libkernel: execv/execve and the setuid/setgroups/setrlimit family are absent (XPSemu, PS5MC); fix branch ready #3558

libkernel lacks execv (KaK52lYTBIs) and execve (kdguLiAheLI), and the identity and limit setters setuid (JVmUZwK-HJU), setreuid (6w8tPp+Yk6E), setregid (ROILLZdYZPc), setgroups (M8VZ3iIlmyg) and setrlimit (4X0QwvuCfjc). From the relink registry order execv/execve are the predicted next stop of XPSemu (PPSA97358) and the setters follow for XPSemu and PS5MC (PPSA99420); not yet observed in a run.

Fix branch ready: screamm:feat/libkernel-exec-setuid (two commits on main a11c8f7f: 57f0fa37 feat(libkernel): implement execv and execve as throwing exports, 1d0b2b61 feat(libkernel): implement the identity and limit setters). I cannot open it as a PR because of the intake check; anyone eligible is welcome to carry it, keeping the commits as they are.

• execv/execve follow the execvp precedent of fix(libkernel): export execvp #2827: FreeBSD's checks first (EFAULT for a null argv or path, ENOENT/ENOTDIR from the guest lookup, EACCES for a non-regular file), then the executable replacement not implemented throw where FreeBSD would load the image; over-long paths and oversized argument blocks throw rather than guess. XPSemu's execv sites run in a fork child and its fork returns ENOSYS, so like execvp they are not reached at start-up (ZiZc3/XPSemu 8eec07b4). Sources: FreeBSD 8c639945 kern_exec.c, exec.c.
• Setters follow the repository's root identity model and FreeBSD kern_prot.c/kern_resource.c: ids 0 (or -1 unchanged) succeed, any other id throws (the console's credential rules are unmeasured); setgroups takes {0} or the empty set (EINVAL above 1024, EFAULT for a null list with a count) and marks issetugid as changed as FreeBSD does; setrlimit returns EFAULT/EINVAL as FreeBSD orders them, treats negatives as infinity, lowers a soft limit above the hard one to it, and succeeds only when the result equals what getrlimit reports (RLIMIT_DATA), otherwise throws because getrlimit stores nothing. Note for XPSemu: its os_setup_limits calls getrlimit(RLIMIT_NOFILE) first, which still throws on main, so that is its real next stop there.
• guest_execvp, guest_credentials, guest_getrlimit gain the new cases.

Tested on Windows 11 Pro 26220, WinLibs MinGW-w64 GCC 15.2.0 r7, Ninja, Release, BUILD_TESTING=ON, full build including libs: the three tests 3/3; ctest -j8 -E "^agc_driver_|^video_out_flip" 533/536, the failures being the known windows_exceptions (#3411), guest_kernel_sys (#3510) and guest_filesystem (#3518); with main's Process.cpp all three extended tests fail; tools/check_conventions.py --base main and git diff --check pass. Not run on Linux.

Written with Claude Code assistance (implementation by a subagent, reviewed and validated in this session).