1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
diff --git a/apps/plugins/bench_scaler.c b/apps/plugins/bench_scaler.c
index c24807d..47ec18a 100644
--- a/apps/plugins/bench_scaler.c
+++ b/apps/plugins/bench_scaler.c
@@ -49,8 +49,8 @@ static void output_row_null(uint32_t row, void * row_in,
 #else
     uint32_t *lim = in + ctx->bm->width;
 #endif
-    for (; in < lim; in++)
-        output = SC_MUL(*in + ctx->round, ctx->divisor);
+    while (in < lim)
+        output = SC_OUT(*in++);
     return;
 }

diff --git a/apps/plugins/lib/grey_draw.c b/apps/plugins/lib/grey_draw.c
index 6315ad9..521643c 100644
--- a/apps/plugins/lib/grey_draw.c
+++ b/apps/plugins/lib/grey_draw.c
@@ -733,7 +733,7 @@ static void output_row_grey_32(uint32_t row, void * row_in,
     uint32_t *qp = (uint32_t*)row_in;
     uint8_t *dest = (uint8_t*)ctx->bm->data + ctx->bm->width * row;
     for (col = 0; col < ctx->bm->width; col++)
-        *dest++ = SC_MUL((*qp++) + ctx->round,ctx->divisor);
+        *dest++ = SC_OUT(*qp++);
 }

 static unsigned int get_size_grey(struct bitmap *bm)
diff --git a/apps/plugins/pictureflow/pictureflow.c b/apps/plugins/pictureflow/pictureflow.c
index c678254..2ad74df 100644
--- a/apps/plugins/pictureflow/pictureflow.c
+++ b/apps/plugins/pictureflow/pictureflow.c
@@ -634,20 +634,19 @@ static void output_row_32_transposed(uint32_t row, void * row_in,
 #ifdef USEGSLIB
     uint32_t *qp = (uint32_t*)row_in;
     for (; dest < end; dest += ctx->bm->height)
-        *dest = SC_MUL((*qp++) + ctx->round, ctx->divisor);
+        *dest = SC_OUT(*qp++);
 #else
     struct uint32_rgb *qp = (struct uint32_rgb*)row_in;
-    uint32_t rb_mul = SCALE_VAL(ctx->divisor, 31),
-             rb_rnd = SCALE_VAL(ctx->round, 31),
-             g_mul = SCALE_VAL(ctx->divisor, 63),
-             g_rnd = SCALE_VAL(ctx->round, 63);
     int r, g, b;
     for (; dest < end; dest += ctx->bm->height)
     {
-        r = SC_MUL(qp->r + rb_rnd, rb_mul);
-        g = SC_MUL(qp->g + g_rnd, g_mul);
-        b = SC_MUL(qp->b + rb_rnd, rb_mul);
+        r = SC_OUT(qp->r);
+        g = SC_OUT(qp->g);
+        b = SC_OUT(qp->b);
         qp++;
+        r = SCALE_VAL(r, 31);
+        g = SCALE_VAL(g, 63);
+        b = SCALE_VAL(b, 31);
         *dest = LCD_RGBPACK_LCD(r,g,b);
     }
 #endif
@@ -663,9 +662,9 @@ static void output_row_32_transposed_fromyuv(uint32_t row, void * row_in,
     for (; dest < end; dest += ctx->bm->height)
     {
         unsigned r, g, b, y, u, v;
-        y = SC_MUL(qp->b + ctx->round, ctx->divisor);
-        u = SC_MUL(qp->g + ctx->round, ctx->divisor);
-        v = SC_MUL(qp->r + ctx->round, ctx->divisor);
+        y = SC_OUT(qp->b);
+        u = SC_OUT(qp->g);
+        v = SC_OUT(qp->r);
         qp++;
         yuv_to_rgb(y, u, v, &r, &g, &b);
         r = (31 * r + (r >> 3) + 127) >> 8;
diff --git a/apps/recorder/resize.c b/apps/recorder/resize.c
index 1e9210e..e6abbe9 100644
--- a/apps/recorder/resize.c
+++ b/apps/recorder/resize.c
@@ -131,14 +131,6 @@ int recalc_dimension(struct dim *dst, struct dim *src)
         return false; \
 }

-/* Set up rounding and scale factors for horizontal area scaler */
-static inline void scale_h_area_setup(struct scaler_context *ctx)
-{
-/* sum is output value * src->width */
-    SDEBUGF("scale_h_area_setup\n");
-    ctx->divisor = ctx->src->width;
-}
-
 /* horizontal area average scaler */
 static bool scale_h_area(void *out_line_ptr,
                          struct scaler_context *ctx, bool accum)
@@ -159,33 +151,86 @@ static bool scale_h_area(void *out_line_ptr,
     mul = 0;
     /* give other tasks a chance to run */
     yield();
+#ifdef CPU_COLDFIRE
+    asm volatile (
+        "move.l #0, %acc0\n\t"
+#ifdef HAVE_LCD_COLOR
+        "move.l #0, %acc1\n\t"
+        "move.l #0, %acc2\n\t"
+#endif
+    );
+#endif
     for (ix = 0; ix < (unsigned int)ctx->src->width; ix++)
     {
-        oxe += ctx->bm->width;
+        oxe += ctx->h_o_val;
         /* end of current area has been reached */
         /* fill buffer if needed */
         FILL_BUF(part,ctx->store_part,ctx->args);
 #ifdef HAVE_LCD_COLOR
-        if (oxe >= (unsigned int)ctx->src->width)
+        if (oxe >= ctx->h_i_val)
         {
             /* "reset" error, which now represents partial coverage of next
                pixel by the next area
             */
-            oxe -= ctx->src->width;
+            oxe -= ctx->h_i_val;

+#ifdef CPU_COLDFIRE
             /* add saved partial pixel from start of area */
-            rgbvalacc.r = rgbvalacc.r * ctx->bm->width + rgbvaltmp.r * mul;
-            rgbvalacc.g = rgbvalacc.g * ctx->bm->width + rgbvaltmp.g * mul;
-            rgbvalacc.b = rgbvalacc.b * ctx->bm->width + rgbvaltmp.b * mul;
+            asm volatile (
+                "mac.l %[racc], %[h_o], %%acc0\n\t"
+                "mac.l %[gacc], %[h_o], %%acc1\n\t"
+                "mac.l %[bacc], %[h_o], %%acc2\n\t"
+                "mac.l %[rtmp], %[mul], %%acc0\n\t"
+                "mac.l %[gtmp], %[mul], %%acc1\n\t"
+                "mac.l %[btmp], %[mul], %%acc2\n\t"
+                :
+                : [racc] "r" (rgbvalacc.r),
+                  [gacc] "r" (rgbvalacc.g),
+                  [bacc] "r" (rgbvalacc.b),
+                  [rtmp] "r" (rgbvaltmp.r),
+                  [gtmp] "r" (rgbvaltmp.g),
+                  [btmp] "r" (rgbvaltmp.b),
+                  [h_o] "r" (ctx->h_o_val),
+                  [mul] "r" (mul)
+            );
+            /* get new pixel , then add its partial coverage to this area */
+            rgbvaltmp.r = part->buf->red;
+            rgbvaltmp.g = part->buf->green;
+            rgbvaltmp.b = part->buf->blue;
+            mul = ctx->h_o_val - oxe;
+            asm volatile (
+                "mac.l %[rtmp], %[mul], %%acc0\n\t"
+                "mac.l %[gtmp], %[mul], %%acc1\n\t"
+                "mac.l %[btmp], %[mul], %%acc2\n\t"
+                "movclr.l %%acc0, %[racc]\n\t"
+                "movclr.l %%acc1, %[gacc]\n\t"
+                "movclr.l %%acc2, %[bacc]\n\t"
+                : [racc] "=r" (rgbvalacc.r),
+                  [gacc] "=r" (rgbvalacc.g),
+                  [bacc] "=r" (rgbvalacc.b)
+                : [rtmp] "r" (rgbvaltmp.r),
+                  [gtmp] "r" (rgbvaltmp.g),
+                  [btmp] "r" (rgbvaltmp.b),
+                  [mul] "r" (mul)
+            );
+#else
+            /* add saved partial pixel from start of area */
+            rgbvalacc.r = rgbvalacc.r * ctx->h_o_val + rgbvaltmp.r * mul;
+            rgbvalacc.g = rgbvalacc.g * ctx->h_o_val + rgbvaltmp.g * mul;
+            rgbvalacc.b = rgbvalacc.b * ctx->h_o_val + rgbvaltmp.b * mul;

             /* get new pixel , then add its partial coverage to this area */
             rgbvaltmp.r = part->buf->red;
             rgbvaltmp.g = part->buf->green;
             rgbvaltmp.b = part->buf->blue;
-            mul = ctx->bm->width - oxe;
+            mul = ctx->h_o_val - oxe;
             rgbvalacc.r += rgbvaltmp.r * mul;
             rgbvalacc.g += rgbvaltmp.g * mul;
             rgbvalacc.b += rgbvaltmp.b * mul;
+#endif
+            rgbvalacc.r = (rgbvalacc.r + (1 << 21)) >> 22;
+            rgbvalacc.g = (rgbvalacc.g + (1 << 21)) >> 22;
+            rgbvalacc.b = (rgbvalacc.b + (1 << 21)) >> 22;
             /* store or accumulate to output row */
             if (accum)
             {
@@ -200,7 +245,7 @@ static bool scale_h_area(void *out_line_ptr,
             rgbvalacc.r = 0;
             rgbvalacc.g = 0;
             rgbvalacc.b = 0;
-            mul = ctx->bm->width - mul;
+            mul = ctx->h_o_val - mul;
             ox += 1;
         /* inside an area */
         } else {
@@ -210,21 +255,45 @@ static bool scale_h_area(void *out_line_ptr,
             rgbvalacc.b += part->buf->blue;
         }
 #else
-        if (oxe >= (unsigned int)ctx->src->width)
+        if (oxe >= ctx->h_i_val)
         {
             /* "reset" error, which now represents partial coverage of next
                pixel by the next area
             */
-            oxe -= ctx->src->width;
+            oxe -= ctx->h_i_val;

+#ifdef CPU_COLDFIRE
+            /* add saved partial pixel from start of area */
+            asm volatile (
+                "mac.l %[acc], %[h_o], %%acc0\n\t"
+                "mac.l %[tmp], %[mul], %%acc0\n\t"
+                :
+                : [acc] "r" (acc),
+                  [tmp] "r" (tmp),
+                  [h_o] "r" (ctx->h_o_val),
+                  [mul] "r" (mul)
+            );
+            /* get new pixel , then add its partial coverage to this area */
+            tmp = *(part->buf);
+            mul = ctx->h_o_val - oxe;
+            asm volatile (
+                "mac.l %[tmp], %[mul], %%acc0\n\t"
+                "movclr.l %%acc0, %[acc]\n\t"
+                : [acc] "=r" (acc)
+                : [tmp] "r" (tmp),
+                  [mul] "r" (mul)
+            );
+#else
             /* add saved partial pixel from start of area */
-            acc = MULUQ(acc, ctx->bm->width) + MULUQ(tmp, mul);
+            acc = (acc * ctx->h_o_val) + (tmp * mul);

             /* get new pixel , then add its partial coverage to this area */
             tmp = *(part->buf);
-            mul = ctx->bm->width - oxe;
-            acc += MULUQ(tmp, mul);
+            mul = ctx->h_o_val - oxe;
+            acc += tmp * mul;
+#endif
             /* round, divide, and either store or accumulate to output row */
+            acc = (acc + (1 << 21)) >> 22;
             if (accum)
             {
                 acc += out_line[ox];
@@ -232,7 +301,7 @@ static bool scale_h_area(void *out_line_ptr,
             out_line[ox] = acc;
             /* reset accumulator */
             acc = 0;
-            mul = ctx->bm->width - mul;
+            mul = ctx->h_o_val - mul;
             ox += 1;
         /* inside an area */
         } else {
@@ -252,9 +321,6 @@ static inline bool scale_v_area(struct rowset *rset, struct scaler_context *ctx)
     uint32_t mul, x, oy, iy, oye;

     /* Set up rounding and scale factors */
-    ctx->divisor *= ctx->src->height;
-    ctx->round = ctx->divisor >> 1;
-    ctx->divisor = 1 + (-((ctx->divisor + 1) >> 1)) / ctx->divisor;
     mul = 0;
     oy = rset->rowstart;
     oye = 0;
@@ -271,28 +337,28 @@ static inline bool scale_v_area(struct rowset *rset, struct scaler_context *ctx)
     /* zero the accumulator and temp rows */
     for (iy = 0; iy < (unsigned int)ctx->src->height; iy++)
     {
-        oye += ctx->bm->height;
+        oye += ctx->v_o_val;
         /* end of current area has been reached */
-        if (oye >= (unsigned int)ctx->src->height)
+        if (oye >= (unsigned int)ctx->v_i_val)
         {
             /* "reset" error, which now represents partial coverage of the next
                row by the next area
             */
-            oye -= ctx->src->height;
+            oye -= ctx->v_i_val;
             /* add stored partial row to accumulator */
 #ifdef HAVE_LCD_COLOR
             for (x = 0; x < 3 * (unsigned int)ctx->bm->width; x++)
 #else
             for (x = 0; x < (unsigned int)ctx->bm->width; x++)
 #endif
-                rowacc[x] = rowacc[x] * ctx->bm->height + mul * rowtmp[x];
+                rowacc[x] = rowacc[x] * ctx->v_o_val + mul * rowtmp[x];
             /* store new scaled row in temp row */
             if(!ctx->h_scaler(rowtmp, ctx, false))
                 return false;
             /* add partial coverage by new row to this area, then round and
                scale to final value
             */
-            mul = ctx->bm->height - oye;
+            mul = ctx->v_o_val - oye;
 #ifdef HAVE_LCD_COLOR
             for (x = 0; x < 3 * (unsigned int)ctx->bm->width; x++)
 #else
@@ -319,15 +385,6 @@ static inline bool scale_v_area(struct rowset *rset, struct scaler_context *ctx)
 }

 #ifdef HAVE_UPSCALER
-/* Set up rounding and scale factors for the horizontal scaler. The divisor
-   is bm->width - 1, so that the first and last pixels in the row align
-   exactly between input and output
-*/
-static inline void scale_h_linear_setup(struct scaler_context *ctx)
-{
-    ctx->divisor = ctx->bm->width - 1;
-}
-
 /* horizontal linear scaler */
 static bool scale_h_linear(void *out_line_ptr, struct scaler_context *ctx,
                            bool accum)
@@ -348,24 +405,24 @@ static bool scale_h_linear(void *out_line_ptr, struct scaler_context *ctx,
     FILL_BUF_INIT(part,ctx->store_part,ctx->args);
     ix = 0;
     /* The error is set so that values are initialized on the first pass. */
-    ixe = ctx->bm->width - 1;
+    ixe = ctx->h_o_val;
     /* give other tasks a chance to run */
     yield();
     for (ox = 0; ox < (uint32_t)ctx->bm->width; ox++)
     {
 #ifdef HAVE_LCD_COLOR
-        if (ixe >= ((uint32_t)ctx->bm->width - 1))
+        if (ixe >= ctx->h_o_val)
         {
             /* Store the new "current" pixel value in rgbval, and the color
                step value in rgbinc.
             */
-            ixe -= (ctx->bm->width - 1);
+            ixe -= ctx->h_o_val;
             rgbinc.r = -(part->buf->red);
             rgbinc.g = -(part->buf->green);
             rgbinc.b = -(part->buf->blue);
-            rgbval.r = (part->buf->red) * (ctx->bm->width - 1);
-            rgbval.g = (part->buf->green) * (ctx->bm->width - 1);
-            rgbval.b = (part->buf->blue) * (ctx->bm->width - 1);
+            rgbval.r = (part->buf->red) * ctx->h_o_val;
+            rgbval.g = (part->buf->green) * ctx->h_o_val;
+            rgbval.b = (part->buf->blue) * ctx->h_o_val;
             ix += 1;
             /* If this wasn't the last pixel, add the next one to rgbinc. */
             if (ix < (uint32_t)ctx->src->width) {
@@ -383,10 +440,10 @@ static bool scale_h_linear(void *out_line_ptr, struct scaler_context *ctx,
                 rgbval.g += rgbinc.g * ixe;
                 rgbval.b += rgbinc.b * ixe;
             }
-            /* Now multiple the color increment to its proper value */
-            rgbinc.r *= ctx->src->width - 1;
-            rgbinc.g *= ctx->src->width - 1;
-            rgbinc.b *= ctx->src->width - 1;
+            /* Now multiply the color increment to its proper value */
+            rgbinc.r *= ctx->h_i_val;
+            rgbinc.g *= ctx->h_i_val;
+            rgbinc.b *= ctx->h_i_val;
         } else {
             rgbval.r += rgbinc.r;
             rgbval.g += rgbinc.g;
@@ -395,24 +452,24 @@ static bool scale_h_linear(void *out_line_ptr, struct scaler_context *ctx,
         /* round and scale values, and accumulate or store to output */
         if (accum)
         {
-            out_line[ox].r += rgbval.r;
-            out_line[ox].g += rgbval.g;
-            out_line[ox].b += rgbval.b;
+            out_line[ox].r += (rgbval.r + (1 << 21)) >> 22;
+            out_line[ox].g += (rgbval.g + (1 << 21)) >> 22;
+            out_line[ox].b += (rgbval.b + (1 << 21)) >> 22;
         } else {
-            out_line[ox].r = rgbval.r;
-            out_line[ox].g = rgbval.g;
-            out_line[ox].b = rgbval.b;
+            out_line[ox].r = (rgbval.r + (1 << 21)) >> 22;
+            out_line[ox].g = (rgbval.g + (1 << 21)) >> 22;
+            out_line[ox].b = (rgbval.b + (1 << 21)) >> 22;
         }
 #else
-        if (ixe >= ((uint32_t)ctx->bm->width - 1))
+        if (ixe >= ctx->h_o_val)
         {
             /* Store the new "current" pixel value in rgbval, and the color
                step value in rgbinc.
             */
-            ixe -= (ctx->bm->width - 1);
+            ixe -= ctx->h_o_val;
             val = *(part->buf);
             inc = -val;
-            val = MULUQ(val, ctx->bm->width - 1);
+            val = val * ctx->h_o_val;
             ix += 1;
             /* If this wasn't the last pixel, add the next one to rgbinc. */
             if (ix < (uint32_t)ctx->src->width) {
@@ -424,21 +481,21 @@ static bool scale_h_linear(void *out_line_ptr, struct scaler_context *ctx,
                 /* Add a partial step to rgbval, in this pixel isn't precisely
                    aligned with the new source pixel
                 */
-                val += MULQ(inc, ixe);
+                val += inc * ixe;
             }
             /* Now multiply the color increment to its proper value */
-            inc = MULQ(inc, ctx->src->width - 1);
+            inc = inc * ctx->h_i_val;
         } else
             val += inc;
         /* round and scale values, and accumulate or store to output */
         if (accum)
         {
-            out_line[ox] += val;
+            out_line[ox] += (val + (1 << 21)) >> 22;
         } else {
-            out_line[ox] = val;
+            out_line[ox] = (val + (1 << 21)) >> 22;
         }
 #endif
-        ixe += ctx->src->width - 1;
+        ixe += ctx->h_i_val;
     }
     return true;
 }
@@ -449,12 +506,8 @@ static inline bool scale_v_linear(struct rowset *rset,
 {
     uint32_t mul, x, iy, iye;
     int32_t oy;
-    /* Set up scale and rounding factors, the divisor is bm->height - 1 */
-    ctx->divisor *= (ctx->bm->height - 1);
-    ctx->round = ctx->divisor >> 1;
-    ctx->divisor = 1 + (-((ctx->divisor + 1) >> 1)) / ctx->divisor;
-    /* Set up our two temp buffers. The names are generic because they'll be
-       swapped each time a new input row is read
+    /* Set up our buffers, to store the increment and current value for each
+       column, and one temp buffer used to read in new rows.
     */
 #ifdef HAVE_LCD_COLOR
     uint32_t *rowinc = (uint32_t *)(ctx->buf),
@@ -469,15 +522,15 @@ static inline bool scale_v_linear(struct rowset *rset,
     SDEBUGF("scale_v_linear\n");
     mul = 0;
     iy = 0;
-    iye = ctx->bm->height - 1;
+    iye = ctx->v_o_val;
     /* get first scaled row in rowtmp */
     if(!ctx->h_scaler((void*)rowtmp, ctx, false))
         return false;
     for (oy = rset->rowstart; oy != rset->rowstop; oy += rset->rowstep)
     {
-        if (iye >= (uint32_t)ctx->bm->height - 1)
+        if (iye >= ctx->v_o_val)
         {
-            iye -= ctx->bm->height - 1;
+            iye -= ctx->v_o_val;
             iy += 1;
 #ifdef HAVE_LCD_COLOR
             for (x = 0; x < 3 * (uint32_t)ctx->bm->width; x++)
@@ -486,7 +539,7 @@ static inline bool scale_v_linear(struct rowset *rset,
 #endif
             {
                 rowinc[x] = -rowtmp[x];
-                rowval[x] = rowtmp[x] * (ctx->bm->height - 1);
+                rowval[x] = rowtmp[x] * ctx->v_o_val;
             }
             if (iy < (uint32_t)ctx->src->height)
             {
@@ -500,7 +553,7 @@ static inline bool scale_v_linear(struct rowset *rset,
                 {
                     rowinc[x] += rowtmp[x];
                     rowval[x] += rowinc[x] * iye;
-                    rowinc[x] *= ctx->src->height - 1;
+                    rowinc[x] *= ctx->v_i_val;
                 }
             }
         } else
@@ -511,7 +564,7 @@ static inline bool scale_v_linear(struct rowset *rset,
 #endif
                     rowval[x] += rowinc[x];
         ctx->output_row(oy, (void*)rowval, ctx);
-        iye += ctx->src->height - 1;
+        iye += ctx->v_i_val;
     }
     return true;
 }
@@ -533,9 +586,9 @@ static void output_row_32_native_fromyuv(uint32_t row, void * row_in,
     for (col = 0; col < ctx->bm->width; col++) {
         if (ctx->dither)
             delta = DITHERXDY(col,dy);
-        y = SC_MUL(qp->b + ctx->round, ctx->divisor);
-        u = SC_MUL(qp->g + ctx->round, ctx->divisor);
-        v = SC_MUL(qp->r + ctx->round, ctx->divisor);
+        y = SC_OUT(qp->b);
+        u = SC_OUT(qp->g);
+        v = SC_OUT(qp->r);
         qp++;
         yuv_to_rgb(y, u, v, &r, &g, &b);
         r = (31 * r + (r >> 3) + delta) >> 8;
@@ -571,7 +624,7 @@ static void output_row_32_native(uint32_t row, void * row_in,
                 for (col = 0; col < ctx->bm->width; col++) {
                     if (ctx->dither)
                         delta = DITHERXDY(col,dy);
-                    bright = SC_MUL((*qp++) + ctx->round,ctx->divisor);
+                    bright = SC_OUT(*qp++);
                     bright = (3 * bright + (bright >> 6) + delta) >> 8;
                     data |= (~bright & 3) << shift;
                     shift -= 2;
@@ -594,7 +647,7 @@ static void output_row_32_native(uint32_t row, void * row_in,
                 for (col = 0; col < ctx->bm->width; col++) {
                     if (ctx->dither)
                         delta = DITHERXDY(col,dy);
-                    bright = SC_MUL((*qp++) + ctx->round, ctx->divisor);
+                    bright = SC_OUT(*qp++);
                     bright = (3 * bright + (bright >> 6) + delta) >> 8;
                     *dest++ |= (~bright & 3) << shift;
                 }
@@ -609,7 +662,7 @@ static void output_row_32_native(uint32_t row, void * row_in,
                 for (col = 0; col < ctx->bm->width; col++) {
                     if (ctx->dither)
                         delta = DITHERXDY(col,dy);
-                    bright = SC_MUL((*qp++) + ctx->round, ctx->divisor);
+                    bright = SC_OUT(*qp++);
                     bright = (3 * bright + (bright >> 6) + delta) >> 8;
                     *dest++ |= vi_pattern[bright] << shift;
                 }
@@ -625,9 +678,9 @@ static void output_row_32_native(uint32_t row, void * row_in,
                     if (ctx->dither)
                         delta = DITHERXDY(col,dy);
                     q0 = *qp++;
-                    r = SC_MUL(q0.r + ctx->round, ctx->divisor);
-                    g = SC_MUL(q0.g + ctx->round, ctx->divisor);
-                    b = SC_MUL(q0.b + ctx->round, ctx->divisor);
+                    r = SC_OUT(q0.r);
+                    g = SC_OUT(q0.g);
+                    b = SC_OUT(q0.b);
                     r = (31 * r + (r >> 3) + delta) >> 8;
                     g = (63 * g + (g >> 2) + delta) >> 8;
                     b = (31 * b + (b >> 3) + delta) >> 8;
@@ -664,13 +717,10 @@ int resize_on_load(struct bitmap *bm, bool dither, struct dim *src,
                    struct img_part* (*store_part)(void *args),
                    void *args)
 {
-
-#ifdef HAVE_UPSCALER
     const int sw = src->width;
     const int sh = src->height;
     const int dw = bm->width;
     const int dh = bm->height;
-#endif
     int ret;
 #ifdef HAVE_LCD_COLOR
     unsigned int needed = sizeof(struct uint32_rgb) * 3 * bm->width;
@@ -740,23 +790,35 @@ int resize_on_load(struct bitmap *bm, bool dither, struct dim *src,
     {
 #endif
         ctx.h_scaler = scale_h_area;
-        scale_h_area_setup(&ctx);
+        uint32_t h_div = (1U << 24) / sw;
+        ctx.h_i_val = sw * h_div;
+        ctx.h_o_val = dw * h_div;
 #ifdef HAVE_UPSCALER
     } else {
         ctx.h_scaler = scale_h_linear;
-        scale_h_linear_setup(&ctx);
+        uint32_t h_div = (1U << 24) / (dw - 1);
+        ctx.h_i_val = (sw - 1) * h_div;
+        ctx.h_o_val = (dw - 1) * h_div;
     }
 #endif
-    SC_MUL_INIT;
 #ifdef HAVE_UPSCALER
     if (sh > dh)
 #endif
+    {
+        uint32_t v_div = (1U << 22) / sh;
+        ctx.v_i_val = sh * v_div;
+        ctx.v_o_val = dh * v_div;
         ret = scale_v_area(rset, &ctx);
+    }
 #ifdef HAVE_UPSCALER
     else
+    {
+        uint32_t v_div = (1U << 22) / dh;
+        ctx.v_i_val = (sh - 1) * v_div;
+        ctx.v_o_val = (dh - 1) * v_div;
         ret = scale_v_linear(rset, &ctx);
+    }
 #endif
-    SC_MUL_END;
 #ifdef HAVE_ADJUSTABLE_CPU_FREQ
     cpu_boost(false);
 #endif
diff --git a/apps/recorder/resize.h b/apps/recorder/resize.h
index 2964fcd..b861e43 100644
--- a/apps/recorder/resize.h
+++ b/apps/recorder/resize.h
@@ -43,68 +43,7 @@
 #define MAX_SC_STACK_ALLOC 0
 #define HAVE_UPSCALER 1

-#if defined(CPU_COLDFIRE)
-#define SC_MUL_INIT \
-    unsigned long macsr_st = coldfire_get_macsr(); \
-    coldfire_set_macsr(EMAC_UNSIGNED);
-#define SC_MUL_END coldfire_set_macsr(macsr_st);
-#define SC_MUL(x, y) \
-({ \
-    unsigned long t; \
-    asm ("mac.l    %[a], %[b], %%acc0\n\t" \
-         "move.l %%accext01, %[t]\n\t" \
-         "move.l #0, %%acc0\n\t" \
-         : [t] "=r" (t) : [a] "r" (x), [b] "r" (y)); \
-    t; \
-})
-#elif (CONFIG_CPU == SH7034)
-/* multiply two unsigned 32 bit values and return the top 32 bit
- * of the 64 bit result */
-static inline unsigned sc_mul32(unsigned a, unsigned b)
-{
-    unsigned r, t1, t2, t3;
-
-    asm (
-        "swap.w  %[a], %[t1]     \n" /* t1 = ba */
-        "mulu    %[t1], %[b]     \n" /* a * d */
-        "swap.w  %[b], %[t3]     \n" /* t3 = dc */
-        "sts     macl, %[t2]     \n" /* t2 = a * d */
-        "mulu    %[t1], %[t3]    \n" /* a * c */
-        "sts     macl, %[r]      \n" /* hi = a * c */
-        "mulu    %[a], %[t3]     \n" /* b * c */
-        "clrt                    \n"
-        "sts     macl, %[t3]     \n" /* t3 = b * c */
-        "addc    %[t2], %[t3]    \n" /* t3 += t2, carry -> t2 */
-        "movt    %[t2]           \n"
-        "mulu    %[a], %[b]      \n" /* b * d */
-        "mov     %[t3], %[t1]    \n" /* t1t3 = t2t3 << 16 */
-        "xtrct   %[t2], %[t1]    \n"
-        "shll16  %[t3]           \n"
-        "sts     macl, %[t2]     \n" /* lo = b * d */
-        "clrt                    \n" /* hi.lo += t1t3 */
-        "addc    %[t3], %[t2]    \n"
-        "addc    %[t1], %[r]     \n"
-        : /* outputs */
-        [r] "=&r"(r),
-        [t1]"=&r"(t1),
-        [t2]"=&r"(t2),
-        [t3]"=&r"(t3)
-        : /* inputs */
-        [a] "r"  (a),
-        [b] "r"  (b)
-    );
-    return r;
-}
-#define SC_MUL(x, y) sc_mul32(x, y)
-#define SC_MUL_INIT
-#define SC_MUL_END
-#endif
-
-#ifndef SC_MUL
-#define SC_MUL(x, y) ((x) * (uint64_t)(y) >> 32)
-#define SC_MUL_INIT
-#define SC_MUL_END
-#endif
+#define SC_OUT(n) (((n) + (1 << 23)) >> 24);

 struct img_part {
     int len;
@@ -130,8 +69,12 @@ struct uint32_rgb {
    horizontal scaler, and row output
 */
 struct scaler_context {
-    uint32_t divisor;
-    uint32_t round;
+    uint32_t h_div;
+    uint32_t h_i_val;
+    uint32_t h_o_val;
+    uint32_t v_div;
+    uint32_t v_i_val;
+    uint32_t v_o_val;
     struct bitmap *bm;
     struct dim *src;
     unsigned char *buf;