Silicon Image 0680 + 3114 on same system ?

[Andre Majorel]

I've bought a Connectland "carte PCI RAID ATA 133 0703014" dual
IDE controller (Silicon Image 0680, PCI 1095:0680 rev 02).
Unfortunately it conflicts with the Silicon Image 3114 on the
motherboard (PCI 1095:3114 rev 02).

If I disable the on-board 3114 in the BIOS, the 0680 works but I
lose 4 SATA ports. If I enable the 3114 in the BIOS, the PC
freezes shortly after power-on.

I'm running Linux. Is it possible to disable the 3114 in the
BIOS and have Linux re-enable it when it boots ?

The I/O ports for the 0680 and 3114 overlap somewhat but maybe
they're assigned dynamically ? I don't need the RAID BIOSes as
I'm doing software RAID only.

0680 :
9c00-9c0f : 0000:01:07.0
a000-a003 : 0000:01:07.0
a400-a407 : 0000:01:07.0
a800-a803 : 0000:01:07.0
ac00-ac07 : 0000:01:07.0

3114 :
9800-980f : 0000:01:0d.0
9c00-9c03 : 0000:01:0d.0
a000-a007 : 0000:01:0d.0
a400-a403 : 0000:01:0d.0
a800-a807 : 0000:01:0d.0

 

[Arno Wagner]

I've bought a Connectland "carte PCI RAID ATA 133 0703014" dual
IDE controller (Silicon Image 0680, PCI 1095:0680 rev 02).
Unfortunately it conflicts with the Silicon Image 3114 on the
motherboard (PCI 1095:3114 rev 02).

If I disable the on-board 3114 in the BIOS, the 0680 works but I
lose 4 SATA ports. If I enable the 3114 in the BIOS, the PC
freezes shortly after power-on.

I'm running Linux. Is it possible to disable the 3114 in the
BIOS and have Linux re-enable it when it boots ?

Not as far as I know.

The I/O ports for the 0680 and 3114 overlap somewhat but maybe
they're assigned dynamically ? I don't need the RAID BIOSes as
I'm doing software RAID only.

0680 :
9c00-9c0f : 0000:01:07.0
a000-a003 : 0000:01:07.0
a400-a407 : 0000:01:07.0
a800-a803 : 0000:01:07.0
ac00-ac07 : 0000:01:07.0

3114 :
9800-980f : 0000:01:0d.0
9c00-9c03 : 0000:01:0d.0
a000-a007 : 0000:01:0d.0
a400-a403 : 0000:01:0d.0
a800-a807 : 0000:01:0d.0

Well, and there seems to be the problem. The ports are likely
allocated by the BIOS and not the kernel. Sorry, but I think Silicon
Image is just being incompatible with itself here and the only
thing that would help is a BIOS update for the controller(s).

Arno

 

[Andre Majorel]

Well, and there seems to be the problem. The ports are likely
allocated by the BIOS and not the kernel. Sorry, but I think Silicon
Image is just being incompatible with itself here and the only
thing that would help is a BIOS update for the controller(s).

Things aren't looking up, then. BIOS updates aren't really
Connectland's style. But thanks.

If someone can recommend a cheap PCI card with two or more IDE
ports, I'm all ears.

 

[Arno Wagner]

Things aren't looking up, then. BIOS updates aren't really
Connectland's style. But thanks.

If someone can recommend a cheap PCI card with two or more IDE
ports, I'm all ears.

I like the Promise ultra 100/133 TX2. They have two poers each and
you can have more than one in a system, the BIOS of the one
frist activated will find the other cards.

I had a 4 disk RAID5 running with a pair of them for several years
under Linux. Never had problems. One thing though: IF you do
RAID with them, use only one disk per IDE channel, otherwise
they get painfully slow. This seems however to be a general
problem with IDE and not with this particular controller only.

Arno

 

[Andre Majorel]

I like the Promise ultra 100/133 TX2. They have two poers each and
you can have more than one in a system, the BIOS of the one
frist activated will find the other cards.

I had a 4 disk RAID5 running with a pair of them for several years
under Linux. Never had problems. One thing though: IF you do
RAID with them, use only one disk per IDE channel, otherwise
they get painfully slow. This seems however to be a general
problem with IDE and not with this particular controller only.

Thanks. I got a Promise Ultra 133 TX2. The chip is a Promise
20269 (PCI ID 105A:4D69 rev 02).

Unfortunately, like the Connectland, it locks up the PC at
power-on unless you disable the on-board SiI 3114.

Curiously, the I/O ports are exactly the same as what is
reported for the SiI 0680 :

9c00-9c0f : 0000:01:07.0
a000-a003 : 0000:01:07.0
a400-a407 : 0000:01:07.0
a800-a803 : 0000:01:07.0
ac00-ac07 : 0000:01:07.0

I tried moving the Promise to a different PCI slot but it makes
no difference (the PC still freezes).

Sheesh. Wasn't PCI supposed to put an end to this kind of
problems ?

 

[Arno Wagner]

Thanks. I got a Promise Ultra 133 TX2. The chip is a Promise
20269 (PCI ID 105A:4D69 rev 02).

Unfortunately, like the Connectland, it locks up the PC at
power-on unless you disable the on-board SiI 3114.

Curiously, the I/O ports are exactly the same as what is
reported for the SiI 0680 :

9c00-9c0f : 0000:01:07.0
a000-a003 : 0000:01:07.0
a400-a407 : 0000:01:07.0
a800-a803 : 0000:01:07.0
ac00-ac07 : 0000:01:07.0

I tried moving the Promise to a different PCI slot but it makes
no difference (the PC still freezes).

Sheesh. Wasn't PCI supposed to put an end to this kind of
problems ?

Yes, indeed. I suspect that the onboard BIOS was programmed
by somebody pretty incompetent. Would nto be the first instance
of that.

Arno

 

[Folkert Rienstra]

Arno Wagner wrote in message news:[email protected]

Yes, indeed. I suspect that the onboard BIOS was programmed
by somebody pretty incompetent.

Someone like yourself, babblebot?

Would nto be the first instance of that.

Pity about that ...

.... earlier. Incompetent babblebot.