Jens, Mark,
This series deals with the problem of generic_file_splice_write() and
ocfs2_file_splice_write() holding i_mutex (and ocfs2 lock) on
destination inode across pipe_wait(), which waits for a user
controlled event.
Lightly tested on ext3. The OCFS2 changes have only been compile
tested.
Please review/test.
Thanks,
Miklos
--
On Tue, Apr 14 2009, Miklos Szeredi wrote:
> Jens, Mark,
>
> This series deals with the problem of generic_file_splice_write() and
> ocfs2_file_splice_write() holding i_mutex (and ocfs2 lock) on
> destination inode across pipe_wait(), which waits for a user
> controlled event.
>
> Lightly tested on ext3. The OCFS2 changes have only been compile
> tested.
>
> Please review/test.
Thanks Miklos, I've integrated it for testing. Looks good!
--
Jens Axboe