cannam@85: /***************************************************************************** cannam@85: * Copyright (C) 2000-2001 Andre McCurdy cannam@85: * cannam@85: * This program is free software. you can redistribute it and/or modify cannam@85: * it under the terms of the GNU General Public License as published by cannam@85: * the Free Software Foundation@ either version 2 of the License, or cannam@85: * (at your option) any later version. cannam@85: * cannam@85: * This program is distributed in the hope that it will be useful, cannam@85: * but WITHOUT ANY WARRANTY, without even the implied warranty of cannam@85: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the cannam@85: * GNU General Public License for more details. cannam@85: * cannam@85: * You should have received a copy of the GNU General Public License cannam@85: * along with this program@ if not, write to the Free Software cannam@85: * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA cannam@85: * cannam@85: ***************************************************************************** cannam@85: * cannam@85: * Notes: cannam@85: * cannam@85: * cannam@85: ***************************************************************************** cannam@85: * cannam@85: * $Id: imdct_l_arm.S,v 1.7 2001/03/25 20:03:34 rob Rel $ cannam@85: * cannam@85: * 2001/03/24: Andre McCurdy cannam@85: * - Corrected PIC unsafe loading of address of 'imdct36_long_karray' cannam@85: * cannam@85: * 2000/09/20: Robert Leslie cannam@85: * - Added a global symbol with leading underscore per suggestion of cannam@85: * Simon Burge to support linking with the a.out format. cannam@85: * cannam@85: * 2000/09/15: Robert Leslie cannam@85: * - Fixed a small bug where flags were changed before a conditional branch. cannam@85: * cannam@85: * 2000/09/15: Andre McCurdy cannam@85: * - Applied Nicolas Pitre's rounding optimisation in all remaining places. cannam@85: * cannam@85: * 2000/09/09: Nicolas Pitre cannam@85: * - Optimized rounding + scaling operations. cannam@85: * cannam@85: * 2000/08/09: Andre McCurdy cannam@85: * - Original created. cannam@85: * cannam@85: ****************************************************************************/ cannam@85: cannam@85: cannam@85: /* cannam@85: On entry: cannam@85: cannam@85: r0 = pointer to 18 element input array cannam@85: r1 = pointer to 36 element output array cannam@85: r2 = windowing block type cannam@85: cannam@85: cannam@85: Stack frame created during execution of the function: cannam@85: cannam@85: Initial Holds: cannam@85: Stack cannam@85: pointer cannam@85: minus: cannam@85: cannam@85: 0 cannam@85: 4 lr cannam@85: 8 r11 cannam@85: 12 r10 cannam@85: 16 r9 cannam@85: 20 r8 cannam@85: 24 r7 cannam@85: 28 r6 cannam@85: 32 r5 cannam@85: 36 r4 cannam@85: cannam@85: 40 r2 : windowing block type cannam@85: cannam@85: 44 ct00 high cannam@85: 48 ct00 low cannam@85: 52 ct01 high cannam@85: 56 ct01 low cannam@85: 60 ct04 high cannam@85: 64 ct04 low cannam@85: 68 ct06 high cannam@85: 72 ct06 low cannam@85: 76 ct05 high cannam@85: 80 ct05 low cannam@85: 84 ct03 high cannam@85: 88 ct03 low cannam@85: 92 -ct05 high cannam@85: 96 -ct05 low cannam@85: 100 -ct07 high cannam@85: 104 -ct07 low cannam@85: 108 ct07 high cannam@85: 112 ct07 low cannam@85: 116 ct02 high cannam@85: 120 ct02 low cannam@85: */ cannam@85: cannam@85: #define BLOCK_MODE_NORMAL 0 cannam@85: #define BLOCK_MODE_START 1 cannam@85: #define BLOCK_MODE_STOP 3 cannam@85: cannam@85: cannam@85: #define X0 0x00 cannam@85: #define X1 0x04 cannam@85: #define X2 0x08 cannam@85: #define X3 0x0C cannam@85: #define X4 0x10 cannam@85: #define X5 0x14 cannam@85: #define X6 0x18 cannam@85: #define X7 0x1c cannam@85: #define X8 0x20 cannam@85: #define X9 0x24 cannam@85: #define X10 0x28 cannam@85: #define X11 0x2c cannam@85: #define X12 0x30 cannam@85: #define X13 0x34 cannam@85: #define X14 0x38 cannam@85: #define X15 0x3c cannam@85: #define X16 0x40 cannam@85: #define X17 0x44 cannam@85: cannam@85: #define x0 0x00 cannam@85: #define x1 0x04 cannam@85: #define x2 0x08 cannam@85: #define x3 0x0C cannam@85: #define x4 0x10 cannam@85: #define x5 0x14 cannam@85: #define x6 0x18 cannam@85: #define x7 0x1c cannam@85: #define x8 0x20 cannam@85: #define x9 0x24 cannam@85: #define x10 0x28 cannam@85: #define x11 0x2c cannam@85: #define x12 0x30 cannam@85: #define x13 0x34 cannam@85: #define x14 0x38 cannam@85: #define x15 0x3c cannam@85: #define x16 0x40 cannam@85: #define x17 0x44 cannam@85: #define x18 0x48 cannam@85: #define x19 0x4c cannam@85: #define x20 0x50 cannam@85: #define x21 0x54 cannam@85: #define x22 0x58 cannam@85: #define x23 0x5c cannam@85: #define x24 0x60 cannam@85: #define x25 0x64 cannam@85: #define x26 0x68 cannam@85: #define x27 0x6c cannam@85: #define x28 0x70 cannam@85: #define x29 0x74 cannam@85: #define x30 0x78 cannam@85: #define x31 0x7c cannam@85: #define x32 0x80 cannam@85: #define x33 0x84 cannam@85: #define x34 0x88 cannam@85: #define x35 0x8c cannam@85: cannam@85: #define K00 0x0ffc19fd cannam@85: #define K01 0x00b2aa3e cannam@85: #define K02 0x0fdcf549 cannam@85: #define K03 0x0216a2a2 cannam@85: #define K04 0x0f9ee890 cannam@85: #define K05 0x03768962 cannam@85: #define K06 0x0f426cb5 cannam@85: #define K07 0x04cfb0e2 cannam@85: #define K08 0x0ec835e8 cannam@85: #define K09 0x061f78aa cannam@85: #define K10 0x0e313245 cannam@85: #define K11 0x07635284 cannam@85: #define K12 0x0d7e8807 cannam@85: #define K13 0x0898c779 cannam@85: #define K14 0x0cb19346 cannam@85: #define K15 0x09bd7ca0 cannam@85: #define K16 0x0bcbe352 cannam@85: #define K17 0x0acf37ad cannam@85: cannam@85: #define minus_K02 0xf0230ab7 cannam@85: cannam@85: #define WL0 0x00b2aa3e cannam@85: #define WL1 0x0216a2a2 cannam@85: #define WL2 0x03768962 cannam@85: #define WL3 0x04cfb0e2 cannam@85: #define WL4 0x061f78aa cannam@85: #define WL5 0x07635284 cannam@85: #define WL6 0x0898c779 cannam@85: #define WL7 0x09bd7ca0 cannam@85: #define WL8 0x0acf37ad cannam@85: #define WL9 0x0bcbe352 cannam@85: #define WL10 0x0cb19346 cannam@85: #define WL11 0x0d7e8807 cannam@85: #define WL12 0x0e313245 cannam@85: #define WL13 0x0ec835e8 cannam@85: #define WL14 0x0f426cb5 cannam@85: #define WL15 0x0f9ee890 cannam@85: #define WL16 0x0fdcf549 cannam@85: #define WL17 0x0ffc19fd cannam@85: cannam@85: cannam@85: @***************************************************************************** cannam@85: cannam@85: cannam@85: .text cannam@85: .align cannam@85: cannam@85: .global III_imdct_l cannam@85: .global _III_imdct_l cannam@85: cannam@85: III_imdct_l: cannam@85: _III_imdct_l: cannam@85: cannam@85: stmdb sp!, { r2, r4 - r11, lr } @ all callee saved regs, plus arg3 cannam@85: cannam@85: ldr r4, =K08 @ r4 = K08 cannam@85: ldr r5, =K09 @ r5 = K09 cannam@85: ldr r8, [r0, #X4] @ r8 = X4 cannam@85: ldr r9, [r0, #X13] @ r9 = X13 cannam@85: rsb r6, r4, #0 @ r6 = -K08 cannam@85: rsb r7, r5, #0 @ r7 = -K09 cannam@85: cannam@85: smull r2, r3, r4, r8 @ r2..r3 = (X4 * K08) cannam@85: smlal r2, r3, r5, r9 @ r2..r3 = (X4 * K08) + (X13 * K09) = ct01 cannam@85: cannam@85: smull r10, lr, r8, r5 @ r10..lr = (X4 * K09) cannam@85: smlal r10, lr, r9, r6 @ r10..lr = (X4 * K09) + (X13 * -K08) = ct00 cannam@85: cannam@85: ldr r8, [r0, #X7] @ r8 = X7 cannam@85: ldr r9, [r0, #X16] @ r9 = X16 cannam@85: cannam@85: stmdb sp!, { r2, r3, r10, lr } @ stack ct00_h, ct00_l, ct01_h, ct01_l cannam@85: cannam@85: add r8, r8, r9 @ r8 = (X7 + X16) cannam@85: ldr r9, [r0, #X1] @ r9 = X1 cannam@85: cannam@85: smlal r2, r3, r6, r8 @ r2..r3 = ct01 + ((X7 + X16) * -K08) cannam@85: smlal r2, r3, r7, r9 @ r2..r3 += (X1 * -K09) cannam@85: cannam@85: ldr r7, [r0, #X10] @ r7 = X10 cannam@85: cannam@85: rsbs r10, r10, #0 cannam@85: rsc lr, lr, #0 @ r10..lr = -ct00 cannam@85: cannam@85: smlal r2, r3, r5, r7 @ r2..r3 += (X10 * K09) = ct06 cannam@85: cannam@85: smlal r10, lr, r9, r6 @ r10..lr = -ct00 + ( X1 * -K08) cannam@85: smlal r10, lr, r8, r5 @ r10..lr += ((X7 + X16) * K09) cannam@85: smlal r10, lr, r7, r4 @ r10..lr += ( X10 * K08) = ct04 cannam@85: cannam@85: stmdb sp!, { r2, r3, r10, lr } @ stack ct04_h, ct04_l, ct06_h, ct06_l cannam@85: cannam@85: @---- cannam@85: cannam@85: ldr r7, [r0, #X0] cannam@85: ldr r8, [r0, #X11] cannam@85: ldr r9, [r0, #X12] cannam@85: sub r7, r7, r8 cannam@85: sub r7, r7, r9 @ r7 = (X0 - X11 -X12) = ct14 cannam@85: cannam@85: ldr r9, [r0, #X3] cannam@85: ldr r8, [r0, #X8] cannam@85: ldr r11, [r0, #X15] cannam@85: sub r8, r8, r9 cannam@85: add r8, r8, r11 @ r8 = (X8 - X3 + X15) = ct16 cannam@85: cannam@85: add r11, r7, r8 @ r11 = ct14 + ct16 = ct18 cannam@85: cannam@85: smlal r2, r3, r6, r11 @ r2..r3 = ct06 + ((X0 - X11 - X3 + X15 + X8 - X12) * -K08) cannam@85: cannam@85: ldr r6, [r0, #X2] cannam@85: ldr r9, [r0, #X9] cannam@85: ldr r12, [r0, #X14] cannam@85: sub r6, r6, r9 cannam@85: sub r6, r6, r12 @ r6 = (X2 - X9 - X14) = ct15 cannam@85: cannam@85: ldr r9, [r0, #X5] cannam@85: ldr r12, [r0, #X6] cannam@85: sub r9, r9, r12 cannam@85: ldr r12, [r0, #X17] cannam@85: sub r9, r9, r12 @ r9 = (X5 - X6 - X17) = ct17 cannam@85: cannam@85: add r12, r9, r6 @ r12 = ct15 + ct17 = ct19 cannam@85: cannam@85: smlal r2, r3, r5, r12 @ r2..r3 += ((X2 - X9 + X5 - X6 - X17 - X14) * K09) cannam@85: cannam@85: smlal r10, lr, r11, r5 @ r10..lr = ct04 + (ct18 * K09) cannam@85: smlal r10, lr, r12, r4 @ r10..lr = ct04 + (ct18 * K09) + (ct19 * K08) cannam@85: cannam@85: movs r2, r2, lsr #28 cannam@85: adc r2, r2, r3, lsl #4 @ r2 = bits[59..28] of r2..r3 cannam@85: str r2, [r1, #x22] @ store result x22 cannam@85: cannam@85: movs r10, r10, lsr #28 cannam@85: adc r10, r10, lr, lsl #4 @ r10 = bits[59..28] of r10..lr cannam@85: str r10, [r1, #x4] @ store result x4 cannam@85: cannam@85: @---- cannam@85: cannam@85: ldmia sp, { r2, r3, r4, r5 } @ r2..r3 = ct06, r4..r5 = ct04 (dont update sp) cannam@85: cannam@85: @ r2..r3 = ct06 cannam@85: @ r4..r5 = ct04 cannam@85: @ r6 = ct15 cannam@85: @ r7 = ct14 cannam@85: @ r8 = ct16 cannam@85: @ r9 = ct17 cannam@85: @ r10 = . cannam@85: @ r11 = . cannam@85: @ r12 = . cannam@85: @ lr = . cannam@85: cannam@85: ldr r10, =K03 @ r10 = K03 cannam@85: ldr lr, =K15 @ lr = K15 cannam@85: cannam@85: smlal r2, r3, r10, r7 @ r2..r3 = ct06 + (ct14 * K03) cannam@85: smlal r4, r5, lr, r7 @ r4..r5 = ct04 + (ct14 * K15) cannam@85: cannam@85: ldr r12, =K14 @ r12 = K14 cannam@85: rsb r10, r10, #0 @ r10 = -K03 cannam@85: cannam@85: smlal r2, r3, lr, r6 @ r2..r3 += (ct15 * K15) cannam@85: smlal r4, r5, r10, r6 @ r4..r5 += (ct15 * -K03) cannam@85: smlal r2, r3, r12, r8 @ r2..r3 += (ct16 * K14) cannam@85: cannam@85: ldr r11, =minus_K02 @ r11 = -K02 cannam@85: rsb r12, r12, #0 @ r12 = -K14 cannam@85: cannam@85: smlal r4, r5, r12, r9 @ r4..r5 += (ct17 * -K14) cannam@85: smlal r2, r3, r11, r9 @ r2..r3 += (ct17 * -K02) cannam@85: smlal r4, r5, r11, r8 @ r4..r5 += (ct16 * -K02) cannam@85: cannam@85: movs r2, r2, lsr #28 cannam@85: adc r2, r2, r3, lsl #4 @ r2 = bits[59..28] of r2..r3 cannam@85: str r2, [r1, #x7] @ store result x7 cannam@85: cannam@85: movs r4, r4, lsr #28 cannam@85: adc r4, r4, r5, lsl #4 @ r4 = bits[59..28] of r4..r5 cannam@85: str r4, [r1, #x1] @ store result x1 cannam@85: cannam@85: @---- cannam@85: cannam@85: ldmia sp, { r2, r3, r4, r5 } @ r2..r3 = ct06, r4..r5 = ct04 (dont update sp) cannam@85: cannam@85: @ r2..r3 = ct06 cannam@85: @ r4..r5 = ct04 cannam@85: @ r6 = ct15 cannam@85: @ r7 = ct14 cannam@85: @ r8 = ct16 cannam@85: @ r9 = ct17 cannam@85: @ r10 = -K03 cannam@85: @ r11 = -K02 cannam@85: @ r12 = -K14 cannam@85: @ lr = K15 cannam@85: cannam@85: rsbs r2, r2, #0 cannam@85: rsc r3, r3, #0 @ r2..r3 = -ct06 cannam@85: cannam@85: smlal r2, r3, r12, r7 @ r2..r3 = -ct06 + (ct14 * -K14) cannam@85: smlal r2, r3, r10, r8 @ r2..r3 += (ct16 * -K03) cannam@85: cannam@85: smlal r4, r5, r12, r6 @ r4..r5 = ct04 + (ct15 * -K14) cannam@85: smlal r4, r5, r10, r9 @ r4..r5 += (ct17 * -K03) cannam@85: smlal r4, r5, lr, r8 @ r4..r5 += (ct16 * K15) cannam@85: smlal r4, r5, r11, r7 @ r4..r5 += (ct14 * -K02) cannam@85: cannam@85: rsb lr, lr, #0 @ lr = -K15 cannam@85: rsb r11, r11, #0 @ r11 = K02 cannam@85: cannam@85: smlal r2, r3, lr, r9 @ r2..r3 += (ct17 * -K15) cannam@85: smlal r2, r3, r11, r6 @ r2..r3 += (ct15 * K02) cannam@85: cannam@85: movs r4, r4, lsr #28 cannam@85: adc r4, r4, r5, lsl #4 @ r4 = bits[59..28] of r4..r5 cannam@85: str r4, [r1, #x25] @ store result x25 cannam@85: cannam@85: movs r2, r2, lsr #28 cannam@85: adc r2, r2, r3, lsl #4 @ r2 = bits[59..28] of r2..r3 cannam@85: str r2, [r1, #x19] @ store result x19 cannam@85: cannam@85: @---- cannam@85: cannam@85: ldr r2, [sp, #16] @ r2 = ct01_l cannam@85: ldr r3, [sp, #20] @ r3 = ct01_h cannam@85: cannam@85: ldr r6, [r0, #X1] cannam@85: ldr r8, [r0, #X7] cannam@85: ldr r9, [r0, #X10] cannam@85: ldr r7, [r0, #X16] cannam@85: cannam@85: rsbs r2, r2, #0 cannam@85: rsc r3, r3, #0 @ r2..r3 = -ct01 cannam@85: cannam@85: mov r4, r2 cannam@85: mov r5, r3 @ r4..r5 = -ct01 cannam@85: cannam@85: @ r2..r3 = -ct01 cannam@85: @ r4..r5 = -ct01 cannam@85: @ r6 = X1 cannam@85: @ r7 = X16 cannam@85: @ r8 = X7 cannam@85: @ r9 = X10 cannam@85: @ r10 = -K03 cannam@85: @ r11 = K02 cannam@85: @ r12 = -K14 cannam@85: @ lr = -K15 cannam@85: cannam@85: smlal r4, r5, r12, r7 @ r4..r5 = -ct01 + (X16 * -K14) cannam@85: smlal r2, r3, lr, r9 @ r2..r3 = -ct01 + (X10 * -K15) cannam@85: cannam@85: smlal r4, r5, r10, r8 @ r4..r5 += (X7 * -K03) cannam@85: smlal r2, r3, r10, r7 @ r2..r3 += (X16 * -K03) cannam@85: cannam@85: smlal r4, r5, r11, r9 @ r4..r5 += (X10 * K02) cannam@85: smlal r2, r3, r12, r8 @ r2..r3 += (X7 * -K14) cannam@85: cannam@85: rsb lr, lr, #0 @ lr = K15 cannam@85: rsb r11, r11, #0 @ r11 = -K02 cannam@85: cannam@85: smlal r4, r5, lr, r6 @ r4..r5 += (X1 * K15) = ct05 cannam@85: smlal r2, r3, r11, r6 @ r2..r3 += (X1 * -K02) = ct03 cannam@85: cannam@85: stmdb sp!, { r2, r3, r4, r5 } @ stack ct05_h, ct05_l, ct03_h, ct03_l cannam@85: cannam@85: rsbs r4, r4, #0 cannam@85: rsc r5, r5, #0 @ r4..r5 = -ct05 cannam@85: cannam@85: stmdb sp!, { r4, r5 } @ stack -ct05_h, -ct05_l cannam@85: cannam@85: ldr r2, [sp, #48] @ r2 = ct00_l cannam@85: ldr r3, [sp, #52] @ r3 = ct00_h cannam@85: cannam@85: rsb r10, r10, #0 @ r10 = K03 cannam@85: cannam@85: rsbs r4, r2, #0 cannam@85: rsc r5, r3, #0 @ r4..r5 = -ct00 cannam@85: cannam@85: @ r2..r3 = ct00 cannam@85: @ r4..r5 = -ct00 cannam@85: @ r6 = X1 cannam@85: @ r7 = X16 cannam@85: @ r8 = X7 cannam@85: @ r9 = X10 cannam@85: @ r10 = K03 cannam@85: @ r11 = -K02 cannam@85: @ r12 = -K14 cannam@85: @ lr = K15 cannam@85: cannam@85: smlal r4, r5, r10, r6 @ r4..r5 = -ct00 + (X1 * K03) cannam@85: smlal r2, r3, r10, r9 @ r2..r3 = ct00 + (X10 * K03) cannam@85: cannam@85: smlal r4, r5, r12, r9 @ r4..r5 += (X10 * -K14) cannam@85: smlal r2, r3, r12, r6 @ r2..r3 += (X1 * -K14) cannam@85: cannam@85: smlal r4, r5, r11, r7 @ r4..r5 += (X16 * -K02) cannam@85: smlal r4, r5, lr, r8 @ r4..r5 += (X7 * K15) = ct07 cannam@85: cannam@85: rsb lr, lr, #0 @ lr = -K15 cannam@85: rsb r11, r11, #0 @ r11 = K02 cannam@85: cannam@85: smlal r2, r3, r11, r8 @ r2..r3 += (X7 * K02) cannam@85: smlal r2, r3, lr, r7 @ r2..r3 += (X16 * -K15) = ct02 cannam@85: cannam@85: rsbs r6, r4, #0 cannam@85: rsc r7, r5, #0 @ r6..r7 = -ct07 cannam@85: cannam@85: stmdb sp!, { r2 - r7 } @ stack -ct07_h, -ct07_l, ct07_h, ct07_l, ct02_h, ct02_l cannam@85: cannam@85: cannam@85: @---- cannam@85: cannam@85: add r2, pc, #(imdct36_long_karray-.-8) @ r2 = base address of Knn array (PIC safe ?) cannam@85: cannam@85: cannam@85: loop: cannam@85: ldr r12, [r0, #X0] cannam@85: cannam@85: ldmia r2!, { r5 - r11 } @ first 7 words from Karray element cannam@85: cannam@85: smull r3, r4, r5, r12 @ sum = (Kxx * X0) cannam@85: ldr r12, [r0, #X2] cannam@85: ldr r5, [r0, #X3] cannam@85: smlal r3, r4, r6, r12 @ sum += (Kxx * X2) cannam@85: ldr r12, [r0, #X5] cannam@85: ldr r6, [r0, #X6] cannam@85: smlal r3, r4, r7, r5 @ sum += (Kxx * X3) cannam@85: smlal r3, r4, r8, r12 @ sum += (Kxx * X5) cannam@85: ldr r12, [r0, #X8] cannam@85: ldr r5, [r0, #X9] cannam@85: smlal r3, r4, r9, r6 @ sum += (Kxx * X6) cannam@85: smlal r3, r4, r10, r12 @ sum += (Kxx * X8) cannam@85: smlal r3, r4, r11, r5 @ sum += (Kxx * X9) cannam@85: cannam@85: ldmia r2!, { r5 - r10 } @ final 6 words from Karray element cannam@85: cannam@85: ldr r11, [r0, #X11] cannam@85: ldr r12, [r0, #X12] cannam@85: smlal r3, r4, r5, r11 @ sum += (Kxx * X11) cannam@85: ldr r11, [r0, #X14] cannam@85: ldr r5, [r0, #X15] cannam@85: smlal r3, r4, r6, r12 @ sum += (Kxx * X12) cannam@85: smlal r3, r4, r7, r11 @ sum += (Kxx * X14) cannam@85: ldr r11, [r0, #X17] cannam@85: smlal r3, r4, r8, r5 @ sum += (Kxx * X15) cannam@85: smlal r3, r4, r9, r11 @ sum += (Kxx * X17) cannam@85: cannam@85: add r5, sp, r10, lsr #16 @ create index back into stack for required ctxx cannam@85: cannam@85: ldmia r5, { r6, r7 } @ r6..r7 = ctxx cannam@85: cannam@85: mov r8, r10, lsl #16 @ push ctxx index off the top end cannam@85: cannam@85: adds r3, r3, r6 @ add low words cannam@85: adc r4, r4, r7 @ add high words, with carry cannam@85: movs r3, r3, lsr #28 cannam@85: adc r3, r3, r4, lsl #4 @ r3 = bits[59..28] of r3..r4 cannam@85: cannam@85: str r3, [r1, r8, lsr #24] @ push completion flag off the bottom end cannam@85: cannam@85: movs r8, r8, lsl #8 @ push result location index off the top end cannam@85: beq loop @ loop back if completion flag not set cannam@85: b imdct_l_windowing @ branch to windowing stage if looping finished cannam@85: cannam@85: imdct36_long_karray: cannam@85: cannam@85: .word K17, -K13, K10, -K06, -K05, K01, -K00, K04, -K07, K11, K12, -K16, 0x00000000 cannam@85: .word K13, K07, K16, K01, K10, -K05, K04, -K11, K00, -K17, K06, -K12, 0x00200800 cannam@85: .word K11, K17, K05, K12, -K01, K06, -K07, K00, -K13, K04, -K16, K10, 0x00200c00 cannam@85: .word K07, K00, -K12, K05, -K16, -K10, K11, -K17, K04, K13, K01, K06, 0x00001400 cannam@85: .word K05, K10, -K00, -K17, K07, -K13, K12, K06, -K16, K01, -K11, -K04, 0x00181800 cannam@85: .word K01, K05, -K07, -K11, K13, K17, -K16, -K12, K10, K06, -K04, -K00, 0x00102000 cannam@85: .word -K16, K12, -K11, K07, K04, -K00, -K01, K05, -K06, K10, K13, -K17, 0x00284800 cannam@85: .word -K12, K06, K17, -K00, -K11, K04, K05, -K10, K01, K16, -K07, -K13, 0x00085000 cannam@85: .word -K10, K16, K04, -K13, -K00, K07, K06, -K01, -K12, -K05, K17, K11, 0x00105400 cannam@85: .word -K06, -K01, K13, K04, K17, -K11, -K10, -K16, -K05, K12, K00, K07, 0x00185c00 cannam@85: .word -K04, -K11, -K01, K16, K06, K12, K13, -K07, -K17, -K00, -K10, -K05, 0x00006000 cannam@85: .word -K00, -K04, -K06, -K10, -K12, -K16, -K17, -K13, -K11, -K07, -K05, -K01, 0x00206801 cannam@85: cannam@85: cannam@85: @---- cannam@85: @------------------------------------------------------------------------- cannam@85: @---- cannam@85: cannam@85: imdct_l_windowing: cannam@85: cannam@85: ldr r11, [sp, #80] @ fetch function parameter 3 from out of the stack cannam@85: ldmia r1!, { r0, r2 - r9 } @ load 9 words from x0, update pointer cannam@85: cannam@85: @ r0 = x0 cannam@85: @ r1 = &x[9] cannam@85: @ r2 = x1 cannam@85: @ r3 = x2 cannam@85: @ r4 = x3 cannam@85: @ r5 = x4 cannam@85: @ r6 = x5 cannam@85: @ r7 = x6 cannam@85: @ r8 = x7 cannam@85: @ r9 = x8 cannam@85: @ r10 = . cannam@85: @ r11 = window mode: (0 == normal), (1 == start block), (3 == stop block) cannam@85: @ r12 = . cannam@85: @ lr = . cannam@85: cannam@85: cmp r11, #BLOCK_MODE_STOP @ setup flags cannam@85: rsb r10, r0, #0 @ r10 = -x0 (DONT change flags !!) cannam@85: beq stop_block_x0_to_x17 cannam@85: cannam@85: cannam@85: @ start and normal blocks are treated the same for x[0]..x[17] cannam@85: cannam@85: normal_block_x0_to_x17: cannam@85: cannam@85: ldr r12, =WL9 @ r12 = window_l[9] cannam@85: cannam@85: rsb r0, r9, #0 @ r0 = -x8 cannam@85: rsb r9, r2, #0 @ r9 = -x1 cannam@85: rsb r2, r8, #0 @ r2 = -x7 cannam@85: rsb r8, r3, #0 @ r8 = -x2 cannam@85: rsb r3, r7, #0 @ r3 = -x6 cannam@85: rsb r7, r4, #0 @ r7 = -x3 cannam@85: rsb r4, r6, #0 @ r4 = -x5 cannam@85: rsb r6, r5, #0 @ r6 = -x4 cannam@85: cannam@85: @ r0 = -x8 cannam@85: @ r1 = &x[9] cannam@85: @ r2 = -x7 cannam@85: @ r3 = -x6 cannam@85: @ r4 = -x5 cannam@85: @ r5 = . cannam@85: @ r6 = -x4 cannam@85: @ r7 = -x3 cannam@85: @ r8 = -x2 cannam@85: @ r9 = -x1 cannam@85: @ r10 = -x0 cannam@85: @ r11 = window mode: (0 == normal), (1 == start block), (3 == stop block) cannam@85: @ r12 = window_l[9] cannam@85: @ lr = . cannam@85: cannam@85: smull r5, lr, r12, r0 @ r5..lr = (window_l[9] * (x[9] == -x[8])) cannam@85: ldr r12, =WL10 @ r12 = window_l[10] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r0, r5, lr, lsl #4 @ r0 = bits[59..28] of windowed x9 cannam@85: cannam@85: smull r5, lr, r12, r2 @ r5..lr = (window_l[10] * (x[10] == -x[7])) cannam@85: ldr r12, =WL11 @ r12 = window_l[11] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r2, r5, lr, lsl #4 @ r2 = bits[59..28] of windowed x10 cannam@85: cannam@85: smull r5, lr, r12, r3 @ r5..lr = (window_l[11] * (x[11] == -x[6])) cannam@85: ldr r12, =WL12 @ r12 = window_l[12] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r3, r5, lr, lsl #4 @ r3 = bits[59..28] of windowed x11 cannam@85: cannam@85: smull r5, lr, r12, r4 @ r5..lr = (window_l[12] * (x[12] == -x[5])) cannam@85: ldr r12, =WL13 @ r12 = window_l[13] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r4, r5, lr, lsl #4 @ r4 = bits[59..28] of windowed x12 cannam@85: cannam@85: smull r5, lr, r12, r6 @ r5..lr = (window_l[13] * (x[13] == -x[4])) cannam@85: ldr r12, =WL14 @ r12 = window_l[14] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r6, r5, lr, lsl #4 @ r6 = bits[59..28] of windowed x13 cannam@85: cannam@85: smull r5, lr, r12, r7 @ r5..lr = (window_l[14] * (x[14] == -x[3])) cannam@85: ldr r12, =WL15 @ r12 = window_l[15] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r7, r5, lr, lsl #4 @ r7 = bits[59..28] of windowed x14 cannam@85: cannam@85: smull r5, lr, r12, r8 @ r5..lr = (window_l[15] * (x[15] == -x[2])) cannam@85: ldr r12, =WL16 @ r12 = window_l[16] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r8, r5, lr, lsl #4 @ r8 = bits[59..28] of windowed x15 cannam@85: cannam@85: smull r5, lr, r12, r9 @ r5..lr = (window_l[16] * (x[16] == -x[1])) cannam@85: ldr r12, =WL17 @ r12 = window_l[17] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r9, r5, lr, lsl #4 @ r9 = bits[59..28] of windowed x16 cannam@85: cannam@85: smull r5, lr, r12, r10 @ r5..lr = (window_l[17] * (x[17] == -x[0])) cannam@85: ldr r12, =WL0 @ r12 = window_l[0] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r10, r5, lr, lsl #4 @ r10 = bits[59..28] of windowed x17 cannam@85: cannam@85: cannam@85: stmia r1, { r0, r2 - r4, r6 - r10 } @ store windowed x[9] .. x[17] cannam@85: ldmdb r1!, { r0, r2 - r9 } @ load 9 words downto (and including) x0 cannam@85: cannam@85: cannam@85: smull r10, lr, r12, r0 @ r10..lr = (window_l[0] * x[0]) cannam@85: ldr r12, =WL1 @ r12 = window_l[1] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r0, r10, lr, lsl #4 @ r0 = bits[59..28] of windowed x0 cannam@85: cannam@85: smull r10, lr, r12, r2 @ r10..lr = (window_l[1] * x[1]) cannam@85: ldr r12, =WL2 @ r12 = window_l[2] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r2, r10, lr, lsl #4 @ r2 = bits[59..28] of windowed x1 cannam@85: cannam@85: smull r10, lr, r12, r3 @ r10..lr = (window_l[2] * x[2]) cannam@85: ldr r12, =WL3 @ r12 = window_l[3] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r3, r10, lr, lsl #4 @ r3 = bits[59..28] of windowed x2 cannam@85: cannam@85: smull r10, lr, r12, r4 @ r10..lr = (window_l[3] * x[3]) cannam@85: ldr r12, =WL4 @ r12 = window_l[4] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r4, r10, lr, lsl #4 @ r4 = bits[59..28] of windowed x3 cannam@85: cannam@85: smull r10, lr, r12, r5 @ r10..lr = (window_l[4] * x[4]) cannam@85: ldr r12, =WL5 @ r12 = window_l[5] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r5, r10, lr, lsl #4 @ r5 = bits[59..28] of windowed x4 cannam@85: cannam@85: smull r10, lr, r12, r6 @ r10..lr = (window_l[5] * x[5]) cannam@85: ldr r12, =WL6 @ r12 = window_l[6] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r6, r10, lr, lsl #4 @ r6 = bits[59..28] of windowed x5 cannam@85: cannam@85: smull r10, lr, r12, r7 @ r10..lr = (window_l[6] * x[6]) cannam@85: ldr r12, =WL7 @ r12 = window_l[7] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r7, r10, lr, lsl #4 @ r7 = bits[59..28] of windowed x6 cannam@85: cannam@85: smull r10, lr, r12, r8 @ r10..lr = (window_l[7] * x[7]) cannam@85: ldr r12, =WL8 @ r12 = window_l[8] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r8, r10, lr, lsl #4 @ r8 = bits[59..28] of windowed x7 cannam@85: cannam@85: smull r10, lr, r12, r9 @ r10..lr = (window_l[8] * x[8]) cannam@85: movs r10, r10, lsr #28 cannam@85: adc r9, r10, lr, lsl #4 @ r9 = bits[59..28] of windowed x8 cannam@85: cannam@85: stmia r1, { r0, r2 - r9 } @ store windowed x[0] .. x[8] cannam@85: cannam@85: cmp r11, #BLOCK_MODE_START cannam@85: beq start_block_x18_to_x35 cannam@85: cannam@85: cannam@85: @---- cannam@85: cannam@85: cannam@85: normal_block_x18_to_x35: cannam@85: cannam@85: ldr r11, =WL3 @ r11 = window_l[3] cannam@85: ldr r12, =WL4 @ r12 = window_l[4] cannam@85: cannam@85: add r1, r1, #(18*4) @ r1 = &x[18] cannam@85: cannam@85: ldmia r1!, { r0, r2 - r4, r6 - r10 } @ load 9 words from x18, update pointer cannam@85: cannam@85: @ r0 = x18 cannam@85: @ r1 = &x[27] cannam@85: @ r2 = x19 cannam@85: @ r3 = x20 cannam@85: @ r4 = x21 cannam@85: @ r5 = . cannam@85: @ r6 = x22 cannam@85: @ r7 = x23 cannam@85: @ r8 = x24 cannam@85: @ r9 = x25 cannam@85: @ r10 = x26 cannam@85: @ r11 = window_l[3] cannam@85: @ r12 = window_l[4] cannam@85: @ lr = . cannam@85: cannam@85: smull r5, lr, r12, r6 @ r5..lr = (window_l[4] * (x[22] == x[31])) cannam@85: movs r5, r5, lsr #28 cannam@85: adc r5, r5, lr, lsl #4 @ r5 = bits[59..28] of windowed x31 cannam@85: cannam@85: smull r6, lr, r11, r4 @ r5..lr = (window_l[3] * (x[21] == x[32])) cannam@85: ldr r12, =WL5 @ r12 = window_l[5] cannam@85: movs r6, r6, lsr #28 cannam@85: adc r6, r6, lr, lsl #4 @ r6 = bits[59..28] of windowed x32 cannam@85: cannam@85: smull r4, lr, r12, r7 @ r4..lr = (window_l[5] * (x[23] == x[30])) cannam@85: ldr r11, =WL1 @ r11 = window_l[1] cannam@85: ldr r12, =WL2 @ r12 = window_l[2] cannam@85: movs r4, r4, lsr #28 cannam@85: adc r4, r4, lr, lsl #4 @ r4 = bits[59..28] of windowed x30 cannam@85: cannam@85: smull r7, lr, r12, r3 @ r7..lr = (window_l[2] * (x[20] == x[33])) cannam@85: ldr r12, =WL6 @ r12 = window_l[6] cannam@85: movs r7, r7, lsr #28 cannam@85: adc r7, r7, lr, lsl #4 @ r7 = bits[59..28] of windowed x33 cannam@85: cannam@85: smull r3, lr, r12, r8 @ r3..lr = (window_l[6] * (x[24] == x[29])) cannam@85: movs r3, r3, lsr #28 cannam@85: adc r3, r3, lr, lsl #4 @ r3 = bits[59..28] of windowed x29 cannam@85: cannam@85: smull r8, lr, r11, r2 @ r7..lr = (window_l[1] * (x[19] == x[34])) cannam@85: ldr r12, =WL7 @ r12 = window_l[7] cannam@85: ldr r11, =WL8 @ r11 = window_l[8] cannam@85: movs r8, r8, lsr #28 cannam@85: adc r8, r8, lr, lsl #4 @ r8 = bits[59..28] of windowed x34 cannam@85: cannam@85: smull r2, lr, r12, r9 @ r7..lr = (window_l[7] * (x[25] == x[28])) cannam@85: ldr r12, =WL0 @ r12 = window_l[0] cannam@85: movs r2, r2, lsr #28 cannam@85: adc r2, r2, lr, lsl #4 @ r2 = bits[59..28] of windowed x28 cannam@85: cannam@85: smull r9, lr, r12, r0 @ r3..lr = (window_l[0] * (x[18] == x[35])) cannam@85: movs r9, r9, lsr #28 cannam@85: adc r9, r9, lr, lsl #4 @ r9 = bits[59..28] of windowed x35 cannam@85: cannam@85: smull r0, lr, r11, r10 @ r7..lr = (window_l[8] * (x[26] == x[27])) cannam@85: ldr r11, =WL16 @ r11 = window_l[16] cannam@85: ldr r12, =WL17 @ r12 = window_l[17] cannam@85: movs r0, r0, lsr #28 cannam@85: adc r0, r0, lr, lsl #4 @ r0 = bits[59..28] of windowed x27 cannam@85: cannam@85: cannam@85: stmia r1, { r0, r2 - r9 } @ store windowed x[27] .. x[35] cannam@85: ldmdb r1!, { r0, r2 - r9 } @ load 9 words downto (and including) x18 cannam@85: cannam@85: cannam@85: smull r10, lr, r12, r0 @ r10..lr = (window_l[17] * x[18]) cannam@85: movs r10, r10, lsr #28 cannam@85: adc r0, r10, lr, lsl #4 @ r0 = bits[59..28] of windowed x0 cannam@85: cannam@85: smull r10, lr, r11, r2 @ r10..lr = (window_l[16] * x[19]) cannam@85: ldr r11, =WL14 @ r11 = window_l[14] cannam@85: ldr r12, =WL15 @ r12 = window_l[15] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r2, r10, lr, lsl #4 @ r2 = bits[59..28] of windowed x1 cannam@85: cannam@85: smull r10, lr, r12, r3 @ r10..lr = (window_l[15] * x[20]) cannam@85: movs r10, r10, lsr #28 cannam@85: adc r3, r10, lr, lsl #4 @ r3 = bits[59..28] of windowed x2 cannam@85: cannam@85: smull r10, lr, r11, r4 @ r10..lr = (window_l[14] * x[21]) cannam@85: ldr r11, =WL12 @ r11 = window_l[12] cannam@85: ldr r12, =WL13 @ r12 = window_l[13] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r4, r10, lr, lsl #4 @ r4 = bits[59..28] of windowed x3 cannam@85: cannam@85: smull r10, lr, r12, r5 @ r10..lr = (window_l[13] * x[22]) cannam@85: movs r10, r10, lsr #28 cannam@85: adc r5, r10, lr, lsl #4 @ r5 = bits[59..28] of windowed x4 cannam@85: cannam@85: smull r10, lr, r11, r6 @ r10..lr = (window_l[12] * x[23]) cannam@85: ldr r11, =WL10 @ r12 = window_l[10] cannam@85: ldr r12, =WL11 @ r12 = window_l[11] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r6, r10, lr, lsl #4 @ r6 = bits[59..28] of windowed x5 cannam@85: cannam@85: smull r10, lr, r12, r7 @ r10..lr = (window_l[11] * x[24]) cannam@85: movs r10, r10, lsr #28 cannam@85: adc r7, r10, lr, lsl #4 @ r7 = bits[59..28] of windowed x6 cannam@85: cannam@85: smull r10, lr, r11, r8 @ r10..lr = (window_l[10] * x[25]) cannam@85: ldr r12, =WL9 @ r12 = window_l[9] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r8, r10, lr, lsl #4 @ r8 = bits[59..28] of windowed x7 cannam@85: cannam@85: smull r10, lr, r12, r9 @ r10..lr = (window_l[9] * x[26]) cannam@85: cannam@85: movs r10, r10, lsr #28 cannam@85: adc r9, r10, lr, lsl #4 @ r9 = bits[59..28] of windowed x8 cannam@85: cannam@85: stmia r1, { r0, r2 - r9 } @ store windowed x[18] .. x[26] cannam@85: cannam@85: @---- cannam@85: @ NB there are 2 possible exits from this function - this is only one of them cannam@85: @---- cannam@85: cannam@85: add sp, sp, #(21*4) @ return stack frame cannam@85: ldmia sp!, { r4 - r11, pc } @ restore callee saved regs, and return cannam@85: cannam@85: @---- cannam@85: cannam@85: cannam@85: stop_block_x0_to_x17: cannam@85: cannam@85: @ r0 = x0 cannam@85: @ r1 = &x[9] cannam@85: @ r2 = x1 cannam@85: @ r3 = x2 cannam@85: @ r4 = x3 cannam@85: @ r5 = x4 cannam@85: @ r6 = x5 cannam@85: @ r7 = x6 cannam@85: @ r8 = x7 cannam@85: @ r9 = x8 cannam@85: @ r10 = -x0 cannam@85: @ r11 = window mode: (0 == normal), (1 == start block), (3 == stop block) cannam@85: @ r12 = . cannam@85: @ lr = . cannam@85: cannam@85: rsb r0, r6, #0 @ r0 = -x5 cannam@85: rsb r6, r2, #0 @ r6 = -x1 cannam@85: rsb r2, r5, #0 @ r2 = -x4 cannam@85: rsb r5, r3, #0 @ r5 = -x2 cannam@85: rsb r3, r4, #0 @ r3 = -x3 cannam@85: cannam@85: add r1, r1, #(3*4) @ r1 = &x[12] cannam@85: stmia r1, { r0, r2, r3, r5, r6, r10 } @ store unchanged x[12] .. x[17] cannam@85: cannam@85: ldr r0, =WL1 @ r0 = window_l[1] == window_s[0] cannam@85: cannam@85: rsb r10, r9, #0 @ r10 = -x8 cannam@85: rsb r12, r8, #0 @ r12 = -x7 cannam@85: rsb lr, r7, #0 @ lr = -x6 cannam@85: cannam@85: @ r0 = WL1 cannam@85: @ r1 = &x[12] cannam@85: @ r2 = . cannam@85: @ r3 = . cannam@85: @ r4 = . cannam@85: @ r5 = . cannam@85: @ r6 = . cannam@85: @ r7 = x6 cannam@85: @ r8 = x7 cannam@85: @ r9 = x8 cannam@85: @ r10 = -x8 cannam@85: @ r11 = window mode: (0 == normal), (1 == start block), (3 == stop block) cannam@85: @ r12 = -x7 cannam@85: @ lr = -x6 cannam@85: cannam@85: smull r5, r6, r0, r7 @ r5..r6 = (window_l[1] * x[6]) cannam@85: ldr r2, =WL4 @ r2 = window_l[4] == window_s[1] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r7, r5, r6, lsl #4 @ r7 = bits[59..28] of windowed x6 cannam@85: cannam@85: smull r5, r6, r2, r8 @ r5..r6 = (window_l[4] * x[7]) cannam@85: ldr r3, =WL7 @ r3 = window_l[7] == window_s[2] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r8, r5, r6, lsl #4 @ r8 = bits[59..28] of windowed x7 cannam@85: cannam@85: smull r5, r6, r3, r9 @ r5..r6 = (window_l[7] * x[8]) cannam@85: ldr r4, =WL10 @ r4 = window_l[10] == window_s[3] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r9, r5, r6, lsl #4 @ r9 = bits[59..28] of windowed x8 cannam@85: cannam@85: smull r5, r6, r4, r10 @ r5..r6 = (window_l[10] * (x[9] == -x[8])) cannam@85: ldr r0, =WL13 @ r0 = window_l[13] == window_s[4] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r10, r5, r6, lsl #4 @ r10 = bits[59..28] of windowed x9 cannam@85: cannam@85: smull r5, r6, r0, r12 @ r5..r6 = (window_l[13] * (x[10] == -x[7])) cannam@85: ldr r2, =WL16 @ r2 = window_l[16] == window_s[5] cannam@85: movs r5, r5, lsr #28 cannam@85: adc r12, r5, r6, lsl #4 @ r10 = bits[59..28] of windowed x9 cannam@85: cannam@85: smull r5, r6, r2, lr @ r5..r6 = (window_l[16] * (x[11] == -x[6])) cannam@85: cannam@85: ldr r0, =0x00 cannam@85: cannam@85: movs r5, r5, lsr #28 cannam@85: adc lr, r5, r6, lsl #4 @ r10 = bits[59..28] of windowed x9 cannam@85: cannam@85: stmdb r1!, { r7 - r10, r12, lr } @ store windowed x[6] .. x[11] cannam@85: cannam@85: ldr r5, =0x00 cannam@85: ldr r6, =0x00 cannam@85: ldr r2, =0x00 cannam@85: ldr r3, =0x00 cannam@85: ldr r4, =0x00 cannam@85: cannam@85: stmdb r1!, { r0, r2 - r6 } @ store windowed x[0] .. x[5] cannam@85: cannam@85: b normal_block_x18_to_x35 cannam@85: cannam@85: cannam@85: @---- cannam@85: cannam@85: cannam@85: start_block_x18_to_x35: cannam@85: cannam@85: ldr r4, =WL1 @ r0 = window_l[1] == window_s[0] cannam@85: cannam@85: add r1, r1, #(24*4) @ r1 = &x[24] cannam@85: cannam@85: ldmia r1, { r0, r2, r3 } @ load 3 words from x24, dont update pointer cannam@85: cannam@85: @ r0 = x24 cannam@85: @ r1 = &x[24] cannam@85: @ r2 = x25 cannam@85: @ r3 = x26 cannam@85: @ r4 = WL1 cannam@85: @ r5 = WL4 cannam@85: @ r6 = WL7 cannam@85: @ r7 = WL10 cannam@85: @ r8 = WL13 cannam@85: @ r9 = WL16 cannam@85: @ r10 = . cannam@85: @ r11 = . cannam@85: @ r12 = . cannam@85: @ lr = . cannam@85: cannam@85: ldr r5, =WL4 @ r5 = window_l[4] == window_s[1] cannam@85: cannam@85: smull r10, r11, r4, r0 @ r10..r11 = (window_l[1] * (x[24] == x[29])) cannam@85: ldr r6, =WL7 @ r6 = window_l[7] == window_s[2] cannam@85: movs r10, r10, lsr #28 cannam@85: adc lr, r10, r11, lsl #4 @ lr = bits[59..28] of windowed x29 cannam@85: cannam@85: smull r10, r11, r5, r2 @ r10..r11 = (window_l[4] * (x[25] == x[28])) cannam@85: ldr r7, =WL10 @ r7 = window_l[10] == window_s[3] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r12, r10, r11, lsl #4 @ r12 = bits[59..28] of windowed x28 cannam@85: cannam@85: smull r10, r11, r6, r3 @ r10..r11 = (window_l[7] * (x[26] == x[27])) cannam@85: ldr r8, =WL13 @ r8 = window_l[13] == window_s[4] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r4, r10, r11, lsl #4 @ r4 = bits[59..28] of windowed x27 cannam@85: cannam@85: smull r10, r11, r7, r3 @ r10..r11 = (window_l[10] * x[26]) cannam@85: ldr r9, =WL16 @ r9 = window_l[16] == window_s[5] cannam@85: movs r10, r10, lsr #28 cannam@85: adc r3, r10, r11, lsl #4 @ r3 = bits[59..28] of windowed x26 cannam@85: cannam@85: smull r10, r11, r8, r2 @ r10..r11 = (window_l[13] * x[25]) cannam@85: ldr r5, =0x00 cannam@85: movs r10, r10, lsr #28 cannam@85: adc r2, r10, r11, lsl #4 @ r2 = bits[59..28] of windowed x25 cannam@85: cannam@85: smull r10, r11, r9, r0 @ r10..r11 = (window_l[16] * x[24]) cannam@85: ldr r6, =0x00 cannam@85: movs r10, r10, lsr #28 cannam@85: adc r0, r10, r11, lsl #4 @ r0 = bits[59..28] of windowed x24 cannam@85: cannam@85: stmia r1!, { r0, r2, r3, r4, r12, lr } @ store windowed x[24] .. x[29] cannam@85: cannam@85: ldr r7, =0x00 cannam@85: ldr r8, =0x00 cannam@85: ldr r9, =0x00 cannam@85: ldr r10, =0x00 cannam@85: cannam@85: stmia r1!, { r5 - r10 } @ store windowed x[30] .. x[35] cannam@85: cannam@85: @---- cannam@85: @ NB there are 2 possible exits from this function - this is only one of them cannam@85: @---- cannam@85: cannam@85: add sp, sp, #(21*4) @ return stack frame cannam@85: ldmia sp!, { r4 - r11, pc } @ restore callee saved regs, and return cannam@85: cannam@85: @---- cannam@85: @END cannam@85: @---- cannam@85: