RE: new n800v dsdt successfully disassembled, patched and install ed (WAS: disasm/compile doesn't work)

RE: new n800v dsdt successfully disassembled, patched and install ed (WAS: disasm/compile doesn't work)

[Moore, Robert]

We have seen the AE_ALREADY_EXISTS error before, but we have never been able
to track it down.  It will probably require a full debug trace of the ACPI
initialization.

The AE_AML_BUFFER_LIMIT is interesting.  It sounds like a Field is being
defined past the end of the buffer, but I would be interested in taking a
look at it.

Please send your AML/ASL and I'll see what I can find.

Thanks,
Bob Moore


-----Original Message-----
From: Bob Lees [mailto:bob-l+PWtdWbHAsG2Il/BtU0GPXRex20P6io@public.gmane.org] 
Sent: Tuesday, October 15, 2002 4:17 AM
To: Mark Pritchard; acpi mail list
Subject: Re: new n800v dsdt successfully disassembled, patched and installed
(WAS: [ACPI] disasm/compile doesn't work)

Hi Mark

Thanks for the response.  I think the 800c, 800v and presario 2800 series
are 
all the same in this area.  It was the external definitions which I was 
missing.  I now have fan control working - yea!

I also have akpi monitoring the battery and ac state.  It needed some
changes 
to the makefile to get it made, I was using akpi-0.3.2.

I still have to figure out how I broke processor performance.  Working on 
this.  I have changed a number of the C??? referernces to the correct
device, 
eg FAN1 etc.  This isn't complete yet.

When the new table is loaded I get a number of errors, which I am still
trying 
to track down.  eg 
evrgnini-0243: *** Error: Could not install Pci_config handler for PCI0, 
AE_ALREADY_EXISTS
where PCI0 is C03C in your version
and
dsopcode-0497 [19] Ds_init_buffer_field  : Field size 1184 exceeds Buffer
siz
e 1088 (bits)
psparse-1155: *** Error: Method execution failed, AE_AML_BUFFER_LIMIT

Have you seen anything like this?

Regards

Bob

On Tuesday 15 October 2002 00:07, Mark Pritchard wrote:
> Hi Bob,
>
> I've been playing with ACPI over the last week or so with my Compaq Evo
> N800v which is very similar to your N800c (in fact, the compaq website
> seems to only differentiate them by maximum processor speed, max hdd and
> screen type... http://www.compaq.com/products/notebooks/index.html)
>
> With some help from this group (and that email you wrote earlier), I've
> successfully disassembled and tweaked my dsdt. I now have battery status
> (which was my goal), but haven't mucked around with the other aspects of
> ACPI (fan, processor speed etc).
>
> The details of what I did to make it work:
>
> 1) 2.4.19 patched to 2.4.20pre8
>
> 2) applied acpi-20021002 2.4.20pre8 patch from
> http://sourceforge.net/projects/acpi (direct download link is
>
http://prdownloads.sourceforge.net/acpi/acpi-20021002-2.4.20-pre8.diff.gz?d
>ownload)
>
> 3) downloaded the iasl release from intel's site at
> http://developer.intel.com/technology/iapc/acpi/ with the direct download
> link being
>
http://developer.intel.com/technology/iapc/acpi/downloads/iasl-linux-200210
>02.tar.gz
>
> 4) disassembled my dsdt by:
>
> 	a) copying /proc/acpi/dsdt to a local file (dsdt.raw)
> 	b) running "iasl -d dsdt.raw" (this gives you the dsdt.dsl file you
tweak
> below)
>
> 5) Tweaked the dsl file by doing the following:
>
> 	a) removed all "*" characters from the object names
> 	b) added external declarations for all missing objects
> 	c) commented out the alias line which complained about the missing
link
> 	d) changed the store line which referenced the above alias line to
use the
> fully qualified path (I did try using the fqp on the alias, but no joy)
> 	e) changed the references from C11C to BAT1 and C11B to BAT2
>
> 6) Recompiled my dsdt with "iasl -tc -p new dsdt.dsl" to give me new.hex
> which is a C formatted representation of the compiled dsl (I ignored the
> two warnings).
>
> 7) Copied new.hex to drivers/acpi/tables/compaq_evo_n800v.h under my
kernel
> source tree.
>
> 8) Edited drivers/acpi/tables/tbgetall.c as follows:
>
> 	a) Included the new dsdt header
> 	b) Declared and set new ACPI_POINTER to hold a physical address
rather
> than the logical so I could pass it to acpi_tb_get_secondary_table()
> 	c) Changed the call to acpi_tb_get_secondary_table() to use the new
dsdt
>
> 9) Configured the kernel with Power Management and all ACPI (except
> toshiba) options selected. Screen shots included in tarball mentioned
> below.
>
> 10) Rebuilt the kernel, crossed my fingers and jumped around for joy when
I
> can now do:
>
> [pritchma@seth:pritchma]$ uname -a
> Linux seth.modus.com.au 2.4.20-pre8 #10 Mon Oct 14 14:39:22 EST 2002 i686
> unknown
> [pritchma@seth:pritchma]$ cat /proc/acpi/battery/BAT1/info
> present:                 yes
> design capacity:         4179 mAh
> last full capacity:      4179 mAh
> battery technology:      rechargeable
> design voltage:          14800 mV
> design capacity warning: 418 mAh
> design capacity low:     0 mAh
> capacity granularity 1:  100 mAh
> capacity granularity 2:  100 mAh
> model number:            Primary
> serial number:           3
> battery type:            4E4F494C
> OEM info:                 COMPAQ
> [pritchma@seth:pritchma]$ cat /proc/acpi/battery/BAT1/state
> present:                 yes
> capacity state:          ok
> charging state:          unknown
> present rate:            0 mA
> remaining capacity:      4179 mAh
> present voltage:         16615 mV
>
> Pull out the power:
>
> [pritchma@seth:pritchma]$ cat /proc/acpi/battery/BAT1/state
> present:                 yes
> capacity state:          ok
> charging state:          discharging
> present rate:            1400 mA
> remaining capacity:      4179 mAh
> present voltage:         16296 mV
>
> Looks good to me.
>
> I've uploaded a tarball to
> http://www.modus.com.au/scratch/compaq_n800v_acpi.tar.bz2
>
> It contains:
>
> * my original, changed and a diff of the disassembled dsdt
> * my original, changed and a diff of tbgetall.c
> * two screenshots of the kernel options I used
>
> Now for the disclaimer :) I knew nothing about ACPI a week ago, and mucked
> around with this until it worked. The above may be everything you need to
> know, or not...but I do hope it sorts your problem out.
>
> I'm now off to work out why kacpi and akpi both hang at startup...or write
> my own battery monitor!
>
> Cheers
>
> Mark Pritchard
>
> On Tue, 15 Oct 2002 06:47, Bob Lees wrote:
> > Hi
> >
> > This may have been asked/aired before, in which case I apologise. 
> > However here goes.
> >
> > I have sort of successfully disassembled/decompiled the dsdt from my evo
> > 800c, using both pacpidump and iasl.  They both produce the same
outcome.
> > By "sort of" I mean that I cannot disassemble and immdeiately recompile
> > without errors.  Some of them are basic to the compaq dsdt in that it
> > references *PNP???? rather than PNP????.  That is easy to fix.  The
other
> > problem is that objects are disassembed which have no definition.  This
> > is much harder as it is not clear how these undefined objects should be
> > defined.  Is this a common problem and does a reasonable fix exist? and
> > if so help!
> >
> > For the present I have commented out the offending references, but this
> > means that I have no fan control!! which may not be good in the long
> > term:)
> >
> > Regards
> >
> > Bob

-- 
Bob Lees
Diamond Consulting Services Ltd
Aylesbury, Bucks, HP17 8UG
Phone: +44 (0) 1296 747667
Fax: +44 (0) 1296 747557
email: bob-l+PWtdWbHAsG2Il/BtU0GPXRex20P6io@public.gmane.org


-------------------------------------------------------
This sf.net email is sponsored by: viaVerio will pay you up to
$1,000 for every account that you consolidate with us.
http://ad.doubleclick.net/clk;4749864;7604308;v?http://www.viaverio.com/
consolidator/osdn.cfm


_______________________________________________
Acpi-devel mailing list
Acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
https://lists.sourceforge.net/lists/listinfo/acpi-devel


-------------------------------------------------------
This sf.net email is sponsored by: viaVerio will pay you up to
$1,000 for every account that you consolidate with us.
http://ad.doubleclick.net/clk;4749864;7604308;v?
http://www.viaverio.com/consolidator/osdn.cfm

RE: new n800v dsdt successfully disassembled, patched and install ed (WAS: disasm/compile doesn't work)

[hugo hallqvist]

tis 2002-10-15 klockan 23.58 skrev Moore, Robert:
> 
> We have seen the AE_ALREADY_EXISTS error before, but we have never been able
> to track it down.  It will probably require a full debug trace of the ACPI
> initialization.
> 
> The AE_AML_BUFFER_LIMIT is interesting.  It sounds like a Field is being
> defined past the end of the buffer, but I would be interested in taking a
> look at it.
> 
> Please send your AML/ASL and I'll see what I can find.
> 
> Thanks,
> Bob Moore

Hi guys,

I think I have a fix for this.

If you have  a method called C04C at about line 495-500,
then add a little before you make a buffer out of it.
At least that is my interpretation of the code, 
with this fix I don't get the AML_BUFFER_LIMIT error anymore, 
however this may not be the right way<tm> to do it..?

//line 494
// addition
            Method (C04C, 0, NotSerialized)
            {
                Acquire (C042, 0xFFFF)
                Multiply (C043, 0x1A, Local1)
                Add (SizeOf (C046), Local1, Local2)
                Add(Local2, 0xB0, Local7)  //add a little to the size
                Store (Buffer (Local7) {}, Local0)  //then make a buffer
//                Store (Buffer (Local2) {}, Local0)    // original code
                Store (0x00, Local1)
                Store (0x00, Local2)

Also, would it be possible to get any of you guys final dsdt's for the
compaq evo n800v? I have messed mine up, have problems with Alias links 
and also the thermal zones/fans. :-(

Thanks in advance,

//Hugo


-------------------------------------------------------
This sf.net email is sponsored by: viaVerio will pay you up to
$1,000 for every account that you consolidate with us.
http://ad.doubleclick.net/clk;4749864;7604308;v?
http://www.viaverio.com/consolidator/osdn.cfm

Re: new n800v dsdt successfully disassembled, patched and install ed (WAS: disasm/compile doesn't work)

[Bob Lees]

[-- Attachment #1: Type: text/plain, Size: 3327 bytes --]

Hi All

Firstly, tried Hugo's suggestion and the AE_AML_BUFFER_LIMIT error goes away.  
Doesn't immediately appear to change any processing.

Attached is my latest dsl file.  Some notes:
1)  I have tried to make sense of a number of the C??? refs and change them to 
"standard(ish)" definitions
2)  As noted in the dsl file, the obvious one of changing C000 to CPU0 breaks 
the speedstep detection.  That doesn't make any sense.
3)  I can't change the processor speed by
<echo 1 > /proc/acpi/processor/C000/performance>
the following error is reported
acpi_processor-1095 [26] acpi_processor_set_per: Transition failed
This happens both before and after trying Hugo's fix
4)  On the AE_ALREADY_EXISTS problem, I have a thought.  The full error 
reported is
evrgnini-0243: *** Error: Could not install Pci_config handler for PCI0, 
AE_ALREADY_EXISTS

There appear to be 3 definitions of OperationRegion C0A1 in devices USB0, USB1 
and USB2 line 1195 onwards.  Could it be that these don't have the correct 
preamble appended by the compiler and so the interpreter sees them as the 
same object??

I hasten to add, that my understanding of ACPI programming is very 
rudermentary, to non-existent:)

Anyway Hugo, usual disclaimers apply:( it works on my machine, but ......

Bob hope some of this helps to track the problem down

For the record, installation  is SuSE 8.1 with a 2.4.19 SuSE kernel with the 
latest ACPI patches for this kernel from SourceForge.  ACPI debug is on and 
all APIC stuff turned off.  PCI interrupt handling is via ACPI, ie pci=noacpi 
is not used.

Regards

Bob Lees

On Tuesday 15 October 2002 23:47, hugo hallqvist wrote:
> tis 2002-10-15 klockan 23.58 skrev Moore, Robert:
> > We have seen the AE_ALREADY_EXISTS error before, but we have never been
> > able to track it down.  It will probably require a full debug trace of
> > the ACPI initialization.
> >
> > The AE_AML_BUFFER_LIMIT is interesting.  It sounds like a Field is being
> > defined past the end of the buffer, but I would be interested in taking a
> > look at it.
> >
> > Please send your AML/ASL and I'll see what I can find.
> >
> > Thanks,
> > Bob Moore
>
> Hi guys,
>
> I think I have a fix for this.
>
> If you have  a method called C04C at about line 495-500,
> then add a little before you make a buffer out of it.
> At least that is my interpretation of the code,
> with this fix I don't get the AML_BUFFER_LIMIT error anymore,
> however this may not be the right way<tm> to do it..?
>
> //line 494
> // addition
>             Method (C04C, 0, NotSerialized)
>             {
>                 Acquire (C042, 0xFFFF)
>                 Multiply (C043, 0x1A, Local1)
>                 Add (SizeOf (C046), Local1, Local2)
>                 Add(Local2, 0xB0, Local7)  //add a little to the size
>                 Store (Buffer (Local7) {}, Local0)  //then make a buffer
> //                Store (Buffer (Local2) {}, Local0)    // original code
>                 Store (0x00, Local1)
>                 Store (0x00, Local2)
>
> Also, would it be possible to get any of you guys final dsdt's for the
> compaq evo n800v? I have messed mine up, have problems with Alias links
> and also the thermal zones/fans. :-(
>
> Thanks in advance,
>
> //Hugo

[-- Attachment #2: bobs-dsl.tgz --]
[-- Type: application/x-tgz, Size: 18325 bytes --]