P3 booting from SATA PCI problem \ VOGONS

P3 booting from SATA PCI problem

Topic actions

First post, by Lostdotfish

Posted on 2024-06-12, 22:52

Lostdotfish Offline

Rank Member

Rank: Member
Posts: 252
Joined: 2019-08-17, 11:49

I'm tinkering with a Gigabyte GA-6VXE7+ (Apollo 133 Pro) system. It's a Windows 98 system.

I added a Sil3114 PCI card and flashed it with the non-raid BIOS.

I booted into GParted via the Ultimate Boot CD and could see my SSD attached to the PCI card. I set the partition (127GB) and then cloned my IDE W98 install over to it .

In the BIOS I have set boot order to SCSI, A, C but I cannot get it to boot from the SATA drive.

I thought maybe I'd messed up the cloning process, so tried to do a fresh install from CD. Upon booting the CD it does not see the SATA drive.

If I boot into my IDE Windows 98 install, it doesn't see the SATA drive.

Boot back into GParted and I can see it fine. I can browse the cloned folder structure and files just fine.

Any ideas what I'm missing in order to get this system to boot from the SATA drive?

Reply 1 of 1, by jakethompson1

Posted on 2024-06-12, 23:28

jakethompson1 Offline

Rank l33t

Rank: l33t
Posts: 2037
Joined: 2015-11-17, 04:16

I have one of those cards also. I have used one on my SE440BX-2 which uses an Intel-branded Phoenix BIOS, and it works perfectly.
I also have an ECS P5HX-A board (a Socket 7). It uses Award 4.51PG and I had the same symptoms you mention. It did have only "A, C" and "C, A" as boot order in addition to the CDROM variants (SCSI is not mentioned). I assumed that might be the reason. But since you are having the same problem, maybe not.
Like you, the SIL311x option ROM BIOS executed as it should, but when booting, I got a "no boot device" style message and it was coming from the Award BIOS, not the SATA BIOS.

I was able to trace down the issue. Somehow in the initialization of the BIOS and option ROM, the number of fixed disks count in the BIOS data area, at 40:75, remained zero. I wrote a boot sector for a floppy disk in assembly that updated that count to one, then waited for you to eject a disk and press a key after which it called the BIOS boot vector again. It worked fine. But you're stuck doing that on every boot. I gave up not just because of this issue, but I was also annoyed that POST took a lot longer when no IDE drives were connected.

In a private conversation with mkarcher, he suggested that the SE440BX-2 Phoenix BIOS was more on the forefront of things and already implemented the BIOS Boot Specification, a more advanced way to boot from RAID controllers, PXE boot, and the like rather than having a chain of option ROMs hooking Int 13h and Int 19h. The suspicion was the SATA option ROM's fallback path for a non-BBS wasn't tested thoroughly and doesn't work, or Award BIOS 4.51 is missing or has a broken implementation of BBS.

Go to top of page Go to top of page

Back to General Old Hardware