/****************************************************************************** Copyright (C) 2000-2001 Andre McCurdy  <armccurdy@yahoo.co.uk>** This program is free software. you can redistribute it and/or modify* it under the terms of the GNU General Public License as published by* the Free Software Foundation@ either version 2 of the License, or* (at your option) any later version.** This program is distributed in the hope that it will be useful,* but WITHOUT ANY WARRANTY, without even the implied warranty of* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the* GNU General Public License for more details.** You should have received a copy of the GNU General Public License* along with this program@ if not, write to the Free Software* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA******************************************************************************** Notes:********************************************************************************* $Id: imdct_l_arm.S,v 1.7 2001/03/25 20:03:34 rob Rel $** 2001/03/24:  Andre McCurdy <armccurdy@yahoo.co.uk>*   - Corrected PIC unsafe loading of address of 'imdct36_long_karray'** 2000/09/20:  Robert Leslie <rob@mars.org>*   - Added a global symbol with leading underscore per suggestion of*     Simon Burge to support linking with the a.out format.** 2000/09/15:  Robert Leslie <rob@mars.org>*   - Fixed a small bug where flags were changed before a conditional branch.** 2000/09/15:  Andre McCurdy <armccurdy@yahoo.co.uk>*   - Applied Nicolas Pitre's rounding optimisation in all remaining places.** 2000/09/09:  Nicolas Pitre <nico@cam.org>*   - Optimized rounding + scaling operations.** 2000/08/09:  Andre McCurdy <armccurdy@yahoo.co.uk>*   - Original created.*****************************************************************************//*   On entry:      r0 = pointer to 18 element input  array      r1 = pointer to 36 element output array      r2 = windowing block type   Stack frame created during execution of the function:   Initial   Holds:   Stack   pointer   minus:       0       4     lr       8     r11      12     r10      16     r9      20     r8      24     r7      28     r6      32     r5      36     r4      40     r2 : windowing block type      44     ct00 high      48     ct00 low      52     ct01 high      56     ct01 low      60     ct04 high      64     ct04 low      68     ct06 high      72     ct06 low      76     ct05 high      80     ct05 low      84     ct03 high      88     ct03 low      92    -ct05 high      96    -ct05 low     100    -ct07 high     104    -ct07 low     108     ct07 high     112     ct07 low     116     ct02 high     120     ct02 low*/#define BLOCK_MODE_NORMAL   0#define BLOCK_MODE_START    1#define BLOCK_MODE_STOP     3#define X0   0x00#define X1   0x04#define X2   0x08#define X3   0x0C#define X4   0x10#define X5   0x14#define X6   0x18#define X7   0x1c#define X8   0x20#define X9   0x24#define X10  0x28#define X11  0x2c#define X12  0x30#define X13  0x34#define X14  0x38#define X15  0x3c#define X16  0x40#define X17  0x44#define x0   0x00#define x1   0x04#define x2   0x08#define x3   0x0C#define x4   0x10#define x5   0x14#define x6   0x18#define x7   0x1c#define x8   0x20#define x9   0x24#define x10  0x28#define x11  0x2c#define x12  0x30#define x13  0x34#define x14  0x38#define x15  0x3c#define x16  0x40#define x17  0x44#define x18  0x48#define x19  0x4c#define x20  0x50#define x21  0x54#define x22  0x58#define x23  0x5c#define x24  0x60#define x25  0x64#define x26  0x68#define x27  0x6c#define x28  0x70#define x29  0x74#define x30  0x78#define x31  0x7c#define x32  0x80#define x33  0x84#define x34  0x88#define x35  0x8c#define K00  0x0ffc19fd#define K01  0x00b2aa3e#define K02  0x0fdcf549#define K03  0x0216a2a2#define K04  0x0f9ee890#define K05  0x03768962#define K06  0x0f426cb5#define K07  0x04cfb0e2#define K08  0x0ec835e8#define K09  0x061f78aa#define K10  0x0e313245#define K11  0x07635284#define K12  0x0d7e8807#define K13  0x0898c779#define K14  0x0cb19346#define K15  0x09bd7ca0#define K16  0x0bcbe352#define K17  0x0acf37ad#define minus_K02 0xf0230ab7#define WL0  0x00b2aa3e#define WL1  0x0216a2a2#define WL2  0x03768962#define WL3  0x04cfb0e2#define WL4  0x061f78aa#define WL5  0x07635284#define WL6  0x0898c779#define WL7  0x09bd7ca0#define WL8  0x0acf37ad#define WL9  0x0bcbe352#define WL10 0x0cb19346#define WL11 0x0d7e8807#define WL12 0x0e313245#define WL13 0x0ec835e8#define WL14 0x0f426cb5#define WL15 0x0f9ee890#define WL16 0x0fdcf549#define WL17 0x0ffc19fd@*****************************************************************************    .text    .align    .global III_imdct_l    .global _III_imdct_lIII_imdct_l:_III_imdct_l:    stmdb   sp!, { r2, r4 - r11, lr }   @ all callee saved regs, plus arg3    ldr     r4, =K08                    @ r4 =  K08    ldr     r5, =K09                    @ r5 =  K09    ldr     r8, [r0, #X4]               @ r8 =  X4    ldr     r9, [r0, #X13]              @ r9 =  X13    rsb     r6, r4, #0                  @ r6 = -K08    rsb     r7, r5, #0                  @ r7 = -K09    smull   r2, r3, r4, r8              @ r2..r3  = (X4 * K08)    smlal   r2, r3, r5, r9              @ r2..r3  = (X4 * K08) + (X13 *  K09) = ct01    smull   r10, lr, r8, r5             @ r10..lr = (X4 * K09)    smlal   r10, lr, r9, r6             @ r10..lr = (X4 * K09) + (X13 * -K08) = ct00    ldr     r8, [r0, #X7]               @ r8 = X7    ldr     r9, [r0, #X16]              @ r9 = X16    stmdb   sp!, { r2, r3, r10, lr }    @ stack ct00_h, ct00_l, ct01_h, ct01_l    add     r8, r8, r9                  @ r8 = (X7 + X16)    ldr     r9, [r0, #X1]               @ r9 = X1    smlal   r2, r3, r6, r8              @ r2..r3  = ct01 + ((X7 + X16) * -K08)    smlal   r2, r3, r7, r9              @ r2..r3 += (X1  * -K09)    ldr     r7, [r0, #X10]              @ r7 = X10    rsbs    r10, r10, #0    rsc     lr, lr, #0                  @ r10..lr  = -ct00    smlal   r2, r3, r5, r7              @ r2..r3  += (X10 *  K09) = ct06    smlal   r10, lr, r9, r6             @ r10..lr  = -ct00 + ( X1        * -K08)    smlal   r10, lr, r8, r5             @ r10..lr +=         ((X7 + X16) *  K09)    smlal   r10, lr, r7, r4             @ r10..lr +=         ( X10       *  K08) = ct04    stmdb   sp!, { r2, r3, r10, lr }    @ stack ct04_h, ct04_l, ct06_h, ct06_l    @----    ldr     r7, [r0, #X0]    ldr     r8, [r0, #X11]    ldr     r9, [r0, #X12]    sub     r7, r7, r8    sub     r7, r7, r9                  @ r7 = (X0 - X11 -X12) = ct14    ldr     r9,  [r0, #X3]    ldr     r8,  [r0, #X8]    ldr     r11, [r0, #X15]    sub     r8, r8, r9    add     r8, r8, r11                 @ r8 = (X8 - X3 + X15) = ct16    add     r11, r7, r8                 @ r11 = ct14 + ct16 = ct18    smlal   r2, r3, r6, r11             @ r2..r3 = ct06 + ((X0 - X11 - X3 + X15 + X8 - X12) * -K08)    ldr     r6,  [r0, #X2]    ldr     r9,  [r0, #X9]    ldr     r12, [r0, #X14]    sub     r6, r6, r9    sub     r6, r6, r12                 @ r6 = (X2 - X9 - X14) = ct15    ldr     r9,  [r0, #X5]    ldr     r12, [r0, #X6]    sub     r9, r9, r12    ldr     r12, [r0, #X17]    sub     r9, r9, r12                 @ r9 = (X5 - X6 - X17) = ct17    add     r12, r9, r6                 @ r12 = ct15 + ct17 = ct19    smlal   r2, r3, r5, r12             @ r2..r3 += ((X2 - X9 + X5 - X6 - X17 - X14) * K09)    smlal   r10, lr, r11, r5            @ r10..lr = ct04 + (ct18 * K09)    smlal   r10, lr, r12, r4            @ r10..lr = ct04 + (ct18 * K09) + (ct19 * K08)    movs    r2, r2, lsr #28    adc     r2, r2, r3, lsl #4          @ r2 = bits[59..28] of r2..r3    str     r2, [r1, #x22]              @ store result x22    movs    r10, r10, lsr #28    adc     r10, r10, lr, lsl #4        @ r10 = bits[59..28] of r10..lr    str     r10, [r1, #x4]              @ store result x4    @----    ldmia   sp, { r2, r3, r4, r5 }      @ r2..r3 = ct06, r4..r5 = ct04 (dont update sp)    @ r2..r3 = ct06    @ r4..r5 = ct04    @ r6     = ct15    @ r7     = ct14    @ r8     = ct16    @ r9     = ct17    @ r10    = .    @ r11    = .    @ r12    = .    @ lr     = .    ldr     r10, =K03                   @ r10 = K03    ldr     lr,  =K15                   @ lr  = K15    smlal   r2, r3, r10, r7             @ r2..r3 = ct06 + (ct14 * K03)    smlal   r4, r5,  lr, r7             @ r4..r5 = ct04 + (ct14 * K15)    ldr     r12, =K14                   @ r12 =  K14    rsb     r10, r10, #0                @ r10 = -K03    smlal   r2, r3,  lr, r6             @ r2..r3 += (ct15 *  K15)    smlal   r4, r5, r10, r6             @ r4..r5 += (ct15 * -K03)    smlal   r2, r3, r12, r8             @ r2..r3 += (ct16 *  K14)    ldr     r11, =minus_K02             @ r11 = -K02    rsb     r12, r12, #0                @ r12 = -K14    smlal   r4, r5, r12, r9             @ r4..r5 += (ct17 * -K14)    smlal   r2, r3, r11, r9             @ r2..r3 += (ct17 * -K02)    smlal   r4, r5, r11, r8             @ r4..r5 += (ct16 * -K02)    movs    r2, r2, lsr #28    adc     r2, r2, r3, lsl #4          @ r2 = bits[59..28] of r2..r3    str     r2, [r1, #x7]               @ store result x7