Openbox re-starting PC?

[melodie]

Hi,

I can't hunt this kind of bug forever. While we were there I used many hours during several days for what ? To see that finally not one of the tests or options submitted by the testers, provided the solution, although they did their best efforts as well as I did. I am sure that if Texstar would know what I was missing he would have told it himself. What makes you think now that it could be related to PXE ? I have not even brought a significant change from the former version to that one. All changes done have been published clearly during the test phase, here on this forum.

I repeat : I thought Openbox was the only version where this issue happened, but in fact it was not:

I started getting similar behavior on my machine that runs PCLOS Gnome - a good portion of the time doing a clean shutdown results in an immediate attempt at rebooting, but not 100% of the time.  I've been working under the assumption it was my hardware doing it (there are some quirks there...) but I will check ACPI and might actually do some more troubleshooting now.

FWIW the motherboard is an ASUS running a socket 939 Athlon 64 x2.

I would not want to get into a sterile discussion about this, just to say I would like to try to stick to minimalist versions (for Bonsai as a base) done formerly from the rpms of the repos (I will also try to learn how to do that, if I get any available time for it during the long winter evenings).

Regards,
Mélodie

[ka9yhd]

I am now also having the reboot issue with my desktop pc running PClinuxOS Gnome 2010.12 fully updated.

It is an older PC with an AMD Sempron 3000+

[melodie]

Hi,

I have installed iasl and start to explore the method described here to investigate; This evening I have read there and at other places, that an acpi compiled (assembler) with ms compiler does not provide the same result as one compiled with the Intel compiler... 

[nightcap]

Seems the PCLinuxOS community is facing a minor technical hurdle. Hopefully there'll be a solution which should benefit all future versions. Regret to say I'm not a code warrior. Never had enough time, nor very likely, the aptitude and patience to cope with meticulous exactitude! I salute you all.   

Some basic ideas. We seem to have no obvious pattern across hardware architectures (e.g., boot ups seem to affect Intel and AMD, various chipsets, multiple BIOS and ACPI versions etc). In order to minimize bug risk, may I suggest, if possible, to avoid as much kernel/system/desktop dependency on ACPI as possible. Perhaps insert fresh code into "kill" instruction set ("terminate process"?) for insuring permanent shutdown? (Apologies for inappropriate coding terminology.)

nightcap   

[Hootiegibbon]

Martin1978,

I did have a look at some potentially helpful documents.

could you try (first - with the issue in place and known on shutdown) open a terminal and type 'poweroff'

does this cause a reboot?

if so try the following boot options one at a time and then poweroff/shutdown

acpi=noirq
acpi=strict
acpi=rdst
pci=noacpi



Jase
 

[Texstar]

New Plymouth shutdown image. lol



Older computers might need in the grub boot line

apci=ht

or

apci=force

[Dragynn]

This worked for me sorta:

Code: [Select]

[dragynn@openbox ~]$ le shutdown s'il vous plait
Error: NO! You are English type, go and boil your bottom you son of a silly person
[dragynn@openbox ~] wth? SHUTDOWN
system@openbox : [fetchez la vache] grep trebuchet | load
[dragynn@openbox ~]...............................................

[dragynn@openbox ~] apt-get install holy-hand-grenade

system@openbox : shutting down
 

[melodie]

New Plymouth shutdown image. lol



Older computers might need in the grub boot line

apci=ht

or

apci=force

Texstar,

Thank you very much !

I don't know the option "ht". Here and here I read things which seem interesting about it:
http://forums.fedoraforum.org/archive/index.php/t-28228.html

http://www.linux-noob.com/forums/index.php?/topic/2601-why-did-acpiht-fix-my-network-problem/

Does the option "ht" stand for hyper threading ? And don't the option "force" be rather used in a machine where the shutdown does not end, such as what I met in my laptop Nec ? (And solved here with the "acpi=force" option ?)

Regards,
Mélodie

[Hootiegibbon]

Mel

I do wish you would attempt to seach for your own answers ...

acpi=ht  allows only enough of the acpi system to be activated to allow hyperthreading

acpi=force is for earlier machines that were acpi able, but were set up for use of the older apm system it forced the use of acpi over the default apm.

you will find more detailed descriptions if you use a search engine...

Jase

[Texstar]

acpi is called to poweroff a computer. Disabling it in the bios just prevents Linux for calling apci to power down the system. Many motherboards have buggy, partially implemented or just plain faulty apci in their bios. Sometimes these kind of issues can be worked around using various kernel parameters if a bios upgrade is not available. acpi has been a constant issue with Linux kernels as the developers continually try to work around faulty acpi. Recent kernels have been doing a better job where as in the past one could expect acpi to work in one kernel be broke in another.

I had an AMD Athon 1700 that required acpi=ht to power off. hyperthreading was not available on AMD so when asked I was told that acpi=ht would only load a very minimum implementation of apci.

Even older systems use to need acpi=off and apmd enabled. I'm not an expert on acpi. I can only tell you what worked for me in the past on old computers.

[djohnston]

This worked for me sorta:

Code: [Select]

[dragynn@openbox ~]$ le shutdown s'il vous plait
Error: NO! You are English type, go and boil your bottom you son of a silly person
[dragynn@openbox ~] wth? SHUTDOWN
system@openbox : [fetchez la vache] grep trebuchet | load
[dragynn@openbox ~]...............................................

[dragynn@openbox ~] apt-get install holy-hand-grenade

system@openbox : shutting down
 

Holy shades of Holy Grail! Now that's funny! 

[melodie]

you will find more detailed descriptions if you use a search engine...

Jase

Beg your pardon ? Where do you think I found the 2 above links ? 

[melodie]

acpi is called to poweroff a computer. Disabling it in the bios just prevents Linux for calling apci to power down the system. Many motherboards have buggy, partially implemented or just plain faulty apci in their bios. Sometimes these kind of issues can be worked around using various kernel parameters if a bios upgrade is not available. acpi has been a constant issue with Linux kernels as the developers continually try to work around faulty acpi. Recent kernels have been doing a better job where as in the past one could expect acpi to work in one kernel be broke in another.

I had an AMD Athon 1700 that required acpi=ht to power off. hyperthreading was not available on AMD so when asked I was told that acpi=ht would only load a very minimum implementation of apci.

Even older systems use to need acpi=off and apmd enabled. I'm not an expert on acpi. I can only tell you what worked for me in the past on old computers.

Hi,

After I read 2 threads on the web about the way it is implemented, I installed and tried the command lines with iasl to be able to read and test a copy of the dstd of my Nec Laptop (which does not have too much problem : we have updated the BIOS just after I bought it from another user), and I have seen that as said in these threads, the content of the lines related to Linux is different from the ones related to the different versions of Windows. I got : "No errors, 5 warnings".

Code: [Select]

$ su
passwd
# cat /sys/firmware/acpi/tables/DSDT > /tmp/dsdt.dat
Make it belong to your user, quit root.

to disassemble it:

Code: [Select]

$ iasl -d dsdt.dat
it produces a file named dsdt.dsl which can be opened in a text editor.

Recompile it to look at the messages:

Code: [Select]

$ iasl -tc DSDT.dsl
I don't know more about it, I read about the above just the past days. I have seen how wicked is the way the acpi of the motherboards are written, and also read that the Intel compiler does a better job than other compilers.

Where the Windows parts of the file can get errors, the OS would be build in a way which deals with the errors, whereas Linux could not... :/

Some guys appearantly recompile their dsdt. (I'm not there yet!)

[nightcap]

As professed earlier I ain't familiar with writing machine code. Maybe I shouldn't butt in here? Nonetheless I've digested brief ACPI views from texstar and others.

ACPI dysfunctionality does not appear to be platform (Intel v. AMD) or mainboard-age dependent. My HP Probook is recent (June 2011, Intel i3-370). HP is the largest supplier of notebooks, and a key collaborator with Intel and M$. If any company has insider privileges and copious resources to nail ACPI, it should be HP, surely?  

Secondly the successor distro loaded onto that laptop (recent kernel, Openbox 3.5) has displayed no such boot up bugs (B.U.B.s), thus far. I also loaded PCLOS LXDE on another brand new laptop - HP Compaq 621 (Intel T4500) with ACPI switched OFF. Hasn't missed a beat in 3 months, and definitely no B.U.B.s. PCLOS LXDE was briefly installed on an AMD Phenom tower also with ACPI switched OFF. No B.U.B.s. I have yet to load PCLOS on (much) older hardware. Will test as soon as practible.

Thirdly HP officially supports Suse Linux and employs a manager to ensure as many of their products are fully Linux compatible. This suggests some corporate programming knowledge of solid Linux shutdowns. (Unqualified Disclaimer - I am in no way employed, or paid by, or connected to HP in any way whatsoever.)

Lastly having installed dozens of other linux distros (including those with OpenBox) on many PCs, this is the first example of sudden B.U.B.s I've encountered. Logic would seem to suggest we need a minor fine-tune within PCLOS.

nightcap

[ka9yhd]

New Plymouth shutdown image. lol



Older computers might need in the grub boot line

apci=ht

or

apci=force

Nice image............ Now quit stealing code from Win 98.