* Enable 'Accurate Multiplication' by default.
* Move 'Disk Shader Cache' to the 'Advanced' tab
* Prevent enabling 'Disk Shader Cache' when 'Enable Hardware Shader' or 'Accurate Multiplication' is disabled.
* Do not load 'Disk Shader Cache' when 'Accurate Multiplication' is disabled.
* Add a tooltip for 'Disk Shader Cache'.
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.
Co-Authored-By: LC <712067+lioncash@users.noreply.github.com>
* Forward declare ui and use unique_ptr
* ConfigureEnhancements: use unique_ptr for ui
* Use make_unique instead of new where applicable
* Move some of the ui includes that already used unique_ptr
* main.cpp: also make use of make_unique on Config
* Address review comments
The deep recursion has caused issues in certain games with large numbers of files, especially with MSVC builds.
Previously the recursion depth is about equal to the number of files present. With this the depth should be about equal to the maximum depth of the directory structure of the RomFS.
Consistency change with how we mark constants in the rest of the
codebase.
Co-Authored-By: LC <712067+lioncash@users.noreply.github.com>
Co-authored-by: LC <712067+lioncash@users.noreply.github.com>
The list of points is returned by const reference, so we don't need to
make a copy of every element in the list.
Co-authored-by: Lioncash <mathew1800@gmail.com>
- In `SetCurrentThreadName`, when on Linux, truncate to 15 bytes, as (at
least on glibc) `pthread_set_name_np` will otherwise return `ERANGE` and
do nothing.
- Also, add logging in case `pthread_set_name_np` returns an error
anyway. This is Linux-specific, as the Apple and BSD versions of
`pthread_set_name_np return `void`.
- Change the name for CPU threads in multi-core mode from
"yuzu:CoreCPUThread_N" (19 bytes) to "yuzu:CPUCore_N" (14 bytes) so it
fits into the Linux limit. Some other thread names are also cut off,
but I didn't bother addressing them as you can guess them from the
truncated versions. For a CPU thread, truncation means you can't see
which core it is!
The general pattern is to mark mutexes as mutable when it comes to
matters of constness, given the mutex acts as a transient member of a
data structure.
Co-Authored-By: LC <lioncash@users.noreply.github.com>
These are intentionally discarded internally, since the rest of the
public API allows querying success. We want all non-internal uses of
these functions to be explicitly checked, so we can signify that we
intentionally want to discard the return values here.
The settings.h file doesn't actually need all of the definitions
on cam.h, only some of the enums. They can, therefore, be separated
into another file, which is included by settings.h instead.
The other changes are fixing files that included settings.h and
depended on indirect includes from includes of includes of cam.h
In all usages of LogSetting(), string literals are provided.
std::string_view is better suited here, as we won't churn a bunch of
string allocations every time the settings are logged out.
While we're at it, we can fold LogSetting() into LogSettings(), given
it's only ever used there.
Co-Authored-By: Mat M. <lioncash@users.noreply.github.com>
In cases where the size is not a known constant when inlining, AlignUp<std::size_t> currently generates two 64-bit div instructions.
This generates one div and a cmov which is significantly cheaper.
Some of the classes in this file already do this, so we can apply this
to the other ones to be consistent.
Allows these classes to play nicely and not churn copies when used with
standard containers or any other API that makes use of
std::move_if_noexcept.
* Change "Toggle Speed Limit" to toggle between 100% and a custom value
This will change the shortcut for "Toggle Speed Limit" to make it swap between 100% and the value of "Limit Speed Percent" in the config. Old functionality is still there, but renamed to "Unthrottle".
* Complete reimplementation of the function
* Fix something that didn't get saved correctly
* Fix missing indentation
* Rewrite to keep only a single QSpinBox
* Second rewrite
* set Unthrottled to 0 in the Qspinbox
* Hotkey for Unthrottle
* minor improvements to the design
* Apply suggestions from code review
Co-authored-by: Ben <bene_thomas@web.de>
* Default slider values
* clang-format fixes
* Prevent the speed slider from changing size
...when an element in its row has variable width.
* Change "Game Speed" to "Emulation Speed"
* Apply suggestions from code review
`game_speed` to` emulation_speed`
Co-authored-by: Valentin Vanelslande <vvanelslandedev@gmail.com>
* Fix for QSliders
* Revert "Prevent the speed slider from changing size"
This reverts commit ddaca2004484f1e024f49d2e6dc99ef5e261f64d.
* clang-format
...doesn't seem to stick to a choice
* Fix 2 for QSliders
Co-authored-by: B3n30 <benediktthomas@gmail.com>
Co-authored-by: Ben <bene_thomas@web.de>
Co-authored-by: Valentin Vanelslande <vvanelslandedev@gmail.com>
We can adjust the API to allow std::size_t indices, which simplifies
operating with standard container types. It also prevents truncation
warnings from occurring in these cases as well.
* archive: Make use of std::pair for OpenFileFromArchive
The tuple only makes use of two elements, so we can use a pair to
simplify the number of underlying template instantiations that need to
be done, while also simplifying the surrounding code.
* archive: Simplify MakeResult calls
MakeResult can deduce the contained result type based off the type of
the variable being passed to it, so explicitly specifying it is a little
verbose.
* Avoid deadlock when stopping video dumping
* Use static_cast, make quit atomic
* One more atomic load
* Use suggested lock instead of atomic
* Fix locking
While we're at it, we can also improve some of the allocations and
copying that would be going on in one case by preallocating and then
emplacing before modifying.
This might be caused by some packet/status delay within the enet protocol, or may be caused by our not properly handling disconnection. Anyway, this situation *can* happen, and we should not crash Citra for a non-critical error.
This change is not HW tested as I do not have 3 3DSes, but this should make sure that all the members of the network hold the same `node_info`, `nodes` and the bitmasks, etc.
Previously, when connecting, the host was using the incorrect node_id to update `node_info`.
This is an attempt to fix tywald's problem with MH, reported on Discord a while ago. I'm not sure if this would actually fix that though.
I'm not sure why we decided to have a boolean here, but apparently that wasn't the correct behaviour. According to HW tests, the Software Keyboard simply displays the default text when the button text provided is empty (**not necessarily all zero**). For example, if you set a text for one of the buttons and leave others empty, the button you set will have your text, while others will have their default texts. Removed the boolean and updated frontend code to make it correct.
* gl_rasterizer_cache: Mark file-scope functions as static where applicable
Prevents -Wmissing-declaration warnings from occurring and also makes
these functions internally linked.
* gl_rasterizer_cache: Remove unused local std::string variable
Despite being unused, compilers are unable to completely remove any code
gen related to the construction and destruction of this variable, since
the destructor of std::string is non-trivial.
Thus, we can remove it to reduce a minor amount of unnecessary code
generation
* gl_rasterizer_cache: Mark hash implementation as noexcept
This shouldn't throw.
* gl_rasterizer_cache: Remove unused variable in ClearAll()
* gl_rasterizer_cache: Make use of const on references explicit
While declared as auto&, these actually behave as const auto& variables,
due to the constness of the container being iterated. We can make this
explicit for readability sake.
* gl_rasterizer_cache: Resolve truncation warnings
The size is forwarded to a std::memset call, which takes a std::size_t
as its size parameter, so we can just make this change to silence the
warnings.
* gl_rasterizer_cache: Resolve variable shadowing warnings
Prevents a -Wshadow warning from occurring.
* GUI: Deadzone controls for sdl engine at configuration input
Co-Authored-By: CJ Bok <cjbok@users.noreply.github.com>
* configure_input: Use slider to edit modifier scale
Co-Authored-By: Kewlan <kewlan@users.noreply.github.com>
* Address minor review comment
Co-Authored-By: Kewlan <kewlan@users.noreply.github.com>
Co-authored-by: CJ Bok <cjbok@users.noreply.github.com>
Co-authored-by: Kewlan <kewlan@users.noreply.github.com>
It's undefined behavior to pass a null pointer to std::fread and
std::fwrite, even if the length passed in is zero, so we must perform
the precondition checking ourselves.
A common case where this can occur is when passing in the data of an
empty std::vector and size, as an empty vector will typically have a
null internal buffer.
While we're at it, we can move the implementation out of line and add
debug checks against passing in nullptr to std::fread and std::fwrite.
Prevents the internal buffer in the std::optional from being zeroed out
unnecessarily and instead sets the validity byte only in some
implementations.
While we're at it, we can make use of std::move to eliminate unnecessary
heap reallocations from occurring.
Allows us to avoid even more string churn by allowing the AddLine
function to make use of fmt formatting so the string is formatted all at
once instead of concatenating multiple strings.
This is similar to how yuzu's decompiler works, which I've made function
the same way in the past.
Quite a few service functions are stubbed but still pop all their
arguments, which can lead to unused variable warnings.
We can mark the unused arguments with [[maybe_unused]] to silence these
warnings until a full implementation of these functions are made.
Allows implementations to allocate the object and the shared_ptr control
block in one allocation instead of needing to do two separate
allocations.
Also looks much nicer to the reader.
* swkbd: Fix a bug where clicking Cancel hangs the game
The text is validated in `Finalize`. If the validation fails, an error is returned and the applet is not actually finalized. This can result in hangs.
This is usually not a problem as the frontend is expected to validate the text passed to `Finalize`. However, when the user clicked on `Cancel`, the text is ignored and the frontend won't do any validation. Therefore, we should skip the validation here as well.
Also fixed a potential data race. All these functions should now be called on the same thread
* Address review comments
Renamed the fields
Remove close button
This class is memcpy-ed and memcpy has the requirement that data passed
to it must be trivially copyable, otherwise the behavior is undefined.
This is trivial to resolve as BitField was made trivially copyable a
while ago, so this explicit copy assignment operator isn't necessary.
This can be removed as it's not used. Even if it were however, it would
be an incorrect forward declaration, as ServiceManager exists within the
Service::SM namespace, not the top-level SM namespace.
Same behavior, no heap allocation.
strings returned from glGetString() are guaranteed to be static strings,
so this is safe to do. They're also guaranteed to be null-terminated.
Some implementations can use the std::nullopt_t constructor of
std::optional to avoid needing to completely zero out the internal
buffer of the optional and instead only set the validity byte within it.
e.g. Consider the following function:
std::optional<std::vector<ShaderDiskCacheRaw>> fn() {
return {};
}
With libc++ this will result in the following code generation on x86-64:
Fn():
mov rax, rdi
vxorps xmm0, xmm0, xmm0
vmovups ymmword ptr [rdi], ymm0
vzeroupper
ret
With libstdc++, we also get the similar equivalent:
Fn():
vpxor xmm0, xmm0, xmm0
mov rax, rdi
vmovdqu XMMWORD PTR [rdi], xmm0
vmovdqu XMMWORD PTR [rdi+16], xmm0
ret
If we change this function to return std::nullopt instead, then this
simplifies both the code gen from libc++ and libstdc++ down to:
Fn():
mov BYTE PTR [rdi+24], 0
mov rax, rdi
ret
Given how little of a change is necessary to result in better code
generation, this is essentially a "free" very minor optimization.
Previously, we were returning a value that was way too big, causing an integer overflow in Fractured Souls.
According to wwylele, the biggest oberserved save size for 3DS is 1MB, so this new value should leave plenty of room, even if games use a bigger size.
Many of these functions are capable of being used within const contexts,
so we can apply the const qualifier in some cases and add const based
overloads for others, which makes the interface a little bit more
flexible and const-correct.
* core/memory: Amend unusual return value of operator=
operator= usually returns a reference to this. Given there's no comment
explaining why void was used, this can be assumed to be an oversight.
* core/memory: Make use of std::move in Entry::operator=
Same behavior, minus the need for an atomic reference count increment
and decrement (since MemoryRef contains a std::shared_ptr).
