diff --git a/initrd/make_img.sh b/initrd/make_img.sh
index 444456d..dd5dd03 100755
--- a/initrd/make_img.sh
+++ b/initrd/make_img.sh
@@ -2,4 +2,5 @@
 
 mkdir -p output
 rm -rf output/*
-find . -depth -print | cpio -o -H newc > output/initrd.img
\ No newline at end of file
+find . -depth -print | cpio -o -H newc > output/initrd.img
+cp output/initrd.img ../isodir/boot
\ No newline at end of file
diff --git a/kernel/common.S b/kernel/common.S
index 5a8dc6c..2b1b563 100644
--- a/kernel/common.S
+++ b/kernel/common.S
@@ -120,4 +120,6 @@ outsw:
 .global fetch_cr3
 fetch_cr3: 
     mov %cr3, %eax 
-    ret
\ No newline at end of file
+    ret
+
+.global set_cr3
diff --git a/kernel/drivers/cpio/cpio.c b/kernel/drivers/cpio/cpio.c
index da39cd9..f5a15ba 100644
--- a/kernel/drivers/cpio/cpio.c
+++ b/kernel/drivers/cpio/cpio.c
@@ -34,7 +34,7 @@ static uint32_t cpio_read(struct vfs_node* node, uint32_t offset, uint32_t size,
     if (offset + size > node->size) {
         size = node->size - offset;
     }
-    memcpy(buffer, (uint8_t*)(node->ptr) + offset, size);
+    memcpy((uint8_t*)(node->data) + offset, buffer, size);
     return size;
 }
 
@@ -80,7 +80,7 @@ vfs_node_t* cpio_read_fs(uint32_t _vaddr) {
 
             current->flags = VFS_FILE;
             current->read = cpio_read;
-            current->ptr = (vfs_node_t*)file_data;
+            current->data = file_data;
         }
 
         uint32_t next_header_offset = file_data_offset + filesize;
diff --git a/kernel/gdt.c b/kernel/gdt.c
index c6bff54..1b18673 100644
--- a/kernel/gdt.c
+++ b/kernel/gdt.c
@@ -1,4 +1,4 @@
-#include <stdint.h>
+#include <gdt.h>
  
 #define SEG_DESCTYPE(x)  ((x) << 0x04) // Descriptor type (0 for system, 1 for code/data)
 #define SEG_PRES(x)      ((x) << 0x07) // Present
@@ -45,41 +45,11 @@ typedef struct __attribute__((packed)) {
     uint16_t limit;
     uint32_t base;
 } gdt_ptr_t;
-
-typedef struct __attribute__((packed)) {
-    uint32_t prev_tss;   
-    uint32_t esp0;       
-    uint32_t ss0;        
-    uint32_t esp1;       
-    uint32_t ss1;       
-    uint32_t esp2;      
-    uint32_t ss2;        
-    uint32_t cr3;       
-    uint32_t eip;      
-    uint32_t eflags;     
-    uint32_t eax;       
-    uint32_t ecx;        
-    uint32_t edx;      
-    uint32_t ebx;     
-    uint32_t esp;       
-    uint32_t ebp;      
-    uint32_t esi;   
-    uint32_t edi;    
-    uint32_t es;         
-    uint32_t cs;         
-    uint32_t ss;         
-    uint32_t ds;      
-    uint32_t fs;        
-    uint32_t gs;
-    uint32_t ldt;
-    uint16_t trap;
-    uint16_t iomap_base;
-} tss_entry;
                      
 uint64_t gdt[6];
 gdt_ptr_t gdtr;
 
-tss_entry tss;
+tss_entry_t tss;
 
 extern void load_gdt();
 extern void tss_flush();
@@ -126,14 +96,14 @@ void init_gdt() {
     gdt[3] = create_descriptor(0, 0x000FFFFF, (GDT_CODE_PL3));
     gdt[4] = create_descriptor(0, 0x000FFFFF, (GDT_DATA_PL3));
 
-    for (uint32_t i = 0; i < sizeof(tss_entry); i++) {
+    for (uint32_t i = 0; i < sizeof(tss_entry_t); i++) {
         ((char*)&tss)[i] = 0;
     }
 
     tss.ss0 = 0x10;
-    tss.iomap_base = sizeof(tss_entry);
+    tss.iomap_base = sizeof(tss_entry_t);
 
-    gdt[5] = create_tss_descriptor((uint32_t)&tss, sizeof(tss_entry) - 1);
+    gdt[5] = create_tss_descriptor((uint32_t)&tss, sizeof(tss_entry_t) - 1);
 
     gdtr.limit = 6 * sizeof(uint64_t) - 1;
     gdtr.base = (uint32_t)&gdt;
diff --git a/kernel/gdt.h b/kernel/gdt.h
index f90e54d..43104cf 100644
--- a/kernel/gdt.h
+++ b/kernel/gdt.h
@@ -1,6 +1,40 @@
 #ifndef KGDT_H
 #define KGDT_H
 
+#include <stdint.h>
+
+typedef struct __attribute__((packed)) {
+    uint32_t prev_tss;   
+    uint32_t esp0;       
+    uint32_t ss0;        
+    uint32_t esp1;       
+    uint32_t ss1;       
+    uint32_t esp2;      
+    uint32_t ss2;        
+    uint32_t cr3;       
+    uint32_t eip;      
+    uint32_t eflags;     
+    uint32_t eax;       
+    uint32_t ecx;        
+    uint32_t edx;      
+    uint32_t ebx;     
+    uint32_t esp;       
+    uint32_t ebp;      
+    uint32_t esi;   
+    uint32_t edi;    
+    uint32_t es;         
+    uint32_t cs;         
+    uint32_t ss;         
+    uint32_t ds;      
+    uint32_t fs;        
+    uint32_t gs;
+    uint32_t ldt;
+    uint16_t trap;
+    uint16_t iomap_base;
+} tss_entry_t;
+
 void init_gdt();
 
+extern tss_entry_t tss;
+
 #endif
\ No newline at end of file
diff --git a/kernel/initrd.c b/kernel/initrd.c
deleted file mode 100644
index 501723d..0000000
--- a/kernel/initrd.c
+++ /dev/null
@@ -1,34 +0,0 @@
-#include "initrd.h"
-
-#include <stdint.h>
-
-#include <lib/print.h>
-#include <lib/memory.h>
-
-void parse_cpio(void *initrd_start) {
-    cpio_header_t *head = (cpio_header_t *)initrd_start;
-
-    while (strncmp(head->c_magic, "070701", 6) == 0) {
-        // If we hit the trailer, we are done
-        char *filename = (char *)(head + 1); 
-        if (strncmp(filename, "TRAILER!!!", 10) == 0) {
-            break; 
-        }
-
-        uint32_t filesize = parse_hex(head->c_filesize);
-        uint32_t namesize = parse_hex(head->c_namesize);
-        
-        kprintf("found file of size %d\n", filesize);
-
-        // Calculate where the data begins (accounting for 4-byte padding)
-        uint32_t header_plus_name = 110 + namesize;
-        header_plus_name = (header_plus_name + 3) & ~3; // Align to 4 bytes
-        
-        char *file_data = (char *)head + header_plus_name;
-
-        uint32_t total_record_size = header_plus_name + filesize;
-        total_record_size = (total_record_size + 3) & ~3; // Align to 4 bytes
-        
-        head = (cpio_header_t *)((char *)head + total_record_size);
-    }
-}
\ No newline at end of file
diff --git a/kernel/initrd.h b/kernel/initrd.h
deleted file mode 100644
index 58981f9..0000000
--- a/kernel/initrd.h
+++ /dev/null
@@ -1,36 +0,0 @@
-#ifndef KINITRD_H
-#define KINITRD_H
-
-#include <stdint.h>
-
-typedef struct {
-    char c_magic[6];
-    char c_ino[8];
-    char c_mode[8];
-    char c_uid[8];
-    char c_gid[8];
-    char c_nlink[8];
-    char c_mtime[8];
-    char c_filesize[8];
-    char c_devmajor[8];
-    char c_devminor[8];
-    char c_rdevmajor[8];
-    char c_rdevminor[8];
-    char c_namesize[8];
-    char c_check[8];
-} cpio_header_t;
-
-static inline uint32_t parse_hex(const char *s) {
-    uint32_t val = 0;
-    for (int i = 0; i < 8; i++) {
-        val <<= 4;
-        if (s[i] >= '0' && s[i] <= '9')      val += (s[i] - '0');
-        else if (s[i] >= 'A' && s[i] <= 'F') val += (s[i] - 'A' + 10);
-        else if (s[i] >= 'a' && s[i] <= 'f') val += (s[i] - 'a' + 10);
-    }
-    return val;
-}
-
-void parse_cpio(void *initrd_start);
-
-#endif
\ No newline at end of file
diff --git a/kernel/interrupts.S b/kernel/interrupts.S
index 00e6d47..a73bcdd 100644
--- a/kernel/interrupts.S
+++ b/kernel/interrupts.S
@@ -44,8 +44,16 @@ ISR_ERRCODE   14   # Page Fault (Has Error Code!)
 .extern send_eoi_master     # drivers/pic/pic.c
 isr32:
     cli
-    
     pushal 
+
+    movw %ds, %ax
+    pushl %eax
+
+    movw $0x10, %ax # Load Kernel Data Segment descriptor
+    movw %ax, %ds
+    movw %ax, %es
+    movw %ax, %fs
+    movw %ax, %gs
     
     pushl %esp 
     call switch_context
@@ -53,6 +61,12 @@ isr32:
 
     call send_eoi_master
 
+    popl %eax
+    movw %ax, %ds
+    movw %ax, %es
+    movw %ax, %fs
+    movw %ax, %gs
+
     popal
     sti
     iret 
diff --git a/kernel/kmain.c b/kernel/kmain.c
index f0c43b4..1b714f7 100644
--- a/kernel/kmain.c
+++ b/kernel/kmain.c
@@ -4,6 +4,7 @@
 #include "scheduler.h"
 #include "common.h"
 #include "vfs.h"
+#include "elf.h"
 
 #include <lib/print.h>
 #include <lib/stream.h>
@@ -51,26 +52,73 @@ vfs_node_t* load_initrd() {
   return cpio_read_fs(initrd_module->mod_start);
 }
 
-void load_origin_program() {
-  vfs_node_t* vfs_root = vfs_create_root();
-  vfs_node_t* initrd_root = load_initrd();
-  
-  if (initrd_root) {
-    mount(vfs_root, initrd_root);
-    kprintf("initrd read, mounted @ /\n");
+void load_program(const char* path) {
+  void* cr3 = create_task_pd();
+  load_pd(cr3);
+  flush_tlb();
 
-    int fd = open("/origin.elf", 0);
-    if (fd == -1) {
-      kprintf("failed to read origin program\n");
-    }
-
-    void* buf = kalloc(512);
-    int rd = read(fd, buf, 512);
-    if (rd > 0) {
-      kprintf("read %d bytes from origin.elf\n", rd);
-    }
+  int fd = open(path, 0);
+  if (fd == -1) {
+    kprintf("failed to open program: %s\n", path);
+    return;;
   }
-  exit();
+  elf_header_t* e_hdr = kalloc(sizeof(elf_header_t));
+  memset((void*)e_hdr, 0, sizeof(elf_header_t));
+
+  uint32_t bytes_rd = read(fd, e_hdr, sizeof(elf_header_t));
+  if (bytes_rd <= 0) {
+    kprintf("failed to read program data for %s\n", path);
+    return;
+  }
+
+  uint32_t ph_table_ptr = e_hdr->e_phoff;
+  elf_program_header_t* ph = kalloc(sizeof(elf_program_header_t));
+  uint32_t heap_start = 0;
+
+  for (uint32_t i = 0; i < e_hdr->e_phnum; i++) {
+    fseek(fd, ph_table_ptr);
+    memset((void*)ph, 0, sizeof(elf_program_header_t));
+    read(fd, ph, sizeof(elf_program_header_t));
+
+    if (ph->p_type == PT_LOAD) {
+      uint32_t vstart = ph->p_vaddr;
+      uint32_t vend = vstart + ph->p_memsz;
+
+      if (vend > heap_start) {
+        heap_start = vend;
+      }
+
+      uint32_t page_start = vstart & ~(PAGE_SIZE - 1);
+      uint32_t page_end = (vend + PAGE_SIZE - 1) & ~(PAGE_SIZE - 1);
+
+      for (uint32_t v = page_start; v < page_end; v += PAGE_SIZE) {
+          void* phys_frame = p_alloc_frame();
+          map_page(v, (uint32_t)phys_frame, 0x07);
+      }
+
+      uint32_t data_ptr = ph->p_offset;
+      fseek(fd, data_ptr);
+      read(fd, (void*)vstart, ph->p_memsz);
+
+      // 0 out bss
+      if (ph->p_memsz > ph->p_filesz) {
+        size_t bss_size = ph->p_memsz - ph->p_filesz;
+        memset((void*)(vstart + ph->p_filesz), 0, bss_size);
+      }
+    }
+
+    ph_table_ptr += sizeof(elf_program_header_t);
+  }
+
+  uint32_t esp = 0xBFFF0000;
+  void* pesp = p_alloc_frame();
+  map_page(esp - PAGE_SIZE, (uint32_t)pesp, 0x07);
+
+  create_task(e_hdr->e_entry, fetch_cr3(), 1);
+}
+
+void start_origin() {
+  load_program("/origin.elf");
 }
 
 void kmain(uint32_t magic, multiboot_info* mbi)  {
@@ -96,9 +144,19 @@ void kmain(uint32_t magic, multiboot_info* mbi)  {
   
   kprintf("basic initializations complete\n");
 
+  vfs_node_t* vfs_root = vfs_create_root();
+
+  vfs_node_t* initrd_root = load_initrd();
+  if (!initrd_root) {
+    kprintf("failed to load initrd\n");
+  }
+  mount(vfs_root, initrd_root);
+  kprintf("initrd read, mounted @ /\n");
+
   disable_interrupts();
   init_scheduler();
   enable_interrupts();
 
-  create_task((uint32_t)load_origin_program, fetch_cr3());
+  kprintf("loading user program\n");
+  start_origin();
 }
diff --git a/kernel/lib/string.h b/kernel/lib/string.h
new file mode 100644
index 0000000..70dcf7d
--- /dev/null
+++ b/kernel/lib/string.h
@@ -0,0 +1,6 @@
+#ifndef KSTRING_H
+#define KSTRING_H
+
+char* strtok_r(char* str, const char* delim, char** saveptr);
+
+#endif
\ No newline at end of file
diff --git a/kernel/lib/strtok.c b/kernel/lib/strtok.c
new file mode 100644
index 0000000..31aa210
--- /dev/null
+++ b/kernel/lib/strtok.c
@@ -0,0 +1,49 @@
+#include <lib/string.h>
+
+#include <stddef.h>
+
+static int is_delim(char c, const char* delim) {
+    while (*delim) {
+        if (c == *delim) {
+            return 1;
+        }
+        delim++;
+    }
+    return 0;
+}
+
+char* strtok_r(char* str, const char* delim, char** saveptr) {
+    if (saveptr == NULL) {
+        return NULL;
+    }
+
+    char* next_token = (str != NULL) ? str : *saveptr;
+
+    if (next_token == NULL || *next_token == '\0') {
+        *saveptr = NULL;
+        return NULL;
+    }
+
+    while (*next_token && is_delim(*next_token, delim)) {
+        next_token++;
+    }
+
+    if (*next_token == '\0') {
+        *saveptr = next_token;
+        return NULL;
+    }
+
+    char* token_start = next_token;
+
+    while (*next_token) {
+        if (is_delim(*next_token, delim)) {
+            *next_token = '\0';        // Destructively null-terminate the token
+            *saveptr = next_token + 1; // Save the position right after the null-terminator
+            return token_start;
+        }
+        next_token++;
+    }
+
+    *saveptr = next_token;
+    return token_start;
+}
\ No newline at end of file
diff --git a/kernel/memory/mm_stubs.S b/kernel/memory/mm.S
similarity index 100%
rename from kernel/memory/mm_stubs.S
rename to kernel/memory/mm.S
diff --git a/kernel/memory/mm.c b/kernel/memory/mm.c
index 85118d9..2d3c175 100644
--- a/kernel/memory/mm.c
+++ b/kernel/memory/mm.c
@@ -8,8 +8,6 @@ extern uint8_t __kernel_start;
 extern uint8_t __kernel_end;
 
 extern uint32_t page_directory[1024];
-extern void flush_tlb();
-extern void disable_pae();
 
 #define PAGE_SIZE 4096
 #define MAX_RAM   0xFFFFFFFF // 4GB
diff --git a/kernel/memory/mm.h b/kernel/memory/mm.h
index 456d8c2..7d73dfc 100644
--- a/kernel/memory/mm.h
+++ b/kernel/memory/mm.h
@@ -24,4 +24,9 @@ void* create_task_pd();
 
 void map_page(uint32_t vaddr, uint32_t paddr, uint32_t flags);
 
+void load_pd(void* paddr);
+
+void flush_tlb();
+void disable_pae();
+
 #endif 
diff --git a/kernel/process.h b/kernel/process.h
index 5962621..689f691 100644
--- a/kernel/process.h
+++ b/kernel/process.h
@@ -13,6 +13,7 @@ typedef enum {
 typedef struct {
     uint32_t pid;
     uint32_t esp;
+    uint32_t esp0;
     uint32_t cr3;
     process_state_t state;
     uint32_t sleep;
diff --git a/kernel/scheduler.c b/kernel/scheduler.c
index 0986a73..ef70e01 100644
--- a/kernel/scheduler.c
+++ b/kernel/scheduler.c
@@ -8,6 +8,7 @@
 #include "common.h"
 #include "process.h"
 #include "scheduler.h"
+#include "gdt.h"
 
 static process_t* process_table;
 
@@ -68,9 +69,10 @@ uint32_t switch_context(uint32_t current_esp) {
 
     current_process = next_process;
 
+    tss.esp0 = process_table[current_process].esp0;
+
     if (fetch_cr3() != process_table[current_process].cr3) {
-        extern void load_pd(uint32_t table);
-        load_pd(process_table[current_process].cr3);
+        load_pd((void*)process_table[current_process].cr3);
     }
 
     return process_table[current_process].esp;
@@ -88,12 +90,12 @@ void init_scheduler() {
     process->sleep = 0;
     total_processes++;
 
-    create_task((uint32_t)idle_task, fetch_cr3());
+    create_task((uint32_t)idle_task, fetch_cr3(), 0);
 
     current_process = 0;
 }
 
-int create_task(uint32_t entry_point, uint32_t cr3) {
+int create_task(uint32_t entry_point, uint32_t cr3, int user) {
     if (total_processes >= 16) {
         return 0;
     }
@@ -117,10 +119,20 @@ int create_task(uint32_t entry_point, uint32_t cr3) {
 
     void* stack_bottom = kalloc(4096);
     uint32_t* esp = (uint32_t*)((uint32_t)stack_bottom + 4096);
+    process->esp0 = (uint32_t)esp;
 
-    *(--esp) = 0x0202;        // EFLAGS   
-    *(--esp) = 0x08;          // CS
-    *(--esp) = entry_point;   // EIP
+    if (user) {
+        *(--esp) = 0x23;            // User Data Segment (SS) with RPL 3 (0x20 | 3)
+        *(--esp) = 0xBFFFF000;      // User Stack Pointer (ESP) - location mapped in user space
+        *(--esp) = 0x0202;          // EFLAGS (Interrupts enabled)
+        *(--esp) = 0x1B;            // User Code Segment (CS) with RPL 3 (0x18 | 3)
+        *(--esp) = entry_point;     // EIP
+    } else {
+        // IRET values for Ring 0
+        *(--esp) = 0x0202;          // EFLAGS
+        *(--esp) = 0x08;            // Kernel Code Segment (CS)
+        *(--esp) = entry_point;     // EIP
+    }
 
     *(--esp) = 0;             // EAX
     *(--esp) = 0;             // ECX
diff --git a/kernel/scheduler.h b/kernel/scheduler.h
index 437290b..20616d0 100644
--- a/kernel/scheduler.h
+++ b/kernel/scheduler.h
@@ -6,7 +6,7 @@
 
 void init_scheduler();
 
-int create_task(uint32_t entry_point, uint32_t cr3);
+int create_task(uint32_t entry_point, uint32_t cr3, int user);
 process_t* current_task();
 
 void sleep(uint32_t ms);
diff --git a/kernel/vfs.c b/kernel/vfs.c
index 4fa7d37..704f1cd 100644
--- a/kernel/vfs.c
+++ b/kernel/vfs.c
@@ -22,22 +22,22 @@ vfs_node_t* vfs_create_root() {
 
 void mount(vfs_node_t* _mnt, vfs_node_t* _fs) {
     if (!_fs || !_mnt || !(_mnt->flags & VFS_DIRECTORY)) return;
-    _mnt->ptr = _fs;
+    _mnt->mnt = _fs;
 }
 
 static vfs_node_t* lookup(const char* path) {
     if (path[0] != '/') return NULL;
     vfs_node_t* current = root;
 
-    if (current->ptr) {
-        current = current->ptr;
+    if (current->mnt) {
+        current = current->mnt;
     }
 
-    const char* component = path + 1;
-    if (*component == '\0') return current;
+    const char* cmp = path + 1;
+    if (*cmp == '\0') return current;
 
     if (current->finddir) {
-        return current->finddir(current, component);
+        return current->finddir(current, cmp);
     }
 
     return NULL;
@@ -72,6 +72,22 @@ int read(int fd, void* buf, size_t sz) {
     return bytes_read;
 }
 
-int fseek(int fd) {
-    
-}
\ No newline at end of file
+int fstat(int fd, file_t* dst) {
+    if (fd < 0 || fd >= MAX_PROCESS_FDS || !fd_table[fd].node) {
+        return 0;
+    }
+
+    file_t* src = &fd_table[fd];
+    memcpy(src, dst, sizeof(file_t));
+    return 1;
+}
+
+int fseek(int fd, size_t offset) {
+    if (fd < 0 || fd >= MAX_PROCESS_FDS || !fd_table[fd].node) {
+        return 0;
+    }
+
+    file_t* file = &fd_table[fd];
+    file->offset = offset;
+    return 1;
+}   
\ No newline at end of file
diff --git a/kernel/vfs.h b/kernel/vfs.h
index d9059e1..4530658 100644
--- a/kernel/vfs.h
+++ b/kernel/vfs.h
@@ -25,7 +25,8 @@ typedef struct vfs_node {
     write_type_t write;
     finddir_type_t finddir; 
 
-    struct vfs_node* ptr;
+    struct vfs_node* mnt;
+    void* data;
 } vfs_node_t;
 
 typedef struct {
@@ -40,4 +41,7 @@ void mount(vfs_node_t* _mnt, vfs_node_t* _fs);
 int open(const char* path, int flags);
 int read(int fd, void* buf, size_t sz);
 
+int fstat(int fd, file_t* _f);
+int fseek(int fd, size_t offset);
+
 #endif
\ No newline at end of file
diff --git a/programs/makefile b/programs/makefile
index c88a74f..9d074de 100644
--- a/programs/makefile
+++ b/programs/makefile
@@ -1,8 +1,8 @@
 CC = i686-elf-gcc
 LD = i686-elf-ld
 
-CFLAGS = -ffreestanding -O2 -Wall -Wextra -I../rocklibc/include
-LDFLAGS = -m elf_i386 -nostdlib
+CFLAGS = -ffreestanding -O2 -Wall -Wextra -ffunction-sections -fdata-sections -I../rocklibc/include
+LDFLAGS = -m elf_i386 -nostdlib --gc-sections
 
 TARGET = origin.elf
 
@@ -16,6 +16,7 @@ main.o: main.c
 
 $(TARGET): crt0.o main.o ../rocklibc/rlibc.a
 	$(LD) $(LDFLAGS) -T linker.ld crt0.o main.o ../rocklibc/rlibc.a $(shell $(CC) -print-libgcc-file-name) -o $(TARGET)
+	cp $(TARGET) ../initrd
 
 clean:
 	rm -f *.o $(TARGET)
diff --git a/rocklibc/makefile b/rocklibc/makefile
index efc4608..f5a54fe 100644
--- a/rocklibc/makefile
+++ b/rocklibc/makefile
@@ -1,6 +1,7 @@
 CC = i686-elf-gcc
 AR = i686-elf-ar
 CFLAGS = -ffreestanding -O2 -Wall -Wextra -Iinclude -DHAVE_MMAP=0 -DLACKS_SYS_MMAN_H=1 -DUSE_LOCKS=0 -DLACKS_SYS_PARAM_H=1
+CFLAGS += -ffunction-sections -fdata-sections
 TARGET = rlibc.a
 
 SRCS = $(shell find src -name "*.c")