Some fanotify groups, especially those like AV scanners, will need to place
lots of marks, particularly ignore marks.  Since ignore marks do not pin
inodes in cache and are cleared if the inode is removed from core (usually
under memory pressure) we expose an interface for listeners, with
CAP_SYS_ADMIN, to override the maximum number of marks and be allowed to
set and 'unlimited' number of marks.  Programs which make use of this
feature will be able to OOM a machine.
Signed-off-by: NEric Paris <eparis@redhat.com>

There is currently no limit on the number of marks a given fanotify group
can have.  Since fanotify is gated on CAP_SYS_ADMIN this was not seen as
a serious DoS threat.  This patch implements a default of 8192, the same as
inotify to work towards removing the CAP_SYS_ADMIN gating and eliminating
the default DoS'able status.
Signed-off-by: NEric Paris <eparis@redhat.com>

fanotify has a defualt max queue depth.  This patch allows processes which
explicitly request it to have an 'unlimited' queue depth.  These processes
need to be very careful to make sure they cannot fall far enough behind
that they OOM the box.  Thus this flag is gated on CAP_SYS_ADMIN.
Signed-off-by: NEric Paris <eparis@redhat.com>

Currently fanotify has no maximum queue depth.  Since fanotify is
CAP_SYS_ADMIN only this does not pose a normal user DoS issue, but it
certianly is possible that an fanotify listener which can't keep up could
OOM the box.  This patch implements a default 16k depth.  This is the same
default depth used by inotify, but given fanotify's better queue merging in
many situations this queue will contain many additional useful events by
comparison.
Signed-off-by: NEric Paris <eparis@redhat.com>

fanotify will clear ignore marks if a task changes the contents of an
inode.  The problem is with the races around when userspace finishes
checking a file and when that result is actually attached to the inode.
This race was described as such:

Consider the following scenario with hostile processes A and B, and
victim process C:
1. Process A opens new file for writing. File check request is generated.
2. File check is performed in userspace. Check result is "file has no malware".
3. The "permit" response is delivered to kernel space.
4. File ignored mark set.
5. Process A writes dummy bytes to the file. File ignored flags are cleared.
6. Process B opens the same file for reading. File check request is generated.
7. File check is performed in userspace. Check result is "file has no malware".
8. Process A writes malware bytes to the file. There is no cached response yet.
9. The "permit" response is delivered to kernel space and is cached in fanotify.
10. File ignored mark set.
11. Now any process C will be permitted to open the malware file.
There is a race between steps 8 and 10

While fanotify makes no strong guarantees about systems with hostile
processes there is no reason we cannot harden against this race.  We do
that by simply ignoring any ignore marks if the inode has open writers (aka
i_writecount > 0).  (We actually do not ignore ignore marks if the
FAN_MARK_SURV_MODIFY flag is set)
Reported-by: NVasily Novikov <vasily.novikov@kaspersky.com>
Signed-off-by: NEric Paris <eparis@redhat.com>

fsnotify perm events do not call fsnotify parent.  That means you cannot
register a perm event on a directory and enforce permissions on all inodes in
that directory.  This patch fixes that situation.
Signed-off-by: NEric Paris <eparis@redhat.com>

When fsnotify groups return errors they are ignored.  For permissions
events these should be passed back up the stack, but for most events these
should continue to be ignored.
Signed-off-by: NEric Paris <eparis@redhat.com>

The fanotify listeners needs to be able to specify what types of operations
they are going to perform so they can be ordered appropriately between other
listeners doing other types of operations.  They need this to be able to make
sure that things like hierarchichal storage managers will get access to inodes
before processes which need the data.  This patch defines 3 possible uses
which groups must indicate in the fanotify_init() flags.

FAN_CLASS_PRE_CONTENT
FAN_CLASS_CONTENT
FAN_CLASS_NOTIF

Groups will receive notification in that order.  The order between 2 groups in
the same class is undeterministic.

FAN_CLASS_PRE_CONTENT is intended to be used by listeners which need access to
the inode before they are certain that the inode contains it's final data.  A
hierarchical storage manager should choose to use this class.

FAN_CLASS_CONTENT is intended to be used by listeners which need access to the
inode after it contains its intended contents.  This would be the appropriate
level for an AV solution or document control system.

FAN_CLASS_NOTIF is intended for normal async notification about access, much the
same as inotify and dnotify.  Syncronous permissions events are not permitted
at this class.
Signed-off-by: NEric Paris <eparis@redhat.com>

fanotify needs to be able to specify that some groups get events before
others.  They use this idea to make sure that a hierarchical storage
manager gets access to files before programs which actually use them.  This
is purely infrastructure.  Everything will have a priority of 0, but the
infrastructure will exist for it to be non-zero.
Signed-off-by: NEric Paris <eparis@redhat.com>

We disabled the ability to build fanotify in commit 7c534773.
This reverts that commit and allows people to build fanotify.
Signed-off-by: NEric Paris <eparis@redhat.com>

Pull removal of fsnotify marks into generic_shutdown_super().
Split umount-time work into a new function - evict_inodes().
Make sure that invalidate_inodes() will be able to cope with
I_FREEING once we change locking in iput().
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Use dget_parent instead of opencoding it.  This simplifies the code, but
more importanly prepares for the more complicated locking for a parent
dget in the dcache scale patch series.

It means we do grab a reference to the parent now if need to be watched,
but not with the specified mask.  If this turns out to be a problem
we'll have to revisit it, but for now let's keep as much as possible
dcache internals inside dcache.[ch].
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

All file_operations should get a .llseek operation so we can make
nonseekable_open the default for future file operations without a
.llseek pointer.

The three cases that we can automatically detect are no_llseek, seq_lseek
and default_llseek. For cases where we can we can automatically prove that
the file offset is always ignored, we use noop_llseek, which maintains
the current behavior of not returning an error from a seek.

New drivers should normally not use noop_llseek but instead use no_llseek
and call nonseekable_open at open time.  Existing drivers can be converted
to do the same when the maintainer knows for certain that no user code
relies on calling seek on the device file.

The generated code is often incorrectly indented and right now contains
comments that clarify for each added line why a specific variant was
chosen. In the version that gets submitted upstream, the comments will
be gone and I will manually fix the indentation, because there does not
seem to be a way to do that using coccinelle.

Some amount of new code is currently sitting in linux-next that should get
the same modifications, which I will do at the end of the merge window.

Many thanks to Julia Lawall for helping me learn to write a semantic
patch that does all this.

===== begin semantic patch =====
// This adds an llseek= method to all file operations,
// as a preparation for making no_llseek the default.
//
// The rules are
// - use no_llseek explicitly if we do nonseekable_open
// - use seq_lseek for sequential files
// - use default_llseek if we know we access f_pos
// - use noop_llseek if we know we don't access f_pos,
//   but we still want to allow users to call lseek
//
@ open1 exists @
identifier nested_open;
@@
nested_open(...)
{
<+...
nonseekable_open(...)
...+>
}

@ open exists@
identifier open_f;
identifier i, f;
identifier open1.nested_open;
@@
int open_f(struct inode *i, struct file *f)
{
<+...
(
nonseekable_open(...)
|
nested_open(...)
)
...+>
}

@ read disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
<+...
(
   *off = E
|
   *off += E
|
   func(..., off, ...)
|
   E = *off
)
...+>
}

@ read_no_fpos disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
... when != off
}

@ write @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
<+...
(
  *off = E
|
  *off += E
|
  func(..., off, ...)
|
  E = *off
)
...+>
}

@ write_no_fpos @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
... when != off
}

@ fops0 @
identifier fops;
@@
struct file_operations fops = {
 ...
};

@ has_llseek depends on fops0 @
identifier fops0.fops;
identifier llseek_f;
@@
struct file_operations fops = {
...
 .llseek = llseek_f,
...
};

@ has_read depends on fops0 @
identifier fops0.fops;
identifier read_f;
@@
struct file_operations fops = {
...
 .read = read_f,
...
};

@ has_write depends on fops0 @
identifier fops0.fops;
identifier write_f;
@@
struct file_operations fops = {
...
 .write = write_f,
...
};

@ has_open depends on fops0 @
identifier fops0.fops;
identifier open_f;
@@
struct file_operations fops = {
...
 .open = open_f,
...
};

// use no_llseek if we call nonseekable_open
////////////////////////////////////////////
@ nonseekable1 depends on !has_llseek && has_open @
identifier fops0.fops;
identifier nso ~= "nonseekable_open";
@@
struct file_operations fops = {
...  .open = nso, ...
+.llseek = no_llseek, /* nonseekable */
};

@ nonseekable2 depends on !has_llseek @
identifier fops0.fops;
identifier open.open_f;
@@
struct file_operations fops = {
...  .open = open_f, ...
+.llseek = no_llseek, /* open uses nonseekable */
};

// use seq_lseek for sequential files
/////////////////////////////////////
@ seq depends on !has_llseek @
identifier fops0.fops;
identifier sr ~= "seq_read";
@@
struct file_operations fops = {
...  .read = sr, ...
+.llseek = seq_lseek, /* we have seq_read */
};

// use default_llseek if there is a readdir
///////////////////////////////////////////
@ fops1 depends on !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier readdir_e;
@@
// any other fop is used that changes pos
struct file_operations fops = {
... .readdir = readdir_e, ...
+.llseek = default_llseek, /* readdir is present */
};

// use default_llseek if at least one of read/write touches f_pos
/////////////////////////////////////////////////////////////////
@ fops2 depends on !fops1 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read.read_f;
@@
// read fops use offset
struct file_operations fops = {
... .read = read_f, ...
+.llseek = default_llseek, /* read accesses f_pos */
};

@ fops3 depends on !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write.write_f;
@@
// write fops use offset
struct file_operations fops = {
... .write = write_f, ...
+	.llseek = default_llseek, /* write accesses f_pos */
};

// Use noop_llseek if neither read nor write accesses f_pos
///////////////////////////////////////////////////////////

@ fops4 depends on !fops1 && !fops2 && !fops3 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
identifier write_no_fpos.write_f;
@@
// write fops use offset
struct file_operations fops = {
...
 .write = write_f,
 .read = read_f,
...
+.llseek = noop_llseek, /* read and write both use no f_pos */
};

@ depends on has_write && !has_read && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write_no_fpos.write_f;
@@
struct file_operations fops = {
... .write = write_f, ...
+.llseek = noop_llseek, /* write uses no f_pos */
};

@ depends on has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
@@
struct file_operations fops = {
... .read = read_f, ...
+.llseek = noop_llseek, /* read uses no f_pos */
};

@ depends on !has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
@@
struct file_operations fops = {
...
+.llseek = noop_llseek, /* no read or write fn */
};
===== End semantic patch =====
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Cc: Julia Lawall <julia@diku.dk>
Cc: Christoph Hellwig <hch@infradead.org>

This patch disables the fanotify syscalls by just not building them and
letting the cond_syscall() statements in kernel/sys_ni.c redirect them
to sys_ni_syscall().

It was pointed out by Tvrtko Ursulin that the fanotify interface did not
include an explicit prioritization between groups.  This is necessary
for fanotify to be usable for hierarchical storage management software,
as they must get first access to the file, before inotify-like notifiers
see the file.

This feature can be added in an ABI compatible way in the next release
(by using a number of bits in the flags field to carry the info) but it
was suggested by Alan that maybe we should just hold off and do it in
the next cycle, likely with an (new) explicit argument to the syscall.
I don't like this approach best as I know people are already starting to
use the current interface, but Alan is all wise and noone on list backed
me up with just using what we have.  I feel this is needlessly ripping
the rug out from under people at the last minute, but if others think it
needs to be a new argument it might be the best way forward.

Three choices:
Go with what we got (and implement the new feature next cycle).  Add a
new field right now (and implement the new feature next cycle).  Wait
till next cycle to release the ABI (and implement the new feature next
cycle).  This is number 3.
Signed-off-by: NEric Paris <eparis@redhat.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The fsnotify main loop has 2 bools which indicated if we processed the
inode or vfsmount mark in that particular pass through the loop.  These
bool can we replaced with the inode_group and vfsmount_group variables
and actually make the code a little easier to understand.
Signed-off-by: NEric Paris <eparis@redhat.com>

Marks were stored on the inode and vfsmonut mark list in order from
highest memory address to lowest memory address.  The code to walk those
lists thought they were in order from lowest to highest with
unpredictable results when trying to match up marks from each.  It was
possible that extra events would be sent to userspace when inode
marks ignoring events wouldn't get matched with the vfsmount marks.

This problem only affected fanotify when using both vfsmount and inode
marks simultaneously.
Signed-off-by: NEric Paris <eparis@redhat.com>

The appropriate error code when privileged operations are denied is
EPERM, not EACCES.
Signed-off-by: NAndreas Gruenbacher <agruen@suse.de>
Signed-off-by: NEric Paris <paris@paris.rdu.redhat.com>

This reminded me... you have two pr_debugs in fanotify_should_send_event
which output redundant information. Maybe you intended it like that so
it is selectable how much log spam you want, or if not you may want to
apply this patch.
Signed-off-by: NTvrtko Ursulin <tvrtko.ursulin@sophos.com>
Signed-off-by: NEric Paris <eparis@redhat.com>

When an fanotify listener is closing it may cause a deadlock between the
listener and the original task doing an fs operation.  If the original task
is waiting for a permissions response it will be holding the srcu lock.  The
listener cannot clean up and exit until after that srcu lock is syncronized.
Thus deadlock.  The fix introduced here is to stop accepting new permissions
events when a listener is shutting down and to grant permission for all
outstanding events.  Thus the original task will eventually release the srcu
lock and the listener can complete shutdown.
Reported-by: NAndreas Gruenbacher <agruen@suse.de>
Cc: Andreas Gruenbacher <agruen@suse.de>
Signed-off-by: NEric Paris <eparis@redhat.com>

The interesting 2 list lockstep walking didn't quite work out if the inode
marks only had ignores and the vfsmount list requested events.  The code to
shortcut list traversal would not run the inode list since it didn't have real
event requests.  This code forces inode list traversal when a vfsmount mark
matches the event type.  Maybe we could add an i_fsnotify_ignored_mask field
to struct inode to get the shortcut back, but it doesn't seem worth it to grow
struct inode again.

I bet with the recent changes to lock the way we do now it would actually not
be a major perf hit to just drop i_fsnotify_mark_mask altogether.  But that is
for another day.
Signed-off-by: NEric Paris <eparis@redhat.com>

The fsnotify main loop has 2 booleans which tell if a particular mark was
sent to the listeners or if it should be processed in the next pass.  The
problem is that the booleans were not reset on each traversal of the loop.
So marks could get skipped even when they were not sent to the notifiers.
Reported-by: NTvrtko Ursulin <tvrtko.ursulin@sophos.com>
Signed-off-by: NEric Paris <eparis@redhat.com>

The fanotify code is supposed to get the group from the mark.  It accidentally
only used the inode_mark.  If the vfsmount_mark was set but not the inode_mark
it would deref the NULL inode_mark.  Get the group from the correct place.
Reported-by: NTvrtko Ursulin <tvrtko.ursulin@sophos.com>
Signed-off-by: NEric Paris <eparis@redhat.com>

This reverts commit 3bcf3860 (and the
accompanying commit c1e5c954 "vfs/fsnotify: fsnotify_close can delay
the final work in fput" that was a horribly ugly hack to make it work at
all).

The 'struct file' approach not only causes that disgusting hack, it
somehow breaks pulseaudio, probably due to some other subtlety with
f_count handling.

Fix up various conflicts due to later fsnotify work.
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

add I_CLEAR instead of replacing I_FREEING with it.  I_CLEAR is
equivalent to I_FREEING for almost all code looking at either;
it's there to keep track of having called clear_inode() exactly
once per inode lifetime, at some point after having set I_FREEING.
I_CLEAR and I_FREEING never get set at the same time with the
current code, so we can switch to setting i_flags to I_FREEING | I_CLEAR
instead of I_CLEAR without loss of information.  As the result of
such change, checks become simpler and the amount of code that needs
to know about I_CLEAR shrinks a lot.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

fanotify currently, when given a vfsmount_mark will look up (if it exists)
the corresponding inode mark.  This patch drops that lookup and uses the
mark provided.
Signed-off-by: NEric Paris <eparis@redhat.com>

should_send_event() and handle_event() will both need to look up the inode
event if they get a vfsmount event.  Lets just pass both at the same time
since we have them both after walking the lists in lockstep.
Signed-off-by: NEric Paris <eparis@redhat.com>

We currently walk the list of marks on an inode followed by the list of
marks on the vfsmount.  These are in order (by the memory address of the
group) so lets walk them both together.  Eventually we can pass both the
inode mark and the vfsmount mark to helpers simultaneously.
Signed-off-by: NEric Paris <eparis@redhat.com>

currently ignored_mark clearing is done in a seperate list traversal
before the actual list traversal to send events.  There is no need for
this.  Do them at the same time.
Signed-off-by: NEric Paris <eparis@redhat.com>

The global fsnotify groups lists were invented as a way to increase the
performance of fsnotify by shortcutting events which were not interesting.
With the changes to walk the object lists rather than global groups lists
these shortcuts are not useful.
Signed-off-by: NEric Paris <eparis@redhat.com>

group->mask is now useless.  It was originally a shortcut for fsnotify to
save on performance.  These checks are now redundant, so we remove them.
Signed-off-by: NEric Paris <eparis@redhat.com>

Because we walk the object->fsnotify_marks list instead of the global
fsnotify groups list we don't need the fsnotify_inode_mask and
fsnotify_vfsmount_mask as these were simply shortcuts in fsnotify() for
performance.  They are now extra checks, rip them out.
Signed-off-by: NEric Paris <eparis@redhat.com>

The change to use srcu and walk the object list rather than the global
fsnotify_group list means that should_send_event is no longer needed for a
number of groups and can be simplified for others.  Do that.
Signed-off-by: NEric Paris <eparis@redhat.com>

fanotify now gets a mark in the should_send_event and handle_event
functions.  Rather than look up the mark themselves fanotify should just use
the mark it was handed.
Signed-off-by: NEric Paris <eparis@redhat.com>

dnotify now gets a mark in the should_send_event and handle_event
functions.  Rather than look up the mark themselves dnotify should just use
the mark it was handed.
Signed-off-by: NEric Paris <eparis@redhat.com>

inotify now gets a mark in the should_send_event and handle_event
functions.  Rather than look up the mark themselves inotify should just use
the mark it was handed.
Signed-off-by: NEric Paris <eparis@redhat.com>

With the change of fsnotify to use srcu walking the marks list instead of
walking the global groups list we now know the mark in question.  The code can
send the mark to the group's handling functions and the groups won't have to
find those marks themselves.
Signed-off-by: NEric Paris <eparis@redhat.com>

Instead of moving list elements from destroy_list to &private_destroy_list,
exchange the list heads.
Signed-off-by: NAndreas Gruenbacher <agruen@suse.de>
Signed-off-by: NEric Paris <eparis@redhat.com>

Currently reading the inode->i_fsnotify_marks or
vfsmount->mnt_fsnotify_marks lists are protected by a spinlock on both the
read and the write side.  This patch protects the read side of those lists
with a new single srcu.
Signed-off-by: NEric Paris <eparis@redhat.com>

Currently fsnotify check is mark->group is NULL to decide if
fsnotify_destroy_mark() has already been called or not.  With the upcoming
rcu work it is a heck of a lot easier to use an explicit flag than worry
about group being set to NULL.
Signed-off-by: NEric Paris <eparis@redhat.com>

In preparation for srcu locking use all _rcu appropiete functions for mark
list addition, removal, and traversal.  The operations are still done under a
spinlock at the end of this patch.
Signed-off-by: NEric Paris <eparis@redhat.com>