ia64/linux-2.6.18-xen.hg

view Documentation/power/swsusp.txt @ 524:7f8b544237bf

netfront: Allow netfront in domain 0.

This is useful if your physical network device is in a utility domain.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
author Keir Fraser <keir.fraser@citrix.com>
date Tue Apr 15 15:18:58 2008 +0100 (2008-04-15)
parents 831230e53067
children
line source
1 Some warnings, first.
3 * BIG FAT WARNING *********************************************************
4 *
5 * If you touch anything on disk between suspend and resume...
6 * ...kiss your data goodbye.
7 *
8 * If you do resume from initrd after your filesystems are mounted...
9 * ...bye bye root partition.
10 * [this is actually same case as above]
11 *
12 * If you have unsupported (*) devices using DMA, you may have some
13 * problems. If your disk driver does not support suspend... (IDE does),
14 * it may cause some problems, too. If you change kernel command line
15 * between suspend and resume, it may do something wrong. If you change
16 * your hardware while system is suspended... well, it was not good idea;
17 * but it will probably only crash.
18 *
19 * (*) suspend/resume support is needed to make it safe.
20 *
21 * If you have any filesystems on USB devices mounted before software suspend,
22 * they won't be accessible after resume and you may lose data, as though
23 * you have unplugged the USB devices with mounted filesystems on them;
24 * see the FAQ below for details. (This is not true for more traditional
25 * power states like "standby", which normally don't turn USB off.)
27 You need to append resume=/dev/your_swap_partition to kernel command
28 line. Then you suspend by
30 echo shutdown > /sys/power/disk; echo disk > /sys/power/state
32 . If you feel ACPI works pretty well on your system, you might try
34 echo platform > /sys/power/disk; echo disk > /sys/power/state
36 . If you have SATA disks, you'll need recent kernels with SATA suspend
37 support. For suspend and resume to work, make sure your disk drivers
38 are built into kernel -- not modules. [There's way to make
39 suspend/resume with modular disk drivers, see FAQ, but you probably
40 should not do that.]
42 If you want to limit the suspend image size to N bytes, do
44 echo N > /sys/power/image_size
46 before suspend (it is limited to 500 MB by default).
49 Article about goals and implementation of Software Suspend for Linux
50 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
51 Author: G‚ábor Kuti
52 Last revised: 2003-10-20 by Pavel Machek
54 Idea and goals to achieve
56 Nowadays it is common in several laptops that they have a suspend button. It
57 saves the state of the machine to a filesystem or to a partition and switches
58 to standby mode. Later resuming the machine the saved state is loaded back to
59 ram and the machine can continue its work. It has two real benefits. First we
60 save ourselves the time machine goes down and later boots up, energy costs
61 are real high when running from batteries. The other gain is that we don't have to
62 interrupt our programs so processes that are calculating something for a long
63 time shouldn't need to be written interruptible.
65 swsusp saves the state of the machine into active swaps and then reboots or
66 powerdowns. You must explicitly specify the swap partition to resume from with
67 ``resume='' kernel option. If signature is found it loads and restores saved
68 state. If the option ``noresume'' is specified as a boot parameter, it skips
69 the resuming.
71 In the meantime while the system is suspended you should not add/remove any
72 of the hardware, write to the filesystems, etc.
74 Sleep states summary
75 ====================
77 There are three different interfaces you can use, /proc/acpi should
78 work like this:
80 In a really perfect world:
81 echo 1 > /proc/acpi/sleep # for standby
82 echo 2 > /proc/acpi/sleep # for suspend to ram
83 echo 3 > /proc/acpi/sleep # for suspend to ram, but with more power conservative
84 echo 4 > /proc/acpi/sleep # for suspend to disk
85 echo 5 > /proc/acpi/sleep # for shutdown unfriendly the system
87 and perhaps
88 echo 4b > /proc/acpi/sleep # for suspend to disk via s4bios
90 Frequently Asked Questions
91 ==========================
93 Q: well, suspending a server is IMHO a really stupid thing,
94 but... (Diego Zuccato):
96 A: You bought new UPS for your server. How do you install it without
97 bringing machine down? Suspend to disk, rearrange power cables,
98 resume.
100 You have your server on UPS. Power died, and UPS is indicating 30
101 seconds to failure. What do you do? Suspend to disk.
104 Q: Maybe I'm missing something, but why don't the regular I/O paths work?
106 A: We do use the regular I/O paths. However we cannot restore the data
107 to its original location as we load it. That would create an
108 inconsistent kernel state which would certainly result in an oops.
109 Instead, we load the image into unused memory and then atomically copy
110 it back to it original location. This implies, of course, a maximum
111 image size of half the amount of memory.
113 There are two solutions to this:
115 * require half of memory to be free during suspend. That way you can
116 read "new" data onto free spots, then cli and copy
118 * assume we had special "polling" ide driver that only uses memory
119 between 0-640KB. That way, I'd have to make sure that 0-640KB is free
120 during suspending, but otherwise it would work...
122 suspend2 shares this fundamental limitation, but does not include user
123 data and disk caches into "used memory" by saving them in
124 advance. That means that the limitation goes away in practice.
126 Q: Does linux support ACPI S4?
128 A: Yes. That's what echo platform > /sys/power/disk does.
130 Q: What is 'suspend2'?
132 A: suspend2 is 'Software Suspend 2', a forked implementation of
133 suspend-to-disk which is available as separate patches for 2.4 and 2.6
134 kernels from swsusp.sourceforge.net. It includes support for SMP, 4GB
135 highmem and preemption. It also has a extensible architecture that
136 allows for arbitrary transformations on the image (compression,
137 encryption) and arbitrary backends for writing the image (eg to swap
138 or an NFS share[Work In Progress]). Questions regarding suspend2
139 should be sent to the mailing list available through the suspend2
140 website, and not to the Linux Kernel Mailing List. We are working
141 toward merging suspend2 into the mainline kernel.
143 Q: A kernel thread must voluntarily freeze itself (call 'refrigerator').
144 I found some kernel threads that don't do it, and they don't freeze
145 so the system can't sleep. Is this a known behavior?
147 A: All such kernel threads need to be fixed, one by one. Select the
148 place where the thread is safe to be frozen (no kernel semaphores
149 should be held at that point and it must be safe to sleep there), and
150 add:
152 try_to_freeze();
154 If the thread is needed for writing the image to storage, you should
155 instead set the PF_NOFREEZE process flag when creating the thread (and
156 be very carefull).
159 Q: What is the difference between between "platform", "shutdown" and
160 "firmware" in /sys/power/disk?
162 A:
164 shutdown: save state in linux, then tell bios to powerdown
166 platform: save state in linux, then tell bios to powerdown and blink
167 "suspended led"
169 firmware: tell bios to save state itself [needs BIOS-specific suspend
170 partition, and has very little to do with swsusp]
172 "platform" is actually right thing to do, but "shutdown" is most
173 reliable.
175 Q: I do not understand why you have such strong objections to idea of
176 selective suspend.
178 A: Do selective suspend during runtime power managment, that's okay. But
179 its useless for suspend-to-disk. (And I do not see how you could use
180 it for suspend-to-ram, I hope you do not want that).
182 Lets see, so you suggest to
184 * SUSPEND all but swap device and parents
185 * Snapshot
186 * Write image to disk
187 * SUSPEND swap device and parents
188 * Powerdown
190 Oh no, that does not work, if swap device or its parents uses DMA,
191 you've corrupted data. You'd have to do
193 * SUSPEND all but swap device and parents
194 * FREEZE swap device and parents
195 * Snapshot
196 * UNFREEZE swap device and parents
197 * Write
198 * SUSPEND swap device and parents
200 Which means that you still need that FREEZE state, and you get more
201 complicated code. (And I have not yet introduce details like system
202 devices).
204 Q: There don't seem to be any generally useful behavioral
205 distinctions between SUSPEND and FREEZE.
207 A: Doing SUSPEND when you are asked to do FREEZE is always correct,
208 but it may be unneccessarily slow. If you want your driver to stay simple,
209 slowness may not matter to you. It can always be fixed later.
211 For devices like disk it does matter, you do not want to spindown for
212 FREEZE.
214 Q: After resuming, system is paging heavilly, leading to very bad interactivity.
216 A: Try running
218 cat `cat /proc/[0-9]*/maps | grep / | sed 's:.* /:/:' | sort -u` > /dev/null
220 after resume. swapoff -a; swapon -a may also be useful.
222 Q: What happens to devices during swsusp? They seem to be resumed
223 during system suspend?
225 A: That's correct. We need to resume them if we want to write image to
226 disk. Whole sequence goes like
228 Suspend part
229 ~~~~~~~~~~~~
230 running system, user asks for suspend-to-disk
232 user processes are stopped
234 suspend(PMSG_FREEZE): devices are frozen so that they don't interfere
235 with state snapshot
237 state snapshot: copy of whole used memory is taken with interrupts disabled
239 resume(): devices are woken up so that we can write image to swap
241 write image to swap
243 suspend(PMSG_SUSPEND): suspend devices so that we can power off
245 turn the power off
247 Resume part
248 ~~~~~~~~~~~
249 (is actually pretty similar)
251 running system, user asks for suspend-to-disk
253 user processes are stopped (in common case there are none, but with resume-from-initrd, noone knows)
255 read image from disk
257 suspend(PMSG_FREEZE): devices are frozen so that they don't interfere
258 with image restoration
260 image restoration: rewrite memory with image
262 resume(): devices are woken up so that system can continue
264 thaw all user processes
266 Q: What is this 'Encrypt suspend image' for?
268 A: First of all: it is not a replacement for dm-crypt encrypted swap.
269 It cannot protect your computer while it is suspended. Instead it does
270 protect from leaking sensitive data after resume from suspend.
272 Think of the following: you suspend while an application is running
273 that keeps sensitive data in memory. The application itself prevents
274 the data from being swapped out. Suspend, however, must write these
275 data to swap to be able to resume later on. Without suspend encryption
276 your sensitive data are then stored in plaintext on disk. This means
277 that after resume your sensitive data are accessible to all
278 applications having direct access to the swap device which was used
279 for suspend. If you don't need swap after resume these data can remain
280 on disk virtually forever. Thus it can happen that your system gets
281 broken in weeks later and sensitive data which you thought were
282 encrypted and protected are retrieved and stolen from the swap device.
283 To prevent this situation you should use 'Encrypt suspend image'.
285 During suspend a temporary key is created and this key is used to
286 encrypt the data written to disk. When, during resume, the data was
287 read back into memory the temporary key is destroyed which simply
288 means that all data written to disk during suspend are then
289 inaccessible so they can't be stolen later on. The only thing that
290 you must then take care of is that you call 'mkswap' for the swap
291 partition used for suspend as early as possible during regular
292 boot. This asserts that any temporary key from an oopsed suspend or
293 from a failed or aborted resume is erased from the swap device.
295 As a rule of thumb use encrypted swap to protect your data while your
296 system is shut down or suspended. Additionally use the encrypted
297 suspend image to prevent sensitive data from being stolen after
298 resume.
300 Q: Why can't we suspend to a swap file?
302 A: Because accessing swap file needs the filesystem mounted, and
303 filesystem might do something wrong (like replaying the journal)
304 during mount.
306 There are few ways to get that fixed:
308 1) Probably could be solved by modifying every filesystem to support
309 some kind of "really read-only!" option. Patches welcome.
311 2) suspend2 gets around that by storing absolute positions in on-disk
312 image (and blocksize), with resume parameter pointing directly to
313 suspend header.
315 Q: Is there a maximum system RAM size that is supported by swsusp?
317 A: It should work okay with highmem.
319 Q: Does swsusp (to disk) use only one swap partition or can it use
320 multiple swap partitions (aggregate them into one logical space)?
322 A: Only one swap partition, sorry.
324 Q: If my application(s) causes lots of memory & swap space to be used
325 (over half of the total system RAM), is it correct that it is likely
326 to be useless to try to suspend to disk while that app is running?
328 A: No, it should work okay, as long as your app does not mlock()
329 it. Just prepare big enough swap partition.
331 Q: What information is useful for debugging suspend-to-disk problems?
333 A: Well, last messages on the screen are always useful. If something
334 is broken, it is usually some kernel driver, therefore trying with as
335 little as possible modules loaded helps a lot. I also prefer people to
336 suspend from console, preferably without X running. Booting with
337 init=/bin/bash, then swapon and starting suspend sequence manually
338 usually does the trick. Then it is good idea to try with latest
339 vanilla kernel.
341 Q: How can distributions ship a swsusp-supporting kernel with modular
342 disk drivers (especially SATA)?
344 A: Well, it can be done, load the drivers, then do echo into
345 /sys/power/disk/resume file from initrd. Be sure not to mount
346 anything, not even read-only mount, or you are going to lose your
347 data.
349 Q: How do I make suspend more verbose?
351 A: If you want to see any non-error kernel messages on the virtual
352 terminal the kernel switches to during suspend, you have to set the
353 kernel console loglevel to at least 4 (KERN_WARNING), for example by
354 doing
356 # save the old loglevel
357 read LOGLEVEL DUMMY < /proc/sys/kernel/printk
358 # set the loglevel so we see the progress bar.
359 # if the level is higher than needed, we leave it alone.
360 if [ $LOGLEVEL -lt 5 ]; then
361 echo 5 > /proc/sys/kernel/printk
362 fi
364 IMG_SZ=0
365 read IMG_SZ < /sys/power/image_size
366 echo -n disk > /sys/power/state
367 RET=$?
368 #
369 # the logic here is:
370 # if image_size > 0 (without kernel support, IMG_SZ will be zero),
371 # then try again with image_size set to zero.
372 if [ $RET -ne 0 -a $IMG_SZ -ne 0 ]; then # try again with minimal image size
373 echo 0 > /sys/power/image_size
374 echo -n disk > /sys/power/state
375 RET=$?
376 fi
378 # restore previous loglevel
379 echo $LOGLEVEL > /proc/sys/kernel/printk
380 exit $RET
382 Q: Is this true that if I have a mounted filesystem on a USB device and
383 I suspend to disk, I can lose data unless the filesystem has been mounted
384 with "sync"?
386 A: That's right ... if you disconnect that device, you may lose data.
387 In fact, even with "-o sync" you can lose data if your programs have
388 information in buffers they haven't written out to a disk you disconnect,
389 or if you disconnect before the device finished saving data you wrote.
391 Software suspend normally powers down USB controllers, which is equivalent
392 to disconnecting all USB devices attached to your system.
394 Your system might well support low-power modes for its USB controllers
395 while the system is asleep, maintaining the connection, using true sleep
396 modes like "suspend-to-RAM" or "standby". (Don't write "disk" to the
397 /sys/power/state file; write "standby" or "mem".) We've not seen any
398 hardware that can use these modes through software suspend, although in
399 theory some systems might support "platform" or "firmware" modes that
400 won't break the USB connections.
402 Remember that it's always a bad idea to unplug a disk drive containing a
403 mounted filesystem. That's true even when your system is asleep! The
404 safest thing is to unmount all filesystems on removable media (such USB,
405 Firewire, CompactFlash, MMC, external SATA, or even IDE hotplug bays)
406 before suspending; then remount them after resuming.
408 Q: I upgraded the kernel from 2.6.15 to 2.6.16. Both kernels were
409 compiled with the similar configuration files. Anyway I found that
410 suspend to disk (and resume) is much slower on 2.6.16 compared to
411 2.6.15. Any idea for why that might happen or how can I speed it up?
413 A: This is because the size of the suspend image is now greater than
414 for 2.6.15 (by saving more data we can get more responsive system
415 after resume).
417 There's the /sys/power/image_size knob that controls the size of the
418 image. If you set it to 0 (eg. by echo 0 > /sys/power/image_size as
419 root), the 2.6.15 behavior should be restored. If it is still too
420 slow, take a look at suspend.sf.net -- userland suspend is faster and
421 supports LZF compression to speed it up further.