ACPI not working

ovihc · 2004-06-16 17:29:39

I am running Arch linux on an HP pavilion ze4400. I'm also running windows XP (since it came with it) but in a small partition because my gf needs to use the computer as well. When I'm using XP, the fan is very efficient. It turns on when I'm doing a lot of things, and it turns off when the computer is not doing anything. I like that since the battery can last longer, and there is no noise at all.

When I previously installed SUSE 9.0, it worked well, but not as well as XP. At least the fan would be a little more efficient in SUSE. It would turn off only when I wouldn't use the computer for a long time with the battery. If it was plugged in, the fan never turned off, but it would at least remain at lower speeds.

Now I'm interested in Arch. I dont need all that KDE crap SUSE likes to use. But i do miss acpi. When I installed Arch, everything went well. When I boot off GRUB, the computer halts. It stops on ACPI *something* IRQ. I can't remember. I have to go into GRUB and type in "acpi=off" and then the computer doesn't halt. But of course, the fan is always on and the computer makes a lot of noise.

What can I do? How can I activate ACPI without having the computer halt? I have ACPI installed. I really want to give Arch a try.

ovihc · 2004-06-16 17:57:00

I went back to check what the computer displays when it halts.

ACPI: Subsystem revision 20040326
ACPI: IRQ9 SCI Level Trigger

Those are the only 2 lines that have anything to do with ACPI.

I really need help.

leejpar · 2004-06-16 19:30:35

Take a look at the the kernel messages during the boot process with dmesg and see if a bunch of ACPI stuff is listed. Also look in /proc for an acpi entry. If you are running the 2.6.x kernel then ACPI should be there. However in order for the ACPI signals to be processed you will need to load the acpi deamon in rc.local (acpid). The deamon will then process scripts that you will need to write yourself or "beg, borrow or steal" from someone else. I asked a few days ago on the forum if someone actually had ACPI funtional via some scripts for acpid and no one answered. So, that leads me to believe that none "really" has this working properly yet. I did find some scripts on the net that I am working on to adapt to my desktop machine. They may just end up being an example to help figure out exactly what is required. The scripts I did find are designed for a laptop's needs. I believe a very brief explaination of what the scripts look like can be seen in the man page for acpid.

hcman · 2004-06-16 20:16:01

What kernel version are we talking about here ('uname -r') because I am running Arch on a compaq presario laptop with a 2.6.5 kernel with ACPI.

On a recently compiled 2.6.6 kernel (with ACPI) I could not get my computer to boot either. So I wonder what changes they made in the ACPI table.

Bobonov · 2004-06-17 05:44:21

I think this is some type of incompatibility, I am having exactly the same experience on a computer I am installing right now.
At the begin I tought it was a problem with bios setting but I tried several configuration without result. it stop during kernel loading at exactly the same point as ovihc.
At this moment I did not have time to investigate the reason deeply. I will look at it later.

ovihc · 2004-06-17 12:39:55

it's kernel 2.6.6. i know acpi is in there! it's just freezing my computer at startup.

should i mention that i have a AMD Athlon-XP-M?

the only way i can boot up my computer succesfully is if i do

acpi=off

on grub. and because of that there's nothing in /proc about acpi. if the computer boots normally (i.e. with acpi enabled) then it freezes since ACPI is freezing it.

by the way, acpid does exist in my computer. even when i type "acpi=off", the acpid script runs, but of course, it spits out that there is no /proc/acpi.

ACPI worked much better is SUSE. why doesn't it work with Arch? should i change something in /etc/modprobe.conf ?

Bobonov · 2004-06-17 13:28:57

No I do not think it is a problem related to the processor, as I wrote I have the same problem with a Pentium 3 800, not either to laptop mine is a normal desktop.
I think it is more related to the chipset/bios, in the sense of modules compiled.
Actually I can't do more test, that computer is at home.
Correct me if I am wrong but your chipset shoul be an ALI, mine is a VIA.

ovihc · 2004-06-17 13:42:28

your'e correct. my chipset is ALI. at least i think so. my sound card module is ALI too. sorry, but I'm not too familiar with modules. i like Arch, because it's also helping me learn about all this.

Leejpar, it seems that you think that I might have acpi activated without me knowing. I do know a little about acpi, and i dont have /proc/acpi. nothing is activated from acpi since i type "acpi=off" in grub. If i dont type that, then my computer freezes.

HCMAN, did acpi work with you on kernel 2.6.5? let me know.

is there a way that I can run Arch on 2.6.5 insted of 2.6.6? The only installation method I've done is through FTP.

***I'm actually thinking of installing SUSE 9.1 and checking every single thing about acpi. Then I can transfer the configurations from SUSE to Arch, and maybe it will work. Although I dont think SUSE is using 2.6.6. for their kernel. I would hate having to go back to SUSE just because of ACPI. I dont want my laptop to toast!

ovihc · 2004-06-17 14:29:06

i'm looking into lm_sensors right now. maybe that can provide a solution for us.

ovihc · 2004-06-17 15:16:58

I'm the only one posting. sorry for all these small posts, but i'm working on this ACPI problem right now. i'm looking for people to help me. so please, if you have a comment or suggestion, post it so that we can find a solution. in the meanwhile, i'm pulling my hairs out!

that said, i tried the lm_sensors but it didn't work. lm_sensors cannot find any sensors on my computer!! does anythng work with this HP!!

although i found something interstesting:

does anyone know if it will work?

From:    Len Brown <len.brown@intel.com>
To:    Linus Torvalds <torvalds@osdl.org>
Subject:    [BKPATCH] ACPI for 2.6
Date:    03 May 2004 12:52:31 -0400
Cc:    "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>, ACPI Developers <acpi-devel@lists.sourceforge.net>

Hi Linus, please do a

bk pull bk://linux-acpi.bkbits.net/linux-acpi-release-2.6.6

A couple of key interrupt configuration fixes in this patch.
thanks,
-Len

ps. a plain patch is also available here:
ftp://ftp.kernel.org/pub/linux/kernel/p … .6.diff.gz

This will update the following files:

through these ChangeSets:

<len.brown@intel.com> (04/05/03 1.1371.707.20)
[ACPI] export symbols to button module

<len.brown@intel.com> (04/05/01 1.1371.707.19)
[ACPI] PCI Interrupt Link fixes
Handle BIOS that reference disabled PCI Interrupt Link Devices
http://bugme.osdl.org/show_bug.cgi?id=1581

Clean up VIA _CRS = 0 BIOS workaround

Handle BIOS returning _CRS outside _PRS
http://bugme.osdl.org/show_bug.cgi?id=2567

delete now unused _SRS retry code
disable redundant console messages

<len.brown@intel.com> (04/04/28 1.1371.707.18)
[ACPI] button build fix

<len.brown@intel.com> (04/04/28 1.1371.707.17)
[ACPI] fix build warning in dmi_scan

<len.brown@intel.com> (04/04/28 1.1371.707.16)
[ACPI] support button driver unload (Luming Yu)
http://bugzilla.kernel.org/show_bug.cgi?id=2281

<len.brown@intel.com> (04/04/28 1.1371.707.15)
[ACPI] toshiba_acpi driver if acpi_disabled (David Shaohua Li)
http://bugzilla.kernel.org/show_bug.cgi?id=2465

<len.brown@intel.com> (04/04/28 1.1371.707.14)
[ACPI] rmmod ACPI modules vs /proc
from Anil S Keshavamurthy and David Shaohua Li
http://bugzilla.kernel.org/show_bug.cgi?id=2457

<len.brown@intel.com> (04/04/28 1.1371.707.13)
[ACPI] pci-link may not always be SHARED (SuSE via Luming Yu)
http://bugzilla.kernel.org/show_bug.cgi?id=2404

<len.brown@intel.com> (04/04/28 1.1371.707.12)
[ACPI] battery "charged" instead of "unknown" (Luming Yu)
http://bugzilla.kernel.org/show_bug.cgi?id=1863

<sziwan@hell.org.pl> (04/04/28 1.1371.707.11)
[PATCH] acpi4asus 0.28 (Karol 'sziwan' Kozimor)
- Added support for Samsung P30
- Fixed an oops triggered by non-standard hardware (Samsung P30)
- Added support for L4400L and M6800N

The patch also removes some superfluous data. It doesn't include the
copy_from_user() conversion, it will be released as a separate patch.

<akpm@osdl.org> (04/04/28 1.1371.707.10)
[PATCH] acpi build fix
setup.c:608: `acpi_skip_timer_override' undeclared

<len.brown@intel.com> (04/04/24 1.1371.749.3)
ACPI irq->gsi naming (Bjorn Helgaas)

<len.brown@intel.com> (04/04/24 1.1371.707.9)
[ACPI] No IRQ known... - using IRQ 255 (Bjarni Rúnar Einarsson)
http://bugzilla.kernel.org/show_bug.cgi?id=2148

<len.brown@intel.com> (04/04/23 1.1371.707.8)
[ACPI] workaround for nForce2 BIOS bug: XT-PIC timer in IOAPIC mode
"acpi_skip_timer_override" boot parameter
dmi_scan for common platforms, may be replaced with PCI-ID in future.
http://bugzilla.kernel.org/show_bug.cgi?id=1203

<len.brown@intel.com> (04/04/23 1.1371.707.7)
[ACPI] Workaround "_BBN 0" BIOS bug
enhance "pci=noacpi" to skip ACPI PCI configuration and interrupt
config
add "acpi=noirq" to skip just ACPI interrupt config (David Shaohua
Li)
http://bugzilla.kernel.org/show_bug.cgi?id=1662

<len.brown@intel.com> (04/04/22 1.1371.707.6)
[ACPI] allow IRQ2 to be used in ACPI/IOAPIC mode
http://bugzilla.kernel.org/show_bug.cgi?id=2564

<len.brown@intel.com> (04/04/22 1.1371.707.5)
[ACPI] if acpi_os_name= is used, print what it finds

<len.brown@intel.com> (04/04/20 1.1371.707.4)
ACPI] Delete IRQ2 "cascade" in ACPI IOAPIC mode
no such concept exists in ACPI, frees IRQ2 for use.

<len.brown@intel.com> (04/04/20 1.1371.707.3)
[ACPI] enhance intr-src-override parsing to handle ES7000
http://bugme.osdl.org/show_bug.cgi?id=2520

<len.brown@intel.com> (04/04/16 1.1371.707.2)
[ACPI] enable 440GX PIRQ router workaround

<len.brown@intel.com> (04/04/14 1.1371.707.1)
[ACPI] fix x86_64 mis-merge

ovihc · 2004-06-17 15:34:03

I am about to apply the patch. it seems like the patch fixes some IRQ errors, which is what i need! when i type "patch -p0 <acpi........", everything goes well, except then it asks me for a file to patch. i have no idea!! can anyone help?

leejpar · 2004-06-17 16:58:15

I would try the new 2.6.7 kernel before I tried to patch version 2.6.6. If the 2.6.6 needed patching, then maybe 2.6.7 was released with the fix already. Arch released the 2.6.7 kernel last night to "current".

hcman · 2004-06-18 08:36:22

ovihc,

yes ACPI works in 2.6.5 but since 2.6.7 (IDE) is in current you might wanna try that.I'll compile 2.6.7 later today and see if that works for me.

Don't worry, ACPI will work in Arch as it did in SuSE.

Bobonov · 2004-06-18 08:52:34

I have the same problem with 2.6.7
I have to say that I do not bother that much for this particular pc since it is a server so all powersaving function should be off, but this is annoing anyway.

Actually I didn't do any test since I directly tried 2.6.7

hecker · 2004-06-18 21:12:29

i'm using kernel 2.6.7. my notebook has a VIA chipset and i boot with pci=noacpi to use pcmcia without problems (otherweise acpi will disrupt pcmcia).

dmesg throws out some interresting messages. e.g...

ACPI: AC Adapter [AC0] (on-line)
ACPI: Battery Slot [BAT0] (battery present)
ACPI: Power Button (FF) [PWRF]
ACPI: Lid Switch [LID]
ACPI: Sleep Button (CM) [SLPB]
ACPI: Processor [CPU1] (supports c1 c3)
ACPI: Thermal Zone [THRM] (75 c)

yeah... overall i would say that ACPI is working for me, but /proc/acpi contains no scripts, seems like that i've to write my own!?

another thing that bug's me...
under gnome i use this sweet little battery display but i can't use it cause: The suspend command '/usr/bin/apm -s' was unsuccessful hm...

edit: ah... /usr/bin/apm and i get No APM support in kernel
modprobe apm -> FATAL: Module apm not found
can ACPI work together with APM ?

ovihc · 2004-06-20 14:11:42

i found a temporary solution to my computer!!

on grub, type in

"nolapic"

it worked! try it! ACPI now works.

Although it fixes the acpi, i'm still curious what it does. does it turn something else off? i want to look more depth into this. i want my computer fully running!

lanrat · 2004-06-20 15:02:36

Just search the forums for noapic. I placed a link to the description of this option some time ago. Basically it changes how kernel handles irqs. It should be quite safe to use this.

Arch Linux

#1 2004-06-16 17:29:39

ACPI not working

#2 2004-06-16 17:57:00

Re: ACPI not working

#3 2004-06-16 19:30:35

Re: ACPI not working

#4 2004-06-16 20:16:01

Re: ACPI not working

#5 2004-06-17 05:44:21

Re: ACPI not working

#6 2004-06-17 12:39:55

Re: ACPI not working

#7 2004-06-17 13:28:57

Re: ACPI not working

#8 2004-06-17 13:42:28

Re: ACPI not working

#9 2004-06-17 14:29:06

Re: ACPI not working

#10 2004-06-17 15:16:58

Re: ACPI not working

#11 2004-06-17 15:34:03

Re: ACPI not working

#12 2004-06-17 16:58:15

Re: ACPI not working

#13 2004-06-18 08:36:22

Re: ACPI not working

#14 2004-06-18 08:52:34

Re: ACPI not working

#15 2004-06-18 21:12:29

Re: ACPI not working

#16 2004-06-20 14:11:42

Re: ACPI not working

#17 2004-06-20 15:02:36

Re: ACPI not working

Board footer