[Intel-gfx] [PATCH] drm/i915: Clean up vblank data after lastclose and unload

[Dave Airlie]

2009/12/18 Kristian Høgsberg <krh at bitplanet.net>:
> Otherwise we end up waking up a freed waitqueue.  Introduced in
> commit 778c902640530371a169ad1c03566e7c51b09874 which adds a
> drm_vblank_off() call in the DPMS off path.

Wrong list ;-), drm core patches to me + dri-devel, this seems like one of those
things that'll bite someone in the ass, so resend it so Michel can look at it.

Dave.
>
> Signed-off-by: Kristian Høgsberg <krh at bitplanet.net>
> ---
>
> As far as I can see, this should be safe.  drm_vblank_cleanup()
> just frees stuff and deletes a timer, so reordering this way should be
> safe.
>
>  drivers/gpu/drm/drm_stub.c |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c
> index ad73e14..3735478 100644
> --- a/drivers/gpu/drm/drm_stub.c
> +++ b/drivers/gpu/drm/drm_stub.c
> @@ -515,8 +515,6 @@ void drm_put_dev(struct drm_device *dev)
>        }
>        driver = dev->driver;
>
> -       drm_vblank_cleanup(dev);
> -
>        drm_lastclose(dev);
>
>        if (drm_core_has_MTRR(dev) && drm_core_has_AGP(dev) &&
> @@ -531,6 +529,8 @@ void drm_put_dev(struct drm_device *dev)
>        if (dev->driver->unload)
>                dev->driver->unload(dev);
>
> +       drm_vblank_cleanup(dev);
> +
>        if (drm_core_has_AGP(dev) && dev->agp) {
>                kfree(dev->agp);
>                dev->agp = NULL;
> --
> 1.6.5.rc2
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx
>