MPCNC/Firmware
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Adjust formatting in stepper.cpp

Browse Source
This commit is contained in:
Scott Lahteine 2018-05-20 08:36:56 -05:00
parent bf61be8ede
commit ddc1a48844
1 changed files with 249 additions and 247 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

496
Marlin/stepper.cpp
View File

@ -50,6 +50,8 @@
* Jerk controlled movements planner added Apr 2018 by Eduardo José Tagle. * Jerk controlled movements planner added Apr 2018 by Eduardo José Tagle.
* Equations based on Synthethos TinyG2 sources, but the fixed-point * Equations based on Synthethos TinyG2 sources, but the fixed-point
* implementation is new, as we are running the ISR with a variable period. * implementation is new, as we are running the ISR with a variable period.
* Also implemented the Bézier velocity curve evaluation in ARM assembler,
* to avoid impacting ISR speed.
*/ */
#include "Marlin.h" #include "Marlin.h"
@ -426,7 +428,7 @@ void Stepper::set_directions() {
* *
* Floating point arithmetic execution time cost is prohibitive, so we will transform the math to * Floating point arithmetic execution time cost is prohibitive, so we will transform the math to
* use fixed point values to be able to evaluate it in realtime. Assuming a maximum of 250000 steps * use fixed point values to be able to evaluate it in realtime. Assuming a maximum of 250000 steps
* per second (driver pulses should at least be 2uS hi/2uS lo), and allocating 2 bits to avoid * per second (driver pulses should at least be 2µS hi/2µS lo), and allocating 2 bits to avoid
* overflows on the evaluation of the Bézier curve, means we can use * overflows on the evaluation of the Bézier curve, means we can use
* *
* t: unsigned Q0.32 (0 <= t < 1) |range 0 to 0xFFFFFFFF unsigned * t: unsigned Q0.32 (0 <= t < 1) |range 0 to 0xFFFFFFFF unsigned
@ -632,65 +634,65 @@ void Stepper::set_directions() {
/* Store initial velocity*/ /* Store initial velocity*/
A("sts bezier_F, %0") A("sts bezier_F, %0")
A("sts bezier_F+1, %1") A("sts bezier_F+1, %1")
A("sts bezier_F+2, %10") /* bezier_F = %10:%1:%0 = v0 */ A("sts bezier_F+2, %10") /* bezier_F = %10:%1:%0 = v0 */
/* Get delta speed */ /* Get delta speed */
A("ldi %2,-1") /* %2 = 0xFF, means A_negative = true */ A("ldi %2,-1") /* %2 = 0xFF, means A_negative = true */
A("clr %8") /* %8 = 0 */ A("clr %8") /* %8 = 0 */
A("sub %0,%3") A("sub %0,%3")
A("sbc %1,%4") A("sbc %1,%4")
A("sbc %10,%5") /* v0 -= v1, C=1 if result is negative */ A("sbc %10,%5") /* v0 -= v1, C=1 if result is negative */
A("brcc 1f") /* branch if result is positive (C=0), that means v0 >= v1 */ A("brcc 1f") /* branch if result is positive (C=0), that means v0 >= v1 */
/* Result was negative, get the absolute value*/ /* Result was negative, get the absolute value*/
A("com %10") A("com %10")
A("com %1") A("com %1")
A("neg %0") A("neg %0")
A("sbc %1,%2") A("sbc %1,%2")
A("sbc %10,%2") /* %10:%1:%0 +1 -> %10:%1:%0 = -(v0 - v1) = (v1 - v0) */ A("sbc %10,%2") /* %10:%1:%0 +1 -> %10:%1:%0 = -(v0 - v1) = (v1 - v0) */
A("clr %2") /* %2 = 0, means A_negative = false */ A("clr %2") /* %2 = 0, means A_negative = false */
/* Store negative flag*/ /* Store negative flag*/
L("1") L("1")
A("sts A_negative, %2") /* Store negative flag */ A("sts A_negative, %2") /* Store negative flag */
/* Compute coefficients A,B and C [20 cycles worst case]*/ /* Compute coefficients A,B and C [20 cycles worst case]*/
A("ldi %9,6") /* %9 = 6 */ A("ldi %9,6") /* %9 = 6 */
A("mul %0,%9") /* r1:r0 = 6*LO(v0-v1) */ A("mul %0,%9") /* r1:r0 = 6*LO(v0-v1) */
A("sts bezier_A, r0") A("sts bezier_A, r0")
A("mov %6,r1") A("mov %6,r1")
A("clr %7") /* %7:%6:r0 = 6*LO(v0-v1) */ A("clr %7") /* %7:%6:r0 = 6*LO(v0-v1) */
A("mul %1,%9") /* r1:r0 = 6*MI(v0-v1) */ A("mul %1,%9") /* r1:r0 = 6*MI(v0-v1) */
A("add %6,r0") A("add %6,r0")
A("adc %7,r1") /* %7:%6:?? += 6*MI(v0-v1) << 8 */ A("adc %7,r1") /* %7:%6:?? += 6*MI(v0-v1) << 8 */
A("mul %10,%9") /* r1:r0 = 6*HI(v0-v1) */ A("mul %10,%9") /* r1:r0 = 6*HI(v0-v1) */
A("add %7,r0") /* %7:%6:?? += 6*HI(v0-v1) << 16 */ A("add %7,r0") /* %7:%6:?? += 6*HI(v0-v1) << 16 */
A("sts bezier_A+1, %6") A("sts bezier_A+1, %6")
A("sts bezier_A+2, %7") /* bezier_A = %7:%6:?? = 6*(v0-v1) [35 cycles worst] */ A("sts bezier_A+2, %7") /* bezier_A = %7:%6:?? = 6*(v0-v1) [35 cycles worst] */
A("ldi %9,15") /* %9 = 15 */ A("ldi %9,15") /* %9 = 15 */
A("mul %0,%9") /* r1:r0 = 5*LO(v0-v1) */ A("mul %0,%9") /* r1:r0 = 5*LO(v0-v1) */
A("sts bezier_B, r0") A("sts bezier_B, r0")
A("mov %6,r1") A("mov %6,r1")
A("clr %7") /* %7:%6:?? = 5*LO(v0-v1) */ A("clr %7") /* %7:%6:?? = 5*LO(v0-v1) */
A("mul %1,%9") /* r1:r0 = 5*MI(v0-v1) */ A("mul %1,%9") /* r1:r0 = 5*MI(v0-v1) */
A("add %6,r0") A("add %6,r0")
A("adc %7,r1") /* %7:%6:?? += 5*MI(v0-v1) << 8 */ A("adc %7,r1") /* %7:%6:?? += 5*MI(v0-v1) << 8 */
A("mul %10,%9") /* r1:r0 = 5*HI(v0-v1) */ A("mul %10,%9") /* r1:r0 = 5*HI(v0-v1) */
A("add %7,r0") /* %7:%6:?? += 5*HI(v0-v1) << 16 */ A("add %7,r0") /* %7:%6:?? += 5*HI(v0-v1) << 16 */
A("sts bezier_B+1, %6") A("sts bezier_B+1, %6")
A("sts bezier_B+2, %7") /* bezier_B = %7:%6:?? = 5*(v0-v1) [50 cycles worst] */ A("sts bezier_B+2, %7") /* bezier_B = %7:%6:?? = 5*(v0-v1) [50 cycles worst] */
A("ldi %9,10") /* %9 = 10 */ A("ldi %9,10") /* %9 = 10 */
A("mul %0,%9") /* r1:r0 = 10*LO(v0-v1) */ A("mul %0,%9") /* r1:r0 = 10*LO(v0-v1) */
A("sts bezier_C, r0") A("sts bezier_C, r0")
A("mov %6,r1") A("mov %6,r1")
A("clr %7") /* %7:%6:?? = 10*LO(v0-v1) */ A("clr %7") /* %7:%6:?? = 10*LO(v0-v1) */
A("mul %1,%9") /* r1:r0 = 10*MI(v0-v1) */ A("mul %1,%9") /* r1:r0 = 10*MI(v0-v1) */
A("add %6,r0") A("add %6,r0")
A("adc %7,r1") /* %7:%6:?? += 10*MI(v0-v1) << 8 */ A("adc %7,r1") /* %7:%6:?? += 10*MI(v0-v1) << 8 */
A("mul %10,%9") /* r1:r0 = 10*HI(v0-v1) */ A("mul %10,%9") /* r1:r0 = 10*HI(v0-v1) */
A("add %7,r0") /* %7:%6:?? += 10*HI(v0-v1) << 16 */ A("add %7,r0") /* %7:%6:?? += 10*HI(v0-v1) << 16 */
A("sts bezier_C+1, %6") A("sts bezier_C+1, %6")
" sts bezier_C+2, %7" /* bezier_C = %7:%6:?? = 10*(v0-v1) [65 cycles worst] */ " sts bezier_C+2, %7" /* bezier_C = %7:%6:?? = 10*(v0-v1) [65 cycles worst] */
: "+r" (r2), : "+r" (r2),
@ -723,357 +725,357 @@ void Stepper::set_directions() {
__asm__ __volatile( __asm__ __volatile(
/* umul24x24to16hi(t, bezier_AV, curr_step); t: Range 0 - 1^16 = 16 bits*/ /* umul24x24to16hi(t, bezier_AV, curr_step); t: Range 0 - 1^16 = 16 bits*/
A("lds %9,bezier_AV") /* %9 = LO(AV)*/ A("lds %9,bezier_AV") /* %9 = LO(AV)*/
A("mul %9,%2") /* r1:r0 = LO(bezier_AV)*LO(curr_step)*/ A("mul %9,%2") /* r1:r0 = LO(bezier_AV)*LO(curr_step)*/
A("mov %7,r1") /* %7 = LO(bezier_AV)*LO(curr_step) >> 8*/ A("mov %7,r1") /* %7 = LO(bezier_AV)*LO(curr_step) >> 8*/
A("clr %8") /* %8:%7 = LO(bezier_AV)*LO(curr_step) >> 8*/ A("clr %8") /* %8:%7 = LO(bezier_AV)*LO(curr_step) >> 8*/
A("lds %10,bezier_AV+1") /* %10 = MI(AV)*/ A("lds %10,bezier_AV+1") /* %10 = MI(AV)*/
A("mul %10,%2") /* r1:r0 = MI(bezier_AV)*LO(curr_step)*/ A("mul %10,%2") /* r1:r0 = MI(bezier_AV)*LO(curr_step)*/
A("add %7,r0") A("add %7,r0")
A("adc %8,r1") /* %8:%7 += MI(bezier_AV)*LO(curr_step)*/ A("adc %8,r1") /* %8:%7 += MI(bezier_AV)*LO(curr_step)*/
A("lds r1,bezier_AV+2") /* r11 = HI(AV)*/ A("lds r1,bezier_AV+2") /* r11 = HI(AV)*/
A("mul r1,%2") /* r1:r0 = HI(bezier_AV)*LO(curr_step)*/ A("mul r1,%2") /* r1:r0 = HI(bezier_AV)*LO(curr_step)*/
A("add %8,r0") /* %8:%7 += HI(bezier_AV)*LO(curr_step) << 8*/ A("add %8,r0") /* %8:%7 += HI(bezier_AV)*LO(curr_step) << 8*/
A("mul %9,%3") /* r1:r0 = LO(bezier_AV)*MI(curr_step)*/ A("mul %9,%3") /* r1:r0 = LO(bezier_AV)*MI(curr_step)*/
A("add %7,r0") A("add %7,r0")
A("adc %8,r1") /* %8:%7 += LO(bezier_AV)*MI(curr_step)*/ A("adc %8,r1") /* %8:%7 += LO(bezier_AV)*MI(curr_step)*/
A("mul %10,%3") /* r1:r0 = MI(bezier_AV)*MI(curr_step)*/ A("mul %10,%3") /* r1:r0 = MI(bezier_AV)*MI(curr_step)*/
A("add %8,r0") /* %8:%7 += LO(bezier_AV)*MI(curr_step) << 8*/ A("add %8,r0") /* %8:%7 += LO(bezier_AV)*MI(curr_step) << 8*/
A("mul %9,%4") /* r1:r0 = LO(bezier_AV)*HI(curr_step)*/ A("mul %9,%4") /* r1:r0 = LO(bezier_AV)*HI(curr_step)*/
A("add %8,r0") /* %8:%7 += LO(bezier_AV)*HI(curr_step) << 8*/ A("add %8,r0") /* %8:%7 += LO(bezier_AV)*HI(curr_step) << 8*/
/* %8:%7 = t*/ /* %8:%7 = t*/
/* uint16_t f = t;*/ /* uint16_t f = t;*/
A("mov %5,%7") /* %6:%5 = f*/ A("mov %5,%7") /* %6:%5 = f*/
A("mov %6,%8") A("mov %6,%8")
/* %6:%5 = f*/ /* %6:%5 = f*/
/* umul16x16to16hi(f, f, t); / Range 16 bits (unsigned) [17] */ /* umul16x16to16hi(f, f, t); / Range 16 bits (unsigned) [17] */
A("mul %5,%7") /* r1:r0 = LO(f) * LO(t)*/ A("mul %5,%7") /* r1:r0 = LO(f) * LO(t)*/
A("mov %9,r1") /* store MIL(LO(f) * LO(t)) in %9, we need it for rounding*/ A("mov %9,r1") /* store MIL(LO(f) * LO(t)) in %9, we need it for rounding*/
A("clr %10") /* %10 = 0*/ A("clr %10") /* %10 = 0*/
A("clr %11") /* %11 = 0*/ A("clr %11") /* %11 = 0*/
A("mul %5,%8") /* r1:r0 = LO(f) * HI(t)*/ A("mul %5,%8") /* r1:r0 = LO(f) * HI(t)*/
A("add %9,r0") /* %9 += LO(LO(f) * HI(t))*/ A("add %9,r0") /* %9 += LO(LO(f) * HI(t))*/
A("adc %10,r1") /* %10 = HI(LO(f) * HI(t))*/ A("adc %10,r1") /* %10 = HI(LO(f) * HI(t))*/
A("adc %11,%0") /* %11 += carry*/ A("adc %11,%0") /* %11 += carry*/
A("mul %6,%7") /* r1:r0 = HI(f) * LO(t)*/ A("mul %6,%7") /* r1:r0 = HI(f) * LO(t)*/
A("add %9,r0") /* %9 += LO(HI(f) * LO(t))*/ A("add %9,r0") /* %9 += LO(HI(f) * LO(t))*/
A("adc %10,r1") /* %10 += HI(HI(f) * LO(t)) */ A("adc %10,r1") /* %10 += HI(HI(f) * LO(t)) */
A("adc %11,%0") /* %11 += carry*/ A("adc %11,%0") /* %11 += carry*/
A("mul %6,%8") /* r1:r0 = HI(f) * HI(t)*/ A("mul %6,%8") /* r1:r0 = HI(f) * HI(t)*/
A("add %10,r0") /* %10 += LO(HI(f) * HI(t))*/ A("add %10,r0") /* %10 += LO(HI(f) * HI(t))*/
A("adc %11,r1") /* %11 += HI(HI(f) * HI(t))*/ A("adc %11,r1") /* %11 += HI(HI(f) * HI(t))*/
A("mov %5,%10") /* %6:%5 = */ A("mov %5,%10") /* %6:%5 = */
A("mov %6,%11") /* f = %10:%11*/ A("mov %6,%11") /* f = %10:%11*/
/* umul16x16to16hi(f, f, t); / Range 16 bits : f = t^3 (unsigned) [17]*/ /* umul16x16to16hi(f, f, t); / Range 16 bits : f = t^3 (unsigned) [17]*/
A("mul %5,%7") /* r1:r0 = LO(f) * LO(t)*/ A("mul %5,%7") /* r1:r0 = LO(f) * LO(t)*/
A("mov %1,r1") /* store MIL(LO(f) * LO(t)) in %1, we need it for rounding*/ A("mov %1,r1") /* store MIL(LO(f) * LO(t)) in %1, we need it for rounding*/
A("clr %10") /* %10 = 0*/ A("clr %10") /* %10 = 0*/
A("clr %11") /* %11 = 0*/ A("clr %11") /* %11 = 0*/
A("mul %5,%8") /* r1:r0 = LO(f) * HI(t)*/ A("mul %5,%8") /* r1:r0 = LO(f) * HI(t)*/
A("add %1,r0") /* %1 += LO(LO(f) * HI(t))*/ A("add %1,r0") /* %1 += LO(LO(f) * HI(t))*/
A("adc %10,r1") /* %10 = HI(LO(f) * HI(t))*/ A("adc %10,r1") /* %10 = HI(LO(f) * HI(t))*/
A("adc %11,%0") /* %11 += carry*/ A("adc %11,%0") /* %11 += carry*/
A("mul %6,%7") /* r1:r0 = HI(f) * LO(t)*/ A("mul %6,%7") /* r1:r0 = HI(f) * LO(t)*/
A("add %1,r0") /* %1 += LO(HI(f) * LO(t))*/ A("add %1,r0") /* %1 += LO(HI(f) * LO(t))*/
A("adc %10,r1") /* %10 += HI(HI(f) * LO(t))*/ A("adc %10,r1") /* %10 += HI(HI(f) * LO(t))*/
A("adc %11,%0") /* %11 += carry*/ A("adc %11,%0") /* %11 += carry*/
A("mul %6,%8") /* r1:r0 = HI(f) * HI(t)*/ A("mul %6,%8") /* r1:r0 = HI(f) * HI(t)*/
A("add %10,r0") /* %10 += LO(HI(f) * HI(t))*/ A("add %10,r0") /* %10 += LO(HI(f) * HI(t))*/
A("adc %11,r1") /* %11 += HI(HI(f) * HI(t))*/ A("adc %11,r1") /* %11 += HI(HI(f) * HI(t))*/
A("mov %5,%10") /* %6:%5 =*/ A("mov %5,%10") /* %6:%5 =*/
A("mov %6,%11") /* f = %10:%11*/ A("mov %6,%11") /* f = %10:%11*/
/* [15 +17*2] = [49]*/ /* [15 +17*2] = [49]*/
/* %4:%3:%2 will be acc from now on*/ /* %4:%3:%2 will be acc from now on*/
/* uint24_t acc = bezier_F; / Range 20 bits (unsigned)*/ /* uint24_t acc = bezier_F; / Range 20 bits (unsigned)*/
A("clr %9") /* "decimal place we get for free"*/ A("clr %9") /* "decimal place we get for free"*/
A("lds %2,bezier_F") A("lds %2,bezier_F")
A("lds %3,bezier_F+1") A("lds %3,bezier_F+1")
A("lds %4,bezier_F+2") /* %4:%3:%2 = acc*/ A("lds %4,bezier_F+2") /* %4:%3:%2 = acc*/
/* if (A_negative) {*/ /* if (A_negative) {*/
A("lds r0,A_negative") A("lds r0,A_negative")
A("or r0,%0") /* Is flag signalling negative? */ A("or r0,%0") /* Is flag signalling negative? */
A("brne 3f") /* If yes, Skip next instruction if A was negative*/ A("brne 3f") /* If yes, Skip next instruction if A was negative*/
A("rjmp 1f") /* Otherwise, jump */ A("rjmp 1f") /* Otherwise, jump */
/* uint24_t v; */ /* uint24_t v; */
/* umul16x24to24hi(v, f, bezier_C); / Range 21bits [29] */ /* umul16x24to24hi(v, f, bezier_C); / Range 21bits [29] */
/* acc -= v; */ /* acc -= v; */
L("3") L("3")
A("lds %10, bezier_C") /* %10 = LO(bezier_C)*/ A("lds %10, bezier_C") /* %10 = LO(bezier_C)*/
A("mul %10,%5") /* r1:r0 = LO(bezier_C) * LO(f)*/ A("mul %10,%5") /* r1:r0 = LO(bezier_C) * LO(f)*/
A("sub %9,r1") A("sub %9,r1")
A("sbc %2,%0") A("sbc %2,%0")
A("sbc %3,%0") A("sbc %3,%0")
A("sbc %4,%0") /* %4:%3:%2:%9 -= HI(LO(bezier_C) * LO(f))*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= HI(LO(bezier_C) * LO(f))*/
A("lds %11, bezier_C+1") /* %11 = MI(bezier_C)*/ A("lds %11, bezier_C+1") /* %11 = MI(bezier_C)*/
A("mul %11,%5") /* r1:r0 = MI(bezier_C) * LO(f)*/ A("mul %11,%5") /* r1:r0 = MI(bezier_C) * LO(f)*/
A("sub %9,r0") A("sub %9,r0")
A("sbc %2,r1") A("sbc %2,r1")
A("sbc %3,%0") A("sbc %3,%0")
A("sbc %4,%0") /* %4:%3:%2:%9 -= MI(bezier_C) * LO(f)*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= MI(bezier_C) * LO(f)*/
A("lds %1, bezier_C+2") /* %1 = HI(bezier_C)*/ A("lds %1, bezier_C+2") /* %1 = HI(bezier_C)*/
A("mul %1,%5") /* r1:r0 = MI(bezier_C) * LO(f)*/ A("mul %1,%5") /* r1:r0 = MI(bezier_C) * LO(f)*/
A("sub %2,r0") A("sub %2,r0")
A("sbc %3,r1") A("sbc %3,r1")
A("sbc %4,%0") /* %4:%3:%2:%9 -= HI(bezier_C) * LO(f) << 8*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= HI(bezier_C) * LO(f) << 8*/
A("mul %10,%6") /* r1:r0 = LO(bezier_C) * MI(f)*/ A("mul %10,%6") /* r1:r0 = LO(bezier_C) * MI(f)*/
A("sub %9,r0") A("sub %9,r0")
A("sbc %2,r1") A("sbc %2,r1")
A("sbc %3,%0") A("sbc %3,%0")
A("sbc %4,%0") /* %4:%3:%2:%9 -= LO(bezier_C) * MI(f)*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= LO(bezier_C) * MI(f)*/
A("mul %11,%6") /* r1:r0 = MI(bezier_C) * MI(f)*/ A("mul %11,%6") /* r1:r0 = MI(bezier_C) * MI(f)*/
A("sub %2,r0") A("sub %2,r0")
A("sbc %3,r1") A("sbc %3,r1")
A("sbc %4,%0") /* %4:%3:%2:%9 -= MI(bezier_C) * MI(f) << 8*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= MI(bezier_C) * MI(f) << 8*/
A("mul %1,%6") /* r1:r0 = HI(bezier_C) * LO(f)*/ A("mul %1,%6") /* r1:r0 = HI(bezier_C) * LO(f)*/
A("sub %3,r0") A("sub %3,r0")
A("sbc %4,r1") /* %4:%3:%2:%9 -= HI(bezier_C) * LO(f) << 16*/ A("sbc %4,r1") /* %4:%3:%2:%9 -= HI(bezier_C) * LO(f) << 16*/
/* umul16x16to16hi(f, f, t); / Range 16 bits : f = t^3 (unsigned) [17]*/ /* umul16x16to16hi(f, f, t); / Range 16 bits : f = t^3 (unsigned) [17]*/
A("mul %5,%7") /* r1:r0 = LO(f) * LO(t)*/ A("mul %5,%7") /* r1:r0 = LO(f) * LO(t)*/
A("mov %1,r1") /* store MIL(LO(f) * LO(t)) in %1, we need it for rounding*/ A("mov %1,r1") /* store MIL(LO(f) * LO(t)) in %1, we need it for rounding*/
A("clr %10") /* %10 = 0*/ A("clr %10") /* %10 = 0*/
A("clr %11") /* %11 = 0*/ A("clr %11") /* %11 = 0*/
A("mul %5,%8") /* r1:r0 = LO(f) * HI(t)*/ A("mul %5,%8") /* r1:r0 = LO(f) * HI(t)*/
A("add %1,r0") /* %1 += LO(LO(f) * HI(t))*/ A("add %1,r0") /* %1 += LO(LO(f) * HI(t))*/
A("adc %10,r1") /* %10 = HI(LO(f) * HI(t))*/ A("adc %10,r1") /* %10 = HI(LO(f) * HI(t))*/
A("adc %11,%0") /* %11 += carry*/ A("adc %11,%0") /* %11 += carry*/
A("mul %6,%7") /* r1:r0 = HI(f) * LO(t)*/ A("mul %6,%7") /* r1:r0 = HI(f) * LO(t)*/
A("add %1,r0") /* %1 += LO(HI(f) * LO(t))*/ A("add %1,r0") /* %1 += LO(HI(f) * LO(t))*/
A("adc %10,r1") /* %10 += HI(HI(f) * LO(t))*/ A("adc %10,r1") /* %10 += HI(HI(f) * LO(t))*/
A("adc %11,%0") /* %11 += carry*/ A("adc %11,%0") /* %11 += carry*/
A("mul %6,%8") /* r1:r0 = HI(f) * HI(t)*/ A("mul %6,%8") /* r1:r0 = HI(f) * HI(t)*/
A("add %10,r0") /* %10 += LO(HI(f) * HI(t))*/ A("add %10,r0") /* %10 += LO(HI(f) * HI(t))*/
A("adc %11,r1") /* %11 += HI(HI(f) * HI(t))*/ A("adc %11,r1") /* %11 += HI(HI(f) * HI(t))*/
A("mov %5,%10") /* %6:%5 =*/ A("mov %5,%10") /* %6:%5 =*/
A("mov %6,%11") /* f = %10:%11*/ A("mov %6,%11") /* f = %10:%11*/
/* umul16x24to24hi(v, f, bezier_B); / Range 22bits [29]*/ /* umul16x24to24hi(v, f, bezier_B); / Range 22bits [29]*/
/* acc += v; */ /* acc += v; */
A("lds %10, bezier_B") /* %10 = LO(bezier_B)*/ A("lds %10, bezier_B") /* %10 = LO(bezier_B)*/
A("mul %10,%5") /* r1:r0 = LO(bezier_B) * LO(f)*/ A("mul %10,%5") /* r1:r0 = LO(bezier_B) * LO(f)*/
A("add %9,r1") A("add %9,r1")
A("adc %2,%0") A("adc %2,%0")
A("adc %3,%0") A("adc %3,%0")
A("adc %4,%0") /* %4:%3:%2:%9 += HI(LO(bezier_B) * LO(f))*/ A("adc %4,%0") /* %4:%3:%2:%9 += HI(LO(bezier_B) * LO(f))*/
A("lds %11, bezier_B+1") /* %11 = MI(bezier_B)*/ A("lds %11, bezier_B+1") /* %11 = MI(bezier_B)*/
A("mul %11,%5") /* r1:r0 = MI(bezier_B) * LO(f)*/ A("mul %11,%5") /* r1:r0 = MI(bezier_B) * LO(f)*/
A("add %9,r0") A("add %9,r0")
A("adc %2,r1") A("adc %2,r1")
A("adc %3,%0") A("adc %3,%0")
A("adc %4,%0") /* %4:%3:%2:%9 += MI(bezier_B) * LO(f)*/ A("adc %4,%0") /* %4:%3:%2:%9 += MI(bezier_B) * LO(f)*/
A("lds %1, bezier_B+2") /* %1 = HI(bezier_B)*/ A("lds %1, bezier_B+2") /* %1 = HI(bezier_B)*/
A("mul %1,%5") /* r1:r0 = MI(bezier_B) * LO(f)*/ A("mul %1,%5") /* r1:r0 = MI(bezier_B) * LO(f)*/
A("add %2,r0") A("add %2,r0")
A("adc %3,r1") A("adc %3,r1")
A("adc %4,%0") /* %4:%3:%2:%9 += HI(bezier_B) * LO(f) << 8*/ A("adc %4,%0") /* %4:%3:%2:%9 += HI(bezier_B) * LO(f) << 8*/
A("mul %10,%6") /* r1:r0 = LO(bezier_B) * MI(f)*/ A("mul %10,%6") /* r1:r0 = LO(bezier_B) * MI(f)*/
A("add %9,r0") A("add %9,r0")
A("adc %2,r1") A("adc %2,r1")
A("adc %3,%0") A("adc %3,%0")
A("adc %4,%0") /* %4:%3:%2:%9 += LO(bezier_B) * MI(f)*/ A("adc %4,%0") /* %4:%3:%2:%9 += LO(bezier_B) * MI(f)*/
A("mul %11,%6") /* r1:r0 = MI(bezier_B) * MI(f)*/ A("mul %11,%6") /* r1:r0 = MI(bezier_B) * MI(f)*/
A("add %2,r0") A("add %2,r0")
A("adc %3,r1") A("adc %3,r1")
A("adc %4,%0") /* %4:%3:%2:%9 += MI(bezier_B) * MI(f) << 8*/ A("adc %4,%0") /* %4:%3:%2:%9 += MI(bezier_B) * MI(f) << 8*/
A("mul %1,%6") /* r1:r0 = HI(bezier_B) * LO(f)*/ A("mul %1,%6") /* r1:r0 = HI(bezier_B) * LO(f)*/
A("add %3,r0") A("add %3,r0")
A("adc %4,r1") /* %4:%3:%2:%9 += HI(bezier_B) * LO(f) << 16*/ A("adc %4,r1") /* %4:%3:%2:%9 += HI(bezier_B) * LO(f) << 16*/
/* umul16x16to16hi(f, f, t); / Range 16 bits : f = t^5 (unsigned) [17]*/ /* umul16x16to16hi(f, f, t); / Range 16 bits : f = t^5 (unsigned) [17]*/
A("mul %5,%7") /* r1:r0 = LO(f) * LO(t)*/ A("mul %5,%7") /* r1:r0 = LO(f) * LO(t)*/
A("mov %1,r1") /* store MIL(LO(f) * LO(t)) in %1, we need it for rounding*/ A("mov %1,r1") /* store MIL(LO(f) * LO(t)) in %1, we need it for rounding*/
A("clr %10") /* %10 = 0*/ A("clr %10") /* %10 = 0*/
A("clr %11") /* %11 = 0*/ A("clr %11") /* %11 = 0*/
A("mul %5,%8") /* r1:r0 = LO(f) * HI(t)*/ A("mul %5,%8") /* r1:r0 = LO(f) * HI(t)*/
A("add %1,r0") /* %1 += LO(LO(f) * HI(t))*/ A("add %1,r0") /* %1 += LO(LO(f) * HI(t))*/
A("adc %10,r1") /* %10 = HI(LO(f) * HI(t))*/ A("adc %10,r1") /* %10 = HI(LO(f) * HI(t))*/
A("adc %11,%0") /* %11 += carry*/ A("adc %11,%0") /* %11 += carry*/
A("mul %6,%7") /* r1:r0 = HI(f) * LO(t)*/ A("mul %6,%7") /* r1:r0 = HI(f) * LO(t)*/
A("add %1,r0") /* %1 += LO(HI(f) * LO(t))*/ A("add %1,r0") /* %1 += LO(HI(f) * LO(t))*/
A("adc %10,r1") /* %10 += HI(HI(f) * LO(t))*/ A("adc %10,r1") /* %10 += HI(HI(f) * LO(t))*/
A("adc %11,%0") /* %11 += carry*/ A("adc %11,%0") /* %11 += carry*/
A("mul %6,%8") /* r1:r0 = HI(f) * HI(t)*/ A("mul %6,%8") /* r1:r0 = HI(f) * HI(t)*/
A("add %10,r0") /* %10 += LO(HI(f) * HI(t))*/ A("add %10,r0") /* %10 += LO(HI(f) * HI(t))*/
A("adc %11,r1") /* %11 += HI(HI(f) * HI(t))*/ A("adc %11,r1") /* %11 += HI(HI(f) * HI(t))*/
A("mov %5,%10") /* %6:%5 =*/ A("mov %5,%10") /* %6:%5 =*/
A("mov %6,%11") /* f = %10:%11*/ A("mov %6,%11") /* f = %10:%11*/
/* umul16x24to24hi(v, f, bezier_A); / Range 21bits [29]*/ /* umul16x24to24hi(v, f, bezier_A); / Range 21bits [29]*/
/* acc -= v; */ /* acc -= v; */
A("lds %10, bezier_A") /* %10 = LO(bezier_A)*/ A("lds %10, bezier_A") /* %10 = LO(bezier_A)*/
A("mul %10,%5") /* r1:r0 = LO(bezier_A) * LO(f)*/ A("mul %10,%5") /* r1:r0 = LO(bezier_A) * LO(f)*/
A("sub %9,r1") A("sub %9,r1")
A("sbc %2,%0") A("sbc %2,%0")
A("sbc %3,%0") A("sbc %3,%0")
A("sbc %4,%0") /* %4:%3:%2:%9 -= HI(LO(bezier_A) * LO(f))*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= HI(LO(bezier_A) * LO(f))*/
A("lds %11, bezier_A+1") /* %11 = MI(bezier_A)*/ A("lds %11, bezier_A+1") /* %11 = MI(bezier_A)*/
A("mul %11,%5") /* r1:r0 = MI(bezier_A) * LO(f)*/ A("mul %11,%5") /* r1:r0 = MI(bezier_A) * LO(f)*/
A("sub %9,r0") A("sub %9,r0")
A("sbc %2,r1") A("sbc %2,r1")
A("sbc %3,%0") A("sbc %3,%0")
A("sbc %4,%0") /* %4:%3:%2:%9 -= MI(bezier_A) * LO(f)*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= MI(bezier_A) * LO(f)*/
A("lds %1, bezier_A+2") /* %1 = HI(bezier_A)*/ A("lds %1, bezier_A+2") /* %1 = HI(bezier_A)*/
A("mul %1,%5") /* r1:r0 = MI(bezier_A) * LO(f)*/ A("mul %1,%5") /* r1:r0 = MI(bezier_A) * LO(f)*/
A("sub %2,r0") A("sub %2,r0")
A("sbc %3,r1") A("sbc %3,r1")
A("sbc %4,%0") /* %4:%3:%2:%9 -= HI(bezier_A) * LO(f) << 8*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= HI(bezier_A) * LO(f) << 8*/
A("mul %10,%6") /* r1:r0 = LO(bezier_A) * MI(f)*/ A("mul %10,%6") /* r1:r0 = LO(bezier_A) * MI(f)*/
A("sub %9,r0") A("sub %9,r0")
A("sbc %2,r1") A("sbc %2,r1")
A("sbc %3,%0") A("sbc %3,%0")
A("sbc %4,%0") /* %4:%3:%2:%9 -= LO(bezier_A) * MI(f)*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= LO(bezier_A) * MI(f)*/
A("mul %11,%6") /* r1:r0 = MI(bezier_A) * MI(f)*/ A("mul %11,%6") /* r1:r0 = MI(bezier_A) * MI(f)*/
A("sub %2,r0") A("sub %2,r0")
A("sbc %3,r1") A("sbc %3,r1")
A("sbc %4,%0") /* %4:%3:%2:%9 -= MI(bezier_A) * MI(f) << 8*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= MI(bezier_A) * MI(f) << 8*/
A("mul %1,%6") /* r1:r0 = HI(bezier_A) * LO(f)*/ A("mul %1,%6") /* r1:r0 = HI(bezier_A) * LO(f)*/
A("sub %3,r0") A("sub %3,r0")
A("sbc %4,r1") /* %4:%3:%2:%9 -= HI(bezier_A) * LO(f) << 16*/ A("sbc %4,r1") /* %4:%3:%2:%9 -= HI(bezier_A) * LO(f) << 16*/
A("jmp 2f") /* Done!*/ A("jmp 2f") /* Done!*/
L("1") L("1")
/* uint24_t v; */ /* uint24_t v; */
/* umul16x24to24hi(v, f, bezier_C); / Range 21bits [29]*/ /* umul16x24to24hi(v, f, bezier_C); / Range 21bits [29]*/
/* acc += v; */ /* acc += v; */
A("lds %10, bezier_C") /* %10 = LO(bezier_C)*/ A("lds %10, bezier_C") /* %10 = LO(bezier_C)*/
A("mul %10,%5") /* r1:r0 = LO(bezier_C) * LO(f)*/ A("mul %10,%5") /* r1:r0 = LO(bezier_C) * LO(f)*/
A("add %9,r1") A("add %9,r1")
A("adc %2,%0") A("adc %2,%0")
A("adc %3,%0") A("adc %3,%0")
A("adc %4,%0") /* %4:%3:%2:%9 += HI(LO(bezier_C) * LO(f))*/ A("adc %4,%0") /* %4:%3:%2:%9 += HI(LO(bezier_C) * LO(f))*/
A("lds %11, bezier_C+1") /* %11 = MI(bezier_C)*/ A("lds %11, bezier_C+1") /* %11 = MI(bezier_C)*/
A("mul %11,%5") /* r1:r0 = MI(bezier_C) * LO(f)*/ A("mul %11,%5") /* r1:r0 = MI(bezier_C) * LO(f)*/
A("add %9,r0") A("add %9,r0")
A("adc %2,r1") A("adc %2,r1")
A("adc %3,%0") A("adc %3,%0")
A("adc %4,%0") /* %4:%3:%2:%9 += MI(bezier_C) * LO(f)*/ A("adc %4,%0") /* %4:%3:%2:%9 += MI(bezier_C) * LO(f)*/
A("lds %1, bezier_C+2") /* %1 = HI(bezier_C)*/ A("lds %1, bezier_C+2") /* %1 = HI(bezier_C)*/
A("mul %1,%5") /* r1:r0 = MI(bezier_C) * LO(f)*/ A("mul %1,%5") /* r1:r0 = MI(bezier_C) * LO(f)*/
A("add %2,r0") A("add %2,r0")
A("adc %3,r1") A("adc %3,r1")
A("adc %4,%0") /* %4:%3:%2:%9 += HI(bezier_C) * LO(f) << 8*/ A("adc %4,%0") /* %4:%3:%2:%9 += HI(bezier_C) * LO(f) << 8*/
A("mul %10,%6") /* r1:r0 = LO(bezier_C) * MI(f)*/ A("mul %10,%6") /* r1:r0 = LO(bezier_C) * MI(f)*/
A("add %9,r0") A("add %9,r0")
A("adc %2,r1") A("adc %2,r1")
A("adc %3,%0") A("adc %3,%0")
A("adc %4,%0") /* %4:%3:%2:%9 += LO(bezier_C) * MI(f)*/ A("adc %4,%0") /* %4:%3:%2:%9 += LO(bezier_C) * MI(f)*/
A("mul %11,%6") /* r1:r0 = MI(bezier_C) * MI(f)*/ A("mul %11,%6") /* r1:r0 = MI(bezier_C) * MI(f)*/
A("add %2,r0") A("add %2,r0")
A("adc %3,r1") A("adc %3,r1")
A("adc %4,%0") /* %4:%3:%2:%9 += MI(bezier_C) * MI(f) << 8*/ A("adc %4,%0") /* %4:%3:%2:%9 += MI(bezier_C) * MI(f) << 8*/
A("mul %1,%6") /* r1:r0 = HI(bezier_C) * LO(f)*/ A("mul %1,%6") /* r1:r0 = HI(bezier_C) * LO(f)*/
A("add %3,r0") A("add %3,r0")
A("adc %4,r1") /* %4:%3:%2:%9 += HI(bezier_C) * LO(f) << 16*/ A("adc %4,r1") /* %4:%3:%2:%9 += HI(bezier_C) * LO(f) << 16*/
/* umul16x16to16hi(f, f, t); / Range 16 bits : f = t^3 (unsigned) [17]*/ /* umul16x16to16hi(f, f, t); / Range 16 bits : f = t^3 (unsigned) [17]*/
A("mul %5,%7") /* r1:r0 = LO(f) * LO(t)*/ A("mul %5,%7") /* r1:r0 = LO(f) * LO(t)*/
A("mov %1,r1") /* store MIL(LO(f) * LO(t)) in %1, we need it for rounding*/ A("mov %1,r1") /* store MIL(LO(f) * LO(t)) in %1, we need it for rounding*/
A("clr %10") /* %10 = 0*/ A("clr %10") /* %10 = 0*/
A("clr %11") /* %11 = 0*/ A("clr %11") /* %11 = 0*/
A("mul %5,%8") /* r1:r0 = LO(f) * HI(t)*/ A("mul %5,%8") /* r1:r0 = LO(f) * HI(t)*/
A("add %1,r0") /* %1 += LO(LO(f) * HI(t))*/ A("add %1,r0") /* %1 += LO(LO(f) * HI(t))*/
A("adc %10,r1") /* %10 = HI(LO(f) * HI(t))*/ A("adc %10,r1") /* %10 = HI(LO(f) * HI(t))*/
A("adc %11,%0") /* %11 += carry*/ A("adc %11,%0") /* %11 += carry*/
A("mul %6,%7") /* r1:r0 = HI(f) * LO(t)*/ A("mul %6,%7") /* r1:r0 = HI(f) * LO(t)*/
A("add %1,r0") /* %1 += LO(HI(f) * LO(t))*/ A("add %1,r0") /* %1 += LO(HI(f) * LO(t))*/
A("adc %10,r1") /* %10 += HI(HI(f) * LO(t))*/ A("adc %10,r1") /* %10 += HI(HI(f) * LO(t))*/
A("adc %11,%0") /* %11 += carry*/ A("adc %11,%0") /* %11 += carry*/
A("mul %6,%8") /* r1:r0 = HI(f) * HI(t)*/ A("mul %6,%8") /* r1:r0 = HI(f) * HI(t)*/
A("add %10,r0") /* %10 += LO(HI(f) * HI(t))*/ A("add %10,r0") /* %10 += LO(HI(f) * HI(t))*/
A("adc %11,r1") /* %11 += HI(HI(f) * HI(t))*/ A("adc %11,r1") /* %11 += HI(HI(f) * HI(t))*/
A("mov %5,%10") /* %6:%5 =*/ A("mov %5,%10") /* %6:%5 =*/
A("mov %6,%11") /* f = %10:%11*/ A("mov %6,%11") /* f = %10:%11*/
/* umul16x24to24hi(v, f, bezier_B); / Range 22bits [29]*/ /* umul16x24to24hi(v, f, bezier_B); / Range 22bits [29]*/
/* acc -= v;*/ /* acc -= v;*/
A("lds %10, bezier_B") /* %10 = LO(bezier_B)*/ A("lds %10, bezier_B") /* %10 = LO(bezier_B)*/
A("mul %10,%5") /* r1:r0 = LO(bezier_B) * LO(f)*/ A("mul %10,%5") /* r1:r0 = LO(bezier_B) * LO(f)*/
A("sub %9,r1") A("sub %9,r1")
A("sbc %2,%0") A("sbc %2,%0")
A("sbc %3,%0") A("sbc %3,%0")
A("sbc %4,%0") /* %4:%3:%2:%9 -= HI(LO(bezier_B) * LO(f))*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= HI(LO(bezier_B) * LO(f))*/
A("lds %11, bezier_B+1") /* %11 = MI(bezier_B)*/ A("lds %11, bezier_B+1") /* %11 = MI(bezier_B)*/
A("mul %11,%5") /* r1:r0 = MI(bezier_B) * LO(f)*/ A("mul %11,%5") /* r1:r0 = MI(bezier_B) * LO(f)*/
A("sub %9,r0") A("sub %9,r0")
A("sbc %2,r1") A("sbc %2,r1")
A("sbc %3,%0") A("sbc %3,%0")
A("sbc %4,%0") /* %4:%3:%2:%9 -= MI(bezier_B) * LO(f)*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= MI(bezier_B) * LO(f)*/
A("lds %1, bezier_B+2") /* %1 = HI(bezier_B)*/ A("lds %1, bezier_B+2") /* %1 = HI(bezier_B)*/
A("mul %1,%5") /* r1:r0 = MI(bezier_B) * LO(f)*/ A("mul %1,%5") /* r1:r0 = MI(bezier_B) * LO(f)*/
A("sub %2,r0") A("sub %2,r0")
A("sbc %3,r1") A("sbc %3,r1")
A("sbc %4,%0") /* %4:%3:%2:%9 -= HI(bezier_B) * LO(f) << 8*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= HI(bezier_B) * LO(f) << 8*/
A("mul %10,%6") /* r1:r0 = LO(bezier_B) * MI(f)*/ A("mul %10,%6") /* r1:r0 = LO(bezier_B) * MI(f)*/
A("sub %9,r0") A("sub %9,r0")
A("sbc %2,r1") A("sbc %2,r1")
A("sbc %3,%0") A("sbc %3,%0")
A("sbc %4,%0") /* %4:%3:%2:%9 -= LO(bezier_B) * MI(f)*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= LO(bezier_B) * MI(f)*/
A("mul %11,%6") /* r1:r0 = MI(bezier_B) * MI(f)*/ A("mul %11,%6") /* r1:r0 = MI(bezier_B) * MI(f)*/
A("sub %2,r0") A("sub %2,r0")
A("sbc %3,r1") A("sbc %3,r1")
A("sbc %4,%0") /* %4:%3:%2:%9 -= MI(bezier_B) * MI(f) << 8*/ A("sbc %4,%0") /* %4:%3:%2:%9 -= MI(bezier_B) * MI(f) << 8*/
A("mul %1,%6") /* r1:r0 = HI(bezier_B) * LO(f)*/ A("mul %1,%6") /* r1:r0 = HI(bezier_B) * LO(f)*/
A("sub %3,r0") A("sub %3,r0")
A("sbc %4,r1") /* %4:%3:%2:%9 -= HI(bezier_B) * LO(f) << 16*/ A("sbc %4,r1") /* %4:%3:%2:%9 -= HI(bezier_B) * LO(f) << 16*/
/* umul16x16to16hi(f, f, t); / Range 16 bits : f = t^5 (unsigned) [17]*/ /* umul16x16to16hi(f, f, t); / Range 16 bits : f = t^5 (unsigned) [17]*/
A("mul %5,%7") /* r1:r0 = LO(f) * LO(t)*/ A("mul %5,%7") /* r1:r0 = LO(f) * LO(t)*/
A("mov %1,r1") /* store MIL(LO(f) * LO(t)) in %1, we need it for rounding*/ A("mov %1,r1") /* store MIL(LO(f) * LO(t)) in %1, we need it for rounding*/
A("clr %10") /* %10 = 0*/ A("clr %10") /* %10 = 0*/
A("clr %11") /* %11 = 0*/ A("clr %11") /* %11 = 0*/
A("mul %5,%8") /* r1:r0 = LO(f) * HI(t)*/ A("mul %5,%8") /* r1:r0 = LO(f) * HI(t)*/
A("add %1,r0") /* %1 += LO(LO(f) * HI(t))*/ A("add %1,r0") /* %1 += LO(LO(f) * HI(t))*/
A("adc %10,r1") /* %10 = HI(LO(f) * HI(t))*/ A("adc %10,r1") /* %10 = HI(LO(f) * HI(t))*/
A("adc %11,%0") /* %11 += carry*/ A("adc %11,%0") /* %11 += carry*/
A("mul %6,%7") /* r1:r0 = HI(f) * LO(t)*/ A("mul %6,%7") /* r1:r0 = HI(f) * LO(t)*/
A("add %1,r0") /* %1 += LO(HI(f) * LO(t))*/ A("add %1,r0") /* %1 += LO(HI(f) * LO(t))*/
A("adc %10,r1") /* %10 += HI(HI(f) * LO(t))*/ A("adc %10,r1") /* %10 += HI(HI(f) * LO(t))*/
A("adc %11,%0") /* %11 += carry*/ A("adc %11,%0") /* %11 += carry*/
A("mul %6,%8") /* r1:r0 = HI(f) * HI(t)*/ A("mul %6,%8") /* r1:r0 = HI(f) * HI(t)*/
A("add %10,r0") /* %10 += LO(HI(f) * HI(t))*/ A("add %10,r0") /* %10 += LO(HI(f) * HI(t))*/
A("adc %11,r1") /* %11 += HI(HI(f) * HI(t))*/ A("adc %11,r1") /* %11 += HI(HI(f) * HI(t))*/
A("mov %5,%10") /* %6:%5 =*/ A("mov %5,%10") /* %6:%5 =*/
A("mov %6,%11") /* f = %10:%11*/ A("mov %6,%11") /* f = %10:%11*/
/* umul16x24to24hi(v, f, bezier_A); / Range 21bits [29]*/ /* umul16x24to24hi(v, f, bezier_A); / Range 21bits [29]*/
/* acc += v; */ /* acc += v; */
A("lds %10, bezier_A") /* %10 = LO(bezier_A)*/ A("lds %10, bezier_A") /* %10 = LO(bezier_A)*/
A("mul %10,%5") /* r1:r0 = LO(bezier_A) * LO(f)*/ A("mul %10,%5") /* r1:r0 = LO(bezier_A) * LO(f)*/
A("add %9,r1") A("add %9,r1")
A("adc %2,%0") A("adc %2,%0")
A("adc %3,%0") A("adc %3,%0")
A("adc %4,%0") /* %4:%3:%2:%9 += HI(LO(bezier_A) * LO(f))*/ A("adc %4,%0") /* %4:%3:%2:%9 += HI(LO(bezier_A) * LO(f))*/
A("lds %11, bezier_A+1") /* %11 = MI(bezier_A)*/ A("lds %11, bezier_A+1") /* %11 = MI(bezier_A)*/
A("mul %11,%5") /* r1:r0 = MI(bezier_A) * LO(f)*/ A("mul %11,%5") /* r1:r0 = MI(bezier_A) * LO(f)*/
A("add %9,r0") A("add %9,r0")
A("adc %2,r1") A("adc %2,r1")
A("adc %3,%0") A("adc %3,%0")
A("adc %4,%0") /* %4:%3:%2:%9 += MI(bezier_A) * LO(f)*/ A("adc %4,%0") /* %4:%3:%2:%9 += MI(bezier_A) * LO(f)*/
A("lds %1, bezier_A+2") /* %1 = HI(bezier_A)*/ A("lds %1, bezier_A+2") /* %1 = HI(bezier_A)*/
A("mul %1,%5") /* r1:r0 = MI(bezier_A) * LO(f)*/ A("mul %1,%5") /* r1:r0 = MI(bezier_A) * LO(f)*/
A("add %2,r0") A("add %2,r0")
A("adc %3,r1") A("adc %3,r1")
A("adc %4,%0") /* %4:%3:%2:%9 += HI(bezier_A) * LO(f) << 8*/ A("adc %4,%0") /* %4:%3:%2:%9 += HI(bezier_A) * LO(f) << 8*/
A("mul %10,%6") /* r1:r0 = LO(bezier_A) * MI(f)*/ A("mul %10,%6") /* r1:r0 = LO(bezier_A) * MI(f)*/
A("add %9,r0") A("add %9,r0")
A("adc %2,r1") A("adc %2,r1")
A("adc %3,%0") A("adc %3,%0")
A("adc %4,%0") /* %4:%3:%2:%9 += LO(bezier_A) * MI(f)*/ A("adc %4,%0") /* %4:%3:%2:%9 += LO(bezier_A) * MI(f)*/
A("mul %11,%6") /* r1:r0 = MI(bezier_A) * MI(f)*/ A("mul %11,%6") /* r1:r0 = MI(bezier_A) * MI(f)*/
A("add %2,r0") A("add %2,r0")
A("adc %3,r1") A("adc %3,r1")
A("adc %4,%0") /* %4:%3:%2:%9 += MI(bezier_A) * MI(f) << 8*/ A("adc %4,%0") /* %4:%3:%2:%9 += MI(bezier_A) * MI(f) << 8*/
A("mul %1,%6") /* r1:r0 = HI(bezier_A) * LO(f)*/ A("mul %1,%6") /* r1:r0 = HI(bezier_A) * LO(f)*/
A("add %3,r0") A("add %3,r0")
A("adc %4,r1") /* %4:%3:%2:%9 += HI(bezier_A) * LO(f) << 16*/ A("adc %4,r1") /* %4:%3:%2:%9 += HI(bezier_A) * LO(f) << 16*/
L("2") L("2")
" clr __zero_reg__" /* C runtime expects r1 = __zero_reg__ = 0 */ " clr __zero_reg__" /* C runtime expects r1 = __zero_reg__ = 0 */
: "+r"(r0), : "+r"(r0),