Based on kernel version 3.15.4. Page generated on 2014-07-07 09:04 EST.
1 Some warnings, first. 2 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.) 26 27 You need to append resume=/dev/your_swap_partition to kernel command 28 line. Then you suspend by 29 30 echo shutdown > /sys/power/disk; echo disk > /sys/power/state 31 32 . If you feel ACPI works pretty well on your system, you might try 33 34 echo platform > /sys/power/disk; echo disk > /sys/power/state 35 36 . If you would like to write hibernation image to swap and then suspend 37 to RAM (provided your platform supports it), you can try 38 39 echo suspend > /sys/power/disk; echo disk > /sys/power/state 40 41 . If you have SATA disks, you'll need recent kernels with SATA suspend 42 support. For suspend and resume to work, make sure your disk drivers 43 are built into kernel -- not modules. [There's way to make 44 suspend/resume with modular disk drivers, see FAQ, but you probably 45 should not do that.] 46 47 If you want to limit the suspend image size to N bytes, do 48 49 echo N > /sys/power/image_size 50 51 before suspend (it is limited to 500 MB by default). 52 53 . The resume process checks for the presence of the resume device, 54 if found, it then checks the contents for the hibernation image signature. 55 If both are found, it resumes the hibernation image. 56 57 . The resume process may be triggered in two ways: 58 1) During lateinit: If resume=/dev/your_swap_partition is specified on 59 the kernel command line, lateinit runs the resume process. If the 60 resume device has not been probed yet, the resume process fails and 61 bootup continues. 62 2) Manually from an initrd or initramfs: May be run from 63 the init script by using the /sys/power/resume file. It is vital 64 that this be done prior to remounting any filesystems (even as 65 read-only) otherwise data may be corrupted. 66 67 Article about goals and implementation of Software Suspend for Linux 68 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 69 Author: GÃ¡bor Kuti 70 Last revised: 2003-10-20 by Pavel Machek 71 72 Idea and goals to achieve 73 74 Nowadays it is common in several laptops that they have a suspend button. It 75 saves the state of the machine to a filesystem or to a partition and switches 76 to standby mode. Later resuming the machine the saved state is loaded back to 77 ram and the machine can continue its work. It has two real benefits. First we 78 save ourselves the time machine goes down and later boots up, energy costs 79 are real high when running from batteries. The other gain is that we don't have to 80 interrupt our programs so processes that are calculating something for a long 81 time shouldn't need to be written interruptible. 82 83 swsusp saves the state of the machine into active swaps and then reboots or 84 powerdowns. You must explicitly specify the swap partition to resume from with 85 ``resume='' kernel option. If signature is found it loads and restores saved 86 state. If the option ``noresume'' is specified as a boot parameter, it skips 87 the resuming. If the option ``hibernate=nocompress'' is specified as a boot 88 parameter, it saves hibernation image without compression. 89 90 In the meantime while the system is suspended you should not add/remove any 91 of the hardware, write to the filesystems, etc. 92 93 Sleep states summary 94 ==================== 95 96 There are three different interfaces you can use, /proc/acpi should 97 work like this: 98 99 In a really perfect world: 100 echo 1 > /proc/acpi/sleep # for standby 101 echo 2 > /proc/acpi/sleep # for suspend to ram 102 echo 3 > /proc/acpi/sleep # for suspend to ram, but with more power conservative 103 echo 4 > /proc/acpi/sleep # for suspend to disk 104 echo 5 > /proc/acpi/sleep # for shutdown unfriendly the system 105 106 and perhaps 107 echo 4b > /proc/acpi/sleep # for suspend to disk via s4bios 108 109 Frequently Asked Questions 110 ========================== 111 112 Q: well, suspending a server is IMHO a really stupid thing, 113 but... (Diego Zuccato): 114 115 A: You bought new UPS for your server. How do you install it without 116 bringing machine down? Suspend to disk, rearrange power cables, 117 resume. 118 119 You have your server on UPS. Power died, and UPS is indicating 30 120 seconds to failure. What do you do? Suspend to disk. 121 122 123 Q: Maybe I'm missing something, but why don't the regular I/O paths work? 124 125 A: We do use the regular I/O paths. However we cannot restore the data 126 to its original location as we load it. That would create an 127 inconsistent kernel state which would certainly result in an oops. 128 Instead, we load the image into unused memory and then atomically copy 129 it back to it original location. This implies, of course, a maximum 130 image size of half the amount of memory. 131 132 There are two solutions to this: 133 134 * require half of memory to be free during suspend. That way you can 135 read "new" data onto free spots, then cli and copy 136 137 * assume we had special "polling" ide driver that only uses memory 138 between 0-640KB. That way, I'd have to make sure that 0-640KB is free 139 during suspending, but otherwise it would work... 140 141 suspend2 shares this fundamental limitation, but does not include user 142 data and disk caches into "used memory" by saving them in 143 advance. That means that the limitation goes away in practice. 144 145 Q: Does linux support ACPI S4? 146 147 A: Yes. That's what echo platform > /sys/power/disk does. 148 149 Q: What is 'suspend2'? 150 151 A: suspend2 is 'Software Suspend 2', a forked implementation of 152 suspend-to-disk which is available as separate patches for 2.4 and 2.6 153 kernels from swsusp.sourceforge.net. It includes support for SMP, 4GB 154 highmem and preemption. It also has a extensible architecture that 155 allows for arbitrary transformations on the image (compression, 156 encryption) and arbitrary backends for writing the image (eg to swap 157 or an NFS share[Work In Progress]). Questions regarding suspend2 158 should be sent to the mailing list available through the suspend2 159 website, and not to the Linux Kernel Mailing List. We are working 160 toward merging suspend2 into the mainline kernel. 161 162 Q: What is the freezing of tasks and why are we using it? 163 164 A: The freezing of tasks is a mechanism by which user space processes and some 165 kernel threads are controlled during hibernation or system-wide suspend (on some 166 architectures). See freezing-of-tasks.txt for details. 167 168 Q: What is the difference between "platform" and "shutdown"? 169 170 A: 171 172 shutdown: save state in linux, then tell bios to powerdown 173 174 platform: save state in linux, then tell bios to powerdown and blink 175 "suspended led" 176 177 "platform" is actually right thing to do where supported, but 178 "shutdown" is most reliable (except on ACPI systems). 179 180 Q: I do not understand why you have such strong objections to idea of 181 selective suspend. 182 183 A: Do selective suspend during runtime power management, that's okay. But 184 it's useless for suspend-to-disk. (And I do not see how you could use 185 it for suspend-to-ram, I hope you do not want that). 186 187 Lets see, so you suggest to 188 189 * SUSPEND all but swap device and parents 190 * Snapshot 191 * Write image to disk 192 * SUSPEND swap device and parents 193 * Powerdown 194 195 Oh no, that does not work, if swap device or its parents uses DMA, 196 you've corrupted data. You'd have to do 197 198 * SUSPEND all but swap device and parents 199 * FREEZE swap device and parents 200 * Snapshot 201 * UNFREEZE swap device and parents 202 * Write 203 * SUSPEND swap device and parents 204 205 Which means that you still need that FREEZE state, and you get more 206 complicated code. (And I have not yet introduce details like system 207 devices). 208 209 Q: There don't seem to be any generally useful behavioral 210 distinctions between SUSPEND and FREEZE. 211 212 A: Doing SUSPEND when you are asked to do FREEZE is always correct, 213 but it may be unnecessarily slow. If you want your driver to stay simple, 214 slowness may not matter to you. It can always be fixed later. 215 216 For devices like disk it does matter, you do not want to spindown for 217 FREEZE. 218 219 Q: After resuming, system is paging heavily, leading to very bad interactivity. 220 221 A: Try running 222 223 cat `cat /proc/[0-9]*/maps | grep / | sed 's:.* /:/:' | sort -u` > /dev/null 224 225 after resume. swapoff -a; swapon -a may also be useful. 226 227 Q: What happens to devices during swsusp? They seem to be resumed 228 during system suspend? 229 230 A: That's correct. We need to resume them if we want to write image to 231 disk. Whole sequence goes like 232 233 Suspend part 234 ~~~~~~~~~~~~ 235 running system, user asks for suspend-to-disk 236 237 user processes are stopped 238 239 suspend(PMSG_FREEZE): devices are frozen so that they don't interfere 240 with state snapshot 241 242 state snapshot: copy of whole used memory is taken with interrupts disabled 243 244 resume(): devices are woken up so that we can write image to swap 245 246 write image to swap 247 248 suspend(PMSG_SUSPEND): suspend devices so that we can power off 249 250 turn the power off 251 252 Resume part 253 ~~~~~~~~~~~ 254 (is actually pretty similar) 255 256 running system, user asks for suspend-to-disk 257 258 user processes are stopped (in common case there are none, but with resume-from-initrd, no one knows) 259 260 read image from disk 261 262 suspend(PMSG_FREEZE): devices are frozen so that they don't interfere 263 with image restoration 264 265 image restoration: rewrite memory with image 266 267 resume(): devices are woken up so that system can continue 268 269 thaw all user processes 270 271 Q: What is this 'Encrypt suspend image' for? 272 273 A: First of all: it is not a replacement for dm-crypt encrypted swap. 274 It cannot protect your computer while it is suspended. Instead it does 275 protect from leaking sensitive data after resume from suspend. 276 277 Think of the following: you suspend while an application is running 278 that keeps sensitive data in memory. The application itself prevents 279 the data from being swapped out. Suspend, however, must write these 280 data to swap to be able to resume later on. Without suspend encryption 281 your sensitive data are then stored in plaintext on disk. This means 282 that after resume your sensitive data are accessible to all 283 applications having direct access to the swap device which was used 284 for suspend. If you don't need swap after resume these data can remain 285 on disk virtually forever. Thus it can happen that your system gets 286 broken in weeks later and sensitive data which you thought were 287 encrypted and protected are retrieved and stolen from the swap device. 288 To prevent this situation you should use 'Encrypt suspend image'. 289 290 During suspend a temporary key is created and this key is used to 291 encrypt the data written to disk. When, during resume, the data was 292 read back into memory the temporary key is destroyed which simply 293 means that all data written to disk during suspend are then 294 inaccessible so they can't be stolen later on. The only thing that 295 you must then take care of is that you call 'mkswap' for the swap 296 partition used for suspend as early as possible during regular 297 boot. This asserts that any temporary key from an oopsed suspend or 298 from a failed or aborted resume is erased from the swap device. 299 300 As a rule of thumb use encrypted swap to protect your data while your 301 system is shut down or suspended. Additionally use the encrypted 302 suspend image to prevent sensitive data from being stolen after 303 resume. 304 305 Q: Can I suspend to a swap file? 306 307 A: Generally, yes, you can. However, it requires you to use the "resume=" and 308 "resume_offset=" kernel command line parameters, so the resume from a swap file 309 cannot be initiated from an initrd or initramfs image. See 310 swsusp-and-swap-files.txt for details. 311 312 Q: Is there a maximum system RAM size that is supported by swsusp? 313 314 A: It should work okay with highmem. 315 316 Q: Does swsusp (to disk) use only one swap partition or can it use 317 multiple swap partitions (aggregate them into one logical space)? 318 319 A: Only one swap partition, sorry. 320 321 Q: If my application(s) causes lots of memory & swap space to be used 322 (over half of the total system RAM), is it correct that it is likely 323 to be useless to try to suspend to disk while that app is running? 324 325 A: No, it should work okay, as long as your app does not mlock() 326 it. Just prepare big enough swap partition. 327 328 Q: What information is useful for debugging suspend-to-disk problems? 329 330 A: Well, last messages on the screen are always useful. If something 331 is broken, it is usually some kernel driver, therefore trying with as 332 little as possible modules loaded helps a lot. I also prefer people to 333 suspend from console, preferably without X running. Booting with 334 init=/bin/bash, then swapon and starting suspend sequence manually 335 usually does the trick. Then it is good idea to try with latest 336 vanilla kernel. 337 338 Q: How can distributions ship a swsusp-supporting kernel with modular 339 disk drivers (especially SATA)? 340 341 A: Well, it can be done, load the drivers, then do echo into 342 /sys/power/resume file from initrd. Be sure not to mount 343 anything, not even read-only mount, or you are going to lose your 344 data. 345 346 Q: How do I make suspend more verbose? 347 348 A: If you want to see any non-error kernel messages on the virtual 349 terminal the kernel switches to during suspend, you have to set the 350 kernel console loglevel to at least 4 (KERN_WARNING), for example by 351 doing 352 353 # save the old loglevel 354 read LOGLEVEL DUMMY < /proc/sys/kernel/printk 355 # set the loglevel so we see the progress bar. 356 # if the level is higher than needed, we leave it alone. 357 if [ $LOGLEVEL -lt 5 ]; then 358 echo 5 > /proc/sys/kernel/printk 359 fi 360 361 IMG_SZ=0 362 read IMG_SZ < /sys/power/image_size 363 echo -n disk > /sys/power/state 364 RET=$? 365 # 366 # the logic here is: 367 # if image_size > 0 (without kernel support, IMG_SZ will be zero), 368 # then try again with image_size set to zero. 369 if [ $RET -ne 0 -a $IMG_SZ -ne 0 ]; then # try again with minimal image size 370 echo 0 > /sys/power/image_size 371 echo -n disk > /sys/power/state 372 RET=$? 373 fi 374 375 # restore previous loglevel 376 echo $LOGLEVEL > /proc/sys/kernel/printk 377 exit $RET 378 379 Q: Is this true that if I have a mounted filesystem on a USB device and 380 I suspend to disk, I can lose data unless the filesystem has been mounted 381 with "sync"? 382 383 A: That's right ... if you disconnect that device, you may lose data. 384 In fact, even with "-o sync" you can lose data if your programs have 385 information in buffers they haven't written out to a disk you disconnect, 386 or if you disconnect before the device finished saving data you wrote. 387 388 Software suspend normally powers down USB controllers, which is equivalent 389 to disconnecting all USB devices attached to your system. 390 391 Your system might well support low-power modes for its USB controllers 392 while the system is asleep, maintaining the connection, using true sleep 393 modes like "suspend-to-RAM" or "standby". (Don't write "disk" to the 394 /sys/power/state file; write "standby" or "mem".) We've not seen any 395 hardware that can use these modes through software suspend, although in 396 theory some systems might support "platform" modes that won't break the 397 USB connections. 398 399 Remember that it's always a bad idea to unplug a disk drive containing a 400 mounted filesystem. That's true even when your system is asleep! The 401 safest thing is to unmount all filesystems on removable media (such USB, 402 Firewire, CompactFlash, MMC, external SATA, or even IDE hotplug bays) 403 before suspending; then remount them after resuming. 404 405 There is a work-around for this problem. For more information, see 406 Documentation/usb/persist.txt. 407 408 Q: Can I suspend-to-disk using a swap partition under LVM? 409 410 A: No. You can suspend successfully, but you'll not be able to 411 resume. uswsusp should be able to work with LVM. See suspend.sf.net. 412 413 Q: I upgraded the kernel from 2.6.15 to 2.6.16. Both kernels were 414 compiled with the similar configuration files. Anyway I found that 415 suspend to disk (and resume) is much slower on 2.6.16 compared to 416 2.6.15. Any idea for why that might happen or how can I speed it up? 417 418 A: This is because the size of the suspend image is now greater than 419 for 2.6.15 (by saving more data we can get more responsive system 420 after resume). 421 422 There's the /sys/power/image_size knob that controls the size of the 423 image. If you set it to 0 (eg. by echo 0 > /sys/power/image_size as 424 root), the 2.6.15 behavior should be restored. If it is still too 425 slow, take a look at suspend.sf.net -- userland suspend is faster and 426 supports LZF compression to speed it up further.