userland-swsusp.txt 9.6 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187
  1. Documentation for userland software suspend interface
  2. (C) 2006 Rafael J. Wysocki <rjw@sisk.pl>
  3. First, the warnings at the beginning of swsusp.txt still apply.
  4. Second, you should read the FAQ in swsusp.txt _now_ if you have not
  5. done it already.
  6. Now, to use the userland interface for software suspend you need special
  7. utilities that will read/write the system memory snapshot from/to the
  8. kernel. Such utilities are available, for example, from
  9. <http://suspend.sourceforge.net>. You may want to have a look at them if you
  10. are going to develop your own suspend/resume utilities.
  11. The interface consists of a character device providing the open(),
  12. release(), read(), and write() operations as well as several ioctl()
  13. commands defined in kernel/power/power.h. The major and minor
  14. numbers of the device are, respectively, 10 and 231, and they can
  15. be read from /sys/class/misc/snapshot/dev.
  16. The device can be open either for reading or for writing. If open for
  17. reading, it is considered to be in the suspend mode. Otherwise it is
  18. assumed to be in the resume mode. The device cannot be open for simultaneous
  19. reading and writing. It is also impossible to have the device open more than
  20. once at a time.
  21. The ioctl() commands recognized by the device are:
  22. SNAPSHOT_FREEZE - freeze user space processes (the current process is
  23. not frozen); this is required for SNAPSHOT_ATOMIC_SNAPSHOT
  24. and SNAPSHOT_ATOMIC_RESTORE to succeed
  25. SNAPSHOT_UNFREEZE - thaw user space processes frozen by SNAPSHOT_FREEZE
  26. SNAPSHOT_ATOMIC_SNAPSHOT - create a snapshot of the system memory; the
  27. last argument of ioctl() should be a pointer to an int variable,
  28. the value of which will indicate whether the call returned after
  29. creating the snapshot (1) or after restoring the system memory state
  30. from it (0) (after resume the system finds itself finishing the
  31. SNAPSHOT_ATOMIC_SNAPSHOT ioctl() again); after the snapshot
  32. has been created the read() operation can be used to transfer
  33. it out of the kernel
  34. SNAPSHOT_ATOMIC_RESTORE - restore the system memory state from the
  35. uploaded snapshot image; before calling it you should transfer
  36. the system memory snapshot back to the kernel using the write()
  37. operation; this call will not succeed if the snapshot
  38. image is not available to the kernel
  39. SNAPSHOT_FREE - free memory allocated for the snapshot image
  40. SNAPSHOT_SET_IMAGE_SIZE - set the preferred maximum size of the image
  41. (the kernel will do its best to ensure the image size will not exceed
  42. this number, but if it turns out to be impossible, the kernel will
  43. create the smallest image possible)
  44. SNAPSHOT_AVAIL_SWAP - return the amount of available swap in bytes (the last
  45. argument should be a pointer to an unsigned int variable that will
  46. contain the result if the call is successful).
  47. SNAPSHOT_GET_SWAP_PAGE - allocate a swap page from the resume partition
  48. (the last argument should be a pointer to a loff_t variable that
  49. will contain the swap page offset if the call is successful)
  50. SNAPSHOT_FREE_SWAP_PAGES - free all swap pages allocated with
  51. SNAPSHOT_GET_SWAP_PAGE
  52. SNAPSHOT_SET_SWAP_FILE - set the resume partition (the last ioctl() argument
  53. should specify the device's major and minor numbers in the old
  54. two-byte format, as returned by the stat() function in the .st_rdev
  55. member of the stat structure)
  56. SNAPSHOT_SET_SWAP_AREA - set the resume partition and the offset (in <PAGE_SIZE>
  57. units) from the beginning of the partition at which the swap header is
  58. located (the last ioctl() argument should point to a struct
  59. resume_swap_area, as defined in kernel/power/power.h, containing the
  60. resume device specification, as for the SNAPSHOT_SET_SWAP_FILE ioctl(),
  61. and the offset); for swap partitions the offset is always 0, but it is
  62. different to zero for swap files (please see
  63. Documentation/swsusp-and-swap-files.txt for details).
  64. The SNAPSHOT_SET_SWAP_AREA ioctl() is considered as a replacement for
  65. SNAPSHOT_SET_SWAP_FILE which is regarded as obsolete. It is
  66. recommended to always use this call, because the code to set the resume
  67. partition may be removed from future kernels
  68. SNAPSHOT_S2RAM - suspend to RAM; using this call causes the kernel to
  69. immediately enter the suspend-to-RAM state, so this call must always
  70. be preceded by the SNAPSHOT_FREEZE call and it is also necessary
  71. to use the SNAPSHOT_UNFREEZE call after the system wakes up. This call
  72. is needed to implement the suspend-to-both mechanism in which the
  73. suspend image is first created, as though the system had been suspended
  74. to disk, and then the system is suspended to RAM (this makes it possible
  75. to resume the system from RAM if there's enough battery power or restore
  76. its state on the basis of the saved suspend image otherwise)
  77. SNAPSHOT_PMOPS - enable the usage of the pmops->prepare, pmops->enter and
  78. pmops->finish methods (the in-kernel swsusp knows these as the "platform
  79. method") which are needed on many machines to (among others) speed up
  80. the resume by letting the BIOS skip some steps or to let the system
  81. recognise the correct state of the hardware after the resume (in
  82. particular on many machines this ensures that unplugged AC
  83. adapters get correctly detected and that kacpid does not run wild after
  84. the resume). The last ioctl() argument can take one of the three
  85. values, defined in kernel/power/power.h:
  86. PMOPS_PREPARE - make the kernel carry out the
  87. pm_ops->prepare(PM_SUSPEND_DISK) operation
  88. PMOPS_ENTER - make the kernel power off the system by calling
  89. pm_ops->enter(PM_SUSPEND_DISK)
  90. PMOPS_FINISH - make the kernel carry out the
  91. pm_ops->finish(PM_SUSPEND_DISK) operation
  92. The device's read() operation can be used to transfer the snapshot image from
  93. the kernel. It has the following limitations:
  94. - you cannot read() more than one virtual memory page at a time
  95. - read()s accross page boundaries are impossible (ie. if ypu read() 1/2 of
  96. a page in the previous call, you will only be able to read()
  97. _at_ _most_ 1/2 of the page in the next call)
  98. The device's write() operation is used for uploading the system memory snapshot
  99. into the kernel. It has the same limitations as the read() operation.
  100. The release() operation frees all memory allocated for the snapshot image
  101. and all swap pages allocated with SNAPSHOT_GET_SWAP_PAGE (if any).
  102. Thus it is not necessary to use either SNAPSHOT_FREE or
  103. SNAPSHOT_FREE_SWAP_PAGES before closing the device (in fact it will also
  104. unfreeze user space processes frozen by SNAPSHOT_UNFREEZE if they are
  105. still frozen when the device is being closed).
  106. Currently it is assumed that the userland utilities reading/writing the
  107. snapshot image from/to the kernel will use a swap parition, called the resume
  108. partition, or a swap file as storage space (if a swap file is used, the resume
  109. partition is the partition that holds this file). However, this is not really
  110. required, as they can use, for example, a special (blank) suspend partition or
  111. a file on a partition that is unmounted before SNAPSHOT_ATOMIC_SNAPSHOT and
  112. mounted afterwards.
  113. These utilities SHOULD NOT make any assumptions regarding the ordering of
  114. data within the snapshot image, except for the image header that MAY be
  115. assumed to start with an swsusp_info structure, as specified in
  116. kernel/power/power.h. This structure MAY be used by the userland utilities
  117. to obtain some information about the snapshot image, such as the size
  118. of the snapshot image, including the metadata and the header itself,
  119. contained in the .size member of swsusp_info.
  120. The snapshot image MUST be written to the kernel unaltered (ie. all of the image
  121. data, metadata and header MUST be written in _exactly_ the same amount, form
  122. and order in which they have been read). Otherwise, the behavior of the
  123. resumed system may be totally unpredictable.
  124. While executing SNAPSHOT_ATOMIC_RESTORE the kernel checks if the
  125. structure of the snapshot image is consistent with the information stored
  126. in the image header. If any inconsistencies are detected,
  127. SNAPSHOT_ATOMIC_RESTORE will not succeed. Still, this is not a fool-proof
  128. mechanism and the userland utilities using the interface SHOULD use additional
  129. means, such as checksums, to ensure the integrity of the snapshot image.
  130. The suspending and resuming utilities MUST lock themselves in memory,
  131. preferrably using mlockall(), before calling SNAPSHOT_FREEZE.
  132. The suspending utility MUST check the value stored by SNAPSHOT_ATOMIC_SNAPSHOT
  133. in the memory location pointed to by the last argument of ioctl() and proceed
  134. in accordance with it:
  135. 1. If the value is 1 (ie. the system memory snapshot has just been
  136. created and the system is ready for saving it):
  137. (a) The suspending utility MUST NOT close the snapshot device
  138. _unless_ the whole suspend procedure is to be cancelled, in
  139. which case, if the snapshot image has already been saved, the
  140. suspending utility SHOULD destroy it, preferrably by zapping
  141. its header. If the suspend is not to be cancelled, the
  142. system MUST be powered off or rebooted after the snapshot
  143. image has been saved.
  144. (b) The suspending utility SHOULD NOT attempt to perform any
  145. file system operations (including reads) on the file systems
  146. that were mounted before SNAPSHOT_ATOMIC_SNAPSHOT has been
  147. called. However, it MAY mount a file system that was not
  148. mounted at that time and perform some operations on it (eg.
  149. use it for saving the image).
  150. 2. If the value is 0 (ie. the system state has just been restored from
  151. the snapshot image), the suspending utility MUST close the snapshot
  152. device. Afterwards it will be treated as a regular userland process,
  153. so it need not exit.
  154. The resuming utility SHOULD NOT attempt to mount any file systems that could
  155. be mounted before suspend and SHOULD NOT attempt to perform any operations
  156. involving such file systems.
  157. For details, please refer to the source code.