diff --git a/src/arch/x86_64/syscall.asm b/src/arch/x86_64/syscall.asm
index 9a72747a..f89398ba 100644
--- a/src/arch/x86_64/syscall.asm
+++ b/src/arch/x86_64/syscall.asm
@@ -14,7 +14,8 @@ extern syscall_write
 extern syscall_stat
 extern syscall_fstat
 extern syscall_fork
-func_count equ 13
+extern syscall_execve
+func_count equ 14
 
 
 func_table:
@@ -31,6 +32,7 @@ func_table:
 	dq syscall_stat
 	dq syscall_fstat
 	dq syscall_fork
+	dq syscall_execve
 section .text
 global asm_syscall_entry
 
diff --git a/src/sys/syscalls/execve.c b/src/sys/syscalls/execve.c
new file mode 100644
index 00000000..c2f2aa26
--- /dev/null
+++ b/src/sys/syscalls/execve.c
@@ -0,0 +1,156 @@
+#include <kernel/syscalls.h>
+#include <kernel/vmm.h>
+#include <errno.h>
+#include <kernel/alloc.h>
+#include <string.h>
+#include <kernel/vfs.h>
+#include <kernel/sched.h>
+#include <arch/spinlock.h>
+#include <kernel/elf.h>
+#include <arch/cls.h>
+#include <arch/interrupt.h>
+
+static void destroyvec(char* vec[], size_t vecsize){
+	for(uintmax_t p = 0; p < vecsize; ++p){
+		if(vec[p])
+			free(vec[p]);
+	}
+
+	free(vec);
+
+}
+
+syscallret syscall_execve(const char* name, char* const argv[], char* const envp[]){
+	
+	syscallret retv;
+	retv.ret = -1;
+	
+	if(name > USER_SPACE_END || argv > USER_SPACE_END || envp > USER_SPACE_END){
+		retv.errno = EFAULT;
+		return retv;
+	}
+
+	size_t namelen = strlen(name); // XXX user strlen
+	size_t argc = 0;
+	size_t envc = 0;
+
+	// this probably should be checked in a safer way
+	
+	while(argv[argc]){
+		++argc;
+	}
+
+	while(envp[envc]){
+		++envc;
+	}
+
+	retv.errno = ENOMEM;
+
+	char** argbuff = alloc((argc+1)*sizeof(char*));
+	if(!argbuff)
+		return retv;
+	
+
+	char** envbuff = alloc((envc+1)*sizeof(char*));
+
+	if(!envbuff){
+		free(argbuff);
+		return retv;
+	}
+
+	char* namebuff = alloc(namelen+1);
+
+	if(!namebuff){
+		free(argbuff);
+		free(envbuff);
+		return retv;
+	}
+
+	strcpy(namebuff, name); // XXX user strcpy
+
+	int err = 0;
+
+	for(uintmax_t arg = 0; arg < argc && err == 0; ++arg){
+		size_t len = strlen(argv[arg]); // XXX user strlen
+		argbuff[arg] = alloc(len+1);
+		if(!argbuff[arg]){
+			err = ENOMEM;
+			break;
+		}
+		strcpy(argbuff[arg], argv[arg]); // XXX user strcpy
+	}
+
+	for(uintmax_t env = 0; env < envc && err == 0; ++env){
+		size_t len = strlen(envp[env]); // XXX user strlen
+		envbuff[env] = alloc(len+1);
+		if(!envbuff[env]){
+			err = ENOMEM;
+			break;
+		}
+		strcpy(envbuff[env], envp[env]); // XXX user strcpy	
+	}
+
+	if(err){
+		retv.errno = err;
+		goto _clean;
+	}
+
+	thread_t* thread = arch_getcls()->thread;
+
+	vmm_context* oldctx = thread->ctx;
+	vmm_context* ctx = vmm_newcontext();
+
+	if(!ctx){
+		retv.errno = ENOMEM;
+		goto _clean;
+	}
+
+	vmm_switchcontext(ctx);
+	
+	spinlock_acquire(&thread->proc->lock);
+
+	vnode_t* node;
+
+	err = vfs_open(&node, namebuff[0] == '/' ? thread->proc->root : thread->proc->cwd, 
+			namebuff);
+	
+	if(err){
+		spinlock_release(&thread->proc->lock);
+		retv.errno = err;
+		return retv;
+	}
+
+	void *entry, *stack;
+
+	err = elf_load(thread, node, argbuff, envbuff, &stack, &entry);
+
+	if(err){
+		vfs_close(node);
+		spinlock_release(&thread->proc->lock);
+		retv.errno = err;
+		goto _clean;
+	}
+
+	arch_interrupt_disable();
+
+	arch_regs_setupuser(thread->regs, entry, stack, true);
+	
+	// TODO stop other threads and destroy old context
+		
+	destroyvec(argbuff, argc+1);
+	destroyvec(envbuff, envc+1);
+	free(namebuff);
+	vfs_close(node);
+	spinlock_release(&thread->proc->lock);
+
+	switch_thread(thread);
+
+	_clean:
+	
+	destroyvec(argbuff, argc+1);
+	destroyvec(envbuff, envc+1);
+	free(namebuff);
+	return retv;
+
+
+}