_k_r2_001c1_gate.nx

_k_r2_001c1_gate.nx -- the K-R2-001c1 gate (virtio-NET MMIO transport HANDSHAKE, the SECOND virtio device). Clones the K-R2-001a blk handshake gate but drives the net (DeviceID=1) device block carved into the sovereign sim at base 0x10002000. NO mocks: runs the REAL nx_virtio_hs_emit (the SAME emitter the blk b-series uses -- data-only on the net spec's base/device-id/banner, so it stays byte-for-byte backward-compatible), then RUNS the emitted image on the SOVEREIGN rv64 emulator (rv64im_min_sim + the second rv64im_min_virtio net block at 0x10002000 -- the PRIMARY, gating lane: Nishi owns the runtime) and asserts the captured serial transcript CONTAINS the emitter's golden ("VNET ACK DRV FEAT OK\n" -- the driver verified magic/version/device-id==1, drove Status through ACK/DRIVER/FEATURES_OK/DRIVER_OK, read HostFeatures, wrote GuestFeatures, re-read Status to confirm FEATURES_OK stuck) AND the sovereign emu reports a clean SiFive-finisher halt. The blk device at 0x10001000 is left BYTE-UNTOUCHED (its own gate, _k_r2_001a, still passes -- proven by the regression re-run alongside this gate). Then the ALIGNMENT lane: qemu-system-riscv64 -machine virt -global virtio-mmio.force-legacy=true with a REAL virtio-blk-device (first) + a REAL virtio-net-device (second). qemu fills its virtio-mmio slots in REVERSE, so the first device (blk) lands at the highest slot 0x10008000 and the SECOND device (net) lands at the next slot DOWN, 0x10007000. The gate authors a SECOND net image from a derived spec whose base = the qemu net slot 0x10007000 and runs THAT on qemu; the SAME handshake driver completes against the REAL legacy virtio-net transport and emits the SAME golden transcript -> lanes AGREE on the handshake (alignment = handshake-completes-on-real-hw, transcript- identical against a real qemu virtio-net-device). Finally a TAMPER test: corrupt the net device-id-EXPECTED constant in the sovereign image -> the driver's device-id verify fails (actual=1 != corrupted-expected) -> it branches PAST the whole handshake straight to the finisher -> the transcript loses its golden -> the gate MUST go RED (proves the gate bites, not a rubber stamp). The device-id-expected li is encoded as lui t4,0 + addi t4,t4,1 over image words 11-12; its addi immediate high byte is image byte 51 -- bumping it changes the expected away from 1 (verified: byte51 0x00->0x01 -> expected 17). Evidence -> knowledge/status/virtio_net.log (NETHSGATE row; the queue row's ||MARK= reads it). Sovereign orchestration (fork/dup3/execve/wait4). license_tier: ORIGINAL