HALT is terminal - stepCPU returns at once when the Halt Flag is up, so a
halted machine does not execute, service devices, or take an interrupt -
and that has to stay true, because every test ends with a halt and "halted"
is how a program says it has finished. The consequence was that SplitBit
had no way to wait at all. Every wait was a spin, and a spin is bus
traffic: 11.5% of Type over a 14K file on a disk of ten thousand cycles,
after read-ahead had already hidden three quarters of the latency.
WAIT is 0xFE, one byte, no operands, sitting under HALT where the
instruction that almost stops the machine belongs. Three decisions in it:
- A line already standing means there is nothing to wait for, so WAIT does
nothing. That is what makes test-then-wait race-free.
- Any line ends the wait, masked or not, so a program can sleep on a device
it has no handler for and read its status afterwards. Masking says who
answers a request, not whether it happened.
- A line that wakes the CPU without being dispatched is taken down by the
WAIT. Left standing it would be found by the next WAIT, which would
return at once - the program would spin exactly as before while looking
as though it slept.
Waiting is NOT a Status bit, and that is the trap avoided rather than a
gap: Status rides into the interrupt frame and comes back out, so a machine
interrupted mid-wait would return from its handler still waiting, and wait
again for what it had already been given. An internal field instead.
Idle cycles are counted apart from bus cycles and the halt line says so
when there are any, which is what makes the difference observable at all -
with the line-clearing removed the total moves by ONE cycle, 20,100 against
20,099, and only the idle half changes, halving to 9,976. A test on
totals could never have seen it. Tests/terminal.sh asks that question,
being the file for things a recorded output cannot see, and fails with the
clear removed while "both reads finished" still passes.
Three collisions, all found by building it:
- 0xFE was the assembler's "not an instruction" sentinel. getOpcode now
answers a negative NOT_AN_OPCODE, which is outside the range of every
possible answer instead of inside the unused part of it.
- 0xFE was also what faultTest and faultResumeTest executed to provoke a
fault. They now use 0xFD and say why, because they did not fail when it
became an instruction - they HUNG, having started sleeping instead.
- Keys.asm has had a label called "wait" for a year, and mnemonics are
matched uppercased. What that reported was "Branch without label" at the
BRQ thirty lines away. The assembler now refuses a label that is already
an instruction, at the label, by name; every instruction added takes a
word out of the space of label names, so this will happen again.
A parent is an entry index PLUS ONE in two bytes, so entry 65535 has no
parent number: adding one wraps to zero, and zero is the root. Eight
entries to a block, so 8192 directory blocks reaches it and SplitDisk
formatted that happily.
It does not fail by refusing, which is why it was worth chasing rather than
reasoning about. Reproduced on a disk built for it: mkdir /deep/child, with
/deep at entry 65535, printed 'Made "/deep/child" as entry 0' and put child
in the ROOT. Listing /deep then showed nothing, because the search is for a
parent of 65536 and the entry carries zero - so the same mkdir succeeded
again, and again, and five entries called /child piled up in the root.
Duplicate names in one directory are the one thing rename refuses outright,
on the grounds that a search answers with whichever it meets first and the
rest can never be reached; this manufactured them one per attempt.
8191 blocks is the most, giving 65528 entries. Refused when formatting and
again when reading, in both implementations, because a disk claiming more
was made by something that never checked. On the machine only the high byte
of the count has to be looked at: anything from 0x20 up is too many.
Three checks, all of which fail with their guard removed. The machine's
disk claims the size rather than having it, so the test image is 64 blocks
that lie rather than sixteen megabytes that do not - mounting is refused at
the geometry, which is read out of block 0.
Saving something that already exists writes a temporary, deletes the
original and gives the temporary its name, so that nothing is lost if the
writing fails. The temporary was told apart from a real file by being
called sbfs.part or sbfs.out - and those are legal names. Starting a save
deleted whatever answered to one as stale scratch, so saving anything at
all in a directory destroyed your own file of that name there, silently.
Flag bit 0x04 now says it. The property is not in the contents - the same
bytes become the finished file the instant the rename lands - so it belongs
in the entry, which is the thing the commit changes. sbfsCreateTempAt is
the door temporaries come in by, the commit writes the flags flat along
with the name, and cleanup wipes what it finds only if the entry says it is
ours. Anything else stops the save instead.
The bit is also the recovery. Both listings show an unfinished write rather
than sizing it, because the size in the entry is the room that was asked
for and not what was written: "<unfinished>" from dir, and a line from
SplitDisk saying the blocks are held and a rename brings the data back.
That was the gap in what the last commit documented - the data survived a
crash and nothing would show you where it was.
Four new agreement checks, three of which fail with the guards removed. The
fourth needed rebuilding first: both tests started on one disk, and the
first save ate the sbfs.part that was the second test's SOURCE, so the copy
failed for want of a file, never opened a stream, and passed while
reporting on nothing. A disk each. The fifth check forges the wreckage by
setting the flag on a finished file, since nothing here can crash a save
half way through.
No version bump: a committed file never carries the bit, so a disk this
writes is byte for byte the disk the old code wrote, which the whole-image
comparisons confirm. Only the wreckage differs, and older code reads that
as an ordinary file - which is what it did before.
The disk's status has always had a bit meaning "still going", and the header
beside it has always said to honour it. Nothing did, because nothing could: the
host finished the transfer inside the instruction that asked for it, so the bit
could never be seen up and asking about it was asking about something that
cannot happen.
--disk-cycles gives it a latency. The command is still checked at once, because
a refusal is not work - a block that is not there fails before any head moves -
but the transfer is remembered and done when the machine has run that far. Until
then the buffer holds the block BEFORE this one.
That last part is the point. A program that does not wait gets the wrong bytes
rather than an error, which is the failure the bit exists to prevent and the one
that would never have shown up. With a latency of two thousand, CosmOS could not
even mount: sbfsMount reads block zero and looks straight at the buffer.
deviceTick is the general shape rather than a disk feature. Called once per
instruction with the machine's clock, it lets anything whose moment has come
finish - which is what a display that refreshes, or a port that waits on the
host, would want in exactly the same way.
The filesystem watches the bit now, in one small routine reached with RCAL. That
is not decoration: what it hands back is the settled status in A, and CALL puts A
back the way it found it, so an ordinary call cannot carry the one thing this
exists to carry. Two bytes of Stack rather than ten, in a routine that runs on
every block the machine ever touches - the first place in the system where the
new call is the right one rather than merely a cheaper one.
The manifest takes a @N after a disk, the way it already takes :ro, so a test can
ask for a slow one. cosmosSlowDisk lists a directory at two thousand cycles a
block and gets the same listing as everything else, which is the whole assertion:
a filesystem that did not wait would print nonsense rather than fail.
Zero is the default and every other test runs at it. What waiting costs, on a
directory heavy run: 229k cycles at zero, 275k at five hundred, 415k at two
thousand, 1.16M at ten thousand.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E2JrLzFvuFX9fgi1LDRjrW
A blit cost ten cycles, which were the five port writes that set it up. The
quarter of a kilobyte that moved cost nothing, and no hardware moves a quarter
of a kilobyte for nothing.
BANKS ARE SEPARATE MEMORIES, AND THAT IS WHAT SETS THE RATE. A move between two
of them can overlap its read and its write - fetch the next byte while the last
one is stored - so it settles at a byte a cycle. A move within one bank cannot,
and costs two. A fill has nothing to read and costs one whatever the banks are.
The odd cycle on each is the pipeline filling.
That is not a modelling choice so much as a reading of the structure the machine
already has: a Program to Data blit is inherently twice the rate of a Data to
Data one, and it is legible why.
Measured: 256 bytes is 297 cycles across banks and 518 within one, both
including the instructions that ask for it.
WHAT IT TAUGHT, which was not what I expected. Charging for movement costs the
native assembler 0.4 per cent and costs directory work 13.4. The assembler reads
a block and then thinks about it for a long time, so the move is amortised into
nothing; the filesystem reads a block in order to look at it and does nothing
else in between.
So the case for a blitter that runs alongside the CPU is weaker than it sounds.
Concurrency pays when there is other work to do during the transfer, and the
place that spends its time moving memory is exactly the place with nothing else
to do - it blits a block precisely so that it can read it. What that workload
wants is a FASTER controller, not a concurrent one: a wider data path halves the
wait, and the machine is waiting either way.
Video is the case that would still want concurrency, since a frame can be moved
while the next one is worked out. That is an argument about software nobody has
written yet, and it is now an argument with numbers on the other side of it.
The byte at a time port is charged too, for the byte it moves beyond reaching
the port. Nothing polls CTRL_STATUS, so the transfer stalls whoever asked for
it, which is the conservative reading and the one the software already assumes.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E2JrLzFvuFX9fgi1LDRjrW
cycleCount used to tick once per instruction, so RSTA cost what SETD cost and a
CALL moving ten bytes of Stack cost what a branch cost. No machine anybody could
build works that way, and the emulator's job is to be the thing the hardware is
designed against.
Every touch of memory now goes through one of four accessors that charge for it:
fetching an opcode, fetching the bytes after it, reading or writing Data Memory,
and reaching a device port. One access, one cycle, nothing overlapped. The
accessors exist so the cost is counted where the access happens rather than in a
table of per instruction costs kept somewhere else - a table like that is a
second copy of what the code does, and the two drift.
The run loop spends a budget of cycles instead of running a count of
instructions, so the emulated rate means something: an instruction costs what it
touches, and a batch ends when the cycles are gone.
What the numbers say now: RSTA 1 and SETD 4, being one byte and four. LDA 3,
DPUA 2. CALL and RET together 24, RCAL and RRET together 8, because the first
pair moves twenty bytes of Stack and the second moves four. The average SplitBit
instruction costs 3.72 of these, measured over the native assembler assembling a
program.
And the measurement that prompted all of this: converting the filesystem's
hottest leaf routine to RCAL is 3.1 per cent cheaper on a directory heavy
workload. The old model said 0.0, which is what a model that cannot see memory
traffic must say about a change that is nothing else.
Three tests moved. settle() strips the cycle count from recorded output, so
nothing should have churned - but it was anchored to the start of a line and
replCalculator's last output has no newline on it, which leaves the halt message
mid line where the pattern never reached. Not anchored any more.
The two Life programs are bounded by a cycle count because they never end, and
that number was rescaled from 3,000,000 to 11,200,000 - the same amount of work
at 3.72 cycles to the instruction. Nothing about either program changed. No limit
reproduces the old output exactly, because the cut now lands elsewhere in a
frame, so they are recorded again rather than tuned to match.
Whether hardware overlaps a fetch with the end of the previous instruction is
left open on purpose. This is the conservative model; pipelining is a decision to
make while drawing the hardware, not one to inherit from an emulator.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E2JrLzFvuFX9fgi1LDRjrW
-S writes every label and the address it was given, in address order.
Nothing else knows that. A program on the disk is bytes; the monitor can
disassemble it but has no idea what any of it is called. So counting which
addresses a program calls says a great deal and names nothing - the answer
arrives as a column of numbers and somebody works out by hand which routine
each one is inside.
It was deferred when the native assembler was planned, as a listing and symbol
dump nobody needed yet. Finding out where the assembler spends its time is what
needed it: the top six call targets were addresses until this existed and are
numStep, numCompare, tokGet, srcNext, numAddByte and clsSameName with it.
Sorted by address rather than by name, because the question asked of it is
always "what is at this address".
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E2JrLzFvuFX9fgi1LDRjrW
The six settled back on the twenty fourth, built now.
RCAL and RRET are a call that puts nothing back. CALL restores A, B and Data
Pointers 0 through 2, which costs ten bytes of Stack and is why a subroutine
here can only hand anything back through Q, DP3 or memory. RCAL costs two and
restores nothing, which is what a short leaf routine wants and is unsafe in
exactly the way the name says.
They are a pair because the frames are different sizes: returning from one
through the other walks the Stack to somewhere that was never a return address.
That was the user's correction to the original proposal, which had a raw call
and no raw return.
DPUA and DPDA offset a Data Pointer by A; DPUW and DPDW by A and B together,
most significant first. DPUP and DPDN take a byte written into the program, so
moving a pointer by something just worked out meant storing it and loading it
back. Down as well as up on symmetry grounds, which was also the user's call -
the argument against it came from counting uses in a corpus written under the
constraint.
The opcodes sit where they belong: 0x16 and 0x1E immediately below CALL and RET,
and 0x4E through 0x51 at the end of the Data Pointer family. All six fit shapes
that already existed, so instructiontable.py needed only set membership and both
machine side copies of the table regenerated from it unchanged.
Checked at every level it exists at: the emulator runs them, the host assembler
encodes them, the monitor disassembles all six with the right lengths, and the
assembler that runs on the machine builds a program using them byte for byte
identically to the host - and that program runs.
The recorded test measures what the two calls COST as well as what they put
back, because an RCAL that quietly did what CALL does would still return to the
right place. It does not survive that: returned through RRET, it hangs.
docs.sh can read a two word number now. The count of instructions taking a Data
Pointer went past twenty, and the pattern only allowed one word, so the check
would have reported that the manual had stopped saying it rather than that the
number was wrong.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E2JrLzFvuFX9fgi1LDRjrW
SBFS version two adds directories out of space each entry had already set
aside: two of the four reserved bytes become a parent, and one of the seven
spare flag bits says an entry is a directory. The entry is still thirty two
bytes, so it still divides two hundred and fifty six and still never straddles
a block, and nothing in the block layer knows anything happened.
A directory is an entry with no blocks. That is what keeps the flat array of
entries the whole allocation map, which is the property the format is built
on: with files laid down contiguously, every block is inside some entry's
range or it is not, and an entry with no range is in nobody's way. There is
still no allocation table to consult and none to keep right.
THE PARENT IS AN INDEX PLUS ONE, so zero means the root. A version one disk has
zeroes in those bytes, and "in the root" is exactly where every file on a flat
disk is - so a version one image is already a valid version two image, with
nothing to convert and no tool to convert it with.
A disk is at the lowest version that describes what is on it. format makes a
version one disk and mkdir is what raises it, so everything built here stays
readable by a reader that has never heard of a directory right up until it
really does have one. That is what lets this land before the machine knows
anything: the whole existing suite passes untouched.
The tool gains mkdir and rmdir, and list, put, get and delete take paths. list
also now reports entries used against entries available, because a disk has two
ceilings and the entry one is the one nobody notices until it bites.
rmdir refuses a directory with anything in it, and that is not politeness:
parents are entry indices, a freed index gets handed out again, and the
children of a removed directory would reappear inside whatever took its place.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E2JrLzFvuFX9fgi1LDRjrW
It always wrote .bin, whatever it had built. So assembling a loadable
program without -o produced Say.bin containing SBEX - a boot image name on a
file the machine cannot boot, in a repository whose whole convention is that
a .bin is started from and a .sbx is loaded.
Successfully wrote SplitBit boot image to "hello.bin".
Successfully wrote SplitBit loadable program to "Say.sbx".
programIsLoadable() already existed and is already what decides which
writer runs; the name now asks it too. Nothing in the build depended on the
old behaviour, because everything that assembles anything passes -o.
THE ASSEMBLER THAT RUNS ON SPLITBIT ALREADY DID IT THIS WAY. Two assemblers
naming their output differently from the same source is exactly the kind of
difference that wastes an afternoon, and the newer one was right.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E2JrLzFvuFX9fgi1LDRjrW
"Binary" was doing three jobs. It meant an SPBT file that the machine starts
from; it meant whatever the assembler happened to produce, which is now
either that or a loadable program; and it meant a compiled host tool. A word
that means three things means none of them, and the first of the three has a
name already - this project has been calling them boot images for a while
and the manuals had not caught up.
Where it means an SPBT file -> boot image
Where it means either output -> output
Where it means a host executable -> left alone
Where it means base two -> left alone
The user facing messages move with it:
Error: No boot image specified.
Usage: ./SplitBit [OPTIONS] <boot image>
Error: This is not a SplitBit boot image.
Error: This boot image is in format version 2, and this emulator reads 1.
Successfully wrote SplitBit boot image to "hello.bin".
The assembler's own help was the interesting case. Its -o writes either
format, so "the binary" there was never right - it is "the output" now, and
the message that names the format is the one that says which it wrote.
No recorded output contained the word, so nothing needed re-blessing.
Checked before starting rather than after.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E2JrLzFvuFX9fgi1LDRjrW
The quotes are gone by the time a token is classified, so checkIfKeyword's
test of token[0] == '#' matched the STRING "#Program" sitting in a program's
Data Segment. It was read as the directive: the segment silently changed in
the middle of the data, the string's nine bytes were charged to the Program
cursor instead of the Data one, and every label defined after it came out
nine bytes wrong - in a file that still had a valid header, a plausible
length, and nothing to say about any of it. The only symptom was a program
that jumped into the middle of an instruction.
This is the FOURTH of the family. A string spelling a mnemonic assembled as
that instruction; a string beginning with a zero was rejected as a malformed
literal; a string in the Program Segment was discarded in silence. The
instruction check and the literal check both carry a "not a STRING" guard
already. This one did not, so it has one now, and it lives inside
checkIfKeyword rather than at the call site so it cannot be left off again.
Nothing had ever triggered it, because nothing had ever needed a directive's
name as data. An assembler written FOR this machine necessarily does: it has
to compare tokens against "#Program" and "#Data". It was found by building
one and watching it fault on its second instruction.
Test stringKeyword puts every directive name in a program's data and prints
a label defined after them. Verified that it bites: without the fix the
assembler refuses the file outright.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E2JrLzFvuFX9fgi1LDRjrW
A token is classified after its quotes have been stripped, so a string
literal reading "ADD" looked exactly like the ADD instruction and was
assembled as one. It failed with "attempting to assemble outside the
Program Segment", a message about a mistake nobody had made.
The literal and label checks were already guarded against strings and the
instruction check was not. Mnemonics match without regard to case, so
"or" and "and" were caught by this too, and those are ordinary enough
words to want in a message.
Third of its family, after a string beginning with '0' being read as a
malformed number and a string in the Program Segment being silently
discarded. All three have the same root.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The loadable program format's header has been in the working tree since
the loader was written and has never been in the repository: the gitignore
rules matched Source/Assembler as a directory, so it was silently
untracked. Anything cloning this repository could not build it, because
secondPass.c includes this file.
This is the failure the previous commit describes, having already happened
once without being noticed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Overflowing/Underflowing A or B with increment or decrement instructions now sets the Carry Flag.
An increment or decrement of A or B that doesn't result in an overflow/underflow now clears the Carry Flag.
Q is no longer saved and restored in subroutine calls, allowing it to be used to return a one byte result.
Documentation updated with more detail.
Added DPUP and DPDN, which take a one byte operand specifying how far up or down to offset the Data Pointer.
Three Fibonacci generators using the print.asm library.
- 8 bit values printing in decimal representation.
- 16 bit values printing in hexadecimal representation.
- 32 bit values printing in hexadecimal representation.
Anachronaut
Anachronaut