Commit Graph
2 Commits
Author SHA1 Message Date
rajamesandClaude Opus 5.5 027b144ec7 chore(lfs): the logs and DoE CSVs committed as plain files go into LFS
.gitattributes has routed logs/**/*.log, *.log and the DoE run CSVs
through LFS since the initial commit, and 3944 of them are pointers.
1003 were committed as plain files instead (from a checkout without
git-lfs), so every fresh clone showed them as modified. git add
--renormalize moves them to LFS like the rest. Content unchanged: each
pointer's oid is the sha256 of the file it replaces.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GE3SDvQn5EoxdGufk84zz2
2026-11-06 00:43:48 +00:00
Robert Allan JamesandClaude Sonnet 5 024f8a84b5 starkernel: Phase 0 acceptance -- three-arch boot + reproducibility (item 0.10)
Adds a one-time boot diagnostic in kernel_main.c, right before sk_repl()
is entered: bounded wait for 3 real heartbeat ticks, then prints tick
count, TIME-TRUST, and variance. Needed because printing immediately
after apic_timer_start() (as first tried) measured 1 tick on amd64 and
0 on riscv64 -- not evidence the heartbeat doesn't work, just that
almost no wall time elapses between arming the timer and that point in
boot; report it honestly rather than let it stand as a false negative.

Verified this session (logs/20260804-001727, -001805, -001850,
-001948, -002021):
- All three architectures boot to ok>.
- Tick count non-zero: amd64 4, riscv64 3, aarch64 3.
- riscv64: trust=Q48_ONE exactly, variance=0 -- architecturally
  invariant counter, as designed.
- amd64: dict_hash=0x3d4e1daf289da94f, identical to the pre-item-0.8
  baseline (logs/20260803-231322) -- unchanged output, satisfying the
  GAP-A1 control.
- Two consecutive amd64 boots produced the identical dict hash --
  reproducible, no wall-clock leakage into patron state.

Phase 0 (Substrate) is complete.

Punch list §25 item 0.10 complete.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-04 00:21:23 -04:00