swsusp.txt 17 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426
  1. Some warnings, first.
  2. * BIG FAT WARNING *********************************************************
  3. *
  4. * If you touch anything on disk between suspend and resume...
  5. * ...kiss your data goodbye.
  6. *
  7. * If you do resume from initrd after your filesystems are mounted...
  8. * ...bye bye root partition.
  9. * [this is actually same case as above]
  10. *
  11. * If you have unsupported (*) devices using DMA, you may have some
  12. * problems. If your disk driver does not support suspend... (IDE does),
  13. * it may cause some problems, too. If you change kernel command line
  14. * between suspend and resume, it may do something wrong. If you change
  15. * your hardware while system is suspended... well, it was not good idea;
  16. * but it will probably only crash.
  17. *
  18. * (*) suspend/resume support is needed to make it safe.
  19. *
  20. * If you have any filesystems on USB devices mounted before software suspend,
  21. * they won't be accessible after resume and you may lose data, as though
  22. * you have unplugged the USB devices with mounted filesystems on them;
  23. * see the FAQ below for details. (This is not true for more traditional
  24. * power states like "standby", which normally don't turn USB off.)
  25. You need to append resume=/dev/your_swap_partition to kernel command
  26. line. Then you suspend by
  27. echo shutdown > /sys/power/disk; echo disk > /sys/power/state
  28. . If you feel ACPI works pretty well on your system, you might try
  29. echo platform > /sys/power/disk; echo disk > /sys/power/state
  30. . If you would like to write hibernation image to swap and then suspend
  31. to RAM (provided your platform supports it), you can try
  32. echo suspend > /sys/power/disk; echo disk > /sys/power/state
  33. . If you have SATA disks, you'll need recent kernels with SATA suspend
  34. support. For suspend and resume to work, make sure your disk drivers
  35. are built into kernel -- not modules. [There's way to make
  36. suspend/resume with modular disk drivers, see FAQ, but you probably
  37. should not do that.]
  38. If you want to limit the suspend image size to N bytes, do
  39. echo N > /sys/power/image_size
  40. before suspend (it is limited to 500 MB by default).
  41. . The resume process checks for the presence of the resume device,
  42. if found, it then checks the contents for the hibernation image signature.
  43. If both are found, it resumes the hibernation image.
  44. . The resume process may be triggered in two ways:
  45. 1) During lateinit: If resume=/dev/your_swap_partition is specified on
  46. the kernel command line, lateinit runs the resume process. If the
  47. resume device has not been probed yet, the resume process fails and
  48. bootup continues.
  49. 2) Manually from an initrd or initramfs: May be run from
  50. the init script by using the /sys/power/resume file. It is vital
  51. that this be done prior to remounting any filesystems (even as
  52. read-only) otherwise data may be corrupted.
  53. Article about goals and implementation of Software Suspend for Linux
  54. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  55. Author: Gábor Kuti
  56. Last revised: 2003-10-20 by Pavel Machek
  57. Idea and goals to achieve
  58. Nowadays it is common in several laptops that they have a suspend button. It
  59. saves the state of the machine to a filesystem or to a partition and switches
  60. to standby mode. Later resuming the machine the saved state is loaded back to
  61. ram and the machine can continue its work. It has two real benefits. First we
  62. save ourselves the time machine goes down and later boots up, energy costs
  63. are real high when running from batteries. The other gain is that we don't have to
  64. interrupt our programs so processes that are calculating something for a long
  65. time shouldn't need to be written interruptible.
  66. swsusp saves the state of the machine into active swaps and then reboots or
  67. powerdowns. You must explicitly specify the swap partition to resume from with
  68. ``resume='' kernel option. If signature is found it loads and restores saved
  69. state. If the option ``noresume'' is specified as a boot parameter, it skips
  70. the resuming. If the option ``hibernate=nocompress'' is specified as a boot
  71. parameter, it saves hibernation image without compression.
  72. In the meantime while the system is suspended you should not add/remove any
  73. of the hardware, write to the filesystems, etc.
  74. Sleep states summary
  75. ====================
  76. There are three different interfaces you can use, /proc/acpi should
  77. work like this:
  78. In a really perfect world:
  79. echo 1 > /proc/acpi/sleep # for standby
  80. echo 2 > /proc/acpi/sleep # for suspend to ram
  81. echo 3 > /proc/acpi/sleep # for suspend to ram, but with more power conservative
  82. echo 4 > /proc/acpi/sleep # for suspend to disk
  83. echo 5 > /proc/acpi/sleep # for shutdown unfriendly the system
  84. and perhaps
  85. echo 4b > /proc/acpi/sleep # for suspend to disk via s4bios
  86. Frequently Asked Questions
  87. ==========================
  88. Q: well, suspending a server is IMHO a really stupid thing,
  89. but... (Diego Zuccato):
  90. A: You bought new UPS for your server. How do you install it without
  91. bringing machine down? Suspend to disk, rearrange power cables,
  92. resume.
  93. You have your server on UPS. Power died, and UPS is indicating 30
  94. seconds to failure. What do you do? Suspend to disk.
  95. Q: Maybe I'm missing something, but why don't the regular I/O paths work?
  96. A: We do use the regular I/O paths. However we cannot restore the data
  97. to its original location as we load it. That would create an
  98. inconsistent kernel state which would certainly result in an oops.
  99. Instead, we load the image into unused memory and then atomically copy
  100. it back to it original location. This implies, of course, a maximum
  101. image size of half the amount of memory.
  102. There are two solutions to this:
  103. * require half of memory to be free during suspend. That way you can
  104. read "new" data onto free spots, then cli and copy
  105. * assume we had special "polling" ide driver that only uses memory
  106. between 0-640KB. That way, I'd have to make sure that 0-640KB is free
  107. during suspending, but otherwise it would work...
  108. suspend2 shares this fundamental limitation, but does not include user
  109. data and disk caches into "used memory" by saving them in
  110. advance. That means that the limitation goes away in practice.
  111. Q: Does linux support ACPI S4?
  112. A: Yes. That's what echo platform > /sys/power/disk does.
  113. Q: What is 'suspend2'?
  114. A: suspend2 is 'Software Suspend 2', a forked implementation of
  115. suspend-to-disk which is available as separate patches for 2.4 and 2.6
  116. kernels from swsusp.sourceforge.net. It includes support for SMP, 4GB
  117. highmem and preemption. It also has a extensible architecture that
  118. allows for arbitrary transformations on the image (compression,
  119. encryption) and arbitrary backends for writing the image (eg to swap
  120. or an NFS share[Work In Progress]). Questions regarding suspend2
  121. should be sent to the mailing list available through the suspend2
  122. website, and not to the Linux Kernel Mailing List. We are working
  123. toward merging suspend2 into the mainline kernel.
  124. Q: What is the freezing of tasks and why are we using it?
  125. A: The freezing of tasks is a mechanism by which user space processes and some
  126. kernel threads are controlled during hibernation or system-wide suspend (on some
  127. architectures). See freezing-of-tasks.txt for details.
  128. Q: What is the difference between "platform" and "shutdown"?
  129. A:
  130. shutdown: save state in linux, then tell bios to powerdown
  131. platform: save state in linux, then tell bios to powerdown and blink
  132. "suspended led"
  133. "platform" is actually right thing to do where supported, but
  134. "shutdown" is most reliable (except on ACPI systems).
  135. Q: I do not understand why you have such strong objections to idea of
  136. selective suspend.
  137. A: Do selective suspend during runtime power management, that's okay. But
  138. it's useless for suspend-to-disk. (And I do not see how you could use
  139. it for suspend-to-ram, I hope you do not want that).
  140. Lets see, so you suggest to
  141. * SUSPEND all but swap device and parents
  142. * Snapshot
  143. * Write image to disk
  144. * SUSPEND swap device and parents
  145. * Powerdown
  146. Oh no, that does not work, if swap device or its parents uses DMA,
  147. you've corrupted data. You'd have to do
  148. * SUSPEND all but swap device and parents
  149. * FREEZE swap device and parents
  150. * Snapshot
  151. * UNFREEZE swap device and parents
  152. * Write
  153. * SUSPEND swap device and parents
  154. Which means that you still need that FREEZE state, and you get more
  155. complicated code. (And I have not yet introduce details like system
  156. devices).
  157. Q: There don't seem to be any generally useful behavioral
  158. distinctions between SUSPEND and FREEZE.
  159. A: Doing SUSPEND when you are asked to do FREEZE is always correct,
  160. but it may be unnecessarily slow. If you want your driver to stay simple,
  161. slowness may not matter to you. It can always be fixed later.
  162. For devices like disk it does matter, you do not want to spindown for
  163. FREEZE.
  164. Q: After resuming, system is paging heavily, leading to very bad interactivity.
  165. A: Try running
  166. cat `cat /proc/[0-9]*/maps | grep / | sed 's:.* /:/:' | sort -u` > /dev/null
  167. after resume. swapoff -a; swapon -a may also be useful.
  168. Q: What happens to devices during swsusp? They seem to be resumed
  169. during system suspend?
  170. A: That's correct. We need to resume them if we want to write image to
  171. disk. Whole sequence goes like
  172. Suspend part
  173. ~~~~~~~~~~~~
  174. running system, user asks for suspend-to-disk
  175. user processes are stopped
  176. suspend(PMSG_FREEZE): devices are frozen so that they don't interfere
  177. with state snapshot
  178. state snapshot: copy of whole used memory is taken with interrupts disabled
  179. resume(): devices are woken up so that we can write image to swap
  180. write image to swap
  181. suspend(PMSG_SUSPEND): suspend devices so that we can power off
  182. turn the power off
  183. Resume part
  184. ~~~~~~~~~~~
  185. (is actually pretty similar)
  186. running system, user asks for suspend-to-disk
  187. user processes are stopped (in common case there are none, but with resume-from-initrd, no one knows)
  188. read image from disk
  189. suspend(PMSG_FREEZE): devices are frozen so that they don't interfere
  190. with image restoration
  191. image restoration: rewrite memory with image
  192. resume(): devices are woken up so that system can continue
  193. thaw all user processes
  194. Q: What is this 'Encrypt suspend image' for?
  195. A: First of all: it is not a replacement for dm-crypt encrypted swap.
  196. It cannot protect your computer while it is suspended. Instead it does
  197. protect from leaking sensitive data after resume from suspend.
  198. Think of the following: you suspend while an application is running
  199. that keeps sensitive data in memory. The application itself prevents
  200. the data from being swapped out. Suspend, however, must write these
  201. data to swap to be able to resume later on. Without suspend encryption
  202. your sensitive data are then stored in plaintext on disk. This means
  203. that after resume your sensitive data are accessible to all
  204. applications having direct access to the swap device which was used
  205. for suspend. If you don't need swap after resume these data can remain
  206. on disk virtually forever. Thus it can happen that your system gets
  207. broken in weeks later and sensitive data which you thought were
  208. encrypted and protected are retrieved and stolen from the swap device.
  209. To prevent this situation you should use 'Encrypt suspend image'.
  210. During suspend a temporary key is created and this key is used to
  211. encrypt the data written to disk. When, during resume, the data was
  212. read back into memory the temporary key is destroyed which simply
  213. means that all data written to disk during suspend are then
  214. inaccessible so they can't be stolen later on. The only thing that
  215. you must then take care of is that you call 'mkswap' for the swap
  216. partition used for suspend as early as possible during regular
  217. boot. This asserts that any temporary key from an oopsed suspend or
  218. from a failed or aborted resume is erased from the swap device.
  219. As a rule of thumb use encrypted swap to protect your data while your
  220. system is shut down or suspended. Additionally use the encrypted
  221. suspend image to prevent sensitive data from being stolen after
  222. resume.
  223. Q: Can I suspend to a swap file?
  224. A: Generally, yes, you can. However, it requires you to use the "resume=" and
  225. "resume_offset=" kernel command line parameters, so the resume from a swap file
  226. cannot be initiated from an initrd or initramfs image. See
  227. swsusp-and-swap-files.txt for details.
  228. Q: Is there a maximum system RAM size that is supported by swsusp?
  229. A: It should work okay with highmem.
  230. Q: Does swsusp (to disk) use only one swap partition or can it use
  231. multiple swap partitions (aggregate them into one logical space)?
  232. A: Only one swap partition, sorry.
  233. Q: If my application(s) causes lots of memory & swap space to be used
  234. (over half of the total system RAM), is it correct that it is likely
  235. to be useless to try to suspend to disk while that app is running?
  236. A: No, it should work okay, as long as your app does not mlock()
  237. it. Just prepare big enough swap partition.
  238. Q: What information is useful for debugging suspend-to-disk problems?
  239. A: Well, last messages on the screen are always useful. If something
  240. is broken, it is usually some kernel driver, therefore trying with as
  241. little as possible modules loaded helps a lot. I also prefer people to
  242. suspend from console, preferably without X running. Booting with
  243. init=/bin/bash, then swapon and starting suspend sequence manually
  244. usually does the trick. Then it is good idea to try with latest
  245. vanilla kernel.
  246. Q: How can distributions ship a swsusp-supporting kernel with modular
  247. disk drivers (especially SATA)?
  248. A: Well, it can be done, load the drivers, then do echo into
  249. /sys/power/resume file from initrd. Be sure not to mount
  250. anything, not even read-only mount, or you are going to lose your
  251. data.
  252. Q: How do I make suspend more verbose?
  253. A: If you want to see any non-error kernel messages on the virtual
  254. terminal the kernel switches to during suspend, you have to set the
  255. kernel console loglevel to at least 4 (KERN_WARNING), for example by
  256. doing
  257. # save the old loglevel
  258. read LOGLEVEL DUMMY < /proc/sys/kernel/printk
  259. # set the loglevel so we see the progress bar.
  260. # if the level is higher than needed, we leave it alone.
  261. if [ $LOGLEVEL -lt 5 ]; then
  262. echo 5 > /proc/sys/kernel/printk
  263. fi
  264. IMG_SZ=0
  265. read IMG_SZ < /sys/power/image_size
  266. echo -n disk > /sys/power/state
  267. RET=$?
  268. #
  269. # the logic here is:
  270. # if image_size > 0 (without kernel support, IMG_SZ will be zero),
  271. # then try again with image_size set to zero.
  272. if [ $RET -ne 0 -a $IMG_SZ -ne 0 ]; then # try again with minimal image size
  273. echo 0 > /sys/power/image_size
  274. echo -n disk > /sys/power/state
  275. RET=$?
  276. fi
  277. # restore previous loglevel
  278. echo $LOGLEVEL > /proc/sys/kernel/printk
  279. exit $RET
  280. Q: Is this true that if I have a mounted filesystem on a USB device and
  281. I suspend to disk, I can lose data unless the filesystem has been mounted
  282. with "sync"?
  283. A: That's right ... if you disconnect that device, you may lose data.
  284. In fact, even with "-o sync" you can lose data if your programs have
  285. information in buffers they haven't written out to a disk you disconnect,
  286. or if you disconnect before the device finished saving data you wrote.
  287. Software suspend normally powers down USB controllers, which is equivalent
  288. to disconnecting all USB devices attached to your system.
  289. Your system might well support low-power modes for its USB controllers
  290. while the system is asleep, maintaining the connection, using true sleep
  291. modes like "suspend-to-RAM" or "standby". (Don't write "disk" to the
  292. /sys/power/state file; write "standby" or "mem".) We've not seen any
  293. hardware that can use these modes through software suspend, although in
  294. theory some systems might support "platform" modes that won't break the
  295. USB connections.
  296. Remember that it's always a bad idea to unplug a disk drive containing a
  297. mounted filesystem. That's true even when your system is asleep! The
  298. safest thing is to unmount all filesystems on removable media (such USB,
  299. Firewire, CompactFlash, MMC, external SATA, or even IDE hotplug bays)
  300. before suspending; then remount them after resuming.
  301. There is a work-around for this problem. For more information, see
  302. Documentation/usb/persist.txt.
  303. Q: Can I suspend-to-disk using a swap partition under LVM?
  304. A: No. You can suspend successfully, but you'll not be able to
  305. resume. uswsusp should be able to work with LVM. See suspend.sf.net.
  306. Q: I upgraded the kernel from 2.6.15 to 2.6.16. Both kernels were
  307. compiled with the similar configuration files. Anyway I found that
  308. suspend to disk (and resume) is much slower on 2.6.16 compared to
  309. 2.6.15. Any idea for why that might happen or how can I speed it up?
  310. A: This is because the size of the suspend image is now greater than
  311. for 2.6.15 (by saving more data we can get more responsive system
  312. after resume).
  313. There's the /sys/power/image_size knob that controls the size of the
  314. image. If you set it to 0 (eg. by echo 0 > /sys/power/image_size as
  315. root), the 2.6.15 behavior should be restored. If it is still too
  316. slow, take a look at suspend.sf.net -- userland suspend is faster and
  317. supports LZF compression to speed it up further.