diff options
Diffstat (limited to 'fs/ext4/inode.c')
| -rw-r--r-- | fs/ext4/inode.c | 790 |
1 files changed, 444 insertions, 346 deletions
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 0188e65e1f5..8a064734e6e 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -38,6 +38,7 @@ #include <linux/slab.h> #include <linux/ratelimit.h> #include <linux/aio.h> +#include <linux/bitops.h> #include "ext4_jbd2.h" #include "xattr.h" @@ -144,8 +145,11 @@ static int ext4_meta_trans_blocks(struct inode *inode, int lblocks, */ static int ext4_inode_is_fast_symlink(struct inode *inode) { - int ea_blocks = EXT4_I(inode)->i_file_acl ? - (inode->i_sb->s_blocksize >> 9) : 0; + int ea_blocks = EXT4_I(inode)->i_file_acl ? + EXT4_CLUSTER_SIZE(inode->i_sb) >> 9 : 0; + + if (ext4_has_inline_data(inode)) + return 0; return (S_ISLNK(inode->i_mode) && inode->i_blocks - ea_blocks == 0); } @@ -214,7 +218,7 @@ void ext4_evict_inode(struct inode *inode) jbd2_complete_transaction(journal, commit_tid); filemap_write_and_wait(&inode->i_data); } - truncate_inode_pages(&inode->i_data, 0); + truncate_inode_pages_final(&inode->i_data); WARN_ON(atomic_read(&EXT4_I(inode)->i_ioend_count)); goto no_delete; @@ -225,7 +229,7 @@ void ext4_evict_inode(struct inode *inode) if (ext4_should_order_data(inode)) ext4_begin_ordered_truncate(inode, 0); - truncate_inode_pages(&inode->i_data, 0); + truncate_inode_pages_final(&inode->i_data); WARN_ON(atomic_read(&EXT4_I(inode)->i_ioend_count)); if (is_bad_inode(inode)) @@ -442,7 +446,7 @@ static void ext4_map_blocks_es_recheck(handle_t *handle, * could be converted. */ if (!(flags & EXT4_GET_BLOCKS_NO_LOCK)) - down_read((&EXT4_I(inode)->i_data_sem)); + down_read(&EXT4_I(inode)->i_data_sem); if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) { retval = ext4_ext_map_blocks(handle, inode, map, flags & EXT4_GET_BLOCKS_KEEP_SIZE); @@ -465,7 +469,7 @@ static void ext4_map_blocks_es_recheck(handle_t *handle, if (es_map->m_lblk != map->m_lblk || es_map->m_flags != map->m_flags || es_map->m_pblk != map->m_pblk) { - printk("ES cache assertation failed for inode: %lu " + printk("ES cache assertion failed for inode: %lu " "es_cached ex [%d/%d/%llu/%x] != " "found ex [%d/%d/%llu/%x] retval %d flags %x\n", inode->i_ino, es_map->m_lblk, es_map->m_len, @@ -488,8 +492,8 @@ static void ext4_map_blocks_es_recheck(handle_t *handle, * Otherwise, call with ext4_ind_map_blocks() to handle indirect mapping * based files * - * On success, it returns the number of blocks being mapped or allocate. - * if create==0 and the blocks are pre-allocated and uninitialized block, + * On success, it returns the number of blocks being mapped or allocated. + * if create==0 and the blocks are pre-allocated and unwritten block, * the result buffer head is unmapped. If the create ==1, it will make sure * the buffer head is mapped. * @@ -503,6 +507,7 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode, { struct extent_status es; int retval; + int ret = 0; #ifdef ES_AGGRESSIVE_TEST struct ext4_map_blocks orig_map; @@ -514,10 +519,19 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode, "logical block %lu\n", inode->i_ino, flags, map->m_len, (unsigned long) map->m_lblk); - ext4_es_lru_add(inode); + /* + * ext4_map_blocks returns an int, and m_len is an unsigned int + */ + if (unlikely(map->m_len > INT_MAX)) + map->m_len = INT_MAX; + + /* We can handle the block number less than EXT_MAX_BLOCKS */ + if (unlikely(map->m_lblk >= EXT_MAX_BLOCKS)) + return -EIO; /* Lookup extent status tree firstly */ if (ext4_es_lookup_extent(inode, map->m_lblk, &es)) { + ext4_es_lru_add(inode); if (ext4_es_is_written(&es) || ext4_es_is_unwritten(&es)) { map->m_pblk = ext4_es_pblock(&es) + map->m_lblk - es.es_lblk; @@ -544,7 +558,7 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode, * file system block. */ if (!(flags & EXT4_GET_BLOCKS_NO_LOCK)) - down_read((&EXT4_I(inode)->i_data_sem)); + down_read(&EXT4_I(inode)->i_data_sem); if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) { retval = ext4_ext_map_blocks(handle, inode, map, flags & EXT4_GET_BLOCKS_KEEP_SIZE); @@ -553,17 +567,15 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode, EXT4_GET_BLOCKS_KEEP_SIZE); } if (retval > 0) { - int ret; - unsigned long long status; + unsigned int status; -#ifdef ES_AGGRESSIVE_TEST - if (retval != map->m_len) { - printk("ES len assertation failed for inode: %lu " - "retval %d != map->m_len %d " - "in %s (lookup)\n", inode->i_ino, retval, - map->m_len, __func__); + if (unlikely(retval != map->m_len)) { + ext4_warning(inode->i_sb, + "ES len assertion failed for inode " + "%lu: retval %d != map->m_len %d", + inode->i_ino, retval, map->m_len); + WARN_ON(1); } -#endif status = map->m_flags & EXT4_MAP_UNWRITTEN ? EXTENT_STATUS_UNWRITTEN : EXTENT_STATUS_WRITTEN; @@ -581,7 +593,7 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode, found: if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED) { - int ret = check_block_validity(inode, map); + ret = check_block_validity(inode, map); if (ret != 0) return ret; } @@ -598,7 +610,13 @@ found: * with buffer head unmapped. */ if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED) - return retval; + /* + * If we need to convert extent to unwritten + * we continue and do the actual work in + * ext4_ext_map_blocks() + */ + if (!(flags & EXT4_GET_BLOCKS_CONVERT_UNWRITTEN)) + return retval; /* * Here we clear m_flags because after allocating an new extent, @@ -607,12 +625,12 @@ found: map->m_flags &= ~EXT4_MAP_FLAGS; /* - * New blocks allocate and/or writing to uninitialized extent + * New blocks allocate and/or writing to unwritten extent * will possibly result in updating i_data, so we take * the write lock of i_data_sem, and call get_blocks() * with create == 1 flag. */ - down_write((&EXT4_I(inode)->i_data_sem)); + down_write(&EXT4_I(inode)->i_data_sem); /* * if the caller is from delayed allocation writeout path @@ -654,17 +672,15 @@ found: ext4_clear_inode_state(inode, EXT4_STATE_DELALLOC_RESERVED); if (retval > 0) { - int ret; - unsigned long long status; + unsigned int status; -#ifdef ES_AGGRESSIVE_TEST - if (retval != map->m_len) { - printk("ES len assertation failed for inode: %lu " - "retval %d != map->m_len %d " - "in %s (allocation)\n", inode->i_ino, retval, - map->m_len, __func__); + if (unlikely(retval != map->m_len)) { + ext4_warning(inode->i_sb, + "ES len assertion failed for inode " + "%lu: retval %d != map->m_len %d", + inode->i_ino, retval, map->m_len); + WARN_ON(1); } -#endif /* * If the extent has been zeroed out, we don't need to update @@ -690,7 +706,7 @@ found: has_zeroout: up_write((&EXT4_I(inode)->i_data_sem)); if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED) { - int ret = check_block_validity(inode, map); + ret = check_block_validity(inode, map); if (ret != 0) return ret; } @@ -730,8 +746,12 @@ static int _ext4_get_block(struct inode *inode, sector_t iblock, ret = ext4_map_blocks(handle, inode, &map, flags); if (ret > 0) { + ext4_io_end_t *io_end = ext4_inode_aio(inode); + map_bh(bh, inode->i_sb, map.m_pblk); bh->b_state = (bh->b_state & ~EXT4_MAP_FLAGS) | map.m_flags; + if (io_end && io_end->flag & EXT4_IO_END_UNWRITTEN) + set_buffer_defer_completion(bh); bh->b_size = inode->i_sb->s_blocksize * map.m_len; ret = 0; } @@ -905,6 +925,7 @@ int do_journal_get_write_access(handle_t *handle, */ if (dirty) clear_buffer_dirty(bh); + BUFFER_TRACE(bh, "get write access"); ret = ext4_journal_get_write_access(handle, bh); if (!ret && dirty) ret = ext4_handle_dirty_metadata(handle, NULL, bh); @@ -972,7 +993,8 @@ retry_journal: ext4_journal_stop(handle); goto retry_grab; } - wait_on_page_writeback(page); + /* In case writeback began while the page was unlocked */ + wait_for_stable_page(page); if (ext4_should_dioread_nolock(inode)) ret = __block_write_begin(page, pos, len, ext4_get_block_write); @@ -1204,7 +1226,6 @@ static int ext4_journalled_write_end(struct file *file, */ static int ext4_da_reserve_metadata(struct inode *inode, ext4_lblk_t lblock) { - int retries = 0; struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); struct ext4_inode_info *ei = EXT4_I(inode); unsigned int md_needed; @@ -1216,7 +1237,6 @@ static int ext4_da_reserve_metadata(struct inode *inode, ext4_lblk_t lblock) * in order to allocate nrblocks * worse case is one extent per block */ -repeat: spin_lock(&ei->i_block_reservation_lock); /* * ext4_calc_metadata_amount() has side effects, which we have @@ -1236,10 +1256,6 @@ repeat: ei->i_da_metadata_calc_len = save_len; ei->i_da_metadata_calc_last_lblock = save_last_lblock; spin_unlock(&ei->i_block_reservation_lock); - if (ext4_should_retry_alloc(inode->i_sb, &retries)) { - cond_resched(); - goto repeat; - } return -ENOSPC; } ei->i_reserved_meta_blocks += md_needed; @@ -1253,7 +1269,6 @@ repeat: */ static int ext4_da_reserve_space(struct inode *inode, ext4_lblk_t lblock) { - int retries = 0; struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); struct ext4_inode_info *ei = EXT4_I(inode); unsigned int md_needed; @@ -1275,7 +1290,6 @@ static int ext4_da_reserve_space(struct inode *inode, ext4_lblk_t lblock) * in order to allocate nrblocks * worse case is one extent per block */ -repeat: spin_lock(&ei->i_block_reservation_lock); /* * ext4_calc_metadata_amount() has side effects, which we have @@ -1295,10 +1309,6 @@ repeat: ei->i_da_metadata_calc_len = save_len; ei->i_da_metadata_calc_last_lblock = save_last_lblock; spin_unlock(&ei->i_block_reservation_lock); - if (ext4_should_retry_alloc(inode->i_sb, &retries)) { - cond_resched(); - goto repeat; - } dquot_release_reservation_block(inode, EXT4_C2B(sbi, 1)); return -ENOSPC; } @@ -1529,14 +1539,12 @@ static int ext4_da_map_blocks(struct inode *inode, sector_t iblock, "logical block %lu\n", inode->i_ino, map->m_len, (unsigned long) map->m_lblk); - ext4_es_lru_add(inode); - /* Lookup extent status tree firstly */ if (ext4_es_lookup_extent(inode, iblock, &es)) { - + ext4_es_lru_add(inode); if (ext4_es_is_hole(&es)) { retval = 0; - down_read((&EXT4_I(inode)->i_data_sem)); + down_read(&EXT4_I(inode)->i_data_sem); goto add_delayed; } @@ -1573,7 +1581,7 @@ static int ext4_da_map_blocks(struct inode *inode, sector_t iblock, * Try to see if we can get the block without requesting a new * file system block. */ - down_read((&EXT4_I(inode)->i_data_sem)); + down_read(&EXT4_I(inode)->i_data_sem); if (ext4_has_inline_data(inode)) { /* * We will soon create blocks for this page, and let @@ -1638,16 +1646,15 @@ add_delayed: set_buffer_delay(bh); } else if (retval > 0) { int ret; - unsigned long long status; + unsigned int status; -#ifdef ES_AGGRESSIVE_TEST - if (retval != map->m_len) { - printk("ES len assertation failed for inode: %lu " - "retval %d != map->m_len %d " - "in %s (lookup)\n", inode->i_ino, retval, - map->m_len, __func__); + if (unlikely(retval != map->m_len)) { + ext4_warning(inode->i_sb, + "ES len assertion failed for inode " + "%lu: retval %d != map->m_len %d", + inode->i_ino, retval, map->m_len); + WARN_ON(1); } -#endif status = map->m_flags & EXT4_MAP_UNWRITTEN ? EXTENT_STATUS_UNWRITTEN : EXTENT_STATUS_WRITTEN; @@ -1766,6 +1773,7 @@ static int __ext4_journalled_writepage(struct page *page, BUG_ON(!ext4_handle_valid(handle)); if (inline_data) { + BUFFER_TRACE(inode_bh, "get write access"); ret = ext4_journal_get_write_access(handle, inode_bh); err = ext4_handle_dirty_metadata(handle, inode, inode_bh); @@ -1785,7 +1793,7 @@ static int __ext4_journalled_writepage(struct page *page, ret = err; if (!ext4_has_inline_data(inode)) - ext4_walk_page_buffers(handle, page_bufs, 0, len, + ext4_walk_page_buffers(NULL, page_bufs, 0, len, NULL, bput_one); ext4_set_inode_state(inode, EXT4_STATE_JDATA); out: @@ -1843,6 +1851,7 @@ static int ext4_writepage(struct page *page, struct buffer_head *page_bufs = NULL; struct inode *inode = page->mapping->host; struct ext4_io_submit io_submit; + bool keep_towrite = false; trace_ext4_writepage(page); size = i_size_read(inode); @@ -1873,6 +1882,7 @@ static int ext4_writepage(struct page *page, unlock_page(page); return 0; } + keep_towrite = true; } if (PageChecked(page) && ext4_should_journal_data(inode)) @@ -1889,19 +1899,39 @@ static int ext4_writepage(struct page *page, unlock_page(page); return -ENOMEM; } - ret = ext4_bio_write_page(&io_submit, page, len, wbc); + ret = ext4_bio_write_page(&io_submit, page, len, wbc, keep_towrite); ext4_io_submit(&io_submit); /* Drop io_end reference we got from init */ ext4_put_io_end_defer(io_submit.io_end); return ret; } +static int mpage_submit_page(struct mpage_da_data *mpd, struct page *page) +{ + int len; + loff_t size = i_size_read(mpd->inode); + int err; + + BUG_ON(page->index != mpd->first_page); + if (page->index == size >> PAGE_CACHE_SHIFT) + len = size & ~PAGE_CACHE_MASK; + else + len = PAGE_CACHE_SIZE; + clear_page_dirty_for_io(page); + err = ext4_bio_write_page(&mpd->io_submit, page, len, mpd->wbc, false); + if (!err) + mpd->wbc->nr_to_write--; + mpd->first_page++; + + return err; +} + #define BH_FLAGS ((1 << BH_Unwritten) | (1 << BH_Delay)) /* * mballoc gives us at most this number of blocks... * XXX: That seems to be only a limitation of ext4_mb_normalize_request(). - * The rest of mballoc seems to handle chunks upto full group size. + * The rest of mballoc seems to handle chunks up to full group size. */ #define MAX_WRITEPAGES_EXTENT_LEN 2048 @@ -1910,82 +1940,94 @@ static int ext4_writepage(struct page *page, * * @mpd - extent of blocks * @lblk - logical number of the block in the file - * @b_state - b_state of the buffer head added + * @bh - buffer head we want to add to the extent * - * the function is used to collect contig. blocks in same state + * The function is used to collect contig. blocks in the same state. If the + * buffer doesn't require mapping for writeback and we haven't started the + * extent of buffers to map yet, the function returns 'true' immediately - the + * caller can write the buffer right away. Otherwise the function returns true + * if the block has been added to the extent, false if the block couldn't be + * added. */ -static int mpage_add_bh_to_extent(struct mpage_da_data *mpd, ext4_lblk_t lblk, - unsigned long b_state) +static bool mpage_add_bh_to_extent(struct mpage_da_data *mpd, ext4_lblk_t lblk, + struct buffer_head *bh) { struct ext4_map_blocks *map = &mpd->map; - /* Don't go larger than mballoc is willing to allocate */ - if (map->m_len >= MAX_WRITEPAGES_EXTENT_LEN) - return 0; + /* Buffer that doesn't need mapping for writeback? */ + if (!buffer_dirty(bh) || !buffer_mapped(bh) || + (!buffer_delay(bh) && !buffer_unwritten(bh))) { + /* So far no extent to map => we write the buffer right away */ + if (map->m_len == 0) + return true; + return false; + } /* First block in the extent? */ if (map->m_len == 0) { map->m_lblk = lblk; map->m_len = 1; - map->m_flags = b_state & BH_FLAGS; - return 1; + map->m_flags = bh->b_state & BH_FLAGS; + return true; } + /* Don't go larger than mballoc is willing to allocate */ + if (map->m_len >= MAX_WRITEPAGES_EXTENT_LEN) + return false; + /* Can we merge the block to our big extent? */ if (lblk == map->m_lblk + map->m_len && - (b_state & BH_FLAGS) == map->m_flags) { + (bh->b_state & BH_FLAGS) == map->m_flags) { map->m_len++; - return 1; + return true; } - return 0; + return false; } -static bool add_page_bufs_to_extent(struct mpage_da_data *mpd, - struct buffer_head *head, - struct buffer_head *bh, - ext4_lblk_t lblk) +/* + * mpage_process_page_bufs - submit page buffers for IO or add them to extent + * + * @mpd - extent of blocks for mapping + * @head - the first buffer in the page + * @bh - buffer we should start processing from + * @lblk - logical number of the block in the file corresponding to @bh + * + * Walk through page buffers from @bh upto @head (exclusive) and either submit + * the page for IO if all buffers in this page were mapped and there's no + * accumulated extent of buffers to map or add buffers in the page to the + * extent of buffers to map. The function returns 1 if the caller can continue + * by processing the next page, 0 if it should stop adding buffers to the + * extent to map because we cannot extend it anymore. It can also return value + * < 0 in case of error during IO submission. + */ +static int mpage_process_page_bufs(struct mpage_da_data *mpd, + struct buffer_head *head, + struct buffer_head *bh, + ext4_lblk_t lblk) { struct inode *inode = mpd->inode; + int err; ext4_lblk_t blocks = (i_size_read(inode) + (1 << inode->i_blkbits) - 1) >> inode->i_blkbits; do { BUG_ON(buffer_locked(bh)); - if (!buffer_dirty(bh) || !buffer_mapped(bh) || - (!buffer_delay(bh) && !buffer_unwritten(bh)) || - lblk >= blocks) { + if (lblk >= blocks || !mpage_add_bh_to_extent(mpd, lblk, bh)) { /* Found extent to map? */ if (mpd->map.m_len) - return false; - if (lblk >= blocks) - return true; - continue; + return 0; + /* Everything mapped so far and we hit EOF */ + break; } - if (!mpage_add_bh_to_extent(mpd, lblk, bh->b_state)) - return false; } while (lblk++, (bh = bh->b_this_page) != head); - return true; -} - -static int mpage_submit_page(struct mpage_da_data *mpd, struct page *page) -{ - int len; - loff_t size = i_size_read(mpd->inode); - int err; - - BUG_ON(page->index != mpd->first_page); - if (page->index == size >> PAGE_CACHE_SHIFT) - len = size & ~PAGE_CACHE_MASK; - else - len = PAGE_CACHE_SIZE; - clear_page_dirty_for_io(page); - err = ext4_bio_write_page(&mpd->io_submit, page, len, mpd->wbc); - if (!err) - mpd->wbc->nr_to_write--; - mpd->first_page++; - - return err; + /* So far everything mapped? Submit the page for IO. */ + if (mpd->map.m_len == 0) { + err = mpage_submit_page(mpd, head->b_page); + if (err < 0) + return err; + } + return lblk < blocks; } /* @@ -1997,7 +2039,7 @@ static int mpage_submit_page(struct mpage_da_data *mpd, struct page *page) * Scan buffers corresponding to changed extent (we expect corresponding pages * to be already locked) and update buffer state according to new extent state. * We map delalloc buffers to their physical location, clear unwritten bits, - * and mark buffers as uninit when we perform writes to uninitialized extents + * and mark buffers as uninit when we perform writes to unwritten extents * and do extent conversion after IO is finished. If the last page is not fully * mapped, we update @map to the next extent in the last page that needs * mapping. Otherwise we submit the page for IO. @@ -2009,8 +2051,6 @@ static int mpage_map_and_submit_buffers(struct mpage_da_data *mpd) struct inode *inode = mpd->inode; struct buffer_head *head, *bh; int bpp_bits = PAGE_CACHE_SHIFT - inode->i_blkbits; - ext4_lblk_t blocks = (i_size_read(inode) + (1 << inode->i_blkbits) - 1) - >> inode->i_blkbits; pgoff_t start, end; ext4_lblk_t lblk; sector_t pblock; @@ -2032,7 +2072,7 @@ static int mpage_map_and_submit_buffers(struct mpage_da_data *mpd) if (page->index > end) break; - /* Upto 'end' pages must be contiguous */ + /* Up to 'end' pages must be contiguous */ BUG_ON(page->index != start); bh = head = page_buffers(page); do { @@ -2045,18 +2085,26 @@ static int mpage_map_and_submit_buffers(struct mpage_da_data *mpd) */ mpd->map.m_len = 0; mpd->map.m_flags = 0; - add_page_bufs_to_extent(mpd, head, bh, - lblk); + /* + * FIXME: If dioread_nolock supports + * blocksize < pagesize, we need to make + * sure we add size mapped so far to + * io_end->size as the following call + * can submit the page for IO. + */ + err = mpage_process_page_bufs(mpd, head, + bh, lblk); pagevec_release(&pvec); - return 0; + if (err > 0) + err = 0; + return err; } if (buffer_delay(bh)) { clear_buffer_delay(bh); bh->b_blocknr = pblock++; } clear_buffer_unwritten(bh); - } while (++lblk < blocks && - (bh = bh->b_this_page) != head); + } while (lblk++, (bh = bh->b_this_page) != head); /* * FIXME: This is going to break if dioread_nolock @@ -2085,12 +2133,12 @@ static int mpage_map_one_extent(handle_t *handle, struct mpage_da_data *mpd) struct inode *inode = mpd->inode; struct ext4_map_blocks *map = &mpd->map; int get_blocks_flags; - int err; + int err, dioread_nolock; trace_ext4_da_write_pages_extent(inode, map); /* * Call ext4_map_blocks() to allocate any delayed allocation blocks, or - * to convert an uninitialized extent to be initialized (in the case + * to convert an unwritten extent to be initialized (in the case * where we have written into one or more preallocated blocks). It is * possible that we're going to need more metadata blocks than * previously reserved. However we must not fail because we're in @@ -2107,7 +2155,8 @@ static int mpage_map_one_extent(handle_t *handle, struct mpage_da_data *mpd) */ get_blocks_flags = EXT4_GET_BLOCKS_CREATE | EXT4_GET_BLOCKS_METADATA_NOFAIL; - if (ext4_should_dioread_nolock(inode)) + dioread_nolock = ext4_should_dioread_nolock(inode); + if (dioread_nolock) get_blocks_flags |= EXT4_GET_BLOCKS_IO_CREATE_EXT; if (map->m_flags & (1 << BH_Delay)) get_blocks_flags |= EXT4_GET_BLOCKS_DELALLOC_RESERVE; @@ -2115,7 +2164,7 @@ static int mpage_map_one_extent(handle_t *handle, struct mpage_da_data *mpd) err = ext4_map_blocks(handle, inode, map, get_blocks_flags); if (err < 0) return err; - if (map->m_flags & EXT4_MAP_UNINIT) { + if (dioread_nolock && (map->m_flags & EXT4_MAP_UNWRITTEN)) { if (!mpd->io_submit.io_end->handle && ext4_handle_valid(handle)) { mpd->io_submit.io_end->handle = handle->h_rsv_handle; @@ -2141,6 +2190,9 @@ static int mpage_map_one_extent(handle_t *handle, struct mpage_da_data *mpd) * * @handle - handle for journal operations * @mpd - extent to map + * @give_up_on_write - we set this to true iff there is a fatal error and there + * is no hope of writing the data. The caller should discard + * dirty pages to avoid infinite loops. * * The function maps extent starting at mpd->lblk of length mpd->len. If it is * delayed, blocks are allocated, if it is unwritten, we may need to convert @@ -2163,7 +2215,7 @@ static int mpage_map_and_submit_extent(handle_t *handle, mpd->io_submit.io_end->offset = ((loff_t)map->m_lblk) << inode->i_blkbits; - while (map->m_len) { + do { err = mpage_map_one_extent(handle, mpd); if (err < 0) { struct super_block *sb = inode->i_sb; @@ -2201,17 +2253,25 @@ static int mpage_map_and_submit_extent(handle_t *handle, err = mpage_map_and_submit_buffers(mpd); if (err < 0) return err; - } + } while (map->m_len); - /* Update on-disk size after IO is submitted */ + /* + * Update on-disk size after IO is submitted. Races with + * truncate are avoided by checking i_size under i_data_sem. + */ disksize = ((loff_t)mpd->first_page) << PAGE_CACHE_SHIFT; - if (disksize > i_size_read(inode)) - disksize = i_size_read(inode); if (disksize > EXT4_I(inode)->i_disksize) { int err2; - - ext4_update_i_disksize(inode, disksize); + loff_t i_size; + + down_write(&EXT4_I(inode)->i_data_sem); + i_size = i_size_read(inode); + if (disksize > i_size) + disksize = i_size; + if (disksize > EXT4_I(inode)->i_disksize) + EXT4_I(inode)->i_disksize = disksize; err2 = ext4_mark_inode_dirty(handle, inode); + up_write(&EXT4_I(inode)->i_data_sem); if (err2) ext4_error(inode->i_sb, "Failed to mark inode %lu dirty", @@ -2225,7 +2285,7 @@ static int mpage_map_and_submit_extent(handle_t *handle, /* * Calculate the total number of credits to reserve for one writepages * iteration. This is called from ext4_writepages(). We map an extent of - * upto MAX_WRITEPAGES_EXTENT_LEN blocks and then we go on and finish mapping + * up to MAX_WRITEPAGES_EXTENT_LEN blocks and then we go on and finish mapping * the last partial page. So in total we can map MAX_WRITEPAGES_EXTENT_LEN + * bpp - 1 blocks in bpp different extents. */ @@ -2260,6 +2320,7 @@ static int mpage_prepare_extent_to_map(struct mpage_da_data *mpd) struct address_space *mapping = mpd->inode->i_mapping; struct pagevec pvec; unsigned int nr_pages; + long left = mpd->wbc->nr_to_write; pgoff_t index = mpd->first_page; pgoff_t end = mpd->last_page; int tag; @@ -2295,6 +2356,17 @@ static int mpage_prepare_extent_to_map(struct mpage_da_data *mpd) if (page->index > end) goto out; + /* + * Accumulated enough dirty pages? This doesn't apply + * to WB_SYNC_ALL mode. For integrity sync we have to + * keep going because someone may be concurrently + * dirtying pages, and we might have synced a lot of + * newly appeared dirty pages, but have not synced all + * of the old dirty pages. + */ + if (mpd->wbc->sync_mode == WB_SYNC_NONE && left <= 0) + goto out; + /* If we can't merge this page, we are done. */ if (mpd->map.m_len > 0 && mpd->next_page != page->index) goto out; @@ -2325,27 +2397,11 @@ static int mpage_prepare_extent_to_map(struct mpage_da_data *mpd) lblk = ((ext4_lblk_t)page->index) << (PAGE_CACHE_SHIFT - blkbits); head = page_buffers(page); - if (!add_page_bufs_to_extent(mpd, head, head, lblk)) - goto out; - /* So far everything mapped? Submit the page for IO. */ - if (mpd->map.m_len == 0) { - err = mpage_submit_page(mpd, page); - if (err < 0) - goto out; - } - - /* - * Accumulated enough dirty pages? This doesn't apply - * to WB_SYNC_ALL mode. For integrity sync we have to - * keep going because someone may be concurrently - * dirtying pages, and we might have synced a lot of - * newly appeared dirty pages, but have not synced all - * of the old dirty pages. - */ - if (mpd->wbc->sync_mode == WB_SYNC_NONE && - mpd->next_page - mpd->first_page >= - mpd->wbc->nr_to_write) + err = mpage_process_page_bufs(mpd, head, head, lblk); + if (err <= 0) goto out; + err = 0; + left--; } pagevec_release(&pvec); cond_resched(); @@ -2389,16 +2445,15 @@ static int ext4_writepages(struct address_space *mapping, * because that could violate lock ordering on umount */ if (!mapping->nrpages || !mapping_tagged(mapping, PAGECACHE_TAG_DIRTY)) - return 0; + goto out_writepages; if (ext4_should_journal_data(inode)) { struct blk_plug plug; - int ret; blk_start_plug(&plug); ret = write_cache_pages(mapping, wbc, __writepage, mapping); blk_finish_plug(&plug); - return ret; + goto out_writepages; } /* @@ -2411,12 +2466,14 @@ static int ext4_writepages(struct address_space *mapping, * *never* be called, so if that ever happens, we would want * the stack trace. */ - if (unlikely(sbi->s_mount_flags & EXT4_MF_FS_ABORTED)) - return -EROFS; + if (unlikely(sbi->s_mount_flags & EXT4_MF_FS_ABORTED)) { + ret = -EROFS; + goto out_writepages; + } if (ext4_should_dioread_nolock(inode)) { /* - * We may need to convert upto one extent per block in + * We may need to convert up to one extent per block in * the page and we may dirty the inode. */ rsv_blocks = 1 + (PAGE_CACHE_SIZE >> inode->i_blkbits); @@ -2532,7 +2589,7 @@ retry: break; } blk_finish_plug(&plug); - if (!ret && !cycled) { + if (!ret && !cycled && wbc->nr_to_write > 0) { cycled = 1; mpd.last_page = writeback_index - 1; mpd.first_page = 0; @@ -2652,7 +2709,7 @@ retry_journal: goto retry_grab; } /* In case writeback began while the page was unlocked */ - wait_on_page_writeback(page); + wait_for_stable_page(page); ret = __block_write_begin(page, pos, len, ext4_da_get_block_prep); if (ret < 0) { @@ -2997,19 +3054,13 @@ static int ext4_get_block_write_nolock(struct inode *inode, sector_t iblock, } static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, - ssize_t size, void *private, int ret, - bool is_async) + ssize_t size, void *private) { - struct inode *inode = file_inode(iocb->ki_filp); ext4_io_end_t *io_end = iocb->private; /* if not async direct IO just return */ - if (!io_end) { - inode_dio_done(inode); - if (is_async) - aio_complete(iocb, ret, 0); + if (!io_end) return; - } ext_debug("ext4_end_io_dio(): io_end 0x%p " "for inode %lu, iocb 0x%p, offset %llu, size %zd\n", @@ -3019,11 +3070,7 @@ static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, iocb->private = NULL; io_end->offset = offset; io_end->size = size; - if (is_async) { - io_end->iocb = iocb; - io_end->result = ret; - } - ext4_put_io_end_defer(io_end); + ext4_put_io_end(io_end); } /* @@ -3031,9 +3078,9 @@ static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, * preallocated extents, and those write extend the file, no need to * fall back to buffered IO. * - * For holes, we fallocate those blocks, mark them as uninitialized + * For holes, we fallocate those blocks, mark them as unwritten * If those blocks were preallocated, we mark sure they are split, but - * still keep the range to write as uninitialized. + * still keep the range to write as unwritten. * * The unwritten extents will be converted to written when DIO is completed. * For async direct IO, since the IO may still pending when return, we @@ -3046,13 +3093,12 @@ static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset, * */ static ssize_t ext4_ext_direct_IO(int rw, struct kiocb *iocb, - const struct iovec *iov, loff_t offset, - unsigned long nr_segs) + struct iov_iter *iter, loff_t offset) { struct file *file = iocb->ki_filp; struct inode *inode = file->f_mapping->host; ssize_t ret; - size_t count = iov_length(iov, nr_segs); + size_t count = iov_iter_count(iter); int overwrite = 0; get_block_t *get_block_func = NULL; int dio_flags = 0; @@ -3061,7 +3107,7 @@ static ssize_t ext4_ext_direct_IO(int rw, struct kiocb *iocb, /* Use the old path for reads and writes beyond i_size. */ if (rw != WRITE || final_size > inode->i_size) - return ext4_ind_direct_IO(rw, iocb, iov, offset, nr_segs); + return ext4_ind_direct_IO(rw, iocb, iter, offset); BUG_ON(iocb->private == NULL); @@ -3085,12 +3131,12 @@ static ssize_t ext4_ext_direct_IO(int rw, struct kiocb *iocb, * We could direct write to holes and fallocate. * * Allocated blocks to fill the hole are marked as - * uninitialized to prevent parallel buffered read to expose + * unwritten to prevent parallel buffered read to expose * the stale data before DIO complete the data IO. * * As to previously fallocated extents, ext4 get_block will * just simply mark the buffer mapped but still keep the - * extents uninitialized. + * extents unwritten. * * For non AIO case, we will convert those unwritten extents * to written after return back from blockdev_direct_IO. @@ -3108,7 +3154,6 @@ static ssize_t ext4_ext_direct_IO(int rw, struct kiocb *iocb, ret = -ENOMEM; goto retake_lock; } - io_end->flag |= EXT4_IO_END_DIRECT; /* * Grab reference for DIO. Will be dropped in ext4_end_io_dio() */ @@ -3129,8 +3174,8 @@ static ssize_t ext4_ext_direct_IO(int rw, struct kiocb *iocb, dio_flags = DIO_LOCKING; } ret = __blockdev_direct_IO(rw, iocb, inode, - inode->i_sb->s_bdev, iov, - offset, nr_segs, + inode->i_sb->s_bdev, iter, + offset, get_block_func, ext4_end_io_dio, NULL, @@ -3153,13 +3198,6 @@ static ssize_t ext4_ext_direct_IO(int rw, struct kiocb *iocb, if (ret <= 0 && ret != -EIOCBQUEUED && iocb->private) { WARN_ON(iocb->private != io_end); WARN_ON(io_end->flag & EXT4_IO_END_UNWRITTEN); - WARN_ON(io_end->iocb); - /* - * Generic code already did inode_dio_done() so we - * have to clear EXT4_IO_END_DIRECT to not do it for - * the second time. - */ - io_end->flag = 0; ext4_put_io_end(io_end); iocb->private = NULL; } @@ -3191,11 +3229,11 @@ retake_lock: } static ssize_t ext4_direct_IO(int rw, struct kiocb *iocb, - const struct iovec *iov, loff_t offset, - unsigned long nr_segs) + struct iov_iter *iter, loff_t offset) { struct file *file = iocb->ki_filp; struct inode *inode = file->f_mapping->host; + size_t count = iov_iter_count(iter); ssize_t ret; /* @@ -3208,13 +3246,12 @@ static ssize_t ext4_direct_IO(int rw, struct kiocb *iocb, if (ext4_has_inline_data(inode)) return 0; - trace_ext4_direct_IO_enter(inode, offset, iov_length(iov, nr_segs), rw); + trace_ext4_direct_IO_enter(inode, offset, count, rw); if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) - ret = ext4_ext_direct_IO(rw, iocb, iov, offset, nr_segs); + ret = ext4_ext_direct_IO(rw, iocb, iter, offset); else - ret = ext4_ind_direct_IO(rw, iocb, iov, offset, nr_segs); - trace_ext4_direct_IO_exit(inode, offset, - iov_length(iov, nr_segs), rw, ret); + ret = ext4_ind_direct_IO(rw, iocb, iter, offset); + trace_ext4_direct_IO_exit(inode, offset, count, rw, ret); return ret; } @@ -3307,33 +3344,13 @@ void ext4_set_aops(struct inode *inode) } /* - * ext4_block_truncate_page() zeroes out a mapping from file offset `from' - * up to the end of the block which corresponds to `from'. - * This required during truncate. We need to physically zero the tail end - * of that block so it doesn't yield old data if the file is later grown. - */ -int ext4_block_truncate_page(handle_t *handle, - struct address_space *mapping, loff_t from) -{ - unsigned offset = from & (PAGE_CACHE_SIZE-1); - unsigned length; - unsigned blocksize; - struct inode *inode = mapping->host; - - blocksize = inode->i_sb->s_blocksize; - length = blocksize - (offset & (blocksize - 1)); - - return ext4_block_zero_page_range(handle, mapping, from, length); -} - -/* * ext4_block_zero_page_range() zeros out a mapping of length 'length' * starting from file offset 'from'. The range to be zero'd must * be contained with in one block. If the specified range exceeds * the end of the block it will be shortened to end of the block * that cooresponds to 'from' */ -int ext4_block_zero_page_range(handle_t *handle, +static int ext4_block_zero_page_range(handle_t *handle, struct address_space *mapping, loff_t from, loff_t length) { ext4_fsblk_t index = from >> PAGE_CACHE_SHIFT; @@ -3423,6 +3440,26 @@ unlock: return err; } +/* + * ext4_block_truncate_page() zeroes out a mapping from file offset `from' + * up to the end of the block which corresponds to `from'. + * This required during truncate. We need to physically zero the tail end + * of that block so it doesn't yield old data if the file is later grown. + */ +static int ext4_block_truncate_page(handle_t *handle, + struct address_space *mapping, loff_t from) +{ + unsigned offset = from & (PAGE_CACHE_SIZE-1); + unsigned length; + unsigned blocksize; + struct inode *inode = mapping->host; + + blocksize = inode->i_sb->s_blocksize; + length = blocksize - (offset & (blocksize - 1)); + + return ext4_block_zero_page_range(handle, mapping, from, length); +} + int ext4_zero_partial_blocks(handle_t *handle, struct inode *inode, loff_t lstart, loff_t length) { @@ -3496,12 +3533,7 @@ int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length) if (!S_ISREG(inode->i_mode)) return -EOPNOTSUPP; - if (EXT4_SB(sb)->s_cluster_ratio > 1) { - /* TODO: Add support for bigalloc file systems */ - return -EOPNOTSUPP; - } - - trace_ext4_punch_hole(inode, offset, length); + trace_ext4_punch_hole(inode, offset, length, 0); /* * Write out all dirty pages to avoid race conditions @@ -3515,15 +3547,6 @@ int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length) } mutex_lock(&inode->i_mutex); - /* It's not possible punch hole on append only file */ - if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) { - ret = -EPERM; - goto out_mutex; - } - if (IS_SWAPFILE(inode)) { - ret = -ETXTBSY; - goto out_mutex; - } /* No need to punch hole beyond i_size */ if (offset >= inode->i_size) @@ -3539,6 +3562,18 @@ int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length) offset; } + if (offset & (sb->s_blocksize - 1) || + (offset + length) & (sb->s_blocksize - 1)) { + /* + * Attach jinode to inode for jbd2 if we do any zeroing of + * partial block + */ + ret = ext4_inode_attach_jinode(inode); + if (ret < 0) + goto out_mutex; + + } + first_block_offset = round_up(offset, sb->s_blocksize); last_block_offset = round_down((offset + length), sb->s_blocksize) - 1; @@ -3592,10 +3627,15 @@ int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length) ret = ext4_free_hole_blocks(handle, inode, first_block, stop_block); - ext4_discard_preallocations(inode); up_write(&EXT4_I(inode)->i_data_sem); if (IS_SYNC(inode)) ext4_handle_sync(handle); + + /* Now release the pages again to reduce race window */ + if (last_block_offset > first_block_offset) + truncate_pagecache_range(inode, first_block_offset, + last_block_offset); + inode->i_mtime = inode->i_ctime = ext4_current_time(inode); ext4_mark_inode_dirty(handle, inode); out_stop: @@ -3607,6 +3647,31 @@ out_mutex: return ret; } +int ext4_inode_attach_jinode(struct inode *inode) +{ + struct ext4_inode_info *ei = EXT4_I(inode); + struct jbd2_inode *jinode; + + if (ei->jinode || !EXT4_SB(inode->i_sb)->s_journal) + return 0; + + jinode = jbd2_alloc_inode(GFP_KERNEL); + spin_lock(&inode->i_lock); + if (!ei->jinode) { + if (!jinode) { + spin_unlock(&inode->i_lock); + return -ENOMEM; + } + ei->jinode = jinode; + jbd2_journal_init_jbd_inode(ei->jinode, inode); + jinode = NULL; + } + spin_unlock(&inode->i_lock); + if (unlikely(jinode != NULL)) + jbd2_free_inode(jinode); + return 0; +} + /* * ext4_truncate() * @@ -3644,7 +3709,7 @@ void ext4_truncate(struct inode *inode) /* * There is a possibility that we're either freeing the inode - * or it completely new indode. In those cases we might not + * or it's a completely new inode. In those cases we might not * have i_mutex locked because it's not necessary. */ if (!(inode->i_state & (I_NEW|I_FREEING))) @@ -3667,6 +3732,12 @@ void ext4_truncate(struct inode *inode) return; } + /* If we zero-out tail of the page, we have to create jinode for jbd2 */ + if (inode->i_size & (inode->i_sb->s_blocksize - 1)) { + if (ext4_inode_attach_jinode(inode) < 0) + return; + } + if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) credits = ext4_writepage_trans_blocks(inode); else @@ -3878,18 +3949,20 @@ int ext4_get_inode_loc(struct inode *inode, struct ext4_iloc *iloc) void ext4_set_inode_flags(struct inode *inode) { unsigned int flags = EXT4_I(inode)->i_flags; + unsigned int new_fl = 0; - inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC); if (flags & EXT4_SYNC_FL) - inode->i_flags |= S_SYNC; + new_fl |= S_SYNC; if (flags & EXT4_APPEND_FL) - inode->i_flags |= S_APPEND; + new_fl |= S_APPEND; if (flags & EXT4_IMMUTABLE_FL) - inode->i_flags |= S_IMMUTABLE; + new_fl |= S_IMMUTABLE; if (flags & EXT4_NOATIME_FL) - inode->i_flags |= S_NOATIME; + new_fl |= S_NOATIME; if (flags & EXT4_DIRSYNC_FL) - inode->i_flags |= S_DIRSYNC; + new_fl |= S_DIRSYNC; + inode_set_flags(inode, new_fl, + S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC); } /* Propagate flags from i_flags to EXT4_I(inode)->i_flags */ @@ -4108,11 +4181,13 @@ struct inode *ext4_iget(struct super_block *sb, unsigned long ino) EXT4_INODE_GET_XTIME(i_atime, inode, raw_inode); EXT4_EINODE_GET_XTIME(i_crtime, ei, raw_inode); - inode->i_version = le32_to_cpu(raw_inode->i_disk_version); - if (EXT4_INODE_SIZE(inode->i_sb) > EXT4_GOOD_OLD_INODE_SIZE) { - if (EXT4_FITS_IN_INODE(raw_inode, ei, i_version_hi)) - inode->i_version |= - (__u64)(le32_to_cpu(raw_inode->i_version_hi)) << 32; + if (likely(!test_opt2(inode->i_sb, HURD_COMPAT))) { + inode->i_version = le32_to_cpu(raw_inode->i_disk_version); + if (EXT4_INODE_SIZE(inode->i_sb) > EXT4_GOOD_OLD_INODE_SIZE) { + if (EXT4_FITS_IN_INODE(raw_inode, ei, i_version_hi)) + inode->i_version |= + (__u64)(le32_to_cpu(raw_inode->i_version_hi)) << 32; + } } ret = 0; @@ -4235,12 +4310,15 @@ static int ext4_do_update_inode(handle_t *handle, struct ext4_inode *raw_inode = ext4_raw_inode(iloc); struct ext4_inode_info *ei = EXT4_I(inode); struct buffer_head *bh = iloc->bh; + struct super_block *sb = inode->i_sb; int err = 0, rc, block; - int need_datasync = 0; + int need_datasync = 0, set_large_file = 0; uid_t i_uid; gid_t i_gid; - /* For fields not not tracking in the in-memory inode, + spin_lock(&ei->i_raw_lock); + + /* For fields not tracked in the in-memory inode, * initialise them to zero for new inodes. */ if (ext4_test_inode_state(inode, EXT4_STATE_NEW)) memset(raw_inode, 0, EXT4_SB(inode->i_sb)->s_inode_size); @@ -4278,12 +4356,13 @@ static int ext4_do_update_inode(handle_t *handle, EXT4_INODE_SET_XTIME(i_atime, inode, raw_inode); EXT4_EINODE_SET_XTIME(i_crtime, ei, raw_inode); - if (ext4_inode_blocks_set(handle, raw_inode, ei)) + if (ext4_inode_blocks_set(handle, raw_inode, ei)) { + spin_unlock(&ei->i_raw_lock); goto out_brelse; + } raw_inode->i_dtime = cpu_to_le32(ei->i_dtime); raw_inode->i_flags = cpu_to_le32(ei->i_flags & 0xFFFFFFFF); - if (EXT4_SB(inode->i_sb)->s_es->s_creator_os != - cpu_to_le32(EXT4_OS_HURD)) + if (likely(!test_opt2(inode->i_sb, HURD_COMPAT))) raw_inode->i_file_acl_high = cpu_to_le16(ei->i_file_acl >> 32); raw_inode->i_file_acl_lo = cpu_to_le32(ei->i_file_acl); @@ -4292,24 +4371,11 @@ static int ext4_do_update_inode(handle_t *handle, need_datasync = 1; } if (ei->i_disksize > 0x7fffffffULL) { - struct super_block *sb = inode->i_sb; if (!EXT4_HAS_RO_COMPAT_FEATURE(sb, EXT4_FEATURE_RO_COMPAT_LARGE_FILE) || EXT4_SB(sb)->s_es->s_rev_level == - cpu_to_le32(EXT4_GOOD_OLD_REV)) { - /* If this is the first large file - * created, add a flag to the superblock. - */ - err = ext4_journal_get_write_access(handle, - EXT4_SB(sb)->s_sbh); - if (err) - goto out_brelse; - ext4_update_dynamic_rev(sb); - EXT4_SET_RO_COMPAT_FEATURE(sb, - EXT4_FEATURE_RO_COMPAT_LARGE_FILE); - ext4_handle_sync(handle); - err = ext4_handle_dirty_super(handle, sb); - } + cpu_to_le32(EXT4_GOOD_OLD_REV)) + set_large_file = 1; } raw_inode->i_generation = cpu_to_le32(inode->i_generation); if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode)) { @@ -4328,22 +4394,37 @@ static int ext4_do_update_inode(handle_t *handle, raw_inode->i_block[block] = ei->i_data[block]; } - raw_inode->i_disk_version = cpu_to_le32(inode->i_version); - if (ei->i_extra_isize) { - if (EXT4_FITS_IN_INODE(raw_inode, ei, i_version_hi)) - raw_inode->i_version_hi = - cpu_to_le32(inode->i_version >> 32); - raw_inode->i_extra_isize = cpu_to_le16(ei->i_extra_isize); + if (likely(!test_opt2(inode->i_sb, HURD_COMPAT))) { + raw_inode->i_disk_version = cpu_to_le32(inode->i_version); + if (ei->i_extra_isize) { + if (EXT4_FITS_IN_INODE(raw_inode, ei, i_version_hi)) + raw_inode->i_version_hi = + cpu_to_le32(inode->i_version >> 32); + raw_inode->i_extra_isize = + cpu_to_le16(ei->i_extra_isize); + } } ext4_inode_csum_set(inode, raw_inode, ei); + spin_unlock(&ei->i_raw_lock); + BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata"); rc = ext4_handle_dirty_metadata(handle, NULL, bh); if (!err) err = rc; ext4_clear_inode_state(inode, EXT4_STATE_NEW); - + if (set_large_file) { + BUFFER_TRACE(EXT4_SB(sb)->s_sbh, "get write access"); + err = ext4_journal_get_write_access(handle, EXT4_SB(sb)->s_sbh); + if (err) + goto out_brelse; + ext4_update_dynamic_rev(sb); + EXT4_SET_RO_COMPAT_FEATURE(sb, + EXT4_FEATURE_RO_COMPAT_LARGE_FILE); + ext4_handle_sync(handle); + err = ext4_handle_dirty_super(handle, sb); + } ext4_update_inode_fsync_trans(handle, inode, need_datasync); out_brelse: brelse(bh); @@ -4356,21 +4437,20 @@ out_brelse: * * We are called from a few places: * - * - Within generic_file_write() for O_SYNC files. + * - Within generic_file_aio_write() -> generic_write_sync() for O_SYNC files. * Here, there will be no transaction running. We wait for any running * transaction to commit. * - * - Within sys_sync(), kupdate and such. - * We wait on commit, if tol to. + * - Within flush work (sys_sync(), kupdate and such). + * We wait on commit, if told to. * - * - Within prune_icache() (PF_MEMALLOC == true) - * Here we simply return. We can't afford to block kswapd on the - * journal commit. + * - Within iput_final() -> write_inode_now() + * We wait on commit, if told to. * * In all cases it is actually safe for us to return without doing anything, * because the inode has been copied into a raw inode buffer in - * ext4_mark_inode_dirty(). This is a correctness thing for O_SYNC and for - * knfsd. + * ext4_mark_inode_dirty(). This is a correctness thing for WB_SYNC_ALL + * writeback. * * Note that we are absolutely dependent upon all inode dirtiers doing the * right thing: they *must* call mark_inode_dirty() after dirtying info in @@ -4382,15 +4462,15 @@ out_brelse: * stuff(); * inode->i_size = expr; * - * is in error because a kswapd-driven write_inode() could occur while - * `stuff()' is running, and the new i_size will be lost. Plus the inode - * will no longer be on the superblock's dirty inode list. + * is in error because write_inode() could occur while `stuff()' is running, + * and the new i_size will be lost. Plus the inode will no longer be on the + * superblock's dirty inode list. */ int ext4_write_inode(struct inode *inode, struct writeback_control *wbc) { int err; - if (current->flags & PF_MEMALLOC) + if (WARN_ON_ONCE(current->flags & PF_MEMALLOC)) return 0; if (EXT4_SB(inode->i_sb)->s_journal) { @@ -4400,7 +4480,12 @@ int ext4_write_inode(struct inode *inode, struct writeback_control *wbc) return -EIO; } - if (wbc->sync_mode != WB_SYNC_ALL) + /* + * No need to force transaction in WB_SYNC_NONE mode. Also + * ext4_sync_fs() will force the commit after everything is + * written. + */ + if (wbc->sync_mode != WB_SYNC_ALL || wbc->for_sync) return 0; err = ext4_force_commit(inode->i_sb); @@ -4410,7 +4495,11 @@ int ext4_write_inode(struct inode *inode, struct writeback_control *wbc) err = __ext4_get_inode_loc(inode, &iloc, 0); if (err) return err; - if (wbc->sync_mode == WB_SYNC_ALL) + /* + * sync(2) will flush the whole buffer cache. No need to do + * it here separately for each inode. + */ + if (wbc->sync_mode == WB_SYNC_ALL && !wbc->for_sync) sync_dirty_buffer(iloc.bh); if (buffer_req(iloc.bh) && !buffer_uptodate(iloc.bh)) { EXT4_ERROR_INODE_BLOCK(inode, iloc.bh->b_blocknr, @@ -4529,7 +4618,8 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr) ext4_journal_stop(handle); } - if (attr->ia_valid & ATTR_SIZE) { + if (attr->ia_valid & ATTR_SIZE && attr->ia_size != inode->i_size) { + handle_t *handle; if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) { struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); @@ -4537,73 +4627,73 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr) if (attr->ia_size > sbi->s_bitmap_maxbytes) return -EFBIG; } - } - - if (S_ISREG(inode->i_mode) && - attr->ia_valid & ATTR_SIZE && - (attr->ia_size < inode->i_size)) { - handle_t *handle; - handle = ext4_journal_start(inode, EXT4_HT_INODE, 3); - if (IS_ERR(handle)) { - error = PTR_ERR(handle); - goto err_out; - } - if (ext4_handle_valid(handle)) { - error = ext4_orphan_add(handle, inode); - orphan = 1; - } - EXT4_I(inode)->i_disksize = attr->ia_size; - rc = ext4_mark_inode_dirty(handle, inode); - if (!error) - error = rc; - ext4_journal_stop(handle); + if (IS_I_VERSION(inode) && attr->ia_size != inode->i_size) + inode_inc_iversion(inode); - if (ext4_should_order_data(inode)) { - error = ext4_begin_ordered_truncate(inode, + if (S_ISREG(inode->i_mode) && + (attr->ia_size < inode->i_size)) { + if (ext4_should_order_data(inode)) { + error = ext4_begin_ordered_truncate(inode, attr->ia_size); - if (error) { - /* Do as much error cleanup as possible */ - handle = ext4_journal_start(inode, - EXT4_HT_INODE, 3); - if (IS_ERR(handle)) { - ext4_orphan_del(NULL, inode); + if (error) goto err_out; - } - ext4_orphan_del(handle, inode); - orphan = 0; - ext4_journal_stop(handle); + } + handle = ext4_journal_start(inode, EXT4_HT_INODE, 3); + if (IS_ERR(handle)) { + error = PTR_ERR(handle); goto err_out; } - } - } - - if (attr->ia_valid & ATTR_SIZE) { - if (attr->ia_size != inode->i_size) { - loff_t oldsize = inode->i_size; - - i_size_write(inode, attr->ia_size); - /* - * Blocks are going to be removed from the inode. Wait - * for dio in flight. Temporarily disable - * dioread_nolock to prevent livelock. - */ - if (orphan) { - if (!ext4_should_journal_data(inode)) { - ext4_inode_block_unlocked_dio(inode); - inode_dio_wait(inode); - ext4_inode_resume_unlocked_dio(inode); - } else - ext4_wait_for_tail_page_commit(inode); + if (ext4_handle_valid(handle)) { + error = ext4_orphan_add(handle, inode); + orphan = 1; } + down_write(&EXT4_I(inode)->i_data_sem); + EXT4_I(inode)->i_disksize = attr->ia_size; + rc = ext4_mark_inode_dirty(handle, inode); + if (!error) + error = rc; /* - * Truncate pagecache after we've waited for commit - * in data=journal mode to make pages freeable. + * We have to update i_size under i_data_sem together + * with i_disksize to avoid races with writeback code + * running ext4_wb_update_i_disksize(). */ - truncate_pagecache(inode, oldsize, inode->i_size); + if (!error) + i_size_write(inode, attr->ia_size); + up_write(&EXT4_I(inode)->i_data_sem); + ext4_journal_stop(handle); + if (error) { + ext4_orphan_del(NULL, inode); + goto err_out; + } + } else + i_size_write(inode, attr->ia_size); + + /* + * Blocks are going to be removed from the inode. Wait + * for dio in flight. Temporarily disable + * dioread_nolock to prevent livelock. + */ + if (orphan) { + if (!ext4_should_journal_data(inode)) { + ext4_inode_block_unlocked_dio(inode); + inode_dio_wait(inode); + ext4_inode_resume_unlocked_dio(inode); + } else + ext4_wait_for_tail_page_commit(inode); } - ext4_truncate(inode); + /* + * Truncate pagecache after we've waited for commit + * in data=journal mode to make pages freeable. + */ + truncate_pagecache(inode, inode->i_size); } + /* + * We want to call ext4_truncate() even if attr->ia_size == + * inode->i_size for cases like truncation of fallocated space + */ + if (attr->ia_valid & ATTR_SIZE) + ext4_truncate(inode); if (!rc) { setattr_copy(inode, attr); @@ -4618,7 +4708,7 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr) ext4_orphan_del(NULL, inode); if (!rc && (ia_valid & ATTR_MODE)) - rc = ext4_acl_chmod(inode); + rc = posix_acl_chmod(inode, inode->i_mode); err_out: ext4_std_error(inode->i_sb, error); @@ -4637,6 +4727,15 @@ int ext4_getattr(struct vfsmount *mnt, struct dentry *dentry, generic_fillattr(inode, stat); /* + * If there is inline data in the inode, the inode will normally not + * have data blocks allocated (it may have an external xattr block). + * Report at least one sector for such files, so tools like tar, rsync, + * others doen't incorrectly think the file is completely sparse. + */ + if (unlikely(ext4_has_inline_data(inode))) + stat->blocks += (stat->size + 511) >> 9; + + /* * We can't update i_blocks if the block allocation is delayed * otherwise in the case of system crash before the real block * allocation is done, we will have i_blocks inconsistent with @@ -4647,9 +4746,8 @@ int ext4_getattr(struct vfsmount *mnt, struct dentry *dentry, * blocks for this file. */ delalloc_blocks = EXT4_C2B(EXT4_SB(inode->i_sb), - EXT4_I(inode)->i_reserved_data_blocks); - - stat->blocks += delalloc_blocks << (inode->i_sb->s_blocksize_bits-9); + EXT4_I(inode)->i_reserved_data_blocks); + stat->blocks += delalloc_blocks << (inode->i_sb->s_blocksize_bits - 9); return 0; } |