Re: [patch 6/8] fuse: clean up setting i_size in write

From: Miklos Szeredi
Date: Tue Mar 18 2008 - 04:17:09 EST


> On Mon, 17 Mar 2008 20:19:14 +0100 Miklos Szeredi <miklos@xxxxxxxxxx> wrote:
>
> > From: Miklos Szeredi <mszeredi@xxxxxxx>
> >
> > Extract common code for setting i_size in write functions into a
> > common helper.
> >
> > Signed-off-by: Miklos Szeredi <mszeredi@xxxxxxx>
> > ---
> > fs/fuse/file.c | 28 +++++++++++++++-------------
> > 1 file changed, 15 insertions(+), 13 deletions(-)
> >
> > Index: linux/fs/fuse/file.c
> > ===================================================================
> > --- linux.orig/fs/fuse/file.c 2008-03-17 18:26:04.000000000 +0100
> > +++ linux/fs/fuse/file.c 2008-03-17 18:26:28.000000000 +0100
> > @@ -610,13 +610,24 @@ static int fuse_write_begin(struct file
> > return 0;
> > }
> >
> > +static void fuse_write_update_size(struct inode *inode, loff_t pos)
> > +{
> > + struct fuse_conn *fc = get_fuse_conn(inode);
> > + struct fuse_inode *fi = get_fuse_inode(inode);
> > +
> > + spin_lock(&fc->lock);
> > + fi->attr_version = ++fc->attr_version;
> > + if (pos > inode->i_size)
> > + i_size_write(inode, pos);
> > + spin_unlock(&fc->lock);
> > +}
> >
> > ...
> >
> > @@ -766,12 +772,8 @@ static ssize_t fuse_direct_io(struct fil
> > }
> > fuse_put_request(fc, req);
> > if (res > 0) {
> > - if (write) {
> > - spin_lock(&fc->lock);
> > - if (pos > inode->i_size)
> > - i_size_write(inode, pos);
> > - spin_unlock(&fc->lock);
> > - }
> > + if (write)
> > + fuse_write_update_size(inode, pos);
>
> We require that i_mutex be held here, to prevent i_size_write() deadlocks.
> Is it held?
>

No, fuse uses the per connection spinlock to protect against
concurrent calls to i_size_write(), because in some cases holding
i_mutex would be difficult.

Fuse already got painfully bitten by the i_size_write() deadlock, so
I'm well aware of the problem :)

Miklos
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/