Solid scene draws whose position reads an object's own uniform blocks (bones c5, place c7, c6) are drawn a second time
with a vertex shader variant (translator patch 0012) that computes last frame's position from last frame's blocks; the
game writes those into two copies that take turns, and the uniform snapshot keeps both (vk_mvprev.c). The draws are
listed during the replay and recorded through the recorder thread with repeated state skipped. Blocks whose last-frame
copies measured wrong (c9 and c11 of walls and cliffs) keep this frame's; positions without an object block (terrain,
grass) keep the camera's motion, which measured as good.
The camera motion is reprojected in double precision on the CPU: the float inverse made noise on far surfaces and a
still camera that DLSS followed into a wobble. DLSS gets Ultra Performance (a third of the output, also in the F1 menu)
and NVIDIA's recommended preset per mode (K, M for Performance, L for Ultra Performance; TOTK_DLSS_PRESET overrides).
Test switches: TOTK_MV_CHECK measures the motion vectors and DLSS's output against the picture, TOTK_DLAA_TIMES times
DLSS and the motion passes on the GPU. A reactive mask pass for transparents is included but off (TOTK_DLSS_REACTIVE):
its mask covers sky and fog too.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
NVIDIA's DLSS guide asks for 8 x (output / scene)^2 jitter phases: 18 in Quality, 24 in Balanced, 32 in
Performance. The sequence was 16 phases in every mode; it keeps 16 as the least (DLAA).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
1.4.3 sets its render size once, at start (main+0x7000e67028 calls
main+0x585bd0(renderer, force)). The function-entry hook catches that call's
renderer, and a TOTK_DLSS change writes the new scene size to the patch page and
has the next nvnQueuePresentTexture call the setup again with force set, on the
game's thread between frames, with the guest registers kept. The heap is sized
for the largest scene a mode asks for. The DLSS texture bias is part of the
sampler cache key and follows the mode and the dlaa switch. The game's scaling
pass is told from the in-place FXAA by its missing depth attachment, not by the
sizes, which change a frame ahead of the drawing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A GPU without ASTC gets the textures decoded to RGBA8, four times the memory of BC3, and on a 4 GiB card
that fills video memory: a GTX 1650 went from 10-23 to 3-12 ms of GPU time a frame with astc_bc = 1
(#87). BC3 loses some precision (normal maps), so it stays off where memory is not short. astc_bc unset
now turns it on when the largest device-local heap is under 5 GiB and logs the choice; 0 and 1 still
decide. The menu row becomes Auto / Off / On, and the description says what Auto does in all four
languages.
On an RX 9070 nothing changes. The Radeon iGPU on Windows reports its shared 10.5 GiB as device-local,
so it keeps RGBA8 as well. All probes pass, probe-astc and probe-cfg included.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
replay_async was off because no timing on our machines showed a gain. Where the game waits for the GPU
it does: a GTX 1650 in the Depths went from 450-686 frames over 50 ms a minute to 10-13, and the 1% low
from 10-12 to 13-24 fps (#88). An RTX 5090 ran slower with it off (#459). Where the replay or the game
threads bound the frame it is neutral (machines A and B, the Steam Deck in #308), and no crash report
ties to it. MacNativeGames has its translation thread on by default for TotK. It adds up to a frame of
latency; replay_async = 0 turns it off.
The fight scene on an RX 9070 with it on: 60.0 fps mean, 1% low 56.3 in the last window, no frame over
50 ms, 0 validation errors.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
GPL was turned off on that driver branch because its device losses at boot (#5) came with a cold
pipeline cache. Their cause turned out to be GPU writes past 3D BC images, fixed hours later by the
3D BC stand-ins in vk_exec.c. With GPL off every new pipeline is built whole on the draw's path: an
RX 6600 log in #381 has frames of 0.5-1.2 s spent in pipeline builds.
On a Radeon iGPU (RDNA 2, driver 26.10.1, the RX 6600 users' driver) the 0.3.4 release with GPL on
booted from a cold cache and ran the fight scene: 921 fast links, none failed, no device loss, 0
validation errors. Pipeline waits in play dropped from 1381 ms (longest 225 ms) with GPL off to 481 ms
(longest 26 ms). TOTK_GPL=0 still turns it off.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A fault in a GPU driver, the C runtime or another DLL, or a call through a null pointer, ended the process with
nothing in totk.log: crash_veh reports only faults whose rip is in this executable (#506, #531, #488, #415).
Windows: an unhandled-exception filter reports every exception no handler took. It runs only after every handler
declined, so exceptions a driver raises and catches itself never reach it, and it costs nothing until a crash. The
report names the module (from its export directory, no loader lock) and the return addresses in totk.exe on the
faulting thread's stack, unwound from the faulting context.
Linux: fault_handler passes faults in libraries (Mesa, SDL) and at address 0 to the crash hook as well, with the
library name from dladdr.
--probe-crash-call (strlen in ucrtbase.dll reads address 8) and --probe-crash-null crash on purpose. Before: exit
0xc0000005 and no FATAL line. After: "FATAL: access violation ... (ucrtbase.dll+0xf2b61)" with host frames
probe_crash and main, exit 71.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The settings menu offers 1920x1280, 2160x1440, 2256x1504, 2880x1920 and 3000x2000 (the 3:2 screens of Surface and
Framework laptops and tablets), and the setup builds romfs-patches-<w>x<h> for each, so the HUD moves to the top and
bottom edges as it does at 16:10 instead of being stretched.
clamp(x, 0.0, x) is undefined for negative x in GLSL.std.450, so the output clamps of the game's FSR pass now become
FMax(x, 0.0), one word shorter, with the rest of the module moved down. When the FSR shader is recognised but none of
its output clamps is found (a translator that stores the colour another way), the log now says the highlights are
still clipped instead of staying silent. The comment also notes that the game takes this pass at 1:1 too.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The camera view draws a half-size RGBA16F target with a half-size D32F depth after the lighting (852x480 under a
1704x960 scene). It matched the light target pattern, so the DLAA point re-armed on it and DLSS scaled it into the
scene's own light target, which left every object black. An armed pair is no longer replaced by a smaller one, and
the scaling pass DLSS writes into must have the output size.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With a DLSS mode the game's own scaling draw covered DLSS's picture with the jittered scene. At a 1280x720 scene and
1920x1080 output the scaling is done by the game's tonemap, not by the FXAA shader vk_dlaa_skip_draw looks for, so no
hash list could catch it. From DLSS's evaluation to the frame's end, passes larger than the light target now sample
DLSS's output in its place (vk_dlaa_swap_img, applied in vk_exec_draw.c tex_info), whichever shader scales.
The motion-vector shader inverted the view-projection in 32-bit floats. With a still camera (identical matrices each
frame) the rounding still gave a few hundredths of a pixel of motion that changed with every jitter step; DLSS
followed it and the whole picture wobbled. mv_pass now makes last frame's view-projection times the inverse of this
frame's in double precision from the uniform snapshot (vk_usnap_host), and the shader only applies it.
Measured on 1.4.3 indoors, still camera, mean per-pixel change between consecutive frames on static areas:
DLSS quality 0.18 -> 0.06, DLAA 0.14 -> 0.05. A camera pan at 10 px per frame shows no ghosting.
Some frames the game upscales its tonemapped scene with its own FSR 1 pass instead of the plain upscale, and that
shader clamps each channel to 0..1, so with TOTK_HDR=1 every highlight above the HUD's white is cut back to it: around
the sun the glare turns into a flat grey-white patch that stays until the camera moves. vk_hdr.c fsr_unclamp finds the
shader by FSR's two reciprocal constants and drops only the clamps whose result goes to its output, leaving the
filter's own weight clamps alone.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A mod's folder name is drawn on the Mods tab (opt_text_row, menu.cpp) and as the Cheats tab's group heading (cheats_tab.cpp), and it comes from disk at runtime like a cheat name, so its glyphs were never in the atlas either.
Both sources are added in the same place in fonts_up(), so one pass covers the whole menu.
After the in-game menu or the text entry closes, input_poll kept the game's input blocked until one poll saw no
button and all four stick axes at 0. An input that never reads as released (a button another program holds down, a
key whose release went to another window, a stick resting outside its dead zone) kept the game deaf until restart
(#462, #494). Closing with Minus+Plus hid this: that poll zeroes the pad before the gate looks. Closing with Esc did
not. The wait now ends after 1 s with input held, and the log names what was held.
Repro: field save, a script holding the right stick at 3000, Esc opens and closes the menu, left-stick pulses after.
Before: 0.00 m/s for the 30 s after the close. After: Link walks from the second poll window on; the log line reads
"input: the menu closed 1 s ago and input is still held (buttons 0, sticks 0 0 3000 0)".
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The motion vectors follow the camera only, so a character, animal or held item that moves on its own drags a faint
copy of itself through DLSS's history. TotK's G-buffer target 0 carries an object class in green: above 64/255 for
characters, creatures and held items (Link 253, others 125, held items 160), 0 for terrain and 32 for foliage, checked
in the field, in Goron City and on horseback. The motion-vector pass now also writes DLSS's bias mask
(DLSS.Input.Bias.Current.Color.Mask) where such an object is now or was the frame before, read through the motion
vector from the previous frame's class mask, so the history that holds the moving object is weighed down there.
dlss_char_mask (0, 35, 65 or 100 percent, default 65; F1 > Screen > "DLSS character trails", live) sets the strength.
A 640x360 pass of the G-buffer's shape, seen while loading, no longer replaces the scene's class buffer in a frame.
The status line also prints after 300 frames, with the frames that had the mask.
dlss_preset (default, k or j; F1 > Screen > "DLSS model", live) sets the render preset hints when a DLSS feature is
created. default leaves the choice to NGX (preset K for DLAA, Quality and Balanced in the 310 releases); j is the
transformer model NVIDIA describes as slightly less ghosting at the cost of more flicker. A change recreates the
features at the next DLAA run. FrameTimeDeltaInMsec now carries the time since the previous evaluation instead of a
fixed 16.67 ms. The menu and totk.ini.example name the DLSS model file of the platform in use.
vk_dlaa.c reached the NGX core only through Windows: the registry, LoadLibraryA and GetModuleFileNameW. On Linux the
NVIDIA driver installs the core as libnvidia-ngx.so.1 with the same Vulkan entry points; it is opened with plat_lib_open.
The parameter object's vtable follows the Itanium C++ ABI there, each overload group in declaration order rather than
MSVC's reverse order, so the slot enum has a Linux variant; vtable_ok confirms it at start. NGX gets the executable's
folder as a 32-bit wchar_t string, converted from UTF-8 without the C locale, and loads libnvidia-ngx-dlss.so.<version>
from there. The Windows path is unchanged.
resolution.c sized the scene for a DLSS mode only on Windows, where it looks for nvngx_dlss.dll; on Linux it now looks
for libnvidia-ngx-dlss.so.<version> next to totk, so quality, balanced and performance work there too. The no-op
stubs vk_dlaa.c had for other platforms go, as Linux now builds the real code.
In mirror mode the replay copied every range a shader wrote from the VRAM mirror back to guest memory at the end of the
submission. TotK's particle emitters write a new particle's whole record into a free slot of a buffer the simulation
shaders update, so the copy put the mirror's stale words over fresh spawns: the waterfall splashes and mist showed about
once a second and the falling rain drifted. Now each written range gets a snapshot in the prologue of the part that
first writes it, after the CPU-to-mirror copies, and at the end of the submission shaders/wb_merge.comp writes only the
32-bit words that differ from the snapshot. Words the GPU left alone keep what the CPU wrote. About 0.1 MiB and three or
four dispatches a frame in rain. TOTK_WB_MERGE=0 restores the copies. The merge's push descriptors go through rec_push
so they keep their place in the recorder's call order.
The device sink loop now writes every output channel; upstream's TOTK_AUDIO_DUMP gets the front pair before clipping.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
While TotK waits for a tag it stops and starts detection several times a second, tens with the menu open: the Amiibo
tab's reader line flipped between ready and searching and the log took two lines per restart. The tab now shows
searching for a second after a stop (amiibo_get_reader_state; the game still reads amiibo_get_device_state), and the
log names the first three restarts, then a count every 200th.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Amiibo tab's folder field took no typing: the menu never passes typed characters to Dear ImGui, so "Open" only
re-read "./amiibo". A missing folder said so at the bottom of the right pane, below its scroll.
"Open" now shows the shell's folder dialog on Windows (IFileOpenDialog on a thread of its own, owned by the game
window) and SDL_ShowOpenFolderDialog on the SDL window thread elsewhere; the tab shows the folder as text and names a
missing one inside the empty list.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The DLSS modes already apply only with dlaa on (3de0839), so the DLL check no longer looks at dlaa, and off Windows,
where vk_dlaa.c is no-op stubs, a DLSS mode no longer shrinks the scene. fxaa = 0 names the FXAA shader it skipped,
or reports a minute of frames without it, so a translator change that moves FXAA_FS shows on any GPU.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With a DLSS mode (a scene smaller than the output) DLSS writes the scaled picture into the scaling pass's target and
vk_dlaa_skip_draw skips the game's own scaling draw, known by its fragment shader's SPIR-V hash. Translator patch
0011 (the shadow filter spread) changed that hash, so the draw ran again and covered DLSS's picture with the scaled
jittered scene: the whole picture shook by a fraction of a pixel each frame, most visible on lit flat surfaces in
shrines. The FXAA switch (fxaa = 0) stopped working for the same reason.
Add the hash of the current translator, and log once when DLSS scaled a frame whose scaling draw was not found, so
the next translator change that moves the hash shows in the log.
Measured on 1.4.3, Ukouh Shrine, output 3200x1800, dlss = quality, mean absolute change between consecutive frames
with a still camera: 1.40 before, 0.07 after (0.054 with dlaa = 0). Outdoors on stone: 4.29 before, 1.08 after.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A DLSS mode (quality, balanced, performance) shrank the scene to 67, 58 or 50 % of the output and sharpened the
material textures' mip bias to match, even when nvngx_dlss.dll was missing and NGX reported DLSS as not available.
The game's plain scaling then stretched the smaller scene, and the negative bias made textures shimmer in motion.
With dlaa = 1 the scene size now waits on nvngx_dlss.dll next to totk.exe (checked before the device exists, as
the size is decided then) and follows render_scale without it. The mip bias applies only when NGX came up with
DLSS. dlaa = 0 keeps the documented size with the game's scaling, without the bias.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
tex_mirrored (vk_texture.c) read t->pool->flags whenever the back end created the host entry of a texture, which can
happen long after nvnTextureInitialize. A game may finalize a memory pool and free the memory the NVNmemoryPool object
lived in while textures from it are still looked up (on a level change, for instance); reading the freed object then
faults. The NVN driver keeps no pointer to the pool either. tex_t now records the pool's flags at Initialize, in the
padding after layout[8], so NVNtexture's size is unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Device-paced frames rendered back to back after each period pull (1024
samples on PipeWire, 4-5 frames), before the game's audio thread could
clear the render event or fill the AUX return ring. Frames now stay at
least 4/5 of a frame apart, as the DSP renders one frame every 5 ms on
hardware. This fixes AUX clicks at frame seams (#426) and the progressive
sound loss of #131 (overlapping render notifications 52.7% -> 0.1%).
Linux prefill default raised to 60 ms: at 20 ms PipeWire underran
thousands of times within minutes in busy scenes.
The reverb ring (#421) and advance_voice stepping parts of this change
landed in dev separately (d22a19c, b44caca).
Fixes#131
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
TOTK_REC_CHECK=1 searches the packets of the part being recorded for each image, view and buffer vk_memory.c defers or
destroys. A deferral stamped below the part's value, or a destruction of an object a packet names, is logged with the
packet op, both stamps and the GPU's value, and counted by kind and op in a summary at exit. put() zeroes each packet's
room while the check is on: packers reserve room for optional arrays they may leave unwritten, and those words held
handles from the ring's previous lap. TOTK_REC_DELAY_US=N makes the recorder wait N us before each batch, as on a slow
machine. rec_begin takes the part's value for the check. Both are off by default; off, the recorder pays one test per
batch and per packet.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
vk_memory.c defer stamped every deferred image, view and buffer with the last submitted timeline value. An object the
replay lets go of while it records a submission can be named by that submission too: vk_texture.c slot_resolve hands a
draw a released entry, and entry_of destroys released entries at that address later in the same submission. The
completion thread then destroyed the object once the GPU reached the earlier value, before the submission naming it
ran, and with the recorder thread behind, before the recorder made the call. The stamp is now the later of the last
submitted value and vk_exec_recording_fence, as LAS Recomp and Savage Wind Recomp stamp their deferrals
(runtime/nvn/vulkan/textures.c defer). TOTK_DEFER_STAMP=0 brings back the old stamp.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The asynchronous preemption stub saved and restored YMM0-15 with vmovdqa and ran vzeroupper in inline asm, which
-march=x86-64-v2 does not reach. The v2 runtime library therefore carried VEX code, and a CPU without AVX stopped
with an illegal instruction the first time a core owner was interrupted, a few seconds into boot (issue #454,
Pentium Gold G6400, totk+0xe7a2b50 = kern_ap_stub+0x20). Without __AVX__ the stub now saves XMM0-15 with movdqa into
the same slots; v2 game code keeps nothing in the upper YMM halves. The v3 stub is unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The replay marks the start and the end of every render pass and dispatch
(vk_exec.c vk_exec_pass_begin and vk_exec_pass_end, vk_exec_draw.c
op_dispatch and op_dispatch_ind), outside the pass so no pass is broken.
With VK_AMD_buffer_marker the GPU writes the last started and finished mark
into words 2 and 3 of the breadcrumb buffer; with
VK_NV_device_diagnostic_checkpoints the same points are checkpoints, read
with vkGetQueueCheckpointData2NV after the loss. A ring of 4096 records maps
a mark to its submission, index, target formats and size, draw count and the
input ids of the first and last draw's shaders. vk_device_fault_report logs
both marks with their records. The calls go through the recorder thread.
TOTK_PASS_CRUMBS (menu: GPU pass marks): unset, AMD buffer marks only; 1,
also NVIDIA checkpoints, whose extension is enabled only then; 0, none.
The design follows Savage Wind Recomp 3fae40c; the code is ours.
--probe-vkexec gets a pass marks part: pass, dispatch, pass in one
submission; the marks read back after the fence name the last pass and the
report text says so.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
SignalAndModifyByWaitingCountIfEqual with count <= 0 and at least one waiter stored value - 1; Horizon stores
value - 2 (mesosphere KAddressArbiter::SignalAndModifyByWaitingCountIfEqual; yuzu, LAS Recomp and nxnative agree).
The word was already updated with compare-and-swap, as LAS Recomp 1144aea made theirs; the count > 0 values were
right.
--probe-kernel arbiter part expected the old value: it now wants 4 for value 6 with one waiter and count -1, and
4 for two waiters and count 0. Before: 5 for both. After: ok.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Ours took a pending CancelSynchronization first, before the signalled objects and before the timeout-0 return, so a
poll or a wait on a signalled object used it up and the blocking wait it was meant for ran on. Horizon checks a
signalled object, then timeout 0, then the cancel (mesosphere KSynchronizationObject::Wait; LAS Recomp 1144aea moved
the poll the same way): both keep it pending. A cancel that wakes a blocked wait still ends it cancelled:
CancelSynchronization now wakes it with R_CANCELLED, which the loop tells apart from a signal's -1.
--probe-kernel events part: a cancel sent before a poll and before a wait on a signalled event; each is followed by a
blocking wait with a 500 ms timeout. Before: the poll and the signalled wait returned 0xec01 (cancelled) and the
blocking waits timed out (0xea01). After: timeout, success, then cancelled both times.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
vk_pool_init gave a size-0 pool with storage a record: the host import of 0 bytes failed (the driver refuses the
memory, VK_ERROR_INVALID_EXTERNAL_HANDLE) and the record went into the pool index as the empty range [mem, mem]. A
lookup at that address then found the empty pool before the pool that starts there and reported it unbacked. TotK
makes such pools now and then (MacNativeGames d3407e29, 1.2.1). Such a pool now gets no record, as a GPU_NO_ACCESS
pool; vk_pool_fini finds none and does nothing. Removal by key already kept a neighbour's range safe at Finalize.
--probe-vkexec gains an empty pool part: a zero-byte pool at the start of another, lookups there before and after
its Finalize. It failed before ("a lookup at the start of a pool finds the empty pool there").
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>