Linux Kernel Architecture

Chapter 17: Data Synchronization

The methods provided forfile_operations->fsyncare very similar for most filesystems. Figure 17-13

shows the code flow diagram for a generalized method.

{fs}_sync_inode

{fs}_sync_file

sync_mapping_buffers

fsync_buffers_list

ll_rw_block

osync_buffers_list

datasync set and Inode

in state l_DIRTY_DATASYNC?

Figure 17-13: Code flow diagram forf_op->fsync.

The code performs two tasks:

1. sync_mapping_bufferswrites back all private inode buffers in theprivate_listof the
   mappinginstance. These normally hold indirection blocks or other internal filesystem data
   that arenotpart of the inode management data but are used to manage the data themselves.
   This function delegates work tofsync_mapping_buffers, which iterates over all buffers. The
   buffer data are written to the block layer by thell_rw_blockfunction with which you are
   familiar from Chapter 6. With the help ofosync_buffers_list, the kernel then waits until
   the write operations have been completed (the block layer also buffers write accesses) and
   then ensures that synchronization of the associated metadata outsidesync_buffers_list
   appears as a synchronous operation.


2. fs_sync_inodewrites back the inode management data (i.e., the data held directly in the
   filesystem-specific inode structure). Note that thedatasyncargument offsyncmust be
   set to (0) to invoke the method. This is the one and only difference betweenfdatasyncand
   fsync.

Since writeback of inode management data is filesystem-specific, see Chapter 9.

17.15.3 Synchronization of Memory Mappings

The kernel provides themsyncsystem call implemented insys_msyncto synchronize memory mappings

in part or in total:

mm/msync.c

asmlinkage long sys_msync(unsigned long start, size_t len, int flags)

startandlenselect an area in the user address space of a process whose mapped data are to be syn-

chronized with the underlying file.