diff options
Diffstat (limited to 'db_drivers')
29 files changed, 16219 insertions, 0 deletions
diff --git a/db_drivers/CMakeLists.txt b/db_drivers/CMakeLists.txt new file mode 100644 index 000000000..5993ed3a2 --- /dev/null +++ b/db_drivers/CMakeLists.txt @@ -0,0 +1,49 @@ +# Copyright (c) 2014-2015, The Monero Project +# +# All rights reserved. +# +# Redistribution and use in source and binary forms, with or without modification, are +# permitted provided that the following conditions are met: +# +# 1. Redistributions of source code must retain the above copyright notice, this list of +# conditions and the following disclaimer. +# +# 2. Redistributions in binary form must reproduce the above copyright notice, this list +# of conditions and the following disclaimer in the documentation and/or other +# materials provided with the distribution. +# +# 3. Neither the name of the copyright holder nor the names of its contributors may be +# used to endorse or promote products derived from this software without specific +# prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY +# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF +# MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL +# THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, +# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, +# STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF +# THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +find_package(LMDB) + +if(NOT LMDB_LIBRARIES OR STATIC) + add_subdirectory(liblmdb) + message(STATUS "lmdb not found, building from src tree") + + set(LMDB_STATIC true PARENT_SCOPE) + set(LMDB_INCLUDE "${CMAKE_CURRENT_SOURCE_DIR}/liblmdb" PARENT_SCOPE) + set(LMDB_LIBRARY "lmdb" PARENT_SCOPE) +else() + message(STATUS "Found liblmdb include (lmdb.h) in ${LMDB_INCLUDE_DIR}") + if(LMDB_LIBRARIES) + message(STATUS "Found liblmdb shared library") + set(LMDB_STATIC false PARENT_SCOPE) + set(LMDB_INCLUDE ${LMDB_INCLUDE_DIR} PARENT_SCOPE) + set(LMDB_LIBRARY ${LMDB_LIBRARIES} PARENT_SCOPE) + set(LMDB_LIBRARY_DIRS "" PARENT_SCOPE) + else() + die("Found liblmdb includes, but could not find liblmdb library. Please make sure you have installed liblmdb or liblmdb-dev or the equivalent") + endif() +endif() diff --git a/db_drivers/liblmdb/.gitignore b/db_drivers/liblmdb/.gitignore new file mode 100644 index 000000000..d5102a87c --- /dev/null +++ b/db_drivers/liblmdb/.gitignore @@ -0,0 +1,23 @@ +mtest +mtest[23456] +testdb +mdb_copy +mdb_stat +mdb_dump +mdb_load +*.lo +*.[ao] +*.so +*.exe +*[~#] +*.bak +*.orig +*.rej +*.gcov +*.gcda +*.gcno +core +core.* +valgrind.* +man/ +html/ diff --git a/db_drivers/liblmdb/CHANGES b/db_drivers/liblmdb/CHANGES new file mode 100644 index 000000000..34049ba7a --- /dev/null +++ b/db_drivers/liblmdb/CHANGES @@ -0,0 +1,112 @@ +LMDB 0.9 Change Log + +LMDB 0.9.14 Release (2014/09/20) + Fix to support 64K page size (ITS#7713) + Fix to persist decreased as well as increased mapsizes (ITS#7789) + Fix cursor bug when deleting last node of a DUPSORT key + Fix mdb_env_info to return FIXEDMAP address + Fix ambiguous error code from writing to closed DBI (ITS#7825) + Fix mdb_copy copying past end of file (ITS#7886) + Fix cursor bugs from page_merge/rebalance + Fix to dirty fewer pages in deletes (mdb_page_loose()) + Fix mdb_dbi_open creating subDBs (ITS#7917) + Fix mdb_cursor_get(_DUP) with single value (ITS#7913) + Fix Windows compat issues in mtests (ITS#7879) + Add compacting variant of mdb_copy + Add BigEndian integer key compare code + Add mdb_dump/mdb_load utilities + +LMDB 0.9.13 Release (2014/06/18) + Fix mdb_page_alloc unlimited overflow page search + Documentation + Re-fix MDB_CURRENT doc (ITS#7793) + Fix MDB_GET_MULTIPLE/MDB_NEXT_MULTIPLE doc + +LMDB 0.9.12 Release (2014/06/13) + Fix MDB_GET_BOTH regression (ITS#7875,#7681) + Fix MDB_MULTIPLE writing multiple keys (ITS#7834) + Fix mdb_rebalance (ITS#7829) + Fix mdb_page_split (ITS#7815) + Fix md_entries count (ITS#7861,#7828,#7793) + Fix MDB_CURRENT (ITS#7793) + Fix possible crash on Windows DLL detach + Misc code cleanup + Documentation + mdb_cursor_put: cursor moves on error (ITS#7771) + + +LMDB 0.9.11 Release (2014/01/15) + Add mdb_env_set_assert() (ITS#7775) + Fix: invalidate txn on page allocation errors (ITS#7377) + Fix xcursor tracking in mdb_cursor_del0() (ITS#7771) + Fix corruption from deletes (ITS#7756) + Fix Windows/MSVC build issues + Raise safe limit of max MDB_MAXKEYSIZE + Misc code cleanup + Documentation + Remove spurious note about non-overlapping flags (ITS#7665) + +LMDB 0.9.10 Release (2013/11/12) + Add MDB_NOMEMINIT option + Fix mdb_page_split() again (ITS#7589) + Fix MDB_NORDAHEAD definition (ITS#7734) + Fix mdb_cursor_del() positioning (ITS#7733) + Partial fix for larger page sizes (ITS#7713) + Fix Windows64/MSVC build issues + +LMDB 0.9.9 Release (2013/10/24) + Add mdb_env_get_fd() + Add MDB_NORDAHEAD option + Add MDB_NOLOCK option + Avoid wasting space in mdb_page_split() (ITS#7589) + Fix mdb_page_merge() cursor fixup (ITS#7722) + Fix mdb_cursor_del() on last delete (ITS#7718) + Fix adding WRITEMAP on existing env (ITS#7715) + Fix nested txns (ITS#7515) + Fix mdb_env_copy() O_DIRECT bug (ITS#7682) + Fix mdb_cursor_set(SET_RANGE) return code (ITS#7681) + Fix mdb_rebalance() cursor fixup (ITS#7701) + Misc code cleanup + Documentation + Note that by default, readers need write access + + +LMDB 0.9.8 Release (2013/09/09) + Allow mdb_env_set_mapsize() on an open environment + Fix mdb_dbi_flags() (ITS#7672) + Fix mdb_page_unspill() in nested txns + Fix mdb_cursor_get(CURRENT|NEXT) after a delete + Fix mdb_cursor_get(DUP) to always return key (ITS#7671) + Fix mdb_cursor_del() to always advance to next item (ITS#7670) + Fix mdb_cursor_set(SET_RANGE) for tree with single page (ITS#7681) + Fix mdb_env_copy() retry open if O_DIRECT fails (ITS#7682) + Tweak mdb_page_spill() to be less aggressive + Documentation + Update caveats since mdb_reader_check() added in 0.9.7 + +LMDB 0.9.7 Release (2013/08/17) + Don't leave stale lockfile on failed RDONLY open (ITS#7664) + Fix mdb_page_split() ref beyond cursor depth + Fix read txn data race (ITS#7635) + Fix mdb_rebalance (ITS#7536, #7538) + Fix mdb_drop() (ITS#7561) + Misc DEBUG macro fixes + Add MDB_NOTLS envflag + Add mdb_env_copyfd() + Add mdb_txn_env() (ITS#7660) + Add mdb_dbi_flags() (ITS#7661) + Add mdb_env_get_maxkeysize() + Add mdb_env_reader_list()/mdb_env_reader_check() + Add mdb_page_spill/unspill, remove hard txn size limit + Use shorter names for semaphores (ITS#7615) + Build + Fix install target (ITS#7656) + Documentation + Misc updates for cursors, DB handles, data lifetime + +LMDB 0.9.6 Release (2013/02/25) + Many fixes/enhancements + +LMDB 0.9.5 Release (2012/11/30) + Renamed from libmdb to liblmdb + Many fixes/enhancements diff --git a/db_drivers/liblmdb/CMakeLists.txt b/db_drivers/liblmdb/CMakeLists.txt new file mode 100644 index 000000000..62b800135 --- /dev/null +++ b/db_drivers/liblmdb/CMakeLists.txt @@ -0,0 +1,39 @@ +# Copyright (c) 2014-2015, The Monero Project +# +# All rights reserved. +# +# Redistribution and use in source and binary forms, with or without modification, are +# permitted provided that the following conditions are met: +# +# 1. Redistributions of source code must retain the above copyright notice, this list of +# conditions and the following disclaimer. +# +# 2. Redistributions in binary form must reproduce the above copyright notice, this list +# of conditions and the following disclaimer in the documentation and/or other +# materials provided with the distribution. +# +# 3. Neither the name of the copyright holder nor the names of its contributors may be +# used to endorse or promote products derived from this software without specific +# prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY +# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF +# MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL +# THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, +# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS +# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, +# STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF +# THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +set (lmdb_sources +mdb.c +midl.c) + +include_directories("${CMAKE_CURRENT_SOURCE_DIR}") + +add_library(lmdb + ${lmdb_sources}) +target_link_libraries(lmdb + LINK_PRIVATE + ${CMAKE_THREAD_LIBS_INIT}) diff --git a/db_drivers/liblmdb/COPYRIGHT b/db_drivers/liblmdb/COPYRIGHT new file mode 100644 index 000000000..e0eb484c7 --- /dev/null +++ b/db_drivers/liblmdb/COPYRIGHT @@ -0,0 +1,20 @@ +Copyright 2011-2014 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/db_drivers/liblmdb/Doxyfile b/db_drivers/liblmdb/Doxyfile new file mode 100644 index 000000000..92d17b09e --- /dev/null +++ b/db_drivers/liblmdb/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 = LMDB + +# 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 = lmdb.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 = tooltag=./man1 + +# 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/db_drivers/liblmdb/LICENSE b/db_drivers/liblmdb/LICENSE new file mode 100644 index 000000000..05ad7571e --- /dev/null +++ b/db_drivers/liblmdb/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/db_drivers/liblmdb/Makefile b/db_drivers/liblmdb/Makefile new file mode 100644 index 000000000..2d0983eff --- /dev/null +++ b/db_drivers/liblmdb/Makefile @@ -0,0 +1,104 @@ +# Makefile for liblmdb (Lightning memory-mapped database library). + +######################################################################## +# Configuration. The compiler options must enable threaded compilation. +# +# Preprocessor macros (for CPPFLAGS) of interest... +# Note that the defaults should already be correct for most +# platforms; you should not need to change any of these. +# Read their descriptions in mdb.c if you do: +# +# - MDB_USE_POSIX_SEM +# - MDB_DSYNC +# - MDB_FDATASYNC +# - MDB_FDATASYNC_WORKS +# - MDB_USE_PWRITEV +# +# There may be other macros in mdb.c of interest. You should +# read mdb.c before changing any of them. +# +CC = gcc +W = -W -Wall -Wno-unused-parameter -Wbad-function-cast -Wuninitialized +THREADS = -pthread +OPT = -O2 -g +CFLAGS = $(THREADS) $(OPT) $(W) $(XCFLAGS) +LDLIBS = +SOLIBS = +prefix = /usr/local + +######################################################################## + +IHDRS = lmdb.h +ILIBS = liblmdb.a liblmdb.so +IPROGS = mdb_stat mdb_copy mdb_dump mdb_load +IDOCS = mdb_stat.1 mdb_copy.1 mdb_dump.1 mdb_load.1 +PROGS = $(IPROGS) mtest mtest2 mtest3 mtest4 mtest5 +all: $(ILIBS) $(PROGS) + +install: $(ILIBS) $(IPROGS) $(IHDRS) + for f in $(IPROGS); do cp $$f $(DESTDIR)$(prefix)/bin; done + for f in $(ILIBS); do cp $$f $(DESTDIR)$(prefix)/lib; done + for f in $(IHDRS); do cp $$f $(DESTDIR)$(prefix)/include; done + for f in $(IDOCS); do cp $$f $(DESTDIR)$(prefix)/man/man1; done + +clean: + rm -rf $(PROGS) *.[ao] *.[ls]o *~ testdb + +test: all + rm -rf testdb && mkdir testdb + ./mtest && ./mdb_stat testdb + +liblmdb.a: mdb.o midl.o + ar rs $@ mdb.o midl.o + +liblmdb.so: mdb.lo midl.lo +# $(CC) $(LDFLAGS) -pthread -shared -Wl,-Bsymbolic -o $@ mdb.o midl.o $(SOLIBS) + $(CC) $(LDFLAGS) -pthread -shared -o $@ mdb.lo midl.lo $(SOLIBS) + +mdb_stat: mdb_stat.o liblmdb.a +mdb_copy: mdb_copy.o liblmdb.a +mdb_dump: mdb_dump.o liblmdb.a +mdb_load: mdb_load.o liblmdb.a +mtest: mtest.o liblmdb.a +mtest2: mtest2.o liblmdb.a +mtest3: mtest3.o liblmdb.a +mtest4: mtest4.o liblmdb.a +mtest5: mtest5.o liblmdb.a +mtest6: mtest6.o liblmdb.a + +mdb.o: mdb.c lmdb.h midl.h + $(CC) $(CFLAGS) $(CPPFLAGS) -c mdb.c + +midl.o: midl.c midl.h + $(CC) $(CFLAGS) $(CPPFLAGS) -c midl.c + +mdb.lo: mdb.c lmdb.h midl.h + $(CC) $(CFLAGS) -fPIC $(CPPFLAGS) -c mdb.c -o $@ + +midl.lo: midl.c midl.h + $(CC) $(CFLAGS) -fPIC $(CPPFLAGS) -c midl.c -o $@ + +%: %.o + $(CC) $(CFLAGS) $(LDFLAGS) $^ $(LDLIBS) -o $@ + +%.o: %.c lmdb.h + $(CC) $(CFLAGS) $(CPPFLAGS) -c $< + +COV_FLAGS=-fprofile-arcs -ftest-coverage +COV_OBJS=xmdb.o xmidl.o + +coverage: xmtest + for i in mtest*.c [0-9]*.c; do j=`basename \$$i .c`; $(MAKE) $$j.o; \ + gcc -o x$$j $$j.o $(COV_OBJS) -pthread $(COV_FLAGS); \ + rm -rf testdb; mkdir testdb; ./x$$j; done + gcov xmdb.c + gcov xmidl.c + +xmtest: mtest.o xmdb.o xmidl.o + gcc -o xmtest mtest.o xmdb.o xmidl.o -pthread $(COV_FLAGS) + +xmdb.o: mdb.c lmdb.h midl.h + $(CC) $(CFLAGS) -fPIC $(CPPFLAGS) -O0 $(COV_FLAGS) -c mdb.c -o $@ + +xmidl.o: midl.c midl.h + $(CC) $(CFLAGS) -fPIC $(CPPFLAGS) -O0 $(COV_FLAGS) -c midl.c -o $@ diff --git a/db_drivers/liblmdb/lmdb.h b/db_drivers/liblmdb/lmdb.h new file mode 100644 index 000000000..ebfbc5dc8 --- /dev/null +++ b/db_drivers/liblmdb/lmdb.h @@ -0,0 +1,1573 @@ +/** @file lmdb.h + * @brief Lightning memory-mapped database library + * + * @mainpage Lightning Memory-Mapped Database Manager (LMDB) + * + * @section intro_sec Introduction + * LMDB is a Btree-based database management library modeled loosely on the + * BerkeleyDB API, but much simplified. The entire database is exposed + * in a 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 when 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, LMDB 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. LMDB 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. + * + * The memory map can be used as a read-only or read-write map. It is + * read-only by default as this provides total immunity to corruption. + * Using read-write mode offers much higher write performance, but adds + * the possibility for stray application writes thru pointers to silently + * corrupt the database. Of course if your application code is known to + * be bug-free (...) then this is not an issue. + * + * @section caveats_sec Caveats + * Troubleshooting the lock file, plus semaphores on BSD systems: + * + * - A broken lockfile can cause sync issues. + * Stale reader transactions left behind by an aborted program + * cause further writes to grow the database quickly, and + * stale locks can block further operation. + * + * Fix: Check for stale readers periodically, using the + * #mdb_reader_check function or the \ref mdb_stat_1 "mdb_stat" tool. + * Stale writers will be cleared automatically on most systems: + * - Windows - automatic + * - BSD, systems using SysV semaphores - automatic + * - Linux, systems using POSIX mutexes with Robust option - automatic + * Otherwise just make all programs using the database close it; + * the lockfile is always reset on first open of the environment. + * + * - On BSD systems or others configured with MDB_USE_SYSV_SEM, + * startup can fail due to semaphores owned by another userid. + * + * Fix: Open and close the database as the user which owns the + * semaphores (likely last user) or as root, while no other + * process is using the database. + * + * Restrictions/caveats (in addition to those listed for some functions): + * + * - Only the database owner should normally use the database on + * BSD systems or when otherwise configured with MDB_USE_POSIX_SEM. + * Multiple users can cause startup to fail later, as noted above. + * + * - There is normally no pure read-only mode, since readers need write + * access to locks and lock file. Exceptions: On read-only filesystems + * or with the #MDB_NOLOCK flag described under #mdb_env_open(). + * + * - By default, in versions before 0.9.10, unused portions of the data + * file might receive garbage data from memory freed by other code. + * (This does not happen when using the #MDB_WRITEMAP flag.) As of + * 0.9.10 the default behavior is to initialize such memory before + * writing to the data file. Since there may be a slight performance + * cost due to this initialization, applications may disable it using + * the #MDB_NOMEMINIT flag. Applications handling sensitive data + * which must not be written should not use this flag. This flag is + * irrelevant when using #MDB_WRITEMAP. + * + * - A thread can only use one transaction at a time, plus any child + * transactions. Each transaction belongs to one thread. See below. + * The #MDB_NOTLS flag changes this for read-only transactions. + * + * - Use an MDB_env* in the process which opened it, without fork()ing. + * + * - Do not have open an LMDB database twice in the same process at + * the same time. Not even from a plain open() call - close()ing it + * breaks flock() advisory locking. + * + * - Avoid long-lived transactions. Read transactions prevent + * reuse of pages freed by newer write transactions, thus the + * database can grow quickly. Write transactions prevent + * other write transactions, since writes are serialized. + * + * - Avoid suspending a process with active transactions. These + * would then be "long-lived" as above. Also read transactions + * suspended when writers commit could sometimes see wrong data. + * + * ...when several processes can use a database concurrently: + * + * - Avoid aborting a process with an active transaction. + * The transaction becomes "long-lived" as above until a check + * for stale readers is performed or the lockfile is reset, + * since the process may not remove it from the lockfile. + * + * This does not apply to write transactions if the system clears + * stale writers, see above. + * + * - If you do that anyway, do a periodic check for stale readers. Or + * close the environment once in a while, so the lockfile can get reset. + * + * - Do not use LMDB databases on remote filesystems, even between + * processes on the same host. This breaks flock() on some OSes, + * possibly memory map sync, and certainly sync between programs + * on different hosts. + * + * - Opening a database can fail if another process is opening or + * closing it at exactly the same time. + * + * @author Howard Chu, Symas Corporation. + * + * @copyright Copyright 2011-2014 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 _LMDB_H_ +#define _LMDB_H_ + +#include <sys/types.h> + +#ifdef __cplusplus +extern "C" { +#endif + +/** Unix permissions for creating files, or dummy definition for Windows */ +#ifdef _MSC_VER +typedef int mdb_mode_t; +#else +typedef mode_t mdb_mode_t; +#endif + +/** An abstraction for a file handle. + * On POSIX systems file handles are small integers. On Windows + * they're opaque pointers. + */ +#ifdef _WIN32 +typedef void *mdb_filehandle_t; +#else +typedef int mdb_filehandle_t; +#endif + +/** @defgroup mdb LMDB API + * @{ + * @brief OpenLDAP Lightning Memory-Mapped Database Manager + */ +/** @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 14 + +/** 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 20, 2014" + +/** A stringifier for the version info */ +#define MDB_VERSTR(a,b,c,d) "LMDB " #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. + * + * Values returned from the database are valid only until a subsequent + * update operation, or the end of the transaction. Do not modify or + * free them, they commonly point into the database itself. + * + * Key sizes must be between 1 and #mdb_env_get_maxkeysize() inclusive. + * The same applies to data sizes in databases with the #MDB_DUPSORT flag. + * Other data items can in theory be from 0 to 0xffffffff bytes long. + */ +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 (experimental) */ +#define MDB_FIXEDMAP 0x01 + /** no environment directory */ +#define MDB_NOSUBDIR 0x4000 + /** don't fsync after commit */ +#define MDB_NOSYNC 0x10000 + /** read only */ +#define MDB_RDONLY 0x20000 + /** don't fsync metapage after commit */ +#define MDB_NOMETASYNC 0x40000 + /** use writable mmap */ +#define MDB_WRITEMAP 0x80000 + /** use asynchronous msync when #MDB_WRITEMAP is used */ +#define MDB_MAPASYNC 0x100000 + /** tie reader locktable slots to #MDB_txn objects instead of to threads */ +#define MDB_NOTLS 0x200000 + /** don't do any locking, caller must manage their own locks */ +#define MDB_NOLOCK 0x400000 + /** don't do readahead (no effect on Windows) */ +#define MDB_NORDAHEAD 0x800000 + /** don't initialize malloc'd memory before writing to datafile */ +#define MDB_NOMEMINIT 0x1000000 +/** @} */ + +/** @defgroup mdb_dbi_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 +/** Duplicate data is being appended, don't split full pages. */ +#define MDB_APPENDDUP 0x40000 +/** Store multiple data items in one call. Only for #MDB_DUPFIXED. */ +#define MDB_MULTIPLE 0x80000 +/* @} */ + +/** @defgroup mdb_copy Copy Flags + * @{ + */ +/** Compacting copy: Omit free space from copy, and renumber all + * pages sequentially. + */ +#define MDB_CP_COMPACT 0x01 +/* @} */ + +/** @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_CURRENT, /**< Return key/data at current cursor position */ + MDB_GET_MULTIPLE, /**< Return key and up to a page of duplicate data items + from current cursor position. Move cursor to prepare + for #MDB_NEXT_MULTIPLE. 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 key and up to a page of duplicate data items + from next cursor position. Move cursor to prepare + for #MDB_NEXT_MULTIPLE. Only for #MDB_DUPFIXED */ + MDB_NEXT_NODUP, /**< Position at first data item of next key */ + 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 */ + MDB_SET, /**< Position at specified key */ + MDB_SET_KEY, /**< Position at specified key, return key + data */ + 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 or environment had fatal error */ +#define MDB_PANIC (-30795) + /** Environment version mismatch */ +#define MDB_VERSION_MISMATCH (-30794) + /** File is not a valid LMDB file */ +#define MDB_INVALID (-30793) + /** Environment mapsize reached */ +#define MDB_MAP_FULL (-30792) + /** Environment maxdbs reached */ +#define MDB_DBS_FULL (-30791) + /** Environment maxreaders reached */ +#define MDB_READERS_FULL (-30790) + /** Too many TLS keys in use - Windows only */ +#define MDB_TLS_FULL (-30789) + /** Txn has too many dirty pages */ +#define MDB_TXN_FULL (-30788) + /** Cursor stack too deep - internal error */ +#define MDB_CURSOR_FULL (-30787) + /** Page has not enough space - internal error */ +#define MDB_PAGE_FULL (-30786) + /** Database contents grew beyond environment mapsize */ +#define MDB_MAP_RESIZED (-30785) + /** MDB_INCOMPATIBLE: Operation and DB incompatible, or DB flags changed */ +#define MDB_INCOMPATIBLE (-30784) + /** Invalid reuse of reader locktable slot */ +#define MDB_BAD_RSLOT (-30783) + /** Transaction cannot recover - it must be aborted */ +#define MDB_BAD_TXN (-30782) + /** Unsupported size of key/DB name/data, or wrong DUPFIXED size */ +#define MDB_BAD_VALSIZE (-30781) + /** The specified DBI was changed unexpectedly */ +#define MDB_BAD_DBI (-30780) + /** The last defined error code */ +#define MDB_LAST_ERRCODE MDB_BAD_DBI +/** @} */ + +/** @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 Information about the environment */ +typedef struct MDB_envinfo { + void *me_mapaddr; /**< Address of map, if fixed */ + size_t me_mapsize; /**< Size of the data memory map */ + size_t me_last_pgno; /**< ID of the last used page */ + size_t me_last_txnid; /**< ID of the last committed transaction */ + unsigned int me_maxreaders; /**< max reader slots in the environment */ + unsigned int me_numreaders; /**< max reader slots used in the environment */ +} MDB_envinfo; + + /** @brief Return the LMDB 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 LMDB library error is + * returned. See @ref errors for a list of LMDB-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 LMDB 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. + * Flags set by mdb_env_set_flags() are also used. + * <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, LMDB 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_RDONLY + * Open the environment in read-only mode. No write operations will be + * allowed. LMDB will still modify the lock file - except on read-only + * filesystems, where LMDB does not use locks. + * <li>#MDB_WRITEMAP + * Use a writeable memory map unless MDB_RDONLY is set. This is faster + * and uses fewer mallocs, but loses protection from application bugs + * like wild pointer writes and other bad updates into the database. + * Incompatible with nested transactions. + * Do not mix processes with and without MDB_WRITEMAP on the same + * environment. This can defeat durability (#mdb_env_sync etc). + * <li>#MDB_NOMETASYNC + * Flush system buffers to disk only once per transaction, omit the + * metadata flush. Defer that until the system flushes files to disk, + * or next non-MDB_RDONLY commit or #mdb_env_sync(). This optimization + * maintains database integrity, but a system crash may undo the last + * committed transaction. I.e. it preserves the ACI (atomicity, + * consistency, isolation) but not D (durability) database property. + * This flag may be changed at any time using #mdb_env_set_flags(). + * <li>#MDB_NOSYNC + * Don't flush system buffers to disk when committing a transaction. + * This optimization means a system crash can corrupt the database or + * lose the last transactions if buffers are not yet flushed to disk. + * The risk is governed by how often the system flushes dirty buffers + * to disk and how often #mdb_env_sync() is called. However, if the + * filesystem preserves write order and the #MDB_WRITEMAP flag is not + * used, transactions exhibit ACI (atomicity, consistency, isolation) + * properties and only lose D (durability). I.e. database integrity + * is maintained, but a system crash may undo the final transactions. + * Note that (#MDB_NOSYNC | #MDB_WRITEMAP) leaves the system with no + * hint for when to write transactions to disk, unless #mdb_env_sync() + * is called. (#MDB_MAPASYNC | #MDB_WRITEMAP) may be preferable. + * This flag may be changed at any time using #mdb_env_set_flags(). + * <li>#MDB_MAPASYNC + * When using #MDB_WRITEMAP, use asynchronous flushes to disk. + * As with #MDB_NOSYNC, a system crash can then corrupt the + * database or lose the last transactions. Calling #mdb_env_sync() + * ensures on-disk database integrity until next commit. + * This flag may be changed at any time using #mdb_env_set_flags(). + * <li>#MDB_NOTLS + * Don't use Thread-Local Storage. Tie reader locktable slots to + * #MDB_txn objects instead of to threads. I.e. #mdb_txn_reset() keeps + * the slot reseved for the #MDB_txn object. A thread may use parallel + * read-only transactions. A read-only transaction may span threads if + * the user synchronizes its use. Applications that multiplex many + * user threads over individual OS threads need this option. Such an + * application must also serialize the write transactions in an OS + * thread, since LMDB's write locking is unaware of the user threads. + * <li>#MDB_NOLOCK + * Don't do any locking. If concurrent access is anticipated, the + * caller must manage all concurrency itself. For proper operation + * the caller must enforce single-writer semantics, and must ensure + * that no readers are using old transactions while a writer is + * active. The simplest approach is to use an exclusive lock so that + * no readers may be active at all when a writer begins. + * <li>#MDB_NORDAHEAD + * Turn off readahead. Most operating systems perform readahead on + * read requests by default. This option turns it off if the OS + * supports it. Turning it off may help random read performance + * when the DB is larger than RAM and system RAM is full. + * The option is not implemented on Windows. + * <li>#MDB_NOMEMINIT + * Don't initialize malloc'd memory before writing to unused spaces + * in the data file. By default, memory for pages written to the data + * file is obtained using malloc. While these pages may be reused in + * subsequent transactions, freshly malloc'd pages will be initialized + * to zeroes before use. This avoids persisting leftover data from other + * code (that used the heap and subsequently freed the memory) into the + * data file. Note that many other system libraries may allocate + * and free memory from the heap for arbitrary uses. E.g., stdio may + * use the heap for file I/O buffers. This initialization step has a + * modest performance cost so some applications may want to disable + * it using this flag. This option can be a problem for applications + * which handle sensitive data like passwords, and it makes memory + * checkers like Valgrind noisy. This flag is not needed with #MDB_WRITEMAP, + * which writes directly to the mmap instead of using malloc for pages. The + * initialization is also skipped if #MDB_RESERVE is used; the + * caller is expected to overwrite all of the memory that was + * reserved in that case. + * This flag may be changed at any time using #mdb_env_set_flags(). + * </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 LMDB library doesn't match the + * version that created the database environment. + * <li>#MDB_INVALID - 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, mdb_mode_t mode); + + /** @brief Copy an LMDB environment to the specified path. + * + * This function may be used to make a backup of an existing environment. + * No lockfile is created, since it gets recreated at need. + * @note This call can trigger significant file size growth if run in + * parallel with write transactions, because it employs a read-only + * transaction. See long-lived transactions under @ref caveats_sec. + * @param[in] env An environment handle returned by #mdb_env_create(). It + * must have already been opened successfully. + * @param[in] path The directory in which the copy will reside. This + * directory must already exist and be writable but must otherwise be + * empty. + * @return A non-zero error value on failure and 0 on success. + */ +int mdb_env_copy(MDB_env *env, const char *path); + + /** @brief Copy an LMDB environment to the specified file descriptor. + * + * This function may be used to make a backup of an existing environment. + * No lockfile is created, since it gets recreated at need. + * @note This call can trigger significant file size growth if run in + * parallel with write transactions, because it employs a read-only + * transaction. See long-lived transactions under @ref caveats_sec. + * @param[in] env An environment handle returned by #mdb_env_create(). It + * must have already been opened successfully. + * @param[in] fd The filedescriptor to write the copy to. It must + * have already been opened for Write access. + * @return A non-zero error value on failure and 0 on success. + */ +int mdb_env_copyfd(MDB_env *env, mdb_filehandle_t fd); + + /** @brief Copy an LMDB environment to the specified path, with options. + * + * This function may be used to make a backup of an existing environment. + * No lockfile is created, since it gets recreated at need. + * @note This call can trigger significant file size growth if run in + * parallel with write transactions, because it employs a read-only + * transaction. See long-lived transactions under @ref caveats_sec. + * @param[in] env An environment handle returned by #mdb_env_create(). It + * must have already been opened successfully. + * @param[in] path The directory in which the copy will reside. This + * directory must already exist and be writable but must otherwise be + * empty. + * @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_CP_COMPACT - Perform compaction while copying: omit free + * pages and sequentially renumber all pages in output. This option + * consumes more CPU and runs more slowly than the default. + * </ul> + * @return A non-zero error value on failure and 0 on success. + */ +int mdb_env_copy2(MDB_env *env, const char *path, unsigned int flags); + + /** @brief Copy an LMDB environment to the specified file descriptor, + * with options. + * + * This function may be used to make a backup of an existing environment. + * No lockfile is created, since it gets recreated at need. See + * #mdb_env_copy2() for further details. + * @note This call can trigger significant file size growth if run in + * parallel with write transactions, because it employs a read-only + * transaction. See long-lived transactions under @ref caveats_sec. + * @param[in] env An environment handle returned by #mdb_env_create(). It + * must have already been opened successfully. + * @param[in] fd The filedescriptor to write the copy to. It must + * have already been opened for Write access. + * @param[in] flags Special options for this operation. + * See #mdb_env_copy2() for options. + * @return A non-zero error value on failure and 0 on success. + */ +int mdb_env_copyfd2(MDB_env *env, mdb_filehandle_t fd, unsigned int flags); + + /** @brief Return statistics about the LMDB 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 Return information about the LMDB environment. + * + * @param[in] env An environment handle returned by #mdb_env_create() + * @param[out] stat The address of an #MDB_envinfo structure + * where the information will be copied + */ +int mdb_env_info(MDB_env *env, MDB_envinfo *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. LMDB always flushes + * the OS buffers upon commit as well, unless the environment was + * opened with #MDB_NOSYNC or in part #MDB_NOMETASYNC. This call is + * not valid if the environment was opened with #MDB_RDONLY. + * @param[in] env An environment handle returned by #mdb_env_create() + * @param[in] force If non-zero, force a synchronous flush. Otherwise + * if the environment has the #MDB_NOSYNC flag set the flushes + * will be omitted, and with #MDB_MAPASYNC they will be asynchronous. + * @return A non-zero error value on failure and 0 on success. Some possible + * errors are: + * <ul> + * <li>EACCES - the environment is read-only. + * <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 in addition to those from + * #mdb_env_open(), or to unset these flags. If several threads + * change the flags at the same time, the result is undefined. + * Most flags cannot be changed after #mdb_env_open(). + * @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 Return the filedescriptor for the given environment. + * + * @param[in] env An environment handle returned by #mdb_env_create() + * @param[out] fd Address of a mdb_filehandle_t to contain the descriptor. + * @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_fd(MDB_env *env, mdb_filehandle_t *fd); + + /** @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 accommodate future growth of the database. + * This function should be called after #mdb_env_create() and before #mdb_env_open(). + * It may be called at later times if no transactions are active in + * this process. Note that the library does not check for this condition, + * the caller must ensure it explicitly. + * + * The new size takes effect immediately for the current process but + * will not be persisted to any others until a write transaction has been + * committed by the current process. Also, only mapsize increases are + * persisted into the environment. + * + * If the mapsize is increased by another process, and data has grown + * beyond the range of the current mapsize, #mdb_txn_begin() will + * return #MDB_MAP_RESIZED. This function may be called with a size + * of zero to adopt the new size. + * + * Any attempt to set a size smaller than the space already consumed + * by the environment will be silently changed to the current size of the used space. + * @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 has + * an active write transaction. + * </ul> + */ +int mdb_env_set_mapsize(MDB_env *env, size_t size); + + /** @brief Set the maximum number of threads/reader slots 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. + * Starting a read-only transaction normally ties a lock table slot to the + * current thread until the environment closes or the thread exits. If + * MDB_NOTLS is in use, #mdb_txn_begin() instead ties the slot to the + * MDB_txn object until it or the #MDB_env object is destroyed. + * 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 reader lock table slots + * @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/reader slots 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 named databases for the environment. + * + * This function is only needed if multiple databases will be used in the + * environment. Simpler applications that use the environment as a single + * unnamed database can ignore this option. + * This function may only be called after #mdb_env_create() and before #mdb_env_open(). + * + * Currently a moderate number of slots are cheap but a huge number gets + * expensive: 7-120 words per transaction, and every #mdb_dbi_open() + * does a linear search of the opened slots. + * @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 Get the maximum size of keys and #MDB_DUPSORT data we can write. + * + * Depends on the compile-time constant #MDB_MAXKEYSIZE. Default 511. + * See @ref MDB_val. + * @param[in] env An environment handle returned by #mdb_env_create() + * @return The maximum size of a key we can write + */ +int mdb_env_get_maxkeysize(MDB_env *env); + + /** @brief Set application information associated with the #MDB_env. + * + * @param[in] env An environment handle returned by #mdb_env_create() + * @param[in] ctx An arbitrary pointer for whatever the application needs. + * @return A non-zero error value on failure and 0 on success. + */ +int mdb_env_set_userctx(MDB_env *env, void *ctx); + + /** @brief Get the application information associated with the #MDB_env. + * + * @param[in] env An environment handle returned by #mdb_env_create() + * @return The pointer set by #mdb_env_set_userctx(). + */ +void *mdb_env_get_userctx(MDB_env *env); + + /** @brief A callback function for most LMDB assert() failures, + * called before printing the message and aborting. + * + * @param[in] env An environment handle returned by #mdb_env_create(). + * @param[in] msg The assertion message, not including newline. + */ +typedef void MDB_assert_func(MDB_env *env, const char *msg); + + /** Set or reset the assert() callback of the environment. + * Disabled if liblmdb is buillt with NDEBUG. + * @note This hack should become obsolete as lmdb's error handling matures. + * @param[in] env An environment handle returned by #mdb_env_create(). + * @param[in] func An #MDB_assert_func function, or 0. + * @return A non-zero error value on failure and 0 on success. + */ +int mdb_env_set_assert(MDB_env *env, MDB_assert_func *func); + + /** @brief Create a transaction for use with the environment. + * + * The transaction handle may be discarded using #mdb_txn_abort() or #mdb_txn_commit(). + * @note A transaction and its cursors must only be used by a single + * thread, and a thread may only have a single transaction at a time. + * If #MDB_NOTLS is in use, this does not apply to read-only transactions. + * @note Cursors may not span transactions. + * @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 and its cursors may not issue any other operations than + * mdb_txn_commit and mdb_txn_abort 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>#MDB_MAP_RESIZED - another process wrote data beyond this MDB_env's + * mapsize and this environment's map must be resized as well. + * See #mdb_env_set_mapsize(). + * <li>#MDB_READERS_FULL - a read-only transaction was requested and + * the reader lock table is full. See #mdb_env_set_maxreaders(). + * <li>ENOMEM - out of memory. + * </ul> + */ +int mdb_txn_begin(MDB_env *env, MDB_txn *parent, unsigned int flags, MDB_txn **txn); + + /** @brief Returns the transaction's #MDB_env + * + * @param[in] txn A transaction handle returned by #mdb_txn_begin() + */ +MDB_env *mdb_txn_env(MDB_txn *txn); + + /** @brief Return the transaction's ID. + * + * This returns the identifier associated with this transaction. For a + * read-only transaction, this corresponds to the snapshot being read; + * concurrent readers will frequently have the same transaction ID. + * + * @param[in] txn A transaction handle returned by #mdb_txn_begin() + * @return A transaction ID, valid if input is an active transaction. + */ +size_t mdb_txn_id(MDB_txn *txn); + + /** @brief Commit all the operations of a transaction into the database. + * + * The transaction handle is freed. It and its cursors must not be used + * again after this call, except with #mdb_cursor_renew(). + * @note Earlier documentation incorrectly said all cursors would be freed. + * Only write-transactions free cursors. + * @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. + * <li>ENOMEM - out of memory. + * </ul> + */ +int mdb_txn_commit(MDB_txn *txn); + + /** @brief Abandon all the operations of the transaction instead of saving them. + * + * The transaction handle is freed. It and its cursors must not be used + * again after this call, except with #mdb_cursor_renew(). + * @note Earlier documentation incorrectly said all cursors would be freed. + * Only write-transactions free cursors. + * @param[in] txn A transaction handle returned by #mdb_txn_begin() + */ +void mdb_txn_abort(MDB_txn *txn); + + /** @brief Reset a read-only transaction. + * + * Abort the transaction like #mdb_txn_abort(), but keep the transaction + * handle. #mdb_txn_renew() may reuse the handle. This saves allocation + * overhead if the process will start a new read-only transaction soon, + * and also locking overhead if #MDB_NOTLS is in use. The reader table + * lock is released, but the table slot stays tied to its thread or + * #MDB_txn. Use mdb_txn_abort() to discard a reset handle, and to free + * its lock table slot if MDB_NOTLS is in use. + * Cursors opened within the transaction must not be used + * again after this call, except with #mdb_cursor_renew(). + * 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); + +/** Compat with version <= 0.9.4, avoid clash with libmdb from MDB Tools project */ +#define mdb_open(txn,name,flags,dbi) mdb_dbi_open(txn,name,flags,dbi) +/** Compat with version <= 0.9.4, avoid clash with libmdb from MDB Tools project */ +#define mdb_close(env,dbi) mdb_dbi_close(env,dbi) + + /** @brief Open a database in the environment. + * + * A database handle denotes the name and parameters of a database, + * independently of whether such a database exists. + * The database handle may be discarded by calling #mdb_dbi_close(). + * The old database handle is returned if the database was already open. + * The handle may only be closed once. + * The database handle will be private to the current transaction until + * the transaction is successfully committed. If the transaction is + * aborted the handle will be closed automatically. + * After a successful commit the + * handle will reside in the shared environment, and may be used + * by other transactions. This function must not be called from + * multiple concurrent transactions in the same process. A transaction + * that uses this function must finish (either commit or abort) before + * any other transaction in the process may use this function. + * + * To use named databases (with name != NULL), #mdb_env_set_maxdbs() + * must be called before opening the environment. Database names + * are kept as keys in the unnamed database. + * @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 environment, 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>#MDB_DBS_FULL - too many databases have been opened. See #mdb_env_set_maxdbs(). + * </ul> + */ +int mdb_dbi_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_dbi_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 Retrieve the DB flags for a database handle. + * + * @param[in] txn A transaction handle returned by #mdb_txn_begin() + * @param[in] dbi A database handle returned by #mdb_dbi_open() + * @param[out] flags Address where the flags will be returned. + * @return A non-zero error value on failure and 0 on success. + */ +int mdb_dbi_flags(MDB_txn *txn, MDB_dbi dbi, unsigned int *flags); + + /** @brief Close a database handle. Normally unnecessary. Use with care: + * + * 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 or one of its cursors any further. Do not close + * a handle if an existing transaction has modified its database. + * Doing so can cause misbehavior from database corruption to errors + * like MDB_BAD_VALSIZE (since the DB name is gone). + * + * Closing a database handle is not necessary, but lets #mdb_dbi_open() + * reuse the handle value. Usually it's better to set a bigger + * #mdb_env_set_maxdbs(), unless that value would be large. + * + * @param[in] env An environment handle returned by #mdb_env_create() + * @param[in] dbi A database handle returned by #mdb_dbi_open() + */ +void mdb_dbi_close(MDB_env *env, MDB_dbi dbi); + + /** @brief Empty or delete+close a database. + * + * See #mdb_dbi_close() for restrictions about closing the DB handle. + * @param[in] txn A transaction handle returned by #mdb_txn_begin() + * @param[in] dbi A database handle returned by #mdb_dbi_open() + * @param[in] del 0 to empty the DB, 1 to delete it from the + * environment and close the DB handle. + * @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 special key flags were specified + * with #mdb_dbi_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_dbi_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_dbi_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_dbi_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_dbi_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_dbi_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. + * @note Values returned from the database are valid only until a + * subsequent update operation, or the end of the transaction. + * @param[in] txn A transaction handle returned by #mdb_txn_begin() + * @param[in] dbi A database handle returned by #mdb_dbi_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_dbi_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. + * <li>#MDB_RESERVE - reserve space for data of the given size, but + * don't copy the given data. Instead, return a pointer to the + * reserved space, which the caller can fill in later - before + * the next update operation or the transaction ends. This saves + * an extra memcpy if the data is being generated later. + * LMDB does nothing else with this memory, the caller is expected + * to modify all of the space requested. + * <li>#MDB_APPEND - append the given key/data pair to the end of the + * database. This option allows fast bulk loading when keys are + * already known to be in the correct order. Loading unsorted keys + * with this flag will cause a #MDB_KEYEXIST error. + * <li>#MDB_APPENDDUP - as above, but for sorted dup data. + * </ul> + * @return A non-zero error value on failure and 0 on success. Some possible + * errors are: + * <ul> + * <li>#MDB_MAP_FULL - the database is full, see #mdb_env_set_mapsize(). + * <li>#MDB_TXN_FULL - the transaction has too many dirty pages. + * <li>EACCES - an attempt was made to write in a read-only transaction. + * <li>EINVAL - an invalid parameter was specified. + * </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_dbi_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>EACCES - 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. + * + * A cursor is associated with a specific transaction and database. + * A cursor cannot be used when its database handle is closed. Nor + * when its transaction has ended, except with #mdb_cursor_renew(). + * It can be discarded with #mdb_cursor_close(). + * A cursor in a write-transaction can be closed before its transaction + * ends, and will otherwise be closed when its transaction ends. + * A cursor in a read-only transaction must be closed explicitly, before + * or after its transaction ends. It can be reused with + * #mdb_cursor_renew() before finally closing it. + * @note Earlier documentation said that cursors in every transaction + * were closed when the transaction committed or aborted. + * @param[in] txn A transaction handle returned by #mdb_txn_begin() + * @param[in] dbi A database handle returned by #mdb_dbi_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. + * Its transaction must still be live if it is a write-transaction. + * @param[in] cursor A cursor handle returned by #mdb_cursor_open() + */ +void mdb_cursor_close(MDB_cursor *cursor); + + /** @brief Renew a cursor handle. + * + * A cursor is associated with a specific transaction and database. + * Cursors that are only used in read-only + * transactions may be re-used, to avoid unnecessary malloc/free overhead. + * The cursor may be associated with a new read-only transaction, and + * referencing the same database handle as it was created with. + * This may be done whether the previous transaction is live or dead. + * @param[in] txn A transaction handle returned by #mdb_txn_begin() + * @param[in] cursor A cursor handle returned by #mdb_cursor_open() + * @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_renew(MDB_txn *txn, 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. + * See #mdb_get() for restrictions on using the output values. + * @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. + * The cursor is positioned at the new item, or on failure usually near it. + * @note Earlier documentation incorrectly said errors would leave the + * state of the cursor unchanged. + * @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 - replace the item at the current cursor position. + * The \b key parameter must still be provided, and must match it. + * If using sorted duplicates (#MDB_DUPSORT) the data item must still + * sort into the same place. This is intended to be used when the + * new data is the same size as the old. Otherwise it will simply + * perform a delete of the old record followed by an insert. + * <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). + * <li>#MDB_RESERVE - reserve space for data of the given size, but + * don't copy the given data. Instead, return a pointer to the + * reserved space, which the caller can fill in later. This saves + * an extra memcpy if the data is being generated later. + * <li>#MDB_APPEND - append the given key/data pair to the end of the + * database. No key comparisons are performed. This option allows + * fast bulk loading when keys are already known to be in the + * correct order. Loading unsorted keys with this flag will cause + * data corruption. + * <li>#MDB_APPENDDUP - as above, but for sorted dup data. + * <li>#MDB_MULTIPLE - store multiple contiguous data elements in a + * single request. This flag may only be specified if the database + * was opened with #MDB_DUPFIXED. The \b data argument must be an + * array of two MDB_vals. The mv_size of the first MDB_val must be + * the size of a single data element. The mv_data of the first MDB_val + * must point to the beginning of the array of contiguous data elements. + * The mv_size of the second MDB_val must be the count of the number + * of data elements to store. On return this field will be set to + * the count of the number of elements actually written. The mv_data + * of the second MDB_val is unused. + * </ul> + * @return A non-zero error value on failure and 0 on success. Some possible + * errors are: + * <ul> + * <li>#MDB_MAP_FULL - the database is full, see #mdb_env_set_mapsize(). + * <li>#MDB_TXN_FULL - the transaction has too many dirty pages. + * <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_dbi_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 + * the specified database. The database must have the #MDB_DUPSORT flag. + * @param[in] txn A transaction handle returned by #mdb_txn_begin() + * @param[in] dbi A database handle returned by #mdb_dbi_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); + + /** @brief A callback function used to print a message from the library. + * + * @param[in] msg The string to be printed. + * @param[in] ctx An arbitrary context pointer for the callback. + * @return < 0 on failure, >= 0 on success. + */ +typedef int (MDB_msg_func)(const char *msg, void *ctx); + + /** @brief Dump the entries in the reader lock table. + * + * @param[in] env An environment handle returned by #mdb_env_create() + * @param[in] func A #MDB_msg_func function + * @param[in] ctx Anything the message function needs + * @return < 0 on failure, >= 0 on success. + */ +int mdb_reader_list(MDB_env *env, MDB_msg_func *func, void *ctx); + + /** @brief Check for stale entries in the reader lock table. + * + * @param[in] env An environment handle returned by #mdb_env_create() + * @param[out] dead Number of stale slots that were cleared + * @return 0 on success, non-zero on failure. + */ +int mdb_reader_check(MDB_env *env, int *dead); +/** @} */ + +#ifdef __cplusplus +} +#endif +/** @page tools LMDB Command Line Tools + The following describes the command line tools that are available for LMDB. + \li \ref mdb_copy_1 + \li \ref mdb_dump_1 + \li \ref mdb_load_1 + \li \ref mdb_stat_1 +*/ + +#endif /* _LMDB_H_ */ diff --git a/db_drivers/liblmdb/mdb.c b/db_drivers/liblmdb/mdb.c new file mode 100644 index 000000000..c4dc26903 --- /dev/null +++ b/db_drivers/liblmdb/mdb.c @@ -0,0 +1,9665 @@ +/** @file mdb.c + * @brief Lightning memory-mapped database library + * + * A Btree-based database management library modeled loosely on the + * BerkeleyDB API, but much simplified. + */ +/* + * Copyright 2011-2014 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. + */ +#ifndef _GNU_SOURCE +#define _GNU_SOURCE 1 +#endif +#ifdef _WIN32 +#include <malloc.h> +#include <windows.h> +/** getpid() returns int; MinGW defines pid_t but MinGW64 typedefs it + * as int64 which is wrong. MSVC doesn't define it at all, so just + * don't use it. + */ +#define MDB_PID_T int +#define MDB_THR_T DWORD +#include <sys/types.h> +#include <sys/stat.h> +#ifdef __GNUC__ +# include <sys/param.h> +#else +# define LITTLE_ENDIAN 1234 +# define BIG_ENDIAN 4321 +# define BYTE_ORDER LITTLE_ENDIAN +# ifndef SSIZE_MAX +# define SSIZE_MAX INT_MAX +# endif +#endif +#else +#include <sys/types.h> +#include <sys/stat.h> +#define MDB_PID_T pid_t +#define MDB_THR_T pthread_t +#include <sys/param.h> +#include <sys/uio.h> +#include <sys/mman.h> +#ifdef HAVE_SYS_FILE_H +#include <sys/file.h> +#endif +#include <fcntl.h> +#endif + +#if defined(__mips) && defined(__linux) +/* MIPS has cache coherency issues, requires explicit cache control */ +#include <asm/cachectl.h> +extern int cacheflush(char *addr, int nbytes, int cache); +#define CACHEFLUSH(addr, bytes, cache) cacheflush(addr, bytes, cache) +#else +#define CACHEFLUSH(addr, bytes, cache) +#endif + +#if defined(__linux) && !defined(MDB_FDATASYNC_WORKS) +/** fdatasync is broken on ext3/ext4fs on older kernels, see + * description in #mdb_env_open2 comments. You can safely + * define MDB_FDATASYNC_WORKS if this code will only be run + * on kernels 3.6 and newer. + */ +#define BROKEN_FDATASYNC +#endif + +#include <errno.h> +#include <limits.h> +#include <stddef.h> +#include <inttypes.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <time.h> +#include <unistd.h> + +#if defined(__sun) || defined(ANDROID) +/* Most platforms have posix_memalign, older may only have memalign */ +#define HAVE_MEMALIGN 1 +#include <malloc.h> +#endif + +#if !(defined(BYTE_ORDER) || defined(__BYTE_ORDER)) +#include <netinet/in.h> +#include <resolv.h> /* defines BYTE_ORDER on HPUX and Solaris */ +#endif + +#if defined(__APPLE__) || defined (BSD) +# define MDB_USE_SYSV_SEM 1 +# define MDB_FDATASYNC fsync +#elif defined(ANDROID) +# define MDB_FDATASYNC fsync +#endif + +#ifndef _WIN32 +#include <pthread.h> +#ifdef MDB_USE_SYSV_SEM +#include <sys/ipc.h> +#include <sys/sem.h> +#ifdef _SEM_SEMUN_UNDEFINED +union semun { + int val; + struct semid_ds *buf; + unsigned short *array; +}; +#endif /* _SEM_SEMUN_UNDEFINED */ +#endif /* MDB_USE_SYSV_SEM */ +#endif /* !_WIN32 */ + +#ifdef USE_VALGRIND +#include <valgrind/memcheck.h> +#define VGMEMP_CREATE(h,r,z) VALGRIND_CREATE_MEMPOOL(h,r,z) +#define VGMEMP_ALLOC(h,a,s) VALGRIND_MEMPOOL_ALLOC(h,a,s) +#define VGMEMP_FREE(h,a) VALGRIND_MEMPOOL_FREE(h,a) +#define VGMEMP_DESTROY(h) VALGRIND_DESTROY_MEMPOOL(h) +#define VGMEMP_DEFINED(a,s) VALGRIND_MAKE_MEM_DEFINED(a,s) +#else +#define VGMEMP_CREATE(h,r,z) +#define VGMEMP_ALLOC(h,a,s) +#define VGMEMP_FREE(h,a) +#define VGMEMP_DESTROY(h) +#define VGMEMP_DEFINED(a,s) +#endif + +#ifndef BYTE_ORDER +# if (defined(_LITTLE_ENDIAN) || defined(_BIG_ENDIAN)) && !(defined(_LITTLE_ENDIAN) && defined(_BIG_ENDIAN)) +/* Solaris just defines one or the other */ +# define LITTLE_ENDIAN 1234 +# define BIG_ENDIAN 4321 +# ifdef _LITTLE_ENDIAN +# define BYTE_ORDER LITTLE_ENDIAN +# else +# define BYTE_ORDER BIG_ENDIAN +# endif +# else +# define BYTE_ORDER __BYTE_ORDER +# endif +#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) || defined(_M_IX86) +#define MISALIGNED_OK 1 +#endif + +#include "lmdb.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 + +#ifdef __GNUC__ +/** Put infrequently used env functions in separate section */ +# ifdef __APPLE__ +# define ESECT __attribute__ ((section("__TEXT,text_env"))) +# else +# define ESECT __attribute__ ((section("text_env"))) +# endif +#else +#define ESECT +#endif + +/** @defgroup internal LMDB Internals + * @{ + */ +/** @defgroup compat 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. + * @{ + */ + + /** Features under development */ +#ifndef MDB_DEVEL +#define MDB_DEVEL 0 +#endif + +#if defined(_WIN32) || defined(MDB_USE_SYSV_SEM) || defined(EOWNERDEAD) +#define MDB_ROBUST_SUPPORTED 1 +#endif + + /** Wrapper around __func__, which is a C99 feature */ +#if __STDC_VERSION__ >= 199901L +# define mdb_func_ __func__ +#elif __GNUC__ >= 2 || _MSC_VER >= 1300 +# define mdb_func_ __FUNCTION__ +#else +/* If a debug message says <mdb_unknown>(), update the #if statements above */ +# define mdb_func_ "<mdb_unknown>" +#endif + +/* Internal error codes, not exposed outside liblmdb */ +#define MDB_NO_ROOT (MDB_LAST_ERRCODE + 10) +#ifdef _WIN32 +#define MDB_OWNERDEAD ((int) WAIT_ABANDONED) +#elif defined MDB_USE_SYSV_SEM +#define MDB_OWNERDEAD (MDB_LAST_ERRCODE + 11) +#else +#define MDB_OWNERDEAD EOWNERDEAD +#endif + +#ifdef _WIN32 +#define MDB_USE_HASH 1 +#define MDB_PIDLOCK 0 +#define THREAD_RET DWORD +#define pthread_t HANDLE +#define pthread_mutex_t HANDLE +#define pthread_cond_t HANDLE +typedef HANDLE mdb_mutex_t; +#define pthread_key_t DWORD +#define pthread_self() GetCurrentThreadId() +#define pthread_key_create(x,y) \ + ((*(x) = TlsAlloc()) == TLS_OUT_OF_INDEXES ? ErrCode() : 0) +#define pthread_key_delete(x) TlsFree(x) +#define pthread_getspecific(x) TlsGetValue(x) +#define pthread_setspecific(x,y) (TlsSetValue(x,y) ? 0 : ErrCode()) +#define pthread_mutex_unlock(x) ReleaseMutex(*x) +#define pthread_mutex_lock(x) WaitForSingleObject(*x, INFINITE) +#define pthread_cond_signal(x) SetEvent(*x) +#define pthread_cond_wait(cond,mutex) do{SignalObjectAndWait(*mutex, *cond, INFINITE, FALSE); WaitForSingleObject(*mutex, INFINITE);}while(0) +#define THREAD_CREATE(thr,start,arg) thr=CreateThread(NULL,0,start,arg,0,NULL) +#define THREAD_FINISH(thr) WaitForSingleObject(thr, INFINITE) +#define MDB_MUTEX(env, rw) ((env)->me_##rw##mutex) +#define LOCK_MUTEX0(mutex) WaitForSingleObject(mutex, INFINITE) +#define UNLOCK_MUTEX(mutex) ReleaseMutex(mutex) +#define mdb_mutex_consistent(mutex) 0 +#define getpid() GetCurrentProcessId() +#define MDB_FDATASYNC(fd) (!FlushFileBuffers(fd)) +#define MDB_MSYNC(addr,len,flags) (!FlushViewOfFile(addr,len)) +#define ErrCode() GetLastError() +#define GET_PAGESIZE(x) {SYSTEM_INFO si; GetSystemInfo(&si); (x) = si.dwPageSize;} +#define close(fd) (CloseHandle(fd) ? 0 : -1) +#define munmap(ptr,len) UnmapViewOfFile(ptr) +#ifdef PROCESS_QUERY_LIMITED_INFORMATION +#define MDB_PROCESS_QUERY_LIMITED_INFORMATION PROCESS_QUERY_LIMITED_INFORMATION +#else +#define MDB_PROCESS_QUERY_LIMITED_INFORMATION 0x1000 +#endif +#define Z "I" +#else +#define THREAD_RET void * +#define THREAD_CREATE(thr,start,arg) pthread_create(&thr,NULL,start,arg) +#define THREAD_FINISH(thr) pthread_join(thr,NULL) +#define Z "z" /**< printf format modifier for size_t */ + + /** For MDB_LOCK_FORMAT: True if readers take a pid lock in the lockfile */ +#define MDB_PIDLOCK 1 + +#ifdef MDB_USE_SYSV_SEM + +typedef struct mdb_mutex { + int semid; + int semnum; + int *locked; +} mdb_mutex_t; + +#define MDB_MUTEX(env, rw) (&(env)->me_##rw##mutex) +#define LOCK_MUTEX0(mutex) mdb_sem_wait(mutex) +#define UNLOCK_MUTEX(mutex) do { \ + struct sembuf sb = { 0, 1, SEM_UNDO }; \ + sb.sem_num = (mutex)->semnum; \ + *(mutex)->locked = 0; \ + semop((mutex)->semid, &sb, 1); \ +} while(0) + +static int +mdb_sem_wait(mdb_mutex_t *sem) +{ + int rc, *locked = sem->locked; + struct sembuf sb = { 0, -1, SEM_UNDO }; + sb.sem_num = sem->semnum; + do { + if (!semop(sem->semid, &sb, 1)) { + rc = *locked ? MDB_OWNERDEAD : MDB_SUCCESS; + *locked = 1; + break; + } + } while ((rc = errno) == EINTR); + return rc; +} + +#define mdb_mutex_consistent(mutex) 0 + +#else + /** Pointer/HANDLE type of shared mutex/semaphore. + */ +typedef pthread_mutex_t mdb_mutex_t; + /** Mutex for the reader table (rw = r) or write transaction (rw = w). + */ +#define MDB_MUTEX(env, rw) (&(env)->me_txns->mti_##rw##mutex) + /** Lock the reader or writer mutex. + * Returns 0 or a code to give #mdb_mutex_failed(), as in #LOCK_MUTEX(). + */ +#define LOCK_MUTEX0(mutex) pthread_mutex_lock(mutex) + /** Unlock the reader or writer mutex. + */ +#define UNLOCK_MUTEX(mutex) pthread_mutex_unlock(mutex) + /** Mark mutex-protected data as repaired, after death of previous owner. + */ +#define mdb_mutex_consistent(mutex) pthread_mutex_consistent(mutex) +#endif /* MDB_USE_SYSV_SEM */ + + /** 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) +#define MNAME_LEN 32 +#elif defined(MDB_USE_SYSV_SEM) +#define MNAME_LEN (sizeof(int)) +#else +#define MNAME_LEN (sizeof(pthread_mutex_t)) +#endif + +#ifdef MDB_USE_SYSV_SEM +#define SYSV_SEM_FLAG 1 /**< SysV sems in lockfile format */ +#else +#define SYSV_SEM_FLAG 0 +#endif + +/** @} */ + +#ifdef MDB_ROBUST_SUPPORTED + /** Lock mutex, handle any error, set rc = result. + * Return 0 on success, nonzero (not rc) on error. + */ +#define LOCK_MUTEX(rc, env, mutex) \ + (((rc) = LOCK_MUTEX0(mutex)) && \ + ((rc) = mdb_mutex_failed(env, mutex, rc))) +static int mdb_mutex_failed(MDB_env *env, mdb_mutex_t *mutex, int rc); +#else +#define LOCK_MUTEX(rc, env, mutex) ((rc) = LOCK_MUTEX0(mutex)) +#define mdb_mutex_failed(env, mutex, rc) (rc) +#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 + +/** Function for flushing the data of a file. Define this to fsync + * if fdatasync() is not supported. + */ +#ifndef MDB_FDATASYNC +# define MDB_FDATASYNC fdatasync +#endif + +#ifndef MDB_MSYNC +# define MDB_MSYNC(addr,len,flags) msync(addr,len,flags) +#endif + +#ifndef MS_SYNC +#define MS_SYNC 1 +#endif + +#ifndef MS_ASYNC +#define MS_ASYNC 0 +#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 MDB_ID pgno_t; + + /** A transaction ID. + * See struct MDB_txn.mt_txnid for details. + */ +typedef MDB_ID txnid_t; + +/** @defgroup debug Debug Macros + * @{ + */ +#ifndef MDB_DEBUG + /** Enable debug output. Needs variable argument macros (a C99 feature). + * 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 MDB_DEBUG 0 +#endif + +#if MDB_DEBUG +static int mdb_debug; +static txnid_t mdb_debug_start; + + /** Print a debug message with printf formatting. + * Requires double parenthesis around 2 or more args. + */ +# define DPRINTF(args) ((void) ((mdb_debug) && DPRINTF0 args)) +# define DPRINTF0(fmt, ...) \ + fprintf(stderr, "%s:%d " fmt "\n", mdb_func_, __LINE__, __VA_ARGS__) +#else +# define DPRINTF(args) ((void) 0) +#endif + /** Print a debug string. + * The string is printed literally, with no format processing. + */ +#define DPUTS(arg) DPRINTF(("%s", arg)) + /** Debuging output value of a cursor DBI: Negative in a sub-cursor. */ +#define DDBI(mc) \ + (((mc)->mc_flags & C_SUB) ? -(int)(mc)->mc_dbi : (int)(mc)->mc_dbi) +/** @} */ + + /** @brief The maximum size of a database page. + * + * It is 32k or 64k, since value-PAGEBASE must fit in + * #MDB_page.%mp_upper. + * + * LMDB will use database pages < OS pages if needed. + * That causes more I/O in write transactions: The OS must + * know (read) the whole page before writing a partial page. + * + * 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 MAX_PAGESIZE (PAGEBASE ? 0x10000 : 0x8000) + + /** 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 LMDB 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 datafile format. */ +#define MDB_DATA_VERSION ((MDB_DEVEL) ? 999 : 1) + /** The version number for a database's lockfile format. */ +#define MDB_LOCK_VERSION ((MDB_DEVEL) ? 999 : 1) + + /** @brief The max size of a key we can write, or 0 for dynamic max. + * + * Define this as 0 to compute the max from the page size. 511 + * is default for backwards compat: liblmdb <= 0.9.10 can break + * when modifying a DB with keys/dupsort data bigger than its max. + * #MDB_DEVEL sets the default to 0. + * + * Data items in an #MDB_DUPSORT database are also limited to + * this size, since they're actually keys of a sub-DB. Keys and + * #MDB_DUPSORT data items must fit on a node in a regular page. + */ +#ifndef MDB_MAXKEYSIZE +#define MDB_MAXKEYSIZE ((MDB_DEVEL) ? 0 : 511) +#endif + + /** The maximum size of a key we can write to the environment. */ +#if MDB_MAXKEYSIZE +#define ENV_MAXKEY(env) (MDB_MAXKEYSIZE) +#else +#define ENV_MAXKEY(env) ((env)->me_maxkey) +#endif + + /** @brief The maximum size of a data item. + * + * We only store a 32 bit value for node sizes. + */ +#define MAXDATASIZE 0xffffffffUL + +#if MDB_DEBUG + /** Key size which fits in a #DKBUF. + * @ingroup debug + */ +#define DKBUF_MAXKEYSIZE ((MDB_MAXKEYSIZE) > 0 ? (MDB_MAXKEYSIZE) : 511) + /** A key buffer. + * @ingroup debug + * This is used for printing a hex dump of a key's contents. + */ +#define DKBUF char kbuf[DKBUF_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 +#define DKEY(x) 0 +#endif + + /** An invalid page number. + * Mainly used to denote an empty tree. + */ +#define P_INVALID (~(pgno_t)0) + + /** Test if the flags \b f are set in a flag word \b w. */ +#define F_ISSET(w, f) (((w) & (f)) == (f)) + + /** Round \b n up to an even number. */ +#define EVEN(n) (((n) + 1U) & -2) /* sign-extending -2 to match n+1U */ + + /** 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. + * + * If #MDB_NOTLS is set, the slot address is not saved in thread-specific data. + * + * No reader table is used if the database is on a read-only filesystem, or + * if #MDB_NOLOCK is set. + * + * 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. + * @{ + */ + /** 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 { + /** Current Transaction ID when this transaction began, or (txnid_t)-1. + * 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. + */ + volatile txnid_t mrb_txnid; + /** The process ID of the process owning this reader txn. */ + volatile MDB_PID_T mrb_pid; + /** The thread ID of the thread owning this txn. */ + volatile MDB_THR_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 LMDB file. It must be set + * to #MDB_MAGIC. */ + uint32_t mtb_magic; + /** Format of this lock file. Must be set to #MDB_LOCK_FORMAT. */ + uint32_t mtb_format; +#if defined(_WIN32) + char mtb_rmname[MNAME_LEN]; +#elif defined(MDB_USE_SYSV_SEM) + int mtb_semid; + int mtb_rlocked; +#else + /** Mutex protecting access to this table. + * This is the #MDB_MUTEX(env,r) reader table lock. + */ + pthread_mutex_t mtb_rmutex; +#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. + */ + volatile 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. + */ + volatile unsigned mtb_numreaders; +} MDB_txbody; + + /** The actual reader table definition. */ +typedef struct MDB_txninfo { + union { + MDB_txbody mtb; +#define mti_magic mt1.mtb.mtb_magic +#define mti_format mt1.mtb.mtb_format +#define mti_rmutex mt1.mtb.mtb_rmutex +#define mti_rmname mt1.mtb.mtb_rmname +#define mti_txnid mt1.mtb.mtb_txnid +#define mti_numreaders mt1.mtb.mtb_numreaders +#ifdef MDB_USE_SYSV_SEM +#define mti_semid mt1.mtb.mtb_semid +#define mti_rlocked mt1.mtb.mtb_rlocked +#endif + char pad[(sizeof(MDB_txbody)+CACHELINE-1) & ~(CACHELINE-1)]; + } mt1; + union { +#if defined(_WIN32) + char mt2_wmname[MNAME_LEN]; +#define mti_wmname mt2.mt2_wmname +#elif defined MDB_USE_SYSV_SEM + int mt2_wlocked; +#define mti_wlocked mt2.mt2_wlocked +#else + pthread_mutex_t mt2_wmutex; +#define mti_wmutex mt2.mt2_wmutex +#endif + char pad[(MNAME_LEN+CACHELINE-1) & ~(CACHELINE-1)]; + } mt2; + MDB_reader mti_readers[1]; +} MDB_txninfo; + + /** Lockfile format signature: version, features and field layout */ +#define MDB_LOCK_FORMAT \ + ((uint32_t) \ + ((MDB_LOCK_VERSION) \ + /* Flags which describe functionality */ \ + + (SYSV_SEM_FLAG << 18) \ + + (((MDB_PIDLOCK) != 0) << 16))) +/** @} */ + +/** 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 */ + struct MDB_page *p_next; /**< for in-memory list of freed pages */ + } 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, also set for #P_SUBP pages */ +#define P_LEAF2 0x20 /**< for #MDB_DUPFIXED records */ +#define P_SUBP 0x40 /**< for #MDB_DUPSORT sub-pages */ +#define P_LOOSE 0x4000 /**< page was dirtied then freed, can be reused */ +#define P_KEEP 0x8000 /**< leave this page alone during spill */ +/** @} */ + 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)) + + /** ITS#7713, change PAGEBASE to handle 65536 byte pages */ +#define PAGEBASE ((MDB_DEVEL) ? PAGEHDRSZ : 0) + + /** Number of nodes on a page */ +#define NUMKEYS(p) (((p)->mp_lower - (PAGEHDRSZ-PAGEBASE)) >> 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) + + /** Link in #MDB_txn.%mt_loose_pgs list */ +#define NEXT_LOOSE_PAGE(p) (*(MDB_page **)((p) + 2)) + + /** Header for a single key/data pair within a page. + * Used in pages of type #P_BRANCH and #P_LEAF without #P_LEAF2. + * We guarantee 2-byte alignment for 'MDB_node's. + */ +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. + */ +#if BYTE_ORDER == LITTLE_ENDIAN + unsigned short mn_lo, mn_hi; /**< part of data size or pgno */ +#else + unsigned short mn_hi, mn_lo; +#endif +/** @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] + PAGEBASE)) + + /** 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) + + /** Copy a page number from src to dst */ +#ifdef MISALIGNED_OK +#define COPY_PGNO(dst,src) dst = src +#else +#if SIZE_MAX > 4294967295UL +#define COPY_PGNO(dst,src) do { \ + unsigned short *s, *d; \ + s = (unsigned short *)&(src); \ + d = (unsigned short *)&(dst); \ + *d++ = *s++; \ + *d++ = *s++; \ + *d++ = *s++; \ + *d = *s; \ +} while (0) +#else +#define COPY_PGNO(dst,src) do { \ + unsigned short *s, *d; \ + s = (unsigned short *)&(src); \ + d = (unsigned short *)&(dst); \ + *d++ = *s++; \ + *d = *s; \ +} while (0) +#endif +#endif + /** 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 keyptr, if requested. */ +#define MDB_GET_KEY(node, keyptr) { if ((keyptr) != NULL) { \ + (keyptr)->mv_size = NODEKSZ(node); (keyptr)->mv_data = NODEKEY(node); } } + + /** Set the \b node's key into \b key. */ +#define MDB_GET_KEY2(node, key) { 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_dbi_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; + + /** mdb_dbi_open flags */ +#define MDB_VALID 0x8000 /**< DB handle is valid, for me_dbflags */ +#define PERSISTENT_FLAGS (0xffff & ~(MDB_VALID)) +#define VALID_FLAGS (MDB_REVERSEKEY|MDB_DUPSORT|MDB_INTEGERKEY|MDB_DUPFIXED|\ + MDB_INTEGERDUP|MDB_REVERSEDUP|MDB_CREATE) + + /** 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. + * A meta page is the start point for accessing a database snapshot. + * Pages 0-1 are meta pages. Transaction N writes meta page #(N % 2). + */ +typedef struct MDB_meta { + /** Stamp identifying this as an LMDB file. It must be set + * to #MDB_MAGIC. */ + uint32_t mm_magic; + /** Version number of this file. Must be set to #MDB_DATA_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 */ + volatile txnid_t mm_txnid; /**< txnid that committed this page */ +} MDB_meta; + + /** Buffer for a stack-allocated meta page. + * The members define size and alignment, and silence type + * aliasing warnings. They are not used directly; that could + * mean incorrectly using several union members in parallel. + */ +typedef union MDB_metabuf { + MDB_page mb_page; + struct { + char mm_pad[PAGEHDRSZ]; + MDB_meta mm_meta; + } mb_metabuf; +} MDB_metabuf; + + /** 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. + */ + MDB_IDL mt_free_pgs; + /** The list of loose pages that became unused and may be reused + * in this transaction, linked through #NEXT_LOOSE_PAGE(page). + */ + MDB_page *mt_loose_pgs; + /* #Number of loose pages (#mt_loose_pgs) */ + int mt_loose_count; + /** The sorted list of dirty pages we temporarily wrote to disk + * because the dirty list was full. page numbers in here are + * shifted left by 1, deleted slots have the LSB set. + */ + MDB_IDL mt_spill_pgs; + union { + /** For write txns: Modified pages. Sorted when not MDB_WRITEMAP. */ + MDB_ID2L dirty_list; + /** For read txns: This thread/txn's reader table slot, or NULL. */ + MDB_reader *reader; + } 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; + /** Array of sequence numbers for each DB handle */ + unsigned int *mt_dbiseqs; +/** @defgroup mt_dbflag Transaction DB Flags + * @ingroup internal + * @{ + */ +#define DB_DIRTY 0x01 /**< DB was modified or is DUPSORT data */ +#define DB_STALE 0x02 /**< Named-DB record is older than txnID */ +#define DB_NEW 0x04 /**< Named-DB handle opened in this txn */ +#define DB_VALID 0x08 /**< DB handle is valid, see also #MDB_VALID */ +/** @} */ + /** In write txns, 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 /**< txn is unusable after an error */ +#define MDB_TXN_DIRTY 0x04 /**< must write, even if dirty list is empty */ +#define MDB_TXN_SPILLS 0x08 /**< txn or a parent has spilled pages */ +/** @} */ + unsigned int mt_flags; /**< @ref mdb_txn */ + /** #dirty_list room: Array size - \#dirty pages visible to this txn. + * Includes ancestor txns' dirty pages not hidden by other txns' + * dirty/spilled pages. Thus commit(nested txn) has room to merge + * dirty_list into mt_parent after freeing hidden mt_parent pages. + */ + unsigned int mt_dirty_room; +}; + +/** 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. + * A cursor holds a path of (page pointer, key index) from the DB + * root to a position in the DB, plus other state. #MDB_DUPSORT + * cursors include an xcursor to the current data item. Write txns + * track their cursors and keep them up to date when data moves. + * Exception: An xcursor's pointer to a #P_SUBP page can be stale. + * (A node with #F_DUPDATA but no #F_SUBDATA contains a subpage). + */ +struct MDB_cursor { + /** Next cursor on this DB in this txn */ + MDB_cursor *mc_next; + /** Backup of the original cursor if this cursor is a shadow */ + MDB_cursor *mc_backup; + /** 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, normally 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_DEL 0x08 /**< last op was a cursor_del */ +#define C_SPLITTING 0x20 /**< Cursor is in page_split */ +#define C_UNTRACK 0x40 /**< Un-track cursor when closing */ +/** @} */ + 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; + + /** State of FreeDB old pages, stored in the MDB_env */ +typedef struct MDB_pgstate { + pgno_t *mf_pghead; /**< Reclaimed freeDB pages, or NULL before use */ + txnid_t mf_pglast; /**< ID of last used record, or 0 if !mf_pghead */ +} MDB_pgstate; + + /** 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 + /** Some fields are initialized. */ +#define MDB_ENV_ACTIVE 0x20000000U + /** me_txkey is set */ +#define MDB_ENV_TXKEY 0x10000000U + /** fdatasync is unreliable */ +#define MDB_FSYNCONLY 0x08000000U + uint32_t me_flags; /**< @ref mdb_env */ + unsigned int me_psize; /**< DB page size, inited from me_os_psize */ + unsigned int me_os_psize; /**< OS page size, from #GET_PAGESIZE */ + unsigned int me_maxreaders; /**< size of the reader table */ + unsigned int me_numreaders; /**< max numreaders set by this env */ + MDB_dbi me_numdbs; /**< number of DBs opened */ + MDB_dbi me_maxdbs; /**< size of the DB table */ + MDB_PID_T me_pid; /**< process ID of this env */ + 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 or NULL */ + MDB_meta *me_metas[2]; /**< pointers to the two meta pages */ + void *me_pbuf; /**< scratch area for DUPSORT put() */ + MDB_txn *me_txn; /**< current write transaction */ + MDB_txn *me_txn0; /**< prealloc'd 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 */ + MDB_dbx *me_dbxs; /**< array of static DB info */ + uint16_t *me_dbflags; /**< array of flags from MDB_db.md_flags */ + unsigned int *me_dbiseqs; /**< array of dbi sequence numbers */ + pthread_key_t me_txkey; /**< thread-key for readers */ + txnid_t me_pgoldest; /**< ID of oldest reader last time we looked */ + MDB_pgstate me_pgstate; /**< state of old pages from freeDB */ +# define me_pglast me_pgstate.mf_pglast +# define me_pghead me_pgstate.mf_pghead + MDB_page *me_dpages; /**< list of malloc'd blocks for re-use */ + /** IDL of pages that became unused in a write txn */ + MDB_IDL me_free_pgs; + /** ID2L of pages written during a write txn. Length MDB_IDL_UM_SIZE. */ + MDB_ID2L me_dirty_list; + /** Max number of freelist items that can fit in a single overflow page */ + int me_maxfree_1pg; + /** Max size of a node on a page */ + unsigned int me_nodemax; +#if !(MDB_MAXKEYSIZE) + unsigned int me_maxkey; /**< max size of a key */ +#endif + int me_live_reader; /**< have liveness lock in reader table */ +#ifdef _WIN32 + int me_pidquery; /**< Used in OpenProcess */ +#endif +#if defined(_WIN32) || defined(MDB_USE_SYSV_SEM) + /* Windows mutexes/SysV semaphores do not reside in shared mem */ + mdb_mutex_t me_rmutex; + mdb_mutex_t me_wmutex; +#endif + void *me_userctx; /**< User-settable context */ + MDB_assert_func *me_assert_func; /**< Callback for assertion failures */ +}; + + /** Nested transaction */ +typedef struct MDB_ntxn { + MDB_txn mnt_txn; /**< the transaction */ + MDB_pgstate mnt_pgstate; /**< parent transaction's saved freestate */ +} MDB_ntxn; + + /** max number of pages to commit in one writev() call */ +#define MDB_COMMIT_PAGES 64 +#if defined(IOV_MAX) && IOV_MAX < MDB_COMMIT_PAGES +#undef MDB_COMMIT_PAGES +#define MDB_COMMIT_PAGES IOV_MAX +#endif + + /** max bytes to write in one call */ +#define MAX_WRITE (0x80000000U >> (sizeof(ssize_t) == 4)) + + /** Check \b txn and \b dbi arguments to a function */ +#define TXN_DBI_EXIST(txn, dbi) \ + ((txn) && (dbi) < (txn)->mt_numdbs && ((txn)->mt_dbflags[dbi] & DB_VALID)) + + /** Check for misused \b dbi handles */ +#define TXN_DBI_CHANGED(txn, dbi) \ + ((txn)->mt_dbiseqs[dbi] != (txn)->mt_env->me_dbiseqs[dbi]) + +static int mdb_page_alloc(MDB_cursor *mc, int num, MDB_page **mp); +static int mdb_page_new(MDB_cursor *mc, uint32_t flags, int num, MDB_page **mp); +static int mdb_page_touch(MDB_cursor *mc); + +static int mdb_page_get(MDB_txn *txn, pgno_t pgno, MDB_page **mp, int *lvl); +static int mdb_page_search_root(MDB_cursor *mc, + MDB_val *key, int modify); +#define MDB_PS_MODIFY 1 +#define MDB_PS_ROOTONLY 2 +#define MDB_PS_FIRST 4 +#define MDB_PS_LAST 8 +static int mdb_page_search(MDB_cursor *mc, + MDB_val *key, int flags); +static int mdb_page_merge(MDB_cursor *csrc, MDB_cursor *cdst); + +#define MDB_SPLIT_REPLACE MDB_APPENDDUP /**< newkey is not new */ +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_pick_meta(const MDB_env *env); +static int mdb_env_write_meta(MDB_txn *txn); +#if !(defined(_WIN32) || defined(MDB_USE_SYSV_SEM)) /* Drop unused excl arg */ +# define mdb_env_close0(env, excl) mdb_env_close1(env) +#endif +static void mdb_env_close0(MDB_env *env, int excl); + +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_cursor *mc, 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_cursor *mc, 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); +static int mdb_del0(MDB_txn *txn, MDB_dbi dbi, MDB_val *key, MDB_val *data, unsigned flags); +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); +static int mdb_reader_check0(MDB_env *env, int rlocked, int *dead); + +/** @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 LMDB @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 or environment had fatal error", + "MDB_VERSION_MISMATCH: Database environment version mismatch", + "MDB_INVALID: File is not an LMDB file", + "MDB_MAP_FULL: Environment mapsize limit reached", + "MDB_DBS_FULL: Environment maxdbs limit reached", + "MDB_READERS_FULL: Environment maxreaders limit reached", + "MDB_TLS_FULL: Thread-local storage keys full - too many environments open", + "MDB_TXN_FULL: Transaction has too many dirty pages - transaction too big", + "MDB_CURSOR_FULL: Internal error - cursor stack limit reached", + "MDB_PAGE_FULL: Internal error - page has no more space", + "MDB_MAP_RESIZED: Database contents grew beyond environment mapsize", + "MDB_INCOMPATIBLE: Operation and DB incompatible, or DB flags changed", + "MDB_BAD_RSLOT: Invalid reuse of reader locktable slot", + "MDB_BAD_TXN: Transaction cannot recover - it must be aborted", + "MDB_BAD_VALSIZE: Unsupported size of key/DB name/data, or wrong DUPFIXED size", + "MDB_BAD_DBI: The specified DBI handle was closed/changed unexpectedly", +}; + +char * +mdb_strerror(int err) +{ +#ifdef _WIN32 + /** HACK: pad 4KB on stack over the buf. Return system msgs in buf. + * This works as long as no function between the call to mdb_strerror + * and the actual use of the message uses more than 4K of stack. + */ + char pad[4096]; + char buf[1024], *ptr = buf; +#endif + int i; + if (!err) + return ("Successful return: 0"); + + if (err >= MDB_KEYEXIST && err <= MDB_LAST_ERRCODE) { + i = err - MDB_KEYEXIST; + return mdb_errstr[i]; + } + +#ifdef _WIN32 + /* These are the C-runtime error codes we use. The comment indicates + * their numeric value, and the Win32 error they would correspond to + * if the error actually came from a Win32 API. A major mess, we should + * have used LMDB-specific error codes for everything. + */ + switch(err) { + case ENOENT: /* 2, FILE_NOT_FOUND */ + case EIO: /* 5, ACCESS_DENIED */ + case ENOMEM: /* 12, INVALID_ACCESS */ + case EACCES: /* 13, INVALID_DATA */ + case EBUSY: /* 16, CURRENT_DIRECTORY */ + case EINVAL: /* 22, BAD_COMMAND */ + case ENOSPC: /* 28, OUT_OF_PAPER */ + return strerror(err); + default: + ; + } + buf[0] = 0; + FormatMessage(FORMAT_MESSAGE_FROM_SYSTEM | + FORMAT_MESSAGE_IGNORE_INSERTS, + NULL, err, 0, ptr, sizeof(buf), (va_list *)pad); + return ptr; +#else + return strerror(err); +#endif +} + +/** assert(3) variant in cursor context */ +#define mdb_cassert(mc, expr) mdb_assert0((mc)->mc_txn->mt_env, expr, #expr) +/** assert(3) variant in transaction context */ +#define mdb_tassert(mc, expr) mdb_assert0((txn)->mt_env, expr, #expr) +/** assert(3) variant in environment context */ +#define mdb_eassert(env, expr) mdb_assert0(env, expr, #expr) + +#ifndef NDEBUG +# define mdb_assert0(env, expr, expr_txt) ((expr) ? (void)0 : \ + mdb_assert_fail(env, expr_txt, mdb_func_, __FILE__, __LINE__)) + +static void +mdb_assert_fail(MDB_env *env, const char *expr_txt, + const char *func, const char *file, int line) +{ + char buf[400]; + sprintf(buf, "%.100s:%d: Assertion '%.200s' failed in %.40s()", + file, line, expr_txt, func); + if (env->me_assert_func) + env->me_assert_func(env, buf); + fprintf(stderr, "%s\n", buf); + abort(); +} +#else +# define mdb_assert0(env, expr, expr_txt) ((void) 0) +#endif /* NDEBUG */ + +#if MDB_DEBUG +/** Return the page number of \b mp which may be sub-page, for debug output */ +static pgno_t +mdb_dbg_pgno(MDB_page *mp) +{ + pgno_t ret; + COPY_PGNO(ret, mp->mp_pgno); + return ret; +} + +/** 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) + return ""; + + if (key->mv_size > DKBUF_MAXKEYSIZE) + return "MDB_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 + buf[0] = '\0'; + 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; +} + +static const char * +mdb_leafnode_type(MDB_node *n) +{ + static char *const tp[2][2] = {{"", ": DB"}, {": sub-page", ": sub-DB"}}; + return F_ISSET(n->mn_flags, F_BIGDATA) ? ": overflow page" : + tp[F_ISSET(n->mn_flags, F_DUPDATA)][F_ISSET(n->mn_flags, F_SUBDATA)]; +} + +/** Display all the keys in the page. */ +void +mdb_page_list(MDB_page *mp) +{ + pgno_t pgno = mdb_dbg_pgno(mp); + const char *type, *state = (mp->mp_flags & P_DIRTY) ? ", dirty" : ""; + MDB_node *node; + unsigned int i, nkeys, nsize, total = 0; + MDB_val key; + DKBUF; + + switch (mp->mp_flags & (P_BRANCH|P_LEAF|P_LEAF2|P_META|P_OVERFLOW|P_SUBP)) { + case P_BRANCH: type = "Branch page"; break; + case P_LEAF: type = "Leaf page"; break; + case P_LEAF|P_SUBP: type = "Sub-page"; break; + case P_LEAF|P_LEAF2: type = "LEAF2 page"; break; + case P_LEAF|P_LEAF2|P_SUBP: type = "LEAF2 sub-page"; break; + case P_OVERFLOW: + fprintf(stderr, "Overflow page %"Z"u pages %u%s\n", + pgno, mp->mp_pages, state); + return; + case P_META: + fprintf(stderr, "Meta-page %"Z"u txnid %"Z"u\n", + pgno, ((MDB_meta *)METADATA(mp))->mm_txnid); + return; + default: + fprintf(stderr, "Bad page %"Z"u flags 0x%u\n", pgno, mp->mp_flags); + return; + } + + nkeys = NUMKEYS(mp); + fprintf(stderr, "%s %"Z"u numkeys %d%s\n", type, pgno, nkeys, state); + + for (i=0; i<nkeys; i++) { + if (IS_LEAF2(mp)) { /* LEAF2 pages have no mp_ptrs[] or node headers */ + key.mv_size = nsize = mp->mp_pad; + key.mv_data = LEAF2KEY(mp, i, nsize); + total += nsize; + fprintf(stderr, "key %d: nsize %d, %s\n", i, nsize, DKEY(&key)); + continue; + } + node = NODEPTR(mp, i); + key.mv_size = node->mn_ksize; + key.mv_data = node->mn_data; + nsize = NODESIZE + key.mv_size; + if (IS_BRANCH(mp)) { + fprintf(stderr, "key %d: page %"Z"u, %s\n", i, NODEPGNO(node), + DKEY(&key)); + total += nsize; + } else { + if (F_ISSET(node->mn_flags, F_BIGDATA)) + nsize += sizeof(pgno_t); + else + nsize += NODEDSZ(node); + total += nsize; + nsize += sizeof(indx_t); + fprintf(stderr, "key %d: nsize %d, %s%s\n", + i, nsize, DKEY(&key), mdb_leafnode_type(node)); + } + total = EVEN(total); + } + fprintf(stderr, "Total: header %d + contents %d + unused %d\n", + IS_LEAF2(mp) ? PAGEHDRSZ : PAGEBASE + mp->mp_lower, total, SIZELEFT(mp)); +} + +void +mdb_cursor_chk(MDB_cursor *mc) +{ + unsigned int i; + MDB_node *node; + MDB_page *mp; + + if (!mc->mc_snum && !(mc->mc_flags & C_INITIALIZED)) return; + for (i=0; i<mc->mc_top; i++) { + mp = mc->mc_pg[i]; + node = NODEPTR(mp, mc->mc_ki[i]); + if (NODEPGNO(node) != mc->mc_pg[i+1]->mp_pgno) + printf("oops!\n"); + } + if (mc->mc_ki[i] >= NUMKEYS(mc->mc_pg[i])) + printf("ack!\n"); +} +#endif + +#if (MDB_DEBUG) > 2 +/** Count all the pages in each DB and in the freelist + * and make sure it matches the actual number of pages + * being used. + * All named DBs must be open for a correct count. + */ +static void mdb_audit(MDB_txn *txn) +{ + MDB_cursor mc; + MDB_val key, data; + MDB_ID freecount, count; + MDB_dbi i; + int rc; + + freecount = 0; + mdb_cursor_init(&mc, txn, FREE_DBI, NULL); + while ((rc = mdb_cursor_get(&mc, &key, &data, MDB_NEXT)) == 0) + freecount += *(MDB_ID *)data.mv_data; + mdb_tassert(txn, rc == MDB_NOTFOUND); + + count = 0; + for (i = 0; i<txn->mt_numdbs; i++) { + MDB_xcursor mx; + if (!(txn->mt_dbflags[i] & DB_VALID)) + continue; + mdb_cursor_init(&mc, txn, i, &mx); + if (txn->mt_dbs[i].md_root == P_INVALID) + continue; + count += txn->mt_dbs[i].md_branch_pages + + txn->mt_dbs[i].md_leaf_pages + + txn->mt_dbs[i].md_overflow_pages; + if (txn->mt_dbs[i].md_flags & MDB_DUPSORT) { + rc = mdb_page_search(&mc, NULL, MDB_PS_FIRST); + for (; rc == MDB_SUCCESS; rc = mdb_cursor_sibling(&mc, 1)) { + unsigned j; + MDB_page *mp; + mp = mc.mc_pg[mc.mc_top]; + for (j=0; j<NUMKEYS(mp); j++) { + MDB_node *leaf = NODEPTR(mp, j); + if (leaf->mn_flags & F_SUBDATA) { + MDB_db db; + memcpy(&db, NODEDATA(leaf), sizeof(db)); + count += db.md_branch_pages + db.md_leaf_pages + + db.md_overflow_pages; + } + } + } + mdb_tassert(txn, rc == MDB_NOTFOUND); + } + } + if (freecount + count + 2 /* metapages */ != txn->mt_next_pgno) { + fprintf(stderr, "audit: %lu freecount: %lu count: %lu total: %lu next_pgno: %lu\n", + txn->mt_txnid, freecount, count+2, freecount+count+2, txn->mt_next_pgno); + } +} +#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) +{ + return txn->mt_dbxs[dbi].md_dcmp(a, b); +} + +/** Allocate memory for a page. + * Re-use old malloc'd pages first for singletons, otherwise just malloc. + */ +static MDB_page * +mdb_page_malloc(MDB_txn *txn, unsigned num) +{ + MDB_env *env = txn->mt_env; + MDB_page *ret = env->me_dpages; + size_t psize = env->me_psize, sz = psize, off; + /* For ! #MDB_NOMEMINIT, psize counts how much to init. + * For a single page alloc, we init everything after the page header. + * For multi-page, we init the final page; if the caller needed that + * many pages they will be filling in at least up to the last page. + */ + if (num == 1) { + if (ret) { + VGMEMP_ALLOC(env, ret, sz); + VGMEMP_DEFINED(ret, sizeof(ret->mp_next)); + env->me_dpages = ret->mp_next; + return ret; + } + psize -= off = PAGEHDRSZ; + } else { + sz *= num; + off = sz - psize; + } + if ((ret = malloc(sz)) != NULL) { + VGMEMP_ALLOC(env, ret, sz); + if (!(env->me_flags & MDB_NOMEMINIT)) { + memset((char *)ret + off, 0, psize); + ret->mp_pad = 0; + } + } else { + txn->mt_flags |= MDB_TXN_ERROR; + } + return ret; +} +/** Free a single page. + * Saves single pages to a list, for future reuse. + * (This is not used for multi-page overflow pages.) + */ +static void +mdb_page_free(MDB_env *env, MDB_page *mp) +{ + mp->mp_next = env->me_dpages; + VGMEMP_FREE(env, mp); + env->me_dpages = mp; +} + +/** Free a dirty page */ +static void +mdb_dpage_free(MDB_env *env, MDB_page *dp) +{ + if (!IS_OVERFLOW(dp) || dp->mp_pages == 1) { + mdb_page_free(env, dp); + } else { + /* large pages just get freed directly */ + VGMEMP_FREE(env, dp); + free(dp); + } +} + +/** Return all dirty pages to dpage list */ +static void +mdb_dlist_free(MDB_txn *txn) +{ + MDB_env *env = txn->mt_env; + MDB_ID2L dl = txn->mt_u.dirty_list; + unsigned i, n = dl[0].mid; + + for (i = 1; i <= n; i++) { + mdb_dpage_free(env, dl[i].mptr); + } + dl[0].mid = 0; +} + +/** Loosen or free a single page. + * Saves single pages to a list for future reuse + * in this same txn. It has been pulled from the freeDB + * and already resides on the dirty list, but has been + * deleted. Use these pages first before pulling again + * from the freeDB. + * + * If the page wasn't dirtied in this txn, just add it + * to this txn's free list. + */ +static int +mdb_page_loose(MDB_cursor *mc, MDB_page *mp) +{ + int loose = 0; + pgno_t pgno = mp->mp_pgno; + MDB_txn *txn = mc->mc_txn; + + if ((mp->mp_flags & P_DIRTY) && mc->mc_dbi != FREE_DBI) { + if (txn->mt_parent) { + MDB_ID2 *dl = txn->mt_u.dirty_list; + /* If txn has a parent, make sure the page is in our + * dirty list. + */ + if (dl[0].mid) { + unsigned x = mdb_mid2l_search(dl, pgno); + if (x <= dl[0].mid && dl[x].mid == pgno) { + if (mp != dl[x].mptr) { /* bad cursor? */ + mc->mc_flags &= ~(C_INITIALIZED|C_EOF); + txn->mt_flags |= MDB_TXN_ERROR; + return MDB_CORRUPTED; + } + /* ok, it's ours */ + loose = 1; + } + } + } else { + /* no parent txn, so it's just ours */ + loose = 1; + } + } + if (loose) { + DPRINTF(("loosen db %d page %"Z"u", DDBI(mc), + mp->mp_pgno)); + NEXT_LOOSE_PAGE(mp) = txn->mt_loose_pgs; + txn->mt_loose_pgs = mp; + txn->mt_loose_count++; + mp->mp_flags |= P_LOOSE; + } else { + int rc = mdb_midl_append(&txn->mt_free_pgs, pgno); + if (rc) + return rc; + } + + return MDB_SUCCESS; +} + +/** Set or clear P_KEEP in dirty, non-overflow, non-sub pages watched by txn. + * @param[in] mc A cursor handle for the current operation. + * @param[in] pflags Flags of the pages to update: + * P_DIRTY to set P_KEEP, P_DIRTY|P_KEEP to clear it. + * @param[in] all No shortcuts. Needed except after a full #mdb_page_flush(). + * @return 0 on success, non-zero on failure. + */ +static int +mdb_pages_xkeep(MDB_cursor *mc, unsigned pflags, int all) +{ + enum { Mask = P_SUBP|P_DIRTY|P_LOOSE|P_KEEP }; + MDB_txn *txn = mc->mc_txn; + MDB_cursor *m3; + MDB_xcursor *mx; + MDB_page *dp, *mp; + MDB_node *leaf; + unsigned i, j; + int rc = MDB_SUCCESS, level; + + /* Mark pages seen by cursors */ + if (mc->mc_flags & C_UNTRACK) + mc = NULL; /* will find mc in mt_cursors */ + for (i = txn->mt_numdbs;; mc = txn->mt_cursors[--i]) { + for (; mc; mc=mc->mc_next) { + if (!(mc->mc_flags & C_INITIALIZED)) + continue; + for (m3 = mc;; m3 = &mx->mx_cursor) { + mp = NULL; + for (j=0; j<m3->mc_snum; j++) { + mp = m3->mc_pg[j]; + if ((mp->mp_flags & Mask) == pflags) + mp->mp_flags ^= P_KEEP; + } + mx = m3->mc_xcursor; + /* Proceed to mx if it is at a sub-database */ + if (! (mx && (mx->mx_cursor.mc_flags & C_INITIALIZED))) + break; + if (! (mp && (mp->mp_flags & P_LEAF))) + break; + leaf = NODEPTR(mp, m3->mc_ki[j-1]); + if (!(leaf->mn_flags & F_SUBDATA)) + break; + } + } + if (i == 0) + break; + } + + if (all) { + /* Mark dirty root pages */ + for (i=0; i<txn->mt_numdbs; i++) { + if (txn->mt_dbflags[i] & DB_DIRTY) { + pgno_t pgno = txn->mt_dbs[i].md_root; + if (pgno == P_INVALID) + continue; + if ((rc = mdb_page_get(txn, pgno, &dp, &level)) != MDB_SUCCESS) + break; + if ((dp->mp_flags & Mask) == pflags && level <= 1) + dp->mp_flags ^= P_KEEP; + } + } + } + + return rc; +} + +static int mdb_page_flush(MDB_txn *txn, int keep); + +/** Spill pages from the dirty list back to disk. + * This is intended to prevent running into #MDB_TXN_FULL situations, + * but note that they may still occur in a few cases: + * 1) our estimate of the txn size could be too small. Currently this + * seems unlikely, except with a large number of #MDB_MULTIPLE items. + * 2) child txns may run out of space if their parents dirtied a + * lot of pages and never spilled them. TODO: we probably should do + * a preemptive spill during #mdb_txn_begin() of a child txn, if + * the parent's dirty_room is below a given threshold. + * + * Otherwise, if not using nested txns, it is expected that apps will + * not run into #MDB_TXN_FULL any more. The pages are flushed to disk + * the same way as for a txn commit, e.g. their P_DIRTY flag is cleared. + * If the txn never references them again, they can be left alone. + * If the txn only reads them, they can be used without any fuss. + * If the txn writes them again, they can be dirtied immediately without + * going thru all of the work of #mdb_page_touch(). Such references are + * handled by #mdb_page_unspill(). + * + * Also note, we never spill DB root pages, nor pages of active cursors, + * because we'll need these back again soon anyway. And in nested txns, + * we can't spill a page in a child txn if it was already spilled in a + * parent txn. That would alter the parent txns' data even though + * the child hasn't committed yet, and we'd have no way to undo it if + * the child aborted. + * + * @param[in] m0 cursor A cursor handle identifying the transaction and + * database for which we are checking space. + * @param[in] key For a put operation, the key being stored. + * @param[in] data For a put operation, the data being stored. + * @return 0 on success, non-zero on failure. + */ +static int +mdb_page_spill(MDB_cursor *m0, MDB_val *key, MDB_val *data) +{ + MDB_txn *txn = m0->mc_txn; + MDB_page *dp; + MDB_ID2L dl = txn->mt_u.dirty_list; + unsigned int i, j, need; + int rc; + + if (m0->mc_flags & C_SUB) + return MDB_SUCCESS; + + /* Estimate how much space this op will take */ + i = m0->mc_db->md_depth; + /* Named DBs also dirty the main DB */ + if (m0->mc_dbi > MAIN_DBI) + i += txn->mt_dbs[MAIN_DBI].md_depth; + /* For puts, roughly factor in the key+data size */ + if (key) + i += (LEAFSIZE(key, data) + txn->mt_env->me_psize) / txn->mt_env->me_psize; + i += i; /* double it for good measure */ + need = i; + + if (txn->mt_dirty_room > i) + return MDB_SUCCESS; + + if (!txn->mt_spill_pgs) { + txn->mt_spill_pgs = mdb_midl_alloc(MDB_IDL_UM_MAX); + if (!txn->mt_spill_pgs) + return ENOMEM; + } else { + /* purge deleted slots */ + MDB_IDL sl = txn->mt_spill_pgs; + unsigned int num = sl[0]; + j=0; + for (i=1; i<=num; i++) { + if (!(sl[i] & 1)) + sl[++j] = sl[i]; + } + sl[0] = j; + } + + /* Preserve pages which may soon be dirtied again */ + if ((rc = mdb_pages_xkeep(m0, P_DIRTY, 1)) != MDB_SUCCESS) + goto done; + + /* Less aggressive spill - we originally spilled the entire dirty list, + * with a few exceptions for cursor pages and DB root pages. But this + * turns out to be a lot of wasted effort because in a large txn many + * of those pages will need to be used again. So now we spill only 1/8th + * of the dirty pages. Testing revealed this to be a good tradeoff, + * better than 1/2, 1/4, or 1/10. + */ + if (need < MDB_IDL_UM_MAX / 8) + need = MDB_IDL_UM_MAX / 8; + + /* Save the page IDs of all the pages we're flushing */ + /* flush from the tail forward, this saves a lot of shifting later on. */ + for (i=dl[0].mid; i && need; i--) { + MDB_ID pn = dl[i].mid << 1; + dp = dl[i].mptr; + if (dp->mp_flags & (P_LOOSE|P_KEEP)) + continue; + /* Can't spill twice, make sure it's not already in a parent's + * spill list. + */ + if (txn->mt_parent) { + MDB_txn *tx2; + for (tx2 = txn->mt_parent; tx2; tx2 = tx2->mt_parent) { + if (tx2->mt_spill_pgs) { + j = mdb_midl_search(tx2->mt_spill_pgs, pn); + if (j <= tx2->mt_spill_pgs[0] && tx2->mt_spill_pgs[j] == pn) { + dp->mp_flags |= P_KEEP; + break; + } + } + } + if (tx2) + continue; + } + if ((rc = mdb_midl_append(&txn->mt_spill_pgs, pn))) + goto done; + need--; + } + mdb_midl_sort(txn->mt_spill_pgs); + + /* Flush the spilled part of dirty list */ + if ((rc = mdb_page_flush(txn, i)) != MDB_SUCCESS) + goto done; + + /* Reset any dirty pages we kept that page_flush didn't see */ + rc = mdb_pages_xkeep(m0, P_DIRTY|P_KEEP, i); + +done: + txn->mt_flags |= rc ? MDB_TXN_ERROR : MDB_TXN_SPILLS; + return rc; +} + +/** Find oldest txnid still referenced. Expects txn->mt_txnid > 0. */ +static txnid_t +mdb_find_oldest(MDB_txn *txn) +{ + int i; + txnid_t mr, oldest = txn->mt_txnid - 1; + if (txn->mt_env->me_txns) { + MDB_reader *r = txn->mt_env->me_txns->mti_readers; + for (i = txn->mt_env->me_txns->mti_numreaders; --i >= 0; ) { + if (r[i].mr_pid) { + mr = r[i].mr_txnid; + if (oldest > mr) + oldest = mr; + } + } + } + return oldest; +} + +/** Add a page to the txn's dirty list */ +static void +mdb_page_dirty(MDB_txn *txn, MDB_page *mp) +{ + MDB_ID2 mid; + int rc, (*insert)(MDB_ID2L, MDB_ID2 *); + + if (txn->mt_env->me_flags & MDB_WRITEMAP) { + insert = mdb_mid2l_append; + } else { + insert = mdb_mid2l_insert; + } + mid.mid = mp->mp_pgno; + mid.mptr = mp; + rc = insert(txn->mt_u.dirty_list, &mid); + mdb_tassert(txn, rc == 0); + txn->mt_dirty_room--; +} + +/** Allocate page numbers and memory for writing. Maintain me_pglast, + * me_pghead and mt_next_pgno. + * + * If there are free pages available from older transactions, they + * are re-used first. Otherwise allocate a new page at mt_next_pgno. + * Do not modify the freedB, just merge freeDB records into me_pghead[] + * and move me_pglast to say which records were consumed. Only this + * function can create me_pghead and move me_pglast/mt_next_pgno. + * @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. + * @param[out] mp Address of the allocated page(s). Requests for multiple pages + * will always be satisfied by a single contiguous chunk of memory. + * @return 0 on success, non-zero on failure. + */ +static int +mdb_page_alloc(MDB_cursor *mc, int num, MDB_page **mp) +{ +#ifdef MDB_PARANOID /* Seems like we can ignore this now */ + /* Get at most <Max_retries> more freeDB records once me_pghead + * has enough pages. If not enough, use new pages from the map. + * If <Paranoid> and mc is updating the freeDB, only get new + * records if me_pghead is empty. Then the freelist cannot play + * catch-up with itself by growing while trying to save it. + */ + enum { Paranoid = 1, Max_retries = 500 }; +#else + enum { Paranoid = 0, Max_retries = INT_MAX /*infinite*/ }; +#endif + int rc, retry = num * 60; + MDB_txn *txn = mc->mc_txn; + MDB_env *env = txn->mt_env; + pgno_t pgno, *mop = env->me_pghead; + unsigned i, j, mop_len = mop ? mop[0] : 0, n2 = num-1; + MDB_page *np; + txnid_t oldest = 0, last; + MDB_cursor_op op; + MDB_cursor m2; + int found_old = 0; + + /* If there are any loose pages, just use them */ + if (num == 1 && txn->mt_loose_pgs) { + np = txn->mt_loose_pgs; + txn->mt_loose_pgs = NEXT_LOOSE_PAGE(np); + txn->mt_loose_count--; + DPRINTF(("db %d use loose page %"Z"u", DDBI(mc), + np->mp_pgno)); + *mp = np; + return MDB_SUCCESS; + } + + *mp = NULL; + + /* If our dirty list is already full, we can't do anything */ + if (txn->mt_dirty_room == 0) { + rc = MDB_TXN_FULL; + goto fail; + } + + for (op = MDB_FIRST;; op = MDB_NEXT) { + MDB_val key, data; + MDB_node *leaf; + pgno_t *idl; + + /* Seek a big enough contiguous page range. Prefer + * pages at the tail, just truncating the list. + */ + if (mop_len > n2) { + i = mop_len; + do { + pgno = mop[i]; + if (mop[i-n2] == pgno+n2) + goto search_done; + } while (--i > n2); + if (--retry < 0) + break; + } + + if (op == MDB_FIRST) { /* 1st iteration */ + /* Prepare to fetch more and coalesce */ + last = env->me_pglast; + oldest = env->me_pgoldest; + mdb_cursor_init(&m2, txn, FREE_DBI, NULL); + if (last) { + op = MDB_SET_RANGE; + key.mv_data = &last; /* will look up last+1 */ + key.mv_size = sizeof(last); + } + if (Paranoid && mc->mc_dbi == FREE_DBI) + retry = -1; + } + if (Paranoid && retry < 0 && mop_len) + break; + + last++; + /* Do not fetch more if the record will be too recent */ + if (oldest <= last) { + if (!found_old) { + oldest = mdb_find_oldest(txn); + env->me_pgoldest = oldest; + found_old = 1; + } + if (oldest <= last) + break; + } + rc = mdb_cursor_get(&m2, &key, NULL, op); + if (rc) { + if (rc == MDB_NOTFOUND) + break; + goto fail; + } + last = *(txnid_t*)key.mv_data; + if (oldest <= last) { + if (!found_old) { + oldest = mdb_find_oldest(txn); + env->me_pgoldest = oldest; + found_old = 1; + } + if (oldest <= last) + break; + } + np = m2.mc_pg[m2.mc_top]; + leaf = NODEPTR(np, m2.mc_ki[m2.mc_top]); + if ((rc = mdb_node_read(txn, leaf, &data)) != MDB_SUCCESS) + return rc; + + idl = (MDB_ID *) data.mv_data; + i = idl[0]; + if (!mop) { + if (!(env->me_pghead = mop = mdb_midl_alloc(i))) { + rc = ENOMEM; + goto fail; + } + } else { + if ((rc = mdb_midl_need(&env->me_pghead, i)) != 0) + goto fail; + mop = env->me_pghead; + } + env->me_pglast = last; +#if (MDB_DEBUG) > 1 + DPRINTF(("IDL read txn %"Z"u root %"Z"u num %u", + last, txn->mt_dbs[FREE_DBI].md_root, i)); + for (j = i; j; j--) + DPRINTF(("IDL %"Z"u", idl[j])); +#endif + /* Merge in descending sorted order */ + mdb_midl_xmerge(mop, idl); + mop_len = mop[0]; + } + + /* Use new pages from the map when nothing suitable in the freeDB */ + i = 0; + pgno = txn->mt_next_pgno; + if (pgno + num >= env->me_maxpg) { + DPUTS("DB size maxed out"); + rc = MDB_MAP_FULL; + goto fail; + } + +search_done: + if (env->me_flags & MDB_WRITEMAP) { + np = (MDB_page *)(env->me_map + env->me_psize * pgno); + } else { + if (!(np = mdb_page_malloc(txn, num))) { + rc = ENOMEM; + goto fail; + } + } + if (i) { + mop[0] = mop_len -= num; + /* Move any stragglers down */ + for (j = i-num; j < mop_len; ) + mop[++j] = mop[++i]; + } else { + txn->mt_next_pgno = pgno + num; + } + np->mp_pgno = pgno; + mdb_page_dirty(txn, np); + *mp = np; + + return MDB_SUCCESS; + +fail: + txn->mt_flags |= MDB_TXN_ERROR; + return rc; +} + +/** Copy the used portions of a non-overflow page. + * @param[in] dst page to copy into + * @param[in] src page to copy from + * @param[in] psize size of a page + */ +static void +mdb_page_copy(MDB_page *dst, MDB_page *src, unsigned int psize) +{ + enum { Align = sizeof(pgno_t) }; + indx_t upper = src->mp_upper, lower = src->mp_lower, unused = upper-lower; + + /* If page isn't full, just copy the used portion. Adjust + * alignment so memcpy may copy words instead of bytes. + */ + if ((unused &= -Align) && !IS_LEAF2(src)) { + upper = (upper + PAGEBASE) & -Align; + memcpy(dst, src, (lower + PAGEBASE + (Align-1)) & -Align); + memcpy((pgno_t *)((char *)dst+upper), (pgno_t *)((char *)src+upper), + psize - upper); + } else { + memcpy(dst, src, psize - unused); + } +} + +/** Pull a page off the txn's spill list, if present. + * If a page being referenced was spilled to disk in this txn, bring + * it back and make it dirty/writable again. + * @param[in] txn the transaction handle. + * @param[in] mp the page being referenced. It must not be dirty. + * @param[out] ret the writable page, if any. ret is unchanged if + * mp wasn't spilled. + */ +static int +mdb_page_unspill(MDB_txn *txn, MDB_page *mp, MDB_page **ret) +{ + MDB_env *env = txn->mt_env; + const MDB_txn *tx2; + unsigned x; + pgno_t pgno = mp->mp_pgno, pn = pgno << 1; + + for (tx2 = txn; tx2; tx2=tx2->mt_parent) { + if (!tx2->mt_spill_pgs) + continue; + x = mdb_midl_search(tx2->mt_spill_pgs, pn); + if (x <= tx2->mt_spill_pgs[0] && tx2->mt_spill_pgs[x] == pn) { + MDB_page *np; + int num; + if (txn->mt_dirty_room == 0) + return MDB_TXN_FULL; + if (IS_OVERFLOW(mp)) + num = mp->mp_pages; + else + num = 1; + if (env->me_flags & MDB_WRITEMAP) { + np = mp; + } else { + np = mdb_page_malloc(txn, num); + if (!np) + return ENOMEM; + if (num > 1) + memcpy(np, mp, num * env->me_psize); + else + mdb_page_copy(np, mp, env->me_psize); + } + if (tx2 == txn) { + /* If in current txn, this page is no longer spilled. + * If it happens to be the last page, truncate the spill list. + * Otherwise mark it as deleted by setting the LSB. + */ + if (x == txn->mt_spill_pgs[0]) + txn->mt_spill_pgs[0]--; + else + txn->mt_spill_pgs[x] |= 1; + } /* otherwise, if belonging to a parent txn, the + * page remains spilled until child commits + */ + + mdb_page_dirty(txn, np); + np->mp_flags |= P_DIRTY; + *ret = np; + break; + } + } + return MDB_SUCCESS; +} + +/** 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], *np; + MDB_txn *txn = mc->mc_txn; + MDB_cursor *m2, *m3; + pgno_t pgno; + int rc; + + if (!F_ISSET(mp->mp_flags, P_DIRTY)) { + if (txn->mt_flags & MDB_TXN_SPILLS) { + np = NULL; + rc = mdb_page_unspill(txn, mp, &np); + if (rc) + goto fail; + if (np) + goto done; + } + if ((rc = mdb_midl_need(&txn->mt_free_pgs, 1)) || + (rc = mdb_page_alloc(mc, 1, &np))) + goto fail; + pgno = np->mp_pgno; + DPRINTF(("touched db %d page %"Z"u -> %"Z"u", DDBI(mc), + mp->mp_pgno, pgno)); + mdb_cassert(mc, mp->mp_pgno != pgno); + mdb_midl_xappend(txn->mt_free_pgs, mp->mp_pgno); + /* Update the parent page, if any, to point to the new page */ + if (mc->mc_top) { + MDB_page *parent = mc->mc_pg[mc->mc_top-1]; + MDB_node *node = NODEPTR(parent, mc->mc_ki[mc->mc_top-1]); + SETPGNO(node, pgno); + } else { + mc->mc_db->md_root = pgno; + } + } else if (txn->mt_parent && !IS_SUBP(mp)) { + MDB_ID2 mid, *dl = txn->mt_u.dirty_list; + pgno = mp->mp_pgno; + /* If txn has a parent, make sure the page is in our + * dirty list. + */ + if (dl[0].mid) { + unsigned x = mdb_mid2l_search(dl, pgno); + if (x <= dl[0].mid && dl[x].mid == pgno) { + if (mp != dl[x].mptr) { /* bad cursor? */ + mc->mc_flags &= ~(C_INITIALIZED|C_EOF); + txn->mt_flags |= MDB_TXN_ERROR; + return MDB_CORRUPTED; + } + return 0; + } + } + mdb_cassert(mc, dl[0].mid < MDB_IDL_UM_MAX); + /* No - copy it */ + np = mdb_page_malloc(txn, 1); + if (!np) + return ENOMEM; + mid.mid = pgno; + mid.mptr = np; + rc = mdb_mid2l_insert(dl, &mid); + mdb_cassert(mc, rc == 0); + } else { + return 0; + } + + mdb_page_copy(np, mp, txn->mt_env->me_psize); + np->mp_pgno = pgno; + np->mp_flags |= P_DIRTY; + +done: + /* Adjust cursors pointing to mp */ + mc->mc_pg[mc->mc_top] = np; + m2 = txn->mt_cursors[mc->mc_dbi]; + if (mc->mc_flags & C_SUB) { + for (; m2; m2=m2->mc_next) { + m3 = &m2->mc_xcursor->mx_cursor; + if (m3->mc_snum < mc->mc_snum) continue; + if (m3->mc_pg[mc->mc_top] == mp) + m3->mc_pg[mc->mc_top] = np; + } + } else { + for (; m2; m2=m2->mc_next) { + if (m2->mc_snum < mc->mc_snum) continue; + if (m2->mc_pg[mc->mc_top] == mp) { + m2->mc_pg[mc->mc_top] = np; + if ((mc->mc_db->md_flags & MDB_DUPSORT) && + IS_LEAF(np) && + m2->mc_ki[mc->mc_top] == mc->mc_ki[mc->mc_top]) + { + MDB_node *leaf = NODEPTR(np, mc->mc_ki[mc->mc_top]); + if (!(leaf->mn_flags & F_SUBDATA)) + m2->mc_xcursor->mx_cursor.mc_pg[0] = NODEDATA(leaf); + } + } + } + } + return 0; + +fail: + txn->mt_flags |= MDB_TXN_ERROR; + return rc; +} + +int +mdb_env_sync(MDB_env *env, int force) +{ + int rc = 0; + if (env->me_flags & MDB_RDONLY) + return EACCES; + if (force || !F_ISSET(env->me_flags, MDB_NOSYNC)) { + if (env->me_flags & MDB_WRITEMAP) { + int flags = ((env->me_flags & MDB_MAPASYNC) && !force) + ? MS_ASYNC : MS_SYNC; + if (MDB_MSYNC(env->me_map, env->me_mapsize, flags)) + rc = ErrCode(); +#ifdef _WIN32 + else if (flags == MS_SYNC && MDB_FDATASYNC(env->me_fd)) + rc = ErrCode(); +#endif + } else { +#ifdef BROKEN_FDATASYNC + if (env->me_flags & MDB_FSYNCONLY) { + if (fsync(env->me_fd)) + rc = ErrCode(); + } else +#endif + if (MDB_FDATASYNC(env->me_fd)) + rc = ErrCode(); + } + } + return rc; +} + +/** Back up parent txn's cursors, then grab the originals for tracking */ +static int +mdb_cursor_shadow(MDB_txn *src, MDB_txn *dst) +{ + MDB_cursor *mc, *bk; + MDB_xcursor *mx; + size_t size; + int i; + + for (i = src->mt_numdbs; --i >= 0; ) { + if ((mc = src->mt_cursors[i]) != NULL) { + size = sizeof(MDB_cursor); + if (mc->mc_xcursor) + size += sizeof(MDB_xcursor); + for (; mc; mc = bk->mc_next) { + bk = malloc(size); + if (!bk) + return ENOMEM; + *bk = *mc; + mc->mc_backup = bk; + mc->mc_db = &dst->mt_dbs[i]; + /* Kill pointers into src - and dst to reduce abuse: The + * user may not use mc until dst ends. Otherwise we'd... + */ + mc->mc_txn = NULL; /* ...set this to dst */ + mc->mc_dbflag = NULL; /* ...and &dst->mt_dbflags[i] */ + if ((mx = mc->mc_xcursor) != NULL) { + *(MDB_xcursor *)(bk+1) = *mx; + mx->mx_cursor.mc_txn = NULL; /* ...and dst. */ + } + mc->mc_next = dst->mt_cursors[i]; + dst->mt_cursors[i] = mc; + } + } + } + return MDB_SUCCESS; +} + +/** Close this write txn's cursors, give parent txn's cursors back to parent. + * @param[in] txn the transaction handle. + * @param[in] merge true to keep changes to parent cursors, false to revert. + * @return 0 on success, non-zero on failure. + */ +static void +mdb_cursors_close(MDB_txn *txn, unsigned merge) +{ + MDB_cursor **cursors = txn->mt_cursors, *mc, *next, *bk; + MDB_xcursor *mx; + int i; + + for (i = txn->mt_numdbs; --i >= 0; ) { + for (mc = cursors[i]; mc; mc = next) { + next = mc->mc_next; + if ((bk = mc->mc_backup) != NULL) { + if (merge) { + /* Commit changes to parent txn */ + mc->mc_next = bk->mc_next; + mc->mc_backup = bk->mc_backup; + mc->mc_txn = bk->mc_txn; + mc->mc_db = bk->mc_db; + mc->mc_dbflag = bk->mc_dbflag; + if ((mx = mc->mc_xcursor) != NULL) + mx->mx_cursor.mc_txn = bk->mc_txn; + } else { + /* Abort nested txn */ + *mc = *bk; + if ((mx = mc->mc_xcursor) != NULL) + *mx = *(MDB_xcursor *)(bk+1); + } + mc = bk; + } + /* Only malloced cursors are permanently tracked. */ + free(mc); + } + cursors[i] = NULL; + } +} + +#if !(MDB_DEBUG) +#define mdb_txn_reset0(txn, act) mdb_txn_reset0(txn) +#endif +static void +mdb_txn_reset0(MDB_txn *txn, const char *act); + +#if !(MDB_PIDLOCK) /* Currently the same as defined(_WIN32) */ +enum Pidlock_op { + Pidset, Pidcheck +}; +#else +enum Pidlock_op { + Pidset = F_SETLK, Pidcheck = F_GETLK +}; +#endif + +/** Set or check a pid lock. Set returns 0 on success. + * Check returns 0 if the process is certainly dead, nonzero if it may + * be alive (the lock exists or an error happened so we do not know). + * + * On Windows Pidset is a no-op, we merely check for the existence + * of the process with the given pid. On POSIX we use a single byte + * lock on the lockfile, set at an offset equal to the pid. + */ +static int +mdb_reader_pid(MDB_env *env, enum Pidlock_op op, MDB_PID_T pid) +{ +#if !(MDB_PIDLOCK) /* Currently the same as defined(_WIN32) */ + int ret = 0; + HANDLE h; + if (op == Pidcheck) { + h = OpenProcess(env->me_pidquery, FALSE, pid); + /* No documented "no such process" code, but other program use this: */ + if (!h) + return ErrCode() != ERROR_INVALID_PARAMETER; + /* A process exists until all handles to it close. Has it exited? */ + ret = WaitForSingleObject(h, 0) != 0; + CloseHandle(h); + } + return ret; +#else + for (;;) { + int rc; + struct flock lock_info; + memset(&lock_info, 0, sizeof(lock_info)); + lock_info.l_type = F_WRLCK; + lock_info.l_whence = SEEK_SET; + lock_info.l_start = pid; + lock_info.l_len = 1; + if ((rc = fcntl(env->me_lfd, op, &lock_info)) == 0) { + if (op == F_GETLK && lock_info.l_type != F_UNLCK) + rc = -1; + } else if ((rc = ErrCode()) == EINTR) { + continue; + } + return rc; + } +#endif +} + +/** 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. + */ +static int +mdb_txn_renew0(MDB_txn *txn) +{ + MDB_env *env = txn->mt_env; + MDB_txninfo *ti = env->me_txns; + MDB_meta *meta; + unsigned int i, nr; + uint16_t x; + int rc, new_notls = 0; + + if (txn->mt_flags & MDB_TXN_RDONLY) { + /* Setup db info */ + txn->mt_numdbs = env->me_numdbs; + txn->mt_dbxs = env->me_dbxs; /* mostly static anyway */ + if (!ti) { + meta = env->me_metas[ mdb_env_pick_meta(env) ]; + txn->mt_txnid = meta->mm_txnid; + txn->mt_u.reader = NULL; + } else { + MDB_reader *r = (env->me_flags & MDB_NOTLS) ? txn->mt_u.reader : + pthread_getspecific(env->me_txkey); + if (r) { + if (r->mr_pid != env->me_pid || r->mr_txnid != (txnid_t)-1) + return MDB_BAD_RSLOT; + } else { + MDB_PID_T pid = env->me_pid; + MDB_THR_T tid = pthread_self(); + mdb_mutex_t *rmutex = MDB_MUTEX(env, r); + + if (!env->me_live_reader) { + rc = mdb_reader_pid(env, Pidset, pid); + if (rc) + return rc; + env->me_live_reader = 1; + } + + if (LOCK_MUTEX(rc, env, rmutex)) + return rc; + nr = ti->mti_numreaders; + for (i=0; i<nr; i++) + if (ti->mti_readers[i].mr_pid == 0) + break; + if (i == env->me_maxreaders) { + UNLOCK_MUTEX(rmutex); + return MDB_READERS_FULL; + } + r = &ti->mti_readers[i]; + r->mr_txnid = (txnid_t)-1; + r->mr_tid = tid; + r->mr_pid = pid; /* should be written last, see ITS#7971. */ + if (i == nr) + ti->mti_numreaders = ++nr; + /* Save numreaders for un-mutexed mdb_env_close() */ + env->me_numreaders = nr; + UNLOCK_MUTEX(rmutex); + + new_notls = (env->me_flags & MDB_NOTLS); + if (!new_notls && (rc=pthread_setspecific(env->me_txkey, r))) { + r->mr_pid = 0; + return rc; + } + } + do /* LY: Retry on a race, ITS#7970. */ + r->mr_txnid = ti->mti_txnid; + while(r->mr_txnid != ti->mti_txnid); + txn->mt_txnid = r->mr_txnid; + txn->mt_u.reader = r; + meta = env->me_metas[txn->mt_txnid & 1]; + } + } else { + if (ti) { + if (LOCK_MUTEX(rc, env, MDB_MUTEX(env, w))) + return rc; + txn->mt_txnid = ti->mti_txnid; + meta = env->me_metas[txn->mt_txnid & 1]; + } else { + meta = env->me_metas[ mdb_env_pick_meta(env) ]; + txn->mt_txnid = meta->mm_txnid; + } + /* Setup db info */ + txn->mt_numdbs = env->me_numdbs; + txn->mt_txnid++; +#if MDB_DEBUG + if (txn->mt_txnid == mdb_debug_start) + mdb_debug = 1; +#endif + txn->mt_flags = 0; + txn->mt_child = NULL; + txn->mt_loose_pgs = NULL; + txn->mt_loose_count = 0; + txn->mt_dirty_room = MDB_IDL_UM_MAX; + 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_spill_pgs = NULL; + env->me_txn = txn; + memcpy(txn->mt_dbiseqs, env->me_dbiseqs, env->me_maxdbs * sizeof(unsigned int)); + } + + /* Copy the DB info and flags */ + memcpy(txn->mt_dbs, meta->mm_dbs, 2 * sizeof(MDB_db)); + + /* Moved to here to avoid a data race in read TXNs */ + txn->mt_next_pgno = meta->mm_last_pg+1; + + for (i=2; i<txn->mt_numdbs; i++) { + x = env->me_dbflags[i]; + txn->mt_dbs[i].md_flags = x & PERSISTENT_FLAGS; + txn->mt_dbflags[i] = (x & MDB_VALID) ? DB_VALID|DB_STALE : 0; + } + txn->mt_dbflags[0] = txn->mt_dbflags[1] = DB_VALID; + + if (env->me_maxpg < txn->mt_next_pgno) { + mdb_txn_reset0(txn, "renew0-mapfail"); + if (new_notls) { + txn->mt_u.reader->mr_pid = 0; + txn->mt_u.reader = NULL; + } + return MDB_MAP_RESIZED; + } + + return MDB_SUCCESS; +} + +int +mdb_txn_renew(MDB_txn *txn) +{ + int rc; + + if (!txn || txn->mt_dbxs) /* A reset txn has mt_dbxs==NULL */ + 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 %"Z"u%c %p on mdbenv %p, root page %"Z"u", + 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; + MDB_ntxn *ntxn; + int rc, size, tsize = sizeof(MDB_txn); + + if (env->me_flags & MDB_FATAL_ERROR) { + DPUTS("environment had fatal error, must shutdown!"); + return MDB_PANIC; + } + if ((env->me_flags & MDB_RDONLY) && !(flags & MDB_RDONLY)) + return EACCES; + if (parent) { + /* Nested transactions: Max 1 child, write txns only, no writemap */ + if (parent->mt_child || + (flags & MDB_RDONLY) || + (parent->mt_flags & (MDB_TXN_RDONLY|MDB_TXN_ERROR)) || + (env->me_flags & MDB_WRITEMAP)) + { + return (parent->mt_flags & MDB_TXN_RDONLY) ? EINVAL : MDB_BAD_TXN; + } + tsize = sizeof(MDB_ntxn); + } + size = tsize; + if (!(flags & MDB_RDONLY)) { + if (!parent) { + txn = env->me_txn0; /* just reuse preallocated write txn */ + goto ok; + } + /* child txns use own copy of cursors */ + size += env->me_maxdbs * sizeof(MDB_cursor *); + } + size += env->me_maxdbs * (sizeof(MDB_db)+1); + + if ((txn = calloc(1, size)) == NULL) { + DPRINTF(("calloc: %s", strerror(errno))); + return ENOMEM; + } + txn->mt_dbs = (MDB_db *) ((char *)txn + tsize); + if (flags & MDB_RDONLY) { + txn->mt_flags |= MDB_TXN_RDONLY; + txn->mt_dbflags = (unsigned char *)(txn->mt_dbs + env->me_maxdbs); + txn->mt_dbiseqs = env->me_dbiseqs; + } else { + txn->mt_cursors = (MDB_cursor **)(txn->mt_dbs + env->me_maxdbs); + if (parent) { + txn->mt_dbiseqs = parent->mt_dbiseqs; + txn->mt_dbflags = (unsigned char *)(txn->mt_cursors + env->me_maxdbs); + } else { + txn->mt_dbiseqs = (unsigned int *)(txn->mt_cursors + env->me_maxdbs); + txn->mt_dbflags = (unsigned char *)(txn->mt_dbiseqs + env->me_maxdbs); + } + } + txn->mt_env = env; + +ok: + if (parent) { + unsigned int i; + txn->mt_u.dirty_list = malloc(sizeof(MDB_ID2)*MDB_IDL_UM_SIZE); + if (!txn->mt_u.dirty_list || + !(txn->mt_free_pgs = mdb_midl_alloc(MDB_IDL_UM_MAX))) + { + free(txn->mt_u.dirty_list); + free(txn); + return ENOMEM; + } + txn->mt_txnid = parent->mt_txnid; + txn->mt_dirty_room = parent->mt_dirty_room; + txn->mt_u.dirty_list[0].mid = 0; + txn->mt_spill_pgs = NULL; + txn->mt_next_pgno = parent->mt_next_pgno; + parent->mt_child = txn; + txn->mt_parent = parent; + txn->mt_numdbs = parent->mt_numdbs; + txn->mt_flags = parent->mt_flags; + txn->mt_dbxs = parent->mt_dbxs; + memcpy(txn->mt_dbs, parent->mt_dbs, txn->mt_numdbs * sizeof(MDB_db)); + /* Copy parent's mt_dbflags, but clear DB_NEW */ + for (i=0; i<txn->mt_numdbs; i++) + txn->mt_dbflags[i] = parent->mt_dbflags[i] & ~DB_NEW; + rc = 0; + ntxn = (MDB_ntxn *)txn; + ntxn->mnt_pgstate = env->me_pgstate; /* save parent me_pghead & co */ + if (env->me_pghead) { + size = MDB_IDL_SIZEOF(env->me_pghead); + env->me_pghead = mdb_midl_alloc(env->me_pghead[0]); + if (env->me_pghead) + memcpy(env->me_pghead, ntxn->mnt_pgstate.mf_pghead, size); + else + rc = ENOMEM; + } + if (!rc) + rc = mdb_cursor_shadow(parent, txn); + if (rc) + mdb_txn_reset0(txn, "beginchild-fail"); + } else { + rc = mdb_txn_renew0(txn); + } + if (rc) { + if (txn != env->me_txn0) + free(txn); + } else { + *ret = txn; + DPRINTF(("begin txn %"Z"u%c %p on mdbenv %p, root page %"Z"u", + txn->mt_txnid, (txn->mt_flags & MDB_TXN_RDONLY) ? 'r' : 'w', + (void *) txn, (void *) env, txn->mt_dbs[MAIN_DBI].md_root)); + } + + return rc; +} + +MDB_env * +mdb_txn_env(MDB_txn *txn) +{ + if(!txn) return NULL; + return txn->mt_env; +} + +size_t +mdb_txn_id(MDB_txn *txn) +{ + if(!txn) return 0; + return txn->mt_txnid; +} + +/** Export or close DBI handles opened in this txn. */ +static void +mdb_dbis_update(MDB_txn *txn, int keep) +{ + int i; + MDB_dbi n = txn->mt_numdbs; + MDB_env *env = txn->mt_env; + unsigned char *tdbflags = txn->mt_dbflags; + + for (i = n; --i >= 2;) { + if (tdbflags[i] & DB_NEW) { + if (keep) { + env->me_dbflags[i] = txn->mt_dbs[i].md_flags | MDB_VALID; + } else { + char *ptr = env->me_dbxs[i].md_name.mv_data; + if (ptr) { + env->me_dbxs[i].md_name.mv_data = NULL; + env->me_dbxs[i].md_name.mv_size = 0; + env->me_dbflags[i] = 0; + env->me_dbiseqs[i]++; + free(ptr); + } + } + } + } + if (keep && env->me_numdbs < n) + env->me_numdbs = n; +} + +/** Common code for #mdb_txn_reset() and #mdb_txn_abort(). + * May be called twice for readonly txns: First reset it, then abort. + * @param[in] txn the transaction handle to reset + * @param[in] act why the transaction is being reset + */ +static void +mdb_txn_reset0(MDB_txn *txn, const char *act) +{ + MDB_env *env = txn->mt_env; + + /* Close any DBI handles opened in this txn */ + mdb_dbis_update(txn, 0); + + DPRINTF(("%s txn %"Z"u%c %p on mdbenv %p, root page %"Z"u", + act, txn->mt_txnid, (txn->mt_flags & MDB_TXN_RDONLY) ? 'r' : 'w', + (void *) txn, (void *)env, txn->mt_dbs[MAIN_DBI].md_root)); + + if (F_ISSET(txn->mt_flags, MDB_TXN_RDONLY)) { + if (txn->mt_u.reader) { + txn->mt_u.reader->mr_txnid = (txnid_t)-1; + if (!(env->me_flags & MDB_NOTLS)) + txn->mt_u.reader = NULL; /* txn does not own reader */ + } + txn->mt_numdbs = 0; /* close nothing if called again */ + txn->mt_dbxs = NULL; /* mark txn as reset */ + } else { + pgno_t *pghead = env->me_pghead; + + mdb_cursors_close(txn, 0); + if (!(env->me_flags & MDB_WRITEMAP)) { + mdb_dlist_free(txn); + } + + if (!txn->mt_parent) { + if (mdb_midl_shrink(&txn->mt_free_pgs)) + env->me_free_pgs = txn->mt_free_pgs; + /* me_pgstate: */ + env->me_pghead = NULL; + env->me_pglast = 0; + + env->me_txn = NULL; + /* The writer mutex was locked in mdb_txn_begin. */ + if (env->me_txns) + UNLOCK_MUTEX(MDB_MUTEX(env, w)); + } else { + txn->mt_parent->mt_child = NULL; + env->me_pgstate = ((MDB_ntxn *)txn)->mnt_pgstate; + mdb_midl_free(txn->mt_free_pgs); + mdb_midl_free(txn->mt_spill_pgs); + free(txn->mt_u.dirty_list); + } + + mdb_midl_free(pghead); + } +} + +void +mdb_txn_reset(MDB_txn *txn) +{ + if (txn == NULL) + return; + + /* This call is only valid for read-only txns */ + if (!(txn->mt_flags & MDB_TXN_RDONLY)) + return; + + mdb_txn_reset0(txn, "reset"); +} + +void +mdb_txn_abort(MDB_txn *txn) +{ + if (txn == NULL) + return; + + if (txn->mt_child) + mdb_txn_abort(txn->mt_child); + + mdb_txn_reset0(txn, "abort"); + /* Free reader slot tied to this txn (if MDB_NOTLS && writable FS) */ + if ((txn->mt_flags & MDB_TXN_RDONLY) && txn->mt_u.reader) + txn->mt_u.reader->mr_pid = 0; + + if (txn != txn->mt_env->me_txn0) + free(txn); +} + +/** Save the freelist as of this transaction to the freeDB. + * This changes the freelist. Keep trying until it stabilizes. + */ +static int +mdb_freelist_save(MDB_txn *txn) +{ + /* env->me_pghead[] can grow and shrink during this call. + * env->me_pglast and txn->mt_free_pgs[] can only grow. + * Page numbers cannot disappear from txn->mt_free_pgs[]. + */ + MDB_cursor mc; + MDB_env *env = txn->mt_env; + int rc, maxfree_1pg = env->me_maxfree_1pg, more = 1; + txnid_t pglast = 0, head_id = 0; + pgno_t freecnt = 0, *free_pgs, *mop; + ssize_t head_room = 0, total_room = 0, mop_len, clean_limit; + + mdb_cursor_init(&mc, txn, FREE_DBI, NULL); + + if (env->me_pghead) { + /* Make sure first page of freeDB is touched and on freelist */ + rc = mdb_page_search(&mc, NULL, MDB_PS_FIRST|MDB_PS_MODIFY); + if (rc && rc != MDB_NOTFOUND) + return rc; + } + + if (!env->me_pghead && txn->mt_loose_pgs) { + /* Put loose page numbers in mt_free_pgs, since + * we may be unable to return them to me_pghead. + */ + MDB_page *mp = txn->mt_loose_pgs; + if ((rc = mdb_midl_need(&txn->mt_free_pgs, txn->mt_loose_count)) != 0) + return rc; + for (; mp; mp = NEXT_LOOSE_PAGE(mp)) + mdb_midl_xappend(txn->mt_free_pgs, mp->mp_pgno); + txn->mt_loose_pgs = NULL; + txn->mt_loose_count = 0; + } + + /* MDB_RESERVE cancels meminit in ovpage malloc (when no WRITEMAP) */ + clean_limit = (env->me_flags & (MDB_NOMEMINIT|MDB_WRITEMAP)) + ? SSIZE_MAX : maxfree_1pg; + + for (;;) { + /* Come back here after each Put() in case freelist changed */ + MDB_val key, data; + pgno_t *pgs; + ssize_t j; + + /* If using records from freeDB which we have not yet + * deleted, delete them and any we reserved for me_pghead. + */ + while (pglast < env->me_pglast) { + rc = mdb_cursor_first(&mc, &key, NULL); + if (rc) + return rc; + pglast = head_id = *(txnid_t *)key.mv_data; + total_room = head_room = 0; + mdb_tassert(txn, pglast <= env->me_pglast); + rc = mdb_cursor_del(&mc, 0); + if (rc) + return rc; + } + + /* Save the IDL of pages freed by this txn, to a single record */ + if (freecnt < txn->mt_free_pgs[0]) { + if (!freecnt) { + /* Make sure last page of freeDB is touched and on freelist */ + rc = mdb_page_search(&mc, NULL, MDB_PS_LAST|MDB_PS_MODIFY); + if (rc && rc != MDB_NOTFOUND) + return rc; + } + free_pgs = txn->mt_free_pgs; + /* Write to last page of freeDB */ + key.mv_size = sizeof(txn->mt_txnid); + key.mv_data = &txn->mt_txnid; + do { + freecnt = free_pgs[0]; + data.mv_size = MDB_IDL_SIZEOF(free_pgs); + rc = mdb_cursor_put(&mc, &key, &data, MDB_RESERVE); + if (rc) + return rc; + /* Retry if mt_free_pgs[] grew during the Put() */ + free_pgs = txn->mt_free_pgs; + } while (freecnt < free_pgs[0]); + mdb_midl_sort(free_pgs); + memcpy(data.mv_data, free_pgs, data.mv_size); +#if (MDB_DEBUG) > 1 + { + unsigned int i = free_pgs[0]; + DPRINTF(("IDL write txn %"Z"u root %"Z"u num %u", + txn->mt_txnid, txn->mt_dbs[FREE_DBI].md_root, i)); + for (; i; i--) + DPRINTF(("IDL %"Z"u", free_pgs[i])); + } +#endif + continue; + } + + mop = env->me_pghead; + mop_len = (mop ? mop[0] : 0) + txn->mt_loose_count; + + /* Reserve records for me_pghead[]. Split it if multi-page, + * to avoid searching freeDB for a page range. Use keys in + * range [1,me_pglast]: Smaller than txnid of oldest reader. + */ + if (total_room >= mop_len) { + if (total_room == mop_len || --more < 0) + break; + } else if (head_room >= maxfree_1pg && head_id > 1) { + /* Keep current record (overflow page), add a new one */ + head_id--; + head_room = 0; + } + /* (Re)write {key = head_id, IDL length = head_room} */ + total_room -= head_room; + head_room = mop_len - total_room; + if (head_room > maxfree_1pg && head_id > 1) { + /* Overflow multi-page for part of me_pghead */ + head_room /= head_id; /* amortize page sizes */ + head_room += maxfree_1pg - head_room % (maxfree_1pg + 1); + } else if (head_room < 0) { + /* Rare case, not bothering to delete this record */ + head_room = 0; + } + key.mv_size = sizeof(head_id); + key.mv_data = &head_id; + data.mv_size = (head_room + 1) * sizeof(pgno_t); + rc = mdb_cursor_put(&mc, &key, &data, MDB_RESERVE); + if (rc) + return rc; + /* IDL is initially empty, zero out at least the length */ + pgs = (pgno_t *)data.mv_data; + j = head_room > clean_limit ? head_room : 0; + do { + pgs[j] = 0; + } while (--j >= 0); + total_room += head_room; + } + + /* Return loose page numbers to me_pghead, though usually none are + * left at this point. The pages themselves remain in dirty_list. + */ + if (txn->mt_loose_pgs) { + MDB_page *mp = txn->mt_loose_pgs; + unsigned count = txn->mt_loose_count; + MDB_IDL loose; + /* Room for loose pages + temp IDL with same */ + if ((rc = mdb_midl_need(&env->me_pghead, 2*count+1)) != 0) + return rc; + mop = env->me_pghead; + loose = mop + MDB_IDL_ALLOCLEN(mop) - count; + for (count = 0; mp; mp = NEXT_LOOSE_PAGE(mp)) + loose[ ++count ] = mp->mp_pgno; + loose[0] = count; + mdb_midl_sort(loose); + mdb_midl_xmerge(mop, loose); + txn->mt_loose_pgs = NULL; + txn->mt_loose_count = 0; + mop_len = mop[0]; + } + + /* Fill in the reserved me_pghead records */ + rc = MDB_SUCCESS; + if (mop_len) { + MDB_val key, data; + + mop += mop_len; + rc = mdb_cursor_first(&mc, &key, &data); + for (; !rc; rc = mdb_cursor_next(&mc, &key, &data, MDB_NEXT)) { + txnid_t id = *(txnid_t *)key.mv_data; + ssize_t len = (ssize_t)(data.mv_size / sizeof(MDB_ID)) - 1; + MDB_ID save; + + mdb_tassert(txn, len >= 0 && id <= env->me_pglast); + key.mv_data = &id; + if (len > mop_len) { + len = mop_len; + data.mv_size = (len + 1) * sizeof(MDB_ID); + } + data.mv_data = mop -= len; + save = mop[0]; + mop[0] = len; + rc = mdb_cursor_put(&mc, &key, &data, MDB_CURRENT); + mop[0] = save; + if (rc || !(mop_len -= len)) + break; + } + } + return rc; +} + +/** Flush (some) dirty pages to the map, after clearing their dirty flag. + * @param[in] txn the transaction that's being committed + * @param[in] keep number of initial pages in dirty_list to keep dirty. + * @return 0 on success, non-zero on failure. + */ +static int +mdb_page_flush(MDB_txn *txn, int keep) +{ + MDB_env *env = txn->mt_env; + MDB_ID2L dl = txn->mt_u.dirty_list; + unsigned psize = env->me_psize, j; + int i, pagecount = dl[0].mid, rc; + size_t size = 0, pos = 0; + pgno_t pgno = 0; + MDB_page *dp = NULL; +#ifdef _WIN32 + OVERLAPPED ov; +#else + struct iovec iov[MDB_COMMIT_PAGES]; + ssize_t wpos = 0, wsize = 0, wres; + size_t next_pos = 1; /* impossible pos, so pos != next_pos */ + int n = 0; +#endif + + j = i = keep; + + if (env->me_flags & MDB_WRITEMAP) { + /* Clear dirty flags */ + while (++i <= pagecount) { + dp = dl[i].mptr; + /* Don't flush this page yet */ + if (dp->mp_flags & (P_LOOSE|P_KEEP)) { + dp->mp_flags &= ~P_KEEP; + dl[++j] = dl[i]; + continue; + } + dp->mp_flags &= ~P_DIRTY; + } + goto done; + } + + /* Write the pages */ + for (;;) { + if (++i <= pagecount) { + dp = dl[i].mptr; + /* Don't flush this page yet */ + if (dp->mp_flags & (P_LOOSE|P_KEEP)) { + dp->mp_flags &= ~P_KEEP; + dl[i].mid = 0; + continue; + } + pgno = dl[i].mid; + /* clear dirty flag */ + dp->mp_flags &= ~P_DIRTY; + pos = pgno * psize; + size = psize; + if (IS_OVERFLOW(dp)) size *= dp->mp_pages; + } +#ifdef _WIN32 + else break; + + /* 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. + */ + DPRINTF(("committing page %"Z"u", pgno)); + memset(&ov, 0, sizeof(ov)); + ov.Offset = pos & 0xffffffff; + ov.OffsetHigh = pos >> 16 >> 16; + if (!WriteFile(env->me_fd, dp, size, NULL, &ov)) { + rc = ErrCode(); + DPRINTF(("WriteFile: %d", rc)); + return rc; + } +#else + /* Write up to MDB_COMMIT_PAGES dirty pages at a time. */ + if (pos!=next_pos || n==MDB_COMMIT_PAGES || wsize+size>MAX_WRITE) { + if (n) { + /* Write previous page(s) */ +#ifdef MDB_USE_PWRITEV + wres = pwritev(env->me_fd, iov, n, wpos); +#else + if (n == 1) { + wres = pwrite(env->me_fd, iov[0].iov_base, wsize, wpos); + } else { + if (lseek(env->me_fd, wpos, SEEK_SET) == -1) { + rc = ErrCode(); + DPRINTF(("lseek: %s", strerror(rc))); + return rc; + } + wres = writev(env->me_fd, iov, n); + } +#endif + if (wres != wsize) { + if (wres < 0) { + rc = ErrCode(); + DPRINTF(("Write error: %s", strerror(rc))); + } else { + rc = EIO; /* TODO: Use which error code? */ + DPUTS("short write, filesystem full?"); + } + return rc; + } + n = 0; + } + if (i > pagecount) + break; + wpos = pos; + wsize = 0; + } + DPRINTF(("committing page %"Z"u", pgno)); + next_pos = pos + size; + iov[n].iov_len = size; + iov[n].iov_base = (char *)dp; + wsize += size; + n++; +#endif /* _WIN32 */ + } + + /* MIPS has cache coherency issues, this is a no-op everywhere else + * Note: for any size >= on-chip cache size, entire on-chip cache is + * flushed. + */ + CACHEFLUSH(env->me_map, txn->mt_next_pgno * env->me_psize, DCACHE); + + for (i = keep; ++i <= pagecount; ) { + dp = dl[i].mptr; + /* This is a page we skipped above */ + if (!dl[i].mid) { + dl[++j] = dl[i]; + dl[j].mid = dp->mp_pgno; + continue; + } + mdb_dpage_free(env, dp); + } + +done: + i--; + txn->mt_dirty_room += i - j; + dl[0].mid = j; + return MDB_SUCCESS; +} + +int +mdb_txn_commit(MDB_txn *txn) +{ + int rc; + unsigned int i; + MDB_env *env; + + if (txn == NULL || txn->mt_env == NULL) + return EINVAL; + + if (txn->mt_child) { + rc = mdb_txn_commit(txn->mt_child); + txn->mt_child = NULL; + if (rc) + goto fail; + } + + env = txn->mt_env; + + if (F_ISSET(txn->mt_flags, MDB_TXN_RDONLY)) { + mdb_dbis_update(txn, 1); + txn->mt_numdbs = 2; /* so txn_abort() doesn't close any new handles */ + 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; + rc = MDB_BAD_TXN; + goto fail; + } + + if (txn->mt_parent) { + MDB_txn *parent = txn->mt_parent; + MDB_page **lp; + MDB_ID2L dst, src; + MDB_IDL pspill; + unsigned x, y, len, ps_len; + + /* Append our free list to parent's */ + rc = mdb_midl_append_list(&parent->mt_free_pgs, txn->mt_free_pgs); + if (rc) + goto fail; + mdb_midl_free(txn->mt_free_pgs); + /* Failures after this must either undo the changes + * to the parent or set MDB_TXN_ERROR in the parent. + */ + + parent->mt_next_pgno = txn->mt_next_pgno; + parent->mt_flags = txn->mt_flags; + + /* Merge our cursors into parent's and close them */ + mdb_cursors_close(txn, 1); + + /* Update parent's DB table. */ + memcpy(parent->mt_dbs, txn->mt_dbs, txn->mt_numdbs * sizeof(MDB_db)); + parent->mt_numdbs = txn->mt_numdbs; + parent->mt_dbflags[0] = txn->mt_dbflags[0]; + parent->mt_dbflags[1] = txn->mt_dbflags[1]; + for (i=2; i<txn->mt_numdbs; i++) { + /* preserve parent's DB_NEW status */ + x = parent->mt_dbflags[i] & DB_NEW; + parent->mt_dbflags[i] = txn->mt_dbflags[i] | x; + } + + dst = parent->mt_u.dirty_list; + src = txn->mt_u.dirty_list; + /* Remove anything in our dirty list from parent's spill list */ + if ((pspill = parent->mt_spill_pgs) && (ps_len = pspill[0])) { + x = y = ps_len; + pspill[0] = (pgno_t)-1; + /* Mark our dirty pages as deleted in parent spill list */ + for (i=0, len=src[0].mid; ++i <= len; ) { + MDB_ID pn = src[i].mid << 1; + while (pn > pspill[x]) + x--; + if (pn == pspill[x]) { + pspill[x] = 1; + y = --x; + } + } + /* Squash deleted pagenums if we deleted any */ + for (x=y; ++x <= ps_len; ) + if (!(pspill[x] & 1)) + pspill[++y] = pspill[x]; + pspill[0] = y; + } + + /* Find len = length of merging our dirty list with parent's */ + x = dst[0].mid; + dst[0].mid = 0; /* simplify loops */ + if (parent->mt_parent) { + len = x + src[0].mid; + y = mdb_mid2l_search(src, dst[x].mid + 1) - 1; + for (i = x; y && i; y--) { + pgno_t yp = src[y].mid; + while (yp < dst[i].mid) + i--; + if (yp == dst[i].mid) { + i--; + len--; + } + } + } else { /* Simplify the above for single-ancestor case */ + len = MDB_IDL_UM_MAX - txn->mt_dirty_room; + } + /* Merge our dirty list with parent's */ + y = src[0].mid; + for (i = len; y; dst[i--] = src[y--]) { + pgno_t yp = src[y].mid; + while (yp < dst[x].mid) + dst[i--] = dst[x--]; + if (yp == dst[x].mid) + free(dst[x--].mptr); + } + mdb_tassert(txn, i == x); + dst[0].mid = len; + free(txn->mt_u.dirty_list); + parent->mt_dirty_room = txn->mt_dirty_room; + if (txn->mt_spill_pgs) { + if (parent->mt_spill_pgs) { + /* TODO: Prevent failure here, so parent does not fail */ + rc = mdb_midl_append_list(&parent->mt_spill_pgs, txn->mt_spill_pgs); + if (rc) + parent->mt_flags |= MDB_TXN_ERROR; + mdb_midl_free(txn->mt_spill_pgs); + mdb_midl_sort(parent->mt_spill_pgs); + } else { + parent->mt_spill_pgs = txn->mt_spill_pgs; + } + } + + /* Append our loose page list to parent's */ + for (lp = &parent->mt_loose_pgs; *lp; lp = &NEXT_LOOSE_PAGE(lp)) + ; + *lp = txn->mt_loose_pgs; + parent->mt_loose_count += txn->mt_loose_count; + + parent->mt_child = NULL; + mdb_midl_free(((MDB_ntxn *)txn)->mnt_pgstate.mf_pghead); + free(txn); + return rc; + } + + if (txn != env->me_txn) { + DPUTS("attempt to commit unknown transaction"); + rc = EINVAL; + goto fail; + } + + mdb_cursors_close(txn, 0); + + if (!txn->mt_u.dirty_list[0].mid && + !(txn->mt_flags & (MDB_TXN_DIRTY|MDB_TXN_SPILLS))) + goto done; + + DPRINTF(("committing txn %"Z"u %p on mdbenv %p, root page %"Z"u", + txn->mt_txnid, (void*)txn, (void*)env, txn->mt_dbs[MAIN_DBI].md_root)); + + /* Update DB root pointers */ + if (txn->mt_numdbs > 2) { + MDB_cursor mc; + 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) { + if (TXN_DBI_CHANGED(txn, i)) { + rc = MDB_BAD_DBI; + goto fail; + } + data.mv_data = &txn->mt_dbs[i]; + rc = mdb_cursor_put(&mc, &txn->mt_dbxs[i].md_name, &data, 0); + if (rc) + goto fail; + } + } + } + + rc = mdb_freelist_save(txn); + if (rc) + goto fail; + + mdb_midl_free(env->me_pghead); + env->me_pghead = NULL; + if (mdb_midl_shrink(&txn->mt_free_pgs)) + env->me_free_pgs = txn->mt_free_pgs; + +#if (MDB_DEBUG) > 2 + mdb_audit(txn); +#endif + + if ((rc = mdb_page_flush(txn, 0)) || + (rc = mdb_env_sync(env, 0)) || + (rc = mdb_env_write_meta(txn))) + goto fail; + + /* Free P_LOOSE pages left behind in dirty_list */ + if (!(env->me_flags & MDB_WRITEMAP)) + mdb_dlist_free(txn); + +done: + env->me_pglast = 0; + env->me_txn = NULL; + mdb_dbis_update(txn, 1); + + if (env->me_txns) + UNLOCK_MUTEX(MDB_MUTEX(env, w)); + if (txn != env->me_txn0) + free(txn); + + return MDB_SUCCESS; + +fail: + mdb_txn_abort(txn); + return rc; +} + +/** 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 ESECT +mdb_env_read_header(MDB_env *env, MDB_meta *meta) +{ + MDB_metabuf pbuf; + MDB_page *p; + MDB_meta *m; + int i, rc, off; + enum { Size = sizeof(pbuf) }; + + /* We don't know the page size yet, so use a minimum value. + * Read both meta pages so we can use the latest one. + */ + + for (i=off=0; i<2; i++, off = meta->mm_psize) { +#ifdef _WIN32 + DWORD len; + OVERLAPPED ov; + memset(&ov, 0, sizeof(ov)); + ov.Offset = off; + rc = ReadFile(env->me_fd, &pbuf, Size, &len, &ov) ? (int)len : -1; + if (rc == -1 && ErrCode() == ERROR_HANDLE_EOF) + rc = 0; +#else + rc = pread(env->me_fd, &pbuf, Size, off); +#endif + if (rc != Size) { + if (rc == 0 && off == 0) + return ENOENT; + rc = rc < 0 ? (int) ErrCode() : MDB_INVALID; + DPRINTF(("read: %s", mdb_strerror(rc))); + return rc; + } + + p = (MDB_page *)&pbuf; + + if (!F_ISSET(p->mp_flags, P_META)) { + DPRINTF(("page %"Z"u not a meta page", p->mp_pgno)); + return MDB_INVALID; + } + + m = METADATA(p); + if (m->mm_magic != MDB_MAGIC) { + DPUTS("meta has invalid magic"); + return MDB_INVALID; + } + + if (m->mm_version != MDB_DATA_VERSION) { + DPRINTF(("database is version %u, expected version %u", + m->mm_version, MDB_DATA_VERSION)); + return MDB_VERSION_MISMATCH; + } + + if (off == 0 || m->mm_txnid > meta->mm_txnid) + *meta = *m; + } + return 0; +} + +/** Fill in most of the zeroed #MDB_meta for an empty database environment */ +static void ESECT +mdb_env_init_meta0(MDB_env *env, MDB_meta *meta) +{ + meta->mm_magic = MDB_MAGIC; + meta->mm_version = MDB_DATA_VERSION; + meta->mm_mapsize = env->me_mapsize; + meta->mm_psize = env->me_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; +} + +/** Write the environment parameters of a freshly created DB environment. + * @param[in] env the environment handle + * @param[in] meta the #MDB_meta to write + * @return 0 on success, non-zero on failure. + */ +static int ESECT +mdb_env_init_meta(MDB_env *env, MDB_meta *meta) +{ + MDB_page *p, *q; + int rc; + unsigned int psize; +#ifdef _WIN32 + DWORD len; + OVERLAPPED ov; + memset(&ov, 0, sizeof(ov)); +#define DO_PWRITE(rc, fd, ptr, size, len, pos) do { \ + ov.Offset = pos; \ + rc = WriteFile(fd, ptr, size, &len, &ov); } while(0) +#else + int len; +#define DO_PWRITE(rc, fd, ptr, size, len, pos) do { \ + len = pwrite(fd, ptr, size, pos); \ + rc = (len >= 0); } while(0) +#endif + + DPUTS("writing new meta page"); + + psize = env->me_psize; + + p = calloc(2, psize); + p->mp_pgno = 0; + p->mp_flags = P_META; + *(MDB_meta *)METADATA(p) = *meta; + + q = (MDB_page *)((char *)p + psize); + q->mp_pgno = 1; + q->mp_flags = P_META; + *(MDB_meta *)METADATA(q) = *meta; + + DO_PWRITE(rc, env->me_fd, p, psize * 2, len, 0); + if (!rc) + rc = ErrCode(); + else if ((unsigned) len == psize * 2) + rc = MDB_SUCCESS; + else + rc = ENOSPC; + 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, *mp; + size_t mapsize; + off_t off; + int rc, len, toggle; + char *ptr; + HANDLE mfd; +#ifdef _WIN32 + OVERLAPPED ov; +#else + int r2; +#endif + + toggle = txn->mt_txnid & 1; + DPRINTF(("writing meta page %d for root page %"Z"u", + toggle, txn->mt_dbs[MAIN_DBI].md_root)); + + env = txn->mt_env; + mp = env->me_metas[toggle]; + mapsize = env->me_metas[toggle ^ 1]->mm_mapsize; + /* Persist any increases of mapsize config */ + if (mapsize < env->me_mapsize) + mapsize = env->me_mapsize; + + if (env->me_flags & MDB_WRITEMAP) { + mp->mm_mapsize = mapsize; + mp->mm_dbs[0] = txn->mt_dbs[0]; + mp->mm_dbs[1] = txn->mt_dbs[1]; + mp->mm_last_pg = txn->mt_next_pgno - 1; +#if !(defined(_MSC_VER) || defined(__i386__) || defined(__x86_64__)) + /* LY: issue a memory barrier, if not x86. ITS#7969 */ + __sync_synchronize(); +#endif + mp->mm_txnid = txn->mt_txnid; + if (!(env->me_flags & (MDB_NOMETASYNC|MDB_NOSYNC))) { + unsigned meta_size = env->me_psize; + rc = (env->me_flags & MDB_MAPASYNC) ? MS_ASYNC : MS_SYNC; + ptr = env->me_map; + if (toggle) { +#ifndef _WIN32 /* POSIX msync() requires ptr = start of OS page */ + if (meta_size < env->me_os_psize) + meta_size += meta_size; + else +#endif + ptr += meta_size; + } + if (MDB_MSYNC(ptr, meta_size, rc)) { + rc = ErrCode(); + goto fail; + } + } + goto done; + } + metab.mm_txnid = env->me_metas[toggle]->mm_txnid; + metab.mm_last_pg = env->me_metas[toggle]->mm_last_pg; + + meta.mm_mapsize = mapsize; + 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; + + off = offsetof(MDB_meta, mm_mapsize); + ptr = (char *)&meta + off; + len = sizeof(MDB_meta) - off; + if (toggle) + off += env->me_psize; + off += PAGEHDRSZ; + + /* Write to the SYNC fd */ + mfd = env->me_flags & (MDB_NOSYNC|MDB_NOMETASYNC) ? + env->me_fd : env->me_mfd; +#ifdef _WIN32 + { + memset(&ov, 0, sizeof(ov)); + ov.Offset = off; + if (!WriteFile(mfd, ptr, len, (DWORD *)&rc, &ov)) + rc = -1; + } +#else + rc = pwrite(mfd, ptr, len, off); +#endif + if (rc != len) { + rc = rc < 0 ? ErrCode() : EIO; + 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 + memset(&ov, 0, sizeof(ov)); + ov.Offset = off; + WriteFile(env->me_fd, ptr, len, NULL, &ov); +#else + r2 = pwrite(env->me_fd, ptr, len, off); + (void)r2; /* Silence warnings. We don't care about pwrite's return value */ +#endif +fail: + env->me_flags |= MDB_FATAL_ERROR; + return rc; + } + /* MIPS has cache coherency issues, this is a no-op everywhere else */ + CACHEFLUSH(env->me_map + off, len, DCACHE); +done: + /* 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. + */ + if (env->me_txns) + env->me_txns->mti_txnid = txn->mt_txnid; + + return MDB_SUCCESS; +} + +/** Check both meta pages to see which one is newer. + * @param[in] env the environment handle + * @return meta toggle (0 or 1). + */ +static int +mdb_env_pick_meta(const MDB_env *env) +{ + return (env->me_metas[0]->mm_txnid < env->me_metas[1]->mm_txnid); +} + +int ESECT +mdb_env_create(MDB_env **env) +{ + MDB_env *e; + + e = calloc(1, sizeof(MDB_env)); + if (!e) + return ENOMEM; + + e->me_maxreaders = DEFAULT_READERS; + e->me_maxdbs = e->me_numdbs = 2; + e->me_fd = INVALID_HANDLE_VALUE; + e->me_lfd = INVALID_HANDLE_VALUE; + e->me_mfd = INVALID_HANDLE_VALUE; +#ifdef MDB_USE_SYSV_SEM + e->me_rmutex.semid = -1; + e->me_wmutex.semid = -1; +#endif + e->me_pid = getpid(); + GET_PAGESIZE(e->me_os_psize); + VGMEMP_CREATE(e,0,0); + *env = e; + return MDB_SUCCESS; +} + +static int ESECT +mdb_env_map(MDB_env *env, void *addr) +{ + MDB_page *p; + unsigned int flags = env->me_flags; +#ifdef _WIN32 + int rc; + HANDLE mh; + LONG sizelo, sizehi; + size_t msize; + + if (flags & MDB_RDONLY) { + /* Don't set explicit map size, use whatever exists */ + msize = 0; + sizelo = 0; + sizehi = 0; + } else { + msize = env->me_mapsize; + sizelo = msize & 0xffffffff; + sizehi = msize >> 16 >> 16; /* only needed on Win64 */ + + /* Windows won't create mappings for zero length files. + * and won't map more than the file size. + * Just set the maxsize right now. + */ + if (SetFilePointer(env->me_fd, sizelo, &sizehi, 0) != (DWORD)sizelo + || !SetEndOfFile(env->me_fd) + || SetFilePointer(env->me_fd, 0, NULL, 0) != 0) + return ErrCode(); + } + + mh = CreateFileMapping(env->me_fd, NULL, flags & MDB_WRITEMAP ? + PAGE_READWRITE : PAGE_READONLY, + sizehi, sizelo, NULL); + if (!mh) + return ErrCode(); + env->me_map = MapViewOfFileEx(mh, flags & MDB_WRITEMAP ? + FILE_MAP_WRITE : FILE_MAP_READ, + 0, 0, msize, addr); + rc = env->me_map ? 0 : ErrCode(); + CloseHandle(mh); + if (rc) + return rc; +#else + int prot = PROT_READ; + if (flags & MDB_WRITEMAP) { + prot |= PROT_WRITE; + if (ftruncate(env->me_fd, env->me_mapsize) < 0) + return ErrCode(); + } + env->me_map = mmap(addr, env->me_mapsize, prot, MAP_SHARED, + env->me_fd, 0); + if (env->me_map == MAP_FAILED) { + env->me_map = NULL; + return ErrCode(); + } + + if (flags & MDB_NORDAHEAD) { + /* Turn off readahead. It's harmful when the DB is larger than RAM. */ +#ifdef MADV_RANDOM + madvise(env->me_map, env->me_mapsize, MADV_RANDOM); +#else +#ifdef POSIX_MADV_RANDOM + posix_madvise(env->me_map, env->me_mapsize, POSIX_MADV_RANDOM); +#endif /* POSIX_MADV_RANDOM */ +#endif /* MADV_RANDOM */ + } +#endif /* _WIN32 */ + + /* Can happen because the address argument to mmap() is just a + * hint. mmap() can pick another, e.g. if the range is in use. + * The MAP_FIXED flag would prevent that, but then mmap could + * instead unmap existing pages to make room for the new map. + */ + if (addr && env->me_map != addr) + return EBUSY; /* TODO: Make a new MDB_* error code? */ + + p = (MDB_page *)env->me_map; + env->me_metas[0] = METADATA(p); + env->me_metas[1] = (MDB_meta *)((char *)env->me_metas[0] + env->me_psize); + + return MDB_SUCCESS; +} + +int ESECT +mdb_env_set_mapsize(MDB_env *env, size_t size) +{ + /* If env is already open, caller is responsible for making + * sure there are no active txns. + */ + if (env->me_map) { + int rc; + MDB_meta *meta; + void *old; + if (env->me_txn) + return EINVAL; + meta = env->me_metas[mdb_env_pick_meta(env)]; + if (!size) + size = meta->mm_mapsize; + { + /* Silently round up to minimum if the size is too small */ + size_t minsize = (meta->mm_last_pg + 1) * env->me_psize; + if (size < minsize) + size = minsize; + } + munmap(env->me_map, env->me_mapsize); + env->me_mapsize = size; + old = (env->me_flags & MDB_FIXEDMAP) ? env->me_map : NULL; + rc = mdb_env_map(env, old); + if (rc) + return rc; + } + env->me_mapsize = size; + if (env->me_psize) + env->me_maxpg = env->me_mapsize / env->me_psize; + return MDB_SUCCESS; +} + +int ESECT +mdb_env_set_maxdbs(MDB_env *env, MDB_dbi dbs) +{ + if (env->me_map) + return EINVAL; + env->me_maxdbs = dbs + 2; /* Named databases + main and free DB */ + return MDB_SUCCESS; +} + +int ESECT +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 ESECT +mdb_env_get_maxreaders(MDB_env *env, unsigned int *readers) +{ + if (!env || !readers) + return EINVAL; + *readers = env->me_maxreaders; + return MDB_SUCCESS; +} + +static int ESECT +mdb_fsize(HANDLE fd, size_t *size) +{ +#ifdef _WIN32 + LARGE_INTEGER fsize; + + if (!GetFileSizeEx(fd, &fsize)) + return ErrCode(); + + *size = fsize.QuadPart; +#else + struct stat st; + + if (fstat(fd, &st)) + return ErrCode(); + + *size = st.st_size; +#endif + return MDB_SUCCESS; +} + +#ifdef BROKEN_FDATASYNC +#include <sys/utsname.h> +#include <sys/vfs.h> +#endif + +/** Further setup required for opening an LMDB environment + */ +static int ESECT +mdb_env_open2(MDB_env *env) +{ + unsigned int flags = env->me_flags; + int i, newenv = 0, rc; + MDB_meta meta; + +#ifdef _WIN32 + /* See if we should use QueryLimited */ + rc = GetVersion(); + if ((rc & 0xff) > 5) + env->me_pidquery = MDB_PROCESS_QUERY_LIMITED_INFORMATION; + else + env->me_pidquery = PROCESS_QUERY_INFORMATION; +#endif /* _WIN32 */ + +#ifdef BROKEN_FDATASYNC + /* ext3/ext4 fdatasync is broken on some older Linux kernels. + * https://lkml.org/lkml/2012/9/3/83 + * Kernels after 3.6-rc6 are known good. + * https://lkml.org/lkml/2012/9/10/556 + * See if the DB is on ext3/ext4, then check for new enough kernel + * Kernels 2.6.32.60, 2.6.34.15, 3.2.30, and 3.5.4 are also known + * to be patched. + */ + { + struct statfs st; + fstatfs(env->me_fd, &st); + while (st.f_type == 0xEF53) { + struct utsname uts; + int i; + uname(&uts); + if (uts.release[0] < '3') { + if (!strncmp(uts.release, "2.6.32.", 7)) { + i = atoi(uts.release+7); + if (i >= 60) + break; /* 2.6.32.60 and newer is OK */ + } else if (!strncmp(uts.release, "2.6.34.", 7)) { + i = atoi(uts.release+7); + if (i >= 15) + break; /* 2.6.34.15 and newer is OK */ + } + } else if (uts.release[0] == '3') { + i = atoi(uts.release+2); + if (i > 5) + break; /* 3.6 and newer is OK */ + if (i == 5) { + i = atoi(uts.release+4); + if (i >= 4) + break; /* 3.5.4 and newer is OK */ + } else if (i == 2) { + i = atoi(uts.release+4); + if (i >= 30) + break; /* 3.2.30 and newer is OK */ + } + } else { /* 4.x and newer is OK */ + break; + } + env->me_flags |= MDB_FSYNCONLY; + break; + } + } +#endif + + if ((i = mdb_env_read_header(env, &meta)) != 0) { + if (i != ENOENT) + return i; + DPUTS("new mdbenv"); + newenv = 1; + env->me_psize = env->me_os_psize; + if (env->me_psize > MAX_PAGESIZE) + env->me_psize = MAX_PAGESIZE; + memset(&meta, 0, sizeof(meta)); + mdb_env_init_meta0(env, &meta); + meta.mm_mapsize = DEFAULT_MAPSIZE; + } else { + env->me_psize = meta.mm_psize; + } + + /* Was a mapsize configured? */ + if (!env->me_mapsize) { + env->me_mapsize = meta.mm_mapsize; + } + { + /* Make sure mapsize >= committed data size. Even when using + * mm_mapsize, which could be broken in old files (ITS#7789). + */ + size_t minsize = (meta.mm_last_pg + 1) * meta.mm_psize; + if (env->me_mapsize < minsize) + env->me_mapsize = minsize; + } + meta.mm_mapsize = env->me_mapsize; + + if (newenv && !(flags & MDB_FIXEDMAP)) { + /* mdb_env_map() may grow the datafile. Write the metapages + * first, so the file will be valid if initialization fails. + * Except with FIXEDMAP, since we do not yet know mm_address. + * We could fill in mm_address later, but then a different + * program might end up doing that - one with a memory layout + * and map address which does not suit the main program. + */ + rc = mdb_env_init_meta(env, &meta); + if (rc) + return rc; + newenv = 0; + } + + rc = mdb_env_map(env, (flags & MDB_FIXEDMAP) ? meta.mm_address : NULL); + if (rc) + return rc; + + if (newenv) { + if (flags & MDB_FIXEDMAP) + meta.mm_address = env->me_map; + i = mdb_env_init_meta(env, &meta); + if (i != MDB_SUCCESS) { + return i; + } + } + + env->me_maxfree_1pg = (env->me_psize - PAGEHDRSZ) / sizeof(pgno_t) - 1; + env->me_nodemax = (((env->me_psize - PAGEHDRSZ) / MDB_MINKEYS) & -2) + - sizeof(indx_t); +#if !(MDB_MAXKEYSIZE) + env->me_maxkey = env->me_nodemax - (NODESIZE + sizeof(MDB_db)); +#endif + env->me_maxpg = env->me_mapsize / env->me_psize; + +#if MDB_DEBUG + { + int toggle = mdb_env_pick_meta(env); + MDB_db *db = &env->me_metas[toggle]->mm_dbs[MAIN_DBI]; + + DPRINTF(("opened database version %u, pagesize %u", + env->me_metas[0]->mm_version, env->me_psize)); + DPRINTF(("using meta page %d", toggle)); + DPRINTF(("depth: %u", db->md_depth)); + DPRINTF(("entries: %"Z"u", db->md_entries)); + DPRINTF(("branch pages: %"Z"u", db->md_branch_pages)); + DPRINTF(("leaf pages: %"Z"u", db->md_leaf_pages)); + DPRINTF(("overflow pages: %"Z"u", db->md_overflow_pages)); + DPRINTF(("root: %"Z"u", db->md_root)); + } +#endif + + return MDB_SUCCESS; +} + + +/** Release a reader thread's slot in the reader lock table. + * This function is called automatically when a thread exits. + * @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_pid = 0; +} + +#ifdef _WIN32 +/** Junk for arranging thread-specific callbacks on Windows. This is + * necessarily platform and compiler-specific. Windows supports up + * to 1088 keys. Let's assume nobody opens more than 64 environments + * in a single process, for now. They can override this if needed. + */ +#ifndef MAX_TLS_KEYS +#define MAX_TLS_KEYS 64 +#endif +static pthread_key_t mdb_tls_keys[MAX_TLS_KEYS]; +static int mdb_tls_nkeys; + +static void NTAPI mdb_tls_callback(PVOID module, DWORD reason, PVOID ptr) +{ + int i; + switch(reason) { + case DLL_PROCESS_ATTACH: break; + case DLL_THREAD_ATTACH: break; + case DLL_THREAD_DETACH: + for (i=0; i<mdb_tls_nkeys; i++) { + MDB_reader *r = pthread_getspecific(mdb_tls_keys[i]); + if (r) { + mdb_env_reader_dest(r); + } + } + break; + case DLL_PROCESS_DETACH: break; + } +} +#ifdef __GNUC__ +#ifdef _WIN64 +const PIMAGE_TLS_CALLBACK mdb_tls_cbp __attribute__((section (".CRT$XLB"))) = mdb_tls_callback; +#else +PIMAGE_TLS_CALLBACK mdb_tls_cbp __attribute__((section (".CRT$XLB"))) = mdb_tls_callback; +#endif +#else +#ifdef _WIN64 +/* Force some symbol references. + * _tls_used forces the linker to create the TLS directory if not already done + * mdb_tls_cbp prevents whole-program-optimizer from dropping the symbol. + */ +#pragma comment(linker, "/INCLUDE:_tls_used") +#pragma comment(linker, "/INCLUDE:mdb_tls_cbp") +#pragma const_seg(".CRT$XLB") +extern const PIMAGE_TLS_CALLBACK mdb_tls_cbp; +const PIMAGE_TLS_CALLBACK mdb_tls_cbp = mdb_tls_callback; +#pragma const_seg() +#else /* _WIN32 */ +#pragma comment(linker, "/INCLUDE:__tls_used") +#pragma comment(linker, "/INCLUDE:_mdb_tls_cbp") +#pragma data_seg(".CRT$XLB") +PIMAGE_TLS_CALLBACK mdb_tls_cbp = mdb_tls_callback; +#pragma data_seg() +#endif /* WIN 32/64 */ +#endif /* !__GNUC__ */ +#endif + +/** Downgrade the exclusive lock on the region back to shared */ +static int ESECT +mdb_env_share_locks(MDB_env *env, int *excl) +{ + int rc = 0, toggle = mdb_env_pick_meta(env); + + 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)); + if (!LockFileEx(env->me_lfd, 0, 0, 1, 0, &ov)) { + rc = ErrCode(); + } else { + UnlockFile(env->me_lfd, 0, 0, 1, 0); + *excl = 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; + while ((rc = fcntl(env->me_lfd, F_SETLK, &lock_info)) && + (rc = ErrCode()) == EINTR) ; + *excl = rc ? -1 : 0; /* error may mean we lost the lock */ + } +#endif + + return rc; +} + +/** Try to get exclusive lock, otherwise shared. + * Maintain *excl = -1: no/unknown lock, 0: shared, 1: exclusive. + */ +static int ESECT +mdb_env_excl_lock(MDB_env *env, int *excl) +{ + int rc = 0; +#ifdef _WIN32 + 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)) { + *excl = 0; + } else { + rc = ErrCode(); + } + } +#else + 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; + while ((rc = fcntl(env->me_lfd, F_SETLK, &lock_info)) && + (rc = ErrCode()) == EINTR) ; + if (!rc) { + *excl = 1; + } else +# ifdef MDB_USE_SYSV_SEM + if (*excl < 0) /* always true when !MDB_USE_SYSV_SEM */ +# endif + { + lock_info.l_type = F_RDLCK; + while ((rc = fcntl(env->me_lfd, F_SETLKW, &lock_info)) && + (rc = ErrCode()) == EINTR) ; + if (rc == 0) + *excl = 0; + } +#endif + return rc; +} + +#ifdef MDB_USE_HASH +/* + * 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] val value 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_val(MDB_val *val, mdb_hash_t hval) +{ + unsigned char *s = (unsigned char *)val->mv_data; /* unsigned string */ + unsigned char *end = s + val->mv_size; + /* + * FNV-1a hash each octet of the string + */ + while (s < end) { + /* 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 encoded hash. + * This uses modified RFC1924 Ascii85 encoding to accommodate systems with + * very short name limits. We don't care about the encoding being reversible, + * we just want to preserve as many bits of the input as possible in a + * small printable string. + * @param[in] str string to hash + * @param[out] encbuf an array of 11 chars to hold the hash + */ +static const char mdb_a85[]= "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz!#$%&()*+-;<=>?@^_`{|}~"; + +static void +mdb_pack85(unsigned long l, char *out) +{ + int i; + + for (i=0; i<5; i++) { + *out++ = mdb_a85[l % 85]; + l /= 85; + } +} + +static void +mdb_hash_enc(MDB_val *val, char *encbuf) +{ + mdb_hash_t h = mdb_hash_val(val, MDB_HASH_INIT); + + mdb_pack85(h, encbuf); + mdb_pack85(h>>32, encbuf+5); + encbuf[10] = '\0'; +} +#endif + +/** Open and/or initialize the lock region for the environment. + * @param[in] env The LMDB 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[in,out] excl In -1, out lock type: -1 none, 0 shared, 1 exclusive + * @return 0 on success, non-zero on failure. + */ +static int ESECT +mdb_env_setup_locks(MDB_env *env, char *lpath, int mode, int *excl) +{ +#ifdef _WIN32 +# define MDB_ERRCODE_ROFS ERROR_WRITE_PROTECT +#else +# define MDB_ERRCODE_ROFS EROFS +#ifdef O_CLOEXEC /* Linux: Open file and set FD_CLOEXEC atomically */ +# define MDB_CLOEXEC O_CLOEXEC +#else + int fdflags; +# define MDB_CLOEXEC 0 +#endif +#endif +#ifdef MDB_USE_SYSV_SEM + int semid; + union semun semu; +#endif + int rc; + off_t size, rsize; + +#ifdef _WIN32 + env->me_lfd = CreateFile(lpath, GENERIC_READ|GENERIC_WRITE, + FILE_SHARE_READ|FILE_SHARE_WRITE, NULL, OPEN_ALWAYS, + FILE_ATTRIBUTE_NORMAL, NULL); +#else + env->me_lfd = open(lpath, O_RDWR|O_CREAT|MDB_CLOEXEC, mode); +#endif + if (env->me_lfd == INVALID_HANDLE_VALUE) { + rc = ErrCode(); + if (rc == MDB_ERRCODE_ROFS && (env->me_flags & MDB_RDONLY)) { + return MDB_SUCCESS; + } + goto fail_errno; + } +#if ! ((MDB_CLOEXEC) || defined(_WIN32)) + /* Lose record locks when exec*() */ + if ((fdflags = fcntl(env->me_lfd, F_GETFD) | FD_CLOEXEC) >= 0) + fcntl(env->me_lfd, F_SETFD, fdflags); +#endif + + if (!(env->me_flags & MDB_NOTLS)) { + rc = pthread_key_create(&env->me_txkey, mdb_env_reader_dest); + if (rc) + goto fail; + env->me_flags |= MDB_ENV_TXKEY; +#ifdef _WIN32 + /* Windows TLS callbacks need help finding their TLS info. */ + if (mdb_tls_nkeys >= MAX_TLS_KEYS) { + rc = MDB_TLS_FULL; + goto fail; + } + mdb_tls_keys[mdb_tls_nkeys++] = env->me_txkey; +#endif + } + + /* Try to get exclusive lock. If we succeed, then + * nobody is using the lock region and we should initialize it. + */ + if ((rc = mdb_env_excl_lock(env, excl))) goto fail; + +#ifdef _WIN32 + size = GetFileSize(env->me_lfd, NULL); +#else + size = lseek(env->me_lfd, 0, SEEK_END); + if (size == -1) goto fail_errno; +#endif + rsize = (env->me_maxreaders-1) * sizeof(MDB_reader) + sizeof(MDB_txninfo); + if (size < rsize && *excl > 0) { +#ifdef _WIN32 + if (SetFilePointer(env->me_lfd, rsize, NULL, FILE_BEGIN) != (DWORD)rsize + || !SetEndOfFile(env->me_lfd)) + goto fail_errno; +#else + if (ftruncate(env->me_lfd, rsize) != 0) goto fail_errno; +#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) goto fail_errno; + env->me_txns = MapViewOfFileEx(mh, FILE_MAP_WRITE, 0, 0, rsize, NULL); + CloseHandle(mh); + if (!env->me_txns) goto fail_errno; +#else + void *m = mmap(NULL, rsize, PROT_READ|PROT_WRITE, MAP_SHARED, + env->me_lfd, 0); + if (m == MAP_FAILED) goto fail_errno; + env->me_txns = m; +#endif + } + if (*excl > 0) { +#ifdef _WIN32 + BY_HANDLE_FILE_INFORMATION stbuf; + struct { + DWORD volume; + DWORD nhigh; + DWORD nlow; + } idbuf; + MDB_val val; + char encbuf[11]; + + 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; + } + if (!GetFileInformationByHandle(env->me_lfd, &stbuf)) goto fail_errno; + idbuf.volume = stbuf.dwVolumeSerialNumber; + idbuf.nhigh = stbuf.nFileIndexHigh; + idbuf.nlow = stbuf.nFileIndexLow; + val.mv_data = &idbuf; + val.mv_size = sizeof(idbuf); + mdb_hash_enc(&val, encbuf); + sprintf(env->me_txns->mti_rmname, "Global\\MDBr%s", encbuf); + sprintf(env->me_txns->mti_wmname, "Global\\MDBw%s", encbuf); + env->me_rmutex = CreateMutex(&mdb_all_sa, FALSE, env->me_txns->mti_rmname); + if (!env->me_rmutex) goto fail_errno; + env->me_wmutex = CreateMutex(&mdb_all_sa, FALSE, env->me_txns->mti_wmname); + if (!env->me_wmutex) goto fail_errno; +#elif defined(MDB_USE_SYSV_SEM) + unsigned short vals[2] = {1, 1}; + semid = semget(IPC_PRIVATE, 2, mode); + if (semid < 0) + goto fail_errno; + semu.array = vals; + if (semctl(semid, 0, SETALL, semu) < 0) + goto fail_errno; + env->me_txns->mti_semid = semid; +#else /* MDB_USE_SYSV_SEM */ + pthread_mutexattr_t mattr; + + if ((rc = pthread_mutexattr_init(&mattr)) + || (rc = pthread_mutexattr_setpshared(&mattr, PTHREAD_PROCESS_SHARED)) +#ifdef MDB_ROBUST_SUPPORTED + || (rc = pthread_mutexattr_setrobust(&mattr, PTHREAD_MUTEX_ROBUST)) +#endif + || (rc = pthread_mutex_init(&env->me_txns->mti_rmutex, &mattr)) + || (rc = pthread_mutex_init(&env->me_txns->mti_wmutex, &mattr))) + goto fail; + pthread_mutexattr_destroy(&mattr); +#endif /* _WIN32 || MDB_USE_SYSV_SEM */ + + env->me_txns->mti_magic = MDB_MAGIC; + env->me_txns->mti_format = MDB_LOCK_FORMAT; + env->me_txns->mti_txnid = 0; + env->me_txns->mti_numreaders = 0; + + } else { +#ifdef MDB_USE_SYSV_SEM + struct semid_ds buf; +#endif + if (env->me_txns->mti_magic != MDB_MAGIC) { + DPUTS("lock region has invalid magic"); + rc = MDB_INVALID; + goto fail; + } + if (env->me_txns->mti_format != MDB_LOCK_FORMAT) { + DPRINTF(("lock region has format+version 0x%x, expected 0x%x", + env->me_txns->mti_format, MDB_LOCK_FORMAT)); + rc = MDB_VERSION_MISMATCH; + goto fail; + } + rc = ErrCode(); + if (rc && rc != EACCES && rc != EAGAIN) { + goto fail; + } +#ifdef _WIN32 + env->me_rmutex = OpenMutex(SYNCHRONIZE, FALSE, env->me_txns->mti_rmname); + if (!env->me_rmutex) goto fail_errno; + env->me_wmutex = OpenMutex(SYNCHRONIZE, FALSE, env->me_txns->mti_wmname); + if (!env->me_wmutex) goto fail_errno; +#elif defined(MDB_USE_SYSV_SEM) + semid = env->me_txns->mti_semid; + semu.buf = &buf; + /* check for read access */ + if (semctl(semid, 0, IPC_STAT, semu) < 0) + goto fail_errno; + /* check for write access */ + if (semctl(semid, 0, IPC_SET, semu) < 0) + goto fail_errno; +#endif + } +#ifdef MDB_USE_SYSV_SEM + env->me_rmutex.semid = semid; + env->me_wmutex.semid = semid; + env->me_rmutex.semnum = 0; + env->me_wmutex.semnum = 1; + env->me_rmutex.locked = &env->me_txns->mti_rlocked; + env->me_wmutex.locked = &env->me_txns->mti_wlocked; +#endif + + return MDB_SUCCESS; + +fail_errno: + rc = ErrCode(); +fail: + 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" + /** 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|MDB_NOMETASYNC|MDB_MAPASYNC|MDB_NOMEMINIT) +#define CHANGELESS (MDB_FIXEDMAP|MDB_NOSUBDIR|MDB_RDONLY| \ + MDB_WRITEMAP|MDB_NOTLS|MDB_NOLOCK|MDB_NORDAHEAD) + +#if VALID_FLAGS & PERSISTENT_FLAGS & (CHANGEABLE|CHANGELESS) +# error "Persistent DB flags & env flags overlap, but both go in mm_flags" +#endif + +int ESECT +mdb_env_open(MDB_env *env, const char *path, unsigned int flags, mdb_mode_t mode) +{ + int oflags, rc, len, excl = -1; + char *lpath, *dpath; + + if (env->me_fd!=INVALID_HANDLE_VALUE || (flags & ~(CHANGEABLE|CHANGELESS))) + return EINVAL; + + 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_SUCCESS; + flags |= env->me_flags; + if (flags & MDB_RDONLY) { + /* silently ignore WRITEMAP when we're only getting read access */ + flags &= ~MDB_WRITEMAP; + } else { + if (!((env->me_free_pgs = mdb_midl_alloc(MDB_IDL_UM_MAX)) && + (env->me_dirty_list = calloc(MDB_IDL_UM_SIZE, sizeof(MDB_ID2))))) + rc = ENOMEM; + } + env->me_flags = flags |= MDB_ENV_ACTIVE; + if (rc) + goto leave; + + env->me_path = strdup(path); + env->me_dbxs = calloc(env->me_maxdbs, sizeof(MDB_dbx)); + env->me_dbflags = calloc(env->me_maxdbs, sizeof(uint16_t)); + env->me_dbiseqs = calloc(env->me_maxdbs, sizeof(unsigned int)); + if (!(env->me_dbxs && env->me_path && env->me_dbflags && env->me_dbiseqs)) { + rc = ENOMEM; + goto leave; + } + + /* For RDONLY, get lockfile after we know datafile exists */ + if (!(flags & (MDB_RDONLY|MDB_NOLOCK))) { + 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; + env->me_fd = CreateFile(dpath, oflags, FILE_SHARE_READ|FILE_SHARE_WRITE, + NULL, len, mode, NULL); +#else + if (F_ISSET(flags, MDB_RDONLY)) + oflags = O_RDONLY; + else + oflags = O_RDWR | O_CREAT; + + env->me_fd = open(dpath, oflags, mode); +#endif + if (env->me_fd == INVALID_HANDLE_VALUE) { + rc = ErrCode(); + goto leave; + } + + if ((flags & (MDB_RDONLY|MDB_NOLOCK)) == MDB_RDONLY) { + rc = mdb_env_setup_locks(env, lpath, mode, &excl); + if (rc) + goto leave; + } + + if ((rc = mdb_env_open2(env)) == MDB_SUCCESS) { + if (flags & (MDB_RDONLY|MDB_WRITEMAP)) { + env->me_mfd = env->me_fd; + } else { + /* Synchronous fd for meta writes. Needed even with + * MDB_NOSYNC/MDB_NOMETASYNC, in case these get reset. + */ +#ifdef _WIN32 + len = OPEN_EXISTING; + env->me_mfd = CreateFile(dpath, oflags, + FILE_SHARE_READ|FILE_SHARE_WRITE, NULL, len, + mode | FILE_FLAG_WRITE_THROUGH, NULL); +#else + oflags &= ~O_CREAT; + env->me_mfd = open(dpath, oflags | MDB_DSYNC, mode); +#endif + if (env->me_mfd == INVALID_HANDLE_VALUE) { + rc = ErrCode(); + goto leave; + } + } + DPRINTF(("opened dbenv %p", (void *) env)); + if (excl > 0) { + rc = mdb_env_share_locks(env, &excl); + if (rc) + goto leave; + } + if (!((flags & MDB_RDONLY) || + (env->me_pbuf = calloc(1, env->me_psize)))) + rc = ENOMEM; + if (!(flags & MDB_RDONLY)) { + MDB_txn *txn; + int tsize = sizeof(MDB_txn), size = tsize + env->me_maxdbs * + (sizeof(MDB_db)+sizeof(MDB_cursor *)+sizeof(unsigned int)+1); + txn = calloc(1, size); + if (txn) { + txn->mt_dbs = (MDB_db *)((char *)txn + tsize); + txn->mt_cursors = (MDB_cursor **)(txn->mt_dbs + env->me_maxdbs); + txn->mt_dbiseqs = (unsigned int *)(txn->mt_cursors + env->me_maxdbs); + txn->mt_dbflags = (unsigned char *)(txn->mt_dbiseqs + env->me_maxdbs); + txn->mt_env = env; + txn->mt_dbxs = env->me_dbxs; + env->me_txn0 = txn; + } else { + rc = ENOMEM; + } + } + } + +leave: + if (rc) { + mdb_env_close0(env, excl); + } + free(lpath); + return rc; +} + +/** Destroy resources from mdb_env_open(), clear our readers & DBIs */ +static void ESECT +mdb_env_close0(MDB_env *env, int excl) +{ + int i; + + if (!(env->me_flags & MDB_ENV_ACTIVE)) + return; + + /* Doing this here since me_dbxs may not exist during mdb_env_close */ + if (env->me_dbxs) { + for (i = env->me_maxdbs; --i > MAIN_DBI; ) + free(env->me_dbxs[i].md_name.mv_data); + free(env->me_dbxs); + } + + free(env->me_pbuf); + free(env->me_dbiseqs); + free(env->me_dbflags); + free(env->me_path); + free(env->me_dirty_list); + free(env->me_txn0); + mdb_midl_free(env->me_free_pgs); + + if (env->me_flags & MDB_ENV_TXKEY) { + pthread_key_delete(env->me_txkey); +#ifdef _WIN32 + /* Delete our key from the global list */ + for (i=0; i<mdb_tls_nkeys; i++) + if (mdb_tls_keys[i] == env->me_txkey) { + mdb_tls_keys[i] = mdb_tls_keys[mdb_tls_nkeys-1]; + mdb_tls_nkeys--; + break; + } +#endif + } + + if (env->me_map) { + munmap(env->me_map, env->me_mapsize); + } + if (env->me_mfd != env->me_fd && env->me_mfd != INVALID_HANDLE_VALUE) + (void) close(env->me_mfd); + if (env->me_fd != INVALID_HANDLE_VALUE) + (void) close(env->me_fd); + if (env->me_txns) { + MDB_PID_T pid = env->me_pid; + /* Clearing readers is done in this function because + * me_txkey with its destructor must be disabled first. + */ + for (i = env->me_numreaders; --i >= 0; ) + if (env->me_txns->mti_readers[i].mr_pid == pid) + env->me_txns->mti_readers[i].mr_pid = 0; +#ifdef _WIN32 + if (env->me_rmutex) { + CloseHandle(env->me_rmutex); + if (env->me_wmutex) CloseHandle(env->me_wmutex); + } + /* Windows automatically destroys the mutexes when + * the last handle closes. + */ +#elif defined(MDB_USE_SYSV_SEM) + if (env->me_rmutex.semid != -1) { + /* If we have the filelock: If we are the + * only remaining user, clean up semaphores. + */ + if (excl == 0) + mdb_env_excl_lock(env, &excl); + if (excl > 0) + semctl(env->me_rmutex.semid, 0, IPC_RMID); + } +#endif + munmap((void *)env->me_txns, (env->me_maxreaders-1)*sizeof(MDB_reader)+sizeof(MDB_txninfo)); + } + if (env->me_lfd != INVALID_HANDLE_VALUE) { +#ifdef _WIN32 + if (excl >= 0) { + /* Unlock the lockfile. Windows would have unlocked it + * after closing anyway, but not necessarily at once. + */ + UnlockFile(env->me_lfd, 0, 0, 1, 0); + } +#endif + (void) close(env->me_lfd); + } + + env->me_flags &= ~(MDB_ENV_ACTIVE|MDB_ENV_TXKEY); +} + +void ESECT +mdb_env_close(MDB_env *env) +{ + MDB_page *dp; + + if (env == NULL) + return; + + VGMEMP_DESTROY(env); + while ((dp = env->me_dpages) != NULL) { + VGMEMP_DEFINED(&dp->mp_next, sizeof(dp->mp_next)); + env->me_dpages = dp->mp_next; + free(dp); + } + + mdb_env_close0(env, 0); + 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 unsigned 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 unsigned 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 + unsigned short *u, *c, *end; + int x; + + end = (unsigned short *) ((char *) a->mv_data + a->mv_size); + u = (unsigned short *)a->mv_data; + c = (unsigned short *)b->mv_data; + do { + x = *u++ - *c++; + } while(!x && u < end); + return x; +#endif +} + +/** Compare two items pointing at size_t's of unknown alignment. */ +#ifdef MISALIGNED_OK +# define mdb_cmp_clong mdb_cmp_long +#else +# define mdb_cmp_clong mdb_cmp_cint +#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 %"Z"u", + nkeys, IS_LEAF(mp) ? "leaf" : "branch", IS_SUBP(mp) ? "sub-" : "", + mdb_dbg_pgno(mp))); + + 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 MDB_DEBUG + if (IS_LEAF(mp)) + DPRINTF(("found leaf index %u [%s], rc = %i", + i, DKEY(&nodekey), rc)); + else + DPRINTF(("found branch index %u [%s -> %"Z"u], 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 && nkeys > 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) +{ + if (mc->mc_snum) { +#if MDB_DEBUG + MDB_page *top = mc->mc_pg[mc->mc_top]; +#endif + mc->mc_snum--; + if (mc->mc_snum) + mc->mc_top--; + + DPRINTF(("popped page %"Z"u off db %d cursor %p", top->mp_pgno, + DDBI(mc), (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 %"Z"u on db %d cursor %p", mp->mp_pgno, + DDBI(mc), (void *) mc)); + + if (mc->mc_snum >= CURSOR_STACK) { + mc->mc_txn->mt_flags |= MDB_TXN_ERROR; + return MDB_CURSOR_FULL; + } + + 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. + * @param[out] lvl dirty_list inheritance level of found page. 1=current txn, 0=mapped page. + * @return 0 on success, non-zero on failure. + */ +static int +mdb_page_get(MDB_txn *txn, pgno_t pgno, MDB_page **ret, int *lvl) +{ + MDB_env *env = txn->mt_env; + MDB_page *p = NULL; + int level; + + if (!((txn->mt_flags & MDB_TXN_RDONLY) | (env->me_flags & MDB_WRITEMAP))) { + MDB_txn *tx2 = txn; + level = 1; + do { + MDB_ID2L dl = tx2->mt_u.dirty_list; + unsigned x; + /* Spilled pages were dirtied in this txn and flushed + * because the dirty list got full. Bring this page + * back in from the map (but don't unspill it here, + * leave that unless page_touch happens again). + */ + if (tx2->mt_spill_pgs) { + MDB_ID pn = pgno << 1; + x = mdb_midl_search(tx2->mt_spill_pgs, pn); + if (x <= tx2->mt_spill_pgs[0] && tx2->mt_spill_pgs[x] == pn) { + p = (MDB_page *)(env->me_map + env->me_psize * pgno); + goto done; + } + } + if (dl[0].mid) { + unsigned x = mdb_mid2l_search(dl, pgno); + if (x <= dl[0].mid && dl[x].mid == pgno) { + p = dl[x].mptr; + goto done; + } + } + level++; + } while ((tx2 = tx2->mt_parent) != NULL); + } + + if (pgno < txn->mt_next_pgno) { + level = 0; + p = (MDB_page *)(env->me_map + env->me_psize * pgno); + } else { + DPRINTF(("page %"Z"u not found", pgno)); + txn->mt_flags |= MDB_TXN_ERROR; + return MDB_PAGE_NOTFOUND; + } + +done: + *ret = p; + if (lvl) + *lvl = level; + return MDB_SUCCESS; +} + +/** Finish #mdb_page_search() / #mdb_page_search_lowest(). + * The cursor is at the root page, set up the rest of it. + */ +static int +mdb_page_search_root(MDB_cursor *mc, MDB_val *key, int flags) +{ + MDB_page *mp = mc->mc_pg[mc->mc_top]; + int rc; + DKBUF; + + while (IS_BRANCH(mp)) { + MDB_node *node; + indx_t i; + + DPRINTF(("branch page %"Z"u has %u keys", mp->mp_pgno, NUMKEYS(mp))); + mdb_cassert(mc, NUMKEYS(mp) > 1); + DPRINTF(("found index 0 to page %"Z"u", NODEPGNO(NODEPTR(mp, 0)))); + + if (flags & (MDB_PS_FIRST|MDB_PS_LAST)) { + i = 0; + if (flags & MDB_PS_LAST) + 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) { + mdb_cassert(mc, i > 0); + i--; + } + } + DPRINTF(("following index %u for key [%s]", i, DKEY(key))); + } + + mdb_cassert(mc, i < NUMKEYS(mp)); + node = NODEPTR(mp, i); + + if ((rc = mdb_page_get(mc->mc_txn, NODEPGNO(node), &mp, NULL)) != 0) + return rc; + + mc->mc_ki[mc->mc_top] = i; + if ((rc = mdb_cursor_push(mc, mp))) + return rc; + + if (flags & MDB_PS_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)); + mc->mc_txn->mt_flags |= MDB_TXN_ERROR; + return MDB_CORRUPTED; + } + + DPRINTF(("found leaf page %"Z"u for key [%s]", mp->mp_pgno, + key ? DKEY(key) : "null")); + mc->mc_flags |= C_INITIALIZED; + mc->mc_flags &= ~C_EOF; + + return MDB_SUCCESS; +} + +/** Search for the lowest key under the current branch page. + * This just bypasses a NUMKEYS check in the current page + * before calling mdb_page_search_root(), because the callers + * are all in situations where the current page is known to + * be underfilled. + */ +static int +mdb_page_search_lowest(MDB_cursor *mc) +{ + MDB_page *mp = mc->mc_pg[mc->mc_top]; + MDB_node *node = NODEPTR(mp, 0); + int rc; + + if ((rc = mdb_page_get(mc->mc_txn, NODEPGNO(node), &mp, NULL)) != 0) + return rc; + + mc->mc_ki[mc->mc_top] = 0; + if ((rc = mdb_cursor_push(mc, mp))) + return rc; + return mdb_page_search_root(mc, NULL, MDB_PS_FIRST); +} + +/** Search for the page a given key should be in. + * Push it and its parent pages on the cursor stack. + * @param[in,out] mc the cursor for this operation. + * @param[in] key the key to search for, or NULL for first/last page. + * @param[in] flags If MDB_PS_MODIFY is set, visited pages in the DB + * are touched (updated with new page numbers). + * If MDB_PS_FIRST or MDB_PS_LAST is set, find first or last leaf. + * This is used by #mdb_cursor_first() and #mdb_cursor_last(). + * If MDB_PS_ROOTONLY set, just fetch root node, no further lookups. + * @return 0 on success, non-zero on failure. + */ +static int +mdb_page_search(MDB_cursor *mc, MDB_val *key, int flags) +{ + int rc; + pgno_t root; + + /* Make sure the txn is still viable, then find the root from + * the txn's db table and set it as the root of the cursor's stack. + */ + if (F_ISSET(mc->mc_txn->mt_flags, MDB_TXN_ERROR)) { + DPUTS("transaction has failed, must abort"); + return MDB_BAD_TXN; + } else { + /* Make sure we're using an up-to-date root */ + if (*mc->mc_dbflag & DB_STALE) { + MDB_cursor mc2; + if (TXN_DBI_CHANGED(mc->mc_txn, mc->mc_dbi)) + return MDB_BAD_DBI; + mdb_cursor_init(&mc2, mc->mc_txn, MAIN_DBI, NULL); + rc = mdb_page_search(&mc2, &mc->mc_dbx->md_name, 0); + if (rc) + return rc; + { + MDB_val data; + int exact = 0; + uint16_t flags; + MDB_node *leaf = mdb_node_search(&mc2, + &mc->mc_dbx->md_name, &exact); + if (!exact) + return MDB_NOTFOUND; + rc = mdb_node_read(mc->mc_txn, leaf, &data); + if (rc) + return rc; + memcpy(&flags, ((char *) data.mv_data + offsetof(MDB_db, md_flags)), + sizeof(uint16_t)); + /* The txn may not know this DBI, or another process may + * have dropped and recreated the DB with other flags. + */ + if ((mc->mc_db->md_flags & PERSISTENT_FLAGS) != flags) + return MDB_INCOMPATIBLE; + memcpy(mc->mc_db, data.mv_data, sizeof(MDB_db)); + } + *mc->mc_dbflag &= ~DB_STALE; + } + root = mc->mc_db->md_root; + + if (root == P_INVALID) { /* Tree is empty. */ + DPUTS("tree is empty"); + return MDB_NOTFOUND; + } + } + + mdb_cassert(mc, root > 1); + if (!mc->mc_pg[0] || mc->mc_pg[0]->mp_pgno != root) + if ((rc = mdb_page_get(mc->mc_txn, root, &mc->mc_pg[0], NULL)) != 0) + return rc; + + mc->mc_snum = 1; + mc->mc_top = 0; + + DPRINTF(("db %d root page %"Z"u has flags 0x%X", + DDBI(mc), root, mc->mc_pg[0]->mp_flags)); + + if (flags & MDB_PS_MODIFY) { + if ((rc = mdb_page_touch(mc))) + return rc; + } + + if (flags & MDB_PS_ROOTONLY) + return MDB_SUCCESS; + + return mdb_page_search_root(mc, key, flags); +} + +static int +mdb_ovpage_free(MDB_cursor *mc, MDB_page *mp) +{ + MDB_txn *txn = mc->mc_txn; + pgno_t pg = mp->mp_pgno; + unsigned x = 0, ovpages = mp->mp_pages; + MDB_env *env = txn->mt_env; + MDB_IDL sl = txn->mt_spill_pgs; + MDB_ID pn = pg << 1; + int rc; + + DPRINTF(("free ov page %"Z"u (%d)", pg, ovpages)); + /* If the page is dirty or on the spill list we just acquired it, + * so we should give it back to our current free list, if any. + * Otherwise put it onto the list of pages we freed in this txn. + * + * Won't create me_pghead: me_pglast must be inited along with it. + * Unsupported in nested txns: They would need to hide the page + * range in ancestor txns' dirty and spilled lists. + */ + if (env->me_pghead && + !txn->mt_parent && + ((mp->mp_flags & P_DIRTY) || + (sl && (x = mdb_midl_search(sl, pn)) <= sl[0] && sl[x] == pn))) + { + unsigned i, j; + pgno_t *mop; + MDB_ID2 *dl, ix, iy; + rc = mdb_midl_need(&env->me_pghead, ovpages); + if (rc) + return rc; + if (!(mp->mp_flags & P_DIRTY)) { + /* This page is no longer spilled */ + if (x == sl[0]) + sl[0]--; + else + sl[x] |= 1; + goto release; + } + /* Remove from dirty list */ + dl = txn->mt_u.dirty_list; + x = dl[0].mid--; + for (ix = dl[x]; ix.mptr != mp; ix = iy) { + if (x > 1) { + x--; + iy = dl[x]; + dl[x] = ix; + } else { + mdb_cassert(mc, x > 1); + j = ++(dl[0].mid); + dl[j] = ix; /* Unsorted. OK when MDB_TXN_ERROR. */ + txn->mt_flags |= MDB_TXN_ERROR; + return MDB_CORRUPTED; + } + } + if (!(env->me_flags & MDB_WRITEMAP)) + mdb_dpage_free(env, mp); +release: + /* Insert in me_pghead */ + mop = env->me_pghead; + j = mop[0] + ovpages; + for (i = mop[0]; i && mop[i] < pg; i--) + mop[j--] = mop[i]; + while (j>i) + mop[j--] = pg++; + mop[0] += ovpages; + } else { + rc = mdb_midl_append_range(&txn->mt_free_pgs, pg, ovpages); + if (rc) + return rc; + } + mc->mc_db->md_overflow_pages -= ovpages; + return 0; +} + +/** 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, NULL)) != 0) { + DPRINTF(("read overflow page %"Z"u 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; + + DPRINTF(("===> get db %u key [%s]", dbi, DKEY(key))); + + if (!key || !data || dbi == FREE_DBI || !TXN_DBI_EXIST(txn, dbi)) + return EINVAL; + + if (txn->mt_flags & MDB_TXN_ERROR) + return MDB_BAD_TXN; + + 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 %"Z"u, 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) { + /* undo cursor_pop before returning */ + mc->mc_top++; + mc->mc_snum++; + 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])); + } + mdb_cassert(mc, 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, NULL)) != 0) { + /* mc will be inconsistent if caller does mc_snum++ as above */ + mc->mc_flags &= ~(C_INITIALIZED|C_EOF); + return rc; + } + + mdb_cursor_push(mc, mp); + if (!move_right) + mc->mc_ki[mc->mc_top] = NUMKEYS(mp)-1; + + 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; + } + + mdb_cassert(mc, 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_NOTFOUND) { + if (rc == MDB_SUCCESS) + MDB_GET_KEY(leaf, key); + return rc; + } + } + } else { + mc->mc_xcursor->mx_cursor.mc_flags &= ~(C_INITIALIZED|C_EOF); + if (op == MDB_NEXT_DUP) + return MDB_NOTFOUND; + } + } + + DPRINTF(("cursor_next: top page is %"Z"u in cursor %p", + mdb_dbg_pgno(mp), (void *) mc)); + if (mc->mc_flags & C_DEL) + goto skip; + + if (mc->mc_ki[mc->mc_top] + 1u >= NUMKEYS(mp)) { + DPUTS("=====> move to next sibling page"); + if ((rc = mdb_cursor_sibling(mc, 1)) != MDB_SUCCESS) { + mc->mc_flags |= C_EOF; + return rc; + } + mp = mc->mc_pg[mc->mc_top]; + DPRINTF(("next page is %"Z"u, key index %u", mp->mp_pgno, mc->mc_ki[mc->mc_top])); + } else + mc->mc_ki[mc->mc_top]++; + +skip: + DPRINTF(("==> cursor points to page %"Z"u with %u keys, key index %u", + mdb_dbg_pgno(mp), 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; + } + + mdb_cassert(mc, 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_GET_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; + + mdb_cassert(mc, 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_PREV || op == MDB_PREV_DUP) { + rc = mdb_cursor_prev(&mc->mc_xcursor->mx_cursor, data, NULL, MDB_PREV); + if (op != MDB_PREV || rc != MDB_NOTFOUND) { + if (rc == MDB_SUCCESS) { + MDB_GET_KEY(leaf, key); + mc->mc_flags &= ~C_EOF; + } + return rc; + } + } + } else { + mc->mc_xcursor->mx_cursor.mc_flags &= ~(C_INITIALIZED|C_EOF); + if (op == MDB_PREV_DUP) + return MDB_NOTFOUND; + } + } + + DPRINTF(("cursor_prev: top page is %"Z"u in cursor %p", + mdb_dbg_pgno(mp), (void *) mc)); + + if (mc->mc_ki[mc->mc_top] == 0) { + DPUTS("=====> move to prev sibling page"); + if ((rc = mdb_cursor_sibling(mc, 0)) != MDB_SUCCESS) { + return rc; + } + mp = mc->mc_pg[mc->mc_top]; + mc->mc_ki[mc->mc_top] = NUMKEYS(mp) - 1; + DPRINTF(("prev page is %"Z"u, 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 %"Z"u with %u keys, key index %u", + mdb_dbg_pgno(mp), 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; + } + + mdb_cassert(mc, 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_GET_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 = NULL; + DKBUF; + + if (key->mv_size == 0) + return MDB_BAD_VALSIZE; + + if (mc->mc_xcursor) + mc->mc_xcursor->mx_cursor.mc_flags &= ~(C_INITIALIZED|C_EOF); + + /* 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_GET_KEY2(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; + 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_GET_KEY2(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; + if (exactp) + *exactp = 1; + goto set1; + } + if (rc < 0) { + if (mc->mc_ki[mc->mc_top] < NUMKEYS(mp)) { + /* This is definitely the right page, skip search_page */ + if (mp->mp_flags & P_LEAF2) { + nodekey.mv_data = LEAF2KEY(mp, + mc->mc_ki[mc->mc_top], nodekey.mv_size); + } else { + leaf = NODEPTR(mp, mc->mc_ki[mc->mc_top]); + MDB_GET_KEY2(leaf, nodekey); + } + rc = mc->mc_dbx->md_cmp(key, &nodekey); + if (rc == 0) { + /* current node was the one we wanted */ + if (exactp) + *exactp = 1; + goto set1; + } + } + 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; + if (op == MDB_SET_RANGE && !exactp) { + rc = 0; + goto set1; + } else + return MDB_NOTFOUND; + } + } + + rc = mdb_page_search(mc, key, 0); + if (rc != MDB_SUCCESS) + return rc; + + mp = mc->mc_pg[mc->mc_top]; + mdb_cassert(mc, 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]; + mdb_cassert(mc, IS_LEAF(mp)); + leaf = NODEPTR(mp, 0); + } + +set1: + mc->mc_flags |= C_INITIALIZED; + mc->mc_flags &= ~C_EOF; + + if (IS_LEAF2(mp)) { + if (op == MDB_SET_RANGE || op == MDB_SET_KEY) { + 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_KEY || 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; + rc = 0; + *data = d2; + } + + } else { + if (mc->mc_xcursor) + mc->mc_xcursor->mx_cursor.mc_flags &= ~(C_INITIALIZED|C_EOF); + 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 || op == MDB_SET_KEY) + MDB_GET_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_xcursor) + mc->mc_xcursor->mx_cursor.mc_flags &= ~(C_INITIALIZED|C_EOF); + + if (!(mc->mc_flags & C_INITIALIZED) || mc->mc_top) { + rc = mdb_page_search(mc, NULL, MDB_PS_FIRST); + if (rc != MDB_SUCCESS) + return rc; + } + mdb_cassert(mc, 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 ((rc = mdb_node_read(mc->mc_txn, leaf, data)) != MDB_SUCCESS) + return rc; + } + } + MDB_GET_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; + + if (mc->mc_xcursor) + mc->mc_xcursor->mx_cursor.mc_flags &= ~(C_INITIALIZED|C_EOF); + + if (!(mc->mc_flags & C_EOF)) { + + if (!(mc->mc_flags & C_INITIALIZED) || mc->mc_top) { + rc = mdb_page_search(mc, NULL, MDB_PS_LAST); + if (rc != MDB_SUCCESS) + return rc; + } + mdb_cassert(mc, IS_LEAF(mc->mc_pg[mc->mc_top])); + + } + mc->mc_ki[mc->mc_top] = NUMKEYS(mc->mc_pg[mc->mc_top]) - 1; + mc->mc_flags |= C_INITIALIZED|C_EOF; + leaf = NODEPTR(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top]); + + 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 ((rc = mdb_node_read(mc->mc_txn, leaf, data)) != MDB_SUCCESS) + return rc; + } + } + + MDB_GET_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; + int (*mfunc)(MDB_cursor *mc, MDB_val *key, MDB_val *data); + + if (mc == NULL) + return EINVAL; + + if (mc->mc_txn->mt_flags & MDB_TXN_ERROR) + return MDB_BAD_TXN; + + switch (op) { + case MDB_GET_CURRENT: + if (!(mc->mc_flags & C_INITIALIZED)) { + rc = EINVAL; + } else { + MDB_page *mp = mc->mc_pg[mc->mc_top]; + int nkeys = NUMKEYS(mp); + if (!nkeys || mc->mc_ki[mc->mc_top] >= nkeys) { + mc->mc_ki[mc->mc_top] = nkeys; + rc = MDB_NOTFOUND; + break; + } + rc = MDB_SUCCESS; + 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); + } else { + MDB_node *leaf = NODEPTR(mp, mc->mc_ki[mc->mc_top]); + MDB_GET_KEY(leaf, key); + if (data) { + if (F_ISSET(leaf->mn_flags, F_DUPDATA)) { + if (mc->mc_flags & C_DEL) + mdb_xcursor_init1(mc, leaf); + rc = mdb_cursor_get(&mc->mc_xcursor->mx_cursor, data, NULL, MDB_GET_CURRENT); + } else { + rc = mdb_node_read(mc->mc_txn, leaf, data); + } + } + } + } + break; + case MDB_GET_BOTH: + case MDB_GET_BOTH_RANGE: + if (data == NULL) { + rc = EINVAL; + break; + } + if (mc->mc_xcursor == NULL) { + rc = MDB_INCOMPATIBLE; + break; + } + /* FALLTHRU */ + case MDB_SET: + case MDB_SET_KEY: + case MDB_SET_RANGE: + if (key == NULL) { + rc = EINVAL; + } else { + rc = mdb_cursor_set(mc, key, data, op, + op == MDB_SET_RANGE ? NULL : &exact); + } + break; + case MDB_GET_MULTIPLE: + if (data == NULL || !(mc->mc_flags & C_INITIALIZED)) { + rc = EINVAL; + break; + } + if (!(mc->mc_db->md_flags & MDB_DUPFIXED)) { + rc = MDB_INCOMPATIBLE; + 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) { + rc = EINVAL; + break; + } + if (!(mc->mc_db->md_flags & MDB_DUPFIXED)) { + rc = MDB_INCOMPATIBLE; + 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)) { + rc = mdb_cursor_last(mc, key, data); + if (rc) + break; + mc->mc_flags |= C_INITIALIZED; + mc->mc_ki[mc->mc_top]++; + } + rc = mdb_cursor_prev(mc, key, data, op); + break; + case MDB_FIRST: + rc = mdb_cursor_first(mc, key, data); + break; + case MDB_FIRST_DUP: + mfunc = mdb_cursor_first; + mmove: + if (data == NULL || !(mc->mc_flags & C_INITIALIZED)) { + rc = EINVAL; + break; + } + if (mc->mc_xcursor == NULL) { + rc = MDB_INCOMPATIBLE; + break; + } + { + MDB_node *leaf = NODEPTR(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top]); + if (!F_ISSET(leaf->mn_flags, F_DUPDATA)) { + MDB_GET_KEY(leaf, key); + rc = mdb_node_read(mc->mc_txn, leaf, data); + break; + } + } + if (!(mc->mc_xcursor->mx_cursor.mc_flags & C_INITIALIZED)) { + rc = EINVAL; + break; + } + rc = mfunc(&mc->mc_xcursor->mx_cursor, data, NULL); + break; + case MDB_LAST: + rc = mdb_cursor_last(mc, key, data); + break; + case MDB_LAST_DUP: + mfunc = mdb_cursor_last; + goto mmove; + default: + DPRINTF(("unhandled/unimplemented cursor operation %u", op)); + rc = EINVAL; + break; + } + + if (mc->mc_flags & C_DEL) + mc->mc_flags ^= C_DEL; + + return rc; +} + +/** Touch all the pages in the cursor stack. Set mc_top. + * 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 = MDB_SUCCESS; + + if (mc->mc_dbi > MAIN_DBI && !(*mc->mc_dbflag & DB_DIRTY)) { + MDB_cursor mc2; + MDB_xcursor mcx; + if (TXN_DBI_CHANGED(mc->mc_txn, mc->mc_dbi)) + return MDB_BAD_DBI; + mdb_cursor_init(&mc2, mc->mc_txn, MAIN_DBI, &mcx); + rc = mdb_page_search(&mc2, &mc->mc_dbx->md_name, MDB_PS_MODIFY); + if (rc) + return rc; + *mc->mc_dbflag |= DB_DIRTY; + } + mc->mc_top = 0; + if (mc->mc_snum) { + do { + rc = mdb_page_touch(mc); + } while (!rc && ++(mc->mc_top) < mc->mc_snum); + mc->mc_top = mc->mc_snum-1; + } + return rc; +} + +/** Do not spill pages to disk if txn is getting full, may fail instead */ +#define MDB_NOSPILL 0x8000 + +int +mdb_cursor_put(MDB_cursor *mc, MDB_val *key, MDB_val *data, + unsigned int flags) +{ + MDB_env *env; + MDB_node *leaf = NULL; + MDB_page *fp, *mp; + uint16_t fp_flags; + MDB_val xdata, *rdata, dkey, olddata; + MDB_db dummy; + int do_sub = 0, insert_key, insert_data; + unsigned int mcount = 0, dcount = 0, nospill; + size_t nsize; + int rc, rc2; + unsigned int nflags; + DKBUF; + + if (mc == NULL || key == NULL) + return EINVAL; + + env = mc->mc_txn->mt_env; + + /* Check this first so counter will always be zero on any + * early failures. + */ + if (flags & MDB_MULTIPLE) { + dcount = data[1].mv_size; + data[1].mv_size = 0; + if (!F_ISSET(mc->mc_db->md_flags, MDB_DUPFIXED)) + return MDB_INCOMPATIBLE; + } + + nospill = flags & MDB_NOSPILL; + flags &= ~MDB_NOSPILL; + + if (mc->mc_txn->mt_flags & (MDB_TXN_RDONLY|MDB_TXN_ERROR)) + return (mc->mc_txn->mt_flags & MDB_TXN_RDONLY) ? EACCES : MDB_BAD_TXN; + + if (key->mv_size-1 >= ENV_MAXKEY(env)) + return MDB_BAD_VALSIZE; + +#if SIZE_MAX > MAXDATASIZE + if (data->mv_size > ((mc->mc_db->md_flags & MDB_DUPSORT) ? ENV_MAXKEY(env) : MAXDATASIZE)) + return MDB_BAD_VALSIZE; +#else + if ((mc->mc_db->md_flags & MDB_DUPSORT) && data->mv_size > ENV_MAXKEY(env)) + return MDB_BAD_VALSIZE; +#endif + + DPRINTF(("==> put db %d key [%s], size %"Z"u, data size %"Z"u", + DDBI(mc), 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) { + /* new database, cursor has nothing to point to */ + mc->mc_snum = 0; + mc->mc_top = 0; + mc->mc_flags &= ~C_INITIALIZED; + rc = MDB_NO_ROOT; + } else { + int exact = 0; + MDB_val d2; + if (flags & MDB_APPEND) { + MDB_val k2; + rc = mdb_cursor_last(mc, &k2, &d2); + if (rc == 0) { + rc = mc->mc_dbx->md_cmp(key, &k2); + if (rc > 0) { + rc = MDB_NOTFOUND; + mc->mc_ki[mc->mc_top]++; + } else { + /* new key is <= last key */ + rc = MDB_KEYEXIST; + } + } + } else { + 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; + } + + if (mc->mc_flags & C_DEL) + mc->mc_flags ^= C_DEL; + + /* Cursor is positioned, check for room in the dirty list */ + if (!nospill) { + if (flags & MDB_MULTIPLE) { + rdata = &xdata; + xdata.mv_size = data->mv_size * dcount; + } else { + rdata = data; + } + if ((rc2 = mdb_page_spill(mc, key, rdata))) + return rc2; + } + + if (rc == MDB_NO_ROOT) { + MDB_page *np; + /* new database, write a root leaf page */ + DPUTS("allocating new root leaf page"); + if ((rc2 = mdb_page_new(mc, P_LEAF, 1, &np))) { + return rc2; + } + 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; + } else { + /* make sure all cursor pages are writable */ + rc2 = mdb_cursor_touch(mc); + if (rc2) + return rc2; + } + + insert_key = insert_data = rc; + if (insert_key) { + /* The key does not exist */ + DPRINTF(("inserting key at index %i", mc->mc_ki[mc->mc_top])); + if ((mc->mc_db->md_flags & MDB_DUPSORT) && + LEAFSIZE(key, data) > env->me_nodemax) + { + /* Too big for a node, insert in sub-DB. Set up an empty + * "old sub-page" for prep_subDB to expand to a full page. + */ + fp_flags = P_LEAF|P_DIRTY; + fp = env->me_pbuf; + fp->mp_pad = data->mv_size; /* used if MDB_DUPFIXED */ + fp->mp_lower = fp->mp_upper = (PAGEHDRSZ-PAGEBASE); + olddata.mv_size = PAGEHDRSZ; + goto prep_subDB; + } + } else { + /* there's only a key anyway, so this is a no-op */ + if (IS_LEAF2(mc->mc_pg[mc->mc_top])) { + char *ptr; + unsigned int ksize = mc->mc_db->md_pad; + if (key->mv_size != ksize) + return MDB_BAD_VALSIZE; + ptr = LEAF2KEY(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top], ksize); + memcpy(ptr, key->mv_data, ksize); +fix_parent: + /* if overwriting slot 0 of leaf, need to + * update branch key if there is a parent page + */ + if (mc->mc_top && !mc->mc_ki[mc->mc_top]) { + unsigned short top = mc->mc_top; + mc->mc_top--; + /* slot 0 is always an empty key, find real slot */ + while (mc->mc_top && !mc->mc_ki[mc->mc_top]) + mc->mc_top--; + if (mc->mc_ki[mc->mc_top]) + rc2 = mdb_update_key(mc, key); + else + rc2 = MDB_SUCCESS; + mc->mc_top = top; + if (rc2) + return rc2; + } + return MDB_SUCCESS; + } + +more: + leaf = NODEPTR(mc->mc_pg[mc->mc_top], mc->mc_ki[mc->mc_top]); + olddata.mv_size = NODEDSZ(leaf); + olddata.mv_data = NODEDATA(leaf); + + /* DB has dups? */ + if (F_ISSET(mc->mc_db->md_flags, MDB_DUPSORT)) { + /* Prepare (sub-)page/sub-DB to accept the new item, + * if needed. fp: old sub-page or a header faking + * it. mp: new (sub-)page. offset: growth in page + * size. xdata: node data with new page or DB. + */ + unsigned i, offset = 0; + mp = fp = xdata.mv_data = env->me_pbuf; + mp->mp_pgno = mc->mc_pg[mc->mc_top]->mp_pgno; + + /* Was a single item before, must convert now */ + if (!F_ISSET(leaf->mn_flags, F_DUPDATA)) { + /* Just overwrite the current item */ + if (flags == MDB_CURRENT) + goto current; + +#if UINT_MAX < SIZE_MAX + if (mc->mc_dbx->md_dcmp == mdb_cmp_int && olddata.mv_size == sizeof(size_t)) + mc->mc_dbx->md_dcmp = mdb_cmp_clong; +#endif + /* does data match? */ + if (!mc->mc_dbx->md_dcmp(data, &olddata)) { + if (flags & MDB_NODUPDATA) + return MDB_KEYEXIST; + /* overwrite it */ + goto current; + } + + /* Back up original data item */ + dkey.mv_size = olddata.mv_size; + dkey.mv_data = memcpy(fp+1, olddata.mv_data, olddata.mv_size); + + /* Make sub-page header for the dup items, with dummy body */ + fp->mp_flags = P_LEAF|P_DIRTY|P_SUBP; + fp->mp_lower = (PAGEHDRSZ-PAGEBASE); + xdata.mv_size = 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; + xdata.mv_size += 2 * data->mv_size; /* leave space for 2 more */ + } else { + xdata.mv_size += 2 * (sizeof(indx_t) + NODESIZE) + + (dkey.mv_size & 1) + (data->mv_size & 1); + } + fp->mp_upper = xdata.mv_size - PAGEBASE; + olddata.mv_size = xdata.mv_size; /* pretend olddata is fp */ + } else if (leaf->mn_flags & F_SUBDATA) { + /* Data is on sub-DB, just store it */ + flags |= F_DUPDATA|F_SUBDATA; + goto put_sub; + } else { + /* Data is on sub-page */ + fp = olddata.mv_data; + switch (flags) { + default: + if (!(mc->mc_db->md_flags & MDB_DUPFIXED)) { + offset = EVEN(NODESIZE + sizeof(indx_t) + + data->mv_size); + break; + } + offset = fp->mp_pad; + if (SIZELEFT(fp) < offset) { + offset *= 4; /* space for 4 more */ + break; + } + /* FALLTHRU: Big enough MDB_DUPFIXED sub-page */ + case MDB_CURRENT: + fp->mp_flags |= P_DIRTY; + COPY_PGNO(fp->mp_pgno, mp->mp_pgno); + mc->mc_xcursor->mx_cursor.mc_pg[0] = fp; + flags |= F_DUPDATA; + goto put_sub; + } + xdata.mv_size = olddata.mv_size + offset; + } + + fp_flags = fp->mp_flags; + if (NODESIZE + NODEKSZ(leaf) + xdata.mv_size > env->me_nodemax) { + /* Too big for a sub-page, convert to sub-DB */ + fp_flags &= ~P_SUBP; +prep_subDB: + if (mc->mc_db->md_flags & MDB_DUPFIXED) { + fp_flags |= P_LEAF2; + dummy.md_pad = fp->mp_pad; + dummy.md_flags = MDB_DUPFIXED; + if (mc->mc_db->md_flags & MDB_INTEGERDUP) + dummy.md_flags |= MDB_INTEGERKEY; + } else { + dummy.md_pad = 0; + dummy.md_flags = 0; + } + dummy.md_depth = 1; + dummy.md_branch_pages = 0; + dummy.md_leaf_pages = 1; + dummy.md_overflow_pages = 0; + dummy.md_entries = NUMKEYS(fp); + xdata.mv_size = sizeof(MDB_db); + xdata.mv_data = &dummy; + if ((rc = mdb_page_alloc(mc, 1, &mp))) + return rc; + offset = env->me_psize - olddata.mv_size; + flags |= F_DUPDATA|F_SUBDATA; + dummy.md_root = mp->mp_pgno; + } + if (mp != fp) { + mp->mp_flags = fp_flags | P_DIRTY; + mp->mp_pad = fp->mp_pad; + mp->mp_lower = fp->mp_lower; + mp->mp_upper = fp->mp_upper + offset; + if (fp_flags & P_LEAF2) { + memcpy(METADATA(mp), METADATA(fp), NUMKEYS(fp) * fp->mp_pad); + } else { + memcpy((char *)mp + mp->mp_upper + PAGEBASE, (char *)fp + fp->mp_upper + PAGEBASE, + olddata.mv_size - fp->mp_upper - PAGEBASE); + for (i=0; i<NUMKEYS(fp); i++) + mp->mp_ptrs[i] = fp->mp_ptrs[i] + offset; + } + } + + rdata = &xdata; + flags |= F_DUPDATA; + do_sub = 1; + if (!insert_key) + mdb_node_del(mc, 0); + goto new_sub; + } +current: + /* overflow page overwrites need special handling */ + if (F_ISSET(leaf->mn_flags, F_BIGDATA)) { + MDB_page *omp; + pgno_t pg; + int level, ovpages, dpages = OVPAGES(data->mv_size, env->me_psize); + + memcpy(&pg, olddata.mv_data, sizeof(pg)); + if ((rc2 = mdb_page_get(mc->mc_txn, pg, &omp, &level)) != 0) + return rc2; + ovpages = omp->mp_pages; + + /* Is the ov page large enough? */ + if (ovpages >= dpages) { + if (!(omp->mp_flags & P_DIRTY) && + (level || (env->me_flags & MDB_WRITEMAP))) + { + rc = mdb_page_unspill(mc->mc_txn, omp, &omp); + if (rc) + return rc; + level = 0; /* dirty in this txn or clean */ + } + /* Is it dirty? */ + if (omp->mp_flags & P_DIRTY) { + /* yes, overwrite it. Note in this case we don't + * bother to try shrinking the page if the new data + * is smaller than the overflow threshold. + */ + if (level > 1) { + /* It is writable only in a parent txn */ + size_t sz = (size_t) env->me_psize * ovpages, off; + MDB_page *np = mdb_page_malloc(mc->mc_txn, ovpages); + MDB_ID2 id2; + if (!np) + return ENOMEM; + id2.mid = pg; + id2.mptr = np; + rc2 = mdb_mid2l_insert(mc->mc_txn->mt_u.dirty_list, &id2); + mdb_cassert(mc, rc2 == 0); + if (!(flags & MDB_RESERVE)) { + /* Copy end of page, adjusting alignment so + * compiler may copy words instead of bytes. + */ + off = (PAGEHDRSZ + data->mv_size) & -sizeof(size_t); + memcpy((size_t *)((char *)np + off), + (size_t *)((char *)omp + off), sz - off); + sz = PAGEHDRSZ; + } + memcpy(np, omp, sz); /* Copy beginning of page */ + omp = np; + } + SETDSZ(leaf, data->mv_size); + if (F_ISSET(flags, MDB_RESERVE)) + data->mv_data = METADATA(omp); + else + memcpy(METADATA(omp), data->mv_data, data->mv_size); + return MDB_SUCCESS; + } + } + if ((rc2 = mdb_ovpage_free(mc, omp)) != MDB_SUCCESS) + return rc2; + } else if (data->mv_size == olddata.mv_size) { + /* same size, just replace it. Note that we could + * also reuse this node if the new data is smaller, + * but instead we opt to shrink the node in that case. + */ + if (F_ISSET(flags, MDB_RESERVE)) + data->mv_data = olddata.mv_data; + else if (!(mc->mc_flags & C_SUB)) + memcpy(olddata.mv_data, data->mv_data, data->mv_size); + else { + memcpy(NODEKEY(leaf), key->mv_data, key->mv_size); + goto fix_parent; + } + return MDB_SUCCESS; + } + mdb_node_del(mc, 0); + } + + rdata = data; + +new_sub: + nflags = flags & NODE_ADD_FLAGS; + nsize = IS_LEAF2(mc->mc_pg[mc->mc_top]) ? key->mv_size : mdb_leaf_size(env, key, rdata); + if (SIZELEFT(mc->mc_pg[mc->mc_top]) < nsize) { + if (( flags & (F_DUPDATA|F_SUBDATA)) == F_DUPDATA ) + nflags &= ~MDB_APPEND; /* sub-page may need room to grow */ + if (!insert_key) + nflags |= MDB_SPLIT_REPLACE; + 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 && insert_key) { + /* 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]; + + 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 || m3->mc_snum < mc->mc_snum) continue; + if (m3->mc_pg[i] == mp && m3->mc_ki[i] >= mc->mc_ki[i]) { + m3->mc_ki[i]++; + } + } + } + } + + if (rc == MDB_SUCCESS) { + /* 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) { + int xflags; + size_t ecount; +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|MDB_NOSPILL; + } else { + mdb_xcursor_init1(mc, leaf); + xflags = (flags & MDB_NODUPDATA) ? + MDB_NOOVERWRITE|MDB_NOSPILL : MDB_NOSPILL; + } + /* converted, write the original data first */ + if (dkey.mv_size) { + rc = mdb_cursor_put(&mc->mc_xcursor->mx_cursor, &dkey, &xdata, xflags); + if (rc) + goto bad_sub; + { + /* 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 || m2->mc_snum < mc->mc_snum) continue; + if (!(m2->mc_flags & C_INITIALIZED)) continue; + if (m2->mc_pg[i] == mp && m2->mc_ki[i] == mc->mc_ki[i]) { + mdb_xcursor_init1(m2, leaf); + } + } + } + /* we've done our job */ + dkey.mv_size = 0; + } + ecount = mc->mc_xcursor->mx_db.md_entries; + if (flags & MDB_APPENDDUP) + xflags |= MDB_APPEND; + rc = mdb_cursor_put(&mc->mc_xcursor->mx_cursor, data, &xdata, xflags); + if (flags & F_SUBDATA) { + void *db = NODEDATA(leaf); + memcpy(db, &mc->mc_xcursor->mx_db, sizeof(MDB_db)); + } + insert_data = mc->mc_xcursor->mx_db.md_entries - ecount; + } + /* Increment count unless we just replaced an existing item. */ + if (insert_data) + mc->mc_db->md_entries++; + if (insert_key) { + /* Invalidate txn if we created an empty sub-DB */ + if (rc) + goto bad_sub; + /* If we succeeded and the key didn't exist before, + * make sure the cursor is marked valid. + */ + mc->mc_flags |= C_INITIALIZED; + } + if (flags & MDB_MULTIPLE) { + if (!rc) { + mcount++; + /* let caller know how many succeeded, if any */ + data[1].mv_size = mcount; + if (mcount < dcount) { + data[0].mv_data = (char *)data[0].mv_data + data[0].mv_size; + insert_key = insert_data = 0; + goto more; + } + } + } + return rc; +bad_sub: + if (rc == MDB_KEYEXIST) /* should not happen, we deleted that item */ + rc = MDB_CORRUPTED; + } + mc->mc_txn->mt_flags |= MDB_TXN_ERROR; + return rc; +} + +int +mdb_cursor_del(MDB_cursor *mc, unsigned int flags) +{ + MDB_node *leaf; + MDB_page *mp; + int rc; + + if (mc->mc_txn->mt_flags & (MDB_TXN_RDONLY|MDB_TXN_ERROR)) + return (mc->mc_txn->mt_flags & MDB_TXN_RDONLY) ? EACCES : MDB_BAD_TXN; + + if (!(mc->mc_flags & C_INITIALIZED)) + return EINVAL; + + if (mc->mc_ki[mc->mc_top] >= NUMKEYS(mc->mc_pg[mc->mc_top])) + return MDB_NOTFOUND; + + if (!(flags & MDB_NOSPILL) && (rc = mdb_page_spill(mc, NULL, NULL))) + return rc; + + rc = mdb_cursor_touch(mc); + if (rc) + return rc; + + mp = mc->mc_pg[mc->mc_top]; + if (IS_LEAF2(mp)) + goto del_key; + leaf = NODEPTR(mp, mc->mc_ki[mc->mc_top]); + + if (F_ISSET(leaf->mn_flags, F_DUPDATA)) { + if (flags & MDB_NODUPDATA) { + /* mdb_cursor_del0() will subtract the final entry */ + mc->mc_db->md_entries -= mc->mc_xcursor->mx_db.md_entries - 1; + } else { + 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, MDB_NOSPILL); + if (rc) + return rc; + /* 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 */ + void *db = NODEDATA(leaf); + memcpy(db, &mc->mc_xcursor->mx_db, sizeof(MDB_db)); + } else { + MDB_cursor *m2; + /* shrink fake page */ + mdb_node_shrink(mp, mc->mc_ki[mc->mc_top]); + leaf = NODEPTR(mp, mc->mc_ki[mc->mc_top]); + mc->mc_xcursor->mx_cursor.mc_pg[0] = NODEDATA(leaf); + /* fix other sub-DB cursors pointed at this fake page */ + for (m2 = mc->mc_txn->mt_cursors[mc->mc_dbi]; m2; m2=m2->mc_next) { + if (m2 == mc || m2->mc_snum < mc->mc_snum) continue; + if (m2->mc_pg[mc->mc_top] == mp && + m2->mc_ki[mc->mc_top] == mc->mc_ki[mc->mc_top]) + m2->mc_xcursor->mx_cursor.mc_pg[0] = NODEDATA(leaf); + } + } + mc->mc_db->md_entries--; + mc->mc_flags |= C_DEL; + 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) + goto fail; + } + } + + /* add overflow pages to free list */ + if (F_ISSET(leaf->mn_flags, F_BIGDATA)) { + MDB_page *omp; + pgno_t pg; + + memcpy(&pg, NODEDATA(leaf), sizeof(pg)); + if ((rc = mdb_page_get(mc->mc_txn, pg, &omp, NULL)) || + (rc = mdb_ovpage_free(mc, omp))) + goto fail; + } + +del_key: + return mdb_cursor_del0(mc); + +fail: + mc->mc_txn->mt_flags |= MDB_TXN_ERROR; + return rc; +} + +/** 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. + * @param[out] mp Address of a page, or NULL on failure. + * @return 0 on success, non-zero on failure. + */ +static int +mdb_page_new(MDB_cursor *mc, uint32_t flags, int num, MDB_page **mp) +{ + MDB_page *np; + int rc; + + if ((rc = mdb_page_alloc(mc, num, &np))) + return rc; + DPRINTF(("allocated new mpage %"Z"u, page size %u", + np->mp_pgno, mc->mc_txn->mt_env->me_psize)); + np->mp_flags = flags | P_DIRTY; + np->mp_lower = (PAGEHDRSZ-PAGEBASE); + np->mp_upper = mc->mc_txn->mt_env->me_psize - PAGEBASE; + + 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; + } + *mp = np; + + return 0; +} + +/** 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 (sz > env->me_nodemax) { + /* put on overflow page */ + sz -= data->mv_size - sizeof(pgno_t); + } + + return EVEN(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_nodemax) { + /* 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>MDB_PAGE_FULL - 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; + ssize_t room; + indx_t ofs; + MDB_node *node; + MDB_page *mp = mc->mc_pg[mc->mc_top]; + MDB_page *ofp = NULL; /* overflow page */ + DKBUF; + + mdb_cassert(mc, mp->mp_upper >= mp->mp_lower); + + DPRINTF(("add to %s %spage %"Z"u index %i, data size %"Z"u key size %"Z"u [%s]", + IS_LEAF(mp) ? "leaf" : "branch", + IS_SUBP(mp) ? "sub-" : "", + mdb_dbg_pgno(mp), 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; + } + + room = (ssize_t)SIZELEFT(mp) - (ssize_t)sizeof(indx_t); + if (key != NULL) + node_size += key->mv_size; + if (IS_LEAF(mp)) { + mdb_cassert(mc, data); + if (F_ISSET(flags, F_BIGDATA)) { + /* Data already on overflow page. */ + node_size += sizeof(pgno_t); + } else if (node_size + data->mv_size > mc->mc_txn->mt_env->me_nodemax) { + int ovpages = OVPAGES(data->mv_size, mc->mc_txn->mt_env->me_psize); + int rc; + /* Put data on overflow page. */ + DPRINTF(("data size is %"Z"u, node would be %"Z"u, put data on overflow page", + data->mv_size, node_size+data->mv_size)); + node_size = EVEN(node_size + sizeof(pgno_t)); + if ((ssize_t)node_size > room) + goto full; + if ((rc = mdb_page_new(mc, P_OVERFLOW, ovpages, &ofp))) + return rc; + DPRINTF(("allocated overflow page %"Z"u", ofp->mp_pgno)); + flags |= F_BIGDATA; + goto update; + } else { + node_size += data->mv_size; + } + } + node_size = EVEN(node_size); + if ((ssize_t)node_size > room) + goto full; + +update: + /* 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; + mdb_cassert(mc, 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)) { + mdb_cassert(mc, 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; + +full: + DPRINTF(("not enough room in page %"Z"u, got %u ptrs", + mdb_dbg_pgno(mp), NUMKEYS(mp))); + DPRINTF(("upper-lower = %u - %u = %"Z"d", mp->mp_upper,mp->mp_lower,room)); + DPRINTF(("node size = %"Z"u", node_size)); + mc->mc_txn->mt_flags |= MDB_TXN_ERROR; + return MDB_PAGE_FULL; +} + +/** Delete the specified node from a page. + * @param[in] mc Cursor pointing to 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_cursor *mc, int ksize) +{ + MDB_page *mp = mc->mc_pg[mc->mc_top]; + indx_t indx = mc->mc_ki[mc->mc_top]; + unsigned int sz; + indx_t i, j, numkeys, ptr; + MDB_node *node; + char *base; + + DPRINTF(("delete node %u on %s page %"Z"u", indx, + IS_LEAF(mp) ? "leaf" : "branch", mdb_dbg_pgno(mp))); + numkeys = NUMKEYS(mp); + mdb_cassert(mc, indx < numkeys); + + if (IS_LEAF2(mp)) { + int x = numkeys - 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 = EVEN(sz); + + ptr = mp->mp_ptrs[indx]; + 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 + PAGEBASE; + 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 nsize, delta; + indx_t i, numkeys, ptr; + + node = NODEPTR(mp, indx); + sp = (MDB_page *)NODEDATA(node); + delta = SIZELEFT(sp); + xp = (MDB_page *)((char *)sp + delta); + + /* shift subpage upward */ + if (IS_LEAF2(sp)) { + nsize = NUMKEYS(sp) * sp->mp_pad; + if (nsize & 1) + return; /* do not make the node uneven-sized */ + memmove(METADATA(xp), METADATA(sp), nsize); + } else { + int i; + 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; + COPY_PGNO(xp->mp_pgno, mp->mp_pgno); + + nsize = NODEDSZ(node) - delta; + SETDSZ(node, nsize); + + /* 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 + PAGEBASE; + 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; + mx->mx_cursor.mc_dbflag = &mx->mx_dbflag; + mx->mx_cursor.mc_snum = 0; + mx->mx_cursor.mc_top = 0; + mx->mx_cursor.mc_flags = C_SUB; + mx->mx_dbx.md_name.mv_size = 0; + mx->mx_dbx.md_name.mv_data = NULL; + 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) { + memcpy(&mx->mx_db, NODEDATA(node), sizeof(MDB_db)); + mx->mx_cursor.mc_pg[0] = 0; + mx->mx_cursor.mc_snum = 0; + mx->mx_cursor.mc_top = 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); + COPY_PGNO(mx->mx_db.md_root, fp->mp_pgno); + mx->mx_cursor.mc_snum = 1; + mx->mx_cursor.mc_top = 0; + mx->mx_cursor.mc_flags = C_INITIALIZED|C_SUB; + 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 root page %"Z"u", mx->mx_cursor.mc_dbi, + mx->mx_db.md_root)); + mx->mx_dbflag = DB_VALID|DB_DIRTY; /* DB_DIRTY guides mdb_cursor_touch */ +#if UINT_MAX < SIZE_MAX + if (mx->mx_dbx.md_cmp == mdb_cmp_int && mx->mx_db.md_pad == sizeof(size_t)) + mx->mx_dbx.md_cmp = mdb_cmp_clong; +#endif +} + +/** 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_next = NULL; + mc->mc_backup = 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_top = 0; + mc->mc_pg[0] = 0; + mc->mc_flags = 0; + if (txn->mt_dbs[dbi].md_flags & MDB_DUPSORT) { + mdb_tassert(txn, mx != NULL); + mc->mc_xcursor = mx; + mdb_xcursor_init0(mc); + } else { + mc->mc_xcursor = NULL; + } + if (*mc->mc_dbflag & DB_STALE) { + mdb_page_search(mc, NULL, MDB_PS_ROOTONLY); + } +} + +int +mdb_cursor_open(MDB_txn *txn, MDB_dbi dbi, MDB_cursor **ret) +{ + MDB_cursor *mc; + size_t size = sizeof(MDB_cursor); + + if (!ret || !TXN_DBI_EXIST(txn, dbi)) + return EINVAL; + + if (txn->mt_flags & MDB_TXN_ERROR) + return MDB_BAD_TXN; + + /* Allow read access to the freelist */ + if (!dbi && !F_ISSET(txn->mt_flags, MDB_TXN_RDONLY)) + return EINVAL; + + if (txn->mt_dbs[dbi].md_flags & MDB_DUPSORT) + size += sizeof(MDB_xcursor); + + if ((mc = malloc(size)) != NULL) { + mdb_cursor_init(mc, txn, dbi, (MDB_xcursor *)(mc + 1)); + if (txn->mt_cursors) { + mc->mc_next = txn->mt_cursors[dbi]; + txn->mt_cursors[dbi] = mc; + mc->mc_flags |= C_UNTRACK; + } + } else { + return ENOMEM; + } + + *ret = mc; + + return MDB_SUCCESS; +} + +int +mdb_cursor_renew(MDB_txn *txn, MDB_cursor *mc) +{ + if (!mc || !TXN_DBI_EXIST(txn, mc->mc_dbi)) + return EINVAL; + + if ((mc->mc_flags & C_UNTRACK) || txn->mt_cursors) + return EINVAL; + + if (txn->mt_flags & MDB_TXN_ERROR) + return MDB_BAD_TXN; + + mdb_cursor_init(mc, txn, mc->mc_dbi, mc->mc_xcursor); + 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_xcursor == NULL) + return MDB_INCOMPATIBLE; + + if (mc->mc_txn->mt_flags & MDB_TXN_ERROR) + return MDB_BAD_TXN; + + if (!(mc->mc_flags & C_INITIALIZED)) + return EINVAL; + + if (!mc->mc_snum || (mc->mc_flags & C_EOF)) + return MDB_NOTFOUND; + + 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 && !mc->mc_backup) { + /* remove from txn, if tracked */ + if ((mc->mc_flags & C_UNTRACK) && 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; + } + 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) +{ + return mc->mc_dbi; +} + +/** Replace the key for a branch node with a new key. + * @param[in] mc Cursor pointing to 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_cursor *mc, MDB_val *key) +{ + MDB_page *mp; + MDB_node *node; + char *base; + size_t len; + int delta, ksize, oksize; + indx_t ptr, i, numkeys, indx; + DKBUF; + + indx = mc->mc_ki[mc->mc_top]; + mp = mc->mc_pg[mc->mc_top]; + node = NODEPTR(mp, indx); + ptr = mp->mp_ptrs[indx]; +#if MDB_DEBUG + { + MDB_val k2; + char kbuf2[DKBUF_MAXKEYSIZE*2+1]; + k2.mv_data = NODEKEY(node); + k2.mv_size = node->mn_ksize; + DPRINTF(("update key %u (ofs %u) [%s] to [%s] on page %"Z"u", + indx, ptr, + mdb_dkey(&k2, kbuf2), + DKEY(key), + mp->mp_pgno)); + } +#endif + + /* Sizes must be 2-byte aligned. */ + ksize = EVEN(key->mv_size); + oksize = EVEN(node->mn_ksize); + delta = ksize - oksize; + + /* Shift node contents if EVEN(key length) changed. */ + if (delta) { + if (delta > 0 && SIZELEFT(mp) < delta) { + pgno_t pgno; + /* not enough space left, do a delete and split */ + DPRINTF(("Not enough room, delta = %d, splitting...", delta)); + pgno = NODEPGNO(node); + mdb_node_del(mc, 0); + return mdb_page_split(mc, key, NULL, pgno, MDB_SPLIT_REPLACE); + } + + 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 + PAGEBASE; + len = ptr - mp->mp_upper + NODESIZE; + memmove(base - delta, base, len); + mp->mp_upper -= delta; + + node = NODEPTR(mp, indx); + } + + /* But even if no shift was needed, update ksize */ + if (node->mn_ksize != key->mv_size) + node->mn_ksize = key->mv_size; + + if (key->mv_size) + memcpy(NODEKEY(node), key->mv_data, key->mv_size); + + return MDB_SUCCESS; +} + +static void +mdb_cursor_copy(const MDB_cursor *csrc, MDB_cursor *cdst); + +/** Move a node from csrc to cdst. + */ +static int +mdb_node_move(MDB_cursor *csrc, MDB_cursor *cdst) +{ + MDB_node *srcnode; + MDB_val key, data; + pgno_t srcpg; + MDB_cursor mn; + int rc; + unsigned short flags; + + 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])) { + 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; + srcpg = 0; + flags = 0; + } else { + srcnode = NODEPTR(csrc->mc_pg[csrc->mc_top], csrc->mc_ki[csrc->mc_top]); + mdb_cassert(csrc, !((size_t)srcnode & 1)); + srcpg = NODEPGNO(srcnode); + flags = srcnode->mn_flags; + 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 */ + rc = mdb_page_search_lowest(csrc); + if (rc) + return rc; + if (IS_LEAF2(csrc->mc_pg[csrc->mc_top])) { + key.mv_size = csrc->mc_db->md_pad; + key.mv_data = LEAF2KEY(csrc->mc_pg[csrc->mc_top], 0, key.mv_size); + } else { + 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); + } + if (IS_BRANCH(cdst->mc_pg[cdst->mc_top]) && cdst->mc_ki[cdst->mc_top] == 0) { + unsigned int snum = cdst->mc_snum; + MDB_node *s2; + MDB_val bkey; + /* must find the lowest key below dst */ + mdb_cursor_copy(cdst, &mn); + rc = mdb_page_search_lowest(&mn); + if (rc) + return rc; + if (IS_LEAF2(mn.mc_pg[mn.mc_top])) { + bkey.mv_size = mn.mc_db->md_pad; + bkey.mv_data = LEAF2KEY(mn.mc_pg[mn.mc_top], 0, bkey.mv_size); + } else { + s2 = NODEPTR(mn.mc_pg[mn.mc_top], 0); + bkey.mv_size = NODEKSZ(s2); + bkey.mv_data = NODEKEY(s2); + } + mn.mc_snum = snum--; + mn.mc_top = snum; + mn.mc_ki[snum] = 0; + rc = mdb_update_key(&mn, &bkey); + if (rc) + return rc; + } + + DPRINTF(("moving %s node %u [%s] on page %"Z"u to node %u on page %"Z"u", + 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, srcpg, flags); + if (rc != MDB_SUCCESS) + return rc; + + /* Delete the node from the source page. + */ + mdb_node_del(csrc, 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]; + + for (m2 = csrc->mc_txn->mt_cursors[dbi]; m2; m2=m2->mc_next) { + if (csrc->mc_flags & C_SUB) + m3 = &m2->mc_xcursor->mx_cursor; + else + m3 = m2; + if (m3 == csrc) continue; + 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 %"Z"u to [%s]", + csrc->mc_pg[csrc->mc_top]->mp_pgno, DKEY(&key))); + mdb_cursor_copy(csrc, &mn); + mn.mc_snum--; + mn.mc_top--; + if ((rc = mdb_update_key(&mn, &key)) != MDB_SUCCESS) + return rc; + } + if (IS_BRANCH(csrc->mc_pg[csrc->mc_top])) { + MDB_val nullkey; + indx_t ix = csrc->mc_ki[csrc->mc_top]; + nullkey.mv_size = 0; + csrc->mc_ki[csrc->mc_top] = 0; + rc = mdb_update_key(csrc, &nullkey); + csrc->mc_ki[csrc->mc_top] = ix; + mdb_cassert(csrc, 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 %"Z"u to [%s]", + cdst->mc_pg[cdst->mc_top]->mp_pgno, DKEY(&key))); + mdb_cursor_copy(cdst, &mn); + mn.mc_snum--; + mn.mc_top--; + if ((rc = mdb_update_key(&mn, &key)) != MDB_SUCCESS) + return rc; + } + if (IS_BRANCH(cdst->mc_pg[cdst->mc_top])) { + MDB_val nullkey; + indx_t ix = cdst->mc_ki[cdst->mc_top]; + nullkey.mv_size = 0; + cdst->mc_ki[cdst->mc_top] = 0; + rc = mdb_update_key(cdst, &nullkey); + cdst->mc_ki[cdst->mc_top] = ix; + mdb_cassert(cdst, 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. + * @return 0 on success, non-zero on failure. + */ +static int +mdb_page_merge(MDB_cursor *csrc, MDB_cursor *cdst) +{ + MDB_page *psrc, *pdst; + MDB_node *srcnode; + MDB_val key, data; + unsigned nkeys; + int rc; + indx_t i, j; + + psrc = csrc->mc_pg[csrc->mc_top]; + pdst = cdst->mc_pg[cdst->mc_top]; + + DPRINTF(("merging page %"Z"u into %"Z"u", psrc->mp_pgno, pdst->mp_pgno)); + + mdb_cassert(csrc, csrc->mc_snum > 1); /* can't merge root page */ + mdb_cassert(csrc, 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(pdst); + if (IS_LEAF2(psrc)) { + key.mv_size = csrc->mc_db->md_pad; + key.mv_data = METADATA(psrc); + for (i = 0; i < NUMKEYS(psrc); 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(psrc); i++, j++) { + srcnode = NODEPTR(psrc, i); + if (i == 0 && IS_BRANCH(psrc)) { + MDB_cursor mn; + MDB_node *s2; + mdb_cursor_copy(csrc, &mn); + /* must find the lowest key below src */ + rc = mdb_page_search_lowest(&mn); + if (rc) + return rc; + if (IS_LEAF2(mn.mc_pg[mn.mc_top])) { + key.mv_size = mn.mc_db->md_pad; + key.mv_data = LEAF2KEY(mn.mc_pg[mn.mc_top], 0, key.mv_size); + } else { + s2 = NODEPTR(mn.mc_pg[mn.mc_top], 0); + key.mv_size = NODEKSZ(s2); + key.mv_data = NODEKEY(s2); + } + } else { + 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 %"Z"u now has %u keys (%.1f%% filled)", + pdst->mp_pgno, NUMKEYS(pdst), + (float)PAGEFILL(cdst->mc_txn->mt_env, pdst) / 10)); + + /* Unlink the src page from parent and add to free list. + */ + csrc->mc_top--; + mdb_node_del(csrc, 0); + if (csrc->mc_ki[csrc->mc_top] == 0) { + key.mv_size = 0; + rc = mdb_update_key(csrc, &key); + if (rc) { + csrc->mc_top++; + return rc; + } + } + csrc->mc_top++; + + psrc = csrc->mc_pg[csrc->mc_top]; + /* If not operating on FreeDB, allow this page to be reused + * in this txn. Otherwise just add to free list. + */ + rc = mdb_page_loose(csrc, psrc); + if (rc) + return rc; + if (IS_LEAF(psrc)) + 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; + + for (m2 = csrc->mc_txn->mt_cursors[dbi]; m2; m2=m2->mc_next) { + if (csrc->mc_flags & C_SUB) + m3 = &m2->mc_xcursor->mx_cursor; + else + m3 = m2; + if (m3 == csrc) continue; + if (m3->mc_snum < csrc->mc_snum) continue; + if (m3->mc_pg[csrc->mc_top] == psrc) { + m3->mc_pg[csrc->mc_top] = pdst; + m3->mc_ki[csrc->mc_top] += nkeys; + } + } + } + { + unsigned int snum = cdst->mc_snum; + uint16_t depth = cdst->mc_db->md_depth; + mdb_cursor_pop(cdst); + rc = mdb_rebalance(cdst); + /* Did the tree shrink? */ + if (depth > cdst->mc_db->md_depth) + snum--; + cdst->mc_snum = snum; + cdst->mc_top = snum-1; + } + return rc; +} + +/** 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, minkeys; + MDB_cursor mn; + indx_t oldki; + + minkeys = 1 + (IS_BRANCH(mc->mc_pg[mc->mc_top])); + DPRINTF(("rebalancing %s page %"Z"u (has %u keys, %.1f%% full)", + IS_LEAF(mc->mc_pg[mc->mc_top]) ? "leaf" : "branch", + mdb_dbg_pgno(mc->mc_pg[mc->mc_top]), 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 && + NUMKEYS(mc->mc_pg[mc->mc_top]) >= minkeys) { + DPRINTF(("no need to rebalance page %"Z"u, above fill threshold", + mdb_dbg_pgno(mc->mc_pg[mc->mc_top]))); + return MDB_SUCCESS; + } + + if (mc->mc_snum < 2) { + MDB_page *mp = mc->mc_pg[0]; + if (IS_SUBP(mp)) { + DPUTS("Can't rebalance a subpage, ignoring"); + return MDB_SUCCESS; + } + 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; + rc = mdb_midl_append(&mc->mc_txn->mt_free_pgs, mp->mp_pgno); + if (rc) + return rc; + /* Adjust cursors pointing to mp */ + mc->mc_snum = 0; + mc->mc_top = 0; + mc->mc_flags &= ~C_INITIALIZED; + { + MDB_cursor *m2, *m3; + MDB_dbi dbi = mc->mc_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_snum < mc->mc_snum) continue; + if (m3->mc_pg[0] == mp) { + m3->mc_snum = 0; + m3->mc_top = 0; + m3->mc_flags &= ~C_INITIALIZED; + } + } + } + } else if (IS_BRANCH(mp) && NUMKEYS(mp) == 1) { + int i; + DPUTS("collapsing root page!"); + rc = mdb_midl_append(&mc->mc_txn->mt_free_pgs, mp->mp_pgno); + if (rc) + return rc; + mc->mc_db->md_root = NODEPGNO(NODEPTR(mp, 0)); + rc = mdb_page_get(mc->mc_txn,mc->mc_db->md_root,&mc->mc_pg[0],NULL); + if (rc) + return rc; + mc->mc_db->md_depth--; + mc->mc_db->md_branch_pages--; + mc->mc_ki[0] = mc->mc_ki[1]; + for (i = 1; i<mc->mc_db->md_depth; i++) { + mc->mc_pg[i] = mc->mc_pg[i+1]; + mc->mc_ki[i] = mc->mc_ki[i+1]; + } + { + /* Adjust other cursors pointing to mp */ + MDB_cursor *m2, *m3; + MDB_dbi dbi = mc->mc_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 || m3->mc_snum < mc->mc_snum) continue; + if (m3->mc_pg[0] == mp) { + m3->mc_snum--; + m3->mc_top--; + for (i=0; i<m3->mc_snum; i++) { + m3->mc_pg[i] = m3->mc_pg[i+1]; + m3->mc_ki[i] = m3->mc_ki[i+1]; + } + } + } + } + } 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; + mdb_cassert(mc, 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; + + oldki = mc->mc_ki[mc->mc_top]; + 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]); + rc = mdb_page_get(mc->mc_txn,NODEPGNO(node),&mn.mc_pg[mn.mc_top],NULL); + if (rc) + 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]); + rc = mdb_page_get(mc->mc_txn,NODEPGNO(node),&mn.mc_pg[mn.mc_top],NULL); + if (rc) + 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 %"Z"u (%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 enough keys, + * move one key from it. Otherwise we should try to merge them. + * (A branch page must never have less than 2 keys.) + */ + minkeys = 1 + (IS_BRANCH(mn.mc_pg[mn.mc_top])); + if (PAGEFILL(mc->mc_txn->mt_env, mn.mc_pg[mn.mc_top]) >= FILL_THRESHOLD && NUMKEYS(mn.mc_pg[mn.mc_top]) > minkeys) { + rc = mdb_node_move(&mn, mc); + if (mc->mc_ki[ptop]) { + oldki++; + } + } else { + if (mc->mc_ki[ptop] == 0) { + rc = mdb_page_merge(&mn, mc); + } else { + oldki += NUMKEYS(mn.mc_pg[mn.mc_top]); + mn.mc_ki[mn.mc_top] += mc->mc_ki[mn.mc_top] + 1; + rc = mdb_page_merge(mc, &mn); + mdb_cursor_copy(&mn, mc); + } + mc->mc_flags &= ~C_EOF; + } + mc->mc_ki[mc->mc_top] = oldki; + return rc; +} + +/** Complete a delete operation started by #mdb_cursor_del(). */ +static int +mdb_cursor_del0(MDB_cursor *mc) +{ + int rc; + MDB_page *mp; + indx_t ki; + unsigned int nkeys; + + ki = mc->mc_ki[mc->mc_top]; + mdb_node_del(mc, mc->mc_db->md_pad); + mc->mc_db->md_entries--; + rc = mdb_rebalance(mc); + + if (rc == MDB_SUCCESS) { + MDB_cursor *m2, *m3; + MDB_dbi dbi = mc->mc_dbi; + + mp = mc->mc_pg[mc->mc_top]; + nkeys = NUMKEYS(mp); + + /* if mc points past last node in page, find next sibling */ + if (mc->mc_ki[mc->mc_top] >= nkeys) { + rc = mdb_cursor_sibling(mc, 1); + if (rc == MDB_NOTFOUND) { + mc->mc_flags |= C_EOF; + rc = MDB_SUCCESS; + } + } + + /* Adjust other cursors pointing to mp */ + for (m2 = mc->mc_txn->mt_cursors[dbi]; !rc && m2; m2=m2->mc_next) { + m3 = (mc->mc_flags & C_SUB) ? &m2->mc_xcursor->mx_cursor : m2; + if (! (m2->mc_flags & m3->mc_flags & C_INITIALIZED)) + continue; + if (m3 == mc || m3->mc_snum < mc->mc_snum) + continue; + if (m3->mc_pg[mc->mc_top] == mp) { + if (m3->mc_ki[mc->mc_top] >= ki) { + m3->mc_flags |= C_DEL; + if (m3->mc_ki[mc->mc_top] > ki) + m3->mc_ki[mc->mc_top]--; + else if (mc->mc_db->md_flags & MDB_DUPSORT) + m3->mc_xcursor->mx_cursor.mc_flags |= C_EOF; + } + if (m3->mc_ki[mc->mc_top] >= nkeys) { + rc = mdb_cursor_sibling(m3, 1); + if (rc == MDB_NOTFOUND) { + m3->mc_flags |= C_EOF; + rc = MDB_SUCCESS; + } + } + } + } + mc->mc_flags |= C_DEL; + } + + if (rc) + 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) +{ + if (!key || dbi == FREE_DBI || !TXN_DBI_EXIST(txn, dbi)) + return EINVAL; + + if (txn->mt_flags & (MDB_TXN_RDONLY|MDB_TXN_ERROR)) + return (txn->mt_flags & MDB_TXN_RDONLY) ? EACCES : MDB_BAD_TXN; + + if (!F_ISSET(txn->mt_dbs[dbi].md_flags, MDB_DUPSORT)) { + /* must ignore any data */ + data = NULL; + } + + return mdb_del0(txn, dbi, key, data, 0); +} + +static int +mdb_del0(MDB_txn *txn, MDB_dbi dbi, + MDB_val *key, MDB_val *data, unsigned flags) +{ + MDB_cursor mc; + MDB_xcursor mx; + MDB_cursor_op op; + MDB_val rdata, *xdata; + int rc, exact = 0; + DKBUF; + + DPRINTF(("====> delete db %u key [%s]", dbi, DKEY(key))); + + mdb_cursor_init(&mc, txn, dbi, &mx); + + if (data) { + op = MDB_GET_BOTH; + rdata = *data; + xdata = &rdata; + } else { + op = MDB_SET; + xdata = NULL; + flags |= MDB_NODUPDATA; + } + rc = mdb_cursor_set(&mc, key, xdata, op, &exact); + if (rc == 0) { + /* let mdb_page_split know about this cursor if needed: + * delete will trigger a rebalance; if it needs to move + * a node from one page to another, it will have to + * update the parent's separator key(s). If the new sepkey + * is larger than the current one, the parent page may + * run out of space, triggering a split. We need this + * cursor to be consistent until the end of the rebalance. + */ + mc.mc_flags |= C_UNTRACK; + mc.mc_next = txn->mt_cursors[dbi]; + txn->mt_cursors[dbi] = &mc; + rc = mdb_cursor_del(&mc, flags); + txn->mt_cursors[dbi] = mc.mc_next; + } + 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. + * @param[in] nflags The #NODE_ADD_FLAGS for the new 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, new_root = 0, did_split = 0; + indx_t newindx; + pgno_t pgno = 0; + int i, j, split_indx, nkeys, pmax; + MDB_env *env = mc->mc_txn->mt_env; + MDB_node *node; + MDB_val sepkey, rkey, xdata, *rdata = &xdata; + MDB_page *copy = NULL; + MDB_page *mp, *rp, *pp; + int ptop; + MDB_cursor mn; + DKBUF; + + mp = mc->mc_pg[mc->mc_top]; + newindx = mc->mc_ki[mc->mc_top]; + nkeys = NUMKEYS(mp); + + DPRINTF(("-----> splitting %s page %"Z"u and adding [%s] at index %i/%i", + IS_LEAF(mp) ? "leaf" : "branch", mp->mp_pgno, + DKEY(newkey), mc->mc_ki[mc->mc_top], nkeys)); + + /* Create a right sibling. */ + if ((rc = mdb_page_new(mc, mp->mp_flags, 1, &rp))) + return rc; + DPRINTF(("new right sibling: page %"Z"u", rp->mp_pgno)); + + if (mc->mc_snum < 2) { + if ((rc = mdb_page_new(mc, P_BRANCH, 1, &pp))) + goto done; + /* 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 = %"Z"u", 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--; + goto done; + } + mc->mc_snum = 2; + mc->mc_top = 1; + ptop = 0; + } else { + ptop = mc->mc_top-1; + DPRINTF(("parent branch page is %"Z"u", mc->mc_pg[ptop]->mp_pgno)); + } + + mc->mc_flags |= C_SPLITTING; + 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; + split_indx = newindx; + nkeys = 0; + } else { + + split_indx = (nkeys+1) / 2; + + if (IS_LEAF2(rp)) { + char *split, *ins; + int x; + unsigned int lsize, rsize, ksize; + /* Move half of the keys to the right sibling */ + 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; + } + } else { + int psize, nsize, k; + /* Maximum free space in an empty page */ + pmax = env->me_psize - PAGEHDRSZ; + if (IS_LEAF(mp)) + nsize = mdb_leaf_size(env, newkey, newdata); + else + nsize = mdb_branch_size(env, newkey); + nsize = EVEN(nsize); + + /* grab a page to hold a temporary copy */ + copy = mdb_page_malloc(mc->mc_txn, 1); + if (copy == NULL) { + rc = ENOMEM; + goto done; + } + copy->mp_pgno = mp->mp_pgno; + copy->mp_flags = mp->mp_flags; + copy->mp_lower = (PAGEHDRSZ-PAGEBASE); + copy->mp_upper = env->me_psize - PAGEBASE; + + /* prepare to insert */ + for (i=0, j=0; i<nkeys; i++) { + if (i == newindx) { + copy->mp_ptrs[j++] = 0; + } + copy->mp_ptrs[j++] = mp->mp_ptrs[i]; + } + + /* When items are relatively large the split point needs + * to be checked, because being off-by-one will make the + * difference between success or failure in mdb_node_add. + * + * It's also relevant if a page happens to be laid out + * such that one half of its nodes are all "small" and + * the other half of its nodes are "large." If the new + * item is also "large" and falls on the half with + * "large" nodes, it also may not fit. + * + * As a final tweak, if the new item goes on the last + * spot on the page (and thus, onto the new page), bias + * the split so the new page is emptier than the old page. + * This yields better packing during sequential inserts. + */ + if (nkeys < 20 || nsize > pmax/16 || newindx >= nkeys) { + /* Find split point */ + psize = 0; + if (newindx <= split_indx || newindx >= nkeys) { + i = 0; j = 1; + k = newindx >= nkeys ? nkeys : split_indx+2; + } else { + i = nkeys; j = -1; + k = split_indx-1; + } + for (; i!=k; i+=j) { + if (i == newindx) { + psize += nsize; + node = NULL; + } else { + node = (MDB_node *)((char *)mp + copy->mp_ptrs[i] + PAGEBASE); + psize += NODESIZE + NODEKSZ(node) + sizeof(indx_t); + if (IS_LEAF(mp)) { + if (F_ISSET(node->mn_flags, F_BIGDATA)) + psize += sizeof(pgno_t); + else + psize += NODEDSZ(node); + } + psize = EVEN(psize); + } + if (psize > pmax || i == k-j) { + split_indx = i + (j<0); + break; + } + } + } + if (split_indx == newindx) { + sepkey.mv_size = newkey->mv_size; + sepkey.mv_data = newkey->mv_data; + } else { + node = (MDB_node *)((char *)mp + copy->mp_ptrs[split_indx] + PAGEBASE); + sepkey.mv_size = node->mn_ksize; + sepkey.mv_data = NODEKEY(node); + } + } + } + + DPRINTF(("separator is %d [%s]", split_indx, DKEY(&sepkey))); + + /* Copy separator key to the parent. + */ + if (SIZELEFT(mn.mc_pg[ptop]) < mdb_branch_size(env, &sepkey)) { + mn.mc_snum--; + mn.mc_top--; + did_split = 1; + rc = mdb_page_split(&mn, &sepkey, NULL, rp->mp_pgno, 0); + if (rc) + goto done; + + /* root split? */ + if (mn.mc_snum == mc->mc_snum) { + mc->mc_pg[mc->mc_snum] = mc->mc_pg[mc->mc_top]; + mc->mc_ki[mc->mc_snum] = mc->mc_ki[mc->mc_top]; + mc->mc_pg[mc->mc_top] = mc->mc_pg[ptop]; + mc->mc_ki[mc->mc_top] = mc->mc_ki[ptop]; + mc->mc_snum++; + mc->mc_top++; + ptop++; + } + /* 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])) { + for (i=0; i<ptop; i++) { + mc->mc_pg[i] = mn.mc_pg[i]; + mc->mc_ki[i] = mn.mc_ki[i]; + } + mc->mc_pg[ptop] = mn.mc_pg[ptop]; + if (mn.mc_ki[ptop]) { + mc->mc_ki[ptop] = mn.mc_ki[ptop] - 1; + } else { + /* find right page's left sibling */ + mc->mc_ki[ptop] = mn.mc_ki[ptop]; + mdb_cursor_sibling(mc, 0); + } + } + } else { + mn.mc_top--; + rc = mdb_node_add(&mn, mn.mc_ki[ptop], &sepkey, NULL, rp->mp_pgno, 0); + mn.mc_top++; + } + mc->mc_flags ^= C_SPLITTING; + if (rc != MDB_SUCCESS) { + goto done; + } + if (nflags & MDB_APPEND) { + mc->mc_pg[mc->mc_top] = rp; + mc->mc_ki[mc->mc_top] = 0; + rc = mdb_node_add(mc, 0, newkey, newdata, newpgno, nflags); + if (rc) + goto done; + for (i=0; i<mc->mc_top; i++) + mc->mc_ki[i] = mn.mc_ki[i]; + } else if (!IS_LEAF2(mp)) { + /* Move nodes */ + mc->mc_pg[mc->mc_top] = rp; + i = split_indx; + j = 0; + do { + if (i == newindx) { + rkey.mv_data = newkey->mv_data; + rkey.mv_size = newkey->mv_size; + if (IS_LEAF(mp)) { + rdata = newdata; + } else + pgno = newpgno; + flags = nflags; + /* Update index for the new key. */ + mc->mc_ki[mc->mc_top] = j; + } else { + node = (MDB_node *)((char *)mp + copy->mp_ptrs[i] + PAGEBASE); + 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; + } + + 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); + if (rc) + goto done; + if (i == nkeys) { + i = 0; + j = 0; + mc->mc_pg[mc->mc_top] = copy; + } else { + i++; + j++; + } + } while (i != split_indx); + + 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), + env->me_psize - copy->mp_upper - PAGEBASE); + + /* 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); + } + } else { + mc->mc_pg[mc->mc_top] = rp; + mc->mc_ki[ptop]++; + /* Make sure mc_ki is still valid. + */ + if (mn.mc_pg[ptop] != mc->mc_pg[ptop] && + mc->mc_ki[ptop] >= NUMKEYS(mc->mc_pg[ptop])) { + for (i=0; i<=ptop; i++) { + mc->mc_pg[i] = mn.mc_pg[i]; + mc->mc_ki[i] = mn.mc_ki[i]; + } + } + } + } + + { + /* Adjust other cursors pointing to mp */ + MDB_cursor *m2, *m3; + MDB_dbi dbi = mc->mc_dbi; + int fixup = NUMKEYS(mp); + + 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 (!(m2->mc_flags & m3->mc_flags & C_INITIALIZED)) + continue; + if (m3->mc_flags & C_SPLITTING) + continue; + if (new_root) { + int k; + /* root split */ + for (k=m3->mc_top; k>=0; k--) { + m3->mc_ki[k+1] = m3->mc_ki[k]; + m3->mc_pg[k+1] = m3->mc_pg[k]; + } + if (m3->mc_ki[0] >= split_indx) { + m3->mc_ki[0] = 1; + } else { + m3->mc_ki[0] = 0; + } + m3->mc_pg[0] = mc->mc_pg[0]; + m3->mc_snum++; + m3->mc_top++; + } + if (m3->mc_top >= mc->mc_top && m3->mc_pg[mc->mc_top] == mp) { + if (m3->mc_ki[mc->mc_top] >= newindx && !(nflags & MDB_SPLIT_REPLACE)) + m3->mc_ki[mc->mc_top]++; + if (m3->mc_ki[mc->mc_top] >= fixup) { + m3->mc_pg[mc->mc_top] = rp; + m3->mc_ki[mc->mc_top] -= fixup; + m3->mc_ki[ptop] = mn.mc_ki[ptop]; + } + } else if (!did_split && m3->mc_top >= ptop && m3->mc_pg[ptop] == mc->mc_pg[ptop] && + m3->mc_ki[ptop] >= mc->mc_ki[ptop]) { + m3->mc_ki[ptop]++; + } + } + } + DPRINTF(("mp left: %d, rp left: %d", SIZELEFT(mp), SIZELEFT(rp))); + +done: + if (copy) /* tmp page */ + mdb_page_free(env, copy); + if (rc) + mc->mc_txn->mt_flags |= MDB_TXN_ERROR; + 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; + + if (!key || !data || dbi == FREE_DBI || !TXN_DBI_EXIST(txn, dbi)) + return EINVAL; + + if ((flags & (MDB_NOOVERWRITE|MDB_NODUPDATA|MDB_RESERVE|MDB_APPEND|MDB_APPENDDUP)) != flags) + return EINVAL; + + mdb_cursor_init(&mc, txn, dbi, &mx); + return mdb_cursor_put(&mc, key, data, flags); +} + +#ifndef MDB_WBUF +#define MDB_WBUF (1024*1024) +#endif + + /** State needed for a compacting copy. */ +typedef struct mdb_copy { + pthread_mutex_t mc_mutex; + pthread_cond_t mc_cond; + char *mc_wbuf[2]; + char *mc_over[2]; + MDB_env *mc_env; + MDB_txn *mc_txn; + int mc_wlen[2]; + int mc_olen[2]; + pgno_t mc_next_pgno; + HANDLE mc_fd; + int mc_status; + volatile int mc_new; + int mc_toggle; + +} mdb_copy; + + /** Dedicated writer thread for compacting copy. */ +static THREAD_RET ESECT +mdb_env_copythr(void *arg) +{ + mdb_copy *my = arg; + char *ptr; + int toggle = 0, wsize, rc; +#ifdef _WIN32 + DWORD len; +#define DO_WRITE(rc, fd, ptr, w2, len) rc = WriteFile(fd, ptr, w2, &len, NULL) +#else + int len; +#define DO_WRITE(rc, fd, ptr, w2, len) len = write(fd, ptr, w2); rc = (len >= 0) +#endif + + pthread_mutex_lock(&my->mc_mutex); + my->mc_new = 0; + pthread_cond_signal(&my->mc_cond); + for(;;) { + while (!my->mc_new) + pthread_cond_wait(&my->mc_cond, &my->mc_mutex); + if (my->mc_new < 0) { + my->mc_new = 0; + break; + } + my->mc_new = 0; + wsize = my->mc_wlen[toggle]; + ptr = my->mc_wbuf[toggle]; +again: + while (wsize > 0) { + DO_WRITE(rc, my->mc_fd, ptr, wsize, len); + if (!rc) { + rc = ErrCode(); + break; + } else if (len > 0) { + rc = MDB_SUCCESS; + ptr += len; + wsize -= len; + continue; + } else { + rc = EIO; + break; + } + } + if (rc) { + my->mc_status = rc; + break; + } + /* If there's an overflow page tail, write it too */ + if (my->mc_olen[toggle]) { + wsize = my->mc_olen[toggle]; + ptr = my->mc_over[toggle]; + my->mc_olen[toggle] = 0; + goto again; + } + my->mc_wlen[toggle] = 0; + toggle ^= 1; + pthread_cond_signal(&my->mc_cond); + } + pthread_cond_signal(&my->mc_cond); + pthread_mutex_unlock(&my->mc_mutex); + return (THREAD_RET)0; +#undef DO_WRITE +} + + /** Tell the writer thread there's a buffer ready to write */ +static int ESECT +mdb_env_cthr_toggle(mdb_copy *my, int st) +{ + int toggle = my->mc_toggle ^ 1; + pthread_mutex_lock(&my->mc_mutex); + if (my->mc_status) { + pthread_mutex_unlock(&my->mc_mutex); + return my->mc_status; + } + while (my->mc_new == 1) + pthread_cond_wait(&my->mc_cond, &my->mc_mutex); + my->mc_new = st; + my->mc_toggle = toggle; + pthread_cond_signal(&my->mc_cond); + pthread_mutex_unlock(&my->mc_mutex); + return 0; +} + + /** Depth-first tree traversal for compacting copy. */ +static int ESECT +mdb_env_cwalk(mdb_copy *my, pgno_t *pg, int flags) +{ + MDB_cursor mc; + MDB_txn *txn = my->mc_txn; + MDB_node *ni; + MDB_page *mo, *mp, *leaf; + char *buf, *ptr; + int rc, toggle; + unsigned int i; + + /* Empty DB, nothing to do */ + if (*pg == P_INVALID) + return MDB_SUCCESS; + + mc.mc_snum = 1; + mc.mc_top = 0; + mc.mc_txn = txn; + + rc = mdb_page_get(my->mc_txn, *pg, &mc.mc_pg[0], NULL); + if (rc) + return rc; + rc = mdb_page_search_root(&mc, NULL, MDB_PS_FIRST); + if (rc) + return rc; + + /* Make cursor pages writable */ + buf = ptr = malloc(my->mc_env->me_psize * mc.mc_snum); + if (buf == NULL) + return ENOMEM; + + for (i=0; i<mc.mc_top; i++) { + mdb_page_copy((MDB_page *)ptr, mc.mc_pg[i], my->mc_env->me_psize); + mc.mc_pg[i] = (MDB_page *)ptr; + ptr += my->mc_env->me_psize; + } + + /* This is writable space for a leaf page. Usually not needed. */ + leaf = (MDB_page *)ptr; + + toggle = my->mc_toggle; + while (mc.mc_snum > 0) { + unsigned n; + mp = mc.mc_pg[mc.mc_top]; + n = NUMKEYS(mp); + + if (IS_LEAF(mp)) { + if (!IS_LEAF2(mp) && !(flags & F_DUPDATA)) { + for (i=0; i<n; i++) { + ni = NODEPTR(mp, i); + if (ni->mn_flags & F_BIGDATA) { + MDB_page *omp; + pgno_t pg; + + /* Need writable leaf */ + if (mp != leaf) { + mc.mc_pg[mc.mc_top] = leaf; + mdb_page_copy(leaf, mp, my->mc_env->me_psize); + mp = leaf; + ni = NODEPTR(mp, i); + } + + memcpy(&pg, NODEDATA(ni), sizeof(pg)); + rc = mdb_page_get(txn, pg, &omp, NULL); + if (rc) + goto done; + if (my->mc_wlen[toggle] >= MDB_WBUF) { + rc = mdb_env_cthr_toggle(my, 1); + if (rc) + goto done; + toggle = my->mc_toggle; + } + mo = (MDB_page *)(my->mc_wbuf[toggle] + my->mc_wlen[toggle]); + memcpy(mo, omp, my->mc_env->me_psize); + mo->mp_pgno = my->mc_next_pgno; + my->mc_next_pgno += omp->mp_pages; + my->mc_wlen[toggle] += my->mc_env->me_psize; + if (omp->mp_pages > 1) { + my->mc_olen[toggle] = my->mc_env->me_psize * (omp->mp_pages - 1); + my->mc_over[toggle] = (char *)omp + my->mc_env->me_psize; + rc = mdb_env_cthr_toggle(my, 1); + if (rc) + goto done; + toggle = my->mc_toggle; + } + memcpy(NODEDATA(ni), &mo->mp_pgno, sizeof(pgno_t)); + } else if (ni->mn_flags & F_SUBDATA) { + MDB_db db; + + /* Need writable leaf */ + if (mp != leaf) { + mc.mc_pg[mc.mc_top] = leaf; + mdb_page_copy(leaf, mp, my->mc_env->me_psize); + mp = leaf; + ni = NODEPTR(mp, i); + } + + memcpy(&db, NODEDATA(ni), sizeof(db)); + my->mc_toggle = toggle; + rc = mdb_env_cwalk(my, &db.md_root, ni->mn_flags & F_DUPDATA); + if (rc) + goto done; + toggle = my->mc_toggle; + memcpy(NODEDATA(ni), &db, sizeof(db)); + } + } + } + } else { + mc.mc_ki[mc.mc_top]++; + if (mc.mc_ki[mc.mc_top] < n) { + pgno_t pg; +again: + ni = NODEPTR(mp, mc.mc_ki[mc.mc_top]); + pg = NODEPGNO(ni); + rc = mdb_page_get(txn, pg, &mp, NULL); + if (rc) + goto done; + mc.mc_top++; + mc.mc_snum++; + mc.mc_ki[mc.mc_top] = 0; + if (IS_BRANCH(mp)) { + /* Whenever we advance to a sibling branch page, + * we must proceed all the way down to its first leaf. + */ + mdb_page_copy(mc.mc_pg[mc.mc_top], mp, my->mc_env->me_psize); + goto again; + } else + mc.mc_pg[mc.mc_top] = mp; + continue; + } + } + if (my->mc_wlen[toggle] >= MDB_WBUF) { + rc = mdb_env_cthr_toggle(my, 1); + if (rc) + goto done; + toggle = my->mc_toggle; + } + mo = (MDB_page *)(my->mc_wbuf[toggle] + my->mc_wlen[toggle]); + mdb_page_copy(mo, mp, my->mc_env->me_psize); + mo->mp_pgno = my->mc_next_pgno++; + my->mc_wlen[toggle] += my->mc_env->me_psize; + if (mc.mc_top) { + /* Update parent if there is one */ + ni = NODEPTR(mc.mc_pg[mc.mc_top-1], mc.mc_ki[mc.mc_top-1]); + SETPGNO(ni, mo->mp_pgno); + mdb_cursor_pop(&mc); + } else { + /* Otherwise we're done */ + *pg = mo->mp_pgno; + break; + } + } +done: + free(buf); + return rc; +} + + /** Copy environment with compaction. */ +static int ESECT +mdb_env_copyfd1(MDB_env *env, HANDLE fd) +{ + MDB_meta *mm; + MDB_page *mp; + mdb_copy my; + MDB_txn *txn = NULL; + pthread_t thr; + int rc; + +#ifdef _WIN32 + my.mc_mutex = CreateMutex(NULL, FALSE, NULL); + my.mc_cond = CreateEvent(NULL, FALSE, FALSE, NULL); + my.mc_wbuf[0] = _aligned_malloc(MDB_WBUF*2, env->me_os_psize); + if (my.mc_wbuf[0] == NULL) + return errno; +#else + pthread_mutex_init(&my.mc_mutex, NULL); + pthread_cond_init(&my.mc_cond, NULL); +#ifdef HAVE_MEMALIGN + my.mc_wbuf[0] = memalign(env->me_os_psize, MDB_WBUF*2); + if (my.mc_wbuf[0] == NULL) + return errno; +#else + rc = posix_memalign((void **)&my.mc_wbuf[0], env->me_os_psize, MDB_WBUF*2); + if (rc) + return rc; +#endif +#endif + memset(my.mc_wbuf[0], 0, MDB_WBUF*2); + my.mc_wbuf[1] = my.mc_wbuf[0] + MDB_WBUF; + my.mc_wlen[0] = 0; + my.mc_wlen[1] = 0; + my.mc_olen[0] = 0; + my.mc_olen[1] = 0; + my.mc_next_pgno = 2; + my.mc_status = 0; + my.mc_new = 1; + my.mc_toggle = 0; + my.mc_env = env; + my.mc_fd = fd; + THREAD_CREATE(thr, mdb_env_copythr, &my); + + rc = mdb_txn_begin(env, NULL, MDB_RDONLY, &txn); + if (rc) + return rc; + + mp = (MDB_page *)my.mc_wbuf[0]; + memset(mp, 0, 2*env->me_psize); + mp->mp_pgno = 0; + mp->mp_flags = P_META; + mm = (MDB_meta *)METADATA(mp); + mdb_env_init_meta0(env, mm); + mm->mm_address = env->me_metas[0]->mm_address; + + mp = (MDB_page *)(my.mc_wbuf[0] + env->me_psize); + mp->mp_pgno = 1; + mp->mp_flags = P_META; + *(MDB_meta *)METADATA(mp) = *mm; + mm = (MDB_meta *)METADATA(mp); + + /* Count the number of free pages, subtract from lastpg to find + * number of active pages + */ + { + MDB_ID freecount = 0; + MDB_cursor mc; + MDB_val key, data; + mdb_cursor_init(&mc, txn, FREE_DBI, NULL); + while ((rc = mdb_cursor_get(&mc, &key, &data, MDB_NEXT)) == 0) + freecount += *(MDB_ID *)data.mv_data; + freecount += txn->mt_dbs[0].md_branch_pages + + txn->mt_dbs[0].md_leaf_pages + + txn->mt_dbs[0].md_overflow_pages; + + /* Set metapage 1 */ + mm->mm_last_pg = txn->mt_next_pgno - freecount - 1; + mm->mm_dbs[1] = txn->mt_dbs[1]; + if (mm->mm_last_pg > 1) { + mm->mm_dbs[1].md_root = mm->mm_last_pg; + mm->mm_txnid = 1; + } else { + mm->mm_dbs[1].md_root = P_INVALID; + } + } + my.mc_wlen[0] = env->me_psize * 2; + my.mc_txn = txn; + pthread_mutex_lock(&my.mc_mutex); + while(my.mc_new) + pthread_cond_wait(&my.mc_cond, &my.mc_mutex); + pthread_mutex_unlock(&my.mc_mutex); + rc = mdb_env_cwalk(&my, &txn->mt_dbs[1].md_root, 0); + if (rc == MDB_SUCCESS && my.mc_wlen[my.mc_toggle]) + rc = mdb_env_cthr_toggle(&my, 1); + mdb_env_cthr_toggle(&my, -1); + pthread_mutex_lock(&my.mc_mutex); + while(my.mc_new) + pthread_cond_wait(&my.mc_cond, &my.mc_mutex); + pthread_mutex_unlock(&my.mc_mutex); + THREAD_FINISH(thr); + + mdb_txn_abort(txn); +#ifdef _WIN32 + CloseHandle(my.mc_cond); + CloseHandle(my.mc_mutex); + _aligned_free(my.mc_wbuf[0]); +#else + pthread_cond_destroy(&my.mc_cond); + pthread_mutex_destroy(&my.mc_mutex); + free(my.mc_wbuf[0]); +#endif + return rc; +} + + /** Copy environment as-is. */ +static int ESECT +mdb_env_copyfd0(MDB_env *env, HANDLE fd) +{ + MDB_txn *txn = NULL; + mdb_mutex_t *wmutex = NULL; + int rc; + size_t wsize; + char *ptr; +#ifdef _WIN32 + DWORD len, w2; +#define DO_WRITE(rc, fd, ptr, w2, len) rc = WriteFile(fd, ptr, w2, &len, NULL) +#else + ssize_t len; + size_t w2; +#define DO_WRITE(rc, fd, ptr, w2, len) len = write(fd, ptr, w2); rc = (len >= 0) +#endif + + /* Do the lock/unlock of the reader mutex before starting the + * write txn. Otherwise other read txns could block writers. + */ + rc = mdb_txn_begin(env, NULL, MDB_RDONLY, &txn); + if (rc) + return rc; + + if (env->me_txns) { + /* We must start the actual read txn after blocking writers */ + mdb_txn_reset0(txn, "reset-stage1"); + + /* Temporarily block writers until we snapshot the meta pages */ + wmutex = MDB_MUTEX(env, w); + if (LOCK_MUTEX(rc, env, wmutex)) + goto leave; + + rc = mdb_txn_renew0(txn); + if (rc) { + UNLOCK_MUTEX(wmutex); + goto leave; + } + } + + wsize = env->me_psize * 2; + ptr = env->me_map; + w2 = wsize; + while (w2 > 0) { + DO_WRITE(rc, fd, ptr, w2, len); + if (!rc) { + rc = ErrCode(); + break; + } else if (len > 0) { + rc = MDB_SUCCESS; + ptr += len; + w2 -= len; + continue; + } else { + /* Non-blocking or async handles are not supported */ + rc = EIO; + break; + } + } + if (wmutex) + UNLOCK_MUTEX(wmutex); + + if (rc) + goto leave; + + w2 = txn->mt_next_pgno * env->me_psize; + { + size_t fsize = 0; + if ((rc = mdb_fsize(env->me_fd, &fsize))) + goto leave; + if (w2 > fsize) + w2 = fsize; + } + wsize = w2 - wsize; + while (wsize > 0) { + if (wsize > MAX_WRITE) + w2 = MAX_WRITE; + else + w2 = wsize; + DO_WRITE(rc, fd, ptr, w2, len); + if (!rc) { + rc = ErrCode(); + break; + } else if (len > 0) { + rc = MDB_SUCCESS; + ptr += len; + wsize -= len; + continue; + } else { + rc = EIO; + break; + } + } + +leave: + mdb_txn_abort(txn); + return rc; +} + +int ESECT +mdb_env_copyfd2(MDB_env *env, HANDLE fd, unsigned int flags) +{ + if (flags & MDB_CP_COMPACT) + return mdb_env_copyfd1(env, fd); + else + return mdb_env_copyfd0(env, fd); +} + +int ESECT +mdb_env_copyfd(MDB_env *env, HANDLE fd) +{ + return mdb_env_copyfd2(env, fd, 0); +} + +int ESECT +mdb_env_copy2(MDB_env *env, const char *path, unsigned int flags) +{ + int rc, len; + char *lpath; + HANDLE newfd = INVALID_HANDLE_VALUE; + + if (env->me_flags & MDB_NOSUBDIR) { + lpath = (char *)path; + } else { + len = strlen(path); + len += sizeof(DATANAME); + lpath = malloc(len); + if (!lpath) + return ENOMEM; + sprintf(lpath, "%s" DATANAME, path); + } + + /* The destination path must exist, but the destination file must not. + * We don't want the OS to cache the writes, since the source data is + * already in the OS cache. + */ +#ifdef _WIN32 + newfd = CreateFile(lpath, GENERIC_WRITE, 0, NULL, CREATE_NEW, + FILE_FLAG_NO_BUFFERING|FILE_FLAG_WRITE_THROUGH, NULL); +#else + newfd = open(lpath, O_WRONLY|O_CREAT|O_EXCL, 0666); +#endif + if (newfd == INVALID_HANDLE_VALUE) { + rc = ErrCode(); + goto leave; + } + + if (env->me_psize >= env->me_os_psize) { +#ifdef O_DIRECT + /* Set O_DIRECT if the file system supports it */ + if ((rc = fcntl(newfd, F_GETFL)) != -1) + (void) fcntl(newfd, F_SETFL, rc | O_DIRECT); +#endif +#ifdef F_NOCACHE /* __APPLE__ */ + rc = fcntl(newfd, F_NOCACHE, 1); + if (rc) { + rc = ErrCode(); + goto leave; + } +#endif + } + + rc = mdb_env_copyfd2(env, newfd, flags); + +leave: + if (!(env->me_flags & MDB_NOSUBDIR)) + free(lpath); + if (newfd != INVALID_HANDLE_VALUE) + if (close(newfd) < 0 && rc == MDB_SUCCESS) + rc = ErrCode(); + + return rc; +} + +int ESECT +mdb_env_copy(MDB_env *env, const char *path) +{ + return mdb_env_copy2(env, path, 0); +} + +int ESECT +mdb_env_set_flags(MDB_env *env, unsigned int flag, int onoff) +{ + if (flag & (env->me_map ? ~CHANGEABLE : ~(CHANGEABLE|CHANGELESS))) + return EINVAL; + if (onoff) + env->me_flags |= flag; + else + env->me_flags &= ~flag; + return MDB_SUCCESS; +} + +int ESECT +mdb_env_get_flags(MDB_env *env, unsigned int *arg) +{ + if (!env || !arg) + return EINVAL; + + *arg = env->me_flags; + return MDB_SUCCESS; +} + +int ESECT +mdb_env_set_userctx(MDB_env *env, void *ctx) +{ + if (!env) + return EINVAL; + env->me_userctx = ctx; + return MDB_SUCCESS; +} + +void * ESECT +mdb_env_get_userctx(MDB_env *env) +{ + return env ? env->me_userctx : NULL; +} + +int ESECT +mdb_env_set_assert(MDB_env *env, MDB_assert_func *func) +{ + if (!env) + return EINVAL; +#ifndef NDEBUG + env->me_assert_func = func; +#endif + return MDB_SUCCESS; +} + +int ESECT +mdb_env_get_path(MDB_env *env, const char **arg) +{ + if (!env || !arg) + return EINVAL; + + *arg = env->me_path; + return MDB_SUCCESS; +} + +int ESECT +mdb_env_get_fd(MDB_env *env, mdb_filehandle_t *arg) +{ + if (!env || !arg) + return EINVAL; + + *arg = env->me_fd; + 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 ESECT +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 ESECT +mdb_env_stat(MDB_env *env, MDB_stat *arg) +{ + int toggle; + + if (env == NULL || arg == NULL) + return EINVAL; + + toggle = mdb_env_pick_meta(env); + + return mdb_stat0(env, &env->me_metas[toggle]->mm_dbs[MAIN_DBI], arg); +} + +int ESECT +mdb_env_info(MDB_env *env, MDB_envinfo *arg) +{ + int toggle; + + if (env == NULL || arg == NULL) + return EINVAL; + + toggle = mdb_env_pick_meta(env); + arg->me_mapaddr = env->me_metas[toggle]->mm_address; + arg->me_mapsize = env->me_mapsize; + arg->me_maxreaders = env->me_maxreaders; + + /* me_numreaders may be zero if this process never used any readers. Use + * the shared numreader count if it exists. + */ + arg->me_numreaders = env->me_txns ? env->me_txns->mti_numreaders : env->me_numreaders; + + arg->me_last_pgno = env->me_metas[toggle]->mm_last_pg; + arg->me_last_txnid = env->me_metas[toggle]->mm_txnid; + return MDB_SUCCESS; +} + +/** 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_dbi_open() + */ +static void +mdb_default_cmp(MDB_txn *txn, MDB_dbi dbi) +{ + uint16_t f = txn->mt_dbs[dbi].md_flags; + + txn->mt_dbxs[dbi].md_cmp = + (f & MDB_REVERSEKEY) ? mdb_cmp_memnr : + (f & MDB_INTEGERKEY) ? mdb_cmp_cint : mdb_cmp_memn; + + txn->mt_dbxs[dbi].md_dcmp = + !(f & MDB_DUPSORT) ? 0 : + ((f & MDB_INTEGERDUP) + ? ((f & MDB_DUPFIXED) ? mdb_cmp_int : mdb_cmp_cint) + : ((f & MDB_REVERSEDUP) ? mdb_cmp_memnr : mdb_cmp_memn)); +} + +int mdb_dbi_open(MDB_txn *txn, const char *name, unsigned int flags, MDB_dbi *dbi) +{ + MDB_val key, data; + MDB_dbi i; + MDB_cursor mc; + MDB_db dummy; + int rc, dbflag, exact; + unsigned int unused = 0, seq; + size_t len; + + if (txn->mt_dbxs[FREE_DBI].md_cmp == NULL) { + mdb_default_cmp(txn, FREE_DBI); + } + + if ((flags & VALID_FLAGS) != flags) + return EINVAL; + if (txn->mt_flags & MDB_TXN_ERROR) + return MDB_BAD_TXN; + + /* main DB? */ + if (!name) { + *dbi = MAIN_DBI; + if (flags & PERSISTENT_FLAGS) { + uint16_t f2 = flags & PERSISTENT_FLAGS; + /* make sure flag changes get committed */ + if ((txn->mt_dbs[MAIN_DBI].md_flags | f2) != txn->mt_dbs[MAIN_DBI].md_flags) { + txn->mt_dbs[MAIN_DBI].md_flags |= f2; + txn->mt_flags |= MDB_TXN_DIRTY; + } + } + 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 (!txn->mt_dbxs[i].md_name.mv_size) { + /* Remember this free slot */ + if (!unused) unused = i; + continue; + } + 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 no free slot and max hit, fail */ + if (!unused && txn->mt_numdbs >= txn->mt_env->me_maxdbs) + return MDB_DBS_FULL; + + /* Cannot mix named databases with some mainDB flags */ + if (txn->mt_dbs[MAIN_DBI].md_flags & (MDB_DUPSORT|MDB_INTEGERKEY)) + return (flags & MDB_CREATE) ? MDB_INCOMPATIBLE : MDB_NOTFOUND; + + /* Find the DB info */ + dbflag = DB_NEW|DB_VALID; + 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 MDB_INCOMPATIBLE; + } else if (rc == MDB_NOTFOUND && (flags & MDB_CREATE)) { + /* Create if requested */ + data.mv_size = sizeof(MDB_db); + data.mv_data = &dummy; + memset(&dummy, 0, sizeof(dummy)); + dummy.md_root = P_INVALID; + dummy.md_flags = flags & PERSISTENT_FLAGS; + rc = mdb_cursor_put(&mc, &key, &data, F_SUBDATA); + dbflag |= DB_DIRTY; + } + + /* OK, got info, add to table */ + if (rc == MDB_SUCCESS) { + unsigned int slot = unused ? unused : txn->mt_numdbs; + txn->mt_dbxs[slot].md_name.mv_data = strdup(name); + txn->mt_dbxs[slot].md_name.mv_size = len; + txn->mt_dbxs[slot].md_rel = NULL; + txn->mt_dbflags[slot] = dbflag; + /* txn-> and env-> are the same in read txns, use + * tmp variable to avoid undefined assignment + */ + seq = ++txn->mt_env->me_dbiseqs[slot]; + txn->mt_dbiseqs[slot] = seq; + + memcpy(&txn->mt_dbs[slot], data.mv_data, sizeof(MDB_db)); + *dbi = slot; + mdb_default_cmp(txn, slot); + if (!unused) { + txn->mt_numdbs++; + } + } + + return rc; +} + +int mdb_stat(MDB_txn *txn, MDB_dbi dbi, MDB_stat *arg) +{ + if (!arg || !TXN_DBI_EXIST(txn, dbi)) + return EINVAL; + + if (txn->mt_flags & MDB_TXN_ERROR) + return MDB_BAD_TXN; + + if (txn->mt_dbflags[dbi] & DB_STALE) { + MDB_cursor mc; + MDB_xcursor mx; + /* Stale, must read the DB's root. cursor_init does it for us. */ + mdb_cursor_init(&mc, txn, dbi, &mx); + } + return mdb_stat0(txn->mt_env, &txn->mt_dbs[dbi], arg); +} + +void mdb_dbi_close(MDB_env *env, MDB_dbi dbi) +{ + char *ptr; + if (dbi <= MAIN_DBI || dbi >= env->me_maxdbs) + return; + ptr = env->me_dbxs[dbi].md_name.mv_data; + /* If there was no name, this was already closed */ + if (ptr) { + env->me_dbxs[dbi].md_name.mv_data = NULL; + env->me_dbxs[dbi].md_name.mv_size = 0; + env->me_dbflags[dbi] = 0; + env->me_dbiseqs[dbi]++; + free(ptr); + } +} + +int mdb_dbi_flags(MDB_txn *txn, MDB_dbi dbi, unsigned int *flags) +{ + /* We could return the flags for the FREE_DBI too but what's the point? */ + if (dbi == FREE_DBI || !TXN_DBI_EXIST(txn, dbi)) + return EINVAL; + *flags = txn->mt_dbs[dbi].md_flags & PERSISTENT_FLAGS; + return MDB_SUCCESS; +} + +/** 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, MDB_PS_FIRST); + if (rc == MDB_SUCCESS) { + MDB_txn *txn = mc->mc_txn; + MDB_node *ni; + MDB_cursor mx; + unsigned int i; + + /* LEAF2 pages have no nodes, cannot have sub-DBs */ + if (IS_LEAF2(mc->mc_pg[mc->mc_top])) + mdb_cursor_pop(mc); + + mdb_cursor_copy(mc, &mx); + while (mc->mc_snum > 0) { + MDB_page *mp = mc->mc_pg[mc->mc_top]; + unsigned n = NUMKEYS(mp); + if (IS_LEAF(mp)) { + for (i=0; i<n; i++) { + ni = NODEPTR(mp, i); + if (ni->mn_flags & F_BIGDATA) { + MDB_page *omp; + pgno_t pg; + memcpy(&pg, NODEDATA(ni), sizeof(pg)); + rc = mdb_page_get(txn, pg, &omp, NULL); + if (rc != 0) + goto done; + mdb_cassert(mc, IS_OVERFLOW(omp)); + rc = mdb_midl_append_range(&txn->mt_free_pgs, + pg, omp->mp_pages); + if (rc) + goto done; + } else if (subs && (ni->mn_flags & F_SUBDATA)) { + mdb_xcursor_init1(mc, ni); + rc = mdb_drop0(&mc->mc_xcursor->mx_cursor, 0); + if (rc) + goto done; + } + } + } else { + if ((rc = mdb_midl_need(&txn->mt_free_pgs, n)) != 0) + goto done; + for (i=0; i<n; i++) { + pgno_t pg; + ni = NODEPTR(mp, i); + pg = NODEPGNO(ni); + /* free it */ + mdb_midl_xappend(txn->mt_free_pgs, pg); + } + } + if (!mc->mc_top) + break; + mc->mc_ki[mc->mc_top] = i; + rc = mdb_cursor_sibling(mc, 1); + if (rc) { + if (rc != MDB_NOTFOUND) + goto done; + /* no more siblings, go back to beginning + * of previous level. + */ + mdb_cursor_pop(mc); + mc->mc_ki[0] = 0; + for (i=1; i<mc->mc_snum; i++) { + mc->mc_ki[i] = 0; + mc->mc_pg[i] = mx.mc_pg[i]; + } + } + } + /* free it */ + rc = mdb_midl_append(&txn->mt_free_pgs, mc->mc_db->md_root); +done: + if (rc) + txn->mt_flags |= MDB_TXN_ERROR; + } else if (rc == MDB_NOTFOUND) { + rc = MDB_SUCCESS; + } + return rc; +} + +int mdb_drop(MDB_txn *txn, MDB_dbi dbi, int del) +{ + MDB_cursor *mc, *m2; + int rc; + + if ((unsigned)del > 1 || dbi == FREE_DBI || !TXN_DBI_EXIST(txn, dbi)) + return EINVAL; + + if (F_ISSET(txn->mt_flags, MDB_TXN_RDONLY)) + return EACCES; + + if (dbi > MAIN_DBI && TXN_DBI_CHANGED(txn, dbi)) + return MDB_BAD_DBI; + + rc = mdb_cursor_open(txn, dbi, &mc); + if (rc) + return rc; + + rc = mdb_drop0(mc, mc->mc_db->md_flags & MDB_DUPSORT); + /* Invalidate the dropped DB's cursors */ + for (m2 = txn->mt_cursors[dbi]; m2; m2 = m2->mc_next) + m2->mc_flags &= ~(C_INITIALIZED|C_EOF); + if (rc) + goto leave; + + /* Can't delete the main DB */ + if (del && dbi > MAIN_DBI) { + rc = mdb_del0(txn, MAIN_DBI, &mc->mc_dbx->md_name, NULL, 0); + if (!rc) { + txn->mt_dbflags[dbi] = DB_STALE; + mdb_dbi_close(txn->mt_env, dbi); + } else { + txn->mt_flags |= MDB_TXN_ERROR; + } + } else { + /* reset the DB record, mark it dirty */ + 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; + + txn->mt_flags |= MDB_TXN_DIRTY; + } +leave: + mdb_cursor_close(mc); + return rc; +} + +int mdb_set_compare(MDB_txn *txn, MDB_dbi dbi, MDB_cmp_func *cmp) +{ + if (dbi == FREE_DBI || !TXN_DBI_EXIST(txn, dbi)) + 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 (dbi == FREE_DBI || !TXN_DBI_EXIST(txn, dbi)) + 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 (dbi == FREE_DBI || !TXN_DBI_EXIST(txn, dbi)) + 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 (dbi == FREE_DBI || !TXN_DBI_EXIST(txn, dbi)) + return EINVAL; + + txn->mt_dbxs[dbi].md_relctx = ctx; + return MDB_SUCCESS; +} + +int ESECT +mdb_env_get_maxkeysize(MDB_env *env) +{ + return ENV_MAXKEY(env); +} + +int ESECT +mdb_reader_list(MDB_env *env, MDB_msg_func *func, void *ctx) +{ + unsigned int i, rdrs; + MDB_reader *mr; + char buf[64]; + int rc = 0, first = 1; + + if (!env || !func) + return -1; + if (!env->me_txns) { + return func("(no reader locks)\n", ctx); + } + rdrs = env->me_txns->mti_numreaders; + mr = env->me_txns->mti_readers; + for (i=0; i<rdrs; i++) { + if (mr[i].mr_pid) { + txnid_t txnid = mr[i].mr_txnid; + sprintf(buf, txnid == (txnid_t)-1 ? + "%10d %"Z"x -\n" : "%10d %"Z"x %"Z"u\n", + (int)mr[i].mr_pid, (size_t)mr[i].mr_tid, txnid); + if (first) { + first = 0; + rc = func(" pid thread txnid\n", ctx); + if (rc < 0) + break; + } + rc = func(buf, ctx); + if (rc < 0) + break; + } + } + if (first) { + rc = func("(no active readers)\n", ctx); + } + return rc; +} + +/** Insert pid into list if not already present. + * return -1 if already present. + */ +static int ESECT +mdb_pid_insert(MDB_PID_T *ids, MDB_PID_T pid) +{ + /* binary search of pid in list */ + 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 = pid - ids[cursor]; + + if( val < 0 ) { + n = pivot; + + } else if ( val > 0 ) { + base = cursor; + n -= pivot + 1; + + } else { + /* found, so it's a duplicate */ + return -1; + } + } + + if( val > 0 ) { + ++cursor; + } + ids[0]++; + for (n = ids[0]; n > cursor; n--) + ids[n] = ids[n-1]; + ids[n] = pid; + return 0; +} + +int ESECT +mdb_reader_check(MDB_env *env, int *dead) +{ + if (!env) + return EINVAL; + if (dead) + *dead = 0; + return env->me_txns ? mdb_reader_check0(env, 0, dead) : MDB_SUCCESS; +} + +/** As #mdb_reader_check(). rlocked = <caller locked the reader mutex>. */ +static int mdb_reader_check0(MDB_env *env, int rlocked, int *dead) +{ + mdb_mutex_t *rmutex = rlocked ? NULL : MDB_MUTEX(env, r); + unsigned int i, j, rdrs; + MDB_reader *mr; + MDB_PID_T *pids, pid; + int rc = MDB_SUCCESS, count = 0; + + rdrs = env->me_txns->mti_numreaders; + pids = malloc((rdrs+1) * sizeof(MDB_PID_T)); + if (!pids) + return ENOMEM; + pids[0] = 0; + mr = env->me_txns->mti_readers; + for (i=0; i<rdrs; i++) { + pid = mr[i].mr_pid; + if (pid && pid != env->me_pid) { + if (mdb_pid_insert(pids, pid) == 0) { + if (!mdb_reader_pid(env, Pidcheck, pid)) { + /* Stale reader found */ + j = i; + if (rmutex) { + if ((rc = LOCK_MUTEX0(rmutex)) != 0) { + if ((rc = mdb_mutex_failed(env, rmutex, rc))) + break; + rdrs = 0; /* the above checked all readers */ + } else { + /* Recheck, a new process may have reused pid */ + if (mdb_reader_pid(env, Pidcheck, pid)) + j = rdrs; + } + } + for (; j<rdrs; j++) + if (mr[j].mr_pid == pid) { + DPRINTF(("clear stale reader pid %u txn %"Z"d", + (unsigned) pid, mr[j].mr_txnid)); + mr[j].mr_pid = 0; + count++; + } + if (rmutex) + UNLOCK_MUTEX(rmutex); + } + } + } + } + free(pids); + if (dead) + *dead = count; + return rc; +} + +#ifdef MDB_ROBUST_SUPPORTED +/** Handle #LOCK_MUTEX0() failure. + * Try to repair the lock file if the mutex owner died. + * @param[in] env the environment handle + * @param[in] mutex LOCK_MUTEX0() mutex + * @param[in] rc LOCK_MUTEX0() error (nonzero) + * @return 0 on success with the mutex locked, or an error code on failure. + */ +static int mdb_mutex_failed(MDB_env *env, mdb_mutex_t *mutex, int rc) +{ + int toggle, rlocked, rc2; + + if (rc == MDB_OWNERDEAD) { + /* We own the mutex. Clean up after dead previous owner. */ + rc = MDB_SUCCESS; + rlocked = (mutex == MDB_MUTEX(env, r)); + if (!rlocked) { + /* Keep mti_txnid updated, otherwise next writer can + * overwrite data which latest meta page refers to. + */ + toggle = mdb_env_pick_meta(env); + env->me_txns->mti_txnid = env->me_metas[toggle]->mm_txnid; + /* env is hosed if the dead thread was ours */ + if (env->me_txn) { + env->me_flags |= MDB_FATAL_ERROR; + env->me_txn = NULL; + rc = MDB_PANIC; + } + } + DPRINTF(("%cmutex owner died, %s", (rlocked ? 'r' : 'w'), + (rc ? "this process' env is hosed" : "recovering"))); + rc2 = mdb_reader_check0(env, rlocked, NULL); + if (rc2 == 0) + rc2 = mdb_mutex_consistent(mutex); + if (rc || (rc = rc2)) { + DPRINTF(("LOCK_MUTEX recovery failed, %s", mdb_strerror(rc))); + UNLOCK_MUTEX(mutex); + } + } else { +#ifdef _WIN32 + rc = ErrCode(); +#endif + DPRINTF(("LOCK_MUTEX failed, %s", mdb_strerror(rc))); + } + + return rc; +} +#endif /* MDB_ROBUST_SUPPORTED */ +/** @} */ diff --git a/db_drivers/liblmdb/mdb_copy.1 b/db_drivers/liblmdb/mdb_copy.1 new file mode 100644 index 000000000..094b26056 --- /dev/null +++ b/db_drivers/liblmdb/mdb_copy.1 @@ -0,0 +1,54 @@ +.TH MDB_COPY 1 "2014/06/20" "LMDB 0.9.14" +.\" Copyright 2012-2014 Howard Chu, Symas Corp. All Rights Reserved. +.\" Copying restrictions apply. See COPYRIGHT/LICENSE. +.SH NAME +mdb_copy \- LMDB environment copy tool +.SH SYNOPSIS +.B mdb_copy +[\c +.BR \-V ] +[\c +.BR \-c ] +[\c +.BR \-n ] +.B srcpath +[\c +.BR dstpath ] +.SH DESCRIPTION +The +.B mdb_copy +utility copies an LMDB environment. The environment can +be copied regardless of whether it is currently in use. +No lockfile is created, since it gets recreated at need. + +If +.I dstpath +is specified it must be the path of an empty directory +for storing the backup. Otherwise, the backup will be +written to stdout. + +.SH OPTIONS +.TP +.BR \-V +Write the library version number to the standard output, and exit. +.TP +.BR \-c +Compact while copying. Only current data pages will be copied; freed +or unused pages will be omitted from the copy. This option will +slow down the backup process as it is more CPU-intensive. +.TP +.BR \-n +Open LDMB environment(s) which do not use subdirectories. + +.SH DIAGNOSTICS +Exit status is zero if no errors occur. +Errors result in a non-zero exit status and +a diagnostic message being written to standard error. +.SH CAVEATS +This utility can trigger significant file size growth if run +in parallel with write transactions, because pages which they +free during copying cannot be reused until the copy is done. +.SH "SEE ALSO" +.BR mdb_stat (1) +.SH AUTHOR +Howard Chu of Symas Corporation <http://www.symas.com> diff --git a/db_drivers/liblmdb/mdb_copy.c b/db_drivers/liblmdb/mdb_copy.c new file mode 100644 index 000000000..af0a94111 --- /dev/null +++ b/db_drivers/liblmdb/mdb_copy.c @@ -0,0 +1,82 @@ +/* mdb_copy.c - memory-mapped database backup tool */ +/* + * Copyright 2012 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>. + */ +#ifdef _WIN32 +#include <windows.h> +#define MDB_STDOUT GetStdHandle(STD_OUTPUT_HANDLE) +#else +#define MDB_STDOUT 1 +#endif +#include <stdio.h> +#include <stdlib.h> +#include <signal.h> +#include "lmdb.h" + +static void +sighandle(int sig) +{ +} + +int main(int argc,char * argv[]) +{ + int rc; + MDB_env *env; + const char *progname = argv[0], *act; + unsigned flags = MDB_RDONLY; + unsigned cpflags = 0; + + for (; argc > 1 && argv[1][0] == '-'; argc--, argv++) { + if (argv[1][1] == 'n' && argv[1][2] == '\0') + flags |= MDB_NOSUBDIR; + else if (argv[1][1] == 'c' && argv[1][2] == '\0') + cpflags |= MDB_CP_COMPACT; + else if (argv[1][1] == 'V' && argv[1][2] == '\0') { + printf("%s\n", MDB_VERSION_STRING); + exit(0); + } else + argc = 0; + } + + if (argc<2 || argc>3) { + fprintf(stderr, "usage: %s [-V] [-c] [-n] srcpath [dstpath]\n", progname); + exit(EXIT_FAILURE); + } + +#ifdef SIGPIPE + signal(SIGPIPE, sighandle); +#endif +#ifdef SIGHUP + signal(SIGHUP, sighandle); +#endif + signal(SIGINT, sighandle); + signal(SIGTERM, sighandle); + + act = "opening environment"; + rc = mdb_env_create(&env); + if (rc == MDB_SUCCESS) { + rc = mdb_env_open(env, argv[1], flags, 0600); + } + if (rc == MDB_SUCCESS) { + act = "copying"; + if (argc == 2) + rc = mdb_env_copyfd2(env, MDB_STDOUT, cpflags); + else + rc = mdb_env_copy2(env, argv[2], cpflags); + } + if (rc) + fprintf(stderr, "%s: %s failed, error %d (%s)\n", + progname, act, rc, mdb_strerror(rc)); + mdb_env_close(env); + + return rc ? EXIT_FAILURE : EXIT_SUCCESS; +} diff --git a/db_drivers/liblmdb/mdb_dump.1 b/db_drivers/liblmdb/mdb_dump.1 new file mode 100644 index 000000000..6fcc93081 --- /dev/null +++ b/db_drivers/liblmdb/mdb_dump.1 @@ -0,0 +1,75 @@ +.TH MDB_DUMP 1 "2014/06/20" "LMDB 0.9.14" +.\" Copyright 2014 Howard Chu, Symas Corp. All Rights Reserved. +.\" Copying restrictions apply. See COPYRIGHT/LICENSE. +.SH NAME +mdb_dump \- LMDB environment export tool +.SH SYNOPSIS +.B mdb_dump +.BR \ envpath +[\c +.BR \-V ] +[\c +.BI \-f \ file\fR] +[\c +.BR \-l ] +[\c +.BR \-n ] +[\c +.BR \-p ] +[\c +.BR \-a \ | +.BI \-s \ subdb\fR] +.SH DESCRIPTION +The +.B mdb_dump +utility reads a database and writes its contents to the +standard output using a portable flat-text format +understood by the +.BR mdb_load (1) +utility. +.SH OPTIONS +.TP +.BR \-V +Write the library version number to the standard output, and exit. +.TP +.BR \-f \ file +Write to the specified file instead of to the standard output. +.TP +.BR \-l +List the databases stored in the environment. Just the +names will be listed, no data will be output. +.TP +.BR \-n +Dump an LMDB database which does not use subdirectories. +.TP +.BR \-p +If characters in either the key or data items are printing characters (as +defined by isprint(3)), output them directly. This option permits users to +use standard text editors and tools to modify the contents of databases. + +Note: different systems may have different notions about what characters +are considered printing characters, and databases dumped in this manner may +be less portable to external systems. +.TP +.BR \-a +Dump all of the subdatabases in the environment. +.TP +.BR \-s \ subdb +Dump a specific subdatabase. If no database is specified, only the main database is dumped. +.SH DIAGNOSTICS +Exit status is zero if no errors occur. +Errors result in a non-zero exit status and +a diagnostic message being written to standard error. + +Dumping and reloading databases that use user-defined comparison functions +will result in new databases that use the default comparison functions. +\fBIn this case it is quite likely that the reloaded database will be +damaged beyond repair permitting neither record storage nor retrieval.\fP + +The only available workaround is to modify the source for the +.BR mdb_load (1) +utility to load the database using the correct comparison functions. +.SH "SEE ALSO" +.BR mdb_load (1) +.SH AUTHOR +Howard Chu of Symas Corporation <http://www.symas.com> diff --git a/db_drivers/liblmdb/mdb_dump.c b/db_drivers/liblmdb/mdb_dump.c new file mode 100644 index 000000000..0eb85fd20 --- /dev/null +++ b/db_drivers/liblmdb/mdb_dump.c @@ -0,0 +1,317 @@ +/* mdb_dump.c - memory-mapped database dump tool */ +/* + * Copyright 2011-2014 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 <errno.h> +#include <stdlib.h> +#include <string.h> +#include <ctype.h> +#include <unistd.h> +#include <signal.h> +#include "lmdb.h" + +#ifdef _WIN32 +#define Z "I" +#else +#define Z "z" +#endif + +#define PRINT 1 +static int mode; + +typedef struct flagbit { + int bit; + char *name; +} flagbit; + +flagbit dbflags[] = { + { MDB_REVERSEKEY, "reversekey" }, + { MDB_DUPSORT, "dupsort" }, + { MDB_INTEGERKEY, "integerkey" }, + { MDB_DUPFIXED, "dupfixed" }, + { MDB_INTEGERDUP, "integerdup" }, + { MDB_REVERSEDUP, "reversedup" }, + { 0, NULL } +}; + +static volatile sig_atomic_t gotsig; + +static void dumpsig( int sig ) +{ + gotsig=1; +} + +static const char hexc[] = "0123456789abcdef"; + +static void hex(unsigned char c) +{ + putchar(hexc[c >> 4]); + putchar(hexc[c & 0xf]); +} + +static void text(MDB_val *v) +{ + unsigned char *c, *end; + + putchar(' '); + c = v->mv_data; + end = c + v->mv_size; + while (c < end) { + if (isprint(*c)) { + putchar(*c); + } else { + putchar('\\'); + hex(*c); + } + c++; + } + putchar('\n'); +} + +static void byte(MDB_val *v) +{ + unsigned char *c, *end; + + putchar(' '); + c = v->mv_data; + end = c + v->mv_size; + while (c < end) { + hex(*c++); + } + putchar('\n'); +} + +/* Dump in BDB-compatible format */ +static int dumpit(MDB_txn *txn, MDB_dbi dbi, char *name) +{ + MDB_cursor *mc; + MDB_stat ms; + MDB_val key, data; + MDB_envinfo info; + unsigned int flags; + int rc, i; + + rc = mdb_dbi_flags(txn, dbi, &flags); + if (rc) return rc; + + rc = mdb_stat(txn, dbi, &ms); + if (rc) return rc; + + rc = mdb_env_info(mdb_txn_env(txn), &info); + if (rc) return rc; + + printf("VERSION=3\n"); + printf("format=%s\n", mode & PRINT ? "print" : "bytevalue"); + if (name) + printf("database=%s\n", name); + printf("type=btree\n"); + printf("mapsize=%" Z "u\n", info.me_mapsize); + if (info.me_mapaddr) + printf("mapaddr=%p\n", info.me_mapaddr); + printf("maxreaders=%u\n", info.me_maxreaders); + + if (flags & MDB_DUPSORT) + printf("duplicates=1\n"); + + for (i=0; dbflags[i].bit; i++) + if (flags & dbflags[i].bit) + printf("%s=1\n", dbflags[i].name); + + printf("db_pagesize=%d\n", ms.ms_psize); + printf("HEADER=END\n"); + + rc = mdb_cursor_open(txn, dbi, &mc); + if (rc) return rc; + + while ((rc = mdb_cursor_get(mc, &key, &data, MDB_NEXT) == MDB_SUCCESS)) { + if (gotsig) { + rc = EINTR; + break; + } + if (mode & PRINT) { + text(&key); + text(&data); + } else { + byte(&key); + byte(&data); + } + } + printf("DATA=END\n"); + if (rc == MDB_NOTFOUND) + rc = MDB_SUCCESS; + + return rc; +} + +static void usage(char *prog) +{ + fprintf(stderr, "usage: %s dbpath [-V] [-f output] [-l] [-n] [-p] [-a|-s subdb]\n", prog); + exit(EXIT_FAILURE); +} + +int main(int argc, char *argv[]) +{ + int i, rc; + MDB_env *env; + MDB_txn *txn; + MDB_dbi dbi; + char *prog = argv[0]; + char *envname; + char *subname = NULL; + int alldbs = 0, envflags = 0, list = 0; + + if (argc < 2) { + usage(prog); + } + + /* -a: dump main DB and all subDBs + * -s: dump only the named subDB + * -n: use NOSUBDIR flag on env_open + * -p: use printable characters + * -f: write to file instead of stdout + * -V: print version and exit + * (default) dump only the main DB + */ + while ((i = getopt(argc, argv, "af:lnps:V")) != EOF) { + switch(i) { + case 'V': + printf("%s\n", MDB_VERSION_STRING); + exit(0); + break; + case 'l': + list = 1; + /*FALLTHROUGH*/; + case 'a': + if (subname) + usage(prog); + alldbs++; + break; + case 'f': + if (freopen(optarg, "w", stdout) == NULL) { + fprintf(stderr, "%s: %s: reopen: %s\n", + prog, optarg, strerror(errno)); + exit(EXIT_FAILURE); + } + break; + case 'n': + envflags |= MDB_NOSUBDIR; + break; + case 'p': + mode |= PRINT; + break; + case 's': + if (alldbs) + usage(prog); + subname = optarg; + break; + default: + usage(prog); + } + } + + if (optind != argc - 1) + usage(prog); + +#ifdef SIGPIPE + signal(SIGPIPE, dumpsig); +#endif +#ifdef SIGHUP + signal(SIGHUP, dumpsig); +#endif + signal(SIGINT, dumpsig); + signal(SIGTERM, dumpsig); + + envname = argv[optind]; + rc = mdb_env_create(&env); + if (rc) { + fprintf(stderr, "mdb_env_create failed, error %d %s\n", rc, mdb_strerror(rc)); + return EXIT_FAILURE; + } + + if (alldbs || subname) { + mdb_env_set_maxdbs(env, 2); + } + + rc = mdb_env_open(env, envname, envflags | MDB_RDONLY, 0664); + if (rc) { + fprintf(stderr, "mdb_env_open failed, error %d %s\n", rc, mdb_strerror(rc)); + goto env_close; + } + + rc = mdb_txn_begin(env, NULL, MDB_RDONLY, &txn); + if (rc) { + fprintf(stderr, "mdb_txn_begin failed, error %d %s\n", rc, mdb_strerror(rc)); + goto env_close; + } + + rc = mdb_open(txn, subname, 0, &dbi); + if (rc) { + fprintf(stderr, "mdb_open failed, error %d %s\n", rc, mdb_strerror(rc)); + goto txn_abort; + } + + if (alldbs) { + MDB_cursor *cursor; + MDB_val key; + int count = 0; + + rc = mdb_cursor_open(txn, dbi, &cursor); + if (rc) { + fprintf(stderr, "mdb_cursor_open failed, error %d %s\n", rc, mdb_strerror(rc)); + goto txn_abort; + } + while ((rc = mdb_cursor_get(cursor, &key, NULL, MDB_NEXT_NODUP)) == 0) { + char *str; + MDB_dbi db2; + if (memchr(key.mv_data, '\0', key.mv_size)) + continue; + count++; + str = malloc(key.mv_size+1); + memcpy(str, key.mv_data, key.mv_size); + str[key.mv_size] = '\0'; + rc = mdb_open(txn, str, 0, &db2); + if (rc == MDB_SUCCESS) { + if (list) { + printf("%s\n", str); + list++; + } else { + rc = dumpit(txn, db2, str); + if (rc) + break; + } + mdb_close(env, db2); + } + free(str); + if (rc) continue; + } + mdb_cursor_close(cursor); + if (!count) { + fprintf(stderr, "%s: %s does not contain multiple databases\n", prog, envname); + rc = MDB_NOTFOUND; + } else if (rc == MDB_NOTFOUND) { + rc = MDB_SUCCESS; + } + } else { + rc = dumpit(txn, dbi, subname); + } + if (rc && rc != MDB_NOTFOUND) + fprintf(stderr, "%s: %s: %s\n", prog, envname, mdb_strerror(rc)); + + mdb_close(env, dbi); +txn_abort: + mdb_txn_abort(txn); +env_close: + mdb_env_close(env); + + return rc ? EXIT_FAILURE : EXIT_SUCCESS; +} diff --git a/db_drivers/liblmdb/mdb_load.1 b/db_drivers/liblmdb/mdb_load.1 new file mode 100644 index 000000000..728024010 --- /dev/null +++ b/db_drivers/liblmdb/mdb_load.1 @@ -0,0 +1,77 @@ +.TH MDB_LOAD 1 "2014/06/20" "LMDB 0.9.14" +.\" Copyright 2014 Howard Chu, Symas Corp. All Rights Reserved. +.\" Copying restrictions apply. See COPYRIGHT/LICENSE. +.SH NAME +mdb_load \- LMDB environment import tool +.SH SYNOPSIS +.B mdb_load +.BR \ envpath +[\c +.BR \-V ] +[\c +.BI \-f \ file\fR] +[\c +.BR \-n ] +[\c +.BI \-s \ subdb\fR] +[\c +.BR \-N ] +[\c +.BR \-T ] +.SH DESCRIPTION +The +.B mdb_load +utility reads from the standard input and loads it into the +LMDB environment +.BR envpath . + +The input to +.B mdb_load +must be in the output format specified by the +.BR mdb_dump (1) +utility or as specified by the +.B -T +option below. +.SH OPTIONS +.TP +.BR \-V +Write the library version number to the standard output, and exit. +.TP +.BR \-f \ file +Read from the specified file instead of from the standard input. +.TP +.BR \-n +Load an LMDB database which does not use subdirectories. +.TP +.BR \-s \ subdb +Load a specific subdatabase. If no database is specified, data is loaded into the main database. +.TP +.BR \-N +Don't overwrite existing records when loading into an already existing database; just skip them. +.TP +.BR \-T +Load data from simple text files. The input must be paired lines of text, where the first +line of the pair is the key item, and the second line of the pair is its corresponding +data item. + +A simple escape mechanism, where newline and backslash (\\) characters are special, is +applied to the text input. Newline characters are interpreted as record separators. +Backslash characters in the text will be interpreted in one of two ways: If the backslash +character precedes another backslash character, the pair will be interpreted as a literal +backslash. If the backslash character precedes any other character, the two characters +following the backslash will be interpreted as a hexadecimal specification of a single +character; for example, \\0a is a newline character in the ASCII character set. + +For this reason, any backslash or newline characters that naturally occur in the text +input must be escaped to avoid misinterpretation by +.BR mdb_load . + +.SH DIAGNOSTICS +Exit status is zero if no errors occur. +Errors result in a non-zero exit status and +a diagnostic message being written to standard error. + +.SH "SEE ALSO" +.BR mdb_dump (1) +.SH AUTHOR +Howard Chu of Symas Corporation <http://www.symas.com> diff --git a/db_drivers/liblmdb/mdb_load.c b/db_drivers/liblmdb/mdb_load.c new file mode 100644 index 000000000..e0d95e13c --- /dev/null +++ b/db_drivers/liblmdb/mdb_load.c @@ -0,0 +1,452 @@ +/* mdb_load.c - memory-mapped database load tool */ +/* + * Copyright 2011-2014 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 <errno.h> +#include <string.h> +#include <ctype.h> +#include <unistd.h> +#include "lmdb.h" + +#define PRINT 1 +#define NOHDR 2 +static int mode; + +static char *subname = NULL; + +static size_t lineno; +static int version; + +static int flags; + +static char *prog; + +static int Eof; + +static MDB_envinfo info; + +static MDB_val kbuf, dbuf; + +#ifdef _WIN32 +#define Z "I" +#else +#define Z "z" +#endif + +#define STRLENOF(s) (sizeof(s)-1) + +typedef struct flagbit { + int bit; + char *name; + int len; +} flagbit; + +#define S(s) s, STRLENOF(s) + +flagbit dbflags[] = { + { MDB_REVERSEKEY, S("reversekey") }, + { MDB_DUPSORT, S("dupsort") }, + { MDB_INTEGERKEY, S("integerkey") }, + { MDB_DUPFIXED, S("dupfixed") }, + { MDB_INTEGERDUP, S("integerdup") }, + { MDB_REVERSEDUP, S("reversedup") }, + { 0, NULL, 0 } +}; + +static void readhdr(void) +{ + char *ptr; + + while (fgets(dbuf.mv_data, dbuf.mv_size, stdin) != NULL) { + lineno++; + if (!strncmp(dbuf.mv_data, "VERSION=", STRLENOF("VERSION="))) { + version=atoi((char *)dbuf.mv_data+STRLENOF("VERSION=")); + if (version > 3) { + fprintf(stderr, "%s: line %" Z "d: unsupported VERSION %d\n", + prog, lineno, version); + exit(EXIT_FAILURE); + } + } else if (!strncmp(dbuf.mv_data, "HEADER=END", STRLENOF("HEADER=END"))) { + break; + } else if (!strncmp(dbuf.mv_data, "format=", STRLENOF("format="))) { + if (!strncmp((char *)dbuf.mv_data+STRLENOF("FORMAT="), "print", STRLENOF("print"))) + mode |= PRINT; + else if (strncmp((char *)dbuf.mv_data+STRLENOF("FORMAT="), "bytevalue", STRLENOF("bytevalue"))) { + fprintf(stderr, "%s: line %" Z "d: unsupported FORMAT %s\n", + prog, lineno, (char *)dbuf.mv_data+STRLENOF("FORMAT=")); + exit(EXIT_FAILURE); + } + } else if (!strncmp(dbuf.mv_data, "database=", STRLENOF("database="))) { + ptr = memchr(dbuf.mv_data, '\n', dbuf.mv_size); + if (ptr) *ptr = '\0'; + if (subname) free(subname); + subname = strdup((char *)dbuf.mv_data+STRLENOF("database=")); + } else if (!strncmp(dbuf.mv_data, "type=", STRLENOF("type="))) { + if (strncmp((char *)dbuf.mv_data+STRLENOF("type="), "btree", STRLENOF("btree"))) { + fprintf(stderr, "%s: line %" Z "d: unsupported type %s\n", + prog, lineno, (char *)dbuf.mv_data+STRLENOF("type=")); + exit(EXIT_FAILURE); + } + } else if (!strncmp(dbuf.mv_data, "mapaddr=", STRLENOF("mapaddr="))) { + int i; + ptr = memchr(dbuf.mv_data, '\n', dbuf.mv_size); + if (ptr) *ptr = '\0'; + i = sscanf((char *)dbuf.mv_data+STRLENOF("mapaddr="), "%p", &info.me_mapaddr); + if (i != 1) { + fprintf(stderr, "%s: line %" Z "d: invalid mapaddr %s\n", + prog, lineno, (char *)dbuf.mv_data+STRLENOF("mapaddr=")); + exit(EXIT_FAILURE); + } + } else if (!strncmp(dbuf.mv_data, "mapsize=", STRLENOF("mapsize="))) { + int i; + ptr = memchr(dbuf.mv_data, '\n', dbuf.mv_size); + if (ptr) *ptr = '\0'; + i = sscanf((char *)dbuf.mv_data+STRLENOF("mapsize="), "%" Z "u", &info.me_mapsize); + if (i != 1) { + fprintf(stderr, "%s: line %" Z "d: invalid mapsize %s\n", + prog, lineno, (char *)dbuf.mv_data+STRLENOF("mapsize=")); + exit(EXIT_FAILURE); + } + } else if (!strncmp(dbuf.mv_data, "maxreaders=", STRLENOF("maxreaders="))) { + int i; + ptr = memchr(dbuf.mv_data, '\n', dbuf.mv_size); + if (ptr) *ptr = '\0'; + i = sscanf((char *)dbuf.mv_data+STRLENOF("maxreaders="), "%u", &info.me_maxreaders); + if (i != 1) { + fprintf(stderr, "%s: line %" Z "d: invalid maxreaders %s\n", + prog, lineno, (char *)dbuf.mv_data+STRLENOF("maxreaders=")); + exit(EXIT_FAILURE); + } + } else { + int i; + for (i=0; dbflags[i].bit; i++) { + if (!strncmp(dbuf.mv_data, dbflags[i].name, dbflags[i].len) && + ((char *)dbuf.mv_data)[dbflags[i].len] == '=') { + flags |= dbflags[i].bit; + break; + } + } + if (!dbflags[i].bit) { + ptr = memchr(dbuf.mv_data, '=', dbuf.mv_size); + if (!ptr) { + fprintf(stderr, "%s: line %" Z "d: unexpected format\n", + prog, lineno); + exit(EXIT_FAILURE); + } else { + *ptr = '\0'; + fprintf(stderr, "%s: line %" Z "d: unrecognized keyword ignored: %s\n", + prog, lineno, (char *)dbuf.mv_data); + } + } + } + } +} + +static void badend(void) +{ + fprintf(stderr, "%s: line %" Z "d: unexpected end of input\n", + prog, lineno); +} + +static int unhex(unsigned char *c2) +{ + int x, c; + x = *c2++ & 0x4f; + if (x & 0x40) + x -= 55; + c = x << 4; + x = *c2 & 0x4f; + if (x & 0x40) + x -= 55; + c |= x; + return c; +} + +static int readline(MDB_val *out, MDB_val *buf) +{ + unsigned char *c1, *c2, *end; + size_t len; + int c; + + if (!(mode & NOHDR)) { + c = fgetc(stdin); + if (c == EOF) { + Eof = 1; + return EOF; + } + if (c != ' ') { + lineno++; + if (fgets(buf->mv_data, buf->mv_size, stdin) == NULL) { +badend: + Eof = 1; + badend(); + return EOF; + } + if (c == 'D' && !strncmp(buf->mv_data, "ATA=END", STRLENOF("ATA=END"))) + return EOF; + goto badend; + } + } + if (fgets(buf->mv_data, buf->mv_size, stdin) == NULL) { + Eof = 1; + return EOF; + } + lineno++; + + c1 = buf->mv_data; + len = strlen((char *)c1); + + /* Is buffer too short? */ + while (c1[len-1] != '\n') { + buf->mv_data = realloc(buf->mv_data, buf->mv_size*2); + if (!buf->mv_data) { + Eof = 1; + fprintf(stderr, "%s: line %" Z "d: out of memory, line too long\n", + prog, lineno); + return EOF; + } + c1 = buf->mv_data; + c1 += buf->mv_size; + if (fgets((char *)c1, buf->mv_size, stdin) == NULL) { + Eof = 1; + badend(); + return EOF; + } + buf->mv_size *= 2; + len = strlen((char *)c1); + } + c1 = c2 = buf->mv_data; + len = strlen((char *)c1); + c1[--len] = '\0'; + end = c1 + len; + + if (mode & PRINT) { + while (c2 < end) { + if (*c2 == '\\') { + if (c2[1] == '\\') { + c1++; c2 += 2; + } else { + if (c2+3 > end || !isxdigit(c2[1]) || !isxdigit(c2[2])) { + Eof = 1; + badend(); + return EOF; + } + *c1++ = unhex(++c2); + c2 += 2; + } + } else { + c1++; c2++; + } + } + } else { + /* odd length not allowed */ + if (len & 1) { + Eof = 1; + badend(); + return EOF; + } + while (c2 < end) { + if (!isxdigit(*c2) || !isxdigit(c2[1])) { + Eof = 1; + badend(); + return EOF; + } + *c1++ = unhex(c2); + c2 += 2; + } + } + c2 = out->mv_data = buf->mv_data; + out->mv_size = c1 - c2; + + return 0; +} + +static void usage(void) +{ + fprintf(stderr, "usage: %s dbpath [-V] [-f input] [-n] [-s name] [-N] [-T]\n", prog); + exit(EXIT_FAILURE); +} + +int main(int argc, char *argv[]) +{ + int i, rc; + MDB_env *env; + MDB_txn *txn; + MDB_cursor *mc; + MDB_dbi dbi; + char *envname; + int envflags = 0, putflags = 0; + int dohdr = 0; + + prog = argv[0]; + + if (argc < 2) { + usage(); + } + + /* -f: load file instead of stdin + * -n: use NOSUBDIR flag on env_open + * -s: load into named subDB + * -N: use NOOVERWRITE on puts + * -T: read plaintext + * -V: print version and exit + */ + while ((i = getopt(argc, argv, "f:ns:NTV")) != EOF) { + switch(i) { + case 'V': + printf("%s\n", MDB_VERSION_STRING); + exit(0); + break; + case 'f': + if (freopen(optarg, "r", stdin) == NULL) { + fprintf(stderr, "%s: %s: reopen: %s\n", + prog, optarg, strerror(errno)); + exit(EXIT_FAILURE); + } + break; + case 'n': + envflags |= MDB_NOSUBDIR; + break; + case 's': + subname = strdup(optarg); + break; + case 'N': + putflags = MDB_NOOVERWRITE|MDB_NODUPDATA; + break; + case 'T': + mode |= NOHDR; + break; + default: + usage(); + } + } + + if (optind != argc - 1) + usage(); + + dbuf.mv_size = 4096; + dbuf.mv_data = malloc(dbuf.mv_size); + + if (!(mode & NOHDR)) + readhdr(); + + envname = argv[optind]; + rc = mdb_env_create(&env); + if (rc) { + fprintf(stderr, "mdb_env_create failed, error %d %s\n", rc, mdb_strerror(rc)); + return EXIT_FAILURE; + } + + mdb_env_set_maxdbs(env, 2); + + if (info.me_maxreaders) + mdb_env_set_maxreaders(env, info.me_maxreaders); + + if (info.me_mapsize) + mdb_env_set_mapsize(env, info.me_mapsize); + + if (info.me_mapaddr) + envflags |= MDB_FIXEDMAP; + + rc = mdb_env_open(env, envname, envflags, 0664); + if (rc) { + fprintf(stderr, "mdb_env_open failed, error %d %s\n", rc, mdb_strerror(rc)); + goto env_close; + } + + kbuf.mv_size = mdb_env_get_maxkeysize(env) * 2 + 2; + kbuf.mv_data = malloc(kbuf.mv_size); + + while(!Eof) { + MDB_val key, data; + int batch = 0; + flags = 0; + + if (!dohdr) { + dohdr = 1; + } else if (!(mode & NOHDR)) + readhdr(); + + rc = mdb_txn_begin(env, NULL, 0, &txn); + if (rc) { + fprintf(stderr, "mdb_txn_begin failed, error %d %s\n", rc, mdb_strerror(rc)); + goto env_close; + } + + rc = mdb_open(txn, subname, flags|MDB_CREATE, &dbi); + if (rc) { + fprintf(stderr, "mdb_open failed, error %d %s\n", rc, mdb_strerror(rc)); + goto txn_abort; + } + + rc = mdb_cursor_open(txn, dbi, &mc); + if (rc) { + fprintf(stderr, "mdb_cursor_open failed, error %d %s\n", rc, mdb_strerror(rc)); + goto txn_abort; + } + + while(1) { + rc = readline(&key, &kbuf); + if (rc == EOF) + break; + if (rc) + goto txn_abort; + + rc = readline(&data, &dbuf); + if (rc) + goto txn_abort; + + rc = mdb_cursor_put(mc, &key, &data, putflags); + if (rc == MDB_KEYEXIST && putflags) + continue; + if (rc) + goto txn_abort; + batch++; + if (batch == 100) { + rc = mdb_txn_commit(txn); + if (rc) { + fprintf(stderr, "%s: line %" Z "d: txn_commit: %s\n", + prog, lineno, mdb_strerror(rc)); + goto env_close; + } + rc = mdb_txn_begin(env, NULL, 0, &txn); + if (rc) { + fprintf(stderr, "mdb_txn_begin failed, error %d %s\n", rc, mdb_strerror(rc)); + goto env_close; + } + rc = mdb_cursor_open(txn, dbi, &mc); + if (rc) { + fprintf(stderr, "mdb_cursor_open failed, error %d %s\n", rc, mdb_strerror(rc)); + goto txn_abort; + } + batch = 0; + } + } + rc = mdb_txn_commit(txn); + txn = NULL; + if (rc) { + fprintf(stderr, "%s: line %" Z "d: txn_commit: %s\n", + prog, lineno, mdb_strerror(rc)); + goto env_close; + } + mdb_dbi_close(env, dbi); + } + +txn_abort: + mdb_txn_abort(txn); +env_close: + mdb_env_close(env); + + return rc ? EXIT_FAILURE : EXIT_SUCCESS; +} diff --git a/db_drivers/liblmdb/mdb_stat.1 b/db_drivers/liblmdb/mdb_stat.1 new file mode 100644 index 000000000..3d8d461d9 --- /dev/null +++ b/db_drivers/liblmdb/mdb_stat.1 @@ -0,0 +1,64 @@ +.TH MDB_STAT 1 "2014/06/20" "LMDB 0.9.14" +.\" Copyright 2012-2014 Howard Chu, Symas Corp. All Rights Reserved. +.\" Copying restrictions apply. See COPYRIGHT/LICENSE. +.SH NAME +mdb_stat \- LMDB environment status tool +.SH SYNOPSIS +.B mdb_stat +.BR \ envpath +[\c +.BR \-V ] +[\c +.BR \-e ] +[\c +.BR \-f [ f [ f ]]] +[\c +.BR \-n ] +[\c +.BR \-r [ r ]] +[\c +.BR \-a \ | +.BI \-s \ subdb\fR] +.SH DESCRIPTION +The +.B mdb_stat +utility displays the status of an LMDB environment. +.SH OPTIONS +.TP +.BR \-V +Write the library version number to the standard output, and exit. +.TP +.BR \-e +Display information about the database environment. +.TP +.BR \-f +Display information about the environment freelist. +If \fB\-ff\fP is given, summarize each freelist entry. +If \fB\-fff\fP is given, display the full list of page IDs in the freelist. +.TP +.BR \-n +Display the status of an LMDB database which does not use subdirectories. +.TP +.BR \-r +Display information about the environment reader table. +Shows the process ID, thread ID, and transaction ID for each active +reader slot. The process ID and transaction ID are in decimal, the +thread ID is in hexadecimal. The transaction ID is displayed as "-" +if the reader does not currently have a read transaction open. +If \fB\-rr\fP is given, check for stale entries in the reader +table and clear them. The reader table will be printed again +after the check is performed. +.TP +.BR \-a +Display the status of all of the subdatabases in the environment. +.TP +.BR \-s \ subdb +Display the status of a specific subdatabase. +.SH DIAGNOSTICS +Exit status is zero if no errors occur. +Errors result in a non-zero exit status and +a diagnostic message being written to standard error. +.SH "SEE ALSO" +.BR mdb_copy (1) +.SH AUTHOR +Howard Chu of Symas Corporation <http://www.symas.com> diff --git a/db_drivers/liblmdb/mdb_stat.c b/db_drivers/liblmdb/mdb_stat.c new file mode 100644 index 000000000..1e9229296 --- /dev/null +++ b/db_drivers/liblmdb/mdb_stat.c @@ -0,0 +1,263 @@ +/* mdb_stat.c - memory-mapped database status tool */ +/* + * Copyright 2011-2013 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 <string.h> +#include <unistd.h> +#include "lmdb.h" + +#ifdef _WIN32 +#define Z "I" +#else +#define Z "z" +#endif + +static void prstat(MDB_stat *ms) +{ +#if 0 + printf(" Page size: %u\n", ms->ms_psize); +#endif + printf(" Tree depth: %u\n", ms->ms_depth); + printf(" Branch pages: %"Z"u\n", ms->ms_branch_pages); + printf(" Leaf pages: %"Z"u\n", ms->ms_leaf_pages); + printf(" Overflow pages: %"Z"u\n", ms->ms_overflow_pages); + printf(" Entries: %"Z"u\n", ms->ms_entries); +} + +static void usage(char *prog) +{ + fprintf(stderr, "usage: %s dbpath [-V] [-n] [-e] [-r[r]] [-f[f[f]]] [-a|-s subdb]\n", prog); + exit(EXIT_FAILURE); +} + +int main(int argc, char *argv[]) +{ + int i, rc; + MDB_env *env; + MDB_txn *txn; + MDB_dbi dbi; + MDB_stat mst; + MDB_envinfo mei; + char *prog = argv[0]; + char *envname; + char *subname = NULL; + int alldbs = 0, envinfo = 0, envflags = 0, freinfo = 0, rdrinfo = 0; + + if (argc < 2) { + usage(prog); + } + + /* -a: print stat of main DB and all subDBs + * -s: print stat of only the named subDB + * -e: print env info + * -f: print freelist info + * -r: print reader info + * -n: use NOSUBDIR flag on env_open + * -V: print version and exit + * (default) print stat of only the main DB + */ + while ((i = getopt(argc, argv, "Vaefnrs:")) != EOF) { + switch(i) { + case 'V': + printf("%s\n", MDB_VERSION_STRING); + exit(0); + break; + case 'a': + if (subname) + usage(prog); + alldbs++; + break; + case 'e': + envinfo++; + break; + case 'f': + freinfo++; + break; + case 'n': + envflags |= MDB_NOSUBDIR; + break; + case 'r': + rdrinfo++; + break; + case 's': + if (alldbs) + usage(prog); + subname = optarg; + break; + default: + usage(prog); + } + } + + if (optind != argc - 1) + usage(prog); + + envname = argv[optind]; + rc = mdb_env_create(&env); + if (rc) { + fprintf(stderr, "mdb_env_create failed, error %d %s\n", rc, mdb_strerror(rc)); + return EXIT_FAILURE; + } + + if (alldbs || subname) { + mdb_env_set_maxdbs(env, 4); + } + + rc = mdb_env_open(env, envname, envflags | MDB_RDONLY, 0664); + if (rc) { + fprintf(stderr, "mdb_env_open failed, error %d %s\n", rc, mdb_strerror(rc)); + goto env_close; + } + + if (envinfo) { + (void)mdb_env_stat(env, &mst); + (void)mdb_env_info(env, &mei); + printf("Environment Info\n"); + printf(" Map address: %p\n", mei.me_mapaddr); + printf(" Map size: %"Z"u\n", mei.me_mapsize); + printf(" Page size: %u\n", mst.ms_psize); + printf(" Max pages: %"Z"u\n", mei.me_mapsize / mst.ms_psize); + printf(" Number of pages used: %"Z"u\n", mei.me_last_pgno+1); + printf(" Last transaction ID: %"Z"u\n", mei.me_last_txnid); + printf(" Max readers: %u\n", mei.me_maxreaders); + printf(" Number of readers used: %u\n", mei.me_numreaders); + } + + if (rdrinfo) { + printf("Reader Table Status\n"); + rc = mdb_reader_list(env, (MDB_msg_func *)fputs, stdout); + if (rdrinfo > 1) { + int dead; + mdb_reader_check(env, &dead); + printf(" %d stale readers cleared.\n", dead); + rc = mdb_reader_list(env, (MDB_msg_func *)fputs, stdout); + } + if (!(subname || alldbs || freinfo)) + goto env_close; + } + + rc = mdb_txn_begin(env, NULL, MDB_RDONLY, &txn); + if (rc) { + fprintf(stderr, "mdb_txn_begin failed, error %d %s\n", rc, mdb_strerror(rc)); + goto env_close; + } + + if (freinfo) { + MDB_cursor *cursor; + MDB_val key, data; + size_t pages = 0, *iptr; + + printf("Freelist Status\n"); + dbi = 0; + rc = mdb_cursor_open(txn, dbi, &cursor); + if (rc) { + fprintf(stderr, "mdb_cursor_open failed, error %d %s\n", rc, mdb_strerror(rc)); + goto txn_abort; + } + rc = mdb_stat(txn, dbi, &mst); + if (rc) { + fprintf(stderr, "mdb_stat failed, error %d %s\n", rc, mdb_strerror(rc)); + goto txn_abort; + } + prstat(&mst); + while ((rc = mdb_cursor_get(cursor, &key, &data, MDB_NEXT)) == 0) { + iptr = data.mv_data; + pages += *iptr; + if (freinfo > 1) { + char *bad = ""; + size_t pg, prev; + ssize_t i, j, span = 0; + j = *iptr++; + for (i = j, prev = 1; --i >= 0; ) { + pg = iptr[i]; + if (pg <= prev) + bad = " [bad sequence]"; + prev = pg; + pg += span; + for (; i >= span && iptr[i-span] == pg; span++, pg++) ; + } + printf(" Transaction %"Z"u, %"Z"d pages, maxspan %"Z"d%s\n", + *(size_t *)key.mv_data, j, span, bad); + if (freinfo > 2) { + for (--j; j >= 0; ) { + pg = iptr[j]; + for (span=1; --j >= 0 && iptr[j] == pg+span; span++) ; + printf(span>1 ? " %9"Z"u[%"Z"d]\n" : " %9"Z"u\n", + pg, span); + } + } + } + } + mdb_cursor_close(cursor); + printf(" Free pages: %"Z"u\n", pages); + } + + rc = mdb_open(txn, subname, 0, &dbi); + if (rc) { + fprintf(stderr, "mdb_open failed, error %d %s\n", rc, mdb_strerror(rc)); + goto txn_abort; + } + + rc = mdb_stat(txn, dbi, &mst); + if (rc) { + fprintf(stderr, "mdb_stat failed, error %d %s\n", rc, mdb_strerror(rc)); + goto txn_abort; + } + printf("Status of %s\n", subname ? subname : "Main DB"); + prstat(&mst); + + if (alldbs) { + MDB_cursor *cursor; + MDB_val key; + + rc = mdb_cursor_open(txn, dbi, &cursor); + if (rc) { + fprintf(stderr, "mdb_cursor_open failed, error %d %s\n", rc, mdb_strerror(rc)); + goto txn_abort; + } + while ((rc = mdb_cursor_get(cursor, &key, NULL, MDB_NEXT_NODUP)) == 0) { + char *str; + MDB_dbi db2; + if (memchr(key.mv_data, '\0', key.mv_size)) + continue; + str = malloc(key.mv_size+1); + memcpy(str, key.mv_data, key.mv_size); + str[key.mv_size] = '\0'; + rc = mdb_open(txn, str, 0, &db2); + if (rc == MDB_SUCCESS) + printf("Status of %s\n", str); + free(str); + if (rc) continue; + rc = mdb_stat(txn, db2, &mst); + if (rc) { + fprintf(stderr, "mdb_stat failed, error %d %s\n", rc, mdb_strerror(rc)); + goto txn_abort; + } + prstat(&mst); + mdb_close(env, db2); + } + mdb_cursor_close(cursor); + } + + if (rc == MDB_NOTFOUND) + rc = MDB_SUCCESS; + + mdb_close(env, dbi); +txn_abort: + mdb_txn_abort(txn); +env_close: + mdb_env_close(env); + + return rc ? EXIT_FAILURE : EXIT_SUCCESS; +} diff --git a/db_drivers/liblmdb/midl.c b/db_drivers/liblmdb/midl.c new file mode 100644 index 000000000..88a3aff10 --- /dev/null +++ b/db_drivers/liblmdb/midl.c @@ -0,0 +1,360 @@ +/** @file midl.c + * @brief ldap bdb back-end ID List functions */ +/* $OpenLDAP$ */ +/* This work is part of OpenLDAP Software <http://www.openldap.org/>. + * + * Copyright 2000-2014 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 <errno.h> +#include <sys/types.h> +#include "midl.h" + +/** @defgroup internal LMDB Internals + * @{ + */ +/** @defgroup idls ID List Management + * @{ + */ +#define CMP(x,y) ( (x) < (y) ? -1 : (x) > (y) ) + +unsigned mdb_midl_search( MDB_IDL ids, MDB_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; +} + +#if 0 /* superseded by append/sort */ +int mdb_midl_insert( MDB_IDL ids, MDB_ID id ) +{ + unsigned x, i; + + 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 ) { + /* no room */ + --ids[0]; + return -2; + + } else { + /* insert id */ + for (i=ids[0]; i>x; i--) + ids[i] = ids[i-1]; + ids[x] = id; + } + + return 0; +} +#endif + +MDB_IDL mdb_midl_alloc(int num) +{ + MDB_IDL ids = malloc((num+2) * sizeof(MDB_ID)); + if (ids) { + *ids++ = num; + *ids = 0; + } + return ids; +} + +void mdb_midl_free(MDB_IDL ids) +{ + if (ids) + free(ids-1); +} + +int mdb_midl_shrink( MDB_IDL *idp ) +{ + MDB_IDL ids = *idp; + if (*(--ids) > MDB_IDL_UM_MAX && + (ids = realloc(ids, (MDB_IDL_UM_MAX+1) * sizeof(MDB_ID)))) + { + *ids++ = MDB_IDL_UM_MAX; + *idp = ids; + return 1; + } + return 0; +} + +static int mdb_midl_grow( MDB_IDL *idp, int num ) +{ + MDB_IDL idn = *idp-1; + /* grow it */ + idn = realloc(idn, (*idn + num + 2) * sizeof(MDB_ID)); + if (!idn) + return ENOMEM; + *idn++ += num; + *idp = idn; + return 0; +} + +int mdb_midl_need( MDB_IDL *idp, unsigned num ) +{ + MDB_IDL ids = *idp; + num += ids[0]; + if (num > ids[-1]) { + num = (num + num/4 + (256 + 2)) & -256; + if (!(ids = realloc(ids-1, num * sizeof(MDB_ID)))) + return ENOMEM; + *ids++ = num - 2; + *idp = ids; + } + return 0; +} + +int mdb_midl_append( MDB_IDL *idp, MDB_ID id ) +{ + MDB_IDL ids = *idp; + /* Too big? */ + if (ids[0] >= ids[-1]) { + if (mdb_midl_grow(idp, MDB_IDL_UM_MAX)) + return ENOMEM; + ids = *idp; + } + ids[0]++; + ids[ids[0]] = id; + return 0; +} + +int mdb_midl_append_list( MDB_IDL *idp, MDB_IDL app ) +{ + MDB_IDL ids = *idp; + /* Too big? */ + if (ids[0] + app[0] >= ids[-1]) { + if (mdb_midl_grow(idp, app[0])) + return ENOMEM; + ids = *idp; + } + memcpy(&ids[ids[0]+1], &app[1], app[0] * sizeof(MDB_ID)); + ids[0] += app[0]; + return 0; +} + +int mdb_midl_append_range( MDB_IDL *idp, MDB_ID id, unsigned n ) +{ + MDB_ID *ids = *idp, len = ids[0]; + /* Too big? */ + if (len + n > ids[-1]) { + if (mdb_midl_grow(idp, n | MDB_IDL_UM_MAX)) + return ENOMEM; + ids = *idp; + } + ids[0] = len + n; + ids += len; + while (n) + ids[n--] = id++; + return 0; +} + +void mdb_midl_xmerge( MDB_IDL idl, MDB_IDL merge ) +{ + MDB_ID old_id, merge_id, i = merge[0], j = idl[0], k = i+j, total = k; + idl[0] = (MDB_ID)-1; /* delimiter for idl scan below */ + old_id = idl[j]; + while (i) { + merge_id = merge[i--]; + for (; old_id < merge_id; old_id = idl[--j]) + idl[k--] = old_id; + idl[k--] = merge_id; + } + idl[0] = total; +} + +/* Quicksort + Insertion sort for small arrays */ + +#define SMALL 8 +#define MIDL_SWAP(a,b) { itmp=(a); (a)=(b); (b)=itmp; } + +void +mdb_midl_sort( MDB_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; + MDB_ID a, itmp; + + ir = (int)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 */ + MIDL_SWAP(ids[k], ids[l+1]); + if (ids[l] < ids[ir]) { + MIDL_SWAP(ids[l], ids[ir]); + } + if (ids[l+1] < ids[ir]) { + MIDL_SWAP(ids[l+1], ids[ir]); + } + if (ids[l] < ids[l+1]) { + MIDL_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; + MIDL_SWAP(ids[i],ids[j]); + } + ids[l+1] = ids[j]; + ids[j] = a; + jstack += 2; + if (ir-i+1 >= j-l) { + 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( MDB_ID2L ids, MDB_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 = (unsigned)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( MDB_ID2L ids, MDB_ID2 *id ) +{ + unsigned x, i; + + x = mdb_mid2l_search( ids, id->mid ); + + 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=(unsigned)ids[0].mid; i>x; i--) + ids[i] = ids[i-1]; + ids[x] = *id; + } + + return 0; +} + +int mdb_mid2l_append( MDB_ID2L ids, MDB_ID2 *id ) +{ + /* Too big? */ + if (ids[0].mid >= MDB_IDL_UM_MAX) { + return -2; + } + ids[0].mid++; + ids[ids[0].mid] = *id; + return 0; +} + +/** @} */ +/** @} */ diff --git a/db_drivers/liblmdb/midl.h b/db_drivers/liblmdb/midl.h new file mode 100644 index 000000000..a7f25026c --- /dev/null +++ b/db_drivers/liblmdb/midl.h @@ -0,0 +1,186 @@ +/** @file midl.h + * @brief LMDB 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 2000-2014 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> + +#ifdef __cplusplus +extern "C" { +#endif + +/** @defgroup internal LMDB Internals + * @{ + */ + +/** @defgroup idls ID List Management + * @{ + */ + /** A generic unsigned ID number. These were entryIDs in back-bdb. + * Preferably it should have the same size as a pointer. + */ +typedef size_t MDB_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 MDB_ID *MDB_IDL; + +/* 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_DB_MAX (MDB_IDL_DB_SIZE-1) +#define MDB_IDL_UM_MAX (MDB_IDL_UM_SIZE-1) + +#define MDB_IDL_SIZEOF(ids) (((ids)[0]+1) * sizeof(MDB_ID)) +#define MDB_IDL_IS_ZERO(ids) ( (ids)[0] == 0 ) +#define MDB_IDL_CPY( dst, src ) (memcpy( dst, src, MDB_IDL_SIZEOF( src ) )) +#define MDB_IDL_FIRST( ids ) ( (ids)[1] ) +#define MDB_IDL_LAST( ids ) ( (ids)[(ids)[0]] ) + + /** Current max length of an #mdb_midl_alloc()ed IDL */ +#define MDB_IDL_ALLOCLEN( ids ) ( (ids)[-1] ) + + /** Append ID to IDL. The IDL must be big enough. */ +#define mdb_midl_xappend(idl, id) do { \ + MDB_ID *xidl = (idl), xlen = ++(xidl[0]); \ + xidl[xlen] = (id); \ + } while (0) + + /** Search for an ID in an IDL. + * @param[in] ids The IDL to search. + * @param[in] id The ID to search for. + * @return The index of the first ID greater than or equal to \b id. + */ +unsigned mdb_midl_search( MDB_IDL ids, MDB_ID id ); + + /** Allocate an IDL. + * Allocates memory for an IDL of the given size. + * @return IDL on success, NULL on failure. + */ +MDB_IDL mdb_midl_alloc(int num); + + /** Free an IDL. + * @param[in] ids The IDL to free. + */ +void mdb_midl_free(MDB_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(MDB_IDL *idp); + + /** Make room for num additional elements in an IDL. + * @param[in,out] idp Address of the IDL. + * @param[in] num Number of elements to make room for. + * @return 0 on success, ENOMEM on failure. + */ +int mdb_midl_need(MDB_IDL *idp, unsigned num); + + /** 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, ENOMEM if the IDL is too large. + */ +int mdb_midl_append( MDB_IDL *idp, MDB_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, ENOMEM if the IDL is too large. + */ +int mdb_midl_append_list( MDB_IDL *idp, MDB_IDL app ); + + /** Append an ID range onto an IDL. + * @param[in,out] idp Address of the IDL to append to. + * @param[in] id The lowest ID to append. + * @param[in] n Number of IDs to append. + * @return 0 on success, ENOMEM if the IDL is too large. + */ +int mdb_midl_append_range( MDB_IDL *idp, MDB_ID id, unsigned n ); + + /** Merge an IDL onto an IDL. The destination IDL must be big enough. + * @param[in] idl The IDL to merge into. + * @param[in] merge The IDL to merge. + */ +void mdb_midl_xmerge( MDB_IDL idl, MDB_IDL merge ); + + /** Sort an IDL. + * @param[in,out] ids The IDL to sort. + */ +void mdb_midl_sort( MDB_IDL ids ); + + /** An ID2 is an ID/pointer pair. + */ +typedef struct MDB_ID2 { + MDB_ID mid; /**< The ID */ + void *mptr; /**< The pointer */ +} MDB_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 MDB_ID2 *MDB_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( MDB_ID2L ids, MDB_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( MDB_ID2L ids, MDB_ID2 *id ); + + /** Append an ID2 into a ID2L. + * @param[in,out] ids The ID2L to append into. + * @param[in] id The ID2 to append. + * @return 0 on success, -2 if the ID2L is too big. + */ +int mdb_mid2l_append( MDB_ID2L ids, MDB_ID2 *id ); + +/** @} */ +/** @} */ +#ifdef __cplusplus +} +#endif +#endif /* _MDB_MIDL_H_ */ diff --git a/db_drivers/liblmdb/mtest.c b/db_drivers/liblmdb/mtest.c new file mode 100644 index 000000000..79b4175e7 --- /dev/null +++ b/db_drivers/liblmdb/mtest.c @@ -0,0 +1,173 @@ +/* mtest.c - memory-mapped database tester/toy */ +/* + * Copyright 2011-2014 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 "lmdb.h" + +#define E(expr) CHECK((rc = (expr)) == MDB_SUCCESS, #expr) +#define RES(err, expr) ((rc = expr) == (err) || (CHECK(!rc, #expr), 0)) +#define CHECK(test, msg) ((test) ? (void)0 : ((void)fprintf(stderr, \ + "%s:%d: %s: %s\n", __FILE__, __LINE__, msg, mdb_strerror(rc)), abort())) + +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, *cur2; + MDB_cursor_op op; + int count; + int *values; + char sval[32] = ""; + + srand(time(NULL)); + + count = (rand()%384) + 64; + values = (int *)malloc(count*sizeof(int)); + + for(i = 0;i<count;i++) { + values[i] = rand()%1024; + } + + E(mdb_env_create(&env)); + E(mdb_env_set_mapsize(env, 10485760)); + E(mdb_env_open(env, "./testdb", MDB_FIXEDMAP /*|MDB_NOSYNC*/, 0664)); + E(mdb_txn_begin(env, NULL, 0, &txn)); + E(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]); + if (RES(MDB_KEYEXIST, mdb_put(txn, dbi, &key, &data, MDB_NOOVERWRITE))) { + j++; + data.mv_size = sizeof(sval); + data.mv_data = sval; + } + } + if (j) printf("%d duplicates skipped\n", j); + E(mdb_txn_commit(txn)); + E(mdb_env_stat(env, &mst)); + + E(mdb_txn_begin(env, NULL, 1, &txn)); + E(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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + mdb_cursor_close(cursor); + mdb_txn_abort(txn); + + j=0; + key.mv_data = sval; + for (i= count - 1; i > -1; i-= (rand()%5)) { + j++; + txn=NULL; + E(mdb_txn_begin(env, NULL, 0, &txn)); + sprintf(sval, "%03x ", values[i]); + if (RES(MDB_NOTFOUND, mdb_del(txn, dbi, &key, NULL))) { + j--; + mdb_txn_abort(txn); + } else { + E(mdb_txn_commit(txn)); + } + } + free(values); + printf("Deleted %d values\n", j); + + E(mdb_env_stat(env, &mst)); + E(mdb_txn_begin(env, NULL, 1, &txn)); + E(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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + printf("Cursor last\n"); + E(mdb_cursor_get(cursor, &key, &data, MDB_LAST)); + 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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + printf("Cursor last/prev\n"); + E(mdb_cursor_get(cursor, &key, &data, MDB_LAST)); + printf("key: %.*s, data: %.*s\n", + (int) key.mv_size, (char *) key.mv_data, + (int) data.mv_size, (char *) data.mv_data); + E(mdb_cursor_get(cursor, &key, &data, MDB_PREV)); + printf("key: %.*s, data: %.*s\n", + (int) key.mv_size, (char *) key.mv_data, + (int) data.mv_size, (char *) data.mv_data); + + mdb_txn_abort(txn); + + printf("Deleting with cursor\n"); + E(mdb_txn_begin(env, NULL, 0, &txn)); + E(mdb_cursor_open(txn, dbi, &cur2)); + for (i=0; i<50; i++) { + if (RES(MDB_NOTFOUND, mdb_cursor_get(cur2, &key, &data, MDB_NEXT))) + break; + 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); + E(mdb_del(txn, dbi, &key, NULL)); + } + + printf("Restarting cursor in txn\n"); + for (op=MDB_FIRST, i=0; i<=32; op=MDB_NEXT, i++) { + if (RES(MDB_NOTFOUND, mdb_cursor_get(cur2, &key, &data, op))) + break; + 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); + E(mdb_txn_commit(txn)); + + printf("Restarting cursor outside txn\n"); + E(mdb_txn_begin(env, NULL, 0, &txn)); + E(mdb_cursor_open(txn, dbi, &cursor)); + for (op=MDB_FIRST, i=0; i<=32; op=MDB_NEXT, i++) { + if (RES(MDB_NOTFOUND, mdb_cursor_get(cursor, &key, &data, op))) + break; + 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_close(env, dbi); + + mdb_txn_abort(txn); + mdb_env_close(env); + + return 0; +} diff --git a/db_drivers/liblmdb/mtest2.c b/db_drivers/liblmdb/mtest2.c new file mode 100644 index 000000000..f1a3dbd6c --- /dev/null +++ b/db_drivers/liblmdb/mtest2.c @@ -0,0 +1,123 @@ +/* mtest2.c - memory-mapped database tester/toy */ +/* + * Copyright 2011-2014 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 */ + +#include <stdio.h> +#include <stdlib.h> +#include <time.h> +#include "lmdb.h" + +#define E(expr) CHECK((rc = (expr)) == MDB_SUCCESS, #expr) +#define RES(err, expr) ((rc = expr) == (err) || (CHECK(!rc, #expr), 0)) +#define CHECK(test, msg) ((test) ? (void)0 : ((void)fprintf(stderr, \ + "%s:%d: %s: %s\n", __FILE__, __LINE__, msg, mdb_strerror(rc)), abort())) + +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] = ""; + + srand(time(NULL)); + + count = (rand()%384) + 64; + values = (int *)malloc(count*sizeof(int)); + + for(i = 0;i<count;i++) { + values[i] = rand()%1024; + } + + E(mdb_env_create(&env)); + E(mdb_env_set_mapsize(env, 10485760)); + E(mdb_env_set_maxdbs(env, 4)); + E(mdb_env_open(env, "./testdb", MDB_FIXEDMAP|MDB_NOSYNC, 0664)); + E(mdb_txn_begin(env, NULL, 0, &txn)); + E(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]); + if (RES(MDB_KEYEXIST, mdb_put(txn, dbi, &key, &data, MDB_NOOVERWRITE))) + j++; + } + if (j) printf("%d duplicates skipped\n", j); + E(mdb_txn_commit(txn)); + E(mdb_env_stat(env, &mst)); + + E(mdb_txn_begin(env, NULL, 1, &txn)); + E(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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + mdb_cursor_close(cursor); + mdb_txn_abort(txn); + + j=0; + key.mv_data = sval; + for (i= count - 1; i > -1; i-= (rand()%5)) { + j++; + txn=NULL; + E(mdb_txn_begin(env, NULL, 0, &txn)); + sprintf(sval, "%03x ", values[i]); + if (RES(MDB_NOTFOUND, mdb_del(txn, dbi, &key, NULL))) { + j--; + mdb_txn_abort(txn); + } else { + E(mdb_txn_commit(txn)); + } + } + free(values); + printf("Deleted %d values\n", j); + + E(mdb_env_stat(env, &mst)); + E(mdb_txn_begin(env, NULL, 1, &txn)); + E(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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + 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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + mdb_cursor_close(cursor); + mdb_close(env, dbi); + + mdb_txn_abort(txn); + mdb_env_close(env); + + return 0; +} diff --git a/db_drivers/liblmdb/mtest3.c b/db_drivers/liblmdb/mtest3.c new file mode 100644 index 000000000..f705c52de --- /dev/null +++ b/db_drivers/liblmdb/mtest3.c @@ -0,0 +1,133 @@ +/* mtest3.c - memory-mapped database tester/toy */ +/* + * Copyright 2011-2014 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 */ +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <time.h> +#include "lmdb.h" + +#define E(expr) CHECK((rc = (expr)) == MDB_SUCCESS, #expr) +#define RES(err, expr) ((rc = expr) == (err) || (CHECK(!rc, #expr), 0)) +#define CHECK(test, msg) ((test) ? (void)0 : ((void)fprintf(stderr, \ + "%s:%d: %s: %s\n", __FILE__, __LINE__, msg, mdb_strerror(rc)), abort())) + +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)]; + + srand(time(NULL)); + + memset(sval, 0, sizeof(sval)); + + count = (rand()%384) + 64; + values = (int *)malloc(count*sizeof(int)); + + for(i = 0;i<count;i++) { + values[i] = rand()%1024; + } + + E(mdb_env_create(&env)); + E(mdb_env_set_mapsize(env, 10485760)); + E(mdb_env_set_maxdbs(env, 4)); + E(mdb_env_open(env, "./testdb", MDB_FIXEDMAP|MDB_NOSYNC, 0664)); + E(mdb_txn_begin(env, NULL, 0, &txn)); + E(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]); + if (RES(MDB_KEYEXIST, mdb_put(txn, dbi, &key, &data, MDB_NODUPDATA))) + j++; + } + if (j) printf("%d duplicates skipped\n", j); + E(mdb_txn_commit(txn)); + E(mdb_env_stat(env, &mst)); + + E(mdb_txn_begin(env, NULL, 1, &txn)); + E(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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + mdb_cursor_close(cursor); + mdb_txn_abort(txn); + + j=0; + + for (i= count - 1; i > -1; i-= (rand()%5)) { + j++; + txn=NULL; + E(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; + if (RES(MDB_NOTFOUND, mdb_del(txn, dbi, &key, &data))) { + j--; + mdb_txn_abort(txn); + } else { + E(mdb_txn_commit(txn)); + } + } + free(values); + printf("Deleted %d values\n", j); + + E(mdb_env_stat(env, &mst)); + E(mdb_txn_begin(env, NULL, 1, &txn)); + E(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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + 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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + mdb_cursor_close(cursor); + mdb_close(env, dbi); + + mdb_txn_abort(txn); + mdb_env_close(env); + + return 0; +} diff --git a/db_drivers/liblmdb/mtest4.c b/db_drivers/liblmdb/mtest4.c new file mode 100644 index 000000000..da5a95304 --- /dev/null +++ b/db_drivers/liblmdb/mtest4.c @@ -0,0 +1,168 @@ +/* mtest4.c - memory-mapped database tester/toy */ +/* + * Copyright 2011-2014 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 */ +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <time.h> +#include "lmdb.h" + +#define E(expr) CHECK((rc = (expr)) == MDB_SUCCESS, #expr) +#define RES(err, expr) ((rc = expr) == (err) || (CHECK(!rc, #expr), 0)) +#define CHECK(test, msg) ((test) ? (void)0 : ((void)fprintf(stderr, \ + "%s:%d: %s: %s\n", __FILE__, __LINE__, msg, mdb_strerror(rc)), abort())) + +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; + } + + E(mdb_env_create(&env)); + E(mdb_env_set_mapsize(env, 10485760)); + E(mdb_env_set_maxdbs(env, 4)); + E(mdb_env_open(env, "./testdb", MDB_FIXEDMAP|MDB_NOSYNC, 0664)); + E(mdb_txn_begin(env, NULL, 0, &txn)); + E(mdb_open(txn, "id4", 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]); + if (RES(MDB_KEYEXIST, mdb_put(txn, dbi, &key, &data, MDB_NODUPDATA))) + j++; + } + if (j) printf("%d duplicates skipped\n", j); + E(mdb_txn_commit(txn)); + E(mdb_env_stat(env, &mst)); + + /* there should be one full page of dups now. + */ + E(mdb_txn_begin(env, NULL, 1, &txn)); + E(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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + 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); + E(mdb_txn_begin(env, NULL, 0, &txn)); + (void)RES(MDB_KEYEXIST, mdb_put(txn, dbi, &key, &data, MDB_NODUPDATA)); + mdb_txn_abort(txn); + + sprintf(sval, "%07x", values[255]+1); + E(mdb_txn_begin(env, NULL, 0, &txn)); + (void)RES(MDB_KEYEXIST, mdb_put(txn, dbi, &key, &data, MDB_NODUPDATA)); + mdb_txn_abort(txn); + + sprintf(sval, "%07x", values[500]+1); + E(mdb_txn_begin(env, NULL, 0, &txn)); + (void)RES(MDB_KEYEXIST, mdb_put(txn, dbi, &key, &data, MDB_NODUPDATA)); + E(mdb_txn_commit(txn)); + + /* Try MDB_NEXT_MULTIPLE */ + E(mdb_txn_begin(env, NULL, 0, &txn)); + E(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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + mdb_cursor_close(cursor); + mdb_txn_abort(txn); + j=0; + + for (i= count - 1; i > -1; i-= (rand()%3)) { + j++; + txn=NULL; + E(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; + if (RES(MDB_NOTFOUND, mdb_del(txn, dbi, &key, &data))) { + j--; + mdb_txn_abort(txn); + } else { + E(mdb_txn_commit(txn)); + } + } + free(values); + printf("Deleted %d values\n", j); + + E(mdb_env_stat(env, &mst)); + E(mdb_txn_begin(env, NULL, 1, &txn)); + E(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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + 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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + mdb_cursor_close(cursor); + mdb_close(env, dbi); + + mdb_txn_abort(txn); + mdb_env_close(env); + + return 0; +} diff --git a/db_drivers/liblmdb/mtest5.c b/db_drivers/liblmdb/mtest5.c new file mode 100644 index 000000000..39a8c728a --- /dev/null +++ b/db_drivers/liblmdb/mtest5.c @@ -0,0 +1,135 @@ +/* mtest5.c - memory-mapped database tester/toy */ +/* + * Copyright 2011-2014 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 */ +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <time.h> +#include "lmdb.h" + +#define E(expr) CHECK((rc = (expr)) == MDB_SUCCESS, #expr) +#define RES(err, expr) ((rc = expr) == (err) || (CHECK(!rc, #expr), 0)) +#define CHECK(test, msg) ((test) ? (void)0 : ((void)fprintf(stderr, \ + "%s:%d: %s: %s\n", __FILE__, __LINE__, msg, mdb_strerror(rc)), abort())) + +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)]; + + srand(time(NULL)); + + memset(sval, 0, sizeof(sval)); + + count = (rand()%384) + 64; + values = (int *)malloc(count*sizeof(int)); + + for(i = 0;i<count;i++) { + values[i] = rand()%1024; + } + + E(mdb_env_create(&env)); + E(mdb_env_set_mapsize(env, 10485760)); + E(mdb_env_set_maxdbs(env, 4)); + E(mdb_env_open(env, "./testdb", MDB_FIXEDMAP|MDB_NOSYNC, 0664)); + E(mdb_txn_begin(env, NULL, 0, &txn)); + E(mdb_open(txn, "id2", MDB_CREATE|MDB_DUPSORT, &dbi)); + E(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]); + if (RES(MDB_KEYEXIST, mdb_cursor_put(cursor, &key, &data, MDB_NODUPDATA))) + j++; + } + if (j) printf("%d duplicates skipped\n", j); + mdb_cursor_close(cursor); + E(mdb_txn_commit(txn)); + E(mdb_env_stat(env, &mst)); + + E(mdb_txn_begin(env, NULL, 1, &txn)); + E(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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + mdb_cursor_close(cursor); + mdb_txn_abort(txn); + + j=0; + + for (i= count - 1; i > -1; i-= (rand()%5)) { + j++; + txn=NULL; + E(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; + if (RES(MDB_NOTFOUND, mdb_del(txn, dbi, &key, &data))) { + j--; + mdb_txn_abort(txn); + } else { + E(mdb_txn_commit(txn)); + } + } + free(values); + printf("Deleted %d values\n", j); + + E(mdb_env_stat(env, &mst)); + E(mdb_txn_begin(env, NULL, 1, &txn)); + E(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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + 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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + mdb_cursor_close(cursor); + mdb_close(env, dbi); + + mdb_txn_abort(txn); + mdb_env_close(env); + + return 0; +} diff --git a/db_drivers/liblmdb/mtest6.c b/db_drivers/liblmdb/mtest6.c new file mode 100644 index 000000000..07d575893 --- /dev/null +++ b/db_drivers/liblmdb/mtest6.c @@ -0,0 +1,137 @@ +/* mtest6.c - memory-mapped database tester/toy */ +/* + * Copyright 2011-2014 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 */ +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <time.h> +#include "lmdb.h" + +#define E(expr) CHECK((rc = (expr)) == MDB_SUCCESS, #expr) +#define RES(err, expr) ((rc = expr) == (err) || (CHECK(!rc, #expr), 0)) +#define CHECK(test, msg) ((test) ? (void)0 : ((void)fprintf(stderr, \ + "%s:%d: %s: %s\n", __FILE__, __LINE__, msg, mdb_strerror(rc)), abort())) + +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; + + srand(time(NULL)); + + E(mdb_env_create(&env)); + E(mdb_env_set_mapsize(env, 10485760)); + E(mdb_env_set_maxdbs(env, 4)); + E(mdb_env_open(env, "./testdb", MDB_FIXEDMAP|MDB_NOSYNC, 0664)); + E(mdb_txn_begin(env, NULL, 0, &txn)); + E(mdb_open(txn, "id6", MDB_CREATE|MDB_INTEGERKEY, &dbi)); + E(mdb_cursor_open(txn, dbi, &cursor)); + E(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); + (void)RES(MDB_KEYEXIST, 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); + (void)RES(MDB_KEYEXIST, 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); + (void)RES(MDB_KEYEXIST, mdb_cursor_put(cursor, &key, &data, MDB_NOOVERWRITE)); + } + E(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); + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + mdb_cursor_close(cursor); + mdb_txn_commit(txn); + +#if 0 + j=0; + + for (i= count - 1; i > -1; i-= (rand()%5)) { + j++; + txn=NULL; + E(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; + if (RES(MDB_NOTFOUND, mdb_del(txn, dbi, &key, &data))) { + j--; + mdb_txn_abort(txn); + } else { + E(mdb_txn_commit(txn)); + } + } + free(values); + printf("Deleted %d values\n", j); + + E(mdb_env_stat(env, &mst)); + E(mdb_txn_begin(env, NULL, 1, &txn)); + E(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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + 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); + } + CHECK(rc == MDB_NOTFOUND, "mdb_cursor_get"); + mdb_cursor_close(cursor); + mdb_close(env, dbi); + + mdb_txn_abort(txn); +#endif + mdb_env_close(env); + + return 0; +} diff --git a/db_drivers/liblmdb/sample-bdb.txt b/db_drivers/liblmdb/sample-bdb.txt new file mode 100644 index 000000000..6a959bd68 --- /dev/null +++ b/db_drivers/liblmdb/sample-bdb.txt @@ -0,0 +1,73 @@ +/* sample-bdb.txt - BerkeleyDB toy/sample + * + * Do a line-by-line comparison of this and sample-mdb.txt + */ +/* + * Copyright 2012 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 <string.h> +#include <db.h> + +int main(int argc,char * argv[]) +{ + int rc; + DB_ENV *env; + DB *dbi; + DBT key, data; + DB_TXN *txn; + DBC *cursor; + char sval[32], kval[32]; + + /* Note: Most error checking omitted for simplicity */ + +#define FLAGS (DB_INIT_LOCK|DB_INIT_LOG|DB_INIT_TXN|DB_INIT_MPOOL|DB_CREATE|DB_THREAD) + rc = db_env_create(&env, 0); + rc = env->open(env, "./testdb", FLAGS, 0664); + rc = db_create(&dbi, env, 0); + rc = env->txn_begin(env, NULL, &txn, 0); + rc = dbi->open(dbi, txn, "test.bdb", NULL, DB_BTREE, DB_CREATE, 0664); + + memset(&key, 0, sizeof(DBT)); + memset(&data, 0, sizeof(DBT)); + key.size = sizeof(int); + key.data = sval; + data.size = sizeof(sval); + data.data = sval; + + sprintf(sval, "%03x %d foo bar", 32, 3141592); + rc = dbi->put(dbi, txn, &key, &data, 0); + rc = txn->commit(txn, 0); + if (rc) { + fprintf(stderr, "txn->commit: (%d) %s\n", rc, db_strerror(rc)); + goto leave; + } + rc = env->txn_begin(env, NULL, &txn, 0); + rc = dbi->cursor(dbi, txn, &cursor, 0); + key.flags = DB_DBT_USERMEM; + key.data = kval; + key.ulen = sizeof(kval); + data.flags = DB_DBT_USERMEM; + data.data = sval; + data.ulen = sizeof(sval); + while ((rc = cursor->c_get(cursor, &key, &data, DB_NEXT)) == 0) { + printf("key: %p %.*s, data: %p %.*s\n", + key.data, (int) key.size, (char *) key.data, + data.data, (int) data.size, (char *) data.data); + } + rc = cursor->c_close(cursor); + rc = txn->abort(txn); +leave: + rc = dbi->close(dbi, 0); + rc = env->close(env, 0); + return rc; +} diff --git a/db_drivers/liblmdb/sample-mdb.txt b/db_drivers/liblmdb/sample-mdb.txt new file mode 100644 index 000000000..a233ec5dd --- /dev/null +++ b/db_drivers/liblmdb/sample-mdb.txt @@ -0,0 +1,62 @@ +/* sample-mdb.txt - MDB toy/sample + * + * Do a line-by-line comparison of this and sample-bdb.txt + */ +/* + * Copyright 2012 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 "lmdb.h" + +int main(int argc,char * argv[]) +{ + int rc; + MDB_env *env; + MDB_dbi dbi; + MDB_val key, data; + MDB_txn *txn; + MDB_cursor *cursor; + char sval[32]; + + /* Note: Most error checking omitted for simplicity */ + + rc = mdb_env_create(&env); + rc = mdb_env_open(env, "./testdb", 0, 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; + + sprintf(sval, "%03x %d foo bar", 32, 3141592); + rc = mdb_put(txn, dbi, &key, &data, 0); + rc = mdb_txn_commit(txn); + if (rc) { + fprintf(stderr, "mdb_txn_commit: (%d) %s\n", rc, mdb_strerror(rc)); + goto leave; + } + rc = mdb_txn_begin(env, NULL, MDB_RDONLY, &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); +leave: + mdb_close(env, dbi); + mdb_env_close(env); + return 0; +} diff --git a/db_drivers/liblmdb/tooltag b/db_drivers/liblmdb/tooltag new file mode 100644 index 000000000..229bf16ba --- /dev/null +++ b/db_drivers/liblmdb/tooltag @@ -0,0 +1,22 @@ +<tagfile> + <compound kind="page"> + <name>mdb_copy_1</name> + <title>mdb_copy - environment copy tool</title> + <filename>mdb_copy.1</filename> + </compound> + <compound kind="page"> + <name>mdb_dump_1</name> + <title>mdb_dump - environment export tool</title> + <filename>mdb_dump.1</filename> + </compound> + <compound kind="page"> + <name>mdb_load_1</name> + <title>mdb_load - environment import tool</title> + <filename>mdb_load.1</filename> + </compound> + <compound kind="page"> + <name>mdb_stat_1</name> + <title>mdb_stat - environment status tool</title> + <filename>mdb_stat.1</filename> + </compound> +</tagfile> |