pixman: Branch 'master' - 2 commits

[GitLab Mirror]

 pixman/pixman-vmx.c |   33 ++++++++-------------------------
 1 file changed, 8 insertions(+), 25 deletions(-)

New commits:
commit 47a1c3d330cc7ea6ba2fa96eb288f4b2291d011c
Author: Havard Eidnes <he at NetBSD.org>
Date:   Wed Aug 30 12:09:23 2023 -0400

    vmx: Reimplement create_mask_32_128 and use it in vmx_fill
    
    Based on suggestion from @siamashka.
    
    This lets the compiler pick the vector instruction to use which is
    usually the best idea.
    
    Use create_mask_32_128() instead of create_mask_1x32_128() in
    vmx_fill(), avoiding loading memory beyond the filler argument on the
    stack.
    
    Remove the now-unused create_mask_1x32_128(). This gets rid of some
    (correct) warnings from the compiler about indexing beyond the variable
    in question.

diff --git a/pixman/pixman-vmx.c b/pixman/pixman-vmx.c
index ab1250c..d4b5dc8 100644
--- a/pixman/pixman-vmx.c
+++ b/pixman/pixman-vmx.c
@@ -278,21 +278,10 @@ save_128_aligned (uint32_t* data,
     STORE_VECTOR(data)
 }
 
-static force_inline vector unsigned int
-create_mask_1x32_128 (const uint32_t *src)
-{
-    vector unsigned int vsrc;
-    DECLARE_SRC_MASK_VAR;
-
-    COMPUTE_SHIFT_MASK (src);
-    LOAD_VECTOR (src);
-    return vec_splat(vsrc, 0);
-}
-
 static force_inline vector unsigned int
 create_mask_32_128 (uint32_t mask)
 {
-    return create_mask_1x32_128(&mask);
+    return (vector unsigned int) {mask, mask, mask, mask};
 }
 
 static force_inline vector unsigned int
@@ -2471,7 +2460,7 @@ vmx_fill (pixman_implementation_t *imp,
 	return FALSE;
     }
 
-    vfiller = create_mask_1x32_128(&filler);
+    vfiller = create_mask_32_128(filler);
 
     while (height--)
     {
commit 634b8196d2ebc1b78adfc3cecca8efa80bb30f7f
Author: Havard Eidnes <he at NetBSD.org>
Date:   Wed Aug 30 12:09:29 2023 -0400

    vmx: Simplify scaled_nearest_scanline_vmx_8888_8888_OVER
    
    Since combine4() does not take vector variables as arguments, there's no
    need to use a vector variable and casts back and forth to normal scalars
    for the arguments.

diff --git a/pixman/pixman-vmx.c b/pixman/pixman-vmx.c
index 1086b28..ab1250c 100644
--- a/pixman/pixman-vmx.c
+++ b/pixman/pixman-vmx.c
@@ -2913,32 +2913,26 @@ scaled_nearest_scanline_vmx_8888_8888_OVER (uint32_t*       pd,
 
     while (w >= 4)
     {
-	vector unsigned int tmp;
-	uint32_t tmp1, tmp2, tmp3, tmp4;
+	uint32_t tmp[4];
 
-	tmp1 = *(ps + pixman_fixed_to_int (vx));
+	tmp[0] = *(ps + pixman_fixed_to_int (vx));
 	vx += unit_x;
 	while (vx >= 0)
 	    vx -= src_width_fixed;
-	tmp2 = *(ps + pixman_fixed_to_int (vx));
+	tmp[1] = *(ps + pixman_fixed_to_int (vx));
 	vx += unit_x;
 	while (vx >= 0)
 	    vx -= src_width_fixed;
-	tmp3 = *(ps + pixman_fixed_to_int (vx));
+	tmp[2] = *(ps + pixman_fixed_to_int (vx));
 	vx += unit_x;
 	while (vx >= 0)
 	    vx -= src_width_fixed;
-	tmp4 = *(ps + pixman_fixed_to_int (vx));
+	tmp[3] = *(ps + pixman_fixed_to_int (vx));
 	vx += unit_x;
 	while (vx >= 0)
 	    vx -= src_width_fixed;
 
-	tmp[0] = tmp1;
-	tmp[1] = tmp2;
-	tmp[2] = tmp3;
-	tmp[3] = tmp4;
-
-	vsrc = combine4 ((const uint32_t *) &tmp, pm);
+	vsrc = combine4 (tmp, pm);
 
 	if (is_opaque (vsrc))
 	{