diff options
Diffstat (limited to 'pixman/pixman/pixman-arm-simd-asm.S')
| -rw-r--r-- | pixman/pixman/pixman-arm-simd-asm.S | 66 |
1 files changed, 52 insertions, 14 deletions
diff --git a/pixman/pixman/pixman-arm-simd-asm.S b/pixman/pixman/pixman-arm-simd-asm.S index d97545c1b..e00836b6c 100644 --- a/pixman/pixman/pixman-arm-simd-asm.S +++ b/pixman/pixman/pixman-arm-simd-asm.S @@ -331,15 +331,29 @@ pixman_asm_function pixman_composite_over_n_8_8888_asm_armv6 .endfunc
/*
- * Note: This function is only using armv4t instructions (not even armv6),
+ * Note: This code is only using armv5te instructions (not even armv6),
* but is scheduled for ARM Cortex-A8 pipeline. So it might need to
* be split into a few variants, tuned for each microarchitecture.
*
* TODO: In order to get good performance on ARM9/ARM11 cores (which don't
* have efficient write combining), it needs to be changed to use 16-byte
* aligned writes using STM instruction.
+ *
+ * Nearest scanline scaler macro template uses the following arguments:
+ * fname - name of the function to generate
+ * bpp_shift - (1 << bpp_shift) is the size of pixel in bytes
+ * t - type suffix for LDR/STR instructions
+ * prefetch_distance - prefetch in the source image by that many
+ * pixels ahead
+ * prefetch_braking_distance - stop prefetching when that many pixels are
+ * remaining before the end of scanline
*/
-pixman_asm_function pixman_scaled_nearest_scanline_0565_0565_SRC_asm_armv6
+
+.macro generate_nearest_scanline_func fname, bpp_shift, t, \
+ prefetch_distance, \
+ prefetch_braking_distance
+
+pixman_asm_function fname
W .req r0
DST .req r1
SRC .req r2
@@ -348,30 +362,46 @@ pixman_asm_function pixman_scaled_nearest_scanline_0565_0565_SRC_asm_armv6 TMP1 .req r4
TMP2 .req r5
VXMASK .req r6
+ PF_OFFS .req r7
ldr UNIT_X, [sp]
push {r4, r5, r6, r7}
- mvn VXMASK, #1
+ mvn VXMASK, #((1 << bpp_shift) - 1)
/* define helper macro */
.macro scale_2_pixels
- ldrh TMP1, [SRC, TMP1]
- and TMP2, VXMASK, VX, lsr #15
+ ldr&t TMP1, [SRC, TMP1]
+ and TMP2, VXMASK, VX, lsr #(16 - bpp_shift)
add VX, VX, UNIT_X
- strh TMP1, [DST], #2
+ str&t TMP1, [DST], #(1 << bpp_shift)
- ldrh TMP2, [SRC, TMP2]
- and TMP1, VXMASK, VX, lsr #15
+ ldr&t TMP2, [SRC, TMP2]
+ and TMP1, VXMASK, VX, lsr #(16 - bpp_shift)
add VX, VX, UNIT_X
- strh TMP2, [DST], #2
+ str&t TMP2, [DST], #(1 << bpp_shift)
.endm
/* now do the scaling */
- and TMP1, VXMASK, VX, lsr #15
+ and TMP1, VXMASK, VX, lsr #(16 - bpp_shift)
add VX, VX, UNIT_X
- subs W, #4
+ subs W, W, #(8 + prefetch_braking_distance)
blt 2f
-1: /* main loop, process 4 pixels per iteration */
+ /* calculate prefetch offset */
+ mov PF_OFFS, #prefetch_distance
+ mla PF_OFFS, UNIT_X, PF_OFFS, VX
+1: /* main loop, process 8 pixels per iteration with prefetch */
+ subs W, W, #8
+ add PF_OFFS, UNIT_X, lsl #3
+ scale_2_pixels
+ scale_2_pixels
+ scale_2_pixels
+ scale_2_pixels
+ pld [SRC, PF_OFFS, lsr #(16 - bpp_shift)]
+ bge 1b
+2:
+ subs W, W, #(4 - 8 - prefetch_braking_distance)
+ blt 2f
+1: /* process the remaining pixels */
scale_2_pixels
scale_2_pixels
subs W, W, #4
@@ -382,8 +412,8 @@ pixman_asm_function pixman_scaled_nearest_scanline_0565_0565_SRC_asm_armv6 scale_2_pixels
2:
tst W, #1
- ldrneh TMP1, [SRC, TMP1]
- strneh TMP1, [DST], #2
+ ldrne&t TMP1, [SRC, TMP1]
+ strne&t TMP1, [DST]
/* cleanup helper macro */
.purgem scale_2_pixels
.unreq DST
@@ -394,7 +424,15 @@ pixman_asm_function pixman_scaled_nearest_scanline_0565_0565_SRC_asm_armv6 .unreq TMP1
.unreq TMP2
.unreq VXMASK
+ .unreq PF_OFFS
/* return */
pop {r4, r5, r6, r7}
bx lr
.endfunc
+.endm
+
+generate_nearest_scanline_func \
+ pixman_scaled_nearest_scanline_0565_0565_SRC_asm_armv6, 1, h, 80, 32
+
+generate_nearest_scanline_func \
+ pixman_scaled_nearest_scanline_8888_8888_SRC_asm_armv6, 2, , 48, 32
|