[hypermoist]

What

These 29 exports were NotImplemented stubs. In the firmware modules (system/common/lib/<library>.sprx, firmware 13.60, read from disassembly, not measured on a console), each is a single constant return that reads no arguments and makes no IPC call or syscall, so the exports now return the same value. This follows #8303, #8305 and #8307. The names are removed from the pending-declaration lists in each library's debt file.

Library	Export	NID	Offset in the module	Firmware body	Now
libSceComposite	sceCompositorCreateIndirectRenderTarget	meXSAmyPBUE	0x6290	mov eax, 0x80d40006; ret	0x80d40006;
libSceComposite	sceCompositorDeleteIndirectRenderTarget	wCqghwW7vQw	0x62a0	mov eax, 0x80d40006; ret	0x80d40006;
libSceComposite	sceCompositorSetIndirectCanvasCommand	-zpfGY1n49E	0x1870	xor eax, eax; ret	0
libSceComposite	sceCompositorSetIndirectCompositionCommand	M6wp6+JUSRI	0x1850	xor eax, eax; ret	0
libSceComposite	sceCompositorSetIndirectRenderTargetConfigCommand	kZ7Xre07w70	0x1860	xor eax, eax; ret	0
libSceFsInternalForVsh	sceFsErriCtl	iLxIuu5Xs2A	0xf5f0	mov eax, 0x8002002d; ret	0x8002002d;
libSceFsInternalForVsh	sceFsErriGetState	-1whCDzGK1E	0x181c0	mov eax, 0x8002002d; ret	0x8002002d;
libSceHttpCache	sceHttpCacheDebugSetUrl	FME0+DPciaU	0x3a5e0	xor eax, eax; ret	0
libScePlayReady	scePlayReadyMemGetStats	iktpef6dl3k	0x1db0	mov eax, 0x80004001; ret	0x80004001;
libSceSdma	sceSdmaInitialize	Xc+bwbBPK00	0x26b0	xor eax, eax; ret	0
libSceSdma	sceSdmaFinalize	mNFbFFiI-7A	0x2960	xor eax, eax; ret	0
libSceSdma	sceSdmaQueryWorkingBufferSize	1FH-82oV3PY	0xd0	mov eax, 0x8000; ret	0x8000;
libSceSdma	sceSdmaFlushAllWithEvent	uHobIZSa1dI	0x2ab0	mov eax, 0x809900c9; ret	0x809900c9;
libSceSdma	sceSdmaWaitForEvent	rzfOSPn1w7k	0x2ac0	mov eax, 0x809900c9; ret	0x809900c9;
libSceSsl	CRYPTO_uninitAsymmetricKey	cW7VCIMCh9A	0x38d0	xor eax, eax; ret	0
libSceSsl	VLONG_freeVlongQueue	wcVuyTUr5ys	0x38e0	xor eax, eax; ret	0
libSceVisionManager	sceVisionManagerOpen	wzr6GTOo9Dw	0x162b0	mov eax, 0x80e70004; ret	0x80e70004;
libSceVisionManager	sceVisionManagerClose	P9+90HStC44	0x162c0	mov eax, 0x80e70004; ret	0x80e70004;
libSceVisionManager	sceVisionManagerRequestFindUser	F8aomqtpD98	0x16600	mov eax, 0x80e70004; ret	0x80e70004;
libSceVisionManager	sceVisionManagerRequestFindUserInRegion	iBbwD8vxSEk	0x16620	mov eax, 0x80e70004; ret	0x80e70004;
libSceVisionManager	sceVisionManagerRequestFindUserWithoutIdentify	CZgPsIdPqSk	0x165c0	mov eax, 0x80e70004; ret	0x80e70004;
libSceVisionManager	sceVisionManagerRequestRegisterUserInRegion	-RJmkkuBhaY	0x165e0	mov eax, 0x80e70004; ret	0x80e70004;
libSceVisionManager	sceVisionManagerGetResultOfFindUser	Nuhv71YjC+g	0x16610	mov eax, 0x80e70004; ret	0x80e70004;
libSceVisionManager	sceVisionManagerGetResultOfFindUserInRegion	Gop3lku7E+Q	0x16630	mov eax, 0x80e70004; ret	0x80e70004;
libSceVisionManager	sceVisionManagerGetResultOfFindUserWithoutIdentify	ES+vLK-z0+U	0x165d0	mov eax, 0x80e70004; ret	0x80e70004;
libSceVisionManager	sceVisionManagerGetResultOfRegisterUserInRegion	GYamKe-3LbY	0x165f0	mov eax, 0x80e70004; ret	0x80e70004;
libSceVisionManager	sceVisionManagerLoadSettings	l9zev-ybPNs	0x178a0	xor eax, eax; ret	0
libmdbg_syscore	sceDebugGetProcessEventCntlFlag	Zi+rAm6czUg	0x4af0	mov eax, 1; ret	1;
libmdbg_syscore	sceDebugSetProcessEventCntlFlag	--I4Ml0ADxQ	0x4b00	mov eax, 1; ret	1;

Left out:

• sceCompositorGetDummyAddress and sceCompositorGetDummySize: the first returns a fixed address in the console's address space.
• Three libSceSaveData constants (sceSaveDataDebugGetThreadId, sceSaveDataGetClientThreadPriority, sceSaveDataGetSavePoint): fix(savedata): build libSceSaveData and libSceSaveDataDialog from the native sources #7780 rewrites that file.

Tested

Linux (CachyOS, GCC), merged with current main: the eight libraries build. The new guest_firmware_constant_returns test (core/libs/tests/GuestFirmwareConstantReturns.cpp) calls all 29 exports and checks each exact return; it passes, as do the 10 existing tests matching these libraries' names. check_conventions passes. pr_overlap could not compare #6039, #8541, #8951 and #9137; I checked their hunks by hand, and none touches these functions or the debt lines edited here. No title was run.

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

🤖 Generated with Claude Code

https://claude.ai/code/session_01RbLgECLQ2bpDM7oe2urGZK

[github-actions]

Progress report (4017310 → f2f5267)

📈 System libraries: 49.84% (+0.23%, +29 implemented functions)

✅ 29 implemented

Library	Name
libSceComposite	sceCompositorCreateIndirectRenderTarget
libSceComposite	sceCompositorDeleteIndirectRenderTarget
libSceComposite	sceCompositorSetIndirectCanvasCommand
libSceComposite	sceCompositorSetIndirectCompositionCommand
libSceComposite	sceCompositorSetIndirectRenderTargetConfigCommand
libSceFsInternalForVsh	sceFsErriCtl
libSceFsInternalForVsh	sceFsErriGetState
libSceHttpCache	sceHttpCacheDebugSetUrl
libScePlayReady	scePlayReadyMemGetStats
libSceSdma	sceSdmaFinalize
libSceSdma	sceSdmaFlushAllWithEvent
libSceSdma	sceSdmaInitialize
libSceSdma	sceSdmaQueryWorkingBufferSize
libSceSdma	sceSdmaWaitForEvent
libSceSsl	CRYPTO_uninitAsymmetricKey
libSceSsl	VLONG_freeVlongQueue
libSceVisionManager	sceVisionManagerClose
libSceVisionManager	sceVisionManagerGetResultOfFindUser
libSceVisionManager	sceVisionManagerGetResultOfFindUserInRegion
libSceVisionManager	sceVisionManagerGetResultOfFindUserWithoutIdentify
libSceVisionManager	sceVisionManagerGetResultOfRegisterUserInRegion
libSceVisionManager	sceVisionManagerLoadSettings
libSceVisionManager	sceVisionManagerOpen
libSceVisionManager	sceVisionManagerRequestFindUser
libSceVisionManager	sceVisionManagerRequestFindUserInRegion
libSceVisionManager	sceVisionManagerRequestFindUserWithoutIdentify
libSceVisionManager	sceVisionManagerRequestRegisterUserInRegion
libmdbg_syscore	sceDebugGetProcessEventCntlFlag
libmdbg_syscore	sceDebugSetProcessEventCntlFlag

⚪ 12 behaviourally empty (constant return or throw)

Library	Name
libSceComposite	sceCompositorSetIndirectCanvasCommand
libSceComposite	sceCompositorSetIndirectCompositionCommand
libSceComposite	sceCompositorSetIndirectRenderTargetConfigCommand
libSceHttpCache	sceHttpCacheDebugSetUrl
libSceSdma	sceSdmaFinalize
libSceSdma	sceSdmaInitialize
libSceSdma	sceSdmaQueryWorkingBufferSize
libSceSsl	CRYPTO_uninitAsymmetricKey
libSceSsl	VLONG_freeVlongQueue
libSceVisionManager	sceVisionManagerLoadSettings
libmdbg_syscore	sceDebugGetProcessEventCntlFlag
libmdbg_syscore	sceDebugSetProcessEventCntlFlag

Changed libraries (implemented exports)

Library	Before	After	Change
libmdbg_syscore	0/143 (0.00%)	2/143 (1.40%)	+1.40 pp
libSceComposite	0/78 (0.00%)	5/78 (6.41%)	+6.41 pp
libSceFsInternalForVsh	0/176 (0.00%)	2/176 (1.14%)	+1.14 pp
libSceHttpCache	2/27 (7.41%)	3/27 (11.11%)	+3.70 pp
libScePlayReady	0/50 (0.00%)	1/50 (2.00%)	+2.00 pp
libSceSdma	0/26 (0.00%)	5/26 (19.23%)	+19.23 pp
libSceSsl	33/55 (60.00%)	35/55 (63.64%)	+3.64 pp
libSceVisionManager	0/56 (0.00%)	11/56 (19.64%)	+19.64 pp

[github-actions]

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/CMakeLists.txt
PR 6039	docs/dev/debt/libSceSsl.md
PR 7201	core/libs/CMakeLists.txt
PR 7275	core/libs/CMakeLists.txt
PR 7780	core/libs/CMakeLists.txt
PR 8285	core/libs/CMakeLists.txt
PR 8650	core/libs/CMakeLists.txt
PR 9284	core/libs/CMakeLists.txt
PR 9293	core/libs/CMakeLists.txt
PR 9297	core/libs/CMakeLists.txt
PR 9302	core/libs/CMakeLists.txt
PR 9535	core/libs/CMakeLists.txt
PR 9536	core/libs/CMakeLists.txt
PR 9571	core/libs/CMakeLists.txt
PR 9822	core/libs/prx/libSceSsl/Export.cpp
PR 9826	core/libs/prx/libSceSsl/Export.cpp
PR 10167	core/libs/prx/libSceSsl/Export.cpp

[MarcZX10]

Review

The implementations are small and consistent with the stated firmware behavior. I did not find a source-level defect in the constant returns. I’m requesting changes because the submitted test counts do not exercise these new exports, and the firmware evidence is not yet specific enough to verify all 29 values.

Direct regression coverage

I searched core/libs/tests for the 29 export names in this change and found no direct references. The reported 24 library tests passing therefore do not demonstrate that these functions now return the claimed values.

Please add focused assertions that call each export and check its return constant (grouping them by library is fine). This will also ensure the changed definitions are linked and callable. Please rerun those tests after updating the branch.

Firmware evidence

The description identifies firmware 12.00 and says these bodies were read from disassembly, but the table only provides the inferred instruction shape and result. Please add concise verification context for each group (module and function offsets or an attached disassembly excerpt) so reviewers can check the 29 constant-return claims. The debt notes should continue to say these behaviors have not been measured on console.

Base and overlap

The head is 104 commits behind current main; the three-way merge check is clean, but CONTRIBUTING asks branches to stay up to date. Please rebase/update before merge and rerun the focused library tests.

The overlap report lists #6039 in docs/dev/debt/libSceSsl.md. I checked its hunk: it edits a separate entry, so I found no dependency or merge-order issue there. The other overlap PRs you checked are likewise described as separate hunks.

Checks

• git diff --check against current main: passed.
• Current conventions check: zero errors (five debt-length warnings in touched library debt files); hosted conventions, eligibility, labels, and report jobs passed.
• Hosted Linux, Windows, register-context, relinker, ASan relinker, and write-watch jobs were skipped.
• I did not build or run the library tests locally in this session. The 24 passing results in the description are author-reported and were run on the earlier base.

The firmware-derived return values are a reasonable implementation direction; direct tests and reviewable source evidence are still needed before approval.

[BunnyFl]

Checked these 29 against firmware 13.42 as well: each is the same constant return as in your description, and the code returns those values.

AI-assisted: yes (Claude Code)

[hypermoist]

@MarcZX10 thanks, all three addressed:

• Tests: guest_firmware_constant_returns (core/libs/tests/GuestFirmwareConstantReturns.cpp, registered in core/libs/CMakeLists.txt with the Windows DLL paths of all eight libraries) calls each of the 29 exports and checks its exact return, grouped by library. It passes on Linux, as do the existing tests for those libraries.
• Evidence: the description table now gives each export's NID and function offset in its .sprx (firmware 12.00) next to its body. It still says the values are read from disassembly and not measured on a console, and the debt files don't claim otherwise.
• Base: I merged current main (7b1c9ca) before running the tests.

[hypermoist]

Merged main (ea8fb12) to clear the core/libs/CMakeLists.txt conflict, keeping the constant-returns test after main's new barrier and pthread tests. Rebuilt fresh: guest_firmware_constant_returns passes on Linux and under Wine (WinLibs 15.2), and all eight patched .prx files build.

Assisted With AI: Yes But all test done locally on actual hardware

[hypermoist]

Correction: I mislabelled the firmware. The decrypted modules I read are firmware 13.60, not 12.00 (each module's module-param segment records SDK 13.60). The behaviour described is unchanged; only the version label was wrong. I've updated the description.

Assisted With AI: Yes But all test done locally on actual hardware

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]

• Rebased onto current main: resolved the core/libs/CMakeLists.txt test-block conflict (both blocks kept) and the libSceHttpCache debt conflict (main now returns zero from sceHttpCacheSystemSendStatistics/sceHttpCacheSystemTerm; sceHttpCacheDebugSetUrl joins that line and leaves the pending list).
• guest_firmware_constant_returns still pins all 29 values; core/libs tests pass on the rebased head; merged.

[zaryabzip]

@hypermoist Some exports your sweeps measured as fixed returns but left out still throw on main, and the reasons for leaving them out have gone:

• fix(libs): return what the firmware returns from 29 more calls it does not implement #9173's three libSceSaveData constants: sceSaveDataDebugGetThreadId, sceSaveDataGetClientThreadPriority and sceSaveDataGetSavePoint. They were left out because fix(savedata): build libSceSaveData and libSceSaveDataDialog from the native sources #7780 rewrote that file, and fix(savedata): build libSceSaveData and libSceSaveDataDialog from the native sources #7780 has merged.
• fix(libs): return what the firmware returns from 31 more calls it does not implement #8307's sceHmdInternalRegisterEventCallback: left out for feat(libSceHmd): answer ten buffer and device queries the way the disconnected-headset path implies #8256, which merged without touching it.
• sceCompositorGetDummySize: if it's a plain constant, unlike the address.

Would you like to take them? If not, could you post the value each one returns, and the firmware you read it on? Then I'll put them in a small PR citing your reading.