Re: [PATCH 09/15] tty: fix kernel-doc for functions in tty.h

["Ilpo Järvinen" <ilpo.jarvinen@linux.intel.com>]

[-- Attachment #1: Type: text/plain, Size: 2550 bytes --]

On Tue, 19 Sep 2023, Jiri Slaby wrote:

> On 19. 09. 23, 12:45, Jiri Slaby wrote:
> > On 19. 09. 23, 12:07, Ilpo Järvinen wrote:
> > > On Tue, 19 Sep 2023, Jiri Slaby (SUSE) wrote:
> > > 
> > > > tty_kref_get() is already included in Documentation, but is not properly
> > > > formatted. Fix this.
> > > > 
> > > > tty_get_baud_rate() is neither properly formatted, nor is included. Fix
> > > > both.
> > > > 
> > > > Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
> > > > ---
> > > >   Documentation/driver-api/tty/tty_ioctl.rst |  3 +++
> > > >   include/linux/tty.h                        | 21 +++++++++------------
> > > >   2 files changed, 12 insertions(+), 12 deletions(-)
> > > > 
> > > > diff --git a/Documentation/driver-api/tty/tty_ioctl.rst
> > > > b/Documentation/driver-api/tty/tty_ioctl.rst
> > > > index 9b0be79fc15e..3ff1ac5e07f1 100644
> > > > --- a/Documentation/driver-api/tty/tty_ioctl.rst
> > > > +++ b/Documentation/driver-api/tty/tty_ioctl.rst
> > > > @@ -5,3 +5,6 @@ TTY IOCTL Helpers
> > > >   =================
> > > >   .. kernel-doc:: drivers/tty/tty_ioctl.c
> > > > +
> > > > +.. kernel-doc:: include/linux/tty.h
> > > > +   :identifiers: tty_get_baud_rate
> > > > diff --git a/include/linux/tty.h b/include/linux/tty.h
> > > > index 59d675f345e9..4b6340ac2af2 100644
> > > > --- a/include/linux/tty.h
> > > > +++ b/include/linux/tty.h
> > > > @@ -390,14 +390,12 @@ int vcs_init(void);
> > > >   extern const struct class tty_class;
> > > >   /**
> > > > - *    tty_kref_get        -    get a tty reference
> > > > - *    @tty: tty device
> > > > + * tty_kref_get - get a tty reference
> > > > + * @tty: tty device
> > > >    *
> > > > - *    Return a new reference to a tty object. The caller must hold
> > > > - *    sufficient locks/counts to ensure that their existing reference
> > > > cannot
> > > > - *    go away
> > > > + * Returns: a new reference to a tty object. The caller must hold
> > > > sufficient
> > > > + * locks/counts to ensure that their existing reference cannot go away
> > > 
> > > Shouldn't this have also Locking: entry instead of hiding the details into
> > > Return?
> > 
> > /me left to fix both in a separate patch.
> 
> Ah, no. The Locking Alan introduced means what _this_ function locks. I am not
> sure we want to extend the meaning to _expected_ locks?

There are plenty of examples with "Must be called with" or "Caller holds"
in Locking. This is for humans anyway so one reading should understand it.


-- 
 i.