Bug 43169 - SV266A with VIA KT266A: suspend to RAM not available
Summary: SV266A with VIA KT266A: suspend to RAM not available
Status: CLOSED INVALID
Alias: None
Product: ACPI
Classification: Unclassified
Component: Power-Sleep-Wake (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: acpi_power-sleep-wake
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-04-27 21:39 UTC by Paul Menzel
Modified: 2013-06-03 06:52 UTC (History)
6 users (show)

See Also:
Kernel Version: 2.6.32
Subsystem:
Regression: No
Bisected commit-id:


Attachments
Output of dmidecode (10.39 KB, text/plain)
2012-04-27 21:39 UTC, Paul Menzel
Details
Output of `biosdecode` (1013 bytes, text/plain)
2012-04-27 21:41 UTC, Paul Menzel
Details

Description Paul Menzel 2012-04-27 21:39:32 UTC
Created attachment 73114 [details]
Output of dmidecode

The option suspend to RAM is not available on the motherboard Syntax SV266A
with a VIA KT266A(?) chipset [1]. Debian Squeeze/stable running Linux kernel 2.6.32-41squeeze2 is used.

        $ echo mem | sudo tee /sys/power/state
        mem
        tee: /sys/power/state: No such device
        $ ls /sys/power/
        disk        image_size  pm_test     resume      state
        $ ls /sys/power/state
        /sys/power/state
        $ sudo su
        # echo mem > /sys/power/state
        bash: echo: Schreibfehler: Kein passendes Gerät gefunden.

I looked around in the BIOS and the option *ACPI functions* is set to
*Enabled*. There is also an option *Suspend Mode* which I can set to
*Disable* or to a number of minutes. But I guess I do not want the BIOS
to suspend my system but using the operating system for that task.

Is Linux looking at the ACPI tables and they are wrong in this case? Can
I force suspend to RAM somehow to test it?

Please tell me if I can provide more information.

There is also a report for this issue in the Debian BTS [2].

[1] http://lists.freedesktop.org/archives/pm-utils/2010-June/002176.html
[2] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=670634
Comment 1 Paul Menzel 2012-04-27 21:41:36 UTC
Created attachment 73115 [details]
Output of `biosdecode`
Comment 2 Lan Tianyu 2012-04-28 06:57:50 UTC
Please check whether the issuse is still existing in the newest upstream kernel.
Comment 3 Alan 2012-05-12 00:49:52 UTC
Thats also a very old board - does the BIOS ACPI even have the needed level of support ?
Comment 4 fpgahardwareengineer 2012-09-29 10:52:37 UTC
Hi Paul,

Can you burn an older version of Fedora (i.e., Fedora 8 should be okay) and boot it from CD/DVD?
When Fedora installer starts up, it does dump quite a bit of information about the system.
There you can see the supported ACPI states.
It is a good diagnostic tool, in my opinion.
If you see a message that looks like "ACPI (Sx Sx Sx Sx)" and you don't see S3 included in the list, then the mainboard does not support Suspend to RAM or Suspend to RAM is not activated in the BIOS setup.
If you don't see a "Suspend to RAM" or "ACPI State" in your BIOS setup's Power section, then I will doubt that you will be able to perform Suspend to RAM with that particular mainboard.
Looking at your debian.org post's lspci information you uploaded, it looks like your mainboard was manufactured by Elite Computer Systems or ECS. 
Looking at the lspci, I saw two devices having an ECS PCI Subvendor ID.
From my own experience of dealing with ECS mainboards, ECS does seem to want to support ACPI S3 State (STR or Suspend To RAM) with mainboards containing SiS or VIA Technologies chipsets.
For example, I have ECS P6S5A (SiS 635), K7S5A Pro (SiS 735), 755-A2 (SiS 755/964), and KT600-A (VIA Technologies KT600/8237) and none of these boards support ACPI S3 State despite the chipsets themselves support ACPI S3 State at the hardware level.
I also have ECS P4M800PRO-M (VIA Technologies PM800/8237) and 761GX-M754 (SiS 761GX/964) mainboards released around 2005, and they identify as supporting ACPI S3 State, but they don't ever enter ACPI S3 State in Linux.
On the other hand, ASUS has sold number of mainboards with SiS and VIA Technologies chipsets, and they have supported ACPI S3 State as early as 2002. (i.e., TUV4X released around 2001/2002 supports STR, although it will always perform Video rePOST.)
ASUS P4S8X mainboard (SiS 648/963 chipset) and P4S8X-MX mainboard (SiS 661GX/964 chipset) support ACPI S3 State.
As you might see, ECS and ASUS use the same southbridge chip (SiS 964), but ECS doesn't at all (755-A2) or is buggy (M761GX-M754) with ACPI S3 State support, but ASUS does diligently supports it (P4S8X-MX.
Hopefully, this information helps.

Regards,

fpgahardwareengineer
Comment 5 Len Brown 2013-01-30 21:26:03 UTC
Linux can't suspend to RAM on an ACPI system that doesn't support S3.

> There is also an option *Suspend Mode* which I can set to
> *Disable* or to a number of minutes.

The "number of minutes" thing is a new one on me.
Usually this is a choice between S1, S3 and disabled...
What happens when you enable this option?

Comment #4 suggests that this type of board may not support S3.

We can confirm that the BIOS is or is not exporting it by looking
at the output of acpidump from your machine -- but I don't
suspect a bug in Linux here.
Comment 6 Aaron Lu 2013-04-07 07:48:12 UTC
Hi Paul,

As Len has said, if the BIOS doesn't claim support of S3, then Linux will not be able to suspend. You should be able to tell by looking at the dmesg, there should be a line like this:
[    0.167122] ACPI: (supports S0 S3 S4 S5)

Please let us know the result, and attach the output of acpidump would also be helpful, thanks.

# acpidump > acpidump.out
Comment 7 Zhang Rui 2013-04-25 02:16:58 UTC
ping...

Note You need to log in before you can comment on or make changes to this bug.