Need help getting started with RO builds (DDE31c, RPCEmu)

If by “on the physical hardware” you mean “the Bush internet TV” then that doesn’t surprise me. The HAL-ification was done after that, so although there may be a few unused switches hanging round in the code none of it will have been enabled since the early 2000’s and the ‘669 is sufficiently different to the ’665 I doubt it’d work out of the box. All those switches selected on Lazarus in Hdr:Machine will have changed the contemporary Kernel, not the HAL-ified one.

A lot of the flags actually seem to still be present but their set values (in the Hdr:Machine.Machine header) are ignored. The big one of those I’d caught was in the VIDC20 video driver: the Bush set-top box has no VIDC VCO, so that flag needs to be set (a RISC PC or A7000 does have the VCO). Problem is the VIDC20 driver doesn’t pull in the Machine header, it just sets DontUseVCO SETL {TRUE} (always use the VCO). I’ve patched that to use the value in the Machine header, so hopefully video might work on the next run. I’m not sure if the minimal VIDC driver in the HAL needs work too.

I’d spotted the 669 Super I/O changes – the Super I/O (the comments sometimes call it the “combo chip”) inititalisation in the contemporary kernel seems much more fleshed-out.

So far the changes I’ve found are…

• The Bush box doesn’t have an RTC. Initialising the system clock from the RTC on boot is disabled (but only if the RTC wasn’t detected) and the system clock defaults to the Unix Epoch (instead of 1/1/1900). The OS_ResyncTime SWI implementation goes away too.
• Monitor type detection is disabled (UpdateMonitorTypeLatch is compiled out). Hsync seems to be forced to be inverted, and a lot of VIDC Init changes related to detecting PAL vs NTSC have been made.
• Super I/O initialisation is completely different (there’s an ifdef STB block which replaces the original init code with the 669-capable code).
• STBs don’t print the “Acorn ADFS” banner on boot. Some error messages also change (notably the boot failure one changes to a “service outage” message) and a variety of floppy-related things go away (Territory module being able to load from floppy, “Floppy Boot” option in error screens, etc.)
• Super I/O Interrupts are set to active-high on the Bush box build — I’m trying to figure out what that means. I expect there will be some IOMD init changes to deal with that.

I did notice the 669 Super I/O is I/O-address configurable, and the Config registers have one of two addresses. If it boots with the alternate config address, it won’t be configured at all (the STB init code deals with this). It’s also address-configurable, which means if the address registers aren’t set (they’re not, because the 665 doesn’t have them).

I’m hoping I can at least get a serial-port console (DebugTerminal TRUE in the kernel) and HAL serial debug, then I’ll worry about VIDC. Though it does worry me that a build I did with one or both of those options on won’t run in RPCEmu. I’m hoping that’s just because the serial port isn’t being emulated. Either way I’m looking through the PortManager code to see if there’s an easily accessible GPIO which I could toggle from code to make the line on my oscilloscope wiggle and work out how far the boot process is getting.

My plan at the minute is to improve the HAL Super I/O initialisation to support the 669 chip, then hope that gets the debug terminal working. If I had a spare RISC PC or 7000 kicking around, I’d program a pair of debug ROMs and see what that did on real hardware. Maybe it works after all?

If you could make use of an A7000 motherboard (ARM7500, not ARM7500FE) I have one spare I could pass along. CMOS battery has been replaced by a capacitor so it doesn’t hold settings for more than a few hours unpowered, but it worked last time I powered it up (~10 years ago). You would need to rig up a suitable PSU.

I’m not sure that would be able to boot your modified ROM though, since it has a 7500 and a ’665 IIRC.

Thanks for the kind offer, Theo – I’ll drop you an email about that.

I’ve made some significant progress getting this all working — this evening the Bush set-top booted into the IOMD HAL for the first time:

IOMD HAL init
0000aa7c = IOMD ID
Probe_DRAM_bank
10000000 = bank address
003fffff = decoded address lines
Probe_DRAM_bank
14000000 = bank address
003fffff = decoded address lines
Initialise_DRAM_A7000 done
003fffff = bank 0
003fffff = bank 1
00000000 = bank 2
00000000 = bank 3
00000070 = DRAMWID
POST started
0000cfae = Vsync time
00004051 = POST result

I should probably start another thread for this given I’ve got the build issues sorted now… :)