src/shaders/post_processing/gen5_6/Makefile
src/shaders/post_processing/gen7/Makefile
src/shaders/post_processing/gen75/Makefile
+ src/shaders/post_processing/gen8/Makefile
src/shaders/render/Makefile
src/shaders/utils/Makefile
src/shaders/vme/Makefile
};
static const uint32_t pp_nv12_load_save_nv12_gen8[][4] = {
-#include "shaders/post_processing/gen7/pl2_to_pl2.g75b"
+#include "shaders/post_processing/gen8/pl2_to_pl2.g8b"
};
static const uint32_t pp_nv12_load_save_pl3_gen8[][4] = {
-SUBDIRS = gen5_6 gen7 gen75
+SUBDIRS = gen5_6 gen7 gen75 gen8
# Extra clean files so that maintainer-clean removes *everything*
MAINTAINERCLEANFILES = Makefile.in
--- /dev/null
+/*
+ * Copyright 2000-2011 Intel Corporation All Rights Reserved
+ *
+ * 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.
+ */
+// 2 // 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=====================================
+
+
+//=======================================================================
+
+
+//r9-r17
+// Define temp space for any usages
+
+
+// Common Buffers
+
+
+// temp space for rotation
+
+.declare fROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+
+.declare udROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+
+.declare uwROBUF Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+.declare ubROBUF Base=r9.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
+
+.declare ub4ROBUF Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
+
+
+// End of common.inc
+
+
+//End of Thread message
+
+mov (8) r127<1>:ud r0.0<8;8,1>:ud
+ send (1) null<1>:d r127 0x27 0x02000010
--- /dev/null
+INTEL_PP_G8B = \
+ pl2_to_pl2.g8b \
+ $(NULL)
+
+INTEL_PP_G8A = \
+ EOT.g8a \
+ PL2_AVS_Buf_0.g8a \
+ PL2_AVS_Buf_1.g8a \
+ PL2_AVS_Buf_2.g8a \
+ PL2_AVS_Buf_3.g8a \
+ Save_AVS_NV12.g8a \
+ Set_AVS_Buf_0123_PL2.g8a \
+ Set_Layer_0.g8a \
+ VP_Setup.g8a \
+ $(NULL)
+
+INTEL_PP_ASM = $(INTEL_PP_G8B:%.g8b=%.asm)
+INTEL_PP_GEN8_ASM = $(INTEL_PP_G8B:%.g8b=%.g8s)
+
+TARGETS =
+if HAVE_GEN4ASM
+TARGETS += $(INTEL_PP_G8B)
+endif
+
+all-local: $(TARGETS)
+
+SUFFIXES = .g8b .g8s .asm
+
+$(INTEL_PP_GEN8_ASM): $(INTEL_PP_ASM) $(INTEL_PP_G8A)
+.asm.g8s:
+ $(AM_V_GEN)cpp $< > _pp0.$@; \
+ ../../gpp.py _pp0.$@ $@; \
+ rm _pp0.$@
+.g8s.g8b:
+ $(AM_V_GEN)intel-gen4asm -a -o $@ -g 8 $<
+
+CLEANFILES = $(INTEL_PP_GEN7_ASM)
+
+EXTRA_DIST = \
+ $(INTEL_PP_G8B)
+
+# Extra clean files so that maintainer-clean removes *everything*
+MAINTAINERCLEANFILES = Makefile.in
--- /dev/null
+/*
+ * Copyright 2000-2011 Intel Corporation All Rights Reserved
+ *
+ * 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.
+ */
+// 44 // 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=====================================
+
+
+//=======================================================================
+
+
+//r9-r17
+// Define temp space for any usages
+
+
+// Common Buffers
+
+
+// temp space for rotation
+
+.declare fROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+
+.declare udROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+
+.declare uwROBUF Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+.declare ubROBUF Base=r9.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
+
+.declare ub4ROBUF Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
+
+
+// End of common.inc
+
+
+// FileName: PL2_AVS_Buf_0.asm
+// Author: Tatiya, Rupesh
+// Description: Loads 8x8 AVS/IEF PL2 data into Buffer 0
+
+
+
+// FileName : PL2_AVS_Buf.asm
+// Author : Tatiya, Rupesh
+// Description : Loads 8x8 AVS/IEF PL2 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=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+ .declare udSCALING_0X_34X_TEMP Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+ .declare ub4SCALING_0X_34X_TEMP Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<1> Type=ub
+ .declare uwSCALING_0X_34X_TEMP Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+ // Sampler ramp is used for Scaling 0X_0.34X
+ .declare fSAMPLER_RAMP Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> Type=f // 1 GRFs, 8 elements
+
+
+ //#define rMSGDSC_UV r23.0
+
+
+//End of _SCALING_
+
+
+ //NOTE: We need offsets for second halfof LAYER 0 - even if we do not load it.
+ //Update the channel offset in the buffers for the lower 8x4 data for BUFFER_0.
+ mov (1) r22.4<1>:ud 0x400040:ud
+
+
+ mov (1) r16.3<1>:ud r0.3<0;1,0>:ud
+
+
+ //AVS_PAYLOAD already has all the data loaded at this point
+ add (1) a0.0<1>:ud r23.5<0;1,0>:ud 0x44EB000:ud //msg desc
+
+ mov (1) r16.2<1>:ud 0x0000D000:ud // Enable Red channel
+
+
+
+ // set the vertical block number
+
+
+ mov (8) r17.0<1>:ud r25.0<8;8,1>:ud // Copy msg payload mirrors to MRFs
+
+ send (1) uwBUFFER_0(0)<1> r16 0x2 a0.0:ud
+ // Returns Y data in 4 GRFs in scrambled order
+
+ add (1) a0.0<1>:ud r23.5<0;1,0>:ud 0x48EB001:ud // msg desc; 1 is added to change BI to UV
+ mov (1) r16.2<1>:ud 0x0000A000:ud // Enable Red+Blue channel
+
+ send (1) uwBUFFER_0(4)<1> r16 0x2 a0.0:ud
+ // Returns UV data in 8 GRFs in scrambled order
+
+SKIP_AVS_LOAD_L0_0_:
+ nop
+
+
--- /dev/null
+/*
+ * Copyright 2000-2011 Intel Corporation All Rights Reserved
+ *
+ * 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.
+ */
+// 42 // 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=====================================
+
+
+//=======================================================================
+
+
+//r9-r17
+// Define temp space for any usages
+
+
+// Common Buffers
+
+
+// temp space for rotation
+
+.declare fROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+
+.declare udROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+
+.declare uwROBUF Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+.declare ubROBUF Base=r9.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
+
+.declare ub4ROBUF Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
+
+
+// End of common.inc
+
+
+// FileName: PL2_AVS_Buf_1.asm
+// Author: Tatiya, Rupesh
+// Description: Loads 8x8 AVS/IEF PL2 data into Buffer 1
+
+
+
+// FileName : PL2_AVS_Buf.asm
+// Author : Tatiya, Rupesh
+// Description : Loads 8x8 AVS/IEF PL2 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=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+ .declare udSCALING_0X_34X_TEMP Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+ .declare ub4SCALING_0X_34X_TEMP Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<1> Type=ub
+ .declare uwSCALING_0X_34X_TEMP Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+ // Sampler ramp is used for Scaling 0X_0.34X
+ .declare fSAMPLER_RAMP Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> Type=f // 1 GRFs, 8 elements
+
+
+ //#define rMSGDSC_UV r23.0
+
+
+//End of _SCALING_
+
+
+ mov (1) r16.3<1>:ud r0.3<0;1,0>:ud
+
+
+ //AVS_PAYLOAD already has all the data loaded at this point
+ add (1) a0.0<1>:ud r23.5<0;1,0>:ud 0x44EB000:ud //msg desc
+
+ mov (1) r16.2<1>:ud 0x0000D000:ud // Enable Red channel
+
+
+ // set the vertical block number
+
+ mov (1) r25.1<1>:ud 1:ud
+
+
+ mov (8) r17.0<1>:ud r25.0<8;8,1>:ud // Copy msg payload mirrors to MRFs
+
+ send (1) uwBUFFER_1(0)<1> r16 0x2 a0.0:ud
+ // Returns Y data in 4 GRFs in scrambled order
+
+ add (1) a0.0<1>:ud r23.5<0;1,0>:ud 0x48EB001:ud // msg desc; 1 is added to change BI to UV
+ mov (1) r16.2<1>:ud 0x0000A000:ud // Enable Red+Blue channel
+
+ send (1) uwBUFFER_1(4)<1> r16 0x2 a0.0:ud
+ // Returns UV data in 8 GRFs in scrambled order
+
+SKIP_AVS_LOAD_L0_1_:
+ nop
+
+
--- /dev/null
+/*
+ * Copyright 2000-2011 Intel Corporation All Rights Reserved
+ *
+ * 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.
+ */
+// 42 // 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=====================================
+
+
+//=======================================================================
+
+
+//r9-r17
+// Define temp space for any usages
+
+
+// Common Buffers
+
+
+// temp space for rotation
+
+.declare fROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+
+.declare udROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+
+.declare uwROBUF Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+.declare ubROBUF Base=r9.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
+
+.declare ub4ROBUF Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
+
+
+// End of common.inc
+
+
+// FileName: PL2_AVS_Buf_2.asm
+// Author: Tatiya, Rupesh
+// Description: Loads 8x8 AVS/IEF PL2 data into Buffer 2
+
+
+
+// FileName : PL2_AVS_Buf.asm
+// Author : Tatiya, Rupesh
+// Description : Loads 8x8 AVS/IEF PL2 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=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+ .declare udSCALING_0X_34X_TEMP Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+ .declare ub4SCALING_0X_34X_TEMP Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<1> Type=ub
+ .declare uwSCALING_0X_34X_TEMP Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+ // Sampler ramp is used for Scaling 0X_0.34X
+ .declare fSAMPLER_RAMP Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> Type=f // 1 GRFs, 8 elements
+
+
+ //#define rMSGDSC_UV r23.0
+
+
+//End of _SCALING_
+
+
+ mov (1) r16.3<1>:ud r0.3<0;1,0>:ud
+
+ //AVS_PAYLOAD already has all the data loaded at this point
+ add (1) a0.0<1>:ud r23.5<0;1,0>:ud 0x44EB000:ud //msg desc
+
+ mov (1) r16.2<1>:ud 0x0000D000:ud // Enable Red channel
+
+
+ // set the vertical block number
+
+
+ mov (1) r25.1<1>:ud 2:ud
+
+
+ mov (8) r17.0<1>:ud r25.0<8;8,1>:ud // Copy msg payload mirrors to MRFs
+
+ send (1) uwBUFFER_2(0)<1> r16 0x2 a0.0:ud
+ // Returns Y data in 4 GRFs in scrambled order
+
+ add (1) a0.0<1>:ud r23.5<0;1,0>:ud 0x48EB001:ud // msg desc; 1 is added to change BI to UV
+ mov (1) r16.2<1>:ud 0x0000A000:ud // Enable Red+Blue channel
+
+ send (1) uwBUFFER_2(4)<1> r16 0x2 a0.0:ud
+ // Returns UV data in 8 GRFs in scrambled order
+
+SKIP_AVS_LOAD_L0_2_:
+ nop
+
+
--- /dev/null
+/*
+ * Copyright 2000-2011 Intel Corporation All Rights Reserved
+ *
+ * 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.
+ */
+// 42 // 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=====================================
+
+
+//=======================================================================
+
+
+//r9-r17
+// Define temp space for any usages
+
+
+// Common Buffers
+
+
+// temp space for rotation
+
+.declare fROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+
+.declare udROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+
+.declare uwROBUF Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+.declare ubROBUF Base=r9.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
+
+.declare ub4ROBUF Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
+
+
+// End of common.inc
+
+
+// FileName: PL2_AVS_Buf_3.asm
+// Author: Tatiya, Rupesh
+// Description: Loads 8x8 AVS/IEF PL2 data into Buffer 0
+
+
+
+// FileName : PL2_AVS_Buf.asm
+// Author : Tatiya, Rupesh
+// Description : Loads 8x8 AVS/IEF PL2 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=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+ .declare udSCALING_0X_34X_TEMP Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+ .declare ub4SCALING_0X_34X_TEMP Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<1> Type=ub
+ .declare uwSCALING_0X_34X_TEMP Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+ // Sampler ramp is used for Scaling 0X_0.34X
+ .declare fSAMPLER_RAMP Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> Type=f // 1 GRFs, 8 elements
+
+
+ //#define rMSGDSC_UV r23.0
+
+
+//End of _SCALING_
+
+
+
+ mov (1) r16.3<1>:ud r0.3<0;1,0>:ud
+
+
+ //AVS_PAYLOAD already has all the data loaded at this point
+ add (1) a0.0<1>:ud r23.5<0;1,0>:ud 0x44EB000:ud //msg desc
+
+ mov (1) r16.2<1>:ud 0x0000D000:ud // Enable Red channel
+
+
+ // set the vertical block number
+
+
+ mov (1) r25.1<1>:ud 3:ud
+
+
+ mov (8) r17.0<1>:ud r25.0<8;8,1>:ud // Copy msg payload mirrors to MRFs
+
+ send (1) uwBUFFER_3(0)<1> r16 0x2 a0.0:ud
+ // Returns Y data in 4 GRFs in scrambled order
+
+ add (1) a0.0<1>:ud r23.5<0;1,0>:ud 0x48EB001:ud // msg desc; 1 is added to change BI to UV
+ mov (1) r16.2<1>:ud 0x0000A000:ud // Enable Red+Blue channel
+
+ send (1) uwBUFFER_3(4)<1> r16 0x2 a0.0:ud
+ // Returns UV data in 8 GRFs in scrambled order
+
+SKIP_AVS_LOAD_L0_3_:
+ nop
+
+
--- /dev/null
+/*
+ * Copyright 2000-2011 Intel Corporation All Rights Reserved
+ *
+ * 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.
+ */
+// 131 // 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=====================================
+
+
+//=======================================================================
+
+
+//r9-r17
+// Define temp space for any usages
+
+
+// Common Buffers
+
+
+// temp space for rotation
+
+.declare fROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+
+.declare udROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+
+.declare uwROBUF Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+.declare ubROBUF Base=r9.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
+
+.declare ub4ROBUF Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
+
+
+// End of common.inc
+
+
+// Module name: Save_AVS_NV12.asm
+//
+// Save NV12 420 frame data block of size 16x16
+//
+// To save 16x16 block (16x16 bytes of Y and 16x8 bytes of interleaved UV), we need 2 send instructions with of size 16x16 and 16x8 each.
+// ---------------
+// | 16x16 |
+// | YUYV |
+// ---------------
+// | 16x8 UV |
+// ---------------
+
+//-----------------------------------------------------------------
+//The layout of data is as follows:
+//mMSGHDR0 : Y data header (16x16)
+//mubMSGPAYLOAD0 : Y data payload (8 GRFs)
+//mMSGHDR1 : U data header (16x8)
+//mubMSGPAYLOAD1 : U data payload (4 GRFs)
+//------------------------------------------------------------------
+
+
+
+// Module name: Save.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.
+
+
+//Msg payload buffers; upto 4 full-size messages can be written
+
+
+.declare mudMSGPAYLOAD0 Base=r29.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
+.declare mudMSGPAYLOAD1 Base=r38.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
+.declare mudMSGPAYLOAD2 Base=r47.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
+.declare mudMSGPAYLOAD3 Base=r56.0 ElementSize=4 SrcRegion=<8;8,1> Type=ud
+
+.declare muwMSGPAYLOAD0 Base=r29.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
+.declare muwMSGPAYLOAD1 Base=r38.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
+.declare muwMSGPAYLOAD2 Base=r47.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
+.declare muwMSGPAYLOAD3 Base=r56.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw
+
+.declare mubMSGPAYLOAD0 Base=r29.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
+.declare mubMSGPAYLOAD1 Base=r38.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
+.declare mubMSGPAYLOAD2 Base=r47.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
+.declare mubMSGPAYLOAD3 Base=r56.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
+.declare mubMSGPAYLOAD4 Base=r32.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
+.declare mubMSGPAYLOAD5 Base=r41.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
+.declare mubMSGPAYLOAD6 Base=r50.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
+.declare mubMSGPAYLOAD7 Base=r59.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub
+
+
+ // the r17 register (nTEMP0) is originally defined from "Common.inc"
+ // instead of re-defining a nTEMP0 here, we use "SAVE_RGB" suffix for its naming
+
+ .declare uwTemp0 Base=r17.0 ElementSize=2 Type=uw
+
+
+//_SAVE_INC_
+
+
+ // At the save module we have all 8 address sub-registers available.
+ // So we will use PING-PONG type of scheme to save the data using
+ // pointers pBUF_CHNL_TOP_8x4 and pBUF_CHNL_BOT_8x4. This will help
+ // reduce dependency. - rT
+
+ //wBUFF_CHNL_PTR points to either buffer 0 or buffer 4.
+ //Add appropriate offsets to get pointers for all buffers (1,2,3 or 5).
+ //Offsets are zero for buffer 0 and buffer 4.
+ add (4) a0.0<1>:uw r22.0<4;4,1>:w 0:uw
+ add (4) a0.4<1>:uw r22.0<4;4,1>:w 512:uw
+
+ //Set up header for Y,U and V data
+ mov (8) r28<1>:ud r27<8;8,1>:ud
+ mov (8) r37<1>:ud r27<8;8,1>:ud
+
+ mov (2) r28.0<1>:d r7.0<2;2,1>:w { NoDDClr } //ORI Y (LUMA) = ORI
+ mov (1) r37.0<1>:d r7.0<0;1,0>:w { NoDDClr } //H ORI (CHROMA) = H ORI
+ shr (1) r37.1<1>:d r7.1<0;1,0>:w 1:w { NoDDClr, NoDDChk } //V ORI (CHROMA) = V ORI/2
+
+ mov (1) r28.2<1>:ud 0xF000F:ud { NoDDChk } // Y Block width and height (16x16)
+ mov (1) r37.2<1>:ud 0x7000F:ud { NoDDChk } // UV Block width and height(16x8)
+
+// Unscramble, and pack data directly to MRFs
+
+// Data 16x16 block is divided as -
+// ---------
+// | 0 |
+// ---------
+// | 1 |
+// ---------
+// | 2 |
+// ---------
+// | 3 |
+// ---------
+// All sub-blocks are of size 16x4
+// 0: ubBUFFER_0
+// 1: ubBUFFER_1, ubBUFFER_0+16
+// 2: ubBUFFER_2
+// 3: ubBUFFER_3, ubBUFFER_2+16
+
+ //Y Rounding 16x4 top part
+ add.sat (16) r[a0.1,0]<1>:uw r[a0.1,0]<16;16,1>:uw 0x0080:uw
+ add.sat (16) r[a0.1,32]<1>:uw r[a0.1,32]<16;16,1>:uw 0x0080:uw
+ add.sat (16) r[a0.1,64]<1>:uw r[a0.1,64]<16;16,1>:uw 0x0080:uw
+ add.sat (16) r[a0.1,96]<1>:uw r[a0.1,96]<16;16,1>:uw 0x0080:uw
+
+ // U Averaging and Rounding, 8x2 top part
+ shr (8) uwBUFFER_5(0,0)<2> r[a0.2,0]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(1,0)<2> r[a0.2,32]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(2,0)<2> r[a0.2,64]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(3,0)<2> r[a0.2,96]<16;8,2>:uw 1:w
+
+ add (8) uwBUFFER_5(0,0)<2> uwBUFFER_5(0,0)<16;8,2> uwBUFFER_5(1,0)<16;8,2>
+ add.sat (8) r[a0.2,0]<2>:uw uwBUFFER_5(0,0)<16;8,2> 0x0080:uw
+
+ add (8) uwBUFFER_5(2,0)<2> uwBUFFER_5(2,0)<16;8,2> uwBUFFER_5(3,0)<16;8,2>
+ add.sat (8) r[a0.2,64]<2>:uw uwBUFFER_5(2,0)<16;8,2> 0x0080:uw
+
+ // V Averaging and Rounding, 8x2 top part
+ shr (8) uwBUFFER_5(4,0)<2> r[a0.0,0]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(5,0)<2> r[a0.0,32]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(6,0)<2> r[a0.0,64]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(7,0)<2> r[a0.0,96]<16;8,2>:uw 1:w
+
+ add (8) uwBUFFER_5(4,0)<2> uwBUFFER_5(4,0)<16;8,2> uwBUFFER_5(5,0)<16;8,2>
+ add.sat (8) r[a0.0,0]<2>:uw uwBUFFER_5(4,0)<16;8,2> 0x0080:uw
+
+ add (8) uwBUFFER_5(6,0)<2> uwBUFFER_5(6,0)<16;8,2> uwBUFFER_5(7,0)<16;8,2>
+ add.sat (8) r[a0.0,64]<2>:uw uwBUFFER_5(6,0)<16;8,2> 0x0080:uw
+
+ add (4) a0.0<1>:uw r22.0<4;4,1>:w 1024:uw //Update Buffer 2 pointers
+
+ //Y Rounding, 16x4 bottom part
+ add.sat (16) r[a0.5,0]<1>:uw r[a0.5,0]<16;16,1>:uw 0x0080:uw
+ add.sat (16) r[a0.5,32]<1>:uw r[a0.5,32]<16;16,1>:uw 0x0080:uw
+ add.sat (16) r[a0.5,64]<1>:uw r[a0.5,64]<16;16,1>:uw 0x0080:uw
+ add.sat (16) r[a0.5,96]<1>:uw r[a0.5,96]<16;16,1>:uw 0x0080:uw
+
+ // U Averaging and Rounding, 8x2 bottom part
+ shr (8) uwBUFFER_5(0,0)<2> r[a0.6,0]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(1,0)<2> r[a0.6,32]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(2,0)<2> r[a0.6,64]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(3,0)<2> r[a0.6,96]<16;8,2>:uw 1:w
+
+ add (8) uwBUFFER_5(0,0)<2> uwBUFFER_5(0,0)<16;8,2> uwBUFFER_5(1,0)<16;8,2>
+ add.sat (8) r[a0.6,0]<2>:uw uwBUFFER_5(0,0)<16;8,2> 0x0080:uw
+
+ add (8) uwBUFFER_5(2,0)<2> uwBUFFER_5(2,0)<16;8,2> uwBUFFER_5(3,0)<16;8,2>
+ add.sat (8) r[a0.6,64]<2>:uw uwBUFFER_5(2,0)<16;8,2> 0x0080:uw
+
+ // V Averaging and Rounding, 8x2 bottom part
+ shr (8) uwBUFFER_5(4,0)<2> r[a0.4,0]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(5,0)<2> r[a0.4,32]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(6,0)<2> r[a0.4,64]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(7,0)<2> r[a0.4,96]<16;8,2>:uw 1:w
+
+ add (8) uwBUFFER_5(4,0)<2> uwBUFFER_5(4,0)<16;8,2> uwBUFFER_5(5,0)<16;8,2>
+ add.sat (8) r[a0.4,0]<2>:uw uwBUFFER_5(4,0)<16;8,2> 0x0080:uw
+
+ add (8) uwBUFFER_5(6,0)<2> uwBUFFER_5(6,0)<16;8,2> uwBUFFER_5(7,0)<16;8,2>
+ add.sat (8) r[a0.4,64]<2>:uw uwBUFFER_5(6,0)<16;8,2> 0x0080:uw
+
+ add (4) a0.4<1>:uw r22.0<4;4,1>:w 1536:uw //Update Buffer 3 pointers
+ //Y Rounding 16x4 top part
+ add.sat (16) r[a0.1,0]<1>:uw r[a0.1,0]<16;16,1>:uw 0x0080:uw
+ add.sat (16) r[a0.1,32]<1>:uw r[a0.1,32]<16;16,1>:uw 0x0080:uw
+ add.sat (16) r[a0.1,64]<1>:uw r[a0.1,64]<16;16,1>:uw 0x0080:uw
+ add.sat (16) r[a0.1,96]<1>:uw r[a0.1,96]<16;16,1>:uw 0x0080:uw
+
+ // U Averaging and Rounding, 8x2 top part
+ shr (8) uwBUFFER_5(0,0)<2> r[a0.2,0]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(1,0)<2> r[a0.2,32]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(2,0)<2> r[a0.2,64]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(3,0)<2> r[a0.2,96]<16;8,2>:uw 1:w
+
+ add (8) uwBUFFER_5(0,0)<2> uwBUFFER_5(0,0)<16;8,2> uwBUFFER_5(1,0)<16;8,2>
+ add.sat (8) r[a0.2,0]<2>:uw uwBUFFER_5(0,0)<16;8,2> 0x0080:uw
+
+ add (8) uwBUFFER_5(2,0)<2> uwBUFFER_5(2,0)<16;8,2> uwBUFFER_5(3,0)<16;8,2>
+ add.sat (8) r[a0.2,64]<2>:uw uwBUFFER_5(2,0)<16;8,2> 0x0080:uw
+
+ // V Averaging and Rounding, 8x2 top part
+ shr (8) uwBUFFER_5(4,0)<2> r[a0.0,0]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(5,0)<2> r[a0.0,32]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(6,0)<2> r[a0.0,64]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(7,0)<2> r[a0.0,96]<16;8,2>:uw 1:w
+
+ add (8) uwBUFFER_5(4,0)<2> uwBUFFER_5(4,0)<16;8,2> uwBUFFER_5(5,0)<16;8,2>
+ add.sat (8) r[a0.0,0]<2>:uw uwBUFFER_5(4,0)<16;8,2> 0x0080:uw
+
+ add (8) uwBUFFER_5(6,0)<2> uwBUFFER_5(6,0)<16;8,2> uwBUFFER_5(7,0)<16;8,2>
+ add.sat (8) r[a0.0,64]<2>:uw uwBUFFER_5(6,0)<16;8,2> 0x0080:uw
+
+ add (4) a0.0<1>:uw r22.0<4;4,1>:w 1024:uw //Update Buffer 2 pointers
+
+ //Y Rounding, 16x4 bottom part
+ add.sat (16) r[a0.5,0]<1>:uw r[a0.5,0]<16;16,1>:uw 0x0080:uw
+ add.sat (16) r[a0.5,32]<1>:uw r[a0.5,32]<16;16,1>:uw 0x0080:uw
+ add.sat (16) r[a0.5,64]<1>:uw r[a0.5,64]<16;16,1>:uw 0x0080:uw
+ add.sat (16) r[a0.5,96]<1>:uw r[a0.5,96]<16;16,1>:uw 0x0080:uw
+
+ // U Averaging and Rounding, 8x2 bottom part
+ shr (8) uwBUFFER_5(0,0)<2> r[a0.6,0]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(1,0)<2> r[a0.6,32]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(2,0)<2> r[a0.6,64]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(3,0)<2> r[a0.6,96]<16;8,2>:uw 1:w
+
+ add (8) uwBUFFER_5(0,0)<2> uwBUFFER_5(0,0)<16;8,2> uwBUFFER_5(1,0)<16;8,2>
+ add.sat (8) r[a0.6,0]<2>:uw uwBUFFER_5(0,0)<16;8,2> 0x0080:uw
+
+ add (8) uwBUFFER_5(2,0)<2> uwBUFFER_5(2,0)<16;8,2> uwBUFFER_5(3,0)<16;8,2>
+ add.sat (8) r[a0.6,64]<2>:uw uwBUFFER_5(2,0)<16;8,2> 0x0080:uw
+
+ // V Averaging and Rounding, 8x2 bottom part
+ shr (8) uwBUFFER_5(4,0)<2> r[a0.4,0]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(5,0)<2> r[a0.4,32]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(6,0)<2> r[a0.4,64]<16;8,2>:uw 1:w
+ shr (8) uwBUFFER_5(7,0)<2> r[a0.4,96]<16;8,2>:uw 1:w
+
+ add (8) uwBUFFER_5(4,0)<2> uwBUFFER_5(4,0)<16;8,2> uwBUFFER_5(5,0)<16;8,2>
+ add.sat (8) r[a0.4,0]<2>:uw uwBUFFER_5(4,0)<16;8,2> 0x0080:uw
+
+ add (8) uwBUFFER_5(6,0)<2> uwBUFFER_5(6,0)<16;8,2> uwBUFFER_5(7,0)<16;8,2>
+ add.sat (8) r[a0.4,64]<2>:uw uwBUFFER_5(6,0)<16;8,2> 0x0080:uw
+
+ add (4) a0.4<1>:uw r22.0<4;4,1>:w 1536:uw //Update Buffer 3 pointers
+ // restore pBUF_CHNL_TOP_8x4 and pBUF_CHNL_BOT_8x4 registers
+ add (4) a0.0<1>:uw r22.0<4;4,1>:w 0:uw
+ add (4) a0.4<1>:uw r22.0<4;4,1>:w 512:uw
+
+//Buffer 0
+//Move Y to msg payload
+ mov (16) mubMSGPAYLOAD0(0,0)<1> r[a0.1, 1]<32;16,2>:ub { NoDDClr }
+ mov (16) mubMSGPAYLOAD0(0,16)<1> r[a0.1, 33]<32;16,2>:ub { NoDDChk }
+ mov (16) mubMSGPAYLOAD0(1,0)<1> r[a0.1, 65]<32;16,2>:ub { NoDDClr }
+ mov (16) mubMSGPAYLOAD0(1,16)<1> r[a0.1, 97]<32;16,2>:ub { NoDDChk }
+
+//Move U to msg payload
+ mov (8) mubMSGPAYLOAD1(0,0)<2> r[a0.2, 1]<32;8,4>:ub { NoDDClr }
+ mov (8) mubMSGPAYLOAD1(0,16)<2> r[a0.2, 65]<32;8,4>:ub { NoDDClr, NoDDChk }
+
+//Move V to msg payload
+ mov (8) mubMSGPAYLOAD1(0,1)<2> r[a0.0, 1]<32;8,4>:ub { NoDDClr, NoDDChk }
+ mov (8) mubMSGPAYLOAD1(0,17)<2> r[a0.0, 65]<32;8,4>:ub { NoDDChk }
+
+ add (4) a0.0<1>:uw r22.0<4;4,1>:w 1024:uw //Update Buffer 2 pointers
+
+//Buffer 1
+ mov (16) mubMSGPAYLOAD0(2,0)<1> r[a0.5, 1]<32;16,2>:ub { NoDDClr }
+ mov (16) mubMSGPAYLOAD0(2,16)<1> r[a0.5, 33]<32;16,2>:ub { NoDDChk }
+ mov (16) mubMSGPAYLOAD0(3,0)<1> r[a0.5, 65]<32;16,2>:ub { NoDDClr }
+ mov (16) mubMSGPAYLOAD0(3,16)<1> r[a0.5, 97]<32;16,2>:ub { NoDDChk }
+
+ mov (8) mubMSGPAYLOAD1(1,0)<2> r[a0.6, 1]<32;8,4>:ub { NoDDClr }
+ mov (8) mubMSGPAYLOAD1(1,16)<2> r[a0.6, 65]<32;8,4>:ub { NoDDClr, NoDDChk }
+
+ mov (8) mubMSGPAYLOAD1(1,1)<2> r[a0.4, 1]<32;8,4>:ub { NoDDClr, NoDDChk }
+ mov (8) mubMSGPAYLOAD1(1,17)<2> r[a0.4, 65]<32;8,4>:ub { NoDDChk }
+
+ add (4) a0.4<1>:uw r22.0<4;4,1>:w 1536:uw //Update Buffer 3 pointers
+
+//Buffer 2
+ mov (16) mubMSGPAYLOAD0(4,0)<1> r[a0.1, 1]<32;16,2>:ub { NoDDClr }
+ mov (16) mubMSGPAYLOAD0(4,16)<1> r[a0.1, 33]<32;16,2>:ub { NoDDChk }
+ mov (16) mubMSGPAYLOAD0(5,0)<1> r[a0.1, 65]<32;16,2>:ub { NoDDClr }
+ mov (16) mubMSGPAYLOAD0(5,16)<1> r[a0.1, 97]<32;16,2>:ub { NoDDChk }
+
+ mov (8) mubMSGPAYLOAD1(2,0)<2> r[a0.2, 1]<32;8,4>:ub { NoDDClr }
+ mov (8) mubMSGPAYLOAD1(2,16)<2> r[a0.2, 65]<32;8,4>:ub { NoDDClr, NoDDChk }
+
+ mov (8) mubMSGPAYLOAD1(2,1)<2> r[a0.0, 1]<32;8,4>:ub { NoDDClr, NoDDChk }
+ mov (8) mubMSGPAYLOAD1(2,17)<2> r[a0.0, 65]<32;8,4>:ub { NoDDChk }
+
+//Buffer 3
+ mov (16) mubMSGPAYLOAD0(6,0)<1> r[a0.5, 1]<32;16,2>:ub { NoDDClr }
+ mov (16) mubMSGPAYLOAD0(6,16)<1> r[a0.5, 33]<32;16,2>:ub { NoDDChk }
+ mov (16) mubMSGPAYLOAD0(7,0)<1> r[a0.5, 65]<32;16,2>:ub { NoDDClr }
+ mov (16) mubMSGPAYLOAD0(7,16)<1> r[a0.5, 97]<32;16,2>:ub { NoDDChk }
+
+ mov (8) mubMSGPAYLOAD1(3,0)<2> r[a0.6, 1]<32;8,4>:ub { NoDDClr }
+ mov (8) mubMSGPAYLOAD1(3,16)<2> r[a0.6, 65]<32;8,4>:ub { NoDDClr, NoDDChk }
+
+ mov (8) mubMSGPAYLOAD1(3,1)<2> r[a0.4, 1]<32;8,4>:ub { NoDDClr, NoDDChk }
+ mov (8) mubMSGPAYLOAD1(3,17)<2> r[a0.4, 65]<32;8,4>:ub { NoDDChk }
+//===========================================================================
+
+send (1) null<1>:d r28 0xc 0x120A8018:ud
+send (1) null<1>:d r37 0xc 0xA0A8019:ud
--- /dev/null
+/*
+ * Copyright 2000-2011 Intel Corporation All Rights Reserved
+ *
+ * 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.
+ */
+// 7 // 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=====================================
+
+
+//=======================================================================
+
+
+//r9-r17
+// Define temp space for any usages
+
+
+// Common Buffers
+
+
+// temp space for rotation
+
+.declare fROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+
+.declare udROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+
+.declare uwROBUF Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+.declare ubROBUF Base=r9.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
+
+.declare ub4ROBUF Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
+
+
+// End of common.inc
+
+
+//Module Name: Set_AVS_Buf_0123_PL2.asm
+
+
+
+//Module Name: Set_Buf_0123_PL2
+
+
+
+
+// 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.
+
+
+ //AVS LAYOUT: (YYUUVVAA)
+ //Assign buffer channel order for Buffer 0123 in the order AUYV a0.3>A, a0.2>U, a0.1>Y, a0.0>V
+ //For PL2-AVS: V = 8, Y= 0, U = 4, A = 12.
+ mov (4) acc0.0<1>:w 0x6EA2:v //Subtract 6 from 0,4,8,12
+ add (4) acc0.0<1>:w acc0<4;4,1>:w 70:uw //add 6 back
+ shl (4) r22.0<1>:w acc0<4;4,1>:w 5:uw //Convert to BYTE address.
+
+ //OPT: wAVS_SU_SHUFFLE_PTR_0 and udAVS_SU_SHUFFLE_OFF_0 are sub-regs of same GRF. -rT
+ //SU LAYOUT:(YUVAYUVA)
+ //V = 4, Y = 0, U = 2, A = 6
+ mov (4) acc0.0<1>:w 0x6204:v
+ add (4) acc0.0<1>:w acc0<4;4,1>:w 64:uw
+ shl (4) r18.0<1>:w acc0<4;4,1>:w 5:uw { NoDDClr } //Convert to BYTE address.
+
+ //OFFSET:
+ mov (1) r18.4<1>:ud 0x1000100:ud { NoDDChk }
+
+
--- /dev/null
+/*
+ * Copyright 2000-2011 Intel Corporation All Rights Reserved
+ *
+ * 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.
+ */
+// 18 // 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 MSG_AVS_SAMPLE 0x00000000
+#define MSG_CONVOLE_SAMPLE 0x10000000
+#define MSG_MINMAX_SAMPLE 0x20000000
+#define MSG_MINMAXF_SAMPLE 0x30000000
+#define MSG_ERODE_SAMPLE 0x40000000
+#define MSG_DILATE_SAMPLE 0x50000000
+#define MSG_BOOLCENT_SAMPLE 0x60000000
+#define MSG_CENTROID_SAMPLE 0x70000000
+
+#define MSG_IEF_BYPASS 0x08000000
+#define MSG_IEF_ENABLE 0x00000000
+
+//16x4 or 8x4 or 16x8 or 4x4
+#define MSG_AVS_164 0x00000000
+#define MSG_AVS_84 0x02000000
+#define MSG_AVS_168 0x04000000
+#define MSG_AVS_44 0x06000000
+
+//#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=====================================
+
+
+//=======================================================================
+
+
+//r9-r17
+// Define temp space for any usages
+
+
+// Common Buffers
+
+
+// temp space for rotation
+
+.declare fROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+
+.declare udROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+
+.declare uwROBUF Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+.declare ubROBUF Base=r9.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
+
+.declare ub4ROBUF Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
+
+
+// End of common.inc
+
+
+
+
+
+
+//Module name: Set_Layer_N.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.
+
+
+//Used to generate LABELS at compile time.
+
+
+//definitions for Expand Mask
+.declare uwMask_Temp1 Base=r17.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw // 1 GRF
+.declare ubMask_Temp1 Base=r17.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub // 1 GRF
+.declare udMask_Temp1 Base=r17.0 ElementSize=4 Type=ud // 1 GRF
+.declare uwMask_Temp2 Base=r16.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw // 1 GRF
+.declare ubMask_Temp2 Base=r16.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub // 1 GRF
+.declare udMask_Temp2 Base=r16.0 ElementSize=4 Type=ud // 1 GRF
+
+.declare uwMask_Temp3 Base=r15.0 ElementSize=2 SrcRegion=<16;16,1> Type=uw // 1 GRF
+.declare ubMask_Temp3 Base=r15.0 ElementSize=1 SrcRegion=<16;16,1> Type=ub // 1 GRF
+
+.declare udALPHA_MASK_REG Base=r21.0 ElementSize=4 Type=ud // 1 GRF
+.declare udALPHA_MASK_REG_TEMP Base=r20.0 ElementSize=4 Type=ud // 1 GRF
+
+
+//Initialize mask reg to FFFF
+
+ mov (16) uwALPHA_MASK_REG(0)<1> 0xFFFF:uw
+
+
+//Fast jump for -
+//LAYER0: we determine whether layer 0 is to be loaded and processed or not based
+// on block mask in module "Set_Layer_0" and store result in f0.1.
+// This flag is then directly used to while loading buf0-3 and colorfill.
+// (So flag f0.1 should not be changed from Set_Layer_0 till Colorfill)
+//
+//LAYER1-7: For all other layers, we compute whether layer is to be loaded and processed
+// based on block mask in module "Set_Layer_1-7" and store result in SKIP_LAYER
+// variable.
+// While Loading buf 4 and 5, we move SKIP_LAYER to f0.0 every time and use it
+// for Loading.
+// For processing though, we move SKIP_LAYER only once to f0.1 in module
+// "Set_Buf0_Buf4" and use f0.1 for deciding whether layer 1-7 (all 4 sub blocks)
+// is to be processed or not.
+// (So flag f0.1) should not be modififed from module "Set_Buf0_Buf4" till module
+// that processess sub-block 3).
+//
+//None of the above fast jumps, apply to CSC modules. We always perform CSC irrespective of mask.
+//
+//Example: (Without going into finer details)
+// Typical Combined kernel:
+//
+// (let var = decision whether to load/process that layer)
+//
+// Set_Layer_0 //f0.1 <- var
+// ..
+// Set_Layer_1 //f0.1 <- var, SKIP_LAYER <- var
+// ..
+// Load buf 0 //use f0.1
+// Load buf 4 //f0.0 <- SKIP_LAYER
+// Load buf 1 //use f0.1
+// Load buf 5 //f0.0 <- SKIP_LAYER
+// Load buf 2 //use f0.1
+// Load buf 3 //use f0.1
+// ..
+// ..
+// Colorfill
+// ..
+// Set_Buf0_Buf4 //f0.1 <- SKIP_LAYER
+// process0-4 //Use f0.1
+// Load buf 4
+// Set_Buf1_Buf5
+// process1-5
+// Load buf 5
+// ..
+// Set_Layer_2 //f0.1 <-var, SKIP_LAYER <- var
+// ..
+// Set_Buf2_Buf4
+// process2-4
+// Load buf 4
+// Set_Buf3_Buf5
+// process3-5
+// Load buf 5
+// ..
+
+
+ and (1) r24.2<1>:ub r2.2<0;1,0>:uw 3:uw
+
+
+ //Copy all AVS Payload data
+ // Setup Message Payload Header for 1st block of Media Sampler 8x8 (16x4 for IVB+)
+ //currently the dx & dy is passed by Constant buffer (zero)
+ mov (1) r25.0<1>:f r7.6<0;1,0>:f //NLAS dy
+ mov (1) r25.6<1>:f r7.5<0;1,0>:f //NLAS dx
+ mov (1) r25.4<1>:f r3.0<0;1,0>:f //Step X
+ mov (1) r25.5<1>:f r4.0<0;1,0>:f //Step Y
+
+
+ mov (1) r25.2<1>:f r6.0<0;1,0>:f //Orig X
+ mov (1) r25.3<1>:f r5.0<0;1,0>:f //Orig Y
+
+ mov (1) r25.7<1>:ud 0:ud
+ add (1) r25.7<1>:ud r25.7<0;1,0>:ud MSG_AVS_SAMPLE + MSG_AVS_164 + MSG_IEF_BYPASS:ud
+
+ //NLAS calculations for 2nd half of blocks of Media Sampler 8x8:
+ // X(i) = X0 + dx*i + ddx*i*(i-1)/2 ==> X(8) = X0 + dx*8 +ddx*28
+ // dx(i)= dx(0) + ddx*i ==> dx(8)= dx + ddx*8
+
+ //OPTIMIZATION: fORIGIN_X_NLAS, fY_OFFSET_2ND_BLOCK, fSTEP_X_NLAS, pMSGDSC_COPY, ubCONST_ALPHA_COPY - are sub registers of same GRF. Use NODDCLR NODDCHK. -rT
+
+ // Calculating X(8)
+ mov (1) acc0.2<1>:f r6.0<0;1,0>:f
+ mac (1) acc0.2<1>:f r3.0<0;1,0>:f 8.0:f
+ mac (1) r23.2<1>:f r7.5<0;1,0>:f 28.0:f { NoDDClr }
+
+ // Calculating Y(4)
+ mul (1) r23.1<1>:f r4.0<0;1,0>:f 4.0:f { NoDDClr, NoDDChk } //dY*4
+
+ // Calculating dx(8)
+ mov (1) acc0.4<1>:f r3.0<0;1,0>:f
+ mac (1) r23.4<1>:f r7.5<0;1,0>:f 8.0:f { NoDDClr, NoDDChk }
+
+ // Binding Index
+ mov (1) r23.5<1>:ud 0:ud { NoDDChk }
+
+
+SKIP_LAYER_L0:
+ nop
+
+
--- /dev/null
+/*
+ * Copyright 2000-2011 Intel Corporation All Rights Reserved
+ *
+ * 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.
+ */
+/*
+ * Copyright 2000-2011 Intel Corporation All Rights Reserved
+ *
+ * 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.
+ * Authors: Zhao Yakui <yakui.zhao@intel.com>
+ */
+
+// 326 // 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=====================================
+
+
+//=======================================================================
+
+
+//r9-r17
+// Define temp space for any usages
+
+
+// Common Buffers
+
+
+// temp space for rotation
+
+.declare fROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=f
+
+.declare udROBUF Base=r9.0 ElementSize=4 SrcRegion=<8;8,1> DstRegion=<1> Type=ud
+
+.declare uwROBUF Base=r9.0 ElementSize=2 SrcRegion=<16;16,1> DstRegion=<1> Type=uw
+
+.declare ubROBUF Base=r9.0 ElementSize=1 SrcRegion=<16;16,1> DstRegion=<1> Type=ub
+
+.declare ub4ROBUF Base=r9.0 ElementSize=1 SrcRegion=<32;8,4> DstRegion=<4> Type=ub
+
+
+// End of common.inc
+
+
+// FileName: VP_Setup.asm
+// Author: Vivek Kumar
+// Description: Sets up all parameters for the Video Processing Kernel
+
+
+
+
+// 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.
+
+
+//Setup pointer to the inline parameter
+
+// Copy MSG HDR
+ mov (8) r27.0<1>:ud r0.0<8;8,1>:ud // Initialize message payload header with R0
+
+// Only one layer is enough
+
+//temp; remove it once unread msg warnings are resolved -vK
+mov (8) r25<1>:ud r0.0<8;8,1>:ud
+mov (8) r26<1>:ud r0.0<8;8,1>:ud
+
+// Calculate StepX for all layers and overwrite it on the ratio
+ mul (8) r3.0<1>:f r3.0<8;8,1>:f r7.4<0;1,0>:f //StepX_ratio = StepX / VideoStepX
+
+ //Normalised Ratio of Horizontal step size with main video for all layers now becomes
+ //Normalised Horizontal step size for all layers
+
+// Calculate block origin for all layers and overwrite it on the frame origin
+ mov (2) r8.5<1>:f r7.0<2;2,1>:w //Convert origin from word to float
+
+ cmp.e.f0.0 (1) null<1>:d r2.26<0;1,0>:ub 1:uw
+
+
+ shr (1) r17.0<1>:uw r2.2<0;1,0>:uw 0:uw
+ and (1) r17.0<1>:uw r17.0<0;1,0>:uw 3:uw
+ cmp.e.f0.1 (1) null<1>:w r17.0<0;1,0>:uw 1:uw
+ (f0.1) jmpi (1) ROTATE_90_L0
+ cmp.e.f0.1 (1) null<1>:w r17.0<0;1,0>:uw 2:uw
+ (f0.1) jmpi (1) ROTATE_180_L0
+ cmp.e.f0.1 (1) null<1>:w r17.0<0;1,0>:uw 3:uw
+ (f0.1) jmpi (1) ROTATE_270_L0
+
+ // rotate 0 degree
+ROTATE_0_L0:
+ (-f0.0)mov (1) acc0.0<1>:f r6.0<0;1,0>:f
+ (-f0.0)mac (1) r6.0<1>:f r3.0<0;1,0>:f r8.5<0;1,0>:f
+
+ mov (1) acc0.0<1>:f r5.0<0;1,0>:f
+ mac (1) r5.0<1>:f r4.0<0;1,0>:f r8.6<0;1,0>:f
+ jmpi (1) END_SRC_BLOCK_ORIG_COMP_L0
+
+ // rotate 90 degree
+ROTATE_90_L0:
+ (-f0.0)mov (1) acc0.0<1>:f r6.0<0;1,0>:f
+ (-f0.0)mac (1) r6.0<1>:f r3.0<0;1,0>:f r8.6<0;1,0>:f
+
+ mov (1) r16.0<1>:f r2.0<0;1,0>:uw
+ add (1) r17.0<1>:f -r8.5<0;1,0>:f r16.0<0;1,0>:f
+ add (1) r17.0<1>:f r17.0<0;1,0>:f -16.0:f
+
+ mov (1) acc0.0<1>:f r5.0<0;1,0>:f
+ mac (1) r5.0<1>:f r4.0<0;1,0>:f r17.0<0;1,0>:f
+ jmpi (1) END_SRC_BLOCK_ORIG_COMP_L0
+
+ // rotate 180 degree
+ROTATE_180_L0:
+ (-f0.0)mov (1) r16.0<1>:f r2.0<0;1,0>:uw
+ (-f0.0)add (1) r17.0<1>:f -r8.5<0;1,0>:f r16.0<0;1,0>:f
+ (-f0.0)add (1) r17.0<1>:f r17.0<0;1,0>:f -16.0:f
+ (-f0.0)mov (1) acc0.0<1>:f r6.0<0;1,0>:f
+ (-f0.0)mac (1) r6.0<1>:f r3.0<0;1,0>:f r17.0<0;1,0>:f
+
+ mov (1) r16.0<1>:f r2.1<0;1,0>:uw
+ add (1) r17.0<1>:f -r8.6<0;1,0>:f r16.0<0;1,0>:f
+ add (1) r17.0<1>:f r17.0<0;1,0>:f -16.0:f
+ mov (1) acc0.0<1>:f r5.0<0;1,0>:f
+ mac (1) r5.0<1>:f r4.0<0;1,0>:f r17.0<0;1,0>:f
+ jmpi (1) END_SRC_BLOCK_ORIG_COMP_L0
+
+ // rotate 270 degree
+ROTATE_270_L0:
+ (-f0.0)mov (1) r16.0<1>:f r2.1<0;1,0>:uw
+ (-f0.0)add (1) r17.0<1>:f -r8.6<0;1,0>:f r16.0<0;1,0>:f
+ (-f0.0)add (1) r17.0<1>:f r17.0<0;1,0>:f -16.0:f
+ (-f0.0)mov (1) acc0.0<1>:f r6.0<0;1,0>:f
+ (-f0.0)mac (1) r6.0<1>:f r3.0<0;1,0>:f r17.0<0;1,0>:f
+
+ mov (1) acc0.0<1>:f r5.0<0;1,0>:f
+ mac (1) r5.0<1>:f r4.0<0;1,0>:f r8.5<0;1,0>:f
+
+END_SRC_BLOCK_ORIG_COMP_L0:
+ nop
--- /dev/null
+// Module name: AVS
+.kernel PL2_TO_PL2
+.code
+
+#include "VP_Setup.g8a"
+#include "Set_Layer_0.g8a"
+#include "Set_AVS_Buf_0123_PL2.g8a"
+#include "PL2_AVS_Buf_0.g8a"
+#include "PL2_AVS_Buf_1.g8a"
+#include "PL2_AVS_Buf_2.g8a"
+#include "PL2_AVS_Buf_3.g8a"
+#include "Save_AVS_NV12.g8a"
+#include "EOT.g8a"
+
+.end_code
+
+.end_kernel
--- /dev/null
+ { 0x00600001, 0x23600208, 0x008d0000, 0x00000000 },
+ { 0x00600001, 0x23200208, 0x008d0000, 0x00000000 },
+ { 0x00600001, 0x23400208, 0x008d0000, 0x00000000 },
+ { 0x00600041, 0x20603ae8, 0x3a8d0060, 0x000000f0 },
+ { 0x00200001, 0x21141ae8, 0x004500e0, 0x00000000 },
+ { 0x01000010, 0x20002220, 0x1600005a, 0x00010001 },
+ { 0x00000008, 0x22201248, 0x16000044, 0x00000000 },
+ { 0x00000005, 0x22201248, 0x16000220, 0x00030003 },
+ { 0x01000010, 0x20001261, 0x16000220, 0x00010001 },
+ { 0x00010020, 0x34000005, 0x0e001400, 0x00000090 },
+ { 0x01000010, 0x20001261, 0x16000220, 0x00020002 },
+ { 0x00010020, 0x34000005, 0x0e001400, 0x000000f0 },
+ { 0x01000010, 0x20001261, 0x16000220, 0x00030003 },
+ { 0x00010020, 0x34000005, 0x0e001400, 0x00000180 },
+ { 0x00110001, 0x24003ae0, 0x000000c0, 0x00000000 },
+ { 0x00110048, 0x20c03ae8, 0x3a000060, 0x00000114 },
+ { 0x00000001, 0x24003ae0, 0x000000a0, 0x00000000 },
+ { 0x00000048, 0x20a03ae8, 0x3a000080, 0x00000118 },
+ { 0x00000020, 0x34000004, 0x0e001400, 0x000001a0 },
+ { 0x00110001, 0x24003ae0, 0x000000c0, 0x00000000 },
+ { 0x00110048, 0x20c03ae8, 0x3a000060, 0x00000118 },
+ { 0x00000001, 0x220012e8, 0x00000040, 0x00000000 },
+ { 0x00000040, 0x22203ae8, 0x3a004114, 0x00000200 },
+ { 0x00000040, 0x22203ae8, 0x3e000220, 0xc1800000 },
+ { 0x00000001, 0x24003ae0, 0x000000a0, 0x00000000 },
+ { 0x00000048, 0x20a03ae8, 0x3a000080, 0x00000220 },
+ { 0x00000020, 0x34000004, 0x0e001400, 0x00000120 },
+ { 0x00110001, 0x220012e8, 0x00000040, 0x00000000 },
+ { 0x00110040, 0x22203ae8, 0x3a004114, 0x00000200 },
+ { 0x00110040, 0x22203ae8, 0x3e000220, 0xc1800000 },
+ { 0x00110001, 0x24003ae0, 0x000000c0, 0x00000000 },
+ { 0x00110048, 0x20c03ae8, 0x3a000060, 0x00000220 },
+ { 0x00000001, 0x220012e8, 0x00000042, 0x00000000 },
+ { 0x00000040, 0x22203ae8, 0x3a004118, 0x00000200 },
+ { 0x00000040, 0x22203ae8, 0x3e000220, 0xc1800000 },
+ { 0x00000001, 0x24003ae0, 0x000000a0, 0x00000000 },
+ { 0x00000048, 0x20a03ae8, 0x3a000080, 0x00000220 },
+ { 0x00000020, 0x34000004, 0x0e001400, 0x00000070 },
+ { 0x00110001, 0x220012e8, 0x00000042, 0x00000000 },
+ { 0x00110040, 0x22203ae8, 0x3a004118, 0x00000200 },
+ { 0x00110040, 0x22203ae8, 0x3e000220, 0xc1800000 },
+ { 0x00110001, 0x24003ae0, 0x000000c0, 0x00000000 },
+ { 0x00110048, 0x20c03ae8, 0x3a000060, 0x00000220 },
+ { 0x00000001, 0x24003ae0, 0x000000a0, 0x00000000 },
+ { 0x00000048, 0x20a03ae8, 0x3a000080, 0x00000114 },
+ { 0x0000007e, 0x00000000, 0x00000000, 0x00000000 },
+ { 0x00800001, 0x22a01648, 0x10000000, 0xffffffff },
+ { 0x00000005, 0x23021288, 0x16000044, 0x00030003 },
+ { 0x00000001, 0x23203ae8, 0x000000f8, 0x00000000 },
+ { 0x00000001, 0x23383ae8, 0x000000f4, 0x00000000 },
+ { 0x00000001, 0x23303ae8, 0x00000060, 0x00000000 },
+ { 0x00000001, 0x23343ae8, 0x00000080, 0x00000000 },
+ { 0x00000001, 0x23283ae8, 0x000000c0, 0x00000000 },
+ { 0x00000001, 0x232c3ae8, 0x000000a0, 0x00000000 },
+ { 0x00000001, 0x233c0608, 0x00000000, 0x00000000 },
+ { 0x00000040, 0x233c0208, 0x0600033c, 0x08000000 },
+ { 0x00000001, 0x24083ae0, 0x000000c0, 0x00000000 },
+ { 0x00000048, 0x24083ae0, 0x3e000060, 0x41000000 },
+ { 0x00000248, 0x22e83ae8, 0x3e0000f4, 0x41e00000 },
+ { 0x00000641, 0x22e43ae8, 0x3e000080, 0x40800000 },
+ { 0x00000001, 0x24103ae0, 0x00000060, 0x00000000 },
+ { 0x00000648, 0x22f03ae8, 0x3e0000f4, 0x41000000 },
+ { 0x00000401, 0x22f40608, 0x00000000, 0x00000000 },
+ { 0x0000007e, 0x00000000, 0x00000000, 0x00000000 },
+ { 0x00400001, 0x24003660, 0x30000000, 0x00006ea2 },
+ { 0x00400040, 0x24001860, 0x16690400, 0x00460046 },
+ { 0x00400009, 0x22c01868, 0x16690400, 0x00050005 },
+ { 0x00400001, 0x24003660, 0x30000000, 0x00006204 },
+ { 0x00400040, 0x24001860, 0x16690400, 0x00400040 },
+ { 0x00400209, 0x22401868, 0x16690400, 0x00050005 },
+ { 0x00000401, 0x22500608, 0x00000000, 0x01000100 },
+ { 0x00000001, 0x22d00608, 0x00000000, 0x00400040 },
+ { 0x00000001, 0x220c0208, 0x0000000c, 0x00000000 },
+ { 0x00000040, 0x22000200, 0x060002f4, 0x044eb000 },
+ { 0x00000001, 0x22080608, 0x00000000, 0x0000d000 },
+ { 0x00600001, 0x22200208, 0x008d0320, 0x00000000 },
+ { 0x02000031, 0x28002248, 0x00000200, 0x00000200 },
+ { 0x00000040, 0x22000200, 0x060002f4, 0x048eb001 },
+ { 0x00000001, 0x22080608, 0x00000000, 0x0000a000 },
+ { 0x02000031, 0x28802248, 0x00000200, 0x00000200 },
+ { 0x0000007e, 0x00000000, 0x00000000, 0x00000000 },
+ { 0x00000001, 0x220c0208, 0x0000000c, 0x00000000 },
+ { 0x00000040, 0x22000200, 0x060002f4, 0x044eb000 },
+ { 0x00000001, 0x22080608, 0x00000000, 0x0000d000 },
+ { 0x00000001, 0x23240608, 0x00000000, 0x00000001 },
+ { 0x00600001, 0x22200208, 0x008d0320, 0x00000000 },
+ { 0x02000031, 0x2a002248, 0x00000200, 0x00000200 },
+ { 0x00000040, 0x22000200, 0x060002f4, 0x048eb001 },
+ { 0x00000001, 0x22080608, 0x00000000, 0x0000a000 },
+ { 0x02000031, 0x2a802248, 0x00000200, 0x00000200 },
+ { 0x0000007e, 0x00000000, 0x00000000, 0x00000000 },
+ { 0x00000001, 0x220c0208, 0x0000000c, 0x00000000 },
+ { 0x00000040, 0x22000200, 0x060002f4, 0x044eb000 },
+ { 0x00000001, 0x22080608, 0x00000000, 0x0000d000 },
+ { 0x00000001, 0x23240608, 0x00000000, 0x00000002 },
+ { 0x00600001, 0x22200208, 0x008d0320, 0x00000000 },
+ { 0x02000031, 0x2c002248, 0x00000200, 0x00000200 },
+ { 0x00000040, 0x22000200, 0x060002f4, 0x048eb001 },
+ { 0x00000001, 0x22080608, 0x00000000, 0x0000a000 },
+ { 0x02000031, 0x2c802248, 0x00000200, 0x00000200 },
+ { 0x0000007e, 0x00000000, 0x00000000, 0x00000000 },
+ { 0x00000001, 0x220c0208, 0x0000000c, 0x00000000 },
+ { 0x00000040, 0x22000200, 0x060002f4, 0x044eb000 },
+ { 0x00000001, 0x22080608, 0x00000000, 0x0000d000 },
+ { 0x00000001, 0x23240608, 0x00000000, 0x00000003 },
+ { 0x00600001, 0x22200208, 0x008d0320, 0x00000000 },
+ { 0x02000031, 0x2e002248, 0x00000200, 0x00000200 },
+ { 0x00000040, 0x22000200, 0x060002f4, 0x048eb001 },
+ { 0x00000001, 0x22080608, 0x00000000, 0x0000a000 },
+ { 0x02000031, 0x2e802248, 0x00000200, 0x00000200 },
+ { 0x0000007e, 0x00000000, 0x00000000, 0x00000000 },
+ { 0x00400040, 0x22001a40, 0x166902c0, 0x00000000 },
+ { 0x00400040, 0x22081a40, 0x166902c0, 0x02000200 },
+ { 0x00600001, 0x23800208, 0x008d0360, 0x00000000 },
+ { 0x00600001, 0x24a00208, 0x008d0360, 0x00000000 },
+ { 0x00200201, 0x23801a28, 0x004500e0, 0x00000000 },
+ { 0x00000201, 0x24a01a28, 0x000000e0, 0x00000000 },
+ { 0x00000608, 0x24a41a28, 0x1e0000e2, 0x00010001 },
+ { 0x00000401, 0x23880608, 0x00000000, 0x000f000f },
+ { 0x00000401, 0x24a80608, 0x00000000, 0x0007000f },
+ { 0x80800040, 0xa2001248, 0x16b18200, 0x00800080 },
+ { 0x80800040, 0xa2201248, 0x16b18220, 0x00800080 },
+ { 0x80800040, 0xa2401248, 0x16b18240, 0x00800080 },
+ { 0x80800040, 0xa2601248, 0x16b18260, 0x00800080 },
+ { 0x00600008, 0x45c01248, 0x1eae8400, 0x00010001 },
+ { 0x00600008, 0x45e01248, 0x1eae8420, 0x00010001 },
+ { 0x00600008, 0x46001248, 0x1eae8440, 0x00010001 },
+ { 0x00600008, 0x46201248, 0x1eae8460, 0x00010001 },
+ { 0x00600040, 0x45c01248, 0x12ae05c0, 0x00ae05e0 },
+ { 0x80600040, 0xc4001248, 0x16ae05c0, 0x00800080 },
+ { 0x00600040, 0x46001248, 0x12ae0600, 0x00ae0620 },
+ { 0x80600040, 0xc4401248, 0x16ae0600, 0x00800080 },
+ { 0x00600008, 0x46401248, 0x1eae8000, 0x00010001 },
+ { 0x00600008, 0x46601248, 0x1eae8020, 0x00010001 },
+ { 0x00600008, 0x46801248, 0x1eae8040, 0x00010001 },
+ { 0x00600008, 0x46a01248, 0x1eae8060, 0x00010001 },
+ { 0x00600040, 0x46401248, 0x12ae0640, 0x00ae0660 },
+ { 0x80600040, 0xc0001248, 0x16ae0640, 0x00800080 },
+ { 0x00600040, 0x46801248, 0x12ae0680, 0x00ae06a0 },
+ { 0x80600040, 0xc0401248, 0x16ae0680, 0x00800080 },
+ { 0x00400040, 0x22001a40, 0x166902c0, 0x04000400 },
+ { 0x80800040, 0xaa001248, 0x16b18a00, 0x00800080 },
+ { 0x80800040, 0xaa201248, 0x16b18a20, 0x00800080 },
+ { 0x80800040, 0xaa401248, 0x16b18a40, 0x00800080 },
+ { 0x80800040, 0xaa601248, 0x16b18a60, 0x00800080 },
+ { 0x00600008, 0x45c01248, 0x1eae8c00, 0x00010001 },
+ { 0x00600008, 0x45e01248, 0x1eae8c20, 0x00010001 },
+ { 0x00600008, 0x46001248, 0x1eae8c40, 0x00010001 },
+ { 0x00600008, 0x46201248, 0x1eae8c60, 0x00010001 },
+ { 0x00600040, 0x45c01248, 0x12ae05c0, 0x00ae05e0 },
+ { 0x80600040, 0xcc001248, 0x16ae05c0, 0x00800080 },
+ { 0x00600040, 0x46001248, 0x12ae0600, 0x00ae0620 },
+ { 0x80600040, 0xcc401248, 0x16ae0600, 0x00800080 },
+ { 0x00600008, 0x46401248, 0x1eae8800, 0x00010001 },
+ { 0x00600008, 0x46601248, 0x1eae8820, 0x00010001 },
+ { 0x00600008, 0x46801248, 0x1eae8840, 0x00010001 },
+ { 0x00600008, 0x46a01248, 0x1eae8860, 0x00010001 },
+ { 0x00600040, 0x46401248, 0x12ae0640, 0x00ae0660 },
+ { 0x80600040, 0xc8001248, 0x16ae0640, 0x00800080 },
+ { 0x00600040, 0x46801248, 0x12ae0680, 0x00ae06a0 },
+ { 0x80600040, 0xc8401248, 0x16ae0680, 0x00800080 },
+ { 0x00400040, 0x22081a40, 0x166902c0, 0x06000600 },
+ { 0x80800040, 0xa2001248, 0x16b18200, 0x00800080 },
+ { 0x80800040, 0xa2201248, 0x16b18220, 0x00800080 },
+ { 0x80800040, 0xa2401248, 0x16b18240, 0x00800080 },
+ { 0x80800040, 0xa2601248, 0x16b18260, 0x00800080 },
+ { 0x00600008, 0x45c01248, 0x1eae8400, 0x00010001 },
+ { 0x00600008, 0x45e01248, 0x1eae8420, 0x00010001 },
+ { 0x00600008, 0x46001248, 0x1eae8440, 0x00010001 },
+ { 0x00600008, 0x46201248, 0x1eae8460, 0x00010001 },
+ { 0x00600040, 0x45c01248, 0x12ae05c0, 0x00ae05e0 },
+ { 0x80600040, 0xc4001248, 0x16ae05c0, 0x00800080 },
+ { 0x00600040, 0x46001248, 0x12ae0600, 0x00ae0620 },
+ { 0x80600040, 0xc4401248, 0x16ae0600, 0x00800080 },
+ { 0x00600008, 0x46401248, 0x1eae8000, 0x00010001 },
+ { 0x00600008, 0x46601248, 0x1eae8020, 0x00010001 },
+ { 0x00600008, 0x46801248, 0x1eae8040, 0x00010001 },
+ { 0x00600008, 0x46a01248, 0x1eae8060, 0x00010001 },
+ { 0x00600040, 0x46401248, 0x12ae0640, 0x00ae0660 },
+ { 0x80600040, 0xc0001248, 0x16ae0640, 0x00800080 },
+ { 0x00600040, 0x46801248, 0x12ae0680, 0x00ae06a0 },
+ { 0x80600040, 0xc0401248, 0x16ae0680, 0x00800080 },
+ { 0x00400040, 0x22001a40, 0x166902c0, 0x04000400 },
+ { 0x80800040, 0xaa001248, 0x16b18a00, 0x00800080 },
+ { 0x80800040, 0xaa201248, 0x16b18a20, 0x00800080 },
+ { 0x80800040, 0xaa401248, 0x16b18a40, 0x00800080 },
+ { 0x80800040, 0xaa601248, 0x16b18a60, 0x00800080 },
+ { 0x00600008, 0x45c01248, 0x1eae8c00, 0x00010001 },
+ { 0x00600008, 0x45e01248, 0x1eae8c20, 0x00010001 },
+ { 0x00600008, 0x46001248, 0x1eae8c40, 0x00010001 },
+ { 0x00600008, 0x46201248, 0x1eae8c60, 0x00010001 },
+ { 0x00600040, 0x45c01248, 0x12ae05c0, 0x00ae05e0 },
+ { 0x80600040, 0xcc001248, 0x16ae05c0, 0x00800080 },
+ { 0x00600040, 0x46001248, 0x12ae0600, 0x00ae0620 },
+ { 0x80600040, 0xcc401248, 0x16ae0600, 0x00800080 },
+ { 0x00600008, 0x46401248, 0x1eae8800, 0x00010001 },
+ { 0x00600008, 0x46601248, 0x1eae8820, 0x00010001 },
+ { 0x00600008, 0x46801248, 0x1eae8840, 0x00010001 },
+ { 0x00600008, 0x46a01248, 0x1eae8860, 0x00010001 },
+ { 0x00600040, 0x46401248, 0x12ae0640, 0x00ae0660 },
+ { 0x80600040, 0xc8001248, 0x16ae0640, 0x00800080 },
+ { 0x00600040, 0x46801248, 0x12ae0680, 0x00ae06a0 },
+ { 0x80600040, 0xc8401248, 0x16ae0680, 0x00800080 },
+ { 0x00400040, 0x22081a40, 0x166902c0, 0x06000600 },
+ { 0x00400040, 0x22001a40, 0x166902c0, 0x00000000 },
+ { 0x00400040, 0x22081a40, 0x166902c0, 0x02000200 },
+ { 0x00800201, 0x23a02288, 0x00d28201, 0x00000000 },
+ { 0x00800401, 0x23b02288, 0x00d28221, 0x00000000 },
+ { 0x00800201, 0x23c02288, 0x00d28241, 0x00000000 },
+ { 0x00800401, 0x23d02288, 0x00d28261, 0x00000000 },
+ { 0x00600201, 0x44c02288, 0x00cf8401, 0x00000000 },
+ { 0x00600601, 0x44d02288, 0x00cf8441, 0x00000000 },
+ { 0x00600601, 0x44c12288, 0x00cf8001, 0x00000000 },
+ { 0x00600401, 0x44d12288, 0x00cf8041, 0x00000000 },
+ { 0x00400040, 0x22001a40, 0x166902c0, 0x04000400 },
+ { 0x00800201, 0x23e02288, 0x00d28a01, 0x00000000 },
+ { 0x00800401, 0x23f02288, 0x00d28a21, 0x00000000 },
+ { 0x00800201, 0x24002288, 0x00d28a41, 0x00000000 },
+ { 0x00800401, 0x24102288, 0x00d28a61, 0x00000000 },
+ { 0x00600201, 0x44e02288, 0x00cf8c01, 0x00000000 },
+ { 0x00600601, 0x44f02288, 0x00cf8c41, 0x00000000 },
+ { 0x00600601, 0x44e12288, 0x00cf8801, 0x00000000 },
+ { 0x00600401, 0x44f12288, 0x00cf8841, 0x00000000 },
+ { 0x00400040, 0x22081a40, 0x166902c0, 0x06000600 },
+ { 0x00800201, 0x24202288, 0x00d28201, 0x00000000 },
+ { 0x00800401, 0x24302288, 0x00d28221, 0x00000000 },
+ { 0x00800201, 0x24402288, 0x00d28241, 0x00000000 },
+ { 0x00800401, 0x24502288, 0x00d28261, 0x00000000 },
+ { 0x00600201, 0x45002288, 0x00cf8401, 0x00000000 },
+ { 0x00600601, 0x45102288, 0x00cf8441, 0x00000000 },
+ { 0x00600601, 0x45012288, 0x00cf8001, 0x00000000 },
+ { 0x00600401, 0x45112288, 0x00cf8041, 0x00000000 },
+ { 0x00800201, 0x24602288, 0x00d28a01, 0x00000000 },
+ { 0x00800401, 0x24702288, 0x00d28a21, 0x00000000 },
+ { 0x00800201, 0x24802288, 0x00d28a41, 0x00000000 },
+ { 0x00800401, 0x24902288, 0x00d28a61, 0x00000000 },
+ { 0x00600201, 0x45202288, 0x00cf8c01, 0x00000000 },
+ { 0x00600601, 0x45302288, 0x00cf8c41, 0x00000000 },
+ { 0x00600601, 0x45212288, 0x00cf8801, 0x00000000 },
+ { 0x00600401, 0x45312288, 0x00cf8841, 0x00000000 },
+ { 0x0c000031, 0x20002220, 0x06000380, 0x120a8018 },
+ { 0x0c000031, 0x20002220, 0x060004a0, 0x0a0a8019 },
+ { 0x00600001, 0x2fe00208, 0x008d0000, 0x00000000 },
+ { 0x07000031, 0x20002220, 0x0e000fe0, 0x82000010 },