From patchwork Wed Jan 17 20:20:31 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Matthew Wilcox (Oracle)" X-Patchwork-Id: 10172617 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id 890F760230 for ; Thu, 18 Jan 2018 09:44:37 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 7802725D99 for ; Thu, 18 Jan 2018 09:44:37 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 6AA7E2623D; Thu, 18 Jan 2018 09:44:37 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-4.1 required=2.0 tests=BAYES_00,DKIM_SIGNED, RCVD_IN_DNSWL_MED,T_DKIM_INVALID autolearn=ham version=3.3.1 Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id 6304025D99 for ; Thu, 18 Jan 2018 09:44:36 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id C79406E58F; Thu, 18 Jan 2018 09:40:19 +0000 (UTC) X-Original-To: intel-gfx@lists.freedesktop.org Delivered-To: intel-gfx@lists.freedesktop.org Received: from bombadil.infradead.org (bombadil.infradead.org [65.50.211.133]) by gabe.freedesktop.org (Postfix) with ESMTPS id 3F1686E1C4 for ; Wed, 17 Jan 2018 20:22:41 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=infradead.org; s=bombadil.20170209; h=References:In-Reply-To:Message-Id: Date:Subject:Cc:To:From:Sender:Reply-To:MIME-Version:Content-Type: Content-Transfer-Encoding:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Id: List-Help:List-Unsubscribe:List-Subscribe:List-Post:List-Owner:List-Archive; bh=Dq/tcYJ6E7elZcH1LmXqlM5mqo6PLyAcOahC8fphEdU=; b=eeVJLwHeKaN0/EYcNzRx6vK3b i9Ev3bXN58bcv2zb2GF7RIZq1JPAGdlZhxcxJuYmdhSaYNtNvWzdRtG/7C5QDVXEUxm5/6pHpuJuB od3NzuiieHQC9e1FL0NO0HWwSf4k248dchlZ/iMv/ltjK8l2BIDD9CVCinq41xNfgoBY4TEW0VmTM EtIVADWg2jN02zAf3m65dtyZy7mVYMlZtmyz410/4O/xuDuC61WSLM50u2PAsert8fcUKmQFQLZoC f4ZN6AXA4mYo3FH8AfDlg0ux+b8mTc1wv//N5dyluky/uD4ZBrmxgy51eIRyrrOsc3BZsPdh6hs0a trNDZau1Q==; Received: from willy by bombadil.infradead.org with local (Exim 4.89 #1 (Red Hat Linux)) id 1ebuE7-0005iO-QU; Wed, 17 Jan 2018 20:22:27 +0000 From: Matthew Wilcox To: linux-kernel@vger.kernel.org Date: Wed, 17 Jan 2018 12:20:31 -0800 Message-Id: <20180117202203.19756-8-willy@infradead.org> X-Mailer: git-send-email 2.14.3 In-Reply-To: <20180117202203.19756-1-willy@infradead.org> References: <20180117202203.19756-1-willy@infradead.org> Cc: linux-s390@vger.kernel.org, David Howells , linux-nilfs@vger.kernel.org, Matthew Wilcox , linux-sh@vger.kernel.org, intel-gfx@lists.freedesktop.org, linux-usb@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-f2fs-devel@lists.sourceforge.net, linux-xfs@vger.kernel.org, linux-mm@kvack.org, iommu@lists.linux-foundation.org, Stefano Stabellini , linux-fsdevel@vger.kernel.org, cgroups@vger.kernel.org, Bjorn Andersson , linux-btrfs@vger.kernel.org Subject: [Intel-gfx] [PATCH v6 07/99] xarray: Add documentation X-BeenThere: intel-gfx@lists.freedesktop.org X-Mailman-Version: 2.1.18 Precedence: list List-Id: Intel graphics driver community testing & development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Errors-To: intel-gfx-bounces@lists.freedesktop.org Sender: "Intel-gfx" X-Virus-Scanned: ClamAV using ClamSMTP From: Matthew Wilcox This is documentation on how to use the XArray, not details about its internal implementation. Signed-off-by: Matthew Wilcox --- Documentation/core-api/index.rst | 1 + Documentation/core-api/xarray.rst | 361 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 362 insertions(+) create mode 100644 Documentation/core-api/xarray.rst diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index d5bbe035316d..eb16ba30aeb6 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -18,6 +18,7 @@ Core utilities local_ops workqueue genericirq + xarray flexible-arrays librs genalloc diff --git a/Documentation/core-api/xarray.rst b/Documentation/core-api/xarray.rst new file mode 100644 index 000000000000..914999c0bf3f --- /dev/null +++ b/Documentation/core-api/xarray.rst @@ -0,0 +1,361 @@ +.. SPDX-License-Identifier: CC-BY-SA-4.0 + +====== +XArray +====== + +:Author: Matthew Wilcox + +Overview +======== + +The XArray is an abstract data type which behaves like a very large array +of pointers. It meets many of the same needs as a hash or a conventional +resizable array. Unlike a hash, it allows you to sensibly go to the +next or previous entry in a cache-efficient manner. In contrast to +a resizable array, there is no need for copying data or changing MMU +mappings in order to grow the array. It is more memory-efficient, +parallelisable and cache friendly than a doubly-linked list. It takes +advantage of RCU to perform lookups without locking. + +The XArray implementation is efficient when the indices used are densely +clustered; hashing the object and using the hash as the index will not +perform well. The XArray is optimised for small indices, but still has +good performance with large indices. If your index can be larger than +``ULONG_MAX`` then the XArray is not the data type for you. The most +important user of the XArray is the page cache. + +A freshly-initialised XArray contains a ``NULL`` pointer at every index. +Each non-``NULL`` entry in the array has three bits associated with it +called tags. Each tag may be set or cleared independently of the others. +You can iterate over entries which are tagged. + +Normal pointers may be stored in the XArray directly. They must be 4-byte +aligned, which is true for any pointer returned from :c:func:`kmalloc` and +:c:func:`alloc_page`. It isn't true for arbitrary user-space pointers, +nor for function pointers. You can store pointers to statically allocated +objects, as long as those objects have an alignment of at least 4. + +You can also store integers between 0 and ``LONG_MAX`` in the XArray. +You must first convert it into an entry using :c:func:`xa_mk_value`. +When you retrieve an entry from the XArray, you can check whether it is +a value entry by calling :c:func:`xa_is_value`, and convert it back to +an integer by calling :c:func:`xa_to_value`. + +The XArray does not support storing :c:func:`IS_ERR` pointers as some +conflict with value entries or internal entries. + +An unusual feature of the XArray is the ability to create entries which +occupy a range of indices. Once stored to, looking up any index in +the range will return the same entry as looking up any other index in +the range. Setting a tag on one index will set it on all of them. +Storing to any index will store to all of them. Multi-index entries can +be explicitly split into smaller entries, or storing ``NULL`` into any +entry will cause the XArray to forget about the range. + +Normal API +========== + +Start by initialising an XArray, either with :c:func:`DEFINE_XARRAY` +for statically allocated XArrays or :c:func:`xa_init` for dynamically +allocated ones. + +You can then set entries using :c:func:`xa_store` and get entries +using :c:func:`xa_load`. xa_store will overwrite any entry with the +new entry and return the previous entry stored at that index. You can +use :c:func:`xa_erase` instead of calling :c:func:`xa_store` with a +%NULL entry. There is no difference between an entry that has never +been stored to and one that has most recently had ``NULL`` stored to it. + +You can conditionally replace an entry at an index by using +:c:func:`xa_cmpxchg`. Like :c:func:`cmpxchg`, it will only succeed if +the entry at that index has the 'old' value. It also returns the entry +which was at that index; if it returns the same entry which was passed as +'old', then :c:func:`xa_cmpxchg` succeeded. + +If you want to only store a new entry to an index if the current entry +at that index is ``NULL``, you can use :c:func:`xa_insert` which +returns ``-EEXIST`` if the entry is not empty. + +Calling :c:func:`xa_reserve` ensures that there is enough memory allocated +to store an entry at the specified index. This is not normally needed, +but some users have a complicated locking scheme. + +You can enquire whether a tag is set on an entry by using +:c:func:`xa_get_tag`. If the entry is not ``NULL``, you can set a tag +on it by using :c:func:`xa_set_tag` and remove the tag from an entry by +calling :c:func:`xa_clear_tag`. You can ask whether any entry in the +XArray has a particular tag set by calling :c:func:`xa_tagged`. + +You can copy entries out of the XArray into a plain array by calling +:c:func:`xa_extract`. Or you can iterate over the present entries in +the XArray by calling :c:func:`xa_for_each`. You may prefer to use +:c:func:`xa_find` or :c:func:`xa_find_after` to move to the next present +entry in the XArray. + +Finally, you can remove all entries from an XArray by calling +:c:func:`xa_destroy`. If the XArray entries are pointers, you may wish +to free the entries first. You can do this by iterating over all present +entries in the XArray using the :c:func:`xa_for_each` iterator. + +Memory allocation +----------------- + +The :c:func:`xa_store`, :c:func:`xa_cmpxchg`, :c:func:`xa_reserve` +and :c:func:`xa_insert` functions take a gfp_t parameter in case +the XArray needs to allocate memory to store this entry. If the entry +being stored is ``NULL``, no memory allocation needs to be performed, +and the GFP flags specified will be ignored. + +It is possible for no memory to be allocatable, particularly if you pass +a restrictive set of GFP flags. In that case, the functions return a +special value which can be turned into an errno using :c:func:`xa_err`. +If you don't need to know exactly which error occurred, using +:c:func:`xa_is_err` is slightly more efficient. + +Locking +------- + +When using the Normal API, you do not have to worry about locking. +The XArray uses RCU and an internal spinlock to synchronise access: + +No lock needed: + * :c:func:`xa_empty` + * :c:func:`xa_tagged` + +Takes RCU read lock: + * :c:func:`xa_load` + * :c:func:`xa_for_each` + * :c:func:`xa_find` + * :c:func:`xa_find_after` + * :c:func:`xa_extract` + * :c:func:`xa_get_tag` + +Takes xa_lock internally: + * :c:func:`xa_store` + * :c:func:`xa_insert` + * :c:func:`xa_erase` + * :c:func:`xa_cmpxchg` + * :c:func:`xa_reserve` + * :c:func:`xa_destroy` + * :c:func:`xa_set_tag` + * :c:func:`xa_clear_tag` + +Assumes xa_lock held on entry: + * :c:func:`__xa_store` + * :c:func:`__xa_insert` + * :c:func:`__xa_erase` + * :c:func:`__xa_cmpxchg` + * :c:func:`__xa_set_tag` + * :c:func:`__xa_clear_tag` + +If you want to take advantage of the lock to protect the data structures +that you are storing in the XArray, you can call :c:func:`xa_lock` +before calling :c:func:`xa_load`, then take a reference count on the +object you have found before calling :c:func:`xa_unlock`. This will +prevent stores from removing the object from the array between looking +up the object and incrementing the refcount. You can also use RCU to +avoid dereferencing freed memory, but an explanation of that is beyond +the scope of this document. + +The XArray does not disable interrupts or softirqs while modifying +the array. It is safe to read the XArray from interrupt or softirq +context as the RCU lock provides enough protection. + +If, for example, you want to store entries in the XArray in process +context and then erase them in softirq context, you can do that this way:: + + foo_init(struct foo *foo) + { + xa_init_flags(&foo->array, XA_FLAGS_LOCK_BH); + } + + foo_store(struct foo *foo, unsigned long index, void *entry) + { + xa_lock_bh(&foo->array); + __xa_store(&foo->array, index, entry, GFP_KERNEL); + foo->count++; + xa_unlock_bh(&foo->array); + } + + /* foo_erase() is only called from softirq context */ + foo_erase(struct foo *foo, unsigned long index) + { + xa_erase(&foo->array, index); + } + +If you are going to modify the XArray from interrupt or softirq context, +you need to initialise the array using :c:func:`xa_init_flags`, passing +``XA_FLAGS_LOCK_IRQ`` or ``XA_FLAGS_LOCK_BH``. + +The above example also shows a common pattern of wanting to extend the +coverage of the xa_lock on the store side to protect some statistics +associated with the array. + +Sharing the XArray with interrupt context is also possible, either +using :c:func:`xa_lock_irqsave` in both the interrupt handler and process +context, or :c:func:`xa_lock_irq` in process context and :c:func:`xa_lock` +in the interrupt handler. + +Sometimes you need to protect access to the XArray with a mutex because +that lock sits above another mutex in the locking hierarchy. That does +not entitle you to use functions like :c:func:`__xa_erase` without taking +the xa_lock; the xa_lock is used for lockdep validation and will be used +for other purposes in the future. + +The :c:func:`__xa_set_tag` and :c:func:`__xa_clear_tag` functions are also +available for situations where you look up an entry and want to atomically +set or clear a tag. It may be more efficient to use the advanced API +in this case, as it will save you from walking the tree twice. + +Advanced API +============ + +The advanced API offers more flexibility and better performance at the +cost of an interface which can be harder to use and has fewer safeguards. +No locking is done for you by the advanced API, and you are required +to use the xa_lock while modifying the array. You can choose whether +to use the xa_lock or the RCU lock while doing read-only operations on +the array. You can mix advanced and normal operations on the same array; +indeed the normal API is implemented in terms of the advanced API. The +advanced API is only available to modules with a GPL-compatible license. + +The advanced API is based around the xa_state. This is an opaque data +structure which you declare on the stack using the :c:func:`XA_STATE` +macro. This macro initialises the xa_state ready to start walking +around the XArray. It is used as a cursor to maintain the position +in the XArray and let you compose various operations together without +having to restart from the top every time. + +The xa_state is also used to store errors. You can call +:c:func:`xas_error` to retrieve the error. All operations check whether +the xa_state is in an error state before proceeding, so there's no need +for you to check for an error after each call; you can make multiple +calls in succession and only check at a convenient point. The only +errors currently generated by the xarray code itself are %ENOMEM and +%EINVAL, but it supports arbitrary errors in case you want to call +:c:func:`xas_set_err` yourself. + +If the xa_state is holding an %ENOMEM error, calling :c:func:`xas_nomem` +will attempt to allocate more memory using the specified gfp flags and +cache it in the xa_state for the next attempt. The idea is that you take +the xa_lock, attempt the operation and drop the lock. The operation +attempts to allocate memory while holding the lock, but it is more +likely to fail. Once you have dropped the lock, :c:func:`xas_nomem` +can try harder to allocate more memory. It will return ``true`` if it +is worth retrying the operation (i.e. that there was a memory error *and* +more memory was allocated). If it has previously allocated memory, and +that memory wasn't used, and there is no error (or some error that isn't +%ENOMEM), then it will free the memory previously allocated. + +Internal Entries +---------------- + +The XArray reserves some entries for its own purposes. These are never +exposed through the normal API, but when using the advanced API, it's +possible to see them. Usually the best way to handle them is to pass them +to :c:func:`xas_retry`, and retry the operation if it returns ``true``. + +.. flat-table:: + :widths: 1 1 6 + + * - Name + - Test + - Usage + + * - Node + - :c:func:`xa_is_node` + - An XArray node. Should never be visible; all functions should recurse + into an XArray node. + + * - Sibling + - :c:func:`xa_is_sibling` + - A non-canonical entry for a multi-index entry. The value indicates + which slot in this node has the canonical entry. + + * - Retry + - :c:func:`xa_is_retry` + - This entry is currently being modified by a thread which has the + xa_lock. The node containing this entry may be freed at the end of + this RCU period. You should restart the lookup from the head of the + array. + +Other internal entries may be added in the future. As far as possible, they +will be handled by :c:func:`xas_retry`. + +Additional functionality +------------------------ + +The :c:func:`xas_create` function ensures that there is somewhere in the +XArray to store an entry. It will store ENOMEM in the xa_state if it +cannot allocate memory. You do not normally need to call this function +yourself as it is called by :c:func:`xas_store`. + +You can use :c:func:`xas_init_tags` to reset the tags on an entry +to their default state. This is usually all tags clear, unless the +XArray is marked with ``XA_FLAGS_TRACK_FREE``, in which case tag 0 is set +and all other tags are clear. Replacing one entry with another using +:c:func:`xas_store` will not reset the tags on that entry; if you want +the tags reset, you should do that explicitly. + +The :c:func:`xas_load` will walk the xa_state as close to the entry +as it can. If you know the xa_state has already been walked to the +entry and need to check that the entry hasn't changed, you can use +:c:func:`xas_reload` to save a function call. + +If you need to move to a different index in the XArray, call +:c:func:`xas_set`. This reinitialises the cursor, which will generally +have the effect of making the next operation walk the cursor to the +desired spot in the tree. If you want to move to the next or previous +index, call :c:func:`xas_next` or :c:func:`xas_prev`. Setting the index +does not walk the cursor around the array so does not require a lock to +be held, while moving to the next or previous index does. + +You can create a multi-index entry by using :c:func:`xas_set_order`. +If a load or find operation finds a multi-index entry, the index in the +xa_state will be the one searched for, and not necessarily the +lowest or highest index used by the entry. +Currently the only supported multi-index entries supported are powers +of two, but there are two potential users of arbitrary ranges, so that +functionality may be added soon. + +You can search for the next present entry using :c:func:`xas_find`. This +is the equivalent of both :c:func:`xa_find` and :c:func:`xa_find_after`; +if the cursor has been walked to an entry, then it will find the next +entry after the one currently referenced. If not, it will return the +entry at the index of the xa_state. Using :c:func:`xas_next_entry` to +move to the next present entry instead of :c:func:`xas_find` will save +a function call in the majority of cases at the expense of emitting more +inline code. + +The :c:func:`xas_find_tag` function is similar, returning the first tagged +entry after the entry referenced by the xa_state if it has already been +walked, and returning the entry at the index of the xa_state if it is +tagged, and the xa_state has not been walked. The :c:func:`xas_next_tag` +function is the equivalent of :c:func:`xas_next_entry`. + +When iterating over a range of the XArray using :c:func:`xas_for_each` +or :c:func:`xas_for_each_tag`, it may be necessary to temporarily stop +the iteration. The :c:func:`xas_pause` function exists for this purpose. +After you have done the necessary work and wish to resume, the xa_state +is in an appropriate state to continue the iteration after the entry +you last processed. If you have interrupts disabled while iterating, +then it is good manners to pause the iteration and reenable interrupts +every ``XA_CHECK_SCHED`` entries. + +The :c:func:`xas_get_tag`, :c:func:`xas_set_tag` and +:c:func:`xas_clear_tag` functions require the xa_state cursor to have +been moved to the appropriate location in the xarray; they will do +nothing if you have called :c:func:`xas_pause` or :c:func:`xas_set` +immediately before. + +You can call :c:func:`xas_set_update` to have a callback function +called each time the XArray updates a node. This is used by the page +cache workingset code to maintain its list of nodes which contain only +shadow entries. + +Functions and structures +======================== + +.. kernel-doc:: include/linux/xarray.h +.. kernel-doc:: lib/xarray.c