[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAMuHMdXWRgL0uxLXt7geDCuHdQ=CDPhOzE7WCa4f4wRFvD50sw@mail.gmail.com>
Date: Fri, 29 Sep 2023 09:33:55 +0200
From: Geert Uytterhoeven <geert@...ux-m68k.org>
To: Thomas Zimmermann <tzimmermann@...e.de>
Cc: Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
Maxime Ripard <mripard@...nel.org>,
David Airlie <airlied@...il.com>,
Daniel Vetter <daniel@...ll.ch>,
Andrzej Hajda <andrzej.hajda@...el.com>,
Neil Armstrong <neil.armstrong@...aro.org>,
Robert Foss <rfoss@...nel.org>,
Laurent Pinchart <Laurent.pinchart@...asonboard.com>,
Jonas Karlman <jonas@...boo.se>,
Jernej Skrabec <jernej.skrabec@...il.com>,
linux-kernel@...r.kernel.org, dri-devel@...ts.freedesktop.org
Subject: Re: [PATCH/RFC 3/3] drm: Split drm_modeset_helper_vtables.h
Hi Thomas,
On Fri, Sep 29, 2023 at 9:11 AM Thomas Zimmermann <tzimmermann@...e.de> wrote:
> Am 28.09.23 um 17:32 schrieb Geert Uytterhoeven:
> > On Thu, Sep 28, 2023 at 3:59 PM Thomas Zimmermann <tzimmermann@...e.de> wrote:
> >> Am 28.09.23 um 14:16 schrieb Geert Uytterhoeven:
> >>> <drm/drm_modeset_helper_vtables.h> is the second largest header file in
> >>> the DRM subsystem, and declares helpers vtables for various DRM
> >>> components. Several vtables contain methods with the same name, and all
> >>> but one vtable do not fit on the screen, making it hard to navigate to
> >>> the actual method one is interested in.
> >>>
> >>> Make it easier for the casual reviewer to keep track by splitting
> >>> <drm/drm_modeset_helper_vtables.h> in multiple header files, one per DRM
> >>> component.
> >>
> >> I never liked this header either, but do we need new header files? Each
> >> struct could be appended to the end of the regular header: struct
> >> drm_plane_helper_funcs to drm_plane.h, drm_connector_helper_func to
> >> drm_connector.h and so on.
> >
> > That would work for me, too. But perhaps we want to maintain a clear
> > separation between core and helpers?
> >
> > Note that moving the contents to *_helper.h would be another option,
> > drm_crtc_helper.h and drm_plane_helper.h already exist.
>
> I've taken a closer look at the users of the _vtables header. There's
> code in drm_atomic_helper.c or drm_probe_helper.c that invokes the
> callback functions.
>
> The drivers fill the pointers with code that often comes from other
> helper modules. That code is in files like drm_plane_helper.c or
> drm_crtc_helper.c. There header files are drm_plane_helper.h, etc.
>
> In that context, the _vtables header makes sense, as it separates the
> callers from the callees. Putting the structs into headers like
> drm_plane_helper.h would move it to the callee side.
>
> I suggest to leave the header as it is. The fallout to the code base
> from refactoring seems worse than the current state.
To clarify: do you mean keeping the single big drm_modeset_helper_vtables.h,
or the split drm_*_helper_vtable.h set?
Thanks!
> >>> Signed-off-by: Geert Uytterhoeven <geert+renesas@...der.be>
> >>> ---
> >>> RFC, a future patch could replace inclusion of
> >>> <drm/drm_modeset_helper_vtables.h> by inclusion of one or more of the
> >>> new files, and reduce compilation time.
> >>> ---
> >>> include/drm/drm_connector_helper_vtable.h | 364 +++++
> >>> include/drm/drm_crtc_helper_vtable.h | 483 ++++++
> >>> include/drm/drm_encoder_helper_vtable.h | 381 +++++
> >>> include/drm/drm_mode_config_helper_vtable.h | 97 ++
> >>> include/drm/drm_modeset_helper_vtables.h | 1466 +------------------
> >>> include/drm/drm_plane_helper_vtable.h | 297 ++++
> >>> 6 files changed, 1627 insertions(+), 1461 deletions(-)
> >>> create mode 100644 include/drm/drm_connector_helper_vtable.h
> >>> create mode 100644 include/drm/drm_crtc_helper_vtable.h
> >>> create mode 100644 include/drm/drm_encoder_helper_vtable.h
> >>> create mode 100644 include/drm/drm_mode_config_helper_vtable.h
> >>> create mode 100644 include/drm/drm_plane_helper_vtable.h
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@...ux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
Powered by blists - more mailing lists