vasilito
1f58a3738c
Fourth-round integrations of the driver-manager migration's D-phase. Three new modules in driver-manager: heartbeat, aer, plus a SharedBlacklist wrapper around the existing Blacklist that supports live reload. heartbeat.rs: - Heartbeat struct with a publishing thread that writes a JSON status line to /var/run/driver-manager.heartbeat.json every 5s - Tracks bound / deferred / spawned / unbound / on_error counters - 3 unit tests cover counter updates, JSON output, and clone semantics aer.rs: - AER (PCI Express Advanced Error Reporting) listener thread - Reads /scheme/acpi/aer for events (parses severity + device bdf) - Routes to bound driver via scheme:driver-manager lookup - Returns RecoveryAction (Handled / ResetDevice / RescanBus) based on severity - Falls back to log-and-no-op when /scheme/acpi is absent - 7 unit tests cover parsing, routing, and severity mapping policy.rs: - Adds SharedBlacklist = Arc<RwLock<Blacklist>> + source_path - Supports live reload via replace() (re-reads from source_path) - is_blacklisted() takes a read lock (lock-free for the probe hot path) - set_global_shared_blacklist in config.rs wires it into the manager - 2 new unit tests cover replace() and snapshot isolation main.rs: - Spawns the heartbeat thread at startup (file at /var/run) - Constructs the SharedBlacklist from the policy directory config.rs: - Replaces GLOBAL_BLACKLIST OnceLock<Blacklist> with OnceLock<SharedBlacklist>; the probe hot path reads via Arc<RwLock>, not the OnceLock directly Docs: - DRIVER-MANAGER-MIGRATION-PLAN.md v1.6 status table (64 tests) - D5-AUDIT.md v1.6 update - HARDWARE-VALIDATION-MATRIX.md adds heartbeat and AER rows - AGENTS.md + docs/README.md pointers to v1.6 Test totals: 64 tests across 4 crates, all passing. § 0.5 audit-no-stubs.py: 0 violations across 37 files. SIGHUP trampoline for the SharedBlacklist is left for a future round; the replace() infrastructure is in place today so an operator can add the signal handler without changing the policy module.
74 lines
6.1 KiB
Markdown
74 lines
6.1 KiB
Markdown
# Hardware Validation Matrix — Red Bear OS
|
|
|
|
| Component | QEMU | Bare Metal | Notes |
|
|
|-----------|------|------------|-------|
|
|
| **Storage** | | | |
|
|
| AHCI SATA | ✅ | 🔲 | NCQ structure present, not runtime-validated |
|
|
| NVMe | 🔲 | 🔲 | Basic driver, no multiqueue |
|
|
| virtio-blk | ✅ | N/A | QEMU only |
|
|
| **Network** | | | |
|
|
| e1000 | 🔲 | 🔲 | ITR structure present, no offload validation |
|
|
| rtl8168 | 🔲 | 🔲 | PHY config present, not runtime-tested |
|
|
| virtio-net | ✅ | N/A | QEMU only |
|
|
| **Audio** | | | |
|
|
| Intel HDA | 🔲 | 🔲 | Codec detection + jack sense added, not validated |
|
|
| AC97 | 🔲 | 🔲 | Basic driver |
|
|
| **Input** | | | |
|
|
| PS/2 | ✅ | 🔲 | QEMU keyboard/mouse work |
|
|
| USB HID | ✅ | 🔲 | Keyboard/mouse autospawn via test-usb-qemu.sh (P1 hardening); not validated on real HW |
|
|
| **USB** | | | |
|
|
| xHCI host controller | ✅ | 🔲 | Interrupt-driven mode proven (test-xhci-irq-qemu.sh); 51-flag quirk table + HCCPARAMS2 gating (P2-A/B); 36-code error recovery (P2-C); not HW-validated |
|
|
| EHCI/UHCI/OHCI controllers | 🔲 | 🔲 | Compile clean; runtime enumeration proofs planned (P8-B) |
|
|
| USB storage (BOT) | ✅ | 🔲 | usbscsid autospawn + bounded sector readback proven (test-usb-storage-qemu.sh); guest-side write verify open |
|
|
| USB hub (non-root) | ✅ | 🔲 | Hub + hub-child enumeration proven in QEMU (test-usb-hub-qemu.sh): hub descriptor/EP1/scan/debounce/reset→enable, and usb-kbd behind hub enumerates to HID subdriver (class 3.1 keyboard); "reactor race" root-caused to startup latency under load (not a deadlock) |
|
|
| **GPU/Display** | | | |
|
|
| VESA | ✅ | 🔲 | QEMU framebuffer works |
|
|
| virtio-gpu | ✅ | N/A | 2D only, QEMU |
|
|
| **CPU/Power** | | | |
|
|
| cpufreqd | 🔲 | 🔲 | Governors implemented, not HW-validated |
|
|
| thermald | 🔲 | 🔲 | ACPI thermal zones, not HW-validated |
|
|
| **SMP** | | | |
|
|
| x2APIC/SMP | ✅ | ✅ | Multi-core works |
|
|
| **Driver Manager (driver-manager vs pcid-spawner)** | | | |
|
|
| D1 P0 capabilities (C1 dynids + C2 remove + C3 enable + C4 BAR) | ✅ | 🔲 | Code complete; 52 tests pass; see `local/docs/evidence/driver-manager/D5-AUDIT.md` |
|
|
| D1 P1 SpawnDecision + format coexistence + policy blacklist | ✅ | 🔲 | `spawn_decision_gate()` consults env/CLI/blacklist; `convert_legacy` reads legacy pcid.d format |
|
|
| D1 P1 PciQuirkFlags wired into spawn | ✅ | 🔲 | `driver-manager/src/quirks.rs` consumes `redox-driver-sys`; NEED_FIRMWARE defers probe; NO_MSIX/NO_MSI/FORCE_LEGACY_IRQ signal intx-fallback via env; DISABLE_ACCEL signals accel-disable |
|
|
| D1 P1 `/etc/driver-manager.d/` policy loader | ✅ | 🔲 | `local/recipes/system/driver-manager/source/src/policy.rs`; reads TOML blacklist; consulted at probe |
|
|
| D2 SMP concurrent probe (worker pool over `std::thread::scope`) | ✅ | 🔲 | `redox-driver-core/src/concurrent.rs` with `CountingSemaphore`; tested on host; opt-in via `--concurrent=N` |
|
|
| D2 PM hooks (Driver::suspend/resume + signal_then_collect) | ✅ | 🔲 | Real SIGTERM/SIGKILL in `DriverConfig`; needs iommu/pcid SetPowerState wiring for actual D3 |
|
|
| D2 AER foundation (Driver::on_error + ErrorSeverity + RecoveryAction) | ✅ | 🔲 | Trait API complete; needs pcid AER wake source |
|
|
| D2 C-state / P-state / IOMMU / MSI-X / NUMA helpers | ✅ Library | 🔲 HW | `redox-driver-core/src/modern_technology.rs` — concrete implementations; no driver-manager wrapper (orchestrator removed at v1.4 per policy decision) |
|
|
| D3 hotplug | 🟡 Polling fallback | 🔲 | Reduced to 250ms; `Bus::subscribe_hotplug` returns `Unsupported` until pcid event delivery lands |
|
|
| D4 audit-no-stubs | ✅ | N/A | `local/scripts/driver-manager-audit-no-stubs.py` scans 34 files; **0 violations** at v1.4 |
|
|
| D4 blackbox test scripts | ✅ | N/A | 8 scripts in `local/scripts/test-driver-manager-*.sh` |
|
|
| D4 heartbeat publisher | ✅ | N/A | `driver-manager/src/heartbeat.rs`; writes JSON to `/var/run/driver-manager.heartbeat.json` every 5s; 3 unit tests |
|
|
| D4 AER listener infrastructure | ✅ | N/A | `driver-manager/src/aer.rs`; polls `/scheme/acpi/aer`; 7 unit tests; falls back to log-and-no-op when scheme is absent |
|
|
| C0 dormant service files | ✅ | N/A | Committed in `local/sources/base` submodule (commit `0407d9cc`); both `ConditionPathExists`-gated |
|
|
| D5 feature-complete gate | ✅ Code | 🔲 HW | All code + tests pass; awaiting real-hardware validation on AMD Threadripper + Intel Alder Lake |
|
|
|
|
✅ = validated 🔲 = implemented, not validated N/A = not applicable
|
|
|
|
### Status note (2026-07-20)
|
|
|
|
- The login-based QEMU proofs (PS/2 + serio, monotonic timer, IOMMU first-use, USB storage BOT)
|
|
now run through `local/scripts/qemu-login-expect.py` (stdlib python) instead of the host
|
|
`expect` tool — no external dependency, per-step timeouts, and safe retry on premature QEMU
|
|
death. See `local/docs/IRQ-AND-LOWLEVEL-CONTROLLERS-ENHANCEMENT-PLAN.md` § "Runtime proof
|
|
status (2026-07-20)".
|
|
- MSI-X (virtio-net) and xHCI interrupt-driven mode were re-confirmed on the current
|
|
`redbear-mini` ISO — once more on 2026-07-20 ~09:10 JST via `test-lowlevel-controllers-qemu.sh`
|
|
(log-grep legs, unaffected by host contention).
|
|
- Re-confirmation of the login-based legs (PS/2, timer, BOT) with the new
|
|
harness is pending an uncontended host — the current build host runs a parallel QEMU-testing
|
|
workload (a concurrent agent session under `~/.claude/jobs/` with its own QMP-controlled
|
|
guest) that SIGKILLs competing QEMU processes; boots reach getty but the
|
|
login prompt is not observed within the timeout there. The ✅ marks above reflect prior
|
|
validation of those legs. A retry loop (`/tmp/opencode/a2-retry.sh`) is armed to run the
|
|
PS/2, timer, and USB-storage legs automatically once the foreign QEMU workload exits.
|
|
- The IOMMU first-use leg currently skips on `redbear-mini` runs because
|
|
`build/x86_64/redbear-full/harddrive.img` is absent — a `redbear-full` build is required
|
|
to re-confirm that leg.
|
|
- Display boots (`-vga std`) can freeze at the fbcond `Performing handoff` step (the
|
|
`reopen_for_handoff()` path can block rather than return `EAGAIN` for retry). Headless
|
|
(`-vga none`) boots avoid it. Tracked separately on the graphics stack.
|