Since Android is a pain when it comes to checking GPU logs in more depth, this is a better way to see what's going on, especially for testers...
This should be expanded to Mali, Xclipse, and Tensor in the future. Since I don't own any of these devices, it's up to developers with similar capabilities to add support for this system.
~~The GPU log sharing button should also be added in the future... For now, they are available in the same location as the traditional logs.~~ Added on 572810e022
Co-authored-by: DraVee <dravee@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3389
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: MrPurple666 <antoniosacramento666usa@gmail.com>
Co-committed-by: MrPurple666 <antoniosacramento666usa@gmail.com>
The `NIGHTLY_BUILD` option changes the app name to "Eden Nightly" and
changes the auto-update URL to use our new Nightly repository.
This needs added to Android, but I can't right now as I have to leave.
Signed-off-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3431
On macOS, adding paths to QFileSystemWatcher can trigger immediate 'directoryChanged' signals (likely due to FSEvent behavior or .DS_Store updates during directory access). This caused a recursive loop where PopulateAsync would trigger itself indefinitely, leading to constant UI refreshing and high CPU usage.
This fix addresses the issue by:
1. Temporarily blocking QFileSystemWatcher signals while the watch list is being populated.
2. Correcting the usage of QList::mid() in the loop (passing length instead of an end index) to avoid massive path duplication.
These changes are isolated to macOS using #ifdef __APPLE__ to ensure no side effects on other platforms.
Authored-by: rayman30
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3418
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-by: crueter <crueter@eden-emu.dev>
Co-authored-by: rayman30 <silentbitdev@gmail.com>
Co-committed-by: rayman30 <silentbitdev@gmail.com>
Metal validation requires fragment shader output types to strictly match the render target format (e.g., writing float to RGBA32Uint is invalid).
This commit:
1. Adds color_output_types to RuntimeInfo.
2. Detects Integer/SignedInteger render targets in the Vulkan backend (MoltenVK only).
3. Updates the SPIR-V emitter to declare the correct output type (Uint/Sint) and bitcast values accordingly.
This fixes the VK_ERROR_INITIALIZATION_FAILED crash on macOS.
Co-authored-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3414
Co-authored-by: rayman30 <silentbitdev@gmail.com>
Co-committed-by: rayman30 <silentbitdev@gmail.com>
"So, found another macOS crash while testing Luigi's mansion 2. It looks like Metal is pretty picky about types and was crashing because the texture gather offsets were being passed as unsigned integers instead of signed ones.
I made a small tweak to the shader recompiler to force them to be signed, and the game boots fine now. Most other drivers usually handle signed offsets anyway, so it should be a safe fix for everyone." - rayman
Authored-by: rayman
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3404
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: crueter <crueter@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Adding and editing users is now done in a single dialog rather than all those other individual buttons and dialogs like before.
Fixed some bugs with profile management too, and made edit/delete a right-click menu.
Signed-off-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3415
About overlays: just as the title suggest. added to avoid misleading user feedbacks, as now captures can bring proof of build and driver in use.
About driver version: Android always showed stock driver version (major.minor.patch) as 0.x.x and that because getSystemDriverInfo() from GpuDriverHelper driver version getter has no proper treatment for qualcomm driver version values. BUT getVulkanDriverVersion from NativeLibrary does! So i just migrated to Native's getter.
A low prio todo would be to merge all VersionString getters into a single centered one inside vk device class. Suit yourselves.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3381
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: xbzk <xbzk@eden-emu.dev>
Co-committed-by: xbzk <xbzk@eden-emu.dev>
This fixes qlaunch "+ Options" :)
More:
- Remove Starter-Applet from menu (Starter is started by qlaunch)
- Stub OLSC cmds and add IStopperObject
- Fail-safe invalid handle return for system applets
- Stub IHomeMenuFunctions::IsSleepEnabled (closes qlaunch now when hitting sleep)
- Lower BuiltInNews timeout from 10s to 2s
- Use proper Event instead of KEvent in npns
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3376
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: Maufeat <sahyno1996@gmail.com>
Co-committed-by: Maufeat <sahyno1996@gmail.com>
Some blit paths could be recorded while a render pass was still active, which is undefined behavior in Vulkan.
Depending on driver behavior and timing, this could result in intermittent visual artifacts, incomplete frames being presented, or issues with capture paths.
Explicitly request an outside render pass operation context before issuing blit commands, ensuring all rendering has fully completed before presentation or readback.
This does not change intended rendering output, but fixes a class of synchronization and command buffer state issues.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3355
Reviewed-by: DraVee <dravee@eden-emu.dev>
Co-authored-by: MaranBr <maranbr@outlook.com>
Co-committed-by: MaranBr <maranbr@outlook.com>
This pr is another set of fixes for the system driver (stock) on Qualcomm devices, mostly taking into consideration drivers from series 512.800.0 and on, reduces unwanted loads in the GPU by allocating properly features that were forced available, when they weren't, generating wrong commands between shader recompiler and Adreno's compiler; also this pr has a bunch of small fixes on games that some textures were displayed wrongly due to double swizzling during the creation of shaders, which allows serveral games to be fixed:
- Zelda: Link's Awakening (no blur, hearts properly displayed, flickering highly reduced on NCE)
- Zelda: Echoes Of Wisdom (no blur, hearts properly displayed, reduced highly flickering, rift properly rendered)
- Mario Kart 8 Deluxe (no longer wrong blue texture on icons of the cups or countdown marker, no longer freezes when compilling shaders on game versions below NCE update to bring more compatibility with CTGP MOD)
- Naruto's Ninja Storm series (no longer blue texturing on Naruto or main character on the screeen)
- Pokemon Scarlet/ Violet (no longer missing characters or npc)
- And many more games that are missing from our scope of testing.
But not only that, but also improved GPU - CPU synchornization for better performance (by 15% compared to previous builds) more stable shader compilations and removed unneeded emulation paths; making games as Breath Of The Wild less stuttery than it was used to; along some cleaning from dead code, allowing mobile devices to go brrrr, always long live to the Snapdragon Master Race, it's needed to explain that devices from A7XX need to use the driver 819.2, meanwhile A8XX can use 842.9 to obtain all the benefits from VK 1.4 allowed in Eden.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3334
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: Maufeat <sahyno1996@gmail.com>
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: CamilleLaVey <camillelavey99@gmail.com>
Co-committed-by: CamilleLaVey <camillelavey99@gmail.com>
first of all it makes UI a tad bit more simple and benefits the end user
second, it allows to add new backends a bit more easily
and also, there was a shader backend option in android... we don't use opengl in android
so may as well save ourselves the trouble, aye

Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3313
Reviewed-by: Maufeat <sahyno1996@gmail.com>
Reviewed-by: DraVee <dravee@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
this revert is intended to bring back the original behavior (throw exception) for untracked bindless textures.
reason: this was a mock intended to extend ender magnolia's mean time to crash.
by reverting it we will be able to identify more games which rely on bindless textures, and avoid to obfuscate visual issues investigation on them.
this is super safe to merge. if any game that was benefitting of it gets visual regression, we will know immediately via the throw message in logs.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3331
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: Maufeat <sahyno1996@gmail.com>
Co-authored-by: xbzk <xbzk@eden-emu.dev>
Co-committed-by: xbzk <xbzk@eden-emu.dev>
As stated by @PavelBARABANOV when joypad overlay was being hidden via quick toggle or autohides (time/input), touchpad was being disabled along (hiding via original show/hide checkbox was working ok).
This PR fixes this by centering all joypad overlay toggling paths, and using correct method to not disable touchpad.
Additionally, mirror state variable was removed for better sync, and also touching screen will only unhide if hiding was triggered via autohides.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3315
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: DraVee <dravee@eden-emu.dev>
Co-authored-by: xbzk <xbzk@eden-emu.dev>
Co-committed-by: xbzk <xbzk@eden-emu.dev>
This pulls eden releases changelog & text from our github releases.
We don't store the msgpack file but rather generate them in-memory for the News Applet.
Uses cache folder. Files generated are:
- cache/news/github_releases.json
- cache/news/eden_logo.jpg
- cache/news/news_read
Additional changes:
- Proper TLV returning for online web applet, to open external URL
- Add applet type `LHub` to properly close, as it also uses TLV return
- qlaunch app sorting, adds another cached .json to track last launched app timestamps and sort them accordingly
Co-authored-by: crueter <crueter@eden-emu.dev>
Co-authored-by: DraVee <dravee@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3308
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Co-authored-by: Maufeat <sahyno1996@gmail.com>
Co-committed-by: Maufeat <sahyno1996@gmail.com>
for core stuff:
just remove unique ptrs that dont need any pointer stability at all (afterall its an allocation within an allocation so yeah)
for fibers:
Main reasoning behind this is because virtualBuffer<> is stupidly fucking expensive and it also clutters my fstat view
ALSO mmap is a syscall, syscalls are bad for performance or whatever
ALSO std::vector<> is better suited for handling this kind of "fixed size thing where its like big but not THAT big" (512 KiB isn't going to kill your memory usage for each fiber...)
for core.cpp stuff
- inlines stuff into std::optional<> as opposed to std::unique_ptr<> (because yknow, we are making the Impl from an unique_ptr, allocating within an allocation is unnecessary)
- reorganizes the structures a bit so padding doesnt screw us up (it's not perfect but eh saves a measly 44 bytes)
- removes unused/dead code
- uses std::vector<> instead of std::deque<>
no perf impact expected, maybe some initialisation boost but very minimal impact nonethless
lto gets rid of most calls anyways - the heavy issue is with shared_ptr and the cache coherency from the atomics... but i clumped them together because well, they kinda do not suffer from cache coherency - hopefully not a mistake
this balloons the size of Impl to about 1.67 MB - which is fine because we throw it in the stack anyways
REST OF INTERFACES: most of them ballooned in size as well, but overhead is ok since its an allocation within an alloc, no stack is used (when it comes to storing these i mean)
Signed-off-by: lizzie lizzie@eden-emu.dev
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3306
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Support for bundled Qt, not through aqtinstall but rather my CI. Multimedia is
implemented too, works on both Windows and Linux, though we don't
actually use it so it doesn't really matter. Contains Declarative and all that so the Quick frontend will work once it becomes a thing.
Some options have changed, notably w.r.t LTO and faster
linker, which are now handled directly in the modules.
CPMUtil also has support for custom dirs (`PackageName_CUSTOM_DIR`) now. Probably most useful for adding external fragment shaders and whatnot.
Signed-off-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3289
- [Added] a new compute shader to handle block-linear unswizzling on the GPU, reducing CPU overhead during texture uploads
- [Implemented] BlockLinearUnswizzle3DPass to take advantage of the new compute shader, unimplemented for OpenGL
- [Implemented] texture streaming and queue system for large sparse textures to prevent hitches
- [Implemented] aggressive garbage collection system to eject large sparse textures to save on memory (Unused)
- [Added] user settings to adjust the streaming unswizzle system for low-end machines
- [Improved] slightly the ASTC GPU decoding system
Co-authored-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Co-authored-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: DraVee <dravee@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3246
Reviewed-by: Maufeat <sahyno1996@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: Forrest Keller <forrestmarkx@outlook.com>
Co-committed-by: Forrest Keller <forrestmarkx@outlook.com>
Raises the size of each page entry to 32 bytes, however, it merges them into a single structure
THEORETICALLY this is better since the access pattern observed corresponds to the program wanting backing_addr/pointers/blocks immediately after one another.
This may improve performance at the cost of some extra memory.
Another implementation would be to structure only backing_addr/blocks within the same virtual buffer.
Alas spamming virtual buffers is evil since each of them is a cache trasher (imagine jumping from wildly different block to wildly different block immediately).
Signed-off-by: lizzie lizzie@eden-emu.dev
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3215
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
This PR brings a feature that has been needed for some time in the Android Switch emulation community: environment variables for Turnip/Freedreno drivers. These are available in PC emulators and can help fix some problems, especially the TU_DEBUG function, which can be set to gmem (thus allowing Adreno 710/720 users to run Turnip correctly), and noubwc, which fixes some problems for OneUI users.
This could also help us debug Turnip in a "better way" in the future.
Attached is a screenshot of a user, Ivan albio, using the gmem function on Adreno 710.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3205
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: DraVee <dravee@eden-emu.dev>
Co-authored-by: MrPurple666 <antoniosacramento666usa@gmail.com>
Co-committed-by: MrPurple666 <antoniosacramento666usa@gmail.com>
This PR rewrites the DynamicState, ExtendedDynamicState and VertexInputDynamicState logic:
- Adds proper handling on how features should be loaded based on driver available features for ExtendedDynamicState/ VertexInputDynamicState.
- Fixes some old regressions with emulated formats for Android.
- Adds better formatting for tiling format features.
- Adds better formatting for format features.
- Adds NonWritable buffers handling for Spir-v.
- Updates Maintenance features calling.
- Adds new features: Multidraw, Robustness2, Image Robustness.
- Removes dead code/ duplicated on Vulkan device related to ExtendedDynamicState handling.
- Adjusts and conditions with better handling for some features callings: SwapchainMaintenance1, ConditionalRendering, ShaderExtencilExport, CustomBorderColor, TransformFeedback, VertexInputDynamicState.
- Removes some older feature ban logic.
- Adds hardware resolve path for MSAA Image Blits on Nvidia cards.
- Adds flat decorations for input interfaces on Spir-v.
- Reduces flushwork within drawcalls.
- Clamps render limits on out-of-area for rasterizer.
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-authored-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Co-authored-by: DraVee <dravee@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3074
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Co-authored-by: CamilleLaVey <camillelavey99@gmail.com>
Co-committed-by: CamilleLaVey <camillelavey99@gmail.com>
This fixes an issue where game cards can stack focus highlights by touching and sliding in Grid/List views. Running and exiting the game by touch leaves a sticky focus that is not cleared. It is again possible to stack focus highlights that way.
The first commit fixes the bug, the second refactors and simplifies the state management in GradientBorderCardView.
WIP for now, until I thoroughly test it.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3242
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: DraVee <dravee@eden-emu.dev>
Co-authored-by: Mike22 <misakupka@gmail.com>
Co-committed-by: Mike22 <misakupka@gmail.com>
This PR fixes the real culprit for the crash mentioned in #3212 (which has its value coz it fixes other layout's minor navigation issues)
The reason was not the carousel, but the joypad overlay autohide instead.
To reproduce the crash, just enable autohide, launch the game, and leave before overlay gets hidden. eden should crash in few seconds.
The fix was adding barriers to ensure fragment is currently attached to its activity and the binding is not null.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3243
Reviewed-by: Maufeat <sahyno1996@gmail.com>
Reviewed-by: DraVee <dravee@eden-emu.dev>
Co-authored-by: xbzk <xbzk@eden-emu.dev>
Co-committed-by: xbzk <xbzk@eden-emu.dev>
- use std::optional instead of std::unique_ptr for the Antialias (FXAA, etc) passes to avoid the extra deref
- use a pattern for deferencing the IR pointer chasing loop as suggested on the intel optimization manual
- this also removes std::vector<> overhead by using boost::container::small_vector<> (not a silver bullet but in the case of this function reduces access times)
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Co-authored-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/2565
Reviewed-by: Maufeat <sahyno1996@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
This reworks the logic to improve performance in many games that heavily rely on DMA. It can help all platforms, but on desktop the performance boost can be noticeable, especially on dedicated GPUs. The option Sync Memory Operations must be enabled.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3179
Co-authored-by: MaranBr <maranbr@outlook.com>
Co-committed-by: MaranBr <maranbr@outlook.com>
- Fast GPU now defaults to 256, removed 128 since it's useless.
- Completely reorganized graphics and CPU settings on both platforms.
Also got rid of Eden's Veil
- Merged some "use ..." settings that weren't really necessary.
- Changed ExtendedDynamicState to be a combo box
Signed-off-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3233
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: DraVee <dravee@eden-emu.dev>
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Yxzx presumes this:
// The NVN driver buffer (index 0) is known to pack the SSBO address followed by its size.
But in MCI i`ve discovered that there are no sizes, both registers are GPU addresses (hence the 2.8gb allocation, it was an address actually)
Method could be much simpler but for safety i`ve routed both old and new worlds.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3206
Reviewed-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Co-authored-by: xbzk <xbzk@eden-emu.dev>
Co-committed-by: xbzk <xbzk@eden-emu.dev>
Hand rolling memcpy like this is always frowned upon because the compiler has more insight on whats going on (plus the code resolves to a worse version of itself on assembly). This removes some branches that are just straight up redundant. May save stuff especially for systems without fastmem enabled.
Signed-off-by: lizzie <lizzie@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/2639
Reviewed-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
This should fix the issue with, for example, ToTK running at 60 FPS when overlay applet is running.
This also should always run the overlay as actual overlay and not in the back.
Stubs RequestListSummaryOverlayNotifications in friends
Syncs Language of the Emulator, when setting language, this is used in Starter Applet
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3123
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Co-authored-by: Maufeat <sahyno1996@gmail.com>
Co-committed-by: Maufeat <sahyno1996@gmail.com>
Setting `HIDE_OVERLAY_ON_CONTROLLER_INPUT` in *Advanced settings → Input Overlay*
**Behavior:**
- First controller input -> hides overlay
- Controller disconnect → shows overlay again
- Subsequent controller inputs → ignored (already hidden, so no retrigger needed)
- Touch screen → does **not** show overlay (so you can use a controller and touchscreen to interact with games)
- Sidebar "Show/Hide controller" button → still works as master toggle
**State reset: The "first input" detection resets when:**
1. Controller disconnects
2. Overlay is shown via sidebar button
3. Controller reconnects
**Interaction with other settings:**
- Requires `SHOW_INPUT_OVERLAY` to be enabled (basicaly a master switch)
- Independent from `ENABLE_INPUT_OVERLAY_AUTO_HIDE` (timer-based hide, was already implemented)
- When both are enabled, touch-to-show is disabled (controller-hide takes precedence)
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3127
Reviewed-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Reviewed-by: Maufeat <sahyno1996@gmail.com>
Co-authored-by: Producdevity <y.gherbi.dev@gmail.com>
Co-committed-by: Producdevity <y.gherbi.dev@gmail.com>
## Overview
Adds guides relating to the multiplayer components of Eden, which will serve as a quick reference for users to get started.
### Guides added
- Multiplayer FAQ
- Hosting a Multiplayer Room
- Joining a Multiplayer Room
- Access Your Multiplayer Room Externally
- Finding Multiplayer Server Info
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3167
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: Evilperson1337 <Evilperson1337@proton.me>
Co-committed-by: Evilperson1337 <Evilperson1337@proton.me>
The GPU Accuracy level is now divided into Performance, Balanced and Accurate.
1. Performance prioritizes speed at all costs. It's faster, but it can be unstable and may have some bugs (which is expected).
2. Balanced maintains excellent performance and is safer against bugs and shader corruption.
3. Accurate is the most precise and the most expensive in terms of hardware. Only a few games still need this level to work properly.
The Release Early Fences toggle has also been removed by @PavelBARABANOV, as it's not needed anymore.
Co-authored-by: PavelBARABANOV <pavelbarabanov94@gmail.com>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3129
Reviewed-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Reviewed-by: Maufeat <sahyno1996@gmail.com>
Co-authored-by: MaranBr <maranbr@outlook.com>
Co-committed-by: MaranBr <maranbr@outlook.com>
From my tests this decreases JIT latency twofold, may be placebo.
saving reg_alloc while having it readily available is certainly a very interesting choice... afterall saving it onto %rdi is way more cheap isn't it? :)
Please test any performance rgressions, I got +20 FPS on Rain World (unlocked) off this change alone
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3150
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
## Overview
Adding Markdown, simplified versions of the docs hosted on [External Site](https://evilperson1337.notion.site/Eden-Guides-2b057c2edaf680149310edee3024565d).
### Guides Added
- Adding Games into Steam with Steam ROM Manager and Eden
- Adding Eden as an emulator into Steam ROM Manager
- How to modify the Date/Time in Eden
- How to configure controller profiles
- How to set controller profiles per game
- Eden fails to launch but does not leave a log (Windows)
- How to configure motion/gyro controls on steam deck
- How to access logs
- How to import saves into Eden (windows)
- How to install Atmosphere mods
- How to install Updates/DLC
- Quick Start Guide (Steam Deck)
- Quick Start Guide (Windows)
- Allowing Eden to Run on MacOS (MacOS)
- Sync Game Saves with Syncthing (General)
- Sync Game Saves with Syncthing (Windows)
- Sync Game Saves with Syncthing (Linux)
- Using Amiibo with Eden (Windows)
- Using Cheats with Eden (Windows)
- Using Cheats with Eden (Android)
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3139
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: Evilperson1337 <Evilperson1337@proton.me>
Co-committed-by: Evilperson1337 <Evilperson1337@proton.me>
Successor to that old MoltenVK PR. Does a lot of cleanups within root CMakeLists.txt, hands over MoltenVK and VulkanValidationLayers to CPMUtil, and separates out common operations into my modules.
Hopefully reduces the monstrosity that is root CMakeLists.txt. Please test:
- builds on all platforms
- VulkanValidationLayers
Signed-off-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3126
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Updates all of our bundled CI deps to support android x86_64, adds a
build flavor thereof (`chromeOS`), and also adds sirit mingw support.
The new FFmpeg package is built in a much better way that actually makes
it identically built to the other CI packages, meaning we now have real
8.0.0 support, no need for libvpx/cpu_features/all that other crap.
PLUS, we can now statically link it! Hooray! It's also built with
MediaCodec support so in the future we can work on that.
Rewrote the android build script too, plus added a copyFlavorTypeOutput
target that assembles and copies the APK. The code behind it sucks
because I'm not great with Gradle but hey, it works.
Testers: please test everything related to video decoding. VP9 and h264, games that normally suck with their prerendered stuff, make sure I didn't nuke it to oblivion, etc.
Signed-off-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3086
Reviewed-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
So when libc starts it has to start at an entry point located into crt0, now most OSes will do "enough" setup to allow mprotect() and mmap() to be called in static ctors (remember they're called BEFORE main)
By some stupid miracle, NetBSD doesn't; this means that using those functions on NetBSD will result in spurious results
The reason why is still unknown to me, but this is also combined with the fact that allocating a big chunk of memory for the JIT will make NetBSD refuse to mprotect()/mmap() it in low memory situations (even when space is available); so I take the same approach as with solaris
Also I now make it so fastmem handlers are NOT registered for OSes that disabled fastmem, this is because they pollute sigsegv and makes debugging stupidier
Signed-off-by: lizzie lizzie@eden-emu.dev
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3092
Reviewed-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Adds fully functional overlay display.
- Enable Overlay Applet via "View" -> "Enable Overlay Display Applet"
- Open the overlay by pressing the home button for over 1s
- Can adjust volume
- Can toggle airplane mode (if on WiFi, maybe if overlay is enabled pretend to be on WiFi?)
- Future TODO(?): Adjust Brightness implementation for host system
- Inputs are properly registered. e.g. if overlay open, application does not register inputs.
You can control volume and airplane mode outside of the emulator window
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3080
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: crueter <crueter@eden-emu.dev>
Co-authored-by: Maufeat <sahyno1996@gmail.com>
Co-committed-by: Maufeat <sahyno1996@gmail.com>
**Aims to dismiss the needing of developers to wait for someone to provide new toggles only to test temporary stuff**
This is a classic debug knob set for development use.
Developers will be able to call Settings::getDebugKnobAt(0 to 15) to pick one of the 16 bits of that setting, allowing users to easily enable or disable multiple features in testing builds, by entering values instructed by the developers.
Co-authored-by: Allison Cunha <allisonbzk@gmail.com>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3076
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Co-authored-by: xbzk <xbzk@eden-emu.dev>
Co-committed-by: xbzk <xbzk@eden-emu.dev>
This issue has like 15 different causes, and I'm surprised it took this
long to pop up.
1. LoadString had a hack *specific* to the AudioEngine enum. Why?
Solving this was easy, just use the explicit type ctor. -_-
2. The LoadString hack was abused in configure_audio.cpp to get around
the canonicalization infrastructure that was explicitly put in to
make this exact operation easier. Why?
3. ToString was also broken because of LoadString's garbage output.
Technically it might work now, but it's better to just use the
canonicalization infrastructure that was made specifically for this
purpose.
Also did a few tiny optimizations in config/settings cuz wynaut.
Signed-off-by: crueter <crueter@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3083
Reviewed-by: Caio Oliveira <caiooliveirafarias0@gmail.com>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
NOTES:
regs.window_origin.flip_y MUST flip the y coordinate of any given FragCoord, we don't emulate this, this is the root cause of the error, but I'll just revert for now since it's easier
DON'T MERGE unless it's near 0.0.4 and I (or someone else) hasn't tackled this yet properly
This reverts commit 17fe74ef11.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3075
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Co-authored-by: lizzie <lizzie@eden-emu.dev>
Co-committed-by: lizzie <lizzie@eden-emu.dev>
Fixes AMD + Windows because it forces barriers to include the fragment test and color output stages explicitly, ensuring that all render pass writes are visible before later commands. Without it, AMD’s driver sometimes skipped synchronization, causing broken rendering in Final Fantasy Tactics.
PR #3069 also fixes this regression by reverting vk_scheduler.cpp in PR #180.
This PR fixes PR #180 and may be the better solution.
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/3071
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: Lizzie <lizzie@eden-emu.dev>
Co-authored-by: John <john@eden-emu.dev>
Co-committed-by: John <john@eden-emu.dev>
about:If you are experiencing an issue with yuzu, and you need tech support, or if you have a general question, try asking in the official yuzu Discord linked here. Piracy is not allowed.
- name:Community forums
url:https://community.citra-emu.org
about:This is an alternative place for tech support, however helpers there are not as active.
From cc15da16e533b2a801934eab2dfeaf3c3949a1dc Mon Sep 17 00:00:00 2001
From: crueter <crueter@eden-emu.dev>
Date: Mon, 8 Sep 2025 12:28:55 -0400
Subject: [PATCH] [cmake] disable NEON runtime check on clang-cl
When enabling runtime NEON checking for clang-cl, the linker would error out with `undefined symbol: __emit`, since clang doesn't actually implement this instruction. Therefore it makes sense to disable the runtime check by default on this platform, until either this is fixed or a clang-cl compatible intrinsic check is added (I don't have enough knowledge of MSVC to do this)
+# master project detection--useful for FetchContent/submodule inclusion
+set(master_project OFF)
+set(subproject ON)
+
+if (NOT DEFINED PROJECT_NAME)
+ set(master_project ON)
+ set(subproject OFF)
+endif()
+
project(spirv-tools)
# Avoid a bug in CMake 3.22.1. By default it will set -std=c++11 for
@@ -135,46 +144,49 @@ if (DEFINED SPIRV_TOOLS_EXTRA_DEFINITIONS)
add_definitions(${SPIRV_TOOLS_EXTRA_DEFINITIONS})
endif()
-# Library build setting definitions:
-#
-# * SPIRV_TOOLS_BUILD_STATIC - ON or OFF - Defaults to ON.
-# If enabled the following targets will be created:
-# ${SPIRV_TOOLS}-static - STATIC library.
-# Has full public symbol visibility.
-# ${SPIRV_TOOLS}-shared - SHARED library.
-# Has default-hidden symbol visibility.
-# ${SPIRV_TOOLS} - will alias to one of above, based on BUILD_SHARED_LIBS.
-# If disabled the following targets will be created:
-# ${SPIRV_TOOLS} - either STATIC or SHARED based on SPIRV_TOOLS_LIBRARY_TYPE.
-# Has full public symbol visibility.
-# ${SPIRV_TOOLS}-shared - SHARED library.
-# Has default-hidden symbol visibility.
-#
-# * SPIRV_TOOLS_LIBRARY_TYPE - SHARED or STATIC.
-# Specifies the library type used for building SPIRV-Tools libraries.
-# Defaults to SHARED when BUILD_SHARED_LIBS=1, otherwise STATIC.
-#
-# * SPIRV_TOOLS_FULL_VISIBILITY - "${SPIRV_TOOLS}-static" or "${SPIRV_TOOLS}"
-# Evaluates to the SPIRV_TOOLS target library name that has no hidden symbols.
-# This is used by internal targets for accessing symbols that are non-public.
-# Note this target provides no API stability guarantees.
-#
-# Ideally, all of these will go away - see https://github.com/KhronosGroup/SPIRV-Tools/issues/3909.
-option(ENABLE_EXCEPTIONS_ON_MSVC "Build SPIRV-TOOLS with c++ exceptions enabled in MSVC" ON)
-option(SPIRV_TOOLS_BUILD_STATIC "Build ${SPIRV_TOOLS}-static target. ${SPIRV_TOOLS} will alias to ${SPIRV_TOOLS}-static or ${SPIRV_TOOLS}-shared based on BUILD_SHARED_LIBS" ON)
<h4 align="center"><b>Eden</b> is an open-source Nintendo Switch emulator, forked from the Yuzu emulator — started by former Citron developer Camille LaVey and the Eden team.
It is written in C++ with portability in mind, and we actively maintain builds for Windows, Linux and Android.
<h4 align="center"><b>Eden</b> is a free and opensource (FOSS) Switch 1 emulator, derived from Yuzu and Sudachi - started by developer Camille LaVey.
It's written in C++ with portability in mind, with builds for Windows, Linux, macOS, Android, FreeBSD and more.
@@ -52,10 +52,10 @@ Check out our [website](https://eden-emu.dev) for the latest news on exciting fe
## Development
Most of the development happens on our Git server. It is also where [our central repository](https://git.eden-emu.dev/eden-emu/eden) is hosted. For development discussions, please join us on [Discord](https://discord.gg/kXAmGCXBGD) or [Revolt](https://rvlt.gg/qKgFEAbH).
You can also follow us on [X (Twitter)](https://x.com/edenemuofficial) for updates and announcements.
Most of the development happens on our Git server. It is also where [our central repository](https://git.eden-emu.dev/eden-emu/eden) is hosted. For development discussions, please join us on [Discord](https://discord.gg/HstXbPch7X) or [Stoat](https://stt.gg/qKgFEAbH).
You can also follow us on [X (Twitter)](https://nitter.poast.org/edenemuofficial) for updates and announcements.
If you would like to contribute, we are open to new developers and pull requests. Please ensure that your work is of a high standard and properly documented. You can also contact any of the developers on Discord or Revolt to learn more about the current state of the emulator.
If you would like to contribute, we are open to new developers and pull requests. Please ensure that your work is of a high standard and properly documented. You can also contact any of the developers on Discord or Stoat to learn more about the current state of the emulator.
See the [sign-up instructions](docs/SIGNUP.md) for information on registration.
@@ -82,7 +82,7 @@ Any donations received will go towards things such as:
* Additional hardware (e.g. GPUs as needed to improve rendering support, other peripherals to add support for, etc.)
* CI Infrastructure
If you would prefer to support us in a different way, please join our [Discord](https://discord.gg/edenemu) and talk to Camille or any of our other developers.
If you would prefer to support us in a different way, please join our [Discord](https://discord.gg/HstXbPch7X) and talk to Camille or any of our other developers.
style="fill:url(#linearGradient12);fill-opacity:1;stroke:url(#linearGradient25);stroke-width:13.7948;stroke-opacity:0.566238;paint-order:stroke fill markers"
id="path8"
cx="256.76401"
cy="253.05354"
r="244.25046"/>
<path
id="rect1-3"
style="fill:#ffffff;fill-opacity:0.5;stroke:none;stroke-width:17.2007;stroke-dasharray:none;stroke-opacity:1;paint-order:stroke fill markers"
d="M 55.701477,114.37179 A 244.25044,244.25044 0 0 0 44.86493,131.57252 h 423.79699 a 244.25044,244.25044 0 0 0 -10.83655,-17.20073 z"/>
<path
style="fill:#ff2bd5;fill-opacity:1;stroke-width:14.3776;stroke-opacity:0.415999;paint-order:stroke fill markers"
d="M 219.6651,499.13816 209.32675,443.18948"
id="path9"/>
<path
id="rect1-2"
style="fill:#ffffff;fill-opacity:0.5;stroke:none;stroke-width:17.2007;stroke-dasharray:none;stroke-opacity:1;paint-order:stroke fill markers"
d="m 58.271509,395.38922 a 244.25044,244.25044 0 0 0 13.54096,17.20073 H 441.71437 a 244.25044,244.25044 0 0 0 13.54097,-17.20073 z"/>
<path
id="rect1-2-5"
style="fill:#ffffff;fill-opacity:0.5;stroke:none;stroke-width:17.2007;stroke-dasharray:none;stroke-opacity:1;paint-order:stroke fill markers"
d="m 24.604933,328.95264 a 244.25044,244.25044 0 0 0 6.347389,17.20074 H 482.57453 a 244.25044,244.25044 0 0 0 6.34738,-17.20074 z"/>
<path
id="rect1"
style="fill:#ffffff;fill-opacity:0.5;stroke:none;stroke-width:17.2007;stroke-dasharray:none;stroke-opacity:1;paint-order:stroke fill markers"
d="m 12.586253,259.07464 a 244.25044,244.25044 0 0 0 1.033053,17.20075 H 499.90754 a 244.25044,244.25044 0 0 0 1.03305,-17.20075 z"/>
<path
id="path15"
style="fill:url(#linearGradient2);fill-opacity:1;stroke:#ffffff;stroke-width:17.2007;stroke-linejoin:round;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1;paint-order:stroke fill markers"
style="fill:url(#linearGradient12);fill-opacity:1;stroke:#ffef91;stroke-width:12.8318;stroke-opacity:1;paint-order:stroke fill markers"
id="path8"
cx="256.2999"
cy="257.2999"
r="227.2"/><path
id="path15"
style="fill:url(#linearGradient2);fill-opacity:1;stroke:#ffef91;stroke-width:15.499;stroke-linejoin:round;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1;paint-order:stroke fill markers"
- MSVC: `Ninja` (preferred) or `Visual Studio 17 2022`
- macOS: `Ninja` (preferred) or `Xcode`
- Others: `Ninja` (preferred) or `UNIX Makefiles`
Available `<BUILD_TYPE>`:
-`Release` (default)
-`RelWithDebInfo` (debug symbols--compiled executable will be large)
-`Debug` (if you are using a debugger and annoyed with stuff getting optimized out)
Caveat for Debug Builds:
- If you're building with CCache, you will need to add the environment variable `CL` with the `/FS` flag ([Reference](https://learn.microsoft.com/pt-br/cpp/build/reference/fs-force-synchronous-pdb-writes?view=msvc-170))
Also see the [Options](Options.md) page for additional CMake options.
Also see the root CMakeLists.txt for more build options. Usually the default will provide the best experience, however.
- *Carefully* re-read the [dependencies guide](Deps.md)
- Clear the CPM cache (`.cache/cpm`) and CMake cache (`<build directory>/CMakeCache.txt`)
- Evaluate the error and find any related settings
- See the [CPM docs](CPM.md) to see if you may need to forcefully bundle any packages
Otherwise, feel free to ask for help in Revolt or Discord.
Otherwise, feel free to ask for help in Stoat or Discord.
## Caveats
@@ -153,20 +161,16 @@ Simply hit Ctrl+B, or the "hammer" icon in the bottom left. To run, hit the "pla
If you are using the `UNIX Makefiles` or `Visual Studio 17 2022` as `<GENERATOR>`, you should also add `--parallel` for faster build times.
```
```sh
cmake --build build
```
Your compiled executable will be in:
-`build/bin/eden.exe` for Windows,
-`build/bin/eden.app/Contents/MacOS/eden` for macOS,
- and `build/bin/eden` for others.
## Scripts
Some platforms have convenience scripts provided for building.
- **[Linux](scripts/Linux.md)**
- **[Windows](scripts/Windows.md)**
macOS scripts will come soon.
Take a look at our [CI scripts](https://github.com/Eden-CI/Workflow). You can use `.ci/common/configure.sh` on any POSIX-compliant shell, but you are heavily encouraged to instead write your own based. It's not really that hard, provided you can read CMake.
* Defaults to `github.com`. Do not include the protocol, as HTTPS is enforced.
- Defaults to `github.com`. Do not include the protocol, as HTTPS is enforced.
-`TAG`: The tag to fetch, if applicable.
-`ARTIFACT`: The name of the artifact, if applicable.
-`SHA`: Commit sha to fetch, if applicable.
@@ -26,23 +26,23 @@
The following configurations are supported, in descending order of precedence:
-`URL`: Bare URL download, useful for custom artifacts
* If this is set, `GIT_URL` or `REPO` should be set to allow the dependency viewer to link to the project's Git repository.
* If this is NOT set, `REPO` must be defined.
- If this is set, `GIT_URL` or `REPO` should be set to allow the dependency viewer to link to the project's Git repository.
- If this is NOT set, `REPO` must be defined.
-`REPO + TAG + ARTIFACT`: GitHub release artifact
* The final download URL will be `https://github.com/${REPO}/releases/download/${TAG}/${ARTIFACT}`
* Useful for prebuilt libraries and prefetched archives
- The final download URL will be `https://github.com/${REPO}/releases/download/${TAG}/${ARTIFACT}`
- Useful for prebuilt libraries and prefetched archives
-`REPO + TAG`: GitHub tag archive
* The final download URL will be `https://github.com/${REPO}/archive/refs/tags/${TAG}.tar.gz`
* Useful for pinning to a specific tag, better for build identification
- The final download URL will be `https://github.com/${REPO}/archive/refs/tags/${TAG}.tar.gz`
- Useful for pinning to a specific tag, better for build identification
-`REPO + SHA`: GitHub commit archive
* The final download URL will be `https://github.com/${REPO}/archive/${SHA}.zip`
* Useful for pinning to a specific commit
- The final download URL will be `https://github.com/${REPO}/archive/${SHA}.zip`
- Useful for pinning to a specific commit
-`REPO + BRANCH`: GitHub branch archive
* The final download URL will be `https://github.com/${REPO}/archive/refs/heads/${BRANCH}.zip`
* Generally not recommended unless the branch is frozen
- The final download URL will be `https://github.com/${REPO}/archive/refs/heads/${BRANCH}.zip`
- Generally not recommended unless the branch is frozen
-`REPO`: GitHub master archive
* The final download URL will be `https://github.com/${REPO}/archive/refs/heads/master.zip`
* Generally not recommended unless the project is dead
- The final download URL will be `https://github.com/${REPO}/archive/refs/heads/master.zip`
- Generally not recommended unless the project is dead
## Hashing
@@ -54,20 +54,20 @@ Hashing strategies, descending order of precedence:
-`HASH`: Bare hash verification, useful for static downloads e.g. commit archives
-`HASH_SUFFIX`: Download the hash as `${DOWNLOAD_URL}.${HASH_SUFFIX}`
* The downloaded hash *must* match the hash algorithm and contain nothing but the hash; no filenames or extra content.
- The downloaded hash *must* match the hash algorithm and contain nothing but the hash; no filenames or extra content.
-`HASH_URL`: Download the hash from a separate URL
## Other Options
-`KEY`: Custom cache key to use (stored as `.cache/cpm/${packagename_lower}/${key}`)
* Default is based on, in descending order of precedence:
- Default is based on, in descending order of precedence:
- First 4 characters of the sha
-`GIT_VERSION`
- Tag
-`VERSION`
- Otherwise, CPM defaults will be used. This is not recommended as it doesn't produce reproducible caches
-`DOWNLOAD_ONLY`: Whether or not to configure the downloaded package via CMake
* Useful to turn `OFF` if the project doesn't use CMake
- Useful to turn `OFF` if the project doesn't use CMake
-`SOURCE_SUBDIR`: Subdirectory of the project containing a CMakeLists.txt file
-`FIND_PACKAGE_ARGUMENTS`: Arguments to pass to the `find_package` call
-`BUNDLED_PACKAGE`: Set to `ON` to default to the bundled package
@@ -80,12 +80,14 @@ Hashing strategies, descending order of precedence:
For each added package, users may additionally force usage of the system/bundled package.
-`${package}_DIR`: Path to a separately-downloaded copy of the package. Note that versioning is not checked!
-`${package}_FORCE_SYSTEM`: Require the package to be installed on the system
-`${package}_FORCE_BUNDLED`: Force the package to be fetched and use the bundled version
## System/Bundled Packages
Descending order of precedence:
- If `${package}_FORCE_SYSTEM` is true, requires the package to be on the system
- If `${package}_FORCE_BUNDLED` is true, forcefully uses the bundled package
- If `CPMUTIL_FORCE_SYSTEM` is true, requires the package to be on the system
@@ -101,8 +103,8 @@ URLs:
-`GIT_URL`
-`REPO` as a Git repository
* You may optionally specify `GIT_HOST` to use a custom host, e.g. `GIT_HOST git.crueter.xyz`. Note that the git host MUST be GitHub-like in its artifact/archive downloads, e.g. Forgejo
* If `GIT_HOST` is unspecified, defaults to `github.com`
- You may optionally specify `GIT_HOST` to use a custom host, e.g. `GIT_HOST git.crueter.xyz`. Note that the git host MUST be GitHub-like in its artifact/archive downloads, e.g. Forgejo
- If `GIT_HOST` is unspecified, defaults to `github.com`
-`URL`
Versions (bundled):
@@ -113,4 +115,4 @@ Versions (bundled):
-`TAG`
- "unknown"
If the package is a system package, AddPackage will attempt to determine the package version and append `(system)` to the identifier. Otherwise, it will be marked as `unknown (system)`
If the package is a system package, AddPackage will attempt to determine the package version and append `(system)` to the identifier. Otherwise, it will be marked as `unknown (system)`
Simply call `AddQt(<Qt Version>)` before any Qt `find_package` calls and everything will be set up for you. On Linux, the bundled Qt library is built as a shared library, and provided you have OpenSSL and X11, everything should just work.
On Windows, MinGW, and MacOS, Qt is bundled as a static library. No further action is needed, as the provided libraries automatically integrate the Windows/Cocoa plugins, alongside the corresponding Multimedia and Network plugins.
## Modules
The following modules are bundled into these Qt builds:
- Base (Gui, Core, Widgets, Network)
- Multimedia
- Declarative (Quick, QML)
- Linux: Wayland client
Each platform has the corresponding QPA built in and set as the default as well. This means you don't need to add `Q_IMPORT_PLUGIN`!
## Example
See an example in the [`tests/qt`](https://git.crueter.xyz/CMake/CPMUtil/src/branch/master/tests/qt/CMakeLists.txt) directory.
## Versions
The following versions have available builds:
- 6.9.3
See [`crueter-ci/Qt`](https://github.com/crueter-ci/Qt) for an updated list at any time.
If you are packaging a project that uses CPMUtil, read this!
### Network Sandbox
For sandboxed environments (e.g. Gentoo, nixOS) you must install all dependencies to the system beforehand and set `-DCPMUTIL_FORCE_SYSTEM=ON`. If a dependency is missing, get creating!
Alternatively, if CPMUtil pulls in a package that has no suitable way to install or use a system version, download it separately and pass `-DPackageName_DIR=/path/to/downloaded/dir` (e.g. shaders)
### Unsandboxed
For others (AUR, MPR, etc). CPMUtil will handle everything for you, including if some of the project's dependencies are missing from your distribution's repositories. That is pretty much half the reason I created this behemoth, after all.
- [Windows 7, Windows 8 and Windows 8.1](#windows-7-windows-8-and-windows-81)
- [Windows Vista and below](#windows-vista-and-below)
- [Windows on ARM](#windows-on-arm)
<!-- /TOC -->
## Arch Linux
@@ -49,6 +52,7 @@ export PATH="$PATH:$PWD"
```
Default MESA is a bit outdated, the following environment variables should be set for a smoother experience:
```sh
exportMESA_GL_VERSION_OVERRIDE=4.6
exportMESA_GLSL_VERSION_OVERRIDE=460
@@ -77,6 +81,8 @@ For this reason this patch is NOT applied to default on all platforms (for obvio
Still will not run flawlessly until `mesa-24` is available. Modify CMakeCache.txt with the `.so` of libGL and libGLESv2 by doing the incredibly difficult task of copy pasting them (`cp /boot/system/lib/libGL.so .`)
If you have `quazip1_qt6_devel`, uninstall it. It may call `Core5Compat` on CMake which is wrongly packaged.
## OpenBSD
After configuration, you may need to modify `externals/ffmpeg/CMakeFiles/ffmpeg-build/build.make` to use `-j$(nproc)` instead of just `-j`.
@@ -91,19 +97,33 @@ Eden is not currently available as a port on FreeBSD, though it is in the works.
The available OpenSSL port (3.0.17) is out-of-date, and using a bundled static library instead is recommended; to do so, add `-DYUZU_USE_BUNDLED_OPENSSL=ON` to your CMake configure command.
Gamepad/controllers may not work on 15.0, this is due to an outdated SDL not responding well to the new `usbhid(2)` driver. To workaround this simply disable `usbhid(2)` (add the following to `/boot/loader.conf`):
```sh
hw.usb.usbhid.enable="0"
```
## NetBSD
Install `pkgin` if not already `pkg_add pkgin`, see also the general [pkgsrc guide](https://www.netbsd.org/docs/pkgsrc/using.html). For NetBSD 10.1 provide `echo 'PKG_PATH="https://cdn.netbsd.org/pub/pkgsrc/packages/NetBSD/x86_64/10.0_2025Q3/All/"' >/etc/pkg_install.conf`. If `pkgin` is taking too much time consider adding the following to `/etc/rc.conf`:
Install `pkgin` if not already `pkg_add pkgin`, see also the general [pkgsrc guide](https://www.netbsd.org/docs/pkgsrc/using.html). For NetBSD 10.1 provide `echo 'PKG_PATH="https://cdn.netbsd.org/pub/pkgsrc/packages/NetBSD/amd64/10.1/All/"' >/etc/pkg_install.conf`. If `pkgin` is taking too much time consider adding the following to `/etc/rc.conf`:
```sh
ip6addrctl=YES
ip6addrctl_policy=ipv4_prefer
```
System provides a default `g++-10` which doesn't support the current C++ codebase; install `clang-19` with `pkgin install clang-19`. Then build with `cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_CXX_COMPILER=clang++ -DCMAKE_C_COMPILER=clang -B build`.
System provides a default `g++-10` which doesn't support the current C++ codebase; install `clang-19` with `pkgin install clang-19`. Or install `gcc14` (or `gcc15` with current pkgsrc). Provided that, the following CMake commands may work:
Make may error out when generating C++ headers of SPIRV shaders, hence it's recommended to use `gmake` over the default system one.
[parallel/spirv-tools](https://iso.us.netbsd.org/pub/pkgsrc/current/pkgsrc/parallel/spirv-tools/index.html) isn't available in binary form and must be built from source.
glslang is not available on NetBSD, to circumvent this simply build glslang by yourself:
However, pkgsrc is highly recommended, see [getting pkgsrc](https://iso.us.netbsd.org/pub/pkgsrc/current/pkgsrc/doc/pkgsrc.html#getting). You must get `current` not the `2025Q2` version.
## DragonFlyBSD
If `libstdc++.so.6` is not found (`GLIBCXX_3.4.30`) then attempt:
This may have unforeseen consequences of which we don't need to worry about for now.
Default `g++` (and the libstdc++) is too outdated - so install `gcc14` and redirect CMake to the new compiler toolchain `-DCMAKE_CXX_COMPILER=gcc14 -DCMAKE_C_COMPILER=g++14`.
@@ -131,13 +155,9 @@ If build hangs, use `hammer2 bulkfree`.
## MSYS2
`qt6-static` isn't supported yet.
Only the `MINGW64` environment is tested (or `CLANGARM64` on ARM); however, all of the others should work (in theory) sans `MINGW32`.
Only the `MINGW64` environment is tested; however, all of the others should work (in theory) sans `MINGW32`.
Currently, only FFmpeg can be used as a system dependency; the others will result in linker errors.
When packaging an MSYS2 build, you will need to copy all dependent DLLs recursively alongside the `windeployqt6`; for example:
When packaging an MSYS2 build that is NOT fully static, you will need to copy all dependent DLLs recursively alongside the `windeployqt6`; for example:
The package install may randomly hang at times, in which case it has to be restarted. ALWAYS do a `sudo pkg update` or the chances of it hanging will be close to 90%. If "multiple" installs fail at once, try installing 1 by 1 the packages.
When CMake invokes certain file syscalls - it may sometimes cause crashes or corruptions on the (kernel?) address space - so reboot the system if there is a "hang" in CMake.
## Windows
### Windows 7, Windows 8 and Windows 8.1
DirectX 12 is not available - simply copy and paste a random DLL and name it `d3d12.dll`.
No support for Windows Vista (or below) is present at the moment. Check back later.
### Windows on ARM
If you're using Snapdragon X or 8CX, use the [the Vulkan translation layer](https://apps.microsoft.com/detail/9nqpsl29bfff?hl=en-us&gl=USE) only if the stock drivers do not work. And of course always keep your system up-to-date.
These are mostly "suggestions", if you feel like your code is readable, comprehensible to others; and most importantly doesn't result in unreadable spaghetti you're fine to go.
But for new developers you may find that following these guidelines will make everything x10 easier.
## Naming conventions
Simply put, types/classes are named as `PascalCase`, same for methods and functions like `AddElement`. Variables are named `like_this_snake_case` and constants are `IN_SCREAMING_CASE`.
Except for Qt MOC where `functionName` is preferred.
Template typenames prefer short names like `T`, `I`, `U`, if a longer name is required either `Iterator` or `perform_action` are fine as well.
Macros must always be in `SCREAMING_CASE`. Do not use short letter macros as systems like Solaris will conflict with them; a good rule of thumb is >5 characters per macro - i.e `THIS_MACRO_IS_GOOD`, `AND_ALSO_THIS_ONE`.
Try not using hungarian notation, if you're able.
## Formatting
Do not put if/while/etc braces after lines:
```c++
// no dont do this
if (thing)
{
some(); // ...
}
// do this
if (thing) {
some(); // ...
}
// or this
if (thing)
some(); // ...
// this is also ok
if (thing) some();
```
Brace rules are lax, if you can get the point across, do it:
```c++
// this is fine
do {
if (thing) {
return 0;
}
} while (other);
// this is also ok --- albeit a bit more dense
do if (thing) return 0; while (other);
// ok as well
do {
if (thing) return 0;
} while (other);
```
There is no 80-column limit but preferably be mindful of other developer's readability (like don't just put everything onto one line).
If you are on desktop and plan to use the Qt frontend, you *must* install Qt 6, and optionally Qt Creator (the **RECOMMENDED** IDE for building)
* On Linux, *BSD and macOS, this can be done by the package manager
- If you wish to use Qt Creator, append `qtcreator` or `qt-creator` to the commands seen below.
* If you wish to use Qt Creator, append `qtcreator` or `qt-creator` to the commands seen below.
* MSVC/clang-cl users on Windows must install through the official [Qt](https://www.qt.io/download-qt-installer-oss) installer
* Linux and macOS users may choose to use the installer as well.
* MSYS2 can also install Qt 6 via the package manager
If you are on Windows, a convenience script to install MSVC, MSYS2, Qt, all necessary packages for MSYS2, and set up a zsh environment with useful keybinds and aliases can be found [here](https://git.crueter.xyz/scripts/windev).
- For help setting up Qt Creator, run `./install.sh -h qtcreator`
* For help setting up Qt Creator, run `./install.sh -h qtcreator`
If you are on **Windows** and building with **MSVC** or **clang-cl**, you may go [back home](Build.md) and continue.
## Externals
The following are handled by Eden's externals:
* [FFmpeg](https://ffmpeg.org/) (should use `-DYUZU_USE_EXTERNAL_FFMPEG=ON`)
@@ -57,6 +59,7 @@ All other dependencies will be downloaded and built by [CPM](https://github.com/
- Using `qt6-static` is possible but currently untested.
- Other environments are entirely untested, but should theoretically work provided you install all the necessary packages.
- GCC is proven to work better with the MinGW environment. If you choose to use Clang, you *may* be better off using the clang64 environment.
- Add `qt-creator` to the `MINGW` variable to install Qt Creator. You can then create a Start Menu shortcut to the MinGW Qt Creator by running `powershell "\$s=(New-Object -COM WScript.Shell).CreateShortcut('C:\\ProgramData\\Microsoft\\Windows\\Start Menu\\Programs\\Qt Creator.lnk');\$s.TargetPath='C:\\msys64\\mingw64\\bin\\qtcreator.exe';\$s.Save()"` in Git Bash or MSYS2.
* Using `qt6-static` is possible as well, provided you build with `-DYUZU_STATIC_BUILD=ON`.
* Other environments are entirely untested, but should theoretically work provided you install all the necessary packages.
* GCC is proven to work better with the MinGW environment. On ARM, only Clang is available through the CLANGARM64 environment, so use that until a GNU ARM environment is available.
* Add `qt-creator` to the `MINGW` variable to install Qt Creator. You can then create a Start Menu shortcut to the MinGW Qt Creator by running `powershell "\$s=(New-Object -COM WScript.Shell).CreateShortcut('C:\\ProgramData\\Microsoft\\Windows\\Start Menu\\Programs\\Qt Creator.lnk');\$s.TargetPath='C:\\msys64\\mingw64\\bin\\qtcreator.exe';\$s.Save()"` in Git Bash or MSYS2.
* Add MinGW binaries to the PATH if they aren't already:
See also: [Dolphin emulator](hhttps://github.com/dolphin-emu/dolphin/blob/cdbea8867df3d0a6fc375e78726dae95612fb1fd/Source/Core/VideoCommon/DriverDetails.h#L84) own list of driver bugs (which are also included here).
## Vulkan
| Vendor/GPU | OS | Drivers | Version | Bug |
|---|---|---|---|---|
| AMD | All | Proprietary | ? | `VK_IMAGE_CREATE_CUBE_COMPATIBLE_BIT` on GCN4 and lower is broken. |
| AMD | All | Proprietary | ? | GCN4 and earlier have broken `VK_EXT_sampler_filter_minmax`. |
| AMD | All | Proprietary | ? | If offset + stride * count is greater than the size, then the last attribute will have the wrong value for vertex buffers. |
| AMD | All | Proprietary | ? | On GCN, 2D mipmapped texture arrays (with width == height) where there are more than 6 layers causes broken corrupt mipmaps. |
| AMD | All | RADV, MESA | ? | Using LLVM emitter causes corrupt FP16, proprietary drivers unaffected. |
| AMD | All | Proprietary | ? | Incorrect implementation of VK_EXT_depth_clamp_control causes incorrect depth values to be written to the depth buffer. |
| NVIDIA | All | Proprietary | ? | Drivers for ampere and newer have broken float16 math. |
| NVIDIA | All | Proprietary | >=510.0.0 | Versions >= 510 do not support MSAA->MSAA image blits. |
| NVIDIA | All | Proprietary | ? | Shader stencil export not supported. |
| NVIDIA | All | Proprietary | ? | Doesn't properly support conditional barriers. |
| NVIDIA | All | Proprietary | ? | GPUs pre-turing doesn't properly work with push descriptors. |
| NVIDIA | All | All | ? | Calling `vkCmdClearAttachments` with a partial rect, or specifying a render area in a render pass with the load op set to clear can cause the GPU to lock up, or raise a bounds violation. This only occurs on MSAA framebuffers, and it seems when there are multiple clears in a single command buffer. Worked around by back to the slow path (drawing quads) when MSAA is enabled. |
| Intel | All | Proprietary | <27.20.100.0 | Intel Windows versions before 27.20.100.0 has broken `VK_EXT_vertex_input_dynamic_state`. |
| Intel | All | Proprietary | ? | Intel proprietary drivers do not support MSAA->MSAA image blits. |
| Intel | All | Proprietary | 0.405.0<br>until<br>0.405.286 | Intel proprietary drivers 0.405.0 until 0.405.286 have broken compute. |
| Intel | macOS | Proprietary | ? | Using dynamic sampler indexing locks up the GPU. |
| Intel | macOS | Proprietary | ? | Using subgroupMax in a shader that can discard results in garbage data. |
| Intel | All | Mesa | ? | Broken lines in geometry shaders when writing to `gl_ClipDistance` in the vertex shader. |
| Qualcomm | All | Proprietary | ? | Using too many samplers (on A8XX is `65536`) can cause heap exhaustion, recommended to use 75% of total available samplers. |
| Qualcomm | All | Proprietary | ? | 64-bit integer extensions (`VK_KHR_SHADER_ATOMIC_INT64_EXTENSION_NAME`) cause nondescriptive crashes and stability issues. |
| Qualcomm | All | All | ? | Driver requires higher-than-reported binding limits (32). |
| Qualcomm | All | All | ? | On Adreno, enabling rasterizer discard somehow corrupts the viewport state; a workaround is forcing it to be updated on next use. |
| Qualcomm | All | Proprietary | ? | Concurrent fence waits are unsupported. |
| Qualcomm | All | Proprietary | ? | `vkCmdCopyImageToBuffer` allocates a staging image when used to copy from an image with optimal tiling; which results in overall slower performance. |
| Qualcomm | All | Proprietary | ? | 32-bit depth clears are broken in the Adreno Vulkan driver, and have no effect. |
| Qualcomm | All | Proprietary | ? | It should be safe to release the resources before actually resetting the VkCommandPool. However, devices running R drivers there was a few months period where the driver had a bug which it incorrectly was accessing objects on the command buffer while it was being reset. If these objects were already destroyed (which is a valid thing to do) it would crash. |
| Qualcomm | All | Proprietary | ? | On Pixel and Pixel2XL's with Adreno 530 and 540s, setting width and height to 10s reliably triggers what appears to be a driver race condition. |
| Qualcomm | All | Proprietary | ? | Non-descriptive broken OpPhi SPIRV lowering, originally using OpPhi to choose the result is crashing on Adreno 4xx. Switched to storing the result in a temp variable as glslang does. |
| Qualcomm | All | Proprietary | ? | See crbug.com/1241134. The bug appears on Adreno 5xx devices with OS PQ3A. It does not repro on the earlier PPR1 version since the extend blend func extension was not present on the older driver. |
| Mali | Android | Proprietary | ? | Non-descriptive green screen on various locations. |
| Mali | Android | Proprietary | ? | Cached memory is significantly slower for readbacks than coherent memory in the Mali Vulkan driver, causing high CPU usage in the `__pi___inval_cache_range` kernel function. |
| Mali | Android | Proprietary | ? | On some old ARM driver versions, dynamic state for stencil write mask doesn't work correctly in the presence of discard or alpha to coverage, if the static state provided when creating the pipeline has a value of 0 (`alphaToCoverageEnable` and `rasterizerDiscardEnable`). |
| Mali | Android | Proprietary | ? | Failing to submit because of a device loss still needs to wait for the fence to signal before deleting. However, there is an ARM bug (b/359822580) where the driver early outs on the fence wait if in a device lost state and thus we can't wait on it. Instead, we just wait on the queue to finish. |
| Mali | Android | Proprietary | ? | With Galaxy S7 we see lots of rendering issues when we suballocate VkImages. |
| Mali | Android | Proprietary | ? | With Galaxy S7 and S9 we see lots of rendering issues with image filters dropping out when using only primary command buffers. We also see issues on the P30 running Android 28. |
| Mali | Android | Proprietary | ? | `RGBA_F32` mipmaps appear to be broken on some Mali devices. |
| Mali | Android | Proprietary | ? | Matrix IR lowering for matrix swizzle, scalar multiplication and unary `(+m)`/`(-m)` present extraneous unexplained bugs with more than 32 matrix temporals. |
| Apple | All | MoltenVK | ? | Driver breaks when using more than 16 vertex attributes/bindings. |
| Apple | macOS | Proprietary | >4 | Some driver and Apple Silicon GPU combinations have problems with fragment discard when early depth test is enabled. Discarded fragments may appear corrupted. |
| Apple | iOS | MoltenVK | ? | Null descriptors cause non-descriptive issues. |
| Apple | iOS | MoltenVK | ? | Push descriptors cause non-descriptive issues. |
| Imagination | Android | Proprietary | ? | Some vulkan implementations don't like the 'clear' loadop renderpass if you try to use a framebuffer with a different load/store op than that which it was created with, despite the spec saying they should be compatible. |
| Intel<br>Qualcomm<br>AMD<br>Apple | All | All<br>MoltenVK (Apple) | ? | Reversed viewport depth range does not work as intended on some Vulkan drivers. The Vulkan spec allows the `minDepth`/`maxDepth` fields in the viewport to be reversed, however the implementation is broken on some drivers. |
To change these options, add `-DOPTION_NAME=NEWVALUE` to the command line.
- On Qt Creator, go to Project -> Current Configuration
Notes:
- Defaults are marked per-platform.
- "Non-UNIX" just means Windows/MSVC and Android (yes, macOS is UNIX
- Android generally doesn't need to change anything; if you do, go to `src/android/app/build.gradle.kts`
- To set a boolean variable to on, use `ON` for the value; to turn it off, use `OFF`
- If a variable is mentioned as being e.g. "ON" for a specific platform(s), that means it is defaulted to OFF on others
- TYPE is always boolean unless otherwise specified
- Format:
*`OPTION_NAME` (TYPE DEFAULT) DESCRIPTION
## Options
-`YUZU_USE_CPM` (ON for non-UNIX) Use CPM to fetch system dependencies (fmt, boost, etc) if needed. Externals will still be fetched. See the [CPM](CPM.md) and [Deps](Deps.md) docs for more info.
-`ENABLE_WEB_SERVICE` (ON) Enable multiplayer service
* [NDK 27+ and CMake 3.22.1](https://developer.android.com/studio/projects/install-ndk#default-version)
* [Git](https://git-scm.com/download)
### WINDOWS ONLY - Additional Dependencies
* **[Visual Studio 2022 Community](https://visualstudio.microsoft.com/downloads/)** - **Make sure to select "Desktop development with C++" support in the installer. Make sure to update to the latest version if already installed.**
* **[Vulkan SDK](https://vulkan.lunarg.com/sdk/home#windows)** - **Make sure to select Latest SDK.**
- A convenience script to install the latest SDK is provided in `.ci\windows\install-vulkan-sdk.ps1`.
## WINDOWS ONLY - Additional Dependencies
* **[Visual Studio 2022 Community](https://visualstudio.microsoft.com/downloads/)** - **Make sure to select "Desktop development with C++" support in the installer. Make sure to update to the latest version if already installed.**
* **[Vulkan SDK](https://vulkan.lunarg.com/sdk/home#windows)** - **Make sure to select Latest SDK.**
* A convenience script to install the latest SDK is provided in `.ci\windows\install-vulkan-sdk.ps1`.
1. Start Android Studio, on the startup dialog select `Open`.
2. Navigate to the `eden/src/android` directory and click on `OK`.
3. In `Build > Select Build Variant`, select `release` or `relWithDebInfo` as the "Active build variant".
4. Build the project with `Build > Make Project` or run it on an Android device with `Run > Run 'app'`.
## Building with Terminal
1. Download the SDK and NDK from Android Studio.
2. Navigate to SDK and NDK paths.
3. Then set ANDROID_SDK_ROOT and ANDROID_NDK_ROOT in terminal via
@@ -38,7 +45,44 @@ Eden by default will be cloned into -
Remember to have a Java SDK installed if not already, on Debian and similar this is done with `sudo apt install openjdk-17-jdk`.
### Script
A convenience script for building is provided in `.ci/android/build.sh`. The built APK can be put into an `artifacts` directory via `.ci/android/package.sh`. On Windows, these must be done in the Git Bash or MinGW terminal.
A convenience script for building is provided in `.ci/android/build.sh`. On Windows, this must be run in Git Bash or MSYS2. This script provides the following options:
This terminal instruction conditionally executes one terminal or another depending
on the run-time state of the ARM flags.
# Register Allocation (x64 Backend)
`HostLoc`s contain values. A `HostLoc` ("host value location") is either a host CPU register or a host spill location.
Values once set cannot be changed. Values can however be moved by the register allocator between `HostLoc`s. This is
handled by the register allocator itself and code that uses the register allocator need not and should not move values
between registers.
The register allocator is based on three concepts: `Use`, `Def` and `Scratch`.
* `Use`: The use of a value.
* `Define`: The definition of a value, this is the only time when a value is set.
* `Scratch`: Allocate a register that can be freely modified as one wishes.
Note that `Use`ing a value decrements its `use_count` by one. When the `use_count` reaches zero the value is discarded and no longer exists.
The member functions on `RegAlloc` are just a combination of the above concepts.
The following registers are reserved for internal use and should NOT participate in register allocation:
- `%xmm0`, `%xmm1`, `%xmm2`: Used as scratch in exclusive memory access.
- `%rsp`: Stack pointer.
- `%r15`: JIT pointer
- `%r14`: Page table pointer.
- `%r13`: Fastmem pointer.
The layout convenes `%r15` as the JIT state pointer - while it may be tempting to turn it into a synthetic pointer, keeping an entire register (out of 12 available) is preferable over inlining a directly computed immediate.
Do NEVER modify `%r15`, we must make it clear that this register is "immutable" for the entirety of the JIT block duration.
At runtime, allocate one of the registers in `desired_locations`. You are free to modify the register. The register is discarded at the end of the allocation scope.
### Pure `Use`
```c++
Xbyak::Reg64 UseGpr(Argument& arg);
Xbyak::Xmm UseXmm(Argument& arg);
OpArg UseOpArg(Argument& arg);
void Use(Argument& arg, HostLoc host_loc);
```
At runtime, the value corresponding to `arg` will be placed a register. The actual register is determined by
which one of the above functions is called. `UseGpr` places it in an unused GPR, `UseXmm` places it
in an unused XMM register, `UseOpArg` might be in a register or might be a memory location, and `Use` allows
you to specify a specific register (GPR or XMM) to use.
This register **must not** have it's value changed.
### `UseScratch`
```c++
Xbyak::Reg64 UseScratchGpr(Argument& arg);
Xbyak::Xmm UseScratchXmm(Argument& arg);
void UseScratch(Argument& arg, HostLoc host_loc);
```
At runtime, the value corresponding to `arg` will be placed a register. The actual register is determined by
which one of the above functions is called. `UseScratchGpr` places it in an unused GPR, `UseScratchXmm` places it
in an unused XMM register, and `UseScratch` allows you to specify a specific register (GPR or XMM) to use.
The return value is the register allocated to you.
You are free to modify the value in the register. The register is discarded at the end of the allocation scope.
### `Define` as register
A `Define` is the defintion of a value. This is the only time when a value may be set.
code->cmp(rbx, qword[code.ABI_JIT_PTR + offsetof(JitState, rsb_location_descriptors) + i * sizeof(u64)]);
code->cmove(rax, qword[code.ABI_JIT_PTR + offsetof(JitState, rsb_codeptrs) + i * sizeof(u64)]);
}
code->jmp(rax);
}
```
In pseudocode:
```c++
rbx := ComputeUniqueHash()
rax := ReturnToDispatch
for (i := 0 .. RSBSize-1)
if (rbx == rsb_location_descriptors[i])
rax = rsb_codeptrs[i]
goto rax
```
# Fast memory (Fastmem)
The main way of accessing memory in JITed programs is via an invoked function, say "Read()" and "Write()". On our translator, such functions usually take a sizable amounts of code space (push + call + pop). Trash the i-cache (due to an indirect call) and overall make code emission more bloated.
The solution? Delegate invalid accesses to a dedicated arena, similar to a swap. The main idea behind such mechanism is to allow the OS to transmit page faults from invalid accesses into the JIT translator directly, bypassing address space calls, while this sacrifices i-cache coherency, it allows for smaller code-size and "faster" throguhput.
Many kernels however, do not support fast signal dispatching (Solaris, OpenBSD, FreeBSD). Only Linux and Windows support relatively "fast" signal dispatching. Hence this feature is better suited for them only.


In x86_64 for example, when a page fault occurs, the CPU will transmit via control registers and the stack (see `IRETQ`) the appropriate arguments for a page fault handler, the OS then will transform that into something that can be sent into userspace.
Most modern OSes implement kernel-page-table-isolation, which means a set of system calls will invoke a context switch (not often used syscalls), whereas others are handled by the same process address space (the smaller kernel portion, often used syscalls) without needing a context switch. This effect can be negated on systems with PCID (up to 4096 unique IDs).
Signal dispatching takes a performance hit from reloading `%cr3` - but Linux does something more clever to avoid reloads: VDSO will take care of the entire thing in the same address space. Making dispatching as costly as an indirect call - without the hazards of increased code size.
The main downside from this is the constant i-cache trashing and pipeline hazards introduced by the VDSO signal handlers. However on most benchmarks fastmem does perform faster than without (Linux only). This also abuses the fact of continous address space emulation by using an arena - which can then be potentially transparently mapped into a hugepage, reducing TLB walk times.
AI is a *tool*, not a replacement or catch-all solution. It is generally okay at a few *very specific* use cases:
- Automation of tedious changes where you have already made the pattern clear and done the necessary groundwork.
- Conversion of code from one paradigm to another.
For everything else, AI is subpar at best, and actively harmful at worst. In general, you are **heavily** encouraged to not use AI at all.
## Why?
AI is notorious for hallucinating facts out of thin air and sometimes outright lying to users. Additionally, code written by LLMs is often needlessly verbose and horrifically inefficient (not to mention the rather ridiculous level of over-commenting). The end result is often one of three things:
- Completely nonfunctional code
- Code that works, but is extraordinarily verbose or not nearly as efficient as it can be
- Code that works well and is written well, but solves a different problem than was intended, or solves the same problem but in a completely incorrect way that will break other things horribly.
Human-written code will, without exception, always be of infinitely higher quality when properly researched and implemented by someone both familiar with the surrounding code and the programming language in use. LLMs may produce a "good enough" result, but this result is often subpar. Keep in mind: all code is held under a standard of excellence. If your code sucks, it will be rejected. AI-generated code just so happens to be a particularly sucky genre of code, and will thus be held to this same standard.
On a lesser-known note, LLM outputs often contain unicode symbols such as emojis or the arrow symbol. Please don't put Unicode symbols in your code. It messes with many an IDE, and the three people viewing your code on Lynx will be very unhappy.
**Learn to code**. It's worth it, we promise!
## Acceptable Use
- As stated previously, AI is good in a few *very specific* cases. In these cases, it's usually fine to use AI, as long as you **explicitly provide notice that it was used**.
- Anything directly outside of the realm of the code written in your PR or patch is none of our business.
- This primarily covers research.
- However, we *still* strongly discourage this for the reasons mentioned above.
- Assistance with cleanups, and minor nitpicks/optimizations.
- This is still discouraged, but it's okay to occasionally use LLMs to catch any minor mistakes you made in your code.
- Debugging
- In general, LLMs are not good at debugging, but if you solve a bug you're facing with help from an AI, and said fix **works properly**, then that's fine.
## Unacceptable Use
Well, everything else. But here are a few examples:
- Commit messages
- LLMs are absolutely horrible at this. They are needlessly verbose, almost never catch the actual intent of the commit, and will almost always hallucinate false information about said changes.
- See the addendum for an example
- Solving problems
- Slapping a few files and a "please fix bug XYZ" into an LLM is a recipe for disaster that will pretty much never work.
- Fully AI-generated code
- This shouldn't need explaining. Do not do this under any circumstance, especially if you don't actually understand what's going on.
- Writing code based on pseudo-instructions
- If you don't know how to write code, don't. If you've figured out the root cause (preferably without feeding random files into an LLM) and actively know what's going on, provide information to other developers or friends of yours who have knowledge of the language and/or the codebase.
## Addendum: Commit Messages
The patchset for pull request [#3422](https://git.eden-emu.dev/eden-emu/eden/pulls/3422) was fed into several LLMs to generate a commit message. All of them sucked, and not a single one caught on to what the commit actually did. For example:
```txt
profile_manager: Refactor user deletion to use index instead of UUID
This commit improves the user profile deletion flow in the following ways:
1. **UI layer (ConfigureProfileManager)**:
- Replaces UUID-based deletion with index-based deletion
- Passes the selected row index from the tree view to the confirmation dialog
- Emits deleteUser(int index) signal from ConfigureProfileManagerDeleteDialog
- Removes lambda capture and std::function callback in favor of direct signal/slot connection
2. **Delete dialog**:
- Changes SetInfo() signature to accept index instead of accept callback
- Stores index internally and emits it via deleteUser signal on acceptance
- Removes duplicated stable_partition + save logic from RemoveUser()
- Centralizes profile removal and save-file writing in RemoveProfileAtIndex()
4. **Minor follow-up**:
- Adds profile_manager.ResetUserSaveFile() call in PopulateUserList() to ensure consistent state when refreshing the user list
These changes eliminate redundant code, reduce the risk of UUID → index mismatch bugs, and make the deletion path more straightforward and maintainable.
Fixes: potential inconsistency when user order changes between UI and backend
```
This is:
- Needlessly verbose (nobody cares about most of these details)
- Doesn't address the actual purpose of the PR (fixing a double-deletion bug in the profile manager)
- Has unicode arrows
- Uses overly corporate and, well, robotic language
As another example:
```txt
profile(manager/ui): switch to index-based deletion and unify removal logic
- Make ConfigureProfileManagerDeleteDialog emit a deleteUser(int) signal instead of taking a callback. Store the target index in the dialog (m_index) and emit it on acceptance.
- Change ConfigureProfileManager to pass an index to SetInfo and to accept an index in DeleteUser. Wire the dialog signal to the new DeleteUser slot.
- Update headers: add the signal and m_index, add TODO to move dialog to a .ui file, and update slot/signature declarations.
- Add ProfileManager::RemoveProfileAtIndex(std::size_t) and refactor RemoveUser(UUID) to call RemoveProfileAtIndex to avoid duplicated removal logic. Ensure the removal path marks saves as needed and writes the user save file.
- Ensure the profile list updates immediately after deletes by calling profile_manager.ResetUserSaveFile() when populating the user list (qlaunch fix).
- Misc: update SPDX copyright year and fix build breakages caused by the API changes.
This consolidates profile removal behavior, fixes potential race conditions in the profile dialog, and removes duplicated removal code.
```
This has all of the same problems as the other one. Needlessly verbose, doesn't address *what* it actually fixes ("consolidates profile removal behavior"... okay, why? What does it fix?), etc. It even has the bonus of totally hallucinating the addition of a method!
These are mostly "suggestions", if you feel like your code is readable, comprehensible to others; and most importantly doesn't result in unreadable spaghetti you're fine to go.
But for new developers you may find that following these guidelines will make everything x10 easier.
## Naming conventions
Simply put, types/classes are named as `PascalCase`, same for methods and functions like `AddElement`. Variables are named `like_this_snake_case` and constants are `IN_SCREAMING_CASE`.
Except for Qt MOC where `functionName` is preferred.
Template typenames prefer short names like `T`, `I`, `U`, if a longer name is required either `Iterator` or `perform_action` are fine as well. Do not use names like `SS` as systems like solaris define it for registers, in general do not use any of the following for short names:
-`SS`, `DS`, `GS`, `FS`: Segment registers, defined by Solaris `<ucontext.h>`
-`EAX`, `EBX`, `ECX`, `EDX`, `ESI`, `EDI`, `ESP`, `EBP`, `EIP`: Registers, defined by Solaris.
-`X`: Defined by some utility headers, avoid.
-`_`: Defined by gettext, avoid.
-`N`, `M`, `S`: Preferably don't use this for types, use it for numeric constants.
-`TR`: Used by some weird `<ucontext.h>` whom define the Task Register as a logical register to provide to the user... (Need to remember which OS in specific).
Macros must always be in `SCREAMING_CASE`. Do not use short letter macros as systems like Solaris will conflict with them; a good rule of thumb is >5 characters per macro - i.e `THIS_MACRO_IS_GOOD`, `AND_ALSO_THIS_ONE`.
Try not using hungarian notation, if you're able.
## Formatting
Formatting is extremelly lax, the general rule of thumb is: Don't add new lines just to increase line count. The less lines we have to look at, the better. This means also packing densely your code while not making it a clusterfuck. Strike a balance of "this is a short and comprehensible piece of code" and "my eyes are actually happy to see this!". Don't just drop the entire thing in a single line and call it "dense code", that's just spaghetti posing as code. In general, be mindful of what other devs need to look at.
Do not put if/while/etc braces after lines:
```c++
// no dont do this
// this is more lines of code for no good reason (why braces need their separate lines?)
// and those take space in someone's screen, cumulatively
if (thing)
{ //<--
some(); // ...
} //<-- 2 lines of code for basically "opening" and "closing" an statment
// do this
if (thing) { //<-- [...] and with your brain you can deduce it's this piece of code
// that's being closed
some(); // ...
} //<-- only one line, and it's clearer since you know its closing something [...]
// or this, albeit the extra line isn't needed (at your discretion of course)
if (thing)
some(); // ...
// this is also ok, keeps things in one line and makes it extremely clear
if (thing) some();
// NOT ok, don't be "clever" and use the comma operator to stash a bunch of statments
// in a single line, doing this will definitely ruin someone's day - just do the thing below
// vvv
if (thing) some(), thing(), a2(a1(), y1(), j1()), do_complex_shit(wa(), wo(), ploo());
// ... and in general don't use the comma operator for "multiple statments", EXCEPT if you think
// that it makes the code more readable (the situation may be rare however)
// Wow so much clearer! Now I can actually see what each statment is meant to do!
if (thing) {
some();
thing();
a2(a1(), y1(), j1());
do_complex_shit(wa(), wo(), ploo());
}
```
Brace rules are lax, if you can get the point across, do it:
```c++
// this is fine
do {
if (thing) {
return 0;
}
} while (other);
// this is also ok --- albeit a bit more dense
do if (thing) return 0; while (other);
// ok as well
do {
if (thing) return 0;
} while (other);
```
There is no 80-column limit but preferably be mindful of other developer's readability (like don't just put everything onto one line).
A note about operators: Use them sparingly, yes, the language is lax on them, but some usages can be... tripping to say the least.
```c++
a, b, c; //<-- NOT OK multiple statments with comma operator is definitely a recipe for disaster
return c ? a : b; //<-- OK ternaries at end of return statments are clear and fine
return a, b; //<-- NOT OK return will take value of `b` but also evaluate `a`, just use a separate statment
void f(int a[]) //<-- OK? if you intend to use the pointer as an array, otherwise just mark it as *
```
And about templates, use them sparingly, don't just do meta-templating for the sake of it, do it when you actually need it. This isn't a competition to see who can make the most complicated and robust meta-templating system. Just use what works, and preferably stick to the standard libary instead of reinventing the wheel. Additionally:
```c++
// NOT OK This will create (T * N * C * P) versions of the same function. DO. NOT. DO. THIS.
* Use `DEVEL=true` to rename app to `Eden Nightly` and disable the update checker
* This should generally not be used unless in a tailor-made packaging environment (e.g. Actions/CI)
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.