=encoding utf8 =head1 NAME guestfs - Library for accessing and modifying virtual machine images =head1 SYNOPSIS #include guestfs_h *g = guestfs_create (); guestfs_add_drive (g, "guest.img"); guestfs_launch (g); guestfs_mount (g, "/dev/sda1", "/"); guestfs_touch (g, "/hello"); guestfs_umount (g, "/"); guestfs_close (g); cc prog.c -o prog -lguestfs or: cc prog.c -o prog `pkg-config libguestfs --cflags --libs` =head1 DESCRIPTION Libguestfs is a library for accessing and modifying guest disk images. Amongst the things this is good for: making batch configuration changes to guests, getting disk used/free statistics (see also: virt-df), migrating between virtualization systems (see also: virt-p2v), performing partial backups, performing partial guest clones, cloning guests and changing registry/UUID/hostname info, and much else besides. Libguestfs uses Linux kernel and qemu code, and can access any type of guest filesystem that Linux and qemu can, including but not limited to: ext2/3/4, btrfs, FAT and NTFS, LVM, many different disk partition schemes, qcow, qcow2, vmdk. Libguestfs provides ways to enumerate guest storage (eg. partitions, LVs, what filesystem is in each LV, etc.). It can also run commands in the context of the guest. Also you can access filesystems over FUSE. Libguestfs is a library that can be linked with C and C++ management programs (or management programs written in OCaml, Perl, Python, Ruby, Java, PHP, Haskell or C#). You can also use it from shell scripts or the command line. You don't need to be root to use libguestfs, although obviously you do need enough permissions to access the disk images. Libguestfs is a large API because it can do many things. For a gentle introduction, please read the L section next. There are also some example programs in the L manual page. =head1 API OVERVIEW This section provides a gentler overview of the libguestfs API. We also try to group API calls together, where that may not be obvious from reading about the individual calls in the main section of this manual. =head2 HANDLES Before you can use libguestfs calls, you have to create a handle. Then you must add at least one disk image to the handle, followed by launching the handle, then performing whatever operations you want, and finally closing the handle. By convention we use the single letter C for the name of the handle variable, although of course you can use any name you want. The general structure of all libguestfs-using programs looks like this: guestfs_h *g = guestfs_create (); /* Call guestfs_add_drive additional times if there are * multiple disk images. */ guestfs_add_drive (g, "guest.img"); /* Most manipulation calls won't work until you've launched * the handle 'g'. You have to do this _after_ adding drives * and _before_ other commands. */ guestfs_launch (g); /* Now you can examine what partitions, LVs etc are available. */ char **partitions = guestfs_list_partitions (g); char **logvols = guestfs_lvs (g); /* To access a filesystem in the image, you must mount it. */ guestfs_mount (g, "/dev/sda1", "/"); /* Now you can perform filesystem actions on the guest * disk image. */ guestfs_touch (g, "/hello"); /* This is only needed for libguestfs < 1.5.24. Since then * it is done automatically when you close the handle. See * discussion of autosync in this page. */ guestfs_sync (g); /* Close the handle 'g'. */ guestfs_close (g); The code above doesn't include any error checking. In real code you should check return values carefully for errors. In general all functions that return integers return C<-1> on error, and all functions that return pointers return C on error. See section L below for how to handle errors, and consult the documentation for each function call below to see precisely how they return error indications. See L for fully worked examples. =head2 DISK IMAGES The image filename (C<"guest.img"> in the example above) could be a disk image from a virtual machine, a L copy of a physical hard disk, an actual block device, or simply an empty file of zeroes that you have created through L. Libguestfs lets you do useful things to all of these. The call you should use in modern code for adding drives is L. To add a disk image, allowing writes, and specifying that the format is raw, do: guestfs_add_drive_opts (g, filename, GUESTFS_ADD_DRIVE_OPTS_FORMAT, "raw", -1); You can add a disk read-only using: guestfs_add_drive_opts (g, filename, GUESTFS_ADD_DRIVE_OPTS_FORMAT, "raw", GUESTFS_ADD_DRIVE_OPTS_READONLY, 1, -1); or by calling the older function L. In either case libguestfs won't modify the file. Be extremely cautious if the disk image is in use, eg. if it is being used by a virtual machine. Adding it read-write will almost certainly cause disk corruption, but adding it read-only is safe. You must add at least one disk image, and you may add multiple disk images. In the API, the disk images are usually referred to as C (for the first one you added), C (for the second one you added), etc. Once L has been called you cannot add any more images. You can call L to get a list of the device names, in the order that you added them. See also L below. =head2 MOUNTING Before you can read or write files, create directories and so on in a disk image that contains filesystems, you have to mount those filesystems using L or L. If you already know that a disk image contains (for example) one partition with a filesystem on that partition, then you can mount it directly: guestfs_mount_options (g, "", "/dev/sda1", "/"); where C means literally the first partition (C<1>) of the first disk image that we added (C). If the disk contains Linux LVM2 logical volumes you could refer to those instead (eg. C). Note that these are libguestfs virtual devices, and are nothing to do with host devices. If you are given a disk image and you don't know what it contains then you have to find out. Libguestfs can do that too: use L and L to list possible partitions and LVs, and either try mounting each to see what is mountable, or else examine them with L or L. To list just filesystems, use L. Libguestfs also has a set of APIs for inspection of unknown disk images (see L below). But you might find it easier to look at higher level programs built on top of libguestfs, in particular L. To mount a filesystem read-only, use L. There are several other variations of the C call. =head2 FILESYSTEM ACCESS AND MODIFICATION The majority of the libguestfs API consists of fairly low-level calls for accessing and modifying the files, directories, symlinks etc on mounted filesystems. There are over a hundred such calls which you can find listed in detail below in this man page, and we don't even pretend to cover them all in this overview. Specify filenames as full paths, starting with C<"/"> and including the mount point. For example, if you mounted a filesystem at C<"/"> and you want to read the file called C<"etc/passwd"> then you could do: char *data = guestfs_cat (g, "/etc/passwd"); This would return C as a newly allocated buffer containing the full content of that file (with some conditions: see also L below), or C if there was an error. As another example, to create a top-level directory on that filesystem called C<"var"> you would do: guestfs_mkdir (g, "/var"); To create a symlink you could do: guestfs_ln_s (g, "/etc/init.d/portmap", "/etc/rc3.d/S30portmap"); Libguestfs will reject attempts to use relative paths and there is no concept of a current working directory. Libguestfs can return errors in many situations: for example if the filesystem isn't writable, or if a file or directory that you requested doesn't exist. If you are using the C API (documented here) you have to check for those error conditions after each call. (Other language bindings turn these errors into exceptions). File writes are affected by the per-handle umask, set by calling L and defaulting to 022. See L. =head2 PARTITIONING Libguestfs contains API calls to read, create and modify partition tables on disk images. In the common case where you want to create a single partition covering the whole disk, you should use the L call: const char *parttype = "mbr"; if (disk_is_larger_than_2TB) parttype = "gpt"; guestfs_part_disk (g, "/dev/sda", parttype); Obviously this effectively wipes anything that was on that disk image before. =head2 LVM2 Libguestfs provides access to a large part of the LVM2 API, such as L and L. It won't make much sense unless you familiarize yourself with the concepts of physical volumes, volume groups and logical volumes. This author strongly recommends reading the LVM HOWTO, online at L. =head2 DOWNLOADING Use L to download small, text only files. This call is limited to files which are less than 2 MB and which cannot contain any ASCII NUL (C<\0>) characters. However the API is very simple to use. L can be used to read files which contain arbitrary 8 bit data, since it returns a (pointer, size) pair. However it is still limited to "small" files, less than 2 MB. L can be used to download any file, with no limits on content or size (even files larger than 4 GB). To download multiple files, see L and L. =head2 UPLOADING It's often the case that you want to write a file or files to the disk image. To write a small file with fixed content, use L. To create a file of all zeroes, use L (sparse) or L (with all disk blocks allocated). There are a variety of other functions for creating test files, for example L and L. To upload a single file, use L. This call has no limits on file content or size (even files larger than 4 GB). To upload multiple files, see L and L. However the fastest way to upload I is to turn them into a squashfs or CD ISO (see L and L), then attach this using L. If you add the drive in a predictable way (eg. adding it last after all other drives) then you can get the device name from L and mount it directly using L. Note that squashfs images are sometimes non-portable between kernel versions, and they don't support labels or UUIDs. If you want to pre-build an image or you need to mount it using a label or UUID, use an ISO image instead. =head2 COPYING There are various different commands for copying between files and devices and in and out of the guest filesystem. These are summarised in the table below. =over 4 =item B to B Use L to copy a single file, or L to copy directories recursively. =item B to B Use L which efficiently uses L to copy between files and devices in the guest. Example: duplicate the contents of an LV: guestfs_dd (g, "/dev/VG/Original", "/dev/VG/Copy"); The destination (C) must be at least as large as the source (C). To copy less than the whole source device, use L. =item B to B Use L. See L above. =item B to B Use L. See L above. =back =head2 UPLOADING AND DOWNLOADING TO PIPES AND FILE DESCRIPTORS Calls like L, L, L, L etc appear to only take filenames as arguments, so it appears you can only upload and download to files. However many Un*x-like hosts let you use the special device files C, C, C and C to read and write from stdin, stdout, stderr, and arbitrary file descriptor N. For example, L writes its output to stdout by doing: guestfs_download (g, filename, "/dev/stdout"); and you can write tar output to a file descriptor C by doing: char devfd[64]; snprintf (devfd, sizeof devfd, "/dev/fd/%d", fd); guestfs_tar_out (g, "/", devfd); =head2 LISTING FILES L is just designed for humans to read (mainly when using the L-equivalent command C). L is a quick way to get a list of files in a directory from programs, as a flat list of strings. L is a programmatic way to get a list of files in a directory, plus additional information about each one. It is more equivalent to using the L call on a local filesystem. L and L can be used to recursively list files. =head2 RUNNING COMMANDS Although libguestfs is primarily an API for manipulating files inside guest images, we also provide some limited facilities for running commands inside guests. There are many limitations to this: =over 4 =item * The kernel version that the command runs under will be different from what it expects. =item * If the command needs to communicate with daemons, then most likely they won't be running. =item * The command will be running in limited memory. =item * The network may not be available unless you enable it (see L). =item * Only supports Linux guests (not Windows, BSD, etc). =item * Architecture limitations (eg. won't work for a PPC guest on an X86 host). =item * For SELinux guests, you may need to enable SELinux and load policy first. See L in this manpage. =item * I It is not safe to run commands from untrusted, possibly malicious guests. These commands may attempt to exploit your program by sending unexpected output. They could also try to exploit the Linux kernel or qemu provided by the libguestfs appliance. They could use the network provided by the libguestfs appliance to bypass ordinary network partitions and firewalls. They could use the elevated privileges or different SELinux context of your program to their advantage. A secure alternative is to use libguestfs to install a "firstboot" script (a script which runs when the guest next boots normally), and to have this script run the commands you want in the normal context of the running guest, network security and so on. For information about other security issues, see L. =back The two main API calls to run commands are L and L (there are also variations). The difference is that L runs commands using the shell, so any shell globs, redirections, etc will work. =head2 CONFIGURATION FILES To read and write configuration files in Linux guest filesystems, we strongly recommend using Augeas. For example, Augeas understands how to read and write, say, a Linux shadow password file or X.org configuration file, and so avoids you having to write that code. The main Augeas calls are bound through the C APIs. We don't document Augeas itself here because there is excellent documentation on the L website. If you don't want to use Augeas (you fool!) then try calling L to get the file as a list of lines which you can iterate over. =head2 SELINUX We support SELinux guests. To ensure that labeling happens correctly in SELinux guests, you need to enable SELinux and load the guest's policy: =over 4 =item 1. Before launching, do: guestfs_set_selinux (g, 1); =item 2. After mounting the guest's filesystem(s), load the policy. This is best done by running the L command in the guest itself: guestfs_sh (g, "/usr/sbin/load_policy"); (Older versions of C require you to specify the name of the policy file). =item 3. Optionally, set the security context for the API. The correct security context to use can only be known by inspecting the guest. As an example: guestfs_setcon (g, "unconfined_u:unconfined_r:unconfined_t:s0"); =back This will work for running commands and editing existing files. When new files are created, you may need to label them explicitly, for example by running the external command C. =head2 UMASK Certain calls are affected by the current file mode creation mask (the "umask"). In particular ones which create files or directories, such as L, L or L. This affects either the default mode that the file is created with or modifies the mode that you supply. The default umask is C<022>, so files are created with modes such as C<0644> and directories with C<0755>. There are two ways to avoid being affected by umask. Either set umask to 0 (call C early after launching). Or call L after creating each file or directory. For more information about umask, see L. =head2 ENCRYPTED DISKS Libguestfs allows you to access Linux guests which have been encrypted using whole disk encryption that conforms to the Linux Unified Key Setup (LUKS) standard. This includes nearly all whole disk encryption systems used by modern Linux guests. Use L to identify LUKS-encrypted block devices (it returns the string C). Then open these devices by calling L. Obviously you will require the passphrase! Opening a LUKS device creates a new device mapper device called C (where C is the string you supply to L). Reads and writes to this mapper device are decrypted from and encrypted to the underlying block device respectively. LVM volume groups on the device can be made visible by calling L followed by L. The logical volume(s) can now be mounted in the usual way. Use the reverse process to close a LUKS device. Unmount any logical volumes on it, deactivate the volume groups by caling C. Then close the mapper device by calling L on the C device (I the underlying encrypted block device). =head2 INSPECTION Libguestfs has APIs for inspecting an unknown disk image to find out if it contains operating systems, an install CD or a live CD. (These APIs used to be in a separate Perl-only library called L but since version 1.5.3 the most frequently used part of this library has been rewritten in C and moved into the core code). Add all disks belonging to the unknown virtual machine and call L in the usual way. Then call L. This function uses other libguestfs calls and certain heuristics, and returns a list of operating systems that were found. An empty list means none were found. A single element is the root filesystem of the operating system. For dual- or multi-boot guests, multiple roots can be returned, each one corresponding to a separate operating system. (Multi-boot virtual machines are extremely rare in the world of virtualization, but since this scenario can happen, we have built libguestfs to deal with it.) For each root, you can then call various C functions to get additional details about that operating system. For example, call L to return the string C or C for Windows and Linux-based operating systems respectively. Un*x-like and Linux-based operating systems usually consist of several filesystems which are mounted at boot time (for example, a separate boot partition mounted on C). The inspection rules are able to detect how filesystems correspond to mount points. Call C to get this mapping. It might return a hash table like this example: /boot => /dev/sda1 / => /dev/vg_guest/lv_root /usr => /dev/vg_guest/lv_usr The caller can then make calls to L to mount the filesystems as suggested. Be careful to mount filesystems in the right order (eg. C before C). Sorting the keys of the hash by length, shortest first, should work. Inspection currently only works for some common operating systems. Contributors are welcome to send patches for other operating systems that we currently cannot detect. Encrypted disks must be opened before inspection. See L for more details. The L function just ignores any encrypted devices. A note on the implementation: The call L performs inspection and caches the results in the guest handle. Subsequent calls to C return this cached information, but I re-read the disks. If you change the content of the guest disks, you can redo inspection by calling L again. (L works a little differently from the other calls and does read the disks. See documentation for that function for details). =head3 INSPECTING INSTALL DISKS Libguestfs (since 1.9.4) can detect some install disks, install CDs, live CDs and more. Call L to return the format of the operating system, which currently can be C (a regular operating system) or C (some sort of install disk). Further information is available about the operating system that can be installed using the regular inspection APIs like L, L etc. Some additional information specific to installer disks is also available from the L, L and L calls. =head2 SPECIAL CONSIDERATIONS FOR WINDOWS GUESTS Libguestfs can mount NTFS partitions. It does this using the L driver. =head3 DRIVE LETTERS AND PATHS DOS and Windows still use drive letters, and the filesystems are always treated as case insensitive by Windows itself, and therefore you might find a Windows configuration file referring to a path like C. When the filesystem is mounted in libguestfs, that directory might be referred to as C. Drive letter mappings can be found using inspection (see L and L) Dealing with separator characters (backslash vs forward slash) is outside the scope of libguestfs, but usually a simple character replacement will work. To resolve the case insensitivity of paths, call L. =head3 ACCESSING THE WINDOWS REGISTRY Libguestfs also provides some help for decoding Windows Registry "hive" files, through the library C which is part of the libguestfs project although ships as a separate tarball. You have to locate and download the hive file(s) yourself, and then pass them to C functions. See also the programs L, L, L and L for more help on this issue. =head3 SYMLINKS ON NTFS-3G FILESYSTEMS Ntfs-3g tries to rewrite "Junction Points" and NTFS "symbolic links" to provide something which looks like a Linux symlink. The way it tries to do the rewriting is described here: L The essential problem is that ntfs-3g simply does not have enough information to do a correct job. NTFS links can contain drive letters and references to external device GUIDs that ntfs-3g has no way of resolving. It is almost certainly the case that libguestfs callers should ignore what ntfs-3g does (ie. don't use L on NTFS volumes). Instead if you encounter a symbolic link on an ntfs-3g filesystem, use L to read the C extended attribute, and read the raw reparse data from that (you can find the format documented in various places around the web). =head3 EXTENDED ATTRIBUTES ON NTFS-3G FILESYSTEMS There are other useful extended attributes that can be read from ntfs-3g filesystems (using L). See: L =head2 USING LIBGUESTFS WITH OTHER PROGRAMMING LANGUAGES Although we don't want to discourage you from using the C API, we will mention here that the same API is also available in other languages. The API is broadly identical in all supported languages. This means that the C call C is C<$g-Eadd_drive_ro($file)> in Perl, C in Python, and C in OCaml. In other words, a straightforward, predictable isomorphism between each language. Error messages are automatically transformed into exceptions if the language supports it. We don't try to "object orientify" parts of the API in OO languages, although contributors are welcome to write higher level APIs above what we provide in their favourite languages if they wish. =over 4 =item B You can use the I header file from C++ programs. The C++ API is identical to the C API. C++ classes and exceptions are not used. =item B The C# bindings are highly experimental. Please read the warnings at the top of C. =item B This is the only language binding that is working but incomplete. Only calls which return simple integers have been bound in Haskell, and we are looking for help to complete this binding. =item B Full documentation is contained in the Javadoc which is distributed with libguestfs. For examples, see L. =item B See L. =item B See L and L. =item B For documentation see C supplied with libguestfs sources or in the php-libguestfs package for your distribution. The PHP binding only works correctly on 64 bit machines. =item B See L. =item B See L. =item B See L. =back =head2 LIBGUESTFS GOTCHAS L: "A feature of a system [...] that works in the way it is documented but is counterintuitive and almost invites mistakes." Since we developed libguestfs and the associated tools, there are several things we would have designed differently, but are now stuck with for backwards compatibility or other reasons. If there is ever a libguestfs 2.0 release, you can expect these to change. Beware of them. =over 4 =item Autosync / forgetting to sync. I Autosync is enabled by default for all API users starting from libguestfs 1.5.24. This section only applies to older versions. When modifying a filesystem from C or another language, you B unmount all filesystems and call L explicitly before you close the libguestfs handle. You can also call: guestfs_set_autosync (g, 1); to have the unmount/sync done automatically for you when the handle 'g' is closed. (This feature is called "autosync", L q.v.) If you forget to do this, then it is entirely possible that your changes won't be written out, or will be partially written, or (very rarely) that you'll get disk corruption. Note that in L autosync is the default. So quick and dirty guestfish scripts that forget to sync will work just fine, which can make this very puzzling if you are trying to debug a problem. =item Mount option C<-o sync> should not be the default. If you use L, then C<-o sync,noatime> are added implicitly. However C<-o sync> does not add any reliability benefit, but does have a very large performance impact. The work around is to use L and set the mount options that you actually want to use. =item Read-only should be the default. In L, I<--ro> should be the default, and you should have to specify I<--rw> if you want to make changes to the image. This would reduce the potential to corrupt live VM images. Note that many filesystems change the disk when you just mount and unmount, even if you didn't perform any writes. You need to use L to guarantee that the disk is not changed. =item guestfish command line is hard to use. C doesn't do what people expect (open C for examination). It tries to run a guestfish command C which doesn't exist, so it fails. In earlier versions of guestfish the error message was also unintuitive, but we have corrected this since. Like the Bourne shell, we should have used C to run commands. =item guestfish megabyte modifiers don't work right on all commands In recent guestfish you can use C<1M> to mean 1 megabyte (and similarly for other modifiers). What guestfish actually does is to multiply the number part by the modifier part and pass the result to the C API. However this doesn't work for a few APIs which aren't expecting bytes, but are already expecting some other unit (eg. megabytes). The most common is L. The guestfish command: lvcreate LV VG 100M does not do what you might expect. Instead because L is already expecting megabytes, this tries to create a 100 I (100 megabytes * megabytes) logical volume. The error message you get from this is also a little obscure. This could be fixed in the generator by specially marking parameters and return values which take bytes or other units. =item Ambiguity between devices and paths There is a subtle ambiguity in the API between a device name (eg. C) and a similar pathname. A file might just happen to be called C in the directory C (consider some non-Unix VM image). In the current API we usually resolve this ambiguity by having two separate calls, for example L and L. Some API calls are ambiguous and (incorrectly) resolve the problem by detecting if the path supplied begins with C. To avoid both the ambiguity and the need to duplicate some calls, we could make paths/devices into structured names. One way to do this would be to use a notation like grub (C), although nobody really likes this aspect of grub. Another way would be to use a structured type, equivalent to this OCaml type: type path = Path of string | Device of int | Partition of int * int which would allow you to pass arguments like: Path "/foo/bar" Device 1 (* /dev/sdb, or perhaps /dev/sda *) Partition (1, 2) (* /dev/sdb2 (or is it /dev/sda2 or /dev/sdb3?) *) Path "/dev/sdb2" (* not a device *) As you can see there are still problems to resolve even with this representation. Also consider how it might work in guestfish. =back =head2 KEYS AND PASSPHRASES Certain libguestfs calls take a parameter that contains sensitive key material, passed in as a C string. In the future we would hope to change the libguestfs implementation so that keys are L-ed into physical RAM, and thus can never end up in swap. However this is I done at the moment, because of the complexity of such an implementation. Therefore you should be aware that any key parameter you pass to libguestfs might end up being written out to the swap partition. If this is a concern, scrub the swap partition or don't use libguestfs on encrypted devices. =head2 MULTIPLE HANDLES AND MULTIPLE THREADS All high-level libguestfs actions are synchronous. If you want to use libguestfs asynchronously then you must create a thread. Only use the handle from a single thread. Either use the handle exclusively from one thread, or provide your own mutex so that two threads cannot issue calls on the same handle at the same time. See the graphical program guestfs-browser for one possible architecture for multithreaded programs using libvirt and libguestfs. =head2 PATH Libguestfs needs a supermin appliance, which it finds by looking along an internal path. By default it looks for these in the directory C<$libdir/guestfs> (eg. C or C). Use L or set the environment variable L to change the directories that libguestfs will search in. The value is a colon-separated list of paths. The current directory is I searched unless the path contains an empty element or C<.>. For example C would search the current directory and then C. =head2 QEMU WRAPPERS If you want to compile your own qemu, run qemu from a non-standard location, or pass extra arguments to qemu, then you can write a shell-script wrapper around qemu. There is one important rule to remember: you I> as the last command in the shell script (so that qemu replaces the shell and becomes the direct child of the libguestfs-using program). If you don't do this, then the qemu process won't be cleaned up correctly. Here is an example of a wrapper, where I have built my own copy of qemu from source: #!/bin/sh - qemudir=/home/rjones/d/qemu exec $qemudir/x86_64-softmmu/qemu-system-x86_64 -L $qemudir/pc-bios "$@" Save this script as C (or wherever), C, and then use it by setting the LIBGUESTFS_QEMU environment variable. For example: LIBGUESTFS_QEMU=/tmp/qemu.wrapper guestfish Note that libguestfs also calls qemu with the -help and -version options in order to determine features. =head2 ATTACHING TO RUNNING DAEMONS I This is B and has a tendency to eat babies. Use with caution. I This section explains how to attach to a running daemon from a low level perspective. For most users, simply using virt tools such as L with the I<--live> option will "just work". =head3 Using guestfs_set_attach_method By calling L you can change how the library connects to the C daemon in L (read L for some background). The normal attach method is C, where a small appliance is created containing the daemon, and then the library connects to this. Setting attach method to C> (where I is the path of a Unix domain socket) causes L to connect to an existing daemon over the Unix domain socket. The normal use for this is to connect to a running virtual machine that contains a C daemon, and send commands so you can read and write files inside the live virtual machine. =head3 Using guestfs_add_domain with live flag L provides some help for getting the correct attach method. If you pass the C option to this function, then (if the virtual machine is running) it will examine the libvirt XML looking for a virtio-serial channel to connect to: ... ... ... L extracts C and sets the attach method to C. Some of the libguestfs tools (including guestfish) support a I<--live> option which is passed through to L thus allowing you to attach to and modify live virtual machines. The virtual machine needs to have been set up beforehand so that it has the virtio-serial channel and so that guestfsd is running inside it. =head2 ABI GUARANTEE We guarantee the libguestfs ABI (binary interface), for public, high-level actions as outlined in this section. Although we will deprecate some actions, for example if they get replaced by newer calls, we will keep the old actions forever. This allows you the developer to program in confidence against the libguestfs API. =head2 BLOCK DEVICE NAMING In the kernel there is now quite a profusion of schemata for naming block devices (in this context, by I I mean a physical or virtual hard drive). The original Linux IDE driver used names starting with C. SCSI devices have historically used a different naming scheme, C. When the Linux kernel I driver became a popular replacement for the old IDE driver (particularly for SATA devices) those devices also used the C scheme. Additionally we now have virtual machines with paravirtualized drivers. This has created several different naming systems, such as C for virtio disks and C for Xen PV disks. As discussed above, libguestfs uses a qemu appliance running an embedded Linux kernel to access block devices. We can run a variety of appliances based on a variety of Linux kernels. This causes a problem for libguestfs because many API calls use device or partition names. Working scripts and the recipe (example) scripts that we make available over the internet could fail if the naming scheme changes. Therefore libguestfs defines C as the I. Internally C names are translated, if necessary, to other names as required. For example, under RHEL 5 which uses the C scheme, any device parameter C is translated to C transparently. Note that this I applies to parameters. The L, L and similar calls return the true names of the devices and partitions as known to the appliance. =head3 ALGORITHM FOR BLOCK DEVICE NAME TRANSLATION Usually this translation is transparent. However in some (very rare) cases you may need to know the exact algorithm. Such cases include where you use L to add a mixture of virtio and IDE devices to the qemu-based appliance, so have a mixture of C and C devices. The algorithm is applied only to I which are known to be either device or partition names. Return values from functions such as L are never changed. =over 4 =item * Is the string a parameter which is a device or partition name? =item * Does the string begin with C? =item * Does the named device exist? If so, we use that device. However if I then we continue with this algorithm. =item * Replace initial C string with C. For example, change C to C. If that named device exists, use it. If not, continue. =item * Replace initial C string with C. If that named device exists, use it. If not, return an error. =back =head3 PORTABILITY CONCERNS WITH BLOCK DEVICE NAMING Although the standard naming scheme and automatic translation is useful for simple programs and guestfish scripts, for larger programs it is best not to rely on this mechanism. Where possible for maximum future portability programs using libguestfs should use these future-proof techniques: =over 4 =item * Use L or L to list actual device names, and then use those names directly. Since those device names exist by definition, they will never be translated. =item * Use higher level ways to identify filesystems, such as LVM names, UUIDs and filesystem labels. =back =head1 SECURITY This section discusses security implications of using libguestfs, particularly with untrusted or malicious guests or disk images. =head2 GENERAL SECURITY CONSIDERATIONS Be careful with any files or data that you download from a guest (by "download" we mean not just the L command but any command that reads files, filenames, directories or anything else from a disk image). An attacker could manipulate the data to fool your program into doing the wrong thing. Consider cases such as: =over 4 =item * the data (file etc) not being present =item * being present but empty =item * being much larger than normal =item * containing arbitrary 8 bit data =item * being in an unexpected character encoding =item * containing homoglyphs. =back =head2 SECURITY OF MOUNTING FILESYSTEMS When you mount a filesystem under Linux, mistakes in the kernel filesystem (VFS) module can sometimes be escalated into exploits by deliberately creating a malicious, malformed filesystem. These exploits are very severe for two reasons. Firstly there are very many filesystem drivers in the kernel, and many of them are infrequently used and not much developer attention has been paid to the code. Linux userspace helps potential crackers by detecting the filesystem type and automatically choosing the right VFS driver, even if that filesystem type is obscure or unexpected for the administrator. Secondly, a kernel-level exploit is like a local root exploit (worse in some ways), giving immediate and total access to the system right down to the hardware level. That explains why you should never mount a filesystem from an untrusted guest on your host kernel. How about libguestfs? We run a Linux kernel inside a qemu virtual machine, usually running as a non-root user. The attacker would need to write a filesystem which first exploited the kernel, and then exploited either qemu virtualization (eg. a faulty qemu driver) or the libguestfs protocol, and finally to be as serious as the host kernel exploit it would need to escalate its privileges to root. This multi-step escalation, performed by a static piece of data, is thought to be extremely hard to do, although we never say 'never' about security issues. In any case callers can reduce the attack surface by forcing the filesystem type when mounting (use L). =head2 PROTOCOL SECURITY The protocol is designed to be secure, being based on RFC 4506 (XDR) with a defined upper message size. However a program that uses libguestfs must also take care - for example you can write a program that downloads a binary from a disk image and executes it locally, and no amount of protocol security will save you from the consequences. =head2 INSPECTION SECURITY Parts of the inspection API (see L) return untrusted strings directly from the guest, and these could contain any 8 bit data. Callers should be careful to escape these before printing them to a structured file (for example, use HTML escaping if creating a web page). Guest configuration may be altered in unusual ways by the administrator of the virtual machine, and may not reflect reality (particularly for untrusted or actively malicious guests). For example we parse the hostname from configuration files like C that we find in the guest, but the guest administrator can easily manipulate these files to provide the wrong hostname. The inspection API parses guest configuration using two external libraries: Augeas (Linux configuration) and hivex (Windows Registry). Both are designed to be robust in the face of malicious data, although denial of service attacks are still possible, for example with oversized configuration files. =head2 RUNNING UNTRUSTED GUEST COMMANDS Be very cautious about running commands from the guest. By running a command in the guest, you are giving CPU time to a binary that you do not control, under the same user account as the library, albeit wrapped in qemu virtualization. More information and alternatives can be found in the section L. =head2 CVE-2010-3851 https://bugzilla.redhat.com/642934 This security bug concerns the automatic disk format detection that qemu does on disk images. A raw disk image is just the raw bytes, there is no header. Other disk images like qcow2 contain a special header. Qemu deals with this by looking for one of the known headers, and if none is found then assuming the disk image must be raw. This allows a guest which has been given a raw disk image to write some other header. At next boot (or when the disk image is accessed by libguestfs) qemu would do autodetection and think the disk image format was, say, qcow2 based on the header written by the guest. This in itself would not be a problem, but qcow2 offers many features, one of which is to allow a disk image to refer to another image (called the "backing disk"). It does this by placing the path to the backing disk into the qcow2 header. This path is not validated and could point to any host file (eg. "/etc/passwd"). The backing disk is then exposed through "holes" in the qcow2 disk image, which of course is completely under the control of the attacker. In libguestfs this is rather hard to exploit except under two circumstances: =over 4 =item 1. You have enabled the network or have opened the disk in write mode. =item 2. You are also running untrusted code from the guest (see L). =back The way to avoid this is to specify the expected disk format when adding disks (the optional C option to L). You should always do this if the disk is raw format, and it's a good idea for other cases too. For disks added from libvirt using calls like L, the format is fetched from libvirt and passed through. For libguestfs tools, use the I<--format> command line parameter as appropriate. =head1 CONNECTION MANAGEMENT =head2 guestfs_h * C is the opaque type representing a connection handle. Create a handle by calling L. Call L to free the handle and release all resources used. For information on using multiple handles and threads, see the section L above. =head2 guestfs_create guestfs_h *guestfs_create (void); Create a connection handle. On success this returns a non-NULL pointer to a handle. On error it returns NULL. You have to "configure" the handle after creating it. This includes calling L (or one of the equivalent calls) on the handle at least once. After configuring the handle, you have to call L. You may also want to configure error handling for the handle. See the L section below. =head2 guestfs_close void guestfs_close (guestfs_h *g); This closes the connection handle and frees up all resources used. If autosync was set on the handle and the handle was launched, then this implicitly calls various functions to unmount filesystems and sync the disk. See L for more details. If a close callback was set on the handle, then it is called. =head1 ERROR HANDLING API functions can return errors. For example, almost all functions that return C will return C<-1> to indicate an error. Additional information is available for errors: an error message string and optionally an error number (errno) if the thing that failed was a system call. You can get at the additional information about the last error on the handle by calling L, L, and/or by setting up an error handler with L. When the handle is created, a default error handler is installed which prints the error message string to C. For small short-running command line programs it is sufficient to do: if (guestfs_launch (g) == -1) exit (EXIT_FAILURE); since the default error handler will ensure that an error message has been printed to C before the program exits. For other programs the caller will almost certainly want to install an alternate error handler or do error handling in-line like this: g = guestfs_create (); /* This disables the default behaviour of printing errors on stderr. */ guestfs_set_error_handler (g, NULL, NULL); if (guestfs_launch (g) == -1) { /* Examine the error message and print it etc. */ char *msg = guestfs_last_error (g); int errnum = guestfs_last_errno (g); fprintf (stderr, "%s\n", msg); /* ... */ } Out of memory errors are handled differently. The default action is to call L. If this is undesirable, then you can set a handler using L. L returns C if the handle cannot be created, and because there is no handle if this happens there is no way to get additional error information. However L is supposed to be a lightweight operation which can only fail because of insufficient memory (it returns NULL in this case). =head2 guestfs_last_error const char *guestfs_last_error (guestfs_h *g); This returns the last error message that happened on C. If there has not been an error since the handle was created, then this returns C. The lifetime of the returned string is until the next error occurs, or L is called. =head2 guestfs_last_errno int guestfs_last_errno (guestfs_h *g); This returns the last error number (errno) that happened on C. If successful, an errno integer not equal to zero is returned. If no error, this returns 0. This call can return 0 in three situations: =over 4 =item 1. There has not been any error on the handle. =item 2. There has been an error but the errno was meaningless. This corresponds to the case where the error did not come from a failed system call, but for some other reason. =item 3. There was an error from a failed system call, but for some reason the errno was not captured and returned. This usually indicates a bug in libguestfs. =back Libguestfs tries to convert the errno from inside the applicance into a corresponding errno for the caller (not entirely trivial: the appliance might be running a completely different operating system from the library and error numbers are not standardized across Un*xen). If this could not be done, then the error is translated to C. In practice this should only happen in very rare circumstances. =head2 guestfs_set_error_handler typedef void (*guestfs_error_handler_cb) (guestfs_h *g, void *opaque, const char *msg); void guestfs_set_error_handler (guestfs_h *g, guestfs_error_handler_cb cb, void *opaque); The callback C will be called if there is an error. The parameters passed to the callback are an opaque data pointer and the error message string. C is not passed to the callback. To get that the callback must call L. Note that the message string C is freed as soon as the callback function returns, so if you want to stash it somewhere you must make your own copy. The default handler prints messages on C. If you set C to C then I handler is called. =head2 guestfs_get_error_handler guestfs_error_handler_cb guestfs_get_error_handler (guestfs_h *g, void **opaque_rtn); Returns the current error handler callback. =head2 guestfs_set_out_of_memory_handler typedef void (*guestfs_abort_cb) (void); int guestfs_set_out_of_memory_handler (guestfs_h *g, guestfs_abort_cb); The callback C will be called if there is an out of memory situation. I. The default is to call L. You cannot set C to C. You can't ignore out of memory situations. =head2 guestfs_get_out_of_memory_handler guestfs_abort_fn guestfs_get_out_of_memory_handler (guestfs_h *g); This returns the current out of memory handler. =head1 API CALLS @ACTIONS@ =head1 STRUCTURES @STRUCTS@ =head1 AVAILABILITY =head2 GROUPS OF FUNCTIONALITY IN THE APPLIANCE Using L you can test availability of the following groups of functions. This test queries the appliance to see if the appliance you are currently using supports the functionality. @AVAILABILITY@ =head2 GUESTFISH supported COMMAND In L there is a handy interactive command C which prints out the available groups and whether they are supported by this build of libguestfs. Note however that you have to do C first. =head2 SINGLE CALLS AT COMPILE TIME Since version 1.5.8, Cguestfs.hE> defines symbols for each C API function, such as: #define LIBGUESTFS_HAVE_DD 1 if L is available. Before version 1.5.8, if you needed to test whether a single libguestfs function is available at compile time, we recommended using build tools such as autoconf or cmake. For example in autotools you could use: AC_CHECK_LIB([guestfs],[guestfs_create]) AC_CHECK_FUNCS([guestfs_dd]) which would result in C being either defined or not defined in your program. =head2 SINGLE CALLS AT RUN TIME Testing at compile time doesn't guarantee that a function really exists in the library. The reason is that you might be dynamically linked against a previous I (dynamic library) which doesn't have the call. This situation unfortunately results in a segmentation fault, which is a shortcoming of the C dynamic linking system itself. You can use L to test if a function is available at run time, as in this example program (note that you still need the compile time check as well): #include #include #include #include #include main () { #ifdef LIBGUESTFS_HAVE_DD void *dl; int has_function; /* Test if the function guestfs_dd is really available. */ dl = dlopen (NULL, RTLD_LAZY); if (!dl) { fprintf (stderr, "dlopen: %s\n", dlerror ()); exit (EXIT_FAILURE); } has_function = dlsym (dl, "guestfs_dd") != NULL; dlclose (dl); if (!has_function) printf ("this libguestfs.so does NOT have guestfs_dd function\n"); else { printf ("this libguestfs.so has guestfs_dd function\n"); /* Now it's safe to call guestfs_dd (g, "foo", "bar"); */ } #else printf ("guestfs_dd function was not found at compile time\n"); #endif } You may think the above is an awful lot of hassle, and it is. There are other ways outside of the C linking system to ensure that this kind of incompatibility never arises, such as using package versioning: Requires: libguestfs >= 1.0.80 =head1 CALLS WITH OPTIONAL ARGUMENTS A recent feature of the API is the introduction of calls which take optional arguments. In C these are declared 3 ways. The main way is as a call which takes variable arguments (ie. C<...>), as in this example: int guestfs_add_drive_opts (guestfs_h *g, const char *filename, ...); Call this with a list of optional arguments, terminated by C<-1>. So to call with no optional arguments specified: guestfs_add_drive_opts (g, filename, -1); With a single optional argument: guestfs_add_drive_opts (g, filename, GUESTFS_ADD_DRIVE_OPTS_FORMAT, "qcow2", -1); With two: guestfs_add_drive_opts (g, filename, GUESTFS_ADD_DRIVE_OPTS_FORMAT, "qcow2", GUESTFS_ADD_DRIVE_OPTS_READONLY, 1, -1); and so forth. Don't forget the terminating C<-1> otherwise Bad Things will happen! =head2 USING va_list FOR OPTIONAL ARGUMENTS The second variant has the same name with the suffix C<_va>, which works the same way but takes a C. See the C manual for details. For the example function, this is declared: int guestfs_add_drive_opts_va (guestfs_h *g, const char *filename, va_list args); =head2 CONSTRUCTING OPTIONAL ARGUMENTS The third variant is useful where you need to construct these calls. You pass in a structure where you fill in the optional fields. The structure has a bitmask as the first element which you must set to indicate which fields you have filled in. For our example function the structure and call are declared: struct guestfs_add_drive_opts_argv { uint64_t bitmask; int readonly; const char *format; /* ... */ }; int guestfs_add_drive_opts_argv (guestfs_h *g, const char *filename, const struct guestfs_add_drive_opts_argv *optargs); You could call it like this: struct guestfs_add_drive_opts_argv optargs = { .bitmask = GUESTFS_ADD_DRIVE_OPTS_READONLY_BITMASK | GUESTFS_ADD_DRIVE_OPTS_FORMAT_BITMASK, .readonly = 1, .format = "qcow2" }; guestfs_add_drive_opts_argv (g, filename, &optargs); Notes: =over 4 =item * The C<_BITMASK> suffix on each option name when specifying the bitmask. =item * You do not need to fill in all fields of the structure. =item * There must be a one-to-one correspondence between fields of the structure that are filled in, and bits set in the bitmask. =back =head2 OPTIONAL ARGUMENTS IN OTHER LANGUAGES In other languages, optional arguments are expressed in the way that is natural for that language. We refer you to the language-specific documentation for more details on that. For guestfish, see L. =head2 SETTING CALLBACKS TO HANDLE EVENTS B This section documents the generic event mechanism introduced in libguestfs 1.10, which you should use in new code if possible. The old functions C, C, C, C and C are no longer documented in this manual page. Because of the ABI guarantee, the old functions continue to work. Handles generate events when certain things happen, such as log messages being generated, progress messages during long-running operations, or the handle being closed. The API calls described below let you register a callback to be called when events happen. You can register multiple callbacks (for the same, different or overlapping sets of events), and individually remove callbacks. If callbacks are not removed, then they remain in force until the handle is closed. In the current implementation, events are only generated synchronously: that means that events (and hence callbacks) can only happen while you are in the middle of making another libguestfs call. The callback is called in the same thread. Events may contain a payload, usually nothing (void), an array of 64 bit unsigned integers, or a message buffer. Payloads are discussed later on. =head3 CLASSES OF EVENTS =over 4 =item GUESTFS_EVENT_CLOSE (payload type: void) The callback function will be called while the handle is being closed (synchronously from L). Note that libguestfs installs an L handler to try to clean up handles that are open when the program exits. This means that this callback might be called indirectly from L, which can cause unexpected problems in higher-level languages (eg. if your HLL interpreter has already been cleaned up by the time this is called, and if your callback then jumps into some HLL function). If no callback is registered: the handle is closed without any callback being invoked. =item GUESTFS_EVENT_SUBPROCESS_QUIT (payload type: void) The callback function will be called when the child process quits, either asynchronously or if killed by L. (This corresponds to a transition from any state to the CONFIG state). If no callback is registered: the event is ignored. =item GUESTFS_EVENT_LAUNCH_DONE (payload type: void) The callback function will be called when the child process becomes ready first time after it has been launched. (This corresponds to a transition from LAUNCHING to the READY state). If no callback is registered: the event is ignored. =item GUESTFS_EVENT_PROGRESS (payload type: array of 4 x uint64_t) Some long-running operations can generate progress messages. If this callback is registered, then it will be called each time a progress message is generated (usually two seconds after the operation started, and three times per second thereafter until it completes, although the frequency may change in future versions). The callback receives in the payload four unsigned 64 bit numbers which are (in order): C, C, C, C. The units of C are not defined, although for some operations C may relate in some way to the amount of data to be transferred (eg. in bytes or megabytes), and C may be the portion which has been transferred. The only defined and stable parts of the API are: =over 4 =item * The callback can display to the user some type of progress bar or indicator which shows the ratio of C:C. =item * 0 E= C E= C =item * If any progress notification is sent during a call, then a final progress notification is always sent when C = C (I the call fails with an error). This is to simplify caller code, so callers can easily set the progress indicator to "100%" at the end of the operation, without requiring special code to detect this case. =item * For some calls we are unable to estimate the progress of the call, but we can still generate progress messages to indicate activity. This is known as "pulse mode", and is directly supported by certain progress bar implementations (eg. GtkProgressBar). For these calls, zero or more progress messages are generated with C and C, followed by a final message with C. As noted above, if the call fails with an error then the final message may not be generated. =back The callback also receives the procedure number (C) and serial number (C) of the call. These are only useful for debugging protocol issues, and the callback can normally ignore them. The callback may want to print these numbers in error messages or debugging messages. If no callback is registered: progress messages are discarded. =item GUESTFS_EVENT_APPLIANCE (payload type: message buffer) The callback function is called whenever a log message is generated by qemu, the appliance kernel, guestfsd (daemon), or utility programs. If the verbose flag (L) is set before launch (L) then additional debug messages are generated. If no callback is registered: the messages are discarded unless the verbose flag is set in which case they are sent to stderr. You can override the printing of verbose messages to stderr by setting up a callback. =item GUESTFS_EVENT_LIBRARY (payload type: message buffer) The callback function is called whenever a log message is generated by the library part of libguestfs. If the verbose flag (L) is set then additional debug messages are generated. If no callback is registered: the messages are discarded unless the verbose flag is set in which case they are sent to stderr. You can override the printing of verbose messages to stderr by setting up a callback. =item GUESTFS_EVENT_TRACE (payload type: message buffer) The callback function is called whenever a trace message is generated. This only applies if the trace flag (L) is set. If no callback is registered: the messages are sent to stderr. You can override the printing of trace messages to stderr by setting up a callback. =back =head3 guestfs_set_event_callback int guestfs_set_event_callback (guestfs_h *g, guestfs_event_callback cb, uint64_t event_bitmask, int flags, void *opaque); This function registers a callback (C) for all event classes in the C. For example, to register for all log message events, you could call this function with the bitmask C. To register a single callback for all possible classes of events, use C. C should always be passed as 0. C is an opaque pointer which is passed to the callback. You can use it for any purpose. The return value is the event handle (an integer) which you can use to delete the callback (see below). If there is an error, this function returns C<-1>, and sets the error in the handle in the usual way (see L etc.) Callbacks remain in effect until they are deleted, or until the handle is closed. In the case where multiple callbacks are registered for a particular event class, all of the callbacks are called. The order in which multiple callbacks are called is not defined. =head3 guestfs_delete_event_callback void guestfs_delete_event_callback (guestfs_h *g, int event_handle); Delete a callback that was previously registered. C should be the integer that was returned by a previous call to C on the same handle. =head3 guestfs_event_callback typedef void (*guestfs_event_callback) ( guestfs_h *g, void *opaque, uint64_t event, int event_handle, int flags, const char *buf, size_t buf_len, const uint64_t *array, size_t array_len); This is the type of the event callback function that you have to provide. The basic parameters are: the handle (C), the opaque user pointer (C), the event class (eg. C), the event handle, and C which in the current API you should ignore. The remaining parameters contain the event payload (if any). Each event may contain a payload, which usually relates to the event class, but for future proofing your code should be written to handle any payload for any event class. C and C contain a message buffer (if C, then there is no message buffer). Note that this message buffer can contain arbitrary 8 bit data, including NUL bytes. C and C is an array of 64 bit unsigned integers. At the moment this is only used for progress messages. =head3 EXAMPLE: CAPTURING LOG MESSAGES One motivation for the generic event API was to allow GUI programs to capture debug and other messages. In libguestfs E 1.8 these were sent unconditionally to C. Events associated with log messages are: C, C and C. (Note that error messages are not events; you must capture error messages separately). Programs have to set up a callback to capture the classes of events of interest: int eh = guestfs_set_event_callback (g, message_callback, GUESTFS_EVENT_LIBRARY|GUESTFS_EVENT_APPLIANCE| GUESTFS_EVENT_TRACE, 0, NULL) == -1) if (eh == -1) { // handle error in the usual way } The callback can then direct messages to the appropriate place. In this example, messages are directed to syslog: static void message_callback ( guestfs_h *g, void *opaque, uint64_t event, int event_handle, int flags, const char *buf, size_t buf_len, const uint64_t *array, size_t array_len) { const int priority = LOG_USER|LOG_INFO; if (buf_len > 0) syslog (priority, "event 0x%lx: %s", event, buf); } =head1 CANCELLING LONG TRANSFERS Some operations can be cancelled by the caller while they are in progress. Currently only operations that involve uploading or downloading data can be cancelled (technically: operations that have C or C parameters in the generator). =head2 guestfs_user_cancel void guestfs_user_cancel (guestfs_h *g); C cancels the current upload or download operation. Unlike most other libguestfs calls, this function is signal safe and thread safe. You can call it from a signal handler or from another thread, without needing to do any locking. The transfer that was in progress (if there is one) will stop shortly afterwards, and will return an error. The errno (see L) is set to C, so you can test for this to find out if the operation was cancelled or failed because of another error. No cleanup is performed: for example, if a file was being uploaded then after cancellation there may be a partially uploaded file. It is the caller's responsibility to clean up if necessary. There are two common places that you might call C. In an interactive text-based program, you might call it from a C signal handler so that pressing C<^C> cancels the current operation. (You also need to call L so that child processes don't receive the C<^C> signal). In a graphical program, when the main thread is displaying a progress bar with a cancel button, wire up the cancel button to call this function. =head1 PRIVATE DATA AREA You can attach named pieces of private data to the libguestfs handle, fetch them by name, and walk over them, for the lifetime of the handle. This is called the private data area and is only available from the C API. To attach a named piece of data, use the following call: void guestfs_set_private (guestfs_h *g, const char *key, void *data); C is the name to associate with this data, and C is an arbitrary pointer (which can be C). Any previous item with the same key is overwritten. You can use any C you want, but your key should I start with an underscore character. Keys beginning with an underscore character are reserved for internal libguestfs purposes (eg. for implementing language bindings). It is recommended that you prefix the key with some unique string to avoid collisions with other users. To retrieve the pointer, use: void *guestfs_get_private (guestfs_h *g, const char *key); This function returns C if either no data is found associated with C, or if the user previously set the C's C pointer to C. Libguestfs does not try to look at or interpret the C pointer in any way. As far as libguestfs is concerned, it need not be a valid pointer at all. In particular, libguestfs does I try to free the data when the handle is closed. If the data must be freed, then the caller must either free it before calling L or must set up a close callback to do it (see L). To walk over all entries, use these two functions: void *guestfs_first_private (guestfs_h *g, const char **key_rtn); void *guestfs_next_private (guestfs_h *g, const char **key_rtn); C returns the first key, pointer pair ("first" does not have any particular meaning -- keys are not returned in any defined order). A pointer to the key is returned in C<*key_rtn> and the corresponding data pointer is returned from the function. C is returned if there are no keys stored in the handle. C returns the next key, pointer pair. The return value of this function is also C is there are no further entries to return. Notes about walking over entries: =over 4 =item * You must not call C while walking over the entries. =item * The handle maintains an internal iterator which is reset when you call C. This internal iterator is invalidated when you call C. =item * If you have set the data pointer associated with a key to C, ie: guestfs_set_private (g, key, NULL); then that C is not returned when walking. =item * C<*key_rtn> is only valid until the next call to C, C or C. =back The following example code shows how to print all keys and data pointers that are associated with the handle C: const char *key; void *data = guestfs_first_private (g, &key); while (data != NULL) { printf ("key = %s, data = %p\n", key, data); data = guestfs_next_private (g, &key); } More commonly you are only interested in keys that begin with an application-specific prefix C. Modify the loop like so: const char *key; void *data = guestfs_first_private (g, &key); while (data != NULL) { if (strncmp (key, "foo_", strlen ("foo_")) == 0) printf ("key = %s, data = %p\n", key, data); data = guestfs_next_private (g, &key); } If you need to modify keys while walking, then you have to jump back to the beginning of the loop. For example, to delete all keys prefixed with C: const char *key; void *data; again: data = guestfs_first_private (g, &key); while (data != NULL) { if (strncmp (key, "foo_", strlen ("foo_")) == 0) { guestfs_set_private (g, key, NULL); /* note that 'key' pointer is now invalid, and so is the internal iterator */ goto again; } data = guestfs_next_private (g, &key); } Note that the above loop is guaranteed to terminate because the keys are being deleted, but other manipulations of keys within the loop might not terminate unless you also maintain an indication of which keys have been visited. =begin html =end html =head1 ARCHITECTURE Internally, libguestfs is implemented by running an appliance (a special type of small virtual machine) using L. Qemu runs as a child process of the main program. ___________________ / \ | main program | | | | | child process / appliance | | __________________________ | | / qemu \ +-------------------+ RPC | +-----------------+ | | libguestfs <--------------------> guestfsd | | | | | +-----------------+ | \___________________/ | | Linux kernel | | | +--^--------------+ | \_________|________________/ | _______v______ / \ | Device or | | disk image | \______________/ The library, linked to the main program, creates the child process and hence the appliance in the L function. Inside the appliance is a Linux kernel and a complete stack of userspace tools (such as LVM and ext2 programs) and a small controlling daemon called L. The library talks to L using remote procedure calls (RPC). There is a mostly one-to-one correspondence between libguestfs API calls and RPC calls to the daemon. Lastly the disk image(s) are attached to the qemu process which translates device access by the appliance's Linux kernel into accesses to the image. A common misunderstanding is that the appliance "is" the virtual machine. Although the disk image you are attached to might also be used by some virtual machine, libguestfs doesn't know or care about this. (But you will care if both libguestfs's qemu process and your virtual machine are trying to update the disk image at the same time, since these usually results in massive disk corruption). =head1 STATE MACHINE libguestfs uses a state machine to model the child process: | guestfs_create | | ____V_____ / \ | CONFIG | \__________/ ^ ^ ^ \ / | \ \ guestfs_launch / | _\__V______ / | / \ / | | LAUNCHING | / | \___________/ / | / / | guestfs_launch / | / ______ / __|____V / \ ------> / \ | BUSY | | READY | \______/ <------ \________/ The normal transitions are (1) CONFIG (when the handle is created, but there is no child process), (2) LAUNCHING (when the child process is booting up), (3) alternating between READY and BUSY as commands are issued to, and carried out by, the child process. The guest may be killed by L, or may die asynchronously at any time (eg. due to some internal error), and that causes the state to transition back to CONFIG. Configuration commands for qemu such as L can only be issued when in the CONFIG state. The API offers one call that goes from CONFIG through LAUNCHING to READY. L blocks until the child process is READY to accept commands (or until some failure or timeout). L internally moves the state from CONFIG to LAUNCHING while it is running. API actions such as L can only be issued when in the READY state. These API calls block waiting for the command to be carried out (ie. the state to transition to BUSY and then back to READY). There are no non-blocking versions, and no way to issue more than one command per handle at the same time. Finally, the child process sends asynchronous messages back to the main program, such as kernel log messages. You can register a callback to receive these messages. =head1 INTERNALS =head2 COMMUNICATION PROTOCOL Don't rely on using this protocol directly. This section documents how it currently works, but it may change at any time. The protocol used to talk between the library and the daemon running inside the qemu virtual machine is a simple RPC mechanism built on top of XDR (RFC 1014, RFC 1832, RFC 4506). The detailed format of structures is in C (note: this file is automatically generated). There are two broad cases, ordinary functions that don't have any C and C parameters, which are handled with very simple request/reply messages. Then there are functions that have any C or C parameters, which use the same request and reply messages, but they may also be followed by files sent using a chunked encoding. =head3 ORDINARY FUNCTIONS (NO FILEIN/FILEOUT PARAMS) For ordinary functions, the request message is: total length (header + arguments, but not including the length word itself) struct guestfs_message_header (encoded as XDR) struct guestfs__args (encoded as XDR) The total length field allows the daemon to allocate a fixed size buffer into which it slurps the rest of the message. As a result, the total length is limited to C bytes (currently 4MB), which means the effective size of any request is limited to somewhere under this size. Note also that many functions don't take any arguments, in which case the C_args> is completely omitted. The header contains the procedure number (C) which is how the receiver knows what type of args structure to expect, or none at all. For functions that take optional arguments, the optional arguments are encoded in the C_args> structure in the same way as ordinary arguments. A bitmask in the header indicates which optional arguments are meaningful. The bitmask is also checked to see if it contains bits set which the daemon does not know about (eg. if more optional arguments were added in a later version of the library), and this causes the call to be rejected. The reply message for ordinary functions is: total length (header + ret, but not including the length word itself) struct guestfs_message_header (encoded as XDR) struct guestfs__ret (encoded as XDR) As above the C_ret> structure may be completely omitted for functions that return no formal return values. As above the total length of the reply is limited to C. In the case of an error, a flag is set in the header, and the reply message is slightly changed: total length (header + error, but not including the length word itself) struct guestfs_message_header (encoded as XDR) struct guestfs_message_error (encoded as XDR) The C structure contains the error message as a string. =head3 FUNCTIONS THAT HAVE FILEIN PARAMETERS A C parameter indicates that we transfer a file I the guest. The normal request message is sent (see above). However this is followed by a sequence of file chunks. total length (header + arguments, but not including the length word itself, and not including the chunks) struct guestfs_message_header (encoded as XDR) struct guestfs__args (encoded as XDR) sequence of chunks for FileIn param #0 sequence of chunks for FileIn param #1 etc. The "sequence of chunks" is: length of chunk (not including length word itself) struct guestfs_chunk (encoded as XDR) length of chunk struct guestfs_chunk (encoded as XDR) ... length of chunk struct guestfs_chunk (with data.data_len == 0) The final chunk has the C field set to zero. Additionally a flag is set in the final chunk to indicate either successful completion or early cancellation. At time of writing there are no functions that have more than one FileIn parameter. However this is (theoretically) supported, by sending the sequence of chunks for each FileIn parameter one after another (from left to right). Both the library (sender) I the daemon (receiver) may cancel the transfer. The library does this by sending a chunk with a special flag set to indicate cancellation. When the daemon sees this, it cancels the whole RPC, does I send any reply, and goes back to reading the next request. The daemon may also cancel. It does this by writing a special word C to the socket. The library listens for this during the transfer, and if it gets it, it will cancel the transfer (it sends a cancel chunk). The special word is chosen so that even if cancellation happens right at the end of the transfer (after the library has finished writing and has started listening for the reply), the "spurious" cancel flag will not be confused with the reply message. This protocol allows the transfer of arbitrary sized files (no 32 bit limit), and also files where the size is not known in advance (eg. from pipes or sockets). However the chunks are rather small (C), so that neither the library nor the daemon need to keep much in memory. =head3 FUNCTIONS THAT HAVE FILEOUT PARAMETERS The protocol for FileOut parameters is exactly the same as for FileIn parameters, but with the roles of daemon and library reversed. total length (header + ret, but not including the length word itself, and not including the chunks) struct guestfs_message_header (encoded as XDR) struct guestfs__ret (encoded as XDR) sequence of chunks for FileOut param #0 sequence of chunks for FileOut param #1 etc. =head3 INITIAL MESSAGE When the daemon launches it sends an initial word (C) which indicates that the guest and daemon is alive. This is what L waits for. =head3 PROGRESS NOTIFICATION MESSAGES The daemon may send progress notification messages at any time. These are distinguished by the normal length word being replaced by C, followed by a fixed size progress message. The library turns them into progress callbacks (see L) if there is a callback registered, or discards them if not. The daemon self-limits the frequency of progress messages it sends (see C). Not all calls generate progress messages. =head1 LIBGUESTFS VERSION NUMBERS Since April 2010, libguestfs has started to make separate development and stable releases, along with corresponding branches in our git repository. These separate releases can be identified by version number: even numbers for stable: 1.2.x, 1.4.x, ... .-------- odd numbers for development: 1.3.x, 1.5.x, ... | v 1 . 3 . 5 ^ ^ | | | `-------- sub-version | `------ always '1' because we don't change the ABI Thus "1.3.5" is the 5th update to the development branch "1.3". As time passes we cherry pick fixes from the development branch and backport those into the stable branch, the effect being that the stable branch should get more stable and less buggy over time. So the stable releases are ideal for people who don't need new features but would just like the software to work. Our criteria for backporting changes are: =over 4 =item * Documentation changes which don't affect any code are backported unless the documentation refers to a future feature which is not in stable. =item * Bug fixes which are not controversial, fix obvious problems, and have been well tested are backported. =item * Simple rearrangements of code which shouldn't affect how it works get backported. This is so that the code in the two branches doesn't get too far out of step, allowing us to backport future fixes more easily. =item * We I backport new features, new APIs, new tools etc, except in one exceptional case: the new feature is required in order to implement an important bug fix. =back A new stable branch starts when we think the new features in development are substantial and compelling enough over the current stable branch to warrant it. When that happens we create new stable and development versions 1.N.0 and 1.(N+1).0 [N is even]. The new dot-oh release won't necessarily be so stable at this point, but by backporting fixes from development, that branch will stabilize over time. =head1 EXTENDING LIBGUESTFS =head2 ADDING A NEW API ACTION Large amounts of boilerplate code in libguestfs (RPC, bindings, documentation) are generated, and this makes it easy to extend the libguestfs API. To add a new API action there are two changes: =over 4 =item 1. You need to add a description of the call (name, parameters, return type, tests, documentation) to C. There are two sorts of API action, depending on whether the call goes through to the daemon in the appliance, or is serviced entirely by the library (see L above). L is an example of the former, since the sync is done in the appliance. L is an example of the latter, since a trace flag is maintained in the handle and all tracing is done on the library side. Most new actions are of the first type, and get added to the C list. Each function has a unique procedure number used in the RPC protocol which is assigned to that action when we publish libguestfs and cannot be reused. Take the latest procedure number and increment it. For library-only actions of the second type, add to the C list. Since these functions are serviced by the library and do not travel over the RPC mechanism to the daemon, these functions do not need a procedure number, and so the procedure number is set to C<-1>. =item 2. Implement the action (in C): For daemon actions, implement the function CnameE> in the C directory. For library actions, implement the function CnameE> (note: double underscore) in the C directory. In either case, use another function as an example of what to do. =back After making these changes, use C to compile. Note that you don't need to implement the RPC, language bindings, manual pages or anything else. It's all automatically generated from the OCaml description. =head2 ADDING TESTS FOR AN API ACTION You can supply zero or as many tests as you want per API call. The tests can either be added as part of the API description (C), or in some rarer cases you may want to drop a script into C. Note that adding a script to C is slower, so if possible use the first method. The following describes the test environment used when you add an API test in C. The test environment has 4 block devices: =over 4 =item C 500MB General block device for testing. =item C 50MB C is an ext2 filesystem used for testing filesystem write operations. =item C 10MB Used in a few tests where two block devices are needed. =item C ISO with fixed content (see C). =back To be able to run the tests in a reasonable amount of time, the libguestfs appliance and block devices are reused between tests. So don't try testing L :-x Each test starts with an initial scenario, selected using one of the C expressions, described in C. These initialize the disks mentioned above in a particular way as documented in C. You should not assume anything about the previous contents of other disks that are not initialized. You can add a prerequisite clause to any individual test. This is a run-time check, which, if it fails, causes the test to be skipped. Useful if testing a command which might not work on all variations of libguestfs builds. A test that has prerequisite of C means to run unconditionally. In addition, packagers can skip individual tests by setting environment variables before running C. SKIP_TEST__=1 eg: C skips test #3 of L. or: SKIP_TEST_=1 eg: C skips all L tests. Packagers can run only certain tests by setting for example: TEST_ONLY="vfs_type zerofree" See C for more details of how these environment variables work. =head2 DEBUGGING NEW API ACTIONS Test new actions work before submitting them. You can use guestfish to try out new commands. Debugging the daemon is a problem because it runs inside a minimal environment. However you can fprintf messages in the daemon to stderr, and they will show up if you use C. =head2 FORMATTING CODE AND OTHER CONVENTIONS Our C source code generally adheres to some basic code-formatting conventions. The existing code base is not totally consistent on this front, but we do prefer that contributed code be formatted similarly. In short, use spaces-not-TABs for indentation, use 2 spaces for each indentation level, and other than that, follow the K&R style. If you use Emacs, add the following to one of one of your start-up files (e.g., ~/.emacs), to help ensure that you get indentation right: ;;; In libguestfs, indent with spaces everywhere (not TABs). ;;; Exceptions: Makefile and ChangeLog modes. (add-hook 'find-file-hook '(lambda () (if (and buffer-file-name (string-match "/libguestfs\\>" (buffer-file-name)) (not (string-equal mode-name "Change Log")) (not (string-equal mode-name "Makefile"))) (setq indent-tabs-mode nil)))) ;;; When editing C sources in libguestfs, use this style. (defun libguestfs-c-mode () "C mode with adjusted defaults for use with libguestfs." (interactive) (c-set-style "K&R") (setq c-indent-level 2) (setq c-basic-offset 2)) (add-hook 'c-mode-hook '(lambda () (if (string-match "/libguestfs\\>" (buffer-file-name)) (libguestfs-c-mode)))) Enable warnings when compiling (and fix any problems this finds): ./configure --enable-gcc-warnings Useful targets are: make syntax-check # checks the syntax of the C code make check # runs the test suite =head2 DAEMON CUSTOM PRINTF FORMATTERS In the daemon code we have created custom printf formatters C<%Q> and C<%R>, which are used to do shell quoting. =over 4 =item %Q Simple shell quoted string. Any spaces or other shell characters are escaped for you. =item %R Same as C<%Q> except the string is treated as a path which is prefixed by the sysroot. =back For example: asprintf (&cmd, "cat %R", path); would produce C I Do I use these when you are passing parameters to the C functions. These parameters do NOT need to be quoted because they are not passed via the shell (instead, straight to exec). You probably want to use the C function however. =head2 SUBMITTING YOUR NEW API ACTIONS Submit patches to the mailing list: L and CC to L. =head2 INTERNATIONALIZATION (I18N) SUPPORT We support i18n (gettext anyhow) in the library. However many messages come from the daemon, and we don't translate those at the moment. One reason is that the appliance generally has all locale files removed from it, because they take up a lot of space. So we'd have to readd some of those, as well as copying our PO files into the appliance. Debugging messages are never translated, since they are intended for the programmers. =head2 SOURCE CODE SUBDIRECTORIES =over 4 =item C The libguestfs appliance, build scripts and so on. =item C Automated tests of the C API. =item C The L, L and L commands and documentation. =item C Safety and liveness tests of components that libguestfs depends upon (not of libguestfs itself). Mainly this is for qemu and the kernel. =item C Outside contributions, experimental parts. =item C The daemon that runs inside the libguestfs appliance and carries out actions. =item C L command and documentation. =item C L command and documentation. =item C C API example code. =item C L, the command-line shell, and various shell scripts built on top such as L, L, L, L. =item C L, FUSE (userspace filesystem) built on top of libguestfs. =item C The crucially important generator, used to automatically generate large amounts of boilerplate C code for things like RPC and bindings. =item C Files used by the test suite. Some "phony" guest images which we test against. =item C L, the virtual machine image inspector. =item C Logo used on the website. The fish is called Arthur by the way. =item C M4 macros used by autoconf. =item C Translations of simple gettext strings. =item C The build infrastructure and PO files for translations of manpages and POD files. Eventually this will be combined with the C directory, but that is rather complicated. =item C Regression tests. =item C L command and documentation. =item C Source code to the C library. =item C Command line tools written in Perl (L and many others). =item C Test tool for end users to test if their qemu/kernel combination will work with libguestfs. =item C =item C =item C =item C =item C =item C =item C =item C Language bindings. =back =head2 MAKING A STABLE RELEASE When we make a stable release, there are several steps documented here. See L for general information about the stable branch policy. =over 4 =item * Check C works on at least Fedora, Debian and Ubuntu. =item * Finalize RELEASE-NOTES. =item * Update ROADMAP. =item * Run C. =item * Push and pull from Transifex. Run: tx push -s to push the latest POT files to Transifex. Then run: ./tx-pull.sh which is a wrapper to pull the latest translated C<*.po> files. =item * Create new stable and development directories under L. =item * Create the branch in git: git tag -a 1.XX.0 -m "Version 1.XX.0 (stable)" git tag -a 1.YY.0 -m "Version 1.YY.0 (development)" git branch stable-1.XX git push origin tag 1.XX.0 1.YY.0 stable-1.XX =back =head1 LIMITS =head2 PROTOCOL LIMITS Internally libguestfs uses a message-based protocol to pass API calls and their responses to and from a small "appliance" (see L for plenty more detail about this). The maximum message size used by the protocol is slightly less than 4 MB. For some API calls you may need to be aware of this limit. The API calls which may be affected are individually documented, with a link back to this section of the documentation. A simple call such as L returns its result (the file data) in a simple string. Because this string is at some point internally encoded as a message, the maximum size that it can return is slightly under 4 MB. If the requested file is larger than this then you will get an error. In order to transfer large files into and out of the guest filesystem, you need to use particular calls that support this. The sections L and L document how to do this. You might also consider mounting the disk image using our FUSE filesystem support (L). =head2 MAXIMUM NUMBER OF DISKS When using virtio disks (the default) the current limit is B<25> disks. Virtio itself consumes 1 virtual PCI slot per disk, and PCI is limited to 31 slots. However febootstrap only understands disks with names C through C (26 letters) and it reserves one disk for its own purposes. We are working to substantially raise this limit in future versions but it requires complex changes to qemu. In future versions of libguestfs it should also be possible to "hot plug" disks (add and remove disks after calling L). This also requires changes to qemu. =head2 MAXIMUM NUMBER OF PARTITIONS PER DISK Virtio limits the maximum number of partitions per disk to B<15>. This is because it reserves 4 bits for the minor device number (thus C, and C through C). If you attach a disk with more than 15 partitions, the extra partitions are ignored by libguestfs. =head2 MAXIMUM SIZE OF A DISK Probably the limit is between 2**63-1 and 2**64-1 bytes. We have tested block devices up to 1 exabyte (2**60 or 1,152,921,504,606,846,976 bytes) using sparse files backed by an XFS host filesystem. Although libguestfs probably does not impose any limit, the underlying host storage will. If you store disk images on a host ext4 filesystem, then the maximum size will be limited by the maximum ext4 file size (currently 16 TB). If you store disk images as host logical volumes then you are limited by the maximum size of an LV. For the hugest disk image files, we recommend using XFS on the host for storage. =head2 MAXIMUM SIZE OF A PARTITION The MBR (ie. classic MS-DOS) partitioning scheme uses 32 bit sector numbers. Assuming a 512 byte sector size, this means that MBR cannot address a partition located beyond 2 TB on the disk. It is recommended that you use GPT partitions on disks which are larger than this size. GPT uses 64 bit sector numbers and so can address partitions which are theoretically larger than the largest disk we could support. =head2 MAXIMUM SIZE OF A FILESYSTEM, FILES, DIRECTORIES This depends on the filesystem type. libguestfs itself does not impose any known limit. Consult Wikipedia or the filesystem documentation to find out what these limits are. =head2 MAXIMUM UPLOAD AND DOWNLOAD The API functions L, L, L, L and the like allow unlimited sized uploads and downloads. =head2 INSPECTION LIMITS The inspection code has several arbitrary limits on things like the size of Windows Registry hive it will read, and the length of product name. These are intended to stop a malicious guest from consuming arbitrary amounts of memory and disk space on the host, and should not be reached in practice. See the source code for more information. =head1 ENVIRONMENT VARIABLES =over 4 =item LIBGUESTFS_APPEND Pass additional options to the guest kernel. =item LIBGUESTFS_DEBUG Set C to enable verbose messages. This has the same effect as calling C. =item LIBGUESTFS_MEMSIZE Set the memory allocated to the qemu process, in megabytes. For example: LIBGUESTFS_MEMSIZE=700 =item LIBGUESTFS_PATH Set the path that libguestfs uses to search for a supermin appliance. See the discussion of paths in section L above. =item LIBGUESTFS_QEMU Set the default qemu binary that libguestfs uses. If not set, then the qemu which was found at compile time by the configure script is used. See also L above. =item LIBGUESTFS_TRACE Set C to enable command traces. This has the same effect as calling C. =item TMPDIR Location of temporary directory, defaults to C except for the cached supermin appliance which defaults to C. If libguestfs was compiled to use the supermin appliance then the real appliance is cached in this directory, shared between all handles belonging to the same EUID. You can use C<$TMPDIR> to configure another directory to use in case C is not large enough. =back =head1 SEE ALSO L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L, L. Tools with a similar purpose: L, L, L, L, L. =head1 BUGS To get a list of bugs against libguestfs use this link: L To report a new bug against libguestfs use this link: L When reporting a bug, please check: =over 4 =item * That the bug hasn't been reported already. =item * That you are testing a recent version. =item * Describe the bug accurately, and give a way to reproduce it. =item * Run libguestfs-test-tool and paste the B output into the bug report. =back =head1 AUTHORS Richard W.M. Jones (C) =head1 COPYRIGHT Copyright (C) 2009-2011 Red Hat Inc. L This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this library; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA