# Fix-kernel builds (in-guest, /usr/src, KERNCONF=X86_64_GENERIC, -j6)
#
# Build 1 (MANGLED patch — hunk header was wrong, patch(1) dropped the
#         added `return;`, leaving `if (cursor->data == NULL)` guarding
#         the nlinks test instead of returning; documented in VERDICT.md):
#   $ cd /usr/src && patch -p1 < fix.diff
#   Patching file sys/vfs/hammer/hammer_prune.c using Plan A...
#   Hunk #1 succeeded at 326 (offset -1 lines).
#   Hmm...  Ignoring the trailing garbage.
#   done
#   $ make -j6 nativekernel KERNCONF=X86_64_GENERIC
#   BUILD_RC=0
#   --------------------------------------------------------------
#   >>> Kernel build for X86_64_GENERIC completed on Tue Sep  8 18:27:08 UTC 2026
#   --------------------------------------------------------------
#   $ make installkernel KERNCONF=X86_64_GENERIC
#   INST_RC=0
#   >>> Kernel install for X86_64_GENERIC completed on Tue Sep  8 18:28:20 UTC 2026
#   -> kernel #1 Sep  8 18:24:10 (still crashed with the folded absolute
#      load `movq 0x48,%rax' — PoC artifact, not a fix failure)
#
# Build 2 (CORRECT patch — pristine hammer_prune.c restored, corrected
#         git-apply-able fix.diff applied: "Hunk #1 succeeded at 329",
#         no warnings):
#   $ cp /home/maxx/hammer_prune.c /usr/src/sys/vfs/hammer/hammer_prune.c
#   $ cd /usr/src && patch -p1 < fix.diff
#   Patching file sys/vfs/hammer/hammer_prune.c using Plan A...
#   Hunk #1 succeeded at 329.
#   done
#   $ make -j6 nativekernel KERNCONF=X86_64_GENERIC
#   BUILD_RC=0
#   --------------------------------------------------------------
#   >>> Kernel build for X86_64_GENERIC completed on Tue Sep  8 18:46:08 UTC 2026
#   --------------------------------------------------------------
#   $ make installkernel KERNCONF=X86_64_GENERIC
#   INST_RC=0
#   -> kernel #2 Sep  8 18:43:09  (PoC re-run: clean, see fixrun.log)