diff --git a/Makefile b/Makefile
index ba05820..f83231b 100644
--- a/Makefile
+++ b/Makefile
@@ -41,13 +41,11 @@ INTERNAL_OBJS := $(CRTI_OBJ) $(OBJS) $(CRTN_OBJ)
 build: all iso run
 
 # === Build all targets and clean ===
-buildc: test all iso run clean
+buildc: all iso run clean
 
 # === Default target ===
 all: $(TARGET)
 
-test:
-	printf $(CRTI_OBJ)
 
 # === Linking ===
 $(TARGET): ./arch/x86/boot/boot.o $(filter-out boot.o, $(OBJ_LINK_LIST))
diff --git a/README.md b/README.md
new file mode 100644
index 0000000..9a25b69
--- /dev/null
+++ b/README.md
@@ -0,0 +1,10 @@
+<--- Espresso. --->
+
+Espresso is a simple operating system built with C and i686(IA32) assembler.
+
+== BUILDING ==
+
+First, you need a cross-compiler and toolchain built, see https://wiki.osdev.org/GCC_Cross-Compiler for instructions.
+
+then, `cd` into the directory with the code (you should see the `kernel`, `include`, and `lib` directories if you're in the right place.) and type `make build`.
+this will also run the OS. type `make clean` after every change to rebuild everything.
diff --git a/arch/x86/boot/boot.s b/arch/x86/boot/boot.s
index 9dd6705..5ae98bf 100644
--- a/arch/x86/boot/boot.s
+++ b/arch/x86/boot/boot.s
@@ -45,29 +45,8 @@ doesn't make sense to return from this function as the bootloader is gone.
 
 .global _start
 .global _kernel_early
-/*.global loadPageDirectory
-.global enablePaging*/
 
 .type _start, @function
-
-/*load_page_directory:
-        push %ebp
-        mov %esp, %ebp
-        mov 8(%esp), %eax
-        mov %eax, %cr3
-        mov %ebp, %esp
-        pop %ebp
-        ret
-
-enable_paging:
-        push %ebp
-        mov %esp, %ebp
-        mov %cr0, %eax
-        or $0x80000000, %eax
-        mov %eax, %cr0
-        mov %ebp, %esp
-        pop %ebp
-        ret*/
         
 enable_sse_asm:
     push %eax
@@ -138,7 +117,7 @@ enable_sse_asm:
 .no_sse:
 .no_cpuid:
     lea sse_initialized, %edi
-    mov $0, (%edi)
+    movl $0, (%edi)
 
 .done:
     pop %edi
@@ -203,9 +182,8 @@ _start:
 	
 	
 	/*
-	Call _kernel_early, early low-level initialization will happen there;
-	please note that while _kernel_early is written in assembler,
-	kernel_early is written in C. (kernel_early is called by _kernel_early, don't be confused. ;) )
+	Call _kernel_early, early low-level initialization will happen there.
+	NOTE: I don't use 'jmp' because it doesn't work. Nice try.
 	*/
 	call _kernel_early
 
diff --git a/arch/x86/new_intrin.c b/arch/x86/new_intrin.c
new file mode 100644
index 0000000..b0ea4ad
--- /dev/null
+++ b/arch/x86/new_intrin.c
@@ -0,0 +1,16 @@
+#include <types.h>
+
+#ifdef __NEW_INTRINSICS_
+
+bool cpu_supports_sse2(void)
+{
+    int k = __buildin_cpu_supports("sse2");
+    return k > 0 ? true : false
+}
+
+bool cpu_supports_sse(void)
+{
+    return __buildin_cpu_supports("sse") > 0 ? true : false;
+}
+
+#endif
diff --git a/drivers/fs/sfs.c b/drivers/fs/sfs.c
index c58dd12..8d104f1 100644
--- a/drivers/fs/sfs.c
+++ b/drivers/fs/sfs.c
@@ -5,7 +5,12 @@
 #include <fs/sfs.h>
 
 
-#define SFS_BEGINNING_LBA 8
+#define SFS_HEADER_START_LBA 8
+#define SFS_HEADER_REGION_SECTORS 1024
+#define SFS_HEADERS_PER_SECTOR 8
+#define TOTAL_SECTORS 32768
+#define SFS_DATA_START_LBA (TOTAL_SECTORS / 4)
+
 
 #pragma pack(push, 1)
 typedef struct sfs_file_header {
@@ -24,38 +29,39 @@ typedef struct sfs_file_header {
 } sfs_file_header_t;
 #pragma pack(pop)
 
-static sfs_file_header_t* base;
+
+uint8_t buffer[512] = { 0 };
+
+/* Global allocator pointer to the next free sector in the data region */
+uint64_t sfs_next_free_data_sector = SFS_DATA_START_LBA;
 
 
 
-
-int32_t sfs_init(bool format)
+int32_t sfs_init(void)
 {
-  uint8_t buffer[512] = { 0 };
+  uint8_t sector_buf[512];
 
-  int32_t kl = read_sector(SFS_BEGINNING_LBA, buffer);
-  
-  if (kl != 0)
+  for (uint64_t lba = SFS_HEADER_START_LBA; lba < SFS_HEADER_START_LBA + SFS_HEADER_REGION_SECTORS; ++lba)
   {
-    return -1;
-  }
-  
-  volatile sfs_file_header_t null_file;
-  memcpy(&null_file, buffer, sizeof(sfs_file_header_t));
-  
-  if ((&null_file)->magic != 0xBEEF)
-  {
-    if (!format)
+    if (read_sector(lba, sector_buf) != 0)
     {
-      return -2;
+      return -1;
     }
-    
-    //sfs_format();
-    return -2;
+
+    sfs_file_header_t* hdr = (sfs_file_header_t*)sector_buf;
+
+    if (hdr->magic == 0xBEEF)
+    {
+      uint64_t end = hdr->data_beginning + hdr->num_sectors;
+      if (end > sfs_next_free_data_sector)
+      {
+        sfs_next_free_data_sector = end;
+      }
+    }
+
+    /* TODO: support multiple headers per sector, loop through all 512/64 = 8 headers here */
   }
-  
-  memcpy(base, &null_file, sizeof(sfs_file_header_t));
-  
+
   return 0;
 }
 
@@ -75,20 +81,210 @@ bool sfs_get_formatted_name(const char* raw_name, char output[37])
   return true;
 }
 
+int32_t sfs_create_file(const char* _name, int32_t size_bytes)
+{
+  if (size_bytes < 0)
+  {
+    return -1;
+  }
 
-/* 
+  char name[37];
+  if (!sfs_get_formatted_name(_name, name))
+  {
+    return -2;
+  }
+
+  uint32_t sectors_needed = (size_bytes + 511) / 512;
+  uint8_t sector_buf[512];
+
+  for (uint64_t lba = SFS_HEADER_START_LBA; lba < SFS_HEADER_START_LBA + SFS_HEADER_REGION_SECTORS; ++lba)
+  {
+    if (read_sector(lba, sector_buf) != 0)
+    {
+      return -3;
+    }
+
+    for (int i = 0; i < SFS_HEADERS_PER_SECTOR; ++i)
+    {
+      sfs_file_header_t* hdr = ((sfs_file_header_t*)sector_buf) + i;
+
+      if (hdr->magic != 0xBEEF)
+      {
+        hdr->magic = 0xBEEF;
+        memcpy(hdr->filename, name, 37);
+        hdr->data_beginning = sfs_next_free_data_sector;
+        hdr->num_sectors = sectors_needed;
+
+        sfs_next_free_data_sector += sectors_needed;
+
+        if (write_sector(lba, sector_buf) != 0)
+        {
+          return -4;
+        }
+
+        return 0;
+      }
+    }
+  }
+
+  return -5;
+}
+
+
+/*
+  num is number of bytes
   Note: if a negative value is passed for num, then read the whole file
 */
-int32_t sfs_read_file(char* _name, void* buffer, uint32_t num)
+int32_t sfs_read_file(char* _name, void* buffer, int32_t num)
 {
   char name[37] = { '\0' };
   
-  if (sfs_get_formatted_name(_name, name))
+  if (!sfs_get_formatted_name(_name, name))
   {
-    return -128;
+    return -1;
   }
   
-  printf("%s", name);
+  memclr(buffer, sizeof(buffer)); /* Zero out buffer */
+  
+  uint64_t j = SFS_HEADER_START_LBA;
+  
+  if (read_sector(j, buffer) != 0)
+  {
+    return -2;
+  }
+  
+  sfs_file_header_t file;
+  
+  memcpy(&file, buffer, sizeof(file));
+  
+  if (file.magic != 0xBEEF)
+  {
+    return -3;
+  }
+  
+  while (strncmp(file.filename, name, 37) != 0)
+  {
+    memclr(buffer, sizeof(buffer));
+    
+    if (read_sector(++j, buffer) != 0)
+    {
+      return -4;
+    }
+    
+    memcpy(&file, buffer, sizeof(file));
+    
+    if (file.magic != 0xBEEF)
+    {
+      return -5;
+    }
+  }
+  
+  /* File found, now to read data */
+  
+  uint32_t sectors_to_read = file.num_sectors;
+  if (num >= 0)
+  {
+    uint32_t max_sectors = (num + 511) / 512;
+    
+    if (max_sectors < sectors_to_read)
+    {
+      sectors_to_read = max_sectors;
+    }
+  }
+
+  for (uint32_t i = 0; i < sectors_to_read; ++i)
+  {
+    if (read_sector((uint64_t)(file.data_beginning) + i, (uint8_t*)buffer + i * 512) != 0)
+    {
+      return -6; /* read error */
+    }
+  }
+
+  return sectors_to_read * 512;
+  
+  memclr(buffer, sizeof(buffer));
   
   return 0;
 }
+
+/*
+  Writes num bytes to an existing file.
+  If the file exists, it will be resized (grown/shrunk) to fit num bytes.
+  Returns number of bytes written (rounded up to sectors), or a negative error code.
+*/
+int32_t sfs_write_file(char* _name, const void* buffer, int32_t num)
+{
+  if (num < 0)
+  {
+    return -1;
+  }
+
+  char name[37] = { '\0' };
+  if (!sfs_get_formatted_name(_name, name))
+  {
+    return -2;
+  }
+
+  uint8_t sector_buf[512];
+  uint64_t header_lba = SFS_HEADER_START_LBA;
+  sfs_file_header_t file;
+
+  while (1) {
+    if (read_sector(header_lba, sector_buf) != 0)
+    {
+      return -3;
+    }
+
+    memcpy(&file, sector_buf, sizeof(file));
+    
+    if (file.magic != 0xBEEF)
+    {
+      return -4;
+    }
+
+    if (strncmp(file.filename, name, 37) == 0)
+    {
+      break;
+    }
+
+    ++header_lba;
+  }
+
+  uint32_t sectors_needed = (num + 511) / 512;
+
+  if (sectors_needed != file.num_sectors)
+  {
+    file.num_sectors = sectors_needed;
+    file.data_beginning = sfs_next_free_data_sector;
+    sfs_next_free_data_sector += sectors_needed;
+  }
+
+  file.magic = 0xBEEF;
+  memcpy(sector_buf, &file, sizeof(file));
+  
+  if (write_sector(header_lba, sector_buf) != 0)
+  {
+    return -5;
+  }
+
+  for (uint32_t i = 0; i < sectors_needed; ++i)
+  {
+    uint8_t temp[512] = {0};
+
+    if (i == sectors_needed - 1 && (num % 512 != 0))
+    {
+      memcpy(temp, (const uint8_t*)buffer + i * 512, num % 512);
+    }
+    else
+    {
+      memcpy(temp, (const uint8_t*)buffer + i * 512, 512);
+    }
+
+    if (write_sector(file.data_beginning + i, temp) != 0)
+    {
+      return -6;
+    }
+  }
+
+  return sectors_needed * 512;
+}
diff --git a/include/fs/sfs.h b/include/fs/sfs.h
index 53c1f85..be8a842 100644
--- a/include/fs/sfs.h
+++ b/include/fs/sfs.h
@@ -3,10 +3,13 @@
 
 #include <types.h>
 
-int32_t sfs_init(bool format);
+int32_t sfs_init(void);
 
 bool sfs_get_formatted_name(const char* raw_name, char output[37]);
 
-int32_t sfs_read_file(char* name, void* buffer, uint32_t num);
+int32_t sfs_create_file(const char* _name, int32_t size_bytes);
+
+int32_t sfs_read_file(char* name, void* buffer, int32_t num);
+int32_t sfs_write_file(char* _name, const void* buffer, int32_t num);
 
 #endif
diff --git a/include/kernel/kconfig.h b/include/kernel/kconfig.h
index 435af3c..7b4673b 100644
--- a/include/kernel/kconfig.h
+++ b/include/kernel/kconfig.h
@@ -13,6 +13,13 @@ extern "C" {
 //#define ENABLE_USB
 //#define ENABLE_NETWORK
 
+typedef struct {
+  bool enable_debug;
+  bool enable_fs;
+  bool enable_ide;
+  bool enable_usb;
+  bool enable_networking;
+} kconf_t;
 
 static inline kconf_t get_kconfig(void)
 {
diff --git a/include/kincl.h b/include/kincl.h
new file mode 100644
index 0000000..6785c75
--- /dev/null
+++ b/include/kincl.h
@@ -0,0 +1,6 @@
+#ifndef _KERNEL_INCLUDES_H
+#define _KERNEL_INCLUDES_H
+
+#define KERNEL_VERSION "0.0.1e"
+
+#endif
diff --git a/include/panic.h b/include/panic.h
deleted file mode 100644
index f96cda6..0000000
--- a/include/panic.h
+++ /dev/null
@@ -1,30 +0,0 @@
-#ifndef _KERNEL_PANIC_H
-#define _KERNEL_PANIC_H
-
-#include <stddef.h>
-#include <tty.h>
-
-void panic(const char* arr)
-{
-  
-  uint8_t color = 0x1F;
-  uint8_t blank = 0xFF;
-  
-  for (size_t y = 0; y < VGA_HEIGHT; y++)
-  {
-    for (size_t x = 0; x < VGA_WIDTH; x++)
-    {
-      const size_t index = y * VGA_WIDTH + x;
-      VGA_MEMORY[index] = (uint16_t) ' ' | (uint16_t) blank << 8;
-    }
-  }
-  
-  int i;
-  for (i = 0; i < 75; i += 2)
-  {
-    VGA_MEMORY[i] = arr[i];
-    VGA_MEMORY[i + 1] = color;
-  }
-}
-
-#endif
diff --git a/include/syscall.h b/include/syscall.h
deleted file mode 100644
index e3f9d1a..0000000
--- a/include/syscall.h
+++ /dev/null
@@ -1,6 +0,0 @@
-#ifndef _SYSCALL_H
-#define _SYSCALL_H
-
-
-
-#endif
diff --git a/kernel/kernel.c b/kernel/kernel.c
index d0a76b0..05a8e5a 100644
--- a/kernel/kernel.c
+++ b/kernel/kernel.c
@@ -45,6 +45,15 @@
 extern void _hang_asm(void);
 extern void _sti_asm(void);
 
+char* espresso_str = ""
+"#######  #####  ######  ######  #######  #####   #####  #######\n"
+"#       #     # #     # #     # #       #     # #     # #     #\n"
+"#       #       #     # #     # #       #       #       #     #\n"
+"#####    #####  ######  ######  #####    #####   #####  #     #\n"
+"#             # #       #   #   #             #       # #     #\n"
+"#       #     # #       #    #  #       #     # #     # #     #\n"
+"#######  #####  #       #     # #######  #####   #####  #######\n";
+
 
 void kernel_main(multiboot_info_t* mbd, uint32_t magic)
 {
@@ -142,16 +151,23 @@ void kernel_main(multiboot_info_t* mbd, uint32_t magic)
   /*pit_sleep(4000);
   begin_anim(espresso_kernel_version);*/
   
-  /*printf("here\n");
+  printf("Guten tag and welcome to Espresso %s\n", espresso_kernel_version);
+ 
+  printf("%s\n", espresso_str);
+ 
+  char buffer[512] = { 0 };
   
-  printf("%i\n", sfs_init(false));
+  int32_t i = sfs_read_file("test.txt", buffer, -1);
   
-  printf("here\n");*/
+  printf("%i\n", i);
   
-  printf("Guten tag and welcome to Espresso\n");
+  if (i == 0)
+  {
+    printf("%s\n", buffer);
+  }
   
   while (true)
   {
-    
+    /* Loop infinitely. We only do something when a syscall happens now. */
   }
 }
diff --git a/lib/mm/paging.c b/lib/mm/paging.c
index 6e45cf5..c20d73f 100644
--- a/lib/mm/paging.c
+++ b/lib/mm/paging.c
@@ -23,7 +23,7 @@ void paging_init(void)
   page_directory = (page_directory_entry_t*)alloc_page();
   memset(page_directory, 0, PAGE_SIZE);
 
-  /* Allocate and set up the first identity-mapped page table (0–4MB) */
+  /* Allocate and set up the first identity-mapped page table (0-4MB) */
   page_tables[0] = (page_table_entry_t*)alloc_page();
   memset(page_tables[0], 0, PAGE_SIZE);
   for (uint32_t i = 0; i < PAGE_TABLE_ENTRIES; i++)
diff --git a/lib/vector_extensions/sse.c b/lib/vector_extensions/sse.c
index befb329..d0ebba5 100644
--- a/lib/vector_extensions/sse.c
+++ b/lib/vector_extensions/sse.c
@@ -99,41 +99,43 @@ void sse2_add_int32_arrays(int32_t *dst, const int32_t *a, const int32_t *b, siz
 __attribute__((force_align_arg_pointer))
 void *sse2_memcpy(void *dst, const void *src, uint32_t n)
 {
-  uintptr_t i;
+  uint8_t *d = (uint8_t *)dst;
+  const uint8_t *s = (const uint8_t *)src;
+  uintptr_t i = 0;
 
-  /* Align to 16 bytes if necessary */
-  while (((uintptr_t)dst & 15) && n > 0)
+  /* Align destination to 16 bytes */
+  while (((uintptr_t)d & 15) && n > 0)
   {
-    *((uint8_t*)dst) = *((uint8_t*)src);
-    dst = (uint8_t*)dst + 1;
-    src = (uint8_t*)src + 1;
+    *d++ = *s++;
     n--;
   }
 
-  /* Copy 16-byte chunks with SSE2 */
-  for (i = 0; i + 15 < n; i += 16)
+  /* Use SSE2 for 16-byte aligned blocks */
+  for (; i + 15 < n; i += 16)
   {
     asm volatile (
-      "movdqa (%1), %%xmm0\n\t"  /* Load 16 bytes from source into xmm0 */
-      "movdqa %%xmm0, (%0)\n\t"  /* Store 16 bytes to destination */
+      "movdqa (%1), %%xmm0\n\t"
+      "movdqa %%xmm0, (%0)\n\t"
       :
-      : "r"(dst + i), "r"(src + i)
+      : "r"(d + i), "r"(s + i)
       : "xmm0", "memory"
     );
   }
 
-  /* Handle remaining bytes (less than 16) */
-  while (n > 0)
+  d += i;
+  s += i;
+  n -= i;
+
+  /* Copy any remaining bytes */
+  while (n--)
   {
-    *((uint8_t*)dst) = *((uint8_t*)src);
-    dst = (uint8_t*)dst + 1;
-    src = (uint8_t*)src + 1;
-    n--;
+    *d++ = *s++;
   }
 
-  return dst;  /* Return pointer to destination */
+  return dst;
 }
 
+
 __attribute__((force_align_arg_pointer))
 char *sse2_strncpy(char *dest, const char *src, uint32_t n)
 {