Re: [Intel-gfx] [PATCH v2 2/7] drm/i915: Mark the ggtt_view structs as packed

2017-01-13 Thread Tvrtko Ursulin


On 13/01/2017 08:47, Chris Wilson wrote:

On Fri, Jan 13, 2017 at 08:44:34AM +, Tvrtko Ursulin wrote:


On 12/01/2017 21:35, Chris Wilson wrote:

In the next few patches, we will depend upon there being no
uninitialised bits inside the ggtt_view. To ensure this we add the
__packed attribute and double check with a build on that gcc hasn't
expanded the struct to include some padding bytes.

Signed-off-by: Chris Wilson 
---
drivers/gpu/drm/i915/i915_gem_gtt.h | 14 --
1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h 
b/drivers/gpu/drm/i915/i915_gem_gtt.h
index 5dd3755a5a45..57cbd532dae1 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.h
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
@@ -156,12 +156,22 @@ struct intel_rotation_info {
/* tiles */
unsigned int width, height, stride, offset;
} plane[2];


Isn't packed theoretically needed on the intel_rotation_plane_info
name in the previous patch as well? Otherwise there could be a hole
between the array elements if the structure got changed in the
future.


Possibly, not too sure on the inner details of __packed. The assert
below will catch accidental holes in future (and if they change the
struct without changing the assert, a nuisance).


+static inline void assert_intel_rotation_info_is_packed(void)
+{
+   BUILD_BUG_ON(sizeof(struct intel_rotation_info) != 8*sizeof(unsigned 
int));
+}




Somehow I only saw the assert for partial info.

Reviewed-by: Tvrtko Ursulin 

Regards,

Tvrtko
___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx


Re: [Intel-gfx] [PATCH v2 2/7] drm/i915: Mark the ggtt_view structs as packed

2017-01-13 Thread Chris Wilson
On Fri, Jan 13, 2017 at 08:44:34AM +, Tvrtko Ursulin wrote:
> 
> On 12/01/2017 21:35, Chris Wilson wrote:
> >In the next few patches, we will depend upon there being no
> >uninitialised bits inside the ggtt_view. To ensure this we add the
> >__packed attribute and double check with a build on that gcc hasn't
> >expanded the struct to include some padding bytes.
> >
> >Signed-off-by: Chris Wilson 
> >---
> > drivers/gpu/drm/i915/i915_gem_gtt.h | 14 --
> > 1 file changed, 12 insertions(+), 2 deletions(-)
> >
> >diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h 
> >b/drivers/gpu/drm/i915/i915_gem_gtt.h
> >index 5dd3755a5a45..57cbd532dae1 100644
> >--- a/drivers/gpu/drm/i915/i915_gem_gtt.h
> >+++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
> >@@ -156,12 +156,22 @@ struct intel_rotation_info {
> > /* tiles */
> > unsigned int width, height, stride, offset;
> > } plane[2];
> 
> Isn't packed theoretically needed on the intel_rotation_plane_info
> name in the previous patch as well? Otherwise there could be a hole
> between the array elements if the structure got changed in the
> future.

Possibly, not too sure on the inner details of __packed. The assert
below will catch accidental holes in future (and if they change the
struct without changing the assert, a nuisance).

> >+static inline void assert_intel_rotation_info_is_packed(void)
> >+{
> >+BUILD_BUG_ON(sizeof(struct intel_rotation_info) != 8*sizeof(unsigned 
> >int));
> >+}

-- 
Chris Wilson, Intel Open Source Technology Centre
___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx


Re: [Intel-gfx] [PATCH v2 2/7] drm/i915: Mark the ggtt_view structs as packed

2017-01-13 Thread Tvrtko Ursulin


On 12/01/2017 21:35, Chris Wilson wrote:

In the next few patches, we will depend upon there being no
uninitialised bits inside the ggtt_view. To ensure this we add the
__packed attribute and double check with a build on that gcc hasn't
expanded the struct to include some padding bytes.

Signed-off-by: Chris Wilson 
---
 drivers/gpu/drm/i915/i915_gem_gtt.h | 14 --
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h 
b/drivers/gpu/drm/i915/i915_gem_gtt.h
index 5dd3755a5a45..57cbd532dae1 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.h
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
@@ -156,12 +156,22 @@ struct intel_rotation_info {
/* tiles */
unsigned int width, height, stride, offset;
} plane[2];


Isn't packed theoretically needed on the intel_rotation_plane_info name 
in the previous patch as well? Otherwise there could be a hole between 
the array elements if the structure got changed in the future.


Regards,

Tvrtko


-};
+} __packed;
+
+static inline void assert_intel_rotation_info_is_packed(void)
+{
+   BUILD_BUG_ON(sizeof(struct intel_rotation_info) != 8*sizeof(unsigned 
int));
+}

 struct intel_partial_info {
u64 offset;
unsigned int size;
-};
+} __packed;
+
+static inline void assert_intel_partial_info_is_packed(void)
+{
+   BUILD_BUG_ON(sizeof(struct intel_partial_info) != sizeof(u64) + 
sizeof(unsigned int));
+}

 struct i915_ggtt_view {
enum i915_ggtt_view_type type;


___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx