Alan Cox [Sun, 28 Dec 2014 23:01:58 +0000 (23:01 +0000)]
makeversion: report Fuzix not UZI+
Alan Cox [Sun, 28 Dec 2014 22:43:50 +0000 (22:43 +0000)]
sysconf/pathconf: Add pathconf, update sysconf
Add load average interface
Add pathconf interfaces, first cut
Alan Cox [Sun, 28 Dec 2014 22:28:52 +0000 (22:28 +0000)]
utsname: add utsname functionality
Alan Cox [Sun, 28 Dec 2014 22:28:16 +0000 (22:28 +0000)]
utsname: tidy header
Alan Cox [Sat, 27 Dec 2014 21:45:47 +0000 (21:45 +0000)]
init: Remove debug stuff
Alan Cox [Sat, 27 Dec 2014 21:41:18 +0000 (21:41 +0000)]
kdata: remove dead time struct
Alan Cox [Sat, 27 Dec 2014 21:40:03 +0000 (21:40 +0000)]
execve: Fix a load of problems in both crt0.s and the kernel
The good news is this fixes a lot of things. The bad news is that all the
old binaries were broken. Oh well.. if we are going to have binary breakage
we might as well do it twice in one day.
Data, bss and break are now all set up correctly (I hope)
Alan Cox [Sat, 27 Dec 2014 21:38:56 +0000 (21:38 +0000)]
timer: Move structures about
All the clock bits are now in the cpu struct together. This also lets us
flip the 6809 clock/mini union the right way up for big endian
Alan Cox [Sat, 27 Dec 2014 21:04:11 +0000 (21:04 +0000)]
size: Add a size tool to report binary sizes
EtchedPixels [Sat, 27 Dec 2014 19:56:32 +0000 (19:56 +0000)]
Merge pull request #31 from willsowerbutts/tickunion
Kernel: Merge 16-bit and 32-bit tick counters into one structure
Will Sowerbutts [Sat, 27 Dec 2014 16:32:15 +0000 (16:32 +0000)]
Kernel: Merge 16-bit and 32-bit tick counters into one structure
EtchedPixels [Sat, 27 Dec 2014 16:31:15 +0000 (16:31 +0000)]
Merge pull request #29 from geijoenr/master
Add keyboard support for msx
Alan Cox [Sat, 27 Dec 2014 16:29:33 +0000 (16:29 +0000)]
README.md: correct
Alan Cox [Sat, 27 Dec 2014 15:48:46 +0000 (15:48 +0000)]
getpass: don't suck in stdio
Alan Cox [Sat, 27 Dec 2014 15:39:16 +0000 (15:39 +0000)]
timer: remove more dead references
Alan Cox [Sat, 27 Dec 2014 15:38:50 +0000 (15:38 +0000)]
timer: remove dead timer reference from other ports
Alan Cox [Sat, 27 Dec 2014 15:27:00 +0000 (15:27 +0000)]
README.md: new ports from Will
EtchedPixels [Sat, 27 Dec 2014 15:23:54 +0000 (15:23 +0000)]
Merge pull request #30 from willsowerbutts/toupstream
Fuzix kernel Z180 support, p112 and n8vem-mark4 platforms, other minor fixes
Alan Cox [Sat, 27 Dec 2014 15:12:17 +0000 (15:12 +0000)]
init: remove use of the malloc package
We can sbrk up our own environment strings and that cuts init down to about
9K. Way more than it should be, but most of what it drags in now needs
fixing in the libraries instead.
Will Sowerbutts [Sat, 27 Dec 2014 14:39:26 +0000 (14:39 +0000)]
Kernel: Bug fix -- bankfixed pagemap_realloc() check was too strict
Will Sowerbutts [Sat, 27 Dec 2014 14:24:55 +0000 (14:24 +0000)]
Kernel: Useful debug routines for Z180 (commented out)
Will Sowerbutts [Sat, 27 Dec 2014 14:39:55 +0000 (14:39 +0000)]
Merge remote-tracking branch 'alan/master' into toupstream
Alan Cox [Sat, 27 Dec 2014 13:59:50 +0000 (13:59 +0000)]
crt0: set the load page to 1
Probably fcc binman should do the needed magic for such things
Alan Cox [Sat, 27 Dec 2014 13:59:03 +0000 (13:59 +0000)]
execve: handle null files
When given an empty file execve was doing a bmap and trying to load block
-1 from the disk. Catch the empty file case earlier and error it.
geijoenr [Sat, 27 Dec 2014 13:21:05 +0000 (14:21 +0100)]
Add keyboard support for msx
Alan Cox [Sat, 27 Dec 2014 13:14:32 +0000 (13:14 +0000)]
ubee: check the right flag on dec b for the crtc loop.
Alan Cox [Sat, 27 Dec 2014 13:13:30 +0000 (13:13 +0000)]
trs80: Check the PV flag not C on 8bit dec
Alan Cox [Sat, 27 Dec 2014 12:51:41 +0000 (12:51 +0000)]
execve: Fix ABI change on unsized binaries
We need to adjust by the load page *after* we have checked for zero.
Will Sowerbutts [Sat, 27 Dec 2014 11:34:19 +0000 (11:34 +0000)]
Kernel: Remove obsolete lines from kernel.def on p112 and n8vem-mark4 platforms
Will Sowerbutts [Sat, 27 Dec 2014 11:26:25 +0000 (11:26 +0000)]
Kernel: Bug fix -- on system call, touch kernel_flag only after switching to the kernel memory bank
Will Sowerbutts [Sat, 27 Dec 2014 11:22:22 +0000 (11:22 +0000)]
Kernel: Tidy up p112 and n8vem-mark4 targets, remove redundant code and merge memory mapping functions into common Z180 code.
Will Sowerbutts [Sat, 27 Dec 2014 02:38:45 +0000 (02:38 +0000)]
Kernel: Add N8VEM Mark 4 and DX-Designs P112 platforms
Will Sowerbutts [Sat, 27 Dec 2014 01:52:51 +0000 (01:52 +0000)]
Kernel: Increment both 16-bit and 32-bit tick counters on every timer interrupt
Will Sowerbutts [Sat, 27 Dec 2014 01:52:26 +0000 (01:52 +0000)]
Kernel: Whitespace fix for start.c
Will Sowerbutts [Sat, 27 Dec 2014 01:42:53 +0000 (01:42 +0000)]
Kernel: lowlevel-z80: Remove unused map_process symbol
Will Sowerbutts [Sat, 27 Dec 2014 01:42:25 +0000 (01:42 +0000)]
Kernel: Allow boot device number to be greater than 9
Will Sowerbutts [Sat, 27 Dec 2014 01:37:45 +0000 (01:37 +0000)]
Kernel: Improve simple timer code. Remove timer.h include where it is not required.
Will Sowerbutts [Sat, 27 Dec 2014 01:26:07 +0000 (01:26 +0000)]
cpm-loader: Fix minor typo in comment
Will Sowerbutts [Sat, 27 Dec 2014 01:12:09 +0000 (01:12 +0000)]
cpm-loader: Tools to make a CP/M application which boots the Fuzix kernel
Will Sowerbutts [Sat, 27 Dec 2014 01:02:49 +0000 (01:02 +0000)]
Kernel support for Z180 CPUs
Alan Cox [Sat, 27 Dec 2014 00:45:09 +0000 (00:45 +0000)]
Oops: revert Will's bits I didn't mean to merge
Alan Cox [Sat, 27 Dec 2014 00:42:05 +0000 (00:42 +0000)]
Merge Will's fixes
Merge branch 'master' of https://github.com/EtchedPixels/FUZIX
Alan Cox [Sat, 27 Dec 2014 00:40:22 +0000 (00:40 +0000)]
syscall-z88dk: initial fix ups
It now looks plausible but isn't yet tested
Alan Cox [Sat, 27 Dec 2014 00:39:56 +0000 (00:39 +0000)]
syscall: 32bytes per dirent needed
Alan Cox [Sat, 27 Dec 2014 00:39:35 +0000 (00:39 +0000)]
syscall_name: Fix wrong entries
Ones found while building the Z88DK test C lib
Alan Cox [Sat, 27 Dec 2014 00:39:09 +0000 (00:39 +0000)]
uname: pass the load average bits down
EtchedPixels [Sat, 27 Dec 2014 00:37:14 +0000 (00:37 +0000)]
Merge pull request #27 from willsowerbutts/master
Library sleep() -- update for time() API change.
Will Sowerbutts [Sat, 27 Dec 2014 00:36:08 +0000 (00:36 +0000)]
Various tweaks to utilities to build cleanly on my system
Will Sowerbutts [Sat, 27 Dec 2014 00:05:02 +0000 (00:05 +0000)]
Library sleep() -- update for time() API change.
Alan Cox [Fri, 26 Dec 2014 22:41:04 +0000 (22:41 +0000)]
Correct spelling of LICENCE. This is a UK based project so there
Alan Cox [Fri, 26 Dec 2014 21:40:55 +0000 (21:40 +0000)]
z88dk: Add first pieces to bootstrap Z88DK for Fuzix
Alan Cox [Fri, 26 Dec 2014 20:42:43 +0000 (20:42 +0000)]
syscall uname: Add the uptime data
While are breaking the prototype ABI lets just stuff the uptime data in
the sysinfo block and be done with it.
Alan Cox [Fri, 26 Dec 2014 20:31:06 +0000 (20:31 +0000)]
syscalls: extra argument for _stime()
Alan Cox [Fri, 26 Dec 2014 20:30:00 +0000 (20:30 +0000)]
time: correct bug, add stime time type
This finishes the basic provision for multiple time sources
Alan Cox [Fri, 26 Dec 2014 20:26:24 +0000 (20:26 +0000)]
execve: change binary format, add a byte for bank base and align fields
Alan Cox [Fri, 26 Dec 2014 20:25:42 +0000 (20:25 +0000)]
crt0: Add a byte for the base of the binary
Next step in the ABI breaking changes
Alan Cox [Fri, 26 Dec 2014 20:25:21 +0000 (20:25 +0000)]
chmem: Update for new format binaries
Alan Cox [Fri, 26 Dec 2014 20:16:01 +0000 (20:16 +0000)]
time: resynch library with API change for time()
Better to do this before we have too many dependancies than to delay it
until we can't change it.
Alan Cox [Fri, 26 Dec 2014 20:11:09 +0000 (20:11 +0000)]
syscall_name: add argument counts to the tables
We will need these to automate creation of some types of call stubs
Alan Cox [Fri, 26 Dec 2014 20:10:36 +0000 (20:10 +0000)]
syscall_proc: change _time to allow multiple kinds of time
Add since boot monotomic time.
Yes this is an ABI break....
Alan Cox [Fri, 26 Dec 2014 20:10:22 +0000 (20:10 +0000)]
syscall: correct arg descriptions in uname
Alan Cox [Tue, 23 Dec 2014 22:49:02 +0000 (22:49 +0000)]
socz80: start reworking MMU code for 16K banking model
Bank switching functions, and switchin/switchout looked at so far. The fork
copying code remains to be done. Here it definitely makes sense to use the
MMU page 17 logic, but we now need to do it in 16K chunks.
Alan Cox [Tue, 23 Dec 2014 22:46:57 +0000 (22:46 +0000)]
socz80: config.h updates
Alan Cox [Tue, 23 Dec 2014 22:46:37 +0000 (22:46 +0000)]
socz80: ramtop has changed type
Start bringing socz80 back into line
Alan Cox [Tue, 23 Dec 2014 21:06:04 +0000 (21:06 +0000)]
TODO: rtc done(ish)
Alan Cox [Mon, 22 Dec 2014 16:54:20 +0000 (16:54 +0000)]
README: update
Alan Cox [Mon, 22 Dec 2014 15:31:35 +0000 (15:31 +0000)]
PORTING: some updates
Alan Cox [Mon, 22 Dec 2014 15:11:08 +0000 (15:11 +0000)]
z80pack-lite: add RTC support
Alan Cox [Mon, 22 Dec 2014 14:54:59 +0000 (14:54 +0000)]
z80pack: turn on RTC support
Alan Cox [Mon, 22 Dec 2014 14:54:42 +0000 (14:54 +0000)]
z80pack: fix vt.c build failure
Alan Cox [Mon, 22 Dec 2014 14:54:22 +0000 (14:54 +0000)]
z80pack: add rtc helper
Alan Cox [Mon, 22 Dec 2014 14:30:41 +0000 (14:30 +0000)]
trs80: ensure the rtc digit wrap doesn't catch us out
Alan Cox [Mon, 22 Dec 2014 14:30:28 +0000 (14:30 +0000)]
nc100: add RTC support
Alan Cox [Mon, 22 Dec 2014 13:04:09 +0000 (13:04 +0000)]
trs80: CRTC reg 0 was not being set
Backported from Microbee port
Alan Cox [Mon, 22 Dec 2014 13:03:54 +0000 (13:03 +0000)]
trs80: turn on the RTC support
Alan Cox [Mon, 22 Dec 2014 12:42:14 +0000 (12:42 +0000)]
ubee: 64 ticks/second and turn on the RTC
We'll need the RTC in fact as 64 doesn't divide by 10
Alan Cox [Mon, 22 Dec 2014 12:39:17 +0000 (12:39 +0000)]
timer: add initial support for RTC seconds locking
We don't try and use the RTC as our actual clock. Many RTCs are not Y2K safe
or have their own interesting sets of problems. We would also have to keep
doing Unix type v RTC time conversion maths.
Instead we treat the RTC seconds as a free running counter and update our
second tick as we see the RTC one shift. This also gives us up to 59 seconds
of 'overrun' which should do for the worst case disk I/O problems.
Alan Cox [Mon, 22 Dec 2014 12:38:30 +0000 (12:38 +0000)]
process: note problem to look at wrt ubee and 64 ticks/second
Alan Cox [Mon, 22 Dec 2014 12:38:12 +0000 (12:38 +0000)]
ubee: clean up some remaining TRS80isms
Alan Cox [Mon, 22 Dec 2014 11:51:50 +0000 (11:51 +0000)]
ubee: enable CRTC loading
Load the CRTC up for 80x25 mode using the table given in the 256TC manuals
Alan Cox [Mon, 22 Dec 2014 11:51:15 +0000 (11:51 +0000)]
trs80: Set a 2ms seek time on the hard disk
Alan Cox [Sun, 21 Dec 2014 22:07:57 +0000 (22:07 +0000)]
ubee,devhd: step rates and the like
Alan Cox [Sun, 21 Dec 2014 21:58:01 +0000 (21:58 +0000)]
ubee: Further floppy driver bashing
512 bytes/sector, double sided
Also set the step to 6ms
Alan Cox [Sun, 21 Dec 2014 21:17:47 +0000 (21:17 +0000)]
ubee: More work on the disk controllers
Set the WD1002 up correctly for and work in 512 bytes/sector mode. This should
bring the devhd side roughly where we need it. This doesn't deal with the
precomp values (can we trust the firmware?) or disk step rates (can we read
the firmware set values somewhere ?)
Alan Cox [Sun, 21 Dec 2014 20:51:29 +0000 (20:51 +0000)]
ubee,devhd: Bring more into line with what is needed
Minor 4-7 are the floppy devices on the WD1002-5 and for those we must set the
side bit in port 0x48
Implement the 0x58 device select logic
Check the device is present on open
Alan Cox [Sun, 21 Dec 2014 20:50:14 +0000 (20:50 +0000)]
ubee,devfd: add device select support for port 0x58
It's probably not present but it costs us nothing much to support
Alan Cox [Sun, 21 Dec 2014 20:16:49 +0000 (20:16 +0000)]
trs80: WD1010 driver FIXME is obsolete
Alan Cox [Sun, 21 Dec 2014 17:55:15 +0000 (17:55 +0000)]
ubee: initial bits cloned from the TRS80 to begin a Microbee 128/256/512 port
Alan Cox [Sun, 21 Dec 2014 17:53:05 +0000 (17:53 +0000)]
README: update for TRS80
Alan Cox [Sun, 21 Dec 2014 16:53:55 +0000 (16:53 +0000)]
trs80: clean up monitor bits
Alan Cox [Sun, 21 Dec 2014 15:32:33 +0000 (15:32 +0000)]
trs80: enable floppy disk write support
Alan Cox [Sat, 20 Dec 2014 00:28:19 +0000 (00:28 +0000)]
trs80, hd: Make the hard disk driver work and add swap
With this we can now swap on the TRS80 platform and things seem to work
acceptably. Disk size and other properties are hardcoded in ways that will
need fixing. Partitioning support also needs to be added.
Alan Cox [Sat, 20 Dec 2014 00:26:58 +0000 (00:26 +0000)]
trs80: serial debug
This makes the serial debug work in the emulator. It's not sufficient for
doing serial debug on a real TRS80
Alan Cox [Fri, 19 Dec 2014 01:30:26 +0000 (01:30 +0000)]
z80fixedbank: Always use the swapstack
If we don't then when we call map_kernel after copying the udata back we
will scribble somewhere on the stack of the process. We could potentially
move the switch back to the point we've fixed up sp, but as we have a swap
stack we might as well use it.
If we get banked boxes without swap then it may be worth looking at the other
path for a lib/z80fixedbanknswap.s
Alan Cox [Fri, 19 Dec 2014 01:13:17 +0000 (01:13 +0000)]
trs80: start tidying up the hd support
Add swap hooks etc
Alan Cox [Fri, 19 Dec 2014 01:12:22 +0000 (01:12 +0000)]
bankfixed: check the right define
Alan Cox [Thu, 18 Dec 2014 23:17:04 +0000 (23:17 +0000)]
process: Correct handling of out of processes case
When the error handling was updated this path was broken. With this fixed
running out of memory/swap now correctly fails with ENOMEM
Alan Cox [Thu, 18 Dec 2014 22:34:52 +0000 (22:34 +0000)]
zx128: We can't yet use the shared banker for this, but prepare anyway
Alan Cox [Thu, 18 Dec 2014 21:34:25 +0000 (21:34 +0000)]
trs80: Add serial, keyboard and interrupt handling
This gives a console you can type into and which is at least good enough
to set the root fs. It follows the standard TRS80 layout but with some extra
ctrl and ctrl-shift hooks for useful keys Tandy forgot to include ({} etc)
Alan Cox [Thu, 18 Dec 2014 21:34:02 +0000 (21:34 +0000)]
trs80: 80x24 not 80x25