PCem

This patch adds two new models to PCem, my first two PCs from 1991 and 1995!

The first is the Hyundai Super-286TR, based on the Chips & Technologies SCAT chipset:
-Award 286 Bios, for Chips and Technologies Single-Chip AT (SCAT) chipset
-AMD 286 12 MHz
-1MB RAM
-VGA OAK OTI-037C 256KB
-PC Speaker Audio
-Floppy drives A: (1.2MB 5 1/4″) and B: (1.44MB 3 1/2″)
-Approx. 40MB HDD – don’t know the model, I was 5 when it broke. Probably a Connor CP3000 or another with the same geometry (type 33 in the BIOS)
-From the chipset: 1x Serial, 1x LPT, 1x AT Keyboard
-System BIOS MD5: 19a0e6a13c3b8df827add273468ebe02
-VGA BIOS MD5: 0fd29240900f1fdc31aacf4c139b3457

(That's my arm as a child in the picture!)

The second is the Itautec Infoway Multimídia, built around the Intel Zappa motherboard.
-Intel Zappa (Advanced/ZP, Triton chipset, 256KB cache)
-Pentium 75MHz Socket 5
-8MB RAM
-VGA Cirrus Logic CL-5434 1MB (Model: VI-720)
-Sound Aztech Sound Galaxy Pro 16 AB (FCC ID I38-MMSN824)
-Floppy drive A: 1.44MB 3 1/2″
-850MB HDD Western Digital Caviar E-IDE WDAC2850 10ms 64K 4500RPM 16.6MB/s (~810MB on base 2)
-CD-ROM 4x speed AZT 468-01I or AZT 468-02I
-From the NS PC87306 I/O chip: 2x serial, 1x LPT, 1x at keyboard
-14.400 modem (USRobotics I think! I don’t have it anymore)
-System BIOS: Get from intel
-VGA BIOS MD5 is 7f2c1fd555c2b8cf4d67e2c4fc13767f for the first 32KB of the 64KB ROM, the upper 32KB is all 0xFF.



I've made a post here a few days ago telling the story of how PCem got so close to them that I decided to make patches for the little that was left: [url]http://eluancm.net/2020/02/15/emulating-my-old-computers/[/url]

Looks good in general. A couple of questions though :

- Do you know / can you check what SuperIO chip the Hyundai machine is using? The BIOS has configuration options for serial/parallel/floppy/IDE which suggests there is a SuperIO on board, but there isn't one mentioned in the patch.
- You've added on board video for both machines; do they both actually have video on the motherboard? The Itautec is confusing me at the moment, because if it's using a Zappa board then it should _not_ have on board video. If you meant that it's another board merely _based_ on the Zappa then that would explain it

Hundai SuperIO:
-The SuperIO chip for the Hyundai is the Chips & Technologies F82C710, one of the companions to the SCAT chipset and I completely forgot about it!
-Luckily, the config process is simple and doing it should be straightforward.

Onboard videos:
-I'm guilty here, I've used them as a way to force a certain config. The good/correct way should be having templates with valid historical configs of the various machines that PCem emulates during config creation and not overloading the onboard video configs (which are purely to allow fixed configs and leave the audio, etc, configs without fixed values anyway.)
-On top of that, there's also the issue of BIOS versions + logos (the ones I've extracted are quite old, it seems) and there is no way to select PCI slots without using the onboard video facilities (to avoid re-configuration during the first boot with a dumped hdd).

I've added the F82C710 UPC. Only the registers needed for the BIOS of the Super-286TR were implemented. Since there are lots of SCAT machines supported by PCem and some of them may use this SuperIO chip, I've added some logging to warn if anything else is configured as an aid.

This chip doesn't have IRQ configuration, so the Super-286TR has jumpers for serial/parallel IRQs. Because of this, I've implemented it as a machine device. There are no other unimplemented jumpers for this machine in PCem. Other machines should just copy the device and add/remove jumper options.

This SuperIO assumes a 16450 UART (its configuration mode init commands are sent to the what is the FIFO control register in the newer 16550). Since PCem only supports the minimum necessary of a 16550, I've just added an option to ignore the FIFO control register in serial.c and kept the old behavior for all other machines, even pre-16550 ones.

I've also implemented the keyboard controller commands to enable/disable turbo, but I don't know where to re-enable turbo on ctrl-alt-del soft reset. This means that if the speed is low, a ctrl-alt-del will make POST run at low speed too. Maybe this is the real behavior (no turbo ON command is sent when doing a ctrl-alt-del soft reset and the keyboard controller reset function isn't called).

I've forgotten to add the new files to the repository before generating the patch, so they are loose.

(And I still didn't think about what to do about the video cards, so they are still the same.)

For the video cards, I'd suggest removing those changes for now. I can have a think on how to handle that for v17.

The rest of the patch looks good for me. Are you okay with me committing with the video changes taken out?

Sure, no problem!

Final question! You've limited the CPUs on the Itautec to P75, 90 and 120. Is there actually some kind of hardware limitation here, or are these just the CPUs this particular model was sold with?

I've used the CPU speeds that the machine was sold with, but this may be related to the difference in supported CPUs in the various Zappa revisions (as stated in the motherboard's manual.)

Later models were available with support for more CPUs. For example, next year's Infoway A96 used 133Mhz CPUS, but it exchanged the Zappa for an Intel Atlantis Advanced/AS motherboard with the ATI-264CT (Mach64 with integrated RAMDAC and clock chip with up to 2 MB DRAM) onboard video and without the onboard sound or cache (a COAST module was needed) offered by the new motherboard.

These computers are a curiosity in Brazil's history. Mine was the minimum multimedia version, but the high-end versions came with all sorts of add-ons like video capture (with 16 simultaneous channels from air), remote control (that could even turn the PC on by means of a hybrid AT/ATX PSU and some custom logic), FM radio, house security alarm, wake-up alarm, video calls, etc.

Committed at revs 1535-1537. I did change the Itautec machine to use the same CPU selection as the existing Zappa board in the end; after all, the user might have upgraded the CPU!

EluanCM wrote: ↑Sat 07 Mar, 2020 1:42 am This patch adds two new models to PCem, my first two PCs from 1991 and 1995!

The first is the Hyundai Super-286TR, based on the Chips & Technologies SCAT chipset:
-Award 286 Bios, for Chips and Technologies Single-Chip AT (SCAT) chipset
-AMD 286 12 MHz
-1MB RAM
-VGA OAK OTI-037C 256KB
-PC Speaker Audio
-Floppy drives A: (1.2MB 5 1/4″) and B: (1.44MB 3 1/2″)
-Approx. 40MB HDD – don’t know the model, I was 5 when it broke. Probably a Connor CP3000 or another with the same geometry (type 33 in the BIOS)
-From the chipset: 1x Serial, 1x LPT, 1x AT Keyboard
-System BIOS MD5: 19a0e6a13c3b8df827add273468ebe02
-VGA BIOS MD5: 0fd29240900f1fdc31aacf4c139b3457

(That's my arm as a child in the picture!)

Hi Eluan, thanks for this!

I have a question: when I try to run this one, I get to the the end of the checks and it just stays there... No booting, no message.

I need to change some RAM amount of something for it to ask to enter setup, and only then to I see the message and can go into setup.

Do you know why this could be?

Thanks.

Thank you Eluan for your work!

I am trying to emulate the Amstrad PC5086. It is an 8086 machine, but it uses the F82C710 UPC. The current code does not work with the Amstrad: the BIOS uses a different sequence to enter configuration mode, and I am working on implementing the missing features to support the amstrad. Do you have a PDF of the datasheet that you can share? I haven't been able to find it on the internet.

EluanCM wrote: ↑Wed 11 Mar, 2020 3:45 am I've added the F82C710 UPC. Only the registers needed for the BIOS of the Super-286TR were implemented. Since there are lots of SCAT machines supported by PCem and some of them may use this SuperIO chip, I've added some logging to warn if anything else is configured as an aid.

This chip doesn't have IRQ configuration, so the Super-286TR has jumpers for serial/parallel IRQs. Because of this, I've implemented it as a machine device. There are no other unimplemented jumpers for this machine in PCem. Other machines should just copy the device and add/remove jumper options.

This SuperIO assumes a 16450 UART (its configuration mode init commands are sent to the what is the FIFO control register in the newer 16550). Since PCem only supports the minimum necessary of a 16550, I've just added an option to ignore the FIFO control register in serial.c and kept the old behavior for all other machines, even pre-16550 ones.