This Clang warning complains when offsetof is used on a
non-standard-layout type (i.e. any class using various C++ features),
even though it works fine (and is not undefined behavior as of C++17).
Removes all remaining usages of the global system instance. After this,
migration can begin to migrate to being constructed and managed entirely
by the various frontends.
Force early fragment tests when the 3D method is enabled.
The established pipeline cache takes care of recompiling if needed.
This is implemented only on Vulkan to avoid invalidating the shader
cache on OpenGL.
error: unknown warning option '-Werror=unused-but-set-parameter'; did you mean '-Werror=unused-parameter'? [-Werror,-Wunknown-warning-option]
error: unknown warning option '-Werror=unused-but-set-variable'; did you mean '-Werror=unused-const-variable'? [-Werror,-Wunknown-warning-option]
- Use .at() instead of raw indexing when dealing with untrusted indices.
- For the special case of WaitFence with syncpoint id UINT32_MAX,
instead of crashing, log an error and ignore. This is what I get when
running Super Mario Maker 2.
EmuWindow::PollEvents was called from the GPU thread (or the CPU thread
in sync-GPU mode) when swapping buffers. It had three implementations:
- In GRenderWindow, it didn't actually poll events, just set a flag and
emit a signal to indicate that a frame was displayed.
- In EmuWindow_SDL2_Hide, it did nothing.
- In EmuWindow_SDL2, it did call SDL_PollEvents, but this is wrong
because SDL_PollEvents is supposed to be called on the thread that set
up video - in this case, the main thread, which was sleeping in a
busyloop (regardless of whether sync-GPU was enabled). On macOS this
causes a crash.
To fix this:
- Rename EmuWindow::PollEvents to OnFrameDisplayed, and give it a
default implementation that does nothing.
- In EmuWindow_SDL2, do not override OnFrameDisplayed, but instead have
the main thread call SDL_WaitEvent in a loop.
Add a std::bit_cast-like function archiving the same runtime results as
the standard function, without compile time support.
This allows us to use bit_cast while we wait for compiler support, it
can be trivially replaced in the future.
VirtualBuffer makes use of VirtualAlloc (on Windows) and mmap() (on
other platforms). Neither of these ensure that non-trivial objects are
properly constructed in the allocated memory.
To prevent potential undefined behavior occurring due to that, we can
add a static assert to loudly complain about cases where that is done.
Makes page tables and virtual buffers able to be moved, but not copied,
making the interface more flexible.
Previously, with the destructor specified, but no move assignment or
constructor specified, they wouldn't be implicitly generated.
Preliminary work for upmixing & general cleanup. Fixes basic issues in games such as Shovel Knight and slightly improves the LEGO games. Upmixing stitll needs to be implemented.
Audio levels in a few games will be fixed as we now use the downmix coefficients when possible instead of supplying our own
Upon further investigation, these commands allow temporary vibrations even when the "Controller Vibration" system setting is disabled. As a result, vibrations are allowed when either the system setting or this flag is set to true. Therefore, we can only block vibrations when both flags are set to false.
Not all controllers have a SDL_GameController binding. This caused controllers not present in the SDL GameController database to have buttons mapped instead of axes.
Furthermore, it was not possible to invert the axes when it could be useful such as emulating a horizontal single joycon or other potential cases. This allows us to invert the axes by reversing the order of mapping (vertical, then horizontal).
Previously we used a vibration filter that filters out amplitudes close to each other. It turns out there are cases where this results into vibrations that are too inaccurate. Remove this and move the 100Hz vibration filter (Only allowing a maximum of 100 vibrations per second) from sdl_impl to npad when enable_accurate_vibrations is set to false.
Some games do not respond to a change in controller type if 1) The controller is not disconnected prior to being reconnected and/or 2) The controller is reconnected instantly after being disconnected.
Since it is not possible to change controllers instantly on hardware and requiring a disconnect prior to connecting a new one, we should emulate this as well with a small delay, fixing the aforementioned issue.
A vibration device is an input device that returns an unsigned byte as status.
It represents whether the vibration device supports vibration or not.
If the status returns 1, it supports vibration. Otherwise, it does not support vibration.
Allows for enabling and modifying vibration and vibration strength per player.
Also adds a toggle for enabling/disabling accurate vibrations.
Co-authored-by: Its-Rei <kupfel@gmail.com>
The implementation of these commands seem incomplete and causes rumble in Super Mario Party to stop working since only EndPermitVibrationSession is called. Thus, these are better off being marked as a stub until this can be investigated more thoroughly.
Sending too many state changes in a short period of time can cause massive performance issues.
As a result, we have to use several heuristics to reduce the number of state changes to minimize/eliminate this performance impact while maintaining the quality of these vibrations as much as possible.
This allows setting the vibration strength percentage anywhere from 1% to 100%.
Also hooks up the remaining motion button and checkbox in the Controller Applet.
Some parameters need to be doubleword aligned due to the presence of the applet_resource_user_id.
Previously, this value was invalid in many commands where it was not doubleword aligned when popped.
The first u32 describes the vibration device type which is a Linear Resonant Actuator used in Nintendo Switch controller hardware.
The second u32 describes the vibration device position, in this case distinguishing between left and right vibration actuators.
Pro Controllers have 2 LRAs each that can vibrate independently of each other, which means they have 2 distinct vibration device handles to distinguish between the two actuators.
Similarly for joycons, the left joycon can be distinguished from the right joycon through the vibration device handle since each joycon has 1 LRA.
RestoreDefaults() now restores the selected devices' mappings using UpdateMappingWithDefaults().
This allows us to move the keyboard mapping from RestoreDefaults() to UpdateMappingWithDefaults().
Previously mouse clicks will not register when touch is disabled.
This rectifies that and allows mouse clicks to be mapped to other buttons if the touchscreen is disabled.
With this, the "Input Devices" combobox should accurately reflect the input device being used and disallows inputs from other input devices unless the input device is set to "Any".
This reduces the overhead of bounds checking on each element.
It won't reduce the cost of allocation because usually this vector's
capacity is usually large enough to hold whatever we push to it.
Changes QMessageBox usages to warnings, as the problems they bring to
light are being safely handled by the application and do not warrant
something of the "critical" level.
Changes LOG_CRITICAL to LOG_ERROR for the same reason. Preferring ERROR
to WARNING as yuzu is denying loading of any guest applications after
checking for these conditions.
Moved logging the GL_RENDERER string into GetUnsupportedGLExtensions()
to make more clear that unsupported extensions were already being
logged. Makes placement of the logs easier to understand later, as well.
Changes the first message to not include the OpenGL version, as the
error is caused by OpenGL failing to load.
Adds a new check for OpenGL version 4.3. This will display a message
with a similar error as well as the GL_RENDERER string. Adds a CRITICAL
log message when triggered. This prevents a crash with yuzu trying to
use older OpenGL versions.
Modifies the unsupported extension message to output the GL_RENDERER
string in the message, as well as logging the string.
Resolves numerous deprecation warnings throughout the codebase due to
inclusion of this header. Now building core should be significantly less
noisy (and also relying on less global state).
This also uncovered quite a few modules that were relying on indirect
includes, which have also been fixed.
The interrupt handler contains a std::atomic_bool, which isn't copyable
or movable, so the special move member functions will always be deleted,
despite being defaulted.
This can resolve warnings on clang and GCC.
Some games like Cave Story+ set invalid values in the ControllerPrivateArg's mode and caller fields.
Use other fields to determine the appropriate mode and caller should either or both fields be invalid.
- This works similiar to GetAlbumContentsFileListForApplication.
- Since we do not implement the album, this should be safe to stub for now.
- Used by Super Smash Bros. Ultimate (newer updates) in World of Light.
Hides all of the implementation details for users of the class. This has
the benefit of reducing includes and also making the fiber classes
movable again.
Unicorn long-since lost most of its use, due to dynarmic gaining support
for handling most instructions. At this point any further issues
encountered should be used to make dynarmic better.
This also allows us to remove our dependency on Python.
Allows our CI to catch more potential bugs. This also removes the
[[nodiscard]] attribute of IOFile's Open member function. There are
cases where a file may want to be opened, but have the status of it
checked at a later time.
It's deprecated in the language to autogenerate these if the destructor
for a type is specified, so we can explicitly specify how we want these
to be generated.
The API of VP9 exposes a WasFrameHidden() function which accesses this
member. Given the constructor previously didn't initialize this member,
it's a potential vector for an uninitialized read.
Instead, we can initialize this to a deterministic value to prevent that
from occurring.
* The web_service http request is now fixed on Windows (R) platform.
* The issue is due to a complicated race-condition in `httplib`, a detailed
explanation is available at https://github.com/yhirose/cpp-httplib/pull/701
* A pending Pull Request on `httplib` has been applied to remedy the
said race-condition.
* The socket availability check is removed due to a behavioral chice of
`httplib` that a socket will not be created before any actual request
is sent.
This implements texture cube arrays with shadow comparisons but doesn't
fix the asserts related to it.
Fixes out of bounds reads on swizzle constructors and makes them use
bounds checked ::at instead of the unsafe operator[].
Previous to this commit, the tests were using operator[] from
unordered_map to query elements but this silently inserts empty elements
when they don't exist. If all threads were executed without concurrency,
this wouldn't be an issue, but the same unordered_map could be written
from two threads at the same time. This is a data race and makes some
previously inserted elements invisible for a short period of time,
causing them to insert and return an empty element. This default
constructed element (a zero) was used to index an array of fibers that
asserted when one of them was nullptr, shutting the test session off.
To address this issue, lock on thread id reads and writes. This could be
a shared mutex to allow concurrent reads, but the definition of
std::this_thread::get_id is fuzzy when using non-standard techniques
like fibers. I opted to use a standard mutex.
While we are at it, fix the included headers.
* A regression was in 39c8d18 and token verification function was
broken.
* The reason being `httplib` now requires OpenSSL 1.1+ API while
LibreSSL 2.x provided OpenSSL 1.0 compatible API.
* The bundled LibreSSL has been updated to 3.2.2 so it now provides
OpenSSL 1.1 compatible API now.
* Also the path hint has been added so that it will find the correct
path to the CA certs on *nix systems.
* An option is provided so that *nix system distributions/providers can
use their own SSL implementations when compiling Yuzu/Citra to
(hopefully) complies with their maintenance guidelines.
* LURLParse is also removed since `httplib` can handle
`scheme:host:port` string itself now.
This commit aims to implement the NVDEC (Nvidia Decoder) functionality, with video frame decoding being handled by the FFmpeg library.
The process begins with Ioctl commands being sent to the NVDEC and VIC (Video Image Composer) emulated devices. These allocate the necessary GPU buffers for the frame data, along with providing information on the incoming video data. A Submit command then signals the GPU to process and decode the frame data.
To decode the frame, the respective codec's header must be manually composed from the information provided by NVDEC, then sent with the raw frame data to the ffmpeg library.
Currently, H264 and VP9 are supported, with VP9 having some minor artifacting issues related mainly to the reference frame composition in its uncompressed header.
Async GPU is not properly implemented at the moment.
Co-Authored-By: David <25727384+ogniK5377@users.noreply.github.com>
It turns out that after a controller is disconnected, there is a chance that events from the previous controller are sent/processed after it has been disconnected.
This causes the previously disconnected controller to reappear as connected due to GetSDLJoystickBySDLID() emplacing this controller back to the map.
Fix this by only returning an SDLJoystick if and only if it exists in the map.
These compiler flags aren't shared with clang, so specifying these flags
unconditionally can lead to a bit of warning spam.
While we're in the area, we can also enable -Wunused-but-set-parameter
given this is almost always a bug.
This emulates the behavior we get on GLSL with regular SSBOs with a
pointer + length pair. It aims to be consistent with the crashes we
might get.
Out of bounds stores are ignored. Atomics are ignored and return zero.
Reads return zero.
Previously, the lower bound wasn't being used and zero was being used as
the lower bound every time this function was called.
This affects the outcome of some of the randomized entries a little bit,
for example, the lower-bound for beard and mustache flags was supposed
to be 1, not 0.
Aside from these cases, the bug didn't affect anything else.
Previously, disconnecting a controller still leaves a null SDLJoystick entry within the vector of SDLJoysticks mapped by GUID.
When a DirectInput device of the same GUID is reconnected, it adds that device to a new port causing non-detectable input.
Furthermore, opening the "Configure" menu would cause yuzu to crash since it first tries to resolve the name of a null SDLJoystick entry that was not removed.
Resolve this by properly erasing the SDLJoystick entry from the vector.
Locks on GetCurrentHostThreadID were causing performance issues
according to Visual Studio's profiler. It was consuming twice the time
as arm_interface.Run(). The cost was not in the function itself but in
the lockinig it required.
Reimplement these functions using atomics and static storage instead of
an unordered_map. This is a side effect to avoid locking and using linked
lists for reads.
Replace unordered_map with a linear search.
Makes our error coverage a little more consistent across the board by
applying it to Linux side of things as well. This also makes it more
consistent with the warning settings in other libraries in the project.
This also updates httplib to 0.7.9, as there are several warning
cleanups made that allow us to enable several warnings as errors.
Vulkan has requirements for primitive topologies that don't play nicely
with yuzu's. Since it's only 4 bits, we can move it to fixed state
without changing the size of the pipeline key.
- Fixes a regression on recent Nvidia drivers on Fire Emblem: Three
Houses.
RDNA devices seem to crash when using VK_EXT_extended_dynamic_state in
the latest 20.9.2 proprietary Windows drivers. As a workaround, for now
we block device names corresponding to current RDNA released products.
TMML takes an array argument that has no known meaning, this one appears
as the first component in gpr8 followed by s, t and r. Skip this
component when arrays are being used. Also implement CUBE texture types.
- Used by Pikmin 3: Deluxe Demo.
The old code had a sort function that was invalid and it didn't work as
expected when the base vector had a different order (e.g. renderdoc was
attached).
This sorts devices as expected and fixes a debug assert on MSVC.
From -fsanitize=address, this code wasn't calling the proper destructor.
Adding virtual destructors for each inherited class and the base class
fixes this bug.
While we are at it, mark the functions as final.
Using the Qt::WindowStaysOnTopHint flag allows these dialogs to show up on top while running in fullscreen. However, if yuzu goes out of focus (by alt-tabbing or otherwise), this flag does not seem to have an effect.
The previous fix only partially solved the issue, as only certain GPUs that needed 9 or less MiB subtracted would work (i.e. GTX 980 Ti, GT 730). This takes from DXVK's example to divide `heap_size` by 2 to determine `allocable_size`. Additionally tested on my Quadro K4200, which previously required setting it to 12 to boot.
While were at it, we can also enable sign conversion warnings and other
common warnings as errors to prevent these from creeping back into the
codebase.
When HEADER_GENERATOR was included in the DEPENDS section of custom
commands, msbuild assumed this was always modified. Changing this file
is not common so we can remove it from there.
Allows some implementations to avoid completely zeroing out the internal
buffer of the optional, and instead only set the validity byte within
the structure.
This also makes it consistent how we return empty optionals.
This is used in multiple games such as:
- Clubhouse Games: 51 Worldwide Classics
- Grandia HD Collection
- XCOM 2 Collection
- Baldur's Gate 1/2
- Dr Kawashima's Brain Training
- Super Mario 3D All-Stars
This is a hack to destroy all HostCounter instances before the base
class destructor is called. The query cache should be redesigned to have
a proper ownership model instead of using shared pointers.
For now, destroy the host counter hierarchy from the derived class
destructor.
This reworks how host<->device synchronization works on the Vulkan
backend. Instead of "protecting" resources with a fence and signalling
these as free when the fence is known to be signalled by the host GPU,
use timeline semaphores.
Vulkan timeline semaphores allow use to work on a subset of D3D12
fences. As far as we are concerned, timeline semaphores are a value set
by the host or the device that can be waited by either of them.
Taking advantange of this, we can have a monolithically increasing
atomic value for each submission to the graphics queue. Instead of
protecting resources with a fence, we simply store the current logical
tick (the atomic value stored in CPU memory). When we want to know if a
resource is free, it can be compared to the current GPU tick.
This greatly simplifies resource management code and the free status of
resources should have less false negatives.
To workaround bugs in validation layers, when these are attached there's
a thread waiting for timeline semaphores.
