Nick Downing [Sun, 21 Apr 2019 05:31:42 +0000 (15:31 +1000)]
Changes to get platform-v8080 running again
Alan Cox [Tue, 16 Apr 2019 22:58:11 +0000 (23:58 +0100)]
zx+3: fix a missing clear and a typo
Alan Cox [Tue, 16 Apr 2019 22:57:39 +0000 (23:57 +0100)]
libc: more optmization on 8080/5
We can't turn this on for apps as it makes the compiler core dump
Alan Cox [Tue, 16 Apr 2019 22:57:09 +0000 (23:57 +0100)]
8085: add hooks for magic Z80 M1 emulation
Alan Cox [Tue, 16 Apr 2019 22:56:07 +0000 (23:56 +0100)]
docs: add a bit more stuff
Alan Cox [Fri, 12 Apr 2019 21:47:08 +0000 (22:47 +0100)]
vt: mention cursor_disable
Alan Cox [Fri, 12 Apr 2019 21:46:58 +0000 (22:46 +0100)]
malloc: remove dead define references
Alan Cox [Fri, 12 Apr 2019 21:46:45 +0000 (22:46 +0100)]
flat: remove dead define references
Alan Cox [Fri, 12 Apr 2019 21:46:32 +0000 (22:46 +0100)]
docs: carry on building some kind of documentation tree
Alan Cox [Fri, 12 Apr 2019 13:04:55 +0000 (14:04 +0100)]
docs: start adding the memory manager docs
Alan Cox [Fri, 12 Apr 2019 11:03:52 +0000 (12:03 +0100)]
INSTALL: mention Ubuntu as well as Debian having broken yacc defaults
Alan Cox [Thu, 11 Apr 2019 23:35:35 +0000 (00:35 +0100)]
docs: Use sphinx and start to bash the documents into a coherent form
Yes they need major surgery on the formatting.
Alan Cox [Thu, 11 Apr 2019 23:34:42 +0000 (00:34 +0100)]
trs80m1: fix typo
Alan Cox [Thu, 11 Apr 2019 23:34:24 +0000 (00:34 +0100)]
scorpion: Fix typo
Alan Cox [Thu, 11 Apr 2019 23:33:56 +0000 (00:33 +0100)]
msx1: note where we need to do work on subslots
Alan Cox [Thu, 11 Apr 2019 23:33:00 +0000 (00:33 +0100)]
fuzix: add the descr file for ack to the tree for now
Alan Cox [Sun, 7 Apr 2019 19:27:55 +0000 (20:27 +0100)]
msx1: ROM probing for cards
We can now autodetect the IDE controller by ROM (if you have the same ROM anyway
other identifiers may well be needed).
Alan Cox [Sun, 7 Apr 2019 19:27:08 +0000 (20:27 +0100)]
msx1: add another debug marker
Alan Cox [Sun, 7 Apr 2019 12:42:17 +0000 (13:42 +0100)]
msx1: update README
Things kind of work now - we have shell and logins and can run stuff, but it
probably still has plenty of bugs left to find
Alan Cox [Sun, 7 Apr 2019 12:38:16 +0000 (13:38 +0100)]
sunrise: redo the bounce buffer logic
We can't just use tmpbuf because there might not be a free buffer and we'd
then get into a mess recursing through our I/O subystem. Instead grab a bit
of spare high memory for the moment.
Once the buffer reclaim logic is added we will turn discard and friends into
buffers but keep the last 512 bytes before FFFF as a bounce buffer that drivers
can share.
(FFFF itself is slot switching so magic and reserved)
Alan Cox [Sun, 7 Apr 2019 12:23:21 +0000 (13:23 +0100)]
msx1: move platform_discard into the right place
It might otherwise discard itself once we add the recovery logic
Alan Cox [Sun, 7 Apr 2019 11:50:20 +0000 (12:50 +0100)]
msx1: use the first disk buffer as the bounce for the vector copy
Alan Cox [Sun, 7 Apr 2019 11:48:46 +0000 (12:48 +0100)]
msx1: custom user copy routines
We know the kernel side will always be mapped so we can do much simpler user
copiers which is important because the bank switch is so incredibly expensive.
The actual code is from SBCv2 so we should probably unify this into a standard
'always mapped' lib helper
Alan Cox [Sun, 7 Apr 2019 11:47:44 +0000 (12:47 +0100)]
msx1: do the vector copy earlier
Alan Cox [Sun, 7 Apr 2019 11:47:10 +0000 (12:47 +0100)]
msx.h: add copy_vectors prototype
Alan Cox [Sun, 7 Apr 2019 10:41:36 +0000 (11:41 +0100)]
sunrise: remove some noisy debug
Alan Cox [Sun, 7 Apr 2019 10:41:14 +0000 (11:41 +0100)]
sunrise: fix incorrect LBA setup
Alan Cox [Sun, 7 Apr 2019 10:40:33 +0000 (11:40 +0100)]
msx1: put the top of process memory back right for swap
We are using split udata so shouldn't account for udata there
Alan Cox [Sun, 7 Apr 2019 10:39:28 +0000 (11:39 +0100)]
msx1: move the constant area into the 16K RAM
This has a cost in space terms but means that anything we might copy to or from
user space is now >= 0xC000, which means it will always be mapped so we can just
ldir blocks when doing uput/uget.
As the bank switching on MSX is so chronically slow this is important
Alan Cox [Sun, 7 Apr 2019 10:39:07 +0000 (11:39 +0100)]
cartman: teach it that bigger is ok
Alan Cox [Sun, 7 Apr 2019 10:38:37 +0000 (11:38 +0100)]
msx1: allow for more load space
Once we've used the ROM services we move our sp up high and save just the stuff
we need at 0xFCxx
Alan Cox [Sat, 6 Apr 2019 23:54:10 +0000 (00:54 +0100)]
sunrise: remove asm debug
Alan Cox [Sat, 6 Apr 2019 23:53:42 +0000 (00:53 +0100)]
msx1: don't need to save AF around the map_kernel_di
Alan Cox [Sat, 6 Apr 2019 23:52:50 +0000 (00:52 +0100)]
msx1: fix register trashing bugs in the map logic
Alan Cox [Sat, 6 Apr 2019 23:52:33 +0000 (00:52 +0100)]
sunrise: take out some debug in the C code
Alan Cox [Sat, 6 Apr 2019 23:52:14 +0000 (00:52 +0100)]
msx1: add the syscall vector
This helps a lot 8)
Alan Cox [Sat, 6 Apr 2019 21:40:01 +0000 (22:40 +0100)]
config: tweaks and fix swap range
Once we get that far we will need to swap the udata not just process image
Alan Cox [Sat, 6 Apr 2019 21:37:54 +0000 (22:37 +0100)]
visualize: add some more area types
Alan Cox [Sat, 6 Apr 2019 21:37:43 +0000 (22:37 +0100)]
cartman: pack the revised kernel format for MSX
Alan Cox [Sat, 6 Apr 2019 21:35:34 +0000 (22:35 +0100)]
msx1: the big rework
- Pull common and discard etc from ROM into C000-FFFF RAM
- Fix missing interrupt disables in mapping functions
- Various mapping bugfixes
- Turn on sunrise (hardcoded location for now for testing)
- Fix the memory layout
With this lot done and some tool updates we can get the point init makes
syscalls although it doesn't end well shortly after that
Alan Cox [Sat, 6 Apr 2019 21:33:57 +0000 (22:33 +0100)]
sunrise: assorted fixes
With these fixed we can mount the root fs and load stuff from the Sunrise
IDE controller. There are a ton of races with switching and interrupts to
fix after this
Alan Cox [Sat, 6 Apr 2019 21:33:12 +0000 (22:33 +0100)]
msx1: we don't have a stash on one process model remove old bits
Alan Cox [Sun, 31 Mar 2019 21:16:09 +0000 (22:16 +0100)]
sc108: need to support ACIA
Alan Cox [Sun, 31 Mar 2019 21:15:47 +0000 (22:15 +0100)]
mtx: update README
Still work to do before we are up and running on the real box and CFII+
Alan Cox [Sun, 31 Mar 2019 21:15:14 +0000 (22:15 +0100)]
8080/8085: optimizations that are useful go to -O4
Alan Cox [Sun, 31 Mar 2019 21:14:44 +0000 (22:14 +0100)]
Merge branch 'master' of https://github.com/EtchedPixels/FUZIX
EtchedPixels [Sun, 31 Mar 2019 21:13:58 +0000 (22:13 +0100)]
Merge pull request #697 from beretta42/tar
Tar take 2
Brett Gordon [Sat, 30 Mar 2019 20:37:28 +0000 (16:37 -0400)]
tar: embolster documentation!
Brett Gordon [Sat, 30 Mar 2019 20:28:47 +0000 (16:28 -0400)]
tar: debug stat problems
* Don't report chown failure. (this would bombard non-priv users tty)
* We're storing too much of fuzix's stat.mode in tar header! (see
header.type)
* got to specify some sort of file mode with O_CREAT !
Brett Gordon [Sat, 30 Mar 2019 19:27:00 +0000 (15:27 -0400)]
tar: support block factor for tape
Tormod Volden [Fri, 15 Mar 2019 21:47:19 +0000 (22:47 +0100)]
syscall_fs2: Add parentheses to fix gcc warning
[Alan's note: This is correct. The condition checks are really confusing and
probably we should have an in_group type test that includes the egid.]
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Fri, 15 Mar 2019 21:47:21 +0000 (22:47 +0100)]
dragon-mooh: Fix signedness mismatch in vc functions
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Fri, 15 Mar 2019 21:47:20 +0000 (22:47 +0100)]
filesys: Fix gcc warning
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Fri, 15 Mar 2019 21:47:18 +0000 (22:47 +0100)]
dragon-nx32: Update custom /etc/rc with mount change
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Alan Cox [Sat, 23 Mar 2019 21:05:52 +0000 (21:05 +0000)]
README: update on 8080/5 front
Alan Cox [Sat, 23 Mar 2019 20:55:18 +0000 (20:55 +0000)]
atarist: Start filling in the drivers
Floppy and ACSI to begin with, along with the IDE adapters often added.
Alan Cox [Sat, 23 Mar 2019 20:51:44 +0000 (20:51 +0000)]
mbr: note we will need to make this return a code at some point
There are systems we might find multiple partition types on and we need to not
install both sets wrongly
Alan Cox [Sat, 23 Mar 2019 20:51:29 +0000 (20:51 +0000)]
devide: fix missing REG in name
Alan Cox [Sat, 23 Mar 2019 20:50:59 +0000 (20:50 +0000)]
ahdi: first cut at Atari partition formats
Alan Cox [Mon, 18 Mar 2019 23:08:35 +0000 (23:08 +0000)]
mtx: Memotech updates
Start putting in place the stuff for the CFII and the Propeller based 80 column
card. Also fill in some of the graphics and other bits. Fix the maximum memory
banks.
There are still some really ugly things to deal with. Do we expose the propeller
directly or via the VT layer. How do we align that with the graphics ioctls.
Stil need to think about UDGs
Alan Cox [Mon, 18 Mar 2019 23:02:44 +0000 (23:02 +0000)]
graphics: add another format
Memotech graphic blocks are 2 x 4 for 256 symbols
Alan Cox [Sun, 17 Mar 2019 14:35:21 +0000 (14:35 +0000)]
sbcv2: update README
Alan Cox [Sun, 17 Mar 2019 00:39:09 +0000 (00:39 +0000)]
atarist: get the fdc probing and add blkdev for the block devices
No actual hard disk drivers yet, and it wil also need a GEMDOS partition table
parser adding.
Alan Cox [Sun, 17 Mar 2019 00:15:53 +0000 (00:15 +0000)]
atarist: first cut at writing a floppy driver for the Atari 1772 + DMA
Heavily cribbed from the GPL EmuTOS codebase
Alan Cox [Sun, 17 Mar 2019 00:15:19 +0000 (00:15 +0000)]
atarist: 200Hz ticker
Alan Cox [Sat, 16 Mar 2019 22:13:31 +0000 (22:13 +0000)]
atarist: get the timers and console working
No device drivers so you can't yet mount a file system
Alan Cox [Sat, 16 Mar 2019 16:26:19 +0000 (16:26 +0000)]
atari-st: this gets us to the point the mono console kind of works-ish
Sufficient to print useful messages. We still need to debug this further and
sort out non mono displays.
Alan Cox [Sat, 16 Mar 2019 16:24:40 +0000 (16:24 +0000)]
68000: add Atari ST bootblock checksumming tool
Alan Cox [Sat, 16 Mar 2019 00:22:26 +0000 (00:22 +0000)]
rc2014: fix up the banking code and update the bootloader
At this point the banked kernel appears to work. Needs a bit more testing and
to be tested on the actual board.
Alan Cox [Sat, 16 Mar 2019 00:21:51 +0000 (00:21 +0000)]
devide: make the generic devide correct for banked and other cases
Only map kernel back if we map something else
Alan Cox [Fri, 15 Mar 2019 22:03:37 +0000 (22:03 +0000)]
rc2014: lay groundwork to make rc2014 run a banked kernel.
RC2014 has so many toys they do not all fit any more and there is still sound,
PIO and TMS9918A support to add!
This breaks everything. The loader needs a rewrite and the code debugging so
if you are doing a build go back before this commit
Alan Cox [Fri, 15 Mar 2019 22:01:47 +0000 (22:01 +0000)]
ds1302: unify all the code and banked versions too
Alan Cox [Fri, 15 Mar 2019 22:01:30 +0000 (22:01 +0000)]
devfd_hw: add a banked version
Alan Cox [Fri, 15 Mar 2019 20:02:35 +0000 (20:02 +0000)]
trs80m1: fix banking to cover CODE4
Alan Cox [Fri, 15 Mar 2019 19:21:40 +0000 (19:21 +0000)]
more type alignments
None of these really matter as uint8_t == uint_fast8_t for Z80 but it's good
to get it and the signed/unsigned clashes tidied up.
Alan Cox [Fri, 15 Mar 2019 19:21:19 +0000 (19:21 +0000)]
dev: sync more types
Alan Cox [Thu, 14 Mar 2019 14:34:10 +0000 (14:34 +0000)]
rc2014-tiny: align types
Alan Cox [Thu, 14 Mar 2019 14:09:50 +0000 (14:09 +0000)]
rc2014-sbc64: align types and fix a Makefile ordering bug
We need boot to be built to generate loader.inc, not to depend on boot.s
Alan Cox [Thu, 14 Mar 2019 14:00:35 +0000 (14:00 +0000)]
rc2014: align types
Alan Cox [Thu, 14 Mar 2019 13:38:36 +0000 (13:38 +0000)]
p112: align types
Alan Cox [Thu, 14 Mar 2019 13:38:22 +0000 (13:38 +0000)]
devscsi: align types
Alan Cox [Wed, 13 Mar 2019 23:42:07 +0000 (23:42 +0000)]
dev: more device type alignment
Alan Cox [Wed, 13 Mar 2019 23:41:48 +0000 (23:41 +0000)]
n8vem-mark4: align types
Alan Cox [Wed, 13 Mar 2019 23:05:49 +0000 (23:05 +0000)]
multicomp09: align types
Alan Cox [Wed, 13 Mar 2019 23:01:35 +0000 (23:01 +0000)]
mtx: align with type changes and update TODO list
Time to move this onto real hardware with CF card support etc
Alan Cox [Wed, 13 Mar 2019 22:50:22 +0000 (22:50 +0000)]
msx2: align types
Alan Cox [Wed, 13 Mar 2019 22:34:50 +0000 (22:34 +0000)]
tty: fix some of the kputchar sign breakages
Alan Cox [Wed, 13 Mar 2019 22:32:36 +0000 (22:32 +0000)]
dragon-nx32: initial type resync
Alan Cox [Wed, 13 Mar 2019 22:27:02 +0000 (22:27 +0000)]
vt: align with type changes
We ought to fix up the internal type usages but a lot of these lead into asm
code so some review is called for.
Alan Cox [Wed, 13 Mar 2019 22:26:03 +0000 (22:26 +0000)]
coco3: initial alignment of type changes
More to do eventually
Alan Cox [Wed, 13 Mar 2019 22:25:50 +0000 (22:25 +0000)]
coco2cart: align with type changes
Alan Cox [Wed, 13 Mar 2019 22:25:37 +0000 (22:25 +0000)]
devsd:fix breakage from type conversions
Alan Cox [Wed, 13 Mar 2019 22:20:10 +0000 (22:20 +0000)]
amstradnc: align with uint_fast8_t and tty unsigned changes
Alan Cox [Wed, 13 Mar 2019 20:52:46 +0000 (20:52 +0000)]
v68: fix the devtty type changes that matter
Alan Cox [Wed, 13 Mar 2019 20:52:28 +0000 (20:52 +0000)]
v85: align with the _fast8_t changes
Alan Cox [Wed, 13 Mar 2019 20:52:15 +0000 (20:52 +0000)]
v8080: align with the _fast8_t changes
Alan Cox [Wed, 13 Mar 2019 20:51:32 +0000 (20:51 +0000)]
dev: bring some of the core devices we need in line with uint_fast8_t
We can defer a lot because only 8080/8085 at this point have it differing
from uint8_t
Alan Cox [Wed, 13 Mar 2019 20:50:34 +0000 (20:50 +0000)]
kernel: core changes to get uint_fast8_t into the drivr layer
This whole set of work saves us a mindboggling 2.2K on 8080/8085 !
Alan Cox [Wed, 13 Mar 2019 20:48:55 +0000 (20:48 +0000)]
usermem: Fix accidentally 16bitisms
In some 32bit cases we use the generic valaddr(). Thus we shouldn't cast things
to 16bit or hardcode 0xFFFF. Fix these.
Alan Cox [Wed, 13 Mar 2019 20:47:57 +0000 (20:47 +0000)]
syscall_exec32: using unsigned for names