docs(libkernel): record select limits as measured on a PS5

Co-Authored-By: Claude Opus 5.5 <[email protected]>

docs/dev/debt/libkernel.md

@@ -235,7 +235,7 @@
 - [File mappings on macOS](../../../core/libs/prx/libkernel/DirectMemory/PosixMemory.cpp) (libkernel) - `mmap` of a regular file, of an anonymous `shm_open` object (an unlinked temporary file on macOS) or of a JIT shared memory descriptor maps the descriptor shared or private as on Linux, with the same access checks. macOS refuses to create an executable file mapping (`EPERM`), so an executable mapping is made without execute and then given it with `mprotect`. Under Rosetta 2 (Apple M2, macOS 26), code rewritten through one view of a file after it ran from another view keeps its old translation until the executable view's protection changes. Named `shm_open` objects still throw on macOS.
 - [sendfile](../../../core/libs/prx/libkernel/Socket/src/Udp.cpp) (libkernel) follows FreeBSD 11's `sendfile` for a connected stream socket: headers, the file from `offset` (to its end when `nbytes` is 0) and trailers, the file offset left unchanged, `EAGAIN` with `*sbytes` set after a partial send on a non-blocking socket. It reads the file through `sceKernelPread` in 64 KiB blocks and sends each with `send`. A socket as the file fails with `EINVAL`, as on FreeBSD, but any other descriptor `sceKernelPread` reads, such as `/dev/urandom`, is sent where FreeBSD requires a regular file or shared memory. `flags` are ignored: reads complete before sending, so `SF_NODISKIO` never fails with `EBUSY`, and `SF_SYNC` returns once the data is queued on the socket rather than sent. PS5MC (PPSA99420) imports it
 - [pthread_kill](../../../core/libs/prx/libkernel/System/src/Signals.cpp) (libkernel) returns the error number like FreeBSD's `pthread_kill`: `EINVAL` for signals outside 0 to 128 and `ESRCH` for a null thread. A signal to the calling thread goes through `raise`, so it shares `raise`'s limits; a signal to another thread is left on the pending set for whichever thread waits for or unblocks it, since it is not delivered to the chosen thread, and a handle that is not a live guest thread returns `ESRCH` without touching it
-- [select](../../../core/libs/prx/libkernel/Socket/src/Socket.cpp) (libkernel) validates `tv_usec` but saturates a valid timeout beyond the remaining `steady_clock` range at `time_point::max()`, so the wait can continue effectively indefinitely. The console's behavior for a valid timeout outside the host clock's range is unknown.
+- [select](../../../core/libs/prx/libkernel/Socket/src/Socket.cpp) (libkernel) fails an `nfds` above 1024 with `EINVAL`, where a payload process on a PS5 (firmware 13.60) accepted 1025 and failed `EBADF` for a set bit above its last open descriptor, as FreeBSD 11.4's `select_check_badfd` does. A timeout beyond the remaining `steady_clock` range saturates at `time_point::max()`, so the wait can continue effectively indefinitely; the console accepted seconds up to `INT64_MAX` (measured only on a descriptor already ready), and failed a negative field or `tv_usec` of 1000000 or more with `EINVAL`, as here
 - [inet_ntop, inet_pton](../../../core/libs/prx/libkernel/Socket/src/AddressConversion.cpp) (libkernel) - an `AF_INET6` result of `inet_ntop` is followed by zero bytes up to `size`. In a payload process on a PS5 (firmware 12.70), the text forms, the order of the checks and this padding were measured through both exports. The padding was sampled at sizes from 3 to 65536, above 64 only at 65, 128, 256, 1024, 4096 and 65536; the sizes that were not sampled, sizes above 65536 and a title process are assumed to behave the same
 - [_fstatfs](../../../core/libs/prx/libkernel/File/src/Stdio.cpp) (libkernel) reports the host volume of a file, not a console mount: block sizes, block counts, name length and the read-only, `nosuid` and `noexec` flags (on Windows only read-only, and `f_files` and `f_ffree` are 0). The type, mount names, owner, filesystem id and I/O counters are 0 or empty. Sockets, pipes and kqueues return `EINVAL`, as FreeBSD's `getvnode` does. A character device, including `/dev/random` and `/dev/urandom`, reports FreeBSD 11.4's [`devfs_statfs`](https://github.com/freebsd/freebsd-src/blob/release/11.4.0/sys/fs/devfs/devfs_vfsops.c#L225-L237) values on every host: no flags, 512-byte blocks, 2 blocks, nothing free and a 255-byte name limit. A Windows descriptor that is none of these throws. Stremio Plus (PPSA74126) reads the free space before a download and only logs the other fields
 - [pthread_spin_lock](../../../core/libs/prx/libkernel/Pthread/Posix/Spinlock.cpp) (libkernel) - a lock taken again by the thread that holds it throws here. On a PS5 (firmware 12.70) that call had not returned after 2 s, as FreeBSD 11.4 libthr [`thr_pspinlock.c`](https://github.com/freebsd/freebsd-src/blob/8c639945a44588f6cfdddfb057bb5aad88735da0/lib/libthr/thread/thr_pspinlock.c) spins forever