build: add rpm + gnupg (Fedora RPM GPG verification) by HarryR · Pull Request #10 · lockboot/stage0

HarryR · 2026-07-10T07:54:47Z

What

Bring the stage0 build image and disk layout in line with the Fedora 44 chain bump, plus stable disk identity.

Changes

rpm + gnupg added to Dockerfile.build so the UKI build can verify Fedora package GPG signatures (image parity with stage1, kept byte-identical). No effect on compiled artifacts.
Stable lockboot GPT GUIDs + FAT volume-id.DISK_GUID / PART_GUID are now constant across releases (4c4f434b-424f-4f54-..., "LOCKBOOT" + a role index), so the initial GPT is byte-stable (stable pre-stage2 PCR5) and the boot disk is identifiable by a fixed GUID rather than by hashing stage0.efi (which changed per release). This is the stable identity stage2's find_boot_device can key on. (stage2 still rewrites the GPT at runtime to add p2/p3, which is why it excludes PCR5 from its key binding.)
Pin vaportpm-attest to v0.3.0 (rev + version guard), matching stage1.
Extract the shared DOCKER_RUN harness into build.mk (same refactor as stage1).

Verification

Reproducible: GUIDs/timestamps fixed; build output byte-stable.
No change to the runtime boot path. Pairs with uki: Fedora 44 kernel 7.0.12 + stub 259.6; GPG-verify + automate deps stage1#19 (Fedora 44 kernel/stub) and stage2: rework /data persistence (XTS, meta_bg) + mkruntime tooling stage2#1.

🤖 Generated with Claude Code

Needed so the UKI build can verify Fedora package GPG signatures (rpm/rpmkeys + gnupg). Image parity: Dockerfile.build is kept byte-identical with stage1, whose tools/build-uki now GPG-verifies the kernel + systemd-boot RPMs it downloads. No effect on compiled artifacts. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Move the docker-images + DOCKER_RUN plumbing block out of the Makefile into a canonical build.mk that the Makefile now includes. This is the source of truth for the shared harness, vendored byte-identically into stage1/vaportpm via the workspace make sync-harness and guarded by make check-harness. Pure relocation -- make -n expands identically; docker-build-harness (stage0-only) stays in the Makefile. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Both stage0 crates floated on vaportpm main (no rev), and stage0's Cargo.lock is untracked, so vaportpm-attest resolved to main HEAD on every build -- no version stability for a measured bootloader. Pin to the v0.3.0 release commit (7041f461) with a version = "=0.3.0" guard, keeping default-features = false (no_std core: Tpm, PcrOps, TpmTransport). Verified: stage0.efi and payload.efi build clean for x86_64-unknown-uefi against v0.3.0. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…entifiable) Replace the stage0.efi-hash-derived DISK_GUID/PART_GUID/VOLUME_ID with fixed constants ("4c4f434b424f4f54" spells "LOCKBOOT"; last group a role index: disk = 0, ESP = partition 1). Constant across releases so the GPT -- and hence the firmware's PCR 5 measurement -- is byte-stable, and a lockboot disk is identifiable by GUID. Dependency of stage2, which locates the boot disk by GUID and rewrites the GPT to add its runtime/data partitions; a stable base GPT keeps PCR 5 predictable. The stage0.efi hash still feeds BUILD_ID. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…#19) ## What Bring the UKI to a current, GPG-verified Fedora 44 base, bundle the kernel modules stage2 needs, and supporting build cleanups. ## Kernel + stub - Bump kernel `6.12.4-200.fc41` -> `7.0.12-201.fc44` and systemd-boot stub `256.17-1.fc41` -> `259.6-1.fc44` (what Fedora CoreOS 44.20260621.3.1 ships). `fc41` is EOL; the kernel was ~18 months behind. - **GPG verification (mandatory):** `download-verify-rpm.sh` fetches from koji's *signed* path (plain `packages/` is unsigned) and verifies each RPM against `keys/RPM-GPG-KEY-fedora-44-primary` (fp `36F6 12DC F27F 7D1A 48A8 35E4 DBFC F71C 6D9F 90A6`) on every build. - **Automated bumps:** pins live in generated `tools/build-uki/fedora-deps.mk`; `make update-fedora-deps FCOS=... SYSTEMD=...` reads the FCOS manifest, downloads + GPG-verifies the signed RPMs, and rewrites the pins. ## Storage modules for stage2 - Bundle + explicitly load (in dependency order, before the `modules_disabled` latch) the modules stage2's loader needs: **dm-crypt** (encrypted /data), **dm-verity** + **reed_solomon** (integrity-checked erofs runtime + FEC dep), **overlay** (ephemeral root), **erofs** + **netfs** (RO image fs + dep), and the **nvme** chain (hkdf -> nvme-auth/keyring -> nvme-core -> nvme; EC2 EBS + the harness disk are NVMe). The payload cannot load modules itself, and an unresolved dep can't be pulled once the latch is set, so deps are listed explicitly. - Module comments describe each dependency's capability generically. **Authenticated /data (dm-integrity + async_xor/async_tx) is documented as an extension point but not shipped** -- stage2 uses confidentiality-only dm-crypt, so those are omitted to keep the measured initramfs minimal. ## Build - Pin `vaportpm-attest` to v0.3.0 (rev + version guard). - Extract the shared `DOCKER_RUN` harness into `build.mk`. - Drop a stale `kernel-hash-%` target. ## Verification - ena (AWS) + gve (GCP) present in 7.0.12; all required `.ko.xz` copy. - `make test-chain-x86_64 SIGN=1 / MANIFEST=1 / SIGN_ARGS=1` boot `Linux 7.0.12-201.fc44` end to end and run stage2. - lockboot/stage2's XTS pivot test passes end to end on this UKI (`PIVOT_TEST=PASS`), dm-integrity absent from the boot. - aarch64 pins are GPG-verified but not boot-tested here. Pairs with lockboot/stage0#10 (image parity: rpm+gnupg, shared harness, vaportpm pin, stable GUIDs). 🤖 Generated with [Claude Code](https://claude.com/claude-code)

HarryR · 2026-07-10T07:54:47Z

What

Bring the stage0 build image and disk layout in line with the Fedora 44 chain bump, plus stable disk identity.

Changes

rpm + gnupg added to Dockerfile.build so the UKI build can verify Fedora package GPG signatures (image parity with stage1, kept byte-identical). No effect on compiled artifacts.
Stable lockboot GPT GUIDs + FAT volume-id.DISK_GUID / PART_GUID are now constant across releases (4c4f434b-424f-4f54-..., "LOCKBOOT" + a role index), so the initial GPT is byte-stable (stable pre-stage2 PCR5) and the boot disk is identifiable by a fixed GUID rather than by hashing stage0.efi (which changed per release). This is the stable identity stage2's find_boot_device can key on. (stage2 still rewrites the GPT at runtime to add p2/p3, which is why it excludes PCR5 from its key binding.)
Pin vaportpm-attest to v0.3.0 (rev + version guard), matching stage1.
Extract the shared DOCKER_RUN harness into build.mk (same refactor as stage1).

Verification

Reproducible: GUIDs/timestamps fixed; build output byte-stable.
No change to the runtime boot path. Pairs with uki: Fedora 44 kernel 7.0.12 + stub 259.6; GPG-verify + automate deps stage1#19 (Fedora 44 kernel/stub) and stage2: rework /data persistence (XTS, meta_bg) + mkruntime tooling stage2#1.

🤖 Generated with Claude Code

Needed so the UKI build can verify Fedora package GPG signatures (rpm/rpmkeys + gnupg). Image parity: Dockerfile.build is kept byte-identical with stage1, whose tools/build-uki now GPG-verifies the kernel + systemd-boot RPMs it downloads. No effect on compiled artifacts. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Move the docker-images + DOCKER_RUN plumbing block out of the Makefile into a canonical build.mk that the Makefile now includes. This is the source of truth for the shared harness, vendored byte-identically into stage1/vaportpm via the workspace make sync-harness and guarded by make check-harness. Pure relocation -- make -n expands identically; docker-build-harness (stage0-only) stays in the Makefile. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Both stage0 crates floated on vaportpm main (no rev), and stage0's Cargo.lock is untracked, so vaportpm-attest resolved to main HEAD on every build -- no version stability for a measured bootloader. Pin to the v0.3.0 release commit (7041f461) with a version = "=0.3.0" guard, keeping default-features = false (no_std core: Tpm, PcrOps, TpmTransport). Verified: stage0.efi and payload.efi build clean for x86_64-unknown-uefi against v0.3.0. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…entifiable) Replace the stage0.efi-hash-derived DISK_GUID/PART_GUID/VOLUME_ID with fixed constants ("4c4f434b424f4f54" spells "LOCKBOOT"; last group a role index: disk = 0, ESP = partition 1). Constant across releases so the GPT -- and hence the firmware's PCR 5 measurement -- is byte-stable, and a lockboot disk is identifiable by GUID. Dependency of stage2, which locates the boot disk by GUID and rewrites the GPT to add its runtime/data partitions; a stable base GPT keeps PCR 5 predictable. The stage0.efi hash still feeds BUILD_ID. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…#19) ## What Bring the UKI to a current, GPG-verified Fedora 44 base, bundle the kernel modules stage2 needs, and supporting build cleanups. ## Kernel + stub - Bump kernel `6.12.4-200.fc41` -> `7.0.12-201.fc44` and systemd-boot stub `256.17-1.fc41` -> `259.6-1.fc44` (what Fedora CoreOS 44.20260621.3.1 ships). `fc41` is EOL; the kernel was ~18 months behind. - **GPG verification (mandatory):** `download-verify-rpm.sh` fetches from koji's *signed* path (plain `packages/` is unsigned) and verifies each RPM against `keys/RPM-GPG-KEY-fedora-44-primary` (fp `36F6 12DC F27F 7D1A 48A8 35E4 DBFC F71C 6D9F 90A6`) on every build. - **Automated bumps:** pins live in generated `tools/build-uki/fedora-deps.mk`; `make update-fedora-deps FCOS=... SYSTEMD=...` reads the FCOS manifest, downloads + GPG-verifies the signed RPMs, and rewrites the pins. ## Storage modules for stage2 - Bundle + explicitly load (in dependency order, before the `modules_disabled` latch) the modules stage2's loader needs: **dm-crypt** (encrypted /data), **dm-verity** + **reed_solomon** (integrity-checked erofs runtime + FEC dep), **overlay** (ephemeral root), **erofs** + **netfs** (RO image fs + dep), and the **nvme** chain (hkdf -> nvme-auth/keyring -> nvme-core -> nvme; EC2 EBS + the harness disk are NVMe). The payload cannot load modules itself, and an unresolved dep can't be pulled once the latch is set, so deps are listed explicitly. - Module comments describe each dependency's capability generically. **Authenticated /data (dm-integrity + async_xor/async_tx) is documented as an extension point but not shipped** -- stage2 uses confidentiality-only dm-crypt, so those are omitted to keep the measured initramfs minimal. ## Build - Pin `vaportpm-attest` to v0.3.0 (rev + version guard). - Extract the shared `DOCKER_RUN` harness into `build.mk`. - Drop a stale `kernel-hash-%` target. ## Verification - ena (AWS) + gve (GCP) present in 7.0.12; all required `.ko.xz` copy. - `make test-chain-x86_64 SIGN=1 / MANIFEST=1 / SIGN_ARGS=1` boot `Linux 7.0.12-201.fc44` end to end and run stage2. - lockboot/stage2's XTS pivot test passes end to end on this UKI (`PIVOT_TEST=PASS`), dm-integrity absent from the boot. - aarch64 pins are GPG-verified but not boot-tested here. Pairs with lockboot/stage0#10 (image parity: rpm+gnupg, shared harness, vaportpm pin, stable GUIDs). 🤖 Generated with [Claude Code](https://claude.com/claude-code)

HarryR · 2026-07-10T07:54:47Z

What

Bring the stage0 build image and disk layout in line with the Fedora 44 chain bump, plus stable disk identity.

Changes

rpm + gnupg added to Dockerfile.build so the UKI build can verify Fedora package GPG signatures (image parity with stage1, kept byte-identical). No effect on compiled artifacts.
Stable lockboot GPT GUIDs + FAT volume-id.DISK_GUID / PART_GUID are now constant across releases (4c4f434b-424f-4f54-..., "LOCKBOOT" + a role index), so the initial GPT is byte-stable (stable pre-stage2 PCR5) and the boot disk is identifiable by a fixed GUID rather than by hashing stage0.efi (which changed per release). This is the stable identity stage2's find_boot_device can key on. (stage2 still rewrites the GPT at runtime to add p2/p3, which is why it excludes PCR5 from its key binding.)
Pin vaportpm-attest to v0.3.0 (rev + version guard), matching stage1.
Extract the shared DOCKER_RUN harness into build.mk (same refactor as stage1).

Verification

Reproducible: GUIDs/timestamps fixed; build output byte-stable.
No change to the runtime boot path. Pairs with uki: Fedora 44 kernel 7.0.12 + stub 259.6; GPG-verify + automate deps stage1#19 (Fedora 44 kernel/stub) and stage2: rework /data persistence (XTS, meta_bg) + mkruntime tooling stage2#1.

🤖 Generated with Claude Code

Needed so the UKI build can verify Fedora package GPG signatures (rpm/rpmkeys + gnupg). Image parity: Dockerfile.build is kept byte-identical with stage1, whose tools/build-uki now GPG-verifies the kernel + systemd-boot RPMs it downloads. No effect on compiled artifacts. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Move the docker-images + DOCKER_RUN plumbing block out of the Makefile into a canonical build.mk that the Makefile now includes. This is the source of truth for the shared harness, vendored byte-identically into stage1/vaportpm via the workspace make sync-harness and guarded by make check-harness. Pure relocation -- make -n expands identically; docker-build-harness (stage0-only) stays in the Makefile. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Both stage0 crates floated on vaportpm main (no rev), and stage0's Cargo.lock is untracked, so vaportpm-attest resolved to main HEAD on every build -- no version stability for a measured bootloader. Pin to the v0.3.0 release commit (7041f461) with a version = "=0.3.0" guard, keeping default-features = false (no_std core: Tpm, PcrOps, TpmTransport). Verified: stage0.efi and payload.efi build clean for x86_64-unknown-uefi against v0.3.0. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…entifiable) Replace the stage0.efi-hash-derived DISK_GUID/PART_GUID/VOLUME_ID with fixed constants ("4c4f434b424f4f54" spells "LOCKBOOT"; last group a role index: disk = 0, ESP = partition 1). Constant across releases so the GPT -- and hence the firmware's PCR 5 measurement -- is byte-stable, and a lockboot disk is identifiable by GUID. Dependency of stage2, which locates the boot disk by GUID and rewrites the GPT to add its runtime/data partitions; a stable base GPT keeps PCR 5 predictable. The stage0.efi hash still feeds BUILD_ID. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…#19) ## What Bring the UKI to a current, GPG-verified Fedora 44 base, bundle the kernel modules stage2 needs, and supporting build cleanups. ## Kernel + stub - Bump kernel `6.12.4-200.fc41` -> `7.0.12-201.fc44` and systemd-boot stub `256.17-1.fc41` -> `259.6-1.fc44` (what Fedora CoreOS 44.20260621.3.1 ships). `fc41` is EOL; the kernel was ~18 months behind. - **GPG verification (mandatory):** `download-verify-rpm.sh` fetches from koji's *signed* path (plain `packages/` is unsigned) and verifies each RPM against `keys/RPM-GPG-KEY-fedora-44-primary` (fp `36F6 12DC F27F 7D1A 48A8 35E4 DBFC F71C 6D9F 90A6`) on every build. - **Automated bumps:** pins live in generated `tools/build-uki/fedora-deps.mk`; `make update-fedora-deps FCOS=... SYSTEMD=...` reads the FCOS manifest, downloads + GPG-verifies the signed RPMs, and rewrites the pins. ## Storage modules for stage2 - Bundle + explicitly load (in dependency order, before the `modules_disabled` latch) the modules stage2's loader needs: **dm-crypt** (encrypted /data), **dm-verity** + **reed_solomon** (integrity-checked erofs runtime + FEC dep), **overlay** (ephemeral root), **erofs** + **netfs** (RO image fs + dep), and the **nvme** chain (hkdf -> nvme-auth/keyring -> nvme-core -> nvme; EC2 EBS + the harness disk are NVMe). The payload cannot load modules itself, and an unresolved dep can't be pulled once the latch is set, so deps are listed explicitly. - Module comments describe each dependency's capability generically. **Authenticated /data (dm-integrity + async_xor/async_tx) is documented as an extension point but not shipped** -- stage2 uses confidentiality-only dm-crypt, so those are omitted to keep the measured initramfs minimal. ## Build - Pin `vaportpm-attest` to v0.3.0 (rev + version guard). - Extract the shared `DOCKER_RUN` harness into `build.mk`. - Drop a stale `kernel-hash-%` target. ## Verification - ena (AWS) + gve (GCP) present in 7.0.12; all required `.ko.xz` copy. - `make test-chain-x86_64 SIGN=1 / MANIFEST=1 / SIGN_ARGS=1` boot `Linux 7.0.12-201.fc44` end to end and run stage2. - lockboot/stage2's XTS pivot test passes end to end on this UKI (`PIVOT_TEST=PASS`), dm-integrity absent from the boot. - aarch64 pins are GPG-verified but not boot-tested here. Pairs with lockboot/stage0#10 (image parity: rpm+gnupg, shared harness, vaportpm pin, stable GUIDs). 🤖 Generated with [Claude Code](https://claude.com/claude-code)

HarryR · 2026-07-10T07:54:47Z

What

Bring the stage0 build image and disk layout in line with the Fedora 44 chain bump, plus stable disk identity.

Changes

rpm + gnupg added to Dockerfile.build so the UKI build can verify Fedora package GPG signatures (image parity with stage1, kept byte-identical). No effect on compiled artifacts.
Stable lockboot GPT GUIDs + FAT volume-id.DISK_GUID / PART_GUID are now constant across releases (4c4f434b-424f-4f54-..., "LOCKBOOT" + a role index), so the initial GPT is byte-stable (stable pre-stage2 PCR5) and the boot disk is identifiable by a fixed GUID rather than by hashing stage0.efi (which changed per release). This is the stable identity stage2's find_boot_device can key on. (stage2 still rewrites the GPT at runtime to add p2/p3, which is why it excludes PCR5 from its key binding.)
Pin vaportpm-attest to v0.3.0 (rev + version guard), matching stage1.
Extract the shared DOCKER_RUN harness into build.mk (same refactor as stage1).

Verification

Reproducible: GUIDs/timestamps fixed; build output byte-stable.
No change to the runtime boot path. Pairs with uki: Fedora 44 kernel 7.0.12 + stub 259.6; GPG-verify + automate deps stage1#19 (Fedora 44 kernel/stub) and stage2: rework /data persistence (XTS, meta_bg) + mkruntime tooling stage2#1.

🤖 Generated with Claude Code

Needed so the UKI build can verify Fedora package GPG signatures (rpm/rpmkeys + gnupg). Image parity: Dockerfile.build is kept byte-identical with stage1, whose tools/build-uki now GPG-verifies the kernel + systemd-boot RPMs it downloads. No effect on compiled artifacts. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Move the docker-images + DOCKER_RUN plumbing block out of the Makefile into a canonical build.mk that the Makefile now includes. This is the source of truth for the shared harness, vendored byte-identically into stage1/vaportpm via the workspace make sync-harness and guarded by make check-harness. Pure relocation -- make -n expands identically; docker-build-harness (stage0-only) stays in the Makefile. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Both stage0 crates floated on vaportpm main (no rev), and stage0's Cargo.lock is untracked, so vaportpm-attest resolved to main HEAD on every build -- no version stability for a measured bootloader. Pin to the v0.3.0 release commit (7041f461) with a version = "=0.3.0" guard, keeping default-features = false (no_std core: Tpm, PcrOps, TpmTransport). Verified: stage0.efi and payload.efi build clean for x86_64-unknown-uefi against v0.3.0. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…entifiable) Replace the stage0.efi-hash-derived DISK_GUID/PART_GUID/VOLUME_ID with fixed constants ("4c4f434b424f4f54" spells "LOCKBOOT"; last group a role index: disk = 0, ESP = partition 1). Constant across releases so the GPT -- and hence the firmware's PCR 5 measurement -- is byte-stable, and a lockboot disk is identifiable by GUID. Dependency of stage2, which locates the boot disk by GUID and rewrites the GPT to add its runtime/data partitions; a stable base GPT keeps PCR 5 predictable. The stage0.efi hash still feeds BUILD_ID. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…#19) ## What Bring the UKI to a current, GPG-verified Fedora 44 base, bundle the kernel modules stage2 needs, and supporting build cleanups. ## Kernel + stub - Bump kernel `6.12.4-200.fc41` -> `7.0.12-201.fc44` and systemd-boot stub `256.17-1.fc41` -> `259.6-1.fc44` (what Fedora CoreOS 44.20260621.3.1 ships). `fc41` is EOL; the kernel was ~18 months behind. - **GPG verification (mandatory):** `download-verify-rpm.sh` fetches from koji's *signed* path (plain `packages/` is unsigned) and verifies each RPM against `keys/RPM-GPG-KEY-fedora-44-primary` (fp `36F6 12DC F27F 7D1A 48A8 35E4 DBFC F71C 6D9F 90A6`) on every build. - **Automated bumps:** pins live in generated `tools/build-uki/fedora-deps.mk`; `make update-fedora-deps FCOS=... SYSTEMD=...` reads the FCOS manifest, downloads + GPG-verifies the signed RPMs, and rewrites the pins. ## Storage modules for stage2 - Bundle + explicitly load (in dependency order, before the `modules_disabled` latch) the modules stage2's loader needs: **dm-crypt** (encrypted /data), **dm-verity** + **reed_solomon** (integrity-checked erofs runtime + FEC dep), **overlay** (ephemeral root), **erofs** + **netfs** (RO image fs + dep), and the **nvme** chain (hkdf -> nvme-auth/keyring -> nvme-core -> nvme; EC2 EBS + the harness disk are NVMe). The payload cannot load modules itself, and an unresolved dep can't be pulled once the latch is set, so deps are listed explicitly. - Module comments describe each dependency's capability generically. **Authenticated /data (dm-integrity + async_xor/async_tx) is documented as an extension point but not shipped** -- stage2 uses confidentiality-only dm-crypt, so those are omitted to keep the measured initramfs minimal. ## Build - Pin `vaportpm-attest` to v0.3.0 (rev + version guard). - Extract the shared `DOCKER_RUN` harness into `build.mk`. - Drop a stale `kernel-hash-%` target. ## Verification - ena (AWS) + gve (GCP) present in 7.0.12; all required `.ko.xz` copy. - `make test-chain-x86_64 SIGN=1 / MANIFEST=1 / SIGN_ARGS=1` boot `Linux 7.0.12-201.fc44` end to end and run stage2. - lockboot/stage2's XTS pivot test passes end to end on this UKI (`PIVOT_TEST=PASS`), dm-integrity absent from the boot. - aarch64 pins are GPG-verified but not boot-tested here. Pairs with lockboot/stage0#10 (image parity: rpm+gnupg, shared harness, vaportpm pin, stable GUIDs). 🤖 Generated with [Claude Code](https://claude.com/claude-code)

HarryR · 2026-07-10T07:54:47Z

What

Bring the stage0 build image and disk layout in line with the Fedora 44 chain bump, plus stable disk identity.

Changes

rpm + gnupg added to Dockerfile.build so the UKI build can verify Fedora package GPG signatures (image parity with stage1, kept byte-identical). No effect on compiled artifacts.
Stable lockboot GPT GUIDs + FAT volume-id.DISK_GUID / PART_GUID are now constant across releases (4c4f434b-424f-4f54-..., "LOCKBOOT" + a role index), so the initial GPT is byte-stable (stable pre-stage2 PCR5) and the boot disk is identifiable by a fixed GUID rather than by hashing stage0.efi (which changed per release). This is the stable identity stage2's find_boot_device can key on. (stage2 still rewrites the GPT at runtime to add p2/p3, which is why it excludes PCR5 from its key binding.)
Pin vaportpm-attest to v0.3.0 (rev + version guard), matching stage1.
Extract the shared DOCKER_RUN harness into build.mk (same refactor as stage1).

Verification

Reproducible: GUIDs/timestamps fixed; build output byte-stable.
No change to the runtime boot path. Pairs with uki: Fedora 44 kernel 7.0.12 + stub 259.6; GPG-verify + automate deps stage1#19 (Fedora 44 kernel/stub) and stage2: rework /data persistence (XTS, meta_bg) + mkruntime tooling stage2#1.

🤖 Generated with Claude Code

Needed so the UKI build can verify Fedora package GPG signatures (rpm/rpmkeys + gnupg). Image parity: Dockerfile.build is kept byte-identical with stage1, whose tools/build-uki now GPG-verifies the kernel + systemd-boot RPMs it downloads. No effect on compiled artifacts. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Move the docker-images + DOCKER_RUN plumbing block out of the Makefile into a canonical build.mk that the Makefile now includes. This is the source of truth for the shared harness, vendored byte-identically into stage1/vaportpm via the workspace make sync-harness and guarded by make check-harness. Pure relocation -- make -n expands identically; docker-build-harness (stage0-only) stays in the Makefile. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Both stage0 crates floated on vaportpm main (no rev), and stage0's Cargo.lock is untracked, so vaportpm-attest resolved to main HEAD on every build -- no version stability for a measured bootloader. Pin to the v0.3.0 release commit (7041f461) with a version = "=0.3.0" guard, keeping default-features = false (no_std core: Tpm, PcrOps, TpmTransport). Verified: stage0.efi and payload.efi build clean for x86_64-unknown-uefi against v0.3.0. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…entifiable) Replace the stage0.efi-hash-derived DISK_GUID/PART_GUID/VOLUME_ID with fixed constants ("4c4f434b424f4f54" spells "LOCKBOOT"; last group a role index: disk = 0, ESP = partition 1). Constant across releases so the GPT -- and hence the firmware's PCR 5 measurement -- is byte-stable, and a lockboot disk is identifiable by GUID. Dependency of stage2, which locates the boot disk by GUID and rewrites the GPT to add its runtime/data partitions; a stable base GPT keeps PCR 5 predictable. The stage0.efi hash still feeds BUILD_ID. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…#19) ## What Bring the UKI to a current, GPG-verified Fedora 44 base, bundle the kernel modules stage2 needs, and supporting build cleanups. ## Kernel + stub - Bump kernel `6.12.4-200.fc41` -> `7.0.12-201.fc44` and systemd-boot stub `256.17-1.fc41` -> `259.6-1.fc44` (what Fedora CoreOS 44.20260621.3.1 ships). `fc41` is EOL; the kernel was ~18 months behind. - **GPG verification (mandatory):** `download-verify-rpm.sh` fetches from koji's *signed* path (plain `packages/` is unsigned) and verifies each RPM against `keys/RPM-GPG-KEY-fedora-44-primary` (fp `36F6 12DC F27F 7D1A 48A8 35E4 DBFC F71C 6D9F 90A6`) on every build. - **Automated bumps:** pins live in generated `tools/build-uki/fedora-deps.mk`; `make update-fedora-deps FCOS=... SYSTEMD=...` reads the FCOS manifest, downloads + GPG-verifies the signed RPMs, and rewrites the pins. ## Storage modules for stage2 - Bundle + explicitly load (in dependency order, before the `modules_disabled` latch) the modules stage2's loader needs: **dm-crypt** (encrypted /data), **dm-verity** + **reed_solomon** (integrity-checked erofs runtime + FEC dep), **overlay** (ephemeral root), **erofs** + **netfs** (RO image fs + dep), and the **nvme** chain (hkdf -> nvme-auth/keyring -> nvme-core -> nvme; EC2 EBS + the harness disk are NVMe). The payload cannot load modules itself, and an unresolved dep can't be pulled once the latch is set, so deps are listed explicitly. - Module comments describe each dependency's capability generically. **Authenticated /data (dm-integrity + async_xor/async_tx) is documented as an extension point but not shipped** -- stage2 uses confidentiality-only dm-crypt, so those are omitted to keep the measured initramfs minimal. ## Build - Pin `vaportpm-attest` to v0.3.0 (rev + version guard). - Extract the shared `DOCKER_RUN` harness into `build.mk`. - Drop a stale `kernel-hash-%` target. ## Verification - ena (AWS) + gve (GCP) present in 7.0.12; all required `.ko.xz` copy. - `make test-chain-x86_64 SIGN=1 / MANIFEST=1 / SIGN_ARGS=1` boot `Linux 7.0.12-201.fc44` end to end and run stage2. - lockboot/stage2's XTS pivot test passes end to end on this UKI (`PIVOT_TEST=PASS`), dm-integrity absent from the boot. - aarch64 pins are GPG-verified but not boot-tested here. Pairs with lockboot/stage0#10 (image parity: rpm+gnupg, shared harness, vaportpm pin, stable GUIDs). 🤖 Generated with [Claude Code](https://claude.com/claude-code)

HarryR · 2026-07-10T07:54:47Z

What

Bring the stage0 build image and disk layout in line with the Fedora 44 chain bump, plus stable disk identity.

Changes

rpm + gnupg added to Dockerfile.build so the UKI build can verify Fedora package GPG signatures (image parity with stage1, kept byte-identical). No effect on compiled artifacts.
Stable lockboot GPT GUIDs + FAT volume-id.DISK_GUID / PART_GUID are now constant across releases (4c4f434b-424f-4f54-..., "LOCKBOOT" + a role index), so the initial GPT is byte-stable (stable pre-stage2 PCR5) and the boot disk is identifiable by a fixed GUID rather than by hashing stage0.efi (which changed per release). This is the stable identity stage2's find_boot_device can key on. (stage2 still rewrites the GPT at runtime to add p2/p3, which is why it excludes PCR5 from its key binding.)
Pin vaportpm-attest to v0.3.0 (rev + version guard), matching stage1.
Extract the shared DOCKER_RUN harness into build.mk (same refactor as stage1).

Verification

Reproducible: GUIDs/timestamps fixed; build output byte-stable.
No change to the runtime boot path. Pairs with uki: Fedora 44 kernel 7.0.12 + stub 259.6; GPG-verify + automate deps stage1#19 (Fedora 44 kernel/stub) and stage2: rework /data persistence (XTS, meta_bg) + mkruntime tooling stage2#1.

🤖 Generated with Claude Code

Needed so the UKI build can verify Fedora package GPG signatures (rpm/rpmkeys + gnupg). Image parity: Dockerfile.build is kept byte-identical with stage1, whose tools/build-uki now GPG-verifies the kernel + systemd-boot RPMs it downloads. No effect on compiled artifacts. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Move the docker-images + DOCKER_RUN plumbing block out of the Makefile into a canonical build.mk that the Makefile now includes. This is the source of truth for the shared harness, vendored byte-identically into stage1/vaportpm via the workspace make sync-harness and guarded by make check-harness. Pure relocation -- make -n expands identically; docker-build-harness (stage0-only) stays in the Makefile. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Both stage0 crates floated on vaportpm main (no rev), and stage0's Cargo.lock is untracked, so vaportpm-attest resolved to main HEAD on every build -- no version stability for a measured bootloader. Pin to the v0.3.0 release commit (7041f461) with a version = "=0.3.0" guard, keeping default-features = false (no_std core: Tpm, PcrOps, TpmTransport). Verified: stage0.efi and payload.efi build clean for x86_64-unknown-uefi against v0.3.0. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…entifiable) Replace the stage0.efi-hash-derived DISK_GUID/PART_GUID/VOLUME_ID with fixed constants ("4c4f434b424f4f54" spells "LOCKBOOT"; last group a role index: disk = 0, ESP = partition 1). Constant across releases so the GPT -- and hence the firmware's PCR 5 measurement -- is byte-stable, and a lockboot disk is identifiable by GUID. Dependency of stage2, which locates the boot disk by GUID and rewrites the GPT to add its runtime/data partitions; a stable base GPT keeps PCR 5 predictable. The stage0.efi hash still feeds BUILD_ID. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…#19) ## What Bring the UKI to a current, GPG-verified Fedora 44 base, bundle the kernel modules stage2 needs, and supporting build cleanups. ## Kernel + stub - Bump kernel `6.12.4-200.fc41` -> `7.0.12-201.fc44` and systemd-boot stub `256.17-1.fc41` -> `259.6-1.fc44` (what Fedora CoreOS 44.20260621.3.1 ships). `fc41` is EOL; the kernel was ~18 months behind. - **GPG verification (mandatory):** `download-verify-rpm.sh` fetches from koji's *signed* path (plain `packages/` is unsigned) and verifies each RPM against `keys/RPM-GPG-KEY-fedora-44-primary` (fp `36F6 12DC F27F 7D1A 48A8 35E4 DBFC F71C 6D9F 90A6`) on every build. - **Automated bumps:** pins live in generated `tools/build-uki/fedora-deps.mk`; `make update-fedora-deps FCOS=... SYSTEMD=...` reads the FCOS manifest, downloads + GPG-verifies the signed RPMs, and rewrites the pins. ## Storage modules for stage2 - Bundle + explicitly load (in dependency order, before the `modules_disabled` latch) the modules stage2's loader needs: **dm-crypt** (encrypted /data), **dm-verity** + **reed_solomon** (integrity-checked erofs runtime + FEC dep), **overlay** (ephemeral root), **erofs** + **netfs** (RO image fs + dep), and the **nvme** chain (hkdf -> nvme-auth/keyring -> nvme-core -> nvme; EC2 EBS + the harness disk are NVMe). The payload cannot load modules itself, and an unresolved dep can't be pulled once the latch is set, so deps are listed explicitly. - Module comments describe each dependency's capability generically. **Authenticated /data (dm-integrity + async_xor/async_tx) is documented as an extension point but not shipped** -- stage2 uses confidentiality-only dm-crypt, so those are omitted to keep the measured initramfs minimal. ## Build - Pin `vaportpm-attest` to v0.3.0 (rev + version guard). - Extract the shared `DOCKER_RUN` harness into `build.mk`. - Drop a stale `kernel-hash-%` target. ## Verification - ena (AWS) + gve (GCP) present in 7.0.12; all required `.ko.xz` copy. - `make test-chain-x86_64 SIGN=1 / MANIFEST=1 / SIGN_ARGS=1` boot `Linux 7.0.12-201.fc44` end to end and run stage2. - lockboot/stage2's XTS pivot test passes end to end on this UKI (`PIVOT_TEST=PASS`), dm-integrity absent from the boot. - aarch64 pins are GPG-verified but not boot-tested here. Pairs with lockboot/stage0#10 (image parity: rpm+gnupg, shared harness, vaportpm pin, stable GUIDs). 🤖 Generated with [Claude Code](https://claude.com/claude-code)

HarryR · 2026-07-10T07:54:47Z

What

Bring the stage0 build image and disk layout in line with the Fedora 44 chain bump, plus stable disk identity.

Changes

rpm + gnupg added to Dockerfile.build so the UKI build can verify Fedora package GPG signatures (image parity with stage1, kept byte-identical). No effect on compiled artifacts.
Stable lockboot GPT GUIDs + FAT volume-id.DISK_GUID / PART_GUID are now constant across releases (4c4f434b-424f-4f54-..., "LOCKBOOT" + a role index), so the initial GPT is byte-stable (stable pre-stage2 PCR5) and the boot disk is identifiable by a fixed GUID rather than by hashing stage0.efi (which changed per release). This is the stable identity stage2's find_boot_device can key on. (stage2 still rewrites the GPT at runtime to add p2/p3, which is why it excludes PCR5 from its key binding.)
Pin vaportpm-attest to v0.3.0 (rev + version guard), matching stage1.
Extract the shared DOCKER_RUN harness into build.mk (same refactor as stage1).

Verification

Reproducible: GUIDs/timestamps fixed; build output byte-stable.
No change to the runtime boot path. Pairs with uki: Fedora 44 kernel 7.0.12 + stub 259.6; GPG-verify + automate deps stage1#19 (Fedora 44 kernel/stub) and stage2: rework /data persistence (XTS, meta_bg) + mkruntime tooling stage2#1.

🤖 Generated with Claude Code

Needed so the UKI build can verify Fedora package GPG signatures (rpm/rpmkeys + gnupg). Image parity: Dockerfile.build is kept byte-identical with stage1, whose tools/build-uki now GPG-verifies the kernel + systemd-boot RPMs it downloads. No effect on compiled artifacts. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Move the docker-images + DOCKER_RUN plumbing block out of the Makefile into a canonical build.mk that the Makefile now includes. This is the source of truth for the shared harness, vendored byte-identically into stage1/vaportpm via the workspace make sync-harness and guarded by make check-harness. Pure relocation -- make -n expands identically; docker-build-harness (stage0-only) stays in the Makefile. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Both stage0 crates floated on vaportpm main (no rev), and stage0's Cargo.lock is untracked, so vaportpm-attest resolved to main HEAD on every build -- no version stability for a measured bootloader. Pin to the v0.3.0 release commit (7041f461) with a version = "=0.3.0" guard, keeping default-features = false (no_std core: Tpm, PcrOps, TpmTransport). Verified: stage0.efi and payload.efi build clean for x86_64-unknown-uefi against v0.3.0. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…entifiable) Replace the stage0.efi-hash-derived DISK_GUID/PART_GUID/VOLUME_ID with fixed constants ("4c4f434b424f4f54" spells "LOCKBOOT"; last group a role index: disk = 0, ESP = partition 1). Constant across releases so the GPT -- and hence the firmware's PCR 5 measurement -- is byte-stable, and a lockboot disk is identifiable by GUID. Dependency of stage2, which locates the boot disk by GUID and rewrites the GPT to add its runtime/data partitions; a stable base GPT keeps PCR 5 predictable. The stage0.efi hash still feeds BUILD_ID. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…#19) ## What Bring the UKI to a current, GPG-verified Fedora 44 base, bundle the kernel modules stage2 needs, and supporting build cleanups. ## Kernel + stub - Bump kernel `6.12.4-200.fc41` -> `7.0.12-201.fc44` and systemd-boot stub `256.17-1.fc41` -> `259.6-1.fc44` (what Fedora CoreOS 44.20260621.3.1 ships). `fc41` is EOL; the kernel was ~18 months behind. - **GPG verification (mandatory):** `download-verify-rpm.sh` fetches from koji's *signed* path (plain `packages/` is unsigned) and verifies each RPM against `keys/RPM-GPG-KEY-fedora-44-primary` (fp `36F6 12DC F27F 7D1A 48A8 35E4 DBFC F71C 6D9F 90A6`) on every build. - **Automated bumps:** pins live in generated `tools/build-uki/fedora-deps.mk`; `make update-fedora-deps FCOS=... SYSTEMD=...` reads the FCOS manifest, downloads + GPG-verifies the signed RPMs, and rewrites the pins. ## Storage modules for stage2 - Bundle + explicitly load (in dependency order, before the `modules_disabled` latch) the modules stage2's loader needs: **dm-crypt** (encrypted /data), **dm-verity** + **reed_solomon** (integrity-checked erofs runtime + FEC dep), **overlay** (ephemeral root), **erofs** + **netfs** (RO image fs + dep), and the **nvme** chain (hkdf -> nvme-auth/keyring -> nvme-core -> nvme; EC2 EBS + the harness disk are NVMe). The payload cannot load modules itself, and an unresolved dep can't be pulled once the latch is set, so deps are listed explicitly. - Module comments describe each dependency's capability generically. **Authenticated /data (dm-integrity + async_xor/async_tx) is documented as an extension point but not shipped** -- stage2 uses confidentiality-only dm-crypt, so those are omitted to keep the measured initramfs minimal. ## Build - Pin `vaportpm-attest` to v0.3.0 (rev + version guard). - Extract the shared `DOCKER_RUN` harness into `build.mk`. - Drop a stale `kernel-hash-%` target. ## Verification - ena (AWS) + gve (GCP) present in 7.0.12; all required `.ko.xz` copy. - `make test-chain-x86_64 SIGN=1 / MANIFEST=1 / SIGN_ARGS=1` boot `Linux 7.0.12-201.fc44` end to end and run stage2. - lockboot/stage2's XTS pivot test passes end to end on this UKI (`PIVOT_TEST=PASS`), dm-integrity absent from the boot. - aarch64 pins are GPG-verified but not boot-tested here. Pairs with lockboot/stage0#10 (image parity: rpm+gnupg, shared harness, vaportpm pin, stable GUIDs). 🤖 Generated with [Claude Code](https://claude.com/claude-code)

HarryR · 2026-07-10T07:54:47Z

What

Bring the stage0 build image and disk layout in line with the Fedora 44 chain bump, plus stable disk identity.

Changes

rpm + gnupg added to Dockerfile.build so the UKI build can verify Fedora package GPG signatures (image parity with stage1, kept byte-identical). No effect on compiled artifacts.
Stable lockboot GPT GUIDs + FAT volume-id.DISK_GUID / PART_GUID are now constant across releases (4c4f434b-424f-4f54-..., "LOCKBOOT" + a role index), so the initial GPT is byte-stable (stable pre-stage2 PCR5) and the boot disk is identifiable by a fixed GUID rather than by hashing stage0.efi (which changed per release). This is the stable identity stage2's find_boot_device can key on. (stage2 still rewrites the GPT at runtime to add p2/p3, which is why it excludes PCR5 from its key binding.)
Pin vaportpm-attest to v0.3.0 (rev + version guard), matching stage1.
Extract the shared DOCKER_RUN harness into build.mk (same refactor as stage1).

Verification

Reproducible: GUIDs/timestamps fixed; build output byte-stable.
No change to the runtime boot path. Pairs with uki: Fedora 44 kernel 7.0.12 + stub 259.6; GPG-verify + automate deps stage1#19 (Fedora 44 kernel/stub) and stage2: rework /data persistence (XTS, meta_bg) + mkruntime tooling stage2#1.

🤖 Generated with Claude Code

Needed so the UKI build can verify Fedora package GPG signatures (rpm/rpmkeys + gnupg). Image parity: Dockerfile.build is kept byte-identical with stage1, whose tools/build-uki now GPG-verifies the kernel + systemd-boot RPMs it downloads. No effect on compiled artifacts. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Move the docker-images + DOCKER_RUN plumbing block out of the Makefile into a canonical build.mk that the Makefile now includes. This is the source of truth for the shared harness, vendored byte-identically into stage1/vaportpm via the workspace make sync-harness and guarded by make check-harness. Pure relocation -- make -n expands identically; docker-build-harness (stage0-only) stays in the Makefile. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Both stage0 crates floated on vaportpm main (no rev), and stage0's Cargo.lock is untracked, so vaportpm-attest resolved to main HEAD on every build -- no version stability for a measured bootloader. Pin to the v0.3.0 release commit (7041f461) with a version = "=0.3.0" guard, keeping default-features = false (no_std core: Tpm, PcrOps, TpmTransport). Verified: stage0.efi and payload.efi build clean for x86_64-unknown-uefi against v0.3.0. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…entifiable) Replace the stage0.efi-hash-derived DISK_GUID/PART_GUID/VOLUME_ID with fixed constants ("4c4f434b424f4f54" spells "LOCKBOOT"; last group a role index: disk = 0, ESP = partition 1). Constant across releases so the GPT -- and hence the firmware's PCR 5 measurement -- is byte-stable, and a lockboot disk is identifiable by GUID. Dependency of stage2, which locates the boot disk by GUID and rewrites the GPT to add its runtime/data partitions; a stable base GPT keeps PCR 5 predictable. The stage0.efi hash still feeds BUILD_ID. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…#19) ## What Bring the UKI to a current, GPG-verified Fedora 44 base, bundle the kernel modules stage2 needs, and supporting build cleanups. ## Kernel + stub - Bump kernel `6.12.4-200.fc41` -> `7.0.12-201.fc44` and systemd-boot stub `256.17-1.fc41` -> `259.6-1.fc44` (what Fedora CoreOS 44.20260621.3.1 ships). `fc41` is EOL; the kernel was ~18 months behind. - **GPG verification (mandatory):** `download-verify-rpm.sh` fetches from koji's *signed* path (plain `packages/` is unsigned) and verifies each RPM against `keys/RPM-GPG-KEY-fedora-44-primary` (fp `36F6 12DC F27F 7D1A 48A8 35E4 DBFC F71C 6D9F 90A6`) on every build. - **Automated bumps:** pins live in generated `tools/build-uki/fedora-deps.mk`; `make update-fedora-deps FCOS=... SYSTEMD=...` reads the FCOS manifest, downloads + GPG-verifies the signed RPMs, and rewrites the pins. ## Storage modules for stage2 - Bundle + explicitly load (in dependency order, before the `modules_disabled` latch) the modules stage2's loader needs: **dm-crypt** (encrypted /data), **dm-verity** + **reed_solomon** (integrity-checked erofs runtime + FEC dep), **overlay** (ephemeral root), **erofs** + **netfs** (RO image fs + dep), and the **nvme** chain (hkdf -> nvme-auth/keyring -> nvme-core -> nvme; EC2 EBS + the harness disk are NVMe). The payload cannot load modules itself, and an unresolved dep can't be pulled once the latch is set, so deps are listed explicitly. - Module comments describe each dependency's capability generically. **Authenticated /data (dm-integrity + async_xor/async_tx) is documented as an extension point but not shipped** -- stage2 uses confidentiality-only dm-crypt, so those are omitted to keep the measured initramfs minimal. ## Build - Pin `vaportpm-attest` to v0.3.0 (rev + version guard). - Extract the shared `DOCKER_RUN` harness into `build.mk`. - Drop a stale `kernel-hash-%` target. ## Verification - ena (AWS) + gve (GCP) present in 7.0.12; all required `.ko.xz` copy. - `make test-chain-x86_64 SIGN=1 / MANIFEST=1 / SIGN_ARGS=1` boot `Linux 7.0.12-201.fc44` end to end and run stage2. - lockboot/stage2's XTS pivot test passes end to end on this UKI (`PIVOT_TEST=PASS`), dm-integrity absent from the boot. - aarch64 pins are GPG-verified but not boot-tested here. Pairs with lockboot/stage0#10 (image parity: rpm+gnupg, shared harness, vaportpm pin, stable GUIDs). 🤖 Generated with [Claude Code](https://claude.com/claude-code)

HarryRand others added 3 commits July 10, 2026 07:54

HarryR force-pushed the fedora44-kernel-bump branch from fcda1b0 to 9e64f6eCompare July 11, 2026 19:43