Repository navigation
recovery: contribute real-device protocol fixes and guarded boot retry - #48
Draft
seanabrahams wants to merge 3 commits into
Draft
seanabrahams wants to merge 3 commits into
seanabrahams wants to merge 3 commits into
Conversation
5 tasks
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem and resulting behavior
The native v0.1.12 recovery path stopped on a MacBookPro14,2 that already lacked usable
EMBEDDEDOSdata. This draft contributes the C/Rust changes used to complete both restores, final boot, EFI publication and sensor-matched import, plus regression tests and the redacted hardware report.Related to #43 and the overlapping protocol observations in #46. This does not close #43 or promote recovery beyond experimental support. The successful hardware run used local trust additions and retry procedures described below, rather than the prescribed unchanged signed-package flow.
Changes to review
FUDData, including a valid empty response when the pinned identity has no FUD components.FDRMemoryCommitsnapshots before ACK; select the final store only after successful restore status, requiring final provisioning/personalization agreement.gg.apple.com:80request over host-verified HTTPS to port 443. Keep device TLS tunnels opaque; retry only eligible connection failures before forwarding device request bytes, never trust failures or interrupted streams.EPIPEhandling forrsepfirmware, finalticket, and blindmemboot. The first two stalls were observed; memboot tolerance has synthetic coverage only. Restore status and stable functional USB remain mandatory.--resume-bootfor two completed, consistent protected restore phases. It performs no restore, reset or Apple request; it validates markers, stores, image/ticket pairing and device identity before the existing boot/publication/import gates. Existing EFI and already-verified boot still block retries.The commits separate the hardware-tested source, additional storage acceptance tests, and documentation. The report in
docs/recovery-macbookpro14-2-report.mdrecords test dates, package versions, deviations and unresolved findings.docs/linux-recovery.mdexplains the new command and limits.Validation
Fresh
make qualitypassed on October 8 (exit 0): formatting, warning-denied workspace/all-feature Clippy, 1,000 Rust tests passed, dependency policy, C tests and ASan/UBSan runs, kernel-module builds against7.2.5-4-omarchy, and packaging verification. Five Rust tests were explicitly ignored: four requiring an external pinned Apple package and one live-hardware preflight. This run used Rust/Cargo 1.98.1 and LLD 22.1.8 from signature-verified cached distro packages extracted into an isolated workspace, plus cargo-deny 0.20.2. No toolchain or module was installed. Documentation links, whitespace and privacy-pattern checks passed.Historical owner-attended hardware evidence (September 23–24): patched final boot exited 0; enrollment, match/non-match, Touch Bar, warm/cold persistence and password fallback passed. A later hibernation-associated fingerprint failure recovered after normal reboot, after which simultaneous 640x480 preview, fingerprint verification, Touch Bar and ALS passed. 720p capture/decode succeeded but preview usability failed before and after reboot. These remain separate lifecycle/application issues relevant to #18 and #20.
Four storage cases previously passed on disposable FAT32 with a clean post-unmount check. That is not power-loss evidence. No live firmware recovery, reset, enrollment change or host reboot was performed while preparing this PR.
Explicit review limits
Please review the shared protocol blockers against #46 first. The draft preserves the tested changes for inspection; it is not a request to merge the broader retry/trust behavior without review.
— GPT on behalf of @seanabrahams