Re: 2.6.0-test4 acpi problems

Re: 2.6.0-test4 acpi problems

[Andrew de Quincey]

On Saturday 30 Aug 2003 10:51 pm, Claas Langbehn wrote:
> To: kernel list <linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>
>
> Hello!
>
>
> I have got an Epox 8K9A9i mainboard with an Athlon XP 1800+ CPU. The
> chipset is an VIA KT400A. The bios is dated 12.05.2003.
> When booting with ACPI enabled, the interrupts dont get enumerated
> properly. with acpi=ht it works, but i dont have enough features.
> Below you see my bootlog with acpi and further down the bootlog with
> acpi=off. Is there a way to use full acpi support and "normal"
> interrupts? pci=noacpi was not really successful.
>
> How can I help acip development and do more debugging?

Aha, this is a known problem with the KT400 chipset. All BIOSes I've 
encountered for the KT400 have a nasty bug in the AML code which breaks ACPI 
IRQ routing.

The only way to make it work is to directly program the hardware itself. Via 
are sending me docs on the KT400, so I will add appropriate code to 
arch/i386/pci/irq.c for this board. As far as I can make out this is how 
Windows makes ACPI work on this board... it has specific hardware support for 
certain chipsets.

Hopefully I'll have something sometime next week.....



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: 2.6.0-test4 acpi problems

[Carlos Morgado]

On Sun, 31 Aug 2003 02:19:41, Andrew de Quincey wrote:

> Aha, this is a known problem with the KT400 chipset. All BIOSes I've
> encountered for the KT400 have a nasty bug in the AML code which breaks  
> ACPI IRQ routing.
>

As far as I can tell my asus a7v8x only doesn't enumerate the ide irqs.


PCI: Probing PCI hardware (bus 00)
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.PCI1._PRT]
PCI: Probing PCI hardware
 pci_irq-0297 [15] acpi_pci_irq_derive   : Unable to derive IRQ for device  
00:11.1
PCI: No IRQ known for interrupt pin A of device 00:11.1 - using IRQ 255
PCI: Using ACPI for IRQ routing

--
Carlos Morgado - chbm(a)ma.ssive.net - http://chbm.net/
0x1FC57F0A FP:0A27 35D3 C448 3641 0573 6876 2A37 4BB2 1FC5 7F0A


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: 2.6.0-test4 acpi problems

[Andrew de Quincey]

On Sunday 31 August 2003 01:35, Carlos Morgado wrote:
> On Sun, 31 Aug 2003 02:19:41, Andrew de Quincey wrote:
> > Aha, this is a known problem with the KT400 chipset. All BIOSes I've
> > encountered for the KT400 have a nasty bug in the AML code which breaks
> > ACPI IRQ routing.
>
> As far as I can tell my asus a7v8x only doesn't enumerate the ide irqs.

Can you send me a complete dmesg from a failed boot please? Also, an lspci 
-vvv, and /proc/acpi/dsdt from a successful boot.



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: 2.6.0-test4 acpi problems

[Alan Cox]

On Sul, 2003-08-31 at 01:35, Carlos Morgado wrote:
> On Sun, 31 Aug 2003 02:19:41, Andrew de Quincey wrote:
> 
> > Aha, this is a known problem with the KT400 chipset. All BIOSes I've
> > encountered for the KT400 have a nasty bug in the AML code which breaks  
> > ACPI IRQ routing.
> >
> 
> As far as I can tell my asus a7v8x only doesn't enumerate the ide irqs.

If the IDE port is in legacy mode (as VIA KT400 ones almost certainly
will be) we shouldnt be asking ACPI anyway as the answer is already
known (14 or 15)



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: 2.6.0-test4 acpi problems

[Andrew de Quincey]

On Sunday 31 August 2003 14:17, Alan Cox wrote:
> On Sul, 2003-08-31 at 01:35, Carlos Morgado wrote:
> > On Sun, 31 Aug 2003 02:19:41, Andrew de Quincey wrote:
> > > Aha, this is a known problem with the KT400 chipset. All BIOSes I've
> > > encountered for the KT400 have a nasty bug in the AML code which breaks
> > > ACPI IRQ routing.
> >
> > As far as I can tell my asus a7v8x only doesn't enumerate the ide irqs.
>
> If the IDE port is in legacy mode (as VIA KT400 ones almost certainly
> will be) we shouldnt be asking ACPI anyway as the answer is already
> known (14 or 15)

Does linux support non-legacy mode IDE? I've noticed in Windows XP my IDE 
devices appear as one, with a a single >15 IRQ, but in linux, it autodetects 
IRQs 14/15.



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: 2.6.0-test4 acpi problems

[Alan Cox]

On Sul, 2003-08-31 at 18:28, Andrew de Quincey wrote:
> Does linux support non-legacy mode IDE? I've noticed in Windows XP my IDE 
> devices appear as one, with a a single >15 IRQ, but in linux, it autodetects 
> IRQs 14/15.

Linux uses the setup the BIOS prefers. You have to know a lot about the
board to change it and many boards don't wire the chips for both. 



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

RE: 2.6.0-test4 acpi problems

[Holger W.]

> Aha, this is a known problem with the KT400 chipset. All BIOSes I've 
> encountered for the KT400 have a nasty bug in the AML code 
> which breaks ACPI IRQ routing.

AFAIK the KT333 chipset has got the same problem. Could you please integrate this chipset into your fix ? Would be really nice :-)

--

Regards,

HOLGER



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: 2.6.0-test4 acpi problems

[Andrew de Quincey]

On Sunday 31 August 2003 14:23, Holger W. wrote:
> > Aha, this is a known problem with the KT400 chipset. All BIOSes I've
> > encountered for the KT400 have a nasty bug in the AML code
> > which breaks ACPI IRQ routing.
>
> AFAIK the KT333 chipset has got the same problem. Could you please
> integrate this chipset into your fix ? Would be really nice :-)

Can you send me a DSDT dump so I can show them theres an issue there as well? 



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf