The Great Gazelle PCI Hack Thread, Part 2

[ried]

Final update from me. Confirming that both USB and FireWire are now working in the TAM (Sonnet Tango 2.0), and hard drives connected to both are mounted on the desktop. Still no FireWire shown in the System Profiler, for some reason. Will update this thread if the bomb that appears during the startup sequence continues to be an issue. As of now, it boots properly after zapping the PRAM.

 

[Byrd]

@ried thanks for posting your progress on this - so this is with the patch, then PRAM reset and all is functional? Can't wait to chuck my Tango card in and fire up an old iPod on it.

 

[ried]

Well, we do have an issue. With my hardware + software configuration, at least. I was unable to boot into 9.2.2 with extensions on unless I zapped the PRAM. If I zapped it, it would startup as normal (with USB + FW functioning).

Then, that zapping the PRAM trick stopped working. Now I cannot boot into 9.2.2 with extensions under any circumstances. I suspect the firmware patch created an incompatibility with either my Ethernet card, the Sonnet Crescendo G3/500, or the OldWorld 1.0.1 hack that enabled 9.2.2 to be installed. I tried disabling extensions for both the Farralon Ethernet card and the Sonnet Crescendo, to no avail. They were not the culprits. The TAM will not boot into 9.2.2 if it has to load any extensions at all, and will only boot with extensions off (holding down shift).

Continuing to investigate.

 

[ried]

Yeah, the firmware patch creates some kind of fundamental incompatibility with Mac OS 9.2.2 on the TAM. Mac OS 9.1 and 9.2.1 seemed to work fine as a fresh install, but upgrading to 9.2.2 results in the same hang during startup.

 

[jeremywork]

I’ve also never seen the cursor become a bomb before, and I thought I’d seen pretty much every form of crash by now!

My TAM is in the same configuration, and while working through the beginnings of this project with @cheesestraws I discovered that my TAM keyboardfolio was strangely incapable of sending the TAM into OF mode, and additionally it wasn’t able to wipe the patch from PRAM, despite the computer rebooting like normal with CMD-OPT-P-R held down. Doing the same thing from a plugged in AppleDesign keyboard resolved the issue for me. If I recall, @cheesestraws’ keyboardfolio was also found to have abnormalities, but not identical to mine.

I also found some combinations of the Farallon driver and Crescendo driver would cause a hang during extensions loading while patched for USB&FW. I think the drivers hosted at 6400 zone are stable for me. No bombshell cursor, though!

 

[cheesestraws]

So, to recap:

• People are likely to need to zap the PRAM after installing the patch—not too surprising, but good to know.
• 9.2.2 doesn't like it at all—again, not too surprising, but good to know.
• Apple System Profiler is as easily confused as I am.

Still no FireWire shown in the System Profiler, for some reason

Yeah, this is what I see too here, even with the "supported" card and the "official" patch. I'm pretty sure it's cosmetic; all I can assume is that Apple System Profiler is a Bear of Very Little Brain and gets confused by unorthodox configurations easily. I mean, don't we all...

Yeah, the firmware patch creates some kind of fundamental incompatibility with Mac OS 9.2.2 on the TAM. Mac OS 9.1 and 9.2.1 seemed to work fine as a fresh install, but upgrading to 9.2.2 results in the same hang during startup.

This isn't totally surprising; as I understand it, 9.2.2 changed a lot of low-level stuff to make it work properly in Classic under OS X, among other things; I don't know what the OldWorld hack does, but it's not totally surprising they're fighting, since presumably both involve OpenFirmware magic. I'm running 9.2.1 on mine, for what it's worth.

If I recall, @cheesestraws’ keyboardfolio was also found to have abnormalities, but not identical to mine.

Yeah, I think they're just a bit shonky. I ended up just setting auto-boot? to false on mine rather than have to wrestle it every time... (and I stuck a post-it on the front of the machine to remind me I'd done it, because I'm an idiot )

 

[cheesestraws]

OldWorld 1.0.1 hack

Can you point me to this hack? Googling this turned up precisely nothing.

 

[ried]

Can you point me to this hack? Googling this turned up precisely nothing.

OS 9 Helper from OWC:

Other World Computing: OS 9 Forever

Upgrade the RAM and storage of nearly any Apple computer at MacSales.com. We also provide refurbished Macs, external storage, docks, accessories, and more!

eshop.macsales.com

Seems we are not the only ones who have been down this road:

USB/Firewire cards (or chipsets) that work in the TAM

there are at least 3 versions of the wings card. I know because i collected them trying to fit one in my G3 AiO, which required one with a special header that's is not present on the other 2 variations.

68kmla.org

 

[cheesestraws]

OS 9 Helper from OWC:

Thanks!

Seems we are not the only ones who have been down this road:

The OrangeMicro patch is identical to the Sonnet one: I strongly suspect that both were provided by a single third party (Apple?) to get the cards working. This would also explain why the patches were never updated, if said third party declined to do so again. So this is good, because it means that this misbehaviour isn't something I have introduced but which is in the original Sonnet patch too.

 

[CC_333]

I’ve also never seen the cursor become a bomb before

I've seen it before, but only on 9.2.2, so it must've been added in that version.

It may also have been in 9.1, 9.2.0 and 9.2.1, but as I recall, I've never crashed either of those in a way that would've given the bomb cursor an opportunity to appear.

c

 

[Franklinstein]

Well, it doesn't get anyone the various IDs, but I would wager the bridge chip on that board is a later variant of the 21052. There's that big Intel 'I' not covered by the sticker.

Intel bought DEC's chip foundry. DEC's PCI-PCI bridge chip was the 21052 (IIRC) . I know Intel made those because I experimented with PPBs about 15 years ago, and I bought some of the Intel flavor. But Hint's 208 pin QFP (or was it 160?) was also pin compatible.

At the time, I swapped about four different variants of the PCI-PCI Bridge chip (later revs. of DEC's, Intel's and Hint's) into the Umax S900 logic board. They all worked fine, but none of them changed the behavior of the S900 -- leading the conclusion that its PCI issues are in firmware.

Yeah it's sad that DEC fell apart so badly that Compaq bought them out, then Compaq murdered discontinued development of the Alpha in favor of the mythical performance of the forthcoming Itanium, a product which turned out to be completely over-hyped, overpriced, and under-performing, ultimately leaving even Intel fanboys severely disappointed in what was supposed to be the Next Big Thing which itself was quietly discontinued following a steady decline in already-low sales numbers (and also HP was no longer paying Intel to prop up the obvious albatross the product had become). Lots of blame to go around there but I lay most of it at Compaq's feet. One of the many things I would "correct" if I had a time machine.

Anyway, the PCI-PCI bridge used by the S900 (and other mid-90s clones) was indeed a 21052. And I actually have an identical but differently badged copy of the Adaptec card upthread (AHA-8945CP): mine is a FireMAX card (AHA-8945/PRMX) and it uses a DEC 21152 PCI-PCI bridge. I don't know what else you'd want to know about it. Maybe a ROM dump? If so let me know if there's a specific utility to use to extract it.
But as mentioned most of the problems in this thread concern overcoming bugs in the Mac's firmware, not faults in DEC's hardware or a specific card's design, so I'm not sure if it would do much good.

 

[trag]

Anyway, the PCI-PCI bridge used by the S900 (and other mid-90s clones) was indeed a 21052. And I actually have an identical but differently badged copy of the Adaptec card upthread (AHA-8945CP): mine is a FireMAX card (AHA-8945/PRMX) and it uses a DEC 21152 PCI-PCI bridge. I don't know what else you'd want to know about it. Maybe a ROM dump? If so let me know if there's a specific utility to use to extract it.
But as mentioned most of the problems in this thread concern overcoming bugs in the Mac's firmware, not faults in DEC's hardware or a specific card's design, so I'm not sure if it would do much good.

If you're card is a Mac specific version, then a ROM dump could be useful if someone wants to convert a PC version to Mac use. I don't know of any dumping utility. I usually just remove the chip and put it in a chip programmer for reading. Although, I think some of the older Adaptec PC firmware update utilties included the ability to read the firmware contents as well as write them. So if you wanted to dump the ROM, installing it in a PC and using an Adaptec firmware updating utility would probably do the trick.

Just to be clear, I'm not requesting that.

Yes, the bridge chip information is academic in this thread -- probably. There's been some discussion about whether the bridge chip affects compatibility with Macs, but I doubt it. As far as I can tell the bridge chips do what they're meant to.

That was kind of the point (although I didn't make it clear) of my message about installing four different bridge chips in the S900. They all behaved exactly the same, and were even pin compatible with each other.

Although, I guess they must be changing something to justify t he different part numbers from the same manufacturer.

 

[cheesestraws]

Yes, the issue with bridge chips is less to do with the chip and all to do with the firmware in the Mac. The chips are, I think, basically interchangeable if your firmware is behaving. Note, though, that this does not mean that the Sonnet patch will work with non-HiNT bridge chips (I tried), because in that case the patch is working around the fact that the firmware isn't working properly, and takes shortcuts. I haven't yet worked up the energy to work out what isn't being initialised right in that siig card you sent me, @trag , for example.

 

[trag]

Note, though, that this does not mean that the Sonnet patch will work with non-HiNT bridge chips (I tried), because in that case the patch is working around the fact that the firmware isn't working properly, and takes shortcuts.

Interesting. Does the Sonnet patch directly use the vendor ID or some such from the HINT chip? Or is it a finer level of detail.

 

[cheesestraws]

It uses the vendor ID and device ID, as I noted above in the "how to come up with your own patch" post. But it also pokes some registers of the bridge chip itself to initialise bits of it (which presumably OF would set up itself on a less dysfunctional system), and when I run that with the vendor/device changed to the siig chip, OF's output looks OK (compared to the beige G3) but MacOS crashes very hard on boot. So something isn't being initialised right. So at the moment the Sonnet-derived patches work (at least on a TAM) for the new Tango 2.0 (thanks to @ried for the testing there!) and the Kalea ST-200 (?) combo, but not the Siig yet.

 

[Trash80toHP_Mini]

Is anyone working on the Sonnet TEMPO trio?

 

[cheesestraws]

Is anyone working on the Sonnet TEMPO trio?

You or someone else asked this upthread . No, or at least not that I'm aware of. I don't have one to play with, for starters. I doubt you'd get all three peripherals working in the same nvramrc, or at least I can't see how to yet, for lack of room.

 

[ried]

Sonnet Tango 1.0 (top) and 2.0 (bottom), just for reference.

 

[ried]

Confirming that the Sonnet Tango 1.0 shows up in Apple System Profiler the exact same way the Tango 2.0 does (after their respective firmware patches are applied). USB bus shown, not attached to the PCI card, and no FireWire bus shown at all. This is Tango 1.0:

 

[cheesestraws]

Thanks for checking that! I hadn't had time to check. Good. This is progress.