ACPI -- why is it such a common source of crashes?

When I mentioned that my Hardy server was locking up, someone
suggested to disable ACPI. Which I did and I hope that it is going to
fix the problem.

But why is that so? Why is ACPI such a common source of troubles?

Is there some design flaw with ACPI itself, or Linux implementation
thereof?

--
Due to extreme spam originating from Google Groups, and their inattention
to spammers, I and many others block all articles originating
from Google Groups. If you want your postings to be seen by
more readers you will need to find a different means of
posting on Usenet.
http://improve-usenet.org/

> But why is that so? Why is ACPI such a common source of troubles?
> Is there some design flaw with ACPI itself, or Linux implementation
> thereof?

A quick google will show you it's one of the least reliable concepts
ever applied to computing since Windows itself.

On 2008-05-12, Colin Wilson <REMOVEEVERYTHINGBUTnewsgroup@phoenixbbsZEROSPAM.c o.uk> wrote:
>> But why is that so? Why is ACPI such a common source of troubles?
>> Is there some design flaw with ACPI itself, or Linux implementation
>> thereof?
>
> A quick google will show you it's one of the least reliable concepts
> ever applied to computing since Windows itself.

I looked up its wikipedia page and, indeed, now I understand. It is
like a whole OS concept foisted on other OSes.

--
Due to extreme spam originating from Google Groups, and their inattention
to spammers, I and many others block all articles originating
from Google Groups. If you want your postings to be seen by
more readers you will need to find a different means of
posting on Usenet.
http://improve-usenet.org/

On Mon, 12 May 2008 12:02:06 -0500, Ignoramus17662 wrote:

> When I mentioned that my Hardy server was locking up, someone suggested
> to disable ACPI. Which I did and I hope that it is going to fix the
> problem.
>
> But why is that so? Why is ACPI such a common source of troubles?
>
> Is there some design flaw with ACPI itself, or Linux implementation
> thereof?

I have always boot my machines with ACPI=OFF. IT literally made my laptop
functioning as it should, and not like a dead animal, which also stinks..

--
Jerry Maguire: Help me... help you. Help me, help you. ;-)

Ignoramus17662 wrote:

> When I mentioned that my Hardy server was locking up, someone
> suggested to disable ACPI. Which I did and I hope that it is going to
> fix the problem.
>
> But why is that so? Why is ACPI such a common source of troubles?
>
> Is there some design flaw with ACPI itself, or Linux implementation
> thereof?

AFAIK, the very complicated ACPI-tables in the BIOS are
often flawed.

Since a PC that doesn't run Windows ist unsalable, the
distributors are testing carefully and developing
workarounds for Windows.

But not so for Linux.

So, the chance that Linux is stumbling over such flaws,
is much higher. But it's not a Linux flaw.

Gunter

In article <tPudnTNcE6CT5bXVnZ2dnUVZ_sjinZ2d@giganews.com>,
Ignoramus17662 <ignoramus17662@NOSPAM.17662.invalid> wrote:
>
> But why is that so? Why is ACPI such a common source of troubles?
>
> Is there some design flaw with ACPI itself, or Linux implementation
> thereof?

At its core, the problem is that PCs are a product of evolution, not
intelligent design. :-)

In the PC world, you've got your hardware companies on one hand, and
your operating system companies on the other (you can read that as
"Microsoft" if you wish--that is a sufficiently close approximation for
most purposes).

So, hardware manufacturers are constrained to keep their advances
compatible with existing operating systems, and operating system
companies are constrained to make their OSes compatible with the
existing hardware.

That's where such abominations as CD-ROM booting working by taking a 1.4
meg file from the CD and making it look to the software like it is a
boot floppy in the A: drive come from. Or USB making the keyboard
appear to be a PS/2 keyboard connected to a PS/2 port.

Look closely at ACPI, and I believe you'll find that it is a product of
PC evolution.

You generally don't see this level of kludge in workstations (e.g., from
Sun or IBM) or in Macs, because the companies that make those are also
making the system software. This lets them make clean breaks: models
before this date work the old way, models after work the new way.

--
--Tim Smith

Ignoramus17662 <ignoramus17662@NOSPAM.17662.invalid> wrote in
news:tPudnTNcE6CT5bXVnZ2dnUVZ_sjinZ2d@giganews.com :

> When I mentioned that my Hardy server was locking up, someone
> suggested to disable ACPI. Which I did and I hope that it is going to
> fix the problem.
>
> But why is that so? Why is ACPI such a common source of troubles?
>
> Is there some design flaw with ACPI itself, or Linux implementation
> thereof?
>

Its the duality of the love-hate relationship. I would love to see ACPI
work and ACPI hates me. Each time I do a new install I seem to find yet
other things which are different, broken, or just not right on my few
year old T43 thinkpad. Like for some reason on Gutsy, it would lock up
with the screensavers and when I came back, the gnome screensaver package
would have completely killed the display and made me have to reboot/power
cycle. On servers I manage, I don't see the need for it. I really don't
want a server to gracefully sleep or suspend when I close its lid or have
a period of time of non-use.

My personal take is that ACPI sucks and its all explained in what it
stands for. I read awhile ago that its just...

Another Crummy Proprietary Interface

But I spend sometimes an inordinate amount of time wrestling with it.

--
Michael Perry | Do or do not. There is no try --Master Yoda
mperry@lnxpowered.org | http://www.lnxpowered.org