Well, well. My original plan was to ship the first step of Shuusou Gyoku
OpenGL support on the next day after this delivery. But unfortunately, the
complications just kept piling up, to a point where the required solutions
definitely blow the current budget for that goal. I'm currently sitting on
over 70 commits that would take at least 5 pushes to deliver as a meaningful
release, and all of that is just rearchitecting work, preparing the
game for a not too Windows-specific OpenGL backend in the first place. I
haven't even written a single line of OpenGL yet… 🥲
This shifts the intended Big Release Month™ to June after all. Now I know
that the next round of Shuusou Gyoku features should better start with the
SC-88Pro recordings, which are much more likely to get done within their
current budget. At least I've already completed the configuration versioning
system required for that goal, which leaves only the actual audio part.
So, TH04 position independence. Thanks to a bit of funding for stage
dialogue RE, non-ASCII translations will soon become viable, which finally
presents a reason to push TH04 to 100% position independence after
📝 TH05 had been there for almost 3 years. I
haven't heard back from Touhou Patch Center about how much they want to be
involved in funding this goal, if at all, but maybe other backers are
interested as well.
And sure, it would be entirely possible to implement non-ASCII translations
in a way that retains the layout of the original binaries and can be easily
compared at a binary level, in case we consider translations to be a
critical piece of infrastructure. This wouldn't even just be an exercise in
needless perfectionism, and we only have to look to Shuusou Gyoku to realize
why: Players expected
that my builds were compatible with existing SpoilerAL SSG files, which
was something I hadn't even considered the need for. I mean, the game is
open-source 📝 and I made it easy to build.
You can just fork the code, implement all the practice features you want in
a much more efficient way, and I'd probably even merge your code into my
builds then?
But I get it – recompiling the game yields just yet another build that can't
be easily compared to the original release. A cheat table is much more
trustworthy in giving players the confidence that they're still practicing
the same original game. And given the current priorities of my backers,
it'll still take a while for me to implement proof by replay validation,
which will ultimately free every part of the community from depending on the
original builds of both Seihou and PC-98 Touhou.
However, such an implementation within the original binary layout would
significantly drive up the budget of non-ASCII translations, and I sure
don't want to constantly maintain this layout during development. So, let's
chase TH04 position independence like it's 2020, and quickly cover a larger
amount of PI-relevant structures and functions at a shallow level. The only
parts I decompiled for now contain calculations whose intent can't be
clearly communicated in ASM. Hitbox visualizations or other more in-depth
research would have to wait until I get to the proper decompilation of these
features.
But even this shallow work left us with a large amount of TH04-exclusive
code that had its worst parts RE'd and could be decompiled fairly quickly.
If you want to see big TH04 finalization% gains, general TH04 progress would
be a very good investment.
The first push went to the often-mentioned stage-specific custom entities
that share a single statically allocated buffer. Back in 2020, I
📝 wrongly claimed that these were a TH05 innovation,
but the system actually originated in TH04. Both games use a 26-byte
structure, but TH04 only allocates a 32-element array rather than TH05's
64-element one. The conclusions from back then still apply, but I also kept
wondering why these games used a static array for these entities to begin
with. You know what they call an area of memory that you can cleanly
repurpose for things? That's right, a heap!
And absolutely no one would mind one additional heap allocation at the start
of a stage, next to the ones for all the sprites and portraits.
However, we are still running in Real Mode with segmented memory. Accessing
anything outside a common data segment involves modifying segment registers,
which has a nonzero CPU cycle cost, and Turbo C++ 4.0J is terrible at
optimizing away the respective instructions. Does this matter? Probably not,
but you don't take "risks" like these if you're in a permanent
micro-optimization mindset…
In TH04, this system is used for:
Kurumi's symmetric bullet spawn rays, fired from her hands towards the left
and right edges of the playfield. These are rather infamous for being the
last thing you see before
📝 the Divide Error crash that can happen in ZUN's original build.
Capped to 6 entities.
The 4 📝 bits used in Marisa's Stage 4 boss
fight. Coincidentally also related to the rare Divide Error
crash in that fight.
Stage 4 Reimu's spinning orbs. Note how the game uses two different sets
of sprites just to have two different outline colors. This was probably
better than messing with the palette, which can easily cause unintended
effects if you only have 16 colors to work with. Heck, I have an entire blog post tag just to highlight
these cases. Capped to the full 32 entities.
The chasing cross bullets, seen in Phase 14 of the same Stage 6 Yuuka
fight. Featuring some smart sprite work, making use of point symmetry to
achieve a fluid animation in just 4 frames. This is
good-code in sprite form. Capped to 31 entities, because the 32nd custom entity during this fight is defined to be…
The single purple pulsating and shrinking safety circle, seen in Phase 4 of
the same fight. The most interesting aspect here is actually still related
to the cross bullets, whose spawn function is wrongly limited to 32 entities
and could theoretically overwrite this circle. This
is strictly landmine territory though:
Yuuka never uses these bullets and the safety circle
simultaneously
She never spawns more than 24 cross bullets
All cross bullets are fast enough to have left the screen by the
time Yuuka restarts the corresponding subpattern
The cross bullets spawn at Yuuka's center position, and assign its
Q12.4 coordinates to structure fields that the safety circle interprets
as raw pixels. The game does try to render the circle afterward, but
since Yuuka's static position during this phase is nowhere near a valid
pixel coordinate, it is immediately clipped.
The flashing lines seen in Phase 5 of the Gengetsu fight,
telegraphing the slightly random bullet columns.
These structures only took 1 push to reverse-engineer rather than the 2 I
needed for their TH05 counterparts because they are much simpler in this
game. The "structure" for Gengetsu's lines literally uses just a single X
position, with the remaining 24 bytes being basically padding. The only
minor bug I found on this shallow level concerns Marisa's bits, which are
clipped at the right and bottom edges of the playfield 16 pixels earlier
than you would expect:
The remaining push went to a bunch of smaller structures and functions:
The structure for the up to 2 "thick" (a.k.a. "Master Spark") lasers. Much
saner than the
📝 madness of TH05's laser system while being
equally customizable in width and duration.
The structure for the various monochrome 16×16 shapes in the background of
the Stage 6 Yuuka fight, drawn on top of the checkerboard.
The rendering code for the three falling stars in the background of Stage 5.
The effect here is entirely palette-related: After blitting the stage tiles,
the 📝 1bpp star image is ORed
into only the 4th VRAM plane, which is equivalent to setting the
highest bit in the palette color index of every pixel within the star-shaped
region. This of course raises the question of how the stage would look like
if it was fully illuminated:
Most code that modifies a stage's tile map, and directly specifies tiles via
their top-left offset in VRAM.
Thanks to code alignment reasons, this forced a much longer detour into the
.STD format loader. Nothing all too noteworthy there since we're still
missing the enemy script and spawn structures before we can call .STD
"reverse-engineered", but maybe still helpful if you're looking for an
overview of the format. Also features a buffer overflow landmine if a .STD
file happens to contain more than 32 enemy scripts… you know, the usual
stuff.
To top off the second push, we've got the vertically scrolling checkerboard
background during the Stage 6 Yuuka fight, made up of 32×32 squares. This
one deserves a special highlight just because of its needless complexity.
You'd think that even a performant implementation would be pretty simple:
Set the GRCG to TDW mode
Set the GRCG tile to one of the two square colors
Start with Y as the current scroll offset, and X
as some indicator of which color is currently shown at the start of each row
of squares
Iterate over all lines of the playfield, filling in all pixels that
should be displayed in the current color, skipping over the other ones
Count down Y for each line drawn
If Y reaches 0, reset it to 32 and flip X
At the bottom of the playfield, change the GRCG tile to the other color,
and repeat with the initial value of X flipped
The most important aspect of this algorithm is how it reduces GRCG state
changes to a minimum, avoiding the costly port I/O that we've identified
time and time again as one of the main bottlenecks in TH01. With just 2
state variables and 3 loops, the resulting code isn't that complex either. A
naive implementation that just drew the squares from top to bottom in a
single pass would barely be simpler, but much slower: By changing the GRCG
tile on every color, such an implementation would burn a low 5-digit number
of CPU cycles per frame for the 12×11.5-square checkerboard used in the
game.
And indeed, ZUN retained all important aspects of this algorithm… but still
implemented it all in ASM, with a ridiculous layer of x86 segment arithmetic
on top? Which blows up the complexity to 4 state
variables, 5 nested loops, and a bunch of constants in unusual units. I'm
not sure what this code is supposed to optimize for, especially with that
rather questionable register allocation that nevertheless leaves one of the
general-purpose registers unused. Fortunately,
the function was still decompilable without too many code generation hacks,
and retains the 5 nested loops in all their goto-connected
glory. If you want to add a checkerboard to your next PC-98
demo, just stick to the algorithm I gave above.
(Using a single XOR for flipping the starting X offset between 32 and 64
pixels is pretty nice though, I have to give him that.)
This makes for a good occasion to talk about the third and final GRCG mode,
completing the series I started with my previous coverage of the
📝 RMW and
📝 TCR modes. The TDW (Tile Data Write) mode
is the simplest of the three and just writes the 8×1 GRCG tile into VRAM
as-is, without applying any alpha bitmask. This makes it perfect for
clearing rectangular areas of pixels – or even all of VRAM by doing a single
memset():
// Set up the GRCG in TDW mode.
outportb(0x7C, 0x80);
// Fill the tile register with color #7 (0111 in binary).
outportb(0x7E, 0xFF); // Plane 0: (B): (********)
outportb(0x7E, 0xFF); // Plane 1: (R): (********)
outportb(0x7E, 0xFF); // Plane 2: (G): (********)
outportb(0x7E, 0x00); // Plane 3: (E): ( )
// Set the 32 pixels at the top-left corner of VRAM to the exact contents of
// the tile register, effectively repeating the tile 4 times. In TDW mode, the
// GRCG ignores the CPU-supplied operand, so we might as well just pass the
// contents of a register with the intended width. This eliminates useless load
// instructions in the compiled assembly, and even sort of signals to readers
// of this code that we do not care about the source value.
*reinterpret_cast<uint32_t far *>(MK_FP(0xA800, 0)) = _EAX;
// Fill the entirety of VRAM with the GRCG tile. A simple C one-liner that will
// probably compile into a single `REP STOS` instruction. Unfortunately, Turbo
// C++ 4.0J only ever generates the 16-bit `REP STOSW` here, even when using
// the `__memset__` intrinsic and when compiling in 386 mode. When targeting
// that CPU and above, you'd ideally want `REP STOSD` for twice the speed.
memset(MK_FP(0xA800, 0), _AL, ((640 / 8) * 400));
However, this might make you wonder why TDW mode is even necessary. If it's
functionally equivalent to RMW mode with a CPU-supplied bitmask made up
entirely of 1 bits (i.e., 0xFF, 0xFFFF, or
0xFFFFFFFF), what's the point? The difference lies in the
hardware implementation: If all you need to do is write tile data to
VRAM, you don't need the read and modify parts of RMW mode
which require additional processing time. The PC-9801 Programmers'
Bible claims a speedup of almost 2× when using TDW mode over equivalent
operations in RMW mode.
And that's the only performance claim I found, because none of these old
PC-98 hardware and programming books did any benchmarks. Then again, it's
not too interesting of a question to benchmark either, as the byte-aligned
nature of TDW blitting severely limits its use in a game engine anyway.
Sure, maybe it makes sense to temporarily switch from RMW to TDW mode
if you've identified a large rectangular and byte-aligned section within a
sprite that could be blitted without a bitmask? But the necessary
identification work likely nullifies the performance gained from TDW mode,
I'd say. In any case, that's pretty deep
micro-optimization territory. Just use TDW mode for the
few cases it's good at, and stick to RMW mode for the rest.
So is this all that can be said about the GRCG? Not quite, because there are
4 bits I haven't talked about yet…
And now we're just 5.37% away from 100% position independence for TH04! From
this point, another 2 pushes should be enough to reach this goal. It might
not look like we're that close based on the current estimate, but a
big chunk of the remaining numbers are false positives from the player shot
control functions. Since we've got a very special deadline to hit, I'm going
to cobble these two pushes together from the two current general
subscriptions and the rest of the backlog. But you can, of course, still
invest in this goal to allow the existing contributions to go to something
else.
… Well, if the store was actually open. So I'd better
continue with a quick task to free up some capacity sooner rather than
later. Next up, therefore: Back to TH02, and its item and player systems.
Shouldn't take that long, I'm not expecting any surprises there. (Yeah, I
know, famous last words…)
Slight change of plans, because we got instructions for
reliably reproducing the TH04 Kurumi Divide Error crash! Major thanks to
Colin Douglas Howell. With those, it also made sense to immediately look at
the crash in the Stage 4 Marisa fight as well. This way, I could release
both of the obligatory bugfix mods at the same time.
Especially since it turned out that I was wrong: Both crashes are entirely
unrelated to the custom entity structure that would have required PI-centric
progress. They are completely specific to Kurumi's and Marisa's
danmaku-pattern code, and really are two separate bugs
with no connection to each other. All of the necessary research nicely fit
into Arandui's 0.5 pushes, with no further deep understanding
required here.
But why were there still three weeks between Colin's message and this blog
post? DMCA distractions aside: There are no easy fixes this time, unlike
📝 back when I looked at the Stage 5 Yuuka crash.
Just like how division by zero is undefined in mathematics, it's also,
literally, undefined what should happen instead of these two
Divide error crashes. This means that any possible "fix" can
only ever be a fanfiction interpretation of the intentions behind ZUN's
code. The gameplay community should be aware of this, and
might decide to handle these cases differently. And if we
have to go into fanfiction territory to work around crashes in the
canon games, we'd better document what exactly we're fixing here and how, as
comprehensible as possible.
With that out of the way, let's look at Kurumi's crash first, since it's way
easier to grasp. This one is known to primarily happen to new players, and
it's easy to see why:
In one of the patterns in her third phase, Kurumi fires a series of 3
aimed rings from both edges of the playfield. By default (that is, on Normal
and with regular rank), these are 6-way rings.
6 happens to be quite a peculiar number here, due to how rings are
(manually) tuned based on the current "rank" value (playperf)
before being fired. The code, abbreviated for clarity:
Let's look at the range of possible playperf values per
difficulty level:
Easy
Normal
Hard
Lunatic
Extra
playperf_min
4
11
20
22
16
playperf_max
16
24
32
34
20
Edit (2022-05-24): This blog post initially had
26 instead of 16 for playperf_min for the Extra Stage. Thanks
to Popfan for pointing out that typo!
Reducing rank to its minimum on Easy mode will therefore result in a
0-ring after tuning.
To calculate the individual angles of each bullet in a ring, ZUN divides
360° (or, more correctly,
📝 0x100) by the total number of
bullets…
Boom, division by zero.
So, what should the workaround look like? Obviously, we want to modify
neither the default number of ring bullets nor the tuning algorithm – that
would change all other non-crashing variations of this pattern on other
difficulties and ranks, creating a fork of the original gameplay. Instead, I
came up with four possible workarounds that all seemed somewhat logical to
me:
Firing no bullet, i.e., interpreting 0-ring literally. This would
create the only constellation in which a call to the bullet group spawn
functions would not spawn at least one new bullet.
Firing a "1-ring", i.e., a single bullet. This would be consistent with
how the bullet spawn functions behave for "0-way" stack and spread
groups.
Firing a "∞-ring", i.e., 200 bullets, which is as much as the game's cap
on 16×16 bullets would allow. This would poke fun at the whole "division by
zero" idea… but given that we're still talking about Easy Mode (and
especially new players) here, it might be a tad too cruel. Certainly the
most trollish interpretation.
Triggering an immediate Game Over, exchanging the hard crash for a
softer and more controlled shutdown. Certainly the option that would be
closest to the behavior of the original games, and perhaps the only one to
be accepted in Serious, High-Level Play™.
As I was writing this post, it felt increasingly wrong for me to make this
decision. So I once again went to Twitter, where 56.3%
voted in favor of the 1-bullet option. Good that I asked! I myself was
more leaning towards the 0-bullet interpretation, which only got 28.7% of
the vote. Also interesting are the 2.3% in favor of the Game Over option but
I get it, low-rank Easy Mode isn't exactly the most competitive mode of
playing TH04.
There are reports of Kurumi crashing on higher difficulties as well, but I
could verify none of them. If they aren't fixed by this workaround, they're
caused by an entirely different bug that we have yet to discover.
Onto the Stage 4 Marisa crash then, which does in fact apply to all
difficulty levels. I was also wrong on this one – it's a hell of a lot more
intricate than being just a division by the number of on-screen bits.
Without having decompiled the entire fight, I can't give a completely
accurate picture of what happens there yet, but here's the rough idea:
Marisa uses different patterns, depending on whether at least one of her
bits is still alive, or all of them have been destroyed.
Destroying the last bit will immediately switch to the bit-less
counterpart of the current pattern.
The bits won't respawn before the pattern ended, which ensures that the
bit-less version is always shown in its entirety after being started or
switched into.
In two of the bit-less patterns, Marisa gradually moves to the point
reflection of her position at the start of the pattern across the playfield
coordinate of (192, 112), or (224, 128) on screen.
The velocity of this movement is determined by both her distance to that
point and the total amount of frames that this instance of the bit-less
pattern will last.
Since this frame amount is directly tied to the frame the player
destroyed the last bit on, it becomes a user-controlled variable. I think
you can see where this is going…
The last 12 frames of this duration, however, are always reserved for a
"braking phase", where Marisa's velocity is halved on each frame.
This part of the code only runs every 4 frames though. This expands the
time window for this crash to 4 frames, rather than just the two frames you
would expect from looking at the division itself.
Both of the broken patterns run for a maximum of 160 frames. Therefore,
the crash will occur when Marisa's last bit is destroyed between frame 152
and 155 inclusive. On these frames, the
last_frame_with_bits_alive variable is set to 148, which is the
crucial 12 duration frames away from the maximum of 160.
Interestingly enough, the calculated velocity is also only
applied every 4 frames, with Marisa actually staying still for the 3 frames
inbetween. As a result, she either moves
too slowly to ever actually reach the yellow point if the last bit
was destroyed early in the pattern (see destruction frames 68 or
112),
or way too quickly, and almost in a jerky, teleporting way (see
destruction frames 144 or 148).
Finally, as you may have already gathered from the formula: Destroying
the last bit between frame 156 and 160 inclusive results in
duration values of 8 or 4. These actually push Marisa
away from the intended point, as the divisor becomes negative.
tl;dr: "Game crashes if last bit destroyed within 4-frame window near end of
two patterns". For an informed decision on a new movement behavior for these
last 8 frames, we definitely need to know all the details behind the crash
though. Here's what I would interpret into the code:
Not moving at all, i.e., interpreting 0 as the middle ground between
positive and negative movement. This would also make sense because a
12-frame duration implies 100% of the movement to consist of
the braking phase – and Marisa wasn't moving before, after all.
Move at maximum speed, i.e., dividing by 1 rather than 0. Since the
movement duration is still 12 in this case, Marisa will immediately start
braking. In total, she will move exactly ¾ of the way from her initial
position to (192, 112) within the 8 frames before the pattern
ends.
Directly warping to (192, 112) on frame 0, and to the
point-reflected target on 4, respectively. This "emulates" the division by
zero by moving Marisa at infinite speed to the exact two points indicated by
the velocity formula. It also fits nicely into the 8 frames we have to fill
here. Sure, Marisa can't reach these points at any other duration, but why
shouldn't she be able to, with infinite speed? Then again, if Marisa
is far away enough from (192, 112), this workaround would warp her
across the entire playfield. Can Marisa teleport according to lore? I
have no idea…
Triggering an immediate Game O– hell no, this is the Stage 4 boss,
people already hate losing runs to this bug!
Asking Twitter worked great for the Kurumi workaround, so let's do it again!
Gotta attach a screenshot of an earlier draft of this blog post though,
since this stuff is impossible to explain in tweets…
…and it went
through the roof, becoming the most successful ReC98 tweet so far?!
Apparently, y'all really like to just look at descriptions of overly complex
bugs that I'd consider way beyond the typical attention span that can be
expected from Twitter. Unfortunately, all those tweet impressions didn't
quite translate into poll turnout. The results
were pretty evenly split between 1) and 2), with option 1) just coming out
slightly ahead at 49.1%, compared to 41.5% of option 2).
(And yes, I only noticed after creating the poll that warping to both the
green and yellow points made more sense than warping to just one of the two.
Let's hope that this additional variant wouldn't have shifted the results
too much. Both warp options only got 9.4% of the vote after all, and no one
else came up with the idea either. In the end,
you can always merge together your preferred combination of workarounds from
the Git branches linked below.)
So here you go: The new definitive version of TH04, containing not only the
community-chosen Kurumi and Stage 4 Marisa workaround variant, but also the
📝 No-EMS bugfix from last year.
Edit (2022-05-31): This package is outdated, 📝 the current version is here!2022-04-18-community-choice-fixes.zip
Oh, and let's also add spaztron64's TH03 GDC clock fix
from 2019 because why not. This binary was built from the community_choice_fixes
branch, and you can find the code for all the individual workarounds on
these branches:
Again, because it can't be stated often enough: These fixes are
fanfiction. The gameplay community should be aware of
this, and might decide to handle these cases differently.
With all of that taking way more time to evaluate and document, this
research really had to become part of a proper push, instead of just being
covered in the quick non-push blog post I initially intended. With ½ of a
push left at the end, TH05's Stage 1-5 boss background rendering functions
fit in perfectly there. If you wonder how these static backdrop images even
need any boss-specific code to begin with, you're right – it's basically the
same function copy-pasted 4 times, differing only in the backdrop image
coordinates and some other inconsequential details.
Only Sara receives a nice variation of the typical
📝 blocky entrance animation: The usually
opaque bitmap data from ST00.BB is instead used as a transition
mask from stage tiles to the backdrop image, by making clever use of the
tile invalidation system:
TH04 uses the same effect a bit more frequently, for its first three bosses.
Next up: Shinki, for real this time! I've already managed to decompile 10 of
her 11 danmaku patterns within a little more than one push – and yes,
that one is included in there. Looks like I've slightly
overestimated the amount of work required for TH04's and TH05's bosses…
Did you know that moving on top of a boss sprite doesn't kill the player in
TH04, only in TH05?
That's the first of only three interesting discoveries in these 3 pushes,
all of which concern TH04. But yeah, 3 for something as seemingly simple as
these shared boss functions… that's still not quite the speed-up I had hoped
for. While most of this can be blamed, again, on TH04 and all of its
hardcoded complexities, there still was a lot of work to be done on the
maintenance front as well. These functions reference a bunch of code I RE'd
years ago and that still had to be brought up to current standards, with the
dependencies reaching from 📝 boss explosions
over 📝 text RAM overlay functionality up to
in-game dialog loading.
The latter provides a good opportunity to talk a bit about x86 memory
segmentation. Many aspiring PC-98 developers these days are very scared
of it, with some even going as far as to rather mess with Protected Mode and
DOS extenders just so that they don't have to deal with it. I wonder where
that fear comes from… Could it be because every modern programming language
I know of assumes memory to be flat, and lacks any standard language-level
features to even express something like segments and offsets? That's why
compilers have a hard time targeting 16-bit x86 these days: Doing anything
interesting on the architecture requires giving the programmer full
control over segmentation, which always comes down to adding the
typical non-standard language extensions of compilers from back in the day.
And as soon as DOS stopped being used, these extensions no longer made sense
and were subsequently removed from newer tools. A good example for this can
be found in an old version of the
NASM manual: The project started as an attempt to make x86 assemblers
simple again by throwing out most of the segmentation features from
MASM-style assemblers, which made complete sense in 1996 when 16-bit DOS and
Windows were already on their way out. But there was a point to all
those features, and that's why ReC98 still has to use the supposedly
inferior TASM.
Not that this fear of segmentation is completely unfounded: All the
segmentation-related keywords, directives, and #pragmas
provided by Borland C++ and TASM absolutely can be the cause of many
weird runtime bugs. Even if the compiler or linker catches them, you are
often left with confusing error messages that aged just as poorly as memory
segmentation itself.
However, embracing the concept does provide quite the opportunity for
optimizations. While it definitely was a very crazy idea, there is a small
bit of brilliance to be gained from making proper use of all these
segmentation features. Case in point: The buffer for the in-game dialog
scripts in TH04 and TH05.
// Thanks to the semantics of `far` pointers, we only need a single 32-bit
// pointer variable for the following code.
extern unsigned char far *dialog_p;
// This master.lib function returns a `void __seg *`, which is a 16-bit
// segment-only pointer. Converting to a `far *` yields a full segment:offset
// pointer to offset 0000h of that segment.
dialog_p = (unsigned char far *)hmem_allocbyte(/* … */);
// Running the dialog script involves pointer arithmetic. On a far pointer,
// this only affects the 16-bit offset part, complete with overflow at 64 KiB,
// from FFFFh back to 0000h.
dialog_p += /* … */;
dialog_p += /* … */;
dialog_p += /* … */;
// Since the segment part of the pointer is still identical to the one we
// allocated above, we can later correctly free the buffer by pulling the
// segment back out of the pointer.
hmem_free((void __seg *)dialog_p);
If dialog_p was a huge pointer, any pointer
arithmetic would have also adjusted the segment part, requiring a second
pointer to store the base address for the hmem_free call. Doing
that will also be necessary for any port to a flat memory model. Depending
on how you look at it, this compression of two logical pointers into a
single variable is either quite nice, or really, really dumb in its
reliance on the precise memory model of one single architecture.
Why look at dialog loading though, wasn't this supposed to be all about
shared boss functions? Well, TH04 unnecessarily puts certain stage-specific
code into the boss defeat function, such as loading the alternate Stage 5
Yuuka defeat dialog before a Bad Ending, or initializing Gengetsu after
Mugetsu's defeat in the Extra Stage.
That's TH04's second core function with an explicit conditional branch for
Gengetsu, after the
📝 dialog exit code we found last year during EMS research.
And I've heard people say that Shinki was the most hardcoded fight in PC-98
Touhou… Really, Shinki is a perfectly regular boss, who makes proper use of
all internal mechanics in the way they were intended, and doesn't blast
holes into the architecture of the game. Even within TH05, it's Mai and Yuki
who rely on hacks and duplicated code, not Shinki.
The worst part about this though? How the function distinguishes Mugetsu
from Gengetsu. Once again, it uses its own global variable to track whether
it is called the first or the second time within TH04's Extra Stage,
unrelated to the same variable used in the dialog exit function. But this
time, it's not just any newly created, single-use variable, oh no. In a
misguided attempt to micro-optimize away a few bytes of conventional memory,
TH04 reserves 16 bytes of "generic boss state", which can (and are) freely
used for anything a boss doesn't want to store in a more dedicated
variable.
It might have been worth it if the bosses actually used most of these
16 bytes, but the majority just use (the same) two, with only Stage 4 Reimu
using a whopping seven different ones. To reverse-engineer the various uses
of these variables, I pretty much had to map out which of the undecompiled
danmaku-pattern functions corresponds to which boss
fight. In the end, I assigned 29 different variable names for each of the
semantically different use cases, which made up another full push on its
own.
Now, 16 bytes of wildly shared state, isn't that the perfect recipe for
bugs? At least during this cursory look, I haven't found any obvious ones
yet. If they do exist, it's more likely that they involve reused state from
earlier bosses – just how the Shinki death glitch in
TH05 is caused by reusing cheeto data from way back in Stage 4 – and
hence require much more boss-specific progress.
And yes, it might have been way too early to look into all these tiny
details of specific boss scripts… but then, this happened:
Looks similar to another
screenshot of a crash in the same fight that was reported in December,
doesn't it? I was too much in a hurry to figure it out exactly, but notice
how both crashes happen right as the last of Marisa's four bits is destroyed.
KirbyComment has suspected
this to be the cause for a while, and now I can pretty much confirm it
to be an unguarded division by the number of on-screen bits in
Marisa-specific pattern code. But what's the cause for Kurumi then?
As for fixing it, I can go for either a fast or a slow option:
Superficially fixing only this crash will probably just take a fraction
of a push.
But I could also go for a deeper understanding by looking at TH04's
version of the 📝 custom entity structure. It
not only stores the data of Marisa's bits, but is also very likely to be
involved in Kurumi's crash, and would get TH04 a lot closer to 100%
PI. Taking that look will probably need at least 2 pushes, and might require
another 3-4 to completely decompile Marisa's fight, and 2-3 to decompile
Kurumi's.
OK, now that that's out of the way, time to finish the boss defeat function…
but not without stumbling over the third of TH04's quirks, relating to the
Clear Bonus for the main game or the Extra Stage:
To achieve the incremental addition effect for the in-game score display
in the HUD, all new points are first added to a score_delta
variable, which is then added to the actual score at a maximum rate of
61,110 points per frame.
There are a fixed 416 frames between showing the score tally and
launching into MAINE.EXE.
As a result, TH04's Clear Bonus is effectively limited to
(416 × 61,110) = 25,421,760 points.
Only TH05 makes sure to commit the entirety of the
score_delta to the actual score before switching binaries,
which fixes this issue.
And after another few collision-related functions, we're now truly,
finally ready to decompile bosses in both TH04 and TH05! Just as the
anything funds were running out… The
remaining ¼ of the third push then went to Shinki's 32×32 ball bullets,
rounding out this delivery with a small self-contained piece of the first
TH05 boss we're probably going to look at.
Next up, though: I'm not sure, actually. Both Shinki and Elis seem just a
little bit larger than the 2¼ or 4 pushes purchased so far, respectively.
Now that there's a bunch of room left in the cap again, I'll just let the
next contribution decide – with a preference for Shinki in case of a tie.
And if it will take longer than usual for the store to sell out again this
time (heh), there's still the
📝 PC-98 text RAM JIS trail word rendering research
waiting to be documented.
EMS memory! The
infamous stopgap measure between the 640 KiB ("ought to be enough for
everyone") of conventional
memory offered by DOS from the very beginning, and the later XMS standard for
accessing all the rest of memory up to 4 GiB in the x86 Protected Mode. With
an optionally active EMS driver, TH04 and TH05 will make use of EMS memory
to preload a bunch of situational .CDG images at the beginning of
MAIN.EXE:
The "eye catch" game title image, shown while stages are loaded
The character-specific background image, shown while bombing
The player character dialog portraits
TH05 additionally stores the boss portraits there, preloading them
at the beginning of each stage. (TH04 instead keeps them in conventional
memory during the entire stage.)
Once these images are needed, they can then be copied into conventional
memory and accessed as usual.
Uh… wait, copied? It certainly would have been possible to map EMS
memory to a regular 16-bit Real Mode segment for direct access,
bank-switching out rarely used system or peripheral memory in exchange for
the EMS data. However, master.lib doesn't expose this functionality, and
only provides functions for copying data from EMS to regular memory and vice
versa.
But even that still makes EMS an excellent fit for the large image files
it's used for, as it's possible to directly copy their pixel data from EMS
to VRAM. (Yes, I tried!) Well… would, because ZUN doesn't do
that either, and always naively copies the images to newly allocated
conventional memory first. In essence, this dumbs down EMS into just another
layer of the memory hierarchy, inserted between conventional memory and
disk: Not quite as slow as disk, but still requiring that
memcpy() to retrieve the data. Most importantly though: Using
EMS in this way does not increase the total amount of memory
simultaneously accessible to the game. After all, some other data will have
to be freed from conventional memory to make room for the newly loaded data.
The most idiomatic way to define the game-specific layout of the EMS area
would be either a struct or an enum.
Unfortunately, the total size of all these images exceeds the range of a
16-bit value, and Turbo C++ 4.0J supports neither 32-bit enums
(which are silently degraded to 16-bit) nor 32-bit structs
(which simply don't compile). That still leaves raw compile-time constants
though, you only have to manually define the offset to each image in terms
of the size of its predecessor. But instead of doing that, ZUN just placed
each image at a nice round decimal offset, each slightly larger than the
actual memory required by the previous image, just to make sure that
everything fits. This results not only in quite
a bit of unnecessary padding, but also in technically the single
biggest amount of "wasted" memory in PC-98 Touhou: Out of the 180,000 (TH04)
and 320,000 (TH05) EMS bytes requested, the game only uses 135,552 (TH04)
and 175,904 (TH05) bytes. But hey, it's EMS, so who cares, right? Out of all
the opportunities to take shortcuts during development, this is among the
most acceptable ones. Any actual PC-98 model that could run these two games
comes with plenty of memory for this to not turn into an actual issue.
On to the EMS-using functions themselves, which are the definition of
"cross-cutting concerns". Most of these have a fallback path for the non-EMS
case, and keep the loaded .CDG images in memory if they are immediately
needed. Which totally makes sense, but also makes it difficult to find names
that reflect all the global state changed by these functions. Every one of
these is also just called from a single place, so inlining
them would have saved me a lot of naming and documentation trouble
there.
The TH04 version of the EMS allocation code was actually displayed on ZUN's monitor in the
2010 MAG・ネット documentary; WindowsTiger already transcribed the low-quality video image
in 2019. By 2015 ReC98 standards, I would have just run with that, but
the current project goal is to write better code than ZUN, so I didn't. 😛
We sure ain't going to use magic numbers for EMS offsets.
The dialog init and exit code then is completely different in both games,
yet equally cross-cutting. TH05 goes even further in saving conventional
memory, loading each individual player or boss portrait into a single .CDG
slot immediately before blitting it to VRAM and freeing the pixel data
again. People who play TH05 without an active EMS driver are surely going to
enjoy the hard drive access lag between each portrait change…
TH04, on the other hand, also abuses the dialog
exit function to preload the Mugetsu defeat / Gengetsu entrance and
Gengetsu defeat portraits, using a static variable to track how often the
function has been called during the Extra Stage… who needs function
parameters anyway, right?
This is also the function in which TH04 infamously crashes after the Stage 5
pre-boss dialog when playing with Reimu and without any active EMS driver.
That crash is what motivated this look into the games' EMS usage… but the
code looks perfectly fine? Oh well, guess the crash is not related to EMS
then. Next u–
OK, of course I can't leave it like that. Everyone is expecting a fix now,
and I still got half of a push left over after decompiling the regular EMS
code. Also, I've now RE'd every function that could possibly be involved in
the crash, and this is very likely to be the last time I'll be looking at
them.
Turns out that the bug has little to do with EMS, and everything to do with
ZUN limiting the amount of conventional RAM that TH04's
MAIN.EXE is allowed to use, and then slightly miscalculating
this upper limit. Playing Stage 5 with Reimu is the most asset-intensive
configuration in this game, due to the combination of
6 player portraits (Marisa has only 5), at 128×128 pixels each
a 288×256 background for the boss fight, tied in size only with the
ones in the Extra Stage
the additional 96×80 image for the vertically scrolling stars during
the stage, wastefully stored as 4 bitplanes rather than a single one.
This image is never freed, not even at the end of the stage.
Remove any single one of the above points, and this crash would have never
occurred. But with all of them combined, the total amount of memory consumed
by TH04's MAIN.EXE just barely exceeds ZUN's limit of 320,000
bytes, by no more than 3,840 bytes, the size of the star image.
But wait: As we established earlier, EMS does nothing to reduce the amount
of conventional memory used by the game. In fact, if you disabled TH04's EMS
handling, you'd still get this crash even if you are running an EMS
driver and loaded DOS into the High Memory Area to free up as much
conventional RAM as possible. How can EMS then prevent this crash in the
first place?
The answer: It's only because ZUN's usage of EMS bypasses the need to load
the cached images back out of the XOR-encrypted 東方幻想.郷
packfile. Leaving aside the general
stupidity of any game data file encryption*, master.lib's decryption
implementation is also quite wasteful: It uses a separate buffer that
receives fixed-size chunks of the file, before decrypting every individual
byte and copying it to its intended destination buffer. That really
resembles the typical slowness of a C fread() implementation
more than it does the highly optimized ASM code that master.lib purports to
be… And how large is this well-hidden decryption buffer? 4 KiB.
So, looking back at the game, here is what happens once the Stage 5
pre-battle dialog ends:
Reimu's bomb background image, which was previously freed to make space
for her dialog portraits, has to be loaded back into conventional memory
from disk
BB0.CDG is found inside the 東方幻想.郷
packfile
file_ropen() ends up allocating a 4 KiB buffer for the
encrypted packfile data, getting us the decisive ~4 KiB closer to the memory
limit
The .CDG loader tries to allocate 52 608 contiguous bytes for the
pixel data of Reimu's bomb image
This would exceed the memory limit, so hmem_allocbyte()
fails and returns a nullptr
ZUN doesn't check for this case (as usual)
The pixel data is loaded to address 0000:0000,
overwriting the Interrupt Vector Table and whatever comes after
The game crashes
The 4 KiB encryption buffer would only be freed by the corresponding
file_close() call, which of course never happens because the
game crashes before it gets there. At one point, I really did suspect the
cause to be some kind of memory leak or fragmentation inside master.lib,
which would have been quite delightful to fix.
Instead, the most straightforward fix here is to bump up that memory limit
by at least 4 KiB. Certainly easier than squeezing in a
cdg_free() call for the star image before the pre-boss dialog
without breaking position dependence.
Or, even better, let's nuke all these memory limits from orbit
because they make little sense to begin with, and fix every other potential
out-of-memory crash that modders would encounter when adding enough data to
any of the 4 games that impose such limits on themselves. Unless you want to
launch other binaries (which need to do their own memory allocations) after
launching the game, there's really no reason to restrict the amount of
memory available to a DOS process. Heck, whenever DOS creates a new one, it
assigns all remaining free memory by default anyway.
Removing the memory limits also removes one of ZUN's few error checks, which
end up quitting the game if there isn't at least a given maximum amount of
conventional RAM available. While it might be tempting to reserve enough
memory at the beginning of execution and then never check any allocation for
a potential failure, that's exactly where something like TH04's crash
comes from.
This game is also still running on DOS, where such an initial allocation
failure is very unlikely to happen – no one fills close to half of
conventional RAM with TSRs and then tries running one of these games. It
might have been useful to detect systems with less than 640 KiB of
actual, physical RAM, but none of the PC-98 models with that little amount
of memory are fast enough to run these games to begin with. How ironic… a
place where ZUN actually added an error check, and then it's mostly
pointless.
Here's an archive that contains both fix variants, just in case. These were
compiled from the th04_noems_crash_fix
and mem_assign_all
branches, and contain as little code changes as possible. Edit (2022-04-18): For TH04, you probably want to download
the 📝 community choice fix package instead,
which contains this fix along with other workarounds for the Divide
error crashes.
2021-11-29-Memory-limit-fixes.zip
So yeah, quite a complex bug, leaving no time for the TH03 scorefile format
research after all. Next up: Raising prices.
Technical debt, part 10… in which two of the PMD-related functions came
with such complex ramifications that they required one full push after
all, leaving no room for the additional decompilations I wanted to do. At
least, this did end up being the final one, completing all
SHARED segments for the time being.
The first one of these functions determines the BGM and sound effect
modes, combining the resident type of the PMD driver with the Option menu
setting. The TH04 and TH05 version is apparently coded quite smartly, as
PC-98 Touhou only needs to distinguish "OPN- /
PC-9801-26K-compatible sound sources handled by PMD.COM"
from "everything else", since all other PMD varieties are
OPNA- / PC-9801-86-compatible.
Therefore, I only documented those two results returned from PMD's
AH=09h function. I'll leave a comprehensive, fully documented
enum to interested contributors, since that would involve research into
basically the entire history of the PC-9800 series, and even the clearly
out-of-scope PC-88VA. After all, distinguishing between more versions of
the PMD driver in the Option menu (and adding new sprites for them!) is
strictly mod territory.
The honor of being the final decompiled function in any SHARED
segment went to TH04's snd_load(). TH04 contains by far the
sanest version of this function: Readable C code, no new ZUN bugs (and
still missing file I/O error handling, of course)… but wait, what about
that actual file read syscall, using the INT 21h, AH=3Fh DOS
file read API? Reading up to a hardcoded number of bytes into PMD's or
MMD's song or sound effect buffer, 20 KiB in TH02-TH04, 64 KiB in
TH05… that's kind of weird. About time we looked closer into this.
Turns out that no, KAJA's driver doesn't give you the full 64 KiB of one
memory segment for these, as especially TH05's code might suggest to
anyone unfamiliar with these drivers. Instead,
you can customize the size of these buffers on its command line. In
GAME.BAT, ZUN allocates 8 KiB for FM songs, 2 KiB for sound
effects, and 12 KiB for MMD files in TH02… which means that the hardcoded
sizes in snd_load() are completely wrong, no matter how you
look at them. Consequently, this read syscall
will overflow PMD's or MMD's song or sound effect buffer if the
given file is larger than the respective buffer size.
Now, ZUN could have simply hardcoded the sizes from GAME.BAT
instead, and it would have been fine. As it also turns out though,
PMD has an API function (AH=22h) to retrieve the actual
buffer sizes, provided for exactly that purpose. There is little excuse
not to use it, as it also gives you PMD's default sizes if you don't
specify any yourself.
(Unless your build process enumerates all PMD files that are part of the
game, and bakes the largest size into both snd_load() and
GAME.BAT. That would even work with MMD, which doesn't have
an equivalent for AH=22h.)
What'd be the consequence of loading a larger file then? Well, since we
don't get a full segment, let's look at the theoretical limit first.
PMD prefers to keep both its driver code and the data buffers in a single
memory segment. As a result, the limit for the combined size of the song,
instrument, and sound effect buffer is determined by the amount of
code in the driver itself. In PMD86 version 4.8o (bundled with TH04
and TH05) for example, the remaining size for these buffers is exactly
45,555 bytes. Being an actually good programmer who doesn't blindly trust
user input, KAJA thankfully validates the sizes given via the
/M, /V, and /E command-line options
before letting the driver reside in memory, and shuts down with an error
message if they exceed 40 KiB. Would have been even better if he calculated
the exact size – even in the current
PMD version 4.8s from
January 2020, it's still a hardcoded value (see line 8581).
Either way: If the file is larger than this maximum, the concrete effect
is down to the INT 21h, AH=3Fh implementation in the
underlying DOS version. DOS 3.3 treats the destination address as linear
and reads past the end of the segment,
DOS
5.0 and DOSBox-X truncate the number of bytes to not exceed the remaining
space in the segment, and maybe there's even a DOS that wraps around
and ends up overwriting the PMD driver code. In any case: You will
overwrite what's after the driver in memory – typically, the game .EXE and
its master.lib functions.
It almost feels like a happy accident that this doesn't cause issues in
the original games. The largest PMD file in any of the 4 games, the -86
version of 幽夢 ~ Inanimate Dream, takes up 8,099 bytes,
just under the 8,192 byte limit for BGM. For modders, I'd really recommend
implementing this properly, with PMD's AH=22h function and
error handling, once position independence has been reached.
Whew, didn't think I'd be doing more research into KAJA's drivers during
regular ReC98 development! That's probably been the final time though, as
all involved functions are now decompiled, and I'm unlikely to iterate
over them again.
And that's it! Repaid the biggest chunk of technical debt, time for some
actual progress again. Next up: Reopening the store tomorrow, and waiting
for new priorities. If we got nothing by Sunday, I'm going to put the
pending [Anonymous] pushes towards some work on the website.
Alright, no more big code maintenance tasks that absolutely need to be
done right now. Time to really focus on parts 6 and 7 of repaying
technical debt, right? Except that we don't get to speed up just yet, as
TH05's barely decompilable PMD file loading function is rather…
complicated.
Fun fact: Whenever I see an unusual sequence of x86 instructions in PC-98
Touhou, I first consult the disassembly of Wolfenstein 3D. That game was
originally compiled with the quite similar Borland C++ 3.0, so it's quite
helpful to compare its ASM to the
officially released source
code. If I find the instructions in question, they mostly come from
that game's ASM code, leading to the amusing realization that "even John
Carmack was unable to get these instructions out of this compiler"
This time though, Wolfenstein 3D did point me
to Borland's intrinsics for common C functions like memcpy()
and strchr(), available via #pragma intrinsic.
Bu~t those unfortunately still generate worse code than what ZUN
micro-optimized here. Commenting how these sequences of instructions
should look in C is unfortunately all I could do here.
The conditional branches in this function did compile quite nicely
though, clarifying the control flow, and clearly exposing a ZUN
bug: TH05's snd_load() will hang in an infinite loop when
trying to load a non-existing -86 BGM file (with a .M2
extension) if the corresponding -26 BGM file (with a .M
extension) doesn't exist either.
Unsurprisingly, the PMD channel monitoring code in TH05's Music Room
remains undecompilable outside the two most "high-level" initialization
and rendering functions. And it's not because there's data in the
middle of the code segment – that would have actually been possible with
some #pragmas to ensure that the data and code segments have
the same name. As soon as the SI and DI registers are referenced
anywhere, Turbo C++ insists on emitting prolog code to save these
on the stack at the beginning of the function, and epilog code to restore
them from there before returning.
Found that out in
September 2019, and confirmed that there's no way around it. All the
small helper functions here are quite simply too optimized, throwing away
any concern for such safety measures. 🤷
Oh well, the two functions that were decompilable at least indicate
that I do try.
Within that same 6th push though, we've finally reached the one function
in TH05 that was blocking further progress in TH04, allowing that game
to finally catch up with the others in terms of separated translation
units. Feels good to finally delete more of those .ASM files we've
decompiled a while ago… finally!
But since that was just getting started, the most satisfying development
in both of these pushes actually came from some more experiments with
macros and inline functions for near-ASM code. By adding
"unused" dummy parameters for all relevant registers, the exact input
registers are made more explicit, which might help future port authors who
then maybe wouldn't have to look them up in an x86 instruction
reference quite as often. At its best, this even allows us to
declare certain functions with the __fastcall convention and
express their parameter lists as regular C, with no additional
pseudo-registers or macros required.
As for output registers, Turbo C++'s code generation turns out to be even
more amazing than previously thought when it comes to returning
pseudo-registers from inline functions. A nice example for
how this can improve readability can be found in this piece of TH02 code
for polling the PC-98 keyboard state using a BIOS interrupt:
inline uint8_t keygroup_sense(uint8_t group) {
_AL = group;
_AH = 0x04;
geninterrupt(0x18);
// This turns the output register of this BIOS call into the return value
// of this function. Surprisingly enough, this does *not* naively generate
// the `MOV AL, AH` instruction you might expect here!
return _AH;
}
void input_sense(void)
{
// As a result, this assignment becomes `_AH = _AH`, which Turbo C++
// never emits as such, giving us only the three instructions we need.
_AH = keygroup_sense(8);
// Whereas this one gives us the one additional `MOV BH, AH` instruction
// we'd expect, and nothing more.
_BH = keygroup_sense(7);
// And now it's obvious what both of these registers contain, from just
// the assignments above.
if(_BH & K7_ARROW_UP || _AH & K8_NUM_8) {
key_det |= INPUT_UP;
}
// […]
}
I love it. No inline assembly, as close to idiomatic C code as something
like this is going to get, yet still compiling into the minimum possible
number of x86 instructions on even a 1994 compiler. This is how I keep
this project interesting for myself during chores like these.
We might have even reached peak
inline already?
And that's 65% of technical debt in the SHARED segment repaid
so far. Next up: Two more of these, which might already complete that
segment? Finally!