[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#589701: ofpath is not compatible with non-builtin disk controllers

On Sep 29, 2010, at 12:12 AM, Milan Kupcevic wrote:

On 09/28/2010 07:26 PM, Benjamin Cama wrote:
Well, even if you didn't mention it explicitly, it was listed in the
lspci dump since the beginning and nobody saw it until now. And one may
also find it disturbing that a piece of software doesn't work with an
addon card on a machine that has free slots to add such cards ! But
bootstrapping an OS on non-original hardware modification is always a
bit hard.

Promise add-on sata controllers do not have proper OpenFirmware support.
Therefore, they will never be able to boot an OpenFirmware machine.
There are very few add-on SATA controllers with proper OpenFirmware
support. Here are four I personally tried out:

Sonnet Tempo SATA PCI [1]
SeriTek/1V4 [2]
SeriTek/1eVE2+2 [3]
SeriTek/1S2 [4]

Other SeriTek cards explicitly described as "compatible with all
PCI-based Power PC Macintosh computers" with "completely self- contained
booting functionality" should probably work equally well.

If you find other add-on sata controller(s) able to boot a PowerMac
machine, please let me know.


 [1] http://www.sonnettech.com/product/tempo_serial_ata.html
 [2] http://www.firmtek.com/seritek/seritek-1v4/
 [3] http://www.firmtek.com/seritek/seritek-1ve2plus2/
 [4] http://www.firmtek.com/seritek/seritek-1s2/

OK... If I'm understanding this correctly, it means that NewWorld Mac's can boot either from the on-board ATA or one of these OF- supporting SATA controllers (possibly with some tweaks to ofpath to deal with any peculiarities of the non-Apple controllers) but *not* from most non-OF-aware controllers.

I, for one, can live with that as long as I've got some small ATA drives I can use for the purpose. All I have to put on it is the "apple bootstrap (HFS) partition" and /boot, right? But root and all the rest can be on the big SATAs, right?

It's not "beautiful" or elegant, but it is a reasonable work-around for a problem that we can't fix in the Debian environment.

Or am I missing something?


Reply to: