如何在xv6中实现内核级线程的克隆

Question

While attempting to test my clone syscall (code below), I keep getting the following errors:

pid 4 thread_test: trap 13 err 0 on cpu 1 eip 0xc54 addr 0x0--kill proc pid 5 thread_test: trap 14 err 4 on cpu 1 eip 0x0 addr 0x28ec83e5--kill proc

Which correspond to a general protection fault and a page fault respectively. Does anyone have any idea what might be causing the new threads to be killed right after creation?

int clone(void *(*func) (void *), void *arg, void *stack)
{
  int i,pid;
  struct proc *np;

  // Allocate process.
  if((np = allocproc()) == 0)
    return -1;

  np->state = UNUSED;
  np->sz = proc->sz;
  np->parent = proc;
  *np->tf = *proc->tf;
  np->pgdir = proc->pgdir;

  np->tf->eax = 0;   // Clear %eax so that fork returns 0 in the child.
  np->tf->eip = (int)func;   //change eip to new function
  np->kstack = stack;  //use given stack

  for(i = 0; i < NOFILE; i++)
    if(proc->ofile[i])
      np->ofile[i] = filedup(proc->ofile[i]);
  np->cwd = idup(proc->cwd);

  np->tf->esp = (uint)(stack+PGSIZE-4); //put esp to right spot on stack
  *((uint*)(np->tf->esp)) = (uint)arg; //arg to function
  *((uint*)(np->tf->esp)-4) = 0xFFFFFFFF; //return to nowhere
  np->tf->esp =(np->tf->esp) -4;

  safestrcpy(np->name, proc->name, sizeof(proc->name));
  pid = np->pid;

  acquire(&ptable.lock);  //lock so writes last
  np->state = RUNNABLE;
  release(&ptable.lock);

  return pid;
}

Answer 1

I found the solution, I was trying to use the passed in stack as the thread's kstack. I needed to create a separate stack in proc.h and then assign the stack to the thread using:

np->stack = (int)stack;