procfs
/* fs/proc/root.c
* This is the root "inode" in the /proc tree..
*/
struct proc_dir_entry proc_root = {
.low_ino = PROC_ROOT_INO,
.namelen = 5,
.mode = S_IFDIR | S_IRUGO | S_IXUGO,
.nlink = 2,
.count = ATOMIC_INIT(1),
.proc_iops = &proc_root_inode_operations,
.proc_fops = &proc_root_operations,
.parent = &proc_root,
.name = "/proc",
};
int proc_fill_super(struct super_block *s)
{
struct inode *root_inode;
s->s_flags |= MS_NODIRATIME | MS_NOSUID | MS_NOEXEC;
s->s_blocksize = 1024;
s->s_blocksize_bits = 10;
s->s_magic = PROC_SUPER_MAGIC;
s->s_op = &proc_sops;
s->s_time_gran = 1;
pde_get(&proc_root);
root_inode = proc_get_inode(s, &proc_root); ///alloc inode for proc dentry
if (!root_inode) {
pr_err("proc_fill_super: get root inode failed\n");
return -ENOMEM;
}
s->s_root = d_make_root(root_inode);
if (!s->s_root) {
pr_err("proc_fill_super: allocate dentry failed\n");
return -ENOMEM;
}
return proc_setup_self(s); ///proc/self
}
- lookup
/*
* Call i_op->lookup on the dentry. The dentry must be negative but may be
* hashed if it was pouplated with DCACHE_NEED_LOOKUP.
*
* dir->d_inode->i_mutex must be held
*/
static struct dentry *lookup_real(struct inode *dir, struct dentry *dentry,
unsigned int flags)
{
struct dentry *old;
/* Don't create child dentry for a dead directory. */
if (unlikely(IS_DEADDIR(dir))) {
dput(dentry);
return ERR_PTR(-ENOENT);
}
///dir is parent inode, dentry is child dentry object
old = dir->i_op->lookup(dir, dentry, flags); ///proc_root_lookup
if (unlikely(old)) {
dput(dentry); ///目录项已经存在，释放新分配的dentry，使用旧的dentry
dentry = old;
}
return dentry;
}
static struct dentry *proc_root_lookup(struct inode * dir, struct dentry * dentry, unsigned int flags)
{
if (!proc_lookup(dir, dentry, flags))
return NULL; ///old dentry is exist
return proc_pid_lookup(dir, dentry, flags); ///alloc inode for dentry
}
- mount
# unshare -m /bin/bash
# mount --make-rslave /
# nsenter --target=14124 --pid -- /bin/bash
# umount -l /proc
# cat /proc/mounts
...
none /var/run/docker/netns/default proc rw,relatime 0 0
none /var/run/docker/netns/96761cbf926b proc rw,relatime 0 0
proc /proc proc rw,relatime 0 0
# umount -l /proc
# cat /proc/mounts
cat: /proc/mounts: No such file or directory
# mount -t proc proc /proc
# ps -ef
UID PID PPID C STIME TTY TIME CMD
root 1 0 0 16:32 ? 00:00:00 /usr/sbin/sshd -D
root 129 0 0 17:40 pts/14 00:00:00 /bin/bash
root 148 129 0 17:41 pts/14 00:00:00 ps -ef
# stat /proc/1
File: `/proc/1'
Size: 0 Blocks: 0 IO Block: 1024 directory
Device: 20h/32d Inode: 1040776094 Links: 7
Access: (0555/dr-xr-xr-x) Uid: ( 0/ root) Gid: ( 0/ root)
Access: 2017-05-19 16:33:00.829295372 +0800
Modify: 2017-05-19 16:33:00.829295372 +0800
Change: 2017-05-19 16:33:00.829295372 +0800
# stat /proc
File: `/proc'
Size: 0 Blocks: 0 IO Block: 1024 directory
Device: 20h/32d Inode: 1 Links: 260
Access: (0555/dr-xr-xr-x) Uid: ( 0/ root) Gid: ( 0/ root)
Access: 2017-05-19 16:33:00.148955354 +0800
Modify: 2017-05-19 16:33:00.148955354 +0800
Change: 2017-05-19 16:33:00.148955354 +0800
# exit
# exit
# stat /proc/14124
File: `/proc/14124'
Size: 0 Blocks: 0 IO Block: 1024 directory
Device: 3h/3d Inode: 1040776053 Links: 7
Access: (0555/dr-xr-xr-x) Uid: ( 0/ root) Gid: ( 0/ root)
Access: 2017-05-19 16:33:00.148955354 +0800
Modify: 2017-05-19 16:33:00.148955354 +0800
Change: 2017-05-19 16:33:00.148955354 +0800
# docker exec -it vm101 /bin/bash
[root@7f5a636ff840 /]# stat /proc/1
File: `/proc/1'
Size: 0 Blocks: 0 IO Block: 1024 directory
Device: 20h/32d Inode: 1040776094 Links: 7
Access: (0555/dr-xr-xr-x) Uid: ( 0/ root) Gid: ( 0/ root)
Access: 2017-05-19 08:33:00.829295372 +0000
Modify: 2017-05-19 08:33:00.829295372 +0000
Change: 2017-05-19 08:33:00.829295372 +0000
# stat /proc
File: `/proc'
Size: 0 Blocks: 0 IO Block: 1024 directory
Device: 20h/32d Inode: 1 Links: 260
Access: (0555/dr-xr-xr-x) Uid: ( 0/ root) Gid: ( 0/ root)
Access: 2017-05-19 08:33:00.148955354 +0000
Modify: 2017-05-19 08:33:00.148955354 +0000
Change: 2017-05-19 08:33:00.148955354 +0000
相同pid namespace看到的/proc/$pid 对应的inode是相同的，而且/proc对应的inode也是相同的。
实际上，每个pid namespace只会对应一个procfs实例：
static struct dentry *proc_mount(struct file_system_type *fs_type,
int flags, const char *dev_name, void *data)
{
int err;
struct super_block *sb;
struct pid_namespace *ns;
char *options;
if (flags & MS_KERNMOUNT) {
ns = (struct pid_namespace *)data;
options = NULL;
} else {
ns = task_active_pid_ns(current); ///pid namespace
options = data;
if (!current_user_ns()->may_mount_proc ||
!ns_capable(ns->user_ns, CAP_SYS_ADMIN))
return ERR_PTR(-EPERM);
}
sb = sget(fs_type, proc_test_super, proc_set_super, flags, ns);
}
struct super_block *sget(struct file_system_type *type,
int (*test)(struct super_block *,void *),
int (*set)(struct super_block *,void *),
int flags,
void *data)
{
struct super_block *s = NULL;
struct super_block *old;
int err;
retry:
spin_lock(&sb_lock);
if (test) {
hlist_for_each_entry(old, &type->fs_supers, s_instances) {
if (!test(old, data))///proc_test_super
continue;
if (!grab_super(old)) ///old pidname == data
goto retry;
if (s) {
up_write(&s->s_umount);
destroy_super(s);
s = NULL;
}
return old;
}
}
///...
}
static int proc_test_super(struct super_block *sb, void *data)
{
return sb->s_fs_info == data;
}
如果pid namespace已经存在一个procfs，直接使用已经存在的procfs实例。
# ./funcgraph proc_mount
Tracing "proc_mount"... Ctrl-C to end.
4) | proc_mount() {
4) 0.102 us | task_active_pid_ns();
4) | ns_capable() {
4) 0.050 us | cap_capable();
4) 0.466 us | }
4) | sget() {
4) 0.056 us | _raw_spin_lock();
4) 0.140 us | proc_test_super();
4) | grab_super() {
4) 0.116 us | down_write();
4) | put_super() {
4) 0.052 us | _raw_spin_lock();
4) 0.065 us | __put_super();
4) 0.819 us | }
4) 1.910 us | }
4) 3.375 us | }
4) 0.124 us | proc_parse_options();
4) 0.084 us | _raw_spin_lock();
4) 6.729 us | }
drop cache
/// proc_sys_call_handler call this
int drop_caches_sysctl_handler(ctl_table *table, int write,
void __user *buffer, size_t *length, loff_t *ppos)
{
int ret;
ret = proc_dointvec_minmax(table, write, buffer, length, ppos);
if (ret)
return ret;
if (write) {
if (sysctl_drop_caches & 1) ///drop page cache
iterate_supers(drop_pagecache_sb, NULL);
if (sysctl_drop_caches & 2)///drop slab
drop_slab();
}
return 0;
}
drop_slab -> shrink_slab -> do_shrinker_shrink -> prune_super -> prune_dcache_sb / prune_icache_sb.