perf: borrow the executing function from its module instance

[matthargett]

Every call or return between two functions cloned the callee's Shared<WasmFunction> and dropped the previous one (two atomic refcount updates each way) and looked the function up in the store.

A module instance now keeps its own functions, which the store allocates contiguously (helping with locality for cache and prefetch). InterpreterRuntime holds the instance for the whole run and the executor borrows the executing function from it, so a call or return inside the instance switches a reference (I think this is what makes cache evict more often than I would expect). A direct call to one of the module's own functions also skips the address table, the host check and the owner check. Execution that moves into another instance's function (through an import, a table, a function reference, a return or an unwinding exception) ends the run and resumes with an executor for that instance. Fuel and time budgets carry over, so budgeted runs should suspend at the same points as before.

Entering a function also skips the value-stack lanes it never uses (most functions only touch the 32-bit lane), and a single-result return moves its result down once instead of popping and pushing it.

tests/cross_instance_calls.rs covers calls, tail calls, table calls, callbacks and exceptions across an instance boundary in both directions, and fuel- and time-budgeted runs that cross it.

Microbenchmark explains the uplift in the larger integrated benchmarks: a loop calling a one-line function drops from 393 to 311 instructions per iteration. Change in cycles per call against next (d1165c2), on the efficiency cores of an iPhone 12 (A14), iPhone XS Max (A12) and iPhone SE (A13), median of five interleaved launches per build:

benchmark	A14	A12	A13
xmrsplayer (1024-frame buffer)	−6.2%	−2.6%	−3.9%
audio DSP (1000 frames × 512)	+0.5%	+3.7%	−0.1%
graphql-validation (AS)	−2.4%	−2.7%	−3.5%
multi-memory twin: one memory	+0.5%	−0.3%	+0.0%
crc32 (64 KB)	+0.1%	−0.2%	−0.1%
convolution 256×256	+0.1%	+0.2%	+0.1%
sieve (10000)	−0.2%	−0.7%	+0.1%
bulk_memory (memory.copy/fill)	−0.1%	+0.0%	+0.0%
matmul relaxed-simd FMA	+1.5%	−5.5%	+4.6%
GC binary trees (~130K struct.new)	−0.9%	−1.0%	−1.6%
fib(30)	−5.5%	−5.0%	−6.0%
tail-call FSM (65536 return_call)	−13.3%	−11.6%	−8.5%
call_indirect (200K)	−6.5%	−10.3%	−7.7%
call_ref (200K)	−12.4%	−14.5%	−9.4%
vtable_poly4 (200K)	−10.8%	−10.7%	−9.3%
EH parser, exnref (4096 stmts, 25% throw)	−18.7%	−18.8%	−16.4%
geomean, cycles	−4.9%	−5.2%	−4.0%
geomean, instructions	−4.2%	−4.8%	−3.9%
geomean, wall time	−4.8%	−4.7%	−4.2%
rows faster (cycles)	11/16	13/16	11/16