 t3hn3rd
|
d34997aa4e
|
feature: implement WASM parser sections and VM opcode handlers
Parser implementations:
- dataSection: parse segments, evaluate i32.const offset expressions,
copy data bytes into linear memory
- globalSection: parse globals with init expression evaluation and
type/mutability handling
- memorySection: parse memory limits (initial/max pages, flags)
- startSection: parse start function index
- wasm.parser: route new section types to their handlers
VM opcode implementations:
- Control: nop, unreachable (trap), drop, select, return
- Variables: local.get/set/tee, global.get/set
- Constants: i32.const, i64.const, f32.const, f64.const
- I32 arithmetic: add, sub, mul, div_s, div_u, rem_s, rem_u,
clz, ctz, popcnt
- I32 bitwise: and, or, xor, shl, shr_s, shr_u, rotl, rotr
- I32 comparison: eqz, eq, ne, lt_s/u, gt_s/u, le_s/u, ge_s/u
- I64 arithmetic: add, sub, mul, div_s, div_u, rem_s, rem_u,
clz, ctz, popcnt
- I64 bitwise: and, or, xor, shl, shr_s, shr_u, rotl, rotr
- I64 comparison: eqz, eq, ne, lt_s/u, gt_s/u, le_s/u, ge_s/u
- F32/F64 comparison: eq, ne, lt, gt, le, ge
- Memory loads: i32/i64/f32/f64.load, i32.load8_s/u, i32.load16_s/u,
i64.load8_s/u, i64.load16_s/u, i64.load32_s/u
- Memory stores: i32/i64/f32/f64.store, i32.store8, i32.store16,
i64.store8, i64.store16, i64.store32
- Memory management: memory.size, memory.grow
Bug fixes:
- memory.grow: guard against uint32 underflow when growing 0 pages
- select: fix off-by-one stack indexing for val1/val2
- i32/i64.shr_s: use manual arithmetic shift right (logical shift +
sign-bit mask fill) instead of RTL intrinsics for bare-metal
kernel compatibility
Type additions:
- wasm.types: new records and pointer types for globals, locals,
and section structures
|
2026-02-24 00:31:36 +00:00 |
|