/*---------------------------------------------------------------------
  $Id: gztxtri.s,v 1.1.1.1 2002/05/02 03:29:12 blythe Exp $
  
  File : gztxtri.s
  
  Coded     by Yoshitaka Yasumoto.   Jun 23, 1997.
  Copyright by Nintendo, Co., Ltd.           1997.
  ---------------------------------------------------------------------*/

 #!Reserved  (0,1,11,12,13,14,15)
 #!Reserved  (16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31)
 #!Reserved  (v0,v1,v29,v30,v31)

	#------ Smooth Shaded Textured Triangle/Quadrangle ------
	#
	#		+-----------+-----------+-----------+-----------+
	#		|    Hdr    | RDP cmd 1 | RDP cmd 2 | RDP cmd 3 |
	#		+-----+-----+--+--+--+--+-----+-----+-----+-----+
	#	Vtx 0	|  X  |  Y  |R |G |B |A |  S  |  T  |invWi|invWf|
	#		+-----+-----+--+--+--+--+-----+-----+-----+-----+
	#	Vtx 1	|  X  |  Y  |R |G |B |A |  S  |  T  |invWi|invWf|
	#		+-----+-----+--+--+--+--+-----+-----+-----+-----+
	#	Vtx 2	|  X  |  Y  |R |G |B |A |  S  |  T  |invWi|invWf|
	#		+-----+-----+--+--+--+--+-----+-----+-----+-----+
	#
	#		+-----+-----+--+--+--+--+-----+-----+-----+-----+
	#	Vtx 3	|  X  |  Y  |R |G |B |A |  S  |  T  |invWi|invWf|
	#		+-----+-----+--+--+--+--+-----+-----+-----+-----+
	#
  AssignForDMAproc
  Assign(objpSave, 2)	
ZObj_TxQuad:
	# objp の保存 & クリア
	lw	objpSave, oRZ_HEADER(zBuf)
	sw	zero,     oRZ_HEADER(zBuf)	
	.symbol	adrs_ZObj_TxQuad, ZObj_TxQuad+0x8000

ZObj_TxTri:
  FixedAssign(vtx0,   v2)
  FixedAssign(vtx1sv, v3)
  Assign(vtx2sv, v4)
  Assign(key0, v5)
  FixedAssign(invWi, v6)
  Assign(invWf, v7)
	#
	#  v2, v3 に zBuf+16, zBuf+32 からの値が代入されている
	#	
	# invWi データ/Y sort Key 作成		# Yield 判定
	vmov	invWi[7], vtx0[6]		andi	sys1, zStat0, SP_STATUS_YIELD
	vadd	key0, vzero, vtx0[1]		bne	sys1, zero, RSPyield_B
						# 頂点データのの取得
	/* 遅延 */				lqv	vtx2sv[0], oRZ_TXTRI_VTX2_X(zBuf)
ZObj_TxTri_0:	
  Assign(cmdsv1, 3)	
	# invW の最大値の取得(1)		# RDPCMD 1 の処理
	vsubc	vtmp,  vtx0,  vtx1sv[7]		lw	gfx1,  oRZ_RDPCMD_1(zBuf)	
	vge	invWi, invWi, vtx1sv[6]		lw	cmdsv1, RSP_STATUS_RDPCMD_1(zero)
  _forward_adrs(adrs_ZObj_TxTri_1, 9)
	vmrg	invWf, vtx0,  vtx1sv[7]		_li	(zStat0, adrs_ZObj_TxTri_1+0xf000) # 終了飛び先(負)
  Assign(keyAl, v8)
  Assign(keyAg, v9)
  Assign(vtxAl, v10)
  Assign(vtxAg, v11)
  Assign(keyBg, v12)
  Assign(vtxBg, v13)
  Assign(cmdsv2, 4)	
	# Y Sort (1)
	vlt	keyAl, key0, vtx1sv[1]		lw	cmdsv2, RSP_STATUS_RDPCMD_2(zero)
	vge	keyAg, key0, vtx1sv[1]		sw	gfx1,   RSP_STATUS_RDPCMD_1(zero) # RDPCMD 1 の保存
	vmrg	vtxAl, vtx1sv, vtx0		bne	gfx1, cmdsv1, case_G_RDPCHUNK_1	  # 新規 CMD なら出力
	vmrg	vtxAg, vtx0,   vtx1sv		/* 遅延 */
  EndAssign(cmdsv1, 3)	
ZObj_TxTri_1:
	# invW の最大値の取得(2)		# RDPCMD 2 の処理
	vsubc	vtmp,  invWf, vtx2sv[7]		lw	gfx1, oRZ_RDPCMD_2(zBuf)
  _forward_adrs(adrs_ZObj_TxTri_2, 9)
	vge	invWi, invWi, vtx2sv[6]		_li	(zStat0, adrs_ZObj_TxTri_2+0xf000) # 終了飛び先(負)
	vmrg	invWf, invWf, vtx2sv[7]		lw	objp, oRZ_HEADER(zBuf)		# 次の ZObject の取得
	# Y Sort (2)
	vge	keyBg, keyAl, vtx2sv[1]		sw	gfx1, RSP_STATUS_RDPCMD_2(zero)	# RDPCMD 2 の保存
	vmrg	vtxBg, vtxAl, vtx2sv		bne	gfx1, cmdsv2, case_G_RDPCHUNK_1	# 新規 CMD なら出力
	vmrg	vtx0,  vtx2sv,  vtxAl		/* 遅延 */
  EndAssign(cmdsv2, 4)	
  Assign(cmdsv3, 3)	
  Assign(cmdmask, 4)	
  Assign(vtx1, v14)
  Assign(vtx2, v15)
ZObj_TxTri_2:
	# nearW の値の取得 (1)			# RDPCMD 3 の処理		
	vrcph	vtmp[0], invWi[7]		lw	gfx1, oRZ_RDPCMD_3(zBuf)
	# Y Sort (3)		
	vge	keyBg, keyAg, keyBg		lw	cmdsv3, RSP_STATUS_RDPCMD_3(zero)
	vmrg	vtx1,  vtxBg, vtxAg			andi	zType,   objp, 7	# 次の zType の取得
	vmrg	vtx2,  vtxAg, vtxBg			sltiu	cmdmask, gfx1, 1	# gfx1 が 0 なら 1
  EndAssign(keyAl, v8)
  EndAssign(keyAg, v9)
  EndAssign(vtxAl, v10)
  EndAssign(vtxAg, v11)
  EndAssign(keyBg, v12)
  EndAssign(vtxBg, v13)
  Assign(atr0i, v8)
  Assign(atr1i, v9)
  Assign(atr2i, v10)
	# nearW の値の取得 (2)
	vrcpl	invWf[0], invWf[7]	sqv	vtx0[0],   oRZ_TXTRI_VTX0_X+ 0(zBuf)
	vrcph	invWi[0], _0x0000	luv	atr0i[12], oRZ_TXTRI_VTX0_X+ 0(zBuf)	# atr0i[0..3]=RGBA0
  Assign(vD10, v11)
  Assign(vD20, v12)
  Assign(vD02, v13)
  Assign(vD21, v16)
  Assign(vD01, v17)
	# 座標差分の計算
	vsub	vD10, vtx1, vtx0	sqv	vtx1[12],  oRZ_TXTRI_VTX1_X+ 0(zBuf)
	vsub	vD20, vtx2, vtx0	luv	atr1i[8],  oRZ_TXTRI_VTX1_X+ 0(zBuf)	# atr1i[0..3]=RGBA1
	vsub	vD02, vtx0, vtx2	sqv	vtx2[0],   oRZ_TXTRI_VTX2_X+ 0(zBuf)
	vsub	vD21, vtx2, vtx1	addi	sys0, zBuf, 4
	vsub	vD01, vtx0, vtx1	ldv	invWi[8],  oRZ_TXTRI_VTX0_X+ 8(sys0)	# invWi[4|6] =iWi0,1
  Assign(invri, v18)
  Assign(invrf, v19)
	# 外積値の計算
	vmudh	vtmp,  vD10,  vD20[0]	ldv	vtx2[6],   oRZ_TXTRI_VTX0_X+ 8(sys0)	# vtx2[4|6]  =iWf0,1
	vmadh	vtmp,  vD02,  vD10[0]	lsv	invWi[14], oRZ_TXTRI_VTX2_X+12(zBuf)	# invWi[7]   =iWi2
	vsar	invri, invri, invri[0]	luv	atr2i[12], oRZ_TXTRI_VTX2_X+ 0(zBuf)	# atr2i[0..3]=RGBA2
	vsar	invrf, invrf, invrf[1]	llv	atr0i[8],  oRZ_TXTRI_VTX0_X+ 8(zBuf)	# atr0i[4|5] =ST0
	# invW/W 値の計算		
	vmudl	vtmp,  vtx2,  invWf[0]	llv	atr1i[8],  oRZ_TXTRI_VTX1_X+12(zBuf)	# atr1i[4|5] =ST1
	vmadm	vtmp,  invWi, invWf[0]	llv	atr2i[8],  oRZ_TXTRI_VTX2_X+ 8(zBuf)	# atr2i[4|5] =ST2
	vmadn	invWf, vtx2,  invWi[0]	lsv	atr0i[12], oRSPZS_0x7fff(vecptr)	# atr0i[6]   =0x7fff
	vmadh	invWi, invWi, invWi[0]			addi	cmdmask, cmdmask, -1	# cmdsv3 のマスク値
  Assign(header, 5)
	# 外積値の逆数計算					
	vrcph	vtmp[0],  invri[1]	lsv	atr1i[12], oRSPZS_0x7fff(vecptr)	# atr1i[6]   =0x7fff
	vrcpl	invrf[0], invrf[1]			mfc2	header, invri[1]	# 外積値の符号取得
	vrcph	invri[0], _0x0000	lsv	atr2i[12], oRSPZS_0x7fff(vecptr)	# atr2i[6]   =0x7fff
  Assign(invDyi, v20)
  Assign(invDyf, v21)
	# Dy の逆数計算 (1)
	vrcp	invDyf[0], vD21[1]	ldv	invWf[0],  oRSPZS_INVWF(vecptr)
	vrcph	invDyi[0], _0x0000	ldv	invWi[0],  oRSPZS_INVWI(vecptr)		
	vrcp	invDyf[2], vD10[1]			and	cmdsv3, cmdsv3, cmdmask	# cmdsv3 のマスク
  _forward_adrs(adrs_ZObj_TxTri_3, 7)
	vrcph	invDyi[2], _0x0000			_li	(zStat0, adrs_ZObj_TxTri_3+0xf000) # 終了飛び先(負)
	vrcp	invDyf[3], vD20[1]			sw	gfx1, RSP_STATUS_RDPCMD_3(zero)	# RDPCMD 3 の保存
	vrcph	invDyi[3], _0x0000			bne	gfx1, cmdsv3, case_G_RDPCHUNK_1	# 新規 CMD なら出力
	vmudm	vtmp,   atr0i,  invWf[0h]		/* 遅延 */
  EndAssign(cmdsv3, 3)	
  EndAssign(cmdmask, 4)	
ZObj_TxTri_3:
  Assign(atr0f, v22)
  Assign(atr1f, v23)
  Assign(atr2f, v24)
	# atr?[0..3]=RGBA atr?[4|5|6]=STW
	# S,T,W に nearW/W を乗じる			# 新規 ZObject の DMA
	vmadn	atr0f,  vconst, _0x0000			_mov	(dmem_adrs, zBuf)
	vmadh	atr0i,  atr0i,  invWi[0h]		and	dram_adrs, objp, zMask
	vmudm	vtmp,   atr1i,  invWf[2h]		lbu	dma_len, RSP_ZOBJ_SIZE(zType)
	vmadn	atr1f,  vconst, _0x0000			_liu	(zStat1, ZObj_Launch_A)		# 終了Jump先(正)
 _forward_adrs(adrs_ZObj_TxTri_4, 6)
	vmadh	atr1i,  atr1i,  invWi[2h]		_liu	(return, adrs_ZObj_TxTri_4)
	vmudm	vtmp,   atr2i,  invWf[3h]		bne	dram_adrs, zero, DMAread	# 読み込み開始
	vmadn	atr2f,  vconst, _0x0000			/* 遅延 */
	/* 条件分岐時の処理 */				_liu	(zStat1, ZObj_NextLink)		# 終了Jump先(正)
ZObj_TxTri_4:					
	vmadh	atr2i,  atr2i,  invWi[3h]	ssv	vD10[0], oRSPZS_VSCRATCH+0(vecptr)	# X 差分をまとめる
  EndAssign(invWf, v7)
  EndAssign(invWi, v6)
	# 1/Dy の桁合わせ
	vmudl	invDyf, invDyf, _0x0008		ssv	vD20[0], oRSPZS_VSCRATCH+2(vecptr)
	vmadm	invDyi, invDyi, _0x0008		llv	vD21[4], oRSPZS_VSCRATCH+0(vecptr)
	vmadn	invDyf, vconst, _0x0000			ssv	vtx2[2],   2(outp)		# データ出力
  Assign(xmin, 3)
  Assign(xmid, 4)
	# 外積値の逆数 桁合わせ	
	vmudn	invrf, invrf, _0x0008				sll	zType, zType, 1
	vmadh	invri, invri, _0x0008			ssv	vtx1[2],   4(outp)
  Assign(DxDyi, v6)
  Assign(DxDyf, v7)
	# Dx の桁合わせ
	vmudm	DxDyi,  vD21,   _0x4000			mfc2	xmin, vtx0[0]
	vmadn	DxDyf,  vconst, _0x0000			mfc2	xmid, vtx1[0]
  EndAssign(vD21, v16)
  EndAssign(vtx1, v14)
  EndAssign(vtx2, v15)
  Assign(aD20f, v14)
  Assign(aD20i, v15)
  Assign(aD10f, v16)
  Assign(aD10i, v25)
	# Da の計算
	vsubc	aD10f, atr1f, atr0f			andi	header, header, 0x80
	vsub	aD10i, atr1i, atr0i			ori	header, header, 0x0e00
	vsubc	aD20f, atr2f, atr0f			ssv	vtx0[2],   6(outp)
	vsub	aD20i, atr2i, atr0i			sh	header,    0(outp)
  EndAssign(atr1i, v9)
  EndAssign(atr1f, v23)
  EndAssign(atr2i, v10)
  EndAssign(atr2f, v24)
	# Dx/Dy の計算						# 4 角形描画時のための vtx3 の読み込み
	vmudm	vtmp,  invDyi, DxDyf				lqv	vtx0[0], oRZ_TXTRI_VTX3_X(zBuf)
	vmadl	vtmp,  invDyf, DxDyf			sll	xmin, xmin, 14
	vmadn	DxDyf, invDyf, DxDyi			sll	xmid, xmid, 14
	vmadh	DxDyi, invDyi, DxDyi			sw	xmid,      8(outp)
  Assign(DaDef, v9)
  Assign(DaDei, v10)
	# Da/De の計算 [A20/y20]
	vmudl	vtmp,  aD20f, invDyf[3]			sw	xmin,     16(outp)
	vmadm	vtmp,  aD20i, invDyf[3]			sw	xmin,     24(outp)
	vmadn	DaDef, aD20f, invDyi[3]			ssv	DxDyf[0], 14(outp)
	vmadh	DaDei, aD20i, invDyi[3]			ssv	DxDyf[6], 22(outp)
  EndAssign(header, 5)
  EndAssign(xmid, 4)
  EndAssign(xmin, 3)
  EndAssign(invDyi, v20)
  EndAssign(invDyf, v21)
	# Da/Dx の計算 (1) [(A20*y10 - A10*y20)/r]
	vmudn	vtmp,  aD20f, vD10[1]		ssv	DxDyf[4], 30(outp)
	vmadh	vtmp,  aD20i, vD10[1]		ssv	DxDyi[0], 12(outp)
	vmadn	vtmp,  aD10f, vD02[1]		ssv	DxDyi[6], 20(outp)
	vmadh	vtmp,  aD10i, vD02[1]		ssv	DxDyi[4], 28(outp)
  EndAssign(vD10, v11)
  EndAssign(vD02, v13)
  EndAssign(DxDyi, v6)
  EndAssign(DxDyf, v7)
  Assign(DaDxi, v6)
  Assign(DaDxf, v7)
	vsar	DaDxf, vconst, vconst[1]	sdv	DaDef[0], 48+32(outp)
	vsar	DaDxi, vconst, vconst[0]	sdv	DaDef[8], 48+96(outp)
	# Da/Dy の計算 (2) [(A10*x20 - A20*x10)/r]
	vmudn	vtmp,  aD10f, vD20[0]		sdv	DaDei[0], 32+32(outp)
	vmadh	vtmp,  aD10i, vD20[0]		sdv	DaDei[8], 32+96(outp)
	vmadn	vtmp,  aD20f, vD01[0]		sdv	vzero[0], 40+32(outp)
	vmadh	vtmp,  aD20i, vD01[0]		sdv	vzero[0], 56+32(outp)
  EndAssign(aD10f, v16)
  EndAssign(aD10i, v25)
  EndAssign(aD20f, v14)
  EndAssign(aD20i, v15)
  EndAssign(DaDef, v9)
  EndAssign(DaDei, v10)
  EndAssign(vD20, v12)
  EndAssign(vD01, v17)
  Assign(DaDyi, v9)
  Assign(DaDyf, v10)
	vsar	DaDyf, vconst, vconst[1]	sdv	atr0i[0],  0+32(outp)
	vsar	DaDyi, vconst, vconst[0]	sdv	vzero[0], 16+32(outp)
	# Da/Dx の計算 (2)
	vmudl	vtmp,  DaDxf, invrf[0]		mfc0	zStat0, SP_STATUS
	vmadm	vtmp,  DaDxi, invrf[0]		sdv	atr0i[8],  0+96(outp)
	vmadn	DaDxf, DaDxf, invri[0]		_mov	(return, zStat1)
	vmadh	DaDxi, DaDxi, invri[0]		mfc0	zStat1, DMA_BUSY
	# Da/Dy の計算 (2)
	vmudl	vtmp,  DaDyf, invrf[0]		sdv	atr0f[8], 16+96(outp)
	vmadm	vtmp,  DaDyi, invrf[0]		addiu	outp, outp, 160
	vmadn	DaDyf, DaDyf, invri[0]		sdv	DaDxf[0], 24+32-160(outp)
	vmadh	DaDyi, DaDyi, invri[0]		sdv	DaDxi[0],  8+32-160(outp)	
  EndAssign(atr0i, v8)
  EndAssign(atr0f, v22)
  EndAssign(invri, v18)
  EndAssign(invrf, v19)
						sdv	DaDxf[8], 24+96-160(outp)
						sdv	DaDxi[8],  8+96-160(outp)	
						sdv	DaDyf[8], 56+96-160(outp)
	bgtz	zObjHdr, OutputClose		# 4 角形のときは Skip
						sdv	DaDyi[8], 40+96-160(outp)
  EndAssign(DaDxf, v7)
  EndAssign(DaDxi, v6)
  EndAssign(DaDyf, v10)
  EndAssign(DaDyi, v9)
  FixedAssign(invWi, v6)
	#------ 4 角形の処理 ------
	# invWi データ作成		# objp を元に戻す
	vmov	invWi[7], vtx0[6]	sw	objpSave, oRZ_HEADER(zBuf)
	# Y sort Key 作成		# 3 角形描画ルーチンへ Jump
	vadd	key0, vzero, vtx0[1]	j	ZObj_TxTri_0
					# 3 角形描画フラグ On
					_li	(zObjHdr, 1)	
  EndAssignForDMAproc
  EndAssign(key0, v5)
  EndAssign(invWi, v6)
  EndAssign(vtx0, v2)
  EndAssign(vtx1sv, v3)
  EndAssign(vtx2sv, v4)
  EndAssign(objpSave, 2)	
				
/*======== End of gztxtri.s ========*/