Commit Graph

2272 Commits

Author SHA1 Message Date
Pacuka 7b722253a4 fix(audiodec): take the M4AAC raw channel layout from the stream
For a raw M4AAC stream the maximum channel count was written into the
AudioSpecificConfig as the channel configuration. The maximum only bounds
the stream: The Smurfs Dreams creates its decoder with 8 channels at most
and sends stereo, so FFmpeg refused configuration 8 (reserved) and a
stereo stream under configuration 6 or 7 decodes no frame. The decoder
now starts from the mono/stereo configuration and on the first access
unit that it cannot decode moves on to the configurations with more
channels, up to the maximum, keeping the first one that decodes.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-08 17:29:46 +02:00
Adrià eb2818f505
Merge pull request #1476 from awake-devel/fix/timedwait-nested-waiter
fix(libkernel): give each nested TimedWait its own waiter on Windows
2026-10-08 16:48:48 +02:00
Adrià 29e533aca0
Merge pull request #1469 from arvindfroi/refactor/coredump-stopinfo-gpu-name
refactor(libkernel): name the Jrs7UUkGOFo Coredump stub by its NID
2026-10-08 16:37:04 +02:00
Adrià 700c0babc6
Merge pull request #1468 from arvindfroi/refactor/ajm-wvorbis-names
refactor(libSceAjm.native): name the WVorbis stubs by their NIDs
2026-10-08 16:36:34 +02:00
Adrià 9f4bd71cb6
Merge pull request #1475 from awake-devel/fix/aio-windows-positioned-io
fix(libkernel): use positioned transfers for Windows AIO requests
2026-10-08 16:28:43 +02:00
Adrià 68cebaee8b
Merge pull request #1474 from popcoder565/fix/agc-queue-worker-precise-poll
fix(agc): poll for completions with a high-resolution timer on Windows
2026-10-08 16:20:23 +02:00
Adrià 4d03ecf4de
Merge pull request #1472 from popcoder565/fix/videoout-vblank-precise-wait
fix(videoout): pace vblanks with a high-resolution timer
2026-10-08 16:13:49 +02:00
Adrià 7a047bc171
Merge pull request #1441 from ToyuAU/fix/windows-import-watch
fix(agc): compare Windows host imports after late driver writes
2026-10-08 16:12:45 +02:00
Adrià fe9beae001
Merge pull request #1452 from mrk1987/fix/posix-mkdir-errors
fix(libkernel): return POSIX errors from mkdir
2026-10-08 16:09:37 +02:00
Adrià 050538528a
Merge pull request #1470 from oneandonlydean/up/waitmemory-race
fix(agc): guard the reported wait-timeout set against concurrent queues
2026-10-08 16:07:49 +02:00
Adrià f7e98dad39
Merge pull request #1453 from mrk1987/fix/posix-pthread-setstack
feat(libkernel): export POSIX pthread_attr_setstack
2026-10-08 16:07:11 +02:00
Adrià f2f7abab71
Merge pull request #1448 from ewraj/fix/net-connect-einprogress
fix(libSceNet): report EINPROGRESS for a non-blocking connect and translate the missing socket errors
2026-10-08 16:02:31 +02:00
Adrià 3c0756ee87
Merge pull request #1467 from arvindfroi/refactor/psml-mfsr2-names
refactor(libScePsml_debug): name the Mfsr2 stubs by their NIDs
2026-10-08 16:00:30 +02:00
Dean 74b5c9103b
Merge pull request #788 from Celegans12/feat/videoout-dcc-visible-keys
fix(agc): store and read DCC keys over the console's metadata extent
2026-10-09 02:59:52 +13:00
Adrià 123a3c0365
Merge pull request #1443 from camezaro-intc/feat/font-dettach-device-cache-buffer
feat(font): implement sceFontDettachDeviceCacheBuffer
2026-10-08 15:58:18 +02:00
Adrià 2b6cc18876
Merge pull request #1460 from LeandroLP23/fix/libc-setvbuf-default-size
fix(libc): give setvbuf a default buffer when the size is 0
2026-10-08 15:56:59 +02:00
Adrià 1e660dd9fe
Merge pull request #1440 from ewraj/test/agc-mesh-tests
test(agc): build the mesh tests with MinGW
2026-10-08 15:53:01 +02:00
Adrià 461fcfc7f3
Merge pull request #1436 from tototomate123/fix/hfw-offline-imports
fix(prx): add missing HFW imports with strict offline handling
2026-10-08 15:52:39 +02:00
Adrià 5fbddcdf53
Merge pull request #1435 from ewraj/fix/linux-guest-varargs
fix(libc): read every variadic argument of the Linux printf and scanf exports
2026-10-08 15:52:17 +02:00
Adrià 50490e4028
Merge pull request #1444 from ewraj/refactor/appcontent-enqueue-download
refactor(libSceAppContent): name the stub the NID database resolves
2026-10-08 15:50:29 +02:00
ToyuAU c01044eeaa fix(memory): restore write tracking for fresh private backing 2026-10-08 15:32:42 +02:00
ToyuAU 0df794548e fix(tests): configure Windows unwind for write-watch coverage 2026-10-08 15:32:42 +02:00
ToyuAU 838d02a88c fix(agc): compare Windows host imports after late driver writes 2026-10-08 15:32:42 +02:00
Adrià b3fcd6d6c0
Merge pull request #1408 from hashaam1217/fix/pin-wide-compute-subgroups
fix(agc): require the widest subgroup for compute programs wider than the device
2026-10-08 15:25:25 +02:00
Adrià 02e741060e
Merge pull request #1394 from gjonn/feat/agc-r10g11b11-float-images
feat(agc): load and store R10_G11_B11 float images
2026-10-08 15:25:08 +02:00
Celegans12 1816aea91c fix(agc): store and read DCC keys over the console's metadata extent
The console's DCC of a single-level 4-byte SW_64KB_R_X surface is
ceil(W/512) * ceil(H/512) * 4096 bytes, 49152 for 1920x1080, but the
driver stored and read one key per 256 surface bytes (34560).
PPSA26344 retiles its render DCC into the displayable copy it
registers (cs_dcc_retile_32bpp); a retile of the stale tail gives the
mixed keys PPSA12544 stops on.

DccKeyCount gives that extent. Color target and storage image key
stores cover it when the DCC is pipe-aligned (CB_COLOR_ATTRIB3 bit 30,
T# word 6 bit 19), whose CB keys reach its last meta block, and throw
where it is not writable; unaligned DCC keeps one key per 256 bytes,
as do texture and color target key reads. DisplayBufferKeys reads the
display DCC over the extent, presents only when every key agrees,
names the first differing key's pixels and checks a pending image
against those keys.
2026-10-08 15:24:43 +02:00
Adrià fd5342b6d6
Merge pull request #1425 from gjonn/fix/agc-reversed-rgba8-color-targets
fix(agc): render 8_8_8_8 and 16_16_16_16 color targets with reversed component orders
2026-10-08 15:18:26 +02:00
Adrià 588c7b7b5e
Merge pull request #1402 from gjonn/claude/sleepy-engelbart-7b6379
fix(recompiler): take point-filtered sampler copies' descriptors from their source sampler
2026-10-08 15:17:38 +02:00
Scott Schuster 550920a1fd fix(agc): render CB_COLOR_CONTROL mode disable as a draw without color
Mode 0 of CB_COLOR_CONTROL turns color writes off, but a draw with a
written color target and this mode was skipped as unsupported. Count no
export as written when the mode is disable, so the draw renders without
a color attachment like the existing attachment-free path, and accept
the mode with any ROP3, which does not apply while writes are off.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-08 15:13:08 +02:00
Adrià ede7ead473
Merge pull request #1419 from TheAggie/fix/agc-accept-tuning-fields
fix(agc): accept RB and primitive-group tuning fields
2026-10-08 15:12:33 +02:00
Adrià f4028921f8
Merge pull request #1412 from TheAggie/fix/recompiler-ngg-vertex-wave-info
fix(recompiler): give vertex-path shaders a full NGG merged wave info
2026-10-08 15:08:22 +02:00
Adrià c68c7576b9
Merge pull request #1399 from hugodearaujo/fix/rtc-parse-datetime-rfc2822-hardware
fix(rtc): parse RFC 2822, RFC 850 and asctime text like the console
2026-10-08 15:05:42 +02:00
Scott Schuster 71e566fafb fix(recompiler): give vertex-path shaders a full NGG merged wave info
The vertex entry set only the vertex and instance index VGPRs, leaving
the NGG merged wave info SGPR s3 at 0. NGG shaders build EXEC from it
(s_bfe_u64 exec, -1, <count in bits 16..22>), so Mesa's NGG vertex
shaders ran with an empty EXEC and rasterized nothing.

Each host invocation on the vertex path is one vertex of a full wave:
when the user data starts at s8 or later, set s3 to the wave size as
both vertex and primitive count, with wave index 0.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-08 15:05:35 +02:00
Adrià Franch 61fbade375 test(agc): pin R10_G11_B11 float image edge values
Store: negative, -0, NaN (quiet and low-payload), +-Inf, overflow to the
largest finite value, truncation and small-float denormals. Load: Inf,
denormal and largest finite codes. Values measured on gfx1035.

Co-Authored-By: Claude <noreply@anthropic.com>
2026-10-08 15:04:47 +02:00
Gjon Gojcevic ede33d56e2 feat(agc): load and store R10_G11_B11 float images
Format 43 packs R as a 10-bit and G and B as 11-bit unsigned floats, which
Vulkan has no format for. Keep it in an R32_UINT image like the other
packed formats without a host equivalent, and convert texels in the shader
on image loads and stores; sampling it still throws.
2026-10-08 15:04:47 +02:00
Adrià 7ad0b2976b
Merge pull request #1406 from ebarr24/codex/batch-demangle-guest-heap
fix(libc): keep guest demangle buffers on the application heap
2026-10-08 15:03:54 +02:00
Adrià 147a9855c0
Merge pull request #1404 from overlordxrz-source/fix/libkernel-small-stacksize-einval
fix(libkernel): return EINVAL for thread stack sizes below 16 KiB
2026-10-08 15:03:07 +02:00
Adrià 79d23eb76c
Merge pull request #1321 from TibixDev/pr/guest-path-case
fix(libc): resolve case differences in guest paths and mounts
2026-10-08 15:01:28 +02:00
Adrià 09a17e963a
Merge pull request #1398 from hugodearaujo/fix/rtc-parse-datetime-iso-hardware
fix(rtc): parse ISO text in sceRtcParseDateTime by the RFC 3339 rules
2026-10-08 14:56:12 +02:00
Gjon Gojcevic 71851d66a3 fix(recompiler): take point-filtered sampler copies' descriptors from their source sampler
When one guest sampler is paired both with an image that needs point
filtering (sint, depth-bits or converted format) and with an ordinary
image, ResourceMaterializer::Apply appends a point-filtered copy of the
sampler after the original samplers. The resource snapshot only holds
the original samplers, so DescriptorBindingBuilder::Populate threw
"guest sampler index is out of range" for the copy, both in the sampler
binding and in the unnormalized-coordinate check.

SamplerResource::copyOf records the sampler a copy was made from, and
the guest sampler descriptor of a copy is read from that sampler; the
point-filter rewrite of word 2 still applies to the copy. The field is
part of the disk cache encoding: cache format version 8.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-08 14:55:18 +02:00
Adrià 377e819612
Merge pull request #1400 from ebarr24/codex/batch-thread-atexit-abi
fix(libc): bridge guest thread destructors to host ABI
2026-10-08 14:52:26 +02:00
Adrià b82c1c2f70
Merge pull request #864 from jeffrey12cali/fix/dlsym-default-scope
fix(module): resolve the default dlsym scope against the main program
2026-10-08 14:52:14 +02:00
Adrià e08c816abc
Merge pull request #1383 from nisbenz/fix/shader-memtime-narrow-subgroup-clock
fix(shader): read s_memtime from the device clock where RADV's subgroup clock is 20 bits
2026-10-08 14:51:41 +02:00
Adrià 0596a7a274
Merge pull request #1385 from gjonn/feat/recompiler-lwe-on-samples
feat(recompiler): accept lwe on MIMG samples and gathers
2026-10-08 14:43:27 +02:00
nisbenz f025a9e506 fix(shader): read s_memtime from the device clock where RADV's subgroup clock is 20 bits
On GFX10.3 to GFX11.7, ACO lowers a subgroup-scope OpReadClockKHR to
s_getreg_b32 of SHADER_CYCLES, 20 bits zero-extended to 64, so s_memtime
wrapped every 2^20 cycles and went backwards. Detect those families from
the RADV device name and read the device clock there instead.

Fixes #985

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-08 14:42:40 +02:00
Adrià 3c4bdda47a
Merge pull request #1379 from gjonn/fix/recompiler-image-load-missing-coordinates
fix(recompiler): read image address components the instruction omits as 0
2026-10-08 14:42:10 +02:00
Adrià 07c2405c35
Merge pull request #1371 from TibixDev/pr/depth-maintenance-precheck
fix(agc): reject depth maintenance before shader compilation
2026-10-08 14:37:01 +02:00
Adrià 072804fd33
Merge pull request #1288 from Celegans12/fix/agc-dcc-filled-keys-refresh
fix(agc): forget a DCC key fill once a refresh sees other keys
2026-10-08 14:35:47 +02:00
Adrià cba8d86da8
Merge pull request #1270 from Celegans12/fix/agc-unaligned-buffer-views
fix(agc): bind guest buffer views that start off a DWORD boundary
2026-10-08 14:34:45 +02:00
Adrià Franch aee56e17f4 test(libc): use generic file names in the path-case test
Co-Authored-By: Claude <noreply@anthropic.com>
2026-10-08 14:33:20 +02:00