Re: nfs MAP_SHARED corruption fix

[Trond Myklebust <trond.myklebust@fys.uio.no>]

>>>>> " " == Andrea Arcangeli <andrea@suse.de> writes:

     > On Tue, May 08, 2001 at 05:21:02PM +0200, Trond Myklebust
     > wrote:
    >> AFAICs this fix will clearly deadlock...

     > yeah, it didn't triggered because it probably needs to be the
     > same page writepaged and in the dirty list at the same time. I
     > hooked it very deep into the writeback logic to keep it generic
     > (it wasn't going to add a significant overhead) but it didn't
     > need to be _that_ deep.

     > Even worse I think it was partly wrong because it was only in
     > the close(2) path but not in the fput path that is the one
     > walked by munmap.

     > This looks better to me, what do you think?

Just 2 comments.

 - You should use nfs_wb_all() here rather than nfs_wb_file() since
writepage() unfortunately can't initialize the NFS writeback structure
with the correct vma->vm_file.

 - Are we allowed to protect the filemap_fdatasync() + nfs_wb_all() by
grabbing the inode->i_sem in nfs_file_close_vma()? If so, this should
be done as well.

 Otherwise, that patch looked fine to me... I'll test it out...

Cheers,
   Trond