Commit Graph
3 Commits
Author SHA1 Message Date
7fd50eb361 fix(zstd): share exact-frame Zstandard decoding across ELF and OTA (#12416)
* fix(zstd): share exact-frame Zstandard decoding across ELF and OTA

Walk RFC 8878 frames once so ELF sections, Android OTA REPLACE_ZSTD, and GPU payloads reject truncated headers, blocks, checksums, and trailing bytes on every supported Node runtime. Bound that walk with the runtime zstd.max_frames setting.

* fix(zstd): keep frame output tight and report GPU frame-cap refusals

Copy each Node 22 frame off its decoder chunk before the next frame is decoded. Map a GPU zstd.max_frames refusal through the resource-limit contract instead of a provider failure.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(zstd): keep frame-cap refusals and bad seeds out of malformed results

Project zstd.max_frames through one resource-limit helper for ELF notes, DWARF, GPU, and artifact reads. An invalid REA_ZSTD_MAX_FRAMES seed stays a configuration failure, and frame-header tests cover content size, window, dictionary id, and RLE blocks.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(zstd): yield GPU frame decoding so cancellation can land

The frame-by-frame fallback awaits a caller pause before each data frame. GPU yields and checks the read signal there, so a many-frame payload does not finish before a disconnect is observed.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(zstd): advertise the frame cap and keep strict GPU decode cancellable

List every operation that enforces zstd.max_frames, including GPU, DWARF, and ELF notes. The strict sequence uses the async decoder and checks the caller signal again after it returns, and a single output buffer is not copied a second time.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(zstd): refresh analysis-setting schemas and GPU frame-cap projection

Regenerate the advertised input summaries for inspect_analysis_settings and update_analysis_settings. GPU inventory now throws the projected zstd.max_frames resource limit from payload decode and #retainFailure, so the adjustable setting survives the public failure.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* style(gpu): format the frame-cap projection throw

oxfmt keeps the zstd.max_frames throw in #retainFailure on one line so CI format:check matches the local formatter.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(zstd): yield after the last frame-by-frame GPU decode

The fallback returned success in the same turn as the final synchronous frame, so a queued abort was still invisible to the caller's signal check. Pause once more before that success, matching the strict sequence.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(zstd): yield ELF, OTA, and DWARF frame decoding on the caller signal

The synchronous frame loop ignored abort signals owned by inspect_elf_notes, inspect_dwarf_units, and Android OTA partition reads. Those callers now use the cooperative decoder and yield before Zstandard work, and 32-frame cancellations cover each path.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(zstd): observe cancellation during structural parsing

* fix(zstd): project section-name frame caps at the ELF directory boundary

readElfDirectoryFromBuffer now requires the caller's operation and throws the adjustable zstd.max_frames resource limit. Hardening, unwind, and the other directory consumers keep that typed refusal instead of an adapter or I/O error.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(zstd): preserve optional ELF evidence when frames are capped

* fix(zip): validate Zstandard members through frame completion (#12461)

fix(zip): require complete Zstandard member frames

* fix(elf): yield hardening section name decoding

* fix(elf): yield section names for unwind and export reads

* fix(zstd): yield ELF directory reads that already own a cancellation signal

Hardening, unwind, initializers, exports, kernel modules, debug link, and Android JNI now read section names through the cooperative directory when their caller has a signal. Readers with no signal stay synchronous.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(elf): share one initializer directory and retain its snapshot

* fix(android): reuse the admitted JNI ELF directory

* fix(elf): retain optional debug candidate evidence through cooperative decoding

* fix(zstd): yield ELF section names for Go, Rust, and IL2CPP

These inspections already own an abort signal. Their ELF directory read now uses the cooperative container decoder so a queued cancellation can land during a many-frame section-name table.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(elf): recognize compressed kernel modules before ELF decoding

The cooperative module reader decoded the bytes as ELF before the container check, so gzip, xz, and zstd modules became malformed images. The compression check now returns the unsupported-container report first.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* refactor(elf): classify each container once before reading it

* fix(il2cpp): preserve admission order during cooperative linking

* fix(elf): reuse section-name decoding in debug-link comparison (#12472)

* style(elf): format updated section-reader test calls

---------

Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: morluto <morluto@users.noreply.github.com>
2026-11-19 19:45:26 +00:00
ShuaixinHuang 884fda99f0 fix(native): index only address-bearing Mach-O initializer symbols (#12070) 2026-11-18 10:53:42 -05:00
8bc01b48b4 feat(native): list ELF and Mach-O initializer and finalizer functions (#9574)
* feat(native): list ELF and Mach-O initializer and finalizer functions

inspect_native_initializers and inspect-native-initializers list the
functions an ELF or Mach-O image runs before and after main: ELF
DT_INIT, DT_FINI, preinit, init and fini arrays and .ctors/.dtors, and
Mach-O __mod_init_func, __mod_term_func and __init_offsets, with slot
addresses, resolved targets, symbols where available and relocation
facts for relocatable objects. Nothing is executed. Slot, symbol,
relocation and note budgets are shared across tables, sentinels are
matched at the image width, and target-derived names are bounded and
marked untrusted.

* fix(native): keep ELF initializer slots unresolved when relocation coverage is partial

A skipped, unreadable or over-budget section relocation table now keeps
every slot it could reach unresolved instead of reading the stored word as
raw or dropping a .ctors slot as a sentinel. Dynamic writes that only
partly overlap a slot, or write it twice, leave it unresolved. Linked
images no longer reapply retained non-SHF_ALLOC link-time records.

The dynamic relocation walk now pauses at the shared reader's step
boundaries and observes cancellation there, resolved targets need only
their first byte mapped, and unreadable symbol names make the report
incomplete.

* fix(native): stop reporting unchecked Mach-O initializer words as resolved

Object files now apply their own section relocations: an UNSIGNED entry
against a defined symbol gives symbol_addend, one against an undefined
symbol gives unresolved_import, and any other entry leaves the slot
unresolved. Linked images without a fixup decoder, such as 32-bit ones,
keep pointer slots unresolved when an LC_DYLD_INFO bind stream or
LC_DYSYMTAB external relocations could replace the stored word.

LC_DYLD_INFO rebases now report dyld_info_rebase instead of
chained_fixup, and unreadable symbol names make the report incomplete.
Section headers keep their declared reloff and nreloc for this.

* docs(native): move initializer guide under docs/guides/native

The documentation guide assigns analyst workflows to docs/guides/<domain>/.
Move the guide there, place its navigation and directory entries in the
Native binaries group, and update links from the docs index, CLI guide,
spec and authored skill. The public /native-initializers route is
unchanged.

The guide and spec also describe per-slot relocation coverage, retained
link-time records, Mach-O object relocations, undecoded bind streams,
the dyld_info_rebase resolution and unreadable names.

* fix(native): keep IFUNC and partly relocated ELF initializer slots unresolved

An absolute relocation against an STT_GNU_IFUNC symbol no longer reports
the resolver address as the target. Section relocation records now use
their real write width, so a record that writes only part of a slot,
starts inside it or repeats it leaves the slot unresolved instead of
reading a raw zero or dropping a .ctors slot as a sentinel.

On 32-bit ARM the Thumb bit is cleared when locating a target's section
and file offset, while value_va keeps the stored callable value.

* fix(native): locate Thumb Mach-O initializer targets and reject arm64_32 with a 64-bit header

On CPU_TYPE_ARM the Thumb bit is cleared when locating and naming a
target, since nlist values do not carry it. An arm64_32 image that
declares a 64-bit header is rejected as malformed rather than read with
8-byte slots. The guide and spec describe IFUNC relocations, partly
relocated section slots, Thumb targets and this rejection.

* fix(native): index initializer address lookups and keep ambiguous mappings unknown

ELF and Mach-O initializer lookups now build the shared file mapping
index once per image, so each slot and target lookup is logarithmic in
the number of load segments rather than a scan of every mapping.

An address that load segments map ambiguously no longer falls back to a
section header offset. Its file offset is reported as unknown and the
report is incomplete; only genuinely unmapped spans use the section
fallback.

* fix(native): report Mach-O bind names over the name budget

A bound import name longer than the 1024-byte name budget was dropped
without a trace, although it is the only identity of an unresolved
initializer target. It now counts toward the over-long name limitation
and makes the report incomplete.

* test(native): count mapping lookups instead of timing the load index regression

The load-mapping regression asserted a wall-clock bound, which is
sensitive to machine load. It now spies on the shared mapping module and
requires the same number of mapping scans and index builds for 8 and 64
slots, so a per-slot scan of every load segment fails deterministically
on a small fixture.

* fix(native): yield while decoding legacy rebase runs

* fix(native): preserve initializer mapping and coverage

* fix(native): index legacy fixup spans

* fix(native): preserve relocation uncertainty and bound bind lookup work

* fix(native): validate ELF slot mappings and yield section relocation scans

---------

Co-authored-by: Matt Van Horn <455140+mvanhorn@users.noreply.github.com>
Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>
2026-11-09 08:58:12 -05:00