On Thu, Sep 22, 2016 at 10:49 AM, Amir Goldstein <amir73il@xxxxxxxxx> wrote:
> Guys, I just hit a lockdep warning on nesting an mnt_want_write() with
> this patch
> because overlay already holds the write lock when getting to ovl_copy_up_data()
> I don't know how I missed it before.
>
> Is it really a problem to nest this lock?
> Should I factor our another set of _locked helpers from the
> vfs_{copy,clone}_file_range helpers?
>
vfs_{copy,clone}_file_range should call sb_start_write() instead of
mnt_want_write_file(), the checks against FMODE_WRITE + S_ISREG ensure
that the __mnt_want_write() part is already held on the file.
That still leaves the lockdep warning. We could do
__vfs_{copy,clone}_file_range() variants without the
sb_start_write()/sb_end_write() and add the non-underscore variants as
static inline to fs.h that do call the sb_start/end_write.
Thanks,
Miklos
--
To unsubscribe from this list: send the line "unsubscribe linux-unionfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
