feat(smolmachines): rewrite Smolfile to smolvm 0.8.0 schema + drop gvproxy (PRD 0023 chunk 2a)
test / unit (pull_request) Successful in 21s
test / integration (pull_request) Successful in 39s

First sub-PR of chunk 2: rewrite the renderer chunk 1 shipped to
match smolvm 0.8.0's actual Smolfile shape, delete the dead
gvproxy renderer + its tests, simplify the prepare flow now that
there's no gvproxy socket + no loopback-port allocation.

Smolfile renderer:
- Old shape (under the abandoned gvproxy design): name = ...,
  command = [...], [[net]] attachment = "unixgram",
  socket = "...".
- New shape (smolvm 0.8.0): env = [...] (sorted K=V pairs),
  [network] allow_cidrs = ["<bundle-ip>/32"]. Nothing else.
  image / entrypoint / cmd come from the .smolmachine artifact
  built in chunk 2b; cpus / memory left at smolvm defaults.
- Tests assert no leakage of TSI's --outbound-localhost-only or
  the old gvproxy/unixgram keys.

util.py:
- smolmachines_gvproxy_subnet → smolmachines_bundle_subnet,
  returning (subnet, gateway, bundle_ip). bundle_ip is always
  at .2 (gateway .1); subnet is /24, third octet derived from
  the slug hash, skipping the docker-default 17 to avoid the
  common 192.168.17.x collision.
- allocate_loopback_port: deleted. The bundle gets a pinned
  docker IP now; the agent dials that IP directly through TSI.
- smolmachines_preflight: dropped the gvproxy check; only
  smolvm is required.

prepare.py:
- Drops the gvproxy.yaml render + the loopback port allocation
  + the gvproxy_socket field on the plan.
- Derives subnet / gateway / bundle_ip from the slug and
  populates the new SmolmachinesBottlePlan fields.
- Agent env now uses IP-literal URLs (http://<bundle-ip>:8888
  etc) since the guest will have no DNS resolver inside TSI's
  allowlist.

bottle_plan.py:
- Old fields: gvproxy_config_path, gvproxy_socket,
  gvproxy_subnet, gvproxy_gateway, host_port_map.
- New fields: bundle_subnet, bundle_gateway, bundle_ip,
  smolfile_path. (smolmachine artifact path lands in chunk 2b.)

Net: -410 lines. Full unit suite: 516 passing.

The VM lifecycle + bundle bringup + launch wiring + smoke tests
land in chunk 2b.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
2026-05-27 04:01:07 -04:00
parent b57256789f
commit c73d717f71
8 changed files with 203 additions and 613 deletions
+28 -64
View File
@@ -1,84 +1,48 @@
"""Slug / preflight helpers for the smolmachines backend (PRD 0023).
Backend-specific utilities the prepare/launch flow reaches for.
Kept in its own module so the renderers can be unit-tested without
importing the docker subprocess paths."""
"""Slug / preflight / subnet helpers for the smolmachines backend
(PRD 0023). Kept in its own module so the renderers can be
unit-tested without importing the docker subprocess paths."""
from __future__ import annotations
import hashlib
import os
import shutil
import socket
from contextlib import closing
from ...log import die
def smolmachines_preflight() -> None:
"""Ensure the binaries the smolmachines backend shells out to are
on PATH. Called from the backend's `_resolve_plan` before any
file writes; gives the operator a clear pointer rather than a
cryptic FileNotFoundError later.
`smolvm` drives the libkrun microVM lifecycle. `gvproxy`
(gvisor-tap-vsock) is the userspace TCP/IP stack the guest's
virtio-net device hooks into."""
missing: list[tuple[str, str]] = []
if shutil.which("smolvm") is None:
missing.append((
"smolvm",
"brew install smolmachines-dev/smolmachines/smolmachines "
"# or build from https://smolmachines.com/",
))
if shutil.which("gvproxy") is None:
missing.append((
"gvproxy",
"go install "
"github.com/containers/gvisor-tap-vsock/cmd/gvproxy@latest "
"# requires Go + ensure $GOPATH/bin on PATH",
))
if not missing:
"""Ensure `smolvm` is on PATH before the launch flow runs.
Called from `_resolve_plan`; gives the operator a clear
install pointer rather than a cryptic FileNotFoundError
later. `gvproxy` is no longer required — see the PRD's design
pivot section."""
if shutil.which("smolvm") is not None:
return
lines = [
f"CLAUDE_BOTTLE_BACKEND=smolmachines requires the following "
f"binar{'y' if len(missing) == 1 else 'ies'} on PATH:",
]
for name, install in missing:
lines.append(f" - {name} install: {install}")
die("\n".join(lines))
die(
"CLAUDE_BOTTLE_BACKEND=smolmachines requires `smolvm` on "
"PATH. Install with: "
"curl -sSL https://smolmachines.com/install.sh | sh"
)
def smolmachines_gvproxy_subnet(slug: str) -> tuple[str, str]:
"""Derive a per-bottle subnet + gateway IP from the slug.
def smolmachines_bundle_subnet(slug: str) -> tuple[str, str, str]:
"""Derive a per-bottle docker subnet + gateway IP + bundle IP
from the slug.
Returns `(subnet_cidr, gateway_ip)`. The third octet comes from
SHA-256 of the slug mod 254 (1..254, skipping 0 and 255 + the
127.x.x.x range) — collision-free for any reasonable concurrent-
bottle count and stable across `start` / `resume` of the same
bottle. Bottles that DO collide would fail at gvproxy bringup;
`launch.py` (chunk 2) detects the conflict and surfaces a clear
error rather than silently reusing another bottle's gateway."""
Returns `(subnet_cidr, gateway_ip, bundle_ip)`. The third
octet comes from SHA-256 of the slug mod 254 (skipping 17 to
avoid the docker-default bridge), so parallel bottles get
distinct /24s and `resume` reuses the same /24. The bundle
container always lands at `.2`; gateway is `.1`; the smolvm
Smolfile's `allow_cidrs` is `<bundle_ip>/32`."""
digest = hashlib.sha256(slug.encode("utf-8")).digest()
octet = (digest[0] % 254) + 1
# Skip docker-default 17 to dodge the usual second-bridge
# collision. Operators with docker's default bridge at
# 172.17.x.x — common Linux setup — would otherwise see a
# collision the moment they run a parallel docker bottle.
# Skip the docker-default bridge to dodge the most common
# collision (operators with `docker0` at 172.17.x.x or a
# 192.168.17.x VPN client).
if octet == 17:
octet = 18
subnet = f"192.168.{octet}.0/24"
gateway = f"192.168.{octet}.1"
return subnet, gateway
def allocate_loopback_port() -> int:
"""Bind / release dance to grab a free TCP port on 127.0.0.1.
Race window: a parallel allocator could grab the same port
between our close and the bundle's bind. The window is small
enough that chunk 1 doesn't address it; chunk 3 will add a
retry loop if it shows up in practice."""
with closing(socket.socket(socket.AF_INET, socket.SOCK_STREAM)) as s:
s.bind(("127.0.0.1", 0))
return int(s.getsockname()[1])
bundle_ip = f"192.168.{octet}.2"
return subnet, gateway, bundle_ip