* [PATCH 0/4] core-api ReST: assoc_array, atomic_ops, local_ops
@ 2016-11-24 12:42 Silvio Fricke
2016-11-24 12:42 ` [PATCH 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke
` (5 more replies)
0 siblings, 6 replies; 51+ messages in thread
From: Silvio Fricke @ 2016-11-24 12:42 UTC (permalink / raw)
To: linux-doc
Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez,
Mauro Carvalho Chehab, linux-kernel, Silvio Fricke
Hi,
Some more ReSTification of core-api's: assoc_array, atomic_ops and local_ops. A
fourth patch removes a warning about a bullet list without ending at
firmware_class.c
Thanks for review.
BR
Silvio
Silvio Fricke (4):
Documentation/assoc_array.txt: convert to ReST markup
Documentation/atomic_ops.txt: convert to ReST markup
Documentation/local_ops.txt: convert to ReST markup
firmware: remove warning at documentation generation time
Documentation/assoc_array.txt | 574 +---------
Documentation/atomic_ops.txt | 640 +-----------
Documentation/core-api/assoc_array.rst | 549 +++++++++-
Documentation/core-api/atomic_ops.rst | 669 +++++++++++-
Documentation/core-api/index.rst | 3 +-
Documentation/core-api/local_ops.rst | 208 +++-
Documentation/local_ops.txt | 191 +---
Documentation/process/volatile-considered-harmful.rst | 3 +-
drivers/base/firmware_class.c | 6 +-
9 files changed, 1435 insertions(+), 1408 deletions(-)
delete mode 100644 Documentation/assoc_array.txt
delete mode 100644 Documentation/atomic_ops.txt
create mode 100644 Documentation/core-api/assoc_array.rst
create mode 100644 Documentation/core-api/atomic_ops.rst
create mode 100644 Documentation/core-api/local_ops.rst
delete mode 100644 Documentation/local_ops.txt
base-commit: 9c240d757658a3ae9968dd309e674c61f07c7f48
--
git-series 0.9.1
^ permalink raw reply [flat|nested] 51+ messages in thread* [PATCH 1/4] Documentation/assoc_array.txt: convert to ReST markup 2016-11-24 12:42 [PATCH 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke @ 2016-11-24 12:42 ` Silvio Fricke 2016-11-24 12:42 ` [PATCH 2/4] Documentation/atomic_ops.txt: " Silvio Fricke ` (4 subsequent siblings) 5 siblings, 0 replies; 51+ messages in thread From: Silvio Fricke @ 2016-11-24 12:42 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Silvio Fricke ... and move to Documentation/core-api folder. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- Documentation/assoc_array.txt | 574 +-------------------------- Documentation/core-api/assoc_array.rst | 549 +++++++++++++++++++++++++- Documentation/core-api/index.rst | 1 +- 3 files changed, 550 insertions(+), 574 deletions(-) delete mode 100644 Documentation/assoc_array.txt create mode 100644 Documentation/core-api/assoc_array.rst diff --git a/Documentation/assoc_array.txt b/Documentation/assoc_array.txt deleted file mode 100644 index 2f2c6cd..0000000 --- a/Documentation/assoc_array.txt +++ /dev/null @@ -1,574 +0,0 @@ - ======================================== - 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 <linux/assoc_array.h>. 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 *object, const void *index_key); - - Return the bit position at which the index key of the specified object - differs from the given index key 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/Documentation/core-api/assoc_array.rst b/Documentation/core-api/assoc_array.rst new file mode 100644 index 0000000..67a3a50 --- /dev/null +++ b/Documentation/core-api/assoc_array.rst @@ -0,0 +1,549 @@ +======================================== +Generic Associative Array Implementation +======================================== + +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 ``<linux/assoc_array.h>``. 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 *object, const void *index_key); + +Return the bit position at which the index key of the specified object +differs from the given index key 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/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index f7ef7fd..480d9a3 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -7,6 +7,7 @@ Kernel and driver related documentation. .. toctree:: :maxdepth: 1 + assoc_array workqueue .. only:: subproject -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* [PATCH 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-24 12:42 [PATCH 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 2016-11-24 12:42 ` [PATCH 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke @ 2016-11-24 12:42 ` Silvio Fricke 2016-11-24 12:42 ` [PATCH 3/4] Documentation/local_ops.txt: " Silvio Fricke ` (3 subsequent siblings) 5 siblings, 0 replies; 51+ messages in thread From: Silvio Fricke @ 2016-11-24 12:42 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Silvio Fricke ... and move to core-api folder. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- Documentation/atomic_ops.txt | 640 +----------- Documentation/core-api/atomic_ops.rst | 669 +++++++++++- Documentation/core-api/index.rst | 1 +- Documentation/process/volatile-considered-harmful.rst | 3 +- 4 files changed, 673 insertions(+), 640 deletions(-) delete mode 100644 Documentation/atomic_ops.txt create mode 100644 Documentation/core-api/atomic_ops.rst diff --git a/Documentation/atomic_ops.txt b/Documentation/atomic_ops.txt deleted file mode 100644 index 6c5e8a9..0000000 --- a/Documentation/atomic_ops.txt +++ /dev/null @@ -1,640 +0,0 @@ - Semantics and Behavior of Atomic and - Bitmask Operations - - David S. Miller - - This document is intended to serve as a guide to Linux port -maintainers on how to implement atomic counter, bitops, and spinlock -interfaces properly. - - The atomic_t type should be defined as a signed integer and -the atomic_long_t type as a signed long integer. Also, they should -be made opaque such that any kind of cast to a normal C integer type -will fail. Something like the following should suffice: - - typedef struct { int counter; } atomic_t; - typedef struct { long counter; } atomic_long_t; - -Historically, counter has been declared volatile. This is now discouraged. -See Documentation/process/volatile-considered-harmful.rst for the complete rationale. - -local_t is very similar to atomic_t. If the counter is per CPU and only -updated by one CPU, local_t is probably more appropriate. Please see -Documentation/local_ops.txt for the semantics of local_t. - -The first operations to implement for atomic_t's are the initializers and -plain reads. - - #define ATOMIC_INIT(i) { (i) } - #define atomic_set(v, i) ((v)->counter = (i)) - -The first macro is used in definitions, such as: - -static atomic_t my_counter = ATOMIC_INIT(1); - -The initializer is atomic in that the return values of the atomic operations -are guaranteed to be correct reflecting the initialized value if the -initializer is used before runtime. If the initializer is used at runtime, a -proper implicit or explicit read memory barrier is needed before reading the -value with atomic_read from another thread. - -As with all of the atomic_ interfaces, replace the leading "atomic_" -with "atomic_long_" to operate on atomic_long_t. - -The second interface can be used at runtime, as in: - - struct foo { atomic_t counter; }; - ... - - struct foo *k; - - k = kmalloc(sizeof(*k), GFP_KERNEL); - if (!k) - return -ENOMEM; - atomic_set(&k->counter, 0); - -The setting is atomic in that the return values of the atomic operations by -all threads are guaranteed to be correct reflecting either the value that has -been set with this operation or set with another operation. A proper implicit -or explicit memory barrier is needed before the value set with the operation -is guaranteed to be readable with atomic_read from another thread. - -Next, we have: - - #define atomic_read(v) ((v)->counter) - -which simply reads the counter value currently visible to the calling thread. -The read is atomic in that the return value is guaranteed to be one of the -values initialized or modified with the interface operations if a proper -implicit or explicit memory barrier is used after possible runtime -initialization by any other thread and the value is modified only with the -interface operations. atomic_read does not guarantee that the runtime -initialization by any other thread is visible yet, so the user of the -interface must take care of that with a proper implicit or explicit memory -barrier. - -*** WARNING: atomic_read() and atomic_set() DO NOT IMPLY BARRIERS! *** - -Some architectures may choose to use the volatile keyword, barriers, or inline -assembly to guarantee some degree of immediacy for atomic_read() and -atomic_set(). This is not uniformly guaranteed, and may change in the future, -so all users of atomic_t should treat atomic_read() and atomic_set() as simple -C statements that may be reordered or optimized away entirely by the compiler -or processor, and explicitly invoke the appropriate compiler and/or memory -barrier for each use case. Failure to do so will result in code that may -suddenly break when used with different architectures or compiler -optimizations, or even changes in unrelated code which changes how the -compiler optimizes the section accessing atomic_t variables. - -*** YOU HAVE BEEN WARNED! *** - -Properly aligned pointers, longs, ints, and chars (and unsigned -equivalents) may be atomically loaded from and stored to in the same -sense as described for atomic_read() and atomic_set(). The READ_ONCE() -and WRITE_ONCE() macros should be used to prevent the compiler from using -optimizations that might otherwise optimize accesses out of existence on -the one hand, or that might create unsolicited accesses on the other. - -For example consider the following code: - - while (a > 0) - do_something(); - -If the compiler can prove that do_something() does not store to the -variable a, then the compiler is within its rights transforming this to -the following: - - tmp = a; - if (a > 0) - for (;;) - do_something(); - -If you don't want the compiler to do this (and you probably don't), then -you should use something like the following: - - while (READ_ONCE(a) < 0) - do_something(); - -Alternatively, you could place a barrier() call in the loop. - -For another example, consider the following code: - - tmp_a = a; - do_something_with(tmp_a); - do_something_else_with(tmp_a); - -If the compiler can prove that do_something_with() does not store to the -variable a, then the compiler is within its rights to manufacture an -additional load as follows: - - tmp_a = a; - do_something_with(tmp_a); - tmp_a = a; - do_something_else_with(tmp_a); - -This could fatally confuse your code if it expected the same value -to be passed to do_something_with() and do_something_else_with(). - -The compiler would be likely to manufacture this additional load if -do_something_with() was an inline function that made very heavy use -of registers: reloading from variable a could save a flush to the -stack and later reload. To prevent the compiler from attacking your -code in this manner, write the following: - - tmp_a = READ_ONCE(a); - do_something_with(tmp_a); - do_something_else_with(tmp_a); - -For a final example, consider the following code, assuming that the -variable a is set at boot time before the second CPU is brought online -and never changed later, so that memory barriers are not needed: - - if (a) - b = 9; - else - b = 42; - -The compiler is within its rights to manufacture an additional store -by transforming the above code into the following: - - b = 42; - if (a) - b = 9; - -This could come as a fatal surprise to other code running concurrently -that expected b to never have the value 42 if a was zero. To prevent -the compiler from doing this, write something like: - - if (a) - WRITE_ONCE(b, 9); - else - WRITE_ONCE(b, 42); - -Don't even -think- about doing this without proper use of memory barriers, -locks, or atomic operations if variable a can change at runtime! - -*** WARNING: READ_ONCE() OR WRITE_ONCE() DO NOT IMPLY A BARRIER! *** - -Now, we move onto the atomic operation interfaces typically implemented with -the help of assembly code. - - void atomic_add(int i, atomic_t *v); - void atomic_sub(int i, atomic_t *v); - void atomic_inc(atomic_t *v); - void atomic_dec(atomic_t *v); - -These four routines add and subtract integral values to/from the given -atomic_t value. The first two routines pass explicit integers by -which to make the adjustment, whereas the latter two use an implicit -adjustment value of "1". - -One very important aspect of these two routines is that they DO NOT -require any explicit memory barriers. They need only perform the -atomic_t counter update in an SMP safe manner. - -Next, we have: - - int atomic_inc_return(atomic_t *v); - int atomic_dec_return(atomic_t *v); - -These routines add 1 and subtract 1, respectively, from the given -atomic_t and return the new counter value after the operation is -performed. - -Unlike the above routines, it is required that these primitives -include explicit memory barriers that are performed before and after -the operation. It must be done such that all memory operations before -and after the atomic operation calls are strongly ordered with respect -to the atomic operation itself. - -For example, it should behave as if a smp_mb() call existed both -before and after the atomic operation. - -If the atomic instructions used in an implementation provide explicit -memory barrier semantics which satisfy the above requirements, that is -fine as well. - -Let's move on: - - int atomic_add_return(int i, atomic_t *v); - int atomic_sub_return(int i, atomic_t *v); - -These behave just like atomic_{inc,dec}_return() except that an -explicit counter adjustment is given instead of the implicit "1". -This means that like atomic_{inc,dec}_return(), the memory barrier -semantics are required. - -Next: - - int atomic_inc_and_test(atomic_t *v); - int atomic_dec_and_test(atomic_t *v); - -These two routines increment and decrement by 1, respectively, the -given atomic counter. They return a boolean indicating whether the -resulting counter value was zero or not. - -Again, these primitives provide explicit memory barrier semantics around -the atomic operation. - - int atomic_sub_and_test(int i, atomic_t *v); - -This is identical to atomic_dec_and_test() except that an explicit -decrement is given instead of the implicit "1". This primitive must -provide explicit memory barrier semantics around the operation. - - int atomic_add_negative(int i, atomic_t *v); - -The given increment is added to the given atomic counter value. A boolean -is return which indicates whether the resulting counter value is negative. -This primitive must provide explicit memory barrier semantics around -the operation. - -Then: - - int atomic_xchg(atomic_t *v, int new); - -This performs an atomic exchange operation on the atomic variable v, setting -the given new value. It returns the old value that the atomic variable v had -just before the operation. - -atomic_xchg must provide explicit memory barriers around the operation. - - int atomic_cmpxchg(atomic_t *v, int old, int new); - -This performs an atomic compare exchange operation on the atomic value v, -with the given old and new values. Like all atomic_xxx operations, -atomic_cmpxchg will only satisfy its atomicity semantics as long as all -other accesses of *v are performed through atomic_xxx operations. - -atomic_cmpxchg must provide explicit memory barriers around the operation, -although if the comparison fails then no memory ordering guarantees are -required. - -The semantics for atomic_cmpxchg are the same as those defined for 'cas' -below. - -Finally: - - int atomic_add_unless(atomic_t *v, int a, int u); - -If the atomic value v is not equal to u, this function adds a to v, and -returns non zero. If v is equal to u then it returns zero. This is done as -an atomic operation. - -atomic_add_unless must provide explicit memory barriers around the -operation unless it fails (returns 0). - -atomic_inc_not_zero, equivalent to atomic_add_unless(v, 1, 0) - - -If a caller requires memory barrier semantics around an atomic_t -operation which does not return a value, a set of interfaces are -defined which accomplish this: - - void smp_mb__before_atomic(void); - void smp_mb__after_atomic(void); - -For example, smp_mb__before_atomic() can be used like so: - - obj->dead = 1; - smp_mb__before_atomic(); - atomic_dec(&obj->ref_count); - -It makes sure that all memory operations preceding the atomic_dec() -call are strongly ordered with respect to the atomic counter -operation. In the above example, it guarantees that the assignment of -"1" to obj->dead will be globally visible to other cpus before the -atomic counter decrement. - -Without the explicit smp_mb__before_atomic() call, the -implementation could legally allow the atomic counter update visible -to other cpus before the "obj->dead = 1;" assignment. - -A missing memory barrier in the cases where they are required by the -atomic_t implementation above can have disastrous results. Here is -an example, which follows a pattern occurring frequently in the Linux -kernel. It is the use of atomic counters to implement reference -counting, and it works such that once the counter falls to zero it can -be guaranteed that no other entity can be accessing the object: - -static void obj_list_add(struct obj *obj, struct list_head *head) -{ - obj->active = 1; - list_add(&obj->list, head); -} - -static void obj_list_del(struct obj *obj) -{ - list_del(&obj->list); - obj->active = 0; -} - -static void obj_destroy(struct obj *obj) -{ - BUG_ON(obj->active); - kfree(obj); -} - -struct obj *obj_list_peek(struct list_head *head) -{ - if (!list_empty(head)) { - struct obj *obj; - - obj = list_entry(head->next, struct obj, list); - atomic_inc(&obj->refcnt); - return obj; - } - return NULL; -} - -void obj_poke(void) -{ - struct obj *obj; - - spin_lock(&global_list_lock); - obj = obj_list_peek(&global_list); - spin_unlock(&global_list_lock); - - if (obj) { - obj->ops->poke(obj); - if (atomic_dec_and_test(&obj->refcnt)) - obj_destroy(obj); - } -} - -void obj_timeout(struct obj *obj) -{ - spin_lock(&global_list_lock); - obj_list_del(obj); - spin_unlock(&global_list_lock); - - if (atomic_dec_and_test(&obj->refcnt)) - obj_destroy(obj); -} - -(This is a simplification of the ARP queue management in the - generic neighbour discover code of the networking. Olaf Kirch - found a bug wrt. memory barriers in kfree_skb() that exposed - the atomic_t memory barrier requirements quite clearly.) - -Given the above scheme, it must be the case that the obj->active -update done by the obj list deletion be visible to other processors -before the atomic counter decrement is performed. - -Otherwise, the counter could fall to zero, yet obj->active would still -be set, thus triggering the assertion in obj_destroy(). The error -sequence looks like this: - - cpu 0 cpu 1 - obj_poke() obj_timeout() - obj = obj_list_peek(); - ... gains ref to obj, refcnt=2 - obj_list_del(obj); - obj->active = 0 ... - ... visibility delayed ... - atomic_dec_and_test() - ... refcnt drops to 1 ... - atomic_dec_and_test() - ... refcount drops to 0 ... - obj_destroy() - BUG() triggers since obj->active - still seen as one - obj->active update visibility occurs - -With the memory barrier semantics required of the atomic_t operations -which return values, the above sequence of memory visibility can never -happen. Specifically, in the above case the atomic_dec_and_test() -counter decrement would not become globally visible until the -obj->active update does. - -As a historical note, 32-bit Sparc used to only allow usage of -24-bits of its atomic_t type. This was because it used 8 bits -as a spinlock for SMP safety. Sparc32 lacked a "compare and swap" -type instruction. However, 32-bit Sparc has since been moved over -to a "hash table of spinlocks" scheme, that allows the full 32-bit -counter to be realized. Essentially, an array of spinlocks are -indexed into based upon the address of the atomic_t being operated -on, and that lock protects the atomic operation. Parisc uses the -same scheme. - -Another note is that the atomic_t operations returning values are -extremely slow on an old 386. - -We will now cover the atomic bitmask operations. You will find that -their SMP and memory barrier semantics are similar in shape and scope -to the atomic_t ops above. - -Native atomic bit operations are defined to operate on objects aligned -to the size of an "unsigned long" C data type, and are least of that -size. The endianness of the bits within each "unsigned long" are the -native endianness of the cpu. - - void set_bit(unsigned long nr, volatile unsigned long *addr); - void clear_bit(unsigned long nr, volatile unsigned long *addr); - void change_bit(unsigned long nr, volatile unsigned long *addr); - -These routines set, clear, and change, respectively, the bit number -indicated by "nr" on the bit mask pointed to by "ADDR". - -They must execute atomically, yet there are no implicit memory barrier -semantics required of these interfaces. - - int test_and_set_bit(unsigned long nr, volatile unsigned long *addr); - int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); - int test_and_change_bit(unsigned long nr, volatile unsigned long *addr); - -Like the above, except that these routines return a boolean which -indicates whether the changed bit was set _BEFORE_ the atomic bit -operation. - -WARNING! It is incredibly important that the value be a boolean, -ie. "0" or "1". Do not try to be fancy and save a few instructions by -declaring the above to return "long" and just returning something like -"old_val & mask" because that will not work. - -For one thing, this return value gets truncated to int in many code -paths using these interfaces, so on 64-bit if the bit is set in the -upper 32-bits then testers will never see that. - -One great example of where this problem crops up are the thread_info -flag operations. Routines such as test_and_set_ti_thread_flag() chop -the return value into an int. There are other places where things -like this occur as well. - -These routines, like the atomic_t counter operations returning values, -must provide explicit memory barrier semantics around their execution. -All memory operations before the atomic bit operation call must be -made visible globally before the atomic bit operation is made visible. -Likewise, the atomic bit operation must be visible globally before any -subsequent memory operation is made visible. For example: - - obj->dead = 1; - if (test_and_set_bit(0, &obj->flags)) - /* ... */; - obj->killed = 1; - -The implementation of test_and_set_bit() must guarantee that -"obj->dead = 1;" is visible to cpus before the atomic memory operation -done by test_and_set_bit() becomes visible. Likewise, the atomic -memory operation done by test_and_set_bit() must become visible before -"obj->killed = 1;" is visible. - -Finally there is the basic operation: - - int test_bit(unsigned long nr, __const__ volatile unsigned long *addr); - -Which returns a boolean indicating if bit "nr" is set in the bitmask -pointed to by "addr". - -If explicit memory barriers are required around {set,clear}_bit() (which do -not return a value, and thus does not need to provide memory barrier -semantics), two interfaces are provided: - - void smp_mb__before_atomic(void); - void smp_mb__after_atomic(void); - -They are used as follows, and are akin to their atomic_t operation -brothers: - - /* All memory operations before this call will - * be globally visible before the clear_bit(). - */ - smp_mb__before_atomic(); - clear_bit( ... ); - - /* The clear_bit() will be visible before all - * subsequent memory operations. - */ - smp_mb__after_atomic(); - -There are two special bitops with lock barrier semantics (acquire/release, -same as spinlocks). These operate in the same way as their non-_lock/unlock -postfixed variants, except that they are to provide acquire/release semantics, -respectively. This means they can be used for bit_spin_trylock and -bit_spin_unlock type operations without specifying any more barriers. - - int test_and_set_bit_lock(unsigned long nr, unsigned long *addr); - void clear_bit_unlock(unsigned long nr, unsigned long *addr); - void __clear_bit_unlock(unsigned long nr, unsigned long *addr); - -The __clear_bit_unlock version is non-atomic, however it still implements -unlock barrier semantics. This can be useful if the lock itself is protecting -the other bits in the word. - -Finally, there are non-atomic versions of the bitmask operations -provided. They are used in contexts where some other higher-level SMP -locking scheme is being used to protect the bitmask, and thus less -expensive non-atomic operations may be used in the implementation. -They have names similar to the above bitmask operation interfaces, -except that two underscores are prefixed to the interface name. - - void __set_bit(unsigned long nr, volatile unsigned long *addr); - void __clear_bit(unsigned long nr, volatile unsigned long *addr); - void __change_bit(unsigned long nr, volatile unsigned long *addr); - int __test_and_set_bit(unsigned long nr, volatile unsigned long *addr); - int __test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); - int __test_and_change_bit(unsigned long nr, volatile unsigned long *addr); - -These non-atomic variants also do not require any special memory -barrier semantics. - -The routines xchg() and cmpxchg() must provide the same exact -memory-barrier semantics as the atomic and bit operations returning -values. - -Note: If someone wants to use xchg(), cmpxchg() and their variants, -linux/atomic.h should be included rather than asm/cmpxchg.h, unless -the code is in arch/* and can take care of itself. - -Spinlocks and rwlocks have memory barrier expectations as well. -The rule to follow is simple: - -1) When acquiring a lock, the implementation must make it globally - visible before any subsequent memory operation. - -2) When releasing a lock, the implementation must make it such that - all previous memory operations are globally visible before the - lock release. - -Which finally brings us to _atomic_dec_and_lock(). There is an -architecture-neutral version implemented in lib/dec_and_lock.c, -but most platforms will wish to optimize this in assembler. - - int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock); - -Atomically decrement the given counter, and if will drop to zero -atomically acquire the given spinlock and perform the decrement -of the counter to zero. If it does not drop to zero, do nothing -with the spinlock. - -It is actually pretty simple to get the memory barrier correct. -Simply satisfy the spinlock grab requirements, which is make -sure the spinlock operation is globally visible before any -subsequent memory operation. - -We can demonstrate this operation more clearly if we define -an abstract atomic operation: - - long cas(long *mem, long old, long new); - -"cas" stands for "compare and swap". It atomically: - -1) Compares "old" with the value currently at "mem". -2) If they are equal, "new" is written to "mem". -3) Regardless, the current value at "mem" is returned. - -As an example usage, here is what an atomic counter update -might look like: - -void example_atomic_inc(long *counter) -{ - long old, new, ret; - - while (1) { - old = *counter; - new = old + 1; - - ret = cas(counter, old, new); - if (ret == old) - break; - } -} - -Let's use cas() in order to build a pseudo-C atomic_dec_and_lock(): - -int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock) -{ - long old, new, ret; - int went_to_zero; - - went_to_zero = 0; - while (1) { - old = atomic_read(atomic); - new = old - 1; - if (new == 0) { - went_to_zero = 1; - spin_lock(lock); - } - ret = cas(atomic, old, new); - if (ret == old) - break; - if (went_to_zero) { - spin_unlock(lock); - went_to_zero = 0; - } - } - - return went_to_zero; -} - -Now, as far as memory barriers go, as long as spin_lock() -strictly orders all subsequent memory operations (including -the cas()) with respect to itself, things will be fine. - -Said another way, _atomic_dec_and_lock() must guarantee that -a counter dropping to zero is never made visible before the -spinlock being acquired. - -Note that this also means that for the case where the counter -is not dropping to zero, there are no memory ordering -requirements. diff --git a/Documentation/core-api/atomic_ops.rst b/Documentation/core-api/atomic_ops.rst new file mode 100644 index 0000000..3ac94ff --- /dev/null +++ b/Documentation/core-api/atomic_ops.rst @@ -0,0 +1,669 @@ +======================================================= +Semantics and Behavior of Atomic and Bitmask Operations +======================================================= + +:Author: David S. Miller + +This document is intended to serve as a guide to Linux port +maintainers on how to implement atomic counter, bitops, and spinlock +interfaces properly. + +Atomic Type And Operations +========================== + +The ``atomic_t type`` should be defined as a signed integer and +the ``atomic_long_t`` type as a signed long integer. Also, they should +be made opaque such that any kind of cast to a normal C integer type +will fail. Something like the following should suffice:: + + typedef struct { int counter; } atomic_t; + typedef struct { long counter; } atomic_long_t; + +Historically, counter has been declared volatile. This is now discouraged. +See :ref:`Documentation/process/volatile-considered-harmful.rst +<volatile_considered_harmful>` for the complete rationale. + +``local_t`` is very similar to ``atomic_t``. If the counter is per CPU and only +updated by one CPU, ``local_t`` is probably more appropriate. Please see +:ref:`Documentation/local_ops.txt <local_ops>` for the semantics of ``local_t``. + +The first operations to implement for ``atomic_t``'s are the initializers and +plain reads. :: + + #define ATOMIC_INIT(i) { (i) } + #define atomic_set(v, i) ((v)->counter = (i)) + +The first macro is used in definitions, such as:: + + static atomic_t my_counter = ATOMIC_INIT(1); + +The initializer is atomic in that the return values of the atomic operations +are guaranteed to be correct reflecting the initialized value if the +initializer is used before runtime. If the initializer is used at runtime, a +proper implicit or explicit read memory barrier is needed before reading the +value with ``atomic_read`` from another thread. + +As with all of the ``atomic_`` interfaces, replace the leading ``atomic_`` +with ``atomic_long_`` to operate on ``atomic_long_t``. + +The second interface can be used at runtime, as in:: + + struct foo { atomic_t counter; }; + ... + + struct foo *k; + + k = kmalloc(sizeof(*k), GFP_KERNEL); + if (!k) + return -ENOMEM; + atomic_set(&k->counter, 0); + +The setting is atomic in that the return values of the atomic operations by +all threads are guaranteed to be correct reflecting either the value that has +been set with this operation or set with another operation. A proper implicit +or explicit memory barrier is needed before the value set with the operation +is guaranteed to be readable with ``atomic_read`` from another thread. + +Next, we have:: + + #define atomic_read(v) ((v)->counter) + +which simply reads the counter value currently visible to the calling thread. +The read is atomic in that the return value is guaranteed to be one of the +values initialized or modified with the interface operations if a proper +implicit or explicit memory barrier is used after possible runtime +initialization by any other thread and the value is modified only with the +interface operations. ``atomic_read`` does not guarantee that the runtime +initialization by any other thread is visible yet, so the user of the +interface must take care of that with a proper implicit or explicit memory +barrier. + +.. warning:: + + ``atomic_read()`` and ``atomic_set()`` DO NOT IMPLY BARRIERS! + + Some architectures may choose to use the ``volatile`` keyword, barriers, or + ``inline`` assembly to guarantee some degree of immediacy for ``atomic_read()`` + and ``atomic_set()``. This is not uniformly guaranteed, and may change in the + future, so all users of ``atomic_t`` should treat ``atomic_read()`` and + ``atomic_set()`` as simple C statements that may be reordered or optimized away + entirely by the compiler or processor, and explicitly invoke the appropriate + compiler and/or memory barrier for each use case. Failure to do so will result + in code that may suddenly break when used with different architectures or + compiler optimizations, or even changes in unrelated code which changes how the + compiler optimizes the section accessing ``atomic_t`` variables. + +Properly aligned pointers, longs, ints, and chars (and unsigned +equivalents) may be atomically loaded from and stored to in the same +sense as described for ``atomic_read()`` and ``atomic_set()``. The +``READ_ONCE()`` and ``WRITE_ONCE()`` macros should be used to prevent the +compiler from using optimizations that might otherwise optimize accesses out of +existence on the one hand, or that might create unsolicited accesses on the +other. + +For example consider the following code:: + + while (a > 0) + do_something(); + +If the compiler can prove that ``do_something()`` does not store to the +variable a, then the compiler is within its rights transforming this to +the following:: + + tmp = a; + if (a > 0) + for (;;) + do_something(); + +If you don't want the compiler to do this (and you probably don't), then +you should use something like the following:: + + while (READ_ONCE(a) < 0) + do_something(); + +Alternatively, you could place a ``barrier()`` call in the loop. + +For another example, consider the following code:: + + tmp_a = a; + do_something_with(tmp_a); + do_something_else_with(tmp_a); + +If the compiler can prove that ``do_something_with()`` does not store to the +variable ``a``, then the compiler is within its rights to manufacture an +additional load as follows:: + + tmp_a = a; + do_something_with(tmp_a); + tmp_a = a; + do_something_else_with(tmp_a); + +This could fatally confuse your code if it expected the same value +to be passed to ``do_something_with()`` and ``do_something_else_with()``. + +The compiler would be likely to manufacture this additional load if +``do_something_with()`` was an inline function that made very heavy use +of registers: reloading from variable a could save a flush to the +stack and later reload. To prevent the compiler from attacking your +code in this manner, write the following:: + + tmp_a = READ_ONCE(a); + do_something_with(tmp_a); + do_something_else_with(tmp_a); + +For a final example, consider the following code, assuming that the +variable ``a`` is set at boot time before the second CPU is brought online +and never changed later, so that memory barriers are not needed:: + + if (a) + b = 9; + else + b = 42; + +The compiler is within its rights to manufacture an additional store +by transforming the above code into the following:: + + b = 42; + if (a) + b = 9; + +This could come as a fatal surprise to other code running concurrently +that expected ``b`` to never have the value ``42`` if ``a`` was zero. To +prevent the compiler from doing this, write something like:: + + if (a) + WRITE_ONCE(b, 9); + else + WRITE_ONCE(b, 42); + +Don't even **think** about doing this without proper use of memory barriers, +locks, or atomic operations if variable ``a`` can change at runtime! + +.. warning:: + + ``READ_ONCE()`` OR ``WRITE_ONCE()`` DO NOT IMPLY A BARRIER! + +Now, we move onto the atomic operation interfaces typically implemented with +the help of assembly code. :: + + void atomic_add(int i, atomic_t *v); + void atomic_sub(int i, atomic_t *v); + void atomic_inc(atomic_t *v); + void atomic_dec(atomic_t *v); + +These four routines add and subtract integral values to/from the given +``atomic_t`` value. The first two routines pass explicit integers by +which to make the adjustment, whereas the latter two use an implicit +adjustment value of "1". + +One very important aspect of these two routines is that they **DO NOT** +require any explicit memory barriers. They need only perform the +``atomic_t`` counter update in an SMP safe manner. + +Next, we have:: + + int atomic_inc_return(atomic_t *v); + int atomic_dec_return(atomic_t *v); + +These routines add 1 and subtract 1, respectively, from the given +``atomic_t`` and return the new counter value after the operation is +performed. + +Unlike the above routines, it is required that these primitives +include explicit memory barriers that are performed before and after +the operation. It must be done such that all memory operations before +and after the atomic operation calls are strongly ordered with respect +to the atomic operation itself. + +For example, it should behave as if a ``smp_mb()`` call existed both +before and after the atomic operation. + +If the atomic instructions used in an implementation provide explicit +memory barrier semantics which satisfy the above requirements, that is +fine as well. + +Let's move on:: + + int atomic_add_return(int i, atomic_t *v); + int atomic_sub_return(int i, atomic_t *v); + +These behave just like ``atomic_{inc,dec}_return()`` except that an +explicit counter adjustment is given instead of the implicit "1". +This means that like ``atomic_{inc,dec}_return()``, the memory barrier +semantics are required. + +Next:: + + int atomic_inc_and_test(atomic_t *v); + int atomic_dec_and_test(atomic_t *v); + +These two routines increment and decrement by 1, respectively, the +given atomic counter. They return a boolean indicating whether the +resulting counter value was zero or not. + +Again, these primitives provide explicit memory barrier semantics around +the atomic operation. :: + + int atomic_sub_and_test(int i, atomic_t *v); + +This is identical to ``atomic_dec_and_test()`` except that an explicit +decrement is given instead of the implicit "1". This primitive must +provide explicit memory barrier semantics around the operation. :: + + int atomic_add_negative(int i, atomic_t *v); + +The given increment is added to the given atomic counter value. A boolean +is return which indicates whether the resulting counter value is negative. +This primitive must provide explicit memory barrier semantics around +the operation. + +Then:: + + int atomic_xchg(atomic_t *v, int new); + +This performs an atomic exchange operation on the atomic variable v, setting +the given new value. It returns the old value that the atomic variable v had +just before the operation. + +``atomic_xchg`` must provide explicit memory barriers around the operation. :: + + int atomic_cmpxchg(atomic_t *v, int old, int new); + +This performs an atomic compare exchange operation on the atomic value v, +with the given old and new values. Like all ``atomic_*`` operations, +``atomic_cmpxchg`` will only satisfy its atomicity semantics as long as all +other accesses of ``*v`` are performed through ``atomic_*`` operations. + +``atomic_cmpxchg`` must provide explicit memory barriers around the operation, +although if the comparison fails then no memory ordering guarantees are +required. + +The semantics for ``atomic_cmpxchg`` are the same as those defined for 'cas' +below. + +Finally:: + + int atomic_add_unless(atomic_t *v, int a, int u); + +If the atomic value v is not equal to u, this function adds a to v, and +returns non zero. If v is equal to u then it returns zero. This is done as +an atomic operation. + +``atomic_add_unless`` must provide explicit memory barriers around the +operation unless it fails (returns 0). + +``atomic_inc_not_zero``, equivalent to ``atomic_add_unless(v, 1, 0)`` + + +If a caller requires memory barrier semantics around an ``atomic_t`` +operation which does not return a value, a set of interfaces are +defined which accomplish this:: + + void smp_mb__before_atomic(void); + void smp_mb__after_atomic(void); + +For example, ``smp_mb__before_atomic()`` can be used like so:: + + obj->dead = 1; + smp_mb__before_atomic(); + atomic_dec(&obj->ref_count); + +It makes sure that all memory operations preceding the ``atomic_dec()`` +call are strongly ordered with respect to the atomic counter +operation. In the above example, it guarantees that the assignment of +"1" to ``obj->dead`` will be globally visible to other cpus before the +atomic counter decrement. + +Without the explicit ``smp_mb__before_atomic()`` call, the +implementation could legally allow the atomic counter update visible +to other cpus before the ``obj->dead = 1;`` assignment. + +A missing memory barrier in the cases where they are required by the +``atomic_t`` implementation above can have disastrous results. Here is +an example, which follows a pattern occurring frequently in the Linux +kernel. It is the use of atomic counters to implement reference +counting, and it works such that once the counter falls to zero it can +be guaranteed that no other entity can be accessing the object: + +.. code-block:: c + + static void obj_list_add(struct obj *obj, struct list_head *head) + { + obj->active = 1; + list_add(&obj->list, head); + } + + static void obj_list_del(struct obj *obj) + { + list_del(&obj->list); + obj->active = 0; + } + + static void obj_destroy(struct obj *obj) + { + BUG_ON(obj->active); + kfree(obj); + } + + struct obj *obj_list_peek(struct list_head *head) + { + if (!list_empty(head)) { + struct obj *obj; + + obj = list_entry(head->next, struct obj, list); + atomic_inc(&obj->refcnt); + return obj; + } + return NULL; + } + + void obj_poke(void) + { + struct obj *obj; + + spin_lock(&global_list_lock); + obj = obj_list_peek(&global_list); + spin_unlock(&global_list_lock); + + if (obj) { + obj->ops->poke(obj); + if (atomic_dec_and_test(&obj->refcnt)) + obj_destroy(obj); + } + } + + void obj_timeout(struct obj *obj) + { + spin_lock(&global_list_lock); + obj_list_del(obj); + spin_unlock(&global_list_lock); + + if (atomic_dec_and_test(&obj->refcnt)) + obj_destroy(obj); + } + +.. note: + + This is a simplification of the ARP queue management in the + generic neighbour discover code of the networking. Olaf Kirch + found a bug wrt. memory barriers in kfree_skb() that exposed + the atomic_t memory barrier requirements quite clearly. + +Given the above scheme, it must be the case that the ``obj->active`` +update done by the obj list deletion be visible to other processors +before the atomic counter decrement is performed. + +Otherwise, the counter could fall to zero, yet ``obj->active`` would still +be set, thus triggering the assertion in ``obj_destroy()``. The error +sequence looks like this:: + + cpu 0 cpu 1 + obj_poke() obj_timeout() + obj = obj_list_peek(); + ... gains ref to obj, refcnt=2 + obj_list_del(obj); + obj->active = 0 ... + ... visibility delayed ... + atomic_dec_and_test() + ... refcnt drops to 1 ... + atomic_dec_and_test() + ... refcount drops to 0 ... + obj_destroy() + BUG() triggers since obj->active + still seen as one + obj->active update visibility occurs + +With the memory barrier semantics required of the atomic_t operations +which return values, the above sequence of memory visibility can never +happen. Specifically, in the above case the ``atomic_dec_and_test()`` +counter decrement would not become globally visible until the +``obj->active`` update does. + +As a historical note, 32-bit Sparc used to only allow usage of +24-bits of its ``atomic_t`` type. This was because it used 8 bits +as a spinlock for SMP safety. Sparc32 lacked a "compare and swap" +type instruction. However, 32-bit Sparc has since been moved over +to a "hash table of spinlocks" scheme, that allows the full 32-bit +counter to be realized. Essentially, an array of spinlocks are +indexed into based upon the address of the ``atomic_t`` being operated +on, and that lock protects the atomic operation. Parisc uses the +same scheme. + +Another note is that the ``atomic_t`` operations returning values are +extremely slow on an old 386. + +Atomic Bitmask +============== + +We will now cover the atomic bitmask operations. You will find that +their SMP and memory barrier semantics are similar in shape and scope +to the ``atomic_t`` ops above. + +Native atomic bit operations are defined to operate on objects aligned +to the size of an ``unsigned long`` C data type, and are least of that +size. The endianness of the bits within each ``unsigned long`` are the +native endianness of the cpu. :: + + void set_bit(unsigned long nr, volatile unsigned long *addr); + void clear_bit(unsigned long nr, volatile unsigned long *addr); + void change_bit(unsigned long nr, volatile unsigned long *addr); + +These routines set, clear, and change, respectively, the bit number +indicated by ``nr`` on the bit mask pointed to by ``addr``. + +They must execute atomically, yet there are no implicit memory barrier +semantics required of these interfaces. :: + + int test_and_set_bit(unsigned long nr, volatile unsigned long *addr); + int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); + int test_and_change_bit(unsigned long nr, volatile unsigned long *addr); + +Like the above, except that these routines return a boolean which +indicates whether the changed bit was set _BEFORE_ the atomic bit +operation. + +.. warning:: + + It is incredibly important that the value be a boolean, + ie. "0" or "1". Do not try to be fancy and save a few instructions by + declaring the above to return "long" and just returning something like + "old_val & mask" because that will not work. + + For one thing, this return value gets truncated to int in many code + paths using these interfaces, so on 64-bit if the bit is set in the + upper 32-bits then testers will never see that. + + One great example of where this problem crops up are the ``thread_info`` + flag operations. Routines such as ``test_and_set_ti_thread_flag()`` chop + the return value into an int. There are other places where things + like this occur as well. + +These routines, like the ``atomic_t`` counter operations returning values, +must provide explicit memory barrier semantics around their execution. +All memory operations before the atomic bit operation call must be +made visible globally before the atomic bit operation is made visible. +Likewise, the atomic bit operation must be visible globally before any +subsequent memory operation is made visible. For example: + +.. code-block:: c + + obj->dead = 1; + if (test_and_set_bit(0, &obj->flags)) + /* ... */; + obj->killed = 1; + +The implementation of ``test_and_set_bit()`` must guarantee that +``obj->dead = 1;`` is visible to cpus before the atomic memory operation +done by ``test_and_set_bit()`` becomes visible. Likewise, the atomic +memory operation done by ``test_and_set_bit()`` must become visible before +``obj->killed = 1;`` is visible. + +Finally there is the basic operation:: + + int test_bit(unsigned long nr, __const__ volatile unsigned long *addr); + +Which returns a boolean indicating if bit ``nr`` is set in the bitmask +pointed to by ``addr``. + +If explicit memory barriers are required around ``{set,clear}_bit()`` (which do +not return a value, and thus does not need to provide memory barrier +semantics), two interfaces are provided:: + + void smp_mb__before_atomic(void); + void smp_mb__after_atomic(void); + +They are used as follows, and are akin to their ``atomic_t`` operation +brothers: + +.. code-block:: c + + /* All memory operations before this call will + * be globally visible before the clear_bit(). + */ + smp_mb__before_atomic(); + clear_bit( ... ); + + /* The clear_bit() will be visible before all + * subsequent memory operations. + */ + smp_mb__after_atomic(); + +There are two special bitops with lock barrier semantics (acquire/release, +same as spinlocks). These operate in the same way as their non-_lock/unlock +postfixed variants, except that they are to provide acquire/release semantics, +respectively. This means they can be used for ``bit_spin_trylock`` and +``bit_spin_unlock`` type operations without specifying any more barriers. :: + + int test_and_set_bit_lock(unsigned long nr, unsigned long *addr); + void clear_bit_unlock(unsigned long nr, unsigned long *addr); + void __clear_bit_unlock(unsigned long nr, unsigned long *addr); + +The ``__clear_bit_unlock`` version is non-atomic, however it still implements +unlock barrier semantics. This can be useful if the lock itself is protecting +the other bits in the word. + +Finally, there are non-atomic versions of the bitmask operations +provided. They are used in contexts where some other higher-level SMP +locking scheme is being used to protect the bitmask, and thus less +expensive non-atomic operations may be used in the implementation. +They have names similar to the above bitmask operation interfaces, +except that two underscores are prefixed to the interface name. :: + + void __set_bit(unsigned long nr, volatile unsigned long *addr); + void __clear_bit(unsigned long nr, volatile unsigned long *addr); + void __change_bit(unsigned long nr, volatile unsigned long *addr); + int __test_and_set_bit(unsigned long nr, volatile unsigned long *addr); + int __test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); + int __test_and_change_bit(unsigned long nr, volatile unsigned long *addr); + +These non-atomic variants also do not require any special memory +barrier semantics. + +The routines ``xchg()`` and ``cmpxchg()`` must provide the same exact +memory-barrier semantics as the atomic and bit operations returning +values. + +.. note:: + + If someone wants to use ``xchg()``, ``cmpxchg()`` and their variants, + ``linux/atomic.h`` should be included rather than ``asm/cmpxchg.h``, unless + the code is in arch/* and can take care of itself. + +Spinlocks and rwlocks have memory barrier expectations as well. +The rule to follow is simple: + +1. When acquiring a lock, the implementation must make it globally + visible before any subsequent memory operation. + +2. When releasing a lock, the implementation must make it such that + all previous memory operations are globally visible before the + lock release. + +Which finally brings us to ``_atomic_dec_and_lock()``. There is an +architecture-neutral version implemented in ``lib/dec_and_lock.c``, +but most platforms will wish to optimize this in assembler. :: + + int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock); + +Atomically decrement the given counter, and if will drop to zero +atomically acquire the given spinlock and perform the decrement +of the counter to zero. If it does not drop to zero, do nothing +with the spinlock. + +It is actually pretty simple to get the memory barrier correct. +Simply satisfy the spinlock grab requirements, which is make +sure the spinlock operation is globally visible before any +subsequent memory operation. + +We can demonstrate this operation more clearly if we define +an abstract atomic operation:: + + long cas(long *mem, long old, long new); + +"``cas``" stands for "compare and swap". It atomically: + +1. Compares "old" with the value currently at "mem". +2. If they are equal, "new" is written to "mem". +3. Regardless, the current value at "mem" is returned. + +As an example usage, here is what an atomic counter update +might look like: + +.. code-block:: c + + void example_atomic_inc(long *counter) + { + long old, new, ret; + + while (1) { + old = *counter; + new = old + 1; + + ret = cas(counter, old, new); + if (ret == old) + break; + } + } + +Let's use cas() in order to build a pseudo-C atomic_dec_and_lock(): + +.. code-block:: c + + int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock) + { + long old, new, ret; + int went_to_zero; + + went_to_zero = 0; + while (1) { + old = atomic_read(atomic); + new = old - 1; + if (new == 0) { + went_to_zero = 1; + spin_lock(lock); + } + ret = cas(atomic, old, new); + if (ret == old) + break; + if (went_to_zero) { + spin_unlock(lock); + went_to_zero = 0; + } + } + + return went_to_zero; + } + +Now, as far as memory barriers go, as long as i``spin_lock()`` +strictly orders all subsequent memory operations (including +the ``cas()``) with respect to itself, things will be fine. + +Said another way, ``_atomic_dec_and_lock()`` must guarantee that +a counter dropping to zero is never made visible before the +spinlock being acquired. + +.. note:: + + Note that this also means that for the case where the counter + is not dropping to zero, there are no memory ordering + requirements. diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index 480d9a3..f3e5f5e 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -8,6 +8,7 @@ Kernel and driver related documentation. :maxdepth: 1 assoc_array + atomic_ops workqueue .. only:: subproject diff --git a/Documentation/process/volatile-considered-harmful.rst b/Documentation/process/volatile-considered-harmful.rst index e0d042a..4934e65 100644 --- a/Documentation/process/volatile-considered-harmful.rst +++ b/Documentation/process/volatile-considered-harmful.rst @@ -1,3 +1,6 @@ + +.. _volatile_considered_harmful: + Why the "volatile" type class should not be used ------------------------------------------------ -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* [PATCH 3/4] Documentation/local_ops.txt: convert to ReST markup 2016-11-24 12:42 [PATCH 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 2016-11-24 12:42 ` [PATCH 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke 2016-11-24 12:42 ` [PATCH 2/4] Documentation/atomic_ops.txt: " Silvio Fricke @ 2016-11-24 12:42 ` Silvio Fricke 2016-11-24 12:42 ` [PATCH 4/4] firmware: remove warning at documentation generation time Silvio Fricke ` (2 subsequent siblings) 5 siblings, 0 replies; 51+ messages in thread From: Silvio Fricke @ 2016-11-24 12:42 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Silvio Fricke ... and move to core-api folder. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- Documentation/core-api/index.rst | 1 +- Documentation/core-api/local_ops.rst | 208 ++++++++++++++++++++++++++++- Documentation/local_ops.txt | 191 +-------------------------- 3 files changed, 209 insertions(+), 191 deletions(-) create mode 100644 Documentation/core-api/local_ops.rst delete mode 100644 Documentation/local_ops.txt diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index f3e5f5e..25b4e4a 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -9,6 +9,7 @@ Kernel and driver related documentation. assoc_array atomic_ops + local_ops workqueue .. only:: subproject diff --git a/Documentation/core-api/local_ops.rst b/Documentation/core-api/local_ops.rst new file mode 100644 index 0000000..01f1880 --- /dev/null +++ b/Documentation/core-api/local_ops.rst @@ -0,0 +1,208 @@ + +.. _local_ops: + +================================================= +Semantics and Behavior of Local Atomic Operations +================================================= + +:Author: Mathieu Desnoyers + + +This document explains the purpose of the local atomic operations, how +to implement them for any given architecture and shows how they can be used +properly. It also stresses on the precautions that must be taken when reading +those local variables across CPUs when the order of memory writes matters. + +.. note:: + + Note that ``local_t`` based operations are not recommended for general + kernel use. Please use the ``this_cpu`` operations instead unless there is + really a special purpose. Most uses of ``local_t`` in the kernel have been + replaced by ``this_cpu`` operations. ``this_cpu`` operations combine the + relocation with the ``local_t`` like semantics in a single instruction and + yield more compact and faster executing code. + + +Purpose of local atomic operations +================================== + +Local atomic operations are meant to provide fast and highly reentrant per CPU +counters. They minimize the performance cost of standard atomic operations by +removing the LOCK prefix and memory barriers normally required to synchronize +across CPUs. + +Having fast per CPU atomic counters is interesting in many cases: it does not +require disabling interrupts to protect from interrupt handlers and it permits +coherent counters in NMI handlers. It is especially useful for tracing purposes +and for various performance monitoring counters. + +Local atomic operations only guarantee variable modification atomicity wrt the +CPU which owns the data. Therefore, care must taken to make sure that only one +CPU writes to the ``local_t`` data. This is done by using per cpu data and +making sure that we modify it from within a preemption safe context. It is +however permitted to read ``local_t`` data from any CPU: it will then appear to +be written out of order wrt other memory writes by the owner CPU. + + +Implementation for a given architecture +======================================= + +It can be done by slightly modifying the standard atomic operations: only +their UP variant must be kept. It typically means removing LOCK prefix (on +i386 and x86_64) and any SMP synchronization barrier. If the architecture does +not have a different behavior between SMP and UP, including +``asm-generic/local.h`` in your architecture's ``local.h`` is sufficient. + +The ``local_t`` type is defined as an opaque ``signed long`` by embedding an +``atomic_long_t`` inside a structure. This is made so a cast from this type to +a ``long`` fails. The definition looks like:: + + typedef struct { atomic_long_t a; } local_t; + + +Rules to follow when using local atomic operations +================================================== + +* Variables touched by local ops must be per cpu variables. +* *Only* the CPU owner of these variables must write to them. +* This CPU can use local ops from any context (process, irq, softirq, nmi, ...) + to update its ``local_t`` variables. +* Preemption (or interrupts) must be disabled when using local ops in + process context to make sure the process won't be migrated to a + different CPU between getting the per-cpu variable and doing the + actual local op. +* When using local ops in interrupt context, no special care must be + taken on a mainline kernel, since they will run on the local CPU with + preemption already disabled. I suggest, however, to explicitly + disable preemption anyway to make sure it will still work correctly on + -rt kernels. +* Reading the local cpu variable will provide the current copy of the + variable. +* Reads of these variables can be done from any CPU, because updates to + "``long``", aligned, variables are always atomic. Since no memory + synchronization is done by the writer CPU, an outdated copy of the + variable can be read when reading some *other* cpu's variables. + + +How to use local atomic operations +================================== + +.. code-block:: c + + #include <linux/percpu.h> + #include <asm/local.h> + + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); + + +Counting +======== + +Counting is done on all the bits of a signed long. + +In preemptible context, use ``get_cpu_var()`` and ``put_cpu_var()`` around +local atomic operations: it makes sure that preemption is disabled around write +access to the per cpu variable. For instance:: + + local_inc(&get_cpu_var(counters)); + put_cpu_var(counters); + +If you are already in a preemption-safe context, you can use +``this_cpu_ptr()`` instead. :: + + local_inc(this_cpu_ptr(&counters)); + + + +Reading the counters +==================== + +Those local counters can be read from foreign CPUs to sum the count. Note that +the data seen by local_read across CPUs must be considered to be out of order +relatively to other memory writes happening on the CPU that owns the data. :: + + long sum = 0; + for_each_online_cpu(cpu) + sum += local_read(&per_cpu(counters, cpu)); + +If you want to use a remote local_read to synchronize access to a resource +between CPUs, explicit ``smp_wmb()`` and ``smp_rmb()`` memory barriers must be used +respectively on the writer and the reader CPUs. It would be the case if you use +the ``local_t`` variable as a counter of bytes written in a buffer: there should +be a ``smp_wmb()`` between the buffer write and the counter increment and also a +``smp_rmb()`` between the counter read and the buffer read. + + +Here is a sample module which implements a basic per cpu counter using +``local.h``. + +.. code-block:: c + + /* test-local.c + * + * Sample module for local.h usage. + */ + + + #include <asm/local.h> + #include <linux/module.h> + #include <linux/timer.h> + + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); + + static struct timer_list test_timer; + + /* IPI called on each CPU. */ + static void test_each(void *info) + { + /* Increment the counter from a non preemptible context */ + printk("Increment on cpu %d\n", smp_processor_id()); + local_inc(this_cpu_ptr(&counters)); + + /* This is what incrementing the variable would look like within a + * preemptible context (it disables preemption) : + * + * local_inc(&get_cpu_var(counters)); + * put_cpu_var(counters); + */ + } + + static void do_test_timer(unsigned long data) + { + int cpu; + + /* Increment the counters */ + on_each_cpu(test_each, NULL, 1); + /* Read all the counters */ + printk("Counters read from CPU %d\n", smp_processor_id()); + for_each_online_cpu(cpu) { + printk("Read : CPU %d, count %ld\n", cpu, + local_read(&per_cpu(counters, cpu))); + } + del_timer(&test_timer); + test_timer.expires = jiffies + 1000; + add_timer(&test_timer); + } + + static int __init test_init(void) + { + /* initialize the timer that will increment the counter */ + init_timer(&test_timer); + test_timer.function = do_test_timer; + test_timer.expires = jiffies + 1; + add_timer(&test_timer); + + return 0; + } + + static void __exit test_exit(void) + { + del_timer_sync(&test_timer); + } + + module_init(test_init); + module_exit(test_exit); + + MODULE_LICENSE("GPL"); + MODULE_AUTHOR("Mathieu Desnoyers"); + MODULE_DESCRIPTION("Local Atomic Ops"); diff --git a/Documentation/local_ops.txt b/Documentation/local_ops.txt deleted file mode 100644 index 407576a..0000000 --- a/Documentation/local_ops.txt +++ /dev/null @@ -1,191 +0,0 @@ - Semantics and Behavior of Local Atomic Operations - - Mathieu Desnoyers - - - This document explains the purpose of the local atomic operations, how -to implement them for any given architecture and shows how they can be used -properly. It also stresses on the precautions that must be taken when reading -those local variables across CPUs when the order of memory writes matters. - -Note that local_t based operations are not recommended for general kernel use. -Please use the this_cpu operations instead unless there is really a special purpose. -Most uses of local_t in the kernel have been replaced by this_cpu operations. -this_cpu operations combine the relocation with the local_t like semantics in -a single instruction and yield more compact and faster executing code. - - -* Purpose of local atomic operations - -Local atomic operations are meant to provide fast and highly reentrant per CPU -counters. They minimize the performance cost of standard atomic operations by -removing the LOCK prefix and memory barriers normally required to synchronize -across CPUs. - -Having fast per CPU atomic counters is interesting in many cases : it does not -require disabling interrupts to protect from interrupt handlers and it permits -coherent counters in NMI handlers. It is especially useful for tracing purposes -and for various performance monitoring counters. - -Local atomic operations only guarantee variable modification atomicity wrt the -CPU which owns the data. Therefore, care must taken to make sure that only one -CPU writes to the local_t data. This is done by using per cpu data and making -sure that we modify it from within a preemption safe context. It is however -permitted to read local_t data from any CPU : it will then appear to be written -out of order wrt other memory writes by the owner CPU. - - -* Implementation for a given architecture - -It can be done by slightly modifying the standard atomic operations : only -their UP variant must be kept. It typically means removing LOCK prefix (on -i386 and x86_64) and any SMP synchronization barrier. If the architecture does -not have a different behavior between SMP and UP, including asm-generic/local.h -in your architecture's local.h is sufficient. - -The local_t type is defined as an opaque signed long by embedding an -atomic_long_t inside a structure. This is made so a cast from this type to a -long fails. The definition looks like : - -typedef struct { atomic_long_t a; } local_t; - - -* Rules to follow when using local atomic operations - -- Variables touched by local ops must be per cpu variables. -- _Only_ the CPU owner of these variables must write to them. -- This CPU can use local ops from any context (process, irq, softirq, nmi, ...) - to update its local_t variables. -- Preemption (or interrupts) must be disabled when using local ops in - process context to make sure the process won't be migrated to a - different CPU between getting the per-cpu variable and doing the - actual local op. -- When using local ops in interrupt context, no special care must be - taken on a mainline kernel, since they will run on the local CPU with - preemption already disabled. I suggest, however, to explicitly - disable preemption anyway to make sure it will still work correctly on - -rt kernels. -- Reading the local cpu variable will provide the current copy of the - variable. -- Reads of these variables can be done from any CPU, because updates to - "long", aligned, variables are always atomic. Since no memory - synchronization is done by the writer CPU, an outdated copy of the - variable can be read when reading some _other_ cpu's variables. - - -* How to use local atomic operations - -#include <linux/percpu.h> -#include <asm/local.h> - -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); - - -* Counting - -Counting is done on all the bits of a signed long. - -In preemptible context, use get_cpu_var() and put_cpu_var() around local atomic -operations : it makes sure that preemption is disabled around write access to -the per cpu variable. For instance : - - local_inc(&get_cpu_var(counters)); - put_cpu_var(counters); - -If you are already in a preemption-safe context, you can use -this_cpu_ptr() instead. - - local_inc(this_cpu_ptr(&counters)); - - - -* Reading the counters - -Those local counters can be read from foreign CPUs to sum the count. Note that -the data seen by local_read across CPUs must be considered to be out of order -relatively to other memory writes happening on the CPU that owns the data. - - long sum = 0; - for_each_online_cpu(cpu) - sum += local_read(&per_cpu(counters, cpu)); - -If you want to use a remote local_read to synchronize access to a resource -between CPUs, explicit smp_wmb() and smp_rmb() memory barriers must be used -respectively on the writer and the reader CPUs. It would be the case if you use -the local_t variable as a counter of bytes written in a buffer : there should -be a smp_wmb() between the buffer write and the counter increment and also a -smp_rmb() between the counter read and the buffer read. - - -Here is a sample module which implements a basic per cpu counter using local.h. - ---- BEGIN --- -/* test-local.c - * - * Sample module for local.h usage. - */ - - -#include <asm/local.h> -#include <linux/module.h> -#include <linux/timer.h> - -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); - -static struct timer_list test_timer; - -/* IPI called on each CPU. */ -static void test_each(void *info) -{ - /* Increment the counter from a non preemptible context */ - printk("Increment on cpu %d\n", smp_processor_id()); - local_inc(this_cpu_ptr(&counters)); - - /* This is what incrementing the variable would look like within a - * preemptible context (it disables preemption) : - * - * local_inc(&get_cpu_var(counters)); - * put_cpu_var(counters); - */ -} - -static void do_test_timer(unsigned long data) -{ - int cpu; - - /* Increment the counters */ - on_each_cpu(test_each, NULL, 1); - /* Read all the counters */ - printk("Counters read from CPU %d\n", smp_processor_id()); - for_each_online_cpu(cpu) { - printk("Read : CPU %d, count %ld\n", cpu, - local_read(&per_cpu(counters, cpu))); - } - del_timer(&test_timer); - test_timer.expires = jiffies + 1000; - add_timer(&test_timer); -} - -static int __init test_init(void) -{ - /* initialize the timer that will increment the counter */ - init_timer(&test_timer); - test_timer.function = do_test_timer; - test_timer.expires = jiffies + 1; - add_timer(&test_timer); - - return 0; -} - -static void __exit test_exit(void) -{ - del_timer_sync(&test_timer); -} - -module_init(test_init); -module_exit(test_exit); - -MODULE_LICENSE("GPL"); -MODULE_AUTHOR("Mathieu Desnoyers"); -MODULE_DESCRIPTION("Local Atomic Ops"); ---- END --- -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* [PATCH 4/4] firmware: remove warning at documentation generation time 2016-11-24 12:42 [PATCH 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke ` (2 preceding siblings ...) 2016-11-24 12:42 ` [PATCH 3/4] Documentation/local_ops.txt: " Silvio Fricke @ 2016-11-24 12:42 ` Silvio Fricke 2016-11-24 19:46 ` [PATCH 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Mauro Carvalho Chehab 2016-11-25 10:02 ` [PATCH v2 " Silvio Fricke 5 siblings, 0 replies; 51+ messages in thread From: Silvio Fricke @ 2016-11-24 12:42 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Silvio Fricke This patch removes following error at for `make htmldocs`. No functional change. ./drivers/base/firmware_class.c:1348: WARNING: Bullet list ends without a blank line; unexpected unindent. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- drivers/base/firmware_class.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c index 22d1760..37b0221 100644 --- a/drivers/base/firmware_class.c +++ b/drivers/base/firmware_class.c @@ -1345,9 +1345,9 @@ static void request_firmware_work_func(struct work_struct *work) * * Asynchronous variant of request_firmware() for user contexts: * - sleep for as small periods as possible since it may - * increase kernel boot time of built-in device drivers - * requesting firmware in their ->probe() methods, if - * @gfp is GFP_KERNEL. + * increase kernel boot time of built-in device drivers + * requesting firmware in their ->probe() methods, if + * @gfp is GFP_KERNEL. * * - can't sleep at all if @gfp is GFP_ATOMIC. **/ -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* Re: [PATCH 0/4] core-api ReST: assoc_array, atomic_ops, local_ops 2016-11-24 12:42 [PATCH 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke ` (3 preceding siblings ...) 2016-11-24 12:42 ` [PATCH 4/4] firmware: remove warning at documentation generation time Silvio Fricke @ 2016-11-24 19:46 ` Mauro Carvalho Chehab 2016-11-25 10:02 ` [PATCH v2 " Silvio Fricke 5 siblings, 0 replies; 51+ messages in thread From: Mauro Carvalho Chehab @ 2016-11-24 19:46 UTC (permalink / raw) To: Silvio Fricke Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel Em Thu, 24 Nov 2016 13:42:16 +0100 Silvio Fricke <silvio.fricke@gmail.com> escreveu: > Hi, > > Some more ReSTification of core-api's: assoc_array, atomic_ops and local_ops. A > fourth patch removes a warning about a bullet list without ending at > firmware_class.c > > Thanks for review. Could you please resend, using git diff -M or adding this to the .git/config file: [diff] renames = true It is very hard to identify what you changed on your patches if you don't enable rename detection ;) Regards, Mauro > > BR > Silvio > > Silvio Fricke (4): > Documentation/assoc_array.txt: convert to ReST markup > Documentation/atomic_ops.txt: convert to ReST markup > Documentation/local_ops.txt: convert to ReST markup > firmware: remove warning at documentation generation time > > Documentation/assoc_array.txt | 574 +--------- > Documentation/atomic_ops.txt | 640 +----------- > Documentation/core-api/assoc_array.rst | 549 +++++++++- > Documentation/core-api/atomic_ops.rst | 669 +++++++++++- > Documentation/core-api/index.rst | 3 +- > Documentation/core-api/local_ops.rst | 208 +++- > Documentation/local_ops.txt | 191 +--- > Documentation/process/volatile-considered-harmful.rst | 3 +- > drivers/base/firmware_class.c | 6 +- > 9 files changed, 1435 insertions(+), 1408 deletions(-) > delete mode 100644 Documentation/assoc_array.txt > delete mode 100644 Documentation/atomic_ops.txt > create mode 100644 Documentation/core-api/assoc_array.rst > create mode 100644 Documentation/core-api/atomic_ops.rst > create mode 100644 Documentation/core-api/local_ops.rst > delete mode 100644 Documentation/local_ops.txt > > base-commit: 9c240d757658a3ae9968dd309e674c61f07c7f48 Thanks, Mauro ^ permalink raw reply [flat|nested] 51+ messages in thread
* [PATCH v2 0/4] core-api ReST: assoc_array, atomic_ops, local_ops 2016-11-24 12:42 [PATCH 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke ` (4 preceding siblings ...) 2016-11-24 19:46 ` [PATCH 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Mauro Carvalho Chehab @ 2016-11-25 10:02 ` Silvio Fricke 2016-11-25 10:02 ` [PATCH v2 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke ` (4 more replies) 5 siblings, 5 replies; 51+ messages in thread From: Silvio Fricke @ 2016-11-25 10:02 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Silvio Fricke Hi, Some more ReSTification of core-api's: assoc_array, atomic_ops and local_ops. A fourth patch removes a warning about a bullet list without ending at firmware_class.c v1 -> v2 * use format-patch with a rename_threshold of 10%, no other changes Thanks for review. BR Silvio Silvio Fricke (4): Documentation/assoc_array.txt: convert to ReST markup Documentation/atomic_ops.txt: convert to ReST markup Documentation/local_ops.txt: convert to ReST markup firmware: remove warning at documentation generation time Documentation/assoc_array.txt => Documentation/core-api/assoc_array.rst | 617 ++++++++++++++++++++++++++++-------------------------------- Documentation/atomic_ops.txt => Documentation/core-api/atomic_ops.rst | 745 +++++++++++++++++++++++++++++++++++++----------------------------------- Documentation/core-api/index.rst | 3 +- Documentation/local_ops.txt => Documentation/core-api/local_ops.rst | 275 ++++++++++++++------------- Documentation/process/volatile-considered-harmful.rst | 3 +- drivers/base/firmware_class.c | 6 +- 6 files changed, 838 insertions(+), 811 deletions(-) base-commit: 9c240d757658a3ae9968dd309e674c61f07c7f48 -- git-series 0.9.1 ^ permalink raw reply [flat|nested] 51+ messages in thread
* [PATCH v2 1/4] Documentation/assoc_array.txt: convert to ReST markup 2016-11-25 10:02 ` [PATCH v2 " Silvio Fricke @ 2016-11-25 10:02 ` Silvio Fricke 2016-11-25 13:03 ` Mauro Carvalho Chehab 2016-11-25 10:02 ` [PATCH v2 2/4] Documentation/atomic_ops.txt: " Silvio Fricke ` (3 subsequent siblings) 4 siblings, 1 reply; 51+ messages in thread From: Silvio Fricke @ 2016-11-25 10:02 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Silvio Fricke ... and move to Documentation/core-api folder. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- Documentation/assoc_array.txt => Documentation/core-api/assoc_array.rst | 617 ++++++++++++++++++++++++++++++++++-------------------------------------- Documentation/core-api/index.rst | 1 +- 2 files changed, 297 insertions(+), 321 deletions(-) diff --git a/Documentation/assoc_array.txt b/Documentation/core-api/assoc_array.rst similarity index 46% rename from Documentation/assoc_array.txt rename to Documentation/core-api/assoc_array.rst index 2f2c6cd..67a3a50 100644 --- a/Documentation/assoc_array.txt +++ b/Documentation/core-api/assoc_array.rst @@ -1,67 +1,46 @@ - ======================================== - GENERIC ASSOCIATIVE ARRAY IMPLEMENTATION - ======================================== +======================================== +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 +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. +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. +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. +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. +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. +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. +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. +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. +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. +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. +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. +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 @@ -71,25 +50,24 @@ 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 ============== -The public API can be found in <linux/assoc_array.h>. The associative array is -rooted on the following structure: +The public API can be found in ``<linux/assoc_array.h>``. The associative +array is rooted on the following structure:: - struct assoc_array { - ... - }; + struct assoc_array { + ... + }; -The code is selected by enabling CONFIG_ASSOCIATIVE_ARRAY. +The code is selected by enabling ``CONFIG_ASSOCIATIVE_ARRAY``. -EDIT SCRIPT +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 +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. @@ -99,246 +77,245 @@ 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: +The script appears as outside of the API as a pointer of the type:: - struct assoc_array_edit; + struct assoc_array_edit; There are two functions for dealing with the script: - (1) Apply an edit script. +1. Apply an edit script. :: - void assoc_array_apply_edit(struct assoc_array_edit *edit); + 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. +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. +2. Cancel an edit script. :: - void assoc_array_cancel_edit(struct assoc_array_edit *edit); + 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. +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 +Operations Table ---------------- -Various functions take a table of operations: +Various functions take a table of operations:: - struct assoc_array_ops { - ... - }; + 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: +1. Get a chunk of index key from caller data:: - unsigned long (*get_key_chunk)(const void *index_key, int level); + 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. +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. +2. Get a chunk of an object's index key. :: - unsigned long (*get_object_key_chunk)(const void *object, int level); + 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. +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. +3. See if this is the object we're looking for. :: - bool (*compare_object)(const void *object, const void *index_key); + 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. +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. +4. Diff the index keys of two objects. :: - int (*diff_objects)(const void *object, const void *index_key); + int (*diff_objects)(const void *object, const void *index_key); - Return the bit position at which the index key of the specified object - differs from the given index key or -1 if they are the same. +Return the bit position at which the index key of the specified object +differs from the given index key or -1 if they are the same. - (5) Free an object. +5. Free an object. :: - void (*free_object)(void *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. +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 +Manipulation Functions ---------------------- There are a number of functions for manipulating an associative array: - (1) Initialise an associative array. +1. Initialise an associative array. :: - void assoc_array_init(struct assoc_array *array); + void assoc_array_init(struct assoc_array *array); - This initialises the base structure for an associative array. It can't - fail. +This initialises the base structure for an associative array. It can't fail. - (2) Insert/replace an object in an associative array. +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); + 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. +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. +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. +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. +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. +The caller should lock exclusively against other modifiers of the array. - (3) Delete an object from an associative 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); + 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. +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. +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. +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. +The caller should lock exclusively against other modifiers of the array. - (4) Delete all objects from an associative 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); + 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 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. +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. +The caller should lock exclusively against other modifiers of the array. - (5) Destroy an associative array, deleting all objects. +5. Destroy an associative array, deleting all objects. :: - void assoc_array_destroy(struct assoc_array *array, - const struct assoc_array_ops *ops); + 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. +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. +The caller should lock exclusively against other modifiers and accessors +of the array. - (6) Garbage collect an associative 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); + 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. +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 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 ``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. +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. +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 +Access Functions ---------------- There are two functions for accessing an associative array: - (1) Iterate over all the objects in 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); + 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 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. +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. +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. +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); + 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 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. +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. +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 +Index Key Form -------------- The index key can be of any form, but since the algorithms aren't told how long @@ -364,8 +341,7 @@ unlikely that more than one word of any particular index key will have to be used. -================= -INTERNAL WORKINGS +Internal Workings ================= The associative array data structure has an internal tree. This tree is @@ -373,82 +349,80 @@ 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. +* 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 +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 | - +---+ +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 +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 +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 @@ -462,23 +436,23 @@ 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: +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 + 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: +and node B:: - 3 PTR TO NODE C 13* - any LEAF 1458952489 + 3 PTR TO NODE C 13* + any LEAF 1458952489 -SHORTCUTS +Shortcuts --------- Shortcuts are metadata records that jump over a piece of keyspace. A shortcut @@ -486,12 +460,13 @@ 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. +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 +Splitting And Collapsing Nodes ------------------------------ Each node has a maximum capacity of 16 leaves and metadata pointers. If the @@ -508,7 +483,7 @@ fewer, then the subtree will be collapsed down to a single node - and this will ripple towards the root if possible. -NON-RECURSIVE ITERATION +Non-Recursive Iteration ----------------------- Each node and shortcut contains a back pointer to its parent and the number of @@ -519,55 +494,55 @@ make sure progress is made without the need for a stack. The backpointers, however, make simultaneous alteration and iteration tricky. -SIMULTANEOUS ALTERATION AND ITERATION +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. +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. +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). +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). +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. + 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). +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. +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. +* 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. + 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 diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index f7ef7fd..480d9a3 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -7,6 +7,7 @@ Kernel and driver related documentation. .. toctree:: :maxdepth: 1 + assoc_array workqueue .. only:: subproject -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* Re: [PATCH v2 1/4] Documentation/assoc_array.txt: convert to ReST markup 2016-11-25 10:02 ` [PATCH v2 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke @ 2016-11-25 13:03 ` Mauro Carvalho Chehab 0 siblings, 0 replies; 51+ messages in thread From: Mauro Carvalho Chehab @ 2016-11-25 13:03 UTC (permalink / raw) To: Silvio Fricke Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel Em Fri, 25 Nov 2016 11:02:39 +0100 Silvio Fricke <silvio.fricke@gmail.com> escreveu: > ... and move to Documentation/core-api folder. > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> > --- > Documentation/assoc_array.txt => Documentation/core-api/assoc_array.rst | 617 ++++++++++++++++++++++++++++++++++-------------------------------------- > Documentation/core-api/index.rst | 1 +- > 2 files changed, 297 insertions(+), 321 deletions(-) > > diff --git a/Documentation/assoc_array.txt b/Documentation/core-api/assoc_array.rst > similarity index 46% > rename from Documentation/assoc_array.txt > rename to Documentation/core-api/assoc_array.rst > index 2f2c6cd..67a3a50 100644 > --- a/Documentation/assoc_array.txt > +++ b/Documentation/core-api/assoc_array.rst > @@ -1,67 +1,46 @@ > - ======================================== > - GENERIC ASSOCIATIVE ARRAY IMPLEMENTATION > - ======================================== > +======================================== > +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 > +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. > +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.** Instead, use: .. note:: Pointers to objects **must** be zero in the least significant bit. There's no underline on ReST. We're converting underlines to bold. > > - (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. > +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. > +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. > +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. > +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. > +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. > +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. > +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. > +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. > +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. > +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 > @@ -71,25 +50,24 @@ 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 > ============== Better to keep API in uppercase. > > -The public API can be found in <linux/assoc_array.h>. The associative array is > -rooted on the following structure: > +The public API can be found in ``<linux/assoc_array.h>``. The associative > +array is rooted on the following structure:: > > - struct assoc_array { > - ... > - }; > + struct assoc_array { > + ... > + }; > > -The code is selected by enabling CONFIG_ASSOCIATIVE_ARRAY. > +The code is selected by enabling ``CONFIG_ASSOCIATIVE_ARRAY``. I would add the command line command used to enable it, e. g.: The code is selected by enabling ``CONFIG_ASSOCIATIVE_ARRAY`` with:: ./scripts/config -e ASSOCIATIVE_ARRAY > > > -EDIT SCRIPT > +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 > +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. > @@ -99,246 +77,245 @@ 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: > +The script appears as outside of the API as a pointer of the type:: > > - struct assoc_array_edit; > + struct assoc_array_edit; > > There are two functions for dealing with the script: > > - (1) Apply an edit script. > +1. Apply an edit script. :: > > - void assoc_array_apply_edit(struct assoc_array_edit *edit); > + 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. > +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. > +2. Cancel an edit script. :: > > - void assoc_array_cancel_edit(struct assoc_array_edit *edit); > + 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. > +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 > +Operations Table > ---------------- > > -Various functions take a table of operations: > +Various functions take a table of operations:: > > - struct assoc_array_ops { > - ... > - }; > + 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: > +1. Get a chunk of index key from caller data:: > > - unsigned long (*get_key_chunk)(const void *index_key, int level); > + 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. > +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. > +2. Get a chunk of an object's index key. :: Better to remove the dot here, E. g: 2. Get a chunk of an object's index key:: > > - unsigned long (*get_object_key_chunk)(const void *object, int level); > + 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. > +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. > +3. See if this is the object we're looking for. :: Same here and on other similar places > > - bool (*compare_object)(const void *object, const void *index_key); > + 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. > +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. > +4. Diff the index keys of two objects. :: > > - int (*diff_objects)(const void *object, const void *index_key); > + int (*diff_objects)(const void *object, const void *index_key); > > - Return the bit position at which the index key of the specified object > - differs from the given index key or -1 if they are the same. > +Return the bit position at which the index key of the specified object > +differs from the given index key or -1 if they are the same. > > > - (5) Free an object. > +5. Free an object. :: > > - void (*free_object)(void *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. > +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 > +Manipulation Functions > ---------------------- > > There are a number of functions for manipulating an associative array: > > - (1) Initialise an associative array. > +1. Initialise an associative array. :: > > - void assoc_array_init(struct assoc_array *array); > + void assoc_array_init(struct assoc_array *array); > > - This initialises the base structure for an associative array. It can't > - fail. > +This initialises the base structure for an associative array. It can't fail. > > > - (2) Insert/replace an object in an associative array. > +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); > + 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. > +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. > +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. > +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. > +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. > +The caller should lock exclusively against other modifiers of the array. > > > - (3) Delete an object from an associative 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); > + 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. > +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. > +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. > +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. > +The caller should lock exclusively against other modifiers of the array. > > > - (4) Delete all objects from an associative 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); > + 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 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. > +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. > +The caller should lock exclusively against other modifiers of the array. > > > - (5) Destroy an associative array, deleting all objects. > +5. Destroy an associative array, deleting all objects. :: > > - void assoc_array_destroy(struct assoc_array *array, > - const struct assoc_array_ops *ops); > + 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. > +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. > +The caller should lock exclusively against other modifiers and accessors > +of the array. > > > - (6) Garbage collect an associative 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); > + 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. > +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 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 ``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. > +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. > +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 > +Access Functions > ---------------- > > There are two functions for accessing an associative array: > > - (1) Iterate over all the objects in 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); > + 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 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. > +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. > +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. > +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); > + 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 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. > +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. > +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 > +Index Key Form > -------------- > > The index key can be of any form, but since the algorithms aren't told how long > @@ -364,8 +341,7 @@ unlikely that more than one word of any particular index key will have to be > used. > > > -================= > -INTERNAL WORKINGS > +Internal Workings > ================= > > The associative array data structure has an internal tree. This tree is > @@ -373,82 +349,80 @@ 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. > +* 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 > +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 | > - +---+ > +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 > +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 > +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 > @@ -462,23 +436,23 @@ 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: > +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 > + 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: > +and node B:: > > - 3 PTR TO NODE C 13* > - any LEAF 1458952489 > + 3 PTR TO NODE C 13* > + any LEAF 1458952489 > > > -SHORTCUTS > +Shortcuts > --------- > > Shortcuts are metadata records that jump over a piece of keyspace. A shortcut > @@ -486,12 +460,13 @@ 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. > +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 > +Splitting And Collapsing Nodes > ------------------------------ > > Each node has a maximum capacity of 16 leaves and metadata pointers. If the > @@ -508,7 +483,7 @@ fewer, then the subtree will be collapsed down to a single node - and this will > ripple towards the root if possible. > > > -NON-RECURSIVE ITERATION > +Non-Recursive Iteration > ----------------------- > > Each node and shortcut contains a back pointer to its parent and the number of > @@ -519,55 +494,55 @@ make sure progress is made without the need for a stack. > The backpointers, however, make simultaneous alteration and iteration tricky. > > > -SIMULTANEOUS ALTERATION AND ITERATION > +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. > +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. > +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). > +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). > +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. > + 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). > +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. > +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: Please use: .. 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. > +* Under some circumstances, we need to simultaneously change the parent ... and remove the asterisk. > + 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. > + 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 > diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst > index f7ef7fd..480d9a3 100644 > --- a/Documentation/core-api/index.rst > +++ b/Documentation/core-api/index.rst > @@ -7,6 +7,7 @@ Kernel and driver related documentation. > .. toctree:: > :maxdepth: 1 > > + assoc_array > workqueue > > .. only:: subproject Thanks, Mauro ^ permalink raw reply [flat|nested] 51+ messages in thread
* [PATCH v2 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-25 10:02 ` [PATCH v2 " Silvio Fricke 2016-11-25 10:02 ` [PATCH v2 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke @ 2016-11-25 10:02 ` Silvio Fricke 2016-11-25 13:25 ` Mauro Carvalho Chehab 2016-11-25 10:02 ` [PATCH v2 3/4] Documentation/local_ops.txt: " Silvio Fricke ` (2 subsequent siblings) 4 siblings, 1 reply; 51+ messages in thread From: Silvio Fricke @ 2016-11-25 10:02 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Silvio Fricke ... and move to core-api folder. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- Documentation/atomic_ops.txt => Documentation/core-api/atomic_ops.rst | 745 +++++++++++++++++++++++++++++++++++++----------------------------------- Documentation/core-api/index.rst | 1 +- Documentation/process/volatile-considered-harmful.rst | 3 +- 3 files changed, 391 insertions(+), 358 deletions(-) diff --git a/Documentation/atomic_ops.txt b/Documentation/core-api/atomic_ops.rst similarity index 47% rename from Documentation/atomic_ops.txt rename to Documentation/core-api/atomic_ops.rst index 6c5e8a9..3ac94ff 100644 --- a/Documentation/atomic_ops.txt +++ b/Documentation/core-api/atomic_ops.rst @@ -1,204 +1,212 @@ - Semantics and Behavior of Atomic and - Bitmask Operations +======================================================= +Semantics and Behavior of Atomic and Bitmask Operations +======================================================= - David S. Miller +:Author: David S. Miller - This document is intended to serve as a guide to Linux port +This document is intended to serve as a guide to Linux port maintainers on how to implement atomic counter, bitops, and spinlock interfaces properly. - The atomic_t type should be defined as a signed integer and -the atomic_long_t type as a signed long integer. Also, they should +Atomic Type And Operations +========================== + +The ``atomic_t type`` should be defined as a signed integer and +the ``atomic_long_t`` type as a signed long integer. Also, they should be made opaque such that any kind of cast to a normal C integer type -will fail. Something like the following should suffice: +will fail. Something like the following should suffice:: - typedef struct { int counter; } atomic_t; - typedef struct { long counter; } atomic_long_t; + typedef struct { int counter; } atomic_t; + typedef struct { long counter; } atomic_long_t; Historically, counter has been declared volatile. This is now discouraged. -See Documentation/process/volatile-considered-harmful.rst for the complete rationale. +See :ref:`Documentation/process/volatile-considered-harmful.rst +<volatile_considered_harmful>` for the complete rationale. -local_t is very similar to atomic_t. If the counter is per CPU and only -updated by one CPU, local_t is probably more appropriate. Please see -Documentation/local_ops.txt for the semantics of local_t. +``local_t`` is very similar to ``atomic_t``. If the counter is per CPU and only +updated by one CPU, ``local_t`` is probably more appropriate. Please see +:ref:`Documentation/local_ops.txt <local_ops>` for the semantics of ``local_t``. -The first operations to implement for atomic_t's are the initializers and -plain reads. +The first operations to implement for ``atomic_t``'s are the initializers and +plain reads. :: - #define ATOMIC_INIT(i) { (i) } - #define atomic_set(v, i) ((v)->counter = (i)) + #define ATOMIC_INIT(i) { (i) } + #define atomic_set(v, i) ((v)->counter = (i)) -The first macro is used in definitions, such as: +The first macro is used in definitions, such as:: -static atomic_t my_counter = ATOMIC_INIT(1); + static atomic_t my_counter = ATOMIC_INIT(1); The initializer is atomic in that the return values of the atomic operations are guaranteed to be correct reflecting the initialized value if the initializer is used before runtime. If the initializer is used at runtime, a proper implicit or explicit read memory barrier is needed before reading the -value with atomic_read from another thread. +value with ``atomic_read`` from another thread. -As with all of the atomic_ interfaces, replace the leading "atomic_" -with "atomic_long_" to operate on atomic_long_t. +As with all of the ``atomic_`` interfaces, replace the leading ``atomic_`` +with ``atomic_long_`` to operate on ``atomic_long_t``. -The second interface can be used at runtime, as in: +The second interface can be used at runtime, as in:: - struct foo { atomic_t counter; }; - ... + struct foo { atomic_t counter; }; + ... - struct foo *k; + struct foo *k; - k = kmalloc(sizeof(*k), GFP_KERNEL); - if (!k) - return -ENOMEM; - atomic_set(&k->counter, 0); + k = kmalloc(sizeof(*k), GFP_KERNEL); + if (!k) + return -ENOMEM; + atomic_set(&k->counter, 0); The setting is atomic in that the return values of the atomic operations by all threads are guaranteed to be correct reflecting either the value that has been set with this operation or set with another operation. A proper implicit or explicit memory barrier is needed before the value set with the operation -is guaranteed to be readable with atomic_read from another thread. +is guaranteed to be readable with ``atomic_read`` from another thread. -Next, we have: +Next, we have:: - #define atomic_read(v) ((v)->counter) + #define atomic_read(v) ((v)->counter) which simply reads the counter value currently visible to the calling thread. The read is atomic in that the return value is guaranteed to be one of the values initialized or modified with the interface operations if a proper implicit or explicit memory barrier is used after possible runtime initialization by any other thread and the value is modified only with the -interface operations. atomic_read does not guarantee that the runtime +interface operations. ``atomic_read`` does not guarantee that the runtime initialization by any other thread is visible yet, so the user of the interface must take care of that with a proper implicit or explicit memory barrier. -*** WARNING: atomic_read() and atomic_set() DO NOT IMPLY BARRIERS! *** +.. warning:: -Some architectures may choose to use the volatile keyword, barriers, or inline -assembly to guarantee some degree of immediacy for atomic_read() and -atomic_set(). This is not uniformly guaranteed, and may change in the future, -so all users of atomic_t should treat atomic_read() and atomic_set() as simple -C statements that may be reordered or optimized away entirely by the compiler -or processor, and explicitly invoke the appropriate compiler and/or memory -barrier for each use case. Failure to do so will result in code that may -suddenly break when used with different architectures or compiler -optimizations, or even changes in unrelated code which changes how the -compiler optimizes the section accessing atomic_t variables. + ``atomic_read()`` and ``atomic_set()`` DO NOT IMPLY BARRIERS! -*** YOU HAVE BEEN WARNED! *** + Some architectures may choose to use the ``volatile`` keyword, barriers, or + ``inline`` assembly to guarantee some degree of immediacy for ``atomic_read()`` + and ``atomic_set()``. This is not uniformly guaranteed, and may change in the + future, so all users of ``atomic_t`` should treat ``atomic_read()`` and + ``atomic_set()`` as simple C statements that may be reordered or optimized away + entirely by the compiler or processor, and explicitly invoke the appropriate + compiler and/or memory barrier for each use case. Failure to do so will result + in code that may suddenly break when used with different architectures or + compiler optimizations, or even changes in unrelated code which changes how the + compiler optimizes the section accessing ``atomic_t`` variables. Properly aligned pointers, longs, ints, and chars (and unsigned equivalents) may be atomically loaded from and stored to in the same -sense as described for atomic_read() and atomic_set(). The READ_ONCE() -and WRITE_ONCE() macros should be used to prevent the compiler from using -optimizations that might otherwise optimize accesses out of existence on -the one hand, or that might create unsolicited accesses on the other. +sense as described for ``atomic_read()`` and ``atomic_set()``. The +``READ_ONCE()`` and ``WRITE_ONCE()`` macros should be used to prevent the +compiler from using optimizations that might otherwise optimize accesses out of +existence on the one hand, or that might create unsolicited accesses on the +other. -For example consider the following code: +For example consider the following code:: - while (a > 0) - do_something(); + while (a > 0) + do_something(); -If the compiler can prove that do_something() does not store to the +If the compiler can prove that ``do_something()`` does not store to the variable a, then the compiler is within its rights transforming this to -the following: +the following:: - tmp = a; - if (a > 0) - for (;;) - do_something(); + tmp = a; + if (a > 0) + for (;;) + do_something(); If you don't want the compiler to do this (and you probably don't), then -you should use something like the following: +you should use something like the following:: - while (READ_ONCE(a) < 0) - do_something(); + while (READ_ONCE(a) < 0) + do_something(); -Alternatively, you could place a barrier() call in the loop. +Alternatively, you could place a ``barrier()`` call in the loop. -For another example, consider the following code: +For another example, consider the following code:: - tmp_a = a; - do_something_with(tmp_a); - do_something_else_with(tmp_a); + tmp_a = a; + do_something_with(tmp_a); + do_something_else_with(tmp_a); -If the compiler can prove that do_something_with() does not store to the -variable a, then the compiler is within its rights to manufacture an -additional load as follows: +If the compiler can prove that ``do_something_with()`` does not store to the +variable ``a``, then the compiler is within its rights to manufacture an +additional load as follows:: - tmp_a = a; - do_something_with(tmp_a); - tmp_a = a; - do_something_else_with(tmp_a); + tmp_a = a; + do_something_with(tmp_a); + tmp_a = a; + do_something_else_with(tmp_a); This could fatally confuse your code if it expected the same value -to be passed to do_something_with() and do_something_else_with(). +to be passed to ``do_something_with()`` and ``do_something_else_with()``. The compiler would be likely to manufacture this additional load if -do_something_with() was an inline function that made very heavy use +``do_something_with()`` was an inline function that made very heavy use of registers: reloading from variable a could save a flush to the stack and later reload. To prevent the compiler from attacking your -code in this manner, write the following: +code in this manner, write the following:: - tmp_a = READ_ONCE(a); - do_something_with(tmp_a); - do_something_else_with(tmp_a); + tmp_a = READ_ONCE(a); + do_something_with(tmp_a); + do_something_else_with(tmp_a); For a final example, consider the following code, assuming that the -variable a is set at boot time before the second CPU is brought online -and never changed later, so that memory barriers are not needed: +variable ``a`` is set at boot time before the second CPU is brought online +and never changed later, so that memory barriers are not needed:: - if (a) - b = 9; - else - b = 42; + if (a) + b = 9; + else + b = 42; The compiler is within its rights to manufacture an additional store -by transforming the above code into the following: +by transforming the above code into the following:: - b = 42; - if (a) - b = 9; + b = 42; + if (a) + b = 9; This could come as a fatal surprise to other code running concurrently -that expected b to never have the value 42 if a was zero. To prevent -the compiler from doing this, write something like: +that expected ``b`` to never have the value ``42`` if ``a`` was zero. To +prevent the compiler from doing this, write something like:: + + if (a) + WRITE_ONCE(b, 9); + else + WRITE_ONCE(b, 42); - if (a) - WRITE_ONCE(b, 9); - else - WRITE_ONCE(b, 42); +Don't even **think** about doing this without proper use of memory barriers, +locks, or atomic operations if variable ``a`` can change at runtime! -Don't even -think- about doing this without proper use of memory barriers, -locks, or atomic operations if variable a can change at runtime! +.. warning:: -*** WARNING: READ_ONCE() OR WRITE_ONCE() DO NOT IMPLY A BARRIER! *** + ``READ_ONCE()`` OR ``WRITE_ONCE()`` DO NOT IMPLY A BARRIER! Now, we move onto the atomic operation interfaces typically implemented with -the help of assembly code. +the help of assembly code. :: - void atomic_add(int i, atomic_t *v); - void atomic_sub(int i, atomic_t *v); - void atomic_inc(atomic_t *v); - void atomic_dec(atomic_t *v); + void atomic_add(int i, atomic_t *v); + void atomic_sub(int i, atomic_t *v); + void atomic_inc(atomic_t *v); + void atomic_dec(atomic_t *v); These four routines add and subtract integral values to/from the given -atomic_t value. The first two routines pass explicit integers by +``atomic_t`` value. The first two routines pass explicit integers by which to make the adjustment, whereas the latter two use an implicit adjustment value of "1". -One very important aspect of these two routines is that they DO NOT +One very important aspect of these two routines is that they **DO NOT** require any explicit memory barriers. They need only perform the -atomic_t counter update in an SMP safe manner. +``atomic_t`` counter update in an SMP safe manner. -Next, we have: +Next, we have:: - int atomic_inc_return(atomic_t *v); - int atomic_dec_return(atomic_t *v); + int atomic_inc_return(atomic_t *v); + int atomic_dec_return(atomic_t *v); These routines add 1 and subtract 1, respectively, from the given -atomic_t and return the new counter value after the operation is +``atomic_t`` and return the new counter value after the operation is performed. Unlike the above routines, it is required that these primitives @@ -207,317 +215,330 @@ the operation. It must be done such that all memory operations before and after the atomic operation calls are strongly ordered with respect to the atomic operation itself. -For example, it should behave as if a smp_mb() call existed both +For example, it should behave as if a ``smp_mb()`` call existed both before and after the atomic operation. If the atomic instructions used in an implementation provide explicit memory barrier semantics which satisfy the above requirements, that is fine as well. -Let's move on: +Let's move on:: - int atomic_add_return(int i, atomic_t *v); - int atomic_sub_return(int i, atomic_t *v); + int atomic_add_return(int i, atomic_t *v); + int atomic_sub_return(int i, atomic_t *v); -These behave just like atomic_{inc,dec}_return() except that an +These behave just like ``atomic_{inc,dec}_return()`` except that an explicit counter adjustment is given instead of the implicit "1". -This means that like atomic_{inc,dec}_return(), the memory barrier +This means that like ``atomic_{inc,dec}_return()``, the memory barrier semantics are required. -Next: +Next:: - int atomic_inc_and_test(atomic_t *v); - int atomic_dec_and_test(atomic_t *v); + int atomic_inc_and_test(atomic_t *v); + int atomic_dec_and_test(atomic_t *v); These two routines increment and decrement by 1, respectively, the given atomic counter. They return a boolean indicating whether the resulting counter value was zero or not. Again, these primitives provide explicit memory barrier semantics around -the atomic operation. +the atomic operation. :: - int atomic_sub_and_test(int i, atomic_t *v); + int atomic_sub_and_test(int i, atomic_t *v); -This is identical to atomic_dec_and_test() except that an explicit +This is identical to ``atomic_dec_and_test()`` except that an explicit decrement is given instead of the implicit "1". This primitive must -provide explicit memory barrier semantics around the operation. +provide explicit memory barrier semantics around the operation. :: - int atomic_add_negative(int i, atomic_t *v); + int atomic_add_negative(int i, atomic_t *v); The given increment is added to the given atomic counter value. A boolean is return which indicates whether the resulting counter value is negative. This primitive must provide explicit memory barrier semantics around the operation. -Then: +Then:: - int atomic_xchg(atomic_t *v, int new); + int atomic_xchg(atomic_t *v, int new); This performs an atomic exchange operation on the atomic variable v, setting the given new value. It returns the old value that the atomic variable v had just before the operation. -atomic_xchg must provide explicit memory barriers around the operation. +``atomic_xchg`` must provide explicit memory barriers around the operation. :: - int atomic_cmpxchg(atomic_t *v, int old, int new); + int atomic_cmpxchg(atomic_t *v, int old, int new); This performs an atomic compare exchange operation on the atomic value v, -with the given old and new values. Like all atomic_xxx operations, -atomic_cmpxchg will only satisfy its atomicity semantics as long as all -other accesses of *v are performed through atomic_xxx operations. +with the given old and new values. Like all ``atomic_*`` operations, +``atomic_cmpxchg`` will only satisfy its atomicity semantics as long as all +other accesses of ``*v`` are performed through ``atomic_*`` operations. -atomic_cmpxchg must provide explicit memory barriers around the operation, +``atomic_cmpxchg`` must provide explicit memory barriers around the operation, although if the comparison fails then no memory ordering guarantees are required. -The semantics for atomic_cmpxchg are the same as those defined for 'cas' +The semantics for ``atomic_cmpxchg`` are the same as those defined for 'cas' below. -Finally: +Finally:: - int atomic_add_unless(atomic_t *v, int a, int u); + int atomic_add_unless(atomic_t *v, int a, int u); If the atomic value v is not equal to u, this function adds a to v, and returns non zero. If v is equal to u then it returns zero. This is done as an atomic operation. -atomic_add_unless must provide explicit memory barriers around the +``atomic_add_unless`` must provide explicit memory barriers around the operation unless it fails (returns 0). -atomic_inc_not_zero, equivalent to atomic_add_unless(v, 1, 0) +``atomic_inc_not_zero``, equivalent to ``atomic_add_unless(v, 1, 0)`` -If a caller requires memory barrier semantics around an atomic_t +If a caller requires memory barrier semantics around an ``atomic_t`` operation which does not return a value, a set of interfaces are -defined which accomplish this: +defined which accomplish this:: - void smp_mb__before_atomic(void); - void smp_mb__after_atomic(void); + void smp_mb__before_atomic(void); + void smp_mb__after_atomic(void); -For example, smp_mb__before_atomic() can be used like so: +For example, ``smp_mb__before_atomic()`` can be used like so:: - obj->dead = 1; - smp_mb__before_atomic(); - atomic_dec(&obj->ref_count); + obj->dead = 1; + smp_mb__before_atomic(); + atomic_dec(&obj->ref_count); -It makes sure that all memory operations preceding the atomic_dec() +It makes sure that all memory operations preceding the ``atomic_dec()`` call are strongly ordered with respect to the atomic counter operation. In the above example, it guarantees that the assignment of -"1" to obj->dead will be globally visible to other cpus before the +"1" to ``obj->dead`` will be globally visible to other cpus before the atomic counter decrement. -Without the explicit smp_mb__before_atomic() call, the +Without the explicit ``smp_mb__before_atomic()`` call, the implementation could legally allow the atomic counter update visible -to other cpus before the "obj->dead = 1;" assignment. +to other cpus before the ``obj->dead = 1;`` assignment. A missing memory barrier in the cases where they are required by the -atomic_t implementation above can have disastrous results. Here is +``atomic_t`` implementation above can have disastrous results. Here is an example, which follows a pattern occurring frequently in the Linux kernel. It is the use of atomic counters to implement reference counting, and it works such that once the counter falls to zero it can be guaranteed that no other entity can be accessing the object: -static void obj_list_add(struct obj *obj, struct list_head *head) -{ - obj->active = 1; - list_add(&obj->list, head); -} - -static void obj_list_del(struct obj *obj) -{ - list_del(&obj->list); - obj->active = 0; -} - -static void obj_destroy(struct obj *obj) -{ - BUG_ON(obj->active); - kfree(obj); -} - -struct obj *obj_list_peek(struct list_head *head) -{ - if (!list_empty(head)) { - struct obj *obj; - - obj = list_entry(head->next, struct obj, list); - atomic_inc(&obj->refcnt); - return obj; - } - return NULL; -} - -void obj_poke(void) -{ - struct obj *obj; - - spin_lock(&global_list_lock); - obj = obj_list_peek(&global_list); - spin_unlock(&global_list_lock); - - if (obj) { - obj->ops->poke(obj); - if (atomic_dec_and_test(&obj->refcnt)) - obj_destroy(obj); - } -} - -void obj_timeout(struct obj *obj) -{ - spin_lock(&global_list_lock); - obj_list_del(obj); - spin_unlock(&global_list_lock); - - if (atomic_dec_and_test(&obj->refcnt)) - obj_destroy(obj); -} - -(This is a simplification of the ARP queue management in the - generic neighbour discover code of the networking. Olaf Kirch - found a bug wrt. memory barriers in kfree_skb() that exposed - the atomic_t memory barrier requirements quite clearly.) - -Given the above scheme, it must be the case that the obj->active +.. code-block:: c + + static void obj_list_add(struct obj *obj, struct list_head *head) + { + obj->active = 1; + list_add(&obj->list, head); + } + + static void obj_list_del(struct obj *obj) + { + list_del(&obj->list); + obj->active = 0; + } + + static void obj_destroy(struct obj *obj) + { + BUG_ON(obj->active); + kfree(obj); + } + + struct obj *obj_list_peek(struct list_head *head) + { + if (!list_empty(head)) { + struct obj *obj; + + obj = list_entry(head->next, struct obj, list); + atomic_inc(&obj->refcnt); + return obj; + } + return NULL; + } + + void obj_poke(void) + { + struct obj *obj; + + spin_lock(&global_list_lock); + obj = obj_list_peek(&global_list); + spin_unlock(&global_list_lock); + + if (obj) { + obj->ops->poke(obj); + if (atomic_dec_and_test(&obj->refcnt)) + obj_destroy(obj); + } + } + + void obj_timeout(struct obj *obj) + { + spin_lock(&global_list_lock); + obj_list_del(obj); + spin_unlock(&global_list_lock); + + if (atomic_dec_and_test(&obj->refcnt)) + obj_destroy(obj); + } + +.. note: + + This is a simplification of the ARP queue management in the + generic neighbour discover code of the networking. Olaf Kirch + found a bug wrt. memory barriers in kfree_skb() that exposed + the atomic_t memory barrier requirements quite clearly. + +Given the above scheme, it must be the case that the ``obj->active`` update done by the obj list deletion be visible to other processors before the atomic counter decrement is performed. -Otherwise, the counter could fall to zero, yet obj->active would still -be set, thus triggering the assertion in obj_destroy(). The error -sequence looks like this: - - cpu 0 cpu 1 - obj_poke() obj_timeout() - obj = obj_list_peek(); - ... gains ref to obj, refcnt=2 - obj_list_del(obj); - obj->active = 0 ... - ... visibility delayed ... - atomic_dec_and_test() - ... refcnt drops to 1 ... - atomic_dec_and_test() - ... refcount drops to 0 ... - obj_destroy() - BUG() triggers since obj->active - still seen as one - obj->active update visibility occurs +Otherwise, the counter could fall to zero, yet ``obj->active`` would still +be set, thus triggering the assertion in ``obj_destroy()``. The error +sequence looks like this:: + + cpu 0 cpu 1 + obj_poke() obj_timeout() + obj = obj_list_peek(); + ... gains ref to obj, refcnt=2 + obj_list_del(obj); + obj->active = 0 ... + ... visibility delayed ... + atomic_dec_and_test() + ... refcnt drops to 1 ... + atomic_dec_and_test() + ... refcount drops to 0 ... + obj_destroy() + BUG() triggers since obj->active + still seen as one + obj->active update visibility occurs With the memory barrier semantics required of the atomic_t operations which return values, the above sequence of memory visibility can never -happen. Specifically, in the above case the atomic_dec_and_test() +happen. Specifically, in the above case the ``atomic_dec_and_test()`` counter decrement would not become globally visible until the -obj->active update does. +``obj->active`` update does. As a historical note, 32-bit Sparc used to only allow usage of -24-bits of its atomic_t type. This was because it used 8 bits +24-bits of its ``atomic_t`` type. This was because it used 8 bits as a spinlock for SMP safety. Sparc32 lacked a "compare and swap" type instruction. However, 32-bit Sparc has since been moved over to a "hash table of spinlocks" scheme, that allows the full 32-bit counter to be realized. Essentially, an array of spinlocks are -indexed into based upon the address of the atomic_t being operated +indexed into based upon the address of the ``atomic_t`` being operated on, and that lock protects the atomic operation. Parisc uses the same scheme. -Another note is that the atomic_t operations returning values are +Another note is that the ``atomic_t`` operations returning values are extremely slow on an old 386. +Atomic Bitmask +============== + We will now cover the atomic bitmask operations. You will find that their SMP and memory barrier semantics are similar in shape and scope -to the atomic_t ops above. +to the ``atomic_t`` ops above. Native atomic bit operations are defined to operate on objects aligned -to the size of an "unsigned long" C data type, and are least of that -size. The endianness of the bits within each "unsigned long" are the -native endianness of the cpu. +to the size of an ``unsigned long`` C data type, and are least of that +size. The endianness of the bits within each ``unsigned long`` are the +native endianness of the cpu. :: - void set_bit(unsigned long nr, volatile unsigned long *addr); - void clear_bit(unsigned long nr, volatile unsigned long *addr); - void change_bit(unsigned long nr, volatile unsigned long *addr); + void set_bit(unsigned long nr, volatile unsigned long *addr); + void clear_bit(unsigned long nr, volatile unsigned long *addr); + void change_bit(unsigned long nr, volatile unsigned long *addr); These routines set, clear, and change, respectively, the bit number -indicated by "nr" on the bit mask pointed to by "ADDR". +indicated by ``nr`` on the bit mask pointed to by ``addr``. They must execute atomically, yet there are no implicit memory barrier -semantics required of these interfaces. +semantics required of these interfaces. :: - int test_and_set_bit(unsigned long nr, volatile unsigned long *addr); - int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); - int test_and_change_bit(unsigned long nr, volatile unsigned long *addr); + int test_and_set_bit(unsigned long nr, volatile unsigned long *addr); + int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); + int test_and_change_bit(unsigned long nr, volatile unsigned long *addr); Like the above, except that these routines return a boolean which indicates whether the changed bit was set _BEFORE_ the atomic bit operation. -WARNING! It is incredibly important that the value be a boolean, -ie. "0" or "1". Do not try to be fancy and save a few instructions by -declaring the above to return "long" and just returning something like -"old_val & mask" because that will not work. +.. warning:: -For one thing, this return value gets truncated to int in many code -paths using these interfaces, so on 64-bit if the bit is set in the -upper 32-bits then testers will never see that. + It is incredibly important that the value be a boolean, + ie. "0" or "1". Do not try to be fancy and save a few instructions by + declaring the above to return "long" and just returning something like + "old_val & mask" because that will not work. -One great example of where this problem crops up are the thread_info -flag operations. Routines such as test_and_set_ti_thread_flag() chop -the return value into an int. There are other places where things -like this occur as well. + For one thing, this return value gets truncated to int in many code + paths using these interfaces, so on 64-bit if the bit is set in the + upper 32-bits then testers will never see that. -These routines, like the atomic_t counter operations returning values, + One great example of where this problem crops up are the ``thread_info`` + flag operations. Routines such as ``test_and_set_ti_thread_flag()`` chop + the return value into an int. There are other places where things + like this occur as well. + +These routines, like the ``atomic_t`` counter operations returning values, must provide explicit memory barrier semantics around their execution. All memory operations before the atomic bit operation call must be made visible globally before the atomic bit operation is made visible. Likewise, the atomic bit operation must be visible globally before any subsequent memory operation is made visible. For example: - obj->dead = 1; - if (test_and_set_bit(0, &obj->flags)) - /* ... */; - obj->killed = 1; +.. code-block:: c + + obj->dead = 1; + if (test_and_set_bit(0, &obj->flags)) + /* ... */; + obj->killed = 1; -The implementation of test_and_set_bit() must guarantee that -"obj->dead = 1;" is visible to cpus before the atomic memory operation -done by test_and_set_bit() becomes visible. Likewise, the atomic -memory operation done by test_and_set_bit() must become visible before -"obj->killed = 1;" is visible. +The implementation of ``test_and_set_bit()`` must guarantee that +``obj->dead = 1;`` is visible to cpus before the atomic memory operation +done by ``test_and_set_bit()`` becomes visible. Likewise, the atomic +memory operation done by ``test_and_set_bit()`` must become visible before +``obj->killed = 1;`` is visible. -Finally there is the basic operation: +Finally there is the basic operation:: - int test_bit(unsigned long nr, __const__ volatile unsigned long *addr); + int test_bit(unsigned long nr, __const__ volatile unsigned long *addr); -Which returns a boolean indicating if bit "nr" is set in the bitmask -pointed to by "addr". +Which returns a boolean indicating if bit ``nr`` is set in the bitmask +pointed to by ``addr``. -If explicit memory barriers are required around {set,clear}_bit() (which do +If explicit memory barriers are required around ``{set,clear}_bit()`` (which do not return a value, and thus does not need to provide memory barrier -semantics), two interfaces are provided: +semantics), two interfaces are provided:: - void smp_mb__before_atomic(void); - void smp_mb__after_atomic(void); + void smp_mb__before_atomic(void); + void smp_mb__after_atomic(void); -They are used as follows, and are akin to their atomic_t operation +They are used as follows, and are akin to their ``atomic_t`` operation brothers: - /* All memory operations before this call will - * be globally visible before the clear_bit(). - */ - smp_mb__before_atomic(); - clear_bit( ... ); +.. code-block:: c - /* The clear_bit() will be visible before all - * subsequent memory operations. - */ - smp_mb__after_atomic(); + /* All memory operations before this call will + * be globally visible before the clear_bit(). + */ + smp_mb__before_atomic(); + clear_bit( ... ); + + /* The clear_bit() will be visible before all + * subsequent memory operations. + */ + smp_mb__after_atomic(); There are two special bitops with lock barrier semantics (acquire/release, same as spinlocks). These operate in the same way as their non-_lock/unlock postfixed variants, except that they are to provide acquire/release semantics, -respectively. This means they can be used for bit_spin_trylock and -bit_spin_unlock type operations without specifying any more barriers. +respectively. This means they can be used for ``bit_spin_trylock`` and +``bit_spin_unlock`` type operations without specifying any more barriers. :: - int test_and_set_bit_lock(unsigned long nr, unsigned long *addr); - void clear_bit_unlock(unsigned long nr, unsigned long *addr); - void __clear_bit_unlock(unsigned long nr, unsigned long *addr); + int test_and_set_bit_lock(unsigned long nr, unsigned long *addr); + void clear_bit_unlock(unsigned long nr, unsigned long *addr); + void __clear_bit_unlock(unsigned long nr, unsigned long *addr); -The __clear_bit_unlock version is non-atomic, however it still implements +The ``__clear_bit_unlock`` version is non-atomic, however it still implements unlock barrier semantics. This can be useful if the lock itself is protecting the other bits in the word. @@ -526,41 +547,43 @@ provided. They are used in contexts where some other higher-level SMP locking scheme is being used to protect the bitmask, and thus less expensive non-atomic operations may be used in the implementation. They have names similar to the above bitmask operation interfaces, -except that two underscores are prefixed to the interface name. +except that two underscores are prefixed to the interface name. :: - void __set_bit(unsigned long nr, volatile unsigned long *addr); - void __clear_bit(unsigned long nr, volatile unsigned long *addr); - void __change_bit(unsigned long nr, volatile unsigned long *addr); - int __test_and_set_bit(unsigned long nr, volatile unsigned long *addr); - int __test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); - int __test_and_change_bit(unsigned long nr, volatile unsigned long *addr); + void __set_bit(unsigned long nr, volatile unsigned long *addr); + void __clear_bit(unsigned long nr, volatile unsigned long *addr); + void __change_bit(unsigned long nr, volatile unsigned long *addr); + int __test_and_set_bit(unsigned long nr, volatile unsigned long *addr); + int __test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); + int __test_and_change_bit(unsigned long nr, volatile unsigned long *addr); These non-atomic variants also do not require any special memory barrier semantics. -The routines xchg() and cmpxchg() must provide the same exact +The routines ``xchg()`` and ``cmpxchg()`` must provide the same exact memory-barrier semantics as the atomic and bit operations returning values. -Note: If someone wants to use xchg(), cmpxchg() and their variants, -linux/atomic.h should be included rather than asm/cmpxchg.h, unless -the code is in arch/* and can take care of itself. +.. note:: + + If someone wants to use ``xchg()``, ``cmpxchg()`` and their variants, + ``linux/atomic.h`` should be included rather than ``asm/cmpxchg.h``, unless + the code is in arch/* and can take care of itself. Spinlocks and rwlocks have memory barrier expectations as well. The rule to follow is simple: -1) When acquiring a lock, the implementation must make it globally +1. When acquiring a lock, the implementation must make it globally visible before any subsequent memory operation. -2) When releasing a lock, the implementation must make it such that +2. When releasing a lock, the implementation must make it such that all previous memory operations are globally visible before the lock release. -Which finally brings us to _atomic_dec_and_lock(). There is an -architecture-neutral version implemented in lib/dec_and_lock.c, -but most platforms will wish to optimize this in assembler. +Which finally brings us to ``_atomic_dec_and_lock()``. There is an +architecture-neutral version implemented in ``lib/dec_and_lock.c``, +but most platforms will wish to optimize this in assembler. :: - int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock); + int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock); Atomically decrement the given counter, and if will drop to zero atomically acquire the given spinlock and perform the decrement @@ -573,68 +596,74 @@ sure the spinlock operation is globally visible before any subsequent memory operation. We can demonstrate this operation more clearly if we define -an abstract atomic operation: +an abstract atomic operation:: - long cas(long *mem, long old, long new); + long cas(long *mem, long old, long new); -"cas" stands for "compare and swap". It atomically: +"``cas``" stands for "compare and swap". It atomically: -1) Compares "old" with the value currently at "mem". -2) If they are equal, "new" is written to "mem". -3) Regardless, the current value at "mem" is returned. +1. Compares "old" with the value currently at "mem". +2. If they are equal, "new" is written to "mem". +3. Regardless, the current value at "mem" is returned. As an example usage, here is what an atomic counter update might look like: -void example_atomic_inc(long *counter) -{ - long old, new, ret; +.. code-block:: c - while (1) { - old = *counter; - new = old + 1; + void example_atomic_inc(long *counter) + { + long old, new, ret; - ret = cas(counter, old, new); - if (ret == old) - break; - } -} + while (1) { + old = *counter; + new = old + 1; + + ret = cas(counter, old, new); + if (ret == old) + break; + } + } Let's use cas() in order to build a pseudo-C atomic_dec_and_lock(): -int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock) -{ - long old, new, ret; - int went_to_zero; - - went_to_zero = 0; - while (1) { - old = atomic_read(atomic); - new = old - 1; - if (new == 0) { - went_to_zero = 1; - spin_lock(lock); - } - ret = cas(atomic, old, new); - if (ret == old) - break; - if (went_to_zero) { - spin_unlock(lock); - went_to_zero = 0; - } - } - - return went_to_zero; -} - -Now, as far as memory barriers go, as long as spin_lock() +.. code-block:: c + + int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock) + { + long old, new, ret; + int went_to_zero; + + went_to_zero = 0; + while (1) { + old = atomic_read(atomic); + new = old - 1; + if (new == 0) { + went_to_zero = 1; + spin_lock(lock); + } + ret = cas(atomic, old, new); + if (ret == old) + break; + if (went_to_zero) { + spin_unlock(lock); + went_to_zero = 0; + } + } + + return went_to_zero; + } + +Now, as far as memory barriers go, as long as i``spin_lock()`` strictly orders all subsequent memory operations (including -the cas()) with respect to itself, things will be fine. +the ``cas()``) with respect to itself, things will be fine. -Said another way, _atomic_dec_and_lock() must guarantee that +Said another way, ``_atomic_dec_and_lock()`` must guarantee that a counter dropping to zero is never made visible before the spinlock being acquired. -Note that this also means that for the case where the counter -is not dropping to zero, there are no memory ordering -requirements. +.. note:: + + Note that this also means that for the case where the counter + is not dropping to zero, there are no memory ordering + requirements. diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index 480d9a3..f3e5f5e 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -8,6 +8,7 @@ Kernel and driver related documentation. :maxdepth: 1 assoc_array + atomic_ops workqueue .. only:: subproject diff --git a/Documentation/process/volatile-considered-harmful.rst b/Documentation/process/volatile-considered-harmful.rst index e0d042a..4934e65 100644 --- a/Documentation/process/volatile-considered-harmful.rst +++ b/Documentation/process/volatile-considered-harmful.rst @@ -1,3 +1,6 @@ + +.. _volatile_considered_harmful: + Why the "volatile" type class should not be used ------------------------------------------------ -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* Re: [PATCH v2 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-25 10:02 ` [PATCH v2 2/4] Documentation/atomic_ops.txt: " Silvio Fricke @ 2016-11-25 13:25 ` Mauro Carvalho Chehab 0 siblings, 0 replies; 51+ messages in thread From: Mauro Carvalho Chehab @ 2016-11-25 13:25 UTC (permalink / raw) To: Silvio Fricke Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel Em Fri, 25 Nov 2016 11:02:40 +0100 Silvio Fricke <silvio.fricke@gmail.com> escreveu: > ... and move to core-api folder. > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> > --- > Documentation/atomic_ops.txt => Documentation/core-api/atomic_ops.rst | 745 +++++++++++++++++++++++++++++++++++++----------------------------------- > Documentation/core-api/index.rst | 1 +- > Documentation/process/volatile-considered-harmful.rst | 3 +- > 3 files changed, 391 insertions(+), 358 deletions(-) > > diff --git a/Documentation/atomic_ops.txt b/Documentation/core-api/atomic_ops.rst > similarity index 47% > rename from Documentation/atomic_ops.txt > rename to Documentation/core-api/atomic_ops.rst > index 6c5e8a9..3ac94ff 100644 > --- a/Documentation/atomic_ops.txt > +++ b/Documentation/core-api/atomic_ops.rst > @@ -1,204 +1,212 @@ > - Semantics and Behavior of Atomic and > - Bitmask Operations > +======================================================= > +Semantics and Behavior of Atomic and Bitmask Operations > +======================================================= > > - David S. Miller > +:Author: David S. Miller > > - This document is intended to serve as a guide to Linux port > +This document is intended to serve as a guide to Linux port > maintainers on how to implement atomic counter, bitops, and spinlock > interfaces properly. > > - The atomic_t type should be defined as a signed integer and > -the atomic_long_t type as a signed long integer. Also, they should > +Atomic Type And Operations > +========================== > + > +The ``atomic_t type`` should be defined as a signed integer and > +the ``atomic_long_t`` type as a signed long integer. Also, they should > be made opaque such that any kind of cast to a normal C integer type > -will fail. Something like the following should suffice: > +will fail. Something like the following should suffice:: > > - typedef struct { int counter; } atomic_t; > - typedef struct { long counter; } atomic_long_t; > + typedef struct { int counter; } atomic_t; > + typedef struct { long counter; } atomic_long_t; > > Historically, counter has been declared volatile. This is now discouraged. > -See Documentation/process/volatile-considered-harmful.rst for the complete rationale. > +See :ref:`Documentation/process/volatile-considered-harmful.rst > +<volatile_considered_harmful>` for the complete rationale. > > -local_t is very similar to atomic_t. If the counter is per CPU and only > -updated by one CPU, local_t is probably more appropriate. Please see > -Documentation/local_ops.txt for the semantics of local_t. > +``local_t`` is very similar to ``atomic_t``. If the counter is per CPU and only > +updated by one CPU, ``local_t`` is probably more appropriate. Please see > +:ref:`Documentation/local_ops.txt <local_ops>` for the semantics of ``local_t``. > > -The first operations to implement for atomic_t's are the initializers and > -plain reads. > +The first operations to implement for ``atomic_t``'s are the initializers and > +plain reads. :: > > - #define ATOMIC_INIT(i) { (i) } > - #define atomic_set(v, i) ((v)->counter = (i)) > + #define ATOMIC_INIT(i) { (i) } > + #define atomic_set(v, i) ((v)->counter = (i)) > > -The first macro is used in definitions, such as: > +The first macro is used in definitions, such as:: > > -static atomic_t my_counter = ATOMIC_INIT(1); > + static atomic_t my_counter = ATOMIC_INIT(1); > > The initializer is atomic in that the return values of the atomic operations > are guaranteed to be correct reflecting the initialized value if the > initializer is used before runtime. If the initializer is used at runtime, a > proper implicit or explicit read memory barrier is needed before reading the > -value with atomic_read from another thread. > +value with ``atomic_read`` from another thread. > > -As with all of the atomic_ interfaces, replace the leading "atomic_" > -with "atomic_long_" to operate on atomic_long_t. > +As with all of the ``atomic_`` interfaces, replace the leading ``atomic_`` > +with ``atomic_long_`` to operate on ``atomic_long_t``. > > -The second interface can be used at runtime, as in: > +The second interface can be used at runtime, as in:: > > - struct foo { atomic_t counter; }; > - ... > + struct foo { atomic_t counter; }; > + ... > > - struct foo *k; > + struct foo *k; > > - k = kmalloc(sizeof(*k), GFP_KERNEL); > - if (!k) > - return -ENOMEM; > - atomic_set(&k->counter, 0); > + k = kmalloc(sizeof(*k), GFP_KERNEL); > + if (!k) > + return -ENOMEM; > + atomic_set(&k->counter, 0); > > The setting is atomic in that the return values of the atomic operations by > all threads are guaranteed to be correct reflecting either the value that has > been set with this operation or set with another operation. A proper implicit > or explicit memory barrier is needed before the value set with the operation > -is guaranteed to be readable with atomic_read from another thread. > +is guaranteed to be readable with ``atomic_read`` from another thread. > > -Next, we have: > +Next, we have:: > > - #define atomic_read(v) ((v)->counter) > + #define atomic_read(v) ((v)->counter) > > which simply reads the counter value currently visible to the calling thread. > The read is atomic in that the return value is guaranteed to be one of the > values initialized or modified with the interface operations if a proper > implicit or explicit memory barrier is used after possible runtime > initialization by any other thread and the value is modified only with the > -interface operations. atomic_read does not guarantee that the runtime > +interface operations. ``atomic_read`` does not guarantee that the runtime > initialization by any other thread is visible yet, so the user of the > interface must take care of that with a proper implicit or explicit memory > barrier. > > -*** WARNING: atomic_read() and atomic_set() DO NOT IMPLY BARRIERS! *** > +.. warning:: > > -Some architectures may choose to use the volatile keyword, barriers, or inline > -assembly to guarantee some degree of immediacy for atomic_read() and > -atomic_set(). This is not uniformly guaranteed, and may change in the future, > -so all users of atomic_t should treat atomic_read() and atomic_set() as simple > -C statements that may be reordered or optimized away entirely by the compiler > -or processor, and explicitly invoke the appropriate compiler and/or memory > -barrier for each use case. Failure to do so will result in code that may > -suddenly break when used with different architectures or compiler > -optimizations, or even changes in unrelated code which changes how the > -compiler optimizes the section accessing atomic_t variables. > + ``atomic_read()`` and ``atomic_set()`` DO NOT IMPLY BARRIERS! > > -*** YOU HAVE BEEN WARNED! *** > + Some architectures may choose to use the ``volatile`` keyword, barriers, or > + ``inline`` assembly to guarantee some degree of immediacy for ``atomic_read()`` > + and ``atomic_set()``. This is not uniformly guaranteed, and may change in the > + future, so all users of ``atomic_t`` should treat ``atomic_read()`` and > + ``atomic_set()`` as simple C statements that may be reordered or optimized away > + entirely by the compiler or processor, and explicitly invoke the appropriate > + compiler and/or memory barrier for each use case. Failure to do so will result > + in code that may suddenly break when used with different architectures or > + compiler optimizations, or even changes in unrelated code which changes how the > + compiler optimizes the section accessing ``atomic_t`` variables. > > Properly aligned pointers, longs, ints, and chars (and unsigned > equivalents) may be atomically loaded from and stored to in the same > -sense as described for atomic_read() and atomic_set(). The READ_ONCE() > -and WRITE_ONCE() macros should be used to prevent the compiler from using > -optimizations that might otherwise optimize accesses out of existence on > -the one hand, or that might create unsolicited accesses on the other. > +sense as described for ``atomic_read()`` and ``atomic_set()``. The > +``READ_ONCE()`` and ``WRITE_ONCE()`` macros should be used to prevent the > +compiler from using optimizations that might otherwise optimize accesses out of > +existence on the one hand, or that might create unsolicited accesses on the > +other. > > -For example consider the following code: > +For example consider the following code:: > > - while (a > 0) > - do_something(); > + while (a > 0) > + do_something(); > > -If the compiler can prove that do_something() does not store to the > +If the compiler can prove that ``do_something()`` does not store to the > variable a, then the compiler is within its rights transforming this to > -the following: > +the following:: > > - tmp = a; > - if (a > 0) > - for (;;) > - do_something(); > + tmp = a; > + if (a > 0) > + for (;;) > + do_something(); > > If you don't want the compiler to do this (and you probably don't), then > -you should use something like the following: > +you should use something like the following:: > > - while (READ_ONCE(a) < 0) > - do_something(); > + while (READ_ONCE(a) < 0) > + do_something(); > > -Alternatively, you could place a barrier() call in the loop. > +Alternatively, you could place a ``barrier()`` call in the loop. > > -For another example, consider the following code: > +For another example, consider the following code:: > > - tmp_a = a; > - do_something_with(tmp_a); > - do_something_else_with(tmp_a); > + tmp_a = a; > + do_something_with(tmp_a); > + do_something_else_with(tmp_a); > > -If the compiler can prove that do_something_with() does not store to the > -variable a, then the compiler is within its rights to manufacture an > -additional load as follows: > +If the compiler can prove that ``do_something_with()`` does not store to the > +variable ``a``, then the compiler is within its rights to manufacture an > +additional load as follows:: > > - tmp_a = a; > - do_something_with(tmp_a); > - tmp_a = a; > - do_something_else_with(tmp_a); > + tmp_a = a; > + do_something_with(tmp_a); > + tmp_a = a; > + do_something_else_with(tmp_a); > > This could fatally confuse your code if it expected the same value > -to be passed to do_something_with() and do_something_else_with(). > +to be passed to ``do_something_with()`` and ``do_something_else_with()``. > > The compiler would be likely to manufacture this additional load if > -do_something_with() was an inline function that made very heavy use > +``do_something_with()`` was an inline function that made very heavy use > of registers: reloading from variable a could save a flush to the > stack and later reload. To prevent the compiler from attacking your > -code in this manner, write the following: > +code in this manner, write the following:: > > - tmp_a = READ_ONCE(a); > - do_something_with(tmp_a); > - do_something_else_with(tmp_a); > + tmp_a = READ_ONCE(a); > + do_something_with(tmp_a); > + do_something_else_with(tmp_a); > > For a final example, consider the following code, assuming that the > -variable a is set at boot time before the second CPU is brought online > -and never changed later, so that memory barriers are not needed: > +variable ``a`` is set at boot time before the second CPU is brought online > +and never changed later, so that memory barriers are not needed:: > > - if (a) > - b = 9; > - else > - b = 42; > + if (a) > + b = 9; > + else > + b = 42; > > The compiler is within its rights to manufacture an additional store > -by transforming the above code into the following: > +by transforming the above code into the following:: > > - b = 42; > - if (a) > - b = 9; > + b = 42; > + if (a) > + b = 9; > > This could come as a fatal surprise to other code running concurrently > -that expected b to never have the value 42 if a was zero. To prevent > -the compiler from doing this, write something like: > +that expected ``b`` to never have the value ``42`` if ``a`` was zero. To > +prevent the compiler from doing this, write something like:: > + > + if (a) > + WRITE_ONCE(b, 9); > + else > + WRITE_ONCE(b, 42); > > - if (a) > - WRITE_ONCE(b, 9); > - else > - WRITE_ONCE(b, 42); > +Don't even **think** about doing this without proper use of memory barriers, > +locks, or atomic operations if variable ``a`` can change at runtime! > > -Don't even -think- about doing this without proper use of memory barriers, > -locks, or atomic operations if variable a can change at runtime! > +.. warning:: > > -*** WARNING: READ_ONCE() OR WRITE_ONCE() DO NOT IMPLY A BARRIER! *** > + ``READ_ONCE()`` OR ``WRITE_ONCE()`` DO NOT IMPLY A BARRIER! > > Now, we move onto the atomic operation interfaces typically implemented with > -the help of assembly code. > +the help of assembly code. :: > > - void atomic_add(int i, atomic_t *v); > - void atomic_sub(int i, atomic_t *v); > - void atomic_inc(atomic_t *v); > - void atomic_dec(atomic_t *v); > + void atomic_add(int i, atomic_t *v); > + void atomic_sub(int i, atomic_t *v); > + void atomic_inc(atomic_t *v); > + void atomic_dec(atomic_t *v); > > These four routines add and subtract integral values to/from the given > -atomic_t value. The first two routines pass explicit integers by > +``atomic_t`` value. The first two routines pass explicit integers by > which to make the adjustment, whereas the latter two use an implicit > adjustment value of "1". > > -One very important aspect of these two routines is that they DO NOT > +One very important aspect of these two routines is that they **DO NOT** > require any explicit memory barriers. They need only perform the > -atomic_t counter update in an SMP safe manner. > +``atomic_t`` counter update in an SMP safe manner. > > -Next, we have: > +Next, we have:: > > - int atomic_inc_return(atomic_t *v); > - int atomic_dec_return(atomic_t *v); > + int atomic_inc_return(atomic_t *v); > + int atomic_dec_return(atomic_t *v); > > These routines add 1 and subtract 1, respectively, from the given > -atomic_t and return the new counter value after the operation is > +``atomic_t`` and return the new counter value after the operation is > performed. > > Unlike the above routines, it is required that these primitives > @@ -207,317 +215,330 @@ the operation. It must be done such that all memory operations before > and after the atomic operation calls are strongly ordered with respect > to the atomic operation itself. > > -For example, it should behave as if a smp_mb() call existed both > +For example, it should behave as if a ``smp_mb()`` call existed both > before and after the atomic operation. > > If the atomic instructions used in an implementation provide explicit > memory barrier semantics which satisfy the above requirements, that is > fine as well. > > -Let's move on: > +Let's move on:: > > - int atomic_add_return(int i, atomic_t *v); > - int atomic_sub_return(int i, atomic_t *v); > + int atomic_add_return(int i, atomic_t *v); > + int atomic_sub_return(int i, atomic_t *v); > > -These behave just like atomic_{inc,dec}_return() except that an > +These behave just like ``atomic_{inc,dec}_return()`` except that an > explicit counter adjustment is given instead of the implicit "1". > -This means that like atomic_{inc,dec}_return(), the memory barrier > +This means that like ``atomic_{inc,dec}_return()``, the memory barrier > semantics are required. > > -Next: > +Next:: > > - int atomic_inc_and_test(atomic_t *v); > - int atomic_dec_and_test(atomic_t *v); > + int atomic_inc_and_test(atomic_t *v); > + int atomic_dec_and_test(atomic_t *v); > > These two routines increment and decrement by 1, respectively, the > given atomic counter. They return a boolean indicating whether the > resulting counter value was zero or not. > > Again, these primitives provide explicit memory barrier semantics around > -the atomic operation. > +the atomic operation. :: Better to remove the dot here: the atomic operation:: And the same on other places. > > - int atomic_sub_and_test(int i, atomic_t *v); > + int atomic_sub_and_test(int i, atomic_t *v); > > -This is identical to atomic_dec_and_test() except that an explicit > +This is identical to ``atomic_dec_and_test()`` except that an explicit > decrement is given instead of the implicit "1". This primitive must > -provide explicit memory barrier semantics around the operation. > +provide explicit memory barrier semantics around the operation. :: > > - int atomic_add_negative(int i, atomic_t *v); > + int atomic_add_negative(int i, atomic_t *v); > > The given increment is added to the given atomic counter value. A boolean > is return which indicates whether the resulting counter value is negative. > This primitive must provide explicit memory barrier semantics around > the operation. > > -Then: > +Then:: > > - int atomic_xchg(atomic_t *v, int new); > + int atomic_xchg(atomic_t *v, int new); > > This performs an atomic exchange operation on the atomic variable v, setting > the given new value. It returns the old value that the atomic variable v had > just before the operation. > > -atomic_xchg must provide explicit memory barriers around the operation. > +``atomic_xchg`` must provide explicit memory barriers around the operation. :: > > - int atomic_cmpxchg(atomic_t *v, int old, int new); > + int atomic_cmpxchg(atomic_t *v, int old, int new); > > This performs an atomic compare exchange operation on the atomic value v, > -with the given old and new values. Like all atomic_xxx operations, > -atomic_cmpxchg will only satisfy its atomicity semantics as long as all > -other accesses of *v are performed through atomic_xxx operations. > +with the given old and new values. Like all ``atomic_*`` operations, > +``atomic_cmpxchg`` will only satisfy its atomicity semantics as long as all > +other accesses of ``*v`` are performed through ``atomic_*`` operations. > > -atomic_cmpxchg must provide explicit memory barriers around the operation, > +``atomic_cmpxchg`` must provide explicit memory barriers around the operation, > although if the comparison fails then no memory ordering guarantees are > required. > > -The semantics for atomic_cmpxchg are the same as those defined for 'cas' > +The semantics for ``atomic_cmpxchg`` are the same as those defined for 'cas' > below. > > -Finally: > +Finally:: > > - int atomic_add_unless(atomic_t *v, int a, int u); > + int atomic_add_unless(atomic_t *v, int a, int u); > > If the atomic value v is not equal to u, this function adds a to v, and > returns non zero. If v is equal to u then it returns zero. This is done as > an atomic operation. > > -atomic_add_unless must provide explicit memory barriers around the > +``atomic_add_unless`` must provide explicit memory barriers around the > operation unless it fails (returns 0). > > -atomic_inc_not_zero, equivalent to atomic_add_unless(v, 1, 0) > +``atomic_inc_not_zero``, equivalent to ``atomic_add_unless(v, 1, 0)`` > > > -If a caller requires memory barrier semantics around an atomic_t > +If a caller requires memory barrier semantics around an ``atomic_t`` > operation which does not return a value, a set of interfaces are > -defined which accomplish this: > +defined which accomplish this:: > > - void smp_mb__before_atomic(void); > - void smp_mb__after_atomic(void); > + void smp_mb__before_atomic(void); > + void smp_mb__after_atomic(void); > > -For example, smp_mb__before_atomic() can be used like so: > +For example, ``smp_mb__before_atomic()`` can be used like so:: > > - obj->dead = 1; > - smp_mb__before_atomic(); > - atomic_dec(&obj->ref_count); > + obj->dead = 1; > + smp_mb__before_atomic(); > + atomic_dec(&obj->ref_count); > > -It makes sure that all memory operations preceding the atomic_dec() > +It makes sure that all memory operations preceding the ``atomic_dec()`` > call are strongly ordered with respect to the atomic counter > operation. In the above example, it guarantees that the assignment of > -"1" to obj->dead will be globally visible to other cpus before the > +"1" to ``obj->dead`` will be globally visible to other cpus before the > atomic counter decrement. > > -Without the explicit smp_mb__before_atomic() call, the > +Without the explicit ``smp_mb__before_atomic()`` call, the > implementation could legally allow the atomic counter update visible > -to other cpus before the "obj->dead = 1;" assignment. > +to other cpus before the ``obj->dead = 1;`` assignment. > > A missing memory barrier in the cases where they are required by the > -atomic_t implementation above can have disastrous results. Here is > +``atomic_t`` implementation above can have disastrous results. Here is > an example, which follows a pattern occurring frequently in the Linux > kernel. It is the use of atomic counters to implement reference > counting, and it works such that once the counter falls to zero it can > be guaranteed that no other entity can be accessing the object: > > -static void obj_list_add(struct obj *obj, struct list_head *head) > -{ > - obj->active = 1; > - list_add(&obj->list, head); > -} > - > -static void obj_list_del(struct obj *obj) > -{ > - list_del(&obj->list); > - obj->active = 0; > -} > - > -static void obj_destroy(struct obj *obj) > -{ > - BUG_ON(obj->active); > - kfree(obj); > -} > - > -struct obj *obj_list_peek(struct list_head *head) > -{ > - if (!list_empty(head)) { > - struct obj *obj; > - > - obj = list_entry(head->next, struct obj, list); > - atomic_inc(&obj->refcnt); > - return obj; > - } > - return NULL; > -} > - > -void obj_poke(void) > -{ > - struct obj *obj; > - > - spin_lock(&global_list_lock); > - obj = obj_list_peek(&global_list); > - spin_unlock(&global_list_lock); > - > - if (obj) { > - obj->ops->poke(obj); > - if (atomic_dec_and_test(&obj->refcnt)) > - obj_destroy(obj); > - } > -} > - > -void obj_timeout(struct obj *obj) > -{ > - spin_lock(&global_list_lock); > - obj_list_del(obj); > - spin_unlock(&global_list_lock); > - > - if (atomic_dec_and_test(&obj->refcnt)) > - obj_destroy(obj); > -} > - > -(This is a simplification of the ARP queue management in the > - generic neighbour discover code of the networking. Olaf Kirch > - found a bug wrt. memory barriers in kfree_skb() that exposed > - the atomic_t memory barrier requirements quite clearly.) > - > -Given the above scheme, it must be the case that the obj->active > +.. code-block:: c > + > + static void obj_list_add(struct obj *obj, struct list_head *head) > + { > + obj->active = 1; > + list_add(&obj->list, head); > + } > + > + static void obj_list_del(struct obj *obj) > + { > + list_del(&obj->list); > + obj->active = 0; > + } > + > + static void obj_destroy(struct obj *obj) > + { > + BUG_ON(obj->active); > + kfree(obj); > + } > + > + struct obj *obj_list_peek(struct list_head *head) > + { > + if (!list_empty(head)) { > + struct obj *obj; > + > + obj = list_entry(head->next, struct obj, list); > + atomic_inc(&obj->refcnt); > + return obj; > + } > + return NULL; > + } > + > + void obj_poke(void) > + { > + struct obj *obj; > + > + spin_lock(&global_list_lock); > + obj = obj_list_peek(&global_list); > + spin_unlock(&global_list_lock); > + > + if (obj) { > + obj->ops->poke(obj); > + if (atomic_dec_and_test(&obj->refcnt)) > + obj_destroy(obj); > + } > + } > + > + void obj_timeout(struct obj *obj) > + { > + spin_lock(&global_list_lock); > + obj_list_del(obj); > + spin_unlock(&global_list_lock); > + > + if (atomic_dec_and_test(&obj->refcnt)) > + obj_destroy(obj); > + } > + > +.. note: > + > + This is a simplification of the ARP queue management in the > + generic neighbour discover code of the networking. Olaf Kirch > + found a bug wrt. memory barriers in kfree_skb() that exposed > + the atomic_t memory barrier requirements quite clearly. > + > +Given the above scheme, it must be the case that the ``obj->active`` > update done by the obj list deletion be visible to other processors > before the atomic counter decrement is performed. > > -Otherwise, the counter could fall to zero, yet obj->active would still > -be set, thus triggering the assertion in obj_destroy(). The error > -sequence looks like this: > - > - cpu 0 cpu 1 > - obj_poke() obj_timeout() > - obj = obj_list_peek(); > - ... gains ref to obj, refcnt=2 > - obj_list_del(obj); > - obj->active = 0 ... > - ... visibility delayed ... > - atomic_dec_and_test() > - ... refcnt drops to 1 ... > - atomic_dec_and_test() > - ... refcount drops to 0 ... > - obj_destroy() > - BUG() triggers since obj->active > - still seen as one > - obj->active update visibility occurs > +Otherwise, the counter could fall to zero, yet ``obj->active`` would still > +be set, thus triggering the assertion in ``obj_destroy()``. The error > +sequence looks like this:: > + > + cpu 0 cpu 1 > + obj_poke() obj_timeout() > + obj = obj_list_peek(); > + ... gains ref to obj, refcnt=2 > + obj_list_del(obj); > + obj->active = 0 ... > + ... visibility delayed ... > + atomic_dec_and_test() > + ... refcnt drops to 1 ... > + atomic_dec_and_test() > + ... refcount drops to 0 ... > + obj_destroy() > + BUG() triggers since obj->active > + still seen as one > + obj->active update visibility occurs > > With the memory barrier semantics required of the atomic_t operations > which return values, the above sequence of memory visibility can never > -happen. Specifically, in the above case the atomic_dec_and_test() > +happen. Specifically, in the above case the ``atomic_dec_and_test()`` > counter decrement would not become globally visible until the > -obj->active update does. > +``obj->active`` update does. > > As a historical note, 32-bit Sparc used to only allow usage of > -24-bits of its atomic_t type. This was because it used 8 bits > +24-bits of its ``atomic_t`` type. This was because it used 8 bits > as a spinlock for SMP safety. Sparc32 lacked a "compare and swap" > type instruction. However, 32-bit Sparc has since been moved over > to a "hash table of spinlocks" scheme, that allows the full 32-bit > counter to be realized. Essentially, an array of spinlocks are > -indexed into based upon the address of the atomic_t being operated > +indexed into based upon the address of the ``atomic_t`` being operated > on, and that lock protects the atomic operation. Parisc uses the > same scheme. > > -Another note is that the atomic_t operations returning values are > +Another note is that the ``atomic_t`` operations returning values are > extremely slow on an old 386. > > +Atomic Bitmask > +============== > + > We will now cover the atomic bitmask operations. You will find that > their SMP and memory barrier semantics are similar in shape and scope > -to the atomic_t ops above. > +to the ``atomic_t`` ops above. > > Native atomic bit operations are defined to operate on objects aligned > -to the size of an "unsigned long" C data type, and are least of that > -size. The endianness of the bits within each "unsigned long" are the > -native endianness of the cpu. > +to the size of an ``unsigned long`` C data type, and are least of that > +size. The endianness of the bits within each ``unsigned long`` are the > +native endianness of the cpu. :: > > - void set_bit(unsigned long nr, volatile unsigned long *addr); > - void clear_bit(unsigned long nr, volatile unsigned long *addr); > - void change_bit(unsigned long nr, volatile unsigned long *addr); > + void set_bit(unsigned long nr, volatile unsigned long *addr); > + void clear_bit(unsigned long nr, volatile unsigned long *addr); > + void change_bit(unsigned long nr, volatile unsigned long *addr); > > These routines set, clear, and change, respectively, the bit number > -indicated by "nr" on the bit mask pointed to by "ADDR". > +indicated by ``nr`` on the bit mask pointed to by ``addr``. > > They must execute atomically, yet there are no implicit memory barrier > -semantics required of these interfaces. > +semantics required of these interfaces. :: > > - int test_and_set_bit(unsigned long nr, volatile unsigned long *addr); > - int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); > - int test_and_change_bit(unsigned long nr, volatile unsigned long *addr); > + int test_and_set_bit(unsigned long nr, volatile unsigned long *addr); > + int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); > + int test_and_change_bit(unsigned long nr, volatile unsigned long *addr); > > Like the above, except that these routines return a boolean which > indicates whether the changed bit was set _BEFORE_ the atomic bit > operation. > > -WARNING! It is incredibly important that the value be a boolean, > -ie. "0" or "1". Do not try to be fancy and save a few instructions by > -declaring the above to return "long" and just returning something like > -"old_val & mask" because that will not work. > +.. warning:: > > -For one thing, this return value gets truncated to int in many code > -paths using these interfaces, so on 64-bit if the bit is set in the > -upper 32-bits then testers will never see that. > + It is incredibly important that the value be a boolean, > + ie. "0" or "1". Do not try to be fancy and save a few instructions by > + declaring the above to return "long" and just returning something like > + "old_val & mask" because that will not work. > > -One great example of where this problem crops up are the thread_info > -flag operations. Routines such as test_and_set_ti_thread_flag() chop > -the return value into an int. There are other places where things > -like this occur as well. > + For one thing, this return value gets truncated to int in many code > + paths using these interfaces, so on 64-bit if the bit is set in the > + upper 32-bits then testers will never see that. > > -These routines, like the atomic_t counter operations returning values, > + One great example of where this problem crops up are the ``thread_info`` > + flag operations. Routines such as ``test_and_set_ti_thread_flag()`` chop > + the return value into an int. There are other places where things > + like this occur as well. > + > +These routines, like the ``atomic_t`` counter operations returning values, > must provide explicit memory barrier semantics around their execution. > All memory operations before the atomic bit operation call must be > made visible globally before the atomic bit operation is made visible. > Likewise, the atomic bit operation must be visible globally before any > subsequent memory operation is made visible. For example: > > - obj->dead = 1; > - if (test_and_set_bit(0, &obj->flags)) > - /* ... */; > - obj->killed = 1; > +.. code-block:: c Instead of using code-block, better to do: subsequent memory operation is made visible. For example:: > + > + obj->dead = 1; > + if (test_and_set_bit(0, &obj->flags)) > + /* ... */; > + obj->killed = 1; > > -The implementation of test_and_set_bit() must guarantee that > -"obj->dead = 1;" is visible to cpus before the atomic memory operation > -done by test_and_set_bit() becomes visible. Likewise, the atomic > -memory operation done by test_and_set_bit() must become visible before > -"obj->killed = 1;" is visible. > +The implementation of ``test_and_set_bit()`` must guarantee that > +``obj->dead = 1;`` is visible to cpus before the atomic memory operation > +done by ``test_and_set_bit()`` becomes visible. Likewise, the atomic > +memory operation done by ``test_and_set_bit()`` must become visible before > +``obj->killed = 1;`` is visible. > > -Finally there is the basic operation: > +Finally there is the basic operation:: > > - int test_bit(unsigned long nr, __const__ volatile unsigned long *addr); > + int test_bit(unsigned long nr, __const__ volatile unsigned long *addr); > > -Which returns a boolean indicating if bit "nr" is set in the bitmask > -pointed to by "addr". > +Which returns a boolean indicating if bit ``nr`` is set in the bitmask > +pointed to by ``addr``. > > -If explicit memory barriers are required around {set,clear}_bit() (which do > +If explicit memory barriers are required around ``{set,clear}_bit()`` (which do > not return a value, and thus does not need to provide memory barrier > -semantics), two interfaces are provided: > +semantics), two interfaces are provided:: > > - void smp_mb__before_atomic(void); > - void smp_mb__after_atomic(void); > + void smp_mb__before_atomic(void); > + void smp_mb__after_atomic(void); > > -They are used as follows, and are akin to their atomic_t operation > +They are used as follows, and are akin to their ``atomic_t`` operation > brothers: > > - /* All memory operations before this call will > - * be globally visible before the clear_bit(). > - */ > - smp_mb__before_atomic(); > - clear_bit( ... ); > +.. code-block:: c > > - /* The clear_bit() will be visible before all > - * subsequent memory operations. > - */ > - smp_mb__after_atomic(); > + /* All memory operations before this call will > + * be globally visible before the clear_bit(). > + */ > + smp_mb__before_atomic(); > + clear_bit( ... ); > + > + /* The clear_bit() will be visible before all > + * subsequent memory operations. > + */ > + smp_mb__after_atomic(); > > There are two special bitops with lock barrier semantics (acquire/release, > same as spinlocks). These operate in the same way as their non-_lock/unlock > postfixed variants, except that they are to provide acquire/release semantics, > -respectively. This means they can be used for bit_spin_trylock and > -bit_spin_unlock type operations without specifying any more barriers. > +respectively. This means they can be used for ``bit_spin_trylock`` and > +``bit_spin_unlock`` type operations without specifying any more barriers. :: > > - int test_and_set_bit_lock(unsigned long nr, unsigned long *addr); > - void clear_bit_unlock(unsigned long nr, unsigned long *addr); > - void __clear_bit_unlock(unsigned long nr, unsigned long *addr); > + int test_and_set_bit_lock(unsigned long nr, unsigned long *addr); > + void clear_bit_unlock(unsigned long nr, unsigned long *addr); > + void __clear_bit_unlock(unsigned long nr, unsigned long *addr); > > -The __clear_bit_unlock version is non-atomic, however it still implements > +The ``__clear_bit_unlock`` version is non-atomic, however it still implements > unlock barrier semantics. This can be useful if the lock itself is protecting > the other bits in the word. > > @@ -526,41 +547,43 @@ provided. They are used in contexts where some other higher-level SMP > locking scheme is being used to protect the bitmask, and thus less > expensive non-atomic operations may be used in the implementation. > They have names similar to the above bitmask operation interfaces, > -except that two underscores are prefixed to the interface name. > +except that two underscores are prefixed to the interface name. :: > > - void __set_bit(unsigned long nr, volatile unsigned long *addr); > - void __clear_bit(unsigned long nr, volatile unsigned long *addr); > - void __change_bit(unsigned long nr, volatile unsigned long *addr); > - int __test_and_set_bit(unsigned long nr, volatile unsigned long *addr); > - int __test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); > - int __test_and_change_bit(unsigned long nr, volatile unsigned long *addr); > + void __set_bit(unsigned long nr, volatile unsigned long *addr); > + void __clear_bit(unsigned long nr, volatile unsigned long *addr); > + void __change_bit(unsigned long nr, volatile unsigned long *addr); > + int __test_and_set_bit(unsigned long nr, volatile unsigned long *addr); > + int __test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); > + int __test_and_change_bit(unsigned long nr, volatile unsigned long *addr); > > These non-atomic variants also do not require any special memory > barrier semantics. > > -The routines xchg() and cmpxchg() must provide the same exact > +The routines ``xchg()`` and ``cmpxchg()`` must provide the same exact > memory-barrier semantics as the atomic and bit operations returning > values. > > -Note: If someone wants to use xchg(), cmpxchg() and their variants, > -linux/atomic.h should be included rather than asm/cmpxchg.h, unless > -the code is in arch/* and can take care of itself. > +.. note:: > + > + If someone wants to use ``xchg()``, ``cmpxchg()`` and their variants, > + ``linux/atomic.h`` should be included rather than ``asm/cmpxchg.h``, unless > + the code is in arch/* and can take care of itself. > > Spinlocks and rwlocks have memory barrier expectations as well. > The rule to follow is simple: > > -1) When acquiring a lock, the implementation must make it globally > +1. When acquiring a lock, the implementation must make it globally > visible before any subsequent memory operation. > > -2) When releasing a lock, the implementation must make it such that > +2. When releasing a lock, the implementation must make it such that > all previous memory operations are globally visible before the > lock release. > > -Which finally brings us to _atomic_dec_and_lock(). There is an > -architecture-neutral version implemented in lib/dec_and_lock.c, > -but most platforms will wish to optimize this in assembler. > +Which finally brings us to ``_atomic_dec_and_lock()``. There is an > +architecture-neutral version implemented in ``lib/dec_and_lock.c``, > +but most platforms will wish to optimize this in assembler. :: > > - int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock); > + int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock); > > Atomically decrement the given counter, and if will drop to zero > atomically acquire the given spinlock and perform the decrement > @@ -573,68 +596,74 @@ sure the spinlock operation is globally visible before any > subsequent memory operation. > > We can demonstrate this operation more clearly if we define > -an abstract atomic operation: > +an abstract atomic operation:: > > - long cas(long *mem, long old, long new); > + long cas(long *mem, long old, long new); > > -"cas" stands for "compare and swap". It atomically: > +"``cas``" stands for "compare and swap". It atomically: > > -1) Compares "old" with the value currently at "mem". > -2) If they are equal, "new" is written to "mem". > -3) Regardless, the current value at "mem" is returned. > +1. Compares "old" with the value currently at "mem". > +2. If they are equal, "new" is written to "mem". > +3. Regardless, the current value at "mem" is returned. > > As an example usage, here is what an atomic counter update > might look like: > > -void example_atomic_inc(long *counter) > -{ > - long old, new, ret; > +.. code-block:: c Same here. Use :: instead of code-block. > > - while (1) { > - old = *counter; > - new = old + 1; > + void example_atomic_inc(long *counter) > + { > + long old, new, ret; > > - ret = cas(counter, old, new); > - if (ret == old) > - break; > - } > -} > + while (1) { > + old = *counter; > + new = old + 1; > + > + ret = cas(counter, old, new); > + if (ret == old) > + break; > + } > + } > > Let's use cas() in order to build a pseudo-C atomic_dec_and_lock(): > > -int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock) > -{ > - long old, new, ret; > - int went_to_zero; > - > - went_to_zero = 0; > - while (1) { > - old = atomic_read(atomic); > - new = old - 1; > - if (new == 0) { > - went_to_zero = 1; > - spin_lock(lock); > - } > - ret = cas(atomic, old, new); > - if (ret == old) > - break; > - if (went_to_zero) { > - spin_unlock(lock); > - went_to_zero = 0; > - } > - } > - > - return went_to_zero; > -} > - > -Now, as far as memory barriers go, as long as spin_lock() > +.. code-block:: c And here > + > + int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock) > + { > + long old, new, ret; > + int went_to_zero; > + > + went_to_zero = 0; > + while (1) { > + old = atomic_read(atomic); > + new = old - 1; > + if (new == 0) { > + went_to_zero = 1; > + spin_lock(lock); > + } > + ret = cas(atomic, old, new); > + if (ret == old) > + break; > + if (went_to_zero) { > + spin_unlock(lock); > + went_to_zero = 0; > + } > + } > + > + return went_to_zero; > + } > + > +Now, as far as memory barriers go, as long as i``spin_lock()`` > strictly orders all subsequent memory operations (including > -the cas()) with respect to itself, things will be fine. > +the ``cas()``) with respect to itself, things will be fine. > > -Said another way, _atomic_dec_and_lock() must guarantee that > +Said another way, ``_atomic_dec_and_lock()`` must guarantee that > a counter dropping to zero is never made visible before the > spinlock being acquired. > > -Note that this also means that for the case where the counter > -is not dropping to zero, there are no memory ordering > -requirements. > +.. note:: > + > + Note that this also means that for the case where the counter > + is not dropping to zero, there are no memory ordering > + requirements. > diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst > index 480d9a3..f3e5f5e 100644 > --- a/Documentation/core-api/index.rst > +++ b/Documentation/core-api/index.rst > @@ -8,6 +8,7 @@ Kernel and driver related documentation. > :maxdepth: 1 > > assoc_array > + atomic_ops > workqueue > > .. only:: subproject > diff --git a/Documentation/process/volatile-considered-harmful.rst b/Documentation/process/volatile-considered-harmful.rst > index e0d042a..4934e65 100644 > --- a/Documentation/process/volatile-considered-harmful.rst > +++ b/Documentation/process/volatile-considered-harmful.rst > @@ -1,3 +1,6 @@ > + > +.. _volatile_considered_harmful: > + > Why the "volatile" type class should not be used > ------------------------------------------------ > Thanks, Mauro ^ permalink raw reply [flat|nested] 51+ messages in thread
* [PATCH v2 3/4] Documentation/local_ops.txt: convert to ReST markup 2016-11-25 10:02 ` [PATCH v2 " Silvio Fricke 2016-11-25 10:02 ` [PATCH v2 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke 2016-11-25 10:02 ` [PATCH v2 2/4] Documentation/atomic_ops.txt: " Silvio Fricke @ 2016-11-25 10:02 ` Silvio Fricke 2016-11-25 13:34 ` Mauro Carvalho Chehab 2016-11-25 10:02 ` [PATCH v2 4/4] firmware: remove warning at documentation generation time Silvio Fricke 2016-11-25 14:59 ` [PATCH v3 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 4 siblings, 1 reply; 51+ messages in thread From: Silvio Fricke @ 2016-11-25 10:02 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Silvio Fricke ... and move to core-api folder. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- Documentation/core-api/index.rst | 1 +- Documentation/local_ops.txt => Documentation/core-api/local_ops.rst | 275 +++---- 2 files changed, 147 insertions(+), 129 deletions(-) diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index f3e5f5e..25b4e4a 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -9,6 +9,7 @@ Kernel and driver related documentation. assoc_array atomic_ops + local_ops workqueue .. only:: subproject diff --git a/Documentation/local_ops.txt b/Documentation/core-api/local_ops.rst similarity index 55% rename from Documentation/local_ops.txt rename to Documentation/core-api/local_ops.rst index 407576a..01f1880 100644 --- a/Documentation/local_ops.txt +++ b/Documentation/core-api/local_ops.rst @@ -1,191 +1,208 @@ - Semantics and Behavior of Local Atomic Operations - Mathieu Desnoyers +.. _local_ops: +================================================= +Semantics and Behavior of Local Atomic Operations +================================================= - This document explains the purpose of the local atomic operations, how +:Author: Mathieu Desnoyers + + +This document explains the purpose of the local atomic operations, how to implement them for any given architecture and shows how they can be used properly. It also stresses on the precautions that must be taken when reading those local variables across CPUs when the order of memory writes matters. -Note that local_t based operations are not recommended for general kernel use. -Please use the this_cpu operations instead unless there is really a special purpose. -Most uses of local_t in the kernel have been replaced by this_cpu operations. -this_cpu operations combine the relocation with the local_t like semantics in -a single instruction and yield more compact and faster executing code. +.. note:: + Note that ``local_t`` based operations are not recommended for general + kernel use. Please use the ``this_cpu`` operations instead unless there is + really a special purpose. Most uses of ``local_t`` in the kernel have been + replaced by ``this_cpu`` operations. ``this_cpu`` operations combine the + relocation with the ``local_t`` like semantics in a single instruction and + yield more compact and faster executing code. -* Purpose of local atomic operations + +Purpose of local atomic operations +================================== Local atomic operations are meant to provide fast and highly reentrant per CPU counters. They minimize the performance cost of standard atomic operations by removing the LOCK prefix and memory barriers normally required to synchronize across CPUs. -Having fast per CPU atomic counters is interesting in many cases : it does not +Having fast per CPU atomic counters is interesting in many cases: it does not require disabling interrupts to protect from interrupt handlers and it permits coherent counters in NMI handlers. It is especially useful for tracing purposes and for various performance monitoring counters. Local atomic operations only guarantee variable modification atomicity wrt the CPU which owns the data. Therefore, care must taken to make sure that only one -CPU writes to the local_t data. This is done by using per cpu data and making -sure that we modify it from within a preemption safe context. It is however -permitted to read local_t data from any CPU : it will then appear to be written -out of order wrt other memory writes by the owner CPU. +CPU writes to the ``local_t`` data. This is done by using per cpu data and +making sure that we modify it from within a preemption safe context. It is +however permitted to read ``local_t`` data from any CPU: it will then appear to +be written out of order wrt other memory writes by the owner CPU. -* Implementation for a given architecture +Implementation for a given architecture +======================================= -It can be done by slightly modifying the standard atomic operations : only +It can be done by slightly modifying the standard atomic operations: only their UP variant must be kept. It typically means removing LOCK prefix (on i386 and x86_64) and any SMP synchronization barrier. If the architecture does -not have a different behavior between SMP and UP, including asm-generic/local.h -in your architecture's local.h is sufficient. +not have a different behavior between SMP and UP, including +``asm-generic/local.h`` in your architecture's ``local.h`` is sufficient. -The local_t type is defined as an opaque signed long by embedding an -atomic_long_t inside a structure. This is made so a cast from this type to a -long fails. The definition looks like : +The ``local_t`` type is defined as an opaque ``signed long`` by embedding an +``atomic_long_t`` inside a structure. This is made so a cast from this type to +a ``long`` fails. The definition looks like:: -typedef struct { atomic_long_t a; } local_t; + typedef struct { atomic_long_t a; } local_t; -* Rules to follow when using local atomic operations +Rules to follow when using local atomic operations +================================================== -- Variables touched by local ops must be per cpu variables. -- _Only_ the CPU owner of these variables must write to them. -- This CPU can use local ops from any context (process, irq, softirq, nmi, ...) - to update its local_t variables. -- Preemption (or interrupts) must be disabled when using local ops in - process context to make sure the process won't be migrated to a +* Variables touched by local ops must be per cpu variables. +* *Only* the CPU owner of these variables must write to them. +* This CPU can use local ops from any context (process, irq, softirq, nmi, ...) + to update its ``local_t`` variables. +* Preemption (or interrupts) must be disabled when using local ops in + process context to make sure the process won't be migrated to a different CPU between getting the per-cpu variable and doing the actual local op. -- When using local ops in interrupt context, no special care must be +* When using local ops in interrupt context, no special care must be taken on a mainline kernel, since they will run on the local CPU with preemption already disabled. I suggest, however, to explicitly disable preemption anyway to make sure it will still work correctly on -rt kernels. -- Reading the local cpu variable will provide the current copy of the +* Reading the local cpu variable will provide the current copy of the variable. -- Reads of these variables can be done from any CPU, because updates to - "long", aligned, variables are always atomic. Since no memory +* Reads of these variables can be done from any CPU, because updates to + "``long``", aligned, variables are always atomic. Since no memory synchronization is done by the writer CPU, an outdated copy of the - variable can be read when reading some _other_ cpu's variables. + variable can be read when reading some *other* cpu's variables. + +How to use local atomic operations +================================== -* How to use local atomic operations +.. code-block:: c -#include <linux/percpu.h> -#include <asm/local.h> + #include <linux/percpu.h> + #include <asm/local.h> -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); -* Counting +Counting +======== Counting is done on all the bits of a signed long. -In preemptible context, use get_cpu_var() and put_cpu_var() around local atomic -operations : it makes sure that preemption is disabled around write access to -the per cpu variable. For instance : +In preemptible context, use ``get_cpu_var()`` and ``put_cpu_var()`` around +local atomic operations: it makes sure that preemption is disabled around write +access to the per cpu variable. For instance:: - local_inc(&get_cpu_var(counters)); - put_cpu_var(counters); + local_inc(&get_cpu_var(counters)); + put_cpu_var(counters); If you are already in a preemption-safe context, you can use -this_cpu_ptr() instead. +``this_cpu_ptr()`` instead. :: - local_inc(this_cpu_ptr(&counters)); + local_inc(this_cpu_ptr(&counters)); -* Reading the counters +Reading the counters +==================== Those local counters can be read from foreign CPUs to sum the count. Note that the data seen by local_read across CPUs must be considered to be out of order -relatively to other memory writes happening on the CPU that owns the data. +relatively to other memory writes happening on the CPU that owns the data. :: - long sum = 0; - for_each_online_cpu(cpu) - sum += local_read(&per_cpu(counters, cpu)); + long sum = 0; + for_each_online_cpu(cpu) + sum += local_read(&per_cpu(counters, cpu)); If you want to use a remote local_read to synchronize access to a resource -between CPUs, explicit smp_wmb() and smp_rmb() memory barriers must be used +between CPUs, explicit ``smp_wmb()`` and ``smp_rmb()`` memory barriers must be used respectively on the writer and the reader CPUs. It would be the case if you use -the local_t variable as a counter of bytes written in a buffer : there should -be a smp_wmb() between the buffer write and the counter increment and also a -smp_rmb() between the counter read and the buffer read. - - -Here is a sample module which implements a basic per cpu counter using local.h. - ---- BEGIN --- -/* test-local.c - * - * Sample module for local.h usage. - */ - - -#include <asm/local.h> -#include <linux/module.h> -#include <linux/timer.h> - -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); - -static struct timer_list test_timer; - -/* IPI called on each CPU. */ -static void test_each(void *info) -{ - /* Increment the counter from a non preemptible context */ - printk("Increment on cpu %d\n", smp_processor_id()); - local_inc(this_cpu_ptr(&counters)); - - /* This is what incrementing the variable would look like within a - * preemptible context (it disables preemption) : - * - * local_inc(&get_cpu_var(counters)); - * put_cpu_var(counters); - */ -} - -static void do_test_timer(unsigned long data) -{ - int cpu; - - /* Increment the counters */ - on_each_cpu(test_each, NULL, 1); - /* Read all the counters */ - printk("Counters read from CPU %d\n", smp_processor_id()); - for_each_online_cpu(cpu) { - printk("Read : CPU %d, count %ld\n", cpu, - local_read(&per_cpu(counters, cpu))); - } - del_timer(&test_timer); - test_timer.expires = jiffies + 1000; - add_timer(&test_timer); -} - -static int __init test_init(void) -{ - /* initialize the timer that will increment the counter */ - init_timer(&test_timer); - test_timer.function = do_test_timer; - test_timer.expires = jiffies + 1; - add_timer(&test_timer); - - return 0; -} - -static void __exit test_exit(void) -{ - del_timer_sync(&test_timer); -} - -module_init(test_init); -module_exit(test_exit); - -MODULE_LICENSE("GPL"); -MODULE_AUTHOR("Mathieu Desnoyers"); -MODULE_DESCRIPTION("Local Atomic Ops"); ---- END --- +the ``local_t`` variable as a counter of bytes written in a buffer: there should +be a ``smp_wmb()`` between the buffer write and the counter increment and also a +``smp_rmb()`` between the counter read and the buffer read. + + +Here is a sample module which implements a basic per cpu counter using +``local.h``. + +.. code-block:: c + + /* test-local.c + * + * Sample module for local.h usage. + */ + + + #include <asm/local.h> + #include <linux/module.h> + #include <linux/timer.h> + + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); + + static struct timer_list test_timer; + + /* IPI called on each CPU. */ + static void test_each(void *info) + { + /* Increment the counter from a non preemptible context */ + printk("Increment on cpu %d\n", smp_processor_id()); + local_inc(this_cpu_ptr(&counters)); + + /* This is what incrementing the variable would look like within a + * preemptible context (it disables preemption) : + * + * local_inc(&get_cpu_var(counters)); + * put_cpu_var(counters); + */ + } + + static void do_test_timer(unsigned long data) + { + int cpu; + + /* Increment the counters */ + on_each_cpu(test_each, NULL, 1); + /* Read all the counters */ + printk("Counters read from CPU %d\n", smp_processor_id()); + for_each_online_cpu(cpu) { + printk("Read : CPU %d, count %ld\n", cpu, + local_read(&per_cpu(counters, cpu))); + } + del_timer(&test_timer); + test_timer.expires = jiffies + 1000; + add_timer(&test_timer); + } + + static int __init test_init(void) + { + /* initialize the timer that will increment the counter */ + init_timer(&test_timer); + test_timer.function = do_test_timer; + test_timer.expires = jiffies + 1; + add_timer(&test_timer); + + return 0; + } + + static void __exit test_exit(void) + { + del_timer_sync(&test_timer); + } + + module_init(test_init); + module_exit(test_exit); + + MODULE_LICENSE("GPL"); + MODULE_AUTHOR("Mathieu Desnoyers"); + MODULE_DESCRIPTION("Local Atomic Ops"); -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* Re: [PATCH v2 3/4] Documentation/local_ops.txt: convert to ReST markup 2016-11-25 10:02 ` [PATCH v2 3/4] Documentation/local_ops.txt: " Silvio Fricke @ 2016-11-25 13:34 ` Mauro Carvalho Chehab 2016-11-25 13:47 ` S. Fricke 0 siblings, 1 reply; 51+ messages in thread From: Mauro Carvalho Chehab @ 2016-11-25 13:34 UTC (permalink / raw) To: Silvio Fricke Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel Em Fri, 25 Nov 2016 11:02:41 +0100 Silvio Fricke <silvio.fricke@gmail.com> escreveu: > ... and move to core-api folder. > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> > --- > Documentation/core-api/index.rst | 1 +- > Documentation/local_ops.txt => Documentation/core-api/local_ops.rst | 275 +++---- > 2 files changed, 147 insertions(+), 129 deletions(-) > > diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst > index f3e5f5e..25b4e4a 100644 > --- a/Documentation/core-api/index.rst > +++ b/Documentation/core-api/index.rst > @@ -9,6 +9,7 @@ Kernel and driver related documentation. > > assoc_array > atomic_ops > + local_ops > workqueue > > .. only:: subproject > diff --git a/Documentation/local_ops.txt b/Documentation/core-api/local_ops.rst > similarity index 55% > rename from Documentation/local_ops.txt > rename to Documentation/core-api/local_ops.rst > index 407576a..01f1880 100644 > --- a/Documentation/local_ops.txt > +++ b/Documentation/core-api/local_ops.rst > @@ -1,191 +1,208 @@ > - Semantics and Behavior of Local Atomic Operations > > - Mathieu Desnoyers > +.. _local_ops: > > +================================================= > +Semantics and Behavior of Local Atomic Operations > +================================================= > > - This document explains the purpose of the local atomic operations, how > +:Author: Mathieu Desnoyers > + > + > +This document explains the purpose of the local atomic operations, how > to implement them for any given architecture and shows how they can be used > properly. It also stresses on the precautions that must be taken when reading > those local variables across CPUs when the order of memory writes matters. > > -Note that local_t based operations are not recommended for general kernel use. > -Please use the this_cpu operations instead unless there is really a special purpose. > -Most uses of local_t in the kernel have been replaced by this_cpu operations. > -this_cpu operations combine the relocation with the local_t like semantics in > -a single instruction and yield more compact and faster executing code. > +.. note:: > > + Note that ``local_t`` based operations are not recommended for general > + kernel use. Please use the ``this_cpu`` operations instead unless there is > + really a special purpose. Most uses of ``local_t`` in the kernel have been > + replaced by ``this_cpu`` operations. ``this_cpu`` operations combine the > + relocation with the ``local_t`` like semantics in a single instruction and > + yield more compact and faster executing code. > > -* Purpose of local atomic operations > + > +Purpose of local atomic operations > +================================== > > Local atomic operations are meant to provide fast and highly reentrant per CPU > counters. They minimize the performance cost of standard atomic operations by > removing the LOCK prefix and memory barriers normally required to synchronize > across CPUs. > > -Having fast per CPU atomic counters is interesting in many cases : it does not > +Having fast per CPU atomic counters is interesting in many cases: it does not > require disabling interrupts to protect from interrupt handlers and it permits > coherent counters in NMI handlers. It is especially useful for tracing purposes > and for various performance monitoring counters. > > Local atomic operations only guarantee variable modification atomicity wrt the > CPU which owns the data. Therefore, care must taken to make sure that only one > -CPU writes to the local_t data. This is done by using per cpu data and making > -sure that we modify it from within a preemption safe context. It is however > -permitted to read local_t data from any CPU : it will then appear to be written > -out of order wrt other memory writes by the owner CPU. > +CPU writes to the ``local_t`` data. This is done by using per cpu data and > +making sure that we modify it from within a preemption safe context. It is > +however permitted to read ``local_t`` data from any CPU: it will then appear to > +be written out of order wrt other memory writes by the owner CPU. > > > -* Implementation for a given architecture > +Implementation for a given architecture > +======================================= > > -It can be done by slightly modifying the standard atomic operations : only > +It can be done by slightly modifying the standard atomic operations: only > their UP variant must be kept. It typically means removing LOCK prefix (on > i386 and x86_64) and any SMP synchronization barrier. If the architecture does > -not have a different behavior between SMP and UP, including asm-generic/local.h > -in your architecture's local.h is sufficient. > +not have a different behavior between SMP and UP, including > +``asm-generic/local.h`` in your architecture's ``local.h`` is sufficient. > > -The local_t type is defined as an opaque signed long by embedding an > -atomic_long_t inside a structure. This is made so a cast from this type to a > -long fails. The definition looks like : > +The ``local_t`` type is defined as an opaque ``signed long`` by embedding an > +``atomic_long_t`` inside a structure. This is made so a cast from this type to > +a ``long`` fails. The definition looks like:: > > -typedef struct { atomic_long_t a; } local_t; > + typedef struct { atomic_long_t a; } local_t; > > > -* Rules to follow when using local atomic operations > +Rules to follow when using local atomic operations > +================================================== > > -- Variables touched by local ops must be per cpu variables. > -- _Only_ the CPU owner of these variables must write to them. > -- This CPU can use local ops from any context (process, irq, softirq, nmi, ...) > - to update its local_t variables. > -- Preemption (or interrupts) must be disabled when using local ops in > - process context to make sure the process won't be migrated to a > +* Variables touched by local ops must be per cpu variables. > +* *Only* the CPU owner of these variables must write to them. > +* This CPU can use local ops from any context (process, irq, softirq, nmi, ...) > + to update its ``local_t`` variables. > +* Preemption (or interrupts) must be disabled when using local ops in > + process context to make sure the process won't be migrated to a > different CPU between getting the per-cpu variable and doing the > actual local op. > -- When using local ops in interrupt context, no special care must be > +* When using local ops in interrupt context, no special care must be > taken on a mainline kernel, since they will run on the local CPU with > preemption already disabled. I suggest, however, to explicitly > disable preemption anyway to make sure it will still work correctly on > -rt kernels. > -- Reading the local cpu variable will provide the current copy of the > +* Reading the local cpu variable will provide the current copy of the > variable. > -- Reads of these variables can be done from any CPU, because updates to > - "long", aligned, variables are always atomic. Since no memory > +* Reads of these variables can be done from any CPU, because updates to > + "``long``", aligned, variables are always atomic. Since no memory > synchronization is done by the writer CPU, an outdated copy of the > - variable can be read when reading some _other_ cpu's variables. > + variable can be read when reading some *other* cpu's variables. > + > > +How to use local atomic operations > +================================== > > -* How to use local atomic operations > +.. code-block:: c Better to use :: instead of code-block. > > -#include <linux/percpu.h> > -#include <asm/local.h> > + #include <linux/percpu.h> > + #include <asm/local.h> > > -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > > > -* Counting > +Counting > +======== > > Counting is done on all the bits of a signed long. > > -In preemptible context, use get_cpu_var() and put_cpu_var() around local atomic > -operations : it makes sure that preemption is disabled around write access to > -the per cpu variable. For instance : > +In preemptible context, use ``get_cpu_var()`` and ``put_cpu_var()`` around > +local atomic operations: it makes sure that preemption is disabled around write > +access to the per cpu variable. For instance:: > > - local_inc(&get_cpu_var(counters)); > - put_cpu_var(counters); > + local_inc(&get_cpu_var(counters)); > + put_cpu_var(counters); > > If you are already in a preemption-safe context, you can use > -this_cpu_ptr() instead. > +``this_cpu_ptr()`` instead. :: > > - local_inc(this_cpu_ptr(&counters)); > + local_inc(this_cpu_ptr(&counters)); > > > > -* Reading the counters > +Reading the counters > +==================== > > Those local counters can be read from foreign CPUs to sum the count. Note that > the data seen by local_read across CPUs must be considered to be out of order > -relatively to other memory writes happening on the CPU that owns the data. > +relatively to other memory writes happening on the CPU that owns the data. :: Please remove the dot at the end. > > - long sum = 0; > - for_each_online_cpu(cpu) > - sum += local_read(&per_cpu(counters, cpu)); > + long sum = 0; > + for_each_online_cpu(cpu) > + sum += local_read(&per_cpu(counters, cpu)); > > If you want to use a remote local_read to synchronize access to a resource > -between CPUs, explicit smp_wmb() and smp_rmb() memory barriers must be used > +between CPUs, explicit ``smp_wmb()`` and ``smp_rmb()`` memory barriers must be used > respectively on the writer and the reader CPUs. It would be the case if you use > -the local_t variable as a counter of bytes written in a buffer : there should > -be a smp_wmb() between the buffer write and the counter increment and also a > -smp_rmb() between the counter read and the buffer read. > - > - > -Here is a sample module which implements a basic per cpu counter using local.h. > - > ---- BEGIN --- > -/* test-local.c > - * > - * Sample module for local.h usage. > - */ > - > - > -#include <asm/local.h> > -#include <linux/module.h> > -#include <linux/timer.h> > - > -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > - > -static struct timer_list test_timer; > - > -/* IPI called on each CPU. */ > -static void test_each(void *info) > -{ > - /* Increment the counter from a non preemptible context */ > - printk("Increment on cpu %d\n", smp_processor_id()); > - local_inc(this_cpu_ptr(&counters)); > - > - /* This is what incrementing the variable would look like within a > - * preemptible context (it disables preemption) : > - * > - * local_inc(&get_cpu_var(counters)); > - * put_cpu_var(counters); > - */ > -} > - > -static void do_test_timer(unsigned long data) > -{ > - int cpu; > - > - /* Increment the counters */ > - on_each_cpu(test_each, NULL, 1); > - /* Read all the counters */ > - printk("Counters read from CPU %d\n", smp_processor_id()); > - for_each_online_cpu(cpu) { > - printk("Read : CPU %d, count %ld\n", cpu, > - local_read(&per_cpu(counters, cpu))); > - } > - del_timer(&test_timer); > - test_timer.expires = jiffies + 1000; > - add_timer(&test_timer); > -} > - > -static int __init test_init(void) > -{ > - /* initialize the timer that will increment the counter */ > - init_timer(&test_timer); > - test_timer.function = do_test_timer; > - test_timer.expires = jiffies + 1; > - add_timer(&test_timer); > - > - return 0; > -} > - > -static void __exit test_exit(void) > -{ > - del_timer_sync(&test_timer); > -} > - > -module_init(test_init); > -module_exit(test_exit); > - > -MODULE_LICENSE("GPL"); > -MODULE_AUTHOR("Mathieu Desnoyers"); > -MODULE_DESCRIPTION("Local Atomic Ops"); > ---- END --- > +the ``local_t`` variable as a counter of bytes written in a buffer: there should > +be a ``smp_wmb()`` between the buffer write and the counter increment and also a > +``smp_rmb()`` between the counter read and the buffer read. > + > + > +Here is a sample module which implements a basic per cpu counter using > +``local.h``. > + > +.. code-block:: c Better to use :: instead of code-block. > + > + /* test-local.c > + * > + * Sample module for local.h usage. > + */ > + > + > + #include <asm/local.h> > + #include <linux/module.h> > + #include <linux/timer.h> > + > + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > + > + static struct timer_list test_timer; > + > + /* IPI called on each CPU. */ > + static void test_each(void *info) > + { > + /* Increment the counter from a non preemptible context */ > + printk("Increment on cpu %d\n", smp_processor_id()); > + local_inc(this_cpu_ptr(&counters)); > + > + /* This is what incrementing the variable would look like within a > + * preemptible context (it disables preemption) : > + * > + * local_inc(&get_cpu_var(counters)); > + * put_cpu_var(counters); > + */ > + } > + > + static void do_test_timer(unsigned long data) > + { > + int cpu; > + > + /* Increment the counters */ > + on_each_cpu(test_each, NULL, 1); > + /* Read all the counters */ > + printk("Counters read from CPU %d\n", smp_processor_id()); > + for_each_online_cpu(cpu) { > + printk("Read : CPU %d, count %ld\n", cpu, > + local_read(&per_cpu(counters, cpu))); > + } > + del_timer(&test_timer); > + test_timer.expires = jiffies + 1000; > + add_timer(&test_timer); > + } > + > + static int __init test_init(void) > + { > + /* initialize the timer that will increment the counter */ > + init_timer(&test_timer); > + test_timer.function = do_test_timer; > + test_timer.expires = jiffies + 1; > + add_timer(&test_timer); > + > + return 0; > + } > + > + static void __exit test_exit(void) > + { > + del_timer_sync(&test_timer); > + } > + > + module_init(test_init); > + module_exit(test_exit); > + > + MODULE_LICENSE("GPL"); > + MODULE_AUTHOR("Mathieu Desnoyers"); > + MODULE_DESCRIPTION("Local Atomic Ops"); Thanks, Mauro ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v2 3/4] Documentation/local_ops.txt: convert to ReST markup 2016-11-25 13:34 ` Mauro Carvalho Chehab @ 2016-11-25 13:47 ` S. Fricke 2016-11-25 13:58 ` Mauro Carvalho Chehab 0 siblings, 1 reply; 51+ messages in thread From: S. Fricke @ 2016-11-25 13:47 UTC (permalink / raw) To: Mauro Carvalho Chehab Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel Hi Mauro, I have a question about the "code-block" and "::". On which situation should I use "code-block" and on which condition a "::"? For now I have used "::" on small one, two or three liners, and "code-block" for "example code" snippets or longer code segments. Thanks for a small clarification, Silvio > Em Fri, 25 Nov 2016 11:02:41 +0100 > Silvio Fricke <silvio.fricke@gmail.com> escreveu: > > > ... and move to core-api folder. > > > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> > > --- > > Documentation/core-api/index.rst | 1 +- > > Documentation/local_ops.txt => Documentation/core-api/local_ops.rst | 275 +++---- > > 2 files changed, 147 insertions(+), 129 deletions(-) > > > > diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst > > index f3e5f5e..25b4e4a 100644 > > --- a/Documentation/core-api/index.rst > > +++ b/Documentation/core-api/index.rst > > @@ -9,6 +9,7 @@ Kernel and driver related documentation. > > > > assoc_array > > atomic_ops > > + local_ops > > workqueue > > > > .. only:: subproject > > diff --git a/Documentation/local_ops.txt b/Documentation/core-api/local_ops.rst > > similarity index 55% > > rename from Documentation/local_ops.txt > > rename to Documentation/core-api/local_ops.rst > > index 407576a..01f1880 100644 > > --- a/Documentation/local_ops.txt > > +++ b/Documentation/core-api/local_ops.rst > > @@ -1,191 +1,208 @@ > > - Semantics and Behavior of Local Atomic Operations > > > > - Mathieu Desnoyers > > +.. _local_ops: > > > > +================================================= > > +Semantics and Behavior of Local Atomic Operations > > +================================================= > > > > - This document explains the purpose of the local atomic operations, how > > +:Author: Mathieu Desnoyers > > + > > + > > +This document explains the purpose of the local atomic operations, how > > to implement them for any given architecture and shows how they can be used > > properly. It also stresses on the precautions that must be taken when reading > > those local variables across CPUs when the order of memory writes matters. > > > > -Note that local_t based operations are not recommended for general kernel use. > > -Please use the this_cpu operations instead unless there is really a special purpose. > > -Most uses of local_t in the kernel have been replaced by this_cpu operations. > > -this_cpu operations combine the relocation with the local_t like semantics in > > -a single instruction and yield more compact and faster executing code. > > +.. note:: > > > > + Note that ``local_t`` based operations are not recommended for general > > + kernel use. Please use the ``this_cpu`` operations instead unless there is > > + really a special purpose. Most uses of ``local_t`` in the kernel have been > > + replaced by ``this_cpu`` operations. ``this_cpu`` operations combine the > > + relocation with the ``local_t`` like semantics in a single instruction and > > + yield more compact and faster executing code. > > > > -* Purpose of local atomic operations > > + > > +Purpose of local atomic operations > > +================================== > > > > Local atomic operations are meant to provide fast and highly reentrant per CPU > > counters. They minimize the performance cost of standard atomic operations by > > removing the LOCK prefix and memory barriers normally required to synchronize > > across CPUs. > > > > -Having fast per CPU atomic counters is interesting in many cases : it does not > > +Having fast per CPU atomic counters is interesting in many cases: it does not > > require disabling interrupts to protect from interrupt handlers and it permits > > coherent counters in NMI handlers. It is especially useful for tracing purposes > > and for various performance monitoring counters. > > > > Local atomic operations only guarantee variable modification atomicity wrt the > > CPU which owns the data. Therefore, care must taken to make sure that only one > > -CPU writes to the local_t data. This is done by using per cpu data and making > > -sure that we modify it from within a preemption safe context. It is however > > -permitted to read local_t data from any CPU : it will then appear to be written > > -out of order wrt other memory writes by the owner CPU. > > +CPU writes to the ``local_t`` data. This is done by using per cpu data and > > +making sure that we modify it from within a preemption safe context. It is > > +however permitted to read ``local_t`` data from any CPU: it will then appear to > > +be written out of order wrt other memory writes by the owner CPU. > > > > > > -* Implementation for a given architecture > > +Implementation for a given architecture > > +======================================= > > > > -It can be done by slightly modifying the standard atomic operations : only > > +It can be done by slightly modifying the standard atomic operations: only > > their UP variant must be kept. It typically means removing LOCK prefix (on > > i386 and x86_64) and any SMP synchronization barrier. If the architecture does > > -not have a different behavior between SMP and UP, including asm-generic/local.h > > -in your architecture's local.h is sufficient. > > +not have a different behavior between SMP and UP, including > > +``asm-generic/local.h`` in your architecture's ``local.h`` is sufficient. > > > > -The local_t type is defined as an opaque signed long by embedding an > > -atomic_long_t inside a structure. This is made so a cast from this type to a > > -long fails. The definition looks like : > > +The ``local_t`` type is defined as an opaque ``signed long`` by embedding an > > +``atomic_long_t`` inside a structure. This is made so a cast from this type to > > +a ``long`` fails. The definition looks like:: > > > > -typedef struct { atomic_long_t a; } local_t; > > + typedef struct { atomic_long_t a; } local_t; > > > > > > -* Rules to follow when using local atomic operations > > +Rules to follow when using local atomic operations > > +================================================== > > > > -- Variables touched by local ops must be per cpu variables. > > -- _Only_ the CPU owner of these variables must write to them. > > -- This CPU can use local ops from any context (process, irq, softirq, nmi, ...) > > - to update its local_t variables. > > -- Preemption (or interrupts) must be disabled when using local ops in > > - process context to make sure the process won't be migrated to a > > +* Variables touched by local ops must be per cpu variables. > > +* *Only* the CPU owner of these variables must write to them. > > +* This CPU can use local ops from any context (process, irq, softirq, nmi, ...) > > + to update its ``local_t`` variables. > > +* Preemption (or interrupts) must be disabled when using local ops in > > + process context to make sure the process won't be migrated to a > > different CPU between getting the per-cpu variable and doing the > > actual local op. > > -- When using local ops in interrupt context, no special care must be > > +* When using local ops in interrupt context, no special care must be > > taken on a mainline kernel, since they will run on the local CPU with > > preemption already disabled. I suggest, however, to explicitly > > disable preemption anyway to make sure it will still work correctly on > > -rt kernels. > > -- Reading the local cpu variable will provide the current copy of the > > +* Reading the local cpu variable will provide the current copy of the > > variable. > > -- Reads of these variables can be done from any CPU, because updates to > > - "long", aligned, variables are always atomic. Since no memory > > +* Reads of these variables can be done from any CPU, because updates to > > + "``long``", aligned, variables are always atomic. Since no memory > > synchronization is done by the writer CPU, an outdated copy of the > > - variable can be read when reading some _other_ cpu's variables. > > + variable can be read when reading some *other* cpu's variables. > > + > > > > +How to use local atomic operations > > +================================== > > > > -* How to use local atomic operations > > +.. code-block:: c > > Better to use :: instead of code-block. > > > > > -#include <linux/percpu.h> > > -#include <asm/local.h> > > + #include <linux/percpu.h> > > + #include <asm/local.h> > > > > -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > > + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > > > > > > -* Counting > > +Counting > > +======== > > > > Counting is done on all the bits of a signed long. > > > > -In preemptible context, use get_cpu_var() and put_cpu_var() around local atomic > > -operations : it makes sure that preemption is disabled around write access to > > -the per cpu variable. For instance : > > +In preemptible context, use ``get_cpu_var()`` and ``put_cpu_var()`` around > > +local atomic operations: it makes sure that preemption is disabled around write > > +access to the per cpu variable. For instance:: > > > > - local_inc(&get_cpu_var(counters)); > > - put_cpu_var(counters); > > + local_inc(&get_cpu_var(counters)); > > + put_cpu_var(counters); > > > > If you are already in a preemption-safe context, you can use > > -this_cpu_ptr() instead. > > +``this_cpu_ptr()`` instead. :: > > > > - local_inc(this_cpu_ptr(&counters)); > > + local_inc(this_cpu_ptr(&counters)); > > > > > > > > -* Reading the counters > > +Reading the counters > > +==================== > > > > Those local counters can be read from foreign CPUs to sum the count. Note that > > the data seen by local_read across CPUs must be considered to be out of order > > -relatively to other memory writes happening on the CPU that owns the data. > > +relatively to other memory writes happening on the CPU that owns the data. :: > > Please remove the dot at the end. > > > > > - long sum = 0; > > - for_each_online_cpu(cpu) > > - sum += local_read(&per_cpu(counters, cpu)); > > + long sum = 0; > > + for_each_online_cpu(cpu) > > + sum += local_read(&per_cpu(counters, cpu)); > > > > If you want to use a remote local_read to synchronize access to a resource > > -between CPUs, explicit smp_wmb() and smp_rmb() memory barriers must be used > > +between CPUs, explicit ``smp_wmb()`` and ``smp_rmb()`` memory barriers must be used > > respectively on the writer and the reader CPUs. It would be the case if you use > > -the local_t variable as a counter of bytes written in a buffer : there should > > -be a smp_wmb() between the buffer write and the counter increment and also a > > -smp_rmb() between the counter read and the buffer read. > > - > > - > > -Here is a sample module which implements a basic per cpu counter using local.h. > > - > > ---- BEGIN --- > > -/* test-local.c > > - * > > - * Sample module for local.h usage. > > - */ > > - > > - > > -#include <asm/local.h> > > -#include <linux/module.h> > > -#include <linux/timer.h> > > - > > -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > > - > > -static struct timer_list test_timer; > > - > > -/* IPI called on each CPU. */ > > -static void test_each(void *info) > > -{ > > - /* Increment the counter from a non preemptible context */ > > - printk("Increment on cpu %d\n", smp_processor_id()); > > - local_inc(this_cpu_ptr(&counters)); > > - > > - /* This is what incrementing the variable would look like within a > > - * preemptible context (it disables preemption) : > > - * > > - * local_inc(&get_cpu_var(counters)); > > - * put_cpu_var(counters); > > - */ > > -} > > - > > -static void do_test_timer(unsigned long data) > > -{ > > - int cpu; > > - > > - /* Increment the counters */ > > - on_each_cpu(test_each, NULL, 1); > > - /* Read all the counters */ > > - printk("Counters read from CPU %d\n", smp_processor_id()); > > - for_each_online_cpu(cpu) { > > - printk("Read : CPU %d, count %ld\n", cpu, > > - local_read(&per_cpu(counters, cpu))); > > - } > > - del_timer(&test_timer); > > - test_timer.expires = jiffies + 1000; > > - add_timer(&test_timer); > > -} > > - > > -static int __init test_init(void) > > -{ > > - /* initialize the timer that will increment the counter */ > > - init_timer(&test_timer); > > - test_timer.function = do_test_timer; > > - test_timer.expires = jiffies + 1; > > - add_timer(&test_timer); > > - > > - return 0; > > -} > > - > > -static void __exit test_exit(void) > > -{ > > - del_timer_sync(&test_timer); > > -} > > - > > -module_init(test_init); > > -module_exit(test_exit); > > - > > -MODULE_LICENSE("GPL"); > > -MODULE_AUTHOR("Mathieu Desnoyers"); > > -MODULE_DESCRIPTION("Local Atomic Ops"); > > ---- END --- > > +the ``local_t`` variable as a counter of bytes written in a buffer: there should > > +be a ``smp_wmb()`` between the buffer write and the counter increment and also a > > +``smp_rmb()`` between the counter read and the buffer read. > > + > > + > > +Here is a sample module which implements a basic per cpu counter using > > +``local.h``. > > + > > +.. code-block:: c > > Better to use :: instead of code-block. > > > + > > + /* test-local.c > > + * > > + * Sample module for local.h usage. > > + */ > > + > > + > > + #include <asm/local.h> > > + #include <linux/module.h> > > + #include <linux/timer.h> > > + > > + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > > + > > + static struct timer_list test_timer; > > + > > + /* IPI called on each CPU. */ > > + static void test_each(void *info) > > + { > > + /* Increment the counter from a non preemptible context */ > > + printk("Increment on cpu %d\n", smp_processor_id()); > > + local_inc(this_cpu_ptr(&counters)); > > + > > + /* This is what incrementing the variable would look like within a > > + * preemptible context (it disables preemption) : > > + * > > + * local_inc(&get_cpu_var(counters)); > > + * put_cpu_var(counters); > > + */ > > + } > > + > > + static void do_test_timer(unsigned long data) > > + { > > + int cpu; > > + > > + /* Increment the counters */ > > + on_each_cpu(test_each, NULL, 1); > > + /* Read all the counters */ > > + printk("Counters read from CPU %d\n", smp_processor_id()); > > + for_each_online_cpu(cpu) { > > + printk("Read : CPU %d, count %ld\n", cpu, > > + local_read(&per_cpu(counters, cpu))); > > + } > > + del_timer(&test_timer); > > + test_timer.expires = jiffies + 1000; > > + add_timer(&test_timer); > > + } > > + > > + static int __init test_init(void) > > + { > > + /* initialize the timer that will increment the counter */ > > + init_timer(&test_timer); > > + test_timer.function = do_test_timer; > > + test_timer.expires = jiffies + 1; > > + add_timer(&test_timer); > > + > > + return 0; > > + } > > + > > + static void __exit test_exit(void) > > + { > > + del_timer_sync(&test_timer); > > + } > > + > > + module_init(test_init); > > + module_exit(test_exit); > > + > > + MODULE_LICENSE("GPL"); > > + MODULE_AUTHOR("Mathieu Desnoyers"); > > + MODULE_DESCRIPTION("Local Atomic Ops"); > > > > Thanks, > Mauro -- -- S. Fricke ---------------------------------------- silvio@port1024.net -- Diplom-Informatiker (FH) Linux-Development Matrix: @silvio:port1024.net ---------------------------------------------------------------------------- ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v2 3/4] Documentation/local_ops.txt: convert to ReST markup 2016-11-25 13:47 ` S. Fricke @ 2016-11-25 13:58 ` Mauro Carvalho Chehab 0 siblings, 0 replies; 51+ messages in thread From: Mauro Carvalho Chehab @ 2016-11-25 13:58 UTC (permalink / raw) To: S. Fricke Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab Em Fri, 25 Nov 2016 14:47:11 +0100 "S. Fricke" <silvio.fricke@gmail.com> escreveu: > Hi Mauro, > > I have a question about the "code-block" and "::". On which situation should > I use "code-block" and on which condition a "::"? > For now I have used "::" on small one, two or three liners, and "code-block" > for "example code" snippets or longer code segments. They're equivalent, but :: makes the file to look nicer if someone is reading the text file directly. There's one difference, though: right now, I guess we're disabling pygments. So, "::" is assuming "code-block:: none". The default can be changed inside a document by adding: .. highlight:: c before using the first "::". That said, IMHO, pygments is crap :) Instead of painting the file with some random colors, Sphinx should instead be doing something useful, e. g. producing cross-references for data structures and functions. So, I wouldn't be using "highlight" there. > > Thanks for a small clarification, > Silvio > > > Em Fri, 25 Nov 2016 11:02:41 +0100 > > Silvio Fricke <silvio.fricke@gmail.com> escreveu: > > > > > ... and move to core-api folder. > > > > > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> > > > --- > > > Documentation/core-api/index.rst | 1 +- > > > Documentation/local_ops.txt => Documentation/core-api/local_ops.rst | 275 +++---- > > > 2 files changed, 147 insertions(+), 129 deletions(-) > > > > > > diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst > > > index f3e5f5e..25b4e4a 100644 > > > --- a/Documentation/core-api/index.rst > > > +++ b/Documentation/core-api/index.rst > > > @@ -9,6 +9,7 @@ Kernel and driver related documentation. > > > > > > assoc_array > > > atomic_ops > > > + local_ops > > > workqueue > > > > > > .. only:: subproject > > > diff --git a/Documentation/local_ops.txt b/Documentation/core-api/local_ops.rst > > > similarity index 55% > > > rename from Documentation/local_ops.txt > > > rename to Documentation/core-api/local_ops.rst > > > index 407576a..01f1880 100644 > > > --- a/Documentation/local_ops.txt > > > +++ b/Documentation/core-api/local_ops.rst > > > @@ -1,191 +1,208 @@ > > > - Semantics and Behavior of Local Atomic Operations > > > > > > - Mathieu Desnoyers > > > +.. _local_ops: > > > > > > +================================================= > > > +Semantics and Behavior of Local Atomic Operations > > > +================================================= > > > > > > - This document explains the purpose of the local atomic operations, how > > > +:Author: Mathieu Desnoyers > > > + > > > + > > > +This document explains the purpose of the local atomic operations, how > > > to implement them for any given architecture and shows how they can be used > > > properly. It also stresses on the precautions that must be taken when reading > > > those local variables across CPUs when the order of memory writes matters. > > > > > > -Note that local_t based operations are not recommended for general kernel use. > > > -Please use the this_cpu operations instead unless there is really a special purpose. > > > -Most uses of local_t in the kernel have been replaced by this_cpu operations. > > > -this_cpu operations combine the relocation with the local_t like semantics in > > > -a single instruction and yield more compact and faster executing code. > > > +.. note:: > > > > > > + Note that ``local_t`` based operations are not recommended for general > > > + kernel use. Please use the ``this_cpu`` operations instead unless there is > > > + really a special purpose. Most uses of ``local_t`` in the kernel have been > > > + replaced by ``this_cpu`` operations. ``this_cpu`` operations combine the > > > + relocation with the ``local_t`` like semantics in a single instruction and > > > + yield more compact and faster executing code. > > > > > > -* Purpose of local atomic operations > > > + > > > +Purpose of local atomic operations > > > +================================== > > > > > > Local atomic operations are meant to provide fast and highly reentrant per CPU > > > counters. They minimize the performance cost of standard atomic operations by > > > removing the LOCK prefix and memory barriers normally required to synchronize > > > across CPUs. > > > > > > -Having fast per CPU atomic counters is interesting in many cases : it does not > > > +Having fast per CPU atomic counters is interesting in many cases: it does not > > > require disabling interrupts to protect from interrupt handlers and it permits > > > coherent counters in NMI handlers. It is especially useful for tracing purposes > > > and for various performance monitoring counters. > > > > > > Local atomic operations only guarantee variable modification atomicity wrt the > > > CPU which owns the data. Therefore, care must taken to make sure that only one > > > -CPU writes to the local_t data. This is done by using per cpu data and making > > > -sure that we modify it from within a preemption safe context. It is however > > > -permitted to read local_t data from any CPU : it will then appear to be written > > > -out of order wrt other memory writes by the owner CPU. > > > +CPU writes to the ``local_t`` data. This is done by using per cpu data and > > > +making sure that we modify it from within a preemption safe context. It is > > > +however permitted to read ``local_t`` data from any CPU: it will then appear to > > > +be written out of order wrt other memory writes by the owner CPU. > > > > > > > > > -* Implementation for a given architecture > > > +Implementation for a given architecture > > > +======================================= > > > > > > -It can be done by slightly modifying the standard atomic operations : only > > > +It can be done by slightly modifying the standard atomic operations: only > > > their UP variant must be kept. It typically means removing LOCK prefix (on > > > i386 and x86_64) and any SMP synchronization barrier. If the architecture does > > > -not have a different behavior between SMP and UP, including asm-generic/local.h > > > -in your architecture's local.h is sufficient. > > > +not have a different behavior between SMP and UP, including > > > +``asm-generic/local.h`` in your architecture's ``local.h`` is sufficient. > > > > > > -The local_t type is defined as an opaque signed long by embedding an > > > -atomic_long_t inside a structure. This is made so a cast from this type to a > > > -long fails. The definition looks like : > > > +The ``local_t`` type is defined as an opaque ``signed long`` by embedding an > > > +``atomic_long_t`` inside a structure. This is made so a cast from this type to > > > +a ``long`` fails. The definition looks like:: > > > > > > -typedef struct { atomic_long_t a; } local_t; > > > + typedef struct { atomic_long_t a; } local_t; > > > > > > > > > -* Rules to follow when using local atomic operations > > > +Rules to follow when using local atomic operations > > > +================================================== > > > > > > -- Variables touched by local ops must be per cpu variables. > > > -- _Only_ the CPU owner of these variables must write to them. > > > -- This CPU can use local ops from any context (process, irq, softirq, nmi, ...) > > > - to update its local_t variables. > > > -- Preemption (or interrupts) must be disabled when using local ops in > > > - process context to make sure the process won't be migrated to a > > > +* Variables touched by local ops must be per cpu variables. > > > +* *Only* the CPU owner of these variables must write to them. > > > +* This CPU can use local ops from any context (process, irq, softirq, nmi, ...) > > > + to update its ``local_t`` variables. > > > +* Preemption (or interrupts) must be disabled when using local ops in > > > + process context to make sure the process won't be migrated to a > > > different CPU between getting the per-cpu variable and doing the > > > actual local op. > > > -- When using local ops in interrupt context, no special care must be > > > +* When using local ops in interrupt context, no special care must be > > > taken on a mainline kernel, since they will run on the local CPU with > > > preemption already disabled. I suggest, however, to explicitly > > > disable preemption anyway to make sure it will still work correctly on > > > -rt kernels. > > > -- Reading the local cpu variable will provide the current copy of the > > > +* Reading the local cpu variable will provide the current copy of the > > > variable. > > > -- Reads of these variables can be done from any CPU, because updates to > > > - "long", aligned, variables are always atomic. Since no memory > > > +* Reads of these variables can be done from any CPU, because updates to > > > + "``long``", aligned, variables are always atomic. Since no memory > > > synchronization is done by the writer CPU, an outdated copy of the > > > - variable can be read when reading some _other_ cpu's variables. > > > + variable can be read when reading some *other* cpu's variables. > > > + > > > > > > +How to use local atomic operations > > > +================================== > > > > > > -* How to use local atomic operations > > > +.. code-block:: c > > > > Better to use :: instead of code-block. > > > > > > > > -#include <linux/percpu.h> > > > -#include <asm/local.h> > > > + #include <linux/percpu.h> > > > + #include <asm/local.h> > > > > > > -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > > > + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > > > > > > > > > -* Counting > > > +Counting > > > +======== > > > > > > Counting is done on all the bits of a signed long. > > > > > > -In preemptible context, use get_cpu_var() and put_cpu_var() around local atomic > > > -operations : it makes sure that preemption is disabled around write access to > > > -the per cpu variable. For instance : > > > +In preemptible context, use ``get_cpu_var()`` and ``put_cpu_var()`` around > > > +local atomic operations: it makes sure that preemption is disabled around write > > > +access to the per cpu variable. For instance:: > > > > > > - local_inc(&get_cpu_var(counters)); > > > - put_cpu_var(counters); > > > + local_inc(&get_cpu_var(counters)); > > > + put_cpu_var(counters); > > > > > > If you are already in a preemption-safe context, you can use > > > -this_cpu_ptr() instead. > > > +``this_cpu_ptr()`` instead. :: > > > > > > - local_inc(this_cpu_ptr(&counters)); > > > + local_inc(this_cpu_ptr(&counters)); > > > > > > > > > > > > -* Reading the counters > > > +Reading the counters > > > +==================== > > > > > > Those local counters can be read from foreign CPUs to sum the count. Note that > > > the data seen by local_read across CPUs must be considered to be out of order > > > -relatively to other memory writes happening on the CPU that owns the data. > > > +relatively to other memory writes happening on the CPU that owns the data. :: > > > > Please remove the dot at the end. > > > > > > > > - long sum = 0; > > > - for_each_online_cpu(cpu) > > > - sum += local_read(&per_cpu(counters, cpu)); > > > + long sum = 0; > > > + for_each_online_cpu(cpu) > > > + sum += local_read(&per_cpu(counters, cpu)); > > > > > > If you want to use a remote local_read to synchronize access to a resource > > > -between CPUs, explicit smp_wmb() and smp_rmb() memory barriers must be used > > > +between CPUs, explicit ``smp_wmb()`` and ``smp_rmb()`` memory barriers must be used > > > respectively on the writer and the reader CPUs. It would be the case if you use > > > -the local_t variable as a counter of bytes written in a buffer : there should > > > -be a smp_wmb() between the buffer write and the counter increment and also a > > > -smp_rmb() between the counter read and the buffer read. > > > - > > > - > > > -Here is a sample module which implements a basic per cpu counter using local.h. > > > - > > > ---- BEGIN --- > > > -/* test-local.c > > > - * > > > - * Sample module for local.h usage. > > > - */ > > > - > > > - > > > -#include <asm/local.h> > > > -#include <linux/module.h> > > > -#include <linux/timer.h> > > > - > > > -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > > > - > > > -static struct timer_list test_timer; > > > - > > > -/* IPI called on each CPU. */ > > > -static void test_each(void *info) > > > -{ > > > - /* Increment the counter from a non preemptible context */ > > > - printk("Increment on cpu %d\n", smp_processor_id()); > > > - local_inc(this_cpu_ptr(&counters)); > > > - > > > - /* This is what incrementing the variable would look like within a > > > - * preemptible context (it disables preemption) : > > > - * > > > - * local_inc(&get_cpu_var(counters)); > > > - * put_cpu_var(counters); > > > - */ > > > -} > > > - > > > -static void do_test_timer(unsigned long data) > > > -{ > > > - int cpu; > > > - > > > - /* Increment the counters */ > > > - on_each_cpu(test_each, NULL, 1); > > > - /* Read all the counters */ > > > - printk("Counters read from CPU %d\n", smp_processor_id()); > > > - for_each_online_cpu(cpu) { > > > - printk("Read : CPU %d, count %ld\n", cpu, > > > - local_read(&per_cpu(counters, cpu))); > > > - } > > > - del_timer(&test_timer); > > > - test_timer.expires = jiffies + 1000; > > > - add_timer(&test_timer); > > > -} > > > - > > > -static int __init test_init(void) > > > -{ > > > - /* initialize the timer that will increment the counter */ > > > - init_timer(&test_timer); > > > - test_timer.function = do_test_timer; > > > - test_timer.expires = jiffies + 1; > > > - add_timer(&test_timer); > > > - > > > - return 0; > > > -} > > > - > > > -static void __exit test_exit(void) > > > -{ > > > - del_timer_sync(&test_timer); > > > -} > > > - > > > -module_init(test_init); > > > -module_exit(test_exit); > > > - > > > -MODULE_LICENSE("GPL"); > > > -MODULE_AUTHOR("Mathieu Desnoyers"); > > > -MODULE_DESCRIPTION("Local Atomic Ops"); > > > ---- END --- > > > +the ``local_t`` variable as a counter of bytes written in a buffer: there should > > > +be a ``smp_wmb()`` between the buffer write and the counter increment and also a > > > +``smp_rmb()`` between the counter read and the buffer read. > > > + > > > + > > > +Here is a sample module which implements a basic per cpu counter using > > > +``local.h``. > > > + > > > +.. code-block:: c > > > > Better to use :: instead of code-block. > > > > > + > > > + /* test-local.c > > > + * > > > + * Sample module for local.h usage. > > > + */ > > > + > > > + > > > + #include <asm/local.h> > > > + #include <linux/module.h> > > > + #include <linux/timer.h> > > > + > > > + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > > > + > > > + static struct timer_list test_timer; > > > + > > > + /* IPI called on each CPU. */ > > > + static void test_each(void *info) > > > + { > > > + /* Increment the counter from a non preemptible context */ > > > + printk("Increment on cpu %d\n", smp_processor_id()); > > > + local_inc(this_cpu_ptr(&counters)); > > > + > > > + /* This is what incrementing the variable would look like within a > > > + * preemptible context (it disables preemption) : > > > + * > > > + * local_inc(&get_cpu_var(counters)); > > > + * put_cpu_var(counters); > > > + */ > > > + } > > > + > > > + static void do_test_timer(unsigned long data) > > > + { > > > + int cpu; > > > + > > > + /* Increment the counters */ > > > + on_each_cpu(test_each, NULL, 1); > > > + /* Read all the counters */ > > > + printk("Counters read from CPU %d\n", smp_processor_id()); > > > + for_each_online_cpu(cpu) { > > > + printk("Read : CPU %d, count %ld\n", cpu, > > > + local_read(&per_cpu(counters, cpu))); > > > + } > > > + del_timer(&test_timer); > > > + test_timer.expires = jiffies + 1000; > > > + add_timer(&test_timer); > > > + } > > > + > > > + static int __init test_init(void) > > > + { > > > + /* initialize the timer that will increment the counter */ > > > + init_timer(&test_timer); > > > + test_timer.function = do_test_timer; > > > + test_timer.expires = jiffies + 1; > > > + add_timer(&test_timer); > > > + > > > + return 0; > > > + } > > > + > > > + static void __exit test_exit(void) > > > + { > > > + del_timer_sync(&test_timer); > > > + } > > > + > > > + module_init(test_init); > > > + module_exit(test_exit); > > > + > > > + MODULE_LICENSE("GPL"); > > > + MODULE_AUTHOR("Mathieu Desnoyers"); > > > + MODULE_DESCRIPTION("Local Atomic Ops"); > > > > > > > > Thanks, > > Mauro > Thanks, Mauro ^ permalink raw reply [flat|nested] 51+ messages in thread
* [PATCH v2 4/4] firmware: remove warning at documentation generation time 2016-11-25 10:02 ` [PATCH v2 " Silvio Fricke ` (2 preceding siblings ...) 2016-11-25 10:02 ` [PATCH v2 3/4] Documentation/local_ops.txt: " Silvio Fricke @ 2016-11-25 10:02 ` Silvio Fricke 2016-11-25 13:34 ` Mauro Carvalho Chehab 2016-11-25 14:59 ` [PATCH v3 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 4 siblings, 1 reply; 51+ messages in thread From: Silvio Fricke @ 2016-11-25 10:02 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Silvio Fricke This patch removes following error at for `make htmldocs`. No functional change. ./drivers/base/firmware_class.c:1348: WARNING: Bullet list ends without a blank line; unexpected unindent. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- drivers/base/firmware_class.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c index 22d1760..37b0221 100644 --- a/drivers/base/firmware_class.c +++ b/drivers/base/firmware_class.c @@ -1345,9 +1345,9 @@ static void request_firmware_work_func(struct work_struct *work) * * Asynchronous variant of request_firmware() for user contexts: * - sleep for as small periods as possible since it may - * increase kernel boot time of built-in device drivers - * requesting firmware in their ->probe() methods, if - * @gfp is GFP_KERNEL. + * increase kernel boot time of built-in device drivers + * requesting firmware in their ->probe() methods, if + * @gfp is GFP_KERNEL. * * - can't sleep at all if @gfp is GFP_ATOMIC. **/ -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* Re: [PATCH v2 4/4] firmware: remove warning at documentation generation time 2016-11-25 10:02 ` [PATCH v2 4/4] firmware: remove warning at documentation generation time Silvio Fricke @ 2016-11-25 13:34 ` Mauro Carvalho Chehab 0 siblings, 0 replies; 51+ messages in thread From: Mauro Carvalho Chehab @ 2016-11-25 13:34 UTC (permalink / raw) To: Silvio Fricke Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel Em Fri, 25 Nov 2016 11:02:42 +0100 Silvio Fricke <silvio.fricke@gmail.com> escreveu: > This patch removes following error at for `make htmldocs`. No functional > change. > > ./drivers/base/firmware_class.c:1348: WARNING: Bullet list ends without a blank line; unexpected unindent. > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> Reviewed-by: Mauro Carvalho Chehab <mchehab@s-opensource.com> > --- > drivers/base/firmware_class.c | 6 +++--- > 1 file changed, 3 insertions(+), 3 deletions(-) > > diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c > index 22d1760..37b0221 100644 > --- a/drivers/base/firmware_class.c > +++ b/drivers/base/firmware_class.c > @@ -1345,9 +1345,9 @@ static void request_firmware_work_func(struct work_struct *work) > * > * Asynchronous variant of request_firmware() for user contexts: > * - sleep for as small periods as possible since it may > - * increase kernel boot time of built-in device drivers > - * requesting firmware in their ->probe() methods, if > - * @gfp is GFP_KERNEL. > + * increase kernel boot time of built-in device drivers > + * requesting firmware in their ->probe() methods, if > + * @gfp is GFP_KERNEL. > * > * - can't sleep at all if @gfp is GFP_ATOMIC. > **/ Thanks, Mauro ^ permalink raw reply [flat|nested] 51+ messages in thread
* [PATCH v3 0/4] core-api ReST: assoc_array, atomic_ops, local_ops 2016-11-25 10:02 ` [PATCH v2 " Silvio Fricke ` (3 preceding siblings ...) 2016-11-25 10:02 ` [PATCH v2 4/4] firmware: remove warning at documentation generation time Silvio Fricke @ 2016-11-25 14:59 ` Silvio Fricke 2016-11-25 14:59 ` [PATCH v3 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke ` (4 more replies) 4 siblings, 5 replies; 51+ messages in thread From: Silvio Fricke @ 2016-11-25 14:59 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula, Silvio Fricke Hi, Thanks Mauro and Jani for reviewing. Some more ReSTification of core-api's: assoc_array, atomic_ops and local_ops. A fourth patch removes a warning about a bullet list without ending at firmware_class.c v2 -> v3 * change ". ::" to "::" * replace all "code-blocks" with "::" * add two ".. notes" declaration in atomic_ops.rst v1 -> v2 * use format-patch with a rename_threshold of 10%, no other changes Thanks for review. BR Silvio Silvio Fricke (4): Documentation/assoc_array.txt: convert to ReST markup Documentation/atomic_ops.txt: convert to ReST markup Documentation/local_ops.txt: convert to ReST markup firmware: remove warning at documentation generation time Documentation/assoc_array.txt => Documentation/core-api/assoc_array.rst | 639 ++++++++++++++++++++++++++++------------------------------- Documentation/atomic_ops.txt => Documentation/core-api/atomic_ops.rst | 777 +++++++++++++++++++++++++++++++++++++----------------------------------- Documentation/core-api/index.rst | 3 +- Documentation/local_ops.txt => Documentation/core-api/local_ops.rst | 273 +++++++++++++------------ Documentation/process/volatile-considered-harmful.rst | 3 +- drivers/base/firmware_class.c | 6 +- 6 files changed, 861 insertions(+), 840 deletions(-) base-commit: 9c240d757658a3ae9968dd309e674c61f07c7f48 -- git-series 0.9.1 ^ permalink raw reply [flat|nested] 51+ messages in thread
* [PATCH v3 1/4] Documentation/assoc_array.txt: convert to ReST markup 2016-11-25 14:59 ` [PATCH v3 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke @ 2016-11-25 14:59 ` Silvio Fricke 2016-11-25 20:53 ` Mauro Carvalho Chehab 2016-11-25 14:59 ` [PATCH v3 2/4] Documentation/atomic_ops.txt: " Silvio Fricke ` (3 subsequent siblings) 4 siblings, 1 reply; 51+ messages in thread From: Silvio Fricke @ 2016-11-25 14:59 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula, Silvio Fricke ... and move to Documentation/core-api folder. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- Documentation/assoc_array.txt => Documentation/core-api/assoc_array.rst | 639 ++++++++++++++++++++++++++++++++++-------------------------------------- Documentation/core-api/index.rst | 1 +- 2 files changed, 309 insertions(+), 331 deletions(-) diff --git a/Documentation/assoc_array.txt b/Documentation/core-api/assoc_array.rst similarity index 46% rename from Documentation/assoc_array.txt rename to Documentation/core-api/assoc_array.rst index 2f2c6cd..dcda7c6 100644 --- a/Documentation/assoc_array.txt +++ b/Documentation/core-api/assoc_array.rst @@ -1,67 +1,46 @@ - ======================================== - GENERIC ASSOCIATIVE ARRAY IMPLEMENTATION - ======================================== +======================================== +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 +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. +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. +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. +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. +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. +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. +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. +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. +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. +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. +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. +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 @@ -71,25 +50,26 @@ 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 -============== -The public API can be found in <linux/assoc_array.h>. The associative array is -rooted on the following structure: +The public API can be found in ``<linux/assoc_array.h>``. The associative +array is rooted on the following structure:: + + struct assoc_array { + ... + }; - struct assoc_array { - ... - }; +The code is selected by enabling ``CONFIG_ASSOCIATIVE_ARRAY`` with:: -The code is selected by enabling CONFIG_ASSOCIATIVE_ARRAY. + ./script/config -e ASSOCIATIVE_ARRAY -EDIT SCRIPT +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 +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. @@ -99,246 +79,245 @@ 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: +The script appears as outside of the API as a pointer of the type:: - struct assoc_array_edit; + struct assoc_array_edit; There are two functions for dealing with the script: - (1) Apply an edit script. +1. Apply an edit script:: - void assoc_array_apply_edit(struct assoc_array_edit *edit); + 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. +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. +2. Cancel an edit script:: - void assoc_array_cancel_edit(struct assoc_array_edit *edit); + 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. +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 +Operations Table ---------------- -Various functions take a table of operations: +Various functions take a table of operations:: - struct assoc_array_ops { - ... - }; + 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: +1. Get a chunk of index key from caller data:: - unsigned long (*get_key_chunk)(const void *index_key, int level); + 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. +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. +2. Get a chunk of an object's index key:: - unsigned long (*get_object_key_chunk)(const void *object, int level); + 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. +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. +3. See if this is the object we're looking for:: - bool (*compare_object)(const void *object, const void *index_key); + 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. +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. +4. Diff the index keys of two objects:: - int (*diff_objects)(const void *object, const void *index_key); + int (*diff_objects)(const void *object, const void *index_key); - Return the bit position at which the index key of the specified object - differs from the given index key or -1 if they are the same. +Return the bit position at which the index key of the specified object +differs from the given index key or -1 if they are the same. - (5) Free an object. +5. Free an object:: - void (*free_object)(void *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. +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 +Manipulation Functions ---------------------- There are a number of functions for manipulating an associative array: - (1) Initialise an associative array. +1. Initialise an associative array:: - void assoc_array_init(struct assoc_array *array); + void assoc_array_init(struct assoc_array *array); - This initialises the base structure for an associative array. It can't - fail. +This initialises the base structure for an associative array. It can't fail. - (2) Insert/replace an object in an associative array. +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); + 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. +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. +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. +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. +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. +The caller should lock exclusively against other modifiers of the array. - (3) Delete an object from an associative 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); + 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. +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. +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. +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. +The caller should lock exclusively against other modifiers of the array. - (4) Delete all objects from an associative 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); + 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 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. +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. +The caller should lock exclusively against other modifiers of the array. - (5) Destroy an associative array, deleting all objects. +5. Destroy an associative array, deleting all objects:: - void assoc_array_destroy(struct assoc_array *array, - const struct assoc_array_ops *ops); + 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. +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. +The caller should lock exclusively against other modifiers and accessors +of the array. - (6) Garbage collect an associative 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); + 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. +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 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 ``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. +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. +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 +Access Functions ---------------- There are two functions for accessing an associative array: - (1) Iterate over all the objects in 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); + 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 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. +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. +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. +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); + 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 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. +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. +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 +Index Key Form -------------- The index key can be of any form, but since the algorithms aren't told how long @@ -364,8 +343,7 @@ unlikely that more than one word of any particular index key will have to be used. -================= -INTERNAL WORKINGS +Internal Workings ================= The associative array data structure has an internal tree. This tree is @@ -373,82 +351,80 @@ 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. +* 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 +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 | - +---+ +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 +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 +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 @@ -462,23 +438,23 @@ 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: +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 + 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: +and node B:: - 3 PTR TO NODE C 13* - any LEAF 1458952489 + 3 PTR TO NODE C 13* + any LEAF 1458952489 -SHORTCUTS +Shortcuts --------- Shortcuts are metadata records that jump over a piece of keyspace. A shortcut @@ -486,12 +462,13 @@ 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. +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 +Splitting And Collapsing Nodes ------------------------------ Each node has a maximum capacity of 16 leaves and metadata pointers. If the @@ -508,7 +485,7 @@ fewer, then the subtree will be collapsed down to a single node - and this will ripple towards the root if possible. -NON-RECURSIVE ITERATION +Non-Recursive Iteration ----------------------- Each node and shortcut contains a back pointer to its parent and the number of @@ -519,55 +496,55 @@ make sure progress is made without the need for a stack. The backpointers, however, make simultaneous alteration and iteration tricky. -SIMULTANEOUS ALTERATION AND ITERATION +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. +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 diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index f7ef7fd..480d9a3 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -7,6 +7,7 @@ Kernel and driver related documentation. .. toctree:: :maxdepth: 1 + assoc_array workqueue .. only:: subproject -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* Re: [PATCH v3 1/4] Documentation/assoc_array.txt: convert to ReST markup 2016-11-25 14:59 ` [PATCH v3 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke @ 2016-11-25 20:53 ` Mauro Carvalho Chehab 0 siblings, 0 replies; 51+ messages in thread From: Mauro Carvalho Chehab @ 2016-11-25 20:53 UTC (permalink / raw) To: Silvio Fricke Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula Em Fri, 25 Nov 2016 15:59:44 +0100 Silvio Fricke <silvio.fricke@gmail.com> escreveu: > ... and move to Documentation/core-api folder. > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> Reviewed-by: Mauro Carvalho Chehab <mchehab@s-opensource.com> > --- > Documentation/assoc_array.txt => Documentation/core-api/assoc_array.rst | 639 ++++++++++++++++++++++++++++++++++-------------------------------------- > Documentation/core-api/index.rst | 1 +- > 2 files changed, 309 insertions(+), 331 deletions(-) > > diff --git a/Documentation/assoc_array.txt b/Documentation/core-api/assoc_array.rst > similarity index 46% > rename from Documentation/assoc_array.txt > rename to Documentation/core-api/assoc_array.rst > index 2f2c6cd..dcda7c6 100644 > --- a/Documentation/assoc_array.txt > +++ b/Documentation/core-api/assoc_array.rst > @@ -1,67 +1,46 @@ > - ======================================== > - GENERIC ASSOCIATIVE ARRAY IMPLEMENTATION > - ======================================== > +======================================== > +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 > +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. > +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. > +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. > +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. > +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. > +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. > +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. > +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. > +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. > +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. > +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. > +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 > @@ -71,25 +50,26 @@ 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 > -============== > > -The public API can be found in <linux/assoc_array.h>. The associative array is > -rooted on the following structure: > +The public API can be found in ``<linux/assoc_array.h>``. The associative > +array is rooted on the following structure:: > + > + struct assoc_array { > + ... > + }; > > - struct assoc_array { > - ... > - }; > +The code is selected by enabling ``CONFIG_ASSOCIATIVE_ARRAY`` with:: > > -The code is selected by enabling CONFIG_ASSOCIATIVE_ARRAY. > + ./script/config -e ASSOCIATIVE_ARRAY > > > -EDIT SCRIPT > +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 > +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. > @@ -99,246 +79,245 @@ 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: > +The script appears as outside of the API as a pointer of the type:: > > - struct assoc_array_edit; > + struct assoc_array_edit; > > There are two functions for dealing with the script: > > - (1) Apply an edit script. > +1. Apply an edit script:: > > - void assoc_array_apply_edit(struct assoc_array_edit *edit); > + 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. > +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. > +2. Cancel an edit script:: > > - void assoc_array_cancel_edit(struct assoc_array_edit *edit); > + 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. > +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 > +Operations Table > ---------------- > > -Various functions take a table of operations: > +Various functions take a table of operations:: > > - struct assoc_array_ops { > - ... > - }; > + 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: > +1. Get a chunk of index key from caller data:: > > - unsigned long (*get_key_chunk)(const void *index_key, int level); > + 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. > +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. > +2. Get a chunk of an object's index key:: > > - unsigned long (*get_object_key_chunk)(const void *object, int level); > + 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. > +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. > +3. See if this is the object we're looking for:: > > - bool (*compare_object)(const void *object, const void *index_key); > + 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. > +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. > +4. Diff the index keys of two objects:: > > - int (*diff_objects)(const void *object, const void *index_key); > + int (*diff_objects)(const void *object, const void *index_key); > > - Return the bit position at which the index key of the specified object > - differs from the given index key or -1 if they are the same. > +Return the bit position at which the index key of the specified object > +differs from the given index key or -1 if they are the same. > > > - (5) Free an object. > +5. Free an object:: > > - void (*free_object)(void *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. > +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 > +Manipulation Functions > ---------------------- > > There are a number of functions for manipulating an associative array: > > - (1) Initialise an associative array. > +1. Initialise an associative array:: > > - void assoc_array_init(struct assoc_array *array); > + void assoc_array_init(struct assoc_array *array); > > - This initialises the base structure for an associative array. It can't > - fail. > +This initialises the base structure for an associative array. It can't fail. > > > - (2) Insert/replace an object in an associative array. > +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); > + 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. > +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. > +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. > +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. > +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. > +The caller should lock exclusively against other modifiers of the array. > > > - (3) Delete an object from an associative 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); > + 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. > +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. > +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. > +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. > +The caller should lock exclusively against other modifiers of the array. > > > - (4) Delete all objects from an associative 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); > + 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 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. > +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. > +The caller should lock exclusively against other modifiers of the array. > > > - (5) Destroy an associative array, deleting all objects. > +5. Destroy an associative array, deleting all objects:: > > - void assoc_array_destroy(struct assoc_array *array, > - const struct assoc_array_ops *ops); > + 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. > +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. > +The caller should lock exclusively against other modifiers and accessors > +of the array. > > > - (6) Garbage collect an associative 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); > + 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. > +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 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 ``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. > +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. > +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 > +Access Functions > ---------------- > > There are two functions for accessing an associative array: > > - (1) Iterate over all the objects in 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); > + 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 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. > +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. > +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. > +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); > + 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 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. > +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. > +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 > +Index Key Form > -------------- > > The index key can be of any form, but since the algorithms aren't told how long > @@ -364,8 +343,7 @@ unlikely that more than one word of any particular index key will have to be > used. > > > -================= > -INTERNAL WORKINGS > +Internal Workings > ================= > > The associative array data structure has an internal tree. This tree is > @@ -373,82 +351,80 @@ 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. > +* 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 > +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 | > - +---+ > +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 > +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 > +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 > @@ -462,23 +438,23 @@ 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: > +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 > + 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: > +and node B:: > > - 3 PTR TO NODE C 13* > - any LEAF 1458952489 > + 3 PTR TO NODE C 13* > + any LEAF 1458952489 > > > -SHORTCUTS > +Shortcuts > --------- > > Shortcuts are metadata records that jump over a piece of keyspace. A shortcut > @@ -486,12 +462,13 @@ 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. > +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 > +Splitting And Collapsing Nodes > ------------------------------ > > Each node has a maximum capacity of 16 leaves and metadata pointers. If the > @@ -508,7 +485,7 @@ fewer, then the subtree will be collapsed down to a single node - and this will > ripple towards the root if possible. > > > -NON-RECURSIVE ITERATION > +Non-Recursive Iteration > ----------------------- > > Each node and shortcut contains a back pointer to its parent and the number of > @@ -519,55 +496,55 @@ make sure progress is made without the need for a stack. > The backpointers, however, make simultaneous alteration and iteration tricky. > > > -SIMULTANEOUS ALTERATION AND ITERATION > +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. > +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 > diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst > index f7ef7fd..480d9a3 100644 > --- a/Documentation/core-api/index.rst > +++ b/Documentation/core-api/index.rst > @@ -7,6 +7,7 @@ Kernel and driver related documentation. > .. toctree:: > :maxdepth: 1 > > + assoc_array > workqueue > > .. only:: subproject Thanks, Mauro ^ permalink raw reply [flat|nested] 51+ messages in thread
* [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-25 14:59 ` [PATCH v3 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 2016-11-25 14:59 ` [PATCH v3 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke @ 2016-11-25 14:59 ` Silvio Fricke 2016-11-25 21:01 ` Mauro Carvalho Chehab 2016-11-25 21:58 ` Peter Zijlstra 2016-11-25 14:59 ` [PATCH v3 3/4] Documentation/local_ops.txt: " Silvio Fricke ` (2 subsequent siblings) 4 siblings, 2 replies; 51+ messages in thread From: Silvio Fricke @ 2016-11-25 14:59 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula, Silvio Fricke ... and move to core-api folder. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- Documentation/atomic_ops.txt => Documentation/core-api/atomic_ops.rst | 777 +++++++++++++++++++++++++++++++++++++----------------------------------- Documentation/core-api/index.rst | 1 +- Documentation/process/volatile-considered-harmful.rst | 3 +- 3 files changed, 404 insertions(+), 377 deletions(-) diff --git a/Documentation/atomic_ops.txt b/Documentation/core-api/atomic_ops.rst similarity index 46% rename from Documentation/atomic_ops.txt rename to Documentation/core-api/atomic_ops.rst index 6c5e8a9..6236951 100644 --- a/Documentation/atomic_ops.txt +++ b/Documentation/core-api/atomic_ops.rst @@ -1,204 +1,212 @@ - Semantics and Behavior of Atomic and - Bitmask Operations +======================================================= +Semantics and Behavior of Atomic and Bitmask Operations +======================================================= - David S. Miller +:Author: David S. Miller - This document is intended to serve as a guide to Linux port +This document is intended to serve as a guide to Linux port maintainers on how to implement atomic counter, bitops, and spinlock interfaces properly. - The atomic_t type should be defined as a signed integer and -the atomic_long_t type as a signed long integer. Also, they should +Atomic Type And Operations +========================== + +The ``atomic_t type`` should be defined as a signed integer and +the ``atomic_long_t`` type as a signed long integer. Also, they should be made opaque such that any kind of cast to a normal C integer type -will fail. Something like the following should suffice: +will fail. Something like the following should suffice:: - typedef struct { int counter; } atomic_t; - typedef struct { long counter; } atomic_long_t; + typedef struct { int counter; } atomic_t; + typedef struct { long counter; } atomic_long_t; Historically, counter has been declared volatile. This is now discouraged. -See Documentation/process/volatile-considered-harmful.rst for the complete rationale. +See :ref:`Documentation/process/volatile-considered-harmful.rst +<volatile_considered_harmful>` for the complete rationale. -local_t is very similar to atomic_t. If the counter is per CPU and only -updated by one CPU, local_t is probably more appropriate. Please see -Documentation/local_ops.txt for the semantics of local_t. +``local_t`` is very similar to ``atomic_t``. If the counter is per CPU and only +updated by one CPU, ``local_t`` is probably more appropriate. Please see +:ref:`Documentation/local_ops.txt <local_ops>` for the semantics of ``local_t``. -The first operations to implement for atomic_t's are the initializers and -plain reads. +The first operations to implement for ``atomic_t``'s are the initializers and +plain reads:: - #define ATOMIC_INIT(i) { (i) } - #define atomic_set(v, i) ((v)->counter = (i)) + #define ATOMIC_INIT(i) { (i) } + #define atomic_set(v, i) ((v)->counter = (i)) -The first macro is used in definitions, such as: +The first macro is used in definitions, such as:: -static atomic_t my_counter = ATOMIC_INIT(1); + static atomic_t my_counter = ATOMIC_INIT(1); The initializer is atomic in that the return values of the atomic operations are guaranteed to be correct reflecting the initialized value if the initializer is used before runtime. If the initializer is used at runtime, a proper implicit or explicit read memory barrier is needed before reading the -value with atomic_read from another thread. +value with ``atomic_read`` from another thread. -As with all of the atomic_ interfaces, replace the leading "atomic_" -with "atomic_long_" to operate on atomic_long_t. +As with all of the ``atomic_`` interfaces, replace the leading ``atomic_`` +with ``atomic_long_`` to operate on ``atomic_long_t``. -The second interface can be used at runtime, as in: +The second interface can be used at runtime, as in:: - struct foo { atomic_t counter; }; - ... + struct foo { atomic_t counter; }; + ... - struct foo *k; + struct foo *k; - k = kmalloc(sizeof(*k), GFP_KERNEL); - if (!k) - return -ENOMEM; - atomic_set(&k->counter, 0); + k = kmalloc(sizeof(*k), GFP_KERNEL); + if (!k) + return -ENOMEM; + atomic_set(&k->counter, 0); The setting is atomic in that the return values of the atomic operations by all threads are guaranteed to be correct reflecting either the value that has been set with this operation or set with another operation. A proper implicit or explicit memory barrier is needed before the value set with the operation -is guaranteed to be readable with atomic_read from another thread. +is guaranteed to be readable with ``atomic_read`` from another thread. -Next, we have: +Next, we have:: - #define atomic_read(v) ((v)->counter) + #define atomic_read(v) ((v)->counter) which simply reads the counter value currently visible to the calling thread. The read is atomic in that the return value is guaranteed to be one of the values initialized or modified with the interface operations if a proper implicit or explicit memory barrier is used after possible runtime initialization by any other thread and the value is modified only with the -interface operations. atomic_read does not guarantee that the runtime +interface operations. ``atomic_read`` does not guarantee that the runtime initialization by any other thread is visible yet, so the user of the interface must take care of that with a proper implicit or explicit memory barrier. -*** WARNING: atomic_read() and atomic_set() DO NOT IMPLY BARRIERS! *** +.. warning:: -Some architectures may choose to use the volatile keyword, barriers, or inline -assembly to guarantee some degree of immediacy for atomic_read() and -atomic_set(). This is not uniformly guaranteed, and may change in the future, -so all users of atomic_t should treat atomic_read() and atomic_set() as simple -C statements that may be reordered or optimized away entirely by the compiler -or processor, and explicitly invoke the appropriate compiler and/or memory -barrier for each use case. Failure to do so will result in code that may -suddenly break when used with different architectures or compiler -optimizations, or even changes in unrelated code which changes how the -compiler optimizes the section accessing atomic_t variables. + ``atomic_read()`` and ``atomic_set()`` DO NOT IMPLY BARRIERS! -*** YOU HAVE BEEN WARNED! *** + Some architectures may choose to use the ``volatile`` keyword, barriers, or + ``inline`` assembly to guarantee some degree of immediacy for ``atomic_read()`` + and ``atomic_set()``. This is not uniformly guaranteed, and may change in the + future, so all users of ``atomic_t`` should treat ``atomic_read()`` and + ``atomic_set()`` as simple C statements that may be reordered or optimized away + entirely by the compiler or processor, and explicitly invoke the appropriate + compiler and/or memory barrier for each use case. Failure to do so will result + in code that may suddenly break when used with different architectures or + compiler optimizations, or even changes in unrelated code which changes how the + compiler optimizes the section accessing ``atomic_t`` variables. Properly aligned pointers, longs, ints, and chars (and unsigned equivalents) may be atomically loaded from and stored to in the same -sense as described for atomic_read() and atomic_set(). The READ_ONCE() -and WRITE_ONCE() macros should be used to prevent the compiler from using -optimizations that might otherwise optimize accesses out of existence on -the one hand, or that might create unsolicited accesses on the other. +sense as described for ``atomic_read()`` and ``atomic_set()``. The +``READ_ONCE()`` and ``WRITE_ONCE()`` macros should be used to prevent the +compiler from using optimizations that might otherwise optimize accesses out of +existence on the one hand, or that might create unsolicited accesses on the +other. -For example consider the following code: +For example consider the following code:: - while (a > 0) - do_something(); + while (a > 0) + do_something(); -If the compiler can prove that do_something() does not store to the +If the compiler can prove that ``do_something()`` does not store to the variable a, then the compiler is within its rights transforming this to -the following: +the following:: - tmp = a; - if (a > 0) - for (;;) - do_something(); + tmp = a; + if (a > 0) + for (;;) + do_something(); If you don't want the compiler to do this (and you probably don't), then -you should use something like the following: +you should use something like the following:: - while (READ_ONCE(a) < 0) - do_something(); + while (READ_ONCE(a) < 0) + do_something(); -Alternatively, you could place a barrier() call in the loop. +Alternatively, you could place a ``barrier()`` call in the loop. -For another example, consider the following code: +For another example, consider the following code:: - tmp_a = a; - do_something_with(tmp_a); - do_something_else_with(tmp_a); + tmp_a = a; + do_something_with(tmp_a); + do_something_else_with(tmp_a); -If the compiler can prove that do_something_with() does not store to the -variable a, then the compiler is within its rights to manufacture an -additional load as follows: +If the compiler can prove that ``do_something_with()`` does not store to the +variable ``a``, then the compiler is within its rights to manufacture an +additional load as follows:: - tmp_a = a; - do_something_with(tmp_a); - tmp_a = a; - do_something_else_with(tmp_a); + tmp_a = a; + do_something_with(tmp_a); + tmp_a = a; + do_something_else_with(tmp_a); This could fatally confuse your code if it expected the same value -to be passed to do_something_with() and do_something_else_with(). +to be passed to ``do_something_with()`` and ``do_something_else_with()``. The compiler would be likely to manufacture this additional load if -do_something_with() was an inline function that made very heavy use +``do_something_with()`` was an inline function that made very heavy use of registers: reloading from variable a could save a flush to the stack and later reload. To prevent the compiler from attacking your -code in this manner, write the following: +code in this manner, write the following:: - tmp_a = READ_ONCE(a); - do_something_with(tmp_a); - do_something_else_with(tmp_a); + tmp_a = READ_ONCE(a); + do_something_with(tmp_a); + do_something_else_with(tmp_a); For a final example, consider the following code, assuming that the -variable a is set at boot time before the second CPU is brought online -and never changed later, so that memory barriers are not needed: +variable ``a`` is set at boot time before the second CPU is brought online +and never changed later, so that memory barriers are not needed:: - if (a) - b = 9; - else - b = 42; + if (a) + b = 9; + else + b = 42; The compiler is within its rights to manufacture an additional store -by transforming the above code into the following: +by transforming the above code into the following:: - b = 42; - if (a) - b = 9; + b = 42; + if (a) + b = 9; This could come as a fatal surprise to other code running concurrently -that expected b to never have the value 42 if a was zero. To prevent -the compiler from doing this, write something like: +that expected ``b`` to never have the value ``42`` if ``a`` was zero. To +prevent the compiler from doing this, write something like:: + + if (a) + WRITE_ONCE(b, 9); + else + WRITE_ONCE(b, 42); - if (a) - WRITE_ONCE(b, 9); - else - WRITE_ONCE(b, 42); +Don't even **think** about doing this without proper use of memory barriers, +locks, or atomic operations if variable ``a`` can change at runtime! -Don't even -think- about doing this without proper use of memory barriers, -locks, or atomic operations if variable a can change at runtime! +.. warning:: -*** WARNING: READ_ONCE() OR WRITE_ONCE() DO NOT IMPLY A BARRIER! *** + ``READ_ONCE()`` OR ``WRITE_ONCE()`` DO NOT IMPLY A BARRIER! Now, we move onto the atomic operation interfaces typically implemented with -the help of assembly code. +the help of assembly code:: - void atomic_add(int i, atomic_t *v); - void atomic_sub(int i, atomic_t *v); - void atomic_inc(atomic_t *v); - void atomic_dec(atomic_t *v); + void atomic_add(int i, atomic_t *v); + void atomic_sub(int i, atomic_t *v); + void atomic_inc(atomic_t *v); + void atomic_dec(atomic_t *v); These four routines add and subtract integral values to/from the given -atomic_t value. The first two routines pass explicit integers by +``atomic_t`` value. The first two routines pass explicit integers by which to make the adjustment, whereas the latter two use an implicit adjustment value of "1". -One very important aspect of these two routines is that they DO NOT +One very important aspect of these two routines is that they **DO NOT** require any explicit memory barriers. They need only perform the -atomic_t counter update in an SMP safe manner. +``atomic_t`` counter update in an SMP safe manner. -Next, we have: +Next, we have:: - int atomic_inc_return(atomic_t *v); - int atomic_dec_return(atomic_t *v); + int atomic_inc_return(atomic_t *v); + int atomic_dec_return(atomic_t *v); These routines add 1 and subtract 1, respectively, from the given -atomic_t and return the new counter value after the operation is +``atomic_t`` and return the new counter value after the operation is performed. Unlike the above routines, it is required that these primitives @@ -207,317 +215,328 @@ the operation. It must be done such that all memory operations before and after the atomic operation calls are strongly ordered with respect to the atomic operation itself. -For example, it should behave as if a smp_mb() call existed both +For example, it should behave as if a ``smp_mb()`` call existed both before and after the atomic operation. If the atomic instructions used in an implementation provide explicit memory barrier semantics which satisfy the above requirements, that is fine as well. -Let's move on: +Let's move on:: - int atomic_add_return(int i, atomic_t *v); - int atomic_sub_return(int i, atomic_t *v); + int atomic_add_return(int i, atomic_t *v); + int atomic_sub_return(int i, atomic_t *v); -These behave just like atomic_{inc,dec}_return() except that an +These behave just like ``atomic_{inc,dec}_return()`` except that an explicit counter adjustment is given instead of the implicit "1". -This means that like atomic_{inc,dec}_return(), the memory barrier +This means that like ``atomic_{inc,dec}_return()``, the memory barrier semantics are required. -Next: +Next:: - int atomic_inc_and_test(atomic_t *v); - int atomic_dec_and_test(atomic_t *v); + int atomic_inc_and_test(atomic_t *v); + int atomic_dec_and_test(atomic_t *v); These two routines increment and decrement by 1, respectively, the given atomic counter. They return a boolean indicating whether the resulting counter value was zero or not. Again, these primitives provide explicit memory barrier semantics around -the atomic operation. +the atomic operation:: - int atomic_sub_and_test(int i, atomic_t *v); + int atomic_sub_and_test(int i, atomic_t *v); -This is identical to atomic_dec_and_test() except that an explicit +This is identical to ``atomic_dec_and_test()`` except that an explicit decrement is given instead of the implicit "1". This primitive must -provide explicit memory barrier semantics around the operation. +provide explicit memory barrier semantics around the operation:: - int atomic_add_negative(int i, atomic_t *v); + int atomic_add_negative(int i, atomic_t *v); The given increment is added to the given atomic counter value. A boolean is return which indicates whether the resulting counter value is negative. This primitive must provide explicit memory barrier semantics around the operation. -Then: +Then:: - int atomic_xchg(atomic_t *v, int new); + int atomic_xchg(atomic_t *v, int new); This performs an atomic exchange operation on the atomic variable v, setting the given new value. It returns the old value that the atomic variable v had just before the operation. -atomic_xchg must provide explicit memory barriers around the operation. +``atomic_xchg`` must provide explicit memory barriers around the operation:: - int atomic_cmpxchg(atomic_t *v, int old, int new); + int atomic_cmpxchg(atomic_t *v, int old, int new); This performs an atomic compare exchange operation on the atomic value v, -with the given old and new values. Like all atomic_xxx operations, -atomic_cmpxchg will only satisfy its atomicity semantics as long as all -other accesses of *v are performed through atomic_xxx operations. +with the given old and new values. Like all ``atomic_*`` operations, +``atomic_cmpxchg`` will only satisfy its atomicity semantics as long as all +other accesses of ``*v`` are performed through ``atomic_*`` operations. -atomic_cmpxchg must provide explicit memory barriers around the operation, +``atomic_cmpxchg`` must provide explicit memory barriers around the operation, although if the comparison fails then no memory ordering guarantees are required. -The semantics for atomic_cmpxchg are the same as those defined for 'cas' +The semantics for ``atomic_cmpxchg`` are the same as those defined for 'cas' below. -Finally: +Finally:: - int atomic_add_unless(atomic_t *v, int a, int u); + int atomic_add_unless(atomic_t *v, int a, int u); If the atomic value v is not equal to u, this function adds a to v, and returns non zero. If v is equal to u then it returns zero. This is done as an atomic operation. -atomic_add_unless must provide explicit memory barriers around the +``atomic_add_unless`` must provide explicit memory barriers around the operation unless it fails (returns 0). -atomic_inc_not_zero, equivalent to atomic_add_unless(v, 1, 0) +``atomic_inc_not_zero``, equivalent to ``atomic_add_unless(v, 1, 0)`` -If a caller requires memory barrier semantics around an atomic_t +If a caller requires memory barrier semantics around an ``atomic_t`` operation which does not return a value, a set of interfaces are -defined which accomplish this: +defined which accomplish this:: - void smp_mb__before_atomic(void); - void smp_mb__after_atomic(void); + void smp_mb__before_atomic(void); + void smp_mb__after_atomic(void); -For example, smp_mb__before_atomic() can be used like so: +For example, ``smp_mb__before_atomic()`` can be used like so:: - obj->dead = 1; - smp_mb__before_atomic(); - atomic_dec(&obj->ref_count); + obj->dead = 1; + smp_mb__before_atomic(); + atomic_dec(&obj->ref_count); -It makes sure that all memory operations preceding the atomic_dec() +It makes sure that all memory operations preceding the ``atomic_dec()`` call are strongly ordered with respect to the atomic counter operation. In the above example, it guarantees that the assignment of -"1" to obj->dead will be globally visible to other cpus before the +"1" to ``obj->dead`` will be globally visible to other cpus before the atomic counter decrement. -Without the explicit smp_mb__before_atomic() call, the +Without the explicit ``smp_mb__before_atomic()`` call, the implementation could legally allow the atomic counter update visible -to other cpus before the "obj->dead = 1;" assignment. +to other cpus before the ``obj->dead = 1;`` assignment. A missing memory barrier in the cases where they are required by the -atomic_t implementation above can have disastrous results. Here is +``atomic_t`` implementation above can have disastrous results. Here is an example, which follows a pattern occurring frequently in the Linux kernel. It is the use of atomic counters to implement reference counting, and it works such that once the counter falls to zero it can -be guaranteed that no other entity can be accessing the object: - -static void obj_list_add(struct obj *obj, struct list_head *head) -{ - obj->active = 1; - list_add(&obj->list, head); -} - -static void obj_list_del(struct obj *obj) -{ - list_del(&obj->list); - obj->active = 0; -} - -static void obj_destroy(struct obj *obj) -{ - BUG_ON(obj->active); - kfree(obj); -} - -struct obj *obj_list_peek(struct list_head *head) -{ - if (!list_empty(head)) { - struct obj *obj; - - obj = list_entry(head->next, struct obj, list); - atomic_inc(&obj->refcnt); - return obj; - } - return NULL; -} - -void obj_poke(void) -{ - struct obj *obj; - - spin_lock(&global_list_lock); - obj = obj_list_peek(&global_list); - spin_unlock(&global_list_lock); - - if (obj) { - obj->ops->poke(obj); - if (atomic_dec_and_test(&obj->refcnt)) - obj_destroy(obj); - } -} - -void obj_timeout(struct obj *obj) -{ - spin_lock(&global_list_lock); - obj_list_del(obj); - spin_unlock(&global_list_lock); - - if (atomic_dec_and_test(&obj->refcnt)) - obj_destroy(obj); -} - -(This is a simplification of the ARP queue management in the - generic neighbour discover code of the networking. Olaf Kirch - found a bug wrt. memory barriers in kfree_skb() that exposed - the atomic_t memory barrier requirements quite clearly.) - -Given the above scheme, it must be the case that the obj->active +be guaranteed that no other entity can be accessing the object:: + + static void obj_list_add(struct obj *obj, struct list_head *head) + { + obj->active = 1; + list_add(&obj->list, head); + } + + static void obj_list_del(struct obj *obj) + { + list_del(&obj->list); + obj->active = 0; + } + + static void obj_destroy(struct obj *obj) + { + BUG_ON(obj->active); + kfree(obj); + } + + struct obj *obj_list_peek(struct list_head *head) + { + if (!list_empty(head)) { + struct obj *obj; + + obj = list_entry(head->next, struct obj, list); + atomic_inc(&obj->refcnt); + return obj; + } + return NULL; + } + + void obj_poke(void) + { + struct obj *obj; + + spin_lock(&global_list_lock); + obj = obj_list_peek(&global_list); + spin_unlock(&global_list_lock); + + if (obj) { + obj->ops->poke(obj); + if (atomic_dec_and_test(&obj->refcnt)) + obj_destroy(obj); + } + } + + void obj_timeout(struct obj *obj) + { + spin_lock(&global_list_lock); + obj_list_del(obj); + spin_unlock(&global_list_lock); + + if (atomic_dec_and_test(&obj->refcnt)) + obj_destroy(obj); + } + +.. note:: + + This is a simplification of the ARP queue management in the + generic neighbour discover code of the networking. Olaf Kirch + found a bug wrt. memory barriers in kfree_skb() that exposed + the atomic_t memory barrier requirements quite clearly. + +Given the above scheme, it must be the case that the ``obj->active`` update done by the obj list deletion be visible to other processors before the atomic counter decrement is performed. -Otherwise, the counter could fall to zero, yet obj->active would still -be set, thus triggering the assertion in obj_destroy(). The error -sequence looks like this: - - cpu 0 cpu 1 - obj_poke() obj_timeout() - obj = obj_list_peek(); - ... gains ref to obj, refcnt=2 - obj_list_del(obj); - obj->active = 0 ... - ... visibility delayed ... - atomic_dec_and_test() - ... refcnt drops to 1 ... - atomic_dec_and_test() - ... refcount drops to 0 ... - obj_destroy() - BUG() triggers since obj->active - still seen as one - obj->active update visibility occurs +Otherwise, the counter could fall to zero, yet ``obj->active`` would still +be set, thus triggering the assertion in ``obj_destroy()``. The error +sequence looks like this:: + + cpu 0 cpu 1 + obj_poke() obj_timeout() + obj = obj_list_peek(); + ... gains ref to obj, refcnt=2 + obj_list_del(obj); + obj->active = 0 ... + ... visibility delayed ... + atomic_dec_and_test() + ... refcnt drops to 1 ... + atomic_dec_and_test() + ... refcount drops to 0 ... + obj_destroy() + BUG() triggers since obj->active + still seen as one + obj->active update visibility occurs With the memory barrier semantics required of the atomic_t operations which return values, the above sequence of memory visibility can never -happen. Specifically, in the above case the atomic_dec_and_test() +happen. Specifically, in the above case the ``atomic_dec_and_test()`` counter decrement would not become globally visible until the -obj->active update does. +``obj->active`` update does. + +.. note:: + + As a historical note, 32-bit Sparc used to only allow usage of + 24-bits of its ``atomic_t`` type. This was because it used 8 bits + as a spinlock for SMP safety. Sparc32 lacked a "compare and swap" + type instruction. However, 32-bit Sparc has since been moved over + to a "hash table of spinlocks" scheme, that allows the full 32-bit + counter to be realized. Essentially, an array of spinlocks are + indexed into based upon the address of the ``atomic_t`` being operated + on, and that lock protects the atomic operation. Parisc uses the + same scheme. -As a historical note, 32-bit Sparc used to only allow usage of -24-bits of its atomic_t type. This was because it used 8 bits -as a spinlock for SMP safety. Sparc32 lacked a "compare and swap" -type instruction. However, 32-bit Sparc has since been moved over -to a "hash table of spinlocks" scheme, that allows the full 32-bit -counter to be realized. Essentially, an array of spinlocks are -indexed into based upon the address of the atomic_t being operated -on, and that lock protects the atomic operation. Parisc uses the -same scheme. +.. note:: -Another note is that the atomic_t operations returning values are -extremely slow on an old 386. + Another note is that the ``atomic_t`` operations returning values are + extremely slow on an old 386. + +Atomic Bitmask +============== We will now cover the atomic bitmask operations. You will find that their SMP and memory barrier semantics are similar in shape and scope -to the atomic_t ops above. +to the ``atomic_t`` ops above. Native atomic bit operations are defined to operate on objects aligned -to the size of an "unsigned long" C data type, and are least of that -size. The endianness of the bits within each "unsigned long" are the -native endianness of the cpu. +to the size of an ``unsigned long`` C data type, and are least of that +size. The endianness of the bits within each ``unsigned long`` are the +native endianness of the cpu:: - void set_bit(unsigned long nr, volatile unsigned long *addr); - void clear_bit(unsigned long nr, volatile unsigned long *addr); - void change_bit(unsigned long nr, volatile unsigned long *addr); + void set_bit(unsigned long nr, volatile unsigned long *addr); + void clear_bit(unsigned long nr, volatile unsigned long *addr); + void change_bit(unsigned long nr, volatile unsigned long *addr); These routines set, clear, and change, respectively, the bit number -indicated by "nr" on the bit mask pointed to by "ADDR". +indicated by ``nr`` on the bit mask pointed to by ``addr``. They must execute atomically, yet there are no implicit memory barrier -semantics required of these interfaces. +semantics required of these interfaces:: - int test_and_set_bit(unsigned long nr, volatile unsigned long *addr); - int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); - int test_and_change_bit(unsigned long nr, volatile unsigned long *addr); + int test_and_set_bit(unsigned long nr, volatile unsigned long *addr); + int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); + int test_and_change_bit(unsigned long nr, volatile unsigned long *addr); Like the above, except that these routines return a boolean which indicates whether the changed bit was set _BEFORE_ the atomic bit operation. -WARNING! It is incredibly important that the value be a boolean, -ie. "0" or "1". Do not try to be fancy and save a few instructions by -declaring the above to return "long" and just returning something like -"old_val & mask" because that will not work. +.. warning:: + + It is incredibly important that the value be a boolean, + ie. "0" or "1". Do not try to be fancy and save a few instructions by + declaring the above to return "long" and just returning something like + "old_val & mask" because that will not work. -For one thing, this return value gets truncated to int in many code -paths using these interfaces, so on 64-bit if the bit is set in the -upper 32-bits then testers will never see that. + For one thing, this return value gets truncated to int in many code + paths using these interfaces, so on 64-bit if the bit is set in the + upper 32-bits then testers will never see that. -One great example of where this problem crops up are the thread_info -flag operations. Routines such as test_and_set_ti_thread_flag() chop -the return value into an int. There are other places where things -like this occur as well. + One great example of where this problem crops up are the ``thread_info`` + flag operations. Routines such as ``test_and_set_ti_thread_flag()`` chop + the return value into an int. There are other places where things + like this occur as well. -These routines, like the atomic_t counter operations returning values, +These routines, like the ``atomic_t`` counter operations returning values, must provide explicit memory barrier semantics around their execution. All memory operations before the atomic bit operation call must be made visible globally before the atomic bit operation is made visible. Likewise, the atomic bit operation must be visible globally before any -subsequent memory operation is made visible. For example: +subsequent memory operation is made visible. For example:: - obj->dead = 1; - if (test_and_set_bit(0, &obj->flags)) - /* ... */; - obj->killed = 1; + obj->dead = 1; + if (test_and_set_bit(0, &obj->flags)) + /* ... */; + obj->killed = 1; -The implementation of test_and_set_bit() must guarantee that -"obj->dead = 1;" is visible to cpus before the atomic memory operation -done by test_and_set_bit() becomes visible. Likewise, the atomic -memory operation done by test_and_set_bit() must become visible before -"obj->killed = 1;" is visible. +The implementation of ``test_and_set_bit()`` must guarantee that +``obj->dead = 1;`` is visible to cpus before the atomic memory operation +done by ``test_and_set_bit()`` becomes visible. Likewise, the atomic +memory operation done by ``test_and_set_bit()`` must become visible before +``obj->killed = 1;`` is visible. -Finally there is the basic operation: +Finally there is the basic operation:: - int test_bit(unsigned long nr, __const__ volatile unsigned long *addr); + int test_bit(unsigned long nr, __const__ volatile unsigned long *addr); -Which returns a boolean indicating if bit "nr" is set in the bitmask -pointed to by "addr". +Which returns a boolean indicating if bit ``nr`` is set in the bitmask +pointed to by ``addr``. -If explicit memory barriers are required around {set,clear}_bit() (which do +If explicit memory barriers are required around ``{set,clear}_bit()`` (which do not return a value, and thus does not need to provide memory barrier -semantics), two interfaces are provided: +semantics), two interfaces are provided:: - void smp_mb__before_atomic(void); - void smp_mb__after_atomic(void); + void smp_mb__before_atomic(void); + void smp_mb__after_atomic(void); -They are used as follows, and are akin to their atomic_t operation -brothers: +They are used as follows, and are akin to their ``atomic_t`` operation +brothers:: - /* All memory operations before this call will - * be globally visible before the clear_bit(). - */ - smp_mb__before_atomic(); - clear_bit( ... ); + /* All memory operations before this call will + * be globally visible before the clear_bit(). + */ + smp_mb__before_atomic(); + clear_bit( ... ); - /* The clear_bit() will be visible before all - * subsequent memory operations. - */ - smp_mb__after_atomic(); + /* The clear_bit() will be visible before all + * subsequent memory operations. + */ + smp_mb__after_atomic(); There are two special bitops with lock barrier semantics (acquire/release, same as spinlocks). These operate in the same way as their non-_lock/unlock postfixed variants, except that they are to provide acquire/release semantics, -respectively. This means they can be used for bit_spin_trylock and -bit_spin_unlock type operations without specifying any more barriers. +respectively. This means they can be used for ``bit_spin_trylock`` and +``bit_spin_unlock`` type operations without specifying any more barriers:: - int test_and_set_bit_lock(unsigned long nr, unsigned long *addr); - void clear_bit_unlock(unsigned long nr, unsigned long *addr); - void __clear_bit_unlock(unsigned long nr, unsigned long *addr); + int test_and_set_bit_lock(unsigned long nr, unsigned long *addr); + void clear_bit_unlock(unsigned long nr, unsigned long *addr); + void __clear_bit_unlock(unsigned long nr, unsigned long *addr); -The __clear_bit_unlock version is non-atomic, however it still implements +The ``__clear_bit_unlock`` version is non-atomic, however it still implements unlock barrier semantics. This can be useful if the lock itself is protecting the other bits in the word. @@ -526,41 +545,43 @@ provided. They are used in contexts where some other higher-level SMP locking scheme is being used to protect the bitmask, and thus less expensive non-atomic operations may be used in the implementation. They have names similar to the above bitmask operation interfaces, -except that two underscores are prefixed to the interface name. +except that two underscores are prefixed to the interface name:: - void __set_bit(unsigned long nr, volatile unsigned long *addr); - void __clear_bit(unsigned long nr, volatile unsigned long *addr); - void __change_bit(unsigned long nr, volatile unsigned long *addr); - int __test_and_set_bit(unsigned long nr, volatile unsigned long *addr); - int __test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); - int __test_and_change_bit(unsigned long nr, volatile unsigned long *addr); + void __set_bit(unsigned long nr, volatile unsigned long *addr); + void __clear_bit(unsigned long nr, volatile unsigned long *addr); + void __change_bit(unsigned long nr, volatile unsigned long *addr); + int __test_and_set_bit(unsigned long nr, volatile unsigned long *addr); + int __test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); + int __test_and_change_bit(unsigned long nr, volatile unsigned long *addr); These non-atomic variants also do not require any special memory barrier semantics. -The routines xchg() and cmpxchg() must provide the same exact +The routines ``xchg()`` and ``cmpxchg()`` must provide the same exact memory-barrier semantics as the atomic and bit operations returning values. -Note: If someone wants to use xchg(), cmpxchg() and their variants, -linux/atomic.h should be included rather than asm/cmpxchg.h, unless -the code is in arch/* and can take care of itself. +.. note:: + + If someone wants to use ``xchg()``, ``cmpxchg()`` and their variants, + ``linux/atomic.h`` should be included rather than ``asm/cmpxchg.h``, unless + the code is in arch/* and can take care of itself. Spinlocks and rwlocks have memory barrier expectations as well. The rule to follow is simple: -1) When acquiring a lock, the implementation must make it globally +1. When acquiring a lock, the implementation must make it globally visible before any subsequent memory operation. -2) When releasing a lock, the implementation must make it such that +2. When releasing a lock, the implementation must make it such that all previous memory operations are globally visible before the lock release. -Which finally brings us to _atomic_dec_and_lock(). There is an -architecture-neutral version implemented in lib/dec_and_lock.c, -but most platforms will wish to optimize this in assembler. +Which finally brings us to ``_atomic_dec_and_lock()``. There is an +architecture-neutral version implemented in ``lib/dec_and_lock.c``, +but most platforms will wish to optimize this in assembler:: - int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock); + int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock); Atomically decrement the given counter, and if will drop to zero atomically acquire the given spinlock and perform the decrement @@ -573,68 +594,70 @@ sure the spinlock operation is globally visible before any subsequent memory operation. We can demonstrate this operation more clearly if we define -an abstract atomic operation: +an abstract atomic operation:: - long cas(long *mem, long old, long new); + long cas(long *mem, long old, long new); -"cas" stands for "compare and swap". It atomically: +"``cas``" stands for "compare and swap". It atomically: -1) Compares "old" with the value currently at "mem". -2) If they are equal, "new" is written to "mem". -3) Regardless, the current value at "mem" is returned. +1. Compares "old" with the value currently at "mem". +2. If they are equal, "new" is written to "mem". +3. Regardless, the current value at "mem" is returned. As an example usage, here is what an atomic counter update -might look like: - -void example_atomic_inc(long *counter) -{ - long old, new, ret; - - while (1) { - old = *counter; - new = old + 1; - - ret = cas(counter, old, new); - if (ret == old) - break; - } -} - -Let's use cas() in order to build a pseudo-C atomic_dec_and_lock(): - -int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock) -{ - long old, new, ret; - int went_to_zero; - - went_to_zero = 0; - while (1) { - old = atomic_read(atomic); - new = old - 1; - if (new == 0) { - went_to_zero = 1; - spin_lock(lock); - } - ret = cas(atomic, old, new); - if (ret == old) - break; - if (went_to_zero) { - spin_unlock(lock); - went_to_zero = 0; - } - } - - return went_to_zero; -} - -Now, as far as memory barriers go, as long as spin_lock() +might look like:: + + void example_atomic_inc(long *counter) + { + long old, new, ret; + + while (1) { + old = *counter; + new = old + 1; + + ret = cas(counter, old, new); + if (ret == old) + break; + } + } + +Let's use ``cas()`` in order to build a pseudo-C ``atomic_dec_and_lock()``:: + + int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock) + { + long old, new, ret; + int went_to_zero; + + went_to_zero = 0; + while (1) { + old = atomic_read(atomic); + new = old - 1; + if (new == 0) { + went_to_zero = 1; + spin_lock(lock); + } + ret = cas(atomic, old, new); + if (ret == old) + break; + if (went_to_zero) { + spin_unlock(lock); + went_to_zero = 0; + } + } + + return went_to_zero; + } + +Now, as far as memory barriers go, as long as i``spin_lock()`` strictly orders all subsequent memory operations (including -the cas()) with respect to itself, things will be fine. +the ``cas()``) with respect to itself, things will be fine. -Said another way, _atomic_dec_and_lock() must guarantee that +Said another way, ``_atomic_dec_and_lock()`` must guarantee that a counter dropping to zero is never made visible before the spinlock being acquired. -Note that this also means that for the case where the counter -is not dropping to zero, there are no memory ordering -requirements. +.. note:: + + Note that this also means that for the case where the counter + is not dropping to zero, there are no memory ordering + requirements. diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index 480d9a3..f3e5f5e 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -8,6 +8,7 @@ Kernel and driver related documentation. :maxdepth: 1 assoc_array + atomic_ops workqueue .. only:: subproject diff --git a/Documentation/process/volatile-considered-harmful.rst b/Documentation/process/volatile-considered-harmful.rst index e0d042a..4934e65 100644 --- a/Documentation/process/volatile-considered-harmful.rst +++ b/Documentation/process/volatile-considered-harmful.rst @@ -1,3 +1,6 @@ + +.. _volatile_considered_harmful: + Why the "volatile" type class should not be used ------------------------------------------------ -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-25 14:59 ` [PATCH v3 2/4] Documentation/atomic_ops.txt: " Silvio Fricke @ 2016-11-25 21:01 ` Mauro Carvalho Chehab 2016-11-25 21:58 ` Peter Zijlstra 1 sibling, 0 replies; 51+ messages in thread From: Mauro Carvalho Chehab @ 2016-11-25 21:01 UTC (permalink / raw) To: Silvio Fricke Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula Em Fri, 25 Nov 2016 15:59:45 +0100 Silvio Fricke <silvio.fricke@gmail.com> escreveu: > ... and move to core-api folder. > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> > --- > Documentation/atomic_ops.txt => Documentation/core-api/atomic_ops.rst | 777 +++++++++++++++++++++++++++++++++++++----------------------------------- > Documentation/core-api/index.rst | 1 +- > Documentation/process/volatile-considered-harmful.rst | 3 +- > 3 files changed, 404 insertions(+), 377 deletions(-) > > diff --git a/Documentation/atomic_ops.txt b/Documentation/core-api/atomic_ops.rst > similarity index 46% > rename from Documentation/atomic_ops.txt > rename to Documentation/core-api/atomic_ops.rst > index 6c5e8a9..6236951 100644 > --- a/Documentation/atomic_ops.txt > +++ b/Documentation/core-api/atomic_ops.rst > @@ -1,204 +1,212 @@ > - Semantics and Behavior of Atomic and > - Bitmask Operations > +======================================================= > +Semantics and Behavior of Atomic and Bitmask Operations > +======================================================= > > - David S. Miller > +:Author: David S. Miller > > - This document is intended to serve as a guide to Linux port > +This document is intended to serve as a guide to Linux port > maintainers on how to implement atomic counter, bitops, and spinlock > interfaces properly. > > - The atomic_t type should be defined as a signed integer and > -the atomic_long_t type as a signed long integer. Also, they should > +Atomic Type And Operations > +========================== > + > +The ``atomic_t type`` should be defined as a signed integer and > +the ``atomic_long_t`` type as a signed long integer. Also, they should > be made opaque such that any kind of cast to a normal C integer type > -will fail. Something like the following should suffice: > +will fail. Something like the following should suffice:: > > - typedef struct { int counter; } atomic_t; > - typedef struct { long counter; } atomic_long_t; > + typedef struct { int counter; } atomic_t; > + typedef struct { long counter; } atomic_long_t; > > Historically, counter has been declared volatile. This is now discouraged. > -See Documentation/process/volatile-considered-harmful.rst for the complete rationale. > +See :ref:`Documentation/process/volatile-considered-harmful.rst > +<volatile_considered_harmful>` for the complete rationale. > > -local_t is very similar to atomic_t. If the counter is per CPU and only > -updated by one CPU, local_t is probably more appropriate. Please see > -Documentation/local_ops.txt for the semantics of local_t. > +``local_t`` is very similar to ``atomic_t``. If the counter is per CPU and only > +updated by one CPU, ``local_t`` is probably more appropriate. Please see > +:ref:`Documentation/local_ops.txt <local_ops>` for the semantics of ``local_t``. Hmm... you're renaming this file on the next patch. IMHO, the best would be to reorder patches 3 and 2 and place the new location on the above reference. In any case, the above reference should be to: :ref:`Documentation/core-api/local_ops.rst <local_ops>` After such change, you can add my review: Reviewed-by: Mauro Carvalho Chehab <mchehab@s-opensource.com> > > -The first operations to implement for atomic_t's are the initializers and > -plain reads. > +The first operations to implement for ``atomic_t``'s are the initializers and > +plain reads:: > > - #define ATOMIC_INIT(i) { (i) } > - #define atomic_set(v, i) ((v)->counter = (i)) > + #define ATOMIC_INIT(i) { (i) } > + #define atomic_set(v, i) ((v)->counter = (i)) > > -The first macro is used in definitions, such as: > +The first macro is used in definitions, such as:: > > -static atomic_t my_counter = ATOMIC_INIT(1); > + static atomic_t my_counter = ATOMIC_INIT(1); > > The initializer is atomic in that the return values of the atomic operations > are guaranteed to be correct reflecting the initialized value if the > initializer is used before runtime. If the initializer is used at runtime, a > proper implicit or explicit read memory barrier is needed before reading the > -value with atomic_read from another thread. > +value with ``atomic_read`` from another thread. > > -As with all of the atomic_ interfaces, replace the leading "atomic_" > -with "atomic_long_" to operate on atomic_long_t. > +As with all of the ``atomic_`` interfaces, replace the leading ``atomic_`` > +with ``atomic_long_`` to operate on ``atomic_long_t``. > > -The second interface can be used at runtime, as in: > +The second interface can be used at runtime, as in:: > > - struct foo { atomic_t counter; }; > - ... > + struct foo { atomic_t counter; }; > + ... > > - struct foo *k; > + struct foo *k; > > - k = kmalloc(sizeof(*k), GFP_KERNEL); > - if (!k) > - return -ENOMEM; > - atomic_set(&k->counter, 0); > + k = kmalloc(sizeof(*k), GFP_KERNEL); > + if (!k) > + return -ENOMEM; > + atomic_set(&k->counter, 0); > > The setting is atomic in that the return values of the atomic operations by > all threads are guaranteed to be correct reflecting either the value that has > been set with this operation or set with another operation. A proper implicit > or explicit memory barrier is needed before the value set with the operation > -is guaranteed to be readable with atomic_read from another thread. > +is guaranteed to be readable with ``atomic_read`` from another thread. > > -Next, we have: > +Next, we have:: > > - #define atomic_read(v) ((v)->counter) > + #define atomic_read(v) ((v)->counter) > > which simply reads the counter value currently visible to the calling thread. > The read is atomic in that the return value is guaranteed to be one of the > values initialized or modified with the interface operations if a proper > implicit or explicit memory barrier is used after possible runtime > initialization by any other thread and the value is modified only with the > -interface operations. atomic_read does not guarantee that the runtime > +interface operations. ``atomic_read`` does not guarantee that the runtime > initialization by any other thread is visible yet, so the user of the > interface must take care of that with a proper implicit or explicit memory > barrier. > > -*** WARNING: atomic_read() and atomic_set() DO NOT IMPLY BARRIERS! *** > +.. warning:: > > -Some architectures may choose to use the volatile keyword, barriers, or inline > -assembly to guarantee some degree of immediacy for atomic_read() and > -atomic_set(). This is not uniformly guaranteed, and may change in the future, > -so all users of atomic_t should treat atomic_read() and atomic_set() as simple > -C statements that may be reordered or optimized away entirely by the compiler > -or processor, and explicitly invoke the appropriate compiler and/or memory > -barrier for each use case. Failure to do so will result in code that may > -suddenly break when used with different architectures or compiler > -optimizations, or even changes in unrelated code which changes how the > -compiler optimizes the section accessing atomic_t variables. > + ``atomic_read()`` and ``atomic_set()`` DO NOT IMPLY BARRIERS! > > -*** YOU HAVE BEEN WARNED! *** > + Some architectures may choose to use the ``volatile`` keyword, barriers, or > + ``inline`` assembly to guarantee some degree of immediacy for ``atomic_read()`` > + and ``atomic_set()``. This is not uniformly guaranteed, and may change in the > + future, so all users of ``atomic_t`` should treat ``atomic_read()`` and > + ``atomic_set()`` as simple C statements that may be reordered or optimized away > + entirely by the compiler or processor, and explicitly invoke the appropriate > + compiler and/or memory barrier for each use case. Failure to do so will result > + in code that may suddenly break when used with different architectures or > + compiler optimizations, or even changes in unrelated code which changes how the > + compiler optimizes the section accessing ``atomic_t`` variables. > > Properly aligned pointers, longs, ints, and chars (and unsigned > equivalents) may be atomically loaded from and stored to in the same > -sense as described for atomic_read() and atomic_set(). The READ_ONCE() > -and WRITE_ONCE() macros should be used to prevent the compiler from using > -optimizations that might otherwise optimize accesses out of existence on > -the one hand, or that might create unsolicited accesses on the other. > +sense as described for ``atomic_read()`` and ``atomic_set()``. The > +``READ_ONCE()`` and ``WRITE_ONCE()`` macros should be used to prevent the > +compiler from using optimizations that might otherwise optimize accesses out of > +existence on the one hand, or that might create unsolicited accesses on the > +other. > > -For example consider the following code: > +For example consider the following code:: > > - while (a > 0) > - do_something(); > + while (a > 0) > + do_something(); > > -If the compiler can prove that do_something() does not store to the > +If the compiler can prove that ``do_something()`` does not store to the > variable a, then the compiler is within its rights transforming this to > -the following: > +the following:: > > - tmp = a; > - if (a > 0) > - for (;;) > - do_something(); > + tmp = a; > + if (a > 0) > + for (;;) > + do_something(); > > If you don't want the compiler to do this (and you probably don't), then > -you should use something like the following: > +you should use something like the following:: > > - while (READ_ONCE(a) < 0) > - do_something(); > + while (READ_ONCE(a) < 0) > + do_something(); > > -Alternatively, you could place a barrier() call in the loop. > +Alternatively, you could place a ``barrier()`` call in the loop. > > -For another example, consider the following code: > +For another example, consider the following code:: > > - tmp_a = a; > - do_something_with(tmp_a); > - do_something_else_with(tmp_a); > + tmp_a = a; > + do_something_with(tmp_a); > + do_something_else_with(tmp_a); > > -If the compiler can prove that do_something_with() does not store to the > -variable a, then the compiler is within its rights to manufacture an > -additional load as follows: > +If the compiler can prove that ``do_something_with()`` does not store to the > +variable ``a``, then the compiler is within its rights to manufacture an > +additional load as follows:: > > - tmp_a = a; > - do_something_with(tmp_a); > - tmp_a = a; > - do_something_else_with(tmp_a); > + tmp_a = a; > + do_something_with(tmp_a); > + tmp_a = a; > + do_something_else_with(tmp_a); > > This could fatally confuse your code if it expected the same value > -to be passed to do_something_with() and do_something_else_with(). > +to be passed to ``do_something_with()`` and ``do_something_else_with()``. > > The compiler would be likely to manufacture this additional load if > -do_something_with() was an inline function that made very heavy use > +``do_something_with()`` was an inline function that made very heavy use > of registers: reloading from variable a could save a flush to the > stack and later reload. To prevent the compiler from attacking your > -code in this manner, write the following: > +code in this manner, write the following:: > > - tmp_a = READ_ONCE(a); > - do_something_with(tmp_a); > - do_something_else_with(tmp_a); > + tmp_a = READ_ONCE(a); > + do_something_with(tmp_a); > + do_something_else_with(tmp_a); > > For a final example, consider the following code, assuming that the > -variable a is set at boot time before the second CPU is brought online > -and never changed later, so that memory barriers are not needed: > +variable ``a`` is set at boot time before the second CPU is brought online > +and never changed later, so that memory barriers are not needed:: > > - if (a) > - b = 9; > - else > - b = 42; > + if (a) > + b = 9; > + else > + b = 42; > > The compiler is within its rights to manufacture an additional store > -by transforming the above code into the following: > +by transforming the above code into the following:: > > - b = 42; > - if (a) > - b = 9; > + b = 42; > + if (a) > + b = 9; > > This could come as a fatal surprise to other code running concurrently > -that expected b to never have the value 42 if a was zero. To prevent > -the compiler from doing this, write something like: > +that expected ``b`` to never have the value ``42`` if ``a`` was zero. To > +prevent the compiler from doing this, write something like:: > + > + if (a) > + WRITE_ONCE(b, 9); > + else > + WRITE_ONCE(b, 42); > > - if (a) > - WRITE_ONCE(b, 9); > - else > - WRITE_ONCE(b, 42); > +Don't even **think** about doing this without proper use of memory barriers, > +locks, or atomic operations if variable ``a`` can change at runtime! > > -Don't even -think- about doing this without proper use of memory barriers, > -locks, or atomic operations if variable a can change at runtime! > +.. warning:: > > -*** WARNING: READ_ONCE() OR WRITE_ONCE() DO NOT IMPLY A BARRIER! *** > + ``READ_ONCE()`` OR ``WRITE_ONCE()`` DO NOT IMPLY A BARRIER! > > Now, we move onto the atomic operation interfaces typically implemented with > -the help of assembly code. > +the help of assembly code:: > > - void atomic_add(int i, atomic_t *v); > - void atomic_sub(int i, atomic_t *v); > - void atomic_inc(atomic_t *v); > - void atomic_dec(atomic_t *v); > + void atomic_add(int i, atomic_t *v); > + void atomic_sub(int i, atomic_t *v); > + void atomic_inc(atomic_t *v); > + void atomic_dec(atomic_t *v); > > These four routines add and subtract integral values to/from the given > -atomic_t value. The first two routines pass explicit integers by > +``atomic_t`` value. The first two routines pass explicit integers by > which to make the adjustment, whereas the latter two use an implicit > adjustment value of "1". > > -One very important aspect of these two routines is that they DO NOT > +One very important aspect of these two routines is that they **DO NOT** > require any explicit memory barriers. They need only perform the > -atomic_t counter update in an SMP safe manner. > +``atomic_t`` counter update in an SMP safe manner. > > -Next, we have: > +Next, we have:: > > - int atomic_inc_return(atomic_t *v); > - int atomic_dec_return(atomic_t *v); > + int atomic_inc_return(atomic_t *v); > + int atomic_dec_return(atomic_t *v); > > These routines add 1 and subtract 1, respectively, from the given > -atomic_t and return the new counter value after the operation is > +``atomic_t`` and return the new counter value after the operation is > performed. > > Unlike the above routines, it is required that these primitives > @@ -207,317 +215,328 @@ the operation. It must be done such that all memory operations before > and after the atomic operation calls are strongly ordered with respect > to the atomic operation itself. > > -For example, it should behave as if a smp_mb() call existed both > +For example, it should behave as if a ``smp_mb()`` call existed both > before and after the atomic operation. > > If the atomic instructions used in an implementation provide explicit > memory barrier semantics which satisfy the above requirements, that is > fine as well. > > -Let's move on: > +Let's move on:: > > - int atomic_add_return(int i, atomic_t *v); > - int atomic_sub_return(int i, atomic_t *v); > + int atomic_add_return(int i, atomic_t *v); > + int atomic_sub_return(int i, atomic_t *v); > > -These behave just like atomic_{inc,dec}_return() except that an > +These behave just like ``atomic_{inc,dec}_return()`` except that an > explicit counter adjustment is given instead of the implicit "1". > -This means that like atomic_{inc,dec}_return(), the memory barrier > +This means that like ``atomic_{inc,dec}_return()``, the memory barrier > semantics are required. > > -Next: > +Next:: > > - int atomic_inc_and_test(atomic_t *v); > - int atomic_dec_and_test(atomic_t *v); > + int atomic_inc_and_test(atomic_t *v); > + int atomic_dec_and_test(atomic_t *v); > > These two routines increment and decrement by 1, respectively, the > given atomic counter. They return a boolean indicating whether the > resulting counter value was zero or not. > > Again, these primitives provide explicit memory barrier semantics around > -the atomic operation. > +the atomic operation:: > > - int atomic_sub_and_test(int i, atomic_t *v); > + int atomic_sub_and_test(int i, atomic_t *v); > > -This is identical to atomic_dec_and_test() except that an explicit > +This is identical to ``atomic_dec_and_test()`` except that an explicit > decrement is given instead of the implicit "1". This primitive must > -provide explicit memory barrier semantics around the operation. > +provide explicit memory barrier semantics around the operation:: > > - int atomic_add_negative(int i, atomic_t *v); > + int atomic_add_negative(int i, atomic_t *v); > > The given increment is added to the given atomic counter value. A boolean > is return which indicates whether the resulting counter value is negative. > This primitive must provide explicit memory barrier semantics around > the operation. > > -Then: > +Then:: > > - int atomic_xchg(atomic_t *v, int new); > + int atomic_xchg(atomic_t *v, int new); > > This performs an atomic exchange operation on the atomic variable v, setting > the given new value. It returns the old value that the atomic variable v had > just before the operation. > > -atomic_xchg must provide explicit memory barriers around the operation. > +``atomic_xchg`` must provide explicit memory barriers around the operation:: > > - int atomic_cmpxchg(atomic_t *v, int old, int new); > + int atomic_cmpxchg(atomic_t *v, int old, int new); > > This performs an atomic compare exchange operation on the atomic value v, > -with the given old and new values. Like all atomic_xxx operations, > -atomic_cmpxchg will only satisfy its atomicity semantics as long as all > -other accesses of *v are performed through atomic_xxx operations. > +with the given old and new values. Like all ``atomic_*`` operations, > +``atomic_cmpxchg`` will only satisfy its atomicity semantics as long as all > +other accesses of ``*v`` are performed through ``atomic_*`` operations. > > -atomic_cmpxchg must provide explicit memory barriers around the operation, > +``atomic_cmpxchg`` must provide explicit memory barriers around the operation, > although if the comparison fails then no memory ordering guarantees are > required. > > -The semantics for atomic_cmpxchg are the same as those defined for 'cas' > +The semantics for ``atomic_cmpxchg`` are the same as those defined for 'cas' > below. > > -Finally: > +Finally:: > > - int atomic_add_unless(atomic_t *v, int a, int u); > + int atomic_add_unless(atomic_t *v, int a, int u); > > If the atomic value v is not equal to u, this function adds a to v, and > returns non zero. If v is equal to u then it returns zero. This is done as > an atomic operation. > > -atomic_add_unless must provide explicit memory barriers around the > +``atomic_add_unless`` must provide explicit memory barriers around the > operation unless it fails (returns 0). > > -atomic_inc_not_zero, equivalent to atomic_add_unless(v, 1, 0) > +``atomic_inc_not_zero``, equivalent to ``atomic_add_unless(v, 1, 0)`` > > > -If a caller requires memory barrier semantics around an atomic_t > +If a caller requires memory barrier semantics around an ``atomic_t`` > operation which does not return a value, a set of interfaces are > -defined which accomplish this: > +defined which accomplish this:: > > - void smp_mb__before_atomic(void); > - void smp_mb__after_atomic(void); > + void smp_mb__before_atomic(void); > + void smp_mb__after_atomic(void); > > -For example, smp_mb__before_atomic() can be used like so: > +For example, ``smp_mb__before_atomic()`` can be used like so:: > > - obj->dead = 1; > - smp_mb__before_atomic(); > - atomic_dec(&obj->ref_count); > + obj->dead = 1; > + smp_mb__before_atomic(); > + atomic_dec(&obj->ref_count); > > -It makes sure that all memory operations preceding the atomic_dec() > +It makes sure that all memory operations preceding the ``atomic_dec()`` > call are strongly ordered with respect to the atomic counter > operation. In the above example, it guarantees that the assignment of > -"1" to obj->dead will be globally visible to other cpus before the > +"1" to ``obj->dead`` will be globally visible to other cpus before the > atomic counter decrement. > > -Without the explicit smp_mb__before_atomic() call, the > +Without the explicit ``smp_mb__before_atomic()`` call, the > implementation could legally allow the atomic counter update visible > -to other cpus before the "obj->dead = 1;" assignment. > +to other cpus before the ``obj->dead = 1;`` assignment. > > A missing memory barrier in the cases where they are required by the > -atomic_t implementation above can have disastrous results. Here is > +``atomic_t`` implementation above can have disastrous results. Here is > an example, which follows a pattern occurring frequently in the Linux > kernel. It is the use of atomic counters to implement reference > counting, and it works such that once the counter falls to zero it can > -be guaranteed that no other entity can be accessing the object: > - > -static void obj_list_add(struct obj *obj, struct list_head *head) > -{ > - obj->active = 1; > - list_add(&obj->list, head); > -} > - > -static void obj_list_del(struct obj *obj) > -{ > - list_del(&obj->list); > - obj->active = 0; > -} > - > -static void obj_destroy(struct obj *obj) > -{ > - BUG_ON(obj->active); > - kfree(obj); > -} > - > -struct obj *obj_list_peek(struct list_head *head) > -{ > - if (!list_empty(head)) { > - struct obj *obj; > - > - obj = list_entry(head->next, struct obj, list); > - atomic_inc(&obj->refcnt); > - return obj; > - } > - return NULL; > -} > - > -void obj_poke(void) > -{ > - struct obj *obj; > - > - spin_lock(&global_list_lock); > - obj = obj_list_peek(&global_list); > - spin_unlock(&global_list_lock); > - > - if (obj) { > - obj->ops->poke(obj); > - if (atomic_dec_and_test(&obj->refcnt)) > - obj_destroy(obj); > - } > -} > - > -void obj_timeout(struct obj *obj) > -{ > - spin_lock(&global_list_lock); > - obj_list_del(obj); > - spin_unlock(&global_list_lock); > - > - if (atomic_dec_and_test(&obj->refcnt)) > - obj_destroy(obj); > -} > - > -(This is a simplification of the ARP queue management in the > - generic neighbour discover code of the networking. Olaf Kirch > - found a bug wrt. memory barriers in kfree_skb() that exposed > - the atomic_t memory barrier requirements quite clearly.) > - > -Given the above scheme, it must be the case that the obj->active > +be guaranteed that no other entity can be accessing the object:: > + > + static void obj_list_add(struct obj *obj, struct list_head *head) > + { > + obj->active = 1; > + list_add(&obj->list, head); > + } > + > + static void obj_list_del(struct obj *obj) > + { > + list_del(&obj->list); > + obj->active = 0; > + } > + > + static void obj_destroy(struct obj *obj) > + { > + BUG_ON(obj->active); > + kfree(obj); > + } > + > + struct obj *obj_list_peek(struct list_head *head) > + { > + if (!list_empty(head)) { > + struct obj *obj; > + > + obj = list_entry(head->next, struct obj, list); > + atomic_inc(&obj->refcnt); > + return obj; > + } > + return NULL; > + } > + > + void obj_poke(void) > + { > + struct obj *obj; > + > + spin_lock(&global_list_lock); > + obj = obj_list_peek(&global_list); > + spin_unlock(&global_list_lock); > + > + if (obj) { > + obj->ops->poke(obj); > + if (atomic_dec_and_test(&obj->refcnt)) > + obj_destroy(obj); > + } > + } > + > + void obj_timeout(struct obj *obj) > + { > + spin_lock(&global_list_lock); > + obj_list_del(obj); > + spin_unlock(&global_list_lock); > + > + if (atomic_dec_and_test(&obj->refcnt)) > + obj_destroy(obj); > + } > + > +.. note:: > + > + This is a simplification of the ARP queue management in the > + generic neighbour discover code of the networking. Olaf Kirch > + found a bug wrt. memory barriers in kfree_skb() that exposed > + the atomic_t memory barrier requirements quite clearly. > + > +Given the above scheme, it must be the case that the ``obj->active`` > update done by the obj list deletion be visible to other processors > before the atomic counter decrement is performed. > > -Otherwise, the counter could fall to zero, yet obj->active would still > -be set, thus triggering the assertion in obj_destroy(). The error > -sequence looks like this: > - > - cpu 0 cpu 1 > - obj_poke() obj_timeout() > - obj = obj_list_peek(); > - ... gains ref to obj, refcnt=2 > - obj_list_del(obj); > - obj->active = 0 ... > - ... visibility delayed ... > - atomic_dec_and_test() > - ... refcnt drops to 1 ... > - atomic_dec_and_test() > - ... refcount drops to 0 ... > - obj_destroy() > - BUG() triggers since obj->active > - still seen as one > - obj->active update visibility occurs > +Otherwise, the counter could fall to zero, yet ``obj->active`` would still > +be set, thus triggering the assertion in ``obj_destroy()``. The error > +sequence looks like this:: > + > + cpu 0 cpu 1 > + obj_poke() obj_timeout() > + obj = obj_list_peek(); > + ... gains ref to obj, refcnt=2 > + obj_list_del(obj); > + obj->active = 0 ... > + ... visibility delayed ... > + atomic_dec_and_test() > + ... refcnt drops to 1 ... > + atomic_dec_and_test() > + ... refcount drops to 0 ... > + obj_destroy() > + BUG() triggers since obj->active > + still seen as one > + obj->active update visibility occurs > > With the memory barrier semantics required of the atomic_t operations > which return values, the above sequence of memory visibility can never > -happen. Specifically, in the above case the atomic_dec_and_test() > +happen. Specifically, in the above case the ``atomic_dec_and_test()`` > counter decrement would not become globally visible until the > -obj->active update does. > +``obj->active`` update does. > + > +.. note:: > + > + As a historical note, 32-bit Sparc used to only allow usage of > + 24-bits of its ``atomic_t`` type. This was because it used 8 bits > + as a spinlock for SMP safety. Sparc32 lacked a "compare and swap" > + type instruction. However, 32-bit Sparc has since been moved over > + to a "hash table of spinlocks" scheme, that allows the full 32-bit > + counter to be realized. Essentially, an array of spinlocks are > + indexed into based upon the address of the ``atomic_t`` being operated > + on, and that lock protects the atomic operation. Parisc uses the > + same scheme. > > -As a historical note, 32-bit Sparc used to only allow usage of > -24-bits of its atomic_t type. This was because it used 8 bits > -as a spinlock for SMP safety. Sparc32 lacked a "compare and swap" > -type instruction. However, 32-bit Sparc has since been moved over > -to a "hash table of spinlocks" scheme, that allows the full 32-bit > -counter to be realized. Essentially, an array of spinlocks are > -indexed into based upon the address of the atomic_t being operated > -on, and that lock protects the atomic operation. Parisc uses the > -same scheme. > +.. note:: > > -Another note is that the atomic_t operations returning values are > -extremely slow on an old 386. > + Another note is that the ``atomic_t`` operations returning values are > + extremely slow on an old 386. > + > +Atomic Bitmask > +============== > > We will now cover the atomic bitmask operations. You will find that > their SMP and memory barrier semantics are similar in shape and scope > -to the atomic_t ops above. > +to the ``atomic_t`` ops above. > > Native atomic bit operations are defined to operate on objects aligned > -to the size of an "unsigned long" C data type, and are least of that > -size. The endianness of the bits within each "unsigned long" are the > -native endianness of the cpu. > +to the size of an ``unsigned long`` C data type, and are least of that > +size. The endianness of the bits within each ``unsigned long`` are the > +native endianness of the cpu:: > > - void set_bit(unsigned long nr, volatile unsigned long *addr); > - void clear_bit(unsigned long nr, volatile unsigned long *addr); > - void change_bit(unsigned long nr, volatile unsigned long *addr); > + void set_bit(unsigned long nr, volatile unsigned long *addr); > + void clear_bit(unsigned long nr, volatile unsigned long *addr); > + void change_bit(unsigned long nr, volatile unsigned long *addr); > > These routines set, clear, and change, respectively, the bit number > -indicated by "nr" on the bit mask pointed to by "ADDR". > +indicated by ``nr`` on the bit mask pointed to by ``addr``. > > They must execute atomically, yet there are no implicit memory barrier > -semantics required of these interfaces. > +semantics required of these interfaces:: > > - int test_and_set_bit(unsigned long nr, volatile unsigned long *addr); > - int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); > - int test_and_change_bit(unsigned long nr, volatile unsigned long *addr); > + int test_and_set_bit(unsigned long nr, volatile unsigned long *addr); > + int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); > + int test_and_change_bit(unsigned long nr, volatile unsigned long *addr); > > Like the above, except that these routines return a boolean which > indicates whether the changed bit was set _BEFORE_ the atomic bit > operation. > > -WARNING! It is incredibly important that the value be a boolean, > -ie. "0" or "1". Do not try to be fancy and save a few instructions by > -declaring the above to return "long" and just returning something like > -"old_val & mask" because that will not work. > +.. warning:: > + > + It is incredibly important that the value be a boolean, > + ie. "0" or "1". Do not try to be fancy and save a few instructions by > + declaring the above to return "long" and just returning something like > + "old_val & mask" because that will not work. > > -For one thing, this return value gets truncated to int in many code > -paths using these interfaces, so on 64-bit if the bit is set in the > -upper 32-bits then testers will never see that. > + For one thing, this return value gets truncated to int in many code > + paths using these interfaces, so on 64-bit if the bit is set in the > + upper 32-bits then testers will never see that. > > -One great example of where this problem crops up are the thread_info > -flag operations. Routines such as test_and_set_ti_thread_flag() chop > -the return value into an int. There are other places where things > -like this occur as well. > + One great example of where this problem crops up are the ``thread_info`` > + flag operations. Routines such as ``test_and_set_ti_thread_flag()`` chop > + the return value into an int. There are other places where things > + like this occur as well. > > -These routines, like the atomic_t counter operations returning values, > +These routines, like the ``atomic_t`` counter operations returning values, > must provide explicit memory barrier semantics around their execution. > All memory operations before the atomic bit operation call must be > made visible globally before the atomic bit operation is made visible. > Likewise, the atomic bit operation must be visible globally before any > -subsequent memory operation is made visible. For example: > +subsequent memory operation is made visible. For example:: > > - obj->dead = 1; > - if (test_and_set_bit(0, &obj->flags)) > - /* ... */; > - obj->killed = 1; > + obj->dead = 1; > + if (test_and_set_bit(0, &obj->flags)) > + /* ... */; > + obj->killed = 1; > > -The implementation of test_and_set_bit() must guarantee that > -"obj->dead = 1;" is visible to cpus before the atomic memory operation > -done by test_and_set_bit() becomes visible. Likewise, the atomic > -memory operation done by test_and_set_bit() must become visible before > -"obj->killed = 1;" is visible. > +The implementation of ``test_and_set_bit()`` must guarantee that > +``obj->dead = 1;`` is visible to cpus before the atomic memory operation > +done by ``test_and_set_bit()`` becomes visible. Likewise, the atomic > +memory operation done by ``test_and_set_bit()`` must become visible before > +``obj->killed = 1;`` is visible. > > -Finally there is the basic operation: > +Finally there is the basic operation:: > > - int test_bit(unsigned long nr, __const__ volatile unsigned long *addr); > + int test_bit(unsigned long nr, __const__ volatile unsigned long *addr); > > -Which returns a boolean indicating if bit "nr" is set in the bitmask > -pointed to by "addr". > +Which returns a boolean indicating if bit ``nr`` is set in the bitmask > +pointed to by ``addr``. > > -If explicit memory barriers are required around {set,clear}_bit() (which do > +If explicit memory barriers are required around ``{set,clear}_bit()`` (which do > not return a value, and thus does not need to provide memory barrier > -semantics), two interfaces are provided: > +semantics), two interfaces are provided:: > > - void smp_mb__before_atomic(void); > - void smp_mb__after_atomic(void); > + void smp_mb__before_atomic(void); > + void smp_mb__after_atomic(void); > > -They are used as follows, and are akin to their atomic_t operation > -brothers: > +They are used as follows, and are akin to their ``atomic_t`` operation > +brothers:: > > - /* All memory operations before this call will > - * be globally visible before the clear_bit(). > - */ > - smp_mb__before_atomic(); > - clear_bit( ... ); > + /* All memory operations before this call will > + * be globally visible before the clear_bit(). > + */ > + smp_mb__before_atomic(); > + clear_bit( ... ); > > - /* The clear_bit() will be visible before all > - * subsequent memory operations. > - */ > - smp_mb__after_atomic(); > + /* The clear_bit() will be visible before all > + * subsequent memory operations. > + */ > + smp_mb__after_atomic(); > > There are two special bitops with lock barrier semantics (acquire/release, > same as spinlocks). These operate in the same way as their non-_lock/unlock > postfixed variants, except that they are to provide acquire/release semantics, > -respectively. This means they can be used for bit_spin_trylock and > -bit_spin_unlock type operations without specifying any more barriers. > +respectively. This means they can be used for ``bit_spin_trylock`` and > +``bit_spin_unlock`` type operations without specifying any more barriers:: > > - int test_and_set_bit_lock(unsigned long nr, unsigned long *addr); > - void clear_bit_unlock(unsigned long nr, unsigned long *addr); > - void __clear_bit_unlock(unsigned long nr, unsigned long *addr); > + int test_and_set_bit_lock(unsigned long nr, unsigned long *addr); > + void clear_bit_unlock(unsigned long nr, unsigned long *addr); > + void __clear_bit_unlock(unsigned long nr, unsigned long *addr); > > -The __clear_bit_unlock version is non-atomic, however it still implements > +The ``__clear_bit_unlock`` version is non-atomic, however it still implements > unlock barrier semantics. This can be useful if the lock itself is protecting > the other bits in the word. > > @@ -526,41 +545,43 @@ provided. They are used in contexts where some other higher-level SMP > locking scheme is being used to protect the bitmask, and thus less > expensive non-atomic operations may be used in the implementation. > They have names similar to the above bitmask operation interfaces, > -except that two underscores are prefixed to the interface name. > +except that two underscores are prefixed to the interface name:: > > - void __set_bit(unsigned long nr, volatile unsigned long *addr); > - void __clear_bit(unsigned long nr, volatile unsigned long *addr); > - void __change_bit(unsigned long nr, volatile unsigned long *addr); > - int __test_and_set_bit(unsigned long nr, volatile unsigned long *addr); > - int __test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); > - int __test_and_change_bit(unsigned long nr, volatile unsigned long *addr); > + void __set_bit(unsigned long nr, volatile unsigned long *addr); > + void __clear_bit(unsigned long nr, volatile unsigned long *addr); > + void __change_bit(unsigned long nr, volatile unsigned long *addr); > + int __test_and_set_bit(unsigned long nr, volatile unsigned long *addr); > + int __test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); > + int __test_and_change_bit(unsigned long nr, volatile unsigned long *addr); > > These non-atomic variants also do not require any special memory > barrier semantics. > > -The routines xchg() and cmpxchg() must provide the same exact > +The routines ``xchg()`` and ``cmpxchg()`` must provide the same exact > memory-barrier semantics as the atomic and bit operations returning > values. > > -Note: If someone wants to use xchg(), cmpxchg() and their variants, > -linux/atomic.h should be included rather than asm/cmpxchg.h, unless > -the code is in arch/* and can take care of itself. > +.. note:: > + > + If someone wants to use ``xchg()``, ``cmpxchg()`` and their variants, > + ``linux/atomic.h`` should be included rather than ``asm/cmpxchg.h``, unless > + the code is in arch/* and can take care of itself. > > Spinlocks and rwlocks have memory barrier expectations as well. > The rule to follow is simple: > > -1) When acquiring a lock, the implementation must make it globally > +1. When acquiring a lock, the implementation must make it globally > visible before any subsequent memory operation. > > -2) When releasing a lock, the implementation must make it such that > +2. When releasing a lock, the implementation must make it such that > all previous memory operations are globally visible before the > lock release. > > -Which finally brings us to _atomic_dec_and_lock(). There is an > -architecture-neutral version implemented in lib/dec_and_lock.c, > -but most platforms will wish to optimize this in assembler. > +Which finally brings us to ``_atomic_dec_and_lock()``. There is an > +architecture-neutral version implemented in ``lib/dec_and_lock.c``, > +but most platforms will wish to optimize this in assembler:: > > - int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock); > + int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock); > > Atomically decrement the given counter, and if will drop to zero > atomically acquire the given spinlock and perform the decrement > @@ -573,68 +594,70 @@ sure the spinlock operation is globally visible before any > subsequent memory operation. > > We can demonstrate this operation more clearly if we define > -an abstract atomic operation: > +an abstract atomic operation:: > > - long cas(long *mem, long old, long new); > + long cas(long *mem, long old, long new); > > -"cas" stands for "compare and swap". It atomically: > +"``cas``" stands for "compare and swap". It atomically: > > -1) Compares "old" with the value currently at "mem". > -2) If they are equal, "new" is written to "mem". > -3) Regardless, the current value at "mem" is returned. > +1. Compares "old" with the value currently at "mem". > +2. If they are equal, "new" is written to "mem". > +3. Regardless, the current value at "mem" is returned. > > As an example usage, here is what an atomic counter update > -might look like: > - > -void example_atomic_inc(long *counter) > -{ > - long old, new, ret; > - > - while (1) { > - old = *counter; > - new = old + 1; > - > - ret = cas(counter, old, new); > - if (ret == old) > - break; > - } > -} > - > -Let's use cas() in order to build a pseudo-C atomic_dec_and_lock(): > - > -int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock) > -{ > - long old, new, ret; > - int went_to_zero; > - > - went_to_zero = 0; > - while (1) { > - old = atomic_read(atomic); > - new = old - 1; > - if (new == 0) { > - went_to_zero = 1; > - spin_lock(lock); > - } > - ret = cas(atomic, old, new); > - if (ret == old) > - break; > - if (went_to_zero) { > - spin_unlock(lock); > - went_to_zero = 0; > - } > - } > - > - return went_to_zero; > -} > - > -Now, as far as memory barriers go, as long as spin_lock() > +might look like:: > + > + void example_atomic_inc(long *counter) > + { > + long old, new, ret; > + > + while (1) { > + old = *counter; > + new = old + 1; > + > + ret = cas(counter, old, new); > + if (ret == old) > + break; > + } > + } > + > +Let's use ``cas()`` in order to build a pseudo-C ``atomic_dec_and_lock()``:: > + > + int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock) > + { > + long old, new, ret; > + int went_to_zero; > + > + went_to_zero = 0; > + while (1) { > + old = atomic_read(atomic); > + new = old - 1; > + if (new == 0) { > + went_to_zero = 1; > + spin_lock(lock); > + } > + ret = cas(atomic, old, new); > + if (ret == old) > + break; > + if (went_to_zero) { > + spin_unlock(lock); > + went_to_zero = 0; > + } > + } > + > + return went_to_zero; > + } > + > +Now, as far as memory barriers go, as long as i``spin_lock()`` > strictly orders all subsequent memory operations (including > -the cas()) with respect to itself, things will be fine. > +the ``cas()``) with respect to itself, things will be fine. > > -Said another way, _atomic_dec_and_lock() must guarantee that > +Said another way, ``_atomic_dec_and_lock()`` must guarantee that > a counter dropping to zero is never made visible before the > spinlock being acquired. > > -Note that this also means that for the case where the counter > -is not dropping to zero, there are no memory ordering > -requirements. > +.. note:: > + > + Note that this also means that for the case where the counter > + is not dropping to zero, there are no memory ordering > + requirements. > diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst > index 480d9a3..f3e5f5e 100644 > --- a/Documentation/core-api/index.rst > +++ b/Documentation/core-api/index.rst > @@ -8,6 +8,7 @@ Kernel and driver related documentation. > :maxdepth: 1 > > assoc_array > + atomic_ops > workqueue > > .. only:: subproject > diff --git a/Documentation/process/volatile-considered-harmful.rst b/Documentation/process/volatile-considered-harmful.rst > index e0d042a..4934e65 100644 > --- a/Documentation/process/volatile-considered-harmful.rst > +++ b/Documentation/process/volatile-considered-harmful.rst > @@ -1,3 +1,6 @@ > + > +.. _volatile_considered_harmful: > + > Why the "volatile" type class should not be used > ------------------------------------------------ > Thanks, Mauro ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-25 14:59 ` [PATCH v3 2/4] Documentation/atomic_ops.txt: " Silvio Fricke 2016-11-25 21:01 ` Mauro Carvalho Chehab @ 2016-11-25 21:58 ` Peter Zijlstra 2016-11-27 14:53 ` S. Fricke 2016-11-27 23:59 ` Jonathan Corbet 1 sibling, 2 replies; 51+ messages in thread From: Peter Zijlstra @ 2016-11-25 21:58 UTC (permalink / raw) To: Silvio Fricke Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula On Fri, Nov 25, 2016 at 03:59:45PM +0100, Silvio Fricke wrote: > ... and move to core-api folder. > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> > --- > Documentation/atomic_ops.txt => Documentation/core-api/atomic_ops.rst | 777 +++++++++++++++++++++++++++++++++++++----------------------------------- > Documentation/core-api/index.rst | 1 +- > Documentation/process/volatile-considered-harmful.rst | 3 +- > 3 files changed, 404 insertions(+), 377 deletions(-) Not a fan of this. The atomic_ops.txt file needs a lot of love, and I wouldn't want to edit a .rst file. Then again, I probably won't actually get around to fixing this document any time soon either. But if and when I would get around to it, I'll have to change it back to a regular .txt file. ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-25 21:58 ` Peter Zijlstra @ 2016-11-27 14:53 ` S. Fricke 2016-11-27 18:36 ` Jani Nikula 2016-11-27 23:59 ` Jonathan Corbet 1 sibling, 1 reply; 51+ messages in thread From: S. Fricke @ 2016-11-27 14:53 UTC (permalink / raw) To: Peter Zijlstra Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula Hi Peter, > On Fri, Nov 25, 2016 at 03:59:45PM +0100, Silvio Fricke wrote: > > ... and move to core-api folder. > > > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> > > --- > > Documentation/atomic_ops.txt => Documentation/core-api/atomic_ops.rst | 777 +++++++++++++++++++++++++++++++++++++----------------------------------- > > Documentation/core-api/index.rst | 1 +- > > Documentation/process/volatile-considered-harmful.rst | 3 +- > > 3 files changed, 404 insertions(+), 377 deletions(-) > > Not a fan of this. The atomic_ops.txt file needs a lot of love, and I > wouldn't want to edit a .rst file. What is the problem with this rst-file? atomic_ops.rst are not so different to the txt variant. I will drop this patch. Best regards, Silvio -- -- S. Fricke ---------------------------------------- silvio@port1024.net -- Diplom-Informatiker (FH) Linux-Development Matrix: @silvio:port1024.net ---------------------------------------------------------------------------- ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-27 14:53 ` S. Fricke @ 2016-11-27 18:36 ` Jani Nikula 0 siblings, 0 replies; 51+ messages in thread From: Jani Nikula @ 2016-11-27 18:36 UTC (permalink / raw) To: S. Fricke, Peter Zijlstra Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel On Sun, 27 Nov 2016, "S. Fricke" <silvio.fricke@gmail.com> wrote: > Hi Peter, > >> On Fri, Nov 25, 2016 at 03:59:45PM +0100, Silvio Fricke wrote: >> > ... and move to core-api folder. >> > >> > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> >> > --- >> > Documentation/atomic_ops.txt => Documentation/core-api/atomic_ops.rst | 777 +++++++++++++++++++++++++++++++++++++----------------------------------- >> > Documentation/core-api/index.rst | 1 +- >> > Documentation/process/volatile-considered-harmful.rst | 3 +- >> > 3 files changed, 404 insertions(+), 377 deletions(-) >> >> Not a fan of this. The atomic_ops.txt file needs a lot of love, and I >> wouldn't want to edit a .rst file. > > What is the problem with this rst-file? atomic_ops.rst are not so > different to the txt variant. There is nothing particularly wrong with the patch. Perhaps you could leave the tabs in place instead of indenting with spaces in the code blocks. It would result in a smaller diff. But other than that, it's fine. I'm sure Peter is capable of editing a file with a .rst extension, and we can clean up any hickups afterwards if getting the formatting right is insurmountable. > I will drop this patch. Please don't. Please let Jon be the judge of that. Thanks, Jani. -- Jani Nikula, Intel Open Source Technology Center ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-25 21:58 ` Peter Zijlstra 2016-11-27 14:53 ` S. Fricke @ 2016-11-27 23:59 ` Jonathan Corbet 2016-11-28 7:26 ` Peter Zijlstra 1 sibling, 1 reply; 51+ messages in thread From: Jonathan Corbet @ 2016-11-27 23:59 UTC (permalink / raw) To: Peter Zijlstra Cc: Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula On Fri, 25 Nov 2016 22:58:14 +0100 Peter Zijlstra <peterz@infradead.org> wrote: > Not a fan of this. The atomic_ops.txt file needs a lot of love, and I > wouldn't want to edit a .rst file. > > Then again, I probably won't actually get around to fixing this document > any time soon either. > > But if and when I would get around to it, I'll have to change it back to > a regular .txt file. Peter, could you please describe what the trouble with RST is? Most of the files in Documentation/ are already almost in RST format; should we change them to something else? :) Silvio, one thing that might help would be to redo the patch with a minimum of changes and, in particular, a minimum of unnecessary markup. Yes, it's better to put literal strings in monospace, but we don't have to, and getting rid of a lot of ``funny quotes`` makes the changes a lot less intrusive. jon ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-27 23:59 ` Jonathan Corbet @ 2016-11-28 7:26 ` Peter Zijlstra 2016-11-28 8:44 ` Daniel Vetter 0 siblings, 1 reply; 51+ messages in thread From: Peter Zijlstra @ 2016-11-28 7:26 UTC (permalink / raw) To: Jonathan Corbet Cc: Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula On Sun, Nov 27, 2016 at 04:59:14PM -0700, Jonathan Corbet wrote: > On Fri, 25 Nov 2016 22:58:14 +0100 > Peter Zijlstra <peterz@infradead.org> wrote: > > > Not a fan of this. The atomic_ops.txt file needs a lot of love, and I > > wouldn't want to edit a .rst file. > > > > Then again, I probably won't actually get around to fixing this document > > any time soon either. > > > > But if and when I would get around to it, I'll have to change it back to > > a regular .txt file. > > Peter, could you please describe what the trouble with RST is? Mostly that its not txt. I don't know what RST is, and I frankly don't care. For me changing this is 'make work', and I really don't need that. Changing this away from txt will simply make want to write documentation less, because it means having to figure out wtf RST is first, which means I'll simply won't do it. > Most of > the files in Documentation/ are already almost in RST format; should we > change them to something else? :) Why change them? What was wrong with txt to begin with? ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-28 7:26 ` Peter Zijlstra @ 2016-11-28 8:44 ` Daniel Vetter 2016-11-28 10:20 ` Peter Zijlstra 0 siblings, 1 reply; 51+ messages in thread From: Daniel Vetter @ 2016-11-28 8:44 UTC (permalink / raw) To: Peter Zijlstra Cc: Jonathan Corbet, Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula On Mon, Nov 28, 2016 at 08:26:26AM +0100, Peter Zijlstra wrote: > On Sun, Nov 27, 2016 at 04:59:14PM -0700, Jonathan Corbet wrote: > > On Fri, 25 Nov 2016 22:58:14 +0100 > > Peter Zijlstra <peterz@infradead.org> wrote: > > > > > Not a fan of this. The atomic_ops.txt file needs a lot of love, and I > > > wouldn't want to edit a .rst file. > > > > > > Then again, I probably won't actually get around to fixing this document > > > any time soon either. > > > > > > But if and when I would get around to it, I'll have to change it back to > > > a regular .txt file. I definitely don't want to make rst/sphinx a pain for anyone, so very much welcome your feedback. > > Peter, could you please describe what the trouble with RST is? > > Mostly that its not txt. I don't know what RST is, and I frankly don't > care. For me changing this is 'make work', and I really don't need that. > > Changing this away from txt will simply make want to write documentation > less, because it means having to figure out wtf RST is first, which > means I'll simply won't do it. I don't think you have to care at all. We've converted thousands of pages of entirely free-form .txt to .rst, and sphinx/rst happily parsed it all. There's the occasional warning, and 0day will report them. But if you don't want to care, then you can choose not to care at all here. The checkpatch/warning police is already picking up on these, so they will get fixed. And if that's too much, then you can also ignore those. > > Most of > > the files in Documentation/ are already almost in RST format; should we > > change them to something else? :) > > Why change them? What was wrong with txt to begin with? In my opinion good docs matter, and one of the key things is to be able to cross reference stuff. And I'd really like to be able to cross-reference to all the core kernel stuff (locks, atomics, workers, timers, all of these) from my driver/subsystem docs. This isn't something you or I ever really will use, but in my experience it is rather useful for people newly reading into a subsystem. And getting new folks on board (at least in subsystems like DRM where we're chronically understaffed for everything) matters. Given that I hope the change in file-ending from .txt to .rst is ok with you, and it won't stop you from updating the docs. If not that'd be bad, since imo anything that makes it harder to type good docs is bad. Another concern some core kernel folks raised is that the .rst markup was too heavy-handed, and makes the text much harder to read. Christoph called it "cat spew". That can be fixed with a much lighter-handed conversion (and 2nd patch iteration was acceptable for Christoph). In this patch the biggest part seems to be the indent shift on the code snippets, and it'd be good to keep those to tell sphinx to fixed-width layout them. But a lot of the other stuff really doesn't need to be if it bothers you. Thanks, Daniel -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-28 8:44 ` Daniel Vetter @ 2016-11-28 10:20 ` Peter Zijlstra 2016-11-28 11:08 ` Mauro Carvalho Chehab ` (2 more replies) 0 siblings, 3 replies; 51+ messages in thread From: Peter Zijlstra @ 2016-11-28 10:20 UTC (permalink / raw) To: Jonathan Corbet, Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula On Mon, Nov 28, 2016 at 09:44:42AM +0100, Daniel Vetter wrote: > > > > Why change them? What was wrong with txt to begin with? > > In my opinion good docs matter, and one of the key things is to be able to > cross reference stuff. Well, good docs begin with useful content; and many docs lack that. Fixing that would be a much more useful thing to do. In any case, I've never had any problems with typing things like: "go read: Documentation/file.txt for more information.". Also, what text editor supports cross references at all then? With the filename I can use 'gf' in vim to open it up in a new buffer and go read that. > Another concern some core kernel folks raised is that the .rst markup was > too heavy-handed, and makes the text much harder to read. Christoph called > it "cat spew". That can be fixed with a much lighter-handed conversion > (and 2nd patch iteration was acceptable for Christoph). Very much agreed, once a file is no longer readable with less or the text editor of your choice, it as good doesn't exist at all. So I very much worry about RST even supporting such heavy markup that the end result is unreadable. Basically, if a file isn't usable from within a 'normal' text editor, it doesn't exist. ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-28 10:20 ` Peter Zijlstra @ 2016-11-28 11:08 ` Mauro Carvalho Chehab 2016-11-28 11:54 ` Peter Zijlstra 2016-11-28 11:16 ` Jani Nikula 2016-11-28 14:13 ` Daniel Vetter 2 siblings, 1 reply; 51+ messages in thread From: Mauro Carvalho Chehab @ 2016-11-28 11:08 UTC (permalink / raw) To: Peter Zijlstra Cc: Jonathan Corbet, Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula Em Mon, 28 Nov 2016 11:20:09 +0100 Peter Zijlstra <peterz@infradead.org> escreveu: > On Mon, Nov 28, 2016 at 09:44:42AM +0100, Daniel Vetter wrote: > > > > > > > Why change them? What was wrong with txt to begin with? > > > > In my opinion good docs matter, and one of the key things is to be able to > > cross reference stuff. > > Well, good docs begin with useful content; and many docs lack that. > Fixing that would be a much more useful thing to do. > > In any case, I've never had any problems with typing things like: "go > read: Documentation/file.txt for more information.". > > Also, what text editor supports cross references at all then? With the > filename I can use 'gf' in vim to open it up in a new buffer and go read > that. > > > Another concern some core kernel folks raised is that the .rst markup was > > too heavy-handed, and makes the text much harder to read. Christoph called > > it "cat spew". That can be fixed with a much lighter-handed conversion > > (and 2nd patch iteration was acceptable for Christoph). > > Very much agreed, once a file is no longer readable with less or the > text editor of your choice, it as good doesn't exist at all. So I very > much worry about RST even supporting such heavy markup that the end > result is unreadable. > > Basically, if a file isn't usable from within a 'normal' text editor, it > doesn't exist. A file "converted" to ReST is still a text file. As Jani mentioned, you can keep doing your normal edit on them as text files, without even noticing that it follows the ReST notation. On most cases, all it takes to convert a file to ReST is: - adjust the chapter/section headers, as they all start on column 1, and all have a line below with a symbol (like "=", "-", "~", ...), E. g: Foo chapter =========== bar section ----------- It will automatically assign the first symbol to chapter, the next one to section and so on. - use *foo* (for italics) or **foo** (for bold) instead of _foo_; - if you need to use cross-references, use :ref:`name <id>`; The way I did cross-references on the documents I converted was to preserve the file name, plus adding a way for the ones reading the documentation in PDF or HTML to have the document cross-referenced. This way, we can still open the cross-referenced document with our favorite text editors, as the file name is there. E. g: :ref:`Documentation/admin-guide/serial-console.rst <serial_console>` - if it contains source examples or ascii artwork diagrams, use "::" on the previous line to create a literal block, e. g.: this is an example:: // whatever I do here, Sphinx will display as-is foo(bar); - adjust whitespaces/new lines where needed, as Sphinx identifies continuation lines if the text is at the same column as the previous line, and blank lines break paragraphs. - if you have something that you want to use a monotonic font on PDF/LaTeX/HTML, use ``foo``. - if you use special characters like '*' in the middle of the text (outsize a literal block), prepend with a \ in order to escape parsing it. Silvio's patch was a little too intrusive, as it touched on block indentation and whitespaces without needing. It also used monotonic fonts for a lot more than it would be required. I'm pretty sure his patch could be reworked to be a lot less intrusive. Btw, a lot of editors recognize ReST files and are capable of parsing the ReST tags while displaying the file, including vim and emacs, using different colors and/or bold to display those tags. So, even for text-editing, converting to ReST brings improvements. Thanks, Mauro ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-28 11:08 ` Mauro Carvalho Chehab @ 2016-11-28 11:54 ` Peter Zijlstra 2016-11-28 13:46 ` Daniel Vetter 0 siblings, 1 reply; 51+ messages in thread From: Peter Zijlstra @ 2016-11-28 11:54 UTC (permalink / raw) To: Mauro Carvalho Chehab Cc: Jonathan Corbet, Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula On Mon, Nov 28, 2016 at 09:08:55AM -0200, Mauro Carvalho Chehab wrote: > - use *foo* (for italics) or **foo** (for bold) instead of _foo_; That's daft, and also you're wrong. The normal convention is: /italic/ *bold* _underlined_ > :ref:`Documentation/admin-guide/serial-console.rst <serial_console>` That seems to work, as in 'gf' doesn't get confused by the spurious characters attached. Is there any way to enforce that? > - if you have something that you want to use a monotonic font on > PDF/LaTeX/HTML, use ``foo``. Bit weird, somewhere in the typewriter age they invented the " symbol so we didn't have to type double quotes anymore. > - if you use special characters like '*' in the middle of the text > (outsize a literal block), prepend with a \ in order to escape > parsing it. That's a bit inconvenient I suppose. > Btw, a lot of editors recognize ReST files and are capable of > parsing the ReST tags while displaying the file, including vim > and emacs, using different colors and/or bold to display those tags. > So, even for text-editing, converting to ReST brings improvements. Doesn't seem to really work though; if I open Documentation/development-process/1.Intro.rst the :ref:s at the start of the document don't work with the regular 'follow-ref' key combo of ^]. ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-28 11:54 ` Peter Zijlstra @ 2016-11-28 13:46 ` Daniel Vetter 2016-11-28 15:14 ` Mauro Carvalho Chehab 0 siblings, 1 reply; 51+ messages in thread From: Daniel Vetter @ 2016-11-28 13:46 UTC (permalink / raw) To: Peter Zijlstra Cc: Mauro Carvalho Chehab, Jonathan Corbet, Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula On Mon, Nov 28, 2016 at 12:54:13PM +0100, Peter Zijlstra wrote: > On Mon, Nov 28, 2016 at 09:08:55AM -0200, Mauro Carvalho Chehab wrote: > > - use *foo* (for italics) or **foo** (for bold) instead of _foo_; > > That's daft, and also you're wrong. The normal convention is: > > /italic/ > *bold* > _underlined_ I dont think anything is lost if we don't use the rst flavour but keep the traditional one. The html output meant for newbies looks a bit more funny, but if that means the old guard is more likely to type the docs then that's more than worth it. > > :ref:`Documentation/admin-guide/serial-console.rst <serial_console>` > > That seems to work, as in 'gf' doesn't get confused by the spurious > characters attached. Is there any way to enforce that? > > > - if you have something that you want to use a monotonic font on > > PDF/LaTeX/HTML, use ``foo``. > > Bit weird, somewhere in the typewriter age they invented the " symbol so > we didn't have to type double quotes anymore. "Quote" just gives you a quote, not fixed-width. The `` noise is what upset Christoph, and I think it's perfectly fine to not use them. > > - if you use special characters like '*' in the middle of the text > > (outsize a literal block), prepend with a \ in order to escape > > parsing it. > > That's a bit inconvenient I suppose. > > > Btw, a lot of editors recognize ReST files and are capable of > > parsing the ReST tags while displaying the file, including vim > > and emacs, using different colors and/or bold to display those tags. > > So, even for text-editing, converting to ReST brings improvements. > > Doesn't seem to really work though; if I open > Documentation/development-process/1.Intro.rst the :ref:s at the start of > the document don't work with the regular 'follow-ref' key combo of ^]. For that we need to integrate some rst/sphinx ctags generator into the kernel's ctags target. Not sure we want to do that (it might upset people who want to follow to a C symbol with clashing naming), at least not by default. But if this desired we can make it happen. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-28 13:46 ` Daniel Vetter @ 2016-11-28 15:14 ` Mauro Carvalho Chehab 0 siblings, 0 replies; 51+ messages in thread From: Mauro Carvalho Chehab @ 2016-11-28 15:14 UTC (permalink / raw) To: Daniel Vetter Cc: Peter Zijlstra, Jonathan Corbet, Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula Em Mon, 28 Nov 2016 14:46:45 +0100 Daniel Vetter <daniel@ffwll.ch> escreveu: > On Mon, Nov 28, 2016 at 12:54:13PM +0100, Peter Zijlstra wrote: > > On Mon, Nov 28, 2016 at 09:08:55AM -0200, Mauro Carvalho Chehab wrote: > > > - use *foo* (for italics) or **foo** (for bold) instead of _foo_; > > > > That's daft, and also you're wrong. The normal convention is: > > > > /italic/ > > *bold* > > _underlined_ > > I dont think anything is lost if we don't use the rst flavour but keep the > traditional one. The html output meant for newbies looks a bit more funny, > but if that means the old guard is more likely to type the docs then > that's more than worth it. Yep. Maybe we could later add a Sphinx extension that would add support for /italic/ and _underlined_ traditional convention. IMHO, it was a bad decision from Sphinx developers to not support underlined texts. > > > > :ref:`Documentation/admin-guide/serial-console.rst <serial_console>` > > > > That seems to work, as in 'gf' doesn't get confused by the spurious > > characters attached. Is there any way to enforce that? > > > > > - if you have something that you want to use a monotonic font on > > > PDF/LaTeX/HTML, use ``foo``. > > > > Bit weird, somewhere in the typewriter age they invented the " symbol so > > we didn't have to type double quotes anymore. > > "Quote" just gives you a quote, not fixed-width. The `` noise is what > upset Christoph, and I think it's perfectly fine to not use them. ``fixed-width`` is worse than "fixed-width", but on the other hand it won't require to add escape codes for " character, if used on some context where fixed width fonts is not required. So, at least for me, ``fixed-width`` is acceptable. Regards, Mauro ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-28 10:20 ` Peter Zijlstra 2016-11-28 11:08 ` Mauro Carvalho Chehab @ 2016-11-28 11:16 ` Jani Nikula 2016-11-28 11:56 ` Peter Zijlstra 2016-11-28 14:13 ` Daniel Vetter 2 siblings, 1 reply; 51+ messages in thread From: Jani Nikula @ 2016-11-28 11:16 UTC (permalink / raw) To: Peter Zijlstra, Jonathan Corbet, Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel On Mon, 28 Nov 2016, Peter Zijlstra <peterz@infradead.org> wrote: > On Mon, Nov 28, 2016 at 09:44:42AM +0100, Daniel Vetter wrote: > >> > >> > Why change them? What was wrong with txt to begin with? >> >> In my opinion good docs matter, and one of the key things is to be able to >> cross reference stuff. > > Well, good docs begin with useful content; and many docs lack that. > Fixing that would be a much more useful thing to do. > > In any case, I've never had any problems with typing things like: "go > read: Documentation/file.txt for more information.". Using rst we can produce decent HTML pages, and make them available at [1], in context. You don't have to read that, but it will be a lot more discoverable for other people, another important quality of good documentation. And perhaps you don't have to tell people to go read it so much. [1] https://www.kernel.org/doc/html/latest/ > Very much agreed, once a file is no longer readable with less or the > text editor of your choice, it as good doesn't exist at all. So I very > much worry about RST even supporting such heavy markup that the end > result is unreadable. The goal is to have the best of both worlds, keeping it pretty much plain text, but adding just enough consistency in formatting that you can generate other formats out of it. We don't have to and we shouldn't go overboard with the markup. Arguably you could call rst a "coding style" for plain text. We have pretty uniform C code, I don't think it's unreasonable to have a little bit of consistency in the plain text. And really, it's not much we're asking. BR, Jani. -- Jani Nikula, Intel Open Source Technology Center ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-28 11:16 ` Jani Nikula @ 2016-11-28 11:56 ` Peter Zijlstra 2016-11-28 12:32 ` Jani Nikula 0 siblings, 1 reply; 51+ messages in thread From: Peter Zijlstra @ 2016-11-28 11:56 UTC (permalink / raw) To: Jani Nikula Cc: Jonathan Corbet, Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel On Mon, Nov 28, 2016 at 01:16:45PM +0200, Jani Nikula wrote: > Using rst we can produce decent HTML pages, and make them available at > [1], in context. You don't have to read that, but it will be a lot more > discoverable for other people, another important quality of good > documentation. And perhaps you don't have to tell people to go read it > so much. > > [1] https://www.kernel.org/doc/html/latest/ *sigh*, basically if I have to touch a browser its broken. > > Very much agreed, once a file is no longer readable with less or the > > text editor of your choice, it as good doesn't exist at all. So I very > > much worry about RST even supporting such heavy markup that the end > > result is unreadable. > > The goal is to have the best of both worlds, keeping it pretty much > plain text, but adding just enough consistency in formatting that you > can generate other formats out of it. We don't have to and we shouldn't > go overboard with the markup. > > Arguably you could call rst a "coding style" for plain text. We have > pretty uniform C code, I don't think it's unreasonable to have a little > bit of consistency in the plain text. And really, it's not much we're > asking. With some decidedly daft conventions though; see my email to Mauro. ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-28 11:56 ` Peter Zijlstra @ 2016-11-28 12:32 ` Jani Nikula 0 siblings, 0 replies; 51+ messages in thread From: Jani Nikula @ 2016-11-28 12:32 UTC (permalink / raw) To: Peter Zijlstra Cc: Jonathan Corbet, Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel On Mon, 28 Nov 2016, Peter Zijlstra <peterz@infradead.org> wrote: > On Mon, Nov 28, 2016 at 01:16:45PM +0200, Jani Nikula wrote: >> Using rst we can produce decent HTML pages, and make them available at >> [1], in context. You don't have to read that, but it will be a lot more >> discoverable for other people, another important quality of good >> documentation. And perhaps you don't have to tell people to go read it >> so much. >> >> [1] https://www.kernel.org/doc/html/latest/ > > *sigh*, basically if I have to touch a browser its broken. Hey, I specifically said *you* don't have to... >> > Very much agreed, once a file is no longer readable with less or the >> > text editor of your choice, it as good doesn't exist at all. So I very >> > much worry about RST even supporting such heavy markup that the end >> > result is unreadable. >> >> The goal is to have the best of both worlds, keeping it pretty much >> plain text, but adding just enough consistency in formatting that you >> can generate other formats out of it. We don't have to and we shouldn't >> go overboard with the markup. >> >> Arguably you could call rst a "coding style" for plain text. We have >> pretty uniform C code, I don't think it's unreasonable to have a little >> bit of consistency in the plain text. And really, it's not much we're >> asking. > > With some decidedly daft conventions though; see my email to Mauro. -- Jani Nikula, Intel Open Source Technology Center ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-28 10:20 ` Peter Zijlstra 2016-11-28 11:08 ` Mauro Carvalho Chehab 2016-11-28 11:16 ` Jani Nikula @ 2016-11-28 14:13 ` Daniel Vetter 2016-11-28 14:29 ` Jonathan Corbet 2 siblings, 1 reply; 51+ messages in thread From: Daniel Vetter @ 2016-11-28 14:13 UTC (permalink / raw) To: Peter Zijlstra Cc: Jonathan Corbet, Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula Hi Peter, On Mon, Nov 28, 2016 at 11:20:09AM +0100, Peter Zijlstra wrote: > On Mon, Nov 28, 2016 at 09:44:42AM +0100, Daniel Vetter wrote: > > > Why change them? What was wrong with txt to begin with? > > > > In my opinion good docs matter, and one of the key things is to be able to > > cross reference stuff. > > Well, good docs begin with useful content; and many docs lack that. > Fixing that would be a much more useful thing to do. Fully agreed, pretty looking docs that don't exist aren't useful at all ;-) Personally I'm pretty happy with typing .rst plain-text, since I mostly ignore all the fancy stuff. Using .rst has also made the in-source kerneldoc a lot more useful, e.g. vtables can now be documented in a reasonable manner and the generated output still looks decent. For an example see include/drm/drm_modeset_helper_vtables.h. > In any case, I've never had any problems with typing things like: "go > read: Documentation/file.txt for more information.". > > Also, what text editor supports cross references at all then? With the > filename I can use 'gf' in vim to open it up in a new buffer and go read > that. Yeah agreed, anything that requires more work for typing docs isn't really useful. The nice thing about the kernel's sphinx toolchain is that a big pile of these references (not all of them yet) are autogenerated. That is of course of 0 use for old hats like us who just browse it all using vim and gf and ^] and all maybe a quickfix list of hits. But in my experience having something pretty to click around in is rather useful for newbies. At least that's been my experience with the drm docs, I have much less to explain on mails and chat since we've started with this all. And excessive amounts of cross-references seem to help a lot in guiding the blind ;-) Anyway, my goal at least is to keep all the plain-text usage perfectly fine, while giving newbies something pretty in there browsers. > > Another concern some core kernel folks raised is that the .rst markup was > > too heavy-handed, and makes the text much harder to read. Christoph called > > it "cat spew". That can be fixed with a much lighter-handed conversion > > (and 2nd patch iteration was acceptable for Christoph). > > Very much agreed, once a file is no longer readable with less or the > text editor of your choice, it as good doesn't exist at all. So I very > much worry about RST even supporting such heavy markup that the end > result is unreadable. > > Basically, if a file isn't usable from within a 'normal' text editor, it > doesn't exist. Yup agreed. Personally I prefer a _much_ more light-handed appraoch with rst markup. Essentially none, except the few things needed to glue all the various docs into a somewhat coherent whole. And I think that's also more-or-less the consensus among many core kernel hackers. Jon, should we document that we want a very light-handed approach to rst markup in kernel docs? This has come up a few times now, and irrespective of what exactly we're going to do with atomic_ops.txt I think it would help with making txt->rst conversions palatable to the core kernel community. And it's knida my own preference too ... Thanks, Daniel -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 2/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-28 14:13 ` Daniel Vetter @ 2016-11-28 14:29 ` Jonathan Corbet 0 siblings, 0 replies; 51+ messages in thread From: Jonathan Corbet @ 2016-11-28 14:29 UTC (permalink / raw) To: Daniel Vetter Cc: Peter Zijlstra, Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula On Mon, 28 Nov 2016 15:13:14 +0100 Daniel Vetter <daniel@ffwll.ch> wrote: > Jon, should we document that we want a very light-handed approach to rst > markup in kernel docs? This has come up a few times now, and irrespective > of what exactly we're going to do with atomic_ops.txt I think it would > help with making txt->rst conversions palatable to the core kernel > community. And it's knida my own preference too ... I think that makes sense, yes. We just have to decide what a light-handed approach actually is... jon ^ permalink raw reply [flat|nested] 51+ messages in thread
* [PATCH v3 3/4] Documentation/local_ops.txt: convert to ReST markup 2016-11-25 14:59 ` [PATCH v3 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 2016-11-25 14:59 ` [PATCH v3 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke 2016-11-25 14:59 ` [PATCH v3 2/4] Documentation/atomic_ops.txt: " Silvio Fricke @ 2016-11-25 14:59 ` Silvio Fricke 2016-11-25 21:10 ` Mauro Carvalho Chehab 2016-11-25 14:59 ` [PATCH v3 4/4] firmware: remove warning at documentation generation time Silvio Fricke 2016-11-28 17:30 ` [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 4 siblings, 1 reply; 51+ messages in thread From: Silvio Fricke @ 2016-11-25 14:59 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula, Silvio Fricke ... and move to core-api folder. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- Documentation/core-api/index.rst | 1 +- Documentation/local_ops.txt => Documentation/core-api/local_ops.rst | 273 +++---- 2 files changed, 145 insertions(+), 129 deletions(-) diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index f3e5f5e..25b4e4a 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -9,6 +9,7 @@ Kernel and driver related documentation. assoc_array atomic_ops + local_ops workqueue .. only:: subproject diff --git a/Documentation/local_ops.txt b/Documentation/core-api/local_ops.rst similarity index 57% rename from Documentation/local_ops.txt rename to Documentation/core-api/local_ops.rst index 407576a..1062ddb 100644 --- a/Documentation/local_ops.txt +++ b/Documentation/core-api/local_ops.rst @@ -1,191 +1,206 @@ - Semantics and Behavior of Local Atomic Operations - Mathieu Desnoyers +.. _local_ops: +================================================= +Semantics and Behavior of Local Atomic Operations +================================================= - This document explains the purpose of the local atomic operations, how +:Author: Mathieu Desnoyers + + +This document explains the purpose of the local atomic operations, how to implement them for any given architecture and shows how they can be used properly. It also stresses on the precautions that must be taken when reading those local variables across CPUs when the order of memory writes matters. -Note that local_t based operations are not recommended for general kernel use. -Please use the this_cpu operations instead unless there is really a special purpose. -Most uses of local_t in the kernel have been replaced by this_cpu operations. -this_cpu operations combine the relocation with the local_t like semantics in -a single instruction and yield more compact and faster executing code. +.. note:: + Note that ``local_t`` based operations are not recommended for general + kernel use. Please use the ``this_cpu`` operations instead unless there is + really a special purpose. Most uses of ``local_t`` in the kernel have been + replaced by ``this_cpu`` operations. ``this_cpu`` operations combine the + relocation with the ``local_t`` like semantics in a single instruction and + yield more compact and faster executing code. -* Purpose of local atomic operations + +Purpose of local atomic operations +================================== Local atomic operations are meant to provide fast and highly reentrant per CPU counters. They minimize the performance cost of standard atomic operations by removing the LOCK prefix and memory barriers normally required to synchronize across CPUs. -Having fast per CPU atomic counters is interesting in many cases : it does not +Having fast per CPU atomic counters is interesting in many cases: it does not require disabling interrupts to protect from interrupt handlers and it permits coherent counters in NMI handlers. It is especially useful for tracing purposes and for various performance monitoring counters. Local atomic operations only guarantee variable modification atomicity wrt the CPU which owns the data. Therefore, care must taken to make sure that only one -CPU writes to the local_t data. This is done by using per cpu data and making -sure that we modify it from within a preemption safe context. It is however -permitted to read local_t data from any CPU : it will then appear to be written -out of order wrt other memory writes by the owner CPU. +CPU writes to the ``local_t`` data. This is done by using per cpu data and +making sure that we modify it from within a preemption safe context. It is +however permitted to read ``local_t`` data from any CPU: it will then appear to +be written out of order wrt other memory writes by the owner CPU. -* Implementation for a given architecture +Implementation for a given architecture +======================================= -It can be done by slightly modifying the standard atomic operations : only +It can be done by slightly modifying the standard atomic operations: only their UP variant must be kept. It typically means removing LOCK prefix (on i386 and x86_64) and any SMP synchronization barrier. If the architecture does -not have a different behavior between SMP and UP, including asm-generic/local.h -in your architecture's local.h is sufficient. +not have a different behavior between SMP and UP, including +``asm-generic/local.h`` in your architecture's ``local.h`` is sufficient. -The local_t type is defined as an opaque signed long by embedding an -atomic_long_t inside a structure. This is made so a cast from this type to a -long fails. The definition looks like : +The ``local_t`` type is defined as an opaque ``signed long`` by embedding an +``atomic_long_t`` inside a structure. This is made so a cast from this type to +a ``long`` fails. The definition looks like:: -typedef struct { atomic_long_t a; } local_t; + typedef struct { atomic_long_t a; } local_t; -* Rules to follow when using local atomic operations +Rules to follow when using local atomic operations +================================================== -- Variables touched by local ops must be per cpu variables. -- _Only_ the CPU owner of these variables must write to them. -- This CPU can use local ops from any context (process, irq, softirq, nmi, ...) - to update its local_t variables. -- Preemption (or interrupts) must be disabled when using local ops in - process context to make sure the process won't be migrated to a +* Variables touched by local ops must be per cpu variables. +* *Only* the CPU owner of these variables must write to them. +* This CPU can use local ops from any context (process, irq, softirq, nmi, ...) + to update its ``local_t`` variables. +* Preemption (or interrupts) must be disabled when using local ops in + process context to make sure the process won't be migrated to a different CPU between getting the per-cpu variable and doing the actual local op. -- When using local ops in interrupt context, no special care must be +* When using local ops in interrupt context, no special care must be taken on a mainline kernel, since they will run on the local CPU with preemption already disabled. I suggest, however, to explicitly disable preemption anyway to make sure it will still work correctly on -rt kernels. -- Reading the local cpu variable will provide the current copy of the +* Reading the local cpu variable will provide the current copy of the variable. -- Reads of these variables can be done from any CPU, because updates to - "long", aligned, variables are always atomic. Since no memory +* Reads of these variables can be done from any CPU, because updates to + "``long``", aligned, variables are always atomic. Since no memory synchronization is done by the writer CPU, an outdated copy of the - variable can be read when reading some _other_ cpu's variables. + variable can be read when reading some *other* cpu's variables. + +How to use local atomic operations +================================== -* How to use local atomic operations +:: -#include <linux/percpu.h> -#include <asm/local.h> + #include <linux/percpu.h> + #include <asm/local.h> -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); -* Counting +Counting +======== Counting is done on all the bits of a signed long. -In preemptible context, use get_cpu_var() and put_cpu_var() around local atomic -operations : it makes sure that preemption is disabled around write access to -the per cpu variable. For instance : +In preemptible context, use ``get_cpu_var()`` and ``put_cpu_var()`` around +local atomic operations: it makes sure that preemption is disabled around write +access to the per cpu variable. For instance:: - local_inc(&get_cpu_var(counters)); - put_cpu_var(counters); + local_inc(&get_cpu_var(counters)); + put_cpu_var(counters); If you are already in a preemption-safe context, you can use -this_cpu_ptr() instead. +``this_cpu_ptr()`` instead:: - local_inc(this_cpu_ptr(&counters)); + local_inc(this_cpu_ptr(&counters)); -* Reading the counters +Reading the counters +==================== Those local counters can be read from foreign CPUs to sum the count. Note that the data seen by local_read across CPUs must be considered to be out of order -relatively to other memory writes happening on the CPU that owns the data. +relatively to other memory writes happening on the CPU that owns the data:: - long sum = 0; - for_each_online_cpu(cpu) - sum += local_read(&per_cpu(counters, cpu)); + long sum = 0; + for_each_online_cpu(cpu) + sum += local_read(&per_cpu(counters, cpu)); If you want to use a remote local_read to synchronize access to a resource -between CPUs, explicit smp_wmb() and smp_rmb() memory barriers must be used +between CPUs, explicit ``smp_wmb()`` and ``smp_rmb()`` memory barriers must be used respectively on the writer and the reader CPUs. It would be the case if you use -the local_t variable as a counter of bytes written in a buffer : there should -be a smp_wmb() between the buffer write and the counter increment and also a -smp_rmb() between the counter read and the buffer read. - - -Here is a sample module which implements a basic per cpu counter using local.h. - ---- BEGIN --- -/* test-local.c - * - * Sample module for local.h usage. - */ - - -#include <asm/local.h> -#include <linux/module.h> -#include <linux/timer.h> - -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); - -static struct timer_list test_timer; - -/* IPI called on each CPU. */ -static void test_each(void *info) -{ - /* Increment the counter from a non preemptible context */ - printk("Increment on cpu %d\n", smp_processor_id()); - local_inc(this_cpu_ptr(&counters)); - - /* This is what incrementing the variable would look like within a - * preemptible context (it disables preemption) : - * - * local_inc(&get_cpu_var(counters)); - * put_cpu_var(counters); - */ -} - -static void do_test_timer(unsigned long data) -{ - int cpu; - - /* Increment the counters */ - on_each_cpu(test_each, NULL, 1); - /* Read all the counters */ - printk("Counters read from CPU %d\n", smp_processor_id()); - for_each_online_cpu(cpu) { - printk("Read : CPU %d, count %ld\n", cpu, - local_read(&per_cpu(counters, cpu))); - } - del_timer(&test_timer); - test_timer.expires = jiffies + 1000; - add_timer(&test_timer); -} - -static int __init test_init(void) -{ - /* initialize the timer that will increment the counter */ - init_timer(&test_timer); - test_timer.function = do_test_timer; - test_timer.expires = jiffies + 1; - add_timer(&test_timer); - - return 0; -} - -static void __exit test_exit(void) -{ - del_timer_sync(&test_timer); -} - -module_init(test_init); -module_exit(test_exit); - -MODULE_LICENSE("GPL"); -MODULE_AUTHOR("Mathieu Desnoyers"); -MODULE_DESCRIPTION("Local Atomic Ops"); ---- END --- +the ``local_t`` variable as a counter of bytes written in a buffer: there should +be a ``smp_wmb()`` between the buffer write and the counter increment and also a +``smp_rmb()`` between the counter read and the buffer read. + + +Here is a sample module which implements a basic per cpu counter using +``local.h``:: + + /* test-local.c + * + * Sample module for local.h usage. + */ + + + #include <asm/local.h> + #include <linux/module.h> + #include <linux/timer.h> + + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); + + static struct timer_list test_timer; + + /* IPI called on each CPU. */ + static void test_each(void *info) + { + /* Increment the counter from a non preemptible context */ + printk("Increment on cpu %d\n", smp_processor_id()); + local_inc(this_cpu_ptr(&counters)); + + /* This is what incrementing the variable would look like within a + * preemptible context (it disables preemption) : + * + * local_inc(&get_cpu_var(counters)); + * put_cpu_var(counters); + */ + } + + static void do_test_timer(unsigned long data) + { + int cpu; + + /* Increment the counters */ + on_each_cpu(test_each, NULL, 1); + /* Read all the counters */ + printk("Counters read from CPU %d\n", smp_processor_id()); + for_each_online_cpu(cpu) { + printk("Read : CPU %d, count %ld\n", cpu, + local_read(&per_cpu(counters, cpu))); + } + del_timer(&test_timer); + test_timer.expires = jiffies + 1000; + add_timer(&test_timer); + } + + static int __init test_init(void) + { + /* initialize the timer that will increment the counter */ + init_timer(&test_timer); + test_timer.function = do_test_timer; + test_timer.expires = jiffies + 1; + add_timer(&test_timer); + + return 0; + } + + static void __exit test_exit(void) + { + del_timer_sync(&test_timer); + } + + module_init(test_init); + module_exit(test_exit); + + MODULE_LICENSE("GPL"); + MODULE_AUTHOR("Mathieu Desnoyers"); + MODULE_DESCRIPTION("Local Atomic Ops"); -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* Re: [PATCH v3 3/4] Documentation/local_ops.txt: convert to ReST markup 2016-11-25 14:59 ` [PATCH v3 3/4] Documentation/local_ops.txt: " Silvio Fricke @ 2016-11-25 21:10 ` Mauro Carvalho Chehab 0 siblings, 0 replies; 51+ messages in thread From: Mauro Carvalho Chehab @ 2016-11-25 21:10 UTC (permalink / raw) To: Silvio Fricke Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula Em Fri, 25 Nov 2016 15:59:46 +0100 Silvio Fricke <silvio.fricke@gmail.com> escreveu: > ... and move to core-api folder. > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> Reviewed-by: Mauro Carvalho Chehab <mchehab@s-opensource.com> > --- > Documentation/core-api/index.rst | 1 +- > Documentation/local_ops.txt => Documentation/core-api/local_ops.rst | 273 +++---- > 2 files changed, 145 insertions(+), 129 deletions(-) > > diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst > index f3e5f5e..25b4e4a 100644 > --- a/Documentation/core-api/index.rst > +++ b/Documentation/core-api/index.rst > @@ -9,6 +9,7 @@ Kernel and driver related documentation. > > assoc_array > atomic_ops > + local_ops > workqueue > > .. only:: subproject > diff --git a/Documentation/local_ops.txt b/Documentation/core-api/local_ops.rst > similarity index 57% > rename from Documentation/local_ops.txt > rename to Documentation/core-api/local_ops.rst > index 407576a..1062ddb 100644 > --- a/Documentation/local_ops.txt > +++ b/Documentation/core-api/local_ops.rst > @@ -1,191 +1,206 @@ > - Semantics and Behavior of Local Atomic Operations > > - Mathieu Desnoyers > +.. _local_ops: > > +================================================= > +Semantics and Behavior of Local Atomic Operations > +================================================= > > - This document explains the purpose of the local atomic operations, how > +:Author: Mathieu Desnoyers > + > + > +This document explains the purpose of the local atomic operations, how > to implement them for any given architecture and shows how they can be used > properly. It also stresses on the precautions that must be taken when reading > those local variables across CPUs when the order of memory writes matters. > > -Note that local_t based operations are not recommended for general kernel use. > -Please use the this_cpu operations instead unless there is really a special purpose. > -Most uses of local_t in the kernel have been replaced by this_cpu operations. > -this_cpu operations combine the relocation with the local_t like semantics in > -a single instruction and yield more compact and faster executing code. > +.. note:: > > + Note that ``local_t`` based operations are not recommended for general > + kernel use. Please use the ``this_cpu`` operations instead unless there is > + really a special purpose. Most uses of ``local_t`` in the kernel have been > + replaced by ``this_cpu`` operations. ``this_cpu`` operations combine the > + relocation with the ``local_t`` like semantics in a single instruction and > + yield more compact and faster executing code. > > -* Purpose of local atomic operations > + > +Purpose of local atomic operations > +================================== > > Local atomic operations are meant to provide fast and highly reentrant per CPU > counters. They minimize the performance cost of standard atomic operations by > removing the LOCK prefix and memory barriers normally required to synchronize > across CPUs. > > -Having fast per CPU atomic counters is interesting in many cases : it does not > +Having fast per CPU atomic counters is interesting in many cases: it does not > require disabling interrupts to protect from interrupt handlers and it permits > coherent counters in NMI handlers. It is especially useful for tracing purposes > and for various performance monitoring counters. > > Local atomic operations only guarantee variable modification atomicity wrt the > CPU which owns the data. Therefore, care must taken to make sure that only one > -CPU writes to the local_t data. This is done by using per cpu data and making > -sure that we modify it from within a preemption safe context. It is however > -permitted to read local_t data from any CPU : it will then appear to be written > -out of order wrt other memory writes by the owner CPU. > +CPU writes to the ``local_t`` data. This is done by using per cpu data and > +making sure that we modify it from within a preemption safe context. It is > +however permitted to read ``local_t`` data from any CPU: it will then appear to > +be written out of order wrt other memory writes by the owner CPU. > > > -* Implementation for a given architecture > +Implementation for a given architecture > +======================================= > > -It can be done by slightly modifying the standard atomic operations : only > +It can be done by slightly modifying the standard atomic operations: only > their UP variant must be kept. It typically means removing LOCK prefix (on > i386 and x86_64) and any SMP synchronization barrier. If the architecture does > -not have a different behavior between SMP and UP, including asm-generic/local.h > -in your architecture's local.h is sufficient. > +not have a different behavior between SMP and UP, including > +``asm-generic/local.h`` in your architecture's ``local.h`` is sufficient. > > -The local_t type is defined as an opaque signed long by embedding an > -atomic_long_t inside a structure. This is made so a cast from this type to a > -long fails. The definition looks like : > +The ``local_t`` type is defined as an opaque ``signed long`` by embedding an > +``atomic_long_t`` inside a structure. This is made so a cast from this type to > +a ``long`` fails. The definition looks like:: > > -typedef struct { atomic_long_t a; } local_t; > + typedef struct { atomic_long_t a; } local_t; > > > -* Rules to follow when using local atomic operations > +Rules to follow when using local atomic operations > +================================================== > > -- Variables touched by local ops must be per cpu variables. > -- _Only_ the CPU owner of these variables must write to them. > -- This CPU can use local ops from any context (process, irq, softirq, nmi, ...) > - to update its local_t variables. > -- Preemption (or interrupts) must be disabled when using local ops in > - process context to make sure the process won't be migrated to a > +* Variables touched by local ops must be per cpu variables. > +* *Only* the CPU owner of these variables must write to them. > +* This CPU can use local ops from any context (process, irq, softirq, nmi, ...) > + to update its ``local_t`` variables. > +* Preemption (or interrupts) must be disabled when using local ops in > + process context to make sure the process won't be migrated to a > different CPU between getting the per-cpu variable and doing the > actual local op. > -- When using local ops in interrupt context, no special care must be > +* When using local ops in interrupt context, no special care must be > taken on a mainline kernel, since they will run on the local CPU with > preemption already disabled. I suggest, however, to explicitly > disable preemption anyway to make sure it will still work correctly on > -rt kernels. > -- Reading the local cpu variable will provide the current copy of the > +* Reading the local cpu variable will provide the current copy of the > variable. > -- Reads of these variables can be done from any CPU, because updates to > - "long", aligned, variables are always atomic. Since no memory > +* Reads of these variables can be done from any CPU, because updates to > + "``long``", aligned, variables are always atomic. Since no memory > synchronization is done by the writer CPU, an outdated copy of the > - variable can be read when reading some _other_ cpu's variables. > + variable can be read when reading some *other* cpu's variables. > + > > +How to use local atomic operations > +================================== > > -* How to use local atomic operations > +:: > > -#include <linux/percpu.h> > -#include <asm/local.h> > + #include <linux/percpu.h> > + #include <asm/local.h> > > -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > > > -* Counting > +Counting > +======== > > Counting is done on all the bits of a signed long. > > -In preemptible context, use get_cpu_var() and put_cpu_var() around local atomic > -operations : it makes sure that preemption is disabled around write access to > -the per cpu variable. For instance : > +In preemptible context, use ``get_cpu_var()`` and ``put_cpu_var()`` around > +local atomic operations: it makes sure that preemption is disabled around write > +access to the per cpu variable. For instance:: > > - local_inc(&get_cpu_var(counters)); > - put_cpu_var(counters); > + local_inc(&get_cpu_var(counters)); > + put_cpu_var(counters); > > If you are already in a preemption-safe context, you can use > -this_cpu_ptr() instead. > +``this_cpu_ptr()`` instead:: > > - local_inc(this_cpu_ptr(&counters)); > + local_inc(this_cpu_ptr(&counters)); > > > > -* Reading the counters > +Reading the counters > +==================== > > Those local counters can be read from foreign CPUs to sum the count. Note that > the data seen by local_read across CPUs must be considered to be out of order > -relatively to other memory writes happening on the CPU that owns the data. > +relatively to other memory writes happening on the CPU that owns the data:: > > - long sum = 0; > - for_each_online_cpu(cpu) > - sum += local_read(&per_cpu(counters, cpu)); > + long sum = 0; > + for_each_online_cpu(cpu) > + sum += local_read(&per_cpu(counters, cpu)); > > If you want to use a remote local_read to synchronize access to a resource > -between CPUs, explicit smp_wmb() and smp_rmb() memory barriers must be used > +between CPUs, explicit ``smp_wmb()`` and ``smp_rmb()`` memory barriers must be used > respectively on the writer and the reader CPUs. It would be the case if you use > -the local_t variable as a counter of bytes written in a buffer : there should > -be a smp_wmb() between the buffer write and the counter increment and also a > -smp_rmb() between the counter read and the buffer read. > - > - > -Here is a sample module which implements a basic per cpu counter using local.h. > - > ---- BEGIN --- > -/* test-local.c > - * > - * Sample module for local.h usage. > - */ > - > - > -#include <asm/local.h> > -#include <linux/module.h> > -#include <linux/timer.h> > - > -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > - > -static struct timer_list test_timer; > - > -/* IPI called on each CPU. */ > -static void test_each(void *info) > -{ > - /* Increment the counter from a non preemptible context */ > - printk("Increment on cpu %d\n", smp_processor_id()); > - local_inc(this_cpu_ptr(&counters)); > - > - /* This is what incrementing the variable would look like within a > - * preemptible context (it disables preemption) : > - * > - * local_inc(&get_cpu_var(counters)); > - * put_cpu_var(counters); > - */ > -} > - > -static void do_test_timer(unsigned long data) > -{ > - int cpu; > - > - /* Increment the counters */ > - on_each_cpu(test_each, NULL, 1); > - /* Read all the counters */ > - printk("Counters read from CPU %d\n", smp_processor_id()); > - for_each_online_cpu(cpu) { > - printk("Read : CPU %d, count %ld\n", cpu, > - local_read(&per_cpu(counters, cpu))); > - } > - del_timer(&test_timer); > - test_timer.expires = jiffies + 1000; > - add_timer(&test_timer); > -} > - > -static int __init test_init(void) > -{ > - /* initialize the timer that will increment the counter */ > - init_timer(&test_timer); > - test_timer.function = do_test_timer; > - test_timer.expires = jiffies + 1; > - add_timer(&test_timer); > - > - return 0; > -} > - > -static void __exit test_exit(void) > -{ > - del_timer_sync(&test_timer); > -} > - > -module_init(test_init); > -module_exit(test_exit); > - > -MODULE_LICENSE("GPL"); > -MODULE_AUTHOR("Mathieu Desnoyers"); > -MODULE_DESCRIPTION("Local Atomic Ops"); > ---- END --- > +the ``local_t`` variable as a counter of bytes written in a buffer: there should > +be a ``smp_wmb()`` between the buffer write and the counter increment and also a > +``smp_rmb()`` between the counter read and the buffer read. > + > + > +Here is a sample module which implements a basic per cpu counter using > +``local.h``:: > + > + /* test-local.c > + * > + * Sample module for local.h usage. > + */ > + > + > + #include <asm/local.h> > + #include <linux/module.h> > + #include <linux/timer.h> > + > + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); > + > + static struct timer_list test_timer; > + > + /* IPI called on each CPU. */ > + static void test_each(void *info) > + { > + /* Increment the counter from a non preemptible context */ > + printk("Increment on cpu %d\n", smp_processor_id()); > + local_inc(this_cpu_ptr(&counters)); > + > + /* This is what incrementing the variable would look like within a > + * preemptible context (it disables preemption) : > + * > + * local_inc(&get_cpu_var(counters)); > + * put_cpu_var(counters); > + */ > + } > + > + static void do_test_timer(unsigned long data) > + { > + int cpu; > + > + /* Increment the counters */ > + on_each_cpu(test_each, NULL, 1); > + /* Read all the counters */ > + printk("Counters read from CPU %d\n", smp_processor_id()); > + for_each_online_cpu(cpu) { > + printk("Read : CPU %d, count %ld\n", cpu, > + local_read(&per_cpu(counters, cpu))); > + } > + del_timer(&test_timer); > + test_timer.expires = jiffies + 1000; > + add_timer(&test_timer); > + } > + > + static int __init test_init(void) > + { > + /* initialize the timer that will increment the counter */ > + init_timer(&test_timer); > + test_timer.function = do_test_timer; > + test_timer.expires = jiffies + 1; > + add_timer(&test_timer); > + > + return 0; > + } > + > + static void __exit test_exit(void) > + { > + del_timer_sync(&test_timer); > + } > + > + module_init(test_init); > + module_exit(test_exit); > + > + MODULE_LICENSE("GPL"); > + MODULE_AUTHOR("Mathieu Desnoyers"); > + MODULE_DESCRIPTION("Local Atomic Ops"); Thanks, Mauro ^ permalink raw reply [flat|nested] 51+ messages in thread
* [PATCH v3 4/4] firmware: remove warning at documentation generation time 2016-11-25 14:59 ` [PATCH v3 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke ` (2 preceding siblings ...) 2016-11-25 14:59 ` [PATCH v3 3/4] Documentation/local_ops.txt: " Silvio Fricke @ 2016-11-25 14:59 ` Silvio Fricke 2016-11-25 20:50 ` Mauro Carvalho Chehab 2016-11-28 17:30 ` [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 4 siblings, 1 reply; 51+ messages in thread From: Silvio Fricke @ 2016-11-25 14:59 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula, Silvio Fricke This patch removes following error at for `make htmldocs`. No functional change. ./drivers/base/firmware_class.c:1348: WARNING: Bullet list ends without a blank line; unexpected unindent. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- drivers/base/firmware_class.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c index 22d1760..37b0221 100644 --- a/drivers/base/firmware_class.c +++ b/drivers/base/firmware_class.c @@ -1345,9 +1345,9 @@ static void request_firmware_work_func(struct work_struct *work) * * Asynchronous variant of request_firmware() for user contexts: * - sleep for as small periods as possible since it may - * increase kernel boot time of built-in device drivers - * requesting firmware in their ->probe() methods, if - * @gfp is GFP_KERNEL. + * increase kernel boot time of built-in device drivers + * requesting firmware in their ->probe() methods, if + * @gfp is GFP_KERNEL. * * - can't sleep at all if @gfp is GFP_ATOMIC. **/ -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* Re: [PATCH v3 4/4] firmware: remove warning at documentation generation time 2016-11-25 14:59 ` [PATCH v3 4/4] firmware: remove warning at documentation generation time Silvio Fricke @ 2016-11-25 20:50 ` Mauro Carvalho Chehab 2016-12-01 16:11 ` Luis R. Rodriguez 0 siblings, 1 reply; 51+ messages in thread From: Mauro Carvalho Chehab @ 2016-11-25 20:50 UTC (permalink / raw) To: Silvio Fricke Cc: linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula Em Fri, 25 Nov 2016 15:59:47 +0100 Silvio Fricke <silvio.fricke@gmail.com> escreveu: > This patch removes following error at for `make htmldocs`. No functional > change. > > ./drivers/base/firmware_class.c:1348: WARNING: Bullet list ends without a blank line; unexpected unindent. > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> Reviewed-by: Mauro Carvalho Chehab <mchehab@s-opensource.com> > --- > drivers/base/firmware_class.c | 6 +++--- > 1 file changed, 3 insertions(+), 3 deletions(-) > > diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c > index 22d1760..37b0221 100644 > --- a/drivers/base/firmware_class.c > +++ b/drivers/base/firmware_class.c > @@ -1345,9 +1345,9 @@ static void request_firmware_work_func(struct work_struct *work) > * > * Asynchronous variant of request_firmware() for user contexts: > * - sleep for as small periods as possible since it may > - * increase kernel boot time of built-in device drivers > - * requesting firmware in their ->probe() methods, if > - * @gfp is GFP_KERNEL. > + * increase kernel boot time of built-in device drivers > + * requesting firmware in their ->probe() methods, if > + * @gfp is GFP_KERNEL. > * > * - can't sleep at all if @gfp is GFP_ATOMIC. > **/ Thanks, Mauro ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 4/4] firmware: remove warning at documentation generation time 2016-11-25 20:50 ` Mauro Carvalho Chehab @ 2016-12-01 16:11 ` Luis R. Rodriguez 2016-12-01 16:45 ` Greg Kroah-Hartman 0 siblings, 1 reply; 51+ messages in thread From: Luis R. Rodriguez @ 2016-12-01 16:11 UTC (permalink / raw) To: Jonathan Corbet, Mauro Carvalho Chehab, Greg Kroah-Hartman Cc: Silvio Fricke, linux-doc, Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula On Fri, Nov 25, 2016 at 06:50:32PM -0200, Mauro Carvalho Chehab wrote: > Em Fri, 25 Nov 2016 15:59:47 +0100 > Silvio Fricke <silvio.fricke@gmail.com> escreveu: > > > This patch removes following error at for `make htmldocs`. No functional > > change. > > > > ./drivers/base/firmware_class.c:1348: WARNING: Bullet list ends without a blank line; unexpected unindent. > > > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> > > Reviewed-by: Mauro Carvalho Chehab <mchehab@s-opensource.com> Acked-by: Luis R. Rodriguez <mcgrof@kernel.org> Jonathan, firmware_class changes have been lately going through Greg's tree, he has a few more pending changes on his driver-core-next branch for firmware_class but this patch not create a conflict. It can go in first or later as such its technically fine for it to go either through your tree or Greg's tree. I'll let Greg decide. I've bounced him a copy of the patch. Luis ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v3 4/4] firmware: remove warning at documentation generation time 2016-12-01 16:11 ` Luis R. Rodriguez @ 2016-12-01 16:45 ` Greg Kroah-Hartman 0 siblings, 0 replies; 51+ messages in thread From: Greg Kroah-Hartman @ 2016-12-01 16:45 UTC (permalink / raw) To: Luis R. Rodriguez Cc: Jonathan Corbet, Mauro Carvalho Chehab, Silvio Fricke, linux-doc, Ming Lei, Mauro Carvalho Chehab, linux-kernel, Jani Nikula On Thu, Dec 01, 2016 at 05:11:54PM +0100, Luis R. Rodriguez wrote: > On Fri, Nov 25, 2016 at 06:50:32PM -0200, Mauro Carvalho Chehab wrote: > > Em Fri, 25 Nov 2016 15:59:47 +0100 > > Silvio Fricke <silvio.fricke@gmail.com> escreveu: > > > > > This patch removes following error at for `make htmldocs`. No functional > > > change. > > > > > > ./drivers/base/firmware_class.c:1348: WARNING: Bullet list ends without a blank line; unexpected unindent. > > > > > > Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> > > > > Reviewed-by: Mauro Carvalho Chehab <mchehab@s-opensource.com> > > Acked-by: Luis R. Rodriguez <mcgrof@kernel.org> > > Jonathan, > > firmware_class changes have been lately going through Greg's tree, he has a few > more pending changes on his driver-core-next branch for firmware_class but this > patch not create a conflict. It can go in first or later as such its > technically fine for it to go either through your tree or Greg's tree. > > I'll let Greg decide. I've bounced him a copy of the patch. I can just take this one in my tree as it's easy to merge as-is. thanks, greg k-h ^ permalink raw reply [flat|nested] 51+ messages in thread
* [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops 2016-11-25 14:59 ` [PATCH v3 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke ` (3 preceding siblings ...) 2016-11-25 14:59 ` [PATCH v3 4/4] firmware: remove warning at documentation generation time Silvio Fricke @ 2016-11-28 17:30 ` Silvio Fricke 2016-11-28 17:30 ` [PATCH v4 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke ` (4 more replies) 4 siblings, 5 replies; 51+ messages in thread From: Silvio Fricke @ 2016-11-28 17:30 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula, Peter Zijlstra, Daniel Vetter, Silvio Fricke Hi, Some more ReSTification of core-api's: assoc_array, atomic_ops and local_ops. A fourth patch removes a warning about a bullet list without ending at firmware_class.c v3 -> v4 * add Review-by comments * rewrite atomic_ops.rst file with fewer inversive changes * reorder patch series old(1,2,3,4) new(1,3,2,4) v2 -> v3 * change ". ::" to "::" * replace all "code-blocks" with "::" * add two ".. notes" declaration in atomic_ops.rst v1 -> v2 * use format-patch with a rename_threshold of 10%, no other changes Thanks for review. BR Silvio Silvio Fricke (4): Documentation/assoc_array.txt: convert to ReST markup Documentation/local_ops.txt: convert to ReST markup Documentation/atomic_ops.txt: convert to ReST markup firmware: remove warning at documentation generation time Documentation/assoc_array.txt => Documentation/core-api/assoc_array.rst | 639 ++++++++++++++++++++++++++++++++++-------------------------------------- Documentation/atomic_ops.txt => Documentation/core-api/atomic_ops.rst | 324 +++++++++++++++++++------------------ Documentation/core-api/index.rst | 3 +- Documentation/local_ops.txt => Documentation/core-api/local_ops.rst | 273 ++++++++++++++++--------------- Documentation/process/volatile-considered-harmful.rst | 3 +- drivers/base/firmware_class.c | 6 +- 6 files changed, 632 insertions(+), 616 deletions(-) base-commit: 9c240d757658a3ae9968dd309e674c61f07c7f48 -- git-series 0.9.1 ^ permalink raw reply [flat|nested] 51+ messages in thread
* [PATCH v4 1/4] Documentation/assoc_array.txt: convert to ReST markup 2016-11-28 17:30 ` [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke @ 2016-11-28 17:30 ` Silvio Fricke 2016-11-28 17:30 ` [PATCH v4 2/4] Documentation/local_ops.txt: " Silvio Fricke ` (3 subsequent siblings) 4 siblings, 0 replies; 51+ messages in thread From: Silvio Fricke @ 2016-11-28 17:30 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula, Peter Zijlstra, Daniel Vetter, Silvio Fricke ... and move to Documentation/core-api folder. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> Reviewed-by: Mauro Carvalho Chehab <mchehab@s-opensource.com> --- Documentation/assoc_array.txt => Documentation/core-api/assoc_array.rst | 639 ++++++++++++++++++++++++++++++++++-------------------------------------- Documentation/core-api/index.rst | 1 +- 2 files changed, 309 insertions(+), 331 deletions(-) diff --git a/Documentation/assoc_array.txt b/Documentation/core-api/assoc_array.rst similarity index 46% rename from Documentation/assoc_array.txt rename to Documentation/core-api/assoc_array.rst index 2f2c6cd..dcda7c6 100644 --- a/Documentation/assoc_array.txt +++ b/Documentation/core-api/assoc_array.rst @@ -1,67 +1,46 @@ - ======================================== - GENERIC ASSOCIATIVE ARRAY IMPLEMENTATION - ======================================== +======================================== +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 +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. +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. +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. +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. +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. +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. +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. +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. +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. +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. +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. +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 @@ -71,25 +50,26 @@ 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 -============== -The public API can be found in <linux/assoc_array.h>. The associative array is -rooted on the following structure: +The public API can be found in ``<linux/assoc_array.h>``. The associative +array is rooted on the following structure:: + + struct assoc_array { + ... + }; - struct assoc_array { - ... - }; +The code is selected by enabling ``CONFIG_ASSOCIATIVE_ARRAY`` with:: -The code is selected by enabling CONFIG_ASSOCIATIVE_ARRAY. + ./script/config -e ASSOCIATIVE_ARRAY -EDIT SCRIPT +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 +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. @@ -99,246 +79,245 @@ 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: +The script appears as outside of the API as a pointer of the type:: - struct assoc_array_edit; + struct assoc_array_edit; There are two functions for dealing with the script: - (1) Apply an edit script. +1. Apply an edit script:: - void assoc_array_apply_edit(struct assoc_array_edit *edit); + 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. +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. +2. Cancel an edit script:: - void assoc_array_cancel_edit(struct assoc_array_edit *edit); + 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. +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 +Operations Table ---------------- -Various functions take a table of operations: +Various functions take a table of operations:: - struct assoc_array_ops { - ... - }; + 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: +1. Get a chunk of index key from caller data:: - unsigned long (*get_key_chunk)(const void *index_key, int level); + 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. +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. +2. Get a chunk of an object's index key:: - unsigned long (*get_object_key_chunk)(const void *object, int level); + 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. +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. +3. See if this is the object we're looking for:: - bool (*compare_object)(const void *object, const void *index_key); + 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. +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. +4. Diff the index keys of two objects:: - int (*diff_objects)(const void *object, const void *index_key); + int (*diff_objects)(const void *object, const void *index_key); - Return the bit position at which the index key of the specified object - differs from the given index key or -1 if they are the same. +Return the bit position at which the index key of the specified object +differs from the given index key or -1 if they are the same. - (5) Free an object. +5. Free an object:: - void (*free_object)(void *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. +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 +Manipulation Functions ---------------------- There are a number of functions for manipulating an associative array: - (1) Initialise an associative array. +1. Initialise an associative array:: - void assoc_array_init(struct assoc_array *array); + void assoc_array_init(struct assoc_array *array); - This initialises the base structure for an associative array. It can't - fail. +This initialises the base structure for an associative array. It can't fail. - (2) Insert/replace an object in an associative array. +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); + 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. +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. +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. +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. +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. +The caller should lock exclusively against other modifiers of the array. - (3) Delete an object from an associative 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); + 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. +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. +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. +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. +The caller should lock exclusively against other modifiers of the array. - (4) Delete all objects from an associative 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); + 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 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. +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. +The caller should lock exclusively against other modifiers of the array. - (5) Destroy an associative array, deleting all objects. +5. Destroy an associative array, deleting all objects:: - void assoc_array_destroy(struct assoc_array *array, - const struct assoc_array_ops *ops); + 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. +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. +The caller should lock exclusively against other modifiers and accessors +of the array. - (6) Garbage collect an associative 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); + 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. +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 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 ``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. +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. +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 +Access Functions ---------------- There are two functions for accessing an associative array: - (1) Iterate over all the objects in 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); + 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 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. +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. +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. +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); + 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 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. +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. +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 +Index Key Form -------------- The index key can be of any form, but since the algorithms aren't told how long @@ -364,8 +343,7 @@ unlikely that more than one word of any particular index key will have to be used. -================= -INTERNAL WORKINGS +Internal Workings ================= The associative array data structure has an internal tree. This tree is @@ -373,82 +351,80 @@ 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. +* 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 +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 | - +---+ +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 +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 +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 @@ -462,23 +438,23 @@ 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: +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 + 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: +and node B:: - 3 PTR TO NODE C 13* - any LEAF 1458952489 + 3 PTR TO NODE C 13* + any LEAF 1458952489 -SHORTCUTS +Shortcuts --------- Shortcuts are metadata records that jump over a piece of keyspace. A shortcut @@ -486,12 +462,13 @@ 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. +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 +Splitting And Collapsing Nodes ------------------------------ Each node has a maximum capacity of 16 leaves and metadata pointers. If the @@ -508,7 +485,7 @@ fewer, then the subtree will be collapsed down to a single node - and this will ripple towards the root if possible. -NON-RECURSIVE ITERATION +Non-Recursive Iteration ----------------------- Each node and shortcut contains a back pointer to its parent and the number of @@ -519,55 +496,55 @@ make sure progress is made without the need for a stack. The backpointers, however, make simultaneous alteration and iteration tricky. -SIMULTANEOUS ALTERATION AND ITERATION +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. +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 diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index f7ef7fd..480d9a3 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -7,6 +7,7 @@ Kernel and driver related documentation. .. toctree:: :maxdepth: 1 + assoc_array workqueue .. only:: subproject -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* [PATCH v4 2/4] Documentation/local_ops.txt: convert to ReST markup 2016-11-28 17:30 ` [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 2016-11-28 17:30 ` [PATCH v4 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke @ 2016-11-28 17:30 ` Silvio Fricke 2016-11-28 17:30 ` [PATCH v4 3/4] Documentation/atomic_ops.txt: " Silvio Fricke ` (2 subsequent siblings) 4 siblings, 0 replies; 51+ messages in thread From: Silvio Fricke @ 2016-11-28 17:30 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula, Peter Zijlstra, Daniel Vetter, Silvio Fricke ... and move to core-api folder. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> Reviewed-by: Mauro Carvalho Chehab <mchehab@s-opensource.com> --- Documentation/core-api/index.rst | 1 +- Documentation/local_ops.txt => Documentation/core-api/local_ops.rst | 273 +++---- 2 files changed, 145 insertions(+), 129 deletions(-) diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index 480d9a3..f53555e 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -8,6 +8,7 @@ Kernel and driver related documentation. :maxdepth: 1 assoc_array + local_ops workqueue .. only:: subproject diff --git a/Documentation/local_ops.txt b/Documentation/core-api/local_ops.rst similarity index 57% rename from Documentation/local_ops.txt rename to Documentation/core-api/local_ops.rst index 407576a..1062ddb 100644 --- a/Documentation/local_ops.txt +++ b/Documentation/core-api/local_ops.rst @@ -1,191 +1,206 @@ - Semantics and Behavior of Local Atomic Operations - Mathieu Desnoyers +.. _local_ops: +================================================= +Semantics and Behavior of Local Atomic Operations +================================================= - This document explains the purpose of the local atomic operations, how +:Author: Mathieu Desnoyers + + +This document explains the purpose of the local atomic operations, how to implement them for any given architecture and shows how they can be used properly. It also stresses on the precautions that must be taken when reading those local variables across CPUs when the order of memory writes matters. -Note that local_t based operations are not recommended for general kernel use. -Please use the this_cpu operations instead unless there is really a special purpose. -Most uses of local_t in the kernel have been replaced by this_cpu operations. -this_cpu operations combine the relocation with the local_t like semantics in -a single instruction and yield more compact and faster executing code. +.. note:: + Note that ``local_t`` based operations are not recommended for general + kernel use. Please use the ``this_cpu`` operations instead unless there is + really a special purpose. Most uses of ``local_t`` in the kernel have been + replaced by ``this_cpu`` operations. ``this_cpu`` operations combine the + relocation with the ``local_t`` like semantics in a single instruction and + yield more compact and faster executing code. -* Purpose of local atomic operations + +Purpose of local atomic operations +================================== Local atomic operations are meant to provide fast and highly reentrant per CPU counters. They minimize the performance cost of standard atomic operations by removing the LOCK prefix and memory barriers normally required to synchronize across CPUs. -Having fast per CPU atomic counters is interesting in many cases : it does not +Having fast per CPU atomic counters is interesting in many cases: it does not require disabling interrupts to protect from interrupt handlers and it permits coherent counters in NMI handlers. It is especially useful for tracing purposes and for various performance monitoring counters. Local atomic operations only guarantee variable modification atomicity wrt the CPU which owns the data. Therefore, care must taken to make sure that only one -CPU writes to the local_t data. This is done by using per cpu data and making -sure that we modify it from within a preemption safe context. It is however -permitted to read local_t data from any CPU : it will then appear to be written -out of order wrt other memory writes by the owner CPU. +CPU writes to the ``local_t`` data. This is done by using per cpu data and +making sure that we modify it from within a preemption safe context. It is +however permitted to read ``local_t`` data from any CPU: it will then appear to +be written out of order wrt other memory writes by the owner CPU. -* Implementation for a given architecture +Implementation for a given architecture +======================================= -It can be done by slightly modifying the standard atomic operations : only +It can be done by slightly modifying the standard atomic operations: only their UP variant must be kept. It typically means removing LOCK prefix (on i386 and x86_64) and any SMP synchronization barrier. If the architecture does -not have a different behavior between SMP and UP, including asm-generic/local.h -in your architecture's local.h is sufficient. +not have a different behavior between SMP and UP, including +``asm-generic/local.h`` in your architecture's ``local.h`` is sufficient. -The local_t type is defined as an opaque signed long by embedding an -atomic_long_t inside a structure. This is made so a cast from this type to a -long fails. The definition looks like : +The ``local_t`` type is defined as an opaque ``signed long`` by embedding an +``atomic_long_t`` inside a structure. This is made so a cast from this type to +a ``long`` fails. The definition looks like:: -typedef struct { atomic_long_t a; } local_t; + typedef struct { atomic_long_t a; } local_t; -* Rules to follow when using local atomic operations +Rules to follow when using local atomic operations +================================================== -- Variables touched by local ops must be per cpu variables. -- _Only_ the CPU owner of these variables must write to them. -- This CPU can use local ops from any context (process, irq, softirq, nmi, ...) - to update its local_t variables. -- Preemption (or interrupts) must be disabled when using local ops in - process context to make sure the process won't be migrated to a +* Variables touched by local ops must be per cpu variables. +* *Only* the CPU owner of these variables must write to them. +* This CPU can use local ops from any context (process, irq, softirq, nmi, ...) + to update its ``local_t`` variables. +* Preemption (or interrupts) must be disabled when using local ops in + process context to make sure the process won't be migrated to a different CPU between getting the per-cpu variable and doing the actual local op. -- When using local ops in interrupt context, no special care must be +* When using local ops in interrupt context, no special care must be taken on a mainline kernel, since they will run on the local CPU with preemption already disabled. I suggest, however, to explicitly disable preemption anyway to make sure it will still work correctly on -rt kernels. -- Reading the local cpu variable will provide the current copy of the +* Reading the local cpu variable will provide the current copy of the variable. -- Reads of these variables can be done from any CPU, because updates to - "long", aligned, variables are always atomic. Since no memory +* Reads of these variables can be done from any CPU, because updates to + "``long``", aligned, variables are always atomic. Since no memory synchronization is done by the writer CPU, an outdated copy of the - variable can be read when reading some _other_ cpu's variables. + variable can be read when reading some *other* cpu's variables. + +How to use local atomic operations +================================== -* How to use local atomic operations +:: -#include <linux/percpu.h> -#include <asm/local.h> + #include <linux/percpu.h> + #include <asm/local.h> -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); -* Counting +Counting +======== Counting is done on all the bits of a signed long. -In preemptible context, use get_cpu_var() and put_cpu_var() around local atomic -operations : it makes sure that preemption is disabled around write access to -the per cpu variable. For instance : +In preemptible context, use ``get_cpu_var()`` and ``put_cpu_var()`` around +local atomic operations: it makes sure that preemption is disabled around write +access to the per cpu variable. For instance:: - local_inc(&get_cpu_var(counters)); - put_cpu_var(counters); + local_inc(&get_cpu_var(counters)); + put_cpu_var(counters); If you are already in a preemption-safe context, you can use -this_cpu_ptr() instead. +``this_cpu_ptr()`` instead:: - local_inc(this_cpu_ptr(&counters)); + local_inc(this_cpu_ptr(&counters)); -* Reading the counters +Reading the counters +==================== Those local counters can be read from foreign CPUs to sum the count. Note that the data seen by local_read across CPUs must be considered to be out of order -relatively to other memory writes happening on the CPU that owns the data. +relatively to other memory writes happening on the CPU that owns the data:: - long sum = 0; - for_each_online_cpu(cpu) - sum += local_read(&per_cpu(counters, cpu)); + long sum = 0; + for_each_online_cpu(cpu) + sum += local_read(&per_cpu(counters, cpu)); If you want to use a remote local_read to synchronize access to a resource -between CPUs, explicit smp_wmb() and smp_rmb() memory barriers must be used +between CPUs, explicit ``smp_wmb()`` and ``smp_rmb()`` memory barriers must be used respectively on the writer and the reader CPUs. It would be the case if you use -the local_t variable as a counter of bytes written in a buffer : there should -be a smp_wmb() between the buffer write and the counter increment and also a -smp_rmb() between the counter read and the buffer read. - - -Here is a sample module which implements a basic per cpu counter using local.h. - ---- BEGIN --- -/* test-local.c - * - * Sample module for local.h usage. - */ - - -#include <asm/local.h> -#include <linux/module.h> -#include <linux/timer.h> - -static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); - -static struct timer_list test_timer; - -/* IPI called on each CPU. */ -static void test_each(void *info) -{ - /* Increment the counter from a non preemptible context */ - printk("Increment on cpu %d\n", smp_processor_id()); - local_inc(this_cpu_ptr(&counters)); - - /* This is what incrementing the variable would look like within a - * preemptible context (it disables preemption) : - * - * local_inc(&get_cpu_var(counters)); - * put_cpu_var(counters); - */ -} - -static void do_test_timer(unsigned long data) -{ - int cpu; - - /* Increment the counters */ - on_each_cpu(test_each, NULL, 1); - /* Read all the counters */ - printk("Counters read from CPU %d\n", smp_processor_id()); - for_each_online_cpu(cpu) { - printk("Read : CPU %d, count %ld\n", cpu, - local_read(&per_cpu(counters, cpu))); - } - del_timer(&test_timer); - test_timer.expires = jiffies + 1000; - add_timer(&test_timer); -} - -static int __init test_init(void) -{ - /* initialize the timer that will increment the counter */ - init_timer(&test_timer); - test_timer.function = do_test_timer; - test_timer.expires = jiffies + 1; - add_timer(&test_timer); - - return 0; -} - -static void __exit test_exit(void) -{ - del_timer_sync(&test_timer); -} - -module_init(test_init); -module_exit(test_exit); - -MODULE_LICENSE("GPL"); -MODULE_AUTHOR("Mathieu Desnoyers"); -MODULE_DESCRIPTION("Local Atomic Ops"); ---- END --- +the ``local_t`` variable as a counter of bytes written in a buffer: there should +be a ``smp_wmb()`` between the buffer write and the counter increment and also a +``smp_rmb()`` between the counter read and the buffer read. + + +Here is a sample module which implements a basic per cpu counter using +``local.h``:: + + /* test-local.c + * + * Sample module for local.h usage. + */ + + + #include <asm/local.h> + #include <linux/module.h> + #include <linux/timer.h> + + static DEFINE_PER_CPU(local_t, counters) = LOCAL_INIT(0); + + static struct timer_list test_timer; + + /* IPI called on each CPU. */ + static void test_each(void *info) + { + /* Increment the counter from a non preemptible context */ + printk("Increment on cpu %d\n", smp_processor_id()); + local_inc(this_cpu_ptr(&counters)); + + /* This is what incrementing the variable would look like within a + * preemptible context (it disables preemption) : + * + * local_inc(&get_cpu_var(counters)); + * put_cpu_var(counters); + */ + } + + static void do_test_timer(unsigned long data) + { + int cpu; + + /* Increment the counters */ + on_each_cpu(test_each, NULL, 1); + /* Read all the counters */ + printk("Counters read from CPU %d\n", smp_processor_id()); + for_each_online_cpu(cpu) { + printk("Read : CPU %d, count %ld\n", cpu, + local_read(&per_cpu(counters, cpu))); + } + del_timer(&test_timer); + test_timer.expires = jiffies + 1000; + add_timer(&test_timer); + } + + static int __init test_init(void) + { + /* initialize the timer that will increment the counter */ + init_timer(&test_timer); + test_timer.function = do_test_timer; + test_timer.expires = jiffies + 1; + add_timer(&test_timer); + + return 0; + } + + static void __exit test_exit(void) + { + del_timer_sync(&test_timer); + } + + module_init(test_init); + module_exit(test_exit); + + MODULE_LICENSE("GPL"); + MODULE_AUTHOR("Mathieu Desnoyers"); + MODULE_DESCRIPTION("Local Atomic Ops"); -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* [PATCH v4 3/4] Documentation/atomic_ops.txt: convert to ReST markup 2016-11-28 17:30 ` [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 2016-11-28 17:30 ` [PATCH v4 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke 2016-11-28 17:30 ` [PATCH v4 2/4] Documentation/local_ops.txt: " Silvio Fricke @ 2016-11-28 17:30 ` Silvio Fricke 2016-11-28 17:30 ` [PATCH v4 4/4] firmware: remove warning at documentation generation time Silvio Fricke 2016-12-01 0:51 ` [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Jonathan Corbet 4 siblings, 0 replies; 51+ messages in thread From: Silvio Fricke @ 2016-11-28 17:30 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula, Peter Zijlstra, Daniel Vetter, Silvio Fricke ... and move to core-api folder. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> --- Documentation/atomic_ops.txt => Documentation/core-api/atomic_ops.rst | 324 ++++++++++++++++++++++++++++++++++++++---------------------------------- Documentation/core-api/index.rst | 1 +- Documentation/process/volatile-considered-harmful.rst | 3 +- 3 files changed, 175 insertions(+), 153 deletions(-) diff --git a/Documentation/atomic_ops.txt b/Documentation/core-api/atomic_ops.rst similarity index 73% rename from Documentation/atomic_ops.txt rename to Documentation/core-api/atomic_ops.rst index 6c5e8a9..55e43f1 100644 --- a/Documentation/atomic_ops.txt +++ b/Documentation/core-api/atomic_ops.rst @@ -1,36 +1,42 @@ - Semantics and Behavior of Atomic and - Bitmask Operations +======================================================= +Semantics and Behavior of Atomic and Bitmask Operations +======================================================= - David S. Miller +:Author: David S. Miller - This document is intended to serve as a guide to Linux port +This document is intended to serve as a guide to Linux port maintainers on how to implement atomic counter, bitops, and spinlock interfaces properly. - The atomic_t type should be defined as a signed integer and +Atomic Type And Operations +========================== + +The atomic_t type should be defined as a signed integer and the atomic_long_t type as a signed long integer. Also, they should be made opaque such that any kind of cast to a normal C integer type -will fail. Something like the following should suffice: +will fail. Something like the following should suffice:: typedef struct { int counter; } atomic_t; typedef struct { long counter; } atomic_long_t; Historically, counter has been declared volatile. This is now discouraged. -See Documentation/process/volatile-considered-harmful.rst for the complete rationale. +See :ref:`Documentation/process/volatile-considered-harmful.rst +<volatile_considered_harmful>` for the complete rationale. local_t is very similar to atomic_t. If the counter is per CPU and only updated by one CPU, local_t is probably more appropriate. Please see -Documentation/local_ops.txt for the semantics of local_t. +:ref:`Documentation/core-api/local_ops.rst <local_ops>` for the semantics of +local_t. The first operations to implement for atomic_t's are the initializers and -plain reads. +plain reads. :: #define ATOMIC_INIT(i) { (i) } #define atomic_set(v, i) ((v)->counter = (i)) -The first macro is used in definitions, such as: +The first macro is used in definitions, such as:: -static atomic_t my_counter = ATOMIC_INIT(1); + static atomic_t my_counter = ATOMIC_INIT(1); The initializer is atomic in that the return values of the atomic operations are guaranteed to be correct reflecting the initialized value if the @@ -38,10 +44,10 @@ initializer is used before runtime. If the initializer is used at runtime, a proper implicit or explicit read memory barrier is needed before reading the value with atomic_read from another thread. -As with all of the atomic_ interfaces, replace the leading "atomic_" -with "atomic_long_" to operate on atomic_long_t. +As with all of the ``atomic_`` interfaces, replace the leading ``atomic_`` +with ``atomic_long_`` to operate on atomic_long_t. -The second interface can be used at runtime, as in: +The second interface can be used at runtime, as in:: struct foo { atomic_t counter; }; ... @@ -59,7 +65,7 @@ been set with this operation or set with another operation. A proper implicit or explicit memory barrier is needed before the value set with the operation is guaranteed to be readable with atomic_read from another thread. -Next, we have: +Next, we have:: #define atomic_read(v) ((v)->counter) @@ -73,20 +79,21 @@ initialization by any other thread is visible yet, so the user of the interface must take care of that with a proper implicit or explicit memory barrier. -*** WARNING: atomic_read() and atomic_set() DO NOT IMPLY BARRIERS! *** +.. warning:: -Some architectures may choose to use the volatile keyword, barriers, or inline -assembly to guarantee some degree of immediacy for atomic_read() and -atomic_set(). This is not uniformly guaranteed, and may change in the future, -so all users of atomic_t should treat atomic_read() and atomic_set() as simple -C statements that may be reordered or optimized away entirely by the compiler -or processor, and explicitly invoke the appropriate compiler and/or memory -barrier for each use case. Failure to do so will result in code that may -suddenly break when used with different architectures or compiler -optimizations, or even changes in unrelated code which changes how the -compiler optimizes the section accessing atomic_t variables. + ``atomic_read()`` and ``atomic_set()`` DO NOT IMPLY BARRIERS! -*** YOU HAVE BEEN WARNED! *** + Some architectures may choose to use the volatile keyword, barriers, or + inline assembly to guarantee some degree of immediacy for atomic_read() + and atomic_set(). This is not uniformly guaranteed, and may change in + the future, so all users of atomic_t should treat atomic_read() and + atomic_set() as simple C statements that may be reordered or optimized + away entirely by the compiler or processor, and explicitly invoke the + appropriate compiler and/or memory barrier for each use case. Failure + to do so will result in code that may suddenly break when used with + different architectures or compiler optimizations, or even changes in + unrelated code which changes how the compiler optimizes the section + accessing atomic_t variables. Properly aligned pointers, longs, ints, and chars (and unsigned equivalents) may be atomically loaded from and stored to in the same @@ -95,14 +102,14 @@ and WRITE_ONCE() macros should be used to prevent the compiler from using optimizations that might otherwise optimize accesses out of existence on the one hand, or that might create unsolicited accesses on the other. -For example consider the following code: +For example consider the following code:: while (a > 0) do_something(); If the compiler can prove that do_something() does not store to the variable a, then the compiler is within its rights transforming this to -the following: +the following:: tmp = a; if (a > 0) @@ -110,14 +117,14 @@ the following: do_something(); If you don't want the compiler to do this (and you probably don't), then -you should use something like the following: +you should use something like the following:: while (READ_ONCE(a) < 0) do_something(); Alternatively, you could place a barrier() call in the loop. -For another example, consider the following code: +For another example, consider the following code:: tmp_a = a; do_something_with(tmp_a); @@ -125,7 +132,7 @@ For another example, consider the following code: If the compiler can prove that do_something_with() does not store to the variable a, then the compiler is within its rights to manufacture an -additional load as follows: +additional load as follows:: tmp_a = a; do_something_with(tmp_a); @@ -139,7 +146,7 @@ The compiler would be likely to manufacture this additional load if do_something_with() was an inline function that made very heavy use of registers: reloading from variable a could save a flush to the stack and later reload. To prevent the compiler from attacking your -code in this manner, write the following: +code in this manner, write the following:: tmp_a = READ_ONCE(a); do_something_with(tmp_a); @@ -147,7 +154,7 @@ code in this manner, write the following: For a final example, consider the following code, assuming that the variable a is set at boot time before the second CPU is brought online -and never changed later, so that memory barriers are not needed: +and never changed later, so that memory barriers are not needed:: if (a) b = 9; @@ -155,7 +162,7 @@ and never changed later, so that memory barriers are not needed: b = 42; The compiler is within its rights to manufacture an additional store -by transforming the above code into the following: +by transforming the above code into the following:: b = 42; if (a) @@ -163,7 +170,7 @@ by transforming the above code into the following: This could come as a fatal surprise to other code running concurrently that expected b to never have the value 42 if a was zero. To prevent -the compiler from doing this, write something like: +the compiler from doing this, write something like:: if (a) WRITE_ONCE(b, 9); @@ -173,10 +180,12 @@ the compiler from doing this, write something like: Don't even -think- about doing this without proper use of memory barriers, locks, or atomic operations if variable a can change at runtime! -*** WARNING: READ_ONCE() OR WRITE_ONCE() DO NOT IMPLY A BARRIER! *** +.. warning:: + + ``READ_ONCE()`` OR ``WRITE_ONCE()`` DO NOT IMPLY A BARRIER! Now, we move onto the atomic operation interfaces typically implemented with -the help of assembly code. +the help of assembly code. :: void atomic_add(int i, atomic_t *v); void atomic_sub(int i, atomic_t *v); @@ -192,7 +201,7 @@ One very important aspect of these two routines is that they DO NOT require any explicit memory barriers. They need only perform the atomic_t counter update in an SMP safe manner. -Next, we have: +Next, we have:: int atomic_inc_return(atomic_t *v); int atomic_dec_return(atomic_t *v); @@ -214,7 +223,7 @@ If the atomic instructions used in an implementation provide explicit memory barrier semantics which satisfy the above requirements, that is fine as well. -Let's move on: +Let's move on:: int atomic_add_return(int i, atomic_t *v); int atomic_sub_return(int i, atomic_t *v); @@ -224,7 +233,7 @@ explicit counter adjustment is given instead of the implicit "1". This means that like atomic_{inc,dec}_return(), the memory barrier semantics are required. -Next: +Next:: int atomic_inc_and_test(atomic_t *v); int atomic_dec_and_test(atomic_t *v); @@ -234,13 +243,13 @@ given atomic counter. They return a boolean indicating whether the resulting counter value was zero or not. Again, these primitives provide explicit memory barrier semantics around -the atomic operation. +the atomic operation:: int atomic_sub_and_test(int i, atomic_t *v); This is identical to atomic_dec_and_test() except that an explicit decrement is given instead of the implicit "1". This primitive must -provide explicit memory barrier semantics around the operation. +provide explicit memory barrier semantics around the operation:: int atomic_add_negative(int i, atomic_t *v); @@ -249,7 +258,7 @@ is return which indicates whether the resulting counter value is negative. This primitive must provide explicit memory barrier semantics around the operation. -Then: +Then:: int atomic_xchg(atomic_t *v, int new); @@ -257,14 +266,14 @@ This performs an atomic exchange operation on the atomic variable v, setting the given new value. It returns the old value that the atomic variable v had just before the operation. -atomic_xchg must provide explicit memory barriers around the operation. +atomic_xchg must provide explicit memory barriers around the operation. :: int atomic_cmpxchg(atomic_t *v, int old, int new); This performs an atomic compare exchange operation on the atomic value v, with the given old and new values. Like all atomic_xxx operations, atomic_cmpxchg will only satisfy its atomicity semantics as long as all -other accesses of *v are performed through atomic_xxx operations. +other accesses of \*v are performed through atomic_xxx operations. atomic_cmpxchg must provide explicit memory barriers around the operation, although if the comparison fails then no memory ordering guarantees are @@ -273,7 +282,7 @@ required. The semantics for atomic_cmpxchg are the same as those defined for 'cas' below. -Finally: +Finally:: int atomic_add_unless(atomic_t *v, int a, int u); @@ -289,12 +298,12 @@ atomic_inc_not_zero, equivalent to atomic_add_unless(v, 1, 0) If a caller requires memory barrier semantics around an atomic_t operation which does not return a value, a set of interfaces are -defined which accomplish this: +defined which accomplish this:: void smp_mb__before_atomic(void); void smp_mb__after_atomic(void); -For example, smp_mb__before_atomic() can be used like so: +For example, smp_mb__before_atomic() can be used like so:: obj->dead = 1; smp_mb__before_atomic(); @@ -315,67 +324,69 @@ atomic_t implementation above can have disastrous results. Here is an example, which follows a pattern occurring frequently in the Linux kernel. It is the use of atomic counters to implement reference counting, and it works such that once the counter falls to zero it can -be guaranteed that no other entity can be accessing the object: - -static void obj_list_add(struct obj *obj, struct list_head *head) -{ - obj->active = 1; - list_add(&obj->list, head); -} - -static void obj_list_del(struct obj *obj) -{ - list_del(&obj->list); - obj->active = 0; -} - -static void obj_destroy(struct obj *obj) -{ - BUG_ON(obj->active); - kfree(obj); -} - -struct obj *obj_list_peek(struct list_head *head) -{ - if (!list_empty(head)) { - struct obj *obj; +be guaranteed that no other entity can be accessing the object:: + + static void obj_list_add(struct obj *obj, struct list_head *head) + { + obj->active = 1; + list_add(&obj->list, head); + } + + static void obj_list_del(struct obj *obj) + { + list_del(&obj->list); + obj->active = 0; + } - obj = list_entry(head->next, struct obj, list); - atomic_inc(&obj->refcnt); - return obj; + static void obj_destroy(struct obj *obj) + { + BUG_ON(obj->active); + kfree(obj); } - return NULL; -} -void obj_poke(void) -{ - struct obj *obj; + struct obj *obj_list_peek(struct list_head *head) + { + if (!list_empty(head)) { + struct obj *obj; + + obj = list_entry(head->next, struct obj, list); + atomic_inc(&obj->refcnt); + return obj; + } + return NULL; + } + + void obj_poke(void) + { + struct obj *obj; + + spin_lock(&global_list_lock); + obj = obj_list_peek(&global_list); + spin_unlock(&global_list_lock); - spin_lock(&global_list_lock); - obj = obj_list_peek(&global_list); - spin_unlock(&global_list_lock); + if (obj) { + obj->ops->poke(obj); + if (atomic_dec_and_test(&obj->refcnt)) + obj_destroy(obj); + } + } + + void obj_timeout(struct obj *obj) + { + spin_lock(&global_list_lock); + obj_list_del(obj); + spin_unlock(&global_list_lock); - if (obj) { - obj->ops->poke(obj); if (atomic_dec_and_test(&obj->refcnt)) obj_destroy(obj); } -} - -void obj_timeout(struct obj *obj) -{ - spin_lock(&global_list_lock); - obj_list_del(obj); - spin_unlock(&global_list_lock); - if (atomic_dec_and_test(&obj->refcnt)) - obj_destroy(obj); -} +.. note:: -(This is a simplification of the ARP queue management in the - generic neighbour discover code of the networking. Olaf Kirch - found a bug wrt. memory barriers in kfree_skb() that exposed - the atomic_t memory barrier requirements quite clearly.) + This is a simplification of the ARP queue management in the generic + neighbour discover code of the networking. Olaf Kirch found a bug wrt. + memory barriers in kfree_skb() that exposed the atomic_t memory barrier + requirements quite clearly. Given the above scheme, it must be the case that the obj->active update done by the obj list deletion be visible to other processors @@ -383,7 +394,7 @@ before the atomic counter decrement is performed. Otherwise, the counter could fall to zero, yet obj->active would still be set, thus triggering the assertion in obj_destroy(). The error -sequence looks like this: +sequence looks like this:: cpu 0 cpu 1 obj_poke() obj_timeout() @@ -420,6 +431,10 @@ same scheme. Another note is that the atomic_t operations returning values are extremely slow on an old 386. + +Atomic Bitmask +============== + We will now cover the atomic bitmask operations. You will find that their SMP and memory barrier semantics are similar in shape and scope to the atomic_t ops above. @@ -427,7 +442,7 @@ to the atomic_t ops above. Native atomic bit operations are defined to operate on objects aligned to the size of an "unsigned long" C data type, and are least of that size. The endianness of the bits within each "unsigned long" are the -native endianness of the cpu. +native endianness of the cpu. :: void set_bit(unsigned long nr, volatile unsigned long *addr); void clear_bit(unsigned long nr, volatile unsigned long *addr); @@ -437,7 +452,7 @@ These routines set, clear, and change, respectively, the bit number indicated by "nr" on the bit mask pointed to by "ADDR". They must execute atomically, yet there are no implicit memory barrier -semantics required of these interfaces. +semantics required of these interfaces. :: int test_and_set_bit(unsigned long nr, volatile unsigned long *addr); int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr); @@ -466,7 +481,7 @@ must provide explicit memory barrier semantics around their execution. All memory operations before the atomic bit operation call must be made visible globally before the atomic bit operation is made visible. Likewise, the atomic bit operation must be visible globally before any -subsequent memory operation is made visible. For example: +subsequent memory operation is made visible. For example:: obj->dead = 1; if (test_and_set_bit(0, &obj->flags)) @@ -479,7 +494,7 @@ done by test_and_set_bit() becomes visible. Likewise, the atomic memory operation done by test_and_set_bit() must become visible before "obj->killed = 1;" is visible. -Finally there is the basic operation: +Finally there is the basic operation:: int test_bit(unsigned long nr, __const__ volatile unsigned long *addr); @@ -488,13 +503,13 @@ pointed to by "addr". If explicit memory barriers are required around {set,clear}_bit() (which do not return a value, and thus does not need to provide memory barrier -semantics), two interfaces are provided: +semantics), two interfaces are provided:: void smp_mb__before_atomic(void); void smp_mb__after_atomic(void); They are used as follows, and are akin to their atomic_t operation -brothers: +brothers:: /* All memory operations before this call will * be globally visible before the clear_bit(). @@ -511,7 +526,7 @@ There are two special bitops with lock barrier semantics (acquire/release, same as spinlocks). These operate in the same way as their non-_lock/unlock postfixed variants, except that they are to provide acquire/release semantics, respectively. This means they can be used for bit_spin_trylock and -bit_spin_unlock type operations without specifying any more barriers. +bit_spin_unlock type operations without specifying any more barriers. :: int test_and_set_bit_lock(unsigned long nr, unsigned long *addr); void clear_bit_unlock(unsigned long nr, unsigned long *addr); @@ -526,7 +541,7 @@ provided. They are used in contexts where some other higher-level SMP locking scheme is being used to protect the bitmask, and thus less expensive non-atomic operations may be used in the implementation. They have names similar to the above bitmask operation interfaces, -except that two underscores are prefixed to the interface name. +except that two underscores are prefixed to the interface name. :: void __set_bit(unsigned long nr, volatile unsigned long *addr); void __clear_bit(unsigned long nr, volatile unsigned long *addr); @@ -542,9 +557,11 @@ The routines xchg() and cmpxchg() must provide the same exact memory-barrier semantics as the atomic and bit operations returning values. -Note: If someone wants to use xchg(), cmpxchg() and their variants, -linux/atomic.h should be included rather than asm/cmpxchg.h, unless -the code is in arch/* and can take care of itself. +.. note:: + + If someone wants to use xchg(), cmpxchg() and their variants, + linux/atomic.h should be included rather than asm/cmpxchg.h, unless the + code is in arch/* and can take care of itself. Spinlocks and rwlocks have memory barrier expectations as well. The rule to follow is simple: @@ -558,7 +575,7 @@ The rule to follow is simple: Which finally brings us to _atomic_dec_and_lock(). There is an architecture-neutral version implemented in lib/dec_and_lock.c, -but most platforms will wish to optimize this in assembler. +but most platforms will wish to optimize this in assembler. :: int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock); @@ -573,7 +590,7 @@ sure the spinlock operation is globally visible before any subsequent memory operation. We can demonstrate this operation more clearly if we define -an abstract atomic operation: +an abstract atomic operation:: long cas(long *mem, long old, long new); @@ -584,48 +601,48 @@ an abstract atomic operation: 3) Regardless, the current value at "mem" is returned. As an example usage, here is what an atomic counter update -might look like: +might look like:: -void example_atomic_inc(long *counter) -{ - long old, new, ret; + void example_atomic_inc(long *counter) + { + long old, new, ret; - while (1) { - old = *counter; - new = old + 1; + while (1) { + old = *counter; + new = old + 1; - ret = cas(counter, old, new); - if (ret == old) - break; - } -} - -Let's use cas() in order to build a pseudo-C atomic_dec_and_lock(): - -int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock) -{ - long old, new, ret; - int went_to_zero; - - went_to_zero = 0; - while (1) { - old = atomic_read(atomic); - new = old - 1; - if (new == 0) { - went_to_zero = 1; - spin_lock(lock); - } - ret = cas(atomic, old, new); - if (ret == old) - break; - if (went_to_zero) { - spin_unlock(lock); - went_to_zero = 0; + ret = cas(counter, old, new); + if (ret == old) + break; } } - return went_to_zero; -} +Let's use cas() in order to build a pseudo-C atomic_dec_and_lock():: + + int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock) + { + long old, new, ret; + int went_to_zero; + + went_to_zero = 0; + while (1) { + old = atomic_read(atomic); + new = old - 1; + if (new == 0) { + went_to_zero = 1; + spin_lock(lock); + } + ret = cas(atomic, old, new); + if (ret == old) + break; + if (went_to_zero) { + spin_unlock(lock); + went_to_zero = 0; + } + } + + return went_to_zero; + } Now, as far as memory barriers go, as long as spin_lock() strictly orders all subsequent memory operations (including @@ -635,6 +652,7 @@ Said another way, _atomic_dec_and_lock() must guarantee that a counter dropping to zero is never made visible before the spinlock being acquired. -Note that this also means that for the case where the counter -is not dropping to zero, there are no memory ordering -requirements. +.. note:: + + Note that this also means that for the case where the counter is not + dropping to zero, there are no memory ordering requirements. diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index f53555e..25b4e4a 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -8,6 +8,7 @@ Kernel and driver related documentation. :maxdepth: 1 assoc_array + atomic_ops local_ops workqueue diff --git a/Documentation/process/volatile-considered-harmful.rst b/Documentation/process/volatile-considered-harmful.rst index e0d042a..4934e65 100644 --- a/Documentation/process/volatile-considered-harmful.rst +++ b/Documentation/process/volatile-considered-harmful.rst @@ -1,3 +1,6 @@ + +.. _volatile_considered_harmful: + Why the "volatile" type class should not be used ------------------------------------------------ -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* [PATCH v4 4/4] firmware: remove warning at documentation generation time 2016-11-28 17:30 ` [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke ` (2 preceding siblings ...) 2016-11-28 17:30 ` [PATCH v4 3/4] Documentation/atomic_ops.txt: " Silvio Fricke @ 2016-11-28 17:30 ` Silvio Fricke 2016-12-01 0:51 ` [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Jonathan Corbet 4 siblings, 0 replies; 51+ messages in thread From: Silvio Fricke @ 2016-11-28 17:30 UTC (permalink / raw) To: linux-doc Cc: Jonathan Corbet, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula, Peter Zijlstra, Daniel Vetter, Silvio Fricke This patch removes following error at for `make htmldocs`. No functional change. ./drivers/base/firmware_class.c:1348: WARNING: Bullet list ends without a blank line; unexpected unindent. Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com> Reviewed-by: Mauro Carvalho Chehab <mchehab@s-opensource.com> --- drivers/base/firmware_class.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c index 22d1760..37b0221 100644 --- a/drivers/base/firmware_class.c +++ b/drivers/base/firmware_class.c @@ -1345,9 +1345,9 @@ static void request_firmware_work_func(struct work_struct *work) * * Asynchronous variant of request_firmware() for user contexts: * - sleep for as small periods as possible since it may - * increase kernel boot time of built-in device drivers - * requesting firmware in their ->probe() methods, if - * @gfp is GFP_KERNEL. + * increase kernel boot time of built-in device drivers + * requesting firmware in their ->probe() methods, if + * @gfp is GFP_KERNEL. * * - can't sleep at all if @gfp is GFP_ATOMIC. **/ -- git-series 0.9.1 ^ permalink raw reply related [flat|nested] 51+ messages in thread
* Re: [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops 2016-11-28 17:30 ` [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke ` (3 preceding siblings ...) 2016-11-28 17:30 ` [PATCH v4 4/4] firmware: remove warning at documentation generation time Silvio Fricke @ 2016-12-01 0:51 ` Jonathan Corbet 2016-12-01 15:49 ` Luis R. Rodriguez 4 siblings, 1 reply; 51+ messages in thread From: Jonathan Corbet @ 2016-12-01 0:51 UTC (permalink / raw) To: Silvio Fricke Cc: linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula, Peter Zijlstra, Daniel Vetter On Mon, 28 Nov 2016 18:30:52 +0100 Silvio Fricke <silvio.fricke@gmail.com> wrote: > Some more ReSTification of core-api's: assoc_array, atomic_ops and local_ops. A > fourth patch removes a warning about a bullet list without ending at > firmware_class.c OK, I've merged patches 1-3. Peter, if you ever want to edit atomic_ops, just send it meward when you're done and I'll deal with any formatting issues. Silvio, if you wanted to take this further, the assoc_array code is nicely decorated with kerneldoc comments; using those in the documentation would be a nice thing. There are some kerneldoc comments for the atomic functions, but that seems rather more erratic. Part 4 (firmware_class.c) is a bit off my turf, so I don't think I can apply it. From what I can tell looking at the history, Andrew Morton might be a logical target for that patch. Thanks, jon ^ permalink raw reply [flat|nested] 51+ messages in thread
* Re: [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops 2016-12-01 0:51 ` [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Jonathan Corbet @ 2016-12-01 15:49 ` Luis R. Rodriguez 0 siblings, 0 replies; 51+ messages in thread From: Luis R. Rodriguez @ 2016-12-01 15:49 UTC (permalink / raw) To: Jonathan Corbet Cc: Silvio Fricke, linux-doc, Ming Lei, Luis R . Rodriguez, Mauro Carvalho Chehab, linux-kernel, Jani Nikula, Peter Zijlstra, Daniel Vetter On Wed, Nov 30, 2016 at 05:51:29PM -0700, Jonathan Corbet wrote: > Part 4 (firmware_class.c) is a bit off my turf, so I don't think I can > apply it. From what I can tell looking at the history, Andrew Morton > might be a logical target for that patch. I'll review that patch, as I've recently taken up helping with maintenance for firmware_class. Provided its fine it could go through you Jonathan as I am not aware of other documentation changes in the pipeline. Will review shortly. Luis ^ permalink raw reply [flat|nested] 51+ messages in thread
end of thread, other threads:[~2016-12-01 16:45 UTC | newest] Thread overview: 51+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2016-11-24 12:42 [PATCH 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 2016-11-24 12:42 ` [PATCH 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke 2016-11-24 12:42 ` [PATCH 2/4] Documentation/atomic_ops.txt: " Silvio Fricke 2016-11-24 12:42 ` [PATCH 3/4] Documentation/local_ops.txt: " Silvio Fricke 2016-11-24 12:42 ` [PATCH 4/4] firmware: remove warning at documentation generation time Silvio Fricke 2016-11-24 19:46 ` [PATCH 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Mauro Carvalho Chehab 2016-11-25 10:02 ` [PATCH v2 " Silvio Fricke 2016-11-25 10:02 ` [PATCH v2 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke 2016-11-25 13:03 ` Mauro Carvalho Chehab 2016-11-25 10:02 ` [PATCH v2 2/4] Documentation/atomic_ops.txt: " Silvio Fricke 2016-11-25 13:25 ` Mauro Carvalho Chehab 2016-11-25 10:02 ` [PATCH v2 3/4] Documentation/local_ops.txt: " Silvio Fricke 2016-11-25 13:34 ` Mauro Carvalho Chehab 2016-11-25 13:47 ` S. Fricke 2016-11-25 13:58 ` Mauro Carvalho Chehab 2016-11-25 10:02 ` [PATCH v2 4/4] firmware: remove warning at documentation generation time Silvio Fricke 2016-11-25 13:34 ` Mauro Carvalho Chehab 2016-11-25 14:59 ` [PATCH v3 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 2016-11-25 14:59 ` [PATCH v3 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke 2016-11-25 20:53 ` Mauro Carvalho Chehab 2016-11-25 14:59 ` [PATCH v3 2/4] Documentation/atomic_ops.txt: " Silvio Fricke 2016-11-25 21:01 ` Mauro Carvalho Chehab 2016-11-25 21:58 ` Peter Zijlstra 2016-11-27 14:53 ` S. Fricke 2016-11-27 18:36 ` Jani Nikula 2016-11-27 23:59 ` Jonathan Corbet 2016-11-28 7:26 ` Peter Zijlstra 2016-11-28 8:44 ` Daniel Vetter 2016-11-28 10:20 ` Peter Zijlstra 2016-11-28 11:08 ` Mauro Carvalho Chehab 2016-11-28 11:54 ` Peter Zijlstra 2016-11-28 13:46 ` Daniel Vetter 2016-11-28 15:14 ` Mauro Carvalho Chehab 2016-11-28 11:16 ` Jani Nikula 2016-11-28 11:56 ` Peter Zijlstra 2016-11-28 12:32 ` Jani Nikula 2016-11-28 14:13 ` Daniel Vetter 2016-11-28 14:29 ` Jonathan Corbet 2016-11-25 14:59 ` [PATCH v3 3/4] Documentation/local_ops.txt: " Silvio Fricke 2016-11-25 21:10 ` Mauro Carvalho Chehab 2016-11-25 14:59 ` [PATCH v3 4/4] firmware: remove warning at documentation generation time Silvio Fricke 2016-11-25 20:50 ` Mauro Carvalho Chehab 2016-12-01 16:11 ` Luis R. Rodriguez 2016-12-01 16:45 ` Greg Kroah-Hartman 2016-11-28 17:30 ` [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Silvio Fricke 2016-11-28 17:30 ` [PATCH v4 1/4] Documentation/assoc_array.txt: convert to ReST markup Silvio Fricke 2016-11-28 17:30 ` [PATCH v4 2/4] Documentation/local_ops.txt: " Silvio Fricke 2016-11-28 17:30 ` [PATCH v4 3/4] Documentation/atomic_ops.txt: " Silvio Fricke 2016-11-28 17:30 ` [PATCH v4 4/4] firmware: remove warning at documentation generation time Silvio Fricke 2016-12-01 0:51 ` [PATCH v4 0/4] core-api ReST: assoc_array, atomic_ops, local_ops Jonathan Corbet 2016-12-01 15:49 ` Luis R. Rodriguez
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).