Commit graph

11970 commits

Author SHA1 Message Date
Bill Currie
9348f7b89c [gamecode] Preserve the return pointer across calls
This required delaying the setting of the return pointer by call until
after the current pointer had been saved, and thus passing the desired
pointer into PR_CallFunction (which does have some advantages for C
functions calling progs functions, but some dangers too (should ensure a
128 byte (32 word) buffer when calling untrusted code (which is any,
really)).
2022-01-17 19:12:28 +09:00
Bill Currie
e9af549195 [gamecode] Fix some goofs in the callchain test
They made it difficult to tell when I got things working :P
2022-01-17 18:51:50 +09:00
Bill Currie
306fcbfbd0 [gamecode] Improve callchain test
Add another function call to further mess up the return value address
and verify just where it is pointing.
2022-01-17 16:55:39 +09:00
Bill Currie
89e120ba34 [gamecode] Make return support 32 words
This took interpreting the lower 5 bits of operand c as size - 1, and
0xffff as void (0 words).
2022-01-17 16:54:27 +09:00
Bill Currie
94ef9931a5 [gamecode] Add two more call/return tests
The two tests check that 32 words can be returned (fails) and that
function calls can be nested and their values returned properly (also
fails).
2022-01-17 16:30:13 +09:00
Bill Currie
736387bc88 [gamecode] Preserve base registers across calls
With this, functions can call other functions without having to worry
about whether the base registers they set up are still valid.
2022-01-17 15:08:58 +09:00
Bill Currie
8da1163a82 [gamecode] Add bases check to callret test
The base registers must be preserved across a function call and they
currently are not, thus the updated test fails again.
2022-01-17 14:59:17 +09:00
Bill Currie
213434b705 [gamecode] Save and restore data stack in call stack
This fixes the issue of the data stack not being restored properly
because the returning function needs to return a value from its local
variables (stored on the stack) and accessing stack data below the stack
pointer is a bad idea (sure, no interrupts yet, but who knows...).
2022-01-17 14:45:14 +09:00
Bill Currie
8e5c2c6534 [gamecode] Correct type of return address in stack
This is another one of those "why signed?" things: can't have negative
return addresses.
2022-01-17 14:43:43 +09:00
Bill Currie
cf3106ce28 [gamecode] Correct call's mode selection
Call's operand c is used to specify where the return value of the
function is to be stored. This gets both the correct function being
called, and the value being returned correctly. Test still fails due to
the stack restoration issue.
2022-01-17 14:29:14 +09:00
Bill Currie
b3909dbe4c [gamecode] Add a test for call and return
It currently fails for two reasons:
 - call's mode selection is incorrect (never updated from when there was
   only the one call instruction and the mode was encoded in operand c)
 - return should automatically restore the stack pointer to the value it
   had on entry to the function, thus allowing local values stored on
   the stack to be safely returned.
2022-01-17 14:26:12 +09:00
Bill Currie
56d8bdee82 [gamecode] Add a fixme about using gcc's vec convert
It might produce better code than the way I'm currently doing it.
2022-01-17 10:32:19 +09:00
Bill Currie
1beadbf871 [gamecode] Add tests for the branch instructions 2022-01-17 09:57:54 +09:00
Bill Currie
1baf17b780 [gamecode] Add tests for the jump instructions 2022-01-17 09:34:16 +09:00
Bill Currie
1397c94ef5 [gamecode] Add tests for the state instructions 2022-01-16 22:17:17 +09:00
Bill Currie
0bd05c71ac [gamecode] Use unsigned for entity values
I don't know why they were ever signed (oversight at id and just
propagated?). Anyway, this resulted in "unsigned" spreading a bit, but
all to reasonable places.
2022-01-16 22:15:18 +09:00
Bill Currie
2b82533526 [gamecode] Add double time state instructions
This has been a long-held wishlist item, really, and I thought I might
as well take the opportunity to add the instructions. The double
versions of STATE require both the nextthink field and time global to be
double (but they're not resolved properly yet: marked with
"FIXME double time" comments).

Also, the frame number for double time state is integer rather than
float.
2022-01-16 21:27:49 +09:00
Bill Currie
d57712975e [gamecode] Add tests for the hops instructions
And, of course, fix a bug in the inverted hops (yay for tests).
2022-01-16 18:46:29 +09:00
Bill Currie
7ea12b3ff9 [gamecode] Implement the HOPS sub-instructions
In the end, I decided any/all/none should be separate from the other
horizontal ops, if I even do them (can be implemented by first
converting to bool, then using the appropriate horizontal operation (& |
etc).
2022-01-16 16:28:48 +09:00
Bill Currie
8050c7bd77 [gamecode] Rearrange several instructions
ANY/ALL/NONE have been temporarily removed until I implement the HOPS
(horizontal operations) sub-instructions, which will all both 32-bit and
64-bit operands and several other operations (eg, horizontal add).

All the fancy addressing modes for the conditional branch instructions
have been permanently removed: I decided the gain was too little for the
cost (24 instructions vs 6). JUMP and CALL retain their addressing
modes, though.

Other instructions have been shuffled around a little to fill most of
the holes in the upper block of 256 instructions: just a single small
7-instruction hole.

Rearrangements in the actual engine are mostly just to keep the code
organized. The only real changes were the various IF statements and
dealing with the resulting changes in their addressing.
2022-01-16 14:22:04 +09:00
Bill Currie
dc4df49fff [gamecode] Add tests for the swizzle instructions
While not always generated (yet?) the tests are at least in. Just not
sure about auto-generated tests that can't be hand-edited.
2022-01-15 22:44:58 +09:00
Bill Currie
49dcd5ef40 [gamecode] Add tests for the with instruction(s)
While mode 4 was tested a lot, none of the rest were. Also added a full
reset instruction.
2022-01-15 18:44:11 +09:00
Bill Currie
38ab0f0243 [gamecode] Move pushregs and popregs into with
This frees up another two instructions.
2022-01-15 16:51:59 +09:00
Bill Currie
a14b9f8252 [gamecode] Move the lea block
It just feels better with it being close the load/store etc
2022-01-15 16:39:52 +09:00
Bill Currie
e133de8c89 [gamecode] Drop B addressing for lea
When creating the tests for lea, I noticed that B was yet another simple
assign, so I decided it was best to drop it and move E into its place
(freeing up another instruction).
2022-01-15 16:27:46 +09:00
Bill Currie
11e89c8c97 [gamecode] Add tests for lea 2022-01-15 16:07:11 +09:00
Bill Currie
fae432f46e [gamecode] Make no-op conversions simple copies
Most useful for 64-bit values as only one instruction is needed to move
the data around rather than two, but could be slightly faster for 32-bit
as the addressing is simpler (needs profiling).
2022-01-15 15:44:53 +09:00
Bill Currie
bffcbfc9fc [gamecode] Add tests for bitops 2022-01-15 13:59:03 +09:00
Bill Currie
7cd398d4a7 [gamecodee] Add tests for move and memset 2022-01-15 13:20:33 +09:00
Bill Currie
6f1f56aea7 [gamecode] Make commented braces match
They annoyed me
2022-01-14 22:51:18 +09:00
Bill Currie
1aa4844bf6 [gamecode] Make string ops mostly conform with bools
The compare/ne operator returns "random" -ve, 0, +ve values (really,
just the numerical difference between the chars of the strings), but all
the rest return -1 for true and 0 for false, as with the rest of the
comparison operators.
2022-01-14 22:44:08 +09:00
Bill Currie
c8362c28fe [gamecode] Add tests for string ops
Does not include string concatenation because I don't feel like messing
with zone init, but all the other operators are tested (currently
failing due to bool convention)
2022-01-14 22:42:56 +09:00
Bill Currie
f4eeed36b7 [gamecode] Add tests for the scale instructions 2022-01-14 19:46:35 +09:00
Bill Currie
1a2ac24d8d [gamecode] Make num_globals() calculate actual globals
It calculating only the size of the array (which was often 4 or 8
globals per element) proved to be a pain when I forgot to alter the size
for the new scale tests. Fixing the size calculation even found a bug in
the shiftop tests.
2022-01-14 19:45:51 +09:00
Bill Currie
a81067603c [gamecode] Mention possibly undefined behavior
It seems casting from float/double to [unsigned] int/long when the value
doesn't fit is undefined (which would explain the inconsistent results).
Mentioning the possibility seems like a good idea should the results for
such casts change and cause the tests to fail.
2022-01-14 16:52:44 +09:00
Bill Currie
a1c1c9fcf0 [gamecode] Add test for conversions to double
And fix an incorrect base index in the bool64 tests.
2022-01-14 15:51:49 +09:00
Bill Currie
1f73b26d24 [gamecode] Correct incorrect bool32-bool64 conversion
And add tests for long, ulong and bool64 conversions.
2022-01-14 14:54:35 +09:00
Bill Currie
4bf934e6b9 [gamecode] Correct incorrect bool64-bool32 conversion
And add tests for float, uint and bool32 conversions.
2022-01-14 13:54:03 +09:00
Bill Currie
e26fb49df7 [gamecode] Get conversion to int working for all types
Bools turned out to be a problem to due to me wanting any non-zero value
to be treated as true thus had to expand them out as well as the
floating point <-> integral conversions.
2022-01-14 12:11:23 +09:00
Bill Currie
cd68455e46 [gamecode] Add tests for converting to int
They currently fail because for vector values, gcc casts the view, not
the value, so vec4 cast to ivec4 simply views the bits as int rather
than doing the actual conversion.
2022-01-14 12:11:22 +09:00
Bill Currie
e8e0a69628 [gamecode] Add conversions to/from bool
For now, from bool results in 0/1, but conversion to bool guarantees
0/-1 and correct interpretations for floating point types.
2022-01-13 17:37:44 +09:00
Bill Currie
ba1d73200f [gamecode] Clean up a pile of duplicate code
All those duplicated formats were getting unwieldy, especially as I want
to add more conversion modes.
2022-01-13 16:10:06 +09:00
Bill Currie
3eb2194343 [gamecode] Invert the meaning of the skip matrix
Rather than specifying that the conversion should be skipped, it now
specifies the mode of the conversions (with 0 being no conversion). This
is in preparation for boolean conversion.
2022-01-13 15:58:12 +09:00
Bill Currie
6f6f47e27e [gamecode] Drop bool ops in favor of long bit ops
I realized that being able to do bit-wise operations with 64-bit values
(and 256-bit vectors) is far more important than some convenient boolean
logic operators. The logic ops can be handled via the bit-wise ops so
long as the values are all properly boolean, and I plan on adding some
boolean conversion ope, so no real loss.
2022-01-13 14:24:11 +09:00
Bill Currie
424bdcbf96 [gamecode] Implement the scale instructions
Both float 2,3,4 vectors and double 2,3,4 vectors (1 would be just a
copy of the mul instructions).

This completes the currently planned instructions. Now for testing.
2022-01-13 13:53:07 +09:00
Bill Currie
3587b13a40 [gamecode] Implement the conversion instructions
Not all possibilities are supported because converting between int and
uint, and long and ulong is essentially a no-op. However, thanks to
Deek's suggestion, not only are all reasonable conversions available,
conversions for all widths are available, so vector conversions are
supported.

The code for the conversions is generated.
2022-01-13 13:51:24 +09:00
Bill Currie
f7181a09b4 [gamecode] Add tests for shiftops
They're in test-unsigned because 2/3 of them are unsigned.
2022-01-12 10:24:59 +09:00
Bill Currie
6229ae8ecc [gamecode] Add tests for unsigned comparisons
And fix the implementation: I had used the wrong macro.
2022-01-11 13:00:54 +09:00
Bill Currie
e7d7ec1989 [gamecode] Add tests for signed comparison ops
Fortunately, they all pass without issues.
2022-01-11 09:37:38 +09:00
Bill Currie
0fb6619585 [gamecode] Compact the convert instructions into one
Thanks to Deek for the suggestion: the mode (ie, src and dst types) are
encoded in st->b. Actual code not written yet, but this frees up 13
instructions: now have 74 available for really interesting stuff :)
2022-01-10 11:53:57 +09:00