On Wed, Jul 22, 2020 at 08:44:32AM +0100, Al Viro wrote:
> On Tue, Jul 21, 2020 at 08:27:01PM +0200, Christoph Hellwig wrote:
> > On Tue, Jul 21, 2020 at 05:55:39PM +0100, Al Viro wrote:
> > > How about fs/for_init.c and putting the damn helpers there? With
> > > calling conventions as close to syscalls as possible, and a fat
> > > comment regarding their intended use being _ONLY_ the setup
> > > in should-have-been-done-in-userland parts of init?
> >
> > Where do you want the prototypes to go? Also do you want devtmpfs
> > use the same helpers, which then't can't be marked __init (mount,
> > chdir, chroot), or separate copies?
>
> Hmm... mount still can be __init (devtmpfs_mount() is), and I suspect
> devtmpfs_setup() could also be made such - just turn devtmpfsd()
> into
> static int __init devtmpfsd(void *p)
> {
> int err = devtmpfs_setup(p);
>
> if (!err)
> devtmpfsd_real(); /* never returns */
> return err;
> }
> and you are done.
Yes, that seems to work. We can obviously call non-__init functions
from __init ones, and kthread_run doesn't seem to care if it gets passed
a __init function.
Here is what I have now:
http://git.infradead.org/users/hch/misc.git/shortlog/refs/heads/init_path
