[xdCloudy]

What

Implement the pwrite, futimesat and symlinkat exports in libkernel_sys. pwrite forwards to libkernel's existing positioned-write implementation. futimesat resolves paths with the supplied directory descriptor and uses libkernel's native timestamp implementation. symlinkat shares the existing symlink implementation and path resolution.

Firmware 13.60 evidence: pwrite at libkernel_sys VA 0x13630 passes the four POSIX arguments to syscall 476. FreeBSD 11.4 syscall definitions support the futimesat and symlinkat prototypes. The current symlink limitations remain: absolute targets and targets containing .. throw as unimplemented. Directory-relative host paths also do not model guest aliases or constrain .. to the guest root. Console results remain unverified; details are recorded in docs/dev/debt/libkernel_sys.md.

Tested

Windows with WinLibs GCC 15.2: configured the full project, built guest_kernel_sys_tests, and passed ctest --test-dir build-win -R '^guest_kernel_sys$' --output-on-failure. The test covers all three exports and error cases.

python3 tools/check_conventions.py --base origin/main passed. python3 tools/pr_overlap.py --branch feat/libkernel-sys-posix-file-ops scanned 1,410 open PRs and found no overlap. Linux was not tested.

Checklist

• Based on current main; no other open PR implements the same functions
• One topic per PR; follow-ups go in a new PR
• No comments in code except technical debt
• Unimplemented paths throw (NotImplemented_nid_no_patch); silent stubs are listed under Silent stubs in their library's technical debt file
• No notes, investigation .md files or images added to the repository (attach them to this PR)
• New third-party code is a submodule built from source
• Generic behaviour, not specific to one title
• Depends on: none
• Fixes: none
• Unblocks: none
• GPUs tested: n/a
• AI-assisted: yes

[github-actions]

Progress report (3a2b919 → e4044fe)

📈 System libraries: 45.92% (+0.01%, +2 implemented functions)

✅ 2 implemented

Library	Name
libkernel_sys	futimesat_nid_postfix
libkernel_sys	pwrite_nid_postfix

Changed libraries (implemented exports)

Library	Before	After	Change
libkernel_sys	116/514 (22.57%)	118/514 (22.96%)	+0.39 pp

[rooklz]

The checklist says no other open PR implements these functions, but two older open PRs do:

• feat(libkernel_sys): implement statfs symlinkat and futimesat #9161 (yabanana; refreshed on main and retested on Windows on 28 Oct) implements futimesat and symlinkat and makes the same symlink split: one shared helper for symlink and symlinkat, resolving through GetDirectoryDescriptorPath_nid_no_patch for a relative path with a descriptor and through ResolvePathError_nid_no_patch for AT_FDCWD or an absolute path. Its futimesat is PosixUtimesAt_nid_no_patch in Stdio.cpp and has the same InvalidTimes → EINVAL check and the same NotFoundError_nid_no_patch mapping.
• feat(libkernel_sys): implement pwrite through the shared POSIX provider #9019 (yabanana) implements pwrite through the shared POSIX provider, and its thread with @Felipebmaitan (feat(libkernel): cancel blocked POSIX writes safely #11969) sets out the cancellation and descriptor-admission order that pwrite has to follow.

I ran python3 tools/pr_overlap.py --branch locally on this head (e4044fef46). It reports Same exports pwrite for #9019 and futimesat/symlink/symlinkat for #9161, plus conflicting lines in Export.Discovery.cpp L68-72/L2154-2158, Export.cpp L772-788/L817 and Api.hpp L102/L547. git merge-tree of this head with #9161 conflicts. Against main 3a2b919816 it is clean. I didn't build or run anything; this is a source comparison only.

Suggestion: drop the three exports from this PR, or close it and offer the parts #9161 lacks there. The readlinkat round-trip, the second-symlinkat EEXIST row and the missing-parent ENOENT row look like test rows #9161 could take. One behaviour difference to settle if they move over: here ENAMETOOLONG is checked only on the descriptor-relative branch (inside ResolveAtPath), but #9161 checks it before choosing a branch, so AT_FDCWD and absolute paths are covered too.

[adriadam10]

Thanks for the work here. Closing because the whole PR duplicates older open PRs:

• futimesat and symlinkat (and the shared symlink split) are implemented in feat(libkernel_sys): implement statfs symlinkat and futimesat #9161.
• pwrite is implemented in feat(libkernel_sys): implement pwrite through the shared POSIX provider #9019.

tools/pr_overlap.py reports the same exports for both, and this head conflicts with #9161 in Export.Discovery.cpp, Export.cpp and Api.hpp, so the overlap can't be removed from this PR.

Some of your test rows look useful for #9161 (readlinkat round-trip, second symlinkat giving EEXIST, missing parent giving ENOENT), and ENAMETOOLONG should be checked before choosing the descriptor-relative branch, as #9161 does. Feel free to offer them there, and to reopen this if those PRs are dropped.

[github-actions]

Other open pull requests touch the same code. Merging one will break the other, or both implement the same thing. If this one builds on them, list them in Depends on.

Pull request	Conflicting files (lines on main)	Same exports	Same new files
PR 9019		pwrite
PR 9161	core/libs/prx/libkernel_sys/Export.Discovery.cpp L68-72, L2154-2158 core/libs/prx/libkernel_sys/Export.cpp L772-776, L787-788, L817 core/libs/prx/libkernel_sys/include/Api.hpp L102, L547	futimesat symlink symlinkat
PR 9162	docs/dev/debt/libkernel_sys.md L14
PR 10332	core/libs/tests/GuestKernelSys.cpp L901
PR 10625	core/libs/prx/libkernel_sys/Export.cpp L781-788

Some comparisons could not be completed because a pull request conflicts with the base branch. Review these shared files manually before concluding that there is no overlap.

Pull request	Shared files not compared
PR 5021	core/libs/prx/libkernel/File/src/Stdio.cpp core/libs/prx/libkernel_sys/Export.cpp
PR 6936	core/libs/prx/libkernel/File/src/Stdio.cpp
PR 7088	core/libs/prx/libkernel/File/include/File.hpp
PR 7121	core/libs/prx/libkernel/File/src/Stdio.cpp
PR 7447	core/libs/prx/libkernel/File/src/Stdio.cpp
PR 9019	core/libs/prx/libkernel/File/include/File.hpp core/libs/prx/libkernel/File/src/Stdio.cpp core/libs/prx/libkernel_sys/Export.Discovery.cpp core/libs/prx/libkernel_sys/Export.cpp core/libs/prx/libkernel_sys/include/Api.hpp docs/dev/debt/libkernel_sys.md
PR 9097	core/libs/prx/libkernel_sys/Export.Discovery.cpp core/libs/prx/libkernel_sys/Export.cpp core/libs/prx/libkernel_sys/include/Api.hpp core/libs/tests/GuestKernelSys.cpp
PR 9156	core/libs/prx/libkernel_sys/Export.Discovery.cpp
PR 9242	core/libs/prx/libkernel_sys/Export.Discovery.cpp core/libs/prx/libkernel_sys/include/Api.hpp
PR 9293	core/libs/prx/libkernel/File/src/Stdio.cpp core/libs/prx/libkernel_sys/Export.Discovery.cpp
PR 9297	core/libs/prx/libkernel/File/src/Stdio.cpp core/libs/prx/libkernel_sys/Export.Discovery.cpp docs/dev/debt/libkernel_sys.md
PR 9302	core/libs/prx/libkernel/File/src/Stdio.cpp core/libs/prx/libkernel_sys/Export.Discovery.cpp docs/dev/debt/libkernel_sys.md
PR 9383	core/libs/prx/libkernel_sys/Export.Discovery.cpp
PR 9389	core/libs/prx/libkernel_sys/Export.Discovery.cpp core/libs/prx/libkernel_sys/Export.cpp
PR 10657	core/libs/prx/libkernel_sys/Export.Discovery.cpp docs/dev/debt/libkernel_sys.md
PR 11031	core/libs/prx/libkernel_sys/Export.Discovery.cpp core/libs/prx/libkernel_sys/Export.cpp core/libs/prx/libkernel_sys/include/Api.hpp core/libs/tests/GuestKernelSys.cpp docs/dev/debt/libkernel_sys.md
PR 11034	core/libs/prx/libkernel_sys/Export.Discovery.cpp core/libs/prx/libkernel_sys/Export.cpp core/libs/prx/libkernel_sys/include/Api.hpp core/libs/tests/GuestKernelSys.cpp docs/dev/debt/libkernel_sys.md
PR 11212	core/libs/prx/libkernel_sys/Export.Discovery.cpp docs/dev/debt/libkernel_sys.md
PR 11216	core/libs/prx/libkernel_sys/Export.Discovery.cpp core/libs/prx/libkernel_sys/Export.cpp core/libs/prx/libkernel_sys/include/Api.hpp core/libs/tests/GuestKernelSys.cpp docs/dev/debt/libkernel_sys.md
PR 11625	core/libs/prx/libkernel/File/src/Stdio.cpp
PR 11969	core/libs/prx/libkernel/File/include/File.hpp core/libs/prx/libkernel/File/src/Stdio.cpp

[xdCloudy]

@rooklz, @adriadam10 Thanks for identifying the duplicate exports. I agree with the closure: #12687’s futimesat, symlinkat and pwrite work duplicated #9161/#9019, so I have not reopened or reintroduced those implementations. I checked the current #9161 test patch: it still lacks the suggested readlinkat round-trip, duplicate-symlinkat EEXIST, and missing-parent ENOENT cases; the reviewer’s long-path concern is already handled before the path branch is selected. I have not added a test-only PR against the still-unmerged implementation.