sceUltGetUlthreadRuntimeInfo returned success when workerThreads was null and workerCapacity was nonzero, while rejecting the same null/nonzero-capacity combination on ulthreads with ULT_ERROR_NULL. Reject null workerThreads with nonzero capacity before touching info, and add test coverage checking the error result and that info is left untouched.
sceUltGetQueueInfo reported every blocked queue thread as a waiting popper because both directions incremented one counter. Track the two wait classes in separate fields and fill both query outputs, copy one waiter record per blocked thread, and cover full-queue and empty-queue waits with a new queue-waiters test. sceUltGetUlthreadRuntimeInfo now throws when a caller asks for worker thread handles, since this runtime starts no worker threads, and the debt entry records the remaining limitations.
The tests required the app list with no room, the running-app queries
for another app ID, sceLncUtilIsAppLaunched for another title, and
sceLncUtilGetAppId (through GetAppIdThrows) for another title to throw.
Those answers aren't measured, and a test shouldn't lock in the
placeholder, so these checks go and GetAppIdThrows is back to main's
form. The calls still refuse those cases, as the debt entries say; the
running title's answers and firmware 13.42's argument checks stay.
The tests checked the exact "not implemented" message from
sceContentDeleteByPath, sceContentDeleteById,
sceNpEntitlementAccessAbortRequest and
sceNpEntitlementAccessDeleteRequest. A test that requires a placeholder
to throw fails as soon as the call is implemented, so those checks go,
with the helper and catch they used. The calls still report themselves
as not implemented, as their debt entries say, and the argument checks
stay.
SSL_get_ex_new_index and SSL_CTX_get_ex_new_index took the new, dup and
free callbacks as void*, which a caller passing real callbacks can't
convert to. They now take LibreSslExNew, LibreSslExDup and
LibreSslExFree, the guest-ABI forms of CRYPTO_EX_new, CRYPTO_EX_dup and
CRYPTO_EX_free from LibreSSL 2.5.4's ossl_typ.h. Reported by @MarcZX10.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
SSLeay() returns 0x20000000, the value measured on a PS5 with firmware
13.40 and the SSLEAY_VERSION_NUMBER of LibreSSL 2.5.4, which the module
is. The guest_libressl_version test checks it.
CONTRIBUTING.md permits silent stubs only when unblocking a title UI path and when backed by title evidence. Since no title ID evidence demonstrates that a constant no-disc return unblocks a UI path, throw NotImplemented_nid_no_patch from sceDiscIdGet and sceDiscIdGet2 without touching caller memory. Tests assert both calls throw on null and non-null outputs and across flags values while leaving caller buffers untouched, and the debt entry records the unresolved hardware contract.
The previous revision rejected a null output with 0x80D50001 and any nonzero flag with 0x80D50002, but no SDK or console evidence establishes those validation rules, the error codes or their precedence, and the flag argument may be a bitfield. Report SCE_DISC_ID_ERROR_NO_DISC for every pointer and flag value without touching the output, and record the unverified contract in the debt entry. Normalize the new test and CMake file to LF line endings.
The previous revision rejected a null output with 0x80D50001 and any nonzero flag with 0x80D50002, but no SDK or console evidence establishes those validation rules, the error codes or their precedence, and the flag argument may be a bitfield. Report SCE_DISC_ID_ERROR_NO_DISC for every pointer and flag value without touching the output, and record the unverified contract in the debt entry.
MarcZX10