[orezraey]

What

Thread-attribute edge cases measured on a retail PS5 (firmware 13.60) with three homebrew probes calling libkernel_sys.sprx from a payload process. The sources and console logs are in this gist (pcond, pcond2, pattr).

The condition-variable change moved to #10177, as requested in review.

scePthreadAttrGet reports the inheritance a thread was created with: 0 for an explicit-scheduling thread and 4 for an inheriting one or one created from a freshly initialised attribute. Before, it always reported 4. pthread_attr_get_np now copies the same inheritance, as FreeBSD's copies the whole attribute; before, it left the caller's value. A null attribute at creation was not measured. The policy and priority it reports already matched:

• the creator's for an inheriting thread (1 and 700 here, even when the attribute said round-robin 700);
• the attribute's for an explicit one.

scePthreadAttrGetinheritsched with a null output returns EINVAL, as scePthreadAttrGetschedpolicy does. Before, it threw. Both also return EINVAL for a null or uninitialized attribute, which AnyPS5 already did. A new attribute holds policy 1, inheritance 4 and priority 700, also as before.

Already matching, so nothing in this diff changes them:

• scePthreadCondTimedwait returns EPERM (0x80020001) for an unowned mutex of each type (default, error-checking, recursive, normal, adaptive), the adaptive static initializer (slot 1), a null slot and a mutex held by another thread.
• It returns EINVAL for the destroyed mutex (slot 2).
• It times out with 0x8002003c, also for 0 µs.
• A recursive mutex locked twice is released completely during the wait (another thread's trylock succeeds) and relocked to depth 2.

The debt entries are updated.

Not modelled: a slot of 2 or 3 crashed the console's condition calls, and so does AnyPS5, which dereferences it. pthread_attr_get_np returned EINVAL for the payload's main thread, so its answer for other threads stays unmeasured.

Tested

Windows 11, MinGW (winlibs), and Linux (Ubuntu on WSL 2, GCC), at head 3345e2047d (main merged in after #9573, keeping its createSuspended field next to this PR's inheritance).

• guest_pthread_attr checks the inheritance scePthreadAttrGet and pthread_attr_get_np report for an explicit and an inheriting thread.
• guest_pthread_attr_errors checks EINVAL for a null scePthreadAttrGetinheritsched output.
• These are the only tests this diff changes. Both fail with main's Pthread sources and pass with this branch's.
• The thread and attribute tests pass three runs in a row on both systems: guest_pthread_attr, guest_pthread_attr_errors, guest_pthread_attr_stack, guest_pthread_create_suspended, guest_pthread_schedparam, guest_pthread_start, guest_pthread_join_detach and guest_kernel_sys_thread_defaults.
• The condition-variable tests, including guest_pthread_destroyed_objects, belong to fix(libkernel): mark a destroyed condition variable with 1 and accept destroying a null one, as measured on a PS5 #10177 and are reported there.
• tools/check_conventions.py passes.

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 (feat(libkernel_sys): create threads suspended and resume them #9573, which edited the same lines, has merged and this PR is refreshed over it)
• Fixes: none
• Unblocks: none
• GPUs tested: n/a
• AI-assisted: yes

[MarcZX10]

The PS5 probe evidence supports the destroyed condition marker, null-destroy behavior, inherited scheduling query, and null-output error changes; GCC 12.2 syntax checks passed for the changed implementation and tests, as did conventions, diff-check, and main-sync (5 ahead/29 behind; merge clean). The author reports the three directly affected tests and 14 related Windows thread/condition tests passing three times. Please split the condition-variable changes (Cond.cpp and destroyed-object tests) from the independent thread-attribute changes (Attr.cpp, inheritance state/query tests) into separate topic branches, as required by the repository’s one-branch-per-topic rule. The attribute portion also conflicts with open PR #10133 in docs/dev/debt/libkernel.md; coordinate a single combined entry and landing order with that author, preserving both sets of measured attribute behavior. Add Depends on if either branch needs the other first, then refresh and rerun focused tests and checks. Hosted policy checks pass; product jobs are skipped by the workflow.

[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 6740	docs/dev/debt/libkernel.md L78-80

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 3986	core/libs/prx/libkernel/Pthread/src/Thread.cpp
PR 4563	core/libs/prx/libkernel/Pthread/include/Pthread.hpp core/libs/prx/libkernel/Pthread/src/Thread.cpp
PR 4630	docs/dev/debt/libkernel.md
PR 5021	core/libs/prx/libkernel/Pthread/include/Pthread.hpp core/libs/prx/libkernel/Pthread/src/Thread.cpp docs/dev/debt/libkernel.md
PR 6149	docs/dev/debt/libkernel.md
PR 6936	docs/dev/debt/libkernel.md
PR 7088	docs/dev/debt/libkernel.md
PR 7105	docs/dev/debt/libkernel.md
PR 7121	docs/dev/debt/libkernel.md
PR 7149	core/libs/prx/libkernel/Pthread/include/Pthread.hpp core/libs/prx/libkernel/Pthread/src/Thread.cpp docs/dev/debt/libkernel.md
PR 7192	docs/dev/debt/libkernel.md
PR 7201	core/libs/prx/libkernel/Pthread/src/Thread.cpp
PR 7447	docs/dev/debt/libkernel.md
PR 7509	docs/dev/debt/libkernel.md
PR 7804	core/libs/prx/libkernel/Pthread/include/Pthread.hpp core/libs/prx/libkernel/Pthread/src/Thread.cpp docs/dev/debt/libkernel.md
PR 8010	docs/dev/debt/libkernel.md
PR 8592	core/libs/prx/libkernel/Pthread/include/Pthread.hpp core/libs/prx/libkernel/Pthread/src/Thread.cpp
PR 8605	docs/dev/debt/libkernel.md
PR 9156	docs/dev/debt/libkernel.md
PR 9199	docs/dev/debt/libkernel.md
PR 9284	docs/dev/debt/libkernel.md
PR 9293	docs/dev/debt/libkernel.md
PR 9297	docs/dev/debt/libkernel.md
PR 9302	docs/dev/debt/libkernel.md
PR 9389	docs/dev/debt/libkernel.md
PR 9491	docs/dev/debt/libkernel.md
PR 9509	docs/dev/debt/libkernel.md
PR 9557	docs/dev/debt/libkernel.md
PR 9566	docs/dev/debt/libkernel.md
PR 9575	docs/dev/debt/libkernel.md
PR 9947	docs/dev/debt/libkernel.md
PR 10105	docs/dev/debt/libkernel.md
PR 10179	docs/dev/debt/libkernel.md
PR 10385	core/libs/prx/libkernel/Pthread/include/Pthread.hpp docs/dev/debt/libkernel.md
PR 10442	docs/dev/debt/libkernel.md
PR 10552	docs/dev/debt/libkernel.md
PR 10621	docs/dev/debt/libkernel.md
PR 10848	docs/dev/debt/libkernel.md
PR 11085	docs/dev/debt/libkernel.md
PR 11264	docs/dev/debt/libkernel.md
PR 11267	docs/dev/debt/libkernel.md

[BunnyFl]

I checked the condition-variable and scePthreadAttrGetinheritsched parts against the PS5 firmware 13.42 libkernel and libkernel_sys modules, running them in an emulator (head e2b8771). Both modules give the same results, and they match this PR.

• scePthreadCondDestroy on a condition object stores 1 in the slot and returns 0. On a null slot it returns 0 and leaves the slot null.
• With the slot holding 1, scePthreadCondDestroy, scePthreadCondSignal, scePthreadCondBroadcast, scePthreadCondWait and scePthreadCondTimedwait all return 0x80020016 and leave the slot at 1.
• scePthreadAttrGetinheritsched returns 0x80020016 for a null attribute, an uninitialized one and a null output. Otherwise it writes the stored inheritance and returns 0.

I didn't run scePthreadAttrGet: its copy of the thread's attribute uses AVX instructions the emulator doesn't support.

AI-assisted: yes (Claude Code).

[orezraey]

@MarcZX10 Split as requested: the condition-variable change is now #10177. For the attribute debt conflict, #10133 now carries this branch and combines both entries, so this one lands first. @BunnyFl thanks for the 13.42 check; #10177 cites it.

[MarcZX10]

Re-review of current head 886a2b249cf0dcbe75f954e89dbdaadea960e1ef. The condition-variable implementation and its test were successfully moved out to #10177, leaving the thread-attribute behavior as this PR's topic. #10133 now declares Depends on: #10130, so the debt-file landing order is recorded. The remaining review issue is that this PR's Tested section still attributes guest_pthread_destroyed_objects and the condition-variable cases to this change, and says all three affected tests fail on main and pass with it. Those condition changes are no longer in this PR's file diff; please update the test report to identify the attribute tests actually changed here and distinguish any checks that now belong to #10177. The focused attribute evidence can remain; the report should match the exact current diff. Hosted product jobs are path-skipped.

[orezraey]

@MarcZX10 Updated the Tested section to the current head 886a2b249c. It now lists only guest_pthread_attr and guest_pthread_attr_errors, the two tests this diff changes; I rebuilt with main's Pthread sources to confirm both fail there and pass with the branch. The condition-variable tests, including guest_pthread_destroyed_objects, are now reported only on #10177. The scePthreadCondTimedwait notes in What are now labelled as already matching, since nothing here changes them.

[orezraey]

@zaryabzip: #9573 changes the same lines as this PR in core/libs/prx/libkernel/Pthread/Posix/Attr.cpp and core/libs/prx/libkernel/Pthread/src/Attr.cpp (git merge-tree of both heads over current main conflicts there). I propose it lands first; I will then merge main into this branch, keeping both sides' behaviour, and rerun its tests. Depends on now lists it (merge order only).

Dismissing a stale review: the requested changes were addressed after it (new head) and the merge queue blocks on it. Please request changes again if anything is still open.

[adriadam10]

Checked the inheritance reported by scePthreadAttrGet and pthread_attr_get_np (explicit 0, inheriting/default 4), the EINVAL for a null scePthreadAttrGetinheritsched output, both tests (they fail on main's sources) and the debt entries against the measurements in the description; the condition-variable change is gone to #10177 and the Tested section matches the head; merged.