On Thu, 8 Feb 2007 10:29:45 +0100 Jan Kara [off-list ref] wrote:

On Wed 07-02-07 12:56:59, Andrew Morton wrote:

quoted

On Wed, 7 Feb 2007 13:46:57 -0700
Andreas Dilger [off-list ref] wrote:

quoted

On Feb 06, 2007  17:35 -0800, Andrew Morton wrote:

quoted

On Mon, 5 Feb 2007 14:12:04 +0100
Jan Kara [off-list ref] wrote:

quoted

quoted

Move the blocks on the temporary inode to the original inode
by a page.
1. Read the file data from the old blocks to the page
2. Move the block on the temporary inode to the original inode
3. Write the file data on the page into the new blocks

  I have one thing - it's probably not good to use page cache for
defragmentation.

Then it is no longer online defragmentation.  The issues with maintaining
correctness and coherency with ongoing VFS activity would be truly ghastly.

If we're worried about pagecache pollution then it would be better to control
that from userspace via fadvise().

It should be possible to have the online defrag tool lock the inode against
any changes,

Sounds easy when you say it fast.  But how do we "lock" against, say, a
read pagefault?  Only by writing back then removing the pagecache page then
reinstantiating it as a locked, not-uptodate page and then removing it from
pagecache afterwards prior to unlocking it.  Or something.

I don't think we want to go there.

  I though Andreas meant "any write changes" - i.e. you check that noone
has open file descriptor for writing and block any new open for writing.
That can be done quite easily.
  Anyway, I agree with you that userspace solution to a possible page
cache pollution is preferable after thinking about it for a while.
As I've been thinking about it, we could actually do the copying
from user space. We could do something like:
  block any writes to file (as I described above)
  craft new inode with blocks allocated as we want (using preallocation,
    we should mostly have the kernel infrastructure we need)
  copy data using splice syscall
  call the kernel to switch data

I don't think we need to block any writes to any file or anything.

To move a page within a file:

	fd = open(file);
	p = mmap(fd);
	the_page_was_in_core = mincore(p, offset);
	munmap(p);
	ioctl(fd, ..., new_block);

			<kernel>
			read_cache_page(inode, offset);
			lock_page(page);
			if (try_to_free_buffers(page)) {
				<relocate the page>
				set_page_dirty(page);
			}
			unlock_page(page);

	if (the_page_was_in_core) {
		sync_file_range(fd, offset SYNC_FILE_RANGE_WAIT_BEFORE|
						SYNC_FILE_RANGE_WRITE|
						SYNC_FILE_RANGE_WAIT_AFTER);
		fadvise(fd, offset, FADV_DONTNEED);
	}

completely coherent with pagecache, quite safe in the presence of mmap,
mlock, O_DIRECT, everything else.  Also fully journallable in-kernel.