userland-swsusp.rst 8.7 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193
  1. =====================================================
  2. Documentation for userland software suspend interface
  3. =====================================================
  4. (C) 2006 Rafael J. Wysocki <rjw@sisk.pl>
  5. 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
  7. done it already.
  8. Now, to use the userland interface for software suspend you need special
  9. utilities that will read/write the system memory snapshot from/to the
  10. kernel. Such utilities are available, for example, from
  11. <http://suspend.sourceforge.net>. You may want to have a look at them if you
  12. are going to develop your own suspend/resume utilities.
  13. The interface consists of a character device providing the open(),
  14. release(), read(), and write() operations as well as several ioctl()
  15. commands defined in include/linux/suspend_ioctls.h . The major and minor
  16. numbers of the device are, respectively, 10 and 231, and they can
  17. be read from /sys/class/misc/snapshot/dev.
  18. The device can be open either for reading or for writing. If open for
  19. reading, it is considered to be in the suspend mode. Otherwise it is
  20. assumed to be in the resume mode. The device cannot be open for simultaneous
  21. reading and writing. It is also impossible to have the device open more than
  22. once at a time.
  23. Even opening the device has side effects. Data structures are
  24. allocated, and PM_HIBERNATION_PREPARE / PM_RESTORE_PREPARE chains are
  25. called.
  26. The ioctl() commands recognized by the device are:
  27. SNAPSHOT_FREEZE
  28. freeze user space processes (the current process is
  29. not frozen); this is required for SNAPSHOT_CREATE_IMAGE
  30. and SNAPSHOT_ATOMIC_RESTORE to succeed
  31. SNAPSHOT_UNFREEZE
  32. thaw user space processes frozen by SNAPSHOT_FREEZE
  33. SNAPSHOT_CREATE_IMAGE
  34. create a snapshot of the system memory; the
  35. last argument of ioctl() should be a pointer to an int variable,
  36. the value of which will indicate whether the call returned after
  37. creating the snapshot (1) or after restoring the system memory state
  38. from it (0) (after resume the system finds itself finishing the
  39. SNAPSHOT_CREATE_IMAGE ioctl() again); after the snapshot
  40. has been created the read() operation can be used to transfer
  41. it out of the kernel
  42. SNAPSHOT_ATOMIC_RESTORE
  43. restore the system memory state from the
  44. uploaded snapshot image; before calling it you should transfer
  45. the system memory snapshot back to the kernel using the write()
  46. operation; this call will not succeed if the snapshot
  47. image is not available to the kernel
  48. SNAPSHOT_FREE
  49. free memory allocated for the snapshot image
  50. SNAPSHOT_PREF_IMAGE_SIZE
  51. set the preferred maximum size of the image
  52. (the kernel will do its best to ensure the image size will not exceed
  53. this number, but if it turns out to be impossible, the kernel will
  54. create the smallest image possible)
  55. SNAPSHOT_GET_IMAGE_SIZE
  56. return the actual size of the hibernation image
  57. (the last argument should be a pointer to a loff_t variable that
  58. will contain the result if the call is successful)
  59. SNAPSHOT_AVAIL_SWAP_SIZE
  60. return the amount of available swap in bytes
  61. (the last argument should be a pointer to a loff_t variable that
  62. will contain the result if the call is successful)
  63. SNAPSHOT_ALLOC_SWAP_PAGE
  64. allocate a swap page from the resume partition
  65. (the last argument should be a pointer to a loff_t variable that
  66. will contain the swap page offset if the call is successful)
  67. SNAPSHOT_FREE_SWAP_PAGES
  68. free all swap pages allocated by
  69. SNAPSHOT_ALLOC_SWAP_PAGE
  70. SNAPSHOT_SET_SWAP_AREA
  71. set the resume partition and the offset (in <PAGE_SIZE>
  72. units) from the beginning of the partition at which the swap header is
  73. located (the last ioctl() argument should point to a struct
  74. resume_swap_area, as defined in kernel/power/suspend_ioctls.h,
  75. containing the resume device specification and the offset); for swap
  76. partitions the offset is always 0, but it is different from zero for
  77. swap files (see Documentation/power/swsusp-and-swap-files.rst for
  78. details).
  79. SNAPSHOT_PLATFORM_SUPPORT
  80. enable/disable the hibernation platform support,
  81. depending on the argument value (enable, if the argument is nonzero)
  82. SNAPSHOT_POWER_OFF
  83. make the kernel transition the system to the hibernation
  84. state (eg. ACPI S4) using the platform (eg. ACPI) driver
  85. SNAPSHOT_S2RAM
  86. 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)
  95. The device's read() operation can be used to transfer the snapshot image from
  96. the kernel. It has the following limitations:
  97. - you cannot read() more than one virtual memory page at a time
  98. - read()s across page boundaries are impossible (ie. if you read() 1/2 of
  99. a page in the previous call, you will only be able to read()
  100. **at most** 1/2 of the page in the next call)
  101. The device's write() operation is used for uploading the system memory snapshot
  102. into the kernel. It has the same limitations as the read() operation.
  103. The release() operation frees all memory allocated for the snapshot image
  104. and all swap pages allocated with SNAPSHOT_ALLOC_SWAP_PAGE (if any).
  105. Thus it is not necessary to use either SNAPSHOT_FREE or
  106. SNAPSHOT_FREE_SWAP_PAGES before closing the device (in fact it will also
  107. unfreeze user space processes frozen by SNAPSHOT_UNFREEZE if they are
  108. still frozen when the device is being closed).
  109. Currently it is assumed that the userland utilities reading/writing the
  110. snapshot image from/to the kernel will use a swap partition, called the resume
  111. partition, or a swap file as storage space (if a swap file is used, the resume
  112. partition is the partition that holds this file). However, this is not really
  113. required, as they can use, for example, a special (blank) suspend partition or
  114. a file on a partition that is unmounted before SNAPSHOT_CREATE_IMAGE and
  115. mounted afterwards.
  116. These utilities MUST NOT make any assumptions regarding the ordering of
  117. data within the snapshot image. The contents of the image are entirely owned
  118. by the kernel and its structure may be changed in future kernel releases.
  119. The snapshot image MUST be written to the kernel unaltered (ie. all of the image
  120. data, metadata and header MUST be written in _exactly_ the same amount, form
  121. and order in which they have been read). Otherwise, the behavior of the
  122. resumed system may be totally unpredictable.
  123. While executing SNAPSHOT_ATOMIC_RESTORE the kernel checks if the
  124. structure of the snapshot image is consistent with the information stored
  125. in the image header. If any inconsistencies are detected,
  126. SNAPSHOT_ATOMIC_RESTORE will not succeed. Still, this is not a fool-proof
  127. mechanism and the userland utilities using the interface SHOULD use additional
  128. means, such as checksums, to ensure the integrity of the snapshot image.
  129. The suspending and resuming utilities MUST lock themselves in memory,
  130. preferably using mlockall(), before calling SNAPSHOT_FREEZE.
  131. The suspending utility MUST check the value stored by SNAPSHOT_CREATE_IMAGE
  132. in the memory location pointed to by the last argument of ioctl() and proceed
  133. in accordance with it:
  134. 1. If the value is 1 (ie. the system memory snapshot has just been
  135. created and the system is ready for saving it):
  136. (a) The suspending utility MUST NOT close the snapshot device
  137. _unless_ the whole suspend procedure is to be cancelled, in
  138. which case, if the snapshot image has already been saved, the
  139. suspending utility SHOULD destroy it, preferably by zapping
  140. its header. If the suspend is not to be cancelled, the
  141. system MUST be powered off or rebooted after the snapshot
  142. image has been saved.
  143. (b) The suspending utility SHOULD NOT attempt to perform any
  144. file system operations (including reads) on the file systems
  145. that were mounted before SNAPSHOT_CREATE_IMAGE has been
  146. called. However, it MAY mount a file system that was not
  147. mounted at that time and perform some operations on it (eg.
  148. use it for saving the image).
  149. 2. If the value is 0 (ie. the system state has just been restored from
  150. the snapshot image), the suspending utility MUST close the snapshot
  151. device. Afterwards it will be treated as a regular userland process,
  152. so it need not exit.
  153. The resuming utility SHOULD NOT attempt to mount any file systems that could
  154. be mounted before suspend and SHOULD NOT attempt to perform any operations
  155. involving such file systems.
  156. For details, please refer to the source code.