diff options
Diffstat (limited to 'sysdeps/i386/i586/lshift.S')
| -rw-r--r-- | sysdeps/i386/i586/lshift.S | 54 |
1 files changed, 30 insertions, 24 deletions
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 |