Artemis genesis stamp: fix a BAM-corrupting offset before it ever ran (FABRIC-3.md §XXVI follow-on, Step 3)

Step 3: one-time artemis_sig_t genesis stamp, written once
kernel_main.c's virtio-blk path confirms Artemis's own disk, so the disk
image is later recognizable generically (repl.c's idle-loop USB-MSC scan,
built in the prior commit) regardless of which bus found it.

Correction made before this ever touched the real disk: the signature's
first design (committed in 29b6789) placed it at a fixed bottom-of-device
forth-block (4, devblock 1) -- copying homeblocks_sig_t's own convention,
which is safe for a raw identity thumbdrive but not for Artemis's own
disk. Artemis's disk is block_subsystem.c's own STFR/v2-formatted volume:
devblock 0 holds that format's header and devblock 1 is the FIRST
DEVBLOCK OF THE LIVE BAM (blk_compute_fresh_geometry(): bam_start = 1).
The original design would have overwritten Artemis's live allocation map
on the very first real boot. Caught via direct cross-reference against
block_subsystem.c before the genesis-stamp call site was ever run against
the real image -- no corruption occurred.

Fixed by moving the header to a fixed offset from the END of the device
instead (ARTEMIS_SIG_DEVBLOCK_FROM_TOP=64), the same top-of-device region
block_subsystem.c's own meta_fence_blocks reservation (128 devblocks)
already carves out for system metadata, and where Zuse's genesis marker/
eligibility list already live -- but computed independently via
blkio_info() rather than through blk_meta_zone_*(), since that accessor
needs an already-attached, format-detected slot, which is exactly the
state pre-attach generic discovery doesn't have yet. Picked well clear of
Zuse's two tenants (devblock_from_top 0 and 1+, open-ended) so the two
subsystems' independent math can never collide.

Also reordered kernel_main.c: rng_init() now runs before the Artemis
virtio-blk block (was after) -- the genesis stamp needs rng_get_bytes()
for disk_uuid, and the original order would have failed the stamp on
every boot.

Verified live: booted amd64 against the real disk/artemis.img twice --
first boot logs "Artemis: genesis signature stamped" (confirmed blank at
the target offset beforehand via a host-side read), second boot on the
now-stamped image logs no re-stamp (idempotent, CRC/read-back verified)
-- both boots and aarch64/riscv64 (against the same now-stamped image)
all still report "Artemis: 22998 data blocks" / "PASS: persist-read"
unchanged, confirming the BAM and data pool were never touched.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EXieurDfDSsDFdnSyusuWo

src/starkernel/kernel_main.c

@@ -63,6 +63,7 @@ EFI_RUNTIME_SERVICES *g_sk_runtime_services = NULL;
 #include "starkernel/capsule_loader.h"
 #include "starkernel/capsule_birth.h"  /* capsule_birth_mama, capsule_find_mama_init */
 #include "starkernel/capsule_zuse_boot.h" /* capsule_zuse_boot_load_root_pubkey */
+#include "starkernel/artemis_sig.h" /* artemis_sig_check/genesis_stamp */
 #include "starkernel/kmalloc.h"
 #include "starkernel/repl.h"
 #include "starkernel/pci.h"
@@ -589,6 +590,30 @@ static void kernel_main_deep(BootInfo *boot_info) {
     console_println("PCI: init...");
     pci_init(boot_info->acpi_table);
 
+    /* Phase 8: entropy. Real per-arch RNG doesn't cover all three
+     * architectures (amd64 RDRAND, riscv64 Zkr, but aarch64 has neither in
+     * QEMU's CPU models -- see vm_uuid.h's identical finding), so signing/
+     * keygen entropy comes from the unified rng_get_bytes() layer, whose
+     * v2.0.0 backend is the paravirtualized virtio-rng device. Unconditional
+     * call site, same graceful-noop precedent as virtio_blk_find_artemis()
+     * below -- boot proceeds either way, the device is only required once
+     * something actually calls rng_get_bytes().
+     *
+     * FABRIC-3.md §XXVI follow-on, 2026-09-13: moved ahead of the Artemis
+     * virtio-blk block below (was after it) -- artemis_sig_genesis_stamp()
+     * needs rng_get_bytes() for disk_uuid, and calling it before rng_init()
+     * ran would have failed the stamp on every single boot forever. Both
+     * calls only need pci_init() above; this reordering has no other
+     * dependency either way. */
+    {
+        int rrc = rng_init();
+        if (rrc == 0) {
+            console_println("entropy: ready");
+        } else {
+            console_println("entropy: not available (continuing without)");
+        }
+    }
+
     {
         static blkio_dev_t artemis_dev;
         int vrc = virtio_blk_find_artemis(&artemis_dev);
@@ -603,28 +628,40 @@ static void kernel_main_deep(BootInfo *boot_info) {
              * capsule_zuse_boot_load_root_pubkey()'s own doc comment for
              * why this is safe and separate from her live-session cert. */
             capsule_zuse_boot_load_root_pubkey((VM *)mama_vm);
+
+            /* FABRIC-3.md §XXVI follow-on, 2026-09-13: one-time
+             * artemis_sig_t genesis stamp, so this exact disk image can
+             * later be recognized generically (by content, not by which
+             * bus/vendor-ID scan happened to find it -- see repl.c's own
+             * idle-loop USB-MSC discovery, the reason this signature
+             * format exists at all). Safe to attempt unconditionally
+             * every boot: virtio_blk_find_artemis() only ever succeeds
+             * against the one dedicated PCI device, so a BLANK read here
+             * unambiguously means "never stamped," not "might be some
+             * other blank drive" -- and artemis_sig_check() returning
+             * anything other than BLANK (already stamped, or a version/
+             * CRC mismatch worth leaving alone rather than overwriting)
+             * skips the stamp. See artemis_sig.h's own doc comment for why
+             * this lands at a fence-relative top-of-device offset now,
+             * not a fixed bottom-of-device forth-block (that first attempt
+             * would have overwritten Artemis's own live BAM -- caught
+             * before ever being run against the real disk). */
+            {
+                artemis_sig_t asig;
+                artemis_sig_result_t art_rc = artemis_sig_check(&artemis_dev, &asig);
+                if (art_rc == ARTEMIS_SIG_BLANK) {
+                    if (artemis_sig_genesis_stamp(&artemis_dev) == 0) {
+                        console_println("Artemis: genesis signature stamped");
+                    } else {
+                        console_println("Artemis: genesis signature stamp FAILED");
+                    }
+                }
+            }
         } else {
             console_println("Artemis: no virtio-blk disk (continuing without)");
         }
     }
 
-    /* Phase 8: entropy. Real per-arch RNG doesn't cover all three
-     * architectures (amd64 RDRAND, riscv64 Zkr, but aarch64 has neither in
-     * QEMU's CPU models -- see vm_uuid.h's identical finding), so signing/
-     * keygen entropy comes from the unified rng_get_bytes() layer, whose
-     * v2.0.0 backend is the paravirtualized virtio-rng device. Unconditional
-     * call site, same graceful-noop precedent as virtio_blk_find_artemis()
-     * above -- boot proceeds either way, the device is only required once
-     * something actually calls rng_get_bytes(). */
-    {
-        int rrc = rng_init();
-        if (rrc == 0) {
-            console_println("entropy: ready");
-        } else {
-            console_println("entropy: not available (continuing without)");
-        }
-    }
-
     /* Zuse identity: SUPERSEDED 2026-08-28 (FABRIC-2.md §F.20/§F.21).
      * The one-shot block-fence mint-or-load that used to run here is
      * gone -- Zuse is thumbdrive-resident now (her seed never touches