From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ville =?iso-8859-1?Q?Syrj=E4l=E4?= Subject: Re: [PATCH] drm: Simplify fb refcounting rules around ->update_plane Date: Tue, 22 Apr 2014 17:28:04 +0300 Message-ID: <20140422142804.GW18465@intel.com> References: <1398157633-31043-1-git-send-email-daniel.vetter@ffwll.ch> <20140422100622.GU18465@intel.com> <20140422140909.GQ10722@phenom.ffwll.local> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Return-path: Received: from mga09.intel.com (mga09.intel.com [134.134.136.24]) by gabe.freedesktop.org (Postfix) with ESMTP id C263A6E89D for ; Tue, 22 Apr 2014 07:28:24 -0700 (PDT) Content-Disposition: inline In-Reply-To: <20140422140909.GQ10722@phenom.ffwll.local> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" To: Daniel Vetter Cc: Daniel Vetter , DRI Development List-Id: dri-devel@lists.freedesktop.org On Tue, Apr 22, 2014 at 04:09:09PM +0200, Daniel Vetter wrote: > On Tue, Apr 22, 2014 at 01:06:22PM +0300, Ville Syrj=E4l=E4 wrote: > > On Tue, Apr 22, 2014 at 11:07:13AM +0200, Daniel Vetter wrote: > > > The introduction of primary planes has apparently caused a bit of fb > > > refcounting fun for people. That makes it a good time to clean up the > > > arcane rules and slight differences between ->update_plane and > > > ->set_config. The new rules are: > > > = > > > - The core holds a reference for both the new and the old fb (if their > > > non-NULL of course) while calling into the driver through either > > > ->update_plane or ->set_config. > > > = > > > - Drivers may not clobber plane->fb if their callback fails. If they > > > do that, they need to store a pointer to the old fb in it again. > > > When calling into the driver plane->fb still points at the current > > > (old) framebuffer. > > > = > > > - The core will update the plane->fb pointer on success. Drivers can > > > do that themselves too. > > > = > > > - The core will update fb refcounts for the plane->fb pointer, > > > presuming the drivers hold up their end of the bargain. > > > = > > > Signed-off-by: Daniel Vetter > > > --- > > > drivers/gpu/drm/drm_crtc.c | 12 +++++------- > > > drivers/gpu/drm/drm_plane_helper.c | 15 --------------- > > > 2 files changed, 5 insertions(+), 22 deletions(-) > > > = > > > diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c > > > index d8b7099abece..966b480ed543 100644 > > > --- a/drivers/gpu/drm/drm_crtc.c > > > +++ b/drivers/gpu/drm/drm_crtc.c > > > @@ -2187,24 +2187,24 @@ int drm_mode_setplane(struct drm_device *dev,= void *data, > > > } > > > = > > > drm_modeset_lock_all(dev); > > > + old_fb =3D plane->fb; > > > ret =3D plane->funcs->update_plane(plane, crtc, fb, > > > plane_req->crtc_x, plane_req->crtc_y, > > > plane_req->crtc_w, plane_req->crtc_h, > > > plane_req->src_x, plane_req->src_y, > > > plane_req->src_w, plane_req->src_h); > > > if (!ret) { > > > - old_fb =3D plane->fb; > > > plane->crtc =3D crtc; > > > plane->fb =3D fb; > > > - fb =3D NULL; > > > } > > > drm_modeset_unlock_all(dev); > > > = > > > out: > > > - if (fb) > > > - drm_framebuffer_unreference(fb); > > > + if (plane->fb) > > > + drm_framebuffer_reference(old_fb); > > ^^^^^^ > > = > > That doesn't look right. > > = > > Also you're now dereferencing plane->fb after you drop the locks. > = > Oops, will fix. > = > > Also i915 fb destruction seems buggy as we do > > intel_fb->obj->framebuffer_references-- w/o holding struct_mutex. > = > Hm, this indeed looks fishy. And the offending patch was actually r-b: > you. Dang. I suck. > Not sure what to do here really. Just s/drm_gem_object_unreference_unlocked/drm_gem_object_unreference/ and grab struct_mutex by hand around both operations? > -Daniel > = > > = > > > if (old_fb) > > > drm_framebuffer_unreference(old_fb); > > > + drm_framebuffer_unreference(fb); > > > = > > > return ret; > > > } > > > @@ -2239,9 +2239,7 @@ int drm_mode_set_config_internal(struct drm_mod= e_set *set) > > > ret =3D crtc->funcs->set_config(set); > > > if (ret =3D=3D 0) { > > > crtc->primary->crtc =3D crtc; > > > - > > > - /* crtc->fb must be updated by ->set_config, enforces this. */ > > > - WARN_ON(fb !=3D crtc->primary->fb); > > > + crtc->primary->fb =3D fb; > > > } > > > = > > > list_for_each_entry(tmp, &crtc->dev->mode_config.crtc_list, head) { > > > diff --git a/drivers/gpu/drm/drm_plane_helper.c b/drivers/gpu/drm/drm= _plane_helper.c > > > index e768d35ff22e..8db129c44fea 100644 > > > --- a/drivers/gpu/drm/drm_plane_helper.c > > > +++ b/drivers/gpu/drm/drm_plane_helper.c > > > @@ -175,22 +175,7 @@ int drm_primary_helper_update(struct drm_plane *= plane, struct drm_crtc *crtc, > > > set.connectors =3D connector_list; > > > set.num_connectors =3D num_connectors; > > > = > > > - /* > > > - * set_config() adjusts crtc->primary->fb; however the DRM setplane > > > - * code that called us expects to handle the framebuffer update and > > > - * reference counting; save and restore the current fb before > > > - * calling it. > > > - * > > > - * N.B., we call set_config() directly here rather than using > > > - * drm_mode_set_config_internal. We're reprogramming the same > > > - * connectors that were already in use, so we shouldn't need the ex= tra > > > - * cross-CRTC fb refcounting to accomodate stealing connectors. > > > - * drm_mode_setplane() already handles the basic refcounting for the > > > - * framebuffers involved in this operation. > > > - */ > > > - tmpfb =3D plane->fb; > > > ret =3D crtc->funcs->set_config(&set); > > > - plane->fb =3D tmpfb; > > > = > > > kfree(connector_list); > > > return ret; > > > -- = > > > 1.9.2 > > > = > > > _______________________________________________ > > > dri-devel mailing list > > > dri-devel@lists.freedesktop.org > > > http://lists.freedesktop.org/mailman/listinfo/dri-devel > > = > > -- = > > Ville Syrj=E4l=E4 > > Intel OTC > = > -- = > Daniel Vetter > Software Engineer, Intel Corporation > +41 (0) 79 365 57 48 - http://blog.ffwll.ch -- = Ville Syrj=E4l=E4 Intel OTC