1Documentation for userland software suspend interface
2    (C) 2006 Rafael J. Wysocki <>
4First, the warnings at the beginning of swsusp.txt still apply.
6Second, you should read the FAQ in swsusp.txt _now_ if you have not
7done it already.
9Now, to use the userland interface for software suspend you need special
10utilities that will read/write the system memory snapshot from/to the
11kernel. Such utilities are available, for example, from
12<>. You may want to have a look at them if you
13are going to develop your own suspend/resume utilities.
15The interface consists of a character device providing the open(),
16release(), read(), and write() operations as well as several ioctl()
17commands defined in include/linux/suspend_ioctls.h . The major and minor
18numbers of the device are, respectively, 10 and 231, and they can
19be read from /sys/class/misc/snapshot/dev.
21The device can be open either for reading or for writing. If open for
22reading, it is considered to be in the suspend mode. Otherwise it is
23assumed to be in the resume mode. The device cannot be open for simultaneous
24reading and writing. It is also impossible to have the device open more than
25once at a time.
27Even opening the device has side effects. Data structures are
31The ioctl() commands recognized by the device are:
33SNAPSHOT_FREEZE - freeze user space processes (the current process is
34    not frozen); this is required for SNAPSHOT_CREATE_IMAGE
35    and SNAPSHOT_ATOMIC_RESTORE to succeed
37SNAPSHOT_UNFREEZE - thaw user space processes frozen by SNAPSHOT_FREEZE
39SNAPSHOT_CREATE_IMAGE - create a snapshot of the system memory; the
40    last argument of ioctl() should be a pointer to an int variable,
41    the value of which will indicate whether the call returned after
42    creating the snapshot (1) or after restoring the system memory state
43    from it (0) (after resume the system finds itself finishing the
44    SNAPSHOT_CREATE_IMAGE ioctl() again); after the snapshot
45    has been created the read() operation can be used to transfer
46    it out of the kernel
48SNAPSHOT_ATOMIC_RESTORE - restore the system memory state from the
49    uploaded snapshot image; before calling it you should transfer
50    the system memory snapshot back to the kernel using the write()
51    operation; this call will not succeed if the snapshot
52    image is not available to the kernel
54SNAPSHOT_FREE - free memory allocated for the snapshot image
56SNAPSHOT_PREF_IMAGE_SIZE - set the preferred maximum size of the image
57    (the kernel will do its best to ensure the image size will not exceed
58    this number, but if it turns out to be impossible, the kernel will
59    create the smallest image possible)
61SNAPSHOT_GET_IMAGE_SIZE - return the actual size of the hibernation image
63SNAPSHOT_AVAIL_SWAP_SIZE - return the amount of available swap in bytes (the
64    last argument should be a pointer to an unsigned int variable that will
65    contain the result if the call is successful).
67SNAPSHOT_ALLOC_SWAP_PAGE - allocate a swap page from the resume partition
68    (the last argument should be a pointer to a loff_t variable that
69    will contain the swap page offset if the call is successful)
71SNAPSHOT_FREE_SWAP_PAGES - free all swap pages allocated by
74SNAPSHOT_SET_SWAP_AREA - set the resume partition and the offset (in <PAGE_SIZE>
75    units) from the beginning of the partition at which the swap header is
76    located (the last ioctl() argument should point to a struct
77    resume_swap_area, as defined in kernel/power/suspend_ioctls.h,
78    containing the resume device specification and the offset); for swap
79    partitions the offset is always 0, but it is different from zero for
80    swap files (see Documentation/swsusp-and-swap-files.txt for details).
82SNAPSHOT_PLATFORM_SUPPORT - enable/disable the hibernation platform support,
83    depending on the argument value (enable, if the argument is nonzero)
85SNAPSHOT_POWER_OFF - make the kernel transition the system to the hibernation
86    state (eg. ACPI S4) using the platform (eg. ACPI) driver
88SNAPSHOT_S2RAM - suspend to RAM; using this call causes the kernel to
89    immediately enter the suspend-to-RAM state, so this call must always
90    be preceded by the SNAPSHOT_FREEZE call and it is also necessary
91    to use the SNAPSHOT_UNFREEZE call after the system wakes up. This call
92    is needed to implement the suspend-to-both mechanism in which the
93    suspend image is first created, as though the system had been suspended
94    to disk, and then the system is suspended to RAM (this makes it possible
95    to resume the system from RAM if there's enough battery power or restore
96    its state on the basis of the saved suspend image otherwise)
98The device's read() operation can be used to transfer the snapshot image from
99the kernel. It has the following limitations:
100- you cannot read() more than one virtual memory page at a time
101- read()s across page boundaries are impossible (ie. if ypu read() 1/2 of
102    a page in the previous call, you will only be able to read()
103    _at_ _most_ 1/2 of the page in the next call)
105The device's write() operation is used for uploading the system memory snapshot
106into the kernel. It has the same limitations as the read() operation.
108The release() operation frees all memory allocated for the snapshot image
109and all swap pages allocated with SNAPSHOT_ALLOC_SWAP_PAGE (if any).
110Thus it is not necessary to use either SNAPSHOT_FREE or
111SNAPSHOT_FREE_SWAP_PAGES before closing the device (in fact it will also
112unfreeze user space processes frozen by SNAPSHOT_UNFREEZE if they are
113still frozen when the device is being closed).
115Currently it is assumed that the userland utilities reading/writing the
116snapshot image from/to the kernel will use a swap partition, called the resume
117partition, or a swap file as storage space (if a swap file is used, the resume
118partition is the partition that holds this file). However, this is not really
119required, as they can use, for example, a special (blank) suspend partition or
120a file on a partition that is unmounted before SNAPSHOT_CREATE_IMAGE and
121mounted afterwards.
123These utilities MUST NOT make any assumptions regarding the ordering of
124data within the snapshot image. The contents of the image are entirely owned
125by the kernel and its structure may be changed in future kernel releases.
127The snapshot image MUST be written to the kernel unaltered (ie. all of the image
128data, metadata and header MUST be written in _exactly_ the same amount, form
129and order in which they have been read). Otherwise, the behavior of the
130resumed system may be totally unpredictable.
132While executing SNAPSHOT_ATOMIC_RESTORE the kernel checks if the
133structure of the snapshot image is consistent with the information stored
134in the image header. If any inconsistencies are detected,
135SNAPSHOT_ATOMIC_RESTORE will not succeed. Still, this is not a fool-proof
136mechanism and the userland utilities using the interface SHOULD use additional
137means, such as checksums, to ensure the integrity of the snapshot image.
139The suspending and resuming utilities MUST lock themselves in memory,
140preferably using mlockall(), before calling SNAPSHOT_FREEZE.
142The suspending utility MUST check the value stored by SNAPSHOT_CREATE_IMAGE
143in the memory location pointed to by the last argument of ioctl() and proceed
144in accordance with it:
1451. If the value is 1 (ie. the system memory snapshot has just been
146    created and the system is ready for saving it):
147    (a) The suspending utility MUST NOT close the snapshot device
148        _unless_ the whole suspend procedure is to be cancelled, in
149        which case, if the snapshot image has already been saved, the
150        suspending utility SHOULD destroy it, preferably by zapping
151        its header. If the suspend is not to be cancelled, the
152        system MUST be powered off or rebooted after the snapshot
153        image has been saved.
154    (b) The suspending utility SHOULD NOT attempt to perform any
155        file system operations (including reads) on the file systems
156        that were mounted before SNAPSHOT_CREATE_IMAGE has been
157        called. However, it MAY mount a file system that was not
158        mounted at that time and perform some operations on it (eg.
159        use it for saving the image).
1602. If the value is 0 (ie. the system state has just been restored from
161    the snapshot image), the suspending utility MUST close the snapshot
162    device. Afterwards it will be treated as a regular userland process,
163    so it need not exit.
165The resuming utility SHOULD NOT attempt to mount any file systems that could
166be mounted before suspend and SHOULD NOT attempt to perform any operations
167involving such file systems.
169For details, please refer to the source code.

Archive Download this file