From a5b4bd2874d9032b42db8cc4880058576c561b06 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 24 Sep 2013 10:35:14 +0100 Subject: KEYS: Use bool in make_key_ref() and is_key_possessed() Make make_key_ref() take a bool possession parameter and make is_key_possessed() return a bool. Signed-off-by: David Howells --- Documentation/security/keys.txt | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'Documentation') diff --git a/Documentation/security/keys.txt b/Documentation/security/keys.txt index 7b4145d00452..9ede67084f0b 100644 --- a/Documentation/security/keys.txt +++ b/Documentation/security/keys.txt @@ -865,15 +865,14 @@ encountered: calling processes has a searchable link to the key from one of its keyrings. There are three functions for dealing with these: - key_ref_t make_key_ref(const struct key *key, - unsigned long possession); + key_ref_t make_key_ref(const struct key *key, bool possession); struct key *key_ref_to_ptr(const key_ref_t key_ref); - unsigned long is_key_possessed(const key_ref_t key_ref); + bool is_key_possessed(const key_ref_t key_ref); The first function constructs a key reference from a key pointer and - possession information (which must be 0 or 1 and not any other value). + possession information (which must be true or false). The second function retrieves the key pointer from a reference and the third retrieves the possession flag. -- cgit v1.2.3 From ccc3e6d9c9aea07a0b60b2b0bfc5b05a704b66d5 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 24 Sep 2013 10:35:16 +0100 Subject: KEYS: Define a __key_get() wrapper to use rather than atomic_inc() Define a __key_get() wrapper to use rather than atomic_inc() on the key usage count as this makes it easier to hook in refcount error debugging. Signed-off-by: David Howells --- Documentation/security/keys.txt | 13 ++++++++----- include/linux/key.h | 10 +++++++--- security/keys/key.c | 2 +- security/keys/keyring.c | 6 +++--- security/keys/process_keys.c | 16 ++++++++-------- 5 files changed, 27 insertions(+), 20 deletions(-) (limited to 'Documentation') diff --git a/Documentation/security/keys.txt b/Documentation/security/keys.txt index 9ede67084f0b..a4c33f1a7c6d 100644 --- a/Documentation/security/keys.txt +++ b/Documentation/security/keys.txt @@ -960,14 +960,17 @@ payload contents" for more information. the argument will not be parsed. -(*) Extra references can be made to a key by calling the following function: +(*) Extra references can be made to a key by calling one of the following + functions: + struct key *__key_get(struct key *key); struct key *key_get(struct key *key); - These need to be disposed of by calling key_put() when they've been - finished with. The key pointer passed in will be returned. If the pointer - is NULL or CONFIG_KEYS is not set then the key will not be dereferenced and - no increment will take place. + Keys so references will need to be disposed of by calling key_put() when + they've been finished with. The key pointer passed in will be returned. + + In the case of key_get(), if the pointer is NULL or CONFIG_KEYS is not set + then the key will not be dereferenced and no increment will take place. (*) A key's serial number can be obtained by calling: diff --git a/include/linux/key.h b/include/linux/key.h index d573e820a23d..ef596c7af585 100644 --- a/include/linux/key.h +++ b/include/linux/key.h @@ -219,13 +219,17 @@ extern void key_revoke(struct key *key); extern void key_invalidate(struct key *key); extern void key_put(struct key *key); -static inline struct key *key_get(struct key *key) +static inline struct key *__key_get(struct key *key) { - if (key) - atomic_inc(&key->usage); + atomic_inc(&key->usage); return key; } +static inline struct key *key_get(struct key *key) +{ + return key ? __key_get(key) : key; +} + static inline void key_ref_put(key_ref_t key_ref) { key_put(key_ref_to_ptr(key_ref)); diff --git a/security/keys/key.c b/security/keys/key.c index 7e6bc396bb23..1e23cc288106 100644 --- a/security/keys/key.c +++ b/security/keys/key.c @@ -644,7 +644,7 @@ found: /* this races with key_put(), but that doesn't matter since key_put() * doesn't actually change the key */ - atomic_inc(&key->usage); + __key_get(key); error: spin_unlock(&key_serial_lock); diff --git a/security/keys/keyring.c b/security/keys/keyring.c index b42f2d4f7f83..87eff32b53f4 100644 --- a/security/keys/keyring.c +++ b/security/keys/keyring.c @@ -479,7 +479,7 @@ not_this_keyring: /* we found a viable match */ found: - atomic_inc(&key->usage); + __key_get(key); key->last_used_at = ctx->now.tv_sec; keyring->last_used_at = ctx->now.tv_sec; while (sp > 0) @@ -573,7 +573,7 @@ key_ref_t __keyring_search_one(key_ref_t keyring_ref, return ERR_PTR(-ENOKEY); found: - atomic_inc(&key->usage); + __key_get(key); keyring->last_used_at = key->last_used_at = current_kernel_time().tv_sec; rcu_read_unlock(); @@ -909,7 +909,7 @@ void __key_link(struct key *keyring, struct key *key, klist = rcu_dereference_locked_keyring(keyring); - atomic_inc(&key->usage); + __key_get(key); keyring->last_used_at = key->last_used_at = current_kernel_time().tv_sec; diff --git a/security/keys/process_keys.c b/security/keys/process_keys.c index e68a3e0e7aa0..68548ea6fe01 100644 --- a/security/keys/process_keys.c +++ b/security/keys/process_keys.c @@ -235,7 +235,7 @@ int install_session_keyring_to_cred(struct cred *cred, struct key *keyring) if (IS_ERR(keyring)) return PTR_ERR(keyring); } else { - atomic_inc(&keyring->usage); + __key_get(keyring); } /* install the keyring */ @@ -544,7 +544,7 @@ try_again: } key = ctx.cred->thread_keyring; - atomic_inc(&key->usage); + __key_get(key); key_ref = make_key_ref(key, 1); break; @@ -562,7 +562,7 @@ try_again: } key = ctx.cred->process_keyring; - atomic_inc(&key->usage); + __key_get(key); key_ref = make_key_ref(key, 1); break; @@ -593,7 +593,7 @@ try_again: rcu_read_lock(); key = rcu_dereference(ctx.cred->session_keyring); - atomic_inc(&key->usage); + __key_get(key); rcu_read_unlock(); key_ref = make_key_ref(key, 1); break; @@ -606,7 +606,7 @@ try_again: } key = ctx.cred->user->uid_keyring; - atomic_inc(&key->usage); + __key_get(key); key_ref = make_key_ref(key, 1); break; @@ -618,7 +618,7 @@ try_again: } key = ctx.cred->user->session_keyring; - atomic_inc(&key->usage); + __key_get(key); key_ref = make_key_ref(key, 1); break; @@ -632,7 +632,7 @@ try_again: if (!key) goto error; - atomic_inc(&key->usage); + __key_get(key); key_ref = make_key_ref(key, 1); break; @@ -648,7 +648,7 @@ try_again: } else { rka = ctx.cred->request_key_auth->payload.data; key = rka->dest_keyring; - atomic_inc(&key->usage); + __key_get(key); } up_read(&ctx.cred->request_key_auth->sem); if (!key) -- cgit v1.2.3 From 3cb989501c2688cacbb7dc4b0d353faf838f53a1 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 24 Sep 2013 10:35:17 +0100 Subject: Add a generic associative array implementation. Add a generic associative array implementation that can be used as the container for keyrings, thereby massively increasing the capacity available whilst also speeding up searching in keyrings that contain a lot of keys. This may also be useful in FS-Cache for tracking cookies. Documentation is added into Documentation/associative_array.txt Some of the properties of the implementation are: (1) Objects are opaque pointers. The implementation does not care where they point (if anywhere) or what they point to (if anything). [!] NOTE: Pointers to objects _must_ be zero in the two least significant bits. (2) Objects do not need to contain linkage blocks for use by the array. This permits an object to be located in multiple arrays simultaneously. Rather, the array is made up of metadata blocks that point to objects. (3) Objects are labelled as being one of two types (the type is a bool value). This information is stored in the array, but has no consequence to the array itself or its algorithms. (4) Objects require index keys to locate them within the array. (5) Index keys must be unique. Inserting an object with the same key as one already in the array will replace the old object. (6) Index keys can be of any length and can be of different lengths. (7) Index keys should encode the length early on, before any variation due to length is seen. (8) Index keys can include a hash to scatter objects throughout the array. (9) The array can iterated over. The objects will not necessarily come out in key order. (10) The array can be iterated whilst it is being modified, provided the RCU readlock is being held by the iterator. Note, however, under these circumstances, some objects may be seen more than once. If this is a problem, the iterator should lock against modification. Objects will not be missed, however, unless deleted. (11) Objects in the array can be looked up by means of their index key. (12) Objects can be looked up whilst the array is being modified, provided the RCU readlock is being held by the thread doing the look up. The implementation uses a tree of 16-pointer nodes internally that are indexed on each level by nibbles from the index key. To improve memory efficiency, shortcuts can be emplaced to skip over what would otherwise be a series of single-occupancy nodes. Further, nodes pack leaf object pointers into spare space in the node rather than making an extra branch until as such time an object needs to be added to a full node. Signed-off-by: David Howells --- Documentation/assoc_array.txt | 574 +++++++++++++ include/linux/assoc_array.h | 92 ++ include/linux/assoc_array_priv.h | 182 ++++ lib/Kconfig | 14 + lib/Makefile | 1 + lib/assoc_array.c | 1745 ++++++++++++++++++++++++++++++++++++++ 6 files changed, 2608 insertions(+) create mode 100644 Documentation/assoc_array.txt create mode 100644 include/linux/assoc_array.h create mode 100644 include/linux/assoc_array_priv.h create mode 100644 lib/assoc_array.c (limited to 'Documentation') diff --git a/Documentation/assoc_array.txt b/Documentation/assoc_array.txt new file mode 100644 index 000000000000..f4faec0f66e4 --- /dev/null +++ b/Documentation/assoc_array.txt @@ -0,0 +1,574 @@ + ======================================== + GENERIC ASSOCIATIVE ARRAY IMPLEMENTATION + ======================================== + +Contents: + + - Overview. + + - The public API. + - Edit script. + - Operations table. + - Manipulation functions. + - Access functions. + - Index key form. + + - Internal workings. + - Basic internal tree layout. + - Shortcuts. + - Splitting and collapsing nodes. + - Non-recursive iteration. + - Simultaneous alteration and iteration. + + +======== +OVERVIEW +======== + +This associative array implementation is an object container with the following +properties: + + (1) Objects are opaque pointers. The implementation does not care where they + point (if anywhere) or what they point to (if anything). + + [!] NOTE: Pointers to objects _must_ be zero in the least significant bit. + + (2) Objects do not need to contain linkage blocks for use by the array. This + permits an object to be located in multiple arrays simultaneously. + Rather, the array is made up of metadata blocks that point to objects. + + (3) Objects require index keys to locate them within the array. + + (4) Index keys must be unique. Inserting an object with the same key as one + already in the array will replace the old object. + + (5) Index keys can be of any length and can be of different lengths. + + (6) Index keys should encode the length early on, before any variation due to + length is seen. + + (7) Index keys can include a hash to scatter objects throughout the array. + + (8) The array can iterated over. The objects will not necessarily come out in + key order. + + (9) The array can be iterated over whilst it is being modified, provided the + RCU readlock is being held by the iterator. Note, however, under these + circumstances, some objects may be seen more than once. If this is a + problem, the iterator should lock against modification. Objects will not + be missed, however, unless deleted. + +(10) Objects in the array can be looked up by means of their index key. + +(11) Objects can be looked up whilst the array is being modified, provided the + RCU readlock is being held by the thread doing the look up. + +The implementation uses a tree of 16-pointer nodes internally that are indexed +on each level by nibbles from the index key in the same manner as in a radix +tree. To improve memory efficiency, shortcuts can be emplaced to skip over +what would otherwise be a series of single-occupancy nodes. Further, nodes +pack leaf object pointers into spare space in the node rather than making an +extra branch until as such time an object needs to be added to a full node. + + +============== +THE PUBLIC API +============== + +The public API can be found in . The associative array is +rooted on the following structure: + + struct assoc_array { + ... + }; + +The code is selected by enabling CONFIG_ASSOCIATIVE_ARRAY. + + +EDIT SCRIPT +----------- + +The insertion and deletion functions produce an 'edit script' that can later be +applied to effect the changes without risking ENOMEM. This retains the +preallocated metadata blocks that will be installed in the internal tree and +keeps track of the metadata blocks that will be removed from the tree when the +script is applied. + +This is also used to keep track of dead blocks and dead objects after the +script has been applied so that they can be freed later. The freeing is done +after an RCU grace period has passed - thus allowing access functions to +proceed under the RCU read lock. + +The script appears as outside of the API as a pointer of the type: + + struct assoc_array_edit; + +There are two functions for dealing with the script: + + (1) Apply an edit script. + + void assoc_array_apply_edit(struct assoc_array_edit *edit); + + This will perform the edit functions, interpolating various write barriers + to permit accesses under the RCU read lock to continue. The edit script + will then be passed to call_rcu() to free it and any dead stuff it points + to. + + (2) Cancel an edit script. + + void assoc_array_cancel_edit(struct assoc_array_edit *edit); + + This frees the edit script and all preallocated memory immediately. If + this was for insertion, the new object is _not_ released by this function, + but must rather be released by the caller. + +These functions are guaranteed not to fail. + + +OPERATIONS TABLE +---------------- + +Various functions take a table of operations: + + struct assoc_array_ops { + ... + }; + +This points to a number of methods, all of which need to be provided: + + (1) Get a chunk of index key from caller data: + + unsigned long (*get_key_chunk)(const void *index_key, int level); + + This should return a chunk of caller-supplied index key starting at the + *bit* position given by the level argument. The level argument will be a + multiple of ASSOC_ARRAY_KEY_CHUNK_SIZE and the function should return + ASSOC_ARRAY_KEY_CHUNK_SIZE bits. No error is possible. + + + (2) Get a chunk of an object's index key. + + unsigned long (*get_object_key_chunk)(const void *object, int level); + + As the previous function, but gets its data from an object in the array + rather than from a caller-supplied index key. + + + (3) See if this is the object we're looking for. + + bool (*compare_object)(const void *object, const void *index_key); + + Compare the object against an index key and return true if it matches and + false if it doesn't. + + + (4) Diff the index keys of two objects. + + int (*diff_objects)(const void *a, const void *b); + + Return the bit position at which the index keys of two objects differ or + -1 if they are the same. + + + (5) Free an object. + + void (*free_object)(void *object); + + Free the specified object. Note that this may be called an RCU grace + period after assoc_array_apply_edit() was called, so synchronize_rcu() may + be necessary on module unloading. + + +MANIPULATION FUNCTIONS +---------------------- + +There are a number of functions for manipulating an associative array: + + (1) Initialise an associative array. + + void assoc_array_init(struct assoc_array *array); + + This initialises the base structure for an associative array. It can't + fail. + + + (2) Insert/replace an object in an associative array. + + struct assoc_array_edit * + assoc_array_insert(struct assoc_array *array, + const struct assoc_array_ops *ops, + const void *index_key, + void *object); + + This inserts the given object into the array. Note that the least + significant bit of the pointer must be zero as it's used to type-mark + pointers internally. + + If an object already exists for that key then it will be replaced with the + new object and the old one will be freed automatically. + + The index_key argument should hold index key information and is + passed to the methods in the ops table when they are called. + + This function makes no alteration to the array itself, but rather returns + an edit script that must be applied. -ENOMEM is returned in the case of + an out-of-memory error. + + The caller should lock exclusively against other modifiers of the array. + + + (3) Delete an object from an associative array. + + struct assoc_array_edit * + assoc_array_delete(struct assoc_array *array, + const struct assoc_array_ops *ops, + const void *index_key); + + This deletes an object that matches the specified data from the array. + + The index_key argument should hold index key information and is + passed to the methods in the ops table when they are called. + + This function makes no alteration to the array itself, but rather returns + an edit script that must be applied. -ENOMEM is returned in the case of + an out-of-memory error. NULL will be returned if the specified object is + not found within the array. + + The caller should lock exclusively against other modifiers of the array. + + + (4) Delete all objects from an associative array. + + struct assoc_array_edit * + assoc_array_clear(struct assoc_array *array, + const struct assoc_array_ops *ops); + + This deletes all the objects from an associative array and leaves it + completely empty. + + This function makes no alteration to the array itself, but rather returns + an edit script that must be applied. -ENOMEM is returned in the case of + an out-of-memory error. + + The caller should lock exclusively against other modifiers of the array. + + + (5) Destroy an associative array, deleting all objects. + + void assoc_array_destroy(struct assoc_array *array, + const struct assoc_array_ops *ops); + + This destroys the contents of the associative array and leaves it + completely empty. It is not permitted for another thread to be traversing + the array under the RCU read lock at the same time as this function is + destroying it as no RCU deferral is performed on memory release - + something that would require memory to be allocated. + + The caller should lock exclusively against other modifiers and accessors + of the array. + + + (6) Garbage collect an associative array. + + int assoc_array_gc(struct assoc_array *array, + const struct assoc_array_ops *ops, + bool (*iterator)(void *object, void *iterator_data), + void *iterator_data); + + This iterates over the objects in an associative array and passes each one + to iterator(). If iterator() returns true, the object is kept. If it + returns false, the object will be freed. If the iterator() function + returns true, it must perform any appropriate refcount incrementing on the + object before returning. + + The internal tree will be packed down if possible as part of the iteration + to reduce the number of nodes in it. + + The iterator_data is passed directly to iterator() and is otherwise + ignored by the function. + + The function will return 0 if successful and -ENOMEM if there wasn't + enough memory. + + It is possible for other threads to iterate over or search the array under + the RCU read lock whilst this function is in progress. The caller should + lock exclusively against other modifiers of the array. + + +ACCESS FUNCTIONS +---------------- + +There are two functions for accessing an associative array: + + (1) Iterate over all the objects in an associative array. + + int assoc_array_iterate(const struct assoc_array *array, + int (*iterator)(const void *object, + void *iterator_data), + void *iterator_data); + + This passes each object in the array to the iterator callback function. + iterator_data is private data for that function. + + This may be used on an array at the same time as the array is being + modified, provided the RCU read lock is held. Under such circumstances, + it is possible for the iteration function to see some objects twice. If + this is a problem, then modification should be locked against. The + iteration algorithm should not, however, miss any objects. + + The function will return 0 if no objects were in the array or else it will + return the result of the last iterator function called. Iteration stops + immediately if any call to the iteration function results in a non-zero + return. + + + (2) Find an object in an associative array. + + void *assoc_array_find(const struct assoc_array *array, + const struct assoc_array_ops *ops, + const void *index_key); + + This walks through the array's internal tree directly to the object + specified by the index key.. + + This may be used on an array at the same time as the array is being + modified, provided the RCU read lock is held. + + The function will return the object if found (and set *_type to the object + type) or will return NULL if the object was not found. + + +INDEX KEY FORM +-------------- + +The index key can be of any form, but since the algorithms aren't told how long +the key is, it is strongly recommended that the index key includes its length +very early on before any variation due to the length would have an effect on +comparisons. + +This will cause leaves with different length keys to scatter away from each +other - and those with the same length keys to cluster together. + +It is also recommended that the index key begin with a hash of the rest of the +key to maximise scattering throughout keyspace. + +The better the scattering, the wider and lower the internal tree will be. + +Poor scattering isn't too much of a problem as there are shortcuts and nodes +can contain mixtures of leaves and metadata pointers. + +The index key is read in chunks of machine word. Each chunk is subdivided into +one nibble (4 bits) per level, so on a 32-bit CPU this is good for 8 levels and +on a 64-bit CPU, 16 levels. Unless the scattering is really poor, it is +unlikely that more than one word of any particular index key will have to be +used. + + +================= +INTERNAL WORKINGS +================= + +The associative array data structure has an internal tree. This tree is +constructed of two types of metadata blocks: nodes and shortcuts. + +A node is an array of slots. Each slot can contain one of four things: + + (*) A NULL pointer, indicating that the slot is empty. + + (*) A pointer to an object (a leaf). + + (*) A pointer to a node at the next level. + + (*) A pointer to a shortcut. + + +BASIC INTERNAL TREE LAYOUT +-------------------------- + +Ignoring shortcuts for the moment, the nodes form a multilevel tree. The index +key space is strictly subdivided by the nodes in the tree and nodes occur on +fixed levels. For example: + + Level: 0 1 2 3 + =============== =============== =============== =============== + NODE D + NODE B NODE C +------>+---+ + +------>+---+ +------>+---+ | | 0 | + NODE A | | 0 | | | 0 | | +---+ + +---+ | +---+ | +---+ | : : + | 0 | | : : | : : | +---+ + +---+ | +---+ | +---+ | | f | + | 1 |---+ | 3 |---+ | 7 |---+ +---+ + +---+ +---+ +---+ + : : : : | 8 |---+ + +---+ +---+ +---+ | NODE E + | e |---+ | f | : : +------>+---+ + +---+ | +---+ +---+ | 0 | + | f | | | f | +---+ + +---+ | +---+ : : + | NODE F +---+ + +------>+---+ | f | + | 0 | NODE G +---+ + +---+ +------>+---+ + : : | | 0 | + +---+ | +---+ + | 6 |---+ : : + +---+ +---+ + : : | f | + +---+ +---+ + | f | + +---+ + +In the above example, there are 7 nodes (A-G), each with 16 slots (0-f). +Assuming no other meta data nodes in the tree, the key space is divided thusly: + + KEY PREFIX NODE + ========== ==== + 137* D + 138* E + 13[0-69-f]* C + 1[0-24-f]* B + e6* G + e[0-57-f]* F + [02-df]* A + +So, for instance, keys with the following example index keys will be found in +the appropriate nodes: + + INDEX KEY PREFIX NODE + =============== ======= ==== + 13694892892489 13 C + 13795289025897 137 D + 13889dde88793 138 E + 138bbb89003093 138 E + 1394879524789 12 C + 1458952489 1 B + 9431809de993ba - A + b4542910809cd - A + e5284310def98 e F + e68428974237 e6 G + e7fffcbd443 e F + f3842239082 - A + +To save memory, if a node can hold all the leaves in its portion of keyspace, +then the node will have all those leaves in it and will not have any metadata +pointers - even if some of those leaves would like to be in the same slot. + +A node can contain a heterogeneous mix of leaves and metadata pointers. +Metadata pointers must be in the slots that match their subdivisions of key +space. The leaves can be in any slot not occupied by a metadata pointer. It +is guaranteed that none of the leaves in a node will match a slot occupied by a +metadata pointer. If the metadata pointer is there, any leaf whose key matches +the metadata key prefix must be in the subtree that the metadata pointer points +to. + +In the above example list of index keys, node A will contain: + + SLOT CONTENT INDEX KEY (PREFIX) + ==== =============== ================== + 1 PTR TO NODE B 1* + any LEAF 9431809de993ba + any LEAF b4542910809cd + e PTR TO NODE F e* + any LEAF f3842239082 + +and node B: + + 3 PTR TO NODE C 13* + any LEAF 1458952489 + + +SHORTCUTS +--------- + +Shortcuts are metadata records that jump over a piece of keyspace. A shortcut +is a replacement for a series of single-occupancy nodes ascending through the +levels. Shortcuts exist to save memory and to speed up traversal. + +It is possible for the root of the tree to be a shortcut - say, for example, +the tree contains at least 17 nodes all with key prefix '1111'. The insertion +algorithm will insert a shortcut to skip over the '1111' keyspace in a single +bound and get to the fourth level where these actually become different. + + +SPLITTING AND COLLAPSING NODES +------------------------------ + +Each node has a maximum capacity of 16 leaves and metadata pointers. If the +insertion algorithm finds that it is trying to insert a 17th object into a +node, that node will be split such that at least two leaves that have a common +key segment at that level end up in a separate node rooted on that slot for +that common key segment. + +If the leaves in a full node and the leaf that is being inserted are +sufficiently similar, then a shortcut will be inserted into the tree. + +When the number of objects in the subtree rooted at a node falls to 16 or +fewer, then the subtree will be collapsed down to a single node - and this will +ripple towards the root if possible. + + +NON-RECURSIVE ITERATION +----------------------- + +Each node and shortcut contains a back pointer to its parent and the number of +slot in that parent that points to it. None-recursive iteration uses these to +proceed rootwards through the tree, going to the parent node, slot N + 1 to +make sure progress is made without the need for a stack. + +The backpointers, however, make simultaneous alteration and iteration tricky. + + +SIMULTANEOUS ALTERATION AND ITERATION +------------------------------------- + +There are a number of cases to consider: + + (1) Simple insert/replace. This involves simply replacing a NULL or old + matching leaf pointer with the pointer to the new leaf after a barrier. + The metadata blocks don't change otherwise. An old leaf won't be freed + until after the RCU grace period. + + (2) Simple delete. This involves just clearing an old matching leaf. The + metadata blocks don't change otherwise. The old leaf won't be freed until + after the RCU grace period. + + (3) Insertion replacing part of a subtree that we haven't yet entered. This + may involve replacement of part of that subtree - but that won't affect + the iteration as we won't have reached the pointer to it yet and the + ancestry blocks are not replaced (the layout of those does not change). + + (4) Insertion replacing nodes that we're actively processing. This isn't a + problem as we've passed the anchoring pointer and won't switch onto the + new layout until we follow the back pointers - at which point we've + already examined the leaves in the replaced node (we iterate over all the + leaves in a node before following any of its metadata pointers). + + We might, however, re-see some leaves that have been split out into a new + branch that's in a slot further along than we were at. + + (5) Insertion replacing nodes that we're processing a dependent branch of. + This won't affect us until we follow the back pointers. Similar to (4). + + (6) Deletion collapsing a branch under us. This doesn't affect us because the + back pointers will get us back to the parent of the new node before we + could see the new node. The entire collapsed subtree is thrown away + unchanged - and will still be rooted on the same slot, so we shouldn't + process it a second time as we'll go back to slot + 1. + +Note: + + (*) Under some circumstances, we need to simultaneously change the parent + pointer and the parent slot pointer on a node (say, for example, we + inserted another node before it and moved it up a level). We cannot do + this without locking against a read - so we have to replace that node too. + + However, when we're changing a shortcut into a node this isn't a problem + as shortcuts only have one slot and so the parent slot number isn't used + when traversing backwards over one. This means that it's okay to change + the slot number first - provided suitable barriers are used to make sure + the parent slot number is read after the back pointer. + +Obsolete blocks and leaves are freed up after an RCU grace period has passed, +so as long as anyone doing walking or iteration holds the RCU read lock, the +old superstructure should not go away on them. diff --git a/include/linux/assoc_array.h b/include/linux/assoc_array.h new file mode 100644 index 000000000000..9a193b84238a --- /dev/null +++ b/include/linux/assoc_array.h @@ -0,0 +1,92 @@ +/* Generic associative array implementation. + * + * See Documentation/assoc_array.txt for information. + * + * Copyright (C) 2013 Red Hat, Inc. All Rights Reserved. + * Written by David Howells (dhowells@redhat.com) + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public Licence + * as published by the Free Software Foundation; either version + * 2 of the Licence, or (at your option) any later version. + */ + +#ifndef _LINUX_ASSOC_ARRAY_H +#define _LINUX_ASSOC_ARRAY_H + +#ifdef CONFIG_ASSOCIATIVE_ARRAY + +#include + +#define ASSOC_ARRAY_KEY_CHUNK_SIZE BITS_PER_LONG /* Key data retrieved in chunks of this size */ + +/* + * Generic associative array. + */ +struct assoc_array { + struct assoc_array_ptr *root; /* The node at the root of the tree */ + unsigned long nr_leaves_on_tree; +}; + +/* + * Operations on objects and index keys for use by array manipulation routines. + */ +struct assoc_array_ops { + /* Method to get a chunk of an index key from caller-supplied data */ + unsigned long (*get_key_chunk)(const void *index_key, int level); + + /* Method to get a piece of an object's index key */ + unsigned long (*get_object_key_chunk)(const void *object, int level); + + /* Is this the object we're looking for? */ + bool (*compare_object)(const void *object, const void *index_key); + + /* How different are two objects, to a bit position in their keys? (or + * -1 if they're the same) + */ + int (*diff_objects)(const void *a, const void *b); + + /* Method to free an object. */ + void (*free_object)(void *object); +}; + +/* + * Access and manipulation functions. + */ +struct assoc_array_edit; + +static inline void assoc_array_init(struct assoc_array *array) +{ + array->root = NULL; + array->nr_leaves_on_tree = 0; +} + +extern int assoc_array_iterate(const struct assoc_array *array, + int (*iterator)(const void *object, + void *iterator_data), + void *iterator_data); +extern void *assoc_array_find(const struct assoc_array *array, + const struct assoc_array_ops *ops, + const void *index_key); +extern void assoc_array_destroy(struct assoc_array *array, + const struct assoc_array_ops *ops); +extern struct assoc_array_edit *assoc_array_insert(struct assoc_array *array, + const struct assoc_array_ops *ops, + const void *index_key, + void *object); +extern void assoc_array_insert_set_object(struct assoc_array_edit *edit, + void *object); +extern struct assoc_array_edit *assoc_array_delete(struct assoc_array *array, + const struct assoc_array_ops *ops, + const void *index_key); +extern struct assoc_array_edit *assoc_array_clear(struct assoc_array *array, + const struct assoc_array_ops *ops); +extern void assoc_array_apply_edit(struct assoc_array_edit *edit); +extern void assoc_array_cancel_edit(struct assoc_array_edit *edit); +extern int assoc_array_gc(struct assoc_array *array, + const struct assoc_array_ops *ops, + bool (*iterator)(void *object, void *iterator_data), + void *iterator_data); + +#endif /* CONFIG_ASSOCIATIVE_ARRAY */ +#endif /* _LINUX_ASSOC_ARRAY_H */ diff --git a/include/linux/assoc_array_priv.h b/include/linux/assoc_array_priv.h new file mode 100644 index 000000000000..711275e6681c --- /dev/null +++ b/include/linux/assoc_array_priv.h @@ -0,0 +1,182 @@ +/* Private definitions for the generic associative array implementation. + * + * See Documentation/assoc_array.txt for information. + * + * Copyright (C) 2013 Red Hat, Inc. All Rights Reserved. + * Written by David Howells (dhowells@redhat.com) + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public Licence + * as published by the Free Software Foundation; either version + * 2 of the Licence, or (at your option) any later version. + */ + +#ifndef _LINUX_ASSOC_ARRAY_PRIV_H +#define _LINUX_ASSOC_ARRAY_PRIV_H + +#ifdef CONFIG_ASSOCIATIVE_ARRAY + +#include + +#define ASSOC_ARRAY_FAN_OUT 16 /* Number of slots per node */ +#define ASSOC_ARRAY_FAN_MASK (ASSOC_ARRAY_FAN_OUT - 1) +#define ASSOC_ARRAY_LEVEL_STEP (ilog2(ASSOC_ARRAY_FAN_OUT)) +#define ASSOC_ARRAY_LEVEL_STEP_MASK (ASSOC_ARRAY_LEVEL_STEP - 1) +#define ASSOC_ARRAY_KEY_CHUNK_MASK (ASSOC_ARRAY_KEY_CHUNK_SIZE - 1) +#define ASSOC_ARRAY_KEY_CHUNK_SHIFT (ilog2(BITS_PER_LONG)) + +/* + * Undefined type representing a pointer with type information in the bottom + * two bits. + */ +struct assoc_array_ptr; + +/* + * An N-way node in the tree. + * + * Each slot contains one of four things: + * + * (1) Nothing (NULL). + * + * (2) A leaf object (pointer types 0). + * + * (3) A next-level node (pointer type 1, subtype 0). + * + * (4) A shortcut (pointer type 1, subtype 1). + * + * The tree is optimised for search-by-ID, but permits reasonable iteration + * also. + * + * The tree is navigated by constructing an index key consisting of an array of + * segments, where each segment is ilog2(ASSOC_ARRAY_FAN_OUT) bits in size. + * + * The segments correspond to levels of the tree (the first segment is used at + * level 0, the second at level 1, etc.). + */ +struct assoc_array_node { + struct assoc_array_ptr *back_pointer; + u8 parent_slot; + struct assoc_array_ptr *slots[ASSOC_ARRAY_FAN_OUT]; + unsigned long nr_leaves_on_branch; +}; + +/* + * A shortcut through the index space out to where a collection of nodes/leaves + * with the same IDs live. + */ +struct assoc_array_shortcut { + struct assoc_array_ptr *back_pointer; + int parent_slot; + int skip_to_level; + struct assoc_array_ptr *next_node; + unsigned long index_key[]; +}; + +/* + * Preallocation cache. + */ +struct assoc_array_edit { + struct rcu_head rcu; + struct assoc_array *array; + const struct assoc_array_ops *ops; + const struct assoc_array_ops *ops_for_excised_subtree; + struct assoc_array_ptr *leaf; + struct assoc_array_ptr **leaf_p; + struct assoc_array_ptr *dead_leaf; + struct assoc_array_ptr *new_meta[3]; + struct assoc_array_ptr *excised_meta[1]; + struct assoc_array_ptr *excised_subtree; + struct assoc_array_ptr **set_backpointers[ASSOC_ARRAY_FAN_OUT]; + struct assoc_array_ptr *set_backpointers_to; + struct assoc_array_node *adjust_count_on; + long adjust_count_by; + struct { + struct assoc_array_ptr **ptr; + struct assoc_array_ptr *to; + } set[2]; + struct { + u8 *p; + u8 to; + } set_parent_slot[1]; + u8 segment_cache[ASSOC_ARRAY_FAN_OUT + 1]; +}; + +/* + * Internal tree member pointers are marked in the bottom one or two bits to + * indicate what type they are so that we don't have to look behind every + * pointer to see what it points to. + * + * We provide functions to test type annotations and to create and translate + * the annotated pointers. + */ +#define ASSOC_ARRAY_PTR_TYPE_MASK 0x1UL +#define ASSOC_ARRAY_PTR_LEAF_TYPE 0x0UL /* Points to leaf (or nowhere) */ +#define ASSOC_ARRAY_PTR_META_TYPE 0x1UL /* Points to node or shortcut */ +#define ASSOC_ARRAY_PTR_SUBTYPE_MASK 0x2UL +#define ASSOC_ARRAY_PTR_NODE_SUBTYPE 0x0UL +#define ASSOC_ARRAY_PTR_SHORTCUT_SUBTYPE 0x2UL + +static inline bool assoc_array_ptr_is_meta(const struct assoc_array_ptr *x) +{ + return (unsigned long)x & ASSOC_ARRAY_PTR_TYPE_MASK; +} +static inline bool assoc_array_ptr_is_leaf(const struct assoc_array_ptr *x) +{ + return !assoc_array_ptr_is_meta(x); +} +static inline bool assoc_array_ptr_is_shortcut(const struct assoc_array_ptr *x) +{ + return (unsigned long)x & ASSOC_ARRAY_PTR_SUBTYPE_MASK; +} +static inline bool assoc_array_ptr_is_node(const struct assoc_array_ptr *x) +{ + return !assoc_array_ptr_is_shortcut(x); +} + +static inline void *assoc_array_ptr_to_leaf(const struct assoc_array_ptr *x) +{ + return (void *)((unsigned long)x & ~ASSOC_ARRAY_PTR_TYPE_MASK); +} + +static inline +unsigned long __assoc_array_ptr_to_meta(const struct assoc_array_ptr *x) +{ + return (unsigned long)x & + ~(ASSOC_ARRAY_PTR_SUBTYPE_MASK | ASSOC_ARRAY_PTR_TYPE_MASK); +} +static inline +struct assoc_array_node *assoc_array_ptr_to_node(const struct assoc_array_ptr *x) +{ + return (struct assoc_array_node *)__assoc_array_ptr_to_meta(x); +} +static inline +struct assoc_array_shortcut *assoc_array_ptr_to_shortcut(const struct assoc_array_ptr *x) +{ + return (struct assoc_array_shortcut *)__assoc_array_ptr_to_meta(x); +} + +static inline +struct assoc_array_ptr *__assoc_array_x_to_ptr(const void *p, unsigned long t) +{ + return (struct assoc_array_ptr *)((unsigned long)p | t); +} +static inline +struct assoc_array_ptr *assoc_array_leaf_to_ptr(const void *p) +{ + return __assoc_array_x_to_ptr(p, ASSOC_ARRAY_PTR_LEAF_TYPE); +} +static inline +struct assoc_array_ptr *assoc_array_node_to_ptr(const struct assoc_array_node *p) +{ + return __assoc_array_x_to_ptr( + p, ASSOC_ARRAY_PTR_META_TYPE | ASSOC_ARRAY_PTR_NODE_SUBTYPE); +} +static inline +struct assoc_array_ptr *assoc_array_shortcut_to_ptr(const struct assoc_array_shortcut *p) +{ + return __assoc_array_x_to_ptr( + p, ASSOC_ARRAY_PTR_META_TYPE | ASSOC_ARRAY_PTR_SHORTCUT_SUBTYPE); +} + +#endif /* CONFIG_ASSOCIATIVE_ARRAY */ +#endif /* _LINUX_ASSOC_ARRAY_PRIV_H */ diff --git a/lib/Kconfig b/lib/Kconfig index b3c8be0da17f..3cb879b1f282 100644 --- a/lib/Kconfig +++ b/lib/Kconfig @@ -322,6 +322,20 @@ config TEXTSEARCH_FSM config BTREE boolean +config ASSOCIATIVE_ARRAY + bool + help + Generic associative array. Can be searched and iterated over whilst + it is being modified. It is also reasonably quick to search and + modify. The algorithms are non-recursive, and the trees are highly + capacious. + + See: + + Documentation/assoc_array.txt + + for more information. + config HAS_IOMEM boolean depends on !NO_IOMEM diff --git a/lib/Makefile b/lib/Makefile index f3bb2cb98adf..1e806477e472 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -51,6 +51,7 @@ CFLAGS_hweight.o = $(subst $(quote),,$(CONFIG_ARCH_HWEIGHT_CFLAGS)) obj-$(CONFIG_GENERIC_HWEIGHT) += hweight.o obj-$(CONFIG_BTREE) += btree.o +obj-$(CONFIG_ASSOCIATIVE_ARRAY) += assoc_array.o obj-$(CONFIG_DEBUG_PREEMPT) += smp_processor_id.o obj-$(CONFIG_DEBUG_LIST) += list_debug.o obj-$(CONFIG_DEBUG_OBJECTS) += debugobjects.o diff --git a/lib/assoc_array.c b/lib/assoc_array.c new file mode 100644 index 000000000000..a0952818f938 --- /dev/null +++ b/lib/assoc_array.c @@ -0,0 +1,1745 @@ +/* Generic associative array implementation. + * + * See Documentation/assoc_array.txt for information. + * + * Copyright (C) 2013 Red Hat, Inc. All Rights Reserved. + * Written by David Howells (dhowells@redhat.com) + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public Licence + * as published by the Free Software Foundation; either version + * 2 of the Licence, or (at your option) any later version. + */ +//#define DEBUG +#include +#include + +/* + * Iterate over an associative array. The caller must hold the RCU read lock + * or better. + */ +static int assoc_array_subtree_iterate(const struct assoc_array_ptr *root, + const struct assoc_array_ptr *stop, + int (*iterator)(const void *leaf, + void *iterator_data), + void *iterator_data) +{ + const struct assoc_array_shortcut *shortcut; + const struct assoc_array_node *node; + const struct assoc_array_ptr *cursor, *ptr, *parent; + unsigned long has_meta; + int slot, ret; + + cursor = root; + +begin_node: + if (assoc_array_ptr_is_shortcut(cursor)) { + /* Descend through a shortcut */ + shortcut = assoc_array_ptr_to_shortcut(cursor); + smp_read_barrier_depends(); + cursor = ACCESS_ONCE(shortcut->next_node); + } + + node = assoc_array_ptr_to_node(cursor); + smp_read_barrier_depends(); + slot = 0; + + /* We perform two passes of each node. + * + * The first pass does all the leaves in this node. This means we + * don't miss any leaves if the node is split up by insertion whilst + * we're iterating over the branches rooted here (we may, however, see + * some leaves twice). + */ + has_meta = 0; + for (; slot < ASSOC_ARRAY_FAN_OUT; slot++) { + ptr = ACCESS_ONCE(node->slots[slot]); + has_meta |= (unsigned long)ptr; + if (ptr && assoc_array_ptr_is_leaf(ptr)) { + /* We need a barrier between the read of the pointer + * and dereferencing the pointer - but only if we are + * actually going to dereference it. + */ + smp_read_barrier_depends(); + + /* Invoke the callback */ + ret = iterator(assoc_array_ptr_to_leaf(ptr), + iterator_data); + if (ret) + return ret; + } + } + + /* The second pass attends to all the metadata pointers. If we follow + * one of these we may find that we don't come back here, but rather go + * back to a replacement node with the leaves in a different layout. + * + * We are guaranteed to make progress, however, as the slot number for + * a particular portion of the key space cannot change - and we + * continue at the back pointer + 1. + */ + if (!(has_meta & ASSOC_ARRAY_PTR_META_TYPE)) + goto finished_node; + slot = 0; + +continue_node: + node = assoc_array_ptr_to_node(cursor); + smp_read_barrier_depends(); + + for (; slot < ASSOC_ARRAY_FAN_OUT; slot++) { + ptr = ACCESS_ONCE(node->slots[slot]); + if (assoc_array_ptr_is_meta(ptr)) { + cursor = ptr; + goto begin_node; + } + } + +finished_node: + /* Move up to the parent (may need to skip back over a shortcut) */ + parent = ACCESS_ONCE(node->back_pointer); + slot = node->parent_slot; + if (parent == stop) + return 0; + + if (assoc_array_ptr_is_shortcut(parent)) { + shortcut = assoc_array_ptr_to_shortcut(parent); + smp_read_barrier_depends(); + cursor = parent; + parent = ACCESS_ONCE(shortcut->back_pointer); + slot = shortcut->parent_slot; + if (parent == stop) + return 0; + } + + /* Ascend to next slot in parent node */ + cursor = parent; + slot++; + goto continue_node; +} + +/** + * assoc_array_iterate - Pass all objects in the array to a callback + * @array: The array to iterate over. + * @iterator: The callback function. + * @iterator_data: Private data for the callback function. + * + * Iterate over all the objects in an associative array. Each one will be + * presented to the iterator function. + * + * If the array is being modified concurrently with the iteration then it is + * possible that some objects in the array will be passed to the iterator + * callback more than once - though every object should be passed at least + * once. If this is undesirable then the caller must lock against modification + * for the duration of this function. + * + * The function will return 0 if no objects were in the array or else it will + * return the result of the last iterator function called. Iteration stops + * immediately if any call to the iteration function results in a non-zero + * return. + * + * The caller should hold the RCU read lock or better if concurrent + * modification is possible. + */ +int assoc_array_iterate(const struct assoc_array *array, + int (*iterator)(const void *object, + void *iterator_data), + void *iterator_data) +{ + struct assoc_array_ptr *root = ACCESS_ONCE(array->root); + + if (!root) + return 0; + return assoc_array_subtree_iterate(root, NULL, iterator, iterator_data); +} + +enum assoc_array_walk_status { + assoc_array_walk_tree_empty, + assoc_array_walk_found_terminal_node, + assoc_array_walk_found_wrong_shortcut, +} status; + +struct assoc_array_walk_result { + struct { + struct assoc_array_node *node; /* Node in which leaf might be found */ + int level; + int slot; + } terminal_node; + struct { + struct assoc_array_shortcut *shortcut; + int level; + int sc_level; + unsigned long sc_segments; + unsigned long dissimilarity; + } wrong_shortcut; +}; + +/* + * Navigate through the internal tree looking for the closest node to the key. + */ +static enum assoc_array_walk_status +assoc_array_walk(const struct assoc_array *array, + const struct assoc_array_ops *ops, + const void *index_key, + struct assoc_array_walk_result *result) +{ + struct assoc_array_shortcut *shortcut; + struct assoc_array_node *node; + struct assoc_array_ptr *cursor, *ptr; + unsigned long sc_segments, dissimilarity; + unsigned long segments; + int level, sc_level, next_sc_level; + int slot; + + pr_devel("-->%s()\n", __func__); + + cursor = ACCESS_ONCE(array->root); + if (!cursor) + return assoc_array_walk_tree_empty; + + level = 0; + + /* Use segments from the key for the new leaf to navigate through the + * internal tree, skipping through nodes and shortcuts that are on + * route to the destination. Eventually we'll come to a slot that is + * either empty or contains a leaf at which point we've found a node in + * which the leaf we're looking for might be found or into which it + * should be inserted. + */ +jumped: + segments = ops->get_key_chunk(index_key, level); + pr_devel("segments[%d]: %lx\n", level, segments); + + if (assoc_array_ptr_is_shortcut(cursor)) + goto follow_shortcut; + +consider_node: + node = assoc_array_ptr_to_node(cursor); + smp_read_barrier_depends(); + + slot = segments >> (level & ASSOC_ARRAY_KEY_CHUNK_MASK); + slot &= ASSOC_ARRAY_FAN_MASK; + ptr = ACCESS_ONCE(node->slots[slot]); + + pr_devel("consider slot %x [ix=%d type=%lu]\n", + slot, level, (unsigned long)ptr & 3); + + if (!assoc_array_ptr_is_meta(ptr)) { + /* The node doesn't have a node/shortcut pointer in the slot + * corresponding to the index key that we have to follow. + */ + result->terminal_node.node = node; + result->terminal_node.level = level; + result->terminal_node.slot = slot; + pr_devel("<--%s() = terminal_node\n", __func__); + return assoc_array_walk_found_terminal_node; + } + + if (assoc_array_ptr_is_node(ptr)) { + /* There is a pointer to a node in the slot corresponding to + * this index key segment, so we need to follow it. + */ + cursor = ptr; + level += ASSOC_ARRAY_LEVEL_STEP; + if ((level & ASSOC_ARRAY_KEY_CHUNK_MASK) != 0) + goto consider_node; + goto jumped; + } + + /* There is a shortcut in the slot corresponding to the index key + * segment. We follow the shortcut if its partial index key matches + * this leaf's. Otherwise we need to split the shortcut. + */ + cursor = ptr; +follow_shortcut: + shortcut = assoc_array_ptr_to_shortcut(cursor); + smp_read_barrier_depends(); + pr_devel("shortcut to %d\n", shortcut->skip_to_level); + sc_level = level + ASSOC_ARRAY_LEVEL_STEP; + BUG_ON(sc_level > shortcut->skip_to_level); + + do { + /* Check the leaf against the shortcut's index key a word at a + * time, trimming the final word (the shortcut stores the index + * key completely from the root to the shortcut's target). + */ + if ((sc_level & ASSOC_ARRAY_KEY_CHUNK_MASK) == 0) + segments = ops->get_key_chunk(index_key, sc_level); + + sc_segments = shortcut->index_key[sc_level >> ASSOC_ARRAY_KEY_CHUNK_SHIFT]; + dissimilarity = segments ^ sc_segments; + + if (round_up(sc_level, ASSOC_ARRAY_KEY_CHUNK_SIZE) > shortcut->skip_to_level) { + /* Trim segments that are beyond the shortcut */ + int shift = shortcut->skip_to_level & ASSOC_ARRAY_KEY_CHUNK_MASK; + dissimilarity &= ~(ULONG_MAX << shift); + next_sc_level = shortcut->skip_to_level; + } else { + next_sc_level = sc_level + ASSOC_ARRAY_KEY_CHUNK_SIZE; + next_sc_level = round_down(next_sc_level, ASSOC_ARRAY_KEY_CHUNK_SIZE); + } + + if (dissimilarity != 0) { + /* This shortcut points elsewhere */ + result->wrong_shortcut.shortcut = shortcut; + result->wrong_shortcut.level = level; + result->wrong_shortcut.sc_level = sc_level; + result->wrong_shortcut.sc_segments = sc_segments; + result->wrong_shortcut.dissimilarity = dissimilarity; + return assoc_array_walk_found_wrong_shortcut; + } + + sc_level = next_sc_level; + } while (sc_level < shortcut->skip_to_level); + + /* The shortcut matches the leaf's index to this point. */ + cursor = ACCESS_ONCE(shortcut->next_node); + if (((level ^ sc_level) & ~ASSOC_ARRAY_KEY_CHUNK_MASK) != 0) { + level = sc_level; + goto jumped; + } else { + level = sc_level; + goto consider_node; + } +} + +/** + * assoc_array_find - Find an object by index key + * @array: The associative array to search. + * @ops: The operations to use. + * @index_key: The key to the object. + * + * Find an object in an associative array by walking through the internal tree + * to the node that should contain the object and then searching the leaves + * there. NULL is returned if the requested object was not found in the array. + * + * The caller must hold the RCU read lock or better. + */ +void *assoc_array_find(const struct assoc_array *array, + const struct assoc_array_ops *ops, + const void *index_key) +{ + struct assoc_array_walk_result result; + const struct assoc_array_node *node; + const struct assoc_array_ptr *ptr; + const void *leaf; + int slot; + + if (assoc_array_walk(array, ops, index_key, &result) != + assoc_array_walk_found_terminal_node) + return NULL; + + node = result.terminal_node.node; + smp_read_barrier_depends(); + + /* If the target key is available to us, it's has to be pointed to by + * the terminal node. + */ + for (slot = 0; slot < ASSOC_ARRAY_FAN_OUT; slot++) { + ptr = ACCESS_ONCE(node->slots[slot]); + if (ptr && assoc_array_ptr_is_leaf(ptr)) { + /* We need a barrier between the read of the pointer + * and dereferencing the pointer - but only if we are + * actually going to dereference it. + */ + leaf = assoc_array_ptr_to_leaf(ptr); + smp_read_barrier_depends(); + if (ops->compare_object(leaf, index_key)) + return (void *)leaf; + } + } + + return NULL; +} + +/* + * Destructively iterate over an associative array. The caller must prevent + * other simultaneous accesses. + */ +static void assoc_array_destroy_subtree(struct assoc_array_ptr *root, + const struct assoc_array_ops *ops) +{ + struct assoc_array_shortcut *shortcut; + struct assoc_array_node *node; + struct assoc_array_ptr *cursor, *parent = NULL; + int slot = -1; + + pr_devel("-->%s()\n", __func__); + + cursor = root; + if (!cursor) { + pr_devel("empty\n"); + return; + } + +move_to_meta: + if (assoc_array_ptr_is_shortcut(cursor)) { + /* Descend through a shortcut */ + pr_devel("[%d] shortcut\n", slot); + BUG_ON(!assoc_array_ptr_is_shortcut(cursor)); + shortcut = assoc_array_ptr_to_shortcut(cursor); + BUG_ON(shortcut->back_pointer != parent); + BUG_ON(slot != -1 && shortcut->parent_slot != slot); + parent = cursor; + cursor = shortcut->next_node; + slot = -1; + BUG_ON(!assoc_array_ptr_is_node(cursor)); + } + + pr_devel("[%d] node\n", slot); + node = assoc_array_ptr_to_node(cursor); + BUG_ON(node->back_pointer != parent); + BUG_ON(slot != -1 && node->parent_slot != slot); + slot = 0; + +continue_node: + pr_devel("Node %p [back=%p]\n", node, node->back_pointer); + for (; slot < ASSOC_ARRAY_FAN_OUT; slot++) { + struct assoc_array_ptr *ptr = node->slots[slot]; + if (!ptr) + continue; + if (assoc_array_ptr_is_meta(ptr)) { + parent = cursor; + cursor = ptr; + goto move_to_meta; + } + + if (ops) { + pr_devel("[%d] free leaf\n", slot); + ops->free_object(assoc_array_ptr_to_leaf(ptr)); + } + } + + parent = node->back_pointer; + slot = node->parent_slot; + pr_devel("free node\n"); + kfree(node); + if (!parent) + return; /* Done */ + + /* Move back up to the parent (may need to free a shortcut on + * the way up) */ + if (assoc_array_ptr_is_shortcut(parent)) { + shortcut = assoc_array_ptr_to_shortcut(parent); + BUG_ON(shortcut->next_node != cursor); + cursor = parent; + parent = shortcut->back_pointer; + slot = shortcut->parent_slot; + pr_devel("free shortcut\n"); + kfree(shortcut); + if (!parent) + return; + + BUG_ON(!assoc_array_ptr_is_node(parent)); + } + + /* Ascend to next slot in parent node */ + pr_devel("ascend to %p[%d]\n", parent, slot); + cursor = parent; + node = assoc_array_ptr_to_node(cursor); + slot++; + goto continue_node; +} + +/** + * assoc_array_destroy - Destroy an associative array + * @array: The array to destroy. + * @ops: The operations to use. + * + * Discard all metadata and free all objects in an associative array. The + * array will be empty and ready to use again upon completion. This function + * cannot fail. + * + * The caller must prevent all other accesses whilst this takes place as no + * attempt is made to adjust pointers gracefully to permit RCU readlock-holding + * accesses to continue. On the other hand, no memory allocation is required. + */ +void assoc_array_destroy(struct assoc_array *array, + const struct assoc_array_ops *ops) +{ + assoc_array_destroy_subtree(array->root, ops); + array->root = NULL; +} + +/* + * Handle insertion into an empty tree. + */ +static bool assoc_array_insert_in_empty_tree(struct assoc_array_edit *edit) +{ + struct assoc_array_node *new_n0; + + pr_devel("-->%s()\n", __func__); + + new_n0 = kzalloc(sizeof(struct assoc_array_node), GFP_KERNEL); + if (!new_n0) + return false; + + edit->new_meta[0] = assoc_array_node_to_ptr(new_n0); + edit->leaf_p = &new_n0->slots[0]; + edit->adjust_count_on = new_n0; + edit->set[0].ptr = &edit->array->root; + edit->set[0].to = assoc_array_node_to_ptr(new_n0); + + pr_devel("<--%s() = ok [no root]\n", __func__); + return true; +} + +/* + * Handle insertion into a terminal node. + */ +static bool assoc_array_insert_into_terminal_node(struct assoc_array_edit *edit, + const struct assoc_array_ops *ops, + const void *index_key, + struct assoc_array_walk_result *result) +{ + struct assoc_array_shortcut *shortcut, *new_s0; + struct assoc_array_node *node, *new_n0, *new_n1, *side; + struct assoc_array_ptr *ptr; + unsigned long dissimilarity, base_seg, blank; + size_t keylen; + bool have_meta; + int level, diff; + int slot, next_slot, free_slot, i, j; + + node = result->terminal_node.node; + level = result->terminal_node.level; + edit->segment_cache[ASSOC_ARRAY_FAN_OUT] = result->terminal_node.slot; + + pr_devel("-->%s()\n", __func__); + + /* We arrived at a node which doesn't have an onward node or shortcut + * pointer that we have to follow. This means that (a) the leaf we + * want must go here (either by insertion or replacement) or (b) we + * need to split this node and insert in one of the fragments. + */ + free_slot = -1; + + /* Firstly, we have to check the leaves in this node to see if there's + * a matching one we should replace in place. + */ + for (i = 0; i < ASSOC_ARRAY_FAN_OUT; i++) { + ptr = node->slots[i]; + if (!ptr) { + free_slot = i; + continue; + } + if (ops->compare_object(assoc_array_ptr_to_leaf(ptr), index_key)) { + pr_devel("replace in slot %d\n", i); + edit->leaf_p = &node->slots[i]; + edit->dead_leaf = node->slots[i]; + pr_devel("<--%s() = ok [replace]\n", __func__); + return true; + } + } + + /* If there is a free slot in this node then we can just insert the + * leaf here. + */ + if (free_slot >= 0) { + pr_devel("insert in free slot %d\n", free_slot); + edit->leaf_p = &node->slots[free_slot]; + edit->adjust_count_on = node; + pr_devel("<--%s() = ok [insert]\n", __func__); + return true; + } + + /* The node has no spare slots - so we're either going to have to split + * it or insert another node before it. + * + * Whatever, we're going to need at least two new nodes - so allocate + * those now. We may also need a new shortcut, but we deal with that + * when we need it. + */ + new_n0 = kzalloc(sizeof(struct assoc_array_node), GFP_KERNEL); + if (!new_n0) + return false; + edit->new_meta[0] = assoc_array_node_to_ptr(new_n0); + new_n1 = kzalloc(sizeof(struct assoc_array_node), GFP_KERNEL); + if (!new_n1) + return false; + edit->new_meta[1] = assoc_array_node_to_ptr(new_n1); + + /* We need to find out how similar the leaves are. */ + pr_devel("no spare slots\n"); + have_meta = false; + for (i = 0; i < ASSOC_ARRAY_FAN_OUT; i++) { + ptr = node->slots[i]; + if (assoc_array_ptr_is_meta(ptr)) { + edit->segment_cache[i] = 0xff; + have_meta = true; + continue; + } + base_seg = ops->get_object_key_chunk( + assoc_array_ptr_to_leaf(ptr), level); + base_seg >>= level & ASSOC_ARRAY_KEY_CHUNK_MASK; + edit->segment_cache[i] = base_seg & ASSOC_ARRAY_FAN_MASK; + } + + if (have_meta) { + pr_devel("have meta\n"); + goto split_node; + } + + /* The node contains only leaves */ + dissimilarity = 0; + base_seg = edit->segment_cache[0]; + for (i = 1; i < ASSOC_ARRAY_FAN_OUT; i++) + dissimilarity |= edit->segment_cache[i] ^ base_seg; + + pr_devel("only leaves; dissimilarity=%lx\n", dissimilarity); + + if ((dissimilarity & ASSOC_ARRAY_FAN_MASK) == 0) { + /* The old leaves all cluster in the same slot. We will need + * to insert a shortcut if the new node wants to cluster with them. + */ + if ((edit->segment_cache[ASSOC_ARRAY_FAN_OUT] ^ base_seg) == 0) + goto all_leaves_cluster_together; + + /* Otherwise we can just insert a new node ahead of the old + * one. + */ + goto present_leaves_cluster_but_not_new_leaf; + } + +split_node: + pr_devel("split node\n"); + + /* We need to split the current node; we know that the node doesn't + * simply contain a full set of leaves that cluster together (it + * contains meta pointers and/or non-clustering leaves). + * + * We need to expel at least two leaves out of a set consisting of the + * leaves in the node and the new leaf. + * + * We need a new node (n0) to replace the current one and a new node to + * take the expelled nodes (n1). + */ + edit->set[0].to = assoc_array_node_to_ptr(new_n0); + new_n0->back_pointer = node->back_pointer; + new_n0->parent_slot = node->parent_slot; + new_n1->back_pointer = assoc_array_node_to_ptr(new_n0); + new_n1->parent_slot = -1; /* Need to calculate this */ + +do_split_node: + pr_devel("do_split_node\n"); + + new_n0->nr_leaves_on_branch = node->nr_leaves_on_branch; + new_n1->nr_leaves_on_branch = 0; + + /* Begin by finding two matching leaves. There have to be at least two + * that match - even if there are meta pointers - because any leaf that + * would match a slot with a meta pointer in it must be somewhere + * behind that meta pointer and cannot be here. Further, given N + * remaining leaf slots, we now have N+1 leaves to go in them. + */ + for (i = 0; i < ASSOC_ARRAY_FAN_OUT; i++) { + slot = edit->segment_cache[i]; + if (slot != 0xff) + for (j = i + 1; j < ASSOC_ARRAY_FAN_OUT + 1; j++) + if (edit->segment_cache[j] == slot) + goto found_slot_for_multiple_occupancy; + } +found_slot_for_multiple_occupancy: + pr_devel("same slot: %x %x [%02x]\n", i, j, slot); + BUG_ON(i >= ASSOC_ARRAY_FAN_OUT); + BUG_ON(j >= ASSOC_ARRAY_FAN_OUT + 1); + BUG_ON(slot >= ASSOC_ARRAY_FAN_OUT); + + new_n1->parent_slot = slot; + + /* Metadata pointers cannot change slot */ + for (i = 0; i < ASSOC_ARRAY_FAN_OUT; i++) + if (assoc_array_ptr_is_meta(node->slots[i])) + new_n0->slots[i] = node->slots[i]; + else + new_n0->slots[i] = NULL; + BUG_ON(new_n0->slots[slot] != NULL); + new_n0->slots[slot] = assoc_array_node_to_ptr(new_n1); + + /* Filter the leaf pointers between the new nodes */ + free_slot = -1; + next_slot = 0; + for (i = 0; i < ASSOC_ARRAY_FAN_OUT; i++) { + if (assoc_array_ptr_is_meta(node->slots[i])) + continue; + if (edit->segment_cache[i] == slot) { + new_n1->slots[next_slot++] = node->slots[i]; + new_n1->nr_leaves_on_branch++; + } else { + do { + free_slot++; + } while (new_n0->slots[free_slot] != NULL); + new_n0->slots[free_slot] = node->slots[i]; + } + } + + pr_devel("filtered: f=%x n=%x\n", free_slot, next_slot); + + if (edit->segment_cache[ASSOC_ARRAY_FAN_OUT] != slot) { + do { + free_slot++; + } while (new_n0->slots[free_slot] != NULL); + edit->leaf_p = &new_n0->slots[free_slot]; + edit->adjust_count_on = new_n0; + } else { + edit->leaf_p = &new_n1->slots[next_slot++]; + edit->adjust_count_on = new_n1; + } + + BUG_ON(next_slot <= 1); + + edit->set_backpointers_to = assoc_array_node_to_ptr(new_n0); + for (i = 0; i < ASSOC_ARRAY_FAN_OUT; i++) { + if (edit->segment_cache[i] == 0xff) { + ptr = node->slots[i]; + BUG_ON(assoc_array_ptr_is_leaf(ptr)); + if (assoc_array_ptr_is_node(ptr)) { + side = assoc_array_ptr_to_node(ptr); + edit->set_backpointers[i] = &side->back_pointer; + } else { + shortcut = assoc_array_ptr_to_shortcut(ptr); + edit->set_backpointers[i] = &shortcut->back_pointer; + } + } + } + + ptr = node->back_pointer; + if (!ptr) + edit->set[0].ptr = &edit->array->root; + else if (assoc_array_ptr_is_node(ptr)) + edit->set[0].ptr = &assoc_array_ptr_to_node(ptr)->slots[node->parent_slot]; + else + edit->set[0].ptr = &assoc_array_ptr_to_shortcut(ptr)->next_node; + edit->excised_meta[0] = assoc_array_node_to_ptr(node); + pr_devel("<--%s() = ok [split node]\n", __func__); + return true; + +present_leaves_cluster_but_not_new_leaf: + /* All the old leaves cluster in the same slot, but the new leaf wants + * to go into a different slot, so we create a new node to hold the new + * leaf and a pointer to a new node holding all the old leaves. + */ + pr_devel("present leaves cluster but not new leaf\n"); + + new_n0->back_pointer = node->back_pointer; + new_n0->parent_slot = node->parent_slot; + new_n0->nr_leaves_on_branch = node->nr_leaves_on_branch; + new_n1->back_pointer = assoc_array_node_to_ptr(new_n0); + new_n1->parent_slot = edit->segment_cache[0]; + new_n1->nr_leaves_on_branch = node->nr_leaves_on_branch; + edit->adjust_count_on = new_n0; + + for (i = 0; i < ASSOC_ARRAY_FAN_OUT; i++) + new_n1->slots[i] = node->slots[i]; + + new_n0->slots[edit->segment_cache[0]] = assoc_array_node_to_ptr(new_n0); + edit->leaf_p = &new_n0->slots[edit->segment_cache[ASSOC_ARRAY_FAN_OUT]]; + + edit->set[0].ptr = &assoc_array_ptr_to_node(node->back_pointer)->slots[node->parent_slot]; + edit->set[0].to = assoc_array_node_to_ptr(new_n0); + edit->excised_meta[0] = assoc_array_node_to_ptr(node); + pr_devel("<--%s() = ok [insert node before]\n", __func__); + return true; + +all_leaves_cluster_together: + /* All the leaves, new and old, want to cluster together in this node + * in the same slot, so we have to replace this node with a shortcut to + * skip over the identical parts of the key and then place a pair of + * nodes, one inside the other, at the end of the shortcut and + * distribute the keys between them. + * + * Firstly we need to work out where the leaves start diverging as a + * bit position into their keys so that we know how big the shortcut + * needs to be. + * + * We only need to make a single pass of N of the N+1 leaves because if + * any keys differ between themselves at bit X then at least one of + * them must also differ with the base key at bit X or before. + */ + pr_devel("all leaves cluster together\n"); + diff = INT_MAX; + for (i = 0; i < ASSOC_ARRAY_FAN_OUT; i++) { + int x = ops->diff_objects(assoc_array_ptr_to_leaf(edit->leaf), + assoc_array_ptr_to_leaf(node->slots[i])); + if (x < diff) { + BUG_ON(x < 0); + diff = x; + } + } + BUG_ON(diff == INT_MAX); + BUG_ON(diff < level + ASSOC_ARRAY_LEVEL_STEP); + + keylen = round_up(diff, ASSOC_ARRAY_KEY_CHUNK_SIZE); + keylen >>= ASSOC_ARRAY_KEY_CHUNK_SHIFT; + + new_s0 = kzalloc(sizeof(struct assoc_array_shortcut) + + keylen * sizeof(unsigned long), GFP_KERNEL); + if (!new_s0) + return false; + edit->new_meta[2] = assoc_array_shortcut_to_ptr(new_s0); + + edit->set[0].to = assoc_array_shortcut_to_ptr(new_s0); + new_s0->back_pointer = node->back_pointer; + new_s0->parent_slot = node->parent_slot; + new_s0->next_node = assoc_array_node_to_ptr(new_n0); + new_n0->back_pointer = assoc_array_shortcut_to_ptr(new_s0); + new_n0->parent_slot = 0; + new_n1->back_pointer = assoc_array_node_to_ptr(new_n0); + new_n1->parent_slot = -1; /* Need to calculate this */ + + new_s0->skip_to_level = level = diff & ~ASSOC_ARRAY_LEVEL_STEP_MASK; + pr_devel("skip_to_level = %d [diff %d]\n", level, diff); + BUG_ON(level <= 0); + + for (i = 0; i < keylen; i++) + new_s0->index_key[i] = + ops->get_key_chunk(index_key, i * ASSOC_ARRAY_KEY_CHUNK_SIZE); + + blank = ULONG_MAX << (level & ASSOC_ARRAY_KEY_CHUNK_MASK); + pr_devel("blank off [%zu] %d: %lx\n", keylen - 1, level, blank); + new_s0->index_key[keylen - 1] &= ~blank; + + /* This now reduces to a node splitting exercise for which we'll need + * to regenerate the disparity table. + */ + for (i = 0; i < ASSOC_ARRAY_FAN_OUT; i++) { + ptr = node->slots[i]; + base_seg = ops->get_object_key_chunk(assoc_array_ptr_to_leaf(ptr), + level); + base_seg >>= level & ASSOC_ARRAY_KEY_CHUNK_MASK; + edit->segment_cache[i] = base_seg & ASSOC_ARRAY_FAN_MASK; + } + + base_seg = ops->get_key_chunk(index_key, level); + base_seg >>= level & ASSOC_ARRAY_KEY_CHUNK_MASK; + edit->segment_cache[ASSOC_ARRAY_FAN_OUT] = base_seg & ASSOC_ARRAY_FAN_MASK; + goto do_split_node; +} + +/* + * Handle insertion into the middle of a shortcut. + */ +static bool assoc_array_insert_mid_shortcut(struct assoc_array_edit *edit, + const struct assoc_array_ops *ops, + struct assoc_array_walk_result *result) +{ + struct assoc_array_shortcut *shortcut, *new_s0, *new_s1; + struct assoc_array_node *node, *new_n0, *side; + unsigned long sc_segments, dissimilarity, blank; + size_t keylen; + int level, sc_level, diff; + int sc_slot; + + shortcut = result->wrong_shortcut.shortcut; + level = result->wrong_shortcut.level; + sc_level = result->wrong_shortcut.sc_level; + sc_segments = result->wrong_shortcut.sc_segments; + dissimilarity = result->wrong_shortcut.dissimilarity; + + pr_devel("-->%s(ix=%d dis=%lx scix=%d)\n", + __func__, level, dissimilarity, sc_level); + + /* We need to split a shortcut and insert a node between the two + * pieces. Zero-length pieces will be dispensed with entirely. + * + * First of all, we need to find out in which level the first + * difference was. + */ + diff = __ffs(dissimilarity); + diff &= ~ASSOC_ARRAY_LEVEL_STEP_MASK; + diff += sc_level & ~ASSOC_ARRAY_KEY_CHUNK_MASK; + pr_devel("diff=%d\n", diff); + + if (!shortcut->back_pointer) { + edit->set[0].ptr = &edit->array->root; + } else if (assoc_array_ptr_is_node(shortcut->back_pointer)) { + node = assoc_array_ptr_to_node(shortcut->back_pointer); + edit->set[0].ptr = &node->slots[shortcut->parent_slot]; + } else { + BUG(); + } + + edit->excised_meta[0] = assoc_array_shortcut_to_ptr(shortcut); + + /* Create a new node now since we're going to need it anyway */ + new_n0 = kzalloc(sizeof(struct assoc_array_node), GFP_KERNEL); + if (!new_n0) + return false; + edit->new_meta[0] = assoc_array_node_to_ptr(new_n0); + edit->adjust_count_on = new_n0; + + /* Insert a new shortcut before the new node if this segment isn't of + * zero length - otherwise we just connect the new node directly to the + * parent. + */ + level += ASSOC_ARRAY_LEVEL_STEP; + if (diff > level) { + pr_devel("pre-shortcut %d...%d\n", level, diff); + keylen = round_up(diff, ASSOC_ARRAY_KEY_CHUNK_SIZE); + keylen >>= ASSOC_ARRAY_KEY_CHUNK_SHIFT; + + new_s0 = kzalloc(sizeof(struct assoc_array_shortcut) + + keylen * sizeof(unsigned long), GFP_KERNEL); + if (!new_s0) + return false; + edit->new_meta[1] = assoc_array_shortcut_to_ptr(new_s0); + edit->set[0].to = assoc_array_shortcut_to_ptr(new_s0); + new_s0->back_pointer = shortcut->back_pointer; + new_s0->parent_slot = shortcut->parent_slot; + new_s0->next_node = assoc_array_node_to_ptr(new_n0); + new_s0->skip_to_level = diff; + + new_n0->back_pointer = assoc_array_shortcut_to_ptr(new_s0); + new_n0->parent_slot = 0; + + memcpy(new_s0->index_key, shortcut->index_key, + keylen * sizeof(unsigned long)); + + blank = ULONG_MAX << (diff & ASSOC_ARRAY_KEY_CHUNK_MASK); + pr_devel("blank off [%zu] %d: %lx\n", keylen - 1, diff, blank); + new_s0->index_key[keylen - 1] &= ~blank; + } else { + pr_devel("no pre-shortcut\n"); + edit->set[0].to = assoc_array_node_to_ptr(new_n0); + new_n0->back_pointer = shortcut->back_pointer; + new_n0->parent_slot = shortcut->parent_slot; + } + + side = assoc_array_ptr_to_node(shortcut->next_node); + new_n0->nr_leaves_on_branch = side->nr_leaves_on_branch; + + /* We need to know which slot in the new node is going to take a + * metadata pointer. + */ + sc_slot = sc_segments >> (diff & ASSOC_ARRAY_KEY_CHUNK_MASK); + sc_slot &= ASSOC_ARRAY_FAN_MASK; + + pr_devel("new slot %lx >> %d -> %d\n", + sc_segments, diff & ASSOC_ARRAY_KEY_CHUNK_MASK, sc_slot); + + /* Determine whether we need to follow the new node with a replacement + * for the current shortcut. We could in theory reuse the current + * shortcut if its parent slot number doesn't change - but that's a + * 1-in-16 chance so not worth expending the code upon. + */ + level = diff + ASSOC_ARRAY_LEVEL_STEP; + if (level < shortcut->skip_to_level) { + pr_devel("post-shortcut %d...%d\n", level, shortcut->skip_to_level); + keylen = round_up(shortcut->skip_to_level, ASSOC_ARRAY_KEY_CHUNK_SIZE); + keylen >>= ASSOC_ARRAY_KEY_CHUNK_SHIFT; + + new_s1 = kzalloc(sizeof(struct assoc_array_shortcut) + + keylen * sizeof(unsigned long), GFP_KERNEL); + if (!new_s1) + return false; + edit->new_meta[2] = assoc_array_shortcut_to_ptr(new_s1); + + new_s1->back_pointer = assoc_array_node_to_ptr(new_n0); + new_s1->parent_slot = sc_slot; + new_s1->next_node = shortcut->next_node; + new_s1->skip_to_level = shortcut->skip_to_level; + + new_n0->slots[sc_slot] = assoc_array_shortcut_to_ptr(new_s1); + + memcpy(new_s1->index_key, shortcut->index_key, + keylen * sizeof(unsigned long)); + + edit->set[1].ptr = &side->back_pointer; + edit->set[1].to = assoc_array_shortcut_to_ptr(new_s1); + } else { + pr_devel("no post-shortcut\n"); + + /* We don't have to replace the pointed-to node as long as we + * use memory barriers to make sure the parent slot number is + * changed before the back pointer (the parent slot number is + * irrelevant to the old parent shortcut). + */ + new_n0->slots[sc_slot] = shortcut->next_node; + edit->set_parent_slot[0].p = &side->parent_slot; + edit->set_parent_slot[0].to = sc_slot; + edit->set[1].ptr = &side->back_pointer; + edit->set[1].to = assoc_array_node_to_ptr(new_n0); + } + + /* Install the new leaf in a spare slot in the new node. */ + if (sc_slot == 0) + edit->leaf_p = &new_n0->slots[1]; + else + edit->leaf_p = &new_n0->slots[0]; + + pr_devel("<--%s() = ok [split shortcut]\n", __func__); + return edit; +} + +/** + * assoc_array_insert - Script insertion of an object into an associative array + * @array: The array to insert into. + * @ops: The operations to use. + * @index_key: The key to insert at. + * @object: The object to insert. + * + * Precalculate and preallocate a script for the insertion or replacement of an + * object in an associative array. This results in an edit script that can + * either be applied or cancelled. + * + * The function returns a pointer to an edit script or -ENOMEM. + * + * The caller should lock against other modifications and must continue to hold + * the lock until assoc_array_apply_edit() has been called. + * + * Accesses to the tree may take place concurrently with this function, + * provided they hold the RCU read lock. + */ +struct assoc_array_edit *assoc_array_insert(struct assoc_array *array, + const struct assoc_array_ops *ops, + const void *index_key, + void *object) +{ + struct assoc_array_walk_result result; + struct assoc_array_edit *edit; + + pr_devel("-->%s()\n", __func__); + + /* The leaf pointer we're given must not have the bottom bit set as we + * use those for type-marking the pointer. NULL pointers are also not + * allowed as they indicate an empty slot but we have to allow them + * here as they can be updated later. + */ + BUG_ON(assoc_array_ptr_is_meta(object)); + + edit = kzalloc(sizeof(struct assoc_array_edit), GFP_KERNEL); + if (!edit) + return ERR_PTR(-ENOMEM); + edit->array = array; + edit->ops = ops; + edit->leaf = assoc_array_leaf_to_ptr(object); + edit->adjust_count_by = 1; + + switch (assoc_array_walk(array, ops, index_key, &result)) { + case assoc_array_walk_tree_empty: + /* Allocate a root node if there isn't one yet */ + if (!assoc_array_insert_in_empty_tree(edit)) + goto enomem; + return edit; + + case assoc_array_walk_found_terminal_node: + /* We found a node that doesn't have a node/shortcut pointer in + * the slot corresponding to the index key that we have to + * follow. + */ + if (!assoc_array_insert_into_terminal_node(edit, ops, index_key, + &result)) + goto enomem; + return edit; + + case assoc_array_walk_found_wrong_shortcut: + /* We found a shortcut that didn't match our key in a slot we + * needed to follow. + */ + if (!assoc_array_insert_mid_shortcut(edit, ops, &result)) + goto enomem; + return edit; + } + +enomem: + /* Clean up after an out of memory error */ + pr_devel("enomem\n"); + assoc_array_cancel_edit(edit); + return ERR_PTR(-ENOMEM); +} + +/** + * assoc_array_insert_set_object - Set the new object pointer in an edit script + * @edit: The edit script to modify. + * @object: The object pointer to set. + * + * Change the object to be inserted in an edit script. The object pointed to + * by the old object is not freed. This must be done prior to applying the + * script. + */ +void assoc_array_insert_set_object(struct assoc_array_edit *edit, void *object) +{ + BUG_ON(!object); + edit->leaf = assoc_array_leaf_to_ptr(object); +} + +struct assoc_array_delete_collapse_context { + struct assoc_array_node *node; + const void *skip_leaf; + int slot; +}; + +/* + * Subtree collapse to node iterator. + */ +static int assoc_array_delete_collapse_iterator(const void *leaf, + void *iterator_data) +{ + struct assoc_array_delete_collapse_context *collapse = iterator_data; + + if (leaf == collapse->skip_leaf) + return 0; + + BUG_ON(collapse->slot >= ASSOC_ARRAY_FAN_OUT); + + collapse->node->slots[collapse->slot++] = assoc_array_leaf_to_ptr(leaf); + return 0; +} + +/** + * assoc_array_delete - Script deletion of an object from an associative array + * @array: The array to search. + * @ops: The operations to use. + * @index_key: The key to the object. + * + * Precalculate and preallocate a script for the deletion of an object from an + * associative array. This results in an edit script that can either be + * applied or cancelled. + * + * The function returns a pointer to an edit script if the object was found, + * NULL if the object was not found or -ENOMEM. + * + * The caller should lock against other modifications and must continue to hold + * the lock until assoc_array_apply_edit() has been called. + * + * Accesses to the tree may take place concurrently with this function, + * provided they hold the RCU read lock. + */ +struct assoc_array_edit *assoc_array_delete(struct assoc_array *array, + const struct assoc_array_ops *ops, + const void *index_key) +{ + struct assoc_array_delete_collapse_context collapse; + struct assoc_array_walk_result result; + struct assoc_array_node *node, *new_n0; + struct assoc_array_edit *edit; + struct assoc_array_ptr *ptr; + bool has_meta; + int slot, i; + + pr_devel("-->%s()\n", __func__); + + edit = kzalloc(sizeof(struct assoc_array_edit), GFP_KERNEL); + if (!edit) + return ERR_PTR(-ENOMEM); + edit->array = array; + edit->ops = ops; + edit->adjust_count_by = -1; + + switch (assoc_array_walk(array, ops, index_key, &result)) { + case assoc_array_walk_found_terminal_node: + /* We found a node that should contain the leaf we've been + * asked to remove - *if* it's in the tree. + */ + pr_devel("terminal_node\n"); + node = result.terminal_node.node; + + for (slot = 0; slot < ASSOC_ARRAY_FAN_OUT; slot++) { + ptr = node->slots[slot]; + if (ptr && + assoc_array_ptr_is_leaf(ptr) && + ops->compare_object(assoc_array_ptr_to_leaf(ptr), + index_key)) + goto found_leaf; + } + case assoc_array_walk_tree_empty: + case assoc_array_walk_found_wrong_shortcut: + default: + assoc_array_cancel_edit(edit); + pr_devel("not found\n"); + return NULL; + } + +found_leaf: + BUG_ON(array->nr_leaves_on_tree <= 0); + + /* In the simplest form of deletion we just clear the slot and release + * the leaf after a suitable interval. + */ + edit->dead_leaf = node->slots[slot]; + edit->set[0].ptr = &node->slots[slot]; + edit->set[0].to = NULL; + edit->adjust_count_on = node; + + /* If that concludes erasure of the last leaf, then delete the entire + * internal array. + */ + if (array->nr_leaves_on_tree == 1) { + edit->set[1].ptr = &array->root; + edit->set[1].to = NULL; + edit->adjust_count_on = NULL; + edit->excised_subtree = array->root; + pr_devel("all gone\n"); + return edit; + } + + /* However, we'd also like to clear up some metadata blocks if we + * possibly can. + * + * We go for a simple algorithm of: if this node has FAN_OUT or fewer + * leaves in it, then attempt to collapse it - and attempt to + * recursively collapse up the tree. + * + * We could also try and collapse in partially filled subtrees to take + * up space in this node. + */ + if (node->nr_leaves_on_branch <= ASSOC_ARRAY_FAN_OUT + 1) { + struct assoc_array_node *parent, *grandparent; + struct assoc_array_ptr *ptr; + + /* First of all, we need to know if this node has metadata so + * that we don't try collapsing if all the leaves are already + * here. + */ + has_meta = false; + for (i = 0; i < ASSOC_ARRAY_FAN_OUT; i++) { + ptr = node->slots[i]; + if (assoc_array_ptr_is_meta(ptr)) { + has_meta = true; + break; + } + } + + pr_devel("leaves: %ld [m=%d]\n", + node->nr_leaves_on_branch - 1, has_meta); + + /* Look further up the tree to see if we can collapse this node + * into a more proximal node too. + */ + parent = node; + collapse_up: + pr_devel("collapse subtree: %ld\n", parent->nr_leaves_on_branch); + + ptr = parent->back_pointer; + if (!ptr) + goto do_collapse; + if (assoc_array_ptr_is_shortcut(ptr)) { + struct assoc_array_shortcut *s = assoc_array_ptr_to_shortcut(ptr); + ptr = s->back_pointer; + if (!ptr) + goto do_collapse; + } + + grandparent = assoc_array_ptr_to_node(ptr); + if (grandparent->nr_leaves_on_branch <= ASSOC_ARRAY_FAN_OUT + 1) { + parent = grandparent; + goto collapse_up; + } + + do_collapse: + /* There's no point collapsing if the original node has no meta + * pointers to discard and if we didn't merge into one of that + * node's ancestry. + */ + if (has_meta || parent != node) { + node = parent; + + /* Create a new node to collapse into */ + new_n0 = kzalloc(sizeof(struct assoc_array_node), GFP_KERNEL); + if (!new_n0) + goto enomem; + edit->new_meta[0] = assoc_array_node_to_ptr(new_n0); + + new_n0->back_pointer = node->back_pointer; + new_n0->parent_slot = node->parent_slot; + new_n0->nr_leaves_on_branch = node->nr_leaves_on_branch; + edit->adjust_count_on = new_n0; + + collapse.node = new_n0; + collapse.skip_leaf = assoc_array_ptr_to_leaf(edit->dead_leaf); + collapse.slot = 0; + assoc_array_subtree_iterate(assoc_array_node_to_ptr(node), + node->back_pointer, + assoc_array_delete_collapse_iterator, + &collapse); + pr_devel("collapsed %d,%lu\n", collapse.slot, new_n0->nr_leaves_on_branch); + BUG_ON(collapse.slot != new_n0->nr_leaves_on_branch - 1); + + if (!node->back_pointer) { + edit->set[1].ptr = &array->root; + } else if (assoc_array_ptr_is_leaf(node->back_pointer)) { + BUG(); + } else if (assoc_array_ptr_is_node(node->back_pointer)) { + struct assoc_array_node *p = + assoc_array_ptr_to_node(node->back_pointer); + edit->set[1].ptr = &p->slots[node->parent_slot]; + } else if (assoc_array_ptr_is_shortcut(node->back_pointer)) { + struct assoc_array_shortcut *s = + assoc_array_ptr_to_shortcut(node->back_pointer); + edit->set[1].ptr = &s->next_node; + } + edit->set[1].to = assoc_array_node_to_ptr(new_n0); + edit->excised_subtree = assoc_array_node_to_ptr(node); + } + } + + return edit; + +enomem: + /* Clean up after an out of memory error */ + pr_devel("enomem\n"); + assoc_array_cancel_edit(edit); + return ERR_PTR(-ENOMEM); +} + +/** + * assoc_array_clear - Script deletion of all objects from an associative array + * @array: The array to clear. + * @ops: The operations to use. + * + * Precalculate and preallocate a script for the deletion of all the objects + * from an associative array. This results in an edit script that can either + * be applied or cancelled. + * + * The function returns a pointer to an edit script if there are objects to be + * deleted, NULL if there are no objects in the array or -ENOMEM. + * + * The caller should lock against other modifications and must continue to hold + * the lock until assoc_array_apply_edit() has been called. + * + * Accesses to the tree may take place concurrently with this function, + * provided they hold the RCU read lock. + */ +struct assoc_array_edit *assoc_array_clear(struct assoc_array *array, + const struct assoc_array_ops *ops) +{ + struct assoc_array_edit *edit; + + pr_devel("-->%s()\n", __func__); + + if (!array->root) + return NULL; + + edit = kzalloc(sizeof(struct assoc_array_edit), GFP_KERNEL); + if (!edit) + return ERR_PTR(-ENOMEM); + edit->array = array; + edit->ops = ops; + edit->set[1].ptr = &array->root; + edit->set[1].to = NULL; + edit->excised_subtree = array->root; + edit->ops_for_excised_subtree = ops; + pr_devel("all gone\n"); + return edit; +} + +/* + * Handle the deferred destruction after an applied edit. + */ +static void assoc_array_rcu_cleanup(struct rcu_head *head) +{ + struct assoc_array_edit *edit = + container_of(head, struct assoc_array_edit, rcu); + int i; + + pr_devel("-->%s()\n", __func__); + + if (edit->dead_leaf) + edit->ops->free_object(assoc_array_ptr_to_leaf(edit->dead_leaf)); + for (i = 0; i < ARRAY_SIZE(edit->excised_meta); i++) + if (edit->excised_meta[i]) + kfree(assoc_array_ptr_to_node(edit->excised_meta[i])); + + if (edit->excised_subtree) { + BUG_ON(assoc_array_ptr_is_leaf(edit->excised_subtree)); + if (assoc_array_ptr_is_node(edit->excised_subtree)) { + struct assoc_array_node *n = + assoc_array_ptr_to_node(edit->excised_subtree); + n->back_pointer = NULL; + } else { + struct assoc_array_shortcut *s = + assoc_array_ptr_to_shortcut(edit->excised_subtree); + s->back_pointer = NULL; + } + assoc_array_destroy_subtree(edit->excised_subtree, + edit->ops_for_excised_subtree); + } + + kfree(edit); +} + +/** + * assoc_array_apply_edit - Apply an edit script to an associative array + * @edit: The script to apply. + * + * Apply an edit script to an associative array to effect an insertion, + * deletion or clearance. As the edit script includes preallocated memory, + * this is guaranteed not to fail. + * + * The edit script, dead objects and dead metadata will be scheduled for + * destruction after an RCU grace period to permit those doing read-only + * accesses on the array to continue to do so under the RCU read lock whilst + * the edit is taking place. + */ +void assoc_array_apply_edit(struct assoc_array_edit *edit) +{ + struct assoc_array_shortcut *shortcut; + struct assoc_array_node *node; + struct assoc_array_ptr *ptr; + int i; + + pr_devel("-->%s()\n", __func__); + + smp_wmb(); + if (edit->leaf_p) + *edit->leaf_p = edit->leaf; + + smp_wmb(); + for (i = 0; i < ARRAY_SIZE(edit->set_parent_slot); i++) + if (edit->set_parent_slot[i].p) + *edit->set_parent_slot[i].p = edit->set_parent_slot[i].to; + + smp_wmb(); + for (i = 0; i < ARRAY_SIZE(edit->set_backpointers); i++) + if (edit->set_backpointers[i]) + *edit->set_backpointers[i] = edit->set_backpointers_to; + + smp_wmb(); + for (i = 0; i < ARRAY_SIZE(edit->set); i++) + if (edit->set[i].ptr) + *edit->set[i].ptr = edit->set[i].to; + + if (edit->array->root == NULL) { + edit->array->nr_leaves_on_tree = 0; + } else if (edit->adjust_count_on) { + node = edit->adjust_count_on; + for (;;) { + node->nr_leaves_on_branch += edit->adjust_count_by; + + ptr = node->back_pointer; + if (!ptr) + break; + if (assoc_array_ptr_is_shortcut(ptr)) { + shortcut = assoc_array_ptr_to_shortcut(ptr); + ptr = shortcut->back_pointer; + if (!ptr) + break; + } + BUG_ON(!assoc_array_ptr_is_node(ptr)); + node = assoc_array_ptr_to_node(ptr); + } + + edit->array->nr_leaves_on_tree += edit->adjust_count_by; + } + + call_rcu(&edit->rcu, assoc_array_rcu_cleanup); +} + +/** + * assoc_array_cancel_edit - Discard an edit script. + * @edit: The script to discard. + * + * Free an edit script and all the preallocated data it holds without making + * any changes to the associative array it was intended for. + * + * NOTE! In the case of an insertion script, this does _not_ release the leaf + * that was to be inserted. That is left to the caller. + */ +void assoc_array_cancel_edit(struct assoc_array_edit *edit) +{ + struct assoc_array_ptr *ptr; + int i; + + pr_devel("-->%s()\n", __func__); + + /* Clean up after an out of memory error */ + for (i = 0; i < ARRAY_SIZE(edit->new_meta); i++) { + ptr = edit->new_meta[i]; + if (ptr) { + if (assoc_array_ptr_is_node(ptr)) + kfree(assoc_array_ptr_to_node(ptr)); + else + kfree(assoc_array_ptr_to_shortcut(ptr)); + } + } + kfree(edit); +} + +/** + * assoc_array_gc - Garbage collect an associative array. + * @array: The array to clean. + * @ops: The operations to use. + * @iterator: A callback function to pass judgement on each object. + * @iterator_data: Private data for the callback function. + * + * Collect garbage from an associative array and pack down the internal tree to + * save memory. + * + * The iterator function is asked to pass judgement upon each object in the + * array. If it returns false, the object is discard and if it returns true, + * the object is kept. If it returns true, it must increment the object's + * usage count (or whatever it needs to do to retain it) before returning. + * + * This function returns 0 if successful or -ENOMEM if out of memory. In the + * latter case, the array is not changed. + * + * The caller should lock against other modifications and must continue to hold + * the lock until assoc_array_apply_edit() has been called. + * + * Accesses to the tree may take place concurrently with this function, + * provided they hold the RCU read lock. + */ +int assoc_array_gc(struct assoc_array *array, + const struct assoc_array_ops *ops, + bool (*iterator)(void *object, void *iterator_data), + void *iterator_data) +{ + struct assoc_array_shortcut *shortcut, *new_s; + struct assoc_array_node *node, *new_n; + struct assoc_array_edit *edit; + struct assoc_array_ptr *cursor, *ptr; + struct assoc_array_ptr *new_root, *new_parent, **new_ptr_pp; + unsigned long nr_leaves_on_tree; + int keylen, slot, nr_free, next_slot, i; + + pr_devel("-->%s()\n", __func__); + + if (!array->root) + return 0; + + edit = kzalloc(sizeof(struct assoc_array_edit), GFP_KERNEL); + if (!edit) + return -ENOMEM; + edit->array = array; + edit->ops = ops; + edit->ops_for_excised_subtree = ops; + edit->set[0].ptr = &array->root; + edit->excised_subtree = array->root; + + new_root = new_parent = NULL; + new_ptr_pp = &new_root; + cursor = array->root; + +descend: + /* If this point is a shortcut, then we need to duplicate it and + * advance the target cursor. + */ + if (assoc_array_ptr_is_shortcut(cursor)) { + shortcut = assoc_array_ptr_to_shortcut(cursor); + keylen = round_up(shortcut->skip_to_level, ASSOC_ARRAY_KEY_CHUNK_SIZE); + keylen >>= ASSOC_ARRAY_KEY_CHUNK_SHIFT; + new_s = kmalloc(sizeof(struct assoc_array_shortcut) + + keylen * sizeof(unsigned long), GFP_KERNEL); + if (!new_s) + goto enomem; + pr_devel("dup shortcut %p -> %p\n", shortcut, new_s); + memcpy(new_s, shortcut, (sizeof(struct assoc_array_shortcut) + + keylen * sizeof(unsigned long))); + new_s->back_pointer = new_parent; + new_s->parent_slot = shortcut->parent_slot; + *new_ptr_pp = new_parent = assoc_array_shortcut_to_ptr(new_s); + new_ptr_pp = &new_s->next_node; + cursor = shortcut->next_node; + } + + /* Duplicate the node at this position */ + node = assoc_array_ptr_to_node(cursor); + new_n = kzalloc(sizeof(struct assoc_array_node), GFP_KERNEL); + if (!new_n) + goto enomem; + pr_devel("dup node %p -> %p\n", node, new_n); + new_n->back_pointer = new_parent; + new_n->parent_slot = node->parent_slot; + *new_ptr_pp = new_parent = assoc_array_node_to_ptr(new_n); + new_ptr_pp = NULL; + slot = 0; + +continue_node: + /* Filter across any leaves and gc any subtrees */ + for (; slot < ASSOC_ARRAY_FAN_OUT; slot++) { + ptr = node->slots[slot]; + if (!ptr) + continue; + + if (assoc_array_ptr_is_leaf(ptr)) { + if (iterator(assoc_array_ptr_to_leaf(ptr), + iterator_data)) + /* The iterator will have done any reference + * counting on the object for us. + */ + new_n->slots[slot] = ptr; + continue; + } + + new_ptr_pp = &new_n->slots[slot]; + cursor = ptr; + goto descend; + } + + pr_devel("-- compress node %p --\n", new_n); + + /* Count up the number of empty slots in this node and work out the + * subtree leaf count. + */ + new_n->nr_leaves_on_branch = 0; + nr_free = 0; + for (slot = 0; slot < ASSOC_ARRAY_FAN_OUT; slot++) { + ptr = new_n->slots[slot]; + if (!ptr) + nr_free++; + else if (assoc_array_ptr_is_leaf(ptr)) + new_n->nr_leaves_on_branch++; + } + pr_devel("free=%d, leaves=%lu\n", nr_free, new_n->nr_leaves_on_branch); + + /* See what we can fold in */ + next_slot = 0; + for (slot = 0; slot < ASSOC_ARRAY_FAN_OUT; slot++) { + struct assoc_array_shortcut *s; + struct assoc_array_node *child; + + ptr = new_n->slots[slot]; + if (!ptr || assoc_array_ptr_is_leaf(ptr)) + continue; + + s = NULL; + if (assoc_array_ptr_is_shortcut(ptr)) { + s = assoc_array_ptr_to_shortcut(ptr); + ptr = s->next_node; + } + + child = assoc_array_ptr_to_node(ptr); + new_n->nr_leaves_on_branch += child->nr_leaves_on_branch; + + if (child->nr_leaves_on_branch <= nr_free + 1) { + /* Fold the child node into this one */ + pr_devel("[%d] fold node %lu/%d [nx %d]\n", + slot, child->nr_leaves_on_branch, nr_free + 1, + next_slot); + + /* We would already have reaped an intervening shortcut + * on the way back up the tree. + */ + BUG_ON(s); + + new_n->slots[slot] = NULL; + nr_free++; + if (slot < next_slot) + next_slot = slot; + for (i = 0; i < ASSOC_ARRAY_FAN_OUT; i++) { + struct assoc_array_ptr *p = child->slots[i]; + if (!p) + continue; + BUG_ON(assoc_array_ptr_is_meta(p)); + while (new_n->slots[next_slot]) + next_slot++; + BUG_ON(next_slot >= ASSOC_ARRAY_FAN_OUT); + new_n->slots[next_slot++] = p; + nr_free--; + } + kfree(child); + } else { + pr_devel("[%d] retain node %lu/%d [nx %d]\n", + slot, child->nr_leaves_on_branch, nr_free + 1, + next_slot); + } + } + + pr_devel("after: %lu\n", new_n->nr_leaves_on_branch); + + nr_leaves_on_tree = new_n->nr_leaves_on_branch; + + /* Excise this node if it is singly occupied by a shortcut */ + if (nr_free == ASSOC_ARRAY_FAN_OUT - 1) { + for (slot = 0; slot < ASSOC_ARRAY_FAN_OUT; slot++) + if ((ptr = new_n->slots[slot])) + break; + + if (assoc_array_ptr_is_meta(ptr) && + assoc_array_ptr_is_shortcut(ptr)) { + pr_devel("excise node %p with 1 shortcut\n", new_n); + new_s = assoc_array_ptr_to_shortcut(ptr); + new_parent = new_n->back_pointer; + slot = new_n->parent_slot; + kfree(new_n); + if (!new_parent) { + new_s->back_pointer = NULL; + new_s->parent_slot = 0; + new_root = ptr; + goto gc_complete; + } + + if (assoc_array_ptr_is_shortcut(new_parent)) { + /* We can discard any preceding shortcut also */ + struct assoc_array_shortcut *s = + assoc_array_ptr_to_shortcut(new_parent); + + pr_devel("excise preceding shortcut\n"); + + new_parent = new_s->back_pointer = s->back_pointer; + slot = new_s->parent_slot = s->parent_slot; + kfree(s); + if (!new_parent) { + new_s->back_pointer = NULL; + new_s->parent_slot = 0; + new_root = ptr; + goto gc_complete; + } + } + + new_s->back_pointer = new_parent; + new_s->parent_slot = slot; + new_n = assoc_array_ptr_to_node(new_parent); + new_n->slots[slot] = ptr; + goto ascend_old_tree; + } + } + + /* Excise any shortcuts we might encounter that point to nodes that + * only contain leaves. + */ + ptr = new_n->back_pointer; + if (!ptr) + goto gc_complete; + + if (assoc_array_ptr_is_shortcut(ptr)) { + new_s = assoc_array_ptr_to_shortcut(ptr); + new_parent = new_s->back_pointer; + slot = new_s->parent_slot; + + if (new_n->nr_leaves_on_branch <= ASSOC_ARRAY_FAN_OUT) { + struct assoc_array_node *n; + + pr_devel("excise shortcut\n"); + new_n->back_pointer = new_parent; + new_n->parent_slot = slot; + kfree(new_s); + if (!new_parent) { + new_root = assoc_array_node_to_ptr(new_n); + goto gc_complete; + } + + n = assoc_array_ptr_to_node(new_parent); + n->slots[slot] = assoc_array_node_to_ptr(new_n); + } + } else { + new_parent = ptr; + } + new_n = assoc_array_ptr_to_node(new_parent); + +ascend_old_tree: + ptr = node->back_pointer; + if (assoc_array_ptr_is_shortcut(ptr)) { + shortcut = assoc_array_ptr_to_shortcut(ptr); + slot = shortcut->parent_slot; + cursor = shortcut->back_pointer; + } else { + slot = node->parent_slot; + cursor = ptr; + } + BUG_ON(!ptr); + node = assoc_array_ptr_to_node(cursor); + slot++; + goto continue_node; + +gc_complete: + edit->set[0].to = new_root; + assoc_array_apply_edit(edit); + edit->array->nr_leaves_on_tree = nr_leaves_on_tree; + return 0; + +enomem: + pr_devel("enomem\n"); + assoc_array_destroy_subtree(new_root, edit->ops); + kfree(edit); + return -ENOMEM; +} -- cgit v1.2.3 From 4c336e4b1556f4b722ba597bc6e3df786968a600 Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Sun, 6 Oct 2013 12:43:13 -0600 Subject: tpm: Add support for the Nuvoton NPCT501 I2C TPM This chip is/was also branded as a Winbond WPCT301. Originally written by Dan Morav and posted to LKML: https://lkml.org/lkml/2011/9/7/206 The original posting was not merged, I have taken it as a starting point, forward ported, tested and revised the driver: - Rework interrupt handling to work properly with level triggered interrupts. The old version just locked up. - Synchronize various items with Peter Huewe's Infineon driver: * Add durations/timeouts sysfs calls * Remove I2C device auto-detection * Don't fiddle with chip->release * Call tpm_dev_vendor_release in the probe error path * Use MODULE_DEVICE_TABLE for the I2C ids * Provide OF compatible strings for DT support * Use SIMPLE_DEV_PM_OPS * Use module_i2c_driver - checkpatch cleanups - Testing on ARM Kirkwood with GPIO interrupts, with this device tree: tpm@57 { compatible = "nuvoton,npct501"; reg = <0x57>; interrupt-parent = <&gpio1>; interrupts = <6 IRQ_TYPE_LEVEL_LOW>; }; Signed-off-by: Dan Morav [jgg: revised and tested] Signed-off-by: Jason Gunthorpe [phuewe: minor whitespace changes, fixed module name in kconfig] Signed-off-by: Peter Huewe --- .../devicetree/bindings/i2c/trivial-devices.txt | 2 + drivers/char/tpm/Kconfig | 10 + drivers/char/tpm/Makefile | 1 + drivers/char/tpm/tpm_i2c_nuvoton.c | 710 +++++++++++++++++++++ 4 files changed, 723 insertions(+) create mode 100644 drivers/char/tpm/tpm_i2c_nuvoton.c (limited to 'Documentation') diff --git a/Documentation/devicetree/bindings/i2c/trivial-devices.txt b/Documentation/devicetree/bindings/i2c/trivial-devices.txt index ad6a73852f08..58454bdfa20e 100644 --- a/Documentation/devicetree/bindings/i2c/trivial-devices.txt +++ b/Documentation/devicetree/bindings/i2c/trivial-devices.txt @@ -44,6 +44,7 @@ mc,rv3029c2 Real Time Clock Module with I2C-Bus national,lm75 I2C TEMP SENSOR national,lm80 Serial Interface ACPI-Compatible Microprocessor System Hardware Monitor national,lm92 ±0.33°C Accurate, 12-Bit + Sign Temperature Sensor and Thermal Window Comparator with Two-Wire Interface +nuvoton,npct501 i2c trusted platform module (TPM) nxp,pca9556 Octal SMBus and I2C registered interface nxp,pca9557 8-bit I2C-bus and SMBus I/O port with reset nxp,pcf8563 Real-time clock/calendar @@ -61,3 +62,4 @@ taos,tsl2550 Ambient Light Sensor with SMBUS/Two Wire Serial Interface ti,tsc2003 I2C Touch-Screen Controller ti,tmp102 Low Power Digital Temperature Sensor with SMBUS/Two Wire Serial Interface ti,tmp275 Digital Temperature Sensor +winbond,wpct301 i2c trusted platform module (TPM) diff --git a/drivers/char/tpm/Kconfig b/drivers/char/tpm/Kconfig index 94c0c74434ea..ade71c134a1a 100644 --- a/drivers/char/tpm/Kconfig +++ b/drivers/char/tpm/Kconfig @@ -44,6 +44,16 @@ config TCG_TIS_I2C_INFINEON To compile this driver as a module, choose M here; the module will be called tpm_tis_i2c_infineon. +config TCG_TIS_I2C_NUVOTON + tristate "TPM Interface Specification 1.2 Interface (I2C - Nuvoton)" + depends on I2C + ---help--- + If you have a TPM security chip with an I2C interface from + Nuvoton Technology Corp. say Yes and it will be accessible + from within Linux. + To compile this driver as a module, choose M here; the module + will be called tpm_i2c_nuvoton. + config TCG_NSC tristate "National Semiconductor TPM Interface" depends on X86 diff --git a/drivers/char/tpm/Makefile b/drivers/char/tpm/Makefile index 15df2f38ae42..07ee67cf3341 100644 --- a/drivers/char/tpm/Makefile +++ b/drivers/char/tpm/Makefile @@ -14,6 +14,7 @@ endif endif obj-$(CONFIG_TCG_TIS) += tpm_tis.o obj-$(CONFIG_TCG_TIS_I2C_INFINEON) += tpm_i2c_infineon.o +obj-$(CONFIG_TCG_TIS_I2C_NUVOTON) += tpm_i2c_nuvoton.o obj-$(CONFIG_TCG_NSC) += tpm_nsc.o obj-$(CONFIG_TCG_ATMEL) += tpm_atmel.o obj-$(CONFIG_TCG_INFINEON) += tpm_infineon.o diff --git a/drivers/char/tpm/tpm_i2c_nuvoton.c b/drivers/char/tpm/tpm_i2c_nuvoton.c new file mode 100644 index 000000000000..6276fea01ff0 --- /dev/null +++ b/drivers/char/tpm/tpm_i2c_nuvoton.c @@ -0,0 +1,710 @@ +/****************************************************************************** + * Nuvoton TPM I2C Device Driver Interface for WPCT301/NPCT501, + * based on the TCG TPM Interface Spec version 1.2. + * Specifications at www.trustedcomputinggroup.org + * + * Copyright (C) 2011, Nuvoton Technology Corporation. + * Dan Morav + * Copyright (C) 2013, Obsidian Research Corp. + * Jason Gunthorpe + * + * This program is free software: you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation, either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see http://www.gnu.org/licenses/>. + * + * Nuvoton contact information: APC.Support@nuvoton.com + *****************************************************************************/ + +#include +#include +#include +#include +#include +#include +#include +#include "tpm.h" + +/* I2C interface offsets */ +#define TPM_STS 0x00 +#define TPM_BURST_COUNT 0x01 +#define TPM_DATA_FIFO_W 0x20 +#define TPM_DATA_FIFO_R 0x40 +#define TPM_VID_DID_RID 0x60 +/* TPM command header size */ +#define TPM_HEADER_SIZE 10 +#define TPM_RETRY 5 +/* + * I2C bus device maximum buffer size w/o counting I2C address or command + * i.e. max size required for I2C write is 34 = addr, command, 32 bytes data + */ +#define TPM_I2C_MAX_BUF_SIZE 32 +#define TPM_I2C_RETRY_COUNT 32 +#define TPM_I2C_BUS_DELAY 1 /* msec */ +#define TPM_I2C_RETRY_DELAY_SHORT 2 /* msec */ +#define TPM_I2C_RETRY_DELAY_LONG 10 /* msec */ + +#define I2C_DRIVER_NAME "tpm_i2c_nuvoton" + +struct priv_data { + unsigned int intrs; +}; + +static s32 i2c_nuvoton_read_buf(struct i2c_client *client, u8 offset, u8 size, + u8 *data) +{ + s32 status; + + status = i2c_smbus_read_i2c_block_data(client, offset, size, data); + dev_dbg(&client->dev, + "%s(offset=%u size=%u data=%*ph) -> sts=%d\n", __func__, + offset, size, (int)size, data, status); + return status; +} + +static s32 i2c_nuvoton_write_buf(struct i2c_client *client, u8 offset, u8 size, + u8 *data) +{ + s32 status; + + status = i2c_smbus_write_i2c_block_data(client, offset, size, data); + dev_dbg(&client->dev, + "%s(offset=%u size=%u data=%*ph) -> sts=%d\n", __func__, + offset, size, (int)size, data, status); + return status; +} + +#define TPM_STS_VALID 0x80 +#define TPM_STS_COMMAND_READY 0x40 +#define TPM_STS_GO 0x20 +#define TPM_STS_DATA_AVAIL 0x10 +#define TPM_STS_EXPECT 0x08 +#define TPM_STS_RESPONSE_RETRY 0x02 +#define TPM_STS_ERR_VAL 0x07 /* bit2...bit0 reads always 0 */ + +#define TPM_I2C_SHORT_TIMEOUT 750 /* ms */ +#define TPM_I2C_LONG_TIMEOUT 2000 /* 2 sec */ + +/* read TPM_STS register */ +static u8 i2c_nuvoton_read_status(struct tpm_chip *chip) +{ + struct i2c_client *client = to_i2c_client(chip->dev); + s32 status; + u8 data; + + status = i2c_nuvoton_read_buf(client, TPM_STS, 1, &data); + if (status <= 0) { + dev_err(chip->dev, "%s() error return %d\n", __func__, + status); + data = TPM_STS_ERR_VAL; + } + + return data; +} + +/* write byte to TPM_STS register */ +static s32 i2c_nuvoton_write_status(struct i2c_client *client, u8 data) +{ + s32 status; + int i; + + /* this causes the current command to be aborted */ + for (i = 0, status = -1; i < TPM_I2C_RETRY_COUNT && status < 0; i++) { + status = i2c_nuvoton_write_buf(client, TPM_STS, 1, &data); + msleep(TPM_I2C_BUS_DELAY); + } + return status; +} + +/* write commandReady to TPM_STS register */ +static void i2c_nuvoton_ready(struct tpm_chip *chip) +{ + struct i2c_client *client = to_i2c_client(chip->dev); + s32 status; + + /* this causes the current command to be aborted */ + status = i2c_nuvoton_write_status(client, TPM_STS_COMMAND_READY); + if (status < 0) + dev_err(chip->dev, + "%s() fail to write TPM_STS.commandReady\n", __func__); +} + +/* read burstCount field from TPM_STS register + * return -1 on fail to read */ +static int i2c_nuvoton_get_burstcount(struct i2c_client *client, + struct tpm_chip *chip) +{ + unsigned long stop = jiffies + chip->vendor.timeout_d; + s32 status; + int burst_count = -1; + u8 data; + + /* wait for burstcount to be non-zero */ + do { + /* in I2C burstCount is 1 byte */ + status = i2c_nuvoton_read_buf(client, TPM_BURST_COUNT, 1, + &data); + if (status > 0 && data > 0) { + burst_count = min_t(u8, TPM_I2C_MAX_BUF_SIZE, data); + break; + } + msleep(TPM_I2C_BUS_DELAY); + } while (time_before(jiffies, stop)); + + return burst_count; +} + +/* + * WPCT301/NPCT501 SINT# supports only dataAvail + * any call to this function which is not waiting for dataAvail will + * set queue to NULL to avoid waiting for interrupt + */ +static bool i2c_nuvoton_check_status(struct tpm_chip *chip, u8 mask, u8 value) +{ + u8 status = i2c_nuvoton_read_status(chip); + return (status != TPM_STS_ERR_VAL) && ((status & mask) == value); +} + +static int i2c_nuvoton_wait_for_stat(struct tpm_chip *chip, u8 mask, u8 value, + u32 timeout, wait_queue_head_t *queue) +{ + if (chip->vendor.irq && queue) { + s32 rc; + DEFINE_WAIT(wait); + struct priv_data *priv = chip->vendor.priv; + unsigned int cur_intrs = priv->intrs; + + enable_irq(chip->vendor.irq); + rc = wait_event_interruptible_timeout(*queue, + cur_intrs != priv->intrs, + timeout); + if (rc > 0) + return 0; + /* At this point we know that the SINT pin is asserted, so we + * do not need to do i2c_nuvoton_check_status */ + } else { + unsigned long ten_msec, stop; + bool status_valid; + + /* check current status */ + status_valid = i2c_nuvoton_check_status(chip, mask, value); + if (status_valid) + return 0; + + /* use polling to wait for the event */ + ten_msec = jiffies + msecs_to_jiffies(TPM_I2C_RETRY_DELAY_LONG); + stop = jiffies + timeout; + do { + if (time_before(jiffies, ten_msec)) + msleep(TPM_I2C_RETRY_DELAY_SHORT); + else + msleep(TPM_I2C_RETRY_DELAY_LONG); + status_valid = i2c_nuvoton_check_status(chip, mask, + value); + if (status_valid) + return 0; + } while (time_before(jiffies, stop)); + } + dev_err(chip->dev, "%s(%02x, %02x) -> timeout\n", __func__, mask, + value); + return -ETIMEDOUT; +} + +/* wait for dataAvail field to be set in the TPM_STS register */ +static int i2c_nuvoton_wait_for_data_avail(struct tpm_chip *chip, u32 timeout, + wait_queue_head_t *queue) +{ + return i2c_nuvoton_wait_for_stat(chip, + TPM_STS_DATA_AVAIL | TPM_STS_VALID, + TPM_STS_DATA_AVAIL | TPM_STS_VALID, + timeout, queue); +} + +/* Read @count bytes into @buf from TPM_RD_FIFO register */ +static int i2c_nuvoton_recv_data(struct i2c_client *client, + struct tpm_chip *chip, u8 *buf, size_t count) +{ + s32 rc; + int burst_count, bytes2read, size = 0; + + while (size < count && + i2c_nuvoton_wait_for_data_avail(chip, + chip->vendor.timeout_c, + &chip->vendor.read_queue) == 0) { + burst_count = i2c_nuvoton_get_burstcount(client, chip); + if (burst_count < 0) { + dev_err(chip->dev, + "%s() fail to read burstCount=%d\n", __func__, + burst_count); + return -EIO; + } + bytes2read = min_t(size_t, burst_count, count - size); + rc = i2c_nuvoton_read_buf(client, TPM_DATA_FIFO_R, + bytes2read, &buf[size]); + if (rc < 0) { + dev_err(chip->dev, + "%s() fail on i2c_nuvoton_read_buf()=%d\n", + __func__, rc); + return -EIO; + } + dev_dbg(chip->dev, "%s(%d):", __func__, bytes2read); + size += bytes2read; + } + + return size; +} + +/* Read TPM command results */ +static int i2c_nuvoton_recv(struct tpm_chip *chip, u8 *buf, size_t count) +{ + struct device *dev = chip->dev; + struct i2c_client *client = to_i2c_client(dev); + s32 rc; + int expected, status, burst_count, retries, size = 0; + + if (count < TPM_HEADER_SIZE) { + i2c_nuvoton_ready(chip); /* return to idle */ + dev_err(dev, "%s() count < header size\n", __func__); + return -EIO; + } + for (retries = 0; retries < TPM_RETRY; retries++) { + if (retries > 0) { + /* if this is not the first trial, set responseRetry */ + i2c_nuvoton_write_status(client, + TPM_STS_RESPONSE_RETRY); + } + /* + * read first available (> 10 bytes), including: + * tag, paramsize, and result + */ + status = i2c_nuvoton_wait_for_data_avail( + chip, chip->vendor.timeout_c, &chip->vendor.read_queue); + if (status != 0) { + dev_err(dev, "%s() timeout on dataAvail\n", __func__); + size = -ETIMEDOUT; + continue; + } + burst_count = i2c_nuvoton_get_burstcount(client, chip); + if (burst_count < 0) { + dev_err(dev, "%s() fail to get burstCount\n", __func__); + size = -EIO; + continue; + } + size = i2c_nuvoton_recv_data(client, chip, buf, + burst_count); + if (size < TPM_HEADER_SIZE) { + dev_err(dev, "%s() fail to read header\n", __func__); + size = -EIO; + continue; + } + /* + * convert number of expected bytes field from big endian 32 bit + * to machine native + */ + expected = be32_to_cpu(*(__be32 *) (buf + 2)); + if (expected > count) { + dev_err(dev, "%s() expected > count\n", __func__); + size = -EIO; + continue; + } + rc = i2c_nuvoton_recv_data(client, chip, &buf[size], + expected - size); + size += rc; + if (rc < 0 || size < expected) { + dev_err(dev, "%s() fail to read remainder of result\n", + __func__); + size = -EIO; + continue; + } + if (i2c_nuvoton_wait_for_stat( + chip, TPM_STS_VALID | TPM_STS_DATA_AVAIL, + TPM_STS_VALID, chip->vendor.timeout_c, + NULL)) { + dev_err(dev, "%s() error left over data\n", __func__); + size = -ETIMEDOUT; + continue; + } + break; + } + i2c_nuvoton_ready(chip); + dev_dbg(chip->dev, "%s() -> %d\n", __func__, size); + return size; +} + +/* + * Send TPM command. + * + * If interrupts are used (signaled by an irq set in the vendor structure) + * tpm.c can skip polling for the data to be available as the interrupt is + * waited for here + */ +static int i2c_nuvoton_send(struct tpm_chip *chip, u8 *buf, size_t len) +{ + struct device *dev = chip->dev; + struct i2c_client *client = to_i2c_client(dev); + u32 ordinal; + size_t count = 0; + int burst_count, bytes2write, retries, rc = -EIO; + + for (retries = 0; retries < TPM_RETRY; retries++) { + i2c_nuvoton_ready(chip); + if (i2c_nuvoton_wait_for_stat(chip, TPM_STS_COMMAND_READY, + TPM_STS_COMMAND_READY, + chip->vendor.timeout_b, NULL)) { + dev_err(dev, "%s() timeout on commandReady\n", + __func__); + rc = -EIO; + continue; + } + rc = 0; + while (count < len - 1) { + burst_count = i2c_nuvoton_get_burstcount(client, + chip); + if (burst_count < 0) { + dev_err(dev, "%s() fail get burstCount\n", + __func__); + rc = -EIO; + break; + } + bytes2write = min_t(size_t, burst_count, + len - 1 - count); + rc = i2c_nuvoton_write_buf(client, TPM_DATA_FIFO_W, + bytes2write, &buf[count]); + if (rc < 0) { + dev_err(dev, "%s() fail i2cWriteBuf\n", + __func__); + break; + } + dev_dbg(dev, "%s(%d):", __func__, bytes2write); + count += bytes2write; + rc = i2c_nuvoton_wait_for_stat(chip, + TPM_STS_VALID | + TPM_STS_EXPECT, + TPM_STS_VALID | + TPM_STS_EXPECT, + chip->vendor.timeout_c, + NULL); + if (rc < 0) { + dev_err(dev, "%s() timeout on Expect\n", + __func__); + rc = -ETIMEDOUT; + break; + } + } + if (rc < 0) + continue; + + /* write last byte */ + rc = i2c_nuvoton_write_buf(client, TPM_DATA_FIFO_W, 1, + &buf[count]); + if (rc < 0) { + dev_err(dev, "%s() fail to write last byte\n", + __func__); + rc = -EIO; + continue; + } + dev_dbg(dev, "%s(last): %02x", __func__, buf[count]); + rc = i2c_nuvoton_wait_for_stat(chip, + TPM_STS_VALID | TPM_STS_EXPECT, + TPM_STS_VALID, + chip->vendor.timeout_c, NULL); + if (rc) { + dev_err(dev, "%s() timeout on Expect to clear\n", + __func__); + rc = -ETIMEDOUT; + continue; + } + break; + } + if (rc < 0) { + /* retries == TPM_RETRY */ + i2c_nuvoton_ready(chip); + return rc; + } + /* execute the TPM command */ + rc = i2c_nuvoton_write_status(client, TPM_STS_GO); + if (rc < 0) { + dev_err(dev, "%s() fail to write Go\n", __func__); + i2c_nuvoton_ready(chip); + return rc; + } + ordinal = be32_to_cpu(*((__be32 *) (buf + 6))); + rc = i2c_nuvoton_wait_for_data_avail(chip, + tpm_calc_ordinal_duration(chip, + ordinal), + &chip->vendor.read_queue); + if (rc) { + dev_err(dev, "%s() timeout command duration\n", __func__); + i2c_nuvoton_ready(chip); + return rc; + } + + dev_dbg(dev, "%s() -> %zd\n", __func__, len); + return len; +} + +static bool i2c_nuvoton_req_canceled(struct tpm_chip *chip, u8 status) +{ + return (status == TPM_STS_COMMAND_READY); +} + +static const struct file_operations i2c_nuvoton_ops = { + .owner = THIS_MODULE, + .llseek = no_llseek, + .open = tpm_open, + .read = tpm_read, + .write = tpm_write, + .release = tpm_release, +}; + +static DEVICE_ATTR(pubek, S_IRUGO, tpm_show_pubek, NULL); +static DEVICE_ATTR(pcrs, S_IRUGO, tpm_show_pcrs, NULL); +static DEVICE_ATTR(enabled, S_IRUGO, tpm_show_enabled, NULL); +static DEVICE_ATTR(active, S_IRUGO, tpm_show_active, NULL); +static DEVICE_ATTR(owned, S_IRUGO, tpm_show_owned, NULL); +static DEVICE_ATTR(temp_deactivated, S_IRUGO, tpm_show_temp_deactivated, NULL); +static DEVICE_ATTR(caps, S_IRUGO, tpm_show_caps, NULL); +static DEVICE_ATTR(cancel, S_IWUSR | S_IWGRP, NULL, tpm_store_cancel); +static DEVICE_ATTR(durations, S_IRUGO, tpm_show_durations, NULL); +static DEVICE_ATTR(timeouts, S_IRUGO, tpm_show_timeouts, NULL); + +static struct attribute *i2c_nuvoton_attrs[] = { + &dev_attr_pubek.attr, + &dev_attr_pcrs.attr, + &dev_attr_enabled.attr, + &dev_attr_active.attr, + &dev_attr_owned.attr, + &dev_attr_temp_deactivated.attr, + &dev_attr_caps.attr, + &dev_attr_cancel.attr, + &dev_attr_durations.attr, + &dev_attr_timeouts.attr, + NULL, +}; + +static struct attribute_group i2c_nuvoton_attr_grp = { + .attrs = i2c_nuvoton_attrs +}; + +static const struct tpm_vendor_specific tpm_i2c = { + .status = i2c_nuvoton_read_status, + .recv = i2c_nuvoton_recv, + .send = i2c_nuvoton_send, + .cancel = i2c_nuvoton_ready, + .req_complete_mask = TPM_STS_DATA_AVAIL | TPM_STS_VALID, + .req_complete_val = TPM_STS_DATA_AVAIL | TPM_STS_VALID, + .req_canceled = i2c_nuvoton_req_canceled, + .attr_group = &i2c_nuvoton_attr_grp, + .miscdev.fops = &i2c_nuvoton_ops, +}; + +/* The only purpose for the handler is to signal to any waiting threads that + * the interrupt is currently being asserted. The driver does not do any + * processing triggered by interrupts, and the chip provides no way to mask at + * the source (plus that would be slow over I2C). Run the IRQ as a one-shot, + * this means it cannot be shared. */ +static irqreturn_t i2c_nuvoton_int_handler(int dummy, void *dev_id) +{ + struct tpm_chip *chip = dev_id; + struct priv_data *priv = chip->vendor.priv; + + priv->intrs++; + wake_up(&chip->vendor.read_queue); + disable_irq_nosync(chip->vendor.irq); + return IRQ_HANDLED; +} + +static int get_vid(struct i2c_client *client, u32 *res) +{ + static const u8 vid_did_rid_value[] = { 0x50, 0x10, 0xfe }; + u32 temp; + s32 rc; + + if (!i2c_check_functionality(client->adapter, I2C_FUNC_SMBUS_BYTE_DATA)) + return -ENODEV; + rc = i2c_nuvoton_read_buf(client, TPM_VID_DID_RID, 4, (u8 *)&temp); + if (rc < 0) + return rc; + + /* check WPCT301 values - ignore RID */ + if (memcmp(&temp, vid_did_rid_value, sizeof(vid_did_rid_value))) { + /* + * f/w rev 2.81 has an issue where the VID_DID_RID is not + * reporting the right value. so give it another chance at + * offset 0x20 (FIFO_W). + */ + rc = i2c_nuvoton_read_buf(client, TPM_DATA_FIFO_W, 4, + (u8 *) (&temp)); + if (rc < 0) + return rc; + + /* check WPCT301 values - ignore RID */ + if (memcmp(&temp, vid_did_rid_value, + sizeof(vid_did_rid_value))) + return -ENODEV; + } + + *res = temp; + return 0; +} + +static int i2c_nuvoton_probe(struct i2c_client *client, + const struct i2c_device_id *id) +{ + int rc; + struct tpm_chip *chip; + struct device *dev = &client->dev; + u32 vid = 0; + + rc = get_vid(client, &vid); + if (rc) + return rc; + + dev_info(dev, "VID: %04X DID: %02X RID: %02X\n", (u16) vid, + (u8) (vid >> 16), (u8) (vid >> 24)); + + chip = tpm_register_hardware(dev, &tpm_i2c); + if (!chip) { + dev_err(dev, "%s() error in tpm_register_hardware\n", __func__); + return -ENODEV; + } + + chip->vendor.priv = devm_kzalloc(dev, sizeof(struct priv_data), + GFP_KERNEL); + init_waitqueue_head(&chip->vendor.read_queue); + init_waitqueue_head(&chip->vendor.int_queue); + + /* Default timeouts */ + chip->vendor.timeout_a = msecs_to_jiffies(TPM_I2C_SHORT_TIMEOUT); + chip->vendor.timeout_b = msecs_to_jiffies(TPM_I2C_LONG_TIMEOUT); + chip->vendor.timeout_c = msecs_to_jiffies(TPM_I2C_SHORT_TIMEOUT); + chip->vendor.timeout_d = msecs_to_jiffies(TPM_I2C_SHORT_TIMEOUT); + + /* + * I2C intfcaps (interrupt capabilitieis) in the chip are hard coded to: + * TPM_INTF_INT_LEVEL_LOW | TPM_INTF_DATA_AVAIL_INT + * The IRQ should be set in the i2c_board_info (which is done + * automatically in of_i2c_register_devices, for device tree users */ + chip->vendor.irq = client->irq; + + if (chip->vendor.irq) { + dev_dbg(dev, "%s() chip-vendor.irq\n", __func__); + rc = devm_request_irq(dev, chip->vendor.irq, + i2c_nuvoton_int_handler, + IRQF_TRIGGER_LOW, + chip->vendor.miscdev.name, + chip); + if (rc) { + dev_err(dev, "%s() Unable to request irq: %d for use\n", + __func__, chip->vendor.irq); + chip->vendor.irq = 0; + } else { + /* Clear any pending interrupt */ + i2c_nuvoton_ready(chip); + /* - wait for TPM_STS==0xA0 (stsValid, commandReady) */ + rc = i2c_nuvoton_wait_for_stat(chip, + TPM_STS_COMMAND_READY, + TPM_STS_COMMAND_READY, + chip->vendor.timeout_b, + NULL); + if (rc == 0) { + /* + * TIS is in ready state + * write dummy byte to enter reception state + * TPM_DATA_FIFO_W <- rc (0) + */ + rc = i2c_nuvoton_write_buf(client, + TPM_DATA_FIFO_W, + 1, (u8 *) (&rc)); + if (rc < 0) + goto out_err; + /* TPM_STS <- 0x40 (commandReady) */ + i2c_nuvoton_ready(chip); + } else { + /* + * timeout_b reached - command was + * aborted. TIS should now be in idle state - + * only TPM_STS_VALID should be set + */ + if (i2c_nuvoton_read_status(chip) != + TPM_STS_VALID) { + rc = -EIO; + goto out_err; + } + } + } + } + + if (tpm_get_timeouts(chip)) { + rc = -ENODEV; + goto out_err; + } + + if (tpm_do_selftest(chip)) { + rc = -ENODEV; + goto out_err; + } + + return 0; + +out_err: + tpm_dev_vendor_release(chip); + tpm_remove_hardware(chip->dev); + return rc; +} + +static int i2c_nuvoton_remove(struct i2c_client *client) +{ + struct device *dev = &(client->dev); + struct tpm_chip *chip = dev_get_drvdata(dev); + + if (chip) + tpm_dev_vendor_release(chip); + tpm_remove_hardware(dev); + kfree(chip); + return 0; +} + + +static const struct i2c_device_id i2c_nuvoton_id[] = { + {I2C_DRIVER_NAME, 0}, + {} +}; +MODULE_DEVICE_TABLE(i2c, i2c_nuvoton_id); + +#ifdef CONFIG_OF +static const struct of_device_id i2c_nuvoton_of_match[] = { + {.compatible = "nuvoton,npct501"}, + {.compatible = "winbond,wpct301"}, + {}, +}; +MODULE_DEVICE_TABLE(of, i2c_nuvoton_of_match); +#endif + +static SIMPLE_DEV_PM_OPS(i2c_nuvoton_pm_ops, tpm_pm_suspend, tpm_pm_resume); + +static struct i2c_driver i2c_nuvoton_driver = { + .id_table = i2c_nuvoton_id, + .probe = i2c_nuvoton_probe, + .remove = i2c_nuvoton_remove, + .driver = { + .name = I2C_DRIVER_NAME, + .owner = THIS_MODULE, + .pm = &i2c_nuvoton_pm_ops, + .of_match_table = of_match_ptr(i2c_nuvoton_of_match), + }, +}; + +module_i2c_driver(i2c_nuvoton_driver); + +MODULE_AUTHOR("Dan Morav (dan.morav@nuvoton.com)"); +MODULE_DESCRIPTION("Nuvoton TPM I2C Driver"); +MODULE_LICENSE("GPL"); -- cgit v1.2.3 From a2871c62e1865c45f87a9343de76f727fb7a0ffd Mon Sep 17 00:00:00 2001 From: Jason Gunthorpe Date: Sun, 6 Oct 2013 12:43:36 -0600 Subject: tpm: Add support for Atmel I2C TPMs This is based on the work of Teddy Reed published on GitHub: https://github.com/theopolis/tpm-i2c-atmel.git 34894b988b67e0ae55088d6388e77b0dbf10c07d That driver was never merged, I have taken it as a starting port, forward ported, tested and revised the driver: - Make it broadly textually similar to the Infineon and Nuvoton I2C driver - Place everything in a format suitable for mainline inclusion - Use high level I2C functions i2c_master_send and i2c_master_recv for data xfer - Use the timeout system from the core code, by faking out a status register - Only I2C transfer the number of bytes in the reply, not a fixed message size. - checkpatch cleanups - Testing on ARM Kirkwood, with this device tree, using a AT97SC3204T-X1A180 tpm@29 { compatible = "atmel,at97sc3204t"; reg = <0x29>; }; Signed-off-by: Teddy Reed [jgg: revised and tested] Signed-off-by: Jason Gunthorpe [phuewe: minor whitespace changes] Signed-off-by: Peter Huewe --- .../devicetree/bindings/i2c/trivial-devices.txt | 1 + drivers/char/tpm/Kconfig | 9 + drivers/char/tpm/Makefile | 1 + drivers/char/tpm/tpm_i2c_atmel.c | 284 +++++++++++++++++++++ 4 files changed, 295 insertions(+) create mode 100644 drivers/char/tpm/tpm_i2c_atmel.c (limited to 'Documentation') diff --git a/Documentation/devicetree/bindings/i2c/trivial-devices.txt b/Documentation/devicetree/bindings/i2c/trivial-devices.txt index 58454bdfa20e..f1fb26eed0e9 100644 --- a/Documentation/devicetree/bindings/i2c/trivial-devices.txt +++ b/Documentation/devicetree/bindings/i2c/trivial-devices.txt @@ -15,6 +15,7 @@ adi,adt7461 +/-1C TDM Extended Temp Range I.C adt7461 +/-1C TDM Extended Temp Range I.C at,24c08 i2c serial eeprom (24cxx) atmel,24c02 i2c serial eeprom (24cxx) +atmel,at97sc3204t i2c trusted platform module (TPM) catalyst,24c32 i2c serial eeprom dallas,ds1307 64 x 8, Serial, I2C Real-Time Clock dallas,ds1338 I2C RTC with 56-Byte NV RAM diff --git a/drivers/char/tpm/Kconfig b/drivers/char/tpm/Kconfig index ade71c134a1a..f9085860b716 100644 --- a/drivers/char/tpm/Kconfig +++ b/drivers/char/tpm/Kconfig @@ -33,6 +33,15 @@ config TCG_TIS from within Linux. To compile this driver as a module, choose M here; the module will be called tpm_tis. +config TCG_TIS_I2C_ATMEL + tristate "TPM Interface Specification 1.2 Interface (I2C - Atmel)" + depends on I2C + ---help--- + If you have an Atmel I2C TPM security chip say Yes and it will be + accessible from within Linux. + To compile this driver as a module, choose M here; the module will + be called tpm_tis_i2c_atmel. + config TCG_TIS_I2C_INFINEON tristate "TPM Interface Specification 1.2 Interface (I2C - Infineon)" depends on I2C diff --git a/drivers/char/tpm/Makefile b/drivers/char/tpm/Makefile index 07ee67cf3341..b80a4000daee 100644 --- a/drivers/char/tpm/Makefile +++ b/drivers/char/tpm/Makefile @@ -13,6 +13,7 @@ ifdef CONFIG_TCG_IBMVTPM endif endif obj-$(CONFIG_TCG_TIS) += tpm_tis.o +obj-$(CONFIG_TCG_TIS_I2C_ATMEL) += tpm_i2c_atmel.o obj-$(CONFIG_TCG_TIS_I2C_INFINEON) += tpm_i2c_infineon.o obj-$(CONFIG_TCG_TIS_I2C_NUVOTON) += tpm_i2c_nuvoton.o obj-$(CONFIG_TCG_NSC) += tpm_nsc.o diff --git a/drivers/char/tpm/tpm_i2c_atmel.c b/drivers/char/tpm/tpm_i2c_atmel.c new file mode 100644 index 000000000000..c3cd7fe481a1 --- /dev/null +++ b/drivers/char/tpm/tpm_i2c_atmel.c @@ -0,0 +1,284 @@ +/* + * ATMEL I2C TPM AT97SC3204T + * + * Copyright (C) 2012 V Lab Technologies + * Teddy Reed + * Copyright (C) 2013, Obsidian Research Corp. + * Jason Gunthorpe + * Device driver for ATMEL I2C TPMs. + * + * Teddy Reed determined the basic I2C command flow, unlike other I2C TPM + * devices the raw TCG formatted TPM command data is written via I2C and then + * raw TCG formatted TPM command data is returned via I2C. + * + * TGC status/locality/etc functions seen in the LPC implementation do not + * seem to be present. + * + * This program is free software: you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation, either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see http://www.gnu.org/licenses/>. + */ +#include +#include +#include +#include +#include +#include "tpm.h" + +#define I2C_DRIVER_NAME "tpm_i2c_atmel" + +#define TPM_I2C_SHORT_TIMEOUT 750 /* ms */ +#define TPM_I2C_LONG_TIMEOUT 2000 /* 2 sec */ + +#define ATMEL_STS_OK 1 + +struct priv_data { + size_t len; + /* This is the amount we read on the first try. 25 was chosen to fit a + * fair number of read responses in the buffer so a 2nd retry can be + * avoided in small message cases. */ + u8 buffer[sizeof(struct tpm_output_header) + 25]; +}; + +static int i2c_atmel_send(struct tpm_chip *chip, u8 *buf, size_t len) +{ + struct priv_data *priv = chip->vendor.priv; + struct i2c_client *client = to_i2c_client(chip->dev); + s32 status; + + priv->len = 0; + + if (len <= 2) + return -EIO; + + status = i2c_master_send(client, buf, len); + + dev_dbg(chip->dev, + "%s(buf=%*ph len=%0zx) -> sts=%d\n", __func__, + (int)min_t(size_t, 64, len), buf, len, status); + return status; +} + +static int i2c_atmel_recv(struct tpm_chip *chip, u8 *buf, size_t count) +{ + struct priv_data *priv = chip->vendor.priv; + struct i2c_client *client = to_i2c_client(chip->dev); + struct tpm_output_header *hdr = + (struct tpm_output_header *)priv->buffer; + u32 expected_len; + int rc; + + if (priv->len == 0) + return -EIO; + + /* Get the message size from the message header, if we didn't get the + * whole message in read_status then we need to re-read the + * message. */ + expected_len = be32_to_cpu(hdr->length); + if (expected_len > count) + return -ENOMEM; + + if (priv->len >= expected_len) { + dev_dbg(chip->dev, + "%s early(buf=%*ph count=%0zx) -> ret=%d\n", __func__, + (int)min_t(size_t, 64, expected_len), buf, count, + expected_len); + memcpy(buf, priv->buffer, expected_len); + return expected_len; + } + + rc = i2c_master_recv(client, buf, expected_len); + dev_dbg(chip->dev, + "%s reread(buf=%*ph count=%0zx) -> ret=%d\n", __func__, + (int)min_t(size_t, 64, expected_len), buf, count, + expected_len); + return rc; +} + +static void i2c_atmel_cancel(struct tpm_chip *chip) +{ + dev_err(chip->dev, "TPM operation cancellation was requested, but is not supported"); +} + +static u8 i2c_atmel_read_status(struct tpm_chip *chip) +{ + struct priv_data *priv = chip->vendor.priv; + struct i2c_client *client = to_i2c_client(chip->dev); + int rc; + + /* The TPM fails the I2C read until it is ready, so we do the entire + * transfer here and buffer it locally. This way the common code can + * properly handle the timeouts. */ + priv->len = 0; + memset(priv->buffer, 0, sizeof(priv->buffer)); + + + /* Once the TPM has completed the command the command remains readable + * until another command is issued. */ + rc = i2c_master_recv(client, priv->buffer, sizeof(priv->buffer)); + dev_dbg(chip->dev, + "%s: sts=%d", __func__, rc); + if (rc <= 0) + return 0; + + priv->len = rc; + + return ATMEL_STS_OK; +} + +static const struct file_operations i2c_atmel_ops = { + .owner = THIS_MODULE, + .llseek = no_llseek, + .open = tpm_open, + .read = tpm_read, + .write = tpm_write, + .release = tpm_release, +}; + +static DEVICE_ATTR(pubek, S_IRUGO, tpm_show_pubek, NULL); +static DEVICE_ATTR(pcrs, S_IRUGO, tpm_show_pcrs, NULL); +static DEVICE_ATTR(enabled, S_IRUGO, tpm_show_enabled, NULL); +static DEVICE_ATTR(active, S_IRUGO, tpm_show_active, NULL); +static DEVICE_ATTR(owned, S_IRUGO, tpm_show_owned, NULL); +static DEVICE_ATTR(temp_deactivated, S_IRUGO, tpm_show_temp_deactivated, NULL); +static DEVICE_ATTR(caps, S_IRUGO, tpm_show_caps, NULL); +static DEVICE_ATTR(cancel, S_IWUSR | S_IWGRP, NULL, tpm_store_cancel); +static DEVICE_ATTR(durations, S_IRUGO, tpm_show_durations, NULL); +static DEVICE_ATTR(timeouts, S_IRUGO, tpm_show_timeouts, NULL); + +static struct attribute *i2c_atmel_attrs[] = { + &dev_attr_pubek.attr, + &dev_attr_pcrs.attr, + &dev_attr_enabled.attr, + &dev_attr_active.attr, + &dev_attr_owned.attr, + &dev_attr_temp_deactivated.attr, + &dev_attr_caps.attr, + &dev_attr_cancel.attr, + &dev_attr_durations.attr, + &dev_attr_timeouts.attr, + NULL, +}; + +static struct attribute_group i2c_atmel_attr_grp = { + .attrs = i2c_atmel_attrs +}; + +static bool i2c_atmel_req_canceled(struct tpm_chip *chip, u8 status) +{ + return 0; +} + +static const struct tpm_vendor_specific i2c_atmel = { + .status = i2c_atmel_read_status, + .recv = i2c_atmel_recv, + .send = i2c_atmel_send, + .cancel = i2c_atmel_cancel, + .req_complete_mask = ATMEL_STS_OK, + .req_complete_val = ATMEL_STS_OK, + .req_canceled = i2c_atmel_req_canceled, + .attr_group = &i2c_atmel_attr_grp, + .miscdev.fops = &i2c_atmel_ops, +}; + +static int i2c_atmel_probe(struct i2c_client *client, + const struct i2c_device_id *id) +{ + int rc; + struct tpm_chip *chip; + struct device *dev = &client->dev; + + if (!i2c_check_functionality(client->adapter, I2C_FUNC_I2C)) + return -ENODEV; + + chip = tpm_register_hardware(dev, &i2c_atmel); + if (!chip) { + dev_err(dev, "%s() error in tpm_register_hardware\n", __func__); + return -ENODEV; + } + + chip->vendor.priv = devm_kzalloc(dev, sizeof(struct priv_data), + GFP_KERNEL); + + /* Default timeouts */ + chip->vendor.timeout_a = msecs_to_jiffies(TPM_I2C_SHORT_TIMEOUT); + chip->vendor.timeout_b = msecs_to_jiffies(TPM_I2C_LONG_TIMEOUT); + chip->vendor.timeout_c = msecs_to_jiffies(TPM_I2C_SHORT_TIMEOUT); + chip->vendor.timeout_d = msecs_to_jiffies(TPM_I2C_SHORT_TIMEOUT); + chip->vendor.irq = 0; + + /* There is no known way to probe for this device, and all version + * information seems to be read via TPM commands. Thus we rely on the + * TPM startup process in the common code to detect the device. */ + if (tpm_get_timeouts(chip)) { + rc = -ENODEV; + goto out_err; + } + + if (tpm_do_selftest(chip)) { + rc = -ENODEV; + goto out_err; + } + + return 0; + +out_err: + tpm_dev_vendor_release(chip); + tpm_remove_hardware(chip->dev); + return rc; +} + +static int i2c_atmel_remove(struct i2c_client *client) +{ + struct device *dev = &(client->dev); + struct tpm_chip *chip = dev_get_drvdata(dev); + + if (chip) + tpm_dev_vendor_release(chip); + tpm_remove_hardware(dev); + kfree(chip); + return 0; +} + +static const struct i2c_device_id i2c_atmel_id[] = { + {I2C_DRIVER_NAME, 0}, + {} +}; +MODULE_DEVICE_TABLE(i2c, i2c_atmel_id); + +#ifdef CONFIG_OF +static const struct of_device_id i2c_atmel_of_match[] = { + {.compatible = "atmel,at97sc3204t"}, + {}, +}; +MODULE_DEVICE_TABLE(of, i2c_atmel_of_match); +#endif + +static SIMPLE_DEV_PM_OPS(i2c_atmel_pm_ops, tpm_pm_suspend, tpm_pm_resume); + +static struct i2c_driver i2c_atmel_driver = { + .id_table = i2c_atmel_id, + .probe = i2c_atmel_probe, + .remove = i2c_atmel_remove, + .driver = { + .name = I2C_DRIVER_NAME, + .owner = THIS_MODULE, + .pm = &i2c_atmel_pm_ops, + .of_match_table = of_match_ptr(i2c_atmel_of_match), + }, +}; + +module_i2c_driver(i2c_atmel_driver); + +MODULE_AUTHOR("Jason Gunthorpe "); +MODULE_DESCRIPTION("Atmel TPM I2C Driver"); +MODULE_LICENSE("GPL"); -- cgit v1.2.3 From adf53a778a0a5a5dc9103509da4a9719046e5310 Mon Sep 17 00:00:00 2001 From: Roberto Sassu Date: Fri, 7 Jun 2013 12:16:29 +0200 Subject: ima: new templates management mechanism MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The original 'ima' template is fixed length, containing the filedata hash and pathname. The filedata hash is limited to 20 bytes (md5/sha1). The pathname is a null terminated string, limited to 255 characters. To overcome these limitations and to add additional file metadata, it is necessary to extend the current version of IMA by defining additional templates. The main reason to introduce this feature is that, each time a new template is defined, the functions that generate and display the measurement list would include the code for handling a new format and, thus, would significantly grow over time. This patch set solves this problem by separating the template management from the remaining IMA code. The core of this solution is the definition of two new data structures: a template descriptor, to determine which information should be included in the measurement list, and a template field, to generate and display data of a given type. To define a new template field, developers define the field identifier and implement two functions, init() and show(), respectively to generate and display measurement entries. Initially, this patch set defines the following template fields (support for additional data types will be added later):  - 'd': the digest of the event (i.e. the digest of a measured file),         calculated with the SHA1 or MD5 hash algorithm;  - 'n': the name of the event (i.e. the file name), with size up to         255 bytes;  - 'd-ng': the digest of the event, calculated with an arbitrary hash            algorithm (field format: [:]digest, where the digest            prefix is shown only if the hash algorithm is not SHA1 or MD5);  - 'n-ng': the name of the event, without size limitations. Defining a new template descriptor requires specifying the template format, a string of field identifiers separated by the '|' character. This patch set defines the following template descriptors:  - "ima": its format is 'd|n';  - "ima-ng" (default): its format is 'd-ng|n-ng' Further details about the new template architecture can be found in Documentation/security/IMA-templates.txt. Changelog: - don't defer calling ima_init_template() - Mimi - don't define ima_lookup_template_desc() until used - Mimi - squashed with documentation patch - Mimi Signed-off-by: Roberto Sassu Signed-off-by: Mimi Zohar --- Documentation/security/00-INDEX | 2 + Documentation/security/IMA-templates.txt | 87 ++++++++++++++++++++++++ security/integrity/ima/Makefile | 2 +- security/integrity/ima/ima.h | 29 ++++++++ security/integrity/ima/ima_init.c | 4 ++ security/integrity/ima/ima_template.c | 112 +++++++++++++++++++++++++++++++ 6 files changed, 235 insertions(+), 1 deletion(-) create mode 100644 Documentation/security/IMA-templates.txt create mode 100644 security/integrity/ima/ima_template.c (limited to 'Documentation') diff --git a/Documentation/security/00-INDEX b/Documentation/security/00-INDEX index 414235c1fcfc..45c82fd3e9d3 100644 --- a/Documentation/security/00-INDEX +++ b/Documentation/security/00-INDEX @@ -22,3 +22,5 @@ keys.txt - description of the kernel key retention service. tomoyo.txt - documentation on the TOMOYO Linux Security Module. +IMA-templates.txt + - documentation on the template management mechanism for IMA. diff --git a/Documentation/security/IMA-templates.txt b/Documentation/security/IMA-templates.txt new file mode 100644 index 000000000000..a777e5f1df5b --- /dev/null +++ b/Documentation/security/IMA-templates.txt @@ -0,0 +1,87 @@ + IMA Template Management Mechanism + + +==== INTRODUCTION ==== + +The original 'ima' template is fixed length, containing the filedata hash +and pathname. The filedata hash is limited to 20 bytes (md5/sha1). +The pathname is a null terminated string, limited to 255 characters. +To overcome these limitations and to add additional file metadata, it is +necessary to extend the current version of IMA by defining additional +templates. For example, information that could be possibly reported are +the inode UID/GID or the LSM labels either of the inode and of the process +that is accessing it. + +However, the main problem to introduce this feature is that, each time +a new template is defined, the functions that generate and display +the measurements list would include the code for handling a new format +and, thus, would significantly grow over the time. + +The proposed solution solves this problem by separating the template +management from the remaining IMA code. The core of this solution is the +definition of two new data structures: a template descriptor, to determine +which information should be included in the measurement list; a template +field, to generate and display data of a given type. + +Managing templates with these structures is very simple. To support +a new data type, developers define the field identifier and implement +two functions, init() and show(), respectively to generate and display +measurement entries. Defining a new template descriptor requires +specifying the template format, a string of field identifiers separated +by the '|' character. While in the current implementation it is possible +to define new template descriptors only by adding their definition in the +template specific code (ima_template.c), in a future version it will be +possible to register a new template on a running kernel by supplying to IMA +the desired format string. In this version, IMA initializes at boot time +all defined template descriptors by translating the format into an array +of template fields structures taken from the set of the supported ones. + +After the initialization step, IMA will call ima_alloc_init_template() +(new function defined within the patches for the new template management +mechanism) to generate a new measurement entry by using the template +descriptor chosen through the kernel configuration or through the newly +introduced 'ima_template=' kernel command line parameter. It is during this +phase that the advantages of the new architecture are clearly shown: +the latter function will not contain specific code to handle a given template +but, instead, it simply calls the init() method of the template fields +associated to the chosen template descriptor and store the result (pointer +to allocated data and data length) in the measurement entry structure. + +The same mechanism is employed to display measurements entries. +The functions ima[_ascii]_measurements_show() retrieve, for each entry, +the template descriptor used to produce that entry and call the show() +method for each item of the array of template fields structures. + + + +==== SUPPORTED TEMPLATE FIELDS AND DESCRIPTORS ==== + +In the following, there is the list of supported template fields +('': description), that can be used to define new template +descriptors by adding their identifier to the format string +(support for more data types will be added later): + + - 'd': the digest of the event (i.e. the digest of a measured file), + calculated with the SHA1 or MD5 hash algorithm; + - 'n': the name of the event (i.e. the file name), with size up to 255 bytes; + - 'd-ng': the digest of the event, calculated with an arbitrary hash + algorithm (field format: [:]digest, where the digest + prefix is shown only if the hash algorithm is not SHA1 or MD5); + - 'n-ng': the name of the event, without size limitations. + + +Below, there is the list of defined template descriptors: + - "ima": its format is 'd|n'; + - "ima-ng" (default): its format is 'd-ng|n-ng'. + + + +==== USE ==== + +To specify the template descriptor to be used to generate measurement entries, +currently the following methods are supported: + + - select a template descriptor among those supported in the kernel + configuration ('ima-ng' is the default choice); + - specify a template descriptor name from the kernel command line through + the 'ima_template=' parameter. diff --git a/security/integrity/ima/Makefile b/security/integrity/ima/Makefile index 56dfee7cbf61..7fe4ae37dbea 100644 --- a/security/integrity/ima/Makefile +++ b/security/integrity/ima/Makefile @@ -6,5 +6,5 @@ obj-$(CONFIG_IMA) += ima.o ima-y := ima_fs.o ima_queue.o ima_init.o ima_main.o ima_crypto.o ima_api.o \ - ima_policy.o + ima_policy.o ima_template.o ima-$(CONFIG_IMA_APPRAISE) += ima_appraise.o diff --git a/security/integrity/ima/ima.h b/security/integrity/ima/ima.h index da03d3389619..c85718f1f623 100644 --- a/security/integrity/ima/ima.h +++ b/security/integrity/ima/ima.h @@ -36,12 +36,39 @@ enum tpm_pcrs { TPM_PCR0 = 0, TPM_PCR8 = 8 }; #define IMA_HASH_BITS 9 #define IMA_MEASURE_HTABLE_SIZE (1 << IMA_HASH_BITS) +#define IMA_TEMPLATE_FIELD_ID_MAX_LEN 16 +#define IMA_TEMPLATE_NUM_FIELDS_MAX 15 + /* set during initialization */ extern int ima_initialized; extern int ima_used_chip; extern int ima_hash_algo; extern int ima_appraise; +/* IMA template field data definition */ +struct ima_field_data { + u8 *data; + u32 len; +}; + +/* IMA template field definition */ +struct ima_template_field { + const char field_id[IMA_TEMPLATE_FIELD_ID_MAX_LEN]; + int (*field_init) (struct integrity_iint_cache *iint, struct file *file, + const unsigned char *filename, + struct ima_field_data *field_data); + void (*field_show) (struct seq_file *m, enum ima_show_type show, + struct ima_field_data *field_data); +}; + +/* IMA template descriptor definition */ +struct ima_template_desc { + char *name; + char *fmt; + int num_fields; + struct ima_template_field **fields; +}; + /* IMA inode template definition */ struct ima_template_data { u8 digest[IMA_DIGEST_SIZE]; /* sha1/md5 measurement hash */ @@ -79,6 +106,8 @@ void ima_add_violation(struct file *file, const unsigned char *filename, const char *op, const char *cause); int ima_init_crypto(void); +int ima_init_template(void); + /* * used to protect h_table and sha_table */ diff --git a/security/integrity/ima/ima_init.c b/security/integrity/ima/ima_init.c index 50e15e6336c4..f84aec5412f3 100644 --- a/security/integrity/ima/ima_init.c +++ b/security/integrity/ima/ima_init.c @@ -99,6 +99,10 @@ int __init ima_init(void) rc = ima_init_crypto(); if (rc) return rc; + rc = ima_init_template(); + if (rc != 0) + return rc; + ima_add_boot_aggregate(); /* boot aggregate must be first entry */ ima_init_policy(); diff --git a/security/integrity/ima/ima_template.c b/security/integrity/ima/ima_template.c new file mode 100644 index 000000000000..7e86783ccf51 --- /dev/null +++ b/security/integrity/ima/ima_template.c @@ -0,0 +1,112 @@ +/* + * Copyright (C) 2013 Politecnico di Torino, Italy + * TORSEC group -- http://security.polito.it + * + * Author: Roberto Sassu + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation, version 2 of the + * License. + * + * File: ima_template.c + * Helpers to manage template descriptors. + */ +#include "ima.h" + +static struct ima_template_desc defined_templates[] = { +}; + +static struct ima_template_field supported_fields[] = { +}; + +static struct ima_template_field *ima_lookup_template_field( + const char *field_id) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(supported_fields); i++) + if (strncmp(supported_fields[i].field_id, field_id, + IMA_TEMPLATE_FIELD_ID_MAX_LEN) == 0) + return &supported_fields[i]; + return NULL; +} + +static int ima_template_fmt_size(char *template_fmt) +{ + char c; + int template_fmt_len = strlen(template_fmt); + int i = 0, j = 0; + + while (i < template_fmt_len) { + c = template_fmt[i]; + if (c == '|') + j++; + i++; + } + + return j + 1; +} + +static int template_desc_init_fields(char *template_fmt, + struct ima_template_field ***fields, + int *num_fields) +{ + char *c, *template_fmt_ptr = template_fmt; + int template_num_fields = ima_template_fmt_size(template_fmt); + int i, result = 0; + + if (template_num_fields > IMA_TEMPLATE_NUM_FIELDS_MAX) + return -EINVAL; + + *fields = kzalloc(template_num_fields * sizeof(*fields), GFP_KERNEL); + if (*fields == NULL) { + result = -ENOMEM; + goto out; + } + for (i = 0; (c = strsep(&template_fmt_ptr, "|")) != NULL && + i < template_num_fields; i++) { + struct ima_template_field *f = ima_lookup_template_field(c); + + if (!f) { + result = -ENOENT; + goto out; + } + (*fields)[i] = f; + } + *num_fields = i; + return 0; +out: + kfree(*fields); + *fields = NULL; + return result; +} + +static int init_defined_templates(void) +{ + int i = 0; + int result = 0; + + /* Init defined templates. */ + for (i = 0; i < ARRAY_SIZE(defined_templates); i++) { + struct ima_template_desc *template = &defined_templates[i]; + + result = template_desc_init_fields(template->fmt, + &(template->fields), + &(template->num_fields)); + if (result < 0) + return result; + } + return result; +} + +int ima_init_template(void) +{ + int result; + + result = init_defined_templates(); + if (result < 0) + return result; + + return 0; +} -- cgit v1.2.3 From 9b9d4ce592d283fc4c01da746c02a840c499bb7e Mon Sep 17 00:00:00 2001 From: Roberto Sassu Date: Fri, 7 Jun 2013 12:16:35 +0200 Subject: ima: define kernel parameter 'ima_template=' to change configured default This patch allows users to specify from the kernel command line the template descriptor, among those defined, that will be used to generate and display measurement entries. If an user specifies a wrong template, IMA reverts to the template descriptor set in the kernel configuration. Signed-off-by: Roberto Sassu Signed-off-by: Mimi Zohar --- Documentation/kernel-parameters.txt | 5 +++++ security/integrity/ima/ima_template.c | 31 +++++++++++++++++++++++++++++++ 2 files changed, 36 insertions(+) (limited to 'Documentation') diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 1a036cd972fb..2b78cb55ac34 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -1190,6 +1190,11 @@ bytes respectively. Such letter suffixes can also be entirely omitted. programs exec'd, files mmap'd for exec, and all files opened for read by uid=0. + ima_template= [IMA] + Select one of defined IMA measurements template formats. + Formats: { "ima" | "ima-ng" } + Default: "ima-ng" + init= [KNL] Format: Run specified binary instead of /sbin/init as init diff --git a/security/integrity/ima/ima_template.c b/security/integrity/ima/ima_template.c index c28ff9bf8f32..000221419f6c 100644 --- a/security/integrity/ima/ima_template.c +++ b/security/integrity/ima/ima_template.c @@ -12,6 +12,8 @@ * File: ima_template.c * Helpers to manage template descriptors. */ +#include + #include "ima.h" #include "ima_template_lib.h" @@ -32,6 +34,35 @@ static struct ima_template_field supported_fields[] = { }; static struct ima_template_desc *ima_template; +static struct ima_template_desc *lookup_template_desc(const char *name); + +static int __init ima_template_setup(char *str) +{ + struct ima_template_desc *template_desc; + int template_len = strlen(str); + + /* + * Verify that a template with the supplied name exists. + * If not, use CONFIG_IMA_DEFAULT_TEMPLATE. + */ + template_desc = lookup_template_desc(str); + if (!template_desc) + return 1; + + /* + * Verify whether the current hash algorithm is supported + * by the 'ima' template. + */ + if (template_len == 3 && strcmp(str, IMA_TEMPLATE_IMA_NAME) == 0 && + ima_hash_algo != HASH_ALGO_SHA1 && ima_hash_algo != HASH_ALGO_MD5) { + pr_err("IMA: template does not support hash alg\n"); + return 1; + } + + ima_template = template_desc; + return 1; +} +__setup("ima_template=", ima_template_setup); static struct ima_template_desc *lookup_template_desc(const char *name) { -- cgit v1.2.3 From e7a2ad7eb6f48ad80c70a22dd8167fb34b409466 Mon Sep 17 00:00:00 2001 From: Mimi Zohar Date: Fri, 7 Jun 2013 12:16:37 +0200 Subject: ima: enable support for larger default filedata hash algorithms The IMA measurement list contains two hashes - a template data hash and a filedata hash. The template data hash is committed to the TPM, which is limited, by the TPM v1.2 specification, to 20 bytes. The filedata hash is defined as 20 bytes as well. Now that support for variable length measurement list templates was added, the filedata hash is not limited to 20 bytes. This patch adds Kconfig support for defining larger default filedata hash algorithms and replacing the builtin default with one specified on the kernel command line. contains a list of hash algorithms. The Kconfig default hash algorithm is a subset of this list, but any hash algorithm included in the list can be specified at boot, using the 'ima_hash=' kernel command line option. Changelog v2: - update Kconfig Changelog: - support hashes that are configured - use generic HASH_ALGO_ definitions - add Kconfig support - hash_setup must be called only once (Dmitry) - removed trailing whitespaces (Roberto Sassu) Signed-off-by: Mimi Zohar Signed-off-by: Roberto Sassu --- Documentation/kernel-parameters.txt | 6 +++++- security/integrity/ima/Kconfig | 35 +++++++++++++++++++++++++++++++++++ security/integrity/ima/ima_main.c | 26 ++++++++++++++++++++++++-- 3 files changed, 64 insertions(+), 3 deletions(-) (limited to 'Documentation') diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 2b78cb55ac34..1e8761c89a2c 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -1181,9 +1181,13 @@ bytes respectively. Such letter suffixes can also be entirely omitted. owned by uid=0. ima_hash= [IMA] - Format: { "sha1" | "md5" } + Format: { md5 | sha1 | rmd160 | sha256 | sha384 + | sha512 | ... } default: "sha1" + The list of supported hash algorithms is defined + in crypto/hash_info.h. + ima_tcb [IMA] Load a policy which meets the needs of the Trusted Computing Base. This means IMA will measure all diff --git a/security/integrity/ima/Kconfig b/security/integrity/ima/Kconfig index de26cc873ae6..351a58ed56ab 100644 --- a/security/integrity/ima/Kconfig +++ b/security/integrity/ima/Kconfig @@ -71,6 +71,41 @@ config IMA_DEFAULT_TEMPLATE default "ima" if IMA_TEMPLATE default "ima-ng" if IMA_NG_TEMPLATE +choice + prompt "Default integrity hash algorithm" + default IMA_DEFAULT_HASH_SHA1 + depends on IMA + help + Select the default hash algorithm used for the measurement + list, integrity appraisal and audit log. The compiled default + hash algorithm can be overwritten using the kernel command + line 'ima_hash=' option. + + config IMA_DEFAULT_HASH_SHA1 + bool "SHA1 (default)" + depends on CRYPTO_SHA1 + + config IMA_DEFAULT_HASH_SHA256 + bool "SHA256" + depends on CRYPTO_SHA256 && !IMA_TEMPLATE + + config IMA_DEFAULT_HASH_SHA512 + bool "SHA512" + depends on CRYPTO_SHA512 && !IMA_TEMPLATE + + config IMA_DEFAULT_HASH_WP512 + bool "WP512" + depends on CRYPTO_WP512 && !IMA_TEMPLATE +endchoice + +config IMA_DEFAULT_HASH + string + depends on IMA + default "sha1" if IMA_DEFAULT_HASH_SHA1 + default "sha256" if IMA_DEFAULT_HASH_SHA256 + default "sha512" if IMA_DEFAULT_HASH_SHA512 + default "wp512" if IMA_DEFAULT_HASH_WP512 + config IMA_APPRAISE bool "Appraise integrity measurements" depends on IMA diff --git a/security/integrity/ima/ima_main.c b/security/integrity/ima/ima_main.c index 0b11bb49ac4f..14d4cb557894 100644 --- a/security/integrity/ima/ima_main.c +++ b/security/integrity/ima/ima_main.c @@ -37,11 +37,32 @@ int ima_appraise; #endif int ima_hash_algo = HASH_ALGO_SHA1; +static int hash_setup_done; static int __init hash_setup(char *str) { - if (strncmp(str, "md5", 3) == 0) - ima_hash_algo = HASH_ALGO_MD5; + struct ima_template_desc *template_desc = ima_template_desc_current(); + int i; + + if (hash_setup_done) + return 1; + + if (strcmp(template_desc->name, IMA_TEMPLATE_IMA_NAME) == 0) { + if (strncmp(str, "sha1", 4) == 0) + ima_hash_algo = HASH_ALGO_SHA1; + else if (strncmp(str, "md5", 3) == 0) + ima_hash_algo = HASH_ALGO_MD5; + goto out; + } + + for (i = 0; i < HASH_ALGO__LAST; i++) { + if (strcmp(str, hash_algo_name[i]) == 0) { + ima_hash_algo = i; + break; + } + } +out: + hash_setup_done = 1; return 1; } __setup("ima_hash=", hash_setup); @@ -306,6 +327,7 @@ static int __init init_ima(void) { int error; + hash_setup(CONFIG_IMA_DEFAULT_HASH); error = ima_init(); if (!error) ima_initialized = 1; -- cgit v1.2.3