interpreter: removes unneeded stack walks

[mathetake]

This makes the interpreter's compilation faster by removing the
unnecessary stack walks. As a result, for some binary, we observe
93% drops in the compilation performance(!) with an additional
but negligible memory usage.

goos: darwin
goarch: arm64
pkg: github.com/tetratelabs/wazero
                      │   old.txt    │              new.txt               │
                      │    sec/op    │   sec/op     vs base               │
Compilation/wazero-10    130.1m ± 1%   124.7m ± 1%   -4.12% (p=0.002 n=6)
Compilation/zig-10      9097.2m ± 0%   549.0m ± 1%  -93.96% (p=0.002 n=6)
Compilation/zz-10         1.159 ± 5%    1.145 ± 3%   -1.20% (p=0.041 n=6)
TinyGo/Compile/container_heap.test-10     16.54m ± 1%   15.60m ± 4%  -5.66% (p=0.002 n=6)
TinyGo/Compile/container_list.test-10     16.36m ± 1%   15.42m ± 1%  -5.74% (p=0.002 n=6)
TinyGo/Compile/container_ring.test-10     16.08m ± 1%   15.17m ± 0%  -5.64% (p=0.002 n=6)
TinyGo/Compile/crypto_des.test-10         16.79m ± 1%   15.84m ± 0%  -5.68% (p=0.002 n=6)
TinyGo/Compile/crypto_md5.test-10         16.61m ± 1%   15.63m ± 0%  -5.88% (p=0.002 n=6)
TinyGo/Compile/crypto_rc4.test-10         15.93m ± 0%   15.05m ± 1%  -5.54% (p=0.002 n=6)
TinyGo/Compile/crypto_sha1.test-10        16.63m ± 0%   15.69m ± 0%  -5.66% (p=0.002 n=6)
TinyGo/Compile/crypto_sha256.test-10      17.15m ± 1%   16.13m ± 1%  -6.00% (p=0.002 n=6)
TinyGo/Compile/crypto_sha512.test-10      17.43m ± 2%   16.30m ± 1%  -6.49% (p=0.002 n=6)
TinyGo/Compile/encoding_ascii85.test-10   16.57m ± 1%   15.66m ± 1%  -5.52% (p=0.002 n=6)