Re: [PATCH 3/4] fs: remove fsparam_path / fs_param_is_path

["Darrick J. Wong" <djwong@kernel.org>]

On Fri, Feb 20, 2026 at 04:24:02PM +0100, Christoph Hellwig wrote:
> On Thu, Feb 19, 2026 at 08:04:28AM -0800, Darrick J. Wong wrote:
> > > diff --git a/Documentation/filesystems/mount_api.rst b/Documentation/filesystems/mount_api.rst
> > > index b4a0f23914a6..e8b94357b4df 100644
> > > --- a/Documentation/filesystems/mount_api.rst
> > > +++ b/Documentation/filesystems/mount_api.rst
> > > @@ -648,7 +648,6 @@ The members are as follows:
> > >  	fs_param_is_enum	Enum value name 	result->uint_32
> > >  	fs_param_is_string	Arbitrary string	param->string
> > >  	fs_param_is_blockdev	Blockdev path		* Needs lookup
> > 
> > Unrelated: should xfs be using fsparam_bdev for its logdev/rtdev mount
> > options?
> 
> Not sure what the point is in having separate string helpers with meaning,
> but maybe I'm missing something/

I'm not sure either -- it'd be one thing if the fsconfig code could
supply us with an open struct file to an O_EXCL bdev, but looking at the
sole user ext4, all it does is sample i_rdev and pass it to
bdev_file_open_by_dev.

> > Or, more crazily, should it grow logfd/rtfd options that use fsparam_fd?
> 
> What would the use case be for that?

I've no idea, I guess if we had mount helper magic then it would be
useful.

--D