DF-2316 baseline kernel build log (reconstructed summary)
=========================================================

The full untrimmed compiler output of the identical build procedure (same
rig, same KERNCONF, same make invocation) is in fix_build.log (37,530 lines,
ends ">>> Kernel install for PNPH completed ... BUILD_KERNEL_OK"); the
baseline build differs from it ONLY by the content of pnpparse.c (pristine
4b13638c... vs patched fde2c7e1...).  The baseline /tmp/buildN.log files were
lost when the guest was reset with-src after the panic; the exact commands
and their results were:

  # cd /usr/src && make -j6 nativekernel KERNCONF=PNPH
  # make installkernel KERNCONF=PNPH

  build attempt 1 (16:44): FAILED - make: don't know how to make
      bus/isa/pnp_harness.c (prep.sh had not installed the harness TU yet)
  build attempt 2 (16:47): FAILED - pnp_harness.c:130: error: "/*" within
      comment [-Werror=comment] (rig comment bug, fixed)
  build attempt 3 (16:49): OK
      ">>> Kernel build for PNPH completed on Sun Oct 18 16:53:15 UTC 2026"
      ">>> Kernel install for PNPH completed on Sun Oct 18 16:54:48 UTC 2026"
      (this boot validated prep + scenarios; superseded by rig fixes)
  build attempt 4 (17:03): OK - IRQ_FORMAT sweep tag byte corrected (0x10 -> 0x20)
      ">>> Kernel install for PNPH completed on Sun Oct 18 17:03:57 UTC 2026"
  build attempt 5 (17:09): OK - wf_stream info bytes corrected
      kernel "#2: Sun Oct 18 17:09:41 UTC 2026"  <-- decisive baseline kernel
      for scenarios 9,20,21,22,23,24,26 (see run.log PART 1/2)
  build attempt 6 (17:20): OK - page-boundary rig switched from pmap_kremove
      to kmem_free of the tail pages (see run_pagebound NOTE)
      kernel used for scenario 25 -> Fatal trap 12 panic (panic.txt)

All six attempts compiled the VERBATIM pnpparse.c (sha gate in prep.sh);
only the rig file pnp_harness.c changed between attempts 3-6, never the
audited code.