The TMS driver would always install a USB keyboard pooling interrupt, despite no keyboard present
This would load the CPU and prevent other I/O operations
Removed ability from boot up (see ez80cpudrv.adm) to allow for setting of
external memory and io timing using eZ80's W/S config. All timings
must now be specified in terms of eZ80's Bus Cycle setting.
This is to increase compatibility with various external RCBus/RC2014
modules.
Due to eZ80 firmware changes, the configuration was always setting
the I/O Bus Cycle to the maximum setting of 7 - typically most system
will work fine at a setting of 4 B/C
current ez80 firmware version: 0.5.0.234 2025-06-21
Per @PeterOGB, added a copy of the HI-TECH C Compiler Messages document. The space-optimized version of the compiler from @Laci1953 omits these messages.
Jörg Linder has disassembled and thoroughly commented a great deal of the BPBIOS binaries. This was an incredible amount of work. I have added all of these to the RomWBW build scripts and will ultimately integrate them more completely.
Update ZAS in User Area 1 of the HITECH-C disk image based on the latest from Ladislau.
Co-Authored-By: ladislau szilagyi <87603175+Laci1953@users.noreply.github.com>
Use a global flag to indicate if interrupts have been enabled during the boot process. Drivers that operate in the pre-interrupt phase can use this to manage interrupt disable bracketing.
This allows restoring the location of interrupt enable in the boot process to it's proper location.
The consenus is that the best default address for Xosera
is $A0, so make that the default.
However, in that case we need to limit the number of UARTs
that are probed to two because the probe for a third UART
writes unlucky values to Xosera (at $A3) and causes it to
reconfigure itself and lock up the bus for a time.
Thanks to Wayne for his help in debugging this.
Corrected DSRTC and BQRTC drivers to properly restore the value of the NVRAM byte used for presence detection.
Thanks and credit to @MartinGieseCelonis for finding this.
KERMIT (for CP/M 2.2 and 3) was originally using a file transfer mode called "DEFAULT". This mode tried to accommodate binary or ASCII files, but could result in file corruption. As suggested by @PeterOGB, the BINARY setting is much safer. This is now the built-in default setting.
Co-Authored-By: PeterOGB <7755057+PeterOGB@users.noreply.github.com>