1 Documentation for userland software suspend interface
2 (C) 2006 Rafael J. Wysocki <rjw@sisk.pl>
4 First, the warnings at the beginning of swsusp.txt still apply.
6 Second, you should read the FAQ in swsusp.txt _now_ if you have not
9 Now, to use the userland interface for software suspend you need special
10 utilities that will read/write the system memory snapshot from/to the
11 kernel. Such utilities are available, for example, from
12 <http://suspend.sourceforge.net>. You may want to have a look at them if you
13 are going to develop your own suspend/resume utilities.
15 The interface consists of a character device providing the open(),
16 release(), read(), and write() operations as well as several ioctl()
17 commands defined in kernel/power/power.h. The major and minor
18 numbers of the device are, respectively, 10 and 231, and they can
19 be read from /sys/class/misc/snapshot/dev.
21 The device can be open either for reading or for writing. If open for
22 reading, it is considered to be in the suspend mode. Otherwise it is
23 assumed to be in the resume mode. The device cannot be open for simultaneous
24 reading and writing. It is also impossible to have the device open more than
27 The ioctl() commands recognized by the device are:
29 SNAPSHOT_FREEZE - freeze user space processes (the current process is
30 not frozen); this is required for SNAPSHOT_ATOMIC_SNAPSHOT
31 and SNAPSHOT_ATOMIC_RESTORE to succeed
33 SNAPSHOT_UNFREEZE - thaw user space processes frozen by SNAPSHOT_FREEZE
35 SNAPSHOT_ATOMIC_SNAPSHOT - create a snapshot of the system memory; the
36 last argument of ioctl() should be a pointer to an int variable,
37 the value of which will indicate whether the call returned after
38 creating the snapshot (1) or after restoring the system memory state
39 from it (0) (after resume the system finds itself finishing the
40 SNAPSHOT_ATOMIC_SNAPSHOT ioctl() again); after the snapshot
41 has been created the read() operation can be used to transfer
44 SNAPSHOT_ATOMIC_RESTORE - restore the system memory state from the
45 uploaded snapshot image; before calling it you should transfer
46 the system memory snapshot back to the kernel using the write()
47 operation; this call will not succeed if the snapshot
48 image is not available to the kernel
50 SNAPSHOT_FREE - free memory allocated for the snapshot image
52 SNAPSHOT_SET_IMAGE_SIZE - set the preferred maximum size of the image
53 (the kernel will do its best to ensure the image size will not exceed
54 this number, but if it turns out to be impossible, the kernel will
55 create the smallest image possible)
57 SNAPSHOT_AVAIL_SWAP - return the amount of available swap in bytes (the last
58 argument should be a pointer to an unsigned int variable that will
59 contain the result if the call is successful).
61 SNAPSHOT_GET_SWAP_PAGE - allocate a swap page from the resume partition
62 (the last argument should be a pointer to a loff_t variable that
63 will contain the swap page offset if the call is successful)
65 SNAPSHOT_FREE_SWAP_PAGES - free all swap pages allocated with
66 SNAPSHOT_GET_SWAP_PAGE
68 SNAPSHOT_SET_SWAP_FILE - set the resume partition (the last ioctl() argument
69 should specify the device's major and minor numbers in the old
70 two-byte format, as returned by the stat() function in the .st_rdev
71 member of the stat structure)
73 SNAPSHOT_SET_SWAP_AREA - set the resume partition and the offset (in <PAGE_SIZE>
74 units) from the beginning of the partition at which the swap header is
75 located (the last ioctl() argument should point to a struct
76 resume_swap_area, as defined in kernel/power/power.h, containing the
77 resume device specification, as for the SNAPSHOT_SET_SWAP_FILE ioctl(),
78 and the offset); for swap partitions the offset is always 0, but it is
79 different to zero for swap files (please see
80 Documentation/swsusp-and-swap-files.txt for details).
81 The SNAPSHOT_SET_SWAP_AREA ioctl() is considered as a replacement for
82 SNAPSHOT_SET_SWAP_FILE which is regarded as obsolete. It is
83 recommended to always use this call, because the code to set the resume
84 partition may be removed from future kernels
86 SNAPSHOT_S2RAM - suspend to RAM; using this call causes the kernel to
87 immediately enter the suspend-to-RAM state, so this call must always
88 be preceded by the SNAPSHOT_FREEZE call and it is also necessary
89 to use the SNAPSHOT_UNFREEZE call after the system wakes up. This call
90 is needed to implement the suspend-to-both mechanism in which the
91 suspend image is first created, as though the system had been suspended
92 to disk, and then the system is suspended to RAM (this makes it possible
93 to resume the system from RAM if there's enough battery power or restore
94 its state on the basis of the saved suspend image otherwise)
96 SNAPSHOT_PMOPS - enable the usage of the hibernation_ops->prepare,
97 hibernate_ops->enter and hibernation_ops->finish methods (the in-kernel
98 swsusp knows these as the "platform method") which are needed on many
99 machines to (among others) speed up the resume by letting the BIOS skip
100 some steps or to let the system recognise the correct state of the
101 hardware after the resume (in particular on many machines this ensures
102 that unplugged AC adapters get correctly detected and that kacpid does
103 not run wild after the resume). The last ioctl() argument can take one
104 of the three values, defined in kernel/power/power.h:
105 PMOPS_PREPARE - make the kernel carry out the
106 hibernation_ops->prepare() operation
107 PMOPS_ENTER - make the kernel power off the system by calling
108 hibernation_ops->enter()
109 PMOPS_FINISH - make the kernel carry out the
110 hibernation_ops->finish() operation
111 Note that the actual constants are misnamed because they surface
112 internal kernel implementation details that have changed.
114 The device's read() operation can be used to transfer the snapshot image from
115 the kernel. It has the following limitations:
116 - you cannot read() more than one virtual memory page at a time
117 - read()s accross page boundaries are impossible (ie. if ypu read() 1/2 of
118 a page in the previous call, you will only be able to read()
119 _at_ _most_ 1/2 of the page in the next call)
121 The device's write() operation is used for uploading the system memory snapshot
122 into the kernel. It has the same limitations as the read() operation.
124 The release() operation frees all memory allocated for the snapshot image
125 and all swap pages allocated with SNAPSHOT_GET_SWAP_PAGE (if any).
126 Thus it is not necessary to use either SNAPSHOT_FREE or
127 SNAPSHOT_FREE_SWAP_PAGES before closing the device (in fact it will also
128 unfreeze user space processes frozen by SNAPSHOT_UNFREEZE if they are
129 still frozen when the device is being closed).
131 Currently it is assumed that the userland utilities reading/writing the
132 snapshot image from/to the kernel will use a swap parition, called the resume
133 partition, or a swap file as storage space (if a swap file is used, the resume
134 partition is the partition that holds this file). However, this is not really
135 required, as they can use, for example, a special (blank) suspend partition or
136 a file on a partition that is unmounted before SNAPSHOT_ATOMIC_SNAPSHOT and
139 These utilities SHOULD NOT make any assumptions regarding the ordering of
140 data within the snapshot image, except for the image header that MAY be
141 assumed to start with an swsusp_info structure, as specified in
142 kernel/power/power.h. This structure MAY be used by the userland utilities
143 to obtain some information about the snapshot image, such as the size
144 of the snapshot image, including the metadata and the header itself,
145 contained in the .size member of swsusp_info.
147 The snapshot image MUST be written to the kernel unaltered (ie. all of the image
148 data, metadata and header MUST be written in _exactly_ the same amount, form
149 and order in which they have been read). Otherwise, the behavior of the
150 resumed system may be totally unpredictable.
152 While executing SNAPSHOT_ATOMIC_RESTORE the kernel checks if the
153 structure of the snapshot image is consistent with the information stored
154 in the image header. If any inconsistencies are detected,
155 SNAPSHOT_ATOMIC_RESTORE will not succeed. Still, this is not a fool-proof
156 mechanism and the userland utilities using the interface SHOULD use additional
157 means, such as checksums, to ensure the integrity of the snapshot image.
159 The suspending and resuming utilities MUST lock themselves in memory,
160 preferrably using mlockall(), before calling SNAPSHOT_FREEZE.
162 The suspending utility MUST check the value stored by SNAPSHOT_ATOMIC_SNAPSHOT
163 in the memory location pointed to by the last argument of ioctl() and proceed
164 in accordance with it:
165 1. If the value is 1 (ie. the system memory snapshot has just been
166 created and the system is ready for saving it):
167 (a) The suspending utility MUST NOT close the snapshot device
168 _unless_ the whole suspend procedure is to be cancelled, in
169 which case, if the snapshot image has already been saved, the
170 suspending utility SHOULD destroy it, preferrably by zapping
171 its header. If the suspend is not to be cancelled, the
172 system MUST be powered off or rebooted after the snapshot
173 image has been saved.
174 (b) The suspending utility SHOULD NOT attempt to perform any
175 file system operations (including reads) on the file systems
176 that were mounted before SNAPSHOT_ATOMIC_SNAPSHOT has been
177 called. However, it MAY mount a file system that was not
178 mounted at that time and perform some operations on it (eg.
179 use it for saving the image).
180 2. If the value is 0 (ie. the system state has just been restored from
181 the snapshot image), the suspending utility MUST close the snapshot
182 device. Afterwards it will be treated as a regular userland process,
185 The resuming utility SHOULD NOT attempt to mount any file systems that could
186 be mounted before suspend and SHOULD NOT attempt to perform any operations
187 involving such file systems.
189 For details, please refer to the source code.