summary |
shortlog | log |
commit |
commitdiff |
tree
first ⋅ prev ⋅ next
Alan Cox [Mon, 10 Nov 2014 01:00:23 +0000 (01:00 +0000)]
tty: fix bracketing
EtchedPixels [Sun, 9 Nov 2014 11:23:58 +0000 (11:23 +0000)]
Merge pull request #11 from atsidaev/zx128-clean
zx128: Initial ZX Spectrum 128 support
Alexander Tsidaev [Sun, 9 Nov 2014 09:51:59 +0000 (14:51 +0500)]
zx128: Initial ZX Spectrum 128 support
Alan Cox [Sat, 8 Nov 2014 23:25:06 +0000 (23:25 +0000)]
msx1: scan for AB rom signatures, hash the first 2K
First steps to detecting the disks etc
Alan Cox [Sat, 8 Nov 2014 23:24:51 +0000 (23:24 +0000)]
devio: use "else" in kprintf
Alan Cox [Sat, 8 Nov 2014 22:40:00 +0000 (22:40 +0000)]
msx1/tricks.s: Fix missing globals
Alan Cox [Sat, 8 Nov 2014 22:34:23 +0000 (22:34 +0000)]
socz80: fix tty, more alignment with the current codebase
Still needs chunks of code writing for the fork copier and actual MMU loading
Alan Cox [Sat, 8 Nov 2014 22:11:32 +0000 (22:11 +0000)]
pcw8256,README: Instructions for build/boot
Tested build and boot instructions to get a bootable kernel image. This is
a raw bootable image so you can boot it on joyce without owning any of the
Locoscript or CP/M 'start of day' discs
Alan Cox [Sat, 8 Nov 2014 22:11:16 +0000 (22:11 +0000)]
pcw8256: comment out debugging for keyboard
Alan Cox [Sat, 8 Nov 2014 22:10:52 +0000 (22:10 +0000)]
pcw8256: update to the new major/minor numbers
Without this we panic with no tty on boot
Alan Cox [Fri, 7 Nov 2014 20:51:12 +0000 (20:51 +0000)]
msx1: memory mapping and other logic
The big TODO for userspace here is to implement forkcopy (copying 32K between
two blocks of MegaRAM)
Alan Cox [Fri, 7 Nov 2014 00:50:09 +0000 (00:50 +0000)]
README: Random ramblings about how a port might work
Alan Cox [Fri, 7 Nov 2014 00:49:29 +0000 (00:49 +0000)]
msx1: sketch out the bits needed to work the brain dead slot system
Plus find the megarams.
All untested.
Alan Cox [Thu, 6 Nov 2014 22:22:54 +0000 (22:22 +0000)]
msx1: Add the initial bits
This is basically in parallel with MSX2 but with the various memory manager
bits stripped out (new one needed) and 40 char vdp code
Alan Cox [Thu, 6 Nov 2014 21:47:47 +0000 (21:47 +0000)]
msx2: fix crash on outchar debug
Alan Cox [Thu, 6 Nov 2014 21:38:39 +0000 (21:38 +0000)]
msx: Add initial (non working) bootrom test, also make it build fuzix.com
Alan Cox [Thu, 6 Nov 2014 21:38:24 +0000 (21:38 +0000)]
binman: minor cleaning/commenting
Alan Cox [Thu, 6 Nov 2014 21:38:03 +0000 (21:38 +0000)]
vdp: scrolling sizes are wrong
Fix some hex/dec confusions
Alan Cox [Thu, 6 Nov 2014 12:15:05 +0000 (12:15 +0000)]
msx2: fix video
With these small vdp corrections we have a proper 80x25 display
Alan Cox [Thu, 6 Nov 2014 00:38:56 +0000 (00:38 +0000)]
platforms: add platform_interrupt_all everywhere
Alan Cox [Thu, 6 Nov 2014 00:31:28 +0000 (00:31 +0000)]
lowlevel-z80: various (important) fixes
1. Exit with reti, most platforms don't care but a few do so avoid
digging holes
2. MSX2 is the first platform we handle where not servicing an IRQ
jams you in a loop. Provide a hook to always service some interrupts
even if we are on an IRQ stack or otherwise detained from calling
platform_interrupt
3. The di() call saved the flags but forgot to DI, this fixes several
rather odd bugs I'm sure
4. The signal processing on irq return is suspect, note it for future
sorting out.
With this and the MSX2 fixes we get closer to a basic MSX2 boot
Alan Cox [Thu, 6 Nov 2014 00:29:17 +0000 (00:29 +0000)]
msx2: various platform side and vdp fixes
With these and the small core fixes to follow we blow up for reasons I don't
yet understand in the uput() call in create_init. Presumably some kind of
mapping setup bug. With that commented we hit the boot device request, at
which point we have no keyboard.
Display is still wonky. Guess it wants an MSXpert to figure out whats wrong
with the 80x25 setup
Alan Cox [Wed, 5 Nov 2014 22:37:17 +0000 (22:37 +0000)]
msx2: Make the cursor work
I was clearly not awake pasting the initial cursor code together.
This gives us a correct text rendering, except for the mysterious double
copy which presumably is a VDP set up error of some kind.
Alan Cox [Wed, 5 Nov 2014 22:29:16 +0000 (22:29 +0000)]
README.md: quick follow up fix
EtchedPixels [Wed, 5 Nov 2014 22:28:34 +0000 (22:28 +0000)]
Merge pull request #8 from marioaugustorama/master
Fix formatting.
Alan Cox [Wed, 5 Nov 2014 20:49:32 +0000 (20:49 +0000)]
msx2: further work on video - closer
We now get the font, and most of the text. The move to a new line keeps
wiping the start of the next, and it appears twice.
Alan Cox [Wed, 5 Nov 2014 19:20:56 +0000 (19:20 +0000)]
syscall_other: optimisations
Cache the inode calculation in execve, otherwise each loop pass must recompute
it as the compiler can't know the functions don't touch it.
Add a couple of staticfasts
Alan Cox [Wed, 5 Nov 2014 19:20:30 +0000 (19:20 +0000)]
filesys/syscall2: optimise a few spots with staticfast
Alan Cox [Wed, 5 Nov 2014 19:18:51 +0000 (19:18 +0000)]
cpu: introduce "staticfast"
On Z80 and to an extent 6502 we gain a lot if we can pin some of our
auto variables to static addresses. Clearly this can only be done for
non-re-entrant functions but our architecture means we have plenty of them
6809 is not really so aflicted so uses auto.
Mario Nunes [Wed, 5 Nov 2014 17:27:29 +0000 (17:27 +0000)]
Fix formatting.
Alan Cox [Wed, 5 Nov 2014 14:06:23 +0000 (14:06 +0000)]
From
2ea7e7e52e5e2411bb91415f20d593dc8131b655 Mon Sep 17 00:00:00 2001
From: root <root@magniette.org>
Date: Wed, 5 Nov 2014 09:53:25 +0100
Subject: Clean some forgotten files in make clean + ignore some generated files in gitignore
Alan Cox [Wed, 5 Nov 2014 13:59:22 +0000 (13:59 +0000)]
pcw8256: add uzi.lnk that was missed from the commits
Noted by github:joyce8256
EtchedPixels [Wed, 5 Nov 2014 13:55:28 +0000 (13:55 +0000)]
Merge pull request #6 from Rappalot/master
Convert zmac format source files to sdasz80
Rappalot [Wed, 5 Nov 2014 01:50:22 +0000 (02:50 +0100)]
Convert zmac format source files to sdasz80
Alan Cox [Wed, 5 Nov 2014 00:08:32 +0000 (00:08 +0000)]
msx2: initial bootstrap debugging
We get as far as enabling interrupts then curl up and die. The VDP
programming also doesn't set text80 but does produce some interesting
flashing blobs
Alan Cox [Tue, 4 Nov 2014 20:36:49 +0000 (20:36 +0000)]
Makefile/link: Update all the other link files to match the previous change
EtchedPixels [Tue, 4 Nov 2014 20:32:35 +0000 (20:32 +0000)]
Merge pull request #4 from gmacd/master
Add condition to build z80pack on Mac
Graham MacDonald [Tue, 4 Nov 2014 08:42:35 +0000 (08:42 +0000)]
Add condition to build z80pack on Mac
Alan Cox [Tue, 4 Nov 2014 00:07:11 +0000 (00:07 +0000)]
msx2: Further random guesswork as we go - memory sizing
Alan Cox [Mon, 3 Nov 2014 21:55:00 +0000 (21:55 +0000)]
msx2: RAM fixup
Apparently RAM is added in the reverse order to tht expected - 3 is at 0x000
2 is at 0x4000 etc. Add them properly then add 0 so that init picks up the
initial 0xC000 bank for common and udata
Alan Cox [Mon, 3 Nov 2014 21:45:13 +0000 (21:45 +0000)]
msx2: Add an initial wild guess at what MSX2 will need
Hopefully that'll be a basis for someone who knows something about MSX2
to fill in the blanks. In particular my VDP code is a wild untested guess
Alan Cox [Mon, 3 Nov 2014 19:26:53 +0000 (19:26 +0000)]
Makefile: call the final result fuzix.bin not uzi.bin
Still need to rename all the temporaries
Alan Cox [Mon, 3 Nov 2014 19:26:00 +0000 (19:26 +0000)]
README: first take at instructions for building
No these are not sufficient yet
EtchedPixels [Mon, 3 Nov 2014 19:24:23 +0000 (19:24 +0000)]
Merge pull request #3 from erkinalp/master
Converted all function definitions into prototype style in mkfs.c
erkinalp [Mon, 3 Nov 2014 16:38:15 +0000 (18:38 +0200)]
Converted all function definitions into prototype style in mkfs.c
Alan Cox [Mon, 3 Nov 2014 00:11:06 +0000 (00:11 +0000)]
pcw8256,fdc765: not the architectural question around banking and swap
Basically the problem is this:
On a 16K banked machine we have 0x0000-0xBFFF holding user code and we can
switch between processes low 48K with gay abandon. However the top 16K holds
process memory, common, *and* our stacks.
What probably needs to happen is that either we bounce such buffers (slow) or
teach the swap device driver in the platform to do bank flipping.
The problem with this last one is that
- it has to be done per platform
- some machines have brain damaged banking where a given address in a bank
cannot be remapped to a different virtual address
Unfortunately it's not obvious how to create a generic solution that doesn't
punish sensibly designed machines.
Alan Cox [Mon, 3 Nov 2014 00:06:33 +0000 (00:06 +0000)]
pcw8256,fdc765.s: Use outi not ini for writes, comment the code fully
This still doesn't explain why we misparse the status errors.
Alan Cox [Sun, 2 Nov 2014 23:52:37 +0000 (23:52 +0000)]
emulator: Add code for lseek() rather than the UZI v6 seek
Not yet tested, but this should be the basis to bootstrap CP/M emulation
code in the OS environment. At that point we can retarget the UZIX native
compiler setup for FUZIX and have a native compiler for userspace (yay!)
Alan Cox [Sun, 2 Nov 2014 18:37:32 +0000 (18:37 +0000)]
px4plus: update comments, remember to swap the uarea
Alan Cox [Sun, 2 Nov 2014 18:37:12 +0000 (18:37 +0000)]
swap: Fix silly bug
Alan Cox [Sat, 1 Nov 2014 17:07:35 +0000 (17:07 +0000)]
nc100/200: update notes, start adding NC200 support
Alan Cox [Sat, 1 Nov 2014 15:45:18 +0000 (15:45 +0000)]
px4 init: Remember to set the video base !
Alan Cox [Sat, 1 Nov 2014 15:40:33 +0000 (15:40 +0000)]
Makefile: introduce the _VIDEO segment, rebalance CODE and CODE2
Not that the linker can bank the things yet...
Alan Cox [Sat, 1 Nov 2014 15:40:13 +0000 (15:40 +0000)]
crt0.s: Prepare for _VIDEO
Alan Cox [Sat, 1 Nov 2014 15:39:36 +0000 (15:39 +0000)]
crt0: Prepare for _VIDEO area
Alan Cox [Sat, 1 Nov 2014 15:38:06 +0000 (15:38 +0000)]
px4plus: Add initial (untested) video, and adjust memory maps etc
Add an init.c for code we can discard.
Alan Cox [Sat, 1 Nov 2014 15:37:24 +0000 (15:37 +0000)]
vt: We are going to make this more bankable, add warnings
Alan Cox [Sat, 1 Nov 2014 15:35:44 +0000 (15:35 +0000)]
binman: Teach binman some new tricks
Allow for discard being low
Allow for initializer being low (indicating we are not doing common packing
and the classic 'copy up' image)
Spot data/const overlapping for cases we force data start (PX4Plus)
Smart end/start identification for complex images rather than the classic
image.
Hopefully classic images still work
Alan Cox [Sat, 1 Nov 2014 15:34:45 +0000 (15:34 +0000)]
start: move more discardables into _DISCARD
Alan Cox [Sat, 1 Nov 2014 15:34:04 +0000 (15:34 +0000)]
px4plus: note the state of play
Alan Cox [Sat, 1 Nov 2014 15:33:01 +0000 (15:33 +0000)]
font8x8: allow 1K font
We don't always want to blow 2K on fonts, so allow a 1K font option
Alan Cox [Sat, 1 Nov 2014 12:34:52 +0000 (12:34 +0000)]
platform-px4: Add the PX4Plus
The question was asked and its an interesting one as the PX4Plus is probably
right on the limit of what we can handle if we use the ROM pods.
This is untested but builds and also not very useful as it doesn't yet have
the vt driver, keyboard, font or floppy. It does seem to prove that if you
breathe in then you can fit the Z80 core into 32K of ROM with 8K for the
common/udata and put the video/floppy/fonts banked into the top of the RAM
area. That would still allow something like 48K binaries to run, and as the
PX4Plus has at best 128K of RAMdisc for swapping 32K is probably a saner
limit anyway.
And no.. I do not think swapping to a 38400 baud serial floppy would rock.
Alan Cox [Fri, 31 Oct 2014 22:59:37 +0000 (22:59 +0000)]
nc200: Initial pass at NC200 support as well as the NC100
Giving the floppy drive a miss as it seems to be completely undocumented
and will need research. May be easier to port CP/M 3 to it first for
investigation.
Alan Cox [Fri, 31 Oct 2014 22:56:54 +0000 (22:56 +0000)]
README.md: mutter mutter formatting crap, whats wrong with ASCII
Alan Cox [Fri, 31 Oct 2014 20:25:33 +0000 (20:25 +0000)]
pcw8256,devtty: Commit some of the shift keys that were missing
Alan Cox [Fri, 31 Oct 2014 19:15:45 +0000 (19:15 +0000)]
PORTING: Clarify map_save/restore
Alan Cox [Fri, 31 Oct 2014 17:07:28 +0000 (17:07 +0000)]
pcw8256: resync with the core updates
Alan Cox [Fri, 31 Oct 2014 16:46:14 +0000 (16:46 +0000)]
Makefile: remove some more bits of legacy uzi.com building
Alan Cox [Fri, 31 Oct 2014 16:32:50 +0000 (16:32 +0000)]
Makefile: remove reference to removed loader code
Alan Cox [Fri, 31 Oct 2014 16:32:24 +0000 (16:32 +0000)]
socz80: Add the SocZ80 platform README
Alan Cox [Fri, 31 Oct 2014 16:26:42 +0000 (16:26 +0000)]
socz80: Add the target
This code isn't yet complete. It's had a first pass conversion to the revised
kernel interfaces but the MMU logic has not yet been wired up to bank16 or
converted to the new format API.
Alan Cox [Fri, 31 Oct 2014 16:24:54 +0000 (16:24 +0000)]
execve: Correct cpm emulator loading
We moved to a real off_t and lseek behaviour but that didn't get propogated
into the CP/M emulator loading logic.
Alan Cox [Fri, 31 Oct 2014 16:24:26 +0000 (16:24 +0000)]
pcw8256: don't define un-needed UAREA bits
This avoids dragging in extra code we won't be using
Alan Cox [Fri, 31 Oct 2014 14:37:20 +0000 (14:37 +0000)]
Standalone: Add the standalone utilities
These are based off the UZI-WRS versions from Will Sowerbutts except for
chmem which is new code.
Alan Cox [Fri, 31 Oct 2014 12:17:21 +0000 (12:17 +0000)]
README.md: put the main notes here
Alan Cox [Fri, 31 Oct 2014 12:14:41 +0000 (12:14 +0000)]
README: remove notes leaving historical bits here
Alan Cox [Fri, 31 Oct 2014 12:13:21 +0000 (12:13 +0000)]
fuzix: Halloween Release
First cut of the development tree. Not terribly useful but hey its hackable
and boots.
EtchedPixels [Fri, 31 Oct 2014 11:46:47 +0000 (11:46 +0000)]
Create README.md
EtchedPixels [Fri, 31 Oct 2014 11:44:15 +0000 (11:44 +0000)]
Initial commit