Alan Cox [Fri, 17 Jul 2015 22:26:06 +0000 (23:26 +0100)]
inode: note a spot we need to optimise
Alan Cox [Fri, 17 Jul 2015 22:25:34 +0000 (23:25 +0100)]
cpu: lay groundwork for 32bit kernel / 16bit user etc
ugetp/uputp for user pointer sizes
uptr_t holds a userspace pointer
Alan Cox [Fri, 17 Jul 2015 22:25:00 +0000 (23:25 +0100)]
trs80: update README
Brett Gordon [Thu, 16 Jul 2015 13:30:07 +0000 (09:30 -0400)]
mkfs: fast format option
--
089e0103e64ae50a2e051afe11ce
Content-Type: text/plain; charset=UTF-8
a new mkfs w/ fast formating (just i blocks) using getopt()
--
Brett M. Gordon,
beretta42@gmail.com
<div dir="ltr">a new mkfs w/ fast formating (just i blocks) using getopt()<br clear="all"><div><br></div>-- <br><div class="gmail_signature">Brett M. Gordon,<br><a href="mailto:beretta42@gmail.com" target="_blank">beretta42@gmail.com</a></div>
</div>
Brett Gordon [Wed, 15 Jul 2015 04:42:23 +0000 (00:42 -0400)]
platform-coco3: sdc driver bugfixes
--
001a11c3371cba70e5051ae2949a
Content-Type: text/plain; charset=UTF-8
* conditionally compile sdc_init(): It kills a becker port DW server
otherwise.
* remove half-arsed rawmode check.
* slightly nicer kernel message on boot-up.
--
Brett M. Gordon,
beretta42@gmail.com
<div dir="ltr"><div>* conditionally compile sdc_init(): It kills a becker port DW server otherwise.</div><div>* remove half-arsed rawmode check.</div><div>* slightly nicer kernel message on boot-up.</div><div><br></div><br clear="all"><div><br></div>-- <br><div class="gmail_signature">Brett M. Gordon,<br><a href="mailto:beretta42@gmail.com" target="_blank">beretta42@gmail.com</a></div>
</div>
Brett Gordon [Sun, 12 Jul 2015 15:48:45 +0000 (11:48 -0400)]
fuzix path
--
001a113d2c22535242051aaf8a38
Content-Type: text/plain; charset=UTF-8
* preliminary CoCoSDC "enhanced" mode drivers written.
--
Brett M. Gordon,
beretta42@gmail.com
<div dir="ltr">* preliminary CoCoSDC "enhanced" mode drivers written.<br clear="all"><div><br></div>-- <br><div class="gmail_signature">Brett M. Gordon,<br><a href="mailto:beretta42@gmail.com" target="_blank">beretta42@gmail.com</a></div>
</div>
Alan Cox [Tue, 7 Jul 2015 14:42:53 +0000 (15:42 +0100)]
From: Hemstra Bart Jeroen <bart-jeroen@hemstra.com>
trs80: move bootloader to use sdasz80/sdldz80
I converted the boot loader to SDASZ80, the result assembles neatly into 256 bytes. I included the file so if you like you can check it and place it in github. I didn't change the (complex) makefile, I assembled it with:
sdasz80 -o trs80load.s
sdldz80 -i trs80load.rel
makebin -s 256 trs80load.ihx trs80load.bin
dd if=trs80load.bin of=mydisk.jv3 bs=1 seek=8704 conv=notrunc
[added Make bits - Alan]
Brett Gordon [Mon, 6 Jul 2015 12:02:42 +0000 (08:02 -0400)]
platform-coco3: implement IDE
--
047d7b41cc5cd5d441051a33ae0e
Content-Type: text/plain; charset=UTF-8
update to platform-coco3:
* now does IDE via blkdev
* shared dragon-nx32's ide transfer routine(s)
* (re)implements mbr.c for coco's partitioning scheme
--
Brett M. Gordon,
beretta42@gmail.com
<div dir="ltr"><div>update to platform-coco3:</div><div><br></div><div>* now does IDE via blkdev</div><div>* shared dragon-nx32's ide transfer routine(s)</div><div>* (re)implements mbr.c for coco's partitioning scheme</div><div><br></div><div><br></div>-- <br><div class="gmail_signature">Brett M. Gordon,<br><a href="mailto:beretta42@gmail.com" target="_blank">beretta42@gmail.com</a></div>
</div>
Brett Gordon [Sat, 4 Jul 2015 14:52:05 +0000 (10:52 -0400)]
IDE bugfix for dragon-nx32
--
94eb2c094236ea23be051a0dd016
Content-Type: text/plain; charset=UTF-8
Here's a bug fix that make IDE work for 6809 stuff.
* bugfix: removed spaces in operand to test the rawflag in blk_op (lwasm
treats as comments)
* bugfix: moved accesses to blk_ops around to before we switch to
userspace: blk_ops is in kernel space to it goes *poof* when we flip to
user-space
The good news is we can share ide.s directly, as these bug fixes make IDE
work for the coco3 platform too.
I haven't tried this on real hardware, but on fake hardware it works great!
:)
--
Brett M. Gordon,
beretta42@gmail.com
<div dir="ltr"><div>Here's a bug fix that make IDE work for 6809 stuff. </div><div>* bugfix: removed spaces in operand to test the rawflag in blk_op (lwasm treats as comments)</div><div>* bugfix: moved accesses to blk_ops around to before we switch to userspace: blk_ops is in kernel space to it goes *poof* when we flip to user-space</div><div><br></div><div>The good news is we can share ide.s directly, as these bug fixes make IDE work for the coco3 platform too.</div><div><br></div><div>I haven't tried this on real hardware, but on fake hardware it works great! :)</div><div><br></div><div><br></div>-- <br><div class="gmail_signature">Brett M. Gordon,<br><a href="mailto:beretta42@gmail.com" target="_blank">beretta42@gmail.com</a></div>
</div>
Alan Cox [Thu, 2 Jul 2015 17:05:50 +0000 (18:05 +0100)]
trs80: fix missed file discard.c
Brett Gordon [Mon, 29 Jun 2015 12:59:13 +0000 (08:59 -0400)]
fuzix: platform-coco3: better drivewire virtual serial port support
* Support for multiple virtual serial and "9server" windows
* Added 'dw' application for simple/testing Drivewire Command/Control API
--
Brett M. Gordon,
beretta42@gmail.com
EtchedPixels [Tue, 30 Jun 2015 10:51:58 +0000 (11:51 +0100)]
Merge pull request #218 from willsowerbutts/una
zeta-v2: _trap_reboot now disables memory paging entirely
Will Sowerbutts [Mon, 29 Jun 2015 20:53:01 +0000 (21:53 +0100)]
zeta-v2: _trap_reboot now disables memory paging entirely
EtchedPixels [Fri, 26 Jun 2015 21:23:44 +0000 (22:23 +0100)]
Merge pull request #216 from willsowerbutts/tinyfixes
A number of small fixes
Will Sowerbutts [Fri, 26 Jun 2015 19:39:08 +0000 (20:39 +0100)]
zeta-v2: Implement _trap_reboot
Will Sowerbutts [Fri, 26 Jun 2015 19:38:56 +0000 (20:38 +0100)]
p112: Implement _trap_reboot
Will Sowerbutts [Fri, 26 Jun 2015 19:38:44 +0000 (20:38 +0100)]
p112: Add missing includes in discard.c
Will Sowerbutts [Fri, 26 Jun 2015 19:38:28 +0000 (20:38 +0100)]
monitor-z80: Add Jump command
Brett Gordon [Fri, 26 Jun 2015 06:35:27 +0000 (02:35 -0400)]
Fuzix: platform-coco3: no more stashing, progtop = 0xe000
I fixed up the coco3 platform's udata/progtop/banking:
* no more udata/stack stashing on switch-in/switch-out
* Progtop now set at 0xe000, the last 8k mmu bank (could go higher)
--
Brett M. Gordon,
beretta42@gmail.com
Alan Cox [Fri, 26 Jun 2015 11:03:34 +0000 (12:03 +0100)]
From: prateemmandal
[FUZIX] make for z80pack fails with error ?ASlink-Warning-Undefined Global '_trap_reboot' referenced by module 'syscall_other' (#215)
I am trying to follow https://sinrega.org/?p=218 to build a Fuzix OS disk image and it fails with following error:
prateem@prateem-VirtualBox:~$ tail -10 fuzix_build_failure.log
Must remake target `sdldz80'.
cc -g -O2 -Wall -Wno-parentheses -DINDEXLIB -DUNIX -I. -o sdldz80 lk_readnl.o lkaomf51.o lkar.o lkarea.o lkdata.o lkelf.o lkeval.o lkhead.o lklex.o lklib.o lklibr.o lklist.o lkmain.o lkmem.o lknoice.o lkout.o lkrel.o lkrloc.o lkrloc3.o lks19.o lksdcclib.o lksym.o sdld.o lksdcdb.o lkbank.o
Successfully remade target file `sdldz80'.
make[1]: Leaving directory `/home/prateem/work/FUZIX/Kernel/tools/bankld'
Successfully remade target file `tools/bankld/sdldz80'.
Must remake target `fuzix.ihx'.
tools/bankld/sdldz80 -n -k /usr/share/sdcc/lib/z80 -f platform-z80pack/fuzix.lnk
?ASlink-Warning-Undefined Global '_trap_reboot' referenced by module 'syscall_other'
make: *** [fuzix.ihx] Error 2
I was able to fix it by making the following changes (but I am not sure if this is the right fix):
Alan Cox [Fri, 26 Jun 2015 11:03:14 +0000 (12:03 +0100)]
trs80: fix white space
Alan Cox [Tue, 23 Jun 2015 21:31:06 +0000 (22:31 +0100)]
bank94: not yet in use or tested
Initial pieces to support the use of 0x94 as the secondary bank control
register as used by some of the TRS80 expansions.
Alan Cox [Tue, 23 Jun 2015 21:30:48 +0000 (22:30 +0100)]
trs80: split out the banking logic
Brett Gordon [Mon, 22 Jun 2015 13:41:04 +0000 (09:41 -0400)]
p_top must be set in process.c as well for bank16k
I'm not exactly sure, but I think I found a bug in the bank16k
implementation. When ptab_alloc() calls pagemap_alloc(), p_top is not
set yet. pagemap_alloc, then proceeds to assign only one bank to the
new processes. Assuming normal copy forking, should p_top be set to
the parent's p_top *before* calling pagemap_alloc() ?
This fix seems to make my coco3 UDATA mods start working
Brett Gordon [Wed, 17 Jun 2015 13:40:33 +0000 (09:40 -0400)]
platform-coco3: Forgot to commit/add link script for Applications/utils
platform-coco3: Forgot to commit/add link script for Applications/utils.
side note: I'll be more careful about trailing white space /
indentation. Do we prefer tabs or spaces in C code? I'll change my
emacs to give the C code a bit more indentation.
--
Brett M. Gordon,
beretta42@gmail.com
Brett Gordon [Wed, 17 Jun 2015 11:25:08 +0000 (12:25 +0100)]
CoCo3 Fuzix (take two)
UDATA/Banking is still non-ideal.
DriveWire virtual terminal still need help.
Stupid Comments fixed
devlpr removed
Alan Cox [Wed, 17 Jun 2015 11:24:11 +0000 (12:24 +0100)]
coco3: trailing space stripping
Alan Cox [Wed, 17 Jun 2015 11:20:27 +0000 (12:20 +0100)]
coco3: more white space
Alan Cox [Wed, 17 Jun 2015 11:19:56 +0000 (12:19 +0100)]
coco3: trimmed trailing white space in a few spots
Alan Cox [Wed, 17 Jun 2015 11:18:50 +0000 (12:18 +0100)]
coco3: indent devtty again
It was looking a bit odd with a mix of tabs and spaces giving weird alignments
No code changes
Brett Gordon [Wed, 17 Jun 2015 11:07:05 +0000 (12:07 +0100)]
CoCo3 Fuzix (take two)
UDATA/Banking is still non-ideal.
DriveWire virtual terminal still need help.
Stupid Comments fixed
devlpr removed
Alan Cox [Wed, 17 Jun 2015 11:03:19 +0000 (12:03 +0100)]
Add 6809 simple to the list for dev work
Alan Cox [Tue, 16 Jun 2015 12:35:37 +0000 (13:35 +0100)]
micropack: straighten out defines
This is our starting point for a minimal build and for Z80 gets us an 0x6000
(strictly 0x5F00) sized user space which we could push up to 0x6200.
Our target is 0x7D00, however 0x6000 is enough to run many of the apps/tools.
Alan Cox [Mon, 15 Jun 2015 22:13:32 +0000 (23:13 +0100)]
micropack: experimenting with some smallness ideas
Alan Cox [Mon, 15 Jun 2015 22:12:55 +0000 (23:12 +0100)]
tty: cannot putc_wait from an interrupt
Tormod Volden [Mon, 15 Jun 2015 21:21:43 +0000 (23:21 +0200)]
sh: Optimize all files for size on 6809
No workaround exception needed for word.c any longer.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 15 Jun 2015 20:47:46 +0000 (22:47 +0200)]
Remove empty toascii.c, also from build.mk
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 15 Jun 2015 20:47:45 +0000 (22:47 +0200)]
Build: Add missing library files
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Mon, 15 Jun 2015 20:47:44 +0000 (22:47 +0200)]
Build: Fix liberror.txt endianness on dragon-nx32
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Alan Cox [Sun, 14 Jun 2015 23:24:39 +0000 (00:24 +0100)]
analysemap: Fix to with with newer SDCC
EtchedPixels [Sun, 14 Jun 2015 11:18:46 +0000 (12:18 +0100)]
Merge pull request #209 from davidgiven/msp430
Fix 6809 Travis setup; add MSP430
Alan Cox [Sun, 14 Jun 2015 10:14:31 +0000 (11:14 +0100)]
string.h: enable SDCC builtins
Shrinks binaries a little bit more
Alan Cox [Sat, 13 Jun 2015 21:39:34 +0000 (22:39 +0100)]
ssh: kill is no longer expensive with tiny atoi
Alan Cox [Sat, 13 Jun 2015 21:38:49 +0000 (22:38 +0100)]
lib: add a smaller atoi
Alan Cox [Sat, 13 Jun 2015 21:38:20 +0000 (22:38 +0100)]
telinit: fix SDCC build
David Given [Sat, 13 Jun 2015 11:59:56 +0000 (13:59 +0200)]
Grr, brain fade.
David Given [Sat, 13 Jun 2015 11:57:03 +0000 (13:57 +0200)]
More dependency fixing.
David Given [Sat, 13 Jun 2015 11:34:32 +0000 (13:34 +0200)]
Fix i386 issues with lwtools.
David Given [Sat, 13 Jun 2015 11:29:09 +0000 (13:29 +0200)]
Fix spilling orrer.
David Given [Sat, 13 Jun 2015 10:36:13 +0000 (12:36 +0200)]
Merge from upstream.
David Given [Sat, 13 Jun 2015 10:33:42 +0000 (12:33 +0200)]
Slowly getting there...
David Given [Sat, 13 Jun 2015 10:28:39 +0000 (12:28 +0200)]
Change MSP430 install path ('Error changing permissions to 0755 in /usr/local/bin/search'?).
David Given [Sat, 13 Jun 2015 10:20:24 +0000 (12:20 +0200)]
Turns out the MSP430 compiler needs the 32-bit environment.
Tormod Volden [Sat, 13 Jun 2015 09:33:26 +0000 (11:33 +0200)]
Build: Add dragon-nx32 to travis CI script
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Alan Cox [Sat, 13 Jun 2015 09:47:01 +0000 (10:47 +0100)]
zx128: add vtattr_cap
Alan Cox [Sat, 13 Jun 2015 09:46:31 +0000 (10:46 +0100)]
px4plus: vtattr_cap addition
Alan Cox [Sat, 13 Jun 2015 09:46:07 +0000 (10:46 +0100)]
mtx: add vtattr_cap and note awkward special case to do
Alan Cox [Sat, 13 Jun 2015 09:45:53 +0000 (10:45 +0100)]
bihx: fix warning
Tormod Volden [Sat, 13 Jun 2015 09:06:56 +0000 (11:06 +0200)]
Build: Add reboot to dragon-nx32
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Sat, 13 Jun 2015 09:06:55 +0000 (11:06 +0200)]
Build: Add init to the root directory
Because of how the parsing is done, it needs a double slash.
Also add the missing inittab.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Sat, 13 Jun 2015 09:06:54 +0000 (11:06 +0200)]
Build: Build disk not floppy image for dragon-nx32
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Sat, 13 Jun 2015 09:06:53 +0000 (11:06 +0200)]
Build: gcc: Don't compile into separate sections
Now that we link libraries correctly, we shouldn't
need the garbage collection of unused objects.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Sat, 13 Jun 2015 09:06:52 +0000 (11:06 +0200)]
Build: Link 6809 libc and libgcc as libraries
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
David Given [Fri, 12 Jun 2015 20:28:33 +0000 (22:28 +0200)]
Tweak msp430 compiler setup.
Tormod Volden [Fri, 12 Jun 2015 17:10:46 +0000 (19:10 +0200)]
Build: Use lwtools ld wrapper for 6809
This is a patched version. We use our own copy until
we get our changes upstream.
Also add a simple objcopy wrapper.
Add a catch-all text segment pattern to the linker script
because the build uses gcc -ffunction-sections.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Tormod Volden [Fri, 12 Jun 2015 17:10:45 +0000 (19:10 +0200)]
Build: Support reversed byte order filesystems
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
David Given [Fri, 12 Jun 2015 19:34:03 +0000 (21:34 +0200)]
Fix typo.
David Given [Fri, 12 Jun 2015 19:22:23 +0000 (21:22 +0200)]
Attempt to build MSP430 userland.
Alan Cox [Thu, 11 Jun 2015 21:14:18 +0000 (22:14 +0100)]
platforms: more vtattr fixups
None of these actually add an attr support yet
Alan Cox [Thu, 11 Jun 2015 19:51:59 +0000 (20:51 +0100)]
dragon: fix build
EtchedPixels [Wed, 10 Jun 2015 22:50:32 +0000 (23:50 +0100)]
Merge pull request #207 from davidgiven/travis
Use precompiled sdcc for Travis builds
Alan Cox [Wed, 10 Jun 2015 22:50:15 +0000 (23:50 +0100)]
trs80: add the needed blank attribute variable
David Given [Wed, 10 Jun 2015 22:33:18 +0000 (00:33 +0200)]
Fix typo.
Alan Cox [Wed, 10 Jun 2015 22:24:05 +0000 (23:24 +0100)]
dragon-nx32: add attribute support
Useful as an example hopefully. Everything supported except flash, which isn't
in the hardware in a practical manner.
Alan Cox [Wed, 10 Jun 2015 22:23:52 +0000 (23:23 +0100)]
decbdragon: fix warning
Alan Cox [Wed, 10 Jun 2015 22:22:35 +0000 (23:22 +0100)]
vt: add bits for attribute setting/reporting
Done with a single escape code to make the parsing tiny. It's then up to the
platform what if anything to offer, but if nothing is supported then the
platform needs the vtattr_cap byte defining.
David Given [Wed, 10 Jun 2015 22:08:22 +0000 (00:08 +0200)]
Use the precompiled sdcc rather than building it ourselves.
Alan Cox [Tue, 9 Jun 2015 21:51:15 +0000 (22:51 +0100)]
trs80: fix incorrect tty struct reference
EtchedPixels [Tue, 9 Jun 2015 21:50:58 +0000 (22:50 +0100)]
Merge pull request #206 from davidgiven/m6809
Some 6809 build system fixes
David Given [Tue, 9 Jun 2015 21:24:33 +0000 (23:24 +0200)]
Made a few minor fixes where the build system was assuming things about
targetgcc platforms that it shouldn't. Add some workaround scripts to
deal with the 6809 platform not actually using binutils as a backend to
gcc.
David Given [Tue, 9 Jun 2015 21:10:35 +0000 (23:10 +0200)]
Switch to using a precompiled cc65.
Alan Cox [Mon, 8 Jun 2015 21:40:43 +0000 (22:40 +0100)]
cpp: build with -Os
Tormod Volden [Mon, 8 Jun 2015 21:20:50 +0000 (23:20 +0200)]
dragon-nx32: Add new build system support
The m6809-unknown-ld wrapper must be patched.
libc-functions.localsrcs doesn't work yet though.
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
Alan Cox [Mon, 8 Jun 2015 21:30:51 +0000 (22:30 +0100)]
cpp: CPP from bcc
48K rather than 32K target. Some thought might be needed for a 32K cpp
however!
Alan Cox [Mon, 8 Jun 2015 21:22:55 +0000 (22:22 +0100)]
ld09: first cut
Needs the output side replacing with a simple binary writer for FUZIX
format headers.
This one actually should fit nicely on a Dragon32 !
Alan Cox [Mon, 8 Jun 2015 21:03:48 +0000 (22:03 +0100)]
as09: 6809 assembler for native asm work
Right now this won't *quite* fit on the Dragon32, but I think it would with
a tiny bit of tidying up and if 6809 is fixed to not suck in stdio always
(or has a stdio free link option as Z80 has for now)
Alan Cox [Mon, 8 Jun 2015 19:33:23 +0000 (20:33 +0100)]
init: add telinit code
Alan Cox [Mon, 8 Jun 2015 19:32:56 +0000 (20:32 +0100)]
gets: correct behaviour of gets_s
Alan Cox [Mon, 8 Jun 2015 14:02:54 +0000 (15:02 +0100)]
From: Brett Gordon <beretta42@gmail.com>
reboot: Fix missing newline
Alan Cox [Sun, 7 Jun 2015 23:35:24 +0000 (00:35 +0100)]
trap_reboot: fix the first batch of platforms to export it
Alan Cox [Sun, 7 Jun 2015 22:45:18 +0000 (23:45 +0100)]
zx128: export trap_reboot
Alan Cox [Sun, 7 Jun 2015 22:45:05 +0000 (23:45 +0100)]
z80: first cut at lowlevel-z80-banked.S for new pre-empt
Alan Cox [Sun, 7 Jun 2015 22:30:13 +0000 (23:30 +0100)]
Merge branch 'master' of https://github.com/EtchedPixels/FUZIX
EtchedPixels [Sun, 7 Jun 2015 22:28:34 +0000 (23:28 +0100)]
Merge pull request #202 from davidgiven/travis
Add Travis continuous integration support
Tormod Volden [Sun, 7 Jun 2015 09:47:26 +0000 (11:47 +0200)]
dragon-nx32: Set BASIC cold start flag
Signed-off-by: Tormod Volden <debian.tormod@gmail.com>
David Given [Sun, 7 Jun 2015 21:19:14 +0000 (23:19 +0200)]
Fixup after merge.
David Given [Sun, 7 Jun 2015 21:07:46 +0000 (23:07 +0200)]
Merge from master.
David Given [Sun, 7 Jun 2015 20:53:15 +0000 (22:53 +0200)]
Update to work with sdcc snapshot. opt-code-size is apparently broken, so use
opt-code-speed instead.
David Given [Sun, 7 Jun 2015 20:22:46 +0000 (22:22 +0200)]
Don't download sdcc directly from SourceForge; instead go via a curated github
repository.
Alan Cox [Sun, 7 Jun 2015 09:24:44 +0000 (10:24 +0100)]
filesystem: add /var/run for utmp