Compare commits

..

16 Commits

Author SHA1 Message Date
MaranBr 7c49b70107 Minor improvement 2026-10-01 10:35:37 -04:00
MaranBr 3f41a04732 Ops 2026-10-01 09:26:42 -04:00
MaranBr 76fa7445d0 Remove Legacy Rescale Pass hack 2026-10-01 09:13:22 -04:00
MaranBr c48b6bccee Fix headers 2026-10-01 09:08:07 -04:00
MaranBr bf3b6886a0 Skip unsupported fragment shader pattern on non-NVIDIA drivers 2026-10-01 09:04:40 -04:00
lizzie d3550c4571 [qt] Filter collates accents (Unicode NFD) (#4380)
Now "POKEMON" and "pokemon" are the same, and search's no longer case sensitive.
Maybe we could transliterate? But that would require icu which is a bit overkill.
Main point is to be able to search "Pokemon" (accent on 'e') without actually typing the accent

https://unicode.org/reports/tr15/

Signed-off-by: lizzie <lizzie@eden-emu.dev>

- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.

-------------------

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4380
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
2026-10-01 07:39:44 +02:00
John 963119e7cc [ffmpeg]Gamescope / Gamemode NVDEC Emulation - GPU Decoding Patches (#4501)
Switch to ffmpeg with liz's gamescope patches.

- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.

-------------------

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4501
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: lizzie <lizzie@eden-emu.dev>
2026-10-01 03:34:57 +02:00
xbzk 2a7e9b53b1 [android,ui] fixed shader cache clearing when changing drivers (#4496)
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.

-------------------
Trivial bug pending on backlog, where changing drivers was not clearing up sharder cache as expected.

The driver flow (broken) was deleting externalFilesDir/shader/<game>, while the visible cache-size control (working) DirectoryInitialization.userDirectory/cache/shader/<game>.

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4496
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: lizzie <lizzie@eden-emu.dev>
2026-10-01 00:28:18 +02:00
xbzk 8e2d26c272 [video_core] buffer cache readback for nce invalidations (#4490)
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.

-------------------

This is a follow up of 4473, a fix for the next Ender Magnolia known crash, the shackled beast's vaper crash, but specific for NCE.

Imma try to be short:

NCE invalidation was dooming whole guest pages when it shouldn't. If the GPU buffer cache owns some bytes on a page, invalidating first discards their ownership, and the retrying CPU can then observe stale data. On Android, this was corrupting some indirect draw packets and causing huge invalid index count reach staging memory allocation.

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4490
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: lizzie <lizzie@eden-emu.dev>
2026-09-30 22:13:37 +02:00
Eden CI f8b7651104 [dist, android] Update translations from Transifex for Sep 28 (#4494)
Automatic translation update for Sep 28

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4494
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
2026-09-30 02:00:44 +02:00
xbzk 3f78512c83 [ro] add TLS patching for runtime NROs (#4492)
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.

-------------------

Aimed to fix Biomutant which stucks in a bootloop due to guest being unable to reach correct thread local storage (TLS)
This fixes runtime NRO loading by properly patching the guest TLS (TPIDR_EL0) path. Steps:
-Read the loaded NRO image into a temporary buffer.
-Run the NCE patcher on it.
-Produce a patch section / relocated code.
-Map that patch section as executable memory in the target process.
Game still has intermittent visual issues which can be solved with GPU level accurate, but that causes a huge performance drop in exchange. Someone else will resume this one.

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4492
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: lizzie <lizzie@eden-emu.dev>
2026-09-30 00:58:29 +02:00
xbzk a82f31a310 [settings,qt] merge of "disable controller applet" toggle with "controller applet mode" picker (#4498)
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.

-------------------
Small good simplification to nuke a toggle which costed me almost an entire day of work when fixing controller applet issues:

There were two settings related to Controller applet:
-"disable controller applet", which is a boolean toggle;
-"controller configuration" applet mode, which is a "Real"/"Custom" picker;

All i did was to nuke the toggle and add a third option "Disabled" to the picker.
Users which used the disable toggle will have to pick Disabled;
Real/Custom settings wont be affected since their indexes remains.
All options tested against diablo 3. All working as intended.

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4498
Reviewed-by: lizzie <lizzie@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
2026-09-30 00:57:23 +02:00
PavelBARABANOV 815325ccec [loader] add ZBIC (zstd variant) NSO decompression support for Switch 22.0+ (#4482)
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.

-------------------
A new NSO compression method was introduced in Switch 22.0.0. This is a
customized variant of zstd and is used when NSO flags have bit 7 set.

Key characteristics:
  - ZSTD_MAGICNUMBER is set to 0x4349425A (b'ZBIC') instead of 0xFD2FB528
  - ZSTD_LEGACY_SUPPORT is set to 0
  - ZSTD_TRACE is set to 1, zstd version used is 1.5.7 (10507)
  - FSE_readNCount is replaced with a BIC (Binary Interpolative Coding)
    version which improves compression of entropy tables significantly

Source: https://switchbrew.org/wiki/22.0.0

Implementation:
  - Detect ZBIC segments via NSO flag bit 7 (NsoFlags_UseZbicCompression)
    and/or ZBIC magic scan in nso.cpp
  - Fall back to LZ4 when ZBIC is not detected or returns unexpected size
  - Handle segment 0 alternate offset (0x100 vs 0x108) for both ZBIC and LZ4

References:
  - Atmosphère loader/strat zstd-zbic integration:
    https://github.com/Atmosphere-NX/Atmosphere/commit/082115187a0509cb6b8a757de639a4e4741b8712
  - nxdumptool ZBIC segment compression support:
    https://github.com/DarkMatterCore/nxdumptool/commit/441e5c0904f29427987433be4eb057a2843d222a
  - STORM_SWITCH ZBIC implementation:
    https://github.com/ReiKatari/STORM_SWITCH/commit/1e09eb82b760aaf340b810031400991c0740c091

Tested with firmware 23.0.0 and ZBIC-compressed NSOs.

Co-authored-by: xbzk <xbzk@eden-emu.dev>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4482
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
Reviewed-by: lizzie <lizzie@eden-emu.dev>
Reviewed-by: Maufeat <sahyno1996@gmail.com>
2026-09-27 12:48:45 +02:00
Exverge 37fe911952 [common/sparse_large_vector] correct Win32 exception handler (#4481)
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.

-------------------
Adds mutex to prevent multiple threads from accessing the vector at the same time and corrects the Windows exception handler to use the correct faulting address used by the exception handler (previously used the instruction address instead of the fault address) and properly shifts the stored values for the handler.
Fixes weird compiler-specific bugs on Windows

Co-authored-by: bruno <protoxseven@gmail.com>
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4481
Reviewed-by: lizzie <lizzie@eden-emu.dev>
Reviewed-by: MaranBr <maranbr@eden-emu.dev>
2026-09-26 16:02:35 +02:00
MaranBr f273423b2b [buffer_cache] Simplify GPU fence synchronization and remove GPU buffer readback (#4477)
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.

-------------------

This is understood to improve GPU synchronization within the buffer cache in certain edge cases.

GPU Fence Strict is no longer necessary. We now apply the stronger synchronization only in the specific case that actually requires it.

The GPU Buffer Readback has been removed, as it is no longer needed following PR #4473.

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4477
Reviewed-by: CamilleLaVey <camillelavey99@gmail.com>
2026-09-26 05:01:27 +02:00
MaranBr 87d2f03c39 [hid_core] Code cleanup (#4461)
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md)
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.

-------------------
This is just a code cleanup. This is no longer necessary due to commit #4457.

Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4461
Reviewed-by: lizzie <lizzie@eden-emu.dev>
2026-09-26 02:21:45 +02:00
68 changed files with 12095 additions and 12549 deletions
+1 -6
View File
@@ -5,11 +5,6 @@
"repo": "lioncash/biscuit",
"version": "v0.19.0"
},
"atmosphere_zstd_bic": {
"hash": "204f7573ea40de2878759c7ff5b2c016111fc29ae8574df50a3a0db9018c3b012f9040ea7466559130424d4335f2dfd6d7abfdc4d4d67e8a34531b8284f0cc6a",
"repo": "Atmosphere-NX/Atmosphere",
"version": "3ace80cc7b87e1837fccca1522801bbf6fddc6aa"
},
"boost": {
"artifact": "boost-%VERSION%.tar.zst",
"find_args": "CONFIG OPTIONAL_COMPONENTS headers context system fiber filesystem",
@@ -84,7 +79,7 @@
"name": "ffmpeg",
"package": "FFmpeg",
"repo": "crueter-ci/FFmpeg",
"version": "9.0.1-1788120736-bf1b838f2a"
"version": "9.0.2-1790726865-946fcce07b"
},
"fmt": {
"hash": "f0da82c545b01692e9fd30fdfb613dbb8dd9716983dcd0ff19ac2a8d36f74beb5540ef38072fdecc1e34191b3682a8542ecbf3a61ef287dbba0a2679d4e023f2",
+455 -473
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+440 -463
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+440 -463
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+440 -463
View File
File diff suppressed because it is too large Load Diff
+440 -463
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
+440 -463
View File
File diff suppressed because it is too large Load Diff
+439 -456
View File
File diff suppressed because it is too large Load Diff
@@ -30,7 +30,6 @@ enum class BooleanSetting(override val key: String) : AbstractBooleanSetting {
RENDERER_REACTIVE_FLUSHING("use_reactive_flushing"),
ENABLE_BUFFER_HISTORY("enable_buffer_history"),
USE_OPTIMIZED_VERTEX_BUFFERS("use_optimized_vertex_buffers"),
ENABLE_GPU_BUFFER_READBACK("enable_gpu_buffer_readback"),
SYNC_MEMORY_OPERATIONS("sync_memory_operations"),
BUFFER_REORDER_DISABLE("disable_buffer_reorder"),
RENDERER_DEBUG("debug"),
@@ -908,13 +908,6 @@ abstract class SettingsItem(
descriptionId = R.string.enable_buffer_history_description
)
)
put(
SwitchSetting(
BooleanSetting.ENABLE_GPU_BUFFER_READBACK,
titleId = R.string.enable_gpu_buffer_readback,
descriptionId = R.string.enable_gpu_buffer_readback_description
)
)
put(
SwitchSetting(
BooleanSetting.USE_OPTIMIZED_VERTEX_BUFFERS,
@@ -549,7 +549,6 @@ class SettingsFragmentPresenter(
add(BooleanSetting.RENDERER_FORCE_MAX_CLOCK.key)
add(BooleanSetting.RENDERER_REACTIVE_FLUSHING.key)
add(BooleanSetting.ENABLE_BUFFER_HISTORY.key)
add(BooleanSetting.ENABLE_GPU_BUFFER_READBACK.key)
add(BooleanSetting.USE_OPTIMIZED_VERTEX_BUFFERS.key)
add(HeaderSetting(R.string.hacks))
@@ -23,6 +23,7 @@ import org.yuzu.yuzu_emu.model.Driver.Companion.toDriver
import org.yuzu.yuzu_emu.utils.GpuDriverHelper
import org.yuzu.yuzu_emu.NativeLibrary
import org.yuzu.yuzu_emu.utils.GpuDriverMetadata
import org.yuzu.yuzu_emu.utils.DirectoryInitialization
import org.yuzu.yuzu_emu.utils.NativeConfig
import java.io.File
@@ -146,11 +147,11 @@ class DriverViewModel : ViewModel() {
private fun wipeGameShaders(game: Game) {
viewModelScope.launch {
withContext(Dispatchers.IO) {
val externalFilesDir = YuzuApplication.appContext.getExternalFilesDir(null)
val userDirectory = DirectoryInitialization.userDirectory
?: return@withContext
val shaderDir = File(
externalFilesDir.absolutePath +
"/shader/" + game.settingsName.lowercase()
userDirectory +
"/cache/shader/" + game.settingsName.lowercase()
)
if (shaderDir.exists()) {
shaderDir.deleteRecursively()
@@ -580,8 +580,6 @@
<string name="renderer_reactive_flushing_description">يحسن دقة العرض في بعض الألعاب على حساب الأداء.</string>
<string name="enable_buffer_history">تمكين سجل التخزين المؤقت</string>
<string name="enable_buffer_history_description">يُتيح هذا الخيار الوصول إلى حالات التخزين المؤقت السابقة. وقد يُحسّن جودة العرض وثبات الأداء في بعض الألعاب.</string>
<string name="enable_gpu_buffer_readback">تفعيل قراءة مخزن وحدة معالجة الرسومات</string>
<string name="enable_gpu_buffer_readback_description">يحافظ هذا النظام على بيانات المخزن المؤقت المُعدّلة بواسطة وحدة معالجة الرسومات عن طريق قراءتها مرة أخرى قبل التحميل. تتطلب بعض الألعاب ذلك لعرض بعض التأثيرات بشكل صحيح. قد يُسبب ذلك مشاكل إذا لم يتمكن الجهاز من التعامل مع عبء العمل الإضافي.</string>
<string name="use_optimized_vertex_buffers">مخازن الرؤوس المُحسّنة</string>
<string name="use_optimized_vertex_buffers_description">يُتيح ربطًا مُحسَّنًا لمخازن الرؤوس لتحسين الأداء. يتطلب برامج تشغيل Mesa 26.0+ Turnip/ برامج تشغيل QCOM. قد يتعطل على برامج تشغيل Turnip القديمة (25.3 وما دون).</string>
@@ -1099,7 +1097,6 @@
<string name="gpu_fence_behavior_immediate">فوري</string>
<string name="gpu_fence_behavior_balanced">متوازن</string>
<string name="gpu_fence_behavior_accurate">دقيق</string>
<string name="gpu_fence_behavior_strict">صارم</string>
<string name="vram_usage_conservative">محافظ</string>
<string name="vram_usage_aggressive">عدواني</string>
@@ -967,7 +967,6 @@ Wirklich fortfahren?</string>
<string name="gpu_fence_behavior_immediate">Direkt</string>
<string name="gpu_fence_behavior_balanced">Ausgewogen</string>
<string name="gpu_fence_behavior_accurate">Genau</string>
<string name="gpu_fence_behavior_strict">Strikt</string>
<string name="vram_usage_conservative">Konservativ</string>
<string name="vram_usage_aggressive">Aggressiv</string>
@@ -524,8 +524,6 @@
<string name="renderer_reactive_flushing_description">Mejora la precisión de renderizado en algunos juegos, pero reduce el rendimiento.</string>
<string name="enable_buffer_history">Activar el historial del búfer</string>
<string name="enable_buffer_history_description">Permite el acceso al estado del búfer anterior. Esta opción puede mejorar la calidad de renderizado y la consistencia en el rendimiento de algunos juegos.</string>
<string name="enable_gpu_buffer_readback">Activar la lectura del buffer de la GPU</string>
<string name="enable_gpu_buffer_readback_description">Conserva los datos del búfer modificados por la GPU leyéndolos antes de subirlos.\nAlgunos juegos requieren esto para renderizar correctamente ciertos efectos.\nPuede causar problemas si el hardware no puede soportar la carga de trabajo adicional.</string>
<string name="use_optimized_vertex_buffers">Búferes de vértices optimizados</string>
<string name="use_optimized_vertex_buffers_description">Permite la optimización del enlace del búfer de vértices para un mejor rendimiento. Requiere controladores Mesa 26.0+ Turnip/ controladores QCOM. Fallará con controladores Turnip más antiguos (versión 25.3 o inferior).</string>
@@ -1038,7 +1036,6 @@
<string name="gpu_fence_behavior_immediate">Inmediato</string>
<string name="gpu_fence_behavior_balanced">Equilibrado</string>
<string name="gpu_fence_behavior_accurate">Preciso</string>
<string name="gpu_fence_behavior_strict">Estricto</string>
<string name="vram_usage_conservative">Conservador</string>
<string name="vram_usage_aggressive">Agresivo</string>
@@ -569,8 +569,6 @@
<string name="renderer_reactive_flushing_description">Повышение точности рендеринга в некоторых играх за счет снижения производительности.</string>
<string name="enable_buffer_history">Включить историю буфера</string>
<string name="enable_buffer_history_description">Позволяет обращаться к предыдущим состояниям буфера. Эта опция может повысить качество рендеринга и стабильность производительности в некоторых играх.</string>
<string name="enable_gpu_buffer_readback">Включить обратное чтение буфера ГПУ</string>
<string name="enable_gpu_buffer_readback_description">Сохраняет измененные ГПУ данные буфера путем чтения их обратно перед выгрузками. Некоторые игры требуют этого, чтобы рендерить определенные эффекты правильно. Может вызывать проблемы если оборудование не может обработать дополнительную рабочую нагрузку.</string>
<string name="use_optimized_vertex_buffers">Оптимизированные вершинные буферы</string>
<string name="use_optimized_vertex_buffers_description">Включает оптимизированную привязку вершинного буфера для повышения производительности. Требует Mesa Turnip 26.0+ / QCOM. Приводит к вылету на старых версиях драйверов Turnip (25.3 и ниже).</string>
@@ -1088,7 +1086,6 @@
<string name="gpu_fence_behavior_immediate">Мгновенный</string>
<string name="gpu_fence_behavior_balanced">Сбалансированный</string>
<string name="gpu_fence_behavior_accurate">Точный</string>
<string name="gpu_fence_behavior_strict">Строгий</string>
<string name="vram_usage_conservative">Консервативный</string>
<string name="vram_usage_aggressive">Агрессивный</string>
@@ -570,8 +570,6 @@
<string name="renderer_reactive_flushing_description">通过牺牲性能来提升某些游戏的渲染精度。</string>
<string name="enable_buffer_history">启用缓冲区历史</string>
<string name="enable_buffer_history_description">启用对先前缓冲区状态的访问。此选项可在某些游戏中提升渲染质量并保持性能的一致性。</string>
<string name="enable_gpu_buffer_readback">启用 GPU 缓冲区回读</string>
<string name="enable_gpu_buffer_readback_description">在上传前回读经由 GPU 修改过的缓冲区数据,以将其保留。一些游戏会用到这项设定以正确渲染某些效果。如果硬件无法处理额外的工作负载,则可能会导致问题。</string>
<string name="use_optimized_vertex_buffers">优化顶点缓冲区</string>
<string name="use_optimized_vertex_buffers_description">启用经过优化的顶点缓冲区绑定以提升性能。需要 Mesa 26.0 及以上版本的 Turnip 或 QCOM 驱动程序。若使用较旧版本的 Turnip 驱动 (25.3 及以下版本) 则会导致崩溃。</string>
@@ -1089,7 +1087,6 @@
<string name="gpu_fence_behavior_immediate">即时</string>
<string name="gpu_fence_behavior_balanced">均衡</string>
<string name="gpu_fence_behavior_accurate">精确</string>
<string name="gpu_fence_behavior_strict">严格</string>
<string name="vram_usage_conservative">保守式</string>
<string name="vram_usage_aggressive">主动式</string>
@@ -561,8 +561,6 @@
<string name="renderer_reactive_flushing_description">犧牲效能,以改善部分遊戲的轉譯準確度</string>
<string name="enable_buffer_history">啟用緩衝區歷史</string>
<string name="enable_buffer_history_description">允許存取先前的緩衝區狀態。此選項可能會改善部分遊戲的渲染品質與效能穩定性</string>
<string name="enable_gpu_buffer_readback">啟用 GPU 緩衝區讀回</string>
<string name="enable_gpu_buffer_readback_description">透過在上傳之前先將 GPU 修改過的緩衝區資料讀回來保存資料,部分遊戲需要啟用此功能才能正常渲染遊戲特效。如果硬體無法負荷可能會導致錯誤</string>
<string name="use_optimized_vertex_buffers">最佳化頂點緩衝區</string>
<string name="use_optimized_vertex_buffers_description">啟用最佳化的頂點緩衝區綁定。需要安裝 Mesa 26.0+ Turnip drivers/Qualcomm drivers。使用舊版 Turnip drivers 會導致當機 (25.3版和更低的版本)</string>
@@ -558,7 +558,6 @@
<item>@string/gpu_fence_behavior_immediate</item>
<item>@string/gpu_fence_behavior_balanced</item>
<item>@string/gpu_fence_behavior_accurate</item>
<item>@string/gpu_fence_behavior_strict</item>
</string-array>
<integer-array name="gpuFenceBehaviorValues">
<item>0</item>
@@ -586,8 +586,6 @@
<string name="renderer_reactive_flushing_description">Improves rendering accuracy in some games at the cost of performance.</string>
<string name="enable_buffer_history">Enable buffer history</string>
<string name="enable_buffer_history_description">Enables access to previous buffer states. This option may improve rendering quality and performance consistency in some games.</string>
<string name="enable_gpu_buffer_readback">Enable GPU Buffer Readback</string>
<string name="enable_gpu_buffer_readback_description">Preserves GPU-modified buffer data by reading it back before uploads. Some games require this to render certain effects properly. May cause issues if the hardware cannot handle the additional workload.</string>
<string name="use_optimized_vertex_buffers">Optimized Vertex Buffers</string>
<string name="use_optimized_vertex_buffers_description">Enables optimized vertex buffer binding for improved performance. Requires Mesa 26.0+ Turnip drivers/ QCOM drivers. Will crash on older Turnip drivers (25.3 and below).</string>
@@ -1138,7 +1136,6 @@
<string name="gpu_fence_behavior_immediate">Immediate</string>
<string name="gpu_fence_behavior_balanced">Balanced</string>
<string name="gpu_fence_behavior_accurate">Accurate</string>
<string name="gpu_fence_behavior_strict">Strict</string>
<!-- ASTC Decoding Method Choices -->
<string name="accelerate_astc_cpu" translatable="false">CPU</string>
-4
View File
@@ -178,10 +178,6 @@ bool IsGPUFenceBehaviorAccurate() {
return values.gpu_fence_behavior.GetValue() == GpuFenceBehavior::Accurate;
}
bool IsGPUFenceBehaviorStrict() {
return values.gpu_fence_behavior.GetValue() == GpuFenceBehavior::Strict;
}
bool IsFastmemEnabled() {
if (values.cpu_accuracy.GetValue() == Settings::CpuAccuracy::Debugging)
return bool(values.cpuopt_fastmem);
+1 -12
View File
@@ -525,7 +525,7 @@ struct Values {
SwitchableSetting<GpuFenceBehavior, true> gpu_fence_behavior{linkage,
GpuFenceBehavior::Default,
GpuFenceBehavior::Default,
GpuFenceBehavior::Strict,
GpuFenceBehavior::Accurate,
"gpu_fence_behavior",
Category::RendererAdvanced,
Specialization::Default,
@@ -653,16 +653,6 @@ struct Values {
SwitchableSetting<bool> emulate_bgr565{linkage, false, "emulate_bgr565",
Category::RendererHacks};
SwitchableSetting<bool> rescale_hack{linkage, false, "rescale_hack",
Category::RendererHacks};
SwitchableSetting<bool> enable_gpu_buffer_readback{linkage,
false,
"enable_gpu_buffer_readback",
Category::RendererAdvanced,
Specialization::Default,
true,
true};
SwitchableSetting<bool> use_asynchronous_shaders{linkage, false, "use_asynchronous_shaders",
Category::RendererHacks};
@@ -979,7 +969,6 @@ bool IsDMALevelSafe();
bool IsGPUFenceBehaviorDefault();
bool IsGPUFenceBehaviorBalanced();
bool IsGPUFenceBehaviorAccurate();
bool IsGPUFenceBehaviorStrict();
bool IsFastmemEnabled();
void SetNceEnabled(bool is_64bit);
+2 -2
View File
@@ -137,7 +137,7 @@ ENUM(VramUsageMode, Conservative, Aggressive);
ENUM(RendererBackend, OpenGL_GLSL, Vulkan, Null, OpenGL_GLASM, OpenGL_SPIRV);
ENUM(GpuAccuracy, Low, High);
ENUM(DmaAccuracy, Default, Unsafe, Safe);
ENUM(GpuFenceBehavior, Default, Immediate, Balanced, Accurate, Strict);
ENUM(GpuFenceBehavior, Default, Immediate, Balanced, Accurate);
ENUM(CpuBackend, Dynarmic, Nce);
ENUM(CpuAccuracy, Auto, Accurate, Unsafe, Paranoid, Debugging);
ENUM(CpuClock, Normal, Boost, Overclock)
@@ -150,7 +150,7 @@ ENUM(ScalingFilter, NearestNeighbor, Bilinear, Bicubic, Gaussian, Lanczos, Scale
ENUM(AntiAliasing, None, Fxaa, Smaa);
ENUM(AspectRatio, R16_9, R4_3, R21_9, R16_10, Stretch);
ENUM(ConsoleMode, Handheld, Docked);
ENUM(AppletMode, HLE, LLE);
ENUM(AppletMode, HLE, LLE, Disabled);
ENUM(SpirvOptimizeMode, Never, OnLoad, Always);
ENUM(GpuClock, Normal, Boost, Overclock)
ENUM(GpuUnswizzleSize, VerySmall, Small, Normal, Large, VeryLarge)
+31 -15
View File
@@ -6,6 +6,7 @@
// SPDX-License-Identifier: GPL-2.0-or-later
#ifdef _WIN32
#include <algorithm>
#include <windows.h>
#include <mutex>
#else
@@ -20,32 +21,36 @@ namespace Common {
#ifdef _WIN32
static std::vector<std::pair<u64, u64>> vector_regions {};
static std::mutex vector_regions_mutex {};
// Workaround for handling non-commited memory accessed by Dynarmic; usually result of an error
static LONG WINAPI FakePageFaultHandler(PEXCEPTION_POINTERS info) {
DWORD code = info->ExceptionRecord->ExceptionCode;
u64 exception_addr = reinterpret_cast<u64>(info->ExceptionRecord->ExceptionAddress);
u64 exception_addr = reinterpret_cast<u64>(info->ExceptionRecord->ExceptionInformation[1]);
if (code != EXCEPTION_ACCESS_VIOLATION) {
if (code != EXCEPTION_ACCESS_VIOLATION || info->ExceptionRecord->ExceptionInformation[0] == 1) {
// Not our problem
return EXCEPTION_CONTINUE_SEARCH;
}
u64 addr = 0, addr2 = 0;
for (auto region: vector_regions) {
auto addr_shifted = exception_addr >> HostPageBits;
if (region.first <= addr_shifted && addr_shifted <= region.second) {
addr = addr_shifted;
}
{
std::lock_guard lock(vector_regions_mutex);
for (auto region: vector_regions) {
auto addr_shifted = exception_addr >> HostPageBits;
if (region.first <= addr_shifted && addr_shifted <= region.second) {
addr = addr_shifted;
}
// Page-boundary accesses
if (auto addr_ = (exception_addr + 0x40) >> HostPageBits; addr_ != addr_shifted && region.first <= addr_ && addr_ <= region.second) {
addr2 = addr_;
}
// Page-boundary accesses
if (auto addr_ = (exception_addr + 0x40) >> HostPageBits; addr_ != addr_shifted && region.first <= addr_ && addr_ <= region.second) {
addr2 = addr_;
}
if (addr != 0 || addr2 != 0) {
break;
if (addr != 0 || addr2 != 0) {
break;
}
}
}
@@ -77,8 +82,16 @@ bool CommitVectorPage(uintptr_t addr, bool write) noexcept {
auto res = VirtualQuery(reinterpret_cast<void*>(addr), &info, sizeof(info));
if (res == 0) {
LOG_CRITICAL(HW_Memory, "Failed to query large buffer region at {:#x} with error {}, will try committing anyway", addr, GetLastError());
} else if (info.State == MEM_COMMIT) {
DWORD old_protect {};
auto perm = write ? PAGE_READWRITE : PAGE_READONLY;
if (!VirtualProtect(reinterpret_cast<void*>(addr), HostPageSize, perm, &old_protect)) {
LOG_ERROR(HW_Memory, "Failed to change permissions of large buffer region at {:#x}, error {}", addr, GetLastError());
return false;
}
return true;
} else if (info.State != MEM_RESERVE) {
LOG_ERROR(HW_Memory, "Tried to commit an unreserved large buffer region at {:#x} that is not mapped or is already committed (state {:#x})", addr, info.State);
LOG_ERROR(HW_Memory, "Tried to commit an unreserved large buffer region at {:#x} that is not mapped (state {:#x})", addr, info.State);
return false;
}
@@ -123,7 +136,8 @@ void* AllocateMemoryPages(std::size_t size) noexcept {
void* base = VirtualAlloc(nullptr, size, MEM_RESERVE, PAGE_READWRITE);
if (base != nullptr) {
vector_regions.emplace_back(reinterpret_cast<u64>(base), reinterpret_cast<u64>(base) + size);
std::lock_guard lock(vector_regions_mutex);
vector_regions.emplace_back(reinterpret_cast<u64>(base) >> HostPageBits, (reinterpret_cast<u64>(base) + size) >> HostPageBits);
static std::once_flag flag;
std::call_once(flag, []() { AddVectoredExceptionHandler(1, FakePageFaultHandler); });
@@ -149,6 +163,8 @@ void FreeMemoryPages(void* base, [[maybe_unused]] std::size_t size) noexcept {
if (!base)
return;
#ifdef _WIN32
std::lock_guard lock(vector_regions_mutex);
std::erase_if(vector_regions, [base](const auto& r) {return r.first == reinterpret_cast<u64>(base); });
ASSERT(VirtualFree(base, 0, MEM_RELEASE));
#else
ASSERT(munmap(base, size) == 0);
+16 -11
View File
@@ -28,9 +28,9 @@ constexpr u64 HostPageBits = 12;
constexpr u64 HostPageMask = ~(HostPageSize - 1);
bool CommitVectorPage(uintptr_t addr, bool write) noexcept;
#else
const u64 HostPageSize = sysconf(_SC_PAGESIZE);
const u64 HostPageBits = std::countr_zero(HostPageSize);
const u64 HostPageMask = ~(HostPageSize - 1);
inline const u64 HostPageSize = sysconf(_SC_PAGESIZE);
inline const u64 HostPageBits = std::countr_zero(HostPageSize);
inline const u64 HostPageMask = ~(HostPageSize - 1);
#endif
void* AllocateMemoryPages(std::size_t size) noexcept;
@@ -81,8 +81,8 @@ public:
UNREACHABLE_MSG("Out of bounds RW access on SparseLargeVector @ {}", index);
}
if (!IsCommittedPage(index)) {
CommitPage(index);
if (!IsCommittedPage(index) && !CommitPage(index)) {
UNREACHABLE_MSG("Cannot access SparseLargeVector index {} with RW permission", index);
}
return base_ptr[index];
}
@@ -103,9 +103,8 @@ public:
LOG_CRITICAL(Common_Memory, "Out of bounds write on SparseLargeVector @ {}", index);
return;
}
if (!IsCommittedPage(index))
CommitPage(index);
base_ptr[index] = value;
if (IsCommittedPage(index) || CommitPage(index))
base_ptr[index] = value;
}
void ZeroRegion(std::size_t start, std::size_t end_) noexcept {
@@ -177,16 +176,22 @@ private:
return (val >> (page & 63)) & 1;
}
constexpr void CommitPage(std::size_t index) noexcept {
constexpr bool CommitPage(std::size_t index) noexcept {
auto page_index = (index * sizeof(T)) >> HostPageBits;
auto page = reinterpret_cast<uintptr_t>(base_ptr + index) & HostPageMask;
#if defined(_WIN32)
CommitVectorPage(page, true);
if (!CommitVectorPage(page, true)) {
return false;
}
#else
mprotect(reinterpret_cast<void*>(page), HostPageSize, PROT_READ | PROT_WRITE);
if (mprotect(reinterpret_cast<void*>(page), HostPageSize, PROT_READ | PROT_WRITE) != 0) {
LOG_ERROR(Common_Memory, "Failed to commit large buffer region at index {}, error {}", index, strerror(errno));
return false;
}
#endif
committed_pages[page_index >> 6].fetch_or(1ULL << (page_index & 63), std::memory_order_release);
return true;
}
constexpr void DecommitPage(std::size_t index) noexcept {
+30 -2
View File
@@ -16,6 +16,10 @@
#include "core/hle/service/ro/ro_types.h"
#include "core/hle/service/server_manager.h"
#include "core/hle/service/service.h"
#ifdef HAS_NCE
#include "core/arm/nce/patcher.h"
#include "core/hle/kernel/k_shared_memory.h"
#endif
namespace Service::RO {
@@ -386,7 +390,7 @@ public:
R_SUCCEED();
}
Result MapManualLoadModuleMemory(u64* out_address, size_t context_id, u64 nro_address,
Result MapManualLoadModuleMemory(Kernel::KernelCore& kernel, u64* out_address, size_t context_id, u64 nro_address,
u64 nro_size, u64 bss_address, u64 bss_size) {
// Get context.
ProcessContext* context = this->GetContextById(context_id);
@@ -421,7 +425,31 @@ public:
R_TRY(context->ValidateNro(std::addressof(nro_info->module_id), std::addressof(rx_size),
std::addressof(ro_size), std::addressof(rw_size),
nro_info->base_address, nro_size, bss_size));
#ifdef HAS_NCE
if (Settings::IsNceEnabled()) {
auto* process = context->GetProcess();
auto& memory = process->GetMemory();
std::vector<u8> image(total_size);
memory.ReadBlock(nro_info->base_address, image.data(), rx_size);
Kernel::CodeSet::Segment code{.size = static_cast<u32>(rx_size)};
Core::NCE::Patcher patch;
patch.PatchText(image, code);
patch.RelocateAndCopy(nro_info->base_address, code, image, nullptr);
const u64 patch_address = nro_info->base_address + total_size;
const size_t patch_size = patch.GetSectionSize();
constexpr auto permission = Kernel::Svc::MemoryPermission::ReadExecute;
auto* patch_memory = Kernel::KSharedMemory::Create(kernel);
R_TRY(patch_memory->Initialize(kernel, kernel.System().DeviceMemory(), process, permission, permission, patch_size));
std::memcpy(patch_memory->GetPointer(), image.data() + total_size, patch_size);
R_TRY(process->AddSharedMemory(kernel, patch_memory, patch_address, patch_size));
R_TRY(patch_memory->Map(*process, patch_address, patch_size, permission));
memory.WriteBlock(nro_info->base_address, image.data(), rx_size);
}
#endif
// Set NRO perms.
R_TRY(SetNroPerms(context->GetProcess(), nro_info->base_address, rx_size, ro_size,
rw_size + bss_size));
@@ -531,7 +559,7 @@ public:
Result MapManualLoadModuleMemory(Out<u64> out_load_address, ClientProcessId client_pid,
u64 nro_address, u64 nro_size, u64 bss_address, u64 bss_size) {
R_TRY(m_ro->ValidateProcess(m_context_id, *client_pid));
R_RETURN(m_ro->MapManualLoadModuleMemory(out_load_address.Get(), m_context_id, nro_address,
R_RETURN(m_ro->MapManualLoadModuleMemory(system.Kernel(), out_load_address.Get(), m_context_id, nro_address,
nro_size, bss_address, bss_size));
}
+1 -1
View File
@@ -757,7 +757,7 @@ struct Memory::Impl {
};
auto& gpu = system.GPU();
gpu_device_memory->ApplyOpOnPointer(
p, scratch_buffers[core], [&](DAddr address) { gpu.InvalidateRegion(address, size); });
p, scratch_buffers[core], [&](DAddr address) { gpu.InvalidateRegion(address, size, true); });
}
Core::System& system;
+4 -18
View File
@@ -761,9 +761,6 @@ void EmulatedController::StartMotionCalibration() {
}
void EmulatedController::SetButton(const Common::Input::CallbackStatus& callback, std::size_t index, Common::UUID uuid) {
const auto player_index = Service::HID::NpadIdTypeToIndex(npad_id_type);
const auto& player = Settings::values.players.GetValue()[player_index];
if (index >= controller.button_values.size()) {
return;
}
@@ -916,21 +913,10 @@ void EmulatedController::SetButton(const Common::Input::CallbackStatus& callback
break;
}
if (!is_connected) {
if (npad_type == NpadStyleIndex::Handheld) {
if (npad_id_type == NpadIdType::Handheld) {
Connect();
controller_connected[player_index] = true;
}
} else if (npad_type != NpadStyleIndex::Handheld) {
if (npad_id_type == NpadIdType::Player1) {
Connect();
controller_connected[player_index] = true;
} else if (player.connected && !controller_connected[player_index]) {
Connect();
controller_connected[player_index] = true;
}
}
const auto player_index = Service::HID::NpadIdTypeToIndex(npad_id_type);
const auto& player = Settings::values.players.GetValue()[player_index];
if (player.connected) {
Connect();
}
TriggerOnChange(ControllerTriggerType::Button, true);
@@ -22,7 +22,6 @@
#include "common/settings.h"
#include "common/vector_math.h"
#include "hid_core/frontend/motion_input.h"
#include "hid_core/hid_core.h"
#include "hid_core/hid_types.h"
#include "hid_core/irsensor/irs_types.h"
@@ -585,7 +584,6 @@ private:
std::array<VibrationValue, 2> last_vibration_value{DEFAULT_VIBRATION_VALUE,
DEFAULT_VIBRATION_VALUE};
std::array<std::chrono::steady_clock::time_point, 2> last_vibration_timepoint{};
std::array<bool, HIDCore::available_controllers> controller_connected{};
// Atomically synched values
std::atomic<HID::NpadStyleIndex> npad_type{HID::NpadStyleIndex::None};
+2 -13
View File
@@ -226,9 +226,7 @@ std::unique_ptr<TranslationMap> InitializeTranslations(QObject* parent) {
INSERT(Settings, dma_accuracy, tr("DMA Accuracy:"),
tr("Controls the DMA read mode.\nUnsafe is faster, while Safe is more stable and can fix issues in some games.\nDefault follows the GPU Accuracy setting."));
INSERT(Settings, gpu_fence_behavior, tr("GPU Fence Behavior:"),
tr("Controls the GPU fence synchronization behavior.\nImmediate is the fastest option, but can introduce some issues.\nBalanced offers better compatibility and may fix issues in some games.\nAccurate further improves compatibility at the cost of some performance.\nStrict is the slowest option, but can fix issues that require stricter synchronization.\nDefault follows the GPU Accuracy setting."));
INSERT(Settings, enable_gpu_buffer_readback, tr("Enable GPU buffer readback"),
tr("Preserves GPU-modified data by reading it back before uploading.\nSome games require this to render certain effects properly."));
tr("Controls the GPU fence synchronization behavior.\nImmediate is the fastest option, but can introduce some issues.\nBalanced offers better compatibility and may fix issues in some games.\nAccurate further improves compatibility at the cost of some performance.\nDefault follows the GPU Mode setting."));
INSERT(Settings, use_asynchronous_shaders, tr("Enable asynchronous shader compilation"),
tr("May reduce shader stutter."));
INSERT(Settings, gpu_clock, tr("GPU Clocks"),
@@ -273,12 +271,6 @@ std::unique_ptr<TranslationMap> InitializeTranslations(QObject* parent) {
"quality and performance consistency in some games."));
INSERT(Settings, fix_bloom_effects, tr("Fix bloom effects"), tr("Removes bloom in Burnout."));
INSERT(Settings, rescale_hack, tr("Enable Legacy Rescale Pass"),
tr("May fix rescale issues in some games by relying on behavior from the previous "
"implementation.\n"
"Legacy behavior workaround that fixes line artifacts on AMD and Intel GPUs, and "
"grey texture flicker on Nvidia GPUs in Luigis Mansion 3."));
// Renderer (Extensions)
INSERT(Settings, dyna_state, tr("Extended Dynamic State"),
tr("Controls the number of features that can be used in Extended Dynamic State.\n"
@@ -344,9 +336,6 @@ std::unique_ptr<TranslationMap> InitializeTranslations(QObject* parent) {
"it bypasses such prompts and directly exits the emulation."));
INSERT(UISettings, hide_mouse, tr("Hide mouse on inactivity"),
tr("Hides the mouse after 2.5s of inactivity."));
INSERT(UISettings, controller_applet_disabled, tr("Disable controller applet"),
tr("Forcibly disables the use of the controller applet in emulated programs.\n"
"When a program attempts to open the controller applet, it is immediately closed."));
INSERT(UISettings, check_for_updates, tr("Check for updates"),
tr("Whether or not to check for updates upon startup."));
@@ -382,6 +371,7 @@ std::unique_ptr<ComboboxTranslationMap> ComboboxEnumeration(QObject* parent) {
{
PAIR(AppletMode, HLE, tr("Custom frontend")),
PAIR(AppletMode, LLE, tr("Real applet")),
PAIR(AppletMode, Disabled, tr("Disabled")),
}});
translations->insert({Settings::EnumMetadata<Settings::SpirvOptimizeMode>::Index(),
@@ -442,7 +432,6 @@ std::unique_ptr<ComboboxTranslationMap> ComboboxEnumeration(QObject* parent) {
PAIR(GpuFenceBehavior, Immediate, tr("Immediate")),
PAIR(GpuFenceBehavior, Balanced, tr("Balanced")),
PAIR(GpuFenceBehavior, Accurate, tr("Accurate")),
PAIR(GpuFenceBehavior, Strict, tr("Strict")),
}});
translations->insert(
{Settings::EnumMetadata<Settings::CpuAccuracy>::Index(),
-2
View File
@@ -127,8 +127,6 @@ struct Values {
Setting<bool> hide_mouse{
linkage, true, "hideInactiveMouse", Category::UiGeneral, Settings::Specialization::Default,
true, true};
Setting<bool> controller_applet_disabled{linkage, false, "disableControllerApplet",
Category::UiGeneral};
// Set when Vulkan is known to crash the application
bool has_broken_vulkan = false;
@@ -301,7 +301,7 @@ IR::Program TranslateProgram(ObjectPool<IR::Inst>& inst_pool, ObjectPool<IR::Blo
Optimization::GlobalMemoryToStorageBufferPass(program, normalized_host_info);
Optimization::TexturePass(env, program, normalized_host_info);
if (Settings::values.resolution_info.active || Settings::values.rescale_hack.GetValue()) {
if (Settings::values.resolution_info.active) {
Optimization::RescalingPass(program);
}
Optimization::DeadCodeEliminationPass(program);
+4
View File
@@ -1,3 +1,6 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2021 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -28,6 +31,7 @@ void TexturePass(Environment& env, IR::Program& program, const HostTranslateInfo
void LayerPass(IR::Program& program, const HostTranslateInfo& host_info);
void VendorWorkaroundPass(IR::Program& program);
void VerificationPass(const IR::Program& program);
bool HasBrokenPattern(const IR::Program& program);
// Dual Vertex
void VertexATransformPass(IR::Program& program);
@@ -67,18 +67,11 @@ void VisitMark(IR::Block& block, IR::Inst& inst) {
if (must_patch_outside) {
const auto it{IR::Block::InstructionList::s_iterator_to(inst)};
IR::IREmitter ir{block, it};
if (Settings::values.rescale_hack.GetValue()) {
const IR::F32 new_inst{&*block.PrependNewInst(it, inst)};
const IR::F32 up_factor{ir.FPRecip(ir.ResolutionDownFactor())};
const IR::Value converted{ir.FPMul(new_inst, up_factor)};
inst.ReplaceUsesWith(converted);
} else {
IR::Inst* const new_inst{&*block.PrependNewInst(it, inst)};
const IR::F32 new_bitcast{ir.ConvertUToF(32, 32, IR::Value{new_inst})};
const IR::F32 up_factor{ir.FPRecip(ir.ResolutionDownFactor())};
const IR::Value converted{ir.FPMul(new_bitcast, up_factor)};
inst.ReplaceUsesWith(converted);
}
IR::Inst* const new_inst{&*block.PrependNewInst(it, inst)};
const IR::F32 new_bitcast{ir.ConvertUToF(32, 32, IR::Value{new_inst})};
const IR::F32 up_factor{ir.FPRecip(ir.ResolutionDownFactor())};
const IR::Value converted{ir.FPMul(new_bitcast, up_factor)};
inst.ReplaceUsesWith(converted);
}
break;
}
@@ -1,3 +1,6 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -76,4 +79,44 @@ void VendorWorkaroundPass(IR::Program& program) {
}
}
bool HasBrokenPattern(const IR::Program& program) {
if (program.stage != Stage::Fragment) {
return false;
}
for (const IR::Block* block : program.post_order_blocks) {
for (const IR::Inst& inst : block->Instructions()) {
switch (inst.GetOpcode()) {
case IR::Opcode::ShuffleIndex:
case IR::Opcode::ShuffleUp:
case IR::Opcode::ShuffleDown:
case IR::Opcode::ShuffleButterfly:
break;
default:
continue;
}
const IR::Value shuffle_arg{inst.Arg(0)};
if (shuffle_arg.IsImmediate()) {
continue;
}
const IR::Inst* bitcast{shuffle_arg.InstRecursive()};
if (!bitcast || bitcast->GetOpcode() != IR::Opcode::BitCastU32F32) {
continue;
}
const IR::Value bitcast_arg{bitcast->Arg(0)};
if (bitcast_arg.IsImmediate()) {
continue;
}
const IR::Inst* attribute{bitcast_arg.InstRecursive()};
if (!attribute || attribute->GetOpcode() != IR::Opcode::GetAttribute) {
continue;
}
const IR::Attribute attr{attribute->Arg(0).Attribute()};
if (attr == IR::Attribute::PositionX || attr == IR::Attribute::PositionY) {
return true;
}
}
}
return false;
}
} // namespace Shader::Optimization
+18 -21
View File
@@ -249,6 +249,10 @@ bool BufferCache<P>::DMACopy(GPUVAddr src_address, GPUVAddr dest_address, u64 am
runtime.CopyBuffer(dest_buffer, src_buffer, copies, true);
if (has_new_downloads) {
memory_tracker.MarkRegionAsGpuModified(*cpu_dest_address, amount);
const bool should_sync = Settings::IsGPUFenceBehaviorBalanced() || Settings::IsGPUFenceBehaviorAccurate();
if (should_sync) {
runtime.Finish();
}
}
Tegra::Memory::DeviceGuestMemoryScoped<u8, Tegra::Memory::GuestMemoryFlags::UnsafeReadWrite>
@@ -1230,7 +1234,7 @@ void BufferCache<P>::BindHostComputeStorageBuffers() {
buffer.MarkUsage(offset, size);
if (is_written) {
MarkWrittenBuffer(binding.buffer_id, binding.device_addr, size);
MarkWrittenBuffer(binding.buffer_id, binding.device_addr, size, true);
}
if constexpr (NEEDS_BIND_STORAGE_INDEX) {
@@ -1516,10 +1520,12 @@ void BufferCache<P>::UpdateComputeTextureBuffers() {
}
template <class P>
void BufferCache<P>::MarkWrittenBuffer(BufferId buffer_id, DAddr device_addr, u32 size) {
void BufferCache<P>::MarkWrittenBuffer(BufferId buffer_id, DAddr device_addr, u32 size, bool needs_sync) {
if constexpr (!IS_OPENGL) {
Buffer& buffer = slot_buffers[buffer_id];
buffer.setWriteTick(runtime.CurrentTick());
if (needs_sync) {
Buffer& buffer = slot_buffers[buffer_id];
buffer.setWriteTick(runtime.CurrentTick());
}
}
memory_tracker.MarkRegionAsGpuModified(device_addr, size);
gpu_modified_ranges.Add(device_addr, size);
@@ -1535,8 +1541,11 @@ BufferId BufferCache<P>::FindBuffer(DAddr device_addr, u32 size, bool sparse_com
const BufferId buffer_id = page_table[page];
if (buffer_id) {
Buffer& buffer = slot_buffers[buffer_id];
WaitForGpuFenceIfNeeded(buffer);
if (buffer.IsInBounds(device_addr, size)) {
const bool should_sync = Settings::IsGPUFenceBehaviorAccurate();
if (should_sync) {
SynchronizeBufferWrites(buffer);
}
bool usable = true;
if constexpr (requires { buffer.IsSparseCompatible(); }) {
if (sparse_compatible && !buffer.IsSparseCompatible()) {
@@ -1552,17 +1561,11 @@ BufferId BufferCache<P>::FindBuffer(DAddr device_addr, u32 size, bool sparse_com
}
template <class P>
void BufferCache<P>::WaitForGpuFenceIfNeeded(Buffer& buffer) {
void BufferCache<P>::SynchronizeBufferWrites(Buffer& buffer) {
if constexpr (!IS_OPENGL) {
const bool gpu_fence_accurate = Settings::IsGPUFenceBehaviorAccurate();
const bool gpu_fence_strict = Settings::IsGPUFenceBehaviorStrict();
if (gpu_fence_accurate || gpu_fence_strict) {
const u64 gpu_tick_delay = gpu_fence_strict ? 0 : 3;
const u64 buffer_tick = buffer.getWriteTick();
const u64 gpu_tick = runtime.KnownGpuTick();
if (buffer_tick > gpu_tick + gpu_tick_delay) {
runtime.Wait(buffer_tick);
}
const u64 buffer_tick = buffer.getWriteTick();
if (!runtime.IsFree(buffer_tick)) {
runtime.Wait(buffer_tick);
}
}
}
@@ -1795,9 +1798,6 @@ void BufferCache<P>::ImmediateUploadMemory([[maybe_unused]] Buffer& buffer,
if (immediate_buffer.empty()) {
immediate_buffer = ImmediateBuffer(largest_copy);
}
if (Settings::values.enable_gpu_buffer_readback.GetValue()) {
DownloadBufferMemory(buffer, device_addr, copy.size);
}
device_memory.ReadBlockUnsafe(device_addr, immediate_buffer.data(), copy.size);
upload_span = immediate_buffer.subspan(0, copy.size);
}
@@ -1816,9 +1816,6 @@ void BufferCache<P>::MappedUploadMemory([[maybe_unused]] Buffer& buffer,
for (BufferCopy& copy : copies) {
u8* const src_pointer = staging_pointer.data() + copy.src_offset;
const DAddr device_addr = buffer.CpuAddr() + copy.dst_offset;
if (Settings::values.enable_gpu_buffer_readback.GetValue()) {
DownloadBufferMemory(buffer, device_addr, copy.size);
}
device_memory.ReadBlockUnsafe(device_addr, src_pointer, copy.size);
// Apply the staging offset
copy.src_offset += upload_staging.offset;
@@ -430,11 +430,11 @@ private:
void UpdateComputeTextureBuffers();
void MarkWrittenBuffer(BufferId buffer_id, DAddr device_addr, u32 size);
void MarkWrittenBuffer(BufferId buffer_id, DAddr device_addr, u32 size, bool needs_sync = false);
[[nodiscard]] BufferId FindBuffer(DAddr device_addr, u32 size, bool sparse_compatible);
void WaitForGpuFenceIfNeeded(Buffer& buffer);
void SynchronizeBufferWrites(Buffer& buffer);
[[nodiscard]] OverlapResult ResolveOverlaps(DAddr device_addr, u32 wanted_size);
+1 -1
View File
@@ -72,7 +72,7 @@ public:
}
void SignalFence(std::function<void()>&& func) {
const bool delay_fence = Settings::IsGPUFenceBehaviorDefault() ? Settings::IsGPULevelHigh() : Settings::IsGPUFenceBehaviorBalanced() || Settings::IsGPUFenceBehaviorAccurate() || Settings::IsGPUFenceBehaviorStrict();
const bool delay_fence = Settings::IsGPUFenceBehaviorDefault() ? Settings::IsGPULevelHigh() : Settings::IsGPUFenceBehaviorBalanced() || Settings::IsGPUFenceBehaviorAccurate();
const bool should_flush = ShouldFlush();
if constexpr (!can_async_check) {
TryReleasePendingFences<false>();
+13 -3
View File
@@ -226,7 +226,17 @@ struct GPU::Impl {
}
/// Notify rasterizer that any caches of the specified region should be invalidated
void InvalidateRegion(DAddr addr, u64 size) {
void InvalidateRegion(DAddr addr, u64 size, bool preserve_gpu_writes) {
VideoCore::RasterizerInterface* rasterizer = renderer->ReadRasterizer();
if (preserve_gpu_writes && rasterizer->MustFlushRegion(addr, size, VideoCommon::CacheType::BufferCache)) {
const u64 fence = RequestSyncOperation([rasterizer, addr, size] {
rasterizer->FlushRegion(addr, size, VideoCommon::CacheType::BufferCache);
rasterizer->OnCacheInvalidation(addr, size);
});
gpu_thread.TickGPU(is_async);
WaitForSyncOperation(fence);
return;
}
gpu_thread.InvalidateRegion(addr, size);
}
@@ -518,8 +528,8 @@ void GPU::FlushRegion(DAddr addr, u64 size) {
impl->FlushRegion(addr, size);
}
void GPU::InvalidateRegion(DAddr addr, u64 size) {
impl->InvalidateRegion(addr, size);
void GPU::InvalidateRegion(DAddr addr, u64 size, bool preserve_gpu_writes) {
impl->InvalidateRegion(addr, size, preserve_gpu_writes);
}
bool GPU::OnCPUWrite(DAddr addr, u64 size) {
+1 -1
View File
@@ -248,7 +248,7 @@ public:
void FlushRegion(DAddr addr, u64 size);
/// Notify rasterizer that any caches of the specified region should be invalidated
void InvalidateRegion(DAddr addr, u64 size);
void InvalidateRegion(DAddr addr, u64 size, bool preserve_gpu_writes = false);
/// Notify rasterizer that CPU is trying to write this area. It returns true if the area is
/// sensible, false otherwise, addr and size must be a valid combination
+1 -1
View File
@@ -260,7 +260,7 @@ void QueryCacheBase<Traits>::CounterReport(GPUVAddr addr, QueryType counter_type
};
u8* pointer = impl->device_memory.template GetPointer<u8>(cpu_addr);
u8* pointer_timestamp = impl->device_memory.template GetPointer<u8>(cpu_addr + 8);
bool is_synced = (Settings::IsGPUFenceBehaviorDefault() ? !Settings::IsGPULevelHigh() : !Settings::IsGPUFenceBehaviorBalanced() && !Settings::IsGPUFenceBehaviorAccurate() && !Settings::IsGPUFenceBehaviorStrict()) && is_fence;
bool is_synced = (Settings::IsGPUFenceBehaviorDefault() ? !Settings::IsGPULevelHigh() : !Settings::IsGPUFenceBehaviorBalanced() && !Settings::IsGPUFenceBehaviorAccurate()) && is_fence;
std::function<void()> operation([this, is_synced, streamer, query_base = query, query_location,
pointer, pointer_timestamp] {
if (True(query_base->flags & QueryFlagBits::IsInvalidated)) {
@@ -419,15 +419,15 @@ void BufferCacheRuntime::TickFrame(Common::SlotVector<Buffer>& slot_buffers) noe
}
u64 BufferCacheRuntime::CurrentTick() {
return scheduler.GetMasterSemaphore().CurrentTick();
return scheduler.CurrentTick();
}
u64 BufferCacheRuntime::KnownGpuTick() {
return scheduler.GetMasterSemaphore().KnownGpuTick();
bool BufferCacheRuntime::IsFree(u64 tick) {
return scheduler.IsFree(tick);
}
void BufferCacheRuntime::Wait(u64 buffer_tick) {
scheduler.Wait(buffer_tick);
void BufferCacheRuntime::Wait(u64 tick) {
scheduler.Wait(tick);
}
void BufferCacheRuntime::Finish() {
@@ -112,9 +112,9 @@ public:
u64 CurrentTick();
u64 KnownGpuTick();
bool IsFree(u64 tick);
void Wait(u64 buffer_tick);
void Wait(u64 tick);
void Finish();
@@ -23,6 +23,7 @@
#include "shader_recompiler/environment.h"
#include "shader_recompiler/frontend/maxwell/control_flow.h"
#include "shader_recompiler/frontend/maxwell/translate_program.h"
#include "shader_recompiler/ir_opt/passes.h"
#include "shader_recompiler/program_header.h"
#include "video_core/engines/kepler_compute.h"
#include "video_core/engines/maxwell_3d.h"
@@ -826,6 +827,11 @@ std::unique_ptr<GraphicsPipeline> PipelineCache::CreateGraphicsPipeline(
programs[index] = MergeDualVertexPrograms(program_va, program_vb, env);
}
if (device.GetDriverID() != VK_DRIVER_ID_NVIDIA_PROPRIETARY && Shader::Optimization::HasBrokenPattern(programs[index])) {
LOG_WARNING(Render_Vulkan, "0x{:016x}: Skipping shader with broken pattern", key.unique_hashes[index]);
return nullptr;
}
if (Settings::values.dump_guest_shaders) {
env.Dump(hash, key.unique_hashes[index]);
}
+3 -1
View File
@@ -1,3 +1,6 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: 2024 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -8,7 +11,6 @@
class QCheckBox;
class QLineEdit;
class QComboBox;
class QDateTimeEdit;
namespace Core {
class System;
+8 -4
View File
@@ -140,7 +140,12 @@ QWidget* Widget::CreateCombobox(std::function<std::string()>& serializer,
if (combobox_enumerations.contains(type)) {
enumeration = &combobox_enumerations.at(type);
for (const auto& [id, name] : *enumeration) {
combobox->addItem(name);
if (type == Settings::EnumMetadata<Settings::AppletMode>::Index() &&
id == static_cast<u32>(Settings::AppletMode::Disabled) &&
setting.Id() != Settings::values.controller_applet_mode.Id()) {
continue;
}
combobox->addItem(name, id);
}
} else {
return combobox;
@@ -158,9 +163,8 @@ QWidget* Widget::CreateCombobox(std::function<std::string()>& serializer,
const u32 setting_value = std::strtoul(setting.ToString().c_str(), nullptr, 0);
combobox->setCurrentIndex(find_index(setting_value));
serializer = [this, enumeration]() {
int current = combobox->currentIndex();
return std::to_string(enumeration->at(current).first);
serializer = [this]() {
return std::to_string(combobox->currentData().toUInt());
};
restore_func = [this, find_index]() {
+15 -11
View File
@@ -12,25 +12,29 @@ namespace Yuzu {
inline bool ContainsAllWords(const QString& haystack, const QString& userinput) {
const QStringList userinput_split = userinput.split(QLatin1Char{' '}, Qt::SkipEmptyParts);
return std::all_of(userinput_split.begin(), userinput_split.end(),
[&haystack](const QString& s) { return haystack.contains(s); });
return std::all_of(userinput_split.begin(), userinput_split.end(), [&haystack](const QString& s) {
return haystack.contains(s);
});
}
inline bool FilterMatches(const QString& filter, const QStandardItem* item) {
if (filter.isEmpty())
return true;
auto const sluggify = [](const QString& s) {
QString o(s.normalized(QString::NormalizationForm_D).toLower());
return o.replace(QRegularExpression(QStringLiteral("[^a-z\\s]")), QStringLiteral(""));
};
const auto norm_filter = sluggify(filter);
const auto program_id = item->data(GameListItemPath::ProgramIdRole).toULongLong();
const QString file_path = item->data(GameListItemPath::FullPathRole).toString().toLower();
const QString file_title = item->data(GameListItemPath::TitleRole).toString().toLower();
const QString file_path = sluggify(item->data(GameListItemPath::FullPathRole).toString());
const QString file_title = sluggify(item->data(GameListItemPath::TitleRole).toString());
const QString file_program_id = QStringLiteral("%1").arg(program_id, 16, 16, QLatin1Char{'0'});
const QString file_name =
file_path.mid(file_path.lastIndexOf(QLatin1Char{'/'}) + 1) + QLatin1Char{' '} + file_title;
return Yuzu::ContainsAllWords(file_name, filter) ||
(file_program_id.size() == 16 && file_program_id.contains(filter));
const QString file_name = sluggify(file_path.mid(file_path.lastIndexOf(QLatin1Char{'/'}) + 1) + QLatin1Char{' '} + file_title);
return (ContainsAllWords(file_name, norm_filter)
|| ContainsAllWords(file_title, norm_filter))
|| (file_program_id.size() == 16 && file_program_id.contains(norm_filter));
}
} // namespace Yuzu
+1 -1
View File
@@ -1786,7 +1786,7 @@ bool MainWindow::LoadROM(const QString& filename, Service::AM::FrontendAppletPar
QtCommon::system->SetFrontendAppletSet({
std::make_unique<QtAmiiboSettings>(*this), // Amiibo Settings
(UISettings::values.controller_applet_disabled.GetValue() == true)
(Settings::values.controller_applet_mode.GetValue() == Settings::AppletMode::Disabled)
? nullptr
: std::make_unique<QtControllerSelector>(*this), // Controller Selector
std::make_unique<QtErrorDisplay>(*this), // Error Display