2005-04-17 06:20:36 +08:00
|
|
|
/*
|
2007-09-20 16:31:38 +08:00
|
|
|
* fs/sysfs/file.c - sysfs regular (text) file implementation
|
|
|
|
*
|
|
|
|
* Copyright (c) 2001-3 Patrick Mochel
|
|
|
|
* Copyright (c) 2007 SUSE Linux Products GmbH
|
|
|
|
* Copyright (c) 2007 Tejun Heo <teheo@suse.de>
|
|
|
|
*
|
|
|
|
* This file is released under the GPLv2.
|
|
|
|
*
|
|
|
|
* Please see Documentation/filesystems/sysfs.txt for more information.
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/kobject.h>
|
2008-03-05 07:09:07 +08:00
|
|
|
#include <linux/kallsyms.h>
|
2008-03-14 10:41:52 +08:00
|
|
|
#include <linux/slab.h>
|
2008-06-16 19:46:47 +08:00
|
|
|
#include <linux/fsnotify.h>
|
2005-06-23 15:09:12 +08:00
|
|
|
#include <linux/namei.h>
|
2006-03-20 14:53:53 +08:00
|
|
|
#include <linux/poll.h>
|
2006-12-20 17:52:44 +08:00
|
|
|
#include <linux/list.h>
|
2007-07-26 19:03:54 +08:00
|
|
|
#include <linux/mutex.h>
|
2007-08-25 07:11:54 +08:00
|
|
|
#include <linux/limits.h>
|
2013-08-22 07:34:59 +08:00
|
|
|
#include <linux/uaccess.h>
|
2013-10-02 05:42:02 +08:00
|
|
|
#include <linux/seq_file.h>
|
2013-10-02 05:42:07 +08:00
|
|
|
#include <linux/mm.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
#include "sysfs.h"
|
|
|
|
|
2007-09-20 15:05:12 +08:00
|
|
|
/*
|
2013-10-02 05:41:59 +08:00
|
|
|
* There's one sysfs_open_file for each open file and one sysfs_open_dirent
|
2013-10-02 05:41:58 +08:00
|
|
|
* for each sysfs_dirent with one or more open files.
|
2007-09-20 15:05:12 +08:00
|
|
|
*
|
2013-10-02 05:41:58 +08:00
|
|
|
* sysfs_dirent->s_attr.open points to sysfs_open_dirent. s_attr.open is
|
|
|
|
* protected by sysfs_open_dirent_lock.
|
|
|
|
*
|
2013-10-02 05:42:02 +08:00
|
|
|
* filp->private_data points to seq_file whose ->private points to
|
|
|
|
* sysfs_open_file. sysfs_open_files are chained at
|
2013-10-02 05:41:59 +08:00
|
|
|
* sysfs_open_dirent->files, which is protected by sysfs_open_file_mutex.
|
2007-09-20 15:05:12 +08:00
|
|
|
*/
|
2007-11-22 06:55:19 +08:00
|
|
|
static DEFINE_SPINLOCK(sysfs_open_dirent_lock);
|
2013-10-02 05:41:58 +08:00
|
|
|
static DEFINE_MUTEX(sysfs_open_file_mutex);
|
2007-09-20 15:05:12 +08:00
|
|
|
|
|
|
|
struct sysfs_open_dirent {
|
|
|
|
atomic_t refcnt;
|
2007-09-20 15:05:12 +08:00
|
|
|
atomic_t event;
|
|
|
|
wait_queue_head_t poll;
|
2013-10-02 05:41:59 +08:00
|
|
|
struct list_head files; /* goes through sysfs_open_file.list */
|
2007-09-20 15:05:12 +08:00
|
|
|
};
|
|
|
|
|
2013-10-02 05:42:02 +08:00
|
|
|
static struct sysfs_open_file *sysfs_of(struct file *file)
|
|
|
|
{
|
|
|
|
return ((struct seq_file *)file->private_data)->private;
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:21 +08:00
|
|
|
/*
|
|
|
|
* Determine the kernfs_ops for the given sysfs_dirent. This function must
|
|
|
|
* be called while holding an active reference.
|
|
|
|
*/
|
|
|
|
static const struct kernfs_ops *kernfs_ops(struct sysfs_dirent *sd)
|
|
|
|
{
|
2013-11-29 03:54:29 +08:00
|
|
|
if (sd->s_flags & SYSFS_FLAG_LOCKDEP)
|
2013-11-29 03:54:21 +08:00
|
|
|
lockdep_assert_held(sd);
|
|
|
|
return sd->s_attr.ops;
|
|
|
|
}
|
|
|
|
|
2013-10-02 05:41:57 +08:00
|
|
|
/*
|
|
|
|
* Determine ktype->sysfs_ops for the given sysfs_dirent. This function
|
|
|
|
* must be called while holding an active reference.
|
|
|
|
*/
|
|
|
|
static const struct sysfs_ops *sysfs_file_ops(struct sysfs_dirent *sd)
|
|
|
|
{
|
2013-11-29 03:54:14 +08:00
|
|
|
struct kobject *kobj = sd->s_parent->priv;
|
2013-10-02 05:41:57 +08:00
|
|
|
|
2013-11-29 03:54:29 +08:00
|
|
|
if (sd->s_flags & SYSFS_FLAG_LOCKDEP)
|
2013-10-14 21:27:11 +08:00
|
|
|
lockdep_assert_held(sd);
|
2013-10-02 05:41:57 +08:00
|
|
|
return kobj->ktype ? kobj->ktype->sysfs_ops : NULL;
|
|
|
|
}
|
|
|
|
|
2013-10-02 05:42:02 +08:00
|
|
|
/*
|
|
|
|
* Reads on sysfs are handled through seq_file, which takes care of hairy
|
|
|
|
* details like buffering and seeking. The following function pipes
|
|
|
|
* sysfs_ops->show() result through seq_file.
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
2013-11-29 03:54:16 +08:00
|
|
|
static int sysfs_kf_seq_show(struct seq_file *sf, void *v)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2013-10-02 05:42:02 +08:00
|
|
|
struct sysfs_open_file *of = sf->private;
|
2013-11-29 03:54:14 +08:00
|
|
|
struct kobject *kobj = of->sd->s_parent->priv;
|
2013-11-29 03:54:16 +08:00
|
|
|
const struct sysfs_ops *ops = sysfs_file_ops(of->sd);
|
2005-04-17 06:20:36 +08:00
|
|
|
ssize_t count;
|
2013-11-29 03:54:16 +08:00
|
|
|
char *buf;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-10-02 05:42:02 +08:00
|
|
|
/* acquire buffer and ensure that it's >= PAGE_SIZE */
|
|
|
|
count = seq_get_buf(sf, &buf);
|
|
|
|
if (count < PAGE_SIZE) {
|
|
|
|
seq_commit(sf, -1);
|
|
|
|
return 0;
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-10-02 05:42:02 +08:00
|
|
|
/*
|
2013-11-29 03:54:16 +08:00
|
|
|
* Invoke show(). Control may reach here via seq file lseek even
|
|
|
|
* if @ops->show() isn't implemented.
|
2013-10-02 05:42:02 +08:00
|
|
|
*/
|
2013-11-29 03:54:16 +08:00
|
|
|
if (ops->show) {
|
2013-11-29 03:54:14 +08:00
|
|
|
count = ops->show(kobj, of->sd->priv, buf);
|
2013-11-29 03:54:16 +08:00
|
|
|
if (count < 0)
|
|
|
|
return count;
|
|
|
|
}
|
2007-06-14 02:45:16 +08:00
|
|
|
|
2007-11-22 06:55:19 +08:00
|
|
|
/*
|
|
|
|
* The code works fine with PAGE_SIZE return but it's likely to
|
|
|
|
* indicate truncated result or overflow in normal use cases.
|
|
|
|
*/
|
2008-03-05 07:09:07 +08:00
|
|
|
if (count >= (ssize_t)PAGE_SIZE) {
|
|
|
|
print_symbol("fill_read_buffer: %s returned bad count\n",
|
|
|
|
(unsigned long)ops->show);
|
|
|
|
/* Try to struggle along */
|
|
|
|
count = PAGE_SIZE - 1;
|
|
|
|
}
|
2013-10-02 05:42:02 +08:00
|
|
|
seq_commit(sf, count);
|
|
|
|
return 0;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:16 +08:00
|
|
|
static ssize_t sysfs_kf_bin_read(struct sysfs_open_file *of, char *buf,
|
|
|
|
size_t count, loff_t pos)
|
2013-10-02 05:42:06 +08:00
|
|
|
{
|
2013-11-29 03:54:14 +08:00
|
|
|
struct bin_attribute *battr = of->sd->priv;
|
|
|
|
struct kobject *kobj = of->sd->s_parent->priv;
|
2013-11-29 03:54:16 +08:00
|
|
|
loff_t size = file_inode(of->file)->i_size;
|
2013-10-02 05:42:06 +08:00
|
|
|
|
2013-11-29 03:54:16 +08:00
|
|
|
if (!count)
|
2013-10-02 05:42:06 +08:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (size) {
|
2013-11-29 03:54:16 +08:00
|
|
|
if (pos > size)
|
2013-10-02 05:42:06 +08:00
|
|
|
return 0;
|
2013-11-29 03:54:16 +08:00
|
|
|
if (pos + count > size)
|
|
|
|
count = size - pos;
|
2013-10-02 05:42:06 +08:00
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:16 +08:00
|
|
|
if (!battr->read)
|
|
|
|
return -EIO;
|
|
|
|
|
|
|
|
return battr->read(of->file, kobj, battr, buf, pos, count);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void *kernfs_seq_start(struct seq_file *sf, loff_t *ppos)
|
|
|
|
{
|
|
|
|
struct sysfs_open_file *of = sf->private;
|
2013-11-29 03:54:26 +08:00
|
|
|
const struct kernfs_ops *ops;
|
2013-11-29 03:54:16 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* @of->mutex nests outside active ref and is just to ensure that
|
|
|
|
* the ops aren't called concurrently for the same open file.
|
|
|
|
*/
|
|
|
|
mutex_lock(&of->mutex);
|
|
|
|
if (!sysfs_get_active(of->sd))
|
|
|
|
return ERR_PTR(-ENODEV);
|
|
|
|
|
2013-11-29 03:54:26 +08:00
|
|
|
ops = kernfs_ops(of->sd);
|
|
|
|
if (ops->seq_start) {
|
|
|
|
return ops->seq_start(sf, ppos);
|
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* The same behavior and code as single_open(). Returns
|
|
|
|
* !NULL if pos is at the beginning; otherwise, NULL.
|
|
|
|
*/
|
|
|
|
return NULL + !*ppos;
|
|
|
|
}
|
2013-11-29 03:54:16 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void *kernfs_seq_next(struct seq_file *sf, void *v, loff_t *ppos)
|
|
|
|
{
|
2013-11-29 03:54:26 +08:00
|
|
|
struct sysfs_open_file *of = sf->private;
|
|
|
|
const struct kernfs_ops *ops = kernfs_ops(of->sd);
|
|
|
|
|
|
|
|
if (ops->seq_next) {
|
|
|
|
return ops->seq_next(sf, v, ppos);
|
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* The same behavior and code as single_open(), always
|
|
|
|
* terminate after the initial read.
|
|
|
|
*/
|
|
|
|
++*ppos;
|
|
|
|
return NULL;
|
|
|
|
}
|
2013-11-29 03:54:16 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void kernfs_seq_stop(struct seq_file *sf, void *v)
|
|
|
|
{
|
|
|
|
struct sysfs_open_file *of = sf->private;
|
2013-11-29 03:54:26 +08:00
|
|
|
const struct kernfs_ops *ops = kernfs_ops(of->sd);
|
|
|
|
|
|
|
|
if (ops->seq_stop)
|
|
|
|
ops->seq_stop(sf, v);
|
2013-11-29 03:54:16 +08:00
|
|
|
|
|
|
|
sysfs_put_active(of->sd);
|
|
|
|
mutex_unlock(&of->mutex);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int kernfs_seq_show(struct seq_file *sf, void *v)
|
|
|
|
{
|
|
|
|
struct sysfs_open_file *of = sf->private;
|
|
|
|
|
|
|
|
of->event = atomic_read(&of->sd->s_attr.open->event);
|
|
|
|
|
2013-11-29 03:54:21 +08:00
|
|
|
return of->sd->s_attr.ops->seq_show(sf, v);
|
2013-11-29 03:54:16 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static const struct seq_operations kernfs_seq_ops = {
|
|
|
|
.start = kernfs_seq_start,
|
|
|
|
.next = kernfs_seq_next,
|
|
|
|
.stop = kernfs_seq_stop,
|
|
|
|
.show = kernfs_seq_show,
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* As reading a bin file can have side-effects, the exact offset and bytes
|
|
|
|
* specified in read(2) call should be passed to the read callback making
|
|
|
|
* it difficult to use seq_file. Implement simplistic custom buffering for
|
|
|
|
* bin files.
|
|
|
|
*/
|
|
|
|
static ssize_t kernfs_file_direct_read(struct sysfs_open_file *of,
|
|
|
|
char __user *user_buf, size_t count,
|
|
|
|
loff_t *ppos)
|
|
|
|
{
|
|
|
|
ssize_t len = min_t(size_t, count, PAGE_SIZE);
|
2013-11-29 03:54:21 +08:00
|
|
|
const struct kernfs_ops *ops;
|
2013-11-29 03:54:16 +08:00
|
|
|
char *buf;
|
|
|
|
|
|
|
|
buf = kmalloc(len, GFP_KERNEL);
|
2013-10-02 05:42:06 +08:00
|
|
|
if (!buf)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
2013-11-29 03:54:16 +08:00
|
|
|
/*
|
|
|
|
* @of->mutex nests outside active ref and is just to ensure that
|
|
|
|
* the ops aren't called concurrently for the same open file.
|
|
|
|
*/
|
2013-10-02 05:42:06 +08:00
|
|
|
mutex_lock(&of->mutex);
|
|
|
|
if (!sysfs_get_active(of->sd)) {
|
2013-11-29 03:54:16 +08:00
|
|
|
len = -ENODEV;
|
2013-10-02 05:42:06 +08:00
|
|
|
mutex_unlock(&of->mutex);
|
|
|
|
goto out_free;
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:21 +08:00
|
|
|
ops = kernfs_ops(of->sd);
|
|
|
|
if (ops->read)
|
|
|
|
len = ops->read(of, buf, len, *ppos);
|
|
|
|
else
|
|
|
|
len = -EINVAL;
|
2013-10-02 05:42:06 +08:00
|
|
|
|
|
|
|
sysfs_put_active(of->sd);
|
|
|
|
mutex_unlock(&of->mutex);
|
|
|
|
|
2013-11-29 03:54:16 +08:00
|
|
|
if (len < 0)
|
2013-10-02 05:42:06 +08:00
|
|
|
goto out_free;
|
|
|
|
|
2013-11-29 03:54:16 +08:00
|
|
|
if (copy_to_user(user_buf, buf, len)) {
|
|
|
|
len = -EFAULT;
|
2013-10-02 05:42:06 +08:00
|
|
|
goto out_free;
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:16 +08:00
|
|
|
*ppos += len;
|
2013-10-02 05:42:06 +08:00
|
|
|
|
|
|
|
out_free:
|
|
|
|
kfree(buf);
|
2013-11-29 03:54:16 +08:00
|
|
|
return len;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* kernfs_file_read - kernfs vfs read callback
|
|
|
|
* @file: file pointer
|
|
|
|
* @user_buf: data to write
|
|
|
|
* @count: number of bytes
|
|
|
|
* @ppos: starting offset
|
|
|
|
*/
|
|
|
|
static ssize_t kernfs_file_read(struct file *file, char __user *user_buf,
|
|
|
|
size_t count, loff_t *ppos)
|
|
|
|
{
|
|
|
|
struct sysfs_open_file *of = sysfs_of(file);
|
|
|
|
|
2013-11-29 03:54:21 +08:00
|
|
|
if (of->sd->s_flags & SYSFS_FLAG_HAS_SEQ_SHOW)
|
2013-11-29 03:54:16 +08:00
|
|
|
return seq_read(file, user_buf, count, ppos);
|
2013-11-29 03:54:21 +08:00
|
|
|
else
|
|
|
|
return kernfs_file_direct_read(of, user_buf, count, ppos);
|
2013-10-02 05:42:06 +08:00
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:17 +08:00
|
|
|
/* kernfs write callback for regular sysfs files */
|
|
|
|
static ssize_t sysfs_kf_write(struct sysfs_open_file *of, char *buf,
|
|
|
|
size_t count, loff_t pos)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2013-11-29 03:54:17 +08:00
|
|
|
const struct sysfs_ops *ops = sysfs_file_ops(of->sd);
|
2013-11-29 03:54:14 +08:00
|
|
|
struct kobject *kobj = of->sd->s_parent->priv;
|
2007-06-14 02:45:16 +08:00
|
|
|
|
2013-11-29 03:54:17 +08:00
|
|
|
if (!count)
|
|
|
|
return 0;
|
2007-06-14 02:45:16 +08:00
|
|
|
|
2013-11-29 03:54:17 +08:00
|
|
|
return ops->store(kobj, of->sd->priv, buf, count);
|
|
|
|
}
|
2013-10-02 05:42:05 +08:00
|
|
|
|
2013-11-29 03:54:17 +08:00
|
|
|
/* kernfs write callback for bin sysfs files */
|
|
|
|
static ssize_t sysfs_kf_bin_write(struct sysfs_open_file *of, char *buf,
|
|
|
|
size_t count, loff_t pos)
|
|
|
|
{
|
|
|
|
struct bin_attribute *battr = of->sd->priv;
|
|
|
|
struct kobject *kobj = of->sd->s_parent->priv;
|
|
|
|
loff_t size = file_inode(of->file)->i_size;
|
2013-10-02 05:42:05 +08:00
|
|
|
|
2013-11-29 03:54:17 +08:00
|
|
|
if (size) {
|
|
|
|
if (size <= pos)
|
|
|
|
return 0;
|
|
|
|
count = min_t(ssize_t, count, size - pos);
|
2013-10-02 05:42:05 +08:00
|
|
|
}
|
2013-11-29 03:54:17 +08:00
|
|
|
if (!count)
|
|
|
|
return 0;
|
2007-06-14 02:45:16 +08:00
|
|
|
|
2013-11-29 03:54:17 +08:00
|
|
|
if (!battr->write)
|
|
|
|
return -EIO;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-11-29 03:54:17 +08:00
|
|
|
return battr->write(of->file, kobj, battr, buf, pos, count);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2013-11-29 03:54:17 +08:00
|
|
|
* kernfs_file_write - kernfs vfs write callback
|
2013-10-02 05:42:01 +08:00
|
|
|
* @file: file pointer
|
|
|
|
* @user_buf: data to write
|
|
|
|
* @count: number of bytes
|
|
|
|
* @ppos: starting offset
|
|
|
|
*
|
2013-11-29 03:54:17 +08:00
|
|
|
* Copy data in from userland and pass it to the matching kernfs write
|
|
|
|
* operation.
|
2005-04-17 06:20:36 +08:00
|
|
|
*
|
2013-10-02 05:42:01 +08:00
|
|
|
* There is no easy way for us to know if userspace is only doing a partial
|
|
|
|
* write, so we don't support them. We expect the entire buffer to come on
|
|
|
|
* the first write. Hint: if you're writing a value, first read the file,
|
|
|
|
* modify only the the value you're changing, then write entire buffer
|
|
|
|
* back.
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
2013-11-29 03:54:17 +08:00
|
|
|
static ssize_t kernfs_file_write(struct file *file, const char __user *user_buf,
|
|
|
|
size_t count, loff_t *ppos)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2013-10-02 05:42:02 +08:00
|
|
|
struct sysfs_open_file *of = sysfs_of(file);
|
2013-10-02 05:42:05 +08:00
|
|
|
ssize_t len = min_t(size_t, count, PAGE_SIZE);
|
2013-11-29 03:54:21 +08:00
|
|
|
const struct kernfs_ops *ops;
|
2013-10-02 05:42:01 +08:00
|
|
|
char *buf;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-10-02 05:42:01 +08:00
|
|
|
buf = kmalloc(len + 1, GFP_KERNEL);
|
|
|
|
if (!buf)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
if (copy_from_user(buf, user_buf, len)) {
|
|
|
|
len = -EFAULT;
|
|
|
|
goto out_free;
|
|
|
|
}
|
|
|
|
buf[len] = '\0'; /* guarantee string termination */
|
|
|
|
|
2013-11-29 03:54:17 +08:00
|
|
|
/*
|
|
|
|
* @of->mutex nests outside active ref and is just to ensure that
|
|
|
|
* the ops aren't called concurrently for the same open file.
|
|
|
|
*/
|
|
|
|
mutex_lock(&of->mutex);
|
|
|
|
if (!sysfs_get_active(of->sd)) {
|
|
|
|
mutex_unlock(&of->mutex);
|
|
|
|
len = -ENODEV;
|
|
|
|
goto out_free;
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:21 +08:00
|
|
|
ops = kernfs_ops(of->sd);
|
|
|
|
if (ops->write)
|
|
|
|
len = ops->write(of, buf, len, *ppos);
|
2013-11-29 03:54:17 +08:00
|
|
|
else
|
2013-11-29 03:54:21 +08:00
|
|
|
len = -EINVAL;
|
2013-11-29 03:54:17 +08:00
|
|
|
|
|
|
|
sysfs_put_active(of->sd);
|
|
|
|
mutex_unlock(&of->mutex);
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
if (len > 0)
|
|
|
|
*ppos += len;
|
2013-10-02 05:42:01 +08:00
|
|
|
out_free:
|
|
|
|
kfree(buf);
|
2005-04-17 06:20:36 +08:00
|
|
|
return len;
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:18 +08:00
|
|
|
static int sysfs_kf_bin_mmap(struct sysfs_open_file *of,
|
|
|
|
struct vm_area_struct *vma)
|
|
|
|
{
|
|
|
|
struct bin_attribute *battr = of->sd->priv;
|
|
|
|
struct kobject *kobj = of->sd->s_parent->priv;
|
|
|
|
|
|
|
|
if (!battr->mmap)
|
|
|
|
return -ENODEV;
|
|
|
|
|
|
|
|
return battr->mmap(of->file, kobj, battr, vma);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void kernfs_vma_open(struct vm_area_struct *vma)
|
2013-10-02 05:42:07 +08:00
|
|
|
{
|
|
|
|
struct file *file = vma->vm_file;
|
|
|
|
struct sysfs_open_file *of = sysfs_of(file);
|
|
|
|
|
|
|
|
if (!of->vm_ops)
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (!sysfs_get_active(of->sd))
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (of->vm_ops->open)
|
|
|
|
of->vm_ops->open(vma);
|
|
|
|
|
|
|
|
sysfs_put_active(of->sd);
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:18 +08:00
|
|
|
static int kernfs_vma_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
|
2013-10-02 05:42:07 +08:00
|
|
|
{
|
|
|
|
struct file *file = vma->vm_file;
|
|
|
|
struct sysfs_open_file *of = sysfs_of(file);
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (!of->vm_ops)
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
|
|
|
|
if (!sysfs_get_active(of->sd))
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
|
|
|
|
ret = VM_FAULT_SIGBUS;
|
|
|
|
if (of->vm_ops->fault)
|
|
|
|
ret = of->vm_ops->fault(vma, vmf);
|
|
|
|
|
|
|
|
sysfs_put_active(of->sd);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:18 +08:00
|
|
|
static int kernfs_vma_page_mkwrite(struct vm_area_struct *vma,
|
|
|
|
struct vm_fault *vmf)
|
2013-10-02 05:42:07 +08:00
|
|
|
{
|
|
|
|
struct file *file = vma->vm_file;
|
|
|
|
struct sysfs_open_file *of = sysfs_of(file);
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (!of->vm_ops)
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
|
|
|
|
if (!sysfs_get_active(of->sd))
|
|
|
|
return VM_FAULT_SIGBUS;
|
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
if (of->vm_ops->page_mkwrite)
|
|
|
|
ret = of->vm_ops->page_mkwrite(vma, vmf);
|
|
|
|
else
|
|
|
|
file_update_time(file);
|
|
|
|
|
|
|
|
sysfs_put_active(of->sd);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:18 +08:00
|
|
|
static int kernfs_vma_access(struct vm_area_struct *vma, unsigned long addr,
|
|
|
|
void *buf, int len, int write)
|
2013-10-02 05:42:07 +08:00
|
|
|
{
|
|
|
|
struct file *file = vma->vm_file;
|
|
|
|
struct sysfs_open_file *of = sysfs_of(file);
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (!of->vm_ops)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (!sysfs_get_active(of->sd))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
ret = -EINVAL;
|
|
|
|
if (of->vm_ops->access)
|
|
|
|
ret = of->vm_ops->access(vma, addr, buf, len, write);
|
|
|
|
|
|
|
|
sysfs_put_active(of->sd);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef CONFIG_NUMA
|
2013-11-29 03:54:18 +08:00
|
|
|
static int kernfs_vma_set_policy(struct vm_area_struct *vma,
|
|
|
|
struct mempolicy *new)
|
2013-10-02 05:42:07 +08:00
|
|
|
{
|
|
|
|
struct file *file = vma->vm_file;
|
|
|
|
struct sysfs_open_file *of = sysfs_of(file);
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (!of->vm_ops)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (!sysfs_get_active(of->sd))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
if (of->vm_ops->set_policy)
|
|
|
|
ret = of->vm_ops->set_policy(vma, new);
|
|
|
|
|
|
|
|
sysfs_put_active(of->sd);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:18 +08:00
|
|
|
static struct mempolicy *kernfs_vma_get_policy(struct vm_area_struct *vma,
|
|
|
|
unsigned long addr)
|
2013-10-02 05:42:07 +08:00
|
|
|
{
|
|
|
|
struct file *file = vma->vm_file;
|
|
|
|
struct sysfs_open_file *of = sysfs_of(file);
|
|
|
|
struct mempolicy *pol;
|
|
|
|
|
|
|
|
if (!of->vm_ops)
|
|
|
|
return vma->vm_policy;
|
|
|
|
|
|
|
|
if (!sysfs_get_active(of->sd))
|
|
|
|
return vma->vm_policy;
|
|
|
|
|
|
|
|
pol = vma->vm_policy;
|
|
|
|
if (of->vm_ops->get_policy)
|
|
|
|
pol = of->vm_ops->get_policy(vma, addr);
|
|
|
|
|
|
|
|
sysfs_put_active(of->sd);
|
|
|
|
return pol;
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:18 +08:00
|
|
|
static int kernfs_vma_migrate(struct vm_area_struct *vma,
|
|
|
|
const nodemask_t *from, const nodemask_t *to,
|
|
|
|
unsigned long flags)
|
2013-10-02 05:42:07 +08:00
|
|
|
{
|
|
|
|
struct file *file = vma->vm_file;
|
|
|
|
struct sysfs_open_file *of = sysfs_of(file);
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (!of->vm_ops)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (!sysfs_get_active(of->sd))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
if (of->vm_ops->migrate)
|
|
|
|
ret = of->vm_ops->migrate(vma, from, to, flags);
|
|
|
|
|
|
|
|
sysfs_put_active(of->sd);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2013-11-29 03:54:18 +08:00
|
|
|
static const struct vm_operations_struct kernfs_vm_ops = {
|
|
|
|
.open = kernfs_vma_open,
|
|
|
|
.fault = kernfs_vma_fault,
|
|
|
|
.page_mkwrite = kernfs_vma_page_mkwrite,
|
|
|
|
.access = kernfs_vma_access,
|
2013-10-02 05:42:07 +08:00
|
|
|
#ifdef CONFIG_NUMA
|
2013-11-29 03:54:18 +08:00
|
|
|
.set_policy = kernfs_vma_set_policy,
|
|
|
|
.get_policy = kernfs_vma_get_policy,
|
|
|
|
.migrate = kernfs_vma_migrate,
|
2013-10-02 05:42:07 +08:00
|
|
|
#endif
|
|
|
|
};
|
|
|
|
|
2013-11-29 03:54:18 +08:00
|
|
|
static int kernfs_file_mmap(struct file *file, struct vm_area_struct *vma)
|
2013-10-02 05:42:07 +08:00
|
|
|
{
|
|
|
|
struct sysfs_open_file *of = sysfs_of(file);
|
2013-11-29 03:54:21 +08:00
|
|
|
const struct kernfs_ops *ops;
|
2013-10-02 05:42:07 +08:00
|
|
|
int rc;
|
|
|
|
|
|
|
|
mutex_lock(&of->mutex);
|
|
|
|
|
|
|
|
rc = -ENODEV;
|
|
|
|
if (!sysfs_get_active(of->sd))
|
|
|
|
goto out_unlock;
|
|
|
|
|
2013-11-29 03:54:21 +08:00
|
|
|
ops = kernfs_ops(of->sd);
|
|
|
|
if (ops->mmap)
|
|
|
|
rc = ops->mmap(of, vma);
|
2013-10-02 05:42:07 +08:00
|
|
|
if (rc)
|
|
|
|
goto out_put;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* PowerPC's pci_mmap of legacy_mem uses shmem_zero_setup()
|
|
|
|
* to satisfy versions of X which crash if the mmap fails: that
|
|
|
|
* substitutes a new vm_file, and we don't then want bin_vm_ops.
|
|
|
|
*/
|
|
|
|
if (vma->vm_file != file)
|
|
|
|
goto out_put;
|
|
|
|
|
|
|
|
rc = -EINVAL;
|
|
|
|
if (of->mmapped && of->vm_ops != vma->vm_ops)
|
|
|
|
goto out_put;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* It is not possible to successfully wrap close.
|
|
|
|
* So error if someone is trying to use close.
|
|
|
|
*/
|
|
|
|
rc = -EINVAL;
|
|
|
|
if (vma->vm_ops && vma->vm_ops->close)
|
|
|
|
goto out_put;
|
|
|
|
|
|
|
|
rc = 0;
|
|
|
|
of->mmapped = 1;
|
|
|
|
of->vm_ops = vma->vm_ops;
|
2013-11-29 03:54:18 +08:00
|
|
|
vma->vm_ops = &kernfs_vm_ops;
|
2013-10-02 05:42:07 +08:00
|
|
|
out_put:
|
|
|
|
sysfs_put_active(of->sd);
|
|
|
|
out_unlock:
|
|
|
|
mutex_unlock(&of->mutex);
|
|
|
|
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
2007-09-20 15:05:12 +08:00
|
|
|
/**
|
|
|
|
* sysfs_get_open_dirent - get or create sysfs_open_dirent
|
|
|
|
* @sd: target sysfs_dirent
|
2013-10-02 05:41:59 +08:00
|
|
|
* @of: sysfs_open_file for this instance of open
|
2007-09-20 15:05:12 +08:00
|
|
|
*
|
|
|
|
* If @sd->s_attr.open exists, increment its reference count;
|
2013-10-02 05:41:59 +08:00
|
|
|
* otherwise, create one. @of is chained to the files list.
|
2007-09-20 15:05:12 +08:00
|
|
|
*
|
|
|
|
* LOCKING:
|
|
|
|
* Kernel thread context (may sleep).
|
|
|
|
*
|
|
|
|
* RETURNS:
|
|
|
|
* 0 on success, -errno on failure.
|
|
|
|
*/
|
|
|
|
static int sysfs_get_open_dirent(struct sysfs_dirent *sd,
|
2013-10-02 05:41:59 +08:00
|
|
|
struct sysfs_open_file *of)
|
2007-09-20 15:05:12 +08:00
|
|
|
{
|
|
|
|
struct sysfs_open_dirent *od, *new_od = NULL;
|
|
|
|
|
|
|
|
retry:
|
2013-10-02 05:41:58 +08:00
|
|
|
mutex_lock(&sysfs_open_file_mutex);
|
2009-09-16 07:05:51 +08:00
|
|
|
spin_lock_irq(&sysfs_open_dirent_lock);
|
2007-09-20 15:05:12 +08:00
|
|
|
|
|
|
|
if (!sd->s_attr.open && new_od) {
|
|
|
|
sd->s_attr.open = new_od;
|
|
|
|
new_od = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
od = sd->s_attr.open;
|
|
|
|
if (od) {
|
|
|
|
atomic_inc(&od->refcnt);
|
2013-10-02 05:41:59 +08:00
|
|
|
list_add_tail(&of->list, &od->files);
|
2007-09-20 15:05:12 +08:00
|
|
|
}
|
|
|
|
|
2009-09-16 07:05:51 +08:00
|
|
|
spin_unlock_irq(&sysfs_open_dirent_lock);
|
2013-10-02 05:41:58 +08:00
|
|
|
mutex_unlock(&sysfs_open_file_mutex);
|
2007-09-20 15:05:12 +08:00
|
|
|
|
|
|
|
if (od) {
|
|
|
|
kfree(new_od);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* not there, initialize a new one and retry */
|
|
|
|
new_od = kmalloc(sizeof(*new_od), GFP_KERNEL);
|
|
|
|
if (!new_od)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
atomic_set(&new_od->refcnt, 0);
|
2007-09-20 15:05:12 +08:00
|
|
|
atomic_set(&new_od->event, 1);
|
|
|
|
init_waitqueue_head(&new_od->poll);
|
2013-10-02 05:41:59 +08:00
|
|
|
INIT_LIST_HEAD(&new_od->files);
|
2007-09-20 15:05:12 +08:00
|
|
|
goto retry;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* sysfs_put_open_dirent - put sysfs_open_dirent
|
|
|
|
* @sd: target sysfs_dirent
|
2013-10-02 05:41:59 +08:00
|
|
|
* @of: associated sysfs_open_file
|
2007-09-20 15:05:12 +08:00
|
|
|
*
|
2013-10-02 05:41:59 +08:00
|
|
|
* Put @sd->s_attr.open and unlink @of from the files list. If
|
|
|
|
* reference count reaches zero, disassociate and free it.
|
2007-09-20 15:05:12 +08:00
|
|
|
*
|
|
|
|
* LOCKING:
|
|
|
|
* None.
|
|
|
|
*/
|
|
|
|
static void sysfs_put_open_dirent(struct sysfs_dirent *sd,
|
2013-10-02 05:41:59 +08:00
|
|
|
struct sysfs_open_file *of)
|
2007-09-20 15:05:12 +08:00
|
|
|
{
|
|
|
|
struct sysfs_open_dirent *od = sd->s_attr.open;
|
2009-09-16 07:05:51 +08:00
|
|
|
unsigned long flags;
|
2007-09-20 15:05:12 +08:00
|
|
|
|
2013-10-02 05:41:58 +08:00
|
|
|
mutex_lock(&sysfs_open_file_mutex);
|
2009-09-16 07:05:51 +08:00
|
|
|
spin_lock_irqsave(&sysfs_open_dirent_lock, flags);
|
2007-09-20 15:05:12 +08:00
|
|
|
|
2013-10-02 05:42:07 +08:00
|
|
|
if (of)
|
|
|
|
list_del(&of->list);
|
|
|
|
|
2007-09-20 15:05:12 +08:00
|
|
|
if (atomic_dec_and_test(&od->refcnt))
|
|
|
|
sd->s_attr.open = NULL;
|
|
|
|
else
|
|
|
|
od = NULL;
|
|
|
|
|
2009-09-16 07:05:51 +08:00
|
|
|
spin_unlock_irqrestore(&sysfs_open_dirent_lock, flags);
|
2013-10-02 05:41:58 +08:00
|
|
|
mutex_unlock(&sysfs_open_file_mutex);
|
2007-09-20 15:05:12 +08:00
|
|
|
|
|
|
|
kfree(od);
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:19 +08:00
|
|
|
static int kernfs_file_open(struct inode *inode, struct file *file)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2007-06-14 02:45:15 +08:00
|
|
|
struct sysfs_dirent *attr_sd = file->f_path.dentry->d_fsdata;
|
2013-11-29 03:54:21 +08:00
|
|
|
const struct kernfs_ops *ops;
|
2013-10-02 05:41:59 +08:00
|
|
|
struct sysfs_open_file *of;
|
2013-11-17 10:17:36 +08:00
|
|
|
bool has_read, has_write, has_mmap;
|
2007-11-02 20:47:53 +08:00
|
|
|
int error = -EACCES;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2010-02-12 07:18:38 +08:00
|
|
|
if (!sysfs_get_active(attr_sd))
|
2007-06-14 02:45:16 +08:00
|
|
|
return -ENODEV;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-11-29 03:54:21 +08:00
|
|
|
ops = kernfs_ops(attr_sd);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-11-29 03:54:21 +08:00
|
|
|
has_read = ops->seq_show || ops->read || ops->mmap;
|
|
|
|
has_write = ops->write || ops->mmap;
|
|
|
|
has_mmap = ops->mmap;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-10-02 05:42:08 +08:00
|
|
|
/* check perms and supported operations */
|
|
|
|
if ((file->f_mode & FMODE_WRITE) &&
|
|
|
|
(!(inode->i_mode & S_IWUGO) || !has_write))
|
|
|
|
goto err_out;
|
|
|
|
|
|
|
|
if ((file->f_mode & FMODE_READ) &&
|
|
|
|
(!(inode->i_mode & S_IRUGO) || !has_read))
|
|
|
|
goto err_out;
|
|
|
|
|
2013-10-02 05:42:02 +08:00
|
|
|
/* allocate a sysfs_open_file for the file */
|
2007-06-14 02:45:16 +08:00
|
|
|
error = -ENOMEM;
|
2013-10-02 05:41:59 +08:00
|
|
|
of = kzalloc(sizeof(struct sysfs_open_file), GFP_KERNEL);
|
|
|
|
if (!of)
|
2007-06-14 02:45:17 +08:00
|
|
|
goto err_out;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-11-17 10:17:36 +08:00
|
|
|
/*
|
|
|
|
* The following is done to give a different lockdep key to
|
|
|
|
* @of->mutex for files which implement mmap. This is a rather
|
|
|
|
* crude way to avoid false positive lockdep warning around
|
|
|
|
* mm->mmap_sem - mmap nests @of->mutex under mm->mmap_sem and
|
|
|
|
* reading /sys/block/sda/trace/act_mask grabs sr_mutex, under
|
|
|
|
* which mm->mmap_sem nests, while holding @of->mutex. As each
|
|
|
|
* open file has a separate mutex, it's okay as long as those don't
|
|
|
|
* happen on the same file. At this point, we can't easily give
|
|
|
|
* each file a separate locking class. Let's differentiate on
|
|
|
|
* whether the file has mmap or not for now.
|
|
|
|
*/
|
|
|
|
if (has_mmap)
|
|
|
|
mutex_init(&of->mutex);
|
|
|
|
else
|
|
|
|
mutex_init(&of->mutex);
|
|
|
|
|
2013-10-02 05:42:00 +08:00
|
|
|
of->sd = attr_sd;
|
|
|
|
of->file = file;
|
2013-10-02 05:42:02 +08:00
|
|
|
|
|
|
|
/*
|
2013-10-02 05:42:08 +08:00
|
|
|
* Always instantiate seq_file even if read access doesn't use
|
|
|
|
* seq_file or is not requested. This unifies private data access
|
|
|
|
* and readable regular files are the vast majority anyway.
|
2013-10-02 05:42:02 +08:00
|
|
|
*/
|
2013-11-29 03:54:21 +08:00
|
|
|
if (ops->seq_show)
|
2013-11-29 03:54:16 +08:00
|
|
|
error = seq_open(file, &kernfs_seq_ops);
|
2013-11-29 03:54:21 +08:00
|
|
|
else
|
|
|
|
error = seq_open(file, NULL);
|
2013-10-02 05:42:02 +08:00
|
|
|
if (error)
|
|
|
|
goto err_free;
|
|
|
|
|
2013-11-29 03:54:16 +08:00
|
|
|
((struct seq_file *)file->private_data)->private = of;
|
|
|
|
|
2013-10-02 05:42:02 +08:00
|
|
|
/* seq_file clears PWRITE unconditionally, restore it if WRITE */
|
|
|
|
if (file->f_mode & FMODE_WRITE)
|
|
|
|
file->f_mode |= FMODE_PWRITE;
|
2007-06-14 02:45:16 +08:00
|
|
|
|
2007-09-20 15:05:12 +08:00
|
|
|
/* make sure we have open dirent struct */
|
2013-10-02 05:41:59 +08:00
|
|
|
error = sysfs_get_open_dirent(attr_sd, of);
|
2007-09-20 15:05:12 +08:00
|
|
|
if (error)
|
2013-10-02 05:42:02 +08:00
|
|
|
goto err_close;
|
2007-09-20 15:05:12 +08:00
|
|
|
|
2007-09-20 15:05:10 +08:00
|
|
|
/* open succeeded, put active references */
|
2010-02-12 07:18:38 +08:00
|
|
|
sysfs_put_active(attr_sd);
|
2007-06-14 02:45:16 +08:00
|
|
|
return 0;
|
|
|
|
|
2013-10-02 05:42:02 +08:00
|
|
|
err_close:
|
2013-11-29 03:54:16 +08:00
|
|
|
seq_release(inode, file);
|
2013-10-02 05:42:02 +08:00
|
|
|
err_free:
|
2013-10-02 05:41:59 +08:00
|
|
|
kfree(of);
|
2013-10-02 05:42:02 +08:00
|
|
|
err_out:
|
2010-02-12 07:18:38 +08:00
|
|
|
sysfs_put_active(attr_sd);
|
2005-04-17 06:20:36 +08:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:19 +08:00
|
|
|
static int kernfs_file_release(struct inode *inode, struct file *filp)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2007-09-20 15:05:12 +08:00
|
|
|
struct sysfs_dirent *sd = filp->f_path.dentry->d_fsdata;
|
2013-10-02 05:42:02 +08:00
|
|
|
struct sysfs_open_file *of = sysfs_of(filp);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-10-02 05:41:59 +08:00
|
|
|
sysfs_put_open_dirent(sd, of);
|
2013-11-29 03:54:16 +08:00
|
|
|
seq_release(inode, filp);
|
2013-10-02 05:41:59 +08:00
|
|
|
kfree(of);
|
2007-09-20 15:05:10 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2013-10-02 05:42:07 +08:00
|
|
|
void sysfs_unmap_bin_file(struct sysfs_dirent *sd)
|
|
|
|
{
|
|
|
|
struct sysfs_open_dirent *od;
|
|
|
|
struct sysfs_open_file *of;
|
|
|
|
|
2013-11-29 03:54:21 +08:00
|
|
|
if (!(sd->s_flags & SYSFS_FLAG_HAS_MMAP))
|
2013-10-02 05:42:07 +08:00
|
|
|
return;
|
|
|
|
|
|
|
|
spin_lock_irq(&sysfs_open_dirent_lock);
|
|
|
|
od = sd->s_attr.open;
|
|
|
|
if (od)
|
|
|
|
atomic_inc(&od->refcnt);
|
|
|
|
spin_unlock_irq(&sysfs_open_dirent_lock);
|
|
|
|
if (!od)
|
|
|
|
return;
|
|
|
|
|
|
|
|
mutex_lock(&sysfs_open_file_mutex);
|
|
|
|
list_for_each_entry(of, &od->files, list) {
|
|
|
|
struct inode *inode = file_inode(of->file);
|
|
|
|
unmap_mapping_range(inode->i_mapping, 0, 0, 1);
|
|
|
|
}
|
|
|
|
mutex_unlock(&sysfs_open_file_mutex);
|
|
|
|
|
|
|
|
sysfs_put_open_dirent(sd, NULL);
|
|
|
|
}
|
|
|
|
|
2006-03-20 14:53:53 +08:00
|
|
|
/* Sysfs attribute files are pollable. The idea is that you read
|
|
|
|
* the content and then you use 'poll' or 'select' to wait for
|
|
|
|
* the content to change. When the content changes (assuming the
|
|
|
|
* manager for the kobject supports notification), poll will
|
|
|
|
* return POLLERR|POLLPRI, and select will return the fd whether
|
|
|
|
* it is waiting for read, write, or exceptions.
|
|
|
|
* Once poll/select indicates that the value has changed, you
|
2008-04-08 06:35:01 +08:00
|
|
|
* need to close and re-open the file, or seek to 0 and read again.
|
2006-03-20 14:53:53 +08:00
|
|
|
* Reminder: this only works for attributes which actively support
|
|
|
|
* it, and it is not possible to test an attribute from userspace
|
2007-08-11 04:51:07 +08:00
|
|
|
* to see if it supports poll (Neither 'poll' nor 'select' return
|
2006-03-20 14:53:53 +08:00
|
|
|
* an appropriate error code). When in doubt, set a suitable timeout value.
|
|
|
|
*/
|
2013-11-29 03:54:19 +08:00
|
|
|
static unsigned int kernfs_file_poll(struct file *filp, poll_table *wait)
|
2006-03-20 14:53:53 +08:00
|
|
|
{
|
2013-10-02 05:42:02 +08:00
|
|
|
struct sysfs_open_file *of = sysfs_of(filp);
|
2007-06-14 02:45:16 +08:00
|
|
|
struct sysfs_dirent *attr_sd = filp->f_path.dentry->d_fsdata;
|
2007-09-20 15:05:12 +08:00
|
|
|
struct sysfs_open_dirent *od = attr_sd->s_attr.open;
|
2007-06-14 02:45:16 +08:00
|
|
|
|
|
|
|
/* need parent for the kobj, grab both */
|
2010-02-12 07:18:38 +08:00
|
|
|
if (!sysfs_get_active(attr_sd))
|
2007-06-14 02:45:16 +08:00
|
|
|
goto trigger;
|
2006-03-20 14:53:53 +08:00
|
|
|
|
2007-09-20 15:05:12 +08:00
|
|
|
poll_wait(filp, &od->poll, wait);
|
2006-03-20 14:53:53 +08:00
|
|
|
|
2010-02-12 07:18:38 +08:00
|
|
|
sysfs_put_active(attr_sd);
|
2007-06-14 02:45:16 +08:00
|
|
|
|
2013-10-02 05:41:59 +08:00
|
|
|
if (of->event != atomic_read(&od->event))
|
2007-06-14 02:45:16 +08:00
|
|
|
goto trigger;
|
2006-03-20 14:53:53 +08:00
|
|
|
|
sysfs: sysfs poll keep the poll rule of regular file.
Currently, following test programs don't finished.
% ruby -e '
Thread.new { sleep }
File.read("/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies")
'
strace expose the reason.
...
open("/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies", O_RDONLY|O_LARGEFILE) = 3
ioctl(3, SNDCTL_TMR_TIMEBASE or TCGETS, 0xbf9fa6b8) = -1 ENOTTY (Inappropriate ioctl for device)
fstat64(3, {st_mode=S_IFREG|0444, st_size=4096, ...}) = 0
_llseek(3, 0, [0], SEEK_CUR) = 0
select(4, [3], NULL, NULL, NULL) = 1 (in [3])
read(3, "1400000 1300000 1200000 1100000 1"..., 4096) = 62
select(4, [3], NULL, NULL, NULL
Because Ruby (the scripting language) VM assume select system-call
against regular file don't block. it because SUSv3 says "Regular files
shall always poll TRUE for reading and writing". see
http://www.opengroup.org/onlinepubs/009695399/functions/poll.html it
seems valid assumption.
But sysfs_poll() don't keep this rule although sysfs file can read and
write always.
This patch restore proper poll behavior to sysfs.
/sys/block/md*/md/sync_action polling application and another sysfs
updating sensitive application still can use POLLERR and POLLPRI.
Cc: Neil Brown <neilb@suse.de>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2009-04-09 12:53:22 +08:00
|
|
|
return DEFAULT_POLLMASK;
|
2007-06-14 02:45:16 +08:00
|
|
|
|
|
|
|
trigger:
|
sysfs: sysfs poll keep the poll rule of regular file.
Currently, following test programs don't finished.
% ruby -e '
Thread.new { sleep }
File.read("/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies")
'
strace expose the reason.
...
open("/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies", O_RDONLY|O_LARGEFILE) = 3
ioctl(3, SNDCTL_TMR_TIMEBASE or TCGETS, 0xbf9fa6b8) = -1 ENOTTY (Inappropriate ioctl for device)
fstat64(3, {st_mode=S_IFREG|0444, st_size=4096, ...}) = 0
_llseek(3, 0, [0], SEEK_CUR) = 0
select(4, [3], NULL, NULL, NULL) = 1 (in [3])
read(3, "1400000 1300000 1200000 1100000 1"..., 4096) = 62
select(4, [3], NULL, NULL, NULL
Because Ruby (the scripting language) VM assume select system-call
against regular file don't block. it because SUSv3 says "Regular files
shall always poll TRUE for reading and writing". see
http://www.opengroup.org/onlinepubs/009695399/functions/poll.html it
seems valid assumption.
But sysfs_poll() don't keep this rule although sysfs file can read and
write always.
This patch restore proper poll behavior to sysfs.
/sys/block/md*/md/sync_action polling application and another sysfs
updating sensitive application still can use POLLERR and POLLPRI.
Cc: Neil Brown <neilb@suse.de>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2009-04-09 12:53:22 +08:00
|
|
|
return DEFAULT_POLLMASK|POLLERR|POLLPRI;
|
2006-03-20 14:53:53 +08:00
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:27 +08:00
|
|
|
/**
|
|
|
|
* kernfs_notify - notify a kernfs file
|
|
|
|
* @sd: file to notify
|
|
|
|
*
|
|
|
|
* Notify @sd such that poll(2) on @sd wakes up.
|
|
|
|
*/
|
|
|
|
void kernfs_notify(struct sysfs_dirent *sd)
|
2008-07-16 06:58:04 +08:00
|
|
|
{
|
|
|
|
struct sysfs_open_dirent *od;
|
2009-09-16 07:05:51 +08:00
|
|
|
unsigned long flags;
|
2008-07-16 06:58:04 +08:00
|
|
|
|
2009-09-16 07:05:51 +08:00
|
|
|
spin_lock_irqsave(&sysfs_open_dirent_lock, flags);
|
2008-07-16 06:58:04 +08:00
|
|
|
|
2013-06-07 22:45:13 +08:00
|
|
|
if (!WARN_ON(sysfs_type(sd) != SYSFS_KOBJ_ATTR)) {
|
|
|
|
od = sd->s_attr.open;
|
|
|
|
if (od) {
|
|
|
|
atomic_inc(&od->event);
|
|
|
|
wake_up_interruptible(&od->poll);
|
|
|
|
}
|
2008-07-16 06:58:04 +08:00
|
|
|
}
|
|
|
|
|
2009-09-16 07:05:51 +08:00
|
|
|
spin_unlock_irqrestore(&sysfs_open_dirent_lock, flags);
|
2008-07-16 06:58:04 +08:00
|
|
|
}
|
2013-11-29 03:54:27 +08:00
|
|
|
EXPORT_SYMBOL_GPL(kernfs_notify);
|
2008-07-16 06:58:04 +08:00
|
|
|
|
2008-09-26 07:45:13 +08:00
|
|
|
void sysfs_notify(struct kobject *k, const char *dir, const char *attr)
|
2006-03-20 14:53:53 +08:00
|
|
|
{
|
2013-11-29 03:54:27 +08:00
|
|
|
struct sysfs_dirent *sd = k->sd, *tmp;
|
2007-06-14 03:27:25 +08:00
|
|
|
|
|
|
|
if (sd && dir)
|
2013-11-29 03:54:30 +08:00
|
|
|
sd = kernfs_find_and_get(sd, dir);
|
2013-11-29 03:54:27 +08:00
|
|
|
else
|
2013-11-29 03:54:30 +08:00
|
|
|
kernfs_get(sd);
|
2013-11-29 03:54:27 +08:00
|
|
|
|
|
|
|
if (sd && attr) {
|
2013-11-29 03:54:30 +08:00
|
|
|
tmp = kernfs_find_and_get(sd, attr);
|
|
|
|
kernfs_put(sd);
|
2013-11-29 03:54:27 +08:00
|
|
|
sd = tmp;
|
|
|
|
}
|
2007-06-14 03:27:25 +08:00
|
|
|
|
2013-11-29 03:54:27 +08:00
|
|
|
if (sd) {
|
|
|
|
kernfs_notify(sd);
|
2013-11-29 03:54:30 +08:00
|
|
|
kernfs_put(sd);
|
2013-11-29 03:54:27 +08:00
|
|
|
}
|
2006-03-20 14:53:53 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(sysfs_notify);
|
|
|
|
|
2013-11-29 03:54:19 +08:00
|
|
|
const struct file_operations kernfs_file_operations = {
|
2013-11-29 03:54:16 +08:00
|
|
|
.read = kernfs_file_read,
|
2013-11-29 03:54:17 +08:00
|
|
|
.write = kernfs_file_write,
|
2013-11-02 01:16:53 +08:00
|
|
|
.llseek = generic_file_llseek,
|
2013-11-29 03:54:18 +08:00
|
|
|
.mmap = kernfs_file_mmap,
|
2013-11-29 03:54:19 +08:00
|
|
|
.open = kernfs_file_open,
|
|
|
|
.release = kernfs_file_release,
|
|
|
|
.poll = kernfs_file_poll,
|
2013-10-02 05:42:05 +08:00
|
|
|
};
|
|
|
|
|
2013-11-29 03:54:21 +08:00
|
|
|
static const struct kernfs_ops sysfs_file_kfops_empty = {
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct kernfs_ops sysfs_file_kfops_ro = {
|
|
|
|
.seq_show = sysfs_kf_seq_show,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct kernfs_ops sysfs_file_kfops_wo = {
|
|
|
|
.write = sysfs_kf_write,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct kernfs_ops sysfs_file_kfops_rw = {
|
|
|
|
.seq_show = sysfs_kf_seq_show,
|
|
|
|
.write = sysfs_kf_write,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct kernfs_ops sysfs_bin_kfops_ro = {
|
|
|
|
.read = sysfs_kf_bin_read,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct kernfs_ops sysfs_bin_kfops_wo = {
|
|
|
|
.write = sysfs_kf_bin_write,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct kernfs_ops sysfs_bin_kfops_rw = {
|
|
|
|
.read = sysfs_kf_bin_read,
|
|
|
|
.write = sysfs_kf_bin_write,
|
|
|
|
.mmap = sysfs_kf_bin_mmap,
|
|
|
|
};
|
|
|
|
|
2013-09-12 10:29:04 +08:00
|
|
|
int sysfs_add_file_mode_ns(struct sysfs_dirent *dir_sd,
|
2013-11-29 03:54:23 +08:00
|
|
|
const struct attribute *attr, bool is_bin,
|
2013-11-29 03:54:24 +08:00
|
|
|
umode_t mode, const void *ns)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2013-11-29 03:54:29 +08:00
|
|
|
struct lock_class_key *key = NULL;
|
2013-11-29 03:54:21 +08:00
|
|
|
const struct kernfs_ops *ops;
|
2007-06-14 02:45:14 +08:00
|
|
|
struct sysfs_dirent *sd;
|
2013-11-29 03:54:22 +08:00
|
|
|
loff_t size;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-11-29 03:54:23 +08:00
|
|
|
if (!is_bin) {
|
2013-11-29 03:54:21 +08:00
|
|
|
struct kobject *kobj = dir_sd->priv;
|
|
|
|
const struct sysfs_ops *sysfs_ops = kobj->ktype->sysfs_ops;
|
|
|
|
|
|
|
|
/* every kobject with an attribute needs a ktype assigned */
|
|
|
|
if (WARN(!sysfs_ops, KERN_ERR
|
|
|
|
"missing sysfs attribute operations for kobject: %s\n",
|
|
|
|
kobject_name(kobj)))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (sysfs_ops->show && sysfs_ops->store)
|
|
|
|
ops = &sysfs_file_kfops_rw;
|
|
|
|
else if (sysfs_ops->show)
|
|
|
|
ops = &sysfs_file_kfops_ro;
|
|
|
|
else if (sysfs_ops->store)
|
|
|
|
ops = &sysfs_file_kfops_wo;
|
|
|
|
else
|
|
|
|
ops = &sysfs_file_kfops_empty;
|
2013-11-29 03:54:22 +08:00
|
|
|
|
|
|
|
size = PAGE_SIZE;
|
2013-11-29 03:54:21 +08:00
|
|
|
} else {
|
|
|
|
struct bin_attribute *battr = (void *)attr;
|
|
|
|
|
|
|
|
if ((battr->read && battr->write) || battr->mmap)
|
|
|
|
ops = &sysfs_bin_kfops_rw;
|
|
|
|
else if (battr->read)
|
|
|
|
ops = &sysfs_bin_kfops_ro;
|
|
|
|
else if (battr->write)
|
|
|
|
ops = &sysfs_bin_kfops_wo;
|
|
|
|
else
|
|
|
|
ops = &sysfs_file_kfops_empty;
|
2013-11-29 03:54:22 +08:00
|
|
|
|
|
|
|
size = battr->size;
|
2013-11-29 03:54:21 +08:00
|
|
|
}
|
|
|
|
|
2013-11-29 03:54:29 +08:00
|
|
|
#ifdef CONFIG_DEBUG_LOCK_ALLOC
|
|
|
|
if (!attr->ignore_lockdep)
|
|
|
|
key = attr->key ?: (struct lock_class_key *)&attr->skey;
|
|
|
|
#endif
|
|
|
|
sd = kernfs_create_file_ns_key(dir_sd, attr->name, mode, size,
|
|
|
|
ops, (void *)attr, ns, key);
|
2013-11-29 03:54:24 +08:00
|
|
|
if (IS_ERR(sd)) {
|
|
|
|
if (PTR_ERR(sd) == -EEXIST)
|
|
|
|
sysfs_warn_dup(dir_sd, attr->name);
|
|
|
|
return PTR_ERR(sd);
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2013-11-29 03:54:29 +08:00
|
|
|
* kernfs_create_file_ns_key - create a file
|
2013-11-29 03:54:24 +08:00
|
|
|
* @parent: directory to create the file in
|
|
|
|
* @name: name of the file
|
|
|
|
* @mode: mode of the file
|
|
|
|
* @size: size of the file
|
|
|
|
* @ops: kernfs operations for the file
|
|
|
|
* @priv: private data for the file
|
|
|
|
* @ns: optional namespace tag of the file
|
2013-11-29 03:54:29 +08:00
|
|
|
* @key: lockdep key for the file's active_ref, %NULL to disable lockdep
|
2013-11-29 03:54:24 +08:00
|
|
|
*
|
|
|
|
* Returns the created node on success, ERR_PTR() value on error.
|
|
|
|
*/
|
2013-11-29 03:54:29 +08:00
|
|
|
struct sysfs_dirent *kernfs_create_file_ns_key(struct sysfs_dirent *parent,
|
|
|
|
const char *name,
|
|
|
|
umode_t mode, loff_t size,
|
|
|
|
const struct kernfs_ops *ops,
|
|
|
|
void *priv, const void *ns,
|
|
|
|
struct lock_class_key *key)
|
2013-11-29 03:54:24 +08:00
|
|
|
{
|
|
|
|
struct sysfs_addrm_cxt acxt;
|
|
|
|
struct sysfs_dirent *sd;
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
sd = sysfs_new_dirent(name, (mode & S_IALLUGO) | S_IFREG,
|
|
|
|
SYSFS_KOBJ_ATTR);
|
2007-06-14 03:27:23 +08:00
|
|
|
if (!sd)
|
2013-11-29 03:54:24 +08:00
|
|
|
return ERR_PTR(-ENOMEM);
|
2011-10-13 05:53:38 +08:00
|
|
|
|
2013-11-29 03:54:21 +08:00
|
|
|
sd->s_attr.ops = ops;
|
2013-11-29 03:54:22 +08:00
|
|
|
sd->s_attr.size = size;
|
2011-10-13 05:53:38 +08:00
|
|
|
sd->s_ns = ns;
|
2013-11-29 03:54:24 +08:00
|
|
|
sd->priv = priv;
|
2013-11-29 03:54:29 +08:00
|
|
|
|
|
|
|
#ifdef CONFIG_DEBUG_LOCK_ALLOC
|
|
|
|
if (key) {
|
|
|
|
lockdep_init_map(&sd->dep_map, "s_active", key, 0);
|
|
|
|
sd->s_flags |= SYSFS_FLAG_LOCKDEP;
|
|
|
|
}
|
|
|
|
#endif
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-11-29 03:54:21 +08:00
|
|
|
/*
|
|
|
|
* sd->s_attr.ops is accesible only while holding active ref. We
|
|
|
|
* need to know whether some ops are implemented outside active
|
|
|
|
* ref. Cache their existence in flags.
|
|
|
|
*/
|
|
|
|
if (ops->seq_show)
|
|
|
|
sd->s_flags |= SYSFS_FLAG_HAS_SEQ_SHOW;
|
|
|
|
if (ops->mmap)
|
|
|
|
sd->s_flags |= SYSFS_FLAG_HAS_MMAP;
|
|
|
|
|
2013-09-19 05:15:35 +08:00
|
|
|
sysfs_addrm_start(&acxt);
|
2013-11-29 03:54:25 +08:00
|
|
|
rc = sysfs_add_one(&acxt, sd, parent);
|
2007-08-02 20:38:03 +08:00
|
|
|
sysfs_addrm_finish(&acxt);
|
2007-06-14 02:45:14 +08:00
|
|
|
|
2013-11-29 03:54:24 +08:00
|
|
|
if (rc) {
|
2013-11-29 03:54:30 +08:00
|
|
|
kernfs_put(sd);
|
2013-11-29 03:54:24 +08:00
|
|
|
return ERR_PTR(rc);
|
|
|
|
}
|
|
|
|
return sd;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2008-03-21 09:47:52 +08:00
|
|
|
int sysfs_add_file(struct sysfs_dirent *dir_sd, const struct attribute *attr,
|
2013-11-29 03:54:23 +08:00
|
|
|
bool is_bin)
|
2008-03-21 09:47:52 +08:00
|
|
|
{
|
2013-11-29 03:54:23 +08:00
|
|
|
return sysfs_add_file_mode_ns(dir_sd, attr, is_bin, attr->mode, NULL);
|
2008-03-21 09:47:52 +08:00
|
|
|
}
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/**
|
2013-09-12 10:29:04 +08:00
|
|
|
* sysfs_create_file_ns - create an attribute file for an object with custom ns
|
|
|
|
* @kobj: object we're creating for
|
|
|
|
* @attr: attribute descriptor
|
|
|
|
* @ns: namespace the new file should belong to
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
2013-09-12 10:29:04 +08:00
|
|
|
int sysfs_create_file_ns(struct kobject *kobj, const struct attribute *attr,
|
|
|
|
const void *ns)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2007-06-14 03:27:22 +08:00
|
|
|
BUG_ON(!kobj || !kobj->sd || !attr);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-11-29 03:54:23 +08:00
|
|
|
return sysfs_add_file_mode_ns(kobj->sd, attr, false, attr->mode, ns);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
}
|
2013-09-12 10:29:04 +08:00
|
|
|
EXPORT_SYMBOL_GPL(sysfs_create_file_ns);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2010-01-05 19:48:01 +08:00
|
|
|
int sysfs_create_files(struct kobject *kobj, const struct attribute **ptr)
|
|
|
|
{
|
|
|
|
int err = 0;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; ptr[i] && !err; i++)
|
|
|
|
err = sysfs_create_file(kobj, ptr[i]);
|
|
|
|
if (err)
|
|
|
|
while (--i >= 0)
|
|
|
|
sysfs_remove_file(kobj, ptr[i]);
|
|
|
|
return err;
|
|
|
|
}
|
2013-08-22 07:17:47 +08:00
|
|
|
EXPORT_SYMBOL_GPL(sysfs_create_files);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2007-02-21 04:02:44 +08:00
|
|
|
/**
|
|
|
|
* sysfs_add_file_to_group - add an attribute file to a pre-existing group.
|
|
|
|
* @kobj: object we're acting for.
|
|
|
|
* @attr: attribute descriptor.
|
|
|
|
* @group: group name.
|
|
|
|
*/
|
|
|
|
int sysfs_add_file_to_group(struct kobject *kobj,
|
|
|
|
const struct attribute *attr, const char *group)
|
|
|
|
{
|
2007-06-14 03:27:22 +08:00
|
|
|
struct sysfs_dirent *dir_sd;
|
2007-02-21 04:02:44 +08:00
|
|
|
int error;
|
|
|
|
|
2013-11-29 03:54:30 +08:00
|
|
|
if (group) {
|
|
|
|
dir_sd = kernfs_find_and_get(kobj->sd, group);
|
|
|
|
} else {
|
|
|
|
dir_sd = kobj->sd;
|
|
|
|
kernfs_get(dir_sd);
|
|
|
|
}
|
2008-01-03 08:44:05 +08:00
|
|
|
|
2007-06-14 03:27:22 +08:00
|
|
|
if (!dir_sd)
|
|
|
|
return -ENOENT;
|
|
|
|
|
2013-11-29 03:54:23 +08:00
|
|
|
error = sysfs_add_file(dir_sd, attr, false);
|
2013-11-29 03:54:30 +08:00
|
|
|
kernfs_put(dir_sd);
|
2007-06-14 03:27:22 +08:00
|
|
|
|
2007-02-21 04:02:44 +08:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(sysfs_add_file_to_group);
|
|
|
|
|
2005-04-19 12:57:32 +08:00
|
|
|
/**
|
|
|
|
* sysfs_chmod_file - update the modified mode value on an object attribute.
|
|
|
|
* @kobj: object we're acting for.
|
|
|
|
* @attr: attribute descriptor.
|
|
|
|
* @mode: file permissions.
|
|
|
|
*
|
|
|
|
*/
|
2010-07-02 22:54:05 +08:00
|
|
|
int sysfs_chmod_file(struct kobject *kobj, const struct attribute *attr,
|
2011-07-24 15:40:40 +08:00
|
|
|
umode_t mode)
|
2005-04-19 12:57:32 +08:00
|
|
|
{
|
2009-11-21 08:08:54 +08:00
|
|
|
struct sysfs_dirent *sd;
|
2005-07-30 03:13:35 +08:00
|
|
|
struct iattr newattrs;
|
2007-06-14 03:27:25 +08:00
|
|
|
int rc;
|
|
|
|
|
2013-11-29 03:54:30 +08:00
|
|
|
sd = kernfs_find_and_get(kobj->sd, attr->name);
|
2009-11-21 08:08:54 +08:00
|
|
|
if (!sd)
|
2013-11-24 06:21:52 +08:00
|
|
|
return -ENOENT;
|
2007-09-20 15:05:10 +08:00
|
|
|
|
2009-11-21 08:08:54 +08:00
|
|
|
newattrs.ia_mode = (mode & S_IALLUGO) | (sd->s_mode & ~S_IALLUGO);
|
2009-11-08 15:27:02 +08:00
|
|
|
newattrs.ia_valid = ATTR_MODE;
|
2007-09-20 15:05:10 +08:00
|
|
|
|
2013-11-24 06:21:52 +08:00
|
|
|
rc = kernfs_setattr(sd, &newattrs);
|
|
|
|
|
2013-11-29 03:54:30 +08:00
|
|
|
kernfs_put(sd);
|
2007-06-14 03:27:25 +08:00
|
|
|
return rc;
|
2005-04-19 12:57:32 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(sysfs_chmod_file);
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/**
|
2013-09-12 10:29:04 +08:00
|
|
|
* sysfs_remove_file_ns - remove an object attribute with a custom ns tag
|
|
|
|
* @kobj: object we're acting for
|
|
|
|
* @attr: attribute descriptor
|
|
|
|
* @ns: namespace tag of the file to remove
|
2005-04-17 06:20:36 +08:00
|
|
|
*
|
2013-09-12 10:29:04 +08:00
|
|
|
* Hash the attribute name and namespace tag and kill the victim.
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
2013-09-12 10:29:04 +08:00
|
|
|
void sysfs_remove_file_ns(struct kobject *kobj, const struct attribute *attr,
|
|
|
|
const void *ns)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2013-09-12 10:29:04 +08:00
|
|
|
struct sysfs_dirent *dir_sd = kobj->sd;
|
2011-10-13 05:53:38 +08:00
|
|
|
|
2013-11-24 06:21:49 +08:00
|
|
|
kernfs_remove_by_name_ns(dir_sd, attr->name, ns);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2013-09-12 10:29:04 +08:00
|
|
|
EXPORT_SYMBOL_GPL(sysfs_remove_file_ns);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-08-22 07:28:26 +08:00
|
|
|
void sysfs_remove_files(struct kobject *kobj, const struct attribute **ptr)
|
2010-01-05 19:48:01 +08:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
for (i = 0; ptr[i]; i++)
|
|
|
|
sysfs_remove_file(kobj, ptr[i]);
|
|
|
|
}
|
2013-08-22 07:17:47 +08:00
|
|
|
EXPORT_SYMBOL_GPL(sysfs_remove_files);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2007-02-21 04:02:44 +08:00
|
|
|
/**
|
|
|
|
* sysfs_remove_file_from_group - remove an attribute file from a group.
|
|
|
|
* @kobj: object we're acting for.
|
|
|
|
* @attr: attribute descriptor.
|
|
|
|
* @group: group name.
|
|
|
|
*/
|
|
|
|
void sysfs_remove_file_from_group(struct kobject *kobj,
|
|
|
|
const struct attribute *attr, const char *group)
|
|
|
|
{
|
2007-06-14 03:27:22 +08:00
|
|
|
struct sysfs_dirent *dir_sd;
|
2007-02-21 04:02:44 +08:00
|
|
|
|
2013-11-29 03:54:30 +08:00
|
|
|
if (group) {
|
|
|
|
dir_sd = kernfs_find_and_get(kobj->sd, group);
|
|
|
|
} else {
|
|
|
|
dir_sd = kobj->sd;
|
|
|
|
kernfs_get(dir_sd);
|
|
|
|
}
|
|
|
|
|
2007-06-14 03:27:22 +08:00
|
|
|
if (dir_sd) {
|
2013-11-24 06:21:49 +08:00
|
|
|
kernfs_remove_by_name(dir_sd, attr->name);
|
2013-11-29 03:54:30 +08:00
|
|
|
kernfs_put(dir_sd);
|
2007-02-21 04:02:44 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(sysfs_remove_file_from_group);
|
|
|
|
|
2013-10-02 05:42:09 +08:00
|
|
|
/**
|
|
|
|
* sysfs_create_bin_file - create binary file for object.
|
|
|
|
* @kobj: object.
|
|
|
|
* @attr: attribute descriptor.
|
|
|
|
*/
|
|
|
|
int sysfs_create_bin_file(struct kobject *kobj,
|
|
|
|
const struct bin_attribute *attr)
|
|
|
|
{
|
|
|
|
BUG_ON(!kobj || !kobj->sd || !attr);
|
|
|
|
|
2013-11-29 03:54:23 +08:00
|
|
|
return sysfs_add_file(kobj->sd, &attr->attr, true);
|
2013-10-02 05:42:09 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(sysfs_create_bin_file);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* sysfs_remove_bin_file - remove binary file for object.
|
|
|
|
* @kobj: object.
|
|
|
|
* @attr: attribute descriptor.
|
|
|
|
*/
|
|
|
|
void sysfs_remove_bin_file(struct kobject *kobj,
|
|
|
|
const struct bin_attribute *attr)
|
|
|
|
{
|
2013-11-24 06:21:49 +08:00
|
|
|
kernfs_remove_by_name(kobj->sd, attr->attr.name);
|
2013-10-02 05:42:09 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(sysfs_remove_bin_file);
|
|
|
|
|
2007-03-16 03:50:34 +08:00
|
|
|
struct sysfs_schedule_callback_struct {
|
2009-03-14 02:07:36 +08:00
|
|
|
struct list_head workq_list;
|
|
|
|
struct kobject *kobj;
|
2007-03-16 03:50:34 +08:00
|
|
|
void (*func)(void *);
|
|
|
|
void *data;
|
2007-04-26 15:12:04 +08:00
|
|
|
struct module *owner;
|
2007-03-16 03:50:34 +08:00
|
|
|
struct work_struct work;
|
|
|
|
};
|
|
|
|
|
sysfs: don't use global workqueue in sysfs_schedule_callback()
A sysfs attribute using sysfs_schedule_callback() to commit suicide
may end up calling device_unregister(), which will eventually call
a driver's ->remove function.
Drivers may call flush_scheduled_work() in their shutdown routines,
in which case lockdep will complain with something like the following:
=============================================
[ INFO: possible recursive locking detected ]
2.6.29-rc8-kk #1
---------------------------------------------
events/4/56 is trying to acquire lock:
(events){--..}, at: [<ffffffff80257fc0>] flush_workqueue+0x0/0xa0
but task is already holding lock:
(events){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
other info that might help us debug this:
3 locks held by events/4/56:
#0: (events){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
#1: (&ss->work){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
#2: (pci_remove_rescan_mutex){--..}, at: [<ffffffff803c10d1>] remove_callback+0x21/0x40
stack backtrace:
Pid: 56, comm: events/4 Not tainted 2.6.29-rc8-kk #1
Call Trace:
[<ffffffff8026dfcd>] validate_chain+0xb7d/0x1260
[<ffffffff8026eade>] __lock_acquire+0x42e/0xa40
[<ffffffff8026f148>] lock_acquire+0x58/0x80
[<ffffffff80257fc0>] ? flush_workqueue+0x0/0xa0
[<ffffffff8025800d>] flush_workqueue+0x4d/0xa0
[<ffffffff80257fc0>] ? flush_workqueue+0x0/0xa0
[<ffffffff80258070>] flush_scheduled_work+0x10/0x20
[<ffffffffa0144065>] e1000_remove+0x55/0xfe [e1000e]
[<ffffffff8033ee30>] ? sysfs_schedule_callback_work+0x0/0x50
[<ffffffff803bfeb2>] pci_device_remove+0x32/0x70
[<ffffffff80441da9>] __device_release_driver+0x59/0x90
[<ffffffff80441edb>] device_release_driver+0x2b/0x40
[<ffffffff804419d6>] bus_remove_device+0xa6/0x120
[<ffffffff8043e46b>] device_del+0x12b/0x190
[<ffffffff8043e4f6>] device_unregister+0x26/0x70
[<ffffffff803ba969>] pci_stop_dev+0x49/0x60
[<ffffffff803baab0>] pci_remove_bus_device+0x40/0xc0
[<ffffffff803c10d9>] remove_callback+0x29/0x40
[<ffffffff8033ee4f>] sysfs_schedule_callback_work+0x1f/0x50
[<ffffffff8025769a>] run_workqueue+0x15a/0x230
[<ffffffff80257648>] ? run_workqueue+0x108/0x230
[<ffffffff8025846f>] worker_thread+0x9f/0x100
[<ffffffff8025bce0>] ? autoremove_wake_function+0x0/0x40
[<ffffffff802583d0>] ? worker_thread+0x0/0x100
[<ffffffff8025b89d>] kthread+0x4d/0x80
[<ffffffff8020d4ba>] child_rip+0xa/0x20
[<ffffffff8020cebc>] ? restore_args+0x0/0x30
[<ffffffff8025b850>] ? kthread+0x0/0x80
[<ffffffff8020d4b0>] ? child_rip+0x0/0x20
Although we know that the device_unregister path will never acquire
a lock that a driver might try to acquire in its ->remove, in general
we should never attempt to flush a workqueue from within the same
workqueue, and lockdep rightly complains.
So as long as sysfs attributes cannot commit suicide directly and we
are stuck with this callback mechanism, put the sysfs callbacks on
their own workqueue instead of the global one.
This has the side benefit that if a suicidal sysfs attribute kicks
off a long chain of ->remove callbacks, we no longer induce a long
delay on the global queue.
This also fixes a missing module_put in the error path introduced
by sysfs-only-allow-one-scheduled-removal-callback-per-kobj.patch.
We never destroy the workqueue, but I'm not sure that's a
problem.
Reported-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Tested-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2009-03-26 05:11:36 +08:00
|
|
|
static struct workqueue_struct *sysfs_workqueue;
|
2009-03-14 02:07:36 +08:00
|
|
|
static DEFINE_MUTEX(sysfs_workq_mutex);
|
|
|
|
static LIST_HEAD(sysfs_workq);
|
2007-03-16 03:50:34 +08:00
|
|
|
static void sysfs_schedule_callback_work(struct work_struct *work)
|
|
|
|
{
|
|
|
|
struct sysfs_schedule_callback_struct *ss = container_of(work,
|
|
|
|
struct sysfs_schedule_callback_struct, work);
|
|
|
|
|
|
|
|
(ss->func)(ss->data);
|
|
|
|
kobject_put(ss->kobj);
|
2007-04-26 15:12:04 +08:00
|
|
|
module_put(ss->owner);
|
2009-03-14 02:07:36 +08:00
|
|
|
mutex_lock(&sysfs_workq_mutex);
|
|
|
|
list_del(&ss->workq_list);
|
|
|
|
mutex_unlock(&sysfs_workq_mutex);
|
2007-03-16 03:50:34 +08:00
|
|
|
kfree(ss);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* sysfs_schedule_callback - helper to schedule a callback for a kobject
|
|
|
|
* @kobj: object we're acting for.
|
|
|
|
* @func: callback function to invoke later.
|
|
|
|
* @data: argument to pass to @func.
|
2007-04-26 15:12:04 +08:00
|
|
|
* @owner: module owning the callback code
|
2007-03-16 03:50:34 +08:00
|
|
|
*
|
|
|
|
* sysfs attribute methods must not unregister themselves or their parent
|
|
|
|
* kobject (which would amount to the same thing). Attempts to do so will
|
|
|
|
* deadlock, since unregistration is mutually exclusive with driver
|
|
|
|
* callbacks.
|
|
|
|
*
|
|
|
|
* Instead methods can call this routine, which will attempt to allocate
|
|
|
|
* and schedule a workqueue request to call back @func with @data as its
|
|
|
|
* argument in the workqueue's process context. @kobj will be pinned
|
|
|
|
* until @func returns.
|
|
|
|
*
|
|
|
|
* Returns 0 if the request was submitted, -ENOMEM if storage could not
|
2009-03-14 02:07:36 +08:00
|
|
|
* be allocated, -ENODEV if a reference to @owner isn't available,
|
|
|
|
* -EAGAIN if a callback has already been scheduled for @kobj.
|
2007-03-16 03:50:34 +08:00
|
|
|
*/
|
|
|
|
int sysfs_schedule_callback(struct kobject *kobj, void (*func)(void *),
|
2007-04-26 15:12:04 +08:00
|
|
|
void *data, struct module *owner)
|
2007-03-16 03:50:34 +08:00
|
|
|
{
|
2009-03-14 02:07:36 +08:00
|
|
|
struct sysfs_schedule_callback_struct *ss, *tmp;
|
2007-03-16 03:50:34 +08:00
|
|
|
|
2007-04-26 15:12:04 +08:00
|
|
|
if (!try_module_get(owner))
|
|
|
|
return -ENODEV;
|
2009-03-14 02:07:36 +08:00
|
|
|
|
|
|
|
mutex_lock(&sysfs_workq_mutex);
|
|
|
|
list_for_each_entry_safe(ss, tmp, &sysfs_workq, workq_list)
|
|
|
|
if (ss->kobj == kobj) {
|
sysfs: don't use global workqueue in sysfs_schedule_callback()
A sysfs attribute using sysfs_schedule_callback() to commit suicide
may end up calling device_unregister(), which will eventually call
a driver's ->remove function.
Drivers may call flush_scheduled_work() in their shutdown routines,
in which case lockdep will complain with something like the following:
=============================================
[ INFO: possible recursive locking detected ]
2.6.29-rc8-kk #1
---------------------------------------------
events/4/56 is trying to acquire lock:
(events){--..}, at: [<ffffffff80257fc0>] flush_workqueue+0x0/0xa0
but task is already holding lock:
(events){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
other info that might help us debug this:
3 locks held by events/4/56:
#0: (events){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
#1: (&ss->work){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
#2: (pci_remove_rescan_mutex){--..}, at: [<ffffffff803c10d1>] remove_callback+0x21/0x40
stack backtrace:
Pid: 56, comm: events/4 Not tainted 2.6.29-rc8-kk #1
Call Trace:
[<ffffffff8026dfcd>] validate_chain+0xb7d/0x1260
[<ffffffff8026eade>] __lock_acquire+0x42e/0xa40
[<ffffffff8026f148>] lock_acquire+0x58/0x80
[<ffffffff80257fc0>] ? flush_workqueue+0x0/0xa0
[<ffffffff8025800d>] flush_workqueue+0x4d/0xa0
[<ffffffff80257fc0>] ? flush_workqueue+0x0/0xa0
[<ffffffff80258070>] flush_scheduled_work+0x10/0x20
[<ffffffffa0144065>] e1000_remove+0x55/0xfe [e1000e]
[<ffffffff8033ee30>] ? sysfs_schedule_callback_work+0x0/0x50
[<ffffffff803bfeb2>] pci_device_remove+0x32/0x70
[<ffffffff80441da9>] __device_release_driver+0x59/0x90
[<ffffffff80441edb>] device_release_driver+0x2b/0x40
[<ffffffff804419d6>] bus_remove_device+0xa6/0x120
[<ffffffff8043e46b>] device_del+0x12b/0x190
[<ffffffff8043e4f6>] device_unregister+0x26/0x70
[<ffffffff803ba969>] pci_stop_dev+0x49/0x60
[<ffffffff803baab0>] pci_remove_bus_device+0x40/0xc0
[<ffffffff803c10d9>] remove_callback+0x29/0x40
[<ffffffff8033ee4f>] sysfs_schedule_callback_work+0x1f/0x50
[<ffffffff8025769a>] run_workqueue+0x15a/0x230
[<ffffffff80257648>] ? run_workqueue+0x108/0x230
[<ffffffff8025846f>] worker_thread+0x9f/0x100
[<ffffffff8025bce0>] ? autoremove_wake_function+0x0/0x40
[<ffffffff802583d0>] ? worker_thread+0x0/0x100
[<ffffffff8025b89d>] kthread+0x4d/0x80
[<ffffffff8020d4ba>] child_rip+0xa/0x20
[<ffffffff8020cebc>] ? restore_args+0x0/0x30
[<ffffffff8025b850>] ? kthread+0x0/0x80
[<ffffffff8020d4b0>] ? child_rip+0x0/0x20
Although we know that the device_unregister path will never acquire
a lock that a driver might try to acquire in its ->remove, in general
we should never attempt to flush a workqueue from within the same
workqueue, and lockdep rightly complains.
So as long as sysfs attributes cannot commit suicide directly and we
are stuck with this callback mechanism, put the sysfs callbacks on
their own workqueue instead of the global one.
This has the side benefit that if a suicidal sysfs attribute kicks
off a long chain of ->remove callbacks, we no longer induce a long
delay on the global queue.
This also fixes a missing module_put in the error path introduced
by sysfs-only-allow-one-scheduled-removal-callback-per-kobj.patch.
We never destroy the workqueue, but I'm not sure that's a
problem.
Reported-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Tested-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2009-03-26 05:11:36 +08:00
|
|
|
module_put(owner);
|
2009-03-14 02:07:36 +08:00
|
|
|
mutex_unlock(&sysfs_workq_mutex);
|
|
|
|
return -EAGAIN;
|
|
|
|
}
|
|
|
|
mutex_unlock(&sysfs_workq_mutex);
|
|
|
|
|
sysfs: don't use global workqueue in sysfs_schedule_callback()
A sysfs attribute using sysfs_schedule_callback() to commit suicide
may end up calling device_unregister(), which will eventually call
a driver's ->remove function.
Drivers may call flush_scheduled_work() in their shutdown routines,
in which case lockdep will complain with something like the following:
=============================================
[ INFO: possible recursive locking detected ]
2.6.29-rc8-kk #1
---------------------------------------------
events/4/56 is trying to acquire lock:
(events){--..}, at: [<ffffffff80257fc0>] flush_workqueue+0x0/0xa0
but task is already holding lock:
(events){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
other info that might help us debug this:
3 locks held by events/4/56:
#0: (events){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
#1: (&ss->work){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
#2: (pci_remove_rescan_mutex){--..}, at: [<ffffffff803c10d1>] remove_callback+0x21/0x40
stack backtrace:
Pid: 56, comm: events/4 Not tainted 2.6.29-rc8-kk #1
Call Trace:
[<ffffffff8026dfcd>] validate_chain+0xb7d/0x1260
[<ffffffff8026eade>] __lock_acquire+0x42e/0xa40
[<ffffffff8026f148>] lock_acquire+0x58/0x80
[<ffffffff80257fc0>] ? flush_workqueue+0x0/0xa0
[<ffffffff8025800d>] flush_workqueue+0x4d/0xa0
[<ffffffff80257fc0>] ? flush_workqueue+0x0/0xa0
[<ffffffff80258070>] flush_scheduled_work+0x10/0x20
[<ffffffffa0144065>] e1000_remove+0x55/0xfe [e1000e]
[<ffffffff8033ee30>] ? sysfs_schedule_callback_work+0x0/0x50
[<ffffffff803bfeb2>] pci_device_remove+0x32/0x70
[<ffffffff80441da9>] __device_release_driver+0x59/0x90
[<ffffffff80441edb>] device_release_driver+0x2b/0x40
[<ffffffff804419d6>] bus_remove_device+0xa6/0x120
[<ffffffff8043e46b>] device_del+0x12b/0x190
[<ffffffff8043e4f6>] device_unregister+0x26/0x70
[<ffffffff803ba969>] pci_stop_dev+0x49/0x60
[<ffffffff803baab0>] pci_remove_bus_device+0x40/0xc0
[<ffffffff803c10d9>] remove_callback+0x29/0x40
[<ffffffff8033ee4f>] sysfs_schedule_callback_work+0x1f/0x50
[<ffffffff8025769a>] run_workqueue+0x15a/0x230
[<ffffffff80257648>] ? run_workqueue+0x108/0x230
[<ffffffff8025846f>] worker_thread+0x9f/0x100
[<ffffffff8025bce0>] ? autoremove_wake_function+0x0/0x40
[<ffffffff802583d0>] ? worker_thread+0x0/0x100
[<ffffffff8025b89d>] kthread+0x4d/0x80
[<ffffffff8020d4ba>] child_rip+0xa/0x20
[<ffffffff8020cebc>] ? restore_args+0x0/0x30
[<ffffffff8025b850>] ? kthread+0x0/0x80
[<ffffffff8020d4b0>] ? child_rip+0x0/0x20
Although we know that the device_unregister path will never acquire
a lock that a driver might try to acquire in its ->remove, in general
we should never attempt to flush a workqueue from within the same
workqueue, and lockdep rightly complains.
So as long as sysfs attributes cannot commit suicide directly and we
are stuck with this callback mechanism, put the sysfs callbacks on
their own workqueue instead of the global one.
This has the side benefit that if a suicidal sysfs attribute kicks
off a long chain of ->remove callbacks, we no longer induce a long
delay on the global queue.
This also fixes a missing module_put in the error path introduced
by sysfs-only-allow-one-scheduled-removal-callback-per-kobj.patch.
We never destroy the workqueue, but I'm not sure that's a
problem.
Reported-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Tested-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2009-03-26 05:11:36 +08:00
|
|
|
if (sysfs_workqueue == NULL) {
|
2009-05-08 03:36:53 +08:00
|
|
|
sysfs_workqueue = create_singlethread_workqueue("sysfsd");
|
sysfs: don't use global workqueue in sysfs_schedule_callback()
A sysfs attribute using sysfs_schedule_callback() to commit suicide
may end up calling device_unregister(), which will eventually call
a driver's ->remove function.
Drivers may call flush_scheduled_work() in their shutdown routines,
in which case lockdep will complain with something like the following:
=============================================
[ INFO: possible recursive locking detected ]
2.6.29-rc8-kk #1
---------------------------------------------
events/4/56 is trying to acquire lock:
(events){--..}, at: [<ffffffff80257fc0>] flush_workqueue+0x0/0xa0
but task is already holding lock:
(events){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
other info that might help us debug this:
3 locks held by events/4/56:
#0: (events){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
#1: (&ss->work){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
#2: (pci_remove_rescan_mutex){--..}, at: [<ffffffff803c10d1>] remove_callback+0x21/0x40
stack backtrace:
Pid: 56, comm: events/4 Not tainted 2.6.29-rc8-kk #1
Call Trace:
[<ffffffff8026dfcd>] validate_chain+0xb7d/0x1260
[<ffffffff8026eade>] __lock_acquire+0x42e/0xa40
[<ffffffff8026f148>] lock_acquire+0x58/0x80
[<ffffffff80257fc0>] ? flush_workqueue+0x0/0xa0
[<ffffffff8025800d>] flush_workqueue+0x4d/0xa0
[<ffffffff80257fc0>] ? flush_workqueue+0x0/0xa0
[<ffffffff80258070>] flush_scheduled_work+0x10/0x20
[<ffffffffa0144065>] e1000_remove+0x55/0xfe [e1000e]
[<ffffffff8033ee30>] ? sysfs_schedule_callback_work+0x0/0x50
[<ffffffff803bfeb2>] pci_device_remove+0x32/0x70
[<ffffffff80441da9>] __device_release_driver+0x59/0x90
[<ffffffff80441edb>] device_release_driver+0x2b/0x40
[<ffffffff804419d6>] bus_remove_device+0xa6/0x120
[<ffffffff8043e46b>] device_del+0x12b/0x190
[<ffffffff8043e4f6>] device_unregister+0x26/0x70
[<ffffffff803ba969>] pci_stop_dev+0x49/0x60
[<ffffffff803baab0>] pci_remove_bus_device+0x40/0xc0
[<ffffffff803c10d9>] remove_callback+0x29/0x40
[<ffffffff8033ee4f>] sysfs_schedule_callback_work+0x1f/0x50
[<ffffffff8025769a>] run_workqueue+0x15a/0x230
[<ffffffff80257648>] ? run_workqueue+0x108/0x230
[<ffffffff8025846f>] worker_thread+0x9f/0x100
[<ffffffff8025bce0>] ? autoremove_wake_function+0x0/0x40
[<ffffffff802583d0>] ? worker_thread+0x0/0x100
[<ffffffff8025b89d>] kthread+0x4d/0x80
[<ffffffff8020d4ba>] child_rip+0xa/0x20
[<ffffffff8020cebc>] ? restore_args+0x0/0x30
[<ffffffff8025b850>] ? kthread+0x0/0x80
[<ffffffff8020d4b0>] ? child_rip+0x0/0x20
Although we know that the device_unregister path will never acquire
a lock that a driver might try to acquire in its ->remove, in general
we should never attempt to flush a workqueue from within the same
workqueue, and lockdep rightly complains.
So as long as sysfs attributes cannot commit suicide directly and we
are stuck with this callback mechanism, put the sysfs callbacks on
their own workqueue instead of the global one.
This has the side benefit that if a suicidal sysfs attribute kicks
off a long chain of ->remove callbacks, we no longer induce a long
delay on the global queue.
This also fixes a missing module_put in the error path introduced
by sysfs-only-allow-one-scheduled-removal-callback-per-kobj.patch.
We never destroy the workqueue, but I'm not sure that's a
problem.
Reported-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Tested-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2009-03-26 05:11:36 +08:00
|
|
|
if (sysfs_workqueue == NULL) {
|
|
|
|
module_put(owner);
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2007-03-16 03:50:34 +08:00
|
|
|
ss = kmalloc(sizeof(*ss), GFP_KERNEL);
|
2007-04-26 15:12:04 +08:00
|
|
|
if (!ss) {
|
|
|
|
module_put(owner);
|
2007-03-16 03:50:34 +08:00
|
|
|
return -ENOMEM;
|
2007-04-26 15:12:04 +08:00
|
|
|
}
|
2007-03-16 03:50:34 +08:00
|
|
|
kobject_get(kobj);
|
|
|
|
ss->kobj = kobj;
|
|
|
|
ss->func = func;
|
|
|
|
ss->data = data;
|
2007-04-26 15:12:04 +08:00
|
|
|
ss->owner = owner;
|
2007-03-16 03:50:34 +08:00
|
|
|
INIT_WORK(&ss->work, sysfs_schedule_callback_work);
|
2009-03-14 02:07:36 +08:00
|
|
|
INIT_LIST_HEAD(&ss->workq_list);
|
|
|
|
mutex_lock(&sysfs_workq_mutex);
|
|
|
|
list_add_tail(&ss->workq_list, &sysfs_workq);
|
|
|
|
mutex_unlock(&sysfs_workq_mutex);
|
sysfs: don't use global workqueue in sysfs_schedule_callback()
A sysfs attribute using sysfs_schedule_callback() to commit suicide
may end up calling device_unregister(), which will eventually call
a driver's ->remove function.
Drivers may call flush_scheduled_work() in their shutdown routines,
in which case lockdep will complain with something like the following:
=============================================
[ INFO: possible recursive locking detected ]
2.6.29-rc8-kk #1
---------------------------------------------
events/4/56 is trying to acquire lock:
(events){--..}, at: [<ffffffff80257fc0>] flush_workqueue+0x0/0xa0
but task is already holding lock:
(events){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
other info that might help us debug this:
3 locks held by events/4/56:
#0: (events){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
#1: (&ss->work){--..}, at: [<ffffffff80257648>] run_workqueue+0x108/0x230
#2: (pci_remove_rescan_mutex){--..}, at: [<ffffffff803c10d1>] remove_callback+0x21/0x40
stack backtrace:
Pid: 56, comm: events/4 Not tainted 2.6.29-rc8-kk #1
Call Trace:
[<ffffffff8026dfcd>] validate_chain+0xb7d/0x1260
[<ffffffff8026eade>] __lock_acquire+0x42e/0xa40
[<ffffffff8026f148>] lock_acquire+0x58/0x80
[<ffffffff80257fc0>] ? flush_workqueue+0x0/0xa0
[<ffffffff8025800d>] flush_workqueue+0x4d/0xa0
[<ffffffff80257fc0>] ? flush_workqueue+0x0/0xa0
[<ffffffff80258070>] flush_scheduled_work+0x10/0x20
[<ffffffffa0144065>] e1000_remove+0x55/0xfe [e1000e]
[<ffffffff8033ee30>] ? sysfs_schedule_callback_work+0x0/0x50
[<ffffffff803bfeb2>] pci_device_remove+0x32/0x70
[<ffffffff80441da9>] __device_release_driver+0x59/0x90
[<ffffffff80441edb>] device_release_driver+0x2b/0x40
[<ffffffff804419d6>] bus_remove_device+0xa6/0x120
[<ffffffff8043e46b>] device_del+0x12b/0x190
[<ffffffff8043e4f6>] device_unregister+0x26/0x70
[<ffffffff803ba969>] pci_stop_dev+0x49/0x60
[<ffffffff803baab0>] pci_remove_bus_device+0x40/0xc0
[<ffffffff803c10d9>] remove_callback+0x29/0x40
[<ffffffff8033ee4f>] sysfs_schedule_callback_work+0x1f/0x50
[<ffffffff8025769a>] run_workqueue+0x15a/0x230
[<ffffffff80257648>] ? run_workqueue+0x108/0x230
[<ffffffff8025846f>] worker_thread+0x9f/0x100
[<ffffffff8025bce0>] ? autoremove_wake_function+0x0/0x40
[<ffffffff802583d0>] ? worker_thread+0x0/0x100
[<ffffffff8025b89d>] kthread+0x4d/0x80
[<ffffffff8020d4ba>] child_rip+0xa/0x20
[<ffffffff8020cebc>] ? restore_args+0x0/0x30
[<ffffffff8025b850>] ? kthread+0x0/0x80
[<ffffffff8020d4b0>] ? child_rip+0x0/0x20
Although we know that the device_unregister path will never acquire
a lock that a driver might try to acquire in its ->remove, in general
we should never attempt to flush a workqueue from within the same
workqueue, and lockdep rightly complains.
So as long as sysfs attributes cannot commit suicide directly and we
are stuck with this callback mechanism, put the sysfs callbacks on
their own workqueue instead of the global one.
This has the side benefit that if a suicidal sysfs attribute kicks
off a long chain of ->remove callbacks, we no longer induce a long
delay on the global queue.
This also fixes a missing module_put in the error path introduced
by sysfs-only-allow-one-scheduled-removal-callback-per-kobj.patch.
We never destroy the workqueue, but I'm not sure that's a
problem.
Reported-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Tested-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2009-03-26 05:11:36 +08:00
|
|
|
queue_work(sysfs_workqueue, &ss->work);
|
2007-03-16 03:50:34 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(sysfs_schedule_callback);
|