On RDNA2 the image, buffer and global float atomics (fmin, fmax,
fcmpswap and the _x2 forms) flush both operands to a signed zero when
bit 0 of the shader's denormal field is clear, as the LDS forms do: the
f32 field for the 32-bit forms, the f16/f64 field for _x2. Only the
LDS forms flushed here. floatAtomicFlush (was sharedFloatFlush) now
sets MemoryInfo::flushDenormals for every float atomic, and
AtomicFloatBits applies it in minMax and compareSwap.
The float atomic tests from #332, #506 and #562 were measured with
denormals kept, so they now set FLOAT_MODE 0xf0.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
sceAudio3dGetSpeakerArrayMemorySize, CreateSpeakerArray, DeleteSpeakerArray
and GetSpeakerArrayMixCoefficients(2) forward to the AudioOut2 speaker-array
calls as the console does: the memory size asks for no ambisonics, create
rejects a null handle or parameters with SCE_AUDIO3D_ERROR_INVALID_PARAMETER
and passes no ambisonics parameters, and MixCoefficients passes height_aware 0
and a 0.1 downmix spread radius. AudioOut2 create no longer rejects a null
ambisonics pointer, which the console accepts.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01RbLgECLQ2bpDM7oe2urGZK
sceAudioInOpenEx and sceAudioInHqOpenEx threw. In the PS5 12.00 module both
call the shared open routine that sceAudioInOpen uses, with one extra
argument in fourth place that only reaches the console's Mbus handle
registration. HqOpenEx sets the high-quality flag but skips the length,
rate and parameter checks of sceAudioInHqOpen, so both open like
sceAudioInOpen here. sceAudioInSetCompressorPreGain and
sceAudioInSetNoiseGateThreshold return 0 on the console without reading
their arguments.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01RbLgECLQ2bpDM7oe2urGZK
Addresses review on #8355: drops g_appLauncherReady and the
unverified init-param size check so both exports are plain silent
stubs, lists them under Silent stubs in the debt file naming
PPSA20397, and moves them out of the unknown-info entry.
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
The call threw for lack of a signature. In the PS5 12.00 libkernel it takes
one int, passes it to syscall 601 (command 29) and returns the raw kernel
result. No dump is produced here, so it returns 0, as
sceCoredumpConfigDumpMode does.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01RbLgECLQ2bpDM7oe2urGZK
sceWebBrowserDialogSetCookie and sceWebBrowserDialogResetCookie threw.
They now check their parameters in the console's order and with its
codes (PS5 12.00 module), then succeed: there is no cookie store. Open
sets the open-dialog flag that makes both return 0x80B80006 until the
dialog is updated to finished, closed or terminated.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01RbLgECLQ2bpDM7oe2urGZK
Follow FreeBSD 11.4 sys_getresuid and sys_getresgid (sys/kern/kern_prot.c):
retrieve real, effective, and saved user and group IDs for the guest process,
which is assumed to run as root with IDs 0. Pointers that are not null receive 0,
null pointers are ignored, and 0 is returned leaving errno unchanged on success.
Cover both in guest_credentials test and update libkernel debt documentation.
Debt entries name the titles they were measured on. Use the title ID, or "a retail title" where the sentence needs no identification. No technical content changes.
Under APS5_LAZY_SHADER_PREPARE=1 registration prepares nothing, but
ResolveShaderAbi still planned and prepared every compute or pixel stage
whose ABI the title resolved, so a title that resolves every shader's
ABI prepared them all at boot whether they ever drew.
With lazy preparation ResolveShaderAbi now plans the stage with the
ABI's state (its register and entry-point checks, as lazy registration
keeps them), records the ABI key, and queues the ABI's context and
primitive registers on the shader. The shader's first draw or dispatch
plans and prepares the queued ABIs with that state before it looks for
its artifact, so a shader that never draws or dispatches stays
unprepared and one that does gets the entries eager resolution would
have made. A stage whose plan fails is deferred as before. Without the
switch nothing changes; graphics-stage ABIs are still prepared when
they are resolved.
Test: agc_driver_lazy_resolved_abi resolves the ABI of four lazily
registered shaders and draws two. None holds prepared entries after
resolution, the two that never drew still hold none, and a drawn one
draws the same bytes as with eager preparation in a child process.
The public SDK's libSceLibcInternal table declares num_get<char>::id, time_get<char>::id and the
num_get<char> vtable as objects; libSceAmpr exported their NIDs as functions that throw, which a
data consumer reads as code (#8875). libc now defines them the way the other facet ids and the
num_put and codecvt vtables are kept: the ids zero, the vtable 15 zeroed slots (two header
entries, the two destructor entries and Dinkumware's eleven do_get), and libSceAmpr drops the
functions. An import qualified to libSceAmpr.prx falls back to libc, so each object has one
identity.
Co-Authored-By: Claude Fable 5.1 <[email protected]>
siginterrupt is a libSceLibcInternal export (the public SDK's stub table has it there, not in the
kernel tables) and FreeBSD 11.4's lib/libc/gen/siginterrupt.c reads the signal's disposition,
sets or clears the signal's bit in _sigintr and SA_RESTART, and installs the disposition again.
libc cannot call libkernel, so, as the pthread hooks do, libc keeps a registered sigaction and
the address of the _sigintr set, and libkernel registers both from a static initializer when it
loads. A call before that throws; the set is written without libkernel's registration lock, as
FreeBSD's libc does. Builds on #8866, which adds the set.
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Dean