Kernel/Ext2FS: Don't hog FS lock when calling base class flush_writes()

Once we've finalized all the file system metadata in flush_writes(),
we no longer need to hold the file system lock during the call to
BlockBasedFileSystem::flush_writes().
This commit is contained in:
Andreas Kling 2021-07-16 02:16:00 +02:00
parent 98c230b370
commit abbd237ec1
Notes: sideshowbarker 2024-07-18 08:57:03 +09:00

View file

@ -688,51 +688,53 @@ void Ext2FS::flush_block_group_descriptor_table()
void Ext2FS::flush_writes()
{
Locker locker(m_lock);
if (m_super_block_dirty) {
flush_super_block();
m_super_block_dirty = false;
}
if (m_block_group_descriptors_dirty) {
flush_block_group_descriptor_table();
m_block_group_descriptors_dirty = false;
}
for (auto& cached_bitmap : m_cached_bitmaps) {
if (cached_bitmap->dirty) {
auto buffer = UserOrKernelBuffer::for_kernel_buffer(cached_bitmap->buffer.data());
if (auto result = write_block(cached_bitmap->bitmap_block_index, buffer, block_size()); result.is_error()) {
dbgln("Ext2FS[{}]::flush_writes(): Failed to write blocks: {}", fsid(), result.error());
}
cached_bitmap->dirty = false;
dbgln_if(EXT2_DEBUG, "Ext2FS[{}]::flush_writes(): Flushed bitmap block {}", fsid(), cached_bitmap->bitmap_block_index);
{
Locker locker(m_lock);
if (m_super_block_dirty) {
flush_super_block();
m_super_block_dirty = false;
}
if (m_block_group_descriptors_dirty) {
flush_block_group_descriptor_table();
m_block_group_descriptors_dirty = false;
}
for (auto& cached_bitmap : m_cached_bitmaps) {
if (cached_bitmap->dirty) {
auto buffer = UserOrKernelBuffer::for_kernel_buffer(cached_bitmap->buffer.data());
if (auto result = write_block(cached_bitmap->bitmap_block_index, buffer, block_size()); result.is_error()) {
dbgln("Ext2FS[{}]::flush_writes(): Failed to write blocks: {}", fsid(), result.error());
}
cached_bitmap->dirty = false;
dbgln_if(EXT2_DEBUG, "Ext2FS[{}]::flush_writes(): Flushed bitmap block {}", fsid(), cached_bitmap->bitmap_block_index);
}
}
// Uncache Inodes that are only kept alive by the index-to-inode lookup cache.
// We don't uncache Inodes that are being watched by at least one InodeWatcher.
// FIXME: It would be better to keep a capped number of Inodes around.
// The problem is that they are quite heavy objects, and use a lot of heap memory
// for their (child name lookup) and (block list) caches.
Vector<InodeIndex> unused_inodes;
for (auto& it : m_inode_cache) {
// NOTE: If we're asked to look up an inode by number (via get_inode) and it turns out
// to not exist, we remember the fact that it doesn't exist by caching a nullptr.
// This seems like a reasonable time to uncache ideas about unknown inodes, so do that.
if (!it.value) {
unused_inodes.append(it.key);
continue;
}
if (it.value->ref_count() != 1)
continue;
if (it.value->has_watchers())
continue;
unused_inodes.append(it.key);
}
for (auto index : unused_inodes)
uncache_inode(index);
}
BlockBasedFileSystem::flush_writes();
// Uncache Inodes that are only kept alive by the index-to-inode lookup cache.
// We don't uncache Inodes that are being watched by at least one InodeWatcher.
// FIXME: It would be better to keep a capped number of Inodes around.
// The problem is that they are quite heavy objects, and use a lot of heap memory
// for their (child name lookup) and (block list) caches.
Vector<InodeIndex> unused_inodes;
for (auto& it : m_inode_cache) {
// NOTE: If we're asked to look up an inode by number (via get_inode) and it turns out
// to not exist, we remember the fact that it doesn't exist by caching a nullptr.
// This seems like a reasonable time to uncache ideas about unknown inodes, so do that.
if (!it.value) {
unused_inodes.append(it.key);
continue;
}
if (it.value->ref_count() != 1)
continue;
if (it.value->has_watchers())
continue;
unused_inodes.append(it.key);
}
for (auto index : unused_inodes)
uncache_inode(index);
}
Ext2FSInode::Ext2FSInode(Ext2FS& fs, InodeIndex index)