- This change avoids clearing the screen every time an OS reset occurs. A non-destructive VDC reset is now performed that leaves the screen contents and cursor position unchanged.
- If an application overwrites the video RAM, then the screen may contain garbage upon return to OS. However, the console will still be functional and the user can just press enter a few times to clear the screen.
- When an OS performs a reset operation, the HBIOS hook has been extended to automatically reset the video hardware of the CRT device (if it exists).
- This should go a long way toward fixing corrupt video after an application is run that reprograms the video hardware.
- An OS reset may or may not be performed when an application exits depending on the behavior of the application. So, if an application exits without initiating a reset, then the video will not be reset. Most applications that muck with the video chip directly will perform the reset at exit, so this is not normally an issue.
- If the OS encounters an error (such as drive not ready when doing a DIR), the error message may be erased by this new behavior depending on the specific scenario.
- The TMS reset function was missing a call to re-initialize the CRT registers. This has been added.
- The SN76489 powers up in a dreadful state where it is emitting loud noise until initialized. Added a PREINIT entry point to initialize the registers earlier. Still not great, but a lot better.
Correct obsolete reference to DiskList.txt in the ReadMe.txt file in the Binary Directory.
Thanks and credit to @shawnshekari (issue #390) for reporting this.
- Generic serial interface client support files have been added. This mode of operation is not documented in the RomWBW User Guide. Refer to <https://github.com/durgadas311/cpnet-z80>.
- Minor patch to NULU so it does not spit out a directory listing every time it is invoked.
Based on input from Martin R, the console takeover mechanism has been modified to require that the space bar be pressed twice in a row on the character unit desiring to be the console. This reduces the chance of a spurious console takeover due to garbage input on serial ports.
- If enabled by AUTOCON config setting, you can press <space> at any character device to take over the console at the boot loader prompt.
- Correction to SIOSEEK documentation in System Guide. Thanks and credit to Martin R.
- Update PR template for new branching under v3.4
- Tweak Makefile.inc for slightly improved performance
- Improve UART driver messaging when bad CTS signal is detected
- Minor doc updates
- Comment updates in ansi.asm and tms.asm
- Correction to build process for ROMless systems
- Fix for ZRC (default configuration) to use all available RAM
- Split out the SD and USB specific support from the main CH37x driver to optimize code space usage.
- Ensure CH37x mode switch is handled properly when there are multiple devices active in a system.
I am encountering some CH376 chips that go haywire after a
reset command. They stop responding for a very long time.
I am seeing this only on "LC Tech" adapters and only on Z80
systems (not Z180). No idea what is going on, so I am
giving up for now and removing the reset.
- Last check-in had a couple of config file errors which are corrected here.
- Also updated Appendix A of the User Guide to reflect new CH37x port addressing.
- Updated to standardize on 0x3E/0x3F for primary CH device and 0x3C/0x3D for secondary CH device. Both devices are optional and detected automatically.
- The post-reset delay of both the IDE and PPIDE drivers has been extended. The SD-IDE adapters need more time to initialize before being ready to behave as proper IDE devices.
- Added support for the FDC section of the Duodyne Disk-IO board.
- Credit and thanks to Martin R for providing a substantial list of suggested fixes and improvements to the User Guide. I have done my best to address them -- others will require more time and will hopefully be addressed in the future.
- The 40 column mode of the TMS driver now conforms to the memory map from the TMS9918 documentation and is also now consistent with the existing TMS9918 video programs from the RC2014 forum.
When using the Z2 memory manager, if the HBIOS exceeds 16K, RTCDEFVAL will not be accessible prior to programming the Z2 memory bank registers. In this corner case the RTC latch could be mis-programmed. This commit introduces a workaround.
Wayne Warthen