08905: attache [cpm]: CP/M won't boot

ID	Category [?]	Severity [?]	Reproducibility	Date Submitted	Last Update
08905	Media Support	Critical (emulation)	Always	Aug 24, 2024, 14:07	Sep 10, 2024, 12:01

Tester	Robbbert	View Status	Public	Platform	MAME (Self-compiled)
Assigned To		Resolution	Open	OS	Windows 10/11 (64-bit)
Status [?]	Confirmed			Driver	skeleton/attache.cpp
Version	0.268	Fixed in Version		Build	64-bit
		Fixed in Git Commit		Github Pull Request #

Summary	08905: attache [cpm]: CP/M won't boot
Description	The system is listed as working, and so is the software. But, it won't boot from the one disk that's in the software list. It just sits there.
Steps To Reproduce	>mame attache cpm
Additional Information	I tried some other disks, and it will point out those that don't have the boot tracks installed. But any proper boot disk just gets stuck.
Github Commit

Flags
Regression Version
Affected Sets / Systems	attache [cpm]

Attached Files

No.22303 Robbbert Senior Tester Sep 4, 2024, 11:29	Debugging showed it running into the weeds. It makes various calls to code at addresses between 1206 and 4606, but there's nothing there. Perhaps there's supposed to be banking? I'm confirming this since nobody else can be bothered.
No.22311 AJR Developer Sep 8, 2024, 16:02	Perhaps the CP/M image isn't for the original Attache but some undumped or incorrectly dumped related system. The hardware has provisions for a 2764 boot EPROM (0x2000 bytes), but all known boot EPROMs are half that size.
No.22312 Pernod Developer Sep 8, 2024, 18:08	After it fails to boot you can press F3 to reset and it will then successfully boot CP/M. This suggests something is not being initialised correctly at first boot, and I'm sure this used to work in the past.
No.22316 Robbbert Senior Tester Sep 10, 2024, 12:01 edited on: Sep 19, 2024, 14:06	Tried the F3 trick - and it works! I was able to test 2 homebrew disks in drive B successfully. Not sure why it's marked as working though. The cursor is missing, and sometimes the bottom line gets stuck. Also only a few variables are save-stated - the majority seem to have been overlooked.