diff mbox

drm/i915: Don't cleanup plane state in intel_plane_destroy()

Message ID 1421421924-9697-1-git-send-email-matthew.d.roper@intel.com (mailing list archive)
State New, archived
Headers show

Commit Message

Matt Roper Jan. 16, 2015, 3:25 p.m. UTC
When we transitioned to the atomic plane helpers in commit:

        commit ea2c67bb4affa84080c616920f3899f123786e56
        Author: Matt Roper <matthew.d.roper@intel.com>
        Date:   Tue Dec 23 10:41:52 2014 -0800

            drm/i915: Move to atomic plane helpers (v9)

one of the changes was to call intel_plane_destroy_state() while tearing
down a plane to prevent leaks when unloading the driver.  That made
sense when the patches were first written, but before they were merged,

        commit 3009c0377f25c29852b218a6933a969d02cbdc5d
        Author:     Thierry Reding <treding@nvidia.com>
        Date:       Tue Nov 25 12:09:49 2014 +0100

            drm: Free atomic state during cleanup

had already landed, which made this the responsibility of the DRM core.
The result was that we were kfree()'ing the state twice, and also
possibly double-unref'ing a framebuffer, leading to memory corruption
when the driver was unloaded.

The fix is to simply not try to cleanup the state in the i915 teardown
code now that the core handles this for us.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=88433
Testcase: igt/drv_module_reload
Root-cause-analysis-by: Ander Conselvan de Oliveira <conselvan2@gmail.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
---
 drivers/gpu/drm/i915/intel_display.c | 1 -
 1 file changed, 1 deletion(-)

Comments

Jani Nikula Jan. 16, 2015, 4:16 p.m. UTC | #1
On Fri, 16 Jan 2015, Matt Roper <matthew.d.roper@intel.com> wrote:
> When we transitioned to the atomic plane helpers in commit:
>
>         commit ea2c67bb4affa84080c616920f3899f123786e56
>         Author: Matt Roper <matthew.d.roper@intel.com>
>         Date:   Tue Dec 23 10:41:52 2014 -0800
>
>             drm/i915: Move to atomic plane helpers (v9)
>
> one of the changes was to call intel_plane_destroy_state() while tearing
> down a plane to prevent leaks when unloading the driver.  That made
> sense when the patches were first written, but before they were merged,
>
>         commit 3009c0377f25c29852b218a6933a969d02cbdc5d
>         Author:     Thierry Reding <treding@nvidia.com>
>         Date:       Tue Nov 25 12:09:49 2014 +0100
>
>             drm: Free atomic state during cleanup
>
> had already landed, which made this the responsibility of the DRM core.
> The result was that we were kfree()'ing the state twice, and also
> possibly double-unref'ing a framebuffer, leading to memory corruption
> when the driver was unloaded.
>
> The fix is to simply not try to cleanup the state in the i915 teardown
> code now that the core handles this for us.
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=88433
> Testcase: igt/drv_module_reload
> Root-cause-analysis-by: Ander Conselvan de Oliveira <conselvan2@gmail.com>

Thanked-by-and-good-weekend-wished-by-and-
Reviewed-by: Jani Nikula <jani.nikula@intel.com>

> Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
> ---
>  drivers/gpu/drm/i915/intel_display.c | 1 -
>  1 file changed, 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> index 91d8ada..cc3b9d8 100644
> --- a/drivers/gpu/drm/i915/intel_display.c
> +++ b/drivers/gpu/drm/i915/intel_display.c
> @@ -11937,7 +11937,6 @@ static void intel_finish_crtc_commit(struct drm_crtc *crtc)
>  void intel_plane_destroy(struct drm_plane *plane)
>  {
>  	struct intel_plane *intel_plane = to_intel_plane(plane);
> -	intel_plane_destroy_state(plane, plane->state);
>  	drm_plane_cleanup(plane);
>  	kfree(intel_plane);
>  }
> -- 
> 1.8.5.1
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Shuang He Jan. 17, 2015, 4:56 a.m. UTC | #2
Tested-By: PRC QA PRTS (Patch Regression Test System Contact: shuang.he@intel.com)
Task id: 5594
-------------------------------------Summary-------------------------------------
Platform          Delta          drm-intel-nightly          Series Applied
PNV                 -1              353/353              352/353
ILK                                  200/200              200/200
SNB                                  400/422              400/422
IVB                                  487/487              487/487
BYT                                  296/296              296/296
HSW                 -1              487/508              486/508
BDW                                  401/402              401/402
-------------------------------------Detailed-------------------------------------
Platform  Test                                drm-intel-nightly          Series Applied
*PNV  igt_gen3_render_linear_blits      PASS(3, M25M23)      CRASH(1, M23)
 HSW  igt_kms_flip_event_leak      NSPT(5, M19)PASS(1, M19)      NSPT(1, M19)
Note: You need to pay more attention to line start with '*'
Daniel Vetter Jan. 19, 2015, 9:05 a.m. UTC | #3
On Fri, Jan 16, 2015 at 06:16:50PM +0200, Jani Nikula wrote:
> On Fri, 16 Jan 2015, Matt Roper <matthew.d.roper@intel.com> wrote:
> > When we transitioned to the atomic plane helpers in commit:
> >
> >         commit ea2c67bb4affa84080c616920f3899f123786e56
> >         Author: Matt Roper <matthew.d.roper@intel.com>
> >         Date:   Tue Dec 23 10:41:52 2014 -0800
> >
> >             drm/i915: Move to atomic plane helpers (v9)
> >
> > one of the changes was to call intel_plane_destroy_state() while tearing
> > down a plane to prevent leaks when unloading the driver.  That made
> > sense when the patches were first written, but before they were merged,
> >
> >         commit 3009c0377f25c29852b218a6933a969d02cbdc5d
> >         Author:     Thierry Reding <treding@nvidia.com>
> >         Date:       Tue Nov 25 12:09:49 2014 +0100
> >
> >             drm: Free atomic state during cleanup
> >
> > had already landed, which made this the responsibility of the DRM core.
> > The result was that we were kfree()'ing the state twice, and also
> > possibly double-unref'ing a framebuffer, leading to memory corruption
> > when the driver was unloaded.
> >
> > The fix is to simply not try to cleanup the state in the i915 teardown
> > code now that the core handles this for us.
> >
> > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=88433
> > Testcase: igt/drv_module_reload
> > Root-cause-analysis-by: Ander Conselvan de Oliveira <conselvan2@gmail.com>
> 
> Thanked-by-and-good-weekend-wished-by-and-
> Reviewed-by: Jani Nikula <jani.nikula@intel.com>

Queued for -next, thanks for the patch.
-Daniel

> 
> > Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_display.c | 1 -
> >  1 file changed, 1 deletion(-)
> >
> > diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> > index 91d8ada..cc3b9d8 100644
> > --- a/drivers/gpu/drm/i915/intel_display.c
> > +++ b/drivers/gpu/drm/i915/intel_display.c
> > @@ -11937,7 +11937,6 @@ static void intel_finish_crtc_commit(struct drm_crtc *crtc)
> >  void intel_plane_destroy(struct drm_plane *plane)
> >  {
> >  	struct intel_plane *intel_plane = to_intel_plane(plane);
> > -	intel_plane_destroy_state(plane, plane->state);
> >  	drm_plane_cleanup(plane);
> >  	kfree(intel_plane);
> >  }
> > -- 
> > 1.8.5.1
> >
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 
> -- 
> Jani Nikula, Intel Open Source Technology Center
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx
diff mbox

Patch

diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 91d8ada..cc3b9d8 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -11937,7 +11937,6 @@  static void intel_finish_crtc_commit(struct drm_crtc *crtc)
 void intel_plane_destroy(struct drm_plane *plane)
 {
 	struct intel_plane *intel_plane = to_intel_plane(plane);
-	intel_plane_destroy_state(plane, plane->state);
 	drm_plane_cleanup(plane);
 	kfree(intel_plane);
 }