Hi.

The problem outline: how to fix a broken DSDT table when the one provided by BIOS is simply an ill-fated joke?

As many of you know, unfortunately, the new boards of Asus have more or less been Kafkaesque for many linux users. Nor has ACPI been an easy issue with any OS.

I am using A8N-VM board but the basic puzzle has been same with related models as well. Asus has not provided any fix for the problem. (This is arrogant stance as the problem _can not be_ troublesome for a motherboard developer.) The newest BIOS for my board is 0702. I want to use the new BIOS, which cancels out the use of available precomplied fixes (e.g. acpi.sourceforge.net supply only a fix for the 0506 BIOS).

My testing distribution has so far been Suse 10.0 x86_64 with 2.6.13-15.8 kernel (and with the inbuilt ACPI support), although the problem is hardly a distro-specific. The weird thing, however, is that with the common "noacpi noapic" etc. Suse could not boot at all. When ACPI is on, the system booted nicely, all nForce/Nvidia stuff was recognized and even the onboard LAN/sound works fine with the proper patches. Whatsoever, ACPI seem to cause most of, if not all, the stability problems / lockouts that have occured so far.

With the <dmesg | grep -i acpi>, one gets the ah-so-familiar-newsgroup-output:

ACPI: Looking for DSDT in initrd... not found!
ACPI-0320: *** Error: ns_search_and_enter: Bad character in ACPI Name: 43045350
ACPI-0304: *** Error: Looking up [0x43045350] (NON-ASCII)
ACPI-0622: *** Warning: During name lookup/catalog, AE_BAD_CHARACTER
ACPI-0127: *** Error: acpi_load_tables: Could not load namespace: AE_BAD_CHARACTER
ACPI-0136: *** Error: acpi_load_tables: Could not load tables: AE_BAD_CHARACTER

Thus, the problem must lie somewhere in the 0x50, 0x53, 0x04, and 0x43. ACPI is completely new to me so I do not really know where one should start to debug. Yet, with the basic <acpidump -t DSDT -o dsdt_output>, one should get the DSDT table from the memory, which can be scanned for the bad point. Running <od -tx1 dsdt_output | grep -B1 '50 53' | more>, gives, to my understanding, two hints:

0065020 35 32 20 35 33 20 20 4c 4e 49 43 30 4c 2e 2f 49
0065040 47 50 53 2e 43 52 53 0a 20 20 31 36 66 30 3a 20

or

0103240 62 20 38 32 20 34 65 20 30 39 20 34 66 20 34 64
0103260 20 35 33 20 20 2e 2e 2e 50 53 2e 43 2e 2e 5b 2e

or what?

Now before I proceed (or start to do anything with the kernel!) some help would be more than welcome. Or even better, give me a precomplied hex table. After all, the issues with ACPI and DSDT should be a common research area for all linux users trying to get new machinery to do their job.

Thanks.

b.

Re-visit the acpi.sourceforge.net site, they have nice helpful guides to help you fix your broken ACPI.

This one help me the most; http://forums.gentoo.org/viewtopic.php?t=122145

And according to http://gaugusch.at/kernel.shtml once you have your custom DSDT follow the howto found here;
http://acpi.sourceforge.net/wiki/ind...ustomDsdtTable

Hi.

Tried to fix the issue for few months but with no success. But I was not alone. Anyone I know from anywhere and from any instance could not get the ACPI working under this mobo.

Yet, small victories (some people put forward an email flooding protest against Asus), and finally the company was able to fix few hex codes in their BIOS. Thus, the new 0902 BIOS fixes the issue.

I found a site that has a file you can patch into your kernel that fixed the power shutdown problem on my Asus A8N-VM CSM board.

Go here and download his patch file.
quaggaspace.org/a8nvm/

Follow his directions as best you can. He leaves out the commands to actually compile, but other than that, this was a fix for me!

Eric

Hi.

Actually contacted few months ago the maintainer of that page for instructions but could not get it to work on my own. Luckily, like said in the above message, the new BIOS corrected the DSDT table. Instead of using some kernel hack to allow shutdown, the new table allows one to even examine other ACPI features aswell.

Thanks anyhow.

Gloomy, you have a A8N-VM board rather than a A8N-VM CSM board?

For the A8N-VM CSM, there is no verision 0902 BIOS on the Asus site, so A8N-VM CSM users are left to apply the DSDT patch.

After I got the ACPI/DSDT working with the patch, I updated the Asus BIOS to version 0702 and it still works.

I can't be certain if Asus version 0702 fixes the DSDT or not. I seem to recall reading forum posts that 0702 does not fix the ACPI/DSDT so the patch is the path for now...

Eric

Yes,

I have A8N-VM, not the CSM.

I am rather certain, following e.g. the useful Nvnews threads, like the one below, that the 0702 did not fix the DSDT of either boards. And release notes of 0702 for CSM do not mention anything about ACPI. Nor did the 0902 for A8N-VM mention anything at all - again strikingly unprofessional from Asus.

http://www.nvnews.net/vbulletin/showthread.php?p=853079

But I guess a new BIOS is coming soon for the CSM aswell. Thus, wait and hope.


Btw I might ask have you been able to get the sensors working? Those are the last unworkable part of this damned board on my setup.

I have not tried to get the sensors to work yet. I'm not sure how to do that.

Asus has fixed the ACPI/DSDT with the newest official version of the A8N-VM CSM bios (ver. 1001)!

Eric

Just in case people hadn't noticed, ASUS released a new bios (1001) today. Since their previous beta bios (901) fixed the DSDT issue (and incidentially broke USB, net, and powersave) this is certainly worth a spin.

I'll try it when I get home - I don't know how to update the bios from within linux so have to boot windows. Maybe I will finally have a working computer!

And, since I seem to be in dreaming mode... maybe this will mean I can boot with both my ivtv and my DVB-S cards plugged in _at the same time_ without freezing the machine! Nah... it'll never happen ;-)

Asus has a built-in BIOS flash utility. You don't need a bootable disk at all. When the board powers up you hit ALT+F2 and it looks on the floppy disk for a file called A8NVMCMS.ROM and loads it into the flash memory. I heard that this also works from a cd-rom with a A8NVMCMS.ROM file on it.

They floppy disk method (OS-free) is detailed in the manual:
http://dlsvr03.asus.com/pub/ASUS/mb/...a8n-vm_csm.pdf

Eric

Thanks guys, I'll try the new BIOS after the labour day is over. (Do not drink and play with the BIOS.)

About the sensors. You have to enable I2C and hardware monitoring support from the kernel. The sensors are, to my understanding, Winbond W83627EHG, supported by the preliminary code for W83627EHF. However, when I type 'sensors-detect' in a Gentoo system, all I get is that "No i2c device files found".

Nevertheless, despite of the lack of voltage values, the W83627EHF gives the right values for MB temp, CPU temp, processor fan and for chassis fan. Or it gave the right values with the 0702 BIOS - the 0902 gives about 10C lower temperature than the ones in the BIOS. The used algorithms are of course easy to modify by hand but haven't got the time or interest yet.

If one wants more information, the page of lm_sensors might be worth reading:

http://secure.netroedge.com/~lm78/kernel26.html