X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=fs%2Fbuffer.c;h=3c40d6382925b12a01f88c60a4ecafda4d2146dc;hb=6a77f38946aaee1cd85eeec6cf4229b204c15071;hp=0b84c0f8470b7a72a604e1cec36b84e2c08cb680;hpb=9213980e6a70d8473e0ffd4b39ab5b6caaba9ff5;p=linux-2.6.git diff --git a/fs/buffer.c b/fs/buffer.c index 0b84c0f84..3c40d6382 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -20,6 +20,7 @@ #include #include +#include #include #include #include @@ -37,40 +38,13 @@ #include #include #include -#include +#include +static int fsync_buffers_list(spinlock_t *lock, struct list_head *list); static void invalidate_bh_lrus(void); #define BH_ENTRY(list) list_entry((list), struct buffer_head, b_assoc_buffers) -struct bh_wait_queue { - struct buffer_head *bh; - wait_queue_t wait; -}; - -#define __DEFINE_BH_WAIT(name, b, f) \ - struct bh_wait_queue name = { \ - .bh = b, \ - .wait = { \ - .task = current, \ - .flags = f, \ - .func = bh_wake_function, \ - .task_list = \ - LIST_HEAD_INIT(name.wait.task_list),\ - }, \ - } -#define DEFINE_BH_WAIT(name, bh) __DEFINE_BH_WAIT(name, bh, 0) -#define DEFINE_BH_WAIT_EXCLUSIVE(name, bh) \ - __DEFINE_BH_WAIT(name, bh, WQ_FLAG_EXCLUSIVE) - -/* - * Hashed waitqueue_head's for wait_on_buffer() - */ -#define BH_WAIT_TABLE_ORDER 7 -static struct bh_wait_queue_head { - wait_queue_head_t wqh; -} ____cacheline_aligned_in_smp bh_wait_queue_heads[1<b_private = private; } -/* - * Return the address of the waitqueue_head to be used for this - * buffer_head - */ -wait_queue_head_t *bh_waitq_head(struct buffer_head *bh) -{ - return &bh_wait_queue_heads[hash_ptr(bh, BH_WAIT_TABLE_ORDER)].wqh; -} -EXPORT_SYMBOL(bh_waitq_head); - -void wake_up_buffer(struct buffer_head *bh) -{ - wait_queue_head_t *wq = bh_waitq_head(bh); - - smp_mb(); - if (waitqueue_active(wq)) - __wake_up(wq, TASK_INTERRUPTIBLE|TASK_UNINTERRUPTIBLE, 1, bh); -} -EXPORT_SYMBOL(wake_up_buffer); - -static int bh_wake_function(wait_queue_t *wait, unsigned mode, - int sync, void *key) -{ - struct buffer_head *bh = key; - struct bh_wait_queue *wq; - - wq = container_of(wait, struct bh_wait_queue, wait); - if (wq->bh != bh || buffer_locked(bh)) - return 0; - else - return autoremove_wake_function(wait, mode, sync, key); -} - -static void sync_buffer(struct buffer_head *bh) +static int sync_buffer(void *word) { struct block_device *bd; + struct buffer_head *bh + = container_of(word, struct buffer_head, b_state); smp_mb(); bd = bh->b_bdev; if (bd) blk_run_address_space(bd->bd_inode->i_mapping); + io_schedule(); + return 0; } void fastcall __lock_buffer(struct buffer_head *bh) { - wait_queue_head_t *wqh = bh_waitq_head(bh); - DEFINE_BH_WAIT_EXCLUSIVE(wait, bh); - - do { - prepare_to_wait_exclusive(wqh, &wait.wait, - TASK_UNINTERRUPTIBLE); - if (buffer_locked(bh)) { - sync_buffer(bh); - io_schedule(); - } - } while (test_set_buffer_locked(bh)); - finish_wait(wqh, &wait.wait); + wait_on_bit_lock(&bh->b_state, BH_Lock, sync_buffer, + TASK_UNINTERRUPTIBLE); } EXPORT_SYMBOL(__lock_buffer); @@ -142,7 +77,7 @@ void fastcall unlock_buffer(struct buffer_head *bh) { clear_buffer_locked(bh); smp_mb__after_clear_bit(); - wake_up_buffer(bh); + wake_up_bit(&bh->b_state, BH_Lock); } /* @@ -152,25 +87,7 @@ void fastcall unlock_buffer(struct buffer_head *bh) */ void __wait_on_buffer(struct buffer_head * bh) { - wait_queue_head_t *wqh = bh_waitq_head(bh); - DEFINE_BH_WAIT(wait, bh); - - do { - prepare_to_wait(wqh, &wait.wait, TASK_UNINTERRUPTIBLE); - if (buffer_locked(bh)) { - sync_buffer(bh); - io_schedule(); - } - } while (buffer_locked(bh)); - finish_wait(wqh, &wait.wait); -} - -static void -__set_page_buffers(struct page *page, struct buffer_head *head) -{ - page_cache_get(page); - SetPagePrivate(page); - page->private = (unsigned long)head; + wait_on_bit(&bh->b_state, BH_Lock, sync_buffer, TASK_UNINTERRUPTIBLE); } static void @@ -213,7 +130,7 @@ void end_buffer_write_sync(struct buffer_head *bh, int uptodate) if (uptodate) { set_buffer_uptodate(bh); } else { - if (printk_ratelimit()) { + if (!buffer_eopnotsupp(bh) && printk_ratelimit()) { buffer_io_error(bh); printk(KERN_WARNING "lost page write due to " "I/O error on %s\n", @@ -394,10 +311,10 @@ int file_fsync(struct file *filp, struct dentry *dentry, int datasync) { struct inode * inode = dentry->d_inode; struct super_block * sb; - int ret; + int ret, err; /* sync the inode to buffers */ - write_inode_now(inode, 0); + ret = write_inode_now(inode, 0); /* sync the superblock to buffers */ sb = inode->i_sb; @@ -407,7 +324,9 @@ int file_fsync(struct file *filp, struct dentry *dentry, int datasync) unlock_super(sb); /* .. finally sync the buffers to disk */ - ret = sync_blockdev(sb->s_bdev); + err = sync_blockdev(sb->s_bdev); + if (!ret) + ret = err; return ret; } @@ -430,18 +349,22 @@ asmlinkage long sys_fsync(unsigned int fd) goto out_putf; } - /* We need to protect against concurrent writers.. */ - down(&mapping->host->i_sem); current->flags |= PF_SYNCWRITE; ret = filemap_fdatawrite(mapping); + + /* + * We need to protect against concurrent writers, + * which could cause livelocks in fsync_buffers_list + */ + down(&mapping->host->i_sem); err = file->f_op->fsync(file, file->f_dentry, 0); if (!ret) ret = err; + up(&mapping->host->i_sem); err = filemap_fdatawait(mapping); if (!ret) ret = err; current->flags &= ~PF_SYNCWRITE; - up(&mapping->host->i_sem); out_putf: fput(file); @@ -466,17 +389,17 @@ asmlinkage long sys_fdatasync(unsigned int fd) mapping = file->f_mapping; - down(&mapping->host->i_sem); current->flags |= PF_SYNCWRITE; ret = filemap_fdatawrite(mapping); + down(&mapping->host->i_sem); err = file->f_op->fsync(file, file->f_dentry, 1); if (!ret) ret = err; + up(&mapping->host->i_sem); err = filemap_fdatawait(mapping); if (!ret) ret = err; current->flags &= ~PF_SYNCWRITE; - up(&mapping->host->i_sem); out_putf: fput(file); @@ -505,6 +428,7 @@ __find_get_block_slow(struct block_device *bdev, sector_t block, int unused) struct buffer_head *bh; struct buffer_head *head; struct page *page; + int all_mapped = 1; index = block >> (PAGE_CACHE_SHIFT - bd_inode->i_blkbits); page = find_get_page(bd_mapping, index); @@ -522,14 +446,23 @@ __find_get_block_slow(struct block_device *bdev, sector_t block, int unused) get_bh(bh); goto out_unlock; } + if (!buffer_mapped(bh)) + all_mapped = 0; bh = bh->b_this_page; } while (bh != head); - printk("__find_get_block_slow() failed. " - "block=%llu, b_blocknr=%llu\n", - (unsigned long long)block, (unsigned long long)bh->b_blocknr); - printk("b_state=0x%08lx, b_size=%u\n", bh->b_state, bh->b_size); - printk("device blocksize: %d\n", 1 << bd_inode->i_blkbits); + /* we might be here because some of the buffers on this page are + * not mapped. This is due to various races between + * file io on the block device and getblk. It gets dealt with + * elsewhere, don't buffer_error if we had some unmapped buffers + */ + if (all_mapped) { + printk("__find_get_block_slow() failed. " + "block=%llu, b_blocknr=%llu\n", + (unsigned long long)block, (unsigned long long)bh->b_blocknr); + printk("b_state=0x%08lx, b_size=%u\n", bh->b_state, bh->b_size); + printk("device blocksize: %d\n", 1 << bd_inode->i_blkbits); + } out_unlock: spin_unlock(&bd_mapping->private_lock); page_cache_release(page); @@ -604,7 +537,7 @@ static void free_more_memory(void) */ static void end_buffer_async_read(struct buffer_head *bh, int uptodate) { - static spinlock_t page_uptodate_lock = SPIN_LOCK_UNLOCKED; + static DEFINE_SPINLOCK(page_uptodate_lock); unsigned long flags; struct buffer_head *tmp; struct page *page; @@ -617,7 +550,8 @@ static void end_buffer_async_read(struct buffer_head *bh, int uptodate) set_buffer_uptodate(bh); } else { clear_buffer_uptodate(bh); - buffer_io_error(bh); + if (printk_ratelimit()) + buffer_io_error(bh); SetPageError(page); } @@ -662,7 +596,7 @@ still_busy: void end_buffer_async_write(struct buffer_head *bh, int uptodate) { char b[BDEVNAME_SIZE]; - static spinlock_t page_uptodate_lock = SPIN_LOCK_UNLOCKED; + static DEFINE_SPINLOCK(page_uptodate_lock); unsigned long flags; struct buffer_head *tmp; struct page *page; @@ -725,12 +659,11 @@ still_busy: * PageLocked prevents anyone from starting writeback of a page which is * under read I/O (PageWriteback is only ever set against a locked page). */ -void mark_buffer_async_read(struct buffer_head *bh) +static void mark_buffer_async_read(struct buffer_head *bh) { bh->b_end_io = end_buffer_async_read; set_buffer_async_read(bh); } -EXPORT_SYMBOL(mark_buffer_async_read); void mark_buffer_async_write(struct buffer_head *bh) { @@ -789,14 +722,6 @@ EXPORT_SYMBOL(mark_buffer_async_write); * b_inode back. */ -void buffer_insert_list(spinlock_t *lock, - struct buffer_head *bh, struct list_head *list) -{ - spin_lock(lock); - list_move_tail(&bh->b_assoc_buffers, list); - spin_unlock(lock); -} - /* * The buffer's backing address_space's private_lock must be held */ @@ -899,9 +824,12 @@ void mark_buffer_dirty_inode(struct buffer_head *bh, struct inode *inode) if (mapping->assoc_mapping != buffer_mapping) BUG(); } - if (list_empty(&bh->b_assoc_buffers)) - buffer_insert_list(&buffer_mapping->private_lock, - bh, &mapping->private_list); + if (list_empty(&bh->b_assoc_buffers)) { + spin_lock(&buffer_mapping->private_lock); + list_move_tail(&bh->b_assoc_buffers, + &mapping->private_list); + spin_unlock(&buffer_mapping->private_lock); + } } EXPORT_SYMBOL(mark_buffer_dirty_inode); @@ -982,7 +910,7 @@ EXPORT_SYMBOL(__set_page_dirty_buffers); * the osync code to catch these locked, dirty buffers without requeuing * any newly dirty buffers for write. */ -int fsync_buffers_list(spinlock_t *lock, struct list_head *list) +static int fsync_buffers_list(spinlock_t *lock, struct list_head *list) { struct buffer_head *bh; struct list_head tmp; @@ -1094,8 +1022,8 @@ int remove_inode_buffers(struct inode *inode) * The retry flag is used to differentiate async IO (paging, swapping) * which may not fail from ordinary buffer allocations. */ -static struct buffer_head * -create_buffers(struct page * page, unsigned long size, int retry) +struct buffer_head *alloc_page_buffers(struct page *page, unsigned long size, + int retry) { struct buffer_head *bh, *head; long offset; @@ -1153,6 +1081,7 @@ no_grow: free_more_memory(); goto try_again; } +EXPORT_SYMBOL_GPL(alloc_page_buffers); static inline void link_dev_buffers(struct page *page, struct buffer_head *head) @@ -1165,7 +1094,7 @@ link_dev_buffers(struct page *page, struct buffer_head *head) bh = bh->b_this_page; } while (bh); tail->b_this_page = head; - __set_page_buffers(page, head); + attach_page_buffers(page, head); } /* @@ -1177,18 +1106,16 @@ init_page_buffers(struct page *page, struct block_device *bdev, { struct buffer_head *head = page_buffers(page); struct buffer_head *bh = head; - unsigned int b_state; - - b_state = 1 << BH_Mapped; - if (PageUptodate(page)) - b_state |= 1 << BH_Uptodate; + int uptodate = PageUptodate(page); do { - if (!(bh->b_state & (1 << BH_Mapped))) { + if (!buffer_mapped(bh)) { init_buffer(bh, NULL, NULL); bh->b_bdev = bdev; bh->b_blocknr = block; - bh->b_state = b_state; + if (uptodate) + set_buffer_uptodate(bh); + set_buffer_mapped(bh); } block++; bh = bh->b_this_page; @@ -1217,8 +1144,10 @@ grow_dev_page(struct block_device *bdev, sector_t block, if (page_has_buffers(page)) { bh = page_buffers(page); - if (bh->b_size == size) + if (bh->b_size == size) { + init_page_buffers(page, bdev, block, size); return page; + } if (!try_to_free_buffers(page)) goto failed; } @@ -1226,7 +1155,7 @@ grow_dev_page(struct block_device *bdev, sector_t block, /* * Allocate some buffers for this page */ - bh = create_buffers(page, size, 0); + bh = alloc_page_buffers(page, size, 0); if (!bh) goto failed; @@ -1420,7 +1349,7 @@ struct bh_lru { struct buffer_head *bhs[BH_LRU_SIZE]; }; -static DEFINE_PER_CPU(struct bh_lru, bh_lrus) = {{0}}; +static DEFINE_PER_CPU(struct bh_lru, bh_lrus) = {{ NULL }}; #ifdef CONFIG_SMP #define bh_lru_lock() local_irq_disable() @@ -1551,6 +1480,7 @@ __getblk(struct block_device *bdev, sector_t block, int size) { struct buffer_head *bh = __find_get_block(bdev, block, size); + might_sleep(); if (bh == NULL) bh = __getblk_slow(bdev, block, size); return bh; @@ -1588,10 +1518,9 @@ __bread(struct block_device *bdev, sector_t block, int size) EXPORT_SYMBOL(__bread); /* - * invalidate_bh_lrus() is called rarely - at unmount. Because it is only for - * unmount it only needs to ensure that all buffers from the target device are - * invalidated on return and it doesn't need to worry about new buffers from - * that device being added - the unmount code has to prevent that. + * invalidate_bh_lrus() is called rarely - but not only at unmount. + * This doesn't race because it runs in each cpu either in irq + * or with preempt disabled. */ static void invalidate_bh_lru(void *arg) { @@ -1732,7 +1661,7 @@ void create_empty_buffers(struct page *page, { struct buffer_head *bh, *head, *tail; - head = create_buffers(page, blocksize, 1); + head = alloc_page_buffers(page, blocksize, 1); bh = head; do { bh->b_state |= b_state; @@ -1752,7 +1681,7 @@ void create_empty_buffers(struct page *page, bh = bh->b_this_page; } while (bh != head); } - __set_page_buffers(page, head); + attach_page_buffers(page, head); spin_unlock(&page->mapping->private_lock); } EXPORT_SYMBOL(create_empty_buffers); @@ -1777,6 +1706,8 @@ void unmap_underlying_metadata(struct block_device *bdev, sector_t block) { struct buffer_head *old_bh; + might_sleep(); + old_bh = __find_get_block_slow(bdev, block, 0); if (old_bh) { clear_buffer_dirty(old_bh); @@ -1807,10 +1738,10 @@ EXPORT_SYMBOL(unmap_underlying_metadata); * state inside lock_buffer(). * * If block_write_full_page() is called for regular writeback - * (called_for_sync() is false) then it will redirty a page which has a locked - * buffer. This only can happen if someone has written the buffer directly, - * with submit_bh(). At the address_space level PageWriteback prevents this - * contention from occurring. + * (wbc->sync_mode == WB_SYNC_NONE) then it will redirty a page which has a + * locked buffer. This only can happen if someone has written the buffer + * directly, with submit_bh(). At the address_space level PageWriteback + * prevents this contention from occurring. */ static int __block_write_full_page(struct inode *inode, struct page *page, get_block_t *get_block, struct writeback_control *wbc) @@ -1899,14 +1830,14 @@ static int __block_write_full_page(struct inode *inode, struct page *page, } } while ((bh = bh->b_this_page) != head); + /* + * The page and its buffers are protected by PageWriteback(), so we can + * drop the bh refcounts early. + */ BUG_ON(PageWriteback(page)); - set_page_writeback(page); /* Keeps try_to_free_buffers() away */ + set_page_writeback(page); unlock_page(page); - /* - * The page may come unlocked any time after the *first* submit_bh() - * call. Be careful with its buffers. - */ do { struct buffer_head *next = bh->b_this_page; if (buffer_async_write(bh)) { @@ -1936,6 +1867,10 @@ done: if (uptodate) SetPageUptodate(page); end_page_writeback(page); + /* + * The page and buffer_heads can be released at any time from + * here on. + */ wbc->pages_skipped++; /* We didn't write this page */ } return err; @@ -2231,7 +2166,7 @@ int generic_cont_expand(struct inode *inode, loff_t size) int err; err = -EFBIG; - limit = current->rlim[RLIMIT_FSIZE].rlim_cur; + limit = current->signal->rlim[RLIMIT_FSIZE].rlim_cur; if (limit != RLIM_INFINITY && size > (loff_t)limit) { send_sig(SIGXFSZ, current, 0); goto out; @@ -2306,8 +2241,7 @@ int cont_prepare_write(struct page *page, unsigned offset, memset(kaddr+zerofrom, 0, PAGE_CACHE_SIZE-zerofrom); flush_dcache_page(new_page); kunmap_atomic(kaddr, KM_USER0); - __block_commit_write(inode, new_page, - zerofrom, PAGE_CACHE_SIZE); + generic_commit_write(NULL, new_page, zerofrom, PAGE_CACHE_SIZE); unlock_page(new_page); page_cache_release(new_page); } @@ -2487,7 +2421,7 @@ int nobh_prepare_write(struct page *page, unsigned from, unsigned to, } bh->b_state = map_bh.b_state; atomic_set(&bh->b_count, 0); - bh->b_this_page = 0; + bh->b_this_page = NULL; bh->b_page = page; bh->b_blocknr = map_bh.b_blocknr; bh->b_size = blocksize; @@ -2723,7 +2657,7 @@ int block_write_full_page(struct page *page, get_block_t *get_block, /* * The page straddles i_size. It must be zeroed out on each and every - * writepage invocation because it may be mmapped. "A file is mapped + * writepage invokation because it may be mmapped. "A file is mapped * in multiples of the page size. For a file that is not a multiple of * the page size, the remaining memory is zeroed when mapped, and * writes to that region are not written out to the file." @@ -2753,21 +2687,33 @@ static int end_bio_bh_io_sync(struct bio *bio, unsigned int bytes_done, int err) if (bio->bi_size) return 1; + if (err == -EOPNOTSUPP) { + set_bit(BIO_EOPNOTSUPP, &bio->bi_flags); + set_bit(BH_Eopnotsupp, &bh->b_state); + } + bh->b_end_io(bh, test_bit(BIO_UPTODATE, &bio->bi_flags)); bio_put(bio); return 0; } -void submit_bh(int rw, struct buffer_head * bh) +int submit_bh(int rw, struct buffer_head * bh) { struct bio *bio; + int ret = 0; BUG_ON(!buffer_locked(bh)); BUG_ON(!buffer_mapped(bh)); BUG_ON(!bh->b_end_io); - /* Only clear out a write error when rewriting */ - if (test_set_buffer_req(bh) && rw == WRITE) + if (buffer_ordered(bh) && (rw == WRITE)) + rw = WRITE_BARRIER; + + /* + * Only clear out a write error when rewriting, should this + * include WRITE_SYNC as well? + */ + if (test_set_buffer_req(bh) && (rw == WRITE || rw == WRITE_BARRIER)) clear_buffer_write_io_error(bh); /* @@ -2789,7 +2735,14 @@ void submit_bh(int rw, struct buffer_head * bh) bio->bi_end_io = end_bio_bh_io_sync; bio->bi_private = bh; + bio_get(bio); submit_bio(rw, bio); + + if (bio_flagged(bio, BIO_EOPNOTSUPP)) + ret = -EOPNOTSUPP; + + bio_put(bio); + return ret; } /** @@ -2848,20 +2801,30 @@ void ll_rw_block(int rw, int nr, struct buffer_head *bhs[]) /* * For a data-integrity writeout, we need to wait upon any in-progress I/O - * and then start new I/O and then wait upon it. + * and then start new I/O and then wait upon it. The caller must have a ref on + * the buffer_head. */ -void sync_dirty_buffer(struct buffer_head *bh) +int sync_dirty_buffer(struct buffer_head *bh) { + int ret = 0; + WARN_ON(atomic_read(&bh->b_count) < 1); lock_buffer(bh); if (test_clear_buffer_dirty(bh)) { get_bh(bh); bh->b_end_io = end_buffer_write_sync; - submit_bh(WRITE, bh); + ret = submit_bh(WRITE, bh); wait_on_buffer(bh); + if (buffer_eopnotsupp(bh)) { + clear_buffer_eopnotsupp(bh); + ret = -EOPNOTSUPP; + } + if (!ret && !buffer_uptodate(bh)) + ret = -EIO; } else { unlock_buffer(bh); } + return ret; } /* @@ -2895,7 +2858,6 @@ drop_buffers(struct page *page, struct buffer_head **buffers_to_free) { struct buffer_head *head = page_buffers(page); struct buffer_head *bh; - int was_uptodate = 1; bh = head; do { @@ -2903,8 +2865,6 @@ drop_buffers(struct page *page, struct buffer_head **buffers_to_free) set_bit(AS_EIO, &page->mapping->flags); if (buffer_busy(bh)) goto failed; - if (!buffer_uptodate(bh) && !buffer_req(bh)) - was_uptodate = 0; bh = bh->b_this_page; } while (bh != head); @@ -3095,14 +3055,11 @@ static int buffer_cpu_notify(struct notifier_block *self, void __init buffer_init(void) { - int i; int nrpages; bh_cachep = kmem_cache_create("buffer_head", sizeof(struct buffer_head), 0, SLAB_PANIC, init_buffer_head, NULL); - for (i = 0; i < ARRAY_SIZE(bh_wait_queue_heads); i++) - init_waitqueue_head(&bh_wait_queue_heads[i].wqh); /* * Limit the bh occupancy to 10% of ZONE_NORMAL @@ -3121,14 +3078,12 @@ EXPORT_SYMBOL(block_read_full_page); EXPORT_SYMBOL(block_sync_page); EXPORT_SYMBOL(block_truncate_page); EXPORT_SYMBOL(block_write_full_page); -EXPORT_SYMBOL(buffer_insert_list); EXPORT_SYMBOL(cont_prepare_write); EXPORT_SYMBOL(end_buffer_async_write); EXPORT_SYMBOL(end_buffer_read_sync); EXPORT_SYMBOL(end_buffer_write_sync); EXPORT_SYMBOL(file_fsync); EXPORT_SYMBOL(fsync_bdev); -EXPORT_SYMBOL(fsync_buffers_list); EXPORT_SYMBOL(generic_block_bmap); EXPORT_SYMBOL(generic_commit_write); EXPORT_SYMBOL(generic_cont_expand);