poweroff problem

poweroff problem

[Anant Aneja]

I've got a problem with my 2.4.2-2 kernel.
after reaching the power down stage i get a :
1. complete listing of the cpu registers
2. a message saying sementaion fault with halt -i -p -d

i contacted the author of the poweroff script but
he says that this means it must be a kernel problem

i know that the kernel is old but i dont want to update
cos my modem has a low download speed

also i cant give u the complete listing of the cpu
registers since it occurs at the last stage
of shutdown and i cant copy it to a file
and am too lazy to write it down

can anyone help me ?
Anant


_______________________________________________________________________
Odomos - the only  mosquito protection outside 4 walls -
Click here to know more!
http://r.rediff.com/r?http://clients.rediff.com/odomos/Odomos.htm&&odomos&&wn

Re: poweroff problem

[Alan Cox]

On Sad, 2003-04-05 at 07:08, Anant Aneja wrote:
> also i cant give u the complete listing of the cpu
> registers since it occurs at the last stage
> of shutdown and i cant copy it to a file
> and am too lazy to write it down

We are too lazy to help you.
Goodbye

Alan

Re: poweroff problem

[Michael Buesch]

> On Sad, 2003-04-05 at 07:08, Anant Aneja wrote:
> > also i cant give u the complete listing of the cpu
> > registers since it occurs at the last stage
> > of shutdown and i cant copy it to a file
> > and am too lazy to write it down

Documentation/serial-console.txt
... but if you're too lazy to read... :)

On Saturday 05 April 2003 18:47, Alan Cox wrote:
> We are too lazy to help you.
> Goodbye
:) *grin*

Regards
Michael Buesch.

-- 
My homepage: http://www.8ung.at/tuxsoft
fighting for peace is like fu**ing for virginity

Re: poweroff problem

[Willy Tarreau]

On Sat, Apr 05, 2003 at 06:08:04AM -0000, Anant Aneja wrote:
 
> i know that the kernel is old but i dont want to update
> cos my modem has a low download speed

you may try to get small chunks of a new one each time you connect to the net
so that after several hours of nearly normal browsing, you'll get a new kernel
as a bonus.

> also i cant give u the complete listing of the cpu
> registers since it occurs at the last stage
> of shutdown and i cant copy it to a file
> and am too lazy to write it down
> 
> can anyone help me ?

at least, try to patch it with kmsgdump to retrieve the last kernel messages
on a floppy. You'll find it easily if you're not too lazy for that ;-)

Willy

RE: poweroff problem

[James Watkins-Harvey]

Hi,


> I've got a problem with my 2.4.2-2 kernel.

  [...]

> i know that the kernel is old but i dont want to update
> cos my modem has a low download speed

  [...]

> can anyone help me ?


For sure!  You can probably help yourself (and people on the list) by
upgrading your kernel to something a bit newer.  No one still have a box
with such installation, anyway.

Common, your modem is that slow?  Ok, let's say you have a 28.8 KBps...
Should be about 2 hours to download the file; give it a try yourself:
http://internetservices.cnet.com/g/bm/dl/0001.asp?filename=Linux+Kernel+2.4.
20+(tar.bz2)&sd=26.1MB&size=27421046 .  Of course, make sure to download the
bzip version, not the gzip one (
ftp://ftp.kernel.org/pub/linux/kernel/v2.4/linux-2.4.20.tar.bz2 ).

Once installed, make sure to keep a clean version of the kernel tree, so in
the future you can regularly download incremental updates; kernel's patch
are about 1 MB per development month.

Give it a try, then if it still not works, it could be useful to mention it
on LKML.  But then, be preapared to do testing by yourself, cause it seems
like you're the only one experiencing such problems.


> also i cant give u the complete listing of the cpu
> registers since it occurs at the last stage
> of shutdown and i cant copy it to a file
> and am too lazy to write it down


Sorry to say it, but i don't think Linux is a good place for laziness;
actually I know about an operating system developped by lazy programmers for
lazy users...  Don't remember the name ;)




James

Re: poweroff problem

[Stephen Rothwell]

On 5 Apr 2003 06:08:04 -0000 "Anant Aneja" <anantaneja@rediffmail.com> wrote:
>
> I've got a problem with my 2.4.2-2 kernel.
> after reaching the power down stage i get a :
> 1. complete listing of the cpu registers
> 2. a message saying sementaion fault with halt -i -p -d

Has it always done this?

> i contacted the author of the poweroff script but
> he says that this means it must be a kernel problem

More likely a BIOS problem that is weel known.

> also i cant give u the complete listing of the cpu
> registers since it occurs at the last stage
> of shutdown and i cant copy it to a file
> and am too lazy to write it down

Write down the first few lines at least ...
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

Re: poweroff problem

[Udo A. Steinberg]

[-- Attachment #1: Type: text/plain, Size: 881 bytes --]

On Sun, 6 Apr 2003 23:33:19 +1000 Stephen Rothwell (SR) wrote:

Hello,

SR> On 5 Apr 2003 06:08:04 -0000 "Anant Aneja" <anantaneja@rediffmail.com> wrote:
SR> >
SR> > I've got a problem with my 2.4.2-2 kernel.
SR> > after reaching the power down stage i get a :
SR> > 1. complete listing of the cpu registers
SR> > 2. a message saying sementaion fault with halt -i -p -d
SR> 
SR> Has it always done this?

I have the same problem with Linux-2.4.21-pre6. I don't know when this started
to happen because I pretty much never halt my machine.

SR> More likely a BIOS problem that is weel known.

It's not a BIOS problem here. halt works pretty well with Linux-2.5.66 here.
It's most likely an ACPI problem. What happens here is that the code to power
down actually does not manage to turn the machine off, instead after a while
the NMI watchdog kicks in and the kernel oopses.


-Udo.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: poweroff problem

[Alan Cox]

On Sul, 2003-04-06 at 16:46, Udo A. Steinberg wrote:
> On Mon, 7 Apr 2003 00:27:03 +1000 Stephen Rothwell (SR) wrote:
> 
> SR> I was asuming the original report was from a kernel using APM not ACPI.
> SR> Did 2.4.2 have ACPI?
> 
> Err no, so he must have been using APM.
> 
> 2.4.2+APM has the problem, 2.4.21-pre+ACPI has the problem. Do APM and
> ACPI both share the same code to power off a machine? If so that seems to
> be the culprit.

We rely on the bios for the power off sequences. Many BIOS vendors do
set it up to share the bios code it seems

Re: poweroff problem

[Stephen Rothwell]

On Sun, 6 Apr 2003 15:58:14 +0200 "Udo A. Steinberg" <us15@os.inf.tu-dresden.de> wrote:
>
> It's not a BIOS problem here. halt works pretty well with Linux-2.5.66 here.
> It's most likely an ACPI problem. What happens here is that the code to power
> down actually does not manage to turn the machine off, instead after a while
> the NMI watchdog kicks in and the kernel oopses.

I was asuming the original report was from a kernel using APM not ACPI.
Did 2.4.2 have ACPI?

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/   APM Maintainer

Re: poweroff problem

[Alan Cox]

On Sul, 2003-04-06 at 17:37, Udo A. Steinberg wrote:
> On 06 Apr 2003 15:24:55 +0100 Alan Cox (AC) wrote:
> 
> AC> We rely on the bios for the power off sequences. Many BIOS vendors do
> AC> set it up to share the bios code it seems
> 
> Ok, but this does not explain why things work with 2.5.66 on the exact same
> machine, unless 2.5 had workarounds for BIOS bugs or didn't use the BIOS for
> power off. Board is an Asus A7V, BIOS version 1011.

Different options for reboot (16 v 32) perhaps, might even be something is
random as which way the carry flag is when the bios code is called. If you
want to be sure for the APM case stick a printk just before we drop into
the BIOS and make sure we oops after and not before..

Re: poweroff problem

[Udo A. Steinberg]

[-- Attachment #1: Type: text/plain, Size: 375 bytes --]

On Mon, 7 Apr 2003 00:27:03 +1000 Stephen Rothwell (SR) wrote:

SR> I was asuming the original report was from a kernel using APM not ACPI.
SR> Did 2.4.2 have ACPI?

Err no, so he must have been using APM.

2.4.2+APM has the problem, 2.4.21-pre+ACPI has the problem. Do APM and
ACPI both share the same code to power off a machine? If so that seems to
be the culprit.

-Udo.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: poweroff problem

[Udo A. Steinberg]

[-- Attachment #1: Type: text/plain, Size: 385 bytes --]

On 06 Apr 2003 15:24:55 +0100 Alan Cox (AC) wrote:

AC> We rely on the bios for the power off sequences. Many BIOS vendors do
AC> set it up to share the bios code it seems

Ok, but this does not explain why things work with 2.5.66 on the exact same
machine, unless 2.5 had workarounds for BIOS bugs or didn't use the BIOS for
power off. Board is an Asus A7V, BIOS version 1011.

-Udo.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: poweroff problem

[Petr Vandrovec]

On Sun, Apr 06, 2003 at 06:37:13PM +0200, Udo A. Steinberg wrote:
> On 06 Apr 2003 15:24:55 +0100 Alan Cox (AC) wrote:
> 
> AC> We rely on the bios for the power off sequences. Many BIOS vendors do
> AC> set it up to share the bios code it seems
> 
> Ok, but this does not explain why things work with 2.5.66 on the exact same
> machine, unless 2.5 had workarounds for BIOS bugs or didn't use the BIOS for
> power off. Board is an Asus A7V, BIOS version 1011.

On A7V with BIOS 1012a01 2.5.40+ ACPI always worked flawlessly, while 2.4.x has
problems. Currently (2.4.21-pre7) is better than before: it writes some panic
to screen, but poweroffs immediately after that (and sorry, nearest serial 
console is about 5km far away, and everybody should use 2.5.x anyway), so I
always suspected that problem is not in poweroff itself, but in some
driver's devexit procedure (I have enabled hotplug in kernel, so __devexit
is not stripped out).

And if you are talking about A7V, for some old ACPI (maybe all 2.4.x, I have
no idea) take a look at http://groups.google.com/groups?q=vandrovec+acpi+poweroff&hl=en&lr=&ie=UTF-8&oe=UTF-8&selm=linux.kernel.20020124184011.GA23785%40vana.vc.cvut.cz&rnum=1
					Best regards,
						Petr Vandrovec
						vandrove@vc.cvut.cz

Re: poweroff problem

[Udo A. Steinberg]

[-- Attachment #1: Type: text/plain, Size: 780 bytes --]

On 06 Apr 2003 16:14:42 +0100 Alan Cox (AC) wrote:

AC> Different options for reboot (16 v 32) perhaps, might even be something is
AC> random as which way the carry flag is when the bios code is called. If you
AC> want to be sure for the APM case stick a printk just before we drop into
AC> the BIOS and make sure we oops after and not before..

I just tried with APM; machine powers down without problems. It's ACPI which
doesn't power down. Last thing it prints during powerdown is:

hwsleep-0178 [-24] Acpi_enter_sleep_state: Entering S5

I can't find any specific A7V workarounds in 2.5.66 ACPI code, so I guess
the ACPI code in 2.4 isn't up-to-date.

The original poster's problem is then probably indeed related to a buggy BIOS
if it doesn't even powerdown with APM.

-Udo.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: poweroff problem

[Carl-Daniel Hailfinger]

Udo A. Steinberg wrote:
> On 06 Apr 2003 16:14:42 +0100 Alan Cox (AC) wrote:
> 
> AC> Different options for reboot (16 v 32) perhaps, might even be something is
> AC> random as which way the carry flag is when the bios code is called. If you
> AC> want to be sure for the APM case stick a printk just before we drop into
> AC> the BIOS and make sure we oops after and not before..
> 
> I just tried with APM; machine powers down without problems. It's ACPI which
> doesn't power down. Last thing it prints during powerdown is:
> 
> hwsleep-0178 [-24] Acpi_enter_sleep_state: Entering S5
> 
> I can't find any specific A7V workarounds in 2.5.66 ACPI code, so I guess
> the ACPI code in 2.4 isn't up-to-date.
> 
> The original poster's problem is then probably indeed related to a buggy BIOS
> if it doesn't even powerdown with APM.

The original poster said his kernel was 2.4.2-2, which looks like a 
RedHat one and could have contained ACPI...

Regards,
Carl-Daniel

Re: poweroff problem

[Alan Cox]

On Sul, 2003-04-06 at 19:20, Carl-Daniel Hailfinger wrote:
> > I can't find any specific A7V workarounds in 2.5.66 ACPI code, so I guess
> > the ACPI code in 2.4 isn't up-to-date.
> > 
> > The original poster's problem is then probably indeed related to a buggy BIOS
> > if it doesn't even powerdown with APM.
> 
> The original poster said his kernel was 2.4.2-2, which looks like a 
> RedHat one and could have contained ACPI...

No Red Hat production kernel so far contains ACPI. As the reliability improves
and the quality of BIOS ACPI code rises that may change.

2.4.2 is also obsolete, and about five erratas back including security
ones.