Read the chip's IrqMask-suppressed count from native radio stats

[NomDeTom]

What this changes, and why

• firmware.RadioStats gains IRQSuppressed and SuppressedReported, read from bytes 43-46 of a radio stats record (kindRadioStats) that is at least 47 bytes long. A shorter record leaves SuppressedReported false.
• node.stats answers irq_suppressed for each node: the count, or null where the node's chip did not report one.
• The Python client's NodeStat gains irq_suppressed: int | None.

The count is the number of IRQ events the chip raised that the firmware's IrqMask kept out of the status register. virtual-sx1262 keeps it from MeshBench/virtual-sx1262#2 onward (irq_suppressed in vsx_counters). The native bridge sends it in MeshBench/meshcore-native#26. Until a node sends a 47-byte record, the field is null.

Without it, a node whose firmware did not enable an interrupt and a node whose chip never raised that interrupt report the same irq_reads, busy_reads and busy_ms.

Manual: MeshBench/docs#40 (regenerated Python and control references).

Verification

• gofmt -l .: empty.
• go vet on internal/firmware, internal/app/session, internal/app/state: clean.
• go test on internal/firmware, internal/app/state, internal/app/session, internal (layers and layout map): pass. The full go test ./... was not run.
• golangci-lint v2.12.2 on the same packages: no findings in changed lines. The pre-existing findings in build.go, native.go, prefs.go and bridge.go (net.Listen) are unchanged.
• tools/verbdoc/verbdoc.py --check: current.
• New tests: internal/firmware/radiostats_test.go (a 47-byte record sets the count and SuppressedReported; a 43-byte record does not). Python: test_a_suppressed_count_is_none_unless_the_chip_kept_one, and an assertion in test_node_stats_carry_the_rows that a stopped node answers None.
• Python client: ruff check and ruff format --check are clean. pytest tests/ against a binary built from this branch: 51 passed, 3 failed (test_pytest_plugin.py: test_the_meshbench_fixture_is_discovered, test_the_scenario_is_cleared_between_tests, test_a_failure_carries_the_provenance). The same three fail with this change stashed; they need the client installed as a package, and it was run from source.
• End to end: native MeshCore 1.16.0 and 1.17.0 builds with bridge: report what the chip's IrqMask kept out meshcore-native#26 against virtual-sx1262 8fb9623, in headless runs on fife-strict (three nodes) and scotland-strict. The parsed counts reached the experiment export through the follow-up PR (Record each node's radio counters in every experiment cell #746): 1.16 nodes (IrqMask 0x0272) reported 24 and 1.17 nodes (0x0276) 12 on the three-node link.

Checklist

• One independent change. If it is two, it is two pull requests.
• A new package updates the layout map in CLAUDE.md in this commit. (No new package.)
• A new dependency is justified here, in one line. (None.)
• Files under 500 lines; new panels and widgets are one type per file.
• A new panel joins auditTargets; a new state is reachable by flag. (No panel.)
• docs/shortcomings.md still tells the truth after this change.
• Every claim here is verified, not assumed — no invented verb, flag, path or result.
• A behaviour change is covered by a test that fails without it.
• Nothing staged that was not written or asked for — no build artifacts, nothing under security-audit/.

[A13xB0]

keeping an eye

[A13xB0]

apologies for pipeline breaking... I'll fix this tomorrow at some point, its just missing tih

      - name: Headers cgo needs
        run: |
          sudo apt-get update
          sudo apt-get install -y --no-install-recommends \
            gcc pkg-config libgl1-mesa-dev xorg-dev \
            libwayland-dev libxkbcommon-dev libxkbcommon-x11-dev \
            libx11-xcb-dev libegl1-mesa-dev libvulkan-dev libffi-dev

feel free to add it if you wish to this PR or I can do it in a seperate one tomorrow