moving some standard library functions to assembly

diff --git a/src/libc/string.c b/src/libc/string.c
index f251f67..ce3c2e7 100644 (file)
--- a/src/libc/string.c
+++ b/src/libc/string.c
@@ -17,6 +17,7 @@ along with this program.  If not, see <https://www.gnu.org/licenses/>.
 */
 #include <string.h>
 
+/*
 void memset(void *s, int c, size_t n)
 {
        char *ptr = s;
@@ -24,10 +25,12 @@ void memset(void *s, int c, size_t n)
                *ptr++ = c;
        }
 }
+*/
 
 /* Does the same thing as memset only with 16bit values.
  * n in this case is the number of values, not the number of bytes.
  */
+/*
 void memset16(void *s, int c, size_t n)
 {
        int16_t *ptr = s;
@@ -35,7 +38,8 @@ void memset16(void *s, int c, size_t n)
                *ptr++ = c;
        }
 }
-
+*/
+/*
 void *memcpy(void *dest, const void *src, size_t n)
 {
        char *dptr = dest;
@@ -46,6 +50,7 @@ void *memcpy(void *dest, const void *src, size_t n)
        }
        return dest;
 }
+*/
 
 void *memmove(void *dest, const void *src, size_t n)
 {

diff --git a/src/libc/string_asm.s b/src/libc/string_asm.s
new file mode 100644 (file)
index 0000000..7bb6f40
--- /dev/null
+++ b/src/libc/string_asm.s
@@ -0,0 +1,179 @@
+# pcboot - bootable PC demo/game kernel
+# Copyright (C) 2018  John Tsiombikas <nuclear@member.fsf.org>
+# 
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+# 
+# This program 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 General Public License for more details.
+# 
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+       # standard C memset
+       .global memset
+memset:
+       push %ebp
+       mov %esp, %ebp
+       push %edi
+
+       mov 8(%ebp), %edi
+       mov 12(%ebp), %al
+       mov %al, %ah
+       mov %ax, %cx
+       shl $16, %eax
+       mov %cx, %ax
+       mov 16(%ebp), %ecx
+
+       cmp $0, %ecx
+       jz msdone
+
+       mov %edi, %edx
+       and $3, %edx
+       jmp *mspre_tab(,%edx,4)
+
+mspre_tab: .long msmain, mspre1, mspre2, mspre3
+mspre1:        stosb
+       dec %ecx
+mspre2:        stosb
+       dec %ecx
+mspre3:        stosb
+       dec %ecx
+       jz msdone
+
+msmain:
+       push %ecx
+       shr $2, %ecx
+       rep stosl
+       pop %ecx
+
+       and $3, %ecx
+       jmp *mspost_tab(,%ecx,4)
+
+mspost_tab: .long msdone, mspost1, mspost2, mspost3
+mspost3:stosb
+mspost2:stosb
+mspost1:stosb
+
+msdone:
+       pop %edi
+       pop %ebp
+       ret
+
+
+       # same as memset but copies 16bit values, and the count is the number
+       # of 16bit values to copy, not bytes.
+       .global memset16
+memset16:
+       push %ebp
+       mov %esp, %ebp
+       push %edi
+
+       mov 8(%ebp), %edi
+       mov 12(%ebp), %ax
+       shl $16, %eax
+       mov 12(%ebp), %ax
+       mov 16(%ebp), %ecx
+
+       cmp $0, %ecx
+       jz ms16done
+
+       mov %edi, %edx
+       and $3, %edx
+       jmp *ms16pre_tab(,%edx,4)
+
+ms16pre_tab: .long ms16main, ms16pre1, ms16pre2, ms16pre3
+ms16pre3:
+       # unaligned by +3:
+       # same as next one, but jump to ms16main instead of falling through
+       mov %al, (%edi)
+       mov %ah, -1(%edi,%ecx,2)
+       rol $8, %eax
+       inc %edi
+       dec %ecx
+       jz ms16done
+       jmp ms16main
+ms16pre1:
+       # unaligned by +1:
+       # - write low byte at edi
+       # - write high byte at the end
+       # - rotate by 8 for the rest
+       # - decrement ecx
+       mov %al, (%edi)
+       mov %ah, -1(%edi,%ecx,2)
+       rol $8, %eax
+       inc %edi
+       dec %ecx
+       jz ms16done
+ms16pre2:
+       # unaligned by +2
+       stosw
+       dec %ecx
+       jz ms16done
+
+ms16main:
+       push %ecx
+       shr $1, %ecx
+       rep stosl
+       pop %ecx
+
+       and $1, %ecx
+       jz ms16done
+       stosw
+ms16done:
+
+       pop %edi
+       pop %ebp
+       ret
+
+       # standard C memcpy
+       .global memcpy
+memcpy:
+       push %ebp
+       mov %esp, %ebp
+       push %edi
+       push %esi
+
+       mov 8(%ebp), %edi
+       mov 12(%ebp), %esi
+       mov 16(%ebp), %ecx
+
+       cmp $0, %ecx
+       jz mcdone
+
+       mov %edi, %edx
+       and $3, %edx
+       jmp *mcpre_tab(,%edx,4)
+
+mcpre_tab: .long mcmain, mcpre1, mcpre2, mcpre3
+mcpre1:        movsw
+       dec %ecx
+mcpre2:        movsw
+       dec %ecx
+mcpre3:        movsw
+       dec %ecx
+       jz mcdone
+
+mcmain:
+       push %ecx
+       shr $2, %ecx
+       rep movsl
+       pop %ecx
+
+       and $3, %ecx
+       jmp *mcpost_tab(,%ecx,4)
+
+mcpost_tab: .long mcdone, mcpost1, mcpost2, mcpost3
+mcpost3:movsb
+mcpost2:movsb
+mcpost1:movsb
+
+mcdone:
+       pop %esi
+       pop %edi
+       pop %ebp
+       ret