1 =============== 1 ===============
2 Pathname lookup 2 Pathname lookup
3 =============== 3 ===============
4 4
5 This write-up is based on three articles publi 5 This write-up is based on three articles published at lwn.net:
6 6
7 - <https://lwn.net/Articles/649115/> Pathname 7 - <https://lwn.net/Articles/649115/> Pathname lookup in Linux
8 - <https://lwn.net/Articles/649729/> RCU-walk: 8 - <https://lwn.net/Articles/649729/> RCU-walk: faster pathname lookup in Linux
9 - <https://lwn.net/Articles/650786/> A walk am 9 - <https://lwn.net/Articles/650786/> A walk among the symlinks
10 10
11 Written by Neil Brown with help from Al Viro a 11 Written by Neil Brown with help from Al Viro and Jon Corbet.
12 It has subsequently been updated to reflect ch 12 It has subsequently been updated to reflect changes in the kernel
13 including: 13 including:
14 14
15 - per-directory parallel name lookup. 15 - per-directory parallel name lookup.
16 - ``openat2()`` resolution restriction flags. <<
17 16
18 Introduction to pathname lookup 17 Introduction to pathname lookup
19 =============================== 18 ===============================
20 19
21 The most obvious aspect of pathname lookup, wh 20 The most obvious aspect of pathname lookup, which very little
22 exploration is needed to discover, is that it 21 exploration is needed to discover, is that it is complex. There are
23 many rules, special cases, and implementation 22 many rules, special cases, and implementation alternatives that all
24 combine to confuse the unwary reader. Compute 23 combine to confuse the unwary reader. Computer science has long been
25 acquainted with such complexity and has tools 24 acquainted with such complexity and has tools to help manage it. One
26 tool that we will make extensive use of is "di 25 tool that we will make extensive use of is "divide and conquer". For
27 the early parts of the analysis we will divide 26 the early parts of the analysis we will divide off symlinks - leaving
28 them until the final part. Well before we get 27 them until the final part. Well before we get to symlinks we have
29 another major division based on the VFS's appr 28 another major division based on the VFS's approach to locking which
30 will allow us to review "REF-walk" and "RCU-wa 29 will allow us to review "REF-walk" and "RCU-walk" separately. But we
31 are getting ahead of ourselves. There are som 30 are getting ahead of ourselves. There are some important low level
32 distinctions we need to clarify first. 31 distinctions we need to clarify first.
33 32
34 There are two sorts of ... 33 There are two sorts of ...
35 -------------------------- 34 --------------------------
36 35
37 .. _openat: http://man7.org/linux/man-pages/ma 36 .. _openat: http://man7.org/linux/man-pages/man2/openat.2.html
38 37
39 Pathnames (sometimes "file names"), used to id 38 Pathnames (sometimes "file names"), used to identify objects in the
40 filesystem, will be familiar to most readers. 39 filesystem, will be familiar to most readers. They contain two sorts
41 of elements: "slashes" that are sequences of o 40 of elements: "slashes" that are sequences of one or more "``/``"
42 characters, and "components" that are sequence 41 characters, and "components" that are sequences of one or more
43 non-"``/``" characters. These form two kinds 42 non-"``/``" characters. These form two kinds of paths. Those that
44 start with slashes are "absolute" and start fr 43 start with slashes are "absolute" and start from the filesystem root.
45 The others are "relative" and start from the c 44 The others are "relative" and start from the current directory, or
46 from some other location specified by a file d !! 45 from some other location specified by a file descriptor given to a
47 "``*at()``" system calls such as `openat() <op !! 46 "``XXXat``" system call such as `openat() <openat_>`_.
48 47
49 .. _execveat: http://man7.org/linux/man-pages/ 48 .. _execveat: http://man7.org/linux/man-pages/man2/execveat.2.html
50 49
51 It is tempting to describe the second kind as 50 It is tempting to describe the second kind as starting with a
52 component, but that isn't always accurate: a p 51 component, but that isn't always accurate: a pathname can lack both
53 slashes and components, it can be empty, in ot 52 slashes and components, it can be empty, in other words. This is
54 generally forbidden in POSIX, but some of thos !! 53 generally forbidden in POSIX, but some of those "xxx``at``" system calls
55 in Linux permit it when the ``AT_EMPTY_PATH`` 54 in Linux permit it when the ``AT_EMPTY_PATH`` flag is given. For
56 example, if you have an open file descriptor o 55 example, if you have an open file descriptor on an executable file you
57 can execute it by calling `execveat() <execvea 56 can execute it by calling `execveat() <execveat_>`_ passing
58 the file descriptor, an empty path, and the `` 57 the file descriptor, an empty path, and the ``AT_EMPTY_PATH`` flag.
59 58
60 These paths can be divided into two sections: 59 These paths can be divided into two sections: the final component and
61 everything else. The "everything else" is the 60 everything else. The "everything else" is the easy bit. In all cases
62 it must identify a directory that already exis 61 it must identify a directory that already exists, otherwise an error
63 such as ``ENOENT`` or ``ENOTDIR`` will be repo 62 such as ``ENOENT`` or ``ENOTDIR`` will be reported.
64 63
65 The final component is not so simple. Not onl 64 The final component is not so simple. Not only do different system
66 calls interpret it quite differently (e.g. som 65 calls interpret it quite differently (e.g. some create it, some do
67 not), but it might not even exist: neither the 66 not), but it might not even exist: neither the empty pathname nor the
68 pathname that is just slashes have a final com 67 pathname that is just slashes have a final component. If it does
69 exist, it could be "``.``" or "``..``" which a 68 exist, it could be "``.``" or "``..``" which are handled quite differently
70 from other components. 69 from other components.
71 70
72 .. _POSIX: https://pubs.opengroup.org/onlinepu !! 71 .. _POSIX: http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap04.html#tag_04_12
73 72
74 If a pathname ends with a slash, such as "``/t 73 If a pathname ends with a slash, such as "``/tmp/foo/``" it might be
75 tempting to consider that to have an empty fin 74 tempting to consider that to have an empty final component. In many
76 ways that would lead to correct results, but n 75 ways that would lead to correct results, but not always. In
77 particular, ``mkdir()`` and ``rmdir()`` each c 76 particular, ``mkdir()`` and ``rmdir()`` each create or remove a directory named
78 by the final component, and they are required 77 by the final component, and they are required to work with pathnames
79 ending in "``/``". According to POSIX_: !! 78 ending in "``/``". According to POSIX_
80 79
81 A pathname that contains at least one non-<s !! 80 A pathname that contains at least one non- <slash> character and
82 that ends with one or more trailing <slash> !! 81 that ends with one or more trailing <slash> characters shall not
83 be resolved successfully unless the last pat 82 be resolved successfully unless the last pathname component before
84 the trailing <slash> characters names an exi 83 the trailing <slash> characters names an existing directory or a
85 directory entry that is to be created for a 84 directory entry that is to be created for a directory immediately
86 after the pathname is resolved. 85 after the pathname is resolved.
87 86
88 The Linux pathname walking code (mostly in ``f 87 The Linux pathname walking code (mostly in ``fs/namei.c``) deals with
89 all of these issues: breaking the path into co 88 all of these issues: breaking the path into components, handling the
90 "everything else" quite separately from the fi 89 "everything else" quite separately from the final component, and
91 checking that the trailing slash is not used w 90 checking that the trailing slash is not used where it isn't
92 permitted. It also addresses the important is 91 permitted. It also addresses the important issue of concurrent
93 access. 92 access.
94 93
95 While one process is looking up a pathname, an 94 While one process is looking up a pathname, another might be making
96 changes that affect that lookup. One fairly e 95 changes that affect that lookup. One fairly extreme case is that if
97 "a/b" were renamed to "a/c/b" while another pr 96 "a/b" were renamed to "a/c/b" while another process were looking up
98 "a/b/..", that process might successfully reso 97 "a/b/..", that process might successfully resolve on "a/c".
99 Most races are much more subtle, and a big par 98 Most races are much more subtle, and a big part of the task of
100 pathname lookup is to prevent them from having 99 pathname lookup is to prevent them from having damaging effects. Many
101 of the possible races are seen most clearly in 100 of the possible races are seen most clearly in the context of the
102 "dcache" and an understanding of that is centr 101 "dcache" and an understanding of that is central to understanding
103 pathname lookup. 102 pathname lookup.
104 103
105 More than just a cache 104 More than just a cache
106 ---------------------- 105 ----------------------
107 106
108 The "dcache" caches information about names in 107 The "dcache" caches information about names in each filesystem to
109 make them quickly available for lookup. Each 108 make them quickly available for lookup. Each entry (known as a
110 "dentry") contains three significant fields: a 109 "dentry") contains three significant fields: a component name, a
111 pointer to a parent dentry, and a pointer to t 110 pointer to a parent dentry, and a pointer to the "inode" which
112 contains further information about the object 111 contains further information about the object in that parent with
113 the given name. The inode pointer can be ``NU 112 the given name. The inode pointer can be ``NULL`` indicating that the
114 name doesn't exist in the parent. While there 113 name doesn't exist in the parent. While there can be linkage in the
115 dentry of a directory to the dentries of the c 114 dentry of a directory to the dentries of the children, that linkage is
116 not used for pathname lookup, and so will not 115 not used for pathname lookup, and so will not be considered here.
117 116
118 The dcache has a number of uses apart from acc 117 The dcache has a number of uses apart from accelerating lookup. One
119 that will be particularly relevant is that it 118 that will be particularly relevant is that it is closely integrated
120 with the mount table that records which filesy 119 with the mount table that records which filesystem is mounted where.
121 What the mount table actually stores is which 120 What the mount table actually stores is which dentry is mounted on top
122 of which other dentry. 121 of which other dentry.
123 122
124 When considering the dcache, we have another o 123 When considering the dcache, we have another of our "two types"
125 distinctions: there are two types of filesyste 124 distinctions: there are two types of filesystems.
126 125
127 Some filesystems ensure that the information i 126 Some filesystems ensure that the information in the dcache is always
128 completely accurate (though not necessarily co 127 completely accurate (though not necessarily complete). This can allow
129 the VFS to determine if a particular file does 128 the VFS to determine if a particular file does or doesn't exist
130 without checking with the filesystem, and mean 129 without checking with the filesystem, and means that the VFS can
131 protect the filesystem against certain races a 130 protect the filesystem against certain races and other problems.
132 These are typically "local" filesystems such a 131 These are typically "local" filesystems such as ext3, XFS, and Btrfs.
133 132
134 Other filesystems don't provide that guarantee 133 Other filesystems don't provide that guarantee because they cannot.
135 These are typically filesystems that are share 134 These are typically filesystems that are shared across a network,
136 whether remote filesystems like NFS and 9P, or 135 whether remote filesystems like NFS and 9P, or cluster filesystems
137 like ocfs2 or cephfs. These filesystems allow 136 like ocfs2 or cephfs. These filesystems allow the VFS to revalidate
138 cached information, and must provide their own 137 cached information, and must provide their own protection against
139 awkward races. The VFS can detect these files 138 awkward races. The VFS can detect these filesystems by the
140 ``DCACHE_OP_REVALIDATE`` flag being set in the 139 ``DCACHE_OP_REVALIDATE`` flag being set in the dentry.
141 140
142 REF-walk: simple concurrency management with r 141 REF-walk: simple concurrency management with refcounts and spinlocks
143 ---------------------------------------------- 142 --------------------------------------------------------------------
144 143
145 With all of those divisions carefully classifi 144 With all of those divisions carefully classified, we can now start
146 looking at the actual process of walking along 145 looking at the actual process of walking along a path. In particular
147 we will start with the handling of the "everyt 146 we will start with the handling of the "everything else" part of a
148 pathname, and focus on the "REF-walk" approach 147 pathname, and focus on the "REF-walk" approach to concurrency
149 management. This code is found in the ``link_ 148 management. This code is found in the ``link_path_walk()`` function, if
150 you ignore all the places that only run when " 149 you ignore all the places that only run when "``LOOKUP_RCU``"
151 (indicating the use of RCU-walk) is set. 150 (indicating the use of RCU-walk) is set.
152 151
153 .. _Meet the Lockers: https://lwn.net/Articles 152 .. _Meet the Lockers: https://lwn.net/Articles/453685/
154 153
155 REF-walk is fairly heavy-handed with locks and 154 REF-walk is fairly heavy-handed with locks and reference counts. Not
156 as heavy-handed as in the old "big kernel lock 155 as heavy-handed as in the old "big kernel lock" days, but certainly not
157 afraid of taking a lock when one is needed. I 156 afraid of taking a lock when one is needed. It uses a variety of
158 different concurrency controls. A background 157 different concurrency controls. A background understanding of the
159 various primitives is assumed, or can be glean 158 various primitives is assumed, or can be gleaned from elsewhere such
160 as in `Meet the Lockers`_. 159 as in `Meet the Lockers`_.
161 160
162 The locking mechanisms used by REF-walk includ 161 The locking mechanisms used by REF-walk include:
163 162
164 dentry->d_lockref 163 dentry->d_lockref
165 ~~~~~~~~~~~~~~~~~ 164 ~~~~~~~~~~~~~~~~~
166 165
167 This uses the lockref primitive to provide bot 166 This uses the lockref primitive to provide both a spinlock and a
168 reference count. The special-sauce of this pr 167 reference count. The special-sauce of this primitive is that the
169 conceptual sequence "lock; inc_ref; unlock;" c 168 conceptual sequence "lock; inc_ref; unlock;" can often be performed
170 with a single atomic memory operation. 169 with a single atomic memory operation.
171 170
172 Holding a reference on a dentry ensures that t 171 Holding a reference on a dentry ensures that the dentry won't suddenly
173 be freed and used for something else, so the v 172 be freed and used for something else, so the values in various fields
174 will behave as expected. It also protects the 173 will behave as expected. It also protects the ``->d_inode`` reference
175 to the inode to some extent. 174 to the inode to some extent.
176 175
177 The association between a dentry and its inode 176 The association between a dentry and its inode is fairly permanent.
178 For example, when a file is renamed, the dentr 177 For example, when a file is renamed, the dentry and inode move
179 together to the new location. When a file is 178 together to the new location. When a file is created the dentry will
180 initially be negative (i.e. ``d_inode`` is ``N 179 initially be negative (i.e. ``d_inode`` is ``NULL``), and will be assigned
181 to the new inode as part of the act of creatio 180 to the new inode as part of the act of creation.
182 181
183 When a file is deleted, this can be reflected 182 When a file is deleted, this can be reflected in the cache either by
184 setting ``d_inode`` to ``NULL``, or by removin 183 setting ``d_inode`` to ``NULL``, or by removing it from the hash table
185 (described shortly) used to look up the name i 184 (described shortly) used to look up the name in the parent directory.
186 If the dentry is still in use the second optio 185 If the dentry is still in use the second option is used as it is
187 perfectly legal to keep using an open file aft 186 perfectly legal to keep using an open file after it has been deleted
188 and having the dentry around helps. If the de 187 and having the dentry around helps. If the dentry is not otherwise in
189 use (i.e. if the refcount in ``d_lockref`` is 188 use (i.e. if the refcount in ``d_lockref`` is one), only then will
190 ``d_inode`` be set to ``NULL``. Doing it this 189 ``d_inode`` be set to ``NULL``. Doing it this way is more efficient for a
191 very common case. 190 very common case.
192 191
193 So as long as a counted reference is held to a 192 So as long as a counted reference is held to a dentry, a non-``NULL`` ``->d_inode``
194 value will never be changed. 193 value will never be changed.
195 194
196 dentry->d_lock 195 dentry->d_lock
197 ~~~~~~~~~~~~~~ 196 ~~~~~~~~~~~~~~
198 197
199 ``d_lock`` is a synonym for the spinlock that 198 ``d_lock`` is a synonym for the spinlock that is part of ``d_lockref`` above.
200 For our purposes, holding this lock protects a 199 For our purposes, holding this lock protects against the dentry being
201 renamed or unlinked. In particular, its paren 200 renamed or unlinked. In particular, its parent (``d_parent``), and its
202 name (``d_name``) cannot be changed, and it ca 201 name (``d_name``) cannot be changed, and it cannot be removed from the
203 dentry hash table. 202 dentry hash table.
204 203
205 When looking for a name in a directory, REF-wa 204 When looking for a name in a directory, REF-walk takes ``d_lock`` on
206 each candidate dentry that it finds in the has 205 each candidate dentry that it finds in the hash table and then checks
207 that the parent and name are correct. So it d 206 that the parent and name are correct. So it doesn't lock the parent
208 while searching in the cache; it only locks ch 207 while searching in the cache; it only locks children.
209 208
210 When looking for the parent for a given name ( 209 When looking for the parent for a given name (to handle "``..``"),
211 REF-walk can take ``d_lock`` to get a stable r 210 REF-walk can take ``d_lock`` to get a stable reference to ``d_parent``,
212 but it first tries a more lightweight approach 211 but it first tries a more lightweight approach. As seen in
213 ``dget_parent()``, if a reference can be claim 212 ``dget_parent()``, if a reference can be claimed on the parent, and if
214 subsequently ``d_parent`` can be seen to have 213 subsequently ``d_parent`` can be seen to have not changed, then there is
215 no need to actually take the lock on the child 214 no need to actually take the lock on the child.
216 215
217 rename_lock 216 rename_lock
218 ~~~~~~~~~~~ 217 ~~~~~~~~~~~
219 218
220 Looking up a given name in a given directory i 219 Looking up a given name in a given directory involves computing a hash
221 from the two values (the name and the dentry o 220 from the two values (the name and the dentry of the directory),
222 accessing that slot in a hash table, and searc 221 accessing that slot in a hash table, and searching the linked list
223 that is found there. 222 that is found there.
224 223
225 When a dentry is renamed, the name and the par 224 When a dentry is renamed, the name and the parent dentry can both
226 change so the hash will almost certainly chang 225 change so the hash will almost certainly change too. This would move the
227 dentry to a different chain in the hash table. 226 dentry to a different chain in the hash table. If a filename search
228 happened to be looking at a dentry that was mo 227 happened to be looking at a dentry that was moved in this way,
229 it might end up continuing the search down the 228 it might end up continuing the search down the wrong chain,
230 and so miss out on part of the correct chain. 229 and so miss out on part of the correct chain.
231 230
232 The name-lookup process (``d_lookup()``) does !! 231 The name-lookup process (``d_lookup()``) does _not_ try to prevent this
233 from happening, but only to detect when it hap 232 from happening, but only to detect when it happens.
234 ``rename_lock`` is a seqlock that is updated w 233 ``rename_lock`` is a seqlock that is updated whenever any dentry is
235 renamed. If ``d_lookup`` finds that a rename 234 renamed. If ``d_lookup`` finds that a rename happened while it
236 unsuccessfully scanned a chain in the hash tab 235 unsuccessfully scanned a chain in the hash table, it simply tries
237 again. 236 again.
238 237
239 ``rename_lock`` is also used to detect and def <<
240 against ``LOOKUP_BENEATH`` and ``LOOKUP_IN_ROO <<
241 the parent directory is moved outside the root <<
242 check). If ``rename_lock`` is updated during t <<
243 a "..", a potential attack occurred and ``hand <<
244 ``-EAGAIN``. <<
245 <<
246 inode->i_rwsem 238 inode->i_rwsem
247 ~~~~~~~~~~~~~~ 239 ~~~~~~~~~~~~~~
248 240
249 ``i_rwsem`` is a read/write semaphore that ser 241 ``i_rwsem`` is a read/write semaphore that serializes all changes to a particular
250 directory. This ensures that, for example, an 242 directory. This ensures that, for example, an ``unlink()`` and a ``rename()``
251 cannot both happen at the same time. It also 243 cannot both happen at the same time. It also keeps the directory
252 stable while the filesystem is asked to look u 244 stable while the filesystem is asked to look up a name that is not
253 currently in the dcache or, optionally, when t 245 currently in the dcache or, optionally, when the list of entries in a
254 directory is being retrieved with ``readdir()` 246 directory is being retrieved with ``readdir()``.
255 247
256 This has a complementary role to that of ``d_l 248 This has a complementary role to that of ``d_lock``: ``i_rwsem`` on a
257 directory protects all of the names in that di 249 directory protects all of the names in that directory, while ``d_lock``
258 on a name protects just one name in a director 250 on a name protects just one name in a directory. Most changes to the
259 dcache hold ``i_rwsem`` on the relevant direct 251 dcache hold ``i_rwsem`` on the relevant directory inode and briefly take
260 ``d_lock`` on one or more the dentries while t 252 ``d_lock`` on one or more the dentries while the change happens. One
261 exception is when idle dentries are removed fr 253 exception is when idle dentries are removed from the dcache due to
262 memory pressure. This uses ``d_lock``, but `` 254 memory pressure. This uses ``d_lock``, but ``i_rwsem`` plays no role.
263 255
264 The semaphore affects pathname lookup in two d 256 The semaphore affects pathname lookup in two distinct ways. Firstly it
265 prevents changes during lookup of a name in a 257 prevents changes during lookup of a name in a directory. ``walk_component()`` uses
266 ``lookup_fast()`` first which, in turn, checks 258 ``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 is 259 using only ``d_lock`` locking. If the name isn't found, then ``walk_component()``
268 falls back to ``lookup_slow()`` which takes a 260 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 261 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 262 definitive answer. A new dentry will be added to the cache regardless of
271 the result. 263 the result.
272 264
273 Secondly, when pathname lookup reaches the fin 265 Secondly, when pathname lookup reaches the final component, it will
274 sometimes need to take an exclusive lock on `` 266 sometimes need to take an exclusive lock on ``i_rwsem`` before performing the last lookup so
275 that the required exclusion can be achieved. 267 that the required exclusion can be achieved. How path lookup chooses
276 to take, or not take, ``i_rwsem`` is one of th 268 to take, or not take, ``i_rwsem`` is one of the
277 issues addressed in a subsequent section. 269 issues addressed in a subsequent section.
278 270
279 If two threads attempt to look up the same nam 271 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 share 272 name that is not yet in the dcache - the shared lock on ``i_rwsem`` will
281 not prevent them both adding new dentries with 273 not prevent them both adding new dentries with the same name. As this
282 would result in confusion an extra level of in 274 would result in confusion an extra level of interlocking is used,
283 based around a secondary hash table (``in_look 275 based around a secondary hash table (``in_lookup_hashtable``) and a
284 per-dentry flag bit (``DCACHE_PAR_LOOKUP``). 276 per-dentry flag bit (``DCACHE_PAR_LOOKUP``).
285 277
286 To add a new dentry to the cache while only ho 278 To add a new dentry to the cache while only holding a shared lock on
287 ``i_rwsem``, a thread must call ``d_alloc_para 279 ``i_rwsem``, a thread must call ``d_alloc_parallel()``. This allocates a
288 dentry, stores the required name and parent in 280 dentry, stores the required name and parent in it, checks if there
289 is already a matching dentry in the primary or 281 is already a matching dentry in the primary or secondary hash
290 tables, and if not, stores the newly allocated 282 tables, and if not, stores the newly allocated dentry in the secondary
291 hash table, with ``DCACHE_PAR_LOOKUP`` set. 283 hash table, with ``DCACHE_PAR_LOOKUP`` set.
292 284
293 If a matching dentry was found in the primary 285 If a matching dentry was found in the primary hash table then that is
294 returned and the caller can know that it lost 286 returned and the caller can know that it lost a race with some other
295 thread adding the entry. If no matching dentr 287 thread adding the entry. If no matching dentry is found in either
296 cache, the newly allocated dentry is returned 288 cache, the newly allocated dentry is returned and the caller can
297 detect this from the presence of ``DCACHE_PAR_ 289 detect this from the presence of ``DCACHE_PAR_LOOKUP``. In this case it
298 knows that it has won any race and now is resp 290 knows that it has won any race and now is responsible for asking the
299 filesystem to perform the lookup and find the 291 filesystem to perform the lookup and find the matching inode. When
300 the lookup is complete, it must call ``d_looku 292 the lookup is complete, it must call ``d_lookup_done()`` which clears
301 the flag and does some other house keeping, in 293 the flag and does some other house keeping, including removing the
302 dentry from the secondary hash table - it will 294 dentry from the secondary hash table - it will normally have been
303 added to the primary hash table already. Note 295 added to the primary hash table already. Note that a ``struct
304 waitqueue_head`` is passed to ``d_alloc_parall 296 waitqueue_head`` is passed to ``d_alloc_parallel()``, and
305 ``d_lookup_done()`` must be called while this 297 ``d_lookup_done()`` must be called while this ``waitqueue_head`` is still
306 in scope. 298 in scope.
307 299
308 If a matching dentry is found in the secondary 300 If a matching dentry is found in the secondary hash table,
309 ``d_alloc_parallel()`` has a little more work 301 ``d_alloc_parallel()`` has a little more work to do. It first waits for
310 ``DCACHE_PAR_LOOKUP`` to be cleared, using a w 302 ``DCACHE_PAR_LOOKUP`` to be cleared, using a wait_queue that was passed
311 to the instance of ``d_alloc_parallel()`` that 303 to the instance of ``d_alloc_parallel()`` that won the race and that
312 will be woken by the call to ``d_lookup_done() 304 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 primar 305 if the dentry has now been added to the primary hash table. If it
314 has, the dentry is returned and the caller jus 306 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 307 race. If it hasn't been added to the primary hash table, the most
316 likely explanation is that some other dentry w 308 likely explanation is that some other dentry was added instead using
317 ``d_splice_alias()``. In any case, ``d_alloc_ 309 ``d_splice_alias()``. In any case, ``d_alloc_parallel()`` repeats all the
318 look ups from the start and will normally retu 310 look ups from the start and will normally return something from the
319 primary hash table. 311 primary hash table.
320 312
321 mnt->mnt_count 313 mnt->mnt_count
322 ~~~~~~~~~~~~~~ 314 ~~~~~~~~~~~~~~
323 315
324 ``mnt_count`` is a per-CPU reference counter o 316 ``mnt_count`` is a per-CPU reference counter on "``mount``" structures.
325 Per-CPU here means that incrementing the count 317 Per-CPU here means that incrementing the count is cheap as it only
326 uses CPU-local memory, but checking if the cou 318 uses CPU-local memory, but checking if the count is zero is expensive as
327 it needs to check with every CPU. Taking a `` 319 it needs to check with every CPU. Taking a ``mnt_count`` reference
328 prevents the mount structure from disappearing 320 prevents the mount structure from disappearing as the result of regular
329 unmount operations, but does not prevent a "la 321 unmount operations, but does not prevent a "lazy" unmount. So holding
330 ``mnt_count`` doesn't ensure that the mount re 322 ``mnt_count`` doesn't ensure that the mount remains in the namespace and,
331 in particular, doesn't stabilize the link to t 323 in particular, doesn't stabilize the link to the mounted-on dentry. It
332 does, however, ensure that the ``mount`` data 324 does, however, ensure that the ``mount`` data structure remains coherent,
333 and it provides a reference to the root dentry 325 and it provides a reference to the root dentry of the mounted
334 filesystem. So a reference through ``->mnt_co 326 filesystem. So a reference through ``->mnt_count`` provides a stable
335 reference to the mounted dentry, but not the m 327 reference to the mounted dentry, but not the mounted-on dentry.
336 328
337 mount_lock 329 mount_lock
338 ~~~~~~~~~~ 330 ~~~~~~~~~~
339 331
340 ``mount_lock`` is a global seqlock, a bit like 332 ``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 333 check if any change has been made to any mount points.
342 334
343 While walking down the tree (away from the roo 335 While walking down the tree (away from the root) this lock is used when
344 crossing a mount point to check that the cross 336 crossing a mount point to check that the crossing was safe. That is,
345 the value in the seqlock is read, then the cod 337 the value in the seqlock is read, then the code finds the mount that
346 is mounted on the current directory, if there 338 is mounted on the current directory, if there is one, and increments
347 the ``mnt_count``. Finally the value in ``mou 339 the ``mnt_count``. Finally the value in ``mount_lock`` is checked against
348 the old value. If there is no change, then th 340 the old value. If there is no change, then the crossing was safe. If there
349 was a change, the ``mnt_count`` is decremented 341 was a change, the ``mnt_count`` is decremented and the whole process is
350 retried. 342 retried.
351 343
352 When walking up the tree (towards the root) by 344 When walking up the tree (towards the root) by following a ".." link,
353 a little more care is needed. In this case th 345 a little more care is needed. In this case the seqlock (which
354 contains both a counter and a spinlock) is ful 346 contains both a counter and a spinlock) is fully locked to prevent
355 any changes to any mount points while stepping 347 any changes to any mount points while stepping up. This locking is
356 needed to stabilize the link to the mounted-on 348 needed to stabilize the link to the mounted-on dentry, which the
357 refcount on the mount itself doesn't ensure. 349 refcount on the mount itself doesn't ensure.
358 350
359 ``mount_lock`` is also used to detect and defe <<
360 against ``LOOKUP_BENEATH`` and ``LOOKUP_IN_ROO <<
361 the parent directory is moved outside the root <<
362 check). If ``mount_lock`` is updated during th <<
363 a "..", a potential attack occurred and ``hand <<
364 ``-EAGAIN``. <<
365 <<
366 RCU 351 RCU
367 ~~~ 352 ~~~
368 353
369 Finally the global (but extremely lightweight) 354 Finally the global (but extremely lightweight) RCU read lock is held
370 from time to time to ensure certain data struc 355 from time to time to ensure certain data structures don't get freed
371 unexpectedly. 356 unexpectedly.
372 357
373 In particular it is held while scanning chains 358 In particular it is held while scanning chains in the dcache hash
374 table, and the mount point hash table. 359 table, and the mount point hash table.
375 360
376 Bringing it together with ``struct nameidata`` 361 Bringing it together with ``struct nameidata``
377 ---------------------------------------------- 362 ----------------------------------------------
378 363
379 .. _First edition Unix: https://minnie.tuhs.or !! 364 .. _First edition Unix: http://minnie.tuhs.org/cgi-bin/utree.pl?file=V1/u2.s
380 365
381 Throughout the process of walking a path, the 366 Throughout the process of walking a path, the current status is stored
382 in a ``struct nameidata``, "namei" being the t 367 in a ``struct nameidata``, "namei" being the traditional name - dating
383 all the way back to `First Edition Unix`_ - of 368 all the way back to `First Edition Unix`_ - of the function that
384 converts a "name" to an "inode". ``struct nam 369 converts a "name" to an "inode". ``struct nameidata`` contains (among
385 other fields): 370 other fields):
386 371
387 ``struct path path`` 372 ``struct path path``
388 ~~~~~~~~~~~~~~~~~~~~ 373 ~~~~~~~~~~~~~~~~~~~~
389 374
390 A ``path`` contains a ``struct vfsmount`` (whi 375 A ``path`` contains a ``struct vfsmount`` (which is
391 embedded in a ``struct mount``) and a ``struct 376 embedded in a ``struct mount``) and a ``struct dentry``. Together these
392 record the current status of the walk. They s 377 record the current status of the walk. They start out referring to the
393 starting point (the current working directory, 378 starting point (the current working directory, the root directory, or some other
394 directory identified by a file descriptor), an 379 directory identified by a file descriptor), and are updated on each
395 step. A reference through ``d_lockref`` and ` 380 step. A reference through ``d_lockref`` and ``mnt_count`` is always
396 held. 381 held.
397 382
398 ``struct qstr last`` 383 ``struct qstr last``
399 ~~~~~~~~~~~~~~~~~~~~ 384 ~~~~~~~~~~~~~~~~~~~~
400 385
401 This is a string together with a length (i.e. !! 386 This is a string together with a length (i.e. _not_ ``nul`` terminated)
402 that is the "next" component in the pathname. 387 that is the "next" component in the pathname.
403 388
404 ``int last_type`` 389 ``int last_type``
405 ~~~~~~~~~~~~~~~~~ 390 ~~~~~~~~~~~~~~~~~
406 391
407 This is one of ``LAST_NORM``, ``LAST_ROOT``, ` !! 392 This is one of ``LAST_NORM``, ``LAST_ROOT``, ``LAST_DOT``, ``LAST_DOTDOT``, or
408 The ``last`` field is only valid if the type i !! 393 ``LAST_BIND``. The ``last`` field is only valid if the type is
>> 394 ``LAST_NORM``. ``LAST_BIND`` is used when following a symlink and no
>> 395 components of the symlink have been processed yet. Others should be
>> 396 fairly self-explanatory.
409 397
410 ``struct path root`` 398 ``struct path root``
411 ~~~~~~~~~~~~~~~~~~~~ 399 ~~~~~~~~~~~~~~~~~~~~
412 400
413 This is used to hold a reference to the effect 401 This is used to hold a reference to the effective root of the
414 filesystem. Often that reference won't be nee 402 filesystem. Often that reference won't be needed, so this field is
415 only assigned the first time it is used, or wh 403 only assigned the first time it is used, or when a non-standard root
416 is requested. Keeping a reference in the ``na 404 is requested. Keeping a reference in the ``nameidata`` ensures that
417 only one root is in effect for the entire path 405 only one root is in effect for the entire path walk, even if it races
418 with a ``chroot()`` system call. 406 with a ``chroot()`` system call.
419 407
420 It should be noted that in the case of ``LOOKU <<
421 ``LOOKUP_BENEATH``, the effective root becomes <<
422 passed to ``openat2()`` (which exposes these ` <<
423 <<
424 The root is needed when either of two conditio 408 The root is needed when either of two conditions holds: (1) either the
425 pathname or a symbolic link starts with a "'/' 409 pathname or a symbolic link starts with a "'/'", or (2) a "``..``"
426 component is being handled, since "``..``" fro 410 component is being handled, since "``..``" from the root must always stay
427 at the root. The value used is usually the cu 411 at the root. The value used is usually the current root directory of
428 the calling process. An alternate root can be 412 the calling process. An alternate root can be provided as when
429 ``sysctl()`` calls ``file_open_root()``, and w 413 ``sysctl()`` calls ``file_open_root()``, and when NFSv4 or Btrfs call
430 ``mount_subtree()``. In each case a pathname 414 ``mount_subtree()``. In each case a pathname is being looked up in a very
431 specific part of the filesystem, and the looku 415 specific part of the filesystem, and the lookup must not be allowed to
432 escape that subtree. It works a bit like a lo 416 escape that subtree. It works a bit like a local ``chroot()``.
433 417
434 Ignoring the handling of symbolic links, we ca 418 Ignoring the handling of symbolic links, we can now describe the
435 "``link_path_walk()``" function, which handles 419 "``link_path_walk()``" function, which handles the lookup of everything
436 except the final component as: 420 except the final component as:
437 421
438 Given a path (``name``) and a nameidata str 422 Given a path (``name``) and a nameidata structure (``nd``), check that the
439 current directory has execute permission an 423 current directory has execute permission and then advance ``name``
440 over one component while updating ``last_ty 424 over one component while updating ``last_type`` and ``last``. If that
441 was the final component, then return, other 425 was the final component, then return, otherwise call
442 ``walk_component()`` and repeat from the to 426 ``walk_component()`` and repeat from the top.
443 427
444 ``walk_component()`` is even easier. If the c 428 ``walk_component()`` is even easier. If the component is ``LAST_DOTS``,
445 it calls ``handle_dots()`` which does the nece 429 it calls ``handle_dots()`` which does the necessary locking as already
446 described. If it finds a ``LAST_NORM`` compon 430 described. If it finds a ``LAST_NORM`` component it first calls
447 "``lookup_fast()``" which only looks in the dc 431 "``lookup_fast()``" which only looks in the dcache, but will ask the
448 filesystem to revalidate the result if it is t 432 filesystem to revalidate the result if it is that sort of filesystem.
449 If that doesn't get a good result, it calls "` 433 If that doesn't get a good result, it calls "``lookup_slow()``" which
450 takes ``i_rwsem``, rechecks the cache, and the 434 takes ``i_rwsem``, rechecks the cache, and then asks the filesystem
451 to find a definitive answer. !! 435 to find a definitive answer. Each of these will call
>> 436 ``follow_managed()`` (as described below) to handle any mount points.
452 437
453 As the last step of walk_component(), step_int !! 438 In the absence of symbolic links, ``walk_component()`` creates a new
454 directly from walk_component() or from handle_ !! 439 ``struct path`` containing a counted reference to the new dentry and a
455 handle_mounts(), to check and handle mount poi !! 440 reference to the new ``vfsmount`` which is only counted if it is
456 ``struct path`` is created containing a counte !! 441 different from the previous ``vfsmount``. It then calls
457 a reference to the new ``vfsmount`` which is o !! 442 ``path_to_nameidata()`` to install the new ``struct path`` in the
458 different from the previous ``vfsmount``. Then !! 443 ``struct nameidata`` and drop the unneeded references.
459 a symbolic link, step_into() calls pick_link() <<
460 otherwise it installs the new ``struct path`` <<
461 drops the unneeded references. <<
462 444
463 This "hand-over-hand" sequencing of getting a 445 This "hand-over-hand" sequencing of getting a reference to the new
464 dentry before dropping the reference to the pr 446 dentry before dropping the reference to the previous dentry may
465 seem obvious, but is worth pointing out so tha 447 seem obvious, but is worth pointing out so that we will recognize its
466 analogue in the "RCU-walk" version. 448 analogue in the "RCU-walk" version.
467 449
468 Handling the final component 450 Handling the final component
469 ---------------------------- 451 ----------------------------
470 452
471 ``link_path_walk()`` only walks as far as sett 453 ``link_path_walk()`` only walks as far as setting ``nd->last`` and
472 ``nd->last_type`` to refer to the final compon 454 ``nd->last_type`` to refer to the final component of the path. It does
473 not call ``walk_component()`` that last time. 455 not call ``walk_component()`` that last time. Handling that final
474 component remains for the caller to sort out. 456 component remains for the caller to sort out. Those callers are
475 path_lookupat(), path_parentat() and !! 457 ``path_lookupat()``, ``path_parentat()``, ``path_mountpoint()`` and
476 path_openat() each of which handles the differ !! 458 ``path_openat()`` each of which handles the differing requirements of
477 different system calls. 459 different system calls.
478 460
479 ``path_parentat()`` is clearly the simplest - 461 ``path_parentat()`` is clearly the simplest - it just wraps a little bit
480 of housekeeping around ``link_path_walk()`` an 462 of housekeeping around ``link_path_walk()`` and returns the parent
481 directory and final component to the caller. 463 directory and final component to the caller. The caller will be either
482 aiming to create a name (via ``filename_create 464 aiming to create a name (via ``filename_create()``) or remove or rename
483 a name (in which case ``user_path_parent()`` i 465 a name (in which case ``user_path_parent()`` is used). They will use
484 ``i_rwsem`` to exclude other changes while the 466 ``i_rwsem`` to exclude other changes while they validate and then
485 perform their operation. 467 perform their operation.
486 468
487 ``path_lookupat()`` is nearly as simple - it i 469 ``path_lookupat()`` is nearly as simple - it is used when an existing
488 object is wanted such as by ``stat()`` or ``ch 470 object is wanted such as by ``stat()`` or ``chmod()``. It essentially just
489 calls ``walk_component()`` on the final compon 471 calls ``walk_component()`` on the final component through a call to
490 ``lookup_last()``. ``path_lookupat()`` return 472 ``lookup_last()``. ``path_lookupat()`` returns just the final dentry.
491 It is worth noting that when flag ``LOOKUP_MOU !! 473
492 path_lookupat() will unset LOOKUP_JUMPED in na !! 474 ``path_mountpoint()`` handles the special case of unmounting which must
493 subsequent path traversal d_weak_revalidate() !! 475 not try to revalidate the mounted filesystem. It effectively
494 This is important when unmounting a filesystem !! 476 contains, through a call to ``mountpoint_last()``, an alternate
>> 477 implementation of ``lookup_slow()`` which skips that step. This is
>> 478 important when unmounting a filesystem that is inaccessible, such as
495 one provided by a dead NFS server. 479 one provided by a dead NFS server.
496 480
497 Finally ``path_openat()`` is used for the ``op 481 Finally ``path_openat()`` is used for the ``open()`` system call; it
498 contains, in support functions starting with " !! 482 contains, in support functions starting with "``do_last()``", all the
499 complexity needed to handle the different subt 483 complexity needed to handle the different subtleties of O_CREAT (with
500 or without O_EXCL), final "``/``" characters, 484 or without O_EXCL), final "``/``" characters, and trailing symbolic
501 links. We will revisit this in the final part 485 links. We will revisit this in the final part of this series, which
502 focuses on those symbolic links. "open_last_l !! 486 focuses on those symbolic links. "``do_last()``" will sometimes, but
503 not always, take ``i_rwsem``, depending on wha 487 not always, take ``i_rwsem``, depending on what it finds.
504 488
505 Each of these, or the functions which call the 489 Each of these, or the functions which call them, need to be alert to
506 the possibility that the final component is no 490 the possibility that the final component is not ``LAST_NORM``. If the
507 goal of the lookup is to create something, the 491 goal of the lookup is to create something, then any value for
508 ``last_type`` other than ``LAST_NORM`` will re 492 ``last_type`` other than ``LAST_NORM`` will result in an error. For
509 example if ``path_parentat()`` reports ``LAST_ 493 example if ``path_parentat()`` reports ``LAST_DOTDOT``, then the caller
510 won't try to create that name. They also chec 494 won't try to create that name. They also check for trailing slashes
511 by testing ``last.name[last.len]``. If there 495 by testing ``last.name[last.len]``. If there is any character beyond
512 the final component, it must be a trailing sla 496 the final component, it must be a trailing slash.
513 497
514 Revalidation and automounts 498 Revalidation and automounts
515 --------------------------- 499 ---------------------------
516 500
517 Apart from symbolic links, there are only two 501 Apart from symbolic links, there are only two parts of the "REF-walk"
518 process not yet covered. One is the handling 502 process not yet covered. One is the handling of stale cache entries
519 and the other is automounts. 503 and the other is automounts.
520 504
521 On filesystems that require it, the lookup rou 505 On filesystems that require it, the lookup routines will call the
522 ``->d_revalidate()`` dentry method to ensure t 506 ``->d_revalidate()`` dentry method to ensure that the cached information
523 is current. This will often confirm validity 507 is current. This will often confirm validity or update a few details
524 from a server. In some cases it may find that 508 from a server. In some cases it may find that there has been change
525 further up the path and that something that wa 509 further up the path and that something that was thought to be valid
526 previously isn't really. When this happens th 510 previously isn't really. When this happens the lookup of the whole
527 path is aborted and retried with the "``LOOKUP 511 path is aborted and retried with the "``LOOKUP_REVAL``" flag set. This
528 forces revalidation to be more thorough. We w 512 forces revalidation to be more thorough. We will see more details of
529 this retry process in the next article. 513 this retry process in the next article.
530 514
531 Automount points are locations in the filesyst 515 Automount points are locations in the filesystem where an attempt to
532 lookup a name can trigger changes to how that 516 lookup a name can trigger changes to how that lookup should be
533 handled, in particular by mounting a filesyste 517 handled, in particular by mounting a filesystem there. These are
534 covered in greater detail in autofs.txt in the 518 covered in greater detail in autofs.txt in the Linux documentation
535 tree, but a few notes specifically related to 519 tree, but a few notes specifically related to path lookup are in order
536 here. 520 here.
537 521
538 The Linux VFS has a concept of "managed" dentr !! 522 The Linux VFS has a concept of "managed" dentries which is reflected
>> 523 in function names such as "``follow_managed()``". There are three
539 potentially interesting things about these den 524 potentially interesting things about these dentries corresponding
540 to three different flags that might be set in 525 to three different flags that might be set in ``dentry->d_flags``:
541 526
542 ``DCACHE_MANAGE_TRANSIT`` 527 ``DCACHE_MANAGE_TRANSIT``
543 ~~~~~~~~~~~~~~~~~~~~~~~~~ 528 ~~~~~~~~~~~~~~~~~~~~~~~~~
544 529
545 If this flag has been set, then the filesystem 530 If this flag has been set, then the filesystem has requested that the
546 ``d_manage()`` dentry operation be called befo 531 ``d_manage()`` dentry operation be called before handling any possible
547 mount point. This can perform two particular 532 mount point. This can perform two particular services:
548 533
549 It can block to avoid races. If an automount 534 It can block to avoid races. If an automount point is being
550 unmounted, the ``d_manage()`` function will us 535 unmounted, the ``d_manage()`` function will usually wait for that
551 process to complete before letting the new loo 536 process to complete before letting the new lookup proceed and possibly
552 trigger a new automount. 537 trigger a new automount.
553 538
554 It can selectively allow only some processes t 539 It can selectively allow only some processes to transit through a
555 mount point. When a server process is managin 540 mount point. When a server process is managing automounts, it may
556 need to access a directory without triggering 541 need to access a directory without triggering normal automount
557 processing. That server process can identify 542 processing. That server process can identify itself to the ``autofs``
558 filesystem, which will then give it a special 543 filesystem, which will then give it a special pass through
559 ``d_manage()`` by returning ``-EISDIR``. 544 ``d_manage()`` by returning ``-EISDIR``.
560 545
561 ``DCACHE_MOUNTED`` 546 ``DCACHE_MOUNTED``
562 ~~~~~~~~~~~~~~~~~~ 547 ~~~~~~~~~~~~~~~~~~
563 548
564 This flag is set on every dentry that is mount 549 This flag is set on every dentry that is mounted on. As Linux
565 supports multiple filesystem namespaces, it is 550 supports multiple filesystem namespaces, it is possible that the
566 dentry may not be mounted on in *this* namespa 551 dentry may not be mounted on in *this* namespace, just in some
567 other. So this flag is seen as a hint, not a 552 other. So this flag is seen as a hint, not a promise.
568 553
569 If this flag is set, and ``d_manage()`` didn't 554 If this flag is set, and ``d_manage()`` didn't return ``-EISDIR``,
570 ``lookup_mnt()`` is called to examine the moun 555 ``lookup_mnt()`` is called to examine the mount hash table (honoring the
571 ``mount_lock`` described earlier) and possibly 556 ``mount_lock`` described earlier) and possibly return a new ``vfsmount``
572 and a new ``dentry`` (both with counted refere 557 and a new ``dentry`` (both with counted references).
573 558
574 ``DCACHE_NEED_AUTOMOUNT`` 559 ``DCACHE_NEED_AUTOMOUNT``
575 ~~~~~~~~~~~~~~~~~~~~~~~~~ 560 ~~~~~~~~~~~~~~~~~~~~~~~~~
576 561
577 If ``d_manage()`` allowed us to get this far, 562 If ``d_manage()`` allowed us to get this far, and ``lookup_mnt()`` didn't
578 find a mount point, then this flag causes the 563 find a mount point, then this flag causes the ``d_automount()`` dentry
579 operation to be called. 564 operation to be called.
580 565
581 The ``d_automount()`` operation can be arbitra 566 The ``d_automount()`` operation can be arbitrarily complex and may
582 communicate with server processes etc. but it 567 communicate with server processes etc. but it should ultimately either
583 report that there was an error, that there was 568 report that there was an error, that there was nothing to mount, or
584 should provide an updated ``struct path`` with 569 should provide an updated ``struct path`` with new ``dentry`` and ``vfsmount``.
585 570
586 In the latter case, ``finish_automount()`` wil 571 In the latter case, ``finish_automount()`` will be called to safely
587 install the new mount point into the mount tab 572 install the new mount point into the mount table.
588 573
589 There is no new locking of import here and it 574 There is no new locking of import here and it is important that no
590 locks (only counted references) are held over 575 locks (only counted references) are held over this processing due to
591 the very real possibility of extended delays. 576 the very real possibility of extended delays.
592 This will become more important next time when 577 This will become more important next time when we examine RCU-walk
593 which is particularly sensitive to delays. 578 which is particularly sensitive to delays.
594 579
595 RCU-walk - faster pathname lookup in Linux 580 RCU-walk - faster pathname lookup in Linux
596 ========================================== 581 ==========================================
597 582
598 RCU-walk is another algorithm for performing p 583 RCU-walk is another algorithm for performing pathname lookup in Linux.
599 It is in many ways similar to REF-walk and the 584 It is in many ways similar to REF-walk and the two share quite a bit
600 of code. The significant difference in RCU-wa 585 of code. The significant difference in RCU-walk is how it allows for
601 the possibility of concurrent access. 586 the possibility of concurrent access.
602 587
603 We noted that REF-walk is complex because ther 588 We noted that REF-walk is complex because there are numerous details
604 and special cases. RCU-walk reduces this comp 589 and special cases. RCU-walk reduces this complexity by simply
605 refusing to handle a number of cases -- it ins 590 refusing to handle a number of cases -- it instead falls back to
606 REF-walk. The difficulty with RCU-walk comes 591 REF-walk. The difficulty with RCU-walk comes from a different
607 direction: unfamiliarity. The locking rules w 592 direction: unfamiliarity. The locking rules when depending on RCU are
608 quite different from traditional locking, so w 593 quite different from traditional locking, so we will spend a little extra
609 time when we come to those. 594 time when we come to those.
610 595
611 Clear demarcation of roles 596 Clear demarcation of roles
612 -------------------------- 597 --------------------------
613 598
614 The easiest way to manage concurrency is to fo 599 The easiest way to manage concurrency is to forcibly stop any other
615 thread from changing the data structures that 600 thread from changing the data structures that a given thread is
616 looking at. In cases where no other thread wo 601 looking at. In cases where no other thread would even think of
617 changing the data and lots of different thread 602 changing the data and lots of different threads want to read at the
618 same time, this can be very costly. Even when 603 same time, this can be very costly. Even when using locks that permit
619 multiple concurrent readers, the simple act of 604 multiple concurrent readers, the simple act of updating the count of
620 the number of current readers can impose an un 605 the number of current readers can impose an unwanted cost. So the
621 goal when reading a shared data structure that 606 goal when reading a shared data structure that no other process is
622 changing is to avoid writing anything to memor 607 changing is to avoid writing anything to memory at all. Take no
623 locks, increment no counts, leave no footprint 608 locks, increment no counts, leave no footprints.
624 609
625 The REF-walk mechanism already described certa 610 The REF-walk mechanism already described certainly doesn't follow this
626 principle, but then it is really designed to w 611 principle, but then it is really designed to work when there may well
627 be other threads modifying the data. RCU-walk 612 be other threads modifying the data. RCU-walk, in contrast, is
628 designed for the common situation where there 613 designed for the common situation where there are lots of frequent
629 readers and only occasional writers. This may 614 readers and only occasional writers. This may not be common in all
630 parts of the filesystem tree, but in many part 615 parts of the filesystem tree, but in many parts it will be. For the
631 other parts it is important that RCU-walk can 616 other parts it is important that RCU-walk can quickly fall back to
632 using REF-walk. 617 using REF-walk.
633 618
634 Pathname lookup always starts in RCU-walk mode 619 Pathname lookup always starts in RCU-walk mode but only remains there
635 as long as what it is looking for is in the ca 620 as long as what it is looking for is in the cache and is stable. It
636 dances lightly down the cached filesystem imag 621 dances lightly down the cached filesystem image, leaving no footprints
637 and carefully watching where it is, to be sure 622 and carefully watching where it is, to be sure it doesn't trip. If it
638 notices that something has changed or is chang 623 notices that something has changed or is changing, or if something
639 isn't in the cache, then it tries to stop grac 624 isn't in the cache, then it tries to stop gracefully and switch to
640 REF-walk. 625 REF-walk.
641 626
642 This stopping requires getting a counted refer 627 This stopping requires getting a counted reference on the current
643 ``vfsmount`` and ``dentry``, and ensuring that 628 ``vfsmount`` and ``dentry``, and ensuring that these are still valid -
644 that a path walk with REF-walk would have foun 629 that a path walk with REF-walk would have found the same entries.
645 This is an invariant that RCU-walk must guaran 630 This is an invariant that RCU-walk must guarantee. It can only make
646 decisions, such as selecting the next step, th 631 decisions, such as selecting the next step, that are decisions which
647 REF-walk could also have made if it were walki 632 REF-walk could also have made if it were walking down the tree at the
648 same time. If the graceful stop succeeds, the 633 same time. If the graceful stop succeeds, the rest of the path is
649 processed with the reliable, if slightly slugg 634 processed with the reliable, if slightly sluggish, REF-walk. If
650 RCU-walk finds it cannot stop gracefully, it s 635 RCU-walk finds it cannot stop gracefully, it simply gives up and
651 restarts from the top with REF-walk. 636 restarts from the top with REF-walk.
652 637
653 This pattern of "try RCU-walk, if that fails t 638 This pattern of "try RCU-walk, if that fails try REF-walk" can be
654 clearly seen in functions like filename_lookup !! 639 clearly seen in functions like ``filename_lookup()``,
655 filename_parentat(), !! 640 ``filename_parentat()``, ``filename_mountpoint()``,
656 do_filp_open(), and do_file_open_root(). Thes !! 641 ``do_filp_open()``, and ``do_file_open_root()``. These five
657 correspond roughly to the three ``path_*()`` f !! 642 correspond roughly to the four ``path_``* functions we met earlier,
658 each of which calls ``link_path_walk()``. The !! 643 each of which calls ``link_path_walk()``. The ``path_*`` functions are
659 called using different mode flags until a mode 644 called using different mode flags until a mode is found which works.
660 They are first called with ``LOOKUP_RCU`` set 645 They are first called with ``LOOKUP_RCU`` set to request "RCU-walk". If
661 that fails with the error ``ECHILD`` they are 646 that fails with the error ``ECHILD`` they are called again with no
662 special flag to request "REF-walk". If either 647 special flag to request "REF-walk". If either of those report the
663 error ``ESTALE`` a final attempt is made with 648 error ``ESTALE`` a final attempt is made with ``LOOKUP_REVAL`` set (and no
664 ``LOOKUP_RCU``) to ensure that entries found i 649 ``LOOKUP_RCU``) to ensure that entries found in the cache are forcibly
665 revalidated - normally entries are only revali 650 revalidated - normally entries are only revalidated if the filesystem
666 determines that they are too old to trust. 651 determines that they are too old to trust.
667 652
668 The ``LOOKUP_RCU`` attempt may drop that flag 653 The ``LOOKUP_RCU`` attempt may drop that flag internally and switch to
669 REF-walk, but will never then try to switch ba 654 REF-walk, but will never then try to switch back to RCU-walk. Places
670 that trip up RCU-walk are much more likely to 655 that trip up RCU-walk are much more likely to be near the leaves and
671 so it is very unlikely that there will be much 656 so it is very unlikely that there will be much, if any, benefit from
672 switching back. 657 switching back.
673 658
674 RCU and seqlocks: fast and light 659 RCU and seqlocks: fast and light
675 -------------------------------- 660 --------------------------------
676 661
677 RCU is, unsurprisingly, critical to RCU-walk m 662 RCU is, unsurprisingly, critical to RCU-walk mode. The
678 ``rcu_read_lock()`` is held for the entire tim 663 ``rcu_read_lock()`` is held for the entire time that RCU-walk is walking
679 down a path. The particular guarantee it prov 664 down a path. The particular guarantee it provides is that the key
680 data structures - dentries, inodes, super_bloc 665 data structures - dentries, inodes, super_blocks, and mounts - will
681 not be freed while the lock is held. They mig 666 not be freed while the lock is held. They might be unlinked or
682 invalidated in one way or another, but the mem 667 invalidated in one way or another, but the memory will not be
683 repurposed so values in various fields will st 668 repurposed so values in various fields will still be meaningful. This
684 is the only guarantee that RCU provides; every 669 is the only guarantee that RCU provides; everything else is done using
685 seqlocks. 670 seqlocks.
686 671
687 As we saw above, REF-walk holds a counted refe 672 As we saw above, REF-walk holds a counted reference to the current
688 dentry and the current vfsmount, and does not 673 dentry and the current vfsmount, and does not release those references
689 before taking references to the "next" dentry 674 before taking references to the "next" dentry or vfsmount. It also
690 sometimes takes the ``d_lock`` spinlock. Thes 675 sometimes takes the ``d_lock`` spinlock. These references and locks are
691 taken to prevent certain changes from happenin 676 taken to prevent certain changes from happening. RCU-walk must not
692 take those references or locks and so cannot p 677 take those references or locks and so cannot prevent such changes.
693 Instead, it checks to see if a change has been 678 Instead, it checks to see if a change has been made, and aborts or
694 retries if it has. 679 retries if it has.
695 680
696 To preserve the invariant mentioned above (tha 681 To preserve the invariant mentioned above (that RCU-walk may only make
697 decisions that REF-walk could have made), it m 682 decisions that REF-walk could have made), it must make the checks at
698 or near the same places that REF-walk holds th 683 or near the same places that REF-walk holds the references. So, when
699 REF-walk increments a reference count or takes 684 REF-walk increments a reference count or takes a spinlock, RCU-walk
700 samples the status of a seqlock using ``read_s 685 samples the status of a seqlock using ``read_seqcount_begin()`` or a
701 similar function. When REF-walk decrements th 686 similar function. When REF-walk decrements the count or drops the
702 lock, RCU-walk checks if the sampled status is 687 lock, RCU-walk checks if the sampled status is still valid using
703 ``read_seqcount_retry()`` or similar. 688 ``read_seqcount_retry()`` or similar.
704 689
705 However, there is a little bit more to seqlock 690 However, there is a little bit more to seqlocks than that. If
706 RCU-walk accesses two different fields in a se 691 RCU-walk accesses two different fields in a seqlock-protected
707 structure, or accesses the same field twice, t 692 structure, or accesses the same field twice, there is no a priori
708 guarantee of any consistency between those acc 693 guarantee of any consistency between those accesses. When consistency
709 is needed - which it usually is - RCU-walk mus 694 is needed - which it usually is - RCU-walk must take a copy and then
710 use ``read_seqcount_retry()`` to validate that 695 use ``read_seqcount_retry()`` to validate that copy.
711 696
712 ``read_seqcount_retry()`` not only checks the 697 ``read_seqcount_retry()`` not only checks the sequence number, but also
713 imposes a memory barrier so that no memory-rea 698 imposes a memory barrier so that no memory-read instruction from
714 *before* the call can be delayed until *after* 699 *before* the call can be delayed until *after* the call, either by the
715 CPU or by the compiler. A simple example of t 700 CPU or by the compiler. A simple example of this can be seen in
716 ``slow_dentry_cmp()`` which, for filesystems w 701 ``slow_dentry_cmp()`` which, for filesystems which do not use simple
717 byte-wise name equality, calls into the filesy 702 byte-wise name equality, calls into the filesystem to compare a name
718 against a dentry. The length and name pointer 703 against a dentry. The length and name pointer are copied into local
719 variables, then ``read_seqcount_retry()`` is c 704 variables, then ``read_seqcount_retry()`` is called to confirm the two
720 are consistent, and only then is ``->d_compare 705 are consistent, and only then is ``->d_compare()`` called. When
721 standard filename comparison is used, ``dentry 706 standard filename comparison is used, ``dentry_cmp()`` is called
722 instead. Notably it does *not* use ``read_seq !! 707 instead. Notably it does _not_ use ``read_seqcount_retry()``, but
723 instead has a large comment explaining why the 708 instead has a large comment explaining why the consistency guarantee
724 isn't necessary. A subsequent ``read_seqcount 709 isn't necessary. A subsequent ``read_seqcount_retry()`` will be
725 sufficient to catch any problem that could occ 710 sufficient to catch any problem that could occur at this point.
726 711
727 With that little refresher on seqlocks out of 712 With that little refresher on seqlocks out of the way we can look at
728 the bigger picture of how RCU-walk uses seqloc 713 the bigger picture of how RCU-walk uses seqlocks.
729 714
730 ``mount_lock`` and ``nd->m_seq`` 715 ``mount_lock`` and ``nd->m_seq``
731 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 716 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
732 717
733 We already met the ``mount_lock`` seqlock when 718 We already met the ``mount_lock`` seqlock when REF-walk used it to
734 ensure that crossing a mount point is performe 719 ensure that crossing a mount point is performed safely. RCU-walk uses
735 it for that too, but for quite a bit more. 720 it for that too, but for quite a bit more.
736 721
737 Instead of taking a counted reference to each 722 Instead of taking a counted reference to each ``vfsmount`` as it
738 descends the tree, RCU-walk samples the state 723 descends the tree, RCU-walk samples the state of ``mount_lock`` at the
739 start of the walk and stores this initial sequ 724 start of the walk and stores this initial sequence number in the
740 ``struct nameidata`` in the ``m_seq`` field. 725 ``struct nameidata`` in the ``m_seq`` field. This one lock and one
741 sequence number are used to validate all acces 726 sequence number are used to validate all accesses to all ``vfsmounts``,
742 and all mount point crossings. As changes to 727 and all mount point crossings. As changes to the mount table are
743 relatively rare, it is reasonable to fall back 728 relatively rare, it is reasonable to fall back on REF-walk any time
744 that any "mount" or "unmount" happens. 729 that any "mount" or "unmount" happens.
745 730
746 ``m_seq`` is checked (using ``read_seqretry()` 731 ``m_seq`` is checked (using ``read_seqretry()``) at the end of an RCU-walk
747 sequence, whether switching to REF-walk for th 732 sequence, whether switching to REF-walk for the rest of the path or
748 when the end of the path is reached. It is al 733 when the end of the path is reached. It is also checked when stepping
749 down over a mount point (in ``__follow_mount_r 734 down over a mount point (in ``__follow_mount_rcu()``) or up (in
750 ``follow_dotdot_rcu()``). If it is ever found 735 ``follow_dotdot_rcu()``). If it is ever found to have changed, the
751 whole RCU-walk sequence is aborted and the pat 736 whole RCU-walk sequence is aborted and the path is processed again by
752 REF-walk. 737 REF-walk.
753 738
754 If RCU-walk finds that ``mount_lock`` hasn't c 739 If RCU-walk finds that ``mount_lock`` hasn't changed then it can be sure
755 that, had REF-walk taken counted references on 740 that, had REF-walk taken counted references on each vfsmount, the
756 results would have been the same. This ensure 741 results would have been the same. This ensures the invariant holds,
757 at least for vfsmount structures. 742 at least for vfsmount structures.
758 743
759 ``dentry->d_seq`` and ``nd->seq`` 744 ``dentry->d_seq`` and ``nd->seq``
760 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 745 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
761 746
762 In place of taking a count or lock on ``d_refl 747 In place of taking a count or lock on ``d_reflock``, RCU-walk samples
763 the per-dentry ``d_seq`` seqlock, and stores t 748 the per-dentry ``d_seq`` seqlock, and stores the sequence number in the
764 ``seq`` field of the nameidata structure, so ` 749 ``seq`` field of the nameidata structure, so ``nd->seq`` should always be
765 the current sequence number of ``nd->dentry``. 750 the current sequence number of ``nd->dentry``. This number needs to be
766 revalidated after copying, and before using, t 751 revalidated after copying, and before using, the name, parent, or
767 inode of the dentry. 752 inode of the dentry.
768 753
769 The handling of the name we have already looke 754 The handling of the name we have already looked at, and the parent is
770 only accessed in ``follow_dotdot_rcu()`` which 755 only accessed in ``follow_dotdot_rcu()`` which fairly trivially follows
771 the required pattern, though it does so for th 756 the required pattern, though it does so for three different cases.
772 757
773 When not at a mount point, ``d_parent`` is fol 758 When not at a mount point, ``d_parent`` is followed and its ``d_seq`` is
774 collected. When we are at a mount point, we i 759 collected. When we are at a mount point, we instead follow the
775 ``mnt->mnt_mountpoint`` link to get a new dent 760 ``mnt->mnt_mountpoint`` link to get a new dentry and collect its
776 ``d_seq``. Then, after finally finding a ``d_ 761 ``d_seq``. Then, after finally finding a ``d_parent`` to follow, we must
777 check if we have landed on a mount point and, 762 check if we have landed on a mount point and, if so, must find that
778 mount point and follow the ``mnt->mnt_root`` l 763 mount point and follow the ``mnt->mnt_root`` link. This would imply a
779 somewhat unusual, but certainly possible, circ 764 somewhat unusual, but certainly possible, circumstance where the
780 starting point of the path lookup was in part 765 starting point of the path lookup was in part of the filesystem that
781 was mounted on, and so not visible from the ro 766 was mounted on, and so not visible from the root.
782 767
783 The inode pointer, stored in ``->d_inode``, is 768 The inode pointer, stored in ``->d_inode``, is a little more
784 interesting. The inode will always need to be 769 interesting. The inode will always need to be accessed at least
785 twice, once to determine if it is NULL and onc 770 twice, once to determine if it is NULL and once to verify access
786 permissions. Symlink handling requires a vali 771 permissions. Symlink handling requires a validated inode pointer too.
787 Rather than revalidating on each access, a cop 772 Rather than revalidating on each access, a copy is made on the first
788 access and it is stored in the ``inode`` field 773 access and it is stored in the ``inode`` field of ``nameidata`` from where
789 it can be safely accessed without further vali 774 it can be safely accessed without further validation.
790 775
791 ``lookup_fast()`` is the only lookup routine t 776 ``lookup_fast()`` is the only lookup routine that is used in RCU-mode,
792 ``lookup_slow()`` being too slow and requiring 777 ``lookup_slow()`` being too slow and requiring locks. It is in
793 ``lookup_fast()`` that we find the important " 778 ``lookup_fast()`` that we find the important "hand over hand" tracking
794 of the current dentry. 779 of the current dentry.
795 780
796 The current ``dentry`` and current ``seq`` num 781 The current ``dentry`` and current ``seq`` number are passed to
797 ``__d_lookup_rcu()`` which, on success, return 782 ``__d_lookup_rcu()`` which, on success, returns a new ``dentry`` and a
798 new ``seq`` number. ``lookup_fast()`` then co 783 new ``seq`` number. ``lookup_fast()`` then copies the inode pointer and
799 revalidates the new ``seq`` number. It then v 784 revalidates the new ``seq`` number. It then validates the old ``dentry``
800 with the old ``seq`` number one last time and 785 with the old ``seq`` number one last time and only then continues. This
801 process of getting the ``seq`` number of the n 786 process of getting the ``seq`` number of the new dentry and then
802 checking the ``seq`` number of the old exactly 787 checking the ``seq`` number of the old exactly mirrors the process of
803 getting a counted reference to the new dentry 788 getting a counted reference to the new dentry before dropping that for
804 the old dentry which we saw in REF-walk. 789 the old dentry which we saw in REF-walk.
805 790
806 No ``inode->i_rwsem`` or even ``rename_lock`` 791 No ``inode->i_rwsem`` or even ``rename_lock``
807 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 792 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
808 793
809 A semaphore is a fairly heavyweight lock that 794 A semaphore is a fairly heavyweight lock that can only be taken when it is
810 permissible to sleep. As ``rcu_read_lock()`` 795 permissible to sleep. As ``rcu_read_lock()`` forbids sleeping,
811 ``inode->i_rwsem`` plays no role in RCU-walk. 796 ``inode->i_rwsem`` plays no role in RCU-walk. If some other thread does
812 take ``i_rwsem`` and modifies the directory in 797 take ``i_rwsem`` and modifies the directory in a way that RCU-walk needs
813 to notice, the result will be either that RCU- 798 to notice, the result will be either that RCU-walk fails to find the
814 dentry that it is looking for, or it will find 799 dentry that it is looking for, or it will find a dentry which
815 ``read_seqretry()`` won't validate. In either 800 ``read_seqretry()`` won't validate. In either case it will drop down to
816 REF-walk mode which can take whatever locks ar 801 REF-walk mode which can take whatever locks are needed.
817 802
818 Though ``rename_lock`` could be used by RCU-wa 803 Though ``rename_lock`` could be used by RCU-walk as it doesn't require
819 any sleeping, RCU-walk doesn't bother. REF-wa 804 any sleeping, RCU-walk doesn't bother. REF-walk uses ``rename_lock`` to
820 protect against the possibility of hash chains 805 protect against the possibility of hash chains in the dcache changing
821 while they are being searched. This can resul 806 while they are being searched. This can result in failing to find
822 something that actually is there. When RCU-wa 807 something that actually is there. When RCU-walk fails to find
823 something in the dentry cache, whether it is r 808 something in the dentry cache, whether it is really there or not, it
824 already drops down to REF-walk and tries again 809 already drops down to REF-walk and tries again with appropriate
825 locking. This neatly handles all cases, so ad 810 locking. This neatly handles all cases, so adding extra checks on
826 rename_lock would bring no significant value. 811 rename_lock would bring no significant value.
827 812
828 ``unlazy walk()`` and ``complete_walk()`` 813 ``unlazy walk()`` and ``complete_walk()``
829 ----------------------------------------- 814 -----------------------------------------
830 815
831 That "dropping down to REF-walk" typically inv 816 That "dropping down to REF-walk" typically involves a call to
832 ``unlazy_walk()``, so named because "RCU-walk" 817 ``unlazy_walk()``, so named because "RCU-walk" is also sometimes
833 referred to as "lazy walk". ``unlazy_walk()`` 818 referred to as "lazy walk". ``unlazy_walk()`` is called when
834 following the path down to the current vfsmoun 819 following the path down to the current vfsmount/dentry pair seems to
835 have proceeded successfully, but the next step 820 have proceeded successfully, but the next step is problematic. This
836 can happen if the next name cannot be found in 821 can happen if the next name cannot be found in the dcache, if
837 permission checking or name revalidation could 822 permission checking or name revalidation couldn't be achieved while
838 the ``rcu_read_lock()`` is held (which forbids 823 the ``rcu_read_lock()`` is held (which forbids sleeping), if an
839 automount point is found, or in a couple of ca 824 automount point is found, or in a couple of cases involving symlinks.
840 It is also called from ``complete_walk()`` whe 825 It is also called from ``complete_walk()`` when the lookup has reached
841 the final component, or the very end of the pa 826 the final component, or the very end of the path, depending on which
842 particular flavor of lookup is used. 827 particular flavor of lookup is used.
843 828
844 Other reasons for dropping out of RCU-walk tha 829 Other reasons for dropping out of RCU-walk that do not trigger a call
845 to ``unlazy_walk()`` are when some inconsisten 830 to ``unlazy_walk()`` are when some inconsistency is found that cannot be
846 handled immediately, such as ``mount_lock`` or 831 handled immediately, such as ``mount_lock`` or one of the ``d_seq``
847 seqlocks reporting a change. In these cases t 832 seqlocks reporting a change. In these cases the relevant function
848 will return ``-ECHILD`` which will percolate u 833 will return ``-ECHILD`` which will percolate up until it triggers a new
849 attempt from the top using REF-walk. 834 attempt from the top using REF-walk.
850 835
851 For those cases where ``unlazy_walk()`` is an 836 For those cases where ``unlazy_walk()`` is an option, it essentially
852 takes a reference on each of the pointers that 837 takes a reference on each of the pointers that it holds (vfsmount,
853 dentry, and possibly some symbolic links) and 838 dentry, and possibly some symbolic links) and then verifies that the
854 relevant seqlocks have not been changed. If t 839 relevant seqlocks have not been changed. If there have been changes,
855 it, too, aborts with ``-ECHILD``, otherwise th 840 it, too, aborts with ``-ECHILD``, otherwise the transition to REF-walk
856 has been a success and the lookup process cont 841 has been a success and the lookup process continues.
857 842
858 Taking a reference on those pointers is not qu 843 Taking a reference on those pointers is not quite as simple as just
859 incrementing a counter. That works to take a 844 incrementing a counter. That works to take a second reference if you
860 already have one (often indirectly through ano 845 already have one (often indirectly through another object), but it
861 isn't sufficient if you don't actually have a 846 isn't sufficient if you don't actually have a counted reference at
862 all. For ``dentry->d_lockref``, it is safe to 847 all. For ``dentry->d_lockref``, it is safe to increment the reference
863 counter to get a reference unless it has been 848 counter to get a reference unless it has been explicitly marked as
864 "dead" which involves setting the counter to ` 849 "dead" which involves setting the counter to ``-128``.
865 ``lockref_get_not_dead()`` achieves this. 850 ``lockref_get_not_dead()`` achieves this.
866 851
867 For ``mnt->mnt_count`` it is safe to take a re 852 For ``mnt->mnt_count`` it is safe to take a reference as long as
868 ``mount_lock`` is then used to validate the re 853 ``mount_lock`` is then used to validate the reference. If that
869 validation fails, it may *not* be safe to just 854 validation fails, it may *not* be safe to just drop that reference in
870 the standard way of calling ``mnt_put()`` - an 855 the standard way of calling ``mnt_put()`` - an unmount may have
871 progressed too far. So the code in ``legitimi 856 progressed too far. So the code in ``legitimize_mnt()``, when it
872 finds that the reference it got might not be s 857 finds that the reference it got might not be safe, checks the
873 ``MNT_SYNC_UMOUNT`` flag to determine if a sim 858 ``MNT_SYNC_UMOUNT`` flag to determine if a simple ``mnt_put()`` is
874 correct, or if it should just decrement the co 859 correct, or if it should just decrement the count and pretend none of
875 this ever happened. 860 this ever happened.
876 861
877 Taking care in filesystems 862 Taking care in filesystems
878 -------------------------- 863 --------------------------
879 864
880 RCU-walk depends almost entirely on cached inf 865 RCU-walk depends almost entirely on cached information and often will
881 not call into the filesystem at all. However 866 not call into the filesystem at all. However there are two places,
882 besides the already-mentioned component-name c 867 besides the already-mentioned component-name comparison, where the
883 file system might be included in RCU-walk, and 868 file system might be included in RCU-walk, and it must know to be
884 careful. 869 careful.
885 870
886 If the filesystem has non-standard permission- 871 If the filesystem has non-standard permission-checking requirements -
887 such as a networked filesystem which may need 872 such as a networked filesystem which may need to check with the server
888 - the ``i_op->permission`` interface might be 873 - the ``i_op->permission`` interface might be called during RCU-walk.
889 In this case an extra "``MAY_NOT_BLOCK``" flag 874 In this case an extra "``MAY_NOT_BLOCK``" flag is passed so that it
890 knows not to sleep, but to return ``-ECHILD`` 875 knows not to sleep, but to return ``-ECHILD`` if it cannot complete
891 promptly. ``i_op->permission`` is given the i 876 promptly. ``i_op->permission`` is given the inode pointer, not the
892 dentry, so it doesn't need to worry about furt 877 dentry, so it doesn't need to worry about further consistency checks.
893 However if it accesses any other filesystem da 878 However if it accesses any other filesystem data structures, it must
894 ensure they are safe to be accessed with only 879 ensure they are safe to be accessed with only the ``rcu_read_lock()``
895 held. This typically means they must be freed 880 held. This typically means they must be freed using ``kfree_rcu()`` or
896 similar. 881 similar.
897 882
898 .. _READ_ONCE: https://lwn.net/Articles/624126 883 .. _READ_ONCE: https://lwn.net/Articles/624126/
899 884
900 If the filesystem may need to revalidate dcach 885 If the filesystem may need to revalidate dcache entries, then
901 ``d_op->d_revalidate`` may be called in RCU-wa 886 ``d_op->d_revalidate`` may be called in RCU-walk too. This interface
902 *is* passed the dentry but does not have acces 887 *is* passed the dentry but does not have access to the ``inode`` or the
903 ``seq`` number from the ``nameidata``, so it n 888 ``seq`` number from the ``nameidata``, so it needs to be extra careful
904 when accessing fields in the dentry. This "ex 889 when accessing fields in the dentry. This "extra care" typically
905 involves using `READ_ONCE() <READ_ONCE_>`_ to 890 involves using `READ_ONCE() <READ_ONCE_>`_ to access fields, and verifying the
906 result is not NULL before using it. This patt 891 result is not NULL before using it. This pattern can be seen in
907 ``nfs_lookup_revalidate()``. 892 ``nfs_lookup_revalidate()``.
908 893
909 A pair of patterns 894 A pair of patterns
910 ------------------ 895 ------------------
911 896
912 In various places in the details of REF-walk a 897 In various places in the details of REF-walk and RCU-walk, and also in
913 the big picture, there are a couple of related 898 the big picture, there are a couple of related patterns that are worth
914 being aware of. 899 being aware of.
915 900
916 The first is "try quickly and check, if that f 901 The first is "try quickly and check, if that fails try slowly". We
917 can see that in the high-level approach of fir 902 can see that in the high-level approach of first trying RCU-walk and
918 then trying REF-walk, and in places where ``un 903 then trying REF-walk, and in places where ``unlazy_walk()`` is used to
919 switch to REF-walk for the rest of the path. 904 switch to REF-walk for the rest of the path. We also saw it earlier
920 in ``dget_parent()`` when following a "``..``" 905 in ``dget_parent()`` when following a "``..``" link. It tries a quick way
921 to get a reference, then falls back to taking 906 to get a reference, then falls back to taking locks if needed.
922 907
923 The second pattern is "try quickly and check, 908 The second pattern is "try quickly and check, if that fails try
924 again - repeatedly". This is seen with the us 909 again - repeatedly". This is seen with the use of ``rename_lock`` and
925 ``mount_lock`` in REF-walk. RCU-walk doesn't 910 ``mount_lock`` in REF-walk. RCU-walk doesn't make use of this pattern -
926 if anything goes wrong it is much safer to jus 911 if anything goes wrong it is much safer to just abort and try a more
927 sedate approach. 912 sedate approach.
928 913
929 The emphasis here is "try quickly and check". 914 The emphasis here is "try quickly and check". It should probably be
930 "try quickly *and carefully*, then check". Th !! 915 "try quickly _and carefully,_ then check". The fact that checking is
931 needed is a reminder that the system is dynami 916 needed is a reminder that the system is dynamic and only a limited
932 number of things are safe at all. The most li 917 number of things are safe at all. The most likely cause of errors in
933 this whole process is assuming something is sa 918 this whole process is assuming something is safe when in reality it
934 isn't. Careful consideration of what exactly 919 isn't. Careful consideration of what exactly guarantees the safety of
935 each access is sometimes necessary. 920 each access is sometimes necessary.
936 921
937 A walk among the symlinks 922 A walk among the symlinks
938 ========================= 923 =========================
939 924
940 There are several basic issues that we will ex 925 There are several basic issues that we will examine to understand the
941 handling of symbolic links: the symlink stack 926 handling of symbolic links: the symlink stack, together with cache
942 lifetimes, will help us understand the overall 927 lifetimes, will help us understand the overall recursive handling of
943 symlinks and lead to the special care needed f 928 symlinks and lead to the special care needed for the final component.
944 Then a consideration of access-time updates an 929 Then a consideration of access-time updates and summary of the various
945 flags controlling lookup will finish the story 930 flags controlling lookup will finish the story.
946 931
947 The symlink stack 932 The symlink stack
948 ----------------- 933 -----------------
949 934
950 There are only two sorts of filesystem objects 935 There are only two sorts of filesystem objects that can usefully
951 appear in a path prior to the final component: 936 appear in a path prior to the final component: directories and symlinks.
952 Handling directories is quite straightforward: 937 Handling directories is quite straightforward: the new directory
953 simply becomes the starting point at which to 938 simply becomes the starting point at which to interpret the next
954 component on the path. Handling symbolic link 939 component on the path. Handling symbolic links requires a bit more
955 work. 940 work.
956 941
957 Conceptually, symbolic links could be handled 942 Conceptually, symbolic links could be handled by editing the path. If
958 a component name refers to a symbolic link, th 943 a component name refers to a symbolic link, then that component is
959 replaced by the body of the link and, if that 944 replaced by the body of the link and, if that body starts with a '/',
960 then all preceding parts of the path are disca 945 then all preceding parts of the path are discarded. This is what the
961 "``readlink -f``" command does, though it also 946 "``readlink -f``" command does, though it also edits out "``.``" and
962 "``..``" components. 947 "``..``" components.
963 948
964 Directly editing the path string is not really 949 Directly editing the path string is not really necessary when looking
965 up a path, and discarding early components is 950 up a path, and discarding early components is pointless as they aren't
966 looked at anyway. Keeping track of all remain 951 looked at anyway. Keeping track of all remaining components is
967 important, but they can of course be kept sepa 952 important, but they can of course be kept separately; there is no need
968 to concatenate them. As one symlink may easil 953 to concatenate them. As one symlink may easily refer to another,
969 which in turn can refer to a third, we may nee 954 which in turn can refer to a third, we may need to keep the remaining
970 components of several paths, each to be proces 955 components of several paths, each to be processed when the preceding
971 ones are completed. These path remnants are k 956 ones are completed. These path remnants are kept on a stack of
972 limited size. 957 limited size.
973 958
974 There are two reasons for placing limits on ho 959 There are two reasons for placing limits on how many symlinks can
975 occur in a single path lookup. The most obvio 960 occur in a single path lookup. The most obvious is to avoid loops.
976 If a symlink referred to itself either directl 961 If a symlink referred to itself either directly or through
977 intermediaries, then following the symlink can 962 intermediaries, then following the symlink can never complete
978 successfully - the error ``ELOOP`` must be ret 963 successfully - the error ``ELOOP`` must be returned. Loops can be
979 detected without imposing limits, but limits a 964 detected without imposing limits, but limits are the simplest solution
980 and, given the second reason for restriction, 965 and, given the second reason for restriction, quite sufficient.
981 966
982 .. _outlined recently: http://thread.gmane.org 967 .. _outlined recently: http://thread.gmane.org/gmane.linux.kernel/1934390/focus=1934550
983 968
984 The second reason was `outlined recently`_ by 969 The second reason was `outlined recently`_ by Linus:
985 970
986 Because it's a latency and DoS issue too. W 971 Because it's a latency and DoS issue too. We need to react well to
987 true loops, but also to "very deep" non-loo 972 true loops, but also to "very deep" non-loops. It's not about memory
988 use, it's about users triggering unreasonab 973 use, it's about users triggering unreasonable CPU resources.
989 974
990 Linux imposes a limit on the length of any pat 975 Linux imposes a limit on the length of any pathname: ``PATH_MAX``, which
991 is 4096. There are a number of reasons for th 976 is 4096. There are a number of reasons for this limit; not letting the
992 kernel spend too much time on just one path is 977 kernel spend too much time on just one path is one of them. With
993 symbolic links you can effectively generate mu 978 symbolic links you can effectively generate much longer paths so some
994 sort of limit is needed for the same reason. 979 sort of limit is needed for the same reason. Linux imposes a limit of
995 at most 40 (MAXSYMLINKS) symlinks in any one p !! 980 at most 40 symlinks in any one path lookup. It previously imposed a
996 a further limit of eight on the maximum depth !! 981 further limit of eight on the maximum depth of recursion, but that was
997 raised to 40 when a separate stack was impleme 982 raised to 40 when a separate stack was implemented, so there is now
998 just the one limit. 983 just the one limit.
999 984
1000 The ``nameidata`` structure that we met in an 985 The ``nameidata`` structure that we met in an earlier article contains a
1001 small stack that can be used to store the rem 986 small stack that can be used to store the remaining part of up to two
1002 symlinks. In many cases this will be suffici 987 symlinks. In many cases this will be sufficient. If it isn't, a
1003 separate stack is allocated with room for 40 988 separate stack is allocated with room for 40 symlinks. Pathname
1004 lookup will never exceed that stack as, once 989 lookup will never exceed that stack as, once the 40th symlink is
1005 detected, an error is returned. 990 detected, an error is returned.
1006 991
1007 It might seem that the name remnants are all 992 It might seem that the name remnants are all that needs to be stored on
1008 this stack, but we need a bit more. To see t 993 this stack, but we need a bit more. To see that, we need to move on to
1009 cache lifetimes. 994 cache lifetimes.
1010 995
1011 Storage and lifetime of cached symlinks 996 Storage and lifetime of cached symlinks
1012 --------------------------------------- 997 ---------------------------------------
1013 998
1014 Like other filesystem resources, such as inod 999 Like other filesystem resources, such as inodes and directory
1015 entries, symlinks are cached by Linux to avoi 1000 entries, symlinks are cached by Linux to avoid repeated costly access
1016 to external storage. It is particularly impo 1001 to external storage. It is particularly important for RCU-walk to be
1017 able to find and temporarily hold onto these 1002 able to find and temporarily hold onto these cached entries, so that
1018 it doesn't need to drop down into REF-walk. 1003 it doesn't need to drop down into REF-walk.
1019 1004
1020 .. _object-oriented design pattern: https://l 1005 .. _object-oriented design pattern: https://lwn.net/Articles/446317/
1021 1006
1022 While each filesystem is free to make its own 1007 While each filesystem is free to make its own choice, symlinks are
1023 typically stored in one of two places. Short 1008 typically stored in one of two places. Short symlinks are often
1024 stored directly in the inode. When a filesys 1009 stored directly in the inode. When a filesystem allocates a ``struct
1025 inode`` it typically allocates extra space to 1010 inode`` it typically allocates extra space to store private data (a
1026 common `object-oriented design pattern`_ in t 1011 common `object-oriented design pattern`_ in the kernel). This will
1027 sometimes include space for a symlink. The o 1012 sometimes include space for a symlink. The other common location is
1028 in the page cache, which normally stores the 1013 in the page cache, which normally stores the content of files. The
1029 pathname in a symlink can be seen as the cont 1014 pathname in a symlink can be seen as the content of that symlink and
1030 can easily be stored in the page cache just l 1015 can easily be stored in the page cache just like file content.
1031 1016
1032 When neither of these is suitable, the next m 1017 When neither of these is suitable, the next most likely scenario is
1033 that the filesystem will allocate some tempor 1018 that the filesystem will allocate some temporary memory and copy or
1034 construct the symlink content into that memor 1019 construct the symlink content into that memory whenever it is needed.
1035 1020
1036 When the symlink is stored in the inode, it h 1021 When the symlink is stored in the inode, it has the same lifetime as
1037 the inode which, itself, is protected by RCU 1022 the inode which, itself, is protected by RCU or by a counted reference
1038 on the dentry. This means that the mechanism 1023 on the dentry. This means that the mechanisms that pathname lookup
1039 uses to access the dcache and icache (inode c 1024 uses to access the dcache and icache (inode cache) safely are quite
1040 sufficient for accessing some cached symlinks 1025 sufficient for accessing some cached symlinks safely. In these cases,
1041 the ``i_link`` pointer in the inode is set to 1026 the ``i_link`` pointer in the inode is set to point to wherever the
1042 symlink is stored and it can be accessed dire 1027 symlink is stored and it can be accessed directly whenever needed.
1043 1028
1044 When the symlink is stored in the page cache 1029 When the symlink is stored in the page cache or elsewhere, the
1045 situation is not so straightforward. A refer 1030 situation is not so straightforward. A reference on a dentry or even
1046 on an inode does not imply any reference on c 1031 on an inode does not imply any reference on cached pages of that
1047 inode, and even an ``rcu_read_lock()`` is not 1032 inode, and even an ``rcu_read_lock()`` is not sufficient to ensure that
1048 a page will not disappear. So for these syml 1033 a page will not disappear. So for these symlinks the pathname lookup
1049 code needs to ask the filesystem to provide a 1034 code needs to ask the filesystem to provide a stable reference and,
1050 significantly, needs to release that referenc 1035 significantly, needs to release that reference when it is finished
1051 with it. 1036 with it.
1052 1037
1053 Taking a reference to a cache page is often p 1038 Taking a reference to a cache page is often possible even in RCU-walk
1054 mode. It does require making changes to memo 1039 mode. It does require making changes to memory, which is best avoided,
1055 but that isn't necessarily a big cost and it 1040 but that isn't necessarily a big cost and it is better than dropping
1056 out of RCU-walk mode completely. Even filesy 1041 out of RCU-walk mode completely. Even filesystems that allocate
1057 space to copy the symlink into can use ``GFP_ 1042 space to copy the symlink into can use ``GFP_ATOMIC`` to often successfully
1058 allocate memory without the need to drop out 1043 allocate memory without the need to drop out of RCU-walk. If a
1059 filesystem cannot successfully get a referenc 1044 filesystem cannot successfully get a reference in RCU-walk mode, it
1060 must return ``-ECHILD`` and ``unlazy_walk()`` 1045 must return ``-ECHILD`` and ``unlazy_walk()`` will be called to return to
1061 REF-walk mode in which the filesystem is allo 1046 REF-walk mode in which the filesystem is allowed to sleep.
1062 1047
1063 The place for all this to happen is the ``i_o !! 1048 The place for all this to happen is the ``i_op->follow_link()`` inode
1064 method. This is called both in RCU-walk and R !! 1049 method. In the present mainline code this is never actually called in
1065 ``dentry*`` argument is NULL, ``->get_link()` !! 1050 RCU-walk mode as the rewrite is not quite complete. It is likely that
1066 RCU-walk. Much like the ``i_op->permission() !! 1051 in a future release this method will be passed an ``inode`` pointer when
1067 looked at previously, ``->get_link()`` would !! 1052 called in RCU-walk mode so it both (1) knows to be careful, and (2) has the
>> 1053 validated pointer. Much like the ``i_op->permission()`` method we
>> 1054 looked at previously, ``->follow_link()`` would need to be careful that
1068 all the data structures it references are saf 1055 all the data structures it references are safe to be accessed while
1069 holding no counted reference, only the RCU lo !! 1056 holding no counted reference, only the RCU lock. Though getting a
1070 ``struct delayed_called`` will be passed to ` !! 1057 reference with ``->follow_link()`` is not yet done in RCU-walk mode, the
1071 file systems can set their own put_link funct !! 1058 code is ready to release the reference when that does happen.
1072 set_delayed_call(). Later on, when VFS wants !! 1059
1073 do_delayed_call() to invoke that callback fun !! 1060 This need to drop the reference to a symlink adds significant
>> 1061 complexity. It requires a reference to the inode so that the
>> 1062 ``i_op->put_link()`` inode operation can be called. In REF-walk, that
>> 1063 reference is kept implicitly through a reference to the dentry, so
>> 1064 keeping the ``struct path`` of the symlink is easiest. For RCU-walk,
>> 1065 the pointer to the inode is kept separately. To allow switching from
>> 1066 RCU-walk back to REF-walk in the middle of processing nested symlinks
>> 1067 we also need the seq number for the dentry so we can confirm that
>> 1068 switching back was safe.
>> 1069
>> 1070 Finally, when providing a reference to a symlink, the filesystem also
>> 1071 provides an opaque "cookie" that must be passed to ``->put_link()`` so that it
>> 1072 knows what to free. This might be the allocated memory area, or a
>> 1073 pointer to the ``struct page`` in the page cache, or something else
>> 1074 completely. Only the filesystem knows what it is.
1074 1075
1075 In order for the reference to each symlink to 1076 In order for the reference to each symlink to be dropped when the walk completes,
1076 whether in RCU-walk or REF-walk, the symlink 1077 whether in RCU-walk or REF-walk, the symlink stack needs to contain,
1077 along with the path remnants: 1078 along with the path remnants:
1078 1079
1079 - the ``struct path`` to provide a reference !! 1080 - the ``struct path`` to provide a reference to the inode in REF-walk
1080 - the ``const char *`` to provide a reference !! 1081 - the ``struct inode *`` to provide a reference to the inode in RCU-walk
1081 - the ``seq`` to allow the path to be safely 1082 - the ``seq`` to allow the path to be safely switched from RCU-walk to REF-walk
1082 - the ``struct delayed_call`` for later invoc !! 1083 - the ``cookie`` that tells ``->put_path()`` what to put.
1083 1084
1084 This means that each entry in the symlink sta 1085 This means that each entry in the symlink stack needs to hold five
1085 pointers and an integer instead of just one p 1086 pointers and an integer instead of just one pointer (the path
1086 remnant). On a 64-bit system, this is about 1087 remnant). On a 64-bit system, this is about 40 bytes per entry;
1087 with 40 entries it adds up to 1600 bytes tota 1088 with 40 entries it adds up to 1600 bytes total, which is less than
1088 half a page. So it might seem like a lot, bu 1089 half a page. So it might seem like a lot, but is by no means
1089 excessive. 1090 excessive.
1090 1091
1091 Note that, in a given stack frame, the path r 1092 Note that, in a given stack frame, the path remnant (``name``) is not
1092 part of the symlink that the other fields ref 1093 part of the symlink that the other fields refer to. It is the remnant
1093 to be followed once that symlink has been ful 1094 to be followed once that symlink has been fully parsed.
1094 1095
1095 Following the symlink 1096 Following the symlink
1096 --------------------- 1097 ---------------------
1097 1098
1098 The main loop in ``link_path_walk()`` iterate 1099 The main loop in ``link_path_walk()`` iterates seamlessly over all
1099 components in the path and all of the non-fin 1100 components in the path and all of the non-final symlinks. As symlinks
1100 are processed, the ``name`` pointer is adjust 1101 are processed, the ``name`` pointer is adjusted to point to a new
1101 symlink, or is restored from the stack, so th 1102 symlink, or is restored from the stack, so that much of the loop
1102 doesn't need to notice. Getting this ``name` 1103 doesn't need to notice. Getting this ``name`` variable on and off the
1103 stack is very straightforward; pushing and po 1104 stack is very straightforward; pushing and popping the references is
1104 a little more complex. 1105 a little more complex.
1105 1106
1106 When a symlink is found, walk_component() cal !! 1107 When a symlink is found, ``walk_component()`` returns the value ``1``
1107 which returns the link from the filesystem. !! 1108 (``0`` is returned for any other sort of success, and a negative number
1108 Providing that operation is successful, the o !! 1109 is, as usual, an error indicator). This causes ``get_link()`` to be
1109 stack, and the new value is used as the ``nam !! 1110 called; it then gets the link from the filesystem. Providing that
>> 1111 operation is successful, the old path ``name`` is placed on the stack,
>> 1112 and the new value is used as the ``name`` for a while. When the end of
1110 the path is found (i.e. ``*name`` is ``'\0'`` 1113 the path is found (i.e. ``*name`` is ``'\0'``) the old ``name`` is restored
1111 off the stack and path walking continues. 1114 off the stack and path walking continues.
1112 1115
1113 Pushing and popping the reference pointers (i 1116 Pushing and popping the reference pointers (inode, cookie, etc.) is more
1114 complex in part because of the desire to hand 1117 complex in part because of the desire to handle tail recursion. When
1115 the last component of a symlink itself points 1118 the last component of a symlink itself points to a symlink, we
1116 want to pop the symlink-just-completed off th 1119 want to pop the symlink-just-completed off the stack before pushing
1117 the symlink-just-found to avoid leaving empty 1120 the symlink-just-found to avoid leaving empty path remnants that would
1118 just get in the way. 1121 just get in the way.
1119 1122
1120 It is most convenient to push the new symlink 1123 It is most convenient to push the new symlink references onto the
1121 stack in ``walk_component()`` immediately whe 1124 stack in ``walk_component()`` immediately when the symlink is found;
1122 ``walk_component()`` is also the last piece o 1125 ``walk_component()`` is also the last piece of code that needs to look at the
1123 old symlink as it walks that last component. 1126 old symlink as it walks that last component. So it is quite
1124 convenient for ``walk_component()`` to releas 1127 convenient for ``walk_component()`` to release the old symlink and pop
1125 the references just before pushing the refere 1128 the references just before pushing the reference information for the
1126 new symlink. It is guided in this by three f !! 1129 new symlink. It is guided in this by two flags; ``WALK_GET``, which
1127 forbids it from following a symlink if it fin !! 1130 gives it permission to follow a symlink if it finds one, and
1128 which indicates that it is yet too early to r !! 1131 ``WALK_PUT``, which tells it to release the current symlink after it has been
1129 current symlink, and ``WALK_TRAILING`` which !! 1132 followed. ``WALK_PUT`` is tested first, leading to a call to
1130 component of the lookup, so we will check use !! 1133 ``put_link()``. ``WALK_GET`` is tested subsequently (by
1131 decide whether follow it when it is a symlink !! 1134 ``should_follow_link()``) leading to a call to ``pick_link()`` which sets
1132 check if we have privilege to follow it. !! 1135 up the stack frame.
1133 1136
1134 Symlinks with no final component 1137 Symlinks with no final component
1135 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1138 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1136 1139
1137 A pair of special-case symlinks deserve a lit 1140 A pair of special-case symlinks deserve a little further explanation.
1138 Both result in a new ``struct path`` (with mo 1141 Both result in a new ``struct path`` (with mount and dentry) being set
1139 up in the ``nameidata``, and result in pick_l !! 1142 up in the ``nameidata``, and result in ``get_link()`` returning ``NULL``.
1140 1143
1141 The more obvious case is a symlink to "``/``" 1144 The more obvious case is a symlink to "``/``". All symlinks starting
1142 with "``/``" are detected in pick_link() whic !! 1145 with "``/``" are detected in ``get_link()`` which resets the ``nameidata``
1143 to point to the effective filesystem root. I 1146 to point to the effective filesystem root. If the symlink only
1144 contains "``/``" then there is nothing more t 1147 contains "``/``" then there is nothing more to do, no components at all,
1145 so ``NULL`` is returned to indicate that the 1148 so ``NULL`` is returned to indicate that the symlink can be released and
1146 the stack frame discarded. 1149 the stack frame discarded.
1147 1150
1148 The other case involves things in ``/proc`` t 1151 The other case involves things in ``/proc`` that look like symlinks but
1149 aren't really (and are therefore commonly ref !! 1152 aren't really::
1150 1153
1151 $ ls -l /proc/self/fd/1 1154 $ ls -l /proc/self/fd/1
1152 lrwx------ 1 neilb neilb 64 Jun 13 10:19 1155 lrwx------ 1 neilb neilb 64 Jun 13 10:19 /proc/self/fd/1 -> /dev/pts/4
1153 1156
1154 Every open file descriptor in any process is 1157 Every open file descriptor in any process is represented in ``/proc`` by
1155 something that looks like a symlink. It is r 1158 something that looks like a symlink. It is really a reference to the
1156 target file, not just the name of it. When y 1159 target file, not just the name of it. When you ``readlink`` these
1157 objects you get a name that might refer to th 1160 objects you get a name that might refer to the same file - unless it
1158 has been unlinked or mounted over. When ``wa 1161 has been unlinked or mounted over. When ``walk_component()`` follows
1159 one of these, the ``->get_link()`` method in !! 1162 one of these, the ``->follow_link()`` method in "procfs" doesn't return
1160 a string name, but instead calls nd_jump_link !! 1163 a string name, but instead calls ``nd_jump_link()`` which updates the
1161 ``nameidata`` in place to point to that targe !! 1164 ``nameidata`` in place to point to that target. ``->follow_link()`` then
1162 returns ``NULL``. Again there is no final co !! 1165 returns ``NULL``. Again there is no final component and ``get_link()``
1163 returns ``NULL``. !! 1166 reports this by leaving the ``last_type`` field of ``nameidata`` as
>> 1167 ``LAST_BIND``.
1164 1168
1165 Following the symlink in the final component 1169 Following the symlink in the final component
1166 -------------------------------------------- 1170 --------------------------------------------
1167 1171
1168 All this leads to ``link_path_walk()`` walkin 1172 All this leads to ``link_path_walk()`` walking down every component, and
1169 following all symbolic links it finds, until 1173 following all symbolic links it finds, until it reaches the final
1170 component. This is just returned in the ``la 1174 component. This is just returned in the ``last`` field of ``nameidata``.
1171 For some callers, this is all they need; they 1175 For some callers, this is all they need; they want to create that
1172 ``last`` name if it doesn't exist or give an 1176 ``last`` name if it doesn't exist or give an error if it does. Other
1173 callers will want to follow a symlink if one 1177 callers will want to follow a symlink if one is found, and possibly
1174 apply special handling to the last component 1178 apply special handling to the last component of that symlink, rather
1175 than just the last component of the original 1179 than just the last component of the original file name. These callers
1176 potentially need to call ``link_path_walk()`` 1180 potentially need to call ``link_path_walk()`` again and again on
1177 successive symlinks until one is found that d 1181 successive symlinks until one is found that doesn't point to another
1178 symlink. 1182 symlink.
1179 1183
1180 This case is handled by relevant callers of l !! 1184 This case is handled by the relevant caller of ``link_path_walk()``, such as
1181 path_lookupat(), path_openat() using a loop t !! 1185 ``path_lookupat()`` using a loop that calls ``link_path_walk()``, and then
1182 and then handles the final component by calli !! 1186 handles the final component. If the final component is a symlink
1183 lookup_last(). If it is a symlink that needs !! 1187 that needs to be followed, then ``trailing_symlink()`` is called to set
1184 open_last_lookups() or lookup_last() will set !! 1188 things up properly and the loop repeats, calling ``link_path_walk()``
1185 return the path so that the loop repeats, cal !! 1189 again. This could loop as many as 40 times if the last component of
1186 link_path_walk() again. This could loop as m !! 1190 each symlink is another symlink.
1187 component of each symlink is another symlink. !! 1191
1188 !! 1192 The various functions that examine the final component and possibly
1189 Of the various functions that examine the fin !! 1193 report that it is a symlink are ``lookup_last()``, ``mountpoint_last()``
1190 open_last_lookups() is the most interesting a !! 1194 and ``do_last()``, each of which use the same convention as
1191 with do_open() for opening a file. Part of o !! 1195 ``walk_component()`` of returning ``1`` if a symlink was found that needs
1192 with ``i_rwsem`` held and this part is in a s !! 1196 to be followed.
1193 !! 1197
1194 Explaining open_last_lookups() and do_open() !! 1198 Of these, ``do_last()`` is the most interesting as it is used for
1195 of this article, but a few highlights should !! 1199 opening a file. Part of ``do_last()`` runs with ``i_rwsem`` held and this
1196 the code. !! 1200 part is in a separate function: ``lookup_open()``.
>> 1201
>> 1202 Explaining ``do_last()`` completely is beyond the scope of this article,
>> 1203 but a few highlights should help those interested in exploring the
>> 1204 code.
1197 1205
1198 1. Rather than just finding the target file, !! 1206 1. Rather than just finding the target file, ``do_last()`` needs to open
1199 open_last_lookup() to open <<
1200 it. If the file was found in the dcache, 1207 it. If the file was found in the dcache, then ``vfs_open()`` is used for
1201 this. If not, then ``lookup_open()`` will 1208 this. If not, then ``lookup_open()`` will either call ``atomic_open()`` (if
1202 the filesystem provides it) to combine the 1209 the filesystem provides it) to combine the final lookup with the open, or
1203 will perform the separate ``i_op->lookup() !! 1210 will perform the separate ``lookup_real()`` and ``vfs_create()`` steps
1204 directly. In the later case the actual "o 1211 directly. In the later case the actual "open" of this newly found or
1205 created file will be performed by vfs_open !! 1212 created file will be performed by ``vfs_open()``, just as if the name
1206 were found in the dcache. 1213 were found in the dcache.
1207 1214
1208 2. vfs_open() can fail with ``-EOPENSTALE`` i !! 1215 2. ``vfs_open()`` can fail with ``-EOPENSTALE`` if the cached information
1209 wasn't quite current enough. If it's in R !! 1216 wasn't quite current enough. Rather than restarting the lookup from
1210 otherwise ``-ESTALE`` is returned. When ` !! 1217 the top with ``LOOKUP_REVAL`` set, ``lookup_open()`` is called instead,
1211 retry with ``LOOKUP_REVAL`` flag set. !! 1218 giving the filesystem a chance to resolve small inconsistencies.
>> 1219 If that doesn't work, only then is the lookup restarted from the top.
1212 1220
1213 3. An open with O_CREAT **does** follow a sym 1221 3. An open with O_CREAT **does** follow a symlink in the final component,
1214 unlike other creation system calls (like ` 1222 unlike other creation system calls (like ``mkdir``). So the sequence::
1215 1223
1216 ln -s bar /tmp/foo 1224 ln -s bar /tmp/foo
1217 echo hello > /tmp/foo 1225 echo hello > /tmp/foo
1218 1226
1219 will create a file called ``/tmp/bar``. T 1227 will create a file called ``/tmp/bar``. This is not permitted if
1220 ``O_EXCL`` is set but otherwise is handled 1228 ``O_EXCL`` is set but otherwise is handled for an O_CREAT open much
1221 like for a non-creating open: lookup_last( !! 1229 like for a non-creating open: ``should_follow_link()`` returns ``1``, and
1222 returns a non ``NULL`` value, and link_pat !! 1230 so does ``do_last()`` so that ``trailing_symlink()`` gets called and the
1223 open process continues on the symlink that 1231 open process continues on the symlink that was found.
1224 1232
1225 Updating the access time 1233 Updating the access time
1226 ------------------------ 1234 ------------------------
1227 1235
1228 We previously said of RCU-walk that it would 1236 We previously said of RCU-walk that it would "take no locks, increment
1229 no counts, leave no footprints." We have sin 1237 no counts, leave no footprints." We have since seen that some
1230 "footprints" can be needed when handling syml 1238 "footprints" can be needed when handling symlinks as a counted
1231 reference (or even a memory allocation) may b 1239 reference (or even a memory allocation) may be needed. But these
1232 footprints are best kept to a minimum. 1240 footprints are best kept to a minimum.
1233 1241
1234 One other place where walking down a symlink 1242 One other place where walking down a symlink can involve leaving
1235 footprints in a way that doesn't affect direc 1243 footprints in a way that doesn't affect directories is in updating access times.
1236 In Unix (and Linux) every filesystem object h 1244 In Unix (and Linux) every filesystem object has a "last accessed
1237 time", or "``atime``". Passing through a dir 1245 time", or "``atime``". Passing through a directory to access a file
1238 within is not considered to be an access for 1246 within is not considered to be an access for the purposes of
1239 ``atime``; only listing the contents of a dir 1247 ``atime``; only listing the contents of a directory can update its ``atime``.
1240 Symlinks are different it seems. Both readin 1248 Symlinks are different it seems. Both reading a symlink (with ``readlink()``)
1241 and looking up a symlink on the way to some o 1249 and looking up a symlink on the way to some other destination can
1242 update the atime on that symlink. 1250 update the atime on that symlink.
1243 1251
1244 .. _clearest statement: https://pubs.opengrou !! 1252 .. _clearest statement: http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap04.html#tag_04_08
1245 1253
1246 It is not clear why this is the case; POSIX h 1254 It is not clear why this is the case; POSIX has little to say on the
1247 subject. The `clearest statement`_ is that, 1255 subject. The `clearest statement`_ is that, if a particular implementation
1248 updates a timestamp in a place not specified 1256 updates a timestamp in a place not specified by POSIX, this must be
1249 documented "except that any changes caused by 1257 documented "except that any changes caused by pathname resolution need
1250 not be documented". This seems to imply that 1258 not be documented". This seems to imply that POSIX doesn't really
1251 care about access-time updates during pathnam 1259 care about access-time updates during pathname lookup.
1252 1260
1253 .. _Linux 1.3.87: https://git.kernel.org/cgit 1261 .. _Linux 1.3.87: https://git.kernel.org/cgit/linux/kernel/git/history/history.git/diff/fs/ext2/symlink.c?id=f806c6db77b8eaa6e00dcfb6b567706feae8dbb8
1254 1262
1255 An examination of history shows that prior to 1263 An examination of history shows that prior to `Linux 1.3.87`_, the ext2
1256 filesystem, at least, didn't update atime whe 1264 filesystem, at least, didn't update atime when following a link.
1257 Unfortunately we have no record of why that b 1265 Unfortunately we have no record of why that behavior was changed.
1258 1266
1259 In any case, access time must now be updated 1267 In any case, access time must now be updated and that operation can be
1260 quite complex. Trying to stay in RCU-walk wh 1268 quite complex. Trying to stay in RCU-walk while doing it is best
1261 avoided. Fortunately it is often permitted t 1269 avoided. Fortunately it is often permitted to skip the ``atime``
1262 update. Because ``atime`` updates cause perf 1270 update. Because ``atime`` updates cause performance problems in various
1263 areas, Linux supports the ``relatime`` mount 1271 areas, Linux supports the ``relatime`` mount option, which generally
1264 limits the updates of ``atime`` to once per d 1272 limits the updates of ``atime`` to once per day on files that aren't
1265 being changed (and symlinks never change once 1273 being changed (and symlinks never change once created). Even without
1266 ``relatime``, many filesystems record ``atime 1274 ``relatime``, many filesystems record ``atime`` with a one-second
1267 granularity, so only one update per second is 1275 granularity, so only one update per second is required.
1268 1276
1269 It is easy to test if an ``atime`` update is 1277 It is easy to test if an ``atime`` update is needed while in RCU-walk
1270 mode and, if it isn't, the update can be skip 1278 mode and, if it isn't, the update can be skipped and RCU-walk mode
1271 continues. Only when an ``atime`` update is 1279 continues. Only when an ``atime`` update is actually required does the
1272 path walk drop down to REF-walk. All of this 1280 path walk drop down to REF-walk. All of this is handled in the
1273 ``get_link()`` function. 1281 ``get_link()`` function.
1274 1282
1275 A few flags 1283 A few flags
1276 ----------- 1284 -----------
1277 1285
1278 A suitable way to wrap up this tour of pathna 1286 A suitable way to wrap up this tour of pathname walking is to list
1279 the various flags that can be stored in the ` 1287 the various flags that can be stored in the ``nameidata`` to guide the
1280 lookup process. Many of these are only meani 1288 lookup process. Many of these are only meaningful on the final
1281 component, others reflect the current state o !! 1289 component, others reflect the current state of the pathname lookup.
1282 apply restrictions to all path components enc <<
1283 <<
1284 And then there is ``LOOKUP_EMPTY``, which doe 1290 And then there is ``LOOKUP_EMPTY``, which doesn't fit conceptually with
1285 the others. If this is not set, an empty pat 1291 the others. If this is not set, an empty pathname causes an error
1286 very early on. If it is set, empty pathnames 1292 very early on. If it is set, empty pathnames are not considered to be
1287 an error. 1293 an error.
1288 1294
1289 Global state flags 1295 Global state flags
1290 ~~~~~~~~~~~~~~~~~~ 1296 ~~~~~~~~~~~~~~~~~~
1291 1297
1292 We have already met two global state flags: ` 1298 We have already met two global state flags: ``LOOKUP_RCU`` and
1293 ``LOOKUP_REVAL``. These select between one o 1299 ``LOOKUP_REVAL``. These select between one of three overall approaches
1294 to lookup: RCU-walk, REF-walk, and REF-walk w 1300 to lookup: RCU-walk, REF-walk, and REF-walk with forced revalidation.
1295 1301
1296 ``LOOKUP_PARENT`` indicates that the final co 1302 ``LOOKUP_PARENT`` indicates that the final component hasn't been reached
1297 yet. This is primarily used to tell the audi 1303 yet. This is primarily used to tell the audit subsystem the full
1298 context of a particular access being audited. 1304 context of a particular access being audited.
1299 1305
1300 ``ND_ROOT_PRESET`` indicates that the ``root` !! 1306 ``LOOKUP_ROOT`` indicates that the ``root`` field in the ``nameidata`` was
1301 provided by the caller, so it shouldn't be re 1307 provided by the caller, so it shouldn't be released when it is no
1302 longer needed. 1308 longer needed.
1303 1309
1304 ``ND_JUMPED`` means that the current dentry w !! 1310 ``LOOKUP_JUMPED`` means that the current dentry was chosen not because
1305 it had the right name but for some other reas 1311 it had the right name but for some other reason. This happens when
1306 following "``..``", following a symlink to `` 1312 following "``..``", following a symlink to ``/``, crossing a mount point
1307 or accessing a "``/proc/$PID/fd/$FD``" symlin !! 1313 or accessing a "``/proc/$PID/fd/$FD``" symlink. In this case the
1308 link"). In this case the filesystem has not b !! 1314 filesystem has not been asked to revalidate the name (with
1309 name (with ``d_revalidate()``). In such case !! 1315 ``d_revalidate()``). In such cases the inode may still need to be
1310 to be revalidated, so ``d_op->d_weak_revalida !! 1316 revalidated, so ``d_op->d_weak_revalidate()`` is called if
1311 ``ND_JUMPED`` is set when the look completes !! 1317 ``LOOKUP_JUMPED`` is set when the look completes - which may be at the
1312 final component or, when creating, unlinking, 1318 final component or, when creating, unlinking, or renaming, at the penultimate component.
1313 <<
1314 Resolution-restriction flags <<
1315 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ <<
1316 <<
1317 In order to allow userspace to protect itself <<
1318 and attack scenarios involving changing path <<
1319 available which apply restrictions to all pat <<
1320 path lookup. These flags are exposed through <<
1321 <<
1322 ``LOOKUP_NO_SYMLINKS`` blocks all symlink tra <<
1323 This is distinctly different from ``LOOKUP_FO <<
1324 relates to restricting the following of trail <<
1325 <<
1326 ``LOOKUP_NO_MAGICLINKS`` blocks all magic-lin <<
1327 ensure that they return errors from ``nd_jump <<
1328 ``LOOKUP_NO_MAGICLINKS`` and other magic-link <<
1329 <<
1330 ``LOOKUP_NO_XDEV`` blocks all ``vfsmount`` tr <<
1331 bind-mounts and ordinary mounts). Note that t <<
1332 lookup is determined by the first mountpoint <<
1333 absolute paths start with the ``vfsmount`` of <<
1334 with the ``dfd``'s ``vfsmount``. Magic-links <<
1335 ``vfsmount`` of the path is unchanged. <<
1336 <<
1337 ``LOOKUP_BENEATH`` blocks any path components <<
1338 starting point of the resolution. This is don <<
1339 as well as blocking ".." if it would jump out <<
1340 ``rename_lock`` and ``mount_lock`` are used t <<
1341 resolution of "..". Magic-links are also bloc <<
1342 <<
1343 ``LOOKUP_IN_ROOT`` resolves all path componen <<
1344 were the filesystem root. ``nd_jump_root()`` <<
1345 the starting point, and ".." at the starting <<
1346 ``LOOKUP_BENEATH``, ``rename_lock`` and ``mou <<
1347 attacks against ".." resolution. Magic-links <<
1348 1319
1349 Final-component flags 1320 Final-component flags
1350 ~~~~~~~~~~~~~~~~~~~~~ 1321 ~~~~~~~~~~~~~~~~~~~~~
1351 1322
1352 Some of these flags are only set when the fin 1323 Some of these flags are only set when the final component is being
1353 considered. Others are only checked for when 1324 considered. Others are only checked for when considering that final
1354 component. 1325 component.
1355 1326
1356 ``LOOKUP_AUTOMOUNT`` ensures that, if the fin 1327 ``LOOKUP_AUTOMOUNT`` ensures that, if the final component is an automount
1357 point, then the mount is triggered. Some ope 1328 point, then the mount is triggered. Some operations would trigger it
1358 anyway, but operations like ``stat()`` delibe 1329 anyway, but operations like ``stat()`` deliberately don't. ``statfs()``
1359 needs to trigger the mount but otherwise beha 1330 needs to trigger the mount but otherwise behaves a lot like ``stat()``, so
1360 it sets ``LOOKUP_AUTOMOUNT``, as does "``quot 1331 it sets ``LOOKUP_AUTOMOUNT``, as does "``quotactl()``" and the handling of
1361 "``mount --bind``". 1332 "``mount --bind``".
1362 1333
1363 ``LOOKUP_FOLLOW`` has a similar function to ` 1334 ``LOOKUP_FOLLOW`` has a similar function to ``LOOKUP_AUTOMOUNT`` but for
1364 symlinks. Some system calls set or clear it 1335 symlinks. Some system calls set or clear it implicitly, while
1365 others have API flags such as ``AT_SYMLINK_FO 1336 others have API flags such as ``AT_SYMLINK_FOLLOW`` and
1366 ``UMOUNT_NOFOLLOW`` to control it. Its effec 1337 ``UMOUNT_NOFOLLOW`` to control it. Its effect is similar to
1367 ``WALK_GET`` that we already met, but it is u 1338 ``WALK_GET`` that we already met, but it is used in a different way.
1368 1339
1369 ``LOOKUP_DIRECTORY`` insists that the final c 1340 ``LOOKUP_DIRECTORY`` insists that the final component is a directory.
1370 Various callers set this and it is also set w 1341 Various callers set this and it is also set when the final component
1371 is found to be followed by a slash. 1342 is found to be followed by a slash.
1372 1343
1373 Finally ``LOOKUP_OPEN``, ``LOOKUP_CREATE``, ` 1344 Finally ``LOOKUP_OPEN``, ``LOOKUP_CREATE``, ``LOOKUP_EXCL``, and
1374 ``LOOKUP_RENAME_TARGET`` are not used directl 1345 ``LOOKUP_RENAME_TARGET`` are not used directly by the VFS but are made
1375 available to the filesystem and particularly 1346 available to the filesystem and particularly the ``->d_revalidate()``
1376 method. A filesystem can choose not to bothe 1347 method. A filesystem can choose not to bother revalidating too hard
1377 if it knows that it will be asked to open or 1348 if it knows that it will be asked to open or create the file soon.
1378 These flags were previously useful for ``->lo 1349 These flags were previously useful for ``->lookup()`` too but with the
1379 introduction of ``->atomic_open()`` they are 1350 introduction of ``->atomic_open()`` they are less relevant there.
1380 1351
1381 End of the road 1352 End of the road
1382 --------------- 1353 ---------------
1383 1354
1384 Despite its complexity, all this pathname loo 1355 Despite its complexity, all this pathname lookup code appears to be
1385 in good shape - various parts are certainly e 1356 in good shape - various parts are certainly easier to understand now
1386 than even a couple of releases ago. But that 1357 than even a couple of releases ago. But that doesn't mean it is
1387 "finished". As already mentioned, RCU-walk 1358 "finished". As already mentioned, RCU-walk currently only follows
1388 symlinks that are stored in the inode so, whi 1359 symlinks that are stored in the inode so, while it handles many ext4
1389 symlinks, it doesn't help with NFS, XFS, or B 1360 symlinks, it doesn't help with NFS, XFS, or Btrfs. That support
1390 is not likely to be long delayed. 1361 is not likely to be long delayed.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.