diff options
Diffstat (limited to 'pixman/pixman/pixman-mips-dspr2-asm.S')
| -rw-r--r-- | pixman/pixman/pixman-mips-dspr2-asm.S | 704 |
1 files changed, 704 insertions, 0 deletions
diff --git a/pixman/pixman/pixman-mips-dspr2-asm.S b/pixman/pixman/pixman-mips-dspr2-asm.S index 48f108ed9..a8fccd5ad 100644 --- a/pixman/pixman/pixman-mips-dspr2-asm.S +++ b/pixman/pixman/pixman-mips-dspr2-asm.S @@ -749,6 +749,645 @@ LEAF_MIPS_DSPR2(pixman_composite_over_n_8_0565_asm_mips) END(pixman_composite_over_n_8_0565_asm_mips) +LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_8888_8888_SRC_asm_mips) +/* + * a0 - *dst + * a1 - *src_top + * a2 - *src_bottom + * a3 - w + * 16(sp) - wt + * 20(sp) - wb + * 24(sp) - vx + * 28(sp) - unit_x + */ + + beqz a3, 1f + nop + + SAVE_REGS_ON_STACK 20, v0, s0, s1, s2, s3, s4, s5, s6, s7 + + lw s0, 36(sp) /* s0 = wt */ + lw s1, 40(sp) /* s1 = wb */ + lw s2, 44(sp) /* s2 = vx */ + lw s3, 48(sp) /* s3 = unit_x */ + li v0, BILINEAR_INTERPOLATION_RANGE + + sll s0, s0, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) + sll s1, s1, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) +0: + andi t4, s2, 0xffff /* t4 = (short)vx */ + srl t4, t4, (16 - BILINEAR_INTERPOLATION_BITS) /* t4 = vx >> 8 */ + subu t5, v0, t4 /* t5 = ( 256 - (vx>>8)) */ + + mul s4, s0, t5 /* s4 = wt*(256-(vx>>8)) */ + mul s5, s0, t4 /* s5 = wt*(vx>>8) */ + mul s6, s1, t5 /* s6 = wb*(256-(vx>>8)) */ + mul s7, s1, t4 /* s7 = wb*(vx>>8) */ + + sra t9, s2, 16 + sll t9, t9, 2 + addiu t8, t9, 4 + lwx t0, t9(a1) /* t0 = tl */ + lwx t1, t8(a1) /* t1 = tr */ + addiu a3, a3, -1 + lwx t2, t9(a2) /* t2 = bl */ + lwx t3, t8(a2) /* t3 = br */ + + BILINEAR_INTERPOLATE_SINGLE_PIXEL t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, s4, s5, s6, s7 + + addu s2, s2, s3 /* vx += unit_x; */ + sw t0, 0(a0) + bnez a3, 0b + addiu a0, a0, 4 + + RESTORE_REGS_FROM_STACK 20, v0, s0, s1, s2, s3, s4, s5, s6, s7 +1: + j ra + nop + +END(pixman_scaled_bilinear_scanline_8888_8888_SRC_asm_mips) + +LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_8888_0565_SRC_asm_mips) +/* + * a0 - *dst + * a1 - *src_top + * a2 - *src_bottom + * a3 - w + * 16(sp) - wt + * 20(sp) - wb + * 24(sp) - vx + * 28(sp) - unit_x + */ + + beqz a3, 1f + nop + + SAVE_REGS_ON_STACK 20, v0, s0, s1, s2, s3, s4, s5, s6, s7 + + lw s0, 36(sp) /* s0 = wt */ + lw s1, 40(sp) /* s1 = wb */ + lw s2, 44(sp) /* s2 = vx */ + lw s3, 48(sp) /* s3 = unit_x */ + li v0, BILINEAR_INTERPOLATION_RANGE + + sll s0, s0, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) + sll s1, s1, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) +0: + andi t4, s2, 0xffff /* t4 = (short)vx */ + srl t4, t4, (16 - BILINEAR_INTERPOLATION_BITS) /* t4 = vx >> 8 */ + subu t5, v0, t4 /* t5 = ( 256 - (vx>>8)) */ + + mul s4, s0, t5 /* s4 = wt*(256-(vx>>8)) */ + mul s5, s0, t4 /* s5 = wt*(vx>>8) */ + mul s6, s1, t5 /* s6 = wb*(256-(vx>>8)) */ + mul s7, s1, t4 /* s7 = wb*(vx>>8) */ + + sra t9, s2, 16 + sll t9, t9, 2 + addiu t8, t9, 4 + lwx t0, t9(a1) /* t0 = tl */ + lwx t1, t8(a1) /* t1 = tr */ + addiu a3, a3, -1 + lwx t2, t9(a2) /* t2 = bl */ + lwx t3, t8(a2) /* t3 = br */ + + BILINEAR_INTERPOLATE_SINGLE_PIXEL t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, s4, s5, s6, s7 + CONVERT_1x8888_TO_1x0565 t0, t1, t2, t3 + + addu s2, s2, s3 /* vx += unit_x; */ + sh t1, 0(a0) + bnez a3, 0b + addiu a0, a0, 2 + + RESTORE_REGS_FROM_STACK 20, v0, s0, s1, s2, s3, s4, s5, s6, s7 +1: + j ra + nop + +END(pixman_scaled_bilinear_scanline_8888_0565_SRC_asm_mips) + +LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_0565_8888_SRC_asm_mips) +/* + * a0 - *dst + * a1 - *src_top + * a2 - *src_bottom + * a3 - w + * 16(sp) - wt + * 20(sp) - wb + * 24(sp) - vx + * 28(sp) - unit_x + */ + + beqz a3, 1f + nop + + SAVE_REGS_ON_STACK 28, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8 + + lw s0, 44(sp) /* s0 = wt */ + lw s1, 48(sp) /* s1 = wb */ + lw s2, 52(sp) /* s2 = vx */ + lw s3, 56(sp) /* s3 = unit_x */ + li v0, BILINEAR_INTERPOLATION_RANGE + li v1, 0x07e007e0 + li s8, 0x001f001f + + sll s0, s0, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) + sll s1, s1, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) +0: + andi t4, s2, 0xffff /* t4 = (short)vx */ + srl t4, t4, (16 - BILINEAR_INTERPOLATION_BITS) /* t4 = vx >> 8 */ + subu t5, v0, t4 /* t5 = ( 256 - (vx>>8)) */ + + mul s4, s0, t5 /* s4 = wt*(256-(vx>>8)) */ + mul s5, s0, t4 /* s5 = wt*(vx>>8) */ + mul s6, s1, t5 /* s6 = wb*(256-(vx>>8)) */ + mul s7, s1, t4 /* s7 = wb*(vx>>8) */ + + sra t9, s2, 16 + sll t9, t9, 1 + addiu t8, t9, 2 + lhx t0, t9(a1) /* t0 = tl */ + lhx t1, t8(a1) /* t1 = tr */ + andi t1, t1, 0xffff + addiu a3, a3, -1 + lhx t2, t9(a2) /* t2 = bl */ + lhx t3, t8(a2) /* t3 = br */ + andi t3, t3, 0xffff + + CONVERT_2x0565_TO_2x8888 t0, t1, t0, t1, v1, s8, t4, t5, t6, t7 + CONVERT_2x0565_TO_2x8888 t2, t3, t2, t3, v1, s8, t4, t5, t6, t7 + BILINEAR_INTERPOLATE_SINGLE_PIXEL t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, s4, s5, s6, s7 + + addu s2, s2, s3 /* vx += unit_x; */ + sw t0, 0(a0) + bnez a3, 0b + addiu a0, a0, 4 + + RESTORE_REGS_FROM_STACK 28, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8 +1: + j ra + nop + +END(pixman_scaled_bilinear_scanline_0565_8888_SRC_asm_mips) + +LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_0565_0565_SRC_asm_mips) +/* + * a0 - *dst + * a1 - *src_top + * a2 - *src_bottom + * a3 - w + * 16(sp) - wt + * 20(sp) - wb + * 24(sp) - vx + * 28(sp) - unit_x + */ + + beqz a3, 1f + nop + + SAVE_REGS_ON_STACK 28, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8 + + lw s0, 44(sp) /* s0 = wt */ + lw s1, 48(sp) /* s1 = wb */ + lw s2, 52(sp) /* s2 = vx */ + lw s3, 56(sp) /* s3 = unit_x */ + li v0, BILINEAR_INTERPOLATION_RANGE + li v1, 0x07e007e0 + li s8, 0x001f001f + + sll s0, s0, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) + sll s1, s1, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) +0: + andi t4, s2, 0xffff /* t4 = (short)vx */ + srl t4, t4, (16 - BILINEAR_INTERPOLATION_BITS) /* t4 = vx >> 8 */ + subu t5, v0, t4 /* t5 = ( 256 - (vx>>8)) */ + + mul s4, s0, t5 /* s4 = wt*(256-(vx>>8)) */ + mul s5, s0, t4 /* s5 = wt*(vx>>8) */ + mul s6, s1, t5 /* s6 = wb*(256-(vx>>8)) */ + mul s7, s1, t4 /* s7 = wb*(vx>>8) */ + + sra t9, s2, 16 + sll t9, t9, 1 + addiu t8, t9, 2 + lhx t0, t9(a1) /* t0 = tl */ + lhx t1, t8(a1) /* t1 = tr */ + andi t1, t1, 0xffff + addiu a3, a3, -1 + lhx t2, t9(a2) /* t2 = bl */ + lhx t3, t8(a2) /* t3 = br */ + andi t3, t3, 0xffff + + CONVERT_2x0565_TO_2x8888 t0, t1, t0, t1, v1, s8, t4, t5, t6, t7 + CONVERT_2x0565_TO_2x8888 t2, t3, t2, t3, v1, s8, t4, t5, t6, t7 + BILINEAR_INTERPOLATE_SINGLE_PIXEL t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, s4, s5, s6, s7 + CONVERT_1x8888_TO_1x0565 t0, t1, t2, t3 + + addu s2, s2, s3 /* vx += unit_x; */ + sh t1, 0(a0) + bnez a3, 0b + addiu a0, a0, 2 + + RESTORE_REGS_FROM_STACK 28, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8 +1: + j ra + nop + +END(pixman_scaled_bilinear_scanline_0565_0565_SRC_asm_mips) + +LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_8888_8888_OVER_asm_mips) +/* + * a0 - *dst + * a1 - *src_top + * a2 - *src_bottom + * a3 - w + * 16(sp) - wt + * 20(sp) - wb + * 24(sp) - vx + * 28(sp) - unit_x + */ + + beqz a3, 1f + nop + + SAVE_REGS_ON_STACK 24, v0, s0, s1, s2, s3, s4, s5, s6, s7, s8 + + lw s0, 40(sp) /* s0 = wt */ + lw s1, 44(sp) /* s1 = wb */ + lw s2, 48(sp) /* s2 = vx */ + lw s3, 52(sp) /* s3 = unit_x */ + li v0, BILINEAR_INTERPOLATION_RANGE + li s8, 0x00ff00ff + + sll s0, s0, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) + sll s1, s1, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) +0: + andi t4, s2, 0xffff /* t4 = (short)vx */ + srl t4, t4, (16 - BILINEAR_INTERPOLATION_BITS) /* t4 = vx >> 8 */ + subu t5, v0, t4 /* t5 = ( 256 - (vx>>8)) */ + + mul s4, s0, t5 /* s4 = wt*(256-(vx>>8)) */ + mul s5, s0, t4 /* s5 = wt*(vx>>8) */ + mul s6, s1, t5 /* s6 = wb*(256-(vx>>8)) */ + mul s7, s1, t4 /* s7 = wb*(vx>>8) */ + + sra t9, s2, 16 + sll t9, t9, 2 + addiu t8, t9, 4 + lwx t0, t9(a1) /* t0 = tl */ + lwx t1, t8(a1) /* t1 = tr */ + addiu a3, a3, -1 + lwx t2, t9(a2) /* t2 = bl */ + lwx t3, t8(a2) /* t3 = br */ + + BILINEAR_INTERPOLATE_SINGLE_PIXEL t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, s4, s5, s6, s7 + lw t1, 0(a0) /* t1 = dest */ + OVER_8888_8888 t0, t1, t2, s8, t3, t4, t5, t6 + + addu s2, s2, s3 /* vx += unit_x; */ + sw t2, 0(a0) + bnez a3, 0b + addiu a0, a0, 4 + + RESTORE_REGS_FROM_STACK 24, v0, s0, s1, s2, s3, s4, s5, s6, s7, s8 +1: + j ra + nop + +END(pixman_scaled_bilinear_scanline_8888_8888_OVER_asm_mips) + +LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_8888_8888_ADD_asm_mips) +/* + * a0 - *dst + * a1 - *src_top + * a2 - *src_bottom + * a3 - w + * 16(sp) - wt + * 20(sp) - wb + * 24(sp) - vx + * 28(sp) - unit_x + */ + + beqz a3, 1f + nop + + SAVE_REGS_ON_STACK 20, v0, s0, s1, s2, s3, s4, s5, s6, s7 + + lw s0, 36(sp) /* s0 = wt */ + lw s1, 40(sp) /* s1 = wb */ + lw s2, 44(sp) /* s2 = vx */ + lw s3, 48(sp) /* s3 = unit_x */ + li v0, BILINEAR_INTERPOLATION_RANGE + + sll s0, s0, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) + sll s1, s1, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) +0: + andi t4, s2, 0xffff /* t4 = (short)vx */ + srl t4, t4, (16 - BILINEAR_INTERPOLATION_BITS) /* t4 = vx >> 8 */ + subu t5, v0, t4 /* t5 = ( 256 - (vx>>8)) */ + + mul s4, s0, t5 /* s4 = wt*(256-(vx>>8)) */ + mul s5, s0, t4 /* s5 = wt*(vx>>8) */ + mul s6, s1, t5 /* s6 = wb*(256-(vx>>8)) */ + mul s7, s1, t4 /* s7 = wb*(vx>>8) */ + + sra t9, s2, 16 + sll t9, t9, 2 + addiu t8, t9, 4 + lwx t0, t9(a1) /* t0 = tl */ + lwx t1, t8(a1) /* t1 = tr */ + addiu a3, a3, -1 + lwx t2, t9(a2) /* t2 = bl */ + lwx t3, t8(a2) /* t3 = br */ + + BILINEAR_INTERPOLATE_SINGLE_PIXEL t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, s4, s5, s6, s7 + lw t1, 0(a0) + addu_s.qb t2, t0, t1 + + addu s2, s2, s3 /* vx += unit_x; */ + sw t2, 0(a0) + bnez a3, 0b + addiu a0, a0, 4 + + RESTORE_REGS_FROM_STACK 20, v0, s0, s1, s2, s3, s4, s5, s6, s7 +1: + j ra + nop + +END(pixman_scaled_bilinear_scanline_8888_8888_ADD_asm_mips) + +LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_8888_8_8888_SRC_asm_mips) +/* + * a0 - *dst + * a1 - *mask + * a2 - *src_top + * a3 - *src_bottom + * 16(sp) - wt + * 20(sp) - wb + * 24(sp) - vx + * 28(sp) - unit_x + * 32(sp) - w + */ + + lw v1, 32(sp) + beqz v1, 1f + nop + + SAVE_REGS_ON_STACK 28, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8 + + lw s0, 44(sp) /* s0 = wt */ + lw s1, 48(sp) /* s1 = wb */ + lw s2, 52(sp) /* s2 = vx */ + lw s3, 56(sp) /* s3 = unit_x */ + li v0, BILINEAR_INTERPOLATION_RANGE + li s8, 0x00ff00ff + + sll s0, s0, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) + sll s1, s1, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) +0: + andi t4, s2, 0xffff /* t4 = (short)vx */ + srl t4, t4, (16 - BILINEAR_INTERPOLATION_BITS) /* t4 = vx >> 8 */ + subu t5, v0, t4 /* t5 = ( 256 - (vx>>8)) */ + + mul s4, s0, t5 /* s4 = wt*(256-(vx>>8)) */ + mul s5, s0, t4 /* s5 = wt*(vx>>8) */ + mul s6, s1, t5 /* s6 = wb*(256-(vx>>8)) */ + mul s7, s1, t4 /* s7 = wb*(vx>>8) */ + + sra t9, s2, 16 + sll t9, t9, 2 + addiu t8, t9, 4 + lwx t0, t9(a2) /* t0 = tl */ + lwx t1, t8(a2) /* t1 = tr */ + addiu v1, v1, -1 + lwx t2, t9(a3) /* t2 = bl */ + lwx t3, t8(a3) /* t3 = br */ + + BILINEAR_INTERPOLATE_SINGLE_PIXEL t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, s4, s5, s6, s7 + lbu t1, 0(a1) /* t1 = mask */ + addiu a1, a1, 1 + MIPS_UN8x4_MUL_UN8 t0, t1, t0, s8, t2, t3, t4 + + addu s2, s2, s3 /* vx += unit_x; */ + sw t0, 0(a0) + bnez v1, 0b + addiu a0, a0, 4 + + RESTORE_REGS_FROM_STACK 28, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8 +1: + j ra + nop + +END(pixman_scaled_bilinear_scanline_8888_8_8888_SRC_asm_mips) + +LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_8888_8_0565_SRC_asm_mips) +/* + * a0 - *dst + * a1 - *mask + * a2 - *src_top + * a3 - *src_bottom + * 16(sp) - wt + * 20(sp) - wb + * 24(sp) - vx + * 28(sp) - unit_x + * 32(sp) - w + */ + + lw v1, 32(sp) + beqz v1, 1f + nop + + SAVE_REGS_ON_STACK 28, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8 + + lw s0, 44(sp) /* s0 = wt */ + lw s1, 48(sp) /* s1 = wb */ + lw s2, 52(sp) /* s2 = vx */ + lw s3, 56(sp) /* s3 = unit_x */ + li v0, BILINEAR_INTERPOLATION_RANGE + li s8, 0x00ff00ff + + sll s0, s0, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) + sll s1, s1, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) +0: + andi t4, s2, 0xffff /* t4 = (short)vx */ + srl t4, t4, (16 - BILINEAR_INTERPOLATION_BITS) /* t4 = vx >> 8 */ + subu t5, v0, t4 /* t5 = ( 256 - (vx>>8)) */ + + mul s4, s0, t5 /* s4 = wt*(256-(vx>>8)) */ + mul s5, s0, t4 /* s5 = wt*(vx>>8) */ + mul s6, s1, t5 /* s6 = wb*(256-(vx>>8)) */ + mul s7, s1, t4 /* s7 = wb*(vx>>8) */ + + sra t9, s2, 16 + sll t9, t9, 2 + addiu t8, t9, 4 + lwx t0, t9(a2) /* t0 = tl */ + lwx t1, t8(a2) /* t1 = tr */ + addiu v1, v1, -1 + lwx t2, t9(a3) /* t2 = bl */ + lwx t3, t8(a3) /* t3 = br */ + + BILINEAR_INTERPOLATE_SINGLE_PIXEL t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, s4, s5, s6, s7 + lbu t1, 0(a1) /* t1 = mask */ + addiu a1, a1, 1 + MIPS_UN8x4_MUL_UN8 t0, t1, t0, s8, t2, t3, t4 + CONVERT_1x8888_TO_1x0565 t0, t1, t2, t3 + + addu s2, s2, s3 /* vx += unit_x; */ + sh t1, 0(a0) + bnez v1, 0b + addiu a0, a0, 2 + + RESTORE_REGS_FROM_STACK 28, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8 +1: + j ra + nop + +END(pixman_scaled_bilinear_scanline_8888_8_0565_SRC_asm_mips) + +LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_0565_8_x888_SRC_asm_mips) +/* + * a0 - *dst + * a1 - *mask + * a2 - *src_top + * a3 - *src_bottom + * 16(sp) - wt + * 20(sp) - wb + * 24(sp) - vx + * 28(sp) - unit_x + * 32(sp) - w + */ + + lw t0, 32(sp) + beqz t0, 1f + nop + + SAVE_REGS_ON_STACK 32, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8, ra + + lw s0, 48(sp) /* s0 = wt */ + lw s1, 52(sp) /* s1 = wb */ + lw s2, 56(sp) /* s2 = vx */ + lw s3, 60(sp) /* s3 = unit_x */ + lw ra, 64(sp) /* ra = w */ + li v0, 0x00ff00ff + li v1, 0x07e007e0 + li s8, 0x001f001f + + sll s0, s0, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) + sll s1, s1, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) +0: + andi t4, s2, 0xffff /* t4 = (short)vx */ + srl t4, t4, (16 - BILINEAR_INTERPOLATION_BITS) /* t4 = vx >> 8 */ + li t5, BILINEAR_INTERPOLATION_RANGE + subu t5, t5, t4 /* t5 = ( 256 - (vx>>8)) */ + + mul s4, s0, t5 /* s4 = wt*(256-(vx>>8)) */ + mul s5, s0, t4 /* s5 = wt*(vx>>8) */ + mul s6, s1, t5 /* s6 = wb*(256-(vx>>8)) */ + mul s7, s1, t4 /* s7 = wb*(vx>>8) */ + + sra t9, s2, 16 + sll t9, t9, 1 + addiu t8, t9, 2 + lhx t0, t9(a2) /* t0 = tl */ + lhx t1, t8(a2) /* t1 = tr */ + andi t1, t1, 0xffff + addiu ra, ra, -1 + lhx t2, t9(a3) /* t2 = bl */ + lhx t3, t8(a3) /* t3 = br */ + andi t3, t3, 0xffff + + CONVERT_2x0565_TO_2x8888 t0, t1, t0, t1, v1, s8, t4, t5, t6, t7 + CONVERT_2x0565_TO_2x8888 t2, t3, t2, t3, v1, s8, t4, t5, t6, t7 + BILINEAR_INTERPOLATE_SINGLE_PIXEL t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, s4, s5, s6, s7 + lbu t1, 0(a1) /* t1 = mask */ + addiu a1, a1, 1 + MIPS_UN8x4_MUL_UN8 t0, t1, t0, v0, t2, t3, t4 + + addu s2, s2, s3 /* vx += unit_x; */ + sw t0, 0(a0) + bnez ra, 0b + addiu a0, a0, 4 + + RESTORE_REGS_FROM_STACK 32, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8, ra +1: + j ra + nop + +END(pixman_scaled_bilinear_scanline_0565_8_x888_SRC_asm_mips) + +LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_0565_8_0565_SRC_asm_mips) +/* + * a0 - *dst + * a1 - *mask + * a2 - *src_top + * a3 - *src_bottom + * 16(sp) - wt + * 20(sp) - wb + * 24(sp) - vx + * 28(sp) - unit_x + * 32(sp) - w + */ + + lw t0, 32(sp) + beqz t0, 1f + nop + + SAVE_REGS_ON_STACK 32, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8, ra + + lw s0, 48(sp) /* s0 = wt */ + lw s1, 52(sp) /* s1 = wb */ + lw s2, 56(sp) /* s2 = vx */ + lw s3, 60(sp) /* s3 = unit_x */ + lw ra, 64(sp) /* ra = w */ + li v0, 0x00ff00ff + li v1, 0x07e007e0 + li s8, 0x001f001f + + sll s0, s0, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) + sll s1, s1, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) +0: + andi t4, s2, 0xffff /* t4 = (short)vx */ + srl t4, t4, (16 - BILINEAR_INTERPOLATION_BITS) /* t4 = vx >> 8 */ + li t5, BILINEAR_INTERPOLATION_RANGE + subu t5, t5, t4 /* t5 = ( 256 - (vx>>8)) */ + + mul s4, s0, t5 /* s4 = wt*(256-(vx>>8)) */ + mul s5, s0, t4 /* s5 = wt*(vx>>8) */ + mul s6, s1, t5 /* s6 = wb*(256-(vx>>8)) */ + mul s7, s1, t4 /* s7 = wb*(vx>>8) */ + + sra t9, s2, 16 + sll t9, t9, 1 + addiu t8, t9, 2 + lhx t0, t9(a2) /* t0 = tl */ + lhx t1, t8(a2) /* t1 = tr */ + andi t1, t1, 0xffff + addiu ra, ra, -1 + lhx t2, t9(a3) /* t2 = bl */ + lhx t3, t8(a3) /* t3 = br */ + andi t3, t3, 0xffff + + CONVERT_2x0565_TO_2x8888 t0, t1, t0, t1, v1, s8, t4, t5, t6, t7 + CONVERT_2x0565_TO_2x8888 t2, t3, t2, t3, v1, s8, t4, t5, t6, t7 + BILINEAR_INTERPOLATE_SINGLE_PIXEL t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, s4, s5, s6, s7 + lbu t1, 0(a1) /* t1 = mask */ + addiu a1, a1, 1 + MIPS_UN8x4_MUL_UN8 t0, t1, t0, v0, t2, t3, t4 + CONVERT_1x8888_TO_1x0565 t0, t1, t2, t3 + + addu s2, s2, s3 /* vx += unit_x; */ + sh t1, 0(a0) + bnez ra, 0b + addiu a0, a0, 2 + + RESTORE_REGS_FROM_STACK 32, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8, ra +1: + j ra + nop + +END(pixman_scaled_bilinear_scanline_0565_8_0565_SRC_asm_mips) + LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_8888_8_8888_OVER_asm_mips) /* * a0 - dst (a8r8g8b8) @@ -815,3 +1454,68 @@ LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_8888_8_8888_OVER_asm_mips) nop END(pixman_scaled_bilinear_scanline_8888_8_8888_OVER_asm_mips) + +LEAF_MIPS_DSPR2(pixman_scaled_bilinear_scanline_8888_8_8888_ADD_asm_mips) +/* + * a0 - *dst + * a1 - *mask + * a2 - *src_top + * a3 - *src_bottom + * 16(sp) - wt + * 20(sp) - wb + * 24(sp) - vx + * 28(sp) - unit_x + * 32(sp) - w + */ + + lw v1, 32(sp) + beqz v1, 1f + nop + + SAVE_REGS_ON_STACK 28, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8 + + lw s0, 44(sp) /* s0 = wt */ + lw s1, 48(sp) /* s1 = wb */ + lw s2, 52(sp) /* s2 = vx */ + lw s3, 56(sp) /* s3 = unit_x */ + li v0, BILINEAR_INTERPOLATION_RANGE + li s8, 0x00ff00ff + + sll s0, s0, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) + sll s1, s1, (2 * (8 - BILINEAR_INTERPOLATION_BITS)) +0: + andi t4, s2, 0xffff /* t4 = (short)vx */ + srl t4, t4, (16 - BILINEAR_INTERPOLATION_BITS) /* t4 = vx >> 8 */ + subu t5, v0, t4 /* t5 = ( 256 - (vx>>8)) */ + + mul s4, s0, t5 /* s4 = wt*(256-(vx>>8)) */ + mul s5, s0, t4 /* s5 = wt*(vx>>8) */ + mul s6, s1, t5 /* s6 = wb*(256-(vx>>8)) */ + mul s7, s1, t4 /* s7 = wb*(vx>>8) */ + + sra t9, s2, 16 + sll t9, t9, 2 + addiu t8, t9, 4 + lwx t0, t9(a2) /* t0 = tl */ + lwx t1, t8(a2) /* t1 = tr */ + addiu v1, v1, -1 + lwx t2, t9(a3) /* t2 = bl */ + lwx t3, t8(a3) /* t3 = br */ + + BILINEAR_INTERPOLATE_SINGLE_PIXEL t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, s4, s5, s6, s7 + lbu t1, 0(a1) /* t1 = mask */ + lw t2, 0(a0) /* t2 = dst */ + addiu a1, a1, 1 + MIPS_UN8x4_MUL_UN8_ADD_UN8x4 t0, t1, t2, t0, s8, t3, t4, t5 + + addu s2, s2, s3 /* vx += unit_x; */ + sw t0, 0(a0) + bnez v1, 0b + addiu a0, a0, 4 + + RESTORE_REGS_FROM_STACK 28, v0, v1, s0, s1, s2, s3, s4, s5, s6, s7, s8 +1: + j ra + nop + +END(pixman_scaled_bilinear_scanline_8888_8_8888_ADD_asm_mips) |