A Moogle  🏳️‍🌈 💻 🎮 🎶 🇸🇪

It then draws two logos using numerous different palette entries. Now, I'm not racist, but I have to assume that Sun's engineers at the time wanted actual logos to appear, and not two black rectangles.

Compounding the issue is that I strongly suspect I know why it's not uploading the correct palette, but I can't quite work out which IRQ line on the SPARC an SBus interrupt is supposed to pull.

I have some vague docs that say that palette uploads should be done during V-sync in order to prevent tearing, and which specify that this should be done by enabling V-sync IRQs, then waiting for the IRQ to trigger.

However, I have other documentation from BitSavers which has a penned-in note saying that IRQ 8 is for "video". If I trigger IRQ 8, then the Solaris 2.5.1 installer explodes with a repeating error message about an unexpected IRQ.

It's not a matter of the board IDing improperly, because if it wasn't properly ID'd, then the installer wouldn't be tickling the exact hardware registers in order to kick drawing commands in the first place. TurboGX is pretty unique.

I've even resorted to checking various BSD distro sources. They do support TGX - badly. In fact, the drivers all say that the info is based on pure reverse-engineering. Moreover, they don't wait for an IRQ to upload palettes, and there's a driver note saying they maybe should.