Alan Cox [Sat, 19 Jan 2019 19:17:03 +0000 (19:17 +0000)]
68000: add some very crude basic register dumping on traps
We need to print the proper trap number and do the frame processing somewhere,
this is for now partly just a hack.
Alan Cox [Sat, 19 Jan 2019 19:16:20 +0000 (19:16 +0000)]
tiny68k: remember to fix memcpy etc
Alan Cox [Sat, 19 Jan 2019 19:15:39 +0000 (19:15 +0000)]
syscall_exec32: set top properly
Alan Cox [Sat, 19 Jan 2019 19:14:56 +0000 (19:14 +0000)]
usermem: fix dumb error where some out of range addresses were not errorred
Alan Cox [Sat, 19 Jan 2019 19:14:12 +0000 (19:14 +0000)]
inode: Fix non 32bit clean path
Alan Cox [Sat, 19 Jan 2019 19:13:48 +0000 (19:13 +0000)]
68000: update to build needed code
Alan Cox [Sat, 19 Jan 2019 16:27:52 +0000 (16:27 +0000)]
tiny68k: TTY driver fixes
Alan Cox [Sat, 19 Jan 2019 16:27:36 +0000 (16:27 +0000)]
tiny68k: fix off by one in swap map
Alan Cox [Sat, 19 Jan 2019 16:27:22 +0000 (16:27 +0000)]
tiny68k: low level IRQ and serial fixes
Alan Cox [Sat, 19 Jan 2019 16:27:04 +0000 (16:27 +0000)]
tiny68k: correct IDE regs for tiny68k
Alan Cox [Sat, 19 Jan 2019 16:26:51 +0000 (16:26 +0000)]
tiny68k: make the loader work
Alan Cox [Sat, 19 Jan 2019 16:26:17 +0000 (16:26 +0000)]
tiny68k: start up fixes
Alan Cox [Sat, 19 Jan 2019 16:25:49 +0000 (16:25 +0000)]
cpu: add a barrier() macro
Alan Cox [Sat, 19 Jan 2019 16:25:02 +0000 (16:25 +0000)]
tiny68k: don't assume a fixed swapfile (we do dynamic now)
Tormod Volden [Wed, 16 Jan 2019 22:43:08 +0000 (23:43 +0100)]
Build V7 and MWC sort commands on 6809
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Thu, 10 Jan 2019 21:44:34 +0000 (22:44 +0100)]
games: Build fortune.dat in Makefile.6809 too
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Thu, 10 Jan 2019 23:46:59 +0000 (00:46 +0100)]
fortune: Count is also fixed endian
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Alan Cox [Wed, 16 Jan 2019 22:19:01 +0000 (22:19 +0000)]
tiny68K; just an initial commit
Alan Cox [Wed, 16 Jan 2019 14:59:52 +0000 (14:59 +0000)]
binnam: handle platforms with an _BOOT section leading
We need in some cases to put a bootstrap entry below everything else and not
have code first. In that case binman needs to back from boot not from code.
Alan Cox [Wed, 16 Jan 2019 14:59:17 +0000 (14:59 +0000)]
syscall_exec32: 32bit pieces for core dump (incomplete)
Just to get us up and running
Alan Cox [Wed, 16 Jan 2019 14:58:44 +0000 (14:58 +0000)]
zxevo: more work in progress getting towards an actual initial port
Alan Cox [Wed, 16 Jan 2019 14:57:55 +0000 (14:57 +0000)]
tbblue: fill in more uart information c/o the zx next forum
Alan Cox [Wed, 16 Jan 2019 14:55:10 +0000 (14:55 +0000)]
sam: fix process size reporting
Alan Cox [Wed, 16 Jan 2019 14:54:34 +0000 (14:54 +0000)]
linc80: switch to the new SIO macros and buffering
Alan Cox [Wed, 16 Jan 2019 14:53:36 +0000 (14:53 +0000)]
linc80: dependancy
Alan Cox [Wed, 16 Jan 2019 14:52:12 +0000 (14:52 +0000)]
z80softspi: from 14K to 19K
Unroll and apply a small optimization from herrw@pluspora to get 19K/sec or
so receive with an 8MHz Z80. At some point I need to test this on a faster
part and see if the signalling is clean enough to get higher still.
Alan Cox [Wed, 16 Jan 2019 14:51:49 +0000 (14:51 +0000)]
z80sio: fix smart serial wrap bug
Alan Cox [Wed, 16 Jan 2019 14:51:28 +0000 (14:51 +0000)]
68000: Fix LEVEL_2 on 32bit systems
Alan Cox [Wed, 16 Jan 2019 14:51:14 +0000 (14:51 +0000)]
credits: update for fortune db
Alan Cox [Wed, 16 Jan 2019 14:50:18 +0000 (14:50 +0000)]
V7: add the V7 sort (ANSIfied)
It's very compact so might be our best 8bit option
Alan Cox [Wed, 16 Jan 2019 14:49:33 +0000 (14:49 +0000)]
MWC/sort: Add another sort candidate
The one we have in util right now isn't useful for an 8bit system, the
MWC one sort of is and scales well
Alan Cox [Wed, 16 Jan 2019 14:49:14 +0000 (14:49 +0000)]
cpuinfo: small updates for T80 and Soviet devices
Alan Cox [Wed, 16 Jan 2019 14:47:22 +0000 (14:47 +0000)]
bank16: we should use mapbase not progbase
Fixes I hope the COCO breakage
Alan Cox [Fri, 11 Jan 2019 00:45:57 +0000 (00:45 +0000)]
zx evolution: initial port sketch
Whilst quite different to tbblue in some ways it's useful to be thinking about
both at the same time.
This is for baseconf. tsconf might be even more interesting ?
Alan Cox [Thu, 10 Jan 2019 14:21:29 +0000 (14:21 +0000)]
fortune: Ship the computer db as a default
There are no shortage of fortune files but this one is a nice bit computing
history and hopefully doesn't contain anything too inappropriate
Alan Cox [Thu, 10 Jan 2019 13:13:03 +0000 (13:13 +0000)]
cromemco: switch to proper IRQ driven serial bufers
Not yet tested
Tormod Volden [Wed, 9 Jan 2019 10:33:56 +0000 (11:33 +0100)]
bank8k: Fix last bank size
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Wed, 9 Jan 2019 10:36:40 +0000 (11:36 +0100)]
lowlevel-6809: Simplify interrupt_handler after kernel_flag was removed
Rewrite the null vector in the same mapping as where we check for it.
Also simplify comments in the syscall glue.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Alan Cox [Thu, 10 Jan 2019 12:32:19 +0000 (12:32 +0000)]
games: build fortune.dat for inclusion
Also make the fortune data file fixed endian so we can cross build it nicely
for everyone
Alan Cox [Wed, 9 Jan 2019 22:19:49 +0000 (22:19 +0000)]
z80sio: tidy up the macros a bit
We still need to think about how IM1 is best done
Alan Cox [Wed, 9 Jan 2019 19:27:50 +0000 (19:27 +0000)]
linc80: tty updates
Alan Cox [Wed, 9 Jan 2019 13:47:50 +0000 (13:47 +0000)]
rc2014: move the default ppide base address
Somewhere along the line RC2014 picked 0x20 rather than the 0x60 that the
retrobrew based systems used.
Remember to change the jumpers if currently using 0x60.
Alan Cox [Wed, 9 Jan 2019 03:08:08 +0000 (03:08 +0000)]
linc80: continued work on proper IM2 support
Add IM2 driven serial interrupts with proper queues. We don't yet do all the
status and error bits right.
This enables IM2 serial but we don't gain much from it because we have yet
to tackle the core changes needed to leave other interrupt lines running for
most cases where we today DI, and to soft count, skip re-entry of the
interrupt_handler/timer proper.
Alan Cox [Tue, 8 Jan 2019 23:39:46 +0000 (23:39 +0000)]
linc80: fix reboot
Alan Cox [Tue, 8 Jan 2019 23:39:28 +0000 (23:39 +0000)]
linc80: fix tty warning
Alan Cox [Tue, 8 Jan 2019 23:39:11 +0000 (23:39 +0000)]
z80softspi: fix timings
Alan Cox [Tue, 8 Jan 2019 23:38:13 +0000 (23:38 +0000)]
kernel: allow per major decision on direct to user space
This is useful if you've got a hard disk and a floppy. For all current
platforms it resolves down to the same code as before but the flexibility is
now there if you need it.
Tormod Volden [Mon, 7 Jan 2019 22:08:41 +0000 (23:08 +0100)]
dragon-mooh: drivewire adaptions to page map switching
Needs a tiny change from the nx32 version.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 22:08:40 +0000 (23:08 +0100)]
dragon-mooh: Only rewrite MMU when user map changes
map_process_a sets up the MMU from the given page map vector.
map_process_always activates the user map.
map_save and map_restore only deals with kernel vs user map, and not
which user map is in use.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 22:08:39 +0000 (23:08 +0100)]
dragon-mooh: Copy only unique banks on fork
Once we reach a duplicated bank in the table, we are done.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 22:08:38 +0000 (23:08 +0100)]
dragon-mooh: Disable interrupts while writing task registers
map_kernel/map_process_always are in a few cases called with interrupt
enabled (e.g. from SPI transfer routines in common). If an interrupt
hits between writing "user" to map_copy and reading the process page
map, the intermediate map_save/map_restore during interrupt will restore
"user" map, and the process page map is gone.
Make map_kernel/map_process_always "atomic" to ensure that map_copy and
task registers are always coherent.
Note that we still rely on map_process_always/map_process_x to be called
with kernel being mapped so that we can read the process page map.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 22:08:37 +0000 (23:08 +0100)]
dragon-mooh: Move udata stash to 0x2000
Also add the missing link file for applications
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Alan Cox [Tue, 8 Jan 2019 17:05:18 +0000 (17:05 +0000)]
bank16: if PROGBASE is non zero we do the maths wrong
None of our bank16k ports have a non zero progbase fortunately. Fix it anyway
Alan Cox [Tue, 8 Jan 2019 17:02:53 +0000 (17:02 +0000)]
bank8k: allow for no zero base
Tormod's patch worked by changing pagemap_realloc to pass top not size. In
fact we want to pass size so that
- We don't break any other banking code
- We can still sanely handle platforms with multiple processes per bank where
pagemap_realloc has a real job. It's not quite that simple there are some
other 32bit changes it would need, but lets not dig a hole to fall down
Tormod Volden [Mon, 7 Jan 2019 21:28:54 +0000 (22:28 +0100)]
bank8k: Run program_vectors() after page map reallocation
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:28:53 +0000 (22:28 +0100)]
bank8k: Redefine page map table to not include LOBANK
The first element corresponds to the first page slot to be mapped.
If LOBANK is not zero, we will have fewer page slots to go through.
Use moving ptr instead of indexing it, to keep all compilers from
messing up.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:28:52 +0000 (22:28 +0100)]
bank8k: Make sure pagemap_free frees something
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:28:51 +0000 (22:28 +0100)]
bank8k: Fix syntax so that it builds again
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Alan Cox [Tue, 8 Jan 2019 16:58:02 +0000 (16:58 +0000)]
Breaks too much. Also messes up any future multi-process per bank arrangement
Revert "syscall_exec16: pagemap_realloc takes top, not size"
This reverts commit
0463011f113a92a0c9b19fb91cbb081c743a2f32.
Tormod Volden [Mon, 7 Jan 2019 21:28:49 +0000 (22:28 +0100)]
syscall_exec16: pagemap_realloc takes top, not size
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:53:06 +0000 (22:53 +0100)]
dragon-nx32: Remove unused (un)map_video calls
Now that we are not sharing vc.c any longer.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:53:05 +0000 (22:53 +0100)]
dragon-mooh: Remaining assembly VT routines for vc
Use own vc.c version calling the assembly routines
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:53:04 +0000 (22:53 +0100)]
dragon-nx32/mooh: Move vc assembly to new vc_asm.s
multihead.s now becomes the same for the two platform.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:53:03 +0000 (22:53 +0100)]
dragon-nx32/mooh: Rename to VC_BASE
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:53:02 +0000 (22:53 +0100)]
dragon-mooh: Squeeze in VC framebuffers at 0x400-0x800
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:53:01 +0000 (22:53 +0100)]
dragon-mooh: Use the (un)map_video helpers in video.s
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:28:48 +0000 (22:28 +0100)]
dragon-mooh: Define TOP_SIZE for swapping
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:28:47 +0000 (22:28 +0100)]
dragon-mooh: Add empty platform copyright
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:28:46 +0000 (22:28 +0100)]
dragon-mooh: Use makeproc
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Alan Cox [Tue, 8 Jan 2019 01:46:47 +0000 (01:46 +0000)]
z80softspi: correct comments
Tormod Volden [Mon, 7 Jan 2019 21:28:45 +0000 (22:28 +0100)]
Standalone/chmem: Correct index for binary magic
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 7 Jan 2019 21:28:44 +0000 (22:28 +0100)]
chmem: Allow values above 32767 also where int is 16-bit
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Alan Cox [Tue, 8 Jan 2019 01:02:04 +0000 (01:02 +0000)]
linc80: support SD cards on PIO port B
5v to 3.3v convertors obviously required!
Alan Cox [Tue, 8 Jan 2019 00:59:51 +0000 (00:59 +0000)]
z80softspi: Software SPI for the Z80
Mainly intended for use with the Z80PIO but ought to work with other hardware
too providing it's got a single register with input and output bits present.
Alan Cox [Tue, 8 Jan 2019 00:59:26 +0000 (00:59 +0000)]
genie-eg64: need_resched needs to be common
Alan Cox [Tue, 8 Jan 2019 00:59:06 +0000 (00:59 +0000)]
devsd: allow asm linkages for SPI helpers
Alan Cox [Tue, 8 Jan 2019 00:58:55 +0000 (00:58 +0000)]
zxmmc: swap was done
Alan Cox [Mon, 7 Jan 2019 11:58:57 +0000 (11:58 +0000)]
sc108: Can also use the loader
Alan Cox [Mon, 7 Jan 2019 01:27:03 +0000 (01:27 +0000)]
linc80: update README
Alan Cox [Mon, 7 Jan 2019 01:26:20 +0000 (01:26 +0000)]
linc80: initial port
Lots of work to do on the IM2 side - serial handlers etc rather than polling
on the timer as an icky hack to get started
Alan Cox [Sun, 6 Jan 2019 21:17:20 +0000 (21:17 +0000)]
rc2014 sio: Fix incorrect calculation causing junk and bad sio setup
Alan Cox [Sat, 5 Jan 2019 22:44:54 +0000 (22:44 +0000)]
filo: tweak it a bit more so we get the right layout
Also load another sector as we will spill into it inevitably. This isn't a big
deal as even in DOS days the loader space was over 60 sectors.
Alan Cox [Sat, 5 Jan 2019 22:41:51 +0000 (22:41 +0000)]
filo: rework some code so we can use it nicely as a boot block loader
Alan Cox [Sat, 5 Jan 2019 22:25:33 +0000 (22:25 +0000)]
sc114: Fix sc108 handling on execve
Alan Cox [Sat, 5 Jan 2019 21:59:15 +0000 (21:59 +0000)]
filo: move constants to right file
Alan Cox [Sat, 5 Jan 2019 21:56:23 +0000 (21:56 +0000)]
filo: build at a given address
Alan Cox [Sat, 5 Jan 2019 21:55:58 +0000 (21:55 +0000)]
sc114: make loader use generic filo parts
Alan Cox [Sat, 5 Jan 2019 21:55:28 +0000 (21:55 +0000)]
filo: first pieces of a generic Z80 loader
Alan Cox [Sat, 5 Jan 2019 21:30:09 +0000 (21:30 +0000)]
sc114: more minor tweaks and README fixups
Alan Cox [Sat, 5 Jan 2019 21:20:47 +0000 (21:20 +0000)]
sc108: fix typo
Alan Cox [Sat, 5 Jan 2019 21:19:15 +0000 (21:19 +0000)]
sc114: get it building
Alan Cox [Sat, 5 Jan 2019 21:18:54 +0000 (21:18 +0000)]
scm_monitor: interface glue
Alan Cox [Sat, 5 Jan 2019 21:06:25 +0000 (21:06 +0000)]
sc114: initial draft of an SC114 port using SCM monitor as much as possible
It's also designed to work on the SC108, except we don't have a suitable bank
aware monitor yet.
Alan Cox [Sat, 5 Jan 2019 21:05:24 +0000 (21:05 +0000)]
sc114: first draft of a real boot loader for Fuzix
Alan Cox [Fri, 4 Jan 2019 20:10:54 +0000 (20:10 +0000)]
rc2014: update notes and documentation
Alan Cox [Fri, 4 Jan 2019 19:30:43 +0000 (19:30 +0000)]
tbblue: add UART minor loader fix
Still working over the code and reviewing, not yet ready for testing
(need to sort a different 8K paged platform first to test the core 8K code
better)
Alan Cox [Fri, 4 Jan 2019 13:51:16 +0000 (13:51 +0000)]
mnakfile: update v65 status to dead for now
Alan Cox [Fri, 4 Jan 2019 13:50:47 +0000 (13:50 +0000)]
nascom: notes update
Alan Cox [Thu, 3 Jan 2019 19:13:31 +0000 (19:13 +0000)]
cromemco: trivial optimization
Alan Cox [Thu, 3 Jan 2019 19:12:14 +0000 (19:12 +0000)]
cromemco: first cut at a fast fork() interbank copier
17 clocks/byte but we can't use quite the same technique for the other paths
we need to optimize (notably udata and disk block copies)