staging: drm/omap: avoid aquiring mutex in atomic context (v2)

omap_gem_roll() could be called by fbcon in atomic context or when
struct_mutext is held.  Avoid aquiring mutex (deadlock), or calling
tiler_pin() (which itself is not safe for atomic context) in these
cases.

Signed-off-by: Rob Clark <rob@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

diff --git a/drivers/staging/omapdrm/omap_gem.c b/drivers/staging/omapdrm/omap_gem.c
index 96848913085f63ba73d66c56925c48c1a4de3686..63490f75350a2da35e9d01e948f51a3eee629003 100644 (file)
--- a/drivers/staging/omapdrm/omap_gem.c
+++ b/drivers/staging/omapdrm/omap_gem.c
@@ -538,10 +538,22 @@ int omap_gem_roll(struct drm_gem_object *obj, uint32_t roll)
                return -EINVAL;
        }
 
-       mutex_lock(&obj->dev->struct_mutex);
-
        omap_obj->roll = roll;
 
+       if (in_atomic() || mutex_is_locked(&obj->dev->struct_mutex)) {
+               /* this can get called from fbcon in atomic context.. so
+                * just ignore it and wait for next time called from
+                * interruptible context to update the PAT.. the result
+                * may be that user sees wrap-around instead of scrolling
+                * momentarily on the screen.  If we wanted to be fancier
+                * we could perhaps schedule some workqueue work at this
+                * point.
+                */
+               return 0;
+       }
+
+       mutex_lock(&obj->dev->struct_mutex);
+
        /* if we aren't mapped yet, we don't need to do anything */
        if (omap_obj->block) {
                struct page **pages;