Hi Emil,
On 2017-05-08 15:43, Emil Velikov wrote:
Hi Marek,
A couple of small nitpicks from UAPI POV.
Thanks for your comments!
On 8 May 2017 at 10:11, Marek Szyprowski <m.szyprowski@xxxxxxxxxxx> wrote:
--- a/include/uapi/drm/exynos_drm.h
+++ b/include/uapi/drm/exynos_drm.h
+struct drm_exynos_pp_get_res {
+ __u64 pp_id_ptr;
+ __u32 count_pps;
Add __u32 pad - sizeof(struct ...) should be multiple of sizeof(__u64).
ok
+struct drm_exynos_pp_get {
+ __u32 pp_id;
+ __u32 capabilities;
+
+ __u32 src_format_count;
+ __u32 dst_format_count;
+ __u32 params_count;
+ __u32 reserved1;
+
Replace with __u32 flags; so that you can extend the struct as applicable.
ok
+ __u64 src_format_type_ptr;
+ __u64 dst_format_type_ptr;
+ __u64 params_ptr;
+ __u64 reserved2;
And now you can drop this piece.
+struct drm_exynos_pp_commit {
+ __u32 id;
+ __u32 flags;
+ __u32 params_count;
+ __u32 reserved;
Why the reserved here - flags should help you extend as needed.
To ensure that the structure will be a multiple of sizeof(u64).
+ __u64 param_ids_ptr;
+ __u64 param_values_ptr;
+ __u64 user_data;
Unused user_data?
Nope. User_data provided here will be reported back to userspace in the
struct
drm_exynos_pp_event, so we cannot drop it.
+struct drm_exynos_pp_event {
+ struct drm_event base;
+ __u64 user_data;
Unused user_data?
See comment above.
+ __u32 tv_sec;
+ __u32 tv_usec;
+ __u32 pp_id;
+ __u32 sequence;
+ __u64 reserved;
Drop in favour of flags?
u64 flags? Probably u32 flags + u32 reserved would make a bit more sense to
keep structure a multiple of sizeof(u64).
Best regards
--
Marek Szyprowski, PhD
Samsung R&D Institute Poland
_______________________________________________
dri-devel mailing list
dri-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/dri-devel
