A7M266-D works?

A7M266-D works?

[Jussi Laako]

Hi,

I'm considering to buy ASUS A7M266-D mobo and wondering if there is fix for
the APIC problems or is there workaround to get that mobo work? If it not
currently, is there some estimate if it will work in near future?


	- Jussi Laako

-- 
PGP key fingerprint: 161D 6FED 6A92 39E2 EB5B  39DD A4DE 63EB C216 1E4B
Available at PGP keyservers

Re: A7M266-D works?

[Alan Cox]

> I'm considering to buy ASUS A7M266-D mobo and wondering if there is fix for
> the APIC problems or is there workaround to get that mobo work? If it not
> currently, is there some estimate if it will work in near future?

The current status on the board I'm using

-	The BIOS appears to misconfigure the PCI setup badly, so badly I've
	been sticking in PCI quirk fixups to make some drivers work
-	MP 1.4 locks solid MP 1.1 is ok - that may be a Linux bug. I have a
	patch to test there
-	The onboard USB was removed and swapped for a plug in card wasting
	a slot (but you do get USB2 now). The manual wasn't updated
-	The disk led appears to be marked wrongly. If you wire it the 
	board won't POST
-	If you put a 33Mhz card in the 64bit slots the board won't POST
	unless you configure the board the way the manual suggests you don't
-	Expect to want > 400W PSU if you have lots of disks. I ended up 
	using a 550W PSU
-	Use a _BIG_ case. With anything but a large tower and actual
	case fans you are not going to get enough ventilation or cable
	room

Having said all that - when you figure through the mess that ASUS calls
QA and documentation (and they should be firing people for shipping it in
its current state quite frankly) it goes like a bat out of hell. With dual
Athlon 1800MP+ processors and hardware raid for the filestore and swap its
building the entirity of XFree86 in 30 minutes wall time (5 of which are
doing non parallelisable work and 5 or so of which are spent in I/O wait)

Nice hardware, lousy docs, dreadful BIOS.

Re: A7M266-D works?

[Martin Josefsson]

On Mon, 11 Feb 2002, Alan Cox wrote:

> The current status on the board I'm using
> 
> -	The BIOS appears to misconfigure the PCI setup badly, so badly I've
> 	been sticking in PCI quirk fixups to make some drivers work

Could you give some examples of which drivers/cards you've seen problems
with. We've been thinking of getting a few of these boards but if there's
this much problems with the BIOS we might have to look at another board :(

I just read on Asus site that this board actually has been approved by
AMD.

"The A7M266-D is the first 760MPX chipset based motherboard to pass all of
AMD's validation requirements, ensuring superior reliability, performance
and quality."

It must have been running an unreleased BIOS at the time of the testing :)

> -	MP 1.4 locks solid MP 1.1 is ok - that may be a Linux bug. I have a
> 	patch to test there

Please test that patch and let us now.

/Martin

Never argue with an idiot. They drag you down to their level, then beat you with experience.

Re: A7M266-D works?

[Alan Cox]

> Could you give some examples of which drivers/cards you've seen problems
> with. We've been thinking of getting a few of these boards but if there's
> this much problems with the BIOS we might have to look at another board :(

I2O fails for one. With the little patch I added I2O seems to work again.
Im just waiting a final AMD confirmation that the docs not the bios are
correct.

Alan

Re: A7M266-D works?

[Maksim Krasnyanskiy]

> > I'm considering to buy ASUS A7M266-D mobo and wondering if there is fix for
> > the APIC problems or is there workaround to get that mobo work? If it not
> > currently, is there some estimate if it will work in near future?
>
>The current status on the board I'm using
>
>-       The BIOS appears to misconfigure the PCI setup badly, so badly I've
>         been sticking in PCI quirk fixups to make some drivers work
Which board rev are you using ?
I have 1.03. No problem with PCI so far. All cards that I tried worked just 
fine.

>-       MP 1.4 locks solid MP 1.1 is ok - that may be a Linux bug. I have a
>         patch to test there
Same for me.

>-       The onboard USB was removed and swapped for a plug in card wasting
>         a slot (but you do get USB2 now). The manual wasn't updated
I replaced that ASUS USB 2.0 card with Adaptec Duo Connect (USB 2.0 / 1394).
Works just fine.

>-       The disk led appears to be marked wrongly. If you wire it the
>         board won't POST
Yeah. You should use another one under JEN jumper.

>-       If you put a 33Mhz card in the 64bit slots the board won't POST
>         unless you configure the board the way the manual suggests you don't
True. It should be Jumperless + DIP SW off/off/off/off.

>-       Expect to want > 400W PSU if you have lots of disks. I ended up
>         using a 550W PSU
I use EnerMax 430W. IBM 40GB + WD 10GB + HITACHI DVD + IOMEGA CDRW.
Works perfectly (with the patch from Linux-IDE folks, otherwise DMA doesn't 
work).

>-       Use a _BIG_ case. With anything but a large tower and actual
>         case fans you are not going to get enough ventilation or cable
>         room
I have minitower in-win case (with the case fan). And I'm very happy with it.
CPU and chipset temperatures are  just fine (41C - 51C).

>Having said all that - when you figure through the mess that ASUS calls
>QA and documentation (and they should be firing people for shipping it in
>its current state quite frankly) it goes like a bat out of hell.
I was pretty happy with packaging.
And yes it's super fast and reliable.

Max

Re: A7M266-D works?

[Alan Cox]

> >-       The BIOS appears to misconfigure the PCI setup badly, so badly I've
> >         been sticking in PCI quirk fixups to make some drivers work
> Which board rev are you using ?
> I have 1.03. No problem with PCI so far. All cards that I tried worked just 
> fine.

Check PCI register 0x4C if bits 1 and 2 are clear your board is not running
in a PCI compliant mode and anything may happen. In paticular since memory
and PCI ordering is not preserved you may see corruption and failures. Most
devices don't have that dependancy but a few do - and break horribly.

Alan

Re: A7M266-D works?

[Bruce Harada]

On Mon, 11 Feb 2002 15:04:57 -0800
Wayne Whitney <whitney@math.berkeley.edu> wrote:

> Lastly, do you know whether the reason that the A7M266-D comes with a
> PCI USB2 card is that the USB support of the AMD768 southbridge is
> borked?  Both the Tyan S2466 and the MSI K7D Master come with PCI USB
> cards.

ASUS and Tyan both delayed shipping their respective boards at the same time
so yes, it seems that way. According to a Japanese site, ASUS told its agents
that "a bug in the AMD768 Southbridge makes on-board USB unusable, and USB2.0
cards will be bundled with all such boards for the foreseeable future". 

Re: A7M266-D works?

[Wayne Whitney]

Hello, on the linux-kernel mailing list, you wrote:

> > > The BIOS appears to misconfigure the PCI setup badly, so badly
> > > I've been sticking in PCI quirk fixups to make some drivers work
>
> Check PCI register 0x4C if bits 1 and 2 are clear your board is not
> running in a PCI compliant mode and anything may happen. 

The PCI quirk fixup you posted (reproduced below) uses the test
"(pcic&6)!=6", so do you mean to say "if bit 1 or bit 2 is clear"?

> In paticular since memory and PCI ordering is not preserved you may
> see corruption and failures. Most devices don't have that dependancy
> but a few do - and break horribly.

Might this cause random hard lockups under a compute intensive load?

Since the fixup applies to the AMD762 northbridge, common to the 760MP
and 760MPX chipsets, this discussion applies to all SMP Athlon
motherboards at present, is that right?

Lastly, do you know whether the reason that the A7M266-D comes with a
PCI USB2 card is that the USB support of the AMD768 southbridge is
borked?  Both the Tyan S2466 and the MSI K7D Master come with PCI USB
cards.

Thanks,
Wayne

--- linux.18p8/drivers/pci/quirks.c	Wed Feb  6 06:44:36 2002
+++ linux.18p8-ac1/drivers/pci/quirks.c	Sun Feb 10 19:41:26 2002
@@ -444,13 +444,15 @@
 static void __init quirk_amd_ordering(struct pci_dev *dev)
 {
 	u32 pcic;
-	
-	pci_read_config_dword(dev, 0x42, &pcic);
-	if((pcic&2)==0)
+	pci_read_config_dword(dev, 0x4C, &pcic);
+	if((pcic&6)!=6)
 	{
-		pcic |= 2;
-		printk(KERN_WARNING "BIOS disabled PCI ordering compliance, so we enabled it again.\n");
-		pci_write_config_dword(dev, 0x42, pcic);		
+		pcic |= 6;
+		printk(KERN_WARNING "BIOS failed to enable PCI standards compliance, fixing this error.\n");
+		pci_write_config_dword(dev, 0x4C, pcic);
+		pci_read_config_dword(dev, 0x84, &pcic);
+		pcic |= (1<<23);	/* Required in this mode */
+		pci_write_config_dword(dev, 0x84, pcic);
 	}
 }
 

Re: A7M266-D works?

[Alan Cox]

> > Check PCI register 0x4C if bits 1 and 2 are clear your board is not
> > running in a PCI compliant mode and anything may happen. 
> 
> The PCI quirk fixup you posted (reproduced below) uses the test
> "(pcic&6)!=6", so do you mean to say "if bit 1 or bit 2 is clear"?

If either bit is clear then the board (according the the AMD 762 BIOS
manual from www.amd.com) is not in PCI compliant mode

> > see corruption and failures. Most devices don't have that dependancy
> > but a few do - and break horribly.
> 
> Might this cause random hard lockups under a compute intensive load?

Hard to say. 

> Since the fixup applies to the AMD762 northbridge, common to the 760MP
> and 760MPX chipsets, this discussion applies to all SMP Athlon
> motherboards at present, is that right?

I believe so. 

> Lastly, do you know whether the reason that the A7M266-D comes with a
> PCI USB2 card is that the USB support of the AMD768 southbridge is
> borked?  Both the Tyan S2466 and the MSI K7D Master come with PCI USB
> cards.

My guess too - but I don't know