XP is very slow on coreboot with SeaBIOS

[Jiang Wang]

Hi there,

My colleagues and me tested Windows XP SP3 on coreboot and SeaBIOS on
ASUS M2V MX-SE motherboard. The XP is very slow when opening a new
window (such as Visual Studio) and when suspending and waking up. We
tried original AMI bios, which is much faster. Also, we noticed that
on coreboot and SeaBIOS, interrupts some times consume about 20-60%
CPU usage, which is very strange.

We tried two coreboot version: 6179 and 6460. Both are slow. The
SeaBIOS version is 0.6.1

I am not sure whether it is coreboot or SeaBIOS problem. Any suggestions?

Regards,

Jiang

[Rudolf Marek]

Hi,

Do you use sata or IDE? Maybe IDE  does not have DMA working in Windows. If you 
use IDE try to find out if PIO is used or DMA.

Thanks
Rudolf

[Jiang Wang]

Hi Rudolf,

We use SATA hard disk. Maybe something else is wrong? Thanks.

Regards,

Jiang

On Fri, Apr 1, 2011 at 3:54 PM, Rudolf Marek  wrote:
> Hi,
>
> Do you use sata or IDE? Maybe IDE  does not have DMA working in Windows. If
> you use IDE try to find out if PIO is used or DMA.
>
> Thanks
> Rudolf

[Scott Duplichan]

]Hi Rudolf,
]
]We use SATA hard disk. Maybe something else is wrong? Thanks.

Though your hardware connection is SATA, the software interface is
probably IDE. WinXP has no in-box AHCI driver for SB700 or SB800.
To make WinXP use AHCI with AMD requires pressing F6 during setup
and giving it a driver floppy.
Check device manager for a DMA mode such as 'Ultra DMA mode 6'
shown in the attached image.

Thanks,
Scott

]Regards,
]
]Jiang
]
]On Fri, Apr 1, 2011 at 3:54 PM, Rudolf Marek  wrote:
]> Hi,
]>
]> Do you use sata or IDE? Maybe IDE  does not have DMA working in Windows. If
]> you use IDE try to find out if PIO is used or DMA.
]>
]> Thanks
]> Rudolf

[Rudolf Marek]

Yeah Scott is right, although VIA SATA is not an AHCI. Please check if you see 
DMA there. If not maybe VIA drivers will help.

Thanks,
Rudolf

[Jiang Wang]

I see. We checked DMA for the hard disk controller and it is not
there. We are searching for VIA chipset drivers.

I am still curious why the original AMI BIOS can support Windows
(probably with DMA) better. Maybe AMI BIOS enables the DMA itself, so
it does not need any driver?

Regards,

Jiang

On Fri, Apr 1, 2011 at 5:37 PM, Rudolf Marek  wrote:
> Yeah Scott is right, although VIA SATA is not an AHCI. Please check if you
> see DMA there. If not maybe VIA drivers will help.
>
> Thanks,
> Rudolf

[Rudolf Marek]

Hi

Most likely because of the ACPI IDE methods which tell OS how to do DMA without 
the driver. I did not need them for Linux so I did not bother to implement them. 
I thit it is like GTM etc check ACPI spec for details. Maybe it can be done 
static way just supply right methods.

Thanks
Rudolf

[Jiang Wang]

I see. Also, we installed VIA SATA driver on XP and the problem is
solved. Thank you all.

Regards,

Jiang

On Fri, Apr 1, 2011 at 5:47 PM, Rudolf Marek  wrote:
> Hi
>
> Most likely because of the ACPI IDE methods which tell OS how to do DMA
> without the driver. I did not need them for Linux so I did not bother to
> implement them. I thit it is like GTM etc check ACPI spec for details. Maybe
> it can be done static way just supply right methods.
>
> Thanks
> Rudolf