heathzenith/h8.cpp: Implement H8 Bus #13560

mgarlanger · 2025-04-05T07:19:52Z

Implements the H8 bus otherwise known as the Benton Harbor Bus.

This moves all of the H8 functionality from h8.cpp on to separate virtual cards which match up with the real H8's architecture.

Added:

CPU8080 - 8080 CPU board
FP - Front Panel board.
H-8-1 - 4k/8k Static RAM memory board
WH-8-64 - 32k/48k/64k Dynamic RAM memory board
H-8-5 - Cassette and serial board

Note:

the H-8-5 was enhanced to have 2 cassette decks instead of 1, and to have the H8 control the decks, so they automatically start when pressing load from the keypad.
there are many more boards to implement, this just keeps parity with the h8.cpp functionality.

mgarlanger · 2025-04-12T21:39:17Z

@Robbbert since you have the copyright on h8.cpp, do you have any feedback on these changes?

Robbbert · 2025-04-12T22:29:33Z

@mgarlanger I have involuntarily abandoned this driver, so you can remove my name from it if you wish. The content is now entirely in your hands.

cuavas · 2025-04-22T20:23:01Z

src/devices/bus/heathzenith/h8/h8bus.h

+#pragma once
+#include "emu.h"


Bad things happen if you #include "emu.h" from headers – it has to be the first thing in each .cpp file (due to the way we use precompiled prefix headers in our build system).

cuavas · 2025-04-22T20:26:18Z

src/devices/bus/heathzenith/h8/h8bus.h

+#include <functional>
+#include <utility>
+#include <vector>
+#include <string.h>


How many of these are actually used in this header? It looks like <utility> and <functional> are used, but I can’t see anything from <string.h> at a glance.

cuavas · 2025-04-22T20:32:15Z

src/devices/bus/heathzenith/h8/h8bus.h

+inline void device_h8bus_card_interface::set_slot_int1(int state)
+{
+	h8bus().set_int1_line(state);
+}
+
+inline void device_h8bus_card_interface::set_slot_int2(int state)
+{
+	h8bus().set_int2_line(state);
+}
+
+inline void device_h8bus_card_interface::set_slot_int3(int state)
+{
+	h8bus().set_int3_line(state);
+}
+
+inline void device_h8bus_card_interface::set_slot_int4(int state)
+{
+	h8bus().set_int4_line(state);
+}
+
+inline void device_h8bus_card_interface::set_slot_int5(int state)
+{
+	h8bus().set_int5_line(state);
+}
+
+inline void device_h8bus_card_interface::set_slot_int6(int state)
+{
+	h8bus().set_int6_line(state);
+}
+
+inline void device_h8bus_card_interface::set_slot_int7(int state)
+{
+	h8bus().set_int7_line(state);
+}


This naïve approach to implementing “party line” signals causes issues if multiple cards assert the same signal simultaneously. Consider what would happen with the code as-is if two cards both assert INT2:

Card 1 asserts INT2, bus sees INT2 asserted (correct)

Card 2 asserts INT2, bus still sees INT2 asserted (correct)

Card 1 deasserts INT2, bus sees INT2 deasserted (incorrect, card 2 still has it asserted)

Card 2 deasserts INT2, bus sees INT2 deasserted (correct)

This is why the INTELLEC 4 bus (src/devices/bus/intellec4) keeps track of which cards have which lines asserted in bit flags. It also facilitates implementing cards that react when any other card asserts a line. I’d suggest doing something similar.

I'm not aware of any Heath software that would properly handle shared interrupts. Would it be ok, to add a TODO in this PR about this issue, and address it in a followup PR?

cuavas · 2025-04-22T20:36:49Z

src/devices/bus/heathzenith/h8/h8bus.cpp

+	// Verify required cards are installed.
+	if (!m_p1_card)
+	{
+		fatalerror("Can't find H-8 P1 - Front panel card\n");
+	}
+
+	if (!m_p2_card)
+	{
+		fatalerror("Can't find H-8 P2 - CPU card\n");
+	}


I don’t think this is a great idea – I realise the system won’t do much without these cards, but MAME doesn’t have a nice way of displaying fatal errors when starting a system from the internal UI, or reporting them back to an external launcher. Letting the user discover the system does nothing is the lesser evil IMO (much like removing the video card from a Mac II or IBM PC clone in MAME).

Ok, I'll remove these. Should I also remove the fatalerror() that is in device_h8bus_card_interface::interface_pre_start() if m_h8bus is not set? Or should it be a throw device_missing_dependencies(); like it is in the intellec4.cpp code?

cuavas · 2025-04-22T20:39:16Z

src/devices/bus/heathzenith/h8/h8bus.cpp

+void h8bus_device::set_int1_line(int state)
+{
+	m_p2_card->int1_w(state);
+}
+
+void h8bus_device::set_int2_line(int state)
+{
+	m_p2_card->int2_w(state);
+}
+
+void h8bus_device::set_int3_line(int state)
+{
+	m_p2_card->int3_w(state);
+}
+
+void h8bus_device::set_int4_line(int state)
+{
+	m_p2_card->int4_w(state);
+}
+
+void h8bus_device::set_int5_line(int state)
+{
+	m_p2_card->int5_w(state);
+}
+
+void h8bus_device::set_int6_line(int state)
+{
+	m_p2_card->int6_w(state);
+}
+
+void h8bus_device::set_int7_line(int state)
+{
+	m_p2_card->int7_w(state);
+}


Remember any card on the bus can see when another card asserts one of these, as they’re all wired together. It’s better to use an approach that allows telling all the cards the state they’d see on these lines.

cuavas · 2025-04-22T20:45:12Z

src/devices/bus/heathzenith/h8/cpu8080.cpp

+	// 2.048 MHz
+	static constexpr XTAL H8_CLOCK = XTAL(18'432'000) / 9;


Please make this a local in the machine configuration function, as it’s only used in one place (it increases static data size if you make it static).

Also, assuming the CPU card drives the bus clock, consider having the CPU card set the bus clock on start so it can propagate to other cards.

cuavas · 2025-04-22T20:52:00Z

src/devices/bus/heathzenith/h8/wh_8_64.cpp

+
+	required_device_array<ram_device, 8> m_ram;


Since you can’t use -ramsize with these anyway, you’re better off just using memory_share_array_creator rather than an array of ram_device.

cuavas · 2025-04-22T20:54:12Z

src/devices/bus/heathzenith/h8/wh_8_64.cpp

+DEFINE_DEVICE_TYPE_PRIVATE(H8BUS_WH_8_64,     device_h8bus_card_interface, wh_8_64_device,     "wh8_h_8_64",     "Heath WH-8-64 64k Dynamic RAM");
+DEFINE_DEVICE_TYPE_PRIVATE(H8BUS_WH_8_64_48K, device_h8bus_card_interface, wh_8_64_48k_device, "wh8_h_8_64_48k", "Heath WH-8-64 48k Dynamic RAM");
+DEFINE_DEVICE_TYPE_PRIVATE(H8BUS_WH_8_64_32K, device_h8bus_card_interface, wh_8_64_32k_device, "wh8_h_8_64_32k", "Heath WH-8-64 32k Dynamic RAM");


Is it worth having separate card devices for the different RAM configurations, or would you be better off just having a machine configuration “switch” to set the amount of RAM present? (For example the MDC4•8/MDC8•24 cards use a fake switch to select the video RAM size.)

Based on this prior comment from @pmackinlay - #13040 (comment) about how to handle the m_connected_tlbc, I split that card into 2 slot options. On the other PR it felt cleaner with the 2 cards, but I'm not sure which would be better with this board. And followed that with this PR. But if you want it handled by a machine config switch, I can change it.

…igurable

…gnals. Other minor cleanup

mgarlanger · 2025-04-27T18:45:33Z

@cuavas all comments should be addressed now, including properly handling "party line" signals.

mgarlanger added 3 commits April 5, 2025 02:07

heathzenith/h8.cpp: Implement H8 Bus

3dc513c

Merge remote-tracking branch 'origin/master' into h8_bus

3eefc54

Fix validate errors, and some minor cleanup

83cd180

Move CPU and FP functionality into cards and some cleanup

65d4759

mgarlanger changed the title ~~heathzenith/h8.cpp: Implement H8 Bus - phase 1~~ heathzenith/h8.cpp: Implement H8 Bus Apr 19, 2025

mgarlanger added 2 commits April 20, 2025 00:29

Add jumper definitions for 8080A CPU board based on schematics

300989e

Minor update to trigger rebuild with fixed master.

cb19ce9

cuavas reviewed Apr 22, 2025

View reviewed changes

mgarlanger added 4 commits April 25, 2025 23:55

Address review comments

fecb08c

Fix compiler error

12fa67c

Add handling of WH_8_64 switches, make amount of memory on cards conf…

06d2122

…igurable

Update bus handling to properly handle multiple cards driving same si…

93b7c9f

…gnals. Other minor cleanup

mgarlanger added 6 commits April 27, 2025 19:07

Disable logging on WH-8-64

3f588de

fix interrupt handling on h8bus and minor cleanup

ba2b227

fix/simplify single instruction interrupt mode

d6e3ab4

Simplify RAM handling in H-8-1

79da6b9

Combine int 1 and int 2 signals on CPU board from p201 and the bus

1a35e0e

use srcclean on files

24302a8

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

heathzenith/h8.cpp: Implement H8 Bus #13560

heathzenith/h8.cpp: Implement H8 Bus #13560

mgarlanger commented Apr 5, 2025 •

edited

Loading

Uh oh!

mgarlanger commented Apr 12, 2025

Uh oh!

Robbbert commented Apr 12, 2025

Uh oh!

cuavas Apr 22, 2025

Uh oh!

cuavas Apr 22, 2025

Uh oh!

cuavas Apr 22, 2025

Uh oh!

mgarlanger Apr 23, 2025

Uh oh!

cuavas Apr 22, 2025

Uh oh!

mgarlanger Apr 23, 2025

Uh oh!

cuavas Apr 22, 2025

Uh oh!

cuavas Apr 22, 2025

Uh oh!

cuavas Apr 22, 2025

Uh oh!

cuavas Apr 22, 2025

Uh oh!

mgarlanger Apr 23, 2025

Uh oh!

mgarlanger commented Apr 27, 2025

Uh oh!

Uh oh!

		// 2.048 MHz
		static constexpr XTAL H8_CLOCK = XTAL(18'432'000) / 9;

		#pragma once
		#include "emu.h"

heathzenith/h8.cpp: Implement H8 Bus #13560

Are you sure you want to change the base?

heathzenith/h8.cpp: Implement H8 Bus #13560

Conversation

mgarlanger commented Apr 5, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mgarlanger commented Apr 12, 2025

Uh oh!

Robbbert commented Apr 12, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mgarlanger commented Apr 27, 2025

Uh oh!

Uh oh!

mgarlanger commented Apr 5, 2025 •

edited

Loading