openbios/openboot/obp/fm/kernel/sparc/multiply.fth
Branch
-
Show log
Commit
-
Hash : 1a08f436
Author :
Date : 2010-08-25T09:23:17
change to well-formed svn directory structure. git-svn-id: svn://coreboot.org/openboot/trunk@2 4486e004-3823-0410-90c7-fb508cc143bb
-
obp/fm/kernel/sparc/multiply.fth
-
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103
id: @(#)multiply.fth 2.6 94/06/18 purpose: copyright: Copyright 1990 Sun Microsystems, Inc. All Rights Reserved \ Copyright 1985-1990 Bradley Forthware decimal \ --- 32*32->64 bit unsigned! multiply ---------------------------------- \ y rs2 y rd code um* ( u1 u2 -- ud[lo hi] ) sp 0 /n* %l0 nget \ l0 <- n1(y) get multiplyer \ (save original y for later) %l0 %g0 wry \ y <- n1 multiplyer \ tos %g0 %l3 or \ l3 <- n2(rs2) multiplicand \ (save original rs2 for later) \ tos %l0 %l1 or \ l2 <- n1.or.n2 (any hi bits) %l1 h# fff %g0 andncc \ cc <- fffff000.and.(n1.or.n2) 0= if \ if neither terms have any \ hi 1's then set do short way. %g0 %g0 %l2 andcc \ N=0, V=0, %l2=0; Neg, oVerflow, \ part-product (delay slot) \ %l2 tos %l2 mulscc \ 1 001 %l2 tos %l2 mulscc \ 2 003 %l2 tos %l2 mulscc \ 3 007 %l2 tos %l2 mulscc \ 4 00f %l2 tos %l2 mulscc \ 5 01f %l2 tos %l2 mulscc \ 6 03f %l2 tos %l2 mulscc \ 7 07f %l2 tos %l2 mulscc \ 8 0ff %l2 tos %l2 mulscc \ 9 1ff %l2 tos %l2 mulscc \ 10 3ff %l2 tos %l2 mulscc \ 11 7ff %l2 tos %l2 mulscc \ 12 fff %l2 tos %l2 mulscc \ 13 1fff %g0 %g0 tos and \ x[hi]-(TOS) <- 0; (msd) %l1 rdy \ l1 <- y (write over saved y \ [not needed for short multiply]) %l2 d# 12 %l2 sll \ l2 <- part-prod << 12; \ msd result (low bits have \ good data) %l1 d# 20 %l1 srl \ l1 <- y >> 20; \ msd result (hi bits have \ good data) \ else \ else do full 32 bit multiply %l1 %l2 %l1 or \ l1 <- mst+lsd 32 bit result (delay) %l2 tos %l2 mulscc \ 1 %l2 tos %l2 mulscc \ 2 %l2 tos %l2 mulscc \ 3 %l2 tos %l2 mulscc \ 4 %l2 tos %l2 mulscc \ 5 %l2 tos %l2 mulscc \ 6 %l2 tos %l2 mulscc \ 7 %l2 tos %l2 mulscc \ 8 %l2 tos %l2 mulscc \ 9 %l2 tos %l2 mulscc \ 10 %l2 tos %l2 mulscc \ 11 %l2 tos %l2 mulscc \ 12 %l2 tos %l2 mulscc \ 13 %l2 tos %l2 mulscc \ 14 %l2 tos %l2 mulscc \ 15 %l2 tos %l2 mulscc \ 16 %l2 tos %l2 mulscc \ 17 %l2 tos %l2 mulscc \ 18 %l2 tos %l2 mulscc \ 19 %l2 tos %l2 mulscc \ 20 %l2 tos %l2 mulscc \ 21 %l2 tos %l2 mulscc \ 22 %l2 tos %l2 mulscc \ 23 %l2 tos %l2 mulscc \ 24 %l2 tos %l2 mulscc \ 25 %l2 tos %l2 mulscc \ 26 %l2 tos %l2 mulscc \ 27 %l2 tos %l2 mulscc \ 28 %l2 tos %l2 mulscc \ 29 %l2 tos %l2 mulscc \ 30 %l2 tos %l2 mulscc \ 31 %l2 tos %l2 mulscc \ 32 %l2 %g0 tos mulscc \ final iteration only shifts; \ move to TOS (delay slot). \ %l3 %g0 %g0 orcc \ was original rs2 negative 0< if nop \ (delay slot) can't put mulscc \ here because N and V flags \ must not be changed. \ %l0 tos tos add \ if so add original y to \ result to adjust for \ signed multiply. \ then %l1 rdy \ get lsd then %l1 sp 0 /n* nput \ x[lo]-(UTOS) <- lsd=l1 (delay slot) c; : * ( n1 n2 -- n3 ) um* drop ;