summaryrefslogtreecommitdiffstats
path: root/drivers/staging/omapdrm/omap_drv.c
diff options
context:
space:
mode:
authorDaniel Vetter <daniel.vetter@ffwll.ch>2013-01-20 15:50:41 +0100
committerDaniel Vetter <daniel.vetter@ffwll.ch>2013-01-20 22:16:53 +0100
commitd5d2636ed7990b93c7216f6a4d323f6b0eee08af (patch)
tree497dec3a9bc20613069422e40aa8f8b40e389621 /drivers/staging/omapdrm/omap_drv.c
parentbbe4b99ff2443e305598768ae8eac6bc3516b7c9 (diff)
downloadlinux-stable-d5d2636ed7990b93c7216f6a4d323f6b0eee08af.tar.gz
linux-stable-d5d2636ed7990b93c7216f6a4d323f6b0eee08af.tar.bz2
linux-stable-d5d2636ed7990b93c7216f6a4d323f6b0eee08af.zip
omapdrm: use modeset_lock_all
I've left the locking in the debugfs code as-is, it's essentially just used to keep the framebuffer object alive (which won't be necessary any more later on). We don't need fb refcounting either, since the new mode_config.fb_lock ensures that the framebuffers can't disappear (once mode_config.mutex doesn't guarantee this any more later on in the series). The fbcon restore needs all modeset locks. The crtc callbacks seem to only need the crtc locks, but I've quickly discussed things with Rob Clark and he's fine with just using modeset_lock_all for those, too. He'll look into converting things over later. Reviewed-by: Rob Clark <rob@ti.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Diffstat (limited to 'drivers/staging/omapdrm/omap_drv.c')
-rw-r--r--drivers/staging/omapdrm/omap_drv.c4
1 files changed, 2 insertions, 2 deletions
diff --git a/drivers/staging/omapdrm/omap_drv.c b/drivers/staging/omapdrm/omap_drv.c
index ae5ecc2efbc7..dfdb4ba1e7c6 100644
--- a/drivers/staging/omapdrm/omap_drv.c
+++ b/drivers/staging/omapdrm/omap_drv.c
@@ -449,9 +449,9 @@ static void dev_lastclose(struct drm_device *dev)
}
}
- mutex_lock(&dev->mode_config.mutex);
+ drm_modeset_lock_all(dev);
ret = drm_fb_helper_restore_fbdev_mode(priv->fbdev);
- mutex_unlock(&dev->mode_config.mutex);
+ drm_modeset_unlock_all(dev);
if (ret)
DBG("failed to restore crtc mode");
}