2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* linux/fs/proc/inode.c
|
|
|
|
*
|
|
|
|
* Copyright (C) 1991, 1992 Linus Torvalds
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/time.h>
|
|
|
|
#include <linux/proc_fs.h>
|
|
|
|
#include <linux/kernel.h>
|
2012-01-11 07:11:27 +08:00
|
|
|
#include <linux/pid_namespace.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/string.h>
|
|
|
|
#include <linux/stat.h>
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
#include <linux/completion.h>
|
2007-09-12 06:23:55 +08:00
|
|
|
#include <linux/poll.h>
|
2013-02-28 09:03:16 +08:00
|
|
|
#include <linux/printk.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <linux/file.h>
|
|
|
|
#include <linux/limits.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/module.h>
|
2008-07-15 20:54:06 +08:00
|
|
|
#include <linux/sysctl.h>
|
2012-01-11 07:11:27 +08:00
|
|
|
#include <linux/seq_file.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
|
|
|
#include <linux/slab.h>
|
2012-01-11 07:11:27 +08:00
|
|
|
#include <linux/mount.h>
|
2013-04-12 06:55:54 +08:00
|
|
|
#include <linux/magic.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
#include <asm/uaccess.h>
|
|
|
|
|
2006-01-08 17:04:16 +08:00
|
|
|
#include "internal.h"
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2010-06-05 10:17:56 +08:00
|
|
|
static void proc_evict_inode(struct inode *inode)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
struct proc_dir_entry *de;
|
2011-03-08 14:25:28 +08:00
|
|
|
struct ctl_table_header *head;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2014-04-04 05:47:49 +08:00
|
|
|
truncate_inode_pages_final(&inode->i_data);
|
2012-05-03 20:48:02 +08:00
|
|
|
clear_inode(inode);
|
2005-09-10 04:01:31 +08:00
|
|
|
|
2006-06-26 15:25:55 +08:00
|
|
|
/* Stop tracking associated processes */
|
2006-06-26 15:25:56 +08:00
|
|
|
put_pid(PROC_I(inode)->pid);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/* Let go of any associated proc directory entry */
|
2015-02-13 07:01:03 +08:00
|
|
|
de = PDE(inode);
|
2009-03-26 03:48:06 +08:00
|
|
|
if (de)
|
2009-12-16 08:45:39 +08:00
|
|
|
pde_put(de);
|
2011-03-08 14:25:28 +08:00
|
|
|
head = PROC_I(inode)->sysctl;
|
|
|
|
if (head) {
|
2014-04-08 06:38:35 +08:00
|
|
|
RCU_INIT_POINTER(PROC_I(inode)->sysctl, NULL);
|
2011-03-08 14:25:28 +08:00
|
|
|
sysctl_head_put(head);
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2006-12-07 12:33:20 +08:00
|
|
|
static struct kmem_cache * proc_inode_cachep;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
static struct inode *proc_alloc_inode(struct super_block *sb)
|
|
|
|
{
|
|
|
|
struct proc_inode *ei;
|
|
|
|
struct inode *inode;
|
|
|
|
|
2006-12-07 12:33:17 +08:00
|
|
|
ei = (struct proc_inode *)kmem_cache_alloc(proc_inode_cachep, GFP_KERNEL);
|
2005-04-17 06:20:36 +08:00
|
|
|
if (!ei)
|
|
|
|
return NULL;
|
2006-06-26 15:25:56 +08:00
|
|
|
ei->pid = NULL;
|
2006-06-26 15:25:44 +08:00
|
|
|
ei->fd = 0;
|
2005-04-17 06:20:36 +08:00
|
|
|
ei->op.proc_get_link = NULL;
|
|
|
|
ei->pde = NULL;
|
2008-07-15 20:54:06 +08:00
|
|
|
ei->sysctl = NULL;
|
|
|
|
ei->sysctl_entry = NULL;
|
2014-11-01 23:10:28 +08:00
|
|
|
ei->ns_ops = NULL;
|
2005-04-17 06:20:36 +08:00
|
|
|
inode = &ei->vfs_inode;
|
|
|
|
return inode;
|
|
|
|
}
|
|
|
|
|
2011-01-07 14:49:49 +08:00
|
|
|
static void proc_i_callback(struct rcu_head *head)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2011-01-07 14:49:49 +08:00
|
|
|
struct inode *inode = container_of(head, struct inode, i_rcu);
|
2005-04-17 06:20:36 +08:00
|
|
|
kmem_cache_free(proc_inode_cachep, PROC_I(inode));
|
|
|
|
}
|
|
|
|
|
2011-01-07 14:49:49 +08:00
|
|
|
static void proc_destroy_inode(struct inode *inode)
|
|
|
|
{
|
|
|
|
call_rcu(&inode->i_rcu, proc_i_callback);
|
|
|
|
}
|
|
|
|
|
2008-07-26 10:45:34 +08:00
|
|
|
static void init_once(void *foo)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
struct proc_inode *ei = (struct proc_inode *) foo;
|
|
|
|
|
2007-05-17 13:10:57 +08:00
|
|
|
inode_init_once(&ei->vfs_inode);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2007-07-20 09:11:58 +08:00
|
|
|
|
2008-10-17 07:43:55 +08:00
|
|
|
void __init proc_init_inodecache(void)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
proc_inode_cachep = kmem_cache_create("proc_inode_cache",
|
|
|
|
sizeof(struct proc_inode),
|
2006-03-24 19:16:06 +08:00
|
|
|
0, (SLAB_RECLAIM_ACCOUNT|
|
2016-01-15 07:18:21 +08:00
|
|
|
SLAB_MEM_SPREAD|SLAB_ACCOUNT|
|
|
|
|
SLAB_PANIC),
|
2007-07-20 09:11:58 +08:00
|
|
|
init_once);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2012-01-11 07:11:27 +08:00
|
|
|
static int proc_show_options(struct seq_file *seq, struct dentry *root)
|
|
|
|
{
|
procfs: add hidepid= and gid= mount options
Add support for mount options to restrict access to /proc/PID/
directories. The default backward-compatible "relaxed" behaviour is left
untouched.
The first mount option is called "hidepid" and its value defines how much
info about processes we want to be available for non-owners:
hidepid=0 (default) means the old behavior - anybody may read all
world-readable /proc/PID/* files.
hidepid=1 means users may not access any /proc/<pid>/ directories, but
their own. Sensitive files like cmdline, sched*, status are now protected
against other users. As permission checking done in proc_pid_permission()
and files' permissions are left untouched, programs expecting specific
files' modes are not confused.
hidepid=2 means hidepid=1 plus all /proc/PID/ will be invisible to other
users. It doesn't mean that it hides whether a process exists (it can be
learned by other means, e.g. by kill -0 $PID), but it hides process' euid
and egid. It compicates intruder's task of gathering info about running
processes, whether some daemon runs with elevated privileges, whether
another user runs some sensitive program, whether other users run any
program at all, etc.
gid=XXX defines a group that will be able to gather all processes' info
(as in hidepid=0 mode). This group should be used instead of putting
nonroot user in sudoers file or something. However, untrusted users (like
daemons, etc.) which are not supposed to monitor the tasks in the whole
system should not be added to the group.
hidepid=1 or higher is designed to restrict access to procfs files, which
might reveal some sensitive private information like precise keystrokes
timings:
http://www.openwall.com/lists/oss-security/2011/11/05/3
hidepid=1/2 doesn't break monitoring userspace tools. ps, top, pgrep, and
conky gracefully handle EPERM/ENOENT and behave as if the current user is
the only user running processes. pstree shows the process subtree which
contains "pstree" process.
Note: the patch doesn't deal with setuid/setgid issues of keeping
preopened descriptors of procfs files (like
https://lkml.org/lkml/2011/2/7/368). We rely on that the leaked
information like the scheduling counters of setuid apps doesn't threaten
anybody's privacy - only the user started the setuid program may read the
counters.
Signed-off-by: Vasiliy Kulikov <segoon@openwall.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Randy Dunlap <rdunlap@xenotime.net>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Greg KH <greg@kroah.com>
Cc: Theodore Tso <tytso@MIT.EDU>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: James Morris <jmorris@namei.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:11:31 +08:00
|
|
|
struct super_block *sb = root->d_sb;
|
|
|
|
struct pid_namespace *pid = sb->s_fs_info;
|
|
|
|
|
2012-02-10 00:48:21 +08:00
|
|
|
if (!gid_eq(pid->pid_gid, GLOBAL_ROOT_GID))
|
|
|
|
seq_printf(seq, ",gid=%u", from_kgid_munged(&init_user_ns, pid->pid_gid));
|
procfs: add hidepid= and gid= mount options
Add support for mount options to restrict access to /proc/PID/
directories. The default backward-compatible "relaxed" behaviour is left
untouched.
The first mount option is called "hidepid" and its value defines how much
info about processes we want to be available for non-owners:
hidepid=0 (default) means the old behavior - anybody may read all
world-readable /proc/PID/* files.
hidepid=1 means users may not access any /proc/<pid>/ directories, but
their own. Sensitive files like cmdline, sched*, status are now protected
against other users. As permission checking done in proc_pid_permission()
and files' permissions are left untouched, programs expecting specific
files' modes are not confused.
hidepid=2 means hidepid=1 plus all /proc/PID/ will be invisible to other
users. It doesn't mean that it hides whether a process exists (it can be
learned by other means, e.g. by kill -0 $PID), but it hides process' euid
and egid. It compicates intruder's task of gathering info about running
processes, whether some daemon runs with elevated privileges, whether
another user runs some sensitive program, whether other users run any
program at all, etc.
gid=XXX defines a group that will be able to gather all processes' info
(as in hidepid=0 mode). This group should be used instead of putting
nonroot user in sudoers file or something. However, untrusted users (like
daemons, etc.) which are not supposed to monitor the tasks in the whole
system should not be added to the group.
hidepid=1 or higher is designed to restrict access to procfs files, which
might reveal some sensitive private information like precise keystrokes
timings:
http://www.openwall.com/lists/oss-security/2011/11/05/3
hidepid=1/2 doesn't break monitoring userspace tools. ps, top, pgrep, and
conky gracefully handle EPERM/ENOENT and behave as if the current user is
the only user running processes. pstree shows the process subtree which
contains "pstree" process.
Note: the patch doesn't deal with setuid/setgid issues of keeping
preopened descriptors of procfs files (like
https://lkml.org/lkml/2011/2/7/368). We rely on that the leaked
information like the scheduling counters of setuid apps doesn't threaten
anybody's privacy - only the user started the setuid program may read the
counters.
Signed-off-by: Vasiliy Kulikov <segoon@openwall.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Randy Dunlap <rdunlap@xenotime.net>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Greg KH <greg@kroah.com>
Cc: Theodore Tso <tytso@MIT.EDU>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: James Morris <jmorris@namei.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-11 07:11:31 +08:00
|
|
|
if (pid->hide_pid != 0)
|
|
|
|
seq_printf(seq, ",hidepid=%u", pid->hide_pid);
|
|
|
|
|
2012-01-11 07:11:27 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-02-12 16:55:41 +08:00
|
|
|
static const struct super_operations proc_sops = {
|
2005-04-17 06:20:36 +08:00
|
|
|
.alloc_inode = proc_alloc_inode,
|
|
|
|
.destroy_inode = proc_destroy_inode,
|
|
|
|
.drop_inode = generic_delete_inode,
|
2010-06-05 10:17:56 +08:00
|
|
|
.evict_inode = proc_evict_inode,
|
2005-04-17 06:20:36 +08:00
|
|
|
.statfs = simple_statfs,
|
2012-01-11 07:11:27 +08:00
|
|
|
.remount_fs = proc_remount,
|
|
|
|
.show_options = proc_show_options,
|
2005-04-17 06:20:36 +08:00
|
|
|
};
|
|
|
|
|
2013-04-04 07:07:30 +08:00
|
|
|
enum {BIAS = -1U<<31};
|
|
|
|
|
|
|
|
static inline int use_pde(struct proc_dir_entry *pde)
|
|
|
|
{
|
2013-04-05 04:28:47 +08:00
|
|
|
return atomic_inc_unless_negative(&pde->in_use);
|
2008-07-25 16:48:29 +08:00
|
|
|
}
|
|
|
|
|
2013-04-04 07:07:30 +08:00
|
|
|
static void unuse_pde(struct proc_dir_entry *pde)
|
2008-07-25 16:48:29 +08:00
|
|
|
{
|
2013-04-05 04:28:47 +08:00
|
|
|
if (atomic_dec_return(&pde->in_use) == BIAS)
|
|
|
|
complete(pde->pde_unload_completion);
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
}
|
|
|
|
|
2013-04-04 07:57:00 +08:00
|
|
|
/* pde is locked */
|
|
|
|
static void close_pdeo(struct proc_dir_entry *pde, struct pde_opener *pdeo)
|
|
|
|
{
|
2013-04-05 04:28:47 +08:00
|
|
|
if (pdeo->closing) {
|
2013-04-04 07:57:00 +08:00
|
|
|
/* somebody else is doing that, just wait */
|
2013-04-05 04:28:47 +08:00
|
|
|
DECLARE_COMPLETION_ONSTACK(c);
|
|
|
|
pdeo->c = &c;
|
2013-04-04 07:57:00 +08:00
|
|
|
spin_unlock(&pde->pde_unload_lock);
|
2013-04-05 04:28:47 +08:00
|
|
|
wait_for_completion(&c);
|
2013-04-04 07:57:00 +08:00
|
|
|
spin_lock(&pde->pde_unload_lock);
|
|
|
|
} else {
|
|
|
|
struct file *file;
|
2013-04-05 04:28:47 +08:00
|
|
|
pdeo->closing = 1;
|
2013-04-04 07:57:00 +08:00
|
|
|
spin_unlock(&pde->pde_unload_lock);
|
|
|
|
file = pdeo->file;
|
|
|
|
pde->proc_fops->release(file_inode(file), file);
|
|
|
|
spin_lock(&pde->pde_unload_lock);
|
|
|
|
list_del_init(&pdeo->lh);
|
2013-04-05 04:28:47 +08:00
|
|
|
if (pdeo->c)
|
|
|
|
complete(pdeo->c);
|
2013-04-04 07:57:00 +08:00
|
|
|
kfree(pdeo);
|
2013-04-05 04:28:47 +08:00
|
|
|
}
|
2013-04-04 07:57:00 +08:00
|
|
|
}
|
|
|
|
|
2013-04-04 07:07:30 +08:00
|
|
|
void proc_entry_rundown(struct proc_dir_entry *de)
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
{
|
2013-04-05 04:28:47 +08:00
|
|
|
DECLARE_COMPLETION_ONSTACK(c);
|
2013-04-04 07:07:30 +08:00
|
|
|
/* Wait until all existing callers into module are done. */
|
2013-04-05 04:28:47 +08:00
|
|
|
de->pde_unload_completion = &c;
|
|
|
|
if (atomic_add_return(BIAS, &de->in_use) != BIAS)
|
|
|
|
wait_for_completion(&c);
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
|
2013-04-05 04:28:47 +08:00
|
|
|
spin_lock(&de->pde_unload_lock);
|
2013-04-04 07:07:30 +08:00
|
|
|
while (!list_empty(&de->pde_openers)) {
|
|
|
|
struct pde_opener *pdeo;
|
|
|
|
pdeo = list_first_entry(&de->pde_openers, struct pde_opener, lh);
|
2013-04-04 07:57:00 +08:00
|
|
|
close_pdeo(de, pdeo);
|
2013-04-04 07:07:30 +08:00
|
|
|
}
|
|
|
|
spin_unlock(&de->pde_unload_lock);
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
}
|
|
|
|
|
2013-04-04 07:07:30 +08:00
|
|
|
static loff_t proc_reg_llseek(struct file *file, loff_t offset, int whence)
|
|
|
|
{
|
|
|
|
struct proc_dir_entry *pde = PDE(file_inode(file));
|
|
|
|
loff_t rv = -EINVAL;
|
|
|
|
if (use_pde(pde)) {
|
|
|
|
loff_t (*llseek)(struct file *, loff_t, int);
|
|
|
|
llseek = pde->proc_fops->llseek;
|
|
|
|
if (!llseek)
|
|
|
|
llseek = default_llseek;
|
|
|
|
rv = llseek(file, offset, whence);
|
|
|
|
unuse_pde(pde);
|
|
|
|
}
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
return rv;
|
|
|
|
}
|
|
|
|
|
2013-04-04 07:07:30 +08:00
|
|
|
static ssize_t proc_reg_read(struct file *file, char __user *buf, size_t count, loff_t *ppos)
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
{
|
2013-04-04 07:07:30 +08:00
|
|
|
ssize_t (*read)(struct file *, char __user *, size_t, loff_t *);
|
2013-01-24 06:07:38 +08:00
|
|
|
struct proc_dir_entry *pde = PDE(file_inode(file));
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
ssize_t rv = -EIO;
|
2013-04-04 07:07:30 +08:00
|
|
|
if (use_pde(pde)) {
|
|
|
|
read = pde->proc_fops->read;
|
|
|
|
if (read)
|
|
|
|
rv = read(file, buf, count, ppos);
|
|
|
|
unuse_pde(pde);
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
}
|
2013-04-04 07:07:30 +08:00
|
|
|
return rv;
|
|
|
|
}
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
|
2013-04-04 07:07:30 +08:00
|
|
|
static ssize_t proc_reg_write(struct file *file, const char __user *buf, size_t count, loff_t *ppos)
|
|
|
|
{
|
|
|
|
ssize_t (*write)(struct file *, const char __user *, size_t, loff_t *);
|
|
|
|
struct proc_dir_entry *pde = PDE(file_inode(file));
|
|
|
|
ssize_t rv = -EIO;
|
|
|
|
if (use_pde(pde)) {
|
|
|
|
write = pde->proc_fops->write;
|
|
|
|
if (write)
|
|
|
|
rv = write(file, buf, count, ppos);
|
|
|
|
unuse_pde(pde);
|
|
|
|
}
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
return rv;
|
|
|
|
}
|
|
|
|
|
|
|
|
static unsigned int proc_reg_poll(struct file *file, struct poll_table_struct *pts)
|
|
|
|
{
|
2013-01-24 06:07:38 +08:00
|
|
|
struct proc_dir_entry *pde = PDE(file_inode(file));
|
2007-09-12 06:23:55 +08:00
|
|
|
unsigned int rv = DEFAULT_POLLMASK;
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
unsigned int (*poll)(struct file *, struct poll_table_struct *);
|
2013-04-04 07:07:30 +08:00
|
|
|
if (use_pde(pde)) {
|
|
|
|
poll = pde->proc_fops->poll;
|
|
|
|
if (poll)
|
|
|
|
rv = poll(file, pts);
|
|
|
|
unuse_pde(pde);
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
}
|
|
|
|
return rv;
|
|
|
|
}
|
|
|
|
|
|
|
|
static long proc_reg_unlocked_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
|
|
|
|
{
|
2013-01-24 06:07:38 +08:00
|
|
|
struct proc_dir_entry *pde = PDE(file_inode(file));
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
long rv = -ENOTTY;
|
2010-07-04 06:15:10 +08:00
|
|
|
long (*ioctl)(struct file *, unsigned int, unsigned long);
|
2013-04-04 07:07:30 +08:00
|
|
|
if (use_pde(pde)) {
|
|
|
|
ioctl = pde->proc_fops->unlocked_ioctl;
|
|
|
|
if (ioctl)
|
|
|
|
rv = ioctl(file, cmd, arg);
|
|
|
|
unuse_pde(pde);
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
}
|
|
|
|
return rv;
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
static long proc_reg_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
|
|
|
|
{
|
2013-01-24 06:07:38 +08:00
|
|
|
struct proc_dir_entry *pde = PDE(file_inode(file));
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
long rv = -ENOTTY;
|
|
|
|
long (*compat_ioctl)(struct file *, unsigned int, unsigned long);
|
2013-04-04 07:07:30 +08:00
|
|
|
if (use_pde(pde)) {
|
|
|
|
compat_ioctl = pde->proc_fops->compat_ioctl;
|
|
|
|
if (compat_ioctl)
|
|
|
|
rv = compat_ioctl(file, cmd, arg);
|
|
|
|
unuse_pde(pde);
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
}
|
|
|
|
return rv;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
static int proc_reg_mmap(struct file *file, struct vm_area_struct *vma)
|
|
|
|
{
|
2013-01-24 06:07:38 +08:00
|
|
|
struct proc_dir_entry *pde = PDE(file_inode(file));
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
int rv = -EIO;
|
|
|
|
int (*mmap)(struct file *, struct vm_area_struct *);
|
2013-04-04 07:07:30 +08:00
|
|
|
if (use_pde(pde)) {
|
|
|
|
mmap = pde->proc_fops->mmap;
|
|
|
|
if (mmap)
|
|
|
|
rv = mmap(file, vma);
|
|
|
|
unuse_pde(pde);
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
}
|
|
|
|
return rv;
|
|
|
|
}
|
|
|
|
|
2013-11-13 07:11:15 +08:00
|
|
|
static unsigned long
|
|
|
|
proc_reg_get_unmapped_area(struct file *file, unsigned long orig_addr,
|
|
|
|
unsigned long len, unsigned long pgoff,
|
|
|
|
unsigned long flags)
|
2013-08-21 03:17:24 +08:00
|
|
|
{
|
|
|
|
struct proc_dir_entry *pde = PDE(file_inode(file));
|
2013-10-17 04:47:04 +08:00
|
|
|
unsigned long rv = -EIO;
|
2013-12-13 09:12:22 +08:00
|
|
|
|
2013-08-21 03:17:24 +08:00
|
|
|
if (use_pde(pde)) {
|
2013-12-13 09:12:22 +08:00
|
|
|
typeof(proc_reg_get_unmapped_area) *get_area;
|
|
|
|
|
|
|
|
get_area = pde->proc_fops->get_unmapped_area;
|
2013-10-17 04:47:05 +08:00
|
|
|
#ifdef CONFIG_MMU
|
2013-12-13 09:12:22 +08:00
|
|
|
if (!get_area)
|
|
|
|
get_area = current->mm->get_unmapped_area;
|
2013-10-17 04:47:05 +08:00
|
|
|
#endif
|
2013-12-13 09:12:22 +08:00
|
|
|
|
2013-11-13 07:11:15 +08:00
|
|
|
if (get_area)
|
|
|
|
rv = get_area(file, orig_addr, len, pgoff, flags);
|
2013-12-13 09:12:22 +08:00
|
|
|
else
|
|
|
|
rv = orig_addr;
|
2013-08-21 03:17:24 +08:00
|
|
|
unuse_pde(pde);
|
|
|
|
}
|
|
|
|
return rv;
|
|
|
|
}
|
|
|
|
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
static int proc_reg_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
struct proc_dir_entry *pde = PDE(inode);
|
|
|
|
int rv = 0;
|
|
|
|
int (*open)(struct inode *, struct file *);
|
2008-07-25 16:48:29 +08:00
|
|
|
int (*release)(struct inode *, struct file *);
|
|
|
|
struct pde_opener *pdeo;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* What for, you ask? Well, we can have open, rmmod, remove_proc_entry
|
|
|
|
* sequence. ->release won't be called because ->proc_fops will be
|
|
|
|
* cleared. Depending on complexity of ->release, consequences vary.
|
|
|
|
*
|
|
|
|
* We can't wait for mercy when close will be done for real, it's
|
|
|
|
* deadlockable: rmmod foo </proc/foo . So, we're going to do ->release
|
|
|
|
* by hand in remove_proc_entry(). For this, save opener's credentials
|
|
|
|
* for later.
|
|
|
|
*/
|
2013-04-05 04:28:47 +08:00
|
|
|
pdeo = kzalloc(sizeof(struct pde_opener), GFP_KERNEL);
|
2008-07-25 16:48:29 +08:00
|
|
|
if (!pdeo)
|
|
|
|
return -ENOMEM;
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
|
2013-04-04 07:07:30 +08:00
|
|
|
if (!use_pde(pde)) {
|
2008-07-25 16:48:29 +08:00
|
|
|
kfree(pdeo);
|
2011-07-27 07:08:37 +08:00
|
|
|
return -ENOENT;
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
}
|
|
|
|
open = pde->proc_fops->open;
|
2008-07-25 16:48:29 +08:00
|
|
|
release = pde->proc_fops->release;
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
|
|
|
|
if (open)
|
|
|
|
rv = open(inode, file);
|
|
|
|
|
2008-07-25 16:48:29 +08:00
|
|
|
if (rv == 0 && release) {
|
|
|
|
/* To know what to release. */
|
|
|
|
pdeo->file = file;
|
|
|
|
/* Strictly for "too late" ->release in proc_reg_release(). */
|
2013-04-05 04:28:47 +08:00
|
|
|
spin_lock(&pde->pde_unload_lock);
|
2008-07-25 16:48:29 +08:00
|
|
|
list_add(&pdeo->lh, &pde->pde_openers);
|
2013-04-05 04:28:47 +08:00
|
|
|
spin_unlock(&pde->pde_unload_lock);
|
2008-07-25 16:48:29 +08:00
|
|
|
} else
|
|
|
|
kfree(pdeo);
|
2013-04-05 04:28:47 +08:00
|
|
|
|
|
|
|
unuse_pde(pde);
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
return rv;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int proc_reg_release(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
struct proc_dir_entry *pde = PDE(inode);
|
2008-07-25 16:48:29 +08:00
|
|
|
struct pde_opener *pdeo;
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
spin_lock(&pde->pde_unload_lock);
|
2013-04-04 07:57:00 +08:00
|
|
|
list_for_each_entry(pdeo, &pde->pde_openers, lh) {
|
|
|
|
if (pdeo->file == file) {
|
|
|
|
close_pdeo(pde, pdeo);
|
|
|
|
break;
|
|
|
|
}
|
2008-07-25 16:48:29 +08:00
|
|
|
}
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
spin_unlock(&pde->pde_unload_lock);
|
2013-04-04 07:57:00 +08:00
|
|
|
return 0;
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static const struct file_operations proc_reg_file_ops = {
|
|
|
|
.llseek = proc_reg_llseek,
|
|
|
|
.read = proc_reg_read,
|
|
|
|
.write = proc_reg_write,
|
|
|
|
.poll = proc_reg_poll,
|
|
|
|
.unlocked_ioctl = proc_reg_unlocked_ioctl,
|
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
.compat_ioctl = proc_reg_compat_ioctl,
|
|
|
|
#endif
|
|
|
|
.mmap = proc_reg_mmap,
|
2013-08-21 03:17:24 +08:00
|
|
|
.get_unmapped_area = proc_reg_get_unmapped_area,
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
.open = proc_reg_open,
|
|
|
|
.release = proc_reg_release,
|
|
|
|
};
|
|
|
|
|
2007-07-28 13:58:37 +08:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
static const struct file_operations proc_reg_file_ops_no_compat = {
|
|
|
|
.llseek = proc_reg_llseek,
|
|
|
|
.read = proc_reg_read,
|
|
|
|
.write = proc_reg_write,
|
|
|
|
.poll = proc_reg_poll,
|
|
|
|
.unlocked_ioctl = proc_reg_unlocked_ioctl,
|
|
|
|
.mmap = proc_reg_mmap,
|
2013-08-21 03:17:24 +08:00
|
|
|
.get_unmapped_area = proc_reg_get_unmapped_area,
|
2007-07-28 13:58:37 +08:00
|
|
|
.open = proc_reg_open,
|
|
|
|
.release = proc_reg_release,
|
|
|
|
};
|
|
|
|
#endif
|
|
|
|
|
2015-12-30 04:58:39 +08:00
|
|
|
static void proc_put_link(void *p)
|
|
|
|
{
|
|
|
|
unuse_pde(p);
|
|
|
|
}
|
|
|
|
|
2015-11-17 23:20:54 +08:00
|
|
|
static const char *proc_get_link(struct dentry *dentry,
|
2015-12-30 04:58:39 +08:00
|
|
|
struct inode *inode,
|
|
|
|
struct delayed_call *done)
|
2015-02-22 11:16:11 +08:00
|
|
|
{
|
2015-11-17 23:20:54 +08:00
|
|
|
struct proc_dir_entry *pde = PDE(inode);
|
2015-02-22 11:16:11 +08:00
|
|
|
if (unlikely(!use_pde(pde)))
|
|
|
|
return ERR_PTR(-EINVAL);
|
2015-12-30 04:58:39 +08:00
|
|
|
set_delayed_call(done, proc_put_link, pde);
|
2015-05-03 01:32:22 +08:00
|
|
|
return pde->data;
|
2015-02-22 11:16:11 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
const struct inode_operations proc_link_inode_operations = {
|
|
|
|
.readlink = generic_readlink,
|
2015-11-17 23:20:54 +08:00
|
|
|
.get_link = proc_get_link,
|
2015-02-22 11:16:11 +08:00
|
|
|
};
|
|
|
|
|
2011-01-13 09:00:33 +08:00
|
|
|
struct inode *proc_get_inode(struct super_block *sb, struct proc_dir_entry *de)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
vfs,proc: guarantee unique inodes in /proc
Dave Jones found another /proc issue with his Trinity tool: thanks to
the namespace model, we can have multiple /proc dentries that point to
the same inode, aliasing directories in /proc/<pid>/net/ for example.
This ends up being a total disaster, because it acts like hardlinked
directories, and causes locking problems. We rely on the topological
sort of the inodes pointed to by dentries, and if we have aliased
directories, that odering becomes unreliable.
In short: don't do this. Multiple dentries with the same (directory)
inode is just a bad idea, and the namespace code should never have
exposed things this way. But we're kind of stuck with it.
This solves things by just always allocating a new inode during /proc
dentry lookup, instead of using "iget_locked()" to look up existing
inodes by superblock and number. That actually simplies the code a bit,
at the cost of potentially doing more inode [de]allocations.
That said, the inode lookup wasn't free either (and did a lot of locking
of inodes), so it is probably not that noticeable. We could easily keep
the old lookup model for non-directory entries, but rather than try to
be excessively clever this just implements the minimal and simplest
workaround for the problem.
Reported-and-tested-by: Dave Jones <davej@redhat.com>
Analyzed-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-03-23 02:44:04 +08:00
|
|
|
struct inode *inode = new_inode_pseudo(sb);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
vfs,proc: guarantee unique inodes in /proc
Dave Jones found another /proc issue with his Trinity tool: thanks to
the namespace model, we can have multiple /proc dentries that point to
the same inode, aliasing directories in /proc/<pid>/net/ for example.
This ends up being a total disaster, because it acts like hardlinked
directories, and causes locking problems. We rely on the topological
sort of the inodes pointed to by dentries, and if we have aliased
directories, that odering becomes unreliable.
In short: don't do this. Multiple dentries with the same (directory)
inode is just a bad idea, and the namespace code should never have
exposed things this way. But we're kind of stuck with it.
This solves things by just always allocating a new inode during /proc
dentry lookup, instead of using "iget_locked()" to look up existing
inodes by superblock and number. That actually simplies the code a bit,
at the cost of potentially doing more inode [de]allocations.
That said, the inode lookup wasn't free either (and did a lot of locking
of inodes), so it is probably not that noticeable. We could easily keep
the old lookup model for non-directory entries, but rather than try to
be excessively clever this just implements the minimal and simplest
workaround for the problem.
Reported-and-tested-by: Dave Jones <davej@redhat.com>
Analyzed-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-03-23 02:44:04 +08:00
|
|
|
if (inode) {
|
|
|
|
inode->i_ino = de->low_ino;
|
2008-02-07 16:15:45 +08:00
|
|
|
inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
|
|
|
|
PROC_I(inode)->pde = de;
|
2008-04-29 16:01:41 +08:00
|
|
|
|
2015-05-12 05:44:25 +08:00
|
|
|
if (is_empty_pde(de)) {
|
|
|
|
make_empty_dir_inode(inode);
|
|
|
|
return inode;
|
|
|
|
}
|
2008-04-29 16:01:41 +08:00
|
|
|
if (de->mode) {
|
|
|
|
inode->i_mode = de->mode;
|
|
|
|
inode->i_uid = de->uid;
|
|
|
|
inode->i_gid = de->gid;
|
|
|
|
}
|
|
|
|
if (de->size)
|
|
|
|
inode->i_size = de->size;
|
|
|
|
if (de->nlink)
|
2011-10-28 20:13:29 +08:00
|
|
|
set_nlink(inode, de->nlink);
|
2013-03-31 09:20:14 +08:00
|
|
|
WARN_ON(!de->proc_iops);
|
|
|
|
inode->i_op = de->proc_iops;
|
2008-04-29 16:01:41 +08:00
|
|
|
if (de->proc_fops) {
|
|
|
|
if (S_ISREG(inode->i_mode)) {
|
2007-07-28 13:58:37 +08:00
|
|
|
#ifdef CONFIG_COMPAT
|
2008-04-29 16:01:41 +08:00
|
|
|
if (!de->proc_fops->compat_ioctl)
|
|
|
|
inode->i_fop =
|
|
|
|
&proc_reg_file_ops_no_compat;
|
|
|
|
else
|
2007-07-28 13:58:37 +08:00
|
|
|
#endif
|
2008-04-29 16:01:41 +08:00
|
|
|
inode->i_fop = &proc_reg_file_ops;
|
|
|
|
} else {
|
|
|
|
inode->i_fop = de->proc_fops;
|
2007-07-28 13:58:37 +08:00
|
|
|
}
|
Fix rmmod/read/write races in /proc entries
Fix following races:
===========================================
1. Write via ->write_proc sleeps in copy_from_user(). Module disappears
meanwhile. Or, more generically, system call done on /proc file, method
supplied by module is called, module dissapeares meanwhile.
pde = create_proc_entry()
if (!pde)
return -ENOMEM;
pde->write_proc = ...
open
write
copy_from_user
pde = create_proc_entry();
if (!pde) {
remove_proc_entry();
return -ENOMEM;
/* module unloaded */
}
*boom*
==========================================
2. bogo-revoke aka proc_kill_inodes()
remove_proc_entry vfs_read
proc_kill_inodes [check ->f_op validness]
[check ->f_op->read validness]
[verify_area, security permissions checks]
->f_op = NULL;
if (file->f_op->read)
/* ->f_op dereference, boom */
NOTE, NOTE, NOTE: file_operations are proxied for regular files only. Let's
see how this scheme behaves, then extend if needed for directories.
Directories creators in /proc only set ->owner for them, so proxying for
directories may be unneeded.
NOTE, NOTE, NOTE: methods being proxied are ->llseek, ->read, ->write,
->poll, ->unlocked_ioctl, ->ioctl, ->compat_ioctl, ->open, ->release.
If your in-tree module uses something else, yell on me. Full audit pending.
[akpm@linux-foundation.org: build fix]
Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-16 14:39:00 +08:00
|
|
|
}
|
2009-03-26 03:48:06 +08:00
|
|
|
} else
|
2009-12-16 08:45:39 +08:00
|
|
|
pde_put(de);
|
2005-04-17 06:20:36 +08:00
|
|
|
return inode;
|
2013-01-26 09:11:22 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2016-06-10 04:32:10 +08:00
|
|
|
int proc_fill_super(struct super_block *s, void *data, int silent)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2016-06-10 04:32:10 +08:00
|
|
|
struct pid_namespace *ns = get_pid_ns(s->s_fs_info);
|
2013-02-20 10:13:32 +08:00
|
|
|
struct inode *root_inode;
|
2014-07-31 18:10:50 +08:00
|
|
|
int ret;
|
2013-02-20 10:13:32 +08:00
|
|
|
|
2016-06-10 04:32:10 +08:00
|
|
|
if (!proc_parse_options(data, ns))
|
|
|
|
return -EINVAL;
|
|
|
|
|
2016-06-10 04:34:02 +08:00
|
|
|
/* User space would break if executables or devices appear on proc */
|
|
|
|
s->s_iflags |= SB_I_USERNS_VISIBLE | SB_I_NOEXEC | SB_I_NODEV;
|
2006-07-16 03:20:05 +08:00
|
|
|
s->s_flags |= MS_NODIRATIME | MS_NOSUID | MS_NOEXEC;
|
2005-04-17 06:20:36 +08:00
|
|
|
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;
|
2016-07-30 06:54:19 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* procfs isn't actually a stacking filesystem; however, there is
|
|
|
|
* too much magic going on inside it to permit stacking things on
|
|
|
|
* top of it
|
|
|
|
*/
|
|
|
|
s->s_stack_depth = FILESYSTEM_MAX_STACK_DEPTH;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2009-12-16 08:45:39 +08:00
|
|
|
pde_get(&proc_root);
|
2013-02-20 10:13:32 +08:00
|
|
|
root_inode = proc_get_inode(s, &proc_root);
|
|
|
|
if (!root_inode) {
|
2013-02-28 09:03:16 +08:00
|
|
|
pr_err("proc_fill_super: get root inode failed\n");
|
2013-02-20 10:13:32 +08:00
|
|
|
return -ENOMEM;
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-02-20 10:13:32 +08:00
|
|
|
s->s_root = d_make_root(root_inode);
|
|
|
|
if (!s->s_root) {
|
2013-02-28 09:03:16 +08:00
|
|
|
pr_err("proc_fill_super: allocate dentry failed\n");
|
2013-02-20 10:13:32 +08:00
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
2014-07-31 18:10:50 +08:00
|
|
|
ret = proc_setup_self(s);
|
|
|
|
if (ret) {
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
return proc_setup_thread_self(s);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|