Boot problem - not the normal type!

[viking777]

When my computer was shutting down yesterday I noticed the following message in the shutdown text:

wake-up enabled by acpi

Now call me an old traditionalist if you like, but when my laptop shuts down I would like it to 'disable' things not 'enable' them, and certainly not any wake-up device. I don't know for sure that this is the cause of these problems (incidentally it hasn't happened for a while), but armed with an error message I was at least able to Google a bit. The first bit of information I came up with was that if you run this command:

Code: Select all

cat /proc/acpi/wakeup

You get a result something like:

Code: Select all

Device   S-state     Status   Sysfs node
HDEF     S3   *disabled  pci:0000:00:1b.0
LANE     S4   *enabled   pci:0000:08:00.0
USB1     S3   *disabled  pci:0000:00:1d.0
USB2     S3   *disabled  pci:0000:00:1d.1
USB3     S3   *disabled  pci:0000:00:1d.2
EHC1     S3   *disabled  pci:0000:00:1d.7


Which quite clearly states that 'LANE' (which I presume refers to wake on lan) is enabled by acpi despite the fact that it is very definitely disabled in my bios. So is acpi overriding bios settings now? I don't know but it would explain the problem.

Obviously my next search was for a way to disable this setting - but it isn't that easy. There is a program that you can download (you wont have it installed) called ethtool which will disable the setting, the command is:

Code: Select all

ethtool -s eth1 wol d

This approach has two major drawbacks - first, is not permanent, the setting is re-enabled at the next boot or probably even if you restart the interface. The second is that it will not configure wireless interfaces and at the moment that is what I am using.

I did find a supposed solution to the first drawback which is to add the line:

Code: Select all

post-up ethtool -s eth1 wol d

To /etc/network/interfaces, but this doesn't work for me presumably because I an not using an ethernet interface.

So that is as far as I have got at the moment. I may be completely on the wrong track, or I may have discovered the cause of the mysterious boots along with a serious bug in acpi I don't really know for sure. If anyone can take my line of research further, or just plain tell me I am wrong, feel free.

[AlbertP]

Actually my ethernet is also listed as enabled in /proc/acpi/wakeup, but my computer does not have mysterious wake-ups.

Code: Select all

PXSX     S4   *enabled   pci:0000:08:00.0

08:00.0 is the ethernet in lspci.

[xircon]

Viking

Think this is a red herring as:

Code: Select all

 cat /proc/acpi/wakeup
Device  S-state   Status   Sysfs node
P0P1      S3    *disabled  pci:0000:00:01.0
P0P2      S3    *disabled 
P0P3      S3    *disabled 
P0P4      S3    *disabled 
P0P5      S3    *disabled 
BR20      S3    *disabled  pci:0000:00:1e.0
PEX0      S3    *disabled  pci:0000:00:1c.0
PEX1      S3    *disabled  pci:0000:00:1c.1
PEX2      S4    *disabled  pci:0000:00:1c.2
PEX3      S3    *disabled 
PEX4      S3    *disabled  pci:0000:00:1c.4
PEX5      S3    *disabled 
PEX6      S3    *disabled 
PEX7      S3    *disabled 
GBE       S4    *disabled 
EHC2      S0    *disabled  pci:0000:00:1a.0
EHCI      S0    *disabled  pci:0000:00:1d.0
LID0      S3    *enabled   


No mention of lane, all disabled except lid.

Cheers

Steve

[viking777]

No mention of lane, all disabled except lid.

Cheers

Steve

Do you know Steve, I am pleased to read that, because trying to get the "lane" disabled permanently in that file is doing my head in

I am going to switch off now and chill out a bit, I don't know how many hours I have been chasing this rainbow.

(I just hope you are right )

[AlbertP]

ACPI remains a complex thing with many hardware and software vendors (including Microsoft) not exactly keeping to the standard.

Code: Select all

$ cat /proc/acpi/wakeup
Device   S-state     Status   Sysfs node
PWRB     S4   *enabled   
PXSX     S4   *enabled   pci:0000:08:00.0
PS2K     S3   *enabled   pnp:00:09
PS2M     S3   *disabled  pnp:00:0a

Everything's enabled except the mouse (PS2M). That's indeed true - keyboard and power button can wake up from suspend, but the mouse doesn't. I haven't tested booting the computer with ethernet (PXSX) as I always use wireless.
And the lid is not mentioned at all, though the computer wakes up from suspend (but not from hibernate) when I open it.

This way you can see that every manufacturer has a different implementation of ACPI. The Linux kernel, for example, contains a lot of hacks to get Toshiba's ACPI implementation working. Microsoft not fully keeping to ACPI standards has also contributed to this mess.

[viking777]

Thanks for the reply AlbertP and I am sure you are right. I have first hand experience of this on an old laptop of mine that would never shut down without the power button. This was 'diagnosed' on another forum as a dsdt problem and I even went as far as compiling my own dsdt into a kernel rebuild, but it still never worked.

As I said when I first introduced this theory a few posts back, I am not sure if acpi is the cause of this restart problem or not. Both AlbertP and xircon, displayed their /proc/acpi/wakeup files, both have items enabled, but one of them is affected by this and one isn't. My guess is that that may be down to the interaction of acpi with an individuals hardware spec, which is why the error appears to be so random.

What annoys me about this is that there should be a tool somewhere called "acpiset" or something of that nature which allows users to tweak individual aspects of that very low level (if it overrides the bios) program. If there is one I cant find it. But this ability to be in control of your own computer is THE central tenet of Linux.

Anyway as there is no fine tuning method that I can find I am for the present resorting to the 'sledge hammer' approach and booting with "acpi=off". How long I have to do this before I prove to myself one way or another that it makes a difference I really don't know.

EDIT. AAAGGGHHH. And you know what side effect this has? It now won't shut down at all (without the power button). It is like beating your head against a brick wall

Still If a use the power button to shut it down at least I know it is shut down

EDIT2. OK I officially give in. Booting with acpi off breaks conky (and I can't be bothered to rewrite the config file and I don't want to do without it) so back to square one.

[AlbertP]

Hmm yes a lot of modern computers can only power down with ACPI. Older ones often support powering down or rebooting through the keyboard controller as well, but your one doesn't seem to support that.

[Anja Noir]

Old thread but...

If you have the checkbox "Save settings for future logins" checked
somewhere - un-check it.

My computer was booting by itself every morning after having clicked
Shutdown at night. Un-checking the above mentioned checkbox solved this
issue for me. On the new LMDE 201204 Xfce edition, the checkbox is
located at the bottom of the Log out choices window, i.e, Shutdown,
Reboot, Suspend, etc...

~Anja

[usbtux]

KISS.

Faulty power switch?

[viking777]

Thanks for the replies people, but as was mentioned, this is an old thread, and it is one of those mysterious things that happens with computers, the problem came along and then it went away again. I haven't the slightest idea why it did either, but it has 'solved' itself

[pacosiquot]

same problem with a Toshiba NB100. System: Ubuntu.
I think if I shut it down holding the I/O button, the problem doesn't occur. But I don't like this method.
Any news from this issue?
excuse my english. I'd appreciate help, I'm really worry bout my computer starting up in it's case!!