X-Git-Url: http://ftp.safe.ca/?p=safe%2Fjmp%2Flinux-2.6;a=blobdiff_plain;f=mm%2Freadahead.c;h=033bc135a41f3885cae863ca0c9f5185d6c5d665;hp=7f9bf588c93630800629b71201939866721ca4ca;hb=e071041be037eca208b62b84469a06bdfc692bea;hpb=f615bfca468c9b80ed2d09be5fdbaf470a32c045 diff --git a/mm/readahead.c b/mm/readahead.c index 7f9bf58..033bc13 100644 --- a/mm/readahead.c +++ b/mm/readahead.c @@ -3,7 +3,7 @@ * * Copyright (C) 2002, Linus Torvalds * - * 09Apr2002 akpm@zip.com.au + * 09Apr2002 Andrew Morton * Initial version. */ @@ -15,27 +15,7 @@ #include #include #include - -void default_unplug_io_fn(struct backing_dev_info *bdi, struct page *page) -{ -} -EXPORT_SYMBOL(default_unplug_io_fn); - -/* - * Convienent macros for min/max read-ahead pages. - * Note that MAX_RA_PAGES is rounded down, while MIN_RA_PAGES is rounded up. - * The latter is necessary for systems with large page size(i.e. 64k). - */ -#define MAX_RA_PAGES (VM_MAX_READAHEAD*1024 / PAGE_CACHE_SIZE) -#define MIN_RA_PAGES DIV_ROUND_UP(VM_MIN_READAHEAD*1024, PAGE_CACHE_SIZE) - -struct backing_dev_info default_backing_dev_info = { - .ra_pages = MAX_RA_PAGES, - .state = 0, - .capabilities = BDI_CAP_MAP_COPY, - .unplug_io_fn = default_unplug_io_fn, -}; -EXPORT_SYMBOL_GPL(default_backing_dev_info); +#include /* * Initialise a struct file's readahead state. Assumes that the caller has @@ -45,88 +25,48 @@ void file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping) { ra->ra_pages = mapping->backing_dev_info->ra_pages; - ra->prev_index = -1; + ra->prev_pos = -1; } EXPORT_SYMBOL_GPL(file_ra_state_init); -/* - * Return max readahead size for this inode in number-of-pages. - */ -static inline unsigned long get_max_readahead(struct file_ra_state *ra) -{ - return ra->ra_pages; -} - -static inline unsigned long get_min_readahead(struct file_ra_state *ra) -{ - return MIN_RA_PAGES; -} - -static inline void reset_ahead_window(struct file_ra_state *ra) -{ - /* - * ... but preserve ahead_start + ahead_size value, - * see 'recheck:' label in page_cache_readahead(). - * Note: We never use ->ahead_size as rvalue without - * checking ->ahead_start != 0 first. - */ - ra->ahead_size += ra->ahead_start; - ra->ahead_start = 0; -} - -static inline void ra_off(struct file_ra_state *ra) -{ - ra->start = 0; - ra->flags = 0; - ra->size = 0; - reset_ahead_window(ra); - return; -} +#define list_to_page(head) (list_entry((head)->prev, struct page, lru)) /* - * Set the initial window size, round to next power of 2 and square - * for small size, x 4 for medium, and x 2 for large - * for 128k (32 page) max ra - * 1-8 page = 32k initial, > 8 page = 128k initial + * see if a page needs releasing upon read_cache_pages() failure + * - the caller of read_cache_pages() may have set PG_private or PG_fscache + * before calling, such as the NFS fs marking pages that are cached locally + * on disk, thus we need to give the fs a chance to clean up in the event of + * an error */ -static unsigned long get_init_ra_size(unsigned long size, unsigned long max) +static void read_cache_pages_invalidate_page(struct address_space *mapping, + struct page *page) { - unsigned long newsize = roundup_pow_of_two(size); - - if (newsize <= max / 32) - newsize = newsize * 4; - else if (newsize <= max / 4) - newsize = newsize * 2; - else - newsize = max; - return newsize; + if (page_has_private(page)) { + if (!trylock_page(page)) + BUG(); + page->mapping = mapping; + do_invalidatepage(page, 0); + page->mapping = NULL; + unlock_page(page); + } + page_cache_release(page); } /* - * Set the new window size, this is called only when I/O is to be submitted, - * not for each call to readahead. If a cache miss occured, reduce next I/O - * size, else increase depending on how close to max we are. + * release a list of pages, invalidating them first if need be */ -static inline unsigned long get_next_ra_size(struct file_ra_state *ra) +static void read_cache_pages_invalidate_pages(struct address_space *mapping, + struct list_head *pages) { - unsigned long max = get_max_readahead(ra); - unsigned long min = get_min_readahead(ra); - unsigned long cur = ra->size; - unsigned long newsize; + struct page *victim; - if (ra->flags & RA_FLAG_MISS) { - ra->flags &= ~RA_FLAG_MISS; - newsize = max((cur - 2), min); - } else if (cur < max / 16) { - newsize = 4 * cur; - } else { - newsize = 2 * cur; + while (!list_empty(pages)) { + victim = list_to_page(pages); + list_del(&victim->lru); + read_cache_pages_invalidate_page(mapping, victim); } - return min(newsize, max); } -#define list_to_page(head) (list_entry((head)->prev, struct page, lru)) - /** * read_cache_pages - populate an address space with some pages & start reads against them * @mapping: the address_space @@ -141,28 +81,25 @@ int read_cache_pages(struct address_space *mapping, struct list_head *pages, int (*filler)(void *, struct page *), void *data) { struct page *page; - struct pagevec lru_pvec; int ret = 0; - pagevec_init(&lru_pvec, 0); - while (!list_empty(pages)) { page = list_to_page(pages); list_del(&page->lru); - if (add_to_page_cache(page, mapping, page->index, GFP_KERNEL)) { - page_cache_release(page); + if (add_to_page_cache_lru(page, mapping, + page->index, GFP_KERNEL)) { + read_cache_pages_invalidate_page(mapping, page); continue; } + page_cache_release(page); + ret = filler(data, page); - if (!pagevec_add(&lru_pvec, page)) - __pagevec_lru_add(&lru_pvec); - if (ret) { - put_pages_list(pages); + if (unlikely(ret)) { + read_cache_pages_invalidate_pages(mapping, pages); break; } task_io_account_read(PAGE_CACHE_SIZE); } - pagevec_lru_add(&lru_pvec); return ret; } @@ -172,7 +109,6 @@ static int read_pages(struct address_space *mapping, struct file *filp, struct list_head *pages, unsigned nr_pages) { unsigned page_idx; - struct pagevec lru_pvec; int ret; if (mapping->a_ops->readpages) { @@ -182,94 +118,27 @@ static int read_pages(struct address_space *mapping, struct file *filp, goto out; } - pagevec_init(&lru_pvec, 0); for (page_idx = 0; page_idx < nr_pages; page_idx++) { struct page *page = list_to_page(pages); list_del(&page->lru); - if (!add_to_page_cache(page, mapping, + if (!add_to_page_cache_lru(page, mapping, page->index, GFP_KERNEL)) { mapping->a_ops->readpage(filp, page); - if (!pagevec_add(&lru_pvec, page)) - __pagevec_lru_add(&lru_pvec); - } else - page_cache_release(page); + } + page_cache_release(page); } - pagevec_lru_add(&lru_pvec); ret = 0; out: return ret; } /* - * Readahead design. - * - * The fields in struct file_ra_state represent the most-recently-executed - * readahead attempt: - * - * start: Page index at which we started the readahead - * size: Number of pages in that read - * Together, these form the "current window". - * Together, start and size represent the `readahead window'. - * prev_index: The page which the readahead algorithm most-recently inspected. - * It is mainly used to detect sequential file reading. - * If page_cache_readahead sees that it is again being called for - * a page which it just looked at, it can return immediately without - * making any state changes. - * offset: Offset in the prev_index where the last read ended - used for - * detection of sequential file reading. - * ahead_start, - * ahead_size: Together, these form the "ahead window". - * ra_pages: The externally controlled max readahead for this fd. - * - * When readahead is in the off state (size == 0), readahead is disabled. - * In this state, prev_index is used to detect the resumption of sequential I/O. - * - * The readahead code manages two windows - the "current" and the "ahead" - * windows. The intent is that while the application is walking the pages - * in the current window, I/O is underway on the ahead window. When the - * current window is fully traversed, it is replaced by the ahead window - * and the ahead window is invalidated. When this copying happens, the - * new current window's pages are probably still locked. So - * we submit a new batch of I/O immediately, creating a new ahead window. - * - * So: - * - * ----|----------------|----------------|----- - * ^start ^start+size - * ^ahead_start ^ahead_start+ahead_size - * - * ^ When this page is read, we submit I/O for the - * ahead window. - * - * A `readahead hit' occurs when a read request is made against a page which is - * the next sequential page. Ahead window calculations are done only when it - * is time to submit a new IO. The code ramps up the size agressively at first, - * but slow down as it approaches max_readhead. - * - * Any seek/ramdom IO will result in readahead being turned off. It will resume - * at the first sequential access. - * - * There is a special-case: if the first page which the application tries to - * read happens to be the first page of the file, it is assumed that a linear - * read is about to happen and the window is immediately set to the initial size - * based on I/O request size and the max_readahead. - * - * This function is to be called for every read request, rather than when - * it is time to perform readahead. It is called only once for the entire I/O - * regardless of size unless readahead is unable to start enough I/O to satisfy - * the request (I/O request > max_readahead). - */ - -/* - * do_page_cache_readahead actually reads a chunk of disk. It allocates all + * __do_page_cache_readahead() actually reads a chunk of disk. It allocates all * the pages first, then submits them all for I/O. This avoids the very bad * behaviour which would occur if page allocations are causing VM writeback. * We really don't want to intermingle reads and writes like that. * * Returns the number of pages requested, or the maximum amount of I/O allowed. - * - * do_page_cache_readahead() returns -1 if it encountered request queue - * congestion. */ static int __do_page_cache_readahead(struct address_space *mapping, struct file *filp, @@ -292,20 +161,19 @@ __do_page_cache_readahead(struct address_space *mapping, struct file *filp, /* * Preallocate as many pages as we will need. */ - read_lock_irq(&mapping->tree_lock); for (page_idx = 0; page_idx < nr_to_read; page_idx++) { pgoff_t page_offset = offset + page_idx; - + if (page_offset > end_index) break; + rcu_read_lock(); page = radix_tree_lookup(&mapping->page_tree, page_offset); + rcu_read_unlock(); if (page) continue; - read_unlock_irq(&mapping->tree_lock); page = page_cache_alloc_cold(mapping); - read_lock_irq(&mapping->tree_lock); if (!page) break; page->index = page_offset; @@ -314,7 +182,6 @@ __do_page_cache_readahead(struct address_space *mapping, struct file *filp, SetPageReadahead(page); ret++; } - read_unlock_irq(&mapping->tree_lock); /* * Now start the IO. We ignore I/O errors - if the page is not @@ -340,6 +207,7 @@ int force_page_cache_readahead(struct address_space *mapping, struct file *filp, if (unlikely(!mapping->a_ops->readpage && !mapping->a_ops->readpages)) return -EINVAL; + nr_to_read = max_sane_readahead(nr_to_read); while (nr_to_read) { int err; @@ -361,234 +229,335 @@ int force_page_cache_readahead(struct address_space *mapping, struct file *filp, } /* - * Check how effective readahead is being. If the amount of started IO is - * less than expected then the file is partly or fully in pagecache and - * readahead isn't helping. - * + * Given a desired number of PAGE_CACHE_SIZE readahead pages, return a + * sensible upper limit. */ -static inline int check_ra_success(struct file_ra_state *ra, - unsigned long nr_to_read, unsigned long actual) +unsigned long max_sane_readahead(unsigned long nr) { - if (actual == 0) { - ra->cache_hit += nr_to_read; - if (ra->cache_hit >= VM_MAX_CACHE_HIT) { - ra_off(ra); - ra->flags |= RA_FLAG_INCACHE; - return 0; - } - } else { - ra->cache_hit=0; - } - return 1; + return min(nr, (node_page_state(numa_node_id(), NR_INACTIVE_FILE) + + node_page_state(numa_node_id(), NR_FREE_PAGES)) / 2); } /* - * This version skips the IO if the queue is read-congested, and will tell the - * block layer to abandon the readahead if request allocation would block. - * - * force_page_cache_readahead() will ignore queue congestion and will block on - * request queues. + * Submit IO for the read-ahead request in file_ra_state. */ -int do_page_cache_readahead(struct address_space *mapping, struct file *filp, - pgoff_t offset, unsigned long nr_to_read) +unsigned long ra_submit(struct file_ra_state *ra, + struct address_space *mapping, struct file *filp) { - if (bdi_read_congested(mapping->backing_dev_info)) - return -1; + int actual; + + actual = __do_page_cache_readahead(mapping, filp, + ra->start, ra->size, ra->async_size); - return __do_page_cache_readahead(mapping, filp, offset, nr_to_read, 0); + return actual; } /* - * Read 'nr_to_read' pages starting at page 'offset'. If the flag 'block' - * is set wait till the read completes. Otherwise attempt to read without - * blocking. - * Returns 1 meaning 'success' if read is successful without switching off - * readahead mode. Otherwise return failure. + * Set the initial window size, round to next power of 2 and square + * for small size, x 4 for medium, and x 2 for large + * for 128k (32 page) max ra + * 1-8 page = 32k initial, > 8 page = 128k initial */ -static int -blockable_page_cache_readahead(struct address_space *mapping, struct file *filp, - pgoff_t offset, unsigned long nr_to_read, - struct file_ra_state *ra, int block) +static unsigned long get_init_ra_size(unsigned long size, unsigned long max) { - int actual; - - if (!block && bdi_read_congested(mapping->backing_dev_info)) - return 0; + unsigned long newsize = roundup_pow_of_two(size); - actual = __do_page_cache_readahead(mapping, filp, offset, nr_to_read, 0); + if (newsize <= max / 32) + newsize = newsize * 4; + else if (newsize <= max / 4) + newsize = newsize * 2; + else + newsize = max; - return check_ra_success(ra, nr_to_read, actual); + return newsize; } -static int make_ahead_window(struct address_space *mapping, struct file *filp, - struct file_ra_state *ra, int force) +/* + * Get the previous window size, ramp it up, and + * return it as the new window size. + */ +static unsigned long get_next_ra_size(struct file_ra_state *ra, + unsigned long max) { - int block, ret; - - ra->ahead_size = get_next_ra_size(ra); - ra->ahead_start = ra->start + ra->size; - - block = force || (ra->prev_index >= ra->ahead_start); - ret = blockable_page_cache_readahead(mapping, filp, - ra->ahead_start, ra->ahead_size, ra, block); - - if (!ret && !force) { - /* A read failure in blocking mode, implies pages are - * all cached. So we can safely assume we have taken - * care of all the pages requested in this call. - * A read failure in non-blocking mode, implies we are - * reading more pages than requested in this call. So - * we safely assume we have taken care of all the pages - * requested in this call. - * - * Just reset the ahead window in case we failed due to - * congestion. The ahead window will any way be closed - * in case we failed due to excessive page cache hits. - */ - reset_ahead_window(ra); - } + unsigned long cur = ra->size; + unsigned long newsize; - return ret; + if (cur < max / 16) + newsize = 4 * cur; + else + newsize = 2 * cur; + + return min(newsize, max); } -/** - * page_cache_readahead - generic adaptive readahead - * @mapping: address_space which holds the pagecache and I/O vectors - * @ra: file_ra_state which holds the readahead state - * @filp: passed on to ->readpage() and ->readpages() - * @offset: start offset into @mapping, in PAGE_CACHE_SIZE units - * @req_size: hint: total size of the read which the caller is performing in - * PAGE_CACHE_SIZE units +/* + * On-demand readahead design. + * + * The fields in struct file_ra_state represent the most-recently-executed + * readahead attempt: + * + * |<----- async_size ---------| + * |------------------- size -------------------->| + * |==================#===========================| + * ^start ^page marked with PG_readahead + * + * To overlap application thinking time and disk I/O time, we do + * `readahead pipelining': Do not wait until the application consumed all + * readahead pages and stalled on the missing page at readahead_index; + * Instead, submit an asynchronous readahead I/O as soon as there are + * only async_size pages left in the readahead window. Normally async_size + * will be equal to size, for maximum pipelining. + * + * In interleaved sequential reads, concurrent streams on the same fd can + * be invalidating each other's readahead state. So we flag the new readahead + * page at (start+size-async_size) with PG_readahead, and use it as readahead + * indicator. The flag won't be set on already cached pages, to avoid the + * readahead-for-nothing fuss, saving pointless page cache lookups. * - * page_cache_readahead() is the main function. It performs the adaptive - * readahead window size management and submits the readahead I/O. + * prev_pos tracks the last visited byte in the _previous_ read request. + * It should be maintained by the caller, and will be used for detecting + * small random reads. Note that the readahead algorithm checks loosely + * for sequential patterns. Hence interleaved reads might be served as + * sequential ones. * - * Note that @filp is purely used for passing on to the ->readpage[s]() - * handler: it may refer to a different file from @mapping (so we may not use - * @filp->f_mapping or @filp->f_path.dentry->d_inode here). - * Also, @ra may not be equal to &@filp->f_ra. + * There is a special-case: if the first page which the application tries to + * read happens to be the first page of the file, it is assumed that a linear + * read is about to happen and the window is immediately set to the initial size + * based on I/O request size and the max_readahead. * + * The code ramps up the readahead size aggressively at first, but slow down as + * it approaches max_readhead. + */ + +/* + * Count contiguously cached pages from @offset-1 to @offset-@max, + * this count is a conservative estimation of + * - length of the sequential read sequence, or + * - thrashing threshold in memory tight systems */ -unsigned long -page_cache_readahead(struct address_space *mapping, struct file_ra_state *ra, - struct file *filp, pgoff_t offset, unsigned long req_size) +static pgoff_t count_history_pages(struct address_space *mapping, + struct file_ra_state *ra, + pgoff_t offset, unsigned long max) { - unsigned long max, newsize; - int sequential; + pgoff_t head; + + rcu_read_lock(); + head = radix_tree_prev_hole(&mapping->page_tree, offset - 1, max); + rcu_read_unlock(); + + return offset - 1 - head; +} + +/* + * page cache context based read-ahead + */ +static int try_context_readahead(struct address_space *mapping, + struct file_ra_state *ra, + pgoff_t offset, + unsigned long req_size, + unsigned long max) +{ + pgoff_t size; + + size = count_history_pages(mapping, ra, offset, max); /* - * We avoid doing extra work and bogusly perturbing the readahead - * window expansion logic. + * no history pages: + * it could be a random read */ - if (offset == ra->prev_index && --req_size) - ++offset; + if (!size) + return 0; - /* Note that prev_index == -1 if it is a first read */ - sequential = (offset == ra->prev_index + 1); - ra->prev_index = offset; - ra->prev_offset = 0; + /* + * starts from beginning of file: + * it is a strong indication of long-run stream (or whole-file-read) + */ + if (size >= offset) + size *= 2; - max = get_max_readahead(ra); - newsize = min(req_size, max); + ra->start = offset; + ra->size = get_init_ra_size(size + req_size, max); + ra->async_size = ra->size; - /* No readahead or sub-page sized read or file already in cache */ - if (newsize == 0 || (ra->flags & RA_FLAG_INCACHE)) - goto out; + return 1; +} - ra->prev_index += newsize - 1; +/* + * A minimal readahead algorithm for trivial sequential/random reads. + */ +static unsigned long +ondemand_readahead(struct address_space *mapping, + struct file_ra_state *ra, struct file *filp, + bool hit_readahead_marker, pgoff_t offset, + unsigned long req_size) +{ + unsigned long max = max_sane_readahead(ra->ra_pages); /* - * Special case - first read at start of file. We'll assume it's - * a whole-file read and grow the window fast. Or detect first - * sequential access + * start of file */ - if (sequential && ra->size == 0) { - ra->size = get_init_ra_size(newsize, max); - ra->start = offset; - if (!blockable_page_cache_readahead(mapping, filp, offset, - ra->size, ra, 1)) - goto out; - - /* - * If the request size is larger than our max readahead, we - * at least want to be sure that we get 2 IOs in flight and - * we know that we will definitly need the new I/O. - * once we do this, subsequent calls should be able to overlap - * IOs,* thus preventing stalls. so issue the ahead window - * immediately. - */ - if (req_size >= max) - make_ahead_window(mapping, filp, ra, 1); + if (!offset) + goto initial_readahead; - goto out; + /* + * It's the expected callback offset, assume sequential access. + * Ramp up sizes, and push forward the readahead window. + */ + if ((offset == (ra->start + ra->size - ra->async_size) || + offset == (ra->start + ra->size))) { + ra->start += ra->size; + ra->size = get_next_ra_size(ra, max); + ra->async_size = ra->size; + goto readit; } /* - * Now handle the random case: - * partial page reads and first access were handled above, - * so this must be the next page otherwise it is random + * Hit a marked page without valid readahead state. + * E.g. interleaved reads. + * Query the pagecache for async_size, which normally equals to + * readahead size. Ramp it up and use it as the new readahead size. */ - if (!sequential) { - ra_off(ra); - blockable_page_cache_readahead(mapping, filp, offset, - newsize, ra, 1); - goto out; + if (hit_readahead_marker) { + pgoff_t start; + + rcu_read_lock(); + start = radix_tree_next_hole(&mapping->page_tree, offset+1,max); + rcu_read_unlock(); + + if (!start || start - offset > max) + return 0; + + ra->start = start; + ra->size = start - offset; /* old async_size */ + ra->size += req_size; + ra->size = get_next_ra_size(ra, max); + ra->async_size = ra->size; + goto readit; } /* - * If we get here we are doing sequential IO and this was not the first - * occurence (ie we have an existing window) + * oversize read */ - if (ra->ahead_start == 0) { /* no ahead window yet */ - if (!make_ahead_window(mapping, filp, ra, 0)) - goto recheck; - } + if (req_size > max) + goto initial_readahead; + + /* + * sequential cache miss + */ + if (offset - (ra->prev_pos >> PAGE_CACHE_SHIFT) <= 1UL) + goto initial_readahead; /* - * Already have an ahead window, check if we crossed into it. - * If so, shift windows and issue a new ahead window. - * Only return the #pages that are in the current window, so that - * we get called back on the first page of the ahead window which - * will allow us to submit more IO. + * Query the page cache and look for the traces(cached history pages) + * that a sequential stream would leave behind. */ - if (ra->prev_index >= ra->ahead_start) { - ra->start = ra->ahead_start; - ra->size = ra->ahead_size; - make_ahead_window(mapping, filp, ra, 0); -recheck: - /* prev_index shouldn't overrun the ahead window */ - ra->prev_index = min(ra->prev_index, - ra->ahead_start + ra->ahead_size - 1); + if (try_context_readahead(mapping, ra, offset, req_size, max)) + goto readit; + + /* + * standalone, small random read + * Read as is, and do not pollute the readahead state. + */ + return __do_page_cache_readahead(mapping, filp, offset, req_size, 0); + +initial_readahead: + ra->start = offset; + ra->size = get_init_ra_size(req_size, max); + ra->async_size = ra->size > req_size ? ra->size - req_size : ra->size; + +readit: + /* + * Will this read hit the readahead marker made by itself? + * If so, trigger the readahead marker hit now, and merge + * the resulted next readahead window into the current one. + */ + if (offset == ra->start && ra->size == ra->async_size) { + ra->async_size = get_next_ra_size(ra, max); + ra->size += ra->async_size; } -out: - return ra->prev_index + 1; + return ra_submit(ra, mapping, filp); } -EXPORT_SYMBOL_GPL(page_cache_readahead); -/* - * handle_ra_miss() is called when it is known that a page which should have - * been present in the pagecache (we just did some readahead there) was in fact - * not found. This will happen if it was evicted by the VM (readahead - * thrashing) +/** + * page_cache_sync_readahead - generic file readahead + * @mapping: address_space which holds the pagecache and I/O vectors + * @ra: file_ra_state which holds the readahead state + * @filp: passed on to ->readpage() and ->readpages() + * @offset: start offset into @mapping, in pagecache page-sized units + * @req_size: hint: total size of the read which the caller is performing in + * pagecache pages * - * Turn on the cache miss flag in the RA struct, this will cause the RA code - * to reduce the RA size on the next read. + * page_cache_sync_readahead() should be called when a cache miss happened: + * it will submit the read. The readahead logic may decide to piggyback more + * pages onto the read request if access patterns suggest it will improve + * performance. */ -void handle_ra_miss(struct address_space *mapping, - struct file_ra_state *ra, pgoff_t offset) +void page_cache_sync_readahead(struct address_space *mapping, + struct file_ra_state *ra, struct file *filp, + pgoff_t offset, unsigned long req_size) { - ra->flags |= RA_FLAG_MISS; - ra->flags &= ~RA_FLAG_INCACHE; - ra->cache_hit = 0; + /* no read-ahead */ + if (!ra->ra_pages) + return; + + /* do read-ahead */ + ondemand_readahead(mapping, ra, filp, false, offset, req_size); } +EXPORT_SYMBOL_GPL(page_cache_sync_readahead); -/* - * Given a desired number of PAGE_CACHE_SIZE readahead pages, return a - * sensible upper limit. +/** + * page_cache_async_readahead - file readahead for marked pages + * @mapping: address_space which holds the pagecache and I/O vectors + * @ra: file_ra_state which holds the readahead state + * @filp: passed on to ->readpage() and ->readpages() + * @page: the page at @offset which has the PG_readahead flag set + * @offset: start offset into @mapping, in pagecache page-sized units + * @req_size: hint: total size of the read which the caller is performing in + * pagecache pages + * + * page_cache_async_ondemand() should be called when a page is used which + * has the PG_readahead flag; this is a marker to suggest that the application + * has used up enough of the readahead window that we should start pulling in + * more pages. */ -unsigned long max_sane_readahead(unsigned long nr) +void +page_cache_async_readahead(struct address_space *mapping, + struct file_ra_state *ra, struct file *filp, + struct page *page, pgoff_t offset, + unsigned long req_size) { - return min(nr, (node_page_state(numa_node_id(), NR_INACTIVE) - + node_page_state(numa_node_id(), NR_FREE_PAGES)) / 2); + /* no read-ahead */ + if (!ra->ra_pages) + return; + + /* + * Same bit is used for PG_readahead and PG_reclaim. + */ + if (PageWriteback(page)) + return; + + ClearPageReadahead(page); + + /* + * Defer asynchronous read-ahead on IO congestion. + */ + if (bdi_read_congested(mapping->backing_dev_info)) + return; + + /* do read-ahead */ + ondemand_readahead(mapping, ra, filp, true, offset, req_size); + +#ifdef CONFIG_BLOCK + /* + * Normally the current page is !uptodate and lock_page() will be + * immediately called to implicitly unplug the device. However this + * is not always true for RAID conifgurations, where data arrives + * not strictly in their submission order. In this case we need to + * explicitly kick off the IO. + */ + if (PageUptodate(page)) + blk_run_backing_dev(mapping->backing_dev_info, NULL); +#endif } +EXPORT_SYMBOL_GPL(page_cache_async_readahead);