VOP3 0x140 was translated as a plain fma, so a zero source times an
infinity or a NaN gave NaN instead of the addend. A zero source now
gives +0 + src2, the same rule v_mul_legacy_f32 and v_mac_legacy_f32
already follow. The opcode also takes neg and abs on its sources.
MIMG 0xe7 is the node test of image_bvh_intersect_ray with a 64-bit node
pointer: the pointer takes the first two address dwords (low, high), so
the instruction addresses 12 dwords, or 9 with A16.
The node index and address are computed from the whole pointer. A zero
base in the T# is a null BVH only for the 32-bit form: with 64-bit
pointers the T# may cover all of memory from base 0 and the pointer then
carries the node address, so only the T# size bounds the node.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
scePthreadKeyDelete frees a key slot and the next scePthreadKeyCreate
hands the same index out again, but every thread kept the value it had
stored under that index. The new key read the previous key's value
instead of NULL, and at thread exit its destructor was called on that
stale pointer.
Each slot now carries a sequence number, as in FreeBSD libthr. A thread
value is visible to scePthreadGetspecific and to the exit destructors
only while its sequence matches the slot's, so a new key starts as NULL
in every thread, as POSIX requires, and a deleted key reads NULL.
scePthreadGetspecific stays lock-free and scePthreadSetspecific no
longer takes the key lock.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Lock the guest pages of the range with VirtualLock or mlock, following FreeBSD mlock(2): the range is widened to guest pages, a zero length succeeds, a wrapping range is EINVAL and a range that is not mapped and accessible is ENOMEM. The Windows working set limits and the Linux RLIMIT_MEMLOCK soft limit are raised when the lock needs it.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
sceAjmBatchJobDecodeSingle runs a single-frame decode job with the stream
sideband. sceAjmBatchJobGetGaplessDecode and sceAjmBatchJobGetCodecInfo
append run jobs without buffers that return the gapless-decode and codec-info
sidebands.
A run job without input or output buffers no longer decodes: it reports the
instance state with a zero result instead of PARTIAL_INPUT on MP3 and Opus.
RUN_GET_CODEC_INFO (bit 11) writes the ATRAC9 codec info after the gapless
sideband. Codec info for other codecs, and codec info together with
RUN_MULTIPLE_FRAMES, throw.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
MUBUF opcode 0x34 subtracts the data register from a buffer dword and
clamps the result to zero, returning the previous value. The form without
glc throws: no GFX10 target is known to provide it.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Controllers other than DualShock 4 and DualSense lack a dedicated touchpad button. Maps the SDL Back button (View on Xbox) to PadButton::TouchPad for non-PlayStation controllers.
Co-Authored-By: saulo benigno <[email protected]>
When an application requests only 2 video buffers and holds one during playback, decode-ahead frames allow the decoder to stay ahead and prevent stalls under SyncModeNone.
Co-Authored-By: saulo benigno <[email protected]>
On MinGW, thread_local storage is emutls: libgcc frees a thread's blocks
from a winpthread key destructor when the thread function returns, and
the CRT runs the thread_local destructors later, from the module's TLS
callback at thread detach. The destructors of the driver's non-trivial
thread_locals (the deferred label queue, packet profile, capture and
revalidation scratch vectors, build profile, memory counters, disk cache
buffers) therefore ran on freed storage. Under the debug heap the flip
test failed on every run at the queue worker's exit after the injected
flip failure; without it the use-after-free showed up as intermittent
heap corruption or a stack overflow on the Windows CI job.
ShaderRecompiler::ThreadOwned keeps such state behind a trivially
destructible thread_local pointer and deletes it from its own pthread key
destructor, which receives the objects directly instead of reading
emutls. The driver no longer registers any thread_local destructor.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A filter_type naming a subset of the PNG filters used only the first
filter of the mask, because stb's writer is either fully adaptive or
forced to one filter. Decoder::Png::Encode now takes the set of allowed
filter types, filters the rows and writes the chunks itself, and keeps
stb for deflate only. Output for the all-filters and single-filter
cases is unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
SHA1RNDS4/NEXTE/MSG1/MSG2 with a memory source made `--to-intel` fail
the whole relink. The stub now spills a scratch xmm, loads the operand
into it with the unaligned load already used for SHA-256 and runs the
register-form lowering. The SHA1RNDS4 immediate is read after the SIB
byte and displacement. RIP-relative operands still fail: the stub runs
at a different address.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Return the kerning between two character codes at the render scale of
the font, the one set by sceFontSetupRenderScalePixel and
sceFontSetupRenderScalePoint, and fail with NOT_BOUND_RENDERER when no
renderer is bound. sceFontGetKerning keeps its behaviour and shares the
FreeType lookup.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The WriteAddressFromCounter_04_00 measure checks only the address and the
counter pair measure only the address and an even start, as in ampr_emu.
TechnicalDebt records that AMPR commands are larger than the console
records, which affects titles that pad with Nop to an exact offset, and
that the at-start flag of the 04_00 writes is ignored.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
s_setreg_b32 has no effect, so a read of the MODE round fields after it
wrote another mode still reports round to nearest even, the rounding the
recompiled code uses. The refused HW_ID test row is named as hardware
register 4, which has no name on gfx10.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Its meaning for the f32 accumulator is not known, like op_sel and neg_hi,
and the integer dots already refuse anything but op_sel_hi set on every
source.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
On a device whose subgroups are narrower than the wave, the recompiled
wave runs as several subgroups, so one subgroup's zeroing write could
land after another's atomics, and a read could run before another
subgroup's atomics. An s_barrier after the zeroing writes and another
after the atomics order them on any subgroup width.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
sceAgcSetPacketPredication sets or clears the predicate bit of a type-3
packet header, and sceAgcSetRangePredication does so for every type-3
packet between two addresses, skipping type-2 fillers. The z-pass, bool
and disable predication size queries return the size of the
SET_PREDICATION packet sceAgcDcbSetPredication writes.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
With a16 the bias is a half that occupies a full address dword and the
coordinates start in the next one. The layout packed the first
coordinate into the upper half of the bias dword, so image_sample_b,
image_sample_b_cl, image_sample_b_o and the biased gathers read their
coordinates from the wrong halves.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Dean