drm: Optimise power-of-two alignments in drm_mm_scan_add_block()

For power-of-two alignments, we can avoid the 64bit divide and do a
simple bitwise add instead.

v2: s/alignment_mask/remainder_mask/

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20161222083641.2691-32-chris@chris-wilson.co.uk

2 files changed, 9 insertions, 1 deletions

diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c
index 189ab84c5a59..0441d84fba74 100644
--- a/drivers/gpu/drm/drm_mm.c
+++ b/drivers/gpu/drm/drm_mm.c
@@ -742,8 +742,12 @@ void drm_mm_scan_init_with_range(struct drm_mm_scan *scan,
 
         scan->mm = mm;
 
+        if (alignment <= 1)
+                alignment = 0;
+
         scan->color = color;
         scan->alignment = alignment;
+        scan->remainder_mask = is_power_of_2(alignment) ? alignment - 1 : 0;
         scan->size = size;
         scan->flags = flags;
 
@@ -811,7 +815,10 @@ bool drm_mm_scan_add_block(struct drm_mm_scan *scan,
         if (scan->alignment) {
                 u64 rem;
 
-                div64_u64_rem(adj_start, scan->alignment, &rem);
+                if (likely(scan->remainder_mask))
+                        rem = adj_start & scan->remainder_mask;
+                else
+                        div64_u64_rem(adj_start, scan->alignment, &rem);
                 if (rem) {
                         adj_start -= rem;
                         if (scan->flags != DRM_MM_CREATE_TOP)
diff --git a/include/drm/drm_mm.h b/include/drm/drm_mm.h
index 606336fc229a..d6701d56ea74 100644
--- a/include/drm/drm_mm.h
+++ b/include/drm/drm_mm.h
@@ -110,6 +110,7 @@ struct drm_mm_scan {
 
         u64 size;
         u64 alignment;
+        u64 remainder_mask;
 
         u64 range_start;
         u64 range_end;