Kernel/Ext2FS: Cache blocks less aggressively

With this patch, we now only cache at most 2^14 blocks, which greatly
relieves memory pressure when working with large files.

Kernel/FileSystem/Ext2FS/BlockView.cpp

@@ -0,0 +1,82 @@
+/*
+ * Copyright (c) 2024, the SerenityOS developers.
+ *
+ * SPDX-License-Identifier: BSD-2-Clause
+ */
+
+#include <Kernel/FileSystem/Ext2FS/BlockView.h>
+#include <Kernel/FileSystem/Ext2FS/Inode.h>
+
+namespace Kernel {
+
+static constexpr size_t max_blocks_in_view = 16384; // 2^14
+
+Ext2FSBlockView::Ext2FSBlockView(Ext2FSInode& inode)
+    : m_inode(inode) {};
+
+ErrorOr<void> Ext2FSBlockView::ensure_block(BlockBasedFileSystem::BlockIndex block)
+{
+    VERIFY(m_block_list_lock.is_locked());
+    if (block >= m_first_block && block <= m_last_block && m_block_list_initialized.was_set())
+        return {};
+
+    BlockBasedFileSystem::BlockIndex new_first_block = (block.value() / max_blocks_in_view) * max_blocks_in_view;
+    BlockBasedFileSystem::BlockIndex new_last_block = new_first_block.value() + max_blocks_in_view - 1;
+
+    m_block_list = TRY(m_inode.compute_block_list(new_first_block, new_last_block));
+
+    m_first_block = new_first_block;
+    m_last_block = new_last_block;
+
+    m_block_list_initialized.set();
+
+    return {};
+}
+
+ErrorOr<BlockBasedFileSystem::BlockIndex> Ext2FSBlockView::get_block(BlockBasedFileSystem::BlockIndex block)
+{
+    MutexLocker block_list_locker(m_block_list_lock);
+    TRY(ensure_block(block));
+
+    auto it = m_block_list.find(block);
+    if (it == m_block_list.end())
+        return BlockBasedFileSystem::BlockIndex { 0 };
+
+    auto on_disk_block = (*it).value;
+    VERIFY(on_disk_block != 0);
+    return on_disk_block;
+}
+
+ErrorOr<BlockBasedFileSystem::BlockIndex> Ext2FSBlockView::get_or_allocate_block(BlockBasedFileSystem::BlockIndex block, bool zero_newly_allocated_block, bool allow_cache)
+{
+    MutexLocker block_list_locker(m_block_list_lock);
+    TRY(ensure_block(block));
+
+    auto it = m_block_list.find(block);
+    if (it != m_block_list.end()) {
+        auto on_disk_block = (*it).value;
+        VERIFY(on_disk_block != 0);
+        return on_disk_block;
+    }
+
+    auto on_disk_block = TRY(m_inode.allocate_block(block, zero_newly_allocated_block, allow_cache));
+    TRY(m_block_list.try_set(block, on_disk_block));
+
+    return on_disk_block;
+}
+
+ErrorOr<void> Ext2FSBlockView::write_block_pointer(BlockBasedFileSystem::BlockIndex logical_block_index, BlockBasedFileSystem::BlockIndex on_disk_index)
+{
+    MutexLocker block_list_locker(m_block_list_lock);
+
+    TRY(m_inode.write_block_pointer(logical_block_index, on_disk_index));
+
+    if (on_disk_index == 0)
+        m_block_list.remove(logical_block_index);
+    else
+        TRY(m_block_list.try_set(logical_block_index, on_disk_index));
+
+    return {};
+}
+
+}