Commit Graph
4 Commits
Author SHA1 Message Date
ARIA e148c83622 feat: experimental NVIDIA power control via RM ioctl interface
Adds an experimental power-cap mode using the undocumented RM ioctl
interface (based on panchovix's LACT PR #1205) to set power limits
below the VBIOS minimum (down to 30 W).

- hal/rm_power.py: RM ioctl power-cap read/write/reset + runtime probe
- limits.py: power_cap_mode (nvml/ioctl) with support detection
- config.py: persist power_cap_mode per GPU
- profiles: record/apply power_cap_mode
- server.py: POST /api/limits validates ioctl support (409 on failure)
- cli.py: profile save falls back to persisted mode
- client.py: power_cap_mode in Limits
- frontend: toggle + warning with panchovix attribution (LACT #1205)
- tests: test_rm_power.py (unit) + integration coverage
- Makefile: add test_rm_power.py to make test

Also includes automated linter reformatting (prettier, ruff, shellcheck,
isort, markdownlint) that the linter would apply anyway.
2026-09-17 22:44:27 +02:00
ARIA 39701c12ff security: harden web server, daemon socket, and write paths
Security review findings, fixed and verified:

Critical
- Fix unauthenticated arbitrary file read: the SPA catch-all route
  joined the raw URL path onto the dist dir without containment, so
  encoded '..' segments (/%2e%2e/etc/passwd) leaked any file readable
  by the root server. Resolve with realpath and reject paths outside
  the dist dir (fail-closed 404).

High
- Daemon socket: serve_start no longer accepts caller-chosen
  host/port. The socket is world-connectable (unprivileged CLI users),
  so callers could previously rebind the root web server to 0.0.0.0.
  The daemon now always binds the operator-configured address and
  reports it in the response; the CLI warns on mismatch.

Medium
- Remove the per-request max_delta_khz override from the API: the
  server-enforced safety cap is now authoritative. CLI direct paths
  (write, profile apply, verify) honor the configured cap; --max-delta
  still overrides for explicit root use.
- Snapshot restore: confine filepath to the snapshot directory
  (realpath containment; blocks symlink escapes).
- Login lockout: honor X-Forwarded-For only for peers listed in the
  new trusted_proxies config (rightmost untrusted hop), so the
  per-IP lockout works behind a reverse proxy. Spoofed headers from
  untrusted peers are ignored.
- /api/shutdown: new allow_api_shutdown config (default true);
  shared systems can disable the API shutdown path.

TLS (opt-in, like auth)
- New ssl_certfile/ssl_keyfile config + CLI flags (serve start,
  service install/configure, --no-ssl to disable). When active:
  HTTPS for UI/API, wss:// for WebSockets, Secure session cookie,
  CLI auto-switches to https://. Cert/key paths are validated up
  front with a clear error instead of a silent uvicorn crash.

Tests & docs
- tests/test_security.py: standalone regression tests (no new deps)
  covering SPA containment, snapshot containment, cap removal,
  client-IP derivation, proxy normalization, TLS scheme detection,
  and daemon host/port hardening.
- README + Usage-Guide: TLS section, new config keys, updated
  security notes.
2026-09-10 16:19:21 +02:00
ARIA bbd692ea2e feat: multi-user authentication (dual mode)
Add optional login protection for the web UI/API, intended for shared
machines (e.g. AI servers). Dual mode: with no users configured the API
and web UI are open (as before); once at least one user exists, every
/api/* and /ws/* endpoint requires a valid session.

- bcrypt password hashing: passwords stored as $2b$ hashes in
  /etc/nvcurve/users.json (0600, root-owned); plaintext never persisted.
- 24-hour sessions: HttpOnly cookie for browsers, Authorization: Bearer
  token for CLI/scripts; in-memory, invalidated on server restart.
- Multi-user: multiple named accounts (no shared-password mode).
- New CLI: nvcurve user add|list|remove|set-password (root for mutating
  ops; password always prompted, never a CLI argument).
- New endpoints: GET /api/ping (public), /api/auth/status|login|logout|users.
- Web UI: sign-in screen when auth is enabled; status bar shows the
  signed-in user with sign-out; expired sessions (401) re-show sign-in.
- Brute-force lockout: 10 failed logins/IP within 5 min -> 15 min lockout.
- New dependency: bcrypt.

Also: LSP config (pyrightconfig.json) pointing at the project .venv, and
small error-handling cleanups in daemon.py/server.py.
2026-09-02 15:21:35 +02:00
ARIA 706df90bc3 Fix doc links: hyphenate filenames for clickable markdown links 2026-05-09 15:23:12 +02:00