blob: 8d197e412ab454a27105887399b30b48fc10cb15 [file] [log] [blame]
/* hppa sub_n -- Sub two limb vectors of the same length > 0 and store
* sum in a third limb vector.
* Copyright (C) 1992, 1994, 1998,
* 2001, 2002 Free Software Foundation, Inc.
* This file is part of Libgcrypt.
* Libgcrypt is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2.1 of
* the License, or (at your option) any later version.
* Libgcrypt is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* GNU Lesser General Public License for more details.
* You should have received a copy of the GNU Lesser General Public
* License along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
* Note: This code is heavily based on the GNU MP Library.
* Actually it's the same code with only minor changes in the
* way the data is stored; this is to support the abstraction
* of an optional secure memory allocation which may be used
* to avoid revealing of sensitive data due to paging etc.
#include "sysdep.h"
#include "asm-syntax.h"
* mpi_limb_t
* _gcry_mpih_sub_n( mpi_ptr_t res_ptr, (gr26)
* mpi_ptr_t s1_ptr, (gr25)
* mpi_ptr_t s2_ptr, (gr24)
* mpi_size_t 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.
.export _gcry_mpih_sub_n
.label _gcry_mpih_sub_n
.callinfo frame=0,no_calls
ldws,ma 4(0,%r25),%r20
ldws,ma 4(0,%r24),%r19
addib,= -1,%r23,L$end ; check for (SIZE == 1)
sub %r20,%r19,%r28 ; subtract first limbs ignoring cy
.label L$loop
ldws,ma 4(0,%r25),%r20
ldws,ma 4(0,%r24),%r19
stws,ma %r28,4(0,%r26)
addib,<> -1,%r23,L$loop
subb %r20,%r19,%r28
.label L$end
stws %r28,0(0,%r26)
addc %r0,%r0,%r28
bv 0(%r2)
subi 1,%r28,%r28