5 This write-up is based on three articles published at lwn.net:
7 - <https://lwn.net/Articles/649115/> Pathname lookup in Linux
8 - <https://lwn.net/Articles/649729/> RCU-walk: faster pathname lookup in Linux
9 - <https://lwn.net/Articles/650786/> A walk among the symlinks
11 Written by Neil Brown with help from Al Viro and Jon Corbet.
12 It has subsequently been updated to reflect changes in the kernel
15 - per-directory parallel name lookup.
16 - ``openat2()`` resolution restriction flags.
18 Introduction to pathname lookup
19 ===============================
21 The most obvious aspect of pathname lookup, which very little
22 exploration is needed to discover, is that it is complex. There are
23 many rules, special cases, and implementation alternatives that all
24 combine to confuse the unwary reader. Computer science has long been
25 acquainted with such complexity and has tools to help manage it. One
26 tool that we will make extensive use of is "divide and conquer". For
27 the early parts of the analysis we will divide off symlinks - leaving
28 them until the final part. Well before we get to symlinks we have
29 another major division based on the VFS's approach to locking which
30 will allow us to review "REF-walk" and "RCU-walk" separately. But we
31 are getting ahead of ourselves. There are some important low level
32 distinctions we need to clarify first.
34 There are two sorts of ...
35 --------------------------
37 .. _openat: http://man7.org/linux/man-pages/man2/openat.2.html
39 Pathnames (sometimes "file names"), used to identify objects in the
40 filesystem, will be familiar to most readers. They contain two sorts
41 of elements: "slashes" that are sequences of one or more "``/``"
42 characters, and "components" that are sequences of one or more
43 non-"``/``" characters. These form two kinds of paths. Those that
44 start with slashes are "absolute" and start from the filesystem root.
45 The others are "relative" and start from the current directory, or
46 from some other location specified by a file descriptor given to
47 "``*at()``" system calls such as `openat() <openat_>`_.
49 .. _execveat: http://man7.org/linux/man-pages/man2/execveat.2.html
51 It is tempting to describe the second kind as starting with a
52 component, but that isn't always accurate: a pathname can lack both
53 slashes and components, it can be empty, in other words. This is
54 generally forbidden in POSIX, but some of those "``*at()``" system calls
55 in Linux permit it when the ``AT_EMPTY_PATH`` flag is given. For
56 example, if you have an open file descriptor on an executable file you
57 can execute it by calling `execveat() <execveat_>`_ passing
58 the file descriptor, an empty path, and the ``AT_EMPTY_PATH`` flag.
60 These paths can be divided into two sections: the final component and
61 everything else. The "everything else" is the easy bit. In all cases
62 it must identify a directory that already exists, otherwise an error
63 such as ``ENOENT`` or ``ENOTDIR`` will be reported.
65 The final component is not so simple. Not only do different system
66 calls interpret it quite differently (e.g. some create it, some do
67 not), but it might not even exist: neither the empty pathname nor the
68 pathname that is just slashes have a final component. If it does
69 exist, it could be "``.``" or "``..``" which are handled quite differently
70 from other components.
72 .. _POSIX: https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap04.html#tag_04_12
74 If a pathname ends with a slash, such as "``/tmp/foo/``" it might be
75 tempting to consider that to have an empty final component. In many
76 ways that would lead to correct results, but not always. In
77 particular, ``mkdir()`` and ``rmdir()`` each create or remove a directory named
78 by the final component, and they are required to work with pathnames
79 ending in "``/``". According to POSIX_:
81 A pathname that contains at least one non-<slash> character and
82 that ends with one or more trailing <slash> characters shall not
83 be resolved successfully unless the last pathname component before
84 the trailing <slash> characters names an existing directory or a
85 directory entry that is to be created for a directory immediately
86 after the pathname is resolved.
88 The Linux pathname walking code (mostly in ``fs/namei.c``) deals with
89 all of these issues: breaking the path into components, handling the
90 "everything else" quite separately from the final component, and
91 checking that the trailing slash is not used where it isn't
92 permitted. It also addresses the important issue of concurrent
95 While one process is looking up a pathname, another might be making
96 changes that affect that lookup. One fairly extreme case is that if
97 "a/b" were renamed to "a/c/b" while another process were looking up
98 "a/b/..", that process might successfully resolve on "a/c".
99 Most races are much more subtle, and a big part of the task of
100 pathname lookup is to prevent them from having damaging effects. Many
101 of the possible races are seen most clearly in the context of the
102 "dcache" and an understanding of that is central to understanding
105 More than just a cache
106 ----------------------
108 The "dcache" caches information about names in each filesystem to
109 make them quickly available for lookup. Each entry (known as a
110 "dentry") contains three significant fields: a component name, a
111 pointer to a parent dentry, and a pointer to the "inode" which
112 contains further information about the object in that parent with
113 the given name. The inode pointer can be ``NULL`` indicating that the
114 name doesn't exist in the parent. While there can be linkage in the
115 dentry of a directory to the dentries of the children, that linkage is
116 not used for pathname lookup, and so will not be considered here.
118 The dcache has a number of uses apart from accelerating lookup. One
119 that will be particularly relevant is that it is closely integrated
120 with the mount table that records which filesystem is mounted where.
121 What the mount table actually stores is which dentry is mounted on top
122 of which other dentry.
124 When considering the dcache, we have another of our "two types"
125 distinctions: there are two types of filesystems.
127 Some filesystems ensure that the information in the dcache is always
128 completely accurate (though not necessarily complete). This can allow
129 the VFS to determine if a particular file does or doesn't exist
130 without checking with the filesystem, and means that the VFS can
131 protect the filesystem against certain races and other problems.
132 These are typically "local" filesystems such as ext3, XFS, and Btrfs.
134 Other filesystems don't provide that guarantee because they cannot.
135 These are typically filesystems that are shared across a network,
136 whether remote filesystems like NFS and 9P, or cluster filesystems
137 like ocfs2 or cephfs. These filesystems allow the VFS to revalidate
138 cached information, and must provide their own protection against
139 awkward races. The VFS can detect these filesystems by the
140 ``DCACHE_OP_REVALIDATE`` flag being set in the dentry.
142 REF-walk: simple concurrency management with refcounts and spinlocks
143 --------------------------------------------------------------------
145 With all of those divisions carefully classified, we can now start
146 looking at the actual process of walking along a path. In particular
147 we will start with the handling of the "everything else" part of a
148 pathname, and focus on the "REF-walk" approach to concurrency
149 management. This code is found in the ``link_path_walk()`` function, if
150 you ignore all the places that only run when "``LOOKUP_RCU``"
151 (indicating the use of RCU-walk) is set.
153 .. _Meet the Lockers: https://lwn.net/Articles/453685/
155 REF-walk is fairly heavy-handed with locks and reference counts. Not
156 as heavy-handed as in the old "big kernel lock" days, but certainly not
157 afraid of taking a lock when one is needed. It uses a variety of
158 different concurrency controls. A background understanding of the
159 various primitives is assumed, or can be gleaned from elsewhere such
160 as in `Meet the Lockers`_.
162 The locking mechanisms used by REF-walk include:
167 This uses the lockref primitive to provide both a spinlock and a
168 reference count. The special-sauce of this primitive is that the
169 conceptual sequence "lock; inc_ref; unlock;" can often be performed
170 with a single atomic memory operation.
172 Holding a reference on a dentry ensures that the dentry won't suddenly
173 be freed and used for something else, so the values in various fields
174 will behave as expected. It also protects the ``->d_inode`` reference
175 to the inode to some extent.
177 The association between a dentry and its inode is fairly permanent.
178 For example, when a file is renamed, the dentry and inode move
179 together to the new location. When a file is created the dentry will
180 initially be negative (i.e. ``d_inode`` is ``NULL``), and will be assigned
181 to the new inode as part of the act of creation.
183 When a file is deleted, this can be reflected in the cache either by
184 setting ``d_inode`` to ``NULL``, or by removing it from the hash table
185 (described shortly) used to look up the name in the parent directory.
186 If the dentry is still in use the second option is used as it is
187 perfectly legal to keep using an open file after it has been deleted
188 and having the dentry around helps. If the dentry is not otherwise in
189 use (i.e. if the refcount in ``d_lockref`` is one), only then will
190 ``d_inode`` be set to ``NULL``. Doing it this way is more efficient for a
193 So as long as a counted reference is held to a dentry, a non-``NULL`` ``->d_inode``
194 value will never be changed.
199 ``d_lock`` is a synonym for the spinlock that is part of ``d_lockref`` above.
200 For our purposes, holding this lock protects against the dentry being
201 renamed or unlinked. In particular, its parent (``d_parent``), and its
202 name (``d_name``) cannot be changed, and it cannot be removed from the
205 When looking for a name in a directory, REF-walk takes ``d_lock`` on
206 each candidate dentry that it finds in the hash table and then checks
207 that the parent and name are correct. So it doesn't lock the parent
208 while searching in the cache; it only locks children.
210 When looking for the parent for a given name (to handle "``..``"),
211 REF-walk can take ``d_lock`` to get a stable reference to ``d_parent``,
212 but it first tries a more lightweight approach. As seen in
213 ``dget_parent()``, if a reference can be claimed on the parent, and if
214 subsequently ``d_parent`` can be seen to have not changed, then there is
215 no need to actually take the lock on the child.
220 Looking up a given name in a given directory involves computing a hash
221 from the two values (the name and the dentry of the directory),
222 accessing that slot in a hash table, and searching the linked list
225 When a dentry is renamed, the name and the parent dentry can both
226 change so the hash will almost certainly change too. This would move the
227 dentry to a different chain in the hash table. If a filename search
228 happened to be looking at a dentry that was moved in this way,
229 it might end up continuing the search down the wrong chain,
230 and so miss out on part of the correct chain.
232 The name-lookup process (``d_lookup()``) does *not* try to prevent this
233 from happening, but only to detect when it happens.
234 ``rename_lock`` is a seqlock that is updated whenever any dentry is
235 renamed. If ``d_lookup`` finds that a rename happened while it
236 unsuccessfully scanned a chain in the hash table, it simply tries
239 ``rename_lock`` is also used to detect and defend against potential attacks
240 against ``LOOKUP_BENEATH`` and ``LOOKUP_IN_ROOT`` when resolving ".." (where
241 the parent directory is moved outside the root, bypassing the ``path_equal()``
242 check). If ``rename_lock`` is updated during the lookup and the path encounters
243 a "..", a potential attack occurred and ``handle_dots()`` will bail out with
249 ``i_rwsem`` is a read/write semaphore that serializes all changes to a particular
250 directory. This ensures that, for example, an ``unlink()`` and a ``rename()``
251 cannot both happen at the same time. It also keeps the directory
252 stable while the filesystem is asked to look up a name that is not
253 currently in the dcache or, optionally, when the list of entries in a
254 directory is being retrieved with ``readdir()``.
256 This has a complementary role to that of ``d_lock``: ``i_rwsem`` on a
257 directory protects all of the names in that directory, while ``d_lock``
258 on a name protects just one name in a directory. Most changes to the
259 dcache hold ``i_rwsem`` on the relevant directory inode and briefly take
260 ``d_lock`` on one or more the dentries while the change happens. One
261 exception is when idle dentries are removed from the dcache due to
262 memory pressure. This uses ``d_lock``, but ``i_rwsem`` plays no role.
264 The semaphore affects pathname lookup in two distinct ways. Firstly it
265 prevents changes during lookup of a name in a directory. ``walk_component()`` uses
266 ``lookup_fast()`` first which, in turn, checks to see if the name is in the cache,
267 using only ``d_lock`` locking. If the name isn't found, then ``walk_component()``
268 falls back to ``lookup_slow()`` which takes a shared lock on ``i_rwsem``, checks again that
269 the name isn't in the cache, and then calls in to the filesystem to get a
270 definitive answer. A new dentry will be added to the cache regardless of
273 Secondly, when pathname lookup reaches the final component, it will
274 sometimes need to take an exclusive lock on ``i_rwsem`` before performing the last lookup so
275 that the required exclusion can be achieved. How path lookup chooses
276 to take, or not take, ``i_rwsem`` is one of the
277 issues addressed in a subsequent section.
279 If two threads attempt to look up the same name at the same time - a
280 name that is not yet in the dcache - the shared lock on ``i_rwsem`` will
281 not prevent them both adding new dentries with the same name. As this
282 would result in confusion an extra level of interlocking is used,
283 based around a secondary hash table (``in_lookup_hashtable``) and a
284 per-dentry flag bit (``DCACHE_PAR_LOOKUP``).
286 To add a new dentry to the cache while only holding a shared lock on
287 ``i_rwsem``, a thread must call ``d_alloc_parallel()``. This allocates a
288 dentry, stores the required name and parent in it, checks if there
289 is already a matching dentry in the primary or secondary hash
290 tables, and if not, stores the newly allocated dentry in the secondary
291 hash table, with ``DCACHE_PAR_LOOKUP`` set.
293 If a matching dentry was found in the primary hash table then that is
294 returned and the caller can know that it lost a race with some other
295 thread adding the entry. If no matching dentry is found in either
296 cache, the newly allocated dentry is returned and the caller can
297 detect this from the presence of ``DCACHE_PAR_LOOKUP``. In this case it
298 knows that it has won any race and now is responsible for asking the
299 filesystem to perform the lookup and find the matching inode. When
300 the lookup is complete, it must call ``d_lookup_done()`` which clears
301 the flag and does some other house keeping, including removing the
302 dentry from the secondary hash table - it will normally have been
303 added to the primary hash table already. Note that a ``struct
304 waitqueue_head`` is passed to ``d_alloc_parallel()``, and
305 ``d_lookup_done()`` must be called while this ``waitqueue_head`` is still
308 If a matching dentry is found in the secondary hash table,
309 ``d_alloc_parallel()`` has a little more work to do. It first waits for
310 ``DCACHE_PAR_LOOKUP`` to be cleared, using a wait_queue that was passed
311 to the instance of ``d_alloc_parallel()`` that won the race and that
312 will be woken by the call to ``d_lookup_done()``. It then checks to see
313 if the dentry has now been added to the primary hash table. If it
314 has, the dentry is returned and the caller just sees that it lost any
315 race. If it hasn't been added to the primary hash table, the most
316 likely explanation is that some other dentry was added instead using
317 ``d_splice_alias()``. In any case, ``d_alloc_parallel()`` repeats all the
318 look ups from the start and will normally return something from the
324 ``mnt_count`` is a per-CPU reference counter on "``mount``" structures.
325 Per-CPU here means that incrementing the count is cheap as it only
326 uses CPU-local memory, but checking if the count is zero is expensive as
327 it needs to check with every CPU. Taking a ``mnt_count`` reference
328 prevents the mount structure from disappearing as the result of regular
329 unmount operations, but does not prevent a "lazy" unmount. So holding
330 ``mnt_count`` doesn't ensure that the mount remains in the namespace and,
331 in particular, doesn't stabilize the link to the mounted-on dentry. It
332 does, however, ensure that the ``mount`` data structure remains coherent,
333 and it provides a reference to the root dentry of the mounted
334 filesystem. So a reference through ``->mnt_count`` provides a stable
335 reference to the mounted dentry, but not the mounted-on dentry.
340 ``mount_lock`` is a global seqlock, a bit like ``rename_lock``. It can be used to
341 check if any change has been made to any mount points.
343 While walking down the tree (away from the root) this lock is used when
344 crossing a mount point to check that the crossing was safe. That is,
345 the value in the seqlock is read, then the code finds the mount that
346 is mounted on the current directory, if there is one, and increments
347 the ``mnt_count``. Finally the value in ``mount_lock`` is checked against
348 the old value. If there is no change, then the crossing was safe. If there
349 was a change, the ``mnt_count`` is decremented and the whole process is
352 When walking up the tree (towards the root) by following a ".." link,
353 a little more care is needed. In this case the seqlock (which
354 contains both a counter and a spinlock) is fully locked to prevent
355 any changes to any mount points while stepping up. This locking is
356 needed to stabilize the link to the mounted-on dentry, which the
357 refcount on the mount itself doesn't ensure.
359 ``mount_lock`` is also used to detect and defend against potential attacks
360 against ``LOOKUP_BENEATH`` and ``LOOKUP_IN_ROOT`` when resolving ".." (where
361 the parent directory is moved outside the root, bypassing the ``path_equal()``
362 check). If ``mount_lock`` is updated during the lookup and the path encounters
363 a "..", a potential attack occurred and ``handle_dots()`` will bail out with
369 Finally the global (but extremely lightweight) RCU read lock is held
370 from time to time to ensure certain data structures don't get freed
373 In particular it is held while scanning chains in the dcache hash
374 table, and the mount point hash table.
376 Bringing it together with ``struct nameidata``
377 ----------------------------------------------
379 .. _First edition Unix: https://minnie.tuhs.org/cgi-bin/utree.pl?file=V1/u2.s
381 Throughout the process of walking a path, the current status is stored
382 in a ``struct nameidata``, "namei" being the traditional name - dating
383 all the way back to `First Edition Unix`_ - of the function that
384 converts a "name" to an "inode". ``struct nameidata`` contains (among
390 A ``path`` contains a ``struct vfsmount`` (which is
391 embedded in a ``struct mount``) and a ``struct dentry``. Together these
392 record the current status of the walk. They start out referring to the
393 starting point (the current working directory, the root directory, or some other
394 directory identified by a file descriptor), and are updated on each
395 step. A reference through ``d_lockref`` and ``mnt_count`` is always
401 This is a string together with a length (i.e. *not* ``nul`` terminated)
402 that is the "next" component in the pathname.
407 This is one of ``LAST_NORM``, ``LAST_ROOT``, ``LAST_DOT`` or ``LAST_DOTDOT``.
408 The ``last`` field is only valid if the type is ``LAST_NORM``.
413 This is used to hold a reference to the effective root of the
414 filesystem. Often that reference won't be needed, so this field is
415 only assigned the first time it is used, or when a non-standard root
416 is requested. Keeping a reference in the ``nameidata`` ensures that
417 only one root is in effect for the entire path walk, even if it races
418 with a ``chroot()`` system call.
420 It should be noted that in the case of ``LOOKUP_IN_ROOT`` or
421 ``LOOKUP_BENEATH``, the effective root becomes the directory file descriptor
422 passed to ``openat2()`` (which exposes these ``LOOKUP_`` flags).
424 The root is needed when either of two conditions holds: (1) either the
425 pathname or a symbolic link starts with a "'/'", or (2) a "``..``"
426 component is being handled, since "``..``" from the root must always stay
427 at the root. The value used is usually the current root directory of
428 the calling process. An alternate root can be provided as when
429 ``sysctl()`` calls ``file_open_root()``, and when NFSv4 or Btrfs call
430 ``mount_subtree()``. In each case a pathname is being looked up in a very
431 specific part of the filesystem, and the lookup must not be allowed to
432 escape that subtree. It works a bit like a local ``chroot()``.
434 Ignoring the handling of symbolic links, we can now describe the
435 "``link_path_walk()``" function, which handles the lookup of everything
436 except the final component as:
438 Given a path (``name``) and a nameidata structure (``nd``), check that the
439 current directory has execute permission and then advance ``name``
440 over one component while updating ``last_type`` and ``last``. If that
441 was the final component, then return, otherwise call
442 ``walk_component()`` and repeat from the top.
444 ``walk_component()`` is even easier. If the component is ``LAST_DOTS``,
445 it calls ``handle_dots()`` which does the necessary locking as already
446 described. If it finds a ``LAST_NORM`` component it first calls
447 "``lookup_fast()``" which only looks in the dcache, but will ask the
448 filesystem to revalidate the result if it is that sort of filesystem.
449 If that doesn't get a good result, it calls "``lookup_slow()``" which
450 takes ``i_rwsem``, rechecks the cache, and then asks the filesystem
451 to find a definitive answer. Each of these will call
452 ``follow_managed()`` (as described below) to handle any mount points.
454 In the absence of symbolic links, ``walk_component()`` creates a new
455 ``struct path`` containing a counted reference to the new dentry and a
456 reference to the new ``vfsmount`` which is only counted if it is
457 different from the previous ``vfsmount``. It then calls
458 ``path_to_nameidata()`` to install the new ``struct path`` in the
459 ``struct nameidata`` and drop the unneeded references.
461 This "hand-over-hand" sequencing of getting a reference to the new
462 dentry before dropping the reference to the previous dentry may
463 seem obvious, but is worth pointing out so that we will recognize its
464 analogue in the "RCU-walk" version.
466 Handling the final component
467 ----------------------------
469 ``link_path_walk()`` only walks as far as setting ``nd->last`` and
470 ``nd->last_type`` to refer to the final component of the path. It does
471 not call ``walk_component()`` that last time. Handling that final
472 component remains for the caller to sort out. Those callers are
473 ``path_lookupat()``, ``path_parentat()``, ``path_mountpoint()`` and
474 ``path_openat()`` each of which handles the differing requirements of
475 different system calls.
477 ``path_parentat()`` is clearly the simplest - it just wraps a little bit
478 of housekeeping around ``link_path_walk()`` and returns the parent
479 directory and final component to the caller. The caller will be either
480 aiming to create a name (via ``filename_create()``) or remove or rename
481 a name (in which case ``user_path_parent()`` is used). They will use
482 ``i_rwsem`` to exclude other changes while they validate and then
483 perform their operation.
485 ``path_lookupat()`` is nearly as simple - it is used when an existing
486 object is wanted such as by ``stat()`` or ``chmod()``. It essentially just
487 calls ``walk_component()`` on the final component through a call to
488 ``lookup_last()``. ``path_lookupat()`` returns just the final dentry.
490 ``path_mountpoint()`` handles the special case of unmounting which must
491 not try to revalidate the mounted filesystem. It effectively
492 contains, through a call to ``mountpoint_last()``, an alternate
493 implementation of ``lookup_slow()`` which skips that step. This is
494 important when unmounting a filesystem that is inaccessible, such as
495 one provided by a dead NFS server.
497 Finally ``path_openat()`` is used for the ``open()`` system call; it
498 contains, in support functions starting with "``do_last()``", all the
499 complexity needed to handle the different subtleties of O_CREAT (with
500 or without O_EXCL), final "``/``" characters, and trailing symbolic
501 links. We will revisit this in the final part of this series, which
502 focuses on those symbolic links. "``do_last()``" will sometimes, but
503 not always, take ``i_rwsem``, depending on what it finds.
505 Each of these, or the functions which call them, need to be alert to
506 the possibility that the final component is not ``LAST_NORM``. If the
507 goal of the lookup is to create something, then any value for
508 ``last_type`` other than ``LAST_NORM`` will result in an error. For
509 example if ``path_parentat()`` reports ``LAST_DOTDOT``, then the caller
510 won't try to create that name. They also check for trailing slashes
511 by testing ``last.name[last.len]``. If there is any character beyond
512 the final component, it must be a trailing slash.
514 Revalidation and automounts
515 ---------------------------
517 Apart from symbolic links, there are only two parts of the "REF-walk"
518 process not yet covered. One is the handling of stale cache entries
519 and the other is automounts.
521 On filesystems that require it, the lookup routines will call the
522 ``->d_revalidate()`` dentry method to ensure that the cached information
523 is current. This will often confirm validity or update a few details
524 from a server. In some cases it may find that there has been change
525 further up the path and that something that was thought to be valid
526 previously isn't really. When this happens the lookup of the whole
527 path is aborted and retried with the "``LOOKUP_REVAL``" flag set. This
528 forces revalidation to be more thorough. We will see more details of
529 this retry process in the next article.
531 Automount points are locations in the filesystem where an attempt to
532 lookup a name can trigger changes to how that lookup should be
533 handled, in particular by mounting a filesystem there. These are
534 covered in greater detail in autofs.txt in the Linux documentation
535 tree, but a few notes specifically related to path lookup are in order
538 The Linux VFS has a concept of "managed" dentries which is reflected
539 in function names such as "``follow_managed()``". There are three
540 potentially interesting things about these dentries corresponding
541 to three different flags that might be set in ``dentry->d_flags``:
543 ``DCACHE_MANAGE_TRANSIT``
544 ~~~~~~~~~~~~~~~~~~~~~~~~~
546 If this flag has been set, then the filesystem has requested that the
547 ``d_manage()`` dentry operation be called before handling any possible
548 mount point. This can perform two particular services:
550 It can block to avoid races. If an automount point is being
551 unmounted, the ``d_manage()`` function will usually wait for that
552 process to complete before letting the new lookup proceed and possibly
553 trigger a new automount.
555 It can selectively allow only some processes to transit through a
556 mount point. When a server process is managing automounts, it may
557 need to access a directory without triggering normal automount
558 processing. That server process can identify itself to the ``autofs``
559 filesystem, which will then give it a special pass through
560 ``d_manage()`` by returning ``-EISDIR``.
565 This flag is set on every dentry that is mounted on. As Linux
566 supports multiple filesystem namespaces, it is possible that the
567 dentry may not be mounted on in *this* namespace, just in some
568 other. So this flag is seen as a hint, not a promise.
570 If this flag is set, and ``d_manage()`` didn't return ``-EISDIR``,
571 ``lookup_mnt()`` is called to examine the mount hash table (honoring the
572 ``mount_lock`` described earlier) and possibly return a new ``vfsmount``
573 and a new ``dentry`` (both with counted references).
575 ``DCACHE_NEED_AUTOMOUNT``
576 ~~~~~~~~~~~~~~~~~~~~~~~~~
578 If ``d_manage()`` allowed us to get this far, and ``lookup_mnt()`` didn't
579 find a mount point, then this flag causes the ``d_automount()`` dentry
580 operation to be called.
582 The ``d_automount()`` operation can be arbitrarily complex and may
583 communicate with server processes etc. but it should ultimately either
584 report that there was an error, that there was nothing to mount, or
585 should provide an updated ``struct path`` with new ``dentry`` and ``vfsmount``.
587 In the latter case, ``finish_automount()`` will be called to safely
588 install the new mount point into the mount table.
590 There is no new locking of import here and it is important that no
591 locks (only counted references) are held over this processing due to
592 the very real possibility of extended delays.
593 This will become more important next time when we examine RCU-walk
594 which is particularly sensitive to delays.
596 RCU-walk - faster pathname lookup in Linux
597 ==========================================
599 RCU-walk is another algorithm for performing pathname lookup in Linux.
600 It is in many ways similar to REF-walk and the two share quite a bit
601 of code. The significant difference in RCU-walk is how it allows for
602 the possibility of concurrent access.
604 We noted that REF-walk is complex because there are numerous details
605 and special cases. RCU-walk reduces this complexity by simply
606 refusing to handle a number of cases -- it instead falls back to
607 REF-walk. The difficulty with RCU-walk comes from a different
608 direction: unfamiliarity. The locking rules when depending on RCU are
609 quite different from traditional locking, so we will spend a little extra
610 time when we come to those.
612 Clear demarcation of roles
613 --------------------------
615 The easiest way to manage concurrency is to forcibly stop any other
616 thread from changing the data structures that a given thread is
617 looking at. In cases where no other thread would even think of
618 changing the data and lots of different threads want to read at the
619 same time, this can be very costly. Even when using locks that permit
620 multiple concurrent readers, the simple act of updating the count of
621 the number of current readers can impose an unwanted cost. So the
622 goal when reading a shared data structure that no other process is
623 changing is to avoid writing anything to memory at all. Take no
624 locks, increment no counts, leave no footprints.
626 The REF-walk mechanism already described certainly doesn't follow this
627 principle, but then it is really designed to work when there may well
628 be other threads modifying the data. RCU-walk, in contrast, is
629 designed for the common situation where there are lots of frequent
630 readers and only occasional writers. This may not be common in all
631 parts of the filesystem tree, but in many parts it will be. For the
632 other parts it is important that RCU-walk can quickly fall back to
635 Pathname lookup always starts in RCU-walk mode but only remains there
636 as long as what it is looking for is in the cache and is stable. It
637 dances lightly down the cached filesystem image, leaving no footprints
638 and carefully watching where it is, to be sure it doesn't trip. If it
639 notices that something has changed or is changing, or if something
640 isn't in the cache, then it tries to stop gracefully and switch to
643 This stopping requires getting a counted reference on the current
644 ``vfsmount`` and ``dentry``, and ensuring that these are still valid -
645 that a path walk with REF-walk would have found the same entries.
646 This is an invariant that RCU-walk must guarantee. It can only make
647 decisions, such as selecting the next step, that are decisions which
648 REF-walk could also have made if it were walking down the tree at the
649 same time. If the graceful stop succeeds, the rest of the path is
650 processed with the reliable, if slightly sluggish, REF-walk. If
651 RCU-walk finds it cannot stop gracefully, it simply gives up and
652 restarts from the top with REF-walk.
654 This pattern of "try RCU-walk, if that fails try REF-walk" can be
655 clearly seen in functions like ``filename_lookup()``,
656 ``filename_parentat()``, ``filename_mountpoint()``,
657 ``do_filp_open()``, and ``do_file_open_root()``. These five
658 correspond roughly to the four ``path_*()`` functions we met earlier,
659 each of which calls ``link_path_walk()``. The ``path_*()`` functions are
660 called using different mode flags until a mode is found which works.
661 They are first called with ``LOOKUP_RCU`` set to request "RCU-walk". If
662 that fails with the error ``ECHILD`` they are called again with no
663 special flag to request "REF-walk". If either of those report the
664 error ``ESTALE`` a final attempt is made with ``LOOKUP_REVAL`` set (and no
665 ``LOOKUP_RCU``) to ensure that entries found in the cache are forcibly
666 revalidated - normally entries are only revalidated if the filesystem
667 determines that they are too old to trust.
669 The ``LOOKUP_RCU`` attempt may drop that flag internally and switch to
670 REF-walk, but will never then try to switch back to RCU-walk. Places
671 that trip up RCU-walk are much more likely to be near the leaves and
672 so it is very unlikely that there will be much, if any, benefit from
675 RCU and seqlocks: fast and light
676 --------------------------------
678 RCU is, unsurprisingly, critical to RCU-walk mode. The
679 ``rcu_read_lock()`` is held for the entire time that RCU-walk is walking
680 down a path. The particular guarantee it provides is that the key
681 data structures - dentries, inodes, super_blocks, and mounts - will
682 not be freed while the lock is held. They might be unlinked or
683 invalidated in one way or another, but the memory will not be
684 repurposed so values in various fields will still be meaningful. This
685 is the only guarantee that RCU provides; everything else is done using
688 As we saw above, REF-walk holds a counted reference to the current
689 dentry and the current vfsmount, and does not release those references
690 before taking references to the "next" dentry or vfsmount. It also
691 sometimes takes the ``d_lock`` spinlock. These references and locks are
692 taken to prevent certain changes from happening. RCU-walk must not
693 take those references or locks and so cannot prevent such changes.
694 Instead, it checks to see if a change has been made, and aborts or
697 To preserve the invariant mentioned above (that RCU-walk may only make
698 decisions that REF-walk could have made), it must make the checks at
699 or near the same places that REF-walk holds the references. So, when
700 REF-walk increments a reference count or takes a spinlock, RCU-walk
701 samples the status of a seqlock using ``read_seqcount_begin()`` or a
702 similar function. When REF-walk decrements the count or drops the
703 lock, RCU-walk checks if the sampled status is still valid using
704 ``read_seqcount_retry()`` or similar.
706 However, there is a little bit more to seqlocks than that. If
707 RCU-walk accesses two different fields in a seqlock-protected
708 structure, or accesses the same field twice, there is no a priori
709 guarantee of any consistency between those accesses. When consistency
710 is needed - which it usually is - RCU-walk must take a copy and then
711 use ``read_seqcount_retry()`` to validate that copy.
713 ``read_seqcount_retry()`` not only checks the sequence number, but also
714 imposes a memory barrier so that no memory-read instruction from
715 *before* the call can be delayed until *after* the call, either by the
716 CPU or by the compiler. A simple example of this can be seen in
717 ``slow_dentry_cmp()`` which, for filesystems which do not use simple
718 byte-wise name equality, calls into the filesystem to compare a name
719 against a dentry. The length and name pointer are copied into local
720 variables, then ``read_seqcount_retry()`` is called to confirm the two
721 are consistent, and only then is ``->d_compare()`` called. When
722 standard filename comparison is used, ``dentry_cmp()`` is called
723 instead. Notably it does *not* use ``read_seqcount_retry()``, but
724 instead has a large comment explaining why the consistency guarantee
725 isn't necessary. A subsequent ``read_seqcount_retry()`` will be
726 sufficient to catch any problem that could occur at this point.
728 With that little refresher on seqlocks out of the way we can look at
729 the bigger picture of how RCU-walk uses seqlocks.
731 ``mount_lock`` and ``nd->m_seq``
732 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
734 We already met the ``mount_lock`` seqlock when REF-walk used it to
735 ensure that crossing a mount point is performed safely. RCU-walk uses
736 it for that too, but for quite a bit more.
738 Instead of taking a counted reference to each ``vfsmount`` as it
739 descends the tree, RCU-walk samples the state of ``mount_lock`` at the
740 start of the walk and stores this initial sequence number in the
741 ``struct nameidata`` in the ``m_seq`` field. This one lock and one
742 sequence number are used to validate all accesses to all ``vfsmounts``,
743 and all mount point crossings. As changes to the mount table are
744 relatively rare, it is reasonable to fall back on REF-walk any time
745 that any "mount" or "unmount" happens.
747 ``m_seq`` is checked (using ``read_seqretry()``) at the end of an RCU-walk
748 sequence, whether switching to REF-walk for the rest of the path or
749 when the end of the path is reached. It is also checked when stepping
750 down over a mount point (in ``__follow_mount_rcu()``) or up (in
751 ``follow_dotdot_rcu()``). If it is ever found to have changed, the
752 whole RCU-walk sequence is aborted and the path is processed again by
755 If RCU-walk finds that ``mount_lock`` hasn't changed then it can be sure
756 that, had REF-walk taken counted references on each vfsmount, the
757 results would have been the same. This ensures the invariant holds,
758 at least for vfsmount structures.
760 ``dentry->d_seq`` and ``nd->seq``
761 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
763 In place of taking a count or lock on ``d_reflock``, RCU-walk samples
764 the per-dentry ``d_seq`` seqlock, and stores the sequence number in the
765 ``seq`` field of the nameidata structure, so ``nd->seq`` should always be
766 the current sequence number of ``nd->dentry``. This number needs to be
767 revalidated after copying, and before using, the name, parent, or
770 The handling of the name we have already looked at, and the parent is
771 only accessed in ``follow_dotdot_rcu()`` which fairly trivially follows
772 the required pattern, though it does so for three different cases.
774 When not at a mount point, ``d_parent`` is followed and its ``d_seq`` is
775 collected. When we are at a mount point, we instead follow the
776 ``mnt->mnt_mountpoint`` link to get a new dentry and collect its
777 ``d_seq``. Then, after finally finding a ``d_parent`` to follow, we must
778 check if we have landed on a mount point and, if so, must find that
779 mount point and follow the ``mnt->mnt_root`` link. This would imply a
780 somewhat unusual, but certainly possible, circumstance where the
781 starting point of the path lookup was in part of the filesystem that
782 was mounted on, and so not visible from the root.
784 The inode pointer, stored in ``->d_inode``, is a little more
785 interesting. The inode will always need to be accessed at least
786 twice, once to determine if it is NULL and once to verify access
787 permissions. Symlink handling requires a validated inode pointer too.
788 Rather than revalidating on each access, a copy is made on the first
789 access and it is stored in the ``inode`` field of ``nameidata`` from where
790 it can be safely accessed without further validation.
792 ``lookup_fast()`` is the only lookup routine that is used in RCU-mode,
793 ``lookup_slow()`` being too slow and requiring locks. It is in
794 ``lookup_fast()`` that we find the important "hand over hand" tracking
795 of the current dentry.
797 The current ``dentry`` and current ``seq`` number are passed to
798 ``__d_lookup_rcu()`` which, on success, returns a new ``dentry`` and a
799 new ``seq`` number. ``lookup_fast()`` then copies the inode pointer and
800 revalidates the new ``seq`` number. It then validates the old ``dentry``
801 with the old ``seq`` number one last time and only then continues. This
802 process of getting the ``seq`` number of the new dentry and then
803 checking the ``seq`` number of the old exactly mirrors the process of
804 getting a counted reference to the new dentry before dropping that for
805 the old dentry which we saw in REF-walk.
807 No ``inode->i_rwsem`` or even ``rename_lock``
808 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
810 A semaphore is a fairly heavyweight lock that can only be taken when it is
811 permissible to sleep. As ``rcu_read_lock()`` forbids sleeping,
812 ``inode->i_rwsem`` plays no role in RCU-walk. If some other thread does
813 take ``i_rwsem`` and modifies the directory in a way that RCU-walk needs
814 to notice, the result will be either that RCU-walk fails to find the
815 dentry that it is looking for, or it will find a dentry which
816 ``read_seqretry()`` won't validate. In either case it will drop down to
817 REF-walk mode which can take whatever locks are needed.
819 Though ``rename_lock`` could be used by RCU-walk as it doesn't require
820 any sleeping, RCU-walk doesn't bother. REF-walk uses ``rename_lock`` to
821 protect against the possibility of hash chains in the dcache changing
822 while they are being searched. This can result in failing to find
823 something that actually is there. When RCU-walk fails to find
824 something in the dentry cache, whether it is really there or not, it
825 already drops down to REF-walk and tries again with appropriate
826 locking. This neatly handles all cases, so adding extra checks on
827 rename_lock would bring no significant value.
829 ``unlazy walk()`` and ``complete_walk()``
830 -----------------------------------------
832 That "dropping down to REF-walk" typically involves a call to
833 ``unlazy_walk()``, so named because "RCU-walk" is also sometimes
834 referred to as "lazy walk". ``unlazy_walk()`` is called when
835 following the path down to the current vfsmount/dentry pair seems to
836 have proceeded successfully, but the next step is problematic. This
837 can happen if the next name cannot be found in the dcache, if
838 permission checking or name revalidation couldn't be achieved while
839 the ``rcu_read_lock()`` is held (which forbids sleeping), if an
840 automount point is found, or in a couple of cases involving symlinks.
841 It is also called from ``complete_walk()`` when the lookup has reached
842 the final component, or the very end of the path, depending on which
843 particular flavor of lookup is used.
845 Other reasons for dropping out of RCU-walk that do not trigger a call
846 to ``unlazy_walk()`` are when some inconsistency is found that cannot be
847 handled immediately, such as ``mount_lock`` or one of the ``d_seq``
848 seqlocks reporting a change. In these cases the relevant function
849 will return ``-ECHILD`` which will percolate up until it triggers a new
850 attempt from the top using REF-walk.
852 For those cases where ``unlazy_walk()`` is an option, it essentially
853 takes a reference on each of the pointers that it holds (vfsmount,
854 dentry, and possibly some symbolic links) and then verifies that the
855 relevant seqlocks have not been changed. If there have been changes,
856 it, too, aborts with ``-ECHILD``, otherwise the transition to REF-walk
857 has been a success and the lookup process continues.
859 Taking a reference on those pointers is not quite as simple as just
860 incrementing a counter. That works to take a second reference if you
861 already have one (often indirectly through another object), but it
862 isn't sufficient if you don't actually have a counted reference at
863 all. For ``dentry->d_lockref``, it is safe to increment the reference
864 counter to get a reference unless it has been explicitly marked as
865 "dead" which involves setting the counter to ``-128``.
866 ``lockref_get_not_dead()`` achieves this.
868 For ``mnt->mnt_count`` it is safe to take a reference as long as
869 ``mount_lock`` is then used to validate the reference. If that
870 validation fails, it may *not* be safe to just drop that reference in
871 the standard way of calling ``mnt_put()`` - an unmount may have
872 progressed too far. So the code in ``legitimize_mnt()``, when it
873 finds that the reference it got might not be safe, checks the
874 ``MNT_SYNC_UMOUNT`` flag to determine if a simple ``mnt_put()`` is
875 correct, or if it should just decrement the count and pretend none of
878 Taking care in filesystems
879 --------------------------
881 RCU-walk depends almost entirely on cached information and often will
882 not call into the filesystem at all. However there are two places,
883 besides the already-mentioned component-name comparison, where the
884 file system might be included in RCU-walk, and it must know to be
887 If the filesystem has non-standard permission-checking requirements -
888 such as a networked filesystem which may need to check with the server
889 - the ``i_op->permission`` interface might be called during RCU-walk.
890 In this case an extra "``MAY_NOT_BLOCK``" flag is passed so that it
891 knows not to sleep, but to return ``-ECHILD`` if it cannot complete
892 promptly. ``i_op->permission`` is given the inode pointer, not the
893 dentry, so it doesn't need to worry about further consistency checks.
894 However if it accesses any other filesystem data structures, it must
895 ensure they are safe to be accessed with only the ``rcu_read_lock()``
896 held. This typically means they must be freed using ``kfree_rcu()`` or
899 .. _READ_ONCE: https://lwn.net/Articles/624126/
901 If the filesystem may need to revalidate dcache entries, then
902 ``d_op->d_revalidate`` may be called in RCU-walk too. This interface
903 *is* passed the dentry but does not have access to the ``inode`` or the
904 ``seq`` number from the ``nameidata``, so it needs to be extra careful
905 when accessing fields in the dentry. This "extra care" typically
906 involves using `READ_ONCE() <READ_ONCE_>`_ to access fields, and verifying the
907 result is not NULL before using it. This pattern can be seen in
908 ``nfs_lookup_revalidate()``.
913 In various places in the details of REF-walk and RCU-walk, and also in
914 the big picture, there are a couple of related patterns that are worth
917 The first is "try quickly and check, if that fails try slowly". We
918 can see that in the high-level approach of first trying RCU-walk and
919 then trying REF-walk, and in places where ``unlazy_walk()`` is used to
920 switch to REF-walk for the rest of the path. We also saw it earlier
921 in ``dget_parent()`` when following a "``..``" link. It tries a quick way
922 to get a reference, then falls back to taking locks if needed.
924 The second pattern is "try quickly and check, if that fails try
925 again - repeatedly". This is seen with the use of ``rename_lock`` and
926 ``mount_lock`` in REF-walk. RCU-walk doesn't make use of this pattern -
927 if anything goes wrong it is much safer to just abort and try a more
930 The emphasis here is "try quickly and check". It should probably be
931 "try quickly *and carefully*, then check". The fact that checking is
932 needed is a reminder that the system is dynamic and only a limited
933 number of things are safe at all. The most likely cause of errors in
934 this whole process is assuming something is safe when in reality it
935 isn't. Careful consideration of what exactly guarantees the safety of
936 each access is sometimes necessary.
938 A walk among the symlinks
939 =========================
941 There are several basic issues that we will examine to understand the
942 handling of symbolic links: the symlink stack, together with cache
943 lifetimes, will help us understand the overall recursive handling of
944 symlinks and lead to the special care needed for the final component.
945 Then a consideration of access-time updates and summary of the various
946 flags controlling lookup will finish the story.
951 There are only two sorts of filesystem objects that can usefully
952 appear in a path prior to the final component: directories and symlinks.
953 Handling directories is quite straightforward: the new directory
954 simply becomes the starting point at which to interpret the next
955 component on the path. Handling symbolic links requires a bit more
958 Conceptually, symbolic links could be handled by editing the path. If
959 a component name refers to a symbolic link, then that component is
960 replaced by the body of the link and, if that body starts with a '/',
961 then all preceding parts of the path are discarded. This is what the
962 "``readlink -f``" command does, though it also edits out "``.``" and
965 Directly editing the path string is not really necessary when looking
966 up a path, and discarding early components is pointless as they aren't
967 looked at anyway. Keeping track of all remaining components is
968 important, but they can of course be kept separately; there is no need
969 to concatenate them. As one symlink may easily refer to another,
970 which in turn can refer to a third, we may need to keep the remaining
971 components of several paths, each to be processed when the preceding
972 ones are completed. These path remnants are kept on a stack of
975 There are two reasons for placing limits on how many symlinks can
976 occur in a single path lookup. The most obvious is to avoid loops.
977 If a symlink referred to itself either directly or through
978 intermediaries, then following the symlink can never complete
979 successfully - the error ``ELOOP`` must be returned. Loops can be
980 detected without imposing limits, but limits are the simplest solution
981 and, given the second reason for restriction, quite sufficient.
983 .. _outlined recently: http://thread.gmane.org/gmane.linux.kernel/1934390/focus=1934550
985 The second reason was `outlined recently`_ by Linus:
987 Because it's a latency and DoS issue too. We need to react well to
988 true loops, but also to "very deep" non-loops. It's not about memory
989 use, it's about users triggering unreasonable CPU resources.
991 Linux imposes a limit on the length of any pathname: ``PATH_MAX``, which
992 is 4096. There are a number of reasons for this limit; not letting the
993 kernel spend too much time on just one path is one of them. With
994 symbolic links you can effectively generate much longer paths so some
995 sort of limit is needed for the same reason. Linux imposes a limit of
996 at most 40 symlinks in any one path lookup. It previously imposed a
997 further limit of eight on the maximum depth of recursion, but that was
998 raised to 40 when a separate stack was implemented, so there is now
1001 The ``nameidata`` structure that we met in an earlier article contains a
1002 small stack that can be used to store the remaining part of up to two
1003 symlinks. In many cases this will be sufficient. If it isn't, a
1004 separate stack is allocated with room for 40 symlinks. Pathname
1005 lookup will never exceed that stack as, once the 40th symlink is
1006 detected, an error is returned.
1008 It might seem that the name remnants are all that needs to be stored on
1009 this stack, but we need a bit more. To see that, we need to move on to
1012 Storage and lifetime of cached symlinks
1013 ---------------------------------------
1015 Like other filesystem resources, such as inodes and directory
1016 entries, symlinks are cached by Linux to avoid repeated costly access
1017 to external storage. It is particularly important for RCU-walk to be
1018 able to find and temporarily hold onto these cached entries, so that
1019 it doesn't need to drop down into REF-walk.
1021 .. _object-oriented design pattern: https://lwn.net/Articles/446317/
1023 While each filesystem is free to make its own choice, symlinks are
1024 typically stored in one of two places. Short symlinks are often
1025 stored directly in the inode. When a filesystem allocates a ``struct
1026 inode`` it typically allocates extra space to store private data (a
1027 common `object-oriented design pattern`_ in the kernel). This will
1028 sometimes include space for a symlink. The other common location is
1029 in the page cache, which normally stores the content of files. The
1030 pathname in a symlink can be seen as the content of that symlink and
1031 can easily be stored in the page cache just like file content.
1033 When neither of these is suitable, the next most likely scenario is
1034 that the filesystem will allocate some temporary memory and copy or
1035 construct the symlink content into that memory whenever it is needed.
1037 When the symlink is stored in the inode, it has the same lifetime as
1038 the inode which, itself, is protected by RCU or by a counted reference
1039 on the dentry. This means that the mechanisms that pathname lookup
1040 uses to access the dcache and icache (inode cache) safely are quite
1041 sufficient for accessing some cached symlinks safely. In these cases,
1042 the ``i_link`` pointer in the inode is set to point to wherever the
1043 symlink is stored and it can be accessed directly whenever needed.
1045 When the symlink is stored in the page cache or elsewhere, the
1046 situation is not so straightforward. A reference on a dentry or even
1047 on an inode does not imply any reference on cached pages of that
1048 inode, and even an ``rcu_read_lock()`` is not sufficient to ensure that
1049 a page will not disappear. So for these symlinks the pathname lookup
1050 code needs to ask the filesystem to provide a stable reference and,
1051 significantly, needs to release that reference when it is finished
1054 Taking a reference to a cache page is often possible even in RCU-walk
1055 mode. It does require making changes to memory, which is best avoided,
1056 but that isn't necessarily a big cost and it is better than dropping
1057 out of RCU-walk mode completely. Even filesystems that allocate
1058 space to copy the symlink into can use ``GFP_ATOMIC`` to often successfully
1059 allocate memory without the need to drop out of RCU-walk. If a
1060 filesystem cannot successfully get a reference in RCU-walk mode, it
1061 must return ``-ECHILD`` and ``unlazy_walk()`` will be called to return to
1062 REF-walk mode in which the filesystem is allowed to sleep.
1064 The place for all this to happen is the ``i_op->follow_link()`` inode
1065 method. In the present mainline code this is never actually called in
1066 RCU-walk mode as the rewrite is not quite complete. It is likely that
1067 in a future release this method will be passed an ``inode`` pointer when
1068 called in RCU-walk mode so it both (1) knows to be careful, and (2) has the
1069 validated pointer. Much like the ``i_op->permission()`` method we
1070 looked at previously, ``->follow_link()`` would need to be careful that
1071 all the data structures it references are safe to be accessed while
1072 holding no counted reference, only the RCU lock. Though getting a
1073 reference with ``->follow_link()`` is not yet done in RCU-walk mode, the
1074 code is ready to release the reference when that does happen.
1076 This need to drop the reference to a symlink adds significant
1077 complexity. It requires a reference to the inode so that the
1078 ``i_op->put_link()`` inode operation can be called. In REF-walk, that
1079 reference is kept implicitly through a reference to the dentry, so
1080 keeping the ``struct path`` of the symlink is easiest. For RCU-walk,
1081 the pointer to the inode is kept separately. To allow switching from
1082 RCU-walk back to REF-walk in the middle of processing nested symlinks
1083 we also need the seq number for the dentry so we can confirm that
1084 switching back was safe.
1086 Finally, when providing a reference to a symlink, the filesystem also
1087 provides an opaque "cookie" that must be passed to ``->put_link()`` so that it
1088 knows what to free. This might be the allocated memory area, or a
1089 pointer to the ``struct page`` in the page cache, or something else
1090 completely. Only the filesystem knows what it is.
1092 In order for the reference to each symlink to be dropped when the walk completes,
1093 whether in RCU-walk or REF-walk, the symlink stack needs to contain,
1094 along with the path remnants:
1096 - the ``struct path`` to provide a reference to the inode in REF-walk
1097 - the ``struct inode *`` to provide a reference to the inode in RCU-walk
1098 - the ``seq`` to allow the path to be safely switched from RCU-walk to REF-walk
1099 - the ``cookie`` that tells ``->put_path()`` what to put.
1101 This means that each entry in the symlink stack needs to hold five
1102 pointers and an integer instead of just one pointer (the path
1103 remnant). On a 64-bit system, this is about 40 bytes per entry;
1104 with 40 entries it adds up to 1600 bytes total, which is less than
1105 half a page. So it might seem like a lot, but is by no means
1108 Note that, in a given stack frame, the path remnant (``name``) is not
1109 part of the symlink that the other fields refer to. It is the remnant
1110 to be followed once that symlink has been fully parsed.
1112 Following the symlink
1113 ---------------------
1115 The main loop in ``link_path_walk()`` iterates seamlessly over all
1116 components in the path and all of the non-final symlinks. As symlinks
1117 are processed, the ``name`` pointer is adjusted to point to a new
1118 symlink, or is restored from the stack, so that much of the loop
1119 doesn't need to notice. Getting this ``name`` variable on and off the
1120 stack is very straightforward; pushing and popping the references is
1121 a little more complex.
1123 When a symlink is found, ``walk_component()`` returns the value ``1``
1124 (``0`` is returned for any other sort of success, and a negative number
1125 is, as usual, an error indicator). This causes ``get_link()`` to be
1126 called; it then gets the link from the filesystem. Providing that
1127 operation is successful, the old path ``name`` is placed on the stack,
1128 and the new value is used as the ``name`` for a while. When the end of
1129 the path is found (i.e. ``*name`` is ``'\0'``) the old ``name`` is restored
1130 off the stack and path walking continues.
1132 Pushing and popping the reference pointers (inode, cookie, etc.) is more
1133 complex in part because of the desire to handle tail recursion. When
1134 the last component of a symlink itself points to a symlink, we
1135 want to pop the symlink-just-completed off the stack before pushing
1136 the symlink-just-found to avoid leaving empty path remnants that would
1137 just get in the way.
1139 It is most convenient to push the new symlink references onto the
1140 stack in ``walk_component()`` immediately when the symlink is found;
1141 ``walk_component()`` is also the last piece of code that needs to look at the
1142 old symlink as it walks that last component. So it is quite
1143 convenient for ``walk_component()`` to release the old symlink and pop
1144 the references just before pushing the reference information for the
1145 new symlink. It is guided in this by two flags; ``WALK_GET``, which
1146 gives it permission to follow a symlink if it finds one, and
1147 ``WALK_PUT``, which tells it to release the current symlink after it has been
1148 followed. ``WALK_PUT`` is tested first, leading to a call to
1149 ``put_link()``. ``WALK_GET`` is tested subsequently (by
1150 ``should_follow_link()``) leading to a call to ``pick_link()`` which sets
1153 Symlinks with no final component
1154 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1156 A pair of special-case symlinks deserve a little further explanation.
1157 Both result in a new ``struct path`` (with mount and dentry) being set
1158 up in the ``nameidata``, and result in ``get_link()`` returning ``NULL``.
1160 The more obvious case is a symlink to "``/``". All symlinks starting
1161 with "``/``" are detected in ``get_link()`` which resets the ``nameidata``
1162 to point to the effective filesystem root. If the symlink only
1163 contains "``/``" then there is nothing more to do, no components at all,
1164 so ``NULL`` is returned to indicate that the symlink can be released and
1165 the stack frame discarded.
1167 The other case involves things in ``/proc`` that look like symlinks but
1168 aren't really (and are therefore commonly referred to as "magic-links")::
1170 $ ls -l /proc/self/fd/1
1171 lrwx------ 1 neilb neilb 64 Jun 13 10:19 /proc/self/fd/1 -> /dev/pts/4
1173 Every open file descriptor in any process is represented in ``/proc`` by
1174 something that looks like a symlink. It is really a reference to the
1175 target file, not just the name of it. When you ``readlink`` these
1176 objects you get a name that might refer to the same file - unless it
1177 has been unlinked or mounted over. When ``walk_component()`` follows
1178 one of these, the ``->follow_link()`` method in "procfs" doesn't return
1179 a string name, but instead calls ``nd_jump_link()`` which updates the
1180 ``nameidata`` in place to point to that target. ``->follow_link()`` then
1181 returns ``NULL``. Again there is no final component and ``get_link()``
1182 reports this by leaving the ``last_type`` field of ``nameidata`` as
1185 Following the symlink in the final component
1186 --------------------------------------------
1188 All this leads to ``link_path_walk()`` walking down every component, and
1189 following all symbolic links it finds, until it reaches the final
1190 component. This is just returned in the ``last`` field of ``nameidata``.
1191 For some callers, this is all they need; they want to create that
1192 ``last`` name if it doesn't exist or give an error if it does. Other
1193 callers will want to follow a symlink if one is found, and possibly
1194 apply special handling to the last component of that symlink, rather
1195 than just the last component of the original file name. These callers
1196 potentially need to call ``link_path_walk()`` again and again on
1197 successive symlinks until one is found that doesn't point to another
1200 This case is handled by the relevant caller of ``link_path_walk()``, such as
1201 ``path_lookupat()`` using a loop that calls ``link_path_walk()``, and then
1202 handles the final component. If the final component is a symlink
1203 that needs to be followed, then ``trailing_symlink()`` is called to set
1204 things up properly and the loop repeats, calling ``link_path_walk()``
1205 again. This could loop as many as 40 times if the last component of
1206 each symlink is another symlink.
1208 The various functions that examine the final component and possibly
1209 report that it is a symlink are ``lookup_last()``, ``mountpoint_last()``
1210 and ``do_last()``, each of which use the same convention as
1211 ``walk_component()`` of returning ``1`` if a symlink was found that needs
1214 Of these, ``do_last()`` is the most interesting as it is used for
1215 opening a file. Part of ``do_last()`` runs with ``i_rwsem`` held and this
1216 part is in a separate function: ``lookup_open()``.
1218 Explaining ``do_last()`` completely is beyond the scope of this article,
1219 but a few highlights should help those interested in exploring the
1222 1. Rather than just finding the target file, ``do_last()`` needs to open
1223 it. If the file was found in the dcache, then ``vfs_open()`` is used for
1224 this. If not, then ``lookup_open()`` will either call ``atomic_open()`` (if
1225 the filesystem provides it) to combine the final lookup with the open, or
1226 will perform the separate ``lookup_real()`` and ``vfs_create()`` steps
1227 directly. In the later case the actual "open" of this newly found or
1228 created file will be performed by ``vfs_open()``, just as if the name
1229 were found in the dcache.
1231 2. ``vfs_open()`` can fail with ``-EOPENSTALE`` if the cached information
1232 wasn't quite current enough. Rather than restarting the lookup from
1233 the top with ``LOOKUP_REVAL`` set, ``lookup_open()`` is called instead,
1234 giving the filesystem a chance to resolve small inconsistencies.
1235 If that doesn't work, only then is the lookup restarted from the top.
1237 3. An open with O_CREAT **does** follow a symlink in the final component,
1238 unlike other creation system calls (like ``mkdir``). So the sequence::
1241 echo hello > /tmp/foo
1243 will create a file called ``/tmp/bar``. This is not permitted if
1244 ``O_EXCL`` is set but otherwise is handled for an O_CREAT open much
1245 like for a non-creating open: ``should_follow_link()`` returns ``1``, and
1246 so does ``do_last()`` so that ``trailing_symlink()`` gets called and the
1247 open process continues on the symlink that was found.
1249 Updating the access time
1250 ------------------------
1252 We previously said of RCU-walk that it would "take no locks, increment
1253 no counts, leave no footprints." We have since seen that some
1254 "footprints" can be needed when handling symlinks as a counted
1255 reference (or even a memory allocation) may be needed. But these
1256 footprints are best kept to a minimum.
1258 One other place where walking down a symlink can involve leaving
1259 footprints in a way that doesn't affect directories is in updating access times.
1260 In Unix (and Linux) every filesystem object has a "last accessed
1261 time", or "``atime``". Passing through a directory to access a file
1262 within is not considered to be an access for the purposes of
1263 ``atime``; only listing the contents of a directory can update its ``atime``.
1264 Symlinks are different it seems. Both reading a symlink (with ``readlink()``)
1265 and looking up a symlink on the way to some other destination can
1266 update the atime on that symlink.
1268 .. _clearest statement: https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap04.html#tag_04_08
1270 It is not clear why this is the case; POSIX has little to say on the
1271 subject. The `clearest statement`_ is that, if a particular implementation
1272 updates a timestamp in a place not specified by POSIX, this must be
1273 documented "except that any changes caused by pathname resolution need
1274 not be documented". This seems to imply that POSIX doesn't really
1275 care about access-time updates during pathname lookup.
1277 .. _Linux 1.3.87: https://git.kernel.org/cgit/linux/kernel/git/history/history.git/diff/fs/ext2/symlink.c?id=f806c6db77b8eaa6e00dcfb6b567706feae8dbb8
1279 An examination of history shows that prior to `Linux 1.3.87`_, the ext2
1280 filesystem, at least, didn't update atime when following a link.
1281 Unfortunately we have no record of why that behavior was changed.
1283 In any case, access time must now be updated and that operation can be
1284 quite complex. Trying to stay in RCU-walk while doing it is best
1285 avoided. Fortunately it is often permitted to skip the ``atime``
1286 update. Because ``atime`` updates cause performance problems in various
1287 areas, Linux supports the ``relatime`` mount option, which generally
1288 limits the updates of ``atime`` to once per day on files that aren't
1289 being changed (and symlinks never change once created). Even without
1290 ``relatime``, many filesystems record ``atime`` with a one-second
1291 granularity, so only one update per second is required.
1293 It is easy to test if an ``atime`` update is needed while in RCU-walk
1294 mode and, if it isn't, the update can be skipped and RCU-walk mode
1295 continues. Only when an ``atime`` update is actually required does the
1296 path walk drop down to REF-walk. All of this is handled in the
1297 ``get_link()`` function.
1302 A suitable way to wrap up this tour of pathname walking is to list
1303 the various flags that can be stored in the ``nameidata`` to guide the
1304 lookup process. Many of these are only meaningful on the final
1305 component, others reflect the current state of the pathname lookup, and some
1306 apply restrictions to all path components encountered in the path lookup.
1308 And then there is ``LOOKUP_EMPTY``, which doesn't fit conceptually with
1309 the others. If this is not set, an empty pathname causes an error
1310 very early on. If it is set, empty pathnames are not considered to be
1316 We have already met two global state flags: ``LOOKUP_RCU`` and
1317 ``LOOKUP_REVAL``. These select between one of three overall approaches
1318 to lookup: RCU-walk, REF-walk, and REF-walk with forced revalidation.
1320 ``LOOKUP_PARENT`` indicates that the final component hasn't been reached
1321 yet. This is primarily used to tell the audit subsystem the full
1322 context of a particular access being audited.
1324 ``LOOKUP_ROOT`` indicates that the ``root`` field in the ``nameidata`` was
1325 provided by the caller, so it shouldn't be released when it is no
1328 ``LOOKUP_JUMPED`` means that the current dentry was chosen not because
1329 it had the right name but for some other reason. This happens when
1330 following "``..``", following a symlink to ``/``, crossing a mount point
1331 or accessing a "``/proc/$PID/fd/$FD``" symlink (also known as a "magic
1332 link"). In this case the filesystem has not been asked to revalidate the
1333 name (with ``d_revalidate()``). In such cases the inode may still need
1334 to be revalidated, so ``d_op->d_weak_revalidate()`` is called if
1335 ``LOOKUP_JUMPED`` is set when the look completes - which may be at the
1336 final component or, when creating, unlinking, or renaming, at the penultimate component.
1338 Resolution-restriction flags
1339 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1341 In order to allow userspace to protect itself against certain race conditions
1342 and attack scenarios involving changing path components, a series of flags are
1343 available which apply restrictions to all path components encountered during
1344 path lookup. These flags are exposed through ``openat2()``'s ``resolve`` field.
1346 ``LOOKUP_NO_SYMLINKS`` blocks all symlink traversals (including magic-links).
1347 This is distinctly different from ``LOOKUP_FOLLOW``, because the latter only
1348 relates to restricting the following of trailing symlinks.
1350 ``LOOKUP_NO_MAGICLINKS`` blocks all magic-link traversals. Filesystems must
1351 ensure that they return errors from ``nd_jump_link()``, because that is how
1352 ``LOOKUP_NO_MAGICLINKS`` and other magic-link restrictions are implemented.
1354 ``LOOKUP_NO_XDEV`` blocks all ``vfsmount`` traversals (this includes both
1355 bind-mounts and ordinary mounts). Note that the ``vfsmount`` which contains the
1356 lookup is determined by the first mountpoint the path lookup reaches --
1357 absolute paths start with the ``vfsmount`` of ``/``, and relative paths start
1358 with the ``dfd``'s ``vfsmount``. Magic-links are only permitted if the
1359 ``vfsmount`` of the path is unchanged.
1361 ``LOOKUP_BENEATH`` blocks any path components which resolve outside the
1362 starting point of the resolution. This is done by blocking ``nd_jump_root()``
1363 as well as blocking ".." if it would jump outside the starting point.
1364 ``rename_lock`` and ``mount_lock`` are used to detect attacks against the
1365 resolution of "..". Magic-links are also blocked.
1367 ``LOOKUP_IN_ROOT`` resolves all path components as though the starting point
1368 were the filesystem root. ``nd_jump_root()`` brings the resolution back to
1369 the starting point, and ".." at the starting point will act as a no-op. As with
1370 ``LOOKUP_BENEATH``, ``rename_lock`` and ``mount_lock`` are used to detect
1371 attacks against ".." resolution. Magic-links are also blocked.
1373 Final-component flags
1374 ~~~~~~~~~~~~~~~~~~~~~
1376 Some of these flags are only set when the final component is being
1377 considered. Others are only checked for when considering that final
1380 ``LOOKUP_AUTOMOUNT`` ensures that, if the final component is an automount
1381 point, then the mount is triggered. Some operations would trigger it
1382 anyway, but operations like ``stat()`` deliberately don't. ``statfs()``
1383 needs to trigger the mount but otherwise behaves a lot like ``stat()``, so
1384 it sets ``LOOKUP_AUTOMOUNT``, as does "``quotactl()``" and the handling of
1387 ``LOOKUP_FOLLOW`` has a similar function to ``LOOKUP_AUTOMOUNT`` but for
1388 symlinks. Some system calls set or clear it implicitly, while
1389 others have API flags such as ``AT_SYMLINK_FOLLOW`` and
1390 ``UMOUNT_NOFOLLOW`` to control it. Its effect is similar to
1391 ``WALK_GET`` that we already met, but it is used in a different way.
1393 ``LOOKUP_DIRECTORY`` insists that the final component is a directory.
1394 Various callers set this and it is also set when the final component
1395 is found to be followed by a slash.
1397 Finally ``LOOKUP_OPEN``, ``LOOKUP_CREATE``, ``LOOKUP_EXCL``, and
1398 ``LOOKUP_RENAME_TARGET`` are not used directly by the VFS but are made
1399 available to the filesystem and particularly the ``->d_revalidate()``
1400 method. A filesystem can choose not to bother revalidating too hard
1401 if it knows that it will be asked to open or create the file soon.
1402 These flags were previously useful for ``->lookup()`` too but with the
1403 introduction of ``->atomic_open()`` they are less relevant there.
1408 Despite its complexity, all this pathname lookup code appears to be
1409 in good shape - various parts are certainly easier to understand now
1410 than even a couple of releases ago. But that doesn't mean it is
1411 "finished". As already mentioned, RCU-walk currently only follows
1412 symlinks that are stored in the inode so, while it handles many ext4
1413 symlinks, it doesn't help with NFS, XFS, or Btrfs. That support
1414 is not likely to be long delayed.
projects / linux-2.6-microblaze.git / blob