lkml.org 
[lkml]   [2009]   [Jan]   [7]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH mmotm] nilfs2: insert checks and hole block allocation in page_mkwrite
From
Date
On Thu, 2009-01-08 at 02:33 +0900, Ryusuke Konishi wrote:
> On Wed, 07 Jan 2009 09:43:02 -0500, Chris Mason wrote:
> > On Wed, 2009-01-07 at 22:06 +0900, Ryusuke Konishi wrote:
> > > Chris Mason told me that page_mkwrite() has some works to do.
> > >
> > > On Wed, 17 Dec 2008 21:52:55 -0500, Chris Mason wrote:
> > > > nilfs_page_mkwrite doesn't seem to dirty the page? block_page_mkwrite
> > > > does more, including checks against i_size and others.
> > >
> > > This will insert i_size check, and hole block allocation code in the
> > > nilfs_page_mkwrite.
> > >
> > > Previously, the hole block allocation was delayed until just before
> > > writing for mmapped pages. This accompanies removal of the delayed
> > > allocation code.
> <snip>
> > > static int nilfs_page_mkwrite(struct vm_area_struct *vma, struct page *page)
> > > {
> > > - if (!(vma->vm_flags & (VM_WRITE | VM_MAYWRITE)))
> > > - return -EPERM;
> > > + struct inode *inode = vma->vm_file->f_dentry->d_inode;
> > > + struct address_space *mapping = inode->i_mapping;
> > > + struct nilfs_transaction_info ti;
> > > + struct buffer_head *bh, *head;
> > > + int fully_mapped = 1;
> > > + int ret = -EINVAL;
> > > +
> > > + /*
> > > + * use i_alloc_sem to stop truncate operations on the inode
> > > + */
> > > + down_read(&inode->i_alloc_sem);
> >
> > I'm not 100% sure this is safe. It seems likely the direct io paths
> > could trigger a mkwrite with the i_alloc_sem already held?
>
> vmtruncate() can change i_size outside the page lock, and it even
> calls unmap_mapping_range(). Does it have any problems?
>

You're right, i_size can change without you knowing about it, but before
vmtruncate does anything to the page itself, it locks the page.

So, we tend to do a somewhat racey check against i_size. If the page is
outside i_size, we happily ignore it. If it is inside i_size, we assume
it is part of the file. Most of the code that does this checks i_size
once after locking the page and saves that value, so if i_size changes
later on the code at least does consistent operations based on a single
value of i_size.

If the page straddles i_size and someone extends the file, they are
expected to lock the page as well (see block_truncate_page and the
places it gets called).

-chris




\
 
 \ /
  Last update: 2009-01-07 18:51    [W:0.054 / U:22.816 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site