diff --git a/app/CMakeLists.txt b/app/CMakeLists.txt
index 7c558cd1..83d06668 100644
--- a/app/CMakeLists.txt
+++ b/app/CMakeLists.txt
@@ -159,6 +159,7 @@ add_library(skyline SHARED
         ${source_DIR}/skyline/gpu/texture_manager.cpp
         ${source_DIR}/skyline/gpu/command_scheduler.cpp
         ${source_DIR}/skyline/gpu/texture/texture.cpp
+        ${source_DIR}/skyline/gpu/buffer.cpp
         ${source_DIR}/skyline/gpu/presentation_engine.cpp
         ${source_DIR}/skyline/gpu/shader_manager.cpp
         ${source_DIR}/skyline/gpu/interconnect/command_executor.cpp
diff --git a/app/src/main/cpp/skyline/gpu/buffer.cpp b/app/src/main/cpp/skyline/gpu/buffer.cpp
new file mode 100644
index 00000000..fc85a345
--- /dev/null
+++ b/app/src/main/cpp/skyline/gpu/buffer.cpp
@@ -0,0 +1,138 @@
+// SPDX-License-Identifier: MPL-2.0
+// Copyright © 2021 Skyline Team and Contributors (https://github.com/skyline-emu/)
+
+#include <gpu.h>
+#include <common/trace.h>
+#include "buffer.h"
+
+namespace skyline::gpu {
+    vk::DeviceSize GuestBuffer::BufferSize() const {
+        vk::DeviceSize size{};
+        for (const auto &buffer : mappings)
+            size += buffer.size_bytes();
+        return size;
+    }
+
+    Buffer::Buffer(GPU &gpu, GuestBuffer guest) : size(guest.BufferSize()), backing(gpu.memory.AllocateBuffer(size)), guest(std::move(guest)) {
+        SynchronizeHost();
+    }
+
+    void Buffer::WaitOnFence() {
+        TRACE_EVENT("gpu", "Buffer::WaitOnFence");
+
+        auto lCycle{cycle.lock()};
+        if (lCycle) {
+            lCycle->Wait();
+            cycle.reset();
+        }
+    }
+
+    void Buffer::SynchronizeHost() {
+        WaitOnFence();
+
+        TRACE_EVENT("gpu", "Buffer::SynchronizeHost");
+
+        auto host{backing.data()};
+        for (auto &mapping : guest.mappings) {
+            auto mappingSize{mapping.size_bytes()};
+            std::memcpy(host, mapping.data(), mappingSize);
+            host += mappingSize;
+        }
+    }
+
+    void Buffer::SynchronizeHostWithCycle(const std::shared_ptr<FenceCycle> &pCycle) {
+        if (pCycle != cycle.lock())
+            WaitOnFence();
+
+        TRACE_EVENT("gpu", "Buffer::SynchronizeHostWithCycle");
+
+        auto host{backing.data()};
+        for (auto &mapping : guest.mappings) {
+            auto mappingSize{mapping.size_bytes()};
+            std::memcpy(host, mapping.data(), mappingSize);
+            host += mappingSize;
+        }
+    }
+
+    void Buffer::SynchronizeGuest() {
+        WaitOnFence();
+
+        TRACE_EVENT("gpu", "Buffer::SynchronizeGuest");
+
+        auto host{backing.data()};
+        for (auto &mapping : guest.mappings) {
+            auto mappingSize{mapping.size_bytes()};
+            std::memcpy(mapping.data(), host, mappingSize);
+            host += mappingSize;
+        }
+    }
+
+    /**
+     * @brief A FenceCycleDependency that synchronizes the contents of a host buffer with the guest buffer
+     */
+    struct BufferGuestSync : public FenceCycleDependency {
+        std::shared_ptr<Buffer> buffer;
+
+        explicit BufferGuestSync(std::shared_ptr<Buffer> buffer) : buffer(std::move(buffer)) {}
+
+        ~BufferGuestSync() {
+            TRACE_EVENT("gpu", "Buffer::BufferGuestSync");
+            buffer->SynchronizeGuest();
+        }
+    };
+
+    void Buffer::SynchronizeGuestWithCycle(const std::shared_ptr<FenceCycle> &pCycle) {
+        if (pCycle != cycle.lock())
+            WaitOnFence();
+
+        pCycle->AttachObject(std::make_shared<BufferGuestSync>(shared_from_this()));
+        cycle = pCycle;
+    }
+
+    std::shared_ptr<BufferView> Buffer::GetView(vk::DeviceSize offset, vk::DeviceSize range, vk::Format format) {
+        for (const auto &viewWeak : views) {
+            auto view{viewWeak.lock()};
+            if (view && view->offset == offset && view->range == range && view->format == format)
+                return view;
+        }
+
+        auto view{std::make_shared<BufferView>(shared_from_this(), offset, range, format)};
+        views.push_back(view);
+        return view;
+    }
+
+    BufferView::BufferView(std::shared_ptr<Buffer> backing, vk::DeviceSize offset, vk::DeviceSize range, vk::Format format) : buffer(std::move(backing)), offset(offset), range(range), format(format) {}
+
+    void BufferView::lock() {
+        auto currentBacking{std::atomic_load(&buffer)};
+        while (true) {
+            currentBacking->lock();
+
+            auto newBacking{std::atomic_load(&buffer)};
+            if (currentBacking == newBacking)
+                return;
+
+            currentBacking->unlock();
+            currentBacking = newBacking;
+        }
+    }
+
+    void BufferView::unlock() {
+        buffer->unlock();
+    }
+
+    bool BufferView::try_lock() {
+        auto currentBacking{std::atomic_load(&buffer)};
+        while (true) {
+            bool success{currentBacking->try_lock()};
+
+            auto newBacking{std::atomic_load(&buffer)};
+            if (currentBacking == newBacking)
+                return success;
+
+            if (success)
+                currentBacking->unlock();
+            currentBacking = newBacking;
+        }
+    }
+}
diff --git a/app/src/main/cpp/skyline/gpu/buffer.h b/app/src/main/cpp/skyline/gpu/buffer.h
new file mode 100644
index 00000000..983adb6b
--- /dev/null
+++ b/app/src/main/cpp/skyline/gpu/buffer.h
@@ -0,0 +1,145 @@
+// SPDX-License-Identifier: MPL-2.0
+// Copyright © 2021 Skyline Team and Contributors (https://github.com/skyline-emu/)
+
+#pragma once
+
+#include "memory_manager.h"
+
+namespace skyline::gpu {
+    /**
+     * @brief A descriptor for a GPU buffer on the guest
+     */
+    struct GuestBuffer {
+        using Mappings = boost::container::small_vector<span<u8>, 3>;
+        Mappings mappings; //!< Spans to CPU memory for the underlying data backing this buffer
+        vk::Format format;
+
+        /**
+         * @return The total size of the buffer by adding up the size of all mappings
+         */
+        vk::DeviceSize BufferSize() const;
+    };
+
+    struct BufferView;
+
+    /**
+     * @brief A buffer which is backed by host constructs while being synchronized with the underlying guest buffer
+     * @note This class conforms to the Lockable and BasicLockable C++ named requirements
+     */
+    class Buffer : public std::enable_shared_from_this<Buffer>, public FenceCycleDependency {
+      private:
+        std::mutex mutex; //!< Synchronizes any mutations to the buffer or its backing
+        vk::DeviceSize size;
+        memory::Buffer backing;
+        GuestBuffer guest;
+
+        std::vector<std::weak_ptr<BufferView>> views; //!< BufferView(s) that are backed by this Buffer, used for repointing to a new Buffer on deletion
+
+        friend BufferView;
+
+      public:
+        std::weak_ptr<FenceCycle> cycle; //!< A fence cycle for when any host operation mutating the buffer has completed, it must be waited on prior to any mutations to the backing
+
+        constexpr vk::Buffer GetBacking() {
+            return backing.vkBuffer;
+        }
+
+        Buffer(GPU &gpu, GuestBuffer guest);
+
+        /**
+         * @brief Acquires an exclusive lock on the texture for the calling thread
+         * @note Naming is in accordance to the BasicLockable named requirement
+         */
+        void lock() {
+            mutex.lock();
+        }
+
+        /**
+         * @brief Relinquishes an existing lock on the texture by the calling thread
+         * @note Naming is in accordance to the BasicLockable named requirement
+         */
+        void unlock() {
+            mutex.unlock();
+        }
+
+        /**
+         * @brief Attempts to acquire an exclusive lock but returns immediately if it's captured by another thread
+         * @note Naming is in accordance to the Lockable named requirement
+         */
+        bool try_lock() {
+            return mutex.try_lock();
+        }
+
+        /**
+         * @brief Waits on a fence cycle if it exists till it's signalled and resets it after
+         * @note The buffer **must** be locked prior to calling this
+         */
+        void WaitOnFence();
+
+        /**
+         * @brief Synchronizes the host buffer with the guest
+         * @note The buffer **must** be locked prior to calling this
+         */
+        void SynchronizeHost();
+
+        /**
+         * @brief Synchronizes the host buffer with the guest
+         * @param cycle A FenceCycle that is checked against the held one to skip waiting on it when equal
+         * @note The buffer **must** be locked prior to calling this
+         */
+        void SynchronizeHostWithCycle(const std::shared_ptr<FenceCycle> &cycle);
+
+        /**
+         * @brief Synchronizes the guest buffer with the host buffer
+         * @note The buffer **must** be locked prior to calling this
+         */
+        void SynchronizeGuest();
+
+        /**
+         * @brief Synchronizes the guest buffer with the host buffer when the FenceCycle is signalled
+         * @note The buffer **must** be locked prior to calling this
+         * @note The guest buffer should not be null prior to calling this
+         */
+        void SynchronizeGuestWithCycle(const std::shared_ptr<FenceCycle> &cycle);
+
+        /**
+         * @return A cached or newly created view into this buffer with the supplied attributes
+         */
+        std::shared_ptr<BufferView> GetView(vk::DeviceSize offset, vk::DeviceSize range, vk::Format format);
+    };
+
+    /**
+     * @brief A contiguous view into a Vulkan Buffer that represents a single guest buffer (as opposed to Buffer objects which contain multiple)
+     * @note The object **must** be locked prior to accessing any members as values will be mutated
+     * @note This class conforms to the Lockable and BasicLockable C++ named requirements
+     */
+    struct BufferView {
+        std::shared_ptr<Buffer> buffer;
+        vk::DeviceSize offset;
+        vk::DeviceSize range;
+        vk::Format format;
+
+        /**
+         * @note A view must **NOT** be constructed directly, it should always be retrieved using Texture::GetView
+         */
+        BufferView(std::shared_ptr<Buffer> backing, vk::DeviceSize offset, vk::DeviceSize range, vk::Format format);
+
+        /**
+         * @brief Acquires an exclusive lock on the buffer for the calling thread
+         * @note Naming is in accordance to the BasicLockable named requirement
+         */
+        void lock();
+
+        /**
+         * @brief Relinquishes an existing lock on the buffer by the calling thread
+         * @note Naming is in accordance to the BasicLockable named requirement
+         */
+        void unlock();
+
+        /**
+         * @brief Attempts to acquire an exclusive lock but returns immediately if it's captured by another thread
+         * @note Naming is in accordance to the Lockable named requirement
+         */
+        bool try_lock();
+    };
+}