proc.c

#include "types.h"
#include "defs.h"
#include "param.h"
#include "memlayout.h"
#include "mmu.h"
#include "x86.h"
#include "proc.h"
#include "spinlock.h"
#include "errno.h"
#include "list.h"
#include "fs.h"
#include "file.h"
#include "stat.h"
#include "err.h"
#include "buf.h"

struct ptable ptable;

static struct proc *initproc;

struct cache_info* proc_cache;
struct cache_info* mm_cache;
struct cache_info* files_struct_cache;
struct cache_info* fs_info_cache;
struct cache_info* mmap_cache;
struct cache_info* mmap_list_cache;

int nextpid = 1;
extern void forkret(void);
extern void trapret(void);

static void wakeup1(void *chan);

void
pinit(void)
{
  initlock(&ptable.lock, "ptable");
  proc_cache = kmem_cache_create(sizeof(struct proc));
  if (proc_cache == 0) {
    panic("Could not allocate proc cache");
  }
  mm_cache = kmem_cache_create(sizeof(struct mm_struct));
  if (mm_cache == 0) {
    panic("Could not allocate mm_struct cache");
  }
  files_struct_cache = kmem_cache_create(sizeof(struct files_struct));
  if (files_struct_cache == 0) {
    panic("Could not allocate files_struct cache");
  }
  fs_info_cache = kmem_cache_create(sizeof(struct fs_info_struct));
  if (fs_info_cache == 0) {
    panic("Could not allocate fs_info_struct cache");
  }
  mmap_cache = kmem_cache_create(sizeof(struct mmap_struct));
  if (mmap_cache == 0) {
    panic("Could not allocate mmap_struct cache");
  }
  mmap_list_cache = kmem_cache_create(sizeof(struct mmap_list));
  if (mmap_list_cache == 0) {
    panic("Could not allocate mmap_list cache");
  }
  INIT_LIST_HEAD(&ptable.list);
}

//PAGEBREAK: 32
// Look in the process table for an UNUSED proc.
// If found, change state to EMBRYO and initialize
// state required to run in the kernel.
// Otherwise return 0.
static struct proc*
allocproc(void)
{
  char *sp;

  struct proc* p = kmem_cache_alloc(proc_cache);
  if (p == 0) {
    return 0;
  }
  acquire(&ptable.lock);
  memset(p, 0, sizeof(*p));

  list_add_tail(&p->list, &ptable.list);

  INIT_LIST_HEAD(&p->children);
  INIT_LIST_HEAD(&p->siblings);
  INIT_LIST_HEAD(&p->thread_group);
  p->state = EMBRYO;
  p->pid = nextpid++;
  release(&ptable.lock);

  // Allocate kernel stack.
  if((p->kstack = kalloc()) == 0){
    p->state = UNUSED;
    list_del(&p->list);
    kmem_cache_free(p);
    return 0;
  }
  sp = p->kstack + KSTACKSIZE;
  
  // Leave room for trap frame.
  sp -= sizeof *p->tf;
  p->tf = (struct trapframe*)sp;
  
  // Set up new context to start executing at forkret,
  // which returns to trapret.
  sp -= 4;
  *(uint*)sp = (uint)trapret;

  sp -= sizeof *p->context;
  p->context = (struct context*)sp;
  memset(p->context, 0, sizeof *p->context);
  p->context->eip = (uint)forkret;

  return p;
}

void
free_mmaps(struct mm_struct* mm)
{
  if (mm->users > 1) {
    return;
  }
  struct list_head *pos, *next;
  list_for_each_safe(pos, next, &mm->mmap_list) {
    struct mmap_list* mmap_list = list_entry(pos, struct mmap_list, list);
    struct mmap_struct* mmap = mmap_list->mmap;
    // Obviously, we need acquire.
    // Placing release right after list_del is safe, because nobody
    // else will know that that mmap exists.
    // Release is needed, because otherwise filewrite would panic when
    // it tries to sleep.
    acquire(&mmap->lock);
    list_del(&mmap_list->list);
    kmem_cache_free(mmap_list);
    --(mmap->users);
    if (mmap->users > 0) {
      release(&mmap->lock);
      for (uint i = (uint)mmap->start;
          i < PGROUNDUP((uint)mmap->start + mmap->length);
          i += PGSIZE) {
        // Mark it as unused, because otherwise we would free it later,
        // because it is in our page table.
        set_pte_permissions(mm->pgdir, (void*)i, 0);
      }
      continue;
    }
    release(&mmap->lock);
    // Write to file.
    for (uint i = (uint)mmap->start;
        i < PGROUNDUP((uint)mmap->start + mmap->length);
        i += PGSIZE) {
      // We don’t have to write changes to the file in these two cases.
      if ((mmap->flags & MAP_SHARED) == 0 ||
          (mmap->flags & MAP_ANONYMOUS) != 0) {
        break;
      }
      int is_last_page =
        (i + PGSIZE >= PGROUNDUP((uint)mmap->start + mmap->length));
      int perm = get_pte_permissions(mm->pgdir, (void*)i);
      if (perm & PTE_D) {
        // Page is dirty, we need to write changes to the file.
        struct file* file = mmap->file;
        uint old_off = file->off;
        file->off = (uint)i - (uint)mmap->start + (uint)mmap->offset;
        uint amount = PGSIZE;
        if (is_last_page) {
          ilock(file->ip);
          amount = file->ip->size;
          iunlock(file->ip);
        }
        filewrite(file, (char*)i, amount);
        file->off = old_off;
      }
    }
    kmem_cache_free(mmap);
  }
}

// Decrease number of users of the memory map,
// free memory when user count reaches 0.
void
free_mm(struct mm_struct* mm)
{
  acquire(&mm->lock);
  if (--mm->users == 0) {
    if (!list_empty(&mm->mmap_list)) {
      panic("mmap list not empty in free_mm");
    }
    if (mm->pgdir != 0) {
      freevm(mm->pgdir);
    }
    mm->pgdir = 0;
    mm->sz = 0;
    release(&mm->lock);
    kmem_cache_free(mm);
  } else {
    release(&mm->lock);
  }
}

// For use in userinit() only.
static struct mm_struct*
setup_mm(void)
{
  struct mm_struct* mm = kmem_cache_alloc(mm_cache);
  if (!mm) return 0;
  initlock(&mm->lock, "proc->mm");
  initlock(&mm->mmap_list_lock, "proc->mm->mmap_list");
  mm->users = 1;
  mm->pgdir = setupkvm();
  mm->sz = PGSIZE;
  INIT_LIST_HEAD(&mm->mmap_list);
  if (!mm->pgdir) {
    free_mm(mm);
    return 0;
  }
  return mm;
}

// Assumes that mm is already locked.
static int
add_mmap_to_mm(struct mm_struct* mm, struct mmap_struct* mmap, pde_t* pgdir)
{
  acquire(&mmap->lock);
  struct mmap_list* mmap_list = kmem_cache_alloc(mmap_list_cache);
  if ((mmap->flags & MAP_SHARED) == 0) {
    struct mmap_struct* copy = kmem_cache_alloc(mmap_cache);
    if (!copy) {
      release(&mmap->lock);
      return -1;
    }
    *copy = *mmap;
    if (copy->file != 0) {
      copy->file = filedup(mmap->file);
    }
    mmap_list->mmap = copy;
  } else {
    mmap->users++;
    mmap_list->mmap = mmap;
  }
  if (mmap->flags & MAP_SHARED) {
    for (uint start = (uint)mmap->start;
        start < (uint)mmap->start + mmap->length;
        start += PGSIZE) {
      pte_t* entry = walkpgdir(pgdir, (void*)start, 0);
      if (entry == 0) {
        panic("add_mmap_to_mm: page table entry does not exist");
      }
      uint addr = PTE_ADDR(*entry);
      uint flags = PTE_FLAGS(*entry);
      mappages(mm->pgdir, (void*)start, PGSIZE, addr, flags);
    }
  }
  release(&mmap->lock);
  list_add_tail(&mmap_list->list, &mm->mmap_list);
  return 0;
}

static int
copy_mm(unsigned int clone_flags, struct proc* p)
{
  struct mm_struct *mm;
  if (!p->mm) return 0;
  if (clone_flags & CLONE_VM) {
    acquire(&p->mm->lock);
    p->mm->users++;
    release(&p->mm->lock);
    return 0;
  }
  mm = kmem_cache_alloc(mm_cache);
  if (!mm) return -ENOMEM;
  initlock(&mm->lock, "proc->mm");
  initlock(&mm->mmap_list_lock, "proc->mmap_list");
  INIT_LIST_HEAD(&mm->mmap_list);
  mm->users = 1;
  acquire(&p->mm->lock);
  mm->pgdir = copyuvm(p->mm->pgdir, p->mm->sz);
  if (mm->pgdir == 0) {
    release(&p->mm->lock);
    free_mm(mm);
    return -ENOMEM;
  }
  mm->sz = p->mm->sz;
  // Copy mmaps
  struct list_head* list;
  acquire(&p->mm->mmap_list_lock);
  list_for_each(list, &p->mm->mmap_list) {
    struct mmap_list* mmap_list = list_entry(list, struct mmap_list, list);
    struct mmap_struct* mmap = mmap_list->mmap;
    if (add_mmap_to_mm(mm, mmap, p->mm->pgdir) < 0) {
      release(&p->mm->mmap_list_lock);
      release(&p->mm->lock);
      free_mm(mm);
      return -ENOMEM;
    }
  }
  release(&p->mm->mmap_list_lock);
  release(&p->mm->lock);
  p->mm = mm;
  return 0;
}

static int
copy_files(unsigned int clone_flags, struct proc* p)
{
  struct files_struct *files;
  if (clone_flags & CLONE_FILES) {
    acquire(&p->files->lock);
    p->files->users++;
    release(&p->files->lock);
    return 0;
  }
  files = kmem_cache_alloc(files_struct_cache);
  if (!files) return -ENOMEM;
  initlock(&files->lock, "proc->files");
  files->users = 1;
  files->fd = (struct file**)kalloc();
  if (!files->fd) {
    kmem_cache_free(files);
    return -ENOMEM;
  }
  memset(files->fd, 0, PGSIZE);
  acquire(&p->files->lock);
  for(int i = 0; i < NOFILE; i++)
    if(p->files->fd[i])
      files->fd[i] = filedup(p->files->fd[i]);
  release(&p->files->lock);
  p->files = files;
  return 0;
}

static void
free_files(struct files_struct* files)
{
  acquire(&files->lock);
  if (--files->users == 0) {
    release(&files->lock);
    for (int fd = 0; fd < NOFILE; fd++){
      if(files->fd[fd]){
        fileclose(files->fd[fd]);
        files->fd[fd] = 0;
      }
    }
    kfree((char*)files->fd);
    kmem_cache_free(files);
  } else {
    release(&files->lock);
  }
}

static int
copy_fs_info(unsigned int clone_flags, struct proc* p)
{
  struct fs_info_struct* fs_info;
  if (clone_flags & CLONE_FS) {
    acquire(&p->fs->lock);
    p->fs->users++;
    release(&p->fs->lock);
    return 0;
  }
  fs_info = kmem_cache_alloc(fs_info_cache);
  if (!fs_info) return -ENOMEM;
  initlock(&fs_info->lock, "proc->fs");
  fs_info->users = 1;
  acquire(&p->fs->lock);
  fs_info->root = idup(p->fs->root);
  fs_info->cwd = idup(p->fs->cwd);
  fs_info->umask = p->fs->umask;
  release(&p->fs->lock);
  p->fs = fs_info;
  return 0;
}

static void
free_fs_info(struct fs_info_struct* fs_info)
{
  acquire(&fs_info->lock);
  if (--fs_info->users == 0) {
    iput(fs_info->cwd);
    iput(fs_info->root);
    fs_info->cwd = 0;
    fs_info->root = 0;
    release(&fs_info->lock);
    kmem_cache_free(fs_info);
  } else {
    release(&fs_info->lock);
  }
}

//PAGEBREAK: 32
// Set up first user process.
void
userinit(void)
{
  struct proc *p;
  extern char _binary_initcode_start[], _binary_initcode_size[];
  
  p = allocproc();
  initproc = p;
  if((p->mm = setup_mm()) == 0)
    panic("userinit: out of memory?");
  inituvm(p->mm->pgdir, _binary_initcode_start, (int)_binary_initcode_size);
  p->mm->sz = PGSIZE;
  p->files = kmem_cache_alloc(files_struct_cache);
  p->files->users = 1;
  p->detached = 0;
  initlock(&p->files->lock, "proc->files");
  p->files->fd = (struct file**)kalloc();
  memset(p->files->fd, 0, PGSIZE);
  memset(p->tf, 0, sizeof(*p->tf));
  p->tf->cs = (SEG_UCODE << 3) | DPL_USER;
  p->tf->ds = (SEG_UDATA << 3) | DPL_USER;
  p->tf->es = p->tf->ds;
  p->tf->ss = p->tf->ds;
  p->tf->eflags = FL_IF;
  p->tf->esp = PGSIZE;
  p->tf->eip = 0;  // beginning of initcode.S

  safestrcpy(p->name, "initcode", sizeof(p->name));
  p->fs = kmem_cache_alloc(fs_info_cache);
  p->fs->root = 0;
  p->fs->cwd = namei("/");
  p->fs->root = namei("/");
  p->fs->users = 1;
  p->fs->umask = 0;
  initlock(&p->fs->lock, "proc->fs");
  p->uid = 0;
  p->euid = 0;
  p->gid = 0;
  p->egid = 0;

  p->state = RUNNABLE;
}

// Grow current process's memory by n bytes.
// Return 0 on success, -1 on failure.
int
growproc(int n)
{
  uint sz;
  
  sz = proc->mm->sz;
  if(n > 0){
    if((sz = allocuvm(proc->mm->pgdir, sz, sz + n, PTE_W | PTE_U)) == 0)
      return -1;
  } else if(n < 0){
    if((sz = deallocuvm(proc->mm->pgdir, sz, sz + n)) == 0)
      return -1;
  }
  proc->mm->sz = sz;
  switchuvm(proc);
  return 0;
}

int
clone(void* child_stack, unsigned int clone_flags)
{
  int pid;
  struct proc *np;

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

  int retval;
  np->mm = proc->mm;
  if ((retval = copy_mm(clone_flags, np)) < 0) {
    np->mm = 0;
    kfree(np->kstack);
    np->kstack = 0;
    np->state = UNUSED;
    return retval;
  }
  np->files = proc->files;
  if ((retval = copy_files(clone_flags, np)) < 0) {
    free_mm(np->mm);
    kfree(np->kstack);
    np->kstack = 0;
    np->state = UNUSED;
    return retval;
  }
  np->fs = proc->fs;
  if ((retval = copy_fs_info(clone_flags, np)) < 0) {
    free_mm(np->mm);
    free_files(np->files);
    kfree(np->kstack);
    np->kstack = 0;
    np->state = UNUSED;
    return retval;
  }
  if (clone_flags & (CLONE_THREAD | CLONE_PARENT)) {
    np->parent = proc->parent;
  } else {
    np->parent = proc;
  }
  *np->tf = *proc->tf;

  // Clear %eax so that clone returns 0 in the child.
  np->tf->eax = 0;

  np->uid = proc->uid;
  np->euid = proc->euid;
  np->suid = proc->suid;
  np->gid = proc->gid;
  np->egid = proc->egid;
  np->sgid = proc->sgid;
  if (child_stack) {
    np->tf->esp = (uint)child_stack;
  }
  if (clone_flags & CLONE_THREAD) {
    np->detached = 1;
  } else {
    np->detached = 0;
  }
  INIT_LIST_HEAD(&np->thread_group);
  if (clone_flags & CLONE_THREAD) {
    np->group_leader = proc->group_leader;
    np->tgid = proc->tgid;
  } else {
    np->group_leader = np;
    np->tgid = np->pid;
  }
  np->ngroups = proc->ngroups;
  for (int i = 0; i < proc->ngroups; ++i) {
    np->groups[i] = proc->groups[i];
  }

  acquire(&ptable.lock);
  if (clone_flags & CLONE_THREAD) {
    list_add_tail(&np->thread_group, &proc->thread_group);
  }
  list_add_tail(&np->siblings, &np->parent->children);
  release(&ptable.lock);

  pid = np->pid;
  np->state = RUNNABLE;
  safestrcpy(np->name, proc->name, sizeof(proc->name));
  return pid;
}

// Exit the specified process.
// An exited process remains in the zombie state
// until its parent calls wait() to find out it exited (unless it was detached).
// ptable.lock must be acquired before entering this function.
void
exit(void)
{
  struct proc *p;

  if(proc == initproc)
    panic("init exiting");

  // Close all open files.
  free_files(proc->files);
  free_fs_info(proc->fs);
  // We need to do free_mmaps here, and not when we are freeing mm,
  // because we are freeing mm with ptable.lock held, which makes filewrite
  // (that may happen in free_mmaps) panic.
  free_mmaps(proc->mm);

  acquire(&ptable.lock);

  // Parent might be sleeping in wait().
  if (!proc->detached) {
    wakeup1(proc->parent);
  }

  // Pass abandoned children to init.
  struct list_head *pos, *next;

  list_for_each_safe(pos, next, &proc->children) {
    p = list_entry(pos, struct proc, siblings);
    list_del(pos);
    if (p->state == UNUSED) continue;
    p->parent = initproc;
    list_add_tail(pos, &initproc->children);
    if (p->state == ZOMBIE && !p->detached)
      wakeup1(initproc);
  }

  if (!proc->detached) {
    proc->state = ZOMBIE;
  } else {
    // Free the process, nobody should wait for it.
    proc->state = UNUSED;
  }
  // Jump into the scheduler, never to return.
  sched();
  panic("zombie exit");
}

void
kill_other_threads_in_group(void)
{
  acquire(&ptable.lock);
  struct proc* p;
  list_for_each_entry(p, &proc->thread_group, thread_group) {
    p->killed = 1;
    if(p->state == SLEEPING)
      p->state = RUNNABLE;
  }
  release(&ptable.lock);
}

// Exit all the threads in the thread group of the current process.
void
exit_group(void)
{
  kill_other_threads_in_group();
  exit();
}

// Wait for a child process to exit and return its pid.
// Return -1 if this process has no children.
int
wait(void)
{
  struct proc *p;
  int havekids, pid;

  acquire(&ptable.lock);
  for(;;){
    // Scan through table looking for zombie children.
    havekids = 0;
    struct list_head *pos, *next;
    list_for_each_safe(pos, next, &proc->children) {
      p = list_entry(pos, struct proc, siblings);
      if (p->detached) continue;
      havekids = 1;
      if(p->state == ZOMBIE){
        // Found one.
        list_del_init(pos);
        release(&ptable.lock);
        pid = p->pid;
        if (p->kstack) {
          kfree(p->kstack);
          p->kstack = 0;
        }
        if (p->mm) {
          free_mm(p->mm);
          p->mm = 0;
        }
        p->state = UNUSED;
        p->pid = 0;
        p->parent = 0;
        p->name[0] = 0;
        p->killed = 0;
        return pid;
      }
    }

    // No point waiting if we don't have any children.
    if(!havekids || proc->killed){
      release(&ptable.lock);
      return -ECHILD;
    }

    // Wait for children to exit.  (See wakeup1 call in proc_exit.)
    sleep(proc, &ptable.lock);  //DOC: wait-sleep
  }
}

//PAGEBREAK: 42
// Per-CPU process scheduler.
// Each CPU calls scheduler() after setting itself up.
// Scheduler never returns.  It loops, doing:
//  - choose a process to run
//  - swtch to start running that process
//  - eventually that process transfers control
//      via swtch back to the scheduler.
void
scheduler(void)
{
  struct proc *p;

  for(;;){
    // Enable interrupts on this processor.
    sti();

    // Loop over process table looking for process to run.
    acquire(&ptable.lock);
    struct list_head *pos;
    // We can’t use list_for_each_safe here, I don’t quite understand
    // why, but I have spent around ten hours debugging this,
    // so you’d better trust me.
    list_for_each(pos, &ptable.list) {
      p = list_entry(pos, struct proc, list);
      if (p->state == UNUSED) {
        if (p->kstack != 0) {
          kfree(p->kstack);
          p->kstack = 0;
        }
        if (p->mm != 0) {
          free_mm(p->mm);
          p->mm = 0;
        }
        struct list_head* prev = pos;
        pos = pos->prev;
        list_del(&p->thread_group);
        list_del(&p->siblings);
        list_del(prev);
        kmem_cache_free(list_entry(prev, struct proc, list));
        continue;
      }
      if (p->state != RUNNABLE)
        continue;

      // Switch to chosen process.  It is the process's job
      // to release ptable.lock and then reacquire it
      // before jumping back to us.
      proc = p;
      switchuvm(p);
      p->state = RUNNING;
      swtch(&cpu->scheduler, proc->context);
      switchkvm();

      // Process is done running for now.
      // It should have changed its p->state before coming back.
      proc = 0;
    }
    release(&ptable.lock);

  }
}

// Enter scheduler.  Must hold only ptable.lock
// and have changed proc->state.
void
sched(void)
{
  int intena;

  if(!holding(&ptable.lock))
    panic("sched ptable.lock");
  if(cpu->ncli != 1)
    panic("sched locks");
  if(proc->state == RUNNING)
    panic("sched running");
  if(readeflags()&FL_IF)
    panic("sched interruptible");
  intena = cpu->intena;
  swtch(&proc->context, cpu->scheduler);
  cpu->intena = intena;
}

// Give up the CPU for one scheduling round.
void
yield(void)
{
  acquire(&ptable.lock);  //DOC: yieldlock
  proc->state = RUNNABLE;
  sched();
  release(&ptable.lock);
}

// A fork child's very first scheduling by scheduler()
// will swtch here.  "Return" to user space.
void
forkret(void)
{
  static int first = 1;
  // Still holding ptable.lock from scheduler.
  release(&ptable.lock);

  if (first) {
    // Some initialization functions must be run in the context
    // of a regular process (e.g., they call sleep), and thus cannot 
    // be run from main().
    first = 0;
    initlog();
  }
  
  // Return to "caller", actually trapret (see allocproc).
}

// Atomically release lock and sleep on chan.
// Reacquires lock when awakened.
void
sleep(void *chan, struct spinlock *lk)
{
  if(proc == 0)
    panic("sleep");

  if(lk == 0)
    panic("sleep without lk");

  // Must acquire ptable.lock in order to
  // change p->state and then call sched.
  // Once we hold ptable.lock, we can be
  // guaranteed that we won't miss any wakeup
  // (wakeup runs with ptable.lock locked),
  // so it's okay to release lk.
  if(lk != &ptable.lock){  //DOC: sleeplock0
    acquire(&ptable.lock);  //DOC: sleeplock1
    release(lk);
  }

  // Go to sleep.
  proc->chan = chan;
  proc->state = SLEEPING;
  sched();

  // Tidy up.
  proc->chan = 0;

  // Reacquire original lock.
  if(lk != &ptable.lock){  //DOC: sleeplock2
    release(&ptable.lock);
    acquire(lk);
  }
}

//PAGEBREAK!
// Wake up all processes sleeping on chan.
// The ptable lock must be held.
static void
wakeup1(void *chan)
{
  struct proc *p;

  /*for(p = ptable.proc; p < &ptable.proc[NPROC]; p++)*/
  struct list_head *pos, *next;
  list_for_each_safe(pos, next, &ptable.list) {
    p = list_entry(pos, struct proc, list);
    if(p->state == SLEEPING && p->chan == chan)
      p->state = RUNNABLE;
  }
}

// Wake up all processes sleeping on chan.
void
wakeup(void *chan)
{
  acquire(&ptable.lock);
  wakeup1(chan);
  release(&ptable.lock);
}

struct proc*
get_proc_by_pid(int pid)
{
  struct proc *p;

  acquire(&ptable.lock);
  list_for_each_entry(p, &ptable.list, list) {
    if(p->pid == pid){
      release(&ptable.lock);
      return p;
    }
  }
  release(&ptable.lock);
  return 0;
}

// Kill the process with the given pid.
// Process won't exit until it returns
// to user space (see trap in trap.c).
int
kill(int pid)
{
  struct proc *p;

  acquire(&ptable.lock);
  struct list_head *pos, *next;
  list_for_each_safe(pos, next, &ptable.list) {
    p = list_entry(pos, struct proc, list);
    if(p->pid == pid){
      if (proc->euid != 0 &&
          proc->uid != p->uid &&
          proc->uid != p->suid &&
          proc->euid != p->uid &&
          proc->euid != p->suid) {
        release(&ptable.lock);
        return -EPERM;
      }
      p->killed = 1;
      // Wake process from sleep if necessary.
      if(p->state == SLEEPING)
        p->state = RUNNABLE;
      release(&ptable.lock);
      return 0;
    }
  }
  release(&ptable.lock);
  return -ESRCH;
}

// addr must be page-aligned.
void*
mmap(void* addr, int length, int prot, int flags, struct file* file,
    int offset)
{
  int current_length = 0;
  if ((flags & (MAP_SHARED | MAP_PRIVATE)) == 0) {
    return ERR_PTR(-EINVAL);
  }
  if ((flags & MAP_ANONYMOUS) == 0) {
    if (((prot & PROT_READ) && !(file->readable)) ||
        ((prot & PROT_WRITE) && !(file->writable))) {
      return ERR_PTR(-EACCES);
    }
    if (file->type != FD_INODE) {
      return ERR_PTR(-EACCES);
    }
  }
  for (current_length = 0; current_length < length;
      current_length += PGSIZE) {
    if (!set_pte_permissions(proc->mm->pgdir, addr + current_length,
          PTE_P | PTE_W)) {
      return ERR_PTR(-ENOMEM);
    }
  }
  memset(addr, 0, current_length);
  struct mmap_struct* mmap = kmem_cache_alloc(mmap_cache);
  *mmap = (struct mmap_struct) {
    .start = addr,
    .length = length,
    .prot = prot,
    .flags = flags,
    .file = (file == 0 ? 0 : filedup(file)),
    .offset = offset,
    .users = 1,
  };
  initlock(&mmap->lock, "mmap");
  if ((flags & MAP_ANONYMOUS) == 0) {
    uint initial_offset = file->off;
    file->off = offset;
    fileread(file, mmap->start, PGROUNDUP(length));
    file->off = initial_offset;
  }
  uint permissions = PTE_P;
  if ((prot & PROT_READ) || (prot & PROT_EXEC)) {
    permissions |= PTE_U;
  }
  if (flags & MAP_SHARED) {
    permissions |= PTE_MMAP;
  }
  for (current_length = 0; current_length < length;
      current_length += PGSIZE) {
    if (!set_pte_permissions(proc->mm->pgdir, addr + current_length,
          permissions)) {
      return ERR_PTR(-ENOMEM);
    }
  }
  struct mmap_list* mmap_list = kmem_cache_alloc(mmap_list_cache);
  mmap_list->mmap = mmap;

  acquire(&proc->mm->lock);
  list_add_tail(&mmap_list->list, &proc->mm->mmap_list);
  release(&proc->mm->lock);

  return addr;
}

int
load_mmap(struct mmap_struct* mmap, uint offset, char* dst,
    int is_write, uint perm)
{
  set_pte_permissions(proc->mm->pgdir, dst, perm);
  return 1;
}

int
handle_pagefault(uint address, uint err)
{
  struct list_head* pos;
  int is_write = (err & 2);
  if (!is_write) {
    return 0;
  }
  acquire(&proc->mm->mmap_list_lock);
  list_for_each(pos, &proc->mm->mmap_list) {
    struct mmap_list* mmap_list = list_entry(pos, struct mmap_list, list);
    struct mmap_struct* mmap = mmap_list->mmap;
    if (mmap->start <= (char*)address &&
        (char*)address < mmap->start + PGROUNDUP(mmap->length)) {
      acquire(&mmap->lock);
      // Found the corresponding mmap.
      // Check permissions.
      if ((is_write && ((mmap->prot & PROT_WRITE) == 0)) ||
          ((mmap->prot & PROT_READ) == 0)) {
        acquire(&mmap->lock);
        release(&proc->mm->mmap_list_lock);
        return 0;
      }
      uint permissions = PTE_P | PTE_U | PTE_W | PTE_D;
      if (mmap->flags & MAP_SHARED) {
        permissions |= PTE_MMAP;
      }
      int retval = load_mmap(mmap, address - (uint)mmap->start,
          (char*)PGROUNDDOWN((uint)address),
          is_write, permissions);
      release(&mmap->lock);
      release(&proc->mm->mmap_list_lock);
      return retval;
    }
  }
  release(&proc->mm->mmap_list_lock);
  return 0;
}

//PAGEBREAK: 36
// Print a process listing to console.  For debugging.
// Runs when user types ^P on console.
// No lock to avoid wedging a stuck machine further.
void
procdump(void)
{
  static char *states[] = {
  [UNUSED]    "unused",
  [EMBRYO]    "embryo",
  [SLEEPING]  "sleep ",
  [RUNNABLE]  "runble",
  [RUNNING]   "run   ",
  [ZOMBIE]    "zombie"
  };
  int i;
  struct proc *p;
  char *state;
  uint pc[10];
  
  struct list_head *pos, *next;
  list_for_each_safe(pos, next, &ptable.list) {
    p = list_entry(pos, struct proc, list);
    if(p->state == UNUSED)
      continue;
    if(p->state >= 0 && p->state < NELEM(states) && states[p->state])
      state = states[p->state];
    else
      state = "???";
    cprintf("%d %s %s %d %d", p->pid, state, p->name, p->uid, p->gid);
    if(p->state == SLEEPING){
      getcallerpcs((uint*)p->context->ebp+2, pc);
      for(i=0; i<10 && pc[i] != 0; i++)
        cprintf(" %p", pc[i]);
    }
    cprintf("\n");
  }
}