0001 =====================================================
0002 Documentation for userland software suspend interface
0003 =====================================================
0004
0005 (C) 2006 Rafael J. Wysocki <rjw@sisk.pl>
0006
0007 First, the warnings at the beginning of swsusp.txt still apply.
0008
0009 Second, you should read the FAQ in swsusp.txt _now_ if you have not
0010 done it already.
0011
0012 Now, to use the userland interface for software suspend you need special
0013 utilities that will read/write the system memory snapshot from/to the
0014 kernel. Such utilities are available, for example, from
0015 <http://suspend.sourceforge.net>. You may want to have a look at them if you
0016 are going to develop your own suspend/resume utilities.
0017
0018 The interface consists of a character device providing the open(),
0019 release(), read(), and write() operations as well as several ioctl()
0020 commands defined in include/linux/suspend_ioctls.h . The major and minor
0021 numbers of the device are, respectively, 10 and 231, and they can
0022 be read from /sys/class/misc/snapshot/dev.
0023
0024 The device can be open either for reading or for writing. If open for
0025 reading, it is considered to be in the suspend mode. Otherwise it is
0026 assumed to be in the resume mode. The device cannot be open for simultaneous
0027 reading and writing. It is also impossible to have the device open more than
0028 once at a time.
0029
0030 Even opening the device has side effects. Data structures are
0031 allocated, and PM_HIBERNATION_PREPARE / PM_RESTORE_PREPARE chains are
0032 called.
0033
0034 The ioctl() commands recognized by the device are:
0035
0036 SNAPSHOT_FREEZE
0037 freeze user space processes (the current process is
0038 not frozen); this is required for SNAPSHOT_CREATE_IMAGE
0039 and SNAPSHOT_ATOMIC_RESTORE to succeed
0040
0041 SNAPSHOT_UNFREEZE
0042 thaw user space processes frozen by SNAPSHOT_FREEZE
0043
0044 SNAPSHOT_CREATE_IMAGE
0045 create a snapshot of the system memory; the
0046 last argument of ioctl() should be a pointer to an int variable,
0047 the value of which will indicate whether the call returned after
0048 creating the snapshot (1) or after restoring the system memory state
0049 from it (0) (after resume the system finds itself finishing the
0050 SNAPSHOT_CREATE_IMAGE ioctl() again); after the snapshot
0051 has been created the read() operation can be used to transfer
0052 it out of the kernel
0053
0054 SNAPSHOT_ATOMIC_RESTORE
0055 restore the system memory state from the
0056 uploaded snapshot image; before calling it you should transfer
0057 the system memory snapshot back to the kernel using the write()
0058 operation; this call will not succeed if the snapshot
0059 image is not available to the kernel
0060
0061 SNAPSHOT_FREE
0062 free memory allocated for the snapshot image
0063
0064 SNAPSHOT_PREF_IMAGE_SIZE
0065 set the preferred maximum size of the image
0066 (the kernel will do its best to ensure the image size will not exceed
0067 this number, but if it turns out to be impossible, the kernel will
0068 create the smallest image possible)
0069
0070 SNAPSHOT_GET_IMAGE_SIZE
0071 return the actual size of the hibernation image
0072 (the last argument should be a pointer to a loff_t variable that
0073 will contain the result if the call is successful)
0074
0075 SNAPSHOT_AVAIL_SWAP_SIZE
0076 return the amount of available swap in bytes
0077 (the last argument should be a pointer to a loff_t variable that
0078 will contain the result if the call is successful)
0079
0080 SNAPSHOT_ALLOC_SWAP_PAGE
0081 allocate a swap page from the resume partition
0082 (the last argument should be a pointer to a loff_t variable that
0083 will contain the swap page offset if the call is successful)
0084
0085 SNAPSHOT_FREE_SWAP_PAGES
0086 free all swap pages allocated by
0087 SNAPSHOT_ALLOC_SWAP_PAGE
0088
0089 SNAPSHOT_SET_SWAP_AREA
0090 set the resume partition and the offset (in <PAGE_SIZE>
0091 units) from the beginning of the partition at which the swap header is
0092 located (the last ioctl() argument should point to a struct
0093 resume_swap_area, as defined in kernel/power/suspend_ioctls.h,
0094 containing the resume device specification and the offset); for swap
0095 partitions the offset is always 0, but it is different from zero for
0096 swap files (see Documentation/power/swsusp-and-swap-files.rst for
0097 details).
0098
0099 SNAPSHOT_PLATFORM_SUPPORT
0100 enable/disable the hibernation platform support,
0101 depending on the argument value (enable, if the argument is nonzero)
0102
0103 SNAPSHOT_POWER_OFF
0104 make the kernel transition the system to the hibernation
0105 state (eg. ACPI S4) using the platform (eg. ACPI) driver
0106
0107 SNAPSHOT_S2RAM
0108 suspend to RAM; using this call causes the kernel to
0109 immediately enter the suspend-to-RAM state, so this call must always
0110 be preceded by the SNAPSHOT_FREEZE call and it is also necessary
0111 to use the SNAPSHOT_UNFREEZE call after the system wakes up. This call
0112 is needed to implement the suspend-to-both mechanism in which the
0113 suspend image is first created, as though the system had been suspended
0114 to disk, and then the system is suspended to RAM (this makes it possible
0115 to resume the system from RAM if there's enough battery power or restore
0116 its state on the basis of the saved suspend image otherwise)
0117
0118 The device's read() operation can be used to transfer the snapshot image from
0119 the kernel. It has the following limitations:
0120
0121 - you cannot read() more than one virtual memory page at a time
0122 - read()s across page boundaries are impossible (ie. if you read() 1/2 of
0123 a page in the previous call, you will only be able to read()
0124 **at most** 1/2 of the page in the next call)
0125
0126 The device's write() operation is used for uploading the system memory snapshot
0127 into the kernel. It has the same limitations as the read() operation.
0128
0129 The release() operation frees all memory allocated for the snapshot image
0130 and all swap pages allocated with SNAPSHOT_ALLOC_SWAP_PAGE (if any).
0131 Thus it is not necessary to use either SNAPSHOT_FREE or
0132 SNAPSHOT_FREE_SWAP_PAGES before closing the device (in fact it will also
0133 unfreeze user space processes frozen by SNAPSHOT_UNFREEZE if they are
0134 still frozen when the device is being closed).
0135
0136 Currently it is assumed that the userland utilities reading/writing the
0137 snapshot image from/to the kernel will use a swap partition, called the resume
0138 partition, or a swap file as storage space (if a swap file is used, the resume
0139 partition is the partition that holds this file). However, this is not really
0140 required, as they can use, for example, a special (blank) suspend partition or
0141 a file on a partition that is unmounted before SNAPSHOT_CREATE_IMAGE and
0142 mounted afterwards.
0143
0144 These utilities MUST NOT make any assumptions regarding the ordering of
0145 data within the snapshot image. The contents of the image are entirely owned
0146 by the kernel and its structure may be changed in future kernel releases.
0147
0148 The snapshot image MUST be written to the kernel unaltered (ie. all of the image
0149 data, metadata and header MUST be written in _exactly_ the same amount, form
0150 and order in which they have been read). Otherwise, the behavior of the
0151 resumed system may be totally unpredictable.
0152
0153 While executing SNAPSHOT_ATOMIC_RESTORE the kernel checks if the
0154 structure of the snapshot image is consistent with the information stored
0155 in the image header. If any inconsistencies are detected,
0156 SNAPSHOT_ATOMIC_RESTORE will not succeed. Still, this is not a fool-proof
0157 mechanism and the userland utilities using the interface SHOULD use additional
0158 means, such as checksums, to ensure the integrity of the snapshot image.
0159
0160 The suspending and resuming utilities MUST lock themselves in memory,
0161 preferably using mlockall(), before calling SNAPSHOT_FREEZE.
0162
0163 The suspending utility MUST check the value stored by SNAPSHOT_CREATE_IMAGE
0164 in the memory location pointed to by the last argument of ioctl() and proceed
0165 in accordance with it:
0166
0167 1. If the value is 1 (ie. the system memory snapshot has just been
0168 created and the system is ready for saving it):
0169
0170 (a) The suspending utility MUST NOT close the snapshot device
0171 _unless_ the whole suspend procedure is to be cancelled, in
0172 which case, if the snapshot image has already been saved, the
0173 suspending utility SHOULD destroy it, preferably by zapping
0174 its header. If the suspend is not to be cancelled, the
0175 system MUST be powered off or rebooted after the snapshot
0176 image has been saved.
0177 (b) The suspending utility SHOULD NOT attempt to perform any
0178 file system operations (including reads) on the file systems
0179 that were mounted before SNAPSHOT_CREATE_IMAGE has been
0180 called. However, it MAY mount a file system that was not
0181 mounted at that time and perform some operations on it (eg.
0182 use it for saving the image).
0183
0184 2. If the value is 0 (ie. the system state has just been restored from
0185 the snapshot image), the suspending utility MUST close the snapshot
0186 device. Afterwards it will be treated as a regular userland process,
0187 so it need not exit.
0188
0189 The resuming utility SHOULD NOT attempt to mount any file systems that could
0190 be mounted before suspend and SHOULD NOT attempt to perform any operations
0191 involving such file systems.
0192
0193 For details, please refer to the source code.
