/*  -*- Mode: Asm -*-  */
;;    Copyright (C) 2012-2020 Free Software Foundation, Inc.
;;    Contributed by Sean D'Epagnier  (sean@depagnier.com)
;;                   Georg-Johann Lay (avr@gjlay.de)

;; This file 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 3, or (at your option) any
;; later version.

;; In addition to the permissions in the GNU General Public License, the
;; Free Software Foundation gives you unlimited permission to link the
;; compiled version of this file into combinations with other programs,
;; and to distribute those combinations without any restriction coming
;; from the use of this file.  (The General Public License restrictions
;; do apply in other respects; for example, they cover modification of
;; the file, and distribution when not linked into a combine
;; executable.)

;; This file 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; see the file COPYING.  If not, write to
;; the Free Software Foundation, 51 Franklin Street, Fifth Floor,
;; Boston, MA 02110-1301, USA.

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; Fixed point library routines for AVR
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

#if defined __AVR_TINY__
#define __zero_reg__ r17
#define __tmp_reg__ r16
#else                                                                                                                                              
#define __zero_reg__ r1
#define __tmp_reg__ r0
#endif

.section .text.libgcc.fixed, "ax", @progbits

#ifndef __AVR_TINY__

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; Conversions to float
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

#if defined (L_fractqqsf)
DEFUN __fractqqsf
    ;; Move in place for SA -> SF conversion
    clr     r22
    mov     r23, r24
    ;; Sign-extend
    lsl     r24
    sbc     r24, r24
    mov     r25, r24
    XJMP    __fractsasf
ENDF __fractqqsf
#endif  /* L_fractqqsf */

#if defined (L_fractuqqsf)
DEFUN __fractuqqsf
    ;; Move in place for USA -> SF conversion
    clr     r22
    mov     r23, r24
    ;; Zero-extend
    clr     r24
    clr     r25
    XJMP    __fractusasf
ENDF __fractuqqsf
#endif  /* L_fractuqqsf */

#if defined (L_fracthqsf)
DEFUN __fracthqsf
    ;; Move in place for SA -> SF conversion
    wmov    22, 24
    ;; Sign-extend
    lsl     r25
    sbc     r24, r24
    mov     r25, r24
    XJMP    __fractsasf
ENDF __fracthqsf
#endif  /* L_fracthqsf */

#if defined (L_fractuhqsf)
DEFUN __fractuhqsf
    ;; Move in place for USA -> SF conversion
    wmov    22, 24
    ;; Zero-extend
    clr     r24
    clr     r25
    XJMP    __fractusasf
ENDF __fractuhqsf
#endif  /* L_fractuhqsf */

#if defined (L_fracthasf)
DEFUN __fracthasf
    ;; Move in place for SA -> SF conversion
    clr     r22
    mov     r23, r24
    mov     r24, r25
    ;; Sign-extend
    lsl     r25
    sbc     r25, r25
    XJMP    __fractsasf
ENDF __fracthasf
#endif  /* L_fracthasf */

#if defined (L_fractuhasf)
DEFUN __fractuhasf
    ;; Move in place for USA -> SF conversion
    clr     r22
    mov     r23, r24
    mov     r24, r25
    ;; Zero-extend
    clr     r25
    XJMP    __fractusasf
ENDF __fractuhasf
#endif  /* L_fractuhasf */


#if defined (L_fractsqsf)
DEFUN __fractsqsf
    XCALL   __floatsisf
    ;; Divide non-zero results by 2^31 to move the
    ;; decimal point into place
    tst     r25
    breq    0f
    subi    r24, exp_lo (31)
    sbci    r25, exp_hi (31)
0:  ret
ENDF __fractsqsf
#endif  /* L_fractsqsf */

#if defined (L_fractusqsf)
DEFUN __fractusqsf
    XCALL   __floatunsisf
    ;; Divide non-zero results by 2^32 to move the
    ;; decimal point into place
    cpse    r25, __zero_reg__
    subi    r25, exp_hi (32)
    ret
ENDF __fractusqsf
#endif  /* L_fractusqsf */

#if defined (L_fractsasf)
DEFUN __fractsasf
    XCALL   __floatsisf
    ;; Divide non-zero results by 2^15 to move the
    ;; decimal point into place
    tst     r25
    breq    0f
    subi    r24, exp_lo (15)
    sbci    r25, exp_hi (15)
0:  ret
ENDF __fractsasf
#endif  /* L_fractsasf */

#if defined (L_fractusasf)
DEFUN __fractusasf
    XCALL   __floatunsisf
    ;; Divide non-zero results by 2^16 to move the
    ;; decimal point into place
    cpse    r25, __zero_reg__
    subi    r25, exp_hi (16)
    ret
ENDF __fractusasf
#endif  /* L_fractusasf */

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; Conversions from float
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

#if defined (L_fractsfqq)
DEFUN __fractsfqq
    ;; Multiply with 2^{24+7} to get a QQ result in r25
    subi    r24, exp_lo (-31)
    sbci    r25, exp_hi (-31)
    XCALL   __fixsfsi
    mov     r24, r25
    ret
ENDF __fractsfqq
#endif  /* L_fractsfqq */

#if defined (L_fractsfuqq)
DEFUN __fractsfuqq
    ;; Multiply with 2^{24+8} to get a UQQ result in r25
    subi    r25, exp_hi (-32)
    XCALL   __fixunssfsi
    mov     r24, r25
    ret
ENDF __fractsfuqq
#endif  /* L_fractsfuqq */

#if defined (L_fractsfha)
DEFUN __fractsfha
    ;; Multiply with 2^{16+7} to get a HA result in r25:r24
    subi    r24, exp_lo (-23)
    sbci    r25, exp_hi (-23)
    XJMP    __fixsfsi
ENDF __fractsfha
#endif  /* L_fractsfha */

#if defined (L_fractsfuha)
DEFUN __fractsfuha
    ;; Multiply with 2^24 to get a UHA result in r25:r24
    subi    r25, exp_hi (-24)
    XJMP    __fixunssfsi
ENDF __fractsfuha
#endif  /* L_fractsfuha */

#if defined (L_fractsfhq)
FALIAS __fractsfsq

DEFUN __fractsfhq
    ;; Multiply with 2^{16+15} to get a HQ result in r25:r24
    ;; resp. with 2^31 to get a SQ result in r25:r22
    subi    r24, exp_lo (-31)
    sbci    r25, exp_hi (-31)
    XJMP    __fixsfsi
ENDF __fractsfhq
#endif  /* L_fractsfhq */

#if defined (L_fractsfuhq)
FALIAS __fractsfusq

DEFUN __fractsfuhq
    ;; Multiply with 2^{16+16} to get a UHQ result in r25:r24
    ;; resp. with 2^32 to get a USQ result in r25:r22
    subi    r25, exp_hi (-32)
    XJMP    __fixunssfsi
ENDF __fractsfuhq
#endif  /* L_fractsfuhq */

#if defined (L_fractsfsa)
DEFUN __fractsfsa
    ;; Multiply with 2^15 to get a SA result in r25:r22
    subi    r24, exp_lo (-15)
    sbci    r25, exp_hi (-15)
    XJMP    __fixsfsi
ENDF __fractsfsa
#endif  /* L_fractsfsa */

#if defined (L_fractsfusa)
DEFUN __fractsfusa
    ;; Multiply with 2^16 to get a USA result in r25:r22
    subi    r25, exp_hi (-16)
    XJMP    __fixunssfsi
ENDF __fractsfusa
#endif  /* L_fractsfusa */


;; For multiplication the functions here are called directly from
;; avr-fixed.md instead of using the standard libcall mechanisms.
;; This can make better code because GCC knows exactly which
;; of the call-used registers (not all of them) are clobbered.  */

/*******************************************************
    Fractional  Multiplication  8 x 8  without MUL
*******************************************************/

#if defined (L_mulqq3) && !defined (__AVR_HAVE_MUL__)
;;; R23 = R24 * R25
;;; Clobbers: __tmp_reg__, R22, R24, R25
;;; Rounding: ???
DEFUN __mulqq3
    XCALL   __fmuls
    ;; TR 18037 requires that  (-1) * (-1)  does not overflow
    ;; The only input that can produce  -1  is  (-1)^2.
    dec     r23
    brvs    0f
    inc     r23
0:  ret
ENDF  __mulqq3
#endif /* L_mulqq3 && ! HAVE_MUL */

/*******************************************************
    Fractional Multiply  .16 x .16  with and without MUL
*******************************************************/

#if defined (L_mulhq3)
;;; Same code with and without MUL, but the interfaces differ:
;;; no MUL: (R25:R24) = (R22:R23) * (R24:R25)
;;;         Clobbers: ABI, called by optabs
;;; MUL:    (R25:R24) = (R19:R18) * (R27:R26)
;;;         Clobbers: __tmp_reg__, R22, R23
;;; Rounding:  -0.5 LSB  <= error  <=  0.5 LSB
DEFUN   __mulhq3
    XCALL   __mulhisi3
    ;; Shift result into place
    lsl     r23
    rol     r24
    rol     r25
    brvs    1f
    ;; Round
    sbrc    r23, 7
    adiw    r24, 1
    ret
1:  ;; Overflow.  TR 18037 requires  (-1)^2  not to overflow
    ldi     r24, lo8 (0x7fff)
    ldi     r25, hi8 (0x7fff)
    ret
ENDF __mulhq3
#endif  /* defined (L_mulhq3) */

#if defined (L_muluhq3)
;;; Same code with and without MUL, but the interfaces differ:
;;; no MUL: (R25:R24) *= (R23:R22)
;;;         Clobbers: ABI, called by optabs
;;; MUL:    (R25:R24) = (R19:R18) * (R27:R26)
;;;         Clobbers: __tmp_reg__, R22, R23
;;; Rounding:  -0.5 LSB  <  error  <=  0.5 LSB
DEFUN   __muluhq3
    XCALL   __umulhisi3
    ;; Round
    sbrc    r23, 7
    adiw    r24, 1
    ret
ENDF __muluhq3
#endif  /* L_muluhq3 */


/*******************************************************
    Fixed  Multiply  8.8 x 8.8  with and without MUL
*******************************************************/

#if defined (L_mulha3)
;;; Same code with and without MUL, but the interfaces differ:
;;; no MUL: (R25:R24) = (R22:R23) * (R24:R25)
;;;         Clobbers: ABI, called by optabs
;;; MUL:    (R25:R24) = (R19:R18) * (R27:R26)
;;;         Clobbers: __tmp_reg__, R22, R23
;;; Rounding:  -0.5 LSB  <=  error  <=  0.5 LSB
DEFUN   __mulha3
    XCALL   __mulhisi3
    lsl     r22
    rol     r23
    rol     r24
    XJMP    __muluha3_round
ENDF __mulha3
#endif  /* L_mulha3 */

#if defined (L_muluha3)
;;; Same code with and without MUL, but the interfaces differ:
;;; no MUL: (R25:R24) *= (R23:R22)
;;;         Clobbers: ABI, called by optabs
;;; MUL:    (R25:R24) = (R19:R18) * (R27:R26)
;;;         Clobbers: __tmp_reg__, R22, R23
;;; Rounding:  -0.5 LSB  <  error  <=  0.5 LSB
DEFUN   __muluha3
    XCALL   __umulhisi3
    XJMP    __muluha3_round
ENDF __muluha3
#endif  /* L_muluha3 */

#if defined (L_muluha3_round)
DEFUN   __muluha3_round
    ;; Shift result into place
    mov     r25, r24
    mov     r24, r23
    ;; Round
    sbrc    r22, 7
    adiw    r24, 1
    ret
ENDF __muluha3_round
#endif  /* L_muluha3_round */


/*******************************************************
    Fixed  Multiplication  16.16 x 16.16
*******************************************************/

;; Bits outside the result (below LSB), used in the signed version
#define GUARD __tmp_reg__

#if defined (__AVR_HAVE_MUL__)

;; Multiplier
#define A0  16
#define A1  A0+1
#define A2  A1+1
#define A3  A2+1

;; Multiplicand
#define B0  20
#define B1  B0+1
#define B2  B1+1
#define B3  B2+1

;; Result
#define C0  24
#define C1  C0+1
#define C2  C1+1
#define C3  C2+1

#if defined (L_mulusa3)
;;; (C3:C0) = (A3:A0) * (B3:B0)
DEFUN __mulusa3
    set
    ;; Fallthru
ENDF  __mulusa3

;;; Round for last digit iff T = 1
;;; Return guard bits in GUARD (__tmp_reg__).
;;; Rounding, T = 0:  -1.0 LSB  <  error  <=  0   LSB
;;; Rounding, T = 1:  -0.5 LSB  <  error  <=  0.5 LSB
DEFUN __mulusa3_round
    ;; Some of the MUL instructions have LSBs outside the result.
    ;; Don't ignore these LSBs in order to tame rounding error.
    ;; Use C2/C3 for these LSBs.

    clr C0
    clr C1
    mul A0, B0  $  movw C2, r0

    mul A1, B0  $  add  C3, r0  $  adc C0, r1
    mul A0, B1  $  add  C3, r0  $  adc C0, r1  $  rol C1

    ;; Round if T = 1.  Store guarding bits outside the result for rounding
    ;; and left-shift by the signed version (function below).
    brtc 0f
    sbrc C3, 7
    adiw C0, 1
0:  push C3

    ;; The following MULs don't have LSBs outside the result.
    ;; C2/C3 is the high part.

    mul  A0, B2  $  add C0, r0  $  adc C1, r1  $  sbc  C2, C2
    mul  A1, B1  $  add C0, r0  $  adc C1, r1  $  sbci C2, 0
    mul  A2, B0  $  add C0, r0  $  adc C1, r1  $  sbci C2, 0
    neg  C2

    mul  A0, B3  $  add C1, r0  $  adc C2, r1  $  sbc  C3, C3
    mul  A1, B2  $  add C1, r0  $  adc C2, r1  $  sbci C3, 0
    mul  A2, B1  $  add C1, r0  $  adc C2, r1  $  sbci C3, 0
    mul  A3, B0  $  add C1, r0  $  adc C2, r1  $  sbci C3, 0
    neg  C3

    mul  A1, B3  $  add C2, r0  $  adc C3, r1
    mul  A2, B2  $  add C2, r0  $  adc C3, r1
    mul  A3, B1  $  add C2, r0  $  adc C3, r1

    mul  A2, B3  $  add C3, r0
    mul  A3, B2  $  add C3, r0

    ;; Guard bits used in the signed version below.
    pop  GUARD
    clr  __zero_reg__
    ret
ENDF __mulusa3_round
#endif /* L_mulusa3 */

#if defined (L_mulsa3)
;;; (C3:C0) = (A3:A0) * (B3:B0)
;;; Clobbers: __tmp_reg__, T
;;; Rounding:  -0.5 LSB  <=  error  <=  0.5 LSB
DEFUN __mulsa3
    clt
    XCALL   __mulusa3_round
    ;; A posteriori sign extension of the operands
    tst     B3
    brpl 1f
    sub     C2, A0
    sbc     C3, A1
1:  sbrs    A3, 7
    rjmp 2f
    sub     C2, B0
    sbc     C3, B1
2:
    ;;  Shift 1 bit left to adjust for 15 fractional bits
    lsl     GUARD
    rol     C0
    rol     C1
    rol     C2
    rol     C3
    ;; Round last digit
    lsl     GUARD
    adc     C0, __zero_reg__
    adc     C1, __zero_reg__
    adc     C2, __zero_reg__
    adc     C3, __zero_reg__
    ret
ENDF __mulsa3
#endif /* L_mulsa3 */

#undef A0
#undef A1
#undef A2
#undef A3
#undef B0
#undef B1
#undef B2
#undef B3
#undef C0
#undef C1
#undef C2
#undef C3

#else /* __AVR_HAVE_MUL__ */

#define A0 18
#define A1 A0+1
#define A2 A0+2
#define A3 A0+3

#define B0 22
#define B1 B0+1
#define B2 B0+2
#define B3 B0+3

#define C0  22
#define C1  C0+1
#define C2  C0+2
#define C3  C0+3

;; __tmp_reg__
#define CC0  0
;; __zero_reg__
#define CC1  1
#define CC2  16
#define CC3  17

#define AA0  26
#define AA1  AA0+1
#define AA2  30
#define AA3  AA2+1

#if defined (L_mulsa3)
;;; (R25:R22)  *=  (R21:R18)
;;; Clobbers: ABI, called by optabs
;;; Rounding:  -1 LSB  <=  error  <=  1 LSB
DEFUN   __mulsa3
    push    B0
    push    B1
    push    B3
    clt
    XCALL   __mulusa3_round
    pop     r30
    ;; sign-extend B
    bst     r30, 7
    brtc 1f
    ;; A1, A0 survived in  R27:R26
    sub     C2, AA0
    sbc     C3, AA1
1:
    pop     AA1  ;; B1
    pop     AA0  ;; B0

    ;; sign-extend A.  A3 survived in  R31
    bst     AA3, 7
    brtc 2f
    sub     C2, AA0
    sbc     C3, AA1
2:
    ;;  Shift 1 bit left to adjust for 15 fractional bits
    lsl     GUARD
    rol     C0
    rol     C1
    rol     C2
    rol     C3
    ;; Round last digit
    lsl     GUARD
    adc     C0, __zero_reg__
    adc     C1, __zero_reg__
    adc     C2, __zero_reg__
    adc     C3, __zero_reg__
    ret
ENDF __mulsa3
#endif  /* L_mulsa3 */

#if defined (L_mulusa3)
;;; (R25:R22)  *=  (R21:R18)
;;; Clobbers: ABI, called by optabs
;;; Rounding:  -1 LSB  <=  error  <=  1 LSB
DEFUN __mulusa3
    set
    ;; Fallthru
ENDF  __mulusa3

;;; A[] survives in 26, 27, 30, 31
;;; Also used by __mulsa3 with T = 0
;;; Round if T = 1
;;; Return Guard bits in GUARD (__tmp_reg__), used by signed version.
DEFUN __mulusa3_round
    push    CC2
    push    CC3
    ; clear result
    clr     __tmp_reg__
    wmov    CC2, CC0
    ; save multiplicand
    wmov    AA0, A0
    wmov    AA2, A2
    rjmp 3f

    ;; Loop the integral part

1:  ;; CC += A * 2^n;  n >= 0
    add  CC0,A0  $  adc CC1,A1  $  adc  CC2,A2  $  adc  CC3,A3

2:  ;; A <<= 1
    lsl  A0      $  rol A1      $  rol  A2      $  rol  A3

3:  ;; IBIT(B) >>= 1
    ;; Carry = n-th bit of B;  n >= 0
    lsr     B3
    ror     B2
    brcs 1b
    sbci    B3, 0
    brne 2b

    ;; Loop the fractional part
    ;; B2/B3 is 0 now, use as guard bits for rounding
    ;; Restore multiplicand
    wmov    A0, AA0
    wmov    A2, AA2
    rjmp 5f

4:  ;; CC += A:Guard * 2^n;  n < 0
    add  B3,B2 $  adc  CC0,A0  $  adc  CC1,A1  $  adc  CC2,A2  $  adc  CC3,A3
5:
    ;; A:Guard >>= 1
    lsr  A3   $  ror  A2  $  ror  A1  $  ror   A0  $   ror  B2

    ;; FBIT(B) <<= 1
    ;; Carry = n-th bit of B;  n < 0
    lsl     B0
    rol     B1
    brcs 4b
    sbci    B0, 0
    brne 5b

    ;; Save guard bits and set carry for rounding
    push    B3
    lsl     B3
    ;; Move result into place
    wmov    C2, CC2
    wmov    C0, CC0
    clr     __zero_reg__
    brtc 6f
    ;; Round iff T = 1
    adc     C0, __zero_reg__
    adc     C1, __zero_reg__
    adc     C2, __zero_reg__
    adc     C3, __zero_reg__
6:
    pop     GUARD
    ;; Epilogue
    pop     CC3
    pop     CC2
    ret
ENDF __mulusa3_round
#endif  /* L_mulusa3 */

#undef A0
#undef A1
#undef A2
#undef A3
#undef B0
#undef B1
#undef B2
#undef B3
#undef C0
#undef C1
#undef C2
#undef C3
#undef AA0
#undef AA1
#undef AA2
#undef AA3
#undef CC0
#undef CC1
#undef CC2
#undef CC3

#endif /* __AVR_HAVE_MUL__ */

#undef GUARD

/***********************************************************
    Fixed  unsigned saturated Multiplication  8.8 x 8.8
***********************************************************/

#define C0  22
#define C1  C0+1
#define C2  C0+2
#define C3  C0+3
#define SS __tmp_reg__

#if defined (L_usmuluha3)
DEFUN __usmuluha3
    ;; Widening multiply
#ifdef __AVR_HAVE_MUL__
    ;; Adjust interface
    movw    R26, R22
    movw    R18, R24
#endif /* HAVE MUL */
    XCALL   __umulhisi3
    tst     C3
    brne .Lmax
    ;; Round, target is in C1..C2
    lsl     C0
    adc     C1, __zero_reg__
    adc     C2, __zero_reg__
    brcs .Lmax
    ;; Move result into place
    mov     C3, C2
    mov     C2, C1
    ret
.Lmax:
    ;; Saturate
    ldi     C2, 0xff
    ldi     C3, 0xff
    ret
ENDF  __usmuluha3
#endif /* L_usmuluha3 */

/***********************************************************
    Fixed signed saturated Multiplication  s8.7 x s8.7
***********************************************************/

#if defined (L_ssmulha3)
DEFUN __ssmulha3
    ;; Widening multiply
#ifdef __AVR_HAVE_MUL__
    ;; Adjust interface
    movw    R26, R22
    movw    R18, R24
#endif /* HAVE MUL */
    XCALL   __mulhisi3
    ;; Adjust decimal point
    lsl     C0
    rol     C1
    rol     C2
    brvs .LsatC3.3
    ;; The 9 MSBs must be the same
    rol     C3
    sbc     SS, SS
    cp      C3, SS
    brne .LsatSS
    ;; Round
    lsl     C0
    adc     C1, __zero_reg__
    adc     C2, __zero_reg__
    brvs .Lmax
    ;; Move result into place
    mov    C3, C2
    mov    C2, C1
    ret
.Lmax:
    ;; Load 0x7fff
    clr     C3
.LsatC3.3:
    ;; C3 <  0 -->  0x8000
    ;; C3 >= 0 -->  0x7fff
    mov     SS, C3
.LsatSS:
    ;; Load min / max value:
    ;; SS = -1  -->  0x8000
    ;; SS =  0  -->  0x7fff
    ldi     C3, 0x7f
    ldi     C2, 0xff
    sbrc    SS, 7
    adiw    C2, 1
    ret
ENDF  __ssmulha3
#endif /* L_ssmulha3 */

#undef C0
#undef C1
#undef C2
#undef C3
#undef SS

/***********************************************************
    Fixed  unsigned saturated Multiplication  16.16 x 16.16
***********************************************************/

#define C0  18
#define C1  C0+1
#define C2  C0+2
#define C3  C0+3
#define C4  C0+4
#define C5  C0+5
#define C6  C0+6
#define C7  C0+7
#define SS __tmp_reg__

#if defined (L_usmulusa3)
;; R22[4] = R22[4] *{ssat} R18[4]
;; Ordinary ABI function
DEFUN __usmulusa3
    ;; Widening multiply
    XCALL   __umulsidi3
    or      C7, C6
    brne .Lmax
    ;; Round, target is in C2..C5
    lsl     C1
    adc     C2, __zero_reg__
    adc     C3, __zero_reg__
    adc     C4, __zero_reg__
    adc     C5, __zero_reg__
    brcs .Lmax
    ;; Move result into place
    wmov    C6, C4
    wmov    C4, C2
    ret
.Lmax:
    ;; Saturate
    ldi     C7, 0xff
    ldi     C6, 0xff
    wmov    C4, C6
    ret
ENDF  __usmulusa3
#endif /* L_usmulusa3 */

/***********************************************************
    Fixed signed saturated Multiplication  s16.15 x s16.15
***********************************************************/

#if defined (L_ssmulsa3)
;; R22[4] = R22[4] *{ssat} R18[4]
;; Ordinary ABI function
DEFUN __ssmulsa3
    ;; Widening multiply
    XCALL   __mulsidi3
    ;; Adjust decimal point
    lsl     C1
    rol     C2
    rol     C3
    rol     C4
    rol     C5
    brvs .LsatC7.7
    ;; The 17 MSBs must be the same
    rol     C6
    rol     C7
    sbc     SS, SS
    cp      C6, SS
    cpc     C7, SS
    brne .LsatSS
    ;; Round
    lsl     C1
    adc     C2, __zero_reg__
    adc     C3, __zero_reg__
    adc     C4, __zero_reg__
    adc     C5, __zero_reg__
    brvs .Lmax
    ;; Move result into place
    wmov    C6, C4
    wmov    C4, C2
    ret

.Lmax:
    ;; Load 0x7fffffff
    clr     C7
.LsatC7.7:
    ;; C7 <  0 -->  0x80000000
    ;; C7 >= 0 -->  0x7fffffff
    lsl     C7
    sbc     SS, SS
.LsatSS:
    ;; Load min / max value:
    ;; SS = -1  -->  0x80000000
    ;; SS =  0  -->  0x7fffffff
    com     SS
    mov     C4, SS
    mov     C5, C4
    wmov    C6, C4
    subi    C7, 0x80
    ret
ENDF  __ssmulsa3
#endif /* L_ssmulsa3 */

#undef C0
#undef C1
#undef C2
#undef C3
#undef C4
#undef C5
#undef C6
#undef C7
#undef SS

/*******************************************************
      Fractional Division 8 / 8
*******************************************************/

#define r_divd  r25     /* dividend */
#define r_quo   r24     /* quotient */
#define r_div   r22     /* divisor */
#define r_sign  __tmp_reg__

#if defined (L_divqq3)
DEFUN   __divqq3
    mov     r_sign, r_divd
    eor     r_sign, r_div
    sbrc    r_div, 7
    neg     r_div
    sbrc    r_divd, 7
    neg     r_divd
    XCALL   __divqq_helper
    lsr     r_quo
    sbrc    r_sign, 7   ; negate result if needed
    neg     r_quo
    ret
ENDF __divqq3
#endif  /* L_divqq3 */

#if defined (L_udivuqq3)
DEFUN   __udivuqq3
    cp      r_divd, r_div
    brsh    0f
    XJMP __divqq_helper
    ;; Result is out of [0, 1)  ==>  Return 1 - eps.
0:  ldi     r_quo, 0xff
    ret
ENDF __udivuqq3
#endif  /* L_udivuqq3 */


#if defined (L_divqq_helper)
DEFUN   __divqq_helper
    clr     r_quo           ; clear quotient
    inc     __zero_reg__    ; init loop counter, used per shift
__udivuqq3_loop:
    lsl     r_divd          ; shift dividend
    brcs    0f              ; dividend overflow
    cp      r_divd,r_div    ; compare dividend & divisor
    brcc    0f              ; dividend >= divisor
    rol     r_quo           ; shift quotient (with CARRY)
    rjmp    __udivuqq3_cont
0:
    sub     r_divd,r_div    ; restore dividend
    lsl     r_quo           ; shift quotient (without CARRY)
__udivuqq3_cont:
    lsl     __zero_reg__    ; shift loop-counter bit
    brne    __udivuqq3_loop
    com     r_quo           ; complement result
                            ; because C flag was complemented in loop
    ret
ENDF __divqq_helper
#endif  /* L_divqq_helper */

#undef  r_divd
#undef  r_quo
#undef  r_div
#undef  r_sign


/*******************************************************
    Fractional Division 16 / 16
*******************************************************/
#define r_divdL 26     /* dividend Low */
#define r_divdH 27     /* dividend Hig */
#define r_quoL  24     /* quotient Low */
#define r_quoH  25     /* quotient High */
#define r_divL  22     /* divisor */
#define r_divH  23     /* divisor */
#define r_cnt   21

#if defined (L_divhq3)
DEFUN   __divhq3
    mov     r0, r_divdH
    eor     r0, r_divH
    sbrs    r_divH, 7
    rjmp    1f
    NEG2    r_divL
1:
    sbrs    r_divdH, 7
    rjmp    2f
    NEG2    r_divdL
2:
    cp      r_divdL, r_divL
    cpc     r_divdH, r_divH
    breq    __divhq3_minus1  ; if equal return -1
    XCALL   __udivuhq3
    lsr     r_quoH
    ror     r_quoL
    brpl    9f
    ;; negate result if needed
    NEG2    r_quoL
9:
    ret
__divhq3_minus1:
    ldi     r_quoH, 0x80
    clr     r_quoL
    ret
ENDF __divhq3
#endif  /* defined (L_divhq3) */

#if defined (L_udivuhq3)
DEFUN   __udivuhq3
    sub     r_quoH,r_quoH   ; clear quotient and carry
    ;; FALLTHRU
ENDF __udivuhq3

DEFUN   __udivuha3_common
    clr     r_quoL          ; clear quotient
    ldi     r_cnt,16        ; init loop counter
__udivuhq3_loop:
    rol     r_divdL         ; shift dividend (with CARRY)
    rol     r_divdH
    brcs    __udivuhq3_ep   ; dividend overflow
    cp      r_divdL,r_divL  ; compare dividend & divisor
    cpc     r_divdH,r_divH
    brcc    __udivuhq3_ep   ; dividend >= divisor
    rol     r_quoL          ; shift quotient (with CARRY)
    rjmp    __udivuhq3_cont
__udivuhq3_ep:
    sub     r_divdL,r_divL  ; restore dividend
    sbc     r_divdH,r_divH
    lsl     r_quoL          ; shift quotient (without CARRY)
__udivuhq3_cont:
    rol     r_quoH          ; shift quotient
    dec     r_cnt           ; decrement loop counter
    brne    __udivuhq3_loop
    com     r_quoL          ; complement result
    com     r_quoH          ; because C flag was complemented in loop
    ret
ENDF __udivuha3_common
#endif  /* defined (L_udivuhq3) */

/*******************************************************
    Fixed Division 8.8 / 8.8
*******************************************************/
#if defined (L_divha3)
DEFUN   __divha3
    mov     r0, r_divdH
    eor     r0, r_divH
    sbrs    r_divH, 7
    rjmp    1f
    NEG2    r_divL
1:
    sbrs    r_divdH, 7
    rjmp    2f
    NEG2    r_divdL
2:
    XCALL   __udivuha3
    lsr     r_quoH  ; adjust to 7 fractional bits
    ror     r_quoL
    sbrs    r0, 7   ; negate result if needed
    ret
    NEG2    r_quoL
    ret
ENDF __divha3
#endif  /* defined (L_divha3) */

#if defined (L_udivuha3)
DEFUN   __udivuha3
    mov     r_quoH, r_divdL
    mov     r_divdL, r_divdH
    clr     r_divdH
    lsl     r_quoH     ; shift quotient into carry
    XJMP    __udivuha3_common ; same as fractional after rearrange
ENDF __udivuha3
#endif  /* defined (L_udivuha3) */

#undef  r_divdL
#undef  r_divdH
#undef  r_quoL
#undef  r_quoH
#undef  r_divL
#undef  r_divH
#undef  r_cnt

/*******************************************************
    Fixed Division 16.16 / 16.16
*******************************************************/

#define r_arg1L  24    /* arg1 gets passed already in place */
#define r_arg1H  25
#define r_arg1HL 26
#define r_arg1HH 27
#define r_divdL  26    /* dividend Low */
#define r_divdH  27
#define r_divdHL 30
#define r_divdHH 31    /* dividend High */
#define r_quoL   22    /* quotient Low */
#define r_quoH   23
#define r_quoHL  24
#define r_quoHH  25    /* quotient High */
#define r_divL   18    /* divisor Low */
#define r_divH   19
#define r_divHL  20
#define r_divHH  21    /* divisor High */
#define r_cnt  __zero_reg__  /* loop count (0 after the loop!) */

#if defined (L_divsa3)
DEFUN   __divsa3
    mov     r0, r_arg1HH
    eor     r0, r_divHH
    sbrs    r_divHH, 7
    rjmp    1f
    NEG4    r_divL
1:
    sbrs    r_arg1HH, 7
    rjmp    2f
    NEG4    r_arg1L
2:
    XCALL   __udivusa3
    lsr     r_quoHH ; adjust to 15 fractional bits
    ror     r_quoHL
    ror     r_quoH
    ror     r_quoL
    sbrs    r0, 7   ; negate result if needed
    ret
    ;; negate r_quoL
    XJMP    __negsi2
ENDF __divsa3
#endif  /* defined (L_divsa3) */

#if defined (L_udivusa3)
DEFUN   __udivusa3
    ldi     r_divdHL, 32    ; init loop counter
    mov     r_cnt, r_divdHL
    clr     r_divdHL
    clr     r_divdHH
    wmov    r_quoL, r_divdHL
    lsl     r_quoHL         ; shift quotient into carry
    rol     r_quoHH
__udivusa3_loop:
    rol     r_divdL         ; shift dividend (with CARRY)
    rol     r_divdH
    rol     r_divdHL
    rol     r_divdHH
    brcs    __udivusa3_ep   ; dividend overflow
    cp      r_divdL,r_divL  ; compare dividend & divisor
    cpc     r_divdH,r_divH
    cpc     r_divdHL,r_divHL
    cpc     r_divdHH,r_divHH
    brcc    __udivusa3_ep   ; dividend >= divisor
    rol     r_quoL          ; shift quotient (with CARRY)
    rjmp    __udivusa3_cont
__udivusa3_ep:
    sub     r_divdL,r_divL  ; restore dividend
    sbc     r_divdH,r_divH
    sbc     r_divdHL,r_divHL
    sbc     r_divdHH,r_divHH
    lsl     r_quoL          ; shift quotient (without CARRY)
__udivusa3_cont:
    rol     r_quoH          ; shift quotient
    rol     r_quoHL
    rol     r_quoHH
    dec     r_cnt           ; decrement loop counter
    brne    __udivusa3_loop
    com     r_quoL          ; complement result
    com     r_quoH          ; because C flag was complemented in loop
    com     r_quoHL
    com     r_quoHH
    ret
ENDF __udivusa3
#endif  /* defined (L_udivusa3) */

#undef  r_arg1L
#undef  r_arg1H
#undef  r_arg1HL
#undef  r_arg1HH
#undef  r_divdL
#undef  r_divdH
#undef  r_divdHL
#undef  r_divdHH
#undef  r_quoL
#undef  r_quoH
#undef  r_quoHL
#undef  r_quoHH
#undef  r_divL
#undef  r_divH
#undef  r_divHL
#undef  r_divHH
#undef  r_cnt


;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; Saturation, 1 Byte
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

;; First Argument and Return Register
#define A0  24

#if defined (L_ssabs_1)
DEFUN __ssabs_1
    sbrs    A0, 7
    ret
    neg     A0
    sbrc    A0,7
    dec     A0
    ret
ENDF __ssabs_1
#endif /* L_ssabs_1 */

#undef A0



;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; Saturation, 2 Bytes
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

;; First Argument and Return Register
#define A0  24
#define A1  A0+1

#if defined (L_ssneg_2)
DEFUN __ssneg_2
    NEG2    A0
    brvc 0f
    sbiw    A0, 1
0:  ret
ENDF __ssneg_2
#endif /* L_ssneg_2 */

#if defined (L_ssabs_2)
DEFUN __ssabs_2
    sbrs    A1, 7
    ret
    XJMP    __ssneg_2
ENDF __ssabs_2
#endif /* L_ssabs_2 */

#undef A0
#undef A1



;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; Saturation, 4 Bytes
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

;; First Argument and Return Register
#define A0  22
#define A1  A0+1
#define A2  A0+2
#define A3  A0+3

#if defined (L_ssneg_4)
DEFUN __ssneg_4
    XCALL   __negsi2
    brvc 0f
    ldi     A3, 0x7f
    ldi     A2, 0xff
    ldi     A1, 0xff
    ldi     A0, 0xff
0:  ret
ENDF __ssneg_4
#endif /* L_ssneg_4 */

#if defined (L_ssabs_4)
DEFUN __ssabs_4
    sbrs    A3, 7
    ret
    XJMP    __ssneg_4
ENDF __ssabs_4
#endif /* L_ssabs_4 */

#undef A0
#undef A1
#undef A2
#undef A3



;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; Saturation, 8 Bytes
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

;; First Argument and Return Register
#define A0  18
#define A1  A0+1
#define A2  A0+2
#define A3  A0+3
#define A4  A0+4
#define A5  A0+5
#define A6  A0+6
#define A7  A0+7

#if defined (L_clr_8)
FALIAS __usneguta2
FALIAS __usneguda2
FALIAS __usnegudq2

;; Clear Carry and all Bytes
DEFUN __clr_8
    ;; Clear Carry and set Z
    sub     A7, A7
    ;; FALLTHRU
ENDF  __clr_8
;; Propagate Carry to all Bytes, Carry unaltered
DEFUN __sbc_8
    sbc     A7, A7
    sbc     A6, A6
    wmov    A4, A6
    wmov    A2, A6
    wmov    A0, A6
    ret
ENDF __sbc_8
#endif /* L_clr_8 */

#if defined (L_ssneg_8)
FALIAS __ssnegta2
FALIAS __ssnegda2
FALIAS __ssnegdq2

DEFUN __ssneg_8
    XCALL   __negdi2
    brvc 0f
    ;; A[] = 0x7fffffff
    sec
    XCALL   __sbc_8
    ldi     A7, 0x7f
0:  ret
ENDF __ssneg_8
#endif /* L_ssneg_8 */

#if defined (L_ssabs_8)
FALIAS __ssabsta2
FALIAS __ssabsda2
FALIAS __ssabsdq2

DEFUN __ssabs_8
    sbrs    A7, 7
    ret
    XJMP    __ssneg_8
ENDF __ssabs_8
#endif /* L_ssabs_8 */

;; Second Argument
#define B0  10
#define B1  B0+1
#define B2  B0+2
#define B3  B0+3
#define B4  B0+4
#define B5  B0+5
#define B6  B0+6
#define B7  B0+7

#if defined (L_usadd_8)
FALIAS __usadduta3
FALIAS __usadduda3
FALIAS __usaddudq3

DEFUN __usadd_8
    XCALL   __adddi3
    brcs 0f
    ret
0:  ;; A[] = 0xffffffff
    XJMP    __sbc_8
ENDF __usadd_8
#endif /* L_usadd_8 */

#if defined (L_ussub_8)
FALIAS __ussubuta3
FALIAS __ussubuda3
FALIAS __ussubudq3

DEFUN __ussub_8
    XCALL   __subdi3
    brcs 0f
    ret
0:  ;; A[] = 0
    XJMP    __clr_8
ENDF __ussub_8
#endif /* L_ussub_8 */

#if defined (L_ssadd_8)
FALIAS __ssaddta3
FALIAS __ssaddda3
FALIAS __ssadddq3

DEFUN __ssadd_8
    XCALL   __adddi3
    brvc 0f
    ;; A = (B >= 0) ? INT64_MAX : INT64_MIN
    cpi     B7, 0x80
    XCALL   __sbc_8
    subi    A7, 0x80
0:  ret
ENDF __ssadd_8
#endif /* L_ssadd_8 */

#if defined (L_sssub_8)
FALIAS __sssubta3
FALIAS __sssubda3
FALIAS __sssubdq3

DEFUN __sssub_8
    XCALL   __subdi3
    brvc 0f
    ;; A = (B < 0) ? INT64_MAX : INT64_MIN
    ldi     A7, 0x7f
    cp      A7, B7
    XCALL   __sbc_8
    subi    A7, 0x80
0:  ret
ENDF __sssub_8
#endif /* L_sssub_8 */

#undef A0
#undef A1
#undef A2
#undef A3
#undef A4
#undef A5
#undef A6
#undef A7
#undef B0
#undef B1
#undef B2
#undef B3
#undef B4
#undef B5
#undef B6
#undef B7


;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; Rounding Helpers
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

#ifdef L_mask1

#define AA 24
#define CC 25

;; R25 = 1 << (R24 & 7)
;; CC  = 1 << (AA  & 7)
;; Clobbers: None
DEFUN __mask1
    ;; CC = 2 ^ AA.1
    ldi     CC, 1 << 2
    sbrs    AA, 1
    ldi     CC, 1 << 0
    ;; CC *= 2 ^ AA.0
    sbrc    AA, 0
    lsl     CC
    ;; CC *= 2 ^ AA.2
    sbrc    AA, 2
    swap    CC
    ret
ENDF __mask1

#undef AA
#undef CC
#endif /* L_mask1 */

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

;; The rounding point. Any bits smaller than
;; 2^{-RP} will be cleared.
#define RP R24

#define A0 22
#define A1 A0 + 1

#define C0 24
#define C1 C0 + 1

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; Rounding, 1 Byte
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

#ifdef L_roundqq3

;; R24 = round (R22, R24)
;; Clobbers: R22, __tmp_reg__
DEFUN  __roundqq3
    mov     __tmp_reg__, C1
    subi    RP, __QQ_FBIT__ - 1
    neg     RP
    ;; R25 = 1 << RP  (Total offset is FBIT-1 - RP)
    XCALL   __mask1
    mov     C0, C1
    ;; Add-Saturate 2^{-RP-1}
    add     A0, C0
    brvc 0f
    ldi     C0, 0x7f
    rjmp 9f
0:  ;; Mask out bits beyond RP
    lsl     C0
    neg     C0
    and     C0, A0
9:  mov     C1, __tmp_reg__
    ret
ENDF  __roundqq3
#endif /* L_roundqq3 */

#ifdef L_rounduqq3

;; R24 = round (R22, R24)
;; Clobbers: R22, __tmp_reg__
DEFUN  __rounduqq3
    mov     __tmp_reg__, C1
    subi    RP, __UQQ_FBIT__ - 1
    neg     RP
    ;; R25 = 1 << RP  (Total offset is FBIT-1 - RP)
    XCALL   __mask1
    mov     C0, C1
    ;; Add-Saturate 2^{-RP-1}
    add     A0, C0
    brcc 0f
    ldi     C0, 0xff
    rjmp 9f
0:  ;; Mask out bits beyond RP
    lsl     C0
    neg     C0
    and     C0, A0
9:  mov     C1, __tmp_reg__
    ret
ENDF  __rounduqq3
#endif /* L_rounduqq3 */

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; Rounding, 2 Bytes
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

#ifdef L_addmask_2

;; [ R25:R24 =  1 << (R24 & 15)
;;   R23:R22 += 1 << (R24 & 15) ]
;; SREG is set according to the addition
DEFUN __addmask_2
    ;; R25 = 1 << (R24 & 7)
    XCALL   __mask1
    cpi     RP, 1 << 3
    sbc     C0, C0
    ;; Swap C0 and C1 if RP.3 was set
    and     C0, C1
    eor     C1, C0
    ;; Finally, add the power-of-two:  A[] += C[]
    add     A0, C0
    adc     A1, C1
    ret
ENDF  __addmask_2
#endif /* L_addmask_2 */

#ifdef L_round_s2

;; R25:R24 = round (R23:R22, R24)
;; Clobbers: R23, R22
DEFUN  __roundhq3
    subi    RP, __HQ_FBIT__ - __HA_FBIT__
ENDF   __roundhq3
DEFUN  __roundha3
    subi    RP, __HA_FBIT__ - 1
    neg     RP
    ;; [ R25:R24  = 1 << (FBIT-1 - RP)
    ;;   R23:R22 += 1 << (FBIT-1 - RP) ]
    XCALL   __addmask_2
    XJMP    __round_s2_const
ENDF  __roundha3

#endif /* L_round_s2 */

#ifdef L_round_u2

;; R25:R24 = round (R23:R22, R24)
;; Clobbers: R23, R22
DEFUN  __rounduhq3
    subi    RP, __UHQ_FBIT__ - __UHA_FBIT__
ENDF   __rounduhq3
DEFUN  __rounduha3
    subi    RP, __UHA_FBIT__ - 1
    neg     RP
    ;; [ R25:R24  = 1 << (FBIT-1 - RP)
    ;;   R23:R22 += 1 << (FBIT-1 - RP) ]
    XCALL   __addmask_2
    XJMP    __round_u2_const 
ENDF  __rounduha3

#endif /* L_round_u2 */


#ifdef L_round_2_const

;; Helpers for 2 byte wide rounding

DEFUN  __round_s2_const
    brvc 2f
    ldi     C1, 0x7f
    rjmp 1f
    ;; FALLTHRU (Barrier)
ENDF  __round_s2_const

DEFUN __round_u2_const
    brcc 2f
    ldi     C1, 0xff
1:
    ldi     C0, 0xff
    rjmp 9f
2:
    ;; Saturation is performed now.
    ;; Currently, we have C[] = 2^{-RP-1}
    ;; C[] = 2^{-RP}
    lsl     C0
    rol     C1
    ;;      
    NEG2    C0
    ;; Clear the bits beyond the rounding point.
    and     C0, A0
    and     C1, A1
9:  ret
ENDF  __round_u2_const

#endif /* L_round_2_const */

#undef A0
#undef A1
#undef C0
#undef C1

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; Rounding, 4 Bytes
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

#define A0 18
#define A1 A0 + 1
#define A2 A0 + 2
#define A3 A0 + 3

#define C0 22
#define C1 C0 + 1
#define C2 C0 + 2
#define C3 C0 + 3

#ifdef L_addmask_4

;; [ R25:R22 =  1 << (R24 & 31)
;;   R21:R18 += 1 << (R24 & 31) ]
;; SREG is set according to the addition
DEFUN __addmask_4
    ;; R25 = 1 << (R24 & 7)
    XCALL   __mask1
    cpi     RP, 1 << 4
    sbc     C0, C0
    sbc     C1, C1
    ;; Swap C2 with C3 if RP.3 is not set
    cpi     RP, 1 << 3
    sbc     C2, C2
    and     C2, C3
    eor     C3, C2
    ;; Swap C3:C2 with C1:C0 if RP.4 is not set
    and     C0, C2  $  eor     C2, C0
    and     C1, C3  $  eor     C3, C1
    ;; Finally, add the power-of-two:  A[] += C[]
    add     A0, C0
    adc     A1, C1
    adc     A2, C2
    adc     A3, C3
    ret
ENDF  __addmask_4
#endif /* L_addmask_4 */

#ifdef L_round_s4

;; R25:R22 = round (R21:R18, R24)
;; Clobbers: R18...R21
DEFUN  __roundsq3
    subi    RP, __SQ_FBIT__ - __SA_FBIT__
ENDF   __roundsq3
DEFUN  __roundsa3
    subi    RP, __SA_FBIT__ - 1
    neg     RP
    ;; [ R25:R22  = 1 << (FBIT-1 - RP)
    ;;   R21:R18 += 1 << (FBIT-1 - RP) ]
    XCALL   __addmask_4
    XJMP    __round_s4_const
ENDF  __roundsa3

#endif /* L_round_s4 */

#ifdef L_round_u4

;; R25:R22 = round (R21:R18, R24)
;; Clobbers: R18...R21
DEFUN  __roundusq3
    subi    RP, __USQ_FBIT__ - __USA_FBIT__
ENDF   __roundusq3
DEFUN  __roundusa3
    subi    RP, __USA_FBIT__ - 1
    neg     RP
    ;; [ R25:R22  = 1 << (FBIT-1 - RP)
    ;;   R21:R18 += 1 << (FBIT-1 - RP) ]
    XCALL   __addmask_4
    XJMP    __round_u4_const 
ENDF  __roundusa3

#endif /* L_round_u4 */


#ifdef L_round_4_const

;; Helpers for 4 byte wide rounding

DEFUN  __round_s4_const
    brvc 2f
    ldi     C3, 0x7f
    rjmp 1f
    ;; FALLTHRU (Barrier)
ENDF  __round_s4_const

DEFUN __round_u4_const
    brcc 2f
    ldi     C3, 0xff
1:
    ldi     C2, 0xff
    ldi     C1, 0xff
    ldi     C0, 0xff
    rjmp 9f
2:
    ;; Saturation is performed now.
    ;; Currently, we have C[] = 2^{-RP-1}
    ;; C[] = 2^{-RP}
    lsl     C0
    rol     C1
    rol     C2
    rol     C3
    XCALL   __negsi2
    ;; Clear the bits beyond the rounding point.
    and     C0, A0
    and     C1, A1
    and     C2, A2
    and     C3, A3
9:  ret
ENDF  __round_u4_const

#endif /* L_round_4_const */

#undef A0
#undef A1
#undef A2
#undef A3
#undef C0
#undef C1
#undef C2
#undef C3

#undef RP

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; Rounding, 8 Bytes
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

#define RP     16
#define FBITm1 31

#define C0 18
#define C1 C0 + 1
#define C2 C0 + 2
#define C3 C0 + 3
#define C4 C0 + 4
#define C5 C0 + 5
#define C6 C0 + 6
#define C7 C0 + 7

#define A0 16
#define A1 17
#define A2 26
#define A3 27
#define A4 28
#define A5 29
#define A6 30
#define A7 31


#ifdef L_rounddq3
;; R25:R18 = round (R25:R18, R16)
;; Clobbers: ABI
DEFUN  __rounddq3
    ldi     FBITm1, __DQ_FBIT__ - 1
    clt
    XJMP    __round_x8
ENDF  __rounddq3
#endif /* L_rounddq3 */

#ifdef L_roundudq3
;; R25:R18 = round (R25:R18, R16)
;; Clobbers: ABI
DEFUN  __roundudq3
    ldi     FBITm1, __UDQ_FBIT__ - 1
    set
    XJMP    __round_x8
ENDF  __roundudq3
#endif /* L_roundudq3 */

#ifdef L_roundda3
;; R25:R18 = round (R25:R18, R16)
;; Clobbers: ABI
DEFUN  __roundda3
    ldi     FBITm1, __DA_FBIT__ - 1
    clt
    XJMP    __round_x8
ENDF  __roundda3
#endif /* L_roundda3 */

#ifdef L_rounduda3
;; R25:R18 = round (R25:R18, R16)
;; Clobbers: ABI
DEFUN  __rounduda3
    ldi     FBITm1, __UDA_FBIT__ - 1
    set
    XJMP    __round_x8
ENDF  __rounduda3
#endif /* L_rounduda3 */

#ifdef L_roundta3
;; R25:R18 = round (R25:R18, R16)
;; Clobbers: ABI
DEFUN  __roundta3
    ldi     FBITm1, __TA_FBIT__ - 1
    clt
    XJMP    __round_x8
ENDF  __roundta3
#endif /* L_roundta3 */

#ifdef L_rounduta3
;; R25:R18 = round (R25:R18, R16)
;; Clobbers: ABI
DEFUN  __rounduta3
    ldi     FBITm1, __UTA_FBIT__ - 1
    set
    XJMP    __round_x8
ENDF  __rounduta3
#endif /* L_rounduta3 */


#ifdef L_round_x8
DEFUN __round_x8
    push r16
    push r17
    push r28
    push r29
    ;; Compute log2 of addend from rounding point
    sub     RP, FBITm1
    neg     RP
    ;; Move input to work register A[]
    push    C0
    mov     A1, C1
    wmov    A2, C2
    wmov    A4, C4
    wmov    A6, C6
    ;; C[] = 1 << (FBIT-1 - RP)
    XCALL   __clr_8
    inc     C0
    XCALL   __ashldi3
    pop     A0
    ;; A[] += C[]
    add     A0, C0
    adc     A1, C1
    adc     A2, C2
    adc     A3, C3
    adc     A4, C4
    adc     A5, C5
    adc     A6, C6
    adc     A7, C7
    brts    1f
    ;; Signed
    brvc    3f
    ;; Signed overflow: A[] = 0x7f...
    brvs    2f
1:  ;; Unsigned
    brcc    3f
    ;; Unsigned overflow: A[] = 0xff...
2:  ldi     C7, 0xff
    ldi     C6, 0xff
    wmov    C0, C6
    wmov    C2, C6
    wmov    C4, C6
    bld     C7, 7
    rjmp 9f
3:
    ;;  C[] = -C[] - C[]
    push    A0
    ldi     r16, 1
    XCALL   __ashldi3
    pop     A0
    XCALL   __negdi2
    ;; Clear the bits beyond the rounding point.
    and     C0, A0
    and     C1, A1
    and     C2, A2
    and     C3, A3
    and     C4, A4
    and     C5, A5
    and     C6, A6
    and     C7, A7
9:  ;; Epilogue
    pop r29
    pop r28
    pop r17
    pop r16
    ret
ENDF  __round_x8

#endif /* L_round_x8 */

#undef A0
#undef A1
#undef A2
#undef A3
#undef A4
#undef A5
#undef A6
#undef A7

#undef C0
#undef C1
#undef C2
#undef C3
#undef C4
#undef C5
#undef C6
#undef C7

#undef RP
#undef FBITm1


;; Supply implementations / symbols for the bit-banging functions
;; __builtin_avr_bitsfx and __builtin_avr_fxbits
#ifdef L_ret
DEFUN __ret
    ret
ENDF  __ret
#endif /* L_ret */

#endif /* if not __AVR_TINY__ */