EtchedPixels [Sun, 15 Feb 2015 00:49:32 +0000 (00:49 +0000)]
Merge pull request #116 from geijoenr/master
move keyboard matrices to discard
geijoenr [Sat, 14 Feb 2015 16:02:47 +0000 (16:02 +0000)]
msx2: move keyboard matrices to discard
EtchedPixels [Sat, 14 Feb 2015 15:15:41 +0000 (15:15 +0000)]
Merge pull request #115 from willsowerbutts/rootdevparse2
Kernel: Generic boot device parsing method
geijoenr [Sat, 14 Feb 2015 15:15:17 +0000 (15:15 +0000)]
msx2: move some code to CODE2
to prevent overlapping of BOOT
Will Sowerbutts [Sat, 14 Feb 2015 14:22:43 +0000 (14:22 +0000)]
Kernel: Generic boot device parsing method
Alan Cox [Fri, 13 Feb 2015 23:19:07 +0000 (23:19 +0000)]
Makefile: add more of the targets
(Not all build right now)
Alan Cox [Fri, 13 Feb 2015 23:15:06 +0000 (23:15 +0000)]
start: tidy up the ifdefs for the boot parser
Alan Cox [Fri, 13 Feb 2015 23:14:44 +0000 (23:14 +0000)]
pcw8256: fix warnings with sdcc 3.4.2
Alan Cox [Fri, 13 Feb 2015 23:14:29 +0000 (23:14 +0000)]
ubee: fix devtty define
Alan Cox [Fri, 13 Feb 2015 23:14:04 +0000 (23:14 +0000)]
px4plus: clean up to current tree
This isn't a functional port yet - more a sanity checker on sizes
Alan Cox [Fri, 13 Feb 2015 23:06:32 +0000 (23:06 +0000)]
micropack: align with current builds
Alan Cox [Fri, 13 Feb 2015 23:06:08 +0000 (23:06 +0000)]
vt: don't assume devtty.h
If we are turned on then its needed, if not then it may be absent
Alan Cox [Fri, 13 Feb 2015 22:50:41 +0000 (22:50 +0000)]
msx2: add new discard.c
Alan Cox [Fri, 13 Feb 2015 22:46:43 +0000 (22:46 +0000)]
msx2: bring in line with discard changes
Alan Cox [Fri, 13 Feb 2015 22:30:23 +0000 (22:30 +0000)]
tres80: Bring in line with __DISCARD changes
Alan Cox [Fri, 13 Feb 2015 22:17:47 +0000 (22:17 +0000)]
z80pack: fix tty for new style writeready
Alan Cox [Fri, 13 Feb 2015 22:17:35 +0000 (22:17 +0000)]
z80pack-lite: Fix build
Alan Cox [Fri, 13 Feb 2015 22:13:03 +0000 (22:13 +0000)]
z80pack: fix build
EtchedPixels [Fri, 13 Feb 2015 22:01:51 +0000 (22:01 +0000)]
Merge pull request #114 from willsowerbutts/rootdevparse
Better root device name parsing
Alan Cox [Fri, 13 Feb 2015 21:45:26 +0000 (21:45 +0000)]
zx128: update README
Alan Cox [Fri, 13 Feb 2015 21:34:51 +0000 (21:34 +0000)]
zx128: update the microdrive code
Move it to device 8 so we can put real floppy controllers at zero. This
is still read only, and still somewhat silly but who cares.
Alan Cox [Fri, 13 Feb 2015 21:11:31 +0000 (21:11 +0000)]
devtty: fix bogus space
Alan Cox [Fri, 13 Feb 2015 21:10:59 +0000 (21:10 +0000)]
zx128: turn on unikey so '£' works
Alan Cox [Fri, 13 Feb 2015 21:10:24 +0000 (21:10 +0000)]
zx128: sort out the keyboard driver
This adds some of the missing keyboard symbols, arrow key handling etc and
also sorts out the inverted handling bug.
Alan Cox [Fri, 13 Feb 2015 21:09:50 +0000 (21:09 +0000)]
zx128: Fix stubs from common to do a bank restore
Turns out this is needed
Will Sowerbutts [Fri, 13 Feb 2015 20:02:08 +0000 (20:02 +0000)]
Kernel Makefile: Fix typo
Will Sowerbutts [Fri, 13 Feb 2015 20:49:33 +0000 (20:49 +0000)]
p112, n8vem-mark4: Provide boot device parsing method
Will Sowerbutts [Fri, 13 Feb 2015 20:44:37 +0000 (20:44 +0000)]
Kernel: Allow platforms to provide boot device parsing method
Alan Cox [Fri, 13 Feb 2015 19:54:06 +0000 (19:54 +0000)]
inode.c: rest of the fix too
Alan Cox [Fri, 13 Feb 2015 18:20:22 +0000 (18:20 +0000)]
TODO: update
Alan Cox [Fri, 13 Feb 2015 18:17:27 +0000 (18:17 +0000)]
inode: allow O_DIRECT reads building from Will's direct change
EtchedPixels [Fri, 13 Feb 2015 17:54:22 +0000 (17:54 +0000)]
Merge pull request #109 from willsowerbutts/readi
Kernel: readi() - when reading whole sectors, transfer them directly into user memory.
Alan Cox [Fri, 13 Feb 2015 17:05:57 +0000 (17:05 +0000)]
zx128: fix banking returns through common and discard
With this little gem fixed we can now get all the way through the IDE probe
in banked code and to the bootdev prompt.
Alan Cox [Fri, 13 Feb 2015 17:03:56 +0000 (17:03 +0000)]
devide: remove escaped debug, ifdef routine not always required
Alan Cox [Fri, 13 Feb 2015 15:46:58 +0000 (15:46 +0000)]
zx128: switch to mainline bankfixed
It now handles the cases that prevented it being used
Alan Cox [Fri, 13 Feb 2015 15:40:28 +0000 (15:40 +0000)]
zx128: use the ROM font
Saves us more memory
Alan Cox [Fri, 13 Feb 2015 15:40:08 +0000 (15:40 +0000)]
zx128: we have two banks so simplify their setup
Will Sowerbutts [Thu, 12 Feb 2015 14:33:57 +0000 (14:33 +0000)]
Kernel: readi() now does direct transfers only if CONFIG_LARGE_IO_DIRECT
is defined in the platform config.h
Will Sowerbutts [Thu, 12 Feb 2015 13:55:24 +0000 (13:55 +0000)]
Kernel: readi(): respect u_sysio flag
Will Sowerbutts [Wed, 11 Feb 2015 19:58:45 +0000 (19:58 +0000)]
Kernel: readi() - when reading whole sectors, transfer them directly
into user memory.
Alan Cox [Fri, 13 Feb 2015 15:25:18 +0000 (15:25 +0000)]
Kernel: revised rules.mk/target.mk
Alan Cox [Fri, 13 Feb 2015 15:24:18 +0000 (15:24 +0000)]
snprintf: first cut at snprintf and friends
Alan Cox [Fri, 13 Feb 2015 14:54:35 +0000 (14:54 +0000)]
process: correct behaviour of kill to process 1
In unixdom this has a magic exception that only handled signals are delivered
to process 1, those that would kill it are ignored.
Alan Cox [Fri, 13 Feb 2015 14:52:28 +0000 (14:52 +0000)]
Revert "Revert "Kernel: cpu-z180/rules.mk can just inherit everything from cpu-z80""
Put back the Z180 change
This reverts commit
0bcaf324642c0b16b34852d48ac5272ef6d4024a.
Alan Cox [Fri, 13 Feb 2015 14:51:13 +0000 (14:51 +0000)]
Makefiles: rework Will's changes in order to make overrides work again
We now use target.mk (mandatory) for per platform initial settings and we use
rules.mk (optional) in the platform directory after the CPU rules are processed
in order to override them
Alan Cox [Fri, 13 Feb 2015 14:34:16 +0000 (14:34 +0000)]
Revert "Kernel: cpu-z180/rules.mk can just inherit everything from cpu-z80"
Reverting the target/platform rules.mk changes - they break too much stuff
This reverts commit
02bcf7af9276326e94d3127786d7535f37baebef.
EtchedPixels [Fri, 13 Feb 2015 13:55:59 +0000 (13:55 +0000)]
Merge pull request #113 from willsowerbutts/z180rules
Kernel: cpu-z180/rules.mk can just inherit everything from cpu-z80
Will Sowerbutts [Fri, 13 Feb 2015 10:17:51 +0000 (10:17 +0000)]
Kernel: cpu-z180/rules.mk can just inherit everything from cpu-z80
Alan Cox [Thu, 12 Feb 2015 23:21:32 +0000 (23:21 +0000)]
zx128: first pieces towards a disciple disk driver
Lots of work needed yet.
Alan Cox [Thu, 12 Feb 2015 22:12:10 +0000 (22:12 +0000)]
zx128: rules.mk for banked
EtchedPixels [Thu, 12 Feb 2015 22:10:14 +0000 (22:10 +0000)]
Merge pull request #111 from willsowerbutts/platformcpu
Kernel: The platform rules.mk should define the CPU used
EtchedPixels [Thu, 12 Feb 2015 22:09:47 +0000 (22:09 +0000)]
Merge pull request #110 from willsowerbutts/discard
ds1302, devsd: Split code for the DISCARD segment into separate files.
Will Sowerbutts [Thu, 12 Feb 2015 21:01:52 +0000 (21:01 +0000)]
Kernel: The platform rules.mk should define the CPU used
This means that only the PLATFORM variable needs to be specified in
Kernel/Makefile, instead both PLATFORM and CPU in combination.
Will Sowerbutts [Thu, 12 Feb 2015 20:49:49 +0000 (20:49 +0000)]
ds1302, devsd: Split code for the DISCARD segment into separate files.
Includes required platform changes for n8vem-mark4, p112.
Alan Cox [Thu, 12 Feb 2015 02:02:30 +0000 (02:02 +0000)]
patches/SDCC: Add another change to put string constants in const not code
Alan Cox [Thu, 12 Feb 2015 01:54:47 +0000 (01:54 +0000)]
z80: split the IDE driver __DISCARD up and remove the feature
We can't have multiple banked and unbanked segments in one object file as
we then have no way to make the banked linker and binmunge fix them up. The
common piece is ok as it doesn't call out to other segments in the same file
likewise calls to common code is ok. Calls betweem banks however don't end
well.
Alan Cox [Thu, 12 Feb 2015 01:53:45 +0000 (01:53 +0000)]
zx128: note bank changes needed
Alan Cox [Thu, 12 Feb 2015 01:53:23 +0000 (01:53 +0000)]
zx128: put the udata *first* so our maps are right
Alan Cox [Thu, 12 Feb 2015 01:52:37 +0000 (01:52 +0000)]
zx128: tricks.s updates
This still needs a lot of work doing but it's a starting point and we don't
crash waiting for input
Alan Cox [Thu, 12 Feb 2015 01:52:20 +0000 (01:52 +0000)]
lkrloc3: typo
Alan Cox [Thu, 12 Feb 2015 01:51:56 +0000 (01:51 +0000)]
zx128: video routine updates for banked mode
Alan Cox [Thu, 12 Feb 2015 01:51:28 +0000 (01:51 +0000)]
zx128: move uarea
We need 0xFFFx free in order to use it for the IM2 stunts needed on the
Spectrum boxes
Alan Cox [Thu, 12 Feb 2015 01:50:41 +0000 (01:50 +0000)]
binmunge: zeroing, intialized data, stub changes etc
Alan Cox [Thu, 12 Feb 2015 01:49:04 +0000 (01:49 +0000)]
zx128: update the helpers and banking stubs
Alan Cox [Thu, 12 Feb 2015 01:48:26 +0000 (01:48 +0000)]
zx128: update to match changes in blkdev and devide
Alan Cox [Thu, 12 Feb 2015 01:48:05 +0000 (01:48 +0000)]
zx128: update the mappings
Alan Cox [Thu, 12 Feb 2015 01:47:13 +0000 (01:47 +0000)]
z80: usermem banked mode uses some new helpers
We may need to generalise these. The good thing is that the save/restore
versions on a non banked kernel are simply the map_kernel/map_process_any
methods
Alan Cox [Thu, 12 Feb 2015 01:46:22 +0000 (01:46 +0000)]
zx128: update README
Alan Cox [Thu, 12 Feb 2015 01:44:16 +0000 (01:44 +0000)]
z80: banked mode need modified helpers so stick them into the tree
If upstream SDCC grows banking support we can tidy this back up but for now
a more generalised banking setup in SDCC is some ways off
Alan Cox [Thu, 12 Feb 2015 01:42:54 +0000 (01:42 +0000)]
z80: update banking documentation
Alan Cox [Wed, 11 Feb 2015 00:29:59 +0000 (00:29 +0000)]
z80: make sure the helpers appear in common
Alan Cox [Wed, 11 Feb 2015 00:28:51 +0000 (00:28 +0000)]
zx128: Further banking fixes
With these in place we get as far as trying to call entries in a function table
load 0xFFFF into HL and jump to it. Probably more binmunge debugging needed
at this point.
Alan Cox [Tue, 10 Feb 2015 23:43:25 +0000 (23:43 +0000)]
patches: SDCC patches for playing with banked mode
Alan Cox [Tue, 10 Feb 2015 23:38:35 +0000 (23:38 +0000)]
zx128: update the various ZX128 bits needed
Alan Cox [Tue, 10 Feb 2015 23:37:52 +0000 (23:37 +0000)]
lowlevel-z80-banked: remove the NULL checker if low memory is not userspace
Alan Cox [Tue, 10 Feb 2015 23:36:14 +0000 (23:36 +0000)]
bin2sna: Add a tool to turn out banked ZX Spectrum 128 binaries into .SNA
This provides a quick and handy way to prove they aren't working yet.
The tool loads the common memory, bank1/2/3.bin into the right banks in the
SNA image and then executes from 0xC003 (0xC000 is the main entry point,
0xC003 is a hack for now for a preloaded and positioned image)
At that point we hit crt0.s and crash somewhere early on trying to bank switch
Alan Cox [Mon, 9 Feb 2015 23:00:09 +0000 (23:00 +0000)]
zx128: banking helpers
Just filling them out for the moment. They try and use logical not real
bank numbers so won't actually work yet
Alan Cox [Sun, 8 Feb 2015 17:04:41 +0000 (17:04 +0000)]
Makefile: allow banked platforms to force different helpers
Alan Cox [Sun, 8 Feb 2015 17:04:00 +0000 (17:04 +0000)]
binmunge: Revised version of the tool
This works with the --external-banker patch to SDCC to do the needed inline
fixups and stubs.
Alan Cox [Sun, 8 Feb 2015 17:01:57 +0000 (17:01 +0000)]
6809: build the decb tool
Alan Cox [Sun, 8 Feb 2015 17:01:29 +0000 (17:01 +0000)]
z80: commit banked helpers
These are annoyingly similar and eventually want folding together
Alan Cox [Sun, 8 Feb 2015 17:00:54 +0000 (17:00 +0000)]
zx128: Make the adjustments in the asm needed for banked binary support
Alan Cox [Sun, 8 Feb 2015 17:00:32 +0000 (17:00 +0000)]
start: Fix warnings if program space is > signed size
Alan Cox [Sun, 8 Feb 2015 15:54:15 +0000 (15:54 +0000)]
Makefile: further tidying
Alan Cox [Sun, 8 Feb 2015 15:30:52 +0000 (15:30 +0000)]
Makefile: Move the per CPU stuff into the CPU directory
Before the Makefiles get out of hand, impose some kind of sanity. We will also
need the platform bits for things like ZX Spectrum 128 where we need to add
our new --external-banker flags and the like.
Alan Cox [Sun, 8 Feb 2015 15:21:25 +0000 (15:21 +0000)]
Z80: update the banking logic based on the SDCC changes made
Alan Cox [Sun, 8 Feb 2015 12:35:55 +0000 (12:35 +0000)]
process.c: scrub final fastcall refs
Alan Cox [Sun, 8 Feb 2015 12:34:40 +0000 (12:34 +0000)]
process.c: remove __fastproc__ now it is obsolete
Alan Cox [Sun, 8 Feb 2015 12:30:56 +0000 (12:30 +0000)]
tty.h: typo fix
Alan Cox [Sun, 8 Feb 2015 12:30:35 +0000 (12:30 +0000)]
tty.h: and clean out CODE1 bits
Alan Cox [Sun, 8 Feb 2015 12:30:03 +0000 (12:30 +0000)]
tty.h: clean CODE2 out
Alan Cox [Sun, 8 Feb 2015 12:24:34 +0000 (12:24 +0000)]
banking: Strip out all the experiments with CODE1 etc
By now it's obvious that banking needs to be done per platform and the logic
really wants to be in the linker. Therefore get rid of all the annoying noise.
In addition with the patches to cc65 we don't need __fastcall__ hacks, and
in fact as we can't fastcall a banked 6502 function we don't want fastcalls
leaking anywhere explicit.
Alan Cox [Sun, 8 Feb 2015 00:35:47 +0000 (00:35 +0000)]
z80: fix if/endifs for non SDCC case
Alan Cox [Fri, 6 Feb 2015 13:41:43 +0000 (13:41 +0000)]
Makefile: switch 68000 toolchain
Alan Cox [Fri, 6 Feb 2015 13:40:31 +0000 (13:40 +0000)]
68000: update the atari-st beginnings, switch tool chain
We now use the atari mint toolchain as it produces TOS binaries and uses the
_ prefixes we expect from compilers.
EtchedPixels [Tue, 3 Feb 2015 22:37:14 +0000 (22:37 +0000)]
Merge pull request #104 from willsowerbutts/monitorbank
monitor-z80: allow platforms to move monitor code to different segments (updates p112, n8vem-mark4)
EtchedPixels [Tue, 3 Feb 2015 22:36:57 +0000 (22:36 +0000)]
Merge pull request #103 from willsowerbutts/prtroot
prtroot: Bug fix, new -i flag to initialise /etc/mtab
Will Sowerbutts [Mon, 2 Feb 2015 23:28:57 +0000 (23:28 +0000)]
monitor-z80: allow moving to different segments (updates p112, n8vem-mark4)
Will Sowerbutts [Tue, 3 Feb 2015 00:35:01 +0000 (00:35 +0000)]
prtroot: Bug fix, new -i flag to initialise /etc/mtab
Bug fix: prtroot would write a 0 byte at the end of its output.
The new "-i" flag causes output to go into /etc/mtab instead of to
stdout.
Alan Cox [Tue, 3 Feb 2015 10:59:39 +0000 (10:59 +0000)]
socz80: fix timer, improve irq code a shade