Talk:APM vs. ACPI

Just wondering...what exactly do headings in this table mean?

• Blank = `Ability to turn off the LCD'
• Sleep = `Ability to (under at least some circumstances, or, `in theory' perform a SUSPEND-TO-RAM)'
• Hibernate = `Ability to perform a SUSPEND-TO-DISK via the Phoenix NoteBIOS'
• Battery = `Ability to get statistics on remaining battery life'

If these are correct, I'd be interested in knowing how hibernation (as described) works under ACPI.

- chris

---

Hei,

i just added explanations of how it is meant to the page. Do you want to know how to make it work under ACPI? Or under APM using the BIOS internal hibernation support? AFAIK in the T4x series models everything works both ways.

Wyrfel 02:10, 15 Mar 2005 (CET)

---

Ah. I was wondering if there was a way to get the internal (BIOS) hibernation to work when ACPI is enabled. If this is what you mean, I'm all for an explanation as it would be the best of both worlds, and too good to be true all at once. If you mean SWSUSP, (although I probably won't use it for reliability reasons), an explaination would be great as well.

Thanks,

- chris

---

There used to be a "S4Bios" state in /proc/acpi/sleep, which is exactly as you describe. My laptop (770x) offers it, but it dissapeared out of the ACPI support at somepoint. Definitely back in Linux 2.4.x it was there, and would trigger the BIOS to hibernate the laptop. Linux didn't make it out of suspend but that was 2.4.x ACPI, so you didn't really expect it to. google://s4bios gives a few things that should tell you more, and a patch. Personally though I could try to get this to work, I'm sticking with swsup2, which is probably what everyone means when they say they're hibernating using ACPI.

lentinj

Thanks for the pointer...I think I'll go ahead and make an account here. This site is pretty helpful.

---

Hrm...guess this machine (T40p) doesn't support s4bios...only S4. I'll give it a try and see how many comprimises it involves. In any event, judging by this ( http://seclists.org/lists/linux-kernel/2005/Mar/2287.html ) it appears that the person who put the support into 2.6 ( http://lwn.net/Articles/2428/ ) is pretty much ready to take it back out, citing instability, and a non-existant user base. If you ask me, there're probably no users because the APM equivalent of s4bios works flawlessly but oh well...

ct

---

See the kernel docs... power/interface.txt. You can do # echo firmware > /sys/power/disk and the next time you suspend-to-disk it will let the BIOS handle it. I didn't try it and never heard of anyone doing so. I don't see the advantages either, swsusp works pretty well for me (last time i used it at least ;-) ). Wyrfel 20:13, 15 Mar 2005 (CET)

---

Agreed,

My stance against swsusp was probably due to my initial insistence on using the binary ATI drivers. At least while I was using them, they did not play well with ACPI suspend (-to-disk or -to-ram); or APM suspend-to-ram. Suspend to disk seemed like magic (in that the BIOS appeared to just pull the rug right out from beneath Linux without it knowing) and only worked while APM was enabled, so I settled on it. I since stopped using those drivers in favour of the much more reliable OpenSource ones, and am just now realizing that that was the only reason why I was still using APM.

SWSUSP appears to be working flawlessly (after the surge of paging-upon-resume), so I think I'll go with ACPI. At last...I'm able to put numbers to these scorching temperatures, and truely know how long I have before the battery's charged. And then there's ibm-acpi...

<sigh>, and thanks for the help,

--chris 21:36, 15 Mar 2005 (CET)

Hi, I own a ThinkPad X40 and Suspend to RAM works under ACPI. You have to provide "acpi_sleep=s3_bios" on the kernel command line, though or the LCD will be dead on resume. Also I have the ability to turn off the LCD by issuing the command:

echo video_switch > /proc/acpi/ibm/video

Not sure if this is what is meant with that point. I haven't tried swsusp yet, but that's on my TODO list. Battery statistics (that is remaining capacity in % and approx. runtime) are available over the acpi or acpitool commands or by looking at /proc/acpi/battery/BAT0/*

--Daniel Willmann 15:09, 24 Mar 2005 (CET)