New (alt) BIOS for Ultimate 1MB/Incognito

[+mytek]

 

Unfortunately not, since the BIOS isn't resetting the state of the caps lock in the first place: it's the OS (which necessarily undergoes a warm reset every time you leave setup) which is resetting the caps lock, as well as everything else which is normally set to default values on reset.

 

Of course this is the way it is . Oh well such is life, but hey at least it lets you off the hook .

 

Take care,

- Michael

[flashjazzcat]

Update

 

Ultimate/Incognito main BIOS v.1.02

Ultimate/Incognito PBI BIOS v.1.62

Ultimate/Incognito/SIDE XEX Loader v.1.05

 

http://atari8.co.uk/firmware/ultimate-1mb/

 

http://atari8.co.uk/firmware/incognito/

 

http://atari8.co.uk/firmware/side/

 

New features include

• Loader: joystick operation augmented by console key home, page down and page up navigation keys (primarily aimed at XEGS with keyboard disconnected)
• Loader: joystick control generally improved with proper debounce and better auto-repeat
• Loader: position in file or option list now preserved when moving between menus
• Loader: minor bug fixes
• Main BIOS: config changes requiring reboot now detected when swapping profiles
• Main BIOS: dual-level HDD write lock (see below)
• PBI BIOS: SIO2BT automatically handled by high speed SIO handler (simply enable HSIO for the appropriate drive)
• PBI BIOS: HDD writes can now be blocked at the partition or physical disk level (meaning writes to the MBR and partition table may be completely blocked once the media is partitioned)
• PBI BIOS: Patch vectors for RAM-based IO driver code

Note: the dual-level HDD write lock required relocation of the old HDD write lock bit in the configuration data, so the new option will only be effective if the main BIOS and PBI BIOS are BOTH updated.

 

Note 2: PBI equates changed dramatically, so partitions will go offline until A8 is power-cycled.

Edited June 8, 2016 by flashjazzcat

[Neo-Rio]

 

Note 2: PBI equates changed dramatically, so partitions will go offline until A8 is power-cycled.

 

Hehe.... noticed something was going on with my SIDE2 as BOOT FAILURE screamed out on the screen after warm reboot.

 

So far, no problems with the updates.

[Neo-Rio]

OK, this is weird.

 

Pac-man arcade XEX used to boot OK on my XEGS with SIDE2 and U1MB. Now, since the BIOS update, I've noticed that it doesn't.

http://atariage.com/forums/topic/196992-8bit-pac-man-arcade-final-version-accepting-orders/

 

Not saying that the two are related... but are they?

Edited June 10, 2016 by Neo-Rio

[Neo-Rio]

OK, this is weird.

 

Pac-man arcade XEX used to boot OK on my XEGS with SIDE2 and U1MB. Now, since the BIOS update, I've noticed that it doesn't.

http://atariage.com/forums/topic/196992-8bit-pac-man-arcade-final-version-accepting-orders/

 

Not saying that the two are related... but are they?

 

Urgh.... never mind. I left BASIC enabled.

[flashjazzcat]

The Montezuma tells me that the SIO2BT auto-detection often times out on the initial speed poll. I tested using my own SIO2BT (there being no-one else to hand to give a second opinion) and it worked with RespeQt under Windows 10 64-bit, but I guess it flattered to deceive. It's a little frustrating since as soon as SIO2BT successfully responds to the speed poll, the long timeout is set for the corresponding drive and everything works just fine. But if the initial speed poll times out (and this is done using the standard timeout), SIO2BT detection will fail and the long timeout not flagged on that drive number. Why not use the long timeout when polling the drive? Simple: polling of offline drives will take too long.

 

In any case, he asked me to make it clear that if SIO2BT auto-detection doesn't work with the PBI BIOS, SIO2BT itself is not to be blamed. I might just drop the feature entirely unless room can be found for some friendly hard-settings in the BIOS menu which don't eat many configuration bits (the idea being that you would explicitly enable the long timeout for - say - D2: if SIO2BT is on drive 2).

[Roydea6]

I would like to know what the version # for the most recent FDISK program is.

 

Thank you..

[flashjazzcat]

4.6 I think?

[Roydea6]

4.6 I think?

OK after searching through 2 pc computers and about 6 Side2, MyIDE, CF cards ++ the downloads from your website. I find that I have two versions of FDISK the ABOUT dialogue on one says 45B and the other 45. Now which of theses two is the most current release. They both work okay but what feature creep or coding change came last.

[flashjazzcat]

OK after searching through 2 pc computers and about 6 Side2, MyIDE, CF cards ++ the downloads from your website. I find that I have two versions of FDISK the ABOUT dialogue on one says 45B and the other 45. Now which of theses two is the most current release. They both work okay but what feature creep or coding change came last.

 

Feature creep? Not aware of any. I think the only recent change was getting rid of some arcane FAT16 partition ID when initialising a new MBR. I was going to wait for the SDX 4.48 release before letting out the latest build, but be assured there's nothing newsworthy in FDISK 4.6.

[Roydea6]

After further testing I only find 4.5B on the CF cards that have MyDos and SDX33a.dos partitions, and the 4.5 FDISK is only on the SDX447 partitions. So this satisfies my confusion.

[Roydea6]

The XEX loader can also run BASIC (or Altirra BASIC) programs. Anyone tried this or had issues with it?

 

Anyone miss the keyclick sound? Anyone try "Drive 1 swap" in the PBI menu (this swaps references in software from D1: to the boot drive if the boot drive is not D1:, supposedly as a quick means of switching to a different OS partition)?

New basic program running from XEX loader throw's an ERR 165 'is the a way to open another file from my basic program'. Just trying to load a PIC file for viewing with the basic program. Also I can not seem to be able to XIO into a different folder in the FAT32 area So my work around is to copy basic rom to each folder that has basic programs in them.

The XEX loader load the RMTplayer programs so I can listen to .rmt files from my RMT folder, and in my CMC folder the cmcplayer plays cmc files.

[flashjazzcat]

New basic program running from XEX loader throw's an ERR 165.

What's the complete filespec which causes the error? Error 165 is "Bad Filename" and is thrown out by the filename parser when it encounters a malformed name. Paths aren't allowed (see below).

 

Also I can not seem to be able to XIO into a different folder in the FAT32 area. So my work around is to copy basic rom to each folder that has basic programs in them.

The FMS provides rudimentary DOS 2.5 compatible read-only access to the working directory of the FAT (the working directory being whichever directory was logged in the loader when the XEX or BASIC program was launched). So you get OPEN, CLOSE, formatted directory access, GET and STATUS - that's all. No XIO of any kind. Filenames must be 8.3, prefixed with "D:", and the only unit number allowed is "1".

 

Handling paths and directories would require a much more sophisticated FMS module, and then the simulated DOS layer would no longer be DOS 2.5 compatible. Really the FMS is just a convenient alternative to loading DOS when (for instance) you just want to run UFLASH or something else which requires basic read-only access to the FAT. For a more complete solution, booting DOS is the answer (and I guess FATs will eventually be writable in SDX).

 

Note that simply hitting enter on a BAS file should start built-in BASIC and cause the interpreter to run the program, so a BASIC ROM file in the FAT shouldn't be needed (unless you're using SIDE without U1MB and want to soft-load some other BASIC).

Edited June 15, 2016 by flashjazzcat

[Roydea6]

What's the complete filespec which causes the error? Error 165 is "Bad Filename" and is thrown out by the filename parser when it encounters a malformed name. Paths aren't allowed (see below).

 

Note that simply hitting enter on a BAS file should start built-in BASIC and cause the interpreter to run the program, so a BASIC ROM file in the FAT shouldn't be needed (unless you're using SIDE without U1MB and want to soft-load some other BASIC).

Well I finally got it to work the filespec it was crashing on was the basic.rom file and a couple of deleted files that the windows machine left ghost filenames like _affy. and any other filename with an extender. Also the filenames ending in COM as in BELLCOM will show up as a runnable program. I renamed it with an COX on the end instead of the COM.. On real hardware my program does not parse the extenders. just D:*. and the dir command and basic skips extenders.

[JoSch]

Is there a way to use the slot that's reserved for the GUI OS in another way?

Edited June 17, 2016 by JoSch

[flashjazzcat]

Is there a way to use the slot that's reserved for the GUI OS in another way?

Sure. I even made the slot name editable (it's right at the top of the ROM). You can put whatever you like in there.

[flashjazzcat]

Well I finally got it to work the filespec it was crashing on was the basic.rom file and a couple of deleted files that the windows machine left ghost filenames like _affy. and any other filename with an extender. Also the filenames ending in COM as in BELLCOM will show up as a runnable program. I renamed it with an COX on the end instead of the COM.. On real hardware my program does not parse the extenders. just D:*. and the dir command and basic skips extenders.

Thanks Roy. I'll double check that deleted aliases aren't appearing. Long filenames are ignored but the short aliases will show up instead, usually with a bunch of underscores in them. I found a bug the other day which is probably connected: directory reads via the CIO handler were completely ignoring the filemask and assuming "*.*" all the time. UFLASH masked the issue because it opens "*.*" and filters the names afterwards (in order to grab any folder names under SDX). I discovered the issue when testing with Altirra BASIC in one of the ROM slots. Fixed now (expect update at the weekend).

 

I also updated the user manual to cover the CIO handler, which got very little coverage in the third edition of the docs.

Edited June 17, 2016 by flashjazzcat

[JoSch]

Sure. I even made the slot name editable (it's right at the top of the ROM). You can put whatever you like in there.

Ok, I should have phrased the question in a different way: How do I access the ROM, I put in this slot? I.e. when I put say Action! in the slot, how can I start it?

Edited June 17, 2016 by JoSch

[flashjazzcat]

Ok, I should have phrased question in a different way: How do I access the ROM, I put in this slot? I.e. when I put say Action! in the slot, how can I start it?

There's the rub, and this was prime reason for writing a new BIOS in the first place: to make the GOS slot bootable. So - it's bootable: just put what you want in it, change the title (use a hex editor for now), and you're good to go... with a couple of limitations.

 

1) Whatever is in the ROM space - if banked - must adhere to the SDX banking model. Banking register is at $D5E0, and GOS slot base bank is at $30 (SDX must be shrunk to 192KB for the slot to become accessible).

2) You can't boot SDX then type CAR: to enter the GOS slot. You could probably boot some disk-based DOS and treat the GOS slot as an external banked cart, though.

 

The SDX shrink requirement would be fiddly to get around (GOS will never be smaller than 128KB), and I haven't made this more flexible simply because I think the likelihood of much else being done with the slot is low. More useful, IMO, is the ability to use a 320KB build of SDX.

Edited June 17, 2016 by flashjazzcat

[JoSch]

The SDX shrink requirement would be fiddly to get around (GOS will never be smaller than 128KB), and I haven't made this more flexible simply because I think the likelihood of much else being done with the slot is low. More useful, IMO, is the ability to use a 320KB build of SDX.

As far I can see it,the SDX Image tool does only use the CAR: space that the image defines. It cannot expand the SDX image.

1.) So how would I build a 320K version of SDX?

2.) Would Uflash handle such a big image?

[flashjazzcat]

As far I can see it,the SDX Image tool does only use the CAR: space that the image defines. It cannot expand the SDX image.

1.) So how would I build a 320K version of SDX?

2.) Would Uflash handle such a big image?

I think Trub was working on the "pro" version of the Imaging tool, which does have the ability to change the CAR image size. In the meantime, I've made some ROMs, so if you want one, drop me a PM.

 

Latest UFLASH is already set up to handle 192KB, 256KB and 320KB SDX ROMs.

[Aking]

having signature failure check

flashing bios 1.02 on ultimate 1 mb (first Batch)

 

if the file for incognito?

 

same problem on rom generator and latest uflash

[flashjazzcat]

Does uflash auto-detect U1MB? If not, the current BIOS is so old it completely lacks any identifying metadata. This version of uflash defaults to the "old" firmware layout when forcing hardware detection:

 

uflash.xex

[Aking]

invalid file signature on bios 1.02 on my ultimate

 

flashed the entire rom 0.24 and it worked

 

added my os slots later

 

but don´t know if those files are the latest

 

do you have a complete bios update for ultimate ?

[Aking]

the old (july 2015) complete 512kb file flashed ok

just need an updated one for ultimate 1mb ...