Merge pull request #2763 from lioncash/map-phys

kernel/vm_manager: Minor cleanup
Merge pull request #2804 from ReinUsesLisp/remove-gs-special
2019-09-09 16:43:54 +10:00 · 2019-09-05 16:03:46 -04:00 · 2019-09-05 15:27:31 -04:00 · 2019-09-05 12:29:50 +10:00 · 2019-09-05 12:26:39 +10:00 · 2019-09-05 12:13:14 +10:00
153 changed files with 3662 additions and 1645 deletions
@@ -10,7 +10,7 @@ jobs:
        BuildSuffix: 'windows-testing'
        ScriptFolder: 'windows'
  steps:
-  - script: sudo apt upgrade python3-pip && pip install requests urllib3
+  - script: sudo apt-get update && sudo apt-get --only-upgrade -y install python3-pip && pip install requests urllib3
    displayName: 'Prepare Environment'
  - task: PythonScript@0
    condition: eq(variables['Build.Reason'], 'PullRequest')
@@ -81,6 +81,7 @@ set(HASH_FILES
    "${VIDEO_CORE}/shader/decode/register_set_predicate.cpp"
    "${VIDEO_CORE}/shader/decode/shift.cpp"
    "${VIDEO_CORE}/shader/decode/video.cpp"
+    "${VIDEO_CORE}/shader/decode/warp.cpp"
    "${VIDEO_CORE}/shader/decode/xmad.cpp"
    "${VIDEO_CORE}/shader/control_flow.cpp"
    "${VIDEO_CORE}/shader/control_flow.h"
@@ -0,0 +1,31 @@
+The icons in this folder and its subfolders have the following licenses:
+
+Icon Name | License | Origin/Author
+--- | --- | ---
+qt_themes/default/icons/16x16/checked.png | Free for non-commercial use
+qt_themes/default/icons/16x16/failed.png | Free for non-commercial use
+qt_themes/default/icons/16x16/lock.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/default/icons/256x256/plus_folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/default/icons/48x48/bad_folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/default/icons/48x48/chip.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/default/icons/48x48/folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/default/icons/48x48/plus.png | CC0 1.0 | Designed by BreadFish64 from the Citra team
+qt_themes/default/icons/48x48/sd_card.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/qdarkstyle/icons/16x16/checked.png | Free for non-commercial use
+qt_themes/qdarkstyle/icons/16x16/failed.png | Free for non-commercial use
+qt_themes/qdarkstyle/icons/16x16/lock.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/qdarkstyle/icons/256x256/plus_folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/qdarkstyle/icons/48x48/bad_folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/qdarkstyle/icons/48x48/chip.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/qdarkstyle/icons/48x48/folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/qdarkstyle/icons/48x48/plus.png | CC0 1.0 | Designed by BreadFish64 from the Citra team
+qt_themes/qdarkstyle/icons/48x48/sd_card.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/16x16/lock.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/256x256/plus_folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/48x48/bad_folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/48x48/chip.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/48x48/folder.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/48x48/plus.png | CC BY-ND 3.0 | https://icons8.com
+qt_themes/colorful/icons/48x48/sd_card.png | CC BY-ND 3.0 | https://icons8.com
+
+<!-- TODO: Add the license of the yuzu icon -->
@@ -0,0 +1,14 @@
+[Icon Theme]
+Name=colorful
+Comment=Colorful theme
+Inherits=default
+Directories=16x16,48x48,256x256
+ 
+[16x16]
+Size=16
+
+[48x48]
+Size=48
+
+[256x256]
+Size=256
@@ -0,0 +1,15 @@
+<RCC>
+    <qresource prefix="icons/colorful">
+        <file alias="index.theme">icons/index.theme</file>
+        <file alias="16x16/lock.png">icons/16x16/lock.png</file>
+        <file alias="48x48/bad_folder.png">icons/48x48/bad_folder.png</file>
+        <file alias="48x48/chip.png">icons/48x48/chip.png</file>
+        <file alias="48x48/folder.png">icons/48x48/folder.png</file>
+        <file alias="48x48/plus.png">icons/48x48/plus.png</file>
+        <file alias="48x48/sd_card.png">icons/48x48/sd_card.png</file>
+        <file alias="256x256/plus_folder.png">icons/256x256/plus_folder.png</file>
+    </qresource>
+    <qresource prefix="colorful">
+        <file>style.qss</file>
+    </qresource>
+</RCC>
@@ -0,0 +1,4 @@
+/*
+    This file is intentionally left blank.
+    We do not want to apply any stylesheet for colorful, only icons.
+*/
@@ -0,0 +1,8 @@
+[Icon Theme]
+Name=colorful_dark
+Comment=Colorful theme (Dark style)
+Inherits=default
+Directories=16x16
+ 
+[16x16]
+Size=16
@@ -0,0 +1,57 @@
+<RCC>
+    <qresource prefix="icons/colorful_dark">
+        <file alias="index.theme">icons/index.theme</file>
+        <file alias="16x16/lock.png">icons/16x16/lock.png</file>
+        <file alias="48x48/bad_folder.png">../colorful/icons/48x48/bad_folder.png</file>
+        <file alias="48x48/chip.png">../colorful/icons/48x48/chip.png</file>
+        <file alias="48x48/folder.png">../colorful/icons/48x48/folder.png</file>
+        <file alias="48x48/plus.png">../colorful/icons/48x48/plus.png</file>
+        <file alias="48x48/sd_card.png">../colorful/icons/48x48/sd_card.png</file>
+        <file alias="256x256/plus_folder.png">../colorful/icons/256x256/plus_folder.png</file>
+    </qresource>
+
+    <qresource prefix="qss_icons">
+        <file alias="rc/up_arrow_disabled.png">../qdarkstyle/rc/up_arrow_disabled.png</file>
+        <file alias="rc/Hmovetoolbar.png">../qdarkstyle/rc/Hmovetoolbar.png</file>
+        <file alias="rc/stylesheet-branch-end.png">../qdarkstyle/rc/stylesheet-branch-end.png</file>
+        <file alias="rc/branch_closed-on.png">../qdarkstyle/rc/branch_closed-on.png</file>
+        <file alias="rc/stylesheet-vline.png">../qdarkstyle/rc/stylesheet-vline.png</file>
+        <file alias="rc/branch_closed.png">../qdarkstyle/rc/branch_closed.png</file>
+        <file alias="rc/branch_open-on.png">../qdarkstyle/rc/branch_open-on.png</file>
+        <file alias="rc/transparent.png">../qdarkstyle/rc/transparent.png</file>
+        <file alias="rc/right_arrow_disabled.png">../qdarkstyle/rc/right_arrow_disabled.png</file>
+        <file alias="rc/sizegrip.png">../qdarkstyle/rc/sizegrip.png</file>
+        <file alias="rc/close.png">../qdarkstyle/rc/close.png</file>
+        <file alias="rc/close-hover.png">../qdarkstyle/rc/close-hover.png</file>
+        <file alias="rc/close-pressed.png">../qdarkstyle/rc/close-pressed.png</file>
+        <file alias="rc/down_arrow.png">../qdarkstyle/rc/down_arrow.png</file>
+        <file alias="rc/Vmovetoolbar.png">../qdarkstyle/rc/Vmovetoolbar.png</file>
+        <file alias="rc/left_arrow.png">../qdarkstyle/rc/left_arrow.png</file>
+        <file alias="rc/stylesheet-branch-more.png">../qdarkstyle/rc/stylesheet-branch-more.png</file>
+        <file alias="rc/up_arrow.png">../qdarkstyle/rc/up_arrow.png</file>
+        <file alias="rc/right_arrow.png">../qdarkstyle/rc/right_arrow.png</file>
+        <file alias="rc/left_arrow_disabled.png">../qdarkstyle/rc/left_arrow_disabled.png</file>
+        <file alias="rc/Hsepartoolbar.png">../qdarkstyle/rc/Hsepartoolbar.png</file>
+        <file alias="rc/branch_open.png">../qdarkstyle/rc/branch_open.png</file>
+        <file alias="rc/Vsepartoolbar.png">../qdarkstyle/rc/Vsepartoolbar.png</file>
+        <file alias="rc/down_arrow_disabled.png">../qdarkstyle/rc/down_arrow_disabled.png</file>
+        <file alias="rc/undock.png">../qdarkstyle/rc/undock.png</file>
+        <file alias="rc/checkbox_checked_disabled.png">../qdarkstyle/rc/checkbox_checked_disabled.png</file>
+        <file alias="rc/checkbox_checked_focus.png">../qdarkstyle/rc/checkbox_checked_focus.png</file>
+        <file alias="rc/checkbox_checked.png">../qdarkstyle/rc/checkbox_checked.png</file>
+        <file alias="rc/checkbox_indeterminate.png">../qdarkstyle/rc/checkbox_indeterminate.png</file>
+        <file alias="rc/checkbox_indeterminate_focus.png">../qdarkstyle/rc/checkbox_indeterminate_focus.png</file>
+        <file alias="rc/checkbox_unchecked_disabled.png">../qdarkstyle/rc/checkbox_unchecked_disabled.png</file>
+        <file alias="rc/checkbox_unchecked_focus.png">../qdarkstyle/rc/checkbox_unchecked_focus.png</file>
+        <file alias="rc/checkbox_unchecked.png">../qdarkstyle/rc/checkbox_unchecked.png</file>
+        <file alias="rc/radio_checked_disabled.png">../qdarkstyle/rc/radio_checked_disabled.png</file>
+        <file alias="rc/radio_checked_focus.png">../qdarkstyle/rc/radio_checked_focus.png</file>
+        <file alias="rc/radio_checked.png">../qdarkstyle/rc/radio_checked.png</file>
+        <file alias="rc/radio_unchecked_disabled.png">../qdarkstyle/rc/radio_unchecked_disabled.png</file>
+        <file alias="rc/radio_unchecked_focus.png">../qdarkstyle/rc/radio_unchecked_focus.png</file>
+        <file alias="rc/radio_unchecked.png">../qdarkstyle/rc/radio_unchecked.png</file>
+    </qresource>
+    <qresource prefix="colorful_dark">
+        <file alias="style.qss">../qdarkstyle/style.qss</file>
+    </qresource>
+</RCC>
@@ -5,7 +5,21 @@
        <file alias="16x16/checked.png">icons/16x16/checked.png</file>

        <file alias="16x16/failed.png">icons/16x16/failed.png</file>
+
+        <file alias="16x16/lock.png">icons/16x16/lock.png</file>
+
+        <file alias="48x48/bad_folder.png">icons/48x48/bad_folder.png</file>
+      
+        <file alias="48x48/chip.png">icons/48x48/chip.png</file>
+
+        <file alias="48x48/folder.png">icons/48x48/folder.png</file>
+
+        <file alias="48x48/plus.png">icons/48x48/plus.png</file>
+      
+        <file alias="48x48/sd_card.png">icons/48x48/sd_card.png</file>
      
        <file alias="256x256/yuzu.png">icons/256x256/yuzu.png</file>
+
+        <file alias="256x256/plus_folder.png">icons/256x256/plus_folder.png</file>
    </qresource>
 </RCC>
@@ -1,10 +1,13 @@
 [Icon Theme]
 Name=default
 Comment=default theme
-Directories=16x16,256x256
+Directories=16x16,48x48,256x256
 
 [16x16]
 Size=16
+
+[48x48]
+Size=48
 
 [256x256]
 Size=256
@@ -2,10 +2,13 @@
 Name=qdarkstyle
 Comment=dark theme
 Inherits=default
-Directories=16x16,256x256
+Directories=16x16,48x48,256x256
 
 [16x16]
 Size=16
- 
+
+[48x48]
+Size=48
+
 [256x256]
 Size=256
@@ -1,6 +1,13 @@
 <RCC>
  <qresource prefix="icons/qdarkstyle">
    <file alias="index.theme">icons/index.theme</file>
+    <file alias="16x16/lock.png">icons/16x16/lock.png</file>
+    <file alias="48x48/bad_folder.png">icons/48x48/bad_folder.png</file>
+    <file alias="48x48/chip.png">icons/48x48/chip.png</file>
+    <file alias="48x48/folder.png">icons/48x48/folder.png</file>
+    <file alias="48x48/plus.png">icons/48x48/plus.png</file>
+    <file alias="48x48/sd_card.png">icons/48x48/sd_card.png</file>
+    <file alias="256x256/plus_folder.png">icons/256x256/plus_folder.png</file>
  </qresource>
  <qresource prefix="qss_icons">
    <file>rc/up_arrow_disabled.png</file>
@@ -337,3 +337,19 @@ proprietary programs.  If your program is a subroutine library, you may
 consider it more useful to permit linking proprietary applications with the
 library.  If this is what you want to do, use the GNU Lesser General
 Public License instead of this License.
+
+
+The icons used in this project have the following licenses:
+
+Icon Name                          | License       | Origin/Author
+---                                | ---           | ---
+checked.png                        | Free for non-commercial use
+failed.png                         | Free for non-commercial use
+lock.png                           | CC BY-ND 3.0  | https://icons8.com
+plus_folder.png                    | CC BY-ND 3.0  | https://icons8.com
+bad_folder.png                     | CC BY-ND 3.0  | https://icons8.com
+chip.png                           | CC BY-ND 3.0  | https://icons8.com
+folder.png                         | CC BY-ND 3.0  | https://icons8.com
+plus.png (Default, Dark)           | CC0 1.0       | Designed by BreadFish64 from the Citra team
+plus.png (Colorful, Colorful Dark) | CC BY-ND 3.0  | https://icons8.com
+sd_card.png | CC BY-ND 3.0 | https://icons8.com
@@ -55,6 +55,7 @@ add_custom_command(OUTPUT scm_rev.cpp
      "${VIDEO_CORE}/shader/decode/register_set_predicate.cpp"
      "${VIDEO_CORE}/shader/decode/shift.cpp"
      "${VIDEO_CORE}/shader/decode/video.cpp"
+      "${VIDEO_CORE}/shader/decode/warp.cpp"
      "${VIDEO_CORE}/shader/decode/xmad.cpp"
      "${VIDEO_CORE}/shader/control_flow.cpp"
      "${VIDEO_CORE}/shader/control_flow.h"
@@ -3,6 +3,7 @@
 #pragma once

 #include <cstddef>
+#include <memory>
 #include <type_traits>

 namespace Common {
@@ -37,4 +38,63 @@ constexpr bool IsWordAligned(T value) {
    return (value & 0b11) == 0;
 }

+template <typename T, std::size_t Align = 16>
+class AlignmentAllocator {
+public:
+    using value_type = T;
+    using size_type = std::size_t;
+    using difference_type = std::ptrdiff_t;
+
+    using pointer = T*;
+    using const_pointer = const T*;
+
+    using reference = T&;
+    using const_reference = const T&;
+
+public:
+    pointer address(reference r) noexcept {
+        return std::addressof(r);
+    }
+
+    const_pointer address(const_reference r) const noexcept {
+        return std::addressof(r);
+    }
+
+    pointer allocate(size_type n) {
+        return static_cast<pointer>(::operator new (n, std::align_val_t{Align}));
+    }
+
+    void deallocate(pointer p, size_type) {
+        ::operator delete (p, std::align_val_t{Align});
+    }
+
+    void construct(pointer p, const value_type& wert) {
+        new (p) value_type(wert);
+    }
+
+    void destroy(pointer p) {
+        p->~value_type();
+    }
+
+    size_type max_size() const noexcept {
+        return size_type(-1) / sizeof(value_type);
+    }
+
+    template <typename T2>
+    struct rebind {
+        using other = AlignmentAllocator<T2, Align>;
+    };
+
+    bool operator!=(const AlignmentAllocator<T, Align>& other) const noexcept {
+        return !(*this == other);
+    }
+
+    // Returns true if and only if storage allocated from *this
+    // can be deallocated from other, and vice versa.
+    // Always returns true for stateless allocators.
+    bool operator==(const AlignmentAllocator<T, Align>& other) const noexcept {
+        return true;
+    }
+};
+
 } // namespace Common
@@ -70,6 +70,8 @@ add_library(core STATIC
    file_sys/sdmc_factory.h
    file_sys/submission_package.cpp
    file_sys/submission_package.h
+    file_sys/system_archive/mii_model.cpp
+    file_sys/system_archive/mii_model.h
    file_sys/system_archive/ng_word.cpp
    file_sys/system_archive/ng_word.h
    file_sys/system_archive/system_archive.cpp
@@ -37,6 +37,7 @@
 namespace Core::Crypto {

 constexpr u64 CURRENT_CRYPTO_REVISION = 0x5;
+constexpr u64 FULL_TICKET_SIZE = 0x400;

 using namespace Common;

@@ -55,6 +56,99 @@ const std::map<std::pair<S128KeyType, u64>, std::string> KEYS_VARIABLE_LENGTH{
    {{S128KeyType::KeyblobMAC, 0}, "keyblob_mac_key_"},
 };

+namespace {
+template <std::size_t Size>
+bool IsAllZeroArray(const std::array<u8, Size>& array) {
+    return std::all_of(array.begin(), array.end(), [](const auto& elem) { return elem == 0; });
+}
+} // namespace
+
+u64 GetSignatureTypeDataSize(SignatureType type) {
+    switch (type) {
+    case SignatureType::RSA_4096_SHA1:
+    case SignatureType::RSA_4096_SHA256:
+        return 0x200;
+    case SignatureType::RSA_2048_SHA1:
+    case SignatureType::RSA_2048_SHA256:
+        return 0x100;
+    case SignatureType::ECDSA_SHA1:
+    case SignatureType::ECDSA_SHA256:
+        return 0x3C;
+    }
+    UNREACHABLE();
+}
+
+u64 GetSignatureTypePaddingSize(SignatureType type) {
+    switch (type) {
+    case SignatureType::RSA_4096_SHA1:
+    case SignatureType::RSA_4096_SHA256:
+    case SignatureType::RSA_2048_SHA1:
+    case SignatureType::RSA_2048_SHA256:
+        return 0x3C;
+    case SignatureType::ECDSA_SHA1:
+    case SignatureType::ECDSA_SHA256:
+        return 0x40;
+    }
+    UNREACHABLE();
+}
+
+SignatureType Ticket::GetSignatureType() const {
+    if (auto ticket = std::get_if<RSA4096Ticket>(&data)) {
+        return ticket->sig_type;
+    }
+    if (auto ticket = std::get_if<RSA2048Ticket>(&data)) {
+        return ticket->sig_type;
+    }
+    if (auto ticket = std::get_if<ECDSATicket>(&data)) {
+        return ticket->sig_type;
+    }
+
+    UNREACHABLE();
+}
+
+TicketData& Ticket::GetData() {
+    if (auto ticket = std::get_if<RSA4096Ticket>(&data)) {
+        return ticket->data;
+    }
+    if (auto ticket = std::get_if<RSA2048Ticket>(&data)) {
+        return ticket->data;
+    }
+    if (auto ticket = std::get_if<ECDSATicket>(&data)) {
+        return ticket->data;
+    }
+
+    UNREACHABLE();
+}
+
+const TicketData& Ticket::GetData() const {
+    if (auto ticket = std::get_if<RSA4096Ticket>(&data)) {
+        return ticket->data;
+    }
+    if (auto ticket = std::get_if<RSA2048Ticket>(&data)) {
+        return ticket->data;
+    }
+    if (auto ticket = std::get_if<ECDSATicket>(&data)) {
+        return ticket->data;
+    }
+
+    UNREACHABLE();
+}
+
+u64 Ticket::GetSize() const {
+    const auto sig_type = GetSignatureType();
+
+    return sizeof(SignatureType) + GetSignatureTypeDataSize(sig_type) +
+           GetSignatureTypePaddingSize(sig_type) + sizeof(TicketData);
+}
+
+Ticket Ticket::SynthesizeCommon(Key128 title_key, const std::array<u8, 16>& rights_id) {
+    RSA2048Ticket out{};
+    out.sig_type = SignatureType::RSA_2048_SHA256;
+    out.data.rights_id = rights_id;
+    out.data.title_key_common = title_key;
+    return Ticket{out};
+}
+
 Key128 GenerateKeyEncryptionKey(Key128 source, Key128 master, Key128 kek_seed, Key128 key_seed) {
    Key128 out{};

@@ -135,6 +229,27 @@ void KeyManager::DeriveGeneralPurposeKeys(std::size_t crypto_revision) {
    }
 }

+RSAKeyPair<2048> KeyManager::GetETicketRSAKey() const {
+    if (IsAllZeroArray(eticket_extended_kek) || !HasKey(S128KeyType::ETicketRSAKek))
+        return {};
+
+    const auto eticket_final = GetKey(S128KeyType::ETicketRSAKek);
+
+    std::vector<u8> extended_iv(eticket_extended_kek.begin(), eticket_extended_kek.begin() + 0x10);
+    std::array<u8, 0x230> extended_dec{};
+    AESCipher<Key128> rsa_1(eticket_final, Mode::CTR);
+    rsa_1.SetIV(extended_iv);
+    rsa_1.Transcode(eticket_extended_kek.data() + 0x10, eticket_extended_kek.size() - 0x10,
+                    extended_dec.data(), Op::Decrypt);
+
+    RSAKeyPair<2048> rsa_key{};
+    std::memcpy(rsa_key.decryption_key.data(), extended_dec.data(), rsa_key.decryption_key.size());
+    std::memcpy(rsa_key.modulus.data(), extended_dec.data() + 0x100, rsa_key.modulus.size());
+    std::memcpy(rsa_key.exponent.data(), extended_dec.data() + 0x200, rsa_key.exponent.size());
+
+    return rsa_key;
+}
+
 Key128 DeriveKeyblobMACKey(const Key128& keyblob_key, const Key128& mac_source) {
    AESCipher<Key128> mac_cipher(keyblob_key, Mode::ECB);
    Key128 mac_key{};
@@ -237,7 +352,7 @@ Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& ke
    return Loader::ResultStatus::Success;
 }

-std::vector<TicketRaw> GetTicketblob(const FileUtil::IOFile& ticket_save) {
+std::vector<Ticket> GetTicketblob(const FileUtil::IOFile& ticket_save) {
    if (!ticket_save.IsOpen())
        return {};

@@ -246,14 +361,14 @@ std::vector<TicketRaw> GetTicketblob(const FileUtil::IOFile& ticket_save) {
        return {};
    }

-    std::vector<TicketRaw> out;
+    std::vector<Ticket> out;
    for (std::size_t offset = 0; offset + 0x4 < buffer.size(); ++offset) {
        if (buffer[offset] == 0x4 && buffer[offset + 1] == 0x0 && buffer[offset + 2] == 0x1 &&
            buffer[offset + 3] == 0x0) {
            out.emplace_back();
            auto& next = out.back();
-            std::memcpy(&next, buffer.data() + offset, sizeof(TicketRaw));
-            offset += next.size();
+            std::memcpy(&next, buffer.data() + offset, sizeof(Ticket));
+            offset += FULL_TICKET_SIZE;
        }
    }

@@ -305,29 +420,23 @@ static std::optional<u64> FindTicketOffset(const std::array<u8, size>& data) {
    return offset;
 }

-std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
+std::optional<std::pair<Key128, Key128>> ParseTicket(const Ticket& ticket,
                                                     const RSAKeyPair<2048>& key) {
-    u32 cert_authority;
-    std::memcpy(&cert_authority, ticket.data() + 0x140, sizeof(cert_authority));
-    if (cert_authority == 0)
+    const auto issuer = ticket.GetData().issuer;
+    if (issuer == std::array<u8, 0x40>{})
        return {};
-    if (cert_authority != Common::MakeMagic('R', 'o', 'o', 't')) {
-        LOG_INFO(Crypto,
-                 "Attempting to parse ticket with non-standard certificate authority {:08X}.",
-                 cert_authority);
+    if (issuer[0] != 'R' || issuer[1] != 'o' || issuer[2] != 'o' || issuer[3] != 't') {
+        LOG_INFO(Crypto, "Attempting to parse ticket with non-standard certificate authority.");
    }

-    Key128 rights_id;
-    std::memcpy(rights_id.data(), ticket.data() + 0x2A0, sizeof(Key128));
+    Key128 rights_id = ticket.GetData().rights_id;

    if (rights_id == Key128{})
        return {};

-    Key128 key_temp{};
-
-    if (!std::any_of(ticket.begin() + 0x190, ticket.begin() + 0x280, [](u8 b) { return b != 0; })) {
-        std::memcpy(key_temp.data(), ticket.data() + 0x180, key_temp.size());
-        return std::make_pair(rights_id, key_temp);
+    if (!std::any_of(ticket.GetData().title_key_common_pad.begin(),
+                     ticket.GetData().title_key_common_pad.end(), [](u8 b) { return b != 0; })) {
+        return std::make_pair(rights_id, ticket.GetData().title_key_common);
    }

    mbedtls_mpi D; // RSA Private Exponent
@@ -342,7 +451,7 @@ std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,

    mbedtls_mpi_read_binary(&D, key.decryption_key.data(), key.decryption_key.size());
    mbedtls_mpi_read_binary(&N, key.modulus.data(), key.modulus.size());
-    mbedtls_mpi_read_binary(&S, ticket.data() + 0x180, 0x100);
+    mbedtls_mpi_read_binary(&S, ticket.GetData().title_key_block.data(), 0x100);

    mbedtls_mpi_exp_mod(&M, &S, &D, &N, nullptr);

@@ -366,6 +475,7 @@ std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
        return {};
    ASSERT(*offset > 0);

+    Key128 key_temp{};
    std::memcpy(key_temp.data(), m_2.data() + *offset, key_temp.size());

    return std::make_pair(rights_id, key_temp);
@@ -450,6 +560,8 @@ void KeyManager::LoadFromFile(const std::string& filename, bool is_title_keys) {

                const auto index = std::stoul(out[0].substr(18, 2), nullptr, 16);
                encrypted_keyblobs[index] = Common::HexStringToArray<0xB0>(out[1]);
+            } else if (out[0].compare(0, 20, "eticket_extended_kek") == 0) {
+                eticket_extended_kek = Common::HexStringToArray<576>(out[1]);
            } else {
                for (const auto& kv : KEYS_VARIABLE_LENGTH) {
                    if (!ValidCryptoRevisionString(out[0], kv.second.size(), 2))
@@ -862,20 +974,19 @@ void KeyManager::DeriveETicket(PartitionDataManager& data) {
    // Titlekeys
    data.DecryptProdInfo(GetBISKey(0));

-    const auto eticket_extended_kek = data.GetETicketExtendedKek();
+    eticket_extended_kek = data.GetETicketExtendedKek();
+    WriteKeyToFile(KeyCategory::Console, "eticket_extended_kek", eticket_extended_kek);
+    PopulateTickets();
+}

-    std::vector<u8> extended_iv(0x10);
-    std::memcpy(extended_iv.data(), eticket_extended_kek.data(), extended_iv.size());
-    std::array<u8, 0x230> extended_dec{};
-    AESCipher<Key128> rsa_1(eticket_final, Mode::CTR);
-    rsa_1.SetIV(extended_iv);
-    rsa_1.Transcode(eticket_extended_kek.data() + 0x10, eticket_extended_kek.size() - 0x10,
-                    extended_dec.data(), Op::Decrypt);
+void KeyManager::PopulateTickets() {
+    const auto rsa_key = GetETicketRSAKey();

-    RSAKeyPair<2048> rsa_key{};
-    std::memcpy(rsa_key.decryption_key.data(), extended_dec.data(), rsa_key.decryption_key.size());
-    std::memcpy(rsa_key.modulus.data(), extended_dec.data() + 0x100, rsa_key.modulus.size());
-    std::memcpy(rsa_key.exponent.data(), extended_dec.data() + 0x200, rsa_key.exponent.size());
+    if (rsa_key == RSAKeyPair<2048>{})
+        return;
+
+    if (!common_tickets.empty() && !personal_tickets.empty())
+        return;

    const FileUtil::IOFile save1(FileUtil::GetUserPath(FileUtil::UserPath::NANDDir) +
                                     "/system/save/80000000000000e1",
@@ -886,19 +997,41 @@ void KeyManager::DeriveETicket(PartitionDataManager& data) {

    const auto blob2 = GetTicketblob(save2);
    auto res = GetTicketblob(save1);
+    const auto idx = res.size();
    res.insert(res.end(), blob2.begin(), blob2.end());

-    for (const auto& raw : res) {
-        const auto pair = ParseTicket(raw, rsa_key);
+    for (std::size_t i = 0; i < res.size(); ++i) {
+        const auto common = i < idx;
+        const auto pair = ParseTicket(res[i], rsa_key);
        if (!pair)
            continue;
        const auto& [rid, key] = *pair;
        u128 rights_id;
        std::memcpy(rights_id.data(), rid.data(), rid.size());
+
+        if (common) {
+            common_tickets[rights_id] = res[i];
+        } else {
+            personal_tickets[rights_id] = res[i];
+        }
+
        SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);
    }
 }

+void KeyManager::SynthesizeTickets() {
+    for (const auto& key : s128_keys) {
+        if (key.first.type != S128KeyType::Titlekey) {
+            continue;
+        }
+        u128 rights_id{key.first.field1, key.first.field2};
+        Key128 rights_id_2;
+        std::memcpy(rights_id_2.data(), rights_id.data(), rights_id_2.size());
+        const auto ticket = Ticket::SynthesizeCommon(key.second, rights_id_2);
+        common_tickets.insert_or_assign(rights_id, ticket);
+    }
+}
+
 void KeyManager::SetKeyWrapped(S128KeyType id, Key128 key, u64 field1, u64 field2) {
    if (key == Key128{})
        return;
@@ -997,6 +1130,46 @@ void KeyManager::PopulateFromPartitionData(PartitionDataManager& data) {
    DeriveBase();
 }

+const std::map<u128, Ticket>& KeyManager::GetCommonTickets() const {
+    return common_tickets;
+}
+
+const std::map<u128, Ticket>& KeyManager::GetPersonalizedTickets() const {
+    return personal_tickets;
+}
+
+bool KeyManager::AddTicketCommon(Ticket raw) {
+    const auto rsa_key = GetETicketRSAKey();
+    if (rsa_key == RSAKeyPair<2048>{})
+        return false;
+
+    const auto pair = ParseTicket(raw, rsa_key);
+    if (!pair)
+        return false;
+    const auto& [rid, key] = *pair;
+    u128 rights_id;
+    std::memcpy(rights_id.data(), rid.data(), rid.size());
+    common_tickets[rights_id] = raw;
+    SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);
+    return true;
+}
+
+bool KeyManager::AddTicketPersonalized(Ticket raw) {
+    const auto rsa_key = GetETicketRSAKey();
+    if (rsa_key == RSAKeyPair<2048>{})
+        return false;
+
+    const auto pair = ParseTicket(raw, rsa_key);
+    if (!pair)
+        return false;
+    const auto& [rid, key] = *pair;
+    u128 rights_id;
+    std::memcpy(rights_id.data(), rid.data(), rid.size());
+    common_tickets[rights_id] = raw;
+    SetKey(S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);
+    return true;
+}
+
 const boost::container::flat_map<std::string, KeyIndex<S128KeyType>> KeyManager::s128_file_id = {
    {"eticket_rsa_kek", {S128KeyType::ETicketRSAKek, 0, 0}},
    {"eticket_rsa_kek_source",
@@ -9,8 +9,10 @@
 #include <optional>
 #include <string>

+#include <variant>
 #include <boost/container/flat_map.hpp>
 #include <fmt/format.h>
+#include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "core/crypto/partition_data_manager.h"
 #include "core/file_sys/vfs_types.h"
@@ -30,7 +32,79 @@ constexpr u64 TICKET_FILE_TITLEKEY_OFFSET = 0x180;
 using Key128 = std::array<u8, 0x10>;
 using Key256 = std::array<u8, 0x20>;
 using SHA256Hash = std::array<u8, 0x20>;
-using TicketRaw = std::array<u8, 0x400>;
+
+enum class SignatureType {
+    RSA_4096_SHA1 = 0x10000,
+    RSA_2048_SHA1 = 0x10001,
+    ECDSA_SHA1 = 0x10002,
+    RSA_4096_SHA256 = 0x10003,
+    RSA_2048_SHA256 = 0x10004,
+    ECDSA_SHA256 = 0x10005,
+};
+
+u64 GetSignatureTypeDataSize(SignatureType type);
+u64 GetSignatureTypePaddingSize(SignatureType type);
+
+enum class TitleKeyType : u8 {
+    Common = 0,
+    Personalized = 1,
+};
+
+struct TicketData {
+    std::array<u8, 0x40> issuer;
+    union {
+        std::array<u8, 0x100> title_key_block;
+
+        struct {
+            Key128 title_key_common;
+            std::array<u8, 0xF0> title_key_common_pad;
+        };
+    };
+
+    INSERT_PADDING_BYTES(0x1);
+    TitleKeyType type;
+    INSERT_PADDING_BYTES(0x3);
+    u8 revision;
+    INSERT_PADDING_BYTES(0xA);
+    u64 ticket_id;
+    u64 device_id;
+    std::array<u8, 0x10> rights_id;
+    u32 account_id;
+    INSERT_PADDING_BYTES(0x14C);
+};
+static_assert(sizeof(TicketData) == 0x2C0, "TicketData has incorrect size.");
+
+struct RSA4096Ticket {
+    SignatureType sig_type;
+    std::array<u8, 0x200> sig_data;
+    INSERT_PADDING_BYTES(0x3C);
+    TicketData data;
+};
+
+struct RSA2048Ticket {
+    SignatureType sig_type;
+    std::array<u8, 0x100> sig_data;
+    INSERT_PADDING_BYTES(0x3C);
+    TicketData data;
+};
+
+struct ECDSATicket {
+    SignatureType sig_type;
+    std::array<u8, 0x3C> sig_data;
+    INSERT_PADDING_BYTES(0x40);
+    TicketData data;
+};
+
+struct Ticket {
+    std::variant<RSA4096Ticket, RSA2048Ticket, ECDSATicket> data;
+
+    SignatureType GetSignatureType() const;
+    TicketData& GetData();
+    const TicketData& GetData() const;
+    u64 GetSize() const;
+
+    static Ticket SynthesizeCommon(Key128 title_key, const std::array<u8, 0x10>& rights_id);
+};

 static_assert(sizeof(Key128) == 16, "Key128 must be 128 bytes big.");
 static_assert(sizeof(Key256) == 32, "Key256 must be 256 bytes big.");
@@ -43,6 +117,19 @@ struct RSAKeyPair {
    std::array<u8, 4> exponent;
 };

+template <size_t bit_size, size_t byte_size>
+bool operator==(const RSAKeyPair<bit_size, byte_size>& lhs,
+                const RSAKeyPair<bit_size, byte_size>& rhs) {
+    return std::tie(lhs.encryption_key, lhs.decryption_key, lhs.modulus, lhs.exponent) ==
+           std::tie(rhs.encryption_key, rhs.decryption_key, rhs.modulus, rhs.exponent);
+}
+
+template <size_t bit_size, size_t byte_size>
+bool operator!=(const RSAKeyPair<bit_size, byte_size>& lhs,
+                const RSAKeyPair<bit_size, byte_size>& rhs) {
+    return !(lhs == rhs);
+}
+
 enum class KeyCategory : u8 {
    Standard,
    Title,
@@ -151,22 +238,35 @@ public:

    static bool KeyFileExists(bool title);

-    // Call before using the sd seed to attempt to derive it if it dosen't exist. Needs system save
-    // 8*43 and the private file to exist.
+    // Call before using the sd seed to attempt to derive it if it dosen't exist. Needs system
+    // save 8*43 and the private file to exist.
    void DeriveSDSeedLazy();

    bool BaseDeriveNecessary() const;
    void DeriveBase();
    void DeriveETicket(PartitionDataManager& data);
+    void PopulateTickets();
+    void SynthesizeTickets();

    void PopulateFromPartitionData(PartitionDataManager& data);

+    const std::map<u128, Ticket>& GetCommonTickets() const;
+    const std::map<u128, Ticket>& GetPersonalizedTickets() const;
+
+    bool AddTicketCommon(Ticket raw);
+    bool AddTicketPersonalized(Ticket raw);
+
 private:
    std::map<KeyIndex<S128KeyType>, Key128> s128_keys;
    std::map<KeyIndex<S256KeyType>, Key256> s256_keys;

+    // Map from rights ID to ticket
+    std::map<u128, Ticket> common_tickets;
+    std::map<u128, Ticket> personal_tickets;
+
    std::array<std::array<u8, 0xB0>, 0x20> encrypted_keyblobs{};
    std::array<std::array<u8, 0x90>, 0x20> keyblobs{};
+    std::array<u8, 576> eticket_extended_kek{};

    bool dev_mode;
    void LoadFromFile(const std::string& filename, bool is_title_keys);
@@ -178,6 +278,8 @@ private:

    void DeriveGeneralPurposeKeys(std::size_t crypto_revision);

+    RSAKeyPair<2048> GetETicketRSAKey() const;
+
    void SetKeyWrapped(S128KeyType id, Key128 key, u64 field1 = 0, u64 field2 = 0);
    void SetKeyWrapped(S256KeyType id, Key256 key, u64 field1 = 0, u64 field2 = 0);

@@ -195,11 +297,11 @@ std::array<u8, 0x90> DecryptKeyblob(const std::array<u8, 0xB0>& encrypted_keyblo
 std::optional<Key128> DeriveSDSeed();
 Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& keys);

-std::vector<TicketRaw> GetTicketblob(const FileUtil::IOFile& ticket_save);
+std::vector<Ticket> GetTicketblob(const FileUtil::IOFile& ticket_save);

-// Returns a pair of {rights_id, titlekey}. Fails if the ticket has no certificate authority (offset
-// 0x140-0x144 is zero)
-std::optional<std::pair<Key128, Key128>> ParseTicket(const TicketRaw& ticket,
+// Returns a pair of {rights_id, titlekey}. Fails if the ticket has no certificate authority
+// (offset 0x140-0x144 is zero)
+std::optional<std::pair<Key128, Key128>> ParseTicket(const Ticket& ticket,
                                                     const RSAKeyPair<2048>& eticket_extended_key);

 } // namespace Core::Crypto
@@ -0,0 +1,46 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "core/file_sys/system_archive/mii_model.h"
+#include "core/file_sys/vfs_vector.h"
+
+namespace FileSys::SystemArchive {
+
+namespace MiiModelData {
+
+constexpr std::array<u8, 0x10> NFTR_STANDARD{'N',  'F',  'T',  'R',  0x01, 0x00, 0x00, 0x00,
+                                             0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00};
+constexpr std::array<u8, 0x10> NFSR_STANDARD{'N',  'F',  'S',  'R',  0x01, 0x00, 0x00, 0x00,
+                                             0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00};
+
+constexpr auto TEXTURE_LOW_LINEAR = NFTR_STANDARD;
+constexpr auto TEXTURE_LOW_SRGB = NFTR_STANDARD;
+constexpr auto TEXTURE_MID_LINEAR = NFTR_STANDARD;
+constexpr auto TEXTURE_MID_SRGB = NFTR_STANDARD;
+constexpr auto SHAPE_HIGH = NFSR_STANDARD;
+constexpr auto SHAPE_MID = NFSR_STANDARD;
+
+} // namespace MiiModelData
+
+VirtualDir MiiModel() {
+    auto out = std::make_shared<VectorVfsDirectory>(std::vector<VirtualFile>{},
+                                                    std::vector<VirtualDir>{}, "data");
+
+    out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::TEXTURE_LOW_LINEAR.size()>>(
+        MiiModelData::TEXTURE_LOW_LINEAR, "NXTextureLowLinear.dat"));
+    out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::TEXTURE_LOW_SRGB.size()>>(
+        MiiModelData::TEXTURE_LOW_SRGB, "NXTextureLowSRGB.dat"));
+    out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::TEXTURE_MID_LINEAR.size()>>(
+        MiiModelData::TEXTURE_MID_LINEAR, "NXTextureMidLinear.dat"));
+    out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::TEXTURE_MID_SRGB.size()>>(
+        MiiModelData::TEXTURE_MID_SRGB, "NXTextureMidSRGB.dat"));
+    out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::SHAPE_HIGH.size()>>(
+        MiiModelData::SHAPE_HIGH, "ShapeHigh.dat"));
+    out->AddFile(std::make_shared<ArrayVfsFile<MiiModelData::SHAPE_MID.size()>>(
+        MiiModelData::SHAPE_MID, "ShapeMid.dat"));
+
+    return std::move(out);
+}
+
+} // namespace FileSys::SystemArchive
@@ -0,0 +1,13 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "core/file_sys/vfs_types.h"
+
+namespace FileSys::SystemArchive {
+
+VirtualDir MiiModel();
+
+} // namespace FileSys::SystemArchive
@@ -4,6 +4,7 @@

 #include "common/logging/log.h"
 #include "core/file_sys/romfs.h"
+#include "core/file_sys/system_archive/mii_model.h"
 #include "core/file_sys/system_archive/ng_word.h"
 #include "core/file_sys/system_archive/system_archive.h"
 #include "core/file_sys/system_archive/system_version.h"
@@ -24,7 +25,7 @@ struct SystemArchiveDescriptor {
 constexpr std::array<SystemArchiveDescriptor, SYSTEM_ARCHIVE_COUNT> SYSTEM_ARCHIVES{{
    {0x0100000000000800, "CertStore", nullptr},
    {0x0100000000000801, "ErrorMessage", nullptr},
-    {0x0100000000000802, "MiiModel", nullptr},
+    {0x0100000000000802, "MiiModel", &MiiModel},
    {0x0100000000000803, "BrowserDll", nullptr},
    {0x0100000000000804, "Help", nullptr},
    {0x0100000000000805, "SharedFont", nullptr},
@@ -8,6 +8,7 @@
 #include <vector>

 #include "common/common_types.h"
+#include "core/hle/kernel/physical_memory.h"

 namespace Kernel {

@@ -77,7 +78,7 @@ struct CodeSet final {
    }

    /// The overall data that backs this code set.
-    std::vector<u8> memory;
+    Kernel::PhysicalMemory memory;

    /// The segments that comprise this code set.
    std::array<Segment, 3> segments;
@@ -0,0 +1,19 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/alignment.h"
+
+namespace Kernel {
+
+// This encapsulation serves 2 purposes:
+// - First, to encapsulate host physical memory under a single type and set an
+// standard for managing it.
+// - Second to ensure all host backing memory used is aligned to 256 bytes due
+// to strict alignment restrictions on GPU memory.
+
+using PhysicalMemory = std::vector<u8, Common::AlignmentAllocator<u8, 256>>;
+
+} // namespace Kernel
@@ -247,7 +247,7 @@ VAddr Process::CreateTLSRegion() {
        ASSERT(region_address.Succeeded());

        const auto map_result = vm_manager.MapMemoryBlock(
-            *region_address, std::make_shared<std::vector<u8>>(Memory::PAGE_SIZE), 0,
+            *region_address, std::make_shared<PhysicalMemory>(Memory::PAGE_SIZE), 0,
            Memory::PAGE_SIZE, MemoryState::ThreadLocal);
        ASSERT(map_result.Succeeded());

@@ -277,7 +277,7 @@ void Process::FreeTLSRegion(VAddr tls_address) {
 }

 void Process::LoadModule(CodeSet module_, VAddr base_addr) {
-    const auto memory = std::make_shared<std::vector<u8>>(std::move(module_.memory));
+    const auto memory = std::make_shared<PhysicalMemory>(std::move(module_.memory));

    const auto MapSegment = [&](const CodeSet::Segment& segment, VMAPermission permissions,
                                MemoryState memory_state) {
@@ -327,7 +327,7 @@ void Process::AllocateMainThreadStack(u64 stack_size) {
    // Allocate and map the main thread stack
    const VAddr mapping_address = vm_manager.GetTLSIORegionEndAddress() - main_thread_stack_size;
    vm_manager
-        .MapMemoryBlock(mapping_address, std::make_shared<std::vector<u8>>(main_thread_stack_size),
+        .MapMemoryBlock(mapping_address, std::make_shared<PhysicalMemory>(main_thread_stack_size),
                        0, main_thread_stack_size, MemoryState::Stack)
        .Unwrap();
 }
@@ -28,7 +28,7 @@ SharedPtr<SharedMemory> SharedMemory::Create(KernelCore& kernel, Process* owner_
    shared_memory->other_permissions = other_permissions;

    if (address == 0) {
-        shared_memory->backing_block = std::make_shared<std::vector<u8>>(size);
+        shared_memory->backing_block = std::make_shared<Kernel::PhysicalMemory>(size);
        shared_memory->backing_block_offset = 0;

        // Refresh the address mappings for the current process.
@@ -59,8 +59,8 @@ SharedPtr<SharedMemory> SharedMemory::Create(KernelCore& kernel, Process* owner_
 }

 SharedPtr<SharedMemory> SharedMemory::CreateForApplet(
-    KernelCore& kernel, std::shared_ptr<std::vector<u8>> heap_block, std::size_t offset, u64 size,
-    MemoryPermission permissions, MemoryPermission other_permissions, std::string name) {
+    KernelCore& kernel, std::shared_ptr<Kernel::PhysicalMemory> heap_block, std::size_t offset,
+    u64 size, MemoryPermission permissions, MemoryPermission other_permissions, std::string name) {
    SharedPtr<SharedMemory> shared_memory(new SharedMemory(kernel));

    shared_memory->owner_process = nullptr;
@@ -10,6 +10,7 @@

 #include "common/common_types.h"
 #include "core/hle/kernel/object.h"
+#include "core/hle/kernel/physical_memory.h"
 #include "core/hle/kernel/process.h"
 #include "core/hle/result.h"

@@ -62,12 +63,10 @@ public:
     * block.
     * @param name Optional object name, used for debugging purposes.
     */
-    static SharedPtr<SharedMemory> CreateForApplet(KernelCore& kernel,
-                                                   std::shared_ptr<std::vector<u8>> heap_block,
-                                                   std::size_t offset, u64 size,
-                                                   MemoryPermission permissions,
-                                                   MemoryPermission other_permissions,
-                                                   std::string name = "Unknown Applet");
+    static SharedPtr<SharedMemory> CreateForApplet(
+        KernelCore& kernel, std::shared_ptr<Kernel::PhysicalMemory> heap_block, std::size_t offset,
+        u64 size, MemoryPermission permissions, MemoryPermission other_permissions,
+        std::string name = "Unknown Applet");

    std::string GetTypeName() const override {
        return "SharedMemory";
@@ -135,7 +134,7 @@ private:
    ~SharedMemory() override;

    /// Backing memory for this shared memory block.
-    std::shared_ptr<std::vector<u8>> backing_block;
+    std::shared_ptr<PhysicalMemory> backing_block;
    /// Offset into the backing block for this shared memory.
    std::size_t backing_block_offset = 0;
    /// Size of the memory block. Page-aligned.
@@ -47,7 +47,7 @@ ResultCode TransferMemory::MapMemory(VAddr address, u64 size, MemoryPermission p
        return ERR_INVALID_STATE;
    }

-    backing_block = std::make_shared<std::vector<u8>>(size);
+    backing_block = std::make_shared<PhysicalMemory>(size);

    const auto map_state = owner_permissions == MemoryPermission::None
                               ? MemoryState::TransferMemoryIsolated
@@ -8,6 +8,7 @@
 #include <vector>

 #include "core/hle/kernel/object.h"
+#include "core/hle/kernel/physical_memory.h"

 union ResultCode;

@@ -82,7 +83,7 @@ private:
    ~TransferMemory() override;

    /// Memory block backing this instance.
-    std::shared_ptr<std::vector<u8>> backing_block;
+    std::shared_ptr<PhysicalMemory> backing_block;

    /// The base address for the memory managed by this instance.
    VAddr base_address = 0;
@@ -5,6 +5,7 @@
 #include <algorithm>
 #include <iterator>
 #include <utility>
+#include "common/alignment.h"
 #include "common/assert.h"
 #include "common/logging/log.h"
 #include "common/memory_hook.h"
@@ -103,7 +104,7 @@ bool VMManager::IsValidHandle(VMAHandle handle) const {
 }

 ResultVal<VMManager::VMAHandle> VMManager::MapMemoryBlock(VAddr target,
-                                                          std::shared_ptr<std::vector<u8>> block,
+                                                          std::shared_ptr<PhysicalMemory> block,
                                                          std::size_t offset, u64 size,
                                                          MemoryState state, VMAPermission perm) {
    ASSERT(block != nullptr);
@@ -260,7 +261,7 @@ ResultVal<VAddr> VMManager::SetHeapSize(u64 size) {

    if (heap_memory == nullptr) {
        // Initialize heap
-        heap_memory = std::make_shared<std::vector<u8>>(size);
+        heap_memory = std::make_shared<PhysicalMemory>(size);
        heap_end = heap_region_base + size;
    } else {
        UnmapRange(heap_region_base, GetCurrentHeapSize());
@@ -295,12 +296,6 @@ ResultVal<VAddr> VMManager::SetHeapSize(u64 size) {
 }

 ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) {
-    const auto end_addr = target + size;
-    const auto last_addr = end_addr - 1;
-    VAddr cur_addr = target;
-
-    ResultCode result = RESULT_SUCCESS;
-
    // Check how much memory we've already mapped.
    const auto mapped_size_result = SizeOfAllocatedVMAsInRange(target, size);
    if (mapped_size_result.Failed()) {
@@ -323,13 +318,16 @@ ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) {

    // Keep track of the memory regions we unmap.
    std::vector<std::pair<u64, u64>> mapped_regions;
+    ResultCode result = RESULT_SUCCESS;

    // Iterate, trying to map memory.
    {
-        cur_addr = target;
+        const auto end_addr = target + size;
+        const auto last_addr = end_addr - 1;
+        VAddr cur_addr = target;

        auto iter = FindVMA(target);
-        ASSERT_MSG(iter != vma_map.end(), "MapPhysicalMemory iter != end");
+        ASSERT(iter != vma_map.end());

        while (true) {
            const auto& vma = iter->second;
@@ -341,7 +339,7 @@ ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) {
            const auto map_size = std::min(end_addr - cur_addr, vma_end - cur_addr);
            if (vma.state == MemoryState::Unmapped) {
                const auto map_res =
-                    MapMemoryBlock(cur_addr, std::make_shared<std::vector<u8>>(map_size, 0), 0,
+                    MapMemoryBlock(cur_addr, std::make_shared<PhysicalMemory>(map_size), 0,
                                   map_size, MemoryState::Heap, VMAPermission::ReadWrite);
                result = map_res.Code();
                if (result.IsError()) {
@@ -359,7 +357,7 @@ ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) {
            // Advance to the next block.
            cur_addr = vma_end;
            iter = FindVMA(cur_addr);
-            ASSERT_MSG(iter != vma_map.end(), "MapPhysicalMemory iter != end");
+            ASSERT(iter != vma_map.end());
        }
    }

@@ -367,7 +365,7 @@ ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) {
    if (result.IsError()) {
        for (const auto [unmap_address, unmap_size] : mapped_regions) {
            ASSERT_MSG(UnmapRange(unmap_address, unmap_size).IsSuccess(),
-                       "MapPhysicalMemory un-map on error");
+                       "Failed to unmap memory range.");
        }

        return result;
@@ -380,12 +378,6 @@ ResultCode VMManager::MapPhysicalMemory(VAddr target, u64 size) {
 }

 ResultCode VMManager::UnmapPhysicalMemory(VAddr target, u64 size) {
-    const auto end_addr = target + size;
-    const auto last_addr = end_addr - 1;
-    VAddr cur_addr = target;
-
-    ResultCode result = RESULT_SUCCESS;
-
    // Check how much memory is currently mapped.
    const auto mapped_size_result = SizeOfUnmappablePhysicalMemoryInRange(target, size);
    if (mapped_size_result.Failed()) {
@@ -400,13 +392,16 @@ ResultCode VMManager::UnmapPhysicalMemory(VAddr target, u64 size) {

    // Keep track of the memory regions we unmap.
    std::vector<std::pair<u64, u64>> unmapped_regions;
+    ResultCode result = RESULT_SUCCESS;

    // Try to unmap regions.
    {
-        cur_addr = target;
+        const auto end_addr = target + size;
+        const auto last_addr = end_addr - 1;
+        VAddr cur_addr = target;

        auto iter = FindVMA(target);
-        ASSERT_MSG(iter != vma_map.end(), "UnmapPhysicalMemory iter != end");
+        ASSERT(iter != vma_map.end());

        while (true) {
            const auto& vma = iter->second;
@@ -433,7 +428,7 @@ ResultCode VMManager::UnmapPhysicalMemory(VAddr target, u64 size) {
            // Advance to the next block.
            cur_addr = vma_end;
            iter = FindVMA(cur_addr);
-            ASSERT_MSG(iter != vma_map.end(), "UnmapPhysicalMemory iter != end");
+            ASSERT(iter != vma_map.end());
        }
    }

@@ -442,10 +437,12 @@ ResultCode VMManager::UnmapPhysicalMemory(VAddr target, u64 size) {
    if (result.IsError()) {
        for (const auto [map_address, map_size] : unmapped_regions) {
            const auto remap_res =
-                MapMemoryBlock(map_address, std::make_shared<std::vector<u8>>(map_size, 0), 0,
-                               map_size, MemoryState::Heap, VMAPermission::None);
-            ASSERT_MSG(remap_res.Succeeded(), "UnmapPhysicalMemory re-map on error");
+                MapMemoryBlock(map_address, std::make_shared<PhysicalMemory>(map_size), 0, map_size,
+                               MemoryState::Heap, VMAPermission::None);
+            ASSERT_MSG(remap_res.Succeeded(), "Failed to remap a memory block.");
        }
+
+        return result;
    }

    // Update mapped amount
@@ -593,7 +590,7 @@ ResultCode VMManager::MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size, Mem
    ASSERT_MSG(vma_offset + size <= vma->second.size,
               "Shared memory exceeds bounds of mapped block");

-    const std::shared_ptr<std::vector<u8>>& backing_block = vma->second.backing_block;
+    const std::shared_ptr<PhysicalMemory>& backing_block = vma->second.backing_block;
    const std::size_t backing_block_offset = vma->second.offset + vma_offset;

    CASCADE_RESULT(auto new_vma,
@@ -606,7 +603,7 @@ ResultCode VMManager::MirrorMemory(VAddr dst_addr, VAddr src_addr, u64 size, Mem
    return RESULT_SUCCESS;
 }

-void VMManager::RefreshMemoryBlockMappings(const std::vector<u8>* block) {
+void VMManager::RefreshMemoryBlockMappings(const PhysicalMemory* block) {
    // If this ever proves to have a noticeable performance impact, allow users of the function to
    // specify a specific range of addresses to limit the scan to.
    for (const auto& p : vma_map) {
@@ -756,20 +753,26 @@ void VMManager::MergeAdjacentVMA(VirtualMemoryArea& left, const VirtualMemoryAre
    // Always merge allocated memory blocks, even when they don't share the same backing block.
    if (left.type == VMAType::AllocatedMemoryBlock &&
        (left.backing_block != right.backing_block || left.offset + left.size != right.offset)) {
+        const auto right_begin = right.backing_block->begin() + right.offset;
+        const auto right_end = right_begin + right.size;
+
        // Check if we can save work.
        if (left.offset == 0 && left.size == left.backing_block->size()) {
            // Fast case: left is an entire backing block.
-            left.backing_block->insert(left.backing_block->end(),
-                                       right.backing_block->begin() + right.offset,
-                                       right.backing_block->begin() + right.offset + right.size);
+            left.backing_block->insert(left.backing_block->end(), right_begin, right_end);
        } else {
            // Slow case: make a new memory block for left and right.
-            auto new_memory = std::make_shared<std::vector<u8>>();
-            new_memory->insert(new_memory->end(), left.backing_block->begin() + left.offset,
-                               left.backing_block->begin() + left.offset + left.size);
-            new_memory->insert(new_memory->end(), right.backing_block->begin() + right.offset,
-                               right.backing_block->begin() + right.offset + right.size);
-            left.backing_block = new_memory;
+            const auto left_begin = left.backing_block->begin() + left.offset;
+            const auto left_end = left_begin + left.size;
+            const auto left_size = static_cast<std::size_t>(std::distance(left_begin, left_end));
+            const auto right_size = static_cast<std::size_t>(std::distance(right_begin, right_end));
+
+            auto new_memory = std::make_shared<PhysicalMemory>();
+            new_memory->reserve(left_size + right_size);
+            new_memory->insert(new_memory->end(), left_begin, left_end);
+            new_memory->insert(new_memory->end(), right_begin, right_end);
+
+            left.backing_block = std::move(new_memory);
            left.offset = 0;
        }

@@ -964,7 +967,7 @@ ResultVal<std::size_t> VMManager::SizeOfAllocatedVMAsInRange(VAddr address,

    VAddr cur_addr = address;
    auto iter = FindVMA(cur_addr);
-    ASSERT_MSG(iter != vma_map.end(), "SizeOfAllocatedVMAsInRange iter != end");
+    ASSERT(iter != vma_map.end());

    while (true) {
        const auto& vma = iter->second;
@@ -985,7 +988,7 @@ ResultVal<std::size_t> VMManager::SizeOfAllocatedVMAsInRange(VAddr address,
        // Advance to the next block.
        cur_addr = vma_end;
        iter = std::next(iter);
-        ASSERT_MSG(iter != vma_map.end(), "SizeOfAllocatedVMAsInRange iter != end");
+        ASSERT(iter != vma_map.end());
    }

    return MakeResult(mapped_size);
@@ -999,7 +1002,7 @@ ResultVal<std::size_t> VMManager::SizeOfUnmappablePhysicalMemoryInRange(VAddr ad

    VAddr cur_addr = address;
    auto iter = FindVMA(cur_addr);
-    ASSERT_MSG(iter != vma_map.end(), "SizeOfUnmappablePhysicalMemoryInRange iter != end");
+    ASSERT(iter != vma_map.end());

    while (true) {
        const auto& vma = iter->second;
@@ -1028,7 +1031,7 @@ ResultVal<std::size_t> VMManager::SizeOfUnmappablePhysicalMemoryInRange(VAddr ad
        // Advance to the next block.
        cur_addr = vma_end;
        iter = std::next(iter);
-        ASSERT_MSG(iter != vma_map.end(), "SizeOfUnmappablePhysicalMemoryInRange iter != end");
+        ASSERT(iter != vma_map.end());
    }

    return MakeResult(mapped_size);
@@ -11,6 +11,7 @@
 #include "common/common_types.h"
 #include "common/memory_hook.h"
 #include "common/page_table.h"
+#include "core/hle/kernel/physical_memory.h"
 #include "core/hle/result.h"
 #include "core/memory.h"

@@ -290,7 +291,7 @@ struct VirtualMemoryArea {

    // Settings for type = AllocatedMemoryBlock
    /// Memory block backing this VMA.
-    std::shared_ptr<std::vector<u8>> backing_block = nullptr;
+    std::shared_ptr<PhysicalMemory> backing_block = nullptr;
    /// Offset into the backing_memory the mapping starts from.
    std::size_t offset = 0;

@@ -348,7 +349,7 @@ public:
     * @param size Size of the mapping.
     * @param state MemoryState tag to attach to the VMA.
     */
-    ResultVal<VMAHandle> MapMemoryBlock(VAddr target, std::shared_ptr<std::vector<u8>> block,
+    ResultVal<VMAHandle> MapMemoryBlock(VAddr target, std::shared_ptr<PhysicalMemory> block,
                                        std::size_t offset, u64 size, MemoryState state,
                                        VMAPermission perm = VMAPermission::ReadWrite);

@@ -453,8 +454,8 @@ public:

    /// Maps memory at a given address.
    ///
-    /// @param addr The virtual address to map memory at.
-    /// @param size The amount of memory to map.
+    /// @param target The virtual address to map memory at.
+    /// @param size   The amount of memory to map.
    ///
    /// @note The destination address must lie within the Map region.
    ///
@@ -467,8 +468,8 @@ public:

    /// Unmaps memory at a given address.
    ///
-    /// @param addr The virtual address to unmap memory at.
-    /// @param size The amount of memory to unmap.
+    /// @param target The virtual address to unmap memory at.
+    /// @param size   The amount of memory to unmap.
    ///
    /// @note The destination address must lie within the Map region.
    ///
@@ -547,7 +548,7 @@ public:
     * Scans all VMAs and updates the page table range of any that use the given vector as backing
     * memory. This should be called after any operation that causes reallocation of the vector.
     */
-    void RefreshMemoryBlockMappings(const std::vector<u8>* block);
+    void RefreshMemoryBlockMappings(const PhysicalMemory* block);

    /// Dumps the address space layout to the log, for debugging
    void LogLayout() const;
@@ -777,7 +778,7 @@ private:
    // the entire virtual address space extents that bound the allocations, including any holes.
    // This makes deallocation and reallocation of holes fast and keeps process memory contiguous
    // in the emulator address space, allowing Memory::GetPointer to be reasonably safe.
-    std::shared_ptr<std::vector<u8>> heap_memory;
+    std::shared_ptr<PhysicalMemory> heap_memory;

    // The end of the currently allocated heap. This is not an inclusive
    // end of the range. This is essentially 'base_address + current_size'.
@@ -1057,6 +1057,7 @@ IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationF
        {120, nullptr, "ExecuteProgram"},
        {121, nullptr, "ClearUserChannel"},
        {122, nullptr, "UnpopToUserChannel"},
+        {130, &IApplicationFunctions::GetGpuErrorDetectedSystemEvent, "GetGpuErrorDetectedSystemEvent"},
        {500, nullptr, "StartContinuousRecordingFlushForDebug"},
        {1000, nullptr, "CreateMovieMaker"},
        {1001, nullptr, "PrepareForJit"},
@@ -1064,6 +1065,10 @@ IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationF
    // clang-format on

    RegisterHandlers(functions);
+
+    auto& kernel = Core::System::GetInstance().Kernel();
+    gpu_error_detected_event = Kernel::WritableEvent::CreateEventPair(
+        kernel, Kernel::ResetType::Manual, "IApplicationFunctions:GpuErrorDetectedSystemEvent");
 }

 IApplicationFunctions::~IApplicationFunctions() = default;
@@ -1285,6 +1290,14 @@ void IApplicationFunctions::GetSaveDataSize(Kernel::HLERequestContext& ctx) {
    rb.Push(size.journal);
 }

+void IApplicationFunctions::GetGpuErrorDetectedSystemEvent(Kernel::HLERequestContext& ctx) {
+    LOG_WARNING(Service_AM, "(STUBBED) called");
+
+    IPC::ResponseBuilder rb{ctx, 2, 1};
+    rb.Push(RESULT_SUCCESS);
+    rb.PushCopyObjects(gpu_error_detected_event.readable);
+}
+
 void InstallInterfaces(SM::ServiceManager& service_manager,
                       std::shared_ptr<NVFlinger::NVFlinger> nvflinger, Core::System& system) {
    auto message_queue = std::make_shared<AppletMessageQueue>();
@@ -242,6 +242,9 @@ private:
    void BeginBlockingHomeButton(Kernel::HLERequestContext& ctx);
    void EndBlockingHomeButton(Kernel::HLERequestContext& ctx);
    void EnableApplicationCrashReport(Kernel::HLERequestContext& ctx);
+    void GetGpuErrorDetectedSystemEvent(Kernel::HLERequestContext& ctx);
+
+    Kernel::EventPair gpu_error_detected_event;
 };

 class IHomeMenuFunctions final : public ServiceFramework<IHomeMenuFunctions> {
@@ -19,16 +19,16 @@

 namespace Service::Audio {

-void InstallInterfaces(SM::ServiceManager& service_manager) {
+void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) {
    std::make_shared<AudCtl>()->InstallAsService(service_manager);
    std::make_shared<AudOutA>()->InstallAsService(service_manager);
-    std::make_shared<AudOutU>()->InstallAsService(service_manager);
+    std::make_shared<AudOutU>(system)->InstallAsService(service_manager);
    std::make_shared<AudInA>()->InstallAsService(service_manager);
    std::make_shared<AudInU>()->InstallAsService(service_manager);
    std::make_shared<AudRecA>()->InstallAsService(service_manager);
    std::make_shared<AudRecU>()->InstallAsService(service_manager);
    std::make_shared<AudRenA>()->InstallAsService(service_manager);
-    std::make_shared<AudRenU>()->InstallAsService(service_manager);
+    std::make_shared<AudRenU>(system)->InstallAsService(service_manager);
    std::make_shared<CodecCtl>()->InstallAsService(service_manager);
    std::make_shared<HwOpus>()->InstallAsService(service_manager);

@@ -4,6 +4,10 @@

 #pragma once

+namespace Core {
+class System;
+}
+
 namespace Service::SM {
 class ServiceManager;
 }
@@ -11,6 +15,6 @@ class ServiceManager;
 namespace Service::Audio {

 /// Registers all Audio services with the specified service manager.
-void InstallInterfaces(SM::ServiceManager& service_manager);
+void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system);

 } // namespace Service::Audio
@@ -40,8 +40,8 @@ enum class AudioState : u32 {

 class IAudioOut final : public ServiceFramework<IAudioOut> {
 public:
-    IAudioOut(AudoutParams audio_params, AudioCore::AudioOut& audio_core, std::string&& device_name,
-              std::string&& unique_name)
+    IAudioOut(Core::System& system, AudoutParams audio_params, AudioCore::AudioOut& audio_core,
+              std::string&& device_name, std::string&& unique_name)
        : ServiceFramework("IAudioOut"), audio_core(audio_core),
          device_name(std::move(device_name)), audio_params(audio_params) {
        // clang-format off
@@ -65,7 +65,6 @@ public:
        RegisterHandlers(functions);

        // This is the event handle used to check if the audio buffer was released
-        auto& system = Core::System::GetInstance();
        buffer_event = Kernel::WritableEvent::CreateEventPair(
            system.Kernel(), Kernel::ResetType::Manual, "IAudioOutBufferReleased");

@@ -212,6 +211,22 @@ private:
    Kernel::EventPair buffer_event;
 };

+AudOutU::AudOutU(Core::System& system_) : ServiceFramework("audout:u"), system{system_} {
+    // clang-format off
+    static const FunctionInfo functions[] = {
+        {0, &AudOutU::ListAudioOutsImpl, "ListAudioOuts"},
+        {1, &AudOutU::OpenAudioOutImpl, "OpenAudioOut"},
+        {2, &AudOutU::ListAudioOutsImpl, "ListAudioOutsAuto"},
+        {3, &AudOutU::OpenAudioOutImpl, "OpenAudioOutAuto"},
+    };
+    // clang-format on
+
+    RegisterHandlers(functions);
+    audio_core = std::make_unique<AudioCore::AudioOut>();
+}
+
+AudOutU::~AudOutU() = default;
+
 void AudOutU::ListAudioOutsImpl(Kernel::HLERequestContext& ctx) {
    LOG_DEBUG(Service_Audio, "called");

@@ -248,7 +263,7 @@ void AudOutU::OpenAudioOutImpl(Kernel::HLERequestContext& ctx) {

    std::string unique_name{fmt::format("{}-{}", device_name, audio_out_interfaces.size())};
    auto audio_out_interface = std::make_shared<IAudioOut>(
-        params, *audio_core, std::move(device_name), std::move(unique_name));
+        system, params, *audio_core, std::move(device_name), std::move(unique_name));

    IPC::ResponseBuilder rb{ctx, 6, 0, 1};
    rb.Push(RESULT_SUCCESS);
@@ -256,20 +271,9 @@ void AudOutU::OpenAudioOutImpl(Kernel::HLERequestContext& ctx) {
    rb.Push<u32>(params.channel_count);
    rb.Push<u32>(static_cast<u32>(AudioCore::Codec::PcmFormat::Int16));
    rb.Push<u32>(static_cast<u32>(AudioState::Stopped));
-    rb.PushIpcInterface<Audio::IAudioOut>(audio_out_interface);
+    rb.PushIpcInterface<IAudioOut>(audio_out_interface);

    audio_out_interfaces.push_back(std::move(audio_out_interface));
 }

-AudOutU::AudOutU() : ServiceFramework("audout:u") {
-    static const FunctionInfo functions[] = {{0, &AudOutU::ListAudioOutsImpl, "ListAudioOuts"},
-                                             {1, &AudOutU::OpenAudioOutImpl, "OpenAudioOut"},
-                                             {2, &AudOutU::ListAudioOutsImpl, "ListAudioOutsAuto"},
-                                             {3, &AudOutU::OpenAudioOutImpl, "OpenAudioOutAuto"}};
-    RegisterHandlers(functions);
-    audio_core = std::make_unique<AudioCore::AudioOut>();
-}
-
-AudOutU::~AudOutU() = default;
-
 } // namespace Service::Audio
@@ -11,6 +11,10 @@ namespace AudioCore {
 class AudioOut;
 }

+namespace Core {
+class System;
+}
+
 namespace Kernel {
 class HLERequestContext;
 }
@@ -21,15 +25,17 @@ class IAudioOut;

 class AudOutU final : public ServiceFramework<AudOutU> {
 public:
-    AudOutU();
+    explicit AudOutU(Core::System& system_);
    ~AudOutU() override;

 private:
+    void ListAudioOutsImpl(Kernel::HLERequestContext& ctx);
+    void OpenAudioOutImpl(Kernel::HLERequestContext& ctx);
+
    std::vector<std::shared_ptr<IAudioOut>> audio_out_interfaces;
    std::unique_ptr<AudioCore::AudioOut> audio_core;

-    void ListAudioOutsImpl(Kernel::HLERequestContext& ctx);
-    void OpenAudioOutImpl(Kernel::HLERequestContext& ctx);
+    Core::System& system;
 };

 } // namespace Service::Audio
@@ -5,6 +5,7 @@
 #include <algorithm>
 #include <array>
 #include <memory>
+#include <string_view>

 #include "audio_core/audio_renderer.h"
 #include "common/alignment.h"
@@ -25,7 +26,7 @@ namespace Service::Audio {

 class IAudioRenderer final : public ServiceFramework<IAudioRenderer> {
 public:
-    explicit IAudioRenderer(AudioCore::AudioRendererParameter audren_params,
+    explicit IAudioRenderer(Core::System& system, AudioCore::AudioRendererParameter audren_params,
                            const std::size_t instance_number)
        : ServiceFramework("IAudioRenderer") {
        // clang-format off
@@ -46,7 +47,6 @@ public:
        // clang-format on
        RegisterHandlers(functions);

-        auto& system = Core::System::GetInstance();
        system_event = Kernel::WritableEvent::CreateEventPair(
            system.Kernel(), Kernel::ResetType::Manual, "IAudioRenderer:SystemEvent");
        renderer = std::make_unique<AudioCore::AudioRenderer>(
@@ -160,28 +160,33 @@ private:

 class IAudioDevice final : public ServiceFramework<IAudioDevice> {
 public:
-    IAudioDevice() : ServiceFramework("IAudioDevice") {
+    explicit IAudioDevice(Core::System& system, u32_le revision_num)
+        : ServiceFramework("IAudioDevice"), revision{revision_num} {
        static const FunctionInfo functions[] = {
            {0, &IAudioDevice::ListAudioDeviceName, "ListAudioDeviceName"},
            {1, &IAudioDevice::SetAudioDeviceOutputVolume, "SetAudioDeviceOutputVolume"},
-            {2, nullptr, "GetAudioDeviceOutputVolume"},
+            {2, &IAudioDevice::GetAudioDeviceOutputVolume, "GetAudioDeviceOutputVolume"},
            {3, &IAudioDevice::GetActiveAudioDeviceName, "GetActiveAudioDeviceName"},
            {4, &IAudioDevice::QueryAudioDeviceSystemEvent, "QueryAudioDeviceSystemEvent"},
            {5, &IAudioDevice::GetActiveChannelCount, "GetActiveChannelCount"},
            {6, &IAudioDevice::ListAudioDeviceName, "ListAudioDeviceNameAuto"},
            {7, &IAudioDevice::SetAudioDeviceOutputVolume, "SetAudioDeviceOutputVolumeAuto"},
-            {8, nullptr, "GetAudioDeviceOutputVolumeAuto"},
+            {8, &IAudioDevice::GetAudioDeviceOutputVolume, "GetAudioDeviceOutputVolumeAuto"},
            {10, &IAudioDevice::GetActiveAudioDeviceName, "GetActiveAudioDeviceNameAuto"},
-            {11, nullptr, "QueryAudioDeviceInputEvent"},
+            {11, &IAudioDevice::QueryAudioDeviceInputEvent, "QueryAudioDeviceInputEvent"},
            {12, &IAudioDevice::QueryAudioDeviceOutputEvent, "QueryAudioDeviceOutputEvent"},
            {13, nullptr, "GetAudioSystemMasterVolumeSetting"},
        };
        RegisterHandlers(functions);

-        auto& kernel = Core::System::GetInstance().Kernel();
+        auto& kernel = system.Kernel();
        buffer_event = Kernel::WritableEvent::CreateEventPair(kernel, Kernel::ResetType::Automatic,
                                                              "IAudioOutBufferReleasedEvent");

+        // Should be similar to audio_output_device_switch_event
+        audio_input_device_switch_event = Kernel::WritableEvent::CreateEventPair(
+            kernel, Kernel::ResetType::Automatic, "IAudioDevice:AudioInputDeviceSwitchedEvent");
+
        // Should only be signalled when an audio output device has been changed, example: speaker
        // to headset
        audio_output_device_switch_event = Kernel::WritableEvent::CreateEventPair(
@@ -189,15 +194,47 @@ public:
    }

 private:
-    void ListAudioDeviceName(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_Audio, "(STUBBED) called");
+    using AudioDeviceName = std::array<char, 256>;
+    static constexpr std::array<std::string_view, 4> audio_device_names{{
+        "AudioStereoJackOutput",
+        "AudioBuiltInSpeakerOutput",
+        "AudioTvOutput",
+        "AudioUsbDeviceOutput",
+    }};
+    enum class DeviceType {
+        AHUBHeadphones,
+        AHUBSpeakers,
+        HDA,
+        USBOutput,
+    };

-        constexpr std::array<char, 15> audio_interface{{"AudioInterface"}};
-        ctx.WriteBuffer(audio_interface);
+    void ListAudioDeviceName(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_Audio, "called");
+
+        const bool usb_output_supported =
+            IsFeatureSupported(AudioFeatures::AudioUSBDeviceOutput, revision);
+        const std::size_t count = ctx.GetWriteBufferSize() / sizeof(AudioDeviceName);
+
+        std::vector<AudioDeviceName> name_buffer;
+        name_buffer.reserve(audio_device_names.size());
+
+        for (std::size_t i = 0; i < count && i < audio_device_names.size(); i++) {
+            const auto type = static_cast<DeviceType>(i);
+
+            if (!usb_output_supported && type == DeviceType::USBOutput) {
+                continue;
+            }
+
+            const auto& device_name = audio_device_names[i];
+            auto& entry = name_buffer.emplace_back();
+            device_name.copy(entry.data(), device_name.size());
+        }
+
+        ctx.WriteBuffer(name_buffer);

        IPC::ResponseBuilder rb{ctx, 3};
        rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(1);
+        rb.Push(static_cast<u32>(name_buffer.size()));
    }

    void SetAudioDeviceOutputVolume(Kernel::HLERequestContext& ctx) {
@@ -213,15 +250,32 @@ private:
        rb.Push(RESULT_SUCCESS);
    }

-    void GetActiveAudioDeviceName(Kernel::HLERequestContext& ctx) {
-        LOG_WARNING(Service_Audio, "(STUBBED) called");
+    void GetAudioDeviceOutputVolume(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};

-        constexpr std::array<char, 12> audio_interface{{"AudioDevice"}};
-        ctx.WriteBuffer(audio_interface);
+        const auto device_name_buffer = ctx.ReadBuffer();
+        const std::string name = Common::StringFromBuffer(device_name_buffer);
+
+        LOG_WARNING(Service_Audio, "(STUBBED) called. name={}", name);

        IPC::ResponseBuilder rb{ctx, 3};
        rb.Push(RESULT_SUCCESS);
-        rb.Push<u32>(1);
+        rb.Push(1.0f);
+    }
+
+    void GetActiveAudioDeviceName(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_Audio, "(STUBBED) called");
+
+        // Currently set to always be TV audio output.
+        const auto& device_name = audio_device_names[2];
+
+        AudioDeviceName out_device_name{};
+        device_name.copy(out_device_name.data(), device_name.size());
+
+        ctx.WriteBuffer(out_device_name);
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
    }

    void QueryAudioDeviceSystemEvent(Kernel::HLERequestContext& ctx) {
@@ -242,6 +296,15 @@ private:
        rb.Push<u32>(1);
    }

+    // Should be similar to QueryAudioDeviceOutputEvent
+    void QueryAudioDeviceInputEvent(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_Audio, "(STUBBED) called");
+
+        IPC::ResponseBuilder rb{ctx, 2, 1};
+        rb.Push(RESULT_SUCCESS);
+        rb.PushCopyObjects(audio_input_device_switch_event.readable);
+    }
+
    void QueryAudioDeviceOutputEvent(Kernel::HLERequestContext& ctx) {
        LOG_DEBUG(Service_Audio, "called");

@@ -250,12 +313,14 @@ private:
        rb.PushCopyObjects(audio_output_device_switch_event.readable);
    }

+    u32_le revision = 0;
    Kernel::EventPair buffer_event;
+    Kernel::EventPair audio_input_device_switch_event;
    Kernel::EventPair audio_output_device_switch_event;

 }; // namespace Audio

-AudRenU::AudRenU() : ServiceFramework("audren:u") {
+AudRenU::AudRenU(Core::System& system_) : ServiceFramework("audren:u"), system{system_} {
    // clang-format off
    static const FunctionInfo functions[] = {
        {0, &AudRenU::OpenAudioRenderer, "OpenAudioRenderer"},
@@ -328,7 +393,7 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) {
    };

    // Calculates the portion of the size related to the mix data (and the sorting thereof).
-    const auto calculate_mix_info_size = [this](const AudioCore::AudioRendererParameter& params) {
+    const auto calculate_mix_info_size = [](const AudioCore::AudioRendererParameter& params) {
        // The size of the mixing info data structure.
        constexpr u64 mix_info_size = 0x940;

@@ -400,7 +465,7 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) {

    // Calculates the part of the size related to the splitter context.
    const auto calculate_splitter_context_size =
-        [this](const AudioCore::AudioRendererParameter& params) -> u64 {
+        [](const AudioCore::AudioRendererParameter& params) -> u64 {
        if (!IsFeatureSupported(AudioFeatures::Splitter, params.revision)) {
            return 0;
        }
@@ -447,7 +512,7 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) {
    };

    // Calculates the part of the size related to performance statistics.
-    const auto calculate_perf_size = [this](const AudioCore::AudioRendererParameter& params) {
+    const auto calculate_perf_size = [](const AudioCore::AudioRendererParameter& params) {
        // Extra size value appended to the end of the calculation.
        constexpr u64 appended = 128;

@@ -474,78 +539,76 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) {
    };

    // Calculates the part of the size that relates to the audio command buffer.
-    const auto calculate_command_buffer_size =
-        [this](const AudioCore::AudioRendererParameter& params) {
-            constexpr u64 alignment = (buffer_alignment_size - 1) * 2;
+    const auto calculate_command_buffer_size = [](const AudioCore::AudioRendererParameter& params) {
+        constexpr u64 alignment = (buffer_alignment_size - 1) * 2;

-            if (!IsFeatureSupported(AudioFeatures::VariadicCommandBuffer, params.revision)) {
-                constexpr u64 command_buffer_size = 0x18000;
+        if (!IsFeatureSupported(AudioFeatures::VariadicCommandBuffer, params.revision)) {
+            constexpr u64 command_buffer_size = 0x18000;

-                return command_buffer_size + alignment;
-            }
+            return command_buffer_size + alignment;
+        }

-            // When the variadic command buffer is supported, this means
-            // the command generator for the audio renderer can issue commands
-            // that are (as one would expect), variable in size. So what we need to do
-            // is determine the maximum possible size for a few command data structures
-            // then multiply them by the amount of present commands indicated by the given
-            // respective audio parameters.
+        // When the variadic command buffer is supported, this means
+        // the command generator for the audio renderer can issue commands
+        // that are (as one would expect), variable in size. So what we need to do
+        // is determine the maximum possible size for a few command data structures
+        // then multiply them by the amount of present commands indicated by the given
+        // respective audio parameters.

-            constexpr u64 max_biquad_filters = 2;
-            constexpr u64 max_mix_buffers = 24;
+        constexpr u64 max_biquad_filters = 2;
+        constexpr u64 max_mix_buffers = 24;

-            constexpr u64 biquad_filter_command_size = 0x2C;
+        constexpr u64 biquad_filter_command_size = 0x2C;

-            constexpr u64 depop_mix_command_size = 0x24;
-            constexpr u64 depop_setup_command_size = 0x50;
+        constexpr u64 depop_mix_command_size = 0x24;
+        constexpr u64 depop_setup_command_size = 0x50;

-            constexpr u64 effect_command_max_size = 0x540;
+        constexpr u64 effect_command_max_size = 0x540;

-            constexpr u64 mix_command_size = 0x1C;
-            constexpr u64 mix_ramp_command_size = 0x24;
-            constexpr u64 mix_ramp_grouped_command_size = 0x13C;
+        constexpr u64 mix_command_size = 0x1C;
+        constexpr u64 mix_ramp_command_size = 0x24;
+        constexpr u64 mix_ramp_grouped_command_size = 0x13C;

-            constexpr u64 perf_command_size = 0x28;
+        constexpr u64 perf_command_size = 0x28;

-            constexpr u64 sink_command_size = 0x130;
+        constexpr u64 sink_command_size = 0x130;

-            constexpr u64 submix_command_max_size =
-                depop_mix_command_size + (mix_command_size * max_mix_buffers) * max_mix_buffers;
+        constexpr u64 submix_command_max_size =
+            depop_mix_command_size + (mix_command_size * max_mix_buffers) * max_mix_buffers;

-            constexpr u64 volume_command_size = 0x1C;
-            constexpr u64 volume_ramp_command_size = 0x20;
+        constexpr u64 volume_command_size = 0x1C;
+        constexpr u64 volume_ramp_command_size = 0x20;

-            constexpr u64 voice_biquad_filter_command_size =
-                biquad_filter_command_size * max_biquad_filters;
-            constexpr u64 voice_data_command_size = 0x9C;
-            const u64 voice_command_max_size =
-                (params.splitter_count * depop_setup_command_size) +
-                (voice_data_command_size + voice_biquad_filter_command_size +
-                 volume_ramp_command_size + mix_ramp_grouped_command_size);
+        constexpr u64 voice_biquad_filter_command_size =
+            biquad_filter_command_size * max_biquad_filters;
+        constexpr u64 voice_data_command_size = 0x9C;
+        const u64 voice_command_max_size =
+            (params.splitter_count * depop_setup_command_size) +
+            (voice_data_command_size + voice_biquad_filter_command_size + volume_ramp_command_size +
+             mix_ramp_grouped_command_size);

-            // Now calculate the individual elements that comprise the size and add them together.
-            const u64 effect_commands_size = params.effect_count * effect_command_max_size;
+        // Now calculate the individual elements that comprise the size and add them together.
+        const u64 effect_commands_size = params.effect_count * effect_command_max_size;

-            const u64 final_mix_commands_size =
-                depop_mix_command_size + volume_command_size * max_mix_buffers;
+        const u64 final_mix_commands_size =
+            depop_mix_command_size + volume_command_size * max_mix_buffers;

-            const u64 perf_commands_size =
-                perf_command_size *
-                (CalculateNumPerformanceEntries(params) + max_perf_detail_entries);
+        const u64 perf_commands_size =
+            perf_command_size * (CalculateNumPerformanceEntries(params) + max_perf_detail_entries);

-            const u64 sink_commands_size = params.sink_count * sink_command_size;
+        const u64 sink_commands_size = params.sink_count * sink_command_size;

-            const u64 splitter_commands_size =
-                params.num_splitter_send_channels * max_mix_buffers * mix_ramp_command_size;
+        const u64 splitter_commands_size =
+            params.num_splitter_send_channels * max_mix_buffers * mix_ramp_command_size;

-            const u64 submix_commands_size = params.submix_count * submix_command_max_size;
+        const u64 submix_commands_size = params.submix_count * submix_command_max_size;

-            const u64 voice_commands_size = params.voice_count * voice_command_max_size;
+        const u64 voice_commands_size = params.voice_count * voice_command_max_size;

-            return effect_commands_size + final_mix_commands_size + perf_commands_size +
-                   sink_commands_size + splitter_commands_size + submix_commands_size +
-                   voice_commands_size + alignment;
-        };
+        return effect_commands_size + final_mix_commands_size + perf_commands_size +
+               sink_commands_size + splitter_commands_size + submix_commands_size +
+               voice_commands_size + alignment;
+    };

    IPC::RequestParser rp{ctx};
    const auto params = rp.PopRaw<AudioCore::AudioRendererParameter>();
@@ -578,12 +641,16 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) {
 }

 void AudRenU::GetAudioDeviceService(Kernel::HLERequestContext& ctx) {
-    LOG_DEBUG(Service_Audio, "called");
+    IPC::RequestParser rp{ctx};
+    const u64 aruid = rp.Pop<u64>();

+    LOG_DEBUG(Service_Audio, "called. aruid={:016X}", aruid);
+
+    // Revisionless variant of GetAudioDeviceServiceWithRevisionInfo that
+    // always assumes the initial release revision (REV1).
    IPC::ResponseBuilder rb{ctx, 2, 0, 1};
-
    rb.Push(RESULT_SUCCESS);
-    rb.PushIpcInterface<Audio::IAudioDevice>();
+    rb.PushIpcInterface<IAudioDevice>(system, Common::MakeMagic('R', 'E', 'V', '1'));
 }

 void AudRenU::OpenAudioRendererAuto(Kernel::HLERequestContext& ctx) {
@@ -593,13 +660,19 @@ void AudRenU::OpenAudioRendererAuto(Kernel::HLERequestContext& ctx) {
 }

 void AudRenU::GetAudioDeviceServiceWithRevisionInfo(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_Audio, "(STUBBED) called");
+    struct Parameters {
+        u32 revision;
+        u64 aruid;
+    };
+
+    IPC::RequestParser rp{ctx};
+    const auto [revision, aruid] = rp.PopRaw<Parameters>();
+
+    LOG_DEBUG(Service_Audio, "called. revision={:08X}, aruid={:016X}", revision, aruid);

    IPC::ResponseBuilder rb{ctx, 2, 0, 1};
-
    rb.Push(RESULT_SUCCESS);
-    rb.PushIpcInterface<Audio::IAudioDevice>(); // TODO(ogniK): Figure out what is different
-                                                // based on the current revision
+    rb.PushIpcInterface<IAudioDevice>(system, revision);
 }

 void AudRenU::OpenAudioRendererImpl(Kernel::HLERequestContext& ctx) {
@@ -608,14 +681,16 @@ void AudRenU::OpenAudioRendererImpl(Kernel::HLERequestContext& ctx) {
    IPC::ResponseBuilder rb{ctx, 2, 0, 1};

    rb.Push(RESULT_SUCCESS);
-    rb.PushIpcInterface<IAudioRenderer>(params, audren_instance_count++);
+    rb.PushIpcInterface<IAudioRenderer>(system, params, audren_instance_count++);
 }

-bool AudRenU::IsFeatureSupported(AudioFeatures feature, u32_le revision) const {
+bool IsFeatureSupported(AudioFeatures feature, u32_le revision) {
    // Byte swap
    const u32_be version_num = revision - Common::MakeMagic('R', 'E', 'V', '0');

    switch (feature) {
+    case AudioFeatures::AudioUSBDeviceOutput:
+        return version_num >= 4U;
    case AudioFeatures::Splitter:
        return version_num >= 2U;
    case AudioFeatures::PerformanceMetricsVersion2:
@@ -6,6 +6,10 @@

 #include "core/hle/service/service.h"

+namespace Core {
+class System;
+}
+
 namespace Kernel {
 class HLERequestContext;
 }
@@ -14,7 +18,7 @@ namespace Service::Audio {

 class AudRenU final : public ServiceFramework<AudRenU> {
 public:
-    explicit AudRenU();
+    explicit AudRenU(Core::System& system_);
    ~AudRenU() override;

 private:
@@ -26,14 +30,19 @@ private:

    void OpenAudioRendererImpl(Kernel::HLERequestContext& ctx);

-    enum class AudioFeatures : u32 {
-        Splitter,
-        PerformanceMetricsVersion2,
-        VariadicCommandBuffer,
-    };
-
-    bool IsFeatureSupported(AudioFeatures feature, u32_le revision) const;
    std::size_t audren_instance_count = 0;
+    Core::System& system;
 };

+// Describes a particular audio feature that may be supported in a particular revision.
+enum class AudioFeatures : u32 {
+    AudioUSBDeviceOutput,
+    Splitter,
+    PerformanceMetricsVersion2,
+    VariadicCommandBuffer,
+};
+
+// Tests if a particular audio feature is supported with a given audio revision.
+bool IsFeatureSupported(AudioFeatures feature, u32_le revision);
+
 } // namespace Service::Audio
@@ -2,32 +2,37 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.

+#include "core/crypto/key_manager.h"
+#include "core/hle/ipc_helpers.h"
 #include "core/hle/service/service.h"

 namespace Service::ES {

+constexpr ResultCode ERROR_INVALID_ARGUMENT{ErrorModule::ETicket, 2};
+constexpr ResultCode ERROR_INVALID_RIGHTS_ID{ErrorModule::ETicket, 3};
+
 class ETicket final : public ServiceFramework<ETicket> {
 public:
    explicit ETicket() : ServiceFramework{"es"} {
        // clang-format off
        static const FunctionInfo functions[] = {
-            {1, nullptr, "ImportTicket"},
+            {1, &ETicket::ImportTicket, "ImportTicket"},
            {2, nullptr, "ImportTicketCertificateSet"},
            {3, nullptr, "DeleteTicket"},
            {4, nullptr, "DeletePersonalizedTicket"},
            {5, nullptr, "DeleteAllCommonTicket"},
            {6, nullptr, "DeleteAllPersonalizedTicket"},
            {7, nullptr, "DeleteAllPersonalizedTicketEx"},
-            {8, nullptr, "GetTitleKey"},
-            {9, nullptr, "CountCommonTicket"},
-            {10, nullptr, "CountPersonalizedTicket"},
-            {11, nullptr, "ListCommonTicket"},
-            {12, nullptr, "ListPersonalizedTicket"},
+            {8, &ETicket::GetTitleKey, "GetTitleKey"},
+            {9, &ETicket::CountCommonTicket, "CountCommonTicket"},
+            {10, &ETicket::CountPersonalizedTicket, "CountPersonalizedTicket"},
+            {11, &ETicket::ListCommonTicket, "ListCommonTicket"},
+            {12, &ETicket::ListPersonalizedTicket, "ListPersonalizedTicket"},
            {13, nullptr, "ListMissingPersonalizedTicket"},
-            {14, nullptr, "GetCommonTicketSize"},
-            {15, nullptr, "GetPersonalizedTicketSize"},
-            {16, nullptr, "GetCommonTicketData"},
-            {17, nullptr, "GetPersonalizedTicketData"},
+            {14, &ETicket::GetCommonTicketSize, "GetCommonTicketSize"},
+            {15, &ETicket::GetPersonalizedTicketSize, "GetPersonalizedTicketSize"},
+            {16, &ETicket::GetCommonTicketData, "GetCommonTicketData"},
+            {17, &ETicket::GetPersonalizedTicketData, "GetPersonalizedTicketData"},
            {18, nullptr, "OwnTicket"},
            {19, nullptr, "GetTicketInfo"},
            {20, nullptr, "ListLightTicketInfo"},
@@ -51,7 +56,212 @@ public:
        };
        // clang-format on
        RegisterHandlers(functions);
+
+        keys.PopulateTickets();
+        keys.SynthesizeTickets();
    }
+
+private:
+    bool CheckRightsId(Kernel::HLERequestContext& ctx, const u128& rights_id) {
+        if (rights_id == u128{}) {
+            LOG_ERROR(Service_ETicket, "The rights ID was invalid!");
+            IPC::ResponseBuilder rb{ctx, 2};
+            rb.Push(ERROR_INVALID_RIGHTS_ID);
+            return false;
+        }
+
+        return true;
+    }
+
+    void ImportTicket(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const auto ticket = ctx.ReadBuffer();
+        const auto cert = ctx.ReadBuffer(1);
+
+        if (ticket.size() < sizeof(Core::Crypto::Ticket)) {
+            LOG_ERROR(Service_ETicket, "The input buffer is not large enough!");
+            IPC::ResponseBuilder rb{ctx, 2};
+            rb.Push(ERROR_INVALID_ARGUMENT);
+            return;
+        }
+
+        Core::Crypto::Ticket raw{};
+        std::memcpy(&raw, ticket.data(), sizeof(Core::Crypto::Ticket));
+
+        if (!keys.AddTicketPersonalized(raw)) {
+            LOG_ERROR(Service_ETicket, "The ticket could not be imported!");
+            IPC::ResponseBuilder rb{ctx, 2};
+            rb.Push(ERROR_INVALID_ARGUMENT);
+            return;
+        }
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void GetTitleKey(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const auto rights_id = rp.PopRaw<u128>();
+
+        LOG_DEBUG(Service_ETicket, "called, rights_id={:016X}{:016X}", rights_id[1], rights_id[0]);
+
+        if (!CheckRightsId(ctx, rights_id))
+            return;
+
+        const auto key =
+            keys.GetKey(Core::Crypto::S128KeyType::Titlekey, rights_id[1], rights_id[0]);
+
+        if (key == Core::Crypto::Key128{}) {
+            LOG_ERROR(Service_ETicket,
+                      "The titlekey doesn't exist in the KeyManager or the rights ID was invalid!");
+            IPC::ResponseBuilder rb{ctx, 2};
+            rb.Push(ERROR_INVALID_RIGHTS_ID);
+            return;
+        }
+
+        ctx.WriteBuffer(key.data(), key.size());
+
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
+    void CountCommonTicket(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_ETicket, "called");
+
+        const auto count = keys.GetCommonTickets().size();
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(count);
+    }
+
+    void CountPersonalizedTicket(Kernel::HLERequestContext& ctx) {
+        LOG_DEBUG(Service_ETicket, "called");
+
+        const auto count = keys.GetPersonalizedTickets().size();
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(count);
+    }
+
+    void ListCommonTicket(Kernel::HLERequestContext& ctx) {
+        u32 out_entries;
+        if (keys.GetCommonTickets().empty())
+            out_entries = 0;
+        else
+            out_entries = ctx.GetWriteBufferSize() / sizeof(u128);
+
+        LOG_DEBUG(Service_ETicket, "called, entries={:016X}", out_entries);
+
+        keys.PopulateTickets();
+        const auto tickets = keys.GetCommonTickets();
+        std::vector<u128> ids;
+        std::transform(tickets.begin(), tickets.end(), std::back_inserter(ids),
+                       [](const auto& pair) { return pair.first; });
+
+        out_entries = std::min<u32>(ids.size(), out_entries);
+        ctx.WriteBuffer(ids.data(), out_entries * sizeof(u128));
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(out_entries);
+    }
+
+    void ListPersonalizedTicket(Kernel::HLERequestContext& ctx) {
+        u32 out_entries;
+        if (keys.GetPersonalizedTickets().empty())
+            out_entries = 0;
+        else
+            out_entries = ctx.GetWriteBufferSize() / sizeof(u128);
+
+        LOG_DEBUG(Service_ETicket, "called, entries={:016X}", out_entries);
+
+        keys.PopulateTickets();
+        const auto tickets = keys.GetPersonalizedTickets();
+        std::vector<u128> ids;
+        std::transform(tickets.begin(), tickets.end(), std::back_inserter(ids),
+                       [](const auto& pair) { return pair.first; });
+
+        out_entries = std::min<u32>(ids.size(), out_entries);
+        ctx.WriteBuffer(ids.data(), out_entries * sizeof(u128));
+
+        IPC::ResponseBuilder rb{ctx, 3};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(out_entries);
+    }
+
+    void GetCommonTicketSize(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const auto rights_id = rp.PopRaw<u128>();
+
+        LOG_DEBUG(Service_ETicket, "called, rights_id={:016X}{:016X}", rights_id[1], rights_id[0]);
+
+        if (!CheckRightsId(ctx, rights_id))
+            return;
+
+        const auto ticket = keys.GetCommonTickets().at(rights_id);
+
+        IPC::ResponseBuilder rb{ctx, 4};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u64>(ticket.GetSize());
+    }
+
+    void GetPersonalizedTicketSize(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const auto rights_id = rp.PopRaw<u128>();
+
+        LOG_DEBUG(Service_ETicket, "called, rights_id={:016X}{:016X}", rights_id[1], rights_id[0]);
+
+        if (!CheckRightsId(ctx, rights_id))
+            return;
+
+        const auto ticket = keys.GetPersonalizedTickets().at(rights_id);
+
+        IPC::ResponseBuilder rb{ctx, 4};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u64>(ticket.GetSize());
+    }
+
+    void GetCommonTicketData(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const auto rights_id = rp.PopRaw<u128>();
+
+        LOG_DEBUG(Service_ETicket, "called, rights_id={:016X}{:016X}", rights_id[1], rights_id[0]);
+
+        if (!CheckRightsId(ctx, rights_id))
+            return;
+
+        const auto ticket = keys.GetCommonTickets().at(rights_id);
+
+        const auto write_size = std::min<u64>(ticket.GetSize(), ctx.GetWriteBufferSize());
+        ctx.WriteBuffer(&ticket, write_size);
+
+        IPC::ResponseBuilder rb{ctx, 4};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u64>(write_size);
+    }
+
+    void GetPersonalizedTicketData(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const auto rights_id = rp.PopRaw<u128>();
+
+        LOG_DEBUG(Service_ETicket, "called, rights_id={:016X}{:016X}", rights_id[1], rights_id[0]);
+
+        if (!CheckRightsId(ctx, rights_id))
+            return;
+
+        const auto ticket = keys.GetPersonalizedTickets().at(rights_id);
+
+        const auto write_size = std::min<u64>(ticket.GetSize(), ctx.GetWriteBufferSize());
+        ctx.WriteBuffer(&ticket, write_size);
+
+        IPC::ResponseBuilder rb{ctx, 4};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u64>(write_size);
+    }
+
+    Core::Crypto::KeyManager keys;
 };

 void InstallInterfaces(SM::ServiceManager& service_manager) {
@@ -5,7 +5,7 @@
 #include <array>
 #include <cstring>
 #include <ctime>
-#include <fmt/time.h>
+#include <fmt/chrono.h>
 #include "common/file_util.h"
 #include "common/logging/log.h"
 #include "common/scm_rev.h"
@@ -636,10 +636,15 @@ Controller_NPad::LedPattern Controller_NPad::GetLedPattern(u32 npad_id) {
        return LedPattern{0, 0, 0, 0};
    };
 }
+
 void Controller_NPad::SetVibrationEnabled(bool can_vibrate) {
    can_controllers_vibrate = can_vibrate;
 }

+bool Controller_NPad::IsVibrationEnabled() const {
+    return can_controllers_vibrate;
+}
+
 void Controller_NPad::ClearAllConnectedControllers() {
    for (auto& controller : connected_controllers) {
        if (controller.is_connected && controller.type != NPadControllerType::None) {
@@ -648,6 +653,7 @@ void Controller_NPad::ClearAllConnectedControllers() {
        }
    }
 }
+
 void Controller_NPad::DisconnectAllConnectedControllers() {
    std::for_each(connected_controllers.begin(), connected_controllers.end(),
                  [](ControllerHolder& controller) { controller.is_connected = false; });
@@ -119,6 +119,7 @@ public:
    void DisconnectNPad(u32 npad_id);
    LedPattern GetLedPattern(u32 npad_id);
    void SetVibrationEnabled(bool can_vibrate);
+    bool IsVibrationEnabled() const;
    void ClearAllConnectedControllers();
    void DisconnectAllConnectedControllers();
    void ConnectAllDisconnectedControllers();
@@ -216,8 +216,8 @@ Hid::Hid() : ServiceFramework("hid") {
        {201, &Hid::SendVibrationValue, "SendVibrationValue"},
        {202, &Hid::GetActualVibrationValue, "GetActualVibrationValue"},
        {203, &Hid::CreateActiveVibrationDeviceList, "CreateActiveVibrationDeviceList"},
-        {204, nullptr, "PermitVibration"},
-        {205, nullptr, "IsVibrationPermitted"},
+        {204, &Hid::PermitVibration, "PermitVibration"},
+        {205, &Hid::IsVibrationPermitted, "IsVibrationPermitted"},
        {206, &Hid::SendVibrationValues, "SendVibrationValues"},
        {207, nullptr, "SendVibrationGcErmCommand"},
        {208, nullptr, "GetActualVibrationGcErmCommand"},
@@ -679,6 +679,27 @@ void Hid::CreateActiveVibrationDeviceList(Kernel::HLERequestContext& ctx) {
    rb.PushIpcInterface<IActiveVibrationDeviceList>();
 }

+void Hid::PermitVibration(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto can_vibrate{rp.Pop<bool>()};
+    applet_resource->GetController<Controller_NPad>(HidController::NPad)
+        .SetVibrationEnabled(can_vibrate);
+
+    LOG_DEBUG(Service_HID, "called, can_vibrate={}", can_vibrate);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
+void Hid::IsVibrationPermitted(Kernel::HLERequestContext& ctx) {
+    LOG_DEBUG(Service_HID, "called");
+
+    IPC::ResponseBuilder rb{ctx, 3};
+    rb.Push(RESULT_SUCCESS);
+    rb.Push(
+        applet_resource->GetController<Controller_NPad>(HidController::NPad).IsVibrationEnabled());
+}
+
 void Hid::ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) {
    IPC::RequestParser rp{ctx};
    const auto applet_resource_user_id{rp.Pop<u64>()};
@@ -114,6 +114,8 @@ private:
    void SetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx);
    void GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx);
    void CreateActiveVibrationDeviceList(Kernel::HLERequestContext& ctx);
+    void PermitVibration(Kernel::HLERequestContext& ctx);
+    void IsVibrationPermitted(Kernel::HLERequestContext& ctx);
    void ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx);
    void StartConsoleSixAxisSensor(Kernel::HLERequestContext& ctx);
    void StopSixAxisSensor(Kernel::HLERequestContext& ctx);
@@ -175,6 +175,10 @@ MiiStoreData ConvertInfoToStoreData(const MiiInfo& info) {
 } // namespace

 std::ostream& operator<<(std::ostream& os, Source source) {
+    if (static_cast<std::size_t>(source) >= SOURCE_NAMES.size()) {
+        return os << "[UNKNOWN SOURCE]";
+    }
+
    os << SOURCE_NAMES.at(static_cast<std::size_t>(source));
    return os;
 }
@@ -77,7 +77,7 @@ enum class LoadState : u32 {
    Done = 1,
 };

-static void DecryptSharedFont(const std::vector<u32>& input, std::vector<u8>& output,
+static void DecryptSharedFont(const std::vector<u32>& input, Kernel::PhysicalMemory& output,
                              std::size_t& offset) {
    ASSERT_MSG(offset + (input.size() * sizeof(u32)) < SHARED_FONT_MEM_SIZE,
               "Shared fonts exceeds 17mb!");
@@ -94,7 +94,7 @@ static void DecryptSharedFont(const std::vector<u32>& input, std::vector<u8>& ou
    offset += transformed_font.size() * sizeof(u32);
 }

-static void EncryptSharedFont(const std::vector<u8>& input, std::vector<u8>& output,
+static void EncryptSharedFont(const std::vector<u8>& input, Kernel::PhysicalMemory& output,
                              std::size_t& offset) {
    ASSERT_MSG(offset + input.size() + 8 < SHARED_FONT_MEM_SIZE, "Shared fonts exceeds 17mb!");
    const u32 KEY = EXPECTED_MAGIC ^ EXPECTED_RESULT;
@@ -121,7 +121,7 @@ struct PL_U::Impl {
        return shared_font_regions.at(index);
    }

-    void BuildSharedFontsRawRegions(const std::vector<u8>& input) {
+    void BuildSharedFontsRawRegions(const Kernel::PhysicalMemory& input) {
        // As we can derive the xor key we can just populate the offsets
        // based on the shared memory dump
        unsigned cur_offset = 0;
@@ -144,7 +144,7 @@ struct PL_U::Impl {
    Kernel::SharedPtr<Kernel::SharedMemory> shared_font_mem;

    /// Backing memory for the shared font data
-    std::shared_ptr<std::vector<u8>> shared_font;
+    std::shared_ptr<Kernel::PhysicalMemory> shared_font;

    // Automatically populated based on shared_fonts dump or system archives.
    std::vector<FontRegion> shared_font_regions;
@@ -166,7 +166,7 @@ PL_U::PL_U() : ServiceFramework("pl:u"), impl{std::make_unique<Impl>()} {
    // Rebuild shared fonts from data ncas
    if (nand->HasEntry(static_cast<u64>(FontArchives::Standard),
                       FileSys::ContentRecordType::Data)) {
-        impl->shared_font = std::make_shared<std::vector<u8>>(SHARED_FONT_MEM_SIZE);
+        impl->shared_font = std::make_shared<Kernel::PhysicalMemory>(SHARED_FONT_MEM_SIZE);
        for (auto font : SHARED_FONTS) {
            const auto nca =
                nand->GetEntry(static_cast<u64>(font.first), FileSys::ContentRecordType::Data);
@@ -207,7 +207,7 @@ PL_U::PL_U() : ServiceFramework("pl:u"), impl{std::make_unique<Impl>()} {
        }

    } else {
-        impl->shared_font = std::make_shared<std::vector<u8>>(
+        impl->shared_font = std::make_shared<Kernel::PhysicalMemory>(
            SHARED_FONT_MEM_SIZE); // Shared memory needs to always be allocated and a fixed size

        const std::string user_path = FileUtil::GetUserPath(FileUtil::UserPath::SysDataDir);
@@ -37,7 +37,7 @@ void nvdisp_disp0::flip(u32 buffer_handle, u32 offset, u32 format, u32 width, u3
        transform, crop_rect};

    system.GetPerfStats().EndGameFrame();
-    system.GPU().SwapBuffers(framebuffer);
+    system.GPU().SwapBuffers(&framebuffer);
 }

 } // namespace Service::Nvidia::Devices
@@ -206,7 +206,7 @@ void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system) {
    AM::InstallInterfaces(*sm, nv_flinger, system);
    AOC::InstallInterfaces(*sm);
    APM::InstallInterfaces(system);
-    Audio::InstallInterfaces(*sm);
+    Audio::InstallInterfaces(*sm, system);
    BCAT::InstallInterfaces(*sm);
    BPC::InstallInterfaces(*sm);
    BtDrv::InstallInterfaces(*sm);
@@ -295,7 +295,7 @@ Kernel::CodeSet ElfReader::LoadInto(VAddr vaddr) {
        }
    }

-    std::vector<u8> program_image(total_image_size);
+    Kernel::PhysicalMemory program_image(total_image_size);
    std::size_t current_image_position = 0;

    Kernel::CodeSet codeset;
@@ -69,7 +69,7 @@ AppLoader::LoadResult AppLoader_KIP::Load(Kernel::Process& process) {

    const VAddr base_address = process.VMManager().GetCodeRegionBaseAddress();
    Kernel::CodeSet codeset;
-    std::vector<u8> program_image;
+    Kernel::PhysicalMemory program_image;

    const auto load_segment = [&program_image](Kernel::CodeSet::Segment& segment,
                                               const std::vector<u8>& data, u32 offset) {
@@ -143,7 +143,7 @@ static bool LoadNroImpl(Kernel::Process& process, const std::vector<u8>& data,
    }

    // Build program image
-    std::vector<u8> program_image(PageAlignSize(nro_header.file_size));
+    Kernel::PhysicalMemory program_image(PageAlignSize(nro_header.file_size));
    std::memcpy(program_image.data(), data.data(), program_image.size());
    if (program_image.size() != PageAlignSize(nro_header.file_size)) {
        return {};
@@ -89,7 +89,7 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process,

    // Build program image
    Kernel::CodeSet codeset;
-    std::vector<u8> program_image;
+    Kernel::PhysicalMemory program_image;
    for (std::size_t i = 0; i < nso_header.segments.size(); ++i) {
        std::vector<u8> data =
            file.ReadBytes(nso_header.segments_compressed_size[i], nso_header.segments[i].offset);
@@ -5,8 +5,8 @@
 #include <ctime>
 #include <fstream>

+#include <fmt/chrono.h>
 #include <fmt/format.h>
-#include <fmt/time.h>
 #include <json.hpp>

 #include "common/file_util.h"
@@ -1,5 +1,7 @@
 add_library(video_core STATIC
-    buffer_cache.h
+    buffer_cache/buffer_block.h
+    buffer_cache/buffer_cache.h
+    buffer_cache/map_interval.h
    dma_pusher.cpp
    dma_pusher.h
    debug_utils/debug_utils.cpp
@@ -100,6 +102,7 @@ add_library(video_core STATIC
    shader/decode/integer_set.cpp
    shader/decode/half_set.cpp
    shader/decode/video.cpp
+    shader/decode/warp.cpp
    shader/decode/xmad.cpp
    shader/decode/other.cpp
    shader/control_flow.cpp
@@ -1,299 +0,0 @@
-// Copyright 2019 yuzu Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include <array>
-#include <memory>
-#include <mutex>
-#include <unordered_map>
-#include <unordered_set>
-#include <utility>
-#include <vector>
-
-#include "common/alignment.h"
-#include "common/common_types.h"
-#include "core/core.h"
-#include "video_core/memory_manager.h"
-#include "video_core/rasterizer_cache.h"
-
-namespace VideoCore {
-class RasterizerInterface;
-}
-
-namespace VideoCommon {
-
-template <typename BufferStorageType>
-class CachedBuffer final : public RasterizerCacheObject {
-public:
-    explicit CachedBuffer(VAddr cpu_addr, u8* host_ptr)
-        : RasterizerCacheObject{host_ptr}, host_ptr{host_ptr}, cpu_addr{cpu_addr} {}
-    ~CachedBuffer() override = default;
-
-    VAddr GetCpuAddr() const override {
-        return cpu_addr;
-    }
-
-    std::size_t GetSizeInBytes() const override {
-        return size;
-    }
-
-    u8* GetWritableHostPtr() const {
-        return host_ptr;
-    }
-
-    std::size_t GetSize() const {
-        return size;
-    }
-
-    std::size_t GetCapacity() const {
-        return capacity;
-    }
-
-    bool IsInternalized() const {
-        return is_internal;
-    }
-
-    const BufferStorageType& GetBuffer() const {
-        return buffer;
-    }
-
-    void SetSize(std::size_t new_size) {
-        size = new_size;
-    }
-
-    void SetInternalState(bool is_internal_) {
-        is_internal = is_internal_;
-    }
-
-    BufferStorageType ExchangeBuffer(BufferStorageType buffer_, std::size_t new_capacity) {
-        capacity = new_capacity;
-        std::swap(buffer, buffer_);
-        return buffer_;
-    }
-
-private:
-    u8* host_ptr{};
-    VAddr cpu_addr{};
-    std::size_t size{};
-    std::size_t capacity{};
-    bool is_internal{};
-    BufferStorageType buffer;
-};
-
-template <typename BufferStorageType, typename BufferType, typename StreamBuffer>
-class BufferCache : public RasterizerCache<std::shared_ptr<CachedBuffer<BufferStorageType>>> {
-public:
-    using Buffer = std::shared_ptr<CachedBuffer<BufferStorageType>>;
-    using BufferInfo = std::pair<const BufferType*, u64>;
-
-    explicit BufferCache(VideoCore::RasterizerInterface& rasterizer, Core::System& system,
-                         std::unique_ptr<StreamBuffer> stream_buffer)
-        : RasterizerCache<Buffer>{rasterizer}, system{system},
-          stream_buffer{std::move(stream_buffer)}, stream_buffer_handle{
-                                                       this->stream_buffer->GetHandle()} {}
-    ~BufferCache() = default;
-
-    void Unregister(const Buffer& entry) override {
-        std::lock_guard lock{RasterizerCache<Buffer>::mutex};
-        if (entry->IsInternalized()) {
-            internalized_entries.erase(entry->GetCacheAddr());
-        }
-        ReserveBuffer(entry);
-        RasterizerCache<Buffer>::Unregister(entry);
-    }
-
-    void TickFrame() {
-        marked_for_destruction_index =
-            (marked_for_destruction_index + 1) % marked_for_destruction_ring_buffer.size();
-        MarkedForDestruction().clear();
-    }
-
-    BufferInfo UploadMemory(GPUVAddr gpu_addr, std::size_t size, std::size_t alignment = 4,
-                            bool internalize = false, bool is_written = false) {
-        std::lock_guard lock{RasterizerCache<Buffer>::mutex};
-
-        auto& memory_manager = system.GPU().MemoryManager();
-        const auto host_ptr = memory_manager.GetPointer(gpu_addr);
-        if (!host_ptr) {
-            return {GetEmptyBuffer(size), 0};
-        }
-        const auto cache_addr = ToCacheAddr(host_ptr);
-
-        // Cache management is a big overhead, so only cache entries with a given size.
-        // TODO: Figure out which size is the best for given games.
-        constexpr std::size_t max_stream_size = 0x800;
-        if (!internalize && size < max_stream_size &&
-            internalized_entries.find(cache_addr) == internalized_entries.end()) {
-            return StreamBufferUpload(host_ptr, size, alignment);
-        }
-
-        auto entry = RasterizerCache<Buffer>::TryGet(cache_addr);
-        if (!entry) {
-            return FixedBufferUpload(gpu_addr, host_ptr, size, internalize, is_written);
-        }
-
-        if (entry->GetSize() < size) {
-            IncreaseBufferSize(entry, size);
-        }
-        if (is_written) {
-            entry->MarkAsModified(true, *this);
-        }
-        return {ToHandle(entry->GetBuffer()), 0};
-    }
-
-    /// Uploads from a host memory. Returns the OpenGL buffer where it's located and its offset.
-    BufferInfo UploadHostMemory(const void* raw_pointer, std::size_t size,
-                                std::size_t alignment = 4) {
-        std::lock_guard lock{RasterizerCache<Buffer>::mutex};
-        return StreamBufferUpload(raw_pointer, size, alignment);
-    }
-
-    void Map(std::size_t max_size) {
-        std::tie(buffer_ptr, buffer_offset_base, invalidated) = stream_buffer->Map(max_size, 4);
-        buffer_offset = buffer_offset_base;
-    }
-
-    /// Finishes the upload stream, returns true on bindings invalidation.
-    bool Unmap() {
-        stream_buffer->Unmap(buffer_offset - buffer_offset_base);
-        return std::exchange(invalidated, false);
-    }
-
-    virtual const BufferType* GetEmptyBuffer(std::size_t size) = 0;
-
-protected:
-    void FlushObjectInner(const Buffer& entry) override {
-        DownloadBufferData(entry->GetBuffer(), 0, entry->GetSize(), entry->GetWritableHostPtr());
-    }
-
-    virtual BufferStorageType CreateBuffer(std::size_t size) = 0;
-
-    virtual const BufferType* ToHandle(const BufferStorageType& storage) = 0;
-
-    virtual void UploadBufferData(const BufferStorageType& buffer, std::size_t offset,
-                                  std::size_t size, const u8* data) = 0;
-
-    virtual void DownloadBufferData(const BufferStorageType& buffer, std::size_t offset,
-                                    std::size_t size, u8* data) = 0;
-
-    virtual void CopyBufferData(const BufferStorageType& src, const BufferStorageType& dst,
-                                std::size_t src_offset, std::size_t dst_offset,
-                                std::size_t size) = 0;
-
-private:
-    BufferInfo StreamBufferUpload(const void* raw_pointer, std::size_t size,
-                                  std::size_t alignment) {
-        AlignBuffer(alignment);
-        const std::size_t uploaded_offset = buffer_offset;
-        std::memcpy(buffer_ptr, raw_pointer, size);
-
-        buffer_ptr += size;
-        buffer_offset += size;
-        return {&stream_buffer_handle, uploaded_offset};
-    }
-
-    BufferInfo FixedBufferUpload(GPUVAddr gpu_addr, u8* host_ptr, std::size_t size,
-                                 bool internalize, bool is_written) {
-        auto& memory_manager = Core::System::GetInstance().GPU().MemoryManager();
-        const auto cpu_addr = memory_manager.GpuToCpuAddress(gpu_addr);
-        ASSERT(cpu_addr);
-
-        auto entry = GetUncachedBuffer(*cpu_addr, host_ptr);
-        entry->SetSize(size);
-        entry->SetInternalState(internalize);
-        RasterizerCache<Buffer>::Register(entry);
-
-        if (internalize) {
-            internalized_entries.emplace(ToCacheAddr(host_ptr));
-        }
-        if (is_written) {
-            entry->MarkAsModified(true, *this);
-        }
-
-        if (entry->GetCapacity() < size) {
-            MarkedForDestruction().push_back(entry->ExchangeBuffer(CreateBuffer(size), size));
-        }
-
-        UploadBufferData(entry->GetBuffer(), 0, size, host_ptr);
-        return {ToHandle(entry->GetBuffer()), 0};
-    }
-
-    void IncreaseBufferSize(Buffer& entry, std::size_t new_size) {
-        const std::size_t old_size = entry->GetSize();
-        if (entry->GetCapacity() < new_size) {
-            const auto& old_buffer = entry->GetBuffer();
-            auto new_buffer = CreateBuffer(new_size);
-
-            // Copy bits from the old buffer to the new buffer.
-            CopyBufferData(old_buffer, new_buffer, 0, 0, old_size);
-            MarkedForDestruction().push_back(
-                entry->ExchangeBuffer(std::move(new_buffer), new_size));
-
-            // This buffer could have been used
-            invalidated = true;
-        }
-        // Upload the new bits.
-        const std::size_t size_diff = new_size - old_size;
-        UploadBufferData(entry->GetBuffer(), old_size, size_diff, entry->GetHostPtr() + old_size);
-
-        // Update entry's size in the object and in the cache.
-        Unregister(entry);
-
-        entry->SetSize(new_size);
-        RasterizerCache<Buffer>::Register(entry);
-    }
-
-    Buffer GetUncachedBuffer(VAddr cpu_addr, u8* host_ptr) {
-        if (auto entry = TryGetReservedBuffer(host_ptr)) {
-            return entry;
-        }
-        return std::make_shared<CachedBuffer<BufferStorageType>>(cpu_addr, host_ptr);
-    }
-
-    Buffer TryGetReservedBuffer(u8* host_ptr) {
-        const auto it = buffer_reserve.find(ToCacheAddr(host_ptr));
-        if (it == buffer_reserve.end()) {
-            return {};
-        }
-        auto& reserve = it->second;
-        auto entry = reserve.back();
-        reserve.pop_back();
-        return entry;
-    }
-
-    void ReserveBuffer(Buffer entry) {
-        buffer_reserve[entry->GetCacheAddr()].push_back(std::move(entry));
-    }
-
-    void AlignBuffer(std::size_t alignment) {
-        // Align the offset, not the mapped pointer
-        const std::size_t offset_aligned = Common::AlignUp(buffer_offset, alignment);
-        buffer_ptr += offset_aligned - buffer_offset;
-        buffer_offset = offset_aligned;
-    }
-
-    std::vector<BufferStorageType>& MarkedForDestruction() {
-        return marked_for_destruction_ring_buffer[marked_for_destruction_index];
-    }
-
-    Core::System& system;
-
-    std::unique_ptr<StreamBuffer> stream_buffer;
-    BufferType stream_buffer_handle{};
-
-    bool invalidated = false;
-
-    u8* buffer_ptr = nullptr;
-    u64 buffer_offset = 0;
-    u64 buffer_offset_base = 0;
-
-    std::size_t marked_for_destruction_index = 0;
-    std::array<std::vector<BufferStorageType>, 4> marked_for_destruction_ring_buffer;
-
-    std::unordered_set<CacheAddr> internalized_entries;
-    std::unordered_map<CacheAddr, std::vector<Buffer>> buffer_reserve;
-};
-
-} // namespace VideoCommon
@@ -0,0 +1,76 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <unordered_set>
+#include <utility>
+
+#include "common/alignment.h"
+#include "common/common_types.h"
+#include "video_core/gpu.h"
+
+namespace VideoCommon {
+
+class BufferBlock {
+public:
+    bool Overlaps(const CacheAddr start, const CacheAddr end) const {
+        return (cache_addr < end) && (cache_addr_end > start);
+    }
+
+    bool IsInside(const CacheAddr other_start, const CacheAddr other_end) const {
+        return cache_addr <= other_start && other_end <= cache_addr_end;
+    }
+
+    u8* GetWritableHostPtr() const {
+        return FromCacheAddr(cache_addr);
+    }
+
+    u8* GetWritableHostPtr(std::size_t offset) const {
+        return FromCacheAddr(cache_addr + offset);
+    }
+
+    std::size_t GetOffset(const CacheAddr in_addr) {
+        return static_cast<std::size_t>(in_addr - cache_addr);
+    }
+
+    CacheAddr GetCacheAddr() const {
+        return cache_addr;
+    }
+
+    CacheAddr GetCacheAddrEnd() const {
+        return cache_addr_end;
+    }
+
+    void SetCacheAddr(const CacheAddr new_addr) {
+        cache_addr = new_addr;
+        cache_addr_end = new_addr + size;
+    }
+
+    std::size_t GetSize() const {
+        return size;
+    }
+
+    void SetEpoch(u64 new_epoch) {
+        epoch = new_epoch;
+    }
+
+    u64 GetEpoch() {
+        return epoch;
+    }
+
+protected:
+    explicit BufferBlock(CacheAddr cache_addr, const std::size_t size) : size{size} {
+        SetCacheAddr(cache_addr);
+    }
+    ~BufferBlock() = default;
+
+private:
+    CacheAddr cache_addr{};
+    CacheAddr cache_addr_end{};
+    std::size_t size{};
+    u64 epoch{};
+};
+
+} // namespace VideoCommon
@@ -0,0 +1,447 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <memory>
+#include <mutex>
+#include <unordered_map>
+#include <unordered_set>
+#include <utility>
+#include <vector>
+
+#include "common/alignment.h"
+#include "common/common_types.h"
+#include "core/core.h"
+#include "video_core/buffer_cache/buffer_block.h"
+#include "video_core/buffer_cache/map_interval.h"
+#include "video_core/memory_manager.h"
+#include "video_core/rasterizer_interface.h"
+
+namespace VideoCommon {
+
+using MapInterval = std::shared_ptr<MapIntervalBase>;
+
+template <typename TBuffer, typename TBufferType, typename StreamBuffer>
+class BufferCache {
+public:
+    using BufferInfo = std::pair<const TBufferType*, u64>;
+
+    BufferInfo UploadMemory(GPUVAddr gpu_addr, std::size_t size, std::size_t alignment = 4,
+                            bool is_written = false) {
+        std::lock_guard lock{mutex};
+
+        auto& memory_manager = system.GPU().MemoryManager();
+        const auto host_ptr = memory_manager.GetPointer(gpu_addr);
+        if (!host_ptr) {
+            return {GetEmptyBuffer(size), 0};
+        }
+        const auto cache_addr = ToCacheAddr(host_ptr);
+
+        // Cache management is a big overhead, so only cache entries with a given size.
+        // TODO: Figure out which size is the best for given games.
+        constexpr std::size_t max_stream_size = 0x800;
+        if (size < max_stream_size) {
+            if (!is_written && !IsRegionWritten(cache_addr, cache_addr + size - 1)) {
+                return StreamBufferUpload(host_ptr, size, alignment);
+            }
+        }
+
+        auto block = GetBlock(cache_addr, size);
+        auto map = MapAddress(block, gpu_addr, cache_addr, size);
+        if (is_written) {
+            map->MarkAsModified(true, GetModifiedTicks());
+            if (!map->IsWritten()) {
+                map->MarkAsWritten(true);
+                MarkRegionAsWritten(map->GetStart(), map->GetEnd() - 1);
+            }
+        } else {
+            if (map->IsWritten()) {
+                WriteBarrier();
+            }
+        }
+
+        const u64 offset = static_cast<u64>(block->GetOffset(cache_addr));
+
+        return {ToHandle(block), offset};
+    }
+
+    /// Uploads from a host memory. Returns the OpenGL buffer where it's located and its offset.
+    BufferInfo UploadHostMemory(const void* raw_pointer, std::size_t size,
+                                std::size_t alignment = 4) {
+        std::lock_guard lock{mutex};
+        return StreamBufferUpload(raw_pointer, size, alignment);
+    }
+
+    void Map(std::size_t max_size) {
+        std::lock_guard lock{mutex};
+
+        std::tie(buffer_ptr, buffer_offset_base, invalidated) = stream_buffer->Map(max_size, 4);
+        buffer_offset = buffer_offset_base;
+    }
+
+    /// Finishes the upload stream, returns true on bindings invalidation.
+    bool Unmap() {
+        std::lock_guard lock{mutex};
+
+        stream_buffer->Unmap(buffer_offset - buffer_offset_base);
+        return std::exchange(invalidated, false);
+    }
+
+    void TickFrame() {
+        ++epoch;
+        while (!pending_destruction.empty()) {
+            if (pending_destruction.front()->GetEpoch() + 1 > epoch) {
+                break;
+            }
+            pending_destruction.pop_front();
+        }
+    }
+
+    /// Write any cached resources overlapping the specified region back to memory
+    void FlushRegion(CacheAddr addr, std::size_t size) {
+        std::lock_guard lock{mutex};
+
+        std::vector<MapInterval> objects = GetMapsInRange(addr, size);
+        std::sort(objects.begin(), objects.end(), [](const MapInterval& a, const MapInterval& b) {
+            return a->GetModificationTick() < b->GetModificationTick();
+        });
+        for (auto& object : objects) {
+            if (object->IsModified() && object->IsRegistered()) {
+                FlushMap(object);
+            }
+        }
+    }
+
+    /// Mark the specified region as being invalidated
+    void InvalidateRegion(CacheAddr addr, u64 size) {
+        std::lock_guard lock{mutex};
+
+        std::vector<MapInterval> objects = GetMapsInRange(addr, size);
+        for (auto& object : objects) {
+            if (object->IsRegistered()) {
+                Unregister(object);
+            }
+        }
+    }
+
+    virtual const TBufferType* GetEmptyBuffer(std::size_t size) = 0;
+
+protected:
+    explicit BufferCache(VideoCore::RasterizerInterface& rasterizer, Core::System& system,
+                         std::unique_ptr<StreamBuffer> stream_buffer)
+        : rasterizer{rasterizer}, system{system}, stream_buffer{std::move(stream_buffer)},
+          stream_buffer_handle{this->stream_buffer->GetHandle()} {}
+
+    ~BufferCache() = default;
+
+    virtual const TBufferType* ToHandle(const TBuffer& storage) = 0;
+
+    virtual void WriteBarrier() = 0;
+
+    virtual TBuffer CreateBlock(CacheAddr cache_addr, std::size_t size) = 0;
+
+    virtual void UploadBlockData(const TBuffer& buffer, std::size_t offset, std::size_t size,
+                                 const u8* data) = 0;
+
+    virtual void DownloadBlockData(const TBuffer& buffer, std::size_t offset, std::size_t size,
+                                   u8* data) = 0;
+
+    virtual void CopyBlock(const TBuffer& src, const TBuffer& dst, std::size_t src_offset,
+                           std::size_t dst_offset, std::size_t size) = 0;
+
+    /// Register an object into the cache
+    void Register(const MapInterval& new_map, bool inherit_written = false) {
+        const CacheAddr cache_ptr = new_map->GetStart();
+        const std::optional<VAddr> cpu_addr =
+            system.GPU().MemoryManager().GpuToCpuAddress(new_map->GetGpuAddress());
+        if (!cache_ptr || !cpu_addr) {
+            LOG_CRITICAL(HW_GPU, "Failed to register buffer with unmapped gpu_address 0x{:016x}",
+                         new_map->GetGpuAddress());
+            return;
+        }
+        const std::size_t size = new_map->GetEnd() - new_map->GetStart();
+        new_map->SetCpuAddress(*cpu_addr);
+        new_map->MarkAsRegistered(true);
+        const IntervalType interval{new_map->GetStart(), new_map->GetEnd()};
+        mapped_addresses.insert({interval, new_map});
+        rasterizer.UpdatePagesCachedCount(*cpu_addr, size, 1);
+        if (inherit_written) {
+            MarkRegionAsWritten(new_map->GetStart(), new_map->GetEnd() - 1);
+            new_map->MarkAsWritten(true);
+        }
+    }
+
+    /// Unregisters an object from the cache
+    void Unregister(MapInterval& map) {
+        const std::size_t size = map->GetEnd() - map->GetStart();
+        rasterizer.UpdatePagesCachedCount(map->GetCpuAddress(), size, -1);
+        map->MarkAsRegistered(false);
+        if (map->IsWritten()) {
+            UnmarkRegionAsWritten(map->GetStart(), map->GetEnd() - 1);
+        }
+        const IntervalType delete_interval{map->GetStart(), map->GetEnd()};
+        mapped_addresses.erase(delete_interval);
+    }
+
+private:
+    MapInterval CreateMap(const CacheAddr start, const CacheAddr end, const GPUVAddr gpu_addr) {
+        return std::make_shared<MapIntervalBase>(start, end, gpu_addr);
+    }
+
+    MapInterval MapAddress(const TBuffer& block, const GPUVAddr gpu_addr,
+                           const CacheAddr cache_addr, const std::size_t size) {
+
+        std::vector<MapInterval> overlaps = GetMapsInRange(cache_addr, size);
+        if (overlaps.empty()) {
+            const CacheAddr cache_addr_end = cache_addr + size;
+            MapInterval new_map = CreateMap(cache_addr, cache_addr_end, gpu_addr);
+            u8* host_ptr = FromCacheAddr(cache_addr);
+            UploadBlockData(block, block->GetOffset(cache_addr), size, host_ptr);
+            Register(new_map);
+            return new_map;
+        }
+
+        const CacheAddr cache_addr_end = cache_addr + size;
+        if (overlaps.size() == 1) {
+            MapInterval& current_map = overlaps[0];
+            if (current_map->IsInside(cache_addr, cache_addr_end)) {
+                return current_map;
+            }
+        }
+        CacheAddr new_start = cache_addr;
+        CacheAddr new_end = cache_addr_end;
+        bool write_inheritance = false;
+        bool modified_inheritance = false;
+        // Calculate new buffer parameters
+        for (auto& overlap : overlaps) {
+            new_start = std::min(overlap->GetStart(), new_start);
+            new_end = std::max(overlap->GetEnd(), new_end);
+            write_inheritance |= overlap->IsWritten();
+            modified_inheritance |= overlap->IsModified();
+        }
+        GPUVAddr new_gpu_addr = gpu_addr + new_start - cache_addr;
+        for (auto& overlap : overlaps) {
+            Unregister(overlap);
+        }
+        UpdateBlock(block, new_start, new_end, overlaps);
+        MapInterval new_map = CreateMap(new_start, new_end, new_gpu_addr);
+        if (modified_inheritance) {
+            new_map->MarkAsModified(true, GetModifiedTicks());
+        }
+        Register(new_map, write_inheritance);
+        return new_map;
+    }
+
+    void UpdateBlock(const TBuffer& block, CacheAddr start, CacheAddr end,
+                     std::vector<MapInterval>& overlaps) {
+        const IntervalType base_interval{start, end};
+        IntervalSet interval_set{};
+        interval_set.add(base_interval);
+        for (auto& overlap : overlaps) {
+            const IntervalType subtract{overlap->GetStart(), overlap->GetEnd()};
+            interval_set.subtract(subtract);
+        }
+        for (auto& interval : interval_set) {
+            std::size_t size = interval.upper() - interval.lower();
+            if (size > 0) {
+                u8* host_ptr = FromCacheAddr(interval.lower());
+                UploadBlockData(block, block->GetOffset(interval.lower()), size, host_ptr);
+            }
+        }
+    }
+
+    std::vector<MapInterval> GetMapsInRange(CacheAddr addr, std::size_t size) {
+        if (size == 0) {
+            return {};
+        }
+
+        std::vector<MapInterval> objects{};
+        const IntervalType interval{addr, addr + size};
+        for (auto& pair : boost::make_iterator_range(mapped_addresses.equal_range(interval))) {
+            objects.push_back(pair.second);
+        }
+
+        return objects;
+    }
+
+    /// Returns a ticks counter used for tracking when cached objects were last modified
+    u64 GetModifiedTicks() {
+        return ++modified_ticks;
+    }
+
+    void FlushMap(MapInterval map) {
+        std::size_t size = map->GetEnd() - map->GetStart();
+        TBuffer block = blocks[map->GetStart() >> block_page_bits];
+        u8* host_ptr = FromCacheAddr(map->GetStart());
+        DownloadBlockData(block, block->GetOffset(map->GetStart()), size, host_ptr);
+        map->MarkAsModified(false, 0);
+    }
+
+    BufferInfo StreamBufferUpload(const void* raw_pointer, std::size_t size,
+                                  std::size_t alignment) {
+        AlignBuffer(alignment);
+        const std::size_t uploaded_offset = buffer_offset;
+        std::memcpy(buffer_ptr, raw_pointer, size);
+
+        buffer_ptr += size;
+        buffer_offset += size;
+        return {&stream_buffer_handle, uploaded_offset};
+    }
+
+    void AlignBuffer(std::size_t alignment) {
+        // Align the offset, not the mapped pointer
+        const std::size_t offset_aligned = Common::AlignUp(buffer_offset, alignment);
+        buffer_ptr += offset_aligned - buffer_offset;
+        buffer_offset = offset_aligned;
+    }
+
+    TBuffer EnlargeBlock(TBuffer buffer) {
+        const std::size_t old_size = buffer->GetSize();
+        const std::size_t new_size = old_size + block_page_size;
+        const CacheAddr cache_addr = buffer->GetCacheAddr();
+        TBuffer new_buffer = CreateBlock(cache_addr, new_size);
+        CopyBlock(buffer, new_buffer, 0, 0, old_size);
+        buffer->SetEpoch(epoch);
+        pending_destruction.push_back(buffer);
+        const CacheAddr cache_addr_end = cache_addr + new_size - 1;
+        u64 page_start = cache_addr >> block_page_bits;
+        const u64 page_end = cache_addr_end >> block_page_bits;
+        while (page_start <= page_end) {
+            blocks[page_start] = new_buffer;
+            ++page_start;
+        }
+        return new_buffer;
+    }
+
+    TBuffer MergeBlocks(TBuffer first, TBuffer second) {
+        const std::size_t size_1 = first->GetSize();
+        const std::size_t size_2 = second->GetSize();
+        const CacheAddr first_addr = first->GetCacheAddr();
+        const CacheAddr second_addr = second->GetCacheAddr();
+        const CacheAddr new_addr = std::min(first_addr, second_addr);
+        const std::size_t new_size = size_1 + size_2;
+        TBuffer new_buffer = CreateBlock(new_addr, new_size);
+        CopyBlock(first, new_buffer, 0, new_buffer->GetOffset(first_addr), size_1);
+        CopyBlock(second, new_buffer, 0, new_buffer->GetOffset(second_addr), size_2);
+        first->SetEpoch(epoch);
+        second->SetEpoch(epoch);
+        pending_destruction.push_back(first);
+        pending_destruction.push_back(second);
+        const CacheAddr cache_addr_end = new_addr + new_size - 1;
+        u64 page_start = new_addr >> block_page_bits;
+        const u64 page_end = cache_addr_end >> block_page_bits;
+        while (page_start <= page_end) {
+            blocks[page_start] = new_buffer;
+            ++page_start;
+        }
+        return new_buffer;
+    }
+
+    TBuffer GetBlock(const CacheAddr cache_addr, const std::size_t size) {
+        TBuffer found{};
+        const CacheAddr cache_addr_end = cache_addr + size - 1;
+        u64 page_start = cache_addr >> block_page_bits;
+        const u64 page_end = cache_addr_end >> block_page_bits;
+        while (page_start <= page_end) {
+            auto it = blocks.find(page_start);
+            if (it == blocks.end()) {
+                if (found) {
+                    found = EnlargeBlock(found);
+                } else {
+                    const CacheAddr start_addr = (page_start << block_page_bits);
+                    found = CreateBlock(start_addr, block_page_size);
+                    blocks[page_start] = found;
+                }
+            } else {
+                if (found) {
+                    if (found == it->second) {
+                        ++page_start;
+                        continue;
+                    }
+                    found = MergeBlocks(found, it->second);
+                } else {
+                    found = it->second;
+                }
+            }
+            ++page_start;
+        }
+        return found;
+    }
+
+    void MarkRegionAsWritten(const CacheAddr start, const CacheAddr end) {
+        u64 page_start = start >> write_page_bit;
+        const u64 page_end = end >> write_page_bit;
+        while (page_start <= page_end) {
+            auto it = written_pages.find(page_start);
+            if (it != written_pages.end()) {
+                it->second = it->second + 1;
+            } else {
+                written_pages[page_start] = 1;
+            }
+            page_start++;
+        }
+    }
+
+    void UnmarkRegionAsWritten(const CacheAddr start, const CacheAddr end) {
+        u64 page_start = start >> write_page_bit;
+        const u64 page_end = end >> write_page_bit;
+        while (page_start <= page_end) {
+            auto it = written_pages.find(page_start);
+            if (it != written_pages.end()) {
+                if (it->second > 1) {
+                    it->second = it->second - 1;
+                } else {
+                    written_pages.erase(it);
+                }
+            }
+            page_start++;
+        }
+    }
+
+    bool IsRegionWritten(const CacheAddr start, const CacheAddr end) const {
+        u64 page_start = start >> write_page_bit;
+        const u64 page_end = end >> write_page_bit;
+        while (page_start <= page_end) {
+            if (written_pages.count(page_start) > 0) {
+                return true;
+            }
+            page_start++;
+        }
+        return false;
+    }
+
+    VideoCore::RasterizerInterface& rasterizer;
+    Core::System& system;
+    std::unique_ptr<StreamBuffer> stream_buffer;
+
+    TBufferType stream_buffer_handle{};
+
+    bool invalidated = false;
+
+    u8* buffer_ptr = nullptr;
+    u64 buffer_offset = 0;
+    u64 buffer_offset_base = 0;
+
+    using IntervalSet = boost::icl::interval_set<CacheAddr>;
+    using IntervalCache = boost::icl::interval_map<CacheAddr, MapInterval>;
+    using IntervalType = typename IntervalCache::interval_type;
+    IntervalCache mapped_addresses{};
+
+    static constexpr u64 write_page_bit{11};
+    std::unordered_map<u64, u32> written_pages{};
+
+    static constexpr u64 block_page_bits{21};
+    static constexpr u64 block_page_size{1 << block_page_bits};
+    std::unordered_map<u64, TBuffer> blocks{};
+
+    std::list<TBuffer> pending_destruction{};
+    u64 epoch{};
+    u64 modified_ticks{};
+
+    std::recursive_mutex mutex;
+};
+
+} // namespace VideoCommon
@@ -0,0 +1,89 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "common/common_types.h"
+#include "video_core/gpu.h"
+
+namespace VideoCommon {
+
+class MapIntervalBase {
+public:
+    MapIntervalBase(const CacheAddr start, const CacheAddr end, const GPUVAddr gpu_addr)
+        : start{start}, end{end}, gpu_addr{gpu_addr} {}
+
+    void SetCpuAddress(VAddr new_cpu_addr) {
+        cpu_addr = new_cpu_addr;
+    }
+
+    VAddr GetCpuAddress() const {
+        return cpu_addr;
+    }
+
+    GPUVAddr GetGpuAddress() const {
+        return gpu_addr;
+    }
+
+    bool IsInside(const CacheAddr other_start, const CacheAddr other_end) const {
+        return (start <= other_start && other_end <= end);
+    }
+
+    bool operator==(const MapIntervalBase& rhs) const {
+        return std::tie(start, end) == std::tie(rhs.start, rhs.end);
+    }
+
+    bool operator!=(const MapIntervalBase& rhs) const {
+        return !operator==(rhs);
+    }
+
+    void MarkAsRegistered(const bool registered) {
+        is_registered = registered;
+    }
+
+    bool IsRegistered() const {
+        return is_registered;
+    }
+
+    CacheAddr GetStart() const {
+        return start;
+    }
+
+    CacheAddr GetEnd() const {
+        return end;
+    }
+
+    void MarkAsModified(const bool is_modified_, const u64 tick) {
+        is_modified = is_modified_;
+        ticks = tick;
+    }
+
+    bool IsModified() const {
+        return is_modified;
+    }
+
+    u64 GetModificationTick() const {
+        return ticks;
+    }
+
+    void MarkAsWritten(const bool is_written_) {
+        is_written = is_written_;
+    }
+
+    bool IsWritten() const {
+        return is_written;
+    }
+
+private:
+    CacheAddr start;
+    CacheAddr end;
+    GPUVAddr gpu_addr;
+    VAddr cpu_addr{};
+    bool is_written{};
+    bool is_modified{};
+    bool is_registered{};
+    u64 ticks{};
+};
+
+} // namespace VideoCommon
@@ -31,6 +31,7 @@ void DmaPusher::DispatchCalls() {
            break;
        }
    }
+    gpu.FlushCommands();
 }

 bool DmaPusher::Step() {
@@ -10,8 +10,7 @@

 namespace Tegra::Engines {

-Fermi2D::Fermi2D(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager)
-    : rasterizer{rasterizer}, memory_manager{memory_manager} {}
+Fermi2D::Fermi2D(VideoCore::RasterizerInterface& rasterizer) : rasterizer{rasterizer} {}

 void Fermi2D::CallMethod(const GPU::MethodCall& method_call) {
    ASSERT_MSG(method_call.method < Regs::NUM_REGS,
@@ -33,7 +33,7 @@ namespace Tegra::Engines {

 class Fermi2D final {
 public:
-    explicit Fermi2D(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager);
+    explicit Fermi2D(VideoCore::RasterizerInterface& rasterizer);
    ~Fermi2D() = default;

    /// Write the value to the register identified by method.
@@ -145,7 +145,6 @@ public:

 private:
    VideoCore::RasterizerInterface& rasterizer;
-    MemoryManager& memory_manager;

    /// Performs the copy from the source surface to the destination surface as configured in the
    /// registers.
@@ -15,7 +15,7 @@
 namespace Tegra::Engines {

 KeplerMemory::KeplerMemory(Core::System& system, MemoryManager& memory_manager)
-    : system{system}, memory_manager{memory_manager}, upload_state{memory_manager, regs.upload} {}
+    : system{system}, upload_state{memory_manager, regs.upload} {}

 KeplerMemory::~KeplerMemory() = default;

@@ -65,7 +65,6 @@ public:

 private:
    Core::System& system;
-    MemoryManager& memory_manager;
    Upload::State upload_state;
 };

@@ -249,16 +249,10 @@ void Maxwell3D::CallMacroMethod(u32 method, std::vector<u32> parameters) {
    executing_macro = 0;

    // Lookup the macro offset
-    const u32 entry{(method - MacroRegistersStart) >> 1};
-    const auto& search{macro_offsets.find(entry)};
-    if (search == macro_offsets.end()) {
-        LOG_CRITICAL(HW_GPU, "macro not found for method 0x{:X}!", method);
-        UNREACHABLE();
-        return;
-    }
+    const u32 entry = ((method - MacroRegistersStart) >> 1) % macro_positions.size();

    // Execute the current macro.
-    macro_interpreter.Execute(search->second, std::move(parameters));
+    macro_interpreter.Execute(macro_positions[entry], std::move(parameters));
 }

 void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) {
@@ -421,7 +415,7 @@ void Maxwell3D::ProcessMacroUpload(u32 data) {
 }

 void Maxwell3D::ProcessMacroBind(u32 data) {
-    macro_offsets[regs.macros.entry] = data;
+    macro_positions[regs.macros.entry++] = data;
 }

 void Maxwell3D::ProcessQueryGet() {
@@ -524,7 +518,7 @@ void Maxwell3D::ProcessQueryCondition() {
 void Maxwell3D::ProcessSyncPoint() {
    const u32 sync_point = regs.sync_info.sync_point.Value();
    const u32 increment = regs.sync_info.increment.Value();
-    const u32 cache_flush = regs.sync_info.unknown.Value();
+    [[maybe_unused]] const u32 cache_flush = regs.sync_info.unknown.Value();
    if (increment) {
        system.GPU().IncrementSyncPoint(sync_point);
    }
@@ -626,10 +620,10 @@ Texture::TICEntry Maxwell3D::GetTICEntry(u32 tic_index) const {
    Texture::TICEntry tic_entry;
    memory_manager.ReadBlockUnsafe(tic_address_gpu, &tic_entry, sizeof(Texture::TICEntry));

-    const auto r_type{tic_entry.r_type.Value()};
-    const auto g_type{tic_entry.g_type.Value()};
-    const auto b_type{tic_entry.b_type.Value()};
-    const auto a_type{tic_entry.a_type.Value()};
+    [[maybe_unused]] const auto r_type{tic_entry.r_type.Value()};
+    [[maybe_unused]] const auto g_type{tic_entry.g_type.Value()};
+    [[maybe_unused]] const auto b_type{tic_entry.b_type.Value()};
+    [[maybe_unused]] const auto a_type{tic_entry.a_type.Value()};

    // TODO(Subv): Different data types for separate components are not supported
    DEBUG_ASSERT(r_type == g_type && r_type == b_type && r_type == a_type);
@@ -1270,7 +1270,7 @@ private:
    MemoryManager& memory_manager;

    /// Start offsets of each macro in macro_memory
-    std::unordered_map<u32, u32> macro_offsets;
+    std::array<u32, 0x80> macro_positions = {};

    /// Memory for macro code
    MacroMemory macro_memory;
@@ -5,18 +5,17 @@
 #include "common/assert.h"
 #include "common/logging/log.h"
 #include "core/core.h"
+#include "core/settings.h"
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/engines/maxwell_dma.h"
 #include "video_core/memory_manager.h"
-#include "video_core/rasterizer_interface.h"
 #include "video_core/renderer_base.h"
 #include "video_core/textures/decoders.h"

 namespace Tegra::Engines {

-MaxwellDMA::MaxwellDMA(Core::System& system, VideoCore::RasterizerInterface& rasterizer,
-                       MemoryManager& memory_manager)
-    : system{system}, rasterizer{rasterizer}, memory_manager{memory_manager} {}
+MaxwellDMA::MaxwellDMA(Core::System& system, MemoryManager& memory_manager)
+    : system{system}, memory_manager{memory_manager} {}

 void MaxwellDMA::CallMethod(const GPU::MethodCall& method_call) {
    ASSERT_MSG(method_call.method < Regs::NUM_REGS,
@@ -84,13 +83,17 @@ void MaxwellDMA::HandleCopy() {
    ASSERT(regs.exec.enable_2d == 1);

    if (regs.exec.is_dst_linear && !regs.exec.is_src_linear) {
-        ASSERT(regs.src_params.size_z == 1);
+        ASSERT(regs.src_params.BlockDepth() == 0);
        // If the input is tiled and the output is linear, deswizzle the input and copy it over.
-        const u32 src_bytes_per_pixel = regs.src_pitch / regs.src_params.size_x;
+        const u32 bytes_per_pixel = regs.dst_pitch / regs.x_count;
        const std::size_t src_size = Texture::CalculateSize(
-            true, src_bytes_per_pixel, regs.src_params.size_x, regs.src_params.size_y,
+            true, bytes_per_pixel, regs.src_params.size_x, regs.src_params.size_y,
            regs.src_params.size_z, regs.src_params.BlockHeight(), regs.src_params.BlockDepth());

+        const std::size_t src_layer_size = Texture::CalculateSize(
+            true, bytes_per_pixel, regs.src_params.size_x, regs.src_params.size_y, 1,
+            regs.src_params.BlockHeight(), regs.src_params.BlockDepth());
+
        const std::size_t dst_size = regs.dst_pitch * regs.y_count;

        if (read_buffer.size() < src_size) {
@@ -104,23 +107,23 @@ void MaxwellDMA::HandleCopy() {
        memory_manager.ReadBlock(source, read_buffer.data(), src_size);
        memory_manager.ReadBlock(dest, write_buffer.data(), dst_size);

-        Texture::UnswizzleSubrect(regs.x_count, regs.y_count, regs.dst_pitch,
-                                  regs.src_params.size_x, src_bytes_per_pixel, read_buffer.data(),
-                                  write_buffer.data(), regs.src_params.BlockHeight(),
-                                  regs.src_params.pos_x, regs.src_params.pos_y);
+        Texture::UnswizzleSubrect(
+            regs.x_count, regs.y_count, regs.dst_pitch, regs.src_params.size_x, bytes_per_pixel,
+            read_buffer.data() + src_layer_size * regs.src_params.pos_z, write_buffer.data(),
+            regs.src_params.BlockHeight(), regs.src_params.pos_x, regs.src_params.pos_y);

        memory_manager.WriteBlock(dest, write_buffer.data(), dst_size);
    } else {
        ASSERT(regs.dst_params.BlockDepth() == 0);

-        const u32 src_bytes_per_pixel = regs.src_pitch / regs.x_count;
+        const u32 bytes_per_pixel = regs.src_pitch / regs.x_count;

        const std::size_t dst_size = Texture::CalculateSize(
-            true, src_bytes_per_pixel, regs.dst_params.size_x, regs.dst_params.size_y,
+            true, bytes_per_pixel, regs.dst_params.size_x, regs.dst_params.size_y,
            regs.dst_params.size_z, regs.dst_params.BlockHeight(), regs.dst_params.BlockDepth());

        const std::size_t dst_layer_size = Texture::CalculateSize(
-            true, src_bytes_per_pixel, regs.dst_params.size_x, regs.dst_params.size_y, 1,
+            true, bytes_per_pixel, regs.dst_params.size_x, regs.dst_params.size_y, 1,
            regs.dst_params.BlockHeight(), regs.dst_params.BlockDepth());

        const std::size_t src_size = regs.src_pitch * regs.y_count;
@@ -133,14 +136,19 @@ void MaxwellDMA::HandleCopy() {
            write_buffer.resize(dst_size);
        }

-        memory_manager.ReadBlock(source, read_buffer.data(), src_size);
-        memory_manager.ReadBlock(dest, write_buffer.data(), dst_size);
+        if (Settings::values.use_accurate_gpu_emulation) {
+            memory_manager.ReadBlock(source, read_buffer.data(), src_size);
+            memory_manager.ReadBlock(dest, write_buffer.data(), dst_size);
+        } else {
+            memory_manager.ReadBlockUnsafe(source, read_buffer.data(), src_size);
+            memory_manager.ReadBlockUnsafe(dest, write_buffer.data(), dst_size);
+        }

        // If the input is linear and the output is tiled, swizzle the input and copy it over.
-        Texture::SwizzleSubrect(regs.x_count, regs.y_count, regs.src_pitch, regs.dst_params.size_x,
-                                src_bytes_per_pixel,
-                                write_buffer.data() + dst_layer_size * regs.dst_params.pos_z,
-                                read_buffer.data(), regs.dst_params.BlockHeight());
+        Texture::SwizzleSubrect(
+            regs.x_count, regs.y_count, regs.src_pitch, regs.dst_params.size_x, bytes_per_pixel,
+            write_buffer.data() + dst_layer_size * regs.dst_params.pos_z, read_buffer.data(),
+            regs.dst_params.BlockHeight(), regs.dst_params.pos_x, regs.dst_params.pos_y);

        memory_manager.WriteBlock(dest, write_buffer.data(), dst_size);
    }
@@ -20,10 +20,6 @@ namespace Tegra {
 class MemoryManager;
 }

-namespace VideoCore {
-class RasterizerInterface;
-}
-
 namespace Tegra::Engines {

 /**
@@ -33,8 +29,7 @@ namespace Tegra::Engines {

 class MaxwellDMA final {
 public:
-    explicit MaxwellDMA(Core::System& system, VideoCore::RasterizerInterface& rasterizer,
-                        MemoryManager& memory_manager);
+    explicit MaxwellDMA(Core::System& system, MemoryManager& memory_manager);
    ~MaxwellDMA() = default;

    /// Write the value to the register identified by method.
@@ -180,8 +175,6 @@ public:
 private:
    Core::System& system;

-    VideoCore::RasterizerInterface& rasterizer;
-
    MemoryManager& memory_manager;

    std::vector<u8> read_buffer;
@@ -538,6 +538,12 @@ enum class PhysicalAttributeDirection : u64 {
    Output = 1,
 };

+enum class VoteOperation : u64 {
+    All = 0, // allThreadsNV
+    Any = 1, // anyThreadNV
+    Eq = 2,  // allThreadsEqualNV
+};
+
 union Instruction {
    Instruction& operator=(const Instruction& instr) {
        value = instr.value;
@@ -564,6 +570,13 @@ union Instruction {
        BitField<13, 1, u64> trigger;
    } nop;

+    union {
+        BitField<48, 2, VoteOperation> operation;
+        BitField<45, 3, u64> dest_pred;
+        BitField<39, 3, u64> value;
+        BitField<42, 1, u64> negate_value;
+    } vote;
+
    union {
        BitField<8, 8, Register> gpr;
        BitField<20, 24, s64> offset;
@@ -873,6 +886,7 @@ union Instruction {
    union {
        BitField<0, 3, u64> pred0;
        BitField<3, 3, u64> pred3;
+        BitField<6, 1, u64> neg_b;
        BitField<7, 1, u64> abs_a;
        BitField<39, 3, u64> pred39;
        BitField<42, 1, u64> neg_pred;
@@ -1006,7 +1020,6 @@ union Instruction {
    } iset;

    union {
-        BitField<41, 2, u64> selector; // i2i and i2f only
        BitField<45, 1, u64> negate_a;
        BitField<49, 1, u64> abs_a;
        BitField<10, 2, Register::Size> src_size;
@@ -1023,8 +1036,6 @@ union Instruction {
        } f2i;

        union {
-            BitField<8, 2, Register::Size> src_size;
-            BitField<10, 2, Register::Size> dst_size;
            BitField<39, 4, u64> rounding;
            // H0, H1 extract for F16 missing
            BitField<41, 1, u64> selector; // Guessed as some games set it, TODO: reverse this value
@@ -1034,6 +1045,13 @@ union Instruction {
            }
        } f2f;

+        union {
+            BitField<41, 2, u64> selector;
+        } int_src;
+
+        union {
+            BitField<41, 1, u64> selector;
+        } float_src;
    } conversion;

    union {
@@ -1489,6 +1507,7 @@ public:
        SYNC,
        BRK,
        DEPBAR,
+        VOTE,
        BFE_C,
        BFE_R,
        BFE_IMM,
@@ -1651,6 +1670,7 @@ public:
        Hfma2,
        Flow,
        Synch,
+        Warp,
        Memory,
        Texture,
        Image,
@@ -1777,6 +1797,7 @@ private:
            INST("111000110100---", Id::BRK, Type::Flow, "BRK"),
            INST("111000110000----", Id::EXIT, Type::Flow, "EXIT"),
            INST("1111000011110---", Id::DEPBAR, Type::Synch, "DEPBAR"),
+            INST("0101000011011---", Id::VOTE, Type::Warp, "VOTE"),
            INST("1110111111011---", Id::LD_A, Type::Memory, "LD_A"),
            INST("1110111101001---", Id::LD_S, Type::Memory, "LD_S"),
            INST("1110111101000---", Id::LD_L, Type::Memory, "LD_L"),
@@ -17,27 +17,15 @@

 namespace Tegra {

-u32 FramebufferConfig::BytesPerPixel(PixelFormat format) {
-    switch (format) {
-    case PixelFormat::ABGR8:
-    case PixelFormat::BGRA8:
-        return 4;
-    default:
-        return 4;
-    }
-
-    UNREACHABLE();
-}
-
 GPU::GPU(Core::System& system, VideoCore::RendererBase& renderer, bool is_async)
    : system{system}, renderer{renderer}, is_async{is_async} {
    auto& rasterizer{renderer.Rasterizer()};
    memory_manager = std::make_unique<Tegra::MemoryManager>(system, rasterizer);
    dma_pusher = std::make_unique<Tegra::DmaPusher>(*this);
    maxwell_3d = std::make_unique<Engines::Maxwell3D>(system, rasterizer, *memory_manager);
-    fermi_2d = std::make_unique<Engines::Fermi2D>(rasterizer, *memory_manager);
+    fermi_2d = std::make_unique<Engines::Fermi2D>(rasterizer);
    kepler_compute = std::make_unique<Engines::KeplerCompute>(system, rasterizer, *memory_manager);
-    maxwell_dma = std::make_unique<Engines::MaxwellDMA>(system, rasterizer, *memory_manager);
+    maxwell_dma = std::make_unique<Engines::MaxwellDMA>(system, *memory_manager);
    kepler_memory = std::make_unique<Engines::KeplerMemory>(system, *memory_manager);
 }

@@ -120,6 +108,10 @@ bool GPU::CancelSyncptInterrupt(const u32 syncpoint_id, const u32 value) {
    return true;
 }

+void GPU::FlushCommands() {
+    renderer.Rasterizer().FlushCommands();
+}
+
 u32 RenderTargetBytesPerPixel(RenderTargetFormat format) {
    ASSERT(format != RenderTargetFormat::NONE);

@@ -19,6 +19,10 @@ inline CacheAddr ToCacheAddr(const void* host_ptr) {
    return reinterpret_cast<CacheAddr>(host_ptr);
 }

+inline u8* FromCacheAddr(CacheAddr cache_addr) {
+    return reinterpret_cast<u8*>(cache_addr);
+}
+
 namespace Core {
 class System;
 }
@@ -91,14 +95,10 @@ class DebugContext;
 struct FramebufferConfig {
    enum class PixelFormat : u32 {
        ABGR8 = 1,
+        RGB565 = 4,
        BGRA8 = 5,
    };

-    /**
-     * Returns the number of bytes per pixel.
-     */
-    static u32 BytesPerPixel(PixelFormat format);
-
    VAddr address;
    u32 offset;
    u32 width;
@@ -153,6 +153,8 @@ public:
    /// Calls a GPU method.
    void CallMethod(const MethodCall& method_call);

+    void FlushCommands();
+
    /// Returns a reference to the Maxwell3D GPU engine.
    Engines::Maxwell3D& Maxwell3D();

@@ -247,8 +249,7 @@ public:
    virtual void PushGPUEntries(Tegra::CommandList&& entries) = 0;

    /// Swap buffers (render frame)
-    virtual void SwapBuffers(
-        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) = 0;
+    virtual void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) = 0;

    /// Notify rasterizer that any caches of the specified region should be flushed to Switch memory
    virtual void FlushRegion(CacheAddr addr, u64 size) = 0;
@@ -279,8 +280,8 @@ private:

 protected:
    std::unique_ptr<Tegra::DmaPusher> dma_pusher;
-    VideoCore::RendererBase& renderer;
    Core::System& system;
+    VideoCore::RendererBase& renderer;

 private:
    std::unique_ptr<Tegra::MemoryManager> memory_manager;
@@ -23,9 +23,8 @@ void GPUAsynch::PushGPUEntries(Tegra::CommandList&& entries) {
    gpu_thread.SubmitList(std::move(entries));
 }

-void GPUAsynch::SwapBuffers(
-    std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) {
-    gpu_thread.SwapBuffers(std::move(framebuffer));
+void GPUAsynch::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) {
+    gpu_thread.SwapBuffers(framebuffer);
 }

 void GPUAsynch::FlushRegion(CacheAddr addr, u64 size) {
@@ -14,15 +14,14 @@ class RendererBase;
 namespace VideoCommon {

 /// Implementation of GPU interface that runs the GPU asynchronously
-class GPUAsynch : public Tegra::GPU {
+class GPUAsynch final : public Tegra::GPU {
 public:
    explicit GPUAsynch(Core::System& system, VideoCore::RendererBase& renderer);
    ~GPUAsynch() override;

    void Start() override;
    void PushGPUEntries(Tegra::CommandList&& entries) override;
-    void SwapBuffers(
-        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) override;
+    void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) override;
    void FlushRegion(CacheAddr addr, u64 size) override;
    void InvalidateRegion(CacheAddr addr, u64 size) override;
    void FlushAndInvalidateRegion(CacheAddr addr, u64 size) override;
@@ -19,9 +19,8 @@ void GPUSynch::PushGPUEntries(Tegra::CommandList&& entries) {
    dma_pusher->DispatchCalls();
 }

-void GPUSynch::SwapBuffers(
-    std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) {
-    renderer.SwapBuffers(std::move(framebuffer));
+void GPUSynch::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) {
+    renderer.SwapBuffers(framebuffer);
 }

 void GPUSynch::FlushRegion(CacheAddr addr, u64 size) {
@@ -13,15 +13,14 @@ class RendererBase;
 namespace VideoCommon {

 /// Implementation of GPU interface that runs the GPU synchronously
-class GPUSynch : public Tegra::GPU {
+class GPUSynch final : public Tegra::GPU {
 public:
    explicit GPUSynch(Core::System& system, VideoCore::RendererBase& renderer);
    ~GPUSynch() override;

    void Start() override;
    void PushGPUEntries(Tegra::CommandList&& entries) override;
-    void SwapBuffers(
-        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) override;
+    void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) override;
    void FlushRegion(CacheAddr addr, u64 size) override;
    void InvalidateRegion(CacheAddr addr, u64 size) override;
    void FlushAndInvalidateRegion(CacheAddr addr, u64 size) override;
@@ -39,7 +39,7 @@ static void RunThread(VideoCore::RendererBase& renderer, Tegra::DmaPusher& dma_p
                dma_pusher.Push(std::move(submit_list->entries));
                dma_pusher.DispatchCalls();
            } else if (const auto data = std::get_if<SwapBuffersCommand>(&next.data)) {
-                renderer.SwapBuffers(std::move(data->framebuffer));
+                renderer.SwapBuffers(data->framebuffer ? &*data->framebuffer : nullptr);
            } else if (const auto data = std::get_if<FlushRegionCommand>(&next.data)) {
                renderer.Rasterizer().FlushRegion(data->addr, data->size);
            } else if (const auto data = std::get_if<InvalidateRegionCommand>(&next.data)) {
@@ -78,9 +78,9 @@ void ThreadManager::SubmitList(Tegra::CommandList&& entries) {
    system.CoreTiming().ScheduleEvent(synchronization_ticks, synchronization_event, fence);
 }

-void ThreadManager::SwapBuffers(
-    std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) {
-    PushCommand(SwapBuffersCommand(std::move(framebuffer)));
+void ThreadManager::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) {
+    PushCommand(SwapBuffersCommand(framebuffer ? *framebuffer
+                                               : std::optional<const Tegra::FramebufferConfig>{}));
 }

 void ThreadManager::FlushRegion(CacheAddr addr, u64 size) {
@@ -110,8 +110,7 @@ public:
    void SubmitList(Tegra::CommandList&& entries);

    /// Swap buffers (render frame)
-    void SwapBuffers(
-        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer);
+    void SwapBuffers(const Tegra::FramebufferConfig* framebuffer);

    /// Notify rasterizer that any caches of the specified region should be flushed to Switch memory
    void FlushRegion(CacheAddr addr, u64 size);
@@ -25,8 +25,8 @@ static void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth

    // With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual
    // pixel values.
-    const u32 tile_size_x{GetDefaultBlockWidth(format)};
-    const u32 tile_size_y{GetDefaultBlockHeight(format)};
+    constexpr u32 tile_size_x{GetDefaultBlockWidth(format)};
+    constexpr u32 tile_size_y{GetDefaultBlockHeight(format)};

    if constexpr (morton_to_linear) {
        Tegra::Texture::UnswizzleTexture(buffer, addr, tile_size_x, tile_size_y, bytes_per_pixel,
@@ -186,99 +186,6 @@ static MortonCopyFn GetSwizzleFunction(MortonSwizzleMode mode, Surface::PixelFor
    return morton_to_linear_fns[static_cast<std::size_t>(format)];
 }

-static u32 MortonInterleave128(u32 x, u32 y) {
-    // 128x128 Z-Order coordinate from 2D coordinates
-    static constexpr u32 xlut[] = {
-        0x0000, 0x0001, 0x0002, 0x0003, 0x0008, 0x0009, 0x000a, 0x000b, 0x0040, 0x0041, 0x0042,
-        0x0043, 0x0048, 0x0049, 0x004a, 0x004b, 0x0800, 0x0801, 0x0802, 0x0803, 0x0808, 0x0809,
-        0x080a, 0x080b, 0x0840, 0x0841, 0x0842, 0x0843, 0x0848, 0x0849, 0x084a, 0x084b, 0x1000,
-        0x1001, 0x1002, 0x1003, 0x1008, 0x1009, 0x100a, 0x100b, 0x1040, 0x1041, 0x1042, 0x1043,
-        0x1048, 0x1049, 0x104a, 0x104b, 0x1800, 0x1801, 0x1802, 0x1803, 0x1808, 0x1809, 0x180a,
-        0x180b, 0x1840, 0x1841, 0x1842, 0x1843, 0x1848, 0x1849, 0x184a, 0x184b, 0x2000, 0x2001,
-        0x2002, 0x2003, 0x2008, 0x2009, 0x200a, 0x200b, 0x2040, 0x2041, 0x2042, 0x2043, 0x2048,
-        0x2049, 0x204a, 0x204b, 0x2800, 0x2801, 0x2802, 0x2803, 0x2808, 0x2809, 0x280a, 0x280b,
-        0x2840, 0x2841, 0x2842, 0x2843, 0x2848, 0x2849, 0x284a, 0x284b, 0x3000, 0x3001, 0x3002,
-        0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x3040, 0x3041, 0x3042, 0x3043, 0x3048, 0x3049,
-        0x304a, 0x304b, 0x3800, 0x3801, 0x3802, 0x3803, 0x3808, 0x3809, 0x380a, 0x380b, 0x3840,
-        0x3841, 0x3842, 0x3843, 0x3848, 0x3849, 0x384a, 0x384b, 0x0000, 0x0001, 0x0002, 0x0003,
-        0x0008, 0x0009, 0x000a, 0x000b, 0x0040, 0x0041, 0x0042, 0x0043, 0x0048, 0x0049, 0x004a,
-        0x004b, 0x0800, 0x0801, 0x0802, 0x0803, 0x0808, 0x0809, 0x080a, 0x080b, 0x0840, 0x0841,
-        0x0842, 0x0843, 0x0848, 0x0849, 0x084a, 0x084b, 0x1000, 0x1001, 0x1002, 0x1003, 0x1008,
-        0x1009, 0x100a, 0x100b, 0x1040, 0x1041, 0x1042, 0x1043, 0x1048, 0x1049, 0x104a, 0x104b,
-        0x1800, 0x1801, 0x1802, 0x1803, 0x1808, 0x1809, 0x180a, 0x180b, 0x1840, 0x1841, 0x1842,
-        0x1843, 0x1848, 0x1849, 0x184a, 0x184b, 0x2000, 0x2001, 0x2002, 0x2003, 0x2008, 0x2009,
-        0x200a, 0x200b, 0x2040, 0x2041, 0x2042, 0x2043, 0x2048, 0x2049, 0x204a, 0x204b, 0x2800,
-        0x2801, 0x2802, 0x2803, 0x2808, 0x2809, 0x280a, 0x280b, 0x2840, 0x2841, 0x2842, 0x2843,
-        0x2848, 0x2849, 0x284a, 0x284b, 0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a,
-        0x300b, 0x3040, 0x3041, 0x3042, 0x3043, 0x3048, 0x3049, 0x304a, 0x304b, 0x3800, 0x3801,
-        0x3802, 0x3803, 0x3808, 0x3809, 0x380a, 0x380b, 0x3840, 0x3841, 0x3842, 0x3843, 0x3848,
-        0x3849, 0x384a, 0x384b, 0x0000, 0x0001, 0x0002, 0x0003, 0x0008, 0x0009, 0x000a, 0x000b,
-        0x0040, 0x0041, 0x0042, 0x0043, 0x0048, 0x0049, 0x004a, 0x004b, 0x0800, 0x0801, 0x0802,
-        0x0803, 0x0808, 0x0809, 0x080a, 0x080b, 0x0840, 0x0841, 0x0842, 0x0843, 0x0848, 0x0849,
-        0x084a, 0x084b, 0x1000, 0x1001, 0x1002, 0x1003, 0x1008, 0x1009, 0x100a, 0x100b, 0x1040,
-        0x1041, 0x1042, 0x1043, 0x1048, 0x1049, 0x104a, 0x104b, 0x1800, 0x1801, 0x1802, 0x1803,
-        0x1808, 0x1809, 0x180a, 0x180b, 0x1840, 0x1841, 0x1842, 0x1843, 0x1848, 0x1849, 0x184a,
-        0x184b, 0x2000, 0x2001, 0x2002, 0x2003, 0x2008, 0x2009, 0x200a, 0x200b, 0x2040, 0x2041,
-        0x2042, 0x2043, 0x2048, 0x2049, 0x204a, 0x204b, 0x2800, 0x2801, 0x2802, 0x2803, 0x2808,
-        0x2809, 0x280a, 0x280b, 0x2840, 0x2841, 0x2842, 0x2843, 0x2848, 0x2849, 0x284a, 0x284b,
-        0x3000, 0x3001, 0x3002, 0x3003, 0x3008, 0x3009, 0x300a, 0x300b, 0x3040, 0x3041, 0x3042,
-        0x3043, 0x3048, 0x3049, 0x304a, 0x304b, 0x3800, 0x3801, 0x3802, 0x3803, 0x3808, 0x3809,
-        0x380a, 0x380b, 0x3840, 0x3841, 0x3842, 0x3843, 0x3848, 0x3849, 0x384a, 0x384b,
-    };
-    static constexpr u32 ylut[] = {
-        0x0000, 0x0004, 0x0010, 0x0014, 0x0020, 0x0024, 0x0030, 0x0034, 0x0080, 0x0084, 0x0090,
-        0x0094, 0x00a0, 0x00a4, 0x00b0, 0x00b4, 0x0100, 0x0104, 0x0110, 0x0114, 0x0120, 0x0124,
-        0x0130, 0x0134, 0x0180, 0x0184, 0x0190, 0x0194, 0x01a0, 0x01a4, 0x01b0, 0x01b4, 0x0200,
-        0x0204, 0x0210, 0x0214, 0x0220, 0x0224, 0x0230, 0x0234, 0x0280, 0x0284, 0x0290, 0x0294,
-        0x02a0, 0x02a4, 0x02b0, 0x02b4, 0x0300, 0x0304, 0x0310, 0x0314, 0x0320, 0x0324, 0x0330,
-        0x0334, 0x0380, 0x0384, 0x0390, 0x0394, 0x03a0, 0x03a4, 0x03b0, 0x03b4, 0x0400, 0x0404,
-        0x0410, 0x0414, 0x0420, 0x0424, 0x0430, 0x0434, 0x0480, 0x0484, 0x0490, 0x0494, 0x04a0,
-        0x04a4, 0x04b0, 0x04b4, 0x0500, 0x0504, 0x0510, 0x0514, 0x0520, 0x0524, 0x0530, 0x0534,
-        0x0580, 0x0584, 0x0590, 0x0594, 0x05a0, 0x05a4, 0x05b0, 0x05b4, 0x0600, 0x0604, 0x0610,
-        0x0614, 0x0620, 0x0624, 0x0630, 0x0634, 0x0680, 0x0684, 0x0690, 0x0694, 0x06a0, 0x06a4,
-        0x06b0, 0x06b4, 0x0700, 0x0704, 0x0710, 0x0714, 0x0720, 0x0724, 0x0730, 0x0734, 0x0780,
-        0x0784, 0x0790, 0x0794, 0x07a0, 0x07a4, 0x07b0, 0x07b4, 0x0000, 0x0004, 0x0010, 0x0014,
-        0x0020, 0x0024, 0x0030, 0x0034, 0x0080, 0x0084, 0x0090, 0x0094, 0x00a0, 0x00a4, 0x00b0,
-        0x00b4, 0x0100, 0x0104, 0x0110, 0x0114, 0x0120, 0x0124, 0x0130, 0x0134, 0x0180, 0x0184,
-        0x0190, 0x0194, 0x01a0, 0x01a4, 0x01b0, 0x01b4, 0x0200, 0x0204, 0x0210, 0x0214, 0x0220,
-        0x0224, 0x0230, 0x0234, 0x0280, 0x0284, 0x0290, 0x0294, 0x02a0, 0x02a4, 0x02b0, 0x02b4,
-        0x0300, 0x0304, 0x0310, 0x0314, 0x0320, 0x0324, 0x0330, 0x0334, 0x0380, 0x0384, 0x0390,
-        0x0394, 0x03a0, 0x03a4, 0x03b0, 0x03b4, 0x0400, 0x0404, 0x0410, 0x0414, 0x0420, 0x0424,
-        0x0430, 0x0434, 0x0480, 0x0484, 0x0490, 0x0494, 0x04a0, 0x04a4, 0x04b0, 0x04b4, 0x0500,
-        0x0504, 0x0510, 0x0514, 0x0520, 0x0524, 0x0530, 0x0534, 0x0580, 0x0584, 0x0590, 0x0594,
-        0x05a0, 0x05a4, 0x05b0, 0x05b4, 0x0600, 0x0604, 0x0610, 0x0614, 0x0620, 0x0624, 0x0630,
-        0x0634, 0x0680, 0x0684, 0x0690, 0x0694, 0x06a0, 0x06a4, 0x06b0, 0x06b4, 0x0700, 0x0704,
-        0x0710, 0x0714, 0x0720, 0x0724, 0x0730, 0x0734, 0x0780, 0x0784, 0x0790, 0x0794, 0x07a0,
-        0x07a4, 0x07b0, 0x07b4, 0x0000, 0x0004, 0x0010, 0x0014, 0x0020, 0x0024, 0x0030, 0x0034,
-        0x0080, 0x0084, 0x0090, 0x0094, 0x00a0, 0x00a4, 0x00b0, 0x00b4, 0x0100, 0x0104, 0x0110,
-        0x0114, 0x0120, 0x0124, 0x0130, 0x0134, 0x0180, 0x0184, 0x0190, 0x0194, 0x01a0, 0x01a4,
-        0x01b0, 0x01b4, 0x0200, 0x0204, 0x0210, 0x0214, 0x0220, 0x0224, 0x0230, 0x0234, 0x0280,
-        0x0284, 0x0290, 0x0294, 0x02a0, 0x02a4, 0x02b0, 0x02b4, 0x0300, 0x0304, 0x0310, 0x0314,
-        0x0320, 0x0324, 0x0330, 0x0334, 0x0380, 0x0384, 0x0390, 0x0394, 0x03a0, 0x03a4, 0x03b0,
-        0x03b4, 0x0400, 0x0404, 0x0410, 0x0414, 0x0420, 0x0424, 0x0430, 0x0434, 0x0480, 0x0484,
-        0x0490, 0x0494, 0x04a0, 0x04a4, 0x04b0, 0x04b4, 0x0500, 0x0504, 0x0510, 0x0514, 0x0520,
-        0x0524, 0x0530, 0x0534, 0x0580, 0x0584, 0x0590, 0x0594, 0x05a0, 0x05a4, 0x05b0, 0x05b4,
-        0x0600, 0x0604, 0x0610, 0x0614, 0x0620, 0x0624, 0x0630, 0x0634, 0x0680, 0x0684, 0x0690,
-        0x0694, 0x06a0, 0x06a4, 0x06b0, 0x06b4, 0x0700, 0x0704, 0x0710, 0x0714, 0x0720, 0x0724,
-        0x0730, 0x0734, 0x0780, 0x0784, 0x0790, 0x0794, 0x07a0, 0x07a4, 0x07b0, 0x07b4,
-    };
-    return xlut[x % 128] + ylut[y % 128];
-}
-
-static u32 GetMortonOffset128(u32 x, u32 y, u32 bytes_per_pixel) {
-    // Calculates the offset of the position of the pixel in Morton order
-    // Framebuffer images are split into 128x128 tiles.
-
-    constexpr u32 block_height = 128;
-    const u32 coarse_x = x & ~127;
-
-    const u32 i = MortonInterleave128(x, y);
-
-    const u32 offset = coarse_x * block_height;
-
-    return (i + offset) * bytes_per_pixel;
-}
-
 void MortonSwizzle(MortonSwizzleMode mode, Surface::PixelFormat format, u32 stride,
                   u32 block_height, u32 height, u32 block_depth, u32 depth, u32 tile_width_spacing,
                   u8* buffer, u8* addr) {
@@ -286,23 +193,4 @@ void MortonSwizzle(MortonSwizzleMode mode, Surface::PixelFormat format, u32 stri
                                     tile_width_spacing, buffer, addr);
 }

-void MortonCopyPixels128(MortonSwizzleMode mode, u32 width, u32 height, u32 bytes_per_pixel,
-                         u32 linear_bytes_per_pixel, u8* morton_data, u8* linear_data) {
-    const bool morton_to_linear = mode == MortonSwizzleMode::MortonToLinear;
-    u8* data_ptrs[2];
-    for (u32 y = 0; y < height; ++y) {
-        for (u32 x = 0; x < width; ++x) {
-            const u32 coarse_y = y & ~127;
-            const u32 morton_offset =
-                GetMortonOffset128(x, y, bytes_per_pixel) + coarse_y * width * bytes_per_pixel;
-            const u32 linear_pixel_index = (x + y * width) * linear_bytes_per_pixel;
-
-            data_ptrs[morton_to_linear ? 1 : 0] = morton_data + morton_offset;
-            data_ptrs[morton_to_linear ? 0 : 1] = &linear_data[linear_pixel_index];
-
-            std::memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel);
-        }
-    }
-}
-
 } // namespace VideoCore