[PATCH] make /proc/mounts pollable
Al Viro [Mon, 7 Nov 2005 22:15:49 +0000 (17:15 -0500)]
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

fs/namespace.c
fs/proc/base.c
include/linux/namespace.h

index 611f777..d1aca68 100644 (file)
@@ -37,7 +37,9 @@ static inline int sysfs_init(void)
 #endif
 
 /* spinlock for vfsmount related operations, inplace of dcache_lock */
- __cacheline_aligned_in_smp DEFINE_SPINLOCK(vfsmount_lock);
+__cacheline_aligned_in_smp DEFINE_SPINLOCK(vfsmount_lock);
+
+static int event;
 
 static struct list_head *mount_hashtable;
 static int hash_mask __read_mostly, hash_bits __read_mostly;
@@ -111,6 +113,22 @@ static inline int check_mnt(struct vfsmount *mnt)
        return mnt->mnt_namespace == current->namespace;
 }
 
+static void touch_namespace(struct namespace *ns)
+{
+       if (ns) {
+               ns->event = ++event;
+               wake_up_interruptible(&ns->poll);
+       }
+}
+
+static void __touch_namespace(struct namespace *ns)
+{
+       if (ns && ns->event != event) {
+               ns->event = event;
+               wake_up_interruptible(&ns->poll);
+       }
+}
+
 static void detach_mnt(struct vfsmount *mnt, struct nameidata *old_nd)
 {
        old_nd->dentry = mnt->mnt_mountpoint;
@@ -384,6 +402,7 @@ static void umount_tree(struct vfsmount *mnt)
        for (p = mnt; p; p = next_mnt(p, mnt)) {
                list_del(&p->mnt_list);
                list_add(&p->mnt_list, &kill);
+               __touch_namespace(p->mnt_namespace);
                p->mnt_namespace = NULL;
        }
 
@@ -473,6 +492,7 @@ static int do_umount(struct vfsmount *mnt, int flags)
 
        down_write(&current->namespace->sem);
        spin_lock(&vfsmount_lock);
+       event++;
 
        retval = -EBUSY;
        if (atomic_read(&mnt->mnt_count) == 2 || flags & MNT_DETACH) {
@@ -634,6 +654,7 @@ static int graft_tree(struct vfsmount *mnt, struct nameidata *nd)
                list_splice(&head, current->namespace->list.prev);
                mntget(mnt);
                err = 0;
+               touch_namespace(current->namespace);
        }
        spin_unlock(&vfsmount_lock);
 out_unlock:
@@ -771,6 +792,7 @@ static int do_move_mount(struct nameidata *nd, char *old_name)
 
        detach_mnt(old_nd.mnt, &parent_nd);
        attach_mnt(old_nd.mnt, nd);
+       touch_namespace(current->namespace);
 
        /* if the mount is moved, it should no longer be expire
         * automatically */
@@ -877,6 +899,7 @@ static void expire_mount(struct vfsmount *mnt, struct list_head *mounts)
                struct nameidata old_nd;
 
                /* delete from the namespace */
+               touch_namespace(mnt->mnt_namespace);
                list_del_init(&mnt->mnt_list);
                mnt->mnt_namespace = NULL;
                detach_mnt(mnt, &old_nd);
@@ -1114,6 +1137,8 @@ int copy_namespace(int flags, struct task_struct *tsk)
        atomic_set(&new_ns->count, 1);
        init_rwsem(&new_ns->sem);
        INIT_LIST_HEAD(&new_ns->list);
+       init_waitqueue_head(&new_ns->poll);
+       new_ns->event = 0;
 
        down_write(&tsk->namespace->sem);
        /* First pass: copy the tree topology */
@@ -1377,6 +1402,7 @@ asmlinkage long sys_pivot_root(const char __user *new_root, const char __user *p
        detach_mnt(user_nd.mnt, &root_parent);
        attach_mnt(user_nd.mnt, &old_nd);     /* mount old root on put_old */
        attach_mnt(new_nd.mnt, &root_parent); /* mount new_root on / */
+       touch_namespace(current->namespace);
        spin_unlock(&vfsmount_lock);
        chroot_fs_refs(&user_nd, &new_nd);
        security_sb_post_pivotroot(&user_nd, &new_nd);
@@ -1413,6 +1439,8 @@ static void __init init_mount_tree(void)
        atomic_set(&namespace->count, 1);
        INIT_LIST_HEAD(&namespace->list);
        init_rwsem(&namespace->sem);
+       init_waitqueue_head(&namespace->poll);
+       namespace->event = 0;
        list_add(&mnt->mnt_list, &namespace->list);
        namespace->root = mnt;
        mnt->mnt_namespace = namespace;
index a170450..634355e 100644 (file)
@@ -70,6 +70,7 @@
 #include <linux/seccomp.h>
 #include <linux/cpuset.h>
 #include <linux/audit.h>
+#include <linux/poll.h>
 #include "internal.h"
 
 /*
@@ -660,26 +661,38 @@ static struct file_operations proc_smaps_operations = {
 #endif
 
 extern struct seq_operations mounts_op;
+struct proc_mounts {
+       struct seq_file m;
+       int event;
+};
+
 static int mounts_open(struct inode *inode, struct file *file)
 {
        struct task_struct *task = proc_task(inode);
-       int ret = seq_open(file, &mounts_op);
+       struct namespace *namespace;
+       struct proc_mounts *p;
+       int ret = -EINVAL;
 
-       if (!ret) {
-               struct seq_file *m = file->private_data;
-               struct namespace *namespace;
-               task_lock(task);
-               namespace = task->namespace;
-               if (namespace)
-                       get_namespace(namespace);
-               task_unlock(task);
-
-               if (namespace)
-                       m->private = namespace;
-               else {
-                       seq_release(inode, file);
-                       ret = -EINVAL;
+       task_lock(task);
+       namespace = task->namespace;
+       if (namespace)
+               get_namespace(namespace);
+       task_unlock(task);
+
+       if (namespace) {
+               ret = -ENOMEM;
+               p = kmalloc(sizeof(struct proc_mounts), GFP_KERNEL);
+               if (p) {
+                       file->private_data = &p->m;
+                       ret = seq_open(file, &mounts_op);
+                       if (!ret) {
+                               p->m.private = namespace;
+                               p->event = namespace->event;
+                               return 0;
+                       }
+                       kfree(p);
                }
+               put_namespace(namespace);
        }
        return ret;
 }
@@ -692,11 +705,30 @@ static int mounts_release(struct inode *inode, struct file *file)
        return seq_release(inode, file);
 }
 
+static unsigned mounts_poll(struct file *file, poll_table *wait)
+{
+       struct proc_mounts *p = file->private_data;
+       struct namespace *ns = p->m.private;
+       unsigned res = 0;
+
+       poll_wait(file, &ns->poll, wait);
+
+       spin_lock(&vfsmount_lock);
+       if (p->event != ns->event) {
+               p->event = ns->event;
+               res = POLLERR;
+       }
+       spin_unlock(&vfsmount_lock);
+
+       return res;
+}
+
 static struct file_operations proc_mounts_operations = {
        .open           = mounts_open,
        .read           = seq_read,
        .llseek         = seq_lseek,
        .release        = mounts_release,
+       .poll           = mounts_poll,
 };
 
 #define PROC_BLOCK_SIZE        (3*1024)                /* 4K page size but our output routines use some slack for overruns */
index 0e5a86f..6f0f25d 100644 (file)
@@ -10,6 +10,8 @@ struct namespace {
        struct vfsmount *       root;
        struct list_head        list;
        struct rw_semaphore     sem;
+       wait_queue_head_t poll;
+       int event;
 };
 
 extern int copy_namespace(int, struct task_struct *);