Compare commits

..

4 Commits

Author SHA1 Message Date
lizzie 3e54fb4c33 fix le crash 2026-04-28 00:52:15 +00:00
lizzie f8689fb451 fx 2026-04-28 00:52:15 +00:00
lizzie 00d745e4e5 coalesce 2026-04-28 00:52:15 +00:00
lizzie f0685cc0c9 [video_core] raise bindless texture array length to 1024
notably fixes TL:LTD grass
Authored-by: Mythrax <mythrax@mytrax-rs.org>

Signed-off-by: lizzie <lizzie@eden-emu.dev>
2026-04-28 00:52:15 +00:00
95 changed files with 9885 additions and 9732 deletions
+9 -4
View File
@@ -268,6 +268,11 @@ if (NOT EXISTS ${PROJECT_BINARY_DIR}/${compat_json})
file(WRITE ${PROJECT_BINARY_DIR}/${compat_json} "")
endif()
if (YUZU_LEGACY)
message(WARNING "Making legacy build. Performance may suffer.")
add_compile_definitions(YUZU_LEGACY)
endif()
if (ARCHITECTURE_arm64 AND (ANDROID OR PLATFORM_LINUX))
set(HAS_NCE 1)
add_compile_definitions(HAS_NCE=1)
@@ -419,10 +424,10 @@ if (zstd_ADDED)
add_library(zstd::libzstd ALIAS libzstd_static)
endif()
# nlohmann
AddJsonPackage(nlohmann)
if (NOT YUZU_STATIC_ROOM)
# nlohmann
AddJsonPackage(nlohmann)
# zlib
AddJsonPackage(zlib)
@@ -480,7 +485,7 @@ endfunction()
# =============================================
if (APPLE)
foreach(fw Carbon Metal Cocoa IOKit CoreVideo CoreMedia Security)
foreach(fw Carbon Metal Cocoa IOKit CoreVideo CoreMedia)
find_library(${fw}_LIBRARY ${fw} REQUIRED)
list(APPEND PLATFORM_LIBRARIES ${${fw}_LIBRARY})
endforeach()
+2 -4
View File
@@ -33,11 +33,9 @@ endif()
set(GIT_DESC ${BUILD_VERSION})
# Generate cpp with Git revision from template
# Also if this is a CI build, add the build name (ie: Nightly, Canary) to the scm_rev file as well
# TODO(crueter): Stable releases feed.
set(BUILD_AUTO_UPDATE_STABLE_REPO "eden-emu/eden")
set(BUILD_AUTO_UPDATE_STABLE_API "git.eden-emu.dev")
set(BUILD_AUTO_UPDATE_STABLE_API_PATH "/api/v1/repos/")
# Auto-updater metadata! Must somewhat mirror GitHub/Forgejo API endpoint
set(BUILD_AUTO_UPDATE_API_PATH "/latest/release.json")
if (NIGHTLY_BUILD)
+310 -294
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+312 -296
View File
File diff suppressed because it is too large Load Diff
+319 -306
View File
File diff suppressed because it is too large Load Diff
+309 -293
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+309 -293
View File
File diff suppressed because it is too large Load Diff
+316 -301
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+309 -293
View File
File diff suppressed because it is too large Load Diff
+309 -293
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+315 -300
View File
File diff suppressed because it is too large Load Diff
+310 -294
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+310 -294
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+306 -293
View File
File diff suppressed because it is too large Load Diff
+308 -294
View File
File diff suppressed because it is too large Load Diff
+308 -293
View File
File diff suppressed because it is too large Load Diff
-10
View File
@@ -25,16 +25,6 @@ if (NIGHTLY_BUILD)
add_compile_definitions(NIGHTLY_BUILD)
endif()
if (YUZU_LEGACY)
message(WARNING "Making legacy build. Performance may suffer.")
add_compile_definitions(YUZU_LEGACY)
endif()
if (GENSHIN_SPOOF)
message(WARNING "Making Genshin spoof build")
add_compile_definitions(GENSHIN_SPOOF)
endif()
# Set compilation flags
if (MSVC AND NOT CXX_CLANG)
set(CMAKE_CONFIGURATION_TYPES Debug Release CACHE STRING "" FORCE)
-7
View File
@@ -88,7 +88,6 @@ android {
"-DBUILD_TESTING=OFF",
"-DYUZU_TESTS=OFF",
"-DDYNARMIC_TESTS=OFF",
"-DENABLE_UPDATE_CHECKER=ON",
*extraCMakeArgs.toTypedArray()
)
)
@@ -193,12 +192,6 @@ android {
manifestPlaceholders += mapOf("appNameBase" to "Eden")
resValue("string", "app_name_suffixed", "Eden")
externalNativeBuild {
cmake {
arguments.add("-DGENSHIN_SPOOF=ON")
}
}
ndk {
abiFilters += listOf("arm64-v8a")
}
@@ -33,18 +33,6 @@ import org.yuzu.yuzu_emu.applets.web.WebBrowser
* with the native side of the Yuzu code.
*/
object NativeLibrary {
data class UpdateResult(
var tag: String = "",
var title: String = "",
var body: String = "",
var url: String = "",
var assets: MutableList<String> = mutableListOf()
) {
fun addAsset(asset: String) {
assets.add(asset)
}
}
@JvmField
var sEmulationActivity = WeakReference<EmulationActivity?>(null)
@@ -252,7 +240,17 @@ object NativeLibrary {
/**
* Checks for available updates.
*/
external fun checkForUpdate(): UpdateResult?
external fun checkForUpdate(): Array<String>?
/**
* Return the URL to the release page
*/
external fun getUpdateUrl(version: String): String
/**
* Return the URL to download the APK for the given version
*/
external fun getUpdateApkUrl(tag: String, artifact: String, packageId: String): String
/**
* Returns whether the update checker is enabled through CMAKE options.
@@ -175,25 +175,25 @@ class MainActivity : AppCompatActivity(), ThemeProvider {
val latestVersion = NativeLibrary.checkForUpdate()
if (latestVersion != null) {
runOnUiThread {
showUpdateDialog(latestVersion)
val tag: String = latestVersion[0]
val name: String = latestVersion[1]
showUpdateDialog(tag, name)
}
}
}.start()
}
// TODO(crueter): body, "View on Forgejo" button
private fun showUpdateDialog(release: NativeLibrary.UpdateResult) {
private fun showUpdateDialog(tag: String, name: String) {
MaterialAlertDialogBuilder(this)
.setTitle(R.string.update_available)
.setMessage(getString(R.string.update_available_description, release.title))
.setMessage(getString(R.string.update_available_description, name))
.setPositiveButton(android.R.string.ok) { _, _ ->
val assets = release.assets
if (assets.isEmpty()) {
openLink(release.url)
} else {
downloadAndInstallUpdate(release)
var artifact = tag
// Nightly builds have a slightly different format
if (NativeLibrary.isNightlyBuild()) {
artifact = tag.substringAfter('.', tag)
}
downloadAndInstallUpdate(tag, artifact)
}
.setNeutralButton(R.string.cancel) { dialog, _ ->
dialog.dismiss()
@@ -206,23 +206,17 @@ class MainActivity : AppCompatActivity(), ThemeProvider {
.show()
}
private fun openLink(link: String) {
val intent = Intent(Intent.ACTION_VIEW, link.toUri())
startActivity(intent)
}
private fun downloadAndInstallUpdate(release: NativeLibrary.UpdateResult) {
private fun downloadAndInstallUpdate(version: String, artifact: String) {
CoroutineScope(Dispatchers.IO).launch {
val packageId = applicationContext.packageName
val asset = release.assets[0]
val artifact = asset.split("/").last()
val apkUrl = NativeLibrary.getUpdateApkUrl(version, artifact, packageId)
val apkFile = File(cacheDir, "update-$artifact.apk")
withContext(Dispatchers.Main) {
showDownloadProgressDialog()
}
val downloader = APKDownloader(asset, apkFile)
val downloader = APKDownloader(apkUrl, apkFile)
downloader.download(
onProgress = { progress ->
runOnUiThread {
@@ -254,7 +248,7 @@ class MainActivity : AppCompatActivity(), ThemeProvider {
} else {
Toast.makeText(
this@MainActivity,
getString(R.string.update_download_failed) + "\n\nURL: $asset",
getString(R.string.update_download_failed) + "\n\nURL: $apkUrl",
Toast.LENGTH_LONG
).show()
}
@@ -283,7 +277,7 @@ class MainActivity : AppCompatActivity(), ThemeProvider {
private fun updateDownloadProgress(progress: Int) {
progressBar?.progress = progress
progressMessage?.text = getString(R.string.percent, progress)
progressMessage?.text = "$progress%"
}
private fun dismissDownloadProgressDialog() {
+58 -58
View File
@@ -1699,76 +1699,76 @@ JNIEXPORT jboolean JNICALL Java_org_yuzu_yuzu_1emu_NativeLibrary_isNightlyBuild(
#ifdef ENABLE_UPDATE_CHECKER
JNIEXPORT jobject JNICALL Java_org_yuzu_yuzu_1emu_NativeLibrary_checkForUpdate(
JNIEXPORT jobjectArray JNICALL Java_org_yuzu_yuzu_1emu_NativeLibrary_checkForUpdate(
JNIEnv* env,
jobject obj) {
std::optional<Common::Net::Release> release = UpdateChecker::GetUpdate();
std::optional<UpdateChecker::Update> release = UpdateChecker::GetUpdate();
if (!release) return nullptr;
const std::string tag = release->tag;
const std::string title = release->title;
const std::string body = release->body;
const std::string url = release->html_url;
const std::string name = release->name;
// Android *should* only ever define a single asset.
// If not, something has gone wrong, but the Kotlin side can handle it.
const auto assets = release->GetPlatformAssets();
jobjectArray result = env->NewObjectArray(2, env->FindClass("java/lang/String"), nullptr);
jclass updateResultClass = env->FindClass("org/yuzu/yuzu_emu/NativeLibrary$UpdateResult");
if (!updateResultClass) {
LOG_ERROR(Frontend, "Could not find UpdateResult class");
return nullptr;
}
jmethodID updateResultCtor = env->GetMethodID(updateResultClass, "<init>", "()V");
if (!updateResultCtor) {
LOG_ERROR(Frontend, "Could not find UpdateResult ctor");
env->DeleteLocalRef(updateResultClass);
return nullptr;
}
jmethodID setTag = env->GetMethodID(updateResultClass, "setTag", "(Ljava/lang/String;)V");
jmethodID setTitle = env->GetMethodID(updateResultClass, "setTitle", "(Ljava/lang/String;)V");
jmethodID setBody = env->GetMethodID(updateResultClass, "setBody", "(Ljava/lang/String;)V");
jmethodID setUrl = env->GetMethodID(updateResultClass, "setUrl", "(Ljava/lang/String;)V");
jmethodID addAsset = env->GetMethodID(updateResultClass, "addAsset", "(Ljava/lang/String;)V");
jobject updateResult = env->NewObject(updateResultClass, updateResultCtor);
LOG_DEBUG(Frontend, "Tag: {}", tag);
LOG_DEBUG(Frontend, "Title: {}", title);
LOG_DEBUG(Frontend, "Body: {}", body);
LOG_DEBUG(Frontend, "Url: {}", url);
const auto jtag = env->NewStringUTF(tag.c_str());
const auto jtitle = env->NewStringUTF(title.c_str());
const auto jbody = env->NewStringUTF(body.c_str());
const auto jurl = env->NewStringUTF(url.c_str());
env->CallVoidMethod(updateResult, setTag, jtag);
env->CallVoidMethod(updateResult, setTitle, jtitle);
env->CallVoidMethod(updateResult, setBody, jbody);
env->CallVoidMethod(updateResult, setUrl, jurl);
// TODO(crueter): Handling for multiple assets?
// Maybe another data class x(
for (const Common::Net::Asset &a : assets) {
const auto jaurl = env->NewStringUTF(a.path.c_str());
env->CallVoidMethod(updateResult, addAsset, jaurl);
env->DeleteLocalRef(jaurl);
}
const jstring jtag = env->NewStringUTF(tag.c_str());
const jstring jname = env->NewStringUTF(name.c_str());
env->SetObjectArrayElement(result, 0, jtag);
env->SetObjectArrayElement(result, 1, jname);
env->DeleteLocalRef(jtag);
env->DeleteLocalRef(jtitle);
env->DeleteLocalRef(jbody);
env->DeleteLocalRef(jurl);
env->DeleteLocalRef(jname);
env->DeleteLocalRef(updateResultClass);
return updateResult;
return result;
}
JNIEXPORT jstring JNICALL Java_org_yuzu_yuzu_1emu_NativeLibrary_getUpdateUrl(
JNIEnv* env,
jobject obj,
jstring version) {
const char* version_str = env->GetStringUTFChars(version, nullptr);
const std::string url = fmt::format("{}/{}",
std::string{Common::g_build_auto_update_api},
version_str);
env->ReleaseStringUTFChars(version, version_str);
return env->NewStringUTF(url.c_str());
}
JNIEXPORT jstring JNICALL Java_org_yuzu_yuzu_1emu_NativeLibrary_getUpdateApkUrl(
JNIEnv* env,
jobject obj,
jstring tag,
jstring artifact,
jstring packageId) {
const char* version_str = env->GetStringUTFChars(tag, nullptr);
const char* artifact_str = env->GetStringUTFChars(artifact, nullptr);
const char* package_id_str = env->GetStringUTFChars(packageId, nullptr);
std::string variant;
std::string package_id(package_id_str);
if (package_id.find("dev.legacy.eden_emulator") != std::string::npos) {
variant = "legacy";
} else if (package_id.find("com.miHoYo.Yuanshen") != std::string::npos) {
variant = "optimized";
} else {
#ifdef ARCHITECTURE_arm64
variant = "standard";
#else
variant = "chromeos";
#endif
}
const std::string apk_filename = fmt::format("Eden-Android-{}-{}.apk", artifact_str, variant);
const std::string url = fmt::format("https://{}/{}/{}",
std::string{Common::g_build_auto_update_api},
version_str, apk_filename);
env->ReleaseStringUTFChars(tag, version_str);
env->ReleaseStringUTFChars(artifact, artifact_str);
env->ReleaseStringUTFChars(packageId, package_id_str);
return env->NewStringUTF(url.c_str());
}
#endif
JNIEXPORT jstring JNICALL Java_org_yuzu_yuzu_1emu_NativeLibrary_getBuildVersion(
@@ -767,7 +767,7 @@
<string name="version">Versión</string>
<string name="copy_details">Copiar detalles</string>
<string name="add_ons">Complementos</string>
<string name="add_ons_description">Alternar mods, actualizaciones y contenido descargable</string>
<string name="add_ons_description">Activa/desactiva mods, actualizaciones y contenidos descargables</string>
<string name="playtime">Tiempo jugado:</string>
<string name="reset_playtime">Borrar tiempo de juego</string>
<string name="reset_playtime_description">Restablecer el tiempo de juego actual a 0 segundos</string>
@@ -1783,6 +1783,5 @@ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
<string name="external_content">External Content</string>
<string name="add_folders">Add Folder</string>
<string name="percent">%1$d%%</string>
</resources>
+2 -3
View File
@@ -147,8 +147,7 @@ add_library(
zstd_compression.h
fs/ryujinx_compat.h fs/ryujinx_compat.cpp
fs/symlink.h fs/symlink.cpp
httplib.h
net/net.h net/net.cpp)
httplib.h)
if(WIN32)
target_sources(common PRIVATE windows/timer_resolution.cpp
@@ -246,7 +245,7 @@ else()
target_link_libraries(common PUBLIC Boost::headers)
endif()
target_link_libraries(common PUBLIC Boost::filesystem Boost::context httplib::httplib nlohmann_json::nlohmann_json)
target_link_libraries(common PUBLIC Boost::filesystem Boost::context httplib::httplib)
if (lz4_ADDED)
target_include_directories(common PRIVATE ${lz4_SOURCE_DIR}/lib)
-287
View File
@@ -1,287 +0,0 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
#include <optional>
#include <boost/algorithm/string/classification.hpp>
#include <boost/algorithm/string/replace.hpp>
#include <boost/algorithm/string/split.hpp>
#include <fmt/format.h>
#include "common/scm_rev.h"
#include "net.h"
#include "common/logging.h"
#include "common/httplib.h"
#ifdef YUZU_BUNDLED_OPENSSL
#include <openssl/cert.h>
#endif
#define QT_TR_NOOP(x) x
namespace Common::Net {
std::vector<Asset> Release::GetPlatformAssets() const {
// TODO(crueter): Need better handling for this as a whole.
#ifdef NIGHTLY_BUILD
std::vector<std::string> result;
boost::algorithm::split(result, tag, boost::is_any_of("."));
if (result.size() != 2)
return {};
const auto ref = result.at(1);
#else
const auto ref = tag;
#endif
std::vector<Asset> found_assets;
// FIXME: This is mildly inefficient.
// Finds assets based on a hierarchy of regex search strings.
const auto find_asset = [&found_assets, ref, this](const std::string& name,
const std::vector<std::string>& suffixes) {
for (const std::string& asset : assets) {
for (const auto& suffix : suffixes) {
if (asset.ends_with(suffix)) {
const std::string_view asset_sv = asset;
const size_t pos = asset_sv.find_last_of('/');
const std::string_view filename =
(pos != std::string_view::npos) ? asset_sv.substr(pos + 1) : asset_sv;
found_assets.emplace_back(Asset{
.name = name,
.url = host,
.path = asset,
.filename = std::string{filename},
});
return;
}
}
}
};
#ifdef _WIN32
#ifdef ARCHITECTURE_x86_64
find_asset("Standard", {"amd64-msvc-standard.exe", "amd64-msvc-standard.zip", "mingw-amd64-gcc-standard.exe", "mingw-amd64-gcc-standard.zip"});
find_asset("PGO", {"mingw-amd64-clang-pgo.exe", "mingw-amd64-clang-pgo.zip"});
#elif defined(ARCHITECTURE_arm64)
find_asset("Standard", {"mingw-arm64-clang-standard.exe", "mingw-arm64-clang-standard.zip"});
find_asset("PGO", {"mingw-arm64-clang-pgo.exe", "mingw-arm64-clang-pgo.zip"});
#endif
#elif defined(__APPLE__)
#ifdef ARCHITECTURE_arm64
find_asset("Standard", {".dmg", ".tar.gz"});
#endif
#elif defined(__ANDROID__)
#ifdef ARCHITECTURE_x86_64
find_asset("Standard", {"chromeos.apk"});
#elif defined(ARCHITECTURE_arm64)
#ifdef YUZU_LEGACY
find_asset("Standard", {"legacy.apk"});
#elif defined(GENSHIN_SPOOF)
find_asset("Standard", {"optimized.apk"});
#else
find_asset("Standard", {"standard.apk"});
#endif
#endif
#endif
return found_assets;
}
static inline u64 ParseIsoTimestamp(const std::string& iso) {
if (iso.empty())
return 0;
std::string buf = iso;
if (buf.back() == 'Z')
buf.pop_back();
std::tm tm{};
std::istringstream ss(buf);
ss >> std::get_time(&tm, "%Y-%m-%dT%H:%M:%S");
if (ss.fail())
return 0;
#ifdef _WIN32
return static_cast<u64>(_mkgmtime(&tm));
#else
return static_cast<u64>(timegm(&tm));
#endif
}
std::optional<Release> Release::FromJson(const nlohmann::json& json, const std::string& host,
const std::string& repo) {
Release rel;
if (!json.is_object())
return std::nullopt;
rel.tag = json.value("tag_name", std::string{});
if (rel.tag.empty())
return std::nullopt;
rel.title = json.value("name", rel.tag);
rel.id = json.value("id", std::hash<std::string>{}(rel.title));
rel.published = ParseIsoTimestamp(json.value("published_at", std::string{}));
rel.prerelease = json.value("prerelease", false);
auto body = json.value("body", rel.title);
boost::replace_all(body, "\\r", "");
boost::replace_all(body, "\\n", "\n");
rel.body = body;
rel.host = host;
const auto release_base =
fmt::format("{}/{}/releases", Common::g_build_auto_update_website, repo);
const auto fallback_html = fmt::format("{}/tag/{}", release_base, rel.tag);
rel.html_url = json.value("html_url", fallback_html);
// This is our own "fake" API.
if (json.contains("base")) {
const auto base = json.value("base", fmt::format("https://{}", Common::g_build_auto_update_api));
rel.base_download_url = fmt::format("{}/{}", base, rel.tag);
// Assets are easy :)
rel.assets = json.value("assets", std::vector<std::string>{});
} else {
const auto base_download_url = fmt::format("/{}/releases/download/{}", repo, rel.tag);
rel.base_download_url = base_download_url;
// assets are a bit more complex here. :(
std::vector<std::string> assets;
const nlohmann::json& arr = json["assets"];
for (const auto &obj : arr) {
const auto url = obj.value("browser_download_url", std::string{});
assets.emplace_back(url);
}
rel.assets = assets;
}
return rel;
}
std::optional<Release> Release::FromJson(const std::string_view& json, const std::string& host,
const std::string& repo) {
try {
return FromJson(nlohmann::json::parse(json), host, repo);
} catch (std::exception& e) {
LOG_WARNING(Common, "Failed to parse JSON: {}", e.what());
}
return {};
}
std::vector<Release> Release::ListFromJson(const nlohmann::json& json, const std::string& host,
const std::string& repo) {
if (!json.is_array())
return {};
std::vector<Release> releases;
for (const auto& obj : json) {
auto rel = Release::FromJson(obj, host, repo);
if (rel)
releases.emplace_back(rel.value());
}
return releases;
}
std::vector<Release> Release::ListFromJson(const std::string_view& json, const std::string& host,
const std::string& repo) {
try {
return ListFromJson(nlohmann::json::parse(json), host, repo);
} catch (std::exception& e) {
LOG_WARNING(Common, "Failed to parse JSON: {}", e.what());
}
return {};
}
std::optional<std::string> MakeRequest(const std::string& url, const std::string& path) {
try {
constexpr std::size_t timeout_seconds = 15;
std::unique_ptr<httplib::Client> client = std::make_unique<httplib::Client>(url);
client->set_connection_timeout(timeout_seconds);
client->set_read_timeout(timeout_seconds);
client->set_write_timeout(timeout_seconds);
#ifdef YUZU_BUNDLED_OPENSSL
client->load_ca_cert_store(kCert, sizeof(kCert));
#endif
if (client == nullptr) {
LOG_ERROR(Common, "Invalid URL {}{}", url, path);
return {};
}
httplib::Request request{
.method = "GET",
.path = path,
};
client->set_follow_location(true);
httplib::Result result = client->send(request);
if (!result) {
LOG_ERROR(Common, "GET to {}{} returned null", url, path);
return {};
}
const auto& response = result.value();
if (response.status >= 400) {
LOG_ERROR(Common, "GET to {}{} returned error status code: {}", url, path,
response.status);
return {};
}
if (!response.headers.contains("content-type")) {
LOG_ERROR(Common, "GET to {}{} returned no content", url, path);
return {};
}
return response.body;
} catch (std::exception& e) {
LOG_ERROR(Common, "GET to {}{} failed during update check: {}", url, path, e.what());
return std::nullopt;
}
}
std::vector<Release> GetReleases() {
const auto body = GetReleasesBody();
if (!body) {
LOG_WARNING(Common, "Failed to get stable releases");
return {};
}
const std::string_view body_str = body.value();
const auto url = fmt::format("https://{}", Common::g_build_auto_update_stable_api);
return Release::ListFromJson(body_str, url, Common::g_build_auto_update_stable_repo);
}
std::optional<Release> GetLatestRelease() {
const auto releases_path = Common::g_build_auto_update_api_path;
const auto url = fmt::format("https://{}", Common::g_build_auto_update_api);
const auto body = MakeRequest(url, releases_path);
if (!body) {
LOG_WARNING(Common, "Failed to get latest release");
return std::nullopt;
}
const std::string_view body_str = body.value();
return Release::FromJson(body_str, url, Common::g_build_auto_update_repo);
}
std::optional<std::string> GetReleasesBody() {
const auto releases_path =
fmt::format("/{}/{}/releases", Common::g_build_auto_update_stable_api_path,
Common::g_build_auto_update_stable_repo);
const auto url = fmt::format("https://{}", Common::g_build_auto_update_stable_api);
return MakeRequest(url, releases_path);
}
} // namespace Common::Net
-56
View File
@@ -1,56 +0,0 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
#pragma once
#include <optional>
#include <string>
#include <vector>
#include <nlohmann/json.hpp>
#include "common/common_types.h"
namespace Common::Net {
typedef struct {
std::string name;
std::string url;
std::string path;
std::string filename;
} Asset;
typedef struct Release {
std::string title;
std::string body;
std::string tag;
std::string base_download_url;
std::string html_url;
std::string host;
std::vector<std::string> assets;
u64 id;
u64 published;
bool prerelease;
// Get the relevant list of assets for the current platform.
std::vector<Asset> GetPlatformAssets() const;
static std::optional<Release> FromJson(const nlohmann::json& json, const std::string &host, const std::string& repo);
static std::optional<Release> FromJson(const std::string_view& json, const std::string &host, const std::string& repo);
static std::vector<Release> ListFromJson(const nlohmann::json &json, const std::string &host, const std::string &repo);
static std::vector<Release> ListFromJson(const std::string_view &json, const std::string &host, const std::string &repo);
} Release;
// Make a request via httplib, and return the response body if applicable.
std::optional<std::string> MakeRequest(const std::string &url, const std::string &path);
// Get all of the latest stable releases.
std::vector<Release> GetReleases();
// Get all of the latest stable releases as text.
std::optional<std::string> GetReleasesBody();
// Get the latest release of the current channel.
std::optional<Release> GetLatestRelease();
}
-6
View File
@@ -22,9 +22,6 @@
#define BUILD_AUTO_UPDATE_API "@BUILD_AUTO_UPDATE_API@"
#define BUILD_AUTO_UPDATE_API_PATH "@BUILD_AUTO_UPDATE_API_PATH@"
#define BUILD_AUTO_UPDATE_REPO "@BUILD_AUTO_UPDATE_REPO@"
#define BUILD_AUTO_UPDATE_STABLE_API "@BUILD_AUTO_UPDATE_STABLE_API@"
#define BUILD_AUTO_UPDATE_STABLE_API_PATH "@BUILD_AUTO_UPDATE_STABLE_API_PATH@"
#define BUILD_AUTO_UPDATE_STABLE_REPO "@BUILD_AUTO_UPDATE_STABLE_REPO@"
#define IS_NIGHTLY_BUILD @IS_NIGHTLY_BUILD@
namespace Common {
@@ -48,8 +45,5 @@ constexpr const char g_build_auto_update_website[] = BUILD_AUTO_UPDATE_WEBSITE;
constexpr const char g_build_auto_update_api[] = BUILD_AUTO_UPDATE_API;
constexpr const char g_build_auto_update_api_path[] = BUILD_AUTO_UPDATE_API_PATH;
constexpr const char g_build_auto_update_repo[] = BUILD_AUTO_UPDATE_REPO;
constexpr const char g_build_auto_update_stable_api[] = BUILD_AUTO_UPDATE_STABLE_API;
constexpr const char g_build_auto_update_stable_api_path[] = BUILD_AUTO_UPDATE_STABLE_API_PATH;
constexpr const char g_build_auto_update_stable_repo[] = BUILD_AUTO_UPDATE_STABLE_REPO;
} // namespace Common
-3
View File
@@ -28,8 +28,5 @@ extern const char g_build_auto_update_website[];
extern const char g_build_auto_update_api[];
extern const char g_build_auto_update_api_path[];
extern const char g_build_auto_update_repo[];
extern const char g_build_auto_update_stable_api[];
extern const char g_build_auto_update_stable_api_path[];
extern const char g_build_auto_update_stable_repo[];
} // namespace Common
@@ -317,14 +317,12 @@ void WebBrowser::Execute() {
return;
}
#ifndef ANDROID
if (Settings::values.disable_web_applet) {
LOG_WARNING(Service_AM, "(STUBBED) called, Web Browser Applet is disabled. shim_kind={}",
web_arg_header.shim_kind);
WebBrowserExit(WebExitReason::EndButtonPressed);
return;
}
#endif
switch (web_arg_header.shim_kind) {
case ShimKind::Shop:
+128 -31
View File
@@ -1,8 +1,6 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
#include "common/net/net.h"
#include "common/scm_rev.h"
#include "core/hle/service/bcat/news/builtin_news.h"
#include "core/hle/service/bcat/news/msgpack.h"
#include "core/hle/service/bcat/news/news_storage.h"
@@ -24,8 +22,10 @@
#include <filesystem>
#include <fstream>
#include <future>
#include <iomanip>
#include <mutex>
#include <optional>
#include <sstream>
#include <thread>
#ifdef YUZU_BUNDLED_OPENSSL
@@ -35,6 +35,9 @@
namespace Service::News {
namespace {
// TODO(crueter): COMPILE DEFINITION
constexpr const char* GitHubAPI_EdenReleases = "/api/v1/repos/eden-emu/eden/releases";
// Cached logo data
std::vector<u8> default_logo_small;
std::vector<u8> default_logo_large;
@@ -63,6 +66,24 @@ u32 HashToNewsId(std::string_view key) {
return static_cast<u32>(std::hash<std::string_view>{}(key) & 0x7FFFFFFF);
}
u64 ParseIsoTimestamp(const std::string& iso) {
if (iso.empty()) return 0;
std::string buf = iso;
if (buf.back() == 'Z') buf.pop_back();
std::tm tm{};
std::istringstream ss(buf);
ss >> std::get_time(&tm, "%Y-%m-%dT%H:%M:%S");
if (ss.fail()) return 0;
#ifdef _WIN32
return static_cast<u64>(_mkgmtime(&tm));
#else
return static_cast<u64>(timegm(&tm));
#endif
}
std::vector<u8> TryLoadFromDisk(const std::filesystem::path& path) {
if (!std::filesystem::exists(path)) return {};
@@ -79,9 +100,8 @@ std::vector<u8> TryLoadFromDisk(const std::filesystem::path& path) {
return data;
}
// TODO(crueter): Migrate to use Common::Net
std::vector<u8> DownloadImage(const std::string& url_path, const std::filesystem::path& cache_path) {
LOG_DEBUG(Service_BCAT, "Downloading image: https://eden-emu.dev{}", url_path);
LOG_INFO(Service_BCAT, "Downloading image: https://eden-emu.dev{}", url_path);
try {
httplib::Client cli("https://eden-emu.dev");
cli.set_follow_location(true);
@@ -206,6 +226,67 @@ void WriteCachedJson(std::string_view json) {
(void)Common::FS::WriteStringToFile(path, Common::FS::FileType::TextFile, json);
}
std::optional<std::string> DownloadReleasesJson() {
try {
#ifdef YUZU_BUNDLED_OPENSSL
const auto url = "https://git.eden-emu.dev";
#else
const auto url = "git.eden-emu.dev";
#endif
// TODO(crueter): This is duplicated between frontend and here.
constexpr auto path = GitHubAPI_EdenReleases;
constexpr std::size_t timeout_seconds = 15;
std::unique_ptr<httplib::Client> client = std::make_unique<httplib::Client>(url);
client->set_connection_timeout(timeout_seconds);
client->set_read_timeout(timeout_seconds);
client->set_write_timeout(timeout_seconds);
#ifdef YUZU_BUNDLED_OPENSSL
client->load_ca_cert_store(kCert, sizeof(kCert));
#endif
if (client == nullptr) {
LOG_ERROR(Service_BCAT, "Invalid URL {}{}", url, path);
return {};
}
httplib::Request request{
.method = "GET",
.path = path,
};
client->set_follow_location(true);
httplib::Result result = client->send(request);
if (!result) {
LOG_ERROR(Service_BCAT, "GET to {}{} returned null", url, path);
return {};
} else if (result->status < 400) {
return result->body;
}
if (result->status >= 400) {
LOG_ERROR(Service_BCAT,
"GET to {}{} returned error status code: {}",
url,
path,
result->status);
return {};
}
if (!result->headers.contains("content-type")) {
LOG_ERROR(Service_BCAT, "GET to {}{} returned no content", url, path);
return {};
}
} catch (...) {
LOG_WARNING(Service_BCAT, " failed to download releases");
}
return std::nullopt;
}
// idk but News App does not render Markdown or HTML, so remove some formatting.
std::string SanitizeMarkdown(std::string_view markdown) {
std::string result;
@@ -261,7 +342,9 @@ std::string SanitizeMarkdown(std::string_view markdown) {
return text;
}
std::string FormatBody(std::string body, const std::string_view &title) {
std::string FormatBody(const nlohmann::json& release, std::string_view title) {
std::string body = release.value("body", std::string{});
if (body.empty()) {
return std::string(title);
}
@@ -292,32 +375,52 @@ std::string FormatBody(std::string body, const std::string_view &title) {
return body;
}
void ImportReleases(const std::vector<Common::Net::Release> &releases) {
void ImportReleases(std::string_view json_text) {
nlohmann::json root;
try {
root = nlohmann::json::parse(json_text);
} catch (...) {
LOG_WARNING(Service_BCAT, "failed to parse JSON");
return;
}
if (!root.is_array()) return;
std::vector<u32> news_ids;
for (const auto& rel : releases) {
const u32 news_id = u32(rel.id & 0x7FFFFFFF);
for (const auto& rel : root) {
if (!rel.is_object()) continue;
std::string title = rel.value("name", rel.value("tag_name", std::string{}));
if (title.empty()) continue;
const u64 release_id = rel.value("id", 0);
const u32 news_id = release_id ? static_cast<u32>(release_id & 0x7FFFFFFF) : HashToNewsId(title);
news_ids.push_back(news_id);
}
PreloadNewsImages(news_ids);
for (const auto& rel : releases) {
const std::string title = rel.title;
const std::string body = rel.body;
const std::string html_url = rel.html_url;
for (const auto& rel : root) {
if (!rel.is_object()) continue;
const u32 news_id = u32(rel.id & 0x7FFFFFFF);
const u64 published = rel.published;
std::string title = rel.value("name", rel.value("tag_name", std::string{}));
if (title.empty()) continue;
const u64 release_id = rel.value("id", 0);
const u32 news_id = release_id ? static_cast<u32>(release_id & 0x7FFFFFFF) : HashToNewsId(title);
const u64 published = ParseIsoTimestamp(rel.value("published_at", std::string{}));
const u64 pickup_limit = published + 600000000;
const u32 priority = rel.prerelease ? 1500 : 2500;
const u32 priority = rel.value("prerelease", false) ? 1500 : 2500;
std::string author = "Eden";
std::string author = "eden";
if (rel.contains("author") && rel["author"].is_object()) {
author = rel["author"].value("login", "eden");
}
auto payload = BuildMsgpack(title, FormatBody(body, title), title, published,
auto payload = BuildMsgpack(title, FormatBody(rel, title), title, published,
pickup_limit, priority, {"en"}, author, {},
html_url, news_id);
rel.value("html_url", std::string{}), news_id);
const std::string news_id_str = fmt::format("LA{:020}", rel.id);
const std::string news_id_str = fmt::format("LA{:020}", news_id);
GithubNewsMeta meta{
.news_id = news_id_str,
@@ -462,21 +565,15 @@ void EnsureBuiltinNewsLoaded() {
LoadDefaultLogos();
if (const auto cached = ReadCachedJson()) {
const std::string_view body = cached.value();
const auto releases = Common::Net::Release::ListFromJson(body, Common::g_build_auto_update_stable_api, Common::g_build_auto_update_stable_repo);
ImportReleases(releases);
LOG_INFO(Service_BCAT, "news: {} entries loaded from cache", NewsStorage::Instance().ListAll().size());
ImportReleases(*cached);
LOG_DEBUG(Service_BCAT, "news: {} entries loaded from cache", NewsStorage::Instance().ListAll().size());
}
std::thread([] {
if (const auto fresh = Common::Net::GetReleasesBody()) {
const std::string_view body = fresh.value();
WriteCachedJson(body);
const auto releases = Common::Net::Release::ListFromJson(body, Common::g_build_auto_update_stable_api, Common::g_build_auto_update_stable_repo);
ImportReleases(releases);
LOG_INFO(Service_BCAT, "news: {} entries updated from Forgejo", NewsStorage::Instance().ListAll().size());
if (const auto fresh = DownloadReleasesJson()) {
WriteCachedJson(*fresh);
ImportReleases(*fresh);
LOG_DEBUG(Service_BCAT, "news: {} entries updated from Forgejo", NewsStorage::Instance().ListAll().size());
}
}).detach();
});
+7 -3
View File
@@ -89,7 +89,7 @@ void NvMap::UnmapHandle(Handle& handle_description) {
// Free and unmap the handle from Host1x GMMU
if (handle_description.pin_virt_address) {
host1x.gmmu_manager.Unmap(static_cast<GPUVAddr>(handle_description.pin_virt_address),
host1x.GMMU().Unmap(static_cast<GPUVAddr>(handle_description.pin_virt_address),
handle_description.aligned_size);
host1x.Allocator().Free(handle_description.pin_virt_address,
static_cast<u32>(handle_description.aligned_size));
@@ -169,8 +169,12 @@ DAddr NvMap::PinHandle(NvMap::Handle::Id handle, bool low_area_pin) {
std::scoped_lock lock(handle_description->mutex);
const auto map_low_area = [&] {
if (handle_description->pin_virt_address == 0) {
u32 address = host1x.Allocator().Allocate(u32(handle_description->aligned_size));
host1x.gmmu_manager.Map(GPUVAddr(address), handle_description->d_address, handle_description->aligned_size);
auto& gmmu_allocator = host1x.Allocator();
auto& gmmu = host1x.GMMU();
u32 address =
gmmu_allocator.Allocate(static_cast<u32>(handle_description->aligned_size));
gmmu.Map(static_cast<GPUVAddr>(address), handle_description->d_address,
handle_description->aligned_size);
handle_description->pin_virt_address = address;
}
};
@@ -10,7 +10,6 @@
#include "common/alignment.h"
#include "common/assert.h"
#include "common/common_types.h"
#include "common/logging.h"
#include "core/core.h"
#include "core/hle/service/nvdrv/core/container.h"
@@ -131,12 +130,12 @@ NvResult nvhost_as_gpu::AllocAsEx(IoctlAllocAsEx& params) {
const auto start_pages{static_cast<u32>(vm.va_range_start >> VM::PAGE_SIZE_BITS)};
const auto end_pages{static_cast<u32>(vm.va_range_split >> VM::PAGE_SIZE_BITS)};
vm.small_page_allocator.emplace(start_pages, end_pages);
vm.small_page_allocator = std::make_shared<VM::Allocator>(start_pages, end_pages);
const auto start_big_pages{static_cast<u32>(vm.va_range_split >> vm.big_page_size_bits)};
const auto end_big_pages{
static_cast<u32>((vm.va_range_end - vm.va_range_split) >> vm.big_page_size_bits)};
vm.big_page_allocator.emplace(start_big_pages, end_big_pages);
vm.big_page_allocator = std::make_unique<VM::Allocator>(start_big_pages, end_big_pages);
gmmu = std::make_shared<Tegra::MemoryManager>(system, max_big_page_bits, vm.va_range_split,
vm.big_page_size_bits, VM::PAGE_SIZE_BITS);
@@ -189,8 +188,8 @@ NvResult nvhost_as_gpu::AllocateSpace(IoctlAllocSpace& params) {
}
allocation_map[params.offset] = {
.mappings{},
.size = size,
.mappings{},
.page_size = params.page_size,
.sparse = (params.flags & MappingFlags::Sparse) != MappingFlags::None,
.big_pages = params.page_size != VM::YUZU_PAGESIZE,
@@ -200,52 +199,71 @@ NvResult nvhost_as_gpu::AllocateSpace(IoctlAllocSpace& params) {
}
void nvhost_as_gpu::FreeMappingLocked(u64 offset) {
auto const it = mapping_map.find(offset);
auto const mapping = it->second;
if (!mapping.fixed) {
auto& allocator{mapping.big_page ? *vm.big_page_allocator : *vm.small_page_allocator};
u32 page_size_bits{mapping.big_page ? vm.big_page_size_bits : VM::PAGE_SIZE_BITS};
u32 page_size{mapping.big_page ? vm.big_page_size : VM::YUZU_PAGESIZE};
u64 aligned_size{Common::AlignUp(mapping.size, page_size)};
allocator.Free(u32(mapping.offset >> page_size_bits), u32(aligned_size >> page_size_bits));
auto mapping{mapping_map.at(offset)};
if (!mapping->fixed) {
auto& allocator{mapping->big_page ? *vm.big_page_allocator : *vm.small_page_allocator};
u32 page_size_bits{mapping->big_page ? vm.big_page_size_bits : VM::PAGE_SIZE_BITS};
u32 page_size{mapping->big_page ? vm.big_page_size : VM::YUZU_PAGESIZE};
u64 aligned_size{Common::AlignUp(mapping->size, page_size)};
allocator.Free(static_cast<u32>(mapping->offset >> page_size_bits),
static_cast<u32>(aligned_size >> page_size_bits));
}
nvmap.UnpinHandle(mapping.handle);
nvmap.UnpinHandle(mapping->handle);
// Sparse mappings shouldn't be fully unmapped, just returned to their sparse state
// Only FreeSpace can unmap them fully
if (mapping.sparse_alloc) {
gmmu->MapSparse(offset, mapping.size, mapping.big_page);
if (mapping->sparse_alloc) {
gmmu->MapSparse(offset, mapping->size, mapping->big_page);
} else {
gmmu->Unmap(offset, mapping.size);
gmmu->Unmap(offset, mapping->size);
}
mapping_map.erase(it);
mapping_map.erase(offset);
}
NvResult nvhost_as_gpu::FreeSpace(IoctlFreeSpace& params) {
LOG_DEBUG(Service_NVDRV, "called, offset={:X}, pages={:X}, page_size={:X}", params.offset, params.pages, params.page_size);
LOG_DEBUG(Service_NVDRV, "called, offset={:X}, pages={:X}, page_size={:X}", params.offset,
params.pages, params.page_size);
std::scoped_lock lock(mutex);
if (!vm.initialised) {
return NvResult::BadValue;
}
if (auto const it = allocation_map.find(params.offset); it != allocation_map.end()) {
auto const allocation = it->second;
if (allocation.page_size != params.page_size || allocation.size != (u64(params.pages) * params.page_size))
return NvResult::BadValue;
for (const auto mapping_offset : allocation.mappings)
FreeMappingLocked(mapping_offset);
try {
auto allocation{allocation_map[params.offset]};
if (allocation.page_size != params.page_size ||
allocation.size != (static_cast<u64>(params.pages) * params.page_size)) {
return NvResult::BadValue;
}
for (const auto& mapping : allocation.mappings) {
FreeMappingLocked(mapping->offset);
}
// Unset sparse flag if required
if (allocation.sparse)
if (allocation.sparse) {
gmmu->Unmap(params.offset, allocation.size);
}
auto& allocator{params.page_size == VM::YUZU_PAGESIZE ? *vm.small_page_allocator : *vm.big_page_allocator};
u32 page_size_bits{params.page_size == VM::YUZU_PAGESIZE ? VM::PAGE_SIZE_BITS : vm.big_page_size_bits};
auto& allocator{params.page_size == VM::YUZU_PAGESIZE ? *vm.small_page_allocator
: *vm.big_page_allocator};
u32 page_size_bits{params.page_size == VM::YUZU_PAGESIZE ? VM::PAGE_SIZE_BITS
: vm.big_page_size_bits};
allocator.Free(u32(params.offset >> page_size_bits), u32(allocation.size >> page_size_bits));
allocator.Free(static_cast<u32>(params.offset >> page_size_bits),
static_cast<u32>(allocation.size >> page_size_bits));
allocation_map.erase(params.offset);
return NvResult::Success;
} catch (const std::out_of_range&) {
return NvResult::BadValue;
}
return NvResult::BadValue;
return NvResult::Success;
}
NvResult nvhost_as_gpu::Remap(std::span<IoctlRemapEntry> entries) {
@@ -309,18 +327,26 @@ NvResult nvhost_as_gpu::MapBufferEx(IoctlMapBufferEx& params) {
// Remaps a subregion of an existing mapping to a different PA
if ((params.flags & MappingFlags::Remap) != MappingFlags::None) {
if (auto const it = mapping_map.find(params.offset); it != mapping_map.end()) {
auto const mapping = it->second;
if (mapping.size < params.mapping_size) {
LOG_WARNING(Service_NVDRV, "Cannot remap a partially mapped GPU address space region: {:#X}", params.offset);
try {
auto mapping{mapping_map.at(params.offset)};
if (mapping->size < params.mapping_size) {
LOG_WARNING(Service_NVDRV,
"Cannot remap a partially mapped GPU address space region: {:#X}",
params.offset);
return NvResult::BadValue;
}
u64 gpu_address = u64(params.offset + params.buffer_offset);
VAddr device_address{mapping.ptr + params.buffer_offset};
gmmu->Map(gpu_address, device_address, params.mapping_size, Tegra::PTEKind(params.kind), mapping.big_page);
u64 gpu_address{static_cast<u64>(params.offset + params.buffer_offset)};
VAddr device_address{mapping->ptr + params.buffer_offset};
gmmu->Map(gpu_address, device_address, params.mapping_size,
static_cast<Tegra::PTEKind>(params.kind), mapping->big_page);
return NvResult::Success;
} else {
LOG_WARNING(Service_NVDRV, "Cannot remap an unmapped GPU address space region: {:#X}", params.offset);
} catch (const std::out_of_range&) {
LOG_WARNING(Service_NVDRV, "Cannot remap an unmapped GPU address space region: {:#X}",
params.offset);
return NvResult::BadValue;
}
}
@@ -330,7 +356,8 @@ NvResult nvhost_as_gpu::MapBufferEx(IoctlMapBufferEx& params) {
return NvResult::BadValue;
}
DAddr device_address = DAddr(nvmap.PinHandle(params.handle, false) + params.buffer_offset);
DAddr device_address{
static_cast<DAddr>(nvmap.PinHandle(params.handle, false) + params.buffer_offset)};
u64 size{params.mapping_size ? params.mapping_size : handle->orig_size};
bool big_page{[&]() {
@@ -354,22 +381,32 @@ NvResult nvhost_as_gpu::MapBufferEx(IoctlMapBufferEx& params) {
}
const bool use_big_pages = alloc->second.big_pages && big_page;
gmmu->Map(params.offset, device_address, size, static_cast<Tegra::PTEKind>(params.kind), use_big_pages);
gmmu->Map(params.offset, device_address, size, static_cast<Tegra::PTEKind>(params.kind),
use_big_pages);
alloc->second.mappings.push_back(params.offset);
mapping_map.insert_or_assign(params.offset, Mapping(params.handle, device_address, params.offset, size, true, use_big_pages, alloc->second.sparse));
auto mapping{std::make_shared<Mapping>(params.handle, device_address, params.offset, size,
true, use_big_pages, alloc->second.sparse)};
alloc->second.mappings.push_back(mapping);
mapping_map[params.offset] = mapping;
} else {
auto& allocator{big_page ? *vm.big_page_allocator : *vm.small_page_allocator};
u32 page_size{big_page ? vm.big_page_size : VM::YUZU_PAGESIZE};
u32 page_size_bits{big_page ? vm.big_page_size_bits : VM::PAGE_SIZE_BITS};
params.offset = u64(allocator.Allocate(u32(Common::AlignUp(size, page_size) >> page_size_bits))) << page_size_bits;
params.offset = static_cast<u64>(allocator.Allocate(
static_cast<u32>(Common::AlignUp(size, page_size) >> page_size_bits)))
<< page_size_bits;
if (!params.offset) {
ASSERT_MSG(false, "Failed to allocate free space in the GPU AS!");
return NvResult::InsufficientMemory;
}
gmmu->Map(params.offset, device_address, Common::AlignUp(size, page_size), Tegra::PTEKind(params.kind), big_page);
mapping_map.insert_or_assign(params.offset, Mapping(params.handle, device_address, params.offset, size, false, big_page, false));
gmmu->Map(params.offset, device_address, Common::AlignUp(size, page_size),
static_cast<Tegra::PTEKind>(params.kind), big_page);
auto mapping{std::make_shared<Mapping>(params.handle, device_address, params.offset, size,
false, big_page, false)};
mapping_map[params.offset] = mapping;
}
map_buffer_offsets.insert(params.offset);
@@ -378,32 +415,37 @@ NvResult nvhost_as_gpu::MapBufferEx(IoctlMapBufferEx& params) {
}
NvResult nvhost_as_gpu::UnmapBuffer(IoctlUnmapBuffer& params) {
std::scoped_lock lock(mutex);
if (auto const offset_it = map_buffer_offsets.find(params.offset); offset_it != map_buffer_offsets.end()) {
if (map_buffer_offsets.find(params.offset) != map_buffer_offsets.end()) {
LOG_DEBUG(Service_NVDRV, "called, offset={:#X}", params.offset);
std::scoped_lock lock(mutex);
if (!vm.initialised) {
return NvResult::BadValue;
}
auto const it = mapping_map.find(params.offset);
auto const mapping = it->second;
if (!mapping.fixed) {
auto& allocator{mapping.big_page ? *vm.big_page_allocator : *vm.small_page_allocator};
u32 page_size_bits{mapping.big_page ? vm.big_page_size_bits : VM::PAGE_SIZE_BITS};
allocator.Free(u32(mapping.offset >> page_size_bits), u32(mapping.size >> page_size_bits));
auto mapping{mapping_map.at(params.offset)};
if (!mapping->fixed) {
auto& allocator{mapping->big_page ? *vm.big_page_allocator : *vm.small_page_allocator};
u32 page_size_bits{mapping->big_page ? vm.big_page_size_bits : VM::PAGE_SIZE_BITS};
allocator.Free(static_cast<u32>(mapping->offset >> page_size_bits),
static_cast<u32>(mapping->size >> page_size_bits));
}
// Sparse mappings shouldn't be fully unmapped, just returned to their sparse state
// Only FreeSpace can unmap them fully
if (mapping.sparse_alloc) {
gmmu->MapSparse(params.offset, mapping.size, mapping.big_page);
if (mapping->sparse_alloc) {
gmmu->MapSparse(params.offset, mapping->size, mapping->big_page);
} else {
gmmu->Unmap(params.offset, mapping.size);
gmmu->Unmap(params.offset, mapping->size);
}
nvmap.UnpinHandle(mapping.handle);
mapping_map.erase(it);
map_buffer_offsets.erase(offset_it);
nvmap.UnpinHandle(mapping->handle);
mapping_map.erase(params.offset);
map_buffer_offsets.erase(params.offset);
}
return NvResult::Success;
}
@@ -436,7 +478,8 @@ void nvhost_as_gpu::GetVARegionsImpl(IoctlGetVaRegions& params) {
}
NvResult nvhost_as_gpu::GetVARegions1(IoctlGetVaRegions& params) {
LOG_DEBUG(Service_NVDRV, "called, buf_addr={:X}, buf_size={:X}", params.buf_addr, params.buf_size);
LOG_DEBUG(Service_NVDRV, "called, buf_addr={:X}, buf_size={:X}", params.buf_addr,
params.buf_size);
std::scoped_lock lock(mutex);
@@ -165,36 +165,35 @@ private:
NvCore::NvMap& nvmap;
struct Mapping {
NvCore::NvMap::Handle::Id handle;
DAddr ptr;
u64 offset;
u64 size;
NvCore::NvMap::Handle::Id handle;
bool fixed : 1;
bool big_page : 1; // Only valid if fixed == false
bool sparse_alloc : 1;
bool fixed;
bool big_page; // Only valid if fixed == false
bool sparse_alloc;
Mapping(NvCore::NvMap::Handle::Id handle_, DAddr ptr_, u64 offset_, u64 size_, bool fixed_, bool big_page_, bool sparse_alloc_)
: ptr(ptr_), offset(offset_), size(size_), handle(handle_)
, fixed(fixed_), big_page(big_page_), sparse_alloc(sparse_alloc_)
{}
Mapping(NvCore::NvMap::Handle::Id handle_, DAddr ptr_, u64 offset_, u64 size_, bool fixed_,
bool big_page_, bool sparse_alloc_)
: handle(handle_), ptr(ptr_), offset(offset_), size(size_), fixed(fixed_),
big_page(big_page_), sparse_alloc(sparse_alloc_) {}
};
struct Allocation {
std::vector<u64> mappings;
u64 size;
std::list<std::shared_ptr<Mapping>> mappings;
u32 page_size;
bool sparse;
bool big_pages;
};
//!< This maps the base addresses of mapped buffers to their total sizes and
//!< mapping type, this is needed as what was originally a single buffer may
//!< have been split into multiple GPU side buffers with the remap flag.
std::map<u64, Mapping> mapping_map;
//!< Holds allocations created by AllocSpace from
//!< which fixed buffers can be mapped into
std::map<u64, Allocation> allocation_map;
std::mutex mutex; //!< Locks all AS operations
std::map<u64, std::shared_ptr<Mapping>>
mapping_map; //!< This maps the base addresses of mapped buffers to their total sizes and
//!< mapping type, this is needed as what was originally a single buffer may
//!< have been split into multiple GPU side buffers with the remap flag.
std::map<u64, Allocation> allocation_map; //!< Holds allocations created by AllocSpace from
//!< which fixed buffers can be mapped into
std::mutex mutex; //!< Locks all AS operations
struct VM {
static constexpr u32 YUZU_PAGESIZE{0x1000};
@@ -214,8 +213,9 @@ private:
using Allocator = Common::FlatAllocator<u32, 0, 32>;
std::optional<Allocator> big_page_allocator;
std::optional<Allocator> small_page_allocator; //! Shared as this is also used by nvhost::GpuChannel
std::unique_ptr<Allocator> big_page_allocator;
std::shared_ptr<Allocator>
small_page_allocator; //! Shared as this is also used by nvhost::GpuChannel
bool initialised{};
} vm;
+130 -29
View File
@@ -5,45 +5,146 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#ifdef NIGHTLY_BUILD
#include <boost/algorithm/string/classification.hpp>
#include <boost/algorithm/string/split.hpp>
#endif
#include <fmt/format.h>
#include "common/net/net.h"
#include "common/logging.h"
#include "common/scm_rev.h"
#include "update_checker.h"
#include "common/logging.h"
#include "common/httplib.h"
std::optional<Common::Net::Release> UpdateChecker::GetUpdate() {
const auto latest = Common::Net::GetLatestRelease();
if (!latest) return std::nullopt;
LOG_INFO(Frontend, "Received update {}", latest->title);
#ifdef NIGHTLY_BUILD
std::vector<std::string> result;
boost::split(result, latest->tag, boost::is_any_of("."));
if (result.size() != 2)
return std::nullopt;
const std::string tag = result[1];
boost::split(result, std::string{Common::g_build_version}, boost::is_any_of("-"));
if (result.empty())
return std::nullopt;
const std::string build = result[0];
#else
const std::string tag = latest->tag;
const std::string build = Common::g_build_version;
#ifdef YUZU_BUNDLED_OPENSSL
#include <openssl/cert.h>
#endif
if (tag != build)
return latest;
#include <nlohmann/json.hpp>
#include <optional>
#include <string>
std::optional<std::string> UpdateChecker::GetResponse(std::string url, std::string path)
{
try {
constexpr std::size_t timeout_seconds = 15;
std::unique_ptr<httplib::Client> client = std::make_unique<httplib::Client>(url);
client->set_connection_timeout(timeout_seconds);
client->set_read_timeout(timeout_seconds);
client->set_write_timeout(timeout_seconds);
#ifdef YUZU_BUNDLED_OPENSSL
client->load_ca_cert_store(kCert, sizeof(kCert));
#endif
if (client == nullptr) {
LOG_ERROR(Frontend, "Invalid URL {}{}", url, path);
return {};
}
httplib::Request request{
.method = "GET",
.path = path,
};
client->set_follow_location(true);
httplib::Result result = client->send(request);
if (!result) {
LOG_ERROR(Frontend, "GET to {}{} returned null", url, path);
return {};
}
const auto &response = result.value();
if (response.status >= 400) {
LOG_ERROR(Frontend,
"GET to {}{} returned error status code: {}",
url,
path,
response.status);
return {};
}
if (!response.headers.contains("content-type")) {
LOG_ERROR(Frontend, "GET to {}{} returned no content", url, path);
return {};
}
return response.body;
} catch (std::exception &e) {
LOG_ERROR(Frontend,
"GET to {}{} failed during update check: {}",
url,
path,
e.what());
return std::nullopt;
}
}
std::optional<UpdateChecker::Update> UpdateChecker::GetLatestRelease() {
#ifdef YUZU_BUNDLED_OPENSSL
const auto update_check_url = fmt::format("https://{}", Common::g_build_auto_update_api);
#else
const auto update_check_url = std::string{Common::g_build_auto_update_api};
#endif
auto update_check_path = std::string{Common::g_build_auto_update_api_path};
try {
const auto response = UpdateChecker::GetResponse(update_check_url, update_check_path);
if (!response)
return {};
const std::string latest_tag = nlohmann::json::parse(response.value()).at("tag_name");
const std::string latest_name = nlohmann::json::parse(response.value()).at("name");
return Update{latest_tag, latest_name};
} catch (nlohmann::detail::out_of_range&) {
LOG_ERROR(Frontend,
"Parsing JSON response from {}{} failed during update check: "
"nlohmann::detail::out_of_range",
update_check_url,
update_check_path);
return {};
} catch (nlohmann::detail::type_error&) {
LOG_ERROR(Frontend,
"Parsing JSON response from {}{} failed during update check: "
"nlohmann::detail::type_error",
update_check_url,
update_check_path);
return {};
}
}
std::optional<UpdateChecker::Update> UpdateChecker::GetUpdate() {
const std::optional<UpdateChecker::Update> latest_release_tag =
UpdateChecker::GetLatestRelease();
if (!latest_release_tag)
goto empty;
{
std::string tag, build;
if (Common::g_is_nightly_build) {
std::vector<std::string> result;
boost::split(result, latest_release_tag->tag, boost::is_any_of("."));
if (result.size() != 2)
goto empty;
tag = result[1];
boost::split(result, std::string{Common::g_build_version}, boost::is_any_of("-"));
if (result.empty())
goto empty;
build = result[0];
} else {
tag = latest_release_tag->tag;
build = Common::g_build_version;
}
if (tag != build)
return latest_release_tag.value();
}
empty:
return std::nullopt;
}
+8 -2
View File
@@ -8,10 +8,16 @@
#pragma once
#include <optional>
#include "common/net/net.h"
#include <string>
namespace UpdateChecker {
std::optional<Common::Net::Release> GetUpdate();
typedef struct {
std::string tag;
std::string name;
} Update;
std::optional<std::string> GetResponse(std::string url, std::string path);
std::optional<Update> GetLatestRelease();
std::optional<Update> GetUpdate();
} // namespace UpdateChecker
@@ -30,8 +30,9 @@ struct TextureInst {
using TextureInstVector = boost::container::small_vector<TextureInst, 24>;
constexpr u32 DESCRIPTOR_SIZE = 8;
constexpr u32 DESCRIPTOR_SIZE_SHIFT = static_cast<u32>(std::countr_zero(DESCRIPTOR_SIZE));
// TODO: We need to implement scatter/gather CBuf stuff
constexpr u32 DESCRIPTOR_SIZE = 1024;
constexpr u32 DESCRIPTOR_SIZE_SHIFT = u32(std::countr_zero(u32(8)));
IR::Opcode IndexedInstruction(const IR::Inst& inst) {
switch (inst.GetOpcode()) {
@@ -362,7 +363,7 @@ std::optional<ConstBufferAddr> TryGetConstBuffer(const IR::Inst* inst, Environme
.secondary_offset = 0,
.secondary_shift_left = 0,
.dynamic_offset = dynamic_offset,
.count = 8,
.count = DESCRIPTOR_SIZE,
.has_secondary = false,
};
}
@@ -406,15 +407,15 @@ u32 GetTextureHandle(Environment& env, const ConstBufferAddr& cbuf) {
return lhs_raw | rhs_raw;
}
[[maybe_unused]]TextureType ReadTextureType(Environment& env, const ConstBufferAddr& cbuf) {
[[maybe_unused]]TextureType ReadTextureType(Environment& env, const ConstBufferAddr& cbuf) {
return env.ReadTextureType(GetTextureHandle(env, cbuf));
}
[[maybe_unused]]TexturePixelFormat ReadTexturePixelFormat(Environment& env, const ConstBufferAddr& cbuf) {
[[maybe_unused]]TexturePixelFormat ReadTexturePixelFormat(Environment& env, const ConstBufferAddr& cbuf) {
return env.ReadTexturePixelFormat(GetTextureHandle(env, cbuf));
}
[[maybe_unused]]bool IsTexturePixelFormatInteger(Environment& env, const ConstBufferAddr& cbuf) {
[[maybe_unused]]bool IsTexturePixelFormatInteger(Environment& env, const ConstBufferAddr& cbuf) {
return env.IsTexturePixelFormatInteger(GetTextureHandle(env, cbuf));
}
@@ -716,8 +717,7 @@ void TexturePass(Environment& env, IR::Program& program, const HostTranslateInfo
const auto insert_point{IR::Block::InstructionList::s_iterator_to(*inst)};
IR::IREmitter ir{*texture_inst.block, insert_point};
const IR::U32 shift{ir.Imm32(DESCRIPTOR_SIZE_SHIFT)};
inst->SetArg(0, ir.UMin(ir.ShiftRightLogical(cbuf.dynamic_offset, shift),
ir.Imm32(DESCRIPTOR_SIZE - 1)));
inst->SetArg(0, ir.UMin(ir.ShiftRightLogical(cbuf.dynamic_offset, shift), ir.Imm32(DESCRIPTOR_SIZE - 1)));
} else {
inst->SetArg(0, IR::Value{});
}
+5 -4
View File
@@ -8,7 +8,7 @@ add_subdirectory(host_shaders)
add_subdirectory(gpu_logging)
if(LIBVA_FOUND)
set_source_files_properties(host1x/ffmpeg.cpp
set_source_files_properties(host1x/ffmpeg/ffmpeg.cpp
PROPERTIES COMPILE_DEFINITIONS LIBVA_FOUND=1)
list(APPEND FFmpeg_LIBRARIES ${LIBVA_LIBRARIES})
endif()
@@ -44,6 +44,7 @@ add_library(video_core STATIC
engines/sw_blitter/converter.h
engines/const_buffer_info.h
engines/draw_manager.cpp
engines/draw_manager.h
engines/engine_interface.h
engines/engine_upload.cpp
engines/engine_upload.h
@@ -71,9 +72,9 @@ add_library(video_core STATIC
host1x/codecs/vp8.h
host1x/codecs/vp9.cpp
host1x/codecs/vp9.h
host1x/codec_types.h
host1x/ffmpeg.cpp
host1x/ffmpeg.h
host1x/codecs/vp9_types.h
host1x/ffmpeg/ffmpeg.cpp
host1x/ffmpeg/ffmpeg.h
host1x/control.cpp
host1x/control.h
host1x/gpu_device_memory_manager.cpp
+19 -13
View File
@@ -356,7 +356,7 @@ void BufferCache<P>::BindHostGeometryBuffers(bool is_indexed) {
if (is_indexed) {
BindHostIndexBuffer();
} else if constexpr (!HAS_FULL_INDEX_AND_PRIMITIVE_SUPPORT) {
const auto& draw_state = maxwell3d->draw_manager.draw_state;
const auto& draw_state = maxwell3d->draw_manager->GetDrawState();
if (draw_state.topology == Maxwell::PrimitiveTopology::Quads ||
draw_state.topology == Maxwell::PrimitiveTopology::QuadStrip) {
runtime.BindQuadIndexBuffer(draw_state.topology, draw_state.vertex_buffer.first,
@@ -740,25 +740,30 @@ void BufferCache<P>::BindHostIndexBuffer() {
TouchBuffer(buffer, channel_state->index_buffer.buffer_id);
const u32 offset = buffer.Offset(channel_state->index_buffer.device_addr);
const u32 size = channel_state->index_buffer.size;
const auto& draw_state = maxwell3d->draw_manager.draw_state;
if (draw_state.inline_index_draw_indexes.empty()) {
SynchronizeBuffer(buffer, channel_state->index_buffer.device_addr, size);
} else {
const auto& draw_state = maxwell3d->draw_manager->GetDrawState();
if (!draw_state.inline_index_draw_indexes.empty()) [[unlikely]] {
if constexpr (USE_MEMORY_MAPS_FOR_UPLOADS) {
auto upload_staging = runtime.UploadStagingBuffer(size);
std::array<BufferCopy, 1> copies{{BufferCopy{.src_offset = upload_staging.offset, .dst_offset = 0, .size = size}}};
std::memcpy(upload_staging.mapped_span.data(), draw_state.inline_index_draw_indexes.data(), size);
std::array<BufferCopy, 1> copies{
{BufferCopy{.src_offset = upload_staging.offset, .dst_offset = 0, .size = size}}};
std::memcpy(upload_staging.mapped_span.data(),
draw_state.inline_index_draw_indexes.data(), size);
runtime.CopyBuffer(buffer, upload_staging.buffer, copies, true);
} else {
buffer.ImmediateUpload(0, draw_state.inline_index_draw_indexes);
}
} else {
SynchronizeBuffer(buffer, channel_state->index_buffer.device_addr, size);
}
if constexpr (HAS_FULL_INDEX_AND_PRIMITIVE_SUPPORT) {
const u32 new_offset = offset + draw_state.index_buffer.first * draw_state.index_buffer.FormatSizeInBytes();
const u32 new_offset =
offset + draw_state.index_buffer.first * draw_state.index_buffer.FormatSizeInBytes();
runtime.BindIndexBuffer(buffer, new_offset, size);
} else {
buffer.MarkUsage(offset, size);
runtime.BindIndexBuffer(draw_state.topology, draw_state.index_buffer.format, draw_state.index_buffer.first, draw_state.index_buffer.count, buffer, offset, size);
runtime.BindIndexBuffer(draw_state.topology, draw_state.index_buffer.format,
draw_state.index_buffer.first, draw_state.index_buffer.count,
buffer, offset, size);
}
}
@@ -940,9 +945,10 @@ void BufferCache<P>::BindHostGraphicsUniformBuffer(size_t stage, u32 index, u32
return alignment > 1 && (offset % alignment) != 0;
}
}();
const bool use_fast_buffer = needs_alignment_stream
|| (has_host_buffer && size <= channel_state->uniform_buffer_skip_cache_size
&& !memory_tracker.IsRegionGpuModified(device_addr, size));
const bool use_fast_buffer = needs_alignment_stream ||
(has_host_buffer &&
size <= channel_state->uniform_buffer_skip_cache_size &&
!memory_tracker.IsRegionGpuModified(device_addr, size));
if (use_fast_buffer) {
if constexpr (IS_OPENGL) {
if (runtime.HasFastBufferSubData()) {
@@ -1223,7 +1229,7 @@ template <class P>
void BufferCache<P>::UpdateIndexBuffer() {
// We have to check for the dirty flags and index count
// The index count is currently changed without updating the dirty flags
const auto& draw_state = maxwell3d->draw_manager.draw_state;
const auto& draw_state = maxwell3d->draw_manager->GetDrawState();
const auto& index_buffer_ref = draw_state.index_buffer;
auto& flags = maxwell3d->dirty.flags;
if (!flags[Dirty::IndexBuffer]) {
@@ -32,7 +32,7 @@
#include "video_core/control/channel_state_cache.h"
#include "video_core/delayed_destruction_ring.h"
#include "video_core/dirty_flags.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/engines/draw_manager.h"
#include "video_core/engines/kepler_compute.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/memory_manager.h"
@@ -305,7 +305,7 @@ public:
[[nodiscard]] bool IsRegionCpuModified(DAddr addr, size_t size);
void SetDrawIndirect(
const Tegra::Engines::Maxwell3D::DrawManager::IndirectParams* current_draw_indirect_) {
const Tegra::Engines::DrawManager::IndirectParams* current_draw_indirect_) {
current_draw_indirect = current_draw_indirect_;
}
@@ -480,7 +480,7 @@ private:
#endif
DelayedDestructionRing<Buffer, TICKS_TO_DESTROY> delayed_destruction_ring;
const Tegra::Engines::Maxwell3D::DrawManager::IndirectParams* current_draw_indirect{};
const Tegra::Engines::DrawManager::IndirectParams* current_draw_indirect{};
u32 last_index_count = 0;
+86 -78
View File
@@ -1,4 +1,4 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Ryujinx Team and Contributors
@@ -20,101 +20,109 @@
namespace Tegra {
CDmaPusher::CDmaPusher(Host1x::Host1x& host1x_, s32 id)
: host_processor(host1x_)
, host1x{host1x_}
, current_class{ChClassId(id)}
{
thread = std::jthread([this](std::stop_token stop_token) {
Common::SetCurrentThreadPriority(Common::ThreadPriority::High);
ChCommandHeaderList command_list{host1x.System().ApplicationMemory(), 0, 0};
u32 count{};
u32 method_offset{};
u32 mask{};
bool incrementing{};
while (!stop_token.stop_requested()) {
{
std::unique_lock l{command_mutex};
command_cv.wait(l, stop_token, [this]() { return command_lists.size() > 0; });
if (stop_token.stop_requested()) {
return;
}
command_list = std::move(command_lists.front());
command_lists.pop_front();
}
size_t i = 0;
for (const auto value : command_list) {
i++;
if (mask != 0) {
const auto lbs = static_cast<u32>(std::countr_zero(mask));
mask &= ~(1U << lbs);
ExecuteCommand(method_offset + lbs, value.raw);
continue;
} else if (count != 0) {
--count;
ExecuteCommand(method_offset, value.raw);
if (incrementing) {
++method_offset;
}
continue;
}
const auto mode = value.submission_mode.Value();
switch (mode) {
case ChSubmissionMode::SetClass: {
mask = value.value & 0x3f;
method_offset = value.method_offset;
current_class = ChClassId((value.value >> 6) & 0x3ff);
break;
}
case ChSubmissionMode::Incrementing:
case ChSubmissionMode::NonIncrementing:
count = value.value;
method_offset = value.method_offset;
incrementing = mode == ChSubmissionMode::Incrementing;
break;
case ChSubmissionMode::Mask:
mask = value.value;
method_offset = value.method_offset;
break;
case ChSubmissionMode::Immediate: {
const u32 data = value.value & 0xfff;
method_offset = value.method_offset;
ExecuteCommand(method_offset, data);
break;
}
default:
LOG_ERROR(HW_GPU, "Bad command at index {} (bytes {:#X}), buffer size {}", i - 1, (i - 1) * sizeof(u32), command_list.size());
UNIMPLEMENTED_MSG("ChSubmission mode {} is not implemented!", u32(mode));
break;
}
}
}
});
: host1x{host1x_}, memory_manager{host1x.GMMU()},
host_processor{std::make_unique<Host1x::Control>(host1x_)}, current_class{
static_cast<ChClassId>(id)} {
thread = std::jthread([this](std::stop_token stop_token) { ProcessEntries(stop_token); });
}
CDmaPusher::~CDmaPusher() = default;
void CDmaPusher::ProcessEntries(std::stop_token stop_token) {
Common::SetCurrentThreadPriority(Common::ThreadPriority::High);
ChCommandHeaderList command_list{host1x.System().ApplicationMemory(), 0, 0};
u32 count{};
u32 method_offset{};
u32 mask{};
bool incrementing{};
while (!stop_token.stop_requested()) {
{
std::unique_lock l{command_mutex};
command_cv.wait(l, stop_token,
[this]() { return command_lists.size() > 0; });
if (stop_token.stop_requested()) {
return;
}
command_list = std::move(command_lists.front());
command_lists.pop_front();
}
size_t i = 0;
for (const auto value : command_list) {
i++;
if (mask != 0) {
const auto lbs = static_cast<u32>(std::countr_zero(mask));
mask &= ~(1U << lbs);
ExecuteCommand(method_offset + lbs, value.raw);
continue;
} else if (count != 0) {
--count;
ExecuteCommand(method_offset, value.raw);
if (incrementing) {
++method_offset;
}
continue;
}
const auto mode = value.submission_mode.Value();
switch (mode) {
case ChSubmissionMode::SetClass: {
mask = value.value & 0x3f;
method_offset = value.method_offset;
current_class = static_cast<ChClassId>((value.value >> 6) & 0x3ff);
break;
}
case ChSubmissionMode::Incrementing:
case ChSubmissionMode::NonIncrementing:
count = value.value;
method_offset = value.method_offset;
incrementing = mode == ChSubmissionMode::Incrementing;
break;
case ChSubmissionMode::Mask:
mask = value.value;
method_offset = value.method_offset;
break;
case ChSubmissionMode::Immediate: {
const u32 data = value.value & 0xfff;
method_offset = value.method_offset;
ExecuteCommand(method_offset, data);
break;
}
default:
LOG_ERROR(HW_GPU, "Bad command at index {} (bytes {:#X}), buffer size {}", i - 1,
(i - 1) * sizeof(u32), command_list.size());
UNIMPLEMENTED_MSG("ChSubmission mode {} is not implemented!",
static_cast<u32>(mode));
break;
}
}
}
}
void CDmaPusher::ExecuteCommand(u32 method, u32 arg) {
switch (current_class) {
case ChClassId::Control:
LOG_TRACE(Service_NVDRV, "Class {} method {:#X} arg 0x{:X}", u32(current_class), method, arg);
host_processor.ProcessMethod(Host1x::Control::Method(method), arg);
LOG_TRACE(Service_NVDRV, "Class {} method {:#X} arg 0x{:X}",
static_cast<u32>(current_class), method, arg);
host_processor->ProcessMethod(static_cast<Host1x::Control::Method>(method), arg);
break;
default:
thi_regs.reg_array[method] = arg;
switch (static_cast<ThiMethod>(method)) {
case ThiMethod::IncSyncpt: {
const auto syncpoint_id = u32(arg & 0xFF);
[[maybe_unused]] const auto cond = u32((arg >> 8) & 0xFF);
LOG_TRACE(Service_NVDRV, "Class {} IncSyncpt Method, syncpt {} cond {}", u32(current_class), syncpoint_id, cond);
const auto syncpoint_id = static_cast<u32>(arg & 0xFF);
[[maybe_unused]] const auto cond = static_cast<u32>((arg >> 8) & 0xFF);
LOG_TRACE(Service_NVDRV, "Class {} IncSyncpt Method, syncpt {} cond {}",
static_cast<u32>(current_class), syncpoint_id, cond);
auto& syncpoint_manager = host1x.GetSyncpointManager();
syncpoint_manager.IncrementGuest(syncpoint_id);
syncpoint_manager.IncrementHost(syncpoint_id);
break;
}
case ThiMethod::SetMethod1:
LOG_TRACE(Service_NVDRV, "Class {} method {:#X} arg 0x{:X}", u32(current_class), u32(thi_regs.method_0), arg);
LOG_TRACE(Service_NVDRV, "Class {} method {:#X} arg 0x{:X}",
static_cast<u32>(current_class), static_cast<u32>(thi_regs.method_0), arg);
ProcessMethod(thi_regs.method_0, arg);
break;
default:
+14 -11
View File
@@ -1,6 +1,3 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -18,11 +15,11 @@
#include "common/common_types.h"
#include "common/polyfill_thread.h"
#include "core/memory.h"
#include "video_core/host1x/control.h"
namespace Tegra {
namespace Host1x {
class Control;
class Host1x;
class Nvdec;
class SyncptIncrManager;
@@ -120,19 +117,25 @@ protected:
virtual void ProcessMethod(u32 method, u32 arg) = 0;
Host1x::Host1x& host1x;
Tegra::MemoryManager& memory_manager;
private:
/// Process the command entry
void ProcessEntries(std::stop_token stop_token);
/// Invoke command class devices to execute the command based on the current state
void ExecuteCommand(u32 state_offset, u32 data);
protected:
ThiRegisters thi_regs{};
std::deque<ChCommandHeaderList> command_lists;
std::condition_variable_any command_cv;
Host1x::Control host_processor;
std::unique_ptr<Host1x::Control> host_processor;
std::mutex command_mutex;
Host1x::Host1x& host1x;
ChClassId current_class;
std::condition_variable_any command_cv;
std::deque<ChCommandHeaderList> command_lists;
std::jthread thread;
ThiRegisters thi_regs{};
ChClassId current_class;
};
} // namespace Tegra
+139 -103
View File
@@ -1,24 +1,23 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
#include "common/settings.h"
#include "video_core/dirty_flags.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/engines/draw_manager.h"
#include "video_core/rasterizer_interface.h"
namespace Tegra::Engines {
DrawManager::DrawManager(Maxwell3D* maxwell3d_) : maxwell3d(maxwell3d_) {}
void Maxwell3D::DrawManager::ProcessMethodCall(Maxwell3D& maxwell3d, u32 method, u32 argument) {
void DrawManager::ProcessMethodCall(u32 method, u32 argument) {
const auto& regs{maxwell3d->regs};
switch (method) {
case MAXWELL3D_REG_INDEX(clear_surface):
return Clear(maxwell3d, 1);
return Clear(1);
case MAXWELL3D_REG_INDEX(draw.begin):
return DrawBegin(maxwell3d);
return DrawBegin();
case MAXWELL3D_REG_INDEX(draw.end):
return DrawEnd(maxwell3d);
return DrawEnd();
case MAXWELL3D_REG_INDEX(vertex_buffer.first):
case MAXWELL3D_REG_INDEX(vertex_buffer.count):
case MAXWELL3D_REG_INDEX(index_buffer.first):
@@ -34,29 +33,33 @@ void Maxwell3D::DrawManager::ProcessMethodCall(Maxwell3D& maxwell3d, u32 method,
case MAXWELL3D_REG_INDEX(index_buffer32_first):
case MAXWELL3D_REG_INDEX(index_buffer16_first):
case MAXWELL3D_REG_INDEX(index_buffer8_first):
return DrawIndexSmall(maxwell3d, argument);
return DrawIndexSmall(argument);
case MAXWELL3D_REG_INDEX(draw_inline_index):
SetInlineIndexBuffer(maxwell3d, argument);
SetInlineIndexBuffer(argument);
break;
case MAXWELL3D_REG_INDEX(inline_index_2x16.even):
SetInlineIndexBuffer(maxwell3d, maxwell3d.regs.inline_index_2x16.even);
SetInlineIndexBuffer(maxwell3d, maxwell3d.regs.inline_index_2x16.odd);
SetInlineIndexBuffer(regs.inline_index_2x16.even);
SetInlineIndexBuffer(regs.inline_index_2x16.odd);
break;
case MAXWELL3D_REG_INDEX(inline_index_4x8.index0):
SetInlineIndexBuffer(maxwell3d, maxwell3d.regs.inline_index_4x8.index0);
SetInlineIndexBuffer(maxwell3d, maxwell3d.regs.inline_index_4x8.index1);
SetInlineIndexBuffer(maxwell3d, maxwell3d.regs.inline_index_4x8.index2);
SetInlineIndexBuffer(maxwell3d, maxwell3d.regs.inline_index_4x8.index3);
SetInlineIndexBuffer(regs.inline_index_4x8.index0);
SetInlineIndexBuffer(regs.inline_index_4x8.index1);
SetInlineIndexBuffer(regs.inline_index_4x8.index2);
SetInlineIndexBuffer(regs.inline_index_4x8.index3);
break;
case MAXWELL3D_REG_INDEX(vertex_array_instance_first):
DrawArrayInstanced(maxwell3d, maxwell3d.regs.vertex_array_instance_first.topology.Value(), maxwell3d.regs.vertex_array_instance_first.start.Value(), maxwell3d.regs.vertex_array_instance_first.count.Value(), false);
DrawArrayInstanced(regs.vertex_array_instance_first.topology.Value(),
regs.vertex_array_instance_first.start.Value(),
regs.vertex_array_instance_first.count.Value(), false);
break;
case MAXWELL3D_REG_INDEX(vertex_array_instance_subsequent): {
DrawArrayInstanced(maxwell3d, maxwell3d.regs.vertex_array_instance_subsequent.topology.Value(), maxwell3d.regs.vertex_array_instance_subsequent.start.Value(), maxwell3d.regs.vertex_array_instance_subsequent.count.Value(), true);
DrawArrayInstanced(regs.vertex_array_instance_subsequent.topology.Value(),
regs.vertex_array_instance_subsequent.start.Value(),
regs.vertex_array_instance_subsequent.count.Value(), true);
break;
}
case MAXWELL3D_REG_INDEX(draw_texture.src_y0): {
DrawTexture(maxwell3d);
DrawTexture();
break;
}
default:
@@ -64,87 +67,101 @@ void Maxwell3D::DrawManager::ProcessMethodCall(Maxwell3D& maxwell3d, u32 method,
}
}
void Maxwell3D::DrawManager::Clear(Maxwell3D& maxwell3d, u32 layer_count) {
if (maxwell3d.ShouldExecute()) {
maxwell3d.rasterizer->Clear(layer_count);
void DrawManager::Clear(u32 layer_count) {
if (maxwell3d->ShouldExecute()) {
maxwell3d->rasterizer->Clear(layer_count);
}
}
void Maxwell3D::DrawManager::DrawDeferred(Maxwell3D& maxwell3d) {
void DrawManager::DrawDeferred() {
if (draw_state.draw_mode != DrawMode::Instance || draw_state.instance_count == 0) {
return;
}
DrawEnd(maxwell3d, draw_state.instance_count + 1, true);
DrawEnd(draw_state.instance_count + 1, true);
draw_state.instance_count = 0;
}
void Maxwell3D::DrawManager::DrawArray(Maxwell3D& maxwell3d, Maxwell3D::Regs::PrimitiveTopology topology, u32 vertex_first, u32 vertex_count, u32 base_instance, u32 num_instances) {
void DrawManager::DrawArray(PrimitiveTopology topology, u32 vertex_first, u32 vertex_count,
u32 base_instance, u32 num_instances) {
draw_state.topology = topology;
draw_state.vertex_buffer.first = vertex_first;
draw_state.vertex_buffer.count = vertex_count;
draw_state.base_instance = base_instance;
ProcessDraw(maxwell3d, false, num_instances);
ProcessDraw(false, num_instances);
}
void Maxwell3D::DrawManager::DrawArrayInstanced(Maxwell3D& maxwell3d, Maxwell3D::Regs::PrimitiveTopology topology, u32 vertex_first, u32 vertex_count, bool subsequent) {
void DrawManager::DrawArrayInstanced(PrimitiveTopology topology, u32 vertex_first, u32 vertex_count,
bool subsequent) {
draw_state.topology = topology;
draw_state.vertex_buffer.first = vertex_first;
draw_state.vertex_buffer.count = vertex_count;
if (!subsequent) {
draw_state.instance_count = 1;
}
draw_state.base_instance = draw_state.instance_count - 1;
draw_state.draw_mode = DrawMode::Instance;
draw_state.instance_count++;
ProcessDraw(maxwell3d, false, 1);
ProcessDraw(false, 1);
}
void Maxwell3D::DrawManager::DrawIndex(Maxwell3D& maxwell3d, Maxwell3D::Regs::PrimitiveTopology topology, u32 index_first, u32 index_count, u32 base_index, u32 base_instance, u32 num_instances) {
void DrawManager::DrawIndex(PrimitiveTopology topology, u32 index_first, u32 index_count,
u32 base_index, u32 base_instance, u32 num_instances) {
const auto& regs{maxwell3d->regs};
draw_state.topology = topology;
draw_state.index_buffer = maxwell3d.regs.index_buffer;
draw_state.index_buffer = regs.index_buffer;
draw_state.index_buffer.first = index_first;
draw_state.index_buffer.count = index_count;
draw_state.base_index = base_index;
draw_state.base_instance = base_instance;
ProcessDraw(maxwell3d, true, num_instances);
ProcessDraw(true, num_instances);
}
void Maxwell3D::DrawManager::DrawArrayIndirect(Maxwell3D& maxwell3d, Maxwell3D::Regs::PrimitiveTopology topology) {
void DrawManager::DrawArrayIndirect(PrimitiveTopology topology) {
draw_state.topology = topology;
ProcessDrawIndirect(maxwell3d);
ProcessDrawIndirect();
}
void Maxwell3D::DrawManager::DrawIndexedIndirect(Maxwell3D& maxwell3d, Maxwell3D::Regs::PrimitiveTopology topology, u32 index_first, u32 index_count) {
void DrawManager::DrawIndexedIndirect(PrimitiveTopology topology, u32 index_first,
u32 index_count) {
const auto& regs{maxwell3d->regs};
draw_state.topology = topology;
draw_state.index_buffer = maxwell3d.regs.index_buffer;
draw_state.index_buffer = regs.index_buffer;
draw_state.index_buffer.first = index_first;
draw_state.index_buffer.count = index_count;
ProcessDrawIndirect(maxwell3d);
ProcessDrawIndirect();
}
void Maxwell3D::DrawManager::SetInlineIndexBuffer(Maxwell3D& maxwell3d, u32 index) {
draw_state.inline_index_draw_indexes.push_back(u8(index & 0x000000ff));
draw_state.inline_index_draw_indexes.push_back(u8((index & 0x0000ff00) >> 8));
draw_state.inline_index_draw_indexes.push_back(u8((index & 0x00ff0000) >> 16));
draw_state.inline_index_draw_indexes.push_back(u8((index & 0xff000000) >> 24));
void DrawManager::SetInlineIndexBuffer(u32 index) {
draw_state.inline_index_draw_indexes.push_back(static_cast<u8>(index & 0x000000ff));
draw_state.inline_index_draw_indexes.push_back(static_cast<u8>((index & 0x0000ff00) >> 8));
draw_state.inline_index_draw_indexes.push_back(static_cast<u8>((index & 0x00ff0000) >> 16));
draw_state.inline_index_draw_indexes.push_back(static_cast<u8>((index & 0xff000000) >> 24));
draw_state.draw_mode = DrawMode::InlineIndex;
}
void Maxwell3D::DrawManager::DrawBegin(Maxwell3D& maxwell3d) {
auto reset_instance_count = maxwell3d.regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::First;
auto increment_instance_count = maxwell3d.regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::Subsequent;
void DrawManager::DrawBegin() {
const auto& regs{maxwell3d->regs};
auto reset_instance_count = regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::First;
auto increment_instance_count =
regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::Subsequent;
if (reset_instance_count) {
DrawDeferred(maxwell3d);
DrawDeferred();
draw_state.instance_count = 0;
draw_state.draw_mode = DrawMode::General;
} else if (increment_instance_count) {
draw_state.instance_count++;
draw_state.draw_mode = DrawMode::Instance;
}
draw_state.topology = maxwell3d.regs.draw.topology;
draw_state.topology = regs.draw.topology;
}
void Maxwell3D::DrawManager::DrawEnd(Maxwell3D& maxwell3d, u32 instance_count, bool force_draw) {
void DrawManager::DrawEnd(u32 instance_count, bool force_draw) {
const auto& regs{maxwell3d->regs};
switch (draw_state.draw_mode) {
case DrawMode::Instance:
if (!force_draw) {
@@ -152,100 +169,119 @@ void Maxwell3D::DrawManager::DrawEnd(Maxwell3D& maxwell3d, u32 instance_count, b
}
[[fallthrough]];
case DrawMode::General:
draw_state.base_instance = maxwell3d.regs.global_base_instance_index;
draw_state.base_index = maxwell3d.regs.global_base_vertex_index;
draw_state.base_instance = regs.global_base_instance_index;
draw_state.base_index = regs.global_base_vertex_index;
if (draw_state.draw_indexed) {
draw_state.index_buffer = maxwell3d.regs.index_buffer;
ProcessDraw(maxwell3d, true, instance_count);
draw_state.index_buffer = regs.index_buffer;
ProcessDraw(true, instance_count);
} else {
draw_state.vertex_buffer = maxwell3d.regs.vertex_buffer;
ProcessDraw(maxwell3d, false, instance_count);
draw_state.vertex_buffer = regs.vertex_buffer;
ProcessDraw(false, instance_count);
}
draw_state.draw_indexed = false;
break;
case DrawMode::InlineIndex:
draw_state.base_instance = maxwell3d.regs.global_base_instance_index;
draw_state.base_index = maxwell3d.regs.global_base_vertex_index;
draw_state.index_buffer = maxwell3d.regs.index_buffer;
draw_state.index_buffer.count = u32(draw_state.inline_index_draw_indexes.size() / 4);
draw_state.base_instance = regs.global_base_instance_index;
draw_state.base_index = regs.global_base_vertex_index;
draw_state.index_buffer = regs.index_buffer;
draw_state.index_buffer.count =
static_cast<u32>(draw_state.inline_index_draw_indexes.size() / 4);
draw_state.index_buffer.format = Maxwell3D::Regs::IndexFormat::UnsignedInt;
maxwell3d.dirty.flags[VideoCommon::Dirty::IndexBuffer] = true;
ProcessDraw(maxwell3d, true, instance_count);
maxwell3d->dirty.flags[VideoCommon::Dirty::IndexBuffer] = true;
ProcessDraw(true, instance_count);
draw_state.inline_index_draw_indexes.clear();
break;
}
}
void Maxwell3D::DrawManager::DrawIndexSmall(Maxwell3D& maxwell3d, u32 argument) {
Maxwell3D::Regs::IndexBufferSmall index_small_params{argument};
draw_state.base_instance = maxwell3d.regs.global_base_instance_index;
draw_state.base_index = maxwell3d.regs.global_base_vertex_index;
draw_state.index_buffer = maxwell3d.regs.index_buffer;
void DrawManager::DrawIndexSmall(u32 argument) {
const auto& regs{maxwell3d->regs};
IndexBufferSmall index_small_params{argument};
draw_state.base_instance = regs.global_base_instance_index;
draw_state.base_index = regs.global_base_vertex_index;
draw_state.index_buffer = regs.index_buffer;
draw_state.index_buffer.first = index_small_params.first;
draw_state.index_buffer.count = index_small_params.count;
draw_state.topology = index_small_params.topology;
maxwell3d.dirty.flags[VideoCommon::Dirty::IndexBuffer] = true;
ProcessDraw(maxwell3d, true, 1);
maxwell3d->dirty.flags[VideoCommon::Dirty::IndexBuffer] = true;
ProcessDraw(true, 1);
}
void Maxwell3D::DrawManager::DrawTexture(Maxwell3D& maxwell3d) {
draw_texture_state.dst_x0 = f32(maxwell3d.regs.draw_texture.dst_x0) / 4096.f;
draw_texture_state.dst_y0 = f32(maxwell3d.regs.draw_texture.dst_y0) / 4096.f;
const auto dst_width = f32(maxwell3d.regs.draw_texture.dst_width) / 4096.f;
const auto dst_height = f32(maxwell3d.regs.draw_texture.dst_height) / 4096.f;
const bool lower_left{maxwell3d.regs.window_origin.mode != Maxwell3D::Regs::WindowOrigin::Mode::UpperLeft};
void DrawManager::DrawTexture() {
const auto& regs{maxwell3d->regs};
draw_texture_state.dst_x0 = static_cast<float>(regs.draw_texture.dst_x0) / 4096.f;
draw_texture_state.dst_y0 = static_cast<float>(regs.draw_texture.dst_y0) / 4096.f;
const auto dst_width = static_cast<float>(regs.draw_texture.dst_width) / 4096.f;
const auto dst_height = static_cast<float>(regs.draw_texture.dst_height) / 4096.f;
const bool lower_left{regs.window_origin.mode !=
Maxwell3D::Regs::WindowOrigin::Mode::UpperLeft};
if (lower_left) {
draw_texture_state.dst_y0 = f32(maxwell3d.regs.surface_clip.height) - draw_texture_state.dst_y0;
draw_texture_state.dst_y0 =
static_cast<f32>(regs.surface_clip.height) - draw_texture_state.dst_y0;
}
draw_texture_state.dst_x1 = draw_texture_state.dst_x0 + dst_width;
draw_texture_state.dst_y1 = draw_texture_state.dst_y0 + dst_height;
draw_texture_state.src_x0 = f32(maxwell3d.regs.draw_texture.src_x0) / 4096.f;
draw_texture_state.src_y0 = f32(maxwell3d.regs.draw_texture.src_y0) / 4096.f;
draw_texture_state.src_x1 = (f32(maxwell3d.regs.draw_texture.dx_du) / 4294967296.f) * dst_width + draw_texture_state.src_x0;
draw_texture_state.src_y1 = (f32(maxwell3d.regs.draw_texture.dy_dv) / 4294967296.f) * dst_height + draw_texture_state.src_y0;
draw_texture_state.src_sampler = maxwell3d.regs.draw_texture.src_sampler;
draw_texture_state.src_texture = maxwell3d.regs.draw_texture.src_texture;
maxwell3d.rasterizer->DrawTexture();
draw_texture_state.src_x0 = static_cast<float>(regs.draw_texture.src_x0) / 4096.f;
draw_texture_state.src_y0 = static_cast<float>(regs.draw_texture.src_y0) / 4096.f;
draw_texture_state.src_x1 =
(static_cast<float>(regs.draw_texture.dx_du) / 4294967296.f) * dst_width +
draw_texture_state.src_x0;
draw_texture_state.src_y1 =
(static_cast<float>(regs.draw_texture.dy_dv) / 4294967296.f) * dst_height +
draw_texture_state.src_y0;
draw_texture_state.src_sampler = regs.draw_texture.src_sampler;
draw_texture_state.src_texture = regs.draw_texture.src_texture;
maxwell3d->rasterizer->DrawTexture();
}
void Maxwell3D::DrawManager::UpdateTopology(Maxwell3D& maxwell3d) {
switch (maxwell3d.regs.primitive_topology_control) {
case Maxwell3D::Regs::PrimitiveTopologyControl::UseInBeginMethods:
void DrawManager::UpdateTopology() {
const auto& regs{maxwell3d->regs};
switch (regs.primitive_topology_control) {
case PrimitiveTopologyControl::UseInBeginMethods:
break;
case Maxwell3D::Regs::PrimitiveTopologyControl::UseSeparateState:
switch (maxwell3d.regs.topology_override) {
case Maxwell3D::Regs::PrimitiveTopologyOverride::None:
case PrimitiveTopologyControl::UseSeparateState:
switch (regs.topology_override) {
case PrimitiveTopologyOverride::None:
break;
case Maxwell3D::Regs::PrimitiveTopologyOverride::Points:
draw_state.topology = Maxwell3D::Regs::PrimitiveTopology::Points;
case PrimitiveTopologyOverride::Points:
draw_state.topology = PrimitiveTopology::Points;
break;
case Maxwell3D::Regs::PrimitiveTopologyOverride::Lines:
draw_state.topology = Maxwell3D::Regs::PrimitiveTopology::Lines;
case PrimitiveTopologyOverride::Lines:
draw_state.topology = PrimitiveTopology::Lines;
break;
case Maxwell3D::Regs::PrimitiveTopologyOverride::LineStrip:
draw_state.topology = Maxwell3D::Regs::PrimitiveTopology::LineStrip;
case PrimitiveTopologyOverride::LineStrip:
draw_state.topology = PrimitiveTopology::LineStrip;
break;
default:
draw_state.topology = Maxwell3D::Regs::PrimitiveTopology(maxwell3d.regs.topology_override);
draw_state.topology = static_cast<PrimitiveTopology>(regs.topology_override);
break;
}
break;
}
}
void Maxwell3D::DrawManager::ProcessDraw(Maxwell3D& maxwell3d, bool draw_indexed, u32 instance_count) {
LOG_TRACE(HW_GPU, "called, topology={}, count={}", draw_state.topology, draw_indexed ? draw_state.index_buffer.count : draw_state.vertex_buffer.count);
UpdateTopology(maxwell3d);
if (maxwell3d.ShouldExecute()) {
maxwell3d.rasterizer->Draw(draw_indexed, instance_count);
void DrawManager::ProcessDraw(bool draw_indexed, u32 instance_count) {
LOG_TRACE(HW_GPU, "called, topology={}, count={}", draw_state.topology,
draw_indexed ? draw_state.index_buffer.count : draw_state.vertex_buffer.count);
UpdateTopology();
if (maxwell3d->ShouldExecute()) {
maxwell3d->rasterizer->Draw(draw_indexed, instance_count);
}
}
void Maxwell3D::DrawManager::ProcessDrawIndirect(Maxwell3D& maxwell3d) {
LOG_TRACE(HW_GPU, "called, topology={}, is_indexed={}, includes_count={}, buffer_size={}, max_draw_count={}", draw_state.topology, indirect_state.is_indexed, indirect_state.include_count, indirect_state.buffer_size, indirect_state.max_draw_counts);
UpdateTopology(maxwell3d);
if (maxwell3d.ShouldExecute()) {
maxwell3d.rasterizer->DrawIndirect();
void DrawManager::ProcessDrawIndirect() {
LOG_TRACE(
HW_GPU,
"called, topology={}, is_indexed={}, includes_count={}, buffer_size={}, max_draw_count={}",
draw_state.topology, indirect_state.is_indexed, indirect_state.include_count,
indirect_state.buffer_size, indirect_state.max_draw_counts);
UpdateTopology();
if (maxwell3d->ShouldExecute()) {
maxwell3d->rasterizer->DrawIndirect();
}
}
} // namespace Tegra::Engines
+117
View File
@@ -0,0 +1,117 @@
// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
#pragma once
#include "common/common_types.h"
#include "video_core/engines/maxwell_3d.h"
namespace VideoCore {
class RasterizerInterface;
}
namespace Tegra::Engines {
using PrimitiveTopologyControl = Maxwell3D::Regs::PrimitiveTopologyControl;
using PrimitiveTopology = Maxwell3D::Regs::PrimitiveTopology;
using PrimitiveTopologyOverride = Maxwell3D::Regs::PrimitiveTopologyOverride;
using IndexBuffer = Maxwell3D::Regs::IndexBuffer;
using VertexBuffer = Maxwell3D::Regs::VertexBuffer;
using IndexBufferSmall = Maxwell3D::Regs::IndexBufferSmall;
class DrawManager {
public:
enum class DrawMode : u32 { General = 0, Instance, InlineIndex };
struct State {
PrimitiveTopology topology{};
DrawMode draw_mode{};
bool draw_indexed{};
u32 base_index{};
VertexBuffer vertex_buffer;
IndexBuffer index_buffer;
u32 base_instance{};
u32 instance_count{};
std::vector<u8> inline_index_draw_indexes;
};
struct DrawTextureState {
f32 dst_x0;
f32 dst_y0;
f32 dst_x1;
f32 dst_y1;
f32 src_x0;
f32 src_y0;
f32 src_x1;
f32 src_y1;
u32 src_sampler;
u32 src_texture;
};
struct IndirectParams {
bool is_byte_count;
bool is_indexed;
bool include_count;
GPUVAddr count_start_address;
GPUVAddr indirect_start_address;
size_t buffer_size;
size_t max_draw_counts;
size_t stride;
};
explicit DrawManager(Maxwell3D* maxwell_3d);
void ProcessMethodCall(u32 method, u32 argument);
void Clear(u32 layer_count);
void DrawDeferred();
void DrawArray(PrimitiveTopology topology, u32 vertex_first, u32 vertex_count,
u32 base_instance, u32 num_instances);
void DrawArrayInstanced(PrimitiveTopology topology, u32 vertex_first, u32 vertex_count,
bool subsequent);
void DrawIndex(PrimitiveTopology topology, u32 index_first, u32 index_count, u32 base_index,
u32 base_instance, u32 num_instances);
void DrawArrayIndirect(PrimitiveTopology topology);
void DrawIndexedIndirect(PrimitiveTopology topology, u32 index_first, u32 index_count);
const State& GetDrawState() const {
return draw_state;
}
const DrawTextureState& GetDrawTextureState() const {
return draw_texture_state;
}
IndirectParams& GetIndirectParams() {
return indirect_state;
}
const IndirectParams& GetIndirectParams() const {
return indirect_state;
}
private:
void SetInlineIndexBuffer(u32 index);
void DrawBegin();
void DrawEnd(u32 instance_count = 1, bool force_draw = false);
void DrawIndexSmall(u32 argument);
void DrawTexture();
void UpdateTopology();
void ProcessDraw(bool draw_indexed, u32 instance_count);
void ProcessDrawIndirect();
Maxwell3D* maxwell3d{};
State draw_state{};
DrawTextureState draw_texture_state{};
IndirectParams indirect_state{};
};
} // namespace Tegra::Engines
+7 -6
View File
@@ -13,7 +13,7 @@
#include "core/core.h"
#include "core/core_timing.h"
#include "video_core/dirty_flags.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/engines/draw_manager.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/gpu.h"
#include "video_core/memory_manager.h"
@@ -26,8 +26,7 @@ namespace Tegra::Engines {
constexpr u32 MacroRegistersStart = 0xE00;
Maxwell3D::Maxwell3D(Core::System& system_, MemoryManager& memory_manager_)
: draw_manager()
, system{system_}
: draw_manager{std::make_unique<DrawManager>(this)}, system{system_}
, memory_manager{memory_manager_}
#ifdef ARCHITECTURE_x86_64
, macro_engine(bool(Settings::values.disable_macro_jit))
@@ -374,7 +373,8 @@ void Maxwell3D::ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argume
case MAXWELL3D_REG_INDEX(sync_info):
return ProcessSyncPoint();
case MAXWELL3D_REG_INDEX(launch_dma):
return upload_state.ProcessExec(regs.launch_dma.memory_layout.Value() == Regs::LaunchDMA::Layout::Pitch);
return upload_state.ProcessExec(regs.launch_dma.memory_layout.Value() ==
Regs::LaunchDMA::Layout::Pitch);
case MAXWELL3D_REG_INDEX(inline_data):
upload_state.ProcessData(argument, is_last_call);
return;
@@ -386,7 +386,7 @@ void Maxwell3D::ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argume
case MAXWELL3D_REG_INDEX(tiled_cache_barrier):
return rasterizer->TiledCacheBarrier();
default:
draw_manager.ProcessMethodCall(*this, method, argument);
draw_manager->ProcessMethodCall(method, argument);
break;
}
}
@@ -401,7 +401,8 @@ void Maxwell3D::CallMacroMethod(u32 method, const std::vector<u32>& parameters)
// Execute the current macro.
macro_engine.Execute(*this, macro_positions[entry], parameters);
draw_manager.DrawDeferred(*this);
draw_manager->DrawDeferred();
}
void Maxwell3D::CallMethod(u32 method, u32 method_argument, bool is_last_call) {
+10 -63
View File
@@ -25,7 +25,6 @@
#include "video_core/gpu.h"
#include "video_core/macro.h"
#include "video_core/textures/texture.h"
#include "video_core/engines/maxwell_3d.h"
namespace Core {
class System;
@@ -41,6 +40,8 @@ class RasterizerInterface;
namespace Tegra::Engines {
class DrawManager;
/**
* This Engine is known as GF100_3D. Documentation can be found in:
* https://github.com/NVIDIA/open-gpu-doc/blob/master/classes/3d/clb197.h
@@ -542,7 +543,7 @@ public:
}
GPUVAddr StorageLimitAddress() const {
return (GPUVAddr{storage_limit_address_high} << 32) |
GPUVAddr{storage_limit_address_low};
GPUVAddr{storage_limit_address_low};
}
};
@@ -818,7 +819,7 @@ public:
u32 Map(std::size_t index) const {
const std::array<u32, NumRenderTargets> maps{target0, target1, target2, target3,
target4, target5, target6, target7};
target4, target5, target6, target7};
ASSERT(index < maps.size());
return maps[index];
}
@@ -1830,7 +1831,7 @@ public:
bool AnyEnabled() const {
return output0_enable || output1_enable || output2_enable || output3_enable ||
output4_enable || output5_enable || output6_enable || output7_enable;
output4_enable || output5_enable || output6_enable || output7_enable;
}
};
@@ -1869,7 +1870,7 @@ public:
bool AnyEnabled() const {
return plane0 || plane1 || plane2 || plane3 || plane4 || plane5 || plane6 ||
plane7;
plane7;
}
};
@@ -3022,7 +3023,8 @@ public:
u32 bindless_texture_const_buffer_slot; ///< 0x2608
u32 trap_handler; ///< 0x260C
INSERT_PADDING_BYTES_NOINIT(0x1F0);
std::array<std::array<StreamOutLayout, 32>, NumTransformFeedbackBuffers> stream_out_layout; ///< 0x2800
std::array<std::array<StreamOutLayout, 32>, NumTransformFeedbackBuffers>
stream_out_layout; ///< 0x2800
INSERT_PADDING_BYTES_NOINIT(0x93C);
ShaderPerformance shader_performance; ///< 0x333C
INSERT_PADDING_BYTES_NOINIT(0x18);
@@ -3033,62 +3035,6 @@ public:
};
// clang-format on
struct DrawManager {
enum class DrawMode : u32 { General = 0, Instance, InlineIndex };
struct State {
Maxwell3D::Regs::PrimitiveTopology topology{};
DrawMode draw_mode{};
bool draw_indexed{};
u32 base_index{};
Maxwell3D::Regs::VertexBuffer vertex_buffer;
Maxwell3D::Regs::IndexBuffer index_buffer;
u32 base_instance{};
u32 instance_count{};
std::vector<u8> inline_index_draw_indexes;
};
struct DrawTextureState {
f32 dst_x0;
f32 dst_y0;
f32 dst_x1;
f32 dst_y1;
f32 src_x0;
f32 src_y0;
f32 src_x1;
f32 src_y1;
u32 src_sampler;
u32 src_texture;
};
struct IndirectParams {
bool is_byte_count;
bool is_indexed;
bool include_count;
GPUVAddr count_start_address;
GPUVAddr indirect_start_address;
size_t buffer_size;
size_t max_draw_counts;
size_t stride;
};
void ProcessMethodCall(Maxwell3D& maxwell3d, u32 method, u32 argument);
void Clear(Maxwell3D& maxwell3d, u32 layer_count);
void DrawDeferred(Maxwell3D& maxwell3d);
void DrawArray(Maxwell3D& maxwell3d, Maxwell3D::Regs::PrimitiveTopology topology, u32 vertex_first, u32 vertex_count, u32 base_instance, u32 num_instances);
void DrawArrayInstanced(Maxwell3D& maxwell3d, Maxwell3D::Regs::PrimitiveTopology topology, u32 vertex_first, u32 vertex_count, bool subsequent);
void DrawIndex(Maxwell3D& maxwell3d, Maxwell3D::Regs::PrimitiveTopology topology, u32 index_first, u32 index_count, u32 base_index, u32 base_instance, u32 num_instances);
void DrawArrayIndirect(Maxwell3D& maxwell3d, Maxwell3D::Regs::PrimitiveTopology topology);
void DrawIndexedIndirect(Maxwell3D& maxwell3d, Maxwell3D::Regs::PrimitiveTopology topology, u32 index_first, u32 index_count);
void SetInlineIndexBuffer(Maxwell3D& maxwell3d, u32 index);
void DrawBegin(Maxwell3D& maxwell3d);
void DrawEnd(Maxwell3D& maxwell3d, u32 instance_count = 1, bool force_draw = false);
void DrawIndexSmall(Maxwell3D& maxwell3d, u32 argument);
void DrawTexture(Maxwell3D& maxwell3d);
void UpdateTopology(Maxwell3D& maxwell3d);
void ProcessDraw(Maxwell3D& maxwell3d, bool draw_indexed, u32 instance_count);
void ProcessDrawIndirect(Maxwell3D& maxwell3d);
State draw_state{};
DrawTextureState draw_texture_state{};
IndirectParams indirect_state{};
};
Regs regs{};
/// Store temporary hw register values, used by some calls to restore state after a operation
@@ -3156,7 +3102,8 @@ public:
Tables tables{};
} dirty;
DrawManager draw_manager;
std::unique_ptr<DrawManager> draw_manager;
friend class DrawManager;
GPUVAddr GetMacroAddress(size_t index) const {
return macro_addresses[index];
+1 -1
View File
@@ -89,7 +89,7 @@ struct GPU::Impl {
renderer = std::move(renderer_);
rasterizer = renderer->ReadRasterizer();
host1x.MemoryManager().BindInterface(rasterizer);
host1x.gmmu_manager.BindRasterizer(rasterizer);
host1x.GMMU().BindRasterizer(rasterizer);
}
/// Flush all current written commands into the host GPU for execution.
+44 -33
View File
@@ -19,45 +19,56 @@
namespace VideoCommon::GPUThread {
/// Runs the GPU thread
static void RunThread(std::stop_token stop_token, Core::System& system,
VideoCore::RendererBase& renderer, Core::Frontend::GraphicsContext& context,
Tegra::Control::Scheduler& scheduler, SynchState& state) {
Common::SetCurrentThreadName("GPU");
Common::SetCurrentThreadPriority(Common::ThreadPriority::Critical);
system.RegisterHostThread();
auto current_context = context.Acquire();
VideoCore::RasterizerInterface* const rasterizer = renderer.ReadRasterizer();
CommandDataContainer next;
while (!stop_token.stop_requested()) {
state.queue.PopWait(next, stop_token);
if (stop_token.stop_requested()) {
break;
}
if (auto* submit_list = std::get_if<SubmitListCommand>(&next.data)) {
scheduler.Push(submit_list->channel, std::move(submit_list->entries));
} else if (std::holds_alternative<GPUTickCommand>(next.data)) {
system.GPU().TickWork();
} else if (const auto* flush = std::get_if<FlushRegionCommand>(&next.data)) {
rasterizer->FlushRegion(flush->addr, flush->size);
} else if (const auto* invalidate = std::get_if<InvalidateRegionCommand>(&next.data)) {
rasterizer->OnCacheInvalidation(invalidate->addr, invalidate->size);
} else {
ASSERT(false);
}
state.signaled_fence.store(next.fence);
if (next.block) {
// We have to lock the write_lock to ensure that the condition_variable wait not get a
// race between the check and the lock itself.
std::scoped_lock lk{state.write_lock};
state.cv.notify_all();
}
}
}
ThreadManager::ThreadManager(Core::System& system_, bool is_async_)
: system{system_}, is_async{is_async_} {}
ThreadManager::~ThreadManager() = default;
void ThreadManager::StartThread(VideoCore::RendererBase& renderer, Core::Frontend::GraphicsContext& context, Tegra::Control::Scheduler& scheduler) {
void ThreadManager::StartThread(VideoCore::RendererBase& renderer,
Core::Frontend::GraphicsContext& context,
Tegra::Control::Scheduler& scheduler) {
rasterizer = renderer.ReadRasterizer();
thread = std::jthread([&](std::stop_token stop_token) {
Common::SetCurrentThreadName("GPU");
Common::SetCurrentThreadPriority(Common::ThreadPriority::Critical);
system.RegisterHostThread();
auto current_context = context.Acquire();
CommandDataContainer next;
while (!stop_token.stop_requested()) {
state.queue.PopWait(next, stop_token);
if (stop_token.stop_requested()) {
break;
}
if (auto* submit_list = std::get_if<SubmitListCommand>(&next.data)) {
scheduler.Push(submit_list->channel, std::move(submit_list->entries));
} else if (std::holds_alternative<GPUTickCommand>(next.data)) {
system.GPU().TickWork();
} else if (const auto* flush = std::get_if<FlushRegionCommand>(&next.data)) {
renderer.ReadRasterizer()->FlushRegion(flush->addr, flush->size);
} else if (const auto* invalidate = std::get_if<InvalidateRegionCommand>(&next.data)) {
renderer.ReadRasterizer()->OnCacheInvalidation(invalidate->addr, invalidate->size);
} else {
ASSERT(false);
}
state.signaled_fence.store(next.fence);
if (next.block) {
// We have to lock the write_lock to ensure that the condition_variable wait not get a
// race between the check and the lock itself.
std::scoped_lock lk{state.write_lock};
state.cv.notify_all();
}
}
});
thread = std::jthread(RunThread, std::ref(system), std::ref(renderer), std::ref(context),
std::ref(scheduler), std::ref(state));
}
void ThreadManager::SubmitList(s32 channel, Tegra::CommandList&& entries) {
+19 -17
View File
@@ -1,6 +1,3 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -16,12 +13,9 @@
namespace Tegra {
Codec::Codec(Host1x::Host1x& host1x_, const Host1x::NvdecCommon::NvdecRegisters& regs)
: host1x(host1x_)
, state{regs}
, h264_decoder(host1x_)
, vp8_decoder(host1x_)
, vp9_decoder(host1x_)
{}
: host1x(host1x_), state{regs}, h264_decoder(std::make_unique<Decoder::H264>(host1x)),
vp8_decoder(std::make_unique<Decoder::VP8>(host1x)),
vp9_decoder(std::make_unique<Decoder::VP9>(host1x)) {}
Codec::~Codec() = default;
@@ -38,11 +32,13 @@ void Codec::SetTargetCodec(Host1x::NvdecCommon::VideoCodec codec) {
void Codec::Decode() {
const bool is_first_frame = !initialized;
if (is_first_frame)
if (is_first_frame) {
Initialize();
}
if (!initialized)
if (!initialized) {
return;
}
// Assemble bitstream.
bool vp9_hidden_frame = false;
@@ -50,13 +46,13 @@ void Codec::Decode() {
const auto packet_data = [&]() {
switch (current_codec) {
case Tegra::Host1x::NvdecCommon::VideoCodec::H264:
return h264_decoder.ComposeFrame(state, &configuration_size, is_first_frame);
return h264_decoder->ComposeFrame(state, &configuration_size, is_first_frame);
case Tegra::Host1x::NvdecCommon::VideoCodec::VP8:
return vp8_decoder.ComposeFrame(state);
return vp8_decoder->ComposeFrame(state);
case Tegra::Host1x::NvdecCommon::VideoCodec::VP9:
vp9_decoder.ComposeFrame(state);
vp9_hidden_frame = vp9_decoder.WasFrameHidden();
return vp9_decoder.GetFrameBytes();
vp9_decoder->ComposeFrame(state);
vp9_hidden_frame = vp9_decoder->WasFrameHidden();
return vp9_decoder->GetFrameBytes();
default:
ASSERT(false);
return std::span<const u8>{};
@@ -85,13 +81,19 @@ void Codec::Decode() {
std::unique_ptr<FFmpeg::Frame> Codec::GetCurrentFrame() {
// Sometimes VIC will request more frames than have been decoded.
// in this case, return a blank frame and don't overwrite previous data.
if (frames.empty())
if (frames.empty()) {
return {};
}
auto frame = std::move(frames.front());
frames.pop();
return frame;
}
Host1x::NvdecCommon::VideoCodec Codec::GetCurrentCodec() const {
return current_codec;
}
std::string_view Codec::GetCurrentCodecName() const {
switch (current_codec) {
case Host1x::NvdecCommon::VideoCodec::None:
+14 -10
View File
@@ -1,6 +1,3 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -11,14 +8,17 @@
#include <string_view>
#include <queue>
#include "common/common_types.h"
#include "video_core/host1x/codecs/h264.h"
#include "video_core/host1x/codecs/vp8.h"
#include "video_core/host1x/codecs/vp9.h"
#include "video_core/host1x/ffmpeg.h"
#include "video_core/host1x/ffmpeg/ffmpeg.h"
#include "video_core/host1x/nvdec_common.h"
namespace Tegra {
namespace Decoder {
class H264;
class VP8;
class VP9;
} // namespace Decoder
namespace Host1x {
class Host1x;
} // namespace Host1x
@@ -40,6 +40,9 @@ public:
/// Returns next decoded frame
[[nodiscard]] std::unique_ptr<FFmpeg::Frame> GetCurrentFrame();
/// Returns the value of current_codec
[[nodiscard]] Host1x::NvdecCommon::VideoCodec GetCurrentCodec() const;
/// Return name of the current codec
[[nodiscard]] std::string_view GetCurrentCodecName() const;
@@ -50,9 +53,10 @@ private:
Host1x::Host1x& host1x;
const Host1x::NvdecCommon::NvdecRegisters& state;
Decoders::H264 h264_decoder;
Decoders::VP8 vp8_decoder;
Decoders::VP9 vp9_decoder;
std::unique_ptr<Decoder::H264> h264_decoder;
std::unique_ptr<Decoder::VP8> vp8_decoder;
std::unique_ptr<Decoder::VP9> vp9_decoder;
std::queue<std::unique_ptr<FFmpeg::Frame>> frames{};
};
+11 -12
View File
@@ -1,4 +1,4 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
@@ -12,11 +12,10 @@
namespace Tegra {
Decoder::Decoder(Host1x::Host1x& host1x_, s32 id_, const Host1x::NvdecCommon::NvdecRegisters& regs_)
: host1x(host1x_)
, regs{regs_}
, id{id_}
{}
Decoder::Decoder(Host1x::Host1x& host1x_, s32 id_, const Host1x::NvdecCommon::NvdecRegisters& regs_,
Host1x::FrameQueue& frame_queue_)
: host1x(host1x_), memory_manager{host1x.GMMU()}, regs{regs_}, id{id_}, frame_queue{
frame_queue_} {}
Decoder::~Decoder() = default;
@@ -54,11 +53,11 @@ void Decoder::Decode() {
}
if (UsingDecodeOrder()) {
host1x.frame_queue.PushDecodeOrder(id, luma_top, std::move(frame));
host1x.frame_queue.PushDecodeOrder(id, luma_bottom, std::move(frame_copy));
frame_queue.PushDecodeOrder(id, luma_top, std::move(frame));
frame_queue.PushDecodeOrder(id, luma_bottom, std::move(frame_copy));
} else {
host1x.frame_queue.PushPresentOrder(id, luma_top, std::move(frame));
host1x.frame_queue.PushPresentOrder(id, luma_bottom, std::move(frame_copy));
frame_queue.PushPresentOrder(id, luma_top, std::move(frame));
frame_queue.PushPresentOrder(id, luma_bottom, std::move(frame_copy));
}
} else {
auto [luma_offset, chroma_offset] = GetProgressiveOffsets();
@@ -69,9 +68,9 @@ void Decoder::Decode() {
}
if (UsingDecodeOrder()) {
host1x.frame_queue.PushDecodeOrder(id, luma_offset, std::move(frame));
frame_queue.PushDecodeOrder(id, luma_offset, std::move(frame));
} else {
host1x.frame_queue.PushPresentOrder(id, luma_offset, std::move(frame));
frame_queue.PushPresentOrder(id, luma_offset, std::move(frame));
}
}
}
+15 -5
View File
@@ -14,7 +14,7 @@
#include <queue>
#include "common/common_types.h"
#include "video_core/host1x/ffmpeg.h"
#include "video_core/host1x/ffmpeg/ffmpeg.h"
#include "video_core/host1x/nvdec_common.h"
namespace Tegra {
@@ -35,23 +35,33 @@ public:
return decode_api.UsingDecodeOrder();
}
/// Returns the value of current_codec
[[nodiscard]] Host1x::NvdecCommon::VideoCodec GetCurrentCodec() const {
return codec;
}
/// Return name of the current codec
[[nodiscard]] virtual std::string_view GetCurrentCodecName() const = 0;
protected:
explicit Decoder(Host1x::Host1x& host1x, s32 id, const Host1x::NvdecCommon::NvdecRegisters& regs);
explicit Decoder(Host1x::Host1x& host1x, s32 id,
const Host1x::NvdecCommon::NvdecRegisters& regs,
Host1x::FrameQueue& frame_queue);
virtual std::span<const u8> ComposeFrame() = 0;
virtual std::tuple<u64, u64> GetProgressiveOffsets() = 0;
virtual std::tuple<u64, u64, u64, u64> GetInterlacedOffsets() = 0;
virtual bool IsInterlaced() = 0;
FFmpeg::DecodeApi decode_api;
Host1x::Host1x& host1x;
Tegra::MemoryManager& memory_manager;
const Host1x::NvdecCommon::NvdecRegisters& regs;
s32 id;
bool initialized : 1 = false;
bool vp9_hidden_frame : 1 = false;
Host1x::FrameQueue& frame_queue;
Host1x::NvdecCommon::VideoCodec codec;
FFmpeg::DecodeApi decode_api;
bool initialized{};
bool vp9_hidden_frame{};
};
} // namespace Tegra
+55 -42
View File
@@ -1,4 +1,4 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
@@ -14,11 +14,25 @@
#include "video_core/memory_manager.h"
namespace Tegra::Decoders {
namespace {
// ZigZag LUTs from libavcodec.
constexpr std::array<u8, 64> zig_zag_direct{
0, 1, 8, 16, 9, 2, 3, 10, 17, 24, 32, 25, 18, 11, 4, 5, 12, 19, 26, 33, 40, 48,
41, 34, 27, 20, 13, 6, 7, 14, 21, 28, 35, 42, 49, 56, 57, 50, 43, 36, 29, 22, 15, 23,
30, 37, 44, 51, 58, 59, 52, 45, 38, 31, 39, 46, 53, 60, 61, 54, 47, 55, 62, 63,
};
H264::H264(Host1x::Host1x& host1x_, const Host1x::NvdecCommon::NvdecRegisters& regs_, s32 id_)
: Decoder{host1x_, id_, regs_}
{
initialized = decode_api.Initialize(Host1x::NvdecCommon::VideoCodec::H264);
constexpr std::array<u8, 16> zig_zag_scan{
0 + 0 * 4, 1 + 0 * 4, 0 + 1 * 4, 0 + 2 * 4, 1 + 1 * 4, 2 + 0 * 4, 3 + 0 * 4, 2 + 1 * 4,
1 + 2 * 4, 0 + 3 * 4, 1 + 3 * 4, 2 + 2 * 4, 3 + 1 * 4, 3 + 2 * 4, 2 + 3 * 4, 3 + 3 * 4,
};
} // Anonymous namespace
H264::H264(Host1x::Host1x& host1x_, const Host1x::NvdecCommon::NvdecRegisters& regs_, s32 id_,
Host1x::FrameQueue& frame_queue_)
: Decoder{host1x_, id_, regs_, frame_queue_} {
codec = Host1x::NvdecCommon::VideoCodec::H264;
initialized = decode_api.Initialize(codec);
}
H264::~H264() = default;
@@ -51,11 +65,14 @@ bool H264::IsInterlaced() {
}
std::span<const u8> H264::ComposeFrame() {
host1x.gmmu_manager.ReadBlock(regs.picture_info_offset.Address(), &current_context, sizeof(H264DecoderContext));
memory_manager.ReadBlock(regs.picture_info_offset.Address(), &current_context,
sizeof(H264DecoderContext));
const s64 frame_number = current_context.h264_parameter_set.frame_number.Value();
if (!is_first_frame && frame_number != 0) {
frame_scratch.resize_destructive(current_context.stream_len);
host1x.gmmu_manager.ReadBlock(regs.frame_bitstream_offset.Address(), frame_scratch.data(), frame_scratch.size());
memory_manager.ReadBlock(regs.frame_bitstream_offset.Address(), frame_scratch.data(),
frame_scratch.size());
return frame_scratch;
}
@@ -157,13 +174,15 @@ std::span<const u8> H264::ComposeFrame() {
for (s32 index = 0; index < 6; index++) {
writer.WriteBit(true);
writer.WriteScalingList(current_context.weight_scale_4x4, index * 16, 16);
std::span<const u8> matrix{current_context.weight_scale_4x4};
writer.WriteScalingList(scan_scratch, matrix, index * 16, 16);
}
if (current_context.h264_parameter_set.transform_8x8_mode_flag) {
for (s32 index = 0; index < 2; index++) {
writer.WriteBit(true);
writer.WriteScalingList(current_context.weight_scale_8x8, index * 64, 64);
std::span<const u8> matrix{current_context.weight_scale_8x8};
writer.WriteScalingList(scan_scratch, matrix, index * 64, 64);
}
}
@@ -177,7 +196,11 @@ std::span<const u8> H264::ComposeFrame() {
const auto& encoded_header = writer.GetByteArray();
frame_scratch.resize(encoded_header.size() + current_context.stream_len);
std::memcpy(frame_scratch.data(), encoded_header.data(), encoded_header.size());
host1x.gmmu_manager.ReadBlock(regs.frame_bitstream_offset.Address(), frame_scratch.data() + encoded_header.size(), current_context.stream_len);
memory_manager.ReadBlock(regs.frame_bitstream_offset.Address(),
frame_scratch.data() + encoded_header.size(),
current_context.stream_len);
return frame_scratch;
}
@@ -206,37 +229,23 @@ void H264BitWriter::WriteBit(bool state) {
WriteBits(state ? 1 : 0, 1);
}
void H264BitWriter::WriteScalingList(std::span<const u8> list, s32 start, s32 count) {
void H264BitWriter::WriteScalingList(Common::ScratchBuffer<u8>& scan, std::span<const u8> list,
s32 start, s32 count) {
scan.resize_destructive(count);
if (count == 16) {
u8 last_scale = 8;
for (s32 index = 0; index < count; index++) {
// libavcodec has a zig zag LUT, but we dont need it, just use a magic
// constant which is a packing of 4 bits for each component of the table
const u8 value = list[start + ((0xfeb7adc963258410 >> (index * 4)) & 0xf)];
const s32 delta_scale = s32(value - last_scale);
WriteSe(delta_scale);
last_scale = value;
}
std::memcpy(scan.data(), zig_zag_scan.data(), scan.size());
} else {
// ZigZag LUTs from libavcodec: this is the famous zigzag pattern found in the ffmpeg logo itself!
static constexpr std::array<u8, 64> scan{
0, 1, 8, 16, 9, 2, 3, 10,
17, 24, 32, 25, 18, 11, 4,
5, 12, 19, 26, 33, 40, 48,
41, 34, 27, 20, 13, 6, 7,
14, 21, 28, 35, 42, 49, 56,
57, 50, 43, 36, 29, 22, 15,
23, 30, 37, 44, 51, 58, 59,
52, 45, 38, 31, 39, 46, 53,
60, 61, 54, 47, 55, 62, 63,
};
u8 last_scale = 8;
for (s32 index = 0; index < count; index++) {
const u8 value = list[start + scan[index]];
const s32 delta_scale = s32(value - last_scale);
WriteSe(delta_scale);
last_scale = value;
}
std::memcpy(scan.data(), zig_zag_direct.data(), scan.size());
}
u8 last_scale = 8;
for (s32 index = 0; index < count; index++) {
const u8 value = list[start + scan[index]];
const s32 delta_scale = static_cast<s32>(value - last_scale);
WriteSe(delta_scale);
last_scale = value;
}
}
@@ -277,15 +286,19 @@ void H264BitWriter::WriteBits(s32 value, s32 bit_count) {
void H264BitWriter::WriteExpGolombCodedInt(s32 value) {
const s32 sign = value <= 0 ? 0 : 1;
if (!sign) value = -value;
WriteExpGolombCodedUInt((value << 1) - sign);
if (value < 0) {
value = -value;
}
value = (value << 1) - sign;
WriteExpGolombCodedUInt(value);
}
void H264BitWriter::WriteExpGolombCodedUInt(u32 value) {
const s32 size = 32 - std::countl_zero(value + 1);
WriteBits(1, size);
value -= (1U << (size - 1)) - 1;
WriteBits(s32(value), size - 1);
WriteBits(static_cast<s32>(value), size - 1);
}
s32 H264BitWriter::GetFreeBufferBits() {
+187 -9
View File
@@ -1,6 +1,3 @@
// 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
@@ -9,10 +6,11 @@
#include <span>
#include <vector>
#include "common/bit_field.h"
#include "common/common_funcs.h"
#include "common/common_types.h"
#include "common/scratch_buffer.h"
#include "video_core/host1x/codecs/decoder.h"
#include "video_core/host1x/codec_types.h"
#include "video_core/host1x/nvdec_common.h"
namespace Tegra {
@@ -42,7 +40,8 @@ public:
/// Based on section 7.3.2.1.1.1 and Table 7-4 in the H.264 specification
/// Writes the scaling matrices of the sream
void WriteScalingList(std::span<const u8> list, s32 start, s32 count);
void WriteScalingList(Common::ScratchBuffer<u8>& scan, std::span<const u8> list, s32 start,
s32 count);
/// Return the bitstream as a vector.
[[nodiscard]] std::vector<u8>& GetByteArray();
@@ -62,9 +61,188 @@ private:
std::vector<u8> byte_array;
};
struct Offset {
constexpr u32 Address() const noexcept {
return offset << 8;
}
private:
u32 offset;
};
static_assert(std::is_trivial_v<Offset>, "Offset must be trivial");
static_assert(sizeof(Offset) == 0x4, "Offset has the wrong size!");
struct H264ParameterSet {
s32 log2_max_pic_order_cnt_lsb_minus4; ///< 0x00
s32 delta_pic_order_always_zero_flag; ///< 0x04
s32 frame_mbs_only_flag; ///< 0x08
u32 pic_width_in_mbs; ///< 0x0C
u32 frame_height_in_mbs; ///< 0x10
union { ///< 0x14
BitField<0, 2, u32> tile_format;
BitField<2, 3, u32> gob_height;
BitField<5, 27, u32> reserved_surface_format;
};
u32 entropy_coding_mode_flag; ///< 0x18
s32 pic_order_present_flag; ///< 0x1C
s32 num_refidx_l0_default_active; ///< 0x20
s32 num_refidx_l1_default_active; ///< 0x24
s32 deblocking_filter_control_present_flag; ///< 0x28
s32 redundant_pic_cnt_present_flag; ///< 0x2C
u32 transform_8x8_mode_flag; ///< 0x30
u32 pitch_luma; ///< 0x34
u32 pitch_chroma; ///< 0x38
Offset luma_top_offset; ///< 0x3C
Offset luma_bot_offset; ///< 0x40
Offset luma_frame_offset; ///< 0x44
Offset chroma_top_offset; ///< 0x48
Offset chroma_bot_offset; ///< 0x4C
Offset chroma_frame_offset; ///< 0x50
u32 hist_buffer_size; ///< 0x54
union { ///< 0x58
union {
BitField<0, 1, u64> mbaff_frame;
BitField<1, 1, u64> direct_8x8_inference;
BitField<2, 1, u64> weighted_pred;
BitField<3, 1, u64> constrained_intra_pred;
BitField<4, 1, u64> ref_pic;
BitField<5, 1, u64> field_pic;
BitField<6, 1, u64> bottom_field;
BitField<7, 1, u64> second_field;
} flags;
BitField<8, 4, u64> log2_max_frame_num_minus4;
BitField<12, 2, u64> chroma_format_idc;
BitField<14, 2, u64> pic_order_cnt_type;
BitField<16, 6, s64> pic_init_qp_minus26;
BitField<22, 5, s64> chroma_qp_index_offset;
BitField<27, 5, s64> second_chroma_qp_index_offset;
BitField<32, 2, u64> weighted_bipred_idc;
BitField<34, 7, u64> curr_pic_idx;
BitField<41, 5, u64> curr_col_idx;
BitField<46, 16, u64> frame_number;
BitField<62, 1, u64> frame_surfaces;
BitField<63, 1, u64> output_memory_layout;
};
};
static_assert(sizeof(H264ParameterSet) == 0x60, "H264ParameterSet is an invalid size");
#define ASSERT_POSITION(field_name, position) \
static_assert(offsetof(H264ParameterSet, field_name) == position, \
"Field " #field_name " has invalid position")
ASSERT_POSITION(log2_max_pic_order_cnt_lsb_minus4, 0x00);
ASSERT_POSITION(delta_pic_order_always_zero_flag, 0x04);
ASSERT_POSITION(frame_mbs_only_flag, 0x08);
ASSERT_POSITION(pic_width_in_mbs, 0x0C);
ASSERT_POSITION(frame_height_in_mbs, 0x10);
ASSERT_POSITION(tile_format, 0x14);
ASSERT_POSITION(entropy_coding_mode_flag, 0x18);
ASSERT_POSITION(pic_order_present_flag, 0x1C);
ASSERT_POSITION(num_refidx_l0_default_active, 0x20);
ASSERT_POSITION(num_refidx_l1_default_active, 0x24);
ASSERT_POSITION(deblocking_filter_control_present_flag, 0x28);
ASSERT_POSITION(redundant_pic_cnt_present_flag, 0x2C);
ASSERT_POSITION(transform_8x8_mode_flag, 0x30);
ASSERT_POSITION(pitch_luma, 0x34);
ASSERT_POSITION(pitch_chroma, 0x38);
ASSERT_POSITION(luma_top_offset, 0x3C);
ASSERT_POSITION(luma_bot_offset, 0x40);
ASSERT_POSITION(luma_frame_offset, 0x44);
ASSERT_POSITION(chroma_top_offset, 0x48);
ASSERT_POSITION(chroma_bot_offset, 0x4C);
ASSERT_POSITION(chroma_frame_offset, 0x50);
ASSERT_POSITION(hist_buffer_size, 0x54);
ASSERT_POSITION(flags, 0x58);
#undef ASSERT_POSITION
struct DpbEntry {
union {
BitField<0, 7, u32> index;
BitField<7, 5, u32> col_idx;
BitField<12, 2, u32> state;
BitField<14, 1, u32> is_long_term;
BitField<15, 1, u32> non_existing;
BitField<16, 1, u32> is_field;
BitField<17, 4, u32> top_field_marking;
BitField<21, 4, u32> bottom_field_marking;
BitField<25, 1, u32> output_memory_layout;
BitField<26, 6, u32> reserved;
} flags;
std::array<u32, 2> field_order_cnt;
u32 frame_idx;
};
static_assert(sizeof(DpbEntry) == 0x10, "DpbEntry has the wrong size!");
struct DisplayParam {
union {
BitField<0, 1, u32> enable_tf_output;
BitField<1, 1, u32> vc1_map_y_flag;
BitField<2, 3, u32> map_y_value;
BitField<5, 1, u32> vc1_map_uv_flag;
BitField<6, 3, u32> map_uv_value;
BitField<9, 8, u32> out_stride;
BitField<17, 3, u32> tiling_format;
BitField<20, 1, u32> output_structure; // 0=frame, 1=field
BitField<21, 11, u32> reserved0;
};
std::array<s32, 2> output_top;
std::array<s32, 2> output_bottom;
union {
BitField<0, 1, u32> enable_histogram;
BitField<1, 12, u32> histogram_start_x;
BitField<13, 12, u32> histogram_start_y;
BitField<25, 7, u32> reserved1;
};
union {
BitField<0, 12, u32> histogram_end_x;
BitField<12, 12, u32> histogram_end_y;
BitField<24, 8, u32> reserved2;
};
};
static_assert(sizeof(DisplayParam) == 0x1C, "DisplayParam has the wrong size!");
struct H264DecoderContext {
INSERT_PADDING_WORDS_NOINIT(13); ///< 0x0000
std::array<u8, 16> eos; ///< 0x0034
u8 explicit_eos_present_flag; ///< 0x0044
u8 hint_dump_en; ///< 0x0045
INSERT_PADDING_BYTES_NOINIT(2); ///< 0x0046
u32 stream_len; ///< 0x0048
u32 slice_count; ///< 0x004C
u32 mbhist_buffer_size; ///< 0x0050
u32 gptimer_timeout_value; ///< 0x0054
H264ParameterSet h264_parameter_set; ///< 0x0058
std::array<s32, 2> curr_field_order_cnt; ///< 0x00B8
std::array<DpbEntry, 16> dpb; ///< 0x00C0
std::array<u8, 0x60> weight_scale_4x4; ///< 0x01C0
std::array<u8, 0x80> weight_scale_8x8; ///< 0x0220
std::array<u8, 2> num_inter_view_refs_lX; ///< 0x02A0
std::array<u8, 14> reserved2; ///< 0x02A2
std::array<std::array<s8, 16>, 2> inter_view_refidx_lX; ///< 0x02B0
union { ///< 0x02D0
BitField<0, 1, u32> lossless_ipred8x8_filter_enable;
BitField<1, 1, u32> qpprime_y_zero_transform_bypass_flag;
BitField<2, 30, u32> reserved3;
};
DisplayParam display_param; ///< 0x02D4
std::array<u32, 3> reserved4; ///< 0x02F0
};
static_assert(sizeof(H264DecoderContext) == 0x2FC, "H264DecoderContext is an invalid size");
#define ASSERT_POSITION(field_name, position) \
static_assert(offsetof(H264DecoderContext, field_name) == position, \
"Field " #field_name " has invalid position")
ASSERT_POSITION(stream_len, 0x48);
ASSERT_POSITION(h264_parameter_set, 0x58);
ASSERT_POSITION(dpb, 0xC0);
ASSERT_POSITION(weight_scale_4x4, 0x1C0);
#undef ASSERT_POSITION
class H264 final : public Decoder {
public:
explicit H264(Host1x::Host1x& host1x, const Host1x::NvdecCommon::NvdecRegisters& regs, s32 id);
explicit H264(Host1x::Host1x& host1x, const Host1x::NvdecCommon::NvdecRegisters& regs, s32 id,
Host1x::FrameQueue& frame_queue);
~H264() override;
H264(const H264&) = delete;
@@ -85,10 +263,10 @@ public:
}
private:
H264DecoderContext current_context{};
std::array<u8, 64> scan_scratch;
Common::ScratchBuffer<u8> frame_scratch;
bool is_first_frame{true};
Common::ScratchBuffer<u8> frame_scratch;
Common::ScratchBuffer<u8> scan_scratch;
H264DecoderContext current_context{};
};
} // namespace Decoders
+26 -23
View File
@@ -1,6 +1,3 @@
// 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
@@ -11,10 +8,11 @@
#include "video_core/memory_manager.h"
namespace Tegra::Decoders {
VP8::VP8(Host1x::Host1x& host1x_, const Host1x::NvdecCommon::NvdecRegisters& regs_, s32 id_)
: Decoder{host1x_, id_, regs_}
{
initialized = decode_api.Initialize(Host1x::NvdecCommon::VideoCodec::VP8);
VP8::VP8(Host1x::Host1x& host1x_, const Host1x::NvdecCommon::NvdecRegisters& regs_, s32 id_,
Host1x::FrameQueue& frame_queue_)
: Decoder{host1x_, id_, regs_, frame_queue_} {
codec = Host1x::NvdecCommon::VideoCodec::VP8;
initialized = decode_api.Initialize(codec);
}
VP8::~VP8() = default;
@@ -27,30 +25,35 @@ std::tuple<u64, u64> VP8::GetProgressiveOffsets() {
std::tuple<u64, u64, u64, u64> VP8::GetInterlacedOffsets() {
auto luma_top{regs.surface_luma_offsets[static_cast<u32>(Vp8SurfaceIndex::Current)].Address()};
auto luma_bottom = regs.surface_luma_offsets[u32(Vp8SurfaceIndex::Current)].Address();
auto chroma_top = regs.surface_chroma_offsets[u32(Vp8SurfaceIndex::Current)].Address();
auto chroma_bottom = regs.surface_chroma_offsets[u32(Vp8SurfaceIndex::Current)].Address();
auto luma_bottom{
regs.surface_luma_offsets[static_cast<u32>(Vp8SurfaceIndex::Current)].Address()};
auto chroma_top{
regs.surface_chroma_offsets[static_cast<u32>(Vp8SurfaceIndex::Current)].Address()};
auto chroma_bottom{
regs.surface_chroma_offsets[static_cast<u32>(Vp8SurfaceIndex::Current)].Address()};
return {luma_top, luma_bottom, chroma_top, chroma_bottom};
}
std::span<const u8> VP8::ComposeFrame() {
host1x.gmmu_manager.ReadBlock(regs.picture_info_offset.Address(), &current_context, sizeof(VP8PictureInfo));
memory_manager.ReadBlock(regs.picture_info_offset.Address(), &current_context,
sizeof(VP8PictureInfo));
const bool is_key_frame = current_context.key_frame == 1u;
const auto bitstream_size = size_t(current_context.vld_buffer_size);
const auto bitstream_size = static_cast<size_t>(current_context.vld_buffer_size);
const size_t header_size = is_key_frame ? 10u : 3u;
frame_scratch.resize(header_size + bitstream_size);
// Based on page 30 of the VP8 specification.
// https://datatracker.ietf.org/doc/rfc6386/
frame_scratch[0] = is_key_frame ? 0u : 1u; // 1-bit frame type (0: keyframe, 1: interframes).
frame_scratch[0] |= u8((current_context.version & 7u) << 1u); // 3-bit version number
frame_scratch[0] |= u8(1u << 4u); // 1-bit show_frame flag
frame_scratch[0] |=
static_cast<u8>((current_context.version & 7u) << 1u); // 3-bit version number
frame_scratch[0] |= static_cast<u8>(1u << 4u); // 1-bit show_frame flag
// The next 19-bits are the first partition size
frame_scratch[0] |= u8((current_context.first_part_size & 7u) << 5u);
frame_scratch[1] = u8((current_context.first_part_size & 0x7f8u) >> 3u);
frame_scratch[2] = u8((current_context.first_part_size & 0x7f800u) >> 11u);
frame_scratch[0] |= static_cast<u8>((current_context.first_part_size & 7u) << 5u);
frame_scratch[1] = static_cast<u8>((current_context.first_part_size & 0x7f8u) >> 3u);
frame_scratch[2] = static_cast<u8>((current_context.first_part_size & 0x7f800u) >> 11u);
if (is_key_frame) {
frame_scratch[3] = 0x9du;
@@ -58,15 +61,15 @@ std::span<const u8> VP8::ComposeFrame() {
frame_scratch[5] = 0x2au;
// TODO(ameerj): Horizontal/Vertical Scale
// 16 bits: (2 bits Horizontal Scale << 14) | Width (14 bits)
frame_scratch[6] = u8(current_context.frame_width & 0xff);
frame_scratch[7] = u8(((current_context.frame_width >> 8) & 0x3f));
frame_scratch[6] = static_cast<u8>(current_context.frame_width & 0xff);
frame_scratch[7] = static_cast<u8>(((current_context.frame_width >> 8) & 0x3f));
// 16 bits:(2 bits Vertical Scale << 14) | Height (14 bits)
frame_scratch[8] = u8(current_context.frame_height & 0xff);
frame_scratch[9] = u8(((current_context.frame_height >> 8) & 0x3f));
frame_scratch[8] = static_cast<u8>(current_context.frame_height & 0xff);
frame_scratch[9] = static_cast<u8>(((current_context.frame_height >> 8) & 0x3f));
}
const u64 bitstream_offset = regs.frame_bitstream_offset.Address();
host1x.gmmu_manager.ReadBlock(bitstream_offset, frame_scratch.data() + header_size, bitstream_size);
memory_manager.ReadBlock(bitstream_offset, frame_scratch.data() + header_size, bitstream_size);
return frame_scratch;
}
+46 -6
View File
@@ -1,6 +1,3 @@
// 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
@@ -14,7 +11,6 @@
#include "common/scratch_buffer.h"
#include "video_core/host1x/codecs/decoder.h"
#include "video_core/host1x/nvdec_common.h"
#include "video_core/host1x/codec_types.h"
namespace Tegra {
@@ -32,7 +28,8 @@ enum class Vp8SurfaceIndex : u32 {
class VP8 final : public Decoder {
public:
explicit VP8(Host1x::Host1x& host1x, const Host1x::NvdecCommon::NvdecRegisters& regs, s32 id);
explicit VP8(Host1x::Host1x& host1x, const Host1x::NvdecCommon::NvdecRegisters& regs, s32 id,
Host1x::FrameQueue& frame_queue);
~VP8() override;
VP8(const VP8&) = delete;
@@ -55,8 +52,51 @@ public:
}
private:
VP8PictureInfo current_context{};
Common::ScratchBuffer<u8> frame_scratch;
struct VP8PictureInfo {
INSERT_PADDING_WORDS_NOINIT(14);
u16 frame_width; // actual frame width
u16 frame_height; // actual frame height
u8 key_frame;
u8 version;
union {
u8 raw;
BitField<0, 2, u8> tile_format;
BitField<2, 3, u8> gob_height;
BitField<5, 3, u8> reserved_surface_format;
};
u8 error_conceal_on; // 1: error conceal on; 0: off
u32 first_part_size; // the size of first partition(frame header and mb header partition)
u32 hist_buffer_size; // in units of 256
u32 vld_buffer_size; // in units of 1
// Current frame buffers
std::array<u32, 2> frame_stride; // [y_c]
u32 luma_top_offset; // offset of luma top field in units of 256
u32 luma_bot_offset; // offset of luma bottom field in units of 256
u32 luma_frame_offset; // offset of luma frame in units of 256
u32 chroma_top_offset; // offset of chroma top field in units of 256
u32 chroma_bot_offset; // offset of chroma bottom field in units of 256
u32 chroma_frame_offset; // offset of chroma frame in units of 256
INSERT_PADDING_BYTES_NOINIT(0x1c); // NvdecDisplayParams
// Decode picture buffer related
s8 current_output_memory_layout;
// output NV12/NV24 setting. index 0: golden; 1: altref; 2: last
std::array<s8, 3> output_memory_layout;
u8 segmentation_feature_data_update;
INSERT_PADDING_BYTES_NOINIT(3);
// ucode return result
u32 result_value;
std::array<u32, 8> partition_offset;
INSERT_PADDING_WORDS_NOINIT(3);
};
static_assert(sizeof(VP8PictureInfo) == 0xc0, "PictureInfo is an invalid size");
VP8PictureInfo current_context{};
};
} // namespace Decoders
+63 -20
View File
@@ -1,11 +1,10 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
#include <algorithm> // for std::copy
#include <bit>
#include <numeric>
#include "common/alignment.h"
@@ -16,11 +15,11 @@
namespace Tegra::Decoders {
namespace {
static constexpr u32 diff_update_probability = 252;
static constexpr u32 frame_sync_code = 0x498342;
constexpr u32 diff_update_probability = 252;
constexpr u32 frame_sync_code = 0x498342;
// Default compressed header probabilities once frame context resets
static constexpr Vp9EntropyProbs default_probs{
constexpr Vp9EntropyProbs default_probs{
.y_mode_prob{
65, 32, 18, 144, 162, 194, 41, 51, 98, 132, 68, 18, 165, 217, 196, 45, 40, 78,
173, 80, 19, 176, 240, 193, 64, 35, 46, 221, 135, 38, 194, 248, 121, 96, 85, 29,
@@ -159,6 +158,34 @@ static constexpr Vp9EntropyProbs default_probs{
.high_precision{128, 128},
};
constexpr std::array<u8, 256> norm_lut{
0, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
};
constexpr std::array<u8, 254> map_lut{
20, 21, 22, 23, 24, 25, 0, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37,
1, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 2, 50, 51, 52, 53, 54,
55, 56, 57, 58, 59, 60, 61, 3, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72,
73, 4, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 5, 86, 87, 88, 89,
90, 91, 92, 93, 94, 95, 96, 97, 6, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107,
108, 109, 7, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 8, 122, 123, 124,
125, 126, 127, 128, 129, 130, 131, 132, 133, 9, 134, 135, 136, 137, 138, 139, 140, 141, 142,
143, 144, 145, 10, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 11, 158, 159,
160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 12, 170, 171, 172, 173, 174, 175, 176, 177,
178, 179, 180, 181, 13, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 14, 194,
195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 15, 206, 207, 208, 209, 210, 211, 212,
213, 214, 215, 216, 217, 16, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 17,
230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 18, 242, 243, 244, 245, 246, 247,
248, 249, 250, 251, 252, 253, 19,
};
// 6.2.14 Tile size calculation
[[nodiscard]] s32 CalcMinLog2TileCols(s32 frame_width) {
@@ -200,17 +227,25 @@ static constexpr Vp9EntropyProbs default_probs{
[[nodiscard]] s32 RemapProbability(s32 new_prob, s32 old_prob) {
new_prob--;
old_prob--;
u8 i = old_prob * 2 <= 0xff
? u8((std::max)(0, RecenterNonNeg(new_prob, old_prob) - 1))
: u8((std::max)(0, RecenterNonNeg(0xff - 1 - new_prob, 0xff - 1 - old_prob) - 1));
return s32((i + 7) % 13 == 0 ? (i + 7) / 13 - 1 : i + 20 - (i + 7) / 13);
std::size_t index{};
if (old_prob * 2 <= 0xff) {
index = static_cast<std::size_t>((std::max)(0, RecenterNonNeg(new_prob, old_prob) - 1));
} else {
index = static_cast<std::size_t>(
(std::max)(0, RecenterNonNeg(0xff - 1 - new_prob, 0xff - 1 - old_prob) - 1));
}
return static_cast<s32>(map_lut[index]);
}
} // Anonymous namespace
VP9::VP9(Host1x::Host1x& host1x_, const Host1x::NvdecCommon::NvdecRegisters& regs_, s32 id_)
: Decoder{host1x_, id_, regs_}
{
initialized = decode_api.Initialize(Host1x::NvdecCommon::VideoCodec::VP9);
VP9::VP9(Host1x::Host1x& host1x_, const Host1x::NvdecCommon::NvdecRegisters& regs_, s32 id_,
Host1x::FrameQueue& frame_queue_)
: Decoder{host1x_, id_, regs_, frame_queue_} {
codec = Host1x::NvdecCommon::VideoCodec::VP9;
initialized = decode_api.Initialize(codec);
}
VP9::~VP9() = default;
@@ -342,7 +377,8 @@ void VP9::WriteSegmentation(VpxBitStreamWriter& writer) {
if (update_map) {
EntropyProbs entropy_probs{};
host1x.gmmu_manager.ReadBlock(regs.vp9_prob_tab_buffer_offset.Address(), &entropy_probs, sizeof(entropy_probs));
memory_manager.ReadBlock(regs.vp9_prob_tab_buffer_offset.Address(), &entropy_probs,
sizeof(entropy_probs));
auto WriteProb = [&](u8 prob) {
bool coded = prob != 255;
@@ -406,7 +442,8 @@ void VP9::WriteSegmentation(VpxBitStreamWriter& writer) {
}
Vp9PictureInfo VP9::GetVp9PictureInfo() {
host1x.gmmu_manager.ReadBlock(regs.picture_info_offset.Address(), &current_picture_info, sizeof(PictureInfo));
memory_manager.ReadBlock(regs.picture_info_offset.Address(), &current_picture_info,
sizeof(PictureInfo));
Vp9PictureInfo vp9_info = current_picture_info.Convert();
InsertEntropy(regs.vp9_prob_tab_buffer_offset.Address(), vp9_info.entropy);
@@ -422,7 +459,7 @@ Vp9PictureInfo VP9::GetVp9PictureInfo() {
void VP9::InsertEntropy(u64 offset, Vp9EntropyProbs& dst) {
EntropyProbs entropy;
host1x.gmmu_manager.ReadBlock(offset, &entropy, sizeof(EntropyProbs));
memory_manager.ReadBlock(offset, &entropy, sizeof(EntropyProbs));
entropy.Convert(dst);
}
@@ -432,7 +469,9 @@ Vp9FrameContainer VP9::GetCurrentFrame() {
// gpu.SyncGuestHost(); epic, why?
current_frame.info = GetVp9PictureInfo();
current_frame.bit_stream.resize(current_frame.info.bitstream_size);
host1x.gmmu_manager.ReadBlock(regs.frame_bitstream_offset.Address(), current_frame.bit_stream.data(), current_frame.info.bitstream_size);
memory_manager.ReadBlock(regs.frame_bitstream_offset.Address(),
current_frame.bit_stream.data(),
current_frame.info.bitstream_size);
}
if (!next_frame.bit_stream.empty()) {
Vp9FrameContainer temp{
@@ -854,9 +893,13 @@ std::span<const u8> VP9::ComposeFrame() {
// Write headers and frame to buffer
frame_scratch.resize(uncompressed_header.size() + compressed_header.size() + bitstream.size());
std::copy(uncompressed_header.begin(), uncompressed_header.end(), frame_scratch.begin());
std::copy(compressed_header.begin(), compressed_header.end(), frame_scratch.begin() + uncompressed_header.size());
std::copy(bitstream.begin(), bitstream.end(), frame_scratch.begin() + uncompressed_header.size() + compressed_header.size());
std::copy(compressed_header.begin(), compressed_header.end(),
frame_scratch.begin() + uncompressed_header.size());
std::copy(bitstream.begin(), bitstream.end(),
frame_scratch.begin() + uncompressed_header.size() + compressed_header.size());
vp9_hidden_frame = WasFrameHidden();
return GetFrameBytes();
}
@@ -886,7 +929,7 @@ void VpxRangeEncoder::Write(bool bit, s32 probability) {
local_range = range - split;
}
s32 shift = s32(local_range == 0 ? 0 : (std::countl_zero<uint32_t>(local_range) - 24));
s32 shift = static_cast<s32>(norm_lut[local_range]);
local_range <<= shift;
count += shift;
+5 -6
View File
@@ -1,6 +1,3 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -14,7 +11,7 @@
#include "common/scratch_buffer.h"
#include "common/stream.h"
#include "video_core/host1x/codecs/decoder.h"
#include "video_core/host1x/codec_types.h"
#include "video_core/host1x/codecs/vp9_types.h"
#include "video_core/host1x/nvdec_common.h"
namespace Tegra {
@@ -116,7 +113,8 @@ private:
class VP9 final : public Decoder {
public:
explicit VP9(Host1x::Host1x& host1x, const Host1x::NvdecCommon::NvdecRegisters& regs, s32 id);
explicit VP9(Host1x::Host1x& host1x, const Host1x::NvdecCommon::NvdecRegisters& regs, s32 id,
Host1x::FrameQueue& frame_queue);
~VP9() override;
VP9(const VP9&) = delete;
@@ -195,10 +193,11 @@ private:
[[nodiscard]] std::vector<u8> ComposeCompressedHeader();
[[nodiscard]] VpxBitStreamWriter ComposeUncompressedHeader();
Common::ScratchBuffer<u8> frame_scratch;
std::array<s8, 4> loop_filter_ref_deltas{};
std::array<s8, 2> loop_filter_mode_deltas{};
Common::ScratchBuffer<u8> frame_scratch;
Vp9FrameContainer next_frame{};
std::array<Vp9EntropyProbs, 4> frame_ctxs{};
bool swap_ref_indices{};
@@ -1,6 +1,3 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
@@ -9,192 +6,12 @@
#include <array>
#include <vector>
#include "common/bit_field.h"
#include "common/common_funcs.h"
#include "common/common_types.h"
namespace Tegra {
namespace Decoders {
struct Offset {
constexpr u32 Address() const noexcept {
return offset << 8;
}
private:
u32 offset;
};
static_assert(std::is_trivial_v<Offset>, "Offset must be trivial");
static_assert(sizeof(Offset) == 0x4, "Offset has the wrong size!");
struct H264ParameterSet {
s32 log2_max_pic_order_cnt_lsb_minus4; ///< 0x00
s32 delta_pic_order_always_zero_flag; ///< 0x04
s32 frame_mbs_only_flag; ///< 0x08
u32 pic_width_in_mbs; ///< 0x0C
u32 frame_height_in_mbs; ///< 0x10
union { ///< 0x14
BitField<0, 2, u32> tile_format;
BitField<2, 3, u32> gob_height;
BitField<5, 27, u32> reserved_surface_format;
};
u32 entropy_coding_mode_flag; ///< 0x18
s32 pic_order_present_flag; ///< 0x1C
s32 num_refidx_l0_default_active; ///< 0x20
s32 num_refidx_l1_default_active; ///< 0x24
s32 deblocking_filter_control_present_flag; ///< 0x28
s32 redundant_pic_cnt_present_flag; ///< 0x2C
u32 transform_8x8_mode_flag; ///< 0x30
u32 pitch_luma; ///< 0x34
u32 pitch_chroma; ///< 0x38
Offset luma_top_offset; ///< 0x3C
Offset luma_bot_offset; ///< 0x40
Offset luma_frame_offset; ///< 0x44
Offset chroma_top_offset; ///< 0x48
Offset chroma_bot_offset; ///< 0x4C
Offset chroma_frame_offset; ///< 0x50
u32 hist_buffer_size; ///< 0x54
union { ///< 0x58
union {
BitField<0, 1, u64> mbaff_frame;
BitField<1, 1, u64> direct_8x8_inference;
BitField<2, 1, u64> weighted_pred;
BitField<3, 1, u64> constrained_intra_pred;
BitField<4, 1, u64> ref_pic;
BitField<5, 1, u64> field_pic;
BitField<6, 1, u64> bottom_field;
BitField<7, 1, u64> second_field;
} flags;
BitField<8, 4, u64> log2_max_frame_num_minus4;
BitField<12, 2, u64> chroma_format_idc;
BitField<14, 2, u64> pic_order_cnt_type;
BitField<16, 6, s64> pic_init_qp_minus26;
BitField<22, 5, s64> chroma_qp_index_offset;
BitField<27, 5, s64> second_chroma_qp_index_offset;
BitField<32, 2, u64> weighted_bipred_idc;
BitField<34, 7, u64> curr_pic_idx;
BitField<41, 5, u64> curr_col_idx;
BitField<46, 16, u64> frame_number;
BitField<62, 1, u64> frame_surfaces;
BitField<63, 1, u64> output_memory_layout;
};
};
static_assert(sizeof(H264ParameterSet) == 0x60, "H264ParameterSet is an invalid size");
#define ASSERT_POSITION(field_name, position) \
static_assert(offsetof(H264ParameterSet, field_name) == position, \
"Field " #field_name " has invalid position")
ASSERT_POSITION(log2_max_pic_order_cnt_lsb_minus4, 0x00);
ASSERT_POSITION(delta_pic_order_always_zero_flag, 0x04);
ASSERT_POSITION(frame_mbs_only_flag, 0x08);
ASSERT_POSITION(pic_width_in_mbs, 0x0C);
ASSERT_POSITION(frame_height_in_mbs, 0x10);
ASSERT_POSITION(tile_format, 0x14);
ASSERT_POSITION(entropy_coding_mode_flag, 0x18);
ASSERT_POSITION(pic_order_present_flag, 0x1C);
ASSERT_POSITION(num_refidx_l0_default_active, 0x20);
ASSERT_POSITION(num_refidx_l1_default_active, 0x24);
ASSERT_POSITION(deblocking_filter_control_present_flag, 0x28);
ASSERT_POSITION(redundant_pic_cnt_present_flag, 0x2C);
ASSERT_POSITION(transform_8x8_mode_flag, 0x30);
ASSERT_POSITION(pitch_luma, 0x34);
ASSERT_POSITION(pitch_chroma, 0x38);
ASSERT_POSITION(luma_top_offset, 0x3C);
ASSERT_POSITION(luma_bot_offset, 0x40);
ASSERT_POSITION(luma_frame_offset, 0x44);
ASSERT_POSITION(chroma_top_offset, 0x48);
ASSERT_POSITION(chroma_bot_offset, 0x4C);
ASSERT_POSITION(chroma_frame_offset, 0x50);
ASSERT_POSITION(hist_buffer_size, 0x54);
ASSERT_POSITION(flags, 0x58);
#undef ASSERT_POSITION
struct DpbEntry {
union {
BitField<0, 7, u32> index;
BitField<7, 5, u32> col_idx;
BitField<12, 2, u32> state;
BitField<14, 1, u32> is_long_term;
BitField<15, 1, u32> non_existing;
BitField<16, 1, u32> is_field;
BitField<17, 4, u32> top_field_marking;
BitField<21, 4, u32> bottom_field_marking;
BitField<25, 1, u32> output_memory_layout;
BitField<26, 6, u32> reserved;
} flags;
std::array<u32, 2> field_order_cnt;
u32 frame_idx;
};
static_assert(sizeof(DpbEntry) == 0x10, "DpbEntry has the wrong size!");
struct DisplayParam {
union {
BitField<0, 1, u32> enable_tf_output;
BitField<1, 1, u32> vc1_map_y_flag;
BitField<2, 3, u32> map_y_value;
BitField<5, 1, u32> vc1_map_uv_flag;
BitField<6, 3, u32> map_uv_value;
BitField<9, 8, u32> out_stride;
BitField<17, 3, u32> tiling_format;
BitField<20, 1, u32> output_structure; // 0=frame, 1=field
BitField<21, 11, u32> reserved0;
};
std::array<s32, 2> output_top;
std::array<s32, 2> output_bottom;
union {
BitField<0, 1, u32> enable_histogram;
BitField<1, 12, u32> histogram_start_x;
BitField<13, 12, u32> histogram_start_y;
BitField<25, 7, u32> reserved1;
};
union {
BitField<0, 12, u32> histogram_end_x;
BitField<12, 12, u32> histogram_end_y;
BitField<24, 8, u32> reserved2;
};
};
static_assert(sizeof(DisplayParam) == 0x1C, "DisplayParam has the wrong size!");
struct H264DecoderContext {
INSERT_PADDING_WORDS_NOINIT(13); ///< 0x0000
std::array<u8, 16> eos; ///< 0x0034
u8 explicit_eos_present_flag; ///< 0x0044
u8 hint_dump_en; ///< 0x0045
INSERT_PADDING_BYTES_NOINIT(2); ///< 0x0046
u32 stream_len; ///< 0x0048
u32 slice_count; ///< 0x004C
u32 mbhist_buffer_size; ///< 0x0050
u32 gptimer_timeout_value; ///< 0x0054
H264ParameterSet h264_parameter_set; ///< 0x0058
std::array<s32, 2> curr_field_order_cnt; ///< 0x00B8
std::array<DpbEntry, 16> dpb; ///< 0x00C0
std::array<u8, 0x60> weight_scale_4x4; ///< 0x01C0
std::array<u8, 0x80> weight_scale_8x8; ///< 0x0220
std::array<u8, 2> num_inter_view_refs_lX; ///< 0x02A0
std::array<u8, 14> reserved2; ///< 0x02A2
std::array<std::array<s8, 16>, 2> inter_view_refidx_lX; ///< 0x02B0
union { ///< 0x02D0
BitField<0, 1, u32> lossless_ipred8x8_filter_enable;
BitField<1, 1, u32> qpprime_y_zero_transform_bypass_flag;
BitField<2, 30, u32> reserved3;
};
DisplayParam display_param; ///< 0x02D4
std::array<u32, 3> reserved4; ///< 0x02F0
};
static_assert(sizeof(H264DecoderContext) == 0x2FC, "H264DecoderContext is an invalid size");
#define ASSERT_POSITION(field_name, position) \
static_assert(offsetof(H264DecoderContext, field_name) == position, \
"Field " #field_name " has invalid position")
ASSERT_POSITION(stream_len, 0x48);
ASSERT_POSITION(h264_parameter_set, 0x58);
ASSERT_POSITION(dpb, 0xC0);
ASSERT_POSITION(weight_scale_4x4, 0x1C0);
#undef ASSERT_POSITION
enum class Vp9SurfaceIndex : u32 {
Last = 0,
Golden = 1,
@@ -470,7 +287,10 @@ struct RefPoolElement {
bool refresh{};
};
#define ASSERT_POSITION(field_name, position) static_assert(offsetof(Vp9EntropyProbs, field_name) == position)
#define ASSERT_POSITION(field_name, position) \
static_assert(offsetof(Vp9EntropyProbs, field_name) == position, \
"Field " #field_name " has invalid position")
ASSERT_POSITION(partition_prob, 0x0024);
ASSERT_POSITION(switchable_interp_prob, 0x0724);
ASSERT_POSITION(sign, 0x0772);
@@ -478,7 +298,10 @@ ASSERT_POSITION(class_0_fr, 0x079E);
ASSERT_POSITION(high_precision, 0x07B2);
#undef ASSERT_POSITION
#define ASSERT_POSITION(field_name, position) static_assert(offsetof(PictureInfo, field_name) == position)
#define ASSERT_POSITION(field_name, position) \
static_assert(offsetof(PictureInfo, field_name) == position, \
"Field " #field_name " has invalid position")
ASSERT_POSITION(bitstream_size, 0x30);
ASSERT_POSITION(last_frame_size, 0x48);
ASSERT_POSITION(first_level, 0x70);
@@ -486,7 +309,9 @@ ASSERT_POSITION(segmentation, 0x80);
ASSERT_POSITION(loop_filter, 0xE4);
#undef ASSERT_POSITION
#define ASSERT_POSITION(field_name, position) static_assert(offsetof(EntropyProbs, field_name) == position)
#define ASSERT_POSITION(field_name, position) \
static_assert(offsetof(EntropyProbs, field_name) == position, \
"Field " #field_name " has invalid position")
ASSERT_POSITION(inter_mode_prob, 0x400);
ASSERT_POSITION(tx_8x8_prob, 0x470);
@@ -496,47 +321,5 @@ ASSERT_POSITION(class_0_fr, 0x560);
ASSERT_POSITION(coef_probs, 0x5A0);
#undef ASSERT_POSITION
struct VP8PictureInfo {
INSERT_PADDING_WORDS_NOINIT(14);
u16 frame_width; // actual frame width
u16 frame_height; // actual frame height
u8 key_frame;
u8 version;
union {
u8 raw;
BitField<0, 2, u8> tile_format;
BitField<2, 3, u8> gob_height;
BitField<5, 3, u8> reserved_surface_format;
};
u8 error_conceal_on; // 1: error conceal on; 0: off
u32 first_part_size; // the size of first partition(frame header and mb header partition)
u32 hist_buffer_size; // in units of 256
u32 vld_buffer_size; // in units of 1
// Current frame buffers
std::array<u32, 2> frame_stride; // [y_c]
u32 luma_top_offset; // offset of luma top field in units of 256
u32 luma_bot_offset; // offset of luma bottom field in units of 256
u32 luma_frame_offset; // offset of luma frame in units of 256
u32 chroma_top_offset; // offset of chroma top field in units of 256
u32 chroma_bot_offset; // offset of chroma bottom field in units of 256
u32 chroma_frame_offset; // offset of chroma frame in units of 256
INSERT_PADDING_BYTES_NOINIT(0x1c); // NvdecDisplayParams
// Decode picture buffer related
s8 current_output_memory_layout;
// output NV12/NV24 setting. index 0: golden; 1: altref; 2: last
std::array<s8, 3> output_memory_layout;
u8 segmentation_feature_data_update;
INSERT_PADDING_BYTES_NOINIT(3);
// ucode return result
u32 result_value;
std::array<u32, 8> partition_offset;
INSERT_PADDING_WORDS_NOINIT(3);
};
static_assert(sizeof(VP8PictureInfo) == 0xc0, "PictureInfo is an invalid size");
}; // namespace Decoders
}; // namespace Tegra
@@ -9,7 +9,7 @@
#include "common/scope_exit.h"
#include "common/settings.h"
#include "core/memory.h"
#include "video_core/host1x/ffmpeg.h"
#include "video_core/host1x/ffmpeg/ffmpeg.h"
#include "video_core/memory_manager.h"
extern "C" {
@@ -34,9 +34,7 @@ constexpr std::array PreferredGpuDecoders = {
AV_HWDEVICE_TYPE_DXVA2,
AV_HWDEVICE_TYPE_D3D12VA,
#elif defined(__FreeBSD__)
AV_HWDEVICE_TYPE_VAAPI,
AV_HWDEVICE_TYPE_VDPAU,
AV_HWDEVICE_TYPE_DRM,
#elif defined(__APPLE__)
AV_HWDEVICE_TYPE_VIDEOTOOLBOX,
#elif defined(ANDROID)
@@ -107,17 +105,18 @@ Frame::~Frame() {
Decoder::Decoder(Tegra::Host1x::NvdecCommon::VideoCodec codec) {
const AVCodecID av_codec = [&] {
switch (codec) {
case Tegra::Host1x::NvdecCommon::VideoCodec::H264:
return AV_CODEC_ID_H264;
case Tegra::Host1x::NvdecCommon::VideoCodec::VP8:
return AV_CODEC_ID_VP8;
case Tegra::Host1x::NvdecCommon::VideoCodec::VP9:
return AV_CODEC_ID_VP9;
default:
UNIMPLEMENTED_MSG("Unknown codec {}", codec);
return AV_CODEC_ID_NONE;
case Tegra::Host1x::NvdecCommon::VideoCodec::H264:
return AV_CODEC_ID_H264;
case Tegra::Host1x::NvdecCommon::VideoCodec::VP8:
return AV_CODEC_ID_VP8;
case Tegra::Host1x::NvdecCommon::VideoCodec::VP9:
return AV_CODEC_ID_VP9;
default:
UNIMPLEMENTED_MSG("Unknown codec {}", codec);
return AV_CODEC_ID_NONE;
}
}();
m_codec = avcodec_find_decoder(av_codec);
}
@@ -1,4 +1,4 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2023 yuzu Emulator Project
+6 -9
View File
@@ -1,4 +1,4 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: 2021 yuzu Emulator Project
@@ -12,12 +12,9 @@
namespace Tegra::Host1x {
Host1x::Host1x(Core::System& system_)
: system{system_}
, syncpoint_manager{}
, memory_manager(system.DeviceMemory())
, gmmu_manager{system, memory_manager, 32, 0, 12}
, allocator{1 << 12}
{}
: system{system_}, syncpoint_manager{},
memory_manager(system.DeviceMemory()), gmmu_manager{system, memory_manager, 32, 0, 12},
allocator{std::make_unique<Common::FlatAllocator<u32, 0, 32>>(1 << 12)} {}
Host1x::~Host1x() = default;
@@ -27,13 +24,13 @@ void Host1x::StartDevice(s32 fd, ChannelType type, u32 syncpt) {
#ifdef YUZU_LEGACY
std::call_once(nvdec_first_init, []() {std::this_thread::sleep_for(std::chrono::milliseconds{500});}); // HACK: For Astroneer
#endif
devices[fd] = std::make_unique<Tegra::Host1x::Nvdec>(*this, fd, syncpt);
devices[fd] = std::make_unique<Tegra::Host1x::Nvdec>(*this, fd, syncpt, frame_queue);
break;
case ChannelType::VIC:
#ifdef YUZU_LEGACY
std::call_once(vic_first_init, []() {std::this_thread::sleep_for(std::chrono::milliseconds{500});}); // HACK: For Astroneer
#endif
devices[fd] = std::make_unique<Tegra::Host1x::Vic>(*this, fd, syncpt);
devices[fd] = std::make_unique<Tegra::Host1x::Vic>(*this, fd, syncpt, frame_queue);
break;
default:
LOG_ERROR(HW_GPU, "Unimplemented host1x device {}", static_cast<u32>(type));
+12 -3
View File
@@ -184,12 +184,20 @@ public:
return memory_manager;
}
Tegra::MemoryManager& GMMU() {
return gmmu_manager;
}
const Tegra::MemoryManager& GMMU() const {
return gmmu_manager;
}
Common::FlatAllocator<u32, 0, 32>& Allocator() {
return allocator;
return *allocator;
}
const Common::FlatAllocator<u32, 0, 32>& Allocator() const {
return allocator;
return *allocator;
}
void StartDevice(s32 fd, ChannelType type, u32 syncpt);
@@ -203,11 +211,12 @@ public:
it->second->PushEntries(std::move(entries));
}
private:
Core::System& system;
SyncpointManager syncpoint_manager;
Tegra::MaxwellDeviceMemoryManager memory_manager;
Tegra::MemoryManager gmmu_manager;
Common::FlatAllocator<u32, 0, 32> allocator;
std::unique_ptr<Common::FlatAllocator<u32, 0, 32>> allocator;
FrameQueue frame_queue;
ankerl::unordered_dense::map<s32, std::unique_ptr<CDmaPusher>> devices;
#ifdef YUZU_LEGACY
+30 -31
View File
@@ -1,10 +1,9 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
#include <variant>
#include "common/assert.h"
#include "common/polyfill_thread.h"
@@ -20,13 +19,10 @@ namespace Tegra::Host1x {
#define NVDEC_REG_INDEX(field_name) \
(offsetof(NvdecCommon::NvdecRegisters, field_name) / sizeof(u64))
Nvdec::Nvdec(Host1x& host1x_, s32 id_, u32 syncpt)
: CDmaPusher{host1x_, id_}
, id{id_}
, syncpoint{syncpt}
{
Nvdec::Nvdec(Host1x& host1x_, s32 id_, u32 syncpt, FrameQueue& frame_queue_)
: CDmaPusher{host1x_, id_}, id{id_}, syncpoint{syncpt}, frame_queue{frame_queue_} {
LOG_INFO(HW_GPU, "Created nvdec {}", id);
host1x.frame_queue.Open(id);
frame_queue.Open(id);
}
Nvdec::~Nvdec() {
@@ -47,22 +43,24 @@ void Nvdec::ProcessMethod(u32 method, u32 argument) {
}
void Nvdec::CreateDecoder(NvdecCommon::VideoCodec codec) {
if (std::holds_alternative<std::monostate>(decoder)) {
switch (codec) {
case NvdecCommon::VideoCodec::H264:
decoder.emplace<Decoders::H264>(host1x, regs, id);
break;
case NvdecCommon::VideoCodec::VP8:
decoder.emplace<Decoders::VP8>(host1x, regs, id);
break;
case NvdecCommon::VideoCodec::VP9:
decoder.emplace<Decoders::VP9>(host1x, regs, id);
break;
default:
break;
}
LOG_INFO(HW_GPU, "Created decoder {} for id {}", codec, id);
if (decoder.get()) {
return;
}
switch (codec) {
case NvdecCommon::VideoCodec::H264:
decoder = std::make_unique<Decoders::H264>(host1x, regs, id, frame_queue);
break;
case NvdecCommon::VideoCodec::VP8:
decoder = std::make_unique<Decoders::VP8>(host1x, regs, id, frame_queue);
break;
case NvdecCommon::VideoCodec::VP9:
decoder = std::make_unique<Decoders::VP9>(host1x, regs, id, frame_queue);
break;
default:
UNIMPLEMENTED_MSG("Codec {}", decoder->GetCurrentCodecName());
break;
}
LOG_INFO(HW_GPU, "Created decoder {} for id {}", decoder->GetCurrentCodecName(), id);
}
void Nvdec::Execute() {
@@ -72,14 +70,15 @@ void Nvdec::Execute() {
std::this_thread::sleep_for(std::chrono::milliseconds(8));
return;
}
if (auto* h264 = std::get_if<Decoders::H264>(&decoder)) {
h264->Decode();
} else if (auto* vp8 = std::get_if<Decoders::VP8>(&decoder)) {
vp8->Decode();
} else if (auto* vp9 = std::get_if<Decoders::VP9>(&decoder)) {
vp9->Decode();
} else {
LOG_ERROR(HW_GPU, "Unrecognized codec executed?");
switch (decoder->GetCurrentCodec()) {
case NvdecCommon::VideoCodec::H264:
case NvdecCommon::VideoCodec::VP8:
case NvdecCommon::VideoCodec::VP9:
decoder->Decode();
break;
default:
UNIMPLEMENTED_MSG("Codec {}", decoder->GetCurrentCodecName());
break;
}
}
+6 -13
View File
@@ -1,4 +1,4 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
@@ -8,14 +8,10 @@
#include <memory>
#include <vector>
#include <variant>
#include "common/common_types.h"
#include "video_core/cdma_pusher.h"
#include "video_core/host1x/codecs/decoder.h"
#include "video_core/host1x/codecs/h264.h"
#include "video_core/host1x/codecs/vp8.h"
#include "video_core/host1x/codecs/vp9.h"
namespace Tegra {
@@ -25,7 +21,7 @@ class FrameQueue;
class Nvdec final : public CDmaPusher {
public:
explicit Nvdec(Host1x& host1x, s32 id, u32 syncpt);
explicit Nvdec(Host1x& host1x, s32 id, u32 syncpt, FrameQueue& frame_queue_);
~Nvdec();
/// Writes the method into the state, Invoke Execute() if encountered
@@ -42,15 +38,12 @@ private:
/// Invoke codec to decode a frame
void Execute();
NvdecCommon::NvdecRegisters regs{};
std::variant<
Decoders::H264,
Decoders::VP8,
Decoders::VP9,
std::monostate
> decoder = std::monostate{};
s32 id;
u32 syncpoint;
FrameQueue& frame_queue;
NvdecCommon::NvdecRegisters regs{};
std::unique_ptr<Decoder> decoder;
};
} // namespace Host1x
+13 -12
View File
@@ -85,17 +85,18 @@ void SwizzleSurface(std::span<u8> output, u32 out_stride, std::span<const u8> in
} // namespace
Vic::Vic(Host1x& host1x_, s32 id_, u32 syncpt) noexcept
: CDmaPusher{host1x_, id_}
Vic::Vic(Host1x& host1x_, s32 id_, u32 syncpt, FrameQueue& frame_queue_) noexcept :
CDmaPusher{host1x_, id_}
, id{id_}
, syncpoint{syncpt}
, frame_queue{frame_queue_}
{
LOG_INFO(HW_GPU, "Created vic {}", id);
}
Vic::~Vic() noexcept {
LOG_INFO(HW_GPU, "Destroying vic {}", id);
host1x.frame_queue.Close(id);
frame_queue.Close(id);
}
void Vic::ProcessMethod(u32 method, u32 arg) noexcept {
@@ -112,7 +113,7 @@ void Vic::ProcessMethod(u32 method, u32 arg) noexcept {
void Vic::Execute() noexcept {
ConfigStruct config{};
host1x.gmmu_manager.ReadBlock(regs.config_struct_offset.Address(), &config, sizeof(ConfigStruct));
memory_manager.ReadBlock(regs.config_struct_offset.Address(), &config, sizeof(ConfigStruct));
auto output_width = config.output_surface_config.out_surface_width + 1;
auto output_height = config.output_surface_config.out_surface_height + 1;
@@ -123,8 +124,8 @@ void Vic::Execute() noexcept {
if (auto& slot_config = config.slot_structs[i]; slot_config.config.slot_enable) {
auto const luma_offset = regs.surfaces[i][SurfaceIndex::Current].luma.Address();
if (nvdec_id == -1)
nvdec_id = host1x.frame_queue.VicFindNvdecFdFromOffset(luma_offset);
if (auto frame = host1x.frame_queue.GetFrame(nvdec_id, luma_offset); frame.get()) {
nvdec_id = frame_queue.VicFindNvdecFdFromOffset(luma_offset);
if (auto frame = frame_queue.GetFrame(nvdec_id, luma_offset); frame.get()) {
switch (frame->GetPixelFormat()) {
case AV_PIX_FMT_YUV420P:
ReadY8__V8U8_N420(slot_config, regs.surfaces[i], std::move(frame), true);
@@ -880,8 +881,8 @@ void Vic::WriteY8__V8U8_N420(const OutputSurfaceConfig& output_surface_config) n
chroma_scratch.resize_destructive(out_chroma_size);
Decode(luma_scratch.data(), chroma_scratch.data());
Tegra::Memory::GpuGuestMemoryScoped<u8, Core::Memory::GuestMemoryFlags::SafeWrite> out_luma(host1x.gmmu_manager, regs.output_surface.luma.Address(), out_luma_swizzle_size, &swizzle_scratch);
Tegra::Memory::GpuGuestMemoryScoped<u8, Core::Memory::GuestMemoryFlags::SafeWrite> out_chroma(host1x.gmmu_manager, regs.output_surface.chroma_u.Address(), out_chroma_swizzle_size, &swizzle_scratch);
Tegra::Memory::GpuGuestMemoryScoped<u8, Core::Memory::GuestMemoryFlags::SafeWrite> out_luma(memory_manager, regs.output_surface.luma.Address(), out_luma_swizzle_size, &swizzle_scratch);
Tegra::Memory::GpuGuestMemoryScoped<u8, Core::Memory::GuestMemoryFlags::SafeWrite> out_chroma(memory_manager, regs.output_surface.chroma_u.Address(), out_chroma_swizzle_size, &swizzle_scratch);
if (block_height == 1) {
SwizzleSurface(out_luma, out_luma_stride, luma_scratch, out_luma_stride, out_luma_height);
SwizzleSurface(out_chroma, out_chroma_stride, chroma_scratch, out_chroma_stride, out_chroma_height);
@@ -909,8 +910,8 @@ void Vic::WriteY8__V8U8_N420(const OutputSurfaceConfig& output_surface_config) n
luma_scratch.resize_destructive(out_luma_size);
chroma_scratch.resize_destructive(out_chroma_size);
Decode(luma_scratch.data(), chroma_scratch.data());
host1x.gmmu_manager.WriteBlock(regs.output_surface.luma.Address(), luma_scratch.data(), out_luma_size);
host1x.gmmu_manager.WriteBlock(regs.output_surface.chroma_u.Address(), chroma_scratch.data(), out_chroma_size);
memory_manager.WriteBlock(regs.output_surface.luma.Address(), luma_scratch.data(), out_luma_size);
memory_manager.WriteBlock(regs.output_surface.chroma_u.Address(), chroma_scratch.data(), out_chroma_size);
} break;
default:
UNREACHABLE();
@@ -1045,7 +1046,7 @@ void Vic::WriteABGR(const OutputSurfaceConfig& output_surface_config, VideoPixel
luma_scratch.resize_destructive(out_luma_size);
Decode(luma_scratch.data(), output_surface.data());
Tegra::Memory::GpuGuestMemoryScoped<u8, Core::Memory::GuestMemoryFlags::SafeWrite> out_luma(host1x.gmmu_manager, regs.output_surface.luma.Address(), out_swizzle_size, &swizzle_scratch);
Tegra::Memory::GpuGuestMemoryScoped<u8, Core::Memory::GuestMemoryFlags::SafeWrite> out_luma(memory_manager, regs.output_surface.luma.Address(), out_swizzle_size, &swizzle_scratch);
if (block_height == 1) {
SwizzleSurface(out_luma, out_luma_stride, luma_scratch, out_luma_stride, out_luma_height);
} else {
@@ -1060,7 +1061,7 @@ void Vic::WriteABGR(const OutputSurfaceConfig& output_surface_config, VideoPixel
surface_stride * surface_height * BytesPerPixel, out_luma_width, out_luma_height,
out_luma_stride, out_luma_size);
luma_scratch.resize_destructive(out_luma_size);
Tegra::Memory::GpuGuestMemoryScoped<u8, Core::Memory::GuestMemoryFlags::SafeWrite> out_luma(host1x.gmmu_manager, regs.output_surface.luma.Address(), out_luma_size, &luma_scratch);
Tegra::Memory::GpuGuestMemoryScoped<u8, Core::Memory::GuestMemoryFlags::SafeWrite> out_luma(memory_manager, regs.output_surface.luma.Address(), out_luma_size, &luma_scratch);
Decode(out_luma.data(), output_surface.data());
} break;
default:
+7 -7
View File
@@ -1,4 +1,4 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project
@@ -15,7 +15,6 @@
#include "common/common_types.h"
#include "common/scratch_buffer.h"
#include "video_core/cdma_pusher.h"
#include "video_core/host1x/host1x.h"
namespace Tegra::Host1x {
class Host1x;
@@ -606,7 +605,7 @@ public:
SetOutputSurfaceChromaUnusedOffset = offsetof(VicRegisters, output_surface.chroma_v)
};
explicit Vic(Host1x& host1x, s32 id, u32 syncpt) noexcept;
explicit Vic(Host1x& host1x, s32 id, u32 syncpt, FrameQueue& frame_queue) noexcept;
~Vic() noexcept;
/// Write to the device state.
@@ -621,17 +620,18 @@ private:
void WriteY8__V8U8_N420(const OutputSurfaceConfig& output_surface_config) noexcept;
void WriteABGR(const OutputSurfaceConfig& output_surface_config, VideoPixelFormat format) noexcept;
s32 id;
s32 nvdec_id{-1};
u32 syncpoint;
VicRegisters regs{};
FrameQueue& frame_queue;
Common::ScratchBuffer<u8> swizzle_scratch;
Common::ScratchBuffer<Pixel> output_surface;
Common::ScratchBuffer<Pixel> slot_surface;
Common::ScratchBuffer<u8> luma_scratch;
Common::ScratchBuffer<u8> chroma_scratch;
s32 id;
s32 nvdec_id{-1};
u32 syncpoint;
};
} // namespace Tegra::Host1x
+18 -14
View File
@@ -31,7 +31,7 @@
#include "common/settings.h"
#include "common/container_hash.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/engines/draw_manager.h"
#include "video_core/dirty_flags.h"
#include "video_core/rasterizer_interface.h"
#include "video_core/macro.h"
@@ -83,7 +83,7 @@ void HLE_DrawArraysIndirect::Execute(Engines::Maxwell3D& maxwell3d, std::span<co
return;
}
auto& params = maxwell3d.draw_manager.indirect_state;
auto& params = maxwell3d.draw_manager->GetIndirectParams();
params.is_byte_count = false;
params.is_indexed = false;
params.include_count = false;
@@ -98,7 +98,7 @@ void HLE_DrawArraysIndirect::Execute(Engines::Maxwell3D& maxwell3d, std::span<co
maxwell3d.SetHLEReplacementAttributeType(0, 0x640, Maxwell3D::HLEReplacementAttributeType::BaseInstance);
}
maxwell3d.draw_manager.DrawArrayIndirect(maxwell3d, topology);
maxwell3d.draw_manager->DrawArrayIndirect(topology);
if (extended) {
maxwell3d.engine_state = Maxwell3D::EngineHint::None;
@@ -127,7 +127,7 @@ void HLE_DrawArraysIndirect::Fallback(Engines::Maxwell3D& maxwell3d, std::span<c
maxwell3d.engine_state = Maxwell3D::EngineHint::OnHLEMacro;
maxwell3d.SetHLEReplacementAttributeType(0, 0x640, Maxwell3D::HLEReplacementAttributeType::BaseInstance);
}
maxwell3d.draw_manager.DrawArray(maxwell3d, topology, vertex_first, vertex_count, base_instance, instance_count);
maxwell3d.draw_manager->DrawArray(topology, vertex_first, vertex_count, base_instance, instance_count);
if (extended) {
maxwell3d.regs.global_base_instance_index = 0;
maxwell3d.engine_state = Maxwell3D::EngineHint::None;
@@ -154,7 +154,7 @@ void HLE_DrawIndexedIndirect::Execute(Engines::Maxwell3D& maxwell3d, std::span<c
maxwell3d.SetHLEReplacementAttributeType(0, 0x640, Maxwell3D::HLEReplacementAttributeType::BaseVertex);
maxwell3d.SetHLEReplacementAttributeType(0, 0x644, Maxwell3D::HLEReplacementAttributeType::BaseInstance);
}
auto& params = maxwell3d.draw_manager.indirect_state;
auto& params = maxwell3d.draw_manager->GetIndirectParams();
params.is_byte_count = false;
params.is_indexed = true;
params.include_count = false;
@@ -164,7 +164,7 @@ void HLE_DrawIndexedIndirect::Execute(Engines::Maxwell3D& maxwell3d, std::span<c
params.max_draw_counts = 1;
params.stride = 0;
maxwell3d.dirty.flags[VideoCommon::Dirty::IndexBuffer] = true;
maxwell3d.draw_manager.DrawIndexedIndirect(maxwell3d, topology, 0, estimate);
maxwell3d.draw_manager->DrawIndexedIndirect(topology, 0, estimate);
maxwell3d.regs.vertex_id_base = 0x0;
maxwell3d.regs.global_base_vertex_index = 0x0;
maxwell3d.regs.global_base_instance_index = 0x0;
@@ -187,7 +187,7 @@ void HLE_DrawIndexedIndirect::Fallback(Engines::Maxwell3D& maxwell3d, std::span<
maxwell3d.SetHLEReplacementAttributeType(0, 0x640, Maxwell3D::HLEReplacementAttributeType::BaseVertex);
maxwell3d.SetHLEReplacementAttributeType(0, 0x644, Maxwell3D::HLEReplacementAttributeType::BaseInstance);
}
maxwell3d.draw_manager.DrawIndex(maxwell3d, Tegra::Maxwell3D::Regs::PrimitiveTopology(parameters[0]), parameters[3], parameters[1], element_base, base_instance, instance_count);
maxwell3d.draw_manager->DrawIndex(Tegra::Maxwell3D::Regs::PrimitiveTopology(parameters[0]), parameters[3], parameters[1], element_base, base_instance, instance_count);
maxwell3d.regs.vertex_id_base = 0x0;
maxwell3d.regs.global_base_vertex_index = 0x0;
maxwell3d.regs.global_base_instance_index = 0x0;
@@ -206,7 +206,7 @@ void HLE_MultiLayerClear::Execute(Engines::Maxwell3D& maxwell3d, std::span<const
ASSERT(clear_params.layer == 0);
maxwell3d.regs.clear_surface.raw = clear_params.raw;
maxwell3d.draw_manager.Clear(maxwell3d, num_layers);
maxwell3d.draw_manager->Clear(num_layers);
}
void HLE_MultiDrawIndexedIndirectCount::Execute(Engines::Maxwell3D& maxwell3d, std::span<const u32> parameters, [[maybe_unused]] u32 method) {
const auto topology = Maxwell3D::Regs::PrimitiveTopology(parameters[2]);
@@ -230,7 +230,7 @@ void HLE_MultiDrawIndexedIndirectCount::Execute(Engines::Maxwell3D& maxwell3d, s
const std::size_t draw_count = end_indirect - start_indirect;
const u32 estimate = static_cast<u32>(maxwell3d.EstimateIndexBufferSize());
maxwell3d.dirty.flags[VideoCommon::Dirty::IndexBuffer] = true;
auto& params = maxwell3d.draw_manager.indirect_state;
auto& params = maxwell3d.draw_manager->GetIndirectParams();
params.is_byte_count = false;
params.is_indexed = true;
params.include_count = true;
@@ -244,7 +244,7 @@ void HLE_MultiDrawIndexedIndirectCount::Execute(Engines::Maxwell3D& maxwell3d, s
maxwell3d.SetHLEReplacementAttributeType(0, 0x640, Maxwell3D::HLEReplacementAttributeType::BaseVertex);
maxwell3d.SetHLEReplacementAttributeType(0, 0x644, Maxwell3D::HLEReplacementAttributeType::BaseInstance);
maxwell3d.SetHLEReplacementAttributeType(0, 0x648, Maxwell3D::HLEReplacementAttributeType::DrawID);
maxwell3d.draw_manager.DrawIndexedIndirect(maxwell3d, topology, 0, estimate);
maxwell3d.draw_manager->DrawIndexedIndirect(topology, 0, estimate);
maxwell3d.engine_state = Maxwell3D::EngineHint::None;
maxwell3d.replace_table.clear();
}
@@ -280,7 +280,7 @@ void HLE_MultiDrawIndexedIndirectCount::Fallback(Engines::Maxwell3D& maxwell3d,
maxwell3d.CallMethod(0x8e3, 0x648, true);
maxwell3d.CallMethod(0x8e4, static_cast<u32>(index), true);
maxwell3d.dirty.flags[VideoCommon::Dirty::IndexBuffer] = true;
maxwell3d.draw_manager.DrawIndex(maxwell3d, topology, parameters[base + 2], parameters[base], base_vertex, base_instance, parameters[base + 1]);
maxwell3d.draw_manager->DrawIndex(topology, parameters[base + 2], parameters[base], base_vertex, base_instance, parameters[base + 1]);
}
}
void HLE_DrawIndirectByteCount::Execute(Engines::Maxwell3D& maxwell3d, std::span<const u32> parameters, [[maybe_unused]] u32 method) {
@@ -290,7 +290,7 @@ void HLE_DrawIndirectByteCount::Execute(Engines::Maxwell3D& maxwell3d, std::span
return;
}
auto topology = Maxwell3D::Regs::PrimitiveTopology(parameters[0] & 0xFFFFU);
auto& params = maxwell3d.draw_manager.indirect_state;
auto& params = maxwell3d.draw_manager->GetIndirectParams();
params.is_byte_count = true;
params.is_indexed = false;
params.include_count = false;
@@ -302,14 +302,18 @@ void HLE_DrawIndirectByteCount::Execute(Engines::Maxwell3D& maxwell3d, std::span
maxwell3d.regs.draw.begin = parameters[0];
maxwell3d.regs.draw_auto_stride = parameters[1];
maxwell3d.regs.draw_auto_byte_count = parameters[2];
maxwell3d.draw_manager.DrawArrayIndirect(maxwell3d, topology);
maxwell3d.draw_manager->DrawArrayIndirect(topology);
}
void HLE_DrawIndirectByteCount::Fallback(Engines::Maxwell3D& maxwell3d, std::span<const u32> parameters) {
maxwell3d.RefreshParameters();
maxwell3d.regs.draw.begin = parameters[0];
maxwell3d.regs.draw_auto_stride = parameters[1];
maxwell3d.regs.draw_auto_byte_count = parameters[2];
maxwell3d.draw_manager.DrawArray(maxwell3d, maxwell3d.regs.draw.topology, 0, maxwell3d.regs.draw_auto_byte_count / maxwell3d.regs.draw_auto_stride, 0, 1);
maxwell3d.draw_manager->DrawArray(
maxwell3d.regs.draw.topology, 0,
maxwell3d.regs.draw_auto_byte_count / maxwell3d.regs.draw_auto_stride, 0, 1);
}
void HLE_C713C83D8F63CCF3::Execute(Engines::Maxwell3D& maxwell3d, std::span<const u32> parameters, [[maybe_unused]] u32 method) {
maxwell3d.RefreshParameters();
@@ -245,7 +245,7 @@ void RasterizerOpenGL::PrepareDraw(bool is_indexed, Func&& draw_func) {
SyncState();
const auto& draw_state = maxwell3d->draw_manager.draw_state;
const auto& draw_state = maxwell3d->draw_manager->GetDrawState();
const GLenum primitive_mode = MaxwellToGL::PrimitiveTopology(draw_state.topology);
BeginTransformFeedback(pipeline, primitive_mode);
@@ -260,12 +260,12 @@ void RasterizerOpenGL::PrepareDraw(bool is_indexed, Func&& draw_func) {
void RasterizerOpenGL::Draw(bool is_indexed, u32 instance_count) {
PrepareDraw(is_indexed, [this, is_indexed, instance_count](GLenum primitive_mode) {
const auto& draw_state = maxwell3d->draw_manager.draw_state;
const GLuint base_instance = GLuint(draw_state.base_instance);
const GLsizei num_instances = GLsizei(instance_count);
const auto& draw_state = maxwell3d->draw_manager->GetDrawState();
const GLuint base_instance = static_cast<GLuint>(draw_state.base_instance);
const GLsizei num_instances = static_cast<GLsizei>(instance_count);
if (is_indexed) {
const GLint base_vertex = GLint(draw_state.base_index);
const GLsizei num_vertices = GLsizei(draw_state.index_buffer.count);
const GLint base_vertex = static_cast<GLint>(draw_state.base_index);
const GLsizei num_vertices = static_cast<GLsizei>(draw_state.index_buffer.count);
const GLvoid* const offset = buffer_cache_runtime.IndexOffset();
const GLenum format = MaxwellToGL::IndexFormat(draw_state.index_buffer.format);
if (num_instances == 1 && base_instance == 0 && base_vertex == 0) {
@@ -302,7 +302,7 @@ void RasterizerOpenGL::Draw(bool is_indexed, u32 instance_count) {
}
void RasterizerOpenGL::DrawIndirect() {
const auto& params = maxwell3d->draw_manager.indirect_state;
const auto& params = maxwell3d->draw_manager->GetIndirectParams();
buffer_cache.SetDrawIndirect(&params);
PrepareDraw(params.is_indexed, [this, &params](GLenum primitive_mode) {
if (params.is_byte_count) {
@@ -358,12 +358,12 @@ void RasterizerOpenGL::DrawTexture() {
SyncState();
const auto& draw_texture_state = maxwell3d->draw_manager.draw_texture_state;
const auto& draw_texture_state = maxwell3d->draw_manager->GetDrawTextureState();
const auto& sampler = texture_cache.GetGraphicsSampler(draw_texture_state.src_sampler);
const auto& texture = texture_cache.GetImageView(draw_texture_state.src_texture);
const auto Scale = [&](auto dim) -> s32 {
return Settings::values.resolution_info.ScaleUp(s32(dim));
return Settings::values.resolution_info.ScaleUp(static_cast<s32>(dim));
};
Region2D dst_region = {
@@ -25,7 +25,7 @@
#include "shader_recompiler/frontend/maxwell/control_flow.h"
#include "shader_recompiler/frontend/maxwell/translate_program.h"
#include "shader_recompiler/profile.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/engines/draw_manager.h"
#include "video_core/engines/kepler_compute.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/memory_manager.h"
@@ -357,7 +357,7 @@ GraphicsPipeline* ShaderCache::CurrentGraphicsPipeline() {
const auto& regs{maxwell3d->regs};
graphics_key.raw = 0;
graphics_key.early_z.Assign(regs.mandated_early_z != 0 ? 1 : 0);
graphics_key.gs_input_topology.Assign(maxwell3d->draw_manager.draw_state.topology);
graphics_key.gs_input_topology.Assign(maxwell3d->draw_manager->GetDrawState().topology);
graphics_key.tessellation_primitive.Assign(regs.tessellation.params.domain_type.Value());
graphics_key.tessellation_spacing.Assign(regs.tessellation.params.spacing.Value());
graphics_key.tessellation_clockwise.Assign(
@@ -397,7 +397,7 @@ GraphicsPipeline* ShaderCache::BuiltPipeline(GraphicsPipeline* pipeline) const n
// If games are using a small index count, we can assume these are full screen quads.
// Usually these shaders are only used once for building textures so we can assume they
// can't be built async
const auto& draw_state = maxwell3d->draw_manager.draw_state;
const auto& draw_state = maxwell3d->draw_manager->GetDrawState();
if (draw_state.index_buffer.count <= 6 || draw_state.vertex_buffer.count <= 6) {
return pipeline;
}
@@ -12,7 +12,7 @@
#include "common/cityhash.h"
#include "common/common_types.h"
#include "common/settings.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/engines/draw_manager.h"
#include "video_core/renderer_vulkan/fixed_pipeline_state.h"
#include "video_core/renderer_vulkan/vk_state_tracker.h"
@@ -54,7 +54,7 @@ void RefreshXfbState(VideoCommon::TransformFeedbackState& state, const Maxwell&
void FixedPipelineState::Refresh(Tegra::Engines::Maxwell3D& maxwell3d, DynamicFeatures& features) {
const Maxwell& regs = maxwell3d.regs;
const auto topology_ = maxwell3d.draw_manager.draw_state.topology;
const auto topology_ = maxwell3d.draw_manager->GetDrawState().topology;
raw1 = 0;
extended_dynamic_state.Assign(features.has_extended_dynamic_state ? 1 : 0);
@@ -127,9 +127,8 @@ void ComputePipeline::Configure(Tegra::Engines::KeplerCompute& kepler_compute,
texture_cache.SynchronizeComputeDescriptors();
static constexpr size_t max_elements = 64;
boost::container::static_vector<VideoCommon::ImageViewInOut, max_elements> views;
boost::container::static_vector<VideoCommon::SamplerId, max_elements> samplers;
boost::container::small_vector<VideoCommon::ImageViewInOut, 64> views;
boost::container::small_vector<VideoCommon::SamplerId, 64> samplers;
const auto& qmd{kepler_compute.launch_description};
const auto& cbufs{qmd.const_buffer_config};
@@ -696,7 +696,7 @@ GraphicsPipeline* PipelineCache::BuiltPipeline(GraphicsPipeline* pipeline) const
// If games are using a small index count, we can assume these are full screen quads.
// Usually these shaders are only used once for building textures so we can assume they
// can't be built async
const auto& draw_state = maxwell3d->draw_manager.draw_state;
const auto& draw_state = maxwell3d->draw_manager->GetDrawState();
if (draw_state.index_buffer.count <= 6 || draw_state.vertex_buffer.count <= 6) {
return pipeline;
}
@@ -16,7 +16,7 @@
#include "video_core/renderer_vulkan/vk_texture_cache.h"
#include "common/bit_util.h"
#include "common/common_types.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/engines/draw_manager.h"
#include "video_core/host1x/gpu_device_memory_manager.h"
#include "video_core/query_cache/query_cache.h"
#include "video_core/rasterizer_interface.h"
@@ -950,7 +950,7 @@ private:
streams_mask = 0; // reset previously recorded streams
runtime.View3DRegs([this](Maxwell3D& maxwell3d) {
buffers_count = 0;
out_topology = maxwell3d.draw_manager.draw_state.topology;
out_topology = maxwell3d.draw_manager->GetDrawState().topology;
patch_vertices = std::max(maxwell3d.regs.patch_vertices, 1U);
if (out_topology == Maxwell3D::Regs::PrimitiveTopology::Patches) {
switch (maxwell3d.regs.tessellation.params.output_primitives.Value()) {
@@ -20,7 +20,7 @@
#include "video_core/buffer_cache/buffer_cache.h"
#include "video_core/gpu_logging/gpu_logging.h"
#include "video_core/control/channel_state.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/engines/draw_manager.h"
#include "video_core/engines/kepler_compute.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/host1x/gpu_device_memory_manager.h"
@@ -46,6 +46,7 @@
namespace Vulkan {
using Maxwell = Tegra::Engines::Maxwell3D::Regs;
using MaxwellDrawState = Tegra::Engines::DrawManager::State;
using VideoCommon::ImageViewId;
using VideoCommon::ImageViewType;
@@ -150,7 +151,7 @@ VkRect2D GetScissorState(const Maxwell& regs, size_t index, u32 up_scale = 1, u3
return scissor;
}
DrawParams MakeDrawParams(const Tegra::Engines::Maxwell3D::DrawManager::State& draw_state, u32 num_instances, bool is_indexed) {
DrawParams MakeDrawParams(const MaxwellDrawState& draw_state, u32 num_instances, bool is_indexed) {
DrawParams params{
.base_instance = draw_state.base_instance,
.num_instances = num_instances,
@@ -248,13 +249,15 @@ void RasterizerVulkan::PrepareDraw(bool is_indexed, Func&& draw_func) {
query_cache.NotifySegment(true);
HandleTransformFeedback();
query_cache.CounterEnable(VideoCommon::QueryType::ZPassPixelCount64, maxwell3d->regs.zpass_pixel_count_enable);
query_cache.CounterEnable(VideoCommon::QueryType::ZPassPixelCount64,
maxwell3d->regs.zpass_pixel_count_enable);
draw_func();
}
void RasterizerVulkan::Draw(bool is_indexed, u32 instance_count) {
PrepareDraw(is_indexed, [this, is_indexed, instance_count] {
const auto& draw_state = maxwell3d->draw_manager.draw_state;
const auto& draw_state = maxwell3d->draw_manager->GetDrawState();
const u32 num_instances{instance_count};
const DrawParams draw_params{MakeDrawParams(draw_state, num_instances, is_indexed)};
@@ -286,7 +289,7 @@ void RasterizerVulkan::Draw(bool is_indexed, u32 instance_count) {
}
void RasterizerVulkan::DrawIndirect() {
const auto& params = maxwell3d->draw_manager.indirect_state;
const auto& params = maxwell3d->draw_manager->GetIndirectParams();
buffer_cache.SetDrawIndirect(&params);
PrepareDraw(params.is_indexed, [this, &params] {
const auto indirect_buffer = buffer_cache.GetDrawIndirectBuffer();
@@ -357,8 +360,9 @@ void RasterizerVulkan::DrawTexture() {
UpdateDynamicStates();
query_cache.NotifySegment(true);
query_cache.CounterEnable(VideoCommon::QueryType::ZPassPixelCount64, maxwell3d->regs.zpass_pixel_count_enable);
const auto& draw_texture_state = maxwell3d->draw_manager.draw_texture_state;
query_cache.CounterEnable(VideoCommon::QueryType::ZPassPixelCount64,
maxwell3d->regs.zpass_pixel_count_enable);
const auto& draw_texture_state = maxwell3d->draw_manager->GetDrawTextureState();
const auto& sampler = texture_cache.GetGraphicsSampler(draw_texture_state.src_sampler);
const auto& texture = texture_cache.GetImageView(draw_texture_state.src_texture);
const auto* framebuffer = texture_cache.GetFramebuffer();
@@ -1010,7 +1014,7 @@ bool AccelerateDMA::BufferToImage(const Tegra::DMA::ImageCopy& copy_info,
void RasterizerVulkan::UpdateDynamicStates() {
auto& regs = maxwell3d->regs;
auto& flags = maxwell3d->dirty.flags;
const auto topology = maxwell3d->draw_manager.draw_state.topology;
const auto topology = maxwell3d->draw_manager->GetDrawState().topology;
if (state_tracker.ChangePrimitiveTopology(topology)) {
flags[Dirty::DepthBiasEnable] = true;
flags[Dirty::PrimitiveRestartEnable] = true;
@@ -1443,7 +1447,8 @@ void RasterizerVulkan::UpdatePrimitiveRestartEnable(Tegra::Engines::Maxwell3D::R
if (device.IsMoltenVK()) {
enable = true;
} else if (enable) {
const auto topology = MaxwellToVK::PrimitiveTopology(device, maxwell3d->draw_manager.draw_state.topology);
const auto topology =
MaxwellToVK::PrimitiveTopology(device, maxwell3d->draw_manager->GetDrawState().topology);
enable = IsPrimitiveRestartSupported(device, topology);
}
@@ -1557,9 +1562,10 @@ void RasterizerVulkan::UpdateDepthBiasEnable(Tegra::Engines::Maxwell3D::Regs& re
regs.polygon_offset_line_enable,
regs.polygon_offset_fill_enable,
};
const u32 topology_index = u32(maxwell3d->draw_manager.draw_state.topology);
const u32 topology_index = static_cast<u32>(maxwell3d->draw_manager->GetDrawState().topology);
const u32 enable = enabled_lut[POLYGON_OFFSET_ENABLE_LUT[topology_index]];
scheduler.Record([enable](vk::CommandBuffer cmdbuf) { cmdbuf.SetDepthBiasEnableEXT(enable != 0); });
scheduler.Record(
[enable](vk::CommandBuffer cmdbuf) { cmdbuf.SetDepthBiasEnableEXT(enable != 0); });
}
void RasterizerVulkan::UpdateLogicOpEnable(Tegra::Engines::Maxwell3D::Regs& regs) {
@@ -500,10 +500,6 @@ Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR
features.shader_atomic_int64.shaderSharedInt64Atomics = false;
features.features.shaderInt64 = false;
LOG_WARNING(Render_Vulkan,
"Qualcomm drivers have a slow VK_KHR_push_descriptor implementation");
RemoveExtension(extensions.push_descriptor, VK_KHR_PUSH_DESCRIPTOR_EXTENSION_NAME);
#if defined(ANDROID) && defined(ARCHITECTURE_arm64)
// BCn patching only safe on Android 9+ (API 28+). Older versions crash on driver load.
const auto major = (properties.properties.driverVersion >> 24) << 2;
-2
View File
@@ -245,8 +245,6 @@ add_executable(yuzu
render/performance_overlay.h render/performance_overlay.cpp render/performance_overlay.ui
libqt_common.h libqt_common.cpp
updater/update_dialog.h updater/update_dialog.cpp updater/update_dialog.ui
)
set_target_properties(yuzu PROPERTIES OUTPUT_NAME "eden")
@@ -330,6 +330,7 @@ void ConfigurePerGameAddons::LoadConfiguration() {
if (is_external_update) {
first_item->setData(static_cast<quint32>(patch.numeric_version), NUMERIC_VERSION);
} else if (is_mod) {
// qDebug() << patch.location;
first_item->setData(QString::fromStdString(patch.location), PATCH_LOCATION);
}
+15 -5
View File
@@ -8,7 +8,6 @@
#include "frontend_common/settings_generator.h"
#include "qt_common/qt_string_lookup.h"
#include "render/performance_overlay.h"
#include "updater/update_dialog.h"
#if defined(QT_STATICPLUGIN) && !defined(__APPLE__)
#undef VMA_IMPLEMENTATION
#endif
@@ -540,7 +539,7 @@ MainWindow::MainWindow(bool has_broken_vulkan)
#ifdef ENABLE_UPDATE_CHECKER
if (UISettings::values.check_for_updates) {
update_future = QtConcurrent::run(
[]() -> std::optional<Common::Net::Release> { return UpdateChecker::GetUpdate(); });
[]() -> std::optional<UpdateChecker::Update> { return UpdateChecker::GetUpdate(); });
update_watcher.connect(&update_watcher, &QFutureWatcher<QString>::finished, this,
&MainWindow::OnEmulatorUpdateAvailable);
update_watcher.setFuture(update_future);
@@ -4219,12 +4218,23 @@ void MainWindow::OnCaptureScreenshot() {
#ifdef ENABLE_UPDATE_CHECKER
void MainWindow::OnEmulatorUpdateAvailable() {
std::optional<Common::Net::Release> version = update_future.result();
std::optional<UpdateChecker::Update> version = update_future.result();
if (!version)
return;
UpdateDialog dialog(version.value(), this);
dialog.exec();
QMessageBox update_prompt(this);
update_prompt.setWindowTitle(tr("Update Available"));
update_prompt.setIcon(QMessageBox::Information);
update_prompt.addButton(QMessageBox::Yes);
update_prompt.addButton(QMessageBox::Ignore);
update_prompt.setText(tr("Download %1?").arg(QString::fromStdString(version->name)));
update_prompt.exec();
if (update_prompt.button(QMessageBox::Yes) == update_prompt.clickedButton()) {
auto const full_url =
fmt::format("{}/{}/releases/tag/", std::string{Common::g_build_auto_update_website},
std::string{Common::g_build_auto_update_repo});
QDesktopServices::openUrl(QUrl(QString::fromStdString(full_url + version->tag)));
}
}
#endif
+2 -2
View File
@@ -493,8 +493,8 @@ private:
std::shared_ptr<InputCommon::InputSubsystem> input_subsystem;
#ifdef ENABLE_UPDATE_CHECKER
QFuture<std::optional<Common::Net::Release>> update_future;
QFutureWatcher<std::optional<Common::Net::Release>> update_watcher;
QFuture<std::optional<UpdateChecker::Update>> update_future;
QFutureWatcher<std::optional<UpdateChecker::Update>> update_watcher;
#endif
MultiplayerState* multiplayer_state = nullptr;
-178
View File
@@ -1,178 +0,0 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
#include <QRadioButton>
#include <QSaveFile>
#include <QStandardPaths>
#include <qdesktopservices.h>
#include "common/logging.h"
#include "qt_common/abstract/frontend.h"
#include "qt_common/abstract/progress.h"
#include "ui_update_dialog.h"
#include "update_dialog.h"
#include "common/httplib.h"
#ifdef YUZU_BUNDLED_OPENSSL
#include <openssl/cert.h>
#endif
#include <QDesktopServices>
#undef GetSaveFileName
UpdateDialog::UpdateDialog(const Common::Net::Release& release, QWidget* parent)
: QDialog(parent), ui(new Ui::UpdateDialog) {
ui->setupUi(this);
ui->version->setText(
tr("%1 is available for download.").arg(QString::fromStdString(release.title)));
ui->url->setText(
tr("<a href=\"%1\">View on Forgejo</a>").arg(QString::fromStdString(release.html_url)));
std::string text{release.body};
if (auto pos = text.find("# Packages"); pos != std::string::npos) {
text = text.substr(0, pos);
}
ui->body->setMarkdown(QString::fromStdString(text));
// TODO(crueter): Find a way to set default
const auto assets = release.GetPlatformAssets();
if (assets.empty()) {
ui->groupBox->setHidden(true);
connect(this, &QDialog::accepted, this, [release]() {
QDesktopServices::openUrl(QUrl{QString::fromStdString(release.html_url)});
});
} else if (assets.size() == 1) {
m_asset = assets[0];
connect(this, &QDialog::accepted, this, &UpdateDialog::Download);
} else {
u32 i = 0;
for (const Common::Net::Asset& a : assets) {
QRadioButton* r = new QRadioButton(tr(a.name.c_str()), this);
connect(r, &QRadioButton::toggled, this, [a, this](bool checked) {
if (checked)
m_asset = a;
});
if (i == 0)
r->setChecked(true);
++i;
ui->radioButtons->addWidget(r);
}
connect(this, &QDialog::accepted, this, &UpdateDialog::Download);
}
}
UpdateDialog::~UpdateDialog() {
delete ui;
}
void UpdateDialog::Download() {
const auto filename = QtCommon::Frontend::GetSaveFileName(
tr("New Version Location"),
qApp->applicationDirPath() % QStringLiteral("/") % QString::fromStdString(m_asset.filename),
tr("All Files (*.*)"));
if (filename.isEmpty())
return;
QSaveFile file(filename);
if (!file.open(QIODevice::Truncate | QIODevice::WriteOnly)) {
LOG_WARNING(Frontend, "Could not open file {}", filename.toStdString());
QtCommon::Frontend::Critical(tr("Failed to save file"),
tr("Could not open file %1 for writing.").arg(filename));
return;
}
// TODO(crueter): Move to net.cpp
constexpr std::size_t timeout_seconds = 15;
std::unique_ptr<httplib::Client> client = std::make_unique<httplib::Client>(m_asset.url);
client->set_connection_timeout(timeout_seconds);
client->set_read_timeout(timeout_seconds);
client->set_write_timeout(timeout_seconds);
#ifdef YUZU_BUNDLED_OPENSSL
client->load_ca_cert_store(kCert, sizeof(kCert));
#endif
if (client == nullptr) {
LOG_ERROR(Frontend, "Invalid URL {}{}", m_asset.url, m_asset.path);
return;
}
auto progress =
QtCommon::Frontend::newProgressDialog(tr("Downloading..."), tr("Cancel"), 0, 100);
progress->show();
QGuiApplication::processEvents();
// Progress dialog.
auto progress_callback = [&](size_t processed_size, size_t total_size) {
QGuiApplication::processEvents();
progress->setValue(static_cast<int>((processed_size * 100) / total_size));
return !progress->wasCanceled();
};
// Write file in chunks.
auto content_receiver = [&file, filename](const char* t_data, size_t data_length) -> bool {
if (file.write(t_data, data_length) == -1) {
LOG_WARNING(Frontend, "Could not write {} bytes to file {}", data_length,
filename.toStdString());
QtCommon::Frontend::Critical(tr("Failed to save file"),
tr("Could not write to file %1.").arg(filename));
return false;
}
return true;
};
// Now send off request
auto result = client->Get(m_asset.path, content_receiver, progress_callback);
progress->close();
// commit to file
if (!file.commit()) {
LOG_WARNING(Frontend, "Could not commit to file {}", filename.toStdString());
QtCommon::Frontend::Critical(tr("Failed to save file"),
tr("Could not commit to file %1.").arg(filename));
}
if (!result) {
LOG_ERROR(Frontend, "GET to {}{} returned null", m_asset.url, m_asset.path);
return;
}
const auto& response = result.value();
if (response.status >= 400) {
LOG_ERROR(Frontend, "GET to {}{} returned error status code: {}", m_asset.url, m_asset.path,
response.status);
QtCommon::Frontend::Critical(tr("Failed to download file"),
tr("Could not download from %1%2\nError code: %3")
.arg(QString::fromStdString(m_asset.url),
QString::fromStdString(m_asset.path),
QString::number(response.status)));
return;
}
if (!response.headers.contains("content-type")) {
LOG_ERROR(Frontend, "GET to {}{} returned no content", m_asset.url, m_asset.path);
return;
}
// Download is complete. User may choose to open in the file manager.
auto button =
QtCommon::Frontend::Question(tr("Download Complete"),
tr("Successfully downloaded %1. Would you like to open it?")
.arg(QString::fromStdString(m_asset.filename)),
QtCommon::Frontend::Yes | QtCommon::Frontend::No);
if (button == QtCommon::Frontend::Yes) {
QDesktopServices::openUrl(QUrl::fromLocalFile(filename));
}
}
-28
View File
@@ -1,28 +0,0 @@
// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
#pragma once
#include <QDialog>
#include "common/net/net.h"
class QRadioButton;
namespace Ui {
class UpdateDialog;
}
class UpdateDialog : public QDialog {
Q_OBJECT
public:
explicit UpdateDialog(const Common::Net::Release &release, QWidget* parent = nullptr);
~UpdateDialog();
private slots:
void Download();
private:
Ui::UpdateDialog* ui;
QList<QRadioButton *> m_buttons;
Common::Net::Asset m_asset;
};
-112
View File
@@ -1,112 +0,0 @@
<?xml version="1.0" encoding="UTF-8"?>
<ui version="4.0">
<class>UpdateDialog</class>
<widget class="QDialog" name="UpdateDialog">
<property name="geometry">
<rect>
<x>0</x>
<y>0</y>
<width>655</width>
<height>551</height>
</rect>
</property>
<property name="windowTitle">
<string>Update Available</string>
</property>
<layout class="QGridLayout" name="gridLayout">
<item row="0" column="1">
<widget class="QLabel" name="url">
<property name="text">
<string>&lt;a href=&quot;%1&quot;&gt;View on Forgejo&lt;/a&gt;</string>
</property>
<property name="alignment">
<set>Qt::AlignmentFlag::AlignRight|Qt::AlignmentFlag::AlignTrailing|Qt::AlignmentFlag::AlignVCenter</set>
</property>
<property name="openExternalLinks">
<bool>true</bool>
</property>
</widget>
</item>
<item row="3" column="0" colspan="2">
<widget class="QLabel" name="label_2">
<property name="text">
<string>Would you like to install this update?</string>
</property>
</widget>
</item>
<item row="4" column="0" colspan="2">
<widget class="QDialogButtonBox" name="buttonBox">
<property name="orientation">
<enum>Qt::Orientation::Horizontal</enum>
</property>
<property name="standardButtons">
<set>QDialogButtonBox::StandardButton::No|QDialogButtonBox::StandardButton::Yes</set>
</property>
</widget>
</item>
<item row="2" column="0" colspan="2">
<widget class="QGroupBox" name="groupBox">
<property name="title">
<string>Available Versions</string>
</property>
<layout class="QVBoxLayout" name="radioButtons"/>
</widget>
</item>
<item row="0" column="0">
<widget class="QLabel" name="version">
<property name="text">
<string>%1 is available for download.</string>
</property>
</widget>
</item>
<item row="1" column="0" colspan="2">
<widget class="QTextBrowser" name="body">
<property name="horizontalScrollBarPolicy">
<enum>Qt::ScrollBarPolicy::ScrollBarAlwaysOff</enum>
</property>
<property name="readOnly">
<bool>true</bool>
</property>
<property name="openExternalLinks">
<bool>true</bool>
</property>
</widget>
</item>
</layout>
</widget>
<resources/>
<connections>
<connection>
<sender>buttonBox</sender>
<signal>accepted()</signal>
<receiver>UpdateDialog</receiver>
<slot>accept()</slot>
<hints>
<hint type="sourcelabel">
<x>248</x>
<y>254</y>
</hint>
<hint type="destinationlabel">
<x>157</x>
<y>274</y>
</hint>
</hints>
</connection>
<connection>
<sender>buttonBox</sender>
<signal>rejected()</signal>
<receiver>UpdateDialog</receiver>
<slot>reject()</slot>
<hints>
<hint type="sourcelabel">
<x>316</x>
<y>260</y>
</hint>
<hint type="destinationlabel">
<x>286</x>
<y>274</y>
</hint>
</hints>
</connection>
</connections>
</ui>