4 messages in org.freebsd.freebsd-hardwareSupport of VIA Apollo Pro in 6.x is c...| From | Sent On | Attachments |
|---|---|---|
| Rashid N. Achilov | Oct 19, 2007 2:23 am | |
| Mark Kirkwood | Oct 21, 2007 8:21 pm | |
| Rashid N. Achilov | Oct 22, 2007 3:08 am | |
| Mark Kirkwood | Oct 22, 2007 9:28 am |
| Subject: | Support of VIA Apollo Pro in 6.x is completely broken? | |
|---|---|---|
| From: | Mark Kirkwood (mar...@paradise.net.nz) | |
| Date: | Oct 21, 2007 8:21:12 pm | |
| List: | org.freebsd.freebsd-hardware | |
Rashid N. Achilov wrote:
I have some boxes, built on motherboards Gigabyte GB-6VA+ (VIA Apollo Pro chipset). Do not say me, that's bullshit, I know :-/
Some of these boxes works under 4.5-RELEASE, some under Windows 2000. Sometime it should be upgraded on to 6.x branch, but I cannot use it - when I install 6.x on these boxes, I cannot bring up any network interface - link leds on card brights, but no any traffic can pass, even on neighbour box! I see
xl0: watchdog timeout xl0: watchdog timeout ...
messages on console.
Later. When connect HDD to motherboard through ordinary (40-wired) cable, 6.x says, that HDD is broken! I see endless messages
DMA_ERROR..., CRC_ERROR... etc
When I change cable to UDMA (80-wired) - all these messages are disapper
Later. When I boot box with "ACPI disabled" network cards works! But box hangs up after about 5-10 minutes of active work. Cooler does not stop but box hangs completely, only reset works.
Hmm interesting - I couldn't actually find any specs for a G*B*-6VA+ (only a GA-6VA+)... is it a VIA Apollo Pro 133A chipset with slot-1 P-II/III?
If so I've run 6.x (a while ago so probably 6.0 or 6.1) on a Tyan Trinity 400 (same chipset and cpu support)... and I used a 3Com office-connect ethernet (xl based) card....without any issues. Maybe see if Gigabyte have a more recent BIOS update for the board and update one of them and re-test!
On the ATA cable front, assuming the chipset is as I guessed, the ATA controller is ATA-66 capable, so yeah, best to use an 80-wire cable and get *some* IO speed out of these old things!
Cheers
Mark