feat(libkernel): declare physhm_open and physhm_unlink

[Nixitov]

What

Declare physhm_open and physhm_unlink in libkernel. Forever Launcher (bizkut/ps5-forever, PPSA99030) stops at physhm_open on current main:

FAIL: unresolved ELF import n371J5cP+uo from libkernel.prx

The PS5 physical shared memory API is not documented, and shadPS4's module generator also declares both with no arguments and returns ORBIS_OK, so the no-argument int signature follows shadPS4 and the console's behaviour is unmeasured. The relinker resolves no call site for either in this title, so both are declared as throwing exports (NotImplemented_nid_no_patch) rather than fabricating a handle, like the declaration-only exports already in this file; recorded in docs/dev/debt/libkernel.md.

Tested

Windows 11, MinGW-w64 GCC 15.2.0 (WinLibs r7), Release.

• Forever Launcher (bizkut/ps5-forever, PPSA99030, v01.000.006), eboot.bin unwrapped and relinked with the built relinker --windows: on current main it stops at n371J5cP+uo (physhm_open); with this change it passes physhm_open and physhm_unlink and stops at b7uXQmnfB2s (sysarch, libkernel), which this PR does not cover.
• python tools/check_conventions.py --base origin/main 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
• Unblocks: Forever Launcher (bizkut/ps5-forever, PPSA99030) — removes its current start-up stop; it then stops at sysarch
• GPUs tested: n/a
• AI-assisted: yes

[github-actions]

Progress report (134a351 → ce8f328)

📉 System libraries: 91.44% (-0.05%, +2 declared functions)

🆕 2 declared as stubs

Library	Name
libkernel	physhm_open_nid_postfix
libkernel	physhm_unlink_nid_postfix

[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	docs/dev/debt/libkernel.md
PR 6149	docs/dev/debt/libkernel.md
PR 6446	docs/dev/debt/libkernel.md
PR 6510	docs/dev/debt/libkernel.md
PR 6519	docs/dev/debt/libkernel.md
PR 6534	docs/dev/debt/libkernel.md
PR 6908	docs/dev/debt/libkernel.md
PR 6936	docs/dev/debt/libkernel.md
PR 6977	core/libs/prx/libkernel/Unimplemented.cpp docs/dev/debt/libkernel.md
PR 7039	docs/dev/debt/libkernel.md
PR 7049	docs/dev/debt/libkernel.md
PR 7075	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 7129	docs/dev/debt/libkernel.md
PR 7149	docs/dev/debt/libkernel.md
PR 7192	docs/dev/debt/libkernel.md
PR 7332	docs/dev/debt/libkernel.md
PR 7509	docs/dev/debt/libkernel.md
PR 7565	docs/dev/debt/libkernel.md
PR 7653	docs/dev/debt/libkernel.md

[milescoviello]

A Linux check of Forever Launcher (PPSA99030, 01.000.006, catalog build), Ubuntu 24.04, GCC 13.3, Release, main 1b1f848:

• On main, it stops at load on n371J5cP+uo (physhm_open).
• With this PR and fix(libkernel): export sysarch so modules that import it load #6028 (sysarch, your next stop), it gets past both. nm then lists three imports that no library exports: the two libSceSystemService PsmKit exports (cltshBrDLC0, 7cTc7seJLfQ, now fix(libSceSystemService): declare the PsmKit local process exports #6939) and K1A7IJp5xYc (sceKeyboardSetProcessPrivilege, feat(libSceKeyboard): export missing firmware functions from database #6787).
• With all four merged, 2 of 2 60 s runs load and reach the next stop, sysctl_nid_postfix: unsupported MIB 1.14.32.3 (kern.proc.vmmap).

AI-assisted: yes

[milescoviello]

An update for the current catalog build of Forever Launcher (PPSA99030, 01.000.012), Linux, main bf20538:

• On main it still stops at load on n371J5cP+uo (physhm_open).
• With this PR, fix(libkernel): export sysarch so modules that import it load #6028, fix(libSceSystemService): declare the PsmKit local process exports #6939 and feat(libSceKeyboard): export missing firmware functions from database #6787 merged locally, it loads and stops at 0 s on sysctl_nid_postfix: unsupported MIB 1.14.32.3 (kern.proc.vmmap), 2 of 2 runs. I opened feat(libkernel): report the guest address-space map through kern.proc.vmmap #7070 for that.
• With feat(libkernel): report the guest address-space map through kern.proc.vmmap #7070 as well, its Wine runtime starts (ntdll.so loads and wineboot.exe -i begins), then exits with status 1 when Wine maps its shared user data at 0x7ffe0000, below the guest arena.

One note for whoever merges second: this PR and #6028 both add a declaration at the same place in libkernel/Unimplemented.cpp, so the second one needs a rebase. I'll rebase #6028 if this one goes in first.

AI-assisted: yes (Claude Code)

[babyhuey]

Windows runs of Forever Launcher (PPSA99030, catalog 01.000.012). Windows 11, RTX 3080, WinLibs GCC 15.2, Release, on main e5f18ea99. Plain merges, no conflicts; Unimplemented.cpp merged cleanly. No warnings in Unimplemented.cpp. This PR adds no test, so none was run.

• main: fails to load (0xC0000139, no window). The Windows loader reports every missing import at once: n371J5cP+uo and AUqJNkobQ1c (libkernel), 7cTc7seJLfQ and cltshBrDLC0 (libSceSystemService), K1A7IJp5xYc (libSceKeyboard).
• This PR alone (44e9c138f): n371J5cP+uo and AUqJNkobQ1c resolve. It still fails to load on the other three, which fix(libSceSystemService): declare the PsmKit local process exports #6939 and feat(libSceKeyboard): export missing firmware functions from database #6787 provide.
• With fix(libSceSystemService): declare the PsmKit local process exports #6939 (f69e6c093), feat(libSceKeyboard): export missing firmware functions from database #6787 (cc941b976) and feat(libkernel): report the guest address-space map through kern.proc.vmmap #7070 (5fcb7e8a8) as well (fix(libkernel): export sysarch so modules that import it load #6028 is already in main), 2 of 2 60 s runs load with every import resolved. Its Wine runtime starts: the psmkit-wine start-up stages all return 0, and wineboot.exe -i begins in /data/forever/prefix. It then exits 1 on err:virtual:virtual_alloc_first_teb wine: failed to map the shared user data: c0000018.

That is the same stop as on Linux: Wine fails to place its shared user data at 0x7ffe0000. A memory trace on Windows shows the guest never asks the emulator for a fixed mapping there: every anonymous map request on this path reaches DoMapAnon with no address and no MAP_FIXED, and each succeeds inside the guest arena. c0000018 is Wine's own result after the page lands elsewhere. (An earlier version of this comment blamed Windows' own KUSER_SHARED_DATA page; the trace doesn't show that.)

AI-assisted: yes

[milescoviello]

@babyhuey thanks for the Windows runs, and sorry for the slow answer. That's the same stop I hit on Linux. On Linux it goes away with #6469 (QRMOE, still a draft), which honours a low mapping hint, so Wine's shared user data lands at 0x7ffe0000.

On main with #6915, #6939, #6787, #7070 and #6469, the next stop is sendmsg with SCM_RIGHTS (#7090). After that, Forever Launcher ran the whole 60 s, 2 of 2, on Linux; NOTE_LOWAT (#7092) and sysarch (#7095) are already in main. It then reaches its own prefix setup and Wine exits with 0xC0000135, because the prefix "must be initialized by payload Wine". On Windows, #7090 still refuses SCM_RIGHTS (control data is not supported on Windows), so that's likely the next stop there after #6469.

AI-assisted: yes (Claude Code)