From: Eric Sandeen <sandeen@redhat.com> Date: Wed, 1 Sep 2010 16:07:53 -0400 Subject: [fs] ext4: quota updates for RHEL5.6 Message-id: <4C7E7A59.5070400@redhat.com> Patchwork-id: 27992 O-Subject: [PATCH RHEL5.6 4/4] ext4: quota updates Bugzilla: 457153 RH-Acked-by: Edward Shishkin <edward@redhat.com> For Bug 457153 - ext4 full support in RHEL5 This patch contains changes to ext4 to bring it up to speed for quota. It contains the following upstream commits: 60e58e0f30e723464c2a7d34b71b8675566c572d ext4: quota reservation for delayed allocation 9f0ccfd8e07d61b413e6536ffa02fbf60d2e20d8 ext4: release reserved quota when block reservation for delalloc retry 8aa6790f876e81f5a2211fe1711a5fe3fe2d7b20 ext4: ext4_get_reserved_space() must return bytes instead of blocks a9e7f4472075fb6937c545af3f6329e9946bbe66 ext4: Convert to generic reserved quota's space management. 39bc680a8160bb9d6743f7873b535d553ff61058 ext4: fix sleep inside spinlock issue with quota and dealloc (#14739) d21cd8f163ac44b15c465aab7306db931c606908 ext4: Fix potential quota deadlock 0637c6f4135f592f094207c7c21e7c0fc5557834 ext4: Patch up how we claim metadata blocks for quota purposes 9d0be50230b333005635967f7ecd4897dbfd181b ext4: Calculate metadata requirements more accurately 1db913823c0f8360fccbd24ca67eb073966a5ffd ext4: Handle -EDQUOT error on write 5f634d064c709ea02c3cdaa850a08323a4a4bf28 ext4: Fix quota accounting error with fallocate 72b8ab9dde211ea518ff27e631b2046ef90c29a2 ext4: don't use quota reservation for speculative metadata blocks 1296cc85c26e94eb865d03f82140f27d598de467 ext4: Drop EXT4_GET_BLOCKS_UPDATE_RESERVE_SPACE flag 62d2b5f2dcd3707b070efb16bbfdf6947c38c194 ext4: Always journal quota file modifications This last patch isn't yet in rhel6, it missed the deadline, but it avoids some scary-sounding quota warnings. I'm aiming for 6.0.z with it. It also adds FS_HAS_GETRESV to ext4's fs_flags so that it can utilize the new reserved space interfaces (and adds FS_HAS_TRYTOFREE to the ext4dev fs_flags; looks like this got missed last update) Thanks, -Eric #endif Signed-off-by: Jarod Wilson <jarod@redhat.com> diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c index 5aaccb8..322b20d 100644 --- a/fs/ext4/balloc.c +++ b/fs/ext4/balloc.c @@ -630,14 +630,15 @@ ext4_fsblk_t ext4_new_meta_blocks(handle_t *handle, struct inode *inode, ret = ext4_mb_new_blocks(handle, &ar, errp); if (count) *count = ar.len; - /* - * Account for the allocated meta blocks + * Account for the allocated meta blocks. We will never + * fail EDQUOT for metdata, but we do account for it. */ if (!(*errp) && EXT4_I(inode)->i_delalloc_reserved_flag) { spin_lock(&EXT4_I(inode)->i_block_reservation_lock); EXT4_I(inode)->i_allocated_meta_blocks += ar.len; spin_unlock(&EXT4_I(inode)->i_block_reservation_lock); + DQUOT_ALLOC_BLOCK_NOFAIL(inode, ar.len); } return ret; } diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 91e054d..fefea9c 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -20,6 +20,7 @@ #include <linux/blkdev.h> #include <linux/magic.h> #include <linux/jbd2.h> +#include <linux/quota.h> #include <linux/rwsem.h> #include <linux/rbtree.h> #include <linux/seqlock.h> @@ -353,14 +354,11 @@ struct ext4_new_group_data { so set the magic i_delalloc_reserve_flag after taking the inode allocation semaphore for */ #define EXT4_GET_BLOCKS_DELALLOC_RESERVE 0x0004 - /* Call ext4_da_update_reserve_space() after successfully - allocating the blocks */ -#define EXT4_GET_BLOCKS_UPDATE_RESERVE_SPACE 0x0008 /* caller is from the direct IO path, request to creation of an unitialized extents if not allocated, split the uninitialized extent if blocks has been preallocated already*/ -#define EXT4_GET_BLOCKS_DIO 0x0010 -#define EXT4_GET_BLOCKS_CONVERT 0x0020 +#define EXT4_GET_BLOCKS_DIO 0x0008 +#define EXT4_GET_BLOCKS_CONVERT 0x0010 #define EXT4_GET_BLOCKS_DIO_CREATE_EXT (EXT4_GET_BLOCKS_DIO|\ EXT4_GET_BLOCKS_CREATE_UNINIT_EXT) /* Convert extent to initialized after direct IO complete */ @@ -690,11 +688,17 @@ struct ext4_inode_info { unsigned int i_reserved_meta_blocks; unsigned int i_allocated_meta_blocks; unsigned short i_delalloc_reserved_flag; + sector_t i_da_metadata_calc_last_lblock; + int i_da_metadata_calc_len; /* on-disk additional length */ __u16 i_extra_isize; spinlock_t i_block_reservation_lock; +#ifdef CONFIG_QUOTA + /* quota space reservation, managed internally by quota code */ + qsize_t i_reserved_quota; +#endif /* completed async DIOs that might need unwritten extents handling */ struct list_head i_aio_dio_complete_list; @@ -1491,7 +1495,10 @@ extern int ext4_chunk_trans_blocks(struct inode *, int nrblocks); extern int ext4_block_truncate_page(handle_t *handle, struct address_space *mapping, loff_t from); extern int ext4_page_mkwrite(struct vm_area_struct *vma, struct page *page); +extern qsize_t *ext4_get_reserved_space(struct inode *inode); extern int flush_aio_dio_completed_IO(struct inode *inode); +extern void ext4_da_update_reserve_space(struct inode *inode, + int used, int quota_claim); /* ioctl.c */ extern long ext4_ioctl(struct file *, unsigned int, unsigned long); extern long ext4_compat_ioctl(struct file *, unsigned int, unsigned long); diff --git a/fs/ext4/ext4_extents.h b/fs/ext4/ext4_extents.h index 2ca6864..bdb6ce7 100644 --- a/fs/ext4/ext4_extents.h +++ b/fs/ext4/ext4_extents.h @@ -225,7 +225,8 @@ static inline void ext4_ext_mark_initialized(struct ext4_extent *ext) ext->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ext)); } -extern int ext4_ext_calc_metadata_amount(struct inode *inode, int blocks); +extern int ext4_ext_calc_metadata_amount(struct inode *inode, + sector_t lblocks); extern ext4_fsblk_t ext_pblock(struct ext4_extent *ex); extern ext4_fsblk_t idx_pblock(struct ext4_extent_idx *); extern void ext4_ext_store_pblock(struct ext4_extent *, ext4_fsblk_t); diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 387b368..7dbeeee 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -293,29 +293,44 @@ static inline int ext4_ext_space_root_idx(struct inode *inode, int check) * to allocate @blocks * Worse case is one block per extent */ -int ext4_ext_calc_metadata_amount(struct inode *inode, int blocks) +int ext4_ext_calc_metadata_amount(struct inode *inode, sector_t lblock) { - int lcap, icap, rcap, leafs, idxs, num; - int newextents = blocks; - - rcap = ext4_ext_space_root_idx(inode, 0); - lcap = ext4_ext_space_block(inode, 0); - icap = ext4_ext_space_block_idx(inode, 0); + struct ext4_inode_info *ei = EXT4_I(inode); + int idxs, num = 0; - /* number of new leaf blocks needed */ - num = leafs = (newextents + lcap - 1) / lcap; + idxs = ((inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header)) + / sizeof(struct ext4_extent_idx)); /* - * Worse case, we need separate index block(s) - * to link all new leaf blocks + * If the new delayed allocation block is contiguous with the + * previous da block, it can share index blocks with the + * previous block, so we only need to allocate a new index + * block every idxs leaf blocks. At ldxs**2 blocks, we need + * an additional index block, and at ldxs**3 blocks, yet + * another index blocks. */ - idxs = (leafs + icap - 1) / icap; - do { - num += idxs; - idxs = (idxs + icap - 1) / icap; - } while (idxs > rcap); + if (ei->i_da_metadata_calc_len && + ei->i_da_metadata_calc_last_lblock+1 == lblock) { + if ((ei->i_da_metadata_calc_len % idxs) == 0) + num++; + if ((ei->i_da_metadata_calc_len % (idxs*idxs)) == 0) + num++; + if ((ei->i_da_metadata_calc_len % (idxs*idxs*idxs)) == 0) { + num++; + ei->i_da_metadata_calc_len = 0; + } else + ei->i_da_metadata_calc_len++; + ei->i_da_metadata_calc_last_lblock++; + return num; + } - return num; + /* + * In the worst case we need a new set of index blocks at + * every level of the inode's extent tree. + */ + ei->i_da_metadata_calc_len = 1; + ei->i_da_metadata_calc_last_lblock = lblock; + return ext_depth(inode) + 1; } static int @@ -3232,7 +3247,19 @@ out: unmap_underlying_metadata_blocks(inode->i_sb->s_bdev, newblock + max_blocks, allocated - max_blocks); + allocated = max_blocks; } + + /* + * If we have done fallocate with the offset that is already + * delayed allocated, we would have block reservation + * and quota reservation done in the delayed write path. + * But fallocate would have already updated quota and block + * count for this offset. So cancel these reservation + */ + if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) + ext4_da_update_reserve_space(inode, allocated, 0); + map_out: set_buffer_mapped(bh_result); out1: @@ -3482,9 +3509,18 @@ int ext4_ext_get_blocks(handle_t *handle, struct inode *inode, /* previous routine could use block we allocated */ newblock = ext_pblock(&newex); allocated = ext4_ext_get_actual_len(&newex); + if (allocated > max_blocks) + allocated = max_blocks; set_buffer_new(bh_result); /* + * Update reserved blocks/metadata blocks after successful + * block allocation which had been deferred till now. + */ + if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) + ext4_da_update_reserve_space(inode, allocated, 1); + + /* * Cache the extent and update transaction to commit on fdatasync only * when it is _not_ an uninitialized extent. */ diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 618cbb7..5daef8e 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -1055,75 +1055,112 @@ out: return err; } +#ifdef CONFIG_QUOTA +qsize_t *ext4_get_reserved_space(struct inode *inode) +{ + return &EXT4_I(inode)->i_reserved_quota; +} +#endif + /* * Calculate the number of metadata blocks need to reserve - * to allocate @blocks for non extent file based file + * to allocate a new block at @lblocks for non extent file based file */ -static int ext4_indirect_calc_metadata_amount(struct inode *inode, int blocks) +static int ext4_indirect_calc_metadata_amount(struct inode *inode, + sector_t lblock) { - int icap = EXT4_ADDR_PER_BLOCK(inode->i_sb); - int ind_blks, dind_blks, tind_blks; - - /* number of new indirect blocks needed */ - ind_blks = (blocks + icap - 1) / icap; + struct ext4_inode_info *ei = EXT4_I(inode); + int dind_mask = EXT4_ADDR_PER_BLOCK(inode->i_sb) - 1; + int blk_bits; - dind_blks = (ind_blks + icap - 1) / icap; + if (lblock < EXT4_NDIR_BLOCKS) + return 0; - tind_blks = 1; + lblock -= EXT4_NDIR_BLOCKS; - return ind_blks + dind_blks + tind_blks; + if (ei->i_da_metadata_calc_len && + (lblock & dind_mask) == ei->i_da_metadata_calc_last_lblock) { + ei->i_da_metadata_calc_len++; + return 0; + } + ei->i_da_metadata_calc_last_lblock = lblock & dind_mask; + ei->i_da_metadata_calc_len = 1; + blk_bits = roundup_pow_of_two(lblock + 1); + return (blk_bits / EXT4_ADDR_PER_BLOCK_BITS(inode->i_sb)) + 1; } /* * Calculate the number of metadata blocks need to reserve - * to allocate given number of blocks + * to allocate a block located at @lblock */ -static int ext4_calc_metadata_amount(struct inode *inode, int blocks) +static int ext4_calc_metadata_amount(struct inode *inode, sector_t lblock) { - if (!blocks) - return 0; - if (EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL) - return ext4_ext_calc_metadata_amount(inode, blocks); + return ext4_ext_calc_metadata_amount(inode, lblock); - return ext4_indirect_calc_metadata_amount(inode, blocks); + return ext4_indirect_calc_metadata_amount(inode, lblock); } -static void ext4_da_update_reserve_space(struct inode *inode, int used) +/* + * Called with i_data_sem down, which is important since we can call + * ext4_discard_preallocations() from here. + */ +void ext4_da_update_reserve_space(struct inode *inode, + int used, int quota_claim) { struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); - int total, mdb, mdb_free; - - spin_lock(&EXT4_I(inode)->i_block_reservation_lock); - /* recalculate the number of metablocks still need to be reserved */ - total = EXT4_I(inode)->i_reserved_data_blocks - used; - mdb = ext4_calc_metadata_amount(inode, total); + struct ext4_inode_info *ei = EXT4_I(inode); - /* figure out how many metablocks to release */ - BUG_ON(mdb > EXT4_I(inode)->i_reserved_meta_blocks); - mdb_free = EXT4_I(inode)->i_reserved_meta_blocks - mdb; + spin_lock(&ei->i_block_reservation_lock); + if (unlikely(used > ei->i_reserved_data_blocks)) { + ext4_msg(inode->i_sb, KERN_NOTICE, "%s: ino %lu, used %d " + "with only %d reserved data blocks\n", + __func__, inode->i_ino, used, + ei->i_reserved_data_blocks); + WARN_ON(1); + used = ei->i_reserved_data_blocks; + } - if (mdb_free) { - /* Account for allocated meta_blocks */ - mdb_free -= EXT4_I(inode)->i_allocated_meta_blocks; + /* Update per-inode reservations */ + ei->i_reserved_data_blocks -= used; + ei->i_reserved_meta_blocks -= ei->i_allocated_meta_blocks; + percpu_counter_sub(&sbi->s_dirtyblocks_counter, + used + ei->i_allocated_meta_blocks); + ei->i_allocated_meta_blocks = 0; - /* update fs dirty blocks counter */ - percpu_counter_sub(&sbi->s_dirtyblocks_counter, mdb_free); - EXT4_I(inode)->i_allocated_meta_blocks = 0; - EXT4_I(inode)->i_reserved_meta_blocks = mdb; + if (ei->i_reserved_data_blocks == 0) { + /* + * We can release all of the reserved metadata blocks + * only when we have written all of the delayed + * allocation blocks. + */ + percpu_counter_sub(&sbi->s_dirtyblocks_counter, + ei->i_reserved_meta_blocks); + ei->i_reserved_meta_blocks = 0; + ei->i_da_metadata_calc_len = 0; } - /* update per-inode reservations */ - BUG_ON(used > EXT4_I(inode)->i_reserved_data_blocks); - EXT4_I(inode)->i_reserved_data_blocks -= used; - spin_unlock(&EXT4_I(inode)->i_block_reservation_lock); + + /* Update quota subsystem for data blocks */ + if (quota_claim) { + DQUOT_CLAIM_BLOCK(inode, used); + } else { + /* + * We did fallocate with an offset that is already delayed + * allocated. So on delayed allocated writeback we should + * not re-claim the quota for fallocated blocks. + */ + DQUOT_RELEASE_RESERVATION_BLOCK(inode, used); + } + /* * If we have done all the pending block allocations and if * there aren't any writers on the inode, we can discard the * inode's preallocations. */ - if (!total && (atomic_read(&inode->i_writecount) == 0)) + if ((ei->i_reserved_data_blocks == 0) && + (atomic_read(&inode->i_writecount) == 0)) ext4_discard_preallocations(inode); } @@ -1257,18 +1294,20 @@ int ext4_get_blocks(handle_t *handle, struct inode *inode, sector_t block, */ ext4_clear_inode_state(inode, EXT4_STATE_EXT_MIGRATE); } - } + /* + * Update reserved blocks/metadata blocks after successful + * block allocation which had been deferred till now. We don't + * support fallocate for non extent files. So we can update + * reserve space here. + */ + if ((retval > 0) && + (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE)) + ext4_da_update_reserve_space(inode, retval, 1); + } if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) EXT4_I(inode)->i_delalloc_reserved_flag = 0; - /* - * Update reserved blocks/metadata blocks after successful - * block allocation which had been deferred till now. - */ - if ((retval > 0) && (flags & EXT4_GET_BLOCKS_UPDATE_RESERVE_SPACE)) - ext4_da_update_reserve_space(inode, retval); - up_write((&EXT4_I(inode)->i_data_sem)); if (retval > 0 && buffer_mapped(bh)) { int ret = check_block_validity(inode, "file system " @@ -1826,11 +1865,15 @@ static int ext4_journalled_write_end(struct file *file, return ret ? ret : copied; } -static int ext4_da_reserve_space(struct inode *inode, int nrblocks) +/* + * Reserve a single block located at lblock + */ +static int ext4_da_reserve_space(struct inode *inode, sector_t lblock) { int retries = 0; - struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); - unsigned long md_needed, mdblocks, total = 0; + struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); + struct ext4_inode_info *ei = EXT4_I(inode); + unsigned long md_needed; /* * recalculate the amount of metadata blocks to reserve @@ -1838,73 +1881,82 @@ static int ext4_da_reserve_space(struct inode *inode, int nrblocks) * worse case is one extent per block */ repeat: - spin_lock(&EXT4_I(inode)->i_block_reservation_lock); - total = EXT4_I(inode)->i_reserved_data_blocks + nrblocks; - mdblocks = ext4_calc_metadata_amount(inode, total); - BUG_ON(mdblocks < EXT4_I(inode)->i_reserved_meta_blocks); + spin_lock(&ei->i_block_reservation_lock); + md_needed = ext4_calc_metadata_amount(inode, lblock); + spin_unlock(&ei->i_block_reservation_lock); - md_needed = mdblocks - EXT4_I(inode)->i_reserved_meta_blocks; - total = md_needed + nrblocks; + /* + * We will charge metadata quota at writeout time; this saves + * us from metadata over-estimation, though we may go over by + * a small amount in the end. Here we just reserve for data. + */ + if (DQUOT_RESERVE_BLOCK(inode, 1)) + return -EDQUOT; - if (ext4_claim_free_blocks(sbi, total)) { - spin_unlock(&EXT4_I(inode)->i_block_reservation_lock); + /* + * We do still charge estimated metadata to the sb though; + * we cannot afford to run out of free blocks. + */ + if (ext4_claim_free_blocks(sbi, md_needed + 1)) { + DQUOT_RELEASE_RESERVATION_BLOCK(inode, 1); if (ext4_should_retry_alloc(inode->i_sb, &retries)) { yield(); goto repeat; } return -ENOSPC; } - EXT4_I(inode)->i_reserved_data_blocks += nrblocks; - EXT4_I(inode)->i_reserved_meta_blocks = mdblocks; + spin_lock(&ei->i_block_reservation_lock); + ei->i_reserved_data_blocks++; + ei->i_reserved_meta_blocks += md_needed; + spin_unlock(&ei->i_block_reservation_lock); - spin_unlock(&EXT4_I(inode)->i_block_reservation_lock); return 0; /* success */ } static void ext4_da_release_space(struct inode *inode, int to_free) { struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); - int total, mdb, mdb_free, release; + struct ext4_inode_info *ei = EXT4_I(inode); if (!to_free) return; /* Nothing to release, exit */ spin_lock(&EXT4_I(inode)->i_block_reservation_lock); - if (!EXT4_I(inode)->i_reserved_data_blocks) { + if (unlikely(to_free > ei->i_reserved_data_blocks)) { /* - * if there is no reserved blocks, but we try to free some - * then the counter is messed up somewhere. - * but since this function is called from invalidate - * page, it's harmless to return without any action + * if there aren't enough reserved blocks, then the + * counter is messed up somewhere. Since this + * function is called from invalidate page, it's + * harmless to return without any action. */ - printk(KERN_INFO "ext4 delalloc try to release %d reserved " - "blocks for inode %lu, but there is no reserved " - "data blocks\n", to_free, inode->i_ino); - spin_unlock(&EXT4_I(inode)->i_block_reservation_lock); - return; + ext4_msg(inode->i_sb, KERN_NOTICE, "ext4_da_release_space: " + "ino %lu, to_free %d with only %d reserved " + "data blocks\n", inode->i_ino, to_free, + ei->i_reserved_data_blocks); + WARN_ON(1); + to_free = ei->i_reserved_data_blocks; } + ei->i_reserved_data_blocks -= to_free; - /* recalculate the number of metablocks still need to be reserved */ - total = EXT4_I(inode)->i_reserved_data_blocks - to_free; - mdb = ext4_calc_metadata_amount(inode, total); - - /* figure out how many metablocks to release */ - BUG_ON(mdb > EXT4_I(inode)->i_reserved_meta_blocks); - mdb_free = EXT4_I(inode)->i_reserved_meta_blocks - mdb; - - release = to_free + mdb_free; - - /* update fs dirty blocks counter for truncate case */ - percpu_counter_sub(&sbi->s_dirtyblocks_counter, release); + if (ei->i_reserved_data_blocks == 0) { + /* + * We can release all of the reserved metadata blocks + * only when we have written all of the delayed + * allocation blocks. + */ + percpu_counter_sub(&sbi->s_dirtyblocks_counter, + ei->i_reserved_meta_blocks); + ei->i_reserved_meta_blocks = 0; + ei->i_da_metadata_calc_len = 0; + } - /* update per-inode reservations */ - BUG_ON(to_free > EXT4_I(inode)->i_reserved_data_blocks); - EXT4_I(inode)->i_reserved_data_blocks -= to_free; + /* update fs dirty data blocks counter */ + percpu_counter_sub(&sbi->s_dirtyblocks_counter, to_free); - BUG_ON(mdb > EXT4_I(inode)->i_reserved_meta_blocks); - EXT4_I(inode)->i_reserved_meta_blocks = mdb; spin_unlock(&EXT4_I(inode)->i_block_reservation_lock); + + DQUOT_RELEASE_RESERVATION_BLOCK(inode, to_free); } static void ext4_da_page_release_reservation(struct page *page, @@ -2220,10 +2272,10 @@ static int mpage_da_map_blocks(struct mpage_da_data *mpd) * variables are updated after the blocks have been allocated. */ new.b_state = 0; - get_blocks_flags = (EXT4_GET_BLOCKS_CREATE | - EXT4_GET_BLOCKS_DELALLOC_RESERVE); + get_blocks_flags = EXT4_GET_BLOCKS_CREATE; if (mpd->b_state & (1 << BH_Delay)) - get_blocks_flags |= EXT4_GET_BLOCKS_UPDATE_RESERVE_SPACE; + get_blocks_flags |= EXT4_GET_BLOCKS_DELALLOC_RESERVE; + blks = ext4_get_blocks(handle, mpd->inode, next, max_blocks, &new, get_blocks_flags); if (blks < 0) { @@ -2521,7 +2573,7 @@ static int ext4_da_get_block_prep(struct inode *inode, sector_t iblock, * XXX: __block_prepare_write() unmaps passed block, * is it OK? */ - ret = ext4_da_reserve_space(inode, 1); + ret = ext4_da_reserve_space(inode, iblock); if (ret) /* not enough space to reserve */ return ret; @@ -4872,6 +4924,9 @@ struct inode *ext4_iget(struct super_block *sb, unsigned long ino) ((__u64)le16_to_cpu(raw_inode->i_file_acl_high)) << 32; inode->i_size = ext4_isize(raw_inode); ei->i_disksize = inode->i_size; +#ifdef CONFIG_QUOTA + ei->i_reserved_quota = 0; +#endif inode->i_generation = le32_to_cpu(raw_inode->i_generation); ei->i_block_group = iloc.block_group; ei->i_last_alloc_group = ~0; diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index a050e96..55752fa 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -3026,9 +3026,6 @@ ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac, if (!(ac->ac_flags & EXT4_MB_DELALLOC_RESERVED)) /* release all the reserved blocks if non delalloc */ percpu_counter_sub(&sbi->s_dirtyblocks_counter, reserv_blks); - else - percpu_counter_sub(&sbi->s_dirtyblocks_counter, - ac->ac_b_ex.fe_len); if (sbi->s_log_groups_per_flex) { ext4_group_t flex_group = ext4_flex_group(sbi, @@ -4498,7 +4495,7 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle, struct ext4_sb_info *sbi; struct super_block *sb; ext4_fsblk_t block = 0; - unsigned int inquota; + unsigned int inquota = 0; unsigned int reserv_blks = 0; sb = ar->inode->i_sb; @@ -4516,9 +4513,17 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle, (unsigned long long) ar->pleft, (unsigned long long) ar->pright); - if (!EXT4_I(ar->inode)->i_delalloc_reserved_flag) { - /* - * With delalloc we already reserved the blocks + /* + * For delayed allocation, we could skip the ENOSPC and + * EDQUOT check, as blocks and quotas have been already + * reserved when data being copied into pagecache. + */ + if (EXT4_I(ar->inode)->i_delalloc_reserved_flag) + ar->flags |= EXT4_MB_DELALLOC_RESERVED; + else { + /* Without delayed allocation we need to verify + * there is enough free blocks to do block allocation + * and verify allocation doesn't exceed the quota limits. */ while (ar->len && ext4_claim_free_blocks(sbi, ar->len)) { /* let others to free the space */ @@ -4530,19 +4535,16 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle, return 0; } reserv_blks = ar->len; + while (ar->len && DQUOT_ALLOC_BLOCK(ar->inode, ar->len)) { + ar->flags |= EXT4_MB_HINT_NOPREALLOC; + ar->len--; + } + inquota = ar->len; + if (ar->len == 0) { + *errp = -EDQUOT; + goto out3; + } } - while (ar->len && DQUOT_ALLOC_BLOCK(ar->inode, ar->len)) { - ar->flags |= EXT4_MB_HINT_NOPREALLOC; - ar->len--; - } - if (ar->len == 0) { - *errp = -EDQUOT; - goto out3; - } - inquota = ar->len; - - if (EXT4_I(ar->inode)->i_delalloc_reserved_flag) - ar->flags |= EXT4_MB_DELALLOC_RESERVED; ac = kmem_cache_alloc(ext4_ac_cachep, GFP_NOFS); if (!ac) { @@ -4609,7 +4611,7 @@ repeat: out2: kmem_cache_free(ext4_ac_cachep, ac); out1: - if (ar->len < inquota) + if (inquota && ar->len < inquota) DQUOT_FREE_BLOCK(ar->inode, inquota - ar->len); out3: if (!ar->len) { diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 2508725..e631779 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -742,8 +742,12 @@ static struct inode *ext4_alloc_inode(struct super_block *sb) ei->i_reserved_data_blocks = 0; ei->i_reserved_meta_blocks = 0; ei->i_allocated_meta_blocks = 0; + ei->i_da_metadata_calc_len = 0; ei->i_delalloc_reserved_flag = 0; spin_lock_init(&(ei->i_block_reservation_lock)); +#ifdef CONFIG_QUOTA + ei->i_reserved_quota = 0; +#endif INIT_LIST_HEAD(&ei->i_aio_dio_complete_list); spin_lock_init(&ei->i_completed_io_lock); ei->cur_aio_dio = NULL; @@ -1045,6 +1049,12 @@ static struct dquot_operations ext4_quota_operations = { .initialize = dquot_initialize, .drop = dquot_drop, .alloc_space = dquot_alloc_space, + .reserve_space = dquot_reserve_space, + .claim_space = dquot_claim_space, + .release_rsv = dquot_release_reserved_space, +#ifdef CONFIG_QUOTA + .get_reserved_space = ext4_get_reserved_space, +#endif .alloc_inode = dquot_alloc_inode, .free_space = dquot_free_space, .free_inode = dquot_free_inode, @@ -3958,7 +3968,6 @@ static ssize_t ext4_quota_write(struct super_block *sb, int type, int err = 0; int offset = off & (sb->s_blocksize - 1); int tocopy; - int journal_quota = EXT4_SB(sb)->s_qf_names[type] != NULL; size_t towrite = len; struct buffer_head *bh; handle_t *handle = journal_current_handle(); @@ -3976,24 +3985,16 @@ static ssize_t ext4_quota_write(struct super_block *sb, int type, bh = ext4_bread(handle, inode, blk, 1, &err); if (!bh) goto out; - if (journal_quota) { - err = ext4_journal_get_write_access(handle, bh); - if (err) { - brelse(bh); - goto out; - } + err = ext4_journal_get_write_access(handle, bh); + if (err) { + brelse(bh); + goto out; } lock_buffer(bh); memcpy(bh->b_data+offset, data, tocopy); flush_dcache_page(bh->b_page); unlock_buffer(bh); - if (journal_quota) - err = ext4_handle_dirty_metadata(handle, NULL, bh); - else { - /* Always do at least ordered writes for quotas */ - err = ext4_jbd2_file_inode(handle, inode); - mark_buffer_dirty(bh); - } + err = ext4_handle_dirty_metadata(handle, NULL, bh); brelse(bh); if (err) goto out; @@ -4031,7 +4032,7 @@ static struct file_system_type ext4_fs_type = { .get_sb = ext4_get_sb, .kill_sb = kill_block_super, .fs_flags = FS_REQUIRES_DEV|FS_HAS_FALLOCATE|FS_HAS_FIEMAP - |FS_HAS_FREEZE|FS_HAS_TRYTOFREE, + |FS_HAS_FREEZE|FS_HAS_TRYTOFREE|FS_HAS_GETRESV, }; #ifdef CONFIG_EXT4DEV_COMPAT @@ -4050,7 +4051,8 @@ static struct file_system_type ext4dev_fs_type = { .name = "ext4dev", .get_sb = ext4dev_get_sb, .kill_sb = kill_block_super, - .fs_flags = FS_REQUIRES_DEV|FS_HAS_FALLOCATE|FS_HAS_FIEMAP|FS_HAS_FREEZE, + .fs_flags = FS_REQUIRES_DEV|FS_HAS_FALLOCATE|FS_HAS_FIEMAP + |FS_HAS_FREEZE|FS_HAS_TRYTOFREE|FS_HAS_GETRESV, }; MODULE_ALIAS("ext4dev"); #endif