diff options
Diffstat (limited to 'sysdeps/i386/i586')
-rw-r--r-- | sysdeps/i386/i586/add_n.S | 54 | ||||
-rw-r--r-- | sysdeps/i386/i586/addmul_1.S | 34 | ||||
-rw-r--r-- | sysdeps/i386/i586/lshift.S | 54 | ||||
-rw-r--r-- | sysdeps/i386/i586/memset.S | 47 | ||||
-rw-r--r-- | sysdeps/i386/i586/mul_1.S | 34 | ||||
-rw-r--r-- | sysdeps/i386/i586/rshift.S | 82 | ||||
-rw-r--r-- | sysdeps/i386/i586/strchr.S | 89 | ||||
-rw-r--r-- | sysdeps/i386/i586/strlen.S | 45 | ||||
-rw-r--r-- | sysdeps/i386/i586/sub_n.S | 54 | ||||
-rw-r--r-- | sysdeps/i386/i586/submul_1.S | 34 |
10 files changed, 265 insertions, 262 deletions
diff --git a/sysdeps/i386/i586/add_n.S b/sysdeps/i386/i586/add_n.S index ac6f2819b2..6d86027b1f 100644 --- a/sysdeps/i386/i586/add_n.S +++ b/sysdeps/i386/i586/add_n.S @@ -1,24 +1,22 @@ /* Pentium __mpn_add_n -- Add two limb vectors of the same length > 0 and store sum in a third limb vector. + Copyright (C) 1992, 1994, 1995, 1996, 1997 Free Software Foundation, Inc. + This file is part of the GNU MP Library. -Copyright (C) 1992, 1994, 1995, 1996 Free Software Foundation, Inc. + 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. -This file is part of the GNU MP Library. + 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 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. - -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 @@ -52,14 +50,14 @@ C_SYMBOL_NAME(__mpn_add_n:) shrl $3,%ecx andl $7,%edx testl %ecx,%ecx /* zero carry flag */ - jz Lend + jz L(end) pushl %edx ALIGN (3) -Loop: movl 28(%edi),%eax /* fetch destination cache line */ +L(oop): movl 28(%edi),%eax /* fetch destination cache line */ leal 32(%edi),%edi -L1: movl (%esi),%eax +L(1): movl (%esi),%eax movl 4(%esi),%edx adcl %ebx,%eax movl 4(%ebp),%ebx @@ -68,7 +66,7 @@ L1: movl (%esi),%eax movl %eax,-32(%edi) movl %edx,-28(%edi) -L2: movl 8(%esi),%eax +L(2): movl 8(%esi),%eax movl 12(%esi),%edx adcl %ebx,%eax movl 12(%ebp),%ebx @@ -77,7 +75,7 @@ L2: movl 8(%esi),%eax movl %eax,-24(%edi) movl %edx,-20(%edi) -L3: movl 16(%esi),%eax +L(3): movl 16(%esi),%eax movl 20(%esi),%edx adcl %ebx,%eax movl 20(%ebp),%ebx @@ -86,7 +84,7 @@ L3: movl 16(%esi),%eax movl %eax,-16(%edi) movl %edx,-12(%edi) -L4: movl 24(%esi),%eax +L(4): movl 24(%esi),%eax movl 28(%esi),%edx adcl %ebx,%eax movl 28(%ebp),%ebx @@ -98,14 +96,14 @@ L4: movl 24(%esi),%eax leal 32(%esi),%esi leal 32(%ebp),%ebp decl %ecx - jnz Loop + jnz L(oop) popl %edx -Lend: +L(end): decl %edx /* test %edx w/o clobbering carry */ - js Lend2 + js L(end2) incl %edx -Loop2: +L(oop2): leal 4(%edi),%edi movl (%esi),%eax adcl %ebx,%eax @@ -114,8 +112,8 @@ Loop2: leal 4(%esi),%esi leal 4(%ebp),%ebp decl %edx - jnz Loop2 -Lend2: + jnz L(oop2) +L(end2): movl (%esi),%eax adcl %ebx,%eax movl %eax,(%edi) diff --git a/sysdeps/i386/i586/addmul_1.S b/sysdeps/i386/i586/addmul_1.S index 7cfa5db687..0d410f6ac3 100644 --- a/sysdeps/i386/i586/addmul_1.S +++ b/sysdeps/i386/i586/addmul_1.S @@ -1,24 +1,22 @@ /* Pentium __mpn_addmul_1 -- Multiply a limb vector with a limb and add the result to a second limb vector. + Copyright (C) 1992, 1994, 1996, 1997 Free Software Foundation, Inc. + This file is part of the GNU MP Library. -Copyright (C) 1992, 1994, 1996 Free Software Foundation, Inc. + 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. -This file is part of the GNU MP Library. + 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 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. - -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 @@ -57,7 +55,7 @@ C_SYMBOL_NAME(__mpn_addmul_1:) INSN2(xor,l ,R(ebx),R(ebx)) ALIGN (3) -Loop: INSN2(adc,l ,R(ebx),$0) +L(oop): INSN2(adc,l ,R(ebx),$0) INSN2(mov,l ,R(eax),MEM_INDEX(s1_ptr,size,4)) INSN1(mul,l ,R(s2_limb)) @@ -72,7 +70,7 @@ Loop: INSN2(adc,l ,R(ebx),$0) INSN1(inc,l ,R(size)) INSN2(mov,l ,R(ebx),R(edx)) - INSN1(jnz, ,Loop) + INSN1(jnz, ,L(oop)) INSN2(adc,l ,R(ebx),$0) INSN2(mov,l ,R(eax),R(ebx)) diff --git a/sysdeps/i386/i586/lshift.S b/sysdeps/i386/i586/lshift.S index bf9b223a0b..1d72fc901e 100644 --- a/sysdeps/i386/i586/lshift.S +++ b/sysdeps/i386/i586/lshift.S @@ -1,5 +1,5 @@ /* Pentium optimized __mpn_lshift -- - Copyright (C) 1992, 1994, 1995, 1996 Free Software Foundation, Inc. + Copyright (C) 1992, 1994, 1995, 1996, 1997 Free Software Foundation, Inc. This file is part of the GNU C Library. The GNU C Library is free software; you can redistribute it and/or modify @@ -44,15 +44,15 @@ C_SYMBOL_NAME(__mpn_lshift:) /* We can use faster code for shift-by-1 under certain conditions. */ cmp $1,%ecx - jne Lnormal + jne L(normal) leal 4(%esi),%eax cmpl %edi,%eax - jnc Lspecial /* jump if s_ptr + 1 >= res_ptr */ + jnc L(special) /* jump if s_ptr + 1 >= res_ptr */ leal (%esi,%ebp,4),%eax cmpl %eax,%edi - jnc Lspecial /* jump if res_ptr >= s_ptr + size */ + jnc L(special) /* jump if res_ptr >= s_ptr + size */ -Lnormal: +L(normal): leal -4(%edi,%ebp,4),%edi leal -4(%esi,%ebp,4),%esi @@ -65,12 +65,12 @@ Lnormal: decl %ebp pushl %ebp shrl $3,%ebp - jz Lend + jz L(end) movl (%edi),%eax /* fetch destination cache line */ ALIGN (2) -Loop: movl -28(%edi),%eax /* fetch destination cache line */ +L(oop): movl -28(%edi),%eax /* fetch destination cache line */ movl %edx,%ebx movl (%esi),%eax @@ -104,21 +104,23 @@ Loop: movl -28(%edi),%eax /* fetch destination cache line */ subl $32,%esi subl $32,%edi decl %ebp - jnz Loop + jnz L(oop) -Lend: popl %ebp +L(end): popl %ebp andl $7,%ebp - jz Lend2 -Loop2: movl (%esi),%eax + jz L(end2) +L(oop2): + movl (%esi),%eax shldl %cl,%eax,%edx movl %edx,(%edi) movl %eax,%edx subl $4,%esi subl $4,%edi decl %ebp - jnz Loop2 + jnz L(oop2) -Lend2: shll %cl,%edx /* compute least significant limb */ +L(end2): + shll %cl,%edx /* compute least significant limb */ movl %edx,(%edi) /* store it */ popl %eax /* pop carry limb */ @@ -134,7 +136,7 @@ Lend2: shll %cl,%edx /* compute least significant limb */ function is documented to work for overlapping source and destination. */ -Lspecial: +L(special): movl (%esi),%edx addl $4,%esi @@ -145,12 +147,13 @@ Lspecial: addl %edx,%edx incl %ebp decl %ebp - jz LLend + jz L(Lend) movl (%edi),%eax /* fetch destination cache line */ ALIGN (2) -LLoop: movl 28(%edi),%eax /* fetch destination cache line */ +L(Loop): + movl 28(%edi),%eax /* fetch destination cache line */ movl %edx,%ebx movl (%esi),%eax @@ -184,14 +187,16 @@ LLoop: movl 28(%edi),%eax /* fetch destination cache line */ leal 32(%esi),%esi /* use leal not to clobber carry */ leal 32(%edi),%edi decl %ebp - jnz LLoop + jnz L(Loop) -LLend: popl %ebp +L(Lend): + popl %ebp sbbl %eax,%eax /* save carry in %eax */ andl $7,%ebp - jz LLend2 + jz L(Lend2) addl %eax,%eax /* restore carry from eax */ -LLoop2: movl %edx,%ebx +L(Loop2): + movl %edx,%ebx movl (%esi),%edx adcl %edx,%edx movl %ebx,(%edi) @@ -199,11 +204,12 @@ LLoop2: movl %edx,%ebx leal 4(%esi),%esi /* use leal not to clobber carry */ leal 4(%edi),%edi decl %ebp - jnz LLoop2 + jnz L(Loop2) - jmp LL1 -LLend2: addl %eax,%eax /* restore carry from eax */ -LL1: movl %edx,(%edi) /* store last limb */ + jmp L(L1) +L(Lend2): + addl %eax,%eax /* restore carry from eax */ +L(L1): movl %edx,(%edi) /* store last limb */ sbbl %eax,%eax negl %eax diff --git a/sysdeps/i386/i586/memset.S b/sysdeps/i386/i586/memset.S index 1e8a9c2111..fb96a369b9 100644 --- a/sysdeps/i386/i586/memset.S +++ b/sysdeps/i386/i586/memset.S @@ -1,25 +1,26 @@ /* memset/bzero -- set memory area to CH/0 -Highly optimized version for ix85, x>=5. -Copyright (C) 1996 Free Software Foundation, Inc. -This file is part of the GNU C Library. -Contributed by Torbjorn Granlund, <tege@matematik.su.se> - -The GNU C 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 C 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 C 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. */ + Highly optimized version for ix85, x>=5. + Copyright (C) 1996, 1997 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Torbjorn Granlund, <tege@matematik.su.se> + + The GNU C 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 C 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 C 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. */ #include <sysdep.h> +#include "asm-syntax.h" /* INPUT PARAMETERS: @@ -52,7 +53,7 @@ ENTRY (memset) /* If less than 36 bytes to write, skip tricky code (it wouldn't work). */ cmpl $36, %edx movl %edx, %ecx /* needed when branch is taken! */ - jl L2 + jl L(2) /* First write 0-3 bytes to make the pointer 32-bit aligned. */ movl %edi, %ecx /* Copy ptr to ecx... */ @@ -66,7 +67,7 @@ ENTRY (memset) movl (%edi), %ecx /* Fetch destination cache line */ .align 2, 0x90 /* supply 0x90 for broken assemblers */ -L1: movl 28(%edi), %ecx /* allocate cache line for destination */ +L(1): movl 28(%edi), %ecx /* allocate cache line for destination */ subl $32, %edx /* decr loop count */ movl %eax, 0(%edi) /* store words pairwise */ movl %eax, 4(%edi) @@ -77,12 +78,12 @@ L1: movl 28(%edi), %ecx /* allocate cache line for destination */ movl %eax, 24(%edi) movl %eax, 28(%edi) leal 32(%edi), %edi /* update destination pointer */ - jge L1 + jge L(1) leal 32(%edx), %ecx /* reset offset count */ /* Write last 0-7 full 32-bit words (up to 8 words if loop was skipped). */ -L2: shrl $2, %ecx /* convert byte count to longword count */ +L(2): shrl $2, %ecx /* convert byte count to longword count */ rep stosl diff --git a/sysdeps/i386/i586/mul_1.S b/sysdeps/i386/i586/mul_1.S index 4ac3050a61..905e65cb58 100644 --- a/sysdeps/i386/i586/mul_1.S +++ b/sysdeps/i386/i586/mul_1.S @@ -1,24 +1,22 @@ /* Pentium __mpn_mul_1 -- Multiply a limb vector with a limb and store the result in a second limb vector. + Copyright (C) 1992, 1994, 1996, 1997 Free Software Foundation, Inc. + This file is part of the GNU MP Library. -Copyright (C) 1992, 1994, 1996 Free Software Foundation, Inc. + 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. -This file is part of the GNU MP Library. + 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 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. - -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 @@ -57,7 +55,7 @@ C_SYMBOL_NAME(__mpn_mul_1:) INSN2(xor,l ,R(ebx),R(ebx)) ALIGN (3) -Loop: INSN2(adc,l ,R(ebx),$0) +L(oop): INSN2(adc,l ,R(ebx),$0) INSN2(mov,l ,R(eax),MEM_INDEX(s1_ptr,size,4)) INSN1(mul,l ,R(s2_limb)) @@ -68,7 +66,7 @@ Loop: INSN2(adc,l ,R(ebx),$0) INSN1(inc,l ,R(size)) INSN2(mov,l ,R(ebx),R(edx)) - INSN1(jnz, ,Loop) + INSN1(jnz, ,L(oop)) INSN2(adc,l ,R(ebx),$0) INSN2(mov,l ,R(eax),R(ebx)) diff --git a/sysdeps/i386/i586/rshift.S b/sysdeps/i386/i586/rshift.S index 1860daf4d3..14ffbb8454 100644 --- a/sysdeps/i386/i586/rshift.S +++ b/sysdeps/i386/i586/rshift.S @@ -1,23 +1,21 @@ /* Pentium optimized __mpn_rshift -- + Copyright (C) 1992, 1994, 1995, 1996, 1997 Free Software Foundation, Inc. + This file is part of the GNU MP Library. -Copyright (C) 1992, 1994, 1995, 1996 Free Software Foundation, Inc. + 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. -This file is part of the GNU MP Library. + 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 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. - -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 @@ -46,15 +44,15 @@ C_SYMBOL_NAME(__mpn_rshift:) /* We can use faster code for shift-by-1 under certain conditions. */ cmp $1,%ecx - jne Lnormal + jne L(normal) leal 4(%edi),%eax cmpl %esi,%eax - jnc Lspecial /* jump if res_ptr + 1 >= s_ptr */ + jnc L(special) /* jump if res_ptr + 1 >= s_ptr */ leal (%edi,%ebp,4),%eax cmpl %eax,%esi - jnc Lspecial /* jump if s_ptr >= res_ptr + size */ + jnc L(special) /* jump if s_ptr >= res_ptr + size */ -Lnormal: +L(normal): movl (%esi),%edx addl $4,%esi xorl %eax,%eax @@ -64,12 +62,12 @@ Lnormal: decl %ebp pushl %ebp shrl $3,%ebp - jz Lend + jz L(end) movl (%edi),%eax /* fetch destination cache line */ ALIGN (2) -Loop: movl 28(%edi),%eax /* fetch destination cache line */ +L(oop): movl 28(%edi),%eax /* fetch destination cache line */ movl %edx,%ebx movl (%esi),%eax @@ -103,21 +101,23 @@ Loop: movl 28(%edi),%eax /* fetch destination cache line */ addl $32,%esi addl $32,%edi decl %ebp - jnz Loop + jnz L(oop) -Lend: popl %ebp +L(end): popl %ebp andl $7,%ebp - jz Lend2 -Loop2: movl (%esi),%eax + jz L(end2) +L(oop2): + movl (%esi),%eax shrdl %cl,%eax,%edx /* compute result limb */ movl %edx,(%edi) movl %eax,%edx addl $4,%esi addl $4,%edi decl %ebp - jnz Loop2 + jnz L(oop2) -Lend2: shrl %cl,%edx /* compute most significant limb */ +L(end2): + shrl %cl,%edx /* compute most significant limb */ movl %edx,(%edi) /* store it */ popl %eax /* pop carry limb */ @@ -133,7 +133,7 @@ Lend2: shrl %cl,%edx /* compute most significant limb */ function is documented to work for overlapping source and destination. */ -Lspecial: +L(special): leal -4(%edi,%ebp,4),%edi leal -4(%esi,%ebp,4),%esi @@ -147,12 +147,13 @@ Lspecial: shrl $1,%edx incl %ebp decl %ebp - jz LLend + jz L(Lend) movl (%edi),%eax /* fetch destination cache line */ ALIGN (2) -LLoop: movl -28(%edi),%eax /* fetch destination cache line */ +L(Loop): + movl -28(%edi),%eax /* fetch destination cache line */ movl %edx,%ebx movl (%esi),%eax @@ -186,14 +187,16 @@ LLoop: movl -28(%edi),%eax /* fetch destination cache line */ leal -32(%esi),%esi /* use leal not to clobber carry */ leal -32(%edi),%edi decl %ebp - jnz LLoop + jnz L(Loop) -LLend: popl %ebp +L(Lend): + popl %ebp sbbl %eax,%eax /* save carry in %eax */ andl $7,%ebp - jz LLend2 + jz L(Lend2) addl %eax,%eax /* restore carry from eax */ -LLoop2: movl %edx,%ebx +L(Loop2): + movl %edx,%ebx movl (%esi),%edx rcrl $1,%edx movl %ebx,(%edi) @@ -201,11 +204,12 @@ LLoop2: movl %edx,%ebx leal -4(%esi),%esi /* use leal not to clobber carry */ leal -4(%edi),%edi decl %ebp - jnz LLoop2 + jnz L(Loop2) - jmp LL1 -LLend2: addl %eax,%eax /* restore carry from eax */ -LL1: movl %edx,(%edi) /* store last limb */ + jmp L(L1) +L(Lend2): + addl %eax,%eax /* restore carry from eax */ +L(L1): movl %edx,(%edi) /* store last limb */ movl $0,%eax rcrl $1,%eax diff --git a/sysdeps/i386/i586/strchr.S b/sysdeps/i386/i586/strchr.S index fda053216f..e351853231 100644 --- a/sysdeps/i386/i586/strchr.S +++ b/sysdeps/i386/i586/strchr.S @@ -1,6 +1,6 @@ -/* strchr -- find character CH in a NUL terminated string. +/* Find character CH in a NUL terminated string. Highly optimized version for ix85, x>=5. - Copyright (C) 1995, 1996 Free Software Foundation, Inc. + Copyright (C) 1995, 1996, 1997 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Ulrich Drepper, <drepper@gnu.ai.mit.edu>. @@ -20,6 +20,7 @@ Boston, MA 02111-1307, USA. */ #include <sysdep.h> +#include "asm-syntax.h" /* This version is especially optimized for the i586 (and following?) processors. This is mainly done by using the two pipelines. The @@ -67,38 +68,38 @@ ENTRY (strchr) orl %ecx, %edx /* and finally c|c|c|c */ andl $3, %edi /* mask alignment bits */ - jz L11 /* alignment is 0 => start loop */ + jz L(11) /* alignment is 0 => start loop */ movb %dl, %cl /* 0 is needed below */ - jp L0 /* exactly two bits set */ + jp L(0) /* exactly two bits set */ xorb (%eax), %cl /* is byte the one we are looking for? */ - jz L2 /* yes => return pointer */ + jz L(2) /* yes => return pointer */ xorb %dl, %cl /* load single byte and test for NUL */ - je L3 /* yes => return NULL */ + je L(3) /* yes => return NULL */ movb 1(%eax), %cl /* load single byte */ incl %eax cmpb %cl, %dl /* is byte == C? */ - je L2 /* aligned => return pointer */ + je L(2) /* aligned => return pointer */ cmpb $0, %cl /* is byte NUL? */ - je L3 /* yes => return NULL */ + je L(3) /* yes => return NULL */ incl %eax decl %edi - jne L11 + jne L(11) -L0: movb (%eax), %cl /* load single byte */ +L(0): movb (%eax), %cl /* load single byte */ cmpb %cl, %dl /* is byte == C? */ - je L2 /* aligned => return pointer */ + je L(2) /* aligned => return pointer */ cmpb $0, %cl /* is byte NUL? */ - je L3 /* yes => return NULL */ + je L(3) /* yes => return NULL */ incl %eax /* increment pointer */ @@ -106,7 +107,7 @@ L0: movb (%eax), %cl /* load single byte */ four instruction up to `L1' will not be executed in the loop because the same code is found at the end of the loop, but there it is executed in parallel with other instructions. */ -L11: movl (%eax), %ecx +L(11): movl (%eax), %ecx movl $magic, %ebp movl $magic, %edi @@ -135,17 +136,17 @@ L11: movl (%eax), %ecx C allows us to overlap the test for 0 in the next word with the test for C. */ -L1: xorl %ecx, %ebp /* (word^magic) */ +L(1): xorl %ecx, %ebp /* (word^magic) */ addl %ecx, %edi /* add magic word */ leal 4(%eax), %eax /* increment pointer */ - jnc L4 /* previous addl caused overflow? */ + jnc L(4) /* previous addl caused overflow? */ movl %ecx, %ebx /* duplicate original word */ orl $magic, %ebp /* (word^magic)|magic */ addl $1, %ebp /* (word^magic)|magic == 0xffffffff? */ - jne L4 /* yes => we found word with NUL */ + jne L(4) /* yes => we found word with NUL */ movl $magic, %esi /* load magic value */ xorl %edx, %ebx /* clear words which are C */ @@ -154,7 +155,7 @@ L1: xorl %ecx, %ebp /* (word^magic) */ addl %ebx, %esi /* (word+magic) */ movl $magic, %edi - jnc L5 /* previous addl caused overflow? */ + jnc L(5) /* previous addl caused overflow? */ movl %edi, %ebp xorl %ebx, %esi /* (word+magic)^word */ @@ -163,19 +164,19 @@ L1: xorl %ecx, %ebp /* (word^magic) */ orl $magic, %esi /* ((word+magic)^word)|magic */ addl $1, %esi /* ((word+magic)^word)|magic==0xf..f?*/ - jne L5 /* yes => we found word with C */ + jne L(5) /* yes => we found word with C */ xorl %ecx, %ebp addl %ecx, %edi leal 4(%eax), %eax - jnc L4 + jnc L(4) movl %ecx, %ebx orl $magic, %ebp addl $1, %ebp - jne L4 + jne L(4) movl $magic, %esi xorl %edx, %ebx @@ -184,7 +185,7 @@ L1: xorl %ecx, %ebp /* (word^magic) */ addl %ebx, %esi movl $magic, %edi - jnc L5 + jnc L(5) movl %edi, %ebp xorl %ebx, %esi @@ -193,19 +194,19 @@ L1: xorl %ecx, %ebp /* (word^magic) */ orl $magic, %esi addl $1, %esi - jne L5 + jne L(5) xorl %ecx, %ebp addl %ecx, %edi leal 4(%eax), %eax - jnc L4 + jnc L(4) movl %ecx, %ebx orl $magic, %ebp addl $1, %ebp - jne L4 + jne L(4) movl $magic, %esi xorl %edx, %ebx @@ -214,7 +215,7 @@ L1: xorl %ecx, %ebp /* (word^magic) */ addl %ebx, %esi movl $magic, %edi - jnc L5 + jnc L(5) movl %edi, %ebp xorl %ebx, %esi @@ -223,19 +224,19 @@ L1: xorl %ecx, %ebp /* (word^magic) */ orl $magic, %esi addl $1, %esi - jne L5 + jne L(5) xorl %ecx, %ebp addl %ecx, %edi leal 4(%eax), %eax - jnc L4 + jnc L(4) movl %ecx, %ebx orl $magic, %ebp addl $1, %ebp - jne L4 + jne L(4) movl $magic, %esi xorl %edx, %ebx @@ -244,7 +245,7 @@ L1: xorl %ecx, %ebp /* (word^magic) */ addl %ebx, %esi movl $magic, %edi - jnc L5 + jnc L(5) movl %edi, %ebp xorl %ebx, %esi @@ -254,29 +255,29 @@ L1: xorl %ecx, %ebp /* (word^magic) */ addl $1, %esi - je L1 + je L(1) /* We know there is no NUL byte but a C byte in the word. %ebx contains NUL in this particular byte. */ -L5: subl $4, %eax /* adjust pointer */ +L(5): subl $4, %eax /* adjust pointer */ testb %bl, %bl /* first byte == C? */ - jz L2 /* yes => return pointer */ + jz L(2) /* yes => return pointer */ incl %eax /* increment pointer */ testb %bh, %bh /* second byte == C? */ - jz L2 /* yes => return pointer */ + jz L(2) /* yes => return pointer */ shrl $16, %ebx /* make upper bytes accessible */ incl %eax /* increment pointer */ cmp $0, %bl /* third byte == C */ - je L2 /* yes => return pointer */ + je L(2) /* yes => return pointer */ incl %eax /* increment pointer */ -L2: popl %ebp /* restore saved registers */ +L(2): popl %ebp /* restore saved registers */ popl %ebx popl %esi @@ -286,38 +287,38 @@ L2: popl %ebp /* restore saved registers */ /* We know there is a NUL byte in the word. But we have to test whether there is an C byte before it in the word. */ -L4: subl $4, %eax /* adjust pointer */ +L(4): subl $4, %eax /* adjust pointer */ cmpb %dl, %cl /* first byte == C? */ - je L2 /* yes => return pointer */ + je L(2) /* yes => return pointer */ cmpb $0, %cl /* first byte == NUL? */ - je L3 /* yes => return NULL */ + je L(3) /* yes => return NULL */ incl %eax /* increment pointer */ cmpb %dl, %ch /* second byte == C? */ - je L2 /* yes => return pointer */ + je L(2) /* yes => return pointer */ cmpb $0, %ch /* second byte == NUL? */ - je L3 /* yes => return NULL */ + je L(3) /* yes => return NULL */ shrl $16, %ecx /* make upper bytes accessible */ incl %eax /* increment pointer */ cmpb %dl, %cl /* third byte == C? */ - je L2 /* yes => return pointer */ + je L(2) /* yes => return pointer */ cmpb $0, %cl /* third byte == NUL? */ - je L3 /* yes => return NULL */ + je L(3) /* yes => return NULL */ incl %eax /* increment pointer */ /* The test four the fourth byte is necessary! */ cmpb %dl, %ch /* fourth byte == C? */ - je L2 /* yes => return pointer */ + je L(2) /* yes => return pointer */ -L3: xorl %eax, %eax /* set return value = NULL */ +L(3): xorl %eax, %eax /* set return value = NULL */ popl %ebp /* restore saved registers */ popl %ebx diff --git a/sysdeps/i386/i586/strlen.S b/sysdeps/i386/i586/strlen.S index e8fb916812..2e6ea680ec 100644 --- a/sysdeps/i386/i586/strlen.S +++ b/sysdeps/i386/i586/strlen.S @@ -1,6 +1,6 @@ /* strlen -- Compute length og NUL terminated string. Highly optimized version for ix86, x>=5. - Copyright (C) 1995, 1996 Free Software Foundation, Inc. + Copyright (C) 1995, 1996, 1997 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Ulrich Drepper, <drepper@gnu.ai.mit.edu>. @@ -20,6 +20,7 @@ Boston, MA 02111-1307, USA. */ #include <sysdep.h> +#include "asm-syntax.h" /* This version is especially optimized for the i586 (and following?) processors. This is mainly done by using the two pipelines. The @@ -46,24 +47,24 @@ ENTRY(strlen) andl %eax, %edx /* separate last two bits of address */ - jz L1 /* aligned => start loop */ - jp L0 /* exactly two bits set */ + jz L(1) /* aligned => start loop */ + jp L(0) /* exactly two bits set */ cmpb %dh, (%eax) /* is byte NUL? */ - je L2 /* yes => return */ + je L(2) /* yes => return */ incl %eax /* increment pointer */ cmpb %dh, (%eax) /* is byte NUL? */ - je L2 /* yes => return */ + je L(2) /* yes => return */ incl %eax /* increment pointer */ xorl $2, %edx - jz L1 + jz L(1) -L0: cmpb %dh, (%eax) /* is byte NUL? */ - je L2 /* yes => return */ +L(0): cmpb %dh, (%eax) /* is byte NUL? */ + je L(2) /* yes => return */ incl %eax /* increment pointer */ xorl %edx, %edx /* We need %edx == 0 for later */ @@ -91,7 +92,7 @@ L0: cmpb %dh, (%eax) /* is byte NUL? */ Note: %edx == 0 in any case here. */ -L1: +L(1): movl (%eax), %ecx /* get word (= 4 bytes) in question */ addl $4, %eax /* adjust pointer for *next* word */ @@ -99,13 +100,13 @@ L1: addl $magic, %ecx /* add magic word */ decl %edx /* complete negation of word */ - jnc L3 /* previous addl caused overflow? */ + jnc L(3) /* previous addl caused overflow? */ xorl %ecx, %edx /* (word+magic)^word */ andl $~magic, %edx /* any of the carry flags set? */ - jne L3 /* yes => determine byte */ + jne L(3) /* yes => determine byte */ movl (%eax), %ecx /* get word (= 4 bytes) in question */ @@ -115,13 +116,13 @@ L1: addl $magic, %ecx /* add magic word */ decl %edx /* complete negation of word */ - jnc L3 /* previous addl caused overflow? */ + jnc L(3) /* previous addl caused overflow? */ xorl %ecx, %edx /* (word+magic)^word */ andl $~magic, %edx /* any of the carry flags set? */ - jne L3 /* yes => determine byte */ + jne L(3) /* yes => determine byte */ movl (%eax), %ecx /* get word (= 4 bytes) in question */ @@ -131,13 +132,13 @@ L1: addl $magic, %ecx /* add magic word */ decl %edx /* complete negation of word */ - jnc L3 /* previous addl caused overflow? */ + jnc L(3) /* previous addl caused overflow? */ xorl %ecx, %edx /* (word+magic)^word */ andl $~magic, %edx /* any of the carry flags set? */ - jne L3 /* yes => determine byte */ + jne L(3) /* yes => determine byte */ movl (%eax), %ecx /* get word (= 4 bytes) in question */ @@ -147,35 +148,35 @@ L1: addl $magic, %ecx /* add magic word */ decl %edx /* complete negation of word */ - jnc L3 /* previous addl caused overflow? */ + jnc L(3) /* previous addl caused overflow? */ xorl %ecx, %edx /* (word+magic)^word */ andl $~magic, %edx /* any of the carry flags set? */ - je L1 /* no => start loop again */ + je L(1) /* no => start loop again */ -L3: subl $4, %eax /* correct too early pointer increment */ +L(3): subl $4, %eax /* correct too early pointer increment */ subl $magic, %ecx cmpb $0, %cl /* lowest byte NUL? */ - jz L2 /* yes => return */ + jz L(2) /* yes => return */ inc %eax /* increment pointer */ testb %ch, %ch /* second byte NUL? */ - jz L2 /* yes => return */ + jz L(2) /* yes => return */ shrl $16, %ecx /* make upper bytes accessible */ incl %eax /* increment pointer */ cmpb $0, %cl /* is third byte NUL? */ - jz L2 /* yes => return */ + jz L(2) /* yes => return */ incl %eax /* increment pointer */ -L2: subl 4(%esp), %eax /* now compute the length as difference +L(2): subl 4(%esp), %eax /* now compute the length as difference between start and terminating NUL character */ diff --git a/sysdeps/i386/i586/sub_n.S b/sysdeps/i386/i586/sub_n.S index d1a2bc0840..758fa86bd3 100644 --- a/sysdeps/i386/i586/sub_n.S +++ b/sysdeps/i386/i586/sub_n.S @@ -1,24 +1,22 @@ /* Pentium __mpn_sub_n -- Subtract two limb vectors of the same length > 0 and store difference in a third limb vector. + Copyright (C) 1992, 1994, 1995, 1996, 1997 Free Software Foundation, Inc. + This file is part of the GNU MP Library. -Copyright (C) 1992, 1994, 1995, 1996 Free Software Foundation, Inc. + 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. -This file is part of the GNU MP Library. + 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 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. - -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 @@ -52,14 +50,14 @@ C_SYMBOL_NAME(__mpn_sub_n:) shrl $3,%ecx andl $7,%edx testl %ecx,%ecx /* zero carry flag */ - jz Lend + jz L(end) pushl %edx ALIGN (3) -Loop: movl 28(%edi),%eax /* fetch destination cache line */ +L(oop): movl 28(%edi),%eax /* fetch destination cache line */ leal 32(%edi),%edi -L1: movl (%esi),%eax +L(1): movl (%esi),%eax movl 4(%esi),%edx sbbl %ebx,%eax movl 4(%ebp),%ebx @@ -68,7 +66,7 @@ L1: movl (%esi),%eax movl %eax,-32(%edi) movl %edx,-28(%edi) -L2: movl 8(%esi),%eax +L(2): movl 8(%esi),%eax movl 12(%esi),%edx sbbl %ebx,%eax movl 12(%ebp),%ebx @@ -77,7 +75,7 @@ L2: movl 8(%esi),%eax movl %eax,-24(%edi) movl %edx,-20(%edi) -L3: movl 16(%esi),%eax +L(3): movl 16(%esi),%eax movl 20(%esi),%edx sbbl %ebx,%eax movl 20(%ebp),%ebx @@ -86,7 +84,7 @@ L3: movl 16(%esi),%eax movl %eax,-16(%edi) movl %edx,-12(%edi) -L4: movl 24(%esi),%eax +L(4): movl 24(%esi),%eax movl 28(%esi),%edx sbbl %ebx,%eax movl 28(%ebp),%ebx @@ -98,14 +96,14 @@ L4: movl 24(%esi),%eax leal 32(%esi),%esi leal 32(%ebp),%ebp decl %ecx - jnz Loop + jnz L(oop) popl %edx -Lend: +L(end): decl %edx /* test %edx w/o clobbering carry */ - js Lend2 + js L(end2) incl %edx -Loop2: +L(oop2): leal 4(%edi),%edi movl (%esi),%eax sbbl %ebx,%eax @@ -114,8 +112,8 @@ Loop2: leal 4(%esi),%esi leal 4(%ebp),%ebp decl %edx - jnz Loop2 -Lend2: + jnz L(oop2) +L(end2): movl (%esi),%eax sbbl %ebx,%eax movl %eax,(%edi) diff --git a/sysdeps/i386/i586/submul_1.S b/sysdeps/i386/i586/submul_1.S index adf2d63e68..20c4d8df0e 100644 --- a/sysdeps/i386/i586/submul_1.S +++ b/sysdeps/i386/i586/submul_1.S @@ -1,24 +1,22 @@ /* Pentium __mpn_submul_1 -- Multiply a limb vector with a limb and subtract the result from a second limb vector. + Copyright (C) 1992, 1994, 1996, 1997 Free Software Foundation, Inc. + This file is part of the GNU MP Library. -Copyright (C) 1992, 1994, 1996 Free Software Foundation, Inc. + 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. -This file is part of the GNU MP Library. + 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 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. - -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 @@ -57,7 +55,7 @@ C_SYMBOL_NAME(__mpn_submul_1:) INSN2(xor,l ,R(ebx),R(ebx)) ALIGN (3) -Loop: INSN2(adc,l ,R(ebx),$0) +L(oop): INSN2(adc,l ,R(ebx),$0) INSN2(mov,l ,R(eax),MEM_INDEX(s1_ptr,size,4)) INSN1(mul,l ,R(s2_limb)) @@ -72,7 +70,7 @@ Loop: INSN2(adc,l ,R(ebx),$0) INSN1(inc,l ,R(size)) INSN2(mov,l ,R(ebx),R(edx)) - INSN1(jnz, ,Loop) + INSN1(jnz, ,L(oop)) INSN2(adc,l ,R(ebx),$0) INSN2(mov,l ,R(eax),R(ebx)) |