-
Notifications
You must be signed in to change notification settings - Fork 0
[Bug] wasm mod capability ceilings are too low for practical worldgen — 4MB memory, 25ms call time #6
Copy link
Copy link
Open
Labels
area:guest-runtimeRuntime-loaded guest execution model, host/runtime behavior, guest session semantics.Runtime-loaded guest execution model, host/runtime behavior, guest session semantics.component:enginefreven-engine: core engine/runtime/simulation/client-server internals.freven-engine: core engine/runtime/simulation/client-server internals.priority:p2Normal priority. Planned work, worthwhile but not urgent.Normal priority. Planned work, worthwhile but not urgent.status:triageNew / untriaged. Needs initial review and classification.New / untriaged. Needs initial review and classification.transport:wasmSpecific to the WASM guest transport/backend.Specific to the WASM guest transport/backend.
Milestone
Description
Summary
Hard caps on wasm capabilities block practical worldgen. Minimal freven_guest_sdk module already uses ~300KB at startup, leaving ~3.7MB headroom. Any noise tables, structure data, or multi-section output hits the ceiling fast. 25ms is also tight for non-trivial generators on slower hardware. Expected: ceilings high enough to allow real worldgen workloads.
Steps to reproduce
# Try setting memory higher than 4MB in mod.toml:
max_linear_memory_bytes = 67108864
# -> Error: out of allowed range (max 4194304)
# Try setting call time higher than 25ms:
max_call_millis = 50
# -> Error: out of allowed range (max 25)
Required: freven_boot --version output
freven_boot 0.1.0
devkit_manifest /Users/yunnwi/Downloads/Games/freven-devkit-v0.1.0-rc5-aarch64-apple-darwin/manifest/devkit_manifest.toml
devkit_version = "0.1.0-rc5"
build_date_utc = "2026-03-11T23:41:29Z"
target = "aarch64-apple-darwin"
net_proto = "v1"
release = true
boot_ref = "freven-boot@devkit-v0.1.0-rc5"
boot_commit = "bbf92e3f6dfdd1a7c9a2416ef7b8de9743693b73"
engine_ref = "freven-engine@v0.1.3-rc3"
engine_commit = "481daf931bc278fde61d2625f69d786f2a54857a"
sdk_ref = "freven-sdk@v0.1.2-rc2"
sdk_commit = "7148f890dae5e64874978f24c4c88cca35cd6e31"
vanilla_ref = "freven-vanilla@v0.1.2-rc2"
vanilla_commit = "0f6ac778217cb514ad08368003f8552de026f0bd"
notes = ""
Logs and artifacts
Error: failed to attach wasm mods: wasm mod 'yunnwi.hello_mod' capability 'max_linear_memory_bytes' value 67108864 is out of allowed range (max 4194304)
Error: failed to attach wasm mods: wasm mod 'yunnwi.hello_mod' capability 'max_call_millis' value 50 is out of allowed range (max 25)
OS
macOS Apple M1 aarch64
GPU / Driver (if relevant)
No response
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
area:guest-runtimeRuntime-loaded guest execution model, host/runtime behavior, guest session semantics.Runtime-loaded guest execution model, host/runtime behavior, guest session semantics.component:enginefreven-engine: core engine/runtime/simulation/client-server internals.freven-engine: core engine/runtime/simulation/client-server internals.priority:p2Normal priority. Planned work, worthwhile but not urgent.Normal priority. Planned work, worthwhile but not urgent.status:triageNew / untriaged. Needs initial review and classification.New / untriaged. Needs initial review and classification.transport:wasmSpecific to the WASM guest transport/backend.Specific to the WASM guest transport/backend.