[FEATURE] Implement original-object namespace mutations and lifetime #27
Labels
No labels
alpha-1
bug
ci
docs
duplicate
enhancement
help wanted
invalid
performance
phase-6
question
refactor
security
wontfix
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
erix/posixd#27
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Problem and motivation
Build tools need directory creation, unlink/removal, hard and symbolic links, raw link reads and atomic rename through their original delegated namespace. The compatibility acquisition scope currently creates regular files but does not mediate this complete operation family. Existing provider pin restrictions also prevent the required lifetime of open objects after namespace changes.
Proposed behavior and scope
Implement mkdir/mkdirat, rmdir/unlink/unlinkat, link/linkat, symlink/symlinkat, readlink/readlinkat and rename/renameat as required by the explicit build workload. Raw paths use each admitted client's original root, cwd or live dirfd; final-link and trailing-directory intent remain explicit. A symbolic-link target is counted data rather than authority. Readlink returns the bounded original target prefix without appending a terminator or using it as a lookup path.
Define conditional native effects over retained original parents and targets, fresh one-use permission revisions and exact basenames. Two-path operations retain both independently resolved originals until one atomic provider effect. Cross-filesystem refusal precedes mutation and compares actual original filesystem custody. Same-object rename succeeds without changing namespace facts. Default ACL, creation mask, group inheritance, sticky-directory checks, link counts and source-qualified event times use authenticated original policy.
Implement the necessary RAM/ext/FAT provider operations and explicit format-supported refusals; build-workload-required behavior must be supported on its selected filesystem. Resolve held-object detachment as a dependency of complete namespace acceptance: unlink or replacement retires the namespace edge while the original opened object remains alive until its final owner retires. Directory moves preserve original custody and never reopen a pathname to recover it.
Authority, security and reliability
The native ceiling is checked before deny-only effective credentials and complete ACLs. UID zero is not a capability. No client-supplied provider endpoint, inode number, mount name, parent path, owner ID or numeric operation key selects new authority. Parent/target checks and conditional publication share the original provider serialization and mutation boundary. Destination absence is not inferred from a generic lookup failure.
Reserve counted custody before effects. A native operation is not replayed after uncertain completion; cancellation retires original callers and pins without unlinking a possibly committed result. Generation/identity retirement prevents slot reuse from reviving old objects. Atomic replacement includes source/destination metadata and retains the provider's original transaction or fail-stop policy. Unexpected replies preserve cleanup. No extra pathname, link-depth or directory-entry ceiling is introduced; real format, protocol and selected storage limits remain explicit.
Acceptance criteria
Checklist
Alternatives and tradeoffs
Reopening an authorized path can select a different object. Relaxing a provider's open-handle checks without retaining detached object lifetime can expose reuse or free live data. These approaches do not satisfy the required original-object contract. Native fixed administrative entry grants and compatibility mutation intents remain distinct, while provider primitives and lifetime accounting should be shared where their semantics agree.
The normative operation basis is the POSIX rename contract, unlink lifetime, link variants and readlink contract. Implementation remains original; specification references do not provide native authority.
Tracking and rollout
File/namespace umbrella, deny-only permission policy, completed status/directory prerequisite, original FAT catalog and phase completion.
Reviews: Posixd, Vfsd, Fatd, Integration, manual. Provider and wire reviews will be linked as their implementation is adopted.
Both full EriX guest generations and external Rust/LLVM rebuilding inside EriX remain required and unperformed.
Detailed scope and contracts
Cross-provider filesystem lifetime — signed source acceptance, 2 October 2026
Two live binding scenarios cover RAM, ext and separately selected FAT32/exFAT: conditional replacement/unlink, retained aliases and original identities, moved/detached directories, name reuse, supported links and read-only, cross-filesystem and unsupported-link refusals. Final media observers require reclaimed allocations, consistent reverse indices and clear undo ownership.
Both full EriX guest generations and external Rust/LLVM guest rebuilding remain required and unperformed.
Phase progress: 19.12% weighted, 7.61% whole acceptance; 27/460 parents and 485/949 sub-items.
Signed lifetime source acceptance:
Current-source acceptance now covers cwd navigation and contextual launch,
exact object/entry and redirection serving, together with compatibility shared
descriptions through namespace changes and original cleanup.
Vfsd's cwd and contextual
serving criteria are accepted; its original ownership and retired-bridge criteria
were already complete. The namespace issue also reconciles its provider criterion
with prior signed RAM/ext/FAT completion. Complete cross-provider live conditional
workloads, resource/refusal, recovery/reuse coverage and coherent runtime graph
adoption remain open. Full EriX guest generations and external Rust/LLVM rebuilding remain
required and unperformed.
Recomputed progress: 19.02% weighted, 6.81% whole acceptance; 25/460 parents and 482/949 sub-items.
Original namespace mutation plan and provider custody — 1 October 2026
The complete namespace feature remains open.
Known pre-effect refusal preserves original pending custody; existing-object touch and absent no-create observations retain their distinct lifetimes.
They cover original-name admission, actual RAM/literal-ext effects, hardlink entry distinction, known refusal, completed-operation retirement, cleanup and namespace reuse.
The namespace umbrella now checks its completed status/directory criterion, matching the already accepted feature and phase master. No new namespace criterion is accepted for the provider prerequisite. Both complete EriX guest generations and external Rust/LLVM rebuilding remain required and unperformed.
RAM open-object lifetime implemented — 1 October 2026
Completed exact-entry cleanup never selects a reused node. Same-name rename is a metadata-preserving no-op; trusted change events remain on held displaced objects and clockless effects invalidate the affected observation.
Both existing VM scenarios require the new OPEN_UNLINK_OK marker. Namespace output capacity is admitted before effects; changed provider controls preserve their original custody/refusal coverage with the new detachment behavior.
The complete namespace mutation and lifetime design remains open: RAM hardlink storage, directory navigation/detachment, ext/FAT object lifetime, complete conditional native contracts and compatibility namespace serving are unfinished. No broader namespace acceptance is inferred from this provider boundary.
Both complete EriX guest generations and external Rust/LLVM rebuilding remain required and unperformed.
RAM entry graph and held-directory design
The next provider boundary separates deployment-owned namespace entries from objects, with independently configured entry capacity and authenticated capacity-layout migration. Every published entry has a nonreused identity; exact-entry custody retains that identity together with its original parent, basename and object, so unlink/relink of the same object cannot revive an old grant. Hardlinks share actual contents, metadata and open-object lifetime. All required storage, link arithmetic and entry identities are admitted before effects.
Held directories retain original parent-object references through detachment, with iterative final reclamation and no path reopening. Directory moves must preserve original object custody while subsequent navigation and enumeration enforce the retained root ceiling. Pending administrative exact-entry grants retain their explicit mutation-pin contract until complete conditional namespace integration. Independent provider, deployment-layout and native workload controls accompany the migration; execution remains deferred. This design does not complete the remaining ext/FAT or compatibility namespace requirements.
Independent RAM entries and retained directories — 1 October 2026
Unlink/relink of the same object and name cannot revive an old exact-entry grant. Creation beneath a removed directory refuses before effects.
Ext/FAT original-object lifetime and complete compatibility namespace serving also remain unfinished in the owning design.
All 37 runtime configurations, packaging, fixtures, CLI/shell routing and reports adopt the new quota; obsolete layouts and missing entry policy are refused. Nineteen Rust and four Python controls, plus two shell policy refusal cases, are authored and unrun. The twenty-operation native VFS transcript requires shared hardlink metadata, retained bytes after both names disappear, and removal of a held empty directory before its final close. Both existing RAM VM scenarios require the corresponding observations.
Both complete EriX guest generations and external Rust/LLVM rebuilding remain required and unperformed.
Conditional retained-entry contract
The next implementation stage binds each namespace selection to an actual retained parent under its original root and a caller-known cleanup key. A selection retains the complete basename, exact binding identity or original absence, original target and a fixed operation role. It cannot be used as an ordinary content handle or substituted into another operation family. Target permission/ACL observations remain available only through the original retained selection; parent and target evidence are compared again at the effect boundary.
Selection submission shares original-root replay fencing and counted cleanup. A namespace effect consumes its admitted selections once, including a definite conditional refusal; cleanup releases only original custody and never unlinks a committed result. Rename and hardlink retain both selections and compare actual provider/mount identity before effects. Missing or malformed replies preserve uncertainty and cannot authorize resubmission. Results distinguish successful completion, definite no-effect refusal and uncertain mutation, with operation-specific filesystem conditions preserved for compatibility errors.
Creation carries explicit authenticated owner/group/mask policy; symlink targets are counted raw bytes without target traversal. Every required resource and output extent is admitted before effects. Directory moves compare actual object ancestry rather than a pathname prefix. Existing legacy cwd and administrative pin contracts remain enforced until their actual consumers have migrated to held-object traversal and effects. Full compatibility serving still requires both original path/dirfd walks, complete ACL/sticky checks, final-link/trailing-separator rules, cross-filesystem refusal and original caller cleanup.
Implementations and independent malformed, replay, refusal, alias, cancellation and native workload controls will be published through a coherent source graph. All execution remains deferred; this design update does not accept an unfinished namespace or cwd requirement.
Conditional retained-entry namespace custody — 1 October 2026
Exact response correlation distinguishes completed effects, definite no-effect refusals and uncertainty.
Missing, malformed or uncorrelated replies preserve cleanup and never permit replay. The installed namespace is currently immutable and append-only; future unmount/reconfiguration requires a persistent nonwrapping epoch.
Ext and FAT return explicit correlated unsupported no-effect dispositions for the new family. Their conditional transactions remain unfinished. Complete compatibility path/dirfd, ACL/sticky, final-link and trailing-separator mediation and legacy cwd migration remain open in the owning design. Required legacy cwd and administrative spelling guards remain.
Thirty-nine new Rust controls are authored and unrun: ten literal wire controls, seventeen RAM provider controls, eight actual Vfsd dispatch controls, and two explicit-refusal controls each for ext and FAT.
Both complete EriX guest generations and external Rust/LLVM rebuilding remain required and unperformed.
Compatibility namespace framing design
Compatibility 1.16 will assign operations 27 through 32 to directory creation, entry removal, hardlink, symbolic-link creation, raw readlink and rename. Each counted request carries descriptive cwd/dirfd bases and exact byte suffixes. Empty paths remain semantic errors for the admitted runtime; embedded NUL is malformed. No provider endpoint, native handle, credential, permission revision or mount identifier is supplied by this ABI.
Use a canonical forty-byte prefix: primary base selector and fd at offsets 0/4, secondary base selector and fd at 8/12, mode or operation-specific boolean at 16, first/second byte counts at 20/24, reserved zero at 28, and readlink maximum at 32. The two exact suffixes follow at 40. Directory creation uses the primary path and permission mode; removal uses the primary path and directory boolean; hardlink uses two paths and a source-final-follow boolean; symbolic-link creation uses raw target bytes followed by the destination path and zeros the unused primary base; readlink uses the primary path and maximum; rename uses two paths. Every unused field is zero. Successful mutations have empty bodies. Readlink replies contain an eight-byte count and the exact raw prefix without a terminator; correlation rejects data beyond the original maximum.
Keep this pure contract separate from operation admission and actual effects. Migrate all current version producers, consumers and independent literal fixtures together. The active serving work must resolve both bases through the original authenticated client, retain two disjoint counted path regions during link expansion, check complete ACL/sticky policy, submit the conditional native effect once, and retire both original owners on all outcomes. A codec-only stage must return explicit unsupported behavior until those serving paths are implemented; it cannot satisfy full namespace acceptance.
Add independent literal success, malformed/unused-field, count/extent, zero-capability, reply-correlation and output-preservation controls. Preserve the current 96-byte status layout using its single shared size definition. Tests and CI remain deferred; source controls do not establish executed success.
Counted namespace intents and original raw readlink — 1 October 2026
Unused fields are canonical zero; names and symbolic targets carry no native authority.
Only the admitted prefix is staged; actual counted storage exhaustion refuses without silently shortening success. Final target bytes are never traversed, no content grant or fd is acquired, and all original callers, parents, links and expansion owners retire before delivery.
Complete mutation permission/effect mediation, two-path rename/link, legacy cwd migration and ext/FAT conditional transactions remain unfinished. Raw readlink is split from the previous combined two-path criterion to record one fully implemented operation boundary; the full namespace parent remains open. The existing native binding workload now requires seven raw-target observations/refusals per peer, including readlinkat and the read-only scope, without new authority or fixture budgets.
Both complete EriX guest generations and external Rust/LLVM rebuilding remain required and unperformed.
Original mutation mediation and errno-domain migration
The next stage connects the existing namespace intents to the retained-entry native families. Each operation preserves original parent custody and complete fresh permission evidence. Two-path operations retain independently resolved source/destination roles and disjoint counted storage, so one path's link expansion cannot overwrite the other path or a raw symbolic target. Complete parent search/write and access ACL checks precede selection; fresh target ownership supplies deny-only sticky-directory policy. Provider defaults are consumed inside the serialized conditional creation boundary.
Preserve operation-specific final-link, trailing-separator and root/dot rules using the POSIX pathname contract and rename contract. Final-link probing cannot promote an entry-selection owner into content access. All original selection and caller owners remain accounted through cancellation and uncertain completion. One admitted mutation is submitted once; cleanup releases custody without removing a possibly committed result.
Compatibility 1.17 will append ENOTEMPTY=28, EXDEV=29, EROFS=30 and EMLINK=31 to the checked errno domain, preserving every earlier assignment and all namespace body layouts. Migrate the Rust/C constants, current consumers and independent version/domain fixtures together. Only correlated proven no-effect native dispositions map to ordinary failures; uncertain completion cannot authorize replay. Full runtime/provider completion and guest rebuilding remain separate acceptance requirements, and all execution remains deferred by instruction.
Original-entry namespace mutation mediation — 1 October 2026
Hardlink source requires parent search; changed directories require write/search.
Ordinary final-link/trailing-directory probes cannot substitute for entry evidence. Cleanup cannot undo a completed mutation.
The independent filesystem model keeps real initial directory entries, so a removed name cannot reappear from a fixture fallback.
They cover independent dirfds, link following, same-object rename, complete ACL/sticky policy, exact refusals, original detached content, cancellation and uncertain completion. The read-only peer requires all five mutation families to refuse.
The exact wire/migration, actual permission/effect mediation and two-path semantic criteria are source-accepted. The full namespace parent stays open: legacy cwd migration, required ext/FAT conditional transactions and cross-provider lifetime remain unfinished. Both complete EriX guest generations and external Rust/LLVM guest rebuilding remain required and unperformed.
Disk namespace transaction and lifetime design — 1 October 2026
The next implementation completes conditional ext/FAT entry effects before migrating legacy cwd consumers. Each selection retains its original root, parent, fixed role, full basename, binding freshness and permission fingerprint. Target observations remain metadata-only. Both roles are consumed once before the conditional effect; stale, cross-filesystem and invalid evidence must refuse without publishing a name.
Ext will use its existing metadata transaction owner and standard persistent orphan representation to retain unlinked inodes until original handles, entry owners and detached directory ancestry retire. Final reclamation and recovery must preserve inode generations, block ownership, checksums, ACL value ownership and quota accounting. FAT will use the selected semantic catalog and undo transaction owner for shared object bindings and deferred storage reclamation. Neither implementation may claim rollback after uncertain media publication.
Creation must preserve explicit mode, owner/group, mask, inherited default/access ACLs, raw symbolic targets and original timestamps. Removal, hardlink and rename must share retained-parent mutation cores with existing administrative operations, including atomic replacement and actual ancestry checks. Keep legacy cwd guards until exact-object/launch and redirection callers use original entry custody on every required provider.
Author independent media/provider and lifecycle controls; maintain inline contracts and component/manual documentation. This is a planned implementation boundary, not accepted completion. Tests, builds, lint checks, VM/profiler execution and CI observation remain deferred by instruction. Owning namespace design; complete phase requirements.
Original ext inode lifetime — 1 October 2026
Cleanup-only rows cannot regain ordinary access after refusal, and uncertain lifetime publication permanently disables endpoint state.
Creation beneath a removed directory refuses before allocation, and complete close response storage precedes persistent cleanup.
The existing independent Integration helper selected by Rootd remains unchanged; these component-local changes require no Cargo edge update.
Full provider criteria stay open. Ext/FAT conditional namespace requests still return typed Unsupported; plain ext2 durable detachment, FAT lifetime and legacy cwd migration remain unfinished. Both full EriX guest generations and external Rust/LLVM rebuilding remain required and unperformed.
Conditional ext entry custody design — 1 October 2026
The next provider boundary binds each fixed namespace role to the original retained parent, basename, target generation or observed absence, navigation ceiling and complete permission fingerprint. Entry handles permit permission observations and original-key cleanup; they convey no content access. A nonwrapping mount namespace epoch advances before every namespace mutation attempt, including legacy entry routes, to reject remove/relink substitutions even when inode identity is unchanged. This conservative epoch can invalidate unrelated selections; it avoids an unbounded binding map and never caches authorization.
Selection preflights response and handle capacity, burns its original acquisition key and consumes the original parent observation before publication. Mutation consumes admitted roles once, compares original parent/target evidence and actual bindings under an exclusive provider borrow, then uses the shared media transaction. Cancellation, stale evidence and failed replies retain original cleanup ownership. Existing retained-inode cleanup and legacy cwd spelling guards remain in force. Creation policy, directory ACL inheritance, raw symbolic targets and two-parent link/rename belong to the complete implementation. Plain ext2 durability, FAT lifetime and legacy cwd migration remain required. Tests will be authored; execution and CI observation remain deferred.
Conditional journaled ext namespace serving — 1 October 2026
Every legacy and native namespace mutation attempt advances a nonwrapping mount epoch, rejecting remove/relink substitutions even when the inode is unchanged. Entry rows cannot grant content, navigation or ordinary metadata mutation.
Native internal names use ext's 255-byte limit; legacy 240-byte reply rows refuse unrepresentable names without truncation.
They cover original custody and permissions, short/lost replies, binding substitution, all effects, cross-mount refusal, retained content/parents, inherited local/external/streamed ACLs, full names/raw targets, allocation/epoch exhaustion, extended directory counts and uncertain publication.
Full provider/lifetime acceptance remains open. Plain ext2 durable transactions, FAT namespace/lifetime and legacy cwd migration remain unfinished; spelling guards stay until their consumers migrate. Filesystem fsync/device-barrier ordering and both full guest EriX generations and external Rust/LLVM rebuilding remain required and unperformed.
Design: explicitly selected original undo for plain ext2
Plain ext2 will retain its filesystem format and use the already configured same-disk external-journal partition as a separately formatted undo store. An immutable, checksummed footer binds the original filesystem UUID and both partition geometries. Neither discovery, formatting, a fallback journal nor an on-media identifier can grant another block route. The selected original journal is recovered before ordinary filesystem parsing. Read-only mounts cannot repair active media.
A private joined block view exposes exactly the held filesystem extent and the selected journal extent to LibBlock’s existing undo owner. Ordinary callbacks receive only the filesystem prefix. Persistence barriers cover both original routes; target effects become durable before the journal is cleared. Semantic refusal rolls back; uncertainty permanently disables ordinary endpoint use. Mount-time repairs must use the same transaction owner, without nested recovery or an ambient bypass.
Shared ext namespace, timestamp and detached-inode transactions will use this owner on admitted plain media. Original content owners and persistent orphan reclamation retain their existing lifetime rules. Authored controls will cover geometry/UUID substitution, unsupported durability, read-only recovery, rollback, lost completion, original ownership and all namespace effects. Integration will explicitly format/select the fixture partition and add VM scenarios. Documentation will specify the independent on-media format. Tests, builds and CI observation remain deferred by instruction.
Original durable undo for plain ext — 1 October 2026
Runtime mounts never discover, format or replace a journal.
Read-only mounts cannot repair active media; semantic refusal rolls back, while uncertain publication permanently disables ordinary endpoint use. Unselected plain media still refuse effects requiring rollback.
Full namespace/lifetime parents, FAT serving, legacy cwd migration, complete filesystem fsync/JBD2 barriers and both full EriX guest generations plus guest Rust/LLVM rebuilding remain unfinished.
Design: conditional original FAT entry operations
The existing selected original catalog and undo owner will serve metadata-only retained-entry selection, attributed regular-file/directory creation, unlink/rmdir and two-parent rename. Original root/acquisition cleanup keys, retained actual parents, complete catalog identity/revision evidence, raw binding observations and one-use roles precede a single filesystem/catalog transaction. No path reconstruction or new provider/device route is permitted. Short replies precede all custody/effects; lost replies retain original cleanup without replay.
The established namespace contract allows explicit format-specific Unsupported. FAT/exFAT have no native hardlink or symbolic-link representation; those effects will return the correlated no-effect disposition. This does not waive any operation required by the selected guest build filesystem. Creation ownership/mode/mask, actual directory links, qualified events and atomic rename/replacement remain required on selected FAT media.
Retained-object detachment and final/restart storage reclamation remain part of this same namespace/lifetime pass. Existing spelling guards stay until their consumers have original-object lifetime. They cannot be silently removed or treated as completed POSIX lifetime. Appropriate provider and VM controls, authority/rollback review, component/manual documentation and coherent signed publication accompany implementation. Tests, builds and CI observation remain deferred by instruction.
Conditional original FAT entries — 1 October 2026
Entry rows cannot supply ordinary I/O, navigation, candidate access or legacy exact mutation. Complete replies are admitted before custody/effects; lost replies retain original-key cleanup, and consumed roles cannot replay.
Legacy 240-byte observations refuse without truncation or cursor publication. Every operation required by the selected build workload still needs implementation on its selected filesystem.
The complete FAT namespace/lifetime items remain open: persistent detached objects, alias updates, original final-close/restart reclamation, legacy cwd migration and additional VM coverage are unfinished. Existing open-object refusal guards remain until that implementation is complete.
Original FAT detached-object design — 1 October 2026
Continue the original namespace implementation through complete FAT/exFAT object lifetime. A zero-link detached record retains its never-reissued original catalog serial, data allocation, independent metadata and actual parent serial. Separate canonical detached binding domains encode only the original serial; deleted native directory slots can be reused without aliasing old objects. Detached records have no namespace entry digest and cannot be opened by pathname. Unknown domains continue to refuse complete catalog admission on older providers.
The same original undo transaction removes or replaces the native name and publishes detached catalog state whenever original handles, entry custody or detached descendants still require the object. Existing live aliases switch to the original detached serial. Subsequent content, truncate, status and permission operations use that held original object and never touch its deleted directory slot. Renaming a live object refreshes original aliases without changing identity. Detached directories retain actual original parents for counted navigation and cleanup; no directory reconstruction or hidden-name namespace is introduced.
Original close, acquisition retirement and anchor retirement must retain cleanup-only custody until required reclamation is positively complete. Reclaim unowned detached leaves before parents under the original undo owner. Mount recovery must admit complete detached state and allocation ownership before reclaiming interrupted lifetimes. In particular, no detached allocation may overlap live objects, other detached objects, the original root, exFAT bitmaps or the name-comparison table. Read-only mounts refuse required cleanup; uncertain cleanup permanently disables ordinary endpoint use. No catalog number or orphan record confers another block route.
Complete the original allocation-error distinctions and remaining format-derived name allocation bounds, then add provider and VM controls for held file/directory content, alias updates, replacement, slot reuse, cleanup failures, restart reclamation and original allocation substitution. Testing and CI observation remain deferred. Full FAT namespace/lifetime acceptance stays open until the complete implementation exists.
Original FAT detached lifetime — 1 October 2026
Known cleanup refusal preserves original custody; uncertain publication permanently disables ordinary endpoint use. Read-only reclamation refuses before mount publication. Earlier providers reject unknown detached binding domains.
Eight new provider controls are authored and unrun for both formats: aliases and slot reuse, content and truncate, rename replacement, retained parents, restart/read-only refusal, cleanup retry, uncertain reclamation, overlapping allocation and original acquisition/anchor keys.
Dedicated VM controls, storage-versus-custody exhaustion, short-alias allocation limits, shared provider-status prose and legacy cwd/exact consumer migration remain open.
Both complete EriX guest build generations and external Rust/LLVM rebuilding remain required and unperformed.
FAT namespace capacity and recovery controls — 1 October 2026
Complete long names allocate all required directory slots. Native hardlink/symlink formats explicitly refuse, so the selected build filesystem must still implement every workload-required operation.
Six host controls and eight dedicated FAT/exFAT VM scenarios cover original detached parent/child restart, undo-before-recovery ordering, read-only policy and overlapping-allocation refusal. Positive media observations require native allocation integrity, catalog and reverse-index retirement, serial nonreuse and clear original undo.
Whole compatibility namespace/lifetime parents remain open, including legacy cwd/exact consumer migration and complete cross-provider controls.
Both full EriX guest build generations and external Rust/LLVM rebuilding remain required and unperformed.
Current FAT parent observations after moves — 1 October 2026
Ordinary selected acquisitions no longer keep obsolete entering handles; pending entry roles and imported media preserve their required custody.
Four additional provider controls cover both FAT and exFAT: moved directories and old-parent reclamation, aliases/repeated moves, detached parents after entering handles retire, root clamping and cyclic ancestry refusal.
Vfsd cwd/launch/exact/redirection consumers still reconstruct path spellings; their migration must precede removal of spelling pins. Complete cross-provider lifetime controls also remain open. Checklist acceptance is unchanged: 18.82% weighted, 6.81% whole-item; 25/460 parents and 473/943 sub-items across 165 decomposed parents.
Both full EriX guest build generations and external Rust/LLVM rebuilding remain required and unperformed.
Current independent FAT directory anchors — 1 October 2026
Imported clones copy only required retained ancestors below their ceiling, carrying no source cleanup keys or broader original roots.
Full Vfsd cwd/launch/exact/redirection migration and complete cross-provider lifetime remain open; legacy spelling guards stay until their consumers migrate.
Both full EriX guest build generations and external Rust/LLVM rebuilding remain required and unperformed.
Equal names share bytes through checked published-owner
counts; copying a descriptive key adds no owner or authority. Native enumeration
returns full raw names; fixed legacy replies explicitly refuse oversized output.
Integration requires independent name_bytes through runtime parser, CLI, shell
forwarding and reporting. Old V2 descriptors are rejected. Runtime catalog adoption remains open; compiled Rootd helpers are unchanged.
Allocator/provider/image controls are authored for complete 601/301-byte raw
names, exact capacity, shared ownership, caller reuse, stale replacement,
hardlinks, symlinks, native enumeration, short replies, original cleanup, mount
reuse, identity exhaustion, literal geometry and overflow/missing-policy refusal.
The containing consumer migration remains open for legacy
directory/consumer frames, complete cross-provider lifetime/refusal/restart
controls and coherent runtime adoption. Full EriX guest generations and external
Rust/LLVM rebuilding inside EriX remain required and unperformed.
State and name module cfg predicates now match their native/test
dependencies; source acceptance does not claim a compiler or CI reproduction.
Related records and contracts
erikinkinen referenced this issue from erix/docs2026-10-01 02:02:46 +02:00
erikinkinen referenced this issue from erix/e2fsd2026-10-01 02:02:47 +02:00
erikinkinen referenced this issue from erix/ramfsd2026-10-01 02:02:47 +02:00
erikinkinen referenced this issue2026-10-01 02:02:48 +02:00
erikinkinen referenced this issue from erix/integration2026-10-01 02:02:48 +02:00
erikinkinen referenced this issue from erix/rootd2026-10-01 02:30:59 +02:00
erikinkinen referenced this issue from erix/blockd2026-10-01 04:14:03 +02:00
erikinkinen referenced this issue from erix/coreutils2026-10-01 04:14:05 +02:00
erikinkinen referenced this issue from erix/crashd2026-10-01 04:14:06 +02:00
erikinkinen referenced this issue from erix/deviced2026-10-01 04:14:08 +02:00
erikinkinen referenced this issue from erix/drv-acpi2026-10-01 04:14:10 +02:00
erikinkinen referenced this issue from erix/drv-i80422026-10-01 04:14:11 +02:00
erikinkinen referenced this issue from erix/drv-ps2-keyboard2026-10-01 04:14:12 +02:00
erikinkinen referenced this issue from erix/drv-ps2-mouse2026-10-01 04:14:13 +02:00
erikinkinen referenced this issue from erix/drv-rtc-cmos2026-10-01 04:14:13 +02:00
erikinkinen referenced this issue from erix/drv-serial2026-10-01 04:14:14 +02:00
erikinkinen referenced this issue from erix/drv-virtio-block2026-10-01 04:14:15 +02:00
erikinkinen referenced this issue from erix/dynlinkd2026-10-01 04:14:16 +02:00
erikinkinen referenced this issue from erix/exsh2026-10-01 04:14:18 +02:00
erikinkinen referenced this issue from erix/extop2026-10-01 04:14:19 +02:00
erikinkinen referenced this issue from erix/fatd2026-10-01 04:14:20 +02:00
erikinkinen referenced this issue from erix/fbcond2026-10-01 04:14:21 +02:00
erikinkinen referenced this issue from erix/hotplugd2026-10-01 04:14:22 +02:00
erikinkinen referenced this issue from erix/inputd2026-10-01 04:14:22 +02:00
erikinkinen referenced this issue from erix/irqd2026-10-01 04:14:24 +02:00
erikinkinen referenced this issue from erix/kernel2026-10-01 04:14:25 +02:00
erikinkinen referenced this issue from erix/keyd2026-10-01 04:14:27 +02:00
erikinkinen referenced this issue from erix/launchd2026-10-01 04:14:28 +02:00
erikinkinen referenced this issue from erix/lib-block2026-10-01 04:14:29 +02:00
erikinkinen referenced this issue from erix/lib-bootstrap2026-10-01 04:14:30 +02:00
erikinkinen referenced this issue from erix/lib-capabi2026-10-01 04:14:31 +02:00
erikinkinen referenced this issue from erix/lib-driver2026-10-01 04:14:32 +02:00
erikinkinen referenced this issue from erix/lib-input2026-10-01 04:14:33 +02:00
erikinkinen referenced this issue from erix/lib-interrupt2026-10-01 04:14:34 +02:00
erikinkinen referenced this issue from erix/lib-ipc2026-10-01 04:14:35 +02:00
erikinkinen referenced this issue from erix/lib-log2026-10-01 04:14:36 +02:00
erikinkinen referenced this issue from erix/lib-posixabi2026-10-01 04:14:38 +02:00
erikinkinen referenced this issue from erix/lib-service2026-10-01 04:14:39 +02:00
erikinkinen referenced this issue from erix/lib-stdio2026-10-01 04:14:39 +02:00
erikinkinen referenced this issue from erix/lib-time2026-10-01 04:14:41 +02:00
erikinkinen referenced this issue from erix/lib-tty2026-10-01 04:14:42 +02:00
erikinkinen referenced this issue from erix/lib-vfs2026-10-01 04:14:43 +02:00
erikinkinen referenced this issue from erix/loaderd2026-10-01 04:14:44 +02:00
erikinkinen referenced this issue from erix/logd2026-10-01 04:14:45 +02:00
erikinkinen referenced this issue from erix/memd2026-10-01 04:14:46 +02:00
erikinkinen referenced this issue from erix/named2026-10-01 04:14:46 +02:00
erikinkinen referenced this issue from erix/pagerd2026-10-01 04:14:47 +02:00
erikinkinen referenced this issue from erix/powerboxd2026-10-01 04:14:48 +02:00
erikinkinen referenced this issue from erix/probed2026-10-01 04:14:48 +02:00
erikinkinen referenced this issue from erix/procd2026-10-01 04:14:49 +02:00
erikinkinen referenced this issue from erix/seriald2026-10-01 04:14:51 +02:00
erikinkinen referenced this issue from erix/timed2026-10-01 04:14:51 +02:00
erikinkinen referenced this issue from erix/ttyd2026-10-01 04:14:52 +02:00
erikinkinen referenced this issue from erix/vfsd2026-10-01 04:14:53 +02:00
erikinkinen referenced this issue from erix/vspaced2026-10-01 04:14:54 +02:00
erikinkinen referenced this issue from erix/boot-tool2026-10-02 02:34:24 +02:00
erikinkinen referenced this issue from erix/bootloader2026-10-02 02:34:26 +02:00
erikinkinen referenced this issue from erix/lib-bootimg2026-10-02 02:34:47 +02:00
erikinkinen referenced this issue from erix/lib-ed255192026-10-02 02:34:50 +02:00
erikinkinen referenced this issue from erix/lib-fs-crypto2026-10-02 02:34:51 +02:00
erikinkinen referenced this issue from erix/lib-partition2026-10-02 02:34:54 +02:00
erikinkinen referenced this issue from erix/sha2-tool2026-10-02 02:35:06 +02:00