Re: 10.0 DMA Issues With SIS-5597/5598 Chipset

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Thu, Jun 13, 2024 at 09:58:34AM -0400, Alexander Jacocks wrote:
> 
> > On Jun 13, 2024, at 8:44?AM, Manuel Bouyer <bouyer%antioche.eu.org@localhost> wrote:
> > 
> > ?On Thu, Jun 13, 2024 at 08:39:37AM -0400, Alexander Jacocks wrote:
> >> 
> >>>> On Jun 13, 2024, at 5:51?AM, Manuel Bouyer <bouyer%antioche.eu.org@localhost> wrote:
> >>> 
> >>> ?On Wed, Jun 12, 2024 at 10:29:59AM -0400, Alexander Jacocks wrote:
> >>>>  And it was pointed out to me that my paste of the DMA error was incorrect.
> >>>>  Here is the correct log:
> >>>>  [     1.000004] pci0 at mainbus0 bus 0: configuration mode 1
> >>>>  [     1.000004] pci0: i/o space, memory space enabled, rd/line, rd/mult,
> >>>>  wr/inv ok
> >>>>  [     1.000004] pchb0 at pci0 dev 0 function 0: Silicon Integrated System
> >>>>  5597/5598 Host Bridge (rev. 0x10)
> >>>>  [     1.000004] pcib0 at pci0 dev 1 function 0: Silicon Integrated System
> >>>>  85C503 or 5597/5598 ISA Bridge (rev. 0x01)
> >>>>  [     1.000004] siside0 at pci0 dev 1 function 1
> >>>>  [     1.000004] siside0: Silicon Integrated Systems 5597/5598 IDE
> >>>>  controller (rev. 0xd0)
> >>>>  [     1.000004] siside0: bus-master DMA support present, but unused
> >>>>  (couldn't map registers)
> >>> 
> >>> The problem here is that the DMA registers are not mappable, so can't be
> >>> used by the driver. Some BIOS settings (like PnP support) may affect this.
> >>> 
> >>>>  [     1.000004] siside0: secondary channel ignored (disabled)
> >>>>  I tried simply forcing DMA on by adding the 0x0001 flag to the pciide
> >>>>  driver, with a minor mod to the GENERIC kernel, but it didn't change the
> >>>>  behavior any. I also attempted to add in a Maxtor SATA-150 PCI IDE
> >>> 
> >>> Yes, that won't work in this case.
> >> 
> >> Interesting. I was able to verify that Windows 98SE is able to enable DMA support, even with the current BIOS settings. This BIOS is pretty limited in what can be configured, unfortunately.
> > 
> > From memory, win98 was the "PnP OS". it had to be disabled for others.
> > 
> >> 
> >> Would disabling ACPI and enabling APM possibly help? Since this board clearly has a broken ACPI configuration, that seems like a possibility.
> > 
> > No APM is something different
> > 
> 
> And PNP is currently disabled, as much as is possible. Is there anything else that I can try to do to make the registers mappable?

Can you send the output of a pcictl dump for this device ?

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--