi dbi, MDB_stat *stat);
+
+ /** @brief Close a database handle.
+ *
+ * This call is not mutex protected. Handles should only be closed by
+ * a single thread, and only if no other threads are going to reference
+ * the database handle any further.
+ * @param[in] env An environment handle returned by #mdb_env_create()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ */
+void mdb_close(MDB_env *env, MDB_dbi dbi);
+
+ /** @brief Delete a database and/or free all its pages.
+ *
+ * If the \b del parameter is non-zero the DB handle will be closed
+ * and the DB will be deleted.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ * @param[in] del non-zero to delete the DB from the environment,
+ * otherwise just free its pages.
+ * @return A non-zero error value on failure and 0 on success.
+ */
+int mdb_drop(MDB_txn *txn, MDB_dbi dbi, int del);
+
+ /** @brief Set a custom key comparison function for a database.
+ *
+ * The comparison function is called whenever it is necessary to compare a
+ * key specified by the application with a key currently stored in the database.
+ * If no comparison function is specified, and no speAGAINcial key flags were specified
+ * with #mdb_open(), the keys are compared lexically, with shorter keys collating
+ * before longer keys.
+ * @warning This function must be called before any data access functions are used,
+ * otherwise data corruption may occur. The same comparison function must be used by every
+ * program accessing the database, every time the database is used.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ * @param[in] cmp A #MDB_cmp_func function
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_set_compare(MDB_txn *txn, MDB_dbi dbi, MDB_cmp_func *cmp);
+
+ /** @brief Set a custom data comparison function for a #MDB_DUPSORT database.
+ *
+ * This comparison function is called whenever it is necessary to compare a data
+ * item specified by the application with a data item currently stored in the database.
+ * This function only takes effect if the database was opened with the #MDB_DUPSORT
+ * flag.
+ * If no comparison function is specified, and no special key flags were specified
+ * with #mdb_open(), the data items are compared lexically, with shorter items collating
+ * before longer items.
+ * @warning This function must be called before any data access functions are used,
+ * otherwise data corruption may occur. The same comparison function must be used by every
+ * program accessing the database, every time the database is used.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ * @param[in] cmp A #MDB_cmp_func function
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_set_dupsort(MDB_txn *txn, MDB_dbi dbi, MDB_cmp_func *cmp);
+
+ /** @brief Set a relocation function for a #MDB_FIXEDMAP database.
+ *
+ * @todo The relocation function is called whenever it is necessary to move the data
+ * of an item to a different position in the database (e.g. through tree
+ * balancing operations, shifts as a result of adds or deletes, etc.). It is
+ * intended to allow address/position-dependent data items to be stored in
+ * a database in an environment opened with the #MDB_FIXEDMAP option.
+ * Currently the relocation feature is unimplemented and setting
+ * this function has no effect.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ * @param[in] rel A #MDB_rel_func function
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_set_relfunc(MDB_txn *txn, MDB_dbi dbi, MDB_rel_func *rel);
+
+ /** @brief Set a context pointer for a #MDB_FIXEDMAP database's relocation function.
+ *
+ * See #mdb_set_relfunc and #MDB_rel_func for more details.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ * @param[in] ctx An arbitrary pointer for whatever the application needs.
+ * It will be passed to the callback function set by #mdb_set_relfunc
+ * as its \b relctx parameter whenever the callback is invoked.
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_set_relctx(MDB_txn *txn, MDB_dbi dbi, void *ctx);
+
+ /** @brief Get items from a database.
+ *
+ * This function retrieves key/data pairs from the database. The address
+ * and length of the data associated with the specified \b key are returned
+ * in the structure to which \b data refers.
+ * If the database supports duplicate keys (#MDB_DUPSORT) then the
+ * first data item for the key will be returned. Retrieval of other
+ * items requires the use of #mdb_cursor_get().
+ *
+ * @note The memory pointed to by the returned values is owned by the
+ * database. The caller need not dispose of the memory, and may not
+ * modify it in any way. For values returned in a read-only transaction
+ * any modification attempts will cause a SIGSEGV.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ * @param[in] key The key to search for in the database
+ * @param[out] data The data corresponding to the key
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>#MDB_NOTFOUND - the key was not in the database.
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_get(MDB_txn *txn, MDB_dbi dbi, MDB_val *key, MDB_val *data);
+
+ /** @brief Store items into a database.
+ *
+ * This function stores key/data pairs in the database. The default behavior
+ * is to enter the new key/data pair, replacing any previously existing key
+ * if duplicates are disallowed, or adding a duplicate data item if
+ * duplicates are allowed (#MDB_DUPSORT).
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ * @param[in] key The key to store in the database
+ * @param[in,out] data The data to store
+ * @param[in] flags Special options for this operation. This parameter
+ * must be set to 0 or by bitwise OR'ing together one or more of the
+ * values described here.
+ * <ul>
+ * <li>#MDB_NODUPDATA - enter the new key/data pair only if it does not
+ * already appear in the database. This flag may only be specified
+ * if the database was opened with #MDB_DUPSORT. The function will
+ * return #MDB_KEYEXIST if the key/data pair already appears in the
+ * database.
+ * <li>#MDB_NOOVERWRITE - enter the new key/data pair only if the key
+ * does not already appear in the database. The function will return
+ * #MDB_KEYEXIST if the key already appears in the database, even if
+ * the database supports duplicates (#MDB_DUPSORT). The \b data
+ * parameter will be set to point to the existing item.
+ * </ul>
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EACCESS - an attempt was made to write in a read-only transaction.
+ * <li>EINVAL - an invalid parameter was specified.
+ * <li>ENOMEM - the database is full, see #mdb_env_set_mapsize().
+ * </ul>
+ */
+int mdb_put(MDB_txn *txn, MDB_dbi dbi, MDB_val *key, MDB_val *data,
+ unsigned int flags);
+
+ /** @brief Delete items from a database.
+ *
+ * This function removes key/data pairs from the database.
+ * If the database does not support sorted duplicate data items
+ * (#MDB_DUPSORT) the data parameter is ignored.
+ * If the database supports sorted duplicates and the data parameter
+ * is NULL, all of the duplicate data items for the key will be
+ * deleted. Otherwise, if the data parameter is non-NULL
+ * only the matching data item will be deleted.
+ * This function will return #MDB_NOTFOUND if the specified key/data
+ * pair is not in the database.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ * @param[in] key The key to delete from the database
+ * @param[in] data The data to delete
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EACCESS - an attempt was made to write in a read-only transaction.
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_del(MDB_txn *txn, MDB_dbi dbi, MDB_val *key, MDB_val *data);
+
+ /** @brief Create a cursor handle.
+ *
+ * Cursors are associated with a specific transaction and database and
+ * may not span threads.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ * @param[out] cursor Address where the new #MDB_cursor handle will be stored
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_cursor_open(MDB_txn *txn, MDB_dbi dbi, MDB_cursor **cursor);
+
+ /** @brief Close a cursor handle.
+ *
+ * The cursor handle will be freed and must not be used again after this call.
+ * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
+ */
+void mdb_cursor_close(MDB_cursor *cursor);
+
+ /** @brief Return the cursor's transaction handle.
+ *
+ * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
+ */
+MDB_txn *mdb_cursor_txn(MDB_cursor *cursor);
+
+ /** @brief Return the cursor's database handle.
+ *
+ * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
+ */
+MDB_dbi mdb_cursor_dbi(MDB_cursor *cursor);
+
+ /** @brief Retrieve by cursor.
+ *
+ * This function retrieves key/data pairs from the database. The address and length
+ * of the key are returned in the object to which \b key refers (except for the
+ * case of the #MDB_SET option, in which the \b key object is unchanged), and
+ * the address and length of the data are returned in the object to which \b data
+ * refers.
+ * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
+ * @param[in,out] key The key for a retrieved item
+ * @param[in,out] data The data of a retrieved item
+ * @param[in] op A cursor operation #MDB_cursor_op
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>#MDB_NOTFOUND - no matching key found.
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_cursor_get(MDB_cursor *cursor, MDB_val *key, MDB_val *data,
+ MDB_cursor_op op);
+
+ /** @brief Store by cursor.
+ *
+ * This function stores key/data pairs into the database.
+ * If the function fails for any reason, the state of the cursor will be
+ * unchanged. If the function succeeds and an item is inserted into the
+ * database, the cursor is always positioned to refer to the newly inserted item.
+ * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
+ * @param[in] key The key operated on.
+ * @param[in] data The data operated on.
+ * @param[in] flags Options for this operation. This parameter
+ * must be set to 0 or one of the values described here.
+ * <ul>
+ * <li>#MDB_CURRENT - overwrite the data of the key/data pair to which
+ * the cursor refers with the specified data item. The \b key
+ * parameter is ignored.
+ * <li>#MDB_NODUPDATA - enter the new key/data pair only if it does not
+ * already appear in the database. This flag may only be specified
+ * if the database was opened with #MDB_DUPSORT. The function will
+ * return #MDB_KEYEXIST if the key/data pair already appears in the
+ * database.
+ * <li>#MDB_NOOVERWRITE - enter the new key/data pair only if the key
+ * does not already appear in the database. The function will return
+ * #MDB_KEYEXIST if the key already appears in the database, even if
+ * the database supports duplicates (#MDB_DUPSORT).
+ * </ul>
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EACCES - an attempt was made to modify a read-only database.
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_cursor_put(MDB_cursor *cursor, MDB_val *key, MDB_val *data,
+ unsigned int flags);
+
+ /** @brief Delete current key/data pair
+ *
+ * This function deletes the key/data pair to which the cursor refers.
+ * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
+ * @param[in] flags Options for this operation. This parameter
+ * must be set to 0 or one of the values described here.
+ * <ul>
+ * <li>#MDB_NODUPDATA - delete all of the data items for the current key.
+ * This flag may only be specified if the database was opened with #MDB_DUPSORT.
+ * </ul>
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EACCES - an attempt was made to modify a read-only database.
+ * <li>EINVAL - an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_cursor_del(MDB_cursor *cursor, unsigned int flags);
+
+ /** @brief Return count of duplicates for current key.
+ *
+ * This call is only valid on databases that support sorted duplicate
+ * data items #MDB_DUPSORT.
+ * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
+ * @param[out] countp Address where the count will be stored
+ * @return A non-zero error value on failure and 0 on success. Some possible
+ * errors are:
+ * <ul>
+ * <li>EINVAL - cursor is not initialized, or an invalid parameter was specified.
+ * </ul>
+ */
+int mdb_cursor_count(MDB_cursor *cursor, size_t *countp);
+
+ /** @brief Compare two data items according to a particular database.
+ *
+ * This returns a comparison as if the two data items were keys in the
+ * specified database.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ * @param[in] a The first item to compare
+ * @param[in] b The second item to compare
+ * @return < 0 if a < b, 0 if a == b, > 0 if a > b
+ */
+int mdb_cmp(MDB_txn *txn, MDB_dbi dbi, const MDB_val *a, const MDB_val *b);
+
+ /** @brief Compare two data items according to a particular database.
+ *
+ * This returns a comparison as if the two items were data items of
+ * a sorted duplicates #MDB_DUPSORT database.
+ * @param[in] txn A transaction handle returned by #mdb_txn_begin()
+ * @param[in] dbi A database handle returned by #mdb_open()
+ * @param[in] a The first item to compare
+ * @param[in] b The second item to compare
+ * @return < 0 if a < b, 0 if a == b, > 0 if a > b
+ */
+int mdb_dcmp(MDB_txn *txn, MDB_dbi dbi, const MDB_val *a, const MDB_val *b);
+/** @} */
+#endif /* _MDB_H_ */
diff --git a/libraries/libmdb/mdb_stat.c b/libraries/libmdb/mdb_stat.c
new file mode 100644
index 0000000000000000000000000000000000000000..a5f484d0637ac873c66532bc4dac85a2e6f9b2c1
--- /dev/null
+++ b/libraries/libmdb/mdb_stat.c
@@ -0,0 +1,64 @@
+/* mdb_stat.c - memory-mapped database status tool */
+/*
+ * Copyright 2011 Howard Chu, Symas Corp.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted only as authorized by the OpenLDAP
+ * Public License.
+ *
+ * A copy of this license is available in the file LICENSE in the
+ * top-level directory of the distribution or, alternatively, at
+ * <http://www.OpenLDAP.org/license.html>.
+ */
+#include <stdio.h>
+#include <stdlib.h>
+#include <time.h>
+#include "mdb.h"
+
+int main(int argc,char * argv[])
+{
+ int rc;
+ MDB_env *env;
+ MDB_txn *txn;
+ MDB_dbi dbi;
+ MDB_stat mst;
+ char *envname = argv[1];
+ char *subname = NULL;
+
+ rc = mdb_env_create(&env);
+
+ if (argc > 2) {
+ mdb_env_set_maxdbs(env, 4);
+ subname = argv[2];
+ }
+
+ rc = mdb_env_open(env, envname, MDB_RDONLY, 0);
+ if (rc) {
+ printf("mdb_env_open failed, error %d\n", rc);
+ exit(1);
+ }
+ rc = mdb_txn_begin(env, NULL, 1, &txn);
+ if (rc) {
+ printf("mdb_txn_begin failed, error %d\n", rc);
+ exit(1);
+ }
+ rc = mdb_open(txn, subname, 0, &dbi);
+ if (rc) {
+ printf("mdb_open failed, error %d\n", rc);
+ exit(1);
+ }
+
+ rc = mdb_stat(txn, dbi, &mst);
+ printf("Page size: %u\n", mst.ms_psize);
+ printf("Tree depth: %u\n", mst.ms_depth);
+ printf("Branch pages: %zu\n", mst.ms_branch_pages);
+ printf("Leaf pages: %zu\n", mst.ms_leaf_pages);
+ printf("Overflow pages: %zu\n", mst.ms_overflow_pages);
+ printf("Entries: %zu\n", mst.ms_entries);
+ mdb_close(env, dbi);
+ mdb_txn_abort(txn);
+ mdb_env_close(env);
+
+ return 0;
+}
diff --git a/libraries/libmdb/midl.c b/libraries/libmdb/midl.c
new file mode 100644
index 0000000000000000000000000000000000000000..500811f7d216dc4251067e5479756b001f364316
--- /dev/null
+++ b/libraries/libmdb/midl.c
@@ -0,0 +1,315 @@
+/** @file midl.c
+ * @brief ldap bdb back-end ID List functions */
+/* $OpenLDAP$ */
+/* This work is part of OpenLDAP Software <http://www.openldap.org/>.
+ *
+ * Copyright 2000-2011 The OpenLDAP Foundation.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted only as authorized by the OpenLDAP
+ * Public License.
+ *
+ * A copy of this license is available in the file LICENSE in the
+ * top-level directory of the distribution or, alternatively, at
+ * <http://www.OpenLDAP.org/license.html>.
+ */
+
+#include <limits.h>
+#include <string.h>
+#include <stdlib.h>
+#include <sys/types.h>
+#include <assert.h>
+#include "midl.h"
+
+/** @defgroup internal MDB Internals
+ * @{
+ */
+/** @defgroup idls ID List Management
+ * @{
+ */
+#define CMP(x,y) ( (x) < (y) ? -1 : (x) > (y) )
+
+#if 0 /* superseded by append/sort */
+static unsigned mdb_midl_search( IDL ids, ID id )
+{
+ /*
+ * binary search of id in ids
+ * if found, returns position of id
+ * if not found, returns first position greater than id
+ */
+ unsigned base = 0;
+ unsigned cursor = 1;
+ int val = 0;
+ unsigned n = ids[0];
+
+ while( 0 < n ) {
+ unsigned pivot = n >> 1;
+ cursor = base + pivot + 1;
+ val = CMP( ids[cursor], id );
+
+ if( val < 0 ) {
+ n = pivot;
+
+ } else if ( val > 0 ) {
+ base = cursor;
+ n -= pivot + 1;
+
+ } else {
+ return cursor;
+ }
+ }
+
+ if( val > 0 ) {
+ ++cursor;
+ }
+ return cursor;
+}
+
+int mdb_midl_insert( IDL ids, ID id )
+{
+ unsigned x, i;
+
+ if (MDB_IDL_IS_RANGE( ids )) {
+ /* if already in range, treat as a dup */
+ if (id >= MDB_IDL_RANGE_FIRST(ids) && id <= MDB_IDL_RANGE_LAST(ids))
+ return -1;
+ if (id < MDB_IDL_RANGE_FIRST(ids))
+ ids[1] = id;
+ else if (id > MDB_IDL_RANGE_LAST(ids))
+ ids[2] = id;
+ return 0;
+ }
+
+ x = mdb_midl_search( ids, id );
+ assert( x > 0 );
+
+ if( x < 1 ) {
+ /* internal error */
+ return -2;
+ }
+
+ if ( x <= ids[0] && ids[x] == id ) {
+ /* duplicate */
+ assert(0);
+ return -1;
+ }
+
+ if ( ++ids[0] >= MDB_IDL_DB_MAX ) {
+ if( id < ids[1] ) {
+ ids[1] = id;
+ ids[2] = ids[ids[0]-1];
+ } else if ( ids[ids[0]-1] < id ) {
+ ids[2] = id;
+ } else {
+ ids[2] = ids[ids[0]-1];
+ }
+ ids[0] = NOID;
+
+ } else {
+ /* insert id */
+ for (i=ids[0]; i>x; i--)
+ ids[i] = ids[i-1];
+ ids[x] = id;
+ }
+
+ return 0;
+}
+#endif
+
+IDL mdb_midl_alloc()
+{
+ IDL ids = malloc((MDB_IDL_UM_MAX+1) * sizeof(ID));
+ *ids++ = MDB_IDL_UM_MAX;
+ return ids;
+}
+
+void mdb_midl_free(IDL ids)
+{
+ free(ids-1);
+}
+
+int mdb_midl_shrink( IDL *idp )
+{
+ IDL ids = *idp;
+ if (ids[-1] > MDB_IDL_UM_MAX) {
+ ids = realloc(ids, (MDB_IDL_UM_MAX+1) * sizeof(ID));
+ *ids++ = MDB_IDL_UM_MAX;
+ *idp = ids;
+ return 1;
+ }
+ return 0;
+}
+
+int mdb_midl_append( IDL *idp, ID id )
+{
+ IDL ids = *idp;
+ /* Too big? */
+ if (ids[0] >= ids[-1]) {
+ IDL idn = ids-1;
+ /* grow it */
+ idn = realloc(idn, (*idn + MDB_IDL_UM_MAX + 1) * sizeof(ID));
+ if (!idn)
+ return -1;
+ *idn++ += MDB_IDL_UM_MAX;
+ ids = idn;
+ *idp = ids;
+ }
+ ids[0]++;
+ ids[ids[0]] = id;
+ return 0;
+}
+
+int mdb_midl_append_list( IDL *idp, IDL app )
+{
+ IDL ids = *idp;
+ /* Too big? */
+ if (ids[0] + app[0] >= ids[-1]) {
+ IDL idn = ids-1;
+ /* grow it */
+ idn = realloc(idn, (*idn + app[-1]) * sizeof(ID));
+ if (!idn)
+ return -1;
+ *idn++ += app[-1];
+ ids = idn;
+ *idp = ids;
+ }
+ memcpy(&ids[ids[0]+1], &app[1], app[0] * sizeof(ID));
+ ids[0] += app[0];
+ return 0;
+}
+
+/* Quicksort + Insertion sort for small arrays */
+
+#define SMALL 8
+#define SWAP(a,b) { itmp=(a); (a)=(b); (b)=itmp; }
+
+void
+mdb_midl_sort( IDL ids )
+{
+ /* Max possible depth of int-indexed tree * 2 items/level */
+ int istack[sizeof(int)*CHAR_BIT * 2];
+ int i,j,k,l,ir,jstack;
+ ID a, itmp;
+
+ ir = ids[0];
+ l = 1;
+ jstack = 0;
+ for(;;) {
+ if (ir - l < SMALL) { /* Insertion sort */
+ for (j=l+1;j<=ir;j++) {
+ a = ids[j];
+ for (i=j-1;i>=1;i--) {
+ if (ids[i] >= a) break;
+ ids[i+1] = ids[i];
+ }
+ ids[i+1] = a;
+ }
+ if (jstack == 0) break;
+ ir = istack[jstack--];
+ l = istack[jstack--];
+ } else {
+ k = (l + ir) >> 1; /* Choose median of left, center, right */
+ SWAP(ids[k], ids[l+1]);
+ if (ids[l] < ids[ir]) {
+ SWAP(ids[l], ids[ir]);
+ }
+ if (ids[l+1] < ids[ir]) {
+ SWAP(ids[l+1], ids[ir]);
+ }
+ if (ids[l] < ids[l+1]) {
+ SWAP(ids[l], ids[l+1]);
+ }
+ i = l+1;
+ j = ir;
+ a = ids[l+1];
+ for(;;) {
+ do i++; while(ids[i] > a);
+ do j--; while(ids[j] < a);
+ if (j < i) break;
+ SWAP(ids[i],ids[j]);
+ }
+ ids[l+1] = ids[j];
+ ids[j] = a;
+ jstack += 2;
+ if (ir-i+1 >= j-1) {
+ istack[jstack] = ir;
+ istack[jstack-1] = i;
+ ir = j-1;
+ } else {
+ istack[jstack] = j-1;
+ istack[jstack-1] = l;
+ l = i;
+ }
+ }
+ }
+}
+
+unsigned mdb_mid2l_search( ID2L ids, ID id )
+{
+ /*
+ * binary search of id in ids
+ * if found, returns position of id
+ * if not found, returns first position greater than id
+ */
+ unsigned base = 0;
+ unsigned cursor = 1;
+ int val = 0;
+ unsigned n = ids[0].mid;
+
+ while( 0 < n ) {
+ unsigned pivot = n >> 1;
+ cursor = base + pivot + 1;
+ val = CMP( id, ids[cursor].mid );
+
+ if( val < 0 ) {
+ n = pivot;
+
+ } else if ( val > 0 ) {
+ base = cursor;
+ n -= pivot + 1;
+
+ } else {
+ return cursor;
+ }
+ }
+
+ if( val > 0 ) {
+ ++cursor;
+ }
+ return cursor;
+}
+
+int mdb_mid2l_insert( ID2L ids, ID2 *id )
+{
+ unsigned x, i;
+
+ x = mdb_mid2l_search( ids, id->mid );
+ assert( x > 0 );
+
+ if( x < 1 ) {
+ /* internal error */
+ return -2;
+ }
+
+ if ( x <= ids[0].mid && ids[x].mid == id->mid ) {
+ /* duplicate */
+ return -1;
+ }
+
+ if ( ids[0].mid >= MDB_IDL_UM_MAX ) {
+ /* too big */
+ return -2;
+
+ } else {
+ /* insert id */
+ ids[0].mid++;
+ for (i=ids[0].mid; i>x; i--)
+ ids[i] = ids[i-1];
+ ids[x] = *id;
+ }
+
+ return 0;
+}
+/** @} */
+/** @} */
diff --git a/libraries/libmdb/midl.h b/libraries/libmdb/midl.h
new file mode 100644
index 0000000000000000000000000000000000000000..4c0e615f1a704f72ef114dd511c3564e9746a549
--- /dev/null
+++ b/libraries/libmdb/midl.h
@@ -0,0 +1,182 @@
+/** @file midl.h
+ * @brief ldap bdb back-end ID List header file.
+ *
+ * This file was originally part of back-bdb but has been
+ * modified for use in libmdb. Most of the macros defined
+ * in this file are unused, just left over from the original.
+ *
+ * This file is only used internally in libmdb and its definitions
+ * are not exposed publicly.
+ */
+/* $OpenLDAP$ */
+/* This work is part of OpenLDAP Software <http://www.openldap.org/>.
+ *
+ * Copyright 2000-2011 The OpenLDAP Foundation.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted only as authorized by the OpenLDAP
+ * Public License.
+ *
+ * A copy of this license is available in the file LICENSE in the
+ * top-level directory of the distribution or, alternatively, at
+ * <http://www.OpenLDAP.org/license.html>.
+ */
+
+#ifndef _MDB_MIDL_H_
+#define _MDB_MIDL_H_
+
+#include <stddef.h>
+
+/** @defgroup internal MDB Internals
+ * @{
+ */
+
+/** @defgroup idls ID List Management
+ * @{
+ */
+ /** A generic ID number. These were entryIDs in back-bdb.
+ * It should be the largest integer type supported on a machine.
+ * It should be equal to the size of a pointer.
+ */
+typedef size_t ID;
+
+ /** An IDL is an ID List, a sorted array of IDs. The first
+ * element of the array is a counter for how many actual
+ * IDs are in the list. In the original back-bdb code, IDLs are
+ * sorted in ascending order. For libmdb IDLs are sorted in
+ * descending order.
+ */
+typedef ID *IDL;
+
+#define NOID (~(ID)0)
+
+/* IDL sizes - likely should be even bigger
+ * limiting factors: sizeof(ID), thread stack size
+ */
+#define MDB_IDL_LOGN 16 /* DB_SIZE is 2^16, UM_SIZE is 2^17 */
+#define MDB_IDL_DB_SIZE (1<<MDB_IDL_LOGN)
+#define MDB_IDL_UM_SIZE (1<<(MDB_IDL_LOGN+1))
+#define MDB_IDL_UM_SIZEOF (MDB_IDL_UM_SIZE * sizeof(ID))
+
+#define MDB_IDL_DB_MAX (MDB_IDL_DB_SIZE-1)
+
+#define MDB_IDL_UM_MAX (MDB_IDL_UM_SIZE-1)
+
+#define MDB_IDL_IS_RANGE(ids) ((ids)[0] == NOID)
+#define MDB_IDL_RANGE_SIZE (3)
+#define MDB_IDL_RANGE_SIZEOF (MDB_IDL_RANGE_SIZE * sizeof(ID))
+#define MDB_IDL_SIZEOF(ids) ((MDB_IDL_IS_RANGE(ids) \
+ ? MDB_IDL_RANGE_SIZE : ((ids)[0]+1)) * sizeof(ID))
+
+#define MDB_IDL_RANGE_FIRST(ids) ((ids)[1])
+#define MDB_IDL_RANGE_LAST(ids) ((ids)[2])
+
+#define MDB_IDL_RANGE( ids, f, l ) \
+ do { \
+ (ids)[0] = NOID; \
+ (ids)[1] = (f); \
+ (ids)[2] = (l); \
+ } while(0)
+
+#define MDB_IDL_ZERO(ids) \
+ do { \
+ (ids)[0] = 0; \
+ (ids)[1] = 0; \
+ (ids)[2] = 0; \
+ } while(0)
+
+#define MDB_IDL_IS_ZERO(ids) ( (ids)[0] == 0 )
+#define MDB_IDL_IS_ALL( range, ids ) ( (ids)[0] == NOID \
+ && (ids)[1] <= (range)[1] && (range)[2] <= (ids)[2] )
+
+#define MDB_IDL_CPY( dst, src ) (memcpy( dst, src, MDB_IDL_SIZEOF( src ) ))
+
+#define MDB_IDL_ID( bdb, ids, id ) MDB_IDL_RANGE( ids, id, ((bdb)->bi_lastid) )
+#define MDB_IDL_ALL( bdb, ids ) MDB_IDL_RANGE( ids, 1, ((bdb)->bi_lastid) )
+
+#define MDB_IDL_FIRST( ids ) ( (ids)[1] )
+#define MDB_IDL_LAST( ids ) ( MDB_IDL_IS_RANGE(ids) \
+ ? (ids)[2] : (ids)[(ids)[0]] )
+
+#define MDB_IDL_N( ids ) ( MDB_IDL_IS_RANGE(ids) \
+ ? ((ids)[2]-(ids)[1])+1 : (ids)[0] )
+
+#if 0 /* superseded by append/sort */
+ /** Insert an ID into an IDL.
+ * @param[in,out] ids The IDL to insert into.
+ * @param[in] id The ID to insert.
+ * @return 0 on success, -1 if the ID was already present in the IDL.
+ */
+int mdb_midl_insert( IDL ids, ID id );
+#endif
+
+ /** Allocate an IDL.
+ * Allocates memory for an IDL of a default size.
+ * @return IDL on success, NULL on failure.
+ */
+IDL mdb_midl_alloc();
+
+ /** Free an IDL.
+ * @param[in] ids The IDL to free.
+ */
+void mdb_midl_free(IDL ids);
+
+ /** Shrink an IDL.
+ * Return the IDL to the default size if it has grown larger.
+ * @param[in,out] idp Address of the IDL to shrink.
+ * @return 0 on no change, non-zero if shrunk.
+ */
+int mdb_midl_shrink(IDL *idp);
+
+ /** Append an ID onto an IDL.
+ * @param[in,out] idp Address of the IDL to append to.
+ * @param[in] id The ID to append.
+ * @return 0 on success, -1 if the IDL is too large.
+ */
+int mdb_midl_append( IDL *idp, ID id );
+
+ /** Append an IDL onto an IDL.
+ * @param[in,out] idp Address of the IDL to append to.
+ * @param[in] app The IDL to append.
+ * @return 0 on success, -1 if the IDL is too large.
+ */
+int mdb_midl_append_list( IDL *idp, IDL app );
+
+ /** Sort an IDL.
+ * @param[in,out] ids The IDL to sort.
+ */
+void mdb_midl_sort( IDL ids );
+
+ /** An ID2 is an ID/pointer pair.
+ */
+typedef struct ID2 {
+ ID mid; /**< The ID */
+ void *mptr; /**< The pointer */
+} ID2;
+
+ /** An ID2L is an ID2 List, a sorted array of ID2s.
+ * The first element's \b mid member is a count of how many actual
+ * elements are in the array. The \b mptr member of the first element is unused.
+ * The array is sorted in ascending order by \b mid.
+ */
+typedef ID2 *ID2L;
+
+ /** Search for an ID in an ID2L.
+ * @param[in] ids The ID2L to search.
+ * @param[in] id The ID to search for.
+ * @return The index of the first ID2 whose \b mid member is greater than or equal to \b id.
+ */
+unsigned mdb_mid2l_search( ID2L ids, ID id );
+
+
+ /** Insert an ID2 into a ID2L.
+ * @param[in,out] ids The ID2L to insert into.
+ * @param[in] id The ID2 to insert.
+ * @return 0 on success, -1 if the ID was already present in the ID2L.
+ */
+int mdb_mid2l_insert( ID2L ids, ID2 *id );
+
+/** @} */
+/** @} */
+#endif /* _MDB_MIDL_H_ */
diff --git a/libraries/libmdb/mtest.c b/libraries/libmdb/mtest.c
new file mode 100644
index 0000000000000000000000000000000000000000..bb5ec816b628b25a967489ce5ca1559409a1cb1c
--- /dev/null
+++ b/libraries/libmdb/mtest.c
@@ -0,0 +1,155 @@
+/* mtest.c - memory-mapped database tester/toy */
+/*
+ * Copyright 2011 Howard Chu, Symas Corp.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted only as authorized by the OpenLDAP
+ * Public License.
+ *
+ * A copy of this license is available in the file LICENSE in the
+ * top-level directory of the distribution or, alternatively, at
+ * <http://www.OpenLDAP.org/license.html>.
+ */
+#define _XOPEN_SOURCE 500 /* srandom(), random() */
+#include <stdio.h>
+#include <stdlib.h>
+#include <time.h>
+#include "mdb.h"
+
+int main(int argc,char * argv[])
+{
+ int i = 0, j = 0, rc;
+ MDB_env *env;
+ MDB_dbi dbi;
+ MDB_val key, data;
+ MDB_txn *txn;
+ MDB_stat mst;
+ MDB_cursor *cursor;
+ int count;
+ int *values;
+ char sval[32];
+
+ srandom(time(NULL));
+
+ count = (random()%384) + 64;
+ values = (int *)malloc(count*sizeof(int));
+
+ for(i = 0;i<count;i++) {
+ values[i] = random()%1024;
+ }
+
+ rc = mdb_env_create(&env);
+ rc = mdb_env_set_mapsize(env, 10485760);
+ rc = mdb_env_open(env, "./testdb", MDB_FIXEDMAP /*|MDB_NOSYNC*/, 0664);
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ rc = mdb_open(txn, NULL, 0, &dbi);
+
+ key.mv_size = sizeof(int);
+ key.mv_data = sval;
+ data.mv_size = sizeof(sval);
+ data.mv_data = sval;
+
+ printf("Adding %d values\n", count);
+ for (i=0;i<count;i++) {
+ sprintf(sval, "%03x %d foo bar", values[i], values[i]);
+ rc = mdb_put(txn, dbi, &key, &data, MDB_NOOVERWRITE);
+ if (rc) j++;
+ }
+ if (j) printf("%d duplicates skipped\n", j);
+ rc = mdb_txn_commit(txn);
+ rc = mdb_env_stat(env, &mst);
+
+ rc = mdb_txn_begin(env, NULL, 1, &txn);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0) {
+ printf("key: %p %.*s, data: %p %.*s\n",
+ key.mv_data, (int) key.mv_size, (char *) key.mv_data,
+ data.mv_data, (int) data.mv_size, (char *) data.mv_data);
+ }
+ mdb_cursor_close(cursor);
+ mdb_txn_abort(txn);
+
+ j=0;
+ key.mv_data = sval;
+ for (i= count - 1; i > -1; i-= (random()%5)) {
+ j++;
+ txn=NULL;
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ sprintf(sval, "%03x ", values[i]);
+ rc = mdb_del(txn, dbi, &key, NULL);
+ if (rc) {
+ j--;
+ mdb_txn_abort(txn);
+ } else {
+ rc = mdb_txn_commit(txn);
+ }
+ }
+ free(values);
+ printf("Deleted %d values\n", j);
+
+ rc = mdb_env_stat(env, &mst);
+ rc = mdb_txn_begin(env, NULL, 1, &txn);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ printf("Cursor next\n");
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+ printf("Cursor prev\n");
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_PREV)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+#if 0
+ /* write ops aren't coordinated with cursors,
+ * this stuff all breaks
+ */
+ printf("Deleting with cursor\n");
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ rc = mdb_cursor_open(db, txn, &cur2);
+ for (i=0; i<50; i++) {
+ rc = mdb_cursor_get(cur2, &key, &data, MDB_NEXT);
+ printf("key: %p %.*s, data: %p %.*s\n",
+ key.mv_data, (int) key.mv_size, (char *) key.mv_data,
+ data.mv_data, (int) data.mv_size, (char *) data.mv_data);
+ rc = mdb_del(db, txn, &key, NULL);
+ }
+
+ printf("Restarting cursor in txn\n");
+ rc = mdb_cursor_get(cur2, &key, &data, MDB_FIRST);
+ printf("key: %p %.*s, data: %p %.*s\n",
+ key.mv_data, (int) key.mv_size, (char *) key.mv_data,
+ data.mv_data, (int) data.mv_size, (char *) data.mv_data);
+ for (i=0; i<32; i++) {
+ rc = mdb_cursor_get(cur2, &key, &data, MDB_NEXT);
+ printf("key: %p %.*s, data: %p %.*s\n",
+ key.mv_data, (int) key.mv_size, (char *) key.mv_data,
+ data.mv_data, (int) data.mv_size, (char *) data.mv_data);
+ }
+ mdb_cursor_close(cur2);
+ rc = mdb_txn_commit(txn);
+ mdb_cursor_close(cursor);
+ printf("Restarting cursor outside txn\n");
+ rc = mdb_cursor_open(db, NULL, &cursor);
+ rc = mdb_cursor_get(cursor, &key, &data, MDB_FIRST);
+ printf("key: %p %.*s, data: %p %.*s\n",
+ key.mv_data, (int) key.mv_size, (char *) key.mv_data,
+ data.mv_data, (int) data.mv_size, (char *) data.mv_data);
+ for (i=0; i<32; i++) {
+ rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT);
+ printf("key: %p %.*s, data: %p %.*s\n",
+ key.mv_data, (int) key.mv_size, (char *) key.mv_data,
+ data.mv_data, (int) data.mv_size, (char *) data.mv_data);
+ }
+#endif
+ mdb_cursor_close(cursor);
+ mdb_close(env, dbi);
+
+ mdb_txn_abort(txn);
+ mdb_env_close(env);
+
+ return 0;
+}
diff --git a/libraries/libmdb/mtest2.c b/libraries/libmdb/mtest2.c
new file mode 100644
index 0000000000000000000000000000000000000000..a0e9914417784fc631282a47409f1b2893133aab
--- /dev/null
+++ b/libraries/libmdb/mtest2.c
@@ -0,0 +1,117 @@
+/* mtest2.c - memory-mapped database tester/toy */
+/*
+ * Copyright 2011 Howard Chu, Symas Corp.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted only as authorized by the OpenLDAP
+ * Public License.
+ *
+ * A copy of this license is available in the file LICENSE in the
+ * top-level directory of the distribution or, alternatively, at
+ * <http://www.OpenLDAP.org/license.html>.
+ */
+
+/* Just like mtest.c, but using a subDB instead of the main DB */
+
+#define _XOPEN_SOURCE 500 /* srandom(), random() */
+#include <stdio.h>
+#include <stdlib.h>
+#include <time.h>
+#include "mdb.h"
+
+int main(int argc,char * argv[])
+{
+ int i = 0, j = 0, rc;
+ MDB_env *env;
+ MDB_dbi dbi;
+ MDB_val key, data;
+ MDB_txn *txn;
+ MDB_stat mst;
+ MDB_cursor *cursor;
+ int count;
+ int *values;
+ char sval[32];
+
+ srandom(time(NULL));
+
+ count = (random()%384) + 64;
+ values = (int *)malloc(count*sizeof(int));
+
+ for(i = 0;i<count;i++) {
+ values[i] = random()%1024;
+ }
+
+ rc = mdb_env_create(&env);
+ rc = mdb_env_set_mapsize(env, 10485760);
+ rc = mdb_env_set_maxdbs(env, 4);
+ rc = mdb_env_open(env, "./testdb", MDB_FIXEDMAP|MDB_NOSYNC, 0664);
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ rc = mdb_open(txn, "id1", MDB_CREATE, &dbi);
+
+ key.mv_size = sizeof(int);
+ key.mv_data = sval;
+ data.mv_size = sizeof(sval);
+ data.mv_data = sval;
+
+ printf("Adding %d values\n", count);
+ for (i=0;i<count;i++) {
+ sprintf(sval, "%03x %d foo bar", values[i], values[i]);
+ rc = mdb_put(txn, dbi, &key, &data, MDB_NOOVERWRITE);
+ if (rc) j++;
+ }
+ if (j) printf("%d duplicates skipped\n", j);
+ rc = mdb_txn_commit(txn);
+ rc = mdb_env_stat(env, &mst);
+
+ rc = mdb_txn_begin(env, NULL, 1, &txn);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0) {
+ printf("key: %p %.*s, data: %p %.*s\n",
+ key.mv_data, (int) key.mv_size, (char *) key.mv_data,
+ data.mv_data, (int) data.mv_size, (char *) data.mv_data);
+ }
+ mdb_cursor_close(cursor);
+ mdb_txn_abort(txn);
+
+ j=0;
+ key.mv_data = sval;
+ for (i= count - 1; i > -1; i-= (random()%5)) {
+ j++;
+ txn=NULL;
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ sprintf(sval, "%03x ", values[i]);
+ rc = mdb_del(txn, dbi, &key, NULL);
+ if (rc) {
+ j--;
+ mdb_txn_abort(txn);
+ } else {
+ rc = mdb_txn_commit(txn);
+ }
+ }
+ free(values);
+ printf("Deleted %d values\n", j);
+
+ rc = mdb_env_stat(env, &mst);
+ rc = mdb_txn_begin(env, NULL, 1, &txn);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ printf("Cursor next\n");
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+ printf("Cursor prev\n");
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_PREV)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+ mdb_cursor_close(cursor);
+ mdb_close(env, dbi);
+
+ mdb_txn_abort(txn);
+ mdb_env_close(env);
+
+ return 0;
+}
diff --git a/libraries/libmdb/mtest3.c b/libraries/libmdb/mtest3.c
new file mode 100644
index 0000000000000000000000000000000000000000..eb9723ba2e58148cdcb769d15765a56a73a3f65c
--- /dev/null
+++ b/libraries/libmdb/mtest3.c
@@ -0,0 +1,127 @@
+/* mtest3.c - memory-mapped database tester/toy */
+/*
+ * Copyright 2011 Howard Chu, Symas Corp.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted only as authorized by the OpenLDAP
+ * Public License.
+ *
+ * A copy of this license is available in the file LICENSE in the
+ * top-level directory of the distribution or, alternatively, at
+ * <http://www.OpenLDAP.org/license.html>.
+ */
+
+/* Tests for sorted duplicate DBs */
+#define _XOPEN_SOURCE 500 /* srandom(), random() */
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+#include "mdb.h"
+
+int main(int argc,char * argv[])
+{
+ int i = 0, j = 0, rc;
+ MDB_env *env;
+ MDB_dbi dbi;
+ MDB_val key, data;
+ MDB_txn *txn;
+ MDB_stat mst;
+ MDB_cursor *cursor;
+ int count;
+ int *values;
+ char sval[32];
+ char kval[sizeof(int)];
+
+ srandom(time(NULL));
+
+ memset(sval, 0, sizeof(sval));
+
+ count = (random()%384) + 64;
+ values = (int *)malloc(count*sizeof(int));
+
+ for(i = 0;i<count;i++) {
+ values[i] = random()%1024;
+ }
+
+ rc = mdb_env_create(&env);
+ rc = mdb_env_set_mapsize(env, 10485760);
+ rc = mdb_env_set_maxdbs(env, 4);
+ rc = mdb_env_open(env, "./testdb", MDB_FIXEDMAP|MDB_NOSYNC, 0664);
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ rc = mdb_open(txn, "id2", MDB_CREATE|MDB_DUPSORT, &dbi);
+
+ key.mv_size = sizeof(int);
+ key.mv_data = kval;
+ data.mv_size = sizeof(sval);
+ data.mv_data = sval;
+
+ printf("Adding %d values\n", count);
+ for (i=0;i<count;i++) {
+ if (!(i & 0x0f))
+ sprintf(kval, "%03x", values[i]);
+ sprintf(sval, "%03x %d foo bar", values[i], values[i]);
+ rc = mdb_put(txn, dbi, &key, &data, MDB_NODUPDATA);
+ if (rc) j++;
+ }
+ if (j) printf("%d duplicates skipped\n", j);
+ rc = mdb_txn_commit(txn);
+ rc = mdb_env_stat(env, &mst);
+
+ rc = mdb_txn_begin(env, NULL, 1, &txn);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0) {
+ printf("key: %p %.*s, data: %p %.*s\n",
+ key.mv_data, (int) key.mv_size, (char *) key.mv_data,
+ data.mv_data, (int) data.mv_size, (char *) data.mv_data);
+ }
+ mdb_cursor_close(cursor);
+ mdb_txn_abort(txn);
+
+ j=0;
+
+ for (i= count - 1; i > -1; i-= (random()%5)) {
+ j++;
+ txn=NULL;
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ sprintf(kval, "%03x", values[i & ~0x0f]);
+ sprintf(sval, "%03x %d foo bar", values[i], values[i]);
+ key.mv_size = sizeof(int);
+ key.mv_data = kval;
+ data.mv_size = sizeof(sval);
+ data.mv_data = sval;
+ rc = mdb_del(txn, dbi, &key, &data);
+ if (rc) {
+ j--;
+ mdb_txn_abort(txn);
+ } else {
+ rc = mdb_txn_commit(txn);
+ }
+ }
+ free(values);
+ printf("Deleted %d values\n", j);
+
+ rc = mdb_env_stat(env, &mst);
+ rc = mdb_txn_begin(env, NULL, 1, &txn);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ printf("Cursor next\n");
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+ printf("Cursor prev\n");
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_PREV)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+ mdb_cursor_close(cursor);
+ mdb_close(env, dbi);
+
+ mdb_txn_abort(txn);
+ mdb_env_close(env);
+
+ return 0;
+}
diff --git a/libraries/libmdb/mtest4.c b/libraries/libmdb/mtest4.c
new file mode 100644
index 0000000000000000000000000000000000000000..b2f5d931c61f5d1c42117e61637dc3b0de892b4b
--- /dev/null
+++ b/libraries/libmdb/mtest4.c
@@ -0,0 +1,161 @@
+/* mtest4.c - memory-mapped database tester/toy */
+/*
+ * Copyright 2011 Howard Chu, Symas Corp.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted only as authorized by the OpenLDAP
+ * Public License.
+ *
+ * A copy of this license is available in the file LICENSE in the
+ * top-level directory of the distribution or, alternatively, at
+ * <http://www.OpenLDAP.org/license.html>.
+ */
+
+/* Tests for sorted duplicate DBs with fixed-size keys */
+#define _XOPEN_SOURCE 500 /* srandom(), random() */
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+#include "mdb.h"
+
+int main(int argc,char * argv[])
+{
+ int i = 0, j = 0, rc;
+ MDB_env *env;
+ MDB_dbi dbi;
+ MDB_val key, data;
+ MDB_txn *txn;
+ MDB_stat mst;
+ MDB_cursor *cursor;
+ int count;
+ int *values;
+ char sval[8];
+ char kval[sizeof(int)];
+
+ memset(sval, 0, sizeof(sval));
+
+ count = 510;
+ values = (int *)malloc(count*sizeof(int));
+
+ for(i = 0;i<count;i++) {
+ values[i] = i*5;
+ }
+
+ rc = mdb_env_create(&env);
+ rc = mdb_env_set_mapsize(env, 10485760);
+ rc = mdb_env_set_maxdbs(env, 4);
+ rc = mdb_env_open(env, "./testdb", MDB_FIXEDMAP|MDB_NOSYNC, 0664);
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ rc = mdb_open(txn, "id2", MDB_CREATE|MDB_DUPSORT|MDB_DUPFIXED, &dbi);
+
+ key.mv_size = sizeof(int);
+ key.mv_data = kval;
+ data.mv_size = sizeof(sval);
+ data.mv_data = sval;
+
+ printf("Adding %d values\n", count);
+ strcpy(kval, "001");
+ for (i=0;i<count;i++) {
+ sprintf(sval, "%07x", values[i]);
+ rc = mdb_put(txn, dbi, &key, &data, MDB_NODUPDATA);
+ if (rc) j++;
+ }
+ if (j) printf("%d duplicates skipped\n", j);
+ rc = mdb_txn_commit(txn);
+ rc = mdb_env_stat(env, &mst);
+
+ /* there should be one full page of dups now.
+ */
+ rc = mdb_txn_begin(env, NULL, 1, &txn);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0) {
+ printf("key: %p %.*s, data: %p %.*s\n",
+ key.mv_data, (int) key.mv_size, (char *) key.mv_data,
+ data.mv_data, (int) data.mv_size, (char *) data.mv_data);
+ }
+ mdb_cursor_close(cursor);
+ mdb_txn_abort(txn);
+
+ /* test all 3 branches of split code:
+ * 1: new key in lower half
+ * 2: new key at split point
+ * 3: new key in upper half
+ */
+
+ key.mv_size = sizeof(int);
+ key.mv_data = kval;
+ data.mv_size = sizeof(sval);
+ data.mv_data = sval;
+
+ sprintf(sval, "%07x", values[3]+1);
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ rc = mdb_put(txn, dbi, &key, &data, MDB_NODUPDATA);
+ mdb_txn_abort(txn);
+
+ sprintf(sval, "%07x", values[255]+1);
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ rc = mdb_put(txn, dbi, &key, &data, MDB_NODUPDATA);
+ mdb_txn_abort(txn);
+
+ sprintf(sval, "%07x", values[500]+1);
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ rc = mdb_put(txn, dbi, &key, &data, MDB_NODUPDATA);
+ rc = mdb_txn_commit(txn);
+
+ /* Try MDB_NEXT_MULTIPLE */
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT_MULTIPLE)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+ mdb_cursor_close(cursor);
+ mdb_txn_abort(txn);
+ j=0;
+
+ for (i= count - 1; i > -1; i-= (random()%3)) {
+ j++;
+ txn=NULL;
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ sprintf(sval, "%07x", values[i]);
+ key.mv_size = sizeof(int);
+ key.mv_data = kval;
+ data.mv_size = sizeof(sval);
+ data.mv_data = sval;
+ rc = mdb_del(txn, dbi, &key, &data);
+ if (rc) {
+ j--;
+ mdb_txn_abort(txn);
+ } else {
+ rc = mdb_txn_commit(txn);
+ }
+ }
+ free(values);
+ printf("Deleted %d values\n", j);
+
+ rc = mdb_env_stat(env, &mst);
+ rc = mdb_txn_begin(env, NULL, 1, &txn);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ printf("Cursor next\n");
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+ printf("Cursor prev\n");
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_PREV)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+ mdb_cursor_close(cursor);
+ mdb_close(env, dbi);
+
+ mdb_txn_abort(txn);
+ mdb_env_close(env);
+
+ return 0;
+}
diff --git a/libraries/libmdb/mtest5.c b/libraries/libmdb/mtest5.c
new file mode 100644
index 0000000000000000000000000000000000000000..c63402ed6bf28ac032fab81e265b95c45fab4dba
--- /dev/null
+++ b/libraries/libmdb/mtest5.c
@@ -0,0 +1,129 @@
+/* mtest5.c - memory-mapped database tester/toy */
+/*
+ * Copyright 2011 Howard Chu, Symas Corp.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted only as authorized by the OpenLDAP
+ * Public License.
+ *
+ * A copy of this license is available in the file LICENSE in the
+ * top-level directory of the distribution or, alternatively, at
+ * <http://www.OpenLDAP.org/license.html>.
+ */
+
+/* Tests for sorted duplicate DBs using cursor_put */
+#define _XOPEN_SOURCE 500 /* srandom(), random() */
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+#include "mdb.h"
+
+int main(int argc,char * argv[])
+{
+ int i = 0, j = 0, rc;
+ MDB_env *env;
+ MDB_dbi dbi;
+ MDB_val key, data;
+ MDB_txn *txn;
+ MDB_stat mst;
+ MDB_cursor *cursor;
+ int count;
+ int *values;
+ char sval[32];
+ char kval[sizeof(int)];
+
+ srandom(time(NULL));
+
+ memset(sval, 0, sizeof(sval));
+
+ count = (random()%384) + 64;
+ values = (int *)malloc(count*sizeof(int));
+
+ for(i = 0;i<count;i++) {
+ values[i] = random()%1024;
+ }
+
+ rc = mdb_env_create(&env);
+ rc = mdb_env_set_mapsize(env, 10485760);
+ rc = mdb_env_set_maxdbs(env, 4);
+ rc = mdb_env_open(env, "./testdb", MDB_FIXEDMAP|MDB_NOSYNC, 0664);
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ rc = mdb_open(txn, "id2", MDB_CREATE|MDB_DUPSORT, &dbi);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+
+ key.mv_size = sizeof(int);
+ key.mv_data = kval;
+ data.mv_size = sizeof(sval);
+ data.mv_data = sval;
+
+ printf("Adding %d values\n", count);
+ for (i=0;i<count;i++) {
+ if (!(i & 0x0f))
+ sprintf(kval, "%03x", values[i]);
+ sprintf(sval, "%03x %d foo bar", values[i], values[i]);
+ rc = mdb_cursor_put(cursor, &key, &data, MDB_NODUPDATA);
+ if (rc) j++;
+ }
+ if (j) printf("%d duplicates skipped\n", j);
+ mdb_cursor_close(cursor);
+ rc = mdb_txn_commit(txn);
+ rc = mdb_env_stat(env, &mst);
+
+ rc = mdb_txn_begin(env, NULL, 1, &txn);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0) {
+ printf("key: %p %.*s, data: %p %.*s\n",
+ key.mv_data, (int) key.mv_size, (char *) key.mv_data,
+ data.mv_data, (int) data.mv_size, (char *) data.mv_data);
+ }
+ mdb_cursor_close(cursor);
+ mdb_txn_abort(txn);
+
+ j=0;
+
+ for (i= count - 1; i > -1; i-= (random()%5)) {
+ j++;
+ txn=NULL;
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ sprintf(kval, "%03x", values[i & ~0x0f]);
+ sprintf(sval, "%03x %d foo bar", values[i], values[i]);
+ key.mv_size = sizeof(int);
+ key.mv_data = kval;
+ data.mv_size = sizeof(sval);
+ data.mv_data = sval;
+ rc = mdb_del(txn, dbi, &key, &data);
+ if (rc) {
+ j--;
+ mdb_txn_abort(txn);
+ } else {
+ rc = mdb_txn_commit(txn);
+ }
+ }
+ free(values);
+ printf("Deleted %d values\n", j);
+
+ rc = mdb_env_stat(env, &mst);
+ rc = mdb_txn_begin(env, NULL, 1, &txn);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ printf("Cursor next\n");
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+ printf("Cursor prev\n");
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_PREV)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+ mdb_cursor_close(cursor);
+ mdb_close(env, dbi);
+
+ mdb_txn_abort(txn);
+ mdb_env_close(env);
+
+ return 0;
+}
diff --git a/libraries/libmdb/mtest6.c b/libraries/libmdb/mtest6.c
new file mode 100644
index 0000000000000000000000000000000000000000..a5bf3a2119779c49914c397126f8a76bda797056
--- /dev/null
+++ b/libraries/libmdb/mtest6.c
@@ -0,0 +1,131 @@
+/* mtest6.c - memory-mapped database tester/toy */
+/*
+ * Copyright 2011 Howard Chu, Symas Corp.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted only as authorized by the OpenLDAP
+ * Public License.
+ *
+ * A copy of this license is available in the file LICENSE in the
+ * top-level directory of the distribution or, alternatively, at
+ * <http://www.OpenLDAP.org/license.html>.
+ */
+
+/* Tests for DB splits and merges */
+#define _XOPEN_SOURCE 500 /* srandom(), random() */
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+#include "mdb.h"
+
+char dkbuf[1024];
+
+int main(int argc,char * argv[])
+{
+ int i = 0, j = 0, rc;
+ MDB_env *env;
+ MDB_dbi dbi;
+ MDB_val key, data;
+ MDB_txn *txn;
+ MDB_stat mst;
+ MDB_cursor *cursor;
+ int count;
+ int *values;
+ long kval;
+ char *sval;
+
+ srandom(time(NULL));
+
+ rc = mdb_env_create(&env);
+ rc = mdb_env_set_mapsize(env, 10485760);
+ rc = mdb_env_set_maxdbs(env, 4);
+ rc = mdb_env_open(env, "./testdb", MDB_FIXEDMAP|MDB_NOSYNC, 0664);
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ rc = mdb_open(txn, "id2", MDB_CREATE|MDB_INTEGERKEY, &dbi);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ rc = mdb_stat(txn, dbi, &mst);
+
+ sval = calloc(1, mst.ms_psize / 4);
+ key.mv_size = sizeof(long);
+ key.mv_data = &kval;
+ data.mv_size = mst.ms_psize / 4 - 30;
+ data.mv_data = sval;
+
+ printf("Adding 12 values, should yield 3 splits\n");
+ for (i=0;i<12;i++) {
+ kval = i*5;
+ sprintf(sval, "%08x", kval);
+ rc = mdb_cursor_put(cursor, &key, &data, MDB_NOOVERWRITE);
+ }
+ printf("Adding 12 more values, should yield 3 splits\n");
+ for (i=0;i<12;i++) {
+ kval = i*5+4;
+ sprintf(sval, "%08x", kval);
+ rc = mdb_cursor_put(cursor, &key, &data, MDB_NOOVERWRITE);
+ }
+ printf("Adding 12 more values, should yield 3 splits\n");
+ for (i=0;i<12;i++) {
+ kval = i*5+1;
+ sprintf(sval, "%08x", kval);
+ rc = mdb_cursor_put(cursor, &key, &data, MDB_NOOVERWRITE);
+ }
+ rc = mdb_cursor_get(cursor, &key, &data, MDB_FIRST);
+
+ do {
+ printf("key: %p %s, data: %p %.*s\n",
+ key.mv_data, mdb_dkey(&key, dkbuf),
+ data.mv_data, (int) data.mv_size, (char *) data.mv_data);
+ } while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0);
+ mdb_cursor_close(cursor);
+ mdb_txn_commit(txn);
+
+#if 0
+ j=0;
+
+ for (i= count - 1; i > -1; i-= (random()%5)) {
+ j++;
+ txn=NULL;
+ rc = mdb_txn_begin(env, NULL, 0, &txn);
+ sprintf(kval, "%03x", values[i & ~0x0f]);
+ sprintf(sval, "%03x %d foo bar", values[i], values[i]);
+ key.mv_size = sizeof(int);
+ key.mv_data = kval;
+ data.mv_size = sizeof(sval);
+ data.mv_data = sval;
+ rc = mdb_del(txn, dbi, &key, &data);
+ if (rc) {
+ j--;
+ mdb_txn_abort(txn);
+ } else {
+ rc = mdb_txn_commit(txn);
+ }
+ }
+ free(values);
+ printf("Deleted %d values\n", j);
+
+ rc = mdb_env_stat(env, &mst);
+ rc = mdb_txn_begin(env, NULL, 1, &txn);
+ rc = mdb_cursor_open(txn, dbi, &cursor);
+ printf("Cursor next\n");
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+ printf("Cursor prev\n");
+ while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_PREV)) == 0) {
+ printf("key: %.*s, data: %.*s\n",
+ (int) key.mv_size, (char *) key.mv_data,
+ (int) data.mv_size, (char *) data.mv_data);
+ }
+ mdb_cursor_close(cursor);
+ mdb_close(txn, dbi);
+
+ mdb_txn_abort(txn);
+#endif
+ mdb_env_close(env);
+
+ return 0;
+}