void ADD_SUFF(IntArgbBmToIntArgbConvert)(BLIT_PARAMS)
{
mlib_s32 dstScan = pDstInfo->scanStride;
mlib_s32 srcScan = pSrcInfo->scanStride;
mlib_d64 dd, dmask, dFF;
mlib_s32 i, i0, j, x, mask;
if (dstScan == 4*width && srcScan == 4*width) {
width *= height;
height = 1;
}
dmask = vis_to_double_dup(0xFFFFFF);
dFF = vis_to_double_dup(0xFFFFFFFF);
for (j = 0; j < height; j++) {
mlib_s32 *src = srcBase;
mlib_s32 *dst = dstBase;
i = i0 = 0;
if ((mlib_s32)dst & 7) {
x = src[i];
dst[i] = (x << 7) >> 7;
i0 = 1;
}
#pragma pipeloop(0)
for (i = i0; i <= (mlib_s32)width - 2; i += 2) {
mlib_u8 *pp0 = (mlib_u8*)(src + i);
mlib_u8 *pp1 = (mlib_u8*)(src + i + 1);
dd = vis_freg_pair(*(mlib_f32*)pp0, *(mlib_f32*)pp1);
dd = vis_fand(dd, dmask);
#if 1
mask = ((*pp0 & 1) << 7) | ((*pp1 & 1) << 3);
*(mlib_d64*)(dst + i) = dd;
vis_pst_8(dFF, dst + i, mask);
#else
mask = ((*pp0 & 1) << 1) | (*pp1 & 1);
dd = vis_for(dd, ((mlib_d64*)vis_amask_arr)[mask]);
*(mlib_d64*)(dst + i) = dd;
#endif
}
if (i < width) {
x = src[i];
dst[i] = (x << 7) >> 7;
}
mlib_status
__mlib_SignalEmphasize_S16S_S16S_Sat(
mlib_s16 *dst,
const mlib_s16 *src,
void *filter,
mlib_s32 n)
{
mlib_emphasize_struct *fist = filter;
mlib_d64 w_maskand0 = vis_to_double(0xFFFFFFFF, 0xFFFF);
mlib_d64 w_maskor0 = vis_freg_pair(0.f, fist->v16_last0);
mlib_d64 w_maskand1 = vis_to_double(0xFFFFFFFF, 0xFFFF0000);
mlib_d64 w_maskor1 = vis_freg_pair(0.f, fist->v16_last1);
mlib_f32 v_mask = vis_to_float(0x80008000);
mlib_f32 v_alpha = fist->v_alpha;
mlib_s16 *fdst = dst + n + n - 1;
mlib_d64 *dpd, *dps, *dsrct1;
mlib_d64 w_dst, w_src, w_src0, w_src1, w_src2, w_lsrc;
mlib_d64 dr0, dr1, dr2, dr3, dr4, dr5, dr6, dr7;
mlib_s32 i, times, t1, t2;
/* check for obvious errors */
if ((fist == NULL) || (n <= 0) || (src == 0) || (dst == 0) ||
(fist->type != MLIB_EMPH)) {
return (MLIB_FAILURE);
}
vis_write_gsr(1 << 3);
w_maskor0 = vis_fand(w_maskor0, w_maskand1);
w_maskor1 = vis_fand(w_maskor1, w_maskand0);
vis_alignaddr((void *)(-(mlib_addr)src), 0);
w_maskand0 = vis_faligndata(w_maskand0, w_maskand0);
w_maskor0 = vis_faligndata(w_maskor0, w_maskor0);
w_maskand1 = vis_faligndata(w_maskand1, w_maskand1);
w_maskor1 = vis_faligndata(w_maskor1, w_maskor1);
dpd = vis_alignaddr(dst, 0);
times = (mlib_d64 *)vis_alignaddr(fdst, 0) - dpd;
t1 = -((mlib_addr)(dst) & 7);
t2 = t1 - 4;
dps = vis_alignaddr((void *)src, t2);
w_src0 = vis_ld_d64_nf(dps);
dps++;
w_src1 = vis_ld_d64_nf(dps);
dps++;
if ((((mlib_addr)dst ^ (mlib_addr)src) & 7)) {
if (((mlib_addr)dps - (mlib_addr)src) >= 6) {
w_src0 = vis_fand(w_maskand0, w_src0);
w_src0 = vis_for(w_maskor0, w_src0);
} else {
w_src1 = vis_fand(w_maskand0, w_src1);
w_src1 = vis_for(w_maskor0, w_src1);
}
if (((mlib_addr)dps - (mlib_addr)src) >= 8) {
w_src0 = vis_fand(w_maskand1, w_src0);
w_src0 = vis_for(w_maskor1, w_src0);
} else {
w_src1 = vis_fand(w_maskand1, w_src1);
w_src1 = vis_for(w_maskor1, w_src1);
}
w_lsrc = vis_faligndata(w_src0, w_src1);
dsrct1 = vis_alignaddr((void *)src, t1);
if (dps - 2 != dsrct1) {
w_src2 = *dps;
dps++;
w_src = vis_faligndata(w_src1, w_src2);
MLIB_MUL8;
if ((mlib_addr)dst & 7) {
times--;
w_src0 = w_src1;
w_src1 = w_src2;
w_src2 = *dps;
vis_alignaddr((void *)src, t2);
w_lsrc = vis_faligndata(w_src0, w_src1);
vis_alignaddr((void *)src, t1);
w_src = vis_faligndata(w_src1, w_src2);
dps++;
MLIB_MIX;
w_dst = vis_fpackfix_pair(dr2, dr3);
vis_pst_16(w_dst, dpd, vis_edge16(dst, fdst));
dpd++;
}
w_src0 = w_src1;
w_src1 = w_src2;
w_src2 = vis_ld_d64_nf(dps);
vis_alignaddr((void *)src, t2);
w_lsrc = vis_faligndata(w_src0, w_src1);
vis_alignaddr((void *)src, t1);
w_src = vis_faligndata(w_src1, w_src2);
MLIB_MIX;
w_dst = vis_fpackfix_pair(dr2, dr3);
dps++;
w_src0 = w_src1;
w_src1 = w_src2;
w_src2 = vis_ld_d64_nf(dps);
vis_alignaddr((void *)src, t2);
w_lsrc = vis_faligndata(w_src0, w_src1);
vis_alignaddr((void *)src, t1);
w_src = vis_faligndata(w_src1, w_src2);
dps++;
for (i = 0; i < times; i++) {
*dpd = w_dst;
MLIB_MIX;
w_dst = vis_fpackfix_pair(dr2, dr3);
w_src0 = w_src1;
w_src1 = w_src2;
w_src2 = vis_ld_d64_nf(dps);
vis_alignaddr((void *)src, t2);
w_lsrc = vis_faligndata(w_src0, w_src1);
vis_alignaddr((void *)src, t1);
w_src = vis_faligndata(w_src1, w_src2);
dpd++;
dps++;
}
} else {
w_src = vis_faligndata(w_src0, w_src1);
MLIB_MUL8;
if ((mlib_addr)dst & 7) {
times--;
w_src0 = w_src1;
w_src1 = vis_ld_d64_nf(dps);
vis_alignaddr((void *)src, t2);
w_lsrc = vis_faligndata(w_src0, w_src1);
vis_alignaddr((void *)src, t1);
w_src = vis_faligndata(w_src0, w_src1);
dps++;
MLIB_MIX;
w_dst = vis_fpackfix_pair(dr2, dr3);
vis_pst_16(w_dst, dpd, vis_edge16(dst, fdst));
dpd++;
}
w_src0 = w_src1;
w_src1 = vis_ld_d64_nf(dps);
vis_alignaddr((void *)src, t2);
w_lsrc = vis_faligndata(w_src0, w_src1);
vis_alignaddr((void *)src, t1);
w_src = vis_faligndata(w_src0, w_src1);
MLIB_MIX;
w_dst = vis_fpackfix_pair(dr2, dr3);
dps++;
w_src0 = w_src1;
w_src1 = vis_ld_d64_nf(dps);
vis_alignaddr((void *)src, t2);
w_lsrc = vis_faligndata(w_src0, w_src1);
vis_alignaddr((void *)src, t1);
w_src = vis_faligndata(w_src0, w_src1);
dps++;
for (i = 0; i < times; i++) {
*dpd = w_dst;
MLIB_MIX;
w_dst = vis_fpackfix_pair(dr2, dr3);
w_src0 = w_src1;
w_src1 = vis_ld_d64_nf(dps);
vis_alignaddr((void *)src, t2);
w_lsrc = vis_faligndata(w_src0, w_src1);
vis_alignaddr((void *)src, t1);
w_src = vis_faligndata(w_src0, w_src1);
dps++;
dpd++;
}
}
} else {
w_src = w_src1;
if ((mlib_addr)src & 7) {
times--;
if (((mlib_addr)src & 7) == 2) {
w_src0 = vis_fand(w_maskand0, w_src0);
w_src0 = vis_for(w_maskor0, w_src0);
} else {
w_src1 = vis_fand(w_maskand0, w_src1);
w_src1 = vis_for(w_maskor0, w_src1);
}
w_src1 = vis_fand(w_maskand1, w_src1);
w_src1 = vis_for(w_maskor1, w_src1);
w_lsrc = vis_faligndata(w_src0, w_src1);
MLIB_MUL8;
w_src0 = w_src1;
w_src1 = *dps;
w_src = w_src1;
w_lsrc = vis_faligndata(w_src0, w_src1);
dps++;
MLIB_MIX;
w_dst = vis_fpackfix_pair(dr2, dr3);
vis_pst_16(w_dst, dpd, vis_edge16(dst, fdst));
dpd++;
} else {
w_src0 = vis_fand(w_maskand0, w_src0);
w_src0 = vis_for(w_maskor0, w_src0);
w_src0 = vis_fand(w_maskand1, w_src0);
w_src0 = vis_for(w_maskor1, w_src0);
w_lsrc = vis_faligndata(w_src0, w_src1);
MLIB_MUL8;
}
w_src = vis_ld_d64_nf(dps);
w_lsrc = vis_faligndata(w_src1, w_src);
MLIB_MIX;
w_src1 = w_src;
w_dst = vis_fpackfix_pair(dr2, dr3);
dps++;
w_src = vis_ld_d64_nf(dps);
w_lsrc = vis_faligndata(w_src1, w_src);
dps++;
for (i = 0; i < times; i++) {
*dpd = w_dst;
MLIB_MIX;
w_src1 = w_src;
w_src = vis_ld_d64_nf(dps);
w_lsrc = vis_faligndata(w_src1, w_src);
w_dst = vis_fpackfix_pair(dr2, dr3);
dps++;
dpd++;
}
}
if (times >= 0) {
vis_pst_16(w_dst, dpd, vis_edge16(dpd, fdst));
}
((mlib_s16 *)&fist->v16_last0)[0] = src[2 * n - 2];
((mlib_s16 *)&fist->v16_last1)[1] = src[2 * n - 1];
return (MLIB_SUCCESS);
}