lkml.org 
[lkml]   [2021]   [Nov]   [15]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Date
SubjectRe: [PATCH 5.13 089/151] ovl: fix deadlock in splice write
Hi Greg,

Looks like upstream commit b91b6b019fd ("ovl: fix deadlock in splice
write") needs to be added to linux-5.4.y.

The reason is that commit 82a763e61e2b ("ovl: simplify file splice")
was backported to v5.4.155, and the above commit fixes this.

Applies cleanly and I reviewed that the backport is correct.

Thanks,
Miklos

On Mon, Aug 16, 2021 at 3:13 PM Greg Kroah-Hartman
<gregkh@linuxfoundation.org> wrote:
>
> From: Miklos Szeredi <mszeredi@redhat.com>
>
> [ Upstream commit 9b91b6b019fda817eb52f728eb9c79b3579760bc ]
>
> There's possibility of an ABBA deadlock in case of a splice write to an
> overlayfs file and a concurrent splice write to a corresponding real file.
>
> The call chain for splice to an overlay file:
>
> -> do_splice [takes sb_writers on overlay file]
> -> do_splice_from
> -> iter_file_splice_write [takes pipe->mutex]
> -> vfs_iter_write
> ...
> -> ovl_write_iter [takes sb_writers on real file]
>
> And the call chain for splice to a real file:
>
> -> do_splice [takes sb_writers on real file]
> -> do_splice_from
> -> iter_file_splice_write [takes pipe->mutex]
>
> Syzbot successfully bisected this to commit 82a763e61e2b ("ovl: simplify
> file splice").
>
> Fix by reverting the write part of the above commit and by adding missing
> bits from ovl_write_iter() into ovl_splice_write().
>
> Fixes: 82a763e61e2b ("ovl: simplify file splice")
> Reported-and-tested-by: syzbot+579885d1a9a833336209@syzkaller.appspotmail.com
> Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
> fs/overlayfs/file.c | 47 ++++++++++++++++++++++++++++++++++++++++++++-
> 1 file changed, 46 insertions(+), 1 deletion(-)
>
> diff --git a/fs/overlayfs/file.c b/fs/overlayfs/file.c
> index 4d53d3b7e5fe..d081faa55e83 100644
> --- a/fs/overlayfs/file.c
> +++ b/fs/overlayfs/file.c
> @@ -392,6 +392,51 @@ out_unlock:
> return ret;
> }
>
> +/*
> + * Calling iter_file_splice_write() directly from overlay's f_op may deadlock
> + * due to lock order inversion between pipe->mutex in iter_file_splice_write()
> + * and file_start_write(real.file) in ovl_write_iter().
> + *
> + * So do everything ovl_write_iter() does and call iter_file_splice_write() on
> + * the real file.
> + */
> +static ssize_t ovl_splice_write(struct pipe_inode_info *pipe, struct file *out,
> + loff_t *ppos, size_t len, unsigned int flags)
> +{
> + struct fd real;
> + const struct cred *old_cred;
> + struct inode *inode = file_inode(out);
> + struct inode *realinode = ovl_inode_real(inode);
> + ssize_t ret;
> +
> + inode_lock(inode);
> + /* Update mode */
> + ovl_copyattr(realinode, inode);
> + ret = file_remove_privs(out);
> + if (ret)
> + goto out_unlock;
> +
> + ret = ovl_real_fdget(out, &real);
> + if (ret)
> + goto out_unlock;
> +
> + old_cred = ovl_override_creds(inode->i_sb);
> + file_start_write(real.file);
> +
> + ret = iter_file_splice_write(pipe, real.file, ppos, len, flags);
> +
> + file_end_write(real.file);
> + /* Update size */
> + ovl_copyattr(realinode, inode);
> + revert_creds(old_cred);
> + fdput(real);
> +
> +out_unlock:
> + inode_unlock(inode);
> +
> + return ret;
> +}
> +
> static int ovl_fsync(struct file *file, loff_t start, loff_t end, int datasync)
> {
> struct fd real;
> @@ -603,7 +648,7 @@ const struct file_operations ovl_file_operations = {
> .fadvise = ovl_fadvise,
> .flush = ovl_flush,
> .splice_read = generic_file_splice_read,
> - .splice_write = iter_file_splice_write,
> + .splice_write = ovl_splice_write,
>
> .copy_file_range = ovl_copy_file_range,
> .remap_file_range = ovl_remap_file_range,
> --
> 2.30.2
>
>
>

\
 
 \ /
  Last update: 2021-11-15 14:56    [W:0.439 / U:0.056 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site