[2/3] drm/omap: avoid aquiring mutex in atomic context

Message ID 1324004717-20595-2-git-send-email-rob.clark@linaro.org
State New
Headers show

Commit Message

Rob Clark Dec. 16, 2011, 3:05 a.m.
From: Rob Clark <rob@ti.com>

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

Signed-off-by: Rob Clark <rob@ti.com>
---
 drivers/staging/omapdrm/omap_gem.c |   16 ++++++++++++++--
 1 files changed, 14 insertions(+), 2 deletions(-)

Comments

Rob Clark Dec. 16, 2011, 5:18 p.m. | #1
On Thu, Dec 15, 2011 at 9:05 PM, Rob Clark <rob.clark@linaro.org> wrote:
> From: Rob Clark <rob@ti.com>
>
> omap_gem_roll() could be called by fbcon in atomic context.  Avoid
> aquiring mutex, or calling tiler_pin() (which itself is not safe
> for atomic context) in these cases.
>
> Signed-off-by: Rob Clark <rob@ti.com>
> ---
>  drivers/staging/omapdrm/omap_gem.c |   16 ++++++++++++++--
>  1 files changed, 14 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/staging/omapdrm/omap_gem.c b/drivers/staging/omapdrm/omap_gem.c
> index 9684891..2bc570a 100644
> --- 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()) {

ugg, this isn't even sufficient if you have any debug printks that can
be enabled within struct_mutex critical section..

updated patch coming

BR,
-R


> +               /* 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;
> --
> 1.7.5.4
>

Patch

diff --git a/drivers/staging/omapdrm/omap_gem.c b/drivers/staging/omapdrm/omap_gem.c
index 9684891..2bc570a 100644
--- 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()) {
+		/* 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;