From 922fd703c9c750fcd09a32f82e8f5280726663fe Mon Sep 17 00:00:00 2001 From: Andreas Kling Date: Mon, 30 Sep 2019 11:20:51 +0200 Subject: [PATCH] Kernel: Convert the DiskBackedFS write API to take "const u8*" This way clients are not required to have instantiated ByteBuffers and can choose whatever memory scheme works best for them. Also converted some of the Ext2FS code to use stack buffers instead. --- Kernel/FileSystem/DiskBackedFileSystem.cpp | 10 ++--- Kernel/FileSystem/DiskBackedFileSystem.h | 4 +- Kernel/FileSystem/Ext2FileSystem.cpp | 45 +++++++++++----------- Kernel/FileSystem/Ext2FileSystem.h | 2 +- 4 files changed, 29 insertions(+), 32 deletions(-) diff --git a/Kernel/FileSystem/DiskBackedFileSystem.cpp b/Kernel/FileSystem/DiskBackedFileSystem.cpp index 9f1385fa57..c8c3be33a6 100644 --- a/Kernel/FileSystem/DiskBackedFileSystem.cpp +++ b/Kernel/FileSystem/DiskBackedFileSystem.cpp @@ -81,15 +81,13 @@ DiskBackedFS::~DiskBackedFS() { } -bool DiskBackedFS::write_block(unsigned index, const ByteBuffer& data) +bool DiskBackedFS::write_block(unsigned index, const u8* data) { #ifdef DBFS_DEBUG kprintf("DiskBackedFileSystem::write_block %u, size=%u\n", index, data.size()); #endif - ASSERT(data.size() == block_size()); - auto& entry = cache().get(index); - memcpy(entry.data, data.data(), data.size()); + memcpy(entry.data, data, block_size()); entry.is_dirty = true; entry.has_data = true; @@ -97,13 +95,13 @@ bool DiskBackedFS::write_block(unsigned index, const ByteBuffer& data) return true; } -bool DiskBackedFS::write_blocks(unsigned index, unsigned count, const ByteBuffer& data) +bool DiskBackedFS::write_blocks(unsigned index, unsigned count, const u8* data) { #ifdef DBFS_DEBUG kprintf("DiskBackedFileSystem::write_blocks %u x%u\n", index, count); #endif for (unsigned i = 0; i < count; ++i) - write_block(index + i, data.slice(i * block_size(), block_size())); + write_block(index + i, data + i * block_size()); return true; } diff --git a/Kernel/FileSystem/DiskBackedFileSystem.h b/Kernel/FileSystem/DiskBackedFileSystem.h index 44096e0762..bfd797ffb6 100644 --- a/Kernel/FileSystem/DiskBackedFileSystem.h +++ b/Kernel/FileSystem/DiskBackedFileSystem.h @@ -22,8 +22,8 @@ protected: bool read_block(unsigned index, u8* buffer) const; bool read_blocks(unsigned index, unsigned count, u8* buffer) const; - bool write_block(unsigned index, const ByteBuffer&); - bool write_blocks(unsigned index, unsigned count, const ByteBuffer&); + bool write_block(unsigned index, const u8*); + bool write_blocks(unsigned index, unsigned count, const u8*); private: DiskCache& cache() const; diff --git a/Kernel/FileSystem/Ext2FileSystem.cpp b/Kernel/FileSystem/Ext2FileSystem.cpp index d7c317b7ab..b21414a117 100644 --- a/Kernel/FileSystem/Ext2FileSystem.cpp +++ b/Kernel/FileSystem/Ext2FileSystem.cpp @@ -10,6 +10,7 @@ //#define EXT2_DEBUG +static const size_t max_block_size = 4096; static const ssize_t max_inline_symlink_length = 60; static u8 to_ext2_file_type(mode_t mode) @@ -121,8 +122,11 @@ bool Ext2FS::initialize() kprintf("ext2fs: desc size = %u\n", EXT2_DESC_SIZE(&super_block)); #endif + set_block_size(EXT2_BLOCK_SIZE(&super_block)); + ASSERT(block_size() <= (int)max_block_size); + m_block_group_count = ceil_div(super_block.s_blocks_count, super_block.s_blocks_per_group); if (m_block_group_count == 0) { @@ -157,16 +161,16 @@ InodeIdentifier Ext2FS::root_inode() const return { fsid(), EXT2_ROOT_INO }; } -ByteBuffer Ext2FS::read_block_containing_inode(unsigned inode, unsigned& block_index, unsigned& offset) const +bool Ext2FS::read_block_containing_inode(unsigned inode, unsigned& block_index, unsigned& offset, u8* buffer) const { LOCKER(m_lock); auto& super_block = this->super_block(); if (inode != EXT2_ROOT_INO && inode < EXT2_FIRST_INO(&super_block)) - return {}; + return false; if (inode > super_block.s_inodes_count) - return {}; + return false; auto& bgd = group_descriptor(group_index_from_inode(inode)); @@ -174,10 +178,7 @@ ByteBuffer Ext2FS::read_block_containing_inode(unsigned inode, unsigned& block_i block_index = bgd.bg_inode_table + (offset >> EXT2_BLOCK_SIZE_BITS(&super_block)); offset &= block_size() - 1; - auto buffer = ByteBuffer::create_uninitialized(block_size()); - if (!read_block(block_index, buffer.data())) - return {}; - return buffer; + return read_block(block_index, buffer); } Ext2FS::BlockListShape Ext2FS::compute_block_list_shape(unsigned blocks) @@ -290,7 +291,7 @@ bool Ext2FS::write_block_list_for_inode(InodeIndex inode_index, ext2_inode& e2in --remaining_blocks; } stream.fill_to_end(0); - bool success = write_block(e2inode.i_block[EXT2_IND_BLOCK], block_contents); + bool success = write_block(e2inode.i_block[EXT2_IND_BLOCK], block_contents.data()); ASSERT(success); } @@ -373,7 +374,7 @@ bool Ext2FS::write_block_list_for_inode(InodeIndex inode_index, ext2_inode& e2in } if (ind_block_dirty) { - bool success = write_block(indirect_block_index, ind_block_contents); + bool success = write_block(indirect_block_index, ind_block_contents.data()); ASSERT(success); } } @@ -385,7 +386,7 @@ bool Ext2FS::write_block_list_for_inode(InodeIndex inode_index, ext2_inode& e2in } if (dind_block_dirty) { - bool success = write_block(e2inode.i_block[EXT2_DIND_BLOCK], dind_block_contents); + bool success = write_block(e2inode.i_block[EXT2_DIND_BLOCK], dind_block_contents.data()); ASSERT(success); } } @@ -509,7 +510,7 @@ void Ext2FS::flush_block_group_descriptor_table() LOCKER(m_lock); unsigned blocks_to_write = ceil_div(m_block_group_count * (unsigned)sizeof(ext2_group_desc), block_size()); unsigned first_block_of_bgdt = block_size() == 1024 ? 2 : 1; - write_blocks(first_block_of_bgdt, blocks_to_write, m_cached_group_descriptor_table); + write_blocks(first_block_of_bgdt, blocks_to_write, m_cached_group_descriptor_table.data()); } Ext2FSInode::Ext2FSInode(Ext2FS& fs, unsigned index) @@ -585,15 +586,15 @@ RefPtr Ext2FS::get_inode(InodeIdentifier inode) const unsigned block_index; unsigned offset; - auto block = read_block_containing_inode(inode.index(), block_index, offset); - if (!block) + u8 block[max_block_size]; + if (!read_block_containing_inode(inode.index(), block_index, offset, block)) return {}; auto it = m_inode_cache.find(inode.index()); if (it != m_inode_cache.end()) return (*it).value; auto new_inode = adopt(*new Ext2FSInode(const_cast(*this), inode.index())); - memcpy(&new_inode->m_raw_inode, reinterpret_cast(block.offset_pointer(offset)), sizeof(ext2_inode)); + memcpy(&new_inode->m_raw_inode, reinterpret_cast(block + offset), sizeof(ext2_inode)); m_inode_cache.set(inode.index(), new_inode); return new_inode; } @@ -644,9 +645,7 @@ ssize_t Ext2FSInode::read_bytes(off_t offset, ssize_t count, u8* buffer, FileDes //kprintf("ok let's do it, read(%u, %u) -> blocks %u thru %u, oifb: %u\n", offset, count, first_block_logical_index, last_block_logical_index, offset_into_first_block); #endif - // Biggest block size should be 4096.. - u8 block[4096]; - ASSERT(block_size < (int)sizeof(block)); + u8 block[max_block_size]; for (int bi = first_block_logical_index; remaining_count && bi <= last_block_logical_index; ++bi) { bool success = fs().read_block(m_block_list[bi], block); @@ -779,7 +778,7 @@ ssize_t Ext2FSInode::write_bytes(off_t offset, ssize_t count, const u8* data, Fi #ifdef EXT2_DEBUG dbgprintf("Ext2FSInode::write_bytes: writing block %u (offset_into_block: %u)\n", m_block_list[bi], offset_into_block); #endif - bool success = fs().write_block(m_block_list[bi], block); + bool success = fs().write_block(m_block_list[bi], block.data()); if (!success) { kprintf("Ext2FSInode::write_bytes: write_block(%u) failed (lbi: %u)\n", m_block_list[bi], bi); ASSERT_NOT_REACHED(); @@ -985,10 +984,10 @@ bool Ext2FS::write_ext2_inode(unsigned inode, const ext2_inode& e2inode) LOCKER(m_lock); unsigned block_index; unsigned offset; - auto block = read_block_containing_inode(inode, block_index, offset); - if (!block) + u8 block[max_block_size]; + if (!read_block_containing_inode(inode, block_index, offset, block)) return false; - memcpy(reinterpret_cast(block.offset_pointer(offset)), &e2inode, inode_size()); + memcpy(reinterpret_cast(block + offset), &e2inode, inode_size()); bool success = write_block(block_index, block); ASSERT(success); return success; @@ -1177,7 +1176,7 @@ bool Ext2FS::set_inode_allocation_state(InodeIndex inode_index, bool new_state) } bitmap.set(bit_index, new_state); - success = write_block(bgd.bg_inode_bitmap, block); + success = write_block(bgd.bg_inode_bitmap, block.data()); ASSERT(success); // Update superblock @@ -1241,7 +1240,7 @@ bool Ext2FS::set_block_allocation_state(BlockIndex block_index, bool new_state) } bitmap.set(bit_index, new_state); - success = write_block(bgd.bg_block_bitmap, block); + success = write_block(bgd.bg_block_bitmap, block.data()); ASSERT(success); // Update superblock diff --git a/Kernel/FileSystem/Ext2FileSystem.h b/Kernel/FileSystem/Ext2FileSystem.h index 3c43f55680..60d6950e82 100644 --- a/Kernel/FileSystem/Ext2FileSystem.h +++ b/Kernel/FileSystem/Ext2FileSystem.h @@ -87,7 +87,7 @@ private: unsigned inode_size() const; bool write_ext2_inode(InodeIndex, const ext2_inode&); - ByteBuffer read_block_containing_inode(InodeIndex inode, BlockIndex& block_index, unsigned& offset) const; + bool read_block_containing_inode(InodeIndex inode, BlockIndex& block_index, unsigned& offset, u8* buffer) const; ByteBuffer read_super_block() const; bool write_super_block(const ext2_super_block&);