-; HP-PA-1.1 __mpn_addmul_1 -- Multiply a limb vector with a limb and
-; add the result to a second limb vector.
+;! HP-PA-1.1 __mpn_addmul_1 -- Multiply a limb vector with a limb and
+;! add the result to a second limb vector.
-; Copyright (C) 1992, 1993, 1994 Free Software Foundation, Inc.
+;! Copyright (C) 1992, 1993, 1994 Free Software Foundation, Inc.
-; This file is part of the GNU MP Library.
+;! This file is part of the GNU MP Library.
-; The GNU MP Library is free software; you can redistribute it and/or modify
-; it under the terms of the GNU Library General Public License as published by
-; the Free Software Foundation; either version 2 of the License, or (at your
-; option) any later version.
+;! The GNU MP Library is free software; you can redistribute it and/or modify
+;! it under the terms of the GNU Library General Public License as published by
+;! the Free Software Foundation; either version 2 of the License, or (at your
+;! option) any later version.
-; The GNU MP Library 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 Library General Public
-; License for more details.
+;! The GNU MP Library 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 Library General Public
+;! License for more details.
-; You should have received a copy of the GNU Library General Public License
-; along with the GNU MP Library; see the file COPYING.LIB. If not, write to
-; the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
-; MA 02111-1307, USA.
+;! You should have received a copy of the GNU Library General Public License
+;! along with the GNU MP Library; see the file COPYING.LIB. If not, write to
+;! the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
+;! MA 02111-1307, USA.
-; INPUT PARAMETERS
-; res_ptr r26
-; s1_ptr r25
-; size r24
-; s2_limb r23
+;! INPUT PARAMETERS
+;! res_ptr r26
+;! s1_ptr r25
+;! size r24
+;! s2_limb r23
-; This runs at 11 cycles/limb on a PA7000. With the used instructions, it
-; can not become faster due to data cache contention after a store. On the
-; PA7100 it runs at 10 cycles/limb, and that can not be improved either,
-; since only the xmpyu does not need the integer pipeline, so the only
-; dual-issue we will get are addc+xmpyu. Unrolling could gain a cycle/limb
-; on the PA7100.
+;! This runs at 11 cycles/limb on a PA7000. With the used instructions, it
+;! can not become faster due to data cache contention after a store. On the
+;! PA7100 it runs at 10 cycles/limb, and that can not be improved either,
+;! since only the xmpyu does not need the integer pipeline, so the only
+;! dual-issue we will get are addc+xmpyu. Unrolling could gain a cycle/limb
+;! on the PA7100.
-; There are some ideas described in mul_1.s that applies to this code too.
+;! There are some ideas described in mul_1.s that applies to this code too.
- .code
+ .text
.export __mpn_addmul_1
-__mpn_addmul_1
+__mpn_addmul_1:
.proc
.callinfo frame=64,no_calls
.entry
ldo 64(%r30),%r30
fldws,ma 4(%r25),%fr5
- stw %r23,-16(%r30) ; move s2_limb ...
+ stw %r23,-16(%r30) ;! move s2_limb ...
addib,= -1,%r24,L$just_one_limb
- fldws -16(%r30),%fr4 ; ... into fr4
- add %r0,%r0,%r0 ; clear carry
+ fldws -16(%r30),%fr4 ;! ... into fr4
+ add %r0,%r0,%r0 ;! clear carry
xmpyu %fr4,%fr5,%fr6
fldws,ma 4(%r25),%fr7
fstds %fr6,-16(%r30)
xmpyu %fr4,%fr7,%fr8
- ldw -12(%r30),%r19 ; least significant limb in product
+ ldw -12(%r30),%r20 ;! least significant limb in product
ldw -16(%r30),%r28
fstds %fr8,-16(%r30)
addib,= -1,%r24,L$end
ldw -12(%r30),%r1
-; Main loop
-L$loop ldws 0(%r26),%r29
+;! Main loop
+L$loop:
+ ldws 0(%r26),%r29
fldws,ma 4(%r25),%fr5
- add %r29,%r19,%r19
- stws,ma %r19,4(%r26)
- addc %r28,%r1,%r19
+ add %r29,%r20,%r20
+ stws,ma %r20,4(%r26)
+ addc %r28,%r1,%r20
xmpyu %fr4,%fr5,%fr6
ldw -16(%r30),%r28
fstds %fr6,-16(%r30)
addib,<> -1,%r24,L$loop
ldw -12(%r30),%r1
-L$end ldw 0(%r26),%r29
- add %r29,%r19,%r19
- stws,ma %r19,4(%r26)
- addc %r28,%r1,%r19
+L$end:
+ ldw 0(%r26),%r29
+ add %r29,%r20,%r20
+ stws,ma %r20,4(%r26)
+ addc %r28,%r1,%r20
ldw -16(%r30),%r28
ldws 0(%r26),%r29
addc %r0,%r28,%r28
- add %r29,%r19,%r19
- stws,ma %r19,4(%r26)
+ add %r29,%r20,%r20
+ stws,ma %r20,4(%r26)
addc %r0,%r28,%r28
bv 0(%r2)
- ldo -64(%r30),%r30
+ ldo -64(%r30),%r30
-L$just_one_limb
+L$just_one_limb:
xmpyu %fr4,%fr5,%fr6
ldw 0(%r26),%r29
fstds %fr6,-16(%r30)
ldw -12(%r30),%r1
ldw -16(%r30),%r28
- add %r29,%r1,%r19
- stw %r19,0(%r26)
+ add %r29,%r1,%r20
+ stw %r20,0(%r26)
addc %r0,%r28,%r28
bv 0(%r2)
ldo -64(%r30),%r30
-; HP-PA-1.1 __mpn_mul_1 -- Multiply a limb vector with a limb and store
-; the result in a second limb vector.
+;! HP-PA-1.1 __mpn_mul_1 -- Multiply a limb vector with a limb and store
+;! the result in a second limb vector.
-; Copyright (C) 1992, 1993, 1994 Free Software Foundation, Inc.
+;! Copyright (C) 1992, 1993, 1994 Free Software Foundation, Inc.
-; This file is part of the GNU MP Library.
+;! This file is part of the GNU MP Library.
-; The GNU MP Library is free software; you can redistribute it and/or modify
-; it under the terms of the GNU Library General Public License as published by
-; the Free Software Foundation; either version 2 of the License, or (at your
-; option) any later version.
+;! The GNU MP Library is free software; you can redistribute it and/or modify
+;! it under the terms of the GNU Library General Public License as published by
+;! the Free Software Foundation; either version 2 of the License, or (at your
+;! option) any later version.
-; The GNU MP Library 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 Library General Public
-; License for more details.
+;! The GNU MP Library 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 Library General Public
+;! License for more details.
-; You should have received a copy of the GNU Library General Public License
-; along with the GNU MP Library; see the file COPYING.LIB. If not, write to
-; the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
-; MA 02111-1307, USA.
+;! You should have received a copy of the GNU Library General Public License
+;! along with the GNU MP Library; see the file COPYING.LIB. If not, write to
+;! the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
+;! MA 02111-1307, USA.
-; INPUT PARAMETERS
-; res_ptr r26
-; s1_ptr r25
-; size r24
-; s2_limb r23
+;! INPUT PARAMETERS
+;! res_ptr r26
+;! s1_ptr r25
+;! size r24
+;! s2_limb r23
-; This runs at 9 cycles/limb on a PA7000. With the used instructions, it can
-; not become faster due to data cache contention after a store. On the
-; PA7100 it runs at 7 cycles/limb, and that can not be improved either, since
-; only the xmpyu does not need the integer pipeline, so the only dual-issue
-; we will get are addc+xmpyu. Unrolling would not help either CPU.
+;! This runs at 9 cycles/limb on a PA7000. With the used instructions, it can
+;! not become faster due to data cache contention after a store. On the
+;! PA7100 it runs at 7 cycles/limb, and that can not be improved either, since
+;! only the xmpyu does not need the integer pipeline, so the only dual-issue
+;! we will get are addc+xmpyu. Unrolling would not help either CPU.
-; We could use fldds to read two limbs at a time from the S1 array, and that
-; could bring down the times to 8.5 and 6.5 cycles/limb for the PA7000 and
-; PA7100, respectively. We don't do that since it does not seem worth the
-; (alignment) troubles...
+;! We could use fldds to read two limbs at a time from the S1 array, and that
+;! could bring down the times to 8.5 and 6.5 cycles/limb for the PA7000 and
+;! PA7100, respectively. We don't do that since it does not seem worth the
+;! (alignment) troubles...
-; At least the PA7100 is rumored to be able to deal with cache-misses
-; without stalling instruction issue. If this is true, and the cache is
-; actually also lockup-free, we should use a deeper software pipeline, and
-; load from S1 very early! (The loads and stores to -12(sp) will surely be
-; in the cache.)
+;! At least the PA7100 is rumored to be able to deal with cache-misses
+;! without stalling instruction issue. If this is true, and the cache is
+;! actually also lockup-free, we should use a deeper software pipeline, and
+;! load from S1 very early; (The loads and stores to -12(sp) will surely be
+;! in the cache.)
- .code
+ .text
.export __mpn_mul_1
-__mpn_mul_1
+__mpn_mul_1:
.proc
.callinfo frame=64,no_calls
.entry
ldo 64(%r30),%r30
fldws,ma 4(%r25),%fr5
- stw %r23,-16(%r30) ; move s2_limb ...
+ stw %r23,-16(%r30) ;! move s2_limb ...
addib,= -1,%r24,L$just_one_limb
- fldws -16(%r30),%fr4 ; ... into fr4
- add %r0,%r0,%r0 ; clear carry
+ fldws -16(%r30),%fr4 ;! ... into fr4
+ add %r0,%r0,%r0 ;! clear carry
xmpyu %fr4,%fr5,%fr6
fldws,ma 4(%r25),%fr7
fstds %fr6,-16(%r30)
xmpyu %fr4,%fr7,%fr8
- ldw -12(%r30),%r19 ; least significant limb in product
+ ldw -12(%r30),%r20 ;! least significant limb in product
ldw -16(%r30),%r28
fstds %fr8,-16(%r30)
addib,= -1,%r24,L$end
ldw -12(%r30),%r1
-; Main loop
-L$loop fldws,ma 4(%r25),%fr5
- stws,ma %r19,4(%r26)
- addc %r28,%r1,%r19
+;! Main loop
+L$loop:
+ fldws,ma 4(%r25),%fr5
+ stws,ma %r20,4(%r26)
+ addc %r28,%r1,%r20
xmpyu %fr4,%fr5,%fr6
ldw -16(%r30),%r28
fstds %fr6,-16(%r30)
addib,<> -1,%r24,L$loop
ldw -12(%r30),%r1
-L$end stws,ma %r19,4(%r26)
- addc %r28,%r1,%r19
+L$end:
+ stws,ma %r20,4(%r26)
+ addc %r28,%r1,%r20
ldw -16(%r30),%r28
- stws,ma %r19,4(%r26)
+ stws,ma %r20,4(%r26)
addc %r0,%r28,%r28
bv 0(%r2)
ldo -64(%r30),%r30
-L$just_one_limb
+L$just_one_limb:
xmpyu %fr4,%fr5,%fr6
fstds %fr6,-16(%r30)
ldw -16(%r30),%r28
-; HP-PA-1.1 __mpn_submul_1 -- Multiply a limb vector with a limb and
-; subtract the result from a second limb vector.
+;! HP-PA-1.1 __mpn_submul_1 -- Multiply a limb vector with a limb and
+;! subtract the result from a second limb vector.
-; Copyright (C) 1992, 1993, 1994 Free Software Foundation, Inc.
+;! Copyright (C) 1992, 1993, 1994 Free Software Foundation, Inc.
-; This file is part of the GNU MP Library.
+;! This file is part of the GNU MP Library.
-; The GNU MP Library is free software; you can redistribute it and/or modify
-; it under the terms of the GNU Library General Public License as published by
-; the Free Software Foundation; either version 2 of the License, or (at your
-; option) any later version.
+;! The GNU MP Library is free software; you can redistribute it and/or modify
+;! it under the terms of the GNU Library General Public License as published by
+;! the Free Software Foundation; either version 2 of the License, or (at your
+;! option) any later version.
-; The GNU MP Library 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 Library General Public
-; License for more details.
+;! The GNU MP Library 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 Library General Public
+;! License for more details.
-; You should have received a copy of the GNU Library General Public License
-; along with the GNU MP Library; see the file COPYING.LIB. If not, write to
-; the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
-; MA 02111-1307, USA.
+;! You should have received a copy of the GNU Library General Public License
+;! along with the GNU MP Library; see the file COPYING.LIB. If not, write to
+;! the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
+;! MA 02111-1307, USA.
-; INPUT PARAMETERS
-; res_ptr r26
-; s1_ptr r25
-; size r24
-; s2_limb r23
+;! INPUT PARAMETERS
+;! res_ptr r26
+;! s1_ptr r25
+;! size r24
+;! s2_limb r23
-; This runs at 12 cycles/limb on a PA7000. With the used instructions, it
-; can not become faster due to data cache contention after a store. On the
-; PA7100 it runs at 11 cycles/limb, and that can not be improved either,
-; since only the xmpyu does not need the integer pipeline, so the only
-; dual-issue we will get are addc+xmpyu. Unrolling could gain a cycle/limb
-; on the PA7100.
+;! This runs at 12 cycles/limb on a PA7000. With the used instructions, it
+;! can not become faster due to data cache contention after a store. On the
+;! PA7100 it runs at 11 cycles/limb, and that can not be improved either,
+;! since only the xmpyu does not need the integer pipeline, so the only
+;! dual-issue we will get are addc+xmpyu. Unrolling could gain a cycle/limb
+;! on the PA7100.
-; There are some ideas described in mul_1.s that applies to this code too.
+;! There are some ideas described in mul_1.s that applies to this code too.
-; It seems possible to make this run as fast as __mpn_addmul_1, if we use
-; sub,>>= %r29,%r19,%r22
-; addi 1,%r28,%r28
-; but that requires reworking the hairy software pipeline...
+;! It seems possible to make this run as fast as __mpn_addmul_1, if we use
+;! sub,>>= %r29,%r20,%r22
+;! addi 1,%r28,%r28
+;! but that requires reworking the hairy software pipeline...
- .code
+ .text
.export __mpn_submul_1
-__mpn_submul_1
+__mpn_submul_1:
.proc
.callinfo frame=64,no_calls
.entry
ldo 64(%r30),%r30
fldws,ma 4(%r25),%fr5
- stw %r23,-16(%r30) ; move s2_limb ...
+ stw %r23,-16(%r30) ;! move s2_limb ...
addib,= -1,%r24,L$just_one_limb
- fldws -16(%r30),%fr4 ; ... into fr4
- add %r0,%r0,%r0 ; clear carry
+ fldws -16(%r30),%fr4 ;! ... into fr4
+ add %r0,%r0,%r0 ;! clear carry
xmpyu %fr4,%fr5,%fr6
fldws,ma 4(%r25),%fr7
fstds %fr6,-16(%r30)
xmpyu %fr4,%fr7,%fr8
- ldw -12(%r30),%r19 ; least significant limb in product
+ ldw -12(%r30),%r20 ;! least significant limb in product
ldw -16(%r30),%r28
fstds %fr8,-16(%r30)
addib,= -1,%r24,L$end
ldw -12(%r30),%r1
-; Main loop
-L$loop ldws 0(%r26),%r29
+;! Main loop
+L$loop:
+ ldws 0(%r26),%r29
fldws,ma 4(%r25),%fr5
- sub %r29,%r19,%r22
- add %r22,%r19,%r0
+ sub %r29,%r20,%r22
+ add %r22,%r20,%r0
stws,ma %r22,4(%r26)
- addc %r28,%r1,%r19
+ addc %r28,%r1,%r20
xmpyu %fr4,%fr5,%fr6
ldw -16(%r30),%r28
fstds %fr6,-16(%r30)
addib,<> -1,%r24,L$loop
ldw -12(%r30),%r1
-L$end ldw 0(%r26),%r29
- sub %r29,%r19,%r22
- add %r22,%r19,%r0
+L$end:
+ ldw 0(%r26),%r29
+ sub %r29,%r20,%r22
+ add %r22,%r20,%r0
stws,ma %r22,4(%r26)
- addc %r28,%r1,%r19
+ addc %r28,%r1,%r20
ldw -16(%r30),%r28
ldws 0(%r26),%r29
addc %r0,%r28,%r28
- sub %r29,%r19,%r22
- add %r22,%r19,%r0
+ sub %r29,%r20,%r22
+ add %r22,%r20,%r0
stws,ma %r22,4(%r26)
addc %r0,%r28,%r28
bv 0(%r2)
ldo -64(%r30),%r30
-L$just_one_limb
+L$just_one_limb:
xmpyu %fr4,%fr5,%fr6
ldw 0(%r26),%r29
fstds %fr6,-16(%r30)
-; HP-PA __udiv_qrnnd division support, used from longlong.h.
-; This version runs fast on PA 7000 and later.
+;! HP-PA __udiv_qrnnd division support, used from longlong.h.
+;! This version runs fast on PA 7000 and later.
-; Copyright (C) 1993, 1994 Free Software Foundation, Inc.
+;! Copyright (C) 1993, 1994 Free Software Foundation, Inc.
-; This file is part of the GNU MP Library.
+;! This file is part of the GNU MP Library.
-; The GNU MP Library is free software; you can redistribute it and/or modify
-; it under the terms of the GNU Library General Public License as published by
-; the Free Software Foundation; either version 2 of the License, or (at your
-; option) any later version.
+;! The GNU MP Library is free software; you can redistribute it and/or modify
+;! it under the terms of the GNU Library General Public License as published by
+;! the Free Software Foundation; either version 2 of the License, or (at your
+;! option) any later version.
-; The GNU MP Library 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 Library General Public
-; License for more details.
+;! The GNU MP Library 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 Library General Public
+;! License for more details.
-; You should have received a copy of the GNU Library General Public License
-; along with the GNU MP Library; see the file COPYING.LIB. If not, write to
-; the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
-; MA 02111-1307, USA.
+;! You should have received a copy of the GNU Library General Public License
+;! along with the GNU MP Library; see the file COPYING.LIB. If not, write to
+;! the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
+;! MA 02111-1307, USA.
-; INPUT PARAMETERS
-; rem_ptr gr26
-; n1 gr25
-; n0 gr24
-; d gr23
+;! INPUT PARAMETERS
+;! rem_ptr gr26
+;! n1 gr25
+;! n0 gr24
+;! d gr23
- .code
-L$0000 .word 0x43f00000
+ .text
+L$0000:
+ .word 0x43f00000
.word 0x0
.export __udiv_qrnnd
-__udiv_qrnnd
+__udiv_qrnnd:
.proc
.callinfo frame=64,no_calls
.entry
ldo 64(%r30),%r30
- stws %r25,-16(0,%r30) ; n_hi
- stws %r24,-12(0,%r30) ; n_lo
- ldil L'L$0000,%r19
- ldo R'L$0000(%r19),%r19
+ stws %r25,-16(0,%r30) ;! n_hi
+ stws %r24,-12(0,%r30) ;! n_lo
+ b,l L$0,%r1
+ ldo L$0000-L$0(%r1),%r1
+L$0:
fldds -16(0,%r30),%fr5
stws %r23,-12(0,%r30)
comib,<= 0,%r25,L$1
fcnvxf,dbl,dbl %fr5,%fr5
- fldds 0(0,%r19),%fr4
+ fldds 0(0,%r1),%fr4
fadd,dbl %fr4,%fr5,%fr5
-L$1
+L$1:
fcpy,sgl %fr0,%fr6L
fldws -12(0,%r30),%fr6R
fcnvxf,dbl,dbl %fr6,%fr4
ldws -12(0,%r30),%r21
ldws -16(0,%r30),%r20
sub %r24,%r21,%r22
- subb %r25,%r20,%r19
- comib,= 0,%r19,L$2
+ subb %r25,%r20,%r1
+ comib,= 0,%r1,L$2
ldo -64(%r30),%r30
add %r22,%r23,%r22
ldo -1(%r28),%r28
-L$2 bv 0(%r2)
+L$2:
+ bv 0(%r2)
stws %r22,0(0,%r26)
.exit
-; HP-PA __mpn_lshift --
+;! HP-PA __mpn_lshift --
-; Copyright (C) 1992, 1994 Free Software Foundation, Inc.
+;! Copyright (C) 1992, 1994 Free Software Foundation, Inc.
-; This file is part of the GNU MP Library.
+;! This file is part of the GNU MP Library.
-; The GNU MP Library is free software; you can redistribute it and/or modify
-; it under the terms of the GNU Library General Public License as published by
-; the Free Software Foundation; either version 2 of the License, or (at your
-; option) any later version.
+;! The GNU MP Library is free software; you can redistribute it and/or modify
+;! it under the terms of the GNU Library General Public License as published by
+;! the Free Software Foundation; either version 2 of the License, or (at your
+;! option) any later version.
-; The GNU MP Library 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 Library General Public
-; License for more details.
+;! The GNU MP Library 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 Library General Public
+;! License for more details.
-; You should have received a copy of the GNU Library General Public License
-; along with the GNU MP Library; see the file COPYING.LIB. If not, write to
-; the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
-; MA 02111-1307, USA.
+;! You should have received a copy of the GNU Library General Public License
+;! along with the GNU MP Library; see the file COPYING.LIB. If not, write to
+;! the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
+;! MA 02111-1307, USA.
-; INPUT PARAMETERS
-; res_ptr gr26
-; s_ptr gr25
-; size gr24
-; cnt gr23
+;! INPUT PARAMETERS
+;! res_ptr gr26
+;! s_ptr gr25
+;! size gr24
+;! cnt gr23
- .code
+ .text
.export __mpn_lshift
-__mpn_lshift
+__mpn_lshift:
.proc
.callinfo frame=64,no_calls
.entry
subi 32,%r23,%r1
mtsar %r1
addib,= -1,%r24,L$0004
- vshd %r0,%r22,%r28 ; compute carry out limb
+ vshd %r0,%r22,%r28 ;! compute carry out limb
ldws,mb -4(0,%r25),%r29
addib,= -1,%r24,L$0002
vshd %r22,%r29,%r20
-L$loop ldws,mb -4(0,%r25),%r22
+L$loop: ldws,mb -4(0,%r25),%r22
stws,mb %r20,-4(0,%r26)
addib,= -1,%r24,L$0003
vshd %r29,%r22,%r20
addib,<> -1,%r24,L$loop
vshd %r22,%r29,%r20
-L$0002 stws,mb %r20,-4(0,%r26)
+L$0002: stws,mb %r20,-4(0,%r26)
vshd %r29,%r0,%r20
bv 0(%r2)
stw %r20,-4(0,%r26)
-L$0003 stws,mb %r20,-4(0,%r26)
-L$0004 vshd %r22,%r0,%r20
+L$0003: stws,mb %r20,-4(0,%r26)
+L$0004: vshd %r22,%r0,%r20
bv 0(%r2)
stw %r20,-4(0,%r26)
-; HP-PA __mpn_rshift --
+;! HP-PA __mpn_rshift --
-; Copyright (C) 1992, 1994 Free Software Foundation, Inc.
+;! Copyright (C) 1992, 1994 Free Software Foundation, Inc.
-; This file is part of the GNU MP Library.
+;! This file is part of the GNU MP Library.
-; The GNU MP Library is free software; you can redistribute it and/or modify
-; it under the terms of the GNU Library General Public License as published by
-; the Free Software Foundation; either version 2 of the License, or (at your
-; option) any later version.
+;! The GNU MP Library is free software; you can redistribute it and/or modify
+;! it under the terms of the GNU Library General Public License as published by
+;! the Free Software Foundation; either version 2 of the License, or (at your
+;! option) any later version.
-; The GNU MP Library 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 Library General Public
-; License for more details.
+;! The GNU MP Library 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 Library General Public
+;! License for more details.
-; You should have received a copy of the GNU Library General Public License
-; along with the GNU MP Library; see the file COPYING.LIB. If not, write to
-; the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
-; MA 02111-1307, USA.
+;! You should have received a copy of the GNU Library General Public License
+;! along with the GNU MP Library; see the file COPYING.LIB. If not, write to
+;! the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
+;! MA 02111-1307, USA.
-; INPUT PARAMETERS
-; res_ptr gr26
-; s_ptr gr25
-; size gr24
-; cnt gr23
+;! INPUT PARAMETERS
+;! res_ptr gr26
+;! s_ptr gr25
+;! size gr24
+;! cnt gr23
- .code
+ .text
.export __mpn_rshift
-__mpn_rshift
+__mpn_rshift:
.proc
.callinfo frame=64,no_calls
.entry
ldws,ma 4(0,%r25),%r22
mtsar %r23
addib,= -1,%r24,L$0004
- vshd %r22,%r0,%r28 ; compute carry out limb
+ vshd %r22,%r0,%r28 ;! compute carry out limb
ldws,ma 4(0,%r25),%r29
addib,= -1,%r24,L$0002
vshd %r29,%r22,%r20
-L$loop ldws,ma 4(0,%r25),%r22
+L$loop: ldws,ma 4(0,%r25),%r22
stws,ma %r20,4(0,%r26)
addib,= -1,%r24,L$0003
vshd %r22,%r29,%r20
addib,<> -1,%r24,L$loop
vshd %r29,%r22,%r20
-L$0002 stws,ma %r20,4(0,%r26)
+L$0002: stws,ma %r20,4(0,%r26)
vshd %r0,%r29,%r20
bv 0(%r2)
stw %r20,0(0,%r26)
-L$0003 stws,ma %r20,4(0,%r26)
-L$0004 vshd %r0,%r22,%r20
+L$0003: stws,ma %r20,4(0,%r26)
+L$0004: vshd %r0,%r22,%r20
bv 0(%r2)
stw %r20,0(0,%r26)
-; HP-PA __mpn_sub_n -- Subtract two limb vectors of the same length > 0 and
-; store difference in a third limb vector.
+;! HP-PA __mpn_sub_n -- Subtract two limb vectors of the same length > 0 and
+;! store difference in a third limb vector.
-; Copyright (C) 1992, 1994 Free Software Foundation, Inc.
+;! Copyright (C) 1992, 1994 Free Software Foundation, Inc.
-; This file is part of the GNU MP Library.
+;! This file is part of the GNU MP Library.
-; The GNU MP Library is free software; you can redistribute it and/or modify
-; it under the terms of the GNU Library General Public License as published by
-; the Free Software Foundation; either version 2 of the License, or (at your
-; option) any later version.
+;! The GNU MP Library is free software; you can redistribute it and/or modify
+;! it under the terms of the GNU Library General Public License as published by
+;! the Free Software Foundation; either version 2 of the License, or (at your
+;! option) any later version.
-; The GNU MP Library 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 Library General Public
-; License for more details.
+;! The GNU MP Library 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 Library General Public
+;! License for more details.
-; You should have received a copy of the GNU Library General Public License
-; along with the GNU MP Library; see the file COPYING.LIB. If not, write to
-; the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
-; MA 02111-1307, USA.
+;! You should have received a copy of the GNU Library General Public License
+;! along with the GNU MP Library; see the file COPYING.LIB. If not, write to
+;! the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
+;! MA 02111-1307, USA.
-; INPUT PARAMETERS
-; res_ptr gr26
-; s1_ptr gr25
-; s2_ptr gr24
-; size gr23
+;! INPUT PARAMETERS
+;! res_ptr gr26
+;! s1_ptr gr25
+;! s2_ptr gr24
+;! size gr23
-; One might want to unroll this as for other processors, but it turns
-; out that the data cache contention after a store makes such
-; unrolling useless. We can't come under 5 cycles/limb anyway.
+;! One might want to unroll this as for other processors, but it turns
+;! out that the data cache contention after a store makes such
+;! unrolling useless. We can't come under 5 cycles/limb anyway.
- .code
+ .text
.export __mpn_sub_n
-__mpn_sub_n
+__mpn_sub_n:
.proc
.callinfo frame=0,no_calls
.entry
- ldws,ma 4(0,%r25),%r20
- ldws,ma 4(0,%r24),%r19
+ ldws,ma 4(0,%r25),%r21
+ ldws,ma 4(0,%r24),%r20
- addib,= -1,%r23,L$end ; check for (SIZE == 1)
- sub %r20,%r19,%r28 ; subtract first limbs ignoring cy
+ addib,= -1,%r23,L$end ;! check for (SIZE == 1)
+ sub %r21,%r20,%r28 ;! subtract first limbs ignoring cy
-L$loop ldws,ma 4(0,%r25),%r20
- ldws,ma 4(0,%r24),%r19
+L$loop: ldws,ma 4(0,%r25),%r21
+ ldws,ma 4(0,%r24),%r20
stws,ma %r28,4(0,%r26)
addib,<> -1,%r23,L$loop
- subb %r20,%r19,%r28
+ subb %r21,%r20,%r28
-L$end stws %r28,0(0,%r26)
+L$end: stws %r28,0(0,%r26)
addc %r0,%r0,%r28
bv 0(%r2)
subi 1,%r28,%r28
-; HP-PA __udiv_qrnnd division support, used from longlong.h.
-; This version runs fast on pre-PA7000 CPUs.
+;! HP-PA __udiv_qrnnd division support, used from longlong.h.
+;! This version runs fast on pre-PA7000 CPUs.
-; Copyright (C) 1993, 1994 Free Software Foundation, Inc.
+;! Copyright (C) 1993, 1994 Free Software Foundation, Inc.
-; This file is part of the GNU MP Library.
+;! This file is part of the GNU MP Library.
-; The GNU MP Library is free software; you can redistribute it and/or modify
-; it under the terms of the GNU Library General Public License as published by
-; the Free Software Foundation; either version 2 of the License, or (at your
-; option) any later version.
+;! The GNU MP Library is free software; you can redistribute it and/or modify
+;! it under the terms of the GNU Library General Public License as published by
+;! the Free Software Foundation; either version 2 of the License, or (at your
+;! option) any later version.
-; The GNU MP Library 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 Library General Public
-; License for more details.
+;! The GNU MP Library 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 Library General Public
+;! License for more details.
-; You should have received a copy of the GNU Library General Public License
-; along with the GNU MP Library; see the file COPYING.LIB. If not, write to
-; the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
-; MA 02111-1307, USA.
+;! You should have received a copy of the GNU Library General Public License
+;! along with the GNU MP Library; see the file COPYING.LIB. If not, write to
+;! the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
+;! MA 02111-1307, USA.
-; INPUT PARAMETERS
-; rem_ptr gr26
-; n1 gr25
-; n0 gr24
-; d gr23
+;! INPUT PARAMETERS
+;! rem_ptr gr26
+;! n1 gr25
+;! n0 gr24
+;! d gr23
-; The code size is a bit excessive. We could merge the last two ds;addc
-; sequences by simply moving the "bb,< Odd" instruction down. The only
-; trouble is the FFFFFFFF code that would need some hacking.
+;! The code size is a bit excessive. We could merge the last two ds;addc
+;! sequences by simply moving the "bb,< Odd" instruction down. The only
+;! trouble is the FFFFFFFF code that would need some hacking.
- .code
+ .text
.export __udiv_qrnnd
-__udiv_qrnnd
+__udiv_qrnnd:
.proc
.callinfo frame=0,no_calls
.entry
comb,< %r23,0,L$largedivisor
- sub %r0,%r23,%r1 ; clear cy as side-effect
+ sub %r0,%r23,%r1 ;! clear cy as side-effect
ds %r0,%r1,%r0
addc %r24,%r24,%r24
ds %r25,%r23,%r25
bv 0(%r2)
addc %r28,%r28,%r28
-L$largedivisor
- extru %r24,31,1,%r19 ; r19 = n0 & 1
+L$largedivisor:
+ extru %r24,31,1,%r20 ;! r20 = n0 & 1
bb,< %r23,31,L$odd
- extru %r23,30,31,%r22 ; r22 = d >> 1
- shd %r25,%r24,1,%r24 ; r24 = new n0
- extru %r25,30,31,%r25 ; r25 = new n1
+ extru %r23,30,31,%r22 ;! r22 = d >> 1
+ shd %r25,%r24,1,%r24 ;! r24 = new n0
+ extru %r25,30,31,%r25 ;! r25 = new n1
sub %r0,%r22,%r21
ds %r0,%r21,%r0
addc %r24,%r24,%r24
ds %r25,%r22,%r25
comclr,>= %r25,%r0,%r0
addl %r25,%r22,%r25
- sh1addl %r25,%r19,%r25
+ sh1addl %r25,%r20,%r25
stws %r25,0(0,%r26)
bv 0(%r2)
addc %r24,%r24,%r28
-L$odd addib,sv,n 1,%r22,L$FF.. ; r22 = (d / 2 + 1)
- shd %r25,%r24,1,%r24 ; r24 = new n0
- extru %r25,30,31,%r25 ; r25 = new n1
+L$odd: addib,sv,n 1,%r22,L$FF.. ;! r22 = (d / 2 + 1)
+ shd %r25,%r24,1,%r24 ;! r24 = new n0
+ extru %r25,30,31,%r25 ;! r25 = new n1
sub %r0,%r22,%r21
ds %r0,%r21,%r0
addc %r24,%r24,%r24
addc %r24,%r24,%r28
comclr,>= %r25,%r0,%r0
addl %r25,%r22,%r25
- sh1addl %r25,%r19,%r25
-; We have computed (n1,,n0) / (d + 1), q' = r28, r' = r25
+ sh1addl %r25,%r20,%r25
+;! We have computed (n1,,n0) / (d + 1), q' = r28, r' = r25
add,nuv %r28,%r25,%r25
addl %r25,%r1,%r25
addc %r0,%r28,%r28
bv 0(%r2)
addc %r0,%r28,%r28
-; This is just a special case of the code above.
-; We come here when d == 0xFFFFFFFF
-L$FF.. add,uv %r25,%r24,%r24
+;! This is just a special case of the code above.
+;! We come here when d == 0xFFFFFFFF
+L$FF..: add,uv %r25,%r24,%r24
sub,<< %r24,%r23,%r0
ldo 1(%r24),%r24
stws %r24,0(0,%r26)
projects / thirdparty / glibc.git / commitdiff
