[lkml]   [2006]   [Dec]   [20]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    Patch in this message
    SubjectRe: [PATCH] mm: fix page_mkclean_one (was: 2.6.19 file content corruption on ext3)
    On Wed, Dec 20, 2006 at 11:50:50AM -0800, Linus Torvalds wrote:
    > On Wed, 20 Dec 2006, Linus Torvalds wrote:
    > >
    > > So that's why I've been harping on the fact that I think we simply do
    > > really wrong things with PG_dirty at times [ ... ]
    > Ok, I'll just put my money where my mouth is, and suggest a patch like
    > THIS instead.
    > This one clears up all the issues I find irritating:
    > - "test_clear_page_dirty()" is insane, both conceptually and as an
    > implementation. "Give me a 'C', give me an 'R', give me an 'A', give me
    > a 'P'".
    > So rip out that mindfart entirely.
    > - "clear_page_dirty()" is badly named, and should be about CANCELLING the
    > dirty bit, and must never be called with pages mapped anyway. So throw
    > that out too, and replace it with a new function:
    > void cancel_dirty_page(struct page *page, unsigned int accounting_size);
    > - "clear_page_dirty_for_io()" is fine.
    > And with that, I then either rip out any old users of
    > "test_clear_page_dirty()" or "clear_page_dirty()", and if appropriate (and
    > it's realy lonly appropriate for "truncate()", I replace them with the new
    > "cancel_dirty_page()". Most of the time, they should just be deleted
    > entirely.
    > NOTE NOTE NOTE! I _only_ did enough to make things compile for my
    > particular configuration. That means that right now the following
    > filesystems are broken with this patch (because they use the totally
    > broken old crap):
    > CIFS, FUSE, JFS, ReiserFS, XFS

    XFS appears to call clear_page_dirty to get the mapping tree dirty
    tag set correctly at the same time the page dirty flag is cleared. I
    note that this can be done by set_page_writeback() if we clear the
    dirty flag on the page first when we are writing back the entire page.

    Hence it seems to me that the XFS call to clear_page_dirty() could
    easily be substituted by clear_page_dirty_for_io() followed by a
    call to set_page_writeback() to get the mapping tree tags set
    correctly after the page has been marked clean.

    Does this make sense (even without the posted patch)?

    fs/xfs/linux-2.6/xfs_aops.c | 4 ++--
    1 file changed, 2 insertions(+), 2 deletions(-)

    Index: 2.6.x-xfs-new/fs/xfs/linux-2.6/xfs_aops.c
    --- 2.6.x-xfs-new.orig/fs/xfs/linux-2.6/xfs_aops.c 2006-12-19 12:22:47.000000000 +1100
    +++ 2.6.x-xfs-new/fs/xfs/linux-2.6/xfs_aops.c 2006-12-21 10:15:04.545375877 +1100
    @@ -340,9 +340,9 @@ xfs_start_page_writeback(
    - set_page_writeback(page);
    if (clear_dirty)
    - clear_page_dirty(page);
    + clear_page_dirty_for_io(page);
    + set_page_writeback(page);
    if (!buffers) {

    Dave Chinner
    Principal Engineer
    SGI Australian Software Group
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to
    More majordomo info at
    Please read the FAQ at

     \ /
      Last update: 2006-12-21 00:29    [W:0.023 / U:0.480 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site