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
|
/*
* Copyright 2000-2011 Intel Corporation All Rights Reserved
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the
* "Software"), to deal in the Software without restriction, including
* without limitation the rights to use, copy, modify, merge, publish,
* distribute, sub license, and/or sell copies of the Software, and to
* permit persons to whom the Software is furnished to do so, subject to
* the following conditions:
*
* The above copyright notice and this permission notice (including the
* next paragraph) shall be included in all copies or substantial portions
* of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
* OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
* IN NO EVENT SHALL PRECISION INSIGHT AND/OR ITS SUPPLIERS BE LIABLE FOR
* ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*
* This file was originally licensed under the following license
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
// 45 // Total instruction count
// 1 // Total kernel count
// Module name: common.inc
//
// Common header file for all Video-Processing kernels
//
.default_execution_size (16)
.default_register_type :ub
.reg_count_total 128
.reg_count_payload 7
//========== Common constants ==========
//========== Macros ==========
//Fast Jump, For more details see "Set_Layer_N.asm"
//========== Defines ====================
//========== Static Parameters (Common To All) ==========
//r1
//r2
// e.g. byte0 byte1 byte2
// YUYV 0 1 3
// YVYU 0 3 1
//Color Pipe (IECP) parameters
//ByteCopy
//r4
// e.g. byte0 byte1 byte2
// YUYV 0 1 3
// YVYU 0 3 1
//========== Inline parameters (Common To All) ===========
//============== Binding Index Table===========
//Common between DNDI and DNUV
//================= Common Message Descriptor =====
// Message descriptor for thread spawning
// Message Descriptors
// = 000 0001 (min message len 1 ) 0,0000 (resp len 0 -add later)
// 0000,0000,0000
// 0001(Spawn a root thread),0001 (Root thread spawn thread)
// = 0x02000011
// Thread Spawner Message Descriptor
// Message descriptor for atomic operation add
// Message Descriptors
// = 000 0110 (min message len 6 ) 0,0000 (resp len 0 -add later)
// 1(header present)001,10(typed atomic operation)0(return enabled)0(slot group, low 8 bits),0111 (AOP_Add)
// 0000,0000 (Binding table index, added later)
// = 0x02000011
// Atomic Operation Add Message Descriptor
// Message descriptor for dataport media write
// Message Descriptors
// = 000 0001 (min message len 1 - add later) 00000 (resp len 0)
// 1 (header present 1) 0 1010 (media block write) 000000
// 00000000 (binding table index - set later)
// = 0x020A8000
// Message Length defines
// Response Length defines
// Block Width and Height Size defines
// Extended Message Descriptors
// Common message descriptors:
//===================== Math Function Control ===================================
//============ Message Registers ===============
// buf4 starts from r28
//#define mMSGHDR_EOT r43 // Dummy Message Register for EOT
.declare mubMSGPAYLOAD Base=r30 ElementSize=1 SrcRegion=<16;16,1> Type=ub
.declare muwMSGPAYLOAD Base=r30 ElementSize=2 SrcRegion=<16;16,1> Type=uw
.declare mudMSGPAYLOAD Base=r30 ElementSize=4 SrcRegion=<8;8,1> Type=ud
.declare mfMSGPAYLOAD Base=r30 ElementSize=4 SrcRegion=<8;8,1> Type=f
//=================== End of thread instruction ===========================
//=====================Pointers Used=====================================
//=======================================================================
//r11-r17
// Define temp space for any usages
// Common Buffers
// temp space for rotation
.declare fROBUF Base=r11.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
.declare udROBUF Base=r11.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
.declare uwROBUF Base=r11.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
.declare ubROBUF Base=r11.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
.declare ub4ROBUF Base=r11.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
// End of common.inc
// FileName: PL3_AVS_Buf_1.asm
// Author: Tatiya, Rupesh
// Description: Loads 8x8 AVS/IEF PL3 data into Buffer 1
// FileName : PL3_AVS_Buf.asm
// Author : Tatiya, Rupesh
// Description : Loads 8x8 AVS/IEF PL3 data into Buffer N
// Module name: Scaling.inc
// Description: Includes all definitions explicit to Fast Composite.
// End of common.inc
//========== GRF partition ==========
// r0 header : r0 (1 GRF)
// Static parameters : r1 - r6 (6 GRFS)
// Inline parameters : r7 - r8 (2 GRFs)
// MSGSRC : r27 (1 GRF)
//===================================
//Interface:
//========== Static Parameters (Explicit To Fast Composite) ==========
//r1
//CSC Set 0
.declare udCSC_CURBE Base=r1.0 ElementSize=4 Type=ud
//Constant alpha
//r2
// Gen7 AVS WA
// WiDi Definitions
//Colorfill
// 0: 0-degree, 1: 90, 2: 180, 3: 270-degree, clockwise.
.declare ubCOLOR_PIXEL_VAL Base=r2.20 ElementSize=1 SrcRegion=<0;1,0> DstRegion=<1> Type=ub
//r3
//Normalised Ratio of Horizontal step size with main video for all layers
//Normalised Ratio of Horizontal step size with main video for all layers becomes
//Normalised Horizontal step size for all layers in VP_Setup.asm
//r4
//Normalised Vertical step size for all layers
//r5
//Normalised Vertical Frame Origin for all layers
//r6
//Normalised Horizontal Frame Origin for all layers
//========== Inline Parameters (Explicit To Fast Composite) ==========
//Main video Step X
//====================== Binding table (Explicit To Fast Composite)=========================================
//Used by Interlaced Scaling Kernels
//========== Sampler State Table Index (Explicit To Fast Composite)==========
//Sampler Index for AVS/IEF messages
//Sampler Index for SIMD16 sampler messages
//=============================================================================
.declare fBUFFER_0 Base=r64.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
.declare fBUFFER_1 Base=r80.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
.declare fBUFFER_2 Base=r96.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
.declare fBUFFER_3 Base=r112.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
.declare fBUFFER_4 Base=r28.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
.declare fBUFFER_5 Base=r46.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
.declare udBUFFER_0 Base=r64.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
.declare udBUFFER_1 Base=r80.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
.declare udBUFFER_2 Base=r96.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
.declare udBUFFER_3 Base=r112.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
.declare udBUFFER_4 Base=r28.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
.declare udBUFFER_5 Base=r46.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
.declare uwBUFFER_0 Base=r64.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
.declare uwBUFFER_1 Base=r80.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
.declare uwBUFFER_2 Base=r96.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
.declare uwBUFFER_3 Base=r112.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
.declare uwBUFFER_4 Base=r28.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
.declare uwBUFFER_5 Base=r46.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
.declare ubBUFFER_0 Base=r64.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
.declare ubBUFFER_1 Base=r80.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
.declare ubBUFFER_2 Base=r96.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
.declare ubBUFFER_3 Base=r112.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
.declare ubBUFFER_4 Base=r28.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
.declare ubBUFFER_5 Base=r46.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
.declare ub4BUFFER_0 Base=r64.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
.declare ub4BUFFER_1 Base=r80.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
.declare ub4BUFFER_2 Base=r96.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
.declare ub4BUFFER_3 Base=r112.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
.declare ub4BUFFER_4 Base=r28.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
.declare ub4BUFFER_5 Base=r46.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
//Pointer to mask reg
//r18
//Always keep Cannel Pointers and Offsets in same GRF, so that we can use
// NODDCLR, NODDCHK flags. -rT
.declare udCSC_COEFF_0 Base=r18.0 ElementSize=4 Type=ud // 1 GRF
//r19
.declare udCSC_COEFF_1 Base=r19.0 ElementSize=4 Type=ud // 1 GRF
//r20
.declare uwALPHA_MASK_REG_TEMP Base=r20.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw // 1 GRF
//r21
.declare uwALPHA_MASK_REG Base=r21.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw // 1 GRF
//r22
//Always keep Cannel Pointers and Offsets in same GRF, so that we can use
// NODDCLR, NODDCHK flags. -rT
//Keep fORIGIN_X_NLAS, fY_OFFSET_2ND_BLOCK, fSTEP_X_NLAS, pMSGDSC_COPY, ubCONST_ALPHA_COPY as
//sub registers of same GRF to enable using NODDCLR NODDCHK. -rT
//r23
//Lumakey
//r24
//r25
//r26
//defines to generate LABELS during compile time.
// Message Header
// m0.7 31:0 Debug
// m0.6 31:0 Debug
// m0.5 31:0 Ignored
// m0.4 31:0 Ignored
// m0.3 31:0 Ignored
// m0.2 31:16 Ignored
// 15 Alpha Write Channel Mask enable=0, disable=1
// 14 Blue Write Channel Mask (U)
// 13 Green Write Channel Mask (Y)
// 12 Red Write Channel Mask (V)
// 11:0 Ignored
// m0.1 Ignored
// m0.0 Ignored
// AVS payload
// m1.7 Group ID Number
// m1.6 U 2nd Derivative ---> NLAS dx
// m1.5 Delta V ---> Step Y
// m1.4 Delta U ---> Step X
// m1.3 Pixel 0 V Address ---> ORIY (Y0)
// m1.2 Pixel 0 U Address ---> ORIX (X0)
// m1.1 Vertical Block Number
// m1.0 Reserved
// Sampler Message Descriptor
// 31:29 Reserved 000
// 28:25 Message length 0010
// 24:20 Response length xxxxx ---> 4GRFs for each enabled channel (AVS), 2GRFs for each enabled channel (sample unorm)
// 19 Header Present 1
// 18:17 SIMD Mode 11 ---> SIMD32/64
// 16:12 Message Type xxxxx ---> 01011 sample_8x8, 01100 (sample_unorm), 01010 (sample_unorm+killpix)
// 11:8 Sampler Index xxxx
// 7:0 Binding Table Index xxxxxxxx
// Msg Header M0.2
// 15:15 Alpha Write Channel Mask, 0: written back, 1: not written back
// 14:14 Blue Write Channel Mask
// 13:13 Green Write Channel Mask
// 12:12 Red Write Channel Mask
//By design, Buffer 0,1,2,3 always have Layer 0 and Buffer 4,5 always have L1-L7
//used to generate LABELS at compile time.
// 18:17 SIMD Mode 10 ---> SIMD16
// 16:12 Message Type xxxxx ---> 00000 (SIMD16)
//r10-17 - 8 GRFs to load SIMD16 data (upto 4 channels)
//r18-19 - 2 GRFs to store sampler ramp.
.declare mfSCALING_0X_34X_PAYLOAD Base=r14.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
.declare muwSCALING_0X_34X_PAYLOAD Base=r14.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
.declare mudCALING_0X_34X_PAYLOAD Base=r14.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
.declare mubCALING_0X_34X_PAYLOAD Base=r14.0 ElementSize=1 SrcRegion=<32;32,1> DstRegion=<1> Type=ub
.declare fSCALING_0X_34X_TEMP Base=r11.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
.declare udSCALING_0X_34X_TEMP Base=r11.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
.declare ub4SCALING_0X_34X_TEMP Base=r11.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<1> Type=ub
.declare uwSCALING_0X_34X_TEMP Base=r11.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
// Sampler ramp is used for Scaling 0X_0.34X
.declare fSAMPLER_RAMP Base=r11.0 ElementSize=4 SrcRegion=<8;8,1> Type=f // 1 GRFs, 8 elements
//#define rMSGDSC_UV r23.0
//End of _SCALING_
//Check if layer is to be skipped
// f0.1 pre-computed in Set_Layer_0
(-f0.1) jmpi (1) SKIP_AVS_LOAD_L0_1_
//AVS_PAYLOAD already has all the data loaded at this point
add (1) a0.0:ud r23.5<0;1,0>:ud 0x44EB400:ud //msg desc
mov (1) r16.2:ud 0x0000D000:ud // Enable Red channel
// set the vertical block number
add (1) r25.1<1>:ud r9.12:uw 1:ud
mov (8) r17.0:ud r25.0<8;8,1>:ud // Copy msg payload mirrors to MRFs
// Gen7 AVS WA Only for YUV packed surfaces, NV12 and Y-channel only for Planar surfaces
// if (((int)(u_left*width + 5.0/256) > (int)(u_left*width))
// {
// modified_u_coord = u_coord – 5.0/(256*width); //floating point
// }
// else if(((int)(u_left*width + 255.0/256) == (int)(u_left*width))
// {
// modified_u_coord = u_coord + 1.0/(256*width); //floating point
// }
// else{
// modified_u_coord = u_coord;
// }
// Where u_left = u – 2*du + 3*ddu for IEF On
// And u_left = u for IEF Off case
//
// check whether Gen7 AVS WA is enabled,
mov (1) r14.8:uw f0.0:uw // save f0.0
mov (1) r14.5:f r17.2<0;1,0>:f // save pixel 0 U for chroma
and.nz.f0.0 (1) null<1>:uw r2.3:uw 0x2:uw
(-f0.0)jmpi (1) GEN7_PL3_AVS_WA_DONE_L0_1_
// Gen7 AVS WA, check if IEF is ON for choosing Gen7 AVS WA formula
and.nz.f0.0 (8) null<1>:uw r2.3<0;1,0>:uw 0x4:uw
(f0.0)mov (8) acc0.0:f r17.2<0;1,0>:f
(f0.0)mac (8) acc0.0:f r17.4<0;1,0>:f -2.0:f
(f0.0)mac (8) acc0.0:f r17.6<0;1,0>:f 3.0:f
(f0.0)mov (1) r14.2:f acc0:f // IEF ON, rTEMP3.2 = u_left
(-f0.0)mov (1) r14.2:f r17.2<0;1,0>:f // IEF OFF, rTEMP3.2 = u_left
and (1) r14.1:ud r2.3:uw 0xFFF8:uw
asr (1) r14.1:ud r14.1:ud 3:d
mov (1) r14.1:f r14.1:ud
// Gen7 AVS WA, if (int)(u_left*width + 5.0/256) > (int)(u_left*width)
mul (1) r14.0:f r14.2:f r14.1:f // rTEMP3.0 = u_left*width
add (1) r14.2:f r14.0:f 0.01953125:f // rTEMP3.2 = u_left*width + 5.0/256
add (1) r14.3:f r14.0:f 0.99609375:f // rTEMP3.3 = u_left*width + 255.0/256
//Check if the values are < 0 and account for (int) cast of negative numbers
//(int)(u_left*width)
cmp.l.f0.0 (1) null<1>:f r14.0:f 0.00000000:f
mov (1) r14.0:d r14.0:f
(f0.0)add (1) r14.0:d r14.0<0;1,0>:d -1:d
//(int)(u_left*width + 5.0/256)
cmp.l.f0.0 (1) null<1>:f r14.2:f 0.00000000:f
mov (1) r14.2:d r14.2:f
(f0.0)add (1) r14.2:d r14.2<0;1,0>:d -1:d
//(int)(u_left*width + 255.0/256)
cmp.l.f0.0 (1) null<1>:f r14.3:f 0.00000000:f
mov (1) r14.3:d r14.3:f
(f0.0)add (1) r14.3:d r14.3<0;1,0>:d -1:d
mov (1) f0.0:uw 0:uw // clear flag
//if (((int)(u_left*width + 5.0/256) > (int)(u_left*width))
cmp.g.f1.0 (1) null<1>:d r14.2:d r14.0:d
// modified_u_coord = u_coord – 5.0/(256*width); //floating point
(f1.0) add (1) r17.2:f r17.2<0;1,0>:f -r2.3:f
//else if(((int)(u_left*width + 255.0/256) == (int)(u_left*width))
(-f1.0) cmp.e.f0.0 (1) null<1>:d r14.3:d r14.0:d
// modified_u_coord = u_coord + 1.0/(256*width); //floating point
(f0.0) add (1) r17.2:f r17.2<0;1,0>:f r2.2:f
GEN7_PL3_AVS_WA_DONE_L0_1_:
mov (1) f0.0:uw r14.8:uw // restore f0.0
send (1) uwBUFFER_1(0)<1> r16 0x2 a0.0:ud
// Returns Y data in 4 GRFs in scrambled order
mov (1) r17.2:f r14.5:f // restore pixel 0 U for chroma, No AVS WA for chroma
add (1) a0.0:ud r23.5<0;1,0>:ud 0x44EB801:ud // msg desc; 1 is added to change BI to UV
mov (1) r16.2:ud 0x0000E000:ud // Enable Red channel
send (1) uwBUFFER_1(4)<1> r16 0x2 a0.0:ud
// Returns U data in 4 GRFs in scrambled order
add (1) a0.0:ud r23.5<0;1,0>:ud 0x44EBC02:ud // msg desc; 1 is added to change BI to UV
mov (1) r16.2:ud 0x0000E000:ud // Enable Red channel
send (1) uwBUFFER_1(8)<1> r16 0x2 a0.0:ud
// Returns V data in 4 GRFs in scrambled order
SKIP_AVS_LOAD_L0_1_:
nop
|
