1 ===============================
2 FS-CACHE NETWORK FILESYSTEM API
3 ===============================
5 There's an API by which a network filesystem can make use of the FS-Cache
6 facilities. This is based around a number of principles:
8 (1) Caches can store a number of different object types. There are two main
9 object types: indices and files. The first is a special type used by
10 FS-Cache to make finding objects faster and to make retiring of groups of
13 (2) Every index, file or other object is represented by a cookie. This cookie
14 may or may not have anything associated with it, but the netfs doesn't
17 (3) Barring the top-level index (one entry per cached netfs), the index
18 hierarchy for each netfs is structured according the whim of the netfs.
20 This API is declared in <linux/fscache.h>.
22 This document contains the following sections:
24 (1) Network filesystem definition
27 (4) Network filesystem (un)registration
29 (6) Index registration
30 (7) Data file registration
31 (8) Miscellaneous object registration
32 (9) Setting the data file size
33 (10) Page alloc/read/write
35 (12) Index and data file consistency
36 (13) Cookie enablement
37 (14) Miscellaneous cookie operations
38 (15) Cookie unregistration
39 (16) Index invalidation
40 (17) Data file invalidation
41 (18) FS-Cache specific page flags.
44 =============================
45 NETWORK FILESYSTEM DEFINITION
46 =============================
48 FS-Cache needs a description of the network filesystem. This is specified
49 using a record of the following structure:
51 struct fscache_netfs {
54 struct fscache_cookie *primary_index;
58 This first two fields should be filled in before registration, and the third
59 will be filled in by the registration function; any other fields should just be
60 ignored and are for internal use only.
64 (1) The name of the netfs (used as the key in the toplevel index).
66 (2) The version of the netfs (if the name matches but the version doesn't, the
67 entire in-cache hierarchy for this netfs will be scrapped and begun
70 (3) The cookie representing the primary index will be allocated according to
71 another parameter passed into the registration function.
73 For example, kAFS (linux/fs/afs/) uses the following definitions to describe
76 struct fscache_netfs afs_cache_netfs = {
86 Indices are used for two purposes:
88 (1) To aid the finding of a file based on a series of keys (such as AFS's
89 "cell", "volume ID", "vnode ID").
91 (2) To make it easier to discard a subset of all the files cached based around
92 a particular key - for instance to mirror the removal of an AFS volume.
94 However, since it's unlikely that any two netfs's are going to want to define
95 their index hierarchies in quite the same way, FS-Cache tries to impose as few
96 restraints as possible on how an index is structured and where it is placed in
97 the tree. The netfs can even mix indices and data files at the same level, but
100 Each index entry consists of a key of indeterminate length plus some auxiliary
101 data, also of indeterminate length.
103 There are some limits on indices:
105 (1) Any index containing non-index objects should be restricted to a single
106 cache. Any such objects created within an index will be created in the
107 first cache only. The cache in which an index is created can be
108 controlled by cache tags (see below).
110 (2) The entry data must be atomically journallable, so it is limited to about
111 400 bytes at present. At least 400 bytes will be available.
113 (3) The depth of the index tree should be judged with care as the search
114 function is recursive. Too many layers will run the kernel out of stack.
121 To define an object, a structure of the following type should be filled out:
123 struct fscache_cookie_def
128 struct fscache_cache_tag *(*select_cache)(
129 const void *parent_netfs_data,
130 const void *cookie_netfs_data);
132 enum fscache_checkaux (*check_aux)(void *cookie_netfs_data,
137 void (*get_context)(void *cookie_netfs_data, void *context);
139 void (*put_context)(void *cookie_netfs_data, void *context);
141 void (*mark_pages_cached)(void *cookie_netfs_data,
142 struct address_space *mapping,
143 struct pagevec *cached_pvec);
146 This has the following fields:
148 (1) The type of the object [mandatory].
150 This is one of the following values:
152 (*) FSCACHE_COOKIE_TYPE_INDEX
154 This defines an index, which is a special FS-Cache type.
156 (*) FSCACHE_COOKIE_TYPE_DATAFILE
158 This defines an ordinary data file.
160 (*) Any other value between 2 and 255
162 This defines an extraordinary object such as an XATTR.
164 (2) The name of the object type (NUL terminated unless all 16 chars are used)
167 (3) A function to select the cache in which to store an index [optional].
169 This function is invoked when an index needs to be instantiated in a cache
170 during the instantiation of a non-index object. Only the immediate index
171 parent for the non-index object will be queried. Any indices above that
172 in the hierarchy may be stored in multiple caches. This function does not
173 need to be supplied for any non-index object or any index that will only
176 If this function is not supplied or if it returns NULL then the first
177 cache in the parent's list will be chosen, or failing that, the first
178 cache in the master list.
180 (4) A function to check the auxiliary data [optional].
182 This function will be called to check that a match found in the cache for
183 this object is valid. For instance with AFS it could check the auxiliary
184 data against the data version number returned by the server to determine
185 whether the index entry in a cache is still valid.
187 If this function is absent, it will be assumed that matching objects in a
188 cache are always valid.
190 The function is also passed the cache's idea of the object size and may
191 use this to manage coherency also.
193 If present, the function should return one of the following values:
195 (*) FSCACHE_CHECKAUX_OKAY - the entry is okay as is
196 (*) FSCACHE_CHECKAUX_NEEDS_UPDATE - the entry requires update
197 (*) FSCACHE_CHECKAUX_OBSOLETE - the entry should be deleted
199 This function can also be used to extract data from the auxiliary data in
200 the cache and copy it into the netfs's structures.
202 (5) A pair of functions to manage contexts for the completion callback
205 The cache read/write functions are passed a context which is then passed
206 to the I/O completion callback function. To ensure this context remains
207 valid until after the I/O completion is called, two functions may be
208 provided: one to get an extra reference on the context, and one to drop a
211 If the context is not used or is a type of object that won't go out of
212 scope, then these functions are not required. These functions are not
213 required for indices as indices may not contain data. These functions may
214 be called in interrupt context and so may not sleep.
216 (6) A function to mark a page as retaining cache metadata [optional].
218 This is called by the cache to indicate that it is retaining in-memory
219 information for this page and that the netfs should uncache the page when
220 it has finished. This does not indicate whether there's data on the disk
221 or not. Note that several pages at once may be presented for marking.
223 The PG_fscache bit is set on the pages before this function would be
224 called, so the function need not be provided if this is sufficient.
226 This function is not required for indices as they're not permitted data.
228 (7) A function to unmark all the pages retaining cache metadata [mandatory].
230 This is called by FS-Cache to indicate that a backing store is being
231 unbound from a cookie and that all the marks on the pages should be
232 cleared to prevent confusion. Note that the cache will have torn down all
233 its tracking information so that the pages don't need to be explicitly
236 This function is not required for indices as they're not permitted data.
239 ===================================
240 NETWORK FILESYSTEM (UN)REGISTRATION
241 ===================================
243 The first step is to declare the network filesystem to the cache. This also
244 involves specifying the layout of the primary index (for AFS, this would be the
247 The registration function is:
249 int fscache_register_netfs(struct fscache_netfs *netfs);
251 It just takes a pointer to the netfs definition. It returns 0 or an error as
254 For kAFS, registration is done as follows:
256 ret = fscache_register_netfs(&afs_cache_netfs);
258 The last step is, of course, unregistration:
260 void fscache_unregister_netfs(struct fscache_netfs *netfs);
267 FS-Cache permits the use of more than one cache. To permit particular index
268 subtrees to be bound to particular caches, the second step is to look up cache
269 representation tags. This step is optional; it can be left entirely up to
270 FS-Cache as to which cache should be used. The problem with doing that is that
271 FS-Cache will always pick the first cache that was registered.
273 To get the representation for a named tag:
275 struct fscache_cache_tag *fscache_lookup_cache_tag(const char *name);
277 This takes a text string as the name and returns a representation of a tag. It
278 will never return an error. It may return a dummy tag, however, if it runs out
279 of memory; this will inhibit caching with this tag.
281 Any representation so obtained must be released by passing it to this function:
283 void fscache_release_cache_tag(struct fscache_cache_tag *tag);
285 The tag will be retrieved by FS-Cache when it calls the object definition
286 operation select_cache().
293 The third step is to inform FS-Cache about part of an index hierarchy that can
294 be used to locate files. This is done by requesting a cookie for each index in
295 the path to the file:
297 struct fscache_cookie *
298 fscache_acquire_cookie(struct fscache_cookie *parent,
299 const struct fscache_object_def *def,
300 const void *index_key,
301 size_t index_key_len,
302 const void *aux_data,
308 This function creates an index entry in the index represented by parent,
309 filling in the index entry by calling the operations pointed to by def.
311 A unique key that represents the object within the parent must be pointed to by
312 index_key and is of length index_key_len.
314 An optional blob of auxiliary data that is to be stored within the cache can be
315 pointed to with aux_data and should be of length aux_data_len. This would
316 typically be used for storing coherency data.
318 The netfs may pass an arbitrary value in netfs_data and this will be presented
319 to it in the event of any calling back. This may also be used in tracing or
322 The cache tracks the size of the data attached to an object and this set to be
323 object_size. For indices, this should be 0. This value will be passed to the
324 ->check_aux() callback.
326 Note that this function never returns an error - all errors are handled
327 internally. It may, however, return NULL to indicate no cookie. It is quite
328 acceptable to pass this token back to this function as the parent to another
329 acquisition (or even to the relinquish cookie, read page and write page
330 functions - see below).
332 Note also that no indices are actually created in a cache until a non-index
333 object needs to be created somewhere down the hierarchy. Furthermore, an index
334 may be created in several different caches independently at different times.
335 This is all handled transparently, and the netfs doesn't see any of it.
337 A cookie will be created in the disabled state if enabled is false. A cookie
338 must be enabled to do anything with it. A disabled cookie can be enabled by
339 calling fscache_enable_cookie() (see below).
341 For example, with AFS, a cell would be added to the primary index. This index
342 entry would have a dependent inode containing volume mappings within this cell:
345 fscache_acquire_cookie(afs_cache_netfs.primary_index,
346 &afs_cell_cache_index_def,
347 cell->name, strlen(cell->name),
351 And then a particular volume could be added to that index by ID, creating
352 another index for vnodes (AFS inode equivalents):
355 fscache_acquire_cookie(volume->cell->cache,
356 &afs_volume_cache_index_def,
357 &volume->vid, sizeof(volume->vid),
362 ======================
363 DATA FILE REGISTRATION
364 ======================
366 The fourth step is to request a data file be created in the cache. This is
367 identical to index cookie acquisition. The only difference is that the type in
368 the object definition should be something other than index type.
371 fscache_acquire_cookie(volume->cache,
372 &afs_vnode_cache_object_def,
375 vnode, vnode->status.size, true);
378 =================================
379 MISCELLANEOUS OBJECT REGISTRATION
380 =================================
382 An optional step is to request an object of miscellaneous type be created in
383 the cache. This is almost identical to index cookie acquisition. The only
384 difference is that the type in the object definition should be something other
385 than index type. Whilst the parent object could be an index, it's more likely
386 it would be some other type of object such as a data file.
389 fscache_acquire_cookie(vnode->cache,
390 &afs_xattr_cache_object_def,
391 &xattr->name, strlen(xattr->name),
393 xattr, strlen(xattr->val), true);
395 Miscellaneous objects might be used to store extended attributes or directory
399 ==========================
400 SETTING THE DATA FILE SIZE
401 ==========================
403 The fifth step is to set the physical attributes of the file, such as its size.
404 This doesn't automatically reserve any space in the cache, but permits the
405 cache to adjust its metadata for data tracking appropriately:
407 int fscache_attr_changed(struct fscache_cookie *cookie);
409 The cache will return -ENOBUFS if there is no backing cache or if there is no
410 space to allocate any extra metadata required in the cache.
412 Note that attempts to read or write data pages in the cache over this size may
413 be rebuffed with -ENOBUFS.
415 This operation schedules an attribute adjustment to happen asynchronously at
416 some point in the future, and as such, it may happen after the function returns
417 to the caller. The attribute adjustment excludes read and write operations.
420 =====================
421 PAGE ALLOC/READ/WRITE
422 =====================
424 And the sixth step is to store and retrieve pages in the cache. There are
425 three functions that are used to do this.
429 (1) A page should not be re-read or re-allocated without uncaching it first.
431 (2) A read or allocated page must be uncached when the netfs page is released
434 (3) A page should only be written to the cache if previous read or allocated.
436 This permits the cache to maintain its page tracking in proper order.
442 Firstly, the netfs should ask FS-Cache to examine the caches and read the
443 contents cached for a particular page of a particular file if present, or else
444 allocate space to store the contents if not:
447 void (*fscache_rw_complete_t)(struct page *page,
451 int fscache_read_or_alloc_page(struct fscache_cookie *cookie,
453 fscache_rw_complete_t end_io_func,
457 The cookie argument must specify a cookie for an object that isn't an index,
458 the page specified will have the data loaded into it (and is also used to
459 specify the page number), and the gfp argument is used to control how any
460 memory allocations made are satisfied.
462 If the cookie indicates the inode is not cached:
464 (1) The function will return -ENOBUFS.
466 Else if there's a copy of the page resident in the cache:
468 (1) The mark_pages_cached() cookie operation will be called on that page.
470 (2) The function will submit a request to read the data from the cache's
471 backing device directly into the page specified.
473 (3) The function will return 0.
475 (4) When the read is complete, end_io_func() will be invoked with:
477 (*) The netfs data supplied when the cookie was created.
479 (*) The page descriptor.
481 (*) The context argument passed to the above function. This will be
482 maintained with the get_context/put_context functions mentioned above.
484 (*) An argument that's 0 on success or negative for an error code.
486 If an error occurs, it should be assumed that the page contains no usable
487 data. fscache_readpages_cancel() may need to be called.
489 end_io_func() will be called in process context if the read is results in
490 an error, but it might be called in interrupt context if the read is
493 Otherwise, if there's not a copy available in cache, but the cache may be able
496 (1) The mark_pages_cached() cookie operation will be called on that page.
498 (2) A block may be reserved in the cache and attached to the object at the
501 (3) The function will return -ENODATA.
503 This function may also return -ENOMEM or -EINTR, in which case it won't have
504 read any data from the cache.
510 Alternatively, if there's not expected to be any data in the cache for a page
511 because the file has been extended, a block can simply be allocated instead:
513 int fscache_alloc_page(struct fscache_cookie *cookie,
517 This is similar to the fscache_read_or_alloc_page() function, except that it
518 never reads from the cache. It will return 0 if a block has been allocated,
519 rather than -ENODATA as the other would. One or the other must be performed
520 before writing to the cache.
522 The mark_pages_cached() cookie operation will be called on the page if
529 Secondly, if the netfs changes the contents of the page (either due to an
530 initial download or if a user performs a write), then the page should be
531 written back to the cache:
533 int fscache_write_page(struct fscache_cookie *cookie,
538 The cookie argument must specify a data file cookie, the page specified should
539 contain the data to be written (and is also used to specify the page number),
540 object_size is the revised size of the object and the gfp argument is used to
541 control how any memory allocations made are satisfied.
543 The page must have first been read or allocated successfully and must not have
544 been uncached before writing is performed.
546 If the cookie indicates the inode is not cached then:
548 (1) The function will return -ENOBUFS.
550 Else if space can be allocated in the cache to hold this page:
552 (1) PG_fscache_write will be set on the page.
554 (2) The function will submit a request to write the data to cache's backing
555 device directly from the page specified.
557 (3) The function will return 0.
559 (4) When the write is complete PG_fscache_write is cleared on the page and
560 anyone waiting for that bit will be woken up.
562 Else if there's no space available in the cache, -ENOBUFS will be returned. It
563 is also possible for the PG_fscache_write bit to be cleared when no write took
564 place if unforeseen circumstances arose (such as a disk error).
566 Writing takes place asynchronously.
572 A facility is provided to read several pages at once, as requested by the
573 readpages() address space operation:
575 int fscache_read_or_alloc_pages(struct fscache_cookie *cookie,
576 struct address_space *mapping,
577 struct list_head *pages,
579 fscache_rw_complete_t end_io_func,
583 This works in a similar way to fscache_read_or_alloc_page(), except:
585 (1) Any page it can retrieve data for is removed from pages and nr_pages and
586 dispatched for reading to the disk. Reads of adjacent pages on disk may
587 be merged for greater efficiency.
589 (2) The mark_pages_cached() cookie operation will be called on several pages
590 at once if they're being read or allocated.
592 (3) If there was an general error, then that error will be returned.
594 Else if some pages couldn't be allocated or read, then -ENOBUFS will be
597 Else if some pages couldn't be read but were allocated, then -ENODATA will
600 Otherwise, if all pages had reads dispatched, then 0 will be returned, the
601 list will be empty and *nr_pages will be 0.
603 (4) end_io_func will be called once for each page being read as the reads
604 complete. It will be called in process context if error != 0, but it may
605 be called in interrupt context if there is no error.
607 Note that a return of -ENODATA, -ENOBUFS or any other error does not preclude
608 some of the pages being read and some being allocated. Those pages will have
609 been marked appropriately and will need uncaching.
612 CANCELLATION OF UNREAD PAGES
613 ----------------------------
615 If one or more pages are passed to fscache_read_or_alloc_pages() but not then
616 read from the cache and also not read from the underlying filesystem then
617 those pages will need to have any marks and reservations removed. This can be
620 void fscache_readpages_cancel(struct fscache_cookie *cookie,
621 struct list_head *pages);
623 prior to returning to the caller. The cookie argument should be as passed to
624 fscache_read_or_alloc_pages(). Every page in the pages list will be examined
625 and any that have PG_fscache set will be uncached.
632 To uncache a page, this function should be called:
634 void fscache_uncache_page(struct fscache_cookie *cookie,
637 This function permits the cache to release any in-memory representation it
638 might be holding for this netfs page. This function must be called once for
639 each page on which the read or write page functions above have been called to
640 make sure the cache's in-memory tracking information gets torn down.
642 Note that pages can't be explicitly deleted from the a data file. The whole
643 data file must be retired (see the relinquish cookie function below).
645 Furthermore, note that this does not cancel the asynchronous read or write
646 operation started by the read/alloc and write functions, so the page
647 invalidation functions must use:
649 bool fscache_check_page_write(struct fscache_cookie *cookie,
652 to see if a page is being written to the cache, and:
654 void fscache_wait_on_page_write(struct fscache_cookie *cookie,
657 to wait for it to finish if it is.
660 When releasepage() is being implemented, a special FS-Cache function exists to
661 manage the heuristics of coping with vmscan trying to eject pages, which may
662 conflict with the cache trying to write pages to the cache (which may itself
663 need to allocate memory):
665 bool fscache_maybe_release_page(struct fscache_cookie *cookie,
669 This takes the netfs cookie, and the page and gfp arguments as supplied to
670 releasepage(). It will return false if the page cannot be released yet for
671 some reason and if it returns true, the page has been uncached and can now be
674 To make a page available for release, this function may wait for an outstanding
675 storage request to complete, or it may attempt to cancel the storage request -
676 in which case the page will not be stored in the cache this time.
679 BULK INODE PAGE UNCACHE
680 -----------------------
682 A convenience routine is provided to perform an uncache on all the pages
683 attached to an inode. This assumes that the pages on the inode correspond on a
684 1:1 basis with the pages in the cache.
686 void fscache_uncache_all_inode_pages(struct fscache_cookie *cookie,
687 struct inode *inode);
689 This takes the netfs cookie that the pages were cached with and the inode that
690 the pages are attached to. This function will wait for pages to finish being
691 written to the cache and for the cache to finish with the page generally. No
695 ===============================
696 INDEX AND DATA FILE CONSISTENCY
697 ===============================
699 To find out whether auxiliary data for an object is up to data within the
700 cache, the following function can be called:
702 int fscache_check_consistency(struct fscache_cookie *cookie,
703 const void *aux_data);
705 This will call back to the netfs to check whether the auxiliary data associated
706 with a cookie is correct; if aux_data is non-NULL, it will update the auxiliary
707 data buffer first. It returns 0 if it is and -ESTALE if it isn't; it may also
708 return -ENOMEM and -ERESTARTSYS.
710 To request an update of the index data for an index or other object, the
711 following function should be called:
713 void fscache_update_cookie(struct fscache_cookie *cookie,
714 const void *aux_data);
716 This function will update the cookie's auxiliary data buffer from aux_data if
717 that is non-NULL and then schedule this to be stored on disk. The update
718 method in the parent index definition will be called to transfer the data.
720 Note that partial updates may happen automatically at other times, such as when
721 data blocks are added to a data file object.
728 Cookies exist in one of two states: enabled and disabled. If a cookie is
729 disabled, it ignores all attempts to acquire child cookies; check, update or
730 invalidate its state; allocate, read or write backing pages - though it is
731 still possible to uncache pages and relinquish the cookie.
733 The initial enablement state is set by fscache_acquire_cookie(), but the cookie
734 can be enabled or disabled later. To disable a cookie, call:
736 void fscache_disable_cookie(struct fscache_cookie *cookie,
737 const void *aux_data,
740 If the cookie is not already disabled, this locks the cookie against other
741 enable and disable ops, marks the cookie as being disabled, discards or
742 invalidates any backing objects and waits for cessation of activity on any
743 associated object before unlocking the cookie.
745 All possible failures are handled internally. The caller should consider
746 calling fscache_uncache_all_inode_pages() afterwards to make sure all page
747 markings are cleared up.
749 Cookies can be enabled or reenabled with:
751 void fscache_enable_cookie(struct fscache_cookie *cookie,
752 const void *aux_data,
754 bool (*can_enable)(void *data),
757 If the cookie is not already enabled, this locks the cookie against other
758 enable and disable ops, invokes can_enable() and, if the cookie is not an index
759 cookie, will begin the procedure of acquiring backing objects.
761 The optional can_enable() function is passed the data argument and returns a
762 ruling as to whether or not enablement should actually be permitted to begin.
764 All possible failures are handled internally. The cookie will only be marked
765 as enabled if provisional backing objects are allocated.
767 The object's data size is updated from object_size and is passed to the
768 ->check_aux() function.
770 In both cases, the cookie's auxiliary data buffer is updated from aux_data if
771 that is non-NULL inside the enablement lock before proceeding.
774 ===============================
775 MISCELLANEOUS COOKIE OPERATIONS
776 ===============================
778 There are a number of operations that can be used to control cookies:
782 int fscache_pin_cookie(struct fscache_cookie *cookie);
783 void fscache_unpin_cookie(struct fscache_cookie *cookie);
785 These operations permit data cookies to be pinned into the cache and to
786 have the pinning removed. They are not permitted on index cookies.
788 The pinning function will return 0 if successful, -ENOBUFS in the cookie
789 isn't backed by a cache, -EOPNOTSUPP if the cache doesn't support pinning,
790 -ENOSPC if there isn't enough space to honour the operation, -ENOMEM or
791 -EIO if there's any other problem.
793 (*) Data space reservation:
795 int fscache_reserve_space(struct fscache_cookie *cookie, loff_t size);
797 This permits a netfs to request cache space be reserved to store up to the
798 given amount of a file. It is permitted to ask for more than the current
799 size of the file to allow for future file expansion.
801 If size is given as zero then the reservation will be cancelled.
803 The function will return 0 if successful, -ENOBUFS in the cookie isn't
804 backed by a cache, -EOPNOTSUPP if the cache doesn't support reservations,
805 -ENOSPC if there isn't enough space to honour the operation, -ENOMEM or
806 -EIO if there's any other problem.
808 Note that this doesn't pin an object in a cache; it can still be culled to
809 make space if it's not in use.
812 =====================
813 COOKIE UNREGISTRATION
814 =====================
816 To get rid of a cookie, this function should be called.
818 void fscache_relinquish_cookie(struct fscache_cookie *cookie,
819 const void *aux_data,
822 If retire is non-zero, then the object will be marked for recycling, and all
823 copies of it will be removed from all active caches in which it is present.
824 Not only that but all child objects will also be retired.
826 If retire is zero, then the object may be available again when next the
827 acquisition function is called. Retirement here will overrule the pinning on a
830 The cookie's auxiliary data will be updated from aux_data if that is non-NULL
831 so that the cache can lazily update it on disk.
833 One very important note - relinquish must NOT be called for a cookie unless all
834 the cookies for "child" indices, objects and pages have been relinquished
842 There is no direct way to invalidate an index subtree. To do this, the caller
843 should relinquish and retire the cookie they have, and then acquire a new one.
846 ======================
847 DATA FILE INVALIDATION
848 ======================
850 Sometimes it will be necessary to invalidate an object that contains data.
851 Typically this will be necessary when the server tells the netfs of a foreign
852 change - at which point the netfs has to throw away all the state it had for an
853 inode and reload from the server.
855 To indicate that a cache object should be invalidated, the following function
858 void fscache_invalidate(struct fscache_cookie *cookie);
860 This can be called with spinlocks held as it defers the work to a thread pool.
861 All extant storage, retrieval and attribute change ops at this point are
862 cancelled and discarded. Some future operations will be rejected until the
863 cache has had a chance to insert a barrier in the operations queue. After
864 that, operations will be queued again behind the invalidation operation.
866 The invalidation operation will perform an attribute change operation and an
867 auxiliary data update operation as it is very likely these will have changed.
869 Using the following function, the netfs can wait for the invalidation operation
870 to have reached a point at which it can start submitting ordinary operations
873 void fscache_wait_on_invalidate(struct fscache_cookie *cookie);
876 ===========================
877 FS-CACHE SPECIFIC PAGE FLAG
878 ===========================
880 FS-Cache makes use of a page flag, PG_private_2, for its own purpose. This is
881 given the alternative name PG_fscache.
883 PG_fscache is used to indicate that the page is known by the cache, and that
884 the cache must be informed if the page is going to go away. It's an indication
885 to the netfs that the cache has an interest in this page, where an interest may
886 be a pointer to it, resources allocated or reserved for it, or I/O in progress
889 The netfs can use this information in methods such as releasepage() to
890 determine whether it needs to uncache a page or update it.
892 Furthermore, if this bit is set, releasepage() and invalidatepage() operations
893 will be called on a page to get rid of it, even if PG_private is not set. This
894 allows caching to attempted on a page before read_cache_pages() to be called
895 after fscache_read_or_alloc_pages() as the former will try and release pages it
896 was given under certain circumstances.
898 This bit does not overlap with such as PG_private. This means that FS-Cache
899 can be used with a filesystem that uses the block buffering code.
901 There are a number of operations defined on this flag:
903 int PageFsCache(struct page *page);
904 void SetPageFsCache(struct page *page)
905 void ClearPageFsCache(struct page *page)
906 int TestSetPageFsCache(struct page *page)
907 int TestClearPageFsCache(struct page *page)
909 These functions are bit test, bit set, bit clear, bit test and set and bit
910 test and clear operations on PG_fscache.