VOGONS

Common searches


First post, by Kahenraz

User metadata
Rank l33t
Rank
l33t

I'm trying to figure out why the installer is hanging here. Unfortunately, this is a Dell BIOS with very little configurable options.

I have disabled the onboard SATA controller to use a PCI IDE controller and a PCI S3 video card instead of the onboard GeForce 6150SE.

I know that I can disable ACPI by installing with the options "/p i", and this gets the installation to finish, but I would like to better understand what the actual problem is and whether it can be resolved some other way.

Edit:

After further research, I've found the problem to be the onboard USB controller. This is a problem, because this is the only way to connect a mouse and keyboard. There are no PS/2 ports on this motherboard. There isn't even a serial port or motherboard header for a mouse.

If I boot I to Safe Mode and remove this device, the system books and the mouse and keyboard is always responsive until it finishes installing. Subsequent reboots that are not into Safe Mode cause the system to hang.

I'm not sure how to work around this, given that I need these ports for mouse and keyboard during the boot process as well, since add-on cards don't work at this early stage of the boot process.

20231029_033602.jpg
Filename
20231029_033602.jpg
File size
997.9 KiB
Views
273 views
File license
CC-BY-4.0

Reply 1 of 2, by Kahenraz

User metadata
Rank l33t
Rank
l33t

I've been working on this for a while and found through trial and error that there are a lot of problems with this motherboard that manifest several issues that affect the ability to run Windows 98 acceptably.

ACPI support is completely broken and hangs the system during installation every time; the only way to install Windows 98 is in APM mode. The IRQ Tables provided by the BIOS have errors, causing IRQ steering for PCI devices to be broken, requiring each "important" PCI device to have a unique IRQ for the system to function.

I need the following devices at a minimum to have their own private IRQ:

* Video
* Sound
* USB (for mouse/keyboard)

Although PCI devices cannot have their IRQs assigned explicitly, they can be manipulated by enabling and disabling different devices in the BIOS. I found that it was possible to get a working system with the previously mentioned devices this way. However, this list does not include a Disk Controller, which means that DMA is missing, and this makes the system very unresponsive when heavy disk operations occur, as well as a degraded overall experience.

Why can't this be solved? I have come to the realization that, despite lots of other IRQs being available, the BIOS will always assign all four of these devices to IRQs 7, 10, and 11 only. The problem here is, specifically, that I want four devices to have a unique interrupt, but there are only ever 3 IRQs available to me. However, I do have a clever workaround.

While examining IRQs on the disk controllers, I found that one of them was always assigned IRQ 14, but that there was never any device attached to it on any of the onboard SATA ports. I have also discovered unpopulated pads for an IDE pin header on the motherboard. I think that if I solder a header onto it, that I will have a working disk on its own IRQ, giving me the 4 unique IRQs for a pleasant Windows 98 experience.

Will this idea work? I have no idea. But I'm crazy enough to try it and find out!

20231103_050733.jpg
Filename
20231103_050733.jpg
File size
872.05 KiB
Views
247 views
File license
CC-BY-4.0

Reply 2 of 2, by Kahenraz

User metadata
Rank l33t
Rank
l33t

I added a header for the onboard IDE controller, and it worked. However, it's not bootable and is only detected from within Windows. I thought that even if it wasn't bootable, I could still bootstrap it from GRUB4DOS, but both it and FreeDOS fail to detect the drive controller.

I wanted to boot Windows 98 from this IDE controller so that I could enable DMA on an IRA that didn't conflict with other devices. But since it's not bootable, then this project to get Windows 98 running on this system is a failure.

I found some information about the IDE drive controller here:

https://linux-hardware.org/index.php?id=pci:1 … -03ec-1019-2609

Here are some photos of the header attachment process, for your enjoyment.

20231120_083144.jpg
Filename
20231120_083144.jpg
File size
1.66 MiB
Views
146 views
File license
CC-BY-4.0
20231120_083213.jpg
Filename
20231120_083213.jpg
File size
1.29 MiB
Views
146 views
File license
CC-BY-4.0
20231120_090459.jpg
Filename
20231120_090459.jpg
File size
1 MiB
Views
146 views
File license
CC-BY-4.0
20231120_091102.jpg
Filename
20231120_091102.jpg
File size
936.55 KiB
Views
146 views
File license
CC-BY-4.0
20231120_092858.jpg
Filename
20231120_092858.jpg
File size
1.2 MiB
Views
146 views
File license
CC-BY-4.0