\input texinfo @c -*-texinfo-*- @c %**start of header @setfilename libcdio.info @include version.texi @settitle GNU @code{libcdio}: Compact Disc Input, Output, and Control Library @c %**end of header @c Karl Berry informs me that this will add straight quotes in @c typewriter text. @c See the "Inserting Quote Characters" node in the Texinfo manual @set txicodequoteundirected @set txicodequotebacktick @copying This manual documents @code{libcdio}, the GNU CD Input, Output, and Control Library. Copyright @copyright{} 2003, 2004, 2005, 2006, 2007, 2008, 2010 Rocky Bernstein and Herbert Valerio Riedel. @quotation Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation; with no Invariant Sections, with no Front-Cover Texts, and with no Back-Cover Texts. A copy of the license is included in the section entitled ``GNU Free Documentation License''. @end quotation @end copying @paragraphindent 0 @exampleindent 0 @set libcdio @code{libcdio} @set program @kbd{libcdio} @c A macro for defining terms variables. @macro term{varname} @c @cindex{\varname\} @emph{\varname\} @end macro @dircategory Software libraries @direntry * libcdio: (libcdio). GNU Compact Disc Input, Output, and Control Library. @end direntry @titlepage @title GNU @code{libcdio} @subtitle GNU Compact Disc Input, Output, and Control Library @subtitle for version @value{VERSION}, @value{UPDATED} @author Rocky Bernstein et al. (@email{bug-libcdio@@gnu.org}) @page @vskip 0pt plus 1filll @insertcopying @end titlepage @contents @ifnottex @node Top @top GNU @value{libcdio} @insertcopying @menu * History:: How this came about * Previous Work:: The problem and previous work * Purpose:: What is in this package (and what's not) * CD Formats:: A tour through the CD-specification spectrum * CD Image Formats:: A tour through various CD-image formats * CD Units:: The units that make up a CD * How to use:: Okay enough babble, lemme at the library! * Utility Programs:: Diagnostic programs that come with this library * CD-ROM Access and Drivers:: CD-ROM access and drivers * Internal Program Organization:: Looking under the hood Appendices * ISO-9660 Character Sets:: * Glossary:: * GNU Free Documentation License:: Indices * General Index:: Overall index @end menu @end ifnottex @node History @chapter History As a result of the repressive Digital Millennium Copyright Act (DMCA) I became aware of Video CD's (VCD's). Video CD's are not subject to the DMCA and therefore enjoy the protection afforded by copyright but no more. But in order for VCD's to be competitive with DVD's, good tools (including GPL tools) are needed for authoring and playing them. And so through VCD's I became aware of the excellent Video CD tools by Herbert Valerio Riedel which form the @kbd{vcdimager} package. Although vcdimager is great for authoring, examining and extracting parts of a Video CD, it is not a VCD player. And when I looked at the state of Video CD handling in existing VCD players: @code{xine}, @code{MPlayer}, and @code{vlc}, I was a bit disappointed. None handled playback control, menu selections, or playing still frames and segments from track 1. Version 0.7.12 of vcdimager was very impressive, however it lacked exportable libraries that could be used in other projects. So with the blessing and encouragement of Herbert Valerio Riedel, I took to extract and create libraries from this code base. The result was two libraries: one to extract information from a VCD which I called libvcdinfo, and another to do the reading and control of a VCD. Well, actually, at this point I should say that a Video CD is really just Video put on a existing well-established Compact Disc or CD format. So the library for this is called @value{libcdio} rather than @kbd{libvcdio}. While on the topic of the name @value{libcdio}, I should also explain that the library really doesn't handle writing or output (the final "o" in the name). However it was felt that if I put @code{libcdi} that might be confused with a particular CD format called CD-I. Later on, the ISO-9660 filesystem handling component from @kbd{vcdimager} was extracted, expanded and made a separate library. Next the ability to add MMC commands was added, and then CD paranoia support. And from there, the rest is history. @node Previous Work @chapter The problem and previous work If around the year 2002 you were to look at the code for a number of free software CD or media players that work on several platforms such as vlc, MPlayer, xine, or xmms to name but a few, you'd find the code to read a CD sprinkled with conditional compilation for this or that platform. That is there was @emph{no} OS-independent programmer library for CD reading and control even though the technology was over 10 years old; yet there are media players which strive for OS independence. One early CD player, @kbd{xmcd} by Ti Kan, was I think a bit better than most in that it tried to @emph{encapsulate} the kinds of CD control mechanisms (SCSI, Linux ioctl, Toshiba, etc.) in a "CD Audio Device Interface Library" called @code{libdi}. However this library is for Audio CD's only and I don't believe this library has been used outside of xmcd. Another project, Simple DirectMedia Layer also encapsulates CD reading. @quotation SDL is a library that allows you portable low-level access to a video framebuffer, audio output, mouse, and keyboard. With SDL, it is easy to write portable games which run on ... @end quotation Many of the media players mentioned above do in fact can make use of the SDL library but for @emph{video} output only. Because the encapsulation is over @emph{many} kinds of I/O (video, joysticks, mice, as well as CD's), I believe that the level of control provided for CD a little bit limited. (However to be fair, it may have only been intended for games and may be suitable for that). Applications that just want the CD reading and control portion I think will find quite a bit overhead. Another related project is J@"org Schilling's SCSI library. You can use that to make a non-SCSI CD-ROM act like one that understands SCSI MMC commands which is a neat thing to do. However it is a little weird to have to install drivers just so you can run a particular user-level program. Installing drivers often requires special privileges and permissions and it is pervasive on a system. It is a little sad that along the way to creating such a SCSI library a library similar to @value{libcdio} wasn't created which could be used. Were that the case, this library certainly never would have been written. At the OS level there is the ``A Linux CD-ROM Standard'' by David van Leeuwen from around 1999. This defines a set of definitions and ioctl's that mask hardware differences of various Compact Disc hardware. It is a great idea, however this ``standard'' lacked adoption on OS's other than GNU/Linux. Or maybe it's the case that the standard on other OS's lacked adoption on GNU/Linux. For example on FreeBSD there is a ``Common Access Method'' (CAM) used for all SCSI access which seems not to be adopted in GNU/Linux.@footnote{And I'm thankful for that since, at least for MMC commands, it is inordinately complicated and in some places arcane.} Finally at the hardware level where a similar chaos exists, there has been an attempt to do something similar with the MMC (multimedia commands). This attempts to provide a uniform command set for CD devices like PostScript does for printer commands.@footnote{I wrote ``attempts'' because over time the command set has changed and now there are several different commands to do a particular function like read a CD table of contents and some hardware understands some of the version of the commands set but might not others} In contrast to PostScript where there one in theory can write a PostScript program in a uniform ASCII representation and send that to a printer, for MMC although there are common internal structures defined, there is no common syntax for representing the structures or an OS-independent library or API for issuing MMC-commands which a programmer would need to use. Instead each Operating System has its own interface. For example Adaptec's ASPI or Microsoft's DeviceIoControl on Microsoft Windows, or IOKit for Apple's OS/X, or FreeBSD's CAM. I've been positively awed at how many different variations and differing levels of complexity there are for doing basically the same thing. How easy it is to issue an MMC command from a program varies from easy to very difficult. And mastering the boilerplate code to issue an MMC command on one OS really doesn't help much in figuring out how to do it on another OS. So in @value{libcdio} we provide a common (and hopefully simple) API to issue MMC commands. @node Purpose @chapter What is in this package (and what's not) The library, @command{libcdio}, encapsulates CD-ROM reading and control. Applications wishing to be oblivious of the OS- and device-dependent properties of a CD-ROM can use this library. Also included is a library, @command{libiso9660}, for working with ISO-9660 filesystems, @command{libcdio_paranoia}, and @command{libcdio_cdda} libraries for applications which want to use cdparanoia's error-correction and jitter detection. Some support for disk-image types like cdrdao's TOC, CDRWIN's BIN/CUE and Ahead Nero's NRG format is available, so applications that use this library also have the ability to read disc images as though they were CDs. @command{libcdio} also provides a way to issue SCSI ``MultiMedia Commands'' (MMC). MMC is supported by many hardware CD-ROM manufacturers; and in some cases where a CD-ROM doesn't understand MMC directly, some Operating Systems (such as GNU/Linux, Solaris, or FreeBSD or Microsoft Windows ASPI to name a few) provide the MMC emulation.@footnote{This concept of software emulation of a common hardware command language is common for printers such as using ghostscript to private postscript emulation for a non-postscript printer.} The first use of the library in this package are the Video CD authoring and ripping tools, VCDImager (@url{http://vcdimager.org}). See @url{http://www.gnu.org/software/libcdio/projects.html} for a list of projects using @command{libcdio}. A version of the CD-DA extraction tool cdparanoia (@url{http://www.xiph.org/paranoia} and its library which corrects for CD-ROM jitter are part of the distribution. Also included in the libcdio package is a utility program @command{cd-info} which displays CD information: number of tracks, CD-format and if possible basic information about the format. If libcddb (@url{http://libcddb.sourceforge.net}) is available, the @command{cd-info} program will display CDDB matches on CD-DA discs. And if a new enough version of libvcdinfo is available (from the vcdimager project), then @command{cd-info} shows basic VCD information. Other utility programs in the libcdio package are: @table @code @item @code{cdda-player} shows off @value{libcdio} audio and CD-ROM control commands. It can play a track, eject or load media and show the the status of a CD-DA that is might be currently played via the audio control commands. It can be run in batch mode or has a simple curses-based interface. If libcddb is available or a CD has CD-Text and your CD-ROM drive supports CD-Text, track/album information about the CD can be shown. @item @code{cd-drive} shows what drivers are available and some basic properties of cd-drives attached to the system. (But media may have to be inserted in order to get this info.) lists out drive capabilities @item cd-read performs low-level block reading of a CD or CD image, @item @code{iso-info} displays ISO-9660 information from an ISO-9660 image. Below is some sample output @smallexample iso-info version 0.82 x86_64-unknown-linux-gnu Copyright (c) 2003, 2004, 2005, 2007, 2008 R. Bernstein This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. __________________________________ ISO 9660 image: ../test/joliet.iso Application: K3B THE CD KREATOR VERSION 0.11.12 (C) 2003 SEBASTIAN TRUEG AND... Preparer : K3b - Version 0.11.12 Publisher : Rocky Bernstein System : LINUX Volume : K3b data project Volume Set : __________________________________ ISO-9660 Information /: Oct 22 2004 19:44 . Oct 22 2004 19:44 .. Oct 22 2004 19:44 libcdio /libcdio/: Oct 22 2004 19:44 . Oct 22 2004 19:44 .. Mar 12 2004 02:18 COPYING Jun 26 2004 07:01 README Aug 12 2004 06:22 README.libcdio Oct 22 2004 19:44 test /libcdio/test/: Oct 22 2004 19:44 . Oct 22 2004 19:44 .. Jul 25 2004 06:52 isofs-m1.cue @end smallexample @item @code{iso-read} extracts files from an ISO-9660 image. @end table Historically, @code{libcdio} did not support write access to drives. In conjunction with additional work in a separate project @code{libburn}, Thomas Schmitt has modified @code{libcdio} to enable sending SCSI write commands on some of the drivers. This enables other programs like @code{libburn} to write to CD's, DVD's and Blu-Ray discs. For the OS drivers which are lacking write access, volunteers are welcome. @node CD Formats @chapter CD Formats Much of what I write in this section can be found elsewhere. See for example @url{http://www.pctechguide.com/08cd-rom.htm} or @url{http://www.pcguide.com/ref/cd/format.htm} We give just enough background here to cover Compact Discs and Compact Disc formats that are handled by this library. The Sony and Philips Corporations invented and Compact Disc (CD) in the early 1980s. The specifications for the layout is often referred to by the color of the cover on the specification. @menu * Red Book:: Red Book (CD-DA) CD Text, CDDB * Yellow Book:: Yellow Book (CD-ROM Digital Data) * Green Book:: Green Book (CD-i) * White Book:: White Book (DV, Video CD) @end menu @node Red Book @section Red Book (CD-DA) @cindex Red Book @menu * CD Text:: CD Text and CD+G * CDDB:: Internet CD Database (CDDB) @end menu The first type of CD that was produced was the Compact Disc Digital Audio (CD-DA) or just plain ``audio CD''. The specification, ICE 60908 (formerly IEC 908) is commonly called the ``Red Book'', @cite{@url{http://en.wikipedia.org/wiki/Red_Book_(audio_CD_standard)}}. Music CD's are recorded in this format which basically allows for around 74 minutes of audio per disc and for that information to be split up into tracks. Tracks are broken up into "sectors" and each sector contains up to 2,352 bytes. To play one 44.1 kHz CD-DA sampled audio second, 75 sectors are used. The minute/second/frame numbering of sectors or MSF format is based on the fact that 75 sectors are used in a second of playing of sound. (And for almost every other CD format and application the MSF format doesn't make that much sense). In @value{libcdio} when you you want to read an audio sector, you call @code{cdio_read_audio_sector()} or @code{cdio_read_audio_sectors()}. @cindex subchannel In addition the the audio data ``channel'' a provision for other information or @term{subchannel} information) can be stored in a sector. Other subchannels include a Media Catalog Number (also abbreviated as MCN and sometimes a UPC), or album meta data (also called CD-Text). Karioke graphics can also be stored in a format called @term{CD+G}. @node CD Text @subsection CD Text, CD+G @cindex CD Text @cindex CD+G CD Text is an extension to the CD-DA standard that adds the ability to album and track meta data (titles, artist/performer names, song titles) and graphical (e.g. Karaoke) information. For an alternative way to get album and track meta-data see @xref{CDDB}. Information is stored in such a way that it doesn't interfere with the normal operation of any CD players or CDROM drives. There are two different parts of the CD where the data can be stored. The first place the information can be recorded is in the R-W sub codes in the lead in area of the CD. This information is stored as a single block of data and is the format. The method for reading this data from a CDROM drive is covered under the Sony proposal to the MMC specification. The format of the data is partially covered in the MMC specification. CD Text information is stored in this area. The format that follows the Interactive Text Transmission System (ITTS) is the same data transmission standard used by such things as Digital Audio Broadcasting (DAB), and virtually the same as the data standard for the MiniDisc. The second place the information can be recorded is in the R-W sub codes in the program area of the CD giving a data capacity of roughly 31MB. CD+G (CD w/graphics) uses this method. The methods for reading this data from a CD-ROM drive were first covered by the programming specs from the individual drive manufacturers. In the case of ATAPI drives, the SFF8020 spec covers the reading of the RW subcodes. Subsequently it has been encorporated into the MMC specifications. Not all drives support reading the RW subcodes from the program area. However for those that do, @value{libcdio} provides a way to get at this information via @code{cdtext_get()} and its friends. There is a separate document in this distribution describing CD-Text information and how it is encoded. @node CDDB @subsection Internet CD Database (CDDB) @cindex CDDB CDDB is an database on the Internet of of CD album/track, artist, and genre information similar to CD Text information. Using track information (number of tracks and length of the tracks), devices that have access to the Internet can query for meta information and contribute information for CD's where there is no existing information. When storage is available (such as you'd expect for any program using @value{libcdio}, the information is often saved for later use when the Internet is not available; people tend request the same information since they via programs play the same music. Obtaining CD meta information when none is encoded in an audio CD is useful in media players or making ones own compilations from audio CDs. There are currently two popular CDDB services on the Internet. The original database has been renamed Gracenote and is a profit making entity. FreeDB (@url{http://freedb.org} is an open source CD information resource that is free for developers and the public to use. As there already is an excellent library for handling CDDB libcddb (@url{http://libcddb.sourceforge.net} we suggest using that. Our utility program @command{cd-info} will make use it if it is available and it's what we use in our applications that need it. @node Yellow Book @section Yellow Book (CD-ROM Digital Data) The CD-ROM specification or the ``Yellow Book'' followed a few years later (Standards ISO/IEC 10149), and describes the extension of CD's to store computer data, i.e. CD-ROM (Compact Disk Read Only Memory). The specification in the Yellow Book defines two modes: Mode 1 and Mode 2. @menu * ISO 9660:: * Mode 1:: Mode 1 Format * Mode 2:: Mode 2 Format @end menu @node ISO 9660 @subsection ISO 9660 @cindex ISO 9660 @menu * ISO 9660 Level 1:: * ISO 9660 Level 2:: * ISO 9660 Level 3:: * Joliet Extensions:: * Rock Ridge Extensions:: @end menu The Yellow Book doesn't specify how data is to be stored on a CD-ROM. It was feared that different companies would implement proprietary data storage formats using this specification, resulting in incompatible data CDs. To prevent this, representatives of major manufacturers met at the High Sierra Hotel and Casino in Lake Tahoe, NV, in 1985, to define a standard for storing data on CDs. This format was nicknamed High Sierra Format. In a slightly modified form it was later adopted as ISO the ISO 9660 standard. This standard is further broken down into 3 "levels", the higher the level, the more permissive. @node ISO 9660 Level 1 @subsubsection ISO 9660 Level 1 Level 1 ISO 9660 defines names in the 8+3 convention so familiar to MS-DOS: eight characters for the filename, a period, and then three characters for the file type, all in upper case. The allowed characters are A-Z, 0-9, ".", and "_".Level 1 ISO 9660 requires that files occupy a contiguous range of sectors. This allows a file to be specified with a start block and a count. The maximum directory depth is 8. For a table of the characters, see @xref{ISO-9660 Character Sets}. @node ISO 9660 Level 2 @subsubsection ISO 9660 Level 2 Level 2 ISO 9660 allows far more flexibility in filenames, but isn't usable on some systems, notably MS-DOS. @node ISO 9660 Level 3 @subsubsection ISO 9660 Level 3 Level 3 ISO-9660 allows non-contiguous files, useful if the file was written in multiple packets with packet-writing software. There have been a number of extensions to the ISO 9660 CD-ROM file format. One extension is Microsoft's Joliet specification, designed to resolve a number of deficiencies in the original ISO 9660 Level 1 file system, and in particular to support the long file names used in Windows 95 and subsequent versions of Windows. Another extension is the Rock Ridge Interchange Protocol (RRIP), which enables the recording of sufficient information to support POSIX File System semantics. @node Joliet Extensions @subsubsection Joliet Extensions @cindex Joliet extensions Joliet extensions were an upward-compatible extension to the ISO 9660 specification that removes the limitation initially put in to deal with the limited filename conventions found in Microsoft DOS OS. In particular, the Joliet specification allows for long filenames and allows for UCS-BE (BigEndian Unicode) encoding of filenames which include mixed case letter, accented characters spaces and various symbols. The way all of this is encoded is by adding a second directory and filesystem structure in addition to or in parallels to original ISO 9600 filesystem. The root node of the ISO 9660 filesystem is found via the @term{Primary Volume Descriptor} or @term{PVD}. The root of the Joliet-encode filesystem is found in a Supplementary Volume Descriptor or @term{SVD} defined in the ISO 9660 specification. The SVD structure is almost identical to a PVD with a couple of unused fields getting used and with the filename encoding changed to UCS-BE. @node Rock Ridge Extensions @subsubsection Rock Ridge Extensions @cindex Rock Ridge extensions Using the Joliet Extension one overcome the limitedness of the original ISO-9660 naming scheme. But another and probably better method is to use the Rock Ridge Extension. Not only can one store a filename as one does in a POSIX OS, but the other file attributes, such as the various timestamps (creation, modification, access), file attributes (user, group, file mode permissions, device type, symbolic links) can be stored. This is much as one would do in XA attributes; however the two are not completely interchangeable in the information they store: XA does @emph{not} address filename limitations, and the Rock Ridge extensions don't indicate if a sector is in Mode 1 or Mode 2 format. The Rock Ridge extension makes use of a hook that was defined as part of the ISO 9660 standard. @node Mode 1 @subsection Mode 1 (2048 data bytes per sector) @cindex Mode 1 Mode 1 is the data storage mode used by to store computer data. There are 3 layers of error correction. A Compact Disc using only this format can hold at most 650 MB. The data is laid out in basically the same way as in and audio CD format, except that the 2,352 bytes of data in each block are broken down further. 2,048 of these bytes are for ``real'' data. The other 304 bytes are used for an additional level of error detecting and correcting code. This is necessary because data CDs cannot tolerate the loss of a handful of bits now and then, the way audio CDs can. In @value{libcdio} when you you want to read a mode1 sector you call the @code{cdio_read_mode1_sector()} or @code{cdio_read_mode1_sectors()}. @node Mode 2 @subsection Mode 2 (2336 data bytes per sector) @cindex Mode 2 Mode 2 data CDs are the same as mode 1 CDs except that the error detecting and correcting codes are omitted. So still there are 2 layers of error correction. A Compact Disc using only this mode can thus hold at most 742 MB. Similar to audio CDs, the mode 2 format provides a more flexible vehicle for storing types of data that do not require high data integrity: for example, graphics and video can use this format. But in contrast to the Red Book standard, different modes can be mixed together; this is the basis for the extensions to the original data CD standards known as CD-ROM Extended Architecture, or CD-ROM XA. CD-ROM XA formats currently in use are CD-I Bridge formats, Photo CD and Video CD plus Sony's Playstation. In @value{libcdio} when you you want to read a mode1 sector you call the @code{cdio_read_mode2_sector()} or @code{cdio_read_mode2_sectors()}. @node Green Book @section Green Book (CD-i) @cindex Green Book This was a CD-ROM format developed by Philips for CD-i (an obsolete embedded CD-ROM application allowing limited user user interaction with films, games and educational applications). The format is ISO 9660 compliant and introduced mode 2 form 2 addressing. It also contains XA (Extended Architecture) attributes. Although some Green Book discs contain CD-i applications which can only be played on a CD-i player, others have films or music videos. Video CDs in Green-Book format are labeled "Digital Video on CD." The Green Book for video is largely superseded by White book CD-ROM which draws on this specification. @node White Book @section White Book (DV, Video CD) @cindex Green Book The White Book was released by Sony, Philips, Matsushita, and JVC in 1993, defines the Video CD specification. The White Book is also known as Digital Video (DV). A Video CD contains one data track recorded in CD-ROM XA Mode 2 Form 2. It is always the first track on the disc (Track 1). The ISO-9660 file structure and a CD-i application program are recorded in this track, as well as the Video CD Information Area which gives general information about the Video Compact Disc. After the data track, video is written in one or more subsequent tracks within the same session. These tracks are also recorded in Mode 2 Form 2. In @value{libcdio} when you you want to read a mode2 format 2 audio sector you call the @code{cdio_read_mode2_sector()} or @code{cdio_read_mode2_sectors()} setting @code{b_form2} to @code{true}. @node CD Image Formats @chapter CD Image Formats @menu * CDRDAO TOC Format:: * CDRWIN BIN/CUE Format:: * NRG Format:: @end menu In both the @command{cdrdao} and bin/cue formats there is one meta-file with extensions @code{.toc} or @code{.cue} respectively and one or more files (often with the extension @code{.bin}) which contains the content of tracks. The format of the track data is often interchangeable between the two formats. For example, in @value{libcdio}'s regression tests we make use of this to reduce the size of the test data and just provide alternate meta-data files (@code{.toc} or @code{.cue}). In contrast to the first two formats, the NRG format consists of a single file. This has the advantage of being a self-contained unit: in the other two formats it is possible for the meta file to refer to a file that can't be found. A disadvantage of the NRG format is that the meta data can't be easily viewed or modified say in a text file as it can be with the first two formats. In conjunction with this disadvantage is another disadvantage that the format is not documented, so how @value{libcdio} interprets an NRG image is based on inference. It is recommended that one of the other forms be used instead of NRG where possible. @node CDRDAO TOC Format @section CDRDAO TOC Format This is @command{cdrdao}'s CD-image description format. Since this program is GPL and everything about it is in the open, it is the preferred format to use. (Alas, at present it isn't as well supported in @value{libcdio} as the BIN/CUE format.) The @emph{toc}-file describes what data is written to the media in the @acronym{CD-ROM}; it allows control over track/index positions, pre-gaps and sub-channel information. It is a text file, so a text editor can be used to create, view or modify it. The @cite{cdrdao(1) manual page}, contains more information about this format. @subsection CDRDAO Grammar Below are the lexical tokens and grammar for a cdrdao TOC. It was taken from the cdrdao's pacct grammar; the token and nonterminal names are the same. @example #lexclass START #token Eof "@@" #token "[\t\r\ ]+" #token Comment "//~[\n@@]*" #token "\n" #token BeginString "\"" #token Integer "[0-9]+" #tokclass AudioFile @{ "AUDIOFILE" "FILE" @} #lexclass STRING #token EndString "\"" #token StringQuote "\\\"" #token StringOctal "\\[0-9][0-9][0-9]" #token String "\\" #token String "[ ]+" #token String "~[\\\n\"\t ]*" @end example @example ::= ( "CATALOG" | )* @{ @} ( )+ Eof ::= "TRACK" @{ @} ( "ISRC" | @{ "NO" @} "COPY" | @{ "NO" @} "PRE_EMPHASIS" | "TWO_CHANNEL_AUDIO" | "FOUR_CHANNEL_AUDIO" )* @{ @} @{ "PREGAP" @} ( | "START" @{ msf @} | "END" @{ msf @} )+ ( "INDEX" )* ::= AudioFile @{ "SWAP" @} @{ "#" @} | "DATAFILE" @{ "#" @{ @} @} | "FIFO" | "SILENCE" | "ZERO" @{ dataMode @} @{ @} ::= BeginString ( String | StringQuote | StringOctal )+ EndString ::= BeginString ( String | StringQuote | StringOctal )* EndString ::= Integer ::= Integer ::= Integer ":" Integer ":" Integer ::= | ::= | ::= "AUDIO" | "MODE0" | "MODE1" | "MODE1_RAW" | "MODE2" | "MODE2_RAW" | "MODE2_FORM1" | "MODE2_FORM2" | "MODE2_FORM_MIX" ::= "AUDIO" | "MODE1" | "MODE1_RAW" | "MODE2" | "MODE2_RAW" | "MODE2_FORM1" | "MODE2_FORM2" | "MODE2_FORM_MIX" ::= "RW" | "RW_RAW" ::= "CD_DA" | "CD_ROM" | "CD_ROM_XA" | "CD_I" ::= "TITLE" | "PERFORMER" | "SONGWRITER" | "COMPOSER" | "ARRANGER" | "MESSAGE" | "DISC_ID" | "GENRE" | "TOC_INFO1" | "TOC_INFO2" | "RESERVED1" | "RESERVED2" | "RESERVED3" | "RESERVED4" | "UPC_EAN" | "ISRC" | "SIZE_INFO" ::= "@{" @{ Integer ( "," Integer )* @} "@}" ::= ( | ) ::= "LANGUAGE" Integer "@{" ( )* "@}" ::= "LANGUAGE_MAP" "@{" ( Integer ":" ( Integer | "EN" ) )+ "@}" ::= "CD_TEXT" "@{" ( )* "@}" ::= "CD_TEXT" "@{" @{ @} ( )* "@}" @end example @node CDRWIN BIN/CUE Format @section CDRWIN BIN/CUE Format @cindex BIN/CUE, CD Image Format The format referred to as @emph{CDRWIN BIN/CUE Format} in this manual is a popular CD image format used in the @acronym{PC} world. Not unlike @command{cdrdao}'s TOC file, the @emph{cue} file describes the track layout, i.e. how the sectors are to be placed on the CD media. The @emph{cue} file usually contains a reference to a file traditionally having the @file{.bin} extension in its filename, the @emph{bin} file. This @emph{bin} file contains the sector data payload which is to be written to the CD medium according to the description in the @emph{cue} file. The following is an attempt to describe the subset of the @file{.cue} file syntax used in @value{libcdio} and vcdimager in an EBNF-like notation: @subsection BIN/CUE Grammar @example @cartouche ::= +( + ) ::= "0" | "1" ... "8" | "9" ::= + ::= ":" ":" ::= "FILE" ::= [ "\"" ] [ "\"" ] | "\"" "\"" ::= "BINARY" ::= [ ] [ ] * [ ] ::= "FLAGS" * ::= "DCP" ::= "TRACK" ::= "PREGAP" ::= "INDEX" ::= "POSTGAP" ::= "AUDIO" | "MODE1/2048" | "MODE1/2352" | "MODE2/2336" | "MODE2/2352" ::= "REM" * @end cartouche @end example @node NRG Format @section NRG Format @cindex Nero NRG, CD-Image format The format referred to as @emph{NRG Format} in this manual is another popular CD image format. It is available only on Nero software on a Microsoft Windows Operating System. It is proprietary and not generally published, so the information we have comes from guessing based on sample CD images. So support for this is incomplete and using this format is not recommended. Unlike @command{cdrdao}'s TOC file the BIN/CUE format everything is contained in one file. that one can edit Meta information such as the number of tracks and track format is contained at the end of the file. This information is not intended to be edited through a text editor. @node CD Units @chapter The units that make up a CD @menu * Tracks:: Tracks * Sectors:: Block addressing (MSF, LSN, LBA) * Pre-gaps:: Track pre-gaps @end menu @node Tracks @section tracks --- disc subdivisions @cindex track @cindex gaps In this section we describe CD properties and terms that we make use of in @value{libcdio}. A CD is formatted into a number of @term{tracks}, and a CD can hold at most 99 such tracks. This is defined by @code{CDIO_CD_MAX_TRACKS} in @file{cdio/sector.h}. Between some tracks CD specifications require a ``2 second'' in gap (called a @term{lead-in gap}. This is unused space with no ``data'' similar to the space between tracks on an old phonograph. The word ``second'' here really refers to a measure of space and not really necessarily an amount of time. However in the special case that the CD encodes an audio CD or CD-DA, the amount of time to play a gap of this size will take 2 seconds. @cindex lead out The beginning (or inner edge) of the CD is supposed to have a ``2 second'' lead-in gap and there is supposed to be another ``2 second'' @term{lead-out} gap at the end (or outer edge) of the CD. People have discovered that they can put useful data in the @term{lead-in} and @term{lead-out} gaps, and their equipment can read this, violating the standards but allowing a CD to store more data. In order to determine the number of tracks on a CD and where they start, commands are used to get this table-of-contents or @term{TOC} information. Asking about the start of the @term{lead-out track} gives the amount of data stored on the Compact Disk. To make it easy to specify this leadout track, special constant 0xAA (decimal 170) is used to indicate it. This is safe since this is higher than the largest legal track position. In @value{libcdio}, @code{CDIO_CDROM_LEADOUT_TRACK} is defined to be this special value. @node Sectors @section block addressing (MSF, LSN, LBA) @cindex MSF @cindex LSN @cindex LBA @cindex sectors @cindex frames A track is broken up into a number of 2352-byte @emph{blocks} which we sometimes call @emph{sectors} or @emph{frames}. Whereas tracks may have a gap between them, a block or sector does not. (In @value{libcdio} the block size constant is defined using @code{CDIO_CD_FRAMESIZE_RAW}). A Compact Disc has a limit on the number of blocks or sectors. This values is defined by constant @code{CDIO_CD_MAX_LSN} in @file{cdio/sector.h}. One can addressing a block in one of three formats. The oldest format is by it's minute/second/frame number, also referred to as @term{MSF} and written in time-like format MM:SS:FF (e.g. 30:01:40). It is best suited in audio (Red Book) applications. In @value{libcdio}, the type @code{msf_t} can be used to declare variables to hold such values. Minute, second and frame values are one byte @emph{and stored BCD notation}.@footnote{Perhaps this is a @value{libcdio} design flaw. It was originally done I guess because it was convenient for VCDs.} There are @value{libcdio} conversion routines @code{cdio_from_bcd8()} and @code{cdio_to_bcd8()} to convert the minute, second, and frame values into or out of integers. If you want to print a field in a BCD-encoded MSF, one can use the format specifier @code{%x} @emph{(not @code{%d})} and things will come out right. In the MSF notation, there are 75 ``frames'' in a ``second,'' and the familiar (if awkward) 60 seconds in a minute. @emph{Frame} here is what we called a @emph{block} above. The CD specification defines ``frame'' to be @emph{another} unit which makes up a block. Very confusing. A frame is also sometimes called a sector, analogous to hard-disk terminology. Even more confusing is using this time-like notation for an address or for a length. Too often people confuse the MSF notation this with an amount of time. A ``second'' (or @code{CDIO_CD_FRAMES_PER_SEC} blocks) in this notation is only a second of playing time for something encoded as CD-DA. It does @emph{not} necessarily represent the amount time that it will take to play a of Video CD---usually you need more blocks than this. Nor does it represent the amount of data used to play a second of an MP3---usually you need fewer blocks than this. It is also not the amount of time your CD-ROM will take to read a ``second'' of data off a Compact Disc: for example a 12x CD player will read 12x @code{CDIO_CD_FRAMES_PER_SEC} @code{CDIO_CD_FRAMSIZE_RAW}-byte blocks in a one second of time. When programming, unless one is working with a CD-DA (and even here, only in a time-like fashion), is generally more cumbersome to use an MSF rather than a LBA or LSN described below, since subtraction of two MSF's has the awkwardness akin to subtraction using Roman Numerals. Probably the simplest way to address a block is to use its @term{LSN} or ``logical sector number.'' This just numbers the blocks usually from 0 on. @emph{fix me: LSNs can be negative up to the pregap size?} The Lead-in and Lead-out gaps described above have LSNs just like any other space on a CD. The last unit of address is a @term{LBA}. It is the same as a LSN but the 150 blocks associated with the initial lead-in is are not counted. So to convert a LBA into an LSN you just add 150. Why the distinction between LBA and LSN? I don't know, perhaps this has something to do with ``multisession'' CDs. @node Pre-gaps @section track pre-gaps -- @acronym{CD-DA} discs and gaps @cindex CD-DA @cindex gaps @cindex lead in @cindex lead out @cindex pre-gap @cindex Q sub-channel Gaps are possibly one of the least understood topics in audio discs. In the case of @acronym{CD-DA} discs, standards require a silent 2 second gap before the first audio track and after the last audio track (in each session.) These are respectively referred to as @term{lead-in} and @term{lead-out} gaps. No other gaps are required. It is important not to confuse the required @term{lead-in} and @term{lead-out} gaps with the optional track @term{pre-gap}s. Track @term{pre-gap}s are the gaps that may occur between audio tracks. Typically, track @term{pre-gap}s are filled with silence so that the listener knows that one song has ended, and the next will soon begin. However, track @term{pre-gap}s do not have to contain silence. One exception is an audio disc of a live performance. Because the performer may seamlessly move from one piece of the performance to the next, it would be unnatural for the disc to contain silence between the two pieces. Instead, the track number updates with no interruption in the performance. This allows the listener to either hear the entire performance without unnatural interruptions, or to conveniently skip to certain pieces of the performance. Finally, some @acronym{CD-DA} discs--whose behavior will be described below--lack track @term{pre-gap}s altogether although they must still include the @term{lead-in} and @term{lead-out} gaps. In order to understand the track @term{pre-gap}s that occur between audio tracks, it is necessary to understand how CD players display the track number and time. Embedded in each block of audio data is non-audio information known as the @term{Q sub-channel}. The @term{Q sub-channel} data tells the CD player what track number and time it should display while it is playing the block of audio data in which the @term{Q sub-channel} data is embedded. Near the end of some tracks, the @term{Q sub-channel} may instruct the CD player to update the track number to the next track, and display a count down to the next track, often starting at -2 seconds and proceeding to zero. This is known as an audio track @term{pre-gap}. It may either contain silence, or as previously discussed--in the case of live performances--it may contain audio. Almost as often as not, there is no @term{pre-gap} whatsoever. Regardless, an audio track @term{pre-gap} is purely determined by the contents of the @term{Q sub-channel}, which is embedded in each audio sector. This has some interesting implications for the track forward button. When the track forward button is pressed on a CD player, the CD player advances to the next track, skipping that track's @term{pre-gap}. This is because the CD player uses the starting address of the track from the disc's table of contents (TOC) to determine where to start playing a track when either the track forward or track backward buttons are pressed. So to hear a @term{pre-gap} for track 4, the listener must either listen to track 3 first, or use the track forward or backward buttons to go to track 4, then use the seek backward button to back up into track 4's @term{pre-gap}, which is really part of track 3, at least according to the TOC. Track 1 @term{pre-gap}s are especially interesting because some commercial discs have audio hidden before the beginning of the first track! The only way to hear this hidden audio with a standard player is to use the seek backward button as soon as track 1 begins playing! Audio track @term{pre-gap}s may be specified in a couple of different ways in the popular cue file format. The first way of specifying a @term{pre-gap} is to use the @command{PREGAP} command. This will place a @term{pre-gap} containing silence before a track. The second way of specifying a @term{pre-gap} is to give a track an @command{INDEX 00} as well as the more normal @command{INDEX 01}. @command{INDEX 01} will be used to specify the start of the track in the disc's TOC, while @command{INDEX 00} will be used to specify the start of the track's @term{pre-gap} as recorded in the @term{Q sub-channel}. @command{INDEX 00} is ordinarily used for specifying track @term{pre-gap}s that contain audio rather than silence. Thus, the cue file format may be used to specify track @term{pre-gap}s with silence or audio, depending on whether the @command{PREGAP} or @command{INDEX 00} commands are specified. If neither type of @term{pre-gap} is specified for a track, no @term{pre-gap} is created for that track, which merely means the absence of @term{pre-gap} information in the @term{Q sub-channel}, and the lack of a short count down to the next track. Various @acronym{CD-DA} ripping programs take various approaches to track @term{pre-gap}s. Some ripping programs ignore track @term{pre-gap}s altogether, relying solely on the disc's TOC to determine where tracks begin and end. If a disc is ripped with such a program, then re-burned later, the resulting disc will lack track @term{pre-gap}s, and thereby lack the playback behavior of counting down to the next track. Other ripping programs detect track @term{pre-gap}s and record them in the popular cue file format among others. Such ripping programs sometimes allow the user to determine whether track @term{pre-gap}s will be appended to the prior track or pre-pended to the track to which they "belong". Note that if a ripping program is ignorant of track @term{pre-gap}s, the track @term{pre-gap}s will be appended to the prior track, because that is where the disc's TOC puts them. Thus, there are many different ways an application may chose to deal with track @term{pre-gap}s. Consequently, @kbd{libcdio} does not dictate the policy a ripping program should use in dealing with track @term{pre-gap}s. Hence, @kbd{libcdio} provides the @code{cdio_get_track_pregap_[lba|lsn]()} interfaces to allow the application to deal with track @term{pre-gap}s as it sees fit. Note that the @code{cdio_get_track_pregap_[lba|lsn]()} interfaces currently only provide information for CDRDAO TOC, CDRWIN BIN/CUE, and NRG images. Getting the track @term{pre-gap}s from a CD drive is a more complicated problem because not all CD drives support reading the @term{Q sub-channel} @emph{directly} at @emph{high} speed, and there is no interface to determine whether or not a drive supports this optional feature, aside from trying to read the @term{Q sub-channel}, and possibly incurring IO errors. However, all drives @emph{do} support reading the @term{Q sub-channel} @emph{indirectly} while playing an audio disc by asking the drive for the current position. Unfortunately, this occurs at normal playback speed, and requires a certain settling time after the disc starts playing. Thus, using this @emph{slow} interface requires a more sophisticated algorithm, such as binary search or some heuristic, like backing up progressively from the end of the prior track to look for the next track's @term{pre-gap}. Note that CD drives seek @emph{slow}ly, so it is better to simply use a drive that can read the @term{Q sub-channel} directly at @emph{high} speed, and avoid complicated software solutions. (Not to mention that if the user has an older system with an analog audio cable hooked up between their soundboard and their drive, and a ripping program uses the @emph{slow} interface, the user will hear bits of the audio on the disc!) Consequently, because there is no good universal solution to the problem of reading the @term{Q sub-channel} from a drive, @kbd{libcdio} currently leaves this problem up to the application, a problem which is readily approachable through either @kbd{libcdio}'s MMC interface or @kbd{libcdio}'s cdda interface. For an example of one such application, see @url{https://gna.org/projects/cued/}. The preceding section on track @term{pre-gaps} and @acronym{CD-DA} was contributed by Robert William Fuller (@email{hydrologiccycle@@gmail.com}). @node How to use @chapter How to use The @value{libcdio} package comes with a number of small example programs in the directory @file{example} which demonstrate different aspects of the library and show how to use the library. The source code to all of the examples here are contained on the package. Other sources for examples would be the larger utility programs @command{cd-drive}, @command{cd-info}, @command{cd-read}, @command{iso-info}, and @command{iso-read} which are all in the @file{src} directory of the @value{libcdio} package. See also @xref{Utility Programs}. @menu * Include problem:: A note about including * Example 1:: list out tracks and LSNs * Example 2:: list drivers available and default CD device * Example 3:: figure out what kind of CD (image) we've got * Example 4:: use libiso9660 to extract a file from an ISO-9660 image * Example 5:: list CD-Text and CD disc mode info * Example 6:: run a MMC INQUIRY command * Example 7:: using the CD Paranoia library for CD-DA reading * All sample programs:: list of all programs in the example directory @end menu @node Include problem @section A note about including @code{} libcdio installs @code{}. This file contains all of the C Preprocessor values from @code{config.h} (created by configure). This header can be used to consult exactly how libcdio was built. Initially I had selected ``interesting'' values, but this became too hard to maintain. One set of values that libdio needs internally is the whether the CPU that was used to compile libcdio is BigEndian or not; it can get this from libcdio's @code{config.h} which is not installed and preferred or @code{cdio/cdio_config.h}. Some of the libcdio programs like the demo programs include @code{config.h} for the generic reasons that the configuration-created @code{config.h} file is used: to figure out what headers are available. For example, do we have @code{}? The file @code{config.h} is generated by an autotools-generated @code{configure} script. It doesn't check to see if it has been included previously. Later, the demo programs include @code{} to get libcdio headers. But because libcdio needs some of the same information like the BigEndian value, this creates a duplicate include. The way I get around this in the demo programs is by defining @code{__CDIO_CONFIG_H__} after including @code{config.h} as follows: @smallexample #ifdef HAVE_CONFIG_H # include "config.h" # define __CDIO_CONFIG_H__ 1 #endif @end smallexample Applications using libcdio may find it handy to do something like this as well. Defining @code{__CDIO_CONFIG_H__} will make sure @code{config_cdio.h} which is internally used, doesn't try to redefine preprocessor symbols. Ok. But now what about the problem that there are common preprocessor symbols in @code{config_cdio.h} that an application may want to define in a different manner, like @code{PACKAGE_NAME}? For this, there is yet another header, @code{}. This file undefines any symbol that @code{config.h} defines. And now we bounce to the problem that there may be symbols that are normally defined (@code{HAVE_UNISTD_H}) and you want to keep that way, but others that you don't. So here is what I suggest: @smallexample // for cdio: #include #include # remove *all* symbols libcdio defines // Add back in the ones you want your program #include @end smallexample The solution isn't the most simple or natural, but programming sometimes can be difficult. If someone has a better solution, let me know. Between header files @code{cdio_config.h} and @code{cdio_unconfig.h} and all the fact that almost all headers@footnote{@code{} is one of the few headers that doesn't set a preprocessor symbol: it does its thing every time it is @code{#included}} define a symbol to indicate they have been included, I think there is enough mechanism to cover most situations that may arise. @node Example 1 @section Example 1: list out tracks and LSNs Here we will give an annotated example which can be found in the distribution as @file{example/tracks.c}. @smallexample 1: #include 2: #include 3: #include 4: int 5: main(int argc, const char *argv[]) 6: @{ 7: CdIo_t *p_cdio = cdio_open ("/dev/cdrom", DRIVER_DEVICE); 8: track_t first_track_num = cdio_get_first_track_num(p_cdio); 9: track_t i_tracks = cdio_get_num_tracks(p_cdio); 10: int j, i=first_track_num; 11: 12: printf("CD-ROM Track List (%i - %i)\n", first_track_num, i_tracks); 13 14: printf(" #: LSN\n"); 15: 16: for (j = 0; j < i_tracks; i++, j++) @{ 17: lsn_t lsn = cdio_get_track_lsn(p_cdio, i); 18: if (CDIO_INVALID_LSN != lsn) 19: printf("%3d: %06d\n", (int) i, lsn); 20: @} 21: printf("%3X: %06d leadout\n", CDIO_CDROM_LEADOUT_TRACK, 22: cdio_get_track_lsn(p_cdio, CDIO_CDROM_LEADOUT_TRACK)); 23: cdio_destroy(p_cdio); 24: return 0; 25: @} @end smallexample Already from the beginning on line 2 we see something odd. The @code{#include } is needed because @value{libcdio} assumes type definitions exist for @code{uint32_t}, @code{uint16_t} and so on. Alternatively you change line 2 to: @smallexample #define HAVE_SYS_TYPES_H @end smallexample and @code{} will insert line 2. If you use GNU autoconf to configure your program, add @code{sys/types.h} to @code{AC_HAVE_HEADERS} and @emph{it} will arrange for @code{HAVE_SYS_TYPES_H} to get defined. If you don't have @code{} but have some other include that defines these types, put that instead of line 2. Or you could roll your own typedefs. (Note: In the future, this will probably get ``fixed'' by requiring glib.h.) Okay after getting over the hurdle of line 2, the next line pretty straightforward: you need to include this to get cdio definitions. One of the types that is defined via line 3 is @code{CdIo_t} and a pointer that is used pretty much in all operations. Line 6 initializes the variable @code{cdio} which we will be using in all of the subsequent libcdio calls. It does this via a call to @code{cdio_open()}. The second parameter of @code{cdio_open} is DRIVER_UNKNOWN. For any given installation a number of Compact Disc device drivers may be available. In particular it's not uncommon to have several drivers that can read CD disk-image formats as well as a driver that handles some CD-ROM piece of hardware. Using DRIVER_UNKNOWN as that second parameter we let the library select a driver amongst those that are available; generally the first hardware driver that is available is the one selected. If there is no CD in any of the CD-ROM drives or one does not have access to the CD-ROM, it is possible that @value{libcdio} will find a CD image in the directory you run this program and will pick a suitable CD-image driver. If this is not what you want, but always want some sort of CD-ROM driver (or failure if none), then use DRIVER_DEVICE instead of DRIVER_UNKNOWN. Note that in contrast to what is typically done using ioctls to read a CD, you don't issue any sort of CD-ROM read TOC command---that is all done by the driver. Of course, the information that you get from reading the TOC is often desired: many tracks are on the CD, or what number the first one is called. This is done through calls on lines 8 and 9. For each track, we call a cdio routine to get the logical sector number, @code{cdio_get_track_lsn()} on line 17 and print the track number and LSN value. Finally we print out the ``lead-out track'' information and we finally call @code{cdio_destroy()} in line 23 to indicate we're done with the CD. @node Example 2 @section Example 2: list drivers available and default CD device One thing that's a bit hockey in Example 1 is hard-coding the name of the device used: @code{/dev/cdrom}. Although often this is the name of a CD-ROM device on GNU/Linux and possibly some other Unix derivatives, there are many OSs for which use a different device name. In the next example, we'll let the driver give us the name of the CD-ROM device that is right for it. @smallexample 1: #include 2: #include 3: #include 4: int 5: main(int argc, const char *argv[]) 6: @{ 7: CdIo_t *p_cdio = cdio_open (NULL, DRIVER_DEVICE); 8: const driver_id_t *driver_id_p; 9: 10: if (NULL != p_cdio) @{ 11: printf("The driver selected is %s\n", cdio_get_driver_name(p_cdio)); 12: printf("The default device for this driver is %s\n\n", 13: cdio_get_default_device(p_cdio)); 14: cdio_destroy(p_cdio); 15: @} else @{ 16: printf("Problem in trying to find a driver.\n\n"); 17: @} 18: 19: for (driver_id_p=cdio_drivers; *driver_id_p!=DRIVER_UNKNOWN; driver_id_p++) 20: if (cdio_have_driver(*driver_id_p)) 21: printf("We have: %s\n", cdio_driver_describe(*driver_id_p)); 22: else 23: printf("We don't have: %s\n", cdio_driver_describe(*driver_id_p)); 24: return 0; 25: @} @end smallexample @node Example 3 @section Example 3: figure out what kind of CD (image) we've got In this example is a somewhat simplified program to show the use of @command{cdio_guess_cd_type()} to figure out the kind of CD image we've got. This can be found in the distribution as @file{example/sample3.c}. @smallexample #ifdef HAVE_CONFIG_H # include "config.h" #endif #include #include #include #include #include static void print_analysis(cdio_iso_analysis_t cdio_iso_analysis, cdio_fs_anal_t fs, int first_data, unsigned int num_audio, track_t i_tracks, track_t first_track_num, CdIo_t *cdio) @{ switch(CDIO_FSTYPE(fs)) @{ case CDIO_FS_AUDIO: break; case CDIO_FS_ISO_9660: printf("CD-ROM with ISO 9660 filesystem"); if (fs & CDIO_FS_ANAL_JOLIET) @{ printf(" and joliet extension level %d", cdio_iso_analysis.joliet_level); @} if (fs & CDIO_FS_ANAL_ROCKRIDGE) printf(" and rockridge extensions"); printf("\n"); break; case CDIO_FS_ISO_9660_INTERACTIVE: printf("CD-ROM with CD-RTOS and ISO 9660 filesystem\n"); break; case CDIO_FS_HIGH_SIERRA: printf("CD-ROM with High Sierra filesystem\n"); break; case CDIO_FS_INTERACTIVE: printf("CD-Interactive%s\n", num_audio > 0 ? "/Ready" : ""); break; case CDIO_FS_HFS: printf("CD-ROM with Macintosh HFS\n"); break; case CDIO_FS_ISO_HFS: printf("CD-ROM with both Macintosh HFS and ISO 9660 filesystem\n"); break; case CDIO_FS_UFS: printf("CD-ROM with Unix UFS\n"); break; case CDIO_FS_EXT2: printf("CD-ROM with Linux second extended filesystem\n"); break; case CDIO_FS_3DO: printf("CD-ROM with Panasonic 3DO filesystem\n"); break; case CDIO_FS_UNKNOWN: printf("CD-ROM with unknown filesystem\n"); break; @} switch(CDIO_FSTYPE(fs)) @{ case CDIO_FS_ISO_9660: case CDIO_FS_ISO_9660_INTERACTIVE: case CDIO_FS_ISO_HFS: printf("ISO 9660: %i blocks, label `%.32s'\n", cdio_iso_analysis.isofs_size, cdio_iso_analysis.iso_label); break; @} if (first_data == 1 && num_audio > 0) printf("mixed mode CD "); if (fs & CDIO_FS_ANAL_XA) printf("XA sectors "); if (fs & CDIO_FS_ANAL_MULTISESSION) printf("Multisession"); if (fs & CDIO_FS_ANAL_HIDDEN_TRACK) printf("Hidden Track "); if (fs & CDIO_FS_ANAL_PHOTO_CD) printf("%sPhoto CD ", num_audio > 0 ? " Portfolio " : ""); if (fs & CDIO_FS_ANAL_CDTV) printf("Commodore CDTV "); if (first_data > 1) printf("CD-Plus/Extra "); if (fs & CDIO_FS_ANAL_BOOTABLE) printf("bootable CD "); if (fs & CDIO_FS_ANAL_VIDEOCD && num_audio == 0) @{ printf("Video CD "); @} if (fs & CDIO_FS_ANAL_SVCD) printf("Super Video CD (SVCD) or Chaoji Video CD (CVD)"); if (fs & CDIO_FS_ANAL_CVD) printf("Chaoji Video CD (CVD)"); printf("\n"); @} int main(int argc, const char *argv[]) @{ CdIo_t *p_cdio = cdio_open (NULL, DRIVER_UNKNOWN); cdio_fs_anal_t fs=0; track_t i_tracks; track_t first_track_num; lsn_t start_track; /* first sector of track */ lsn_t data_start =0; /* start of data area */ int first_data = -1; /* # of first data track */ int first_audio = -1; /* # of first audio track */ unsigned int num_data = 0; /* # of data tracks */ unsigned int num_audio = 0; /* # of audio tracks */ unsigned int i; if (NULL == p_cdio) @{ printf("Problem in trying to find a driver.\n\n"); return 1; @} first_track_num = cdio_get_first_track_num(p_cdio); i_tracks = cdio_get_num_tracks(p_cdio); /* Count the number of data and audio tracks. */ for (i = first_track_num; i <= i_tracks; i++) @{ if (TRACK_FORMAT_AUDIO == cdio_get_track_format(p_cdio, i)) @{ num_audio++; if (-1 == first_audio) first_audio = i; @} else @{ num_data++; if (-1 == first_data) first_data = i; @} @} /* try to find out what sort of CD we have */ if (0 == num_data) @{ printf("Audio CD\n"); @} else @{ /* we have data track(s) */ int j; cdio_iso_analysis_t cdio_iso_analysis; memset(&cdio_iso_analysis, 0, sizeof(cdio_iso_analysis)); for (j = 2, i = first_data; i <= i_tracks; i++) @{ lsn_t lsn; track_format_t track_format = cdio_get_track_format(p_cdio, i); lsn = cdio_get_track_lsn(p_cdio, i); switch ( track_format ) @{ case TRACK_FORMAT_AUDIO: case TRACK_FORMAT_ERROR: break; case TRACK_FORMAT_CDI: case TRACK_FORMAT_XA: case TRACK_FORMAT_DATA: case TRACK_FORMAT_PSX: ; @} start_track = (i == 1) ? 0 : lsn; /* save the start of the data area */ if (i == first_data) data_start = start_track; /* skip tracks which belong to the current walked session */ if (start_track < data_start + cdio_iso_analysis.isofs_size) continue; fs = cdio_guess_cd_type(p_cdio, start_track, i, &cdio_iso_analysis); print_analysis(cdio_iso_analysis, fs, first_data, num_audio, i_tracks, first_track_num, p_cdio); if ( !(CDIO_FSTYPE(fs) == CDIO_FS_ISO_9660 || CDIO_FSTYPE(fs) == CDIO_FS_ISO_HFS || CDIO_FSTYPE(fs) == CDIO_FS_ISO_9660_INTERACTIVE) ) /* no method for non-ISO9660 multisessions */ break; @} @} cdio_destroy(p_cdio); return 0; @} @end smallexample @node Example 4 @section Example 4: use libiso9660 to extract a file from an ISO-9660 image Next a program to show using @command{libiso9660} to extract a file from an ISO-9660 image. This can be found in the distribution as @file{example/isofile.c}. A more complete and expanded version of this is @command{iso-read}, part of this distribution. @smallexample /* This is the ISO 9660 image. */ #define ISO9660_IMAGE_PATH "../" #define ISO9660_IMAGE ISO9660_IMAGE_PATH "test/copying.iso" #define LOCAL_FILENAME "copying" #ifdef HAVE_CONFIG_H # include "config.h" #endif #include #include #include #include #ifdef HAVE_ERRNO_H #include #endif #ifdef HAVE_STRING_H #include #endif #ifdef HAVE_UNISTD_H #include #endif #ifdef HAVE_SYS_TYPES_H #include #endif #define my_exit(rc) \ fclose (p_outfd); \ free(p_statbuf); \ iso9660_close(p_iso); \ return rc; \ int main(int argc, const char *argv[]) @{ iso9660_stat_t *p_statbuf; FILE *p_outfd; int i; iso9660_t *p_iso = iso9660_open (ISO9660_IMAGE); if (NULL == p_iso) @{ fprintf(stderr, "Sorry, couldn't open ISO 9660 image %s\n", ISO9660_IMAGE); return 1; @} p_statbuf = iso9660_ifs_stat_translate (p_iso, LOCAL_FILENAME); if (NULL == p_statbuf) @{ fprintf(stderr, "Could not get ISO-9660 file information for file %s\n", LOCAL_FILENAME); iso9660_close(p_iso); return 2; @} if (!(p_outfd = fopen (LOCAL_FILENAME, "wb"))) @{ perror ("fopen()"); free(p_statbuf); iso9660_close(p_iso); return 3; @} /* Copy the blocks from the ISO-9660 filesystem to the local filesystem. */ for (i = 0; i < p_statbuf->size; i += ISO_BLOCKSIZE) @{ char buf[ISO_BLOCKSIZE]; memset (buf, 0, ISO_BLOCKSIZE); if ( ISO_BLOCKSIZE != iso9660_iso_seek_read (p_iso, buf, p_statbuf->lsn + (i / ISO_BLOCKSIZE), 1) ) @{ fprintf(stderr, "Error reading ISO 9660 file at lsn %lu\n", (long unsigned int) p_statbuf->lsn + (i / ISO_BLOCKSIZE)); my_exit(4); @} fwrite (buf, ISO_BLOCKSIZE, 1, p_outfd); if (ferror (p_outfd)) @{ perror ("fwrite()"); my_exit(5); @} @} fflush (p_outfd); /* Make sure the file size has the exact same byte size. Without the truncate below, the file will a multiple of ISO_BLOCKSIZE. */ if (ftruncate (fileno (p_outfd), p_statbuf->size)) perror ("ftruncate()"); my_exit(0); @} @end smallexample @node Example 5 @section Example 5: list CD-Text and disc mode info Next a program to show using @command{libcdio} to list CD-TEXT data. This can be found in the distribution as @file{example/cdtext.c}. @smallexample /* Simple program to list CD-Text info of a Compact Disc using libcdio. */ #ifdef HAVE_CONFIG_H # include "config.h" #endif #include #include #include #include static void print_cdtext_track_info(CdIo_t *p_cdio, track_t i_track, const char *message) @{ const cdtext_t *cdtext = cdio_get_cdtext(p_cdio, 0); if (NULL != cdtext) @{ cdtext_field_t i; printf("%s\n", message); for (i=0; i < MAX_CDTEXT_FIELDS; i++) @{ if (cdtext->field[i]) @{ printf("\t%s: %s\n", cdtext_field2str(i), cdtext->field[i]); @} @} @} @} static void print_disc_info(CdIo_t *p_cdio, track_t i_tracks, track_t i_first_track) @{ track_t i_last_track = i_first_track+i_tracks; discmode_t cd_discmode = cdio_get_discmode(p_cdio); printf("%s\n", discmode2str[cd_discmode]); print_cdtext_track_info(p_cdio, 0, "\nCD-Text for Disc:"); for ( ; i_first_track < i_last_track; i_first_track++ ) @{ char psz_msg[50]; sprintf(msg, "CD-Text for Track %d:", i_first_track); print_cdtext_track_info(p_cdio, i_first_track, psz_msg); @} @} int main(int argc, const char *argv[]) @{ track_t i_first_track; track_t i_tracks; CdIo_t *p_cdio; cdio = cdio_open (NULL, DRIVER_UNKNOWN); i_first_track = cdio_get_first_track_num(p_cdio); i_tracks = cdio_get_num_tracks(p_cdio); if (NULL == p_cdio) @{ printf("Couldn't find CD\n"); return 1; @} else @{ print_disc_info(p_cdio, i_tracks, i_first_track); @} cdio_destroy(p_cdio); return 0; @} @end smallexample @node Example 6 @section Example 6: Using MMC to run an @code{INQURY} command Now a program to show issuing a simple MMC command (@code{INQUIRY}). This MMC command retrieves the vendor, model and firmware revision number of a CD drive. For this command to work, usually a CD to be loaded into the drive; odd since the CD itself is not used. This can be found in the distribution as @file{example/mmc1.c}. @smallexample #ifdef HAVE_CONFIG_H # include "config.h" # define __CDIO_CONFIG_H__ 1 /* assumes config.h is libcdio's config.h / #endif #include #include #include #include #include /* Set how long to wait for MMC commands to complete */ #define DEFAULT_TIMEOUT_MS 10000 int main(int argc, const char *argv[]) @{ CdIo_t *p_cdio; p_cdio = cdio_open (NULL, DRIVER_UNKNOWN); if (NULL == p_cdio) @{ printf("Couldn't find CD\n"); return 1; @} else @{ int i_status; /* Result of MMC command */ char buf[36] = @{ 0, @}; /* Place to hold returned data */ scsi_mmc_cdb_t cdb = @{@{0, @}@}; /* Command Descriptor Buffer */ CDIO_MMC_SET_COMMAND(cdb.field, CDIO_MMC_GPCMD_INQUIRY); cdb.field[4] = sizeof(buf); i_status = scsi_mmc_run_cmd(p_cdio, DEFAULT_TIMEOUT_MS, &cdb, SCSI_MMC_DATA_READ, sizeof(buf), &buf); if (i_status == 0) @{ char psz_vendor[CDIO_MMC_HW_VENDOR_LEN+1]; char psz_model[CDIO_MMC_HW_MODEL_LEN+1]; char psz_rev[CDIO_MMC_HW_REVISION_LEN+1]; memcpy(psz_vendor, buf + 8, sizeof(psz_vendor)-1); psz_vendor[sizeof(psz_vendor)-1] = '\0'; memcpy(psz_model, buf + 8 + CDIO_MMC_HW_VENDOR_LEN, sizeof(psz_model)-1); psz_model[sizeof(psz_model)-1] = '\0'; memcpy(psz_rev, buf + 8 + CDIO_MMC_HW_VENDOR_LEN +CDIO_MMC_HW_MODEL_LEN, sizeof(psz_rev)-1); psz_rev[sizeof(psz_rev)-1] = '\0'; printf("Vendor: %s\nModel: %s\nRevision: %s\n", psz_vendor, psz_model, psz_rev); @} else @{ printf("Couldn't get INQUIRY data (vendor, model, and revision\n"); @} @} cdio_destroy(p_cdio); return 0; @} @end smallexample Note the include of @code{#define} of @code{__CDIO_CONFIG_H__} towards the beginning. This is useful if the prior @code{#include} of @code{config.h} refers to libcdio's configuration header. It indicates that libcdio's configuration settings have been used. Without it, you may get messages about C Preprocessor symbols getting redefined in the @code{#include} of @code{}. @node Example 7 @section Example 7: Using the CD Paranoia library for CD-DA reading The below program reads CD-DA data. For a more complete program to add a WAV header so that the CD can be played from a copy on a hard disk, see the corresponding distribution program. This can be found in the distribution as @file{example/paranoia.c}. @smallexample #ifdef HAVE_CONFIG_H # include "config.h" # define __CDIO_CONFIG_H__ 1 /* assumes config.h is libcdio's config.h / #endif #include #include #include #ifdef HAVE_STDLIB_H #include #endif int main(int argc, const char *argv[]) @{ cdrom_drive_t *d = NULL; /* Place to store handle given by cd-paranoia. */ char **ppsz_cd_drives; /* List of all drives with a loaded CDDA in it. */ /* See if we can find a device with a loaded CD-DA in it. */ ppsz_cd_drives = cdio_get_devices_with_cap(NULL, CDIO_FS_AUDIO, false); if (ppsz_cd_drives) @{ /* Found such a CD-ROM with a CD-DA loaded. Use the first drive in the list. */ d=cdio_cddap_identify(*ppsz_cd_drives, 1, NULL); @} else @{ printf("Unable find or access a CD-ROM drive with an audio CD in it.\n"); exit(1); @} /* Don't need a list of CD's with CD-DA's any more. */ cdio_free_device_list(ppsz_cd_drives); /* We'll set for verbose paranoia messages. */ cdio_cddap_verbose_set(d, CDDA_MESSAGE_PRINTIT, CDDA_MESSAGE_PRINTIT); if ( 0 != cdio_cddap_open(d) ) @{ printf("Unable to open disc.\n"); exit(1); @} /* Okay now set up to read up to the first 300 frames of the first audio track of the Audio CD. */ @{ cdrom_paranoia_t *p = cdio_paranoia_init(d); lsn_t i_first_lsn = cdio_cddap_disc_firstsector(d); if ( -1 == i_first_lsn ) @{ printf("Trouble getting starting LSN\n"); @} else @{ lsn_t i_cursor; track_t i_track = cdio_cddap_sector_gettrack(d, i_first_lsn); lsn_t i_last_lsn = cdio_cddap_track_lastsector(d, i_track); /* For demo purposes we'll read only 300 frames (about 4 seconds). We don't want this to take too long. On the other hand, I suppose it should be something close to a real test. */ if ( i_last_lsn - i_first_lsn > 300) i_last_lsn = i_first_lsn + 299; printf("Reading track %d from LSN %ld to LSN %ld\n", i_track, (long int) i_first_lsn, (long int) i_last_lsn); /* Set reading mode for full paranoia, but allow skipping sectors. */ paranoia_modeset(p, PARANOIA_MODE_FULL^PARANOIA_MODE_NEVERSKIP); paranoia_seek(p, i_first_lsn, SEEK_SET); for ( i_cursor = i_first_lsn; i_cursor <= i_last_lsn; i_cursor ++) @{ /* read a sector */ int16_t *p_readbuf=cdio_paranoia_read(p, NULL); char *psz_err=cdio_cddap_errors(d); char *psz_mes=cdio_cddap_messages(d); if (psz_mes || psz_err) printf("%s%s\n", psz_mes ? psz_mes: "", psz_err ? psz_err: ""); if (psz_err) free(psz_err); if (psz_mes) free(psz_mes); if( !p_readbuf ) @{ printf("paranoia read error. Stopping.\n"); break; @} @} @} cdio_paranoia_free(p); @} cdio_cdda_close(d); exit(0); @} @end smallexample Those who are die-hard cdparanoia programmers will notice that the @value{libcdio} paranoia names are similar but a little bit different. In particular instead of @code{paranoia_read} we have above @code{cdio_paranoia_read} and instead of @code{cdda_open} we have @code{cdio_cddap_open}. This was done intentionally so that it is possible for the original paranoia program can co-exist both in source code and linked libraries and not conflict with @value{libcdio}'s paranoia source and libraries. In general in place of any paranoia routine that begins @code{paranoia_}, use @code{cdio_paranoia_} and in place of any paranoia routine that begins @code{cdda_}, use @code{cdio_cddap_}. But for a limited time @value{libcdio} will accept the old paranoia names which may be useful for legacy paranoia code. The way this magic works is by defining the old paranoia name to be the @value{libcdio} name. In the unusual case where you do want to use both the original paranoia and @value{libcdio} routines in a single source, the C preprocessor symbol @code{DO_NOT_WANT_PARANOIA_COMPATIBILITY} can be @code{define}'d and this disables the @code{#define} substitution done automatically. The may still be a problem with conflicting structure definitions like @code{cdrom_drive_t}. @node All sample programs @section A list of all sample programs in the @code{example} directory The @code{example} directory contains some simple examples of the use of the @value{libcdio} library. A larger more-complicated example are the @command{cd-drive}, @command{cd-info}, @command{cd-read}, @command{iso-info} and @command{iso-info} programs in the @command{src} directory. Descriptions of the sample are as follows... @table @code @item @code{audio.c} A program to show audio controls. @item @code{cdchange.c} A program to test if a CD has been changed since the last change test. @item @code{cd-eject.c} A a stripped-down "eject" command to open or close a CD-ROM tray. @item @code{cdtext.c} A program to show CD-Text and CD disc mode info. @item @code{drives.c} A program to show drivers installed and what the default CD-ROM drive is and what CD drives are available. @item @code{eject.c} A program eject a CD from a CD-ROM drive and then close the door again. @item @code{isofile.c} A program to show using libiso9660 to extract a file from an ISO-9660 image. @item @code{isofile2.c} A program to show using libiso9660 to extract a file from a CDRWIN cue/bin CD image. @item @code{C++/isofile2.cpp} The same program as @code{isofile2.c} written in C++. @item @code{isofuzzy.c} A program showing fuzzy ISO-9660 detection/reading. @item @code{isolist.c} A program to show using @code{libiso9660} to list files in a directory of an ISO-9660 image. @item @code{C++/isolist.cpp} The same program as @code{isolist.c} written in C++. @item @code{C++/isolist.cpp} The same program as @code{isolist.c} written in C++. @item @code{isofuzzy.c} A program showing fuzzy ISO-9660 detection/reading. @item @code{mmc1.c} A program to show issuing a simple MMC command (@code{INQUIRY}). @item @code{C++/mmc1.cpp} The same program as @code{mmc1.c} written in C++. @item @code{mmc2.c} A more involved MMC command to list CD and drive features from a SCSI-MMC @code{GET_CONFIGURATION} command. @item @code{mmc2a.c} Prints MMC @command{MODE_SENSE} page 2A parameters. Page 2a are the CD/DVD Capabilities and Mechanical Status. @item @code{C++/mmc2.cpp} The same program as @code{mmc2.c} written in C++. @item @code{paranoia.c} A program to show using libcdio's version of the CD-DA paranoia. @item @code{paranoia2.c} A program to show using libcdio's version of the CD-DA paranoia library. But in this version, we'll open a cdio object before calling paranoia's open. I imagine in many cases such as media players this may be what will be done since, one may want to get CDDB/CD-Text info beforehand. @item @code{tracks.c} A simple program to list track numbers and logical sector numbers of a Compact Disc using @value{libcdio}. @item @code{sample2.c} A simple program to show drivers installed and what the default CD-ROM drive is. @item @code{sample3.c} A simple program to show the use of @code{cdio_guess_cd_type()}. Figures out the kind of CD image we've got. @item @code{sample4.c} A slightly improved sample3 program: we handle cdio logging and take an optional CD-location. @item @code{udf1.c} A program to show using libudf to list files in a directory of an UDF image. @item @code{udf2.c} A program to show using libudf to extract a file from an UDF image. @end table @node Utility Programs @chapter Diagnostic programs: @command{cd-drive}, @command{cd-info}, @command{cd-read}, @command{iso-info}, @command{iso-read} @menu * cd-drive:: list out CD-ROM drive information * cd-info:: list out CD or CD-image information * cd-read:: read blocks of a CD or CD image * iso-info:: list out ISO-9600 image information * iso-read:: extract a file from an ISO 9660 image @end menu @node cd-drive @section @samp{cd-drive} @samp{cd-drive} lists out drive information, what features drive supports, and information about what hardware drivers are available. @node cd-info @section @samp{cd-info} @samp{cd-info} will print out the structure of a CD medium which could either be a Compact Disc in a CD ROM or an CD image. It can try to analyze the medium to give characteristics of the medium, such as how many tracks are in the CD and the format of each track, whether a CD contains a Video CD, CD-DA, PhotoCD, whether a track has an ISO-9660 filesystem. @node cd-read @section @samp{cd-read} @samp{cd-info} can be used to read blocks a CD medium which could either be a Compact Disc in a CD ROM or an CD image. You specify the beginning and ending LSN and what mode format to use in the reading. @node iso-info @section @samp{iso-info} @samp{iso-info} can be used to print out the structure of an ISO 9660 image. @node iso-read @section @samp{iso-read} @samp{iso-read} can be used to extract a file in an ISO-9660 image. @node CD-ROM Access and Drivers @chapter CD-ROM Access and Drivers @menu * SCSI mess:: SCSI, SCSI commands, and MMC commands * Access Modes:: Access Modes * Accessing Driver Parameters:: Accessing Driver Parameters * GNU/Linux:: GNU/Linux ioctl * Microsoft:: Microsoft Windows ioctl and ASPI * Solaris:: Solaris ATAPI and SCSI * FreeBSD:: FreeBSD ioctl and CAM * OS X:: OSX (non-exclussive access) @end menu @node SCSI mess @section SCSI, SCSI commands, and MMC commands Historically, SCSI referred to a class of hardware devices and device controllers, bus technology and the data cables and protocols which attached to such devices. This is now called ``Parallel SCSI''. A specification standard grew out of the @emph{commands} that controlled such SCSI devices, but now covers a wider variety of bus technologies including Parallel SCSI, ATA/ATAPI, Serial ATA, Universal Serial Bus (USB versions 1.1 and 2.0), and High Performance Serial Bus (IEEE 1394, 1394A, and 1394B). Another similar class of hardware devices and controllers is called ATA and a command interface to that is called ATAPI (ATA Packetized Interface). ATAPI provides a mechanism for transferring and executing SCSI commands. MMC (Multimedia commands) is a specification which adds special SCSI commands for CD, DVD, Blu-Ray devices. If your optical drive understands MMC commands as most do nowadays, this probably gives the most flexibility in control. SCSI and ATAPI CD-ROM devices generally support a fairly large set of MMC commands. Unfortunately, on most Operating Systems one may need to do some additional setup, such as install drivers or modules, to allow access in this manner. The name ``SCSI MMC'' is often found in the literature in specifications and on the Internet. The ``SCSI'' part is probably a little bit misleading because a drive can understand ``SCSI MMC'' commands but not use a SCSI bus protocol---ATAPI CD-ROMs are one such broad class of examples. In fact there are drivers to ``encapsulate'' non-SCSI drives to make them act like more like SCSI drives, such as by adding SCSI address naming. For clarity and precision we will use the term ``MMC'' rather than ``SCSI MMC''. One of the problems with MMC is that there are so many different ``standards''. In particular: @itemize @item MMC --- @url{ftp://ftp.t10.org/t10/drafts/mmc/}, @item MMC 2 --- @url{ftp://ftp.t10.org/t10/drafts/mmc2/} @item MMC 3 --- @url{ftp://ftp.t10.org/t10/drafts/mmc3/} @item MMC 4 --- @url{ftp://ftp.t10.org/t10/drafts/mmc4/} @item MMC 5 --- @url{ftp://ftp.t10.org/t10/drafts/mmc5/} @end itemize along with the several ``drafts'' of these. Another problem with the MMC commands related to the variations in standards is the variation in the commands themselves and there are perhaps two or three ways to do many of the basic commands like read a CD frame. There seems to be a fascination with the number of bytes a command takes in the MMC-specification world. (Size matters?) So often the name of an operation will have a suffix with the number of bytes of the command (actually in MMC jargon this is called a ``CDB'' @cindex CDB (Command Descriptor Block) or command descriptor block). So for example there is a 6-byte ``MODE SELECT'' often called ``MODE SELECT 6'' and a 10-byte ``MODE SELECT'' often called ``MODE SELECT 10''. Presumably the 6-byte command came first and it was discovered that there was some deficiency causing the longer command. In @value{libcdio} where there are two formats we add the suffix in the name, e.g. @code{CDIO_MMC_GPCMD_MODE_SELECT_6} or @code{CDIO_MMC_GPCMD_MODE_SELECT_10}. If the fascination and emphasis in the MMC specifications of CDB size is a bit odd, equally so is the fact that this too often has bled through at the OS programming API. However in @value{libcdio}, you just give the opcode in @code{scsi_mmc_run_cmd()} and we'll do the work to figure out how many bytes of the CDB are used. Down the line it is hoped that @value{libcdio} will have a way to remove a distinction between the various alternative and alternative-size MMC commands. In @code{cdio/scsi-mmc.h} you will find a little bit of this for example via the routine @code{scsi_mmc_get_drive_cap()}. However much more work is needed. Finally, in @code{libcdio} there is a driver access mode (not a driver) called ``MMC''. It tells the specific drivers to use MMC commands instead of other OS-specific mechanisms. @node Access Modes @section Access Modes There are serveral ways that you can open a CD-ROM drive for subsequent use. Each way is called an @emph{access mode}. Historically libcdio only supported a reading kind of access. Adding the abilty to writing to a drive for ``burning'' is being added by Thomas Schmitt, and this is accomplished by opening the drive in a read-write mode. Currently writing modes are only supported via the MMC command interface. Under this, one can get exclusive read-write access or non-exclusive read-write access. The names of these two modes are @code{MMC_RDWR_EXCL} and @code{MMC_RDWR} respectively. On various OS's often there are two kinds of read modes that are supported, one which uses MMC commands and one which uses some sort of OS-specific native command interface. For example on Unix, there is often a access mode associated with issuing an device-specific @code{ioctl}'s that the OS supports. To specify a particular kind of access mode, use @code{cdio_open_am} which is like @code{cdio_open} but it requires one to specify an access mode. @node Accessing Driver Parameters @section Accessing Driver Parameters --- @code{cdio_get_arg} Once a driver is opened, you can use call @code{cdio_get_arg} to get information about the driver. Each driver can have specific features that can be queried, but there are features that are common to all drivers. These are listed below: @table @code @item @code{access-mode} This returns a string which is the name of the access mode in use. @item @code{mmc-supported?} This returns a string ``true'' or ``false'' depending whether the driver with this access mode support MMC commands. @item @code{scsi-tuple} On drivers that support MMC commands, this returns the SCSI name or a faked-up SCSI name that ripping front ends typically use. @end table @node GNU/Linux @section GNU/Linux The GNU/Linux uses a hybrid of methods. Somethings are done via ioctl and some things via MMC. GNU/Linux has a rather nice and complete ioctl mechanism. On the other hand, the MMC mechanism is more universal. There are other ``access modes'' listed which are not really access modes and should probably be redone/rethought. They are just different ways to run the read command. But for completeness These are ``READ_CD'' and ``READ_10''. Writing/burning to a drive is supported via access modes @code{MMC_RDWR_EXCL} or @code{MMC_RDWR}. @node Microsoft @section Microsoft Windows ioctl and ASPI There are two CD drive access methods on Microsoft Windows platforms: ioctl and ASPI. The ASPI interface specification was developed by Adaptec for sending commands to a SCSI host adapter (such as those controlling CD and DVD drives) and used on Window 9x/NT and later. Emulation for ATAPI drives was added so that the same sets of commands worked those even though the drives might not be SCSI nor might there even be a SCSI controller attached. The DLL is not part of Microsoft Windows and has to be downloaded and installed separately. However in Windows NT/2K/XP, Microsoft provides their Win32 ioctl interface, and has taken steps to make using ASPI more inaccessible (e.g. requiring administrative access to use ASPI). @node Solaris @section Solaris ATAPI and SCSI There is currently only one CD drive access methods in Solaris: SCSI (called ``USCSI'' or ``user SCSI'' in Solaris). There used to be an ATAPI method and it could be resurrected if needed. USCSI was preferred since on newer releases of Solaris and Solaris environments one need to have root access for ATAPI. @node FreeBSD @section FreeBSD ioctl and CAM There are two classes of access methods on FreeBSD: ioctl and CAM (common access method). CAM is preferred when possible, especially on newer releases. However CAM is right now sort of a hybrid and includes some ioctl code. Writing/burning to a drive is supported via access modes @code{MMC_RDWR_EXCL} or @code{MMC_RDWR} which underneath use CAM access. @node OS X @section OS X (non-exclusive access) A problem with libcdio on OS/X is that if the OS thinks it understands the drive, it will get exclusive access to the drive and thus prevents a library like this from obtaining non-exclusive access. Currently @value{libcdio} access the CD-ROM non-exclusively. However in order to be able to issue MMC, the current belief is that exclusive access is needed. Probably in a future @value{libcdio}, there will be some way to specify which kind of access is desired (with the inherent consequences of each). More work on this driver is needed. Volunteers? @node Internal Program Organization @chapter Internal Program Organization @menu * File Organization:: * Library Organization:: * Programming Conventions:: @end menu @node File Organization @section File Organization Here is a list of @value{libcdio} directories. @itemize @item @code{include/cdio} This contains the headers that are public. One that will probably be used quite a bit is @code{}. @item @code{lib} Code for installed libraries. See below for further breakout @item @code{lib/driver} Code for various OS-specific CD-ROM drivers, image drivers, and common MMC routines. This code comprises @code{libcdio.a} (or the shared version of it). @item @code{lib/iso9660} Code for to extract or query ISO-9660 images. This code comprises @code{libiso9660.a} (or the shared version of it). @item @code{lib/paranoia} This is from cdparanoia. It is the OS- and hardware- dependent code to detect and correct jitter for CD-DA CDs. @item @code{lib/cdda_interface} This is also from cdparanoia. It is the OS- and hardware- independent code to detect and correct jitter for CD-DA CDs. @item @code{doc} A home for fine documentation such as this masterpiece. @item @code{example} Here you will find various small example programs using @value{libcdio} which are largely for pedagogical purposes. You might be able to find one that is similar to what you want to do that could be extended. In fact some these are contain the kernel ideas behind of some of the larger programs in @file{src}. @item @code{src} Various stand-alone utility programs. See below. @item @code{src/paranoia} @value{libcdio}'s version of @code{cdparanoia}. Except for the fact that the back-end CD-reading code has been replaced by @value{libcdio}'s routines the code is pretty much identical. @item @code{test} Regression tests @item @code{test/data} Disk images and image meta-data used in tests @item @code{test/driver} Unit tests centered around the libcdio library (@code{libcdio}, source location @code{lib/driver} @end itemize @node Library Organization @section Library Organization @menu * libcdio:: * libcdio_cdda:: Access to CD-DA via the CD Paranoia library * libcdio_paranoia:: Access to the CD Paranoia library * libiso9660:: Access to ISO 9660 file systems and structures * libudf:: Access to UDF file systems and structures @end menu @node libcdio @subsection @samp{libcdio} @value{libcdio} exports one opaque type @code{CdIo_t}. Internally this a structure containing an enumeration for the driver, a structure containing function pointers and a generic ``environment'' pointer which is passed as a parameter on a function call. See @file{lib/driver/cdio_private.h}. The initialization routine for each driver sets up the function pointers and allocates memory for the environment. When a particular user-level cdio routine is called (e.g @code{cdio_get_first_track_num} for lib/driver/track.c), the environment pointer is passed to a device-specific routine which will then cast this pointer into something of the appropriate type. Because function pointers are used, there can be and is quite a bit of sharing of common routines. Some of the common routines are found in the file @file{lib/driver/_cdio_generic.c}. Another set of routines that one is likely to find shared amongst drivers are the MMC commands. These are located in @file{lib/driver/scsi_mmc.c}. There is not only an attempt to share functions but we've tried to create a generic CD structure @code{generic_img_private_t} of file @file{lib/driver/generic.h}. By putting information into a common structure, we increase the likelihood of being able to have a common routine to perform some sort of function. The generic CD structure would also be useful in a utility to convert one CD-image format to another. Basically the first image format is ``parsed'' into the common internal format and then from this structure it is not parsed. @node libcdio_cdda @subsection @samp{libcdio_cdda} This library is intended to give access CD-DA disks using Monty's cd-paranoia library underneath. To be completed.... @node libcdio_paranoia @subsection @samp{libcdio_paranoia} This library is intended to give access Monty's cd-paranoia library. It is the gap detection and jitter correction part without the part dealing with CD-DA reading. To be completed.... @node libiso9660 @subsection @samp{libiso9660} This library is intended to give access and manipulate a ISO-9600 file image. One part of it is concerned with the the entire ISO-9660 file system image, and the other part access routines for manipulating data structures and fields that go into such an image. To be completed.... @node libudf @subsection @samp{libudf} This library is intended to give access and manipulate a UDF file image. To be completed.... @node Programming Conventions @section Programming Conventions @menu * Coding Conventions:: * Namespace Conventions:: @end menu @node Coding Conventions @subsection Coding Conventions In @value{libcdio} there are a number of conventions used. If you understand some of these conventions it may facilitate understanding the code a little. @node Namespace Conventions @subsection Namespace Conventions For the most part, the visible external @value{libcdio} names follow conventions so as not to be confused with other applications or libraries. If you understand these conventions, there will be little or no chance that the names you use will conflict with @value{libcdio} and @code{libiso9660} and vice versa. All of the external @value{libcdio} C routines start out with @code{cdio_}, e.g. @code{cdio_open}; as a corollary, the @value{libcdio} CD-Paranoia routines start @code{cdio_cddap_}, e.g. @code{cdio_cddap_open}. @code{libiso9660} routines start @code{iso9660_}, e.g. @code{iso9660_open}. @value{libcdio} C-Preprocessor names generally start @code{CDIO_}, for example @code{CDIO_CD_FRAMESIZE_RAW}; @code{libiso9660} C-preprocessor names start @code{ISO9660_}, e.g. @code{ISO9660_FRAMESIZE}. @subsubsection suffixes (type and structure names) A few suffixes are used in type and structure names: @itemize @item @code{_e} An enumeration tag. Generally though the same name will appear with the @code{_t} suffix and probably that should be used instead. @item @code{_s} A structure tag. Generally though the same name will appear with the @code{_t} suffix and probably that should be used instead. @item @code{_t} A type suffix. @end itemize @subsubsection prefixes (variable names) A number of prefixes are used in variable names here's what they mean @itemize @item @code{i_} An integer type of some sort. A variable of this ilk one might find being iterated over in @code{for} loops or used as the index of an array for example. @item @code{b_} A Boolean type of some sort. A variable of this ilk one might find being in an @code{if} condition for example. @item @code{p_} A pointer of some sort. A variable of this ilk, say @code{p_foo} one is like likely to see @code{*p_foo} or @code{p_foo->...}. @item @code{pp_} A pointer to a pointer of some sort. A variable of this ilk, say @code{pp_foo} one is like likely to see @code{**p_foo} or @code{p_foo[x][y]} for example @item @code{psz_} A @code{char *} pointer of some sort. A variable of this ilk, say @code{psz_foo} may be used in a string operation. For example @code{printf(%s\n", psz_foo)} or @code{strdup(psz_foo)}. @item @code{ppsz_} A pointer to a @code{char *} pointer of some sort. A variable of this ilk, say @code{ppsz_foo} is used for example to return a list of CD-ROM device names @end itemize There are a some other naming conventions. Generally if a routine name starts @code{cdio_}, e.g. @code{cdio_open}, then it is an externally visible routine in @code{libcdio}. If a name starts @code{iso9660_}, e.g. @code{iso9660_is_dchar} then it is an externally visible routine in @code{libiso9660}. If a name starts @code{scsi_mmc_}, e.g. @code{scsi_mmc_get_discmode}, then it is an externally visible MMC routine. (We don't have a separate library for this yet. Names using entirely capital letters and that start @code{CDIO_} are externally visible @code{#defines}. @node ISO-9660 Character Sets @appendix ISO-9660 Character Sets For a description of where are used see @xref{ISO 9660 Level 1}. @menu * ISO646 d-Characters:: * ISO646 a-Characters:: @end menu @node ISO646 d-Characters @appendixsec ISO646 d-Characters @example | 0 1 2 3 4 5 6 7 --+----------------- 0 | 0 P 1 | 1 A Q 2 | 2 B R 3 | 3 C S 4 | 4 D T 5 | 5 E U 6 | 6 F V 7 | 7 G W 8 | 8 H X 9 | 9 I Y a | J Z b | K c | L d | M e | N f | O _ @end example @node ISO646 a-Characters @appendixsec ISO646 a-Characters @example | 0 1 2 3 4 5 6 7 --+----------------- 0 | 0 P 1 | ! 1 A Q 2 | " 2 B R 3 | 3 C S 4 | 4 D T 5 | % 5 E U 6 | & 6 F V 7 | ' 7 G W 8 | ( 8 H X 9 | ) 9 I Y a | * : J Z b | + ; K c | , < L d | - = M e | . > N f | / ? O _ @end example @node Glossary @appendix Glossary @include glossary.texi @node GNU Free Documentation License @appendix GNU Free Documentation License @cindex FDL, GNU Free Documentation License @include fdl.texi @node General Index @unnumbered General Index @printindex cp @bye