home *** CD-ROM | disk | FTP | other *** search
- |
- | single floating point divide routine
- |
- #ifndef __M68881__
- .text
- .even
- .globl __divsf3, ___divsf3
- #ifdef ERROR_CHECK
- #include "errbase.h"
- .globl __infinitysf
-
- LC0:
- .ascii "floating point division by 0\12\15\0"
- .even
- #endif ERROR_CHECK
-
- __divsf3:
- ___divsf3:
-
- #ifdef ERROR_CHECK
- tstl sp@(8) | check if divisor is 0
- bne no_exception
-
- pea pc@(LC0)
- pea Stderr
- jbsr _fprintf
- addql #8,sp
-
- moveq #Erange,d0 | set _errno to ERANGE
- Emove d0,Errno
- movel __infinitysf,d0 | return signed infinity
- btst #31,sp@(4) | transfer sign of dividend
- beq clear | (mjr++)
- bset #31,d0 |
- clear: |
- rts
-
- no_exception:
- #endif ERROR_CHECK
-
- #ifndef sfp004
- |
- | written by Kai-Uwe Bloem (I5110401@dbstu1.bitnet).
- | Based on a 80x86 floating point packet from comp.os.minix, written by P.Housel
- | Revision 1.2.4 michal 05-93 (ntomczak@vm.ucs.ualberta.ca)
- | + resynchro with errno codes
- | + code smoothing
- | + removed extra code 'retz:' and 'divz:'; these cases are
- | going back through 'retzero:' and 'retinf:'
- |
- | patched by Olaf Flebbe (flebbe@tat.physik.uni-tuebingen.de)
- |
- | Revision 1.2.3 olaf 5-93
- | + correct sign bug for retinf and retzero
- |
- | Revision 1.2.2 olaf 12-92
- | + added support for NaN and Infinites
- | + added support for -0
- |
- | Revision 1.2.1 olaf 11-92
- | + prevent the tie rounding case if dividing is not exact.
- | > paranoia now says: "Division appears to round correctly"
- | ** requires _normsf Version 1.4.2 or later
- |
- | Revision 1.2, kub 01-90 :
- | added support for denormalized numbers
- |
- | Revision 1.1, kub 12-89 :
- | Created single float version for 68000
- |
- | Revision 1.0:
- | original 8088 code from P.S.Housel for double floats
-
- BIAS4 = 0x7F-1
-
- lea sp@(4),a0 | pointer to parameters u and v
- moveml d2-d5,sp@- | save registers
- moveml a0@,d4/d5 | d4 = u, d5 = v
-
- movel #0x7fffff,d3
- movel d4,d0 | d0 = u.exp
- andl d3,d4 | remove exponent from u.mantissa
- swap d0
- movew d0,d2 | d2 = u.sign
-
- movel d5,d1 | d1 = v.exp
- andl d3,d5 | remove exponent from v.mantissa
- swap d1
- eorw d1,d2 | d2 = u.sign ^ v.sign (in bit 15)
-
- moveq #15,d3
- bclr d3,d0 | kill sign bit
- bclr d3,d1 | kill sign bit
- lsrw #7,d0
- lsrw #7,d1
- |
- |
- |
- moveq #-1,d3
- cmpb d3,d0 | comparison with #0xff
- beq 0f | u == NaN || u== Inf
- cmpb d3,d1
- beq 1f | v == NaN || v == Inf
- tstb d0
- bne 3f | u not zero nor denorm
- tstl d4
- beq 2f | 0/ ?
-
- 3: tstw d1
- bne nospec
-
- tstl d5
- bne nospec
- bra retinf | x/0 -> +/- Inf
-
- 0: tstl d4 | u == NaN ?
- bne retnan | NaN/ x
- cmpb d3,d1
- beq retnan | Inf/Inf or Inf/NaN
- | bra retinf | Inf/x | x != Inf && x != NaN
- |
- | Return Infinity with correct sign
- |
- retinf: movel #0xff000000,d0
- lslw #1,d2
- roxrl #1,d0 | shift in high bit as given by d2
- return: moveml sp@+,d2-d5
- rts
-
- 1: tstl d5
- bne retnan | x/NaN
- | bra retzero | x/Inf -> +/- 0
- |
- | Return correct signed zero
- |
- retzero:moveq #0,d0 | zero destination
- lslw #1,d2 | set X bit accordingly
- roxrl #1,d0
- bra return
-
- 2: tstw d1
- bne retzero | 0/x ->+/- 0
- tstl d4
- bne retzero | 0/x
- | bra retnan | 0/0
- |
- | Return NaN
- |
- retnan: movel d3,d0 | d3 contains 0xffffffff
- lsrl #1,d0
- bra return
- |
- | End of special handling
- |
- nospec: moveq #23,d3
- bset d3,d4 | restore implied leading "1"
- tstw d0 | check for zero exponent - no leading "1"
- bne 1f
- bclr d3,d4 | remove it
- addw #1,d0 | "normalize" exponent
- 1:
- tstl d4
- beq retzero | dividing zero
-
- bset d3,d5 | restore implied leading "1"
- tstw d1 | check for zero exponent - no leading "1"
- bne 1f
- bclr d3,d5 | remove it
- addw #1,d1 | "normalize" exponent
- 1:
- # ifndef ERROR_CHECK
- tstl d5
- beq retinf | divide by zero
- # endif ERROR_CHECK
-
- subw d1,d0 | subtract exponents,
- addw #BIAS4-8+1,d0 | add bias back in, account for shift
- addw #34,d0 | add loop offset, +2 for extra rounding bits
- | for denormalized numbers (2 implied by dbra)
- movew #27,d1 | bit number for "implied" pos (+4 for rounding)
- moveql #-1,d3 | zero quotient (for speed a one''s complement)
- subl d5,d4 | initial subtraction, u = u - v
- 2:
- btst d1,d3 | divide until 1 in implied position
- beq 5f
-
- addl d4,d4
- bcs 4f | if carry is set, add, else subtract
-
- addxl d3,d3 | shift quotient and set bit zero
- subl d5,d4 | subtract, u = u - v
- dbra d0,2b | give up if result is denormalized
- bra 5f
- 4:
- addxl d3,d3 | shift quotient and clear bit zero
- addl d5,d4 | add (restore), u = u + v
- dbra d0,2b | give up if result is denormalized
- 5: subqw #2,d0 | remove rounding offset for denormalized nums
- notl d3 | invert quotient to get it right
-
- clrl d1 | zero rounding bits
- tstl d4 | check for exact result
- beq 1f
- moveql #-1,d1 | prevent tie case
- 1: movel d3,d4 | save quotient mantissa
- jmp norm_sf | (registers on stack removed by norm_sf)
-
- #if 0
- | this is dead code right now - retzero and retinf used (mj)
- # ifndef ERROR_CHECK
- retz: moveq #0,d0 | zero destination
- moveml sp@+,d2-d5
- rts | no normalization needed
-
- divz: movel __infinitysf,d0 | return infinity value
- moveml sp@+,d2-d5 | should really cause trap ?!?
- btst #31,sp@(4) | transfer sign of dividend
- beq clear | (mjr++)
- bset #31,d0 |
- rts |
- clear: |
- bclr #31,d0 |
- rts
- # endif ERROR_CHECK
- #endif
-
- #else
-
- | single precision floating point stuff for Atari-gcc using the SFP004
- | or compatible boards with a memory mapped 68881
- | developed with gas
- |
- | single floating point divide routine
- |
- | M. Ritzert (mjr at dmzrzu71)
- | (ritzert@dfg.dbp.de)
- | 4.10.1990
- |
- | +_infinitysf returned instead of a NAN
- | the DOMAIN exception is not supported yet. In case of an exception
- | _errno is always set to ERANGE
-
- | addresses of the 68881 data port. This choice is fastest when much data is
- | transferred between the two processors.
-
- comm = -6
- resp = -16
- zahl = 0
-
- | waiting loop ...
- |
- | wait:
- | ww: cmpiw #0x8900,a0@(resp)
- | beq ww
- | is coded directly by
- | .long 0x0c688900, 0xfff067f8
-
- lea 0xfffffa50:w,a0
- movew #0x4400,a0@(comm) | load first argument to fp0
- cmpiw #0x8900,a0@(resp) | check
- movel sp@(4),a0@
- movew #0x4424,a0@(comm)
- .long 0x0c688900, 0xfff067f8
- movel sp@(8),a0@
- movew #0x6400,a0@(comm) | result to d0
- .long 0x0c688900, 0xfff067f8
- movel a0@,d0 | REMARK: 0/0 returns a NAN
- rts | if ERROR_CHECK is disabled
-
- #endif sfp004
- #endif /* !__M68881__ */
-