[kallmek3nny]

What

sceUserServiceGetNpAccountId now returns USER_SERVICE_ERROR_INVALID_ARGUMENT (0x80960005) for a null output before it checks whether the user is logged in. Before, a user who is not logged in got USER_SERVICE_ERROR_UNKNOWN_USER (0x80960105) even with a null output. With a non-null output nothing changes: a user who is not logged in still gets 0x80960105 with the output untouched, and a logged-in user still gets account id 0.

Behaviour source: firmware 13.42's libSceUserService.sprx, read from its disassembly and run natively on x86-64 with its imports stubbed, not measured on a console: #12165 (comment). Its handler returns 0x80960002 while the module is not initialized, then 0x80960005 for a null output, and only after that looks at the user; once initialized, null-output calls returned 0x80960005 for users 0x10000000, 0x10000001, 1, 123, 255, -1 and 0. The non-null result for a user who is not logged in stays as measured on a PS4 Pro (firmware 10.71). The not-initialized check that comes first is not modelled; the sceUserServiceGetNpAccountId entry in docs/dev/debt/libSceUserService.md now names this source and that gap.

The four null-output calls in core/libs/tests/GuestUserService.cpp for users that are not logged in now expect USER_SERVICE_ERROR_INVALID_ARGUMENT instead of USER_SERVICE_ERROR_UNKNOWN_USER.

Tested

Microsoft Windows [Version 10.0.26300.9711], gcc 16.1.0 (MinGW-W64 x86_64-ucrt-posix-seh, built by Brecht Sanders, r4), cmake 4.0.3. Commit ed71528 on main at 5bae1a4. Not tested with a title.

• git submodule update --init --recursive --jobs 8 && cmake -S . -B build -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++ -DCMAKE_C_COMPILER_LAUNCHER=ccache -DCMAKE_CXX_COMPILER_LAUNCHER=ccache -DBUILD_TESTING=ON -DPython3_EXECUTABLE=C:/Python314/python.exe && cmake --build build --parallel 4 on main at 5bae1a4: exit 0. After the changes, cmake --build build --parallel 4: exit 0, with no error: or FAILED: lines.
• ctest --test-dir build -R '[Uu]ser_?[Ss]ervice' --no-tests=error --output-on-failure --timeout 60: guest_user_service, guest_user_service_user_name and guest_user_service_lifecycle passed, 3/3.
• python tools/check_conventions.py --base origin/main: "All conventions checks passed."
• python tools/pr_overlap.py --branch fix/libSceUserService-npaccountid-null-output: exit 2. It reported 15 open PRs that share files or exports with this branch and merged 15/15 into the base, but couldn't compare feat(libSceUserService): take the user name from ANYPS5_USER_NAME #9345 on core/libs/prx/libSceUserService/Export.cpp or docs(debt): separate entries with blank lines and drop derived export totals #14212 on docs/dev/debt/libSceUserService.md, because both conflict with main. I compared them by hand: in Export.cpp, feat(libSceUserService): take the user name from ANYPS5_USER_NAME #9345's only hunk is in sceUserServiceGetUserName, and none of its changed lines mention the account id. In the debt file, docs(debt): separate entries with blank lines and drop derived export totals #14212 separates entries with blank lines and drops the derived export totals; none of its changed lines name sceUserServiceGetNpAccountId.

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

[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 9345	core/libs/prx/libSceUserService/Export.cpp
PR 12860	core/libs/prx/libSceUserService/Export.cpp core/libs/tests/GuestUserService.cpp docs/dev/debt/libSceUserService.md
PR 12900	core/libs/prx/libSceUserService/Export.cpp core/libs/tests/GuestUserService.cpp docs/dev/debt/libSceUserService.md
PR 14212	docs/dev/debt/libSceUserService.md

[BunnyFl]

Reviewed at ed71528da4, merged locally into main 7fad683776 (clean merge; GitHub also reports it mergeable). Windows 11 Pro 10.0.26300, WinLibs GCC 15.2.0, Release, Ninja.

Firmware check (disassembly only, not run on a console). libSceUserService.sprx from firmware 13.42:

• sceUserServiceGetNpAccountId (6dfDreosXGY, 0x4940) is a tail call through a handler table at 0x1c230, slot +0x10, with setting 0x100 in esi and the output in rdx.
• That table is filled at 0x47d6 from the static table at 0x18000; its +0x10 entry relocates to 0x3e10:

  3e17: mov  0x1c220(%rip),%rdi   ; service object
  3e1e: test %rdi,%rdi ; je -> mov $0x80960002,%eax ; ret
  3e23: test %rcx,%rcx ; jne 0x410            ; the per-user query
  3e2c: mov  $0x80960005,%eax ; ret
  

  So once initialized, a null output returns 0x80960005 before 0x410 looks at the user, as this PR does.
• The only other writer of 0x1c230 (0x1ad8, in the shared initialize body at 0x1930) copies a caller-supplied table, and only when its fourth argument is non-null. sceUserServiceInitialize (0x2850) and sceUserServiceInitialize2 (0x2880) both pass rcx = 0; only sceUserServiceInitializeForShellCore (0x28a0) passes one. So the 0x3e10 handler is the one titles get.

What I ran

• ctest -R '^guest_user_service' --repeat until-fail:3: guest_user_service, guest_user_service_user_name, guest_user_service_lifecycle pass, 3 of 3 runs.
• Negative control: main's libSceUserService/Export.cpp with this PR's test: guest_user_service aborts (0xc0000409).
• python tools/check_conventions.py --base upstream/main: passes.

The debt entry keeps the PS4 10.71 measurement for a non-null output, adds this source and records the unmodelled 0x80960002 gap. Looks good.

AI-assisted: yes (Claude Code)