From f30be3ddd5a4f52523379b5d179a4462c20aae17 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Borna=20=C5=A0o=C5=A1tari=C4=87?=
 <sostaric.borna.dev@hotmail.com>
Date: Sat, 27 Dec 2025 11:19:42 +0100
Subject: [PATCH] initial implementation of klibc

fix linker error about ctx_switch
---
 .gitignore                                    |   2 +
 Makefile                                      |  21 ++-
 .../{context_switch.s => context_switch.asm}  |   2 +-
 kernel/memory.c                               | 123 +-----------------
 kernel/types.h                                |   1 -
 kernel/utils.c                                |   7 -
 kernel/utils.h                                |   2 +
 klibc/include/stdarg.h                        |  14 ++
 klibc/include/stdbool.h                       |   6 +
 klibc/include/stddef.h                        |  10 ++
 klibc/include/stdint.h                        |  16 +++
 klibc/include/stdio.h                         |   4 +
 klibc/include/stdlib.h                        |   4 +
 klibc/include/string.h                        |  14 ++
 klibc/src/string.c                            | 107 +++++++++++++++
 15 files changed, 198 insertions(+), 135 deletions(-)
 rename kernel/{context_switch.s => context_switch.asm} (98%)
 create mode 100644 klibc/include/stdarg.h
 create mode 100644 klibc/include/stdbool.h
 create mode 100644 klibc/include/stddef.h
 create mode 100644 klibc/include/stdint.h
 create mode 100644 klibc/include/stdio.h
 create mode 100644 klibc/include/stdlib.h
 create mode 100644 klibc/include/string.h
 create mode 100644 klibc/src/string.c

diff --git a/.gitignore b/.gitignore
index 7012c23..432f88b 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,3 +1,5 @@
 .build.env
 build
 cross
+.cache/
+compile_commands.json
diff --git a/Makefile b/Makefile
index a99cafe..1787090 100644
--- a/Makefile
+++ b/Makefile
@@ -6,6 +6,7 @@ QEMU= qemu-system-i386
 OBJCOPY = i386-elf-objcopy
 
 BUILD_DIR = build
+CROSS_DIR = cross
 DISK_IMG = $(BUILD_DIR)/disk.img
 STAGE2_SIZE = 2048
 
@@ -14,6 +15,9 @@ KERNEL_ASM_SRC = $(wildcard kernel/*.asm)
 KERNEL_OBJ = $(patsubst kernel/%.c, $(BUILD_DIR)/%.o, $(KERNEL_C_SRC))
 KERNEL_OBJ += $(patsubst kernel/%.asm, $(BUILD_DIR)/asm_%.o, $(KERNEL_ASM_SRC))
 
+KLIBC_SRC = $(wildcard klibc/src/*.c)
+KLIBC_OBJ = $(patsubst klibc/src/%.c, $(BUILD_DIR)/klibc/%.o, $(KLIBC_SRC))
+
 all: $(DISK_IMG)
 
 .PHONY: stage1 stage2 kernel run gdb clean
@@ -26,7 +30,7 @@ stage1: $(BUILD_DIR)
 # Alternatively, convey the final stage2 size through other means to stage1.
 stage2: $(BUILD_DIR)
 	$(AS) $(ASFLAGS) -o $(BUILD_DIR)/stage2.o bootloader/stage2.asm
-	$(CC) -std=c11 -ffreestanding -nostdlib -fno-stack-protector -m32 -g -c -o $(BUILD_DIR)/stage2_load.o bootloader/stage2_load.c
+	$(CC) -std=c11 -ffreestanding -nostdlib -nostdinc -fno-stack-protector -m32 -Iklibc/include -g -c -o $(BUILD_DIR)/stage2_load.o bootloader/stage2_load.c
 	$(LD) -Tbootloader/stage2.ld -melf_i386 -o $(BUILD_DIR)/$@.elf $(BUILD_DIR)/stage2.o $(BUILD_DIR)/stage2_load.o
 	$(OBJCOPY) -O binary $(BUILD_DIR)/$@.elf $(BUILD_DIR)/$@.bin
 	truncate -s $(STAGE2_SIZE) $(BUILD_DIR)/$@.bin
@@ -35,10 +39,13 @@ $(BUILD_DIR)/asm_%.o: kernel/%.asm
 	$(AS) $(ASFLAGS) -o $@ $<
 
 $(BUILD_DIR)/%.o: kernel/%.c
-	$(CC) -std=c11 -ffreestanding -nostdlib -fno-stack-protector -m32 -g -c -o $@ $<
+	$(CC) -std=c11 -ffreestanding -nostdlib -nostdinc -fno-stack-protector -m32 -Iklibc/include -g -c -o $@ $<
 
-kernel: $(KERNEL_OBJ) | $(BUILD_DIR)
-	$(LD) -melf_i386 -Tkernel/linker.ld -o $(BUILD_DIR)/kernel.elf $(KERNEL_OBJ)
+$(BUILD_DIR)/klibc/%.o: klibc/src/%.c
+	$(CC) -std=c11 -ffreestanding -nostdlib -nostdinc -fno-stack-protector -m32 -Iklibc/include -g -c -o $@ $<
+
+kernel: $(KERNEL_OBJ) | $(BUILD_DIR) $(KLIBC_OBJ)
+	$(LD) -melf_i386 -Tkernel/linker.ld -o $(BUILD_DIR)/kernel.elf $(KERNEL_OBJ) $(KLIBC_OBJ)
 
 $(DISK_IMG): stage1 stage2 kernel
 	dd if=$(BUILD_DIR)/stage1.bin of=$@
@@ -48,6 +55,7 @@ $(DISK_IMG): stage1 stage2 kernel
 
 $(BUILD_DIR):
 	mkdir -p $@
+	mkdir -p $(BUILD_DIR)/klibc
 
 run:
 	qemu-system-i386 -s -S $(DISK_IMG)
@@ -57,4 +65,9 @@ gdb:
 
 clean:
 	rm -rf $(BUILD_DIR)
+
+clean-cross:
 	rm -rf $(CROSS_DIR)
+	rm -rf .build.env
+
+clean-all: clean clean-cross
diff --git a/kernel/context_switch.s b/kernel/context_switch.asm
similarity index 98%
rename from kernel/context_switch.s
rename to kernel/context_switch.asm
index 0e68648..89a910d 100644
--- a/kernel/context_switch.s
+++ b/kernel/context_switch.asm
@@ -1,4 +1,4 @@
-.global ctx_switch
+global ctx_switch
 
 ; void ctx_switch(uint32_t **old_sp_ptr, uint32_t *new_sp);
 ; Arguments on stack (cdecl convention):
diff --git a/kernel/memory.c b/kernel/memory.c
index 9f4d02e..dc6aae2 100644
--- a/kernel/memory.c
+++ b/kernel/memory.c
@@ -1,6 +1,6 @@
 #include "memory.h"
 
-/* note: this is a stub, please use care as theres duplicate functions in utils implementation 
+/* note: this is a stub, please use care as theres duplicate functions in utils implementation
 /* --------------------------------------------------------------------- *
  *  Helper: copy a single byte (used by both memcpy and memmove)
  * --------------------------------------------------------------------- */
@@ -15,124 +15,3 @@ static inline void byte_copy_backward(uint8_t *dst, const uint8_t *src, size_t n
     while (n--) *--dst = *--src;
 }
 
-/* --------------------------------------------------------------------- *
- *  memcpy  –  no overlap allowed (behaviour undefined if overlap)
- * --------------------------------------------------------------------- */
-void *memcpy(void *restrict dst, const void *restrict src, size_t n)
-{
-    uint8_t *d = (uint8_t *)dst;
-    const uint8_t *s = (const uint8_t *)src;
-
-#if defined(MEMORY_OPTIMIZED)
-    /* Align destination to 4-byte boundary */
-    size_t align = (uintptr_t)d & 3U;
-    if (align) {
-        size_t head = 4 - align;
-        if (head > n) head = n;
-        byte_copy_forward(d, s, head);
-        d += head; s += head; n -= head;
-    }
-
-    /* 32-bit word copy – safe because we already aligned dst */
-    {
-        uint32_t *d32 = (uint32_t *)d;
-        const uint32_t *s32 = (const uint32_t *)s;
-        size_t words = n / 4;
-        while (words--) *d32++ = *s32++;
-        d = (uint8_t *)d32;
-        s = (const uint8_t *)s32;
-        n &= 3;
-    }
-#endif
-
-    byte_copy_forward(d, s, n);
-    return dst;
-}
-
-/* --------------------------------------------------------------------- *
- *  memmove – handles overlapping regions correctly
- * --------------------------------------------------------------------- */
-void *memmove(void *dst, const void *src, size_t n)
-{
-    uint8_t *d = (uint8_t *)dst;
-    const uint8_t *s = (const uint8_t *)src;
-
-    if (n == 0 || dst == src)
-        return dst;
-
-    if (d < s) {                     /* copy forward */
-#if defined(MEMORY_OPTIMIZED)
-        /* Same fast path as memcpy when no overlap */
-        size_t align = (uintptr_t)d & 3U;
-        if (align) {
-            size_t head = 4 - align;
-            if (head > n) head = n;
-            byte_copy_forward(d, s, head);
-            d += head; s += head; n -= head;
-        }
-        {
-            uint32_t *d32 = (uint32_t *)d;
-            const uint32_t *s32 = (const uint32_t *)s;
-            size_t words = n / 4;
-            while (words--) *d32++ = *s32++;
-            d = (uint8_t *)d32;
-            s = (const uint8_t *)s32;
-            n &= 3;
-        }
-#endif
-        byte_copy_forward(d, s, n);
-    } else {                         /* copy backward */
-        byte_copy_backward(d, s, n);
-    }
-    return dst;
-}
-
-/* --------------------------------------------------------------------- *
- *  memcmp  – lexicographical compare
- * --------------------------------------------------------------------- */
-int memcmp(const void *s1, const void *s2, size_t n)
-{
-    const uint8_t *a = (const uint8_t *)s1;
-    const uint8_t *b = (const uint8_t *)s2;
-
-#if defined(MEMORY_OPTIMIZED)
-    /* Align to 4-byte boundary */
-    size_t align = (uintptr_t)a & 3U;
-    if (align && align == ((uintptr_t)b & 3U)) {
-        size_t head = 4 - align;
-        if (head > n) head = n;
-        while (head--) {
-            int diff = *a++ - *b++;
-            if (diff) return diff;
-        }
-        n -= head;
-    }
-
-    {
-        const uint32_t *a32 = (const uint32_t *)a;
-        const uint32_t *b32 = (const uint32_t *)b;
-        size_t words = n / 4;
-        while (words--) {
-            uint32_t va = *a32++, vb = *b32++;
-            if (va != vb) {
-                /* byte-wise fallback for the differing word */
-                const uint8_t *pa = (const uint8_t *)(a32 - 1);
-                const uint8_t *pb = (const uint8_t *)(b32 - 1);
-                for (int i = 0; i < 4; ++i) {
-                    int diff = pa[i] - pb[i];
-                    if (diff) return diff;
-                }
-            }
-        }
-        a = (const uint8_t *)a32;
-        b = (const uint8_t *)b32;
-        n &= 3;
-    }
-#endif
-
-    while (n--) {
-        int diff = *a++ - *b++;
-        if (diff) return diff;
-    }
-    return 0;
-}
diff --git a/kernel/types.h b/kernel/types.h
index 32fc7ff..332aa4e 100644
--- a/kernel/types.h
+++ b/kernel/types.h
@@ -27,7 +27,6 @@ typedef enum { false = 0, true = 1 } bool;
 // ----------------------------
 // OS subsystem types
 // ----------------------------
-typedef uint32_t size_t;
 typedef int32_t  ssize_t;
 
 typedef uint32_t phys_addr_t; // Physical address
diff --git a/kernel/utils.c b/kernel/utils.c
index 0cb4b04..a0f2035 100644
--- a/kernel/utils.c
+++ b/kernel/utils.c
@@ -76,10 +76,3 @@ char* utoa(unsigned int value, char* str, int base) {
     reverse(str, i);
     return str;
 }
-
-void *memset(void *dest, int value, size_t len) {
-    unsigned char *ptr = (unsigned char *)dest;
-    while (len-- > 0)
-        *ptr++ = (unsigned char)value;
-    return dest;
-}
diff --git a/kernel/utils.h b/kernel/utils.h
index 802db8c..53fab8a 100644
--- a/kernel/utils.h
+++ b/kernel/utils.h
@@ -1,6 +1,8 @@
 #ifndef UTILS_H
 #define UTILS_H
 
+#include <stddef.h>
+
 #include "types.h"
 
 // Convert integer to string (base is typically 10, 16, etc.)
diff --git a/klibc/include/stdarg.h b/klibc/include/stdarg.h
new file mode 100644
index 0000000..07d50ca
--- /dev/null
+++ b/klibc/include/stdarg.h
@@ -0,0 +1,14 @@
+#ifndef CLASSICOS_KLIBC_STDARG_H
+#define CLASSICOS_KLIBC_STDARG_H
+
+typedef __builtin_va_list va_list;
+
+#ifndef va_start
+#define va_start(ap, param) __builtin_va_start(ap, param)
+#endif
+
+#define va_end(ap) __builtin_va_end(ap)
+#define va_arg(ap, type) __builtin_va_arg(ap, type)
+#define va_copy(dest, src) __builtin_va_copy(dest, src)
+
+#endif  // CLASSICOS_KLIBC_STDARG_H
diff --git a/klibc/include/stdbool.h b/klibc/include/stdbool.h
new file mode 100644
index 0000000..0cd5393
--- /dev/null
+++ b/klibc/include/stdbool.h
@@ -0,0 +1,6 @@
+#ifndef CLASSICOS_KLIBC_STDBOOL_H
+#define CLASSICOS_KLIBC_STDBOOL_H
+
+typedef enum { false = 0, true = 1 } bool;
+
+#endif  // CLASSICOS_KLIBC_STDBOOL_H
diff --git a/klibc/include/stddef.h b/klibc/include/stddef.h
new file mode 100644
index 0000000..3b67bd7
--- /dev/null
+++ b/klibc/include/stddef.h
@@ -0,0 +1,10 @@
+#ifndef CLASSICOS_KLIBC_STDDEF_H
+#define CLASSICOS_KLIBC_STDDEF_H
+
+typedef __SIZE_TYPE__ size_t;
+typedef __PTRDIFF_TYPE__ ptrdiff_t;
+
+#undef NULL
+#define NULL ((void*)0)
+
+#endif  // CLASSICOS_KLIBC_STDDEF_H
diff --git a/klibc/include/stdint.h b/klibc/include/stdint.h
new file mode 100644
index 0000000..a0db679
--- /dev/null
+++ b/klibc/include/stdint.h
@@ -0,0 +1,16 @@
+#ifndef CLASSICOS_KLIBC_STDINT_H
+#define CLASSICOS_KLIBC_STDINT_H
+
+typedef signed char int8_t;
+typedef short int int16_t;
+typedef int int32_t;
+typedef long long int int64_t;
+
+typedef unsigned char uint8_t;
+typedef unsigned short int uint16_t;
+typedef unsigned int uint32_t;
+typedef unsigned long long int uint64_t;
+
+typedef unsigned int uintptr_t;
+
+#endif  // CLASSICOS_KLIBC_STDINT_H
diff --git a/klibc/include/stdio.h b/klibc/include/stdio.h
new file mode 100644
index 0000000..af112bf
--- /dev/null
+++ b/klibc/include/stdio.h
@@ -0,0 +1,4 @@
+#ifndef CLASSICOS_KLIBC_STDIO_H
+#define CLASSICOS_KLIBC_STDIO_H
+
+#endif  // CLASSICOS_KLIBC_STDIO_H
diff --git a/klibc/include/stdlib.h b/klibc/include/stdlib.h
new file mode 100644
index 0000000..5046538
--- /dev/null
+++ b/klibc/include/stdlib.h
@@ -0,0 +1,4 @@
+#ifndef CLASSICOS_KLIBC_STDLIB_H
+#define CLASSICOS_KLIBC_STDLIB_H
+
+#endif  // CLASSICOS_KLIBC_STDLIB_H
diff --git a/klibc/include/string.h b/klibc/include/string.h
new file mode 100644
index 0000000..4f01fc1
--- /dev/null
+++ b/klibc/include/string.h
@@ -0,0 +1,14 @@
+#ifndef CLASSICOS_KLIBC_STRING_H
+#define CLASSICOS_KLIBC_STRING_H
+
+#include <stddef.h>
+
+extern int memcmp(const void* s1, const void* s2, size_t n);
+extern void* memmove(void* dst, const void* src, size_t n);
+extern void* memcpy(void* dst, const void* src, size_t n);
+extern void* memset(void* dst, int c, size_t n);
+
+extern size_t strlen(const char* s);
+extern int strcmp(const char* s1, const char* s2);
+
+#endif  // CLASSICOS_KLIBC_STRING_H
diff --git a/klibc/src/string.c b/klibc/src/string.c
new file mode 100644
index 0000000..a9a8a84
--- /dev/null
+++ b/klibc/src/string.c
@@ -0,0 +1,107 @@
+#include <string.h>
+
+int memcmp(const void* s1, const void* s2, size_t n) {
+    const unsigned char* c1 = s1;
+    const unsigned char* c2 = s2;
+    int d = 0;
+
+    while (n--) {
+        d = (int)*c1++ - (int)*c2++;
+        if (d) break;
+    }
+
+    return d;
+}
+
+void* memmove(void* dst, const void* src, size_t n) {
+    const char* p = src;
+    char* q = dst;
+#if defined(__i386__) || defined(__x86_64__)
+    if (q < p) {
+        __asm__ volatile("cld; rep; movsb" : "+c"(n), "+S"(p), "+D"(q));
+    } else {
+        p += (n - 1);
+        q += (n - 1);
+        __asm__ volatile("std; rep; movsb; cld" : "+c"(n), "+S"(p), "+D"(q));
+    }
+#else
+    if (q < p) {
+        while (n--) {
+            *q++ = *p++;
+        }
+    } else {
+        p += n;
+        q += n;
+        while (n--) {
+            *--q = *--p;
+        }
+    }
+#endif
+
+    return dst;
+}
+
+void* memcpy(void* dst, const void* src, size_t n) {
+    const char* p = src;
+    char* q = dst;
+#if defined(__i386__)
+    size_t nl = n >> 2;
+    __asm__ volatile("cld ; rep ; movsl ; movl %3,%0 ; rep ; movsb"
+                    : "+c"(nl), "+S"(p), "+D"(q)
+                    : "r"(n & 3));
+#elif defined(__x86_64__)
+    size_t nq = n >> 3;
+    __asm__ volatile("cld ; rep ; movsq ; movl %3,%%ecx ; rep ; movsb"
+                    : "+c"(nq), "+S"(p), "+D"(q)
+                    : "r"((uint32_t)(n & 7)));
+#else
+    while (n--) {
+        *q++ = *p++;
+    }
+#endif
+
+    return dst;
+}
+
+void* memset(void* dst, int c, size_t n) {
+    char* q = dst;
+
+#if defined(__i386__)
+    size_t nl = n >> 2;
+    __asm__ volatile("cld ; rep ; stosl ; movl %3,%0 ; rep ; stosb"
+                    : "+c"(nl), "+D"(q)
+                    : "a"((unsigned char)c * 0x01010101U), "r"(n & 3));
+#elif defined(__x86_64__)
+    size_t nq = n >> 3;
+    __asm__ volatile("cld ; rep ; stosq ; movl %3,%%ecx ; rep ; stosb"
+                    : "+c"(nq), "+D"(q)
+                    : "a"((unsigned char)c * 0x0101010101010101U),
+                      "r"((uint32_t)n & 7));
+#else
+    while (n--) {
+        *q++ = c;
+    }
+#endif
+
+    return dst;
+}
+
+size_t strlen(const char* s) {
+    const char* ss = s;
+    while (*ss) ss++;
+    return ss - s;
+}
+
+int strcmp(const char* s1, const char* s2) {
+    const unsigned char* c1 = (const unsigned char*)s1;
+    const unsigned char* c2 = (const unsigned char*)s2;
+    unsigned char ch;
+    int d = 0;
+
+    while (1) {
+        d = (int)(ch = *c1++) - (int)*c2++;
+        if (d || !ch) break;
+    }
+
+    return d;
+}